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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
5e086d4b-5841-4cf1-91be-baccd421bafb | on-the-convergence-of-policy-gradient-methods | 2210.08857 | null | https://arxiv.org/abs/2210.08857v1 | https://arxiv.org/pdf/2210.08857v1.pdf | On the convergence of policy gradient methods to Nash equilibria in general stochastic games | Learning in stochastic games is a notoriously difficult problem because, in addition to each other's strategic decisions, the players must also contend with the fact that the game itself evolves over time, possibly in a very complicated manner. Because of this, the convergence properties of popular learning algorithms - like policy gradient and its variants - are poorly understood, except in specific classes of games (such as potential or two-player, zero-sum games). In view of this, we examine the long-run behavior of policy gradient methods with respect to Nash equilibrium policies that are second-order stationary (SOS) in a sense similar to the type of sufficiency conditions used in optimization. Our first result is that SOS policies are locally attracting with high probability, and we show that policy gradient trajectories with gradient estimates provided by the REINFORCE algorithm achieve an $\mathcal{O}(1/\sqrt{n})$ distance-squared convergence rate if the method's step-size is chosen appropriately. Subsequently, specializing to the class of deterministic Nash policies, we show that this rate can be improved dramatically and, in fact, policy gradient methods converge within a finite number of iterations in that case. | ['Emmanouil-Vasileios Vlatakis-Gkaragkounis', 'Panayotis Mertikopoulos', 'Kyriakos Lotidis', 'Angeliki Giannou'] | 2022-10-17 | null | null | null | null | ['policy-gradient-methods'] | ['methodology'] | [-1.50914818e-01 1.35146931e-01 -2.14980721e-01 2.48502359e-01
-5.89051902e-01 -9.16143298e-01 2.59671450e-01 8.49030092e-02
-8.68149638e-01 1.26751709e+00 7.62590487e-03 -6.70842409e-01
-5.01696050e-01 -7.82478094e-01 -6.67839587e-01 -9.54347849e-01
-2.48831332e-01 4.37061369e-01 2.69115716e-01 -5.35107434e-01
3.41875464e-01 3.42871308e-01 -1.28028548e+00 -6.88300610e-01
1.19376111e+00 7.59032607e-01 -4.74280678e-02 9.31211948e-01
-1.29610866e-01 8.70637476e-01 -4.82459873e-01 -3.99192244e-01
4.45085675e-01 -7.50638902e-01 -6.77204788e-01 1.64990332e-02
2.56539714e-02 -1.04193985e-01 -2.66267926e-01 1.57598877e+00
6.06391251e-01 5.30622482e-01 6.84366524e-01 -1.04408169e+00
-1.66985482e-01 5.24009764e-01 -5.18562257e-01 3.42561662e-01
1.49846360e-01 1.80922344e-01 9.76801217e-01 -3.82908285e-01
4.44682479e-01 1.05519450e+00 5.96124649e-01 7.59961545e-01
-1.28296471e+00 -5.14353752e-01 5.29658973e-01 -2.53467470e-01
-9.63635206e-01 -1.47799700e-01 4.98374194e-01 -4.76615101e-01
4.27968562e-01 3.37125689e-01 7.83443570e-01 7.44960964e-01
1.81522250e-01 7.57992268e-01 1.29169893e+00 -3.79998952e-01
5.83757818e-01 1.06910281e-01 -2.03164548e-01 7.52810836e-01
3.40326905e-01 3.10375988e-01 -6.49863482e-02 -2.64502823e-01
8.76851797e-01 -2.41944000e-01 -4.11879361e-01 -5.00720441e-01
-7.70629287e-01 1.15818155e+00 8.73688683e-02 3.48279625e-01
-4.92822796e-01 3.77123475e-01 1.51953027e-01 5.97153485e-01
5.37631035e-01 5.86873651e-01 -2.70521820e-01 -4.72801119e-01
-8.35712373e-01 5.75897694e-01 1.20082831e+00 5.08224905e-01
4.54959422e-01 4.23922300e-01 -7.04075545e-02 5.39455295e-01
9.71291214e-03 6.39447570e-01 2.65019953e-01 -1.02809572e+00
3.77520502e-01 6.91482201e-02 6.68778658e-01 -9.34832275e-01
-3.13915193e-01 -7.77465820e-01 -7.02151775e-01 6.15170121e-01
1.06957662e+00 -6.85393214e-01 -2.84150600e-01 2.17395687e+00
3.39959323e-01 6.14705356e-03 5.28648915e-03 8.57804000e-01
-2.17833325e-01 6.36789680e-01 -8.45239833e-02 -7.31492639e-01
7.64980853e-01 -4.14665788e-01 -4.23388988e-01 -7.49473572e-02
4.29039657e-01 -4.36280608e-01 1.17147326e+00 3.50794911e-01
-1.26759100e+00 1.64858326e-01 -6.70166135e-01 6.89317644e-01
5.02112098e-02 -3.56392801e-01 4.43875372e-01 7.55601048e-01
-1.20951772e+00 8.04911315e-01 -6.28310800e-01 -3.12942713e-01
2.36371875e-01 4.21238273e-01 2.02102035e-01 3.62807781e-01
-1.13882732e+00 8.40572834e-01 2.23766834e-01 1.92783251e-02
-1.04673028e+00 -6.87483609e-01 -3.79499286e-01 1.69497088e-01
8.63037467e-01 -3.34719598e-01 1.54723227e+00 -1.25638056e+00
-1.77635336e+00 3.69794607e-01 -9.98447090e-02 -3.81383598e-01
1.15958965e+00 6.31037280e-02 1.43852323e-01 7.45156556e-02
1.86154068e-01 -2.51900125e-02 8.44672024e-01 -1.10218728e+00
-8.74179363e-01 -3.08507442e-01 5.83250225e-01 5.43199658e-01
-3.56316954e-01 -1.98024139e-02 2.50316352e-01 -3.17584693e-01
-3.42108667e-01 -9.83919442e-01 -8.73613477e-01 -8.00934955e-02
-9.92329493e-02 -2.95323223e-01 4.13116723e-01 -1.49142325e-01
1.20336664e+00 -1.87776053e+00 2.87486583e-01 4.41071808e-01
3.02623630e-01 2.05967963e-01 1.89970121e-01 4.38239902e-01
6.46392480e-02 2.22858593e-01 -1.86077029e-01 1.28361017e-01
7.90761113e-02 1.18533429e-02 -2.95554280e-01 6.89753830e-01
-2.93574750e-01 5.84589601e-01 -1.29451394e+00 -2.04625174e-01
-2.54208818e-02 -1.97989076e-01 -6.96821153e-01 -2.51088560e-01
-3.65860820e-01 3.98240268e-01 -8.68471801e-01 2.07599178e-02
1.82346731e-01 -2.95333743e-01 2.52512395e-01 7.11789012e-01
-3.52506220e-01 -6.78330287e-02 -1.54237211e+00 1.06845331e+00
-3.40302020e-01 5.41478872e-01 4.72750992e-01 -1.22201180e+00
3.34375322e-01 3.76772881e-01 8.18870962e-01 -2.05093443e-01
2.46368408e-01 5.01751423e-01 3.44139822e-02 -2.46584713e-01
3.37110996e-01 -5.75043201e-01 1.34328946e-01 6.24972999e-01
-2.35165343e-01 -7.75356218e-02 4.41060841e-01 2.76823759e-01
1.00674736e+00 -2.33427271e-01 2.91576505e-01 -7.60346234e-01
3.98427248e-01 1.02535702e-01 5.77576637e-01 1.07770157e+00
-3.03688675e-01 2.83716004e-02 1.10876811e+00 -6.13457412e-02
-1.03703105e+00 -9.38448191e-01 2.51462609e-01 1.08068538e+00
1.38681129e-01 -6.86846673e-02 -7.47359455e-01 -4.93065298e-01
1.99670687e-01 5.57380497e-01 -7.46713817e-01 -2.52881080e-01
-4.66421932e-01 -7.24988520e-01 2.77593255e-01 2.09804147e-01
4.39727724e-01 -9.49482501e-01 -5.88498354e-01 5.77187121e-01
2.51680136e-01 -4.48222727e-01 -6.27223432e-01 1.07595049e-01
-8.89242232e-01 -1.30034256e+00 -1.09149039e+00 -4.32626784e-01
6.18373454e-01 9.30116177e-02 9.16763723e-01 -1.30080208e-01
2.24936098e-01 6.38940394e-01 -7.35197635e-03 -5.17428875e-01
-4.04676765e-01 7.01491013e-02 3.24446261e-01 -5.91532663e-02
-1.14705220e-01 -5.53246081e-01 -6.51741862e-01 2.03167081e-01
-8.36189032e-01 -3.54958564e-01 7.73892254e-02 8.41771662e-01
2.25157470e-01 2.05091581e-01 5.52958846e-01 -8.30071926e-01
1.12657094e+00 -5.19597232e-01 -1.33537114e+00 1.10452786e-01
-5.91256380e-01 1.43717647e-01 8.62245142e-01 -7.21880436e-01
-7.82841504e-01 -1.61448643e-01 2.19809055e-01 -3.44745696e-01
3.46597105e-01 6.12760484e-01 2.96431780e-01 -2.17427328e-01
8.27317655e-01 4.42069739e-01 3.52480918e-01 -1.42224312e-01
1.99302614e-01 2.47479036e-01 2.13766113e-01 -9.10131037e-01
8.37351859e-01 3.92108351e-01 1.58493876e-01 -8.05304527e-01
-7.02420592e-01 -3.96222323e-02 5.03105596e-02 -5.84916592e-01
5.83138585e-01 -4.79050934e-01 -1.34322393e+00 5.00202894e-01
-6.89676106e-01 -7.41920531e-01 -6.03877842e-01 5.35436094e-01
-8.45168650e-01 7.72246793e-02 -6.05778158e-01 -1.51971507e+00
1.21311016e-01 -9.44341063e-01 1.41996250e-01 6.55340433e-01
1.66628063e-01 -1.29114771e+00 2.03352913e-01 -2.25262061e-01
6.59565091e-01 1.95925593e-01 7.73953915e-01 -4.74597782e-01
-2.58301377e-01 1.66590849e-03 1.48634717e-01 3.18283021e-01
-5.64014353e-02 -2.63041127e-02 -2.54480809e-01 -4.51823771e-01
1.98256567e-01 -1.50963575e-01 6.59766018e-01 8.85521352e-01
6.60613716e-01 -5.83510697e-01 -5.99071234e-02 1.59887224e-01
1.58597541e+00 3.71107876e-01 1.82914019e-01 3.01317871e-01
4.05063599e-01 5.26360691e-01 3.21181685e-01 6.24889255e-01
1.08402237e-01 2.50250787e-01 5.17977417e-01 5.53686768e-02
6.34636879e-01 -1.93641290e-01 4.44255531e-01 3.84942770e-01
-3.06285799e-01 -5.75965568e-02 -6.69735909e-01 5.64343333e-01
-2.14406180e+00 -1.37158155e+00 3.79565582e-02 2.55745482e+00
1.02360761e+00 2.04879403e-01 6.53073549e-01 -7.07578519e-03
8.03485215e-01 9.93627608e-02 -9.08972025e-01 -5.02368152e-01
-1.88545421e-01 2.90010810e-01 9.20240879e-01 9.37528014e-01
-7.37999201e-01 9.83353674e-01 6.63910627e+00 9.88512039e-01
-1.25501823e+00 6.62998036e-02 5.79067230e-01 -4.43064272e-01
-3.84295285e-01 5.09126531e-03 -4.16614264e-01 6.28232896e-01
6.85622692e-01 -6.51998580e-01 8.27331901e-01 9.34426546e-01
6.29193723e-01 -3.96840721e-01 -4.95924711e-01 6.54392242e-01
-5.31980217e-01 -1.24264789e+00 -6.70109987e-01 4.40588385e-01
1.09363818e+00 -1.76846310e-01 1.90336019e-01 3.10261190e-01
1.22053301e+00 -9.11217034e-01 7.79383302e-01 2.11284339e-01
5.37466407e-01 -1.14324355e+00 3.80842537e-01 5.18768013e-01
-9.34477329e-01 -5.17323971e-01 -3.93959016e-01 -3.49088877e-01
1.03885263e-01 6.32421076e-01 -2.25843102e-01 2.64454693e-01
3.81840795e-01 1.87674507e-01 1.46394327e-01 1.29530525e+00
-2.13527828e-01 8.45120072e-01 -5.88742852e-01 -6.00216210e-01
7.22355962e-01 -6.18792832e-01 8.05263281e-01 6.51911855e-01
2.71907568e-01 4.47874486e-01 2.47960895e-01 6.53847694e-01
-2.16361620e-02 2.92973578e-01 -6.47703350e-01 -7.12224394e-02
3.14091712e-01 9.87880409e-01 -7.71440446e-01 -1.42570645e-01
-1.58982232e-01 6.57324195e-01 3.53307724e-01 6.69446826e-01
-6.64723575e-01 -4.50060159e-01 1.00605261e+00 8.52474794e-02
3.35047126e-01 -3.60080779e-01 6.02674335e-02 -1.15247464e+00
1.51685655e-01 -6.90135241e-01 3.02684873e-01 -1.99130446e-01
-1.13615561e+00 1.68484494e-01 -2.16789357e-02 -8.50021362e-01
-5.31938195e-01 -5.49301982e-01 -6.92378342e-01 6.91713572e-01
-1.07716167e+00 -3.01092356e-01 4.76187795e-01 7.40670800e-01
2.37280697e-01 -1.58672124e-01 3.34276736e-01 3.14486511e-02
-4.20206070e-01 3.21154624e-01 8.30170751e-01 3.45210731e-02
1.14087395e-01 -1.57873440e+00 -2.90111184e-01 8.74597967e-01
-4.37690839e-02 2.53284186e-01 1.13605976e+00 -3.60219777e-01
-1.38667202e+00 -6.01448178e-01 5.01366496e-01 -1.08626613e-03
1.10077834e+00 -1.83500394e-01 -5.17881036e-01 3.14704180e-01
-8.36019367e-02 -4.66333441e-02 7.86625147e-02 1.48935109e-01
2.33838618e-01 -2.44528838e-02 -1.01525104e+00 1.05688775e+00
9.06381011e-01 -2.89781898e-01 1.84488408e-02 3.44191492e-01
3.51135701e-01 -4.22638863e-01 -4.96269822e-01 -9.55971181e-02
3.75575989e-01 -8.79712284e-01 5.44258356e-01 -1.11475468e+00
1.42776608e-01 -1.62372604e-01 5.55104502e-02 -1.52750468e+00
-1.95620105e-01 -1.29461420e+00 1.69106960e-01 7.26212740e-01
3.68573040e-01 -1.00397336e+00 1.16161168e+00 5.65854549e-01
2.31349215e-01 -8.90985727e-01 -1.24605489e+00 -1.09568536e+00
6.70020998e-01 -2.88581192e-01 6.86415359e-02 7.18318582e-01
2.21575245e-01 1.00826383e-01 -4.39780444e-01 -1.32993773e-01
5.58335185e-01 -1.16324194e-01 4.58443224e-01 -1.12114906e+00
-5.09117663e-01 -1.04379177e+00 -1.36708140e-01 -1.06081021e+00
2.51745313e-01 -5.54192781e-01 1.16444089e-01 -1.21137369e+00
-2.88574602e-02 -7.60628223e-01 -3.74406785e-01 9.14313644e-02
-4.05186802e-01 -3.26981336e-01 4.18857515e-01 8.36242288e-02
-5.19535840e-01 6.85973108e-01 1.29442644e+00 2.42764186e-02
-4.93940026e-01 5.87716639e-01 -8.48106444e-01 6.47952318e-01
8.41358602e-01 -6.02243066e-01 -6.22612298e-01 -1.41235953e-02
5.09544551e-01 5.06605148e-01 1.43671185e-01 -7.89109051e-01
3.41531128e-01 -7.64929712e-01 -2.54270583e-01 9.23003554e-02
-2.96189953e-02 -4.62594718e-01 -1.99231416e-01 7.65661240e-01
-6.07883871e-01 2.01201901e-01 -7.89216980e-02 7.39258945e-01
1.79124326e-01 -5.93734741e-01 8.87398839e-01 -2.48387709e-01
-1.46591023e-01 4.10840273e-01 -9.75890875e-01 6.20971918e-01
9.99138892e-01 -1.54118195e-01 -2.16504410e-02 -1.10839319e+00
-6.15023553e-01 4.25710917e-01 3.45731080e-01 -2.20681041e-01
1.74404919e-01 -1.08236313e+00 -6.92059100e-01 -4.92631674e-01
-5.24411261e-01 -1.72942549e-01 8.95931646e-02 1.04491198e+00
-3.24406743e-01 7.02891275e-02 6.99012130e-02 -2.02765703e-01
-7.89065361e-01 3.33725035e-01 7.73634732e-01 -4.95757997e-01
-3.37977320e-01 8.39882135e-01 8.35302398e-02 -1.81759834e-01
2.04054013e-01 6.77514961e-03 1.07824050e-01 7.54184797e-02
4.24183607e-01 4.43727046e-01 -4.18554574e-01 -1.02466859e-01
-7.54307806e-02 2.80089825e-01 1.14023715e-01 -7.62572348e-01
1.21407402e+00 1.88347381e-02 -1.99382100e-02 3.20525497e-01
8.72932911e-01 2.15830997e-01 -1.66211283e+00 -2.57295877e-01
8.81453231e-02 -2.93384880e-01 -2.71325976e-01 -4.59233552e-01
-1.21939003e+00 5.54389179e-01 4.23466027e-01 6.26633704e-01
8.35819185e-01 -3.15732569e-01 1.93541065e-01 3.26627314e-01
4.25518245e-01 -1.39003778e+00 4.16412018e-02 7.07900465e-01
4.66141492e-01 -9.47561026e-01 -2.08176225e-01 5.20159937e-02
-5.65572202e-01 9.95924711e-01 3.13634932e-01 -5.59993148e-01
6.86313868e-01 1.62377544e-02 -9.40449014e-02 8.46285895e-02
-6.90179110e-01 -4.43739712e-01 -2.04027474e-01 5.64889729e-01
1.98726937e-01 2.10775621e-02 -7.37800181e-01 1.70594722e-01
-1.60980225e-01 1.14679271e-02 7.40858912e-01 9.24989402e-01
-7.26640999e-01 -1.01548111e+00 -2.30097041e-01 5.67882359e-01
-7.20896959e-01 3.33179384e-02 -1.63641229e-01 8.85843039e-01
-6.14867806e-01 1.04198110e+00 -1.32366329e-01 7.56371170e-02
2.04833150e-01 -1.94216102e-01 5.46209931e-01 -2.80220777e-01
-6.27708673e-01 8.95606056e-02 -9.18216109e-02 -4.52861935e-01
-1.71947777e-01 -7.23674297e-01 -1.05143869e+00 -6.59107685e-01
-2.98088551e-01 6.29222453e-01 4.43789124e-01 1.16452241e+00
-6.50674775e-02 1.45272359e-01 8.05664480e-01 -5.94432175e-01
-1.24374008e+00 -6.37349784e-01 -8.91269088e-01 9.33142230e-02
2.51229554e-01 -5.52641451e-01 -4.91349697e-01 -5.26923239e-01] | [4.2360520362854, 2.6311147212982178] |
b88d4b47-82f9-4d05-8faf-66b9ccc90e28 | duo-segnet-adversarial-dual-views-for-semi | 2108.11154 | null | https://arxiv.org/abs/2108.11154v1 | https://arxiv.org/pdf/2108.11154v1.pdf | Duo-SegNet: Adversarial Dual-Views for Semi-Supervised Medical Image Segmentation | Segmentation of images is a long-standing challenge in medical AI. This is mainly due to the fact that training a neural network to perform image segmentation requires a significant number of pixel-level annotated data, which is often unavailable. To address this issue, we propose a semi-supervised image segmentation technique based on the concept of multi-view learning. In contrast to the previous art, we introduce an adversarial form of dual-view training and employ a critic to formulate the learning problem in multi-view training as a min-max problem. Thorough quantitative and qualitative evaluations on several datasets indicate that our proposed method outperforms state-of-the-art medical image segmentation algorithms consistently and comfortably. The code is publicly available at https://github.com/himashi92/Duo-SegNet | ['Mehrtash Harandi', 'Gary Egan', 'Zhaolin Chen', 'Himashi Peiris'] | 2021-08-25 | null | null | null | null | ['semi-supervised-medical-image-segmentation', 'multi-view-learning'] | ['computer-vision', 'computer-vision'] | [ 3.23530346e-01 2.86984056e-01 -3.15211266e-01 -3.97268206e-01
-1.02341878e+00 -5.72140992e-01 2.68528331e-02 -2.96456486e-01
-5.44850469e-01 6.46615267e-01 -1.54753432e-01 -2.56639183e-01
3.43343168e-01 -6.69761479e-01 -7.56207943e-01 -8.48327100e-01
3.92991304e-01 3.62205654e-01 7.90674463e-02 -2.08781660e-01
3.59536409e-02 1.45862341e-01 -7.85448670e-01 1.68852136e-01
8.90182972e-01 7.77776897e-01 8.94794241e-02 6.36913121e-01
1.79436609e-01 4.60324198e-01 -3.91956866e-01 -5.23155451e-01
3.78191799e-01 -5.32552481e-01 -1.09034240e+00 4.52837676e-01
2.33091876e-01 -4.21343446e-01 2.96758078e-02 1.26041555e+00
5.89899540e-01 -1.21041648e-01 3.28070611e-01 -1.13600683e+00
-4.48571682e-01 3.66606742e-01 -8.61171126e-01 2.98947513e-01
1.27071634e-01 -5.28784953e-02 8.64683688e-01 -6.32127583e-01
6.41592324e-01 8.03479016e-01 5.75080752e-01 8.67731333e-01
-1.14118612e+00 -4.82668221e-01 2.47252405e-01 -1.78411454e-01
-1.12373936e+00 -2.08118901e-01 9.63650167e-01 -3.92525792e-01
3.42955977e-01 2.12049678e-01 7.38720536e-01 1.11027515e+00
2.90400386e-01 1.29151440e+00 1.32914424e+00 -1.50888264e-01
1.68648988e-01 1.18445465e-02 -1.22140780e-01 9.38637495e-01
-3.33259851e-02 -1.24168873e-01 -1.07962510e-03 2.28409860e-02
9.34721708e-01 1.47814855e-01 -2.97907382e-01 -4.63991195e-01
-9.57418561e-01 9.28879976e-01 5.14208972e-01 1.38579011e-01
-3.66013587e-01 6.57669380e-02 4.96206254e-01 -5.19361086e-02
7.59826243e-01 2.92590380e-01 -5.74789941e-01 1.25085011e-01
-9.43452775e-01 6.03157952e-02 5.62034369e-01 6.64356589e-01
5.66053748e-01 1.74351502e-02 2.99264103e-01 7.64180720e-01
2.90211588e-01 2.79961556e-01 4.63626564e-01 -1.11872673e+00
4.30640072e-01 6.60378754e-01 -1.57337978e-01 -6.13103628e-01
-2.85071641e-01 -5.56131363e-01 -9.99359846e-01 3.49090636e-01
5.20852387e-01 -3.31145495e-01 -1.11254203e+00 1.46717715e+00
5.48471451e-01 1.57070071e-01 1.44785926e-01 8.97061825e-01
8.34894896e-01 4.16974157e-01 1.33864060e-02 -2.05098003e-01
1.28570282e+00 -1.35592365e+00 -6.98325038e-01 -3.59415859e-01
4.55390513e-01 -5.13230443e-01 9.52297866e-01 5.36699772e-01
-1.46793032e+00 -3.47580314e-01 -9.55005825e-01 -5.41407848e-03
-3.45132113e-01 -3.17207389e-02 6.63619518e-01 6.43440247e-01
-7.89828777e-01 4.12904024e-01 -1.13336742e+00 -2.41879933e-02
8.98096979e-01 3.49612534e-01 -4.12701190e-01 -9.02591944e-02
-1.04959953e+00 5.47050178e-01 5.10747731e-01 2.67152190e-01
-8.10571492e-01 -6.01665199e-01 -8.73288631e-01 -3.25574517e-01
6.04575336e-01 -9.59803522e-01 1.30963945e+00 -1.30220723e+00
-1.40860593e+00 1.20960820e+00 1.45671470e-02 -2.54943162e-01
9.22613621e-01 -3.17767173e-01 -3.76792066e-02 4.25685078e-01
2.60668516e-01 7.35510349e-01 6.99978173e-01 -1.70207942e+00
-5.10619938e-01 -5.81952155e-01 2.42449492e-01 3.23952466e-01
6.94124773e-02 -1.64442927e-01 -8.24613929e-01 -7.85405755e-01
8.31335261e-02 -1.09626579e+00 -6.29409969e-01 -2.98207086e-02
-6.73426986e-01 1.82252228e-02 7.96468496e-01 -8.07720006e-01
9.60604012e-01 -1.97737443e+00 4.21733707e-01 -7.85534903e-02
2.97902822e-01 2.73387879e-01 3.45975459e-01 -2.52153128e-02
-6.42008260e-02 2.37012461e-01 -7.07296729e-01 -4.05435085e-01
-5.11810482e-01 2.33451292e-01 -5.92394872e-03 6.46274090e-01
-6.34260327e-02 1.09027743e+00 -1.00479221e+00 -7.27450728e-01
2.48412341e-01 4.39188957e-01 -5.06582737e-01 2.84537405e-01
-2.33177304e-01 1.01121569e+00 -8.27125669e-01 8.73883963e-01
5.90158284e-01 -6.16229236e-01 -3.07750590e-02 -1.17902212e-01
1.53395340e-01 -5.08603096e-01 -8.59957635e-01 2.18355465e+00
-3.44668716e-01 9.12019759e-02 2.41163313e-01 -1.37203372e+00
3.18316549e-01 4.52281326e-01 8.50699127e-01 -2.85796911e-01
4.46925581e-01 1.47646755e-01 -3.01205665e-01 -6.23996615e-01
1.57479584e-01 -3.53353918e-01 -9.64602157e-02 2.67577857e-01
-5.29990010e-02 -2.95165598e-01 2.06610993e-01 1.63447708e-01
6.73949659e-01 4.20710802e-01 3.39193761e-01 -1.29551049e-02
7.44613647e-01 5.27365468e-02 8.43342662e-01 3.13170046e-01
-4.13896918e-01 9.23419297e-01 5.82584739e-01 -2.74888575e-01
-8.25609982e-01 -1.02913916e+00 -1.49916902e-01 7.08662391e-01
2.52360582e-01 3.65567654e-02 -1.23656487e+00 -1.02912819e+00
-2.56431490e-01 4.75356489e-01 -7.70543873e-01 1.23032205e-01
-5.90604067e-01 -7.99308360e-01 4.91607010e-01 6.53369606e-01
5.95702231e-01 -9.73884463e-01 -8.55682015e-01 1.29931159e-02
-3.15250874e-01 -1.32352102e+00 -7.22147882e-01 4.39696386e-02
-1.07000971e+00 -1.18133795e+00 -1.19293427e+00 -7.87332058e-01
1.13474107e+00 1.08270943e-01 1.12077796e+00 1.34549454e-01
-4.17750448e-01 4.34316427e-01 -3.08274329e-01 -4.71197635e-01
-3.90716732e-01 1.11904182e-01 -4.03160602e-01 -1.33176185e-02
-9.57469717e-02 -5.13464391e-01 -9.85500634e-01 2.56635904e-01
-1.05436015e+00 4.53417271e-01 6.36544824e-01 9.72765863e-01
1.14269316e+00 -2.40517676e-01 5.62246978e-01 -1.49145603e+00
2.86061287e-01 -3.86300474e-01 -5.92387080e-01 2.18981147e-01
-5.31459212e-01 -3.43448818e-01 5.67605019e-01 -4.42232937e-02
-8.48427892e-01 4.79981393e-01 -3.60071391e-01 -6.47147357e-01
-3.57272565e-01 3.14796954e-01 -9.61817503e-02 -5.33748567e-02
1.84337988e-01 2.77435690e-01 2.24531278e-01 -1.62372157e-01
4.43348050e-01 4.77007598e-01 4.57246214e-01 -3.17016691e-01
5.21309674e-01 8.77695918e-01 -2.06133127e-01 -4.67686892e-01
-1.24204004e+00 -4.98554349e-01 -8.05440187e-01 -4.01822537e-01
1.40520501e+00 -9.54640150e-01 -5.67272842e-01 5.56431174e-01
-9.23931301e-01 -4.88566399e-01 -1.85873210e-01 1.67262748e-01
-9.12369430e-01 3.38515967e-01 -6.40359223e-01 -3.38992029e-01
-5.84335983e-01 -1.56737232e+00 1.14689744e+00 3.72501105e-01
1.18957184e-01 -1.26344264e+00 -1.27125487e-01 1.12662113e+00
3.69892754e-02 7.74271488e-01 6.25717998e-01 -4.74044949e-01
-4.52217340e-01 -1.18418835e-01 -1.78504232e-02 6.38327599e-01
2.27949172e-01 -2.19236523e-01 -8.78189743e-01 -4.91867483e-01
2.00512692e-01 -5.81013739e-01 7.55840242e-01 6.87364161e-01
1.49093390e+00 1.05084879e-02 -3.81786913e-01 9.35994089e-01
1.52298558e+00 2.24758789e-01 4.39137399e-01 2.53617853e-01
8.97235096e-01 4.40149426e-01 5.94156325e-01 2.85949498e-01
4.45530415e-01 3.98306638e-01 5.08024871e-01 -6.66419268e-01
1.51026458e-01 -1.63501605e-01 -1.71003699e-01 7.02705085e-01
-3.53482664e-02 -3.95023823e-01 -1.03070748e+00 5.16396642e-01
-1.80750048e+00 -7.71772265e-01 1.10830784e-01 1.68948972e+00
8.23814631e-01 1.00527495e-01 1.76619112e-01 9.42410156e-02
5.70776284e-01 1.67916924e-01 -8.35243344e-01 -1.86726823e-01
2.21171752e-01 -6.83099106e-02 6.66781783e-01 4.29052174e-01
-1.41779733e+00 1.01724195e+00 5.87050915e+00 6.50729001e-01
-1.30378342e+00 3.10919642e-01 1.28443718e+00 7.11523220e-02
-7.86511227e-02 -4.35657620e-01 -3.85792643e-01 4.22278851e-01
4.13221240e-01 1.93362609e-01 1.86736733e-01 7.76557148e-01
1.83839664e-01 -2.21131623e-01 -8.95463586e-01 9.13647115e-01
2.15654388e-01 -1.25513244e+00 -1.99120328e-01 8.64881203e-02
9.69347239e-01 6.70619830e-02 2.23595336e-01 -1.23448759e-01
1.07896097e-01 -1.00254178e+00 4.06846941e-01 2.08129182e-01
6.65806353e-01 -7.46159196e-01 5.81302643e-01 4.03915226e-01
-8.93795490e-01 1.69041514e-01 7.63781071e-02 4.48811203e-01
4.83826548e-01 4.26834464e-01 -4.06547308e-01 7.50199497e-01
5.79687655e-01 7.88840473e-01 -4.20090288e-01 6.62718058e-01
-2.83160686e-01 5.98212838e-01 2.79309265e-02 6.23067379e-01
3.86296809e-01 -4.20750469e-01 3.50805014e-01 8.88426721e-01
-1.41010478e-01 1.83108330e-01 4.10607904e-01 8.06568861e-01
-1.78023592e-01 5.83489984e-02 -4.99555916e-01 1.21124081e-01
-8.43621269e-02 1.43015337e+00 -1.25228488e+00 -2.80140281e-01
-6.48964226e-01 1.35918093e+00 1.29659057e-01 3.04539204e-01
-9.37061489e-01 6.06979728e-02 1.05845340e-01 -5.89757748e-02
1.72070697e-01 -4.46155779e-02 -5.44173479e-01 -1.21384585e+00
-4.40428294e-02 -9.11692858e-01 4.17843223e-01 -4.98089135e-01
-1.13749826e+00 6.09146655e-01 3.58990170e-02 -1.38875234e+00
-1.87296703e-01 -3.73511672e-01 -5.72456598e-01 5.88942170e-01
-1.55528879e+00 -1.38595152e+00 -2.85048455e-01 5.37464201e-01
7.74296224e-01 -1.51025638e-01 7.26301074e-01 3.95419627e-01
-7.21154034e-01 5.89646161e-01 -9.34453448e-04 4.44785297e-01
4.31619495e-01 -1.44659233e+00 3.87951165e-01 7.62858629e-01
-2.50871666e-02 2.85728455e-01 4.67806995e-01 -5.34078002e-01
-1.32018948e+00 -1.10079920e+00 1.12789892e-01 -3.96387100e-01
4.12084639e-01 -1.40109003e-01 -8.17763746e-01 8.15200925e-01
4.49697405e-01 5.33883512e-01 8.38486731e-01 -4.02866215e-01
1.02606155e-01 4.85916063e-02 -1.32430971e+00 3.31055671e-01
7.39926219e-01 -2.14335859e-01 -3.73623759e-01 4.51587379e-01
6.37158692e-01 -9.64084625e-01 -9.91242766e-01 3.90422761e-01
3.44922900e-01 -9.82877612e-01 1.08986008e+00 -3.47697407e-01
7.08004594e-01 -1.62680790e-01 1.53365120e-01 -1.34577322e+00
1.54849589e-01 -3.78697157e-01 1.22418001e-01 9.53643620e-01
4.08414721e-01 -6.19747162e-01 1.11641276e+00 3.71017545e-01
-1.17183857e-01 -1.46068037e+00 -8.30276966e-01 -5.16924441e-01
3.13649625e-01 5.12242392e-02 -1.35155559e-01 1.04934788e+00
-4.21693504e-01 1.41030282e-01 -2.09566310e-01 2.15452164e-01
8.03454757e-01 2.36787528e-01 4.15179342e-01 -7.99922526e-01
-3.55964243e-01 -1.91004157e-01 -1.15741536e-01 -7.59977221e-01
3.41334432e-01 -1.00137293e+00 9.22035202e-02 -1.65415466e+00
3.15378934e-01 -2.96042204e-01 -2.14603961e-01 3.35433990e-01
-2.93396592e-01 5.03150344e-01 3.42169404e-02 9.07742903e-02
-5.33365786e-01 3.53310138e-01 1.81577444e+00 -7.54954219e-02
-6.56201690e-02 3.42069060e-01 -7.74672627e-01 9.91995692e-01
1.09676850e+00 -4.51664746e-01 -5.21783233e-01 -5.10655463e-01
5.67626730e-02 4.33688819e-01 3.39970082e-01 -7.01672018e-01
2.48660654e-01 -1.71979502e-01 3.63386869e-01 -5.72577357e-01
3.90485138e-01 -8.17010403e-01 -1.34125680e-01 7.45486557e-01
-1.95395038e-01 1.62736610e-01 1.43383980e-01 5.94613016e-01
-3.61062288e-01 -1.66125849e-01 1.02286649e+00 -6.59768462e-01
-5.54052472e-01 4.64931101e-01 -7.54578188e-02 5.25043011e-01
1.30988586e+00 -2.01898709e-01 -5.47913574e-02 -1.26933709e-01
-7.77530074e-01 5.50079584e-01 4.62154180e-01 3.33015919e-01
5.74275970e-01 -8.95793557e-01 -5.53319275e-01 2.32077669e-02
-1.28494143e-01 5.72638512e-01 3.80266279e-01 9.26645517e-01
-5.99821806e-01 4.43940982e-02 -1.62656158e-01 -6.51902020e-01
-1.26048052e+00 5.02823532e-01 5.39864302e-01 -3.41993541e-01
-7.08690047e-01 9.16538477e-01 3.32838297e-01 -4.52377319e-01
7.55055994e-02 -8.28057602e-02 -1.52992204e-01 -1.78464204e-02
1.52656808e-01 2.53089756e-01 -2.32254624e-01 -6.40103936e-01
-2.29300067e-01 6.60870612e-01 -1.58733398e-01 -8.73565227e-02
1.36166716e+00 -7.21734166e-02 -9.44063216e-02 4.57696617e-01
1.26652050e+00 -2.86714852e-01 -1.38831639e+00 -1.21710543e-02
-2.35902265e-01 -2.01142415e-01 1.57474175e-01 -8.10170054e-01
-1.54019356e+00 8.48974407e-01 8.25268626e-01 -3.90530266e-02
1.13226116e+00 5.74533781e-03 1.15195096e+00 -3.29703502e-02
2.10055351e-01 -1.26179767e+00 2.53687948e-01 1.13298625e-01
7.61137307e-01 -1.62463045e+00 1.52852118e-01 -5.41122973e-01
-9.49799776e-01 8.39478076e-01 8.67436290e-01 -2.16187939e-01
6.87498033e-01 3.41192514e-01 4.92576003e-01 -2.23760828e-01
-1.97670162e-01 1.24645671e-02 1.54618576e-01 1.01895764e-01
4.77574706e-01 -2.64254361e-02 -2.34965995e-01 4.44271356e-01
8.00113231e-02 3.11710592e-02 4.47724789e-01 1.10273063e+00
1.48813725e-02 -1.21703982e+00 -1.20774321e-01 3.39883655e-01
-1.08887768e+00 9.60906968e-02 -8.82286131e-02 6.83931530e-01
1.43081740e-01 5.67001045e-01 -2.10847422e-01 -1.22613490e-01
1.97008416e-01 -5.66247664e-02 4.21728760e-01 -5.91920674e-01
-6.57246530e-01 3.32362145e-01 -3.65715504e-01 -5.10537982e-01
-7.84492612e-01 -6.87942147e-01 -1.55814815e+00 1.79678276e-01
-2.47310206e-01 -1.17598316e-02 6.43018782e-01 8.84655654e-01
8.64063054e-02 9.36205149e-01 6.56712413e-01 -7.71338701e-01
-4.64350015e-01 -4.79061335e-01 -3.38322401e-01 5.20800292e-01
4.10906434e-01 -3.87729198e-01 -1.32748961e-01 3.46316934e-01] | [14.68226146697998, -2.0678367614746094] |
9c849682-8cac-4f5f-8526-d56df5c0049a | relation-order-histograms-as-a-network | null | null | https://link.springer.com/chapter/10.1007/978-3-030-77964-1_18 | https://www.iccs-meeting.org/archive/iccs2021/papers/127430217.pdf | Relation order histograms as a network embedding tool | In this work, we introduce a novel graph embedding technique called NERO (Network Embedding based on Relation Order histograms). Its performance is assessed using a number of well-known classification problems and a newly introduced benchmark dealing with detailed laminae venation networks. The proposed algorithm achieves results surpassing those attained by other kernel-type methods and comparable with many state-of-the-art GNNs while requiring no GPU support and being able to handle relatively large input data. It is also demonstrated that the produced representation can be easily paired with existing model interpretation techniques to provide an overview of the individual edge and vertex influence on the investigated process. | ['Michał Idzik', 'Radosław Łazaz'] | 2021-06-09 | null | null | null | international-conference-on-computational-5 | ['network-embedding'] | ['methodology'] | [ 1.65138185e-01 3.23114067e-01 -1.59996241e-01 1.48959339e-01
2.77615458e-01 -4.22032446e-01 1.04530597e+00 7.73942351e-01
-4.31640357e-01 6.02701128e-01 -1.66910142e-01 -5.01966357e-01
-7.41611898e-01 -1.14162695e+00 -2.21153870e-01 -6.65574431e-01
-4.51619536e-01 8.70390892e-01 5.00307083e-01 -4.90799129e-01
4.49054718e-01 1.17969930e+00 -1.59349906e+00 -1.92963198e-01
4.81092542e-01 8.75730872e-01 -3.13386410e-01 6.15930259e-01
-9.78423208e-02 2.82122970e-01 -5.49156785e-01 -4.54965144e-01
-4.32192236e-02 6.14355616e-02 -6.59945428e-01 -9.81157571e-02
4.30887192e-01 2.72135288e-01 -5.71720183e-01 7.48850107e-01
3.43759686e-01 2.31413603e-01 1.11497843e+00 -1.30519640e+00
-2.67974973e-01 4.25279170e-01 -2.45380923e-01 4.48675543e-01
4.09494489e-01 -7.34756738e-02 9.60753441e-01 -4.84491408e-01
9.21404183e-01 1.11749589e+00 6.07798398e-01 -1.18168688e-03
-1.69377661e+00 -1.22573063e-01 -2.34678373e-01 4.45027560e-01
-1.21755850e+00 2.82105953e-01 1.03416884e+00 -4.92954671e-01
1.04188848e+00 5.40335357e-01 8.76331389e-01 9.46373343e-01
4.94708180e-01 2.67961800e-01 1.43080235e+00 -4.87087220e-01
2.41303667e-01 3.15740466e-01 5.12468040e-01 9.56485450e-01
6.46409273e-01 1.08116828e-01 -1.83428749e-01 -3.15371513e-01
6.74726605e-01 -4.26705360e-01 -8.92394930e-02 -7.37137675e-01
-1.10366321e+00 7.84669518e-01 6.17662072e-01 7.05574036e-01
-2.10600197e-01 3.10781896e-01 6.66772366e-01 1.68836474e-01
5.89068353e-01 3.81340504e-01 -1.08037449e-01 -1.34736961e-02
-7.93437123e-01 1.02945484e-01 1.25261283e+00 5.92354059e-01
5.26783407e-01 2.86707878e-01 1.05909951e-01 4.54951823e-01
2.08035946e-01 -5.46466745e-02 3.61344218e-03 -3.10596287e-01
3.61485220e-02 1.02410877e+00 -4.62470055e-01 -1.43841171e+00
-6.17822886e-01 -5.50638258e-01 -9.52476263e-01 7.28371620e-01
4.75178540e-01 2.44068921e-01 -7.03929484e-01 9.31733787e-01
3.78414959e-01 1.61376655e-01 2.77602579e-03 4.47286159e-01
7.58916438e-01 7.14906096e-01 4.94196452e-02 7.79834911e-02
1.38312173e+00 -1.03053987e+00 -7.23765075e-01 4.03239816e-01
2.92417377e-01 -7.92083263e-01 4.24245328e-01 6.47909701e-01
-7.64806032e-01 -5.35954535e-01 -1.28950238e+00 7.59396702e-02
-1.14165521e+00 5.94527908e-02 9.82381701e-01 8.39195013e-01
-1.11988568e+00 1.19371009e+00 -7.22384989e-01 -6.38164282e-01
1.96576685e-01 5.00982761e-01 -6.67212486e-01 1.56600326e-01
-9.60802317e-01 1.12926567e+00 5.12167633e-01 3.08263838e-01
-5.54025650e-01 -6.41979873e-01 -9.21372414e-01 2.70322651e-01
2.91368902e-01 -6.08833253e-01 3.22653770e-01 -2.23091230e-01
-1.63658893e+00 6.28634453e-01 3.84585768e-01 -4.23521250e-01
6.81888521e-01 3.17380399e-01 -6.57724321e-01 4.66343552e-01
-6.09760761e-01 2.73672551e-01 6.54033303e-01 -1.27444518e+00
1.44187482e-02 -2.52809703e-01 1.94167018e-01 -1.09142892e-01
-5.33539653e-01 -2.25204408e-01 -1.19114801e-01 -6.61869705e-01
1.61582753e-01 -7.53377616e-01 -2.50962943e-01 3.62312257e-01
-4.21327025e-01 -1.93662003e-01 1.08563745e+00 -6.89339101e-01
1.19245923e+00 -1.74584186e+00 6.82477713e-01 8.04554343e-01
4.25539941e-01 4.71755445e-01 7.10739493e-02 1.19046652e+00
-3.31610978e-01 1.63826987e-01 -2.94048548e-01 -3.89795341e-02
1.88663125e-01 2.39790052e-01 3.30397129e-01 8.28001618e-01
-3.49768586e-02 5.46942949e-01 -7.64319301e-01 -4.43342358e-01
8.47594023e-01 1.00160670e+00 -7.16731176e-02 -1.83759853e-01
2.18054488e-01 1.88439533e-01 -3.79480012e-02 4.48490024e-01
5.76090813e-01 1.89366981e-01 4.19949442e-01 -5.67615807e-01
8.15569460e-02 -1.77973673e-01 -1.44435942e+00 1.09364283e+00
-5.54625750e-01 8.92144978e-01 -2.19834670e-01 -1.06706405e+00
1.22859561e+00 3.04742008e-01 2.95317531e-01 -6.20843880e-02
4.57491279e-01 1.40436351e-01 3.65587205e-01 -1.81741446e-01
4.31547523e-01 1.58312663e-01 2.98124582e-01 2.73333490e-01
4.80609149e-01 -4.15739343e-02 4.99466240e-01 1.83247313e-01
1.21838319e+00 -2.42867813e-01 5.35232723e-01 -8.25193226e-01
9.48839545e-01 -2.21203670e-01 -2.72490889e-01 1.97426483e-01
6.22839574e-03 -5.93643170e-03 1.01690245e+00 -4.92122978e-01
-1.15411615e+00 -1.20683217e+00 -3.99870425e-01 3.37179393e-01
1.89297080e-01 -5.25357425e-01 -8.47117066e-01 -6.43209815e-01
2.98899651e-01 5.63296378e-01 -9.79026258e-01 6.36214241e-02
-3.77062976e-01 -6.25852525e-01 5.91398358e-01 5.08927047e-01
2.95420051e-01 -1.02172410e+00 -1.84684828e-01 2.31359780e-01
7.53276527e-01 -8.86843085e-01 3.88142914e-01 1.34237975e-01
-1.14630771e+00 -1.34358454e+00 -5.03882825e-01 -8.46790612e-01
9.16221857e-01 -4.40999150e-01 1.11956120e+00 3.15592229e-01
-7.34599352e-01 4.71137434e-01 -3.18060428e-01 8.54539797e-02
-6.43296063e-01 3.59580278e-01 -9.45095420e-02 4.98917587e-02
2.02949479e-01 -8.11591029e-01 -3.09580058e-01 1.96318001e-01
-1.05942941e+00 -2.22072721e-01 4.89747584e-01 8.91002715e-01
3.21898192e-01 5.67388535e-01 2.23293573e-01 -9.97964263e-01
7.75671005e-01 -1.98848754e-01 -8.20877373e-01 1.56717926e-01
-9.96709287e-01 2.81251550e-01 6.01643026e-01 -2.67870903e-01
-6.47709906e-01 -3.48837346e-01 -1.35044515e-01 -1.04370147e-01
-2.78591007e-01 3.30074251e-01 -8.15958306e-02 -7.45131969e-01
1.96938232e-01 4.55847345e-02 2.02998921e-01 -6.38943672e-01
2.52805024e-01 2.43052617e-01 2.56052166e-01 -4.47441638e-01
1.03430462e+00 5.07635474e-01 8.24133277e-01 -1.18621290e+00
1.60821632e-01 -4.93872881e-01 -8.88492286e-01 -5.12827575e-01
6.99212432e-01 -3.95469069e-02 -1.03020227e+00 3.33678544e-01
-1.14448321e+00 5.54090776e-02 -1.57173708e-01 3.90697360e-01
-4.07502741e-01 4.44979668e-01 -9.12615836e-01 -8.42720926e-01
-2.98470527e-01 -9.33831275e-01 7.61380613e-01 5.68449460e-02
-8.56863186e-02 -1.78043377e+00 5.89301027e-02 8.06412920e-02
4.15283620e-01 6.88102067e-01 1.16182530e+00 -6.36512995e-01
-4.51068193e-01 -6.11180663e-01 -2.36219168e-01 3.72938722e-01
-1.92661420e-01 4.60746467e-01 -7.46589303e-01 -3.99724334e-01
-2.25726545e-01 3.03310901e-01 8.03903103e-01 -1.01274721e-01
9.56490636e-01 7.01183975e-02 -5.60942888e-01 3.95044267e-01
2.00531816e+00 -1.50261566e-01 8.55043828e-01 5.35246670e-01
6.78131938e-01 8.16051185e-01 2.07334787e-01 9.03448835e-02
-3.51460576e-01 7.14852214e-01 6.68009460e-01 -3.21140647e-01
-9.25978571e-02 -4.53984737e-02 -1.24964051e-01 1.09507000e+00
-4.97682065e-01 -4.12758768e-01 -9.66107845e-01 4.78698850e-01
-1.61932814e+00 -7.85927236e-01 -6.65728509e-01 1.79720306e+00
4.75067925e-03 5.30445397e-01 -8.13034326e-02 7.60874808e-01
7.86664486e-01 4.70101178e-01 4.52010036e-02 -8.79581332e-01
-2.33288612e-02 6.49011016e-01 8.01632822e-01 7.61204541e-01
-1.06896198e+00 5.89285970e-01 6.88808632e+00 9.70833898e-01
-7.43718445e-01 5.68305440e-02 2.51584142e-01 5.35713434e-01
-2.46840671e-01 8.36086720e-02 -4.62264299e-01 8.95423889e-02
9.18322623e-01 -2.64139146e-01 3.20739001e-01 6.81116998e-01
-2.43048012e-01 -2.66952395e-01 -1.00000286e+00 8.11875641e-01
3.23039412e-01 -1.32555091e+00 1.06901918e-02 2.30902627e-01
4.41479683e-01 -3.44539762e-01 -9.47441384e-02 -6.93374127e-02
-1.95459902e-01 -1.06134105e+00 2.07939982e-01 4.60204482e-01
4.70885873e-01 -9.64066148e-01 1.01641965e+00 -2.90495008e-01
-1.63566780e+00 6.37358725e-02 -3.64928156e-01 -1.82288915e-01
1.93642363e-01 6.27626717e-01 -8.11828852e-01 1.17731297e+00
2.23574132e-01 6.04170442e-01 -9.87220168e-01 1.25426888e+00
-4.18900847e-01 5.81135154e-01 -4.41120923e-01 -3.50886494e-01
4.36443806e-01 -4.71721530e-01 9.42308187e-01 1.23117304e+00
1.73523828e-01 -4.86506701e-01 -2.96214521e-01 7.19805360e-01
2.04869285e-01 3.20890665e-01 -9.71330404e-01 -1.47459179e-01
4.92217131e-02 1.72323072e+00 -1.46643877e+00 -2.98822552e-01
-1.75820872e-01 9.28890228e-01 8.28506052e-02 1.88483313e-01
-7.67789483e-01 -6.45090938e-01 3.90009671e-01 2.47036189e-01
5.05006194e-01 -3.30824852e-01 6.29201299e-04 -7.55383432e-01
-1.45248100e-01 -3.29028130e-01 2.04406649e-01 -4.59223300e-01
-1.30471408e+00 9.12469864e-01 2.41342545e-01 -1.11875761e+00
1.12114539e-02 -1.49512744e+00 -6.07706368e-01 6.98004067e-01
-1.20676827e+00 -1.15262628e+00 -1.30716324e-01 1.46699309e-01
-6.18952587e-02 -2.46575341e-01 1.00995934e+00 5.13612926e-01
-5.37740707e-01 2.54864037e-01 3.54024440e-01 -1.01132795e-01
1.81681380e-01 -1.60791492e+00 3.42604309e-01 4.65682864e-01
3.52965951e-01 3.86117697e-01 9.00097966e-01 -5.30316889e-01
-1.36699224e+00 -7.03853667e-01 8.21121454e-01 -2.58088529e-01
9.46229815e-01 -7.20858276e-01 -7.87345946e-01 4.15903568e-01
4.82513070e-01 2.01993421e-01 5.87500095e-01 1.02988593e-01
-1.31119773e-01 -1.75181985e-01 -1.18239248e+00 6.39678478e-01
7.93041289e-01 -4.08175081e-01 -6.44214451e-01 1.93219796e-01
1.99522287e-01 -1.49093077e-01 -1.32355797e+00 3.75600934e-01
5.24357975e-01 -9.76985991e-01 1.06955862e+00 -4.15214241e-01
1.58065155e-01 -2.75511712e-01 1.78705618e-01 -1.20865214e+00
-2.38852099e-01 -4.74164814e-01 -2.34743953e-01 1.29687428e+00
2.65082598e-01 -8.83084595e-01 8.85059476e-01 -2.84476615e-02
2.19307184e-01 -1.05317569e+00 -1.03010452e+00 -1.05597234e+00
-2.52032757e-01 -3.35670598e-02 4.18163925e-01 1.03632033e+00
-2.55274121e-02 3.10177803e-01 -4.07969989e-02 1.61224946e-01
7.98403919e-01 -1.34571180e-01 4.52814281e-01 -1.65143156e+00
-2.88904637e-01 -5.32881021e-01 -1.53633916e+00 -3.29591446e-02
2.29738086e-01 -1.13512433e+00 -7.75745928e-01 -1.63162792e+00
-1.74952731e-01 -1.39113620e-01 -4.32176918e-01 6.00959249e-02
3.47219169e-01 6.00829422e-01 6.55314401e-02 -4.35275167e-01
-1.28785685e-01 4.40794080e-01 1.11759949e+00 -2.29461566e-01
3.12378585e-01 -3.83257568e-01 1.43209128e-02 7.24610209e-01
7.15649903e-01 -4.14055377e-01 -3.08794320e-01 2.93752700e-01
6.12246282e-02 -1.68631867e-01 5.52904129e-01 -1.39189994e+00
1.89615667e-01 6.15220547e-01 3.36481005e-01 -6.23655021e-01
3.77079993e-01 -1.16832602e+00 4.15645570e-01 7.84882069e-01
-4.25455682e-02 3.00938874e-01 4.93201613e-01 7.21179843e-01
-3.61797869e-01 -4.51608926e-01 5.01663327e-01 3.59154940e-01
-7.95035779e-01 1.82185769e-02 -5.76383173e-01 -4.99635845e-01
1.45213664e+00 -3.46814156e-01 -4.47486460e-01 -1.00099564e-01
-9.87972677e-01 -1.65434271e-01 6.00567281e-01 1.57980844e-01
3.78411502e-01 -1.42922306e+00 -3.78163427e-01 1.26873806e-01
9.41803530e-02 -7.34670877e-01 3.14984471e-01 7.51145482e-01
-1.39752877e+00 3.90104741e-01 -6.95042253e-01 -5.07502317e-01
-1.45288301e+00 6.44570112e-01 1.32053018e-01 -7.12579727e-01
-6.91547930e-01 3.06613922e-01 -4.38055366e-01 -6.81540966e-01
-6.13461658e-02 -2.25317299e-01 -4.63513404e-01 1.09730572e-01
7.02497214e-02 8.81662011e-01 2.36946478e-01 -6.22094274e-01
-4.21160042e-01 6.09516919e-01 3.18182111e-01 7.52119645e-02
1.43918955e+00 3.55602384e-01 -5.62911272e-01 6.69181526e-01
1.25337684e+00 -7.93936998e-02 -7.39337623e-01 3.12455863e-01
6.74847374e-03 -2.18283579e-01 2.26887107e-01 -6.35772943e-01
-1.11485767e+00 9.38169181e-01 9.43248570e-01 6.51941001e-01
8.14121068e-01 -2.55919933e-01 3.96331370e-01 4.46639240e-01
4.34763521e-01 -9.95227039e-01 -2.11555272e-01 7.15317726e-02
8.29468429e-01 -7.14673340e-01 5.22563696e-01 -1.14604044e+00
-7.42977262e-02 1.68476951e+00 3.09651196e-01 -5.27213871e-01
9.66107547e-01 3.79662156e-01 -2.92955428e-01 -4.17770624e-01
-4.18580830e-01 -1.95717126e-01 5.83427072e-01 5.94213486e-01
3.61044884e-01 4.58094850e-02 -8.92876267e-01 -7.89614320e-02
-4.04897109e-02 -4.29284096e-01 5.52185893e-01 8.53407502e-01
4.85643595e-02 -1.62972486e+00 -2.21054986e-01 6.21041000e-01
-2.86657006e-01 1.69118538e-01 -3.65806580e-01 1.51661170e+00
3.99727281e-03 4.48750883e-01 1.31022349e-01 -2.77059227e-01
3.19504291e-01 1.41534731e-02 7.32979417e-01 -4.45111156e-01
-8.57286334e-01 -3.05088431e-01 4.32745010e-01 -6.29544437e-01
-4.45744216e-01 -4.24121976e-01 -7.31887102e-01 -5.37151039e-01
-5.27693331e-01 1.21337503e-01 8.74993682e-01 5.67179799e-01
-2.17802189e-02 1.10814273e+00 2.06440687e-01 -1.15271032e+00
-4.15164139e-03 -8.19860041e-01 -1.02211189e+00 2.85022080e-01
-2.69496888e-01 -1.22340477e+00 -6.56540453e-01 -4.37952131e-01] | [7.047555923461914, 5.907862663269043] |
22026cf4-19e8-48dd-bde0-b0d97d0c367e | motion-aware-dynamic-graph-neural-network-for | 2203.00387 | null | https://arxiv.org/abs/2203.00387v1 | https://arxiv.org/pdf/2203.00387v1.pdf | Motion-aware Dynamic Graph Neural Network for Video Compressive Sensing | Video snapshot compressive imaging (SCI) utilizes a 2D detector to capture sequential video frames and compresses them into a single measurement. Various reconstruction methods have been developed to recover the high-speed video frames from the snapshot measurement. However, most existing reconstruction methods are incapable of capturing long-range spatial and temporal dependencies, which are critical for video processing. In this paper, we propose a flexible and robust approach based on graph neural network (GNN) to efficiently model non-local interactions between pixels in space as well as time regardless of the distance. Specifically, we develop a motion-aware dynamic GNN for better video representation, i.e., represent each pixel as the aggregation of relative nodes under the guidance of frame-by-frame motions, which consists of motion-aware dynamic sampling, cross-scale node sampling and graph aggregation. Extensive results on both simulation and real data demonstrate both the effectiveness and efficiency of the proposed approach, and the visualization clearly illustrates the intrinsic dynamic sampling operations of our proposed model for boosting the video SCI reconstruction results. The code and models will be released to the public. | ['Xin Yuan', 'Bo Chen', 'Ziheng Cheng', 'Ruiying Lu'] | 2022-03-01 | null | null | null | null | ['video-compressive-sensing'] | ['computer-vision'] | [ 3.17389458e-01 -6.59735918e-01 -4.90750335e-02 -4.65806015e-02
-2.42942959e-01 -2.58876383e-01 2.35070124e-01 -3.11015427e-01
3.53768654e-02 4.93079424e-01 2.97949970e-01 -1.83860749e-01
-2.59008139e-01 -6.21917844e-01 -7.19628513e-01 -8.30688059e-01
-3.96280318e-01 -3.47748816e-01 3.56301516e-01 1.62083045e-01
1.40926450e-01 4.86309826e-01 -1.09938180e+00 8.72755703e-03
5.01215577e-01 1.02331531e+00 5.24622798e-01 8.27196181e-01
2.57995844e-01 1.24018145e+00 -2.82640249e-01 3.69397849e-01
3.18268239e-01 -6.67289078e-01 2.51073167e-02 4.00489002e-01
2.31735870e-01 -9.66490865e-01 -1.12422729e+00 1.17296636e+00
5.31708241e-01 1.43542588e-01 1.39128491e-01 -9.25000250e-01
-5.62911689e-01 3.31826657e-01 -9.12519932e-01 6.71602249e-01
5.11813760e-01 7.96414018e-02 3.13220829e-01 -7.76334524e-01
6.55351758e-01 1.11497998e+00 4.86233324e-01 1.38685063e-01
-6.48320794e-01 -6.48960412e-01 9.12450403e-02 6.25323892e-01
-1.53852534e+00 -5.45066893e-01 1.25812578e+00 -1.47012815e-01
5.41777551e-01 2.10301474e-01 8.47849369e-01 7.40507722e-01
3.27497572e-01 7.54048169e-01 9.04058337e-01 -4.63195652e-01
9.11300331e-02 -6.79148614e-01 -7.04479590e-02 6.44869804e-01
3.00004542e-01 1.73695311e-01 -7.40485728e-01 -1.92878425e-01
1.28130913e+00 4.80579585e-01 -8.74230266e-01 -3.85952652e-01
-1.37469208e+00 3.50681275e-01 2.10223064e-01 1.72585279e-01
-6.28351092e-01 3.38258594e-01 2.64159411e-01 1.64836064e-01
3.02589744e-01 -3.78976852e-01 2.10514843e-01 -1.96676422e-03
-1.12857878e+00 -1.30147755e-01 4.82227713e-01 1.05707312e+00
6.19299948e-01 4.16833758e-01 -9.78401676e-02 4.87091005e-01
3.44531417e-01 6.87639832e-01 2.97948241e-01 -1.23587024e+00
3.32714885e-01 1.22776374e-01 -1.92338997e-03 -1.72504115e+00
-1.74284264e-01 -4.04144228e-01 -1.28771675e+00 -9.80270505e-02
-1.73176020e-01 -1.71763763e-01 -4.85100567e-01 1.63415468e+00
4.00386393e-01 1.12122083e+00 -1.04859173e-01 1.13493192e+00
6.12293482e-01 1.05358458e+00 -3.06559235e-01 -9.12280262e-01
9.12518322e-01 -7.21755087e-01 -1.01590538e+00 9.08142626e-02
1.02773517e-01 -6.28281415e-01 2.82082736e-01 3.40754777e-01
-1.28955567e+00 -6.29581094e-01 -1.29822540e+00 1.39435604e-01
4.17959601e-01 -1.50183523e-02 2.35215500e-01 2.34730259e-01
-1.09106517e+00 3.51002336e-01 -1.02565229e+00 -5.44927642e-02
2.24482864e-01 1.25366390e-01 -1.13189824e-01 -6.26772106e-01
-1.13348567e+00 1.71694741e-01 2.70576894e-01 4.05617416e-01
-1.00181520e+00 -4.34709400e-01 -6.08813822e-01 7.15143606e-02
5.83723009e-01 -6.76641762e-01 6.95322514e-01 -7.51197219e-01
-1.13661075e+00 1.55741751e-01 -2.52762973e-01 -3.58481795e-01
3.31937313e-01 2.45329756e-02 -5.20371258e-01 7.22074986e-01
-9.60184932e-02 1.16344646e-01 9.60904241e-01 -1.09570420e+00
-4.72168386e-01 -3.24688733e-01 6.09079003e-02 4.58307028e-01
-3.80891323e-01 -1.10113837e-01 -9.22730684e-01 -5.94528317e-01
3.35504293e-01 -8.12395811e-01 -1.53164193e-01 1.03007041e-01
-2.47535333e-01 3.94514024e-01 1.15671325e+00 -9.31141198e-01
1.57089949e+00 -2.26293755e+00 1.59887329e-01 1.08392268e-01
4.94877934e-01 3.27916473e-01 2.39676479e-02 4.14888293e-01
4.76015471e-02 -4.33069736e-01 -1.57427356e-01 -5.92317581e-02
-6.14836156e-01 2.25226089e-04 -9.57852155e-02 7.18189716e-01
-4.02058989e-01 6.46869779e-01 -1.02980816e+00 -4.79069084e-01
5.73945999e-01 7.02361524e-01 -3.39083254e-01 2.52220899e-01
1.43208042e-01 4.80991513e-01 -6.95755780e-01 6.78496242e-01
1.00376248e+00 -3.61039281e-01 4.01552409e-01 -5.60841680e-01
-1.20937385e-01 -2.31680632e-01 -1.38493824e+00 1.68047035e+00
-3.36298458e-02 6.00120604e-01 3.74339461e-01 -9.91883516e-01
6.09791934e-01 2.86504507e-01 8.49214256e-01 -7.32723236e-01
1.09155560e-02 4.09018099e-02 -3.32155287e-01 -7.05721319e-01
5.02436996e-01 2.65396416e-01 4.32588488e-01 4.24032629e-01
-3.78901094e-01 4.77713019e-01 8.96392167e-02 3.49427402e-01
1.25500417e+00 1.04003489e-01 3.73140812e-01 -3.93452030e-03
6.33957803e-01 -3.21860522e-01 6.33628786e-01 6.22885048e-01
-3.73768985e-01 6.81901991e-01 1.37831315e-01 -3.57902080e-01
-1.01148272e+00 -8.68712783e-01 3.10835540e-01 5.61515868e-01
7.77804017e-01 -3.44708204e-01 -5.86806774e-01 -3.40634473e-02
-3.46442521e-01 4.16165918e-01 -8.66626874e-02 -1.58536270e-01
-8.87139201e-01 -6.44981623e-01 1.81456268e-01 1.55697569e-01
7.78330207e-01 -6.87056482e-01 -7.70823121e-01 4.34426337e-01
-5.23483872e-01 -1.39379454e+00 -7.50678420e-01 -5.05174637e-01
-8.61803412e-01 -1.09621215e+00 -7.80541539e-01 -7.60779917e-01
5.56456745e-01 1.31827092e+00 6.59557045e-01 3.80117804e-01
-1.03216246e-01 5.09262443e-01 -3.65114808e-01 2.64915973e-01
-1.14219733e-01 -6.12887144e-01 1.88498925e-02 4.06992525e-01
-1.55294994e-02 -9.63111818e-01 -1.01505244e+00 1.91934034e-01
-1.14297032e+00 3.52754474e-01 5.57496905e-01 6.63065135e-01
7.08533049e-01 2.27993697e-01 1.48865148e-01 -5.14456809e-01
5.54436088e-01 -7.31807709e-01 -4.06987906e-01 2.03279331e-01
-2.85804510e-01 -4.60187316e-01 6.00722432e-01 -4.04221147e-01
-8.80445302e-01 1.95131656e-02 2.46310592e-01 -1.09655666e+00
2.15792432e-01 6.71614587e-01 -1.29993632e-01 -3.40451539e-01
3.69906537e-02 7.64172375e-01 1.54913709e-01 -8.23550299e-02
7.31430277e-02 4.73270684e-01 6.07878089e-01 -3.17051381e-01
6.49467289e-01 7.29735553e-01 2.99990356e-01 -1.09381771e+00
-1.96561828e-01 -4.64049280e-01 -2.84326077e-01 -6.76222980e-01
7.15054095e-01 -1.24362195e+00 -5.75626075e-01 5.95604897e-01
-1.04600561e+00 -6.12085462e-02 1.49870455e-01 7.52866507e-01
-3.98801923e-01 1.04814255e+00 -8.77808928e-01 -7.25374877e-01
-2.58728355e-01 -1.06459272e+00 8.94693732e-01 2.86869466e-01
4.11500275e-01 -9.45562005e-01 -9.45856497e-02 3.84609550e-02
4.08041924e-01 4.38975275e-01 6.45208359e-01 1.54464424e-01
-1.31064939e+00 -1.18019968e-01 -3.81652027e-01 2.24924192e-01
1.43337995e-01 -6.15569055e-02 -4.32528883e-01 -5.02298594e-01
3.95833522e-01 2.27791369e-01 5.07489741e-01 8.77809107e-01
1.12337828e+00 -2.59572089e-01 -3.05864394e-01 9.55855012e-01
1.68259656e+00 4.12751436e-01 7.31358647e-01 6.76191524e-02
9.72652316e-01 1.08563229e-01 4.85002518e-01 6.01292133e-01
3.43907952e-01 5.92431247e-01 5.58404744e-01 3.93567793e-02
-3.43229920e-01 -2.23250195e-01 3.89686197e-01 1.30586874e+00
-2.37028092e-01 -6.00070238e-01 -5.47697723e-01 3.64289761e-01
-1.89926076e+00 -1.30171525e+00 -1.90289974e-01 1.93697762e+00
1.81357935e-01 -1.61474749e-01 -2.54488528e-01 -3.12793665e-02
1.04047298e+00 6.67585194e-01 -6.67274177e-01 4.04411167e-01
-2.93904483e-01 -2.86224812e-01 5.34186482e-01 4.43185717e-01
-7.18699515e-01 3.65348250e-01 5.75535536e+00 8.28905523e-01
-1.18660510e+00 4.72022593e-02 5.91634512e-01 -1.12956189e-01
-2.21349403e-01 1.96084455e-02 -2.53763646e-01 7.33520389e-01
6.21643186e-01 -3.14617246e-01 6.71239018e-01 4.09167230e-01
6.18265927e-01 -9.81554314e-02 -5.81977546e-01 1.26919627e+00
2.32858032e-01 -1.54811954e+00 2.04776064e-01 2.93150432e-02
6.90161049e-01 -1.33063301e-01 -1.28126845e-01 -2.66648710e-01
-2.43423358e-01 -4.54436868e-01 5.98366737e-01 7.28255689e-01
7.83234060e-01 -5.38462341e-01 4.50578213e-01 4.73247558e-01
-1.40633214e+00 -4.92407829e-02 -3.39808077e-01 -1.19705401e-01
5.90877712e-01 6.77540541e-01 -2.43080020e-01 9.44977760e-01
5.52218258e-01 1.17414284e+00 -2.44092152e-01 9.91406739e-01
1.56730920e-01 6.44357920e-01 -3.14270079e-01 1.51180670e-01
1.60897732e-01 -5.62455952e-01 8.08626592e-01 8.68986905e-01
8.98399591e-01 6.18918478e-01 2.92267174e-01 4.77046490e-01
1.30843714e-01 -1.44824967e-01 -6.14971697e-01 1.27305940e-01
6.38604581e-01 9.72832441e-01 -5.50892472e-01 -3.92936379e-01
-7.14089870e-01 9.00419950e-01 -3.55382822e-02 6.74247086e-01
-7.92623818e-01 -5.95988706e-02 3.30234259e-01 2.85990059e-01
4.23930049e-01 -6.39399886e-01 1.49238050e-01 -1.24472654e+00
2.19972715e-01 -7.64425278e-01 2.56401896e-01 -1.05928326e+00
-8.07743073e-01 3.51173162e-01 1.14153355e-01 -1.62226188e+00
-1.38283476e-01 -9.45326909e-02 -5.88797867e-01 5.60351610e-01
-1.57103801e+00 -8.92567694e-01 -7.51777947e-01 9.22765434e-01
4.28312212e-01 -5.04022604e-03 2.20664963e-01 6.04674697e-01
-5.57929575e-01 1.16689540e-01 2.98031002e-01 8.57894421e-02
4.60950702e-01 -3.43799293e-01 8.01070482e-02 1.34099126e+00
-1.49748430e-01 5.14598727e-01 6.37115180e-01 -8.67709219e-01
-2.00037503e+00 -9.63018119e-01 2.93310553e-01 4.07933712e-01
5.11700988e-01 5.78415841e-02 -8.54053915e-01 6.36792660e-01
2.72789508e-01 4.16766167e-01 2.36818448e-01 -7.12647855e-01
3.16505954e-02 -2.89654672e-01 -8.99221063e-01 5.38676322e-01
1.26977170e+00 -5.46323061e-01 -1.21591389e-01 2.56860524e-01
6.83487713e-01 -5.53286731e-01 -5.39366663e-01 4.10051942e-01
5.23109734e-01 -1.35751283e+00 1.18156791e+00 1.77252501e-01
4.66856241e-01 -7.00468123e-01 -3.65603983e-01 -8.76024783e-01
-5.30487120e-01 -7.73415864e-01 -6.13409281e-01 9.29536104e-01
-5.41507065e-01 -4.80262041e-01 6.69502616e-01 9.22004953e-02
-1.49787396e-01 -6.39434159e-01 -9.73702013e-01 -4.70348388e-01
-8.56631517e-01 -3.78108293e-01 4.41251844e-01 9.53083932e-01
-3.63879055e-01 9.16907042e-02 -9.24408138e-01 6.80987954e-01
1.08014369e+00 1.30510181e-01 5.80247819e-01 -5.06483674e-01
-4.73041177e-01 -3.42550799e-02 -7.28376031e-01 -1.57795537e+00
-1.68603629e-01 -4.20104325e-01 -8.98113474e-02 -1.48635697e+00
4.02756125e-01 -1.98470280e-01 -4.35334265e-01 -4.06436563e-01
-2.70806044e-01 1.51501834e-01 2.97182947e-01 6.09237850e-01
-6.49665236e-01 4.53207344e-01 1.43917942e+00 -4.08478193e-02
4.18987907e-02 -2.35457465e-01 -2.97705352e-01 4.82888162e-01
4.26828355e-01 -2.34220892e-01 -7.50091732e-01 -8.10635388e-01
-4.68155742e-02 9.44727898e-01 5.03608346e-01 -1.22662771e+00
5.88279605e-01 -2.63653606e-01 3.77192199e-01 -7.73885429e-01
4.58292603e-01 -9.68263626e-01 6.74885809e-01 4.74807262e-01
-1.16654718e-03 1.94743723e-01 -1.69162527e-01 1.11068559e+00
-2.87748307e-01 9.67020616e-02 5.78517795e-01 -1.52560741e-01
-9.43911433e-01 7.26722062e-01 -1.08752713e-01 -2.80171633e-01
1.17951655e+00 -3.81698698e-01 -2.94745833e-01 -8.46290231e-01
-2.97852367e-01 9.00550410e-02 5.27609110e-01 6.90306202e-02
8.85832369e-01 -1.46213365e+00 -7.36531734e-01 3.87556225e-01
-2.48461559e-01 -1.18477151e-01 9.70997989e-01 9.47562516e-01
-9.77141976e-01 1.43285781e-01 -1.99365973e-01 -7.86934674e-01
-1.22605956e+00 7.41370916e-01 1.29712537e-01 -1.08147517e-01
-8.56922209e-01 5.41168749e-01 2.06054553e-01 3.33404750e-01
7.27459490e-02 -4.10905033e-02 -6.43183887e-02 -3.54302377e-01
7.78710544e-01 5.08671045e-01 -5.03667235e-01 -7.13136077e-01
-1.70611367e-01 7.49254584e-01 2.34548911e-01 6.73739091e-02
1.26552081e+00 -7.54325747e-01 -1.76446125e-01 2.71166176e-01
1.28082383e+00 1.43960431e-01 -1.53359127e+00 -4.16448295e-01
-6.67903662e-01 -7.54819930e-01 2.61274189e-01 5.17266570e-04
-1.29550958e+00 6.33107007e-01 5.51965415e-01 2.36525610e-01
1.48630965e+00 -4.78722811e-01 1.17170107e+00 -2.06095949e-02
5.67543626e-01 -7.01126873e-01 -7.12918304e-03 2.10068733e-01
4.86740321e-01 -7.63623059e-01 4.79769886e-01 -5.10117173e-01
-1.77298918e-01 1.32277346e+00 1.85700893e-01 -3.77267241e-01
6.63705528e-01 2.05632195e-01 -2.72258282e-01 -5.16932458e-02
-6.69521213e-01 3.00579697e-01 -1.16004191e-01 5.18244624e-01
1.13556562e-02 -5.98902740e-02 -2.17243657e-01 -4.44550328e-02
5.08649230e-01 3.46529484e-01 8.15768242e-01 9.87805724e-01
-3.71721387e-01 -4.93244082e-01 -4.88356709e-01 2.99463898e-01
-1.64789066e-01 5.56090800e-03 3.21984380e-01 5.95266342e-01
-5.70628792e-02 9.49065804e-01 -3.90762910e-02 -5.66706002e-01
-7.60803232e-03 -5.53782463e-01 4.60795164e-01 -9.37087983e-02
2.24582508e-01 3.50738108e-01 -1.60932407e-01 -7.34580100e-01
-7.82942593e-01 -7.32205808e-01 -1.02793992e+00 -5.92911303e-01
-6.90694600e-02 -1.16212115e-01 4.27687347e-01 6.92947984e-01
4.86268580e-01 6.43680036e-01 9.01594698e-01 -1.00298667e+00
-2.70439267e-01 -5.93682468e-01 -7.59350896e-01 3.38471353e-01
6.78830445e-01 -3.27668011e-01 -5.80845416e-01 1.46500155e-01] | [11.029342651367188, -2.068798780441284] |
292b4713-77de-4654-be14-ccaa0c99e870 | sirfyn-single-image-relighting-from-your | 2112.04497 | null | https://arxiv.org/abs/2112.04497v1 | https://arxiv.org/pdf/2112.04497v1.pdf | SIRfyN: Single Image Relighting from your Neighbors | We show how to relight a scene, depicted in a single image, such that (a) the overall shading has changed and (b) the resulting image looks like a natural image of that scene. Applications for such a procedure include generating training data and building authoring environments. Naive methods for doing this fail. One reason is that shading and albedo are quite strongly related; for example, sharp boundaries in shading tend to appear at depth discontinuities, which usually apparent in albedo. The same scene can be lit in different ways, and established theory shows the different lightings form a cone (the illumination cone). Novel theory shows that one can use similar scenes to estimate the different lightings that apply to a given scene, with bounded expected error. Our method exploits this theory to estimate a representation of the available lighting fields in the form of imputed generators of the illumination cone. Our procedure does not require expensive "inverse graphics" datasets, and sees no ground truth data of any kind. Qualitative evaluation suggests the method can erase and restore soft indoor shadows, and can "steer" light around a scene. We offer a summary quantitative evaluation of the method with a novel application of the FID. An extension of the FID allows per-generated-image evaluation. Furthermore, we offer qualitative evaluation with a user study, and show that our method produces images that can successfully be used for data augmentation. | ['YuXiong Wang', 'Yuanyi Zhong', 'Pranav Asthana', 'Anand Bhattad', 'D. A. Forsyth'] | 2021-12-08 | null | null | null | null | ['image-relighting'] | ['computer-vision'] | [ 7.31029928e-01 1.42908514e-01 7.03893185e-01 -6.96871758e-01
-3.22433710e-01 -5.76816916e-01 6.36926949e-01 -4.13277149e-01
8.25673118e-02 7.46441960e-01 2.27012664e-01 -3.58803034e-01
3.73516530e-01 -9.62023735e-01 -9.12769139e-01 -7.33340383e-01
2.25684181e-01 4.18752462e-01 1.85499251e-01 -5.02532840e-01
1.54628515e-01 5.89498460e-01 -1.84601212e+00 3.74954492e-01
8.12428534e-01 4.07878041e-01 4.49846387e-01 9.91248488e-01
-1.88527748e-01 6.56191051e-01 -8.15816879e-01 -1.26962245e-01
3.82539243e-01 -6.60740435e-01 -6.35944724e-01 6.79324865e-01
8.89111817e-01 -4.14095640e-01 3.44286077e-02 8.33607376e-01
3.34420204e-01 1.41878322e-01 4.95714247e-01 -1.05804491e+00
-4.35591966e-01 -2.43535772e-01 -5.54712653e-01 -4.07568127e-01
9.38696384e-01 1.93447992e-01 5.18423438e-01 -7.80565083e-01
6.67907059e-01 1.17795742e+00 7.45410144e-01 3.88266295e-01
-1.70002222e+00 -1.33757502e-01 -1.67058527e-01 -3.49189132e-01
-1.20043993e+00 -5.66754639e-01 7.81812727e-01 -3.65089744e-01
5.16907811e-01 8.58797967e-01 8.55224669e-01 7.86026299e-01
4.46781665e-02 3.89186352e-01 1.66758609e+00 -8.69448960e-01
2.70653665e-01 6.43256366e-01 -2.73748398e-01 8.59747529e-01
1.95554085e-02 1.74692512e-01 -4.93725389e-01 -9.97963473e-02
7.29481220e-01 -2.08185181e-01 -7.00509071e-01 -5.00299931e-01
-9.70644653e-01 4.68127489e-01 5.75687766e-01 1.80972032e-02
6.05856702e-02 1.55043676e-01 -2.01056868e-01 1.24823563e-01
5.64693987e-01 3.98687720e-01 -2.14627936e-01 3.62541288e-01
-7.54275739e-01 2.47766346e-01 9.70530987e-01 7.89933503e-01
1.18675816e+00 3.47663723e-02 2.91860580e-01 7.93100119e-01
2.72020608e-01 8.67463231e-01 4.76817675e-02 -1.34004927e+00
1.19218543e-01 2.08446741e-01 3.06742907e-01 -1.13619184e+00
-2.29307353e-01 1.10189065e-01 -5.63533545e-01 9.59489167e-01
5.56582153e-01 -1.00249220e-02 -1.12106550e+00 1.54136193e+00
3.88353914e-01 1.18721381e-01 1.33009849e-03 7.87220597e-01
6.23790205e-01 7.06767559e-01 -6.11804545e-01 1.99243575e-02
1.21952283e+00 -5.08598506e-01 -6.65344834e-01 -5.06600142e-01
4.80025709e-01 -9.60654199e-01 1.54802084e+00 5.99079132e-01
-1.10800743e+00 -4.18401808e-01 -1.00676477e+00 -2.09036365e-01
-2.81089932e-01 -1.59623817e-01 8.64891291e-01 1.01669610e+00
-1.27995694e+00 5.63137054e-01 -4.51840371e-01 -3.91186416e-01
1.32260211e-02 6.78075524e-03 -1.56923786e-01 -2.99064487e-01
-6.98619068e-01 8.04593503e-01 -3.69981304e-02 1.57286331e-01
-6.80217743e-01 -6.37514830e-01 -9.33579326e-01 -2.10886464e-01
-1.61332544e-02 -7.47632623e-01 1.10548139e+00 -1.31853998e+00
-1.46326172e+00 9.99830961e-01 -4.33585376e-01 1.42200530e-01
3.87816697e-01 -1.41050830e-01 -2.44111061e-01 1.43207470e-02
2.00661668e-03 5.33101439e-01 7.23282456e-01 -2.16110253e+00
-2.57132381e-01 -3.18501920e-01 1.28973201e-01 5.09076178e-01
8.34212750e-02 -2.16531187e-01 -3.36059064e-01 -2.55366713e-01
3.42624635e-01 -7.18665421e-01 -3.70679826e-01 2.37509906e-01
-6.70821488e-01 8.18850219e-01 7.89098442e-01 -5.99262238e-01
7.25999713e-01 -2.14782929e+00 -4.21986759e-01 5.83794177e-01
-1.25406981e-01 -2.10234672e-01 1.26856983e-01 3.08642536e-01
-1.71886891e-01 3.09471954e-02 -5.42549372e-01 -5.84561229e-01
-1.90511107e-01 5.66124022e-01 -6.60686493e-01 5.41225433e-01
-3.34031671e-01 3.47342670e-01 -8.55081856e-01 -4.25775021e-01
5.70083261e-01 7.83061206e-01 -5.74401200e-01 2.72426665e-01
-2.42992371e-01 5.94147027e-01 5.83198257e-02 2.37395227e-01
8.89566600e-01 1.00572996e-01 2.20261127e-01 -1.46224022e-01
-2.38262847e-01 1.67619348e-01 -1.44000399e+00 1.57089448e+00
-6.43047631e-01 1.04296744e+00 5.65286055e-02 -3.31308633e-01
9.70689893e-01 6.41837791e-02 9.29460078e-02 -6.96117043e-01
-5.94609939e-02 8.85728002e-02 -4.18166727e-01 -4.39008206e-01
5.92415094e-01 -4.27729428e-01 3.52859616e-01 3.99269521e-01
-6.41397357e-01 -1.17337561e+00 -1.68417707e-01 1.80042505e-01
9.58947778e-01 4.83371168e-01 6.24797009e-02 -1.86899856e-01
1.85969293e-01 1.24625675e-02 1.83541387e-01 6.35293484e-01
4.77170765e-01 1.22192907e+00 2.55580515e-01 -5.10773003e-01
-1.15969801e+00 -9.76678491e-01 -3.89567435e-01 6.55967414e-01
2.09517583e-01 -3.52600329e-02 -1.00298715e+00 -1.04124121e-01
-1.28815174e-01 1.19291770e+00 -6.72315240e-01 3.27112198e-01
-3.95349741e-01 -7.37283826e-01 1.58867970e-01 -1.22024417e-02
5.22790968e-01 -1.02098286e+00 -8.81176054e-01 -1.82536662e-01
-2.22481385e-01 -1.00778615e+00 -2.33209968e-01 5.40172197e-02
-8.64497125e-01 -1.05912769e+00 -4.66657311e-01 -6.15104198e-01
1.18444872e+00 6.28026247e-01 1.45322359e+00 3.43363881e-01
-5.10351717e-01 6.56031013e-01 -5.65897413e-02 -3.74482274e-01
-6.11099482e-01 -8.41593146e-01 -2.67114431e-01 9.39951316e-02
-1.99604765e-01 -7.33004987e-01 -6.79974437e-01 4.69587266e-01
-1.09683251e+00 5.30961990e-01 -1.48152942e-02 5.68532228e-01
5.13999641e-01 7.61664361e-02 -3.43868017e-01 -1.25398147e+00
2.55657703e-01 1.45488055e-02 -7.12021112e-01 1.90769404e-01
-4.18747663e-01 -1.30641852e-02 5.71228325e-01 7.62470812e-02
-1.66250300e+00 3.56767744e-01 9.27655101e-02 -1.06218725e-01
-4.51418251e-01 -3.02369539e-02 -2.88374215e-01 -2.50932187e-01
1.10806918e+00 -9.63479504e-02 -1.85195222e-01 -4.44528073e-01
6.05237603e-01 3.81518394e-01 7.92965531e-01 -6.48764729e-01
1.01244652e+00 1.04859817e+00 2.89747298e-01 -1.16568077e+00
-8.20158005e-01 -1.28810301e-01 -5.43390572e-01 -3.06798398e-01
5.62678456e-01 -5.49933970e-01 -6.10986650e-01 3.10649902e-01
-1.25982261e+00 -8.36991549e-01 -6.90729856e-01 1.29785389e-01
-6.69826388e-01 4.01233554e-01 -7.98361078e-02 -1.09559345e+00
3.74435306e-01 -8.59166384e-01 1.23840046e+00 2.46636391e-01
-2.90689338e-02 -1.21285915e+00 3.56354952e-01 3.14833969e-01
9.54240859e-02 5.70547879e-01 7.07220256e-01 3.20094764e-01
-7.61987209e-01 1.36647418e-01 -2.40423381e-01 4.84763503e-01
3.84993225e-01 2.67491013e-01 -1.60893369e+00 -1.31182909e-01
8.77292082e-02 -1.27114236e-01 5.43231964e-01 3.77285153e-01
1.14357901e+00 -3.59559804e-01 -1.44173488e-01 8.96510839e-01
1.84974170e+00 9.96862054e-02 1.26194108e+00 2.96282828e-01
8.85896623e-01 9.44282055e-01 3.37200463e-01 3.30469310e-01
1.98095143e-01 7.62211084e-01 4.71055835e-01 -8.44909549e-01
-4.20417517e-01 -1.07979752e-01 2.24731505e-01 3.62956911e-01
-1.55712351e-01 -3.01240653e-01 -8.27620625e-01 3.10126990e-01
-1.35826802e+00 -9.12632525e-01 -6.98381960e-01 2.60008240e+00
8.39133084e-01 -9.77596045e-02 -4.46611166e-01 1.69906244e-01
3.21893424e-01 1.43533587e-01 -1.60996631e-01 -6.45305336e-01
-3.29975933e-01 2.82994360e-01 6.49480999e-01 1.16812086e+00
-5.57823241e-01 6.08674645e-01 7.07558203e+00 2.09692121e-01
-7.74894118e-01 -2.32527137e-01 8.25100482e-01 3.31039459e-01
-9.67435837e-01 4.72471058e-01 -3.29945803e-01 1.85487628e-01
6.08608246e-01 1.75899804e-01 6.97770953e-01 6.53688908e-01
4.22836959e-01 -7.84262180e-01 -1.12267005e+00 8.55714083e-01
3.21112335e-01 -1.09170318e+00 -2.05791950e-01 1.23886548e-01
9.51431692e-01 -3.52393419e-01 -2.64647771e-02 -2.71373093e-01
4.99262363e-01 -1.01654053e+00 5.34927428e-01 8.41894090e-01
1.03476405e+00 -4.22604740e-01 2.82920688e-01 2.31254920e-01
-7.51206756e-01 3.80330652e-01 -4.13856566e-01 -1.46149531e-01
2.49966070e-01 7.69067943e-01 -1.25548601e+00 4.04524356e-01
5.62618256e-01 1.42994925e-01 -6.57960296e-01 1.00504053e+00
-4.19462502e-01 3.52462143e-01 -5.71783125e-01 3.36770236e-01
-2.83097774e-01 -6.24920309e-01 4.00512308e-01 1.18119967e+00
3.89785439e-01 1.66481391e-01 6.93370327e-02 8.11934233e-01
1.12830438e-01 -2.58865841e-02 -1.14064515e+00 6.80572927e-01
2.05959566e-02 1.11478317e+00 -8.03004444e-01 -3.25690389e-01
-2.35851735e-01 1.26345551e+00 -2.18175948e-01 7.56390750e-01
-5.82409918e-01 -2.35926971e-01 4.08544809e-01 4.56268191e-01
-2.16167226e-01 -3.30022275e-02 -6.24344409e-01 -9.82488573e-01
-1.05022304e-01 -7.63382733e-01 -2.21988916e-01 -1.77763379e+00
-8.25498402e-01 4.08042699e-01 3.13782729e-02 -1.08417726e+00
-2.09566817e-01 -4.96206552e-01 -5.91351628e-01 1.03906417e+00
-1.39110231e+00 -9.17326152e-01 -9.05025363e-01 5.28221250e-01
4.11909938e-01 3.96288306e-01 1.04188991e+00 5.89870103e-02
-1.11089729e-01 -2.39605512e-02 3.44689429e-01 -3.03144962e-01
5.76583982e-01 -1.65542650e+00 3.82342637e-01 9.69900787e-01
3.83053094e-01 3.77992153e-01 1.29335070e+00 -5.19290209e-01
-1.17739344e+00 -6.96341872e-01 5.14651716e-01 -6.96672142e-01
9.82764587e-02 -4.39533114e-01 -9.46324050e-01 6.20020568e-01
2.34598547e-01 -5.40864728e-02 3.25822443e-01 -3.45646627e-02
-3.44532907e-01 -2.08507672e-01 -1.32389879e+00 8.63675535e-01
8.64205182e-01 -4.23707962e-01 -1.23992413e-01 7.89314449e-01
5.74936330e-01 -6.92729115e-01 -2.18804464e-01 1.18661784e-01
5.18231034e-01 -1.81744790e+00 1.10104525e+00 -1.43805251e-01
3.48411113e-01 -5.56234479e-01 -4.39029038e-01 -1.38154101e+00
1.00383945e-01 -7.03320086e-01 3.44102830e-01 1.10225391e+00
1.72494575e-01 -6.18249595e-01 7.58372843e-01 9.79711592e-01
-2.93618262e-01 -3.28075409e-01 -5.17575800e-01 -4.95176733e-01
-5.16996622e-01 -4.84952122e-01 6.24819934e-01 8.16868544e-01
-6.39187455e-01 2.13048816e-01 -3.58979404e-01 3.48953366e-01
7.40670800e-01 1.38311535e-01 1.28066611e+00 -1.08831787e+00
-5.02754569e-01 -1.69895217e-02 -1.35035589e-01 -9.69355643e-01
-2.84369022e-01 -5.01899004e-01 4.59618747e-01 -1.73893893e+00
4.78541963e-02 -8.18242550e-01 4.32770073e-01 4.02226478e-01
4.41200137e-02 4.22526300e-01 -6.14848640e-03 9.08531249e-02
1.68945625e-01 3.15301806e-01 1.29679537e+00 2.11507425e-01
-3.26650083e-01 -9.32787880e-02 -4.97701794e-01 1.16513920e+00
6.67083859e-01 -2.17194006e-01 -6.49547994e-01 -5.19551218e-01
4.93203223e-01 -1.14125833e-01 7.28767455e-01 -9.70212042e-01
-2.66870618e-01 -2.08679080e-01 6.41220391e-01 -3.41858804e-01
6.97972119e-01 -1.03182793e+00 6.37250781e-01 1.63515553e-01
-4.20067366e-03 -2.96506554e-01 1.04469925e-01 3.32046837e-01
2.88297474e-01 -4.09562171e-01 8.37091148e-01 -2.36610487e-01
-4.52490121e-01 -2.78257132e-01 -2.51670275e-02 -1.70261621e-01
5.02442241e-01 -6.98647320e-01 -3.19693863e-01 -7.67314136e-01
-3.69618863e-01 -2.76211619e-01 1.06890643e+00 -3.42103064e-01
5.93365312e-01 -1.10291946e+00 -4.67921227e-01 3.58157218e-01
-1.43247455e-01 2.69387126e-01 6.82223216e-02 3.50333750e-01
-1.01025724e+00 -3.19788992e-01 7.57069588e-02 -6.61970913e-01
-1.35798311e+00 2.51572102e-01 5.87187111e-01 2.29201451e-01
-8.48909736e-01 7.07629859e-01 8.40893090e-01 -4.66399729e-01
-1.70529131e-02 -3.01065862e-01 2.18317181e-01 -2.27161974e-01
5.07163525e-01 5.51935695e-02 1.25255883e-01 -4.73215371e-01
-1.43667534e-01 6.45643413e-01 5.21142185e-01 -5.82271278e-01
1.29775310e+00 -3.08526784e-01 -2.24574968e-01 7.19211519e-01
1.02108490e+00 5.44966757e-01 -1.38382709e+00 3.93465720e-02
-5.58461547e-01 -1.20677888e+00 -1.49612147e-02 -8.71255219e-01
-7.50942469e-01 9.24834669e-01 6.10432863e-01 5.49438179e-01
1.45295405e+00 -2.25439519e-01 4.30464536e-01 3.81350011e-01
5.10806859e-01 -9.09025073e-01 -3.60733643e-03 1.57336235e-01
1.07343745e+00 -1.01154411e+00 5.09097338e-01 -6.21466339e-01
-5.01159370e-01 1.12762129e+00 2.82916903e-01 6.48398772e-02
3.32305789e-01 4.56676871e-01 3.08324963e-01 -3.80295187e-01
-1.90512300e-01 -5.79957440e-02 1.72410578e-01 8.11757565e-01
3.87150377e-01 7.18733966e-02 3.58072281e-01 -4.42821264e-01
-5.29277682e-01 -2.49138281e-01 1.01120555e+00 7.99625516e-01
-5.42858660e-01 -9.61254835e-01 -9.58404124e-01 1.05365902e-01
2.14878321e-01 -1.38381496e-01 -2.94943213e-01 7.62148380e-01
9.22610983e-02 8.43878269e-01 6.28996417e-02 -4.39340249e-02
4.39696312e-01 -1.08049452e-01 7.26298332e-01 -6.63632393e-01
-1.88499764e-01 1.93496361e-01 2.69593805e-01 -6.68701649e-01
-5.98354340e-01 -6.22472584e-01 -1.10234213e+00 -4.16100562e-01
-3.46810132e-01 -1.35354072e-01 1.01965427e+00 6.08338237e-01
-4.24429066e-02 5.00265539e-01 9.72406447e-01 -1.22640407e+00
2.61576951e-01 -4.50464636e-01 -7.09183156e-01 5.84543407e-01
4.84714329e-01 -4.09682751e-01 -6.75673783e-01 6.22728765e-01] | [9.821457862854004, -2.986156702041626] |
ec4f17e7-345b-4cf7-9ed1-d33c38322145 | image-restoration-using-convolutional-auto | 1606.08921 | null | http://arxiv.org/abs/1606.08921v3 | http://arxiv.org/pdf/1606.08921v3.pdf | Image Restoration Using Convolutional Auto-encoders with Symmetric Skip Connections | Image restoration, including image denoising, super resolution, inpainting,
and so on, is a well-studied problem in computer vision and image processing,
as well as a test bed for low-level image modeling algorithms. In this work, we
propose a very deep fully convolutional auto-encoder network for image
restoration, which is a encoding-decoding framework with symmetric
convolutional-deconvolutional layers. In other words, the network is composed
of multiple layers of convolution and de-convolution operators, learning
end-to-end mappings from corrupted images to the original ones. The
convolutional layers capture the abstraction of image contents while
eliminating corruptions. Deconvolutional layers have the capability to upsample
the feature maps and recover the image details. To deal with the problem that
deeper networks tend to be more difficult to train, we propose to symmetrically
link convolutional and deconvolutional layers with skip-layer connections, with
which the training converges much faster and attains better results. | ['Yu-Bin Yang', 'Xiao-Jiao Mao', 'Chunhua Shen'] | 2016-06-29 | null | null | null | null | ['jpeg-artifact-correction'] | ['computer-vision'] | [ 3.07263613e-01 -7.57953674e-02 2.71434605e-01 -4.12469745e-01
-3.64361078e-01 5.42980433e-02 3.66072148e-01 -4.61393774e-01
-2.39476621e-01 6.65872455e-01 4.57884163e-01 -5.29551171e-02
2.16155022e-01 -7.28381038e-01 -1.07234418e+00 -8.05964470e-01
2.26804018e-01 -1.72463968e-01 -3.01426109e-02 -2.65563518e-01
6.70307577e-02 5.17727494e-01 -1.56011713e+00 6.77358806e-01
9.11473572e-01 9.25055623e-01 4.84503686e-01 8.04449081e-01
-1.20060623e-01 1.37876368e+00 -4.02561873e-01 -4.53311592e-01
4.94020805e-02 -4.28652436e-01 -7.81978309e-01 2.54239142e-01
3.70550692e-01 -8.08344543e-01 -8.60372722e-01 1.36652720e+00
3.20596933e-01 -1.51125655e-01 2.73680717e-01 -6.91643178e-01
-1.23189664e+00 4.68570143e-01 -4.32199150e-01 2.10959196e-01
-5.01427837e-02 4.66556735e-02 4.94136870e-01 -9.97595727e-01
3.83245677e-01 1.42586696e+00 8.93902242e-01 6.94697499e-01
-1.27332020e+00 -3.92885447e-01 -9.78960916e-02 3.09423447e-01
-1.06712735e+00 -3.77136379e-01 6.07407987e-01 -3.26896459e-01
5.25189459e-01 3.40406448e-02 3.50956738e-01 1.10172462e+00
2.65715837e-01 7.13762760e-01 8.60497594e-01 -2.17164740e-01
-1.36032775e-01 -1.40588935e-02 -9.64057967e-02 5.11607528e-01
1.96782082e-01 1.37056381e-01 -3.45357060e-01 3.44252825e-01
1.25897491e+00 5.64710319e-01 -5.89602828e-01 1.10154226e-01
-9.49857593e-01 5.87021530e-01 7.30547309e-01 3.27182263e-01
-7.79573262e-01 4.20732081e-01 2.41433457e-01 4.36381996e-01
6.14167094e-01 1.18724249e-01 -3.79907489e-01 3.12988698e-01
-1.12288201e+00 2.52354234e-01 2.67938524e-01 5.84472597e-01
8.33044469e-01 3.50159138e-01 -1.34111106e-01 9.82943356e-01
1.47479549e-01 8.97776335e-02 6.34979963e-01 -1.19569552e+00
2.36409515e-01 3.34034741e-01 1.23985596e-01 -6.25474036e-01
-9.01773870e-02 -6.16101503e-01 -1.56542683e+00 5.90992928e-01
3.63535471e-02 -1.10149026e-01 -1.07008696e+00 1.59475124e+00
-1.28932238e-01 4.94426876e-01 2.90977001e-01 1.10658336e+00
7.81058669e-01 8.85953128e-01 -2.18149591e-02 -7.49200359e-02
1.29022670e+00 -1.08187318e+00 -9.71538901e-01 -3.57033908e-01
2.57424247e-02 -7.94979036e-01 7.97052741e-01 4.27060783e-01
-1.42601335e+00 -1.03078270e+00 -1.04047513e+00 -7.12356150e-01
-8.09503719e-02 2.36497834e-01 4.97499257e-01 -6.76868334e-02
-1.18926597e+00 1.17777956e+00 -8.96105170e-01 4.54328284e-02
7.07146466e-01 6.12575449e-02 -6.96503222e-01 -4.10398960e-01
-1.13694215e+00 9.85200584e-01 4.43133652e-01 4.96557444e-01
-1.19111145e+00 -8.86073947e-01 -1.04785502e+00 3.33928704e-01
-8.98841098e-02 -7.97262788e-01 9.87669945e-01 -1.34129632e+00
-1.47643185e+00 6.48524523e-01 -1.55304521e-01 -6.47412896e-01
4.96970445e-01 -4.94276702e-01 -3.11576009e-01 1.24708852e-02
-9.26225260e-02 4.54395384e-01 1.26033974e+00 -1.33719397e+00
-4.21437383e-01 -2.27919221e-01 -3.70197594e-02 -4.99476045e-02
-2.92285413e-01 6.78178063e-03 -3.13028514e-01 -9.79875028e-01
2.04321414e-01 -2.87526578e-01 -4.72317547e-01 2.76307464e-01
-2.11631224e-01 2.75842905e-01 8.41103971e-01 -1.35324442e+00
7.99687684e-01 -2.45966315e+00 4.95473862e-01 -2.96871603e-01
2.77816296e-01 5.40259719e-01 -3.82892758e-01 1.72469541e-01
-5.11325121e-01 -1.16241373e-01 -5.16949058e-01 -7.14919806e-01
-3.24757457e-01 4.41054702e-01 -4.72579718e-01 4.04093891e-01
5.29578507e-01 8.60389829e-01 -6.31678581e-01 2.46778727e-02
4.00737554e-01 1.05648398e+00 -6.11302376e-01 5.41670978e-01
-7.68440291e-02 6.79027796e-01 -8.33265856e-02 2.61501163e-01
1.09589720e+00 -1.84585571e-01 -1.93446308e-01 -4.84043092e-01
-2.23769158e-01 7.14774355e-02 -1.11756957e+00 1.84155846e+00
-7.08236039e-01 8.35484803e-01 4.53402191e-01 -1.20449686e+00
7.18034506e-01 3.43203366e-01 1.80387393e-01 -8.18367422e-01
1.16563685e-01 2.22185478e-01 -2.95542866e-01 -6.92961872e-01
4.68244791e-01 -2.77757823e-01 4.04079378e-01 5.87151423e-02
2.08907679e-01 1.16510302e-01 3.11658182e-03 -1.23690985e-01
9.41230416e-01 2.11720139e-01 -8.97736847e-02 -4.64398377e-02
6.78585231e-01 -3.60020846e-01 4.70529795e-01 4.08803433e-01
3.84432673e-01 1.09308267e+00 4.25980330e-01 -7.31635988e-01
-1.36227238e+00 -8.96510839e-01 -2.94701308e-02 5.41818798e-01
1.10067613e-02 -5.93910180e-02 -1.07018661e+00 3.00886761e-02
-1.12236694e-01 4.97351140e-01 -6.35975122e-01 -5.35110533e-01
-7.18876362e-01 -6.91827655e-01 3.97817552e-01 5.97588718e-01
9.75701392e-01 -1.21803701e+00 -3.08545589e-01 3.94014746e-01
-3.55520725e-01 -1.11803377e+00 -3.59029889e-01 2.71846801e-01
-1.06411684e+00 -9.40725744e-01 -1.10882759e+00 -1.13801956e+00
8.35562289e-01 4.24751312e-01 1.00043547e+00 3.89789373e-01
-3.20194453e-01 -3.77137810e-01 -3.03976655e-01 6.76975548e-02
-5.33427894e-01 -4.32413369e-01 -3.42398256e-01 3.24238032e-01
-2.14989990e-01 -8.73489439e-01 -7.43256807e-01 -9.25706252e-02
-1.49342179e+00 2.02574477e-01 8.26810718e-01 1.21548462e+00
6.33057356e-01 3.60773265e-01 1.48313299e-01 -8.69236469e-01
5.80964327e-01 -1.94064632e-01 -5.64831913e-01 3.05917319e-02
-2.73077965e-01 3.72579068e-01 1.00371456e+00 -1.85858399e-01
-1.21270156e+00 5.60413226e-02 -5.97970665e-01 -8.08406532e-01
-9.85161588e-02 4.50141400e-01 -1.96847603e-01 -1.95427805e-01
5.38030624e-01 5.57863116e-01 1.37459636e-01 -1.09839940e+00
4.08105791e-01 5.54478526e-01 1.04660964e+00 -6.88070804e-02
7.87197292e-01 5.55337846e-01 -2.25604236e-01 -6.28556728e-01
-1.01546836e+00 -6.38107955e-02 -5.88082671e-01 1.98236018e-01
1.05863428e+00 -1.27525723e+00 -4.39665765e-01 8.49200726e-01
-1.53085339e+00 -4.82102871e-01 -4.10394669e-01 4.19580728e-01
-3.70240748e-01 4.85059351e-01 -1.15606976e+00 -2.34030291e-01
-2.06419364e-01 -1.23907816e+00 9.19269621e-01 3.58834177e-01
4.39034224e-01 -8.37069631e-01 -1.67346776e-01 2.00843066e-01
6.04060888e-01 2.06608847e-01 7.82900810e-01 2.11860150e-01
-6.90063775e-01 1.09029952e-02 -5.78244925e-01 1.27744937e+00
1.33224353e-02 -3.17689002e-01 -1.08367324e+00 -3.82442296e-01
3.40190619e-01 -1.05847947e-01 1.48707080e+00 4.96069312e-01
1.73977089e+00 -5.20257533e-01 1.04462966e-01 1.22248137e+00
1.62011087e+00 -1.80921048e-01 1.34620440e+00 2.65478730e-01
7.71523178e-01 4.40864056e-01 -3.40507925e-02 2.42246121e-01
2.72362500e-01 2.68495739e-01 7.01540589e-01 -6.23293638e-01
-7.02140987e-01 -1.73907578e-01 3.88734967e-01 7.04453468e-01
-1.96766004e-01 9.10562277e-03 -3.00312072e-01 5.64873517e-01
-1.62397850e+00 -1.01804018e+00 -2.77516037e-01 1.97157347e+00
9.01188374e-01 -1.84587151e-01 -5.06499946e-01 1.52768955e-01
7.31239438e-01 1.78250581e-01 -4.25796241e-01 -2.64824808e-01
-1.79955840e-01 4.45026487e-01 4.81392592e-01 7.53867507e-01
-1.07745957e+00 8.46749246e-01 6.05842113e+00 6.44448698e-01
-1.21215904e+00 2.15693086e-01 7.57885575e-01 3.58829588e-01
-1.62602812e-01 -1.22621149e-01 -1.99963301e-01 5.35230398e-01
7.62230754e-01 3.57159406e-01 7.98333228e-01 5.40692925e-01
2.06136554e-01 2.47222975e-01 -9.73910213e-01 1.13675332e+00
-7.18166179e-04 -1.51636887e+00 2.82443315e-01 -2.30818555e-01
8.92538726e-01 -1.09107204e-01 6.83344677e-02 4.16863039e-02
1.82384267e-01 -1.34231770e+00 8.20861638e-01 7.67502129e-01
8.59080434e-01 -7.15506911e-01 9.30356443e-01 3.41486961e-01
-7.70055771e-01 -1.84672773e-01 -7.53527462e-01 -1.41906053e-01
2.16395676e-01 7.70003974e-01 3.98789458e-02 6.28239632e-01
8.94609451e-01 1.03227365e+00 -1.88832775e-01 1.04219961e+00
-4.80293781e-01 2.75055498e-01 3.49496782e-01 7.48181999e-01
4.58700284e-02 -2.60021389e-01 2.10753784e-01 1.22987533e+00
3.58604282e-01 1.39265925e-01 -2.19855428e-01 1.06606209e+00
-4.11448121e-01 -6.34067953e-01 -2.58750111e-01 1.20008871e-01
-1.64635539e-01 1.23161685e+00 -1.58075944e-01 -3.73940378e-01
-3.88752848e-01 1.53093064e+00 2.58786261e-01 6.42386496e-01
-7.17971265e-01 -5.22921562e-01 9.61369097e-01 1.63496390e-01
5.08270919e-01 -1.79653510e-01 -3.92228901e-01 -1.42465889e+00
5.61710037e-02 -8.50241125e-01 -1.09521285e-01 -1.04658377e+00
-1.24581373e+00 8.76698077e-01 -5.94118416e-01 -1.13219738e+00
3.51552144e-02 -7.93646991e-01 -6.90625310e-01 1.22258425e+00
-2.16436076e+00 -9.77768898e-01 -6.45475447e-01 9.12801802e-01
5.34827888e-01 1.53748900e-01 6.93588674e-01 8.16735804e-01
-5.48634052e-01 2.84239620e-01 4.09630537e-01 3.11083227e-01
4.62349057e-01 -1.09195399e+00 4.00827527e-01 1.15954089e+00
-2.31894240e-01 4.83083695e-01 7.25110650e-01 -2.99074948e-01
-1.17003560e+00 -1.39670539e+00 7.52822816e-01 2.24536240e-01
3.49509239e-01 -2.35712990e-01 -1.32535946e+00 7.33837068e-01
3.26838017e-01 3.56296510e-01 -4.63356562e-02 -4.18653667e-01
-4.20765728e-01 -3.30646127e-01 -1.13202083e+00 5.10424852e-01
7.77199924e-01 -7.01145291e-01 -4.16737229e-01 3.79020184e-01
8.44454110e-01 -5.26134968e-01 -7.55275905e-01 1.34721622e-01
1.59173831e-01 -1.16092455e+00 1.34131551e+00 -3.69359225e-01
1.02822900e+00 -4.99721080e-01 5.62302908e-03 -1.41679740e+00
-6.02340877e-01 -2.98846185e-01 5.87588139e-02 1.07103372e+00
1.54751599e-01 -4.68439132e-01 5.18389285e-01 1.65723592e-01
-4.69687343e-01 -4.66254979e-01 -6.88975573e-01 -4.45877165e-01
-1.41889909e-02 1.99469067e-02 4.99979198e-01 7.82758772e-01
-6.20676517e-01 8.95400941e-02 -7.89691567e-01 3.96506995e-01
8.09380889e-01 -1.09584369e-02 3.81621748e-01 -1.06037343e+00
-2.41400927e-01 -3.25908214e-01 -2.25863010e-01 -1.28086507e+00
5.96085787e-02 -6.86859369e-01 2.20682576e-01 -1.77667713e+00
2.34670311e-01 7.15544373e-02 -2.34890163e-01 4.54084158e-01
-1.67702824e-01 5.60049117e-01 -9.08736289e-02 2.33010009e-01
-1.61902517e-01 7.15958178e-01 1.67452550e+00 -2.06754431e-01
1.28064394e-01 -5.04636206e-02 -8.11394095e-01 6.34189367e-01
5.30160725e-01 -3.28687698e-01 -1.24596924e-01 -1.08190155e+00
-1.49541572e-01 1.91996068e-01 8.69092584e-01 -9.21198785e-01
2.28358760e-01 1.76617041e-01 7.87885666e-01 -2.79278100e-01
3.55806112e-01 -8.37188780e-01 3.89239222e-01 7.18376338e-01
-3.84318054e-01 -1.93210438e-01 5.87088689e-02 3.91823739e-01
-8.06810915e-01 -2.88756937e-01 1.13328433e+00 -5.09900808e-01
-7.27127314e-01 4.91648555e-01 -2.43143827e-01 -3.77215832e-01
6.08276904e-01 -8.35027769e-02 -1.23408899e-01 -3.40298504e-01
-8.04210424e-01 -4.62216437e-02 3.66277456e-01 2.54003048e-01
8.97640109e-01 -1.26843631e+00 -1.04028547e+00 4.43374962e-01
-4.58179921e-01 3.57953995e-01 6.75525248e-01 6.04907751e-01
-9.63972747e-01 -6.30564243e-02 -5.73856950e-01 -3.13962877e-01
-8.68005931e-01 5.79812169e-01 6.42879367e-01 -1.79607689e-01
-9.72767234e-01 9.41572905e-01 3.08479160e-01 -1.52359843e-01
3.10835123e-01 -4.12800759e-01 -2.65452087e-01 -3.02866817e-01
1.02848446e+00 2.50257492e-01 1.50963757e-02 -4.52486575e-01
4.63280194e-02 4.97245669e-01 -1.50956601e-01 3.19876492e-01
1.85425866e+00 -2.10018262e-01 -6.73230529e-01 -1.49528563e-01
1.35287809e+00 -4.33221221e-01 -1.82108760e+00 -3.60747188e-01
-2.02262029e-01 -4.63610858e-01 4.77581441e-01 -6.37095928e-01
-1.59562409e+00 1.24495041e+00 6.24160767e-01 4.68882620e-02
1.41279566e+00 -2.77248472e-01 9.83592749e-01 -3.59531678e-02
-4.32460904e-02 -7.90970147e-01 2.17272311e-01 5.31948745e-01
1.16696763e+00 -1.06285644e+00 -1.17961153e-01 -2.83015400e-01
-3.28271151e-01 1.40735054e+00 4.38074112e-01 -5.95255971e-01
5.68459928e-01 2.98947185e-01 -1.56342953e-01 3.93916965e-02
-5.49266040e-01 -1.76317915e-01 4.88234349e-02 5.43219984e-01
3.56818885e-01 -2.25342587e-01 -8.15963559e-03 6.63804770e-01
1.35628790e-01 3.69199812e-01 6.49893880e-01 4.22650546e-01
-4.05452996e-01 -1.01445496e+00 -4.21391726e-01 5.43794334e-02
-6.29981577e-01 -4.19225663e-01 2.44079590e-01 3.54166895e-01
3.71985286e-01 7.20184505e-01 1.13666102e-01 -1.95989221e-01
3.98526281e-01 -3.34404200e-01 3.87210041e-01 -4.31941777e-01
-5.00518501e-01 -2.64416579e-02 -3.63844246e-01 -6.46723509e-01
-3.83999139e-01 -2.49483302e-01 -1.10098016e+00 -4.63849306e-01
4.02303692e-03 -6.17905296e-02 6.24705434e-01 9.37774658e-01
2.89375186e-01 1.14347029e+00 6.48228526e-01 -1.10457170e+00
-5.20869732e-01 -9.29184377e-01 -7.32092559e-01 5.01087546e-01
9.80260134e-01 -2.44027466e-01 -2.60474294e-01 4.52732295e-01] | [11.281986236572266, -2.2546942234039307] |
d3af98e9-643a-4d19-aaf8-06e7727410b2 | text-only-training-for-image-captioning-using | 2211.00575 | null | https://arxiv.org/abs/2211.00575v1 | https://arxiv.org/pdf/2211.00575v1.pdf | Text-Only Training for Image Captioning using Noise-Injected CLIP | We consider the task of image-captioning using only the CLIP model and additional text data at training time, and no additional captioned images. Our approach relies on the fact that CLIP is trained to make visual and textual embeddings similar. Therefore, we only need to learn how to translate CLIP textual embeddings back into text, and we can learn how to do this by learning a decoder for the frozen CLIP text encoder using only text. We argue that this intuition is "almost correct" because of a gap between the embedding spaces, and propose to rectify this via noise injection during training. We demonstrate the effectiveness of our approach by showing SOTA zero-shot image captioning across four benchmarks, including style transfer. Code, data, and models are available on GitHub. | ['Amir Globerson', 'Ron Mokady', 'David Nukrai'] | 2022-11-01 | null | null | null | null | ['semi-supervised-learning-for-image-captioning'] | ['computer-vision'] | [ 5.68847954e-01 4.52295601e-01 -2.18085766e-01 -3.95285159e-01
-9.33527231e-01 -7.24337697e-01 5.05214930e-01 -1.48199692e-01
-3.61835063e-01 6.61071897e-01 4.65416104e-01 -2.33887821e-01
6.67154968e-01 -3.47330719e-01 -1.34231925e+00 -2.65082866e-01
4.22884375e-01 3.65837336e-01 8.28542095e-03 -7.63285980e-02
6.42432347e-02 -7.09184706e-02 -9.68662262e-01 5.17576873e-01
5.87992907e-01 7.29263604e-01 1.87545180e-01 8.87901723e-01
-1.77382529e-01 8.56754422e-01 -3.74387145e-01 -8.34681571e-01
3.13256711e-01 -6.26304507e-01 -7.18307614e-01 2.92455554e-01
7.55948186e-01 -7.41027653e-01 -6.55241311e-01 9.32764411e-01
3.05788636e-01 -2.21092463e-01 6.56307340e-01 -1.32316959e+00
-1.21995902e+00 5.33658743e-01 -5.57568014e-01 -3.35843533e-01
2.25192353e-01 8.45319852e-02 1.13626754e+00 -1.09030950e+00
1.01628876e+00 8.66019607e-01 5.50317049e-01 9.40375566e-01
-1.47863352e+00 -3.78047347e-01 6.07521879e-03 -2.88807824e-02
-1.01027656e+00 -5.68749666e-01 6.39134884e-01 -5.82017839e-01
7.37512171e-01 8.75033997e-03 5.71331859e-01 1.39733362e+00
-1.50522860e-02 8.53690505e-01 6.73160672e-01 -5.09451270e-01
1.45679414e-01 3.82592648e-01 -2.23265037e-01 5.26433110e-01
9.82126594e-02 -5.62957413e-02 -5.05680323e-01 1.20466582e-01
7.35252738e-01 -6.74961209e-02 -3.72781038e-01 -7.57437825e-01
-1.00832808e+00 8.94053519e-01 5.17909586e-01 -1.79821625e-02
2.70098746e-02 6.46194816e-01 5.16879022e-01 2.53011495e-01
5.50275803e-01 4.12524492e-01 -2.95790285e-01 -2.42799237e-01
-9.07610416e-01 7.44594727e-04 7.08659649e-01 1.34484255e+00
7.03899086e-01 -1.31355956e-01 -2.37807006e-01 7.79532433e-01
-2.09476147e-02 7.14452803e-01 2.06087992e-01 -1.10210168e+00
6.38212442e-01 1.70752201e-02 3.18297505e-01 -6.55984163e-01
1.76431686e-01 -5.24277724e-02 -3.99179518e-01 1.35390341e-01
2.43043438e-01 -3.72786075e-01 -1.30866480e+00 1.89064300e+00
-3.75917137e-01 1.77191168e-01 5.84339947e-02 8.65162969e-01
3.35037649e-01 8.96965146e-01 1.00564837e-01 1.79494590e-01
1.17204988e+00 -1.24038720e+00 -8.35470796e-01 -6.81016505e-01
7.53010511e-01 -6.92777693e-01 1.30143785e+00 -2.80516129e-02
-1.18717885e+00 -2.88008124e-01 -1.17684221e+00 -6.09112978e-01
-3.22365910e-01 1.69420883e-01 2.28556171e-01 1.91484734e-01
-1.07776248e+00 4.38114673e-01 -7.47643352e-01 -5.92544913e-01
5.04127324e-01 4.32872698e-02 -5.30098557e-01 -3.07896048e-01
-1.00196874e+00 8.48323047e-01 2.26108789e-01 -1.18217349e-01
-8.60802233e-01 -6.16394520e-01 -1.06906104e+00 2.80527323e-01
1.64853603e-01 -7.11718202e-01 1.38277698e+00 -1.59405613e+00
-1.27461076e+00 9.52486455e-01 -3.04102093e-01 -4.66791213e-01
6.86908722e-01 -3.92176211e-01 6.30082637e-02 4.83023793e-01
7.49365315e-02 1.21455741e+00 9.05742705e-01 -1.56681013e+00
-3.18885326e-01 9.80932638e-02 2.04332218e-01 1.62478641e-01
-4.87972230e-01 -5.32185063e-02 -8.30233216e-01 -6.68698370e-01
-4.44938809e-01 -1.15358496e+00 -1.90794580e-02 6.56585336e-01
-1.95208177e-01 3.70208710e-01 8.51801276e-01 -9.85550404e-01
8.02244961e-01 -2.46179986e+00 1.82300642e-01 -1.34010226e-01
7.61624351e-02 2.02133387e-01 -4.40261275e-01 4.41008240e-01
-2.50103593e-01 3.50023776e-01 -5.60922086e-01 -7.67342746e-01
5.23254648e-02 2.99463063e-01 -8.05056810e-01 1.71293080e-01
4.38386738e-01 1.29133141e+00 -8.68611634e-01 -5.75322688e-01
2.37970755e-01 6.53462350e-01 -8.19360614e-01 3.31036657e-01
-4.66262013e-01 1.33984819e-01 -1.71221912e-01 1.05072983e-01
7.39616036e-01 -3.02010596e-01 2.10530549e-01 -2.40350589e-01
9.14266929e-02 1.14076599e-01 -7.02057540e-01 2.05606222e+00
-5.71228206e-01 1.18871856e+00 -2.98595317e-02 -8.56329203e-01
4.98233527e-01 4.48911756e-01 1.82286829e-01 -7.39981413e-01
1.13171332e-01 1.93142518e-01 -5.34751773e-01 -5.77513278e-01
3.69001955e-01 -2.75335819e-01 8.40158761e-03 5.42747796e-01
1.96037292e-01 -2.61564493e-01 -9.28396732e-03 4.94621605e-01
9.30714667e-01 5.29237449e-01 -3.75789940e-01 1.51560724e-01
-3.95625606e-02 1.53902426e-01 1.29799381e-01 5.90004444e-01
8.78859535e-02 1.17521584e+00 7.98109829e-01 -2.20383823e-01
-1.82094049e+00 -1.24010789e+00 1.35622159e-01 9.76668715e-01
4.16656807e-02 -4.58420813e-01 -9.70366478e-01 -7.01087356e-01
-1.73199043e-01 9.42698359e-01 -1.00935686e+00 -1.13072880e-01
-6.64501190e-01 7.04550222e-02 4.05759573e-01 7.67008364e-01
2.20517933e-01 -9.16645944e-01 -3.82905930e-01 1.71536263e-02
-2.87764639e-01 -1.31754518e+00 -9.94512618e-01 1.25458866e-01
-7.65188634e-01 -7.37982094e-01 -1.05839181e+00 -1.09310460e+00
1.00785303e+00 3.42953265e-01 1.12985849e+00 2.16516644e-01
-1.58211768e-01 4.60327059e-01 -3.44012171e-01 -2.53531158e-01
-4.42601800e-01 1.29826501e-01 -6.61789775e-01 -8.91784504e-02
6.95865080e-02 -5.25771439e-01 -5.28917074e-01 -2.54782736e-02
-1.17783391e+00 4.87663269e-01 6.75621748e-01 9.02766466e-01
2.77192026e-01 -7.80512035e-01 2.58685619e-01 -9.56672907e-01
3.65959704e-01 -2.83825397e-01 -5.17746091e-01 4.26623076e-01
-2.18321860e-01 2.45691165e-01 8.55620861e-01 -3.70798558e-01
-9.10191059e-01 3.18972737e-01 -1.10462762e-01 -7.85019696e-01
6.07740227e-03 2.38730952e-01 1.61124356e-02 1.94652572e-01
3.83267969e-01 1.53931394e-01 2.71035749e-02 -3.88466388e-01
9.01103854e-01 7.44808376e-01 6.34821773e-01 -6.70374930e-01
9.76990700e-01 6.25491560e-01 -4.20042396e-01 -5.02650201e-01
-9.91184831e-01 -5.53685687e-02 -5.34076035e-01 -1.33572258e-02
1.21673560e+00 -1.05941451e+00 -3.59577686e-01 4.01790924e-02
-1.57442117e+00 -5.49802423e-01 -5.07494748e-01 2.93552846e-01
-8.71688962e-01 2.52381444e-01 -7.24406004e-01 -3.86776328e-01
-3.46257210e-01 -9.54355776e-01 1.15454519e+00 2.57384162e-02
-8.02285820e-02 -1.07516038e+00 6.26665130e-02 1.28473163e-01
3.95163327e-01 1.36636272e-01 8.82743239e-01 -3.51370513e-01
-6.20661974e-01 -2.14866489e-01 -6.07733369e-01 5.79184949e-01
-4.93480600e-02 -9.87835750e-02 -1.04411435e+00 -3.70744675e-01
-1.34946331e-01 -6.66623294e-01 9.87562358e-01 1.42693490e-01
1.08992052e+00 -3.97610605e-01 -3.33287686e-01 8.60150397e-01
1.77117169e+00 -7.91580826e-02 8.74204457e-01 1.46751478e-01
9.44705427e-01 5.29286444e-01 2.25007981e-01 1.37473255e-01
3.88064355e-01 6.33334756e-01 2.91984379e-01 -3.36767703e-01
-3.51615995e-01 -9.02634740e-01 4.22204465e-01 6.60630822e-01
4.52456206e-01 -4.75185961e-01 -6.11392081e-01 7.97539353e-01
-1.84340453e+00 -9.32984948e-01 2.36505732e-01 2.06343365e+00
9.71893907e-01 2.30248168e-01 -2.44661480e-01 -3.54257256e-01
6.50781691e-01 1.09046333e-01 -5.26798248e-01 -6.45042777e-01
-4.70448062e-02 2.33728468e-01 8.69828999e-01 7.88033664e-01
-9.12946403e-01 1.03606415e+00 6.92743015e+00 4.59322751e-01
-1.34428132e+00 2.86882550e-01 6.69626772e-01 -1.50047883e-01
-5.23335516e-01 3.63781482e-01 -2.99164563e-01 5.40935099e-01
8.75169456e-01 -3.45687926e-01 4.94831830e-01 7.39181042e-01
1.47907555e-01 3.96176837e-02 -1.43861151e+00 7.98233271e-01
4.22003835e-01 -1.48577321e+00 3.05854410e-01 -2.19753534e-01
8.54573548e-01 1.80585764e-03 2.13351056e-01 2.10351050e-01
2.11905107e-01 -1.08955872e+00 8.33720744e-01 2.97551364e-01
1.12598455e+00 -2.98537821e-01 4.86120552e-01 -1.30616695e-01
-8.18848729e-01 1.69893906e-01 -3.52784634e-01 -2.12289789e-03
4.23862755e-01 2.07207933e-01 -6.47022784e-01 4.38851029e-01
2.85910398e-01 8.97253454e-01 -6.87598348e-01 9.23037946e-01
-4.65343922e-01 5.07097185e-01 -9.61515307e-02 3.35989654e-01
1.04002394e-01 -2.07164697e-02 1.48012295e-01 1.27538419e+00
4.15225416e-01 -1.64749444e-01 -1.58883154e-01 1.02935958e+00
-3.95065337e-01 -9.03008357e-02 -6.77336395e-01 -2.34629318e-01
2.11733878e-01 1.02302480e+00 -5.07261515e-01 -4.98221159e-01
-7.43053615e-01 1.57541013e+00 2.92769790e-01 5.62154651e-01
-1.16449893e+00 -5.89427233e-01 5.28124213e-01 1.98798403e-01
5.79030156e-01 -3.48813236e-01 -4.65879291e-01 -1.37773013e+00
2.16617033e-01 -6.44374371e-01 -6.76171184e-02 -1.42888737e+00
-1.05611300e+00 4.84206617e-01 -1.69473320e-01 -1.20239699e+00
-7.52278790e-02 -6.37300670e-01 -6.47717655e-01 6.96365476e-01
-1.51189280e+00 -1.31671762e+00 -3.35120201e-01 3.27109873e-01
6.34058416e-01 5.76514840e-01 6.54704571e-01 3.27398807e-01
-2.15448245e-01 8.55468750e-01 2.24766985e-01 4.50574100e-01
1.21849227e+00 -1.14286005e+00 6.80164278e-01 6.95767283e-01
1.27485588e-01 3.63165021e-01 1.02957273e+00 -3.88084173e-01
-1.31145275e+00 -1.16372299e+00 8.56643558e-01 -6.25909090e-01
6.51510060e-01 -9.20720994e-01 -9.07052755e-01 1.15629935e+00
7.41499305e-01 6.19446449e-02 3.90268892e-01 -3.72677118e-01
-6.83070838e-01 1.43126711e-01 -7.43968189e-01 7.34887183e-01
9.89092171e-01 -8.25458467e-01 -6.40608788e-01 4.88751203e-01
1.23831844e+00 -4.41586941e-01 -5.54332972e-01 -1.32275805e-01
5.30506134e-01 -5.34520268e-01 7.42235005e-01 -5.41531265e-01
1.04328716e+00 -1.90984309e-01 -9.91363004e-02 -1.32451546e+00
4.18807045e-02 -4.62659627e-01 2.27635175e-01 1.08025229e+00
7.38252103e-01 -3.57354611e-01 7.71001041e-01 7.55586207e-01
-1.48762271e-01 -5.54353237e-01 -7.22767293e-01 -7.82638788e-01
4.42550182e-01 -1.82148874e-01 2.72166222e-01 8.59203815e-01
1.21780686e-01 6.79535508e-01 -7.50795245e-01 -9.57876742e-02
4.49591428e-01 -1.21909067e-01 7.89643884e-01 -6.27081156e-01
-3.87907356e-01 -5.92350103e-02 -9.38468650e-02 -1.35449529e+00
2.28549451e-01 -8.51556718e-01 2.83774525e-01 -1.70404780e+00
5.28513372e-01 -9.31746960e-02 3.53967138e-02 7.01500654e-01
-1.09879106e-01 5.97596943e-01 6.19740009e-01 3.08319628e-01
-6.50949419e-01 6.69727683e-01 1.28605723e+00 -1.79623783e-01
1.96114764e-01 -7.12761462e-01 -7.52695739e-01 3.32806289e-01
5.69263577e-01 -5.29755414e-01 -4.35542941e-01 -1.07109833e+00
3.37742180e-01 3.14661325e-03 5.88947713e-01 -8.68241608e-01
1.67277277e-01 4.73688543e-03 4.41756248e-01 -1.48013785e-01
5.59305847e-01 -9.02751207e-01 -9.87877250e-02 2.42367819e-01
-7.53172219e-01 -9.49539244e-02 3.51222724e-01 4.62152630e-01
-2.10768700e-01 -3.78765941e-01 7.98154891e-01 1.78607568e-01
-3.98449868e-01 2.31532529e-01 -4.44562770e-02 2.21789673e-01
9.15518105e-01 -3.82596366e-02 -5.31870008e-01 -8.11026990e-01
-8.81507754e-01 2.73911357e-01 1.11614907e+00 5.64945996e-01
5.44037938e-01 -1.40059566e+00 -5.94938934e-01 1.20160371e-01
2.35340118e-01 -2.49148801e-01 9.23079476e-02 5.97382069e-01
-6.10141635e-01 3.24454695e-01 -2.81258732e-01 -4.32665735e-01
-8.68728757e-01 8.55807722e-01 3.05179417e-01 1.22019269e-01
-6.57654881e-01 5.07575154e-01 6.03935659e-01 -4.74829115e-02
2.59243608e-01 -6.85717165e-02 4.03447092e-01 -3.20850402e-01
4.75250930e-01 -3.73613983e-01 -3.27615559e-01 -3.16706061e-01
-6.79028705e-02 7.75855660e-01 -2.05879599e-01 -5.64768851e-01
1.20259798e+00 -3.42629194e-01 1.76663399e-01 3.28155667e-01
1.69043684e+00 7.56620755e-03 -1.72119641e+00 -1.73365436e-02
-2.32630789e-01 -5.11499345e-01 -3.76075566e-01 -7.03514934e-01
-9.28256273e-01 1.31518674e+00 4.91597831e-01 -1.43726304e-01
1.09290504e+00 6.72346400e-03 9.50822353e-01 2.61222064e-01
-6.44960850e-02 -1.12891543e+00 2.92620927e-01 4.47718054e-01
1.03697848e+00 -1.20468187e+00 -1.84918538e-01 -2.73675680e-01
-8.31456006e-01 1.08127737e+00 5.53214133e-01 -3.69019747e-01
4.23457593e-01 2.85559863e-01 3.35321715e-03 1.98402748e-01
-9.41781461e-01 -2.70133112e-02 -1.98193073e-01 5.71047723e-01
5.00773966e-01 -2.30145723e-01 -8.21318924e-02 2.98406094e-01
-1.82946231e-02 4.09137696e-01 8.20381403e-01 9.68807936e-01
-2.92340338e-01 -1.21530485e+00 -2.95902807e-02 1.93468958e-01
-2.99095929e-01 -3.21824938e-01 -4.40938354e-01 5.69013596e-01
-5.93247376e-02 6.37803555e-01 2.80340374e-01 -3.03610086e-01
2.08533779e-01 2.07335949e-01 4.95322794e-01 -6.33701563e-01
1.60283484e-02 -4.65377830e-02 -4.02508229e-02 -4.55270529e-01
-4.90787812e-02 -1.71871334e-01 -1.04757547e+00 -2.58634120e-01
-1.14526041e-01 1.56714469e-01 8.61714900e-01 7.25355029e-01
6.06571317e-01 3.98819208e-01 5.73570013e-01 -8.89441848e-01
-2.71296084e-01 -7.56486952e-01 -1.73743933e-01 7.30518162e-01
3.89358342e-01 -2.58719325e-01 -6.05381906e-01 6.70437753e-01] | [11.142629623413086, 0.730832576751709] |
97aab74f-c3ce-4b19-a8a8-4443401b49d8 | how-well-do-multi-hop-reading-comprehension | null | null | https://openreview.net/forum?id=Ougcw7mdk8u | https://openreview.net/pdf?id=Ougcw7mdk8u | How Well Do Multi-hop Reading Comprehension Models Understand Date Information? | Many previous works demonstrated that existing multi-hop reading comprehension datasets (e.g., HotpotQA) contain reasoning shortcuts, where the questions can be answered without performing multi-hop reasoning. Recently, several multi-hop datasets have been proposed to solve the reasoning shortcut problem or evaluate the internal reasoning process. However, the design of the reasoning chain for comparison questions in R4C and 2WikiMultiHopQA does not fully explain the answer; meanwhile, MuSiQue only focuses on bridge questions. Therefore, it is unclear about the ability of a model to perform step-by-step reasoning when finding an answer for a comparison question that requires comparison and numerical reasoning skills. To evaluate the model completely in a hierarchical manner, we first propose a dataset, HieraDate, created by reusing and enhancing two previous multi-hop datasets, HotpotQA and 2WikiMultiHopQA. Our dataset focuses on comparison questions on date information that require multi-hop reasoning for solving. We then evaluate the ability of existing models to understand date at three levels: extraction, reasoning, and robustness. Our experimental results reveal that the multi-hop models fail at the reasoning level. Comparison reasoning and numerical reasoning (e.g., subtraction) are key challenges that need to be addressed in future works. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['multi-hop-reading-comprehension'] | ['natural-language-processing'] | [-2.13638797e-01 5.56322694e-01 8.77104606e-03 -5.21515250e-01
-8.57799113e-01 -8.13944101e-01 2.93858349e-01 4.95842546e-01
-1.35542929e-01 5.48900604e-01 8.03219974e-02 -8.54043126e-01
-8.14622819e-01 -1.30477333e+00 -8.08525741e-01 2.57662654e-01
4.26465780e-01 6.84256971e-01 5.35757184e-01 -7.84002483e-01
6.49440229e-01 -1.13333594e-02 -1.72688663e+00 7.36578941e-01
1.53743815e+00 6.82418585e-01 9.18964297e-02 9.06085134e-01
-6.79633915e-01 1.50208390e+00 -9.13420737e-01 -7.13013530e-01
6.91761300e-02 -5.98559856e-01 -1.56262946e+00 -8.02375019e-01
5.90838790e-01 -7.20253170e-01 -1.50445148e-01 8.93506646e-01
1.82679713e-01 5.37672965e-03 2.36334905e-01 -1.58217108e+00
-1.04071975e+00 1.25792181e+00 -1.82061613e-01 2.46362761e-01
1.17414320e+00 1.64492857e-02 1.23163509e+00 -4.73002344e-01
6.34621084e-01 1.41168833e+00 7.18243539e-01 4.30271566e-01
-4.29729879e-01 -5.84442198e-01 3.66524547e-01 1.06217575e+00
-9.64013934e-01 9.54692140e-02 6.94645286e-01 -1.60417706e-01
1.09876585e+00 7.16391146e-01 4.62367892e-01 7.14034915e-01
-4.66285907e-02 9.53037798e-01 1.25319469e+00 -3.54493976e-01
5.00983894e-02 -1.99532494e-01 7.66069114e-01 8.31126392e-01
1.34263024e-01 -4.41677749e-01 -7.95999706e-01 2.99912423e-01
4.63072479e-01 -4.82952595e-01 -3.77783626e-01 1.53907582e-01
-1.08089113e+00 6.01478577e-01 4.20165718e-01 2.19292104e-01
-6.46701381e-02 -1.70699209e-01 1.42627865e-01 7.51713336e-01
-3.89236003e-01 8.83831620e-01 -8.64216506e-01 -3.98697227e-01
-8.24172616e-01 6.53521538e-01 1.18153918e+00 1.26148403e+00
6.29642487e-01 -6.14688218e-01 -3.43298852e-01 3.40751022e-01
1.83697894e-01 4.41564411e-01 1.87046647e-01 -1.44094574e+00
9.15820599e-01 1.20363331e+00 1.49330452e-01 -1.07502234e+00
-7.10912168e-01 -1.33653879e-01 -2.50082731e-01 -1.24088973e-01
7.89683402e-01 -1.89017039e-02 -5.00341356e-01 1.73450780e+00
2.99920768e-01 -1.32853806e-01 3.93155426e-01 9.45808113e-01
1.61751521e+00 5.00916898e-01 -1.94851011e-01 8.57197866e-02
1.67928040e+00 -1.21496320e+00 -1.04050994e+00 -1.62314638e-01
7.30024576e-01 -6.74350083e-01 1.49536622e+00 5.32876313e-01
-1.54945779e+00 -5.62179565e-01 -1.07856500e+00 -8.05519760e-01
-7.95601249e-01 -1.83108851e-01 6.17727757e-01 3.94559860e-01
-7.57566571e-01 3.55527312e-01 -2.77393192e-01 -3.00732255e-01
-1.42400235e-01 -1.27744496e-01 4.37966175e-02 -4.21790183e-01
-1.82067025e+00 1.24883878e+00 5.23270249e-01 -6.12474903e-02
-2.70169437e-01 -1.15195227e+00 -8.35244775e-01 2.89456129e-01
8.95840406e-01 -8.91709507e-01 1.62238073e+00 -5.19662164e-02
-1.18438029e+00 5.84852397e-01 -1.33487388e-01 -3.33782375e-01
7.47453153e-01 -4.36095864e-01 -4.53126132e-01 5.41037261e-01
3.93065631e-01 6.93443358e-01 8.92077461e-02 -1.01154327e+00
-7.20694661e-01 -3.98651809e-01 1.10819328e+00 1.35861531e-01
1.46123976e-01 -1.35730222e-01 -4.69556779e-01 -4.01042193e-01
4.90262359e-01 -3.84891599e-01 4.19719249e-01 3.99849005e-03
-3.62794459e-01 -5.21836877e-01 7.68361449e-01 -1.09328377e+00
1.64899421e+00 -1.45216632e+00 2.01374337e-01 -7.84363225e-02
3.53143543e-01 -2.53116898e-02 -2.87513018e-01 6.41860604e-01
1.42086223e-01 3.85914117e-01 -1.20759029e-02 3.57400984e-01
3.31330866e-01 2.60229230e-01 -4.36580181e-01 -5.35156131e-01
1.79794684e-01 1.16480756e+00 -1.02560055e+00 -7.42717206e-01
-4.92337719e-02 -1.84145242e-01 -7.31672227e-01 1.95928007e-01
-8.34360778e-01 -3.89332026e-02 -4.96926337e-01 9.25990283e-01
8.43861282e-01 -2.35338315e-01 7.73621872e-02 -2.52965242e-01
-8.52498934e-02 5.72425127e-01 -1.26330090e+00 1.98635662e+00
-4.41212088e-01 4.35104549e-01 -3.56369406e-01 -5.42005420e-01
8.06687772e-01 -3.60360369e-02 1.43010365e-02 -1.19183373e+00
-1.90927476e-01 3.45609635e-01 2.05819100e-01 -9.74383235e-01
8.08189809e-01 4.01867062e-01 -2.10318893e-01 4.64028090e-01
-2.76928872e-01 -4.20558602e-01 8.45593452e-01 6.64384604e-01
1.36472774e+00 3.40763867e-01 4.89200354e-02 1.94350351e-02
7.29894996e-01 5.53624213e-01 4.93070781e-01 9.80752349e-01
-7.27779567e-02 4.55894232e-01 8.57719719e-01 -1.91573516e-01
-5.17435431e-01 -1.09626007e+00 1.82230353e-01 9.07896936e-01
5.05446494e-01 -9.26839411e-01 -7.85356462e-01 -8.10566008e-01
-1.10656880e-01 1.28329384e+00 -3.61611933e-01 -4.40818705e-02
-6.90887809e-01 -9.50460732e-02 9.56439376e-01 7.47756600e-01
1.14324498e+00 -7.84947097e-01 -1.14256871e+00 1.62649825e-01
-9.82145965e-01 -1.24926913e+00 2.32757717e-01 -1.29721165e-01
-6.14271641e-01 -1.69584203e+00 -5.49743623e-02 -5.41301429e-01
4.74347860e-01 3.17694187e-01 1.45264316e+00 8.06661308e-01
7.48709589e-02 6.60169363e-01 -7.40185678e-01 -6.45382643e-01
-2.39622787e-01 1.20091870e-01 -5.83216548e-01 -1.13250530e+00
7.05318272e-01 -2.34443918e-01 -4.66693372e-01 5.18330932e-01
-8.37083340e-01 3.16677153e-01 4.44620639e-01 4.14864630e-01
1.51900724e-01 3.86234194e-01 5.65916240e-01 -8.23997676e-01
1.01564753e+00 -4.16532010e-01 -4.61774200e-01 1.04277825e+00
-5.76326072e-01 2.46808738e-01 7.54117608e-01 -1.38592795e-02
-1.27565432e+00 -8.26909006e-01 -1.58416212e-01 1.44397080e-01
1.09443389e-01 8.58729184e-01 -2.12317929e-01 3.54785949e-01
4.65364188e-01 -7.32937753e-02 -4.35804278e-01 -2.70505220e-01
5.67698896e-01 3.36406499e-01 7.24465072e-01 -1.12449729e+00
9.46305931e-01 1.17614716e-02 -2.57003725e-01 -1.72353581e-01
-1.23230195e+00 -2.00991318e-01 -4.16825950e-01 -3.43705565e-01
9.13367212e-01 -6.22429550e-01 -1.26944137e+00 2.58932799e-01
-1.40175450e+00 -2.80671209e-01 -1.90216377e-01 2.23369971e-01
-6.19648278e-01 2.51920313e-01 -5.99332988e-01 -5.36450326e-01
-1.60092801e-01 -9.30590272e-01 7.06165016e-01 4.80567008e-01
-7.03050077e-01 -8.19889128e-01 -3.08058977e-01 1.14021707e+00
3.51209223e-01 1.20118016e-03 1.75356853e+00 -5.11713684e-01
-9.63666618e-01 1.97410256e-01 -3.32042813e-01 -2.01506540e-01
-2.04335392e-01 1.59321174e-01 -6.48882449e-01 4.17224616e-01
3.23152617e-02 -6.06593728e-01 4.16832596e-01 -7.82893226e-02
1.25615060e+00 -1.45558670e-01 1.38307467e-01 1.70396477e-01
1.10797250e+00 2.78813273e-01 1.04969704e+00 7.34540045e-01
2.08295673e-01 8.61128330e-01 9.82168555e-01 1.34437844e-01
1.23083341e+00 1.97171032e-01 3.68407279e-01 3.77387375e-01
-8.66785944e-02 -5.63546777e-01 -7.57183880e-02 9.84458387e-01
1.07683182e-01 -1.83169618e-01 -1.20343363e+00 4.75963533e-01
-2.01039553e+00 -9.68387127e-01 -6.00171089e-01 1.49522102e+00
8.71849120e-01 2.15002552e-01 -3.07979584e-01 5.96234798e-01
1.91284373e-01 -9.94843021e-02 -6.71175241e-01 -6.44744992e-01
-2.25706145e-01 1.84215665e-01 -2.53000528e-01 5.26500225e-01
-3.41063172e-01 9.37707126e-01 5.88592339e+00 3.41940850e-01
-4.89788264e-01 -2.85305351e-01 -1.30874254e-02 2.24458769e-01
-8.68228436e-01 3.91267598e-01 -6.78403497e-01 -1.12103559e-01
5.89946747e-01 -1.79900512e-01 4.26191181e-01 3.43648165e-01
-9.25852433e-02 -5.51168799e-01 -1.41069138e+00 7.67890394e-01
9.59955677e-02 -1.39946449e+00 2.78196335e-01 -7.48573601e-01
4.23652351e-01 -9.27992821e-01 -1.51290894e-01 9.39278483e-01
3.73148322e-01 -9.77071404e-01 8.78653526e-01 7.89573133e-01
1.20034426e-01 -6.27629995e-01 7.51202047e-01 6.67787671e-01
-1.11678982e+00 -2.54958421e-01 -1.74835145e-01 -4.66424644e-01
9.30004716e-02 3.09701622e-01 -3.75632584e-01 1.13398159e+00
9.22816694e-01 3.02728236e-01 -1.05101228e+00 7.71408856e-01
-1.01796854e+00 1.87806666e-01 -4.49671261e-02 -1.83711842e-01
1.41132534e-01 -3.28580961e-02 1.72652468e-01 5.32600820e-01
4.68009025e-01 5.21927893e-01 -8.86758938e-02 1.19931602e+00
-7.63300508e-02 -1.00247711e-01 2.62441323e-03 -4.03655916e-02
8.25374663e-01 9.09719408e-01 -5.07946908e-01 -3.68170857e-01
-4.78049755e-01 5.77417195e-01 5.20930469e-01 2.81468809e-01
-1.22978139e+00 -7.20328629e-01 4.06701058e-01 8.45770463e-02
-2.56375782e-02 -1.57909274e-01 -6.44770503e-01 -1.15888464e+00
5.56479692e-01 -1.30506814e+00 8.16723943e-01 -1.64336145e+00
-1.12561679e+00 1.51295230e-01 3.39774668e-01 -7.76333570e-01
-2.21999615e-01 -5.39551318e-01 -6.72572136e-01 7.94475317e-01
-1.79671645e+00 -9.63936269e-01 -8.87811720e-01 5.87508142e-01
5.12376249e-01 4.24648821e-01 7.41518259e-01 1.56674176e-01
-4.44526583e-01 4.95770574e-01 -8.37730825e-01 7.27909878e-02
6.02792561e-01 -1.48367763e+00 2.19809860e-01 8.66654754e-01
-1.87639967e-01 1.01662290e+00 7.29015350e-01 -7.24715412e-01
-1.66233313e+00 -5.01952529e-01 1.13666594e+00 -6.25544786e-01
7.90717781e-01 9.35617313e-02 -1.22561800e+00 6.50317490e-01
4.20062155e-01 -9.01033342e-01 4.66678351e-01 1.95413798e-01
-6.55374587e-01 1.50708975e-02 -1.26475036e+00 9.05000031e-01
1.22229826e+00 -5.56463361e-01 -1.54875493e+00 2.93281734e-01
1.08887327e+00 -8.12029839e-01 -9.12417650e-01 4.69968796e-01
4.32966560e-01 -1.16892374e+00 8.49930346e-01 -7.73891151e-01
1.07500982e+00 -5.27747691e-01 -1.49979830e-01 -9.99999106e-01
8.81226510e-02 -3.76232982e-01 -3.47861737e-01 1.32236695e+00
6.27124250e-01 -3.74017835e-01 5.60345173e-01 1.03948855e+00
-1.30983561e-01 -6.93415463e-01 -4.07655269e-01 -6.69318557e-01
3.04891944e-01 -7.36296713e-01 1.13273680e+00 8.91198039e-01
2.04694554e-01 2.36623600e-01 2.59103656e-01 5.22310019e-01
1.35917723e-01 4.61331785e-01 8.12941611e-01 -1.20468521e+00
-1.24000706e-01 -4.12483692e-01 1.65390790e-01 -1.24520350e+00
1.40405670e-01 -7.36622989e-01 -2.52249122e-01 -2.44576859e+00
-1.71804786e-01 -1.97899327e-01 1.87692508e-01 5.04185677e-01
-3.31407845e-01 -5.21293402e-01 3.47468495e-01 -3.46445478e-02
-6.67406440e-01 7.59696066e-02 1.53017044e+00 -2.71487087e-01
-4.33628634e-03 -3.54532927e-01 -1.02713990e+00 8.40887606e-01
8.73105586e-01 -1.69761658e-01 -8.62844050e-01 -9.10160363e-01
9.80676293e-01 4.95205343e-01 4.62433964e-01 -9.78526175e-01
8.19861233e-01 -3.33166331e-01 1.73352256e-01 -1.11215174e+00
1.44938082e-01 -6.83501959e-01 -2.91016519e-01 3.55963111e-01
-4.89380926e-01 6.98894262e-01 2.30816126e-01 -6.39697686e-02
-4.24197942e-01 -6.45726264e-01 1.57244951e-01 -3.68466854e-01
-1.01453102e+00 -4.75085378e-01 -2.14873910e-01 6.48514986e-01
1.03947473e+00 -1.35164768e-01 -1.12057447e+00 -4.17693913e-01
-4.46856529e-01 9.58170772e-01 5.48971146e-02 5.25150836e-01
7.59016335e-01 -1.04941165e+00 -5.39618909e-01 -2.55151242e-01
2.81538099e-01 2.60832489e-01 4.01867330e-01 6.41268551e-01
-8.21544051e-01 4.96009767e-01 -4.56111312e-01 -2.45234296e-01
-1.17168176e+00 5.50147295e-01 3.15815151e-01 -4.95724887e-01
-3.45758885e-01 7.67242193e-01 -6.21066809e-01 -9.33387756e-01
2.81795412e-01 -9.13991392e-01 -3.80624503e-01 2.84308463e-01
6.16789758e-01 6.86385751e-01 1.75066203e-01 2.49360129e-01
-2.68531412e-01 7.39686072e-01 -2.02466547e-03 -3.41971703e-02
9.98460114e-01 -1.50339022e-01 -4.38958079e-01 3.83019060e-01
6.07252657e-01 2.80105341e-02 -4.49006855e-01 -4.59218360e-02
2.98307151e-01 -3.23641926e-01 -3.95156801e-01 -1.43958783e+00
-5.85170984e-01 8.05571198e-01 -2.44019821e-01 3.70811462e-01
1.24656200e+00 -7.04607889e-02 1.04552364e+00 9.32709455e-01
4.07624722e-01 -1.32815123e+00 2.83545047e-01 8.97847712e-01
1.11465657e+00 -1.18590510e+00 -2.58623362e-02 -7.40579665e-01
-3.27989578e-01 1.39249671e+00 1.45649731e+00 6.36022925e-01
1.19499937e-01 2.64343262e-01 2.89225280e-01 -5.61029434e-01
-1.02404761e+00 -2.39658684e-01 2.12757483e-01 3.51080984e-01
3.41752291e-01 -2.20978156e-01 -5.44619083e-01 8.68521214e-01
-9.90265250e-01 1.08907454e-01 7.85832942e-01 1.37233233e+00
-3.71761531e-01 -9.08468664e-01 -5.23334384e-01 1.89680234e-01
8.47965330e-02 -1.01441734e-01 -9.12071288e-01 1.16083729e+00
8.28802362e-02 1.48362970e+00 -1.45786390e-01 -2.41922244e-01
8.10942411e-01 2.50066429e-01 6.37138128e-01 -2.55671769e-01
-8.77053797e-01 -1.01815033e+00 2.46224314e-01 -5.76995134e-01
-3.91654313e-01 -9.64179114e-02 -1.91212595e+00 -7.08709002e-01
-4.69552055e-02 1.97356746e-01 4.09280449e-01 1.13113451e+00
6.21730387e-02 9.75006580e-01 -1.31912023e-01 4.76456493e-01
-5.53810060e-01 -6.24778092e-01 9.61848944e-02 3.20876956e-01
-1.45326614e-01 -3.75051856e-01 -2.39189506e-01 -3.49531472e-01] | [10.939240455627441, 7.917413711547852] |
ecf8d616-59b3-4b35-9c72-08bf8a83cbfe | multilingual-lexicalized-constituency-parsing | null | null | https://aclanthology.org/E17-2053 | https://aclanthology.org/E17-2053.pdf | Multilingual Lexicalized Constituency Parsing with Word-Level Auxiliary Tasks | We introduce a constituency parser based on a bi-LSTM encoder adapted from recent work (Cross and Huang, 2016b; Kiperwasser and Goldberg, 2016), which can incorporate a lower level character biLSTM (Ballesteros et al., 2015; Plank et al., 2016). We model two important interfaces of constituency parsing with auxiliary tasks supervised at the word level: (i) part-of-speech (POS) and morphological tagging, (ii) functional label prediction. On the SPMRL dataset, our parser obtains above state-of-the-art results on constituency parsing without requiring either predicted POS or morphological tags, and outputs labelled dependency trees. | ["Beno{\\^\\i}t Crabb{\\'e}", 'Maximin Coavoux'] | 2017-04-01 | null | null | null | eacl-2017-4 | ['morphological-tagging'] | ['natural-language-processing'] | [ 3.48433167e-01 7.33820856e-01 -3.89906794e-01 -5.11481941e-01
-1.13497174e+00 -9.56666768e-01 3.24415624e-01 5.37239671e-01
-5.31776249e-01 8.24620724e-01 5.63497007e-01 -9.23539042e-01
6.30794585e-01 -7.92564631e-01 -8.77114534e-01 -2.45162904e-01
-2.29899898e-01 1.00131541e-01 3.92593473e-01 -2.35760674e-01
-8.66031274e-02 3.60859782e-01 -9.99097347e-01 7.56852031e-01
4.44655389e-01 8.72893035e-01 4.03114378e-01 1.08046365e+00
-2.51996607e-01 1.27761531e+00 -3.25971484e-01 -7.80236542e-01
-1.39878824e-01 -8.98559345e-04 -1.04015553e+00 -5.31406045e-01
4.25183088e-01 -5.49789332e-02 -1.41760275e-01 9.08478200e-01
1.11422025e-01 -3.09645921e-01 1.54518768e-01 -4.28073406e-01
-8.43234599e-01 1.28639436e+00 -1.97235987e-01 3.86680543e-01
1.15754113e-01 4.83098673e-03 1.61997604e+00 -7.98263431e-01
6.64915264e-01 1.20377946e+00 8.73180211e-01 6.79864883e-01
-1.36802292e+00 -5.29622972e-01 5.21700025e-01 -5.32534756e-02
-5.36001623e-01 -3.54748666e-01 5.98423719e-01 -2.05327287e-01
1.62964213e+00 -1.67673483e-01 3.84620905e-01 1.31050432e+00
6.24141753e-01 8.66398156e-01 1.17140043e+00 -7.34440684e-01
3.97572517e-02 -5.14049232e-01 6.58621073e-01 8.12818408e-01
2.58318726e-02 1.39246866e-01 -3.82548362e-01 -6.93345368e-02
6.12381697e-01 -5.21710813e-01 4.19448525e-01 1.65196374e-01
-9.51280594e-01 1.02376342e+00 2.59467334e-01 6.69632792e-01
-3.93138200e-01 2.80480891e-01 7.43522525e-01 1.23435892e-01
8.23401690e-01 2.45874807e-01 -1.20871830e+00 -2.82163471e-01
-5.82164049e-01 -1.91650197e-01 9.09980237e-01 1.04268503e+00
5.65825105e-01 -1.16743125e-01 -3.25366288e-01 7.63540089e-01
3.40216070e-01 6.39519244e-02 3.98812264e-01 -8.47854376e-01
7.82340229e-01 1.62596092e-01 -3.00976604e-01 -1.11642249e-01
-6.13874316e-01 -1.55967638e-01 -3.82263571e-01 -1.42208710e-01
3.88288528e-01 -5.14158964e-01 -1.03788960e+00 2.31554008e+00
3.91418971e-02 2.13168740e-01 3.45689625e-01 1.93688139e-01
1.01731074e+00 8.12644720e-01 9.94093359e-01 -3.07047933e-01
1.60567784e+00 -1.22598064e+00 -7.12893963e-01 -6.64752126e-01
8.72725904e-01 -7.03112721e-01 6.52542949e-01 2.31839523e-01
-1.48026633e+00 -6.98636293e-01 -9.06135798e-01 -4.24243152e-01
-3.08687747e-01 2.58489400e-01 9.70495522e-01 6.49997890e-01
-1.38868523e+00 7.52166390e-01 -1.23016441e+00 -1.69239148e-01
1.69327989e-01 2.99758494e-01 -6.25659883e-01 1.43579811e-01
-1.36397755e+00 1.16856658e+00 6.46072090e-01 7.02147856e-02
-4.37445462e-01 -4.50201124e-01 -1.25222731e+00 3.02326512e-02
1.57195167e-03 -3.09773266e-01 1.50706720e+00 -9.07018721e-01
-1.77910364e+00 1.16702735e+00 -5.64362146e-02 -5.56451499e-01
-3.06305885e-01 -3.02654952e-01 -4.42764699e-01 -4.43038829e-02
1.87863901e-01 8.33014131e-01 1.59981385e-01 -9.15827870e-01
-8.82372618e-01 -2.41676867e-01 1.39579073e-01 -3.09197363e-02
6.21085637e-04 6.90901339e-01 4.13039662e-02 -6.26303792e-01
-1.92703232e-02 -7.49844968e-01 -2.29331538e-01 -5.83611071e-01
-3.65725815e-01 -6.50239170e-01 2.37431809e-01 -1.31563282e+00
1.12541294e+00 -1.92667377e+00 -8.59975740e-02 -4.96042401e-01
-4.75087427e-02 4.91929442e-01 -5.74897885e-01 5.34123600e-01
-3.71698946e-01 3.28083992e-01 -4.56954092e-01 -7.93415904e-01
-1.08061861e-02 4.03321534e-01 -2.41550446e-01 1.25046447e-01
6.94851398e-01 1.14897943e+00 -8.89993191e-01 -4.88526046e-01
-3.55188269e-04 3.92355651e-01 -3.75775844e-01 1.77247271e-01
-3.34925085e-01 3.90758663e-01 -7.69071728e-02 7.23966599e-01
5.40474832e-01 1.36239395e-01 6.77029252e-01 -3.24760675e-02
-6.09075725e-01 1.48426366e+00 -6.12780929e-01 1.72938859e+00
-9.70538080e-01 4.45272595e-01 3.70589495e-01 -9.69202578e-01
7.61472702e-01 6.76045477e-01 8.19637440e-03 -3.80270779e-01
2.46663019e-01 3.09472948e-01 3.41633826e-01 -4.25491855e-02
4.63947445e-01 -2.51856089e-01 -6.09462976e-01 1.58552438e-01
6.64200664e-01 2.02970609e-01 4.65021491e-01 -7.73061290e-02
1.52474117e+00 6.11403465e-01 6.27269447e-01 -3.28834116e-01
3.38534087e-01 -1.89203173e-02 8.64968240e-01 4.93177980e-01
-3.70419353e-01 2.56325096e-01 6.09555662e-01 -4.45615232e-01
-8.90148997e-01 -1.35615087e+00 -3.96704674e-01 1.76488554e+00
-6.04701281e-01 -2.83194482e-01 -8.86589587e-01 -9.21227276e-01
-3.90720487e-01 7.74322271e-01 -7.95904875e-01 1.04671337e-01
-1.36585438e+00 -6.21211827e-01 8.41289043e-01 9.90401566e-01
4.84264642e-02 -1.89909899e+00 -5.83469987e-01 8.93611729e-01
-3.03005707e-02 -1.41942286e+00 -1.36582971e-01 1.15603042e+00
-1.01279068e+00 -7.69080579e-01 -1.59453988e-01 -1.42293954e+00
4.95873466e-02 -4.94313091e-01 1.52555132e+00 -8.01131949e-02
1.33370310e-01 -2.70061910e-01 -5.88214815e-01 -2.82185972e-01
-9.12263870e-01 4.20381159e-01 -2.36646891e-01 -5.82002044e-01
3.05535287e-01 -6.56800508e-01 -8.89290124e-02 -4.19641644e-01
-5.30639410e-01 1.26015320e-01 1.06948316e+00 8.16541255e-01
8.02150905e-01 -7.36111701e-01 5.50102711e-01 -1.04995859e+00
1.93279326e-01 -3.63935024e-01 -4.72239763e-01 2.29508087e-01
-8.80146995e-02 5.55832088e-02 9.76945043e-01 -1.90198973e-01
-1.11961389e+00 1.71107575e-01 -8.74200881e-01 4.99275208e-01
-5.34863412e-01 2.98141897e-01 -3.83708358e-01 3.48185420e-01
6.23691194e-02 2.24369857e-03 -9.61983263e-01 -7.40278482e-01
6.97642148e-01 4.89525706e-01 6.20774448e-01 -7.49230206e-01
3.37685734e-01 -3.21859829e-02 -1.48955937e-02 -3.49014014e-01
-1.40442824e+00 -1.07709505e-01 -1.17350566e+00 3.91495079e-01
1.42184162e+00 -1.04139280e+00 -4.63195115e-01 2.70849735e-01
-1.69057918e+00 -7.54374862e-01 -2.38672122e-01 2.67642945e-01
-5.47836006e-01 2.72950709e-01 -1.61891198e+00 -5.77749312e-01
-5.64252853e-01 -7.45656967e-01 1.01220381e+00 1.41712781e-02
-8.76938701e-02 -1.07205379e+00 4.01802272e-01 -3.92145589e-02
3.26176756e-03 2.90419579e-01 1.22562027e+00 -7.53915846e-01
-1.05710454e-01 1.31348863e-01 -2.39877775e-01 5.12794495e-01
-9.46416520e-03 -1.48107320e-01 -9.87901926e-01 -4.71685305e-02
-1.19427651e-01 -5.46280563e-01 1.14470899e+00 4.81830329e-01
7.21647263e-01 -2.71786273e-01 -1.19124450e-01 5.39458990e-01
1.12231481e+00 1.68193519e-01 4.22482759e-01 3.39241177e-01
6.21190131e-01 6.79200411e-01 5.82536697e-01 -1.64250165e-01
7.26978362e-01 2.35887766e-01 3.29966903e-01 -3.90234329e-02
-2.62679309e-01 -5.07362127e-01 6.54715419e-01 1.11817336e+00
-5.99947385e-02 -3.46490622e-01 -7.88203418e-01 6.88696802e-01
-1.83817565e+00 -6.18124723e-01 -3.99155676e-01 1.50377941e+00
1.20092082e+00 4.80820239e-01 -1.34751529e-01 -1.06053196e-01
9.27685678e-01 3.47840697e-01 -3.65013331e-02 -1.19662166e+00
-1.67926624e-01 1.06671369e+00 7.19234824e-01 5.40912032e-01
-1.51439071e+00 1.78166556e+00 5.68260479e+00 4.99304593e-01
-8.62580836e-01 5.22177756e-01 8.08011532e-01 4.89751101e-01
-1.34956360e-01 3.46707135e-01 -1.12916481e+00 3.44743550e-01
1.51179016e+00 6.79202199e-01 5.40639497e-02 8.60551298e-01
-4.96887594e-01 2.59479582e-02 -1.09536529e+00 1.85237706e-01
-2.11155072e-01 -1.13732207e+00 -5.45794785e-01 -1.67083666e-01
4.43020791e-01 6.30852699e-01 -3.53295565e-01 6.81877017e-01
1.03358865e+00 -7.69539714e-01 1.00550473e+00 -1.04275746e-02
7.17406631e-01 -5.65241218e-01 6.92041397e-01 4.09429908e-01
-1.46072900e+00 7.64966458e-02 -4.28562999e-01 -3.07601720e-01
6.67808294e-01 5.08506060e-01 -2.00907841e-01 2.27035686e-01
6.73300385e-01 4.78912562e-01 -3.73677045e-01 9.05119106e-02
-1.20037544e+00 1.20095956e+00 -2.63054162e-01 -1.80138778e-02
4.19796646e-01 1.80560108e-02 1.71343639e-01 1.74800062e+00
-5.07111363e-02 1.79972306e-01 2.53373355e-01 6.28643751e-01
-2.43708327e-01 2.00708508e-01 -3.40951495e-02 -1.73322991e-01
5.86977839e-01 1.48948729e+00 -9.72645164e-01 -4.00367141e-01
-6.31583214e-01 8.19315195e-01 1.00771189e+00 -1.32416606e-01
-5.01608431e-01 -1.38579026e-01 4.87925142e-01 -4.32059854e-01
8.67123544e-01 -3.53210330e-01 -4.96264726e-01 -9.22137082e-01
2.27971375e-02 -4.91679937e-01 5.04090130e-01 -1.76249623e-01
-1.36251152e+00 8.48823011e-01 -2.89788693e-01 -5.30833542e-01
-1.90929979e-01 -1.04467452e+00 -9.35306370e-01 9.97299433e-01
-1.74260962e+00 -1.80585408e+00 6.54445291e-01 -5.02735078e-02
4.98004049e-01 1.42050952e-01 1.25452018e+00 9.54352096e-02
-6.19168460e-01 4.22317117e-01 -3.41293693e-01 5.99854052e-01
3.58183742e-01 -1.55707324e+00 1.53367984e+00 1.06312752e+00
2.85278201e-01 4.51694340e-01 2.04209700e-01 -6.77684605e-01
-9.60000038e-01 -1.25561011e+00 1.75716305e+00 -6.68566227e-01
9.89231706e-01 -9.88367558e-01 -7.26776659e-01 1.27591217e+00
4.28626269e-01 2.43604258e-01 7.55718410e-01 3.40178072e-01
-4.07418966e-01 4.78869319e-01 -1.06148279e+00 2.49640584e-01
1.22450125e+00 -4.48587060e-01 -8.20610285e-01 1.60262153e-01
1.28138542e+00 -3.75628561e-01 -9.93118048e-01 5.03687441e-01
3.96113753e-01 -1.01990533e+00 6.14813268e-01 -7.42873073e-01
7.04370737e-01 3.56504619e-01 -3.38033825e-01 -1.13511872e+00
-7.07945347e-01 -2.59799004e-01 -1.54251039e-01 1.61448371e+00
8.45077217e-01 -3.64814341e-01 4.26441699e-01 2.00531676e-01
-8.96702707e-01 -8.21343541e-01 -1.11274743e+00 -6.38255894e-01
7.37931132e-01 -5.99749446e-01 2.96395808e-01 6.42030001e-01
1.45985857e-01 6.85374916e-01 -3.83135676e-02 1.96589828e-01
2.71289378e-01 -7.42696598e-02 -3.83102670e-02 -1.23505700e+00
-6.52975261e-01 -3.39341164e-01 9.34247579e-03 -8.77306819e-01
8.57196331e-01 -1.01921570e+00 4.66618061e-01 -1.87438703e+00
-5.57301976e-02 -3.92912358e-01 -5.56429148e-01 9.85969961e-01
-4.16473240e-01 3.61269563e-01 3.09665680e-01 -2.90671468e-01
-6.21691108e-01 1.63912967e-01 5.85711300e-01 2.56054431e-01
3.64571437e-02 -2.19433039e-01 -6.33442104e-01 9.25593197e-01
8.87169719e-01 -6.98955894e-01 5.24273932e-01 -5.79289079e-01
1.17727198e-01 3.43099684e-01 -1.29228666e-01 -6.35710835e-01
-2.15095714e-01 9.88353789e-02 2.32044488e-01 -5.18202126e-01
2.82655567e-01 -2.00225189e-01 -8.30640256e-01 7.70676315e-01
-3.36103231e-01 5.08610487e-01 4.89476800e-01 1.12137474e-01
-1.52814403e-01 -4.85135585e-01 7.51076639e-01 -3.23032707e-01
-6.52463734e-01 1.49056450e-01 -6.03232145e-01 2.36613363e-01
5.50079107e-01 2.24513784e-01 -4.95899469e-01 1.37227312e-01
-8.07704508e-01 -8.50512907e-02 1.68412924e-02 2.80316889e-01
-4.88259606e-02 -1.09258699e+00 -7.24835813e-01 -6.47341534e-02
-1.27120376e-01 -5.78569360e-02 -5.17988019e-02 4.51943457e-01
-3.40089977e-01 5.65693378e-01 -1.17371932e-01 -2.33729798e-02
-9.28329766e-01 5.49232125e-01 -2.02706560e-01 -1.03449154e+00
-4.34900045e-01 1.16692662e+00 2.12333336e-01 -7.76274085e-01
-1.92548528e-01 -8.21790040e-01 -2.67631739e-01 -1.45673584e-02
3.25578660e-01 -7.34356046e-02 2.30649814e-01 -7.37457812e-01
-6.23021126e-01 2.95270503e-01 -1.11815885e-01 -7.68429190e-02
1.61373103e+00 1.02552809e-01 -2.84193397e-01 4.88152951e-01
1.14352393e+00 1.14216976e-01 -1.34315836e+00 -1.76090032e-01
5.68533421e-01 5.12871623e-01 4.30874750e-02 -1.04658103e+00
-5.80550253e-01 1.23417234e+00 1.48923904e-01 3.08469385e-01
8.75305831e-01 4.29487199e-01 1.39027548e+00 1.72813386e-01
3.24547023e-01 -1.15258527e+00 -3.95543903e-01 1.25247061e+00
4.33925033e-01 -1.14639473e+00 -5.88612080e-01 -5.85474849e-01
-4.47179407e-01 1.16220200e+00 5.62193811e-01 -3.49857509e-01
6.32569194e-01 7.83907175e-01 2.27419868e-01 2.42490977e-01
-1.03324676e+00 -5.17754674e-01 7.39437938e-02 4.71383691e-01
1.04777503e+00 3.66718769e-01 -7.73390353e-01 1.13502347e+00
-2.77088970e-01 -5.94683349e-01 3.07236046e-01 9.06716585e-01
-5.13208807e-01 -1.69819248e+00 2.99525429e-02 2.40920812e-01
-1.08499086e+00 -8.90242398e-01 -3.26774180e-01 6.39464676e-01
3.66167486e-01 1.01083875e+00 3.32378238e-01 -1.05047174e-01
3.41088831e-01 3.34908336e-01 6.73204958e-01 -1.40378940e+00
-1.04425991e+00 8.94710198e-02 6.33028865e-01 -3.47389907e-01
-3.77458990e-01 -8.01619112e-01 -1.48149467e+00 3.22757721e-01
-2.13971823e-01 8.09005648e-02 7.12226868e-01 1.21007097e+00
-3.42041999e-03 7.16900826e-01 4.58946824e-01 -8.88766110e-01
-2.62529463e-01 -1.51848435e+00 -2.95340151e-01 4.11804505e-02
1.78877071e-01 -2.50304967e-01 -1.11771822e-01 1.14191905e-01] | [10.360333442687988, 9.671757698059082] |
f2a03ce8-cfce-43aa-a7ee-c72baacf4295 | collaborating-domain-shared-and-target | 2207.09767 | null | https://arxiv.org/abs/2207.09767v1 | https://arxiv.org/pdf/2207.09767v1.pdf | Collaborating Domain-shared and Target-specific Feature Clustering for Cross-domain 3D Action Recognition | In this work, we consider the problem of cross-domain 3D action recognition in the open-set setting, which has been rarely explored before. Specifically, there is a source domain and a target domain that contain the skeleton sequences with different styles and categories, and our purpose is to cluster the target data by utilizing the labeled source data and unlabeled target data. For such a challenging task, this paper presents a novel approach dubbed CoDT to collaboratively cluster the domain-shared features and target-specific features. CoDT consists of two parallel branches. One branch aims to learn domain-shared features with supervised learning in the source domain, while the other is to learn target-specific features using contrastive learning in the target domain. To cluster the features, we propose an online clustering algorithm that enables simultaneous promotion of robust pseudo label generation and feature clustering. Furthermore, to leverage the complementarity of domain-shared features and target-specific features, we propose a novel collaborative clustering strategy to enforce pair-wise relationship consistency between the two branches. We conduct extensive experiments on multiple cross-domain 3D action recognition datasets, and the results demonstrate the effectiveness of our method. | ['Zilei Wang', 'Qinying Liu'] | 2022-07-20 | null | null | null | null | ['online-clustering', 'video-domain-adapation', '3d-human-action-recognition'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 3.29504341e-01 -1.89884409e-01 -4.65231717e-01 -4.02227193e-01
-8.46492946e-01 -5.37269711e-01 3.60766917e-01 -4.19218391e-01
-6.00353256e-03 4.04787362e-01 2.98986018e-01 2.16126576e-01
-3.25857580e-01 -2.59917885e-01 -3.54706138e-01 -8.89907420e-01
-4.32361476e-03 5.40362656e-01 3.50954503e-01 1.20211542e-01
2.28997603e-01 2.35175282e-01 -1.56158125e+00 4.85294968e-01
8.56467605e-01 1.19353008e+00 4.30282243e-02 1.30027562e-01
-2.16842405e-02 8.83113265e-01 -4.58333015e-01 -2.01463141e-02
5.46882331e-01 -6.61716521e-01 -8.28310847e-01 9.09503579e-01
2.96028852e-01 5.93953347e-03 -2.62809426e-01 8.94768596e-01
3.74837905e-01 1.17224224e-01 8.76971364e-01 -1.67650843e+00
-3.43857050e-01 2.47931555e-01 -9.46616173e-01 -1.59355327e-01
3.25331837e-01 2.34132007e-01 9.20230567e-01 -6.77056074e-01
6.52008057e-01 1.18633699e+00 3.93820703e-01 7.20255673e-01
-1.22125840e+00 -8.39420915e-01 4.37282443e-01 3.04054499e-01
-1.42959893e+00 -3.27157557e-01 1.23237169e+00 -6.75161302e-01
3.22535694e-01 -1.09813184e-01 6.43909335e-01 1.19367552e+00
-3.00961345e-01 1.09742308e+00 1.30462885e+00 -4.27246481e-01
3.98075819e-01 -1.24088019e-01 8.81309211e-02 6.34593368e-01
-1.24780506e-01 5.65026999e-02 -7.11350203e-01 -2.89056242e-01
6.74392462e-01 1.88826859e-01 -2.22606272e-01 -1.01588595e+00
-1.32513070e+00 6.99272871e-01 6.40279949e-02 3.00705910e-01
-1.18076399e-01 -3.46145958e-01 4.51245070e-01 2.32166320e-01
4.48419034e-01 2.38210801e-02 -5.24295211e-01 -1.25223145e-01
-6.38902247e-01 1.04040369e-01 6.37733161e-01 1.21294296e+00
7.72977173e-01 -5.78234136e-01 -3.56702805e-02 1.08440685e+00
4.35520262e-01 3.87022674e-01 7.95518517e-01 -9.25687850e-01
7.04040408e-01 9.96218443e-01 -2.96552461e-02 -9.12297487e-01
-2.25676671e-01 -4.65873927e-02 -7.27290869e-01 2.37925231e-01
6.91247165e-01 9.67424130e-04 -8.06501329e-01 1.85028708e+00
7.73762822e-01 2.77717441e-01 1.92212328e-01 1.03088963e+00
4.49412107e-01 1.57038122e-01 -2.14910164e-01 -2.21770525e-01
1.00732470e+00 -1.08219457e+00 -3.84683847e-01 -4.03475761e-02
8.19128990e-01 -5.30645907e-01 8.62713933e-01 1.36808887e-01
-5.98738194e-01 -5.60146809e-01 -9.39467490e-01 1.92745402e-01
-5.74552314e-03 2.48993114e-01 3.73108357e-01 3.89396578e-01
-3.17498684e-01 4.40160960e-01 -7.47990131e-01 -3.83028090e-01
7.09813237e-01 2.29183719e-01 -5.66057503e-01 -2.62429774e-01
-8.09046268e-01 3.94436717e-01 5.75585783e-01 -3.15366268e-01
-8.24407995e-01 -3.91033590e-01 -7.33087122e-01 -4.20441717e-01
7.96881139e-01 -3.69127333e-01 1.01613104e+00 -1.21684432e+00
-1.55323076e+00 1.10669470e+00 -2.22137826e-03 1.11286722e-01
5.77790678e-01 2.83918679e-02 -4.10991222e-01 2.10094213e-01
4.19485331e-01 4.18756723e-01 1.03840828e+00 -1.32306123e+00
-9.79346395e-01 -8.05016398e-01 -1.20723531e-01 3.80842507e-01
-5.33575356e-01 -1.59307599e-01 -7.16765702e-01 -6.62749469e-01
4.33601499e-01 -1.06439018e+00 -1.22881390e-01 1.96789831e-01
-4.24139738e-01 -3.56020182e-01 8.81774068e-01 -1.99912533e-01
8.47602844e-01 -2.39329624e+00 3.97202015e-01 5.25610566e-01
2.87449628e-01 6.69582784e-02 -1.65566325e-01 2.10741356e-01
-2.14944556e-01 -3.31296593e-01 -4.64201123e-01 -1.47698998e-01
3.69621664e-02 3.91177714e-01 -1.39358014e-01 5.76116204e-01
2.25520805e-01 5.55969477e-01 -9.61193144e-01 -8.94816220e-01
1.04922056e-01 -7.87174031e-02 -4.41578686e-01 3.97720218e-01
-2.49594897e-01 7.82005072e-01 -9.21512187e-01 8.87097001e-01
6.06090188e-01 -4.45708841e-01 5.17880976e-01 -2.73312956e-01
1.17260911e-01 -1.10854581e-01 -1.52747798e+00 1.94486392e+00
3.56952958e-02 2.47270390e-02 -1.65110230e-01 -1.32633567e+00
1.01508951e+00 7.35764429e-02 9.65902269e-01 -4.87637579e-01
2.11013854e-01 2.24695385e-01 -2.01091304e-01 -5.62119305e-01
-1.63494870e-02 -1.49631858e-01 -2.66757220e-01 8.00247669e-01
4.86256704e-02 2.19424680e-01 -7.83338249e-02 1.14892507e-02
1.10628343e+00 5.20902753e-01 4.09374386e-01 -1.89369582e-02
6.37239099e-01 1.27701879e-01 1.03061354e+00 2.80094594e-01
-5.74322641e-01 6.67203844e-01 5.14381111e-01 -1.46052793e-01
-6.33803070e-01 -9.77045178e-01 1.79061100e-01 9.74630356e-01
4.84977275e-01 -3.95541579e-01 -6.09880388e-01 -1.40661991e+00
7.81960413e-02 -5.44829294e-02 -7.39954889e-01 -3.26445043e-01
-3.50337744e-01 -3.13658744e-01 3.71229559e-01 6.75420403e-01
6.17000461e-01 -6.76929593e-01 -6.07731938e-01 -8.87442306e-02
-3.23752820e-01 -1.04630589e+00 -8.09380651e-01 2.95873672e-01
-9.16472852e-01 -1.51147151e+00 -7.91232646e-01 -9.55997288e-01
7.90058911e-01 4.76157129e-01 8.19417357e-01 -1.02472581e-01
-1.48108438e-01 7.31272817e-01 -7.46069431e-01 -1.73773766e-02
-9.92287248e-02 -1.07784033e-01 1.42482623e-01 5.96642673e-01
5.35194695e-01 -6.92891002e-01 -3.82067859e-01 8.65542293e-01
-8.83888841e-01 -3.30658257e-02 6.47776008e-01 9.85175610e-01
8.42641234e-01 1.15898527e-01 3.47981572e-01 -6.93920195e-01
1.86859623e-01 -5.31425118e-01 -2.86639422e-01 3.17307562e-01
-3.25570583e-01 -5.71527658e-03 5.34916818e-01 -7.80916154e-01
-9.71401274e-01 6.36911392e-01 4.01017725e-01 -9.30592656e-01
-4.45366532e-01 4.07305956e-01 -9.09291446e-01 5.58516718e-02
5.18711865e-01 3.00642997e-01 2.94861108e-01 -5.69225967e-01
3.94047946e-01 7.98360765e-01 5.55621445e-01 -8.00708234e-01
8.77967954e-01 7.41018176e-01 -1.59267575e-01 -4.80729967e-01
-1.02446806e+00 -7.39745975e-01 -1.17060888e+00 -3.50316614e-01
9.02313471e-01 -9.24220085e-01 -2.98816264e-01 6.92526519e-01
-6.75948679e-01 -2.86303192e-01 -5.04982054e-01 5.75808406e-01
-9.85535443e-01 6.93490028e-01 2.72467434e-02 -5.31874537e-01
3.06902796e-01 -1.02561307e+00 1.10181451e+00 1.14334434e-01
-7.98960775e-02 -8.38658154e-01 2.83202201e-01 6.21872365e-01
-4.24798340e-01 3.97909194e-01 6.95234776e-01 -1.06101680e+00
-3.03494692e-01 9.68167931e-03 -1.55397549e-01 4.37460840e-01
5.94363928e-01 -3.60520780e-01 -6.96329474e-01 -1.64338529e-01
-5.60342595e-02 -7.96644270e-01 6.09349728e-01 -7.21270591e-03
1.19406140e+00 6.64184541e-02 -5.62037230e-01 4.32315469e-01
9.93089974e-01 2.38260791e-01 2.95358390e-01 2.12303385e-01
7.48427927e-01 8.69407952e-01 1.17399657e+00 8.09933543e-01
4.10469174e-01 8.60009730e-01 1.42578974e-01 7.28180334e-02
-2.07677465e-02 -3.57785463e-01 3.87118310e-01 7.30900407e-01
1.66842505e-01 2.49317482e-01 -7.82366514e-01 5.06177425e-01
-2.16371179e+00 -1.02649999e+00 1.77875787e-01 2.02105331e+00
1.01550353e+00 -8.36427733e-02 6.44122243e-01 2.67240077e-01
1.00138605e+00 1.33233339e-01 -9.54344988e-01 4.83576268e-01
4.21100892e-02 -5.03867008e-02 1.51798576e-01 -1.64254338e-01
-1.37012947e+00 6.76265657e-01 5.09988499e+00 1.02480400e+00
-8.27655315e-01 3.92935835e-02 2.37692371e-01 -8.17834437e-02
1.44774318e-01 7.64658004e-02 -6.43468380e-01 7.14816689e-01
3.01977277e-01 -8.71308669e-02 1.02503799e-01 9.71677303e-01
5.98112568e-02 4.51636016e-02 -1.34170902e+00 1.18520093e+00
2.95990646e-01 -9.50117707e-01 -1.23899378e-01 2.22404599e-01
7.21806467e-01 -4.38120782e-01 -8.59144777e-02 2.20990941e-01
4.57215101e-01 -4.49661851e-01 6.25858426e-01 4.23021168e-01
7.36589372e-01 -7.55980015e-01 1.27617002e-01 5.42231619e-01
-1.39491332e+00 -1.88188329e-01 -2.44842798e-01 1.35915920e-01
-7.16543570e-02 4.47011173e-01 -3.16489726e-01 7.69579709e-01
7.23699570e-01 1.42549717e+00 -3.21365029e-01 9.47002947e-01
-2.81696357e-02 3.44001055e-01 -2.04935875e-02 3.13219041e-01
1.09104499e-01 -4.43743289e-01 3.65921617e-01 6.68818414e-01
6.42072186e-02 2.65369177e-01 7.63028383e-01 5.59139907e-01
1.21802308e-01 1.27064243e-01 -5.67637801e-01 2.86095534e-02
4.19194341e-01 1.03033113e+00 -6.65064871e-01 -2.33840555e-01
-5.44174790e-01 1.20480299e+00 4.22728181e-01 8.75057932e-03
-9.92666602e-01 -1.89675197e-01 7.93160617e-01 -2.08001554e-01
5.48457146e-01 -1.58799529e-01 -1.54183164e-01 -1.36541271e+00
3.06567907e-01 -1.05801702e+00 8.29752862e-01 -3.69681984e-01
-1.80770874e+00 1.63298890e-01 -2.18116976e-02 -2.03409123e+00
1.19373715e-02 -4.72906142e-01 -3.78258288e-01 4.41652358e-01
-1.38970196e+00 -1.34661078e+00 -3.23479176e-01 1.34055972e+00
4.91708159e-01 -4.03790534e-01 5.01123786e-01 2.73997068e-01
-5.60842812e-01 6.01649940e-01 3.88035029e-01 3.08847606e-01
9.38697517e-01 -1.00036955e+00 -2.05491215e-01 4.90461022e-01
1.58255160e-01 2.83036590e-01 1.58411145e-01 -7.40009487e-01
-1.25749338e+00 -1.37441063e+00 4.31915373e-01 -3.15638572e-01
6.41042054e-01 -3.33990186e-01 -6.83065295e-01 6.56008065e-01
-3.51913452e-01 2.59979159e-01 1.18252289e+00 4.89294939e-02
-7.39079952e-01 -8.41806382e-02 -1.04415703e+00 3.83373618e-01
1.57104182e+00 -4.01571393e-01 -7.85869658e-01 5.83385050e-01
3.59518856e-01 -4.28030998e-01 -1.04132450e+00 3.33701104e-01
6.29025280e-01 -9.33128774e-01 7.06451654e-01 -8.65778744e-01
5.59962213e-01 -5.62708616e-01 -4.90410894e-01 -1.24478102e+00
-2.63805151e-01 -3.35309029e-01 -3.65510769e-03 1.50905621e+00
1.80625364e-01 -5.26557326e-01 9.78749216e-01 4.05373275e-01
-2.95142829e-01 -5.76236069e-01 -1.03995919e+00 -1.32792759e+00
8.67142826e-02 -4.56789672e-01 4.63146269e-01 1.06726241e+00
1.30961701e-01 2.24407271e-01 -5.32301545e-01 9.83658805e-02
6.97363675e-01 8.82863641e-01 1.10102820e+00 -1.33166730e+00
-4.21704352e-01 -2.06611961e-01 -5.83545625e-01 -1.37963450e+00
5.39700866e-01 -1.01922011e+00 1.89239204e-01 -1.07071614e+00
5.74352443e-01 -7.30733275e-01 -2.49242902e-01 7.30214298e-01
-5.83430529e-02 9.54589024e-02 1.89437076e-01 6.07697308e-01
-1.15320778e+00 8.53553712e-01 1.23854160e+00 -1.73800558e-01
-2.33696729e-01 2.54786387e-02 -6.98325455e-01 6.37729108e-01
4.89903063e-01 -5.54384530e-01 -6.17998123e-01 -1.18515179e-01
-5.78072846e-01 -9.35668796e-02 3.77831668e-01 -1.09117031e+00
1.59652084e-01 -4.81171131e-01 3.55160147e-01 -6.34203374e-01
2.73907691e-01 -1.15578175e+00 -1.34407103e-01 1.95579395e-01
-4.03675616e-01 -5.33502758e-01 -3.49816561e-01 9.74707425e-01
-2.93550342e-01 9.83538032e-02 9.29071784e-01 -3.21428962e-02
-9.78295922e-01 5.57477474e-01 -1.11676037e-01 3.95677358e-01
1.59995186e+00 -5.12354374e-01 -8.55208337e-02 -3.03795673e-02
-8.80237937e-01 4.93842751e-01 6.68508291e-01 6.75952256e-01
6.46216214e-01 -1.74923301e+00 -4.25555021e-01 3.11983198e-01
7.01042891e-01 -6.78058043e-02 2.45865360e-01 8.94590795e-01
3.24156553e-01 3.10146995e-02 -3.29472035e-01 -9.69817936e-01
-1.45548952e+00 7.00239241e-01 2.84120321e-01 -2.16493681e-01
-6.76460445e-01 6.50716782e-01 3.20656121e-01 -6.44710422e-01
3.02239925e-01 1.48010314e-01 -2.28296489e-01 1.55215949e-01
3.81373972e-01 3.66325289e-01 -3.16129863e-01 -7.67452896e-01
-5.08014619e-01 8.33356142e-01 -1.48399621e-01 9.19118226e-02
1.27269435e+00 -2.10964665e-01 6.87684938e-02 5.36155701e-01
1.25921726e+00 -3.10632437e-01 -1.55914700e+00 -6.21216476e-01
2.33556539e-01 -8.39488745e-01 -4.38035309e-01 -7.24234760e-01
-1.30733776e+00 5.35149634e-01 6.75448418e-01 -2.31929675e-01
1.36920524e+00 4.02567208e-01 6.94953620e-01 7.50142634e-02
5.21429181e-01 -1.45999360e+00 6.31711543e-01 4.55242068e-01
5.30622780e-01 -1.29885995e+00 -2.09740475e-01 -5.78068137e-01
-9.72713947e-01 9.33582842e-01 8.46394837e-01 1.80684458e-02
6.00355208e-01 -6.41340613e-02 8.92717093e-02 -3.35386336e-01
-5.89826941e-01 -5.21443248e-01 1.57508478e-01 9.38956261e-01
-3.04100718e-02 1.92463398e-02 -2.52324462e-01 8.60429704e-01
2.89058208e-01 1.12870455e-01 1.15262620e-01 1.35256374e+00
-2.64057785e-01 -1.44816232e+00 -4.85986471e-01 3.78175467e-01
4.46756855e-02 4.78895068e-01 -8.04568946e-01 7.16212451e-01
3.51993918e-01 1.11052668e+00 -1.07464030e-01 -8.44810188e-01
4.14060682e-01 1.14220917e-01 5.04001141e-01 -6.43941581e-01
-1.86072037e-01 1.93317026e-01 -1.05330557e-01 -7.57099807e-01
-8.58579218e-01 -1.09413588e+00 -1.26412833e+00 2.32053310e-01
-2.87462652e-01 4.51963171e-02 2.97556162e-01 1.13545871e+00
5.27642250e-01 1.33009493e-01 1.21600759e+00 -5.86438358e-01
-7.69158781e-01 -5.71977675e-01 -9.99227941e-01 7.53139019e-01
9.06909928e-02 -1.33140457e+00 -3.24729770e-01 3.58797640e-01] | [8.381820678710938, 0.9197940230369568] |
df235761-13cb-4528-b06a-04b33b397b5b | on-the-transferability-of-pre-trained-1 | 2204.09653 | null | https://arxiv.org/abs/2204.09653v1 | https://arxiv.org/pdf/2204.09653v1.pdf | On the Transferability of Pre-trained Language Models for Low-Resource Programming Languages | A recent study by Ahmed and Devanbu reported that using a corpus of code written in multilingual datasets to fine-tune multilingual Pre-trained Language Models (PLMs) achieves higher performance as opposed to using a corpus of code written in just one programming language. However, no analysis was made with respect to fine-tuning monolingual PLMs. Furthermore, some programming languages are inherently different and code written in one language usually cannot be interchanged with the others, i.e., Ruby and Java code possess very different structure. To better understand how monolingual and multilingual PLMs affect different programming languages, we investigate 1) the performance of PLMs on Ruby for two popular Software Engineering tasks: Code Summarization and Code Search, 2) the strategy (to select programming languages) that works well on fine-tuning multilingual PLMs for Ruby, and 3) the performance of the fine-tuned PLMs on Ruby given different code lengths. In this work, we analyze over a hundred of pre-trained and fine-tuned models. Our results show that 1) multilingual PLMs have a lower Performance-to-Time Ratio (the BLEU, METEOR, or MRR scores over the fine-tuning duration) as compared to monolingual PLMs, 2) our proposed strategy to select target programming languages to fine-tune multilingual PLMs is effective: it reduces the time to fine-tune yet achieves higher performance in Code Summarization and Code Search tasks, and 3) our proposed strategy consistently shows good performance on different code lengths. | ['Timofey Bryksin', 'David Lo', 'Fatemeh Fard', 'Fuxiang Chen'] | 2022-04-05 | null | null | null | null | ['code-search', 'code-search'] | ['computer-code', 'computer-vision'] | [-2.76255250e-01 -2.95170367e-01 -3.40303510e-01 -2.22924143e-01
-1.12717354e+00 -9.56961036e-01 4.81639028e-01 3.12850147e-01
-2.77422667e-01 4.29865718e-01 2.24069670e-01 -6.01280212e-01
-7.27567682e-03 -5.13445020e-01 -6.57444656e-01 -1.93234503e-01
3.29795405e-02 2.78655678e-01 1.64124250e-01 -3.35102856e-01
7.11621642e-01 1.77175298e-01 -1.28089130e+00 3.43917727e-01
1.58849549e+00 -2.16085523e-01 8.68767023e-01 9.29913461e-01
-5.65555871e-01 9.96227324e-01 -8.28163445e-01 -5.44631362e-01
-6.09072149e-02 -3.54290098e-01 -1.04138720e+00 -3.38042885e-01
6.14766359e-01 2.76677281e-01 1.13492265e-01 1.29503834e+00
3.64130110e-01 -3.70602190e-01 4.34201986e-01 -9.01430309e-01
-7.68159211e-01 1.18862605e+00 -7.95055807e-01 1.54072851e-01
4.21303093e-01 2.23039072e-02 9.80124176e-01 -7.61814654e-01
5.93912184e-01 1.19037271e+00 7.13994682e-01 4.16839898e-01
-1.28301656e+00 -6.93503559e-01 -7.43243173e-02 -3.14784378e-01
-1.65122867e+00 -2.98789948e-01 4.04873163e-01 -1.06313992e+00
1.54022241e+00 2.38194168e-01 6.15548752e-02 5.09878337e-01
7.67265797e-01 4.09299850e-01 8.32337379e-01 -7.66009092e-01
-3.85067433e-01 4.52127308e-01 3.27407688e-01 9.35085177e-01
2.41388276e-01 -3.93631846e-01 -1.76804978e-02 -5.42954504e-01
2.10616305e-01 -1.99912205e-01 -7.77573362e-02 -1.38802364e-01
-1.39723086e+00 1.03556800e+00 -6.43250495e-02 8.07962954e-01
-3.12683396e-02 1.69854000e-01 9.30922389e-01 5.21446466e-01
1.22867584e-01 9.08816338e-01 -7.42559552e-01 -3.20233732e-01
-1.17293727e+00 1.19097650e-01 8.55982184e-01 1.24623537e+00
1.15354514e+00 1.91286206e-01 4.10362072e-02 1.37145531e+00
1.93210214e-01 8.06594491e-01 8.54923844e-01 -8.15961599e-01
9.24808145e-01 8.23548138e-01 -3.72288674e-01 -7.28865266e-01
-3.01799804e-01 -1.58356577e-01 -3.64075482e-01 2.66377348e-02
1.99393898e-01 -6.40173405e-02 -3.30448747e-01 1.59085596e+00
-4.49504495e-01 -7.82361269e-01 2.37185150e-01 1.81427166e-01
8.98295939e-01 7.26690829e-01 2.32201125e-02 -1.22765571e-01
1.23235333e+00 -1.31954157e+00 -2.91296393e-01 -4.24073517e-01
9.90984857e-01 -1.41788185e+00 1.46982110e+00 2.41258502e-01
-1.04400110e+00 -6.93051159e-01 -9.61299002e-01 -5.90029871e-03
-1.99604735e-01 6.13836467e-01 5.22929490e-01 7.44220316e-01
-1.22036862e+00 1.29492715e-01 -6.89868808e-01 -6.96351051e-01
-3.91732246e-01 1.33212000e-01 -2.14333609e-01 -5.00004627e-02
-8.04094374e-01 9.61993575e-01 5.96185565e-01 -7.34516144e-01
-1.04173779e+00 -6.00960016e-01 -8.63172293e-01 4.68102619e-02
2.36324832e-01 -4.51486915e-01 1.25065684e+00 -9.35387611e-01
-1.21101010e+00 1.00299585e+00 -1.76952809e-01 -4.39977683e-02
8.77981782e-02 -1.21953890e-01 -3.41106057e-01 -3.76230061e-01
3.96627635e-01 3.10596436e-01 4.56524789e-01 -1.04804313e+00
-7.23066926e-01 8.51791278e-02 3.50968182e-01 -1.07926361e-01
-3.07541877e-01 5.51523149e-01 -6.51785374e-01 -7.25645840e-01
-3.81200105e-01 -1.10167921e+00 -8.53824392e-02 -1.06636226e+00
-2.60221392e-01 -2.45959148e-01 1.96977645e-01 -7.73594022e-01
1.91030622e+00 -2.01464105e+00 4.02872890e-01 -9.50013697e-02
9.20434296e-02 2.08878040e-01 -5.57579041e-01 7.30292857e-01
-1.56169444e-01 4.52892631e-01 -3.17900181e-01 2.65454233e-01
-1.61618609e-02 8.82792845e-02 -1.64933383e-01 8.56603161e-02
-2.37559587e-01 6.79018617e-01 -8.90438378e-01 -5.68082809e-01
-4.43227403e-03 9.39753801e-02 -9.71985221e-01 4.41481248e-02
-7.78165460e-02 1.25438318e-01 -4.09176290e-01 5.90497315e-01
4.23196584e-01 -1.22177802e-01 3.12512785e-01 2.05789834e-01
-6.02236271e-01 3.40891480e-01 -7.69888580e-01 1.77844524e+00
-1.33685327e+00 6.96297467e-01 -2.33634979e-01 -3.65646303e-01
1.02335107e+00 1.26619026e-01 1.38683021e-01 -5.69455683e-01
-3.69536310e-01 5.97661316e-01 -1.82523392e-02 -8.60735357e-01
7.33243465e-01 3.29893768e-01 -6.96693003e-01 7.43903160e-01
1.07791781e-01 -3.97659063e-01 7.20950603e-01 2.56838650e-01
1.20952666e+00 -5.01914769e-02 6.84947610e-01 -7.97343254e-01
1.06065476e+00 1.18537590e-01 3.34453195e-01 8.60004485e-01
2.40928337e-01 2.75722235e-01 4.74777281e-01 -2.99125463e-01
-1.00539613e+00 -9.06578302e-01 -2.57496275e-02 1.66930807e+00
-1.94278494e-01 -9.38735366e-01 -8.46994817e-01 -6.77685976e-01
-1.85957819e-01 8.68996918e-01 -1.85170427e-01 -9.41850536e-04
-9.18087780e-01 -8.42232704e-01 1.02676654e+00 4.12084311e-02
2.94641703e-01 -1.07893908e+00 -4.78979498e-01 2.01702103e-01
-4.91044432e-01 -7.94034719e-01 -1.05432057e+00 1.58262730e-01
-7.64242351e-01 -1.18280685e+00 -4.08932000e-01 -1.13241363e+00
6.05036318e-01 1.80165231e-01 1.72187102e+00 3.22993606e-01
-9.59923398e-03 3.53675187e-01 -4.52493697e-01 -1.53177127e-01
-1.33851779e+00 6.47870719e-01 -2.50319123e-01 -8.05901408e-01
4.22869295e-01 -4.30787891e-01 8.87174085e-02 1.59391522e-01
-9.37868118e-01 -1.38353705e-01 8.24293494e-01 5.87109745e-01
5.48116378e-02 -1.01824097e-01 4.84846920e-01 -1.12580228e+00
1.06161666e+00 -4.91502434e-01 -8.14850509e-01 8.53193641e-01
-7.94892848e-01 4.06740576e-01 6.97642744e-01 -4.58019495e-01
-1.01581633e+00 -7.37168267e-02 -2.53101945e-01 2.44173899e-01
2.71839499e-01 9.17469382e-01 2.51919627e-01 -2.90362298e-01
1.00035632e+00 4.23562258e-01 -5.54947019e-01 -5.82662940e-01
3.24491024e-01 8.60583961e-01 3.75077695e-01 -1.13997769e+00
7.11123407e-01 -1.78227350e-01 -8.09026539e-01 -6.85051203e-01
-2.54711330e-01 -4.79710788e-01 -6.65352702e-01 8.01713914e-02
7.63508558e-01 -9.52926040e-01 -1.48830190e-01 2.40767717e-01
-1.48726141e+00 -2.04658628e-01 2.55121678e-01 4.16260868e-01
-4.87203956e-01 6.17884338e-01 -7.38811910e-01 -3.74596387e-01
-5.84106922e-01 -1.80859756e+00 9.47462142e-01 5.12650944e-02
-5.44810951e-01 -1.10410917e+00 6.16734922e-01 3.55445087e-01
5.16751945e-01 -1.07490130e-01 1.62430513e+00 -4.00278032e-01
-3.43669295e-01 1.51676401e-01 -1.49241701e-01 3.48412871e-01
2.39964649e-01 4.98429626e-01 -5.04706144e-01 -5.93540967e-01
-1.35663494e-01 -2.65170902e-01 5.01388609e-01 1.95296183e-01
6.74132645e-01 -3.90114546e-01 1.69838301e-03 4.24573123e-01
1.66544318e+00 2.66882718e-01 3.68914217e-01 5.73778868e-01
8.47623348e-01 5.19584000e-01 4.53347385e-01 3.77851576e-01
6.44696534e-01 6.70030653e-01 6.43004617e-03 3.43621075e-01
-7.18336776e-02 -5.16683422e-02 1.02396762e+00 1.93041968e+00
1.26677170e-01 2.77075648e-01 -1.40451217e+00 8.12697828e-01
-1.64034760e+00 -4.99415100e-01 -2.97854900e-01 2.24390268e+00
1.29561865e+00 -1.22541137e-01 2.13953257e-02 -6.37985229e-01
6.56738698e-01 4.80101295e-02 -1.52890056e-01 -9.39974904e-01
-4.20334153e-02 1.32379234e-01 4.61622864e-01 7.22344398e-01
-7.04553545e-01 1.06409347e+00 6.24893618e+00 1.02570879e+00
-1.26448858e+00 3.95240128e-01 -7.63816088e-02 1.97522089e-01
-4.90226477e-01 3.22744340e-01 -7.97102392e-01 4.10321265e-01
1.13490784e+00 -7.46679008e-01 8.37557435e-01 1.16673172e+00
-6.05757050e-02 -7.28203207e-02 -1.22995138e+00 8.89395893e-01
1.69057682e-01 -1.23921430e+00 2.43752047e-01 -2.79464960e-01
1.19633973e+00 6.94115281e-01 -4.43923295e-01 8.95028710e-01
6.23589039e-01 -8.32181811e-01 9.16629970e-01 2.15584993e-01
7.57758439e-01 -7.38532364e-01 7.16151893e-01 5.61442196e-01
-1.42766201e+00 6.89661456e-03 -5.32422841e-01 3.00060779e-01
-3.64442319e-01 3.87091279e-01 -6.83329284e-01 6.46890759e-01
6.89696789e-01 4.26673830e-01 -1.10549629e+00 8.23069215e-01
-1.00186877e-01 4.50788528e-01 2.73523688e-01 7.16557875e-02
1.91707790e-01 -1.07060507e-01 4.99911308e-01 1.93149447e+00
5.16300142e-01 -7.52108991e-01 3.46350461e-01 1.05781019e+00
-4.93737729e-03 5.96391797e-01 -6.25911534e-01 -2.99360663e-01
4.49458599e-01 1.08756566e+00 -4.39996213e-01 -3.69997591e-01
-8.69167626e-01 5.86193621e-01 2.83941180e-01 3.35666239e-01
-7.51118124e-01 -7.37111747e-01 4.22112912e-01 -1.60399333e-01
-3.03390641e-02 -2.65524834e-01 -9.96839553e-02 -1.31665063e+00
3.05155758e-02 -1.50767589e+00 4.60867226e-01 -6.88924432e-01
-1.03319204e+00 1.06297207e+00 7.48550100e-03 -1.07660007e+00
-3.95137429e-01 -7.35467151e-02 -3.42927128e-01 1.11148489e+00
-1.45324707e+00 -9.87836301e-01 -1.41554996e-01 3.01637977e-01
9.15032923e-01 -5.13422370e-01 8.49623919e-01 7.50854015e-01
-2.62662470e-01 7.50868440e-01 3.81933361e-01 1.44463763e-01
9.73548532e-01 -1.15422356e+00 3.65997642e-01 9.76880312e-01
1.50541291e-01 1.38737428e+00 6.43042684e-01 -6.68608606e-01
-1.29635739e+00 -1.14397299e+00 1.22542763e+00 -6.84554994e-01
8.92786205e-01 -1.50111288e-01 -9.97473836e-01 6.49157584e-01
4.70938742e-01 -7.03886151e-01 5.17063439e-01 1.82661980e-01
-5.60707331e-01 -7.29876710e-03 -6.63580835e-01 4.73790467e-01
3.47174376e-01 -9.39741969e-01 -6.28000438e-01 3.02839935e-01
6.40832543e-01 -2.97775030e-01 -1.14535069e+00 2.59017706e-01
5.15688837e-01 -9.07076061e-01 6.11591995e-01 -3.63435835e-01
5.77442348e-01 -4.62924480e-01 -3.49957615e-01 -1.32790017e+00
-2.47193098e-01 -4.67151374e-01 5.03109515e-01 1.44637704e+00
5.90479791e-01 -4.62928802e-01 -1.32156402e-01 1.52221890e-02
-1.94288284e-01 -4.45896357e-01 -3.72190952e-01 -8.78626049e-01
6.28014684e-01 -2.39998698e-01 5.21036565e-01 1.17479229e+00
8.00297260e-02 4.35427159e-01 -2.17975363e-01 -4.31435406e-02
6.53618574e-02 4.16310936e-01 9.94733393e-01 -9.85942960e-01
-4.71265256e-01 -7.42107689e-01 3.43236551e-02 -7.86870241e-01
5.08505940e-01 -1.45570815e+00 1.17520802e-01 -1.28789020e+00
9.01970088e-01 -4.89717066e-01 -1.02981450e-02 5.39276004e-01
-2.19539851e-01 -7.99327269e-02 3.25032949e-01 5.34377813e-01
-5.99150121e-01 6.34683594e-02 7.30577826e-01 -4.47254539e-01
-2.66400188e-01 -1.38868138e-01 -8.24761569e-01 7.74912417e-01
4.74946409e-01 -7.81397939e-01 -3.13099444e-01 -8.48869681e-01
6.71080232e-01 4.06917900e-01 -4.07903105e-01 -7.53357351e-01
5.60052432e-02 -2.90799409e-01 -4.90811139e-01 -2.06263214e-01
-8.01689386e-01 -3.98136377e-01 2.96416968e-01 6.27251029e-01
-3.00386339e-01 6.33104265e-01 5.02277792e-01 -1.16030239e-01
-3.29269618e-01 -9.70231652e-01 9.43712592e-01 -3.15431744e-01
-8.12384546e-01 -2.99346328e-01 -7.23480582e-01 3.84089708e-01
6.79026186e-01 1.71779662e-01 -5.26133478e-01 4.77155484e-02
-1.47032425e-01 1.58322096e-01 8.79885972e-01 9.57741737e-01
1.34874377e-02 -1.13871932e+00 -9.40961123e-01 1.61152128e-02
6.08204722e-01 -5.07047832e-01 -3.82493511e-02 8.01022530e-01
-1.04890931e+00 7.17185438e-01 -2.28737026e-01 -6.74928367e-01
-1.52281952e+00 4.55285043e-01 2.54264027e-01 -7.12427616e-01
-1.48464784e-01 4.88117129e-01 5.54198265e-01 -1.24539113e+00
-6.40726388e-02 -3.43947172e-01 -2.09408343e-01 -1.00838907e-01
3.24333131e-01 2.61075348e-01 2.71984696e-01 -7.52097845e-01
-5.82110107e-01 9.44149971e-01 -8.15316588e-02 3.12614024e-01
1.26657569e+00 -7.27234110e-02 -9.58334208e-01 6.34467125e-01
1.30242860e+00 8.87058914e-01 -4.31292117e-01 -1.77642494e-01
5.57106614e-01 -3.06918651e-01 -4.13939238e-01 -8.06870818e-01
-9.78639245e-01 6.43751681e-01 2.37615600e-01 3.41385370e-03
9.52801287e-01 1.97989270e-01 4.35491472e-01 5.20749331e-01
7.39115357e-01 -8.46408010e-01 -6.89026564e-02 9.80397165e-01
8.76343727e-01 -1.06184196e+00 -2.45163478e-02 -3.26597802e-02
-6.53667808e-01 1.37632298e+00 6.87651992e-01 1.10934317e-01
3.20458502e-01 4.68330801e-01 2.58118004e-01 -1.38833448e-01
-7.72147954e-01 2.12390080e-01 3.68602306e-01 9.23696682e-02
1.08465266e+00 9.00391936e-02 -6.95058525e-01 4.13093865e-01
-3.14442217e-01 -2.79748857e-01 8.77389848e-01 8.78292263e-01
-3.57244313e-01 -1.39722586e+00 -4.39317524e-01 3.99431884e-01
-7.25125790e-01 -5.41212022e-01 -3.22373211e-01 7.15955079e-01
1.47297278e-01 8.72197449e-01 -4.73140597e-01 -4.25240844e-01
2.74363816e-01 9.01979506e-02 4.08652991e-01 -1.21781552e+00
-1.06841302e+00 -1.05929576e-01 -4.76423837e-02 -3.96900773e-01
-3.03157657e-01 -4.94028747e-01 -1.09075809e+00 -4.30373371e-01
-4.28703666e-01 5.23408055e-01 7.31460690e-01 7.87612975e-01
2.11288020e-01 7.54305065e-01 4.13022995e-01 -5.24585664e-01
-4.36288208e-01 -9.75416303e-01 -2.05521822e-01 6.81391209e-02
2.28307545e-01 -2.39894375e-01 -3.19156319e-01 3.56236964e-01] | [7.628871917724609, 7.928544521331787] |
35243e67-92ee-4112-97cb-4cab35a220e1 | id-mixgcl-identity-mixup-for-graph | 2304.10045 | null | https://arxiv.org/abs/2304.10045v1 | https://arxiv.org/pdf/2304.10045v1.pdf | ID-MixGCL: Identity Mixup for Graph Contrastive Learning | Recently developed graph contrastive learning (GCL) approaches compare two different "views" of the same graph in order to learn node/graph representations. The core assumption of these approaches is that by graph augmentation, it is possible to generate several structurally different but semantically similar graph structures, and therefore, the identity labels of the original and augmented graph/nodes should be identical. However, in this paper, we observe that this assumption does not always hold, for example, any perturbation to nodes or edges in a molecular graph will change the graph labels to some degree. Therefore, we believe that augmenting the graph structure should be accompanied by an adaptation of the labels used for the contrastive loss. Based on this idea, we propose ID-MixGCL, which allows for simultaneous modulation of both the input graph and the corresponding identity labels, with a controllable degree of change, leading to the capture of fine-grained representations from unlabeled graphs. Experimental results demonstrate that ID-MixGCL improves performance on graph classification and node classification tasks, as demonstrated by significant improvements on the Cora, IMDB-B, and IMDB-M datasets compared to state-of-the-art techniques, by 3-29% absolute points. | ['Chuan Zhou', 'Tingwen Liu', 'Xinghua Zhang', 'Jiangxia Cao', 'Bowen Yu', 'Gehang Zhang'] | 2023-04-20 | null | null | null | null | ['graph-classification'] | ['graphs'] | [ 4.55354691e-01 4.28693295e-01 -2.61702597e-01 -1.88183516e-01
-3.06328893e-01 -7.39753425e-01 5.97832322e-01 7.17627704e-01
-6.37825057e-02 7.61176884e-01 -1.53269887e-01 -3.49252075e-01
5.80016449e-02 -1.02969623e+00 -9.35099781e-01 -9.56124544e-01
-9.91113335e-02 4.82728213e-01 2.45390818e-01 -2.77674109e-01
-4.34546061e-02 8.39362025e-01 -1.24636924e+00 1.61906108e-01
6.39367998e-01 6.10339046e-01 -6.45241514e-02 4.17071283e-01
-1.50893733e-01 6.58601105e-01 -2.76265770e-01 -5.01293659e-01
2.49654934e-01 -4.96355355e-01 -8.85386527e-01 1.28700286e-01
6.87549770e-01 2.36273527e-01 -5.96631944e-01 1.26940835e+00
3.17080826e-01 1.37555167e-01 9.51127887e-01 -1.31395245e+00
-5.69280207e-01 5.82628071e-01 -7.35810280e-01 6.11418486e-02
2.96349287e-01 -3.68212499e-02 1.33380330e+00 -5.64371169e-01
9.26846921e-01 1.30131316e+00 2.99100339e-01 5.88371992e-01
-1.85707009e+00 -4.74139780e-01 5.20601034e-01 7.21659884e-02
-1.41044772e+00 -8.42553377e-02 1.07782006e+00 -4.06481206e-01
6.46644413e-01 3.32279265e-01 6.30396843e-01 1.01232839e+00
2.64761925e-01 3.21164608e-01 1.15303624e+00 -5.88430524e-01
2.03457370e-01 -2.05387194e-02 1.92446396e-01 1.15506530e+00
4.47248101e-01 3.39003727e-02 -2.86572009e-01 -2.00795680e-01
3.89773279e-01 -1.90904960e-02 -3.80843818e-01 -9.16884959e-01
-8.76295686e-01 8.71838450e-01 9.65701938e-01 4.12975997e-01
-1.55056193e-01 2.75270957e-02 4.84297812e-01 4.87602025e-01
5.13202786e-01 5.68749368e-01 -1.41085843e-02 7.32128024e-01
-2.79020816e-01 -1.05618034e-02 6.93462193e-01 8.84100378e-01
1.11217999e+00 -9.32652969e-03 -1.18524991e-01 3.96766454e-01
2.08855435e-01 1.40687451e-01 1.03152767e-01 -4.11202937e-01
2.77873814e-01 9.62319016e-01 -3.69685203e-01 -1.15394926e+00
-5.41980028e-01 -5.93613088e-01 -1.05979264e+00 2.41958782e-01
4.55859393e-01 2.79213399e-01 -1.08236265e+00 2.17897987e+00
3.22497696e-01 1.50036022e-01 -1.00971423e-01 5.32476366e-01
8.59132111e-01 3.80406767e-01 2.97836721e-01 -2.36636519e-01
1.17474830e+00 -7.75555372e-01 -5.19981265e-01 -1.81553245e-01
8.16543877e-01 -2.45070130e-01 1.19537354e+00 4.13952395e-02
-7.78607190e-01 -4.81772929e-01 -1.20337403e+00 6.15637898e-02
-6.27606153e-01 -3.47572356e-01 6.29541993e-01 5.93612254e-01
-9.21055794e-01 7.80555189e-01 -5.40603161e-01 -3.18475425e-01
2.57749021e-01 3.53101790e-01 -6.62089169e-01 -3.00531000e-01
-1.02478087e+00 4.31233913e-01 5.96267045e-01 -1.41082302e-01
-8.23665380e-01 -6.44068360e-01 -9.55088854e-01 2.54656464e-01
4.44302559e-01 -6.02858007e-01 4.98903394e-01 -9.26695406e-01
-9.61378753e-01 1.16587937e+00 -1.74196735e-02 -4.41271186e-01
3.82502168e-01 4.97484028e-01 -4.12830234e-01 2.77619720e-01
-9.72948000e-02 6.62911892e-01 8.12380135e-01 -1.41465187e+00
-4.05601636e-02 -6.08452022e-01 3.51795316e-01 1.73818663e-01
-2.38172397e-01 -4.48114693e-01 -3.50915253e-01 -6.19830906e-01
2.67071635e-01 -1.09636855e+00 -2.14743048e-01 -1.81054545e-03
-5.62686324e-01 1.24205440e-01 7.59018421e-01 -4.06281918e-01
9.25977707e-01 -2.27115989e+00 4.64615911e-01 4.62223470e-01
7.53253520e-01 1.13050148e-01 -3.84925187e-01 5.49472034e-01
-5.07216990e-01 3.27206224e-01 -2.77558178e-01 -1.79033384e-01
-1.98545620e-01 1.68807402e-01 -6.74868599e-02 6.82280779e-01
1.37075335e-01 9.61960435e-01 -1.05512989e+00 -4.14098322e-01
2.85732001e-01 4.24963534e-01 -4.47064817e-01 3.78685258e-02
-4.26587820e-01 6.84910834e-01 -2.56795973e-01 4.29840386e-01
6.35437310e-01 -6.56020284e-01 8.46967876e-01 -3.54184061e-01
3.41364861e-01 3.27798724e-01 -8.23891640e-01 1.41517317e+00
-7.09925145e-02 2.28043109e-01 -2.10429817e-01 -1.22154343e+00
9.58174467e-01 1.63200915e-01 4.47330117e-01 -7.44053125e-01
-8.94048735e-02 -4.69478033e-02 1.56578362e-01 2.34648764e-01
1.61946863e-01 -1.59236610e-01 -5.44894338e-02 3.33708644e-01
3.60355973e-02 -6.06448427e-02 2.81485885e-01 4.77248043e-01
1.14704239e+00 -2.54045397e-01 3.92398208e-01 -3.47318113e-01
7.46078670e-01 -2.45233640e-01 4.44916010e-01 6.91139698e-01
1.43577652e-02 3.00230831e-01 9.20800507e-01 -3.53394777e-01
-8.78558159e-01 -1.07423103e+00 1.28914937e-01 9.23612475e-01
2.09942460e-01 -5.14445961e-01 -8.32323313e-01 -1.05278099e+00
7.26149767e-04 3.99495453e-01 -6.85779154e-01 -5.85643053e-01
-5.60227990e-01 -7.53210604e-01 2.33953997e-01 1.39876753e-01
2.68156081e-01 -8.71521294e-01 1.48878336e-01 8.10101479e-02
2.34029084e-01 -1.07508326e+00 -4.72979486e-01 2.07371235e-01
-8.76164675e-01 -1.38446879e+00 -2.85958797e-01 -9.89903331e-01
1.14996421e+00 4.29496258e-01 1.19257700e+00 4.20011640e-01
-2.56750911e-01 1.85718760e-01 -3.33846420e-01 1.74368620e-01
-7.96729445e-01 3.26253653e-01 -2.48878285e-01 1.90573335e-01
-2.14689717e-01 -8.67552400e-01 -3.91979247e-01 2.94214557e-03
-1.03457928e+00 1.23041943e-01 4.01146263e-01 9.02592957e-01
8.85138094e-01 7.76085854e-02 4.63724703e-01 -1.44803452e+00
3.97556484e-01 -2.95503438e-01 -7.20862329e-01 5.01221299e-01
-1.00796926e+00 3.85256529e-01 9.13414061e-01 -2.81586319e-01
-6.48176789e-01 3.37739065e-02 3.04172561e-02 -4.91699487e-01
7.68960714e-02 6.22301638e-01 -5.46911955e-01 -4.41792816e-01
6.52949333e-01 1.31862417e-01 8.67414661e-03 -2.97168106e-01
7.25704193e-01 -2.58366652e-02 3.64389926e-01 -5.13439357e-01
8.90203238e-01 5.08417666e-01 6.85041308e-01 -8.01375687e-01
-5.20577848e-01 -2.83745795e-01 -6.20975852e-01 -1.49145886e-01
6.34386778e-01 -6.75450206e-01 -6.52420223e-01 4.28755522e-01
-7.66134381e-01 -1.32818446e-01 -3.32301378e-01 1.22030802e-01
-3.94282013e-01 6.88091874e-01 -6.83671534e-01 -3.15147668e-01
-2.81344742e-01 -1.21026146e+00 8.18124115e-01 -2.17766017e-02
1.35535732e-01 -1.26553524e+00 3.25793698e-02 4.95779840e-03
1.97211225e-02 6.23704553e-01 1.71343851e+00 -7.78539419e-01
-7.54253030e-01 -2.60254174e-01 -1.70460299e-01 2.35251471e-01
4.66272026e-01 -1.66361064e-01 -7.05476940e-01 -7.87529886e-01
-3.71518552e-01 -3.36838573e-01 8.90431821e-01 1.27642795e-01
1.07219434e+00 -3.47494185e-01 -5.61890602e-01 5.14007628e-01
1.44346154e+00 -7.81380087e-02 5.81962347e-01 -1.57185793e-01
1.06898069e+00 5.33711135e-01 2.04390839e-01 -2.48194230e-03
1.07818142e-01 7.60835588e-01 7.02010453e-01 -2.96708822e-01
-3.10624987e-01 -5.78204513e-01 1.24556154e-01 7.85295606e-01
2.65533328e-01 -5.91249526e-01 -5.07928073e-01 1.68697342e-01
-1.55584157e+00 -6.62970722e-01 -2.78156042e-01 2.52163744e+00
5.56206465e-01 1.83285117e-01 4.26996760e-02 1.27412930e-01
1.07850945e+00 6.31013811e-01 -8.13132405e-01 -2.02223539e-01
-2.03852043e-01 2.96069771e-01 5.53432524e-01 7.12077558e-01
-9.83854413e-01 9.03269231e-01 6.23750973e+00 7.15705633e-01
-1.34154773e+00 -6.86764717e-02 8.38051677e-01 3.25586528e-01
-6.44288063e-01 2.74009198e-01 -4.70290035e-01 3.47145796e-01
7.03732252e-01 -3.33458006e-01 5.21625996e-01 6.07687354e-01
-1.91179082e-01 4.86979574e-01 -1.09832346e+00 6.77997231e-01
-7.21078292e-02 -1.39651227e+00 4.72667158e-01 3.24850947e-01
6.02991581e-01 -1.14566654e-01 -6.18668925e-03 1.88224405e-01
2.87842959e-01 -9.91267323e-01 3.29550326e-01 3.69010150e-01
8.01174045e-01 -7.70760655e-01 3.86911929e-01 1.03652939e-01
-1.36199582e+00 4.51277375e-01 -2.22121477e-01 2.77562410e-01
-2.03421965e-01 6.72986865e-01 -8.16595256e-01 8.94510746e-01
1.78521648e-01 6.85620546e-01 -6.91548705e-01 4.95463639e-01
-4.59452182e-01 4.07035917e-01 6.14821166e-02 2.18162745e-01
7.17698857e-02 -5.18424869e-01 5.57897389e-01 8.04521918e-01
-1.33819655e-01 -2.41589487e-01 4.77237552e-01 9.64332640e-01
-6.76732063e-01 2.52276331e-01 -8.14238846e-01 -3.48541021e-01
4.96870369e-01 1.31303513e+00 -7.58996308e-01 -2.38205478e-01
-4.64038014e-01 9.69170034e-01 7.71522462e-01 3.41079056e-01
-6.78229034e-01 -8.00886229e-02 4.69445705e-01 2.92391181e-01
1.95309401e-01 -3.34439352e-02 3.08984846e-01 -9.57108557e-01
-1.45477066e-02 -9.08206224e-01 5.85627556e-01 -3.99485767e-01
-1.31331360e+00 5.99479437e-01 -2.05912173e-01 -9.20563579e-01
-8.84441361e-02 -6.13178194e-01 -3.81190836e-01 8.45036149e-01
-1.50852168e+00 -1.17855895e+00 -3.43360275e-01 4.29481775e-01
-1.35615274e-01 1.86542813e-02 9.20436800e-01 2.16847390e-01
-4.71676856e-01 6.87597334e-01 5.97003475e-02 4.08576243e-02
5.02176702e-01 -1.41439319e+00 7.03496754e-01 7.09513545e-01
4.69951332e-01 4.86280829e-01 6.67981029e-01 -6.52688622e-01
-1.44575417e+00 -1.28244245e+00 5.70475459e-01 -9.83562041e-03
5.52480221e-01 -5.83774686e-01 -1.15621400e+00 8.30755711e-01
-1.09665945e-01 4.37848777e-01 5.30621290e-01 1.77573022e-02
-6.92473829e-01 -1.04690365e-01 -1.32428646e+00 7.18678176e-01
1.26280129e+00 -7.37681925e-01 -5.82444891e-02 6.19149745e-01
9.41590309e-01 -1.65057421e-01 -9.68112946e-01 5.79702258e-01
2.00455993e-01 -8.89432669e-01 9.70612168e-01 -9.69686270e-01
-2.98287533e-02 -2.99467325e-01 -1.62870765e-01 -1.36700559e+00
-4.92965370e-01 -5.35495281e-01 -1.55026019e-01 9.74346101e-01
3.84390026e-01 -8.35100710e-01 9.52739298e-01 2.61846751e-01
-3.32647674e-02 -4.89878267e-01 -1.00087428e+00 -8.62888575e-01
9.35732052e-02 1.65963992e-01 6.09182417e-01 1.09412432e+00
-1.88437402e-01 7.24259794e-01 -2.55086869e-01 3.01982582e-01
5.62311232e-01 1.56390667e-01 7.25945413e-01 -1.35460973e+00
-4.81522411e-01 -4.05334949e-01 -7.50834703e-01 -8.12276304e-01
5.09398520e-01 -1.49561441e+00 -4.25303102e-01 -1.29216838e+00
1.76030383e-01 -4.08211976e-01 -4.18599874e-01 3.82979482e-01
-2.93502331e-01 1.15379058e-01 9.89869535e-02 1.21231310e-01
-3.95388514e-01 6.04593039e-01 1.22722089e+00 -4.58332539e-01
-9.72476825e-02 -1.86121926e-01 -7.92147994e-01 3.48906428e-01
6.50513232e-01 -4.91922408e-01 -5.86599112e-01 1.01433247e-01
1.80241838e-01 -7.95770844e-04 3.26823175e-01 -7.06897974e-01
-2.02089846e-01 7.10476339e-02 -2.67423131e-02 -1.34127796e-01
3.28886569e-01 -6.48218632e-01 3.83110136e-01 7.55763888e-01
-4.34009403e-01 -7.94173256e-02 6.98323846e-02 8.59539330e-01
-1.25651538e-01 -2.10202232e-01 1.02479815e+00 -2.24907100e-01
-4.33809102e-01 5.54778337e-01 2.79977955e-02 2.72515304e-02
9.17841136e-01 3.64844166e-02 -5.35212159e-01 -2.98955917e-01
-7.87683427e-01 -1.35711342e-01 8.32917213e-01 1.38507038e-01
3.24534118e-01 -1.29733205e+00 -5.58422029e-01 2.47823760e-01
4.90735471e-01 -1.47636741e-01 1.46417826e-01 5.60235083e-01
-4.32176530e-01 2.01037243e-01 -9.87306759e-02 -4.83018368e-01
-1.33149612e+00 1.08498788e+00 4.07184452e-01 -6.58632040e-01
-5.52848399e-01 5.57744741e-01 6.25537038e-01 -5.61706543e-01
1.81823422e-03 -8.38778168e-02 -8.30539539e-02 -1.59197822e-01
1.27461493e-01 5.95672205e-02 2.66385913e-01 -5.68906724e-01
-2.51684308e-01 4.37040299e-01 -5.16458988e-01 3.67325693e-01
1.02405834e+00 9.62802023e-02 -1.91273481e-01 3.29103023e-01
1.23429310e+00 1.94993705e-01 -1.12079859e+00 -4.17747766e-01
-3.44932005e-02 -3.14646304e-01 -1.27932891e-01 -4.86321777e-01
-1.29213309e+00 6.44594967e-01 4.93520260e-01 4.24856931e-01
1.07203650e+00 2.88440317e-01 2.33053505e-01 3.52922380e-01
5.33025086e-01 -3.79081637e-01 1.68878242e-01 -6.86363131e-02
7.94341803e-01 -1.02609134e+00 7.09738210e-02 -9.47144926e-01
-3.45686764e-01 7.34993458e-01 4.99413371e-01 -1.66273221e-01
3.42369944e-01 -1.41793549e-01 -3.70245516e-01 -5.81276894e-01
-5.70281208e-01 -2.05289811e-01 4.19217199e-01 5.82765877e-01
3.65530014e-01 2.92204291e-01 -1.86162010e-01 -1.65265575e-01
1.79202959e-01 -6.80687666e-01 4.62833524e-01 7.48664558e-01
-2.25530699e-01 -1.50158989e+00 1.24189556e-02 3.99018556e-01
-2.04737708e-01 -1.91770807e-01 -7.87878692e-01 1.05882847e+00
-2.91271597e-01 6.36344314e-01 -1.39946416e-01 -3.56805414e-01
2.86441118e-01 1.09674506e-01 7.67822385e-01 -8.29583108e-01
-2.55953163e-01 -8.63344148e-02 1.18070943e-02 -2.38177449e-01
-3.40770572e-01 -3.89801413e-01 -1.33068645e+00 -4.19488609e-01
-3.46929908e-01 2.13424489e-01 3.15820068e-01 6.98480666e-01
5.07472813e-01 7.32274830e-01 7.85646260e-01 -5.12145638e-01
-3.40101212e-01 -7.88719594e-01 -8.73159528e-01 7.20607519e-01
1.01893120e-01 -7.40541875e-01 -5.81677973e-01 -1.71626389e-01] | [7.191321849822998, 6.270619869232178] |
e6c2f0e2-8503-4a23-ab56-2967e128cd49 | deep-multimodal-representation-learning-from | 1704.03152 | null | http://arxiv.org/abs/1704.03152v1 | http://arxiv.org/pdf/1704.03152v1.pdf | Deep Multimodal Representation Learning from Temporal Data | In recent years, Deep Learning has been successfully applied to multimodal
learning problems, with the aim of learning useful joint representations in
data fusion applications. When the available modalities consist of time series
data such as video, audio and sensor signals, it becomes imperative to consider
their temporal structure during the fusion process. In this paper, we propose
the Correlational Recurrent Neural Network (CorrRNN), a novel temporal fusion
model for fusing multiple input modalities that are inherently temporal in
nature. Key features of our proposed model include: (i) simultaneous learning
of the joint representation and temporal dependencies between modalities, (ii)
use of multiple loss terms in the objective function, including a maximum
correlation loss term to enhance learning of cross-modal information, and (iii)
the use of an attention model to dynamically adjust the contribution of
different input modalities to the joint representation. We validate our model
via experimentation on two different tasks: video- and sensor-based activity
classification, and audio-visual speech recognition. We empirically analyze the
contributions of different components of the proposed CorrRNN model, and
demonstrate its robustness, effectiveness and state-of-the-art performance on
multiple datasets. | ['Palghat Ramesh', 'Jiebo Luo', 'Sriganesh Madhvanath', 'Edgar A. Bernal', 'Xitong Yang', 'Radha Chitta'] | 2017-04-11 | deep-multimodal-representation-learning-from-1 | http://openaccess.thecvf.com/content_cvpr_2017/html/Yang_Deep_Multimodal_Representation_CVPR_2017_paper.html | http://openaccess.thecvf.com/content_cvpr_2017/papers/Yang_Deep_Multimodal_Representation_CVPR_2017_paper.pdf | cvpr-2017-7 | ['audio-visual-speech-recognition'] | ['speech'] | [ 4.10227567e-01 -5.21007240e-01 -7.64274299e-02 -3.21550429e-01
-1.06980503e+00 -2.10970193e-01 8.04325938e-01 2.63222486e-01
-5.50922751e-01 5.32960176e-01 4.93537009e-01 2.55815238e-02
-5.06582916e-01 -2.24267438e-01 -6.13619149e-01 -8.00930679e-01
-2.81863451e-01 -1.66907668e-01 2.72403866e-01 -1.12665944e-01
-4.14875746e-02 3.44450951e-01 -1.78116572e+00 3.42614144e-01
6.90085948e-01 1.37386453e+00 1.92318246e-01 5.85136712e-01
5.08020818e-02 1.00457311e+00 -4.99158770e-01 -9.66421440e-02
-8.41429085e-02 -3.99001807e-01 -5.61328769e-01 1.33436233e-01
-6.05576746e-02 1.47965193e-01 -5.31778157e-01 7.33493268e-01
4.48575348e-01 6.87458217e-01 5.40174603e-01 -1.37402654e+00
-1.23982199e-01 6.30602837e-01 -7.54288495e-01 4.75561947e-01
4.09950346e-01 -7.30615947e-03 9.75518763e-01 -7.28753626e-01
2.64102995e-01 1.14043748e+00 4.05514956e-01 3.42355847e-01
-1.04749966e+00 -4.56714064e-01 4.99119490e-01 4.93031174e-01
-1.35846865e+00 -7.25312889e-01 1.02241087e+00 -4.65440243e-01
9.47558522e-01 1.34921640e-01 3.55577171e-01 1.25039291e+00
2.41896868e-01 1.03479242e+00 8.91814947e-01 -3.50123554e-01
1.12598352e-01 -1.70722172e-01 1.00493893e-01 3.16437244e-01
-4.07700062e-01 7.85009935e-02 -1.05839550e+00 -2.26573274e-01
4.59945947e-01 3.17854911e-01 -2.98504233e-01 -1.46336183e-01
-1.16525829e+00 5.19285619e-01 3.89715970e-01 6.36048317e-01
-6.14381969e-01 2.37473026e-01 4.72868949e-01 2.91942656e-01
5.68419337e-01 -2.99504232e-02 -3.21701854e-01 -2.03412592e-01
-7.77919948e-01 -7.80569091e-02 2.40844145e-01 3.49380374e-01
3.36596489e-01 9.06478465e-02 -2.96399593e-01 1.09749210e+00
5.00688434e-01 2.76464552e-01 8.24901640e-01 -8.14514577e-01
6.86023831e-01 4.09029096e-01 1.49989324e-02 -8.85242224e-01
-4.76158381e-01 -2.21782848e-01 -9.40459430e-01 -9.99252871e-02
5.69762066e-02 -2.43729517e-01 -8.01445484e-01 2.19522572e+00
2.18861088e-01 4.86298263e-01 2.43457168e-01 8.37207735e-01
6.65438116e-01 6.76525772e-01 3.06936979e-01 -5.94395816e-01
1.15844727e+00 -6.79214239e-01 -1.04188335e+00 -7.40198269e-02
2.63091892e-01 -5.86880147e-01 4.16373879e-01 3.17785829e-01
-1.14501774e+00 -6.59818649e-01 -1.03980517e+00 1.26835063e-01
-3.41001004e-01 4.38614301e-02 4.64891493e-01 6.65049031e-02
-8.30051899e-01 5.77885270e-01 -1.08510947e+00 -2.85191655e-01
2.08925918e-01 5.85347354e-01 -3.81421030e-01 8.22119862e-02
-1.21793067e+00 6.06324553e-01 4.79865551e-01 3.18141729e-01
-8.11118364e-01 -3.69351268e-01 -8.87876034e-01 1.37552634e-01
5.43217480e-01 -5.71539998e-01 1.14682400e+00 -1.21899831e+00
-1.48842812e+00 1.47273824e-01 -3.03688705e-01 -4.46418166e-01
1.92784160e-01 -2.99919724e-01 -7.22795486e-01 1.84379563e-01
-3.99517953e-01 4.66687977e-01 9.56959367e-01 -9.75439370e-01
-7.51153290e-01 -4.05111879e-01 -7.46823847e-02 3.44813287e-01
-4.43441153e-01 4.13418859e-02 -3.23972672e-01 -8.78389060e-01
1.16148345e-01 -6.80275500e-01 -5.61780110e-02 -2.33646721e-01
2.95053907e-02 -4.51114237e-01 9.72866356e-01 -7.75783479e-01
1.26607430e+00 -2.26849365e+00 7.97934115e-01 1.87630743e-01
8.35124627e-02 2.44668305e-01 -3.37645113e-01 5.12368441e-01
-1.20039299e-01 -2.84899563e-01 -1.64871216e-01 -7.31309652e-01
-2.23294929e-01 1.58099651e-01 -2.38820106e-01 3.17135513e-01
2.59728730e-01 7.03457296e-01 -6.81310534e-01 -3.34932566e-01
3.90927047e-01 8.24384451e-01 1.21523701e-01 3.66497248e-01
-1.93552911e-01 4.62142378e-01 -2.57278204e-01 5.19332945e-01
1.95144877e-01 -1.24972671e-01 1.41456217e-01 -5.16665399e-01
-2.70959288e-02 1.88483775e-01 -1.19941771e+00 1.90887618e+00
-4.43553001e-01 5.75664580e-01 7.40833655e-02 -1.00206280e+00
6.69488251e-01 7.53699720e-01 8.33034098e-01 -8.72676253e-01
3.16655010e-01 7.67974630e-02 -1.04000621e-01 -5.89257598e-01
5.01140535e-01 3.87608223e-02 -3.15662399e-02 4.75453556e-01
4.55343485e-01 4.36790705e-01 1.50458708e-01 5.25251776e-02
1.01689601e+00 -5.21158874e-02 1.21648580e-01 3.23584020e-01
7.38610804e-01 -7.66982734e-01 5.63682377e-01 4.62015122e-01
-5.50615601e-02 4.28206742e-01 2.10759908e-01 -8.32108036e-02
-5.88149250e-01 -7.73059964e-01 2.47016281e-01 1.24170256e+00
-2.23349240e-02 -2.83122241e-01 -2.03131273e-01 -5.77341020e-01
-2.01656073e-01 3.31467122e-01 -6.59561276e-01 -3.53187233e-01
-3.71406347e-01 -6.85384333e-01 4.04204935e-01 7.71798074e-01
1.83308244e-01 -8.91947508e-01 -6.64400816e-01 2.74984449e-01
-5.65491915e-01 -1.20195651e+00 -5.07956147e-01 3.69623899e-01
-9.14215088e-01 -8.96142840e-01 -5.73087633e-01 -4.17807996e-01
2.19709501e-01 4.51564580e-01 5.62152267e-01 -2.60085255e-01
1.30880186e-02 9.04849350e-01 -5.44338942e-01 -1.50287837e-01
-7.17992857e-02 -3.47616300e-02 1.09338105e-01 7.21567512e-01
-4.06830870e-02 -7.52727151e-01 -3.72136354e-01 2.70992041e-01
-1.18301809e+00 -8.34680200e-02 4.36287493e-01 9.13482547e-01
4.96808231e-01 9.94163230e-02 6.34241343e-01 -1.64878175e-01
7.21817017e-01 -5.94316006e-01 -4.06388253e-01 5.52145004e-01
-1.68901652e-01 1.88970849e-01 2.76298285e-01 -7.95632184e-01
-1.19463849e+00 -7.49562383e-02 2.99789328e-02 -8.74157369e-01
1.52866825e-01 9.44618702e-01 -1.06965214e-01 7.14144781e-02
2.20837310e-01 1.94637284e-01 1.38926832e-02 -4.90136027e-01
2.42398292e-01 5.04049838e-01 5.79601645e-01 -5.23558617e-01
2.67167330e-01 3.55744034e-01 -5.69463782e-02 -8.12753201e-01
-5.93733191e-01 -5.84805846e-01 -4.68211532e-01 -3.71892095e-01
7.27868617e-01 -1.09088445e+00 -8.06347549e-01 7.39116549e-01
-1.07569468e+00 7.98253343e-03 -2.07757473e-01 8.47458124e-01
-4.46443260e-01 3.91142547e-01 -4.87628132e-01 -1.17217910e+00
-3.26766610e-01 -1.01366794e+00 9.84095514e-01 4.10543144e-01
-3.14922221e-02 -9.83184874e-01 9.02827978e-02 3.95788521e-01
3.44631165e-01 2.70291954e-01 6.13181949e-01 -6.18267477e-01
-2.65408367e-01 -1.27086908e-01 3.23852003e-02 4.11353618e-01
3.77468795e-01 4.42208350e-02 -1.11960304e+00 -3.98984402e-01
-2.57635899e-02 -3.93222600e-01 1.15191698e+00 4.70098436e-01
1.07639194e+00 -8.38110372e-02 -2.65944153e-01 3.26083332e-01
1.12324095e+00 5.60372055e-01 4.07327592e-01 -2.12916151e-01
6.42084599e-01 6.75384283e-01 5.29487193e-01 7.11551189e-01
3.37616384e-01 8.88550997e-01 4.80510086e-01 1.64076477e-01
1.13919944e-01 1.96638983e-02 6.39212966e-01 1.19675732e+00
-2.05863953e-01 -4.67600316e-01 -6.32177889e-01 6.17502332e-01
-2.46532202e+00 -1.07374287e+00 4.54695642e-01 2.31616879e+00
4.68780249e-01 -6.10621758e-02 3.09324771e-01 2.60060489e-01
5.69679320e-01 3.02890509e-01 -5.62813938e-01 4.57974635e-02
-3.05956483e-01 3.79704051e-02 -8.43791142e-02 3.56703192e-01
-1.35652053e+00 3.91527414e-01 5.78264761e+00 7.72284687e-01
-1.26466918e+00 2.40148500e-01 3.25212717e-01 -3.08756262e-01
-1.81222379e-01 -3.32405299e-01 -3.00854921e-01 4.65747863e-01
1.00873590e+00 1.57940798e-02 5.71288109e-01 4.24003080e-02
2.58551955e-01 -1.35362267e-01 -1.05901659e+00 1.20286870e+00
1.63125455e-01 -1.02741408e+00 -8.21039006e-02 -1.39811680e-01
4.86707151e-01 1.10476442e-01 1.56473815e-01 1.17684864e-01
8.85626581e-03 -8.74873936e-01 8.09060633e-01 9.58851635e-01
2.09156916e-01 -8.59251499e-01 7.76925385e-01 1.59875140e-01
-1.40869689e+00 -4.98028755e-01 2.98795521e-01 1.84252962e-01
2.84829497e-01 3.84701520e-01 -2.77431786e-01 1.01172316e+00
6.26876056e-01 9.75562513e-01 -4.51721191e-01 9.79060531e-01
1.51577905e-01 3.82877231e-01 -4.90641445e-01 2.08004400e-01
1.30725935e-01 3.24819125e-02 4.83910859e-01 1.02020156e+00
4.24379259e-01 1.39106467e-01 2.11437553e-01 3.60160202e-01
-8.84969309e-02 -9.01930779e-02 -3.74090046e-01 -2.82400727e-01
4.21727508e-01 1.00346112e+00 -3.63487452e-01 3.30246054e-02
-5.47273934e-01 5.61856389e-01 2.55995691e-01 4.58021134e-01
-7.88107336e-01 -1.38278812e-01 7.44205058e-01 -3.70975941e-01
5.39429128e-01 -3.08790624e-01 1.00046292e-01 -1.10756350e+00
1.92334071e-01 -6.93262398e-01 7.13192403e-01 -8.22548866e-01
-1.20799172e+00 6.42957330e-01 1.97081238e-01 -1.39376175e+00
-3.86683494e-01 -1.85509965e-01 -3.85460168e-01 7.19369650e-01
-1.29819930e+00 -1.25523472e+00 -7.80379474e-02 9.85633135e-01
4.79950577e-01 -3.66276741e-01 4.26892400e-01 6.26467824e-01
-6.80731297e-01 4.90340143e-01 1.26540689e-02 -1.16182968e-01
5.91895461e-01 -6.98802710e-01 -2.96395183e-01 9.21382308e-01
2.51883864e-01 4.48590487e-01 5.91470897e-01 -3.40223968e-01
-1.47886407e+00 -1.02822232e+00 5.72144866e-01 -2.63704415e-02
6.62092805e-01 -1.22719578e-01 -9.96024370e-01 4.91528571e-01
3.80353481e-01 3.28891561e-03 7.34884799e-01 1.32964551e-01
-4.75335330e-01 -4.36201811e-01 -8.77175212e-01 3.10448527e-01
6.22774541e-01 -8.08727801e-01 -5.58665454e-01 -7.03466265e-03
8.51542950e-01 -2.45132163e-01 -9.28271949e-01 5.55580556e-01
7.71197617e-01 -6.29387975e-01 8.36439729e-01 -4.79419231e-01
1.20075941e-01 -4.87267345e-01 -4.25208658e-01 -1.21280921e+00
-1.14771768e-01 -5.26955783e-01 -5.90478420e-01 1.35856295e+00
1.32044762e-01 -3.64043802e-01 6.73341081e-02 4.40256596e-01
-5.77316619e-02 -6.71785414e-01 -1.35531354e+00 -4.17128414e-01
-6.42389596e-01 -7.02899277e-01 3.22669268e-01 7.60951102e-01
8.00758228e-02 4.34097737e-01 -8.70433033e-01 2.04877868e-01
3.00863147e-01 -1.59058258e-01 3.05233806e-01 -1.19576716e+00
-4.33622330e-01 -4.09055889e-01 -4.39365476e-01 -8.36820424e-01
1.08566493e-01 -4.73778039e-01 8.61328933e-03 -1.28560424e+00
8.97461623e-02 -5.50224865e-03 -9.68350470e-01 6.75426960e-01
-1.47742778e-02 -1.53128773e-01 4.03636545e-01 1.83535665e-01
-8.49516928e-01 9.86697555e-01 6.26445293e-01 -3.58644098e-01
-3.97638083e-01 1.12768531e-01 -4.91480827e-01 4.76516515e-01
4.30745363e-01 -3.57624203e-01 -6.02561474e-01 -4.27857697e-01
5.11385463e-02 5.25458694e-01 3.85568053e-01 -1.07493365e+00
4.82372999e-01 -8.26106146e-02 2.04500049e-01 -6.23528242e-01
8.78404796e-01 -1.08227921e+00 3.90124857e-01 1.88347965e-01
-4.54954177e-01 5.07854298e-02 3.31624866e-01 8.26425791e-01
-5.47555804e-01 1.32967621e-01 6.16984785e-01 3.64102811e-01
-7.10452616e-01 2.41740659e-01 -3.83608311e-01 -3.77637684e-01
7.97271609e-01 5.39095365e-02 -1.47834972e-01 -4.11321133e-01
-7.86816180e-01 3.99486601e-01 -1.58346578e-01 9.20452893e-01
7.61177003e-01 -1.51102340e+00 -4.69864875e-01 1.00663051e-01
1.66415602e-01 -4.70797002e-01 6.31294429e-01 1.16391659e+00
3.55883300e-01 2.21388221e-01 -1.05432853e-01 -6.97612524e-01
-1.55530715e+00 5.64828098e-01 3.44332933e-01 -3.49166751e-01
-1.94113955e-01 5.69157720e-01 6.20580316e-02 -3.94003280e-02
6.50481462e-01 -3.64228994e-01 -5.32646716e-01 4.60822940e-01
6.06981754e-01 4.14312392e-01 9.11870673e-02 -7.73452103e-01
-5.05750239e-01 4.72255409e-01 -1.06911570e-01 -4.14543927e-01
1.36182725e+00 -3.52092326e-01 -9.17777792e-03 9.57236111e-01
1.08414018e+00 -5.01705647e-01 -1.29619634e+00 -7.85365760e-01
-1.14817940e-01 -1.26657680e-01 1.59865618e-01 -7.87823081e-01
-1.19774437e+00 8.31822217e-01 9.28857505e-01 1.46327555e-01
1.58135748e+00 7.94616155e-03 6.26907051e-01 1.11987188e-01
2.46175528e-01 -1.06882000e+00 3.11545670e-01 5.41325450e-01
8.29515100e-01 -1.11804485e+00 -1.22703128e-01 1.60215974e-01
-6.77420974e-01 1.02565289e+00 3.04040283e-01 3.62494946e-01
7.81993091e-01 7.56500289e-02 -1.98920742e-01 -2.53049331e-03
-1.25163329e+00 -3.61460358e-01 5.75016141e-01 3.27283710e-01
3.64151061e-01 -1.02110244e-01 -1.94123551e-01 6.92348003e-01
6.22383714e-01 8.33871141e-02 8.08919519e-02 1.07427251e+00
-1.18996792e-01 -9.63654757e-01 -3.64415675e-01 2.59166270e-01
-4.76800442e-01 3.51412952e-01 -1.07764676e-01 3.36871058e-01
3.58210988e-02 1.25047970e+00 -4.50230055e-02 -6.93807125e-01
3.17872435e-01 1.91834643e-01 5.17967880e-01 1.59074869e-02
-7.18309045e-01 4.86057758e-01 -1.67186737e-01 -5.21514416e-01
-1.12872577e+00 -9.45430934e-01 -8.87606084e-01 5.61368428e-02
-2.61073679e-01 1.48315728e-01 6.01225257e-01 1.33964944e+00
5.47078013e-01 8.86304140e-01 7.59813368e-01 -9.96151090e-01
-2.53593773e-01 -1.05532503e+00 -4.86506373e-01 3.90184969e-01
7.04281390e-01 -8.78063023e-01 -2.37017214e-01 5.85682020e-02] | [13.340847969055176, 5.037857532501221] |
80ec76b2-f3c0-4536-af5d-68f3e3dd1830 | using-hankel-matrices-for-dynamics-based | 1506.05001 | null | http://arxiv.org/abs/1506.05001v1 | http://arxiv.org/pdf/1506.05001v1.pdf | Using Hankel Matrices for Dynamics-based Facial Emotion Recognition and Pain Detection | This paper proposes a new approach to model the temporal dynamics of a
sequence of facial expressions. To this purpose, a sequence of Face Image
Descriptors (FID) is regarded as the output of a Linear Time Invariant (LTI)
system. The temporal dynamics of such sequence of descriptors are represented
by means of a Hankel matrix. The paper presents different strategies to compute
dynamics-based representation of a sequence of FID, and reports classification
accuracy values of the proposed representations within different standard
classification frameworks. The representations have been validated in two very
challenging application domains: emotion recognition and pain detection.
Experiments on two publicly available benchmarks and comparison with
state-of-the-art approaches demonstrate that the dynamics-based FID
representation attains competitive performance when off-the-shelf
classification tools are adopted. | ['Liliana Lo Presti', 'Marco La Cascia'] | 2015-06-16 | null | null | null | null | ['facial-emotion-recognition'] | ['computer-vision'] | [ 1.23773478e-01 -4.63960439e-01 -3.05105031e-01 -5.08094847e-01
-1.75831988e-01 -4.48940098e-01 1.00913835e+00 -2.47991383e-01
-1.79002717e-01 4.14896876e-01 -2.37459078e-01 2.92028695e-01
-3.19702834e-01 -1.44328758e-01 -1.53357461e-01 -9.13661480e-01
-6.96893394e-01 2.43515447e-02 -4.82998878e-01 -4.27804023e-01
1.29110023e-01 8.09569418e-01 -1.94309151e+00 1.82516783e-01
-1.05557172e-03 1.56243336e+00 -5.55274904e-01 6.17323875e-01
-4.19156328e-02 8.52328718e-01 -5.76379061e-01 -1.44313142e-01
1.43476740e-01 -4.21136081e-01 -6.46242499e-01 4.80779819e-02
3.22712421e-01 1.41950741e-01 -3.64781737e-01 8.73417556e-01
2.70497084e-01 2.96141058e-01 1.04052210e+00 -1.51698601e+00
-4.51255262e-01 -2.29550496e-01 -3.61708373e-01 5.37111104e-01
6.57893181e-01 -2.39900425e-01 5.32515824e-01 -9.14963245e-01
1.01437163e+00 1.36806655e+00 6.33526921e-01 6.70582950e-01
-1.40769553e+00 -6.77403212e-01 1.50291756e-01 4.43003625e-01
-1.56521702e+00 -5.66865385e-01 1.08235371e+00 -6.70307219e-01
9.15380001e-01 3.12519163e-01 8.21778536e-01 1.35845435e+00
6.95056677e-01 4.49483395e-01 1.37367034e+00 -2.29311511e-01
1.93210810e-01 3.17362659e-02 3.49686980e-01 9.58595812e-01
-5.00993013e-01 2.39039063e-01 -8.38048100e-01 -5.29257476e-01
3.89655322e-01 -2.25609258e-01 9.81878564e-02 -3.34582955e-01
-4.34496939e-01 8.56325567e-01 3.16984653e-02 7.79331446e-01
-5.68281353e-01 3.19328845e-01 8.81326854e-01 6.51136339e-01
7.33349383e-01 -3.83009732e-01 -1.58601135e-01 -3.68487626e-01
-8.39808881e-01 1.68625727e-01 7.74850011e-01 4.60069507e-01
4.61681873e-01 2.20029891e-01 -1.53192043e-01 5.88716328e-01
8.63324329e-02 3.85856509e-01 6.92876816e-01 -7.21389472e-01
-2.24073663e-01 3.45391870e-01 3.47730704e-02 -1.44791794e+00
-3.66050988e-01 -2.81700697e-02 -9.95156169e-01 3.06686312e-01
1.78707123e-01 1.11791454e-01 -8.10313165e-01 1.85831237e+00
3.52145016e-01 6.45807207e-01 1.14558749e-02 7.26980746e-01
4.73286420e-01 9.14837480e-01 1.90213859e-01 -8.20882976e-01
1.14562440e+00 -5.04946589e-01 -1.13247871e+00 6.03830934e-01
2.32810155e-01 -5.93126118e-01 3.95454198e-01 7.48488605e-01
-7.46892571e-01 -5.86797893e-01 -7.72254288e-01 3.99448335e-01
-4.78695482e-01 2.71769762e-01 6.35339856e-01 4.35697556e-01
-1.18710232e+00 7.98117578e-01 -8.34318519e-01 -4.63471591e-01
-7.79983923e-02 5.61623812e-01 -5.83357573e-01 5.43116748e-01
-1.08749318e+00 9.03151691e-01 -1.94877326e-01 4.72751498e-01
-9.55946624e-01 -3.72115701e-01 -7.72800148e-01 -2.57638276e-01
-2.83890098e-01 -1.16151907e-01 1.20593154e+00 -1.44396019e+00
-1.85511947e+00 1.10535800e+00 -3.91369224e-01 -2.43524924e-01
5.79778671e-01 5.11670634e-02 -8.25937569e-01 3.81204098e-01
-2.82807797e-01 1.57024816e-01 1.62128294e+00 -8.49380970e-01
-1.04665376e-01 -4.54334050e-01 -4.14502084e-01 -1.62952855e-01
-5.51760495e-01 4.11470950e-01 -4.09212969e-02 -5.12177467e-01
-2.29119077e-01 -1.19173169e+00 6.67533055e-02 1.28532529e-01
1.00080989e-01 -5.07693529e-01 1.38156402e+00 -4.80106235e-01
1.48577201e+00 -2.28224373e+00 4.48840529e-01 2.24268332e-01
-3.40517163e-01 3.38487089e-01 -2.20701903e-01 4.63847667e-01
-5.58395386e-01 -1.89327821e-01 5.58902510e-03 -3.75157297e-01
-1.63814247e-01 2.73123622e-01 -4.83319700e-01 9.38357294e-01
2.49849647e-01 5.08873224e-01 -6.40080750e-01 -3.04581225e-01
3.17831576e-01 8.26926053e-01 -7.67818764e-02 1.17034860e-01
4.10056338e-02 7.08872616e-01 -4.97184515e-01 4.88554150e-01
4.12398696e-01 2.72002399e-01 2.24718079e-01 -2.34379157e-01
-1.29450411e-01 -3.42892200e-01 -1.15721977e+00 1.61697805e+00
-5.74137390e-01 6.88129306e-01 -1.00901708e-01 -1.33219624e+00
1.21674752e+00 7.48881519e-01 1.00006747e+00 -6.93772733e-01
3.71181160e-01 -2.26261211e-03 -1.05014622e-01 -4.73763973e-01
1.99770480e-02 -1.25994250e-01 -1.60448521e-01 3.58804494e-01
4.01948094e-01 1.91273868e-01 1.72120303e-01 -3.40230823e-01
1.08907843e+00 1.73844814e-01 5.29055715e-01 -4.59414244e-01
1.09857416e+00 -2.52810776e-01 4.27169472e-01 1.70941100e-01
-4.56729621e-01 -1.53287396e-01 6.09439611e-01 -9.31648612e-01
-6.82018995e-01 -7.50040531e-01 -3.88526529e-01 1.01468658e+00
-2.54484057e-01 -3.17426205e-01 -6.57829821e-01 -3.38499725e-01
2.69553930e-01 8.60962793e-02 -1.27180648e+00 -3.22003663e-01
-4.91423488e-01 -6.16563857e-01 5.79701900e-01 2.95415998e-01
3.26329529e-01 -1.03545475e+00 -1.02805889e+00 3.32016051e-01
2.37062201e-01 -9.53122556e-01 -2.26921707e-01 6.87978193e-02
-6.48558497e-01 -9.29233432e-01 -6.61132812e-01 -6.24320090e-01
3.81873816e-01 -1.91056952e-01 7.98738718e-01 -2.07108960e-01
-8.82919788e-01 9.75857317e-01 -1.60769403e-01 -2.95791537e-01
-2.66011924e-01 -5.58543205e-01 4.36412543e-01 8.11382592e-01
1.46838307e-01 -6.25385165e-01 -4.80051517e-01 2.43215501e-01
-8.37516129e-01 -6.52454436e-01 -1.42581342e-02 9.93560195e-01
6.09737158e-01 -8.02124143e-02 3.48752499e-01 -3.40146840e-01
8.54961216e-01 -4.79416102e-01 -7.69344449e-01 2.25418389e-01
-4.85730499e-01 1.76905498e-01 5.54964542e-01 -9.55077529e-01
-1.12025332e+00 4.35053259e-01 3.17659855e-01 -1.01522219e+00
7.70052522e-02 4.77697670e-01 5.87721884e-01 -4.70023662e-01
4.42988634e-01 2.92813599e-01 3.06012064e-01 -3.92428219e-01
3.09494376e-01 5.18327773e-01 5.23959875e-01 -5.65242887e-01
5.15510142e-01 6.24063075e-01 4.45550412e-01 -9.66913581e-01
-4.70945746e-01 -5.15001476e-01 -7.34119296e-01 -6.39232635e-01
5.52325785e-01 -6.56913638e-01 -1.14904606e+00 6.43576443e-01
-1.11636388e+00 -2.79097706e-02 -1.14456108e-02 1.39098898e-01
-8.85889411e-01 -2.95690335e-02 -7.64118671e-01 -1.27560270e+00
-5.39309323e-01 -7.32796192e-01 1.20715797e+00 1.63267314e-01
-2.62854606e-01 -9.61921275e-01 5.71130037e-01 -4.72659469e-01
3.90283197e-01 8.03995669e-01 6.31411195e-01 -2.36601695e-01
1.64535254e-01 -4.46458459e-01 4.86519575e-01 5.17325699e-01
1.41190797e-01 6.49587214e-01 -1.07515264e+00 -2.32258767e-01
2.41467372e-01 -3.28846931e-01 5.03094614e-01 2.18401715e-01
1.06454706e+00 -3.28048378e-01 -2.84136087e-01 5.15241921e-01
1.47671640e+00 6.62224948e-01 2.94846296e-01 -7.65543431e-02
-8.73408094e-02 6.07096672e-01 8.69670093e-01 9.11816180e-01
-2.17055157e-01 1.08610511e+00 1.99499235e-01 3.27558339e-01
4.26950336e-01 9.12227109e-02 7.44388342e-01 8.01813722e-01
-3.69157583e-01 1.36505857e-01 -8.77062380e-01 3.12072307e-01
-1.82704031e+00 -1.24930584e+00 3.01918745e-01 1.71753705e+00
4.73531723e-01 -4.51256663e-01 3.09000731e-01 2.21020356e-01
5.21984339e-01 2.65179276e-01 -4.98601526e-01 -1.09474218e+00
3.87645178e-02 6.40152097e-01 -2.15724632e-01 2.22738281e-01
-1.31906044e+00 8.92439842e-01 7.01444483e+00 5.63097060e-01
-1.88784075e+00 2.15919882e-01 5.09999573e-01 8.15991834e-02
5.77449024e-01 -4.73813742e-01 -5.63683152e-01 3.64177376e-01
1.41636622e+00 -7.66707778e-01 3.68477732e-01 1.02134502e+00
2.88191438e-01 -5.19237630e-02 -1.04526043e+00 1.57041585e+00
1.64668307e-01 -1.02114403e+00 -1.28032640e-01 -1.22694999e-01
5.22395015e-01 -4.06310827e-01 6.15503430e-01 1.62483156e-01
-2.69574672e-01 -1.20973969e+00 7.02069342e-01 1.03613520e+00
9.88921165e-01 -8.44111919e-01 3.55612725e-01 4.05344367e-03
-1.65065849e+00 -3.64732862e-01 -2.15439811e-01 -1.38687059e-01
-8.41509849e-02 9.07170251e-02 -5.10213673e-01 4.82648045e-01
6.19131804e-01 1.04475534e+00 -4.56667572e-01 4.71987545e-01
3.21717292e-01 4.54038531e-01 -2.21680850e-01 3.71996202e-02
3.28191429e-01 -4.11980987e-01 4.58958924e-01 1.32257664e+00
4.17552620e-01 2.01403216e-01 -1.05788186e-01 4.59638655e-01
3.19748253e-01 2.96859771e-01 -1.12360013e+00 -4.33646709e-01
-9.58465263e-02 1.50532699e+00 -3.86752605e-01 -2.69320101e-01
-1.70126408e-01 1.17891788e+00 1.25558689e-01 3.57902683e-02
-8.82728040e-01 8.24147463e-02 1.14879286e+00 -3.31914365e-01
1.72113165e-01 -2.53558546e-01 5.14605105e-01 -1.11410606e+00
-1.25923738e-01 -7.37122595e-01 5.26249468e-01 -6.55383825e-01
-1.12546885e+00 1.05944324e+00 2.06714794e-01 -1.38919842e+00
-8.10398519e-01 -6.97999716e-01 -4.93207127e-01 5.52825272e-01
-1.15453339e+00 -9.07280684e-01 -3.58347476e-01 1.09825802e+00
3.16130191e-01 -2.04984441e-01 1.34417844e+00 3.02013546e-01
-6.13480747e-01 4.24959421e-01 7.23793879e-02 -5.00997566e-02
3.03239793e-01 -7.27614164e-01 -1.11047208e-01 3.04000139e-01
2.89595097e-01 5.18766642e-01 7.97334492e-01 -1.85622483e-01
-1.59326386e+00 -8.59156728e-01 5.57906985e-01 -1.54121548e-01
8.23234797e-01 -1.69221133e-01 -7.56506145e-01 4.91346031e-01
7.08866417e-02 6.99316144e-01 4.86418694e-01 -2.63299763e-01
-4.52797800e-01 -3.60361308e-01 -1.13477826e+00 1.49807066e-01
9.98798549e-01 -7.49752581e-01 -2.87264854e-01 3.33269209e-01
-2.62005180e-01 -1.76597565e-01 -9.74515676e-01 5.30433714e-01
9.82840002e-01 -8.50415468e-01 7.37274766e-01 -9.32536244e-01
-3.23434733e-02 -4.47164252e-02 -9.85605866e-02 -1.12406564e+00
-1.56983763e-01 -9.49548781e-01 -4.20640200e-01 9.85154867e-01
-4.73829240e-01 -4.56328720e-01 3.04317325e-01 1.63064986e-01
5.05078137e-01 -9.58977878e-01 -1.39682901e+00 -1.10170114e+00
-2.24070802e-01 -2.00172067e-01 1.41727343e-01 8.36231589e-01
6.86508045e-02 8.89877900e-02 -4.99688715e-01 -1.98544905e-01
3.41034174e-01 3.89411114e-02 5.03103435e-01 -1.34069729e+00
3.25132787e-01 -3.25383335e-01 -1.38634562e+00 7.51310075e-03
8.53191614e-01 -8.68652523e-01 -3.75911951e-01 -4.47022736e-01
3.86930518e-02 -1.33495331e-02 -6.10748708e-01 4.54592377e-01
5.99436998e-01 1.52485967e-01 3.01837802e-01 -1.68450996e-02
-2.19935670e-01 7.44255185e-01 6.69610977e-01 -1.56802669e-01
-9.16408524e-02 -2.17491344e-01 2.04391375e-01 7.46762216e-01
5.36349356e-01 -4.65449542e-01 -4.27332163e-01 3.26605231e-01
-2.60651857e-01 4.26037878e-01 4.36707318e-01 -1.12578428e+00
1.95860327e-03 -1.88772157e-01 2.23600909e-01 -2.43784040e-01
8.09798300e-01 -8.26766849e-01 7.16991186e-01 6.61515117e-01
-2.47890815e-01 4.28401142e-01 3.86545748e-01 5.90756536e-01
-6.82626545e-01 1.90985739e-01 1.11080539e+00 1.25786275e-01
-8.65252912e-01 4.01058376e-01 -4.06217933e-01 -2.91097015e-01
1.39192581e+00 -1.06175831e-02 1.18336163e-01 -5.07616878e-01
-1.16020036e+00 -5.33225358e-01 6.32332936e-02 6.63121581e-01
6.18015230e-01 -1.61588168e+00 -3.08585912e-01 3.31283629e-01
8.50813650e-03 -1.23325229e+00 1.82898894e-01 1.03918040e+00
-2.02430114e-01 7.46700883e-01 -7.71052957e-01 -6.97179019e-01
-1.84328306e+00 5.18671870e-01 7.23395407e-01 -2.88682729e-01
-5.33961773e-01 5.75183094e-01 -1.02960140e-01 1.30681962e-01
2.21533567e-01 -2.22929358e-01 -3.43809128e-01 5.92002034e-01
5.51026642e-01 2.31278434e-01 1.36218250e-01 -1.30098307e+00
-7.22751439e-01 9.28279221e-01 2.93777257e-01 -2.60500997e-01
1.40601289e+00 1.85372889e-01 -1.93339646e-01 8.48160446e-01
1.71075106e+00 -5.25445580e-01 -8.77882183e-01 -2.03914493e-02
1.94533721e-01 -3.08821648e-01 -1.15936063e-01 -4.28822130e-01
-1.08020175e+00 8.29176426e-01 1.26318383e+00 7.18559604e-03
1.46509564e+00 -3.69807631e-01 2.32131764e-01 3.63263637e-01
6.14999473e-01 -1.01912224e+00 2.73735851e-01 4.93995428e-01
1.23388183e+00 -6.78633094e-01 -2.43198618e-01 -3.10989141e-01
-6.90864027e-01 1.49835920e+00 2.08067030e-01 -4.18937057e-01
1.14729869e+00 2.29597300e-01 1.47026941e-01 -2.19328478e-01
-1.25369847e+00 -1.42736495e-01 5.75639248e-01 2.32217729e-01
4.23965156e-01 -1.33239344e-01 -5.98500907e-01 3.47260714e-01
9.57760066e-02 3.27924877e-01 1.18726425e-01 1.03935063e+00
2.02113643e-01 -9.82645869e-01 -2.70229936e-01 9.65564232e-03
-5.07009625e-01 5.84581792e-01 -3.53810608e-01 8.79551589e-01
2.88778115e-02 7.07219899e-01 7.48969913e-02 -4.06874985e-01
2.44033471e-01 5.82800388e-01 7.46396542e-01 -2.42091134e-01
-5.47572196e-01 -4.42826375e-02 -2.67127514e-01 -1.04945886e+00
-8.40735972e-01 -8.71477544e-01 -9.67763722e-01 2.25316994e-02
1.18861057e-01 -1.67279467e-02 7.87467897e-01 7.30089903e-01
4.50839281e-01 1.65222675e-01 1.05890632e+00 -9.76167202e-01
-7.20159471e-01 -7.67487347e-01 -7.66973913e-01 7.92399049e-01
4.14186597e-01 -1.26298344e+00 -4.13451523e-01 2.88622469e-01] | [13.647475242614746, 1.8394198417663574] |
ce6ec311-ac68-4fbe-831b-6e7717bbb298 | on-exploring-undetermined-relationships-for | 1905.01595 | null | https://arxiv.org/abs/1905.01595v1 | https://arxiv.org/pdf/1905.01595v1.pdf | On Exploring Undetermined Relationships for Visual Relationship Detection | In visual relationship detection, human-notated relationships can be regarded as determinate relationships. However, there are still large amount of unlabeled data, such as object pairs with less significant relationships or even with no relationships. We refer to these unlabeled but potentially useful data as undetermined relationships. Although a vast body of literature exists, few methods exploit these undetermined relationships for visual relationship detection. In this paper, we explore the beneficial effect of undetermined relationships on visual relationship detection. We propose a novel multi-modal feature based undetermined relationship learning network (MF-URLN) and achieve great improvements in relationship detection. In detail, our MF-URLN automatically generates undetermined relationships by comparing object pairs with human-notated data according to a designed criterion. Then, the MF-URLN extracts and fuses features of object pairs from three complementary modals: visual, spatial, and linguistic modals. Further, the MF-URLN proposes two correlated subnetworks: one subnetwork decides the determinate confidence, and the other predicts the relationships. We evaluate the MF-URLN on two datasets: the Visual Relationship Detection (VRD) and the Visual Genome (VG) datasets. The experimental results compared with state-of-the-art methods verify the significant improvements made by the undetermined relationships, e.g., the top-50 relation detection recall improves from 19.5% to 23.9% on the VRD dataset. | ['DaCheng Tao', 'Yibing Zhan', 'Jun Yu', 'Ting Yu'] | 2019-05-05 | on-exploring-undetermined-relationships-for-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Zhan_On_Exploring_Undetermined_Relationships_for_Visual_Relationship_Detection_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Zhan_On_Exploring_Undetermined_Relationships_for_Visual_Relationship_Detection_CVPR_2019_paper.pdf | cvpr-2019-6 | ['visual-relationship-detection'] | ['computer-vision'] | [ 1.24479465e-01 5.35875522e-02 -5.69873512e-01 -4.74049866e-01
-8.48378316e-02 -4.70021904e-01 5.96908748e-01 7.27512240e-02
-1.46490827e-01 7.16764569e-01 -2.24226899e-02 -1.91800758e-01
-2.64892131e-01 -9.48413253e-01 -6.41004324e-01 -5.62121451e-01
1.51744746e-02 4.76854295e-01 6.12155616e-01 -2.49461353e-01
-2.30134595e-02 4.99463797e-01 -1.66070902e+00 5.40561795e-01
8.35399568e-01 8.47123981e-01 8.80949497e-02 3.31471056e-01
5.61434142e-02 6.68954730e-01 -6.18471384e-01 -4.75599110e-01
-1.13215581e-01 -4.95593160e-01 -4.15896297e-01 -1.46452785e-01
2.90552139e-01 -9.95324552e-02 -3.88065070e-01 1.02455449e+00
3.09147775e-01 -1.29184380e-01 8.62257957e-01 -1.73161805e+00
-1.18328738e+00 5.84438443e-01 -1.15369344e+00 3.14745158e-01
5.21576166e-01 -5.00127934e-02 1.18169844e+00 -1.02409172e+00
7.55065739e-01 1.65830255e+00 4.21323836e-01 1.48302108e-01
-1.01024497e+00 -1.17483103e+00 2.39383399e-01 4.02210236e-01
-1.78333664e+00 -1.68724269e-01 5.36167741e-01 -5.45108557e-01
1.04642332e+00 3.58938217e-01 6.00414097e-01 6.61770523e-01
8.39259177e-02 6.90535963e-01 6.88078046e-01 -4.44651306e-01
-2.87415177e-01 2.26241097e-01 2.28393331e-01 7.96611488e-01
6.87875211e-01 3.45952868e-01 -5.61028719e-01 -4.05175872e-02
9.16526258e-01 -1.86392039e-01 -3.58625263e-01 -5.63672900e-01
-1.32203066e+00 5.65750360e-01 7.43282080e-01 3.44831169e-01
1.71221018e-01 -3.75298858e-01 2.73492008e-01 2.08918169e-01
2.88980514e-01 1.77960217e-01 -3.15507382e-01 4.21135634e-01
-1.80271223e-01 -2.14013726e-01 4.86441821e-01 1.28037035e+00
7.52169490e-01 -4.83483404e-01 -3.17012876e-01 1.07904482e+00
6.21296227e-01 3.00422758e-01 2.25074127e-01 -4.10172313e-01
5.34422100e-01 1.02939630e+00 -9.84237567e-02 -1.70587015e+00
-4.51259077e-01 -2.83150256e-01 -1.00875390e+00 -2.20308006e-01
1.63423270e-01 1.68512702e-01 -6.72324061e-01 1.74010372e+00
4.75153267e-01 2.62810308e-02 1.37942031e-01 1.19203436e+00
1.77727187e+00 5.93377173e-01 -7.37342462e-02 -3.50583285e-01
1.33421695e+00 -8.62834692e-01 -8.73059928e-01 -1.35532856e-01
3.00250322e-01 -7.13873506e-01 7.71883607e-01 1.38646647e-01
-4.51680928e-01 -7.80995309e-01 -1.06685698e+00 1.47620782e-01
-3.99092883e-01 4.10017073e-01 9.12526190e-01 3.56907427e-01
-7.90563881e-01 4.45055626e-02 -3.56671482e-01 -5.57608604e-01
4.31458443e-01 4.43318337e-01 -6.33573472e-01 2.05206480e-02
-1.35359848e+00 8.20927322e-01 7.69482613e-01 3.09079528e-01
-5.66072166e-01 -8.86930898e-02 -8.85523975e-01 -1.60367385e-01
6.83246195e-01 -6.08421326e-01 5.64707935e-01 -5.20227015e-01
-5.03484190e-01 1.10152411e+00 -1.95835203e-01 2.52944399e-02
3.94543856e-01 5.60052432e-02 -9.13660586e-01 1.44460186e-01
2.21947864e-01 6.60069287e-01 4.58264202e-01 -1.62789309e+00
-8.32081795e-01 -2.02236488e-01 9.55848172e-02 3.16071332e-01
-4.67180341e-01 8.23963135e-02 -9.72788930e-01 -6.18069112e-01
4.12223369e-01 -7.52404690e-01 2.71083504e-01 3.97564292e-01
-6.46819472e-01 -6.13032997e-01 9.43289995e-01 -3.95961285e-01
1.16137886e+00 -2.11045194e+00 6.57148734e-02 2.44318068e-01
5.94790518e-01 4.13007170e-01 -2.44459048e-01 -2.30900850e-02
-3.37993503e-01 1.18619464e-01 3.95689718e-02 1.98447198e-01
-3.03409278e-01 3.36777538e-01 1.55790329e-01 4.14604932e-01
3.36713701e-01 9.95161712e-01 -1.15747249e+00 -1.01422358e+00
2.26359159e-01 4.21237946e-01 2.64043789e-02 4.25078690e-01
3.75659950e-02 3.51958454e-01 -4.20372635e-01 9.66756225e-01
6.65833950e-01 -3.86309862e-01 4.53100950e-01 -8.24240565e-01
1.45752281e-01 -2.89518535e-01 -8.71548533e-01 1.07380748e+00
-3.20753902e-02 7.07999706e-01 -4.55759197e-01 -8.12419057e-01
1.40438747e+00 3.24149430e-02 1.27500758e-01 -6.55695617e-01
2.35968038e-01 1.05402686e-01 1.91224620e-01 -7.12019980e-01
4.11830217e-01 4.36671786e-02 1.22871563e-01 -2.37654634e-02
1.64477248e-03 2.57745355e-01 2.85767227e-01 3.78818274e-01
8.14706326e-01 3.03456545e-01 7.23926544e-01 8.51508081e-02
6.85566962e-01 -1.64963782e-01 8.96319151e-01 4.82294798e-01
-3.67336303e-01 5.97025812e-01 7.77858794e-01 -1.15295745e-01
-3.64802331e-01 -1.17464590e+00 -2.20533490e-01 7.21106768e-01
8.84480238e-01 -3.73570025e-01 1.90209180e-01 -1.11337698e+00
2.64756262e-01 3.94724131e-01 -6.79484010e-01 -4.89530742e-01
-4.22731489e-01 -7.75983393e-01 3.70900273e-01 7.00052202e-01
5.98324597e-01 -1.08906686e+00 -1.89168788e-02 -3.21149647e-01
-1.44152388e-01 -1.30081558e+00 -3.05469841e-01 -1.29993856e-01
-4.82704699e-01 -1.30403745e+00 -4.14528221e-01 -1.12850988e+00
8.75009418e-01 5.55917978e-01 1.24564052e+00 2.31988043e-01
-1.09048352e-01 -1.02299161e-01 -4.26556885e-01 -1.20773926e-01
-2.66652763e-01 -2.67440319e-01 -9.17978808e-02 8.65981951e-02
5.69664001e-01 -3.69422734e-01 -3.23255360e-01 8.26363504e-01
-4.37109202e-01 1.70156181e-01 7.97400415e-01 8.58798444e-01
7.95345724e-01 2.66560733e-01 6.48133695e-01 -8.35840762e-01
3.11799526e-01 -6.05953991e-01 -4.43153203e-01 5.97289920e-01
-6.07852876e-01 -1.50050715e-01 3.84116858e-01 -7.95659006e-01
-9.67286468e-01 1.24223260e-02 3.12780619e-01 -5.78143656e-01
-8.20823908e-02 4.03039366e-01 -5.50251245e-01 -5.73832169e-02
5.73706567e-01 -1.38944030e-01 -2.91967213e-01 -1.30712956e-01
4.42917019e-01 6.72746837e-01 7.46804237e-01 -3.06624919e-01
8.68341565e-01 3.10810179e-01 5.51350489e-02 -3.94977927e-01
-9.08601820e-01 -5.43533742e-01 -5.28308451e-01 -5.52029729e-01
7.19177544e-01 -7.59665489e-01 -8.11902344e-01 6.56487867e-02
-1.21774256e+00 3.37111503e-01 6.17696457e-02 5.76354682e-01
-1.51698545e-01 3.72379482e-01 -5.56443214e-01 -8.69043469e-01
-1.99056506e-01 -8.45453978e-01 9.13442135e-01 5.82582712e-01
-3.04824173e-01 -7.32271433e-01 -2.62717098e-01 4.20909673e-01
-2.10483864e-01 4.70486045e-01 1.03289366e+00 -6.87606215e-01
-4.36848372e-01 -9.72793251e-02 -1.00570154e+00 -1.08092867e-01
5.33416152e-01 1.66571021e-01 -7.81204700e-01 2.14651991e-02
-7.06753671e-01 -3.44292730e-01 9.54822361e-01 8.74759331e-02
7.13139832e-01 8.29653516e-02 -8.30755770e-01 3.08926731e-01
1.03315818e+00 5.84464908e-01 3.98072749e-01 2.86790412e-02
1.10710406e+00 7.66069174e-01 1.35313880e+00 2.30318949e-01
5.72155893e-01 6.84924781e-01 6.10580921e-01 -2.62321293e-01
-2.85305232e-01 -3.01466644e-01 -1.01099230e-01 6.90075636e-01
-2.57265598e-01 -5.52854717e-01 -7.93366194e-01 6.97583973e-01
-2.23196578e+00 -7.06365585e-01 -5.88560045e-01 1.98001063e+00
1.02480924e+00 2.34870911e-01 9.71072987e-02 7.10399076e-02
1.39345670e+00 -1.74372923e-02 -7.29137123e-01 1.22458927e-01
-5.12181282e-01 -4.54273343e-01 -3.98966223e-02 1.76244929e-01
-1.23125529e+00 9.23852623e-01 5.05877638e+00 9.14455175e-01
-6.67579770e-01 -1.25154048e-01 5.90691328e-01 2.82768875e-01
-2.71076977e-01 -7.31213391e-02 -9.28979576e-01 1.71144679e-01
1.79185107e-01 -2.33539995e-02 2.43760627e-02 7.23584652e-01
-9.16170478e-02 -1.23544253e-01 -1.18719149e+00 1.26607192e+00
2.67012686e-01 -1.00129211e+00 2.02426270e-01 -1.15546353e-01
4.76441443e-01 -5.67482650e-01 3.51808593e-03 3.27335626e-01
1.98193565e-01 -9.71928000e-01 6.73011780e-01 5.94472945e-01
9.64617550e-01 -9.16625082e-01 9.21303868e-01 1.48558274e-01
-1.78270137e+00 2.27745906e-01 -4.25404936e-01 1.23136617e-01
-3.06706056e-02 7.12655723e-01 -7.77645707e-01 8.86985838e-01
7.52261102e-01 1.14405489e+00 -7.96288252e-01 9.97426331e-01
-4.60096061e-01 2.65006155e-01 -2.14308366e-01 -2.88480729e-01
-3.42679530e-01 -1.35623693e-01 4.70751703e-01 9.93204772e-01
-4.45210934e-02 1.92245662e-01 6.18481599e-02 9.20479536e-01
-3.40935588e-02 1.49121121e-01 -5.97340882e-01 -4.62926403e-02
7.30650306e-01 1.45893025e+00 -1.00278091e+00 -1.93860307e-01
-5.13899148e-01 7.04931378e-01 4.90170270e-01 3.17574590e-01
-1.02320719e+00 -4.16170061e-01 1.75938904e-01 -2.26747289e-01
1.97741538e-01 1.94951236e-01 -1.06291227e-01 -1.06819689e+00
7.46059939e-02 -6.34536564e-01 8.40115488e-01 -1.07827997e+00
-1.65593755e+00 7.72461116e-01 1.81993008e-01 -1.44458818e+00
4.13325764e-02 -4.56652492e-01 -1.32778853e-01 7.12399900e-01
-1.40519583e+00 -1.33201265e+00 -6.73409462e-01 3.96650374e-01
9.02763009e-02 -3.35267812e-01 5.92764556e-01 4.20758158e-01
-7.52920926e-01 7.61591971e-01 -6.08629704e-01 3.50624561e-01
7.74430871e-01 -8.87906671e-01 1.67899370e-01 7.41224706e-01
1.76469758e-01 7.64273643e-01 4.36263591e-01 -9.92314875e-01
-1.06322825e+00 -1.28895426e+00 9.81391191e-01 -2.17903316e-01
5.84731877e-01 -2.78938383e-01 -8.99364054e-01 7.86617339e-01
-1.29633486e-01 4.50959533e-01 6.00300431e-01 6.63513318e-02
-6.79289222e-01 -1.60502478e-01 -1.14038563e+00 7.32708037e-01
1.53384471e+00 -3.63187373e-01 -7.12596178e-01 3.40515077e-01
9.34746146e-01 -2.49863848e-01 -7.47678578e-01 9.28072631e-01
7.70111501e-01 -1.03948510e+00 1.22209120e+00 -3.47851843e-01
3.92984450e-01 -7.20645249e-01 -1.90737128e-01 -8.49701285e-01
-5.20497382e-01 -1.68274775e-01 -4.21414196e-01 1.71755314e+00
4.31657165e-01 -7.58039057e-01 5.51694751e-01 1.18329845e-01
1.06915429e-01 -8.49883139e-01 -5.99973559e-01 -9.55310225e-01
-5.45063138e-01 -1.68929905e-01 5.63828290e-01 1.26304913e+00
-1.43558696e-01 6.88574791e-01 -3.62529069e-01 2.93930709e-01
4.66773033e-01 5.65597057e-01 6.32155120e-01 -1.27858973e+00
-4.09386158e-01 -2.13960096e-01 -6.80965126e-01 -1.06294334e+00
1.27983049e-01 -8.93073738e-01 2.43173018e-02 -1.60456836e+00
6.99024022e-01 -5.81772566e-01 -4.28427666e-01 7.51419127e-01
-5.33998430e-01 4.39572096e-01 4.98760454e-02 2.47928485e-01
-8.19586992e-01 4.94143933e-01 1.44301617e+00 -3.16311777e-01
-2.24645302e-01 -1.15999229e-01 -6.89041734e-01 8.46700013e-01
5.35764694e-01 -2.88927615e-01 -5.94928384e-01 -5.48398457e-02
3.02674145e-01 1.01326495e-01 4.14915085e-01 -5.37802935e-01
4.66607921e-02 -2.75555849e-01 6.21155500e-01 -1.12039053e+00
2.77412742e-01 -7.26086676e-01 3.07160228e-01 2.21554294e-01
-2.13399261e-01 -2.74965703e-01 -6.65546686e-05 7.41772830e-01
-2.24735245e-01 3.84732187e-02 5.47266304e-01 1.67405933e-01
-1.02751994e+00 5.75728193e-02 1.23173743e-01 -1.56193227e-02
1.19934595e+00 -3.51652920e-01 -7.64570236e-01 -3.12693417e-01
-6.33175790e-01 3.37876350e-01 2.02441528e-01 6.98293507e-01
9.62024689e-01 -1.58242011e+00 -6.70596242e-01 9.27742124e-02
6.49679542e-01 4.73067611e-02 6.06123768e-02 5.88473856e-01
-8.79669413e-02 1.97780326e-01 9.46159475e-03 -8.77302170e-01
-1.65288532e+00 8.34769428e-01 1.34555817e-01 -8.15246329e-02
-2.32998550e-01 1.06860375e+00 5.12118876e-01 -5.17649353e-01
1.28955618e-01 -7.27666691e-02 -8.86844695e-01 2.83490688e-01
2.52760798e-01 4.09073718e-02 -2.35076293e-01 -1.03622758e+00
-7.74825513e-01 9.56749260e-01 2.11902447e-02 4.48252082e-01
9.98397171e-01 -5.94441220e-03 -3.57821167e-01 5.85834563e-01
1.20007360e+00 -1.14937276e-01 -9.22357380e-01 -3.34307969e-01
-1.22566402e-01 -6.67518914e-01 -3.58113557e-01 -7.97153831e-01
-1.31590986e+00 4.37661767e-01 4.94529903e-01 2.60717392e-01
1.28508282e+00 4.99998569e-01 4.00627404e-01 1.43493965e-01
5.49651802e-01 -3.92384499e-01 2.21699342e-01 2.17057019e-01
1.03834295e+00 -1.46025634e+00 3.90740842e-01 -1.37923801e+00
-7.17280447e-01 8.68513107e-01 1.22426832e+00 3.19152266e-01
5.04450262e-01 1.94002241e-01 -9.01810303e-02 -2.72260189e-01
-6.31528318e-01 -4.02918071e-01 7.88728118e-01 9.25958037e-01
3.92177284e-01 6.02433942e-02 -5.80992103e-01 7.34977841e-01
1.38009302e-02 -1.72856867e-01 -3.59223522e-02 8.02271307e-01
-1.66294113e-01 -7.09591568e-01 -4.42283511e-01 6.44465864e-01
1.28876060e-01 -5.45315444e-02 -7.59278238e-01 9.45484161e-01
5.35624504e-01 1.29379880e+00 1.30806267e-01 -9.14628744e-01
3.70376945e-01 -7.33726382e-01 4.09348577e-01 -5.31109571e-01
-4.62509513e-01 7.57514536e-02 5.06551266e-01 -3.31936687e-01
-7.21184075e-01 -3.15196991e-01 -1.63898349e+00 -2.51619853e-02
-8.68832588e-01 -8.45861509e-02 -2.02183008e-01 8.16027462e-01
1.43969655e-01 6.79736614e-01 4.68078315e-01 -3.51627052e-01
2.47278884e-01 -7.69639671e-01 -4.09165651e-01 4.90707844e-01
3.02203260e-02 -1.10629892e+00 -3.36610019e-01 -2.39904094e-02] | [10.227730751037598, 1.6997777223587036] |
a184687e-7a33-4221-8cbe-a04a4be7bc1e | tumornet-lung-nodule-characterization-using | 1703.00645 | null | http://arxiv.org/abs/1703.00645v1 | http://arxiv.org/pdf/1703.00645v1.pdf | TumorNet: Lung Nodule Characterization Using Multi-View Convolutional Neural Network with Gaussian Process | Characterization of lung nodules as benign or malignant is one of the most
important tasks in lung cancer diagnosis, staging and treatment planning. While
the variation in the appearance of the nodules remains large, there is a need
for a fast and robust computer aided system. In this work, we propose an
end-to-end trainable multi-view deep Convolutional Neural Network (CNN) for
nodule characterization. First, we use median intensity projection to obtain a
2D patch corresponding to each dimension. The three images are then
concatenated to form a tensor, where the images serve as different channels of
the input image. In order to increase the number of training samples, we
perform data augmentation by scaling, rotating and adding noise to the input
image. The trained network is used to extract features from the input image
followed by a Gaussian Process (GP) regression to obtain the malignancy score.
We also empirically establish the significance of different high level nodule
attributes such as calcification, sphericity and others for malignancy
determination. These attributes are found to be complementary to the deep
multi-view CNN features and a significant improvement over other methods is
obtained. | ['Sarfaraz Hussein', 'Qi Song', 'Ulas Bagci', 'Kunlin Cao', 'Robert Gillies'] | 2017-03-02 | null | null | null | null | ['lung-cancer-diagnosis'] | ['medical'] | [ 1.45814374e-01 6.67962357e-02 -6.62266687e-02 -2.37537339e-01
-7.54281640e-01 -4.83308375e-01 4.96801406e-01 5.04431389e-02
-2.58162498e-01 1.70011505e-01 3.30112278e-01 -3.43197703e-01
-1.52497247e-01 -7.56378293e-01 -3.55335623e-01 -9.88431275e-01
2.26371035e-01 5.93021989e-01 3.71773303e-01 7.08620772e-02
-6.88937455e-02 8.90940011e-01 -1.12704897e+00 2.74808139e-01
3.58604819e-01 1.16961658e+00 1.79284781e-01 8.85076106e-01
1.24016225e-01 5.88876128e-01 1.06626287e-01 -1.49062485e-01
3.48881692e-01 -2.89521515e-01 -9.23828006e-01 5.36270320e-01
2.27397382e-01 -3.98485690e-01 -2.65687823e-01 7.60120869e-01
4.89172846e-01 -3.06665659e-01 8.95670712e-01 -6.56023920e-01
-2.57759988e-01 1.51301056e-01 -5.84298730e-01 1.20757744e-01
-1.73119813e-01 -5.37299626e-02 9.07984614e-01 -1.01710355e+00
4.80520427e-01 7.75243998e-01 5.89933395e-01 3.97906780e-01
-8.29177439e-01 -6.03752807e-02 -5.27128518e-01 -4.21141721e-02
-9.14631546e-01 -2.45865751e-02 5.66736877e-01 -5.56708395e-01
2.80629098e-01 5.19721866e-01 9.04571891e-01 6.84976518e-01
3.76576394e-01 4.62672979e-01 8.97301316e-01 -2.51310796e-01
-1.21771172e-01 2.34779209e-01 -3.78588766e-01 1.08087516e+00
3.06721807e-01 -1.39647469e-01 2.53157765e-01 -3.41819286e-01
1.04560483e+00 4.20466065e-01 -9.61806029e-02 -6.89086616e-01
-1.69754148e+00 7.55290806e-01 8.71431887e-01 3.75027180e-01
-7.13759065e-01 2.68017799e-01 3.26790154e-01 -2.11403638e-01
3.51501167e-01 3.38724643e-01 -1.94074571e-01 2.18008831e-01
-7.39640296e-01 -1.53999105e-01 5.94395578e-01 2.38435388e-01
5.21951318e-01 -2.62971401e-01 -4.85163629e-01 8.83168519e-01
3.25921983e-01 2.66558081e-01 7.19066978e-01 -7.42592990e-01
1.30384043e-01 9.40590024e-01 -1.86926022e-01 -6.78787231e-01
-6.46721721e-01 -5.68342447e-01 -1.12516224e+00 -3.44734192e-02
4.00282502e-01 1.22574121e-02 -1.09696114e+00 9.62833524e-01
4.07067716e-01 -5.30233011e-02 -1.44584114e-02 8.10521781e-01
7.01719284e-01 1.10196106e-01 -9.10401270e-02 2.69197524e-01
1.70113862e+00 -9.96520042e-01 -5.67509085e-02 -7.52383173e-02
8.94841790e-01 -8.43183994e-01 8.01176310e-01 -1.45132365e-02
-1.03323102e+00 -2.33212471e-01 -8.39218438e-01 -1.17920496e-01
-9.92506668e-02 7.19681323e-01 5.57105482e-01 4.61583138e-01
-8.30259740e-01 4.14239526e-01 -1.16455293e+00 -3.92039925e-01
4.65103090e-01 5.04995346e-01 -6.57548964e-01 -1.34394720e-01
-5.82233250e-01 7.21195281e-01 2.80947000e-01 1.37971237e-01
-7.73962319e-01 -6.55259490e-01 -5.72283268e-01 1.14320733e-01
1.44650906e-01 -1.11550808e+00 1.11304212e+00 -7.13070154e-01
-1.19096732e+00 9.38505054e-01 -5.52022569e-02 -1.77489847e-01
6.27443254e-01 3.79335433e-01 2.33336166e-02 2.31119499e-01
1.24797158e-01 5.54182768e-01 7.83935606e-01 -1.08733797e+00
-6.77921414e-01 -2.96762735e-01 -2.12305740e-01 2.70351261e-01
-4.23799187e-01 -2.02377841e-01 -7.26196349e-01 -4.08868819e-01
6.55948341e-01 -1.30873513e+00 -4.73582417e-01 2.72062689e-01
-6.24409378e-01 -1.64424386e-02 9.29367661e-01 -7.40183771e-01
8.63490820e-01 -2.13954639e+00 2.69309610e-01 5.24438262e-01
5.93479156e-01 -2.77981967e-01 1.94651321e-01 -5.13592772e-02
-7.66228735e-02 1.70305848e-01 -7.29632303e-02 -1.21074775e-02
-4.41962898e-01 1.85635358e-01 6.52331293e-01 6.03653312e-01
4.13766980e-01 1.06862307e+00 -5.78105628e-01 -7.15942442e-01
2.31953159e-01 4.33664203e-01 -4.33227479e-01 2.06541345e-01
3.30858515e-03 4.62528288e-01 -5.29014945e-01 5.11246800e-01
3.33961904e-01 -7.55920053e-01 -5.27428277e-02 -4.21225429e-01
2.16575652e-01 -1.74019132e-02 -8.79050374e-01 1.07155502e+00
-5.26001990e-01 4.59610462e-01 -6.00714535e-02 -7.80557334e-01
7.31860816e-01 6.01007640e-01 9.46745336e-01 1.16349027e-01
2.46164843e-01 3.40939373e-01 4.86470282e-01 -7.43653178e-01
2.30417639e-01 -1.89577267e-01 2.56930858e-01 5.11871874e-01
-2.66461760e-01 -3.69881600e-01 2.00109810e-01 4.30901013e-02
1.03713918e+00 -5.32315254e-01 4.31976080e-01 -1.39618903e-01
7.67851472e-01 1.25633469e-02 7.50017092e-02 6.32744282e-02
2.46432535e-02 9.60389018e-01 8.52715731e-01 -5.02542675e-01
-1.26534736e+00 -8.87636542e-01 -1.60379201e-01 5.20235777e-01
-1.85831890e-01 2.10439637e-01 -5.59242606e-01 -8.74582410e-01
-9.62712802e-03 6.33257180e-02 -8.55258942e-01 5.46892593e-03
-6.68589592e-01 -7.75891006e-01 2.87856042e-01 8.28226447e-01
5.19671023e-01 -6.44417703e-01 -6.36654317e-01 4.81240861e-02
-2.33700126e-01 -1.03772259e+00 -4.82999533e-01 3.60995531e-01
-1.15423179e+00 -1.17019582e+00 -9.42338586e-01 -8.65894020e-01
1.06981111e+00 3.58688176e-01 7.21501946e-01 2.15926483e-01
-5.54366469e-01 2.26895720e-01 -5.08843325e-02 -8.63196105e-02
-7.28764772e-01 2.53558189e-01 -3.06063920e-01 7.04706162e-02
-5.68683296e-02 -2.92121619e-01 -8.50771368e-01 3.48312110e-01
-9.33302820e-01 9.62283462e-02 1.28958333e+00 9.04662251e-01
9.29257631e-01 -3.61970477e-02 -8.58167633e-02 -9.02373314e-01
5.25239706e-01 -3.60524535e-01 -1.72187328e-01 1.50158018e-01
-2.87182897e-01 1.08284436e-01 5.31682253e-01 -3.91469657e-01
-6.28332675e-01 4.47297901e-01 -5.71477264e-02 -4.57287192e-01
-2.05753848e-01 6.75424099e-01 2.75105119e-01 -3.31900835e-01
7.07184434e-01 1.57380968e-01 3.42001140e-01 -3.87484655e-02
1.42552927e-01 6.15504265e-01 2.23171130e-01 3.12241185e-02
8.49506140e-01 7.34758794e-01 4.27439243e-01 -9.10495579e-01
-8.48701298e-01 -8.44004989e-01 -8.86870682e-01 -2.36628219e-01
1.02990150e+00 -6.54505610e-01 -2.60883212e-01 1.21460229e-01
-7.72114277e-01 9.27669033e-02 -2.77139187e-01 7.56713986e-01
-5.12528777e-01 3.60997707e-01 -6.56441808e-01 -1.10576972e-01
-3.95302415e-01 -1.43179202e+00 1.06569755e+00 1.24306574e-01
6.14100471e-02 -9.38210189e-01 -1.71697199e-01 6.13979220e-01
5.65768540e-01 1.72799349e-01 1.11130476e+00 -7.49699712e-01
-4.79044110e-01 -7.22976685e-01 -4.28598166e-01 3.59158367e-01
3.54446113e-01 2.74249732e-01 -7.45690227e-01 -1.44628093e-01
1.83116645e-01 -2.05544904e-01 7.22704530e-01 5.79958916e-01
1.33313954e+00 -8.56958106e-02 -4.95182842e-01 7.06258893e-01
1.26287818e+00 -6.00337721e-02 4.01472688e-01 2.62266338e-01
7.98997104e-01 3.37421596e-01 1.00668952e-01 3.03354919e-01
4.80807573e-02 1.92339495e-01 6.57116175e-01 -3.94905150e-01
-1.98890612e-01 1.04504086e-01 -2.21861973e-01 8.22720647e-01
-3.26765597e-01 -1.18489554e-02 -1.03979290e+00 3.70445073e-01
-1.24961615e+00 -6.13507390e-01 -2.98829496e-01 2.00529408e+00
2.18079031e-01 4.81523909e-02 -5.55405878e-02 1.10715047e-01
4.70601052e-01 -6.30267411e-02 -3.95443439e-01 7.03161061e-02
2.80524969e-01 -1.36985302e-01 7.49298871e-01 6.38461336e-02
-1.30334771e+00 3.47477794e-01 5.80006075e+00 5.75907588e-01
-1.64160836e+00 -1.25624552e-01 1.08378327e+00 2.85383910e-01
-1.50231227e-01 -2.71067858e-01 -4.38579738e-01 -2.16883376e-01
4.96512383e-01 1.69312984e-01 1.20769024e-01 8.43281686e-01
1.34456739e-01 -2.04321623e-01 -9.96470749e-01 6.98960483e-01
7.87456054e-03 -1.14774430e+00 -6.57752305e-02 3.87813777e-01
7.53796816e-01 3.40773165e-01 1.09829932e-01 -2.88841426e-01
-1.80557042e-01 -9.52951670e-01 8.05002823e-02 4.80676621e-01
6.54910862e-01 -4.25363630e-01 1.01703238e+00 2.35763028e-01
-1.18905652e+00 2.30887216e-02 -2.72239089e-01 6.99702322e-01
-2.09432930e-01 4.65296775e-01 -1.44924021e+00 5.15454412e-01
3.32795084e-01 5.25673211e-01 -9.78591859e-01 1.12301958e+00
1.83149189e-01 4.26493287e-01 -4.56801414e-01 -2.29203522e-01
3.40775073e-01 -1.35174274e-01 2.84323424e-01 8.28646123e-01
5.23895741e-01 -1.09916091e-01 6.55081794e-02 6.55944884e-01
-7.50861242e-02 2.47956946e-01 -3.52758050e-01 -2.32716948e-01
-4.87218462e-02 1.81159520e+00 -1.21743274e+00 -3.37040097e-01
-5.39811909e-01 9.50721800e-01 4.29289788e-02 1.03038535e-01
-5.91638505e-01 7.33681163e-03 1.13103446e-03 4.08157617e-01
2.99274772e-01 -7.23374337e-02 -2.63726801e-01 -9.87270296e-01
-7.56335780e-02 -6.02279603e-01 2.00721219e-01 -5.11829615e-01
-1.18565679e+00 6.30662084e-01 -3.28574210e-01 -1.54134035e+00
-3.77871156e-01 -7.15888023e-01 -7.59757221e-01 7.23102212e-01
-1.16888249e+00 -1.24328196e+00 -7.90504634e-01 3.72943550e-01
4.01954561e-01 -1.98736906e-01 7.68581092e-01 2.86202785e-02
-2.62591302e-01 1.73524573e-01 2.02183470e-01 3.76411766e-01
5.58792889e-01 -1.39532232e+00 1.26402387e-02 3.67151231e-01
-1.65191188e-01 9.78660062e-02 1.79416284e-01 -4.94193882e-01
-1.30767369e+00 -1.22881925e+00 5.52706122e-01 -1.62062481e-01
6.79981351e-01 1.16880454e-01 -6.50653899e-01 5.28480947e-01
-1.02817558e-01 6.24778390e-01 6.16908967e-01 -4.18339908e-01
8.28809440e-02 4.24407423e-02 -9.57525015e-01 4.90371883e-01
2.46897385e-01 -2.95177877e-01 5.96786253e-02 4.89165753e-01
2.73412883e-01 -4.65103179e-01 -1.17079008e+00 4.79767799e-01
5.63502669e-01 -9.60378170e-01 9.80913162e-01 -4.28041458e-01
6.63682044e-01 -1.73682213e-01 5.81817105e-02 -1.11419082e+00
-6.19898081e-01 1.55994296e-01 4.48167711e-01 4.59665269e-01
7.85119474e-01 -4.17319328e-01 1.32325077e+00 4.52833712e-01
-1.21023715e-01 -1.32900989e+00 -7.70867467e-01 -2.64135361e-01
8.30858126e-02 2.89480630e-02 6.65356293e-02 6.07871771e-01
-3.78001571e-01 1.54186174e-01 1.84973851e-01 1.64357170e-01
2.97949761e-01 1.15009181e-01 5.00355005e-01 -1.22513270e+00
-2.70707756e-01 -4.59194720e-01 -4.74885583e-01 -6.45732343e-01
-5.22894681e-01 -9.86476123e-01 -8.59999433e-02 -1.82710040e+00
5.30422211e-01 -4.07558411e-01 -1.16507195e-01 1.68124795e-01
-1.56137675e-01 2.99467593e-01 -1.45739511e-01 4.63793308e-01
-1.08975358e-01 2.47054711e-01 1.76337099e+00 -5.98010011e-02
-5.73052187e-03 6.72693074e-01 -5.16095698e-01 7.16893733e-01
9.35853422e-01 -2.98343182e-01 -3.19390781e-02 -1.67553782e-01
-1.29826233e-01 3.53771597e-01 4.96518850e-01 -1.08426845e+00
2.09449992e-01 1.51954610e-02 7.46520579e-01 -7.52910972e-01
5.19870996e-01 -1.04245150e+00 5.07907104e-03 9.48523223e-01
-2.14606702e-01 7.49662612e-03 -1.80472046e-01 5.47721624e-01
-2.64299929e-01 -4.59471524e-01 9.35264349e-01 -4.97994930e-01
3.19495164e-02 5.42474926e-01 -3.82034034e-01 -4.31438744e-01
1.19975471e+00 -2.48400912e-01 4.51804698e-02 -2.76216745e-01
-7.64298499e-01 -2.51709204e-03 3.31028789e-01 -4.09364216e-02
5.94708860e-01 -1.21694493e+00 -7.85973907e-01 2.41735637e-01
5.23190983e-02 3.22198868e-01 2.01328825e-02 1.33403563e+00
-9.32242990e-01 4.75404203e-01 -5.78303002e-02 -8.46807182e-01
-1.46722198e+00 3.90395969e-01 7.71643460e-01 -5.44207335e-01
-4.65369344e-01 7.55097508e-01 1.51740998e-01 -3.15361053e-01
1.95964724e-01 -5.67183852e-01 -2.87173778e-01 -7.16097057e-02
-2.86249369e-02 2.24641040e-01 2.39075229e-01 -4.68703061e-01
-2.71449964e-02 5.74473262e-01 -1.76115602e-01 -1.48462951e-02
1.37872910e+00 2.00001508e-01 -3.80399942e-01 1.90403372e-01
1.43985736e+00 -7.78342411e-02 -7.67824829e-01 -1.67545855e-01
3.24793495e-02 -1.68064892e-01 2.24515513e-01 -3.77855808e-01
-1.27350998e+00 9.20473635e-01 8.67282987e-01 3.22233528e-01
9.33940291e-01 3.14804345e-01 6.59270108e-01 4.95201826e-01
-5.12423337e-01 -4.81272042e-01 4.45230722e-01 3.18636745e-01
7.89476275e-01 -1.39656568e+00 1.79349303e-01 -5.37408710e-01
-5.87088108e-01 1.57398546e+00 5.84015191e-01 -1.58680901e-01
8.15901935e-01 1.80873245e-01 1.67790309e-01 -4.35553253e-01
-5.66517889e-01 -2.22045392e-01 5.11784375e-01 1.01647280e-01
6.69999838e-01 9.66699868e-02 -9.06943604e-02 1.18331544e-01
-1.20009713e-01 -1.32390618e-01 4.75474775e-01 7.10689068e-01
-5.71850002e-01 -8.67550731e-01 -4.64757025e-01 1.10402393e+00
-7.04921484e-01 1.46175981e-01 -4.87659544e-01 7.87815630e-01
-8.38980600e-02 2.55562514e-01 2.22330224e-02 -2.90721238e-01
2.33885124e-01 -5.63781429e-03 3.56127590e-01 -6.07002974e-01
-6.45613313e-01 5.23684978e-01 -6.74949139e-02 -6.68545887e-02
-2.66781420e-01 -6.82415903e-01 -1.03731406e+00 3.34713757e-02
-5.07350028e-01 -1.52161121e-01 7.91701853e-01 7.20948100e-01
-9.52542573e-02 7.14310884e-01 8.57431412e-01 -7.48486757e-01
-6.54672980e-01 -9.91538227e-01 -3.39111835e-01 3.82862002e-01
2.54685551e-01 -3.05262387e-01 -4.61724281e-01 9.94604602e-02] | [15.419922828674316, -2.1384196281433105] |
8e97547e-cf24-4479-9cf6-8cefea84af01 | unlimiformer-long-range-transformers-with | 2305.01625 | null | https://arxiv.org/abs/2305.01625v2 | https://arxiv.org/pdf/2305.01625v2.pdf | Unlimiformer: Long-Range Transformers with Unlimited Length Input | Since the proposal of transformers, these models have been limited to bounded input lengths, because of their need to attend to every token in the input. In this work, we propose Unlimiformer: a general approach that wraps any existing pretrained encoder-decoder transformer, and offloads the cross-attention computation to a single k-nearest-neighbor (kNN) index, while the returned kNN distances are the attention dot-product scores. This kNN index can be kept on either the GPU or CPU memory and queried in sub-linear time; this way, we can index practically unlimited input sequences, while every attention head in every decoder layer retrieves its top-k keys, instead of attending to every key. We evaluate Unlimiformer on several long-document and book-summarization benchmarks, showing that it can process even 500k token-long inputs from the BookSum dataset, without any input truncation at test time. We demonstrate that Unlimiformer improves pretrained models such as BART and Longformer by extending them to unlimited inputs without additional learned weights and without modifying their code. We make our code and models publicly available at https://github.com/abertsch72/unlimiformer . | ['Matthew R. Gormley', 'Graham Neubig', 'Uri Alon', 'Amanda Bertsch'] | 2023-05-02 | null | null | null | null | ['multi-document-summarization', 'document-summarization'] | ['natural-language-processing', 'natural-language-processing'] | [ 1.65504292e-01 5.80424629e-02 -1.95763648e-01 -7.07559437e-02
-1.29065442e+00 -8.80822837e-01 4.27627265e-01 4.20212090e-01
-7.41642714e-01 5.74594259e-01 5.14410257e-01 -6.09832168e-01
-7.67610967e-02 -7.39193082e-01 -1.04918051e+00 -4.94696110e-01
4.80113924e-03 8.37126791e-01 2.89011121e-01 7.39680603e-02
4.85124648e-01 2.18076736e-01 -1.42506897e+00 6.68211222e-01
6.95515633e-01 1.02941871e+00 2.73646653e-01 1.16223145e+00
2.00449415e-02 6.08889222e-01 -4.35330302e-01 -7.89740562e-01
1.23671360e-01 -1.35474414e-01 -1.11702287e+00 -5.97503066e-01
6.38152003e-01 -6.49035096e-01 -5.15277743e-01 6.63244069e-01
7.72059441e-01 8.01195204e-02 4.20258224e-01 -8.58263850e-01
-1.03257239e+00 1.12916398e+00 -2.72954822e-01 4.55904782e-01
2.35841006e-01 6.14831112e-02 1.53829515e+00 -1.21554446e+00
4.14653808e-01 9.64925408e-01 6.74546242e-01 2.35381752e-01
-1.21491611e+00 -5.32455623e-01 9.35739726e-02 4.77343291e-01
-1.32158756e+00 -4.92279828e-01 1.73252851e-01 -2.17135787e-01
1.59648192e+00 4.54575896e-01 5.04384160e-01 8.61610889e-01
7.74578303e-02 1.03256238e+00 4.75265920e-01 -3.59055161e-01
1.27975876e-02 -3.81253511e-01 4.43776608e-01 6.07514203e-01
4.23819833e-02 -2.17638791e-01 -6.24683917e-01 -1.62204295e-01
4.88956094e-01 5.86490333e-02 -4.02992308e-01 8.35074782e-02
-1.37668955e+00 6.62592173e-01 4.81457710e-01 1.98891908e-01
-2.61910796e-01 6.13655388e-01 8.11900318e-01 3.41680020e-01
4.89625156e-01 3.91394496e-01 -6.52655125e-01 -5.60872674e-01
-9.23460603e-01 9.54468399e-02 8.04127812e-01 1.16823316e+00
7.59827316e-01 -3.67377937e-01 -5.41977942e-01 7.44750857e-01
-3.96504492e-01 5.16189277e-01 6.92660570e-01 -7.68037200e-01
7.79588461e-01 3.07119489e-01 -7.93105960e-02 -6.47655487e-01
-1.69645280e-01 -4.08868909e-01 -8.26122046e-01 -4.64094520e-01
2.25629419e-01 -8.20184946e-02 -9.03601408e-01 1.41650343e+00
4.71135788e-02 1.93996206e-02 -1.03092790e-01 8.77491355e-01
5.83658755e-01 1.05112493e+00 -3.09462160e-01 4.47866991e-02
1.48661685e+00 -1.40464330e+00 -3.36613297e-01 -2.53689915e-01
9.20675874e-01 -7.63624132e-01 1.41515064e+00 5.49068689e-01
-1.50420928e+00 -3.81838620e-01 -7.93825209e-01 -8.57401013e-01
-3.58201593e-01 2.49868184e-01 3.71645421e-01 1.37631461e-01
-1.25170898e+00 7.78588057e-01 -8.18365216e-01 -1.35693073e-01
2.06959412e-01 3.62210959e-01 -1.50598496e-01 -3.11334021e-02
-1.08441925e+00 7.55698740e-01 5.90610504e-01 1.82963282e-01
-8.04995537e-01 -7.92215466e-01 -5.61726570e-01 5.11426389e-01
4.74837929e-01 -7.13579237e-01 1.63079572e+00 -5.61442137e-01
-1.27501893e+00 5.87523282e-01 -4.41988021e-01 -6.81944609e-01
2.62491643e-01 -6.07744575e-01 5.55664264e-02 -4.61230204e-02
-1.80652350e-01 5.63160658e-01 6.16145253e-01 -5.90359807e-01
-4.59753484e-01 -3.08430552e-01 1.22771703e-01 3.05641353e-01
-5.33573389e-01 -3.03321965e-02 -1.01855135e+00 -6.96710706e-01
-1.25687718e-01 -9.40992117e-01 -2.16836073e-02 -2.77020395e-01
-5.90170205e-01 -2.84154207e-01 3.63415986e-01 -6.71154916e-01
1.58084762e+00 -2.10073996e+00 3.23503315e-01 -1.53586911e-02
3.21308851e-01 3.63748550e-01 -2.64239013e-01 8.28342557e-01
1.37039244e-01 1.08011797e-01 -3.20227146e-01 -4.68760014e-01
2.76264548e-01 2.59067506e-01 -5.57653844e-01 2.77488351e-01
-1.99915096e-02 1.23643112e+00 -9.21578765e-01 -2.61790037e-01
-1.37605429e-01 4.22874749e-01 -9.07527268e-01 1.40162170e-01
-2.76496500e-01 -3.78190756e-01 -1.97977036e-01 2.63047814e-01
4.62830186e-01 -4.00922716e-01 -2.30621994e-02 -1.93967074e-01
-6.48535788e-02 8.84851933e-01 -7.42147088e-01 1.75031090e+00
-8.78408313e-01 7.44851828e-01 -2.59219110e-01 -7.41589904e-01
5.71541309e-01 9.44055393e-02 8.70637745e-02 -9.01738286e-01
4.42214571e-02 4.10827160e-01 -2.77459532e-01 -2.30383992e-01
8.91877413e-01 3.05272162e-01 -2.05478162e-01 6.97208643e-01
4.97595593e-02 1.47138089e-01 3.23540926e-01 4.98094022e-01
1.18184280e+00 -1.81449860e-01 -1.39232455e-02 -1.03500620e-01
3.71837676e-01 -2.72448957e-01 9.79200453e-02 8.12914371e-01
5.20043015e-01 6.56537414e-01 7.56634712e-01 -3.88106763e-01
-1.29067874e+00 -1.02208149e+00 5.45070805e-02 1.66982114e+00
-2.14415878e-01 -8.60593379e-01 -8.49632978e-01 -5.91994762e-01
1.30879462e-01 6.54428422e-01 -6.50913477e-01 -1.23969749e-01
-7.75614977e-01 -3.66619110e-01 8.78259540e-01 7.96081722e-01
6.22815080e-02 -1.07968199e+00 -8.50713968e-01 3.00380200e-01
-1.45501047e-01 -8.17856431e-01 -1.07367730e+00 5.76936126e-01
-7.44067848e-01 -8.39586616e-01 -8.50309908e-01 -7.04169393e-01
6.08622074e-01 -6.42310083e-02 1.24260843e+00 1.43529251e-01
-1.55276924e-01 -4.04147059e-02 -4.24019754e-01 -9.16465819e-02
-7.19842613e-02 8.48801374e-01 -2.98282862e-01 -2.41600320e-01
2.37496212e-01 -5.69316506e-01 -6.76689863e-01 6.04740009e-02
-8.75481963e-01 2.53121853e-01 7.39544630e-01 1.01660824e+00
6.84537888e-01 -5.20624280e-01 3.83714020e-01 -9.57629442e-01
7.46054769e-01 -3.53055298e-01 -7.26132274e-01 1.55535981e-01
-4.68946815e-01 4.25855249e-01 9.24005508e-01 -4.23664987e-01
-3.42668235e-01 -2.46897995e-01 -4.69381332e-01 -4.45784688e-01
3.98592442e-01 7.06814885e-01 6.77672476e-02 4.65095967e-01
4.44811046e-01 5.33659518e-01 -2.90957034e-01 -6.51347995e-01
6.30474687e-01 7.44660616e-01 7.56340921e-01 -7.11310744e-01
4.75564718e-01 3.61769870e-02 -5.34255385e-01 -4.88119751e-01
-8.59900117e-01 -4.36309338e-01 -3.89239818e-01 2.84354776e-01
6.04071438e-01 -6.38745725e-01 -9.53564584e-01 3.50329578e-01
-1.38613307e+00 -7.21575379e-01 -3.72650385e-01 2.37090349e-01
-4.60706085e-01 1.84549987e-01 -9.60341573e-01 -4.69909936e-01
-8.85120630e-01 -1.17593694e+00 1.22319722e+00 -2.98951805e-01
-2.66924351e-01 -7.16302037e-01 -8.40541273e-02 7.89904669e-02
5.46920478e-01 -4.13004011e-01 1.14721310e+00 -8.82820487e-01
-7.40543544e-01 -2.55203336e-01 -3.92011762e-01 3.80000025e-01
-2.95658916e-01 -3.32502127e-01 -8.51992130e-01 -4.30237234e-01
-3.94204021e-01 -5.67445755e-01 1.22542346e+00 -1.27418861e-02
1.58195972e+00 -7.34149039e-01 -1.35964617e-01 8.07853520e-01
1.30146945e+00 -1.07298180e-01 5.62982440e-01 2.44120538e-01
6.11897826e-01 -4.02651094e-02 1.71978742e-01 5.71774542e-01
4.50690418e-01 6.06988370e-01 4.04963434e-01 1.23609275e-01
-1.10370785e-01 -3.94857615e-01 3.63637596e-01 1.26199555e+00
-1.95211470e-02 -5.23151338e-01 -9.55789685e-01 7.67137051e-01
-1.79137087e+00 -8.24563801e-01 5.20741381e-02 2.27551675e+00
1.07357919e+00 1.50530517e-01 -1.37024105e-01 1.58321142e-01
4.07566607e-01 1.38256475e-01 -7.52532780e-01 -8.46833050e-01
9.48062465e-02 5.44005752e-01 8.88223052e-01 7.08669424e-01
-8.09750617e-01 1.03232110e+00 5.92038059e+00 1.00818658e+00
-1.01950169e+00 2.36861572e-01 6.44441724e-01 -6.48622751e-01
-4.35204744e-01 -1.01244435e-01 -9.51541901e-01 4.99790788e-01
1.20009613e+00 -3.66358697e-01 8.08088362e-01 7.70519316e-01
-2.36170262e-01 -1.42080277e-01 -1.40987802e+00 9.00642395e-01
3.10484823e-02 -1.41662621e+00 1.31089270e-01 9.99066047e-03
4.70628500e-01 3.68634820e-01 1.98329479e-01 4.37799007e-01
3.19792211e-01 -1.04217386e+00 8.52241993e-01 4.37560827e-01
9.51502085e-01 -9.45354819e-01 6.86507940e-01 3.67544591e-01
-1.13370323e+00 -1.71658739e-01 -5.09335756e-01 8.48684832e-03
1.92288697e-01 5.93068779e-01 -8.28755736e-01 4.34184730e-01
7.16071427e-01 4.76114780e-01 -5.05031943e-01 9.02002692e-01
-1.00443423e-01 7.46584833e-01 -4.85377192e-01 -1.59757242e-01
4.94362295e-01 1.42608017e-01 1.57031566e-01 1.50883555e+00
6.67219639e-01 1.02101132e-01 -9.20522064e-02 4.67837691e-01
-5.06489515e-01 1.02425098e-01 -1.71845734e-01 3.50758657e-02
6.97633862e-01 9.21053767e-01 -4.36616868e-01 -5.66792488e-01
-2.88418055e-01 1.33871198e+00 7.08692729e-01 2.48498946e-01
-1.00501800e+00 -8.70175004e-01 5.88288784e-01 2.38577724e-01
7.89935648e-01 -1.44723520e-01 -2.00397313e-01 -9.84528124e-01
2.68715411e-01 -9.32830870e-01 5.24895430e-01 -8.26078892e-01
-9.41557348e-01 7.28823185e-01 -1.41095817e-01 -6.83305383e-01
-1.91060156e-01 -4.84280229e-01 -2.68964410e-01 9.75173056e-01
-1.40312064e+00 -8.96417499e-01 -5.06246090e-02 4.55317199e-01
5.31906664e-01 3.52918446e-01 7.64829814e-01 5.08717060e-01
-4.75378513e-01 1.05089819e+00 3.54678214e-01 1.08523831e-01
6.40322566e-01 -1.28932297e+00 8.95817399e-01 5.19062042e-01
1.99190378e-01 7.18795776e-01 4.11049634e-01 -2.45101333e-01
-1.69251096e+00 -9.88853574e-01 1.24353099e+00 -3.81798238e-01
8.52024734e-01 -6.32658303e-01 -9.70544219e-01 9.80346262e-01
4.37608421e-01 1.03302017e-01 4.34114158e-01 1.97502241e-01
-4.26828086e-01 -1.74079597e-01 -4.34421390e-01 5.48650742e-01
1.17551148e+00 -7.47091591e-01 -4.68031585e-01 3.45310271e-01
1.01986945e+00 -7.93120146e-01 -8.42045844e-01 8.87982696e-02
5.12220502e-01 -7.95505285e-01 8.42799544e-01 -4.49372709e-01
5.91745436e-01 -8.17631781e-02 -9.25396159e-02 -1.12228239e+00
-3.21505666e-01 -6.81237161e-01 -4.65490222e-01 9.75325346e-01
8.01414490e-01 -6.24324024e-01 6.25064492e-01 3.14040601e-01
-5.58220744e-01 -1.23415887e+00 -9.60980296e-01 -5.78727365e-01
1.77378803e-01 -6.46985531e-01 8.10426533e-01 4.30602580e-01
4.82130572e-02 4.79631662e-01 -4.24819887e-01 1.22207940e-01
9.61521789e-02 2.02047199e-01 7.22198486e-01 -7.01463997e-01
-5.85550606e-01 -5.85153222e-01 -5.87497056e-02 -1.59458327e+00
-4.87521738e-02 -1.36505842e+00 5.38245216e-02 -1.67036045e+00
2.82612324e-01 -3.43868554e-01 -2.40832806e-01 8.46085489e-01
-1.53758734e-01 2.88541406e-01 2.90669799e-01 6.65677786e-02
-7.27051973e-01 5.67710161e-01 1.00822484e+00 -1.47304565e-01
8.25458467e-02 -2.90659964e-01 -6.93839371e-01 4.07120407e-01
6.36181593e-01 -5.62143922e-01 -4.54233706e-01 -1.06563687e+00
5.79398751e-01 -1.12250187e-02 3.48973095e-01 -9.88229215e-01
5.62468767e-01 3.90005141e-01 2.15829790e-01 -9.11002815e-01
2.78389663e-01 -4.77002054e-01 -9.78617668e-02 3.42312813e-01
-7.45591402e-01 5.80496013e-01 3.06750774e-01 2.74012625e-01
8.93662311e-03 -3.22054714e-01 4.22317326e-01 7.39425272e-02
-2.13781044e-01 3.88625652e-01 -2.58773744e-01 4.20857519e-01
6.77149534e-01 2.79859789e-02 -4.43754673e-01 -1.38725370e-01
-6.55021489e-01 2.80113190e-01 4.01380658e-01 1.93583488e-01
5.11864364e-01 -1.12815368e+00 -6.21601105e-01 2.10236773e-01
-6.88849837e-02 2.52450466e-01 2.31074959e-01 8.03774416e-01
-6.61283016e-01 8.42043996e-01 1.99937731e-01 -2.88675070e-01
-1.11744964e+00 8.33009601e-01 9.05045271e-02 -5.68627119e-01
-8.02160621e-01 1.08515823e+00 5.00402711e-02 -3.27037156e-01
4.69090998e-01 -8.13580751e-01 3.25842202e-01 1.74141854e-01
6.63118899e-01 3.83717269e-01 4.10860360e-01 -2.80171156e-01
-2.50948548e-01 5.17758250e-01 -3.89847368e-01 -1.09746054e-01
1.25465965e+00 9.37454253e-02 -2.33799905e-01 3.08376253e-01
1.48322773e+00 5.55599369e-02 -1.02181935e+00 -3.67291331e-01
5.54471686e-02 -2.57428259e-01 3.84352636e-03 -7.30278373e-01
-1.06303954e+00 1.00034404e+00 5.59702888e-02 1.37947634e-01
1.29606175e+00 2.36462113e-02 1.31235266e+00 7.62115121e-01
2.40245640e-01 -9.17240262e-01 5.06350622e-02 9.23367023e-01
9.32756305e-01 -4.92282450e-01 -1.34971708e-01 9.81326401e-02
-5.91659307e-01 1.00537086e+00 3.31368506e-01 -6.44102022e-02
3.11130494e-01 5.79883575e-01 -3.56546223e-01 5.64004630e-02
-1.34673965e+00 -9.89331678e-02 1.46150693e-01 3.70132625e-02
5.78803301e-01 -9.99628752e-02 -1.77549005e-01 7.22378373e-01
-6.14736676e-01 4.08566520e-02 3.97432506e-01 8.13345790e-01
-3.99833441e-01 -1.08482969e+00 -1.20704532e-01 6.68484986e-01
-5.79687834e-01 -7.59047031e-01 -7.72410706e-02 3.34182441e-01
-1.46158427e-01 3.95563364e-01 3.31198633e-01 -3.02073896e-01
4.16962236e-01 2.20218897e-01 2.67055124e-01 -5.03466249e-01
-8.78088593e-01 -1.28206909e-01 1.16928488e-01 -6.45510197e-01
4.69570845e-01 -4.21287090e-01 -1.28668118e+00 -5.96354485e-01
-2.71873951e-01 3.88123721e-01 4.31418389e-01 5.19206166e-01
7.40602434e-01 5.31156480e-01 2.56043792e-01 -8.44374716e-01
-7.39850163e-01 -9.65711534e-01 -9.48051512e-02 3.62312458e-02
4.93731439e-01 1.70462485e-02 -1.30759507e-01 -1.51333958e-03] | [10.921186447143555, 7.5249342918396] |
700f78ca-13bc-40a9-a188-02751bd3b3d8 | shared-coupling-bridge-for-weakly-supervised | 2212.07047 | null | https://arxiv.org/abs/2212.07047v1 | https://arxiv.org/pdf/2212.07047v1.pdf | Shared Coupling-bridge for Weakly Supervised Local Feature Learning | Sparse local feature extraction is usually believed to be of important significance in typical vision tasks such as simultaneous localization and mapping, image matching and 3D reconstruction. At present, it still has some deficiencies needing further improvement, mainly including the discrimination power of extracted local descriptors, the localization accuracy of detected keypoints, and the efficiency of local feature learning. This paper focuses on promoting the currently popular sparse local feature learning with camera pose supervision. Therefore, it pertinently proposes a Shared Coupling-bridge scheme with four light-weight yet effective improvements for weakly-supervised local feature (SCFeat) learning. It mainly contains: i) the \emph{Feature-Fusion-ResUNet Backbone} (F2R-Backbone) for local descriptors learning, ii) a shared coupling-bridge normalization to improve the decoupling training of description network and detection network, iii) an improved detection network with peakiness measurement to detect keypoints and iv) the fundamental matrix error as a reward factor to further optimize feature detection training. Extensive experiments prove that our SCFeat improvement is effective. It could often obtain a state-of-the-art performance on classic image matching and visual localization. In terms of 3D reconstruction, it could still achieve competitive results. For sharing and communication, our source codes are available at https://github.com/sunjiayuanro/SCFeat.git. | ['Luping Ji', 'Jiewen Zhu', 'Jiayuan Sun'] | 2022-12-14 | null | null | null | null | ['simultaneous-localization-and-mapping', 'camera-localization', 'visual-localization', 'image-matching'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [-9.55046415e-02 -4.71537709e-01 -5.30405164e-01 -2.66673833e-01
-1.14297879e+00 -2.44722977e-01 5.95236123e-01 -1.26390502e-01
-3.38542610e-01 2.02438563e-01 5.19031584e-02 6.92325681e-02
-3.38451594e-01 -4.09842193e-01 -8.03691566e-01 -8.95467043e-01
-1.42748162e-01 -4.27081482e-03 2.78258145e-01 -1.25618741e-01
3.80012244e-01 9.27641332e-01 -1.60889697e+00 -3.49431410e-02
6.17897868e-01 1.38701475e+00 4.88955736e-01 2.12240592e-01
9.98963714e-02 7.13366270e-01 -2.32077882e-01 -6.51407614e-02
6.03362799e-01 -9.12785232e-02 -5.03302991e-01 -1.31804431e-02
7.22796738e-01 -3.19115341e-01 -5.51181734e-01 1.16540658e+00
7.98178494e-01 1.46240667e-02 3.38366389e-01 -1.29670775e+00
-5.42468905e-01 1.54496342e-01 -6.88174307e-01 1.87375754e-01
4.42482561e-01 9.38146561e-02 1.02033234e+00 -1.42494047e+00
4.89995599e-01 8.78501892e-01 8.15201342e-01 1.77803189e-01
-7.94340968e-01 -8.84474695e-01 1.69428378e-01 5.60153425e-01
-1.97317576e+00 -5.90209663e-01 8.68996918e-01 -2.02030122e-01
8.06553006e-01 1.98528707e-01 5.47895133e-01 8.37901175e-01
5.30844107e-02 7.78838694e-01 7.85049796e-01 -4.15239036e-01
-1.46726176e-01 8.75425246e-03 -1.65534645e-01 9.64676142e-01
8.34788680e-02 3.27605069e-01 -7.48473704e-01 -1.14236370e-01
1.07417333e+00 3.38821799e-01 -4.46315497e-01 -7.23575354e-01
-1.40803063e+00 8.44548106e-01 9.26334083e-01 4.71895009e-01
-1.98052451e-01 1.28294632e-01 1.56045392e-01 3.37761432e-01
3.18787694e-01 2.07435444e-01 -4.19775337e-01 5.35960011e-02
-7.93287396e-01 1.02754094e-01 4.28714305e-01 1.18630278e+00
1.14389503e+00 -4.34837490e-02 -5.05379178e-02 8.64570856e-01
2.65498221e-01 8.39259744e-01 4.46595848e-01 -9.02988136e-01
3.50189358e-01 5.65548360e-01 7.72871217e-03 -1.31458676e+00
-5.16902983e-01 -6.29723132e-01 -9.40109015e-01 3.59046645e-02
3.17186192e-02 1.43930644e-01 -5.83589733e-01 1.51903558e+00
3.18260908e-01 4.09354597e-01 -3.77183467e-01 1.17107058e+00
8.89305651e-01 5.22879124e-01 -4.56983805e-01 6.54642358e-02
1.14898634e+00 -9.22624886e-01 -3.91088337e-01 -1.96037814e-01
5.77539623e-01 -1.00883329e+00 5.73564410e-01 -1.46646705e-02
-8.10688555e-01 -7.62051702e-01 -9.95526254e-01 -2.32961133e-01
-2.47587994e-01 3.92236471e-01 7.63960540e-01 7.84977525e-02
-9.86790299e-01 3.14803690e-01 -6.99930310e-01 -4.69557166e-01
3.42303932e-01 5.41557789e-01 -7.54688144e-01 -3.19341540e-01
-9.09415901e-01 8.14122200e-01 1.31565481e-01 1.38515726e-01
-6.86437070e-01 -6.64920092e-01 -1.09912765e+00 -1.34476587e-01
5.32917380e-01 -7.86319613e-01 8.74824464e-01 -4.37269419e-01
-1.32110167e+00 9.50897336e-01 -3.99465144e-01 -1.62532806e-01
2.27011651e-01 -1.22522533e-01 -2.35694051e-01 2.05280095e-01
4.13356751e-01 6.21843815e-01 9.32415664e-01 -1.04104936e+00
-7.95436800e-01 -4.72851068e-01 -3.24665010e-01 2.57716179e-01
-2.60038435e-01 -1.11774802e-02 -8.11181188e-01 -6.51910722e-01
5.40107191e-01 -9.17404890e-01 -2.21590161e-01 3.45306993e-01
-1.22041307e-01 -4.43442643e-01 6.15882158e-01 -3.54517758e-01
1.01568222e+00 -2.36941910e+00 5.04069217e-02 4.64882612e-01
2.08366796e-01 1.25347942e-01 -3.65454555e-01 3.04289579e-01
-7.65544400e-02 -3.72372240e-01 1.02320679e-01 -3.72061908e-01
-1.85644880e-01 -1.08197935e-01 -1.66235670e-01 1.03018308e+00
1.43738881e-01 1.09370267e+00 -6.78632915e-01 -4.48183954e-01
5.55940986e-01 5.81739187e-01 -3.38213682e-01 1.19452000e-01
5.28578818e-01 3.00921053e-01 -7.18217552e-01 1.04769135e+00
8.06589782e-01 -3.82490188e-01 -4.47309583e-01 -6.34988368e-01
-3.88339996e-01 -8.99918079e-02 -1.50408959e+00 2.11341453e+00
-3.07239860e-01 3.31065327e-01 1.77320704e-01 -1.06168163e+00
1.14244771e+00 -8.39365274e-02 6.35698140e-01 -8.90416980e-01
1.96098521e-01 5.07559240e-01 -4.67577428e-01 -2.21564695e-01
2.63335913e-01 2.28656307e-01 -5.17859794e-02 -1.30657747e-01
3.21061134e-01 -5.72919361e-02 -2.07176641e-01 7.11798966e-02
9.25022066e-01 1.03526682e-01 4.40260172e-01 -3.33836496e-01
8.72244537e-01 -2.36635551e-01 5.89990854e-01 7.33623743e-01
-2.82187641e-01 7.48573720e-01 -9.05699432e-02 -2.78637916e-01
-8.01397085e-01 -7.84425616e-01 -4.47897255e-01 8.98942590e-01
7.24395216e-01 -5.30821562e-01 -2.28536606e-01 -4.98842150e-01
2.61773735e-01 -8.73783678e-02 -3.61972690e-01 -3.22712630e-01
-5.02759218e-01 -2.14098662e-01 2.86928415e-01 5.55794239e-01
5.85989356e-01 -5.93771398e-01 -3.57948422e-01 -2.35506669e-01
-3.62843089e-02 -1.15627360e+00 -7.65168011e-01 3.63263041e-01
-6.87564790e-01 -1.05198717e+00 -8.86428475e-01 -1.05362809e+00
7.72924185e-01 9.69777107e-01 5.48679292e-01 2.09350020e-01
-4.76235062e-01 5.02059877e-01 -3.56512785e-01 1.12554687e-03
4.52235818e-01 1.02666140e-01 1.43461049e-01 3.28696929e-02
4.34688836e-01 -4.64597672e-01 -7.20219493e-01 5.75578868e-01
-4.93574381e-01 -2.29243100e-01 9.40899789e-01 1.02895975e+00
8.45461011e-01 -2.67960012e-01 1.10979900e-01 -2.94997364e-01
1.58799678e-01 -2.74797559e-01 -6.91517293e-01 1.34905308e-01
-6.04950607e-01 -1.33324498e-02 5.44132054e-01 -2.83565760e-01
-6.02971971e-01 2.00836718e-01 -2.32833058e-01 -6.85173512e-01
-1.09767929e-01 2.82170862e-01 -1.17554061e-01 -9.31677461e-01
5.22992671e-01 5.55561543e-01 1.20219953e-01 -4.60037023e-01
2.90201604e-01 6.00441158e-01 5.52089512e-01 -4.94612098e-01
1.20137978e+00 5.52990735e-01 2.03801155e-01 -7.15226352e-01
-8.01185727e-01 -1.02405274e+00 -5.89454770e-01 2.43243314e-02
3.38484317e-01 -1.44917989e+00 -8.65702748e-01 5.09118199e-01
-8.98245394e-01 1.66687652e-01 -2.41754681e-01 6.58504784e-01
-5.29035628e-01 4.91147220e-01 -3.80176544e-01 -3.61276835e-01
-2.67367631e-01 -1.15613711e+00 1.33751857e+00 3.01130891e-01
3.17098379e-01 -8.08667421e-01 -4.29945886e-02 2.85669476e-01
5.61272919e-01 9.05162841e-02 3.08623999e-01 -3.83959800e-01
-8.62619817e-01 -5.00836849e-01 -5.00097871e-01 3.56335819e-01
-4.66170311e-02 -3.50510448e-01 -8.43407631e-01 -6.60437047e-01
-1.37058333e-01 -3.33044976e-01 8.91766489e-01 3.85416716e-01
1.10888255e+00 8.49156454e-03 -4.92632300e-01 1.34828997e+00
1.47677374e+00 -3.72833461e-02 3.56384695e-01 4.69709754e-01
7.47924566e-01 3.07818264e-01 6.77551746e-01 5.35971820e-01
4.32562083e-01 9.52627063e-01 4.35691625e-01 -2.15734228e-01
-3.55226308e-01 -3.17865461e-01 4.40550089e-01 8.57110798e-01
-1.24321811e-01 3.96721333e-01 -6.14104927e-01 5.12511790e-01
-1.97706532e+00 -8.74528289e-01 1.57780096e-01 2.20995569e+00
5.25883138e-01 -2.20491931e-01 -1.24973617e-01 -1.91773072e-01
7.67023683e-01 3.69598776e-01 -5.80260932e-01 3.60007346e-01
-3.70387256e-01 1.22928724e-01 7.78022408e-01 4.90144849e-01
-1.18820238e+00 9.11056340e-01 4.87624454e+00 1.27477431e+00
-1.26420057e+00 2.25471959e-01 2.30514869e-01 9.57828462e-02
9.63334180e-03 8.23069885e-02 -1.16688871e+00 3.46795887e-01
3.70623559e-01 1.11363558e-02 3.76074672e-01 9.47789550e-01
-7.86686763e-02 7.75786415e-02 -8.35631013e-01 1.58641791e+00
4.66954827e-01 -1.38942063e+00 6.99838251e-02 7.68591464e-02
5.77746570e-01 3.85620773e-01 1.05845697e-01 1.00002475e-01
-4.96358544e-01 -7.80522525e-01 7.15606689e-01 5.20806730e-01
9.06758964e-01 -8.45802248e-01 8.52250457e-01 3.42635840e-01
-1.57452369e+00 -2.33012572e-01 -6.55659437e-01 2.14321733e-01
-1.28927618e-01 6.87809467e-01 -2.32442334e-01 8.82014990e-01
7.82805622e-01 1.26146674e+00 -7.54636884e-01 1.30743694e+00
-2.59275377e-01 -5.66820428e-02 -4.65478182e-01 1.51000381e-01
2.14087561e-01 -1.80938289e-01 5.73620260e-01 1.05135369e+00
5.08703530e-01 -1.30269364e-01 4.53181535e-01 6.10505223e-01
2.47320533e-03 1.82542980e-01 -6.90708518e-01 4.75031376e-01
7.10351586e-01 1.46965694e+00 -4.33931679e-01 2.06485251e-03
-6.15931153e-01 9.68928814e-01 3.46018881e-01 1.32897019e-01
-6.61602378e-01 -6.46457732e-01 7.01898277e-01 1.06635183e-01
5.42937577e-01 -3.55951905e-01 1.30749360e-01 -1.47217786e+00
2.80960411e-01 -7.17112422e-01 1.13240480e-01 -5.52146196e-01
-1.26477861e+00 4.75197762e-01 -1.61188349e-01 -1.61155367e+00
-9.29654390e-03 -7.95585155e-01 -4.59969193e-01 7.73460507e-01
-1.93180275e+00 -1.45033050e+00 -7.44317293e-01 1.29661572e+00
2.33035654e-01 -2.76296735e-01 5.40884197e-01 5.37624419e-01
-4.61558998e-01 8.88187706e-01 1.70922503e-01 1.76426828e-01
1.05546558e+00 -8.37591350e-01 5.05374894e-02 9.91621912e-01
2.67271727e-01 6.79212868e-01 2.27216378e-01 -3.54763031e-01
-1.78939772e+00 -9.76616263e-01 7.58635163e-01 -3.60005587e-01
6.50980294e-01 -4.22731161e-01 -5.50542355e-01 4.15487975e-01
-2.58300155e-01 5.35219789e-01 3.88296425e-01 9.33816656e-03
-4.40380603e-01 -5.78621805e-01 -8.45272243e-01 2.37782389e-01
1.33043754e+00 -8.61242592e-01 -2.47848511e-01 4.85759377e-01
5.12452126e-01 -5.57217479e-01 -9.40392852e-01 5.33098340e-01
5.43921113e-01 -1.08021665e+00 1.29767954e+00 1.34106308e-01
-1.42886788e-01 -5.54326415e-01 -4.25200254e-01 -8.08719218e-01
-6.52385056e-01 -5.36623716e-01 -1.81393698e-01 1.11476886e+00
-5.22029586e-02 -7.76918352e-01 6.29922450e-01 4.57736384e-03
-2.28641450e-01 -9.30939555e-01 -1.17396450e+00 -9.13312495e-01
-3.18176031e-01 -3.23417693e-01 4.17261720e-01 8.82172167e-01
-2.48165578e-01 3.13986599e-01 -5.09179592e-01 3.82472038e-01
7.76139319e-01 3.66643965e-01 9.09723759e-01 -1.10206854e+00
-5.26350066e-02 -4.38002199e-01 -8.67490232e-01 -1.61746979e+00
4.30426225e-02 -1.07379472e+00 -1.16173007e-01 -1.29362404e+00
3.32777381e-01 -3.90081644e-01 -4.45145786e-01 5.72153747e-01
7.13320151e-02 4.14693475e-01 2.89514929e-01 6.44937575e-01
-8.38021278e-01 7.00438857e-01 1.19570959e+00 9.10258815e-02
5.86651303e-02 6.11922629e-02 -7.65871048e-01 4.88500446e-01
3.93635482e-01 -4.18491840e-01 -8.37809220e-02 -3.36436212e-01
2.10919790e-02 -1.70330286e-01 7.68670559e-01 -1.04834390e+00
7.82342553e-01 8.50932673e-02 5.83983004e-01 -4.92170155e-01
3.94232601e-01 -9.28785443e-01 -1.03207611e-01 4.99635816e-01
3.67706753e-02 1.28433583e-02 -9.48349386e-02 5.29097080e-01
-6.08777046e-01 -1.00782655e-01 7.35696137e-01 -5.39617911e-02
-1.17932820e+00 6.86945796e-01 3.61335725e-01 -2.91826159e-01
1.01598859e+00 -4.51516896e-01 -3.50133300e-01 -2.34872550e-01
-2.21304983e-01 2.97215134e-01 5.84701478e-01 4.45841700e-01
8.46283555e-01 -1.59989345e+00 -6.83466554e-01 6.87840462e-01
5.20300806e-01 -2.47958433e-02 3.33243906e-01 1.27177775e+00
-3.36664766e-01 6.09321594e-01 -7.28232563e-02 -8.22639763e-01
-1.01197076e+00 4.35110867e-01 2.55228996e-01 1.91745922e-01
-8.16725671e-01 9.87286031e-01 1.88034400e-01 -4.14636254e-01
4.37125504e-01 -2.07660079e-01 -1.27864694e-02 -1.28292535e-02
5.31736970e-01 3.94369662e-01 1.31499141e-01 -8.28055263e-01
-7.13409007e-01 1.39474225e+00 -8.58648214e-03 5.64038754e-01
1.26293659e+00 -3.28171551e-01 -2.18639955e-01 5.84047958e-02
1.66010737e+00 1.43295228e-01 -1.23471749e+00 -5.29051423e-01
-1.68682382e-01 -7.86287606e-01 3.13547075e-01 -3.23175043e-01
-1.14930189e+00 6.03207409e-01 8.72293770e-01 -2.84223974e-01
1.17468190e+00 2.74537593e-01 7.50350237e-01 6.02165222e-01
7.94304371e-01 -6.91165566e-01 -1.81225836e-02 5.36989868e-01
9.62479770e-01 -1.52913666e+00 2.38061905e-01 -2.47434616e-01
-2.86572337e-01 1.08999598e+00 3.76031548e-01 -4.79778945e-01
8.31639767e-01 1.55692041e-01 1.10472040e-03 -2.16456875e-01
-2.87678540e-01 -3.99846345e-01 5.87268412e-01 4.75239277e-01
1.10670887e-01 -2.55313396e-01 1.52194444e-02 3.50249112e-01
1.53812096e-01 -2.70141631e-01 -1.20338015e-01 7.62649298e-01
-4.98355269e-01 -9.45736349e-01 -4.40696627e-01 2.17328057e-01
-5.14496788e-02 -1.49771020e-01 -9.98622645e-03 8.46866667e-01
3.91121894e-01 5.94758570e-01 -1.67483836e-01 -4.71769571e-01
4.83727485e-01 -3.31578076e-01 5.33114612e-01 -4.08140689e-01
-3.45716596e-01 2.35348076e-01 -4.68711466e-01 -1.12487197e+00
-6.79784000e-01 -6.31620228e-01 -9.72545981e-01 -3.05898666e-01
-6.43356144e-01 1.46102667e-01 6.46768689e-01 7.15547621e-01
6.50107682e-01 6.76801577e-02 9.22330856e-01 -1.16896069e+00
-7.19409287e-01 -6.99214697e-01 -8.25278461e-01 2.19530433e-01
5.61631143e-01 -8.47168386e-01 -6.07700467e-01 -3.22937101e-01] | [7.8910298347473145, -1.9964605569839478] |
fd20bee2-cb5a-4735-aa9b-25232bedb55d | mad-a-scalable-dataset-for-language-grounding | 2112.00431 | null | https://arxiv.org/abs/2112.00431v2 | https://arxiv.org/pdf/2112.00431v2.pdf | MAD: A Scalable Dataset for Language Grounding in Videos from Movie Audio Descriptions | The recent and increasing interest in video-language research has driven the development of large-scale datasets that enable data-intensive machine learning techniques. In comparison, limited effort has been made at assessing the fitness of these datasets for the video-language grounding task. Recent works have begun to discover significant limitations in these datasets, suggesting that state-of-the-art techniques commonly overfit to hidden dataset biases. In this work, we present MAD (Movie Audio Descriptions), a novel benchmark that departs from the paradigm of augmenting existing video datasets with text annotations and focuses on crawling and aligning available audio descriptions of mainstream movies. MAD contains over 384,000 natural language sentences grounded in over 1,200 hours of videos and exhibits a significant reduction in the currently diagnosed biases for video-language grounding datasets. MAD's collection strategy enables a novel and more challenging version of video-language grounding, where short temporal moments (typically seconds long) must be accurately grounded in diverse long-form videos that can last up to three hours. We have released MAD's data and baselines code at https://github.com/Soldelli/MAD. | ['Bernard Ghanem', 'Silvio Giancola', 'Chen Zhao', 'Fabian Caba Heilbron', 'Juan León Alcázar', 'Alejandro Pardo', 'Mattia Soldan'] | 2021-12-01 | mad-a-scalable-dataset-for-language-grounding-1 | http://openaccess.thecvf.com//content/CVPR2022/html/Soldan_MAD_A_Scalable_Dataset_for_Language_Grounding_in_Videos_From_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Soldan_MAD_A_Scalable_Dataset_for_Language_Grounding_in_Videos_From_CVPR_2022_paper.pdf | cvpr-2022-1 | ['moment-retrieval', 'natural-language-moment-retrieval'] | ['computer-vision', 'computer-vision'] | [ 1.71205133e-01 -1.94707304e-01 -4.60466921e-01 -4.74192739e-01
-1.28399026e+00 -6.91614866e-01 7.77917087e-01 1.45903811e-01
-3.59829962e-01 4.79614943e-01 6.52835011e-01 -1.15359225e-03
2.14109957e-01 -2.73539603e-01 -8.77278090e-01 -3.17620128e-01
-2.87257582e-01 2.30096877e-01 2.43434981e-01 -8.31942111e-02
1.16601899e-01 -1.01321436e-01 -1.63348663e+00 8.01808536e-01
1.71981025e-02 1.02918196e+00 1.34386122e-01 9.12141681e-01
1.20911293e-01 1.16002166e+00 -4.65562820e-01 -5.62258661e-01
1.91606939e-01 -5.51022470e-01 -9.29309726e-01 1.68386370e-01
1.15992141e+00 -5.61260164e-01 -7.75696695e-01 7.22476602e-01
5.65519691e-01 1.41420871e-01 1.34725988e-01 -1.65919721e+00
-6.47209942e-01 8.85598779e-01 -2.14167550e-01 5.36431491e-01
9.04925287e-01 6.13712668e-02 1.46055830e+00 -8.28815579e-01
1.01595831e+00 1.08822525e+00 8.23849082e-01 5.92675149e-01
-1.00019228e+00 -6.71360195e-01 8.02894905e-02 2.95040101e-01
-1.60485888e+00 -8.08666766e-01 6.21341705e-01 -7.67630041e-01
9.69281256e-01 2.44057596e-01 5.65779507e-01 1.82835186e+00
-2.48073444e-01 1.04315889e+00 6.25311255e-01 -3.71351928e-01
-9.82558914e-03 -5.17241396e-02 1.33847808e-02 6.07778788e-01
2.41095990e-01 -2.77065456e-01 -1.07114530e+00 -2.59760082e-01
4.06615645e-01 -1.87297121e-01 -1.54804423e-01 -1.75869554e-01
-1.70652735e+00 7.59761691e-01 -1.63274437e-01 3.82635921e-01
-6.38588518e-02 2.53887117e-01 7.75118470e-01 2.97300965e-01
6.22161090e-01 4.95305717e-01 -1.58411115e-01 -7.96137869e-01
-1.31774414e+00 6.85379863e-01 6.77139819e-01 1.49082875e+00
5.86643577e-01 -8.72842222e-02 -2.71154583e-01 5.79894662e-01
-3.95186022e-02 3.55723679e-01 5.39135337e-01 -1.23084700e+00
8.32587779e-01 2.96112955e-01 5.83401360e-02 -1.33117151e+00
-1.03338532e-01 1.26848087e-01 -4.30303723e-01 -7.23478019e-01
3.35649669e-01 6.17195852e-02 -3.94188106e-01 1.67901909e+00
7.53811374e-02 4.98318940e-01 -2.13501513e-01 8.15977097e-01
1.07054877e+00 5.11272252e-01 -2.07871750e-01 2.21725702e-02
1.08485436e+00 -8.79582465e-01 -7.26261914e-01 -1.97806552e-01
8.07659745e-01 -7.86921561e-01 1.50249505e+00 4.37762052e-01
-1.07680273e+00 -5.74127734e-01 -1.01359248e+00 -2.32818380e-01
-2.30011016e-01 -1.77059010e-01 4.95509028e-01 2.88545609e-01
-1.03328884e+00 4.67916489e-01 -6.40550077e-01 -6.49233162e-01
4.69624698e-01 -8.15036073e-02 -6.33386791e-01 -2.31557816e-01
-1.15075362e+00 3.26522648e-01 3.16638499e-01 -1.55666009e-01
-1.22517419e+00 -7.00637579e-01 -9.02368188e-01 -6.09709859e-01
6.53268576e-01 -2.93885916e-01 1.45388746e+00 -9.81106222e-01
-9.38160658e-01 1.30117953e+00 -9.27331224e-02 -6.44903421e-01
4.76564020e-01 -6.22891128e-01 -6.06068194e-01 5.18144906e-01
4.04861838e-01 9.04541433e-01 8.30001950e-01 -8.43996167e-01
-6.60696268e-01 2.92851850e-02 4.36999202e-01 -1.22859828e-01
-6.65606201e-01 3.37086946e-01 -6.63807690e-01 -1.00941741e+00
-2.98631847e-01 -1.16352105e+00 2.64684737e-01 -2.34249115e-01
-4.92758602e-01 -4.65629250e-03 6.42530322e-01 -6.31618381e-01
1.60491574e+00 -2.29644537e+00 2.78551251e-01 -4.00465339e-01
3.93720508e-01 -2.30909556e-01 -2.91319013e-01 8.02978277e-01
8.99773762e-02 4.13898975e-01 1.63419783e-01 -5.03756464e-01
1.75238833e-01 2.43513420e-01 -5.67045569e-01 4.72577989e-01
3.78692038e-02 7.13065982e-01 -1.08875990e+00 -7.45690882e-01
-1.89588413e-01 1.57530025e-01 -9.57345784e-01 3.55234206e-01
-2.91708618e-01 3.75785768e-01 -1.96751624e-01 8.48334074e-01
-9.86603741e-03 -3.20857972e-01 7.41115585e-02 -1.49475828e-01
-1.21593580e-01 4.46306795e-01 -8.78817499e-01 2.27701545e+00
-4.27968241e-02 1.25723064e+00 -2.49728277e-01 -8.11536908e-01
5.32440603e-01 5.17390192e-01 9.15777504e-01 -4.66075271e-01
6.25485852e-02 2.70012040e-02 -3.22319835e-01 -8.77534330e-01
8.41778815e-01 7.59830549e-02 -4.12190318e-01 3.60667199e-01
2.94905841e-01 -1.39099747e-01 5.18742025e-01 5.58472633e-01
1.17196476e+00 2.13810503e-01 1.45558611e-01 9.84687451e-03
2.20704600e-01 3.02078664e-01 3.85163039e-01 7.55077899e-01
-2.26040840e-01 9.05949891e-01 4.54046071e-01 -5.16177595e-01
-1.26001215e+00 -7.93076217e-01 -4.30765636e-02 1.42365801e+00
-7.05598369e-02 -1.14782035e+00 -7.60563791e-01 -5.29361844e-01
1.53472740e-02 3.57397199e-01 -5.74944198e-01 -1.10624647e-02
-4.84681278e-01 -2.39104494e-01 1.11120951e+00 4.75904137e-01
2.57668525e-01 -8.69453728e-01 -3.84137124e-01 1.10408716e-01
-7.60979235e-01 -1.78254342e+00 -4.91488129e-01 -4.08602208e-01
-6.80371225e-01 -1.15034235e+00 -5.37028432e-01 -5.26632905e-01
2.04396382e-01 4.98087496e-01 1.62632620e+00 1.47879541e-01
-1.92161858e-01 7.44959772e-01 -9.16545212e-01 -3.82787049e-01
-3.67781192e-01 2.76423305e-01 4.12937671e-01 3.58138345e-02
7.31261969e-01 -2.77891040e-01 -2.31728166e-01 3.78167659e-01
-9.52623367e-01 1.36829168e-01 1.77102759e-01 5.03256619e-01
6.38378263e-01 -3.10493290e-01 4.02522653e-01 -6.84430301e-01
3.36288601e-01 -9.10603464e-01 -3.06282490e-01 -1.29488394e-01
-1.06324427e-01 -3.17928731e-01 1.91050529e-01 -5.41571319e-01
-4.03012812e-01 -1.63135946e-01 9.61767957e-02 -7.52668917e-01
-2.11772218e-01 5.71377933e-01 1.18807286e-01 3.00803691e-01
5.67645252e-01 -9.17714909e-02 -2.69195914e-01 -4.89896446e-01
3.32960278e-01 7.83962250e-01 7.70157158e-01 -8.53009045e-01
6.82116568e-01 5.99028468e-01 -3.00891876e-01 -1.03330231e+00
-1.22226655e+00 -7.16643631e-01 -8.27482522e-01 -4.47557271e-01
9.61308539e-01 -1.23964751e+00 -2.06316158e-01 2.72630394e-01
-8.71026039e-01 -5.10471940e-01 -2.67631143e-01 4.89716709e-01
-8.39298964e-01 4.55402493e-01 -5.62068820e-01 -3.89278412e-01
7.24265948e-02 -8.68464053e-01 1.31021225e+00 -4.63070035e-01
-7.00917363e-01 -6.82952642e-01 2.08137795e-01 6.19354010e-01
-2.86659673e-02 5.25849044e-01 3.12141627e-01 -7.10431516e-01
-5.36133528e-01 -3.74324620e-01 1.24662161e-01 3.58995259e-01
5.18159475e-03 2.50105947e-01 -1.07351530e+00 -5.00582039e-01
-1.90040782e-01 -7.29971051e-01 5.90994775e-01 2.06570119e-01
1.16748428e+00 -5.30559003e-01 -2.10691579e-02 5.94026804e-01
1.10108614e+00 -1.52390987e-01 3.65825981e-01 5.89033902e-01
7.88430512e-01 4.76774096e-01 8.84724855e-01 6.66506350e-01
4.60259885e-01 8.93166840e-01 3.45425904e-01 2.49130785e-01
-6.58904612e-02 -6.21392012e-01 5.80677927e-01 1.19629872e+00
-1.54302508e-01 -4.40722972e-01 -1.12164426e+00 7.91003823e-01
-1.77858937e+00 -1.47081125e+00 2.01239400e-02 2.02168489e+00
7.94767618e-01 1.38266131e-01 4.25824493e-01 1.88373551e-01
5.88932633e-01 4.92994964e-01 -1.80093706e-01 9.51924250e-02
-2.64159799e-01 -4.22174096e-01 1.30785123e-01 1.43125713e-01
-1.37573111e+00 7.65629113e-01 6.83186960e+00 5.50724864e-01
-9.11948502e-01 1.61070287e-01 4.38722432e-01 -7.36815810e-01
-2.40266666e-01 -5.54258823e-02 -9.07610595e-01 6.07946396e-01
1.35228515e+00 -2.47817606e-01 4.19248730e-01 7.99184561e-01
2.68524826e-01 4.68820408e-02 -1.70041323e+00 1.23825908e+00
5.02296746e-01 -1.60974503e+00 2.63100505e-01 7.34173646e-03
8.16963732e-01 3.22437078e-01 2.49662586e-02 2.82514840e-01
-1.05352238e-01 -7.90216267e-01 1.39633739e+00 3.33441287e-01
1.08183289e+00 -3.74175996e-01 5.61174214e-01 2.09743634e-01
-1.20656216e+00 -1.11006968e-01 -3.24497044e-01 -2.69151062e-01
2.42077440e-01 3.07398438e-01 -7.44510293e-01 3.70432973e-01
1.16082394e+00 1.30413234e+00 -7.48073399e-01 5.97208917e-01
2.73439050e-01 6.99935377e-01 -1.46684602e-01 2.21153781e-01
5.87439179e-01 1.48291290e-01 6.47008359e-01 1.49731195e+00
3.85074079e-01 -5.48076630e-02 3.81826252e-01 4.41179216e-01
-3.56941700e-01 8.08621421e-02 -1.12074220e+00 -6.93939030e-01
5.88331878e-01 1.01186287e+00 -5.35352051e-01 -5.28899312e-01
-7.65460193e-01 6.56446755e-01 1.35569662e-01 1.24311328e-01
-1.20155382e+00 -3.02696247e-02 7.86573052e-01 4.29567486e-01
5.28365597e-02 -4.27158624e-01 3.91915500e-01 -1.47636878e+00
1.22257516e-01 -1.28201699e+00 5.94157398e-01 -1.01303589e+00
-1.29463792e+00 6.87870264e-01 3.13794702e-01 -1.53065801e+00
-4.64918315e-01 -3.62253308e-01 -8.73526931e-03 4.10933420e-02
-1.10614371e+00 -1.10890174e+00 -7.00620770e-01 8.24506521e-01
9.20290232e-01 -3.97228062e-01 6.88072503e-01 7.05438614e-01
-5.13834655e-01 5.01789808e-01 -9.28499475e-02 3.47411036e-01
1.16119742e+00 -8.54914486e-01 3.85147870e-01 9.07255530e-01
7.81554282e-01 4.49916720e-01 8.03498268e-01 -4.74611193e-01
-1.50894547e+00 -1.08712506e+00 9.48988318e-01 -1.13625789e+00
1.21374559e+00 -7.01355875e-01 -5.29927671e-01 1.18377471e+00
2.35374451e-01 5.17176576e-02 9.20666754e-01 6.51168544e-03
-5.93176901e-01 -3.32845971e-02 -5.83747447e-01 4.75408107e-01
1.33993423e+00 -1.11467004e+00 -6.48033261e-01 5.53163528e-01
8.30414593e-01 -4.84116405e-01 -9.16137040e-01 1.57374769e-01
6.89436018e-01 -8.59571517e-01 8.87925327e-01 -1.01206589e+00
8.01205039e-01 -7.69169033e-02 -6.82504714e-01 -6.83846831e-01
-1.53829575e-01 -1.09010351e+00 -2.05537930e-01 1.60646605e+00
1.30739212e-01 1.00628495e-01 6.56497717e-01 3.43465030e-01
-2.78785139e-01 -4.87020373e-01 -7.66270339e-01 -9.41760898e-01
-2.78053373e-01 -8.84014964e-01 4.40655768e-01 1.29016089e+00
1.13644011e-01 1.47697702e-01 -6.78431332e-01 -7.38021061e-02
4.25778925e-01 -5.75461537e-02 1.10868180e+00 -9.75632548e-01
-2.06945002e-01 -1.95802823e-01 -6.73618257e-01 -9.91285264e-01
3.24404836e-01 -8.35712433e-01 2.27395464e-02 -1.13810682e+00
4.13960963e-01 -8.94235596e-02 -2.47037977e-01 3.72633576e-01
1.96051136e-01 6.66357040e-01 2.83358604e-01 3.91579032e-01
-1.11341894e+00 1.16140954e-01 7.19034553e-01 -2.60142833e-01
2.71258235e-01 -4.14462537e-01 -5.72237730e-01 8.68165851e-01
5.26098847e-01 -5.69016159e-01 -4.44614410e-01 -8.31922889e-01
6.05099857e-01 -1.39441997e-01 5.00890851e-01 -1.15574253e+00
2.32944489e-02 -6.46038353e-02 -7.54275322e-02 -5.92175424e-01
3.74485523e-01 -6.41707361e-01 2.08171189e-01 -1.29644796e-01
-7.36045122e-01 5.48438311e-01 2.25692675e-01 6.11082733e-01
-5.72107852e-01 -7.14568794e-02 3.21510166e-01 -1.84230089e-01
-9.73377645e-01 4.44780350e-01 -4.36049074e-01 7.21985996e-01
9.53869164e-01 -2.67268866e-01 -3.72741580e-01 -6.12508774e-01
-6.11873627e-01 -1.15793265e-01 5.98801017e-01 9.32947040e-01
4.13457543e-01 -1.55300629e+00 -8.43627810e-01 -1.80514336e-01
5.16609430e-01 -2.31729254e-01 1.91475525e-01 8.45557153e-01
-5.46178699e-01 6.11518323e-01 -2.05296546e-01 -8.64616930e-01
-1.41506517e+00 5.81596136e-01 -1.67119369e-01 2.69426674e-01
-7.23249435e-01 9.87134278e-01 -9.28448811e-02 2.34348312e-01
2.83538133e-01 -4.80046868e-01 3.38426195e-02 4.26059514e-01
6.51201367e-01 4.61427242e-01 -8.53862092e-02 -1.00381804e+00
-3.74538332e-01 3.37836593e-01 1.15244046e-01 -8.89244601e-02
1.37930632e+00 -2.91845769e-01 8.72391462e-02 9.70266223e-01
1.29881907e+00 2.42627591e-01 -1.13326120e+00 -1.58055499e-01
2.71034300e-01 -6.41578734e-01 -6.51571527e-02 -1.94277316e-01
-7.44668365e-01 7.30436563e-01 1.01024404e-01 3.99469942e-01
9.04321432e-01 2.74893105e-01 1.02491927e+00 4.99079973e-01
5.81165552e-01 -9.87687171e-01 4.49124038e-01 5.02252996e-01
9.79660451e-01 -1.30230141e+00 1.11901067e-01 -5.83339436e-03
-7.09188342e-01 8.93898308e-01 4.35777783e-01 -4.21619639e-02
3.08766186e-01 3.14976871e-01 1.61774957e-03 -2.39524275e-01
-9.73405540e-01 1.06242232e-01 4.82735127e-01 3.31809402e-01
5.96916199e-01 -1.30254805e-01 1.98236436e-01 6.87602103e-01
-3.82118106e-01 4.25452702e-02 6.56037390e-01 9.36410904e-01
-1.77738562e-01 -8.52337420e-01 -8.73700082e-02 2.20273271e-01
-7.64153421e-01 -1.46706730e-01 -4.99070674e-01 9.34543610e-01
6.19897433e-02 1.00584674e+00 2.35853106e-01 -5.81092775e-01
1.19532585e-01 3.75156514e-02 3.85397077e-01 -8.11890423e-01
-3.21890503e-01 -7.29438663e-03 4.94484752e-01 -1.13531315e+00
-9.34680641e-01 -8.85675490e-01 -8.15213025e-01 -4.66270089e-01
5.61864302e-02 1.45907655e-01 3.47026139e-01 7.81405330e-01
3.52003276e-01 1.90744475e-01 3.06264937e-01 -1.18906534e+00
-2.18337014e-01 -8.94151926e-01 -3.07632089e-01 8.47064734e-01
3.70317757e-01 -7.23898590e-01 -4.34873223e-01 6.86067224e-01] | [10.428953170776367, 0.8274787068367004] |
d33d057c-ad18-4a5b-aa81-690a161c74b1 | simple-yet-surprisingly-effective-training | 2212.13918 | null | https://arxiv.org/abs/2212.13918v1 | https://arxiv.org/pdf/2212.13918v1.pdf | Simple Yet Surprisingly Effective Training Strategies for LSTMs in Sensor-Based Human Activity Recognition | Human Activity Recognition (HAR) is one of the core research areas in mobile and wearable computing. With the application of deep learning (DL) techniques such as CNN, recognizing periodic or static activities (e.g, walking, lying, cycling, etc.) has become a well studied problem. What remains a major challenge though is the sporadic activity recognition (SAR) problem, where activities of interest tend to be non periodic, and occur less frequently when compared with the often large amount of irrelevant background activities. Recent works suggested that sequential DL models (such as LSTMs) have great potential for modeling nonperiodic behaviours, and in this paper we studied some LSTM training strategies for SAR. Specifically, we proposed two simple yet effective LSTM variants, namely delay model and inverse model, for two SAR scenarios (with and without time critical requirement). For time critical SAR, the delay model can effectively exploit predefined delay intervals (within tolerance) in form of contextual information for improved performance. For regular SAR task, the second proposed, inverse model can learn patterns from the time series in an inverse manner, which can be complementary to the forward model (i.e.,LSTM), and combining both can boost the performance. These two LSTM variants are very practical, and they can be deemed as training strategies without alteration of the LSTM fundamentals. We also studied some additional LSTM training strategies, which can further improve the accuracy. We evaluated our models on two SAR and one non-SAR datasets, and the promising results demonstrated the effectiveness of our approaches in HAR applications. | ['Thomas Ploetz', 'Paolo Missier', 'Xin Guan', 'Yu Guan', 'Shuai Shao'] | 2022-12-23 | null | null | null | null | ['human-activity-recognition', 'human-activity-recognition'] | ['computer-vision', 'time-series'] | [ 5.14420509e-01 -2.12227300e-01 -3.74623328e-01 -1.61966383e-02
-1.82535276e-01 -1.45453826e-01 5.44434965e-01 -4.79162186e-02
-5.14357686e-01 8.74304056e-01 2.47920290e-01 -1.66752636e-01
-2.68003047e-01 -7.74208486e-01 -6.28636897e-01 -9.33505654e-01
-4.62743372e-01 -2.61450320e-01 4.62757915e-01 -6.93137422e-02
5.41356392e-02 4.46961880e-01 -1.70790184e+00 3.84916246e-01
8.91971707e-01 1.25116241e+00 1.69611707e-01 3.40359241e-01
-2.11860895e-01 1.23626983e+00 -7.34695613e-01 3.25751871e-01
-2.13234857e-01 -5.59748352e-01 -4.85526413e-01 -1.09647773e-01
-3.79122168e-01 6.86553866e-02 -3.51399630e-01 4.62301970e-01
6.94159448e-01 4.69254702e-01 9.64769349e-02 -9.95580733e-01
-9.79271308e-02 4.46121603e-01 -5.30862391e-01 5.98400056e-01
4.55473602e-01 -1.29124239e-01 3.52154434e-01 -5.95214605e-01
-4.45411839e-02 7.71516979e-01 8.88452411e-01 4.61906910e-01
-8.35992277e-01 -6.08653069e-01 2.67855316e-01 6.07018411e-01
-1.32307041e+00 -4.97128129e-01 9.38595533e-01 -1.18360072e-01
1.10856175e+00 3.11083227e-01 8.73156846e-01 1.60925233e+00
4.36404765e-01 1.13768327e+00 1.13127649e+00 -3.68933380e-01
3.45578045e-01 -2.21024379e-01 2.38699671e-02 2.94210047e-01
2.73390561e-02 2.47504842e-02 -5.93893647e-01 -8.87781605e-02
7.93239534e-01 3.76881659e-01 -3.99272949e-01 8.61630589e-02
-1.40831053e+00 3.66197467e-01 2.81736791e-01 8.52290094e-01
-5.41883290e-01 9.86201242e-02 5.62520325e-01 1.70515344e-01
4.64693934e-01 1.18166722e-01 -4.44658190e-01 -5.55553377e-01
-8.61918628e-01 -1.22089028e-01 6.81098223e-01 5.82264006e-01
3.90230566e-01 4.13270205e-01 -5.21680236e-01 9.62647915e-01
4.14700620e-03 2.45215937e-01 1.02458704e+00 -4.66159374e-01
4.47181880e-01 5.54276645e-01 1.04125135e-01 -1.01172876e+00
-6.18133664e-01 -7.04843640e-01 -1.22761106e+00 -5.72201252e-01
2.51118213e-01 -2.16163516e-01 -8.22060168e-01 1.70225060e+00
3.16518582e-02 7.61301458e-01 -9.80213359e-02 5.74389756e-01
5.29505491e-01 8.11513841e-01 2.17974439e-01 -7.39641845e-01
1.13586295e+00 -8.49848986e-01 -1.14015436e+00 -4.48466659e-01
7.60053515e-01 -3.71627659e-01 9.55440283e-01 4.99110132e-01
-7.65536487e-01 -5.95008969e-01 -1.13819170e+00 3.70816350e-01
-3.08194637e-01 2.15403572e-01 7.07465589e-01 6.89963639e-01
-8.37847650e-01 7.02050865e-01 -1.20516932e+00 -4.82186764e-01
2.88152128e-01 3.93697500e-01 -3.78112644e-02 1.92311794e-01
-1.51215363e+00 6.13316476e-01 4.47210491e-01 6.23782039e-01
-7.29234040e-01 -1.87403545e-01 -7.67374098e-01 -1.67092010e-02
4.38383102e-01 -1.57275259e-01 1.16589355e+00 -1.27094913e+00
-1.49626386e+00 3.69455844e-01 -2.74252892e-01 -8.01622093e-01
4.15820330e-01 -3.30989242e-01 -9.06647027e-01 -8.17924961e-02
-2.80658394e-01 4.83510531e-02 7.95110345e-01 -6.76318586e-01
-4.96827453e-01 -2.30038702e-01 -1.05885081e-01 1.20635062e-01
-7.06738114e-01 -5.91178164e-02 -9.39501002e-02 -7.91508019e-01
1.50692267e-02 -7.55913258e-01 -1.27280608e-01 -3.73910338e-01
-7.82728866e-02 -2.95656800e-01 1.00888288e+00 -7.72699296e-01
1.83385217e+00 -2.05020618e+00 -2.82444030e-01 1.39045075e-01
-1.51377857e-01 6.04987264e-01 1.81813300e-01 5.34896374e-01
-1.72691084e-02 -8.91326666e-02 -3.27965379e-01 2.29192581e-02
-3.76900226e-01 5.44256628e-01 -1.37373820e-01 3.18500072e-01
6.72632977e-02 9.10673141e-01 -8.84171188e-01 -3.36254865e-01
1.25704840e-01 5.40553331e-01 9.30126756e-02 1.90455481e-01
-7.52218068e-02 5.75869620e-01 -4.62411255e-01 5.26190996e-01
2.60930389e-01 -1.91219330e-01 2.53200352e-01 -1.80328771e-01
-2.39469275e-01 2.75569111e-01 -1.22175872e+00 1.37947261e+00
-5.75160384e-01 6.15530849e-01 -5.02372801e-01 -1.39526081e+00
1.03669727e+00 5.28355360e-01 7.39405036e-01 -1.23430300e+00
-1.25847803e-02 1.05214126e-01 1.53009191e-01 -7.20408678e-01
2.08075956e-01 2.45708734e-01 1.44653484e-01 4.51635987e-01
-2.26664126e-01 1.00383854e+00 3.89095135e-02 -5.45263350e-01
1.23738050e+00 4.06756908e-01 5.77756464e-01 -7.79229254e-02
5.96740127e-01 -3.76387775e-01 8.25536966e-01 6.48337185e-01
-1.21983573e-01 2.17848465e-01 2.01424941e-01 -6.24240875e-01
-4.71068412e-01 -6.68095708e-01 2.24466339e-01 1.09598851e+00
8.11210573e-02 -2.21067652e-01 -5.49028158e-01 -4.11381185e-01
-6.42768145e-01 4.93806362e-01 -4.52903271e-01 -4.42651898e-01
-1.09743488e+00 -1.08364248e+00 8.40400159e-01 9.66156185e-01
1.19729781e+00 -1.35763478e+00 -9.72552419e-01 6.57013834e-01
-3.68714184e-01 -1.07578814e+00 -3.47163826e-01 4.54319894e-01
-1.08825779e+00 -8.66255462e-01 -8.38204265e-01 -8.29177856e-01
1.79531604e-01 3.77064645e-01 7.32013106e-01 -5.27041219e-03
1.30964801e-01 3.21519971e-01 -5.04527032e-01 -3.76028895e-01
1.54501840e-01 1.59611270e-01 1.08330674e-01 5.87076485e-01
2.54800081e-01 -8.55324209e-01 -8.72237802e-01 5.55907071e-01
-9.09067512e-01 -4.20885123e-02 7.84807861e-01 7.71036029e-01
6.40928090e-01 1.38780221e-01 7.88591206e-01 -6.93439782e-01
8.01288605e-01 -5.20400405e-01 8.50884691e-02 3.40480059e-01
-3.77164066e-01 -6.49872124e-02 8.52055490e-01 -1.06705058e+00
-1.10138190e+00 -1.01864740e-01 -3.09942067e-01 -3.60744715e-01
9.38516296e-03 7.44765520e-01 -3.46702427e-01 2.35041678e-02
5.04459620e-01 8.05703819e-01 -3.41077179e-01 -5.83520353e-01
-3.65175366e-01 6.51707947e-01 1.43284649e-01 -5.79050303e-01
6.19627386e-02 3.69238853e-01 -1.84305921e-01 -1.14125276e+00
-5.81935406e-01 -2.54406154e-01 -3.11892867e-01 -2.95455396e-01
6.40450239e-01 -7.57176220e-01 -6.13532782e-01 7.82488644e-01
-9.71515417e-01 -6.39424145e-01 -2.66537398e-01 5.18651783e-01
-4.15624171e-01 3.59412134e-01 -5.55003107e-01 -1.16606021e+00
-3.86246264e-01 -6.87442601e-01 6.73879445e-01 4.11401868e-01
-3.38067740e-01 -1.26984608e+00 -6.47666007e-02 -7.84517303e-02
5.82910895e-01 5.59741437e-01 6.90887272e-01 -7.98848689e-01
-2.47976050e-01 -1.83580935e-01 2.13178724e-01 3.64977002e-01
3.09766650e-01 -3.60843748e-01 -1.08393669e+00 -2.08938941e-01
4.22619432e-01 8.37205499e-02 5.97538352e-01 4.35182244e-01
1.49145043e+00 -5.67772269e-01 -4.46559250e-01 2.90171653e-01
9.52555358e-01 8.75722229e-01 1.19157183e+00 3.86632383e-01
5.82499743e-01 2.36332402e-01 6.71521246e-01 4.39958096e-01
1.96003363e-01 8.60573232e-01 9.04261395e-02 -3.41487765e-01
1.49425238e-01 -2.03170300e-01 7.32911289e-01 1.04542124e+00
-5.97592473e-01 -2.84188241e-01 -8.04900885e-01 4.83403713e-01
-2.18950725e+00 -1.25502908e+00 -1.09211616e-01 2.38208818e+00
5.79514265e-01 2.83836663e-01 3.05210054e-01 7.03881621e-01
5.78531206e-01 4.96615946e-01 -4.98439431e-01 -4.26117510e-01
-1.65027976e-01 4.08429265e-01 3.47302616e-01 -1.96430236e-01
-1.15442967e+00 4.52257216e-01 5.79531288e+00 1.11489367e+00
-1.30482471e+00 3.31025004e-01 5.87777197e-01 4.18161042e-02
1.54921979e-01 -3.10806572e-01 -6.02529943e-01 9.30590928e-01
1.25065684e+00 6.85619190e-02 1.58932194e-01 6.67415977e-01
6.63507044e-01 -2.00114861e-01 -9.95068073e-01 1.13013351e+00
-7.26019889e-02 -9.91550684e-01 -1.82444364e-01 3.44341807e-02
5.10116756e-01 -3.48081470e-01 -2.43454322e-01 5.80590487e-01
-3.71703684e-01 -9.48123932e-01 3.94755900e-01 7.86256313e-01
5.14535010e-01 -6.55325174e-01 9.50004697e-01 6.31428480e-01
-1.66387403e+00 -2.71026611e-01 4.39821742e-02 -3.35599840e-01
2.07622990e-01 5.67525804e-01 -1.35560349e-01 7.16541052e-01
9.37168241e-01 1.12043142e+00 -4.69438851e-01 1.09797347e+00
-9.72773805e-02 8.58434916e-01 -3.35242659e-01 -3.33569318e-01
1.59092769e-01 2.82635540e-02 3.24998319e-01 1.15931141e+00
5.36094427e-01 3.94760780e-02 1.66614935e-01 3.17231238e-01
2.72264957e-01 -7.19707906e-02 -6.23874843e-01 -5.90757690e-02
3.31701845e-01 9.25611019e-01 -7.07083166e-01 -2.49735028e-01
-3.62548143e-01 7.78480113e-01 -8.66704062e-02 4.93496448e-01
-1.14228892e+00 -3.84270877e-01 2.65904605e-01 2.49452740e-01
9.06390920e-02 -3.96615952e-01 -1.55764237e-01 -1.09655738e+00
3.30432445e-01 -8.54769945e-01 5.13703763e-01 -5.52090704e-01
-8.72939587e-01 5.13118207e-01 6.19293898e-02 -1.61524451e+00
-2.55406350e-01 -2.55427271e-01 -9.49956238e-01 3.48869622e-01
-1.32034147e+00 -9.15403187e-01 -4.44665670e-01 9.26897943e-01
7.55768776e-01 1.70217723e-01 6.52635694e-01 6.52748704e-01
-7.51962721e-01 5.10734200e-01 -6.49501607e-02 -6.10401630e-02
2.36557722e-01 -7.65325010e-01 -3.86954993e-02 9.55635488e-01
1.31395757e-01 6.16352320e-01 4.49505210e-01 -3.92095864e-01
-1.24894011e+00 -1.22547245e+00 7.92660296e-01 9.68233570e-02
4.40108776e-01 -3.64116937e-01 -1.10364842e+00 7.27730751e-01
6.77919760e-03 -1.60646588e-01 7.95874715e-01 -1.57724231e-01
3.95527571e-01 -3.71629477e-01 -8.36685956e-01 7.30887234e-01
1.40325797e+00 -2.66794860e-01 -5.03881276e-01 8.99237096e-02
5.31862676e-01 -3.10489595e-01 -8.81541491e-01 6.63645208e-01
7.03442872e-01 -9.37252581e-01 8.77988815e-01 -2.80426562e-01
-6.18464164e-02 -3.05909246e-01 -2.22655423e-02 -1.07536447e+00
-1.27312303e-01 -6.99893177e-01 -9.73641753e-01 1.14913082e+00
1.82320643e-02 -8.73319030e-01 6.72104776e-01 1.91575661e-01
-3.51578116e-01 -1.01177597e+00 -1.00483692e+00 -1.16060972e+00
-7.23530829e-01 -4.82732803e-01 4.44981784e-01 8.66346419e-01
-9.08552557e-02 1.04027323e-01 -8.72160912e-01 -1.98864579e-01
5.38174324e-02 -9.91371498e-02 4.55475539e-01 -9.82939541e-01
-2.80386209e-01 -1.64462477e-01 -3.64767015e-01 -1.31822062e+00
-2.66981900e-01 -4.07263041e-01 -5.36151305e-02 -1.53603506e+00
-2.50344425e-01 -2.17849329e-01 -7.59890974e-01 7.70734012e-01
1.20966665e-01 -2.51063295e-02 -2.61930317e-01 1.79862201e-01
-5.49853027e-01 7.36974716e-01 9.89200830e-01 -1.87288135e-01
-6.75615966e-01 5.74829936e-01 -3.91533375e-01 6.25638843e-01
1.03477311e+00 -2.62264222e-01 -6.30118310e-01 -2.12696105e-01
5.29161952e-02 1.48782387e-01 3.18876386e-01 -1.46152735e+00
3.21842790e-01 -1.62683725e-01 4.95669663e-01 -4.79010850e-01
4.17928159e-01 -7.84289598e-01 4.17366266e-01 7.28015721e-01
-1.59433499e-01 1.52042076e-01 3.36081535e-01 6.42447591e-01
-3.12958926e-01 2.06321672e-01 4.09249067e-01 -6.96605965e-02
-9.69224274e-01 2.34878898e-01 -7.70834088e-01 -2.09008172e-01
1.05622375e+00 -6.93797290e-01 -1.57092720e-01 -3.64926577e-01
-9.10622060e-01 7.66757652e-02 -2.87824720e-01 5.89226604e-01
6.23714209e-01 -1.49675190e+00 -8.46499130e-02 1.74049243e-01
-4.05856147e-02 -2.74161249e-01 5.22011936e-01 1.42983246e+00
-1.58271641e-01 4.65120286e-01 -2.09280103e-01 -6.68629944e-01
-9.94978786e-01 4.94943738e-01 5.24094403e-01 -5.20134211e-01
-5.74541867e-01 1.83699027e-01 -6.87422901e-02 4.90025319e-02
4.72167343e-01 -5.65584362e-01 -6.51421905e-01 7.34214410e-02
7.01981246e-01 6.43965602e-01 2.90900558e-01 -2.93340564e-01
-5.75036764e-01 4.63585287e-01 1.25280902e-01 2.75619119e-01
1.25561309e+00 7.43192388e-03 1.40885144e-01 8.32161367e-01
8.64341497e-01 -2.91476816e-01 -1.06701171e+00 -3.09867799e-01
5.63323975e-01 -7.19302669e-02 -1.53283268e-01 -5.75805247e-01
-1.02809501e+00 1.00169492e+00 9.12581325e-01 4.10304129e-01
1.61873102e+00 -4.14422095e-01 1.08323693e+00 3.29700589e-01
5.71390867e-01 -1.12922287e+00 2.33879179e-01 4.30483609e-01
5.73135555e-01 -8.55759263e-01 -2.84915358e-01 8.58760905e-03
-5.85135400e-01 1.11700654e+00 6.63895667e-01 3.62428613e-02
5.00972807e-01 1.76466480e-01 -3.42912406e-01 9.89232585e-02
-7.08055496e-01 -3.00636142e-01 2.95223713e-01 6.28626764e-01
4.67550755e-01 -9.35018510e-02 -6.10352516e-01 8.86157155e-01
3.23487997e-01 3.87163281e-01 2.37897232e-01 1.26876366e+00
-3.61293107e-01 -8.74225140e-01 -1.93329021e-01 7.27107227e-01
-4.89973575e-01 1.63861647e-01 1.10040549e-02 7.14320660e-01
3.87211114e-01 9.95630622e-01 -2.54322086e-02 -7.02792168e-01
5.51163912e-01 8.71054009e-02 2.92530954e-01 -3.20237309e-01
-6.31816447e-01 1.40823022e-01 2.09297463e-01 -8.12722743e-01
-7.89101601e-01 -5.35217285e-01 -1.01510406e+00 -1.16938114e-01
-1.09765343e-01 1.51283275e-02 1.92312270e-01 1.27305281e+00
4.22808766e-01 8.38295698e-01 6.12917721e-01 -8.57799590e-01
-2.11862415e-01 -1.00679600e+00 -5.52562118e-01 5.91720976e-02
1.76775619e-01 -7.37699807e-01 -1.63936302e-01 -6.60707988e-03] | [7.5750298500061035, 0.8106054663658142] |
d3778b35-0625-4e2f-b6f1-d6209fc41af4 | a-spectral-approach-to-off-policy-evaluation | 2109.10502 | null | https://arxiv.org/abs/2109.10502v1 | https://arxiv.org/pdf/2109.10502v1.pdf | A Spectral Approach to Off-Policy Evaluation for POMDPs | We consider off-policy evaluation (OPE) in Partially Observable Markov Decision Processes, where the evaluation policy depends only on observable variables but the behavior policy depends on latent states (Tennenholtz et al. (2020a)). Prior work on this problem uses a causal identification strategy based on one-step observable proxies of the hidden state, which relies on the invertibility of certain one-step moment matrices. In this work, we relax this requirement by using spectral methods and extending one-step proxies both into the past and future. We empirically compare our OPE methods to existing ones and demonstrate their improved prediction accuracy and greater generality. Lastly, we derive a separate Importance Sampling (IS) algorithm which relies on rank, distinctness, and positivity conditions, and not on the strict sufficiency conditions of observable trajectories with respect to the reward and hidden-state structure required by Tennenholtz et al. (2020a). | ['Nan Jiang', 'Yash Nair'] | 2021-09-22 | null | null | null | null | ['causal-identification'] | ['reasoning'] | [ 3.06183517e-01 2.94612765e-01 -6.67127550e-01 6.67100102e-02
-2.37936750e-01 -7.78705001e-01 8.58003974e-01 -1.06257468e-01
-3.68090183e-01 9.74646151e-01 4.37783182e-01 -6.73000038e-01
-6.13426745e-01 -4.36734170e-01 -3.25063646e-01 -6.97782874e-01
-2.95306176e-01 4.91259158e-01 1.07289687e-01 1.29759103e-01
-1.44020631e-03 4.88880605e-01 -1.17042768e+00 -2.91831285e-01
6.26047850e-01 6.99368536e-01 -4.04827110e-02 9.68886018e-01
4.74949479e-01 9.85364556e-01 -6.81867972e-02 -1.73410237e-01
3.47291797e-01 -4.40775096e-01 -7.46283710e-01 1.21259801e-01
-3.83822143e-01 -5.64791620e-01 -3.09362590e-01 1.12868953e+00
1.29437938e-01 1.81946278e-01 8.93186212e-01 -1.52238894e+00
-4.22997564e-01 4.85616356e-01 -2.16798276e-01 2.50749826e-01
1.72635764e-01 3.87773931e-01 1.38176119e+00 -3.47111821e-01
4.44079190e-01 1.17773962e+00 4.96780783e-01 4.82249886e-01
-1.52372456e+00 -3.65247875e-01 4.72564787e-01 2.14416951e-01
-7.74342239e-01 -3.85105282e-01 3.82255018e-01 -7.62519419e-01
6.36021078e-01 2.50517398e-01 6.95330977e-01 1.60128868e+00
2.02928156e-01 7.94801712e-01 1.29079795e+00 -3.03203940e-01
5.72547197e-01 -1.55965447e-01 4.08922017e-01 4.80064362e-01
3.81890148e-01 7.23869860e-01 -3.08345169e-01 -5.80605090e-01
1.06206179e+00 1.98527381e-01 -7.61075243e-02 -4.95830536e-01
-1.19081545e+00 9.19720471e-01 -4.91580039e-01 -4.57192287e-02
-7.95842648e-01 1.73068643e-01 8.21732804e-02 3.61483127e-01
1.74995825e-01 4.25702155e-01 -3.62517148e-01 -4.84579712e-01
-7.11745441e-01 9.31847095e-02 1.00023317e+00 7.60272682e-01
5.43419600e-01 3.49408872e-02 -7.48223007e-01 -7.36039728e-02
2.16686532e-01 6.72123790e-01 1.33451059e-01 -1.44536161e+00
2.87317008e-01 1.28646687e-01 8.72484922e-01 -4.12190795e-01
-2.66187847e-01 -6.47767305e-01 -5.41425228e-01 2.37248316e-02
6.09773040e-01 -4.17463094e-01 -8.17844450e-01 2.17859364e+00
2.58340538e-01 2.45733261e-01 7.34452233e-02 7.41092265e-01
-5.06527841e-01 4.76819217e-01 -3.77710047e-03 -6.69111252e-01
9.16217625e-01 -5.33241451e-01 -6.47184968e-01 -1.17588826e-02
3.47818226e-01 -2.89610296e-01 9.23236728e-01 3.28700811e-01
-9.70881402e-01 -8.03952292e-02 -5.67302287e-01 4.01365787e-01
3.34757596e-01 3.10100526e-01 7.17120290e-01 7.69146562e-01
-1.05989623e+00 6.84505045e-01 -1.14347506e+00 -2.36237049e-01
-2.10178554e-01 2.35702381e-01 1.62566721e-01 4.20699924e-01
-1.25331485e+00 7.47296453e-01 -4.49245749e-03 1.57977212e-02
-1.50594389e+00 -4.27190453e-01 -3.42023641e-01 1.86248794e-01
9.01256382e-01 -7.33115971e-01 1.61236513e+00 -8.80950153e-01
-1.84120095e+00 4.13281322e-02 -4.94840518e-02 -6.33987486e-01
8.49522650e-01 -3.20984811e-01 -2.86290526e-01 4.47356254e-01
2.04746932e-01 -1.85696810e-01 9.99414921e-01 -8.44477594e-01
-6.33102059e-01 -3.12379330e-01 3.90844315e-01 1.71759039e-01
-1.15976959e-01 -6.85667247e-02 -8.29420611e-02 -3.80223244e-01
-1.42060384e-01 -1.23554361e+00 -5.03494143e-01 -5.28211653e-01
-5.23120642e-01 -5.34132496e-02 2.36419871e-01 -6.75925732e-01
1.20643675e+00 -1.89254713e+00 1.90394267e-01 4.78619456e-01
1.24091402e-01 -1.38298228e-01 -5.91771975e-02 6.85250103e-01
-2.64445543e-01 1.38687953e-01 -2.17644617e-01 -3.31502646e-01
3.06544185e-01 3.47894281e-01 -8.24883699e-01 7.78784156e-01
2.02204093e-01 6.55086577e-01 -1.01867092e+00 -1.18588015e-01
1.19881436e-01 -1.32081479e-01 -5.05190313e-01 -7.08969906e-02
-2.86526680e-01 6.72310770e-01 -6.39825702e-01 3.72118056e-01
-1.70089751e-02 -3.83449584e-01 5.46513796e-01 3.91236007e-01
-4.67900038e-01 5.22961318e-01 -1.28576910e+00 8.61148000e-01
-2.80661941e-01 1.49953261e-01 6.76073283e-02 -8.88575017e-01
8.63421410e-02 6.46211088e-01 8.63782465e-01 -3.82066667e-01
3.14243883e-02 -4.82064337e-02 1.38476565e-01 -2.55175054e-01
3.81647527e-01 -3.75151753e-01 -6.50850087e-02 6.96614146e-01
-1.12735912e-01 5.55150092e-01 4.24154788e-01 2.89066553e-01
1.18233359e+00 3.29731822e-01 4.72527444e-01 -5.19916892e-01
2.25822642e-01 -1.22836955e-01 8.79193246e-01 1.01431692e+00
-3.59737515e-01 -3.54141816e-02 1.17182052e+00 2.47349873e-01
-1.14251471e+00 -1.12962401e+00 2.27527823e-02 8.91317427e-01
-2.40068033e-01 -2.08154738e-01 -5.38357556e-01 -4.65984434e-01
1.01182155e-01 8.92683148e-01 -9.80437577e-01 -1.91019118e-01
-1.70935616e-01 -6.40835106e-01 3.92570496e-01 5.34911215e-01
1.01515219e-01 -6.03699028e-01 -8.01074922e-01 3.53985369e-01
-8.83385837e-02 -7.62388289e-01 -3.94277900e-01 2.82569051e-01
-8.42141151e-01 -8.67015183e-01 -8.40176702e-01 2.75549948e-01
5.63092351e-01 -5.87804057e-02 7.98535407e-01 -6.33424878e-01
4.27796930e-01 8.57335150e-01 4.98667583e-02 -2.42135316e-01
-3.37757200e-01 -3.46412808e-02 3.92032266e-01 3.68237227e-01
-1.11808956e-01 -6.76634431e-01 -4.76139665e-01 2.07845628e-01
-6.22884274e-01 -1.56472612e-03 7.80642211e-01 8.53061259e-01
5.47348082e-01 -1.34920403e-01 2.06527948e-01 -7.44442403e-01
6.16835415e-01 -5.73418200e-01 -1.05696857e+00 3.26344192e-01
-1.05183434e+00 5.75919688e-01 4.59980786e-01 -6.86411798e-01
-1.19879043e+00 1.62803352e-01 4.74690795e-01 -5.13488650e-01
1.84558526e-01 3.67704660e-01 -8.63026455e-03 5.99110901e-01
1.24374971e-01 1.85189545e-01 -1.74136072e-01 -4.86786604e-01
3.65623325e-01 1.45507932e-01 3.67433637e-01 -9.20032203e-01
7.89903641e-01 7.34020650e-01 3.95734698e-01 -4.81836587e-01
-7.85402477e-01 -4.30603355e-01 -5.38139045e-01 -2.88010657e-01
7.35051870e-01 -8.08930099e-01 -1.32991767e+00 -1.11445256e-01
-7.40285933e-01 -6.01316094e-01 -5.08258402e-01 9.60229337e-01
-8.93178821e-01 4.17017162e-01 -7.92762816e-01 -1.74670315e+00
3.38322610e-01 -9.20561612e-01 8.12722683e-01 -1.91698730e-01
-2.72582561e-01 -9.73184347e-01 4.54377472e-01 -1.44978821e-01
-2.83584464e-02 -9.06316657e-03 7.75038719e-01 -3.10130864e-01
-7.24531770e-01 1.15931788e-02 7.62270540e-02 9.65490639e-02
-9.46169570e-02 7.22360089e-02 -4.68500435e-01 -1.87351674e-01
2.39919841e-01 2.14289710e-01 8.86256993e-01 5.89929521e-01
6.34641409e-01 -7.73136318e-01 -1.94225937e-01 8.42689797e-02
1.12473953e+00 2.63572305e-01 1.35476530e-01 2.25485370e-01
4.02112782e-01 8.42315853e-01 5.49727082e-01 8.55241001e-01
3.40066761e-01 3.90690923e-01 3.45301896e-01 2.20705107e-01
4.42937106e-01 -6.15046442e-01 9.02940929e-01 6.07125223e-01
-3.94184947e-01 -8.17216709e-02 -7.50197709e-01 7.03510821e-01
-2.17819953e+00 -1.31756973e+00 -2.17133775e-01 2.61166310e+00
8.67797196e-01 8.44265074e-02 5.46275198e-01 -3.20706852e-02
5.98095000e-01 -1.94804952e-01 -8.73861790e-01 -2.36028284e-01
-1.32356090e-02 1.08435079e-01 1.13568604e+00 7.48719633e-01
-8.23027790e-01 3.93319070e-01 6.72583389e+00 3.78519952e-01
-7.78188705e-01 2.98328161e-01 3.64531904e-01 -1.62846252e-01
-3.74273092e-01 5.98707438e-01 -9.06391799e-01 4.77441043e-01
1.16659534e+00 -2.97309458e-01 5.22649527e-01 7.47358561e-01
4.39558774e-01 -1.98592946e-01 -1.18262005e+00 4.52210188e-01
-6.42355859e-01 -6.40422642e-01 -3.47904682e-01 6.62382424e-01
8.34168553e-01 -2.77123481e-01 2.52316147e-01 2.11916432e-01
1.07713592e+00 -6.29645348e-01 1.00786602e+00 7.40637362e-01
6.09636962e-01 -7.06811428e-01 2.34195203e-01 5.01437306e-01
-9.91390169e-01 -5.69319785e-01 9.27834734e-02 -5.83566785e-01
3.15266609e-01 7.49542058e-01 -5.16428947e-01 4.32547808e-01
8.68049413e-02 7.16323376e-01 -2.52234414e-02 4.21482444e-01
-7.40303755e-01 1.08919919e+00 -3.11403662e-01 9.69800577e-02
3.45678240e-01 -5.51942289e-01 8.35837781e-01 8.32233489e-01
1.94952458e-01 6.92223087e-02 2.30702698e-01 8.09064388e-01
4.61701393e-01 -4.77640808e-01 -3.21014017e-01 -4.06124353e-01
1.62144870e-01 7.95852005e-01 -4.85286146e-01 -4.75425065e-01
-3.71900141e-01 7.62525201e-01 -2.12510765e-01 8.40458930e-01
-7.92205334e-01 1.92046076e-01 9.87954855e-01 -1.13109134e-01
3.47288787e-01 -5.85950434e-01 -1.76799133e-01 -1.32854939e+00
-5.11449203e-02 -8.27550948e-01 3.47245187e-01 -1.87558711e-01
-1.17691147e+00 -2.65059918e-02 2.64762878e-01 -1.04037702e+00
-8.04143190e-01 -3.62405956e-01 -2.43257284e-01 9.69425738e-01
-1.37370932e+00 -6.81294918e-01 8.25233698e-01 7.01619387e-01
3.72181870e-02 4.39309359e-01 5.77722669e-01 -4.92573678e-02
-9.49285388e-01 -4.30617817e-02 5.48365295e-01 -2.03936651e-01
1.69958979e-01 -1.35911274e+00 2.69279867e-01 1.16740131e+00
-3.87147143e-02 7.49091566e-01 8.43072891e-01 -9.04499710e-01
-1.38898945e+00 -8.56686294e-01 8.23408663e-01 -5.62283456e-01
1.15659368e+00 -2.37181410e-01 -3.87408167e-01 1.10237122e+00
-7.01941773e-02 -5.29561996e-01 3.21858376e-01 4.20183271e-01
-7.91938677e-02 1.81472301e-01 -5.24046540e-01 7.98921466e-01
1.10069132e+00 -7.28083253e-01 -2.94514388e-01 1.20124191e-01
6.20768666e-01 1.31910831e-01 -6.47534966e-01 3.59002292e-01
6.52287841e-01 -8.86584580e-01 6.99963808e-01 -1.05156946e+00
1.70973822e-01 -2.08304316e-01 -1.87619984e-01 -9.34506238e-01
-6.66110635e-01 -1.15428746e+00 -4.73126531e-01 8.65394652e-01
3.42762262e-01 -9.84282672e-01 3.67275774e-01 8.39757025e-01
1.98762760e-01 -4.02178794e-01 -9.99408424e-01 -1.28787196e+00
-1.19608149e-01 -3.70754093e-01 4.80024159e-01 8.40379417e-01
6.87135234e-02 5.02949059e-01 -6.88336492e-01 4.85710025e-01
7.62784779e-01 2.85455644e-01 3.86417985e-01 -1.34228659e+00
-1.00920117e+00 -4.13974226e-01 2.02703208e-01 -1.12629592e+00
2.60513157e-01 -3.61828536e-01 1.90282539e-02 -1.13510931e+00
4.38322902e-01 -2.80404031e-01 -5.81554294e-01 3.77082169e-01
-1.19875915e-01 -4.73657399e-01 2.88292527e-01 5.74064851e-01
-5.60550690e-01 7.89432108e-01 1.01220143e+00 3.35871696e-01
-2.31782898e-01 4.09834445e-01 -4.04598087e-01 6.17314935e-01
7.79550314e-01 -4.31387484e-01 -6.74107492e-01 1.97918974e-02
3.88513714e-01 6.74902797e-01 6.11107111e-01 -6.35083914e-01
-1.36334211e-01 -8.16988587e-01 -1.25297204e-01 -4.69826192e-01
3.44503224e-01 -4.44910467e-01 4.15747672e-01 8.35029364e-01
-6.82278752e-01 1.31857306e-01 -3.00187528e-01 1.10349810e+00
3.46427590e-01 -2.60021001e-01 3.78234744e-01 1.10258078e-02
-4.26730633e-01 3.87317449e-01 -7.04246581e-01 -1.03113294e-01
8.13823998e-01 1.99549928e-01 8.67241472e-02 -7.45499432e-01
-9.66797054e-01 7.42855966e-02 2.35232428e-01 -1.73038185e-01
1.71799697e-02 -1.03945935e+00 -3.86967272e-01 -1.98649392e-01
-3.30259293e-01 -8.02959859e-01 3.34801003e-02 1.38551962e+00
4.60204810e-01 7.89201617e-01 -6.84347749e-02 -2.05412328e-01
-8.92709374e-01 8.11687350e-01 3.37622017e-02 -6.42439246e-01
-3.59355092e-01 1.60789222e-01 3.34551573e-01 5.57751246e-02
-5.54281734e-02 -4.68289554e-01 1.91828180e-02 2.00993251e-02
1.71948925e-01 7.56390870e-01 -5.11678815e-01 -2.91145086e-01
-2.01178059e-01 -1.15747247e-02 3.15375835e-01 -1.06676114e+00
9.58890736e-01 -4.99267548e-01 2.19069868e-01 9.44887638e-01
5.89671791e-01 9.30926502e-02 -1.75193655e+00 -3.54848802e-01
1.68703705e-01 -2.86960807e-02 -1.18182059e-02 -6.54970765e-01
-5.38596153e-01 8.91997218e-01 3.02523643e-01 4.10877198e-01
8.86340320e-01 -2.48067722e-01 3.99134070e-01 2.80761749e-01
3.78214478e-01 -1.14242029e+00 -1.57688737e-01 5.82598746e-01
4.54707742e-01 -6.92474663e-01 -3.93547677e-02 -4.97454852e-02
-7.16442704e-01 6.71712756e-01 -1.25173301e-01 6.94874972e-02
6.11945152e-01 -9.31368396e-02 -5.16580224e-01 2.20167905e-01
-1.12522674e+00 -7.34455645e-01 1.27398791e-02 3.77759486e-01
1.16745457e-01 5.95567703e-01 -4.66118097e-01 6.26402497e-01
9.96505544e-02 8.20272043e-02 5.52042067e-01 8.16499352e-01
-1.69366419e-01 -1.18323624e+00 -3.43320757e-01 3.14841479e-01
-4.53101784e-01 -1.60796389e-01 -3.90536666e-01 5.92546821e-01
-3.89855802e-01 1.03311145e+00 -5.65165505e-02 -1.17420100e-01
1.12133458e-01 1.72919184e-01 3.81097496e-01 -5.36369860e-01
-1.31413162e-01 3.14128190e-01 1.19502641e-01 -6.36093140e-01
-3.52384776e-01 -1.15723622e+00 -8.30133140e-01 -2.63953805e-01
-2.05837324e-01 1.87903762e-01 3.85852039e-01 1.09868324e+00
4.65476334e-01 4.32284623e-01 8.63156796e-01 -3.79731357e-01
-1.41393268e+00 -8.55182767e-01 -7.58038640e-01 7.15629458e-02
5.29512167e-01 -8.34497690e-01 -4.56270814e-01 9.85010415e-02] | [4.219958305358887, 2.624936103820801] |
48d6979b-fdb0-40bf-9d05-40299a505b20 | maupqa-massive-automatically-created-polish | 2305.05486 | null | https://arxiv.org/abs/2305.05486v1 | https://arxiv.org/pdf/2305.05486v1.pdf | MAUPQA: Massive Automatically-created Polish Question Answering Dataset | Recently, open-domain question answering systems have begun to rely heavily on annotated datasets to train neural passage retrievers. However, manually annotating such datasets is both difficult and time-consuming, which limits their availability for less popular languages. In this work, we experiment with several methods for automatically collecting weakly labeled datasets and show how they affect the performance of the neural passage retrieval models. As a result of our work, we publish the MAUPQA dataset, consisting of nearly 400,000 question-passage pairs for Polish, as well as the HerBERT-QA neural retriever. | ['Piotr Rybak'] | 2023-05-09 | null | null | null | null | ['passage-retrieval', 'open-domain-question-answering'] | ['natural-language-processing', 'natural-language-processing'] | [-3.57095391e-01 -1.48596361e-01 2.05575526e-01 -2.66593874e-01
-1.51413870e+00 -8.88121545e-01 4.86644983e-01 5.00188470e-01
-8.77545774e-01 1.20138681e+00 3.61701190e-01 -4.11207050e-01
-5.78681864e-02 -9.94238257e-01 -6.40734017e-01 -1.63360491e-01
1.59925506e-01 9.81117785e-01 6.07644141e-01 -6.90260828e-01
3.35692093e-02 -1.23166256e-02 -1.26095724e+00 5.29175639e-01
9.99574542e-01 7.32608497e-01 1.10994741e-01 7.61323452e-01
-4.31722194e-01 1.05848265e+00 -7.83815444e-01 -9.06821311e-01
-1.04370728e-01 -5.04816890e-01 -1.48558903e+00 -8.25710893e-01
6.16982520e-01 -3.76223892e-01 -4.62348402e-01 6.75236166e-01
6.83239818e-01 4.57677305e-01 6.53795660e-01 -7.23821104e-01
-1.18940115e+00 8.59397292e-01 1.15693629e-01 4.73751813e-01
6.61536694e-01 -4.68953662e-02 1.50799143e+00 -8.24404061e-01
9.02252018e-01 6.57761037e-01 5.86706638e-01 6.64177716e-01
-8.82787585e-01 -2.93667585e-01 -4.41793293e-01 5.07832468e-01
-1.42321777e+00 -5.48383176e-01 5.61074913e-01 4.50736172e-02
1.21485591e+00 3.10618311e-01 1.81050062e-01 1.11429143e+00
-3.04199159e-01 9.81275082e-01 5.81459284e-01 -7.15999067e-01
1.75624378e-02 1.14310972e-01 5.69840550e-01 4.65315074e-01
2.15072021e-01 -2.96760619e-01 -3.00593466e-01 -4.01591182e-01
2.92785048e-01 -4.58491474e-01 -6.48079157e-01 -7.22061237e-03
-9.38921034e-01 7.77848601e-01 5.56337893e-01 6.61354601e-01
-2.08018079e-01 -2.30194494e-01 5.74562132e-01 7.66953468e-01
6.12374902e-01 1.12053251e+00 -7.03912079e-01 -3.26004535e-01
-8.67560267e-01 6.53014123e-01 1.35474908e+00 8.51989508e-01
5.78067303e-01 -7.34098911e-01 -4.55746353e-01 1.37939715e+00
4.00857851e-02 4.50245529e-01 5.17069936e-01 -7.49898374e-01
7.49356210e-01 5.10742009e-01 4.45363253e-01 -6.90852821e-01
-3.90796721e-01 2.21509440e-03 -4.42773283e-01 -8.35838556e-01
7.64191628e-01 -2.17831224e-01 -5.44999719e-01 1.53039861e+00
2.79967068e-03 -4.00385708e-01 4.63951468e-01 9.46962416e-01
1.37837374e+00 9.31701541e-01 3.44293386e-01 2.77675167e-02
1.31075227e+00 -1.23500729e+00 -7.17080951e-01 -8.07526633e-02
8.98743868e-01 -8.62957954e-01 1.51422286e+00 2.01096442e-02
-1.30924058e+00 -2.23136663e-01 -7.22147524e-01 -6.88104749e-01
-6.74458742e-01 2.60041859e-02 4.11508143e-01 2.47485042e-01
-1.06110883e+00 5.31937718e-01 -5.16329885e-01 -7.26512790e-01
1.74838275e-01 -1.64858028e-01 -2.23022074e-01 -5.41225493e-01
-1.75828087e+00 1.24124634e+00 4.40614939e-01 -1.94580033e-02
-6.20777309e-01 -8.21636260e-01 -6.38892293e-01 3.14858228e-01
1.60053447e-01 -7.44595110e-01 1.71365523e+00 -4.89641279e-01
-1.37787187e+00 1.08884990e+00 6.38107955e-02 -5.82251489e-01
1.93617702e-03 -3.63279819e-01 -6.27921581e-01 3.25923145e-01
6.10524509e-03 6.17731750e-01 1.17604695e-01 -8.96932960e-01
-4.98919666e-01 -1.52841792e-01 5.15625536e-01 2.60343850e-01
-3.85246158e-01 1.88043669e-01 -4.82680500e-01 -2.95071930e-01
-3.15489173e-01 -4.96668786e-01 -1.10673033e-01 -3.24987262e-01
-5.70195466e-02 -9.84866917e-01 3.05714279e-01 -9.72636104e-01
1.16233051e+00 -1.97226083e+00 -1.24355540e-01 -2.94742614e-01
-4.41655777e-02 4.21298593e-01 -5.34073949e-01 8.84263337e-01
4.84833032e-01 1.04354031e-01 -1.96120888e-01 -5.58216870e-02
2.01650351e-01 3.05435240e-01 -6.38812006e-01 -1.10038951e-01
2.22934604e-01 1.12134075e+00 -1.13516808e+00 -6.43097460e-01
-3.18411320e-01 1.19928867e-01 -3.93420279e-01 5.48634648e-01
-6.77424014e-01 5.68931065e-02 -5.96806288e-01 6.41198754e-01
1.26880690e-01 -2.62075365e-01 -3.34523261e-01 2.00004801e-01
2.05045551e-01 9.94087636e-01 -2.58196890e-01 2.07313275e+00
-5.37682474e-01 6.71268404e-01 -2.64962226e-01 -6.55697823e-01
9.37834442e-01 6.39561176e-01 1.62654057e-01 -1.00936830e+00
-9.59931873e-03 5.27104080e-01 -2.03730911e-01 -7.74506271e-01
1.10125148e+00 -2.34529048e-01 -2.27855995e-01 6.36527240e-01
4.69269127e-01 -4.84069228e-01 7.17491269e-01 2.84094751e-01
1.38730299e+00 1.66486934e-01 2.71924362e-02 -2.53108412e-01
5.95972359e-01 5.06808400e-01 3.16033512e-02 8.26604724e-01
-2.80764043e-01 5.25414884e-01 1.13208937e-02 -4.53218848e-01
-1.06128895e+00 -1.11559463e+00 -2.25917920e-01 1.26487732e+00
-1.33689716e-01 -3.49882603e-01 -8.09767067e-01 -6.31086528e-01
-2.03676581e-01 8.28474700e-01 -2.39625931e-01 -2.14620948e-01
-8.21932375e-01 -3.54707509e-01 1.11944544e+00 4.27442074e-01
3.73943269e-01 -1.52272546e+00 -1.83188230e-01 4.27025527e-01
-8.21114480e-01 -1.02657521e+00 -2.20830172e-01 6.93605095e-02
-7.24286020e-01 -1.03788495e+00 -1.17395163e+00 -1.11159074e+00
2.01098546e-01 -1.81052685e-02 1.99884045e+00 3.09273988e-01
1.63271725e-02 5.42684197e-01 -8.01277816e-01 -3.32514614e-01
-2.64109731e-01 8.25162053e-01 -3.04123312e-01 -6.53727412e-01
9.01473999e-01 -4.00500596e-01 -3.68045211e-01 1.78065315e-01
-9.37898159e-01 -4.12126839e-01 2.62840122e-01 8.78807843e-01
3.78586471e-01 -5.31820595e-01 1.03681886e+00 -7.56346464e-01
1.28447151e+00 -6.47473574e-01 -5.41777611e-01 8.06209505e-01
9.27118491e-03 7.60470703e-03 7.17398643e-01 -2.21713275e-01
-1.19143319e+00 -4.80216384e-01 -5.47686636e-01 -3.53555605e-02
-1.25845492e-01 7.69298792e-01 1.91844314e-01 1.93917781e-01
1.02517056e+00 -5.94694121e-03 -6.20854259e-01 -6.84306443e-01
6.94195867e-01 9.61515546e-01 5.29814005e-01 -8.88448596e-01
6.95614874e-01 -7.84057081e-02 -7.77862310e-01 -9.10178602e-01
-1.32430637e+00 -7.31301546e-01 -4.11965549e-01 -2.81311255e-02
7.86570251e-01 -7.51540542e-01 -2.53090888e-01 2.02157125e-01
-1.27858162e+00 -4.57964092e-01 -4.84182119e-01 3.93970609e-01
-4.29631174e-01 1.99725479e-01 -1.07482696e+00 -3.85978311e-01
-8.27969790e-01 -6.06888890e-01 6.91654682e-01 2.40705267e-01
-4.13204789e-01 -1.04274571e+00 7.57734537e-01 7.15879321e-01
5.96940517e-01 -4.02212501e-01 9.63708937e-01 -9.82626081e-01
-4.83089864e-01 -3.80222201e-01 -1.40124440e-01 2.57340908e-01
-2.66858339e-01 -3.60352010e-01 -9.44118798e-01 -2.05184355e-01
-1.91032022e-01 -1.25734961e+00 7.52922833e-01 -8.79301429e-02
8.66399407e-01 -1.37558401e-01 4.11860496e-02 2.93964241e-03
1.28854334e+00 -1.50834426e-01 5.52345514e-01 4.22100842e-01
2.24761635e-01 6.61737323e-01 5.10953665e-01 -9.39578712e-02
5.98692477e-01 1.96486786e-01 -1.60848975e-01 1.81564510e-01
-4.80884686e-02 -4.96792853e-01 -7.86066055e-04 1.40226591e+00
-3.35495286e-02 -5.15135050e-01 -1.18256748e+00 1.04891539e+00
-1.60202312e+00 -7.16654658e-01 8.51753503e-02 1.94601226e+00
1.38834369e+00 -2.67459631e-01 -2.51072109e-01 -2.50906110e-01
1.43853039e-01 7.47119486e-02 -2.61642128e-01 -3.21899623e-01
-1.43965885e-01 6.15929484e-01 -1.69315338e-02 3.25121164e-01
-8.81569922e-01 1.10371602e+00 6.62984085e+00 6.41446590e-01
-3.66736203e-01 1.64836437e-01 3.00513893e-01 -6.55815750e-02
-3.07507396e-01 -1.56317174e-01 -6.34220541e-01 7.32043386e-02
1.36332619e+00 -2.80872732e-01 4.39111918e-01 5.54361641e-01
-4.23689246e-01 -1.17732391e-01 -1.26837838e+00 6.29860163e-01
2.93975413e-01 -1.18261552e+00 3.33163410e-01 -6.32108867e-01
6.64766967e-01 5.59718013e-01 -5.52595854e-01 1.06595826e+00
6.04215026e-01 -9.11266625e-01 3.14415872e-01 5.47266662e-01
3.45402747e-01 -5.89428604e-01 1.00079870e+00 4.91324544e-01
-6.79790437e-01 2.46796846e-01 -8.74147713e-01 -9.78103280e-02
2.84904987e-01 3.08987588e-01 -4.96863991e-01 5.17003119e-01
9.72326398e-01 2.35669836e-01 -6.86440825e-01 1.26026630e+00
-5.12984514e-01 8.32562804e-01 -3.12547386e-01 -5.69919527e-01
3.72592658e-01 -6.35411171e-03 2.29867011e-01 1.01154387e+00
3.48292924e-02 4.24035162e-01 9.11958441e-02 7.83571661e-01
-6.17981613e-01 5.03851056e-01 -5.40780663e-01 -3.14214021e-01
4.14832950e-01 1.04354596e+00 -2.32300952e-01 -4.01666909e-01
-6.56627059e-01 8.84676814e-01 1.11201632e+00 5.03289282e-01
-4.30163532e-01 -8.68557632e-01 4.12014246e-01 -1.54143199e-01
9.03484179e-04 -1.09065332e-01 2.04004094e-01 -1.48818481e+00
2.71033645e-01 -9.46805477e-01 8.56415093e-01 -1.01020873e+00
-1.88285673e+00 8.32002401e-01 -2.46860087e-01 -7.90131807e-01
-3.60533774e-01 -4.27792758e-01 -3.72346818e-01 9.52154338e-01
-1.83430243e+00 -9.78239655e-01 7.98758566e-02 7.35582829e-01
5.48549116e-01 -7.44575858e-02 1.27465248e+00 7.21839726e-01
-1.19155772e-01 5.72190762e-01 2.82326728e-01 5.90235770e-01
1.03944957e+00 -1.04348886e+00 3.27213645e-01 4.05106038e-01
6.92552984e-01 7.98507512e-01 5.08228421e-01 -1.07866891e-01
-1.25237107e+00 -7.05793619e-01 1.42679799e+00 -8.37563097e-01
1.02199316e+00 -1.87120855e-01 -1.40370071e+00 6.14456415e-01
6.08135164e-01 -1.58730298e-01 8.69310141e-01 5.33592463e-01
-4.37452793e-01 3.02166808e-02 -8.98075342e-01 6.94478750e-01
7.42933810e-01 -8.62645566e-01 -1.35223043e+00 5.56444466e-01
1.06868100e+00 -4.09395516e-01 -1.11199069e+00 3.41157615e-01
6.27871901e-02 -3.22182685e-01 9.34397101e-01 -1.02826512e+00
4.91771251e-01 -4.12619933e-02 -1.46290153e-01 -1.32508123e+00
-5.77325821e-02 -8.18864852e-02 -8.68856460e-02 1.33273208e+00
9.51634526e-01 -4.55305964e-01 4.29607183e-01 6.37843490e-01
-1.13700673e-01 -5.51367700e-01 -1.08539951e+00 -5.89055717e-01
7.64594853e-01 -2.44197160e-01 5.89477658e-01 1.05476320e+00
4.18443918e-01 7.77692497e-01 1.67263374e-01 -7.85963386e-02
6.44856244e-02 2.07688659e-01 5.03858626e-01 -1.08567858e+00
-2.07869530e-01 -3.72505784e-01 3.74215208e-02 -1.07186484e+00
2.34548450e-01 -9.28063929e-01 3.38469565e-01 -1.68729722e+00
2.27285936e-01 -5.73567450e-01 -4.59659845e-01 4.87387896e-01
-3.52920741e-01 4.48581487e-01 -1.45526439e-01 3.17988694e-01
-1.09667420e+00 5.94221056e-01 1.15037930e+00 -3.12405139e-01
-6.65957108e-02 -3.04499865e-01 -4.67158496e-01 4.64013457e-01
1.06951094e+00 -3.91124338e-01 -3.22505623e-01 -1.10984921e+00
6.91800952e-01 3.23947072e-01 1.55073807e-01 -9.30388391e-01
2.92594492e-01 2.26912573e-01 7.35756904e-02 -5.89969397e-01
2.35212937e-01 -5.18668056e-01 -6.92448914e-01 2.03589462e-02
-6.41961217e-01 9.21013802e-02 3.93567920e-01 1.55949146e-01
-5.87878108e-01 -6.88079298e-01 3.28612208e-01 -3.48011672e-01
-5.31958938e-01 2.04811051e-01 -3.97210538e-01 9.14077342e-01
4.79920983e-01 6.13427877e-01 -6.49580717e-01 -5.00584543e-01
-3.98206949e-01 5.29401660e-01 2.42574736e-01 6.15175486e-01
3.27409953e-01 -1.14417028e+00 -9.37092304e-01 -3.09612274e-01
5.74222744e-01 -4.32017483e-02 2.41313428e-01 2.61028558e-01
-6.07670009e-01 8.91643465e-01 -1.08116586e-02 -1.57696173e-01
-8.41361880e-01 4.18781668e-01 1.85985863e-01 -7.09714830e-01
-5.45288682e-01 8.60995770e-01 -5.16796291e-01 -1.02493036e+00
2.36527234e-01 -2.60962158e-01 -6.51846588e-01 -5.49163185e-02
7.04585075e-01 1.46351695e-01 2.29746059e-01 -2.48306498e-01
8.32127258e-02 1.07779711e-01 -2.60886401e-01 -2.31986493e-01
1.17722619e+00 -3.16051431e-02 -3.97293985e-01 5.33250690e-01
1.12566936e+00 -2.28283703e-01 -2.88203359e-01 -6.10788405e-01
5.08299351e-01 2.25198660e-02 -1.17459111e-01 -1.12055922e+00
-4.65635538e-01 7.90541172e-01 3.00402492e-02 3.45398217e-01
9.15458441e-01 2.83308834e-01 1.39598727e+00 1.31264436e+00
6.19653344e-01 -9.22945976e-01 -2.65280992e-01 1.12088609e+00
8.62410724e-01 -1.14210463e+00 -3.29499274e-01 1.52420759e-01
-2.15202153e-01 8.19258451e-01 6.31099641e-01 8.96896888e-03
3.33302319e-01 -3.77227575e-01 4.39563811e-01 -3.78190577e-01
-8.00042808e-01 -2.64988422e-01 3.97864670e-01 4.37792867e-01
6.96037889e-01 -8.36637616e-02 -6.26233757e-01 7.36526549e-01
-5.12069166e-01 2.89352208e-01 3.71737659e-01 1.02063596e+00
-4.57471818e-01 -1.21770489e+00 -1.29847944e-01 4.19972628e-01
-6.31923735e-01 -5.75895488e-01 -5.08872569e-01 7.12361872e-01
-4.85012293e-01 1.05767334e+00 1.26336794e-02 2.87549466e-01
4.78178144e-01 5.03062665e-01 4.96844232e-01 -6.47913873e-01
-6.99342668e-01 -9.16916430e-01 6.22460127e-01 -2.75302768e-01
-3.36598814e-01 -2.66789585e-01 -9.54714775e-01 -1.10877812e-01
-4.34516758e-01 9.18663144e-01 2.74489939e-01 8.41258824e-01
5.91591634e-02 1.27686143e-01 5.79521693e-02 -1.56644002e-01
-6.85915470e-01 -1.41559935e+00 -3.71435076e-01 4.89706337e-01
9.20372233e-02 -2.16113627e-01 -3.05485159e-01 -6.05625622e-02] | [11.386475563049316, 7.954854488372803] |
48c4e8dc-aecd-4cef-b4c3-81280eb85b4e | advancing-volumetric-medical-image | 2306.08913 | null | https://arxiv.org/abs/2306.08913v1 | https://arxiv.org/pdf/2306.08913v1.pdf | Advancing Volumetric Medical Image Segmentation via Global-Local Masked Autoencoder | Masked autoencoder (MAE) has emerged as a promising self-supervised pretraining technique to enhance the representation learning of a neural network without human intervention. To adapt MAE onto volumetric medical images, existing methods exhibit two challenges: first, the global information crucial for understanding the clinical context of the holistic data is lacked; second, there was no guarantee of stabilizing the representations learned from the randomly masked inputs. To tackle these limitations, we proposed Global-Local Masked AutoEncoder (GL-MAE), a simple yet effective self-supervised pre-training strategy. GL-MAE reconstructs both the masked global and masked local volumes, which enables learning the essential local details as well as the global context. We further introduced global-to-global consistency and local-to-global correspondence via global-guided consistency learning to enhance and stabilize the representation learning of the masked volumes. Finetuning results on multiple datasets illustrate the superiority of our method over other state-of-the-art self-supervised algorithms, demonstrating its effectiveness on versatile volumetric medical image segmentation tasks, even when annotations are scarce. Codes and models will be released upon acceptance. | ['Hao Chen', 'Luyang Luo', 'Jia-Xin Zhuang'] | 2023-06-15 | null | null | null | null | ['volumetric-medical-image-segmentation'] | ['medical'] | [ 1.48511127e-01 4.12567258e-01 -8.12646970e-02 -5.88095844e-01
-6.62644267e-01 -8.34690854e-02 3.29821467e-01 2.63846755e-01
-3.00932229e-01 7.17631340e-01 2.83159643e-01 1.36250749e-01
-2.10223228e-01 -6.38006568e-01 -7.85454750e-01 -9.07455623e-01
-3.74295294e-01 5.79611063e-01 3.26259851e-01 -2.81190395e-01
-2.61285514e-01 4.40724641e-01 -1.47489464e+00 5.53054571e-01
9.52291250e-01 1.16662276e+00 4.92853254e-01 3.16400588e-01
-4.98196185e-02 9.24861908e-01 -3.36744308e-01 -1.02313817e-01
7.91219324e-02 -5.61502218e-01 -8.02580237e-01 1.59413546e-01
4.66723084e-01 -3.53995502e-01 -3.27358156e-01 9.62665617e-01
5.15504658e-01 -3.80528755e-02 5.75675905e-01 -7.75619566e-01
-7.35480487e-01 6.81428313e-01 -3.94987315e-01 5.12074709e-01
-3.87705922e-01 4.53548878e-02 6.13407254e-01 -8.51183295e-01
7.38977671e-01 7.63240755e-01 7.76427448e-01 7.06002057e-01
-1.30884552e+00 -4.88501757e-01 3.04441284e-02 8.55164975e-02
-1.31578255e+00 -3.43421847e-01 8.82276297e-01 -7.06357241e-01
6.56180859e-01 1.27221361e-01 6.93160534e-01 9.35698032e-01
5.82699597e-01 6.43059373e-01 1.41150415e+00 -2.03205481e-01
2.66882867e-01 1.22789457e-01 -9.85608324e-02 1.00144708e+00
-7.03101754e-02 3.56588095e-01 -5.12174547e-01 8.31335131e-03
1.02773929e+00 -5.12610897e-02 -2.13834047e-01 -5.97042680e-01
-1.25994861e+00 9.10318494e-01 8.75427902e-01 6.85690403e-01
-5.76299548e-01 -2.12956280e-01 3.05759937e-01 7.18073770e-02
6.17165387e-01 3.87071729e-01 -4.01551217e-01 4.24409002e-01
-1.26993215e+00 -1.92993999e-01 1.11024424e-01 4.70744342e-01
7.46895969e-01 3.72310579e-01 -2.76317328e-01 7.21668363e-01
1.19370244e-01 2.04452410e-01 9.58406448e-01 -5.96247733e-01
1.54128581e-01 6.88323736e-01 -3.91258657e-01 -1.09838414e+00
-7.10822463e-01 -9.34442937e-01 -1.30480349e+00 3.20809036e-01
2.03689858e-01 5.09214289e-02 -1.10017645e+00 1.81656659e+00
4.45266783e-01 6.02272600e-02 -1.74872335e-02 9.05654192e-01
1.09818554e+00 2.39512742e-01 1.06818005e-01 -3.27669412e-01
1.11870730e+00 -8.75687897e-01 -8.18710923e-01 -3.81945610e-01
2.95168430e-01 -3.89623940e-01 8.33278239e-01 3.30007635e-02
-1.20322073e+00 -7.80070722e-01 -1.10589051e+00 1.14388749e-01
-2.58498579e-01 3.27924192e-02 6.58081710e-01 2.57174134e-01
-1.12909997e+00 9.83181596e-01 -1.29475594e+00 1.57102957e-01
7.99523830e-01 4.90966588e-01 -5.44303656e-01 6.46996573e-02
-1.09790671e+00 9.58606243e-01 4.82562721e-01 2.22282037e-01
-9.01334703e-01 -9.80327010e-01 -9.18919683e-01 -7.15740100e-02
4.43043336e-02 -4.15885687e-01 8.19528043e-01 -1.07774198e+00
-1.34190583e+00 1.08111548e+00 -4.96940613e-02 -5.08079529e-01
5.59223711e-01 1.11947402e-01 -4.58508432e-01 4.96123880e-01
2.39382759e-01 7.75456369e-01 1.06447458e+00 -1.46450090e+00
-1.10274881e-01 -5.14295936e-01 -6.68806732e-01 4.13519554e-02
-2.23401144e-01 -4.16012704e-01 -6.10629246e-02 -9.87738729e-01
6.23434722e-01 -5.66926003e-01 -3.60405058e-01 6.83808699e-02
-1.21453986e-01 1.57961398e-01 6.75608337e-01 -1.02627027e+00
9.76560652e-01 -2.06623507e+00 2.40252644e-01 2.72015989e-01
6.23901069e-01 1.46506697e-01 2.76304930e-02 -1.24509424e-01
-4.82705861e-01 -2.68991798e-01 -6.62220001e-01 -4.69119489e-01
-5.77370167e-01 3.61335337e-01 -8.13864022e-02 6.27817869e-01
3.14553648e-01 1.14978707e+00 -9.93715107e-01 -6.74966812e-01
4.18938369e-01 6.18595004e-01 -4.34748977e-01 4.93110716e-01
7.30810016e-02 1.16242230e+00 -4.36176546e-02 6.10490978e-01
6.00601375e-01 -5.51048458e-01 5.80825880e-02 -3.40413690e-01
2.57630259e-01 -1.40976563e-01 -9.14014935e-01 1.65398645e+00
-3.65303993e-01 3.16542387e-01 2.13964701e-01 -1.26380825e+00
8.00341606e-01 3.45012814e-01 9.20153320e-01 -9.67705071e-01
3.26513648e-01 1.79051295e-01 -3.33498530e-02 -4.28131372e-01
1.16754122e-01 -5.50852478e-01 3.21892470e-01 3.50608468e-01
4.48846966e-01 5.35276420e-02 -2.68117100e-01 8.81961063e-02
6.92716002e-01 -6.00356981e-02 3.42830777e-01 -5.56293011e-01
3.69875014e-01 -1.09957896e-01 4.98492718e-01 7.16922283e-01
-3.61587822e-01 1.05959761e+00 6.77383840e-02 -6.32795095e-01
-9.13727283e-01 -1.06415272e+00 -4.13442641e-01 8.23344767e-01
5.34630287e-03 -1.39195085e-01 -8.97888958e-01 -6.48572922e-01
-1.34502053e-01 2.79174447e-01 -1.11355913e+00 -2.71778643e-01
-5.56339502e-01 -8.42919350e-01 1.54761106e-01 7.88434029e-01
4.88471955e-01 -1.16095185e+00 -5.71630538e-01 4.60487932e-01
-2.60148495e-01 -1.23499382e+00 -4.07895535e-01 3.69029552e-01
-1.28751850e+00 -9.74805474e-01 -8.40414166e-01 -8.29325140e-01
1.03263164e+00 -9.16496217e-02 1.12332606e+00 2.48562753e-01
-3.47372025e-01 2.25846708e-01 -2.78839804e-02 -9.73143131e-02
-6.88931584e-01 -1.61371812e-01 4.16279137e-02 1.64859965e-01
-2.22121090e-01 -8.23345304e-01 -8.56119573e-01 2.47476995e-01
-9.14276719e-01 7.01565370e-02 7.01270342e-01 1.18155277e+00
1.01181591e+00 2.25766860e-02 4.95711416e-01 -1.03987503e+00
2.73464799e-01 -3.61549854e-01 -2.23738864e-01 2.18211889e-01
-6.88420177e-01 1.47471756e-01 4.91631359e-01 -4.82433826e-01
-8.62024128e-01 2.11756781e-01 -3.01068276e-01 -5.59462905e-01
1.33214705e-02 6.22504354e-01 4.92446162e-02 -2.19152197e-01
8.37048292e-01 4.27127272e-01 4.07397687e-01 -2.84394234e-01
2.93088138e-01 2.86090702e-01 9.34791148e-01 -5.12746572e-01
5.14614820e-01 7.72490263e-01 -3.65157239e-02 -5.42979479e-01
-7.92534053e-01 -4.42477256e-01 -1.03617895e+00 -1.41471833e-01
1.05773556e+00 -9.85269725e-01 -3.76063913e-01 5.48122585e-01
-6.65594459e-01 -5.80751479e-01 -6.22934520e-01 3.16754073e-01
-5.29092491e-01 3.53352189e-01 -6.87530816e-01 -3.19032103e-01
-6.13588274e-01 -1.26360106e+00 9.53549147e-01 4.73639295e-02
-2.15263024e-01 -1.16604364e+00 -5.73759191e-02 2.87273228e-01
7.11932778e-01 4.78095144e-01 8.60680878e-01 -4.31771398e-01
-2.25295991e-01 -9.91016328e-02 -9.26565379e-03 5.89892566e-01
4.18481946e-01 -4.63256538e-01 -1.06260645e+00 -4.34166849e-01
3.80512953e-01 -4.42091197e-01 8.70597124e-01 6.79045200e-01
1.48274958e+00 -4.08090293e-01 -1.07933015e-01 7.69161344e-01
1.29739881e+00 -2.30663449e-01 6.01587415e-01 2.47708723e-01
7.23655462e-01 4.20359880e-01 1.13688067e-01 2.76519716e-01
2.99488872e-01 3.86000305e-01 3.20763499e-01 -5.55257142e-01
-4.78092611e-01 -2.49560356e-01 -1.47961214e-01 1.10769904e+00
-8.74785706e-02 5.03656924e-01 -7.90592849e-01 6.43925309e-01
-1.67563605e+00 -6.93685770e-01 2.04774439e-01 2.04455709e+00
1.14784896e+00 7.92054906e-02 -6.80842577e-03 2.41306022e-01
5.95129848e-01 2.72344112e-01 -5.10011673e-01 1.34864911e-01
-3.11334401e-01 5.62363803e-01 2.84316599e-01 4.89014983e-01
-1.32163930e+00 8.43058288e-01 6.58684874e+00 7.11226165e-01
-1.51491868e+00 6.32128596e-01 8.64283681e-01 1.90409452e-01
-8.04763809e-02 -3.65494967e-01 -2.39666298e-01 4.16690886e-01
8.08972776e-01 3.52735013e-01 1.84392542e-01 9.91915762e-01
-5.76413758e-02 -4.07485217e-02 -9.41189826e-01 9.07392263e-01
1.37618840e-01 -1.55564713e+00 -6.73610792e-02 -1.20657466e-01
1.02571285e+00 7.65026687e-03 2.00578406e-01 1.28527373e-01
6.05314272e-03 -1.15489137e+00 6.68450534e-01 5.96975684e-01
9.74821150e-01 -3.90651286e-01 7.91525245e-01 1.28130093e-01
-1.03264534e+00 5.76982610e-02 -2.34995678e-01 3.52860302e-01
-1.08540438e-01 6.87143743e-01 -7.95796931e-01 5.20086169e-01
7.60454535e-01 5.83822787e-01 -8.42648149e-01 6.32923305e-01
-2.22572625e-01 4.94264275e-01 -1.36330411e-01 5.40516675e-01
4.88041043e-02 1.70963526e-01 3.34049851e-01 1.08593631e+00
-8.49755928e-02 1.64875969e-01 1.84611470e-01 1.09236085e+00
1.78613275e-01 7.41318911e-02 -1.44642219e-01 1.56575814e-01
-3.76370959e-02 1.21873999e+00 -9.97143924e-01 -4.40853924e-01
-2.43607014e-02 9.95519519e-01 4.12754506e-01 2.33184740e-01
-5.65371692e-01 1.90849870e-01 9.92623195e-02 3.89449418e-01
3.91145319e-01 -1.21930793e-01 -6.39972627e-01 -1.09737539e+00
-1.39425388e-02 -9.15144026e-01 6.14564538e-01 -5.77400446e-01
-1.35305548e+00 1.02084053e+00 -8.23760405e-02 -1.15490437e+00
-1.41175807e-01 -4.03364360e-01 -4.55109894e-01 5.72580516e-01
-1.38076925e+00 -1.35245395e+00 -5.86613297e-01 8.50709915e-01
2.33460277e-01 -3.56821448e-01 1.00900912e+00 1.84705928e-01
-3.58568221e-01 7.08910704e-01 3.65439020e-02 3.04026872e-01
5.62763631e-01 -1.16401577e+00 -1.35359839e-01 7.58800983e-01
5.74309044e-02 6.61008120e-01 5.29045582e-01 -7.40227163e-01
-1.00383842e+00 -1.11454213e+00 4.44815308e-01 -3.61818135e-01
5.09742975e-01 -2.12357551e-01 -1.31905425e+00 7.41720200e-01
9.73215625e-02 6.75345778e-01 5.90949535e-01 -9.85742137e-02
-9.32767019e-02 -2.40073621e-01 -1.25458479e+00 2.62268305e-01
6.86712146e-01 -5.96909225e-01 -7.25422859e-01 3.41587007e-01
5.21825075e-01 -9.27204311e-01 -1.18832636e+00 7.15816915e-01
2.56919891e-01 -1.05363739e+00 9.21930671e-01 -4.24023032e-01
6.40123248e-01 -1.79280430e-01 -1.47839263e-01 -1.27015746e+00
-4.33184892e-01 -2.42952958e-01 -5.12188554e-01 9.40182030e-01
1.54429883e-01 -5.04427195e-01 8.73566091e-01 5.74457109e-01
-3.46422136e-01 -1.09825885e+00 -1.27805889e+00 -5.37595034e-01
3.41446459e-01 -2.18952641e-01 4.69518244e-01 1.11593938e+00
-1.69351876e-01 1.59717891e-02 -3.93349260e-01 2.83216864e-01
7.92549372e-01 -2.10208707e-02 2.14296654e-01 -1.00303221e+00
-4.01624173e-01 -4.78079945e-01 -4.97054636e-01 -5.88913441e-01
4.20011021e-02 -1.17964637e+00 9.88271236e-02 -1.40184724e+00
2.34074250e-01 -4.29867327e-01 -6.49915993e-01 5.88043034e-01
-4.05033678e-01 4.22107995e-01 -1.56798452e-01 4.08183008e-01
-4.55574304e-01 7.74118900e-01 1.59760070e+00 -3.10584068e-01
-3.24938089e-01 -8.23623613e-02 -7.05389977e-01 4.88980919e-01
6.38142347e-01 -5.23864686e-01 -3.73275965e-01 -2.18757242e-01
-3.59478265e-01 1.20788235e-02 4.44822371e-01 -1.23404551e+00
2.76334167e-01 1.29630685e-01 8.82543087e-01 -6.11300647e-01
1.45603478e-01 -7.50830710e-01 -3.44744287e-02 5.32139778e-01
-2.59958059e-01 -1.46805927e-01 3.56320441e-01 5.18808722e-01
-4.21902239e-01 6.62349761e-02 1.13237822e+00 -4.15882021e-01
-6.16427422e-01 4.74403232e-01 -7.79666230e-02 7.76392817e-02
8.01922441e-01 -1.83449492e-01 2.74067819e-02 -1.93264350e-01
-1.20394552e+00 1.18386438e-02 9.70837697e-02 2.87565559e-01
8.71908009e-01 -1.23325276e+00 -6.42378449e-01 5.23948908e-01
2.42573489e-02 3.80170703e-01 7.98222005e-01 1.08143306e+00
-4.83319461e-01 1.58080831e-01 -4.24035072e-01 -1.05442750e+00
-1.03562760e+00 7.79467404e-01 6.92767501e-01 -3.59333217e-01
-1.07941103e+00 8.79660368e-01 2.48831615e-01 -4.70940739e-01
2.61818916e-01 -3.42417806e-01 -1.84143081e-01 -1.76772580e-01
6.11860573e-01 -8.04128349e-02 4.09635276e-01 -7.50144482e-01
-5.02821684e-01 5.70118070e-01 -7.01388940e-02 1.31031156e-01
1.72939694e+00 -1.57600135e-01 -2.54277587e-01 4.01968956e-01
1.14772558e+00 -1.91189170e-01 -1.38760245e+00 -5.48145592e-01
-2.31110245e-01 -1.35455966e-01 4.06363904e-01 -7.44260013e-01
-1.58424425e+00 9.10791576e-01 1.09346557e+00 -2.67450094e-01
1.13188899e+00 1.23307221e-01 7.00340092e-01 -7.26992190e-02
2.37760514e-01 -9.77568150e-01 3.18061113e-01 8.40578377e-02
1.16237187e+00 -1.55268133e+00 2.57480443e-01 -2.74118900e-01
-9.12520111e-01 1.06868041e+00 5.68870246e-01 -1.57843813e-01
9.64493752e-01 1.27958626e-01 2.90869564e-01 -4.36097503e-01
-2.12260112e-01 -1.15303382e-01 6.44810140e-01 7.14000106e-01
2.86335260e-01 5.03949597e-02 1.16588920e-01 8.86856675e-01
-3.35286945e-01 -9.94447768e-02 8.22623968e-02 8.67962778e-01
-9.63703990e-02 -7.20639169e-01 -3.26004595e-01 3.97531182e-01
-3.61490399e-01 -5.57222366e-02 -4.70237322e-02 8.63034964e-01
3.44062120e-01 5.87716818e-01 7.52563998e-02 -4.04583246e-01
2.64277961e-02 -4.85642217e-02 5.55764496e-01 -5.63278377e-01
-6.07129693e-01 2.68507004e-01 -5.21493196e-01 -5.31876028e-01
-4.45349902e-01 -4.35771435e-01 -1.33862257e+00 4.80830334e-02
-1.16690502e-01 -9.10249278e-02 4.41050887e-01 1.09275317e+00
3.89432847e-01 9.14995432e-01 6.27300739e-01 -9.63157952e-01
-4.39336151e-01 -9.51901913e-01 -4.97367233e-01 7.03381121e-01
6.14436567e-01 -7.60398924e-01 -1.66331023e-01 7.46337026e-02] | [14.589520454406738, -2.1473593711853027] |
3defd28a-72bc-4486-b3f4-0a5ed2db189e | integrating-tick-level-data-and-periodical | 2306.17179 | null | https://arxiv.org/abs/2306.17179v1 | https://arxiv.org/pdf/2306.17179v1.pdf | Integrating Tick-level Data and Periodical Signal for High-frequency Market Making | We focus on the problem of market making in high-frequency trading. Market making is a critical function in financial markets that involves providing liquidity by buying and selling assets. However, the increasing complexity of financial markets and the high volume of data generated by tick-level trading makes it challenging to develop effective market making strategies. To address this challenge, we propose a deep reinforcement learning approach that fuses tick-level data with periodic prediction signals to develop a more accurate and robust market making strategy. Our results of market making strategies based on different deep reinforcement learning algorithms under the simulation scenarios and real data experiments in the cryptocurrency markets show that the proposed framework outperforms existing methods in terms of profitability and risk management. | ['Can Yang', 'Cong Zheng', 'Jiafa He'] | 2023-06-19 | null | null | null | null | ['management'] | ['miscellaneous'] | [-7.86148846e-01 -5.19831240e-01 3.36427465e-02 -2.02401385e-01
-5.60936034e-01 -7.00163841e-01 5.79853117e-01 1.02509491e-01
-4.45605457e-01 9.68074560e-01 -1.02070525e-01 -6.23182356e-01
-2.95201600e-01 -1.37795317e+00 -4.27676499e-01 -5.76628566e-01
-5.74388504e-01 7.08934665e-01 -6.90649450e-02 -4.88283962e-01
8.31980646e-01 2.59284377e-01 -1.02950895e+00 3.59039493e-02
6.84266746e-01 1.72864676e+00 -1.24495514e-01 3.02621067e-01
-3.90254945e-01 1.01684785e+00 -7.99513459e-01 -2.43582025e-01
1.11591256e+00 -3.06534708e-01 -1.05566069e-01 -4.41538058e-02
-5.57235479e-01 -9.01855469e-01 1.54817617e-02 8.15212309e-01
4.33467597e-01 4.99826372e-02 4.20932770e-01 -1.11319625e+00
-4.47028846e-01 9.20453489e-01 -7.64105618e-01 8.25544536e-01
-1.67938307e-01 2.27961794e-01 1.23818588e+00 -6.11591399e-01
1.18206076e-01 9.16965067e-01 2.75895000e-01 2.19321489e-01
-7.80653179e-01 -1.03648043e+00 1.51621193e-01 2.98893042e-02
-6.14722013e-01 2.45390728e-01 7.78871477e-01 -4.63643968e-01
1.02993071e+00 -2.60211706e-01 1.00391996e+00 3.49505395e-01
8.48372400e-01 7.02007651e-01 1.33162022e+00 -1.58067420e-01
5.86935282e-01 -2.46593043e-01 -1.71296865e-01 -3.62304986e-01
3.26353401e-01 7.83625722e-01 -3.93481374e-01 -2.67258376e-01
1.13906944e+00 1.14783682e-01 2.76599109e-01 -1.11243919e-01
-9.70232546e-01 1.57636476e+00 4.76211339e-01 1.62005231e-01
-9.81971383e-01 2.48838812e-01 4.59145010e-01 1.03518116e+00
3.45163435e-01 5.99674463e-01 -5.51365256e-01 -1.40470818e-01
-1.05839336e+00 8.78144801e-01 9.52331126e-01 4.30886865e-01
2.23015368e-01 7.41771340e-01 8.43274295e-02 3.47050220e-01
4.48945910e-01 5.06340623e-01 8.30502808e-01 -8.25094879e-01
4.96834487e-01 2.51113772e-01 5.38853884e-01 -7.31239259e-01
-4.01776463e-01 -4.97715771e-01 -5.94933331e-01 6.40878201e-01
3.26284409e-01 -6.71276927e-01 -3.45333159e-01 1.22032630e+00
7.86180422e-02 1.36016846e-01 3.24860722e-01 8.20619702e-01
-1.84511077e-02 7.98282087e-01 -1.36544704e-01 -5.83091497e-01
1.15132391e+00 -7.15516746e-01 -9.50019658e-01 2.27639049e-01
5.11618294e-02 -7.69696772e-01 2.90882856e-01 6.94090962e-01
-1.05316377e+00 -3.34955186e-01 -1.24762559e+00 7.85600603e-01
-2.23713472e-01 -5.41948140e-01 7.86681890e-01 2.29620203e-01
-7.97212780e-01 7.14801013e-01 -5.42688727e-01 5.68704903e-01
1.13701321e-01 4.88798440e-01 3.72381538e-01 9.57707107e-01
-1.80605137e+00 8.03405106e-01 6.05368435e-01 2.03689739e-01
-7.86356807e-01 -5.81098199e-01 -4.13637549e-01 1.69466347e-01
6.82796896e-01 -1.59436092e-01 1.86672521e+00 -9.41012323e-01
-1.50699162e+00 1.58869624e-01 8.97916257e-01 -1.43042529e+00
8.23776662e-01 -4.13138419e-01 -3.51822168e-01 -1.21920153e-01
-1.86864793e-01 1.24791779e-01 8.81363928e-01 -4.67282712e-01
-1.04927182e+00 -7.67482296e-02 -5.24455532e-02 2.89525419e-01
1.30422607e-01 -4.82133925e-02 6.67816579e-01 -1.18156087e+00
-1.68026477e-01 -4.22245651e-01 -2.86916077e-01 -7.81266034e-01
1.90813348e-01 -1.96235910e-01 6.07655585e-01 -8.24837148e-01
1.27024853e+00 -1.62828863e+00 -4.12100881e-01 4.15844530e-01
7.25801811e-02 9.91548821e-02 3.40139419e-01 8.07736099e-01
-2.05533430e-01 1.81226432e-01 7.05635771e-02 6.15082204e-01
2.94316530e-01 4.58841436e-02 -7.87879169e-01 -7.03989044e-02
3.46683949e-01 9.20845747e-01 -5.61463654e-01 -4.36585806e-02
2.53412426e-01 -4.04437572e-01 -3.44198912e-01 4.94949281e-01
-4.92805541e-01 1.79832518e-01 -8.78361166e-01 8.34543705e-01
7.68254220e-01 -2.18902647e-01 5.35871647e-02 3.76495749e-01
-3.07985872e-01 2.71967173e-01 -1.63003218e+00 6.60318196e-01
-2.33297467e-01 1.30377203e-01 2.34078104e-03 -1.23980355e+00
1.17344558e+00 4.42361355e-01 6.01403296e-01 -1.36260164e+00
2.24381030e-01 6.56553209e-01 3.31619591e-01 -1.11398399e-02
3.67913812e-01 -8.71435702e-01 -2.44487688e-01 1.29059780e+00
-3.44465822e-01 -3.52719307e-01 3.26168507e-01 -3.79066348e-01
7.91353285e-01 2.73631094e-03 4.08940107e-01 -3.22724730e-01
2.98192650e-01 -1.97421327e-01 1.01559305e+00 4.25577670e-01
-3.84364486e-01 -5.56225032e-02 7.04591095e-01 -9.90581155e-01
-1.05187988e+00 -9.84300733e-01 -1.09963775e-01 6.64525270e-01
-3.57548475e-01 4.99581933e-01 -5.98283112e-01 -4.36605930e-01
8.26489031e-01 6.15730941e-01 -4.50421840e-01 2.68620253e-01
-4.42643762e-01 -1.09978509e+00 -1.69138521e-01 5.95674217e-01
8.52528453e-01 -1.64531112e+00 -1.03648722e+00 9.64090645e-01
3.95552963e-01 -6.06643140e-01 -2.73820430e-01 3.59469801e-01
-8.75768304e-01 -1.20713687e+00 -8.30065012e-01 -3.08789700e-01
-5.26132435e-02 -9.46251824e-02 1.21407723e+00 -1.02935486e-01
1.31609127e-01 -1.22451752e-01 -2.94453591e-01 -8.94382596e-01
-2.84812719e-01 -1.45048186e-01 7.36866295e-02 2.29282435e-02
4.80234802e-01 -2.41951838e-01 -8.92503083e-01 1.40311658e-01
-1.02575338e+00 -3.85455668e-01 6.61568463e-01 1.11650741e+00
5.00893414e-01 6.28490448e-01 1.40639079e+00 -6.75569415e-01
1.30335951e+00 -5.00950873e-01 -1.51873016e+00 2.00790301e-01
-1.08537257e+00 2.13969037e-01 4.10837680e-01 -1.64255634e-01
-1.00353599e+00 -5.60936868e-01 2.03701258e-01 -5.20759113e-02
7.30816483e-01 7.51122534e-01 3.74950230e-01 2.03629553e-01
-9.60683897e-02 2.36689270e-01 2.48933226e-01 -3.45180959e-01
-1.29690349e-01 4.19173568e-01 9.87682715e-02 -4.98298585e-01
8.27464581e-01 -7.03110471e-02 -2.37523109e-01 -9.64081958e-02
-1.89757496e-01 4.14489545e-02 -3.08267236e-01 -1.62050977e-01
5.86538255e-01 -8.89609933e-01 -1.05974329e+00 9.04898047e-01
-5.86868167e-01 -2.11187720e-01 -1.92697436e-01 6.25016749e-01
-8.64420950e-01 -9.29812044e-02 -1.03950131e+00 -1.11725903e+00
-7.47396171e-01 -9.62303936e-01 4.11123484e-01 4.55129445e-01
-2.87803598e-02 -1.08358502e+00 4.52665538e-01 1.71300828e-01
6.56444132e-01 6.01048112e-01 8.38716149e-01 -1.04593706e+00
-9.13170040e-01 -1.61404327e-01 -5.83247468e-02 5.92826247e-01
3.10874552e-01 -8.91743526e-02 -6.82807565e-01 -2.85918325e-01
5.39254606e-01 -4.30600613e-01 6.53438687e-01 6.48026705e-01
5.15509307e-01 -1.87452257e-01 5.14665067e-01 2.58100450e-01
1.52025688e+00 1.05903697e+00 2.08032474e-01 1.15456903e+00
-1.22303061e-01 6.82283580e-01 1.28878021e+00 1.12521064e+00
9.16096643e-02 9.34034958e-02 4.83750343e-01 2.18332514e-01
9.19902980e-01 -1.35562569e-01 2.35002130e-01 5.81045747e-01
4.40765806e-02 1.50029376e-01 -8.80937278e-01 2.71992266e-01
-2.08466697e+00 -1.20348668e+00 5.10602057e-01 2.05772257e+00
8.36953700e-01 6.65245771e-01 3.77922326e-01 1.64211541e-01
5.71136177e-01 1.99947283e-01 -1.06801319e+00 -7.05909789e-01
-2.79255092e-01 4.72166896e-01 6.74757063e-01 2.11360365e-01
-1.23543870e+00 7.48109877e-01 6.82509995e+00 5.88332891e-01
-1.27002585e+00 -4.82976913e-01 1.09447896e+00 -1.73313543e-01
-3.76508743e-01 -2.32211903e-01 -6.22983575e-01 7.38775909e-01
9.80514765e-01 -8.41367841e-01 4.00526613e-01 6.72993064e-01
5.26493192e-01 -1.46124065e-01 -9.04110968e-01 7.39616036e-01
-8.24239016e-01 -1.80664361e+00 -4.05244900e-05 2.35458806e-01
6.06546700e-01 -2.54167557e-01 2.77449936e-01 4.20172453e-01
5.72312593e-01 -1.07640493e+00 7.64000475e-01 6.63416326e-01
-2.62963355e-01 -1.27921069e+00 1.24674046e+00 1.33334100e-01
-1.14118636e+00 -4.63113010e-01 -2.38422632e-01 -4.50741053e-01
1.13624439e-01 4.38533604e-01 -7.68079340e-01 5.18761694e-01
6.71622634e-01 5.34335196e-01 6.49983436e-02 8.13875020e-01
5.23570552e-02 4.62189853e-01 -3.49877477e-01 -2.24441364e-01
6.88782036e-01 -5.61412573e-01 -1.80498064e-01 4.02676523e-01
5.35027325e-01 5.94563000e-02 2.55179554e-01 9.01466608e-01
-8.59447643e-02 1.32873952e-01 -4.68579710e-01 -3.49247426e-01
4.44520444e-01 8.64573836e-01 -8.08083951e-01 -1.73360810e-01
-4.88481253e-01 2.35167444e-01 -2.08090648e-01 1.57358721e-01
-7.02030063e-01 -2.21105054e-01 4.71639365e-01 1.46661267e-01
5.07888138e-01 -3.24093968e-01 -3.43633235e-01 -8.94378543e-01
1.61418810e-01 -1.20747876e+00 7.12714970e-01 -1.55875176e-01
-1.50437939e+00 2.41144121e-01 -2.57916421e-01 -1.43702793e+00
-1.04237187e+00 -7.18718708e-01 -8.83894622e-01 9.33914721e-01
-1.97731221e+00 -4.34183061e-01 6.33775294e-01 4.89474863e-01
6.06118262e-01 -8.59601974e-01 3.79102141e-01 -1.76486686e-01
-2.85553396e-01 1.29987895e-01 5.42179108e-01 2.66242504e-01
3.07031959e-01 -1.47228026e+00 6.30003810e-01 5.56079686e-01
-1.12103604e-01 2.85486937e-01 5.30061543e-01 -8.36690903e-01
-1.18004405e+00 -4.54758316e-01 4.13324267e-01 5.25972135e-02
1.30582368e+00 4.81667370e-02 -9.35978949e-01 3.38388652e-01
6.97748542e-01 -4.72716391e-01 6.52499318e-01 -2.74790347e-01
1.03028394e-01 -5.28335333e-01 -1.12419081e+00 1.34835005e-01
-1.51968271e-01 -6.22679889e-02 -1.14484763e+00 -8.29176903e-02
5.29509246e-01 -9.96394381e-02 -1.01263273e+00 3.30123007e-01
4.99182910e-01 -9.56551194e-01 5.43579221e-01 -2.42206305e-01
5.40776849e-02 1.24866389e-01 9.60040763e-02 -1.33483255e+00
9.57028419e-02 -1.03500843e+00 -1.22507349e-01 8.42701077e-01
3.28108937e-01 -9.78350043e-01 7.79643893e-01 7.79126704e-01
6.35318041e-01 -7.25930393e-01 -1.16346991e+00 -7.09350824e-01
5.60086310e-01 -8.51762518e-02 1.26026785e+00 9.18376148e-01
-1.06982909e-01 -1.29530877e-01 -4.81136292e-01 -3.75125408e-01
8.58895063e-01 6.72666192e-01 4.28428322e-01 -1.14141190e+00
-3.50638509e-01 -7.83723831e-01 8.05381278e-04 -4.47568655e-01
-1.02239065e-01 -2.33839691e-01 -4.40910667e-01 -9.92258966e-01
-2.86449701e-01 -4.13538218e-01 -1.07723272e+00 2.37408690e-02
1.17300071e-01 -4.94536817e-01 4.48455423e-01 3.32473248e-01
-1.13525033e-01 7.41842210e-01 1.29842782e+00 -9.88688171e-02
-1.57565773e-01 3.48525554e-01 -5.66736042e-01 4.81398374e-01
1.10363960e+00 6.30756095e-02 -3.24121207e-01 -2.59287972e-02
5.05475223e-01 5.99788427e-01 -1.46075621e-01 -7.63406813e-01
-1.02832103e-02 -5.89794517e-01 5.41104317e-01 -1.04656601e+00
-2.69013703e-01 -5.85546732e-01 -4.66392040e-02 9.94969070e-01
-2.01468378e-01 8.54973733e-01 2.44981036e-01 4.59151685e-01
-7.68320680e-01 -7.91187957e-02 5.01190960e-01 -3.20890516e-01
-9.37322855e-01 2.95169532e-01 -4.15717959e-01 1.23090118e-01
1.39840746e+00 -7.34301060e-02 1.05230929e-02 -6.30089819e-01
-5.63503325e-01 9.99837697e-01 -2.15295516e-02 6.41354978e-01
8.21745634e-01 -1.53379643e+00 -8.95385981e-01 3.12651634e-01
-3.78673792e-01 -1.93918779e-01 -5.94831631e-02 2.60197580e-01
-9.27346587e-01 5.80647171e-01 -7.02486932e-01 -8.40061456e-02
-4.64174271e-01 3.85545462e-01 7.93767035e-01 -8.44939828e-01
-3.83442283e-01 4.70062137e-01 2.32630298e-01 -7.54065216e-02
5.33805527e-02 -6.45711064e-01 -3.19161594e-01 6.52069330e-01
7.35200644e-01 2.72007376e-01 1.11842729e-01 1.49618834e-03
-2.37652324e-02 4.27812487e-01 -3.53317142e-01 -4.85480458e-01
1.62813473e+00 2.82612830e-01 1.37188151e-01 5.25105059e-01
4.36746031e-01 -5.00608385e-01 -1.58295667e+00 -1.93306193e-01
8.84755254e-01 -4.30739820e-01 1.66801345e-02 -7.43598878e-01
-1.28392184e+00 8.02359641e-01 5.73349655e-01 7.62098610e-01
1.09377778e+00 -8.15352321e-01 1.13543403e+00 3.80542487e-01
6.62874103e-01 -1.97294438e+00 4.89549488e-02 4.90143776e-01
1.03592372e+00 -1.35216689e+00 -1.74109161e-01 5.97071707e-01
-7.52625823e-01 1.27209318e+00 4.37778503e-01 -8.46668422e-01
1.18324864e+00 4.53948021e-01 6.59682155e-01 7.32740387e-02
-1.07144928e+00 1.85952023e-01 -2.33802348e-01 2.03292847e-01
3.96796495e-01 2.92067200e-01 -6.53842866e-01 8.58499587e-01
-3.80067259e-01 1.86192364e-01 5.78378618e-01 1.20698500e+00
-6.74739540e-01 -1.31697905e+00 -6.47476435e-01 5.96692264e-01
-8.87979209e-01 5.76767623e-02 -2.63583779e-01 1.18424964e+00
-3.90255332e-01 6.99377656e-01 3.29821527e-01 -1.22828960e-01
3.08315843e-01 7.83051178e-03 -3.01465951e-02 -2.66978174e-01
-1.04958951e+00 7.61986911e-01 -2.60337293e-01 -4.27791089e-01
-3.13830703e-01 -7.45888650e-01 -1.71458292e+00 -4.52037305e-01
-4.91393432e-02 4.86605555e-01 2.87261009e-01 7.93013275e-01
1.88681222e-02 4.91636276e-01 1.32398021e+00 -4.53495115e-01
-1.60925853e+00 -8.00694525e-01 -1.39213789e+00 3.04035991e-01
5.41248798e-01 -9.32522655e-01 -2.72755295e-01 -5.32848179e-01] | [4.417545318603516, 3.926316022872925] |
8d6bebef-7994-439d-a85a-a5898c381587 | variational-quantum-cloning-improving | 2012.11424 | null | https://arxiv.org/abs/2012.11424v1 | https://arxiv.org/pdf/2012.11424v1.pdf | Variational Quantum Cloning: Improving Practicality for Quantum Cryptanalysis | Cryptanalysis on standard quantum cryptographic systems generally involves finding optimal adversarial attack strategies on the underlying protocols. The core principle of modelling quantum attacks in many cases reduces to the adversary's ability to clone unknown quantum states which facilitates the extraction of some meaningful secret information. Explicit optimal attack strategies typically require high computational resources due to large circuit depths or, in many cases, are unknown. In this work, we propose variational quantum cloning (VQC), a quantum machine learning based cryptanalysis algorithm which allows an adversary to obtain optimal (approximate) cloning strategies with short depth quantum circuits, trained using hybrid classical-quantum techniques. The algorithm contains operationally meaningful cost functions with theoretical guarantees, quantum circuit structure learning and gradient descent based optimisation. Our approach enables the end-to-end discovery of hardware efficient quantum circuits to clone specific families of quantum states, which in turn leads to an improvement in cloning fidelites when implemented on quantum hardware: the Rigetti Aspen chip. Finally, we connect these results to quantum cryptographic primitives, in particular quantum coin flipping. We derive attacks on two protocols as examples, based on quantum cloning and facilitated by VQC. As a result, our algorithm can improve near term attacks on these protocols, using approximate quantum cloning as a resource. | ['Niraj Kumar', 'Elham Kashefi', 'Mina Doosti', 'Brian Coyle'] | 2020-12-21 | null | null | null | null | ['cryptanalysis'] | ['miscellaneous'] | [ 4.89986241e-01 1.03685394e-01 9.84875858e-02 4.33385149e-02
-1.34555721e+00 -9.83290553e-01 4.28842396e-01 -6.55501708e-02
-5.25911987e-01 4.93439913e-01 -3.63298237e-01 -9.42830622e-01
8.50582495e-03 -1.31111348e+00 -1.00131631e+00 -1.07018089e+00
-1.42731756e-01 4.44011986e-01 -6.82345852e-02 -6.66842103e-01
8.96408021e-01 3.40636402e-01 -1.17444849e+00 5.75186312e-01
4.49237317e-01 3.90763193e-01 -3.96898359e-01 1.26440537e+00
-3.66316773e-02 5.83992302e-01 -3.67917240e-01 -8.13226581e-01
6.01950943e-01 -7.49647617e-01 -1.17315125e+00 -6.37367547e-01
9.90440845e-02 -2.11020052e-01 -7.22950578e-01 1.41705275e+00
4.64857519e-01 -4.19012457e-01 4.67059255e-01 -1.09297681e+00
-7.45252147e-02 1.14334166e+00 2.84049481e-01 -1.91999912e-01
3.94595325e-01 3.32350433e-01 1.43499565e+00 -2.06835885e-02
6.02696061e-01 1.10888541e+00 7.22384751e-01 7.47663319e-01
-1.82700264e+00 -9.28350031e-01 -1.05147374e+00 5.21938562e-01
-1.79119444e+00 -3.41885954e-01 4.22948718e-01 2.51185223e-02
1.26982093e+00 1.93889380e-01 3.60203773e-01 7.42796838e-01
7.60185182e-01 4.61286277e-01 1.02942419e+00 -7.52426922e-01
5.96196532e-01 9.81161594e-02 -4.03478071e-02 5.21128237e-01
1.69868618e-01 9.60148811e-01 -1.35545462e-01 -7.20215976e-01
9.30655971e-02 -2.35473156e-01 -6.06762469e-02 -3.20594311e-01
-1.04288638e+00 1.19501638e+00 3.33330601e-01 4.27779734e-01
9.50498655e-02 7.06048489e-01 4.93437350e-01 1.11343634e+00
-6.41297698e-01 8.04859042e-01 -5.10247588e-01 -1.84296265e-01
-4.82617766e-01 4.84145999e-01 1.42593753e+00 7.67369628e-01
1.28483164e+00 -5.30670047e-01 2.13039488e-01 -3.46497387e-01
-4.90506105e-02 6.99188590e-01 -1.68224648e-01 -9.46754217e-01
1.49212927e-01 1.37830973e-01 -5.86505607e-03 -7.44169354e-02
-5.15266478e-01 7.08364969e-05 -8.08558822e-01 2.62709558e-01
5.30618012e-01 -6.12317324e-02 -4.44512010e-01 1.70872295e+00
2.03250825e-01 3.38171780e-01 7.27458060e-01 5.22606730e-01
7.04662036e-03 8.08891237e-01 -2.31413573e-01 1.02245443e-01
1.47837031e+00 -3.58869106e-01 -3.30133229e-01 1.90710843e-01
1.72085917e+00 -8.33656251e-01 4.94211465e-01 5.49874902e-01
-8.65503371e-01 -1.80038840e-01 -1.37240815e+00 1.36644125e-01
-2.27213308e-01 -6.68434799e-01 9.78174806e-01 1.50525081e+00
-9.23198104e-01 1.36894059e+00 -6.39356792e-01 1.81425080e-01
2.65503764e-01 1.20875478e+00 -3.99306476e-01 -2.49003977e-01
-1.44625628e+00 6.85372770e-01 6.90739632e-01 -1.70990616e-01
-6.98479593e-01 -5.99152029e-01 -5.27820170e-01 1.18260689e-01
2.64040887e-01 -9.29502487e-01 1.40616202e+00 -8.31943154e-01
-1.99434650e+00 8.18467557e-01 2.12676898e-01 -8.18411887e-01
1.01017766e-02 6.16962850e-01 -5.93200661e-02 -2.26319861e-02
-3.72707605e-01 2.76014507e-01 5.84933460e-01 -4.72866178e-01
-6.19248688e-01 -4.04071897e-01 5.50567865e-01 -2.39270836e-01
4.99014884e-01 -1.69740960e-01 3.45682681e-01 2.32277080e-01
2.48120666e-01 -1.56507981e+00 -5.65307498e-01 -6.07742846e-01
-4.06699091e-01 -6.55184910e-02 1.50156796e-01 -1.51047166e-02
7.46357501e-01 -2.02051544e+00 5.88754117e-01 6.79108858e-01
7.76734129e-02 1.36280298e-01 -1.50360748e-01 7.57623076e-01
2.43504822e-01 -2.85008568e-02 -4.07408357e-01 1.01321348e-04
4.41697329e-01 2.81416982e-01 -2.03445166e-01 1.01139772e+00
8.78203437e-02 1.05798244e+00 -7.99883068e-01 -6.99072238e-03
7.00302720e-02 1.19085655e-01 -1.12418795e+00 2.56505180e-02
-1.37169957e-01 5.92729092e-01 -4.84065562e-01 4.85978067e-01
1.16344786e+00 7.30704367e-02 4.38336015e-01 1.14413410e-01
-1.60119459e-01 6.13586009e-01 -1.48105490e+00 2.20093918e+00
-5.22125602e-01 1.72023907e-01 1.42620549e-01 -1.09909976e+00
3.41819078e-01 3.17006379e-01 2.29713619e-02 -7.94906914e-01
7.72437215e-01 7.79763520e-01 4.65236872e-01 -2.12973636e-02
3.57959628e-01 -6.63355470e-01 -6.24559343e-01 8.85469794e-01
3.83890197e-02 -7.07481325e-01 -1.44942343e-01 2.27812797e-01
1.46101189e+00 -1.95017263e-01 1.58408731e-01 -2.68017471e-01
1.09611094e+00 9.06639695e-02 2.27982089e-01 1.08140552e+00
-1.77981302e-01 3.47652882e-02 6.77993238e-01 -3.17411065e-01
-1.50463271e+00 -1.03022742e+00 -2.81902432e-01 8.65046918e-01
3.94598752e-01 -5.11030495e-01 -9.18549716e-01 -2.24294156e-01
-1.66447774e-01 6.70000315e-01 -4.34790105e-01 -7.25473940e-01
-6.51703596e-01 -8.42362583e-01 1.27353525e+00 -3.30060460e-02
9.65265632e-02 -8.71146619e-01 -9.08631086e-02 4.11804765e-01
4.07620907e-01 -8.66247356e-01 -3.00800856e-02 6.76065981e-01
-8.01428199e-01 -1.11059904e+00 1.20055847e-01 -6.21128082e-01
2.93638647e-01 -1.19199911e-02 8.52025628e-01 3.35669249e-01
-4.96470183e-01 2.46065836e-02 -2.02481672e-01 -9.84021695e-04
-1.57472527e+00 2.28459507e-01 6.70762286e-02 -2.65263915e-01
6.12580478e-01 -5.39167345e-01 -5.16173840e-01 -8.72473568e-02
-1.08132470e+00 -3.04544657e-01 1.04907036e+00 1.21564591e+00
4.33967382e-01 9.35631767e-02 -2.57905483e-01 -1.08173490e+00
-2.49237027e-02 -1.51009306e-01 -1.29932761e+00 1.92354947e-01
-4.60545957e-01 9.97019947e-01 9.55388188e-01 -2.81250238e-01
-4.02324826e-01 2.24555805e-01 -4.28169698e-01 1.45993114e-01
5.30071110e-02 -1.04860313e-01 -1.98190033e-01 -9.39191878e-01
1.06034005e+00 4.72423822e-01 5.40362718e-03 1.40013456e-01
7.40407944e-01 6.90562725e-01 2.89903641e-01 -8.89822900e-01
1.19964647e+00 4.86750245e-01 8.07263792e-01 -5.92988729e-01
-3.28478396e-01 -4.72442001e-01 -7.78516591e-01 7.86704719e-01
5.41270614e-01 -7.63147175e-01 -1.55891204e+00 4.46612477e-01
-1.11708200e+00 -1.97252810e-01 2.45776493e-02 5.30412257e-01
-7.32825637e-01 6.53768003e-01 -8.92801940e-01 -9.70488667e-01
-5.35537362e-01 -1.59600520e+00 1.33471000e+00 3.62941506e-03
1.02574565e-01 -8.08714867e-01 4.00386125e-01 1.94465339e-01
2.34105185e-01 -2.08265871e-01 9.58140492e-01 -6.12964451e-01
-1.36584175e+00 -5.28542340e-01 1.31432731e-02 2.55531460e-01
-5.48675299e-01 -5.04334569e-01 -1.08497953e+00 -6.66354656e-01
-7.73483142e-02 -4.46851730e-01 6.47495687e-01 -1.15025111e-01
6.14404023e-01 1.01865314e-01 -2.10083961e-01 9.28948760e-01
1.67998767e+00 7.46339858e-02 1.11802280e+00 4.03019398e-01
4.10673350e-01 1.71739161e-01 1.26436874e-01 2.72094667e-01
-5.03126420e-02 6.52445495e-01 2.44497925e-01 5.38692594e-01
6.08604848e-01 4.81874980e-02 4.81348485e-01 7.29644537e-01
9.95953977e-02 4.71483141e-01 -5.51586509e-01 -8.59118104e-02
-1.20758557e+00 -1.11405075e+00 -2.13106766e-01 2.50309277e+00
8.51766229e-01 1.45560235e-01 -1.72336951e-01 4.74476904e-01
4.20986533e-01 -5.09137094e-01 -1.50622919e-01 -9.63380396e-01
-1.44795403e-01 1.34868574e+00 1.21534324e+00 7.42727280e-01
-1.06473982e+00 9.61474478e-01 4.97643328e+00 9.64750409e-01
-8.98502469e-01 3.86207253e-01 1.79830547e-02 3.71118963e-01
-4.50330824e-01 8.91484857e-01 -6.95665896e-01 2.92067081e-02
1.59216094e+00 -2.41962060e-01 1.19345725e+00 6.41925335e-01
-5.99788785e-01 8.09693858e-02 -1.40854406e+00 9.50349450e-01
-5.75551808e-01 -1.62200284e+00 -4.25788134e-01 3.69113117e-01
7.23567545e-01 4.18702587e-02 -8.55522603e-02 7.72476315e-01
7.81839713e-02 -1.01565051e+00 5.83378792e-01 -1.40757218e-01
7.66577482e-01 -1.15724266e+00 1.06715977e+00 3.79222661e-01
-9.34251547e-01 -2.61837780e-01 -8.32747877e-01 -2.35463545e-01
-2.99546003e-01 -2.88269937e-01 -7.84645140e-01 5.51135600e-01
3.20373513e-02 -1.71004862e-01 4.93700504e-02 7.07361877e-01
-2.47351781e-01 4.37242508e-01 -4.33346570e-01 -3.95626634e-01
6.11412466e-01 -6.36755466e-01 2.91254938e-01 1.02792442e+00
3.21783036e-01 6.18986785e-01 -2.76561659e-02 1.08223903e+00
-2.17100576e-01 1.67004034e-01 -5.87991714e-01 -1.26131043e-01
4.57941145e-01 9.59977627e-01 -4.89119709e-01 -1.27110347e-01
-4.83140051e-01 1.24311054e+00 -6.12253882e-02 -2.24759474e-01
-6.42049551e-01 -7.52622843e-01 8.11041474e-01 -5.05182028e-01
2.66118109e-01 4.69216444e-02 -3.03307772e-01 -1.26796472e+00
-5.40013194e-01 -1.10901308e+00 2.01833099e-01 2.80913979e-01
-8.41325045e-01 4.93073575e-02 -7.88658857e-01 -1.02678323e+00
-2.98959583e-01 -8.38398635e-01 -6.72135830e-01 1.00491166e+00
-1.35297322e+00 -9.12920177e-01 5.56342959e-01 4.74412769e-01
-3.64301503e-01 -3.45487185e-02 1.36275589e+00 2.56490380e-01
-2.61549145e-01 1.11530423e+00 7.05134690e-01 6.14443980e-02
2.85114974e-01 -1.35809445e+00 8.38719785e-01 6.19539440e-01
3.77797812e-01 7.21619368e-01 8.16931546e-01 -2.50423625e-02
-2.76941442e+00 -5.34760714e-01 6.88820243e-01 -5.92398405e-01
1.03869438e+00 -5.87776899e-01 -7.74879217e-01 4.75735664e-01
-2.06029594e-01 -2.65334427e-01 8.50136459e-01 1.35292083e-01
-5.53151667e-01 2.08416179e-01 -1.18842506e+00 3.78415793e-01
6.85562789e-01 -1.24133515e+00 -3.80653232e-01 6.04594529e-01
5.55596471e-01 -5.59464931e-01 -8.92009556e-01 -2.02601448e-01
6.85929716e-01 -8.68861318e-01 1.06240940e+00 -7.19854593e-01
-1.75187081e-01 -3.26589018e-01 -4.25926566e-01 -7.58283377e-01
-7.73181543e-02 -1.35574281e+00 1.97889417e-01 3.46265376e-01
4.89674389e-01 -6.44934773e-01 1.12122321e+00 3.39381993e-01
1.59284934e-01 -7.15822354e-02 -1.47614861e+00 -7.89947987e-01
8.87690902e-01 -7.68361866e-01 7.96767056e-01 7.50363946e-01
3.36726189e-01 3.37649018e-01 -2.18671009e-01 7.41967201e-01
9.49954689e-01 3.86716366e-01 1.16761184e+00 -9.33429062e-01
-9.02502775e-01 -7.31972516e-01 -1.09643745e+00 -7.65165269e-01
4.43560690e-01 -1.29970109e+00 -4.51439321e-02 -3.51799637e-01
2.01511949e-01 -5.64232349e-01 -2.64801681e-01 -1.14623718e-01
1.17553756e-01 2.06214577e-01 -1.10809185e-01 2.31022965e-02
-4.20652211e-01 2.26637945e-01 8.19733381e-01 -5.13276696e-01
-2.02236295e-01 4.31732118e-01 -3.78493786e-01 3.79298590e-02
6.28467619e-01 -9.28084075e-01 1.66880190e-01 2.63385981e-01
7.23221362e-01 2.46903121e-01 1.95197761e-01 -1.05154955e+00
3.35491896e-01 1.20125532e-01 -5.33893287e-01 5.51412031e-02
1.93212032e-01 -5.05386174e-01 3.72633964e-01 1.44609904e+00
4.33523059e-02 -4.61710602e-01 9.80137736e-02 3.86279255e-01
1.16638735e-01 -5.78231871e-01 1.06661427e+00 -2.62786838e-04
-6.38269067e-01 2.46765435e-01 -3.24980080e-01 -1.96881846e-01
7.51824021e-01 6.42150119e-02 -1.67926654e-01 -7.14809895e-02
-5.47352254e-01 -1.13922983e-01 8.38802576e-01 -2.59852856e-01
1.36080816e-01 -9.80718493e-01 -8.12890828e-01 3.48786235e-01
3.12066644e-01 -4.68501747e-01 4.55994725e-01 8.98341417e-01
-1.21345878e+00 7.48053849e-01 -5.36958724e-02 -5.77611625e-01
-1.12563503e+00 8.21057558e-01 6.31678522e-01 -2.26541176e-01
-3.95228356e-01 1.09235692e+00 4.74578105e-02 -7.18565404e-01
-1.74261123e-01 -3.31855088e-01 6.94308341e-01 -6.14499867e-01
7.00821102e-01 1.48037687e-01 2.16214508e-01 -6.37842238e-01
-2.12149903e-01 6.48889542e-01 6.11335821e-02 -1.49096966e-01
7.36082077e-01 2.66280890e-01 -4.14339542e-01 -1.50268510e-01
1.72995996e+00 -1.19025387e-01 -5.37131488e-01 -4.25177395e-01
1.98778495e-01 -2.27295548e-01 -1.61915496e-02 -1.40439972e-01
-5.11342347e-01 1.03141999e+00 9.18771088e-01 1.36391278e-02
8.93206298e-01 -9.96301696e-02 1.12223911e+00 1.44071925e+00
1.18178856e+00 -7.53768981e-01 -7.25836873e-01 5.67172885e-01
-2.84883171e-01 -1.11307621e+00 -1.84538700e-02 -2.08492965e-01
3.45112458e-02 1.48189151e+00 -3.51543009e-01 -2.15527773e-01
3.49299163e-01 3.95908922e-01 -1.67945568e-02 1.90644525e-02
-4.97063488e-01 -2.37100527e-01 -8.78798813e-02 2.50675529e-01
-6.46840408e-02 4.55912322e-01 -2.92652458e-01 4.18970793e-01
-4.20917392e-01 -3.43594372e-01 7.79905975e-01 1.16350293e+00
-3.83954108e-01 -2.12882018e+00 -7.52248049e-01 -1.25757948e-01
-7.06952631e-01 -3.97632867e-01 -5.99587820e-02 6.49968445e-01
2.00078115e-01 6.99036837e-01 -3.26056749e-01 -6.62016988e-01
2.68404990e-01 1.37428969e-01 1.13294661e+00 -4.12040770e-01
-7.94152141e-01 -2.98187971e-01 -1.90816328e-01 -6.87640250e-01
-3.88333127e-02 -6.57394946e-01 -1.51548445e+00 -1.14627993e+00
-9.16957021e-01 4.67810839e-01 9.36097324e-01 1.09975433e+00
1.86059698e-01 3.82283852e-02 1.03583932e+00 -8.30658674e-01
-1.29869604e+00 -5.21394849e-01 -9.42425549e-01 2.07387075e-01
8.62473249e-02 -1.03494100e-01 -5.31931102e-01 -5.01225650e-01] | [5.606878280639648, 4.98549747467041] |
aa87bb15-31c0-44d1-88e8-cc9e3551fe5b | unsupervised-speech-representation-learning | 1901.08810 | null | https://arxiv.org/abs/1901.08810v2 | https://arxiv.org/pdf/1901.08810v2.pdf | Unsupervised speech representation learning using WaveNet autoencoders | We consider the task of unsupervised extraction of meaningful latent representations of speech by applying autoencoding neural networks to speech waveforms. The goal is to learn a representation able to capture high level semantic content from the signal, e.g.\ phoneme identities, while being invariant to confounding low level details in the signal such as the underlying pitch contour or background noise. Since the learned representation is tuned to contain only phonetic content, we resort to using a high capacity WaveNet decoder to infer information discarded by the encoder from previous samples. Moreover, the behavior of autoencoder models depends on the kind of constraint that is applied to the latent representation. We compare three variants: a simple dimensionality reduction bottleneck, a Gaussian Variational Autoencoder (VAE), and a discrete Vector Quantized VAE (VQ-VAE). We analyze the quality of learned representations in terms of speaker independence, the ability to predict phonetic content, and the ability to accurately reconstruct individual spectrogram frames. Moreover, for discrete encodings extracted using the VQ-VAE, we measure the ease of mapping them to phonemes. We introduce a regularization scheme that forces the representations to focus on the phonetic content of the utterance and report performance comparable with the top entries in the ZeroSpeech 2017 unsupervised acoustic unit discovery task. | ['Aäron van den Oord', 'Samy Bengio', 'Ron J. Weiss', 'Jan Chorowski'] | 2019-01-25 | null | null | null | null | ['acoustic-unit-discovery'] | ['speech'] | [ 2.70036906e-01 5.09964406e-01 1.21132873e-01 -2.90681839e-01
-9.03881788e-01 -6.59210861e-01 5.44779599e-01 7.76574761e-02
-2.68002331e-01 4.46724713e-01 6.72791183e-01 -5.15105687e-02
-9.17186029e-03 -6.73838317e-01 -9.95931923e-01 -8.94461274e-01
-4.42249402e-02 2.87075192e-01 -1.73654243e-01 6.42621517e-02
-1.84386119e-01 2.84516990e-01 -1.71678650e+00 3.29285324e-01
3.74907553e-01 1.00754690e+00 2.22681344e-01 7.68207312e-01
-1.31531814e-02 4.83008057e-01 -5.88287115e-01 -4.81279977e-02
9.62521732e-02 -5.78382790e-01 -5.77223182e-01 1.29788846e-01
7.34902248e-02 -9.47189778e-02 -3.33446175e-01 1.04417551e+00
3.15959573e-01 2.62481630e-01 1.04352558e+00 -7.09356904e-01
-3.78902912e-01 8.64819765e-01 3.24870080e-01 1.36058882e-01
3.52845192e-02 -1.15659229e-01 1.17841053e+00 -1.00728238e+00
6.59596801e-01 1.06606996e+00 4.84442115e-01 4.40985680e-01
-1.55812144e+00 -3.20558578e-01 -3.92926037e-02 1.24206297e-01
-1.35219157e+00 -1.20244801e+00 9.17541504e-01 -7.26609766e-01
1.01128280e+00 9.07725468e-02 1.47766232e-01 1.35518503e+00
-3.63904028e-03 3.07894111e-01 7.88910151e-01 -5.82031786e-01
6.28746510e-01 2.81677663e-01 4.97626141e-04 3.93148571e-01
-2.60623068e-01 1.41885489e-01 -6.28044844e-01 -1.67596951e-01
5.40215731e-01 -5.20256042e-01 -4.44158405e-01 -3.93455774e-01
-8.58224332e-01 9.16440427e-01 -2.72089057e-02 4.57012653e-01
-6.86402678e-01 9.06531364e-02 4.86696184e-01 3.38100135e-01
3.68189454e-01 3.17288339e-01 -5.70140183e-01 -2.29012355e-01
-9.73544180e-01 -1.74361125e-01 7.50179350e-01 5.81629813e-01
7.37746060e-01 7.15035677e-01 1.82058662e-02 8.85269701e-01
9.54527482e-02 3.04053575e-01 7.97544539e-01 -1.05636156e+00
2.26277545e-01 -1.71751767e-01 -1.60338372e-01 -6.06663585e-01
-9.62189287e-02 -4.62597489e-01 -5.83701074e-01 5.26285097e-02
9.78427380e-02 -2.28698909e-01 -8.13157678e-01 2.19528127e+00
-4.28344607e-02 2.39541918e-01 3.81484419e-01 6.14245296e-01
4.98125464e-01 1.05400753e+00 -1.71129167e-01 -4.63965893e-01
1.18656802e+00 -4.37424421e-01 -8.42198014e-01 -1.90347418e-01
3.88184041e-01 -4.37584907e-01 9.77089763e-01 3.99888545e-01
-1.21371174e+00 -8.16143394e-01 -1.03020310e+00 -1.28859119e-03
-3.63131285e-01 2.06625521e-01 5.04440777e-02 5.39850235e-01
-8.98538411e-01 7.48885214e-01 -7.48200893e-01 -1.70635581e-02
-4.05116454e-02 2.55370915e-01 -4.79454637e-01 3.69317114e-01
-1.20378220e+00 6.87160909e-01 6.59964859e-01 -2.70636261e-01
-1.19423938e+00 -7.20537364e-01 -9.66267645e-01 5.84811747e-01
1.08846696e-02 -2.10348085e-01 1.01968360e+00 -9.90023196e-01
-1.72206080e+00 4.51547354e-01 -4.00356293e-01 -6.20770693e-01
-6.06990233e-02 1.23011637e-02 -5.67501247e-01 3.65692139e-01
-1.56322196e-01 4.10079747e-01 1.32067978e+00 -9.78596389e-01
-2.91384339e-01 -5.42003079e-04 -4.03785527e-01 2.28727572e-02
-4.44675595e-01 -3.14729482e-01 -1.16945274e-01 -7.78468847e-01
1.79461926e-01 -8.30224037e-01 2.32896864e-01 -4.69440132e-01
-2.46999428e-01 -1.91164151e-01 6.01539195e-01 -1.00125635e+00
1.08898401e+00 -2.73873758e+00 5.66539407e-01 1.85469568e-01
-1.36599615e-01 -3.50528322e-02 -2.41650596e-01 4.88675416e-01
-3.00285071e-01 -6.13131886e-03 -3.33023787e-01 -5.53363502e-01
1.71678126e-01 3.80348772e-01 -8.17434311e-01 4.33811307e-01
4.16410953e-01 6.47842109e-01 -5.49725175e-01 -7.80296624e-02
1.58093482e-01 8.60681415e-01 -8.15302253e-01 2.42261648e-01
-2.53920615e-01 4.84846383e-01 1.03909664e-01 -3.13204750e-02
1.39981359e-01 1.18193999e-01 1.81011647e-01 -2.84070015e-01
-9.13461596e-02 1.05772269e+00 -1.05225205e+00 1.74755669e+00
-6.69906199e-01 7.68534482e-01 2.74940103e-01 -1.12656331e+00
8.20086420e-01 8.72124434e-01 4.94748652e-01 -4.40881401e-01
8.69925469e-02 2.03786552e-01 9.08775181e-02 -1.39558434e-01
3.07451725e-01 -3.91669452e-01 3.49446535e-02 2.84417331e-01
7.01887369e-01 -4.22843806e-02 -2.31677353e-01 -4.55133058e-02
8.99101317e-01 -1.79202378e-01 8.83573070e-02 -2.55258113e-01
3.22931349e-01 -5.96391857e-01 4.27055210e-01 5.40297925e-01
1.64485395e-01 6.12680018e-01 5.15508890e-01 1.88142508e-01
-1.17565715e+00 -1.38316023e+00 -3.92291963e-01 1.06345880e+00
-5.24727762e-01 -5.31706333e-01 -7.76490688e-01 -2.97704518e-01
-2.25675032e-01 1.07263553e+00 -5.67613423e-01 -4.86704379e-01
-5.87155879e-01 -3.01459193e-01 5.89938581e-01 3.36088598e-01
-3.05911154e-01 -1.11684895e+00 -3.27545822e-01 4.82715994e-01
-3.23185444e-01 -1.28432441e+00 -2.82426089e-01 6.73158765e-01
-6.54640079e-01 -4.54577655e-01 -3.69214654e-01 -6.42755449e-01
2.47016206e-01 -4.96723682e-01 9.05091941e-01 -6.16924763e-01
7.85805658e-03 3.97561312e-01 -4.33108658e-01 -2.08863661e-01
-7.28445470e-01 -1.69391111e-02 4.16574121e-01 3.59082967e-01
1.76256791e-01 -8.98777604e-01 -1.65937841e-01 -1.42381310e-01
-7.98870444e-01 -2.81193584e-01 3.63006711e-01 7.44497061e-01
7.45817006e-01 5.03315508e-01 5.38322270e-01 -5.50129652e-01
4.63350505e-01 -4.59576309e-01 -4.31655258e-01 -2.17165247e-01
-3.49807233e-01 4.91588831e-01 7.49835610e-01 -6.14612460e-01
-7.98586905e-01 1.14835314e-01 -4.90110964e-01 -6.79412127e-01
-1.77634910e-01 3.51598173e-01 -2.66310871e-01 4.32796955e-01
4.93098110e-01 5.33094406e-01 -1.31628752e-01 -7.20063686e-01
6.14040732e-01 6.71723783e-01 7.09596038e-01 -4.30953473e-01
5.51128745e-01 1.37665644e-01 -4.49805826e-01 -1.34596884e+00
-5.52944660e-01 -2.25380555e-01 -5.57126403e-01 1.61740258e-01
1.14210975e+00 -8.60143244e-01 -4.87202138e-01 -1.22370478e-02
-1.18073440e+00 -1.30830586e-01 -7.65081942e-01 7.45019495e-01
-7.78981745e-01 1.53823525e-01 -7.59865761e-01 -9.22514677e-01
-1.11645065e-01 -1.13129687e+00 8.84006083e-01 -4.39084202e-01
-3.77674311e-01 -8.82514656e-01 1.39783666e-01 -1.36867404e-01
3.00169617e-01 5.14104366e-02 1.31106687e+00 -6.08588994e-01
-2.48663068e-01 -7.82233626e-02 3.72714221e-01 8.54513586e-01
1.86440066e-01 -1.69093981e-01 -1.63991761e+00 -2.93265551e-01
4.15468484e-01 -1.65283263e-01 1.09819078e+00 6.42115116e-01
9.94556904e-01 -7.39138901e-01 1.83547940e-02 5.99419117e-01
1.17595863e+00 1.80851981e-01 5.22877097e-01 -4.09511358e-01
4.89557326e-01 7.89243340e-01 -2.96566069e-01 2.30419517e-01
-6.55717403e-03 6.61482632e-01 1.83595479e-01 3.44983250e-01
-1.68565989e-01 -5.31307399e-01 6.94604635e-01 1.72602773e+00
1.95461288e-01 -7.76284337e-02 -5.19957900e-01 6.63718164e-01
-1.35936272e+00 -8.80857468e-01 3.60731542e-01 2.28032064e+00
8.14625442e-01 4.14009571e-01 -1.04540624e-01 5.60967982e-01
4.22436476e-01 2.20797539e-01 -2.86856234e-01 -6.73000038e-01
-3.32855172e-02 5.39256275e-01 1.84630170e-01 7.39525557e-01
-9.16715980e-01 7.01456726e-01 6.50915861e+00 7.45995283e-01
-1.14581001e+00 3.93428743e-01 1.05452441e-01 -8.95197168e-02
-3.94135356e-01 -5.62522970e-02 -5.32488465e-01 5.30474424e-01
1.60814536e+00 1.55728787e-01 7.33868122e-01 6.29086912e-01
1.65889606e-01 4.75839257e-01 -1.44521046e+00 7.48807847e-01
-1.47600546e-01 -1.17056191e+00 8.97909030e-02 2.02720344e-01
3.99391383e-01 -1.36973504e-02 3.46637934e-01 4.25723732e-01
-4.59428020e-02 -1.24415159e+00 1.02605379e+00 5.65504074e-01
8.87264907e-01 -6.32223964e-01 2.83649921e-01 4.16254848e-01
-1.15114498e+00 -9.07123461e-02 -4.75216031e-01 6.97906390e-02
9.03331712e-02 5.67768216e-01 -1.00704050e+00 2.75635988e-01
3.76529515e-01 3.71479183e-01 -1.91624165e-02 3.33558649e-01
-4.11493368e-02 1.07506633e+00 -2.60592818e-01 4.59936082e-01
1.83896586e-01 -1.43860042e-01 6.88969970e-01 1.23505223e+00
3.58342111e-01 6.57250285e-02 -1.48052245e-01 1.00361276e+00
-1.23069793e-01 8.03708658e-02 -5.50870776e-01 -5.82929909e-01
6.05870664e-01 7.39013791e-01 -3.25980514e-01 -2.40738735e-01
-3.03686917e-01 1.01262200e+00 2.81392157e-01 5.71656406e-01
-4.62518513e-01 -2.58034885e-01 9.26564872e-01 4.52290811e-02
8.20040345e-01 -4.19207394e-01 1.26425236e-01 -1.02876604e+00
-4.37304564e-02 -7.80039728e-01 -1.04348443e-01 -6.29232287e-01
-9.88076270e-01 7.99044847e-01 -1.80837587e-01 -1.01684868e+00
-9.57094014e-01 -5.21240473e-01 -3.62180322e-01 1.10493517e+00
-1.30322564e+00 -6.54512823e-01 4.89512473e-01 4.92937237e-01
7.11933255e-01 -3.08323234e-01 1.20681536e+00 2.36622244e-01
-2.33098224e-01 4.87253189e-01 2.99423039e-01 2.56750166e-01
3.32940072e-01 -1.18027449e+00 3.69751871e-01 6.62869394e-01
6.57391965e-01 6.46399856e-01 8.25306535e-01 -2.81943202e-01
-1.19885182e+00 -1.00093842e+00 9.51571345e-01 -5.34034789e-01
5.07091463e-01 -8.40989292e-01 -1.17399740e+00 7.60533810e-01
-1.04118325e-02 7.34832436e-02 6.53410733e-01 9.03345421e-02
-4.74539101e-01 -4.13314253e-02 -5.58178127e-01 1.62483767e-01
5.44138730e-01 -1.39624262e+00 -8.69007051e-01 -4.47939523e-02
1.06974423e+00 -6.18542656e-02 -8.04296851e-01 1.81443319e-01
4.36423570e-01 -8.56666088e-01 1.04838467e+00 -6.07461751e-01
3.55230123e-01 1.20738917e-03 -5.88271379e-01 -1.36937892e+00
-3.72513980e-01 -4.54505324e-01 -3.11691016e-01 1.40045238e+00
4.57661539e-01 -3.58405888e-01 5.26177347e-01 4.72082458e-02
-1.78515002e-01 -3.64064902e-01 -1.14909863e+00 -7.15958834e-01
2.41829187e-01 -6.97888613e-01 4.65189785e-01 8.76463234e-01
-8.52544755e-02 4.77080137e-01 -4.51059967e-01 5.65567732e-01
3.95992875e-01 -1.46802515e-01 2.38821268e-01 -1.20863819e+00
-7.87956238e-01 -3.88924658e-01 -3.22967947e-01 -9.10783887e-01
4.61382687e-01 -1.02872431e+00 2.39294931e-01 -9.88858879e-01
-4.91077930e-01 9.76387132e-03 -5.98689497e-01 1.55050680e-01
3.86151582e-01 3.13660093e-02 1.23106420e-01 1.53636143e-01
5.78403808e-02 6.66555583e-01 6.15599394e-01 -1.10223390e-01
-2.85892665e-01 2.13599000e-02 -2.67611593e-01 7.24749446e-01
5.98209500e-01 -5.99861145e-01 -4.92070585e-01 -3.26274574e-01
-4.51378943e-03 4.24371690e-01 3.42726290e-01 -1.01833165e+00
9.71850157e-02 3.18362892e-01 2.69557953e-01 -2.01764107e-01
8.95098805e-01 -6.34356499e-01 1.95491686e-01 1.43091574e-01
-6.12004876e-01 -3.76444906e-01 1.13718137e-01 4.71806675e-01
-4.29080993e-01 -3.79478842e-01 6.82524562e-01 7.63700111e-03
-4.12582457e-01 -2.02428875e-03 -6.60468102e-01 6.67026266e-02
3.04903358e-01 -4.78031598e-02 2.93665260e-01 -5.47627270e-01
-1.00650275e+00 -5.67363262e-01 9.18911695e-02 3.95821959e-01
4.25016612e-01 -1.36986983e+00 -6.36556745e-01 6.78779781e-01
-1.60293013e-03 -5.02866447e-01 6.93953335e-02 4.17055845e-01
2.52010554e-01 6.32375181e-01 3.04399412e-02 -4.21650052e-01
-7.69857824e-01 5.89429080e-01 3.42825294e-01 2.26180241e-01
-8.03001225e-01 8.12232375e-01 4.83409852e-01 -1.74566880e-01
4.41885233e-01 -5.66980600e-01 -2.37816244e-01 2.34063044e-01
3.01071107e-01 8.11778232e-02 1.77120805e-01 -9.72957909e-01
-2.70790160e-01 3.70393991e-01 3.11430365e-01 -4.52055007e-01
1.26537609e+00 -2.19365209e-01 9.12166983e-02 1.06863832e+00
1.62442935e+00 3.17298055e-01 -1.34204865e+00 -2.56813824e-01
4.16005328e-02 1.54154271e-01 3.54894638e-01 -3.49072039e-01
-6.68156624e-01 1.13161314e+00 5.98361552e-01 3.81687671e-01
8.58818114e-01 5.09720668e-02 6.07608199e-01 1.85132876e-01
5.18923365e-02 -1.11862385e+00 -1.46784689e-02 6.18323326e-01
7.81024337e-01 -8.45374703e-01 -5.15617132e-01 -1.07651487e-01
-5.30235350e-01 1.11119223e+00 -2.35935777e-01 -2.05065414e-01
9.20707047e-01 3.29937279e-01 -8.21581576e-03 -5.46253026e-02
-7.89346039e-01 -1.86871812e-01 6.04975283e-01 4.64450538e-01
2.43125841e-01 6.73922971e-02 2.52708405e-01 8.07875991e-01
-7.53897607e-01 -6.17103279e-01 4.02921200e-01 2.93849289e-01
-5.64389527e-01 -8.41953814e-01 -2.22286940e-01 2.75908411e-01
-4.88698542e-01 -2.94711918e-01 -2.13736176e-01 3.67995054e-01
1.34408370e-01 8.66568327e-01 3.98739278e-01 -3.31609488e-01
1.90479726e-01 8.03806365e-01 2.58792818e-01 -7.43590951e-01
-1.61116645e-01 4.65832055e-01 1.08455785e-03 -4.09708381e-01
-1.13380544e-01 -5.81230283e-01 -1.15479827e+00 2.30791956e-01
-2.12438047e-01 4.44010288e-01 9.01590407e-01 1.06732512e+00
1.36406153e-01 8.09862673e-01 5.31932652e-01 -8.09128046e-01
-6.66875124e-01 -9.67829466e-01 -8.07810247e-01 4.03531700e-01
8.83081734e-01 -4.01860625e-01 -5.78697741e-01 3.79355699e-01] | [14.912330627441406, 6.416104793548584] |
e492177b-d37a-4b02-ab58-8c190cdd07a9 | my-health-sensor-my-classifier-adapting-a | 2009.10799 | null | https://arxiv.org/abs/2009.10799v1 | https://arxiv.org/pdf/2009.10799v1.pdf | My Health Sensor, my Classifier: Adapting a Trained Classifier to Unlabeled End-User Data | In this work, we present an approach for unsupervised domain adaptation (DA) with the constraint, that the labeled source data are not directly available, and instead only access to a classifier trained on the source data is provided. Our solution, iteratively labels only high confidence sub-regions of the target data distribution, based on the belief of the classifier. Then it iteratively learns new classifiers from the expanding high-confidence dataset. The goal is to apply the proposed approach on DA for the task of sleep apnea detection and achieve personalization based on the needs of the patient. In a series of experiments with both open and closed sleep monitoring datasets, the proposed approach is applied to data from different sensors, for DA between the different datasets. The proposed approach outperforms in all experiments the classifier trained in the source domain, with an improvement of the kappa coefficient that varies from 0.012 to 0.242. Additionally, our solution is applied to digit classification DA between three well established digit datasets, to investigate the generalizability of the approach, and to allow for comparison with related work. Even without direct access to the source data, it achieves good results, and outperforms several well established unsupervised DA methods. | ['Lars Aakerøy', 'Britt Øverland', 'Knut Liestøl', 'Mohan Kankanhalli', 'Stein Kristiansen', 'Sigurd Steinshamn', 'Konstantinos Nikolaidis', 'Vera Goebel', 'Thomas Plagemann', 'Harriet Akre', 'Gunn Marit Traaen'] | 2020-09-22 | null | null | null | null | ['sleep-apnea-detection'] | ['medical'] | [ 2.42969885e-01 4.18564647e-01 -2.57371545e-01 -4.45619076e-01
-6.12719655e-01 -4.07491952e-01 1.95756137e-01 6.78868115e-01
-6.72516584e-01 1.05919659e+00 2.19127964e-02 5.83858117e-02
-4.08805788e-01 -6.40289187e-01 -1.95298269e-01 -6.47558868e-01
1.24875695e-01 7.47849822e-01 4.62306827e-01 1.27841130e-01
1.09874951e-02 3.74767989e-01 -1.71521235e+00 2.64856033e-02
9.83570695e-01 1.25268984e+00 1.94339231e-01 3.37952077e-01
1.23232929e-02 5.83130419e-01 -6.92535043e-01 -4.65754010e-02
8.30237493e-02 -3.49025548e-01 -6.47926569e-01 3.17433178e-01
1.22480117e-01 2.78976038e-02 2.84974903e-01 8.99534881e-01
5.17925084e-01 9.81892124e-02 7.69973338e-01 -8.06490600e-01
-1.39523014e-01 5.17440677e-01 2.76093185e-02 2.43913278e-01
4.60008025e-01 -1.48813084e-01 6.45885646e-01 -4.73154366e-01
4.73310649e-01 5.88502586e-01 5.89628100e-01 6.35772586e-01
-1.34510529e+00 -4.89796162e-01 -1.35112166e-01 2.95373201e-01
-1.42739987e+00 -4.83471096e-01 7.17117727e-01 -5.60562253e-01
6.38710260e-01 1.16386209e-02 5.18977702e-01 1.07557869e+00
-1.10436171e-01 3.35790783e-01 1.31860220e+00 -7.03992128e-01
9.53821659e-01 8.33265901e-01 4.79717225e-01 4.18895036e-01
4.05737072e-01 1.36527777e-01 -5.22667110e-01 -2.79013813e-01
1.15920879e-01 -1.63372442e-01 -6.79996833e-02 -5.97515285e-01
-8.86809766e-01 7.08109319e-01 1.64165199e-01 7.64742553e-01
-6.93005025e-01 -7.19216704e-01 3.57458651e-01 8.47448334e-02
3.81203979e-01 3.98691386e-01 -4.75249231e-01 -2.56768018e-01
-1.16518927e+00 -1.23725861e-01 1.07034004e+00 7.75155127e-01
6.13719106e-01 -2.34701782e-01 -9.87204257e-03 7.70785153e-01
2.88872033e-01 3.32389295e-01 9.65799391e-01 -5.68719685e-01
2.39872724e-01 8.40929568e-01 9.49551985e-02 -5.90408146e-01
-6.27212167e-01 -4.57333565e-01 -6.52028203e-01 1.88643768e-01
5.00309646e-01 -1.96134821e-01 -8.15795541e-01 1.56109118e+00
2.73371994e-01 -1.10699102e-01 4.31314915e-01 4.31152612e-01
6.31721020e-01 2.64259696e-01 8.45383778e-02 -6.47162914e-01
1.07776999e+00 -4.62093681e-01 -7.52973258e-01 -1.86205193e-01
2.98588127e-01 -4.45252717e-01 9.01264906e-01 8.98023605e-01
-7.77515829e-01 -6.98881567e-01 -1.28052807e+00 2.83159465e-01
-4.60407078e-01 4.68966722e-01 -1.50995150e-01 8.09946358e-01
-8.65015626e-01 3.69785428e-01 -8.18805456e-01 -8.30157280e-01
2.99004823e-01 5.31699538e-01 -1.80379212e-01 2.71586124e-02
-8.38109314e-01 1.04460847e+00 8.15255642e-01 -1.65736258e-01
-6.34857476e-01 -3.32267612e-01 -6.06795669e-01 2.69695255e-03
2.15519860e-01 -4.00700837e-01 9.67224360e-01 -8.85439634e-01
-1.49594486e+00 7.72425711e-01 -6.45109490e-02 -8.59366775e-01
5.64044237e-01 -1.75050855e-01 -6.48818374e-01 1.75268903e-01
-8.04031268e-03 4.18898582e-01 7.84536779e-01 -1.17998242e+00
-7.19974935e-01 -5.54768562e-01 -2.44180501e-01 -1.00283191e-01
-6.33553743e-01 -3.52258444e-01 -1.66277736e-01 -4.17226762e-01
-1.61503866e-01 -8.71203125e-01 -7.95128867e-02 -1.68794766e-01
-1.58249378e-01 -2.48000458e-01 5.21045506e-01 -5.98012865e-01
1.50209439e+00 -2.43625569e+00 2.69631773e-01 5.49524128e-01
4.92819771e-02 1.98752910e-01 5.43568432e-01 1.98877975e-01
-1.88460536e-02 -4.97893661e-01 -8.83676589e-01 -3.91652316e-01
-2.47971952e-01 5.94658673e-01 9.46822241e-02 2.92342484e-01
-1.69337362e-01 -4.84395511e-02 -7.55359232e-01 -6.53499186e-01
3.64267349e-01 1.99360609e-01 -3.22345644e-01 3.80990088e-01
-1.57202497e-01 5.36518872e-01 -3.22159976e-01 2.68982947e-01
4.45997804e-01 -8.74261409e-02 3.21661562e-01 -1.51787087e-01
-1.44015387e-01 7.56070241e-02 -1.25502920e+00 1.73827505e+00
-4.43761200e-01 3.36326003e-01 -3.59178364e-01 -1.24164844e+00
1.40208840e+00 4.27415133e-01 7.02112734e-01 -4.43700373e-01
7.87368044e-02 4.01480138e-01 3.50108854e-02 -6.25524700e-01
1.04620419e-01 -1.94521248e-01 4.68176529e-02 1.92778230e-01
3.67122918e-01 4.60710712e-02 2.97011703e-01 -1.01887770e-01
9.67362523e-01 -1.93596154e-01 8.26285005e-01 -4.13587511e-01
8.89915705e-01 -3.58146094e-02 3.01047027e-01 4.82366920e-01
-1.30347535e-01 3.24031740e-01 2.74339408e-01 -6.88640624e-02
-7.88047254e-01 -1.07842839e+00 -4.84713942e-01 5.75120091e-01
-5.90702184e-02 -3.00815850e-01 -8.68481159e-01 -9.78206754e-01
-1.01457708e-01 1.06645644e+00 -7.36085892e-01 -3.19187075e-01
-1.28809139e-01 -6.87052190e-01 4.16397870e-01 4.60672855e-01
5.34134865e-01 -1.03496933e+00 -1.05101454e+00 2.61956096e-01
-8.05955529e-02 -1.01219869e+00 1.69086307e-01 5.44547856e-01
-1.14490080e+00 -1.17190421e+00 -4.31265503e-01 -5.42994022e-01
5.25274932e-01 -5.78795493e-01 8.53772938e-01 -2.03820333e-01
-1.13942184e-01 6.38605177e-01 -5.43849289e-01 -3.86303902e-01
-7.18986750e-01 2.93880850e-01 2.12541103e-01 2.64640331e-01
7.24246502e-01 -6.82446659e-01 -3.47000986e-01 3.79690528e-01
-8.21610272e-01 -6.64755583e-01 6.03973091e-01 5.87457061e-01
5.58254242e-01 1.86190560e-01 8.55109572e-01 -8.99498940e-01
5.41886866e-01 -6.68219805e-01 -5.08657992e-01 7.83073157e-02
-1.25908315e+00 2.24176273e-01 6.07029736e-01 -3.78994048e-01
-1.00430965e+00 3.63258272e-01 -2.62888610e-01 -2.37711608e-01
-7.02573001e-01 3.67604077e-01 -1.91296756e-01 3.79970610e-01
1.09423268e+00 3.01029861e-01 1.04122639e-01 -7.88302243e-01
5.03742807e-02 9.72472429e-01 6.31495237e-01 -4.45284277e-01
3.81050557e-01 2.31149197e-01 -1.15223221e-01 -8.32533300e-01
-6.88084841e-01 -5.57097971e-01 -1.01073098e+00 -6.52122032e-03
9.93145704e-01 -6.60241783e-01 -1.72924578e-01 2.66317874e-01
-6.37972891e-01 -3.07959974e-01 -7.45447159e-01 7.58579552e-01
-6.50034606e-01 4.62450862e-01 5.83624057e-02 -9.36379433e-01
-1.40977547e-01 -8.76308978e-01 5.07780790e-01 1.19594574e-01
-4.35663640e-01 -1.06187570e+00 2.47294798e-01 2.09012955e-01
2.21712470e-01 1.93489879e-01 8.03198934e-01 -1.28669727e+00
1.89122602e-01 -3.46567929e-01 2.80564696e-01 8.74013364e-01
6.11404061e-01 -2.61246890e-01 -1.14405012e+00 -2.34874353e-01
4.07281846e-01 -1.84116662e-01 6.83010995e-01 2.00642928e-01
1.08768189e+00 -8.10791850e-02 -3.07396650e-01 -7.82407522e-02
1.24258363e+00 4.33448255e-01 4.81297314e-01 3.56918037e-01
8.99604261e-02 5.64336300e-01 8.39829147e-01 6.78673148e-01
3.25277865e-01 6.87402070e-01 3.04005951e-01 5.30893318e-02
-5.25104627e-02 8.96285102e-02 3.38974267e-01 5.21196485e-01
7.90484473e-02 1.78910736e-02 -7.64942050e-01 6.84792340e-01
-1.68146873e+00 -5.63649178e-01 8.32157359e-02 2.66175437e+00
6.95630729e-01 4.20540273e-01 6.77021146e-01 6.07817888e-01
6.47191525e-01 -3.63308042e-01 -5.67508101e-01 -3.01272184e-01
2.65327305e-01 5.39993703e-01 3.26759160e-01 4.51406300e-01
-9.82678175e-01 3.03818256e-01 6.17087793e+00 3.68145853e-01
-9.98151481e-01 2.90712684e-01 2.63581961e-01 -9.98690277e-02
3.52364242e-01 -3.56866390e-01 -6.64333522e-01 7.87881017e-01
1.26552033e+00 -1.22633986e-01 2.73807406e-01 9.44221854e-01
1.38481691e-01 -4.56261903e-01 -1.37951398e+00 8.69005799e-01
2.43577078e-01 -6.27105951e-01 -4.66273934e-01 2.80904230e-02
4.61042851e-01 -1.03917301e-01 -1.13712892e-01 2.05295295e-01
-7.52520040e-02 -6.39884412e-01 5.40954649e-01 6.73973024e-01
5.55016100e-01 -7.58001089e-01 9.80869412e-01 5.31173050e-01
-9.22642767e-01 -4.99722302e-01 -1.90169960e-01 9.09220800e-02
-1.76871359e-01 7.04482138e-01 -1.08967292e+00 5.22477388e-01
8.28642011e-01 7.03221202e-01 -7.29458094e-01 1.06927836e+00
-1.94281250e-01 8.05080533e-01 -4.76835847e-01 -5.57267889e-02
-1.47132456e-01 3.62812467e-02 3.30264390e-01 1.15845811e+00
4.27423000e-01 1.76221933e-02 7.55425990e-02 6.75967038e-01
3.68536860e-01 3.33591193e-01 -5.73638916e-01 3.17043364e-01
4.12166387e-01 9.60981369e-01 -4.86855417e-01 -5.50868988e-01
-3.08144152e-01 8.16566765e-01 1.41178429e-01 2.39856709e-02
-5.37873089e-01 -2.51882344e-01 5.91034926e-02 1.83231503e-01
5.52752972e-01 8.54697973e-02 -2.51072109e-01 -9.27498281e-01
1.35703042e-01 -7.87665606e-01 8.87869239e-01 -5.19571364e-01
-1.36672056e+00 7.25565076e-01 3.68171751e-01 -1.42817497e+00
-4.74451870e-01 -5.33658206e-01 -2.96385229e-01 6.56561852e-01
-1.16363418e+00 -6.32442594e-01 -4.26229626e-01 8.16477001e-01
3.84677380e-01 -3.23609322e-01 9.93377805e-01 5.84590793e-01
-3.64519387e-01 5.48684537e-01 1.88503638e-01 -2.76925117e-01
8.95765007e-01 -1.30287564e+00 -4.56348032e-01 5.54403961e-01
9.01529416e-02 3.61658871e-01 6.42018795e-01 -5.61480343e-01
-6.26023829e-01 -9.87929106e-01 6.71825230e-01 -4.92061585e-01
3.42056572e-01 -5.01002446e-02 -1.07566798e+00 4.24004823e-01
5.80683202e-02 -1.83980897e-01 9.82725024e-01 1.05999701e-01
1.14716543e-02 -5.28864264e-01 -1.65768802e+00 -2.96590865e-01
6.25288308e-01 -1.94653302e-01 -9.82198060e-01 1.82604581e-01
1.92257032e-01 -2.28819966e-01 -1.22930121e+00 2.87465513e-01
3.55638683e-01 -1.07197440e+00 6.82341635e-01 -2.31979236e-01
9.95459482e-02 -4.04287308e-01 -3.23682457e-01 -1.33315182e+00
-1.28684212e-02 -4.25669365e-02 -2.22177699e-01 1.28571475e+00
5.53381562e-01 -6.42495811e-01 5.71385324e-01 5.07069170e-01
-1.18161298e-01 -3.84795547e-01 -1.15826535e+00 -5.77366590e-01
-1.42343149e-01 -4.28205132e-01 3.94211143e-01 7.33635306e-01
1.90595523e-01 2.94206381e-01 -2.26812616e-01 3.45974177e-01
5.62147081e-01 -1.58481553e-01 5.62694967e-01 -1.56218004e+00
-5.41943192e-01 3.75101492e-02 -5.94908655e-01 -4.88902271e-01
-9.13825259e-02 -7.28928804e-01 6.42569736e-02 -1.32550001e+00
-2.91283280e-02 -4.19844776e-01 -6.80320919e-01 5.02180755e-01
1.14549153e-01 2.02320561e-01 -1.21938139e-01 1.72125861e-01
-3.05849940e-01 2.90088713e-01 4.40168649e-01 -4.19395827e-02
-5.43399632e-01 4.00851578e-01 -5.46398163e-01 8.25353682e-01
9.87349212e-01 -5.78205287e-01 -6.46967530e-01 1.51312426e-01
-1.90669328e-01 -1.07499510e-01 3.04921508e-01 -1.57251847e+00
1.44494355e-01 2.50849903e-01 4.55167770e-01 -5.88000357e-01
2.92460084e-01 -1.36247241e+00 1.16367117e-01 6.47166848e-01
-3.76889139e-01 -2.72554785e-01 1.91181451e-01 5.60861111e-01
-1.26245826e-01 -5.45888782e-01 1.01082850e+00 8.71580392e-02
-6.11078680e-01 -2.56603181e-01 -4.17670876e-01 -5.87626211e-02
1.11082006e+00 -2.94019818e-01 1.98008209e-01 -1.38193876e-01
-1.48552585e+00 -1.02323689e-01 4.25594658e-01 1.02398545e-01
3.36016417e-01 -9.94593740e-01 -4.02397811e-01 3.55075330e-01
5.19738555e-01 1.12180389e-03 -9.56907570e-02 7.79820919e-01
-5.12522198e-02 3.10427487e-01 -2.94637918e-01 -7.69133627e-01
-1.23471439e+00 8.48248959e-01 3.17354172e-01 -1.25410795e-01
-5.39171815e-01 2.77987212e-01 -2.48890027e-01 -2.13977262e-01
3.87624055e-01 -7.16798544e-01 -4.90132809e-01 2.01484159e-01
3.34898591e-01 4.71549958e-01 4.15719241e-01 -4.57359344e-01
-4.20604289e-01 5.40263236e-01 1.43359140e-01 -9.13544893e-02
1.16269696e+00 -1.80343464e-01 -2.03973585e-04 8.19376290e-01
7.92291999e-01 2.15336993e-01 -9.10802007e-01 -2.66767651e-01
2.96461731e-01 -1.63126037e-01 -2.13131055e-01 -1.05167103e+00
-7.59012640e-01 7.05625296e-01 1.27698827e+00 3.14295322e-01
1.48514354e+00 1.62060052e-01 1.72833309e-01 5.08477807e-01
3.53482753e-01 -1.05266917e+00 -3.83800603e-02 -3.83758061e-02
5.11404812e-01 -1.05764234e+00 1.04824930e-01 -1.09393239e-01
-7.47959137e-01 1.22015440e+00 5.45739949e-01 6.85999496e-03
7.73490250e-01 2.44728364e-02 -2.21948745e-03 -4.36550528e-02
-3.58578384e-01 -1.59361318e-01 2.59604454e-01 7.97386110e-01
1.94983631e-01 4.03212868e-02 -6.29558206e-01 6.96369350e-01
-2.48199701e-01 4.04473633e-01 4.20136720e-01 6.69705451e-01
-5.78893423e-01 -1.40710652e+00 -4.23737645e-01 5.55991709e-01
-3.45101118e-01 2.46992707e-01 -4.42013472e-01 8.25502276e-01
5.75261950e-01 1.14940393e+00 6.18049987e-02 -2.03333616e-01
5.55171013e-01 3.66331935e-01 3.86366457e-01 -7.30412960e-01
-3.53464782e-01 -4.56836224e-02 2.48784781e-01 -2.76693434e-01
-7.66266227e-01 -9.64196980e-01 -1.14297271e+00 3.82298082e-01
-1.89696878e-01 5.12576699e-01 6.81144655e-01 9.44266975e-01
2.16928288e-01 4.22970116e-01 6.99956894e-01 -2.82178044e-01
-6.09487891e-01 -1.12936592e+00 -5.60088754e-01 3.64678860e-01
4.28578973e-01 -8.57984960e-01 -3.77347171e-01 2.96724766e-01] | [10.071670532226562, 3.277552366256714] |
dabfe3ec-fd0e-4dd8-8239-1d5b50adadeb | effective-audio-classification-network-based | 2211.02940 | null | https://arxiv.org/abs/2211.02940v4 | https://arxiv.org/pdf/2211.02940v4.pdf | Effective Audio Classification Network Based on Paired Inverse Pyramid Structure and Dense MLP Block | Recently, massive architectures based on Convolutional Neural Network (CNN) and self-attention mechanisms have become necessary for audio classification. While these techniques are state-of-the-art, these works' effectiveness can only be guaranteed with huge computational costs and parameters, large amounts of data augmentation, transfer from large datasets and some other tricks. By utilizing the lightweight nature of audio, we propose an efficient network structure called Paired Inverse Pyramid Structure (PIP) and a network called Paired Inverse Pyramid Structure MLP Network (PIPMN). The PIPMN reaches 96\% of Environmental Sound Classification (ESC) accuracy on the UrbanSound8K dataset and 93.2\% of Music Genre Classification (MGC) on the GTAZN dataset, with only 1 million parameters. Both of the results are achieved without data augmentation or model transfer. Public code is available at: https://github.com/JNAIC/PIPMN | ['Jianlu Shen', 'Zhen Ren', 'Yifan Huang', 'Lifang Chen', 'Zihui Yan', 'Yunjie Zhu', 'Yunhao Chen'] | 2022-11-05 | null | null | null | null | ['environmental-sound-classification', 'sound-classification', 'genre-classification'] | ['audio', 'audio', 'computer-vision'] | [ 3.97156328e-02 -1.61894321e-01 8.88071731e-02 -1.69962212e-01
-7.94904470e-01 -2.68697530e-01 7.23191425e-02 -1.70775890e-01
-4.77727413e-01 5.35265803e-01 1.51941106e-01 -1.69399217e-01
-9.38371420e-02 -8.09590280e-01 -8.52603137e-01 -5.28777421e-01
-2.14437157e-01 1.09041564e-01 1.51506290e-01 -3.39352190e-02
1.12054236e-02 1.25648826e-01 -1.84659016e+00 4.38050568e-01
6.64676189e-01 1.49870872e+00 1.00358926e-01 8.83632481e-01
5.52211590e-02 7.59263575e-01 -4.86933738e-01 -2.35589027e-01
1.88821048e-01 7.10620955e-02 -8.85938644e-01 -5.43269217e-01
3.29494327e-01 -2.04993308e-01 -4.82419401e-01 8.85305464e-01
1.04156077e+00 1.90996140e-01 1.52026072e-01 -1.50097382e+00
-5.82422972e-01 9.95210886e-01 -3.73075396e-01 2.78589875e-01
-1.52282938e-01 1.66297629e-01 1.10497761e+00 -8.34217668e-01
-1.17750429e-01 1.05766737e+00 1.13320673e+00 3.90465051e-01
-9.50451434e-01 -1.26463282e+00 -1.00057594e-01 5.01488745e-01
-1.46631086e+00 -7.21662045e-01 8.15148056e-01 -3.79526109e-01
1.04417408e+00 2.76834965e-01 7.42643178e-01 1.13130772e+00
-3.59342724e-01 6.68497503e-01 7.36909568e-01 -3.10712993e-01
9.92075428e-02 -1.17438249e-01 6.41812161e-02 4.88043576e-01
-2.46943250e-01 6.35126233e-02 -7.72072136e-01 -2.04231739e-01
6.47222340e-01 -3.05131704e-01 -3.31370533e-01 5.27774215e-01
-9.98000383e-01 5.26547551e-01 5.50467432e-01 2.34963119e-01
-3.05843890e-01 5.68970740e-01 7.25907981e-01 3.92874658e-01
5.38556933e-01 2.92847812e-01 -7.49238908e-01 -5.85347176e-01
-6.48711383e-01 1.51706055e-01 6.08863950e-01 9.83536303e-01
4.90916282e-01 5.64269006e-01 2.68355697e-01 1.23444927e+00
8.72315317e-02 4.04210865e-01 6.39607370e-01 -1.07773864e+00
6.62040651e-01 4.55470622e-01 -2.78642684e-01 -1.00320005e+00
-4.54658628e-01 -8.12408030e-01 -1.25637972e+00 -8.44045877e-02
1.56163901e-01 -4.03592587e-01 -4.75712806e-01 1.69919765e+00
7.39822686e-02 5.22222519e-01 -3.29851173e-02 5.82353294e-01
1.14318705e+00 8.93539488e-01 2.65904181e-02 3.18650603e-01
1.28233171e+00 -1.22118235e+00 -4.67914909e-01 -5.25621660e-02
4.38050389e-01 -7.97205269e-01 1.47161019e+00 4.84129250e-01
-9.99322712e-01 -1.06211686e+00 -8.92584622e-01 -3.11461128e-02
-3.35241675e-01 3.44688922e-01 5.35707057e-01 5.87701380e-01
-1.05770290e+00 7.14453161e-01 -7.19249308e-01 2.72615813e-02
6.99583054e-01 6.07860386e-01 -2.51974463e-01 3.58905286e-01
-1.22289109e+00 5.90948649e-02 4.34953421e-01 2.91989863e-01
-1.06766272e+00 -1.18090487e+00 -4.28997248e-01 4.14905638e-01
3.29838060e-02 -4.74536836e-01 1.42350149e+00 -9.08995926e-01
-1.67140210e+00 3.61641735e-01 3.34891111e-01 -5.43435633e-01
2.10265338e-01 -7.13852406e-01 -5.18998265e-01 -8.87386501e-03
-1.46803245e-01 8.83105516e-01 6.29284501e-01 -7.86113024e-01
-7.49754190e-01 -4.86368425e-02 -5.98948076e-02 -6.47318410e-03
-6.61255896e-01 2.55717516e-01 -2.65231162e-01 -7.00807571e-01
-1.89326718e-01 -8.74814510e-01 -4.79344055e-02 -1.42783359e-01
-4.65062082e-01 -3.34025085e-01 9.16147947e-01 -7.79536724e-01
1.34277594e+00 -2.48903275e+00 -2.18416587e-01 -1.91451028e-01
-1.08917423e-01 5.48706412e-01 -4.65636969e-01 2.45903015e-01
-2.11437404e-01 2.92326480e-01 -1.82078540e-01 -4.31452900e-01
1.17953904e-01 -5.93433678e-02 -4.66991991e-01 5.04459664e-02
1.60889104e-01 6.74296260e-01 -5.92659771e-01 -1.37356564e-01
-8.35561100e-03 6.23163164e-01 -1.00600576e+00 1.70247436e-01
-9.94661376e-02 4.82002944e-01 -2.01030195e-01 7.72516787e-01
5.36983252e-01 -1.64088309e-01 -3.17070514e-01 -3.19904298e-01
-3.00807834e-01 3.27208042e-01 -1.13692284e+00 1.73377264e+00
-5.80921292e-01 6.48263812e-01 2.66738862e-01 -9.57289279e-01
8.08172584e-01 7.57763743e-01 4.63927865e-01 -4.47120756e-01
2.44748846e-01 4.36140835e-01 1.67984992e-01 -5.50511062e-01
2.26856932e-01 1.59598678e-01 8.16139206e-02 3.13590527e-01
3.70498091e-01 1.18571013e-01 -6.42454401e-02 -2.69868433e-01
1.08640957e+00 -4.19370979e-02 -2.22751454e-01 -2.58154660e-01
3.51091862e-01 -2.96507508e-01 8.40183020e-01 5.38550794e-01
-1.63843766e-01 6.26118004e-01 3.41249466e-01 -5.71019351e-01
-1.17206538e+00 -5.13372362e-01 -2.51411557e-01 1.26330566e+00
-5.60715377e-01 -6.80130661e-01 -8.73952091e-01 -1.09956779e-01
-1.16541579e-01 2.70393819e-01 -3.56637776e-01 -1.66934311e-01
-5.97406864e-01 -9.79230285e-01 1.29414022e+00 7.90914178e-01
1.09726846e+00 -1.28830242e+00 -4.43927974e-01 3.97673965e-01
-4.39679921e-01 -1.13012910e+00 -1.17462270e-01 2.50612676e-01
-6.58977389e-01 -8.16767216e-01 -5.29367268e-01 -7.05596745e-01
-3.78631614e-02 -1.34453341e-01 1.15163577e+00 -2.12381817e-02
-1.60556555e-01 6.15224354e-02 -4.51553345e-01 -8.19887340e-01
-1.04503490e-01 5.43494165e-01 3.46139044e-01 6.69767484e-02
2.94716489e-02 -1.34603727e+00 -6.64310992e-01 1.12276919e-01
-5.75185895e-01 1.19529262e-01 4.09587473e-01 7.37070143e-01
5.12161911e-01 2.00770929e-01 9.54682827e-01 -3.74305218e-01
3.69477183e-01 -4.52174723e-01 -4.89795536e-01 -2.62177080e-01
-2.79311985e-01 -5.22590339e-01 8.27054024e-01 -4.96414661e-01
-6.84454858e-01 -8.57375562e-02 -8.22393775e-01 -6.04718685e-01
-5.42180300e-01 3.05379122e-01 -1.13010764e-01 8.18875730e-02
5.79188406e-01 2.12547824e-01 -3.19104314e-01 -1.04060578e+00
-4.91834758e-03 1.11329079e+00 4.55029994e-01 -5.34869194e-01
4.57020611e-01 2.66286373e-01 -2.80143172e-01 -9.63270724e-01
-1.01450944e+00 -1.28694430e-01 -6.25521839e-01 -5.66477366e-02
9.48621929e-01 -1.16105592e+00 -8.38454962e-01 7.81425893e-01
-1.04965901e+00 -7.12959349e-01 -3.58976424e-01 5.28722823e-01
-3.92766505e-01 -4.33403552e-02 -8.15308094e-01 -8.53977919e-01
-7.99728394e-01 -9.54549432e-01 8.98804367e-01 6.66367039e-02
-5.34668285e-03 -4.51303929e-01 -5.12062199e-02 4.19051170e-01
8.00247133e-01 1.90379322e-02 1.08845294e+00 -5.43885469e-01
-3.57836902e-01 -1.97600573e-01 -2.92083144e-01 8.37654948e-01
-5.89812808e-02 -1.26047409e-03 -1.66344750e+00 -1.24654599e-01
-2.63780028e-01 -4.59926993e-01 7.53803432e-01 4.04303432e-01
1.88905215e+00 -4.02844965e-01 1.30331039e-01 9.13134038e-01
1.15032411e+00 4.11095738e-01 5.90526402e-01 3.10214877e-01
8.00941527e-01 1.81126937e-01 1.77219380e-02 4.92721349e-01
2.87355870e-01 5.71624100e-01 7.01272547e-01 -1.16068058e-01
-2.29676768e-01 -1.56634361e-01 3.19637120e-01 1.44934726e+00
-4.97357279e-01 -7.09528476e-02 -1.03284621e+00 4.93017614e-01
-1.53446007e+00 -8.49661291e-01 -2.12782457e-01 1.83950198e+00
8.76792133e-01 -1.55721921e-02 2.14002818e-01 7.86926091e-01
5.78417897e-01 8.07381198e-02 -2.52762228e-01 -2.35374048e-01
-3.81663926e-02 3.39153320e-01 7.64749423e-02 9.17931199e-02
-1.35763025e+00 8.45177770e-01 6.14330339e+00 1.11964488e+00
-1.15091407e+00 3.29376042e-01 6.55936658e-01 -2.22424194e-01
2.96362072e-01 -3.61626178e-01 -8.14526320e-01 7.05828905e-01
1.45798934e+00 3.56571287e-01 4.38599199e-01 9.29101884e-01
5.33969589e-02 4.28128213e-01 -5.84055722e-01 1.15819812e+00
-3.84458959e-01 -1.27264774e+00 -6.43704683e-02 -5.99239990e-02
4.38691109e-01 4.98066694e-01 1.20446570e-01 5.88008344e-01
-5.15146405e-02 -8.73368680e-01 7.15143502e-01 3.50484967e-01
8.86083961e-01 -1.00395274e+00 8.97771120e-01 2.00628236e-01
-1.50762010e+00 -5.24453819e-01 -5.76959312e-01 -3.10638875e-01
-4.90861461e-02 5.10033548e-01 -4.30505157e-01 5.75473428e-01
1.44896245e+00 6.09639108e-01 -5.23868918e-01 1.14242661e+00
5.47083877e-02 1.34642935e+00 -4.82577890e-01 1.68078557e-01
2.51258165e-01 1.55203566e-02 4.78845954e-01 1.17605674e+00
4.44498867e-01 -1.20630391e-01 -6.83876351e-02 5.66150904e-01
-3.44672143e-01 1.91536993e-01 -2.52778292e-01 -1.60806812e-02
4.98031139e-01 1.37595868e+00 -3.22268516e-01 -3.47096056e-01
-3.45863104e-01 5.30726492e-01 2.41906449e-01 1.06405981e-01
-9.44119275e-01 -6.80566907e-01 8.50314319e-01 -3.09555940e-02
2.18025804e-01 5.12611419e-02 -1.98137313e-01 -8.83853555e-01
2.53543705e-02 -9.13896024e-01 2.71176010e-01 -8.62832725e-01
-1.20128441e+00 9.20896769e-01 -3.45661908e-01 -1.43936670e+00
3.43255609e-01 -6.15020156e-01 -7.66306281e-01 5.14541388e-01
-1.47565043e+00 -1.08603072e+00 -5.43186128e-01 6.56242251e-01
5.42063832e-01 -4.92320001e-01 1.20067501e+00 1.08600950e+00
-8.15079749e-01 6.67270422e-01 4.86489423e-02 4.18208659e-01
5.19033849e-01 -1.09978604e+00 3.45822304e-01 3.60110611e-01
1.40357062e-01 1.89776331e-01 1.90479293e-01 -7.33416120e-04
-1.02263665e+00 -1.38699293e+00 6.23248577e-01 -4.49373238e-02
8.25199544e-01 -4.54828471e-01 -9.72331285e-01 5.75342774e-01
3.43243748e-01 1.10156983e-01 1.12283432e+00 2.91778296e-01
-3.98360610e-01 -5.89714229e-01 -6.81421220e-01 4.12235290e-01
1.09518850e+00 -5.06809413e-01 -5.33876158e-02 3.30930412e-01
1.06480110e+00 -2.94007838e-01 -1.14418113e+00 6.16279185e-01
6.18194640e-01 -7.61000395e-01 9.73241568e-01 -7.07364798e-01
5.44866264e-01 -2.46935427e-01 -4.91281927e-01 -1.10523546e+00
-3.93977046e-01 -5.06219804e-01 -1.13955535e-01 1.48825812e+00
5.31805396e-01 -6.02116883e-01 5.71293235e-01 6.65729716e-02
-7.49632299e-01 -8.23951602e-01 -1.23450768e+00 -8.93876135e-01
1.05254754e-01 -9.64519739e-01 9.29129422e-01 1.08200586e+00
-2.54821211e-01 3.23137224e-01 -5.44361115e-01 2.33029574e-01
3.68994176e-01 -1.10027775e-01 7.91120291e-01 -1.35364580e+00
-4.00070280e-01 -4.53595072e-01 -4.88489956e-01 -6.26200974e-01
5.17140143e-02 -8.22215438e-01 -1.74476638e-01 -1.16386735e+00
-1.37040108e-01 -7.46014595e-01 -6.86587811e-01 9.78587985e-01
2.99676239e-01 6.54048800e-01 2.28860348e-01 1.19006857e-01
-4.32356417e-01 8.12231719e-01 9.53785598e-01 -8.90685990e-02
-2.43104920e-01 1.50176719e-01 -5.61850429e-01 9.99847472e-01
1.45493817e+00 -4.68957126e-01 -2.18173221e-01 -7.73202538e-01
1.36437476e-01 -1.21287830e-01 5.75840116e-01 -1.70686126e+00
1.67902678e-01 3.30821216e-01 3.06693524e-01 -5.95352054e-01
6.22694373e-01 -6.80569768e-01 2.83381969e-01 4.83598232e-01
-3.06824565e-01 9.02798623e-02 6.11629725e-01 3.45183939e-01
-4.51962531e-01 -5.82971387e-02 7.83260226e-01 2.66657025e-02
-3.97056371e-01 5.15823364e-01 -3.20716083e-01 -6.68091327e-02
4.49715585e-01 1.84569091e-01 -2.96509653e-01 -2.66177326e-01
-7.82171786e-01 4.07293141e-02 -3.24850142e-01 4.32482988e-01
3.42995524e-01 -1.57488596e+00 -8.68221402e-01 3.10249448e-01
-2.04204142e-01 1.99965179e-01 7.18814552e-01 9.64090168e-01
-6.01182759e-01 3.97462696e-01 -2.80759066e-01 -5.89507163e-01
-1.23990512e+00 1.18217096e-01 4.39959049e-01 -1.94000322e-02
-8.03030431e-01 1.19725668e+00 1.27207056e-01 -8.04791689e-01
5.03391206e-01 -3.53746235e-01 -1.92276761e-01 3.72017249e-02
5.26040256e-01 5.73410094e-01 9.51800495e-02 -3.81709725e-01
-1.52826324e-01 5.50218463e-01 2.64673531e-01 8.73857141e-02
1.72288585e+00 2.40208209e-01 -1.10571519e-01 5.21146536e-01
1.05180502e+00 -1.75800979e-01 -1.18093526e+00 -1.38101414e-01
-3.07180911e-01 -2.71614462e-01 1.90899134e-01 -6.82949305e-01
-1.47685671e+00 1.43506789e+00 8.92442644e-01 3.98746073e-01
1.33346796e+00 -1.44872278e-01 9.57863390e-01 5.64275205e-01
5.32760732e-02 -1.06791282e+00 4.60755602e-02 7.19363391e-01
1.17661846e+00 -9.18948174e-01 -4.73899871e-01 -1.92487881e-01
-4.13637370e-01 9.37338948e-01 7.15320528e-01 -2.24483430e-01
1.06273139e+00 4.78777856e-01 1.00493215e-01 1.50269553e-01
-9.76748824e-01 2.49728262e-02 1.12493418e-01 4.79450315e-01
4.85346168e-01 5.15443236e-02 2.98530966e-01 1.15240514e+00
-7.68022001e-01 -7.27524310e-02 1.17716573e-01 6.38571143e-01
-3.39372247e-01 -7.68633783e-01 -1.57965168e-01 4.95876640e-01
-8.64734888e-01 -4.25354004e-01 1.78684816e-02 5.99556506e-01
4.21862483e-01 8.32304537e-01 2.28608519e-01 -7.23672271e-01
3.97621036e-01 2.75028586e-01 1.70330890e-02 -3.75575840e-01
-9.18457925e-01 2.42765442e-01 1.11258700e-01 -3.68201047e-01
-3.17652524e-01 -4.01231766e-01 -9.53642607e-01 -3.58548641e-01
-3.62325937e-01 2.20048010e-01 5.28339386e-01 4.80466396e-01
8.04293036e-01 9.72315431e-01 5.14630139e-01 -1.08253253e+00
-2.67481565e-01 -1.45346534e+00 -6.31593645e-01 -1.37926325e-01
1.99992269e-01 -4.33262944e-01 -3.37269038e-01 2.00975552e-01] | [15.159199714660645, 5.218752384185791] |
38101bb6-1266-4052-b876-ad3c5ed8ed92 | generalizing-successor-features-to-continuous | null | null | https://openreview.net/forum?id=0m4c9ZfDrDt | https://openreview.net/pdf?id=0m4c9ZfDrDt | Generalizing Successor Features to continuous domains for Multi-task Learning | The deep reinforcement learning (RL) framework has shown great promise to tackle sequential decision-making problems, where the agent learns to behave optimally through interactions with the environment and receiving rewards.
The ability of an RL agent to learn different reward functions concurrently has many benefits, such as the decomposition of task rewards and skill reuse. One obstacle for achieving this, is the amount of data required as well as the capacity of the model for solving multiple tasks. In this paper, we consider the problem of continuous control for various robot manipulation tasks with an explicit representation that promotes skill reuse while learning multiple tasks, related through the reward function. Our approach relies on two key concepts: successor features (SF), a value function representation that decouples the dynamics of the environment from the rewards, and an actor-critic framework that incorporates the learned SF representations.
We propose a practical implementation of successor features in continuous action spaces. We first show how to learn a decomposable representation required by SF. Our proposed methods, is able to learn decoupled state and reward features representations. We study this approach on a non-trivial continuous control problems with compositional structure built into the reward functions of various tasks. | ['Animesh Garg', 'Fabio Ramos', 'David Meger', 'Dieter Fox', 'Melissa Mozifian'] | 2021-09-29 | null | null | null | null | ['robot-manipulation'] | ['robots'] | [ 2.77626395e-01 2.49303356e-01 -1.78497180e-01 -5.88399693e-02
-5.33148587e-01 -3.97960842e-01 7.77437627e-01 2.08252251e-01
-7.40604818e-01 1.02603936e+00 -2.25460138e-02 9.97215584e-02
-6.01054728e-01 -5.87054431e-01 -6.04925156e-01 -8.83366704e-01
-4.78979826e-01 4.96602833e-01 2.09771693e-01 -5.63972950e-01
3.86622071e-01 6.17368877e-01 -1.85816312e+00 -3.88900600e-02
6.23552799e-01 1.11111331e+00 6.78667486e-01 9.14339542e-01
1.83384612e-01 1.31124449e+00 -4.73534405e-01 4.56090540e-01
4.73169506e-01 -4.14691389e-01 -9.99982357e-01 1.70706987e-01
-3.03200573e-01 -4.37621146e-01 -9.46081579e-02 6.32066965e-01
3.46579492e-01 6.37421846e-01 7.15514004e-01 -1.48949277e+00
-2.09043533e-01 4.67743039e-01 -1.44228607e-01 -9.26078483e-02
2.21430063e-01 3.41308266e-01 1.04401350e+00 -1.35592043e-01
5.52051187e-01 1.33340979e+00 -7.64328754e-03 7.62744248e-01
-1.21292698e+00 -2.56057940e-02 3.03402454e-01 3.40047538e-01
-5.45263886e-01 -1.50280610e-01 4.11169499e-01 -4.95703876e-01
1.16620111e+00 -7.85581321e-02 6.76402330e-01 1.01183951e+00
4.15686399e-01 9.55377281e-01 1.32946956e+00 -4.16789412e-01
5.21200836e-01 4.83844243e-02 -9.55943167e-02 4.05711114e-01
-7.54626989e-02 6.90355241e-01 -2.50218987e-01 -5.89152239e-02
1.01179039e+00 3.21734607e-01 -4.26910967e-02 -9.34395254e-01
-1.16117465e+00 8.75226021e-01 3.93576622e-01 4.21845138e-01
-7.03263581e-01 6.47764027e-01 4.43948776e-01 8.71266842e-01
-2.72246093e-01 8.10264945e-01 -6.15612388e-01 -1.85583621e-01
-3.46233100e-01 5.73513508e-01 7.19113171e-01 7.91454911e-01
8.71163666e-01 1.62993103e-01 -3.80939394e-01 4.37743187e-01
2.08244815e-01 4.40704465e-01 6.28085792e-01 -1.30260229e+00
3.85376751e-01 2.69206733e-01 5.67002952e-01 -5.52219093e-01
-5.31207919e-01 -2.70512134e-01 -4.06882614e-01 8.89029801e-01
3.75106543e-01 -2.88762510e-01 -5.35990834e-01 2.02411437e+00
2.79870898e-01 -1.65946707e-01 4.83677745e-01 6.86019003e-01
-1.15432002e-01 5.77846825e-01 5.10059968e-02 -2.07616195e-01
8.84915054e-01 -1.08058357e+00 -6.28341615e-01 -3.24350685e-01
5.30299246e-01 -1.28305554e-01 8.43874276e-01 5.23534954e-01
-1.40704107e+00 -5.69338381e-01 -9.78150070e-01 5.45365810e-02
-2.02540278e-01 1.64714321e-01 4.00468230e-01 6.04587095e-03
-1.17835891e+00 1.25597370e+00 -9.09635246e-01 -2.21148819e-01
1.50999472e-01 7.13227153e-01 -1.88982397e-01 1.97929993e-01
-1.09856570e+00 1.34070241e+00 5.87199628e-01 2.61104591e-02
-1.46737659e+00 -5.33847101e-02 -8.04435611e-01 3.70730430e-01
8.01901042e-01 -5.87421834e-01 1.56970787e+00 -1.23468387e+00
-2.05080175e+00 1.85971618e-01 4.53436762e-01 -5.63850164e-01
5.66555738e-01 -2.62121320e-01 1.90560445e-01 1.34382471e-01
1.43319489e-02 3.61744255e-01 1.19475543e+00 -1.06503642e+00
-8.96166027e-01 -3.74207467e-01 5.12690127e-01 4.74567384e-01
-8.59956592e-02 -2.37288713e-01 4.34979439e-01 -3.50858390e-01
-5.83254576e-01 -1.00384068e+00 -7.46962249e-01 -1.54519543e-01
1.92784131e-01 -2.87551969e-01 4.09468800e-01 -1.39484376e-01
6.55401468e-01 -2.10018754e+00 8.98605049e-01 1.55621856e-01
6.82484582e-02 -6.82325289e-02 -5.45896113e-01 6.75739944e-01
3.90934721e-02 -4.74325895e-01 -1.28044665e-01 -1.93444356e-01
2.55299449e-01 6.22993112e-01 -2.43165195e-01 3.92615467e-01
4.30021346e-01 6.99039936e-01 -1.18842101e+00 -7.44264899e-03
8.58275220e-02 7.71127790e-02 -4.72268134e-01 5.81683099e-01
-4.87991452e-01 4.76585954e-01 -8.47265899e-01 4.41379026e-02
2.12602079e-01 3.74580398e-02 5.00116825e-01 5.53723037e-01
-3.66759509e-01 3.24329257e-01 -1.49594069e+00 1.73925233e+00
-5.67815244e-01 2.20822673e-02 4.91912544e-01 -1.26903284e+00
9.37768102e-01 3.25687826e-01 6.45172536e-01 -6.64124966e-01
2.78580636e-01 3.43651086e-01 2.46693064e-02 -4.21649635e-01
5.22091806e-01 -1.71294540e-01 -3.01027834e-01 6.50477290e-01
2.82292664e-01 -3.18494380e-01 2.84420401e-01 -9.37583670e-02
1.30070722e+00 5.03983557e-01 5.38738370e-01 -2.88752645e-01
6.14075482e-01 6.73970357e-02 3.82260919e-01 7.56409228e-01
-1.13810949e-01 -8.78723115e-02 8.46372843e-01 -3.09264690e-01
-8.80296767e-01 -8.26138496e-01 4.46608692e-01 1.34720612e+00
-8.56011435e-02 2.48098206e-02 -2.40617648e-01 -5.62397957e-01
2.33028963e-01 5.27459085e-01 -7.04405963e-01 -4.00028974e-01
-7.31468022e-01 -1.08212523e-01 -5.69404252e-02 4.82615173e-01
3.75517085e-03 -1.52392089e+00 -1.47088754e+00 4.68929410e-01
3.28964025e-01 -5.72425902e-01 -2.25355566e-01 8.97818208e-01
-1.04967201e+00 -1.11076820e+00 -7.63557315e-01 -6.90369308e-01
4.42414671e-01 8.57165605e-02 7.53428996e-01 -1.62081018e-01
-2.06348851e-01 8.74412596e-01 -2.79748589e-01 -2.36323461e-01
-4.67187941e-01 -5.31236939e-02 1.54592201e-01 -1.12940632e-01
-3.19562078e-01 -6.60417140e-01 -2.92648435e-01 2.31602825e-02
-1.03663671e+00 -1.12637378e-01 8.47695947e-01 1.18038738e+00
3.06567699e-01 -1.75512761e-01 6.94031239e-01 -5.72870910e-01
8.52049351e-01 -4.49182987e-01 -7.41069794e-01 1.94564328e-01
-5.64543366e-01 7.90705144e-01 8.37399483e-01 -5.13364017e-01
-8.49971831e-01 3.20507467e-01 2.52950042e-01 -2.71640480e-01
9.10600834e-03 1.99844301e-01 5.02942875e-02 7.53737614e-02
5.35071433e-01 2.73983121e-01 4.53804642e-01 -3.62808675e-01
5.54929554e-01 3.71834397e-01 2.43199781e-01 -9.42074537e-01
3.84919673e-01 -5.98853901e-02 4.05160934e-01 -3.63728195e-01
-3.08492184e-01 -3.98732007e-01 -5.87132692e-01 -2.06016406e-01
6.17020786e-01 -6.46123767e-01 -1.17868578e+00 2.96466202e-01
-1.07683790e+00 -6.70179725e-01 -9.24672484e-01 4.75489825e-01
-1.49620414e+00 1.28637210e-01 -4.35042560e-01 -1.00115180e+00
1.51890365e-03 -1.21518171e+00 7.84896374e-01 1.65614739e-01
1.84129015e-01 -6.94365501e-01 3.54104489e-01 -3.23392153e-01
5.20582378e-01 4.41864997e-01 9.06449080e-01 -5.75216234e-01
-6.00081921e-01 2.40706682e-01 1.87741920e-01 3.53094846e-01
6.64446503e-02 -5.43675303e-01 -6.65590823e-01 -6.30077958e-01
2.89939016e-01 -9.96360660e-01 7.59556174e-01 2.65832722e-01
7.16127157e-01 -3.00426334e-01 -2.18592864e-02 -3.95216933e-03
1.52641761e+00 4.69935387e-01 2.99762607e-01 5.38387775e-01
2.75210887e-01 8.61862421e-01 8.89825702e-01 8.47817540e-01
3.00652027e-01 6.58774972e-01 8.34310532e-01 4.04939890e-01
2.18062967e-01 -5.96353747e-02 7.19983876e-01 3.18337739e-01
-3.42472732e-01 2.56426096e-01 -6.31150663e-01 3.60784292e-01
-2.45538521e+00 -9.84989762e-01 3.43367696e-01 2.31157184e+00
7.31283486e-01 -8.93348008e-02 3.63620132e-01 1.32238165e-01
2.92493969e-01 -5.14904335e-02 -8.74267042e-01 -8.04703593e-01
3.79236013e-01 2.48794854e-01 4.27106887e-01 5.83857536e-01
-8.50968778e-01 6.45420671e-01 6.62533569e+00 5.51766515e-01
-8.94266605e-01 -2.18398031e-02 1.46218032e-01 -1.61116540e-01
1.49963945e-02 -1.21926025e-01 -6.27441406e-01 4.56966162e-02
9.74757433e-01 -1.99433222e-01 9.80711997e-01 1.12225199e+00
1.23765789e-01 -2.09121138e-01 -1.44479656e+00 6.44911051e-01
-3.29704046e-01 -9.04990315e-01 -3.75168115e-01 9.21028014e-03
4.97149587e-01 -1.08260185e-01 7.80265778e-02 8.43923867e-01
7.65650451e-01 -9.17039871e-01 9.23566580e-01 5.65949798e-01
4.88297850e-01 -7.29829192e-01 4.41077381e-01 6.83715165e-01
-1.09026086e+00 -8.04319561e-01 -3.64344358e-01 -5.53855598e-01
-2.52657831e-01 -1.70632005e-01 -7.56745994e-01 6.34421408e-01
2.16386318e-01 8.49452496e-01 -1.17519051e-01 6.37819469e-01
-2.57806331e-01 -3.71129006e-01 -7.06865191e-02 -4.69153792e-01
6.47332609e-01 -1.50937572e-01 3.72527599e-01 7.43049920e-01
2.44028971e-01 -1.41670972e-01 5.86812913e-01 8.15519929e-01
4.94915515e-01 -3.13573889e-02 -7.72257030e-01 -1.24945991e-01
-6.61486164e-02 1.09689867e+00 -3.49480689e-01 -1.33476138e-01
-1.58522338e-01 7.28523791e-01 8.88720691e-01 2.31719017e-01
-5.71463466e-01 -2.63993293e-01 7.05281734e-01 -2.53267080e-01
4.36293960e-01 -4.88916934e-01 2.12159544e-01 -9.78646278e-01
-8.33601505e-02 -9.64720428e-01 2.29501471e-01 -4.79171008e-01
-9.70044613e-01 5.13212621e-01 2.92306617e-02 -1.22370374e+00
-7.79029667e-01 -8.04857194e-01 -2.59950250e-01 9.25886095e-01
-1.82294524e+00 -7.23839462e-01 1.08056732e-01 9.06819940e-01
5.75270414e-01 -2.42009819e-01 8.07880521e-01 -2.33632967e-01
-3.32978219e-01 -1.50716882e-02 4.31084752e-01 -4.73954022e-01
4.00072932e-01 -1.60810876e+00 -1.41877085e-01 3.15039814e-01
-3.80381554e-01 1.98218092e-01 6.20757580e-01 -3.57996017e-01
-1.63523388e+00 -7.20304251e-01 4.58066612e-01 -1.02120727e-01
8.10304105e-01 -1.06682695e-01 -4.45338398e-01 6.46899104e-01
1.39462963e-01 -7.99439922e-02 2.12885171e-01 -9.00664926e-02
1.03377327e-01 -1.63939968e-01 -1.05987465e+00 4.34958041e-01
7.17752993e-01 -1.73824787e-01 -8.08160067e-01 7.50768781e-02
5.65405905e-01 -2.34313115e-01 -6.91461205e-01 -6.01326004e-02
4.91728753e-01 -7.88064420e-01 8.62495065e-01 -9.11373854e-01
5.40154397e-01 -1.21483557e-01 -1.06741190e-01 -1.70570338e+00
-5.72892070e-01 -6.71082437e-01 -3.19568336e-01 5.99330068e-01
1.56072900e-01 -6.18506491e-01 4.39087689e-01 6.46402836e-01
-7.89010376e-02 -8.62198532e-01 -9.71660078e-01 -9.27406669e-01
1.01719357e-01 8.11789557e-02 4.85820562e-01 4.60846603e-01
2.63448477e-01 3.25108796e-01 -5.90554416e-01 -2.03394651e-01
3.56042534e-01 2.04666391e-01 6.82824850e-01 -1.19659507e+00
-8.97155344e-01 -4.42110002e-01 -2.10352093e-01 -1.10177219e+00
2.16011465e-01 -8.01868737e-01 3.40461493e-01 -1.45633960e+00
-3.54709267e-03 -5.90890169e-01 -5.57577848e-01 7.06908762e-01
1.45535499e-01 -8.84164572e-01 7.76268125e-01 1.97248340e-01
-7.30518818e-01 8.55646908e-01 1.59756374e+00 -8.90787169e-02
-4.18553799e-01 1.63409263e-01 -4.09103274e-01 4.45599258e-01
9.29745197e-01 -5.82968056e-01 -6.12010002e-01 -4.61223185e-01
2.15785623e-01 6.73596323e-01 2.19219357e-01 -1.00385666e+00
8.79948363e-02 -5.89431345e-01 1.10844694e-01 -5.89536466e-02
4.44778562e-01 -1.11732864e+00 -1.42736658e-01 9.72574413e-01
-7.48992801e-01 2.04466566e-01 -3.90291400e-02 8.10823143e-01
-7.57657662e-02 -6.30932450e-01 7.08527863e-01 -4.95445758e-01
-6.68015778e-01 9.44862887e-02 -5.99593639e-01 -4.06331718e-02
1.32543755e+00 1.73544601e-01 8.76354054e-03 -1.15014061e-01
-1.06615150e+00 5.87777078e-01 1.63892657e-01 3.93960238e-01
5.10709763e-01 -1.18627155e+00 -4.24402863e-01 7.81990066e-02
-1.73580855e-01 -1.52297989e-01 -6.42889924e-03 6.68870986e-01
-3.82482931e-02 3.12039346e-01 -9.54695344e-01 -2.34216154e-01
-9.19819295e-01 7.96421409e-01 3.36239010e-01 -9.05897796e-01
-4.87461686e-01 3.52274477e-01 -8.00502002e-02 -2.91852236e-01
3.13985765e-01 -7.17201531e-01 -5.75775385e-01 1.90628365e-01
4.42044467e-01 5.36498964e-01 -2.62624294e-01 -1.06649324e-01
-1.29988352e-02 3.92818034e-01 -6.65630028e-02 -4.19562936e-01
1.60052752e+00 -6.49488941e-02 5.95205557e-03 6.36909008e-01
7.77359664e-01 -7.50655949e-01 -1.61636603e+00 -2.77746528e-01
1.45021096e-01 -3.48088562e-01 -8.38175863e-02 -7.92178094e-01
-6.80133045e-01 8.04199874e-01 4.96031880e-01 3.66816044e-01
1.09181821e+00 -3.22882563e-01 4.26200718e-01 8.24984610e-01
8.36562812e-01 -1.32976234e+00 5.27570784e-01 8.53338361e-01
1.09118283e+00 -1.05883706e+00 -2.26448461e-01 2.65536517e-01
-9.06984270e-01 1.45303643e+00 4.25302446e-01 -6.02273643e-01
3.07963401e-01 1.93535447e-01 -4.27698731e-01 1.00501463e-01
-1.08825946e+00 -5.68312645e-01 -2.01924562e-01 7.14619637e-01
1.89591143e-02 1.29922271e-01 -3.17468256e-01 5.88485718e-01
3.33702505e-01 1.52984023e-01 6.22691035e-01 1.36929095e+00
-8.09826493e-01 -1.46749580e+00 -3.60218644e-01 3.08149844e-01
-1.02176391e-01 3.73461783e-01 -9.72567797e-02 6.65492892e-01
2.82518789e-02 7.16477811e-01 -2.27932334e-01 -2.30319649e-02
4.05162603e-01 -2.76951808e-02 8.03515613e-01 -8.22425365e-01
-8.17536354e-01 -6.04778938e-02 -7.85967410e-02 -8.67408812e-01
-4.51115191e-01 -8.05051684e-01 -1.38535237e+00 1.55287206e-01
1.77582443e-01 1.74518228e-01 5.69893837e-01 9.23248053e-01
2.50214517e-01 8.08160663e-01 1.07295620e+00 -1.14893723e+00
-1.49194288e+00 -7.38455117e-01 -9.99914408e-01 2.56864071e-01
7.21219182e-01 -9.84042943e-01 -1.52759194e-01 -3.41252118e-01] | [4.2016072273254395, 1.7755732536315918] |
899aa16a-fb69-485f-a318-2e3052eb517f | supertagging-combinatory-categorial-grammar | 2010.06115 | null | https://arxiv.org/abs/2010.06115v2 | https://arxiv.org/pdf/2010.06115v2.pdf | Supertagging Combinatory Categorial Grammar with Attentive Graph Convolutional Networks | Supertagging is conventionally regarded as an important task for combinatory categorial grammar (CCG) parsing, where effective modeling of contextual information is highly important to this task. However, existing studies have made limited efforts to leverage contextual features except for applying powerful encoders (e.g., bi-LSTM). In this paper, we propose attentive graph convolutional networks to enhance neural CCG supertagging through a novel solution of leveraging contextual information. Specifically, we build the graph from chunks (n-grams) extracted from a lexicon and apply attention over the graph, so that different word pairs from the contexts within and across chunks are weighted in the model and facilitate the supertagging accordingly. The experiments performed on the CCGbank demonstrate that our approach outperforms all previous studies in terms of both supertagging and parsing. Further analyses illustrate the effectiveness of each component in our approach to discriminatively learn from word pairs to enhance CCG supertagging. | ['Fei Xia', 'Yan Song', 'Yuanhe Tian'] | 2020-10-13 | null | https://aclanthology.org/2020.emnlp-main.487 | https://aclanthology.org/2020.emnlp-main.487.pdf | emnlp-2020-11 | ['ccg-supertagging'] | ['natural-language-processing'] | [ 2.52699554e-01 6.31958961e-01 -1.01396538e-01 -6.09173000e-01
-8.36664677e-01 -3.45711648e-01 4.84150887e-01 2.92363524e-01
-4.24699396e-01 4.95827258e-01 6.35930419e-01 -5.55126011e-01
1.39919028e-01 -9.45231616e-01 -8.12281430e-01 -4.75794613e-01
-2.23103151e-01 2.29607597e-01 2.53384486e-02 -2.09136978e-01
1.41656831e-01 -1.79874435e-01 -1.11416042e+00 2.60045111e-01
1.23357630e+00 7.85508037e-01 7.55896568e-01 4.76190597e-01
-5.66355109e-01 7.98011601e-01 -6.84700310e-01 -7.07817733e-01
-1.47165343e-01 -4.59551185e-01 -8.43591988e-01 -2.00338379e-01
4.57020521e-01 -2.27124885e-01 -2.08362207e-01 1.07757199e+00
4.31522012e-01 1.99765027e-01 2.43756503e-01 -6.93480074e-01
-9.83393013e-01 1.60550809e+00 -1.52309209e-01 1.49667457e-01
1.76186010e-03 -2.44286209e-01 1.72572231e+00 -6.83925092e-01
3.61277014e-01 1.45105612e+00 6.41603649e-01 6.68298066e-01
-6.88830972e-01 -6.48513734e-01 7.50592291e-01 3.74518484e-02
-8.29725564e-01 -1.70867577e-01 8.39957356e-01 -1.49048701e-01
1.22156012e+00 8.08052123e-02 6.51198804e-01 8.81949246e-01
2.45920554e-01 9.30531383e-01 5.92974484e-01 -6.07337773e-01
9.83124971e-02 -5.50545633e-01 4.77406770e-01 8.71195555e-01
3.07121336e-01 -6.76974356e-02 -5.11598170e-01 9.77529809e-02
5.91093719e-01 -1.81278214e-01 8.45847428e-02 7.18669966e-02
-8.88632596e-01 1.12743449e+00 7.11256325e-01 5.94693363e-01
-2.50176340e-01 6.06052220e-01 4.98097062e-01 4.69908975e-02
8.57643068e-01 6.00777209e-01 -5.53383887e-01 -4.24293578e-02
-2.18373701e-01 -7.12189451e-02 4.90628481e-01 1.23833692e+00
8.21605682e-01 1.50037214e-01 -2.12960735e-01 9.35715199e-01
6.15198612e-01 1.64244398e-01 5.60091197e-01 -5.32173991e-01
8.94064009e-01 8.74859571e-01 -6.34119630e-01 -8.15622389e-01
-4.89476442e-01 -6.65917397e-01 -5.71955383e-01 -6.01035953e-01
1.31199747e-01 -3.54859293e-01 -1.13198996e+00 2.23090911e+00
1.56566471e-01 2.70570904e-01 -3.06650531e-02 6.59537673e-01
1.25361538e+00 4.70544487e-01 8.61295283e-01 1.66946337e-01
1.43031454e+00 -8.74754906e-01 -7.89846063e-01 -8.38552415e-01
1.22611821e+00 -3.85961741e-01 1.26125979e+00 -1.27787381e-01
-9.20940518e-01 -4.56072509e-01 -9.08706725e-01 -3.93706471e-01
-4.01765555e-01 -1.11426108e-01 1.05519092e+00 3.89154732e-01
-1.04763937e+00 5.44247687e-01 -8.86370957e-01 -3.74747425e-01
2.35358492e-01 2.80315220e-01 -2.45370120e-01 -9.24391113e-03
-1.65683103e+00 7.33285129e-01 7.92426884e-01 2.46582031e-01
-5.18171191e-01 -5.07468879e-01 -1.23967040e+00 2.03264937e-01
3.21294993e-01 -6.36555493e-01 1.27069902e+00 -7.43971646e-01
-1.37368345e+00 7.83164859e-01 -2.58572530e-02 -5.44599831e-01
-3.52480918e-01 -3.21914941e-01 -1.05866387e-01 -7.71684274e-02
9.80611965e-02 6.83591664e-01 4.54993248e-01 -8.51372182e-01
-6.50871634e-01 -5.83177507e-01 3.89584810e-01 5.08670151e-01
-3.39201927e-01 1.17617682e-01 -4.95380372e-01 -8.18689287e-01
3.57810557e-01 -7.63181746e-01 -3.45185727e-01 -1.00085509e+00
-6.13599658e-01 -7.32822299e-01 5.80065668e-01 -9.59911227e-01
1.54289341e+00 -2.02864933e+00 2.70126879e-01 -2.33494118e-01
4.41572547e-01 2.30027407e-01 -1.84561178e-01 6.09223247e-01
2.31054664e-01 4.78872865e-01 -1.22506738e-01 -5.97482622e-01
2.95491964e-01 4.00343210e-01 -1.07281044e-01 -1.72109324e-02
1.98894903e-01 1.38684535e+00 -1.10631227e+00 -4.22336787e-01
1.13997996e-01 1.44103393e-01 -5.91294944e-01 3.44374686e-01
-6.04241669e-01 1.67163938e-01 -6.39202595e-01 4.22626406e-01
4.05837148e-01 -2.33835638e-01 7.09641218e-01 3.53298225e-02
1.44037724e-01 9.90321279e-01 -3.56695771e-01 1.80069232e+00
-8.33748758e-01 2.35620096e-01 -2.21472532e-01 -1.04600239e+00
8.05297136e-01 1.21856377e-01 -1.36984766e-01 -7.79253483e-01
1.98661745e-01 2.34133616e-01 3.81013662e-01 -2.49584705e-01
7.32779801e-01 -6.06837384e-02 -6.02615535e-01 4.55683589e-01
3.64426434e-01 6.74858131e-03 -3.69995907e-02 5.57081878e-01
1.14479029e+00 6.92410320e-02 4.54993278e-01 -2.50164777e-01
2.97318809e-02 -3.40703815e-01 6.35383427e-01 6.88309789e-01
2.89542563e-02 2.15378508e-01 5.77063799e-01 -3.06787372e-01
-6.72056377e-01 -5.52319348e-01 3.42876911e-01 1.72862959e+00
9.45466105e-04 -7.38093317e-01 -9.68875647e-01 -1.08018589e+00
-9.39963683e-02 7.35871077e-01 -7.86146939e-01 -3.27877998e-01
-1.05075395e+00 -7.45502591e-01 5.15880466e-01 8.58096540e-01
4.18099582e-01 -1.32408667e+00 -2.03946188e-01 5.70174992e-01
-6.27319589e-02 -1.22535169e+00 -5.81076324e-01 4.65052724e-01
-9.57044125e-01 -8.21054816e-01 2.70906067e-03 -1.11076915e+00
4.09935027e-01 2.84728408e-01 1.44746745e+00 4.14280027e-01
2.10190162e-01 -9.47127789e-02 -7.91780114e-01 -2.91997463e-01
-4.10733372e-01 6.67354763e-01 -5.34753501e-01 -3.74335110e-01
5.50909460e-01 -5.01124561e-01 -3.97517741e-01 -1.49232775e-01
-7.55272746e-01 3.03012341e-01 6.35586977e-01 8.78243566e-01
3.08208436e-01 -5.45730233e-01 8.41975272e-01 -1.51245368e+00
7.11493790e-01 -5.90622962e-01 -5.27051389e-01 2.78620422e-01
-2.77570873e-01 1.56262413e-01 7.12605357e-01 -7.62074739e-02
-9.95642245e-01 -2.09444344e-01 -4.11973566e-01 1.63462967e-01
1.15744404e-01 9.63100135e-01 -7.11470723e-01 1.22541845e-01
8.94648656e-02 -3.93772125e-02 -3.82322580e-01 -5.75621903e-01
7.33497679e-01 6.35568619e-01 1.72335431e-01 -6.03483200e-01
3.72563511e-01 -1.34457082e-01 -3.13435107e-01 -5.64987123e-01
-1.10420740e+00 -2.27549598e-01 -5.18362105e-01 1.51443845e-02
1.06816947e+00 -7.30181515e-01 -3.33951771e-01 5.67110658e-01
-1.15207100e+00 -5.92817128e-01 2.15899900e-01 4.65074182e-01
-1.03803955e-01 3.85937214e-01 -8.49950969e-01 -5.88811755e-01
-5.48396647e-01 -9.41037118e-01 1.28750515e+00 8.15483928e-02
1.85870826e-02 -1.44746077e+00 -1.70700401e-01 2.48563394e-01
4.42134023e-01 3.55486162e-02 1.39770854e+00 -9.01948392e-01
-6.22138560e-01 8.87392759e-02 -2.62529254e-01 1.61833867e-01
2.35547662e-01 -4.04711276e-01 -9.87698734e-01 -1.30900830e-01
-4.56134200e-01 -2.48654708e-01 1.06722188e+00 4.36789781e-01
1.08378875e+00 -4.39895123e-01 -3.05540472e-01 6.73558116e-01
1.25590265e+00 2.49705762e-01 5.50587893e-01 1.56507060e-01
1.17626679e+00 5.98092556e-01 6.57814145e-01 2.81456206e-02
8.51957500e-01 6.07930899e-01 7.57510424e-01 1.03647172e-01
-3.27480257e-01 -5.94592929e-01 2.52463877e-01 1.32884598e+00
1.55932039e-01 -4.69581217e-01 -8.00725043e-01 6.28629029e-01
-1.81144500e+00 -3.73038054e-01 -5.12102768e-02 1.79197907e+00
8.01042259e-01 6.16611391e-02 -3.72158080e-01 -1.81398094e-01
9.42318439e-01 5.34250021e-01 -3.27961147e-01 -5.09376824e-01
6.14719167e-02 3.78453135e-01 3.19327772e-01 6.04359627e-01
-1.18321216e+00 1.76725793e+00 5.50305748e+00 8.26108754e-01
-8.73262107e-01 2.67260224e-01 5.62872887e-01 4.00498986e-01
-7.23854959e-01 1.31092235e-01 -9.64698732e-01 5.36552072e-01
1.01977324e+00 8.58143419e-02 2.42913112e-01 7.64843762e-01
-1.34129852e-01 1.30446389e-01 -7.81457126e-01 4.50803429e-01
-8.18997324e-02 -1.10442448e+00 7.82490820e-02 2.57647578e-02
3.98308873e-01 1.42094940e-01 -1.07875243e-01 7.34073937e-01
8.57181132e-01 -8.42246294e-01 6.32752597e-01 1.26834575e-03
6.16940796e-01 -5.82365990e-01 8.17303181e-01 1.55222505e-01
-1.43446898e+00 2.88597792e-02 -5.85753798e-01 -2.16097638e-01
2.30689734e-01 5.32648563e-01 -1.06955981e+00 5.85490644e-01
3.53764325e-01 7.00114429e-01 -5.81494749e-01 5.19981682e-01
-7.34544396e-01 1.07959366e+00 4.92172455e-03 -4.04844403e-01
5.74946702e-01 -1.90129817e-01 2.34505907e-01 1.21218312e+00
3.24093878e-01 1.31418183e-01 2.36578494e-01 5.00084221e-01
-3.90713066e-01 3.38886946e-01 -6.87804222e-01 -2.70578772e-01
7.93210745e-01 1.22102129e+00 -5.67767680e-01 -4.07503307e-01
-5.74252427e-01 6.18131280e-01 1.03150952e+00 1.22877516e-01
-6.83077574e-01 -4.02720273e-01 6.83501720e-01 -2.49376982e-01
3.59703422e-01 -4.14990336e-01 -3.52358073e-01 -1.09239829e+00
-1.55426756e-01 -5.39130628e-01 7.00624883e-01 -5.79928398e-01
-1.17939675e+00 8.97099197e-01 -1.08611859e-01 -6.83694422e-01
-4.12952006e-01 -5.25049031e-01 -6.40127003e-01 8.42939556e-01
-1.45736897e+00 -1.75361443e+00 -1.52503058e-01 2.30899289e-01
5.10558844e-01 4.51621823e-02 7.36983895e-01 1.81245506e-01
-5.40873885e-01 8.02685797e-01 -4.37230080e-01 3.06730717e-01
3.10112387e-01 -1.60234869e+00 1.21164513e+00 1.18028760e+00
3.00558925e-01 9.15185869e-01 3.05770427e-01 -8.94280851e-01
-1.27239382e+00 -1.46356261e+00 1.30390680e+00 -1.94302395e-01
7.77637124e-01 -7.98908770e-01 -1.08683729e+00 1.06607890e+00
2.62500703e-01 -1.70486972e-01 6.50984883e-01 6.16307795e-01
-3.78932804e-01 1.79921955e-01 -6.36000454e-01 5.78347683e-01
1.61063302e+00 -6.01561368e-01 -7.78177500e-01 8.05572197e-02
1.46835291e+00 -4.57143128e-01 -7.36733198e-01 2.15982676e-01
2.54139125e-01 -6.07464790e-01 4.89795327e-01 -8.83401453e-01
4.57072198e-01 6.74649850e-02 -2.21531868e-01 -1.74334884e+00
-4.90165085e-01 -5.13954937e-01 -1.79079324e-01 1.34563208e+00
4.81377035e-01 -7.93189824e-01 7.01904893e-01 2.88576126e-01
-1.03979635e+00 -6.73393428e-01 -8.00620973e-01 -5.35422027e-01
2.60499507e-01 -5.38979828e-01 1.03729153e+00 8.31124306e-01
2.96457440e-01 7.19277322e-01 -3.61049175e-01 2.35137045e-01
2.98421264e-01 1.91348016e-01 4.71656680e-01 -1.04989874e+00
-2.60496527e-01 -1.57923356e-01 -3.93560141e-01 -1.19713461e+00
5.39498627e-01 -1.24106288e+00 2.99501538e-01 -1.86432111e+00
1.05973355e-01 -7.18908072e-01 -3.07297766e-01 6.84492767e-01
-8.98727179e-01 5.44389337e-02 1.86903238e-01 -3.37148070e-01
-7.93664694e-01 4.50350493e-01 1.25662184e+00 4.78366055e-02
5.68110123e-02 -3.05318683e-01 -1.23386228e+00 6.64745092e-01
8.66879225e-01 -2.72105426e-01 -4.79517788e-01 -1.08817506e+00
4.35986131e-01 -5.89874387e-02 6.49241060e-02 -5.79342127e-01
9.80540179e-03 1.44412061e-02 -2.49451652e-01 -3.47554624e-01
7.42239207e-02 -3.39755684e-01 -2.90576547e-01 2.01308474e-01
-3.82849514e-01 1.20481685e-01 1.32097155e-02 5.21725833e-01
-2.84322083e-01 -2.67661870e-01 3.46468627e-01 -3.15113008e-01
-9.75730181e-01 4.29412425e-01 -1.16544858e-01 4.17010069e-01
5.42652667e-01 9.97833610e-02 -6.71566665e-01 -3.54978353e-01
-6.06391907e-01 1.45766795e-01 1.22219309e-01 4.14925903e-01
4.84498918e-01 -1.30291450e+00 -5.07885754e-01 2.01178417e-01
2.52662957e-01 1.23324484e-01 2.96658009e-01 3.57345670e-01
-2.27371797e-01 5.26561081e-01 1.99440777e-01 -1.62804961e-01
-1.12854779e+00 3.94517839e-01 1.68578759e-01 -5.91447711e-01
-6.47979498e-01 1.20150685e+00 6.02355838e-01 -5.79924285e-01
1.71717495e-01 -9.05158043e-01 -4.81709063e-01 -6.18804023e-02
2.63118446e-01 -1.36872962e-01 3.50311905e-01 -5.58477879e-01
-2.02664092e-01 3.81587446e-01 -2.54498869e-01 3.18646491e-01
1.39092124e+00 -2.31375724e-01 -1.75794065e-01 1.83712095e-01
1.14594841e+00 8.67496654e-02 -1.07208204e+00 -3.39217097e-01
4.36151236e-01 -1.96042940e-01 -8.63351971e-02 -6.30845726e-01
-1.08525014e+00 1.01633656e+00 -1.69507101e-01 5.22266626e-01
9.45750892e-01 3.91805202e-01 1.19287491e+00 7.61164427e-02
5.13780534e-01 -7.50846982e-01 -9.66243669e-02 9.92848516e-01
6.59051120e-01 -1.11078179e+00 -5.51269412e-01 -7.65563965e-01
-5.94850302e-01 8.12268496e-01 7.19646692e-01 -1.93698704e-01
4.26990718e-01 1.05818987e-01 1.05237700e-01 -3.15176219e-01
-9.30669427e-01 -5.10862231e-01 2.91436553e-01 5.00260234e-01
7.53554702e-01 4.08400327e-01 -6.22107208e-01 1.01792455e+00
-4.74678159e-01 -5.53461432e-01 2.62389749e-01 7.68578291e-01
-5.13459921e-01 -1.28460860e+00 2.90342480e-01 5.93302906e-01
-7.07125127e-01 -6.82247162e-01 -3.92404944e-01 8.35652769e-01
2.38025393e-02 9.51367140e-01 -3.85335311e-02 -5.46428442e-01
5.73379397e-02 8.91263336e-02 3.10840696e-01 -1.22758508e+00
-8.86636198e-01 5.91510674e-03 5.95161617e-01 -5.30694664e-01
-2.89698333e-01 -4.15011138e-01 -1.33371425e+00 5.32875434e-02
-4.60642427e-01 3.49853247e-01 5.43302059e-01 1.10370553e+00
4.54966247e-01 8.61878574e-01 2.96533316e-01 -5.26802123e-01
-3.32546085e-01 -1.46375740e+00 -6.85413480e-01 3.58477056e-01
-9.24482662e-03 -6.71923280e-01 -2.35168189e-01 -3.69274318e-01] | [10.505233764648438, 9.374051094055176] |
1e79d3e6-1872-44e4-8eb3-de56d30976c7 | multi-plane-program-induction-with-3d-box-1 | 2011.10007 | null | https://arxiv.org/abs/2011.10007v2 | https://arxiv.org/pdf/2011.10007v2.pdf | Multi-Plane Program Induction with 3D Box Priors | We consider two important aspects in understanding and editing images: modeling regular, program-like texture or patterns in 2D planes, and 3D posing of these planes in the scene. Unlike prior work on image-based program synthesis, which assumes the image contains a single visible 2D plane, we present Box Program Induction (BPI), which infers a program-like scene representation that simultaneously models repeated structure on multiple 2D planes, the 3D position and orientation of the planes, and camera parameters, all from a single image. Our model assumes a box prior, i.e., that the image captures either an inner view or an outer view of a box in 3D. It uses neural networks to infer visual cues such as vanishing points, wireframe lines to guide a search-based algorithm to find the program that best explains the image. Such a holistic, structured scene representation enables 3D-aware interactive image editing operations such as inpainting missing pixels, changing camera parameters, and extrapolate the image contents. | ['Joshua B. Tenenbaum', 'William T. Freeman', 'Jiajun Wu', 'Noah Snavely', 'Xiuming Zhang', 'Jiayuan Mao', 'Yikai Li'] | 2020-11-19 | multi-plane-program-induction-with-3d-box | http://proceedings.neurips.cc/paper/2020/hash/5301c4d888f5204274439e6dcf5fdb54-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/5301c4d888f5204274439e6dcf5fdb54-Paper.pdf | neurips-2020-12 | ['program-induction'] | ['computer-code'] | [ 7.75780976e-01 2.06994593e-01 -1.07626982e-01 -5.07722020e-01
-9.17802900e-02 -7.39274621e-01 6.26912415e-01 -1.01620428e-01
1.67312786e-01 -1.27484888e-01 1.67341873e-01 -5.53750038e-01
4.01157558e-01 -7.66281843e-01 -1.38886511e+00 8.67555477e-03
3.65079761e-01 3.82356763e-01 1.62383810e-01 1.25707716e-01
7.76213408e-01 8.19292128e-01 -1.53470004e+00 6.18653655e-01
2.47664601e-01 4.56266433e-01 3.78964514e-01 1.24529827e+00
-2.76022553e-01 8.72677684e-01 -6.14638984e-01 -1.95822984e-01
3.99578482e-01 -4.21493888e-01 -5.44589758e-01 8.64145041e-01
8.07956874e-01 -8.29754174e-01 -3.77415746e-01 9.86527920e-01
-1.94246396e-01 -1.39487192e-01 2.65541375e-01 -1.04273677e+00
-6.42737567e-01 1.31467748e-02 -8.95769000e-01 -4.93120164e-01
1.02840841e+00 3.81975800e-01 6.78240478e-01 -9.53010023e-01
1.09569633e+00 1.13021278e+00 6.92411125e-01 1.92981616e-01
-1.71529365e+00 3.86862159e-02 1.23899072e-01 -3.75344306e-01
-1.23616600e+00 -3.81477326e-01 1.07573080e+00 -8.48214567e-01
9.96892154e-01 3.01084042e-01 9.19643879e-01 5.65250337e-01
4.91706401e-01 5.48015833e-01 8.71032000e-01 -9.67803717e-01
7.55651817e-02 2.17600182e-01 -7.40565583e-02 8.59977484e-01
-2.12328956e-01 2.43608728e-01 -6.77412331e-01 -1.44939631e-01
1.59370935e+00 5.08646555e-02 -3.28269005e-01 -1.02767181e+00
-1.28708959e+00 4.26259190e-01 -2.93536838e-02 -3.32906604e-01
-3.39558035e-01 2.43721560e-01 4.72577773e-02 1.43326730e-01
2.33777508e-01 6.59860313e-01 -3.89794350e-01 2.90421173e-02
-1.17145491e+00 5.58187962e-01 7.06570566e-01 1.45634055e+00
1.08455765e+00 7.25107267e-02 1.81948707e-01 2.59904325e-01
9.65065733e-02 2.90679455e-01 -2.79928446e-01 -1.57597244e+00
3.89606297e-01 6.90537751e-01 1.81674048e-01 -1.03777719e+00
-1.17956087e-01 2.47931674e-01 -1.38927519e-01 7.83893049e-01
2.63252079e-01 8.62106532e-02 -8.93804491e-01 1.22398889e+00
3.07263553e-01 -1.71319708e-01 -3.04886013e-01 6.45938158e-01
6.25873804e-01 7.89815187e-01 -6.47255361e-01 1.70799613e-01
1.13478303e+00 -8.91836703e-01 -4.73439902e-01 -4.56283271e-01
5.05460262e-01 -9.88575459e-01 1.23528123e+00 3.57530653e-01
-1.53733122e+00 -6.37977660e-01 -1.00745809e+00 -4.55065161e-01
1.80652775e-02 1.24743097e-01 5.52209318e-01 4.32760984e-01
-1.22038412e+00 4.61522818e-01 -6.33543193e-01 -1.58647090e-01
2.05912560e-01 7.03743771e-02 -4.01276767e-01 -1.54932171e-01
-2.22094223e-01 5.72912097e-01 1.75700560e-01 -1.13539629e-01
-8.96987379e-01 -9.64669228e-01 -8.95504177e-01 -2.79229647e-03
4.15908754e-01 -9.73799050e-01 1.32801831e+00 -1.28722596e+00
-1.51234829e+00 1.35362983e+00 -6.85163856e-01 -1.12233154e-01
3.21297407e-01 -3.47591251e-01 3.45510334e-01 2.20752969e-01
4.14629988e-02 8.82422745e-01 8.96929860e-01 -1.59969103e+00
-1.96041793e-01 -5.88472188e-01 4.80573297e-01 5.19027770e-01
3.85154784e-01 1.35968775e-01 -8.99678707e-01 -3.62520427e-01
4.47309643e-01 -6.93219006e-01 -1.23125605e-01 7.05918670e-01
-7.99504578e-01 4.38148677e-01 1.06993604e+00 -8.22278023e-01
8.97919595e-01 -2.14744353e+00 3.07196289e-01 3.67108166e-01
5.60058132e-02 -3.61085445e-01 7.51150697e-02 4.29537684e-01
-3.36061180e-01 1.76848277e-01 -1.06872834e-01 -5.05747855e-01
-3.05411100e-01 8.33170861e-02 -4.06427413e-01 2.65378863e-01
1.68480068e-01 5.25922775e-01 -4.78496283e-01 -2.15373471e-01
5.93156457e-01 3.83193165e-01 -9.73155260e-01 5.06153822e-01
-6.40993416e-01 4.70080644e-01 -1.32601723e-01 4.83528346e-01
8.00467789e-01 -9.73332152e-02 6.05325624e-02 -2.78618276e-01
-4.64089990e-01 1.49503127e-01 -1.10341156e+00 1.93625557e+00
-4.29991186e-01 1.02575028e+00 1.45042151e-01 -4.81352419e-01
9.06934679e-01 -1.64654236e-02 2.12158918e-01 -4.74791467e-01
-1.86773747e-01 -4.70718890e-01 -6.50449216e-01 -5.28338790e-01
8.51460040e-01 4.28383738e-01 3.22937500e-03 7.53069282e-01
-3.80723417e-01 -1.12111270e+00 -5.27660474e-02 2.92152613e-01
8.30443680e-01 8.35525811e-01 2.78980762e-01 -5.40406890e-02
4.23117541e-02 3.51223826e-01 2.33607497e-02 9.44601595e-01
2.98375607e-01 9.26131785e-01 9.22818720e-01 -7.17688322e-01
-1.58937788e+00 -8.07205915e-01 2.21875176e-01 7.59907484e-01
3.45409065e-01 -4.74766821e-01 -9.43608463e-01 -1.11814499e-01
-2.78338373e-01 1.14588523e+00 -7.19539940e-01 2.49846280e-01
-6.81865752e-01 1.84156120e-01 4.52460237e-02 2.08962500e-01
2.56216586e-01 -5.54292738e-01 -1.19586670e+00 -2.96144754e-01
2.12435592e-02 -1.01517379e+00 -9.44896996e-01 1.35288239e-01
-1.07445145e+00 -1.12215006e+00 -2.76575655e-01 -8.41527164e-01
1.26735544e+00 6.16749346e-01 1.39405429e+00 1.68532357e-02
-2.87233144e-01 7.12210417e-01 6.54520914e-02 -1.87179551e-01
-4.94843334e-01 -6.89890862e-01 -6.56530619e-01 -2.84367204e-01
1.12828054e-02 -5.23221493e-01 -2.97794610e-01 1.29634053e-01
-9.06845808e-01 9.74829376e-01 4.80950773e-02 5.50690770e-01
9.07114863e-01 1.56265840e-01 -7.91575968e-01 -1.10872853e+00
2.21221089e-01 7.12313950e-02 -9.18787360e-01 1.82151794e-01
5.34380414e-02 -4.19332869e-02 4.44041818e-01 -2.40008533e-01
-1.36203122e+00 4.66057360e-01 3.66548806e-01 -7.57007539e-01
-4.65728790e-01 2.31014907e-01 -2.63753533e-01 1.83849946e-01
7.48525083e-01 2.70442754e-01 -1.57916337e-01 -9.40947905e-02
6.52468741e-01 3.84164080e-02 7.33043432e-01 -7.37959802e-01
8.35889816e-01 8.56581986e-01 -2.49541879e-01 -1.04371953e+00
-4.51615542e-01 -1.52241141e-01 -1.24721241e+00 -1.77183121e-01
9.88268733e-01 -7.70758688e-01 -4.54574436e-01 3.73759896e-01
-1.89432228e+00 -6.33618176e-01 -2.88893491e-01 -3.75461020e-02
-1.00601649e+00 1.73829690e-01 -3.11960578e-01 -6.07666016e-01
2.30229557e-01 -1.47536528e+00 1.46102667e+00 5.90313450e-02
-6.83736026e-01 -8.83295417e-01 -1.33504465e-01 2.54719794e-01
-1.48113504e-01 4.19347346e-01 1.46027732e+00 3.43447775e-01
-1.18229103e+00 -2.75660425e-01 -1.77467093e-01 1.03737945e-02
-1.27532735e-01 6.98770702e-01 -7.59164274e-01 4.00958389e-01
3.09879463e-02 1.06954694e-01 -1.14524933e-04 7.12150574e-01
1.42152619e+00 -2.11479768e-01 -3.42851341e-01 1.06470001e+00
1.62504303e+00 7.85852790e-01 7.53517807e-01 2.78373510e-01
9.04733360e-01 7.99291134e-01 2.30771914e-01 4.98403192e-01
3.17965776e-01 6.97727084e-01 4.73825365e-01 -3.01012784e-01
-1.37395382e-01 -7.70512819e-01 1.76105052e-01 1.49000287e-01
2.49906093e-01 -6.13106899e-02 -9.30356026e-01 3.40036303e-01
-1.41498148e+00 -8.23892593e-01 -2.57966906e-01 2.39815903e+00
5.01329899e-01 2.27920949e-01 -1.59350157e-01 -3.16580564e-01
7.75003791e-01 2.01757342e-01 -6.92623556e-01 -8.97969246e-01
-3.27125043e-02 -7.42578357e-02 5.40753901e-01 7.66712725e-01
-7.36194432e-01 9.26997185e-01 7.02440262e+00 4.72079329e-02
-9.78855252e-01 -3.61675888e-01 7.11741984e-01 -3.96248102e-02
-8.46135378e-01 6.73634291e-01 -5.91757357e-01 -3.11314929e-02
3.10194045e-01 3.86534296e-02 8.73703659e-01 8.89892697e-01
5.48930645e-01 -4.66811627e-01 -1.63668668e+00 9.21082139e-01
3.44124287e-01 -1.66608644e+00 1.74702704e-01 -2.66378727e-02
1.06462228e+00 -3.29608142e-01 -2.63347235e-02 -2.56589204e-01
4.01767671e-01 -7.65837848e-01 1.20466042e+00 5.48605323e-01
9.90433753e-01 -4.63952214e-01 -1.35041460e-01 6.03075743e-01
-7.06085503e-01 4.14592415e-01 -3.05536222e-02 -2.03024566e-01
2.35580847e-01 4.47434574e-01 -8.31133425e-01 -4.12168168e-02
6.27651095e-01 4.95773852e-01 -3.51664215e-01 5.87851882e-01
-3.24535251e-01 -3.82363945e-02 -1.39583349e-01 4.11996394e-01
-1.76930115e-01 -4.34913695e-01 6.42681599e-01 8.38004947e-01
2.78653830e-01 -1.31329335e-02 7.14737326e-02 1.57014072e+00
3.19454493e-03 -2.36017317e-01 -1.22558308e+00 1.37747675e-01
4.06283230e-01 6.37336671e-01 -8.26513767e-01 -2.68674761e-01
-5.83603799e-01 1.32328868e+00 -5.38120531e-02 6.89132333e-01
-8.10403526e-01 -4.19277310e-01 3.02422285e-01 3.97188306e-01
4.43081528e-01 -5.38104236e-01 -8.86896014e-01 -1.01293433e+00
7.74379149e-02 -7.87009418e-01 -4.59512830e-01 -1.79335070e+00
-2.03239202e-01 3.13563675e-01 2.18735754e-01 -1.12260020e+00
-1.76387787e-01 -5.31949341e-01 -8.78950119e-01 9.56494927e-01
-1.06901634e+00 -1.09368002e+00 -2.79249817e-01 3.65700990e-01
7.66818047e-01 2.66449064e-01 7.61963785e-01 -3.51390302e-01
-2.89048105e-01 5.96416295e-02 -5.91929033e-02 -7.74490908e-02
3.77320707e-01 -1.05589020e+00 7.38277495e-01 1.00330412e+00
4.21243859e-03 9.20264840e-01 7.58752882e-01 -6.84645236e-01
-1.88052654e+00 -6.73647344e-01 5.65249979e-01 -9.32905316e-01
1.93209141e-01 -4.92992193e-01 -5.96270263e-01 1.15140164e+00
2.56930918e-01 -3.18557739e-01 2.85082698e-01 -1.23277485e-01
-4.75062132e-01 1.27456173e-01 -8.76687765e-01 9.75875258e-01
8.99547875e-01 -8.49882841e-01 -2.92805076e-01 3.86341780e-01
7.52380788e-01 -1.06059444e+00 -5.68654597e-01 -1.49566948e-01
7.44725466e-01 -1.40556931e+00 1.09430647e+00 -3.68376881e-01
1.13693416e+00 -4.65467095e-01 -1.75344288e-01 -9.84232485e-01
1.36818066e-01 -6.47660077e-01 7.74823055e-02 6.64169550e-01
1.25312507e-01 9.95590910e-03 1.00770950e+00 9.91567433e-01
-4.06613022e-01 -6.23394907e-01 -2.71519840e-01 5.02090203e-04
-2.46318787e-01 -6.67549670e-01 5.98230183e-01 7.71314502e-01
-2.27082223e-01 8.38238671e-02 -2.87052095e-01 3.35862041e-01
4.92951632e-01 4.55766827e-01 1.49220419e+00 -8.29729140e-01
-5.37681222e-01 -4.60506320e-01 -3.22073512e-02 -1.66384244e+00
-3.66232693e-02 -4.36459363e-01 1.90585792e-01 -1.41368222e+00
2.28675440e-01 -1.34486943e-01 8.71471465e-01 1.09571837e-01
1.76474437e-01 -3.78419727e-01 9.87730548e-02 2.36573741e-01
-2.00918630e-01 -3.79573256e-02 1.41530657e+00 -1.99060868e-02
-5.26683033e-01 -1.17256366e-01 -6.96413219e-01 1.21990287e+00
3.83961916e-01 -2.30005518e-01 -6.36122227e-01 -1.18502676e+00
3.55268538e-01 8.00526917e-01 4.51524943e-01 -5.36984861e-01
3.54049891e-01 -4.28200632e-01 6.61509752e-01 -1.09460378e+00
5.24258256e-01 -8.66613269e-01 4.59090054e-01 7.86018372e-02
-6.20792508e-01 1.19624972e-01 2.34558627e-01 3.57586950e-01
2.78098825e-02 -5.22925913e-01 4.38867629e-01 -7.23573506e-01
-7.71146953e-01 1.25918329e-01 -4.29685175e-01 -3.52103680e-01
9.50947404e-01 -1.04313970e+00 -9.67130363e-02 -5.83383441e-01
-5.77319801e-01 -1.60619974e-01 1.24150217e+00 1.64725661e-01
8.94247234e-01 -9.84104097e-01 -2.62968361e-01 6.95167542e-01
4.63206507e-03 4.79469419e-01 4.94492084e-01 1.14599146e-01
-1.26937342e+00 3.38151187e-01 -4.67041917e-02 -1.10859871e+00
-1.41689706e+00 4.53681827e-01 3.22623998e-01 1.53340146e-01
-6.33423567e-01 8.28494072e-01 8.18343401e-01 -6.56492710e-01
1.51885167e-01 -3.55091870e-01 5.62751830e-01 -7.04236031e-01
4.67184752e-01 -5.05429842e-02 -2.49776289e-01 -4.26696628e-01
1.95839912e-01 8.83972943e-01 -1.03998058e-01 -2.66831666e-01
1.06992519e+00 -1.82291940e-01 -3.87939721e-01 4.87225384e-01
1.04771101e+00 1.51325643e-01 -1.75991821e+00 -2.39961028e-01
-4.06570435e-01 -8.90376627e-01 2.36341089e-01 -4.77137536e-01
-4.86350656e-01 1.20691037e+00 9.50851962e-02 -2.24233959e-02
9.78440404e-01 -4.81389277e-02 3.95250171e-01 1.99189991e-01
4.13652718e-01 -9.11706626e-01 2.20933467e-01 4.87760961e-01
9.94786084e-01 -8.26372147e-01 3.48772258e-01 -7.03925073e-01
-4.17297572e-01 1.53633165e+00 8.45549107e-01 -1.30070550e-02
3.99240702e-01 4.99309063e-01 -2.43190214e-01 -3.83843482e-01
-5.18173695e-01 4.87884402e-01 1.08876772e-01 6.84768379e-01
3.70283246e-01 5.29479980e-03 6.83797181e-01 -2.31696278e-01
-1.52223766e-01 7.77360564e-03 1.00761259e+00 1.06479430e+00
-2.43974909e-01 -7.74959683e-01 -6.81973398e-01 1.95276171e-01
1.10063083e-01 -1.04132831e-01 -2.63543278e-01 7.51744270e-01
1.48013502e-01 3.61059636e-01 4.23969656e-01 -2.43803591e-01
5.48223317e-01 -1.19925000e-01 8.44911993e-01 -1.08315349e+00
-4.07204404e-03 1.99943781e-01 -1.85482964e-01 -6.73835814e-01
-3.58473510e-01 -7.20573068e-01 -1.00123918e+00 -2.81218469e-01
-1.36532843e-01 -5.80884397e-01 8.62099469e-01 6.35699451e-01
5.12971401e-01 3.52670074e-01 6.06036127e-01 -1.35269368e+00
1.89467311e-01 -2.02959314e-01 -3.03157419e-01 3.14536005e-01
4.32568580e-01 -1.98261440e-01 -1.48894906e-01 8.21531296e-01] | [9.100329399108887, -3.132481575012207] |
ac627e81-d7f9-45f6-892b-5fce2cbcc3c6 | generalized-reference-kernel-for-one-class | 2205.00534 | null | https://arxiv.org/abs/2205.00534v2 | https://arxiv.org/pdf/2205.00534v2.pdf | Generalized Reference Kernel for One-class Classification | In this paper, we formulate a new generalized reference kernel hoping to improve the original base kernel using a set of reference vectors. Depending on the selected reference vectors, our formulation shows similarities to approximate kernels, random mappings, and Non-linear Projection Trick. Focusing on small-scale one-class classification, our analysis and experimental results show that the new formulation provides approaches to regularize, adjust the rank, and incorporate additional information into the kernel itself, leading to improved one-class classification accuracy. | ['Alexandros Iosifidis', 'Jenni Raitoharju'] | 2022-05-01 | null | null | null | null | ['one-class-classification'] | ['miscellaneous'] | [-3.22548598e-02 -2.20002279e-01 -3.87309372e-01 -3.76946479e-01
-6.17664278e-01 -5.16450763e-01 6.34435356e-01 -1.99985318e-02
-2.94818759e-01 6.35515392e-01 2.94646740e-01 -1.39178231e-01
-5.43357968e-01 -6.04007959e-01 -2.75215536e-01 -6.91363215e-01
-6.35203868e-02 2.45586202e-01 5.09143591e-01 -2.84722120e-01
4.37899292e-01 7.77969956e-01 -1.10193765e+00 4.04708743e-01
9.81345773e-01 7.95455635e-01 -1.16963997e-01 5.16986668e-01
5.21983989e-02 6.78994954e-01 -2.77339637e-01 -3.83456349e-01
4.07324612e-01 -8.61581266e-02 -6.89036429e-01 -1.46034196e-01
5.18247724e-01 -5.64651601e-02 -5.48446000e-01 1.24587023e+00
1.50179029e-01 3.45900267e-01 1.10363007e+00 -1.16546023e+00
-1.28117311e+00 6.02202475e-01 -2.63748497e-01 -5.71391881e-02
3.68172437e-01 -3.73657048e-01 7.57866144e-01 -1.52500868e+00
3.38274896e-01 9.40683067e-01 1.27667785e+00 3.56195539e-01
-1.53580070e+00 -3.28572869e-01 -1.14755712e-01 5.24547219e-01
-1.84682739e+00 -1.05118498e-01 9.01474535e-01 -4.59875852e-01
6.22154176e-01 4.78348315e-01 5.37312329e-01 7.96170294e-01
7.34830722e-02 5.64117789e-01 1.48765945e+00 -6.96095109e-01
2.06434697e-01 8.48211884e-01 6.50957286e-01 8.81000161e-01
3.06544781e-01 9.21444073e-02 -1.88080847e-01 -8.12561870e-01
5.32492697e-01 1.42271325e-01 -7.33433187e-01 -9.79893208e-01
-1.21332633e+00 1.14822340e+00 3.13353449e-01 3.26335460e-01
-1.56767577e-01 -2.36824676e-01 2.33197883e-01 2.21424714e-01
4.81382310e-01 4.88156736e-01 -3.69841367e-01 -2.50960886e-02
-9.53343451e-01 -2.53823325e-02 1.07163119e+00 7.13233054e-01
1.08362889e+00 -5.12494333e-02 -3.66607368e-01 1.09877765e+00
-1.06540650e-01 6.90733254e-01 6.28925323e-01 -5.63127220e-01
1.41465172e-01 5.16934514e-01 1.69588819e-01 -9.94249463e-01
-3.14410895e-01 -2.88520724e-01 -8.15690696e-01 1.56120822e-01
2.53029048e-01 4.45107371e-01 -9.10876989e-01 1.39617515e+00
1.13608964e-01 5.16343951e-01 5.88582568e-02 8.83892715e-01
3.28962564e-01 5.54075599e-01 -2.65860349e-01 2.86291596e-02
1.02636003e+00 -1.10545552e+00 -4.98800933e-01 4.03336316e-01
6.58658683e-01 -9.33770299e-01 1.30753505e+00 3.03552002e-01
-4.25367743e-01 -5.49158931e-01 -1.20773685e+00 2.74642766e-01
-9.05945420e-01 6.06876433e-01 7.71247208e-01 8.19075644e-01
-1.14558864e+00 6.98935807e-01 -5.31790495e-01 -3.38117093e-01
-1.09840499e-03 3.67876709e-01 -6.45385146e-01 -4.59310934e-02
-1.13265038e+00 1.11926758e+00 4.48168248e-01 -1.21290021e-01
-6.93991929e-02 -9.37753499e-01 -6.50593996e-01 -8.38064253e-02
4.58867587e-02 -2.43109584e-01 6.69486165e-01 -4.08691972e-01
-1.78886974e+00 5.46677828e-01 5.11549264e-02 -3.06513637e-01
1.86346412e-01 -3.07241380e-01 -5.63152790e-01 7.28680566e-02
-6.90430403e-02 2.62910366e-01 1.01174736e+00 -8.99373889e-01
-3.13843042e-01 -1.03556640e-01 -2.14265630e-01 7.15578124e-02
-8.75296772e-01 -1.16692692e-01 -2.78582275e-01 -9.37528431e-01
-1.89337283e-01 -1.12053633e+00 -2.72006959e-01 -7.37569705e-02
-2.35105664e-01 -2.25317553e-01 1.03455210e+00 -3.89829159e-01
1.22172105e+00 -2.44343257e+00 3.12250763e-01 6.67663574e-01
2.52241880e-01 5.60667217e-01 -2.41960675e-01 4.03711081e-01
-3.44070435e-01 -1.53641760e-01 -2.84551740e-01 1.66377679e-01
1.79176161e-03 1.51882842e-01 -6.49678826e-01 5.90181172e-01
5.85274845e-02 8.90608788e-01 -9.02697444e-01 -1.89735293e-01
4.10839409e-01 8.54361057e-01 -2.20266104e-01 1.62110999e-01
7.79120028e-01 -3.94734949e-01 -3.14280808e-01 5.68577468e-01
8.69202554e-01 -2.66719967e-01 -3.35058242e-01 -7.33896255e-01
2.03653991e-01 -2.29630530e-01 -1.47258794e+00 1.36605108e+00
-2.32716262e-01 6.03202224e-01 -2.28989586e-01 -1.10635448e+00
8.94487739e-01 3.26775685e-02 4.35890645e-01 7.97670148e-03
-3.91012222e-01 2.25616977e-01 -4.45412964e-01 8.02506953e-02
4.95963961e-01 2.33645290e-01 1.49039745e-01 2.99958497e-01
1.69141859e-01 -4.74181324e-02 1.41655216e-02 1.79641455e-01
1.06299591e+00 -1.61888197e-01 5.85398376e-01 -5.41651607e-01
9.43476319e-01 9.77538303e-02 8.61867219e-02 7.05327868e-01
-1.15494065e-01 5.59435248e-01 1.78672656e-01 -6.51107490e-01
-7.70996273e-01 -1.49608195e+00 -6.02724671e-01 1.05728102e+00
1.76132545e-02 -3.42856765e-01 -6.58354819e-01 -8.15628827e-01
3.47271055e-01 5.77856541e-01 -1.04761350e+00 -5.91552675e-01
-3.75908017e-01 -1.13571548e+00 7.16058910e-01 4.50336009e-01
1.55189812e-01 -5.56237102e-01 3.90634425e-02 1.13753401e-01
3.25042039e-01 -9.59284544e-01 -1.01106763e+00 3.68868768e-01
-8.65258932e-01 -1.15911067e+00 -8.98121476e-01 -6.30889952e-01
8.92855227e-01 4.48233128e-01 6.75154746e-01 -4.09647048e-01
-2.95806915e-01 6.04613006e-01 -3.58817816e-01 1.32211089e-01
-1.89686611e-01 1.65014908e-01 4.67334867e-01 2.19661728e-01
4.57623512e-01 -3.77673954e-01 -2.61274815e-01 5.04781067e-01
-8.24292123e-01 -3.18338990e-01 7.23288178e-01 1.27479136e+00
4.80475485e-01 -9.99661088e-02 2.55350709e-01 -9.11773324e-01
1.19754446e+00 -3.56122077e-01 -5.00334144e-01 6.23498023e-01
-1.34146178e+00 4.94034857e-01 8.38772595e-01 -1.26885951e+00
-7.96096802e-01 1.75289720e-01 3.37992668e-01 -8.23190391e-01
1.84837863e-01 1.88862458e-01 2.45356306e-01 -1.03634024e+00
1.03058290e+00 6.41500771e-01 1.04297169e-01 -5.03921568e-01
9.19089973e-01 4.51349229e-01 4.58858162e-01 -6.30612195e-01
1.10820961e+00 3.57146502e-01 1.62848439e-02 -6.57165349e-01
-7.63885319e-01 -6.92249417e-01 -1.02098095e+00 3.00341427e-01
3.01150888e-01 -6.19455516e-01 -2.52779663e-01 2.72718251e-01
-7.48126388e-01 -1.38487041e-01 -6.13976479e-01 8.19197953e-01
-5.96222281e-01 2.84464180e-01 -7.04234719e-01 -3.01858127e-01
-4.80557948e-01 -9.90586996e-01 7.66583145e-01 1.30152762e-01
-3.10762096e-02 -1.27997971e+00 5.25063276e-01 -3.80120009e-01
7.72978246e-01 -3.79985929e-01 7.27889359e-01 -9.69694078e-01
-1.32048935e-01 -5.81972003e-01 -5.60105562e-01 6.75895035e-01
4.46638018e-01 -2.61554956e-01 -7.07837880e-01 -3.28191251e-01
-1.42310113e-01 2.51554307e-02 8.28449667e-01 1.22623555e-01
9.44826722e-01 -2.56288469e-01 -4.31507081e-01 1.15665066e+00
1.32157254e+00 -2.48058021e-01 4.70891029e-01 1.33663669e-01
8.47871482e-01 1.57739297e-01 4.31843251e-01 1.47438258e-01
1.43075988e-01 7.34337807e-01 -3.65591913e-01 -1.69320643e-01
-1.28279150e-01 -9.59691256e-02 3.70914370e-01 9.23721135e-01
-1.75601676e-01 7.14662015e-01 -4.76256281e-01 3.93702000e-01
-1.97726560e+00 -9.64172363e-01 5.22527322e-02 2.30549908e+00
8.34689438e-01 -3.46483409e-01 2.94866180e-03 3.85239162e-03
7.64961481e-01 -2.35124361e-02 -3.84149611e-01 -3.81955653e-01
-2.17560962e-01 4.46883559e-01 7.98399925e-01 6.41734362e-01
-1.12229717e+00 1.10595298e+00 8.31826019e+00 1.23208487e+00
-1.25506330e+00 1.02797426e-01 9.95189473e-02 8.48670304e-02
-1.40733114e-02 4.60440554e-02 -8.09922338e-01 2.90619493e-01
7.94933259e-01 -5.35631597e-01 8.22456121e-01 1.06825936e+00
-2.95160413e-01 3.52073938e-01 -1.05068958e+00 1.00215161e+00
3.26949239e-01 -1.29510629e+00 2.93195933e-01 4.06338414e-03
6.62061155e-01 -1.67303249e-01 4.17379811e-02 4.52432185e-01
2.27464601e-01 -8.35901260e-01 2.42526397e-01 7.85106421e-01
7.59288430e-01 -7.58939862e-01 7.42573380e-01 -1.43458799e-01
-1.41621780e+00 2.10520282e-01 -7.79025972e-01 1.51638076e-01
-2.36344561e-01 7.47891784e-01 -7.98726857e-01 3.22683007e-01
4.13791835e-01 5.64990640e-01 -9.98044074e-01 9.59333837e-01
-1.31498709e-01 5.17081559e-01 -3.15921128e-01 -1.86693743e-01
9.80289727e-02 -4.50239986e-01 4.91462857e-01 1.54155314e+00
2.51793802e-01 6.67523295e-02 3.27780396e-01 7.01659679e-01
3.18952918e-01 5.28331101e-01 -7.46176779e-01 2.52576143e-01
2.54741758e-01 1.48417389e+00 -5.80303133e-01 -4.98713970e-01
-6.08159840e-01 1.49986029e+00 7.20293224e-01 6.45742536e-01
-9.31417406e-01 -9.23382223e-01 8.33787978e-01 -2.05955915e-02
2.93945491e-01 -2.17920065e-01 -7.33682439e-02 -1.40091431e+00
1.25723749e-01 -4.76549596e-01 3.37393194e-01 -5.11703014e-01
-1.58044612e+00 6.28374100e-01 3.06390941e-01 -1.26629794e+00
-2.91640759e-02 -1.02597368e+00 -4.95765954e-01 9.08513129e-01
-1.25336277e+00 -1.16817427e+00 -4.40183729e-02 8.35670292e-01
-1.17376991e-01 -4.89073336e-01 1.14236975e+00 2.30828598e-01
-2.70656139e-01 9.91366625e-01 3.74533474e-01 8.07027146e-02
1.09469151e+00 -1.34383035e+00 -2.65598446e-02 5.18361628e-01
4.21772122e-01 8.62219810e-01 3.95957381e-01 -3.57839137e-01
-1.39522183e+00 -1.08121336e+00 3.30018669e-01 -4.29191887e-01
1.21111882e+00 -5.32861575e-02 -1.14902186e+00 4.21233296e-01
-2.36444697e-01 6.99468315e-01 8.96566689e-01 2.59683907e-01
-8.16306114e-01 -3.05911899e-01 -1.18337440e+00 3.29873234e-01
7.60824621e-01 -1.03081930e+00 -7.43483603e-01 3.07379246e-01
4.44886833e-01 -1.83313385e-01 -1.02315879e+00 4.46790397e-01
6.96639359e-01 -5.69146812e-01 1.47131252e+00 -7.35609889e-01
-4.25114930e-01 -3.99842292e-01 -2.50466943e-01 -1.58593333e+00
-9.61077929e-01 -2.79489160e-01 -4.37987030e-01 7.61412084e-01
3.51893961e-01 -1.05681622e+00 7.44097292e-01 5.37629545e-01
4.92117368e-02 -8.94499242e-01 -7.49582589e-01 -1.04397345e+00
5.10744378e-02 -2.17365876e-01 3.76320988e-01 1.13405299e+00
1.65946126e-01 -1.99706107e-02 -6.54806852e-01 4.08367872e-01
9.50878263e-01 1.55683219e-01 5.21035731e-01 -1.24268305e+00
-3.17446351e-01 -5.40685833e-01 -8.87173951e-01 -7.85494268e-01
2.41291240e-01 -1.08276856e+00 -2.84278750e-01 -9.26181436e-01
3.85666579e-01 -5.43726563e-01 -9.31293607e-01 5.88020146e-01
-4.92675036e-01 5.92893302e-01 -9.24510956e-02 2.63308793e-01
-4.61201161e-01 5.36280096e-01 7.74088025e-01 -2.38881409e-01
-3.81597131e-01 4.43448350e-02 -6.74968660e-01 5.81021667e-01
6.60480917e-01 -3.92445832e-01 -5.11821389e-01 1.92394555e-01
-2.94111818e-01 -5.13220608e-01 2.87704647e-01 -1.09955299e+00
3.07389170e-01 -3.96472633e-01 5.62955141e-01 -2.04284877e-01
4.57774311e-01 -6.68782592e-01 1.65312458e-02 3.78245443e-01
-2.91839451e-01 1.27044162e-02 1.44565701e-01 6.93556964e-01
-2.56442159e-01 -3.54811341e-01 9.01450872e-01 3.57336104e-01
-8.25794697e-01 2.95593619e-01 -1.94764659e-01 -8.55313465e-02
1.20953834e+00 -3.60933542e-01 -2.17139766e-01 -2.03466751e-02
-8.67817521e-01 -3.07619750e-01 6.08320236e-01 4.54160184e-01
8.42233181e-01 -1.87449932e+00 -6.12748325e-01 3.43964666e-01
4.91959363e-01 -8.88572454e-01 -2.71477699e-01 7.76487768e-01
-1.70037240e-01 5.88367283e-01 -2.74004072e-01 -4.39516425e-01
-1.06961381e+00 8.40271950e-01 2.23555729e-01 -3.48523647e-01
-7.06990182e-01 7.28151202e-01 3.42207365e-02 -6.14399791e-01
1.48049027e-01 -3.46548468e-01 -1.82950646e-01 -1.47559360e-01
7.57890761e-01 5.31517923e-01 -4.67406176e-02 -5.70028603e-01
-4.43507731e-01 1.02345824e+00 -1.86872810e-01 1.41896270e-02
9.95177627e-01 2.13764772e-01 -2.31114656e-01 4.98223931e-01
1.44732511e+00 3.41820896e-01 -8.27922046e-01 -4.88644958e-01
1.19680785e-01 -6.89377666e-01 -6.83030533e-03 -3.79346609e-01
-7.41953790e-01 3.93375874e-01 8.37992251e-01 1.01181343e-01
1.10048139e+00 -1.11797936e-01 2.79373735e-01 8.99825871e-01
4.11388218e-01 -9.66130793e-01 -1.20888948e-01 5.20730555e-01
8.02234232e-01 -1.08223009e+00 3.37630659e-01 -6.24808252e-01
-5.61103344e-01 1.38268554e+00 5.09290576e-01 -3.83425206e-01
1.26554835e+00 2.42161453e-01 2.58063134e-02 1.23950988e-01
-4.46969658e-01 -7.97403827e-02 9.35445666e-01 8.46651435e-01
2.15424016e-01 3.99612874e-01 -3.86190951e-01 3.86881799e-01
3.83156203e-02 -1.35258064e-01 3.03667307e-01 3.70305806e-01
-3.27147126e-01 -1.40718997e+00 -5.46448410e-01 6.40851021e-01
5.91703355e-02 -2.01549649e-01 -2.41149530e-01 6.94437623e-01
-1.78093478e-01 2.62154520e-01 -3.08475882e-01 -7.70919442e-01
4.35371220e-01 3.32629263e-01 5.28685629e-01 -5.70040166e-01
-1.44793674e-01 -4.53466117e-01 -2.69777000e-01 -6.83384955e-01
-1.99295461e-01 -1.93513453e-01 -9.61465895e-01 -1.59522891e-01
-6.12722814e-01 5.95224917e-01 4.81959552e-01 7.35160470e-01
4.14636552e-01 4.84297648e-02 7.28921175e-01 -8.07183743e-01
-1.10487378e+00 -8.92367423e-01 -8.12561929e-01 6.01247668e-01
8.10089037e-02 -9.14440930e-01 -5.63943803e-01 -1.11675924e-02] | [7.7664031982421875, 3.999070405960083] |
42e4de7b-0123-42f7-b36b-1a09989b4993 | towards-a-robust-sensor-fusion-step-for-3d | 2306.07344 | null | https://arxiv.org/abs/2306.07344v1 | https://arxiv.org/pdf/2306.07344v1.pdf | Towards a Robust Sensor Fusion Step for 3D Object Detection on Corrupted Data | Multimodal sensor fusion methods for 3D object detection have been revolutionizing the autonomous driving research field. Nevertheless, most of these methods heavily rely on dense LiDAR data and accurately calibrated sensors which is often not the case in real-world scenarios. Data from LiDAR and cameras often come misaligned due to the miscalibration, decalibration, or different frequencies of the sensors. Additionally, some parts of the LiDAR data may be occluded and parts of the data may be missing due to hardware malfunction or weather conditions. This work presents a novel fusion step that addresses data corruptions and makes sensor fusion for 3D object detection more robust. Through extensive experiments, we demonstrate that our method performs on par with state-of-the-art approaches on normal data and outperforms them on misaligned data. | ['Patric Jensfelt', 'Marko Thiel', 'Viktor Karefjards', 'Maciej K. Wozniak'] | 2023-06-12 | null | null | null | null | ['3d-object-detection'] | ['computer-vision'] | [ 1.11304753e-01 -4.31049407e-01 5.20148762e-02 -5.62702775e-01
-6.29762948e-01 -6.37591183e-01 5.59739470e-01 3.53227645e-01
-4.99013305e-01 6.67817593e-01 -1.74096212e-01 -1.95327669e-01
1.78339824e-01 -7.07885385e-01 -8.79217446e-01 -5.14080763e-01
2.94621289e-01 6.85703635e-01 6.70623660e-01 -2.14139819e-01
1.16100267e-01 8.33274841e-01 -1.98871410e+00 -3.01877737e-01
6.58839941e-01 1.03224885e+00 1.65952265e-01 4.93942648e-01
-2.27973893e-01 1.09397620e-02 -6.58461154e-01 -1.85449511e-01
5.31411350e-01 3.96557510e-01 1.24910146e-01 2.50876158e-01
9.66245770e-01 -4.49786097e-01 -3.37761015e-01 1.32928216e+00
2.55923659e-01 -2.67527342e-01 3.47903073e-01 -1.58135915e+00
-3.20151402e-03 -1.38491377e-01 -9.18567061e-01 9.08529162e-02
3.73807609e-01 1.01220436e-01 4.73005891e-01 -9.50888574e-01
3.65247339e-01 1.43457007e+00 6.64073527e-01 -1.81254372e-02
-1.27658284e+00 -8.72793198e-01 3.45807858e-02 3.62602174e-01
-1.40133584e+00 -6.48046136e-01 6.96516693e-01 -3.65416259e-01
4.59104508e-01 1.10422499e-01 5.59698462e-01 9.29369628e-01
1.56112581e-01 4.58784491e-01 1.06865060e+00 -6.97712377e-02
9.81367454e-02 -1.96042825e-02 1.75820649e-01 3.68939906e-01
1.13680804e+00 4.63774234e-01 -7.39157021e-01 -3.05451788e-02
2.06331670e-01 2.81081527e-01 -7.48982430e-02 -8.29907179e-01
-1.25734460e+00 5.58996439e-01 3.61555696e-01 -2.40926415e-01
-4.69709784e-01 -6.60996735e-02 1.52776033e-01 1.24157213e-01
3.37169841e-02 -8.99793804e-02 -7.00100139e-02 -8.56537372e-02
-8.84927630e-01 4.61248040e-01 3.71465564e-01 1.29165363e+00
1.12682807e+00 1.06359720e-01 3.25730562e-01 4.24465090e-01
4.46511447e-01 1.08115542e+00 -9.37394518e-03 -7.79653847e-01
6.18513107e-01 7.53575504e-01 4.05361682e-01 -8.34331751e-01
-4.08838630e-01 -1.19079024e-01 -7.23671973e-01 6.08593166e-01
3.76867086e-01 1.03285730e-01 -1.09052610e+00 1.18254447e+00
5.21799147e-01 3.01138461e-01 1.17466941e-01 1.08012223e+00
8.00264537e-01 1.03206508e-01 -2.63228297e-01 9.47634969e-03
1.22826099e+00 -1.62108809e-01 -8.59332979e-01 -7.33189166e-01
8.85515064e-02 -1.02958429e+00 4.98233855e-01 3.20591807e-01
-5.35614669e-01 -7.05526829e-01 -1.58853042e+00 6.23628609e-02
-4.29509014e-01 -1.35610133e-01 2.47489408e-01 7.46367455e-01
-4.78796691e-01 -7.56083354e-02 -7.81300128e-01 -4.09465879e-01
3.08666974e-01 3.24325502e-01 -6.23825967e-01 -6.56535506e-01
-8.05918455e-01 1.34593236e+00 3.74604225e-01 2.48670414e-01
-6.05743229e-01 -5.27811527e-01 -1.08547068e+00 -4.63723838e-01
5.75771928e-01 -4.44682181e-01 1.07770205e+00 5.12261875e-02
-8.41167390e-01 6.93364143e-01 -3.10380071e-01 -6.42536998e-01
6.36356413e-01 -5.05538166e-01 -4.68209058e-01 -3.10647011e-01
7.85812587e-02 6.34043217e-01 8.99296582e-01 -1.55465972e+00
-9.81657445e-01 -7.89360166e-01 -3.08999628e-01 1.30793825e-01
3.91264915e-01 -5.17345071e-01 -3.30512971e-01 -1.31265670e-01
5.53102970e-01 -1.03865373e+00 -1.46145552e-01 2.56845448e-02
-4.04099166e-01 1.58110768e-01 1.49572361e+00 -3.79111953e-02
5.21712005e-01 -2.37222147e+00 -2.17425600e-01 1.68633372e-01
1.57612354e-01 3.21954578e-01 6.10981919e-02 2.99854010e-01
2.18306407e-01 -2.23950744e-01 -1.96346641e-01 -5.74935675e-01
-4.08309512e-02 8.16639364e-01 -4.50452358e-01 7.80309200e-01
4.67814393e-02 4.95879084e-01 -6.68539524e-01 -3.18852693e-01
9.01109457e-01 5.63883781e-01 8.95578861e-02 1.51338041e-01
1.18980609e-01 4.51954305e-01 -1.65824577e-01 9.12318408e-01
1.20322549e+00 5.95642626e-01 -2.82882512e-01 -4.72568572e-01
-3.72459024e-01 3.05905730e-01 -1.67697597e+00 1.54865766e+00
1.42252473e-02 7.50066996e-01 2.84649223e-01 -7.64439702e-01
1.00187433e+00 -2.09604986e-02 5.19173503e-01 -6.75199747e-01
1.08514316e-01 2.98658699e-01 -1.33350685e-01 -3.50757629e-01
1.01506400e+00 2.22492423e-02 -6.09668195e-02 -1.00666009e-01
-3.65536034e-01 -8.06800723e-01 1.10992432e-01 -1.17976822e-01
1.02293599e+00 -3.43334191e-02 1.93763494e-01 2.72117555e-01
3.15122426e-01 2.13588148e-01 8.15176606e-01 9.14348364e-01
-4.36881453e-01 6.30325198e-01 -1.15001641e-01 -1.15433365e-01
-1.00528550e+00 -1.23571908e+00 -3.85830522e-01 2.24261478e-01
6.51010513e-01 -3.24001491e-01 -2.34435916e-01 -2.80501753e-01
6.46421015e-01 6.69026375e-01 -2.50287563e-01 -2.38881573e-01
-2.43103728e-01 -4.76564616e-01 4.51525003e-01 4.27811146e-01
5.78836083e-01 -7.72625282e-02 -1.05970478e+00 2.09055483e-01
-9.10264701e-02 -1.78768945e+00 1.15792967e-01 1.38352171e-01
-8.99450719e-01 -1.21590590e+00 1.15562826e-01 -6.98778406e-02
5.80021918e-01 1.19281387e+00 7.98810422e-01 -4.39881645e-02
-1.95795834e-01 3.53872597e-01 -2.31395960e-01 -1.06715727e+00
-1.97336406e-01 -2.82965302e-01 3.83022279e-01 -2.45286115e-02
7.39271343e-01 -3.12165588e-01 -2.54494846e-01 4.11961317e-01
-7.95732558e-01 -1.89559624e-01 4.20156300e-01 4.23906505e-01
6.89578891e-01 2.83521622e-01 4.95854113e-03 -5.08199692e-01
3.11486006e-01 -2.24744484e-01 -1.09752345e+00 -2.87446052e-01
-6.29121959e-01 -6.65161684e-02 -2.74692267e-01 -2.34043121e-01
-6.26949549e-01 5.51788330e-01 3.20196092e-01 -6.61632895e-01
-6.31150007e-01 2.76303738e-01 -3.36435348e-01 -2.30126396e-01
4.82022524e-01 -2.47777343e-01 2.52724290e-01 -5.18160105e-01
1.49931997e-01 8.65205944e-01 9.25225973e-01 -1.94711044e-01
1.30598485e+00 9.38124299e-01 4.09552306e-01 -1.01590085e+00
-5.81693590e-01 -6.17188513e-01 -8.35248947e-01 -3.16848904e-01
5.86495876e-01 -1.16750944e+00 -3.55048299e-01 6.38219774e-01
-1.14038229e+00 4.12407577e-01 -1.32870585e-01 7.08255112e-01
-1.17720082e-01 4.30809736e-01 2.03104332e-01 -9.84131992e-01
3.39599699e-01 -1.45828700e+00 1.35868561e+00 2.27287576e-01
4.71308306e-02 -2.29360938e-01 -4.01885509e-02 4.81448978e-01
2.79908299e-01 3.96113038e-01 2.66243309e-01 -2.29808241e-01
-8.63105953e-01 -8.48665416e-01 -2.17541784e-01 4.51045893e-02
2.95425028e-01 2.06594512e-01 -9.52466786e-01 -1.38985991e-01
-2.20652550e-01 5.39321303e-02 8.10615659e-01 1.17164716e-01
3.62349153e-01 4.27741617e-01 -4.97455269e-01 2.93104708e-01
1.17590368e+00 8.15774277e-02 5.37541032e-01 2.53812969e-01
6.12386465e-01 5.62762439e-01 9.84311163e-01 3.54889125e-01
7.88939178e-01 7.85244584e-01 1.13932860e+00 -1.07909255e-02
-5.27674705e-02 -1.41235590e-01 2.09460840e-01 1.28944561e-01
1.26345247e-01 8.84458423e-02 -9.65512335e-01 5.10350645e-01
-1.90932322e+00 -7.36278713e-01 -7.58916676e-01 2.60004234e+00
1.95108578e-01 4.33295548e-01 -4.61117588e-02 4.67544883e-01
7.95518577e-01 2.04195734e-02 -6.48218811e-01 4.76520285e-02
-3.07693720e-01 -1.65285870e-01 1.15344214e+00 5.28104126e-01
-1.00252938e+00 6.81709647e-01 6.26312542e+00 -2.40362948e-03
-9.17976201e-01 2.12187059e-02 -3.52577478e-01 -4.90475483e-02
-2.49692127e-02 1.72250062e-01 -1.17576742e+00 3.79956990e-01
6.11980557e-01 1.08318508e-01 -5.03866759e-04 4.77293044e-01
1.72636449e-01 -7.84556329e-01 -1.09885728e+00 1.34380841e+00
1.26934513e-01 -9.31457877e-01 -4.24038440e-01 4.62340236e-01
3.36884737e-01 5.25883913e-01 -2.34602913e-01 1.31028965e-02
4.43095922e-01 -6.19506180e-01 9.21620786e-01 4.16344672e-01
4.06354845e-01 -5.91966987e-01 9.26193714e-01 5.07820845e-01
-1.18297672e+00 4.17319126e-02 -4.07607853e-01 -1.84975863e-01
4.48820889e-01 1.06953287e+00 -1.06228697e+00 5.84200382e-01
8.94233823e-01 6.48250461e-01 -7.24326253e-01 1.37906671e+00
-3.39829594e-01 2.24280447e-01 -9.51391160e-01 2.58308083e-01
-8.16777200e-02 -1.13727406e-01 1.02267838e+00 6.72567844e-01
5.17282188e-01 -1.38885617e-01 3.34320247e-01 4.50212091e-01
1.63240388e-01 -5.41525960e-01 -1.12591815e+00 3.35612416e-01
8.16239357e-01 1.08106208e+00 -2.66678303e-01 -2.25327805e-01
-7.91272402e-01 3.68651241e-01 -1.37983233e-01 2.73002863e-01
-6.74348831e-01 1.41374562e-02 1.19468403e+00 2.41781548e-01
4.73794073e-01 -9.30182755e-01 -4.44479674e-01 -9.85141158e-01
1.56519711e-01 -3.94254833e-01 1.95073768e-01 -9.27438319e-01
-1.09116840e+00 1.39621571e-01 2.59126395e-01 -1.60226536e+00
-1.43513486e-01 -4.01313305e-01 -7.81983361e-02 6.04642689e-01
-1.80818534e+00 -1.05926144e+00 -7.17500806e-01 4.48552728e-01
2.32153103e-01 -9.89349708e-02 4.37347740e-01 4.60473627e-01
-4.11842078e-01 -5.40901013e-02 -1.84323415e-01 -2.27839082e-01
9.61318135e-01 -8.99352551e-01 2.76389509e-01 1.31713772e+00
1.28470898e-01 1.59023523e-01 1.13825655e+00 -6.78490937e-01
-1.72934115e+00 -9.33870018e-01 6.48683488e-01 -4.43161756e-01
4.72039491e-01 -4.42593277e-01 -8.45006227e-01 6.44468784e-01
1.16819562e-02 9.98356044e-02 3.87086004e-01 -1.64054744e-02
-4.68891978e-01 -6.12236559e-01 -1.22445452e+00 3.10645342e-01
7.75535047e-01 -2.81401187e-01 -7.28680670e-01 1.06481798e-01
2.40244359e-01 -9.95326638e-01 -5.28597772e-01 8.58454227e-01
5.22925556e-01 -1.04906380e+00 1.04587567e+00 -1.24607608e-01
-4.87493247e-01 -9.58827436e-01 -6.75284803e-01 -1.09555268e+00
2.66551673e-02 -3.20773050e-02 -2.14644402e-01 1.21900976e+00
-1.43122479e-01 -7.32709885e-01 6.03495300e-01 6.85639203e-01
-2.52936989e-01 3.09009373e-01 -1.47949386e+00 -7.57187247e-01
-5.53525448e-01 -1.00733089e+00 8.40509295e-01 5.72135031e-01
-5.89127421e-01 3.62141788e-01 -2.91547120e-01 6.90094948e-01
1.02947581e+00 2.15088814e-01 1.36570716e+00 -1.68171036e+00
5.02557874e-01 -3.55331242e-01 -1.06389236e+00 -8.24678779e-01
-2.10527219e-02 -2.32879922e-01 4.06790167e-01 -1.50490487e+00
-3.93024206e-01 -3.91836882e-01 5.13986312e-02 4.03894395e-01
-7.71188065e-02 4.08881396e-01 2.26381913e-01 2.29793102e-01
-1.53414339e-01 4.26786542e-01 6.97493911e-01 -2.86868989e-01
-1.16925919e-02 1.66406646e-01 -4.97397423e-01 6.51180625e-01
5.34237444e-01 -5.05591333e-01 -1.41042724e-01 -6.55661762e-01
2.18776658e-01 -2.41431668e-01 6.49785936e-01 -1.35194981e+00
4.36366051e-01 -2.18257487e-01 3.43731552e-01 -1.47182250e+00
6.85093045e-01 -1.63013864e+00 3.29015017e-01 3.04977328e-01
5.07975996e-01 2.56221145e-01 3.53550971e-01 7.20903814e-01
-3.26741248e-01 -1.89079100e-03 7.75566638e-01 1.35045692e-01
-9.03217256e-01 2.04665154e-01 -3.75834346e-01 -4.53420132e-01
1.14389646e+00 -6.16972744e-01 -3.79571259e-01 -2.33954683e-01
-9.65856239e-02 4.54012185e-01 6.50441229e-01 7.82012761e-01
7.53382921e-01 -1.45709884e+00 -6.44268692e-01 6.08924270e-01
4.43405330e-01 6.16608322e-01 -1.25158317e-02 9.08744037e-01
-1.80345848e-01 3.01821440e-01 -1.37217551e-01 -1.14007235e+00
-1.56661665e+00 3.37593973e-01 2.06246912e-01 5.10362685e-01
-4.13620472e-01 2.94760048e-01 -3.61203641e-01 -3.35165650e-01
3.02698821e-01 -4.72304553e-01 5.88823184e-02 2.56784558e-01
4.93392289e-01 3.62795949e-01 4.65218067e-01 -1.11625981e+00
-5.85617602e-01 7.26586342e-01 7.63365626e-02 -2.08028182e-01
9.82402444e-01 -5.01200616e-01 1.61416754e-01 6.03208065e-01
5.47087610e-01 -1.80816390e-02 -1.27481711e+00 -4.95200127e-01
-2.80118018e-01 -8.30170989e-01 2.71817178e-01 -2.99416780e-01
-9.30167139e-01 9.55255568e-01 9.35757279e-01 2.39602312e-01
8.11290622e-01 -1.25078619e-01 6.93137228e-01 4.08840328e-01
6.29647553e-01 -1.10092545e+00 -5.25972426e-01 5.31174183e-01
6.11901462e-01 -1.60263288e+00 4.33956802e-01 -4.38494235e-01
-2.99444228e-01 9.35722470e-01 5.97296536e-01 9.98584479e-02
6.19462967e-01 6.04145169e-01 2.76516348e-01 -8.06494579e-02
-4.14539665e-01 -6.84721231e-01 5.15080504e-02 9.11725283e-01
-2.59486884e-01 9.28244516e-02 1.77354619e-01 -7.03529716e-02
-2.74424672e-01 -2.04936132e-01 4.99049336e-01 1.24486136e+00
-7.04071045e-01 -1.10072982e+00 -1.14976227e+00 3.26346397e-01
1.41427964e-01 3.37748885e-01 -4.07914162e-01 9.16445315e-01
3.95681322e-01 1.26974332e+00 2.89631277e-01 -5.03148854e-01
9.60346162e-01 -7.45003019e-03 5.17059803e-01 -4.14278418e-01
1.07784621e-01 1.67848900e-01 -1.02883577e-02 -6.03369117e-01
-6.56563878e-01 -1.12355852e+00 -1.31952870e+00 -4.49141383e-01
-4.00859207e-01 -2.89309531e-01 1.25736701e+00 1.01541233e+00
5.39542317e-01 2.90150762e-01 3.77431780e-01 -1.13421774e+00
-3.29772919e-01 -8.67108166e-01 -3.52966696e-01 2.49361247e-01
8.31760526e-01 -1.34304273e+00 -2.10677251e-01 -3.32522206e-02] | [7.711618423461914, -2.3321995735168457] |
3dee9598-e815-4e35-9c70-cae20065db45 | wildrefer-3d-object-localization-in-large | 2304.05645 | null | https://arxiv.org/abs/2304.05645v1 | https://arxiv.org/pdf/2304.05645v1.pdf | WildRefer: 3D Object Localization in Large-scale Dynamic Scenes with Multi-modal Visual Data and Natural Language | We introduce the task of 3D visual grounding in large-scale dynamic scenes based on natural linguistic descriptions and online captured multi-modal visual data, including 2D images and 3D LiDAR point clouds. We present a novel method, WildRefer, for this task by fully utilizing the appearance features in images, the location and geometry features in point clouds, and the dynamic features in consecutive input frames to match the semantic features in language. In particular, we propose two novel datasets, STRefer and LifeRefer, which focus on large-scale human-centric daily-life scenarios with abundant 3D object and natural language annotations. Our datasets are significant for the research of 3D visual grounding in the wild and has huge potential to boost the development of autonomous driving and service robots. Extensive comparisons and ablation studies illustrate that our method achieves state-of-the-art performance on two proposed datasets. Code and dataset will be released when the paper is published. | ['Yuexin Ma', 'Sibei Yang', 'Xinge Zhu', 'Yuenan Hou', 'Peishan Cong', 'Xidong Peng', 'Zhenxiang Lin'] | 2023-04-12 | null | null | null | null | ['visual-grounding', 'object-localization'] | ['computer-vision', 'computer-vision'] | [-4.07081217e-01 -2.46099696e-01 -2.08661467e-01 -7.68606305e-01
-4.02035087e-01 -5.37136137e-01 6.81790709e-01 -1.01389393e-01
-3.88324231e-01 3.31974149e-01 -1.79801747e-01 -6.85741007e-02
3.25755924e-01 -5.87304056e-01 -8.09160471e-01 -8.00563842e-02
-1.29852101e-01 6.93312943e-01 7.60448813e-01 -6.55597568e-01
2.60557324e-01 8.20645332e-01 -2.17708254e+00 9.91674662e-02
4.62358862e-01 1.09206820e+00 7.75892913e-01 3.51323396e-01
-4.86143529e-01 4.70881402e-01 2.47221161e-02 -1.98601440e-01
4.73466396e-01 2.24508151e-01 -5.66651762e-01 2.58922458e-01
8.41635108e-01 -3.19281787e-01 -3.92270267e-01 9.91165698e-01
3.08696300e-01 2.25740373e-01 4.26753640e-01 -1.85909903e+00
-7.73765326e-01 -3.19599926e-01 -7.32004225e-01 -1.49462270e-02
6.86647296e-01 2.71564573e-01 6.84703410e-01 -1.19420159e+00
9.43501174e-01 1.57307673e+00 6.28814220e-01 4.66426104e-01
-6.72272921e-01 -6.52564824e-01 2.48253852e-01 4.19649273e-01
-1.49661493e+00 -4.96766150e-01 7.83783853e-01 -7.09007680e-01
1.06846905e+00 -2.23032355e-01 8.68898928e-01 1.05765092e+00
5.39940521e-02 6.99902892e-01 1.00485909e+00 -2.39076674e-01
6.60297573e-02 8.91178697e-02 -4.12513688e-02 1.19525039e+00
2.55329788e-01 3.34489495e-01 -8.98009360e-01 -8.54856428e-03
5.92620850e-01 3.90280724e-01 2.34336480e-01 -9.50716555e-01
-1.52204716e+00 6.58025861e-01 8.12657356e-01 -8.43875110e-02
-2.27745369e-01 3.83735508e-01 2.48278022e-01 -8.34999904e-02
4.60698009e-01 -2.44381741e-01 -3.19231212e-01 -1.69681199e-02
-5.64695001e-01 4.94168937e-01 2.18793809e-01 1.75418532e+00
1.26179743e+00 -1.10800400e-01 2.26719379e-01 6.59375191e-01
5.51785827e-01 1.24964213e+00 2.11706206e-01 -1.22773290e+00
5.56550622e-01 7.42377341e-01 3.90127271e-01 -1.17264748e+00
-4.06943649e-01 2.69120723e-01 -3.88284266e-01 2.44886696e-01
-8.37293640e-02 3.93647671e-01 -9.38223302e-01 1.41456544e+00
5.53329945e-01 1.64525196e-01 8.47634971e-02 1.16055882e+00
1.14589560e+00 5.29460430e-01 -1.08414497e-02 2.83404469e-01
1.28692603e+00 -9.40763414e-01 -6.81446135e-01 -7.80583501e-01
4.26285386e-01 -5.05568147e-01 1.17469239e+00 -2.32594043e-01
-7.55218089e-01 -7.50978351e-01 -1.10431302e+00 -3.72607350e-01
-7.80290425e-01 2.14660883e-01 5.80035746e-01 8.33413303e-02
-1.11392105e+00 -1.71031710e-02 -7.49939919e-01 -9.59571421e-01
4.62648720e-01 -5.16710319e-02 -7.88321793e-01 -3.56293052e-01
-8.77712309e-01 9.96954679e-01 3.54960084e-01 8.50406140e-02
-9.84406769e-01 -4.47274208e-01 -1.37915742e+00 -7.42300212e-01
3.00959170e-01 -8.65707397e-01 1.06694245e+00 -1.87948361e-01
-9.23376679e-01 1.83314562e+00 -3.18666309e-01 -4.15706098e-01
6.68917775e-01 -2.97528744e-01 -2.51615793e-01 6.58639222e-02
6.12542868e-01 1.15590560e+00 4.85885412e-01 -1.56307983e+00
-8.66943777e-01 -7.45988488e-01 1.80475667e-01 2.60281622e-01
2.51142710e-01 -1.30260915e-01 -6.96114600e-01 -1.72851458e-02
4.68515486e-01 -1.11032808e+00 -1.77320287e-01 6.32441521e-01
-1.41034141e-01 -3.38784218e-01 1.02235353e+00 -2.10899353e-01
1.43980652e-01 -2.21126890e+00 -1.36963308e-01 -2.82944679e-01
2.87637766e-02 -6.10050410e-02 -1.25090614e-01 3.58194321e-01
5.97013474e-01 -1.48303851e-01 8.95424187e-02 -6.13884032e-01
1.04773685e-01 6.58205807e-01 -4.04940963e-01 6.59342408e-01
2.06727177e-01 1.09614992e+00 -1.27067935e+00 -8.17458808e-01
7.74387956e-01 2.69543171e-01 -4.07697290e-01 2.99715072e-01
-2.54636079e-01 4.39653933e-01 -7.12789714e-01 8.26089442e-01
6.56183004e-01 -6.86783716e-02 -4.80884999e-01 -2.66073436e-01
-3.44763756e-01 -1.75901115e-01 -8.68468165e-01 2.55121112e+00
-3.63739461e-01 5.71303904e-01 -1.57746330e-01 -8.04971278e-01
1.17032206e+00 -1.41850308e-01 4.29835200e-01 -9.20018435e-01
2.93943956e-02 2.50841528e-01 -7.89972663e-01 -6.69566035e-01
7.81844556e-01 1.26054347e-01 -2.93711185e-01 -2.29346991e-01
-1.21413423e-02 -8.62357557e-01 1.33160785e-01 2.70802498e-01
6.15377545e-01 6.54436111e-01 2.63804406e-01 -1.48274034e-01
3.10366362e-01 6.60512209e-01 4.01896983e-01 6.12536907e-01
-6.55552268e-01 6.65250540e-01 -3.77172738e-01 -6.92197800e-01
-1.23136008e+00 -1.32753384e+00 -8.16644430e-02 8.36175203e-01
9.58436310e-01 -2.43462101e-01 -9.63396803e-02 -3.83615434e-01
4.68303770e-01 5.18426180e-01 -5.45668900e-01 3.51053476e-02
-3.75756323e-01 6.17449246e-02 2.88481921e-01 5.27695060e-01
8.09692562e-01 -1.10035682e+00 -1.12882531e+00 -2.04187661e-01
-3.72761041e-01 -1.84108305e+00 -2.45346755e-01 -7.29269162e-02
-6.08803332e-01 -1.03838503e+00 -3.70407879e-01 -8.73844206e-01
4.58255559e-01 9.86437798e-01 1.06675935e+00 -1.17146699e-02
-4.77029711e-01 6.46905661e-01 -4.40841764e-01 -6.52179897e-01
-1.18321747e-01 -3.86693627e-01 4.43046600e-01 -2.41384730e-01
5.23370206e-01 -4.16131377e-01 -3.68798316e-01 4.51849014e-01
-2.48749465e-01 3.20237577e-01 3.40684801e-01 4.37537283e-01
9.81509984e-01 -5.78741610e-01 1.95240408e-01 -3.20682406e-01
-1.49074554e-01 -3.55875313e-01 -6.76788270e-01 -5.96619844e-02
-3.21476460e-01 -1.31889671e-01 2.32574400e-02 -7.27203935e-02
-6.91346943e-01 3.40303093e-01 1.62339419e-01 -1.03097379e+00
-4.70818162e-01 1.19760260e-01 2.07874216e-02 -1.52547941e-01
7.20156550e-01 1.19436216e-02 7.88258985e-02 -3.95216644e-01
8.50989401e-01 7.11089194e-01 1.12405002e+00 -5.03977835e-01
9.65624452e-01 1.05095458e+00 9.74535495e-02 -6.95579708e-01
-1.04090631e+00 -1.03200340e+00 -1.04515564e+00 -4.45259959e-01
9.92287099e-01 -1.40314054e+00 -4.63571280e-01 2.89837867e-01
-1.34592402e+00 -1.49531484e-01 -2.70596057e-01 3.15022469e-01
-1.12422097e+00 2.45769575e-01 1.80655904e-02 -8.87099147e-01
-5.22426218e-02 -9.29520547e-01 1.76974618e+00 1.09207526e-01
1.30795702e-01 -6.32997215e-01 -5.29527524e-03 4.39642757e-01
-4.76981997e-02 5.71102917e-01 4.70111877e-01 -9.92625132e-02
-7.53487051e-01 -1.36754975e-01 -4.43214059e-01 -8.62102136e-02
1.77729145e-01 -2.12628022e-01 -9.67772424e-01 -1.33633465e-01
-5.14672577e-01 -5.61847389e-01 6.04481637e-01 1.52706563e-01
8.18843722e-01 2.99082279e-01 -6.27385437e-01 6.82977557e-01
1.34412467e+00 6.48619980e-02 1.94105759e-01 4.89007562e-01
9.02327836e-01 8.01234066e-01 1.22819781e+00 4.23637867e-01
8.49351287e-01 8.92963052e-01 8.81494999e-01 -1.27982721e-01
-2.71209985e-01 -6.96924508e-01 2.61220992e-01 6.64895535e-01
8.35681055e-03 6.64884895e-02 -1.26293147e+00 9.52957153e-01
-2.07183576e+00 -9.44806933e-01 -2.85275936e-01 1.88058662e+00
2.17969462e-01 -1.33261383e-02 -6.69855028e-02 -3.61696780e-01
6.85729325e-01 1.28418326e-01 -6.64388537e-01 -3.91966738e-02
-3.79667819e-01 -4.83096927e-01 5.53252339e-01 2.14202210e-01
-1.05791104e+00 1.39781618e+00 6.14493656e+00 3.60347867e-01
-8.12689424e-01 3.66484702e-01 -5.42262420e-02 4.52444032e-02
-1.01361372e-01 -3.27669904e-02 -8.74276936e-01 9.23684537e-02
5.50143898e-01 -3.13369811e-01 2.48076424e-01 1.19255114e+00
3.82619947e-01 -5.82350194e-02 -1.17502773e+00 1.57106805e+00
2.15538144e-01 -1.38326025e+00 2.88850516e-02 8.40258449e-02
7.00673342e-01 7.76721597e-01 -3.01703941e-02 5.11379689e-02
6.12406671e-01 -5.92944503e-01 1.45546353e+00 4.80129033e-01
9.63714838e-01 -4.26212579e-01 4.48997617e-01 5.77147007e-01
-1.52118349e+00 -8.20752382e-02 -5.47741711e-01 -8.75886679e-02
4.55080330e-01 1.86626315e-01 -7.70698130e-01 7.53939152e-01
1.31526315e+00 1.42272592e+00 -6.45846426e-01 9.55172539e-01
9.18559823e-03 -2.21158445e-01 -4.10242051e-01 -2.58228695e-03
3.25318336e-01 -1.47226349e-01 5.43644667e-01 9.87850964e-01
4.09737945e-01 -4.98407939e-03 7.46311963e-01 1.00521028e+00
3.65410984e-01 -3.62340314e-03 -1.34735513e+00 5.74451908e-02
5.75612903e-01 1.20509148e+00 -5.12257218e-01 -4.29382324e-01
-4.57855612e-01 8.35695386e-01 4.74059969e-01 2.45399833e-01
-6.72435462e-01 -7.26203397e-02 9.86059129e-01 3.12564820e-01
6.92607462e-02 -9.52651024e-01 -1.25458313e-03 -1.12661791e+00
1.48606047e-01 -2.20941361e-02 1.19661145e-01 -1.63513052e+00
-1.31239581e+00 5.75120091e-01 4.54072684e-01 -1.49340820e+00
-3.19559008e-01 -6.70358241e-01 -5.12542054e-02 6.75890267e-01
-1.93636000e+00 -1.64398932e+00 -1.13408577e+00 8.68400156e-01
7.69938350e-01 -1.66671827e-01 4.88611996e-01 6.23377487e-02
1.35658070e-01 -5.39343767e-02 -1.85709655e-01 -1.09125011e-01
5.95448077e-01 -8.31406593e-01 8.30988467e-01 6.64141774e-01
4.21084374e-01 3.16455960e-01 6.36558235e-01 -6.85034037e-01
-1.71895051e+00 -1.35891569e+00 7.40960360e-01 -8.48863602e-01
5.93175113e-01 -7.13817418e-01 -4.76734906e-01 7.43415594e-01
-1.34911835e-01 5.06326556e-01 -4.56142016e-02 -3.02641422e-01
-3.25617760e-01 -2.22569536e-02 -1.28064215e+00 4.83904481e-01
1.93876731e+00 -6.56198978e-01 -7.19710052e-01 5.14819980e-01
1.08479881e+00 -8.13467145e-01 -3.68550241e-01 5.43044984e-01
4.41456318e-01 -7.83486128e-01 1.22789609e+00 -4.08188134e-01
-6.89186379e-02 -6.28944755e-01 -7.73701489e-01 -1.00354719e+00
-1.96151540e-01 -2.59870440e-01 1.42978847e-01 8.80795956e-01
-1.58046380e-01 -4.90236729e-01 6.29560888e-01 2.29278505e-01
-5.96032858e-01 -1.98735312e-01 -1.05263996e+00 -9.38903987e-01
-3.77118230e-01 -7.02836514e-01 6.33132875e-01 6.46638274e-01
-4.61719006e-01 1.53937995e-01 -2.98356205e-01 2.95129925e-01
8.18218768e-01 3.97790343e-01 1.19063604e+00 -1.41023302e+00
5.88692605e-01 1.71847850e-01 -1.12147272e+00 -1.24682009e+00
6.13254011e-01 -9.46705937e-01 4.24541444e-01 -1.64560425e+00
6.40530586e-02 -6.92030311e-01 1.20892271e-01 6.11755610e-01
3.12432319e-01 4.78509843e-01 2.81627148e-01 4.00184989e-01
-8.07070851e-01 9.19863343e-01 1.12211525e+00 -4.04650807e-01
5.90673275e-02 -4.48388666e-01 -8.28096643e-02 7.78367400e-01
3.78998131e-01 -2.65508354e-01 -4.83462811e-01 -6.47670388e-01
1.24262974e-01 -2.43567437e-01 8.91046822e-01 -1.05349445e+00
2.51609176e-01 -6.46576226e-01 4.23020758e-02 -1.24840760e+00
8.46236527e-01 -1.00950265e+00 1.17711827e-01 1.36500746e-01
7.24088550e-02 5.86021423e-01 3.38892758e-01 8.32552075e-01
3.00074986e-04 1.88703313e-01 5.28362155e-01 -3.96870404e-01
-1.73694980e+00 7.66745508e-01 8.10648426e-02 1.44306675e-01
1.17379475e+00 -5.09377778e-01 -3.21718067e-01 -1.64302602e-01
-4.94346768e-01 5.92793703e-01 9.85574841e-01 1.15897846e+00
1.00654078e+00 -1.71723604e+00 -6.41859293e-01 3.67286295e-01
9.21676099e-01 3.08014393e-01 1.87675312e-01 5.35453498e-01
-7.13130891e-01 4.57398176e-01 -4.24033254e-01 -1.30490553e+00
-1.22619390e+00 6.83471739e-01 3.08064401e-01 7.23607957e-01
-1.13988745e+00 6.31626189e-01 3.41528863e-01 -6.85780168e-01
-4.14747298e-02 -5.04145443e-01 5.94728403e-02 -3.56544346e-01
2.97492325e-01 -4.84069698e-02 6.93183616e-02 -1.43533218e+00
-7.58623719e-01 1.03881562e+00 4.16790128e-01 -1.48754925e-01
1.04351628e+00 -6.82980239e-01 7.10056424e-02 1.00832713e+00
1.22251177e+00 -3.82005215e-01 -1.33381903e+00 -5.38475215e-01
-1.05170161e-01 -9.07884419e-01 3.70822139e-02 -2.89472848e-01
-7.57907927e-01 9.38991904e-01 9.04900610e-01 -2.68096924e-01
4.93472964e-01 5.08208156e-01 6.70927882e-01 5.76297641e-01
1.41507435e+00 -8.48512650e-01 3.02018136e-01 6.12640381e-01
1.08714187e+00 -1.55466580e+00 -6.51476532e-02 -4.43728626e-01
-6.30920351e-01 6.95538998e-01 6.77619517e-01 -8.77304077e-02
5.09898424e-01 -7.29165599e-03 3.04871798e-01 -4.53713089e-01
-5.40518522e-01 -5.97387254e-01 8.62181783e-02 1.25991786e+00
-3.33105147e-01 1.06711455e-01 4.28814828e-01 4.33702953e-02
-4.02490020e-01 -1.19801186e-01 1.98053226e-01 1.05161381e+00
-5.96291900e-01 -6.71180010e-01 -1.95202902e-01 -9.31780934e-02
4.53889698e-01 1.68264210e-01 -3.90581965e-01 8.10280561e-01
3.63402456e-01 1.01115322e+00 3.60547185e-01 -3.58671129e-01
6.29771888e-01 -6.62983209e-02 6.98623717e-01 -6.68251753e-01
2.16955677e-01 -5.62519610e-01 5.59920780e-02 -1.00848699e+00
-6.91406727e-01 -7.78352261e-01 -1.51056254e+00 -2.29700953e-01
-2.15783324e-02 -2.49875218e-01 1.14362669e+00 9.27029788e-01
5.78880608e-01 2.75747031e-02 4.27032799e-01 -1.45153892e+00
-1.71416521e-01 -5.91855884e-01 -5.74486315e-01 7.51178682e-01
4.85963970e-01 -1.45976853e+00 -1.33914188e-01 3.36223125e-01] | [7.917912483215332, -2.2702484130859375] |
eb5a6873-7b12-436b-a10e-048cae61bc63 | automatic-spoken-language-identification | null | null | https://core.ac.uk/download/pdf/10900425.pdf | https://core.ac.uk/download/pdf/10900425.pdf | Automatic Spoken Language Identification Utilizing Acoustic and Phonetic Speech Information | Automatic spoken Language Identification (LID) is the process of identifying the language spoken within an utterance. The challenge that this task presents is that no prior information is available indicating the content of the utterance or the identity of the speaker. The trend of globalization and the pervasive popularity of the Internet will amplify the need for the capabilities spoken language identification systems provide. A prominent application arises in call centers dealing with speakers speaking different languages. Another important application is to index or search huge speech data archives and corpora that contain multiple languages. The aim of this research is to develop techniques targeted at producing a fast and more accurate automatic spoken LID system compared to the previous National Institute of Standards and Technology (NIST) Language Recognition Evaluation.
Acoustic and phonetic speech information are targeted as the most suitable features for representing the characteristics of a language. To model the acoustic speech features a Gaussian Mixture Model based approach is employed. Phonetic speech information is extracted using existing speech recognition technology. Various techniques to improve LID accuracy are also studied. One approach examined is the employment of Vocal Tract Length Normalization to reduce the speech variation caused by different speakers. A linear data fusion technique is adopted to combine the various aspects of information extracted from speech.
As a result of this research, a LID system was implemented and presented for evaluation in the 2003 Language Recognition Evaluation conducted by the NIST. | ['BIT', 'BEng(Hons)', 'Kim-Yung Eddie Wong'] | 2004-06-01 | null | null | null | null | ['spoken-language-identification'] | ['speech'] | [-1.38969094e-01 -3.30171734e-01 3.78159918e-02 -5.25128603e-01
-9.02081847e-01 -5.56725800e-01 6.47809565e-01 1.13224022e-01
-4.30575907e-01 3.54856193e-01 5.18958271e-01 -1.16109222e-01
-4.11853613e-03 -1.45159900e-01 1.04816422e-01 -7.34667480e-01
5.01572609e-01 3.62597257e-01 -1.91978797e-01 -2.05564961e-01
5.04943311e-01 8.05205703e-01 -1.98612118e+00 1.43961936e-01
5.97668648e-01 7.39303470e-01 5.74778199e-01 8.55711281e-01
-7.27584481e-01 4.18189555e-01 -7.80137122e-01 3.67836863e-01
6.54687583e-02 -5.25909066e-01 -6.88474000e-01 4.95306313e-01
4.79730293e-02 -1.65543303e-01 8.33853632e-02 9.78077412e-01
7.03347921e-01 3.33745509e-01 5.96539915e-01 -7.92166591e-01
-1.03424333e-01 4.64304388e-01 3.03456020e-02 3.87416840e-01
5.53440213e-01 -1.62280798e-01 4.46558565e-01 -1.06050527e+00
2.54680991e-01 1.45496821e+00 2.11863860e-01 3.79110903e-01
-9.99597490e-01 -4.44019377e-01 -1.95947349e-01 1.06836043e-01
-1.65177357e+00 -9.04548943e-01 7.80870676e-01 -5.11118889e-01
1.02606618e+00 4.01729137e-01 7.22238496e-02 4.88383323e-01
-9.74288881e-02 2.81480104e-01 1.13600934e+00 -1.12905324e+00
2.11395368e-01 8.10682237e-01 5.75767338e-01 2.42628410e-01
-1.54204890e-01 -1.42941430e-01 -3.58632058e-01 -1.81156769e-01
1.22179866e-01 -4.46600080e-01 -1.69579573e-02 1.32787943e-01
-5.59900403e-01 8.59892130e-01 -4.27746445e-01 1.11274219e+00
-6.34899080e-01 -6.09549820e-01 5.32894611e-01 1.98615149e-01
3.43878746e-01 5.42616472e-02 -5.28450869e-02 -3.76717836e-01
-1.02683735e+00 2.86974874e-03 9.85475004e-01 4.58385259e-01
5.10446250e-01 2.27894604e-01 2.34847009e-01 1.24914777e+00
8.58862579e-01 3.85619134e-01 9.41075742e-01 -7.83541262e-01
8.18641037e-02 6.94334447e-01 -7.60159642e-02 -7.18352258e-01
-1.81083694e-01 6.11611493e-02 -3.12216341e-01 2.29538336e-01
3.15527111e-01 -1.47123914e-02 -9.40255344e-01 1.34267497e+00
3.89193058e-01 -3.89432102e-01 4.86477315e-01 6.69502318e-01
8.32653403e-01 9.48692977e-01 -1.04134351e-01 -4.05455858e-01
1.55455542e+00 -4.54395056e-01 -1.00082850e+00 -2.64518298e-02
3.45327437e-01 -1.42727470e+00 8.45215797e-01 4.92441118e-01
-6.64455593e-01 -6.80094361e-01 -9.04197514e-01 1.96881413e-01
-5.09531379e-01 3.78630877e-01 -2.94860434e-02 1.12632191e+00
-9.29086924e-01 -3.02240551e-01 -4.47081715e-01 -6.09761119e-01
-4.14630473e-01 4.86909389e-01 -3.84300351e-01 -1.05154224e-01
-9.15743649e-01 9.92082775e-01 2.67844349e-01 4.58607934e-02
-3.38488221e-01 -9.22333375e-02 -7.79003143e-01 -1.26387790e-01
-7.85379559e-02 1.67730585e-01 1.32966518e+00 -9.42389667e-01
-1.53059542e+00 8.04374099e-01 -6.05545282e-01 -2.49121726e-01
6.91321418e-02 2.51228809e-01 -7.24664152e-01 8.23109299e-02
-5.99485449e-02 1.50310308e-01 7.27636099e-01 -1.09943271e+00
-8.37605894e-01 -6.20424092e-01 -6.99610412e-01 3.56330544e-01
-1.38662100e-01 6.63733244e-01 -2.18751460e-01 -3.73460770e-01
2.51197457e-01 -7.81003296e-01 2.91857660e-01 -7.89376020e-01
-6.07636049e-02 -5.29152274e-01 1.19452715e+00 -1.13903582e+00
1.26798618e+00 -2.36226535e+00 -2.85531074e-01 4.60914284e-01
-3.04478735e-01 4.52257156e-01 2.72543699e-01 7.02673912e-01
2.03825124e-02 -7.71429315e-02 7.38447607e-02 -5.60252964e-01
-1.11622913e-02 3.31330031e-01 -4.70999293e-02 3.24165165e-01
-2.49006391e-01 -1.45812675e-01 -1.44625694e-01 -4.40429449e-01
5.76725960e-01 7.83809006e-01 8.09198692e-02 2.89833248e-01
2.31329978e-01 4.59323585e-01 -1.80382863e-01 4.47339326e-01
5.13697624e-01 6.57274902e-01 -1.30526811e-01 -1.01647995e-01
-6.40207410e-01 6.51212752e-01 -1.68286967e+00 1.12876225e+00
-4.69330013e-01 6.24444008e-01 6.41487062e-01 -9.50992823e-01
1.42974651e+00 9.79193509e-01 3.93641889e-01 -3.81637871e-01
3.83170575e-01 5.75752974e-01 2.78571725e-01 -6.56681418e-01
6.06836081e-01 -1.36299103e-01 1.01398773e-01 3.70356053e-01
5.72777316e-02 -2.32163727e-01 1.99741051e-01 -1.41240031e-01
3.24609399e-01 -6.10573828e-01 3.92949998e-01 -3.62172663e-01
1.17683244e+00 -1.82476208e-01 3.74079496e-01 2.10728571e-01
-3.73248547e-01 2.26532310e-01 -2.69262820e-01 -5.29125333e-02
-9.85833049e-01 -5.23957014e-01 -1.72022015e-01 8.76919389e-01
-4.34647292e-01 2.91133579e-02 -9.42946672e-01 1.31127164e-01
-3.11894238e-01 8.13264310e-01 1.33962974e-01 8.63626301e-02
-3.72927994e-01 -2.89118737e-01 6.48685813e-01 -2.05639318e-01
3.54663730e-01 -1.20141184e+00 -2.00207680e-01 4.20551449e-01
-1.61811218e-01 -9.83071089e-01 -6.44016862e-01 4.73891012e-02
-4.84989792e-01 -6.27105236e-01 -6.35850728e-01 -1.05821085e+00
3.75178099e-01 1.67774975e-01 2.74924695e-01 -4.53805983e-01
-4.41936851e-01 6.86375260e-01 -3.22896123e-01 -7.06465423e-01
-1.24688780e+00 -1.17909871e-02 4.54674959e-01 2.80448407e-01
7.12782800e-01 -1.20965563e-01 -1.25563428e-01 2.72549599e-01
-7.55423605e-01 -6.46802485e-01 2.69665211e-01 4.92767036e-01
2.84146160e-01 5.16313970e-01 8.14092159e-01 -1.34913385e-01
1.20247018e+00 -2.01158419e-01 -5.23383498e-01 2.62019962e-01
-5.70465028e-01 -6.36214688e-02 4.12190884e-01 -3.23227406e-01
-1.18665481e+00 2.27270886e-01 -4.88265991e-01 -2.95678657e-02
-7.83263147e-01 4.83768255e-01 -4.15680498e-01 -1.40890047e-01
3.55902761e-01 5.94450653e-01 6.56526566e-01 -6.76239252e-01
-1.40057474e-01 1.60601711e+00 5.01691997e-01 -2.42648095e-01
1.74320281e-01 -6.55167252e-02 -4.90791678e-01 -1.63122714e+00
-8.95361230e-02 -1.28958821e+00 -5.52955031e-01 -4.02768463e-01
8.05110931e-01 -6.72998726e-01 -5.94581425e-01 7.66539812e-01
-1.06302500e+00 3.94130886e-01 5.36418781e-02 9.48073506e-01
-3.08203220e-01 5.28800249e-01 -2.36937940e-01 -1.54625273e+00
-6.17460072e-01 -1.36810923e+00 6.37464881e-01 3.27743024e-01
-4.88301426e-01 -7.54836202e-01 8.76562521e-02 6.81176066e-01
5.23195446e-01 -5.50438881e-01 5.93858421e-01 -1.03551805e+00
-1.82583109e-02 -5.84421694e-01 3.53691906e-01 9.73244965e-01
7.98121512e-01 1.53549045e-01 -1.11819029e+00 -1.02601886e-01
4.80954319e-01 5.52580804e-02 3.01812649e-01 3.62736732e-01
3.29197347e-01 -3.10132205e-01 2.23671541e-01 -2.51755387e-01
1.10764635e+00 9.22491491e-01 4.12746698e-01 9.46622342e-02
3.55187237e-01 1.03173161e+00 5.97200096e-01 3.20278019e-01
2.51547992e-01 7.72898614e-01 -1.58076853e-01 2.23851413e-01
-1.22445211e-01 2.18153372e-01 5.21163285e-01 1.12765205e+00
3.55916977e-01 -1.16785727e-01 -1.05083060e+00 5.35727024e-01
-1.19581199e+00 -1.06780005e+00 -3.05957403e-02 2.40654445e+00
4.69233543e-01 -1.41232476e-01 3.50214183e-01 7.59407759e-01
9.99257624e-01 -2.68811762e-01 -1.29658967e-01 -9.97173548e-01
-7.05982596e-02 1.22588590e-01 3.07353050e-01 9.61071253e-01
-9.29579496e-01 6.25156045e-01 5.77868557e+00 7.87497878e-01
-1.40671384e+00 -1.34837225e-01 3.24992210e-01 3.10373753e-01
1.15962386e-01 -3.10002327e-01 -1.11433566e+00 4.35940027e-01
1.25465834e+00 -4.06596452e-01 4.29394662e-01 6.08595848e-01
8.55427146e-01 -4.34543490e-01 -6.49447560e-01 1.12637520e+00
3.95798624e-01 -7.20791519e-01 -7.05034211e-02 2.72510588e-01
2.30305061e-01 -2.89435163e-02 1.35866016e-01 2.41461415e-02
-3.01154405e-01 -8.05482864e-01 6.24901414e-01 3.83094996e-01
2.80125767e-01 -9.42980886e-01 8.40012610e-01 6.33630931e-01
-1.09900570e+00 -7.30510503e-02 -5.79086132e-02 6.44106194e-02
1.82931691e-01 8.95377323e-02 -1.53432357e+00 2.60872185e-01
3.65997523e-01 -2.23278999e-01 -1.57288522e-01 1.05317473e+00
3.75774324e-01 8.15288424e-01 -6.04195058e-01 -1.55158997e-01
2.85568267e-01 -3.50701690e-01 7.64217496e-01 1.30879974e+00
6.71434879e-01 -1.39204729e-02 9.26626846e-02 5.47182083e-01
3.56050730e-01 7.86461711e-01 -6.54959261e-01 -3.95681560e-01
6.21511877e-01 9.89563584e-01 -5.64558327e-01 -2.43763462e-01
-4.82029170e-01 6.71352327e-01 -5.28629899e-01 1.23838075e-01
2.94202212e-02 -1.66974306e-01 7.28285789e-01 1.82838291e-01
6.01238981e-02 -3.65022659e-01 -1.07009858e-01 -3.42833430e-01
-9.30183232e-02 -1.09382880e+00 1.72083944e-01 -2.65730172e-01
-9.85469937e-01 6.50842428e-01 2.97610071e-02 -9.22057807e-01
-8.69742393e-01 -4.60279822e-01 -5.44553161e-01 1.55634880e+00
-8.49829912e-01 -9.58556831e-01 4.15866300e-02 4.11794841e-01
1.04646230e+00 -8.42009366e-01 1.10977566e+00 3.99103522e-01
-4.37097222e-01 3.11276585e-01 3.30417782e-01 -2.32327338e-02
5.13592720e-01 -8.85305166e-01 -1.60014644e-01 8.07857394e-01
1.24534518e-01 5.32118559e-01 8.65120232e-01 -4.95907009e-01
-1.18350148e+00 -5.18530071e-01 1.45103252e+00 1.82098210e-01
4.00094688e-01 -8.96075293e-02 -1.05849338e+00 9.37950835e-02
2.89383799e-01 -6.89245105e-01 9.07652199e-01 -2.83287793e-01
1.94044337e-02 -2.02296659e-01 -1.31266594e+00 1.69904187e-01
-1.77152559e-01 -9.00922239e-01 -7.69888580e-01 -1.16534077e-01
2.65658647e-01 1.37500450e-01 -6.77287042e-01 -8.39362368e-02
3.27392936e-01 -7.92487800e-01 5.52976966e-01 1.07964398e-02
-5.39812505e-01 -2.75026619e-01 -4.35627460e-01 -1.07889974e+00
1.62359416e-01 -3.92285615e-01 6.53538346e-01 1.85535741e+00
2.59194493e-01 -7.95884907e-01 5.94888926e-01 8.96899223e-01
5.68221905e-04 -1.06434599e-02 -1.16980791e+00 -6.88607097e-01
-2.92717010e-01 -3.74351919e-01 3.32386196e-01 6.98424816e-01
3.15450574e-03 3.91147316e-01 -1.21742308e-01 1.87340155e-01
5.87105811e-01 -4.20423090e-01 5.80063403e-01 -1.23613155e+00
8.08753818e-02 -5.20477593e-01 -5.48903823e-01 -4.30666238e-01
2.08055198e-01 -6.23965383e-01 2.45638669e-01 -1.42729461e+00
-2.91384786e-01 -1.40302777e-01 -8.95011052e-03 3.29426043e-02
2.08235487e-01 -2.76824534e-01 2.37047464e-01 1.97748885e-01
3.03409100e-01 2.33292729e-01 5.48656344e-01 -4.60038446e-02
-5.77965617e-01 5.44492900e-01 -2.72765338e-01 5.84256053e-01
1.12392092e+00 -2.49127537e-01 -4.74666715e-01 1.85460404e-01
-6.61489427e-01 3.76042314e-02 -4.11451571e-02 -9.53564942e-01
4.90613639e-01 1.81266338e-01 -2.30084673e-01 -8.78299892e-01
5.69060922e-01 -9.53141809e-01 2.66344070e-01 2.60013252e-01
-1.35300279e-01 2.68649310e-02 3.02911282e-01 -7.69075155e-02
-7.15828419e-01 -6.71168804e-01 8.42820704e-01 -6.08196631e-02
-8.73232722e-01 -2.93079287e-01 -1.10519958e+00 -6.44385159e-01
9.47819412e-01 -4.87171054e-01 1.73574314e-01 -6.78500056e-01
-7.90538490e-01 -2.02691421e-01 1.13106303e-01 5.92580974e-01
5.13189018e-01 -9.50401306e-01 -6.49451554e-01 4.83968139e-01
5.37167229e-02 -3.93547058e-01 2.11089551e-01 6.11923039e-01
-4.88441944e-01 8.65710497e-01 -1.47569641e-01 -3.26666683e-01
-2.19908500e+00 2.15443522e-01 3.83779883e-01 2.41902143e-01
-1.13834679e-01 5.36386073e-01 -2.72170603e-01 -3.83955508e-01
4.98377532e-01 -9.06286389e-02 -6.70563042e-01 2.96420723e-01
7.10154533e-01 4.55480963e-01 2.85624236e-01 -1.35214376e+00
-3.43233109e-01 3.06363553e-01 -1.20583043e-01 -6.32147908e-01
9.84385967e-01 -6.33771002e-01 -2.78724343e-01 8.91867280e-01
1.21631813e+00 2.26432070e-01 -3.69318008e-01 -1.72468632e-01
3.13651383e-01 -1.26339853e-01 3.43136877e-01 -6.45659089e-01
-3.51404011e-01 7.86872029e-01 1.05269349e+00 3.91163051e-01
1.03386068e+00 -1.01743326e-01 3.26514393e-01 2.93949038e-01
1.67934507e-01 -1.52537119e+00 -5.51291227e-01 5.97219050e-01
8.85188043e-01 -1.21660674e+00 -5.62401652e-01 -3.97092223e-01
-5.46472251e-01 1.13199222e+00 1.77121252e-01 3.57180685e-01
7.43082941e-01 4.38712269e-01 5.79692245e-01 1.66111737e-01
-2.61667371e-01 -4.11942035e-01 4.31541562e-01 4.94783103e-01
8.00880075e-01 6.93499297e-03 -8.09683859e-01 3.86344701e-01
-2.53543049e-01 -1.98970616e-01 3.14771175e-01 7.68301666e-01
-8.15551400e-01 -1.51833653e+00 -1.14101338e+00 2.96318471e-01
-7.08095610e-01 9.03209448e-02 -6.54767513e-01 3.92789751e-01
-1.25416424e-02 1.49662852e+00 -7.90490285e-02 -1.58678457e-01
1.87796310e-01 6.23897135e-01 1.16531722e-01 -6.55917883e-01
-5.67325056e-01 5.31689763e-01 1.52202919e-01 9.24680606e-02
-4.91726041e-01 -9.42052782e-01 -1.45235467e+00 -1.73614323e-01
-1.68635413e-01 6.22262836e-01 1.66661251e+00 1.00548959e+00
-2.15547383e-02 3.11156482e-01 7.57413208e-01 -6.75574064e-01
-7.09194541e-01 -1.17443228e+00 -7.92959630e-01 1.35799572e-01
2.91002333e-01 -2.72962689e-01 -4.73320574e-01 1.92837343e-01] | [14.445891380310059, 6.080620288848877] |
03f11192-11a9-4e98-9836-b5fcf712b78f | nev-ncd-negative-learning-entropy-and | 2304.07354 | null | https://arxiv.org/abs/2304.07354v1 | https://arxiv.org/pdf/2304.07354v1.pdf | NEV-NCD: Negative Learning, Entropy, and Variance regularization based novel action categories discovery | Novel Categories Discovery (NCD) facilitates learning from a partially annotated label space and enables deep learning (DL) models to operate in an open-world setting by identifying and differentiating instances of novel classes based on the labeled data notions. One of the primary assumptions of NCD is that the novel label space is perfectly disjoint and can be equipartitioned, but it is rarely realized by most NCD approaches in practice. To better align with this assumption, we propose a novel single-stage joint optimization-based NCD method, Negative learning, Entropy, and Variance regularization NCD (NEV-NCD). We demonstrate the efficacy of NEV-NCD in previously unexplored NCD applications of video action recognition (VAR) with the public UCF101 dataset and a curated in-house partial action-space annotated multi-view video dataset. We perform a thorough ablation study by varying the composition of final joint loss and associated hyper-parameters. During our experiments with UCF101 and multi-view action dataset, NEV-NCD achieves ~ 83% classification accuracy in test instances of labeled data. NEV-NCD achieves ~ 70% clustering accuracy over unlabeled data outperforming both naive baselines (by ~ 40%) and state-of-the-art pseudo-labeling-based approaches (by ~ 3.5%) over both datasets. Further, we propose to incorporate optional view-invariant feature learning with the multiview dataset to identify novel categories from novel viewpoints. Our additional view-invariance constraint improves the discriminative accuracy for both known and unknown categories by ~ 10% for novel viewpoints. | ['Nirmalya Roy', 'Hyungtae Lee', 'Heesung Kwon', 'Sanjay Purushotham', 'Abu Zaher Md Faridee', 'Masud Ahmed', 'Zahid Hasan'] | 2023-04-14 | null | null | null | null | ['action-recognition-in-videos', 'action-recognition'] | ['computer-vision', 'computer-vision'] | [ 4.16261405e-01 2.79754512e-02 -5.95948756e-01 -5.36281705e-01
-1.27372766e+00 -7.64280558e-01 6.31704211e-01 -2.59576797e-01
-4.28216368e-01 6.47328973e-01 3.79752159e-01 9.88028795e-02
1.47943720e-02 -1.69246256e-01 -7.66813695e-01 -5.80231965e-01
-1.12966202e-01 5.26672721e-01 8.10223892e-02 4.49526936e-01
-1.03219308e-01 4.44203913e-01 -1.79076731e+00 5.85758388e-01
4.74624962e-01 1.02794600e+00 -1.94126919e-01 4.07971978e-01
3.14402759e-01 6.79608285e-01 -2.76298255e-01 -2.29327485e-01
7.91954994e-01 -2.87126005e-01 -7.80824184e-01 4.58731294e-01
1.00888443e+00 -4.97403651e-01 -2.17269704e-01 7.97165155e-01
4.50073332e-01 4.22537893e-01 9.10472572e-01 -1.65707767e+00
-5.40595353e-01 2.93854237e-01 -7.51101196e-01 1.76359311e-01
3.15747917e-01 3.86619687e-01 1.23246026e+00 -9.85742688e-01
1.07575893e+00 1.20299554e+00 7.43473172e-01 8.29321623e-01
-1.56164801e+00 -6.60224795e-01 4.21877503e-01 2.52180129e-01
-1.32850218e+00 -5.17760456e-01 5.33294499e-01 -6.07663453e-01
9.79275286e-01 7.08842054e-02 4.63878185e-01 1.47977352e+00
-1.80923268e-01 1.15403688e+00 1.07420886e+00 -2.17689112e-01
5.05845785e-01 -1.14190280e-02 4.13752422e-02 6.22011185e-01
3.76032852e-02 -2.04151887e-02 -5.12847602e-01 -1.67856961e-01
4.41755205e-01 1.32489085e-01 -1.85757115e-01 -1.05293202e+00
-1.26557422e+00 7.57115126e-01 1.40781850e-01 -4.62446846e-02
-5.09679578e-02 1.37428597e-01 7.21683979e-01 2.26130247e-01
5.85463881e-01 5.42042851e-01 -8.85063648e-01 -3.01791221e-01
-7.68105030e-01 2.11893886e-01 4.64615732e-01 1.04555333e+00
5.30429780e-01 -7.10206777e-02 -2.66459167e-01 7.90656209e-01
2.21659280e-02 4.64781523e-01 3.96183699e-01 -1.36334777e+00
3.63077492e-01 6.55768692e-01 -7.61958286e-02 -4.87835765e-01
-5.06204486e-01 -4.61924702e-01 -5.66654325e-01 1.92117810e-01
4.65924472e-01 2.54433248e-02 -1.00814903e+00 2.11419249e+00
5.02857625e-01 2.57402003e-01 1.29021063e-01 6.99792624e-01
6.31150544e-01 6.61268458e-02 5.54434676e-03 -2.75937378e-01
8.62294674e-01 -1.01775479e+00 -3.74415189e-01 3.99008580e-02
1.01667333e+00 -2.42297053e-01 1.23487258e+00 3.63161594e-01
-5.47304392e-01 -5.28135061e-01 -7.76493728e-01 -4.52645645e-02
-1.75275147e-01 3.22382063e-01 5.50724447e-01 3.90571892e-01
-8.08955789e-01 4.05605108e-01 -9.77669537e-01 -5.69503248e-01
1.01656830e+00 2.94009686e-01 -7.71078169e-01 -4.13986534e-01
-6.45895541e-01 5.42219281e-01 3.29241246e-01 -2.92885512e-01
-1.54867387e+00 -1.01061523e+00 -9.45132136e-01 -2.64821291e-01
7.21243501e-01 -6.04440033e-01 1.04181886e+00 -1.12463069e+00
-9.24583793e-01 1.09248471e+00 -1.40249386e-01 -4.99169171e-01
6.17059827e-01 -3.56455535e-01 -3.55615377e-01 3.06803882e-01
5.16839623e-01 1.03118694e+00 7.47036219e-01 -1.39458966e+00
-5.98248065e-01 -4.33938801e-01 2.45976359e-01 4.03966665e-01
-3.29386562e-01 -5.06826282e-01 -2.52354592e-01 -5.82526624e-01
1.32787019e-01 -1.21971786e+00 -1.75212417e-02 3.17887872e-01
-2.99533844e-01 -3.27935547e-01 9.96706188e-01 -4.53374624e-01
8.04806352e-01 -2.36162686e+00 2.47399718e-01 -1.05431952e-01
3.03364128e-01 1.00716017e-01 -2.85149217e-01 -3.27457711e-02
-2.58063495e-01 -7.35157430e-02 -2.22247422e-01 -4.81350869e-01
4.81709652e-02 2.46466577e-01 -8.38770047e-02 6.82454526e-01
1.92905501e-01 7.81981885e-01 -9.90202427e-01 -2.91994482e-01
3.09517264e-01 2.17756361e-01 -8.10862541e-01 1.02995336e-01
-3.36870492e-01 5.52189708e-01 -7.04237074e-02 1.03327966e+00
5.27064085e-01 -3.88920307e-01 3.51070076e-01 -4.34358567e-01
3.20317179e-01 -1.59495354e-01 -1.22454250e+00 2.03478408e+00
-1.74028784e-01 5.80149293e-01 -1.95884258e-01 -9.92314279e-01
4.81002003e-01 1.66053578e-01 7.85592258e-01 -3.44014734e-01
-1.10113958e-03 -8.73731636e-03 -2.62630492e-01 -5.48343360e-01
6.21105693e-02 -1.66737944e-01 -5.57143316e-02 5.22102833e-01
5.33732057e-01 3.75633419e-01 1.07090883e-01 3.55948597e-01
1.44616473e+00 4.39503282e-01 2.97165155e-01 -2.17465878e-01
3.00781816e-01 -7.37161934e-02 9.53566015e-01 7.45624244e-01
-6.67405486e-01 6.37665093e-01 4.54835445e-01 -4.18968141e-01
-8.92633498e-01 -1.22181857e+00 -2.72283733e-01 1.21110320e+00
-6.90582581e-03 -2.67849475e-01 -5.64571321e-01 -1.51606882e+00
1.90569729e-01 8.04816365e-01 -8.69235694e-01 -2.60412306e-01
-2.38801941e-01 -5.05672157e-01 4.56192076e-01 7.32587516e-01
4.56743926e-01 -6.72927201e-01 -2.99720943e-01 -1.55147463e-01
-2.21968949e-01 -1.36394918e+00 -3.01801443e-01 3.71952951e-01
-7.99593627e-01 -1.36893082e+00 -3.92611563e-01 -5.87055624e-01
7.19150126e-01 3.33328843e-01 1.13088834e+00 -3.67266089e-01
-3.03745776e-01 9.31653559e-01 -5.08787811e-01 6.73303520e-03
-2.66072005e-01 9.80646014e-02 3.94691229e-01 1.09563597e-01
6.65398955e-01 -6.01244390e-01 -4.72452551e-01 3.39010715e-01
-7.54995644e-01 -8.00146088e-02 4.20725822e-01 9.01687205e-01
7.85274327e-01 -1.40704721e-01 7.18764782e-01 -7.52469480e-01
-2.75989860e-01 -6.20437860e-01 -2.43382692e-01 2.31523708e-01
-4.96840537e-01 -3.52031551e-02 5.25808156e-01 -5.69764912e-01
-9.01358306e-01 3.28275174e-01 1.99630022e-01 -1.00195372e+00
-5.11799276e-01 8.64752308e-02 -5.16970873e-01 1.93344370e-01
6.92242146e-01 -7.27129579e-02 1.52519243e-02 -4.09011215e-01
5.35737634e-01 4.84781682e-01 4.00140673e-01 -5.56906760e-01
6.13470852e-01 8.57701182e-01 -5.47005981e-02 -4.59473163e-01
-1.22798526e+00 -6.41247451e-01 -1.20098591e+00 -2.93025404e-01
1.14783037e+00 -1.41042233e+00 -4.67231303e-01 3.11256051e-01
-5.25372684e-01 -4.43273008e-01 -6.83733225e-01 6.54672563e-01
-7.52680063e-01 4.47852910e-01 -2.66228408e-01 -5.57976782e-01
3.73007841e-02 -1.15115571e+00 1.31258547e+00 -1.80576056e-01
-2.61753172e-01 -8.79134119e-01 -1.96978394e-02 8.11705053e-01
-1.26982972e-01 4.58321303e-01 9.07933652e-01 -1.12392211e+00
-5.57397723e-01 -1.46448724e-02 -1.19587958e-01 7.71550775e-01
2.64336675e-01 -3.36066514e-01 -1.15707219e+00 -5.83268762e-01
-2.94565678e-01 -1.01387846e+00 1.16213918e+00 3.37880552e-01
1.15418637e+00 1.27761185e-01 -3.91997486e-01 6.21694565e-01
1.28907156e+00 7.77207837e-02 1.98970988e-01 3.41870189e-01
9.18220997e-01 2.83618510e-01 8.68540883e-01 6.06264591e-01
4.33429062e-01 6.28253341e-01 6.38271391e-01 1.16947152e-01
-2.02761665e-01 -1.73847407e-01 4.62540746e-01 2.70713508e-01
2.76063085e-01 -2.23812535e-01 -7.60947108e-01 6.67749941e-01
-1.81701112e+00 -1.02523017e+00 2.98566967e-01 2.22521472e+00
6.62643850e-01 1.24130890e-01 2.18722314e-01 -1.24743529e-01
6.78083897e-01 1.81390822e-01 -1.10496974e+00 7.33218864e-02
-1.51388332e-01 -1.41593158e-01 3.71038258e-01 1.20381288e-01
-1.69692135e+00 8.03300440e-01 5.73196363e+00 8.79081726e-01
-7.00426102e-01 2.04788893e-01 6.61364913e-01 -6.15998805e-01
-1.87853590e-01 -5.24694361e-02 -8.70055139e-01 1.50557831e-01
6.83962643e-01 2.19437405e-01 2.70662397e-01 1.11084020e+00
1.89358182e-02 -1.06958702e-01 -1.61921227e+00 1.07285702e+00
4.45821792e-01 -1.14870846e+00 7.03815818e-02 7.80576915e-02
1.07335842e+00 2.95640171e-01 7.51590878e-02 5.96726477e-01
4.12665814e-01 -7.30224967e-01 4.99374479e-01 2.31683761e-01
1.11729801e+00 -6.47514880e-01 5.07838190e-01 1.99638575e-01
-1.01269054e+00 -4.69883114e-01 -2.08192617e-01 8.93012658e-02
7.65907690e-02 4.52792794e-01 -7.16909587e-01 4.57731605e-01
8.35782945e-01 1.37638557e+00 -7.05157220e-01 5.47728717e-01
6.41301796e-02 7.07135737e-01 -2.98475921e-01 6.00750387e-01
2.99289107e-01 -1.29661085e-02 5.94338596e-01 7.84752548e-01
-4.74857837e-02 5.77224419e-02 4.17433619e-01 6.52139902e-01
-3.67872775e-01 -2.04338476e-01 -6.86879933e-01 7.14727342e-02
3.94028902e-01 1.07996476e+00 -6.82326496e-01 -3.79893303e-01
-5.82630694e-01 9.61582243e-01 3.45953971e-01 3.82784337e-01
-9.97178137e-01 3.79222810e-01 8.54691803e-01 -1.39460359e-02
6.00064754e-01 8.25089365e-02 4.77934116e-03 -1.59365642e+00
8.55590925e-02 -1.17757165e+00 6.70719504e-01 -5.03426790e-01
-1.57018268e+00 3.29996258e-01 1.68324947e-01 -1.59612191e+00
-7.65616447e-02 -4.80036914e-01 5.27279451e-02 1.99093416e-01
-1.18159688e+00 -1.31266546e+00 -1.80267528e-01 6.30753338e-01
8.13091159e-01 -4.63150740e-01 8.60714734e-01 3.49141210e-01
-6.81410968e-01 7.81748116e-01 5.51648140e-01 1.91145062e-01
1.03856754e+00 -1.23292673e+00 1.64157674e-01 8.15730870e-01
2.55976677e-01 3.14122677e-01 3.05680215e-01 -6.50108874e-01
-1.14875174e+00 -1.57210302e+00 1.99362978e-01 -9.27587450e-01
5.09694040e-01 -6.29877329e-01 -6.36486948e-01 1.09928453e+00
-2.23382533e-01 4.86902535e-01 1.12207580e+00 1.80672780e-01
-8.18492770e-01 -1.15695126e-01 -1.31579554e+00 4.43441153e-01
1.61748528e+00 -5.69332421e-01 -4.97920930e-01 6.41984403e-01
8.66566181e-01 -9.88720730e-02 -9.77904797e-01 8.20379674e-01
6.65861964e-01 -1.02773845e+00 8.98798823e-01 -9.29915965e-01
3.47729117e-01 -3.31241518e-01 -7.34111488e-01 -1.01800787e+00
-2.82519698e-01 -2.21722811e-01 -2.16059163e-01 1.27395427e+00
2.65970588e-01 -2.98310697e-01 9.14014220e-01 6.55036390e-01
-2.65698314e-01 -8.34029138e-01 -1.04665399e+00 -9.82119739e-01
-4.40482050e-03 -6.40863001e-01 8.73835981e-02 1.19927049e+00
-4.17240024e-01 2.71629065e-01 -3.21678221e-01 4.48848940e-02
8.08180690e-01 1.53795302e-01 9.23805833e-01 -1.02974284e+00
-3.45482409e-01 1.42462999e-02 -7.06552923e-01 -9.68975902e-01
6.45722866e-01 -1.07289219e+00 -2.50011951e-01 -1.34761691e+00
5.96665680e-01 -4.12911236e-01 -4.59772646e-01 7.46869981e-01
-1.75153948e-02 3.97261113e-01 2.05338985e-01 1.92165852e-01
-1.18169963e+00 8.04676354e-01 8.16711605e-01 -2.44260103e-01
-1.26216382e-01 -2.36805558e-01 -5.44402719e-01 8.58443439e-01
4.68453884e-01 -5.37455916e-01 -6.84833109e-01 -3.24047536e-01
-1.63295209e-01 -3.89626503e-01 4.72539812e-01 -1.04539657e+00
-3.36626261e-01 -1.52902514e-01 5.82854867e-01 -6.97600067e-01
3.73843998e-01 -9.89046156e-01 1.74615514e-02 1.47672996e-01
-7.41136014e-01 -3.65000755e-01 9.26487371e-02 9.49241221e-01
6.90629855e-02 1.62039369e-01 8.06681991e-01 -7.06388205e-02
-1.11469305e+00 3.65733683e-01 -1.24867275e-01 5.02874792e-01
1.30574572e+00 -3.99348766e-01 -3.60580921e-01 -2.00531855e-01
-9.48803008e-01 2.54833728e-01 7.26658583e-01 4.94376272e-01
4.96608198e-01 -1.50787044e+00 -5.51804602e-01 2.27091834e-01
5.17383754e-01 -5.80663905e-02 4.80767787e-01 1.00422108e+00
-8.69093165e-02 1.33077055e-01 -1.66062698e-01 -1.00616097e+00
-1.42634702e+00 7.02020586e-01 2.40810618e-01 -1.63220122e-01
-6.27745211e-01 9.62157071e-01 4.37876403e-01 -7.22323298e-01
4.13199127e-01 -9.78152752e-02 -4.64705825e-02 1.88581303e-01
3.30119520e-01 6.24902487e-01 -7.40229413e-02 -7.37805784e-01
-5.97456157e-01 3.90275985e-01 -2.80017853e-01 9.70133990e-02
1.34673691e+00 -2.56117105e-01 2.33240753e-01 6.57078385e-01
1.64717734e+00 -3.57768595e-01 -1.71829498e+00 -2.95264661e-01
-2.08113462e-01 -5.76786935e-01 -1.69272155e-01 -9.35845912e-01
-1.01454568e+00 5.62640846e-01 8.84706676e-01 -4.16479617e-01
1.11254883e+00 2.72235751e-01 5.53045809e-01 4.67350632e-01
3.92571032e-01 -1.14632797e+00 3.81715149e-01 3.56563359e-01
6.35951936e-01 -1.57613885e+00 2.54099876e-01 -2.26226807e-01
-9.38512087e-01 7.87471354e-01 9.55457747e-01 6.94711208e-02
6.17634118e-01 -1.93602089e-02 1.11614443e-01 -1.32735237e-01
-9.48136628e-01 -2.50855446e-01 3.12020302e-01 7.53214061e-01
4.21434008e-02 -3.29867899e-02 1.10196941e-01 5.01999438e-01
3.37045163e-01 -7.33774006e-02 3.97464722e-01 1.04032218e+00
-8.13862607e-02 -7.91131198e-01 9.82274339e-02 7.02831388e-01
-3.00007790e-01 7.16099665e-02 -2.78118461e-01 8.84107113e-01
4.14180130e-01 7.73746192e-01 1.16696268e-01 -5.97241461e-01
1.51831135e-01 2.98493326e-01 3.89730573e-01 -7.99767375e-01
-1.58176124e-01 -3.84381320e-03 1.25212222e-01 -8.51774216e-01
-8.57149780e-01 -1.18896651e+00 -1.21969271e+00 1.53221250e-01
-4.35094059e-01 -2.64916897e-01 2.00128779e-01 1.06318808e+00
6.84875786e-01 2.74979830e-01 6.82797194e-01 -6.23156309e-01
-6.81466877e-01 -7.79661059e-01 -6.37624443e-01 7.08179653e-01
1.51266381e-01 -1.14587748e+00 -7.19035864e-01 2.96686769e-01] | [8.826106071472168, 1.0871411561965942] |
9162934c-3bb1-4c3a-802b-82f5c31e583d | data-questeval-a-referenceless-metric-for | 2104.07555 | null | https://arxiv.org/abs/2104.07555v3 | https://arxiv.org/pdf/2104.07555v3.pdf | Data-QuestEval: A Referenceless Metric for Data-to-Text Semantic Evaluation | QuestEval is a reference-less metric used in text-to-text tasks, that compares the generated summaries directly to the source text, by automatically asking and answering questions. Its adaptation to Data-to-Text tasks is not straightforward, as it requires multimodal Question Generation and Answering systems on the considered tasks, which are seldom available. To this purpose, we propose a method to build synthetic multimodal corpora enabling to train multimodal components for a data-QuestEval metric. The resulting metric is reference-less and multimodal; it obtains state-of-the-art correlations with human judgment on the WebNLG and WikiBio benchmarks. We make data-QuestEval's code and models available for reproducibility purpose, as part of the QuestEval project. | ['Patrick Gallinari', 'Geoffrey Scoutheeten', 'Jacopo Staiano', 'Sylvain Lamprier', 'Benjamin Piwowarski', 'Laure Soulier', 'Thomas Scialom', 'Clément Rebuffel'] | 2021-04-15 | null | https://aclanthology.org/2021.emnlp-main.633 | https://aclanthology.org/2021.emnlp-main.633.pdf | emnlp-2021-11 | ['data-to-text-generation'] | ['natural-language-processing'] | [ 8.86074081e-02 3.99631172e-01 8.78406763e-02 -4.12734389e-01
-1.42291725e+00 -7.27371633e-01 1.35029447e+00 7.24318564e-01
-6.04210556e-01 9.70051169e-01 6.20519817e-01 -4.87439819e-02
-3.47617090e-01 -5.69619656e-01 -3.78022820e-01 -1.82830229e-01
2.24560052e-01 1.21552753e+00 1.01507813e-01 -6.39179766e-01
4.57651705e-01 -8.46193433e-02 -1.71142697e+00 8.49299550e-01
1.00571132e+00 8.71049047e-01 9.17986706e-02 1.11449206e+00
-6.00487471e-01 9.27004397e-01 -7.64357626e-01 -8.60713720e-01
-4.79197711e-01 -7.67343879e-01 -1.18695438e+00 -2.77435720e-01
6.47977114e-01 1.53856665e-01 7.03914687e-02 7.25816309e-01
6.67413771e-01 9.89522859e-02 1.24193311e+00 -1.08799362e+00
-5.35301626e-01 8.21927905e-01 2.16907203e-01 -1.40228271e-01
1.04928029e+00 -8.62529327e-04 1.15762401e+00 -1.06596398e+00
8.68350565e-01 1.38005853e+00 2.68550277e-01 6.52767122e-01
-1.45218730e+00 1.20767519e-01 -6.23047352e-01 2.27126569e-01
-9.81607020e-01 -6.26245081e-01 4.52448547e-01 -5.14058471e-01
7.27057993e-01 6.53966784e-01 -4.56204563e-02 1.59135246e+00
-2.13959590e-01 8.03980231e-01 1.16502452e+00 -4.78980035e-01
1.39499098e-01 4.11987692e-01 2.48603061e-01 4.32628423e-01
-1.24043047e-01 -3.49722892e-01 -8.77399802e-01 -8.04653019e-02
1.78087445e-03 -7.89808929e-01 -3.87428135e-01 -1.26646593e-01
-1.63052249e+00 8.26969326e-01 -4.14140560e-02 5.79759240e-01
-3.90130550e-01 -2.24989489e-01 7.13732243e-01 7.32763708e-01
4.40682471e-01 7.59915888e-01 -2.26398543e-01 -5.78649044e-01
-8.58647466e-01 6.33548737e-01 1.44711673e+00 8.43528688e-01
6.03284240e-01 -5.19128203e-01 -7.51608849e-01 9.13091183e-01
2.45227545e-01 6.12346411e-01 6.07116044e-01 -9.37962651e-01
9.07847583e-01 6.75808668e-01 3.87099624e-01 -8.40260506e-01
-6.02056205e-01 -1.46221872e-02 -7.09909558e-01 -9.22294185e-02
7.35914826e-01 -3.00350934e-01 -3.28722417e-01 1.45250583e+00
1.69421569e-01 -4.07223493e-01 6.94877267e-01 7.36050427e-01
1.65619600e+00 6.67698562e-01 -9.02349502e-02 -2.53634691e-01
1.41963518e+00 -8.06606233e-01 -1.06063223e+00 3.66529226e-02
8.55067372e-01 -1.04379416e+00 1.38831222e+00 4.22185630e-01
-1.19694698e+00 -5.78176677e-01 -7.30926812e-01 -2.50111878e-01
-7.59211540e-01 3.84622693e-01 -3.96878161e-02 5.00632763e-01
-9.64953184e-01 5.21087885e-01 -1.35487914e-01 -7.09251463e-01
-2.00088516e-01 -1.77126870e-01 -5.25588810e-01 1.73826382e-01
-1.27001274e+00 1.14868820e+00 4.75081712e-01 -1.92443684e-01
-7.48709202e-01 -3.01580667e-01 -8.20016086e-01 -1.23122297e-01
4.97542173e-01 -8.01855385e-01 1.48382068e+00 -5.81843376e-01
-1.54152989e+00 1.15857375e+00 -1.96762189e-01 -3.24893266e-01
8.64744663e-01 -1.91202715e-01 -5.20149291e-01 2.05799609e-01
1.19823730e-02 6.09214664e-01 7.18969464e-01 -1.33234465e+00
-1.15337156e-01 -2.02027321e-01 -1.06281862e-01 1.41776010e-01
-2.70642042e-01 -1.61103979e-01 -3.55885535e-01 -3.71020436e-01
-5.32457411e-01 -3.98043543e-01 2.07525030e-01 -4.57946301e-01
-7.76064098e-01 -5.38046360e-01 5.38243115e-01 -6.56383812e-01
1.19834030e+00 -1.80075049e+00 4.77359951e-01 4.75137495e-03
1.03952065e-01 2.25501209e-01 -6.86071575e-01 1.17392862e+00
1.95346206e-01 2.53209304e-02 -3.32203329e-01 -8.25883210e-01
4.70894188e-01 2.40682028e-02 -2.34268874e-01 -8.89502987e-02
1.54729053e-01 9.89242733e-01 -1.03964162e+00 -8.49424720e-01
2.16831237e-01 1.37352720e-01 -1.95163518e-01 6.67717874e-01
-6.40833557e-01 3.84414613e-01 -3.51436406e-01 2.58258641e-01
2.46712074e-01 1.43221505e-02 -9.15790573e-02 -2.31774524e-01
-2.29847834e-01 2.59308994e-01 -8.07647109e-01 2.06382871e+00
-5.17290950e-01 7.39759266e-01 -1.79720566e-01 -6.07127905e-01
1.37085140e+00 5.17648458e-01 1.65843278e-01 -8.22607517e-01
2.94217259e-01 4.01949912e-01 -3.94203395e-01 -1.09428596e+00
9.39636052e-01 -2.05085310e-03 -2.18213171e-01 5.58303177e-01
4.32442725e-01 -6.33308589e-01 9.30558324e-01 2.57800162e-01
1.04846978e+00 2.36317456e-01 1.72432780e-01 -1.79483280e-01
8.91123950e-01 7.96737596e-02 -6.07774198e-01 7.29841888e-01
3.61645341e-01 7.11170733e-01 8.87813687e-01 7.81198740e-02
-8.06477606e-01 -9.32401121e-01 5.84458336e-02 1.11346090e+00
-2.42463961e-01 -6.21576965e-01 -1.04631221e+00 -8.20300460e-01
-2.84354180e-01 1.30095363e+00 -6.77232385e-01 1.51736587e-01
4.98213358e-02 -3.62852991e-01 7.12094486e-01 -8.26915950e-02
-1.11426832e-02 -1.20694304e+00 -4.44949985e-01 2.97363669e-01
-6.07343853e-01 -1.24166131e+00 -2.11294353e-01 3.04397419e-02
-5.92890799e-01 -1.21885383e+00 -9.75883126e-01 -3.46340001e-01
1.23303674e-01 -4.23736334e-01 1.75819659e+00 -2.20660076e-01
3.72253507e-02 7.12134182e-01 -7.82071054e-01 -3.59583288e-01
-1.07217467e+00 4.75304663e-01 -5.35898685e-01 1.26812711e-01
2.43763924e-01 -3.45051140e-01 -2.31415749e-01 3.47056895e-01
-1.27937663e+00 6.47275746e-02 5.33038795e-01 6.04507267e-01
8.16007331e-02 -7.73205578e-01 8.27936828e-01 -8.05857718e-01
1.35811174e+00 -4.34941292e-01 -2.25380108e-01 3.86768579e-01
-2.84489572e-01 3.67727548e-01 5.97254276e-01 -2.13656723e-01
-9.15107369e-01 -2.95297414e-01 -4.97560203e-01 3.71751711e-02
-4.83572215e-01 6.67845905e-01 -2.02331707e-01 4.36299413e-01
9.59884346e-01 6.39745072e-02 -1.11305319e-01 -7.33427525e-01
7.92835772e-01 8.60627651e-01 6.91895902e-01 -5.63463807e-01
5.32103717e-01 -7.42621273e-02 -4.76110540e-02 -9.35488284e-01
-7.55611777e-01 -4.07093495e-01 -4.75817919e-01 -4.23608482e-01
1.05751848e+00 -5.47062159e-01 -9.18833852e-01 1.24583907e-01
-1.66137850e+00 -3.21039766e-01 -3.64367723e-01 3.32626134e-01
-8.05121839e-01 2.72050291e-01 -1.98679209e-01 -7.75768638e-01
-4.31049377e-01 -8.57039750e-01 1.24161172e+00 7.91773349e-02
-5.94292402e-01 -1.09373641e+00 7.45740056e-01 5.32899976e-01
3.81888568e-01 4.23835814e-01 9.13530469e-01 -1.21759033e+00
4.04615607e-03 -2.79662669e-01 -3.25468123e-01 3.68277222e-01
-1.63835600e-01 1.95552021e-01 -1.08334017e+00 2.15928346e-01
-3.12583029e-01 -8.75344336e-01 8.09475720e-01 -2.30831310e-01
7.89966464e-01 -3.80541891e-01 9.60950330e-02 -2.17830688e-01
1.25431645e+00 -3.15706372e-01 6.71628356e-01 -1.85169633e-02
3.63669842e-01 1.24945080e+00 7.05390811e-01 5.12234807e-01
7.57936537e-01 8.19031537e-01 5.49899876e-01 1.50162309e-01
-2.38507599e-01 -1.90850973e-01 3.20138991e-01 9.92039979e-01
4.28357385e-02 -7.61880279e-01 -9.58181202e-01 5.72023809e-01
-2.11940765e+00 -9.49817717e-01 -8.19515526e-01 2.10394979e+00
9.85341072e-01 -7.32888430e-02 1.54955328e-01 -1.64285451e-02
3.17231208e-01 2.68281877e-01 2.12138712e-01 -5.83072960e-01
-3.22527319e-01 9.24652219e-02 -3.03452492e-01 7.33867168e-01
-7.57616401e-01 5.41581690e-01 5.67588568e+00 9.23638582e-01
-5.72308302e-01 2.54326344e-01 1.94992140e-01 1.57909155e-01
-5.99876523e-01 -2.56810665e-01 -5.66319525e-01 2.00147554e-01
1.32102203e+00 -1.80844799e-01 2.95156300e-01 4.17360544e-01
2.99989969e-01 -2.74292201e-01 -1.75902212e+00 1.12086844e+00
4.72384602e-01 -1.13917196e+00 2.36497313e-01 -3.64669353e-01
3.89588237e-01 -8.62820968e-02 -1.96231544e-01 3.95660818e-01
-4.62136529e-02 -1.08603895e+00 7.53469110e-01 1.06539118e+00
5.17062485e-01 -7.46588230e-01 1.04173219e+00 4.36269194e-01
-7.23635733e-01 3.15070421e-01 -1.58572987e-01 2.85523206e-01
4.94041502e-01 6.11386001e-01 -8.62969160e-01 8.73037755e-01
1.21990614e-01 4.75868136e-01 -1.08282912e+00 7.53521204e-01
-7.65518174e-02 3.15372825e-01 1.41518444e-01 -6.26252592e-01
2.81159848e-01 -1.90446585e-01 8.50208282e-01 1.60886240e+00
5.54810464e-01 -4.83925462e-01 -1.71913251e-01 9.21990871e-01
-3.12581748e-01 7.59341121e-01 -8.15422177e-01 -3.29003751e-01
1.03729956e-01 1.54028833e+00 -1.74581334e-01 -4.05581832e-01
-1.49609432e-01 9.75708425e-01 3.01984221e-01 2.02659115e-01
-4.07310814e-01 -7.38126814e-01 -1.39468655e-01 -8.02464038e-02
-1.83916971e-01 -1.67058110e-01 -1.57505766e-01 -9.51888621e-01
3.66449617e-02 -1.14012659e+00 3.69483322e-01 -1.06225824e+00
-1.49578428e+00 1.01087952e+00 2.14930534e-01 -1.05956030e+00
-9.32439744e-01 -6.54228628e-01 -5.50948441e-01 7.39953578e-01
-1.19708693e+00 -9.98156607e-01 -6.23460352e-01 6.40772998e-01
4.01521295e-01 -3.40728581e-01 1.13495970e+00 2.70720333e-01
-1.38509169e-01 3.71864498e-01 -2.83629540e-02 -6.56661466e-02
1.21197128e+00 -1.53175175e+00 1.25583947e-01 3.18004161e-01
3.07015836e-01 1.32625997e-01 1.31368017e+00 -2.73218721e-01
-1.42528546e+00 -8.17113400e-01 1.53162074e+00 -9.42159593e-01
9.81995046e-01 -3.04848433e-01 -9.77977693e-01 2.41610166e-02
1.03673565e+00 -7.54374385e-01 7.61594057e-01 -6.14973307e-02
-2.22179770e-01 3.60751599e-02 -8.88217390e-01 5.69889843e-01
6.52674794e-01 -7.03476071e-01 -9.98699427e-01 4.97177422e-01
6.62869692e-01 -1.83896601e-01 -1.08983541e+00 2.41034836e-01
2.05787852e-01 -1.16853714e+00 5.35346746e-01 -5.24568141e-01
8.27992618e-01 -2.46831805e-01 -3.90133530e-01 -1.50561213e+00
6.33024156e-01 -7.86420524e-01 -6.67513758e-02 1.71975470e+00
9.01415884e-01 -4.47484314e-01 2.44148761e-01 4.21240062e-01
-1.01975679e-01 -2.75015622e-01 -1.05470777e+00 -5.44000804e-01
2.52235401e-02 -5.36898792e-01 5.64683557e-01 6.50940537e-01
9.46071669e-02 9.55985367e-01 -2.28102058e-01 -4.10485238e-01
3.68850112e-01 -1.01117797e-01 1.20773530e+00 -1.25924492e+00
-2.18088597e-01 -6.47064507e-01 1.64129622e-02 -6.61479414e-01
2.22193822e-01 -8.76961648e-01 1.10058174e-01 -1.74758577e+00
3.13984044e-02 3.21807742e-01 3.34747314e-01 2.44386584e-01
1.79773588e-02 2.32051268e-01 3.25433880e-01 -8.14218000e-02
-9.40730989e-01 8.17385912e-01 1.22690749e+00 -1.72333002e-01
-9.61120874e-02 -2.15490937e-01 -4.38539386e-01 4.06624883e-01
5.13870776e-01 -4.18634266e-01 -2.67472357e-01 -2.19627738e-01
6.42240107e-01 4.68358338e-01 5.34549296e-01 -9.86527920e-01
7.63259381e-02 1.51887774e-01 -1.27888650e-01 -8.68658781e-01
4.57415819e-01 -6.03579104e-01 -3.65888439e-02 1.14703789e-01
-8.10009420e-01 3.10034126e-01 -1.25810150e-02 7.29482695e-02
-6.10872805e-01 -8.88877749e-01 2.80601829e-01 1.44234553e-01
-3.27699006e-01 -1.73677236e-01 -4.43027854e-01 6.32743001e-01
5.77000737e-01 3.87198210e-01 -6.81800723e-01 -8.36134791e-01
-5.52497149e-01 2.18592986e-01 3.37745428e-01 4.35564071e-01
5.38122475e-01 -1.29794168e+00 -1.35229361e+00 -4.71986622e-01
6.92155123e-01 -3.90536457e-01 1.76570937e-01 9.42938328e-01
-4.17932510e-01 5.72562516e-01 1.38996327e-02 -5.18412232e-01
-1.38004780e+00 4.39869910e-01 2.00852081e-01 -5.18032253e-01
2.67186295e-02 2.99464077e-01 -2.86542028e-01 -8.64860177e-01
2.04328895e-01 -2.75498748e-01 -6.85097933e-01 7.38077700e-01
7.26224661e-01 6.57110393e-01 1.69367135e-01 -4.97167706e-01
-5.66192791e-02 4.33220744e-01 5.55015981e-01 -8.91754568e-01
9.39751863e-01 -2.94468731e-01 -2.29095533e-01 7.71561563e-01
1.21767485e+00 7.37892687e-02 -5.23145854e-01 -9.27066803e-02
3.63676280e-01 -1.16411094e-02 -3.11231345e-01 -1.09793389e+00
-2.30018035e-01 1.12418520e+00 4.48836654e-01 8.29110324e-01
8.00286293e-01 2.24686369e-01 4.02917594e-01 8.34343791e-01
7.20283948e-03 -1.00797617e+00 3.46219212e-01 7.88893342e-01
1.58320236e+00 -1.36144006e+00 -3.78780276e-01 -1.80227924e-02
-8.41845095e-01 1.38965821e+00 2.09595218e-01 4.79717463e-01
1.12285860e-01 -1.30523294e-01 2.49474496e-01 -3.71356994e-01
-9.80658174e-01 -5.32689095e-01 9.07809496e-01 6.87229812e-01
7.76996851e-01 -3.77076305e-03 -6.30175948e-01 5.22135258e-01
-3.05691630e-01 3.05776056e-02 5.11842370e-01 3.67466033e-01
-3.38272601e-01 -1.34733760e+00 -4.69735712e-01 2.31999591e-01
-2.50439972e-01 -1.56769365e-01 -1.01493514e+00 7.90030539e-01
-2.72424161e-01 1.46729910e+00 -3.25550407e-01 -2.25742996e-01
6.35549426e-01 1.88812822e-01 4.25764769e-01 -5.05913198e-01
-9.01705980e-01 -2.49776661e-01 9.73044336e-01 -4.75506872e-01
-7.14999557e-01 -4.63567734e-01 -9.48754847e-01 -2.35642105e-01
8.63106251e-02 5.30382872e-01 9.67585087e-01 1.04146886e+00
2.75597870e-01 4.62509066e-01 3.74177337e-01 -9.21237051e-01
-4.48683143e-01 -1.38961554e+00 -2.80831218e-01 6.74560428e-01
1.89695328e-01 -2.02940091e-01 -2.48029917e-01 -3.52764018e-02] | [11.761666297912598, 8.431695938110352] |
3dc32b39-380d-4113-97f7-8cd98519297f | swatac-a-sentiment-analyzer-using-one-vs-rest | null | null | https://aclanthology.org/S15-2106 | https://aclanthology.org/S15-2106.pdf | SWATAC: A Sentiment Analyzer using One-Vs-Rest Logistic Regression | null | ['Yousef Alhessi', 'Richard Wicentowski'] | 2015-06-01 | null | null | null | semeval-2015-6 | ['negation-detection'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.193156719207764, 3.611880302429199] |
45b8bb15-b9d4-4a26-b54a-6afd189855d8 | applying-artificial-intelligence-for-age | 2201.03045 | null | https://arxiv.org/abs/2201.03045v1 | https://arxiv.org/pdf/2201.03045v1.pdf | Applying Artificial Intelligence for Age Estimation in Digital Forensic Investigations | The precise age estimation of child sexual abuse and exploitation (CSAE) victims is one of the most significant digital forensic challenges. Investigators often need to determine the age of victims by looking at images and interpreting the sexual development stages and other human characteristics. The main priority - safeguarding children -- is often negatively impacted by a huge forensic backlog, cognitive bias and the immense psychological stress that this work can entail. This paper evaluates existing facial image datasets and proposes a new dataset tailored to the needs of similar digital forensic research contributions. This small, diverse dataset of 0 to 20-year-old individuals contains 245 images and is merged with 82 unique images from the FG-NET dataset, thus achieving a total of 327 images with high image diversity and low age range density. The new dataset is tested on the Deep EXpectation (DEX) algorithm pre-trained on the IMDB-WIKI dataset. The overall results for young adolescents aged 10 to 15 and older adolescents/adults aged 16 to 20 are very encouraging -- achieving MAEs as low as 1.79, but also suggest that the accuracy for children aged 0 to 10 needs further work. In order to determine the efficacy of the prototype, valuable input of four digital forensic experts, including two forensic investigators, has been taken into account to improve age estimation results. Further research is required to extend datasets both concerning image density and the equal distribution of factors such as gender and racial diversity. | ['Harjinder Singh Lallie', 'Thomas Grubl'] | 2022-01-09 | null | null | null | null | ['age-estimation', 'age-estimation'] | ['computer-vision', 'miscellaneous'] | [-6.00205176e-02 2.14985356e-01 -3.21848951e-02 -5.57337821e-01
-2.81030387e-01 -2.22042054e-01 3.23380977e-01 3.36949140e-01
-8.35643113e-01 3.98165226e-01 1.50450151e-02 -1.36364043e-01
-1.54738382e-01 -7.49857306e-01 -2.62215227e-01 -5.51637948e-01
-3.48639876e-01 6.94971621e-01 -8.49420875e-02 2.32698381e-01
4.86571223e-01 7.86413968e-01 -1.73357666e+00 1.06127657e-01
6.53016925e-01 8.37096095e-01 -2.97264487e-01 5.95556378e-01
-1.54409362e-02 6.00231290e-01 -7.17866123e-01 -1.33247459e+00
3.02914083e-01 1.97888762e-01 -5.66392958e-01 -1.49486922e-02
1.18962276e+00 -1.06552541e+00 -1.81727827e-01 9.19247210e-01
1.01832187e+00 -7.14319870e-02 6.08321011e-01 -1.39047992e+00
-5.96267164e-01 6.56009197e-01 -1.04623008e+00 5.54947555e-01
2.62814552e-01 2.29671210e-01 3.00310761e-01 -6.93306863e-01
7.92188406e-01 1.52310824e+00 7.89535105e-01 7.33462334e-01
-7.11079359e-01 -1.37274873e+00 -3.40552449e-01 6.67662203e-01
-1.50394058e+00 -5.62879980e-01 7.36742616e-01 -5.17639160e-01
6.20302081e-01 -2.40501136e-01 5.73574424e-01 1.40281689e+00
-3.34900677e-01 3.41415584e-01 1.13848209e+00 -1.04080245e-01
1.07517794e-01 4.88427356e-02 9.93755013e-02 3.31425786e-01
6.88906252e-01 -1.18629389e-01 -5.73210716e-01 -6.90686107e-02
4.13317740e-01 -2.38731444e-01 3.34310710e-01 -3.52408970e-03
-3.82958889e-01 8.21923316e-01 -1.57047182e-01 9.03750733e-02
-3.51505578e-01 -1.83290675e-01 8.69409680e-01 3.45602371e-02
4.12252784e-01 -1.60490032e-02 1.77423641e-01 -6.10922992e-01
-1.35777819e+00 4.39805090e-01 3.46801639e-01 4.82843846e-01
3.23437333e-01 1.01391256e-01 3.33282322e-01 8.42545033e-01
1.39590427e-01 4.36621785e-01 1.02922004e-02 -1.22186303e+00
7.11663365e-01 3.93985599e-01 -3.69417399e-01 -1.24946439e+00
-2.69110769e-01 2.28642430e-02 -5.84558666e-01 4.72460926e-01
7.86041200e-01 5.91392778e-02 -8.92710507e-01 1.61886680e+00
6.14597380e-01 -4.97277491e-02 -4.12426472e-01 6.34400785e-01
9.05024469e-01 7.05448538e-02 5.41700900e-01 -1.24409974e-01
1.54699576e+00 1.22817837e-01 -4.66838449e-01 -3.48139882e-01
8.44034404e-02 -6.08882904e-01 4.15889770e-01 6.71845853e-01
-1.47634840e+00 -1.97558045e-01 -7.44955778e-01 -1.18201040e-01
-3.49828303e-01 -1.53708621e-03 7.51614034e-01 1.40461767e+00
-8.79528046e-01 6.67613626e-01 -6.49612546e-01 -4.86249298e-01
1.13620961e+00 4.94126111e-01 -9.32746828e-01 -3.04230213e-01
-7.98160970e-01 9.01566684e-01 2.55854487e-01 -2.97686420e-02
-6.08773053e-01 -1.01983643e+00 -8.76551986e-01 -1.56373218e-01
-1.26156658e-02 -1.76086992e-01 7.11547971e-01 -5.07916987e-01
-5.72799444e-01 1.59637141e+00 2.25760013e-01 -4.07493919e-01
7.08693922e-01 -8.67559835e-02 -4.87240970e-01 8.29555750e-01
4.04153585e-01 9.72279429e-01 7.69441903e-01 -9.92838919e-01
-4.42889154e-01 -1.18507159e+00 -2.58037239e-01 -1.62407428e-01
-5.94314337e-01 8.84804368e-01 -2.11124539e-01 -4.75115448e-01
-1.36742294e-01 -6.80240035e-01 2.82668173e-01 3.17683846e-01
6.67279586e-02 2.21411958e-02 7.89268434e-01 -1.34570241e+00
1.16602349e+00 -2.16141391e+00 -3.78144413e-01 2.01677367e-01
3.13776463e-01 5.12994885e-01 7.34158233e-02 1.84870511e-01
-5.14348507e-01 1.32076263e-01 -2.97729555e-03 -6.31842375e-01
-1.20211966e-01 2.04343625e-04 7.99237117e-02 9.36416328e-01
2.50792205e-01 3.62708449e-01 -7.92174637e-01 -9.38445091e-01
1.17469709e-02 6.22385502e-01 -2.84836233e-01 1.12250008e-01
6.36910558e-01 2.95962691e-02 -6.83943406e-02 1.07617462e+00
1.45226741e+00 7.06307888e-01 -9.81856957e-02 -4.76011969e-02
4.90908399e-02 -2.47971654e-01 -1.19609714e+00 1.00215662e+00
-2.44096629e-02 6.07351184e-01 5.30428052e-01 -8.53034139e-01
1.03249133e+00 6.81866780e-02 5.02950847e-01 -8.42637241e-01
3.95816356e-01 1.83226749e-01 1.97120175e-01 -7.57344604e-01
5.45719922e-01 -2.16704324e-01 2.10700423e-01 5.40150583e-01
2.21979186e-01 2.38948584e-01 5.50842822e-01 2.60607809e-01
7.93850720e-01 -2.34889701e-01 -2.30075806e-01 -2.77760439e-02
1.92326069e-01 -5.82493663e-01 6.33773386e-01 4.22774434e-01
-6.01945043e-01 7.69100010e-01 7.79275417e-01 -4.76001561e-01
-1.26835668e+00 -1.23290551e+00 -4.13676351e-01 7.99475968e-01
-5.42173743e-01 -1.66042060e-01 -9.70161617e-01 -4.73536104e-01
1.22803487e-01 7.18584239e-01 -8.90744328e-01 -2.42281497e-01
-4.41755205e-01 -7.51590729e-01 1.02248847e+00 5.16145468e-01
4.52702790e-01 -9.99849498e-01 -7.63592422e-01 -1.66689083e-01
1.15954995e-01 -1.22294080e+00 1.72184256e-04 -4.25893784e-01
-7.36438751e-01 -1.29274118e+00 -9.98834252e-01 -3.40342820e-01
7.49334037e-01 -1.94428161e-01 8.43390226e-01 1.51981980e-01
-7.33971298e-01 6.04582608e-01 -3.90957564e-01 -6.50783539e-01
-3.17337304e-01 -1.88897282e-01 2.41566464e-01 -1.81794286e-01
8.23348582e-01 -9.53802884e-01 -7.75020182e-01 5.41887023e-02
-1.05717039e+00 -4.41448867e-01 3.16712856e-01 3.04734588e-01
-1.58424467e-01 -7.94345960e-02 3.78854066e-01 -6.48761749e-01
3.92915547e-01 -6.13333941e-01 -1.81099147e-01 2.96294279e-02
-5.66030204e-01 -5.77753365e-01 1.03440568e-01 -7.68191397e-01
-1.04900157e+00 -5.16525924e-01 -1.55149475e-01 -6.19948506e-01
-4.53767180e-01 -1.69921696e-01 -2.05557644e-01 6.14836477e-02
5.22231400e-01 -1.86662734e-01 3.44968587e-01 -6.03459299e-01
-1.98512152e-01 8.57782066e-01 9.77913499e-01 -7.74327457e-01
5.23963690e-01 5.74033678e-01 2.55069226e-01 -9.46253359e-01
-2.84096539e-01 -2.59442478e-01 -7.56387115e-01 -5.85339546e-01
7.03992963e-01 -8.05066407e-01 -9.09798145e-01 1.09738386e+00
-9.21342731e-01 1.53404206e-01 1.01427138e-01 5.14134824e-01
-4.51760218e-02 8.36818933e-01 -7.03688502e-01 -1.23843646e+00
-6.43355012e-01 -9.95008349e-01 8.37893605e-01 2.23110169e-01
-2.59423822e-01 -6.95017219e-01 -1.53777122e-01 9.32151556e-01
1.48184463e-01 6.03414297e-01 6.48573518e-01 -7.14097202e-01
2.45384350e-01 -3.90810251e-01 -5.41238666e-01 4.84150916e-01
-3.71214360e-01 3.61607194e-01 -1.01336038e+00 -2.50933319e-01
-2.63788164e-01 -4.80396599e-01 6.62237048e-01 2.56089717e-01
9.78002310e-01 5.22045642e-02 1.58806473e-01 3.81192982e-01
1.09115243e+00 1.30742699e-01 1.05913651e+00 4.40981001e-01
6.27061307e-01 1.19591701e+00 5.70658743e-01 8.69377017e-01
6.07635736e-01 2.61142910e-01 6.24595761e-01 2.81112850e-01
-2.52506703e-01 -3.64252478e-02 3.10198277e-01 1.61968097e-01
-3.56037229e-01 2.43910745e-01 -1.13738716e+00 9.30788457e-01
-1.08869469e+00 -1.42539632e+00 -2.28995994e-01 2.25845051e+00
5.83386362e-01 -1.73391365e-02 6.42176211e-01 3.76900494e-01
1.13352525e+00 -1.06252858e-03 -3.94405514e-01 -9.15022254e-01
-6.33393154e-02 6.66387022e-01 4.07982200e-01 -1.46546215e-01
-7.42591798e-01 5.83499432e-01 6.13656378e+00 1.02443767e+00
-1.11275709e+00 -5.78187928e-02 9.67570961e-01 -4.10316616e-01
1.21892519e-01 -1.17237799e-01 -8.29764307e-01 8.48857582e-01
1.07809055e+00 4.74690758e-02 2.12606132e-01 7.91106343e-01
1.02800928e-01 -5.44665575e-01 -7.14085877e-01 1.13234401e+00
4.92119104e-01 -7.92082608e-01 -3.15081507e-01 3.95250380e-01
8.40214714e-02 -4.09926265e-01 3.41082186e-01 5.47917746e-02
-2.05133587e-01 -1.15344691e+00 8.92791092e-01 3.14528167e-01
1.10485804e+00 -1.18142676e+00 7.26590991e-01 3.67921405e-02
-7.01170087e-01 -3.21911752e-01 -4.48871017e-01 -2.36048684e-01
1.16853975e-01 5.80105305e-01 -7.66765237e-01 9.42083970e-02
1.01261806e+00 2.76453257e-01 -9.45156753e-01 1.00194788e+00
6.71266951e-03 6.04713202e-01 -5.78872561e-01 4.46176976e-01
-9.06974375e-02 -7.92952925e-02 1.00859530e-01 1.28919530e+00
6.04152977e-01 1.28044680e-01 -7.50610709e-01 5.79934299e-01
5.36923110e-02 -9.48426202e-02 -5.23962498e-01 -8.50410089e-02
6.29673481e-01 1.47828853e+00 -7.96070516e-01 -7.90333748e-03
-3.18744570e-01 4.50542837e-01 3.15309614e-01 -2.11037725e-01
-5.82589865e-01 -1.93319708e-01 8.01343083e-01 6.68272018e-01
3.64393711e-01 -4.55554090e-02 -2.84387946e-01 -5.92945695e-01
-7.91630521e-03 -8.87124300e-01 6.45707726e-01 -7.09373832e-01
-1.31374872e+00 -1.48123754e-02 3.57225776e-01 -7.71908164e-01
-3.31125140e-01 -7.19724238e-01 -8.19529116e-01 6.11738443e-01
-8.23039412e-01 -1.43980038e+00 -1.00619838e-01 4.08890992e-01
2.28900611e-02 -2.10262582e-01 3.48787129e-01 9.39907014e-01
-7.56127834e-01 1.11900294e+00 -4.94365722e-01 3.47794622e-01
8.44644189e-01 -9.48108137e-01 1.00304663e-01 8.37056100e-01
-5.23794949e-01 7.29332566e-01 6.84551775e-01 -9.14464474e-01
-1.01921690e+00 -4.97866124e-01 9.14986253e-01 -3.84842485e-01
5.14817655e-01 -2.12716818e-01 -7.88680553e-01 3.77916276e-01
4.84287776e-02 -3.25722307e-01 1.02030540e+00 4.64488417e-02
-5.67371249e-01 -7.63265118e-02 -1.87981987e+00 3.51789147e-01
1.05452943e+00 -3.40039760e-01 -4.95360643e-01 -3.11717182e-01
-3.02147150e-01 -2.23659962e-01 -1.22195888e+00 3.94851536e-01
9.17141616e-01 -1.53456318e+00 1.21862745e+00 -2.88419425e-01
7.19147980e-01 4.00670320e-01 2.87458539e-01 -5.81382632e-01
1.32984787e-01 -7.61432061e-03 -1.47140548e-02 1.99110568e+00
-1.55850247e-01 -2.17436448e-01 1.17433512e+00 1.21982360e+00
3.20841610e-01 -7.28728712e-01 -1.16293979e+00 -5.91868162e-01
3.87591988e-01 -8.46199751e-01 7.29731858e-01 9.89107311e-01
-2.23037273e-01 -3.67388695e-01 -3.03285480e-01 7.95349181e-02
1.09794116e+00 -6.59883738e-01 8.48179698e-01 -1.11202908e+00
2.08464757e-01 -5.06968200e-01 -9.44828033e-01 3.59636724e-01
2.37959757e-01 -4.17334557e-01 -9.22806859e-01 -1.08032966e+00
5.97276866e-01 -1.40063271e-01 2.44847998e-01 3.61520529e-01
3.97567302e-02 5.01358032e-01 3.23447824e-01 -3.07939798e-01
-2.23580062e-01 5.07598929e-02 9.55086410e-01 8.95867124e-02
5.10274529e-01 -1.02165304e-01 -7.78541923e-01 7.80470371e-01
6.31984055e-01 -4.66074705e-01 -2.22577065e-01 -2.03111589e-01
1.74747914e-01 -8.48932713e-02 3.90435070e-01 -9.30968821e-01
1.44548669e-01 -1.21597657e-02 7.95511305e-01 -7.44251013e-01
6.28580391e-01 -4.79841292e-01 2.24600270e-01 2.66135335e-01
2.48322219e-01 2.01087922e-01 1.54580817e-01 -2.24739239e-01
-1.66084785e-02 -8.03602576e-01 7.80766845e-01 -1.40970349e-01
-6.26730740e-01 5.05609810e-01 -1.61057353e-01 1.44065306e-01
1.03680885e+00 -7.20997989e-01 -2.83673346e-01 -4.12305057e-01
-6.42980814e-01 1.36676371e-01 7.57599890e-01 2.67402768e-01
6.58190906e-01 -1.15204155e+00 -8.59550536e-01 1.01104915e-01
-2.36078389e-02 -3.88276368e-01 6.78760171e-01 7.85096109e-01
-7.76243687e-01 -2.35106736e-01 -8.64410341e-01 -1.69442087e-01
-1.83930290e+00 4.56137091e-01 -2.33088732e-01 3.28325778e-02
-2.78455526e-01 8.84079278e-01 -4.11811739e-01 -1.40267640e-01
2.45456412e-01 6.07684791e-01 -4.34368670e-01 7.81191409e-01
8.88659537e-01 1.16236115e+00 -9.38250497e-02 -1.19075131e+00
-4.09961611e-01 5.82813740e-01 -3.48258823e-01 -1.39832646e-01
1.65539885e+00 5.85942119e-02 -2.90728033e-01 -3.75323415e-01
1.20660937e+00 2.38815397e-02 -8.65900099e-01 3.74805242e-01
-1.55228138e-01 -1.09484673e+00 -1.12487018e-01 -3.17537278e-01
-1.38954306e+00 1.35010266e+00 9.04023707e-01 -1.48247421e-01
1.11992753e+00 2.19892357e-02 7.25430489e-01 -2.69636005e-01
2.60160029e-01 -1.44168317e+00 4.01563421e-02 5.88834956e-02
4.84913468e-01 -1.12684989e+00 4.47305620e-01 -6.73672929e-02
-5.13396263e-01 1.18308783e+00 7.97561526e-01 1.08418956e-01
1.91760331e-01 2.61306107e-01 5.16278483e-02 -3.64679694e-01
-1.37492195e-01 -3.22811364e-04 -5.72758764e-02 1.19952083e+00
2.57999271e-01 -1.76401079e-01 -6.39633060e-01 7.00931311e-01
-5.21753430e-01 -5.51480614e-02 5.43004692e-01 7.06661165e-01
5.53268902e-02 -1.35602880e+00 -8.72691035e-01 6.71377540e-01
-1.11295688e+00 2.60030597e-01 -4.89296585e-01 9.14468884e-01
7.15786576e-01 7.07292557e-01 3.98998916e-01 -1.35044619e-01
-4.33033109e-02 -4.74636108e-02 8.30487251e-01 -2.05979049e-01
-6.36280000e-01 -5.12639880e-01 3.24481547e-01 -3.30102593e-01
-2.82176703e-01 -1.36690593e+00 -8.48882020e-01 -9.33022976e-01
2.13058107e-02 -3.59557629e-01 9.52675223e-01 5.03503025e-01
6.98820055e-02 -3.12780499e-01 2.85553604e-01 -9.80097473e-01
-3.12902212e-01 -8.62603307e-01 -9.28547621e-01 6.78685606e-01
-1.66253909e-01 -8.79192770e-01 -4.29061949e-01 -2.26694405e-01] | [13.1021146774292, 1.041684627532959] |
c17d91ae-f1db-4765-89d0-928ed9051caf | defuzz-deep-learning-guided-directed-fuzzing | 2010.12149 | null | https://arxiv.org/abs/2010.12149v1 | https://arxiv.org/pdf/2010.12149v1.pdf | DeFuzz: Deep Learning Guided Directed Fuzzing | Fuzzing is one of the most effective technique to identify potential software vulnerabilities. Most of the fuzzers aim to improve the code coverage, and there is lack of directedness (e.g., fuzz the specified path in a software). In this paper, we proposed a deep learning (DL) guided directed fuzzing for software vulnerability detection, named DeFuzz. DeFuzz includes two main schemes: (1) we employ a pre-trained DL prediction model to identify the potentially vulnerable functions and the locations (i.e., vulnerable addresses). Precisely, we employ Bidirectional-LSTM (BiLSTM) to identify attention words, and the vulnerabilities are associated with these attention words in functions. (2) then we employ directly fuzzing to fuzz the potential vulnerabilities by generating inputs that tend to arrive the predicted locations. To evaluate the effectiveness and practical of the proposed DeFuzz technique, we have conducted experiments on real-world data sets. Experimental results show that our DeFuzz can discover coverage more and faster than AFL. Moreover, DeFuzz exposes 43 more bugs than AFL on real-world applications. | ['Yang Xiang', 'Camtepe Seyit', 'Jun Zhang', 'Sheng Wen', 'Xian Li', 'Shigang Liu', 'Xiaogang Zhu'] | 2020-10-23 | null | null | null | null | ['vulnerability-detection'] | ['miscellaneous'] | [ 7.90996477e-02 -1.27460986e-01 -2.09838405e-01 -2.80318379e-01
-5.21523833e-01 -4.63497788e-01 4.57881093e-02 9.46386904e-02
1.78069025e-01 4.81476307e-01 -6.54697418e-04 -7.39299953e-01
3.77844423e-02 -1.17642307e+00 -9.60000575e-01 -1.98828533e-01
-1.79386050e-01 -1.10884689e-01 5.49804866e-01 -2.97496080e-01
6.23133659e-01 1.40756860e-01 -1.41912019e+00 5.40160120e-01
1.33410311e+00 9.01925147e-01 2.95866311e-01 3.77206564e-01
-3.35907489e-01 7.99822688e-01 -1.13115072e+00 -4.64072615e-01
-3.56455892e-01 -1.55157626e-01 -6.31949127e-01 -4.13456708e-01
-8.69866554e-03 -3.35541040e-01 -7.37150982e-02 1.83921015e+00
-9.42958519e-02 -2.48346791e-01 2.34935790e-01 -1.23316145e+00
-9.95062172e-01 1.22105336e+00 -5.81014633e-01 5.49431562e-01
2.80298978e-01 1.77360594e-01 7.70124555e-01 -6.22062325e-01
1.98399974e-03 1.52946639e+00 5.03245592e-01 4.27179515e-01
-5.69776952e-01 -5.98444819e-01 3.88787538e-01 7.81103373e-02
-1.42192519e+00 -8.69846791e-02 8.60568404e-01 -3.69862795e-01
1.18753600e+00 1.43584326e-01 8.80552232e-02 7.62788892e-01
7.06126392e-01 4.69909340e-01 5.35306931e-01 -4.34211642e-01
1.48865908e-01 -1.70688238e-02 3.57874632e-01 9.10831511e-01
4.37677175e-01 2.23221704e-01 7.42368773e-02 -2.91321248e-01
4.72552717e-01 4.20446545e-01 -4.94437069e-01 5.92698932e-01
-1.21524048e+00 8.31034720e-01 6.71896100e-01 4.67476100e-01
-6.38020262e-02 3.31593275e-01 6.44203126e-01 2.92594522e-01
3.12657267e-01 4.76844400e-01 -5.37718654e-01 -2.57461667e-01
-6.16283476e-01 -3.06482404e-03 4.33784783e-01 8.32369328e-01
8.72750759e-01 4.75380480e-01 -1.60897300e-01 2.90279716e-01
4.59732175e-01 4.43462849e-01 7.01452911e-01 -3.72114211e-01
7.85216212e-01 1.08907378e+00 -1.17184266e-01 -1.31391013e+00
8.61261040e-02 -2.20379785e-01 -6.10697210e-01 4.15822953e-01
-2.12905690e-01 -4.26698387e-01 -9.31916237e-01 1.57899106e+00
-2.53544241e-01 6.26954436e-01 8.91819671e-02 7.63853490e-01
5.82355738e-01 1.01592350e+00 -7.12613687e-02 7.70776719e-02
1.21336234e+00 -6.79753065e-01 -7.06710458e-01 -3.72633129e-01
5.34645975e-01 -5.06196320e-01 1.34663463e+00 6.34182870e-01
-7.73256540e-01 -6.58310890e-01 -1.43361664e+00 4.36814398e-01
-5.60294330e-01 6.43710420e-02 4.91768718e-01 7.31118619e-01
-9.23398077e-01 5.12663543e-01 -9.56372678e-01 3.57730418e-01
3.25051844e-01 1.54379070e-01 2.21363321e-01 7.53545538e-02
-1.63535714e+00 6.18052661e-01 7.97686994e-01 2.09286079e-01
-1.23605275e+00 -4.63118076e-01 -1.14513838e+00 5.22426486e-01
5.79274654e-01 1.21267587e-01 1.17946231e+00 -9.26163733e-01
-9.04127836e-01 1.01771951e-01 -5.05104177e-02 -3.98715377e-01
-1.95066795e-01 -2.34305367e-01 -7.73604691e-01 -7.21415058e-02
-1.38532698e-01 1.62244335e-01 7.34628320e-01 -9.48488474e-01
-7.73988366e-01 -4.09789309e-02 5.33366442e-01 -4.36250240e-01
-6.31154954e-01 1.40123695e-01 -6.96101189e-02 -5.41433632e-01
-2.99494743e-01 -1.92673534e-01 -2.50208098e-02 -6.03025854e-01
-6.25014722e-01 -3.09973538e-01 8.38863730e-01 -8.55759561e-01
1.98933828e+00 -2.24649882e+00 -3.66223343e-02 2.93291479e-01
4.19588447e-01 6.00424528e-01 2.89221406e-02 5.11750095e-02
-1.08872727e-01 6.15607321e-01 -3.91430765e-01 2.78184593e-01
-1.41175659e-02 5.08387163e-02 -6.61641896e-01 -5.27682230e-02
6.35475814e-01 9.31723475e-01 -9.06421304e-01 -3.59167248e-01
7.27776764e-03 1.81018412e-01 -5.30556321e-01 4.83397618e-02
-8.94739926e-01 -2.98601776e-01 -6.34477794e-01 8.41349661e-01
6.95968688e-01 -8.13388601e-02 -2.93430030e-01 -8.92677382e-02
-1.50142565e-01 4.39067662e-01 -8.65699530e-01 1.26587570e+00
-5.25200546e-01 4.80771750e-01 -4.22952056e-01 -6.94168687e-01
1.05343735e+00 1.96489945e-01 -4.08897728e-01 -4.90085125e-01
2.04323441e-01 2.06690133e-01 9.71410945e-02 -6.88122869e-01
8.75881612e-02 1.17568560e-01 -3.01453680e-01 4.19346154e-01
-1.30411431e-01 4.45070744e-01 -6.69596121e-02 -4.19248566e-02
1.21468616e+00 -1.48702860e-01 4.86734062e-02 -2.53065825e-01
8.23476315e-01 -3.12019959e-02 6.05302393e-01 3.05113792e-01
-4.20794934e-02 2.86293533e-02 9.95091200e-01 -3.37336451e-01
-4.75443780e-01 -8.60501647e-01 2.24461049e-01 6.79152548e-01
1.98106557e-01 -3.94263983e-01 -1.10516095e+00 -1.03295803e+00
-2.10701719e-01 9.49271262e-01 -6.27799571e-01 -8.43465626e-01
-5.99226773e-01 -2.39116982e-01 8.08830976e-01 6.88524485e-01
6.20886266e-01 -1.43979752e+00 -7.16436148e-01 1.00726888e-01
-1.57303050e-01 -4.43766505e-01 -7.46861398e-01 4.27482985e-02
-6.23609304e-01 -1.03005540e+00 -1.52872592e-01 -1.04307818e+00
6.32608414e-01 3.16239335e-02 9.96210277e-01 6.58055127e-01
-1.76680312e-02 -4.43267614e-01 -4.54816580e-01 -2.36797750e-01
-3.57732981e-01 -1.69856325e-02 -2.34103546e-01 -2.22903937e-01
6.72778070e-01 -2.61105716e-01 -1.66395798e-01 1.89713761e-01
-1.13640201e+00 -5.80247760e-01 6.35883868e-01 7.16108561e-01
5.87728441e-01 9.07919228e-01 7.02966690e-01 -8.50783706e-01
1.14468479e+00 -6.24369204e-01 -9.75664973e-01 3.89107704e-01
-7.96045244e-01 3.45249385e-01 7.95065641e-01 -6.42499745e-01
-1.16829503e+00 -3.98209125e-01 -3.85908425e-01 -6.55413091e-01
1.33348517e-02 1.27185154e+00 -6.53394103e-01 -1.71993658e-01
6.79857552e-01 9.52365398e-02 -2.35366017e-01 -3.98494363e-01
1.60341904e-01 7.68187404e-01 4.80255872e-01 -6.49509966e-01
6.57088876e-01 -1.61675826e-01 -4.21139002e-01 -2.47896705e-02
-3.57897222e-01 4.90867853e-01 4.11773752e-03 -2.95906980e-02
7.61266470e-01 -2.88697749e-01 -9.28758442e-01 2.68183976e-01
-1.24690318e+00 -1.51652947e-01 2.88725883e-01 1.11591637e-01
-3.42078805e-02 4.67104405e-01 -6.14061356e-01 -8.27709675e-01
-5.73114395e-01 -1.48715329e+00 8.73639286e-01 5.47980487e-01
2.70743426e-02 -8.07754755e-01 -8.31859857e-02 -5.50454199e-01
3.73464406e-01 3.39520514e-01 1.39155197e+00 -5.25784016e-01
-5.80455482e-01 -2.42469683e-01 -1.42429054e-01 4.77964789e-01
3.98624808e-01 4.76975381e-01 -7.27766454e-01 -2.71514598e-02
1.51040912e-01 2.85565928e-02 6.49945378e-01 1.72704473e-01
1.48077273e+00 -6.15491688e-01 -5.20721853e-01 5.10812938e-01
1.45223665e+00 9.41553116e-01 9.06493008e-01 1.24266580e-01
6.53962135e-01 1.55045658e-01 7.80322552e-01 2.14758426e-01
8.93479586e-02 1.18341498e-01 9.35835123e-01 1.63548380e-01
3.12576175e-01 -4.93463337e-01 6.35367692e-01 5.79610288e-01
3.48517478e-01 -2.49070510e-01 -1.10064220e+00 4.55282688e-01
-1.63570082e+00 -8.25735748e-01 -1.58506110e-02 2.16070294e+00
9.04191375e-01 9.16436672e-01 -2.28571892e-01 3.53412688e-01
9.51744676e-01 -1.08036749e-01 -6.15861118e-01 -6.14243090e-01
2.81245321e-01 3.76713812e-01 1.81860551e-01 7.51586139e-01
-7.99381435e-01 9.11786616e-01 5.19668102e+00 8.84457409e-01
-1.40964401e+00 9.92303193e-02 6.99802160e-01 2.24575356e-01
-6.29320264e-01 -1.02743976e-01 -6.50175989e-01 8.92530441e-01
9.23409820e-01 -3.96347433e-01 6.55099273e-01 1.04882383e+00
-1.73015475e-01 2.80250579e-01 -8.99907529e-01 5.15858352e-01
9.95203182e-02 -1.25331724e+00 7.33580962e-02 -3.01319808e-01
3.84247363e-01 -4.29223746e-01 1.88637331e-01 6.46710992e-01
1.77104041e-01 -1.06142235e+00 7.69373238e-01 7.65018284e-01
8.56517673e-01 -1.25130975e+00 9.30051327e-01 3.71974915e-01
-1.23401976e+00 -2.46138200e-01 -5.17272353e-01 -2.08988599e-02
-1.75354183e-01 6.80265069e-01 -5.75900435e-01 3.98120284e-01
7.88915038e-01 5.38261652e-01 -7.32183397e-01 9.32186782e-01
-5.31405568e-01 7.18483984e-01 -3.31734976e-04 -4.50726241e-01
1.84943452e-01 4.16951403e-02 1.07787512e-01 8.70909035e-01
5.82107365e-01 -2.20464468e-01 -2.23641526e-02 1.43960929e+00
9.83016007e-03 -5.74549437e-01 -4.34591025e-01 -4.51306492e-01
7.32197046e-01 9.01193023e-01 -5.52480638e-01 -2.90639430e-01
-4.17269379e-01 6.23086035e-01 1.30075857e-01 4.44387019e-01
-1.23629498e+00 -1.26591957e+00 7.49386191e-01 -6.82600513e-02
3.32106292e-01 3.10376305e-02 -4.55355644e-01 -1.00073516e+00
2.06676543e-01 -9.81447458e-01 4.85453576e-01 -9.56009924e-01
-1.10567915e+00 1.04090858e+00 -4.60917838e-02 -1.03198576e+00
-3.84245701e-02 -5.67030966e-01 -1.18034184e+00 1.27490425e+00
-1.41952157e+00 -7.49835908e-01 -2.79954344e-01 3.80729914e-01
3.98671985e-01 -5.58390208e-02 5.20893097e-01 3.63932371e-01
-8.02298725e-01 8.20262671e-01 -4.78447706e-01 2.02602819e-01
3.93158019e-01 -1.16431248e+00 8.67192447e-01 1.34607840e+00
-2.57149339e-01 1.14922643e+00 3.14614564e-01 -1.14799154e+00
-1.25715232e+00 -1.36035001e+00 7.89447963e-01 -2.78625965e-01
7.02950001e-01 -3.07318956e-01 -1.16540670e+00 9.88906801e-01
1.65099785e-01 -3.84268373e-01 1.44936413e-01 -3.36617619e-01
-3.54153931e-01 -2.25763574e-01 -1.25228333e+00 6.13054812e-01
5.52709103e-01 -4.63652641e-01 -7.65347898e-01 4.18905308e-03
1.54514658e+00 -4.60134685e-01 -4.66740638e-01 3.82456154e-01
-1.41322315e-02 -8.33033204e-01 8.18664312e-01 -4.22604769e-01
6.50108278e-01 -7.00194359e-01 -2.57380996e-02 -1.25547314e+00
-4.29570198e-01 -3.68392676e-01 -4.05814528e-01 1.38232899e+00
5.67465305e-01 -8.79676402e-01 3.64027470e-01 2.28624508e-01
-5.64613581e-01 -8.59698594e-01 -6.15128160e-01 -4.46000189e-01
-8.34610313e-02 -3.30702305e-01 1.36422360e+00 8.21898758e-01
3.63657981e-01 2.91000325e-02 1.49946744e-02 5.71134329e-01
1.43292025e-01 1.22052170e-01 -8.21180269e-02 -1.01305652e+00
-4.04420286e-01 -6.98666215e-01 -3.38717133e-01 -7.87726283e-01
2.95987248e-01 -6.07837915e-01 2.41470337e-01 -1.16911006e+00
-2.16090411e-01 -6.17927201e-02 -5.54040849e-01 9.19292331e-01
-6.44411564e-01 -3.62754434e-01 -5.15599966e-01 -2.15999514e-01
-2.56808996e-01 2.34180048e-01 8.99758339e-01 -6.48589134e-01
-1.15614660e-01 1.40526146e-01 -9.73803937e-01 4.42899853e-01
8.36621821e-01 -3.52862239e-01 -6.62993073e-01 -6.83885872e-01
5.65873504e-01 2.05597162e-01 1.81049034e-01 -9.85484481e-01
4.76204883e-03 -2.63304353e-01 5.49886338e-02 -4.79074150e-01
-3.74104768e-01 -6.85044408e-01 -1.80545092e-01 8.62776816e-01
-2.03931123e-01 5.02821267e-01 4.67859715e-01 3.61433148e-01
-3.08421016e-01 -6.89089537e-01 4.14091349e-01 -1.04755042e-02
-8.39344800e-01 2.56554842e-01 -3.72059792e-01 -2.79046353e-02
9.72149730e-01 1.98504165e-01 -6.07115090e-01 -4.48385924e-02
-2.84934670e-01 2.94172943e-01 2.12106943e-01 4.78385746e-01
1.03528488e+00 -1.37922466e+00 -2.37179026e-01 6.02887928e-01
1.72178715e-01 1.00512005e-01 4.91117314e-02 2.27336287e-01
-7.26344705e-01 3.63113523e-01 -1.51965082e-01 -1.77506879e-01
-9.76373672e-01 1.24645317e+00 6.24049187e-01 4.43579592e-02
-3.87645274e-01 1.01038229e+00 5.71525842e-02 -9.84290764e-02
3.88876915e-01 -1.10392725e+00 -3.94103736e-01 -5.19543827e-01
8.67678285e-01 1.58339053e-01 1.42853111e-01 4.53360192e-02
-5.61381340e-01 2.33938053e-01 -1.46278948e-01 1.70411155e-01
8.67022097e-01 3.99426937e-01 -5.54950714e-01 1.27102211e-01
8.10334027e-01 2.58681197e-02 -9.60299253e-01 -3.95011157e-02
1.58916876e-01 -3.91540259e-01 5.28424904e-02 -1.06733644e+00
-1.24043334e+00 1.28926241e+00 4.57631379e-01 5.26652336e-01
1.41701722e+00 -2.18471840e-01 1.02356017e+00 2.17585608e-01
5.18362463e-01 -2.90678263e-01 -5.05921058e-02 4.36442465e-01
5.39160132e-01 -6.19505942e-01 -6.92477167e-01 -2.36809105e-01
-3.79410058e-01 1.19153416e+00 1.09991634e+00 -1.95309684e-01
6.01954877e-01 6.28024995e-01 -1.75684780e-01 -3.05641055e-01
-6.28366530e-01 1.22391850e-01 -1.47000672e-02 5.82394958e-01
2.28181735e-01 2.40745042e-02 -1.07164927e-01 1.21254408e+00
1.44800227e-02 -2.91445153e-03 5.68400145e-01 9.36556935e-01
-7.50383794e-01 -8.73012066e-01 -4.44875717e-01 5.81664443e-01
-4.22087580e-01 -3.29075843e-01 -3.26966673e-01 3.15502405e-01
3.93884569e-01 1.17456949e+00 -8.21914822e-02 -1.05271018e+00
3.59743148e-01 -2.76107304e-02 1.06224246e-01 -8.08302283e-01
-5.79566121e-01 -1.74439758e-01 -3.14411759e-01 -4.72893417e-01
5.83558321e-01 -1.13190323e-01 -1.59035468e+00 -3.02808046e-01
-5.44328570e-01 3.66334260e-01 1.71815723e-01 1.00978911e+00
3.64880055e-01 1.25079370e+00 6.41141117e-01 -2.29434833e-01
-6.76528335e-01 -9.45165515e-01 -2.21416786e-01 -2.46413425e-02
5.58454514e-01 -7.08754957e-01 -4.97120231e-01 -2.28489633e-03] | [7.084081649780273, 7.781159400939941] |
d22ab5d7-f9ce-4f5e-adf3-5ad947eb9bec | location-free-scene-graph-generation | 2303.10944 | null | https://arxiv.org/abs/2303.10944v1 | https://arxiv.org/pdf/2303.10944v1.pdf | Location-Free Scene Graph Generation | Scene Graph Generation (SGG) is a challenging visual understanding task. It combines the detection of entities and relationships between them in a scene. Both previous works and existing evaluation metrics rely on bounding box labels, even though many downstream scene graph applications do not need location information. The need for localization labels significantly increases the annotation cost and hampers the creation of more and larger scene graph datasets. We suggest breaking the dependency of scene graphs on bounding box labels by proposing location-free scene graph generation (LF-SGG). This new task aims at predicting instances of entities, as well as their relationships, without spatial localization. To objectively evaluate the task, the predicted and ground truth scene graphs need to be compared. We solve this NP-hard problem through an efficient algorithm using branching. Additionally, we design the first LF-SGG method, Pix2SG, using autoregressive sequence modeling. Our proposed method is evaluated on Visual Genome and 4D-OR. Although using significantly fewer labels during training, we achieve 74.12\% of the location-supervised SOTA performance on Visual Genome and even outperform the best method on 4D-OR. | ['Benjamin Busam', 'Nassir Navab', 'Tobias Czempiel', 'Felix Holm', 'Ege Özsoy'] | 2023-03-20 | null | null | null | null | ['scene-graph-generation'] | ['computer-vision'] | [ 2.86444724e-01 1.95640430e-01 1.50414139e-01 -3.50710690e-01
-5.17451704e-01 -8.01189005e-01 5.54838121e-01 5.14044881e-01
-1.42349511e-01 4.93039697e-01 -4.92992699e-02 -4.42157924e-01
1.40834125e-02 -1.03188205e+00 -8.73525381e-01 -4.78016615e-01
-8.65818858e-02 4.80459332e-01 5.92327118e-01 1.63100436e-01
1.91596925e-01 3.80431920e-01 -1.48806918e+00 1.95227623e-01
9.38153625e-01 8.31191599e-01 6.31545246e-01 7.56320536e-01
-3.54977578e-01 9.61635172e-01 -5.66735387e-01 -4.04153138e-01
2.51245081e-01 -2.68684715e-01 -7.66278327e-01 1.30409762e-01
7.70985365e-01 -7.34971417e-03 -1.29812077e-01 9.95575607e-01
3.94965410e-01 6.56282455e-02 6.90443397e-01 -1.57190025e+00
-6.02148116e-01 4.52388853e-01 -6.99752927e-01 1.24909960e-01
4.32156861e-01 -1.44578636e-01 1.08185923e+00 -9.88925219e-01
9.54525411e-01 9.49103773e-01 6.18608832e-01 1.47417903e-01
-1.00956142e+00 -1.94715977e-01 3.00627291e-01 3.06047469e-01
-1.56168234e+00 -2.24547908e-01 7.13097334e-01 -7.50196934e-01
1.08661878e+00 2.82239854e-01 4.04190481e-01 6.59382522e-01
-2.96854615e-01 6.74936771e-01 8.71154010e-01 -5.65389752e-01
2.77997911e-01 -9.51004103e-02 2.20312178e-01 1.03481841e+00
3.34511101e-01 -4.75674659e-01 -5.69468141e-01 9.55736861e-02
6.74713790e-01 -1.21010371e-01 -1.88227132e-01 -6.09851062e-01
-1.28027415e+00 4.69425857e-01 5.70496082e-01 3.52278277e-02
-1.12998508e-01 1.72745511e-01 3.42983842e-01 -2.60811299e-01
5.47349393e-01 3.42780530e-01 -2.44497791e-01 1.38454542e-01
-7.84900486e-01 3.02516427e-02 4.76222068e-01 1.42317677e+00
9.25965905e-01 -1.84150070e-01 -2.86978632e-01 7.23015010e-01
1.73137963e-01 4.43578809e-01 -4.18020524e-02 -5.30135930e-01
6.61827326e-01 1.09761155e+00 -5.83620295e-02 -1.49769497e+00
-5.65874279e-01 -4.38271165e-01 -7.31728673e-01 -3.07310045e-01
5.56403756e-01 -1.23104295e-02 -1.19224977e+00 1.56386518e+00
6.19511068e-01 4.70301718e-01 -1.04594082e-01 8.52701545e-01
1.13736224e+00 7.56514311e-01 1.64280072e-01 3.17993499e-02
1.26528049e+00 -1.14032209e+00 -4.90779161e-01 -4.55387563e-01
1.03904223e+00 -5.36977172e-01 1.16841888e+00 7.10614771e-02
-6.20960236e-01 -5.03960371e-01 -8.23131323e-01 -3.02648187e-01
-5.62380493e-01 4.85621810e-01 8.86118770e-01 5.29207885e-01
-1.09263980e+00 1.27174512e-01 -6.19542480e-01 -5.67739844e-01
3.34554940e-01 -1.16951100e-03 -5.01567245e-01 -1.92784771e-01
-7.17577100e-01 5.60110867e-01 5.38469553e-01 1.68245584e-01
-8.48514378e-01 -5.91618240e-01 -1.26617658e+00 1.49885342e-01
6.06768668e-01 -5.85071087e-01 8.38502944e-01 -3.57686818e-01
-7.25411713e-01 1.10636854e+00 -4.46980387e-01 -2.21156552e-01
2.99456209e-01 -6.03968427e-02 -1.92522794e-01 1.77150723e-02
5.45187235e-01 7.25457728e-01 4.44395781e-01 -1.38498795e+00
-7.28374720e-01 -3.20930928e-01 1.76155150e-01 1.85889512e-01
-9.66140907e-03 -3.20194900e-01 -9.72288609e-01 -4.92471576e-01
2.72030085e-01 -9.11522686e-01 -2.50643790e-01 -1.70555294e-01
-9.74004924e-01 -4.01212811e-01 9.29015219e-01 -7.20581293e-01
1.19040942e+00 -2.11502481e+00 -1.12801343e-01 2.60751516e-01
3.22582662e-01 3.64318043e-02 -1.88473672e-01 6.04578018e-01
-1.56227082e-01 1.52144045e-01 -2.02709943e-01 -3.12315702e-01
-1.56444445e-01 6.74373209e-02 -4.30777907e-01 4.06269044e-01
1.22817457e-01 8.63764405e-01 -1.01924729e+00 -7.05950022e-01
2.56962508e-01 9.77916047e-02 -6.98585272e-01 1.38551116e-01
-4.77156550e-01 2.47269928e-01 -3.45288247e-01 7.99504519e-01
7.52051413e-01 -8.20855439e-01 2.67234445e-01 -2.93653697e-01
-6.95747063e-02 6.57250434e-02 -1.18792093e+00 1.80714941e+00
-1.95778638e-01 7.12907434e-01 -4.86519098e-01 -1.02501583e+00
1.03733337e+00 -1.88194707e-01 2.72219419e-01 -5.44052660e-01
-2.28857562e-01 -6.00656047e-02 -3.08397293e-01 -5.06190121e-01
5.13480306e-01 3.80641550e-01 -3.18579853e-01 -3.66498455e-02
-7.64582604e-02 -7.92406723e-02 4.74907041e-01 6.05454922e-01
1.34787750e+00 4.02596563e-01 3.87210190e-01 -7.41863698e-02
3.67257476e-01 5.52680731e-01 7.00768232e-01 8.30432892e-01
-9.66568105e-03 6.42570615e-01 7.20817745e-01 -3.13224852e-01
-8.80907476e-01 -9.33195829e-01 1.13692366e-01 9.01159465e-01
6.82074308e-01 -8.96038711e-01 -7.73315430e-01 -9.63845849e-01
-8.88519809e-02 7.74105191e-01 -4.00463283e-01 2.52259552e-01
-4.37052220e-01 -6.39856398e-01 5.08686244e-01 5.09815454e-01
4.76021498e-01 -6.71066046e-01 -4.50541288e-01 -9.63547826e-02
-3.37974787e-01 -1.51420593e+00 -2.83677697e-01 -5.34392074e-02
-3.34074408e-01 -1.23546958e+00 -4.32669461e-01 -8.76618326e-01
1.06280947e+00 3.95585835e-01 1.16286469e+00 5.43768704e-02
-5.62234402e-01 3.86754781e-01 -3.94152075e-01 -3.55223298e-01
9.49577708e-03 -1.08111072e-02 -2.80858994e-01 -7.42681250e-02
1.96242005e-01 -3.61496001e-01 -5.36426008e-01 3.86074007e-01
-4.42309678e-01 6.59974694e-01 3.73540580e-01 6.51593804e-01
9.46062922e-01 9.09736678e-02 4.07601386e-01 -1.19095635e+00
2.29375958e-01 -3.64235967e-01 -8.87667954e-01 5.68274677e-01
-3.41601878e-01 -2.05345340e-02 6.02241755e-01 -2.99837766e-03
-1.02962995e+00 4.48287874e-01 1.77230880e-01 -4.21329230e-01
-1.97803110e-01 5.94216883e-01 -2.47565359e-01 1.07957117e-01
6.94149137e-01 1.36858135e-01 -7.48894930e-01 -2.85543442e-01
7.28736460e-01 3.30177039e-01 6.30506694e-01 -2.90560812e-01
8.28281581e-01 5.52910089e-01 1.98398411e-01 -8.82245541e-01
-9.62900221e-01 -8.63782823e-01 -7.20870316e-01 -3.79119754e-01
1.02430832e+00 -9.84786093e-01 -5.49566090e-01 2.30635494e-01
-1.33677197e+00 -3.59861970e-01 -4.47108224e-03 1.31945401e-01
-4.53136712e-01 5.60906947e-01 -2.10451126e-01 -8.26586425e-01
-6.43820465e-02 -8.81082654e-01 1.60610664e+00 -5.00119179e-02
-1.14984810e-01 -7.42700398e-01 5.05077727e-02 3.97481680e-01
-1.42617330e-01 5.64404845e-01 1.00804985e+00 -5.27758956e-01
-1.05873930e+00 -2.08495080e-01 -6.59683824e-01 -2.09729686e-01
1.32598370e-01 -4.97558787e-02 -9.70761538e-01 -6.15923032e-02
-5.87912619e-01 -2.69007143e-02 7.38221765e-01 2.60765433e-01
1.29978311e+00 -1.39460474e-01 -7.98552930e-01 6.52973413e-01
1.57357073e+00 1.44181684e-01 4.07688528e-01 1.30766541e-01
1.21529043e+00 7.81130254e-01 9.63724196e-01 4.50690717e-01
6.38150036e-01 7.16048539e-01 5.95085204e-01 -2.81551331e-01
-4.10027146e-01 -7.20442176e-01 -5.34501076e-02 7.02959120e-01
9.83758494e-02 -8.21765304e-01 -1.33759415e+00 7.78889239e-01
-2.12583423e+00 -6.84553266e-01 -6.67482734e-01 1.96865201e+00
3.36023062e-01 -9.49225575e-02 -1.61477149e-01 4.85941470e-02
7.80416012e-01 1.02350600e-01 -4.52368855e-01 9.96080935e-02
-2.50119299e-01 -2.16478154e-01 6.69168651e-01 4.46545631e-01
-1.31515348e+00 1.31696820e+00 5.71271563e+00 8.55639637e-01
-9.07926381e-01 -5.39375097e-02 7.03328252e-01 1.92204103e-01
-3.08767766e-01 2.73172677e-01 -9.76059675e-01 3.29097033e-01
3.73070747e-01 -5.72328903e-02 1.96778491e-01 1.12560940e+00
2.55702198e-01 -3.89078408e-01 -1.07182992e+00 1.20803142e+00
2.28220791e-01 -1.39598811e+00 5.24490215e-02 -1.36656865e-01
7.06823289e-01 -3.54985632e-02 -4.07416165e-01 1.66851372e-01
3.26578140e-01 -8.55140686e-01 7.55023897e-01 4.40866023e-01
9.65691447e-01 -3.51968765e-01 4.75671411e-01 4.80516553e-01
-1.51558423e+00 2.29443118e-01 -3.52345496e-01 8.31177831e-02
3.93356293e-01 8.59159529e-01 -1.46844661e+00 7.77383745e-01
6.56867981e-01 8.52525353e-01 -8.28105927e-01 1.23655355e+00
-4.66053545e-01 6.39160335e-01 -3.52985293e-01 -1.06328249e-01
-2.72213165e-02 -1.21236376e-01 4.53651488e-01 1.36822402e+00
4.15723473e-01 -1.56900629e-01 4.08750296e-01 8.84425700e-01
-1.23095542e-01 2.13511556e-01 -1.00102949e+00 -3.37747559e-02
4.96271521e-01 1.31003964e+00 -1.22617757e+00 -3.56665969e-01
-2.12098658e-01 1.12072301e+00 6.31399095e-01 3.68126214e-01
-1.14662278e+00 -4.19494659e-01 2.34961018e-01 1.56558231e-01
1.51641190e-01 -4.55104917e-01 -2.98860937e-01 -1.14038324e+00
1.71427131e-01 -4.36709106e-01 5.80214679e-01 -1.17597222e+00
-1.16109741e+00 4.55007225e-01 7.78229162e-02 -1.09764850e+00
-2.33398452e-01 -4.91346896e-01 -4.21264529e-01 5.37101805e-01
-1.37682819e+00 -1.41697824e+00 -8.38190079e-01 5.00647604e-01
4.64183807e-01 2.80733019e-01 6.15494013e-01 4.33689505e-01
-7.23295808e-01 3.58899057e-01 -3.27682704e-01 1.33936763e-01
4.86443847e-01 -1.38670456e+00 6.17638409e-01 1.21084678e+00
3.81800681e-01 3.67384344e-01 4.67538714e-01 -1.02022672e+00
-1.30511475e+00 -1.54789758e+00 1.05814290e+00 -4.93727505e-01
3.45873088e-01 -8.63996565e-01 -6.39647126e-01 7.83075392e-01
-1.32532120e-01 1.58020034e-01 3.22533637e-01 5.02364524e-02
-3.75039041e-01 6.15877286e-02 -9.09844279e-01 7.39786506e-01
1.79481363e+00 -5.79238951e-01 -8.32375661e-02 7.70353794e-01
9.28077698e-01 -4.67103571e-01 -4.67233717e-01 5.96067369e-01
-4.33495939e-02 -8.57765615e-01 9.38583851e-01 -4.20658350e-01
2.58849889e-01 -9.67338145e-01 -2.06383809e-01 -9.88511682e-01
-1.77674070e-01 -4.13744390e-01 5.71348444e-02 1.36343229e+00
5.58191955e-01 -3.29389840e-01 1.02719569e+00 4.05383676e-01
-2.96280682e-01 -4.53674048e-01 -5.76585948e-01 -8.76928687e-01
-7.71217883e-01 -5.79818189e-01 4.41746026e-01 1.22825623e+00
-1.27442166e-01 5.20463347e-01 -2.85697281e-01 7.22365439e-01
6.07377410e-01 3.88510078e-01 1.11492872e+00 -9.14552808e-01
-1.83562115e-01 -6.70460835e-02 -7.68796206e-01 -1.26308334e+00
-4.54030894e-02 -1.09347224e+00 9.85897630e-02 -2.11011696e+00
1.93226814e-01 -6.93948209e-01 2.33248919e-01 8.08057368e-01
-2.23506421e-01 1.29215464e-01 4.96709161e-03 1.48165962e-02
-9.07903671e-01 4.22403842e-01 9.69847322e-01 -1.92614987e-01
-7.40532428e-02 -4.57751304e-01 -2.68593639e-01 8.32901657e-01
6.51814044e-01 -2.41158515e-01 -7.33318329e-01 -4.91563141e-01
3.95434469e-01 -9.14246589e-02 4.15960401e-01 -9.71902728e-01
4.74922776e-01 -1.33659169e-01 1.66339353e-01 -1.04544556e+00
3.73064041e-01 -6.50237918e-01 3.29750389e-01 3.50077301e-02
-1.90316187e-03 7.73676336e-02 1.01120666e-01 8.48668039e-01
-1.51272133e-01 -1.55968010e-01 3.28241110e-01 -7.08782077e-02
-1.33844101e+00 1.45164013e-01 5.40304231e-03 3.93703394e-02
1.43195891e+00 -2.86741138e-01 -6.68796420e-01 -2.45574787e-01
-7.63612390e-01 2.28914991e-01 5.41415393e-01 2.73295373e-01
7.33149946e-01 -1.03035629e+00 -4.49759394e-01 5.33514433e-02
5.45935154e-01 5.17669797e-01 3.53125930e-01 5.94085515e-01
-9.43768203e-01 4.44098711e-01 3.59649301e-01 -8.86023879e-01
-1.43449879e+00 6.80060506e-01 1.17678791e-01 -2.10896790e-01
-5.69472730e-01 1.00898921e+00 7.86758065e-01 -4.97420371e-01
1.64573699e-01 -4.01508719e-01 -2.60112852e-01 -2.04377905e-01
4.51103747e-02 3.43811244e-01 -1.58865601e-02 -5.66993535e-01
-4.41994220e-01 6.04920506e-01 2.26860821e-01 3.47172797e-01
1.16176903e+00 -9.69965160e-02 -1.63378090e-01 2.86275148e-01
8.49200130e-01 1.91260904e-01 -9.37192261e-01 1.23278193e-01
3.80375952e-01 -5.21115124e-01 -1.62601307e-01 -7.66663849e-01
-7.96731770e-01 8.53951871e-01 3.43443960e-01 4.28624570e-01
1.03620231e+00 3.51925194e-01 5.45136034e-01 3.05912524e-01
5.82872748e-01 -6.48743689e-01 6.38087466e-02 3.94181490e-01
7.52811313e-01 -1.15852726e+00 8.46572220e-02 -1.37264907e+00
-5.14922082e-01 6.21859193e-01 9.18628454e-01 6.14497252e-02
3.36979985e-01 9.40891355e-02 -1.85425147e-01 -5.64531326e-01
-5.56803346e-01 -5.44797778e-01 4.55224723e-01 6.93504691e-01
2.26726085e-01 4.77379709e-02 -3.44160348e-02 1.97574526e-01
-2.04776466e-01 -2.84233302e-01 3.37939620e-01 7.14131296e-01
-2.16321021e-01 -8.27732205e-01 -9.07218270e-03 4.21006709e-01
-1.16707936e-01 -2.52303898e-01 -5.85493445e-01 7.38574147e-01
2.99681365e-01 1.01809013e+00 1.33041039e-01 -3.39551002e-01
3.44373345e-01 -2.84028620e-01 1.40101478e-01 -8.86285424e-01
2.73414198e-02 -3.82041372e-02 5.71382284e-01 -5.79035282e-01
-3.40418398e-01 -3.39907885e-01 -1.35585237e+00 -5.95934242e-02
-5.80320239e-01 -7.23588392e-02 6.36404872e-01 6.20352685e-01
7.69589365e-01 5.08504510e-01 3.87016505e-01 -4.67897207e-01
2.17349008e-01 -6.56362891e-01 -3.97085011e-01 5.18006027e-01
-3.59862186e-02 -6.29103899e-01 -9.49037001e-02 2.99840480e-01] | [10.33240795135498, 1.63335120677948] |
60554e1e-82b1-4dcf-acf7-f0eef0b0c514 | adaptation-of-domain-specific-transformer | 2209.10966 | null | https://arxiv.org/abs/2209.10966v2 | https://arxiv.org/pdf/2209.10966v2.pdf | Adaptation of domain-specific transformer models with text oversampling for sentiment analysis of social media posts on Covid-19 vaccines | Covid-19 has spread across the world and several vaccines have been developed to counter its surge. To identify the correct sentiments associated with the vaccines from social media posts, we fine-tune various state-of-the-art pre-trained transformer models on tweets associated with Covid-19 vaccines. Specifically, we use the recently introduced state-of-the-art pre-trained transformer models RoBERTa, XLNet and BERT, and the domain-specific transformer models CT-BERT and BERTweet that are pre-trained on Covid-19 tweets. We further explore the option of text augmentation by oversampling using Language Model based Oversampling Technique (LMOTE) to improve the accuracies of these models, specifically, for small sample datasets where there is an imbalanced class distribution among the positive, negative and neutral sentiment classes. Our results summarize our findings on the suitability of text oversampling for imbalanced small sample datasets that are used to fine-tune state-of-the-art pre-trained transformer models, and the utility of domain-specific transformer models for the classification task. | ['Seba Susan', 'Anubhav Sharma', 'Arjun Choudhry', 'Anmol Bansal'] | 2022-09-22 | null | null | null | null | ['text-augmentation'] | ['natural-language-processing'] | [ 2.87910551e-01 1.03932194e-01 -3.58709842e-01 -6.58257127e-01
-8.71085763e-01 -5.39540231e-01 8.62430334e-01 7.03502893e-01
-2.32148379e-01 8.19974601e-01 4.40621644e-01 -5.20586431e-01
8.66329968e-02 -9.32538211e-01 -7.09243953e-01 -3.16553712e-01
-7.96021298e-02 1.04969561e+00 -1.64382905e-01 -7.69411087e-01
3.88377183e-03 1.45752519e-01 -1.26299143e+00 9.25263941e-01
9.82834339e-01 8.51431310e-01 -4.23129767e-01 5.49502850e-01
-4.27747726e-01 7.68747509e-01 -8.44270349e-01 -5.84752917e-01
6.13753982e-02 -8.57808739e-02 -4.29736584e-01 -3.46460462e-01
3.40530306e-01 -2.77681649e-01 2.59180248e-01 8.99146676e-01
3.90580058e-01 -4.70761091e-01 8.77390921e-01 -1.23452044e+00
-3.71294260e-01 4.91181195e-01 -5.45606852e-01 3.78997803e-01
2.94023424e-01 1.66682884e-01 8.21690738e-01 -5.47608674e-01
7.59878457e-01 1.29295135e+00 1.26760745e+00 3.44068110e-01
-1.10579967e+00 -1.21210837e+00 5.23448251e-02 -2.44797215e-01
-9.12995636e-01 8.42666533e-03 5.43988049e-01 -9.22794759e-01
1.39048052e+00 2.13953659e-01 7.65907049e-01 1.23205090e+00
8.60383987e-01 7.35853374e-01 1.58906531e+00 -4.49990891e-02
3.35751995e-02 4.66661543e-01 2.47279242e-01 3.48085552e-01
2.65844584e-01 -1.00458674e-01 -5.34458399e-01 -6.75785780e-01
2.55973339e-01 -2.93431361e-03 2.50907511e-01 7.33257011e-02
-8.92967880e-01 1.55018544e+00 1.43540189e-01 1.59429818e-01
-7.64406621e-01 -6.61397159e-01 9.02511477e-01 3.76594961e-01
1.47419667e+00 6.18923128e-01 -1.04413843e+00 8.65701213e-02
-9.43809628e-01 5.51200747e-01 6.87711298e-01 7.01452434e-01
6.19928181e-01 -1.48969367e-01 -4.80677277e-01 6.99876189e-01
3.74020562e-02 7.65022159e-01 5.88565528e-01 3.48351983e-04
5.63046217e-01 7.79495358e-01 1.48766443e-01 -7.21271813e-01
-5.31863868e-01 -6.91533923e-01 -5.37128925e-01 -4.07944411e-01
1.32030234e-01 -4.45855975e-01 -1.35864198e+00 1.71310437e+00
4.22635823e-01 -6.17675111e-02 1.37429282e-01 3.97329658e-01
8.62336159e-01 6.24956608e-01 4.83861387e-01 9.52084363e-03
1.65754569e+00 -5.55056989e-01 -9.35504019e-01 -3.88609469e-01
9.18323994e-01 -1.05104029e+00 8.54240358e-01 3.66324991e-01
-7.08522797e-01 -1.26991138e-01 -9.39976633e-01 2.32193336e-01
-8.74741077e-01 -3.55211571e-02 6.09384179e-01 7.15124726e-01
-8.67874086e-01 2.20025405e-01 -5.69591284e-01 -4.29220408e-01
5.13555408e-01 5.94507396e-01 -1.91180587e-01 -1.54423798e-02
-1.66963255e+00 1.29682744e+00 -5.94273908e-03 -2.21841335e-01
-8.50311458e-01 -1.16499901e+00 -8.37552905e-01 -4.38241243e-01
-4.32362437e-01 -5.78162074e-01 1.15092504e+00 -1.17861664e+00
-1.17692578e+00 1.07620895e+00 -1.32694036e-01 -6.04021430e-01
2.31734991e-01 -3.55184637e-02 -5.58442652e-01 -2.32060671e-01
1.58476308e-01 4.07545686e-01 5.15137136e-01 -7.67082334e-01
-5.23005903e-01 -3.36315960e-01 -1.39136195e-01 -4.05970700e-02
-3.43147457e-01 4.68238145e-01 4.75787610e-01 -7.97661006e-01
-6.76278412e-01 -7.97103882e-01 -1.61501020e-01 -6.61616325e-01
-3.39217663e-01 -3.43711257e-01 7.83123970e-01 -9.36068296e-01
8.49482298e-01 -1.86985743e+00 -5.47658920e-01 3.14560205e-01
1.53273523e-01 6.08416617e-01 -4.76433635e-01 6.19585156e-01
-2.73791075e-01 2.85338610e-03 3.20886374e-01 -1.77343518e-01
-1.96484953e-01 2.07348749e-01 -5.21229327e-01 4.99326259e-01
5.25122344e-01 4.91749823e-01 -9.21512961e-01 -1.31174177e-01
1.50239304e-01 7.89334536e-01 -6.93715274e-01 3.33635688e-01
-6.47627532e-01 1.45253390e-01 -5.97429156e-01 3.04286659e-01
9.34615314e-01 -2.65989423e-01 1.86038017e-01 -4.11154568e-01
-1.16580324e-02 8.24330151e-01 -3.53959233e-01 8.63507867e-01
-2.09310159e-01 4.69022304e-01 -5.12069166e-02 -6.79047167e-01
9.22198713e-01 3.26262712e-01 6.10499501e-01 -8.13829482e-01
3.44943941e-01 1.69488251e-01 -6.90906569e-02 -4.13700908e-01
6.74256742e-01 -6.91502571e-01 4.06561680e-02 5.02140880e-01
6.91497847e-02 -2.70987824e-02 5.60162216e-02 1.59173712e-01
9.41326737e-01 -3.06000441e-01 1.20677419e-01 -4.70519423e-01
2.31635019e-01 4.10671294e-01 3.21611196e-01 4.19117182e-01
-1.45084366e-01 2.23490447e-01 4.01516557e-01 -4.42943305e-01
-9.39805925e-01 -7.28897035e-01 -2.71161884e-01 1.42453861e+00
-6.55323684e-01 -3.98199797e-01 -8.07827294e-01 -6.37492478e-01
-1.02802180e-02 6.94015563e-01 -1.14714015e+00 8.06678832e-02
-2.33365357e-01 -1.35002434e+00 6.01130605e-01 2.11459026e-01
2.87638698e-02 -6.72760308e-01 -2.09687755e-01 2.81105429e-01
-5.00282168e-01 -1.03652453e+00 -3.52681458e-01 3.81402493e-01
-5.66485882e-01 -1.14350498e+00 -5.32268584e-01 -5.77297330e-01
5.70655644e-01 -2.33879939e-01 1.25771725e+00 -4.52304721e-01
1.47792935e-01 -1.08248377e-02 -2.77915627e-01 -1.11243832e+00
-5.71137547e-01 3.52989227e-01 -1.05295889e-01 -1.67362750e-01
1.01603734e+00 1.49948671e-01 -5.60308933e-01 5.79432072e-03
-9.31714594e-01 7.48464763e-02 2.07087874e-01 6.63818955e-01
3.72912288e-01 -6.95308819e-02 7.76285052e-01 -1.55822909e+00
8.32742035e-01 -8.03182960e-01 -1.70714781e-01 5.03729135e-02
-4.68995452e-01 -7.23647997e-02 5.56603968e-01 -6.38556540e-01
-8.48680973e-01 -4.79470611e-01 -4.70217586e-01 -1.24354318e-01
3.15654248e-01 7.93024838e-01 4.58927959e-01 5.36844842e-02
7.79601872e-01 -1.69788212e-01 2.57582903e-01 -1.89125910e-01
6.47430271e-02 1.07994175e+00 -1.87726989e-01 -9.62546989e-02
3.47505480e-01 3.86709809e-01 -5.75129807e-01 -5.52098453e-01
-1.23142767e+00 -4.65376914e-01 6.05464876e-02 1.04213037e-01
7.20169902e-01 -1.39646399e+00 -2.16804430e-01 7.60502338e-01
-1.08279932e+00 -4.33075786e-01 -1.01100300e-02 2.60543793e-01
-1.79386735e-02 -3.48629892e-01 -9.52950597e-01 -5.85695267e-01
-1.05552471e+00 -1.07741261e+00 1.23228061e+00 -8.97949785e-02
-7.83374846e-01 -1.23913217e+00 5.92305839e-01 5.94841123e-01
8.15561652e-01 5.07612467e-01 1.25985253e+00 -1.54196668e+00
2.91335553e-01 -4.42555368e-01 -1.77881286e-01 2.22070977e-01
4.67115939e-01 8.71179551e-02 -1.00187361e+00 -7.07606435e-01
5.82128018e-02 -4.19171721e-01 3.49731773e-01 5.80028653e-01
4.56738353e-01 -4.91010964e-01 -5.00864148e-01 8.17109272e-02
1.28408015e+00 -4.93370108e-02 3.76253039e-01 2.45034844e-01
4.20311183e-01 6.84950054e-01 8.03252876e-01 5.26931226e-01
8.01611185e-01 3.00218195e-01 5.75368479e-02 -4.89583343e-01
2.43536741e-01 -2.81881601e-01 2.62508690e-01 9.03338909e-01
5.55671573e-01 -8.07130337e-02 -7.58874178e-01 8.36943209e-01
-1.21816313e+00 -5.59754670e-01 -2.67504193e-02 1.87615693e+00
1.47296083e+00 2.98442125e-01 3.69852692e-01 -1.03681341e-01
5.40441513e-01 1.80834204e-01 -3.97827119e-01 -8.75179827e-01
-9.83123407e-02 7.35934734e-01 7.44687319e-01 4.02228862e-01
-1.16807616e+00 7.27928340e-01 7.53104353e+00 8.78133595e-01
-1.64071584e+00 2.85801679e-01 8.29899788e-01 -1.48469865e-01
-6.19014859e-01 -4.26995516e-01 -9.33889627e-01 5.11420012e-01
1.38784969e+00 -4.94305380e-02 1.19999669e-01 4.91818786e-01
3.63406867e-01 5.11418171e-02 -8.99383962e-01 3.67493451e-01
2.12866426e-01 -1.32148242e+00 1.10082842e-01 -9.20184329e-02
9.61298645e-01 6.97983980e-01 2.63421416e-01 4.81505245e-01
5.64081490e-01 -9.70274985e-01 4.50273156e-01 -5.90495020e-02
8.42171252e-01 -6.09354317e-01 1.20645142e+00 6.16188534e-02
-5.97869337e-01 3.27987522e-01 1.73023939e-02 -1.08084626e-01
-2.26869062e-01 9.61179018e-01 -1.49133849e+00 2.77008742e-01
6.66480243e-01 7.84955740e-01 -5.82698822e-01 4.43954796e-01
3.55977893e-01 7.37239778e-01 -4.57087219e-01 -2.49910429e-01
3.70638669e-01 1.61258712e-01 8.70557502e-02 1.53395200e+00
5.26268631e-02 -1.48188263e-01 -5.37192412e-02 3.47445160e-01
1.31153211e-01 1.73416376e-01 -5.51269710e-01 -2.89830953e-01
3.24458927e-01 9.50492680e-01 -4.91043478e-01 -6.46322846e-01
-3.17242384e-01 1.66049719e-01 2.78662071e-02 2.24728569e-01
-7.27222502e-01 -1.91961244e-01 1.02171206e+00 5.58673084e-01
4.66719449e-01 4.56098408e-01 -4.12433714e-01 -9.20810461e-01
-5.26918471e-01 -1.59921575e+00 6.10638976e-01 -6.64546192e-01
-1.68064153e+00 7.66131282e-01 7.30396947e-04 -1.07100213e+00
-3.48012924e-01 -6.19263530e-01 -4.08022285e-01 1.00176108e+00
-1.74613976e+00 -1.32823336e+00 3.40069607e-02 4.68551576e-01
4.09207910e-01 -1.36421859e-01 9.87533987e-01 3.70194227e-01
-2.32510969e-01 6.92785919e-01 2.08276883e-01 1.88381821e-01
9.72956240e-01 -7.69147873e-01 4.09342349e-01 2.75074959e-01
-8.48755419e-01 7.63085544e-01 1.21956813e+00 -9.82876778e-01
-1.11006045e+00 -1.42714870e+00 1.32782662e+00 -5.80306113e-01
8.64233196e-01 -6.61757410e-01 -6.62953377e-01 8.93010139e-01
4.43243921e-01 -5.49252689e-01 1.17926931e+00 2.71347672e-01
-5.46310544e-01 -5.81094958e-02 -1.75637186e+00 3.01362783e-01
1.86083913e-01 -8.14751744e-01 -6.17954969e-01 6.36838615e-01
5.72870791e-01 -5.10920584e-01 -1.01438296e+00 6.17636621e-01
4.57062870e-01 -7.20517278e-01 8.09751093e-01 -8.91658247e-01
5.91955066e-01 7.67137706e-02 3.37395556e-02 -1.70647323e+00
-7.73258805e-02 -3.25456947e-01 5.16239882e-01 1.11920846e+00
6.13812983e-01 -1.01860332e+00 7.76690841e-01 3.02727073e-01
3.12624663e-01 -9.92553711e-01 -6.60237730e-01 -1.54352978e-01
2.96789944e-01 -2.38706738e-01 8.79048288e-01 1.21834135e+00
1.55152380e-01 3.82128090e-01 -2.67958164e-01 -2.17516050e-01
3.39397937e-01 -6.14202172e-02 6.42875612e-01 -1.01721168e+00
3.81741375e-02 -7.35127479e-02 -2.27580428e-01 -4.50728416e-01
3.21458280e-03 -6.17316127e-01 -1.99220464e-01 -1.32649243e+00
3.72316778e-01 -6.81006610e-01 -2.64230132e-01 5.99320829e-01
-1.30048260e-01 3.20462018e-01 -1.42243654e-01 -1.71825513e-01
-1.89022452e-01 3.21254164e-01 1.09177458e+00 -4.36311603e-01
6.48627207e-02 -6.14848025e-02 -1.04976106e+00 4.42075521e-01
7.89280772e-01 -6.93565130e-01 -5.13145566e-01 -2.39569962e-01
5.62827289e-01 -2.31028035e-01 -1.36571378e-01 -2.24715859e-01
-2.34145746e-01 -1.76142842e-01 3.41280133e-01 -9.01030838e-01
2.24124417e-01 -5.43936729e-01 2.40294248e-01 6.43670619e-01
-5.01256287e-01 4.17711407e-01 6.69410527e-01 1.54491290e-01
-1.00388788e-01 1.82450071e-01 6.37004912e-01 2.59270519e-02
2.72344083e-01 7.02940002e-02 -9.06418443e-01 4.74899232e-01
6.65555537e-01 1.01223968e-01 -7.90596187e-01 -2.36876950e-01
-1.24095529e-01 1.91818461e-01 2.63823032e-01 5.66921294e-01
3.45654190e-01 -1.08953369e+00 -9.44394827e-01 6.86896563e-01
1.69243515e-01 -2.41811067e-01 1.12038620e-01 7.32155621e-01
-5.76828122e-01 7.18717456e-01 -3.86311561e-01 -5.63082516e-01
-1.27381420e+00 4.50400651e-01 4.79074717e-01 -8.90519619e-01
1.09012559e-01 6.47986531e-01 3.85019720e-01 -8.57653022e-01
-7.06792995e-02 -7.14302361e-01 -3.62489700e-01 1.75276771e-01
5.21108508e-01 -2.88292449e-02 4.55851406e-01 -4.65886712e-01
-6.33273780e-01 2.48068482e-01 -5.03105521e-01 2.09719434e-01
1.42624140e+00 3.65394413e-01 -3.78900498e-01 2.78662831e-01
1.27338672e+00 -5.58771901e-02 -6.17224336e-01 -1.22617677e-01
-3.58079374e-01 -1.18635617e-01 1.10778116e-01 -1.21645820e+00
-9.19165254e-01 4.47303861e-01 6.90681756e-01 1.27577797e-01
1.00250602e+00 -3.92618984e-01 1.01402903e+00 -2.98739940e-01
1.50051415e-01 -1.01920199e+00 -1.71814039e-01 8.06740940e-01
4.87507612e-01 -1.17525768e+00 2.46636011e-03 -3.66779357e-01
-6.53153718e-01 3.55913311e-01 3.65533173e-01 -7.84253031e-02
8.51440132e-01 6.13589704e-01 4.65903580e-01 -4.83746380e-01
-1.07290983e+00 2.63620228e-01 2.02337489e-01 7.26384044e-01
8.28515530e-01 3.02879661e-01 -3.68474185e-01 5.36621690e-01
-4.29259568e-01 3.17291290e-01 3.74297172e-01 7.79885948e-01
8.96760821e-02 -1.01136839e+00 -3.18043649e-01 1.16539669e+00
-8.77382696e-01 -4.76943135e-01 -4.07984585e-01 4.73790735e-01
4.60851043e-02 1.17538750e+00 1.97997689e-01 -5.97259820e-01
3.36592615e-01 -1.04529196e-02 3.46844733e-01 -5.74994147e-01
-1.38596165e+00 -6.83137178e-02 6.04765177e-01 -1.13040172e-01
-5.26576579e-01 -6.54950619e-01 -8.54333043e-01 -6.11564398e-01
-4.27902013e-01 2.82570899e-01 6.10173523e-01 1.23703539e+00
4.67289865e-01 4.61920857e-01 5.38336933e-01 -4.95504469e-01
-5.22837341e-01 -1.28597260e+00 -1.19657479e-01 5.74169874e-01
6.60856366e-01 -5.53735733e-01 -2.32866094e-01 -1.78928047e-01] | [8.558506965637207, 9.386434555053711] |
7a099e72-c0c6-4404-8297-1f61be1a30c4 | catching-both-gray-and-black-swans-open-set | 2203.14506 | null | https://arxiv.org/abs/2203.14506v1 | https://arxiv.org/pdf/2203.14506v1.pdf | Catching Both Gray and Black Swans: Open-set Supervised Anomaly Detection | Despite most existing anomaly detection studies assume the availability of normal training samples only, a few labeled anomaly examples are often available in many real-world applications, such as defect samples identified during random quality inspection, lesion images confirmed by radiologists in daily medical screening, etc. These anomaly examples provide valuable knowledge about the application-specific abnormality, enabling significantly improved detection of similar anomalies in some recent models. However, those anomalies seen during training often do not illustrate every possible class of anomaly, rendering these models ineffective in generalizing to unseen anomaly classes. This paper tackles open-set supervised anomaly detection, in which we learn detection models using the anomaly examples with the objective to detect both seen anomalies (`gray swans') and unseen anomalies (`black swans'). We propose a novel approach that learns disentangled representations of abnormalities illustrated by seen anomalies, pseudo anomalies, and latent residual anomalies (i.e., samples that have unusual residuals compared to the normal data in a latent space), with the last two abnormalities designed to detect unseen anomalies. Extensive experiments on nine real-world anomaly detection datasets show superior performance of our model in detecting seen and unseen anomalies under diverse settings. Code and data are available at: https://github.com/choubo/DRA. | ['Chunhua Shen', 'Guansong Pang', 'Choubo Ding'] | 2022-03-28 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Ding_Catching_Both_Gray_and_Black_Swans_Open-Set_Supervised_Anomaly_Detection_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Ding_Catching_Both_Gray_and_Black_Swans_Open-Set_Supervised_Anomaly_Detection_CVPR_2022_paper.pdf | cvpr-2022-1 | ['supervised-anomaly-detection'] | ['computer-vision'] | [ 4.86764401e-01 1.04982808e-01 1.24555439e-01 -4.91240531e-01
-8.03030729e-01 -3.34986061e-01 4.03791845e-01 5.15989125e-01
2.09377110e-01 3.37945133e-01 -5.29971607e-02 -3.32498223e-01
-2.71432668e-01 -5.35068333e-01 -5.92779636e-01 -9.59029496e-01
-4.47577149e-01 3.84891987e-01 3.04237362e-02 1.26461640e-01
1.86152413e-01 3.25150043e-01 -1.76197040e+00 3.77913684e-01
9.89494741e-01 8.47284317e-01 -3.64713579e-01 7.08225787e-01
-9.75808054e-02 6.61779404e-01 -6.89200580e-01 -8.04571956e-02
3.66376549e-01 -6.26144230e-01 -5.01210093e-01 3.67147088e-01
5.44964373e-01 -4.18624818e-01 -2.91990340e-01 1.10242176e+00
2.60515064e-01 -4.59992997e-02 6.47214711e-01 -1.71026611e+00
-6.89253628e-01 2.50257462e-01 -8.91917467e-01 8.19185019e-01
4.13798898e-01 1.50833040e-01 1.11103964e+00 -8.05092692e-01
2.13841707e-01 8.48128140e-01 3.08266282e-01 8.80511343e-01
-1.15907276e+00 -5.58776677e-01 2.17778265e-01 2.71961182e-01
-9.56080496e-01 -3.46512944e-01 8.43631566e-01 -4.41603184e-01
6.18457139e-01 6.79516733e-01 4.77619141e-01 1.39027452e+00
3.03003401e-01 1.07533622e+00 6.35099351e-01 -1.95045277e-01
1.77186057e-01 -1.40178367e-01 4.63754654e-01 7.16898978e-01
7.08820701e-01 1.19135782e-01 -3.41945052e-01 -6.79922938e-01
4.58468497e-01 7.77767658e-01 -4.01648015e-01 -5.52192628e-01
-1.17348063e+00 6.91650748e-01 1.37649506e-01 1.81862503e-01
-4.79242533e-01 -2.63454258e-01 4.14558440e-01 5.94287753e-01
5.34685314e-01 3.49995822e-01 -4.71371889e-01 8.93700346e-02
-4.02741283e-01 2.23505739e-02 4.74603266e-01 6.11919463e-01
5.40811241e-01 2.71932423e-01 -3.84802297e-02 8.76709759e-01
3.71902764e-01 4.24859315e-01 8.86504769e-01 -4.79061663e-01
3.76700044e-01 9.73564208e-01 -1.60701349e-01 -9.36065793e-01
-2.39855394e-01 -4.35601056e-01 -9.50495601e-01 3.06072850e-02
5.57690680e-01 1.67177003e-02 -1.14756441e+00 1.41113484e+00
2.05390528e-01 7.03947663e-01 2.77374506e-01 6.71420872e-01
6.57159984e-01 3.48684043e-01 -2.82464713e-01 -1.73539910e-02
1.28614438e+00 -6.43038273e-01 -7.35653400e-01 -3.62011850e-01
9.56728995e-01 -5.51275432e-01 9.61640298e-01 5.52837610e-01
-6.62075341e-01 -1.02109775e-01 -8.02907765e-01 3.64525229e-01
-2.94544876e-01 -1.01158954e-01 4.76064056e-01 4.70481217e-01
-6.54526591e-01 3.71464223e-01 -1.24640715e+00 -2.16394082e-01
6.02619529e-01 8.02767873e-02 -6.19919598e-01 -3.71211410e-01
-6.89102590e-01 3.33857328e-01 2.24854380e-01 2.46326968e-01
-9.51609552e-01 -6.14339113e-01 -1.15696347e+00 -2.52641998e-02
7.71056354e-01 -2.46208444e-01 1.06329095e+00 -8.72391582e-01
-4.96630490e-01 1.03528810e+00 -3.18025798e-01 -4.50436890e-01
4.32164282e-01 -4.47289377e-01 -9.65718150e-01 -3.59478779e-03
1.90790385e-01 -1.70147493e-01 9.29607928e-01 -1.09307373e+00
-6.88495994e-01 -5.26813269e-01 -2.48708546e-01 -2.61320710e-01
-1.93039373e-01 -2.77062535e-01 1.30614564e-01 -8.27405751e-01
7.53460109e-01 -8.35562527e-01 -4.64954704e-01 7.00825900e-02
-9.03249443e-01 -2.53238261e-01 1.26687682e+00 -5.66555440e-01
1.05773234e+00 -2.58133864e+00 -2.46668443e-01 4.81228501e-01
6.18283749e-01 2.49162056e-02 -1.18678600e-01 2.38919556e-01
-6.40030444e-01 -1.12652937e-02 -7.30880380e-01 -7.85656422e-02
-2.75077224e-01 6.24400377e-01 -4.21501815e-01 6.47221208e-01
6.07019246e-01 5.30239940e-01 -1.09766376e+00 -1.96308613e-01
5.59092499e-02 -1.46736830e-01 -5.83207250e-01 4.35490400e-01
3.53291184e-02 8.13581347e-01 -5.77431619e-01 1.09970558e+00
4.42278802e-01 -3.03834409e-01 -3.49762738e-01 3.63556147e-01
3.52510601e-01 -1.13936372e-01 -1.23989630e+00 1.22984099e+00
4.26935926e-02 5.63373923e-01 -3.87488693e-01 -1.23995900e+00
8.26504111e-01 4.57184941e-01 5.41590631e-01 -4.55715299e-01
-3.57378542e-01 4.70144153e-01 4.83016521e-01 -8.40730011e-01
4.61069457e-02 1.06796131e-01 1.55162916e-01 5.70832431e-01
-1.27158776e-01 4.54739064e-01 2.28330180e-01 3.11931551e-01
1.52008736e+00 -4.20696527e-01 3.96060139e-01 1.40037015e-01
5.41377902e-01 -2.02134341e-01 8.93468022e-01 8.49993527e-01
-3.00865620e-01 9.31102574e-01 9.38342631e-01 -7.31656373e-01
-8.33675504e-01 -1.40548229e+00 -3.42446566e-01 7.77998626e-01
-1.08982466e-01 -1.87353447e-01 -1.25255913e-01 -1.08751512e+00
3.81061248e-02 6.35659099e-01 -7.39528358e-01 -5.01817584e-01
-5.81495106e-01 -1.00066316e+00 3.89940023e-01 6.09671175e-01
1.47073651e-02 -1.16590798e+00 -6.18454158e-01 -1.54444888e-01
-6.66041300e-02 -7.27964342e-01 -2.49167517e-01 1.28402993e-01
-9.10991967e-01 -1.77134991e+00 -4.01647806e-01 -5.08935452e-01
1.34117126e+00 1.07895881e-01 1.04152608e+00 4.87487048e-01
-1.02276015e+00 6.08464956e-01 -5.35793543e-01 -7.38173723e-01
-4.94156241e-01 -7.13583529e-01 2.73564577e-01 5.30386746e-01
6.02708220e-01 -4.41956550e-01 -5.97533584e-01 3.17477524e-01
-1.24424696e+00 -5.87069154e-01 7.39404321e-01 1.03417969e+00
7.58813858e-01 -6.22016899e-02 4.81452912e-01 -1.24307847e+00
4.02821004e-01 -1.22643697e+00 -2.72580206e-01 -7.04183094e-06
-5.70530355e-01 6.01966083e-02 5.55205703e-01 -4.92733449e-01
-6.27045929e-01 -2.71569639e-01 -2.75485963e-03 -6.75820053e-01
-7.81826258e-01 2.59632230e-01 8.41782312e-04 4.74993408e-01
7.67947435e-01 4.52419102e-01 1.90347657e-01 -5.64157426e-01
-3.43477070e-01 3.38482946e-01 4.65442657e-01 -3.03831100e-01
9.57269907e-01 5.20659506e-01 2.42226422e-02 -7.66065300e-01
-9.21990216e-01 -8.45739126e-01 -4.66088384e-01 2.25187823e-01
4.89603102e-01 -6.49502158e-01 3.41257378e-02 4.47513819e-01
-5.02837241e-01 1.70019031e-01 -7.53180206e-01 4.22470003e-01
-3.02243173e-01 5.42811811e-01 -2.52847463e-01 -8.87059033e-01
-3.70850563e-02 -9.62827325e-01 1.12269342e+00 6.27715886e-02
-3.32191974e-01 -1.04094923e+00 1.62541568e-01 4.65602875e-02
2.03614399e-01 7.30912566e-01 1.21006882e+00 -1.56945801e+00
-4.69660729e-01 -7.25059450e-01 1.72492921e-01 5.22661984e-01
6.11496150e-01 1.24616019e-01 -1.00274634e+00 -3.95382613e-01
8.37599933e-02 -9.39212441e-02 8.05224419e-01 1.86154962e-01
1.65428960e+00 -3.08945805e-01 -3.50534409e-01 5.76704681e-01
9.28547680e-01 2.68090278e-01 3.73605043e-01 5.55742756e-02
7.58197427e-01 5.03214955e-01 4.56817061e-01 4.74031031e-01
-1.48710638e-01 1.21797740e-01 6.89361155e-01 -1.37302727e-01
4.00381535e-01 -1.72253549e-02 2.43984431e-01 5.16795516e-01
1.17792105e-02 -1.63129464e-01 -1.22713017e+00 7.69801736e-01
-1.83883357e+00 -9.60861146e-01 -4.05987561e-01 2.27358866e+00
1.89878955e-01 1.50096878e-01 -7.21568391e-02 5.97838819e-01
7.11301446e-01 1.51501209e-01 -9.48265731e-01 -2.86793441e-01
-4.44838591e-02 -6.79718936e-03 -1.86682537e-01 -8.69235769e-02
-1.11555481e+00 6.99171703e-03 5.60110140e+00 2.24017784e-01
-8.59616280e-01 -1.33442953e-01 6.77929342e-01 -1.65891156e-01
-2.68663853e-01 -2.49440715e-01 -2.01155707e-01 5.04371881e-01
7.76958704e-01 -1.83721647e-01 -2.45874524e-01 9.87183988e-01
-1.11551709e-01 -2.50286218e-02 -1.57892585e+00 8.94698441e-01
1.68723777e-01 -7.45759070e-01 3.01789820e-01 1.23258084e-01
4.85436440e-01 -1.09280691e-01 4.32152420e-01 3.48223776e-01
-6.16609082e-02 -8.50890577e-01 -1.33702815e-01 4.85363901e-01
5.49483716e-01 -4.44327086e-01 1.04934180e+00 4.05788422e-01
-6.87561333e-01 -3.84237975e-01 -2.23788276e-01 1.63470596e-01
-2.37305552e-01 6.76655769e-01 -8.66379738e-01 5.20981431e-01
8.67895484e-01 9.99150991e-01 -7.93336451e-01 1.12180281e+00
-1.06092446e-01 9.35097694e-01 -1.93749413e-01 5.36972463e-01
1.35202408e-01 -1.33097395e-01 1.06711495e+00 7.25300431e-01
5.33859432e-01 9.89004523e-02 5.06424122e-02 7.25872993e-01
2.80821383e-01 8.27759579e-02 -1.02963221e+00 -2.61676451e-03
-4.60599698e-02 1.08425343e+00 -6.59783542e-01 -6.76495731e-02
-6.92272544e-01 9.31557298e-01 -1.89343125e-01 4.02093679e-01
-5.00100493e-01 -2.99191654e-01 8.51909816e-01 1.08971074e-01
-1.23619564e-01 4.03118849e-01 5.95425852e-02 -1.50102949e+00
4.19960797e-01 -9.60425317e-01 9.64173973e-01 -2.76088297e-01
-1.84699702e+00 4.81388509e-01 -8.21151435e-02 -1.85736156e+00
-6.00485802e-01 -7.50969946e-01 -1.18912518e+00 6.36317372e-01
-1.19182825e+00 -6.12600505e-01 -3.84344041e-01 6.91867232e-01
6.52520001e-01 -5.00422418e-01 9.85142052e-01 2.05588773e-01
-8.68795156e-01 6.79825544e-01 1.67475700e-01 5.50339282e-01
6.49870753e-01 -1.57332122e+00 4.32817310e-01 1.19813728e+00
3.75821948e-01 4.52654064e-01 6.68766618e-01 -5.71498871e-01
-9.99558449e-01 -1.11648762e+00 3.55402797e-01 -7.46809959e-01
7.78218687e-01 -1.41948804e-01 -1.57778835e+00 1.06515014e+00
-3.99103671e-01 8.71326506e-01 1.15449667e+00 1.12483829e-01
-3.80966038e-01 2.61719108e-01 -1.22644699e+00 4.56768811e-01
9.40876901e-01 -1.80511042e-01 -6.93794847e-01 4.00320739e-01
3.82066578e-01 -5.11986017e-01 -3.65941077e-01 5.84874988e-01
1.91507399e-01 -1.02072012e+00 7.59239674e-01 -1.20767653e+00
6.27039969e-01 -1.02263048e-01 -7.68575966e-02 -1.32666647e+00
4.41081077e-03 -1.75721571e-01 -7.20873296e-01 8.31834078e-01
3.90385270e-01 -1.08500624e+00 7.18722880e-01 4.70684469e-01
-3.75950605e-01 -1.13200855e+00 -9.89896834e-01 -5.91719627e-01
-3.91741812e-01 -3.96994591e-01 5.29644430e-01 1.26768780e+00
-2.28161886e-01 -3.41595501e-01 -1.73793614e-01 7.83138335e-01
6.44074023e-01 4.48136404e-02 5.30069530e-01 -1.50981975e+00
-2.35289350e-01 -2.23327190e-01 -1.07298458e+00 -3.65885794e-01
-1.05518721e-01 -9.19907093e-01 -1.61854371e-01 -1.14621663e+00
1.90441981e-01 -3.53968054e-01 -7.57483184e-01 6.68286264e-01
-5.27036548e-01 1.15178801e-01 -3.62970889e-01 2.60069877e-01
-4.07704055e-01 3.18043917e-01 8.44455540e-01 -7.43542612e-02
-6.50670901e-02 6.02763891e-01 -7.23059833e-01 1.17188203e+00
7.08006561e-01 -6.63257480e-01 -4.64407265e-01 -2.37272516e-01
3.35892439e-02 1.93920098e-02 6.44602120e-01 -9.21779454e-01
-1.38137102e-01 -3.07301152e-02 4.97357100e-01 -5.01976430e-01
1.46305934e-02 -9.83734012e-01 -3.33326131e-01 6.20339751e-01
-2.84806758e-01 3.48620445e-01 1.33464500e-01 1.15484130e+00
-5.06434381e-01 -2.60102749e-01 4.36699867e-01 -3.12210042e-02
-7.46715188e-01 4.96856689e-01 -2.30792105e-01 1.89945742e-01
1.38698161e+00 -1.94530100e-01 -4.52728540e-01 -3.25431108e-01
-1.08259380e+00 4.21157241e-01 9.23521891e-02 6.60884678e-01
1.01219475e+00 -1.12637007e+00 -9.48953927e-01 8.74719143e-01
6.59325659e-01 4.20341492e-01 4.81124520e-01 9.18147266e-01
-3.40704679e-01 -7.44125992e-02 -1.22112848e-01 -9.85599697e-01
-1.31368279e+00 4.60934550e-01 3.40057343e-01 -2.09118053e-01
-1.02050829e+00 7.21368968e-01 3.65332514e-01 -3.55732858e-01
2.49066934e-01 -2.42438778e-01 -1.25784054e-01 -2.35119134e-01
6.59028471e-01 3.33962828e-01 6.46970347e-02 -3.95678282e-01
-2.60456622e-01 1.70918882e-01 -5.10942519e-01 5.56411922e-01
1.16485798e+00 2.25406677e-01 -1.47992596e-01 9.33242381e-01
9.54850912e-01 4.66984836e-03 -8.20438921e-01 -4.99669611e-01
3.60046268e-01 -6.52621508e-01 -5.41322768e-01 -5.19469857e-01
-1.28388274e+00 9.33100998e-01 9.00509357e-01 5.22522271e-01
1.22792804e+00 2.82529891e-01 5.76506197e-01 6.28020287e-01
7.60455504e-02 -4.41110551e-01 3.54986101e-01 6.78124949e-02
7.99070179e-01 -1.69181502e+00 -2.62709886e-01 -3.15490097e-01
-7.08881676e-01 9.97523606e-01 9.13410068e-01 -2.81371266e-01
6.07327342e-01 -1.13894660e-02 3.40229958e-01 -4.80247915e-01
-8.18209648e-01 7.06225261e-02 4.79989797e-01 5.58678091e-01
3.21837246e-01 1.40544549e-01 1.77718669e-01 4.95367616e-01
2.82967463e-02 -7.42764115e-01 8.57305765e-01 1.13002157e+00
-1.99399367e-01 -7.55879998e-01 -5.14132977e-01 1.35176921e+00
-8.74762118e-01 1.80712640e-01 -1.57271683e-01 7.32479274e-01
6.44785240e-02 6.95979416e-01 5.76908886e-01 1.31361336e-02
5.17691553e-01 4.52605337e-01 -1.39073163e-01 -9.40705836e-01
-5.04928641e-02 -1.45398527e-01 -2.71662235e-01 -7.56093919e-01
2.88222614e-03 -8.68220448e-01 -1.19423413e+00 3.29749852e-01
-2.63572246e-01 2.49687567e-01 1.27519056e-01 9.96737003e-01
2.14549154e-01 7.93762922e-01 6.94254696e-01 -2.55393326e-01
-5.79626620e-01 -8.22023630e-01 -7.65732348e-01 9.86460924e-01
1.17332566e+00 -6.41760111e-01 -7.97959566e-01 1.40176376e-03] | [7.623922824859619, 2.290506362915039] |
fc7b65e5-3b06-4951-a84a-b73e8b36f67d | introduction-to-discourse-relation-parsing | null | null | https://aclanthology.org/W19-2701 | https://aclanthology.org/W19-2701.pdf | Introduction to Discourse Relation Parsing and Treebanking (DISRPT): 7th Workshop on Rhetorical Structure Theory and Related Formalisms | This overview summarizes the main contributions of the accepted papers at the 2019 workshop on Discourse Relation Parsing and Treebanking (DISRPT 2019). Co-located with NAACL 2019 in Minneapolis, the workshop{'}s aim was to bring together researchers working on corpus-based and computational approaches to discourse relations. In addition to an invited talk, eighteen papers outlined below were presented, four of which were submitted as part of a shared task on elementary discourse unit segmentation and connective detection. | ['Erick Galani Maziero', 'Mikel Iruskieta', 'Juliano Antonio', 'Debopam Das', 'Amir Zeldes'] | 2019-06-01 | null | null | null | ws-2019-6 | ['connective-detection'] | ['natural-language-processing'] | [ 3.38544518e-01 1.12340057e+00 -4.58784431e-01 -4.46107835e-01
-9.60645020e-01 -8.15023243e-01 9.43493724e-01 7.25977600e-01
-3.20287526e-01 1.12515199e+00 9.00120735e-01 -7.51449525e-01
8.77027884e-02 -3.90110910e-01 -2.67977655e-01 -5.09688072e-02
-1.33438274e-01 5.43968022e-01 4.42046136e-01 -5.55193007e-01
5.13290763e-01 8.04418325e-02 -1.05967605e+00 1.14758813e+00
6.56369090e-01 3.10362756e-01 4.34789509e-02 1.07620227e+00
-4.43360656e-01 1.41406250e+00 -1.05288875e+00 -3.86606216e-01
-5.03886640e-01 -8.39724243e-01 -1.75091398e+00 2.35544238e-02
1.81289166e-01 1.74198687e-01 -2.70283431e-01 9.14175451e-01
2.53017157e-01 2.91703194e-01 3.15083534e-01 -6.51661754e-01
-3.17921519e-01 1.58728540e+00 -1.33759186e-01 8.61858070e-01
1.22674048e+00 -7.37196624e-01 1.42987812e+00 -7.74255812e-01
1.16619349e+00 1.63424027e+00 5.10046542e-01 4.07670707e-01
-9.38676596e-01 2.36508995e-02 5.47547281e-01 4.46450710e-01
-8.89091969e-01 -7.62449741e-01 4.88056958e-01 -3.92760307e-01
1.77385068e+00 7.47696757e-01 5.67463160e-01 8.44847560e-01
-2.62021840e-01 1.18248105e+00 1.09891558e+00 -1.11114311e+00
3.27580050e-02 1.88610945e-02 6.86033666e-01 1.78464592e-01
-5.95564721e-03 -5.39187789e-01 -5.90405464e-01 -2.46232644e-01
5.07298410e-01 -1.50532115e+00 -1.34134799e-01 2.80300260e-01
-1.37439060e+00 8.29426944e-01 -1.82913348e-01 1.01006818e+00
3.05065773e-02 -4.68490839e-01 8.21426213e-01 3.65954131e-01
5.54920733e-01 4.63678628e-01 -1.09804414e-01 -4.60604429e-01
-2.61869431e-01 7.43626177e-01 1.19189978e+00 9.04856265e-01
-9.45793241e-02 -2.76838690e-01 -3.02116692e-01 8.42194796e-01
3.55806530e-01 -2.64351547e-01 7.52736926e-02 -1.25697672e+00
1.18767023e+00 3.85159582e-01 1.89526111e-01 -8.95407259e-01
-4.78273153e-01 5.72388291e-01 -7.89539665e-02 -4.61289525e-01
6.61668956e-01 -6.52793109e-01 -1.18641183e-01 1.23698604e+00
4.96063977e-01 -7.05009162e-01 4.76391435e-01 5.48906028e-01
1.59384286e+00 7.04919279e-01 4.44076717e-01 -8.53794396e-01
1.60182023e+00 -8.79414976e-01 -1.40321505e+00 -1.99502259e-02
1.04654872e+00 -1.28194654e+00 4.15464222e-01 1.38373608e-02
-1.72197080e+00 -2.09366709e-01 -9.99742091e-01 -3.29757929e-01
-2.31188014e-01 -2.18024611e-01 7.57386923e-01 5.70304334e-01
-8.50142062e-01 2.86725648e-02 -6.09449327e-01 -4.69795465e-01
3.70168537e-02 3.23993117e-02 2.17570923e-02 5.89402378e-01
-1.47342837e+00 1.40217018e+00 8.38771284e-01 7.90759102e-02
2.98578024e-01 -2.88560420e-01 -8.40692401e-01 -6.15225911e-01
3.67634088e-01 -9.24634859e-02 1.76293075e+00 -4.11700457e-01
-1.76918697e+00 1.57513905e+00 -4.43350613e-01 -6.14313483e-01
2.82006949e-01 -5.75667799e-01 -6.19213641e-01 1.75913870e-01
1.39978053e-02 1.11733653e-01 -3.07531238e-01 -9.60363030e-01
-1.04345572e+00 -1.26324058e-01 5.26323199e-01 6.86935723e-01
3.27469110e-01 1.32618022e+00 3.60827148e-02 -6.14227593e-01
2.68430173e-01 -5.13460934e-01 5.12341745e-02 -1.13140452e+00
-6.86577320e-01 -9.96459186e-01 6.97560549e-01 -9.27690446e-01
1.61675692e+00 -1.58016360e+00 4.22247678e-01 -3.98756742e-01
3.80829982e-02 5.77656746e-01 4.63345915e-01 1.02616501e+00
-2.69174010e-01 2.03082770e-01 -8.90117884e-02 -6.43575415e-02
2.16966256e-01 2.83278227e-01 -3.58132005e-01 2.73386776e-01
6.98194429e-02 1.02480996e+00 -9.79746819e-01 -7.68046379e-01
4.36283022e-01 6.59270585e-02 2.13867992e-01 1.18941642e-01
-5.66132545e-01 5.96008182e-01 -5.40483296e-01 2.47540861e-01
2.58573711e-01 1.53965190e-01 9.73243475e-01 2.34179109e-01
-7.24727333e-01 1.40458786e+00 -1.00501966e+00 1.64669049e+00
1.56665817e-02 9.93727267e-01 4.53175873e-01 -1.23449731e+00
7.40955532e-01 1.06875408e+00 1.00383349e-01 -3.13872129e-01
4.73978966e-01 1.85389504e-01 2.07773641e-01 -6.80152595e-01
9.81583476e-01 -1.04776226e-01 -3.88602674e-01 5.47000408e-01
-1.76904157e-01 -1.62616745e-01 6.24769807e-01 3.88879448e-01
8.16941082e-01 1.44058824e-01 9.10005689e-01 -5.80789804e-01
1.04252076e+00 4.19242173e-01 2.80037105e-01 3.75916362e-01
-2.23033682e-01 2.50393063e-01 7.83999622e-01 -4.38189030e-01
-7.61693478e-01 -6.73232734e-01 -4.24631238e-01 1.14512038e+00
-3.46143618e-02 -8.98668170e-01 -8.59581530e-01 -4.89255905e-01
-6.30982041e-01 9.11619127e-01 -5.14454961e-01 8.32264602e-01
-1.69827259e+00 -8.48973811e-01 8.23165596e-01 3.96596044e-01
4.78230953e-01 -1.30758095e+00 -7.56825686e-01 3.96287173e-01
-8.95712912e-01 -1.45696139e+00 2.07522169e-01 1.73647508e-01
-5.71164787e-01 -1.34165502e+00 -4.67177123e-01 -1.29877961e+00
9.87461135e-02 -2.14128003e-01 1.09113085e+00 -8.21153261e-03
1.00475714e-01 1.57709643e-01 -7.17323601e-01 -4.35538232e-01
-7.66593158e-01 2.66521245e-01 -5.02778411e-01 -7.30765820e-01
6.09681249e-01 -8.17696378e-02 4.00651097e-02 -2.63494641e-01
-1.87721848e-01 -1.11159794e-01 -2.72257358e-01 5.55908322e-01
1.83875158e-01 -5.13719976e-01 6.73127592e-01 -1.48280001e+00
1.08709490e+00 -1.02041401e-01 -2.41932958e-01 5.13301551e-01
2.72285137e-02 -5.41589975e-01 -4.49768342e-02 -2.90000513e-02
-1.74814141e+00 -5.11671603e-01 -6.16982162e-01 1.05597830e+00
-2.64647126e-01 5.82880199e-01 -3.57077122e-01 3.18367302e-01
6.26029968e-01 -5.95011771e-01 -4.31049466e-01 -3.16037357e-01
5.29604435e-01 5.51713705e-01 6.08746827e-01 -9.23925638e-01
7.01234192e-02 -2.51344085e-01 -4.48446006e-01 -1.10590065e+00
-1.00878763e+00 -4.60106343e-01 -9.37575519e-01 -2.33949214e-01
1.24132276e+00 -6.69282198e-01 -5.24263322e-01 2.60393441e-01
-1.84155118e+00 -3.85755211e-01 -6.08087480e-01 3.94696862e-01
-3.86978418e-01 5.27850330e-01 -1.13312840e+00 -1.08876431e+00
-1.53946653e-01 -5.23622990e-01 5.06962657e-01 5.16895473e-01
-1.09610403e+00 -1.51577091e+00 2.36163348e-01 5.36857426e-01
-1.95798203e-01 5.89999795e-01 9.33426082e-01 -8.13003361e-01
-4.49153259e-02 4.65316651e-03 -1.92312196e-01 -6.03080261e-03
1.07879639e-01 5.30465320e-02 -8.42737913e-01 1.83175549e-01
5.91786690e-02 -5.51541209e-01 4.29447412e-01 4.66330618e-01
2.48069182e-01 -2.58015960e-01 -4.01989281e-01 -4.04465109e-01
8.07105541e-01 6.43916309e-01 6.66588426e-01 7.75589764e-01
4.00350541e-01 1.07677686e+00 9.35210884e-01 8.49505588e-02
7.28775501e-01 5.94511509e-01 -1.70377612e-01 4.21145171e-01
-4.25631344e-01 2.18219802e-01 1.66804627e-01 1.44211650e+00
-6.04070187e-01 -3.64198446e-01 -1.09471607e+00 6.54984415e-01
-2.05415773e+00 -9.95674431e-01 -9.43302095e-01 1.60229087e+00
1.16728866e+00 2.45330885e-01 3.92073661e-01 3.17118466e-01
9.97882307e-01 7.28912175e-01 2.92272568e-01 -1.04464686e+00
-6.52834415e-01 4.31401312e-01 -2.22200930e-01 1.20627439e+00
-1.30756676e+00 1.08059239e+00 8.04496574e+00 2.76453257e-01
-3.22602540e-01 1.95377424e-01 2.76956528e-01 4.57407862e-01
-3.12663645e-01 2.00541481e-01 -9.08980787e-01 -2.79063702e-01
1.26526332e+00 -6.17673039e-01 -1.67405561e-01 2.95056850e-01
-2.14358911e-01 -4.48197126e-01 -9.15637732e-01 2.90381908e-01
1.91445407e-02 -1.82914197e+00 -5.68159759e-01 -3.07261705e-01
5.15472889e-01 -1.59583524e-01 -5.06832004e-01 2.51450062e-01
4.76378322e-01 -1.06105232e+00 6.92337632e-01 -5.35715930e-02
2.21572921e-01 -5.41114450e-01 7.49024510e-01 1.02779783e-01
-1.26284635e+00 3.17539483e-01 -9.35608447e-02 -8.42148304e-01
7.52140403e-01 1.90466419e-01 -7.59545147e-01 8.68309557e-01
3.26440781e-01 2.84868360e-01 9.64985713e-02 5.94157159e-01
-6.81315422e-01 7.35056758e-01 -2.34499504e-03 -3.94986331e-01
1.09834895e-01 1.10289291e-01 9.40899551e-01 1.73542154e+00
-6.91962600e-01 9.54654336e-01 3.13354671e-01 2.25531459e-01
1.64916396e-01 1.95875689e-01 -8.46070796e-02 -2.50321273e-02
7.27040708e-01 8.82563293e-01 -1.08017313e+00 -5.64384639e-01
-3.02283764e-01 3.94522130e-01 2.37003282e-01 1.17438219e-01
-6.04922056e-01 -3.98146510e-01 9.63368118e-02 -9.17550474e-02
9.72518325e-02 -3.35253835e-01 -6.57719731e-01 -6.98125958e-01
4.99328785e-02 -9.85775232e-01 6.59515202e-01 -1.53129980e-01
-1.02053142e+00 8.83279383e-01 6.34249389e-01 -4.16980445e-01
-4.19298291e-01 -5.03147483e-01 -6.20134532e-01 1.03313732e+00
-8.98806930e-01 -9.66286361e-01 3.00942302e-01 -9.62513760e-02
6.29477203e-01 6.38084412e-02 1.43672669e+00 1.56838462e-01
-3.89209926e-01 2.79367775e-01 -3.57539862e-01 2.94285446e-01
6.09360516e-01 -1.49542201e+00 7.94417262e-01 6.83055580e-01
1.83433592e-01 5.88409066e-01 9.59961295e-01 -5.52495718e-01
-8.25296223e-01 -4.68234032e-01 2.03355527e+00 -5.37924767e-01
1.12157142e+00 -1.25384644e-01 -9.62206364e-01 1.05905902e+00
1.12737024e+00 -4.37998980e-01 7.28121281e-01 4.33824539e-01
1.25772104e-01 7.56346703e-01 -8.37034643e-01 2.62633085e-01
9.62771475e-01 -6.50138378e-01 -1.60507596e+00 7.07265496e-01
9.01056528e-01 -1.44144285e+00 -1.22517467e+00 4.43866879e-01
2.82172918e-01 -6.55999422e-01 7.57151663e-01 -6.68920755e-01
4.16464865e-01 -6.11373670e-02 -3.32982749e-01 -4.98474836e-01
3.98764834e-02 -1.10340333e+00 -1.79750741e-01 1.50736189e+00
5.69065452e-01 -4.62462902e-01 6.51896596e-01 5.49164057e-01
-5.21827281e-01 -4.78456914e-01 -1.19693363e+00 -2.46197864e-01
6.04954839e-01 -2.46434778e-01 -6.76260376e-03 9.97419000e-01
1.13892639e+00 1.05948544e+00 1.91483766e-01 -4.36054826e-01
3.01937222e-01 9.38213095e-02 3.43277276e-01 -1.07048464e+00
-1.43466681e-01 -5.23390055e-01 -5.56678586e-02 -1.17652714e+00
1.51968047e-01 -7.25470006e-01 9.20757577e-02 -1.99796546e+00
-3.47735941e-01 -3.83794069e-01 2.60733128e-01 1.49513289e-01
-1.67918131e-01 -5.97493500e-02 3.42026353e-01 9.24416929e-02
-8.35038424e-01 1.11872889e-01 1.06014645e+00 -3.97248566e-02
-3.97865176e-01 2.05219761e-01 -6.57808483e-01 8.14916432e-01
8.22023153e-01 -8.69611874e-02 -2.64175802e-01 -4.65225905e-01
-2.15508752e-02 4.77892905e-01 -1.20524213e-01 -3.79457891e-01
2.70048469e-01 -2.05588251e-01 -1.34274825e-01 -1.04682446e+00
4.04793262e-01 2.60795176e-01 -2.56269127e-01 1.59553185e-01
-6.41901135e-01 2.64621884e-01 4.30068046e-01 -5.96155375e-02
-6.42262995e-01 -4.86895919e-01 4.17766422e-01 -3.49062055e-01
-6.97417438e-01 -7.51535177e-01 -9.94740427e-01 7.25261033e-01
1.27534544e+00 -3.23775932e-02 -7.70582736e-01 -7.30541795e-02
-1.01465917e+00 2.26110607e-01 -2.32258230e-01 3.81015927e-01
2.34900162e-01 -8.85108054e-01 -9.05538380e-01 -8.24616313e-01
-3.50639164e-01 1.43551558e-01 -3.39638591e-02 7.50034690e-01
-6.74686432e-01 1.18714249e+00 9.20460448e-02 -2.94616312e-01
-1.66638637e+00 5.46641462e-02 -7.19779804e-02 -6.47123158e-01
-9.28968847e-01 1.35449064e+00 -4.60015774e-01 -2.38676861e-01
4.97558594e-01 -1.80444047e-01 -8.07935357e-01 3.32115710e-01
8.33211362e-01 3.53350848e-01 2.10407928e-01 -8.21421742e-01
-5.78145325e-01 7.97215383e-03 -1.33769721e-01 -6.11889064e-01
1.09898663e+00 -4.41913605e-01 -6.74865127e-01 1.25637865e+00
5.41103899e-01 1.56968936e-01 -4.30375963e-01 -2.95539767e-01
9.63650048e-01 1.73358053e-01 -3.78422678e-01 -8.01153123e-01
7.27306828e-02 5.66287458e-01 -2.18537644e-01 7.50837564e-01
5.23822129e-01 3.06180924e-01 7.91658521e-01 3.02767038e-01
8.05325955e-02 -1.57308424e+00 -3.76510769e-01 1.16972148e+00
1.08734298e+00 -7.78599024e-01 3.26648742e-01 -1.10767508e+00
-6.00343168e-01 1.40874016e+00 4.38491940e-01 -9.94472802e-02
6.24540448e-01 4.59243327e-01 1.55923620e-01 -3.12785685e-01
-6.44599915e-01 1.27572477e-01 6.41632378e-02 9.03603971e-01
1.34804285e+00 1.15445457e-01 -1.21667075e+00 5.23938298e-01
-6.25175714e-01 -4.62699771e-01 4.79636520e-01 1.26745188e+00
-6.31051660e-01 -1.74049568e+00 -5.52788079e-01 -1.99983895e-01
-5.88546693e-01 2.03191768e-03 -9.42332447e-01 1.23140657e+00
-8.01750645e-02 1.27227771e+00 7.91870728e-02 -2.83533726e-02
5.92172086e-01 1.47368118e-01 9.17102993e-01 -9.87908900e-01
-1.03914285e+00 -4.60681580e-02 1.68037415e+00 -1.58508822e-01
-1.24557149e+00 -1.30974567e+00 -1.68281317e+00 -5.05653024e-01
-2.05672652e-01 8.22846353e-01 1.17533796e-01 1.23528504e+00
-5.29664755e-01 9.62045431e-01 -4.51753363e-02 -5.64591408e-01
-1.16244800e-01 -1.28107226e+00 -2.57673353e-01 -2.99698979e-01
-3.76254469e-02 -3.63364339e-01 2.05956623e-02 2.12916732e-01] | [10.745706558227539, 9.488287925720215] |
9790fbb8-81b0-490d-a0e2-aac5540986b5 | winodict-probing-language-models-for-in | 2209.12153 | null | https://arxiv.org/abs/2209.12153v1 | https://arxiv.org/pdf/2209.12153v1.pdf | WinoDict: Probing language models for in-context word acquisition | We introduce a new in-context learning paradigm to measure Large Language Models' (LLMs) ability to learn novel words during inference. In particular, we rewrite Winograd-style co-reference resolution problems by replacing the key concept word with a synthetic but plausible word that the model must understand to complete the task. Solving this task requires the model to make use of the dictionary definition of the new word given in the prompt. This benchmark addresses word acquisition, one important aspect of the diachronic degradation known to afflict LLMs. As LLMs are frozen in time at the moment they are trained, they are normally unable to reflect the way language changes over time. We show that the accuracy of LLMs compared to the original Winograd tasks decreases radically in our benchmark, thus identifying a limitation of current models and providing a benchmark to measure future improvements in LLMs ability to do in-context learning. | ['William W. Cohen', 'Fangyu Liu', 'Jeremy R. Cole', 'Julian Martin Eisenschlos'] | 2022-09-25 | null | null | null | null | ['probing-language-models'] | ['natural-language-processing'] | [ 4.34754491e-01 1.46311373e-01 -2.33312264e-01 -1.29945174e-01
-9.63964939e-01 -8.28791916e-01 1.24206007e+00 3.04589868e-01
-8.38602901e-01 9.34985816e-01 4.95439649e-01 -5.19862592e-01
4.08041142e-02 -6.89786375e-01 -8.47380459e-01 -3.17852765e-01
1.26597509e-01 7.35953689e-01 6.90768659e-02 -4.55905646e-01
3.25276911e-01 4.48277444e-01 -1.33906770e+00 2.20879003e-01
6.64195895e-01 1.54096127e-01 6.33919716e-01 7.58433521e-01
-4.02524889e-01 6.75361812e-01 -7.75071740e-01 -3.99991453e-01
-1.82120398e-01 -3.76396328e-01 -1.06961584e+00 -6.18274570e-01
6.05657637e-01 1.07141435e-01 -2.14319646e-01 9.26069617e-01
3.40160072e-01 3.66773963e-01 6.60866499e-01 -6.06122732e-01
-6.84525132e-01 1.19357562e+00 1.62741005e-01 7.64334261e-01
5.90720475e-01 1.18583575e-01 1.15474987e+00 -1.10499275e+00
9.15774405e-01 1.55354941e+00 6.26945734e-01 7.44387865e-01
-1.53964794e+00 -5.42260945e-01 4.76170778e-01 3.74236822e-01
-1.27506924e+00 -8.01259995e-01 4.79599774e-01 -2.64175892e-01
1.52530670e+00 2.72930115e-01 5.27787626e-01 1.46680498e+00
3.32831353e-01 4.00176764e-01 9.80975449e-01 -1.00384748e+00
1.76678553e-01 4.72966060e-02 3.14554304e-01 3.90172124e-01
4.34992105e-01 3.98750454e-01 -7.70392179e-01 -4.27136011e-02
2.53019899e-01 -4.82441366e-01 -1.67498976e-01 8.12210366e-02
-1.34331048e+00 5.33321023e-01 -1.44732356e-01 8.73644829e-01
1.64882392e-02 2.61506110e-01 3.36412936e-01 4.53634739e-01
2.87040621e-01 1.05853832e+00 -8.46748948e-01 -3.76086891e-01
-9.24023867e-01 4.50917125e-01 7.90739834e-01 6.57354891e-01
4.62157130e-01 9.65421125e-02 -1.37128681e-01 7.37638891e-01
-5.97884096e-02 5.69330096e-01 1.00391221e+00 -7.98585534e-01
2.12143883e-01 1.06496044e-01 -3.19414586e-02 -6.49369836e-01
-2.23770067e-01 -6.27630770e-01 -3.73501748e-01 -2.09008932e-01
4.82262671e-01 1.81050241e-01 -6.18177116e-01 2.20160294e+00
-1.72418356e-01 4.66022760e-01 2.27878124e-01 1.09050594e-01
4.58672673e-01 6.32946670e-01 2.39061281e-01 -7.06769168e-01
9.69624579e-01 -6.28668487e-01 -6.55669570e-01 -6.18209660e-01
8.43574047e-01 -5.78085601e-01 1.60269380e+00 5.87912023e-01
-1.09100366e+00 -7.26905227e-01 -1.18807709e+00 -1.42634213e-01
-5.92191160e-01 -7.25137830e-01 4.66889232e-01 3.13429385e-01
-9.04373884e-01 8.12633634e-01 -4.89912480e-01 -3.98848712e-01
-1.43398985e-01 1.91403870e-02 -1.27300471e-01 -9.33613926e-02
-1.74478030e+00 1.56601489e+00 7.25678325e-01 -3.43856812e-01
-9.51148927e-01 -9.89529073e-01 -7.41427958e-01 -1.51457891e-01
3.91081363e-01 -5.47095001e-01 1.54452944e+00 -8.09887886e-01
-1.05158842e+00 9.82871473e-01 -4.97970283e-01 -5.92520535e-01
3.01386982e-01 -3.55813950e-01 -8.11840177e-01 -5.21840036e-01
4.59878482e-02 3.06465447e-01 6.67777956e-01 -1.24024272e+00
-5.08263588e-01 -1.45119861e-01 3.23210955e-02 4.51505631e-02
2.18592081e-02 -1.55081525e-01 -1.19868353e-01 -8.72623384e-01
-2.42502078e-01 -6.76797807e-01 6.62944615e-02 -5.92656910e-01
-4.87126643e-03 -4.39472497e-01 4.92361307e-01 -5.74294329e-01
1.62882566e+00 -1.80268705e+00 2.97764957e-01 1.06610870e-02
4.51437607e-02 5.88089287e-01 -5.25610626e-01 5.87540984e-01
-2.31308967e-01 4.79596049e-01 -1.43648013e-01 -4.61852789e-01
-3.58050466e-02 5.20517707e-01 -8.47198784e-01 3.61174438e-03
8.22695047e-02 1.11280239e+00 -1.19614828e+00 -2.27007985e-01
-1.59348957e-02 1.95795432e-01 -3.50223511e-01 2.37955712e-02
-6.50949955e-01 2.11793646e-01 1.96273133e-01 4.40937169e-02
5.93689606e-02 2.55984925e-02 3.31414521e-01 1.21393986e-01
1.03553489e-01 8.47524643e-01 -9.90591764e-01 1.90018201e+00
-9.45903778e-01 5.99752069e-01 -5.87795377e-01 -9.54404593e-01
6.91642225e-01 1.89341441e-01 -2.26363122e-01 -9.02754486e-01
-2.46191680e-01 1.60854712e-01 2.39183068e-01 -2.77536154e-01
5.76170862e-01 -5.29967606e-01 -7.96078667e-02 5.07553577e-01
1.68254837e-01 -3.18747550e-01 3.55487287e-01 3.19890320e-01
9.46274519e-01 1.29430946e-02 5.39108276e-01 -3.69391590e-01
6.69618189e-01 -1.73134580e-01 4.46991980e-01 1.12144160e+00
3.00102949e-01 8.69624019e-02 1.90075874e-01 -4.39347059e-01
-9.73351121e-01 -1.32472992e+00 -2.14907199e-01 1.30332863e+00
-2.49566615e-01 -7.21308351e-01 -4.85799134e-01 -5.58832884e-01
-1.49774373e-01 1.66438460e+00 -6.17318869e-01 -4.93386328e-01
-8.80593061e-01 -3.95239353e-01 6.98543787e-01 3.89439881e-01
-1.60803735e-01 -1.19548631e+00 -3.53135049e-01 4.97031420e-01
-2.38768771e-01 -8.48579586e-01 -4.81888622e-01 2.55311400e-01
-9.28603709e-01 -9.86891985e-01 -1.30330741e-01 -7.54556477e-01
2.28891104e-01 -2.55432308e-01 1.59845948e+00 1.49401560e-01
-1.75546572e-01 2.56985962e-01 -1.99183315e-01 -4.83241290e-01
-9.33279455e-01 3.79135787e-01 4.05711532e-01 -6.37508512e-01
6.59190655e-01 -6.97149336e-01 1.30190849e-01 -2.76041031e-01
-9.45258021e-01 -6.77455142e-02 3.28983426e-01 9.95264232e-01
5.49803913e-01 -2.67199166e-02 8.56330037e-01 -1.17938292e+00
9.35565412e-01 -3.63540739e-01 -5.96328318e-01 4.76124495e-01
-1.08897364e+00 7.00546503e-01 6.80330038e-01 -7.50629306e-01
-8.51975203e-01 -4.18227941e-01 -1.29748359e-01 7.76202753e-02
4.91036698e-02 7.56139278e-01 -1.32896930e-01 4.50385928e-01
9.26117778e-01 4.65659738e-01 -2.56631702e-01 -6.23210430e-01
7.29319930e-01 2.59663999e-01 7.43594646e-01 -9.83166695e-01
9.14932013e-01 -2.14378417e-01 -2.47906744e-01 -6.80742741e-01
-1.09958088e+00 -9.57196876e-02 -8.49122405e-01 2.61974514e-01
3.32081646e-01 -8.13136816e-01 -3.24734449e-01 1.35028511e-01
-1.37331319e+00 -4.53546137e-01 -6.34644806e-01 1.90244019e-01
-5.02613962e-01 3.28848153e-01 -2.61133254e-01 -6.45068109e-01
-3.72614443e-01 -7.05365539e-01 6.77769601e-01 -8.21821466e-02
-7.90089488e-01 -1.42109621e+00 5.04879534e-01 -1.40102938e-01
4.84226823e-01 -6.27673343e-02 1.59127235e+00 -8.74919713e-01
-8.44047442e-02 9.44879279e-02 3.93330932e-01 5.03461123e-01
1.99127749e-01 -1.77847996e-01 -9.26937521e-01 -3.75710547e-01
1.59504548e-01 -2.77850956e-01 8.90408933e-01 -2.34718978e-01
9.79596734e-01 -5.56972921e-01 -4.21287894e-01 3.01971227e-01
1.32333469e+00 1.65269583e-01 5.14173746e-01 2.71846384e-01
3.54897082e-01 2.68737376e-01 2.91609138e-01 -8.71249959e-02
2.58705020e-01 6.31253481e-01 -2.70136416e-01 5.42621195e-01
-4.19009477e-01 -6.27162039e-01 6.97894156e-01 1.33517909e+00
2.64244616e-01 -2.18328685e-02 -1.23256731e+00 7.71619737e-01
-1.44247258e+00 -1.01398420e+00 2.89809853e-01 2.35494566e+00
1.55265641e+00 6.31607771e-01 -3.83844465e-01 1.12069421e-01
1.97039932e-01 2.73305506e-01 -5.51572680e-01 -5.62303364e-01
-3.37274283e-01 7.65996039e-01 2.70907208e-02 1.20751560e+00
-4.57095295e-01 1.29131186e+00 7.18314791e+00 7.91220844e-01
-9.76939797e-01 2.78674245e-01 1.11387454e-01 -1.96699217e-01
-7.50664711e-01 1.03943482e-01 -8.48660529e-01 4.06924397e-01
1.54813671e+00 -6.73221767e-01 7.01707780e-01 4.95279729e-01
-2.34827455e-02 -1.33256599e-01 -1.74098051e+00 8.49440932e-01
3.78542215e-01 -1.30095828e+00 5.55324972e-01 -3.53437364e-01
6.27584040e-01 5.60636707e-02 -4.09298427e-02 6.08616114e-01
4.03287917e-01 -1.27523351e+00 6.24322474e-01 9.21664059e-01
1.03476501e+00 -6.89745545e-01 3.26249033e-01 6.47177041e-01
-8.08040619e-01 2.37143096e-02 -3.55664998e-01 -3.16631228e-01
5.52890599e-02 4.48254317e-01 -9.31901336e-01 1.30879432e-01
-4.18558419e-02 5.16921818e-01 -9.77569878e-01 3.86328846e-01
-5.21133542e-01 8.41117382e-01 -8.29877108e-02 -3.60162184e-02
-1.08391248e-01 2.45806873e-01 6.69472098e-01 1.32341242e+00
1.71976715e-01 -5.98146021e-02 1.94801297e-02 8.63227367e-01
-1.46864071e-01 -3.42724472e-02 -6.80028081e-01 -2.65563875e-01
8.28430474e-01 6.39886677e-01 -3.80212329e-02 -5.93843222e-01
-1.79444283e-01 8.11481178e-01 5.35126328e-01 3.23223203e-01
-3.75498086e-01 -2.62410611e-01 6.86474860e-01 1.43748060e-01
1.76856946e-02 -5.11806369e-01 -1.31238952e-01 -1.33303952e+00
-1.72752917e-01 -1.08478093e+00 1.78535238e-01 -7.31314778e-01
-1.33165431e+00 3.80143225e-01 2.57718712e-01 -3.74864072e-01
-7.61772215e-01 -5.73497117e-01 -5.24897873e-01 1.13199842e+00
-1.31891489e+00 -8.69159400e-01 2.23967060e-01 3.90520811e-01
9.05631006e-01 -2.43028894e-01 1.25167608e+00 -1.02171332e-01
-2.07002491e-01 7.37589359e-01 9.90622342e-02 -3.51920351e-02
7.99741030e-01 -1.28262675e+00 9.57071424e-01 9.79318559e-01
6.97044909e-01 1.13997006e+00 9.04005706e-01 -7.30679274e-01
-1.28757775e+00 -8.32267702e-01 1.63329363e+00 -1.12244761e+00
8.40703845e-01 -6.69390500e-01 -1.14533079e+00 8.98202360e-01
1.92513764e-02 -2.23871648e-01 5.39442480e-01 4.80436355e-01
-7.31232941e-01 -1.65933654e-01 -6.33449435e-01 7.90920675e-01
1.29311132e+00 -1.10038507e+00 -1.66891706e+00 2.84333080e-01
1.21192634e+00 -1.24127299e-01 -5.59074759e-01 1.82831958e-01
4.25819695e-01 -2.74583489e-01 9.00739610e-01 -1.15507436e+00
2.25362927e-01 -1.54859766e-01 -3.64830136e-01 -1.56549883e+00
-3.93394381e-01 -7.38377213e-01 -4.47333544e-01 1.20585752e+00
5.95924079e-01 -5.32077134e-01 6.05794340e-02 3.99075270e-01
1.02386534e-01 -3.29535842e-01 -1.04051638e+00 -1.03319526e+00
8.36638451e-01 -8.47264409e-01 6.06349230e-01 1.05512631e+00
7.78111964e-02 8.34274590e-01 4.39953059e-02 -2.46897042e-01
4.15187478e-01 -3.16626936e-01 4.18936223e-01 -1.41967738e+00
-4.47838128e-01 -4.93347406e-01 3.46725062e-02 -7.94315755e-01
5.00117719e-01 -1.29664910e+00 -5.75767718e-02 -1.10699272e+00
3.05102207e-02 -3.64146441e-01 -4.96543437e-01 3.40367109e-01
-3.47835302e-01 -1.58727273e-01 2.57818252e-01 9.72812846e-02
-4.70671296e-01 2.68212229e-01 7.83380926e-01 -2.22768903e-01
-7.03604147e-02 -3.33991468e-01 -7.91895628e-01 6.24111056e-01
6.12887919e-01 -6.17488980e-01 -5.57723165e-01 -5.75777471e-01
6.58975124e-01 -2.93183059e-01 1.82222337e-01 -8.36864948e-01
2.64155477e-01 -3.14503580e-01 3.61428380e-01 -1.73932165e-01
2.78135994e-03 -5.53376973e-01 1.61770120e-01 5.74618995e-01
-8.16430569e-01 4.05729741e-01 4.36406434e-01 4.49836373e-01
5.35998382e-02 -4.24442977e-01 6.88787997e-01 -4.66300756e-01
-9.43067253e-01 -1.22316919e-01 -4.29096699e-01 6.50178075e-01
4.55144376e-01 1.14048772e-01 -2.37994090e-01 -8.03210288e-02
-8.96926165e-01 -9.57445130e-02 4.48569894e-01 7.41100609e-01
4.15269583e-01 -1.27873266e+00 -6.89617038e-01 1.59734026e-01
2.98949420e-01 -1.35837346e-01 -2.38920197e-01 2.66400695e-01
-4.06929329e-02 6.68225050e-01 1.98024869e-01 -1.45966932e-01
-8.91894102e-01 7.75862753e-01 4.57784683e-01 -5.01467109e-01
-4.28245842e-01 7.15898573e-01 -5.59269302e-02 -4.27394569e-01
2.87188590e-01 -4.30108517e-01 -2.24375706e-02 2.63766617e-01
8.31016243e-01 5.70626184e-02 2.99406171e-01 -3.16303730e-01
-3.36615473e-01 3.32566440e-01 -4.03215706e-01 -4.40031052e-01
1.10847163e+00 -2.62263060e-01 -1.76592082e-01 1.21842265e+00
1.00621867e+00 3.04896444e-01 -6.46334052e-01 -7.60918379e-01
5.60994089e-01 -1.63613290e-01 -1.70115039e-01 -1.39375889e+00
-8.74663889e-02 7.72182465e-01 5.41896045e-01 -2.54447162e-01
7.02495217e-01 1.48773501e-02 6.53463125e-01 9.47177172e-01
5.15103936e-01 -1.19686151e+00 5.12267165e-02 1.17713940e+00
1.17552507e+00 -9.16742682e-01 -1.11020073e-01 3.55428308e-01
-2.55688071e-01 9.76626456e-01 4.51469392e-01 1.00220546e-01
5.89928269e-01 2.69997567e-01 -2.32579969e-02 5.61318509e-02
-1.21607757e+00 1.20900974e-01 3.16092670e-01 7.17746317e-01
4.75674808e-01 2.62709558e-02 -3.73236924e-01 5.89007437e-01
-6.54821634e-01 -1.17997818e-01 5.50778449e-01 5.93326926e-01
-5.95736086e-01 -1.49281025e+00 -6.08508997e-02 4.36305821e-01
-2.40227833e-01 -6.77155018e-01 -4.23580438e-01 7.50826418e-01
2.87519157e-01 6.79872334e-01 2.01814920e-02 -1.96312875e-01
3.89892131e-01 9.07354116e-01 6.10517383e-01 -1.17259228e+00
-4.25192386e-01 -4.53466743e-01 1.78303882e-01 -5.01740873e-01
-1.24073923e-01 -6.76137209e-01 -1.20776427e+00 -1.92751303e-01
-1.37802318e-01 1.49271026e-01 4.47960585e-01 1.35358381e+00
7.24643618e-02 4.38589603e-01 1.89996764e-01 -3.98818165e-01
-6.67483509e-01 -1.17150068e+00 -2.42408991e-01 4.83633280e-01
5.03650904e-01 -4.76707488e-01 -4.06479031e-01 1.66897014e-01] | [10.713622093200684, 8.639735221862793] |
6985dfa9-2a43-4fd2-8af6-60ce541a3288 | detecting-multiword-expression-type-helps | 2005.05692 | null | https://arxiv.org/abs/2005.05692v1 | https://arxiv.org/pdf/2005.05692v1.pdf | Detecting Multiword Expression Type Helps Lexical Complexity Assessment | Multiword expressions (MWEs) represent lexemes that should be treated as single lexical units due to their idiosyncratic nature. Multiple NLP applications have been shown to benefit from MWE identification, however the research on lexical complexity of MWEs is still an under-explored area. In this work, we re-annotate the Complex Word Identification Shared Task 2018 dataset of Yimam et al. (2017), which provides complexity scores for a range of lexemes, with the types of MWEs. We release the MWE-annotated dataset with this paper, and we believe this dataset represents a valuable resource for the text simplification community. In addition, we investigate which types of expressions are most problematic for native and non-native readers. Finally, we show that a lexical complexity assessment system benefits from the information about MWE types. | ['Sian Gooding', 'Ekaterina Kochmar', 'Matthew Shardlow'] | 2020-05-12 | detecting-multiword-expression-type-helps-1 | https://aclanthology.org/2020.lrec-1.545 | https://aclanthology.org/2020.lrec-1.545.pdf | lrec-2020-5 | ['complex-word-identification'] | ['natural-language-processing'] | [ 1.00568332e-01 1.04329251e-01 -2.44546518e-01 -2.43485078e-01
-6.45527065e-01 -8.02636147e-01 4.46640283e-01 5.64173698e-01
-6.88390732e-01 5.19515693e-01 6.32834554e-01 -3.98318589e-01
-1.34775057e-01 -6.74047768e-01 -3.01604509e-01 -3.00239235e-01
3.44469339e-01 2.14714840e-01 -3.38370591e-01 -3.88413906e-01
2.68032491e-01 2.05970109e-01 -1.65838230e+00 4.44749653e-01
1.37998784e+00 5.78597069e-01 1.64185062e-01 2.59789318e-01
-5.21975636e-01 3.37941319e-01 -6.91131413e-01 -1.05569386e+00
2.25082710e-01 -2.85215139e-01 -7.57903874e-01 -4.15325016e-01
9.18790460e-01 1.18157089e-01 -1.08324336e-02 1.17228854e+00
6.28169298e-01 7.27305412e-02 7.73567319e-01 -7.42488265e-01
-6.23229861e-01 1.23146570e+00 -7.98042715e-02 1.92858949e-01
4.99255478e-01 1.11157075e-01 1.38792324e+00 -9.52508390e-01
9.69811797e-01 1.11233687e+00 8.31241071e-01 4.75380898e-01
-1.14035761e+00 -8.59572828e-01 2.57311523e-01 3.82600039e-01
-1.42644608e+00 -4.85055149e-01 4.62622195e-01 -4.58405048e-01
1.49459171e+00 3.97141218e-01 8.36560130e-01 1.14855397e+00
-3.89521532e-02 8.03329587e-01 1.24526024e+00 -8.28249574e-01
-3.27821374e-01 -1.23109017e-02 5.94721437e-01 4.98955220e-01
6.70309126e-01 -2.33436406e-01 -6.21306002e-01 1.93740055e-01
3.44797857e-02 -5.73287308e-01 -2.06517875e-01 3.27354908e-01
-1.06180215e+00 7.04170048e-01 -1.51811451e-01 4.85879332e-01
-1.40620083e-01 -4.84166265e-01 6.81488693e-01 4.57340479e-01
5.40519655e-01 9.06357288e-01 -6.46337330e-01 -5.42252421e-01
-6.38315678e-01 2.33503446e-01 8.74299169e-01 1.16733813e+00
4.40994263e-01 7.24840388e-02 -2.94876605e-01 1.26371074e+00
-8.79295170e-02 4.94711697e-01 6.01494551e-01 -4.74213421e-01
5.53454816e-01 8.68228972e-01 -2.88456887e-01 -7.03370512e-01
-6.64086163e-01 -2.03433633e-01 -6.00925267e-01 -1.10915236e-01
4.12042111e-01 -2.62262404e-01 -3.30846250e-01 2.00953007e+00
2.56761014e-02 -4.77676898e-01 -7.99784586e-02 2.89087415e-01
1.00084329e+00 5.57603836e-01 5.52554369e-01 -1.84299171e-01
1.63539732e+00 -7.32288182e-01 -1.10213399e+00 -7.47423992e-02
1.15539551e+00 -1.05187368e+00 1.60731602e+00 6.09821379e-01
-1.19005442e+00 -3.58045757e-01 -9.44228649e-01 -4.01520222e-01
-9.44298267e-01 4.68162090e-01 7.45617628e-01 1.09713280e+00
-5.31671584e-01 5.04074097e-01 -5.46561778e-01 -4.27446097e-01
2.65109777e-01 6.96624666e-02 -5.41718662e-01 1.87984362e-01
-1.44202495e+00 1.47183132e+00 5.92069566e-01 -2.78683066e-01
1.13476485e-01 -1.03695941e+00 -9.42817986e-01 -6.36539375e-03
1.88894242e-01 -4.69159812e-01 1.08296919e+00 -8.25886548e-01
-1.42012656e+00 1.15892553e+00 -2.57226944e-01 -1.77324787e-01
9.81834233e-02 -4.71798241e-01 -6.25406861e-01 -3.56279135e-01
6.21117428e-02 3.00812036e-01 3.21634144e-01 -6.79124832e-01
-8.97605836e-01 -1.35748446e-01 2.84082685e-02 3.48787606e-01
-7.44903028e-01 4.37860072e-01 -1.05423495e-01 -1.09340560e+00
-3.71175021e-01 -8.58518422e-01 4.80206341e-01 -5.07571995e-01
-4.43761617e-01 -6.18843496e-01 1.22294500e-01 -9.15398777e-01
2.13017201e+00 -1.85744262e+00 3.22815657e-01 2.33650692e-02
2.09839955e-01 5.39386451e-01 -2.14083672e-01 6.18140757e-01
-2.55458891e-01 7.65874207e-01 -2.74362266e-01 -5.33789873e-01
4.94029015e-01 -1.13942459e-01 -7.53996447e-02 2.05846235e-01
1.89224735e-01 1.21303105e+00 -7.43790686e-01 -3.12054485e-01
2.66225159e-01 6.97105825e-02 -3.43282312e-01 -4.89300825e-02
1.00087943e-02 -7.33013526e-02 2.08722353e-01 6.51350319e-01
6.78590715e-01 3.94206107e-01 1.62500545e-01 -3.52949411e-01
-5.94227016e-01 7.87061810e-01 -1.20004940e+00 1.23630226e+00
-7.08205879e-01 8.12538207e-01 -2.40262836e-01 -4.27252233e-01
3.34233314e-01 1.07139200e-01 5.29217795e-02 -7.14085400e-01
2.90266037e-01 2.45443001e-01 5.37388027e-01 -4.87878650e-01
8.42180431e-01 -1.27478272e-01 -4.43047017e-01 4.84885484e-01
2.18330428e-01 -2.58094877e-01 7.17486858e-01 1.03886768e-01
9.99186873e-01 2.69575510e-02 9.45212126e-01 -5.82198441e-01
3.58023554e-01 -1.35147095e-01 6.38127506e-01 6.70774400e-01
-4.86831106e-02 1.89051777e-01 4.43450391e-01 -6.46025538e-02
-1.06645632e+00 -9.38554943e-01 -6.53102875e-01 1.30890620e+00
-3.57201159e-01 -1.22632623e+00 -8.51210654e-01 -4.32934821e-01
-9.29974988e-02 1.05894935e+00 -5.48444211e-01 -1.98360100e-01
-6.94310188e-01 -6.95566237e-01 9.67330277e-01 3.98042023e-01
-8.22216645e-02 -1.22358179e+00 -2.61337459e-01 1.14956148e-01
-3.60320896e-01 -1.09436190e+00 -5.54982364e-01 1.60855010e-01
-3.18588078e-01 -1.11218131e+00 -2.86261827e-01 -8.96314979e-01
2.61899859e-01 9.81129631e-02 1.31440413e+00 1.28628269e-01
-2.81592399e-01 2.67643571e-01 -8.41462791e-01 -9.67346549e-01
-4.61529970e-01 5.55286765e-01 1.70368373e-01 -4.86943364e-01
7.37650216e-01 -1.73533350e-01 6.52212426e-02 -1.50236428e-01
-9.19027269e-01 3.80375274e-02 4.32563543e-01 7.37550795e-01
5.85168898e-01 -1.20271139e-01 4.39567149e-01 -1.44646657e+00
1.07419026e+00 -5.06960273e-01 -4.05835211e-01 4.76840496e-01
-6.22472405e-01 -2.56819222e-02 8.06204736e-01 -6.84456527e-01
-1.19154525e+00 -2.94183254e-01 -4.29834664e-01 4.56109136e-01
-1.90964460e-01 7.25971758e-01 -5.71112394e-01 -1.12950593e-01
6.00379288e-01 -1.96631417e-01 -5.58819175e-01 -4.52468365e-01
4.92316067e-01 6.83175623e-01 2.74955422e-01 -8.90565753e-01
7.13851452e-01 -1.09746560e-01 -2.48362973e-01 -1.14082003e+00
-9.30563867e-01 -4.38820660e-01 -7.59763539e-01 2.23502442e-02
7.38646269e-01 -8.27415347e-01 -5.19130349e-01 5.94763398e-01
-1.30058634e+00 -2.22238615e-01 -2.02844292e-01 4.30029690e-01
-8.69968757e-02 7.35291898e-01 -5.91293395e-01 -4.89697605e-01
-4.81461465e-01 -9.50244367e-01 8.88013780e-01 1.12663656e-01
-1.09063578e+00 -1.06254983e+00 -2.49435157e-02 1.56537257e-02
3.26291740e-01 1.32814899e-01 1.35455084e+00 -7.50627458e-01
1.81812961e-02 1.78592913e-02 7.27001159e-03 2.92689830e-01
6.01948947e-02 2.41506502e-01 -6.59803391e-01 1.44052088e-01
-1.39108092e-01 -2.57916391e-01 7.58101761e-01 3.39987814e-01
1.02233183e+00 -3.23867649e-01 7.77804554e-02 8.75985146e-01
1.07241046e+00 -1.56334415e-01 5.63945770e-01 6.57176256e-01
7.09371328e-01 8.31851006e-01 6.09989047e-01 4.39417541e-01
5.79671800e-01 6.62375271e-01 -7.14066774e-02 8.00988227e-02
-3.38549495e-01 -1.47091031e-01 5.59249699e-01 1.47536206e+00
1.39688641e-01 -4.54908729e-01 -1.00783503e+00 4.18752909e-01
-1.42535996e+00 -7.00287044e-01 -5.87863564e-01 2.03111196e+00
1.22510564e+00 1.50357308e-02 2.57479772e-02 1.28569528e-01
5.67569911e-01 -3.02359313e-02 -7.24222958e-02 -6.88654244e-01
-8.17594111e-01 7.71734476e-01 4.91118491e-01 5.40114164e-01
-1.00438893e+00 1.60225141e+00 6.00611925e+00 1.30558455e+00
-7.05190718e-01 -7.12382793e-02 -9.51449946e-02 6.04701303e-02
-5.78696728e-01 4.03355472e-02 -1.06448114e+00 4.84038264e-01
9.65587974e-01 -5.03753960e-01 4.31605965e-01 4.00621802e-01
-1.49855986e-01 -1.27964720e-01 -1.05944335e+00 1.03716123e+00
2.45903060e-01 -6.85718775e-01 3.27701777e-01 -2.76350111e-01
6.88898742e-01 -9.22537670e-02 2.69664582e-02 5.79218328e-01
1.65313005e-01 -1.13263702e+00 8.83940041e-01 4.90016043e-01
9.17665839e-01 -7.94778883e-01 7.06460655e-01 1.68417484e-01
-1.32493031e+00 7.95546398e-02 -4.48174477e-01 -2.65020519e-01
2.65192956e-01 6.47611976e-01 -1.85499489e-01 2.86489129e-01
5.42930365e-01 5.95469713e-01 -1.05147815e+00 9.50622916e-01
-6.64362907e-01 8.19505870e-01 -2.84933954e-01 -2.95122057e-01
-2.43391007e-01 -4.13047016e-01 6.52657211e-01 1.73235130e+00
3.72137487e-01 9.25025865e-02 -1.77296475e-01 6.20959759e-01
-1.44336253e-01 9.63921189e-01 -4.39873159e-01 -3.73746306e-01
8.96947980e-01 1.46954238e+00 -6.15588903e-01 -3.39001894e-01
-5.30254841e-01 8.24595988e-01 7.69554019e-01 -1.29520863e-01
-5.24429381e-01 -7.95334876e-01 9.20065343e-01 -2.40994960e-01
-8.91854316e-02 -2.39461362e-01 -7.77863741e-01 -1.37526274e+00
5.68138324e-02 -1.10235238e+00 4.05493110e-01 -5.33834219e-01
-1.75083828e+00 5.16667485e-01 2.97515634e-02 -1.09794009e+00
1.16779946e-01 -1.08820641e+00 -4.62592363e-01 9.63979304e-01
-1.46686316e+00 -1.11604285e+00 -1.80726558e-01 1.73356608e-01
5.18637121e-01 -1.87681317e-01 1.16546977e+00 4.35864955e-01
-8.79124403e-01 1.11651361e+00 -5.76682435e-03 8.25002193e-02
1.00931072e+00 -1.44462156e+00 6.01766646e-01 8.31195533e-01
2.30029494e-01 1.08149028e+00 4.44390923e-01 -8.59992385e-01
-9.34154093e-01 -1.04700541e+00 1.49598312e+00 -8.10675085e-01
9.24301445e-01 -5.89664757e-01 -1.00830305e+00 7.41579175e-01
5.01656890e-01 -7.16852784e-01 1.26307440e+00 5.71625233e-01
-3.67767006e-01 3.52905303e-01 -8.35389256e-01 1.04054475e+00
1.34383237e+00 -6.15003943e-01 -8.75318348e-01 1.91338211e-01
6.79901958e-01 -4.45558548e-01 -1.10259378e+00 2.26613536e-01
5.70771635e-01 -7.09206104e-01 7.26783693e-01 -5.73418319e-01
4.26066369e-01 1.15128875e-01 1.61244288e-01 -1.81057763e+00
-4.61017281e-01 -5.14551222e-01 7.34150857e-02 1.59163129e+00
6.40923619e-01 -4.98416960e-01 8.00489560e-02 6.40660822e-01
-4.75098699e-01 -4.63407665e-01 -8.28515232e-01 -8.57539415e-01
5.88971317e-01 -7.61877656e-01 6.57271504e-01 1.23616362e+00
4.72178906e-01 4.52488840e-01 -1.44773602e-01 -3.92313659e-01
1.09958917e-01 -3.17898184e-01 5.61215162e-01 -1.33930731e+00
-4.33210917e-02 -1.12891018e+00 -1.47584647e-01 -5.35382390e-01
8.60558152e-01 -1.53435707e+00 -4.21550691e-01 -1.26689434e+00
1.18025400e-01 -1.19657844e-01 1.66641563e-01 6.10452652e-01
-6.22822940e-01 1.52749252e-02 4.58724022e-01 -7.83269852e-02
-5.08870304e-01 5.38009644e-01 7.49084055e-01 1.06021836e-01
-9.66697782e-02 -3.82113069e-01 -8.17717075e-01 9.24006104e-01
1.16725659e+00 -4.63691980e-01 -8.17254782e-02 -6.42993391e-01
6.59215629e-01 -7.54882991e-01 -4.05770242e-01 -6.66998148e-01
4.43999544e-02 -1.51944205e-01 1.99976593e-01 -3.24196726e-01
-2.21332479e-02 -5.52915514e-01 -6.51743338e-02 1.07067816e-01
-2.95005113e-01 4.53072727e-01 6.31761074e-01 -3.14866245e-01
-3.26168872e-02 -7.81951189e-01 5.65739393e-01 1.43701788e-02
-5.70859134e-01 1.86195169e-02 -7.87190259e-01 5.02774775e-01
8.44757199e-01 -2.17829913e-01 -3.54175150e-01 -1.58732221e-01
-3.20191562e-01 4.71186526e-02 3.93155754e-01 5.11879802e-01
2.74671108e-01 -1.31394410e+00 -7.59406805e-01 2.00845562e-02
5.81490397e-01 -3.98958594e-01 1.68057203e-01 5.37817299e-01
-4.43238646e-01 3.00401866e-01 -2.63936788e-01 2.40033612e-01
-1.33183646e+00 3.82311612e-01 -5.30399010e-02 -3.98214519e-01
-4.21289116e-01 7.04012454e-01 -1.59349233e-01 -8.28579128e-01
2.30181411e-01 -4.17343765e-01 -5.53026199e-01 6.57026827e-01
6.17798626e-01 7.92452633e-01 2.54997939e-01 -8.78882587e-01
-3.51644196e-02 5.97640097e-01 -1.49243400e-01 1.58651173e-02
1.18490076e+00 -2.71058857e-01 -5.26962638e-01 6.27147436e-01
1.15360665e+00 7.23242044e-01 -2.27507368e-01 -1.82380751e-01
5.11772037e-01 -1.48003682e-01 -2.65819311e-01 -9.33709502e-01
-4.20339733e-01 8.06765974e-01 2.25091591e-01 -8.36228952e-02
1.07323122e+00 -2.06557915e-01 8.91322732e-01 5.00467896e-01
1.06773768e-02 -1.66171014e+00 -5.05789399e-01 1.26991701e+00
8.18622112e-01 -1.00813949e+00 -1.21343797e-02 -7.37229049e-01
-7.18083143e-01 1.23424602e+00 7.23409712e-01 2.44698882e-01
5.08032382e-01 3.83860379e-01 -3.73274013e-02 -9.09220204e-02
-5.91374695e-01 -6.51513755e-01 5.78268647e-01 6.42857134e-01
1.05181563e+00 3.56495202e-01 -1.36773348e+00 1.36833751e+00
-9.82446790e-01 -5.28202474e-01 5.41714668e-01 5.39510667e-01
-2.26473361e-01 -1.64231777e+00 -1.30663157e-01 8.21081102e-01
-4.73592758e-01 -8.69591892e-01 -7.47287750e-01 9.17142689e-01
3.41108531e-01 7.77135909e-01 -2.19890937e-01 -3.85786325e-01
6.61058962e-01 4.01522845e-01 5.86782038e-01 -9.42230046e-01
-1.03097510e+00 -4.13296044e-01 3.82835478e-01 -1.91197142e-01
-5.42899258e-02 -7.21033156e-01 -1.22516513e+00 -4.08668607e-01
-2.34225929e-01 -1.71721593e-01 4.97513980e-01 8.76903355e-01
1.38674691e-01 5.46057522e-01 1.88469177e-03 -5.00018120e-01
-3.36718172e-01 -1.29241800e+00 -5.71742535e-01 8.61194134e-01
-2.69532502e-01 -7.07769334e-01 -3.10580254e-01 -7.30617642e-02] | [10.858498573303223, 10.386322975158691] |
e736fac1-f1aa-47ff-939f-ffb317afa036 | pragmatic-competence-of-pre-trained-language | 2109.12951 | null | https://arxiv.org/abs/2109.12951v1 | https://arxiv.org/pdf/2109.12951v1.pdf | Pragmatic competence of pre-trained language models through the lens of discourse connectives | As pre-trained language models (LMs) continue to dominate NLP, it is increasingly important that we understand the depth of language capabilities in these models. In this paper, we target pre-trained LMs' competence in pragmatics, with a focus on pragmatics relating to discourse connectives. We formulate cloze-style tests using a combination of naturally-occurring data and controlled inputs drawn from psycholinguistics. We focus on testing models' ability to use pragmatic cues to predict discourse connectives, models' ability to understand implicatures relating to connectives, and the extent to which models show humanlike preferences regarding temporal dynamics of connectives. We find that although models predict connectives reasonably well in the context of naturally-occurring data, when we control contexts to isolate high-level pragmatic cues, model sensitivity is much lower. Models also do not show substantial humanlike temporal preferences. Overall, the findings suggest that at present, dominant pre-training paradigms do not result in substantial pragmatic competence in our models. | ['Allyson Ettinger', 'Yan Cong', 'Lalchand Pandia'] | 2021-09-27 | null | https://aclanthology.org/2021.conll-1.29 | https://aclanthology.org/2021.conll-1.29.pdf | conll-emnlp-2021-11 | ['implicatures'] | ['natural-language-processing'] | [-5.40447794e-02 6.87396228e-01 -7.87312239e-02 -5.16371727e-01
-6.25466228e-01 -6.98004067e-01 7.56771624e-01 3.89607906e-01
-5.81368506e-01 3.54286075e-01 6.97716832e-01 -7.95352817e-01
-3.60582680e-01 -4.51612741e-01 -4.81785893e-01 -6.50160983e-02
1.38188556e-01 4.77051228e-01 3.02382916e-01 -4.24971730e-01
3.42026830e-01 2.90409416e-01 -1.17539096e+00 6.90118432e-01
1.02773249e+00 1.68952018e-01 4.65844244e-01 6.45137727e-01
1.29056394e-01 1.22532940e+00 -4.41468239e-01 -1.02243282e-01
-2.17611924e-01 -5.48423350e-01 -1.19138038e+00 -2.23327771e-01
4.56733489e-03 -2.95606285e-01 -1.17952256e-02 7.20744073e-01
1.20888032e-01 3.41689549e-02 7.87886739e-01 -4.53041077e-01
-7.71620512e-01 9.01997626e-01 1.14683084e-01 4.75144804e-01
5.81417561e-01 6.13541365e-01 1.42012167e+00 -4.90329862e-01
8.39909077e-01 1.59717834e+00 5.73895276e-01 5.62290490e-01
-1.42291451e+00 1.26589313e-01 4.76848930e-01 2.05764472e-01
-7.11883843e-01 -6.72712088e-01 6.63806796e-01 -5.87147415e-01
1.08958328e+00 2.05443203e-01 8.33629012e-01 1.01317203e+00
5.41401878e-02 7.02455640e-01 1.73496509e+00 -7.78474331e-01
6.72862008e-02 3.76855552e-01 3.59204322e-01 3.17106009e-01
-1.03228390e-01 5.35152256e-01 -7.74092317e-01 8.69882107e-02
5.75767279e-01 -9.03895557e-01 -1.45508781e-01 1.63233057e-01
-1.21900749e+00 7.71307945e-01 3.98175508e-01 6.53010428e-01
-2.85895526e-01 -1.78170398e-01 2.06825584e-01 4.88411874e-01
1.57909486e-02 1.16246140e+00 -6.99059308e-01 -4.83106107e-01
-1.73050836e-01 4.10760045e-01 8.97505283e-01 7.49416232e-01
1.04995720e-01 -1.87500522e-01 -6.48897663e-02 1.18714356e+00
7.39200771e-01 2.99103081e-01 5.40040910e-01 -1.32942748e+00
3.25780421e-01 5.36330581e-01 2.30608627e-01 -8.10457706e-01
-7.04255164e-01 3.75923127e-01 4.16486084e-01 -1.14781782e-01
8.48757267e-01 -2.16051310e-01 -3.71931821e-01 2.18671632e+00
1.69178590e-01 -6.66762054e-01 2.73650050e-01 7.71840751e-01
6.15872443e-01 6.04080141e-01 6.43186629e-01 -5.31832933e-01
1.40861487e+00 -4.80642170e-01 -4.97386545e-01 -6.77216232e-01
9.79388356e-01 -8.32457602e-01 1.95118022e+00 1.85225099e-01
-1.38178611e+00 -3.89464438e-01 -9.63945210e-01 -4.39882785e-01
5.40514775e-02 -2.99757987e-01 9.23441827e-01 5.26460648e-01
-1.02664137e+00 6.95660651e-01 -8.87120485e-01 -6.19316101e-01
-1.60548776e-01 1.24533109e-01 -7.37631693e-02 8.37366953e-02
-1.17120492e+00 1.48956883e+00 3.82404536e-01 2.95142233e-01
-4.89814430e-01 -2.28444740e-01 -9.14819658e-01 -1.23941951e-01
3.17352824e-02 -2.58847624e-01 1.95461071e+00 -1.01812398e+00
-1.69998407e+00 1.04492271e+00 -1.13827921e-01 -1.40070766e-01
5.61969019e-02 -9.90570784e-02 -4.61050540e-01 2.17455789e-01
5.84102757e-02 7.37527072e-01 1.71914488e-01 -1.04245877e+00
-6.24722302e-01 -1.12481289e-01 5.27677417e-01 4.82562631e-01
-1.45267084e-01 5.00043631e-01 2.87459970e-01 -2.29394630e-01
3.93441617e-01 -9.78860557e-01 2.14022361e-02 -1.48603857e-01
-1.37159452e-01 -5.99486947e-01 1.63175389e-01 -4.64981437e-01
1.28613758e+00 -2.02672005e+00 6.64013475e-02 -4.48907912e-02
-7.11695999e-02 -1.13712132e-01 -1.54285997e-01 5.38097560e-01
9.97485667e-02 3.33372474e-01 -4.30242680e-02 -7.07182810e-02
4.36050624e-01 3.30173016e-01 -1.27354190e-01 2.11157784e-01
4.56895173e-01 8.76717925e-01 -8.45892489e-01 -6.10654831e-01
1.29427493e-01 -1.02276526e-01 -1.07952929e+00 2.85195243e-02
-6.70464516e-01 4.69793826e-01 -2.03179508e-01 4.53994781e-01
-1.42173856e-01 -1.89642638e-01 8.35863829e-01 2.82068759e-01
-2.91561663e-01 1.44256580e+00 -7.16686249e-01 1.22987962e+00
-5.35853624e-01 8.53027344e-01 2.95547664e-01 -6.19921446e-01
2.57700771e-01 3.86779130e-01 -3.62200260e-01 -7.72350192e-01
2.33691394e-01 5.61200380e-01 1.36979902e+00 -8.86877835e-01
2.99004614e-01 -8.08426142e-01 -7.32385889e-02 4.15982664e-01
-4.02964294e-01 -6.14609301e-01 3.13386708e-01 1.07710306e-02
8.12630415e-01 1.79517537e-01 1.68503344e-01 -7.48443902e-01
3.11951041e-01 3.56406808e-01 5.91325104e-01 6.28814399e-01
-2.55513549e-01 6.39646202e-02 7.24863946e-01 -3.33161131e-02
-8.23056042e-01 -8.85673702e-01 -5.75135291e-01 1.23414898e+00
-4.39537056e-02 -3.09702247e-01 -7.41872430e-01 -2.33179048e-01
-3.41006100e-01 1.38982916e+00 -4.50911909e-01 6.73096336e-04
-9.76119757e-01 -6.74869120e-01 3.06090862e-01 4.01652634e-01
-1.49390653e-01 -1.68110371e+00 -9.13591802e-01 3.78271252e-01
-8.65168944e-02 -1.05623209e+00 -6.62082210e-02 6.79022074e-02
-1.04824007e+00 -1.06961584e+00 1.36934474e-01 -9.56772506e-01
2.95324147e-01 -1.57078668e-01 9.90795672e-01 3.11092347e-01
4.89211231e-01 3.78504574e-01 -3.51818770e-01 -4.08434719e-01
-9.91707563e-01 -1.32876366e-01 -7.65729621e-02 -8.05604517e-01
6.64595962e-01 -6.06370091e-01 -2.40186125e-01 1.28399014e-01
-6.84081316e-01 7.95883387e-02 4.42891479e-01 8.80254626e-01
7.86640216e-03 -4.16660279e-01 4.96599078e-01 -1.03448045e+00
1.13502300e+00 -3.35669875e-01 -3.41991782e-01 6.98958337e-02
-5.48347116e-01 7.80504430e-03 3.72401774e-01 -6.62231684e-01
-1.23349488e+00 -6.85762584e-01 -1.55260518e-01 4.33919847e-01
-3.51598233e-01 9.98972893e-01 8.37666318e-02 3.82051855e-01
1.07097387e+00 -3.34787995e-01 3.02520186e-01 -4.80479628e-01
2.82529801e-01 6.68664694e-01 1.14991352e-01 -1.43839443e+00
4.37093735e-01 -1.95398584e-01 -4.48328257e-01 -1.10831428e+00
-9.15503323e-01 1.32797018e-01 -3.44720244e-01 5.75056486e-02
8.93303096e-01 -8.59615445e-01 -7.27756083e-01 2.36553997e-01
-1.04294765e+00 -1.02066934e+00 -3.20950121e-01 8.08226585e-01
-8.41313183e-01 2.52233922e-01 -1.21529055e+00 -7.87573457e-01
2.63737679e-01 -1.36606014e+00 4.29166347e-01 2.74866950e-02
-1.13453794e+00 -1.19723749e+00 -9.17598903e-02 4.01811332e-01
1.92133948e-01 -7.18259364e-02 1.74312150e+00 -9.41218972e-01
-4.24972087e-01 3.34901839e-01 5.97207695e-02 3.04711908e-01
4.75654043e-02 2.15825185e-01 -8.68071616e-01 2.35929549e-01
4.14409131e-01 -6.59873486e-01 9.06777754e-02 3.08422953e-01
4.44572806e-01 -4.42270190e-01 -2.41623763e-02 7.93804005e-02
9.81609523e-01 4.10128266e-01 3.47059965e-01 2.97323734e-01
1.88033357e-01 1.29266405e+00 7.09358454e-01 -2.19892487e-01
7.22433269e-01 5.14154255e-01 -3.15499246e-01 5.90816617e-01
4.12010141e-02 -4.11737800e-01 5.24170578e-01 1.14963162e+00
3.35495949e-01 -5.21129705e-02 -1.16676271e+00 8.37666631e-01
-1.59243011e+00 -6.88055694e-01 -2.69720435e-01 1.81939209e+00
1.37744021e+00 5.81548989e-01 7.33598545e-02 3.37627605e-02
4.18178350e-01 1.90860361e-01 -2.56407380e-01 -8.04000139e-01
6.86242357e-02 1.35561720e-01 -3.08997929e-01 1.15741730e+00
-5.53345084e-01 1.32765746e+00 7.02958584e+00 2.47244835e-01
-9.01757061e-01 -9.48641524e-02 5.66882133e-01 5.21449037e-02
-6.26713574e-01 1.71532929e-01 -4.52561021e-01 2.07960218e-01
1.26839232e+00 2.18295991e-01 5.35747468e-01 6.17730677e-01
4.92954880e-01 -3.80133241e-01 -1.72405005e+00 1.49164712e-02
-4.34870958e-01 -9.08304572e-01 -4.44636494e-01 -1.21546641e-01
4.52438384e-01 -1.35634452e-01 -1.92321226e-01 6.02867842e-01
4.34488505e-01 -1.19754827e+00 1.07864380e+00 1.98585819e-02
5.28880179e-01 -1.18610278e-01 4.18547183e-01 6.96027517e-01
-3.88224602e-01 -2.61879325e-01 -2.81646073e-01 -9.44176972e-01
2.32690781e-01 -3.30451101e-01 -9.74247038e-01 -5.30253112e-01
3.20351213e-01 1.86607137e-01 -3.81447464e-01 4.12866175e-01
-7.65797257e-01 8.34192634e-01 -7.59414315e-01 -6.56965613e-01
2.56155163e-01 -1.03062972e-01 3.70423764e-01 8.17196727e-01
-2.18615502e-01 7.25052893e-01 -9.00915787e-02 7.68371820e-01
4.14938539e-01 3.67947310e-01 -5.99910617e-01 -2.87820160e-01
8.93501222e-01 6.28300488e-01 -7.25745142e-01 3.05673312e-02
-5.78359067e-01 2.43219137e-01 5.38154483e-01 2.77055591e-01
-2.99159795e-01 2.72080928e-01 5.34770370e-01 2.06570923e-01
-1.85760900e-01 -2.91022897e-01 -6.54715836e-01 -9.20203030e-01
-2.84241438e-02 -1.23345852e+00 4.75903898e-02 -8.78357351e-01
-1.25045407e+00 1.23809621e-01 3.63561690e-01 -5.78008413e-01
-5.93848407e-01 -1.02241516e+00 -6.35831118e-01 9.86784697e-01
-1.06146121e+00 -7.94107139e-01 4.59720373e-01 1.86194599e-01
5.53480625e-01 4.22452539e-01 9.13991928e-01 -3.60142618e-01
-2.63224602e-01 2.15117112e-01 -3.22723091e-01 1.02678895e-01
6.06567800e-01 -1.32228076e+00 3.11686277e-01 6.44851983e-01
3.50521840e-02 1.30434060e+00 1.05504155e+00 -5.58204830e-01
-9.53170836e-01 1.25133470e-02 1.15576196e+00 -7.49902666e-01
1.00174785e+00 -4.93482351e-02 -9.19358134e-01 9.63418663e-01
4.09892917e-01 -4.03157562e-01 7.82905996e-01 7.96282351e-01
-1.57489270e-01 3.05647194e-01 -9.90507424e-01 1.05377877e+00
1.15426803e+00 -9.17565107e-01 -1.54087603e+00 5.27897060e-01
8.50251436e-01 -5.50854802e-01 -8.78914416e-01 3.69817436e-01
3.13961685e-01 -8.46920848e-01 5.60017288e-01 -9.50464308e-01
7.00097799e-01 -4.02875990e-03 -2.57808179e-01 -1.16620553e+00
-1.90407872e-01 -5.24107099e-01 1.93053231e-01 1.00659597e+00
8.88823211e-01 -4.83729511e-01 4.22158659e-01 1.13916373e+00
-4.11226422e-01 -7.41808534e-01 -8.63250852e-01 -2.52290249e-01
7.56269217e-01 -5.70261002e-01 6.86048344e-02 8.94190848e-01
9.26375091e-01 7.99513221e-01 4.00067210e-01 1.13106407e-01
-9.76040736e-02 -2.22622845e-02 4.65649188e-01 -1.25186539e+00
-6.18926585e-01 -6.23300850e-01 2.58279890e-01 -1.14357483e+00
1.47374049e-01 -7.03437150e-01 4.97228712e-01 -1.42836678e+00
-2.59726167e-01 -7.62599170e-01 2.74959028e-01 2.61604726e-01
-1.40380964e-01 -3.49723488e-01 4.84741956e-01 2.18976870e-01
-6.57150522e-02 3.84112179e-01 1.37703562e+00 4.56129581e-01
-6.28301024e-01 -1.19930081e-01 -1.18897843e+00 1.29860687e+00
8.33788931e-01 -1.64610103e-01 -7.81499803e-01 -7.56685078e-01
2.76698887e-01 2.97280520e-01 1.45511210e-01 -6.58334136e-01
3.71829718e-02 -7.30600178e-01 -1.63358897e-02 1.82517901e-01
4.62995499e-01 -4.16537970e-01 -2.42702916e-01 3.03165644e-01
-6.89246714e-01 2.44981796e-01 6.27040088e-01 1.03421643e-01
6.26404658e-02 -5.48984408e-01 7.31461704e-01 -2.95774609e-01
-4.03645962e-01 -6.50840044e-01 -7.50759363e-01 5.39260924e-01
6.05899632e-01 -5.66480979e-02 -3.71932328e-01 -2.43655249e-01
-1.18836069e+00 1.53026596e-01 6.30173862e-01 3.18789631e-01
2.37215653e-01 -8.39152694e-01 -3.10829729e-01 -1.17221050e-01
1.46146277e-02 -1.73998568e-02 -1.98995192e-02 9.84278858e-01
-7.41147280e-01 5.70266604e-01 2.66883373e-01 -4.39640164e-01
-9.10708845e-01 4.55776095e-01 4.13542539e-01 -2.94461083e-02
-4.04009700e-01 1.05729294e+00 1.92632496e-01 -6.80041671e-01
-6.00906461e-02 -8.15429568e-01 -2.86832601e-01 1.07746214e-01
2.19095647e-01 -1.46538973e-01 -5.09976625e-01 -6.14482582e-01
-2.26984128e-01 2.23528832e-01 -1.78415477e-01 -6.59432232e-01
1.19306362e+00 -9.56233665e-02 -6.55803010e-02 1.07729185e+00
6.68277264e-01 4.96528298e-01 -1.26797116e+00 -1.36408713e-02
4.53913867e-01 -4.93978076e-02 -4.00439441e-01 -1.00550044e+00
-2.31188592e-02 8.64509642e-01 -1.24124496e-03 2.97621161e-01
4.63742971e-01 2.16196910e-01 2.01968417e-01 6.26226366e-01
2.20565245e-01 -1.42073035e+00 -1.03331000e-01 7.47721910e-01
9.40763474e-01 -9.76889491e-01 -1.12434790e-01 -5.01619339e-01
-6.86879694e-01 7.57991135e-01 1.00294006e+00 -4.30322029e-02
6.74787104e-01 -3.05084232e-02 3.49325866e-01 -2.59515345e-01
-1.19758451e+00 -9.88726169e-02 -1.11178495e-01 6.27905071e-01
8.62885714e-01 1.16303608e-01 -8.92597914e-01 4.70131844e-01
-8.66618991e-01 -3.90450150e-01 7.55988955e-01 8.97949457e-01
-5.91693163e-01 -1.07285762e+00 -3.21150988e-01 3.97706181e-01
-3.51842523e-01 -5.96900105e-01 -4.81755108e-01 1.23205340e+00
-1.77632928e-01 1.11300170e+00 1.12880737e-01 -8.37076530e-02
4.21127170e-01 4.32949543e-01 4.45230365e-01 -1.16556990e+00
-6.26797616e-01 1.35004476e-01 7.14726806e-01 -2.68452495e-01
-4.60085750e-01 -1.21122420e+00 -1.31259799e+00 -3.11018378e-01
-3.73523474e-01 1.87007546e-01 1.53593004e-01 1.34351146e+00
-8.65646899e-02 2.08012387e-01 -5.60470074e-02 -4.61766988e-01
-8.95756423e-01 -1.59048057e+00 -3.41991603e-01 2.17130840e-01
1.45764470e-01 -4.69587684e-01 -5.99917591e-01 -2.70388544e-01] | [10.487058639526367, 8.763480186462402] |
b2e35d7d-d99a-4376-8b34-404506f3f2be | sparse-interventions-in-language-models-with | 2112.06837 | null | https://arxiv.org/abs/2112.06837v1 | https://arxiv.org/pdf/2112.06837v1.pdf | Sparse Interventions in Language Models with Differentiable Masking | There has been a lot of interest in understanding what information is captured by hidden representations of language models (LMs). Typically, interpretation methods i) do not guarantee that the model actually uses the encoded information, and ii) do not discover small subsets of neurons responsible for a considered phenomenon. Inspired by causal mediation analysis, we propose a method that discovers within a neural LM a small subset of neurons responsible for a particular linguistic phenomenon, i.e., subsets causing a change in the corresponding token emission probabilities. We use a differentiable relaxation to approximately search through the combinatorial space. An $L_0$ regularization term ensures that the search converges to discrete and sparse solutions. We apply our method to analyze subject-verb number agreement and gender bias detection in LSTMs. We observe that it is fast and finds better solutions than the alternative (REINFORCE). Our experiments confirm that each of these phenomenons is mediated through a small subset of neurons that do not play any other discernible role. | ['Ivan Titov', 'Dieuwke Hupkes', 'Leon Schmid', 'Nicola De Cao'] | 2021-12-13 | null | null | null | null | ['gender-bias-detection', 'gender-bias-detection'] | ['miscellaneous', 'natural-language-processing'] | [ 4.03858304e-01 5.37070572e-01 -6.95977807e-01 -4.47447479e-01
-4.62252080e-01 -3.53529364e-01 5.22383869e-01 3.05806816e-01
-4.71849203e-01 8.95845652e-01 4.63311464e-01 -3.67145419e-01
-2.36888051e-01 -6.94092929e-01 -1.08715117e+00 -7.72398233e-01
-3.95095646e-02 5.45299172e-01 -4.00905102e-01 9.24017727e-02
1.93509445e-01 1.31659329e-01 -1.41278780e+00 5.51152289e-01
7.66728461e-01 5.99493504e-01 4.53112572e-01 -3.32221314e-02
-4.73620594e-01 8.67116868e-01 -5.34764707e-01 -1.67788006e-02
-2.69696355e-01 -7.36349404e-01 -1.00932848e+00 -8.77782032e-02
2.18493342e-01 2.68496722e-01 -2.90270001e-02 1.18611634e+00
-2.92910524e-02 2.07088694e-01 8.79894435e-01 -1.16452456e+00
-7.73394823e-01 1.40954375e+00 -7.33895719e-01 2.88704962e-01
3.09515912e-02 -2.49210134e-01 1.37602758e+00 -7.47321308e-01
6.95498884e-01 1.68100965e+00 3.39450419e-01 6.84717774e-01
-1.65570629e+00 -7.70065904e-01 6.18885934e-01 5.67053258e-02
-1.27196455e+00 -3.88128132e-01 7.67473042e-01 -7.87489489e-02
1.00090551e+00 3.35851282e-01 3.98023307e-01 1.30225718e+00
2.50914037e-01 8.13712776e-01 1.00074267e+00 -6.18125021e-01
3.63208383e-01 2.66279250e-01 2.93242991e-01 7.27007449e-01
3.32379907e-01 -2.41694108e-01 -8.29238594e-01 -4.32224721e-01
5.12769938e-01 -1.68924555e-01 -2.43347391e-01 2.23749597e-02
-1.14910650e+00 1.18313909e+00 4.49135423e-01 7.05008566e-01
-5.08239388e-01 6.07146919e-01 9.23623815e-02 1.27072617e-01
2.48703957e-01 6.01477027e-01 -6.83421791e-01 1.15975447e-01
-8.94184411e-01 2.18862385e-01 6.04402363e-01 3.72354209e-01
8.57481539e-01 7.43182302e-02 1.01021960e-01 8.16674948e-01
3.92736137e-01 2.72886366e-01 5.62040329e-01 -9.06540394e-01
4.32433635e-01 5.76437294e-01 3.38647626e-02 -1.15255213e+00
-4.96423960e-01 -3.40534627e-01 -7.51939297e-01 -3.10918957e-01
5.77584326e-01 -4.23188321e-02 -6.82474315e-01 2.58737659e+00
2.01836936e-02 2.10897345e-02 -7.32546300e-02 7.94880807e-01
5.66623628e-01 8.67581189e-01 4.01261687e-01 -5.15673399e-01
1.17592537e+00 -2.23911285e-01 -9.92825329e-01 -4.99556810e-01
7.24280179e-01 -3.10569942e-01 1.12183332e+00 7.84289390e-02
-1.22687507e+00 -2.89797246e-01 -6.47989810e-01 -4.52726334e-02
-2.07441792e-01 1.12195566e-01 8.66073310e-01 4.06819075e-01
-8.26427281e-01 3.99454653e-01 -6.34314954e-01 -1.91362679e-01
2.90214241e-01 6.18405938e-01 -9.05612409e-02 3.18468750e-01
-1.42043912e+00 7.32045531e-01 4.59448785e-01 3.09671313e-01
-7.29172945e-01 -5.01164854e-01 -6.30217612e-01 2.64567822e-01
4.24860835e-01 -4.86273617e-01 8.37956667e-01 -1.44213164e+00
-1.11285472e+00 9.06251311e-01 -8.38821411e-01 -4.36752915e-01
-8.54356661e-02 5.78285418e-02 -2.68220812e-01 -1.42191589e-01
2.43106321e-01 9.20463562e-01 8.70399117e-01 -1.50928926e+00
-4.36482787e-01 -4.51830566e-01 -1.40228182e-01 1.03656175e-02
-5.67635894e-01 2.87211150e-01 5.13905138e-02 -6.25672281e-01
4.27061856e-01 -7.93110132e-01 -2.34181225e-01 -4.09718484e-01
-8.25414479e-01 -5.07044852e-01 3.82956803e-01 -3.73846203e-01
1.18254054e+00 -1.91525269e+00 4.62580651e-01 4.66006637e-01
3.43278974e-01 -2.08676636e-01 -1.00114845e-01 3.37262183e-01
-3.52942199e-01 7.17300415e-01 -4.97116536e-01 -4.88694996e-01
-4.94216830e-02 6.17673874e-01 -9.55370307e-01 4.78674203e-01
2.01430976e-01 8.80632997e-01 -7.29040444e-01 -3.41063619e-01
-2.39424840e-01 3.21870357e-01 -6.95903420e-01 -1.53167397e-01
-5.47825634e-01 3.55893850e-01 -4.60774243e-01 3.70165706e-01
3.75753701e-01 -3.73500019e-01 6.23054862e-01 -3.85654122e-02
-2.25845814e-01 5.16517937e-01 -1.24540806e+00 1.38566494e+00
-3.64003778e-01 6.03574276e-01 -1.11630097e-01 -1.22064793e+00
9.36066568e-01 2.87437439e-01 1.01202719e-01 -4.35654193e-01
1.90338090e-01 3.68268579e-01 3.16365063e-01 -3.49058032e-01
3.46312881e-01 -3.69769782e-01 -9.87683386e-02 6.65052474e-01
8.40782523e-02 6.30041897e-01 -3.80079262e-03 1.26383826e-01
7.98047185e-01 -2.96822250e-01 2.81199515e-01 -5.56571186e-01
1.13173507e-01 -6.33264333e-02 8.33280921e-01 1.07900190e+00
3.31595600e-01 1.58093497e-01 9.31053638e-01 -2.38056600e-01
-6.52395606e-01 -9.81964350e-01 -9.79603380e-02 1.15462387e+00
-9.66074839e-02 -6.24660552e-02 -5.75669527e-01 -2.37187505e-01
-2.08672181e-01 1.15527177e+00 -1.03069830e+00 -3.26353729e-01
-7.62454689e-01 -1.04207456e+00 5.21684051e-01 3.92937481e-01
3.26506644e-01 -1.42614710e+00 -6.23011947e-01 2.05569401e-01
-5.08086145e-01 -8.42065334e-01 -1.33690238e-01 5.61745286e-01
-1.14774740e+00 -7.33502269e-01 -1.25249863e-01 -9.11898494e-01
1.04176366e+00 -9.19383243e-02 8.94475937e-01 3.81720513e-02
1.11292467e-01 -1.05808698e-01 3.32252681e-02 -2.75233656e-01
-3.03795516e-01 2.02748075e-01 1.92300111e-01 1.87640652e-01
5.77748656e-01 -5.31448364e-01 -7.47508928e-02 -1.41842306e-01
-7.27666557e-01 -9.10109803e-02 5.72006166e-01 7.80361414e-01
5.57851017e-01 1.97729126e-01 5.79009831e-01 -9.10436928e-01
8.30345869e-01 -7.54545033e-01 -3.89850706e-01 1.24518588e-01
-3.95452887e-01 5.63738048e-01 5.44126809e-01 -6.35375500e-01
-1.09839380e+00 1.39004529e-01 3.03000867e-01 -2.56117791e-01
-3.29961270e-01 9.88388658e-01 -2.76300639e-01 5.80554187e-01
5.22261262e-01 8.17753375e-02 -2.27676436e-01 -6.27648294e-01
3.69062722e-01 1.35109693e-01 1.88593671e-01 -7.06275165e-01
4.17015612e-01 6.20132267e-01 -3.22606005e-02 -7.83321679e-01
-8.86059284e-01 -9.69132632e-02 -4.42496270e-01 7.32306540e-02
7.47009635e-01 -4.65157092e-01 -9.17776346e-01 -7.77981654e-02
-1.49776602e+00 -1.83264107e-01 -3.38813245e-01 5.70548236e-01
-3.66368800e-01 -1.99068531e-01 -4.63793337e-01 -1.07428455e+00
1.84149817e-01 -8.64709139e-01 6.45506263e-01 -4.48865891e-02
-7.22993314e-01 -9.43025351e-01 -2.82463897e-03 8.51844263e-04
-8.76976475e-02 -1.10641897e-01 1.56935608e+00 -7.39598513e-01
-3.24827760e-01 1.68411508e-01 -2.87131444e-02 -2.43979469e-01
1.76746622e-01 -7.89345577e-02 -8.91645670e-01 6.97114021e-02
2.03886896e-01 -1.84618011e-01 1.36664367e+00 9.29895759e-01
1.35006440e+00 -8.19429934e-01 -4.93397683e-01 2.80121773e-01
1.11804307e+00 2.08479688e-01 4.80929375e-01 6.40903264e-02
6.13680124e-01 8.85809958e-01 4.84045297e-02 1.79735512e-01
2.53666997e-01 5.66814363e-01 4.28579301e-01 -2.09533691e-01
4.16965097e-01 -4.29596782e-01 4.86669093e-01 5.90246916e-01
-1.46904856e-01 -3.57793421e-01 -8.37484121e-01 7.98866510e-01
-1.97839415e+00 -1.14877582e+00 -3.22517693e-01 2.15177393e+00
1.26190794e+00 2.91428920e-02 -1.17964379e-01 9.30866078e-02
7.80646443e-01 1.40660763e-01 -5.04515409e-01 -6.82979882e-01
-3.85479152e-01 2.67433673e-01 4.04271543e-01 7.40629911e-01
-6.09263659e-01 1.18828547e+00 6.50480890e+00 7.19378650e-01
-1.02758336e+00 1.03564426e-01 8.96527469e-01 -9.97113213e-02
-7.88509190e-01 1.65201072e-02 -9.39381838e-01 1.69334874e-01
1.00688064e+00 -4.83219326e-02 4.22752827e-01 3.73678356e-01
6.83789492e-01 -1.62010461e-01 -1.18296862e+00 6.29703045e-01
-1.41492382e-01 -1.42674685e+00 4.65892971e-01 4.11036998e-01
7.04733253e-01 -3.02067697e-01 2.72557467e-01 9.28787980e-03
3.87405157e-01 -1.43841934e+00 9.40198541e-01 4.64899927e-01
3.14616650e-01 -8.24943542e-01 2.47919574e-01 5.68933666e-01
-8.42250764e-01 -2.28719756e-01 -6.45241857e-01 -2.05439016e-01
-3.93679179e-02 6.56741619e-01 -8.46727312e-01 -1.35955468e-01
4.06700313e-01 3.38360131e-01 -9.71816480e-02 4.23805386e-01
-6.18316829e-01 1.15131879e+00 -2.40973085e-01 -3.33155125e-01
3.34062070e-01 -4.86976989e-02 5.71844399e-01 1.03359997e+00
5.99553771e-02 1.09573252e-01 -1.31643817e-01 1.70567060e+00
-2.54658997e-01 1.90034524e-01 -6.11181080e-01 -2.62553185e-01
5.54037809e-01 7.95170724e-01 -9.46242690e-01 -4.12620194e-02
-1.36861533e-01 6.32686138e-01 4.27460641e-01 7.01124251e-01
-7.40482092e-01 1.66691124e-01 5.60666859e-01 -9.01146308e-02
3.13020527e-01 -2.36448810e-01 -4.63465571e-01 -9.14644480e-01
-1.79300860e-01 -7.33369350e-01 2.97793418e-01 -5.36991179e-01
-1.07849026e+00 3.34831476e-01 3.53824720e-02 -4.03104097e-01
-4.34027165e-01 -3.04231405e-01 -4.94169325e-01 9.63500261e-01
-1.25420010e+00 -7.74823785e-01 5.34558415e-01 4.68229860e-01
4.65081722e-01 5.12568280e-02 9.81038213e-01 -5.55156842e-02
-7.03601778e-01 3.73716801e-01 -8.95421058e-02 7.83352926e-02
9.81144682e-02 -9.63256240e-01 -1.23661133e-02 6.90398097e-01
4.63041544e-01 1.20813143e+00 9.51746881e-01 -7.41143346e-01
-1.10054255e+00 -9.51338828e-01 1.57096303e+00 -2.56573319e-01
5.97200990e-01 -4.04814541e-01 -1.05619073e+00 1.03052580e+00
2.42816165e-01 -4.46734935e-01 7.47931838e-01 6.08998716e-01
-4.09976810e-01 1.86454907e-01 -9.00801897e-01 6.93143368e-01
9.58429515e-01 -4.78452772e-01 -8.51979911e-01 1.28539026e-01
6.60849750e-01 1.13096885e-01 -1.95416600e-01 1.30282715e-01
2.52585381e-01 -4.25698936e-01 8.20724249e-01 -1.22202408e+00
5.78785777e-01 -4.09490578e-02 -3.19546163e-01 -1.29697049e+00
-2.09560737e-01 -4.19145882e-01 -1.53092504e-01 1.15377140e+00
8.96385968e-01 -7.15700507e-01 5.37823319e-01 6.27865076e-01
1.54716164e-01 -6.25927508e-01 -1.20895529e+00 -5.14254987e-01
2.91014493e-01 -5.61067641e-01 4.31051672e-01 1.27839053e+00
1.25562474e-01 4.08451527e-01 -3.64686131e-01 2.31356829e-01
7.03019500e-01 4.24576789e-01 7.30564743e-02 -1.19181514e+00
-2.56943315e-01 -3.54076743e-01 1.57986164e-01 -9.95476842e-01
7.32358098e-01 -1.03735399e+00 9.03771296e-02 -1.37809575e+00
5.12606025e-01 -5.00948012e-01 -5.28929114e-01 8.32074761e-01
5.85370697e-02 -1.12007752e-01 -1.16156444e-01 2.58859962e-01
-1.89134449e-01 5.26063681e-01 8.04385900e-01 -1.49875358e-01
-3.93608183e-01 -2.02801317e-01 -9.89467919e-01 9.16752160e-01
9.85037565e-01 -1.07121205e+00 -2.42646694e-01 -5.12134790e-01
7.79527366e-01 -2.15957873e-02 5.99975526e-01 -2.35199586e-01
4.61822897e-02 -5.10378599e-01 3.05426300e-01 -2.64501393e-01
3.35708857e-01 -4.61117566e-01 -1.26196563e-01 5.54572582e-01
-1.06338966e+00 -7.33009279e-02 1.19528636e-01 4.16841000e-01
-9.16695967e-02 -5.31768322e-01 5.51215529e-01 -3.37457806e-01
-5.30346155e-01 -1.31784514e-01 -8.38035285e-01 1.54996589e-01
3.16634506e-01 6.57626539e-02 -4.16910127e-02 -3.76161486e-01
-9.10140455e-01 -6.90819398e-02 -1.36822686e-01 4.51718062e-01
6.55045092e-01 -1.20364201e+00 -4.97367769e-01 -3.23429294e-02
-2.68219769e-01 -4.17761385e-01 -1.69749394e-01 6.59708261e-01
3.94827187e-01 6.49393439e-01 5.66938594e-02 -3.45313996e-01
-1.10311055e+00 2.51879662e-01 4.59061742e-01 -1.41051784e-01
-3.45647037e-01 1.12552845e+00 4.93076414e-01 -3.38579267e-01
1.25235781e-01 -7.11881995e-01 -4.15209800e-01 3.05854559e-01
3.15134495e-01 1.81110471e-01 -3.11833978e-01 -7.69447327e-01
-4.14063573e-01 2.50310719e-01 -1.14483498e-02 -3.87713045e-01
1.49155879e+00 -4.30714041e-02 -4.99016196e-01 1.03531480e+00
1.11594534e+00 9.71461833e-02 -7.93579876e-01 -2.08612293e-01
2.39176676e-01 -2.90267915e-01 2.37485319e-01 -7.07152724e-01
-1.12229538e+00 7.67904043e-01 2.55515814e-01 3.07603717e-01
8.02383065e-01 3.88365746e-01 4.17170495e-01 4.55653578e-01
1.35188088e-01 -1.10233235e+00 -1.52996495e-01 4.44165647e-01
9.47742164e-01 -7.51280963e-01 -3.38411897e-01 -3.69589031e-01
-4.89803225e-01 1.02646220e+00 3.26332390e-01 -3.91861098e-03
4.03661013e-01 1.31546900e-01 -2.02702865e-01 -3.50613713e-01
-1.08592629e+00 -1.19595185e-01 8.44988376e-02 2.26150244e-01
6.11764252e-01 1.89270884e-01 -4.81595695e-01 7.87419081e-01
-1.79638594e-01 -2.65662044e-01 4.44524914e-01 4.21157688e-01
-4.69950944e-01 -9.72143233e-01 -4.81987834e-01 3.92168552e-01
-4.72839206e-01 -4.15674806e-01 -5.73254526e-01 6.60503626e-01
2.50300437e-01 1.06116414e+00 2.46926084e-01 4.55685370e-02
-5.34256399e-02 1.60476223e-01 4.47525173e-01 -6.74717724e-01
-3.97241443e-01 2.21742347e-01 -2.87104566e-02 -5.42582154e-01
-5.40394962e-01 -1.07038522e+00 -1.92328095e+00 -1.32158205e-01
-2.89876759e-01 2.93848217e-01 5.51411390e-01 1.23883164e+00
5.42020053e-02 5.26715100e-01 4.00453925e-01 -2.98826814e-01
-4.20058757e-01 -7.75753617e-01 -6.32149875e-01 -1.45063165e-03
2.18718037e-01 -6.90113246e-01 -5.38942397e-01 -5.27680758e-03] | [8.995101928710938, 6.1492719650268555] |
cbc17f37-d116-4685-ab02-4ae54530be59 | drug-drug-interaction-extraction-via-1 | null | null | https://academic.oup.com/bioinformatics/article/34/5/828/4565590 | https://academic.oup.com/bioinformatics/article-pdf/34/5/828/25117737/btx659.pdf | Drug–drug interaction extraction via hierarchical RNNs on sequence and shortest dependency paths | Motivation
Adverse events resulting from drug-drug interactions (DDI) pose a serious health issue. The ability to automatically extract DDIs described in the biomedical literature could further efforts for ongoing pharmacovigilance. Most of neural networks-based methods typically focus on sentence sequence to identify these DDIs, however the shortest dependency path (SDP) between the two entities contains valuable syntactic and semantic information. Effectively exploiting such information may improve DDI extraction.
Results
In this article, we present a hierarchical recurrent neural networks (RNNs)-based method to integrate the SDP and sentence sequence for DDI extraction task. Firstly, the sentence sequence is divided into three subsequences. Then, the bottom RNNs model is employed to learn the feature representation of the subsequences and SDP, and the top RNNs model is employed to learn the feature representation of both sentence sequence and SDP. Furthermore, we introduce the embedding attention mechanism to identify and enhance keywords for the DDI extraction task. We evaluate our approach using the DDI extraction 2013 corpus. Our method is competitive or superior in performance as compared with other state-of-the-art methods. Experimental results show that the sentence sequence and SDP are complementary to each other. Integrating the sentence sequence with SDP can effectively improve the DDI extraction performance. | ['Wei Zheng', 'Jian Wang', 'Yijia Zhang', 'Zhihao Yang', 'Michel Dumontier', 'Hongfei Lin'] | 2017-10-25 | null | null | null | bioinformatics-2017-10 | ['drug-drug-interaction-extraction'] | ['natural-language-processing'] | [ 4.45760041e-01 -1.10987306e-01 -6.73733473e-01 -3.82789105e-01
-7.76811004e-01 -3.47060978e-01 3.07848364e-01 6.13861442e-01
-5.21760166e-01 8.54273021e-01 6.24024332e-01 -4.26383823e-01
-1.65056348e-01 -6.71714306e-01 -6.38088465e-01 -5.73381305e-01
-2.89411247e-01 9.52249989e-02 -3.45141679e-01 6.38242960e-02
2.55202144e-01 6.31621659e-01 -7.85879195e-01 4.02161986e-01
1.15818226e+00 6.64855480e-01 3.94615293e-01 2.92040080e-01
-5.67974150e-01 9.32653964e-01 -7.55504668e-01 -7.46495798e-02
-2.11460337e-01 -5.24874508e-01 -9.01901782e-01 -4.55235332e-01
-3.43173087e-01 -3.73744220e-01 -4.45883095e-01 1.17012072e+00
9.29303944e-01 -1.71549916e-01 7.32291222e-01 -7.27654338e-01
-7.86340535e-01 7.21618056e-01 -5.86728752e-01 4.75337565e-01
5.51666200e-01 -5.03716096e-02 1.13207042e+00 -1.03532219e+00
9.66519892e-01 1.13891828e+00 4.88003701e-01 5.48682690e-01
-7.17283845e-01 -6.63887203e-01 1.18141323e-01 5.00239193e-01
-1.25707269e+00 -1.60974354e-01 8.28167975e-01 -2.66258985e-01
1.66926908e+00 -3.58600989e-02 5.26186943e-01 1.29819083e+00
7.00137854e-01 1.31398094e+00 2.91927695e-01 -3.31719309e-01
-1.62956528e-02 1.10981269e-02 4.99571949e-01 3.00382793e-01
2.90808231e-01 1.70049481e-02 -2.83099979e-01 -2.50598550e-01
2.89478272e-01 3.93598676e-01 -3.06279898e-01 4.24835056e-01
-9.76240814e-01 1.06305051e+00 2.80218393e-01 5.06073952e-01
-9.07922745e-01 -3.23871791e-01 7.77233779e-01 -8.97159725e-02
5.32230854e-01 5.21440387e-01 -9.11356151e-01 7.24645406e-02
-7.33989060e-01 1.20176144e-01 8.83657396e-01 8.69402230e-01
1.64493874e-01 -2.10635334e-01 -4.93362069e-01 9.65249062e-01
3.50394875e-01 4.27916139e-01 6.54256940e-01 -1.92575425e-01
6.92908287e-01 6.81073368e-01 -1.81680679e-01 -1.03896677e+00
-6.23760343e-01 -3.81725490e-01 -1.07460976e+00 -7.24151134e-01
-3.22247714e-01 -5.65930903e-01 -9.94433641e-01 1.53251624e+00
3.71265084e-01 6.02071285e-02 3.60964328e-01 4.72969592e-01
1.50109124e+00 1.00648260e+00 3.82190704e-01 -4.26190048e-01
1.61441362e+00 -6.58963263e-01 -1.33475912e+00 5.92582859e-03
9.72680986e-01 -6.96675062e-01 2.48420253e-01 9.08263475e-02
-9.36597526e-01 -1.44613504e-01 -1.21868777e+00 -6.99540377e-02
-5.20172477e-01 2.36191869e-01 2.24684790e-01 1.03643969e-01
-4.05274630e-01 7.95550704e-01 -6.70577168e-01 4.55781706e-02
7.21443057e-01 5.44187248e-01 -9.89511088e-02 -4.01214696e-02
-2.03416896e+00 1.23393750e+00 5.76084018e-01 3.06654155e-01
-5.51616669e-01 -1.03753030e+00 -1.25439656e+00 1.75746799e-01
2.13024154e-01 -7.70571172e-01 1.14108598e+00 -3.51378769e-01
-1.27867758e+00 3.67452443e-01 -4.64119375e-01 -7.16093540e-01
-2.99650878e-01 -2.67551720e-01 -4.57321405e-01 2.81591982e-01
2.92590588e-01 4.37405497e-01 3.31923515e-01 -5.55543244e-01
-5.64272583e-01 -5.02367973e-01 -3.56571138e-01 2.83813983e-01
-2.73170888e-01 3.50287795e-01 -1.97238892e-01 -7.17437923e-01
-3.27127576e-01 -6.42846763e-01 -4.05072272e-01 -6.11396790e-01
-8.50722909e-01 -6.91002846e-01 4.10635859e-01 -9.68209326e-01
1.74183178e+00 -1.88303828e+00 1.64580226e-01 1.03779241e-01
4.01668370e-01 6.30030870e-01 -4.17983890e-01 7.31230438e-01
-4.75796372e-01 3.39726686e-01 -1.89174920e-01 2.85816081e-02
-2.35686645e-01 1.03970222e-01 -6.14051111e-02 3.44238788e-01
6.84661746e-01 1.06422746e+00 -9.70624030e-01 -5.98322868e-01
-9.69771892e-02 7.23509431e-01 -2.19755784e-01 2.80501753e-01
-1.82174444e-01 4.01800841e-01 -8.03180575e-01 5.22011936e-01
6.19335175e-01 -4.29200888e-01 3.56681705e-01 -3.44630033e-01
1.59590542e-02 1.03560734e+00 -6.61249518e-01 1.61383033e+00
-3.00394058e-01 2.08778903e-01 -6.33673012e-01 -1.01733387e+00
6.49099231e-01 7.06584811e-01 7.49724746e-01 -8.29066396e-01
1.05449356e-01 1.60059243e-01 2.47020900e-01 -9.09864247e-01
-9.93900746e-02 -1.65529847e-01 1.15291318e-02 2.79231161e-01
1.06258601e-01 5.93904674e-01 3.45961958e-01 2.30073869e-01
1.12629175e+00 1.81847520e-03 6.41958833e-01 5.55364452e-02
8.17070961e-01 -1.70757204e-01 9.74721372e-01 4.00427878e-01
3.84439416e-02 1.97649434e-01 6.51870370e-01 -5.70458174e-01
-8.18242848e-01 -5.38893342e-01 -3.57478678e-01 3.53955418e-01
-2.13182002e-01 -5.88760257e-01 -7.78082073e-01 -9.40328658e-01
-1.37152076e-01 6.79766059e-01 -6.47987127e-01 -1.89468518e-01
-6.56321645e-01 -1.15008044e+00 5.97706556e-01 6.06849492e-01
3.41777772e-01 -1.38015556e+00 -1.15280747e-01 6.60122633e-01
-2.65808463e-01 -9.33673561e-01 -7.77252555e-01 2.79226750e-01
-6.81896210e-01 -1.08899808e+00 -8.56441975e-01 -8.60792279e-01
4.86592621e-01 -1.07695665e-02 9.27648485e-01 -1.64414868e-01
-4.28015113e-01 -3.22368145e-01 -3.21237504e-01 -9.16066170e-01
-3.28854948e-01 2.51264215e-01 -5.77799901e-02 -5.08408070e-01
1.09193492e+00 -3.88087660e-01 -7.06916392e-01 -3.58908445e-01
-9.49976981e-01 -9.74517018e-02 8.24348927e-01 7.00005770e-01
5.48351109e-01 -1.37881070e-01 9.35239196e-01 -1.26966846e+00
9.15709436e-01 -7.91613281e-01 -1.84021801e-01 1.98800951e-01
-5.54479539e-01 4.79706109e-01 5.88892877e-01 -2.28039503e-01
-9.62724984e-01 6.59716725e-02 -6.91969872e-01 5.95861152e-02
-2.46647805e-01 1.24005377e+00 -3.83455783e-01 5.92143476e-01
2.45543599e-01 3.56261343e-01 -8.15885961e-02 -5.64566135e-01
1.84980308e-04 9.11460876e-01 -2.05003530e-01 -5.54150417e-02
1.46569952e-01 -1.40014410e-01 -1.38820559e-01 -8.89225602e-01
-1.10148132e+00 -5.36304832e-01 -4.77999419e-01 5.62710464e-01
1.04949868e+00 -7.80417740e-01 -1.04711020e+00 1.54797837e-01
-1.91362572e+00 4.85462993e-01 1.97607711e-01 8.19970191e-01
-1.16368078e-01 5.97761810e-01 -1.10533905e+00 -5.95426619e-01
-1.09709692e+00 -1.28796959e+00 8.89189899e-01 2.35069409e-01
-5.56103051e-01 -9.22999084e-01 4.95344341e-01 -9.14115980e-02
-5.52117191e-02 1.72507048e-01 1.37280476e+00 -1.38556397e+00
-1.34000123e-01 -2.88097918e-01 -3.08911324e-01 2.13469014e-01
5.83994567e-01 -2.82436430e-01 -5.29703975e-01 1.64377034e-01
2.20565960e-01 2.32475519e-01 7.45653629e-01 7.58426130e-01
1.03341329e+00 -7.91360557e-01 -5.83202660e-01 3.60207945e-01
1.33315992e+00 9.40601528e-01 6.22811735e-01 3.59947592e-01
8.28750968e-01 7.20754921e-01 3.29358488e-01 3.92986536e-01
3.40912491e-01 3.44098389e-01 -6.00804351e-02 -3.34151417e-01
5.34133092e-02 -9.08334851e-02 3.00198942e-01 9.85089362e-01
2.24310622e-01 -1.34968504e-01 -9.33010578e-01 5.24159789e-01
-1.69045830e+00 -7.76411414e-01 -1.11465439e-01 1.80492544e+00
1.43619633e+00 -4.84949760e-02 -2.40812767e-02 -1.06884606e-01
6.63579285e-01 -2.19825543e-02 -6.91025436e-01 -6.07030690e-01
-4.50183963e-03 4.32594359e-01 4.09076303e-01 3.05191100e-01
-1.07216454e+00 6.59524977e-01 5.47047281e+00 8.40919554e-01
-7.27821589e-01 -1.70936286e-01 4.38398123e-01 -7.22291693e-02
-2.71484017e-01 -5.19817472e-01 -1.32520747e+00 7.31837869e-01
1.44068956e+00 -2.46586218e-01 -1.63982436e-01 4.39965874e-01
5.85260093e-01 3.90956521e-01 -1.22491848e+00 7.91778445e-01
1.04579188e-01 -1.67446625e+00 3.47685486e-01 1.23870790e-01
3.79613996e-01 1.09242342e-01 -9.76144150e-02 1.10939823e-01
1.10052273e-01 -1.18068171e+00 -2.29174927e-01 4.72276747e-01
4.95757073e-01 -1.20712090e+00 1.31224132e+00 2.57351846e-01
-1.09961891e+00 2.08717927e-01 -3.31030190e-01 4.90496963e-01
7.40925744e-02 8.11704457e-01 -1.01566660e+00 1.06410682e+00
3.69742393e-01 1.47453308e+00 -1.99520513e-01 9.51212287e-01
-4.96171087e-01 3.87686789e-01 -3.40769812e-02 -4.55099910e-01
5.09979367e-01 -1.97774485e-01 4.90625799e-01 1.58376944e+00
6.62528500e-02 2.27507710e-01 -9.80428606e-02 7.64983654e-01
-3.89422327e-01 5.29145420e-01 -8.64795923e-01 -5.60549676e-01
3.75493526e-01 8.53227735e-01 -3.19365174e-01 -5.03818989e-01
-6.03265941e-01 8.96247625e-01 -4.32921723e-02 3.64093900e-01
-6.99410737e-01 -1.10017169e+00 7.44124830e-01 -4.47587848e-01
3.90643895e-01 2.48121262e-01 -1.91307560e-01 -1.07229829e+00
6.60992265e-02 -1.02931404e+00 5.78386009e-01 -3.16445023e-01
-1.44061255e+00 8.30860078e-01 -2.30579466e-01 -1.36533034e+00
-3.67505848e-01 -4.40538198e-01 -5.08536935e-01 1.20746148e+00
-1.77305400e+00 -8.52068245e-01 7.35373259e-01 1.09333143e-01
7.58780420e-01 -7.46194646e-02 1.03125966e+00 5.95459998e-01
-1.07466078e+00 3.67777109e-01 2.46352598e-01 5.82833946e-01
5.67810297e-01 -9.96541560e-01 4.81962353e-01 3.67722601e-01
9.08717792e-03 1.18602657e+00 4.57419485e-01 -1.10885108e+00
-1.21624553e+00 -1.22233057e+00 1.60370862e+00 -2.41146252e-01
5.25468171e-01 -7.05033913e-02 -1.05351555e+00 4.13543880e-01
5.17172515e-01 -5.84379017e-01 1.32053709e+00 9.99777764e-02
-1.10224448e-01 2.53587246e-01 -7.90519714e-01 4.99478191e-01
6.39171124e-01 -5.83950400e-01 -1.04233408e+00 6.59685373e-01
1.04891694e+00 -3.53253901e-01 -1.16299498e+00 5.20553291e-01
4.30183232e-01 -4.11028080e-02 1.27490366e+00 -1.22675443e+00
9.89467025e-01 -1.28727049e-01 3.14640522e-01 -1.27269328e+00
-1.69809923e-01 -3.11659217e-01 -2.24602073e-01 1.07428133e+00
1.03599238e+00 -4.24045205e-01 5.27784348e-01 4.60552245e-01
-8.98807198e-02 -1.35192156e+00 -6.26990736e-01 -5.58033526e-01
1.02702692e-01 -2.59915620e-01 6.51043236e-01 8.61329377e-01
3.77914399e-01 1.03815401e+00 -3.71584654e-01 7.82947689e-02
3.09494942e-01 1.17034435e-01 -9.35930759e-02 -1.01184273e+00
-2.89603081e-02 -3.05844456e-01 -1.02392010e-01 -9.09077942e-01
3.36343825e-01 -1.17432356e+00 -5.95440939e-02 -1.86669505e+00
5.75358927e-01 4.49736148e-01 -5.70506096e-01 5.37964940e-01
-4.81409818e-01 -5.36890924e-01 -1.22080848e-01 2.98238806e-02
-3.39230865e-01 6.18462563e-01 1.16491377e+00 -4.95787233e-01
-5.85328102e-01 -7.72605985e-02 -7.09187508e-01 7.15485752e-01
7.11746812e-01 -9.23088670e-01 -2.56506920e-01 -1.29225641e-01
2.56422371e-01 1.56107754e-01 -3.08970779e-01 -7.25248754e-02
3.93844903e-01 -6.77032908e-03 6.28621638e-01 -1.05700850e+00
-1.24083169e-01 -5.23308039e-01 -3.50178808e-01 7.63199091e-01
-6.31372333e-01 -1.36187166e-01 5.10645092e-01 5.92076719e-01
-3.27939034e-01 -5.37839472e-01 2.89961189e-01 -2.33986825e-01
-1.00937963e-01 5.80575764e-01 -6.57846034e-01 -5.01196831e-02
5.25737524e-01 3.69758904e-02 -6.84460551e-02 2.85033751e-02
-7.63807237e-01 4.67333972e-01 -4.92836416e-01 4.05362904e-01
1.05599380e+00 -1.31569898e+00 -8.63623321e-01 3.82739701e-03
-1.67681230e-03 -1.40655741e-01 3.87723923e-01 7.97570646e-01
-2.55247533e-01 9.70950842e-01 1.23639092e-01 -1.84053093e-01
-1.51286960e+00 8.16444755e-01 8.97718072e-02 -8.16246152e-01
-6.61542833e-01 7.40852833e-01 3.71714711e-01 -3.17527354e-01
3.29450935e-01 -2.78497398e-01 -8.24817121e-01 3.20203483e-01
1.00139964e+00 -5.76414317e-02 1.66330919e-01 -3.17381769e-01
-6.72808886e-01 3.72806579e-01 -8.60951066e-01 3.46366525e-01
1.89270425e+00 3.54456864e-02 -4.77661610e-01 1.10693052e-01
1.78998077e+00 -4.51768875e-01 -3.94397527e-01 -3.91037852e-01
5.31330109e-01 1.22017048e-01 -1.10070093e-03 -8.73396337e-01
-8.56361389e-01 9.73688066e-01 3.03460151e-01 -1.83573574e-01
9.22683895e-01 -8.72376114e-02 1.44284534e+00 4.48735803e-01
-2.93612510e-01 -9.64851499e-01 -2.51881927e-01 6.23344064e-01
8.37935448e-01 -1.00324416e+00 8.09073150e-02 -3.42756152e-01
-7.22640932e-01 1.17412090e+00 8.49161148e-02 5.46493605e-02
7.60734379e-01 2.07889453e-01 -1.83881730e-01 -5.51276088e-01
-6.41402543e-01 3.25381868e-02 6.02177441e-01 2.35488638e-01
7.77467489e-01 -2.67741740e-01 -6.80580020e-01 1.00361419e+00
3.76565427e-01 8.90105441e-02 3.63058686e-01 1.03549337e+00
-1.22062713e-01 -1.49932885e+00 1.26029477e-01 6.50250852e-01
-1.05039179e+00 -7.29582608e-01 -4.25734371e-01 4.83054638e-01
-3.88588607e-02 8.07971358e-01 -3.88747990e-01 -1.66467428e-01
5.35999358e-01 2.70885557e-01 1.51904777e-01 -9.20516849e-01
-8.34912419e-01 3.48470718e-01 1.51593640e-01 -5.13286531e-01
-5.66956460e-01 -4.79980707e-01 -1.55474985e+00 1.37977108e-01
-3.23560476e-01 3.46990883e-01 6.86299026e-01 1.16422319e+00
7.23773658e-01 8.82011056e-01 7.71452010e-01 -1.80495679e-01
-3.33765268e-01 -1.02249908e+00 -3.13393444e-01 7.90629536e-02
6.52328372e-01 -2.79690206e-01 2.43266419e-04 -1.28374752e-02] | [8.356643676757812, 8.660877227783203] |
0c93f3f7-ead7-4c55-acc1-ca7d573356bd | rfnet-recurrent-forward-network-for-dense | null | null | http://openaccess.thecvf.com//content/ICCV2021/html/Huang_RFNet_Recurrent_Forward_Network_for_Dense_Point_Cloud_Completion_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Huang_RFNet_Recurrent_Forward_Network_for_Dense_Point_Cloud_Completion_ICCV_2021_paper.pdf | RFNet: Recurrent Forward Network for Dense Point Cloud Completion | Point cloud completion is an interesting and challenging task in 3D vision, aiming to recover complete shapes from sparse and incomplete point clouds. Existing learning-based methods often require vast computation cost to achieve excellent performance, which limits their practical applications. In this paper, we propose a novel Recurrent Forward Network (RFNet), which is composed of three modules: Recurrent Feature Extraction (RFE), Forward Dense Completion (FDC) and Raw Shape Protection (RSP). The RFE extracts multiple global features from the incomplete point clouds for different recurrent levels, and the FDC generates point clouds in a coarse-to-fine pipeline. The RSP introduces details from the original incomplete models to refine the completion results. Besides, we propose a Sampling Chamfer Distance to better capture the shapes of models and a new Balanced Expansion Constraint to restrict the expansion distances from coarse to fine. According to the experiments on ShapeNet and KITTI, our network can achieve the state-of-the-art with lower memory cost and faster convergence. | ['Yong liu', 'Yifan Xu', 'Yi Yuan', 'Jiangning Zhang', 'Xiangrui Zhao', 'Mengmeng Wang', 'Xuemeng Yang', 'Jinhao Cui', 'Hao Zou', 'Tianxin Huang'] | 2021-01-01 | null | null | null | iccv-2021-1 | ['point-cloud-completion'] | ['computer-vision'] | [-1.68210298e-01 -4.12422776e-01 2.10716173e-01 -2.76706725e-01
-7.30782568e-01 -4.08787549e-01 4.89230901e-01 -2.57449120e-01
-1.10821120e-01 3.32574770e-02 -2.20829062e-02 -1.20014362e-02
-1.06042691e-01 -8.48099470e-01 -7.50561655e-01 -4.70120579e-01
2.09003195e-01 7.02911377e-01 5.15833437e-01 -5.56978770e-03
2.37029523e-01 1.10067165e+00 -1.57889283e+00 1.60820350e-01
9.27581906e-01 1.11120057e+00 5.47778368e-01 1.69801578e-01
-4.10881668e-01 2.81082779e-01 -1.21667556e-01 -7.24202022e-02
5.65826058e-01 4.72626656e-01 -2.97899604e-01 5.21811433e-02
3.11403215e-01 -5.24843872e-01 -3.84185791e-01 9.56260622e-01
5.48275828e-01 -8.31111223e-02 3.72360438e-01 -1.21685731e+00
-5.64954937e-01 2.05648273e-01 -8.78089249e-01 -4.52742279e-01
6.93240389e-02 1.43335819e-01 5.70899725e-01 -1.60179114e+00
6.05514884e-01 1.39068794e+00 9.93215561e-01 3.49584252e-01
-9.39414382e-01 -8.67135108e-01 2.05197632e-01 -6.96028918e-02
-1.73053360e+00 -2.75290519e-01 1.08207440e+00 -3.91681880e-01
8.62317920e-01 2.13308334e-01 8.17442536e-01 4.83878762e-01
-1.71848565e-01 6.23300135e-01 6.99551642e-01 1.00393202e-02
1.71916157e-01 -4.45817590e-01 -1.71561092e-01 5.94605982e-01
2.74621904e-01 3.15185040e-01 -1.82732895e-01 -3.06506306e-01
1.33943093e+00 5.61878145e-01 -2.13613003e-01 -5.95336676e-01
-1.24973357e+00 6.20197356e-01 8.33882391e-01 7.28757605e-02
-6.46342039e-01 1.87698290e-01 4.95872498e-02 -6.42744750e-02
5.19119024e-01 -5.99999242e-02 -6.49032533e-01 1.31195545e-01
-9.43660378e-01 4.23559040e-01 5.50861716e-01 1.28093255e+00
1.09247136e+00 -1.44401163e-01 1.53181497e-02 8.82464707e-01
5.89404643e-01 7.72966981e-01 -7.04915524e-02 -1.10291958e+00
5.56860089e-01 9.78497386e-01 2.57499684e-02 -1.15813661e+00
-3.32077235e-01 -4.76846874e-01 -1.09475601e+00 3.01805794e-01
-1.16581805e-01 1.94442928e-01 -1.03910434e+00 1.10617900e+00
6.68572724e-01 5.24585366e-01 -4.19718742e-01 1.07294905e+00
1.20193505e+00 7.44628012e-01 -2.71747619e-01 -9.73529294e-02
1.01753950e+00 -8.13674390e-01 -2.69880563e-01 -1.06219545e-01
1.88893422e-01 -9.46228027e-01 9.02131617e-01 2.56714553e-01
-1.06482899e+00 -6.63179040e-01 -7.97801673e-01 -5.19240141e-01
1.26232356e-01 5.66113591e-01 6.65715575e-01 -1.86954081e-01
-8.47864687e-01 6.38799369e-01 -9.89904106e-01 1.04577385e-01
8.22027206e-01 3.49248379e-01 -3.79317939e-01 -4.44667667e-01
-4.42567766e-01 4.37785000e-01 -1.55508118e-02 4.83203828e-01
-7.42783606e-01 -1.04137993e+00 -7.39252806e-01 1.25129402e-01
2.43253380e-01 -9.79963303e-01 9.80696678e-01 -2.39635929e-01
-1.37189627e+00 8.38731766e-01 -1.90659374e-01 -6.38565347e-02
5.59400439e-01 -2.48712763e-01 1.40925333e-01 -7.49116838e-02
1.08030394e-01 7.70864487e-01 9.57589149e-01 -1.35118806e+00
-5.11131346e-01 -6.19378924e-01 -2.24538580e-01 2.27444187e-01
8.44821185e-02 -2.96603858e-01 -8.25125754e-01 -6.73522592e-01
8.49504054e-01 -9.01313901e-01 -4.31510329e-01 4.29828644e-01
-1.15845688e-01 -5.29572785e-01 1.06068802e+00 -5.40486097e-01
8.86690021e-01 -2.30596900e+00 1.93670601e-01 2.56685764e-01
4.39461559e-01 1.78242028e-01 -3.62173289e-01 2.24962264e-01
8.26220494e-03 -1.41506925e-01 -3.33147049e-01 -7.74239898e-01
-7.57814348e-02 2.94138253e-01 -4.17217195e-01 5.02450824e-01
1.45878538e-01 1.00683415e+00 -6.23327315e-01 -3.54374141e-01
4.88923758e-01 6.69435382e-01 -5.30234396e-01 2.09442988e-01
-2.89636612e-01 3.66683900e-01 -7.40481675e-01 8.90929580e-01
1.43003905e+00 -2.59350568e-01 -4.89181668e-01 -4.68253285e-01
-5.33609927e-01 -3.17533202e-02 -1.30011940e+00 2.22558260e+00
-3.38626653e-01 -3.13818790e-02 3.67428631e-01 -4.49393630e-01
1.35075080e+00 -4.51711342e-02 6.37378812e-01 -2.13131413e-01
3.20660835e-03 3.89077872e-01 -3.68353784e-01 -2.00624973e-01
3.51752937e-01 -1.08060548e-02 3.38440120e-01 8.40493441e-02
-3.26578736e-01 -5.80325723e-01 -3.55489135e-01 2.36943946e-03
8.48903537e-01 4.36024249e-01 -6.38854206e-02 -5.96893532e-03
6.95676088e-01 8.71201083e-02 9.18333590e-01 1.24739818e-01
1.58077136e-01 9.32017744e-01 -3.53370374e-03 -7.76623726e-01
-1.12446821e+00 -9.96683240e-01 6.97315298e-03 4.70152199e-01
3.77450436e-01 -3.91139537e-01 -3.63024443e-01 -5.05235851e-01
2.94265389e-01 2.36664832e-01 -2.08588138e-01 6.40162900e-02
-7.27663457e-01 -6.33964241e-02 -3.63746174e-02 6.44604743e-01
5.29312909e-01 -1.04244637e+00 -3.01645547e-01 -2.64207702e-02
-6.55458868e-03 -1.08069503e+00 -6.26489937e-01 -4.22381848e-01
-1.35359752e+00 -1.03363478e+00 -7.24307418e-01 -8.30363572e-01
9.21894789e-01 6.67555571e-01 9.50030625e-01 3.15077782e-01
-9.50367078e-02 8.71440172e-02 -2.35540286e-01 -3.82894665e-01
1.32523537e-01 1.53144017e-01 -1.33757994e-01 -1.42622381e-01
2.65504569e-01 -9.75794256e-01 -6.06208324e-01 2.26741537e-01
-8.51556838e-01 3.23193222e-01 8.93805146e-01 6.14630461e-01
1.11824882e+00 -1.25806272e-01 8.12671632e-02 -4.67053294e-01
2.97021121e-01 -2.13545799e-01 -8.47982109e-01 1.09410174e-01
-3.13766629e-01 -1.66041344e-01 4.85094368e-01 -2.48699188e-01
-8.58755112e-01 5.65154612e-01 -4.22385097e-01 -1.34925461e+00
-1.67096168e-01 4.20681357e-01 -3.11534554e-01 -3.41352195e-01
3.56615871e-01 3.66960317e-01 1.47970691e-01 -9.31002736e-01
3.40391785e-01 4.64045376e-01 5.37025392e-01 -3.70824605e-01
1.20309246e+00 7.76756585e-01 1.95590571e-01 -7.13350177e-01
-7.12635994e-01 -5.67311585e-01 -8.82193446e-01 -8.14962536e-02
5.19195616e-01 -1.16183162e+00 -7.81996787e-01 4.98571843e-01
-1.57456517e+00 -2.74125338e-02 -5.00824392e-01 4.25204009e-01
-5.33841252e-01 3.71855080e-01 -5.74019074e-01 -4.79582310e-01
-6.54166281e-01 -9.89906609e-01 1.44871223e+00 1.46995366e-01
4.47105199e-01 -4.57832843e-01 8.24118331e-02 1.27532765e-01
1.89626873e-01 3.17974150e-01 4.97498453e-01 -4.25125211e-02
-1.17557740e+00 -1.17167644e-01 -5.51295280e-01 2.31409580e-01
6.60414174e-02 8.42295960e-02 -7.99498916e-01 -4.53244865e-01
2.07860649e-01 -1.87372137e-02 8.72594535e-01 4.56067234e-01
1.29978514e+00 -9.72300619e-02 -4.51578736e-01 1.12716663e+00
1.51040328e+00 -2.47394666e-01 6.70477629e-01 1.24129402e-02
1.01890934e+00 3.44693810e-01 7.47566938e-01 5.87898672e-01
4.78216022e-01 4.72468197e-01 7.39105880e-01 -9.68349651e-02
-2.97716588e-01 -5.83349824e-01 -2.77163312e-02 1.11675560e+00
-2.71526784e-01 5.89158297e-01 -1.03609109e+00 4.44417119e-01
-1.88420880e+00 -6.76829278e-01 -3.77525121e-01 2.09290910e+00
4.68372345e-01 -1.16698584e-02 -2.54700840e-01 -7.15939477e-02
6.76545382e-01 9.96063873e-02 -7.05335021e-01 3.56449425e-01
1.15368336e-01 1.29767537e-01 2.37109810e-01 4.81392711e-01
-8.54561746e-01 1.01476276e+00 4.97731590e+00 1.06840241e+00
-1.06477261e+00 7.54751861e-02 2.07441658e-01 7.09064752e-02
-4.91074055e-01 1.68858439e-01 -7.71812856e-01 2.82993346e-01
-2.56915186e-02 8.00664946e-02 4.56093580e-01 1.07150435e+00
1.35910913e-01 2.82005280e-01 -7.72843957e-01 1.48772693e+00
-6.44927174e-02 -1.54550064e+00 1.30249947e-01 8.66756514e-02
7.79660583e-01 5.54279268e-01 -2.41674349e-01 1.59132704e-01
8.85841921e-02 -8.45909357e-01 6.54876411e-01 9.09924567e-01
1.06403220e+00 -8.79215956e-01 4.98848319e-01 8.44320536e-01
-1.63834059e+00 4.74232435e-03 -9.43947434e-01 8.44476931e-03
1.67231023e-01 1.04171884e+00 -5.15917540e-01 8.13005865e-01
7.85408020e-01 1.08290517e+00 -3.65452111e-01 1.12447155e+00
-1.99853554e-01 -4.87532206e-02 -6.56441689e-01 2.77576029e-01
4.54588607e-02 -4.76912290e-01 7.29957461e-01 7.69028425e-01
5.09803176e-01 3.99377972e-01 5.20467639e-01 1.10531998e+00
-6.84895217e-02 -1.17254019e-01 -4.82494295e-01 3.92897159e-01
8.30280006e-01 1.62436020e+00 -5.85604668e-01 -1.12266824e-01
-2.51906514e-01 6.75561011e-01 4.48533863e-01 1.24564208e-01
-4.50617611e-01 -1.61423519e-01 5.98977447e-01 3.04785907e-01
4.29656029e-01 -6.39334977e-01 -4.91853237e-01 -1.23876429e+00
3.02047402e-01 -4.26493138e-01 -9.39917937e-03 -9.17461276e-01
-1.44203925e+00 5.30881822e-01 -2.00455815e-01 -1.63887203e+00
2.52112925e-01 -2.88758725e-01 -6.31891012e-01 1.09761322e+00
-1.75662529e+00 -1.62334061e+00 -8.93955469e-01 7.01141775e-01
6.39138460e-01 1.93407033e-02 3.95290077e-01 2.95751691e-01
-1.58298999e-01 -6.88269269e-03 -2.76666611e-01 5.12983240e-02
3.07012111e-01 -6.92340612e-01 7.53630102e-01 6.23721182e-01
-6.25133514e-02 5.95396280e-01 3.78666185e-02 -9.93810713e-01
-1.63329709e+00 -1.39581370e+00 6.61577761e-01 -1.63822755e-01
1.24680638e-01 -2.18325838e-01 -9.62359786e-01 4.73099828e-01
-5.68640649e-01 5.67754745e-01 -5.78778237e-02 -1.87729463e-01
-2.50879824e-01 -3.06256026e-01 -1.11645520e+00 3.07657450e-01
1.38967776e+00 -3.46338481e-01 -4.79052573e-01 2.03886762e-01
1.01616275e+00 -7.36329198e-01 -8.61788034e-01 7.05286443e-01
4.36569273e-01 -9.00148928e-01 1.19879866e+00 -4.80168313e-02
3.78418475e-01 -7.65413284e-01 -8.38996992e-02 -1.08195603e+00
-7.37638354e-01 -4.33919400e-01 -3.43544364e-01 1.16314197e+00
-7.59714171e-02 -4.17399108e-01 1.00124764e+00 3.39837432e-01
-5.51574171e-01 -1.09771812e+00 -9.37528491e-01 -6.03495300e-01
-6.55925646e-02 -5.60070336e-01 1.13340259e+00 7.89888382e-01
-7.16334581e-01 1.04419589e-01 -1.00863762e-02 3.72934163e-01
8.09050500e-01 5.48827469e-01 1.01463974e+00 -1.71885478e+00
2.69855976e-01 -2.70517319e-01 -1.88367844e-01 -1.50485635e+00
7.42368549e-02 -9.35227036e-01 -8.36530793e-03 -1.80486679e+00
5.87681904e-02 -1.02931428e+00 2.57901847e-01 4.09538925e-01
9.63107049e-02 1.84586138e-01 4.51640636e-01 7.17571974e-01
-2.64516294e-01 9.19326365e-01 1.54701161e+00 -7.05048954e-03
-3.78641456e-01 2.05586731e-01 -3.35666329e-01 9.88960266e-01
6.47358060e-01 -3.73012245e-01 -2.50885308e-01 -8.22950780e-01
1.61841065e-01 4.92189191e-02 5.26569128e-01 -9.77261424e-01
6.73506200e-01 -1.76939711e-01 5.61168849e-01 -1.55326891e+00
5.82026005e-01 -1.12539983e+00 4.01275039e-01 3.86436939e-01
2.73783267e-01 1.11023553e-01 8.83014873e-02 5.53204596e-01
-1.35742038e-01 2.64838971e-02 6.75760150e-01 -2.89905369e-01
-4.54435706e-01 1.14301610e+00 6.92576647e-01 -3.58784199e-01
8.95651162e-01 -3.01615685e-01 -2.73361336e-03 8.67901966e-02
-6.08558774e-01 4.51390773e-01 7.22674370e-01 4.48771268e-01
1.33205235e+00 -1.65302467e+00 -8.84286344e-01 6.58294797e-01
-3.48697379e-02 9.78553653e-01 5.62779486e-01 7.16168761e-01
-7.85048544e-01 2.33665302e-01 6.25021607e-02 -1.09631991e+00
-1.15493309e+00 4.79498118e-01 2.37661824e-01 -1.21840239e-01
-9.08728063e-01 8.25301111e-01 2.17875898e-01 -9.06529367e-01
2.01817602e-01 -4.32930261e-01 -4.19289656e-02 -2.39685535e-01
3.06464493e-01 5.70617020e-01 2.33279884e-01 -4.72981095e-01
-3.36585402e-01 1.16769981e+00 5.60855456e-02 1.79464012e-01
1.62862372e+00 4.53274883e-02 -4.82615888e-01 3.31091098e-02
1.06031585e+00 -9.41547041e-05 -1.58722425e+00 -3.67273450e-01
-3.14844102e-01 -8.39470744e-01 1.72954112e-01 -3.96294922e-01
-1.20750475e+00 8.74941230e-01 3.38574052e-01 -2.35657990e-01
1.10920990e+00 4.04202472e-03 1.05009913e+00 1.94104418e-01
7.19576180e-01 -7.53182113e-01 -2.27357388e-01 7.16415703e-01
1.44314361e+00 -8.09199989e-01 2.74109006e-01 -8.63162279e-01
-2.37704754e-01 1.17033195e+00 6.10873878e-01 -5.42360723e-01
9.51268435e-01 2.53711641e-01 -2.72884846e-01 -3.70097816e-01
-6.75491333e-01 -3.57927307e-02 3.92339349e-01 6.00316405e-01
-1.71206251e-01 -5.10990433e-02 -1.79494143e-01 5.34762800e-01
-1.98503822e-01 2.98381120e-01 3.80433425e-02 6.98079228e-01
-4.82589364e-01 -9.11185980e-01 -5.57220995e-01 4.65047151e-01
1.87202081e-01 1.46647155e-01 -2.43953526e-01 5.06278396e-01
3.95207405e-01 4.73698556e-01 2.36714512e-01 -4.73454922e-01
6.06631100e-01 -4.75014240e-01 2.70377696e-01 -7.10120499e-01
-4.83257502e-01 3.37066293e-01 -5.46940207e-01 -7.81400979e-01
-3.62152994e-01 -7.18629181e-01 -1.35001838e+00 -3.85597825e-01
-3.85663569e-01 -1.42254541e-02 7.27630675e-01 6.45094633e-01
7.87628055e-01 2.59823233e-01 9.60179865e-01 -1.34945762e+00
-5.90115488e-01 -9.89918470e-01 -4.64783639e-01 1.98461786e-01
2.80128211e-01 -6.46621048e-01 -2.44584382e-01 -2.06986189e-01] | [8.319242477416992, -3.5721988677978516] |
f8754c65-020e-4085-9c49-125f61f65f82 | layoutnet-reconstructing-the-3d-room-layout | 1803.08999 | null | http://arxiv.org/abs/1803.08999v1 | http://arxiv.org/pdf/1803.08999v1.pdf | LayoutNet: Reconstructing the 3D Room Layout from a Single RGB Image | We propose an algorithm to predict room layout from a single image that
generalizes across panoramas and perspective images, cuboid layouts and more
general layouts (e.g. L-shape room). Our method operates directly on the
panoramic image, rather than decomposing into perspective images as do recent
works. Our network architecture is similar to that of RoomNet, but we show
improvements due to aligning the image based on vanishing points, predicting
multiple layout elements (corners, boundaries, size and translation), and
fitting a constrained Manhattan layout to the resulting predictions. Our method
compares well in speed and accuracy to other existing work on panoramas,
achieves among the best accuracy for perspective images, and can handle both
cuboid-shaped and more general Manhattan layouts. | ['Qi Shan', 'Derek Hoiem', 'Chuhang Zou', 'Alex Colburn'] | 2018-03-23 | layoutnet-reconstructing-the-3d-room-layout-1 | http://openaccess.thecvf.com/content_cvpr_2018/html/Zou_LayoutNet_Reconstructing_the_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Zou_LayoutNet_Reconstructing_the_CVPR_2018_paper.pdf | cvpr-2018-6 | ['3d-room-layouts-from-a-single-rgb-panorama'] | ['computer-vision'] | [ 3.03714752e-01 -5.36310039e-02 9.10976231e-02 -4.49832112e-01
-2.89268523e-01 -1.01411414e+00 7.07342207e-01 -9.59439129e-02
5.81120923e-02 2.82657862e-01 6.55668795e-01 -7.27290392e-01
-1.66202739e-01 -7.37554848e-01 -9.58094180e-01 -3.94663274e-01
-2.29919806e-01 5.56602597e-01 1.48419783e-01 -6.00042976e-02
4.83728588e-01 8.38076591e-01 -1.32440925e+00 3.57700765e-01
3.79872352e-01 7.40879416e-01 2.34411567e-01 1.11124885e+00
2.21448362e-01 6.75001919e-01 -2.82332152e-01 1.48301437e-01
5.61359584e-01 -5.89934029e-02 -6.81968093e-01 3.56108367e-01
1.58293915e+00 -4.31542546e-01 -5.48772097e-01 3.77553344e-01
2.50704557e-01 2.42642611e-01 5.26796043e-01 -9.72493410e-01
-7.67074764e-01 2.86250561e-01 -5.65727651e-01 -2.75340348e-01
4.29025918e-01 -1.67729616e-01 1.13555503e+00 -9.04448748e-01
6.50321424e-01 1.50578690e+00 7.89642155e-01 2.26560190e-01
-1.27372444e+00 -3.09539795e-01 5.69393933e-01 -2.10807294e-01
-1.14403498e+00 -4.64525372e-01 6.90738320e-01 -1.88898116e-01
9.25439298e-01 5.73067486e-01 7.25881934e-01 9.09171343e-01
2.74678916e-01 7.27364063e-01 1.33109188e+00 -5.70374370e-01
2.53216803e-01 -2.38083676e-01 -1.91973239e-01 9.71475542e-01
1.32121995e-01 -3.66493315e-01 4.28087004e-02 5.28739719e-03
1.23218012e+00 2.63227463e-01 -2.73598790e-01 -1.16332948e+00
-1.46015608e+00 3.12591374e-01 7.24664211e-01 -2.73397546e-02
7.23870993e-02 2.28990227e-01 -2.27510054e-02 -2.31714740e-01
2.51007304e-02 9.04786289e-01 -3.57368052e-01 1.21833868e-01
-1.03027809e+00 4.28196698e-01 7.93096781e-01 1.51462662e+00
6.89895928e-01 -1.49578974e-01 7.14526549e-02 8.29142869e-01
-6.87111169e-02 5.15910745e-01 -1.74483702e-01 -1.50505042e+00
6.85252428e-01 5.26566744e-01 1.41479909e-01 -1.01524186e+00
-9.24510241e-01 5.62302768e-02 -8.11951339e-01 1.47511452e-01
5.91865659e-01 2.50399485e-02 -1.25352669e+00 1.37122619e+00
-1.24949522e-01 -1.29952207e-01 -3.67499113e-01 7.81401873e-01
5.51681399e-01 7.75130510e-01 -5.09196043e-01 4.03429627e-01
1.31544888e+00 -1.61629319e+00 -2.54971802e-01 -7.39170909e-01
2.57717043e-01 -9.67471242e-01 1.46367538e+00 5.25347710e-01
-1.05042922e+00 -4.63114947e-01 -1.03680241e+00 -5.58823407e-01
-3.72360051e-01 2.11350694e-01 5.79655647e-01 7.18678832e-01
-1.46786964e+00 3.59341204e-01 -4.50597167e-01 -6.51942015e-01
1.02291971e-01 4.26887989e-01 -4.44251448e-01 -3.84972513e-01
-2.76121676e-01 7.75836110e-01 6.70683235e-02 1.20470732e-01
-7.40874112e-01 -4.38157886e-01 -1.19001830e+00 3.56753111e-01
4.00026262e-01 -7.28487790e-01 1.09415424e+00 -7.53243446e-01
-1.30988574e+00 5.27326047e-01 -2.57627189e-01 -1.05704375e-01
3.26719016e-01 -1.93157032e-01 -2.25674003e-01 -3.01377147e-01
-2.65575260e-01 9.75839734e-01 3.82997841e-01 -1.63402462e+00
-2.92157441e-01 -2.80493855e-01 3.68884623e-01 6.37954354e-01
-1.94743171e-01 -6.13269389e-01 -5.31237483e-01 -8.42416752e-03
4.09898818e-01 -1.24410307e+00 -5.27643323e-01 3.65786515e-02
-9.18278754e-01 2.88875669e-01 8.03441286e-01 -5.98301888e-01
1.15022469e+00 -1.95497918e+00 -9.43185911e-02 7.69151330e-01
1.37888283e-01 -3.65622967e-01 -1.59352809e-01 5.26073158e-01
-4.68567461e-02 3.76736186e-02 -1.28559783e-01 -3.86519283e-01
1.04099900e-01 1.73110932e-01 -2.23212570e-01 3.27980757e-01
-3.53827745e-01 7.35007226e-01 -7.42806435e-01 -3.66731077e-01
6.08461261e-01 3.21945667e-01 -6.56763613e-01 -6.99836090e-02
-1.06201209e-01 9.42276493e-02 4.11933623e-02 6.29692912e-01
7.13404298e-01 -3.31411004e-01 6.10615253e-01 2.93124020e-02
-4.33201075e-01 4.36516047e-01 -1.07313645e+00 1.78695643e+00
-9.07142103e-01 1.16079986e+00 1.92717582e-01 -2.79670328e-01
1.03998971e+00 -1.14402175e-01 2.46129498e-01 -5.06408095e-01
-1.47052899e-01 -6.06242493e-02 -1.90296069e-01 -2.80280411e-01
1.08271348e+00 4.37795132e-01 -2.70931005e-01 2.79626995e-01
-3.65992278e-01 -5.72674155e-01 9.46399644e-02 7.62686357e-02
9.74119067e-01 3.78043622e-01 2.79131293e-01 -4.29631770e-01
8.26756954e-02 1.07772369e-02 -4.79976274e-02 8.61346900e-01
2.35902041e-01 1.11867857e+00 5.87680876e-01 -1.02324080e+00
-1.88636816e+00 -1.14744544e+00 6.67143017e-02 1.10333109e+00
3.06226552e-01 -6.53460681e-01 -8.73697996e-01 -4.21966404e-01
-2.61145979e-01 6.68115497e-01 -6.34048581e-01 4.48231071e-01
-1.11497140e+00 -1.56870171e-01 2.43582755e-01 7.34445333e-01
3.68390381e-01 -9.25785065e-01 -7.43167460e-01 -2.57619560e-01
-1.23739056e-01 -1.23089981e+00 -8.22956622e-01 4.11778420e-01
-8.51842165e-01 -1.06708467e+00 -4.42249298e-01 -1.10362315e+00
9.67202246e-01 7.98472822e-01 1.38519979e+00 -2.11473182e-01
-2.38204911e-01 4.43354875e-01 2.90538847e-01 -1.27913147e-01
1.10602982e-01 1.96238652e-01 -1.70367777e-01 -6.25647187e-01
-9.38468128e-02 -5.04951179e-01 -8.58624876e-01 3.13800961e-01
-5.38340986e-01 5.44429123e-01 4.84900862e-01 5.53513944e-01
4.61566448e-01 6.89348066e-03 -5.96946895e-01 -9.73432779e-01
3.50094736e-01 -9.93442982e-02 -8.14744532e-01 2.45624125e-01
-5.39625049e-01 -1.67632595e-01 1.13457000e+00 -1.03406549e-01
-9.23325717e-01 3.17779958e-01 2.49502629e-01 -3.55581433e-01
-5.78702509e-01 -3.14575285e-01 -2.37291411e-01 4.58885059e-02
6.70827985e-01 3.88600561e-03 -3.39075893e-01 -1.30790919e-01
8.39023232e-01 2.92918354e-01 4.72400576e-01 -6.53346419e-01
8.89761925e-01 7.84965515e-01 2.65774995e-01 -1.14398932e+00
-6.80655837e-01 -2.93244064e-01 -1.21460450e+00 -3.86869498e-02
8.65116775e-01 -7.35094488e-01 -7.37539411e-01 4.10727412e-02
-1.04384041e+00 -6.46231651e-01 -3.08203790e-02 1.14231229e-01
-6.45290017e-01 1.16487235e-01 -5.38373649e-01 -7.68968105e-01
-6.38506562e-02 -8.84923637e-01 1.25160289e+00 2.36223832e-01
-3.84087056e-01 -1.24004734e+00 1.77055478e-01 2.84702539e-01
3.07172030e-01 1.83834702e-01 1.15033841e+00 -1.02190554e-01
-1.27026892e+00 -5.14762327e-02 -2.30188772e-01 -9.02317837e-02
-5.90030402e-02 9.70692039e-02 -9.58937109e-01 -1.24094963e-01
-5.86848497e-01 -1.15861662e-01 7.83139288e-01 8.13979268e-01
1.57198656e+00 -4.87930059e-01 -4.89400804e-01 8.58097732e-01
1.37146854e+00 2.31535941e-01 8.53279650e-01 7.50675738e-01
1.10668683e+00 7.06467509e-01 2.64155120e-01 1.32059269e-02
7.41910219e-01 3.48431319e-01 5.50736189e-01 -4.26731706e-01
1.10952422e-01 -6.87229514e-01 1.85058296e-01 5.20351648e-01
-5.40834740e-02 -4.31107461e-01 -9.35366213e-01 6.27439618e-01
-1.70720732e+00 -9.44945097e-01 -5.94487377e-02 2.20643997e+00
1.35991005e-02 1.30885527e-01 5.28340451e-02 -3.40742499e-01
4.98338014e-01 7.22492039e-01 -4.02532101e-01 -9.66025114e-01
-1.38088122e-01 8.51256773e-02 1.01196265e+00 6.78605497e-01
-1.23156679e+00 1.09487724e+00 8.03178215e+00 2.78530926e-01
-9.18244123e-01 -4.41230208e-01 9.80575204e-01 -2.38777444e-01
-4.95586216e-01 1.85660735e-01 -6.18851721e-01 -2.19933301e-01
4.94042665e-01 7.75743246e-01 1.01613712e+00 1.14064682e+00
7.20690042e-02 -1.15969166e-01 -1.22629023e+00 9.56606627e-01
5.10219693e-01 -1.50344884e+00 -2.14032888e-01 2.64006346e-01
1.13945961e+00 3.81398648e-02 2.63127059e-01 4.17508557e-02
5.48833966e-01 -1.38385201e+00 7.27605760e-01 1.27197549e-01
7.87724197e-01 -5.74624717e-01 1.42196730e-01 2.08037212e-01
-1.32247138e+00 -3.64542186e-01 -5.20677865e-01 -3.62238765e-01
-1.64208919e-01 -4.36311923e-02 -1.20991802e+00 1.25456467e-01
8.58327270e-01 4.61806238e-01 -7.35547960e-01 8.22170556e-01
-6.10397309e-02 8.22118521e-02 -5.07268667e-01 -4.35618162e-02
5.43757915e-01 -5.06115496e-01 1.01272359e-01 1.17571580e+00
4.49184597e-01 -4.27422792e-01 2.82066345e-01 7.80025005e-01
-8.89512070e-04 -8.65851715e-02 -1.23747110e+00 5.03377676e-01
6.18687749e-01 1.58861518e+00 -1.14840806e+00 -3.27388912e-01
-2.66820163e-01 7.97872245e-01 3.14923823e-01 5.05198479e-01
-5.68301260e-01 -2.28483200e-01 2.02333778e-01 5.41749179e-01
4.78202343e-01 -6.37253642e-01 -7.80102730e-01 -9.49128211e-01
-1.34247005e-01 -7.40694463e-01 -2.89397296e-02 -1.10685301e+00
-6.22571528e-01 5.85777342e-01 7.01832324e-02 -1.09245801e+00
6.48874566e-02 -1.07560503e+00 -7.86898255e-01 3.35790962e-01
-1.18712044e+00 -1.51831532e+00 -6.35065198e-01 2.48098612e-01
5.89716852e-01 3.00719082e-01 9.44315076e-01 -5.87164834e-02
-2.83602268e-01 1.83532923e-01 3.45151514e-01 8.83303732e-02
8.55599642e-01 -1.72385502e+00 9.68340218e-01 6.95910037e-01
2.82721579e-01 9.17896926e-01 6.45726144e-01 -1.54328108e-01
-1.30923223e+00 -1.02437294e+00 5.16590416e-01 -8.84908736e-01
3.11410993e-01 -1.12065101e+00 -2.74676770e-01 1.20361865e+00
5.95952451e-01 -2.24380076e-01 3.68186295e-01 6.33840501e-01
-6.09872758e-01 -1.88125327e-01 -6.18595421e-01 1.14456642e+00
1.31538630e+00 -4.16207194e-01 -6.42341971e-02 3.55115116e-01
4.03173447e-01 -8.77347648e-01 -5.61177671e-01 -3.19727287e-02
9.84716296e-01 -1.23327732e+00 1.27579606e+00 -1.34238645e-01
6.62860215e-01 -3.44052404e-01 -2.45880544e-01 -1.43639648e+00
-7.26042926e-01 -4.06010747e-01 1.83265015e-01 4.51393127e-01
4.55376565e-01 -7.72576854e-02 9.86568570e-01 5.15229583e-01
-4.79488701e-01 -8.10566306e-01 -3.80097240e-01 -4.19444025e-01
-4.37465571e-02 -2.51602605e-02 7.66940951e-01 7.59669960e-01
-7.39862248e-02 4.61586237e-01 -7.62489259e-01 1.41397446e-01
4.48878020e-01 5.14546931e-01 1.30555165e+00 -9.67599988e-01
-4.46440391e-02 -4.84241635e-01 9.85815600e-02 -1.57499552e+00
-3.78849320e-02 -6.33612394e-01 2.36768022e-01 -1.87309408e+00
1.24053648e-02 -5.79359531e-01 3.33041959e-02 3.53805512e-01
2.53480464e-01 3.58994871e-01 2.14130908e-01 -9.12563577e-02
-8.37878525e-01 1.82702988e-01 1.47171462e+00 1.62177235e-02
-3.63748372e-01 -2.81538129e-01 -5.69736302e-01 9.52110469e-01
7.78379500e-01 2.24898219e-01 -4.25327539e-01 -5.89822710e-01
3.67444962e-01 -1.15058504e-01 2.87658781e-01 -1.14045155e+00
2.08808199e-01 -2.83854842e-01 1.12762165e+00 -9.28875983e-01
7.37567127e-01 -9.07187700e-01 -8.96075293e-02 4.62356992e-02
-4.15426016e-01 6.26681030e-01 2.15493858e-01 3.65461946e-01
2.84731746e-01 4.09021527e-01 4.40035850e-01 -4.75298613e-01
-6.61280751e-01 1.51450902e-01 -3.00019979e-01 -3.28396738e-01
7.92425871e-01 -6.81537926e-01 -9.17949796e-01 -5.81609190e-01
-3.97251695e-01 3.39809686e-01 1.16686225e+00 6.67578280e-01
7.61721969e-01 -1.23796082e+00 8.89617503e-02 4.63998288e-01
8.41904432e-02 4.19930518e-01 2.27922156e-01 5.13851404e-01
-1.22450185e+00 7.40638554e-01 -3.45646441e-01 -7.16329992e-01
-1.27022994e+00 6.37648582e-01 3.00526083e-01 -1.67195186e-01
-6.91015303e-01 4.86867756e-01 6.81110382e-01 -9.90776062e-01
2.02778608e-01 -6.71597838e-01 -4.48025279e-02 -5.32568097e-01
1.70512661e-01 4.04399782e-01 -3.14923525e-01 -5.04869998e-01
-2.09744900e-01 8.59920204e-01 7.94587955e-02 -5.91632687e-02
1.22134256e+00 -2.98780411e-01 -3.15942764e-01 7.56447375e-01
8.90723169e-01 4.29288536e-01 -1.60456681e+00 1.73137024e-01
-2.79648632e-01 -5.09322584e-01 -3.47289443e-01 -6.29876792e-01
-4.38419282e-01 1.04916072e+00 4.23703164e-01 6.93825856e-02
7.43225098e-01 -1.66876435e-01 5.65202475e-01 6.36167467e-01
2.39597693e-01 -1.16411734e+00 1.85010210e-01 7.46590853e-01
6.26462042e-01 -8.48393679e-01 2.52496123e-01 -2.54722357e-01
-4.58431810e-01 1.39506447e+00 1.00709748e+00 -4.63854104e-01
4.79509085e-01 4.67828423e-01 8.75980183e-02 -2.49601781e-01
-5.70957303e-01 4.08026814e-01 4.21594113e-01 4.37446415e-01
3.91538680e-01 4.21290874e-01 7.18804359e-01 -4.60187405e-01
-6.82232320e-01 -6.73670173e-01 6.23350084e-01 7.35054731e-01
-5.97522855e-01 -7.57270277e-01 -8.80019784e-01 4.21443671e-01
1.40597383e-02 -2.99602151e-01 -4.43660736e-01 1.04218662e+00
1.87411889e-01 5.74546278e-01 6.68532610e-01 -2.85611212e-01
4.30750489e-01 -2.12147802e-01 7.59865761e-01 -5.11399031e-01
-2.34305352e-01 3.57445270e-01 1.79397509e-01 -6.70524299e-01
-7.66870603e-02 -3.79269481e-01 -8.17867756e-01 -6.13481939e-01
-6.79361671e-02 -2.78849214e-01 7.38811672e-01 5.85595191e-01
3.44600797e-01 5.45942008e-01 4.54252392e-01 -1.23645115e+00
2.52487123e-01 -7.81796157e-01 -3.85026544e-01 2.33566537e-01
5.41948736e-01 -1.48980975e-01 1.39137299e-03 -8.33542868e-02] | [8.770343780517578, -2.839613676071167] |
b0ffb43c-eb75-4c78-b08e-50d8d55466cc | learning-semantic-representations-in-a-bigram | null | null | https://aclanthology.org/W13-0210 | https://aclanthology.org/W13-0210.pdf | Learning Semantic Representations in a Bigram Language Model | null | ['Jeff Mitchell'] | 2013-03-01 | null | null | null | ws-2013-3 | ['learning-semantic-representations'] | ['methodology'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.295959949493408, 3.663486957550049] |
e223e0dd-fb0f-404c-b9e1-cac19198b502 | aasist-audio-anti-spoofing-using-integrated | 2110.01200 | null | https://arxiv.org/abs/2110.01200v1 | https://arxiv.org/pdf/2110.01200v1.pdf | AASIST: Audio Anti-Spoofing using Integrated Spectro-Temporal Graph Attention Networks | Artefacts that differentiate spoofed from bona-fide utterances can reside in spectral or temporal domains. Their reliable detection usually depends upon computationally demanding ensemble systems where each subsystem is tuned to some specific artefacts. We seek to develop an efficient, single system that can detect a broad range of different spoofing attacks without score-level ensembles. We propose a novel heterogeneous stacking graph attention layer which models artefacts spanning heterogeneous temporal and spectral domains with a heterogeneous attention mechanism and a stack node. With a new max graph operation that involves a competitive mechanism and an extended readout scheme, our approach, named AASIST, outperforms the current state-of-the-art by 20% relative. Even a lightweight variant, AASIST-L, with only 85K parameters, outperforms all competing systems. | ['Nicholas Evans', 'Ha-Jin Yu', 'Bong-Jin Lee', 'Joon Son Chung', 'Hye-jin Shim', 'Hemlata Tak', 'Hee-Soo Heo', 'Jee-weon Jung'] | 2021-10-04 | null | null | null | null | ['voice-anti-spoofing'] | ['audio'] | [ 4.67651844e-01 -2.86444008e-01 2.68015236e-01 -8.98001790e-02
-1.11035228e+00 -7.06817389e-01 7.52139866e-01 6.95373416e-02
-1.74242944e-01 2.08625108e-01 -4.98632193e-02 -6.48302913e-01
-5.30975778e-03 -2.07725212e-01 -5.86189806e-01 -5.85550010e-01
-3.11861992e-01 2.87721157e-01 7.70088971e-01 -4.96973991e-01
8.12924951e-02 2.24997371e-01 -9.65682983e-01 4.34199959e-01
7.63832629e-01 9.24487710e-01 -1.62984446e-01 1.26863968e+00
2.76972741e-01 8.87638271e-01 -1.06126499e+00 -8.41633797e-01
-9.93888155e-02 -4.54227209e-01 -5.20743072e-01 -3.43835920e-01
5.45098901e-01 -5.18670380e-02 -4.01637614e-01 1.17728174e+00
9.02429581e-01 -1.26225278e-01 4.17629808e-01 -1.07388532e+00
-5.95262945e-01 9.26838756e-01 -5.51775455e-01 5.42352438e-01
4.55159843e-01 3.78027976e-01 1.00166965e+00 -7.82603264e-01
2.72039711e-01 1.26071393e+00 9.75504994e-01 4.30917501e-01
-1.41862142e+00 -7.28562176e-01 5.37519567e-02 3.56774330e-01
-1.24773335e+00 -7.68601775e-01 9.00824189e-01 -1.90881491e-01
1.48075807e+00 5.38415492e-01 1.61553770e-01 1.91832662e+00
8.40087458e-02 4.61970031e-01 9.66416955e-01 -1.25356168e-01
4.36825529e-02 1.78132784e-02 4.23129141e-01 7.22774148e-01
4.75347221e-01 1.83862761e-01 -7.60571778e-01 -4.28328902e-01
-1.21193398e-02 -3.70471954e-01 -2.60048747e-01 -9.08660218e-02
-9.84851837e-01 7.88254976e-01 3.71033669e-01 4.04757440e-01
-1.54815227e-01 3.81068170e-01 7.61497438e-01 5.45955062e-01
3.52588356e-01 6.94132507e-01 -2.66417742e-01 3.74255329e-02
-1.10716271e+00 -2.07322598e-01 8.30480754e-01 7.39405394e-01
3.96401793e-01 3.63069862e-01 -4.03382689e-01 5.50570369e-01
3.20330352e-01 5.09457409e-01 3.63861322e-01 -2.40566015e-01
6.25898719e-01 1.64780855e-01 -1.41127333e-01 -1.00771964e+00
-6.60373092e-01 -9.16445553e-01 -6.31849170e-01 -2.59671360e-01
1.00843191e-01 -2.62942731e-01 -1.13285458e+00 1.66356206e+00
2.44958431e-01 5.91842830e-01 -2.39340827e-01 7.07637072e-01
6.98003650e-01 4.18848693e-01 -1.28940167e-02 -1.54311970e-01
1.47467041e+00 -9.77837384e-01 -7.42753983e-01 -1.97344750e-01
6.14555538e-01 -7.97803521e-01 7.53849149e-01 5.02445579e-01
-8.49050164e-01 -3.08674991e-01 -1.38567853e+00 -5.83486371e-02
-5.17194271e-01 -2.58364201e-01 2.74488866e-01 1.52962708e+00
-1.28336084e+00 6.21202469e-01 -5.52669048e-01 -1.64407462e-01
1.49714142e-01 4.37432289e-01 -2.54421622e-01 4.89250362e-01
-1.61618769e+00 9.18126702e-01 4.92333025e-01 -4.57055345e-02
-1.13496363e+00 -3.76547873e-01 -7.47866750e-01 1.23566054e-01
4.04702365e-01 -7.67266214e-01 9.95732486e-01 -4.88937467e-01
-1.85219884e+00 5.89801967e-01 3.05104166e-01 -8.35953891e-01
3.65114748e-01 -2.25050300e-01 -9.30010021e-01 1.31536037e-01
-2.59516776e-01 -2.27229163e-01 1.46689260e+00 -1.07567954e+00
4.21552248e-02 -1.58118606e-01 -2.31958464e-01 -2.24684119e-01
-4.51637655e-01 6.48840129e-01 -3.40538710e-01 -7.04946518e-01
-4.29145582e-02 -1.03755224e+00 2.54585892e-02 -9.67263818e-01
-8.48562598e-01 3.46284918e-02 1.14680493e+00 -8.32229793e-01
1.74319911e+00 -2.07637835e+00 1.49153158e-01 2.55997121e-01
2.92931855e-01 8.95244598e-01 -2.18659759e-01 5.30139089e-01
-2.74216115e-01 4.30401951e-01 -2.69979954e-01 -7.63969362e-01
1.96751446e-01 -3.93724620e-01 -3.23999733e-01 6.87616229e-01
5.62298536e-01 7.63780773e-01 -1.05293536e+00 -3.72274727e-01
2.80884355e-01 6.63754761e-01 -7.90223658e-01 1.66003197e-01
-5.66299185e-02 2.15536162e-01 -2.15145275e-01 4.60229874e-01
7.65474916e-01 -4.20523703e-01 3.91245693e-01 -7.69303665e-02
1.81051895e-01 7.82475710e-01 -9.61017966e-01 1.68018889e+00
-3.09932739e-01 7.31614113e-01 2.73754597e-01 -8.02081406e-01
8.84997547e-01 5.52468598e-01 -4.09366563e-02 -2.60231584e-01
3.17680359e-01 5.71149111e-01 1.94739431e-01 -4.36577260e-01
7.16783404e-01 1.77959025e-01 -3.53292048e-01 4.52734560e-01
6.40868783e-01 -2.12589785e-01 -1.02473572e-01 5.20677269e-01
1.81059778e+00 -2.18201682e-01 -2.09185947e-02 -2.36970633e-01
4.78187084e-01 -6.56190574e-01 2.33084828e-01 1.14923882e+00
-3.63323331e-01 5.29976070e-01 6.47274196e-01 -2.68157661e-01
-8.16909015e-01 -8.23111713e-01 5.95147498e-02 1.42393804e+00
1.19830683e-01 -6.79188609e-01 -8.29118550e-01 -9.53984737e-01
-1.45142034e-01 6.41903877e-01 -3.47160995e-01 -3.04054648e-01
-8.35988402e-01 -1.00630736e+00 1.16330969e+00 1.08680157e-02
3.94018859e-01 -9.03765559e-01 -4.00288105e-01 2.95396656e-01
-1.23190448e-01 -1.25632977e+00 -7.34628499e-01 4.75097954e-01
-1.57423660e-01 -7.57379234e-01 -2.70239830e-01 -3.60894501e-01
3.19637731e-02 2.94733912e-01 1.19716644e+00 1.48791254e-01
-1.11426937e-03 -2.53921449e-01 -5.46240628e-01 -3.48140836e-01
-6.51832581e-01 3.32224429e-01 1.03694156e-01 4.98614639e-01
1.07677922e-01 -5.74088454e-01 -3.74112040e-01 3.72353867e-02
-7.98235893e-01 -3.66396993e-01 4.77472931e-01 9.65031803e-01
-4.01208937e-01 -2.28623196e-01 4.05951679e-01 -8.72514129e-01
7.87796080e-01 -6.87109828e-01 -3.55123818e-01 1.31052643e-01
-4.32644755e-01 3.72772545e-01 5.30610502e-01 -4.53930050e-01
-6.80489182e-01 -4.20229048e-01 -1.85272962e-01 -6.18030906e-01
1.28210112e-01 1.03739321e-01 -1.26070425e-01 -1.87541798e-01
8.82820725e-01 1.09595163e-02 -3.98905098e-01 -4.56576645e-01
2.77317673e-01 7.87463188e-01 5.56203246e-01 -2.86821187e-01
9.00472283e-01 -3.30499895e-02 -2.72808701e-01 -9.12930667e-01
-2.94008106e-01 -5.25654554e-01 -1.67703092e-01 -6.66356981e-02
6.16170347e-01 -7.02086091e-01 -7.55422890e-01 1.02071500e+00
-1.63404942e+00 -1.50659233e-01 1.47066444e-01 2.78397560e-01
-1.17817655e-01 6.84458911e-01 -1.10071087e+00 -1.15057623e+00
-6.21634364e-01 -1.30641937e+00 1.41496146e+00 -2.75932163e-01
-2.39179134e-01 -6.41835213e-01 1.85101584e-01 3.15086931e-01
6.58173561e-01 1.77142829e-01 3.63070488e-01 -1.20046031e+00
-2.94748604e-01 -2.29710415e-01 -3.22377741e-01 3.22351635e-01
-2.38162190e-01 8.96014795e-02 -1.40432429e+00 -5.67188382e-01
1.25236167e-02 -1.87145352e-01 1.50094378e+00 -2.50412673e-02
9.10256028e-01 -4.59964663e-01 -2.60989010e-01 7.25848734e-01
1.04472172e+00 1.20100178e-01 3.69915783e-01 1.40676824e-02
7.64551401e-01 2.69589543e-01 -2.86810875e-01 2.17545062e-01
6.33855239e-02 9.86432672e-01 5.52174687e-01 1.93865940e-01
-2.09867448e-01 -6.53620586e-02 9.33764279e-01 1.32107854e+00
1.07730225e-01 -9.16673183e-01 -8.83510888e-01 3.99315506e-01
-1.61211050e+00 -1.27306974e+00 -1.20557502e-01 2.14238429e+00
5.40325403e-01 3.48252863e-01 1.43834680e-01 4.87781674e-01
1.08316386e+00 6.82655931e-01 -2.10685328e-01 -8.29519629e-01
-2.50292450e-01 4.51276630e-01 5.80404758e-01 6.53364658e-01
-1.47589517e+00 8.08291078e-01 6.89188385e+00 1.06618214e+00
-1.11064029e+00 4.25363362e-01 2.72318035e-01 -4.29017767e-02
-1.73961297e-01 -2.32215554e-01 -4.72782791e-01 7.66617954e-01
1.60970724e+00 4.82635409e-01 4.56620693e-01 4.58251715e-01
-2.17440099e-01 6.99414492e-01 -6.66301131e-01 7.47867942e-01
2.74493039e-01 -9.72158134e-01 -4.00579512e-01 4.37862203e-02
5.08185506e-01 4.52544600e-01 3.32113206e-01 2.71377623e-01
8.79672229e-01 -8.76446366e-01 7.77519107e-01 5.37114963e-02
8.91413689e-01 -5.16755104e-01 7.05121636e-01 1.05309010e-01
-1.12456059e+00 -1.65463120e-01 2.41059035e-01 4.23548043e-01
1.78695068e-01 5.57894111e-01 -1.07780576e+00 8.27748775e-01
5.21856666e-01 2.20808670e-01 -5.29438198e-01 8.72189522e-01
-2.35555798e-01 9.01988804e-01 -4.04261768e-01 -2.21944451e-01
3.03847313e-01 4.38773155e-01 1.03422892e+00 1.89351046e+00
5.52203298e-01 -4.07231033e-01 -3.75551209e-02 3.66731495e-01
-3.84913087e-02 -2.62239844e-01 -5.55180430e-01 -1.16627239e-01
6.94932640e-01 1.06812882e+00 -3.39816481e-01 -3.72193336e-01
-7.60041475e-02 1.11331379e+00 3.74736995e-01 2.02160671e-01
-1.10255408e+00 -5.48884988e-01 6.68713689e-01 -2.78125107e-01
5.51154613e-01 -1.75863892e-01 -3.55155505e-02 -1.30867195e+00
-2.07440898e-01 -1.20838523e+00 4.87467587e-01 -4.60609674e-01
-1.09237933e+00 1.18000007e+00 -3.93867880e-01 -9.94291306e-01
-3.74765873e-01 -7.09625065e-01 -5.60544670e-01 8.16063583e-01
-1.18674922e+00 -1.02437937e+00 -6.65583163e-02 5.30419528e-01
2.22141460e-01 -2.54480422e-01 8.60841572e-01 3.30973923e-01
-1.01490080e+00 9.10166442e-01 -1.16207495e-01 1.62311703e-01
6.63684249e-01 -1.28496623e+00 1.10905433e+00 1.33297658e+00
4.41700995e-01 6.67167187e-01 1.07119405e+00 -5.48865676e-01
-1.29692864e+00 -1.00221574e+00 9.63086128e-01 -6.23626173e-01
9.17491913e-01 -5.19221902e-01 -9.66742337e-01 5.06499112e-01
6.13922238e-01 2.28976399e-01 5.19157231e-01 2.92321384e-01
-1.21397245e+00 1.45911843e-01 -1.05651784e+00 3.56389761e-01
1.25899839e+00 -1.03988016e+00 -2.90165901e-01 1.73040390e-01
1.10954642e+00 -5.29449165e-01 -4.75787580e-01 4.07989591e-01
3.97836387e-01 -1.25102258e+00 9.42622006e-01 -6.18735790e-01
-4.23641831e-01 -3.88213456e-01 -1.65069029e-02 -1.48726940e+00
-7.34219909e-01 -1.55326962e+00 -6.58848882e-01 9.26999748e-01
5.32153130e-01 -1.00550079e+00 3.34409237e-01 -4.00123820e-02
-3.94458026e-01 -2.07220703e-01 -1.22268546e+00 -8.90957057e-01
-3.55305165e-01 -8.16316485e-01 9.58633780e-01 8.56364429e-01
1.79397553e-01 5.68180859e-01 -8.63306642e-01 5.07291555e-01
5.78793645e-01 -3.73078644e-01 4.85362232e-01 -9.24867809e-01
-7.26330400e-01 -7.26435840e-01 -6.33575618e-01 -8.14789414e-01
-1.40867114e-01 -7.89949298e-01 -1.22539207e-01 -5.97309172e-01
-3.47295627e-02 -1.30414113e-01 -8.47134948e-01 1.39328092e-01
-4.30789888e-01 4.87875372e-01 2.83524126e-01 -4.64711413e-02
-9.22638297e-01 1.38518885e-01 5.39886355e-01 -3.13684523e-01
4.13261950e-02 -2.22358346e-01 -6.14209473e-01 3.79838407e-01
7.51516700e-01 -5.06122470e-01 -1.06757566e-01 -4.81227577e-01
1.69818193e-01 1.35250390e-02 3.66280735e-01 -1.12710106e+00
2.73236454e-01 4.08909798e-01 -6.00857846e-02 -4.14771378e-01
6.09092712e-01 -4.14158732e-01 3.63408208e-01 6.11978233e-01
-3.14590037e-01 1.63780481e-01 7.72154555e-02 5.99858344e-01
4.56574373e-02 -5.15559921e-03 7.34197140e-01 2.37672478e-01
-2.11928561e-01 2.78284699e-01 -1.12974681e-01 1.54409856e-02
7.17787802e-01 1.80213839e-01 -8.44086349e-01 -2.73925900e-01
-4.65171963e-01 -1.60501003e-01 2.72249907e-01 5.52886128e-01
3.43236685e-01 -1.24649191e+00 -9.94206250e-01 5.30064106e-01
1.76435374e-02 -7.43899643e-01 3.09805721e-01 8.34232211e-01
-4.03892219e-01 6.44329309e-01 2.57530749e-01 -4.44006532e-01
-1.26596677e+00 8.50039959e-01 2.51130402e-01 -8.04835498e-01
-2.52942026e-01 1.23014295e+00 -1.50071114e-01 -3.20914268e-01
-5.16176224e-02 1.09948907e-02 -1.96361952e-02 5.63838966e-02
5.43143690e-01 5.56112230e-01 5.66021740e-01 -1.10039461e+00
-7.62574971e-01 9.91194770e-02 4.40298729e-02 -1.00868441e-01
8.38537395e-01 1.19678602e-01 2.02865154e-01 7.16418773e-02
1.42003274e+00 2.36386463e-01 -8.08781803e-01 -3.09338570e-01
2.24114627e-01 -3.48469108e-01 9.95583832e-02 -7.07077026e-01
-8.21922779e-01 8.88847351e-01 3.95808041e-01 1.05741405e+00
1.08394504e+00 -7.08023608e-02 7.87975490e-01 -1.46481335e-01
5.16392529e-01 -7.53003299e-01 2.44708255e-01 6.58276856e-01
8.70438218e-01 -1.12410903e+00 -3.47826719e-01 -3.15038294e-01
-5.07454991e-01 7.30085492e-01 1.43229514e-01 -9.72044021e-02
4.41295862e-01 4.13721561e-01 -1.36006385e-01 -2.89233327e-01
-1.02763259e+00 -2.33825117e-01 3.95949602e-01 5.54069161e-01
4.58545774e-01 4.17212069e-01 6.62868749e-03 3.68967861e-01
-1.79795817e-01 -7.32046604e-01 1.79832742e-01 3.46323073e-01
-3.39668691e-01 -1.11321855e+00 -4.60144132e-01 3.85263115e-01
-8.52084637e-01 -2.94184387e-01 -4.47516829e-01 3.67244869e-01
5.60165159e-02 1.34695029e+00 -2.86596954e-01 -1.19083524e+00
1.68149576e-01 1.92372456e-01 1.71305120e-01 -4.61073816e-01
-1.23819578e+00 1.48440987e-01 5.80940247e-01 -7.09859252e-01
-1.55570835e-01 -7.75067568e-01 -3.80173773e-01 -4.52768624e-01
-9.38627601e-01 1.61034092e-01 5.78188539e-01 5.28306603e-01
7.64678836e-01 6.84617639e-01 8.44533622e-01 -9.43275094e-01
-7.61324525e-01 -1.30192280e+00 -3.67760479e-01 3.68847728e-01
9.31629062e-01 -5.94372272e-01 -6.11018658e-01 -5.62807202e-01] | [14.03677749633789, 5.859729290008545] |
2356d00d-9e82-4032-b6d1-9f26f16de0bb | generative-gaitnet | 2201.12044 | null | https://arxiv.org/abs/2201.12044v1 | https://arxiv.org/pdf/2201.12044v1.pdf | Generative GaitNet | Understanding the relation between anatomy andgait is key to successful predictive gait simulation. Inthis paper, we present Generative GaitNet, which isa novel network architecture based on deep reinforce-ment learning for controlling a comprehensive, full-body, musculoskeletal model with 304 Hill-type mus-culotendons. The Generative Gait is a pre-trained, in-tegrated system of artificial neural networks learnedin a 618-dimensional continuous domain of anatomyconditions (e.g., mass distribution, body proportion,bone deformity, and muscle deficits) and gait condi-tions (e.g., stride and cadence). The pre-trained Gait-Net takes anatomy and gait conditions as input andgenerates a series of gait cycles appropriate to theconditions through physics-based simulation. We willdemonstrate the efficacy and expressive power of Gen-erative GaitNet to generate a variety of healthy andpathologic human gaits in real-time physics-based sim-ulation. | ['Jehee Lee', 'Moonseok Park', 'Jaedong Lee', 'Phil Sik Chang', 'Sehee Min', 'Jungnam Park'] | 2022-01-28 | null | null | null | null | ['gait-identification'] | ['computer-vision'] | [-3.88101816e-01 1.44955024e-01 -2.07000881e-01 2.13103727e-01
-3.53986025e-01 2.99712215e-02 2.49984398e-01 -7.04110339e-02
-2.64762163e-01 1.13067198e+00 1.55680016e-01 -2.24540561e-01
-3.36793870e-01 -1.04333937e+00 -1.00540340e+00 -6.20738506e-01
-7.04503357e-01 8.08169484e-01 9.80735794e-02 -5.97259760e-01
-2.89336681e-01 2.35437736e-01 -1.39454496e+00 -2.20713481e-01
8.52551758e-01 4.55499172e-01 4.01750132e-02 8.91708612e-01
8.11079979e-01 5.00334859e-01 -5.24228871e-01 1.77399784e-01
-1.13475278e-01 -7.32456982e-01 -4.93195474e-01 -1.61218598e-01
-9.39270183e-02 -1.74212471e-01 -6.97689116e-01 4.77493107e-01
7.35822022e-01 3.48408431e-01 8.78886342e-01 -1.18399739e+00
-5.08637607e-01 5.24157405e-01 -2.77277231e-01 2.74696171e-01
3.29804391e-01 4.59871292e-01 5.75342953e-01 -5.22485793e-01
6.70523167e-01 1.18494189e+00 1.10891271e+00 6.36610746e-01
-1.36314881e+00 -4.81688917e-01 -5.16691387e-01 7.27784112e-02
-1.03539431e+00 1.85369357e-01 7.41339743e-01 -3.11327606e-01
8.12590301e-01 8.55656713e-02 1.50558317e+00 1.35929704e+00
1.48603761e+00 3.13820183e-01 9.75375473e-01 -1.72401488e-01
4.75108713e-01 -1.12932765e+00 -4.76016760e-01 7.78510213e-01
3.44543010e-01 7.53385484e-01 -4.81872797e-01 1.60874445e-02
1.37571871e+00 -3.95255923e-01 1.41039103e-01 -2.45715961e-01
-9.85908508e-01 3.72959405e-01 4.69177336e-01 6.71657175e-02
-6.78138435e-01 1.17672050e+00 3.07199329e-01 4.89702970e-02
2.91338135e-02 6.73522174e-01 -3.28497201e-01 -2.74784237e-01
-7.49525189e-01 1.12639129e+00 6.12183809e-01 2.63753802e-01
2.56674945e-01 8.42186451e-01 -1.78676378e-02 6.42038524e-01
3.51277977e-01 9.87812281e-01 4.96842921e-01 -1.27486551e+00
-2.49647275e-02 1.43231705e-01 -1.24024875e-01 -1.00992417e+00
-9.13769305e-01 -3.25929672e-01 -1.00848162e+00 4.51405764e-01
2.27303848e-01 -9.04960811e-01 -1.34108603e+00 1.81734478e+00
3.09905767e-01 3.48715544e-01 -2.68722832e-01 1.08400631e+00
5.38614213e-01 4.39103216e-01 6.74834847e-01 3.53083074e-01
1.03704095e+00 -2.43990064e-01 -4.44474697e-01 -2.88442224e-01
5.00630140e-01 -2.85675466e-01 8.68406653e-01 5.05837463e-02
-1.73699141e+00 -3.55302930e-01 -1.12355006e+00 3.83789331e-01
-1.01440683e-01 -2.30609849e-01 6.00073040e-01 3.77657264e-02
-1.16590166e+00 1.20343006e+00 -1.40490448e+00 -2.97394663e-01
1.47656396e-01 4.25394982e-01 -2.31655553e-01 2.67575830e-01
-1.63759983e+00 1.20649254e+00 4.79831606e-01 1.12238908e-02
-1.14800608e+00 -1.09689975e+00 -8.94062996e-01 8.97494480e-02
-2.01279502e-02 -1.38597715e+00 1.31717849e+00 -4.21051472e-01
-1.46339023e+00 6.01776481e-01 8.46907020e-01 -5.13620496e-01
4.23914343e-01 -3.77297737e-02 -2.62518257e-01 2.13283703e-01
6.06907681e-02 8.32649171e-01 3.72812897e-01 -1.13976634e+00
2.15485781e-01 4.13542166e-02 -3.20879221e-01 5.26823699e-01
3.08669865e-01 -3.58335108e-01 -2.55822055e-02 -1.12345231e+00
-3.84490713e-02 -1.00584590e+00 -4.05059367e-01 2.29730159e-01
-5.64797282e-01 4.60132927e-01 5.61393142e-01 -7.43516088e-01
8.89081299e-01 -1.71113873e+00 6.18972957e-01 3.17737460e-01
2.19568405e-02 3.13993096e-02 -2.85849664e-02 7.27292061e-01
-4.54887152e-01 -2.90185865e-02 -4.58871573e-01 1.83127567e-01
-9.50551480e-02 6.45617604e-01 4.00136471e-01 3.07319880e-01
2.09451318e-01 1.13792729e+00 -1.09615147e+00 -8.46503139e-01
3.27940702e-01 4.32741165e-01 -7.45019197e-01 1.15384579e-01
-4.11702693e-01 5.64396083e-01 -5.34734011e-01 6.70789063e-01
1.70588225e-01 6.49168417e-02 2.79444069e-01 -5.95431717e-04
9.56380665e-02 -2.56758928e-01 -6.47154391e-01 1.58579862e+00
-4.97653574e-01 1.24992728e-01 2.95145422e-01 -8.48931491e-01
6.14405572e-01 4.42982763e-01 1.01253402e+00 -7.48492301e-01
7.97487915e-01 2.42764264e-01 5.00701427e-01 -5.29958487e-01
1.35507628e-01 -6.12818480e-01 -5.01892745e-01 3.81302416e-01
1.94629475e-01 -7.72604823e-01 3.24761569e-01 -3.24754268e-01
1.28604066e+00 4.02824402e-01 2.65756510e-02 -6.86024487e-01
-3.09256557e-02 3.11245054e-01 3.43512326e-01 3.29360515e-01
-1.85032502e-01 5.09067774e-01 2.98023462e-01 -2.64299512e-01
-1.48233676e+00 -1.58437335e+00 3.38991173e-02 7.32165396e-01
-5.29668480e-02 -2.01299712e-01 -9.37418640e-01 4.78014857e-01
3.22223574e-01 5.15875876e-01 -8.19459796e-01 -9.62398887e-01
-9.59037602e-01 -9.63191807e-01 6.39434993e-01 8.67999971e-01
6.08536661e-01 -1.49810374e+00 -9.18285429e-01 5.84262908e-01
-4.46205512e-02 -4.53218132e-01 -1.63687050e-01 3.35214883e-01
-1.00865817e+00 -1.14665151e+00 -7.43790567e-01 -9.83274102e-01
4.23691362e-01 -7.35275567e-01 1.24093413e+00 3.78217995e-01
-5.95658898e-01 3.30084920e-01 3.70223843e-03 -1.92137554e-01
-8.00148368e-01 9.27253515e-02 2.78892010e-01 -1.00938570e+00
-4.80723202e-01 -1.19942701e+00 -8.84245634e-01 2.19370231e-01
-6.62403584e-01 3.35545599e-01 6.74553514e-01 1.03053677e+00
1.00908792e+00 5.75481076e-03 5.83152115e-01 -2.53638774e-01
9.24837947e-01 -6.11329854e-01 -5.34183197e-02 -5.38288802e-02
-1.41392693e-01 -1.20542243e-01 6.55124009e-01 -4.35553044e-01
-7.42866337e-01 -1.28705725e-02 -5.50341189e-01 -3.60342115e-01
-7.07960427e-02 8.10409546e-01 1.61499888e-01 2.29307953e-02
9.12074506e-01 3.42237353e-01 2.92780966e-01 -3.95455882e-02
4.36367452e-01 -3.25856686e-01 9.77756023e-01 -1.25762010e+00
7.90487230e-01 -9.68690738e-02 6.30072176e-01 -5.18406570e-01
-2.26688683e-02 4.41963434e-01 -5.07518709e-01 -5.87716401e-01
9.85544682e-01 -4.69595373e-01 -1.07195938e+00 9.63927329e-01
-5.26503563e-01 -1.23812973e+00 -5.46536088e-01 3.65187436e-01
-1.38936388e+00 -3.90033759e-02 -1.12351394e+00 -2.16187254e-01
-6.26131892e-01 -6.52567387e-01 1.02391946e+00 3.13146889e-01
-6.12649977e-01 -1.17965925e+00 4.90357727e-01 -3.82994324e-01
5.44904947e-01 1.12363887e+00 1.23549569e+00 1.77293897e-01
-4.39757444e-02 -1.68111086e-01 2.69171834e-01 -7.65509233e-02
1.37153164e-01 1.35401394e-02 -9.61422324e-02 -2.21356004e-01
-5.32558143e-01 -5.55454075e-01 1.04395874e-01 8.36236537e-01
1.16950560e+00 -4.01801690e-02 -2.91902155e-01 5.00336051e-01
1.31413043e+00 1.76297277e-01 7.19049931e-01 3.34701121e-01
5.06509840e-01 6.95158467e-02 4.02089745e-01 4.70044166e-01
1.02177396e-01 5.09075999e-01 4.49773014e-01 -3.11640918e-01
-5.24604380e-01 -4.83290613e-01 -1.86611861e-01 7.65874505e-01
-5.49904168e-01 -1.28050536e-01 -1.11303020e+00 4.35030401e-01
-1.47883308e+00 -8.85674417e-01 3.01552534e-01 1.76935482e+00
1.07143641e+00 4.01629269e-01 4.12368178e-01 -6.84875902e-03
5.82231522e-01 -1.82630733e-01 -8.22420359e-01 -2.19555199e-01
9.63827670e-02 6.47970855e-01 4.80584085e-01 1.36631846e-01
-7.36408174e-01 8.35206926e-01 7.60070848e+00 3.21752459e-01
-1.24019372e+00 -1.80696636e-01 5.36620021e-01 -2.36353632e-02
-7.72806033e-02 -5.56938708e-01 3.43920253e-02 6.24719501e-01
1.53821695e+00 -6.81832135e-01 4.28613096e-01 5.51057935e-01
5.05714953e-01 -1.66487485e-01 -6.48141503e-01 2.49267876e-01
-2.78745919e-01 -1.37079191e+00 -3.03213596e-01 -2.90060163e-01
5.50078869e-01 -2.43023280e-02 1.77130789e-01 3.84263575e-01
7.54463971e-01 -1.15905726e+00 6.77105010e-01 6.27758801e-01
1.23049128e+00 -8.38351548e-01 5.51512599e-01 5.69522753e-02
-1.08885634e+00 -1.92865655e-02 1.73266217e-01 1.88118875e-01
6.39643133e-01 7.77125880e-02 -5.22412002e-01 5.63316345e-01
7.13077545e-01 6.56044841e-01 -5.47972284e-02 1.22682464e+00
-2.30450630e-01 7.47439086e-01 -4.45702195e-01 6.63575456e-02
7.14405021e-03 -9.93348509e-02 4.62531924e-01 7.72594392e-01
4.24877077e-01 2.38815233e-01 2.48595774e-01 8.73760760e-01
1.36842519e-01 -1.81040555e-01 -3.95521551e-01 -1.48305893e-01
2.94899702e-01 8.33755672e-01 -8.00244629e-01 -4.65487577e-02
5.85619926e-01 4.79367733e-01 -5.89920394e-02 4.13136363e-01
-1.07459295e+00 -1.94893554e-01 7.21247792e-01 6.65090799e-01
-5.90682551e-02 -4.95351493e-01 -3.43123108e-01 -5.67535996e-01
-5.32370150e-01 -6.98815286e-01 2.63203353e-01 -1.24012077e+00
-1.11814821e+00 2.01359242e-01 6.48598611e-01 -9.74999964e-01
-6.78146541e-01 -4.46497440e-01 -9.30743039e-01 8.03835392e-01
-7.25276291e-01 -1.02916348e+00 -4.40619551e-02 4.62553442e-01
2.04280332e-01 2.91088134e-01 8.21198285e-01 2.60567784e-01
-5.53214371e-01 3.74779046e-01 -1.21564135e-01 6.31138012e-02
4.01165158e-01 -1.29428041e+00 6.11134708e-01 4.12451506e-01
-1.00331843e+00 4.38781708e-01 1.35570896e+00 -1.33106244e+00
-1.34140313e+00 -1.07067716e+00 2.01427832e-01 1.70909420e-01
9.20804024e-01 4.67144027e-02 -7.41708815e-01 5.39887011e-01
-2.56363899e-02 2.78567919e-03 2.67591387e-01 -2.62935609e-01
6.12379432e-01 2.41431817e-01 -1.15220261e+00 1.01386786e+00
1.19435072e+00 4.60010059e-02 -5.97305477e-01 3.96057725e-01
3.87407869e-01 -1.20003164e+00 -1.20805371e+00 3.75424176e-01
6.19930089e-01 -2.73473561e-01 1.36771405e+00 -8.93275559e-01
1.12000859e+00 1.53646633e-01 2.64425188e-01 -1.84619999e+00
-3.60260546e-01 -2.79663622e-01 -2.90468216e-01 1.01958595e-01
3.76463607e-02 -2.85295308e-01 1.17624009e+00 2.69024909e-01
-7.46660352e-01 -1.31418085e+00 -1.15600646e+00 -9.06171441e-01
7.38962352e-01 -7.19262734e-02 5.25838673e-01 6.09947979e-01
-1.68460943e-02 -4.36723009e-02 -6.08819723e-01 -1.73799172e-01
4.77260172e-01 -3.62520695e-01 3.86792988e-01 -9.83303964e-01
-4.30164367e-01 -3.09711695e-01 -4.85700637e-01 -1.74881175e-01
4.69969632e-03 -5.17892897e-01 3.46970052e-01 -1.83184266e+00
-3.36073071e-01 -4.15871292e-01 -4.80537713e-02 7.80982673e-01
-1.37398556e-01 7.24406782e-05 -3.83778028e-02 -1.16879068e-01
4.70164232e-02 8.57040346e-01 1.95115542e+00 1.26904190e-01
-2.23165348e-01 -2.56843328e-01 -1.41684920e-01 4.27630454e-01
9.35361207e-01 -4.05107647e-01 -6.38741612e-01 5.27637973e-02
2.34338213e-02 7.81851470e-01 5.98189235e-01 -1.49768639e+00
-1.25246108e-01 -4.16251808e-01 7.39285767e-01 -4.91858453e-01
4.87093747e-01 -3.94225597e-01 6.85450315e-01 1.20821071e+00
-1.34697989e-01 5.36882758e-01 5.80348372e-01 4.71728235e-01
6.90838099e-02 4.86841530e-01 6.94534361e-01 -2.69245625e-01
-4.05350178e-01 3.96629930e-01 -8.74326825e-01 2.13443458e-01
8.78843725e-01 -3.50461990e-01 -1.35835633e-01 -3.65839124e-01
-1.29354918e+00 3.42782408e-01 4.97095585e-01 1.73608661e-01
6.44517481e-01 -1.52761173e+00 -6.81645572e-01 6.28833994e-02
-3.22101831e-01 -2.25759163e-01 6.67084634e-01 6.36227012e-01
-1.20537353e+00 -4.80136573e-02 -1.24478519e+00 -3.62447172e-01
-7.08973587e-01 7.58873820e-02 9.29951668e-01 -3.60239059e-01
-8.14643681e-01 4.61422294e-01 -2.52683878e-01 -7.30945647e-01
-5.30432522e-01 -4.11289960e-01 9.51409899e-03 -8.31333518e-01
-2.34512553e-01 3.51101100e-01 -1.49883643e-01 -4.32325065e-01
-1.59581721e-01 3.88665617e-01 6.24736547e-01 -1.27825871e-01
1.30143738e+00 4.57749844e-01 2.07471728e-01 2.21649379e-01
4.77148145e-01 -9.46353555e-01 -1.41017044e+00 6.40380383e-01
-5.65815866e-01 1.08822831e-03 -2.40671828e-01 -9.25429106e-01
-1.32727969e+00 3.88574600e-01 6.01691842e-01 -4.00519043e-01
9.32202280e-01 -1.62498176e-01 1.08604848e+00 1.34095699e-01
5.37018001e-01 -1.20664430e+00 4.39100146e-01 3.68561774e-01
1.17000234e+00 -5.87782979e-01 1.31839469e-01 -4.10392463e-01
-1.22820795e-01 9.80015755e-01 9.94096875e-01 -7.44619846e-01
9.19092119e-01 5.59476197e-01 -1.06098175e-01 -5.50474346e-01
-5.02817035e-01 3.08852553e-01 9.85875875e-02 8.79398286e-01
3.41046214e-01 3.75525892e-01 -5.73953390e-01 2.22943842e-01
-7.21313953e-01 3.16834927e-01 3.94194692e-01 1.39102805e+00
-3.79796863e-01 -9.24733341e-01 -4.72467095e-01 5.22431433e-01
-2.10100248e-01 2.41939604e-01 1.19077899e-02 1.39155757e+00
1.66722804e-01 1.85910136e-01 8.44921619e-02 -4.50093210e-01
5.00040710e-01 -2.58517295e-01 7.38650739e-01 -6.25496864e-01
-5.39473474e-01 -1.82718918e-01 2.57612556e-01 -4.66610551e-01
4.33610007e-02 -6.49360359e-01 -1.77620053e+00 -6.73526764e-01
2.15526238e-01 -2.75398403e-01 5.96501350e-01 8.09961975e-01
1.48602545e-01 1.19995821e+00 -3.41951512e-02 -1.22581971e+00
-3.07958722e-01 -8.06623816e-01 -6.84115231e-01 4.84961033e-01
-3.00758541e-01 -1.16373491e+00 4.84392419e-02 -5.51511161e-02] | [6.927558898925781, 0.04730941727757454] |
aca47fdb-6990-4037-b9a5-5e14bd8ef2fc | progressive-cross-camera-soft-label-learning | 1908.05669 | null | https://arxiv.org/abs/1908.05669v2 | https://arxiv.org/pdf/1908.05669v2.pdf | Progressive Cross-camera Soft-label Learning for Semi-supervised Person Re-identification | In this paper, we focus on the semi-supervised person re-identification (Re-ID) case, which only has the intra-camera (within-camera) labels but not inter-camera (cross-camera) labels. In real-world applications, these intra-camera labels can be readily captured by tracking algorithms or few manual annotations, when compared with cross-camera labels. In this case, it is very difficult to explore the relationships between cross-camera persons in the training stage due to the lack of cross-camera label information. To deal with this issue, we propose a novel Progressive Cross-camera Soft-label Learning (PCSL) framework for the semi-supervised person Re-ID task, which can generate cross-camera soft-labels and utilize them to optimize the network. Concretely, we calculate an affinity matrix based on person-level features and adapt them to produce the similarities between cross-camera persons (i.e., cross-camera soft-labels). To exploit these soft-labels to train the network, we investigate the weighted cross-entropy loss and the weighted triplet loss from the classification and discrimination perspectives, respectively. Particularly, the proposed framework alternately generates progressive cross-camera soft-labels and gradually improves feature representations in the whole learning course. Extensive experiments on five large-scale benchmark datasets show that PCSL significantly outperforms the state-of-the-art unsupervised methods that employ labeled source domains or the images generated by the GAN-based models. Furthermore, the proposed method even has a competitive performance with respect to deep supervised Re-ID methods. | ['Yang Gao', 'Lei Qi', 'Jing Huo', 'Yinghuan Shi', 'Lei Wang'] | 2019-08-15 | null | null | null | null | ['semi-supervised-person-re-identification'] | ['computer-vision'] | [ 3.45887035e-01 -3.56854945e-01 -3.80713306e-02 -6.45660520e-01
-5.66155136e-01 -5.91938257e-01 6.77956939e-01 -1.62033856e-01
-6.16157055e-01 7.20963597e-01 1.59036174e-01 2.95947164e-01
-1.08520836e-01 -4.96533364e-01 -4.88579988e-01 -7.92366028e-01
5.85646212e-01 6.80880606e-01 -2.09645316e-01 2.76685029e-01
-2.04748616e-01 2.27708489e-01 -1.58104861e+00 -5.91657534e-02
9.28031802e-01 8.52437973e-01 -2.20229197e-02 1.42106593e-01
2.06213564e-01 4.64570314e-01 -3.55674952e-01 -9.16990459e-01
3.87239099e-01 -3.89907479e-01 -5.68786323e-01 4.13362384e-01
6.61321342e-01 -2.39374265e-01 -2.84628391e-01 1.31843972e+00
5.26710272e-01 3.18660617e-01 5.73551536e-01 -1.26220310e+00
-7.68637598e-01 2.25778684e-01 -7.82534301e-01 -4.09519449e-02
5.30573964e-01 7.36537874e-02 7.74644256e-01 -7.63408244e-01
3.33504915e-01 1.24546123e+00 8.47988248e-01 8.45815837e-01
-1.22372198e+00 -9.49833453e-01 3.20638418e-01 2.39359573e-01
-1.63078988e+00 -3.25461060e-01 9.10223424e-01 -5.68942070e-01
2.72696137e-01 1.53533101e-01 5.51821291e-01 1.34748518e+00
-4.07793820e-01 5.79062283e-01 1.37053597e+00 -4.30155367e-01
-1.33740336e-01 5.07884681e-01 2.04322278e-01 6.24377191e-01
1.91122115e-01 2.88073540e-01 -2.80904919e-01 8.59446675e-02
5.52798450e-01 4.90844339e-01 -3.67814869e-01 -2.54167080e-01
-1.36170197e+00 5.20619512e-01 3.54219615e-01 1.84782654e-01
-6.01685420e-02 -3.32321852e-01 3.33869994e-01 -6.69863522e-02
3.71299297e-01 1.80702806e-02 -1.49082616e-01 1.13356479e-01
-9.40981448e-01 -8.60112309e-02 3.77300143e-01 1.10411668e+00
8.45421016e-01 -2.62966335e-01 -2.55039126e-01 1.14693594e+00
3.05040091e-01 5.45979381e-01 5.11901677e-01 -5.49572587e-01
6.90374792e-01 6.40347481e-01 2.37139255e-01 -1.06779706e+00
-2.88933426e-01 -5.92139065e-01 -1.29917371e+00 -2.14863688e-01
5.63598514e-01 -2.62768298e-01 -7.94862568e-01 1.90041804e+00
3.45454931e-01 6.20841026e-01 1.10598626e-02 9.91354346e-01
7.23854542e-01 4.63219762e-01 2.31800139e-01 -2.24701971e-01
1.39838290e+00 -1.36321175e+00 -5.89181840e-01 -3.26297939e-01
4.05579925e-01 -4.80261832e-01 8.43750775e-01 1.30838960e-01
-6.94071293e-01 -1.09552681e+00 -8.77736390e-01 2.00469092e-01
-4.87775385e-01 7.45523572e-01 1.45992637e-01 7.92171776e-01
-7.57915318e-01 3.68212521e-01 -4.40422237e-01 -5.39979279e-01
2.84204930e-01 4.22113895e-01 -6.83670819e-01 -3.59989852e-01
-1.14891493e+00 6.11388862e-01 4.34829384e-01 3.02759588e-01
-8.31075490e-01 -5.28494179e-01 -8.28183413e-01 6.61338568e-02
4.88250226e-01 -7.00099051e-01 7.33785033e-01 -1.36957109e+00
-1.44162512e+00 1.15890086e+00 -4.06202115e-02 4.40608636e-02
6.71532512e-01 -6.63625374e-02 -7.10165083e-01 4.00714651e-02
2.77081668e-01 6.57678962e-01 7.72580028e-01 -1.46330166e+00
-8.07920575e-01 -4.31004196e-01 1.34211704e-01 3.16835195e-01
-6.30277157e-01 1.83983296e-02 -6.98767483e-01 -8.28217804e-01
-2.97891021e-01 -1.34995031e+00 1.05324097e-01 -2.96307772e-01
-7.15798020e-01 -3.28428984e-01 5.08031964e-01 -7.11529315e-01
9.30068433e-01 -2.06298447e+00 1.10922135e-01 2.80213535e-01
-4.39691916e-02 5.44012487e-01 -4.07360159e-02 3.71548459e-02
-3.96687567e-01 -7.60003850e-02 -2.56389380e-01 -9.07221675e-01
8.04216489e-02 -9.82951969e-02 1.43230841e-01 4.87993836e-01
-1.60221756e-01 6.48661554e-01 -8.38526607e-01 -7.14122474e-01
3.38897616e-01 6.12515628e-01 -1.81234702e-01 4.62659419e-01
4.45494294e-01 8.19870889e-01 -2.46283069e-01 7.26123214e-01
8.34445059e-01 -3.53128195e-01 1.96203172e-01 -4.48094130e-01
1.87318966e-01 -3.47395539e-01 -1.20306325e+00 1.65330446e+00
-4.94965106e-01 1.75881058e-01 -2.57946342e-01 -1.00299609e+00
8.68106246e-01 2.32353359e-01 4.27738518e-01 -5.60309231e-01
8.25069249e-02 3.54460329e-02 -5.21371782e-01 -2.09688097e-01
1.84815109e-01 -1.43331513e-01 -6.85417205e-02 4.91578847e-01
3.65462936e-02 8.33077610e-01 1.47419438e-01 -8.06322172e-02
5.23803413e-01 1.96175277e-01 2.39800915e-01 -5.01032211e-02
1.18002868e+00 -4.17373806e-01 8.11522007e-01 4.91337776e-01
-3.37256014e-01 7.40580440e-01 -1.38666525e-01 -3.86117190e-01
-9.00996625e-01 -9.97319281e-01 -8.04398507e-02 8.69699776e-01
6.05256975e-01 -3.39768916e-01 -9.75969136e-01 -1.17374420e+00
-1.11718960e-01 3.59047264e-01 -7.01190412e-01 -2.24482626e-01
-4.24973011e-01 -7.36071885e-01 5.10015070e-01 6.41790688e-01
9.58379626e-01 -7.32760191e-01 4.26972732e-02 -4.70038094e-02
-3.73472393e-01 -1.35095668e+00 -1.00813472e+00 -2.00449914e-01
-3.51660758e-01 -1.11811280e+00 -1.13005674e+00 -1.03659904e+00
1.12207222e+00 4.06415731e-01 7.55878091e-01 1.97221246e-03
-5.88328168e-02 6.44221187e-01 -4.25002635e-01 2.72266507e-01
5.42008728e-02 1.71978530e-02 3.38347793e-01 7.81817079e-01
6.59015477e-01 -3.54793787e-01 -5.98333895e-01 7.80763686e-01
-5.59967816e-01 6.78998306e-02 4.57823426e-01 1.14014578e+00
8.61345291e-01 3.88305366e-01 4.97168571e-01 -9.17406797e-01
2.72921056e-01 -3.43876958e-01 -5.27258933e-01 7.41959751e-01
-6.67042613e-01 -1.95571527e-01 8.29176009e-01 -7.10186481e-01
-1.35731769e+00 3.87291819e-01 9.51052830e-02 -5.55123329e-01
-4.73910123e-01 9.39029753e-02 -6.00927114e-01 -1.69297174e-01
1.20478071e-01 2.95758307e-01 -4.45584893e-01 -5.70828736e-01
1.13553077e-01 8.88228357e-01 6.66718304e-01 -5.94033360e-01
1.12982285e+00 5.72935879e-01 -2.38602579e-01 -2.89169103e-01
-1.04821885e+00 -6.79522574e-01 -9.16942477e-01 -3.38227838e-01
1.14552331e+00 -1.24338937e+00 -7.20968902e-01 7.71959960e-01
-9.72142756e-01 2.44973861e-02 -1.18342362e-01 5.57539582e-01
-2.82934904e-01 5.76827288e-01 -4.09063697e-01 -7.01539040e-01
-2.08016321e-01 -1.23140550e+00 1.12455344e+00 6.94472373e-01
1.26713127e-01 -1.19098747e+00 1.11268461e-01 7.30212212e-01
-7.06934109e-02 1.01991624e-01 4.04013902e-01 -6.90016210e-01
-2.95478582e-01 -3.57391775e-01 -5.10546803e-01 5.26994705e-01
3.24944735e-01 -4.98972982e-01 -1.07300293e+00 -5.94284475e-01
-3.92631024e-01 -3.83389026e-01 6.37278855e-01 1.98611300e-02
1.24179542e+00 -1.98775247e-01 -5.51362991e-01 8.54265034e-01
1.37547803e+00 1.51664801e-02 3.97760808e-01 3.10155213e-01
1.15653300e+00 7.84402251e-01 6.28415346e-01 5.18177450e-01
7.48161495e-01 9.96815383e-01 -4.22912762e-02 -2.25975752e-01
-6.29462227e-02 -5.81342638e-01 2.91259646e-01 7.50691295e-01
-3.20823491e-01 -2.80319452e-01 -6.86692297e-01 3.56875390e-01
-1.84686148e+00 -9.29094315e-01 1.23943686e-01 2.37324572e+00
5.89819014e-01 -1.83555961e-01 3.21491450e-01 -4.78118435e-02
1.44995141e+00 -2.57927086e-02 -6.11359835e-01 2.90950388e-01
-5.30987158e-02 -2.65341640e-01 6.43306375e-01 2.52037227e-01
-1.48506498e+00 7.10811913e-01 4.88223505e+00 9.64931130e-01
-7.39814043e-01 4.37289178e-01 7.65463531e-01 3.71621884e-02
9.02742520e-02 -1.97567001e-01 -1.24717081e+00 9.92260039e-01
5.70676565e-01 2.08363488e-01 6.54875815e-01 8.32361341e-01
-1.84643716e-01 1.51553884e-01 -1.38386202e+00 1.65111780e+00
5.33125281e-01 -6.80705488e-01 -7.89739471e-03 1.33428305e-01
9.53356564e-01 -5.11940897e-01 8.14050157e-03 2.23575279e-01
3.03591907e-01 -7.87016809e-01 6.20147109e-01 6.26966000e-01
1.16930735e+00 -9.01022017e-01 1.09535396e+00 3.36694360e-01
-1.50422990e+00 -3.01157087e-01 -3.52135777e-01 3.78574252e-01
3.08503419e-01 4.27422971e-01 -2.40927473e-01 8.13789189e-01
9.16030347e-01 1.04145277e+00 -7.99130559e-01 8.75778794e-01
-1.74186796e-01 2.49117166e-01 -1.71223387e-01 4.13728416e-01
-2.56449450e-03 -4.24420059e-01 1.83683172e-01 1.25686514e+00
4.91061002e-01 -2.93092728e-02 4.28585291e-01 7.69123018e-01
-2.41465136e-01 -9.23109949e-02 -2.66844511e-01 2.45306700e-01
4.30754632e-01 1.32217288e+00 -5.22810400e-01 -3.54280502e-01
-6.64145291e-01 1.36177027e+00 3.79165709e-01 4.98718262e-01
-9.70136225e-01 -2.20635101e-01 3.82132411e-01 4.76662479e-02
1.19273923e-01 1.77738339e-01 1.23898305e-01 -1.53929293e+00
1.34935603e-01 -7.37441003e-01 6.23547435e-01 -6.04381621e-01
-1.83796465e+00 6.60883784e-01 3.79196964e-02 -1.54456341e+00
-1.12567030e-01 -3.93540323e-01 -4.74556357e-01 7.81194687e-01
-1.66421103e+00 -1.63360596e+00 -7.39248574e-01 1.01635194e+00
2.33293280e-01 -4.40191507e-01 5.98168492e-01 8.05319667e-01
-9.24415588e-01 1.34909058e+00 2.99710959e-01 4.56435651e-01
1.02153349e+00 -1.05235636e+00 -5.45470119e-02 9.39277112e-01
1.83808301e-02 5.66126168e-01 -1.79695114e-02 -5.42566538e-01
-8.08768570e-01 -1.36810398e+00 8.03231537e-01 -3.35676581e-01
2.29468524e-01 -4.56053942e-01 -5.46243012e-01 8.82332146e-01
-8.22323635e-02 1.90676808e-01 9.39938486e-01 6.44995645e-02
-4.37331498e-01 -5.55075169e-01 -1.25731325e+00 2.48109773e-01
1.29618859e+00 -7.82582641e-01 -3.55937958e-01 4.36591178e-01
3.70777726e-01 -1.07730538e-01 -8.45539272e-01 3.06691587e-01
5.49451351e-01 -8.62142444e-01 1.20460796e+00 -2.54632801e-01
6.67169765e-02 -4.62373823e-01 1.54210851e-01 -1.28177297e+00
-4.88898873e-01 -2.97873646e-01 2.83661425e-01 1.95907259e+00
-5.04135042e-02 -7.27960765e-01 7.52951145e-01 7.35696137e-01
1.24531157e-01 -2.96558052e-01 -8.86489153e-01 -8.64180505e-01
-3.27784061e-01 1.42875925e-01 7.85206497e-01 1.12007165e+00
-3.09944630e-01 1.13786966e-01 -8.36203754e-01 2.48127460e-01
9.38126326e-01 6.91446811e-02 7.28774965e-01 -1.37332308e+00
-4.18685436e-01 -4.97911423e-02 -3.74240428e-01 -1.13265514e+00
6.13191426e-01 -9.72392440e-01 1.33204404e-02 -1.16517961e+00
6.81430101e-01 -8.31122041e-01 -5.61988175e-01 4.60629672e-01
-3.95939291e-01 5.53524435e-01 2.95344263e-01 5.05601823e-01
-9.27903473e-01 6.36828721e-01 1.10760689e+00 -3.53623420e-01
3.02718133e-02 7.34057352e-02 -7.60350764e-01 7.16757774e-01
5.22404432e-01 -4.84646261e-01 -3.69446516e-01 -3.87517780e-01
-3.88092361e-02 -2.98814476e-01 5.63033819e-01 -1.28851712e+00
4.38974679e-01 2.17163637e-02 6.13115728e-01 -3.94973457e-01
3.55745971e-01 -1.10683095e+00 3.12745780e-01 -6.44162437e-03
-4.21262622e-01 -9.95257497e-02 -3.01639795e-01 8.19808185e-01
-2.92113751e-01 -3.58714759e-01 9.50700760e-01 -1.81302637e-01
-5.49822092e-01 5.44543564e-01 1.60708323e-01 -1.53149525e-02
1.24707842e+00 -3.47035855e-01 -1.95327967e-01 -2.78523773e-01
-7.48934567e-01 3.66232485e-01 6.79315031e-01 2.60531902e-01
3.87386560e-01 -1.72427225e+00 -5.42631149e-01 2.59751290e-01
3.55543017e-01 -2.67439503e-02 6.21055126e-01 6.75109684e-01
-2.52267998e-02 4.25968796e-01 -2.94345856e-01 -5.89345574e-01
-1.27312338e+00 8.41763794e-01 3.55020106e-01 -5.07565260e-01
-3.11765999e-01 7.45136201e-01 5.49577773e-01 -7.24597633e-01
3.13163370e-01 3.67190808e-01 -4.01651055e-01 1.01847246e-01
6.68864489e-01 3.94836426e-01 -3.45996439e-01 -1.20717847e+00
-4.37470227e-01 1.18681264e+00 -1.29524365e-01 2.02159494e-01
9.90674496e-01 -3.93295109e-01 7.82849044e-02 2.28358597e-01
1.32090318e+00 -7.28922486e-02 -1.46104968e+00 -5.47161937e-01
-2.45009571e-01 -6.11653328e-01 -3.04084688e-01 -7.86915898e-01
-1.32974529e+00 8.31523120e-01 1.03283620e+00 -2.28283465e-01
1.22870767e+00 -1.12283431e-01 9.87251401e-01 2.50655830e-01
4.46711957e-01 -1.29087150e+00 1.49349242e-01 -6.72562495e-02
3.33795518e-01 -1.50805438e+00 4.86695953e-02 -5.31323910e-01
-8.06084335e-01 8.54829490e-01 8.33288312e-01 6.19916990e-02
5.68051815e-01 -2.98236609e-01 -1.20825730e-01 3.12775612e-01
1.43439220e-02 -2.53779173e-01 3.40797335e-01 8.81650925e-01
-2.95351930e-02 2.76522368e-01 1.64325222e-01 7.55471706e-01
1.15252964e-01 5.80534525e-02 1.42140552e-01 3.63164216e-01
2.28741929e-01 -1.31036544e+00 -5.08971989e-01 1.89292058e-01
-1.61434665e-01 4.02212776e-02 -3.15670311e-01 4.58210707e-01
7.86733031e-01 1.13314378e+00 3.49877886e-02 -6.95911348e-01
1.50718451e-01 -9.89125520e-02 3.44300508e-01 -3.85007471e-01
-5.93230426e-01 -1.08031258e-01 -3.16679776e-02 -2.08187118e-01
-9.58535492e-01 -7.52128124e-01 -8.02742124e-01 -3.63097340e-01
-5.06411314e-01 8.75127017e-02 3.28547746e-01 1.01542604e+00
3.05958360e-01 3.05696219e-01 8.67487848e-01 -8.20140541e-01
-3.10488403e-01 -8.96066666e-01 -6.47392035e-01 9.07670975e-01
9.11041275e-02 -1.02604842e+00 -4.72219139e-01 3.18823278e-01] | [14.77426815032959, 1.0384069681167603] |
05e67cae-c93f-441b-b897-9d31bcb9fac7 | a-fast-semi-automatic-method-for | 2004.08690 | null | https://arxiv.org/abs/2004.08690v1 | https://arxiv.org/pdf/2004.08690v1.pdf | A fast semi-automatic method for classification and counting the number and types of blood cells in an image | A novel and fast semi-automatic method for segmentation, locating and counting blood cells in an image is proposed. In this method, thresholding is used to separate the nucleus from the other parts. We also use Hough transform for circles to locate the center of white cells. Locating and counting of red cells is performed using template matching. We make use of finding local maxima, labeling and mean value computation in order to shrink the areas obtained after applying Hough transform or template matching, to a single pixel as representative of location of each region. The proposed method is very fast and computes the number and location of white cells accurately. It is also capable of locating and counting the red cells with a small error. | ['Shahram Shirani', 'Hamed Sadeghi', 'David W. Capson'] | 2020-04-18 | null | null | null | null | ['template-matching'] | ['computer-vision'] | [-3.08772270e-02 -3.10351044e-01 2.22620279e-01 -3.33236530e-02
-2.45345592e-01 -5.11576593e-01 1.55104339e-01 8.62610519e-01
-8.46853256e-01 6.83228314e-01 -3.15036625e-01 1.31746501e-01
3.94030243e-01 -9.98092413e-01 1.20016381e-01 -8.57063770e-01
8.34509507e-02 7.89081037e-01 6.47207201e-01 3.90255928e-01
5.28065860e-01 1.11901569e+00 -1.07474470e+00 -2.62488216e-01
7.12889552e-01 6.72837377e-01 -1.38357997e-01 9.94363070e-01
-3.86285543e-01 5.86905122e-01 -7.99646974e-01 3.61594632e-02
3.16934437e-01 -5.73771417e-01 -5.85280001e-01 5.03551781e-01
1.24956787e-01 -3.59760344e-01 4.18476760e-01 1.00574553e+00
4.83493388e-01 1.23858333e-01 1.00975049e+00 -6.34356558e-01
-2.50870604e-02 5.68218902e-02 -1.28145027e+00 5.68592906e-01
2.30765864e-01 -2.35325601e-02 2.64884353e-01 -6.60065234e-01
5.41844785e-01 8.55662405e-01 4.94397432e-01 1.94337741e-01
-9.77207661e-01 -4.89336163e-01 -8.11099708e-01 -3.89573164e-02
-1.86786830e+00 -2.23680779e-01 2.69768298e-01 -5.11881649e-01
6.58131778e-01 3.71987641e-01 7.73139536e-01 -5.21107137e-01
2.15562627e-01 3.84006768e-01 1.17948437e+00 -9.20715988e-01
2.84768909e-01 9.19186473e-02 2.23593757e-01 9.21755314e-01
7.13205397e-01 -2.98615634e-01 2.04294667e-01 -1.66138098e-01
1.02460098e+00 2.08441496e-01 5.57827279e-02 6.36947006e-02
-1.12855065e+00 6.73532605e-01 1.16973575e-02 8.72111499e-01
-6.47319674e-01 8.27404037e-02 2.71777779e-01 -4.16115850e-01
1.39318228e-01 8.90538022e-02 3.28876637e-02 -3.04065011e-02
-1.42382014e+00 -1.97181761e-01 7.64479637e-01 6.59297585e-01
7.83191979e-01 -1.61519304e-01 -2.20475256e-01 4.97857898e-01
-2.25286465e-03 4.94212717e-01 4.05781448e-01 -1.08174002e+00
-1.63684115e-01 1.02925980e+00 1.96969420e-01 -8.98963392e-01
-3.92151743e-01 1.92663118e-01 -7.31729388e-01 7.11560845e-01
1.13703334e+00 -1.92025661e-01 -9.90616679e-01 7.83239365e-01
6.74121797e-01 1.31110713e-01 -1.73674554e-01 7.27508008e-01
5.07993758e-01 7.34997690e-01 1.73763335e-01 -4.58064646e-01
1.75018215e+00 -6.05725765e-01 -7.15379834e-01 2.99841493e-01
5.25482476e-01 -1.09409964e+00 3.88556451e-01 2.31661245e-01
-1.09802020e+00 -4.00576711e-01 -8.10732543e-01 -3.04742269e-02
-4.82330441e-01 5.29070437e-01 2.99794465e-01 6.71327412e-01
-8.05000603e-01 3.76370013e-01 -9.03992355e-01 -6.46326602e-01
2.22974911e-01 4.38684702e-01 -3.74519825e-01 3.49520832e-01
-3.29700410e-01 8.94015551e-01 4.78875011e-01 1.67734280e-01
-1.11076139e-01 -2.27400556e-01 -6.81344032e-01 -6.99496269e-03
-2.64833532e-02 -4.68589574e-01 6.64596438e-01 -8.13594937e-01
-1.21117318e+00 1.19925845e+00 -4.88063753e-01 -4.75623965e-01
4.78921473e-01 4.88085479e-01 -1.35511741e-01 7.94150174e-01
5.27462661e-02 5.06154656e-01 6.05789483e-01 -8.24074209e-01
-8.96435797e-01 -4.60555106e-01 -7.43865430e-01 9.64016914e-02
2.53471792e-01 3.10510725e-01 -4.32811320e-01 -3.60215843e-01
3.55220675e-01 -3.50320578e-01 -2.58941531e-01 1.72929689e-02
-2.28085369e-01 -3.58538687e-01 8.87001097e-01 -9.03535962e-01
1.17720962e+00 -2.04994678e+00 -7.13489592e-01 6.10584378e-01
5.43610930e-01 1.94531620e-01 5.40547192e-01 5.72858006e-02
2.85735965e-01 -1.16751872e-01 -3.61306012e-01 1.45712942e-02
-4.84627634e-01 6.48800954e-02 2.94146955e-01 7.36616850e-01
3.54537964e-02 4.68676716e-01 -6.29302919e-01 -1.35667634e+00
5.44773996e-01 7.44445324e-01 1.19210564e-01 -1.23693906e-02
5.58567524e-01 2.70751327e-01 -3.76195818e-01 7.57719457e-01
8.66514623e-01 -2.20857263e-02 -1.65840447e-01 -1.01712324e-01
-5.01952827e-01 -4.25640225e-01 -1.54188371e+00 6.45291567e-01
-6.74940944e-02 6.21478796e-01 1.49603769e-01 -6.66680336e-01
1.36978412e+00 4.22565758e-01 8.30756307e-01 -3.05989742e-01
6.49322212e-01 3.07779342e-01 -1.40424266e-01 -3.57683003e-01
3.81855518e-01 -3.40884298e-01 4.56864357e-01 7.68189967e-01
-2.44716913e-01 -1.22296371e-01 9.99079525e-01 3.70718315e-02
5.90652227e-01 -2.96561658e-01 9.39241648e-01 -2.88678408e-01
9.88730073e-01 1.31076142e-01 5.28767407e-01 2.46238768e-01
-5.01579583e-01 5.70006728e-01 3.40879261e-01 -4.67715949e-01
-9.08349454e-01 -7.16816902e-01 -4.96071011e-01 4.41671908e-01
2.55018950e-01 2.79940873e-01 -1.22889233e+00 -2.03575477e-01
-9.77627560e-03 4.62688237e-01 -5.53227603e-01 5.76058209e-01
-8.61534417e-01 -7.91486144e-01 1.17873579e-01 2.33331174e-01
6.94403291e-01 -1.18414652e+00 -1.33308208e+00 3.46666366e-01
5.83509915e-02 -8.49949181e-01 -4.84593570e-01 -1.55963793e-01
-1.16020262e+00 -1.27554667e+00 -1.20105302e+00 -9.74003732e-01
1.28945971e+00 2.23319441e-01 6.62943542e-01 5.28693497e-01
-8.58774006e-01 1.07460923e-01 -3.34091187e-01 -1.17146395e-01
-4.06709760e-01 -4.39345956e-01 -3.61942530e-01 1.44352075e-02
7.22842813e-01 -6.44205362e-02 -7.59159923e-01 3.18183511e-01
-5.22483945e-01 -4.69969749e-01 4.98359889e-01 3.20211619e-01
8.72684538e-01 3.70626867e-01 -7.44461566e-02 -8.51012647e-01
4.26998407e-01 -2.17452087e-02 -7.91262507e-01 2.43292093e-01
-4.66203332e-01 -2.58424848e-01 5.06533802e-01 -1.58498496e-01
-6.97495341e-01 2.68620670e-01 2.38010138e-01 -2.07213033e-03
-4.97836143e-01 -2.64310658e-01 4.30035532e-01 -4.36960131e-01
4.83415902e-01 2.72667468e-01 1.05059728e-01 -4.09811914e-01
1.45458028e-01 7.29178369e-01 7.72402763e-01 -8.59985948e-02
5.05849898e-01 8.12004507e-01 3.38281453e-01 -9.92328286e-01
3.58852819e-02 -9.31639791e-01 -1.05580246e+00 -3.24203014e-01
1.09622133e+00 -2.08152875e-01 -8.82990956e-01 3.60175014e-01
-1.06525385e+00 9.65774581e-02 -3.93012315e-01 6.23150527e-01
-2.24721238e-01 8.05951834e-01 -8.23644698e-01 -1.14175248e+00
-8.56469452e-01 -8.77419174e-01 5.87513268e-01 8.43039751e-01
-1.87214315e-01 -1.06431234e+00 1.63673796e-02 4.56811339e-02
2.84453005e-01 4.08178359e-01 5.88535905e-01 -6.14262104e-01
-2.88335174e-01 -8.94187093e-01 -3.50124866e-01 8.56512040e-03
2.08154678e-01 6.00424111e-01 -4.66116995e-01 -7.84417242e-02
5.42513542e-02 3.97588581e-01 7.25394964e-01 8.52926254e-01
6.05683029e-01 -1.28506467e-01 -6.07312381e-01 4.31875437e-01
1.67617714e+00 6.64624572e-01 7.79561460e-01 8.46997350e-02
2.15631574e-01 6.30317926e-01 6.65333092e-01 5.82109809e-01
1.55440755e-02 6.50507212e-02 -1.31034613e-01 -5.78057289e-01
-1.58263117e-01 2.53881365e-01 -2.14987844e-01 4.78748649e-01
-3.74852777e-01 1.30518794e-01 -8.37817430e-01 6.25176132e-01
-1.17264438e+00 -8.90974641e-01 -4.53752875e-01 2.43272042e+00
6.85809374e-01 -7.15552866e-02 5.12613118e-01 4.73300189e-01
1.23206699e+00 -4.66122597e-01 1.70662105e-02 -4.99050200e-01
5.81043661e-02 6.09526217e-01 9.27830875e-01 8.32130194e-01
-1.03224361e+00 7.91193724e-01 6.73218012e+00 7.80773878e-01
-1.12448633e+00 -5.85110709e-02 8.33178878e-01 4.48220104e-01
4.95275468e-01 8.93261209e-02 -8.74436498e-01 5.18111229e-01
2.01000825e-01 -8.44403133e-02 1.92012608e-01 3.61592650e-01
2.91290492e-01 -1.28747165e+00 -4.79241967e-01 9.13717985e-01
-2.78043579e-02 -9.80208695e-01 -2.64611989e-01 -7.47084245e-02
6.08006716e-01 -7.46177256e-01 -4.29835677e-01 -3.61847401e-01
-2.33559702e-02 -6.97131097e-01 3.08665454e-01 6.78423464e-01
5.64906836e-01 -7.26020277e-01 9.45068240e-01 1.43528074e-01
-1.49346840e+00 3.07081580e-01 -5.63995063e-01 -2.40811016e-02
8.32936317e-02 6.59157753e-01 -1.18464148e+00 -1.75662309e-01
2.25417614e-01 -3.21077295e-02 -5.74014604e-01 1.70675910e+00
2.12434176e-02 2.95757473e-01 -7.94273078e-01 -2.52979159e-01
1.03513919e-01 -8.27817678e-01 2.71511614e-01 1.50376773e+00
4.22915936e-01 2.33422235e-01 -3.11168525e-02 9.87104833e-01
3.21961939e-01 7.50008821e-01 -1.57588094e-01 1.70637250e-01
7.50465930e-01 1.54911077e+00 -1.83713257e+00 -7.37757206e-01
-1.37774134e-02 7.53310323e-01 -2.26451576e-01 2.19914213e-01
-4.37954754e-01 -1.01682210e+00 -4.29644942e-01 3.30312103e-01
1.02394968e-01 -2.97556847e-01 -6.00660264e-01 -3.31785589e-01
-2.63871759e-01 -8.65014568e-02 6.78367615e-01 -4.08966035e-01
-4.33890104e-01 3.83421838e-01 -2.24438775e-02 -1.00651336e+00
-2.89651722e-01 -4.08384949e-01 -1.09088886e+00 1.02692735e+00
-1.03137016e+00 -8.49863291e-01 -3.19089502e-01 3.94853204e-01
1.81929305e-01 -4.19734744e-03 4.72859114e-01 3.30913849e-02
-6.35101974e-01 2.11618528e-01 4.15829197e-02 4.00142401e-01
5.44202566e-01 -1.51511168e+00 -1.54380873e-01 9.99526560e-01
-1.63792416e-01 6.91008687e-01 7.01150298e-01 -6.92230523e-01
-5.79959035e-01 -3.64084005e-01 1.16481841e+00 9.98597518e-02
1.69443741e-01 1.01330027e-01 -6.50696874e-01 2.18178347e-01
1.32361263e-01 1.26435816e-01 5.33517361e-01 -4.57748204e-01
3.18565726e-01 -1.70303077e-01 -1.66120994e+00 2.60511100e-01
-2.90539086e-01 6.04506284e-02 -2.74901271e-01 3.23673040e-01
-2.01574355e-01 -2.82174498e-01 -8.97582293e-01 -8.33372474e-02
5.27899444e-01 -1.14151621e+00 7.46653378e-01 3.72318000e-01
-3.20968211e-01 -7.48882651e-01 6.35622144e-01 -5.45461416e-01
-1.62653536e-01 -5.65774918e-01 4.54412311e-01 1.14510965e+00
2.93046296e-01 -5.72276592e-01 9.46806371e-01 4.23786581e-01
3.89253110e-01 -4.77826834e-01 -1.01898634e+00 -1.72496721e-01
-3.20959598e-01 4.61803705e-01 2.38836691e-01 7.32048869e-01
1.92481294e-01 -8.25637206e-02 5.34354337e-02 -2.38238811e-01
9.66769993e-01 2.22803280e-01 5.52281499e-01 -1.23143268e+00
1.16411291e-01 -5.94553530e-01 -6.40118122e-01 -5.84332943e-01
-4.43129659e-01 -4.02762502e-01 -1.20008335e-01 -1.67589068e+00
2.26303309e-01 -3.59129161e-01 1.01739027e-01 3.97406250e-01
-2.21753195e-01 6.17446184e-01 4.47714329e-02 2.11724982e-01
-1.85318351e-01 -5.66541910e-01 1.36991799e+00 3.62509608e-01
-3.70311081e-01 1.48538381e-01 -1.41309679e-01 6.65272236e-01
9.01820838e-01 -3.97755325e-01 3.31712604e-01 2.36144379e-01
-4.05247390e-01 6.93552420e-02 1.42861068e-01 -1.07414758e+00
4.75111514e-01 -9.35168639e-02 7.21428692e-01 -9.42815006e-01
2.09866185e-02 -8.28253090e-01 -1.26714736e-01 8.90054822e-01
5.22349179e-02 4.60341619e-03 -9.09612179e-02 6.99891746e-02
-2.58580092e-02 -9.40087855e-01 1.37840593e+00 -3.75948012e-01
-6.66147947e-01 -1.45464018e-01 -7.59030104e-01 -4.50304031e-01
1.50124002e+00 -9.12657380e-01 -7.63827981e-03 -1.48951605e-01
-7.53068388e-01 1.07009269e-01 4.63101029e-01 -8.29494774e-01
5.39778411e-01 -8.02565217e-01 -6.68041170e-01 2.44262218e-01
-2.56491959e-01 -1.56692922e-01 2.09784564e-02 1.23354793e+00
-1.70881164e+00 3.57947230e-01 -3.51273209e-01 -3.33554476e-01
-1.66107857e+00 5.45771718e-01 4.82591182e-01 -1.35223165e-01
-5.81450045e-01 7.56269515e-01 -3.04783612e-01 5.23171067e-01
1.28262907e-01 -4.54216480e-01 -5.83509386e-01 -1.13363251e-01
7.72252679e-01 1.08069777e+00 -3.30644101e-01 -9.10969079e-01
-3.49434763e-01 1.20272887e+00 2.71180749e-01 -1.19034000e-01
5.82332969e-01 -2.20908299e-01 -5.92586875e-01 2.70790100e-01
9.90244269e-01 5.17926812e-01 -8.52444291e-01 -7.80949974e-03
4.24678624e-02 -5.47525346e-01 2.67225523e-02 -5.46726108e-01
-1.05590343e+00 1.13077557e+00 7.98731089e-01 2.65360773e-01
1.26375115e+00 -2.35948861e-01 7.88742185e-01 -2.50190884e-01
1.36899367e-01 -1.30241454e+00 -4.08260345e-01 8.12415853e-02
1.48311839e-01 -8.64705026e-01 4.29645926e-01 -6.36259615e-01
-5.29330254e-01 1.70320880e+00 4.07431781e-01 -3.90371352e-01
4.55949754e-01 6.05399966e-01 1.70705408e-01 -1.06601439e-01
-5.57517931e-02 -5.90200007e-01 9.91705433e-02 5.64616203e-01
6.99894905e-01 9.45085585e-02 -1.02177274e+00 -1.97199732e-01
1.72008917e-01 1.62870690e-01 4.80991751e-01 8.82955492e-01
-1.22305512e+00 -7.39644349e-01 -9.99202013e-01 3.85861933e-01
-6.91336632e-01 2.47677520e-01 -4.08853859e-01 9.36237633e-01
4.37335521e-01 9.46465492e-01 5.94470978e-01 2.04835996e-01
-3.62918414e-02 -4.69511598e-02 6.42665923e-01 -1.96589053e-01
-6.21659636e-01 7.62071550e-01 -3.18560749e-01 -3.28353830e-02
-2.27806866e-01 -5.53013563e-01 -1.92207098e+00 -3.01093370e-01
-4.37648475e-01 4.56686318e-01 6.20353281e-01 6.28131926e-01
-4.13806051e-01 -2.02188537e-01 5.05216360e-01 -3.81983966e-01
-2.12922722e-01 -1.01328886e+00 -1.16428256e+00 2.53977329e-01
1.01688102e-01 -2.88042665e-01 -4.20739084e-01 3.91228259e-01] | [14.710326194763184, -3.053816795349121] |
14d427bd-0ff5-436f-90cc-d3af314ef531 | imagearg-a-multi-modal-tweet-dataset-for | 2209.06416 | null | https://arxiv.org/abs/2209.06416v1 | https://arxiv.org/pdf/2209.06416v1.pdf | ImageArg: A Multi-modal Tweet Dataset for Image Persuasiveness Mining | The growing interest in developing corpora of persuasive texts has promoted applications in automated systems, e.g., debating and essay scoring systems; however, there is little prior work mining image persuasiveness from an argumentative perspective. To expand persuasiveness mining into a multi-modal realm, we present a multi-modal dataset, ImageArg, consisting of annotations of image persuasiveness in tweets. The annotations are based on a persuasion taxonomy we developed to explore image functionalities and the means of persuasion. We benchmark image persuasiveness tasks on ImageArg using widely-used multi-modal learning methods. The experimental results show that our dataset offers a useful resource for this rich and challenging topic, and there is ample room for modeling improvement. | ['Diane Litman', 'Yue Dai', 'Meiqi Guo', 'Zhexiong Liu'] | 2022-09-14 | null | https://aclanthology.org/2022.argmining-1.1 | https://aclanthology.org/2022.argmining-1.1.pdf | argmining-acl-2022-10 | ['argument-mining'] | ['natural-language-processing'] | [ 7.98390985e-01 3.78974944e-01 -6.24227166e-01 -4.78574514e-01
-8.49444330e-01 -3.40247214e-01 1.32451749e+00 4.74499285e-01
-7.26843059e-01 5.98389447e-01 7.67140090e-01 -6.16168797e-01
-3.24012414e-02 -6.95381701e-01 -6.46601737e-01 -4.95607376e-01
4.04211760e-01 2.09580466e-01 4.27939557e-02 -4.42213356e-01
6.69423103e-01 -1.75061166e-01 -1.71834993e+00 9.36373949e-01
6.98386610e-01 7.24869967e-01 2.58723974e-01 6.95134342e-01
-1.20766886e-01 1.05341733e+00 -7.98629105e-01 -5.42931437e-01
-1.08485579e-01 -2.98372984e-01 -1.08082652e+00 1.76868752e-01
9.67157125e-01 -1.44197926e-01 7.10316822e-02 9.86274660e-01
5.18905580e-01 -1.03906617e-02 9.23339069e-01 -1.27784085e+00
-1.15765476e+00 1.12091398e+00 -8.15142035e-01 6.43350363e-01
3.24010164e-01 8.76756534e-02 1.29878926e+00 -9.63384867e-01
7.25076020e-01 1.74411809e+00 4.38209176e-01 4.97465432e-01
-1.11530304e+00 -5.83009839e-01 2.46085823e-01 9.13949788e-01
-7.11097538e-01 -2.00056851e-01 9.96991932e-01 -3.56705606e-01
3.21481138e-01 5.30663967e-01 8.17793787e-01 1.90551376e+00
-1.00776747e-01 1.41119063e+00 1.79103875e+00 -6.64224327e-01
1.33874267e-01 2.12828919e-01 6.07645273e-01 8.26379836e-01
8.99590105e-02 -2.04555064e-01 -6.01761341e-01 2.40915138e-02
1.79783881e-01 -3.41955036e-01 -1.16752662e-01 1.59128625e-02
-1.47030818e+00 1.57648766e+00 4.35502231e-01 -1.37830749e-02
-1.67748541e-01 8.95369127e-02 6.33224905e-01 1.16266534e-01
8.28834832e-01 1.91805795e-01 -2.70049404e-02 -2.95368396e-02
-4.54331398e-01 3.32496583e-01 5.20427883e-01 3.92831415e-01
6.50260925e-01 -4.56511155e-02 -3.88324767e-01 9.74669218e-01
2.69772142e-01 7.47730553e-01 5.03410995e-01 -6.84796512e-01
4.55264330e-01 6.63381219e-01 -1.70565143e-01 -1.41269791e+00
-4.66965228e-01 1.38209522e-01 -7.65929163e-01 3.57054055e-01
2.87904829e-01 -1.24611877e-01 -4.67512637e-01 1.38408935e+00
4.82271969e-01 -6.88898787e-02 -4.24260050e-02 1.14621115e+00
1.49959242e+00 6.77615821e-01 7.92701244e-01 -5.87623827e-02
2.01349592e+00 -1.12157309e+00 -5.06226182e-01 -5.74928403e-01
5.42248011e-01 -8.92242551e-01 2.04786992e+00 6.21636026e-02
-7.34672308e-01 -4.70642477e-01 -1.17459714e+00 -1.56541303e-01
-3.89458805e-01 1.78646892e-02 7.49729335e-01 6.83933914e-01
-4.10919487e-01 -1.73248410e-01 1.01307809e-01 -8.09583068e-02
9.84879017e-01 -5.39896548e-01 -8.95135403e-02 4.49612411e-03
-1.48821473e+00 9.93137538e-01 2.44364813e-01 -6.76718891e-01
-9.13000941e-01 -7.41348684e-01 -9.19442475e-01 -1.53954014e-01
6.28013074e-01 -8.19968045e-01 1.23294246e+00 -1.46774566e+00
-1.15516770e+00 1.20704544e+00 3.41735512e-01 -6.07428849e-01
5.98354220e-01 -2.41679296e-01 -4.20943618e-01 4.53628540e-01
4.34423119e-01 1.01770234e+00 1.36807096e+00 -1.10543966e+00
-8.31658483e-01 -3.59833598e-01 4.37727571e-01 5.96339703e-01
-7.83965468e-01 1.57847166e-01 1.41529903e-01 -6.02797091e-01
-8.48906457e-01 -6.43867671e-01 -7.65609369e-02 -1.18026309e-01
-5.99259138e-01 -6.87416732e-01 1.29325461e+00 -1.49891898e-01
9.00443196e-01 -1.86339593e+00 -3.07793617e-01 -7.04913661e-02
6.64912581e-01 4.43796739e-02 -1.47117957e-01 -2.16084972e-01
-1.57058369e-02 1.89266384e-01 3.65606427e-01 4.57686186e-02
2.12846231e-02 8.30470100e-02 -2.73677289e-01 6.49911702e-01
-1.47308663e-01 1.07373250e+00 -9.72212553e-01 -1.13639987e+00
3.92068803e-01 2.79226065e-01 -1.34581268e-01 -3.09167892e-01
-5.74022830e-01 -8.92007798e-02 -9.30735171e-01 3.86308730e-01
2.85285115e-01 -8.31887007e-01 -1.15893766e-01 -2.52870023e-01
-1.99262530e-01 2.73518741e-01 -6.04018629e-01 1.37803519e+00
-2.89273262e-01 1.01474595e+00 -9.06947181e-02 -1.03729522e+00
4.74912256e-01 -9.15762857e-02 1.29327714e-01 -1.10039639e+00
4.72766995e-01 -4.38290656e-01 -6.76256046e-02 -5.15235126e-01
7.87690520e-01 -1.64115101e-01 -3.01464468e-01 1.01097453e+00
-7.85242617e-01 2.47171661e-03 3.59681278e-01 7.78251290e-01
8.82679880e-01 -4.37882066e-01 4.56293643e-01 -5.93267798e-01
6.79186046e-01 7.99461603e-01 -4.15068537e-01 9.90704119e-01
-1.21851206e-01 2.63920486e-01 1.79053217e-01 -5.55282414e-01
-1.11171770e+00 -7.76077032e-01 -2.49987140e-01 1.70297146e+00
4.07490939e-01 -4.94367808e-01 -8.17382693e-01 -1.05576563e+00
-8.90545174e-02 8.47535968e-01 -9.79834437e-01 1.54984117e-01
-3.38913798e-01 -9.68087792e-01 4.67639834e-01 1.20316461e-01
8.23631108e-01 -1.27004230e+00 -8.87670994e-01 -1.79595396e-01
-5.56654871e-01 -9.75602269e-01 -5.49654245e-01 -3.46585602e-01
-2.74564862e-01 -1.37649715e+00 -3.55205506e-01 -7.42943048e-01
3.94872516e-01 9.88585830e-01 1.55032659e+00 2.92100251e-01
-4.62465733e-01 9.38651085e-01 -4.04777586e-01 -1.06600904e+00
-7.89658964e-01 -1.08437978e-01 -2.96345055e-01 -2.41141528e-01
4.57107335e-01 1.66166395e-01 -9.46840763e-01 3.72963339e-01
-8.98666203e-01 5.91932356e-01 5.36334872e-01 8.73939037e-01
7.16813505e-01 -4.31085765e-01 8.26231539e-01 -1.24741817e+00
1.33239257e+00 -5.12121379e-01 -1.69713423e-01 8.21915828e-03
-9.29863811e-01 -2.48773366e-01 3.25585306e-01 -8.07979107e-01
-1.30811858e+00 -4.14761871e-01 -8.98815617e-02 1.95857406e-01
-2.63967425e-01 3.94489527e-01 4.28654663e-02 2.66011804e-01
1.25846088e+00 -1.56820267e-01 2.28388667e-01 2.84003001e-02
1.10514188e+00 6.83879197e-01 4.39657986e-01 -5.67328334e-01
3.53947401e-01 8.02469790e-01 -2.08793268e-01 -1.15743566e+00
-1.56252778e+00 -4.08015072e-01 7.08508864e-02 -6.83921874e-01
8.84456158e-01 -7.81407654e-01 -9.81991231e-01 -1.51040465e-01
-9.61068571e-01 -1.12380654e-01 -3.38343382e-01 9.36116576e-02
-4.43921745e-01 7.51220882e-01 -2.69883007e-01 -4.31684554e-01
-8.85447502e-01 -8.66436243e-01 9.81402218e-01 1.67813465e-01
-6.30887687e-01 -1.25639915e+00 2.05813209e-03 1.13913846e+00
1.08323053e-01 4.77973744e-02 9.40324783e-01 -7.46696293e-01
-1.48597598e-01 2.69079581e-02 -5.86221337e-01 3.28759812e-02
-3.73325408e-01 -3.07688534e-01 -9.36557353e-01 2.70907074e-01
-2.32461914e-01 -8.82657230e-01 9.82526183e-01 4.98219222e-01
1.44728947e+00 -7.14224458e-01 -3.05716783e-01 -3.08242023e-01
8.16521943e-01 -5.18063426e-01 7.93082297e-01 1.11870563e+00
6.73644245e-01 4.76177543e-01 1.04954159e+00 3.96575689e-01
8.50138128e-01 2.86148816e-01 4.79169399e-01 -5.81229985e-01
-4.24906313e-01 -2.19804108e-01 3.94748986e-01 2.37141550e-01
-2.05424372e-02 -3.45535845e-01 -5.54832220e-01 4.59560633e-01
-1.88580716e+00 -1.67862558e+00 -7.05768943e-01 1.28391039e+00
9.44764555e-01 2.34217029e-02 2.23141447e-01 3.08895648e-01
7.32847273e-01 5.76757491e-01 -1.95972189e-01 -5.33122480e-01
-6.74061477e-01 -8.19715858e-02 1.49492770e-01 4.03431773e-01
-1.38548982e+00 8.13469052e-01 6.61002064e+00 1.36016583e+00
-8.81679177e-01 3.04340124e-01 1.20979428e+00 1.86833411e-01
-6.53724909e-01 -3.22377771e-01 -6.80825055e-01 3.49982917e-01
6.04619324e-01 -7.45578408e-02 -2.89673150e-01 1.25901949e+00
3.75403017e-01 -4.96497452e-01 -8.31484139e-01 9.77960944e-01
5.95028162e-01 -1.99668849e+00 2.42028475e-01 -9.44847390e-02
4.87693876e-01 -4.99996357e-02 4.78591472e-01 1.97980568e-01
3.26180309e-01 -1.01309526e+00 9.54030871e-01 -1.59575999e-01
6.81144893e-01 -4.90395188e-01 2.86771148e-01 4.07149285e-01
-4.18165028e-01 2.08606914e-01 -3.11990440e-01 -3.87830079e-01
2.02026933e-01 7.72362530e-01 -1.27233362e+00 -3.32563460e-01
7.54409373e-01 7.94522882e-01 -8.38584423e-01 1.50995597e-01
-5.73535383e-01 8.50297511e-01 4.10497561e-02 -8.45859826e-01
2.62884468e-01 3.73117737e-02 5.32603860e-01 1.56952977e+00
-2.75882453e-01 9.93828401e-02 3.05351615e-01 6.52049363e-01
-2.82591820e-01 4.34051871e-01 -4.88574862e-01 1.58786345e-02
1.80700585e-01 1.90150321e+00 -8.29139292e-01 -7.32587695e-01
-6.22393668e-01 6.31878376e-01 3.42494905e-01 -2.08158299e-01
-1.17284322e+00 3.16572905e-01 2.78554827e-01 1.51363775e-01
-2.17615247e-01 1.51010752e-01 -7.66143203e-01 -8.74456882e-01
-5.17040074e-01 -1.17990613e+00 9.29533243e-01 -9.39387202e-01
-1.58520651e+00 2.83414066e-01 2.50582218e-01 -6.37255967e-01
-4.20011841e-02 -7.18503475e-01 -6.60329521e-01 3.64916295e-01
-1.79844356e+00 -1.38510227e+00 -5.86669445e-01 7.90901184e-01
1.02959740e+00 -7.72505850e-02 7.54011869e-01 -1.56567767e-02
-1.82821795e-01 2.73488849e-01 -5.52575707e-01 -3.53715867e-01
9.04521883e-01 -1.17349601e+00 5.77513613e-02 4.53942329e-01
4.66812819e-01 1.78911373e-01 9.34732080e-01 -4.32266444e-01
-1.38996804e+00 -8.58163595e-01 3.40527147e-01 -7.47147799e-01
1.09554458e+00 -4.80426699e-02 -6.29205704e-01 2.52745897e-01
7.48700440e-01 -6.70497000e-01 1.01604497e+00 3.67765576e-01
-6.07209027e-01 4.92168337e-01 -1.02920699e+00 9.98266101e-01
1.02916718e+00 -1.69129953e-01 -9.12389815e-01 5.76154530e-01
2.55264819e-01 -9.61574763e-02 -4.51087117e-01 6.05389103e-02
3.90715748e-01 -7.44259596e-01 1.52600551e+00 -5.35690188e-01
1.19955659e+00 -1.10514648e-02 1.04241766e-01 -1.36076784e+00
-5.05175948e-01 -1.17883757e-02 -4.78437319e-02 1.05359328e+00
6.15067124e-01 -6.78723305e-02 8.94569755e-01 4.29338843e-01
-1.43283382e-01 -4.23176199e-01 -6.44362390e-01 9.87500697e-02
-5.97759075e-02 -5.91593266e-01 1.78529680e-01 1.07898164e+00
1.94470391e-01 1.66682994e+00 -4.40087825e-01 -3.08706641e-01
7.34909058e-01 5.87106526e-01 1.08385110e+00 -1.14442134e+00
-2.97639761e-02 -7.64011979e-01 9.37958285e-02 -1.10384917e+00
5.21821141e-01 -9.25994098e-01 -3.38410512e-02 -1.59941244e+00
1.01290166e+00 -2.92986363e-01 2.54517436e-01 4.91684765e-01
-6.48618817e-01 6.79279506e-01 -2.09425241e-01 8.07846561e-02
-1.19906032e+00 4.16693687e-01 1.76044142e+00 -7.42370725e-01
3.22848439e-01 -5.47974050e-01 -1.35592020e+00 1.21029389e+00
9.16509092e-01 1.49689186e-02 -5.27631044e-01 -3.81066054e-01
2.14073837e-01 -5.53008258e-01 5.60616374e-01 -4.64616306e-02
-8.15534666e-02 -4.96342987e-01 5.93883157e-01 -6.40943170e-01
3.68184119e-01 -4.50189769e-01 -6.49065316e-01 1.87551916e-01
-6.02549911e-01 1.25510395e-01 -8.93750042e-03 7.87243068e-01
1.36974603e-01 -1.20737039e-01 8.80265594e-01 -5.78956716e-02
-1.26752305e+00 3.30006406e-02 -7.57562757e-01 3.61961156e-01
9.39717710e-01 -2.19866291e-01 -1.10841846e+00 -5.76685786e-01
-2.29234084e-01 1.53598741e-01 2.02401236e-01 7.36647248e-01
9.08618987e-01 -1.27518117e+00 -1.19255364e+00 -4.99257416e-01
5.71815670e-01 -5.11086762e-01 4.11684841e-01 7.39109516e-01
-2.79617310e-02 1.29673416e-02 -1.35596380e-01 -7.78698504e-01
-1.99660015e+00 6.34658694e-01 -3.18449885e-01 5.57618849e-02
-1.07548869e+00 6.08660221e-01 5.29140174e-01 1.52404830e-01
-7.27146715e-02 2.91867495e-01 -7.84528553e-01 2.04080015e-01
1.14313936e+00 4.03135598e-01 -3.35214764e-01 -7.98408568e-01
-2.27445718e-02 -2.08170246e-02 -1.53704569e-01 -5.07464036e-02
1.22927630e+00 -3.16294491e-01 1.00679338e-01 6.26652911e-02
7.52920985e-01 1.76963195e-01 -8.93265843e-01 -2.20945124e-02
-6.04507774e-02 -6.35796070e-01 2.25906029e-01 -9.82106626e-01
-4.63348240e-01 5.88525057e-01 4.54483330e-01 5.40564477e-01
6.98030949e-01 6.03248417e-01 4.10563976e-01 3.06312501e-01
-1.57293454e-01 -1.43468082e+00 6.76740229e-01 4.61695343e-01
1.19674206e+00 -1.64305234e+00 3.17897022e-01 -5.92679739e-01
-1.07335114e+00 6.69328451e-01 4.92922723e-01 1.60362720e-01
4.43485379e-01 1.16090573e-01 3.29750150e-01 -7.50636339e-01
-5.93612671e-01 4.44301777e-02 5.66997588e-01 7.43211985e-01
5.97419381e-01 4.10360008e-01 -8.57468843e-01 7.82694399e-01
-4.07002717e-01 -5.57299674e-01 5.60044467e-01 3.94646376e-01
-8.24467182e-01 -7.74524510e-01 -6.33086622e-01 5.50442219e-01
-4.93802130e-01 -3.19694012e-01 -8.70815217e-01 1.01882517e+00
-1.49656326e-01 1.18863297e+00 -1.68877319e-01 -5.67474626e-02
1.36517137e-01 -3.15099567e-01 6.46844923e-01 -3.95440549e-01
-3.38836670e-01 -1.71196252e-01 9.15899813e-01 -1.52673557e-01
-9.93502676e-01 -6.10690534e-01 -1.23226929e+00 -7.40597963e-01
-2.34812915e-01 9.07884836e-02 8.22313607e-01 1.03920805e+00
1.35031775e-01 5.38857758e-01 2.69901574e-01 -7.75782347e-01
-4.39336896e-01 -9.45134401e-01 -1.90121099e-01 9.74930525e-01
-2.07047209e-01 -6.16387248e-01 -2.24926770e-01 2.42999032e-01] | [8.456421852111816, 10.39653491973877] |
813325b5-e682-4f1c-ac58-8e992873cf1f | gs3d-an-efficient-3d-object-detection | 1903.10955 | null | http://arxiv.org/abs/1903.10955v2 | http://arxiv.org/pdf/1903.10955v2.pdf | GS3D: An Efficient 3D Object Detection Framework for Autonomous Driving | We present an efficient 3D object detection framework based on a single RGB
image in the scenario of autonomous driving. Our efforts are put on extracting
the underlying 3D information in a 2D image and determining the accurate 3D
bounding box of the object without point cloud or stereo data. Leveraging the
off-the-shelf 2D object detector, we propose an artful approach to efficiently
obtain a coarse cuboid for each predicted 2D box. The coarse cuboid has enough
accuracy to guide us to determine the 3D box of the object by refinement. In
contrast to previous state-of-the-art methods that only use the features
extracted from the 2D bounding box for box refinement, we explore the 3D
structure information of the object by employing the visual features of visible
surfaces. The new features from surfaces are utilized to eliminate the problem
of representation ambiguity brought by only using a 2D bounding box. Moreover,
we investigate different methods of 3D box refinement and discover that a
classification formulation with quality aware loss has much better performance
than regression. Evaluated on the KITTI benchmark, our approach outperforms
current state-of-the-art methods for single RGB image based 3D object
detection. | ['Wanli Ouyang', 'Buyu Li', 'Xingyu Zeng', 'Lu Sheng', 'Xiaogang Wang'] | 2019-03-26 | gs3d-an-efficient-3d-object-detection-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Li_GS3D_An_Efficient_3D_Object_Detection_Framework_for_Autonomous_Driving_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Li_GS3D_An_Efficient_3D_Object_Detection_Framework_for_Autonomous_Driving_CVPR_2019_paper.pdf | cvpr-2019-6 | ['vehicle-pose-estimation'] | ['computer-vision'] | [ 4.88044918e-02 2.38409340e-01 1.49978727e-01 -2.20282927e-01
-7.79918194e-01 -3.38578224e-01 4.41684455e-01 5.58789186e-02
-6.21204734e-01 1.17538549e-01 -5.17383099e-01 -4.21163797e-01
2.35624582e-01 -7.11372197e-01 -1.00142121e+00 -5.65671921e-01
1.31763995e-01 8.54434848e-01 9.46188092e-01 -3.83924037e-01
5.50051570e-01 9.23924208e-01 -2.19022465e+00 -7.94788450e-02
6.23540223e-01 1.71398437e+00 2.73379415e-01 5.00095427e-01
-3.68759096e-01 -1.91777460e-02 -5.09298742e-01 -6.43234253e-02
8.59987199e-01 3.19631770e-02 -2.31952220e-01 3.32523912e-01
6.63568139e-01 -5.44700682e-01 1.05159611e-01 9.86328900e-01
2.17945457e-01 -9.04511586e-02 7.29554474e-01 -1.15521479e+00
2.42759630e-01 -3.95993382e-01 -8.56083095e-01 -1.99132800e-01
4.07768965e-01 1.28661597e-03 7.35593915e-01 -1.24723113e+00
6.71738088e-01 1.31895375e+00 6.38477206e-01 3.98559958e-01
-1.02366459e+00 -7.52545714e-01 3.77354622e-01 1.63315505e-01
-1.49158692e+00 -2.50743061e-01 9.93669450e-01 -5.41012228e-01
1.02183950e+00 2.28626743e-01 8.15110326e-01 4.21332985e-01
7.82585070e-02 7.18745053e-01 1.42067266e+00 -5.69821000e-01
2.84034550e-01 2.64078856e-01 2.91676491e-01 9.97231662e-01
4.76728112e-01 4.09916610e-01 -5.54570079e-01 -4.68461215e-02
5.99358976e-01 -1.89442933e-02 1.44039929e-01 -1.06283176e+00
-7.81039655e-01 7.26443708e-01 5.16113102e-01 -4.33904141e-01
-2.55268335e-01 6.22190461e-02 5.54995500e-02 -2.23855793e-01
7.35809326e-01 1.74681351e-01 -3.41574669e-01 2.89739501e-02
-8.39253664e-01 4.54725116e-01 5.44112325e-01 1.11725438e+00
1.25712264e+00 -4.29088920e-01 2.50830352e-01 2.34908134e-01
5.51263154e-01 7.28968084e-01 -2.05363184e-01 -9.47236240e-01
4.62888807e-01 1.23275697e+00 4.84300554e-01 -5.64027548e-01
-4.90626872e-01 -1.11977145e-01 -1.65871918e-01 1.20581663e+00
3.61681044e-01 3.85407358e-01 -1.23861074e+00 7.84635782e-01
9.57517028e-01 -2.85879314e-01 -6.78800642e-02 9.49472368e-01
8.55133176e-01 4.13522899e-01 -6.93233430e-01 2.77807951e-01
1.39235866e+00 -7.47444034e-01 -8.79859477e-02 -3.70005637e-01
3.84031802e-01 -4.61959809e-01 7.32881665e-01 6.11175299e-01
-8.36515844e-01 -6.13728762e-01 -1.36815572e+00 -1.17853366e-01
-4.98719603e-01 1.28611088e-01 3.82895797e-01 7.19061017e-01
-6.96670473e-01 2.59514332e-01 -8.86135101e-01 -2.51496673e-01
5.86349666e-01 3.95646125e-01 -4.96782511e-01 -5.50193489e-02
-3.58002126e-01 1.39277768e+00 3.38084996e-01 -5.73804565e-02
-8.30722690e-01 -6.24939740e-01 -9.05414641e-01 -3.96883935e-01
7.94521749e-01 -5.23890793e-01 9.69825625e-01 -3.32112342e-01
-1.23570395e+00 1.04960442e+00 -3.86119038e-01 -5.47433019e-01
7.66040742e-01 -3.87248248e-01 3.72531146e-01 2.66930938e-01
1.99258234e-02 6.66529596e-01 1.20790684e+00 -1.72658551e+00
-1.02963531e+00 -9.16135609e-01 -3.57961543e-02 2.99655944e-01
2.01656461e-01 -3.26458871e-01 -4.40611959e-01 1.39283270e-01
7.59352028e-01 -9.00352716e-01 -3.45094860e-01 4.24862295e-01
-4.46150422e-01 -2.90698767e-01 1.07572055e+00 -3.22155088e-01
5.76188922e-01 -2.12668514e+00 -1.16534792e-01 2.63160914e-01
4.30100322e-01 4.06285077e-02 3.40902090e-01 -2.24137679e-02
4.53921974e-01 -1.84723027e-02 -3.08875859e-01 -6.68699384e-01
7.87871927e-02 1.17086813e-01 -2.87993938e-01 7.95809686e-01
4.95214194e-01 3.92471641e-01 -5.36275208e-01 -5.00336647e-01
5.79487264e-01 5.47271848e-01 -4.72031146e-01 3.49857509e-01
-1.75160050e-01 2.75041163e-01 -7.76838005e-01 8.38493407e-01
1.05717897e+00 1.02021992e-01 -6.34317756e-01 -1.07962877e-01
-4.57252800e-01 4.22325343e-01 -1.36822748e+00 1.51167631e+00
-2.15086624e-01 4.10593241e-01 1.42303497e-01 -6.73106611e-01
1.40896571e+00 -2.11077020e-01 3.77551645e-01 -5.15484869e-01
3.17849107e-02 4.34616178e-01 -3.63376081e-01 -4.87866886e-02
4.60546970e-01 -7.92643130e-02 -4.12443765e-02 1.71939194e-01
-2.43333980e-01 -7.95263231e-01 -1.38030306e-01 -1.64708182e-01
1.02230394e+00 6.03756487e-01 3.86193067e-01 -5.97750619e-02
3.16440195e-01 5.62920749e-01 3.35666418e-01 8.20033431e-01
-2.96428710e-01 7.87404835e-01 2.61435479e-01 -6.12656415e-01
-1.17624164e+00 -8.40335190e-01 -4.89885509e-01 4.34663862e-01
7.39339232e-01 -2.44937241e-01 -6.31278336e-01 -7.16162860e-01
5.21445811e-01 5.63483000e-01 -8.20183039e-01 1.33012459e-01
-4.74382907e-01 -5.57199299e-01 5.02285585e-02 4.13326591e-01
4.51431900e-01 -3.52515012e-01 -1.59931993e+00 -1.99687686e-02
2.32799277e-01 -1.26142800e+00 4.67407070e-02 6.57569587e-01
-1.03185987e+00 -1.20560265e+00 -1.92292720e-01 -1.57962695e-01
7.95784712e-01 5.33921599e-01 9.65790808e-01 1.48592480e-02
-5.28460741e-01 4.38308060e-01 -5.28491080e-01 -8.97552967e-01
-2.02675685e-01 -3.96850199e-01 7.87322000e-02 -7.83195272e-02
4.78340805e-01 -1.31770298e-01 -5.52907348e-01 4.85805392e-01
-2.42869645e-01 2.71302700e-01 5.87090969e-01 5.21205664e-01
1.00609648e+00 -4.12092246e-02 -1.95899010e-01 -5.30454755e-01
-1.68719649e-01 4.35206704e-02 -1.23642552e+00 -2.99491674e-01
-7.52947271e-01 1.93073571e-01 1.10363569e-02 -2.19086111e-01
-7.15147138e-01 8.12865436e-01 5.91505468e-02 -6.77788019e-01
-3.88899237e-01 -3.40246916e-01 5.08511700e-02 -3.75023484e-01
5.53841472e-01 9.12502334e-02 1.88575044e-01 -7.41261840e-01
1.98495328e-01 5.42163908e-01 2.56111026e-01 -3.35129291e-01
1.06560624e+00 1.11998880e+00 4.93413419e-01 -7.52582192e-01
-8.92012954e-01 -9.65133190e-01 -1.14682150e+00 -4.10174429e-01
8.11047435e-01 -9.45778370e-01 -8.64943981e-01 1.87015012e-01
-1.44654918e+00 8.56546611e-02 -3.23922902e-01 3.19236040e-01
-6.73106968e-01 8.66406709e-02 -2.16142759e-02 -1.41903591e+00
-2.29889616e-01 -1.42337453e+00 1.87826276e+00 -4.13406081e-02
1.57161951e-01 -1.86437592e-02 -4.82264757e-02 4.82959300e-01
-5.24107106e-02 4.22578454e-01 5.68415523e-01 -3.47700179e-01
-1.00323379e+00 -5.50595462e-01 -2.81741142e-01 6.38772473e-02
-3.00875723e-01 -5.12709469e-02 -1.30968988e+00 8.33909512e-02
1.44130975e-01 -2.38115601e-02 1.07878864e+00 1.49922863e-01
7.98219144e-01 2.56708741e-01 -5.94411671e-01 7.55439878e-01
1.39688838e+00 1.03815235e-01 2.49718204e-01 6.61133766e-01
6.35536671e-01 7.75078893e-01 1.01253641e+00 4.38722968e-01
4.36226398e-01 7.36883581e-01 1.16412199e+00 6.80980855e-04
-1.46401912e-01 -2.00038359e-01 2.50525504e-01 -7.56831691e-02
-3.10431361e-01 2.31010467e-01 -9.85390842e-01 2.26635039e-01
-1.58960664e+00 -3.62734586e-01 -4.12106782e-01 2.29272866e+00
2.78411657e-01 6.99791849e-01 3.81814927e-01 4.23120558e-01
4.98654008e-01 -1.00401610e-01 -7.36513913e-01 -1.56197175e-01
1.38542518e-01 -3.62458304e-02 1.01236236e+00 4.37996805e-01
-1.07816422e+00 9.58146036e-01 5.82914591e+00 5.37344992e-01
-7.69363761e-01 -8.44400190e-03 2.46227056e-01 -7.80857056e-02
1.33064901e-02 1.03792183e-01 -1.69905519e+00 1.54636158e-02
3.87603521e-01 3.71429503e-01 6.32037073e-02 1.18939543e+00
7.95968100e-02 -6.53736413e-01 -1.16999137e+00 1.00335610e+00
2.13846266e-01 -1.05367124e+00 -1.57914251e-01 3.12703222e-01
4.10477847e-01 3.07121634e-01 -1.51122540e-01 -6.41563311e-02
7.09073292e-03 -7.41262078e-01 1.32389939e+00 4.35281724e-01
6.54431403e-01 -5.70989311e-01 6.20788693e-01 7.18135357e-01
-1.25724828e+00 -4.92839105e-02 -4.86210972e-01 -1.52289435e-01
-1.58409372e-01 5.92094958e-01 -1.28625345e+00 2.43554950e-01
1.06896996e+00 3.46438318e-01 -8.25719953e-01 1.19451237e+00
-1.74959868e-01 1.88078731e-01 -7.50085711e-01 -2.30455115e-01
2.34676674e-01 -2.21628785e-01 8.52833390e-01 7.88644016e-01
3.33957434e-01 1.47086933e-01 1.98243991e-01 1.14463556e+00
2.89250165e-01 -2.13381164e-02 -7.43442893e-01 5.91548204e-01
2.21745282e-01 1.12915409e+00 -9.01777744e-01 -2.31461897e-01
-4.48474199e-01 5.87806821e-01 2.83069551e-01 -2.52279378e-02
-4.37385738e-01 -2.61377364e-01 6.56656742e-01 5.02766669e-01
8.34684551e-01 -5.14964998e-01 -6.38671517e-01 -7.02517450e-01
4.27359015e-01 -3.21084052e-01 -1.29981801e-01 -9.44490612e-01
-8.73880386e-01 7.05179632e-01 2.04713881e-01 -1.42539632e+00
3.40332799e-02 -1.17292595e+00 -1.75690323e-01 9.16852951e-01
-1.82301271e+00 -1.03761351e+00 -5.39685071e-01 2.98229665e-01
4.21488106e-01 2.63830751e-01 6.25576138e-01 -3.30617070e-01
-1.16055004e-01 -4.13534306e-02 -1.60825953e-01 -1.33982688e-01
2.95345783e-01 -1.23311925e+00 4.94849771e-01 6.78159893e-01
-2.77888160e-02 1.31190091e-01 6.76397085e-01 -8.10019732e-01
-1.73276174e+00 -8.22224259e-01 5.00238657e-01 -1.08119857e+00
2.40084156e-01 -9.10371482e-01 -7.19298840e-01 2.30326518e-01
-6.02440000e-01 3.02481622e-01 1.53224654e-02 -1.22270323e-01
-5.33306658e-01 -3.05304825e-01 -1.18580413e+00 1.51092872e-01
1.14878154e+00 -1.94980383e-01 -6.43542290e-01 8.62690657e-02
7.50052512e-01 -8.38067770e-01 -4.49305326e-01 7.88996398e-01
5.91360033e-01 -1.13064897e+00 1.15718067e+00 -2.90580273e-01
3.28347608e-02 -6.77569866e-01 -1.69388920e-01 -7.58575618e-01
2.18005985e-01 -3.92173469e-01 -1.33427098e-01 4.54921603e-01
3.26435000e-01 -4.54776943e-01 1.14748669e+00 5.55272043e-01
-4.89381939e-01 -6.57161593e-01 -1.36471641e+00 -7.42451668e-01
-3.15263838e-01 -9.04226482e-01 5.96082211e-01 -1.20652048e-02
-4.30316865e-01 1.14736795e-01 9.92417783e-02 4.36218441e-01
8.74613225e-01 6.46971524e-01 1.20756257e+00 -1.70247364e+00
1.66126370e-01 -3.81337136e-01 -7.94686973e-01 -1.18813968e+00
-8.03404152e-02 -4.21695232e-01 5.21717429e-01 -1.35994196e+00
6.70494139e-02 -5.88477671e-01 2.02513412e-02 2.70059168e-01
6.75628185e-02 4.65256244e-01 2.48087138e-01 1.05268419e-01
-6.18899763e-01 5.11408865e-01 1.16722178e+00 -1.31460978e-02
-2.02605411e-01 1.27901137e-01 -2.68661916e-01 9.11079168e-01
4.72604394e-01 -6.47402823e-01 2.11713091e-01 -1.58152655e-01
8.70678723e-02 -2.08902642e-01 6.50522947e-01 -9.79088545e-01
1.19602725e-01 3.10861021e-02 5.30323386e-01 -1.44457173e+00
1.02551234e+00 -1.21318603e+00 -3.52789164e-01 4.66607004e-01
1.92811236e-01 -3.25449139e-01 3.13387483e-01 6.08449221e-01
2.15677440e-01 -3.27778846e-01 8.62104118e-01 -3.05343747e-01
-7.47551322e-01 3.16876411e-01 -1.18263729e-01 -3.45391542e-01
1.14644718e+00 -7.31910646e-01 3.18074375e-02 7.87007511e-02
-4.51295823e-01 1.04286447e-02 8.49386394e-01 2.90942669e-01
1.00037408e+00 -9.54015970e-01 -5.86045027e-01 6.23236954e-01
4.56921011e-01 5.79503000e-01 -3.46988350e-01 5.16013622e-01
-5.21571636e-01 4.90009993e-01 -9.92169008e-02 -1.08800399e+00
-1.31886208e+00 5.47294199e-01 4.91377264e-01 1.90292820e-01
-8.81418467e-01 8.85428429e-01 3.39083940e-01 -3.69450867e-01
3.84079695e-01 -6.94571614e-01 -3.02306321e-02 -9.69581902e-02
4.33345735e-01 3.95677894e-01 5.48755944e-01 -8.81784618e-01
-5.66376925e-01 1.04285800e+00 1.46418557e-01 -2.23681033e-02
1.38014388e+00 -1.43441483e-01 1.75539672e-01 5.90697885e-01
8.65646362e-01 -1.66291162e-01 -1.71939576e+00 -5.96561544e-02
-3.86312194e-02 -7.62603104e-01 3.73614043e-01 -4.87110198e-01
-6.23120010e-01 1.09801614e+00 8.33473802e-01 1.22762136e-01
8.12827408e-01 2.97607541e-01 3.51736218e-01 4.64552522e-01
7.39648163e-01 -9.60847437e-01 -8.68726969e-02 5.09042978e-01
8.78177762e-01 -1.69066346e+00 3.40328515e-01 -6.59638941e-01
-2.66726673e-01 1.26338911e+00 5.98516166e-01 -3.30068767e-01
5.21999359e-01 2.90299594e-01 -1.82777971e-01 -4.29166466e-01
-3.64685059e-01 -7.11686015e-01 6.82928741e-01 6.49676204e-01
-3.17937762e-01 -6.10837266e-02 2.98941344e-01 3.67089182e-01
-2.38342598e-01 -4.96815503e-01 3.63763720e-01 9.44127321e-01
-9.76478457e-01 -7.62543857e-01 -9.71418917e-01 3.86652768e-01
3.32375504e-02 1.22700922e-01 -5.77327728e-01 1.06624961e+00
4.08461630e-01 7.16216922e-01 2.67096102e-01 -3.49965960e-01
6.04548693e-01 -5.24964882e-03 6.82644904e-01 -8.14395726e-01
-1.10934921e-01 1.35952070e-01 -5.67631051e-02 -8.94052565e-01
-3.62091660e-01 -7.75816083e-01 -1.35985649e+00 2.19818845e-01
-6.84732258e-01 -1.86402977e-01 1.32684922e+00 8.15449059e-01
1.84394076e-01 -3.18759829e-02 5.66309929e-01 -1.46260190e+00
-4.95205700e-01 -6.22372508e-01 -5.33200085e-01 -1.44619206e-02
7.55131364e-01 -1.21189046e+00 -5.08212984e-01 -2.13671193e-01] | [7.682435989379883, -2.6143765449523926] |
0bfb23a8-8077-402f-b043-88b93dbb1c01 | ecml-an-ensemble-cascade-metric-learning | 2007.05720 | null | https://arxiv.org/abs/2007.05720v1 | https://arxiv.org/pdf/2007.05720v1.pdf | ECML: An Ensemble Cascade Metric Learning Mechanism towards Face Verification | Face verification can be regarded as a 2-class fine-grained visual recognition problem. Enhancing the feature's discriminative power is one of the key problems to improve its performance. Metric learning technology is often applied to address this need, while achieving a good tradeoff between underfitting and overfitting plays the vital role in metric learning. Hence, we propose a novel ensemble cascade metric learning (ECML) mechanism. In particular, hierarchical metric learning is executed in the cascade way to alleviate underfitting. Meanwhile, at each learning level, the features are split into non-overlapping groups. Then, metric learning is executed among the feature groups in the ensemble manner to resist overfitting. Considering the feature distribution characteristics of faces, a robust Mahalanobis metric learning method (RMML) with closed-form solution is additionally proposed. It can avoid the computation failure issue on inverse matrix faced by some well-known metric learning approaches (e.g., KISSME). Embedding RMML into the proposed ECML mechanism, our metric learning paradigm (EC-RMML) can run in the one-pass learning manner. Experimental results demonstrate that EC-RMML is superior to state-of-the-art metric learning methods for face verification. And, the proposed ensemble cascade metric learning mechanism is also applicable to other metric learning approaches. | ['Yancheng Wang', 'Joey Tianyi Zhou', 'Zhiguo Cao', 'Fu Xiong', 'Yang Xiao', 'Jianxi Wu'] | 2020-07-11 | null | null | null | null | ['fine-grained-visual-recognition'] | ['computer-vision'] | [-7.26959780e-02 -5.41397393e-01 1.10732829e-02 -5.44729233e-01
-6.71975851e-01 -1.90707505e-01 3.63103658e-01 -3.44187737e-01
-1.18566409e-01 4.69413429e-01 -1.40022859e-02 -1.39517233e-01
-6.22568369e-01 -6.36281013e-01 -1.83040410e-01 -1.01177609e+00
2.89027840e-01 9.45972204e-02 -1.03391752e-01 -1.70094997e-01
3.21102351e-01 5.73037505e-01 -1.40141845e+00 -1.52193550e-02
1.05642664e+00 1.32372844e+00 -8.48729461e-02 1.83920309e-01
-1.23519301e-01 7.07879066e-01 -3.44222009e-01 -5.33804834e-01
2.50973374e-01 -5.99974215e-01 -2.56064653e-01 5.57747968e-02
3.00877601e-01 -4.89309877e-01 -4.44816500e-01 1.06363332e+00
5.54163873e-01 -1.24216238e-02 8.79580021e-01 -1.73642647e+00
-5.76868594e-01 3.07609379e-01 -8.74784589e-01 -1.03927955e-01
1.55389920e-01 -7.42103606e-02 9.22877789e-01 -1.40267360e+00
-6.32307604e-02 1.50225151e+00 6.44223034e-01 4.54110533e-01
-1.08093464e+00 -1.18895304e+00 1.13931127e-01 7.31139541e-01
-1.87208509e+00 -3.93118411e-01 1.08138323e+00 -5.48188627e-01
1.26032472e-01 1.98007107e-01 2.25093603e-01 7.48333097e-01
9.42924917e-02 5.86503625e-01 1.33138025e+00 -1.83124021e-01
2.01581568e-01 4.14395034e-02 -1.55239686e-01 8.71455967e-01
1.77412376e-01 2.48932824e-01 -4.57731336e-01 -2.10500419e-01
5.21921515e-01 2.35047400e-01 -1.71185598e-01 -2.85228521e-01
-9.53876257e-01 8.21941495e-01 4.81885254e-01 2.43195832e-01
8.27406421e-02 -1.13699369e-01 3.51622999e-01 5.02193153e-01
2.82285810e-01 -2.94300392e-02 -3.02276284e-01 3.67366672e-02
-1.00842583e+00 -5.65293431e-02 5.32575846e-01 7.22288966e-01
1.01248622e+00 1.17487498e-02 -1.79764271e-01 8.17327380e-01
6.69761360e-01 4.84318942e-01 5.06911814e-01 -5.58742702e-01
5.36428988e-01 1.02112639e+00 -9.92498323e-02 -1.48666501e+00
-3.15343201e-01 -2.73349434e-01 -1.19727945e+00 2.96738684e-01
2.30475962e-01 -4.04872745e-02 -5.11922121e-01 1.68265390e+00
5.54173052e-01 6.49701297e-01 -2.32553959e-01 7.39590764e-01
6.13035500e-01 5.08165479e-01 -4.69397195e-02 -4.76216495e-01
1.14494276e+00 -4.49165165e-01 -6.11858487e-01 2.85113782e-01
6.43137097e-01 -9.54610467e-01 8.08826685e-01 3.90120149e-01
-2.97628671e-01 -7.18395114e-01 -1.41929710e+00 2.00078547e-01
-2.89058536e-01 4.72269744e-01 4.93092030e-01 8.69758964e-01
-7.84380376e-01 5.98552585e-01 -5.05865157e-01 1.99887548e-02
5.95638216e-01 4.62086588e-01 -4.76188809e-01 -4.02685225e-01
-1.02796984e+00 8.35922599e-01 2.90998667e-01 5.09537578e-01
-8.20333838e-01 -6.02031827e-01 -8.42361569e-01 -1.41003802e-01
2.17090428e-01 -2.14828968e-01 9.09949481e-01 -5.22840679e-01
-1.39526689e+00 4.61740375e-01 -1.15477957e-01 2.26853132e-01
5.36774039e-01 2.26584852e-01 -6.58454955e-01 -1.39099315e-01
-1.71563357e-01 5.71997873e-02 1.10154021e+00 -1.07017982e+00
-5.20825207e-01 -6.53474927e-01 -1.37461409e-01 4.45345901e-02
-8.40077519e-01 1.68200836e-01 7.27870166e-02 -6.95375144e-01
4.19910312e-01 -6.46773636e-01 4.77911532e-02 3.21425617e-01
-1.15649626e-01 -4.95688856e-01 1.18785632e+00 -8.13580453e-01
1.68743122e+00 -2.09675097e+00 2.40452617e-01 6.05334103e-01
2.84974039e-01 2.83366323e-01 -2.09631309e-01 2.61782736e-01
-6.17993064e-02 2.71428898e-02 -4.54650104e-01 -2.07685098e-01
1.17402703e-01 -5.23710959e-02 6.20572045e-02 6.83580935e-01
2.03859732e-01 5.89885712e-01 -7.08577394e-01 -8.61420393e-01
3.19575369e-01 4.97695744e-01 -3.90466481e-01 3.64171237e-01
2.77278006e-01 3.81204158e-01 -5.01410186e-01 9.31307375e-01
1.14163542e+00 7.10715577e-02 5.02985194e-02 -6.20355666e-01
3.43937576e-02 -3.40581328e-01 -1.67469823e+00 1.58254194e+00
-5.88547468e-01 1.20092884e-01 -6.16269745e-02 -1.08176959e+00
1.25169945e+00 2.15161502e-01 6.40347421e-01 -4.52798665e-01
2.51908809e-01 3.63894790e-01 2.18084753e-01 -3.64786267e-01
-1.97733298e-01 -9.71584916e-02 4.28813957e-02 6.18750274e-01
8.23720396e-02 2.36105904e-01 -9.12351161e-03 -6.15575574e-02
7.71847606e-01 7.54908547e-02 3.96553010e-01 -1.93993673e-01
1.24604440e+00 -6.43876851e-01 9.56766248e-01 1.00996569e-01
-3.83563429e-01 3.23425800e-01 1.84399545e-01 -2.59748280e-01
-7.49402344e-01 -9.52640712e-01 -2.87274212e-01 9.14411545e-01
2.72761166e-01 -4.82956737e-01 -7.41739690e-01 -8.42313290e-01
1.66622549e-01 5.33948168e-02 -7.00285852e-01 -6.16677403e-01
-5.51620543e-01 -8.44139040e-01 5.05677223e-01 5.65708220e-01
9.29556668e-01 -7.34754860e-01 7.27219954e-02 2.64309287e-01
-2.95793125e-03 -6.83006763e-01 -8.11418414e-01 -3.83872539e-01
-6.93543375e-01 -1.44053543e+00 -4.49430943e-01 -8.23176801e-01
7.26736844e-01 5.07783294e-01 5.34050405e-01 3.52830470e-01
-3.30204129e-01 1.13581307e-01 -3.83785486e-01 -6.88656718e-02
-2.39007205e-01 -2.14395851e-01 4.02300864e-01 7.72717953e-01
6.78810239e-01 -6.97817147e-01 -8.65895927e-01 7.15108454e-01
-7.51549304e-01 -3.41731668e-01 7.56825864e-01 1.18456531e+00
4.47673261e-01 4.95480970e-02 8.63830209e-01 -2.91194230e-01
4.69894946e-01 -4.36439544e-01 -6.76988363e-01 6.13874912e-01
-7.99906254e-01 1.01270853e-02 7.67544329e-01 -5.20106673e-01
-9.80676711e-01 -3.77321690e-02 -7.71757914e-03 -3.75502080e-01
2.71748871e-01 4.11563069e-01 -8.95000160e-01 -6.65116310e-01
4.37827915e-01 3.60095918e-01 1.49114385e-01 -4.85450447e-01
2.88020015e-01 1.00859141e+00 1.24525391e-01 -6.44486845e-01
1.29521918e+00 1.75391376e-01 3.11626434e-01 -4.82343704e-01
-6.81845427e-01 -3.55592877e-01 -6.64967358e-01 -4.21210408e-01
6.07511044e-01 -6.93091810e-01 -9.90666389e-01 6.49744451e-01
-9.13389325e-01 3.12878311e-01 3.11133176e-01 5.20717323e-01
-5.04239738e-01 4.26746964e-01 -2.96240985e-01 -9.17860746e-01
-4.03965592e-01 -1.09032154e+00 7.60597765e-01 5.57226777e-01
1.77035034e-01 -8.70787084e-01 -3.18140611e-02 2.19503939e-01
3.57179433e-01 3.34363133e-01 7.78646588e-01 -6.42182648e-01
-2.60537118e-01 -1.91505373e-01 -4.84718800e-01 7.03678846e-01
4.84344304e-01 6.67970628e-02 -9.72011864e-01 -5.35703480e-01
2.11247727e-01 -2.56167501e-01 5.77815771e-01 -1.62089929e-01
1.38376510e+00 -4.29881454e-01 -2.67113209e-01 6.90543890e-01
1.36220729e+00 2.46027127e-01 3.99393916e-01 1.27714142e-01
8.87932122e-01 4.99636859e-01 8.98194969e-01 6.50793016e-01
4.12667096e-01 6.98858619e-01 2.90986598e-01 7.73053430e-03
6.51424006e-02 -4.98304695e-01 3.67070019e-01 1.00536621e+00
1.59135774e-01 3.65829229e-01 -6.52645648e-01 -5.60648181e-02
-1.84723723e+00 -9.72719550e-01 1.67234659e-01 2.30870485e+00
7.55080223e-01 -2.50627846e-01 8.17710981e-02 8.10711563e-01
8.97927284e-01 9.83597413e-02 -6.35496616e-01 -2.29520962e-01
-4.24837284e-02 1.05586492e-01 -8.40304047e-02 2.41321921e-01
-1.19685113e+00 6.26841009e-01 4.32139730e+00 1.19838655e+00
-9.39515352e-01 2.13889793e-01 5.82149565e-01 3.91620785e-01
-5.80685996e-02 -7.45697841e-02 -7.17527986e-01 7.95454144e-01
5.02055824e-01 -3.37913215e-01 5.96571326e-01 8.03391933e-01
2.86183059e-01 3.67188394e-01 -1.16630268e+00 1.56869578e+00
2.75921226e-01 -8.41803014e-01 9.38968360e-02 3.31284404e-02
6.20140672e-01 -6.95568681e-01 1.02439538e-01 4.18928772e-01
-6.18684404e-02 -1.11008060e+00 2.50882030e-01 5.55165112e-01
1.09143353e+00 -1.09429681e+00 7.50194311e-01 2.79097050e-01
-1.70295727e+00 -4.25819606e-01 -6.14472926e-01 2.46583819e-01
-1.54567793e-01 7.51568794e-01 -4.07791942e-01 7.80252457e-01
4.73701954e-01 8.90656769e-01 -7.76422679e-01 1.29833663e+00
-1.44428775e-01 3.46728295e-01 -1.15764886e-01 7.45536909e-02
-6.71361610e-02 -5.43618262e-01 2.30165303e-01 8.82722139e-01
6.23487175e-01 2.59545058e-01 2.37958208e-01 4.45903629e-01
-9.05045792e-02 4.84861404e-01 -4.65645760e-01 2.61405110e-01
7.32313097e-01 1.44239056e+00 -2.11411789e-01 6.03436977e-02
-5.07699132e-01 9.15424824e-01 3.27324510e-01 1.17763259e-01
-8.78671467e-01 -6.78876579e-01 9.11673486e-01 -2.18452483e-01
1.56220421e-01 -1.06883638e-01 -2.01856315e-01 -1.07929695e+00
3.03147137e-01 -1.01827669e+00 5.99469721e-01 -2.34111249e-01
-1.58270538e+00 3.81647915e-01 -3.72615367e-01 -1.76544321e+00
3.52812372e-02 -7.20170856e-01 -7.59188533e-01 8.51189315e-01
-1.60257602e+00 -1.34995508e+00 -5.60563624e-01 9.58877146e-01
1.39711633e-01 -2.91603237e-01 8.14419329e-01 5.97597122e-01
-6.92611575e-01 1.25690782e+00 1.50196150e-01 2.75177509e-01
8.45651686e-01 -9.98682737e-01 -2.92299420e-01 7.95011044e-01
1.90368995e-01 5.14177680e-01 5.18913642e-02 -3.09121192e-01
-1.48667896e+00 -1.24741018e+00 6.62584782e-01 -1.58495709e-01
6.00252509e-01 -3.17988455e-01 -8.91006649e-01 1.39210939e-01
-3.84391129e-01 1.75888419e-01 1.01480317e+00 -7.30373263e-02
-8.85449171e-01 -8.87551665e-01 -1.42222548e+00 4.38589633e-01
1.02768576e+00 -7.23923266e-01 -4.02596623e-01 1.92305788e-01
3.66488427e-01 -1.11712329e-01 -1.11253595e+00 6.96152329e-01
7.17081726e-01 -9.43141341e-01 6.99541628e-01 -4.65285808e-01
8.80723670e-02 -8.42515826e-01 -3.06030959e-01 -1.30780184e+00
-3.32456142e-01 -5.38395107e-01 -5.64194441e-01 1.68463004e+00
5.91170602e-02 -7.04516053e-01 5.57918966e-01 3.35805148e-01
8.06539506e-02 -9.35853243e-01 -1.23742223e+00 -9.19730961e-01
-3.70458141e-02 -1.35796368e-01 8.56454730e-01 8.36112916e-01
-6.32428601e-02 6.57823961e-03 -4.26304460e-01 1.30123019e-01
9.29972649e-01 1.45945579e-01 6.89152479e-01 -1.41789687e+00
-8.95377696e-02 -6.06503725e-01 -8.55649769e-01 -8.18998992e-01
2.68316507e-01 -1.02328157e+00 -8.26970562e-02 -1.17820001e+00
3.36547166e-01 -6.05333865e-01 -8.14804018e-01 3.85037005e-01
-4.30065036e-01 3.38890404e-01 1.53528035e-01 1.09437190e-01
-3.51199538e-01 9.05975997e-01 1.24708164e+00 -3.20651144e-01
1.01227455e-01 4.22098190e-02 -8.43730152e-01 4.29300904e-01
7.83851027e-01 -3.63967299e-01 -5.08413076e-01 -1.17486633e-01
-2.26887241e-01 -2.43802771e-01 1.23259120e-01 -1.11792564e+00
3.70467901e-01 -2.23369807e-01 6.35010242e-01 -4.45562422e-01
9.87326577e-02 -9.54317808e-01 -3.06630302e-02 4.72865552e-01
1.60449848e-01 1.68950021e-01 -1.37870505e-01 5.68163931e-01
-4.09052372e-01 9.04624909e-02 9.54160810e-01 3.31958860e-01
-5.12825429e-01 9.50470865e-01 3.49840939e-01 -1.12691864e-01
1.19360483e+00 -2.83339500e-01 -1.10321723e-01 1.51580349e-02
-4.49207395e-01 3.67299348e-01 9.87269804e-02 4.98431474e-01
9.65289831e-01 -1.91697454e+00 -1.02175033e+00 2.49894068e-01
3.35381657e-01 -3.11241984e-01 3.82544070e-01 8.45840812e-01
-1.69845745e-02 7.21144006e-02 -7.87881315e-02 -6.34386957e-01
-1.21329296e+00 4.93217111e-01 3.98701042e-01 -1.87843770e-01
-8.49304795e-02 7.12799370e-01 -9.79559869e-02 -4.80234593e-01
2.27011114e-01 2.22111613e-01 -2.46752858e-01 3.00856858e-01
8.51002038e-01 6.71070993e-01 2.00776324e-01 -7.33400047e-01
-5.75559378e-01 1.02252400e+00 -1.57393306e-01 1.79934815e-01
1.02674675e+00 -4.87333983e-02 -3.49354863e-01 1.00604713e-01
1.55473053e+00 -2.17130497e-01 -1.25342929e+00 -3.11111659e-01
2.47276679e-01 -6.81831539e-01 -9.33340415e-02 -4.90349978e-01
-1.11849737e+00 8.26428592e-01 8.97237897e-01 -3.78910869e-01
1.41479087e+00 -4.22275126e-01 6.00047171e-01 2.48246431e-01
5.51430881e-01 -9.69685972e-01 2.82976419e-01 2.35661909e-01
9.49359298e-01 -1.52931273e+00 1.31833047e-01 -2.71693647e-01
-1.50021225e-01 1.22328401e+00 9.85668063e-01 -4.35755104e-02
1.05380678e+00 2.17308234e-02 -1.51481358e-02 2.13727236e-01
-3.26549768e-01 4.66229729e-02 4.46631610e-01 5.41527867e-01
2.53432781e-01 4.37452942e-02 -5.78255773e-01 8.34255040e-01
-1.02508999e-01 -1.58399582e-01 9.08794906e-03 5.98575950e-01
-2.51112998e-01 -1.30262053e+00 -3.54297400e-01 4.52201128e-01
1.22607008e-01 7.27885589e-02 -2.24760354e-01 5.59816599e-01
4.41358387e-01 1.36174655e+00 -3.26702118e-01 -1.00269234e+00
2.32439265e-01 -5.46431839e-02 5.06894648e-01 -3.46792459e-01
-2.93220431e-01 -4.04689789e-01 -5.48850596e-01 -7.32275486e-01
-2.93061107e-01 -5.87882996e-01 -9.58552361e-01 -3.57427090e-01
-4.11524594e-01 1.49354950e-01 5.23237944e-01 1.03021598e+00
1.08249880e-01 8.67696702e-02 1.58782303e+00 -5.37923574e-01
-1.05312693e+00 -1.08435106e+00 -5.24751008e-01 3.22493494e-01
2.49026135e-01 -9.74082053e-01 -5.38043201e-01 -2.76477814e-01] | [13.12364387512207, 0.6458603739738464] |
16e463ee-c477-42f5-8841-55a23963f321 | evading-forensic-classifiers-with-attribute-1 | 2306.13091 | null | https://arxiv.org/abs/2306.13091v1 | https://arxiv.org/pdf/2306.13091v1.pdf | Evading Forensic Classifiers with Attribute-Conditioned Adversarial Faces | The ability of generative models to produce highly realistic synthetic face images has raised security and ethical concerns. As a first line of defense against such fake faces, deep learning based forensic classifiers have been developed. While these forensic models can detect whether a face image is synthetic or real with high accuracy, they are also vulnerable to adversarial attacks. Although such attacks can be highly successful in evading detection by forensic classifiers, they introduce visible noise patterns that are detectable through careful human scrutiny. Additionally, these attacks assume access to the target model(s) which may not always be true. Attempts have been made to directly perturb the latent space of GANs to produce adversarial fake faces that can circumvent forensic classifiers. In this work, we go one step further and show that it is possible to successfully generate adversarial fake faces with a specified set of attributes (e.g., hair color, eye size, race, gender, etc.). To achieve this goal, we leverage the state-of-the-art generative model StyleGAN with disentangled representations, which enables a range of modifications without leaving the manifold of natural images. We propose a framework to search for adversarial latent codes within the feature space of StyleGAN, where the search can be guided either by a text prompt or a reference image. We also propose a meta-learning based optimization strategy to achieve transferable performance on unknown target models. Extensive experiments demonstrate that the proposed approach can produce semantically manipulated adversarial fake faces, which are true to the specified attribute set and can successfully fool forensic face classifiers, while remaining undetectable by humans. Code: https://github.com/koushiksrivats/face_attribute_attack. | ['Karthik Nandakumar', 'Koushik Srivatsan', 'Fahad Shamshad'] | 2023-06-22 | evading-forensic-classifiers-with-attribute | http://openaccess.thecvf.com//content/CVPR2023/html/Shamshad_Evading_Forensic_Classifiers_With_Attribute-Conditioned_Adversarial_Faces_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Shamshad_Evading_Forensic_Classifiers_With_Attribute-Conditioned_Adversarial_Faces_CVPR_2023_paper.pdf | cvpr-2023-1 | ['meta-learning'] | ['methodology'] | [ 5.94686568e-01 3.61279756e-01 1.42221510e-01 -1.65322363e-01
-7.51185954e-01 -1.04791057e+00 6.99187994e-01 -7.19876230e-01
2.23276138e-01 6.30389631e-01 -4.19959992e-01 -3.58265609e-01
2.93604821e-01 -9.24571455e-01 -9.19668913e-01 -8.90863597e-01
1.96040615e-01 3.40465367e-01 -3.11661750e-01 -8.84915888e-02
1.40412256e-01 7.42629290e-01 -1.50552750e+00 3.56131285e-01
6.94348752e-01 6.07541680e-01 -4.91763115e-01 7.38373816e-01
1.73533484e-01 6.27479613e-01 -9.90250170e-01 -1.03264093e+00
6.55108869e-01 -7.43704855e-01 -5.34902096e-01 4.14178148e-02
7.56400466e-01 -5.64765871e-01 -5.24584711e-01 1.44016886e+00
4.13525611e-01 -4.63940173e-01 7.28700101e-01 -1.79673660e+00
-9.35713232e-01 3.55084300e-01 -4.63426620e-01 -1.74651399e-01
3.99885625e-01 8.30117822e-01 4.78850096e-01 -7.37243593e-01
5.23243725e-01 1.60057652e+00 3.37390006e-01 1.07213581e+00
-1.22578359e+00 -1.34313977e+00 -4.35165137e-01 5.03512146e-03
-1.27257729e+00 -8.51503909e-01 9.62493718e-01 -4.03413087e-01
5.53962030e-02 4.31776404e-01 4.33412820e-01 1.89970672e+00
1.46465778e-01 4.16053593e-01 1.58653498e+00 -2.76515454e-01
1.10895805e-01 2.77829438e-01 -5.51460743e-01 9.28517401e-01
5.06826103e-01 5.89292824e-01 -4.31825131e-01 -5.46904504e-01
7.56066859e-01 -1.84638977e-01 -4.54299957e-01 -2.43154377e-01
-7.76853502e-01 1.09340751e+00 1.47222579e-01 -9.05886516e-02
-1.32778272e-01 2.74027556e-01 1.23905726e-01 3.96265149e-01
2.03256831e-01 6.55845106e-01 -9.44076031e-02 2.04582915e-01
-8.29878092e-01 1.77601337e-01 7.46908009e-01 7.19256461e-01
6.35978341e-01 4.76096183e-01 -1.27499208e-01 3.32378775e-01
1.94188640e-01 9.84906614e-01 1.82141691e-01 -9.23819125e-01
1.79835454e-01 2.84081846e-01 -4.41899104e-03 -1.26882672e+00
3.71231347e-01 5.36138080e-02 -7.04134881e-01 6.60256386e-01
5.59979737e-01 -1.82264224e-01 -1.02964056e+00 1.75645792e+00
4.78739381e-01 3.64194602e-01 1.57446191e-01 6.44952416e-01
5.59259295e-01 4.82851803e-01 -1.14697501e-01 -4.00245823e-02
1.26378942e+00 -4.33522493e-01 -6.14513397e-01 -3.07736754e-01
3.24567884e-01 -7.86923051e-01 1.00493062e+00 2.39014566e-01
-9.33892608e-01 -2.52948552e-01 -1.09861910e+00 3.76494408e-01
-3.92339617e-01 1.48622720e-02 5.68459511e-01 1.53938711e+00
-8.81944060e-01 4.44796264e-01 -7.13165045e-01 9.02778748e-03
9.60858166e-01 5.44913054e-01 -5.34400702e-01 -2.62047708e-01
-1.07252514e+00 6.47445858e-01 7.17807561e-02 2.00561583e-01
-1.40197766e+00 -5.88663816e-01 -8.34071100e-01 -5.80433384e-02
4.28034306e-01 -4.58920658e-01 6.42646194e-01 -1.38449717e+00
-1.61396515e+00 1.10585582e+00 8.90168846e-02 -3.06134790e-01
9.07768428e-01 1.28007531e-01 -6.58142805e-01 3.20674121e-01
-1.73575524e-02 5.90460598e-01 1.71009433e+00 -1.55078971e+00
-2.50314847e-02 -4.21912760e-01 8.00559074e-02 -3.62319261e-01
-4.63495731e-01 2.63774365e-01 1.01170890e-01 -6.82817400e-01
-4.43968505e-01 -1.30719960e+00 1.63290203e-01 2.20441595e-01
-8.64577293e-01 4.85369951e-01 1.18304634e+00 -6.47105634e-01
5.50881147e-01 -2.03616714e+00 -1.47787973e-01 2.49364570e-01
3.26505899e-01 5.75838566e-01 -2.56202459e-01 1.89013761e-02
-1.98090851e-01 6.39133096e-01 -3.17645520e-01 -2.03443274e-01
-1.96764380e-01 5.90142980e-02 -6.50904715e-01 9.24432933e-01
5.20603955e-01 9.59507763e-01 -7.57748365e-01 -3.82403880e-01
-1.32634817e-02 8.12119246e-01 -5.53012311e-01 3.63533646e-01
-7.03913942e-02 7.41089761e-01 -3.52377146e-01 9.40694809e-01
1.00907433e+00 1.36149824e-01 1.70543253e-01 -4.74059694e-02
5.42720377e-01 -4.19824496e-02 -7.88866222e-01 7.84617662e-01
-2.49607518e-01 6.36125028e-01 1.98121876e-01 -7.25071907e-01
9.43627596e-01 2.92078227e-01 1.51978120e-01 -2.12922961e-01
4.06838596e-01 1.17081575e-01 1.80845976e-01 -3.30224395e-01
8.30700919e-02 -2.01765403e-01 -1.29375039e-02 5.43352902e-01
8.19399022e-03 -1.07595302e-01 -3.43694329e-01 -7.60782659e-02
1.13202047e+00 9.32247192e-02 -6.93686455e-02 5.49463034e-02
5.22809625e-01 -1.97374642e-01 4.68283206e-01 7.53562689e-01
-2.54628450e-01 5.78201354e-01 7.06912279e-01 -2.97371358e-01
-1.13819814e+00 -1.08938527e+00 1.22044064e-01 6.33058310e-01
-1.45334443e-02 2.86574364e-02 -9.88268435e-01 -1.05771542e+00
6.46649525e-02 5.63342571e-01 -8.85354817e-01 -6.50434613e-01
-6.23337388e-01 -5.65104544e-01 1.25614083e+00 2.35726219e-02
5.05346417e-01 -9.84348655e-01 -4.71533000e-01 -3.72484386e-01
1.77307025e-01 -1.20099378e+00 -3.33912075e-01 -3.39904457e-01
-4.44107324e-01 -1.40018165e+00 -3.60691220e-01 -3.26120019e-01
1.03147399e+00 3.95102501e-02 7.64937818e-01 4.51667994e-01
-3.74958485e-01 4.36775118e-01 -2.02060238e-01 -4.74760652e-01
-1.16738749e+00 -1.94093883e-01 1.66350022e-01 4.58605587e-01
1.07228234e-01 -4.71974730e-01 -4.55457181e-01 4.23517734e-01
-1.08925438e+00 -1.58444613e-01 3.97257805e-01 9.17002439e-01
6.07290827e-02 -1.61722135e-02 4.84502584e-01 -1.19778204e+00
5.35567462e-01 -4.66854930e-01 -8.27318728e-01 3.23419571e-01
-4.80583102e-01 -6.59243539e-02 8.66910934e-01 -8.83608222e-01
-8.64018381e-01 1.04383724e-02 6.73346594e-02 -9.97998416e-01
-3.01257104e-01 -1.69990003e-01 -6.54656827e-01 -5.70440829e-01
8.26687574e-01 3.19864541e-01 4.06802863e-01 -3.10512409e-02
4.28026885e-01 5.61497271e-01 5.04891634e-01 -6.53043687e-01
1.55346990e+00 7.32188821e-01 1.54397786e-01 -7.03454673e-01
-5.70971012e-01 4.01173413e-01 -4.57249552e-01 -3.67972016e-01
5.52458584e-01 -6.97952390e-01 -7.76685178e-01 6.14423931e-01
-1.13665783e+00 -1.16263002e-01 2.42724251e-02 -1.32018954e-01
-4.05770183e-01 4.24259275e-01 -3.59061241e-01 -8.85397851e-01
-1.56117216e-01 -1.31655264e+00 9.61289346e-01 6.31419718e-02
-9.01589245e-02 -8.03646684e-01 -3.60455513e-01 6.30625367e-01
3.84252042e-01 8.20584118e-01 7.21566916e-01 -6.92910790e-01
-6.78359628e-01 -4.37161714e-01 6.64894506e-02 4.22691613e-01
4.32288766e-01 2.52807826e-01 -1.19697106e+00 -5.48377335e-01
1.49910435e-01 -4.49585050e-01 4.87604141e-01 -1.63699239e-01
1.17238271e+00 -8.43172252e-01 -2.49585301e-01 1.00586116e+00
1.06333053e+00 2.13913620e-01 8.72553408e-01 7.18985572e-02
9.06086862e-01 6.22391343e-01 1.88042343e-01 1.79728985e-01
-1.53127819e-01 5.19680619e-01 7.12385654e-01 2.47116815e-02
-2.07252935e-01 -4.28241700e-01 7.31789827e-01 -9.66329575e-02
1.65294692e-01 -3.48976165e-01 -6.57571256e-01 3.18101019e-01
-1.22551394e+00 -1.16924202e+00 1.00587390e-01 2.25545192e+00
6.69728875e-01 6.16292842e-02 -4.60779807e-03 1.08721860e-01
9.53677416e-01 8.07639435e-02 -7.39662528e-01 -5.39301336e-01
-2.18649015e-01 4.43382293e-01 6.56456590e-01 3.03175181e-01
-9.95271266e-01 1.28324568e+00 5.62519026e+00 6.58954203e-01
-1.48957336e+00 2.13418975e-01 8.35990012e-01 -2.06165180e-01
-4.89022940e-01 8.17217827e-02 -5.89174807e-01 6.77604556e-01
8.91353428e-01 -1.03995472e-01 7.41915941e-01 8.03607047e-01
-5.96382283e-02 4.17294770e-01 -1.01110756e+00 1.00596428e+00
3.96323889e-01 -1.35400987e+00 2.62750268e-01 5.31539977e-01
6.78932428e-01 -5.98231733e-01 6.22218549e-01 1.21143073e-01
4.19810027e-01 -1.55733180e+00 6.81564331e-01 2.13282883e-01
1.35106611e+00 -8.83460820e-01 3.41856569e-01 1.45600706e-01
-4.94659156e-01 -1.32988706e-01 -3.71623307e-01 3.65887672e-01
-1.96002230e-01 2.17604503e-01 -9.82925653e-01 1.64184317e-01
3.53303730e-01 1.22367904e-01 -6.29353881e-01 4.19905931e-01
-6.07819259e-01 8.58315825e-01 -2.58288741e-01 2.97814518e-01
-1.65538210e-02 -6.75359070e-02 7.06850708e-01 7.32875764e-01
5.32503843e-01 3.51698301e-03 -2.13456631e-01 1.39034998e+00
-4.25447643e-01 -2.21630007e-01 -1.13033438e+00 -3.22122395e-01
5.26321828e-01 1.18748891e+00 -5.40354908e-01 -5.79361953e-02
1.11750819e-01 1.17991102e+00 7.21777976e-02 2.99538434e-01
-1.00073481e+00 -7.61639848e-02 8.78658354e-01 2.51967371e-01
2.14926064e-01 6.96116686e-02 -1.85448542e-01 -1.31320488e+00
-1.50789591e-02 -1.40808070e+00 1.32855922e-01 -5.13141274e-01
-1.24989998e+00 6.64452791e-01 -2.21247330e-01 -9.82238412e-01
-3.16213369e-01 -6.30039394e-01 -6.88077033e-01 8.21783423e-01
-1.29493225e+00 -1.66603100e+00 -2.10031122e-01 7.84101844e-01
1.37750804e-01 -5.77723086e-01 9.03738260e-01 -3.72231267e-02
-5.65587282e-01 1.06640553e+00 -1.25018463e-01 4.83320028e-01
7.73837328e-01 -8.19239378e-01 4.74811256e-01 1.00837743e+00
1.56158417e-01 6.05398357e-01 8.38984549e-01 -7.66124308e-01
-1.73909032e+00 -1.15831947e+00 1.40194118e-01 -6.55660987e-01
4.71058667e-01 -6.75859332e-01 -8.08315456e-01 7.17405975e-01
-4.15227638e-04 3.59253913e-01 7.59814322e-01 -4.52560157e-01
-8.40641737e-01 1.20048195e-01 -1.75278521e+00 6.94912553e-01
8.63346636e-01 -7.23454654e-01 -1.85990721e-01 3.94528180e-01
3.30379039e-01 -2.17051208e-01 -4.54045027e-01 2.47621819e-01
6.34854615e-01 -1.11196721e+00 1.07840848e+00 -7.97989428e-01
5.51738977e-01 -2.61829197e-01 5.65156378e-02 -1.17327559e+00
1.53114125e-01 -1.05298340e+00 -2.62893945e-01 1.44420207e+00
2.17121124e-01 -8.65510523e-01 9.61236596e-01 5.86169660e-01
4.63155389e-01 -4.51807380e-01 -1.00229084e+00 -9.49469149e-01
3.04241359e-01 -2.38701627e-01 9.19166028e-01 1.19180906e+00
-4.73462552e-01 -1.79932743e-01 -8.60780895e-01 5.77537417e-01
9.97332513e-01 -3.53798158e-02 1.07720006e+00 -1.01768637e+00
-2.27788106e-01 -2.92703897e-01 -6.04792893e-01 -3.54948819e-01
7.60714471e-01 -7.93280184e-01 -2.84665734e-01 -5.13517618e-01
3.19111869e-02 -3.94109040e-01 1.50552258e-01 6.21983051e-01
-5.27830757e-02 6.22760773e-01 2.88097233e-01 2.26960227e-01
1.80003583e-01 3.32176596e-01 1.20373225e+00 -3.12776953e-01
1.78973407e-01 -8.50823745e-02 -9.24944639e-01 5.63142836e-01
9.15367305e-01 -8.83329332e-01 -2.60156065e-01 -1.91343114e-01
8.38468522e-02 1.81922968e-02 8.10710907e-01 -8.15762103e-01
-7.91959837e-02 -2.73920089e-01 5.12668848e-01 1.85789123e-01
5.69544077e-01 -7.31290221e-01 3.42559904e-01 5.09314358e-01
-1.40124813e-01 -1.52019724e-01 1.19961411e-01 4.96418446e-01
1.25142530e-01 -3.57355252e-02 1.15864980e+00 -2.08232969e-01
6.08709902e-02 4.06558484e-01 -1.38532817e-01 -8.53357837e-02
1.19117010e+00 -3.14753234e-01 -5.69620848e-01 -5.28648317e-01
-3.42799067e-01 -1.99083239e-01 9.98724937e-01 4.25511628e-01
7.42581189e-01 -1.23385286e+00 -8.62722397e-01 3.89973193e-01
4.29133028e-02 -4.72588539e-01 4.34085838e-02 2.74390846e-01
-5.05592167e-01 -1.62393019e-01 -5.01739562e-01 -3.84476840e-01
-1.44158757e+00 8.82437766e-01 5.37119448e-01 1.84938654e-01
-2.86049277e-01 6.76863253e-01 1.89605847e-01 -3.91321391e-01
-3.47640157e-01 4.04197127e-01 1.70180216e-01 -1.91836983e-01
5.03825366e-01 4.30024639e-02 -2.60683417e-01 -1.03962660e+00
-2.63381571e-01 3.36578816e-01 8.17188546e-02 1.00291640e-01
1.14908779e+00 2.77534217e-01 -1.05449311e-01 -2.19492540e-01
1.23526192e+00 3.49507362e-01 -1.27003372e+00 2.44809389e-01
-3.61773938e-01 -1.13021803e+00 -2.00178638e-01 -6.52198553e-01
-1.46937907e+00 8.47178400e-01 7.17060626e-01 2.68742442e-01
9.24024045e-01 -1.46276340e-01 7.40488052e-01 3.91822457e-02
3.34618121e-01 -4.60625082e-01 1.89974695e-01 -6.51320145e-02
7.41023481e-01 -1.40254033e+00 -2.07770020e-01 -6.11265421e-01
-4.52016920e-01 1.14812887e+00 6.27608538e-01 -1.37640521e-01
3.17179799e-01 3.06995422e-01 1.03811525e-01 -2.55176574e-01
-3.42877775e-01 2.47562066e-01 8.35120603e-02 9.25502598e-01
-9.21156853e-02 1.58491552e-01 2.42807105e-01 1.91427141e-01
-5.77454090e-01 -3.69983941e-01 8.50551665e-01 7.25613177e-01
1.46134362e-01 -1.29040885e+00 -9.63408649e-01 2.94872314e-01
-8.37848127e-01 7.73982778e-02 -6.81793451e-01 6.41726732e-01
1.89868554e-01 1.05305111e+00 -3.01601231e-01 -3.68749976e-01
-1.31009281e-01 1.41541302e-01 6.54684901e-01 -4.42931145e-01
-5.68561316e-01 -3.14835250e-01 -1.67620137e-01 -4.89345372e-01
-1.26996368e-01 -6.12414420e-01 -5.74042261e-01 -5.62384665e-01
-2.66640306e-01 -1.42783821e-01 6.16398096e-01 8.73378515e-01
3.49972337e-01 2.39330642e-02 1.04296148e+00 -6.63044691e-01
-6.52227461e-01 -6.20825052e-01 -2.93179035e-01 5.08218825e-01
4.32153016e-01 -6.82347536e-01 -7.54560828e-01 2.24775016e-01] | [12.653654098510742, 0.9638271331787109] |
3aee6623-a5ec-4087-a70f-104649beb130 | gca-net-utilizing-gated-context-attention-for | 2112.04298 | null | https://arxiv.org/abs/2112.04298v3 | https://arxiv.org/pdf/2112.04298v3.pdf | GCA-Net : Utilizing Gated Context Attention for Improving Image Forgery Localization and Detection | Forensic analysis of manipulated pixels requires the identification of various hidden and subtle features from images. Conventional image recognition models generally fail at this task because they are biased and more attentive toward the dominant local and spatial features. In this paper, we propose a novel Gated Context Attention Network (GCA-Net) that utilizes non-local attention in conjunction with a gating mechanism in order to capture the finer image discrepancies and better identify forged regions. The proposed framework uses high dimensional embeddings to filter and aggregate the relevant context from coarse feature maps at various stages of the decoding process. This improves the network's understanding of global differences and reduces false-positive localizations. Our evaluation on standard image forensic benchmarks shows that GCA-Net can both compete against and improve over state-of-the-art networks by an average of 4.7% AUC. Additional ablation studies also demonstrate the method's robustness against attributions and resilience to false-positive predictions. | ['Md. Ruhul Amin', 'Md. Saiful Islam', 'Sowmen Das'] | 2021-12-08 | null | null | null | null | ['image-forensics'] | ['computer-vision'] | [ 3.21497083e-01 -2.39483401e-01 1.14659965e-01 -4.33854222e-01
-7.76160002e-01 -4.63326365e-01 6.28159702e-01 3.62996429e-01
-5.33553302e-01 3.86536032e-01 1.61667541e-01 -2.43425921e-01
6.34802580e-02 -6.64742470e-01 -4.92387772e-01 -6.69920266e-01
-1.19469404e-01 2.35611331e-02 3.33394140e-01 9.04599428e-02
7.19953358e-01 7.65224040e-01 -1.31587672e+00 6.94137752e-01
6.29952490e-01 1.12348878e+00 7.72776678e-02 3.98132712e-01
-7.90466461e-03 8.64372015e-01 -8.15074742e-01 -7.12073982e-01
1.78800926e-01 -1.92344770e-01 -5.94237030e-01 -1.28913090e-01
7.77140617e-01 -6.10401213e-01 -3.56770039e-01 1.17775965e+00
5.64165533e-01 -7.18767866e-02 3.47108006e-01 -1.02245986e+00
-1.03887427e+00 2.15026557e-01 -7.49843121e-01 9.76981401e-01
1.36215612e-01 5.51688910e-01 9.25505638e-01 -1.23901749e+00
5.60647190e-01 1.26612055e+00 8.21905553e-01 4.55630779e-01
-1.30197060e+00 -8.83533180e-01 -3.59047242e-02 6.87533319e-01
-1.41506386e+00 -4.24286842e-01 8.40746164e-01 -2.39251122e-01
1.08015287e+00 1.20771319e-01 1.86692894e-01 1.51300037e+00
3.70042771e-01 4.52624947e-01 1.28798747e+00 -3.26225430e-01
2.20960975e-01 -1.64619699e-01 4.39720601e-01 5.94448745e-01
6.43004894e-01 1.05493940e-01 -8.78387153e-01 -2.34837219e-01
6.12779319e-01 6.27192333e-02 -1.07471667e-01 1.35370150e-01
-8.66527617e-01 9.64601696e-01 5.39778650e-01 3.22678775e-01
-6.19048893e-01 2.06570625e-01 3.65753025e-01 -7.33691081e-02
3.37258011e-01 5.47438204e-01 2.55525976e-01 -3.10119782e-02
-1.19692850e+00 4.78208922e-02 2.17920780e-01 2.81655163e-01
7.12292671e-01 2.35085890e-01 -4.83553022e-01 5.93912840e-01
5.90849519e-02 2.28105411e-01 3.40023190e-01 -8.06723058e-01
4.36417669e-01 6.77873135e-01 -2.16973975e-01 -1.42387640e+00
-1.97934434e-01 -5.02851546e-01 -5.27475357e-01 6.11765325e-01
4.24063712e-01 2.69906640e-01 -1.13949835e+00 1.60704172e+00
3.87529247e-02 3.12674791e-01 -3.88898462e-01 1.07560098e+00
4.22808975e-01 2.78422982e-01 3.47173601e-01 2.85133570e-01
1.33108675e+00 -5.40833414e-01 -5.07082939e-01 -5.97082317e-01
1.28715202e-01 -4.73904938e-01 9.09315765e-01 3.07763726e-01
-8.66690934e-01 -6.31742060e-01 -1.06405449e+00 -3.18824470e-01
-5.36246300e-01 -1.21218094e-03 4.66078490e-01 9.66378629e-01
-9.98156309e-01 7.80458629e-01 -5.70836544e-01 -2.09040403e-01
1.14428496e+00 4.59094495e-01 -4.74556774e-01 -1.55420795e-01
-9.10236835e-01 8.71266365e-01 1.44101977e-01 3.51979494e-01
-1.02429569e+00 -3.91366959e-01 -6.56156003e-01 1.88755229e-01
1.81804180e-01 -1.77791610e-01 4.64368194e-01 -8.76935005e-01
-8.68483603e-01 9.38796997e-01 -1.34335876e-01 -6.29436910e-01
4.51159358e-01 -2.68797517e-01 -3.16876501e-01 6.48454487e-01
3.23145539e-01 7.08240211e-01 1.06247413e+00 -1.20874631e+00
-3.15253109e-01 -5.80789149e-01 -6.81399927e-02 -3.51682395e-01
-3.11388075e-01 2.98426688e-01 -1.88423797e-01 -9.57227528e-01
-1.55150592e-01 -6.09749258e-01 2.00716078e-01 2.27662489e-01
-3.35867763e-01 6.73595071e-02 1.32588756e+00 -9.45903122e-01
1.03985929e+00 -2.15821719e+00 -2.68427461e-01 2.46274754e-01
3.92431945e-01 6.23826385e-01 -2.16671944e-01 1.89872324e-01
-2.53587931e-01 3.09306413e-01 -3.90300751e-01 -6.33199036e-01
-1.50709614e-01 1.37768373e-01 -3.93759042e-01 8.30258429e-01
7.50923634e-01 1.04817474e+00 -8.75598729e-01 -4.42576051e-01
3.56916994e-01 6.61395371e-01 -5.67216218e-01 -1.97064634e-02
2.53431529e-01 4.86025780e-01 -7.95321241e-02 1.05236530e+00
6.39887869e-01 -2.01235101e-01 1.07308246e-01 -2.36581638e-01
1.99069932e-01 1.50414005e-01 -7.12490320e-01 1.57383478e+00
-1.63894638e-01 1.13071001e+00 4.60665412e-02 -9.67104137e-01
9.83716309e-01 -1.94151122e-02 -2.89133191e-03 -1.01975799e+00
3.77545029e-01 4.70916070e-02 1.52550280e-01 -5.95253289e-01
6.70307934e-01 -1.05751634e-01 2.13217288e-02 6.01839066e-01
2.06538036e-01 8.64653885e-01 -2.23691180e-01 1.65603653e-01
1.20591962e+00 -7.00721741e-02 -4.68462519e-02 -1.85247108e-01
4.67129827e-01 -4.05678719e-01 7.28968143e-01 9.74374294e-01
-5.88675737e-01 7.18052864e-01 5.66952288e-01 -5.60481131e-01
-8.86735737e-01 -9.83249366e-01 1.67175457e-02 1.01770747e+00
2.41564468e-01 -2.19240502e-01 -8.03746641e-01 -8.68335426e-01
7.24903196e-02 9.46049154e-01 -1.19629300e+00 -3.60580236e-01
-6.30942583e-01 -7.64387190e-01 8.10069263e-01 7.06861496e-01
6.02836728e-01 -1.21203101e+00 -8.10710907e-01 4.15876098e-02
-7.75018707e-02 -1.23128557e+00 -2.12106541e-01 1.24342591e-01
-6.79539025e-01 -1.22024429e+00 -2.12452650e-01 -3.13243717e-01
6.46442413e-01 3.90118450e-01 7.45655954e-01 3.08536202e-01
-5.97533703e-01 1.12599157e-01 -4.46378946e-01 -7.31441453e-02
-2.30997756e-01 -1.33080170e-01 -2.44375631e-01 5.58500826e-01
5.90045512e-01 -6.63476884e-01 -7.58004189e-01 1.90385580e-01
-1.09574616e+00 -3.98731321e-01 6.68523490e-01 1.06649363e+00
2.90586919e-01 -2.07074717e-01 6.06867850e-01 -7.24170983e-01
7.50651181e-01 -7.08903253e-01 -1.36441365e-01 2.20036849e-01
-5.78457773e-01 2.86982898e-02 5.72117865e-01 -3.28673065e-01
-9.78730679e-01 -2.97713101e-01 3.01638812e-01 -7.45559394e-01
-8.56116191e-02 1.01600014e-01 -1.49471983e-01 -5.50286531e-01
8.09979916e-01 1.66610733e-01 -1.43366560e-01 -4.89495844e-01
1.85453221e-01 4.99433279e-01 7.47830570e-01 -5.18027306e-01
6.72998428e-01 7.83308685e-01 6.35141358e-02 -6.54632270e-01
-4.33526307e-01 -2.21222147e-01 -7.88557231e-01 -2.48718977e-01
8.85245562e-01 -6.65070474e-01 -5.77737570e-01 5.17050803e-01
-1.22388685e+00 -3.53281014e-02 -7.31860772e-02 -1.83949098e-02
-5.92629015e-02 7.92541981e-01 -6.11101329e-01 -9.22341526e-01
-5.19664586e-01 -1.22044599e+00 9.67200279e-01 2.81995416e-01
-2.35522971e-01 -7.57146835e-01 -3.39639187e-01 5.35591900e-01
5.39362252e-01 5.37111461e-01 7.18277156e-01 -7.35404789e-01
-5.27044535e-01 -3.66091818e-01 -7.46698201e-01 5.41568100e-01
-8.84073153e-02 -2.62746304e-01 -1.44548094e+00 -2.12257743e-01
-1.31447926e-01 -2.33782172e-01 1.19530475e+00 -2.50773709e-02
1.38187993e+00 -1.84904337e-01 -1.59430102e-01 6.46745265e-01
1.40795386e+00 2.51774061e-02 1.02423537e+00 5.14559507e-01
6.69422150e-01 4.88329142e-01 3.22320461e-01 3.31263691e-01
-6.55402318e-02 5.01210630e-01 8.40903223e-01 1.31550953e-01
-3.36243600e-01 -1.62162170e-01 3.09197426e-01 9.04705301e-02
-1.09417187e-02 -1.38554931e-01 -1.03372800e+00 7.53254950e-01
-1.51347458e+00 -1.22235835e+00 -9.08186361e-02 2.09893107e+00
4.50993627e-01 2.55417377e-01 -5.73518127e-03 1.21715285e-01
1.17197418e+00 3.98734838e-01 -5.25830209e-01 -6.04713976e-01
-2.97564536e-01 6.33633137e-01 6.36498749e-01 1.98827311e-01
-1.04402459e+00 9.00551021e-01 6.36798716e+00 8.72942984e-01
-1.21454096e+00 4.37537760e-01 8.19305599e-01 -2.33959317e-01
3.35150175e-02 -1.86645370e-02 -4.79385525e-01 8.25973809e-01
9.42323864e-01 4.91808295e-01 3.68970484e-01 7.47122705e-01
-1.45271346e-02 -2.25940779e-01 -7.48395443e-01 1.04518878e+00
3.86728436e-01 -1.50630343e+00 1.41389802e-01 3.78628731e-01
3.88776153e-01 -2.50745807e-02 2.70914435e-01 1.01323007e-02
-7.71780238e-02 -1.20667529e+00 8.35345268e-01 4.79437202e-01
7.26651669e-01 -8.18371117e-01 7.88156271e-01 -1.59412652e-01
-7.44181335e-01 -5.46341479e-01 -3.19608241e-01 -1.28897890e-01
5.91192655e-02 4.05934215e-01 -7.85450280e-01 3.10983211e-01
7.61418760e-01 4.38593775e-01 -1.09789026e+00 9.04879630e-01
-3.20806384e-01 7.35344827e-01 -9.47633758e-02 3.36833000e-01
3.91537458e-01 1.77607343e-01 5.68816960e-01 1.52237427e+00
7.04549029e-02 -1.08707733e-01 -2.87385404e-01 1.28684092e+00
-1.63641199e-01 -3.52073282e-01 -3.87396187e-01 5.80087490e-02
5.85239470e-01 1.29122055e+00 -9.55885828e-01 -2.76637584e-01
-1.82784319e-01 1.21479905e+00 4.68467474e-01 1.63015485e-01
-9.41573322e-01 -4.18460965e-01 7.27719963e-01 1.76200464e-01
7.00132549e-01 -2.14112476e-01 -5.53691566e-01 -1.03240120e+00
7.36709386e-02 -7.49585927e-01 4.14085716e-01 -8.06340933e-01
-1.29641044e+00 7.27981567e-01 -1.93401173e-01 -8.57981026e-01
4.59151901e-02 -6.78265095e-01 -9.43715751e-01 8.97537947e-01
-1.68067801e+00 -1.37394273e+00 -4.17457968e-01 6.75769806e-01
2.69275755e-01 -1.76763222e-01 7.08783925e-01 3.63136977e-01
-7.99439728e-01 8.23567152e-01 -1.03930734e-01 5.13660431e-01
7.69735515e-01 -9.63986099e-01 5.39014637e-01 1.41093588e+00
1.05899908e-02 8.63766372e-01 4.41890806e-01 -7.98651516e-01
-9.61754262e-01 -1.00254166e+00 6.39193058e-01 -5.58175504e-01
5.90025365e-01 -3.98067564e-01 -1.09572875e+00 5.00423968e-01
2.12170348e-01 4.21163529e-01 7.26184905e-01 -1.49854198e-01
-1.11212146e+00 -5.75063340e-02 -1.44968808e+00 2.26066664e-01
8.95299911e-01 -9.72748995e-01 -6.34215057e-01 -2.83425361e-01
1.84369311e-02 8.15148130e-02 -5.03470957e-01 1.96621697e-02
5.49701512e-01 -1.39521754e+00 1.15143406e+00 -4.42077130e-01
4.73825425e-01 -1.96356013e-01 -3.06606740e-01 -6.61329687e-01
-6.51022255e-01 -4.91368294e-01 -7.61913061e-02 1.22195804e+00
-1.70509413e-01 -5.49848020e-01 6.82960808e-01 4.58525568e-01
2.25411337e-02 -5.12763202e-01 -1.29801714e+00 -7.48608053e-01
-2.93170754e-02 -5.52096426e-01 3.63858074e-01 1.06865799e+00
-3.31777781e-01 1.25830337e-01 -4.38786983e-01 3.55566412e-01
9.88588512e-01 1.79768484e-02 2.73762703e-01 -1.01372540e+00
-1.30748913e-01 -5.47888041e-01 -9.58045304e-01 -2.09470466e-01
6.56175986e-02 -8.13002288e-01 -2.72957355e-01 -1.03166592e+00
3.14651638e-01 -1.35325536e-01 -5.58946669e-01 6.74107611e-01
-3.82458180e-01 8.14072132e-01 4.81886059e-01 3.14537108e-01
-7.11710632e-01 2.18356177e-01 6.74803436e-01 -1.13831960e-01
3.44580799e-01 -8.37610662e-01 -8.86222899e-01 4.43489879e-01
8.39450657e-01 -6.56388342e-01 2.10623816e-03 -5.33982098e-01
-1.78553522e-01 -5.20908296e-01 1.14076471e+00 -1.22070205e+00
1.46153286e-01 1.37365222e-01 7.83819020e-01 -3.54135305e-01
4.98682708e-01 -6.16928637e-01 -1.40902042e-01 3.31204265e-01
-1.55325875e-01 1.64842680e-01 3.44143897e-01 7.99745977e-01
1.31223574e-02 -2.10611269e-01 7.74973392e-01 -3.97919826e-02
-8.82141411e-01 -5.41703887e-02 -2.15839982e-01 -2.06958786e-01
1.00785983e+00 -6.06238186e-01 -6.88082397e-01 -2.02117145e-01
-6.01611555e-01 -2.65124023e-01 4.98273849e-01 3.78644317e-01
7.90575802e-01 -1.14590502e+00 -4.62832779e-01 2.91276425e-01
-1.97873265e-02 -5.82593858e-01 4.15622860e-01 8.04724216e-01
-6.50275409e-01 2.19064757e-01 -6.52609289e-01 -5.39689183e-01
-1.22495806e+00 5.80086052e-01 2.57849097e-01 -2.48611137e-01
-5.56776345e-01 8.18162620e-01 8.91731754e-02 1.63076684e-01
4.63300645e-02 -1.23905182e-01 1.55141810e-02 8.64546299e-02
8.22049379e-01 4.82721835e-01 2.81991437e-03 -1.06121325e+00
-6.66950226e-01 3.74405622e-01 -3.00852358e-01 1.06033981e-01
1.50552022e+00 4.95233350e-02 -5.61856478e-02 -1.26432553e-01
1.25252497e+00 -6.77842572e-02 -1.47077131e+00 -1.53174758e-01
7.70067051e-02 -1.03599608e+00 4.51356500e-01 -9.83395278e-01
-1.30195189e+00 1.29306829e+00 8.63122642e-01 -1.09392905e-03
1.14258611e+00 -2.66622037e-01 7.51508713e-01 -1.41265437e-01
1.97597861e-01 -9.46142495e-01 3.11728477e-01 1.12178899e-01
6.70773804e-01 -1.20598161e+00 1.09364547e-01 -2.86089834e-02
-5.96704245e-01 1.33106029e+00 4.59986359e-01 -4.59938407e-01
2.84503937e-01 -1.57214012e-02 -1.14970982e-01 -4.31580812e-01
-3.83822113e-01 -7.46305510e-02 9.65940282e-02 8.14991057e-01
6.96754456e-02 -2.25099921e-01 -7.43307499e-03 5.67329168e-01
2.53502280e-01 -2.88984507e-01 3.48789245e-01 9.20289576e-01
-3.34091097e-01 -9.38661098e-01 -6.99532509e-01 3.22732478e-01
-9.44717348e-01 -9.27124321e-02 -6.01852119e-01 6.96032822e-01
4.15489197e-01 9.40819919e-01 1.57064080e-01 -5.51055729e-01
-7.59872869e-02 1.54179692e-01 3.66432637e-01 -3.76240402e-01
-9.93575871e-01 -2.45838642e-01 -1.06106676e-01 -9.64297473e-01
-3.34193081e-01 -8.97939980e-01 -8.93203437e-01 -4.19115454e-01
-3.28760624e-01 -3.29165339e-01 6.41897738e-01 8.51752758e-01
7.76171207e-01 3.94422203e-01 3.78982872e-01 -9.48119938e-01
-3.55115145e-01 -7.80391693e-01 -2.47106478e-01 6.64781034e-01
3.46782446e-01 -8.29688787e-01 -3.58441114e-01 -2.18104735e-01] | [12.386143684387207, 1.0156409740447998] |
0deb9cf2-6a02-4c7f-aecb-a12da50f7f4a | data-augmentation-with-paraphrase-generation | 2205.04006 | null | https://arxiv.org/abs/2205.04006v1 | https://arxiv.org/pdf/2205.04006v1.pdf | Data Augmentation with Paraphrase Generation and Entity Extraction for Multimodal Dialogue System | Contextually aware intelligent agents are often required to understand the users and their surroundings in real-time. Our goal is to build Artificial Intelligence (AI) systems that can assist children in their learning process. Within such complex frameworks, Spoken Dialogue Systems (SDS) are crucial building blocks to handle efficient task-oriented communication with children in game-based learning settings. We are working towards a multimodal dialogue system for younger kids learning basic math concepts. Our focus is on improving the Natural Language Understanding (NLU) module of the task-oriented SDS pipeline with limited datasets. This work explores the potential benefits of data augmentation with paraphrase generation for the NLU models trained on small task-specific datasets. We also investigate the effects of extracting entities for conceivably further data expansion. We have shown that paraphrasing with model-in-the-loop (MITL) strategies using small seed data is a promising approach yielding improved performance results for the Intent Recognition task. | ['Lama Nachman', 'Saurav Sahay', 'Eda Okur'] | 2022-05-09 | null | https://aclanthology.org/2022.lrec-1.437 | https://aclanthology.org/2022.lrec-1.437.pdf | lrec-2022-6 | ['paraphrase-generation', 'intent-recognition', 'paraphrase-generation', 'spoken-dialogue-systems'] | ['computer-code', 'natural-language-processing', 'natural-language-processing', 'speech'] | [ 3.54775071e-01 8.17534983e-01 2.34138682e-01 -5.87031126e-01
-4.74971443e-01 -6.42274618e-01 8.73300552e-01 5.50650477e-01
-5.09559333e-01 5.79767227e-01 6.13083303e-01 -5.39120257e-01
1.15506098e-01 -8.52718174e-01 -2.48507395e-01 -2.31113899e-02
1.07698783e-01 8.61330688e-01 4.29669142e-01 -8.67360294e-01
2.75525808e-01 2.19234139e-01 -2.02418327e+00 6.57491982e-01
1.19566870e+00 1.25452042e-01 3.38794112e-01 1.29725242e+00
-6.26949668e-01 1.28618586e+00 -6.25638306e-01 -4.57442224e-01
-2.42900163e-01 -4.03932065e-01 -1.25310612e+00 -2.43857279e-01
4.93604183e-01 -4.93719369e-01 -3.22529189e-02 5.76052248e-01
5.44281244e-01 3.61918509e-01 5.88123500e-01 -1.08856666e+00
-2.28220344e-01 8.20350707e-01 8.24697837e-02 -2.20370919e-01
1.05716312e+00 6.68113381e-02 7.91737735e-01 -5.99764943e-01
6.43186748e-01 1.05703175e+00 3.72905552e-01 1.11797082e+00
-9.65523899e-01 -2.26979420e-01 -8.65526125e-02 3.27531189e-01
-9.89580989e-01 -6.45739138e-01 6.51853561e-01 -3.14507276e-01
1.42612278e+00 4.76636142e-01 8.24727297e-01 9.04895186e-01
-5.89300215e-01 1.39955819e+00 7.23268688e-01 -9.95055974e-01
2.20373958e-01 2.93778419e-01 4.54693139e-01 6.51646733e-01
-2.32745394e-01 -3.49889576e-01 -9.07398880e-01 1.64269611e-01
3.64501178e-01 -4.30627286e-01 -7.49136582e-02 -4.53257561e-01
-8.35640609e-01 1.04590809e+00 -1.17207929e-01 4.62553322e-01
-1.25612589e-02 -2.11880028e-01 6.24253392e-01 5.82138717e-01
5.78626096e-01 9.94304180e-01 -6.32048726e-01 -1.03451431e+00
-6.62135363e-01 4.22382772e-01 1.36792076e+00 1.25985706e+00
3.72238398e-01 -4.17824537e-01 -1.51580185e-01 1.36325812e+00
4.06111956e-01 8.15719962e-02 5.75601220e-01 -7.94604659e-01
6.67984009e-01 1.08402979e+00 -2.49504521e-02 -2.75044680e-01
-5.63810885e-01 3.29831392e-02 -3.83338332e-01 -5.80798797e-02
6.23786271e-01 -5.04496932e-01 -6.32868230e-01 1.62337804e+00
6.47514045e-01 -5.84446965e-03 5.59996426e-01 2.01970756e-01
1.57613146e+00 5.07811844e-01 4.05537546e-01 6.71532676e-02
1.31067514e+00 -9.69667852e-01 -6.63101137e-01 -3.79403114e-01
1.43977892e+00 -6.85746193e-01 1.11423349e+00 3.98013741e-01
-1.22316444e+00 -3.09254289e-01 -9.12859321e-01 -4.35184330e-01
-6.93624735e-01 4.41729240e-02 8.45751822e-01 1.01587582e+00
-1.19679499e+00 1.72615290e-01 -6.55450702e-01 -8.39802265e-01
6.19908161e-02 3.93416524e-01 -3.90049696e-01 -8.78021419e-02
-1.18698502e+00 1.05790424e+00 5.19869447e-01 -3.79354000e-01
-3.31894547e-01 -7.55624712e-01 -1.29420257e+00 -9.39451680e-02
1.66342750e-01 -5.10044932e-01 1.65199065e+00 -5.76866567e-01
-1.85891306e+00 1.14369392e+00 6.12464920e-02 -4.18924034e-01
3.36904138e-01 -3.52349222e-01 1.92893714e-01 -1.15524037e-02
6.22912683e-02 1.01942515e+00 1.85421541e-01 -7.11930752e-01
-9.07433271e-01 -4.15881336e-01 2.24722296e-01 8.05934191e-01
-3.70327204e-01 2.40295738e-01 -2.56097019e-01 -2.62036592e-01
1.36469409e-01 -8.43075693e-01 -2.04808578e-01 -5.38020194e-01
2.35999897e-02 -7.29140580e-01 4.53303188e-01 -6.18982255e-01
1.24027383e+00 -1.63280010e+00 6.83534890e-02 -1.21634595e-01
-5.95502816e-02 8.40313077e-01 -2.46138275e-01 8.05485010e-01
1.04854137e-01 -2.08543867e-01 3.61351758e-01 -4.71812785e-01
8.36970955e-02 2.19992071e-01 7.51520917e-02 -4.55540657e-01
1.13181852e-01 9.72231507e-01 -1.07540226e+00 -3.84470940e-01
5.23680151e-01 3.31304595e-02 -6.41227245e-01 8.69260192e-01
-4.59862500e-01 2.65256554e-01 -4.19292659e-01 1.68408900e-01
2.17057064e-01 2.90269464e-01 -8.24658498e-02 4.90262479e-01
-3.67744416e-01 7.02275813e-01 -1.02983344e+00 2.03206563e+00
-8.14626634e-01 7.70679951e-01 3.51623669e-02 -7.71752954e-01
7.33734667e-01 4.53330487e-01 1.50046453e-01 -6.65777206e-01
-1.09810308e-01 -2.58088429e-02 1.30618095e-01 -9.96285021e-01
7.00994730e-01 2.26120010e-01 -8.34418535e-02 6.13838732e-01
2.66008079e-01 -6.54599428e-01 3.25926691e-01 4.56815660e-01
1.33043694e+00 1.08863704e-01 5.33961356e-01 -1.08058207e-01
6.40243709e-01 2.10558042e-01 -2.70728588e-01 9.96847749e-01
-9.67232361e-02 3.62943679e-01 2.74434417e-01 -5.04055619e-01
-8.38473201e-01 -5.80805779e-01 3.13271403e-01 1.62916195e+00
-5.16949713e-01 -6.39453888e-01 -9.64964867e-01 -5.54799974e-01
-5.45224786e-01 1.16552293e+00 -2.59079725e-01 -2.43079104e-03
-4.08504426e-01 -1.51713282e-01 7.93540359e-01 3.81880373e-01
5.40980935e-01 -1.12450433e+00 -8.76685262e-01 5.80320358e-01
-1.64359972e-01 -1.10656810e+00 1.59365416e-01 1.11400358e-01
-4.74169701e-01 -8.94688368e-01 -6.40727282e-01 -9.26559210e-01
4.28932309e-01 4.08431701e-02 1.13551354e+00 2.06871480e-01
-1.78686470e-01 8.95574808e-01 -9.02408242e-01 -9.15018022e-01
-9.51101899e-01 4.25990850e-01 -2.12497413e-01 -8.01275253e-01
8.63802552e-01 -5.08993745e-01 -3.01234514e-01 -1.02312751e-01
-6.47024274e-01 9.42162275e-01 2.00882167e-01 6.73192203e-01
-3.94653916e-01 -3.44877034e-01 5.34720242e-01 -9.99237835e-01
1.04249907e+00 -2.81085372e-01 -2.33372927e-01 4.98940736e-01
-8.96856040e-02 2.22912401e-01 2.69988477e-01 -5.63172102e-01
-1.25632465e+00 2.53991812e-01 -7.80230701e-01 4.59004164e-01
-9.69777584e-01 3.03120226e-01 -2.39687800e-01 -1.22783743e-01
7.54928589e-01 2.63129890e-01 -9.37229674e-03 -2.67260224e-01
7.11804569e-01 1.16081965e+00 3.73190850e-01 -9.51613963e-01
7.89083466e-02 -3.14787060e-01 -2.81179219e-01 -1.37501609e+00
-6.81293249e-01 -8.09456527e-01 -8.55554461e-01 -4.91433173e-01
7.96671450e-01 -9.15512919e-01 -8.29558372e-01 5.37706733e-01
-1.40751112e+00 -6.87709391e-01 -2.30776057e-01 2.41873458e-01
-6.06311679e-01 4.23624814e-02 -4.58338618e-01 -1.06216013e+00
-3.84992927e-01 -1.00760579e+00 9.55804467e-01 7.38300443e-01
-7.69727945e-01 -1.03186023e+00 4.48088109e-01 1.11402643e+00
2.80521929e-01 -3.06501836e-01 9.59488571e-01 -1.66060245e+00
-1.75463393e-01 -2.11693887e-02 1.47663325e-03 1.18188962e-01
-1.08690500e-01 -7.16454834e-02 -1.11630332e+00 1.91945508e-01
-2.86844015e-01 -7.55004346e-01 8.85623470e-02 -1.55982420e-01
4.82877374e-01 -3.62461567e-01 2.25346535e-02 -5.83827868e-02
8.89815331e-01 3.37892115e-01 4.42231745e-01 2.06326798e-01
7.58519948e-01 1.31522048e+00 7.63970017e-01 3.09024096e-01
9.47870195e-01 6.62329197e-01 -1.01265796e-01 2.44136140e-01
-1.11173563e-01 -3.64307702e-01 1.60960510e-01 6.67953491e-01
1.30437464e-01 -2.07914814e-01 -1.49366617e+00 1.00181711e+00
-1.96881151e+00 -6.38454974e-01 -3.61803144e-01 1.83390200e+00
1.21977019e+00 -1.40635267e-01 1.51174068e-01 -3.83319110e-02
2.41486952e-01 -1.42566323e-01 -6.63068295e-02 -8.99273753e-01
5.55437386e-01 5.20507634e-01 -3.93029489e-02 7.77989447e-01
-8.41822088e-01 1.52037573e+00 5.29279804e+00 8.28954458e-01
-5.03639519e-01 -1.09782510e-01 4.79549825e-01 1.59178466e-01
-2.24837273e-01 -3.96924525e-01 -8.53428066e-01 5.69477230e-02
1.12372887e+00 1.01597048e-01 3.32393825e-01 7.88966715e-01
2.13983715e-01 -5.62103212e-01 -1.37826550e+00 6.29560828e-01
4.40611653e-02 -1.21160781e+00 -1.75814629e-01 -3.18409562e-01
4.45121020e-01 1.19609684e-01 -3.80318731e-01 6.18074954e-01
6.88992620e-01 -9.70752060e-01 1.68934867e-01 2.54787803e-01
4.06251162e-01 -7.84521818e-01 6.51160896e-01 8.51550996e-01
-9.06438291e-01 1.86486661e-01 1.52605083e-02 -5.26148498e-01
-3.24809432e-01 -2.08461404e-01 -1.69689155e+00 7.53732324e-02
4.01590019e-01 1.70260429e-01 -7.69779384e-01 8.91897678e-01
-4.37288582e-01 5.61496317e-01 -5.51026583e-01 -8.88771236e-01
2.31841177e-01 -2.21205249e-01 4.94582415e-01 1.26777542e+00
2.02558320e-02 7.95075417e-01 -2.13906974e-01 4.66174871e-01
2.97844142e-01 5.35667598e-01 -1.08513880e+00 -1.47197247e-01
4.13742304e-01 1.07025063e+00 -5.39078653e-01 -1.39510483e-01
-6.47453427e-01 9.01777864e-01 5.85130572e-01 2.55327066e-03
-8.41730833e-02 -4.13164884e-01 7.10309088e-01 -2.10540257e-02
-1.38462365e-01 -2.49764040e-01 -3.43458921e-01 -9.53256249e-01
-3.50007743e-01 -1.14447260e+00 3.16550523e-01 -9.33604240e-01
-7.57540107e-01 1.19557887e-01 1.72711477e-01 -7.29852617e-01
-9.16333318e-01 -5.58518887e-01 -9.79575515e-01 6.22029483e-01
-7.87159324e-01 -1.42185843e+00 -1.12612374e-01 4.18573052e-01
1.13842940e+00 -3.35371286e-01 1.22070312e+00 -2.07675889e-01
-3.74694437e-01 4.75854009e-01 -1.71139717e-01 1.76058605e-01
4.70666111e-01 -1.60658157e+00 5.13905585e-01 7.03354061e-01
3.97187650e-01 6.99429691e-01 1.09329581e+00 -7.55529404e-01
-1.11931336e+00 -4.83537853e-01 1.19002700e+00 -5.81004083e-01
7.96773672e-01 -5.92969000e-01 -8.62680435e-01 3.66555780e-01
6.39539599e-01 -9.30950046e-01 1.10396802e+00 4.84373510e-01
1.30803794e-01 4.58858848e-01 -1.12734604e+00 8.76816332e-01
9.57734585e-01 -5.72663963e-01 -1.09049404e+00 2.48952657e-01
8.96282196e-01 -7.30409741e-01 -8.63092363e-01 3.03410292e-01
3.50552946e-01 -8.76732171e-01 8.79829347e-01 -9.32772279e-01
5.28523147e-01 2.25160003e-01 2.55451530e-01 -1.36725652e+00
4.95958149e-01 -8.34686518e-01 6.89844787e-02 1.62403309e+00
5.25618434e-01 -2.87808567e-01 1.30427063e+00 1.60115910e+00
1.35765048e-02 -6.04765534e-01 -7.48984158e-01 2.48697773e-02
1.06799871e-01 -8.60868692e-01 5.12700915e-01 8.37585628e-01
8.40340257e-01 7.29648709e-01 1.00853555e-01 8.47563222e-02
2.51585450e-02 -4.00068730e-01 1.06812882e+00 -1.09522378e+00
-1.77029856e-02 -3.06252956e-01 -3.65454257e-01 -1.03800070e+00
2.14787409e-01 -4.16943431e-01 1.00401014e-01 -1.67159891e+00
-2.19014570e-01 -2.96463549e-01 5.71944892e-01 5.69929838e-01
-1.56987846e-01 -4.44976896e-01 1.44420579e-01 -4.80878711e-01
-9.42730844e-01 4.83666122e-01 9.70453799e-01 8.74782261e-03
-4.58714724e-01 5.45882761e-01 -4.86929566e-01 9.17016447e-01
7.63446808e-01 -8.32226127e-02 -9.59311783e-01 -3.31422925e-01
3.71333271e-01 2.07212523e-01 -1.24087602e-01 -9.70884562e-01
8.64944875e-01 -1.56294405e-01 -1.36536971e-01 -7.72406101e-01
1.73730865e-01 -7.03140974e-01 -4.65763420e-01 1.36732116e-01
-7.74601221e-01 -2.85198689e-01 4.69250143e-01 -1.80683568e-01
-1.87962770e-01 -8.93611550e-01 2.52355725e-01 -1.90683380e-01
-9.86084223e-01 -2.29257658e-01 -8.81106079e-01 2.50547796e-01
1.00930011e+00 -2.80038208e-01 -3.07705998e-01 -8.03383708e-01
-7.42746890e-01 7.90735245e-01 3.35900992e-01 4.85614568e-01
6.24808133e-01 -8.10199976e-01 -6.16414845e-01 4.12861466e-01
4.48959112e-01 4.81051624e-01 1.19191095e-01 -1.05280627e-03
-8.66360545e-01 8.51199508e-01 -1.71265841e-01 -2.85947680e-01
-2.12767458e+00 -1.28194481e-01 1.90698326e-01 -3.99983853e-01
-4.81333174e-02 1.41258955e+00 -1.82929739e-01 -1.05970931e+00
6.43576741e-01 -2.78371871e-01 -9.28117156e-01 2.53773391e-01
8.49833727e-01 4.02844459e-01 1.30381227e-01 -3.64181131e-01
8.26847479e-02 -6.62731528e-02 -3.72539043e-01 -4.66089487e-01
1.54914939e+00 -2.06561357e-01 -2.57366919e-03 2.61751175e-01
7.89987564e-01 -1.37342378e-01 -7.92318165e-01 -1.88579232e-01
3.48770559e-01 3.46477819e-03 -2.87108034e-01 -9.10378456e-01
-1.63627177e-01 1.00587475e+00 3.71380061e-01 5.26522398e-01
8.89235735e-01 2.47376621e-01 3.20060819e-01 8.42760444e-01
4.92097586e-01 -1.13152266e+00 -4.85437140e-02 9.55596924e-01
7.05303788e-01 -1.46852279e+00 -3.11226379e-02 -4.19231772e-01
-8.68175983e-01 1.12586462e+00 1.07015800e+00 6.20529711e-01
1.07675835e-01 1.56774133e-01 1.53838187e-01 -2.99440593e-01
-1.06013048e+00 -6.17440403e-01 3.92169803e-01 1.00746870e+00
8.51496696e-01 5.36446162e-02 -3.01266134e-01 6.31826997e-01
-5.99498391e-01 -1.83280930e-01 7.55239189e-01 1.09641540e+00
-4.93899524e-01 -1.54277217e+00 9.74906050e-03 2.97461748e-01
-1.26178250e-01 -4.04690623e-01 -1.00453150e+00 8.41444969e-01
5.64022083e-03 1.36346018e+00 -1.78962454e-01 -2.04389125e-01
4.55428481e-01 4.67043996e-01 5.79237938e-01 -1.41028821e+00
-9.42149639e-01 -6.03533506e-01 7.85931945e-01 -4.59540814e-01
-2.77158171e-01 -7.87873387e-01 -1.25393939e+00 1.53653011e-01
-2.75116712e-01 2.40873948e-01 8.01685333e-01 1.23832119e+00
4.31469157e-02 9.98411402e-02 1.11742079e-01 -5.18220961e-01
-3.69514376e-02 -1.08631706e+00 4.05899547e-02 4.99947928e-02
6.88406006e-02 -1.67121105e-02 2.18807116e-01 -2.35518619e-01] | [12.584705352783203, 7.9947075843811035] |
1c1b8e5d-a41a-43f6-8c65-80ec1efb04f6 | image-to-image-regression-with-distribution | 2202.05265 | null | https://arxiv.org/abs/2202.05265v1 | https://arxiv.org/pdf/2202.05265v1.pdf | Image-to-Image Regression with Distribution-Free Uncertainty Quantification and Applications in Imaging | Image-to-image regression is an important learning task, used frequently in biological imaging. Current algorithms, however, do not generally offer statistical guarantees that protect against a model's mistakes and hallucinations. To address this, we develop uncertainty quantification techniques with rigorous statistical guarantees for image-to-image regression problems. In particular, we show how to derive uncertainty intervals around each pixel that are guaranteed to contain the true value with a user-specified confidence probability. Our methods work in conjunction with any base machine learning model, such as a neural network, and endow it with formal mathematical guarantees -- regardless of the true unknown data distribution or choice of model. Furthermore, they are simple to implement and computationally inexpensive. We evaluate our procedure on three image-to-image regression tasks: quantitative phase microscopy, accelerated magnetic resonance imaging, and super-resolution transmission electron microscopy of a Drosophila melanogaster brain. | ['Yaniv Romano', 'Srigokul Upadhyayula', 'Thayer Alshaabi', 'Jitendra Malik', 'Michael I Jordan', 'Stephen Bates', 'Amit P Kohli', 'Anastasios N Angelopoulos'] | 2022-02-10 | null | null | null | null | ['prediction-intervals'] | ['miscellaneous'] | [ 4.30726379e-01 3.46694171e-01 5.03005125e-02 -5.47703743e-01
-1.04561365e+00 -1.99310467e-01 2.51223087e-01 1.71964377e-01
-7.16396272e-01 1.18350971e+00 -5.42743087e-01 -1.19890332e-01
-7.60985166e-02 -3.47326964e-01 -9.85188305e-01 -1.04203248e+00
-1.17883980e-01 5.81105828e-01 1.10109307e-01 5.12036026e-01
5.25933921e-01 4.02665108e-01 -9.88429606e-01 -2.27824617e-02
6.73912048e-01 1.01213622e+00 1.22765295e-01 7.46514916e-01
3.51995438e-01 8.92419159e-01 -3.87728959e-01 -2.34932810e-01
2.74396949e-02 -3.42426121e-01 -5.78945398e-01 9.52573344e-02
2.42616400e-01 -3.84606928e-01 3.50121632e-02 1.47752619e+00
4.08222616e-01 -1.54672712e-01 8.40429008e-01 -1.38070774e+00
-5.05138695e-01 4.51700181e-01 -9.22898948e-01 1.92108795e-01
-6.02939390e-02 3.23010944e-02 5.80760598e-01 -5.94879687e-01
7.28382945e-01 1.00843406e+00 7.76147187e-01 8.11371863e-01
-1.75190604e+00 -5.90516806e-01 -1.62015319e-01 -1.31152868e-01
-1.37634981e+00 -5.83535075e-01 5.36744535e-01 -7.88134277e-01
5.77033460e-01 -4.52298000e-02 2.36966595e-01 8.49744856e-01
7.63696790e-01 5.56000292e-01 1.34483731e+00 -4.51921344e-01
4.68061984e-01 2.89061069e-01 1.01271737e-02 8.08271527e-01
2.32674658e-01 1.38786733e-01 -5.30585110e-01 -3.54515970e-01
1.12363780e+00 -3.68302613e-01 -4.56316918e-01 -4.82892275e-01
-1.04235506e+00 7.33335555e-01 -3.12311808e-03 1.94576904e-02
-1.55284345e-01 5.02503455e-01 1.44350395e-01 4.47600894e-02
6.33841991e-01 3.04129034e-01 -4.50808555e-01 1.91232413e-01
-9.26852345e-01 2.75592625e-01 5.30282438e-01 6.02700651e-01
6.22096717e-01 7.10845739e-02 2.32134417e-01 5.07725716e-01
3.73241007e-01 5.62025309e-01 2.72538096e-01 -1.57518208e+00
-1.47244751e-01 -3.00146073e-01 3.22856605e-01 -8.75665009e-01
-3.98670584e-01 -3.04364413e-01 -9.62562382e-01 9.20171261e-01
5.06489635e-01 -1.58429131e-01 -6.30350232e-01 2.08816624e+00
1.65712371e-01 2.88655251e-01 -1.45179406e-01 8.16980839e-01
1.95689768e-01 3.29468936e-01 -7.07915574e-02 -7.35432267e-01
9.33638334e-01 -2.53075868e-01 -6.28966451e-01 5.71084619e-02
3.47730547e-01 -3.03426743e-01 7.56356895e-01 7.43264258e-01
-1.28259146e+00 -8.47360194e-02 -1.09420013e+00 9.52629298e-02
1.86715484e-01 -1.92484364e-01 4.51219857e-01 4.94353056e-01
-9.80498314e-01 9.67034817e-01 -1.14028156e+00 1.15839750e-01
6.00875258e-01 4.32130933e-01 -5.46307445e-01 3.79779786e-01
-5.21013141e-01 1.03423667e+00 2.73636401e-01 -1.30863383e-01
-9.26570296e-01 -8.19975972e-01 -7.06407547e-01 -2.22322777e-01
3.42532918e-02 -8.05426717e-01 1.30078363e+00 -1.04347777e+00
-1.37856019e+00 1.19560349e+00 -2.20803127e-01 -7.63384104e-01
6.26597822e-01 6.69242516e-02 1.46947771e-01 2.89653331e-01
1.59919351e-01 8.84391963e-01 1.15338862e+00 -1.41055787e+00
-3.05557221e-01 -6.40653610e-01 -3.50425929e-01 -1.67205244e-01
1.82555523e-02 -8.03096443e-02 2.31068823e-02 -2.62381017e-01
2.61998296e-01 -7.77339935e-01 -5.22700727e-01 5.62301815e-01
-5.40109694e-01 2.47559741e-01 2.24959314e-01 -4.93746191e-01
6.40287161e-01 -2.07274652e+00 9.97125506e-02 2.34735787e-01
3.92590761e-01 -2.87889063e-01 1.38080850e-01 -2.54531175e-01
-1.25383005e-01 1.50440007e-01 -5.36136627e-01 -5.33434153e-01
-1.84893876e-01 1.84273183e-01 -1.79572240e-01 1.06661296e+00
2.80332804e-01 5.89156151e-01 -7.34978735e-01 -9.28479850e-01
2.79617310e-01 5.75692892e-01 -4.94771481e-01 3.86690609e-02
-2.53073335e-01 7.63311625e-01 -4.23465550e-01 2.77450740e-01
6.46483719e-01 -5.31108022e-01 6.12831973e-02 -9.89559516e-02
-2.22216267e-02 -2.76157409e-01 -8.47638607e-01 1.40856814e+00
-1.25341773e-01 6.07378244e-01 3.24565351e-01 -1.14298439e+00
5.37913084e-01 1.66125536e-01 7.70492375e-01 -3.80147815e-01
1.27571195e-01 1.81329638e-01 -1.82419822e-01 -2.67423749e-01
-2.31934458e-01 -6.95105791e-01 3.67045730e-01 4.02970344e-01
1.66890159e-01 -4.87125665e-01 -1.29864365e-01 4.22223471e-02
9.74053621e-01 1.95118293e-01 4.33070600e-01 -5.40562809e-01
2.43397608e-01 -1.53935879e-01 8.18808675e-01 7.80114055e-01
-3.69920343e-01 9.50108528e-01 8.22218299e-01 -2.69157678e-01
-1.39002061e+00 -1.26702952e+00 -5.56485057e-01 3.62195671e-01
-9.07425135e-02 1.21972889e-01 -1.04716337e+00 -2.72132546e-01
-8.29395428e-02 5.89716911e-01 -6.88111663e-01 2.04723373e-01
6.55489862e-02 -8.14021051e-01 3.95456433e-01 1.49264023e-01
6.77696243e-02 -9.28286433e-01 -7.13727653e-01 1.59055889e-01
5.42355105e-02 -1.16376126e+00 -1.56631455e-01 3.89284641e-01
-9.68257666e-01 -1.16077137e+00 -6.18700922e-01 -3.64122421e-01
9.74938035e-01 -2.08559424e-01 1.11692774e+00 -1.08204233e-02
-4.61512238e-01 4.18350518e-01 2.93517590e-01 -5.51631689e-01
-4.83782798e-01 -6.21063292e-01 2.15149239e-01 -3.45774889e-01
5.16759371e-03 -7.43866801e-01 -3.58048856e-01 9.79885757e-02
-1.00032854e+00 9.61077586e-02 2.59854436e-01 1.07324421e+00
1.07580364e+00 3.99999827e-01 3.11511189e-01 -9.33970988e-01
3.89714062e-01 -2.33269230e-01 -1.10393453e+00 2.01118872e-01
-7.11251736e-01 2.06680253e-01 4.13865447e-01 -3.99408966e-01
-7.62515604e-01 3.06519449e-01 4.92838547e-02 -5.40622115e-01
-6.05578087e-02 4.26391125e-01 1.84264094e-01 -4.20946240e-01
8.47074449e-01 1.09117411e-01 4.26931113e-01 -4.34567258e-02
1.52139403e-02 2.56707281e-01 9.32101786e-01 -7.48521268e-01
6.34329975e-01 7.73118079e-01 4.14444059e-01 -8.88871789e-01
-8.20075929e-01 -1.30009294e-01 -5.70637405e-01 -2.49558806e-01
8.96832943e-01 -7.57909834e-01 -9.45800185e-01 5.33257484e-01
-1.14719903e+00 -4.19157535e-01 -6.63268417e-02 6.49006963e-01
-1.25548220e+00 4.79976654e-01 -6.92926705e-01 -1.20894587e+00
-1.58425689e-01 -1.13821101e+00 1.01351571e+00 1.12295881e-01
1.04892841e-02 -1.03393555e+00 -4.00781296e-02 6.17780909e-02
1.97137281e-01 4.79895741e-01 8.81417334e-01 -2.71189749e-01
-4.74704325e-01 2.26972974e-03 -2.94661373e-01 5.73433161e-01
-1.65916055e-01 2.89728791e-01 -1.05708504e+00 -1.77483737e-01
5.57362199e-01 -5.67452312e-01 8.95611227e-01 1.11258256e+00
1.54170191e+00 -3.16566527e-01 -1.63914248e-01 6.25363827e-01
1.41708362e+00 -1.11448370e-01 6.41907096e-01 9.12338272e-02
1.48697481e-01 7.99092948e-01 4.12194908e-01 6.01381361e-01
-1.14691541e-01 3.72537613e-01 5.25506616e-01 1.11576401e-01
3.22884470e-01 -1.34837283e-02 2.27143154e-01 4.57933128e-01
7.43135065e-02 -4.97304089e-03 -6.25587225e-01 3.04419994e-01
-1.86207414e+00 -1.01711631e+00 -1.02331294e-02 2.47816849e+00
1.02030313e+00 3.81052107e-01 -1.28518641e-01 3.28761823e-02
6.61107540e-01 -2.61303604e-01 -7.94503212e-01 3.98557447e-02
-2.33542338e-01 -6.91498630e-04 6.19848788e-01 8.26253235e-01
-1.01787233e+00 4.61153358e-01 7.19563532e+00 8.22477639e-01
-9.41262484e-01 6.44091368e-02 1.09249938e+00 2.18127538e-02
-4.42575276e-01 -1.31770432e-01 -4.32869345e-01 3.49428296e-01
8.04901719e-01 -1.49022713e-01 4.61455107e-01 7.89117575e-01
2.87008703e-01 -3.81864130e-01 -1.30193877e+00 1.13093591e+00
-7.18573183e-02 -1.47701073e+00 -3.59411448e-01 8.17833692e-02
4.88584787e-01 2.55047753e-02 2.32498407e-01 -2.88215548e-01
3.32152247e-01 -1.26816869e+00 6.16417944e-01 6.87075734e-01
6.94550395e-01 -7.98934579e-01 4.65414256e-01 5.24191201e-01
-2.22213998e-01 3.04582000e-01 -6.76558793e-01 2.20203415e-01
3.46773341e-02 1.06387281e+00 -5.76569736e-01 -1.05917148e-01
6.52419269e-01 4.07426536e-01 -1.35246724e-01 8.41869295e-01
1.15008038e-02 3.64727229e-01 -5.14483809e-01 1.53170720e-01
-1.90452218e-01 -2.23410606e-01 5.49362242e-01 1.09329963e+00
2.91242421e-01 1.26495600e-01 -1.60796851e-01 1.30487454e+00
7.65122548e-02 -2.46972024e-01 -5.90579152e-01 -4.48155813e-02
2.28356168e-01 1.15512860e+00 -7.73565769e-01 1.47546288e-02
-2.03856960e-01 8.04680645e-01 3.04228127e-01 4.05090243e-01
-6.97683871e-01 -4.82420288e-02 5.94371736e-01 2.32425049e-01
-5.77303991e-02 -1.59482941e-01 -4.44974869e-01 -1.18516266e+00
-4.87393923e-02 -8.19921196e-01 1.65587831e-02 -1.21622670e+00
-1.55824280e+00 1.41050339e-01 6.36716336e-02 -6.59533203e-01
-3.58334363e-01 -9.74259377e-01 -3.12645674e-01 8.26627612e-01
-1.40347004e+00 -8.28177571e-01 2.79588073e-01 5.75903356e-01
3.72742862e-02 3.47195081e-02 8.70531738e-01 -2.24053994e-01
-3.61335248e-01 2.48858139e-01 2.43816972e-01 8.53823125e-02
6.29883707e-01 -1.38026512e+00 -1.75062969e-01 7.73075700e-01
-4.30348217e-02 5.64636290e-01 1.22054803e+00 -5.17194450e-01
-1.15619588e+00 -7.27402091e-01 5.55747330e-01 -3.62776488e-01
8.31796765e-01 -8.18131343e-02 -9.08778191e-01 8.21920633e-01
-8.29705223e-02 4.16036397e-01 6.08984053e-01 -9.58361179e-02
-3.40815902e-01 6.14457093e-02 -1.58040977e+00 4.86276925e-01
4.47468072e-01 -7.39042163e-01 -3.50316674e-01 5.15935063e-01
4.33431506e-01 -3.85973960e-01 -1.03855002e+00 3.98438931e-01
6.90344095e-01 -1.11733198e+00 9.66073096e-01 -6.30191267e-01
6.08743906e-01 -2.73635685e-01 -3.48420471e-01 -1.12775779e+00
-5.16419522e-02 -7.01256156e-01 -5.60385641e-04 7.69889176e-01
3.38823795e-01 -5.15637994e-01 8.56636882e-01 9.52013314e-01
1.61025017e-01 -5.31813502e-01 -1.17205155e+00 -6.89077497e-01
3.55564743e-01 -6.15537345e-01 -1.33787304e-01 8.03632438e-01
1.64443925e-01 7.86025673e-02 -4.78624314e-01 3.70321602e-01
1.46475422e+00 -2.87091911e-01 4.08526570e-01 -1.23827624e+00
-4.61341351e-01 -5.39681137e-01 -7.22144246e-01 -6.38593495e-01
4.62771833e-01 -4.19100136e-01 3.97936583e-01 -8.40206206e-01
5.98581612e-01 -2.79119760e-01 -3.28594595e-01 2.06826150e-01
1.63806453e-02 1.53598830e-01 -2.06602797e-01 1.32843271e-01
-5.88336170e-01 2.56018609e-01 1.09504259e+00 -5.57604730e-02
3.11182380e-01 6.60434440e-02 -5.11926115e-01 1.16776562e+00
7.93509245e-01 -7.28829265e-01 -3.85728449e-01 -2.53659219e-01
3.96132916e-01 4.77345943e-01 6.02361619e-01 -9.10124540e-01
3.78866732e-01 -2.62523115e-01 5.86776137e-01 -2.85751939e-01
3.07354182e-01 -7.15944588e-01 1.53918862e-01 5.18845558e-01
-6.21415794e-01 -3.73202682e-01 1.07066534e-01 5.48527598e-01
2.22483501e-02 -6.02751613e-01 1.34993672e+00 -1.54643506e-01
-2.28201658e-01 3.52415949e-01 -3.92986417e-01 9.37068537e-02
8.53684068e-01 1.71984136e-01 -1.58385262e-01 -5.94692409e-01
-9.11176383e-01 8.64290223e-02 5.53102255e-01 -3.31754029e-01
7.90892899e-01 -9.53884304e-01 -8.52023840e-01 1.24483190e-01
-1.37666147e-02 -6.69345185e-02 1.30893141e-01 9.55354333e-01
-3.85917157e-01 1.79518744e-01 -2.41547272e-01 -9.18638468e-01
-1.15865016e+00 5.46909094e-01 6.60740435e-01 1.72590949e-02
-3.61414284e-01 8.34390104e-01 3.54815841e-01 -3.53721052e-01
2.05545515e-01 -3.77816021e-01 2.78500110e-01 -5.56157112e-01
7.42741704e-01 7.91741163e-02 -3.42784882e-01 -3.79155844e-01
-1.87439904e-01 5.34116030e-01 1.76373228e-01 -5.33632159e-01
1.26900029e+00 -3.08198959e-01 -3.07616115e-01 7.88230896e-01
9.93273616e-01 -2.86875576e-01 -1.63427842e+00 -1.25042915e-01
-1.60200540e-02 -3.38106036e-01 2.82193542e-01 -5.96533716e-01
-9.42727804e-01 1.05921495e+00 5.48433840e-01 1.55792028e-01
8.80739510e-01 -1.64069369e-01 1.69544786e-01 5.50312698e-01
5.01200080e-01 -1.05220187e+00 -7.05280155e-03 6.76208548e-03
7.59046674e-01 -1.61682808e+00 2.09749907e-01 -1.75807789e-01
-4.04733121e-01 1.07458854e+00 3.05825710e-01 -2.42135808e-01
8.91789258e-01 7.97120988e-01 -5.96157275e-02 2.48053707e-02
-7.93404400e-01 2.39666983e-01 1.98539048e-02 9.88286614e-01
3.49189758e-01 -1.36855930e-01 -5.46827074e-03 3.23018909e-01
1.72679156e-01 4.79472041e-01 5.89327693e-01 6.18854284e-01
-7.24965572e-01 -7.11087048e-01 -3.27147603e-01 4.05382127e-01
-7.61947870e-01 -1.06284223e-01 5.96475378e-02 6.63201153e-01
-1.60393640e-01 7.50522196e-01 6.15578741e-02 4.80428338e-02
-3.17456067e-01 -1.05320476e-01 1.02112448e+00 -3.94215226e-01
1.74262390e-01 2.24298537e-01 -2.66881168e-01 -5.97257137e-01
-6.74271584e-01 -6.29256546e-01 -1.48411047e+00 -5.04485130e-01
-4.43177044e-01 4.35266457e-02 7.94039249e-01 1.11226261e+00
-8.04920867e-02 1.11501351e-01 3.60211700e-01 -8.18961918e-01
-7.52573967e-01 -6.92727327e-01 -9.10434127e-01 4.61797304e-02
6.33165002e-01 -4.42066163e-01 -6.29204214e-01 2.05105722e-01] | [11.758296966552734, -1.753294587135315] |
d1b0eaa5-7d6c-4b81-ad8f-adff93b4d890 | skepxels-spatio-temporal-image-representation | 1711.05941 | null | http://arxiv.org/abs/1711.05941v4 | http://arxiv.org/pdf/1711.05941v4.pdf | Skepxels: Spatio-temporal Image Representation of Human Skeleton Joints for Action Recognition | Human skeleton joints are popular for action analysis since they can be
easily extracted from videos to discard background noises. However, current
skeleton representations do not fully benefit from machine learning with CNNs.
We propose "Skepxels" a spatio-temporal representation for skeleton sequences
to fully exploit the "local" correlations between joints using the 2D
convolution kernels of CNN. We transform skeleton videos into images of
flexible dimensions using Skepxels and develop a CNN-based framework for
effective human action recognition using the resulting images. Skepxels encode
rich spatio-temporal information about the skeleton joints in the frames by
maximizing a unique distance metric, defined collaboratively over the distinct
joint arrangements used in the skeletal image. Moreover, they are flexible in
encoding compound semantic notions such as location and speed of the joints.
The proposed action recognition exploits the representation in a hierarchical
manner by first capturing the micro-temporal relations between the skeleton
joints with the Skepxels and then exploiting their macro-temporal relations by
computing the Fourier Temporal Pyramids over the CNN features of the skeletal
images. We extend the Inception-ResNet CNN architecture with the proposed
method and improve the state-of-the-art accuracy by 4.4% on the large scale NTU
human activity dataset. On the medium-sized N-UCLA and UTH-MHAD datasets, our
method outperforms the existing results by 5.7% and 9.3% respectively. | ['Ajmal Mian', 'Naveed Akhtar', 'Jian Liu'] | 2017-11-16 | null | null | null | null | ['action-analysis'] | ['computer-vision'] | [ 1.42835438e-01 -2.68822968e-01 -3.79902810e-01 -1.21449515e-01
-3.89238507e-01 -1.17418267e-01 4.80250865e-01 -4.43438411e-01
-6.57236218e-01 4.02029812e-01 7.72475958e-01 5.84570169e-01
-3.57317775e-01 -5.77230573e-01 -7.96949565e-01 -7.37912595e-01
-3.49536657e-01 -7.69705558e-03 5.28841019e-01 -7.39683658e-02
9.39818025e-02 6.89583719e-01 -1.55619490e+00 5.79102814e-01
2.27278441e-01 1.22407401e+00 -4.07893257e-03 8.44013274e-01
3.02526802e-01 1.10529304e+00 -4.28044349e-01 -4.47427407e-02
4.03756589e-01 -3.90104026e-01 -7.86749423e-01 2.74208188e-01
4.21223134e-01 -4.99055564e-01 -1.00143015e+00 6.84929371e-01
4.42609727e-01 4.06297863e-01 3.41759443e-01 -1.08944738e+00
-6.60008252e-01 2.36376509e-01 -8.41579854e-01 3.91349107e-01
2.82148868e-01 2.08068609e-01 1.08071101e+00 -7.76085794e-01
7.20142245e-01 1.28870070e+00 7.65851736e-01 4.32148427e-01
-7.46440589e-01 -5.66921234e-01 4.34101634e-02 8.22476685e-01
-1.26964366e+00 -3.07069898e-01 6.27531230e-01 -3.57970983e-01
1.26475024e+00 6.68005496e-02 1.04037118e+00 1.31265295e+00
4.12708014e-01 1.04917502e+00 5.60612619e-01 -1.29808739e-01
-1.23841330e-01 -1.13328993e+00 -6.54777363e-02 1.02345502e+00
-3.49033028e-01 -9.21357945e-02 -1.03652644e+00 1.21499196e-01
1.34010041e+00 4.62310791e-01 -1.47235170e-01 -2.87950724e-01
-1.60406029e+00 6.14122450e-01 4.33448136e-01 3.03909838e-01
-5.62701106e-01 6.61699891e-01 6.87319696e-01 -1.06683727e-02
8.46457556e-02 5.18400175e-03 -5.73310912e-01 -5.09814203e-01
-6.92553639e-01 1.88757122e-01 2.05956474e-01 8.10053408e-01
4.42578346e-01 4.00985479e-02 -3.44320357e-01 8.51439357e-01
1.46439061e-01 2.81236082e-01 7.86700904e-01 -1.48500550e+00
4.75836933e-01 6.67229295e-01 -7.06305131e-02 -1.11916184e+00
-3.96683365e-01 -2.17175603e-01 -9.67240930e-01 2.23288182e-02
5.08536935e-01 1.24144576e-01 -8.35980415e-01 1.65619612e+00
3.48750234e-01 5.04737437e-01 -7.13261068e-02 7.91376948e-01
3.46454531e-01 4.77590621e-01 6.22003153e-02 3.29710054e-03
1.53031719e+00 -1.30566597e+00 -5.50394952e-01 6.38155788e-02
5.68161905e-01 -5.26140630e-01 8.19362760e-01 3.49646479e-01
-1.00150478e+00 -9.23301995e-01 -9.65452373e-01 -2.70500064e-01
5.05097471e-02 5.87506711e-01 5.21718740e-01 6.24935254e-02
-7.61439204e-01 9.59177256e-01 -1.46302462e+00 -2.90874094e-01
5.50190985e-01 3.72391552e-01 -6.37849867e-01 1.15567572e-01
-1.11884904e+00 4.76269931e-01 3.33115429e-01 3.01039279e-01
-9.45239782e-01 -4.21669900e-01 -9.55395103e-01 -6.40053824e-02
4.32746291e-01 -5.25869668e-01 1.02430630e+00 -9.46605742e-01
-1.56805503e+00 4.66207564e-01 -1.00662097e-01 -6.64155185e-01
6.04473233e-01 -7.15423107e-01 -3.20034117e-01 7.59059131e-01
2.28332177e-01 6.48035705e-01 8.30761075e-01 -3.96908939e-01
-8.22725296e-01 -4.91955757e-01 1.07779711e-01 1.62506267e-01
-3.77492428e-01 -4.39010113e-02 -7.78784037e-01 -1.05028367e+00
2.41390318e-01 -9.15694714e-01 -1.51105687e-01 4.44668382e-01
-1.69152245e-01 -2.69497126e-01 1.19604468e+00 -8.10138524e-01
9.47723687e-01 -2.13215065e+00 3.47356021e-01 -1.23419948e-01
8.50202665e-02 2.92933315e-01 -1.78094864e-01 3.44309419e-01
-1.07141554e-01 -4.80595380e-01 8.68793279e-02 -2.68370628e-01
-1.38521001e-01 6.64139807e-01 1.46796526e-02 7.13732064e-01
1.75772846e-01 1.05902326e+00 -9.26963151e-01 -6.57577693e-01
4.79410946e-01 7.32585132e-01 -4.25276846e-01 5.66735491e-02
4.76941168e-02 5.69555342e-01 -6.56954050e-01 5.69542468e-01
8.53901207e-02 -2.05075875e-01 3.94996032e-02 -4.41766053e-01
-1.61503218e-02 -9.06830207e-02 -1.06598079e+00 2.32061291e+00
-1.69087589e-01 5.00602484e-01 -2.48540714e-01 -1.22712994e+00
6.37679338e-01 3.22668225e-01 1.17645669e+00 -7.10807025e-01
3.57347801e-02 -1.20236173e-01 -1.71872690e-01 -7.18315423e-01
3.11965853e-01 1.99767530e-01 1.12250531e-02 2.96657771e-01
3.36156696e-01 7.25981832e-01 3.34685296e-01 -3.46027948e-02
1.40232015e+00 5.24750829e-01 3.29844326e-01 -2.30207164e-02
6.32306039e-01 -4.27842945e-01 7.39559591e-01 3.88804853e-01
-3.92305374e-01 6.61172390e-01 3.70619982e-01 -6.76530659e-01
-9.35592830e-01 -1.20151830e+00 2.88033932e-01 1.18498600e+00
-1.84246615e-01 -5.46737432e-01 -7.06533611e-01 -7.05470264e-01
-1.54721960e-01 -8.11687186e-02 -9.10933018e-01 -2.94878691e-01
-9.14317071e-01 -2.41662562e-01 8.69800210e-01 1.20721459e+00
1.20729244e+00 -1.08598089e+00 -1.11478078e+00 2.37757966e-01
-3.24138880e-01 -1.44416153e+00 -6.36370957e-01 -6.64382800e-02
-9.48689640e-01 -1.28055251e+00 -9.16682422e-01 -5.61668277e-01
2.68180847e-01 1.98238030e-01 5.51491857e-01 -5.27902655e-02
-6.07805431e-01 5.75954676e-01 -5.24870634e-01 3.62274110e-01
2.59829581e-01 -2.03491136e-01 1.20053232e-01 4.71216381e-01
1.94644272e-01 -9.06179845e-01 -1.01065278e+00 4.93089408e-01
-7.76975393e-01 1.09337941e-01 6.68550611e-01 8.56458366e-01
7.07009912e-01 9.43975449e-02 2.18568102e-01 -2.15899095e-01
-5.60341356e-03 -2.33598590e-01 3.32516395e-02 1.36435553e-01
1.30009010e-01 1.12300754e-01 5.49331427e-01 -5.24881065e-01
-1.00886703e+00 4.05823201e-01 -8.86748806e-02 -8.28085840e-01
-1.72875911e-01 1.23972133e-01 -1.33284226e-01 1.86075941e-01
5.02765417e-01 4.19384271e-01 -7.68386796e-02 -4.79429543e-01
3.35618615e-01 2.41580129e-01 9.03076053e-01 -7.46599853e-01
3.91384959e-01 1.00272536e+00 1.96739852e-01 -9.18522835e-01
-9.15399969e-01 -6.19736195e-01 -1.12069142e+00 -2.28531316e-01
1.45950294e+00 -9.37078834e-01 -7.46806920e-01 8.03420305e-01
-1.11563742e+00 -1.19770944e-01 -4.58023131e-01 7.51083970e-01
-9.25283194e-01 9.82135832e-01 -1.00091994e+00 -4.20930505e-01
-2.05286667e-01 -9.88336027e-01 1.26033556e+00 8.29516575e-02
-3.07367980e-01 -6.46024585e-01 3.54792215e-02 5.80234110e-01
-8.07052776e-02 5.25290787e-01 5.91114342e-01 -2.38631770e-01
-4.01318163e-01 -9.30609778e-02 -1.83717549e-01 5.87625027e-01
1.99518159e-01 -7.45145278e-03 -6.38411343e-01 -1.13205232e-01
-2.27734715e-01 -2.25516185e-01 9.88229811e-01 4.46025997e-01
1.42828751e+00 -2.27414072e-01 -1.08204670e-01 5.03254890e-01
9.44173992e-01 8.19840655e-02 7.09937572e-01 3.46178979e-01
1.00030375e+00 3.69360656e-01 5.23861408e-01 7.21834600e-01
1.44711658e-01 8.05979013e-01 4.21528935e-01 2.74493486e-01
-2.76667744e-01 -3.20164472e-01 6.46042228e-01 7.73161829e-01
-8.64453673e-01 2.47689098e-01 -5.52574873e-01 6.23191357e-01
-2.36733150e+00 -1.19711864e+00 -5.84012270e-02 1.81783831e+00
6.27442598e-01 4.89037447e-02 2.10932940e-01 1.84823006e-01
6.04393542e-01 4.67262715e-01 -5.42897940e-01 1.33512035e-01
-7.40759866e-03 5.14538407e-01 4.67518359e-01 -2.46462077e-02
-1.25915694e+00 7.95221090e-01 5.51194239e+00 9.85394299e-01
-8.50052774e-01 1.69013232e-01 5.56452908e-02 -3.35616916e-01
5.85153401e-01 -2.97790080e-01 -5.64375162e-01 3.77036661e-01
8.07658136e-01 3.24934423e-01 1.29211619e-01 8.70905757e-01
3.51877004e-01 -3.81597467e-02 -1.04421294e+00 1.10105932e+00
-4.58701812e-02 -1.28294206e+00 6.33880347e-02 1.89600382e-02
6.37094200e-01 -8.26242268e-02 -9.83510241e-02 1.71636697e-02
1.25098705e-01 -1.02408850e+00 7.68349767e-01 7.55358994e-01
5.11838675e-01 -8.53583038e-01 4.12494659e-01 2.26669863e-01
-1.67594624e+00 -3.28124046e-01 -3.81642789e-01 -2.12220326e-01
2.09103137e-01 2.27738708e-01 -2.64406472e-01 6.40855312e-01
1.04691684e+00 1.43083930e+00 -4.10426795e-01 6.62971497e-01
-2.71413654e-01 3.22073907e-01 -2.14328334e-01 3.34483445e-01
4.49043930e-01 -1.43618863e-02 2.18996301e-01 1.14007056e+00
3.66960734e-01 1.47419691e-01 5.39946742e-02 4.00666416e-01
9.43247899e-02 -2.19756171e-01 -3.21400106e-01 -1.75069690e-01
-4.60799970e-02 1.02615178e+00 -6.05791032e-01 -4.51260865e-01
-5.31957567e-01 1.30348110e+00 2.41269425e-01 3.23595047e-01
-1.04149520e+00 -2.39678949e-01 9.29231107e-01 -3.90459336e-02
6.13720775e-01 -7.27936506e-01 1.95342988e-01 -1.18435836e+00
2.35296905e-01 -8.28936994e-01 5.54915369e-01 -7.57990897e-01
-9.85122859e-01 2.35808790e-01 1.31386608e-01 -1.30335963e+00
-2.29068771e-01 -7.68339515e-01 -4.24848169e-01 3.21545601e-01
-9.65760469e-01 -1.33812308e+00 -4.12790924e-01 1.24363530e+00
9.85785723e-01 -1.50171518e-01 7.58619010e-01 3.93611938e-01
-4.23274249e-01 2.46831968e-01 -1.53017178e-01 7.35184431e-01
3.03143591e-01 -9.27568257e-01 4.13300425e-01 7.61886001e-01
3.74119788e-01 5.26335597e-01 2.02638403e-01 -5.98789930e-01
-1.33917940e+00 -1.11467803e+00 5.80167532e-01 -3.02390665e-01
7.55821526e-01 -7.36135468e-02 -6.45161808e-01 7.70642519e-01
3.58196534e-02 4.31371272e-01 5.11564016e-01 -1.89986274e-01
-6.03543043e-01 -4.60265838e-02 -7.19642460e-01 6.13626719e-01
1.60392523e+00 -6.07582450e-01 -7.57316113e-01 3.74354661e-01
4.87503648e-01 -3.54386508e-01 -1.05171204e+00 3.67623776e-01
9.46103215e-01 -1.09788215e+00 1.30037212e+00 -8.52995336e-01
7.19176412e-01 -3.54594529e-01 -3.52470964e-01 -7.65617311e-01
-4.16803509e-01 -3.56690735e-01 -3.67084622e-01 5.81158638e-01
-3.18139166e-01 -2.28084847e-01 9.66799438e-01 1.61307395e-01
-1.69693485e-01 -9.63900566e-01 -1.37784076e+00 -7.87868142e-01
-4.02213603e-01 -7.62410522e-01 2.09312916e-01 6.61876500e-01
-2.56155670e-01 3.95381004e-02 -7.12838173e-01 -8.16841144e-03
6.05152488e-01 -9.20612961e-02 7.15126514e-01 -9.00419354e-01
-5.79405367e-01 -2.97208786e-01 -9.61105287e-01 -1.45070279e+00
2.88016707e-01 -4.93637681e-01 -2.25652188e-01 -1.20916581e+00
2.57105708e-01 3.18738997e-01 -4.66423959e-01 7.32863486e-01
3.09532404e-01 3.89791071e-01 1.42644346e-01 2.56251574e-01
-7.63081729e-01 7.73283362e-01 1.41871607e+00 -1.44388318e-01
2.07258940e-01 -1.22747026e-01 6.26370236e-02 1.04051888e+00
5.32416701e-01 -1.76008612e-01 -3.80819976e-01 -4.66213107e-01
-2.39632204e-01 1.79617479e-01 7.94519424e-01 -1.47123122e+00
2.00274140e-01 -1.53690189e-01 8.28732669e-01 -6.81960583e-01
7.76964247e-01 -7.54186630e-01 2.25435555e-01 5.88882744e-01
-3.83620113e-01 2.11010590e-01 -1.54889986e-01 9.65459645e-01
-1.81284085e-01 1.56228900e-01 7.19231606e-01 -4.17535335e-01
-9.06656086e-01 5.85234880e-01 -3.32668513e-01 -5.44996150e-02
1.11856043e+00 -4.73884553e-01 -1.13117741e-03 -2.24270090e-01
-1.05730367e+00 -6.19703680e-02 8.82280022e-02 5.34855127e-01
7.83096969e-01 -1.59002411e+00 -2.81411320e-01 1.69223323e-01
-4.68741693e-02 -1.62760183e-01 6.26539469e-01 1.13028216e+00
-5.08881688e-01 3.25516909e-01 -9.04863417e-01 -6.92663312e-01
-1.40727246e+00 2.63634533e-01 3.98378909e-01 -3.47523093e-01
-1.06708801e+00 7.74740577e-01 2.37504482e-01 5.98021373e-02
2.13486299e-01 -5.06131351e-01 -2.25610450e-01 -5.63702323e-02
4.74697322e-01 6.42582595e-01 -1.91368774e-01 -1.05210280e+00
-3.18185300e-01 8.47499490e-01 1.06725588e-01 -9.66615304e-02
1.47634900e+00 7.93943405e-02 5.75773744e-03 2.74784476e-01
1.52176404e+00 -3.44534367e-01 -1.73057449e+00 -4.23613489e-01
-6.63793981e-02 -7.05996037e-01 -3.35663915e-01 -2.35337585e-01
-1.55553961e+00 1.02281404e+00 5.26478827e-01 -3.81975621e-01
1.13826632e+00 -6.41585663e-02 1.21137857e+00 4.25493896e-01
3.02021474e-01 -1.38894081e+00 8.58052373e-01 5.74597538e-01
9.31627572e-01 -6.75793827e-01 8.46108701e-03 -1.56869203e-01
-6.58191979e-01 1.41713476e+00 5.53510606e-01 -4.04222876e-01
5.92493653e-01 -7.76101202e-02 -2.40919769e-01 -3.58686775e-01
-4.83159870e-01 -1.62968487e-01 3.77181381e-01 5.30310452e-01
2.71347582e-01 2.76648495e-02 -2.90447772e-01 5.31188011e-01
1.19415857e-01 8.42321739e-02 9.15660635e-02 1.07652700e+00
-3.70783985e-01 -1.00945282e+00 -1.85501799e-01 2.52584726e-01
-5.00636816e-01 3.55386704e-01 -2.05756083e-01 8.90110433e-01
3.63613635e-01 5.37782133e-01 3.30284536e-02 -5.87228060e-01
5.21364152e-01 2.89445240e-02 6.26671731e-01 -3.82956535e-01
-2.63502270e-01 1.57870367e-01 -3.91981713e-02 -1.22211039e+00
-8.07778895e-01 -8.73094738e-01 -1.64729226e+00 7.52248149e-03
3.05416614e-01 -2.76021838e-01 1.22411616e-01 1.10264945e+00
4.28969026e-01 5.82396507e-01 1.98546752e-01 -1.03143394e+00
-5.67576766e-01 -8.74056220e-01 -7.33938575e-01 5.80801487e-01
1.87078059e-01 -1.04900610e+00 4.29190733e-02 3.30056012e-01] | [7.863264560699463, 0.357166588306427] |
1f925c57-e5d4-4246-abc4-3915be6f5947 | pop-mining-potential-performance-of-new | 2207.11001 | null | https://arxiv.org/abs/2207.11001v1 | https://arxiv.org/pdf/2207.11001v1.pdf | POP: Mining POtential Performance of new fashion products via webly cross-modal query expansion | We propose a data-centric pipeline able to generate exogenous observation data for the New Fashion Product Performance Forecasting (NFPPF) problem, i.e., predicting the performance of a brand-new clothing probe with no available past observations. Our pipeline manufactures the missing past starting from a single, available image of the clothing probe. It starts by expanding textual tags associated with the image, querying related fashionable or unfashionable images uploaded on the web at a specific time in the past. A binary classifier is robustly trained on these web images by confident learning, to learn what was fashionable in the past and how much the probe image conforms to this notion of fashionability. This compliance produces the POtential Performance (POP) time series, indicating how performing the probe could have been if it were available earlier. POP proves to be highly predictive for the probe's future performance, ameliorating the sales forecasts of all state-of-the-art models on the recent VISUELLE fast-fashion dataset. We also show that POP reflects the ground-truth popularity of new styles (ensembles of clothing items) on the Fashion Forward benchmark, demonstrating that our webly-learned signal is a truthful expression of popularity, accessible by everyone and generalizable to any time of analysis. Forecasting code, data and the POP time series are available at: https://github.com/HumaticsLAB/POP-Mining-POtential-Performance | ['Marco Cristani', 'Geri Skenderi', 'Christian Joppi'] | 2022-07-22 | null | null | null | null | ['new-product-sales-forecasting'] | ['time-series'] | [ 6.31624386e-02 -1.94158673e-01 -4.86266375e-01 -6.19350791e-01
-6.58096969e-01 -9.49917018e-01 8.59442174e-01 3.42070907e-01
-2.21361704e-02 4.40166295e-01 3.91420841e-01 2.52957474e-02
-1.92057770e-02 -8.68658662e-01 -1.35958600e+00 -4.59336102e-01
-1.64694473e-01 5.71593106e-01 -2.07716569e-01 -2.19540507e-01
-6.36675954e-02 -3.52324136e-02 -1.66784120e+00 6.25453532e-01
2.08866909e-01 1.50734985e+00 1.14411071e-01 6.36265337e-01
5.16622126e-01 5.36369383e-01 -2.86409140e-01 -1.05639839e+00
5.83260715e-01 8.08025524e-02 -4.07516569e-01 1.71967044e-01
8.03282559e-01 -4.79476422e-01 -3.63274902e-01 8.74790430e-01
-2.08844915e-01 -1.70604572e-01 3.65409344e-01 -1.26432693e+00
-1.24837601e+00 8.36751282e-01 -4.05829906e-01 3.54513079e-01
5.00632703e-01 5.08375347e-01 1.25497615e+00 -5.86664259e-01
1.16086864e+00 9.31561172e-01 8.10761452e-01 -6.93758056e-02
-1.56188250e+00 -6.38127863e-01 3.91506404e-01 2.16522574e-01
-1.08421898e+00 -2.23360151e-01 9.15051460e-01 -7.79505730e-01
3.90884101e-01 5.39899349e-01 1.16456008e+00 1.94112110e+00
4.18171346e-01 9.12446320e-01 1.41886890e+00 9.07911435e-02
2.14143500e-01 2.45335937e-01 3.49247940e-02 4.62267965e-01
-1.88260397e-03 3.19870353e-01 -6.16366267e-01 1.00696646e-02
5.00870287e-01 1.13402568e-01 2.69878477e-01 -3.70337553e-02
-1.25580800e+00 5.93957841e-01 5.15990257e-01 1.26265183e-01
-7.99908221e-01 4.56526950e-02 3.01230192e-01 4.99791026e-01
9.51216578e-01 4.11435664e-01 -1.07602596e+00 -2.93835580e-01
-1.00431752e+00 6.38448477e-01 7.13405848e-01 8.22104394e-01
5.52245021e-01 -3.35596561e-01 -2.12137364e-02 4.59169149e-01
-1.91591129e-01 5.89731753e-01 2.14346334e-01 -1.05202425e+00
8.95034745e-02 4.12261665e-01 3.38953048e-01 -1.16744983e+00
-1.71340287e-01 -7.22159147e-01 -3.71522665e-01 -2.61007816e-01
7.58860052e-01 1.52550519e-01 -8.40378940e-01 1.87492633e+00
1.92919031e-01 8.51466507e-02 -5.11065543e-01 9.62328255e-01
4.40228492e-01 8.23878109e-01 3.57282609e-01 -2.52805948e-01
1.42194593e+00 -6.14406765e-01 -3.00484180e-01 -2.32941136e-01
8.21221918e-02 -8.12301338e-01 1.12757480e+00 7.49434531e-01
-8.56659830e-01 -9.72471178e-01 -9.59844887e-01 2.65810668e-01
-4.81869727e-01 -5.88410022e-03 9.88512158e-01 2.77677387e-01
-7.07668781e-01 8.54065120e-01 -6.32310033e-01 -4.73396212e-01
3.37682962e-01 3.33357230e-02 -4.23038661e-01 -1.83085069e-01
-1.04814851e+00 5.86668551e-01 1.47994652e-01 -7.13875815e-02
-1.10446990e+00 -1.06982946e+00 -5.26439548e-01 -1.13408580e-01
4.01409835e-01 -5.70235610e-01 1.21605706e+00 -1.60607350e+00
-9.96453106e-01 1.04544103e+00 1.41469985e-01 -6.51073277e-01
7.30765820e-01 -2.37971157e-01 -1.07030857e+00 -4.29649472e-01
1.97246313e-01 6.84016645e-01 1.02493405e+00 -1.23245060e+00
-7.14659035e-01 -5.11415482e-01 3.16942018e-03 -3.34689915e-01
-8.46402794e-02 -2.77265340e-01 -4.93591666e-01 -7.77362704e-01
-7.37792328e-02 -1.30888915e+00 2.26218328e-02 -1.64903626e-01
-2.41756603e-01 -1.64079085e-01 2.05685660e-01 -1.19920599e+00
1.21251261e+00 -2.28467989e+00 -1.48584127e-01 1.78037718e-01
3.83771211e-02 -4.25272137e-01 -2.42002487e-01 4.64648634e-01
-5.24756573e-02 -1.07656457e-01 5.00818670e-01 -1.46426857e-01
1.99601620e-01 2.61166900e-01 -5.25651336e-01 4.49390948e-01
6.00358471e-02 1.07896066e+00 -1.03187132e+00 -2.28677448e-02
8.58996287e-02 1.09667152e-01 -4.96274173e-01 1.83769926e-01
-8.97038639e-01 6.07219994e-01 -1.37756735e-01 9.57376778e-01
6.93072200e-01 -4.13791597e-01 3.50708097e-01 -6.93147719e-01
-1.86807856e-01 -1.65014461e-01 -7.29860783e-01 1.61294472e+00
-2.99966961e-01 3.49965602e-01 -2.72674352e-01 -3.46634984e-01
8.65777373e-01 -2.25567237e-01 6.57208145e-01 -1.08076525e+00
1.07091844e-01 4.43714932e-02 -2.31766522e-01 -4.79580283e-01
7.23741531e-01 8.53239447e-02 -5.15582263e-01 -2.60251686e-02
2.08335668e-01 4.18472558e-01 2.93050796e-01 -4.73109493e-03
1.21026719e+00 2.43958145e-01 -1.91687614e-01 -1.37619644e-01
-1.19237453e-01 2.19206259e-01 7.41349041e-01 5.60782313e-01
-1.44672275e-01 2.27721527e-01 4.15624440e-01 -8.44446063e-01
-1.64296699e+00 -1.54143310e+00 -1.13174751e-01 1.49572444e+00
1.04646804e-02 -6.00024104e-01 -2.35521525e-01 -6.10688031e-01
4.91583765e-01 8.75797570e-01 -1.07596767e+00 1.68692935e-02
-2.87101299e-01 -4.01281565e-01 -1.20798193e-01 3.03289026e-01
1.01658329e-01 -7.27989793e-01 -7.76997730e-02 2.70094752e-01
-2.17143461e-01 -1.16313171e+00 -3.66751552e-01 -7.44145289e-02
-4.94385689e-01 -8.36326778e-01 -4.51057911e-01 -3.69644165e-01
3.27516973e-01 -3.72647226e-01 1.49844146e+00 -3.75413030e-01
-4.16775607e-02 3.61505270e-01 -3.20080489e-01 -2.66586512e-01
-5.83999574e-01 3.44420895e-02 2.86602825e-01 5.23152053e-01
4.10920769e-01 -6.73666656e-01 -9.36447501e-01 4.35558200e-01
-5.44306695e-01 2.82596558e-01 4.28838730e-01 6.31417274e-01
7.95118034e-01 1.83878422e-01 3.36926669e-01 -6.81832612e-01
2.90487528e-01 -9.91331458e-01 -6.37688339e-01 1.46211848e-01
-7.78662384e-01 -2.33493075e-01 6.54677153e-01 -7.38173842e-01
-8.34188163e-01 2.03988284e-01 -1.28704727e-01 -5.73759377e-01
-2.56045222e-01 4.32269394e-01 1.67545930e-01 6.61491573e-01
6.26207888e-01 -3.34863961e-02 -2.64218479e-01 -8.43973935e-01
6.88179553e-01 3.40530336e-01 9.16173100e-01 -6.68104470e-01
6.03959620e-01 4.91977096e-01 -2.43585154e-01 -3.56625676e-01
-1.12448132e+00 -3.60861987e-01 -5.73428810e-01 -6.84914827e-01
7.23469019e-01 -1.12681532e+00 -1.04341042e+00 3.53043377e-01
-6.04109228e-01 -3.04687589e-01 -4.75048631e-01 2.60417968e-01
-5.62839329e-01 -2.40980864e-01 -6.23583853e-01 -8.06325376e-01
-9.26276892e-02 -6.15072489e-01 9.38018084e-01 1.97364837e-02
-6.45622730e-01 -7.20132351e-01 2.87002437e-02 4.94726241e-01
1.96832657e-01 7.60718703e-01 8.64876688e-01 -5.57748973e-01
-5.54606497e-01 -4.29886550e-01 5.26722595e-02 3.51018369e-01
-1.97464138e-01 2.24943310e-01 -9.26341355e-01 -2.70206481e-01
-2.95505911e-01 -1.76531494e-01 7.73139954e-01 4.62643802e-01
1.09474039e+00 -8.65525663e-01 -1.46464869e-01 4.03107762e-01
1.44548810e+00 -1.68671191e-01 2.84464657e-01 3.44502330e-01
1.94491997e-01 5.13956606e-01 8.52898717e-01 5.74391484e-01
5.10651946e-01 7.52693951e-01 5.56966841e-01 6.74107969e-02
2.74212845e-02 -9.81170118e-01 5.08929670e-01 7.31180787e-01
-1.69597492e-01 -1.63169503e-01 -5.43760180e-01 6.92541897e-01
-1.91145205e+00 -1.02008331e+00 -1.85283393e-01 1.98906255e+00
7.38699436e-01 5.24436355e-01 5.17509043e-01 -1.55100018e-01
3.42035949e-01 1.33534864e-01 -8.49767148e-01 -3.36743146e-01
-3.62128496e-01 -2.52221283e-02 8.17779541e-01 5.93913309e-02
-1.13757133e+00 6.97634041e-01 6.24907207e+00 7.27417290e-01
-9.16882217e-01 4.54451323e-01 1.05520725e+00 -3.64556253e-01
-3.33665639e-01 3.96391191e-02 -6.38337731e-01 9.83686864e-01
1.32485664e+00 -1.27658531e-01 7.82222211e-01 1.13199115e+00
4.15519141e-02 -2.07439467e-01 -1.30931246e+00 8.94219458e-01
-1.34509373e-02 -1.39480925e+00 -1.24353327e-01 1.47243842e-01
7.62879133e-01 2.03702420e-01 6.38092995e-01 3.99285316e-01
5.07506490e-01 -6.05569124e-01 1.34064257e+00 9.41251516e-01
8.93357813e-01 -2.92242020e-01 4.08242464e-01 3.10550600e-01
-8.04941773e-01 -5.20469248e-01 -7.64079615e-02 -1.46280214e-01
3.44750464e-01 7.77990282e-01 -8.36765528e-01 3.86498302e-01
9.30196762e-01 9.60650742e-01 -1.05699515e+00 4.08039689e-01
1.82505235e-01 8.17063987e-01 -3.37839782e-01 2.27275163e-01
2.69289408e-02 -8.56052712e-02 5.08815944e-01 9.08257306e-01
4.40508425e-01 -2.08516002e-01 2.21111283e-01 9.72931743e-01
-2.03642040e-01 -9.53290388e-02 -3.98370296e-01 -3.38618547e-01
1.12022035e-01 1.40925753e+00 -3.94928068e-01 -2.35108927e-01
-3.02060187e-01 1.04866147e+00 7.98679423e-03 1.49973124e-01
-1.02567899e+00 8.20908427e-01 7.61734188e-01 7.50951409e-01
3.89748245e-01 -2.37064481e-01 -1.42028853e-01 -1.05522704e+00
4.38435562e-02 -8.00211728e-01 5.36719322e-01 -1.18179214e+00
-1.93623543e+00 4.07190531e-01 -1.35602355e-01 -1.17768228e+00
1.20201008e-02 -6.78953707e-01 -4.19518836e-02 4.60719645e-01
-9.36866343e-01 -1.73366308e+00 -2.22094372e-01 3.30687493e-01
6.13237917e-01 -9.93210450e-02 6.38948441e-01 2.18808249e-01
-2.05618128e-01 5.15339673e-01 1.50805980e-01 -1.64863482e-01
8.44450474e-01 -9.94713306e-01 6.68667436e-01 4.53984261e-01
4.09926683e-01 3.46830666e-01 1.15755260e+00 -1.02364874e+00
-1.64368165e+00 -1.00717247e+00 6.87777162e-01 -1.03396690e+00
1.17884874e+00 -5.58252752e-01 -4.52750921e-01 9.52074766e-01
-1.15625635e-01 -2.78016567e-01 5.46658456e-01 5.98691463e-01
-5.75692117e-01 -5.57835519e-01 -9.40741718e-01 3.52059275e-01
1.14268923e+00 -3.29034001e-01 -3.98787200e-01 5.47537565e-01
6.75795138e-01 -2.51771361e-01 -1.39240241e+00 4.07352954e-01
1.15368116e+00 -8.50213706e-01 8.74584138e-01 -7.54117489e-01
8.16887140e-01 -4.62074019e-02 -6.73533916e-01 -1.20094931e+00
-7.69941449e-01 -3.60671490e-01 -3.22157770e-01 1.38536417e+00
4.77431357e-01 -8.17244202e-02 7.41063952e-01 6.08819246e-01
1.72987089e-01 -7.34239459e-01 -8.33248138e-01 -8.67461562e-01
-2.71574229e-01 -7.10448623e-01 8.56235087e-01 1.05274808e+00
-3.56458902e-01 1.58675507e-01 -6.84625506e-01 1.59756362e-01
6.20543182e-01 2.90749133e-01 8.79601836e-01 -1.28690326e+00
-6.64691806e-01 -1.58332184e-01 -4.26960021e-01 -7.85100698e-01
-8.87720510e-02 -9.83848572e-01 -4.76718038e-01 -9.91477430e-01
3.60395104e-01 -5.19703507e-01 -5.67573965e-01 3.75490963e-01
2.06945360e-01 4.64588940e-01 3.63365054e-01 4.14489806e-01
-7.15032160e-01 -4.22983570e-03 1.20059609e+00 -3.71183664e-01
1.02822147e-01 7.99432620e-02 -6.96799040e-01 6.71770334e-01
5.48723102e-01 -3.42610180e-01 3.04865763e-02 -1.49423048e-01
7.61440396e-01 -3.79865207e-02 7.30965197e-01 -9.68014479e-01
-1.24172650e-01 -2.22345874e-01 9.65927720e-01 -5.96180320e-01
5.19828379e-01 -9.61468816e-01 1.12290895e+00 3.18032503e-01
-3.68524164e-01 3.47529680e-01 -3.05200126e-02 8.57043266e-01
2.97196358e-01 4.22549605e-01 3.07926744e-01 -1.99838847e-01
-1.01366842e+00 3.82579595e-01 1.56080529e-01 -3.43070060e-01
9.47892308e-01 6.48566633e-02 -3.68358970e-01 -3.03539664e-01
-1.27510726e+00 -1.16806351e-01 7.04718292e-01 9.46036756e-01
-1.65054370e-02 -1.40922940e+00 -8.00830662e-01 3.52496088e-01
4.40580308e-01 -9.99416769e-01 5.81697345e-01 5.99088430e-01
2.70410981e-02 8.67302492e-02 -4.78045344e-01 -7.04775035e-01
-1.00904238e+00 1.15525913e+00 -2.50482023e-01 -3.28609288e-01
-4.20263231e-01 6.75799549e-01 4.05183770e-02 -4.43625189e-02
-2.00117633e-01 -4.80847061e-01 3.38859588e-01 2.09686622e-01
2.54986167e-01 1.97665483e-01 -2.14015767e-02 -6.37144566e-01
-1.46617759e-02 2.57936686e-01 -1.53842092e-01 9.57988724e-02
1.56114960e+00 -2.22109675e-01 1.30099431e-01 1.00138080e+00
1.04321659e+00 -1.24216927e-02 -1.76902723e+00 -2.50511616e-01
-7.47039020e-02 -8.11765313e-01 -1.50274038e-01 -1.53524876e+00
-9.33065355e-01 1.27522781e-01 8.31963181e-01 4.77291405e-01
1.02162087e+00 3.94954413e-01 1.00448370e+00 -3.49312067e-01
7.01672256e-01 -1.26903331e+00 -1.31435022e-01 1.40570272e-02
1.15308082e+00 -1.25417829e+00 1.41818807e-01 -9.56433192e-02
-7.50654340e-01 7.02697575e-01 3.17194402e-01 -1.01123489e-01
8.12640369e-01 -3.15127298e-02 -3.01817954e-01 -2.68134594e-01
-1.00771701e+00 9.80553627e-02 4.89932507e-01 3.50694209e-01
7.55945221e-02 6.87985182e-01 -4.35575470e-02 1.16313279e+00
-7.42729247e-01 1.72270373e-01 -5.33088483e-02 3.03408921e-01
6.95876330e-02 -7.90452003e-01 -1.49202242e-01 7.82295585e-01
-3.06651026e-01 7.92118162e-02 -1.66854605e-01 4.85412031e-01
7.82658756e-01 7.09028661e-01 3.64189625e-01 -7.74859428e-01
6.54896975e-01 -6.96293369e-04 6.29318893e-01 4.15134523e-03
-5.79770207e-01 6.33305013e-02 4.25780356e-01 -7.14819431e-01
-3.63547616e-02 -9.33738172e-01 -1.95542529e-01 -8.90330255e-01
8.21302086e-02 -2.54905254e-01 7.73459494e-01 3.75100374e-01
3.91903013e-01 2.23298103e-01 7.91962385e-01 -8.74074340e-01
-2.90299416e-01 -7.99220622e-01 -7.45478272e-01 1.12262070e+00
1.20025232e-01 -5.44650316e-01 -1.34992614e-01 5.66906333e-01] | [7.198333740234375, 2.8070719242095947] |
e0250228-1d48-4bbc-b43e-48fc310720fa | xai-in-the-context-of-predictive-process | 2202.08265 | null | https://arxiv.org/abs/2202.08265v1 | https://arxiv.org/pdf/2202.08265v1.pdf | XAI in the context of Predictive Process Monitoring: Too much to Reveal | Predictive Process Monitoring (PPM) has been integrated into process mining tools as a value-adding task. PPM provides useful predictions on the further execution of the running business processes. To this end, machine learning-based techniques are widely employed in the context of PPM. In order to gain stakeholders trust and advocacy of PPM predictions, eXplainable Artificial Intelligence (XAI) methods are employed in order to compensate for the lack of transparency of most efficient predictive models. Even when employed under the same settings regarding data, preprocessing techniques, and ML models, explanations generated by multiple XAI methods differ profoundly. A comparison is missing to distinguish XAI characteristics or underlying conditions that are deterministic to an explanation. To address this gap, we provide a framework to enable studying the effect of different PPM-related settings and ML model-related choices on characteristics and expressiveness of resulting explanations. In addition, we compare how different explainability methods characteristics can shape resulting explanations and enable reflecting underlying model reasoning process | ['Manfred Reichert', 'Mervat Abuelkheir', 'Ghada ElKhawaga'] | 2022-02-16 | null | null | null | null | ['predictive-process-monitoring'] | ['time-series'] | [ 4.50205922e-01 6.13981247e-01 -4.45477486e-01 -5.16074300e-01
8.19102377e-02 -4.04212981e-01 1.00288880e+00 5.55044651e-01
2.38520354e-01 3.05350304e-01 4.05997545e-01 -7.29986370e-01
-7.39887834e-01 -7.87295878e-01 -3.84791851e-01 -1.46600515e-01
1.15478233e-01 4.42931831e-01 -3.33097905e-01 2.61725426e-01
6.22377276e-01 5.13282061e-01 -1.52156878e+00 5.94710886e-01
9.01385605e-01 7.48642802e-01 -1.33311629e-01 4.09872442e-01
-5.75614989e-01 1.19906676e+00 -4.45309430e-01 -5.11249840e-01
3.23796242e-01 -4.00460958e-01 -6.46741331e-01 8.46479088e-02
-1.78091988e-01 7.48193860e-02 1.42926633e-01 5.54295599e-01
-2.62364298e-01 -2.93495864e-01 9.01296258e-01 -1.57764709e+00
-4.41818327e-01 1.07530856e+00 -2.20399737e-01 -4.28102762e-02
5.22330284e-01 5.87937891e-01 1.02015460e+00 -4.13657904e-01
6.34231389e-01 1.31772435e+00 4.09777850e-01 3.89853477e-01
-1.39427269e+00 -5.27134597e-01 3.66657019e-01 1.21376328e-01
-8.56127560e-01 -3.67265582e-01 6.73491836e-01 -5.79443693e-01
8.15658808e-01 6.52165890e-01 7.71122098e-01 1.06409121e+00
4.59593445e-01 4.29537624e-01 1.32821929e+00 -6.49653852e-01
4.92032588e-01 4.79272962e-01 3.92093241e-01 1.67666495e-01
7.90426791e-01 1.47339880e-01 -6.38814688e-01 -4.03596669e-01
5.33407748e-01 3.46459121e-01 -1.74467787e-01 -2.50230968e-01
-9.85348165e-01 8.36566448e-01 -5.97106032e-02 3.99595201e-01
-4.76274431e-01 4.14882489e-02 1.20830312e-01 5.32309294e-01
2.50025112e-02 1.23195636e+00 -6.28257811e-01 -3.11277330e-01
-6.79864109e-01 4.93484348e-01 1.24865270e+00 8.54261041e-01
6.66325688e-01 -5.05178757e-02 -4.84228075e-01 -3.83099318e-02
6.86669052e-01 -7.48904943e-02 2.69690722e-01 -1.21082354e+00
3.38512868e-01 1.33758891e+00 3.08949202e-01 -7.61350036e-01
-2.50016540e-01 -4.13059443e-01 -4.01714176e-01 4.10627425e-01
5.04021764e-01 1.08738102e-01 -5.17519414e-01 1.18767631e+00
-1.29710525e-01 -3.47315431e-01 4.96472698e-03 4.90282774e-01
2.40519896e-01 3.19173723e-01 4.48768049e-01 -3.92207861e-01
1.19740522e+00 -7.08762527e-01 -8.22517514e-01 -3.45269680e-01
6.07365966e-01 -3.92257780e-01 1.19517887e+00 4.90914583e-01
-1.10195708e+00 -4.25233245e-01 -9.24240470e-01 5.25769770e-01
-1.85688466e-01 -3.00819784e-01 8.19862783e-01 7.85965323e-01
-4.97674167e-01 9.34831262e-01 -7.95291483e-01 -3.53369743e-01
3.15469503e-01 2.47718439e-01 -6.22772388e-02 8.61426592e-02
-7.83145487e-01 1.16554284e+00 1.24981701e-01 -3.08727492e-02
-4.44719255e-01 -1.04254341e+00 -4.81766760e-01 5.50845027e-01
5.32952905e-01 -9.39512074e-01 1.26993406e+00 -1.21766472e+00
-1.21993721e+00 3.16735089e-01 -1.55924350e-01 -8.19495320e-01
1.03013599e+00 -1.76189706e-01 -3.98745477e-01 -3.42340380e-01
-1.91477895e-01 9.44869518e-02 6.15325570e-01 -1.45643640e+00
-5.61726213e-01 -5.14742672e-01 -2.93408986e-02 -1.25343278e-01
1.61085963e-01 1.19004421e-01 5.37198447e-02 -5.27666546e-02
9.05848071e-02 -7.24236488e-01 -5.06922841e-01 -3.58017772e-01
-4.17016029e-01 3.30158211e-02 5.89163780e-01 -2.97925502e-01
1.40894341e+00 -1.80832756e+00 -2.45714098e-01 5.31088114e-01
3.45331967e-01 -1.40180767e-01 3.42612684e-01 6.75862908e-01
9.10855606e-02 6.99444711e-01 6.99232072e-02 -2.69170195e-01
4.43009466e-01 2.54913092e-01 -3.73175502e-01 1.18528530e-02
4.64589924e-01 6.49346471e-01 -4.01773661e-01 -2.30250835e-01
5.70174098e-01 1.18363976e-01 -3.51732641e-01 2.97139615e-01
-5.13424695e-01 7.30118632e-01 -4.87456471e-01 6.53967679e-01
3.28239799e-01 -3.19720298e-01 4.54281688e-01 1.58175275e-01
-3.84698361e-01 3.24192941e-01 -1.19776106e+00 8.39216530e-01
-4.44904536e-01 3.51180345e-01 -9.43138599e-02 -4.91107613e-01
1.19333291e+00 3.02398384e-01 3.45505059e-01 -4.82105702e-01
5.10211587e-02 8.47719088e-02 4.30586636e-01 -5.57728887e-01
3.21980476e-01 -1.48114756e-01 3.31444174e-01 6.07889414e-01
-5.60391366e-01 -3.53485160e-02 8.99624377e-02 -1.09550186e-01
1.28656113e+00 1.73883840e-01 7.74688184e-01 -2.89712280e-01
3.92785966e-01 2.86528111e-01 8.20385933e-01 8.73682141e-01
-1.45621970e-01 3.55676949e-01 8.65770817e-01 -8.12779486e-01
-8.68251562e-01 -5.33596277e-01 -5.45774736e-02 5.03658533e-01
3.02791856e-02 -5.86104155e-01 -4.84034300e-01 -6.16998374e-01
2.35525474e-01 1.44470286e+00 -7.34746337e-01 -4.34094481e-02
-1.43853128e-01 -4.62471664e-01 1.88811049e-01 5.03781617e-01
1.59180909e-01 -1.10008097e+00 -1.18791687e+00 4.09427792e-01
3.38345259e-01 -8.62745404e-01 3.99614513e-01 3.74811083e-01
-1.02636874e+00 -1.31060064e+00 2.07884967e-01 4.03991461e-01
3.49944651e-01 -2.01978475e-01 1.39919412e+00 1.17862128e-01
4.22730260e-02 3.45153838e-01 -3.37874740e-01 -1.03521132e+00
-9.70755875e-01 5.97359799e-02 -2.67323941e-01 -1.75058201e-01
5.51588058e-01 -5.36810637e-01 -4.22659010e-01 2.54663259e-01
-7.57549465e-01 1.61046594e-01 7.20312834e-01 2.58085907e-01
3.24335903e-01 5.75332120e-02 3.72606844e-01 -1.23637128e+00
9.69861984e-01 -4.64168727e-01 -4.05257553e-01 5.89893937e-01
-1.53415394e+00 4.14602160e-01 4.12039578e-01 -3.48205328e-01
-1.19476855e+00 -2.52433002e-01 4.35208291e-01 -2.35193133e-01
-5.94645381e-01 7.64064014e-01 -3.21528614e-01 4.10390794e-01
5.85796714e-01 -2.49756262e-01 1.79261550e-01 -2.46455833e-01
1.84378818e-01 2.78835267e-01 2.58213072e-03 -5.19311368e-01
5.71407795e-01 4.16491598e-01 3.37551124e-02 -1.05145834e-01
-5.09296000e-01 -2.40762196e-02 -3.85039210e-01 -1.89111203e-01
6.08206868e-01 -3.01927507e-01 -7.20285773e-01 -3.94905269e-01
-1.02215350e+00 -3.25539201e-01 -5.28467596e-01 3.70593458e-01
-6.14862382e-01 -4.09743711e-02 -1.86799720e-01 -1.27982891e+00
-3.43344897e-01 -1.28273106e+00 3.97957116e-01 2.83820808e-01
-1.09786510e+00 -8.89062405e-01 -8.35079476e-02 4.98066545e-01
5.55944383e-01 3.72679740e-01 1.37576354e+00 -1.26984453e+00
-6.68303668e-01 -1.96727082e-01 -9.41130519e-02 -8.66430998e-03
1.37966886e-01 5.49788415e-01 -8.87517989e-01 3.99542987e-01
1.97767183e-01 5.63461244e-01 1.38066456e-01 4.06360537e-01
1.05886936e+00 -6.65907800e-01 -3.24531555e-01 1.60482362e-01
1.37529218e+00 3.84459972e-01 5.58430612e-01 1.01852560e+00
3.25389266e-01 1.21278799e+00 7.43940353e-01 6.59537375e-01
1.81563929e-01 4.75161105e-01 4.99114066e-01 1.58785626e-01
3.11817318e-01 -1.89770579e-01 2.89919972e-01 -3.63902748e-02
-3.37224752e-01 1.11808456e-01 -1.30619121e+00 1.12680346e-01
-2.03646207e+00 -1.12150919e+00 -6.13654733e-01 2.15444040e+00
2.99907148e-01 5.33626974e-01 1.46638015e-02 4.35096055e-01
3.23622584e-01 -1.96169853e-01 -3.92392576e-01 -8.97457659e-01
1.29927605e-01 -1.97431296e-01 4.83119369e-01 3.25368762e-01
-4.24933732e-01 3.69970679e-01 6.40365362e+00 -1.82390097e-03
-8.00734937e-01 -2.78398573e-01 7.17622399e-01 6.63108677e-02
-8.46785486e-01 4.09254551e-01 -5.39955854e-01 3.99051309e-01
1.15320969e+00 -3.99496168e-01 3.11015129e-01 1.20438182e+00
9.05739963e-01 1.36834115e-01 -1.55621552e+00 4.75922137e-01
-5.64732969e-01 -1.67144501e+00 1.86175898e-01 4.56785858e-01
5.54883778e-01 -6.01556957e-01 1.86857360e-03 1.48351446e-01
3.02052885e-01 -1.38700390e+00 9.53240097e-01 7.19226480e-01
-1.95217133e-03 -6.93716526e-01 8.91800284e-01 1.84705824e-01
-7.96778977e-01 -6.93031013e-01 -7.10371882e-02 -7.38735914e-01
-3.05383950e-02 6.96547389e-01 -1.20751548e+00 6.54568136e-01
6.35712385e-01 2.41040379e-01 -4.68002737e-01 6.44331574e-01
-3.27629119e-01 9.38439131e-01 2.71024376e-01 1.26836166e-01
-1.56928729e-02 -4.30904806e-01 3.73767287e-01 1.10817826e+00
3.15877706e-01 -2.29462638e-01 -1.82261825e-01 1.58582628e+00
5.39528728e-01 6.36276528e-02 -7.77218580e-01 -1.89436883e-01
7.74636209e-01 9.64631259e-01 -5.37221849e-01 -2.48329312e-01
-5.55387259e-01 3.77909213e-01 -1.52865693e-01 3.81749421e-01
-5.26851118e-01 5.29305100e-01 9.05044496e-01 8.20307255e-01
-1.80893332e-01 -1.20782321e-02 -1.37300134e+00 -6.87598109e-01
-1.42139748e-01 -1.31625748e+00 2.90439010e-01 -6.72054529e-01
-1.23901415e+00 3.79564166e-01 5.91872633e-02 -8.79896820e-01
-4.17322546e-01 -3.10471535e-01 -1.12916481e+00 9.10680711e-01
-1.38993001e+00 -1.14109921e+00 -4.87336844e-01 1.31575778e-01
4.14315611e-01 -1.44916847e-01 7.77816057e-01 -5.44752419e-01
-5.65335035e-01 -4.95093204e-02 -2.31566355e-01 -4.94522959e-01
5.74211717e-01 -1.36261964e+00 2.72412419e-01 6.92676365e-01
-3.22290976e-03 9.77751315e-01 1.09354782e+00 -7.45748281e-01
-1.43088579e+00 -8.59779000e-01 1.00694358e+00 -6.89293265e-01
6.92882419e-01 1.91496551e-01 -1.00968707e+00 7.71785498e-01
1.73810929e-01 -6.27017081e-01 9.47759330e-01 4.80757982e-01
-2.08792746e-01 -9.95827168e-02 -1.14522934e+00 7.23450780e-01
8.84735703e-01 -1.38737544e-01 -6.19758189e-01 -2.59062469e-01
5.04117787e-01 1.81704223e-01 -1.19148719e+00 4.62159663e-01
3.42795908e-01 -1.33500135e+00 6.19834840e-01 -6.32878602e-01
5.47202885e-01 -2.56156057e-01 -9.14959386e-02 -8.33531320e-01
-4.33891714e-01 -7.17562318e-01 -3.41385275e-01 1.35162675e+00
7.29736030e-01 -7.86357105e-01 7.02396631e-01 1.70409811e+00
8.91516730e-02 -8.94569874e-01 -3.69590074e-01 -4.54030067e-01
-2.54560947e-01 -7.54025519e-01 1.24784398e+00 8.71650100e-01
2.10296363e-01 3.03842723e-02 5.39844520e-02 1.51870742e-01
4.50433612e-01 4.62615699e-01 1.06189108e+00 -1.65076923e+00
-3.31405401e-01 -6.73867643e-01 -2.95377910e-01 -9.63333249e-02
-3.55556756e-02 -5.14997005e-01 -5.81777513e-01 -1.56655204e+00
1.90646172e-01 -6.00628912e-01 -2.32819393e-01 4.33146179e-01
-2.70510525e-01 -7.07687914e-01 5.69443047e-01 6.38063431e-01
-8.26752931e-02 1.33327246e-01 9.08183515e-01 3.45610119e-02
-6.41228676e-01 3.17032605e-01 -1.21161973e+00 8.19786012e-01
9.92176175e-01 -4.82222259e-01 -5.05584538e-01 -2.58179866e-02
4.11309630e-01 6.80900216e-02 3.36012751e-01 -9.57940042e-01
1.13891870e-01 -6.48662984e-01 2.80511916e-01 -7.54670873e-02
-1.60620719e-01 -1.06805170e+00 9.89582479e-01 7.35113561e-01
-7.04894960e-01 1.43755808e-01 -5.91910742e-02 4.94124115e-01
-2.41067961e-01 -4.18945879e-01 2.19857574e-01 -2.08405748e-01
-4.09494698e-01 -2.40303446e-02 -5.18914342e-01 -4.91521835e-01
1.10102081e+00 -7.64904380e-01 -1.60225376e-01 -3.94328237e-01
-6.27757788e-01 -1.41131934e-02 8.19910347e-01 2.09345087e-01
1.55025512e-01 -6.82078779e-01 -6.22358859e-01 7.02841729e-02
2.22337291e-01 -1.55255385e-02 -1.64711818e-01 8.13388407e-01
-2.54000664e-01 5.01824617e-01 -2.33377680e-01 -2.70624995e-01
-1.08817124e+00 5.83708227e-01 2.02436879e-01 -5.95527172e-01
-5.67120373e-01 8.69257376e-02 6.51392015e-03 -2.79919356e-01
6.04049861e-02 -6.11780763e-01 -1.76640019e-01 -1.90558270e-01
4.16193306e-01 5.61324418e-01 5.99148348e-02 2.43149593e-01
-2.16673762e-01 -1.89709336e-01 5.54467514e-02 -1.20820768e-01
1.37000871e+00 -1.41997084e-01 -2.20781425e-03 6.61805689e-01
-1.88152923e-03 7.80972987e-02 -1.41967654e+00 3.09548408e-01
7.20197976e-01 -5.13555348e-01 -2.71736681e-01 -1.05378079e+00
-6.20789230e-01 6.60059810e-01 1.41901031e-01 6.02644861e-01
7.58923709e-01 4.27604876e-02 -2.52465338e-01 2.24592183e-02
1.78997025e-01 -8.91561687e-01 -3.79669875e-01 -2.30675861e-01
9.04787481e-01 -1.05704939e+00 1.68531820e-01 -7.11495876e-01
-1.06577873e+00 1.23077893e+00 5.86010277e-01 3.93999130e-01
1.91842020e-01 3.97019953e-01 1.54774651e-01 -3.36243033e-01
-1.09402394e+00 2.96688944e-01 -9.18052439e-03 7.48864472e-01
6.36351168e-01 3.39784265e-01 -2.90068179e-01 1.11372566e+00
-1.64617553e-01 2.01334685e-01 5.44822812e-01 9.44795966e-01
-1.72510296e-01 -1.32371902e+00 -6.86978400e-01 7.31109321e-01
-4.81266737e-01 1.82001412e-01 -7.13602126e-01 1.11162257e+00
7.36026391e-02 1.29854739e+00 -2.22705320e-01 -3.09871048e-01
4.17248756e-01 3.77601653e-01 1.21020712e-01 -5.46810925e-01
-1.08071434e+00 -3.42967540e-01 3.49856317e-01 -6.66133881e-01
-1.97320834e-01 -9.53930080e-01 -1.34713757e+00 -4.52988297e-01
-2.75329322e-01 1.66166276e-01 6.99593782e-01 1.17435205e+00
6.05730534e-01 7.68426418e-01 2.29752794e-01 -3.43605310e-01
-8.30420613e-01 -8.97502124e-01 -2.94194251e-01 6.06266439e-01
-3.25311720e-01 -5.32810748e-01 -3.98815095e-01 5.62228709e-02] | [8.703179359436035, 5.947995662689209] |
feee920a-a7ab-4ca0-85ba-ed51333fd7a7 | finding-the-optimal-currency-composition-of | 2303.01909 | null | https://arxiv.org/abs/2303.01909v1 | https://arxiv.org/pdf/2303.01909v1.pdf | Finding the Optimal Currency Composition of Foreign Exchange Reserves with a Quantum Computer | Portfolio optimization is an inseparable part of strategic asset allocation at the Czech National Bank. Quantum computing is a new technology offering algorithms for that problem. The capabilities and limitations of quantum computers with regard to portfolio optimization should therefore be investigated. In this paper, we focus on applications of quantum algorithms to dynamic portfolio optimization based on the Markowitz model. In particular, we compare algorithms for universal gate-based quantum computers (the QAOA, the VQE and Grover adaptive search), single-purpose quantum annealers, the classical exact branch and bound solver and classical heuristic algorithms (simulated annealing and genetic optimization). To run the quantum algorithms we use the IBM Quantum\textsuperscript{TM} gate-based quantum computer. We also employ the quantum annealer offered by D-Wave. We demonstrate portfolio optimization on finding the optimal currency composition of the CNB's FX reserves. A secondary goal of the paper is to provide staff of central banks and other financial market regulators with literature on quantum optimization algorithms, because financial firms are active in finding possible applications of quantum computing. | ['Martin Vesely'] | 2023-03-03 | null | null | null | null | ['portfolio-optimization'] | ['time-series'] | [-2.04369426e-01 -1.60928831e-01 2.10847929e-01 -6.95950240e-02
-5.48346281e-01 -7.52447963e-01 3.75094622e-01 -2.86618173e-01
-7.57152677e-01 1.20436728e+00 -3.48952234e-01 -7.97062218e-01
-5.30357480e-01 -1.46798873e+00 4.58126329e-02 -9.31452096e-01
-4.31971103e-02 1.16148484e+00 -2.01043099e-01 -8.80668044e-01
9.17991579e-01 8.23329091e-01 -1.11141646e+00 1.55753866e-01
6.30883574e-01 9.14739609e-01 -2.85275042e-01 8.20884645e-01
2.07426567e-02 3.14709336e-01 -2.04511210e-01 -1.03580320e+00
8.40088964e-01 -6.86353087e-01 -6.71089590e-01 -5.13869047e-01
-4.79147285e-01 -1.23167917e-01 -6.42383635e-01 1.51762080e+00
7.99353063e-01 3.17795753e-01 5.24377704e-01 -7.04363644e-01
-2.62611419e-01 7.68470228e-01 1.46735176e-01 6.11761093e-01
1.62994072e-01 3.18165600e-01 1.56925249e+00 -2.84536988e-01
5.52928388e-01 8.82361591e-01 1.92059875e-01 3.85178566e-01
-9.34306204e-01 -3.73656631e-01 -9.05761540e-01 7.00918853e-01
-1.42734492e+00 -3.79892737e-01 3.75912100e-01 -1.70254093e-02
1.42688799e+00 6.31510139e-01 7.27516592e-01 8.53157267e-02
6.82339311e-01 1.66948184e-01 1.43588138e+00 -8.49345684e-01
7.38181651e-01 8.68558809e-02 5.28488122e-02 6.06463492e-01
4.07837063e-01 1.08077037e+00 -5.93827605e-01 -5.24863362e-01
5.18146217e-01 -5.40129721e-01 -1.54259622e-01 -1.52259126e-01
-9.78087842e-01 1.23759568e+00 1.28658684e-02 4.38426882e-01
-5.93429565e-01 3.77526581e-01 -7.19040930e-02 6.12764776e-01
-1.70820400e-01 9.47872460e-01 -2.17094094e-01 -4.13048297e-01
-6.92692637e-01 6.74820960e-01 1.31242442e+00 7.20462024e-01
5.79978764e-01 -1.05347857e-01 -4.73566279e-02 -2.59319633e-01
3.88842463e-01 8.38165581e-01 8.96146595e-02 -9.83226359e-01
3.88567448e-01 -2.04500124e-01 3.94650131e-01 -4.33901668e-01
-6.48248792e-01 -3.34404886e-01 -4.37154621e-01 3.31415147e-01
3.54349315e-01 -3.01269114e-01 -2.04319343e-01 1.03039300e+00
2.15238884e-01 -2.63836294e-01 4.22006726e-01 8.45905066e-01
2.47891635e-01 5.47069013e-01 -5.45686960e-01 -6.41461313e-01
1.34521317e+00 -5.73364675e-01 -7.29320109e-01 3.62238646e-01
7.67794788e-01 -1.04295051e+00 -7.54022971e-02 6.01537228e-01
-1.23355913e+00 3.23075324e-01 -1.12229323e+00 5.28054059e-01
-4.71674144e-01 -5.77613950e-01 1.15556765e+00 1.55766320e+00
-1.04005563e+00 1.35677922e+00 -6.67878687e-01 6.17071427e-02
-5.35604283e-02 8.88963699e-01 1.94261044e-01 3.12682360e-01
-1.69755232e+00 1.36037898e+00 7.69471943e-01 1.68628231e-01
-1.00164756e-01 6.92147836e-02 -2.14187384e-01 1.49159059e-01
1.94542542e-01 -9.84400272e-01 1.05152678e+00 -3.63955319e-01
-2.02639198e+00 8.08652878e-01 5.25218725e-01 -8.06822240e-01
2.71474987e-01 8.04991007e-01 -4.46899176e-01 1.89461797e-01
-2.03285053e-01 -2.98709720e-01 1.68181211e-01 2.04352051e-01
-5.52565098e-01 -4.88454223e-01 1.36059999e-01 3.87028724e-01
3.36599797e-01 3.74778837e-01 4.59085017e-01 -2.75282741e-01
3.46050143e-01 -1.16636693e+00 -5.18199742e-01 -1.05362427e+00
-1.05408773e-01 1.52478497e-02 -3.22995722e-01 -3.65369856e-01
1.48588252e+00 -1.43150401e+00 3.21672797e-01 8.12783122e-01
-3.89683902e-01 -2.42829785e-01 1.72578260e-01 9.00958955e-01
-2.19580442e-01 -3.24213766e-02 -2.07287624e-01 4.91093397e-01
7.15799868e-01 1.37483582e-01 -2.10172340e-01 8.12514961e-01
-5.43730319e-01 1.15704834e+00 -6.23960257e-01 -2.24009007e-01
-1.14019867e-02 -5.82753539e-01 -7.27950096e-01 -3.83917838e-01
-1.29332036e-01 9.42233726e-02 -8.13532054e-01 8.66729736e-01
6.38552070e-01 -1.06378958e-01 2.63520956e-01 1.03783444e-01
-4.22470123e-01 4.19606507e-01 -1.44510663e+00 1.37634480e+00
1.89619631e-01 1.89356789e-01 -4.39138375e-02 -9.18045163e-01
5.15457809e-01 2.73850560e-01 5.14267027e-01 -9.37732100e-01
3.65270495e-01 8.36001754e-01 6.01176739e-01 -6.72109947e-02
8.80207717e-01 -1.02086139e+00 -1.94215089e-01 1.08750308e+00
-8.21702406e-02 -5.81126690e-01 4.87846464e-01 -1.14874683e-01
1.19810319e+00 -5.72422892e-02 4.65961546e-01 -4.90249783e-01
5.80645621e-01 1.06840581e-01 2.96499282e-01 8.77443671e-01
-4.33840841e-01 1.61877856e-01 4.94809479e-01 -5.58711529e-01
-1.24784076e+00 -8.69054437e-01 -4.19396907e-01 5.92156529e-01
2.63573021e-01 -4.22845095e-01 -6.67936027e-01 1.41767606e-01
3.51244695e-02 1.11267340e+00 -1.65185764e-01 3.81806977e-02
-5.24635017e-01 -1.58944130e+00 1.75030425e-01 -2.83949792e-01
4.39359993e-01 -1.12399626e+00 -6.92052603e-01 7.38767624e-01
2.39155844e-01 -7.24814355e-01 6.68374971e-02 4.41823304e-01
-8.96475196e-01 -1.02183938e+00 -4.68755871e-01 2.77590573e-01
4.83884709e-03 -3.39284182e-01 1.05532432e+00 -3.09797764e-01
-2.99157947e-01 2.10769385e-01 -3.91078800e-01 -2.63059288e-01
-6.51281238e-01 -9.09180269e-02 1.74252227e-01 -3.06928933e-01
5.87495387e-01 -3.96431893e-01 -5.11790574e-01 1.51150256e-01
-4.76567507e-01 -4.76825774e-01 4.63545054e-01 1.08715713e+00
5.61802804e-01 4.01311934e-01 1.67149387e-03 -7.14779198e-01
6.29411638e-01 -7.86035694e-03 -1.39161885e+00 4.42426920e-01
-8.16070378e-01 4.84816521e-01 2.41900951e-01 5.15233278e-01
-7.52769589e-01 -2.14963078e-01 -1.15010165e-01 1.51928425e-01
6.87824965e-01 4.22304928e-01 -1.60792861e-02 -8.59456003e-01
6.44571900e-01 1.56766266e-01 -1.72002688e-01 -1.15219913e-02
3.28234583e-01 5.57867765e-01 -9.86476243e-02 -6.04844153e-01
6.54379368e-01 2.81430930e-01 7.65714347e-01 -6.72659516e-01
-3.73551160e-01 -2.43067816e-01 -3.15100223e-01 1.98924430e-02
9.55002189e-01 -3.33699584e-01 -1.08286560e+00 2.30283141e-01
-1.01288438e+00 1.73612371e-01 -3.19177389e-01 8.64061773e-01
-1.16329920e+00 5.35511971e-01 -5.91163397e-01 -1.20792651e+00
-5.93760252e-01 -1.22737813e+00 5.13935506e-01 4.15364414e-01
1.94607481e-01 -8.49273980e-01 2.82906681e-01 4.34303194e-01
3.72467637e-01 1.20912477e-01 6.67675138e-01 -7.68328309e-01
-1.23369312e+00 -3.08837235e-01 2.91100889e-01 -5.52714728e-02
-6.29327536e-01 -2.77083278e-01 -6.16102695e-01 -2.68954486e-01
4.40624774e-01 1.56634673e-01 5.63238084e-01 3.95515591e-01
4.15450037e-01 5.94952004e-03 6.57844096e-02 6.67406797e-01
1.57637239e+00 5.69513440e-01 8.67008090e-01 9.87315714e-01
-2.06695527e-01 2.10251659e-01 8.18269849e-01 8.24616015e-01
-1.49243534e-01 8.06852877e-01 3.54212403e-01 1.00002623e+00
8.66343677e-01 5.56896269e-01 2.56330311e-01 7.51095593e-01
-5.94508827e-01 -5.77015355e-02 -7.64365017e-01 -1.42006516e-01
-1.71076000e+00 -1.41919196e+00 -1.64150015e-01 2.36557484e+00
4.54891145e-01 2.85800368e-01 -8.41495767e-02 -9.79951844e-02
7.80436397e-01 -2.49780402e-01 -1.95735633e-01 -1.00065577e+00
-2.94766456e-01 1.17910588e+00 1.24335396e+00 5.12037158e-01
-9.47506905e-01 9.37755287e-01 6.01276445e+00 1.16086316e+00
-4.81081277e-01 2.51731604e-01 3.10097039e-01 -2.87494630e-01
-5.84747612e-01 5.65670073e-01 -8.02508414e-01 5.04403412e-01
1.39940095e+00 -7.18995094e-01 1.16781616e+00 4.88086194e-01
-3.12099066e-02 -3.13867897e-01 -9.73872960e-01 1.09057903e+00
-5.11330664e-01 -1.79355192e+00 -4.54886734e-01 6.24659121e-01
9.19252813e-01 1.14255778e-01 7.06459060e-02 3.24327677e-01
-2.39067692e-02 -9.72237766e-01 7.83720672e-01 7.28079617e-01
4.51278001e-01 -1.28301263e+00 1.23292255e+00 2.14851707e-01
-5.82975268e-01 -9.36360210e-02 -7.59823442e-01 -3.26134503e-01
4.20983225e-01 1.35756135e-01 -4.04731095e-01 9.41618323e-01
3.14634800e-01 -1.81424797e-01 7.57044926e-02 1.17079961e+00
-2.65347678e-02 1.66499496e-01 -7.06579804e-01 -7.03688145e-01
6.40583277e-01 -1.40450621e+00 6.83339596e-01 4.64964032e-01
6.90008283e-01 1.04286385e+00 -4.92650956e-01 1.07384062e+00
2.72998363e-01 1.61874682e-01 -4.50706095e-01 -6.46207392e-01
2.07171768e-01 9.31456268e-01 -8.17986727e-01 -1.17923329e-02
-1.82026684e-01 7.50492513e-01 -3.77573371e-01 1.64791942e-02
-7.22829759e-01 -7.33050644e-01 4.56984043e-01 -2.83859938e-01
3.19842547e-01 -1.60012558e-01 -5.79468250e-01 -1.21969819e+00
-4.67150301e-01 -9.43423808e-01 4.74346071e-01 -4.13092881e-01
-9.52161372e-01 2.14917704e-01 -3.13284487e-01 -9.46439803e-01
-4.66501176e-01 -9.57665026e-01 -4.72439855e-01 1.24362528e+00
-1.20445049e+00 -2.95103639e-01 6.73701763e-01 2.19927400e-01
-4.25944865e-01 -7.68638194e-01 9.87233520e-01 -1.48298796e-02
-2.92340964e-01 4.51666355e-01 1.02062857e+00 -2.91018546e-01
8.62369463e-02 -1.27448452e+00 3.72412801e-01 6.85881734e-01
2.89287776e-01 6.12914681e-01 1.16762483e+00 -5.31378329e-01
-1.96184504e+00 -9.60172266e-02 7.54186511e-01 -4.77447331e-01
1.25946319e+00 1.50650442e-01 -5.43611720e-02 2.07920969e-01
1.25337675e-01 -5.90958118e-01 9.19827163e-01 -1.45646304e-01
5.42408288e-01 1.87969059e-02 -1.24039972e+00 4.99130875e-01
6.30429685e-01 -6.52643442e-01 -5.38199604e-01 9.50769961e-01
2.68644005e-01 -5.83663523e-01 -8.53578687e-01 5.59013374e-02
4.66686040e-01 -1.40217185e+00 9.64493334e-01 -5.81413627e-01
-5.32217836e-03 2.29285359e-02 -4.43070889e-01 -8.42135251e-01
-1.85879454e-01 -1.53070152e+00 2.77640879e-01 2.63305843e-01
4.71500278e-01 -1.19130790e+00 1.13017118e+00 9.61698949e-01
9.90515724e-02 -5.97184658e-01 -1.61051679e+00 -8.51926386e-01
4.20751929e-01 -2.79731572e-01 7.27128088e-01 5.75740337e-01
3.29791307e-01 -4.88402322e-02 -2.27453664e-01 7.87369683e-02
1.06259310e+00 6.74611747e-01 2.55187243e-01 -6.24293327e-01
-1.03843486e+00 -7.92549849e-01 -6.22087836e-01 -3.34155828e-01
-8.21254626e-02 -1.13296974e+00 -5.82932115e-01 -8.42290640e-01
-2.14265361e-02 -4.54146028e-01 -4.86302137e-01 -3.75300318e-01
2.83350021e-01 -1.00397289e-01 3.62784475e-01 1.40161693e-01
-4.28478032e-01 4.94510233e-01 1.16007137e+00 -4.17728052e-02
6.93998262e-02 5.84025860e-01 -3.37418467e-01 2.83377826e-01
7.08234966e-01 -6.24609828e-01 3.07327330e-01 2.56212950e-01
9.16957438e-01 7.99165785e-01 -1.22535139e-01 -7.01137781e-01
5.06003320e-01 -3.32188696e-01 -2.00538531e-01 -6.79285169e-01
2.39975736e-01 -3.84059131e-01 5.65711856e-01 9.48024213e-01
2.91584194e-01 2.41846472e-01 -2.86837041e-01 3.15859258e-01
-1.06119193e-01 -1.36862290e+00 8.04054201e-01 -6.51714921e-01
-5.82483947e-01 3.36133271e-01 -1.86076403e-01 -1.87809497e-01
1.25782704e+00 -1.42764568e-01 -2.50549734e-01 -2.38736108e-01
-9.48246062e-01 1.61309868e-01 6.97254360e-01 -5.74219167e-01
2.97657996e-01 -1.12153661e+00 -5.93427777e-01 -1.32973552e-01
-3.76326889e-01 -8.74069154e-01 2.68775165e-01 7.76717782e-01
-1.45562804e+00 1.11239505e+00 -3.59255165e-01 1.29565671e-01
-6.49287343e-01 4.84720707e-01 8.08814347e-01 -5.76144218e-01
-5.34223579e-02 1.06809342e+00 -4.48342711e-01 -1.97912037e-01
-5.21792769e-01 3.15441996e-01 1.87768713e-01 9.60966572e-02
3.71731877e-01 7.03043997e-01 2.34714285e-01 -4.05744582e-01
-4.64288563e-01 3.65431041e-01 3.61451328e-01 -6.85668886e-01
1.38701153e+00 1.40034080e-01 -7.62179673e-01 -3.83919179e-02
8.06267023e-01 -1.41594872e-01 -2.37697139e-01 1.77193254e-01
3.20083410e-01 -2.91695833e-01 1.71263456e-01 -4.95376676e-01
-5.27425885e-01 7.98635423e-01 6.62574828e-01 3.54856819e-01
8.91209364e-01 -5.74048579e-01 6.75152838e-01 9.65026021e-01
1.31965864e+00 -1.51008022e+00 -7.00997412e-01 6.12853110e-01
3.13610673e-01 -9.19482768e-01 5.96156299e-01 6.30296692e-02
-3.86476934e-01 1.73046851e+00 -5.25957823e-01 -2.11855575e-01
5.94580531e-01 -6.64397925e-02 -5.20156085e-01 -2.89494008e-01
-5.97166598e-01 -6.25395894e-01 1.19084027e-02 -1.21474162e-01
-2.74451841e-02 4.70255226e-01 -1.16169322e+00 4.33435082e-01
-6.45871520e-01 -1.99958906e-01 9.91653383e-01 1.02639139e+00
-6.24486387e-01 -1.78901052e+00 -6.98359013e-01 4.22883987e-01
-6.54341757e-01 -5.65861523e-01 -1.27808377e-01 7.05533564e-01
-1.72806621e-01 6.02476120e-01 -3.44300330e-01 -1.49690151e-01
1.22097626e-01 4.45702896e-02 1.03803062e+00 -5.54873824e-01
-8.27342570e-01 1.31060854e-02 4.38427031e-01 -5.84038556e-01
-1.96824715e-01 -1.06855929e+00 -1.27119982e+00 -7.12236762e-01
-6.94870174e-01 8.73636127e-01 9.40307140e-01 9.12748456e-01
-1.59638345e-01 5.37035614e-02 5.66913188e-01 -5.37999451e-01
-1.72511947e+00 -3.61093909e-01 -1.57163608e+00 -4.68801409e-01
-5.43994248e-01 -5.03341079e-01 -4.90026832e-01 -8.35117996e-01] | [5.557889461517334, 4.926635265350342] |
a85bd6c3-0e1c-4206-8b52-243b4f8e11c6 | merry-go-round-rotate-a-frame-and-fool-a-dnn | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/Thapar_Merry_Go_Round_Rotate_a_Frame_and_Fool_a_DNN_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Thapar_Merry_Go_Round_Rotate_a_Frame_and_Fool_a_DNN_CVPR_2022_paper.pdf | Merry Go Round: Rotate a Frame and Fool a DNN | A large proportion of videos captured today are first per-son videos shot from wearable cameras. Similar to other computer vision tasks, Deep Neural Networks (DNNs) are the workhorse for most state-of-the-art (SOTA) egocentric vision techniques. On the other hand DNNs are known to be susceptible to Adversarial Attacks (AAs) which add im-perceptible noise to the input. Both black-box, as well as white-box attacks on image as well as video analysis tasks have been shown. We observe that most AA techniques basically add intensity perturbation to an image. Even for videos, the same process is essentially repeated for each frame independently. We note that the definition of imperceptibility used for images may not be applicable for videos, where a small intensity change happening randomly in two consecutive frames may still be perceptible. In this paper we make a key novel suggestion to use perturbation in optical flow to carry out AAs on a video analysis system. Such perturbation is especially useful for egocentric videos, because there is a lot of shake in the egocentric videos anyways, and adding a little more, keeps it highly imperceptible. In general, our idea can be seen as adding structured, para-metric noise as the adversarial perturbation. Our implementation of the idea by adding 3D rotations to the frames reveal that using our technique, one can mount a black-box AA on an egocentric activity detection system in one-third of the queries compared to the SOTA AA technique. | ['Chetan Arora', 'Aditya Nigam', 'Daksh Thapar'] | 2022-01-01 | null | null | null | cvpr-2022-1 | ['activity-detection'] | ['computer-vision'] | [ 2.43724018e-01 1.19032443e-01 3.03333431e-01 1.35111839e-01
-1.85995400e-02 -8.91083837e-01 5.18703997e-01 -5.31892836e-01
-6.26608849e-01 5.20614088e-01 1.98152199e-01 -8.93262029e-02
2.53199250e-01 -6.32425070e-01 -1.00063670e+00 -9.53216851e-01
-2.65158355e-01 -3.65564853e-01 5.24863064e-01 -3.34272921e-01
1.36565506e-01 6.66363358e-01 -1.40246534e+00 7.04334257e-03
2.18403518e-01 6.58770621e-01 -1.80860206e-01 1.10383272e+00
5.27056456e-01 1.16661835e+00 -7.99418986e-01 -6.26794577e-01
5.93179524e-01 -2.62507409e-01 -6.65639699e-01 3.66617180e-02
6.25415981e-01 -7.86177576e-01 -9.78442430e-01 1.42104149e+00
5.54988503e-01 1.59893960e-01 3.48808169e-01 -1.32856476e+00
-2.45365992e-01 3.21058869e-01 -5.39471269e-01 5.66646457e-01
6.01808131e-01 3.69937241e-01 2.61174083e-01 -4.41101432e-01
7.47623384e-01 1.10044432e+00 5.07050931e-01 7.05293536e-01
-9.67179596e-01 -5.43300152e-01 -1.23134211e-01 5.22436678e-01
-1.15084207e+00 -4.91997868e-01 1.03825176e+00 -3.28874290e-01
6.38743997e-01 4.65334445e-01 7.13809729e-01 1.42637849e+00
4.49955821e-01 6.14784896e-01 8.10448647e-01 -3.01623434e-01
2.81951278e-01 3.87937725e-02 -3.05513620e-01 4.32844460e-01
3.64701837e-01 1.18160136e-01 -2.03206062e-01 -5.46598397e-02
8.65670085e-01 4.65597101e-02 -6.09304368e-01 -4.92998213e-01
-1.24566579e+00 6.19867802e-01 3.83855999e-01 2.06639528e-01
-2.38577098e-01 4.93241817e-01 6.46083951e-01 5.27330577e-01
1.01394095e-01 4.11194742e-01 -3.47780660e-02 -4.37490761e-01
-5.54650426e-01 1.06947102e-01 7.32338190e-01 7.34276414e-01
6.07241511e-01 4.37802613e-01 1.59877732e-01 1.68614507e-01
-2.40775153e-01 6.21028125e-01 6.24686897e-01 -1.18463933e+00
2.08885804e-01 3.90167627e-03 3.00425068e-02 -1.39221203e+00
-3.52822691e-01 -1.76249564e-01 -8.11668575e-01 5.97916365e-01
5.46641946e-01 -4.88576084e-01 -5.39735258e-01 1.87960458e+00
2.49824598e-01 5.13386250e-01 9.18067396e-02 1.10300493e+00
3.91786128e-01 4.82833505e-01 -2.71938145e-01 -2.66433924e-01
1.29106975e+00 -5.15903473e-01 -8.13538551e-01 -1.80777814e-02
5.98286331e-01 -8.12867045e-01 7.71390676e-01 5.29976070e-01
-9.73433435e-01 -4.38632250e-01 -1.27216434e+00 1.63647324e-01
-3.71719003e-01 -4.96848941e-01 2.65048325e-01 1.15985811e+00
-9.89645720e-01 5.24880171e-01 -8.47419977e-01 -4.39098090e-01
1.16877995e-01 4.21018839e-01 -7.98990905e-01 -1.45067452e-02
-1.28181732e+00 8.68472755e-01 2.51262128e-01 8.51415247e-02
-1.13517499e+00 -3.56129915e-01 -8.35476220e-01 -1.96416378e-02
6.21572316e-01 -6.23337209e-01 9.02300954e-01 -1.39923954e+00
-1.63967621e+00 8.15571666e-01 5.40950857e-02 -7.08406806e-01
7.68534839e-01 -4.19887125e-01 -5.54413736e-01 7.27251768e-01
-3.26516151e-01 4.59473044e-01 1.32504988e+00 -9.81813788e-01
-2.11061668e-02 -2.75308192e-01 6.35267079e-01 1.34275019e-01
-4.21856850e-01 1.77407295e-01 -2.00396866e-01 -8.30704391e-01
-3.18817765e-01 -9.91491795e-01 3.19890194e-02 1.63991511e-01
-3.30886602e-01 2.77338654e-01 1.45862234e+00 -3.33968133e-01
1.04766738e+00 -2.33327913e+00 2.73187701e-02 1.48572018e-02
3.55802923e-01 5.48654079e-01 -5.55251874e-02 4.45501536e-01
-3.86954635e-01 -7.49432147e-02 1.86020155e-02 2.13719532e-03
-2.48416960e-01 1.85824499e-01 -5.15509605e-01 1.00066829e+00
-9.09883901e-02 7.78939188e-01 -1.02553117e+00 -3.05399634e-02
4.25523728e-01 5.05711138e-01 -7.32510626e-01 1.82124317e-01
3.07014525e-01 1.42959967e-01 -1.18673623e-01 3.34045887e-01
7.40017653e-01 3.18483472e-01 -1.23931788e-01 -2.71050900e-01
6.09485731e-02 -2.00735927e-01 -1.13173676e+00 1.39268112e+00
-1.82102293e-01 1.20885348e+00 -4.55488823e-02 -1.12775385e+00
5.35493255e-01 5.20904243e-01 4.08072948e-01 -3.64793748e-01
3.73624653e-01 -9.73908678e-02 2.13644907e-01 -8.28950763e-01
4.20887202e-01 -3.37032229e-02 1.46596625e-01 3.15165669e-01
8.60298499e-02 1.33978158e-01 7.90312961e-02 3.00519556e-01
1.37396550e+00 4.32199426e-02 3.22602779e-01 -1.29500210e-01
7.24210680e-01 -3.75869900e-01 4.32532996e-01 8.99653792e-01
-5.36539137e-01 8.61057699e-01 6.53950691e-01 -4.93612796e-01
-1.14497626e+00 -9.70118582e-01 2.93148130e-01 4.63025361e-01
4.64763582e-01 -4.72327530e-01 -1.21292245e+00 -6.69358432e-01
-4.50777799e-01 3.32975864e-01 -5.95617831e-01 -5.88222206e-01
-6.48855627e-01 -3.62311125e-01 1.00227535e+00 3.50284427e-01
8.32388461e-01 -9.62733209e-01 -9.13668513e-01 2.91874874e-02
-3.97809371e-02 -1.45140243e+00 -4.82681215e-01 -2.48280570e-01
-5.94963074e-01 -1.32782805e+00 -8.37356031e-01 -3.73989582e-01
6.05795443e-01 6.65540874e-01 6.92066848e-01 -1.75917208e-01
-2.33442739e-01 7.53934741e-01 -4.48131949e-01 -2.77324408e-01
-4.02323484e-01 -5.71786165e-01 5.58344126e-01 3.37457746e-01
4.98215467e-01 -7.32062519e-01 -8.22351992e-01 4.23351228e-01
-1.25195706e+00 -3.81323308e-01 1.31996721e-01 5.46147466e-01
-1.14200458e-01 1.37587845e-01 1.06956899e-01 -7.07128227e-01
2.91748017e-01 -2.72465944e-01 -3.67139131e-01 -1.75541580e-01
9.14192721e-02 -3.57587844e-01 1.14526272e+00 -6.71614289e-01
-5.67077816e-01 -8.54058564e-02 3.10413875e-02 -1.08917379e+00
-3.85325015e-01 -3.68010402e-02 -2.56619900e-01 -4.53392297e-01
7.90169597e-01 1.81721687e-01 1.06012866e-01 -2.05264073e-02
3.94087791e-01 4.76117492e-01 7.59414434e-01 -6.56178743e-02
1.18895900e+00 1.09239447e+00 2.22381324e-01 -1.34056950e+00
-1.62657872e-01 -3.94963384e-01 -3.63950521e-01 -5.20316780e-01
9.42304909e-01 -7.78844774e-01 -1.01687932e+00 9.03600931e-01
-1.17123461e+00 -4.62179445e-02 -3.74608606e-01 6.39561236e-01
-7.13274598e-01 9.77481186e-01 -2.90242821e-01 -6.74078226e-01
1.35821909e-01 -1.26052141e+00 6.00757778e-01 1.64243057e-01
-3.38824801e-02 -1.01687634e+00 5.98836038e-03 1.42033383e-01
3.87884587e-01 2.58023858e-01 3.54079515e-01 -4.47355032e-01
-4.42293704e-01 -4.87539738e-01 1.35019109e-01 6.03633523e-01
2.72214383e-01 8.11039656e-03 -1.30956542e+00 -5.37800848e-01
5.64366758e-01 -3.28237526e-02 6.41343117e-01 4.00506586e-01
1.05325472e+00 -5.15832067e-01 1.48813754e-01 8.08926821e-01
1.25273693e+00 3.53887707e-01 1.24707282e+00 4.73917723e-01
7.55690396e-01 4.69479918e-01 3.20334792e-01 2.52128512e-01
-2.30225146e-01 8.80059659e-01 8.69902492e-01 -1.33724242e-01
2.07151230e-02 -3.28435116e-02 9.38941777e-01 2.68396318e-01
-3.10016215e-01 -4.91849244e-01 -4.08996940e-01 2.17439145e-01
-1.49724090e+00 -1.49480140e+00 1.17533663e-02 2.38782907e+00
2.36666515e-01 2.17199877e-01 2.94024825e-01 4.47398871e-01
8.63800943e-01 4.62944061e-01 -2.49163911e-01 -4.71268445e-01
-2.13618979e-01 -1.44506112e-01 8.14896762e-01 3.33045065e-01
-1.36698747e+00 6.72827721e-01 6.11421633e+00 6.75420344e-01
-1.48054969e+00 5.90813793e-02 3.44165891e-01 -2.21633568e-01
1.15473509e-01 3.07294466e-02 -1.82754919e-01 7.25433052e-01
7.00943053e-01 -1.41763568e-01 4.49857175e-01 7.92387366e-01
5.07310808e-01 -2.40188986e-01 -9.41990316e-01 1.37310219e+00
5.04296184e-01 -1.07590127e+00 1.38695627e-01 7.04398677e-02
5.39355338e-01 -2.43861139e-01 1.58777311e-01 -1.61819071e-01
-1.25035465e-01 -8.09612334e-01 5.27534783e-01 2.02273116e-01
6.92095995e-01 -7.54558802e-01 8.72612178e-01 1.57124832e-01
-8.47632885e-01 -9.12140980e-02 -5.39786398e-01 -2.29664922e-01
2.02917874e-01 2.61125565e-01 -4.20666963e-01 4.38256353e-01
6.62201703e-01 6.33925855e-01 -3.60119969e-01 1.03267205e+00
-2.84883052e-01 5.81740022e-01 -4.10048693e-01 2.07069591e-01
3.51466984e-01 -2.81703502e-01 1.17647290e+00 1.10488296e+00
2.34036997e-01 -4.14350890e-02 -4.54619676e-01 3.43437970e-01
-3.98766957e-02 -2.19585210e-01 -1.29956543e+00 2.37461418e-01
2.01692190e-02 1.00928497e+00 -6.56302094e-01 -3.46857041e-01
-5.74535787e-01 1.44707835e+00 -2.79139787e-01 5.72507679e-01
-8.35470140e-01 -6.43529475e-01 8.60078752e-01 4.96270843e-02
3.29050153e-01 -3.10303092e-01 2.33918756e-01 -1.40322495e+00
4.87177670e-02 -9.54855740e-01 1.42841980e-01 -9.97787595e-01
-8.32210600e-01 3.77507657e-01 -7.10768104e-02 -1.79159021e+00
-3.86241764e-01 -6.76317692e-01 -7.67409444e-01 3.10653359e-01
-1.14943969e+00 -5.98534584e-01 -4.52358782e-01 1.02885330e+00
4.96725827e-01 -2.23645523e-01 5.00574589e-01 3.61892283e-01
-4.74608988e-01 6.44723654e-01 1.32349119e-01 4.94163781e-01
7.74550080e-01 -1.05054867e+00 3.06202829e-01 1.48294985e+00
2.57856101e-01 7.37642765e-01 1.15591502e+00 -2.65106857e-01
-1.67941785e+00 -7.72929370e-01 1.98831514e-01 -3.11883628e-01
8.33701372e-01 -3.64310175e-01 -6.24460518e-01 7.42670119e-01
3.68798554e-01 1.79260314e-01 2.93495238e-01 -5.68392992e-01
-3.66847008e-01 -2.09601074e-01 -1.02451527e+00 8.98083568e-01
8.22917879e-01 -7.43695796e-01 -6.65427387e-01 1.97019190e-01
4.91022617e-01 -2.78956532e-01 -3.93651396e-01 1.62971869e-01
6.82844639e-01 -1.38517797e+00 9.59488273e-01 -4.98930454e-01
2.92932481e-01 -5.65142393e-01 -1.49819925e-01 -1.20872629e+00
1.88444871e-02 -1.30153203e+00 -2.73336917e-01 8.39474618e-01
-3.90126407e-01 -7.40704000e-01 7.58242965e-01 3.01800013e-01
8.27992037e-02 -3.02255936e-02 -1.16484320e+00 -1.12093496e+00
-3.20183992e-01 -5.08496463e-01 4.36001495e-02 9.74627852e-01
3.75496373e-02 -6.82668611e-02 -8.18045080e-01 3.87325853e-01
6.44015372e-01 -7.23307014e-01 1.23778450e+00 -5.84663808e-01
-3.89871120e-01 -2.50001222e-01 -1.34365201e+00 -1.02706110e+00
-1.06168315e-01 -3.04305464e-01 -4.25797433e-01 -6.21418059e-01
-6.32420778e-02 4.09577727e-01 -2.33520702e-01 6.49713306e-03
-6.97448850e-03 6.71341896e-01 4.30938482e-01 -1.98232681e-02
-1.71474054e-01 2.54752755e-01 9.94834483e-01 -1.52842194e-01
7.54502416e-03 4.84323464e-02 -2.69782454e-01 8.43257785e-01
5.71436048e-01 -5.47593355e-01 -6.12115204e-01 -2.14258909e-01
1.93043157e-01 -1.79398417e-01 7.40001321e-01 -1.28763533e+00
3.29088748e-01 1.71490893e-01 7.79523104e-02 -1.46520168e-01
6.07480168e-01 -1.12231493e+00 2.28718877e-01 6.91009760e-01
1.52797118e-01 7.67391399e-02 1.06647663e-01 6.12833083e-01
-3.29489052e-01 -3.38099867e-01 8.54402423e-01 -2.47039944e-01
-7.17120230e-01 2.63894629e-02 -6.40904844e-01 -8.59533250e-02
1.23466492e+00 -7.01994777e-01 -5.87003171e-01 -8.24596345e-01
-5.00827193e-01 -2.90706933e-01 6.90767109e-01 2.95781732e-01
5.71057260e-01 -9.43049192e-01 -3.31474930e-01 3.47689986e-01
-2.27355272e-01 -3.88683856e-01 5.79560578e-01 9.85085428e-01
-1.01321876e+00 2.34183162e-01 -5.48608243e-01 -5.06736040e-01
-1.41847849e+00 1.02340841e+00 4.70479876e-01 3.14815372e-01
-8.63313854e-01 6.04126751e-01 6.50277376e-01 4.71197784e-01
1.34712785e-01 -4.47450578e-02 -2.34870940e-01 -9.81152430e-03
7.15341628e-01 5.53746581e-01 -9.05706733e-02 -6.59884036e-01
-2.53790915e-01 7.13985801e-01 -6.53558448e-02 -4.96385135e-02
9.45867300e-01 -2.21968845e-01 8.07949007e-02 3.20715368e-01
1.25987756e+00 3.38878334e-01 -1.38375401e+00 1.55751377e-01
-6.04750872e-01 -6.66755378e-01 -3.39008379e-03 -1.18021235e-01
-1.18088722e+00 1.01941764e+00 6.96965158e-01 5.60304821e-01
1.26865649e+00 -4.37149793e-01 6.39871478e-01 5.46702325e-01
3.61736149e-01 -9.19881344e-01 1.35386214e-01 2.57823497e-01
6.68783545e-01 -9.21270728e-01 9.80907585e-03 -3.40785831e-01
-5.25109768e-01 1.24135911e+00 3.84110838e-01 -6.09817147e-01
4.83117014e-01 2.17520863e-01 2.02471107e-01 5.81732318e-02
-4.46514547e-01 7.92834312e-02 -2.00355917e-01 8.19662869e-01
-2.22612381e-01 -3.45990151e-01 -9.21644941e-02 -1.20326459e-01
-8.61963704e-02 -2.14045271e-01 1.36954260e+00 8.11223686e-01
-1.85352281e-01 -6.48710072e-01 -6.42535448e-01 -1.26240477e-01
-7.97334850e-01 1.15246959e-01 -2.68616259e-01 9.97714698e-01
-5.91240786e-02 8.84499609e-01 9.91659835e-02 -5.87368965e-01
2.30169415e-01 -1.90316603e-01 3.59057575e-01 -2.88102310e-02
-4.47073191e-01 -1.44054219e-02 -1.39637947e-01 -9.54507053e-01
-6.79849982e-01 -4.54808861e-01 -6.99713469e-01 -6.49409473e-01
2.18794439e-02 -1.39193103e-01 4.99571770e-01 8.37839186e-01
6.55513927e-02 3.27046573e-01 9.62204516e-01 -1.06375575e+00
-2.84266442e-01 -7.55609572e-01 -6.91067636e-01 6.38748288e-01
6.42258704e-01 -7.59971678e-01 -9.28764701e-01 1.31647035e-01] | [5.44708776473999, 7.9567131996154785] |
485bc078-487a-4d49-8e7f-630fc07400ef | an-empirical-study-of-the-expressiveness-of | null | null | https://openreview.net/forum?id=CJmMqnXthgX | https://openreview.net/pdf?id=CJmMqnXthgX | An Empirical Study of the Expressiveness of Graph Kernels and Graph Neural Networks | Graph neural networks and graph kernels have achieved great success in solving machine learning problems on graphs. Recently, there has been considerable interest in determining the expressive power mainly of graph neural networks and of graph kernels, to a lesser extent. Most studies have focused on the ability of these approaches to distinguish non-isomorphic graphs or to identify specific graph properties. However, there is often a need for algorithms whose produced graph representations can accurately capture similarity/distance of graphs. This paper studies the expressive power of graph neural networks and graph kernels from an empirical perspective. Specifically, we compare the graph representations and similarities produced by these algorithms against those generated by a well-accepted, but intractable graph similarity function. We also investigate the impact of node attributes on the performance of the different models and kernels. Our results reveal interesting findings. For instance, we find that theoretically more powerful models do not necessarily yield higher-quality representations, while graph kernels are shown to be very competitive with graph neural networks. | ['Michalis Vazirgiannis', 'George Panagopoulos', 'Giannis Nikolentzos'] | 2021-01-01 | null | null | null | null | ['graph-similarity'] | ['graphs'] | [ 1.15874233e-02 2.69333303e-01 -1.34449527e-01 -2.16445237e-01
-1.06137790e-01 -6.81272984e-01 6.64371967e-01 7.83298969e-01
-1.56160071e-01 1.72038600e-01 1.88715786e-01 -5.10122776e-01
-5.25072098e-01 -1.07212007e+00 -3.83582741e-01 -4.14279968e-01
-6.89922392e-01 4.05888259e-01 1.43033341e-01 -8.91793668e-02
1.50362179e-01 7.13120222e-01 -1.32398355e+00 -2.06617713e-01
8.92690420e-01 6.71218634e-01 -2.51989812e-01 8.63379359e-01
-7.19099194e-02 7.45690763e-01 -3.22678357e-01 -6.88439488e-01
1.68553010e-01 -3.48148406e-01 -7.09945440e-01 -4.78423126e-02
6.52874529e-01 1.91446662e-01 -7.93431342e-01 1.27692628e+00
2.75976688e-01 8.48060474e-02 8.76701891e-01 -1.52853441e+00
-1.07849240e+00 7.75034189e-01 -3.84269476e-01 4.13012505e-01
4.46635932e-01 -2.91522741e-02 1.33037448e+00 -1.89732328e-01
4.23026055e-01 1.15834939e+00 9.88654792e-01 4.89226058e-02
-1.48304319e+00 -3.97647262e-01 -3.18939574e-02 9.58948955e-02
-1.33554268e+00 -3.54096033e-02 8.18251073e-01 -3.59344602e-01
8.37121964e-01 3.27763081e-01 5.27836323e-01 5.83419263e-01
2.51966596e-01 3.07006598e-01 9.73209083e-01 -6.36457860e-01
8.68832618e-02 1.41269982e-01 5.34277260e-01 1.10426438e+00
8.29223216e-01 1.90508097e-01 -5.36480062e-02 -4.12324965e-01
8.37921619e-01 -4.71037850e-02 -4.80128139e-01 -8.44109297e-01
-8.81116867e-01 1.09711254e+00 6.87283218e-01 7.60657072e-01
-2.51100570e-01 3.03354383e-01 4.61554199e-01 7.02196479e-01
6.76388800e-01 8.97655785e-01 -5.61775006e-02 8.63249376e-02
-4.97719973e-01 5.05654551e-02 1.09576154e+00 7.61599004e-01
7.87710786e-01 7.24504441e-02 9.04498100e-02 6.58089399e-01
-1.51891753e-01 2.80554798e-02 2.02610627e-01 -4.05351579e-01
1.20737024e-01 8.97608221e-01 -3.67648542e-01 -1.67902362e+00
-5.97473860e-01 -1.93089291e-01 -9.04221892e-01 1.90195478e-02
7.43154526e-01 1.40653476e-01 -6.57741904e-01 1.69787407e+00
-1.12499572e-01 1.38762623e-01 1.24707213e-02 6.20957494e-01
7.28883386e-01 2.93682367e-01 9.97025892e-02 1.80857286e-01
1.06549180e+00 -6.33213699e-01 -2.55307525e-01 -1.65340662e-01
1.04996014e+00 -4.48217690e-01 9.84294474e-01 -7.18383715e-02
-9.40817893e-01 -3.10024410e-01 -8.70469451e-01 1.84762686e-01
-6.22110128e-01 -2.17601974e-02 1.20661891e+00 9.87405956e-01
-1.28839052e+00 1.06491256e+00 -6.31335735e-01 -8.15328419e-01
7.16807842e-02 4.46173489e-01 -5.98438621e-01 -2.28428379e-01
-1.05488777e+00 1.08914459e+00 6.13279223e-01 -2.03110874e-01
-2.34343797e-01 -4.96719718e-01 -1.04322028e+00 3.74742508e-01
3.87834430e-01 -3.60737532e-01 7.21919894e-01 -1.25250101e+00
-8.62216294e-01 7.62206316e-01 4.19838727e-01 -6.85641587e-01
7.95289502e-02 3.86842787e-01 -4.33287948e-01 1.13398276e-01
-1.89067110e-01 1.80948958e-01 4.23861027e-01 -8.17862630e-01
-1.67605340e-01 -3.37378830e-01 3.90061051e-01 -4.87570427e-02
-5.89847505e-01 -3.89522351e-02 -1.50730401e-01 -5.26927173e-01
-1.44671410e-01 -1.06481433e+00 -2.20105961e-01 -2.06210345e-01
-3.67925704e-01 -4.44064021e-01 3.91464859e-01 -4.34643835e-01
1.15639377e+00 -1.99002576e+00 -7.30386972e-02 4.23411429e-01
7.00753689e-01 1.92715898e-01 -5.12495518e-01 6.97496593e-01
-4.36430156e-01 2.91846991e-01 2.27093231e-02 3.99207324e-01
4.74516302e-02 2.26966739e-01 -7.29466453e-02 7.46218741e-01
2.32131541e-01 1.06307960e+00 -9.41379666e-01 -1.88671380e-01
1.35282069e-01 5.79019308e-01 -4.15740073e-01 1.18979774e-02
1.82552010e-01 -3.08717459e-01 -4.24214929e-01 4.56838101e-01
2.64752716e-01 -6.88036144e-01 5.81465542e-01 -7.67765790e-02
3.69793415e-01 5.57740927e-02 -8.35566878e-01 1.04616594e+00
-1.60877898e-01 8.72878492e-01 -2.98606038e-01 -1.29525125e+00
8.62817228e-01 1.29154310e-01 3.06861728e-01 -6.27282441e-01
1.47975236e-01 -4.28802930e-02 4.11726058e-01 -1.55820459e-01
4.00778443e-01 -1.44239888e-01 9.17453170e-02 6.68507040e-01
5.60808033e-02 8.36034045e-02 3.70842814e-01 2.46744931e-01
1.51002443e+00 -3.76160413e-01 4.28533942e-01 -4.33129102e-01
2.97171146e-01 2.81316712e-02 -4.74190041e-02 8.38098466e-01
6.66203862e-03 3.15678835e-01 9.25726891e-01 -6.78334057e-01
-1.01722276e+00 -1.11819005e+00 2.27346331e-01 1.12843859e+00
1.42123252e-01 -5.94687581e-01 -6.35685086e-01 -6.79410040e-01
3.45579803e-01 5.21502316e-01 -9.06936109e-01 -7.30828285e-01
-3.26821417e-01 -7.68940330e-01 7.25703180e-01 6.89123929e-01
-9.38214064e-02 -9.03936625e-01 -2.24672839e-01 -1.07361704e-01
3.97253364e-01 -1.11290312e+00 -3.92372340e-01 1.59243852e-01
-1.00511849e+00 -1.53170228e+00 -5.33845186e-01 -8.19079459e-01
8.89086306e-01 4.68439579e-01 1.37619424e+00 4.13404644e-01
-2.03916460e-01 6.70856178e-01 -4.27564442e-01 -1.17024839e-01
-6.24021232e-01 4.39862221e-01 8.92345533e-02 -9.10789371e-02
4.26532894e-01 -8.39554906e-01 -6.86342046e-02 1.98801085e-01
-9.96816337e-01 -3.05124134e-01 7.51478136e-01 5.69099784e-01
4.06513922e-02 3.38094622e-01 3.59708071e-01 -1.08940053e+00
1.44840121e+00 -4.99277681e-01 -5.32921314e-01 4.34680969e-01
-1.06797814e+00 4.85258222e-01 8.68542492e-01 -6.64548576e-01
-3.32148701e-01 -2.45868415e-01 4.75600481e-01 -4.83010560e-01
-7.42729157e-02 8.00633192e-01 2.35926524e-01 -6.30997539e-01
8.84608209e-01 1.40680492e-01 1.25254691e-01 -2.07708344e-01
6.06025040e-01 1.05518185e-01 3.52343291e-01 -6.37143075e-01
9.55157697e-01 2.10323319e-01 1.30208269e-01 -9.58490908e-01
-4.42179292e-01 -3.66227865e-01 -5.15939355e-01 -2.20768470e-02
6.14215851e-01 -3.96537900e-01 -4.40094322e-01 1.04996122e-01
-7.68321395e-01 -2.48989359e-01 -1.86486155e-01 5.01738787e-01
-5.07606149e-01 7.51789212e-01 -7.39896417e-01 -5.69431663e-01
-2.37775519e-01 -6.99442625e-01 5.32434702e-01 1.64233357e-01
-2.33357534e-01 -1.45170951e+00 2.67400265e-01 -9.39996764e-02
5.26965857e-01 4.17045385e-01 1.36729455e+00 -1.18516290e+00
-3.58023375e-01 -6.14315510e-01 -6.92213476e-01 1.41972214e-01
2.98058540e-01 -7.61826104e-03 -6.58368409e-01 -4.10006762e-01
-6.14132583e-01 -2.35423103e-01 8.57790112e-01 4.23461050e-01
7.70359457e-01 -2.30276182e-01 -3.82565767e-01 5.81164002e-01
1.46850324e+00 -3.14586796e-02 5.07122517e-01 5.34167178e-02
7.81147480e-01 6.89193666e-01 -1.53496787e-01 -1.06564328e-01
2.31993020e-01 3.65985930e-01 2.05181763e-01 -7.94435441e-02
1.66666299e-01 -2.67770410e-01 3.01598087e-02 8.77915442e-01
-3.92771900e-01 -3.87429923e-01 -1.12812090e+00 3.97145122e-01
-1.74400222e+00 -7.90040135e-01 -1.15499504e-01 2.31106329e+00
1.01564460e-01 2.38570496e-01 3.53666335e-01 5.16964830e-02
9.00288403e-01 3.15245867e-01 -1.85482413e-01 -5.01220047e-01
-7.72867575e-02 2.26632223e-01 6.48776531e-01 1.89349562e-01
-9.61675584e-01 8.25522840e-01 7.17106819e+00 5.72415113e-01
-1.03856516e+00 -3.06322426e-01 5.31604528e-01 2.96616435e-01
-2.95357049e-01 1.72485158e-01 2.81606894e-02 9.36197117e-02
1.13420391e+00 -6.21306896e-01 5.23894608e-01 9.68781590e-01
-4.30337042e-01 1.35720164e-01 -1.25012743e+00 9.25127387e-01
2.28844807e-01 -1.18130744e+00 2.28210096e-03 3.37262660e-01
4.15795833e-01 -1.83742177e-02 -3.92080024e-02 3.67020041e-01
5.93066394e-01 -1.42460954e+00 1.17910452e-01 3.62243503e-01
4.93655503e-01 -8.64766121e-01 7.46324658e-01 5.91211393e-03
-1.26056755e+00 4.26792428e-02 -6.79103732e-01 -3.22925240e-01
-3.61090153e-01 4.19938475e-01 -9.69368279e-01 5.61948895e-01
3.28530133e-01 5.41266739e-01 -9.43624854e-01 1.01006961e+00
2.13837922e-02 5.17960310e-01 -2.17249051e-01 -4.11766320e-01
3.98235798e-01 -4.38888222e-01 1.78370252e-01 1.12910497e+00
1.38493022e-02 -1.83437839e-01 2.38901496e-01 8.89201760e-01
-6.92073852e-02 2.93332666e-01 -1.17280114e+00 -9.56005335e-01
1.85796976e-01 1.32138729e+00 -1.13034117e+00 -1.41428947e-01
-5.90946674e-01 8.56419563e-01 8.21678221e-01 2.98526466e-01
-6.40805423e-01 -5.04159987e-01 6.12337232e-01 2.80944347e-01
7.47188702e-02 -4.03031886e-01 1.85111180e-01 -1.09598231e+00
-1.85689256e-01 -6.85785472e-01 5.92167795e-01 -6.78017199e-01
-1.60113311e+00 6.61805153e-01 -3.41948308e-02 -5.87327540e-01
-2.96122968e-01 -7.90218949e-01 -8.04143667e-01 7.13968515e-01
-1.07523549e+00 -9.71516252e-01 -3.24653506e-01 3.63475263e-01
-2.61462271e-01 2.69825719e-02 9.29122210e-01 1.63220335e-02
-2.65683025e-01 4.60122436e-01 1.77988216e-01 5.39448380e-01
2.58358985e-01 -1.41491592e+00 7.39646375e-01 4.94816810e-01
7.03898609e-01 5.12552142e-01 5.85394800e-01 -6.40966594e-01
-1.53756320e+00 -7.88059175e-01 7.05302835e-01 -4.84833479e-01
9.29932594e-01 -3.05909991e-01 -1.06468523e+00 8.00426722e-01
-8.45929161e-02 1.98106319e-01 5.99140406e-01 5.42558491e-01
-7.51563668e-01 1.65237859e-01 -9.03369427e-01 6.16743505e-01
1.05182886e+00 -8.04247677e-01 -3.45191807e-01 2.69395381e-01
4.79781568e-01 7.47769997e-02 -9.86323059e-01 2.55647510e-01
5.14362156e-01 -1.13853264e+00 1.00094390e+00 -1.04392171e+00
1.83623031e-01 2.13163540e-01 -3.35035473e-02 -1.41767466e+00
-7.66031921e-01 -2.40328789e-01 1.19091347e-01 7.78901815e-01
4.48537529e-01 -8.95033717e-01 1.04078662e+00 6.65056765e-01
3.17985207e-01 -7.39332557e-01 -6.09938681e-01 -1.05801356e+00
1.03123143e-01 -1.99223205e-01 5.15673876e-01 1.46358919e+00
1.93468302e-01 5.14882684e-01 -7.94809908e-02 1.00704819e-01
4.07654345e-01 2.46680781e-01 7.16199338e-01 -1.62391019e+00
-1.98817059e-01 -9.09939587e-01 -1.18224311e+00 -3.33342671e-01
3.66108358e-01 -1.26335323e+00 -5.91260374e-01 -1.55690730e+00
1.62896276e-01 -3.60456616e-01 -4.30682600e-01 4.31105822e-01
-1.21512748e-01 3.95392403e-02 8.52119848e-02 -4.60762419e-02
-4.39849734e-01 2.32386485e-01 9.00066435e-01 -8.11750144e-02
-2.30181977e-01 -1.32519573e-01 -8.53632152e-01 6.72298074e-01
9.22330797e-01 -2.38058239e-01 -6.51922345e-01 -1.65430173e-01
4.45588887e-01 -4.45446931e-02 3.51494461e-01 -1.06833386e+00
1.91477805e-01 7.95285627e-02 1.70396850e-01 2.03649506e-01
-3.53624560e-02 -6.22638404e-01 1.65915489e-01 4.10513222e-01
-3.85277838e-01 5.21126330e-01 6.01984560e-02 8.03484142e-01
-1.04320897e-02 -1.73881263e-01 5.18890142e-01 -1.89463690e-01
-5.67587376e-01 4.34850931e-01 -7.27706775e-02 2.06544697e-01
8.53167593e-01 -5.20789027e-01 -2.52132446e-01 -6.69033587e-01
-4.66082305e-01 -1.31087884e-01 6.99329853e-01 5.38555801e-01
4.12222058e-01 -1.24045908e+00 -5.36170065e-01 3.61788571e-02
2.70647019e-01 -6.51713431e-01 -1.57886729e-01 8.13950062e-01
-6.28345191e-01 4.08653051e-01 -2.18275532e-01 -2.11010039e-01
-1.26481640e+00 9.32122648e-01 3.80590349e-01 -3.58809292e-01
-7.20426679e-01 5.29884994e-01 3.59979391e-01 -6.04533017e-01
1.12129357e-02 -5.04252017e-01 5.53785563e-02 -6.67310357e-02
1.65712953e-01 4.75952774e-01 -1.70192733e-01 -5.31178176e-01
-3.24699938e-01 4.05071318e-01 -6.35400936e-02 4.54692811e-01
1.28436124e+00 4.85532075e-01 -1.00117922e-01 3.18662256e-01
1.16349375e+00 -2.58167069e-02 -7.35994399e-01 -1.56179652e-01
4.00025576e-01 -3.00418884e-01 -9.28175896e-02 -3.76761407e-01
-1.07530308e+00 6.55217230e-01 3.21887851e-01 1.10407591e+00
9.08045828e-01 2.45063394e-01 3.91557902e-01 3.70365500e-01
2.64739275e-01 -6.58779979e-01 -4.70350571e-02 3.11458766e-01
4.98232871e-01 -1.09277809e+00 1.83009237e-01 -5.82067490e-01
-3.39419395e-01 1.30804837e+00 4.44209903e-01 -5.29199898e-01
5.85098684e-01 -7.25520700e-02 -2.55081058e-01 -5.33043981e-01
-5.48624218e-01 -3.34936082e-01 6.19391561e-01 9.34164345e-01
5.59202611e-01 2.58335590e-01 -7.45899230e-02 1.71477094e-01
-2.71139652e-01 -2.82541454e-01 4.77592349e-01 6.06506407e-01
-3.73057723e-01 -9.48022008e-01 -5.47502972e-02 8.57300460e-01
-2.45358974e-01 -2.18590349e-01 -9.73489821e-01 1.21833992e+00
-5.22908211e-01 5.82332194e-01 2.36164182e-02 -6.61206841e-01
1.67833462e-01 1.00026960e-02 7.15097308e-01 -5.32437146e-01
-4.66474026e-01 -5.73498011e-01 3.47348481e-01 -5.05730450e-01
-2.04471692e-01 -1.35401219e-01 -8.80432546e-01 -6.03415430e-01
-6.32955074e-01 4.22848433e-01 3.61145556e-01 6.23546004e-01
4.33415979e-01 2.59152621e-01 2.26071641e-01 -5.18022954e-01
-5.08153617e-01 -8.30158651e-01 -1.03738022e+00 5.24206519e-01
-1.06203012e-01 -5.42506516e-01 -5.77359915e-01 -5.44703662e-01] | [6.9605231285095215, 6.175890922546387] |
563211c2-7662-487a-a2fa-ac32deaad660 | grit-general-robust-image-task-benchmark | 2204.13653 | null | https://arxiv.org/abs/2204.13653v2 | https://arxiv.org/pdf/2204.13653v2.pdf | GRIT: General Robust Image Task Benchmark | Computer vision models excel at making predictions when the test distribution closely resembles the training distribution. Such models have yet to match the ability of biological vision to learn from multiple sources and generalize to new data sources and tasks. To facilitate the development and evaluation of more general vision systems, we introduce the General Robust Image Task (GRIT) benchmark. GRIT evaluates the performance, robustness, and calibration of a vision system across a variety of image prediction tasks, concepts, and data sources. The seven tasks in GRIT are selected to cover a range of visual skills: object categorization, object localization, referring expression grounding, visual question answering, segmentation, human keypoint detection, and surface normal estimation. GRIT is carefully designed to enable the evaluation of robustness under image perturbations, image source distribution shift, and concept distribution shift. By providing a unified platform for thorough assessment of skills and concepts learned by a vision model, we hope GRIT catalyzes the development of performant and robust general purpose vision systems. | ['Derek Hoiem', 'Aniruddha Kembhavi', 'Ryan Marten', 'Tanmay Gupta'] | 2022-04-28 | null | null | null | null | ['surface-normals-estimation', 'object-categorization'] | ['computer-vision', 'computer-vision'] | [ 5.18688023e-01 -2.80154526e-01 2.01590806e-01 -4.75218683e-01
-6.87110722e-01 -6.17997706e-01 6.27739906e-01 2.27668285e-01
-4.19954568e-01 1.47288099e-01 -3.42432469e-01 -1.40689597e-01
-1.15721874e-01 -2.17830703e-01 -7.36288428e-01 -4.50028121e-01
2.11991578e-01 3.37242216e-01 5.63967884e-01 -1.53873593e-01
5.64680040e-01 6.43074512e-01 -1.82908738e+00 3.87105435e-01
7.09568501e-01 1.03395295e+00 4.83706385e-01 8.77113342e-01
2.95724481e-01 7.50253737e-01 -7.13195562e-01 -3.89727414e-01
2.44821593e-01 -2.38420755e-01 -7.17173755e-01 2.86255389e-01
1.11453199e+00 6.40182644e-02 -2.20114943e-02 1.12710297e+00
5.10107398e-01 2.49655545e-01 9.14140940e-01 -1.38400209e+00
-1.00198853e+00 -4.66015795e-03 -5.86173236e-01 6.74279332e-01
3.02701205e-01 6.56182885e-01 8.19370031e-01 -9.45382893e-01
5.94451785e-01 1.30759966e+00 8.09274375e-01 6.82599008e-01
-1.25066268e+00 -3.14469516e-01 1.68703228e-01 1.89937606e-01
-9.87880409e-01 -4.70883638e-01 4.29108262e-01 -7.01192498e-01
9.44259822e-01 1.14253089e-01 4.51147258e-01 1.03870344e+00
2.53563464e-01 8.44859481e-01 1.26400840e+00 -2.93197453e-01
1.58154294e-01 1.00026138e-01 2.78240800e-01 7.76672542e-01
2.60681093e-01 2.85008669e-01 -5.16065121e-01 1.52639925e-01
5.79688966e-01 -4.71727610e-01 -4.37755942e-01 -3.06644648e-01
-1.19179714e+00 4.92268324e-01 6.15191400e-01 4.29623500e-02
-7.64957592e-02 1.48632780e-01 2.48285979e-01 7.54197761e-02
5.09706773e-02 8.80610645e-01 -4.42187041e-01 1.35166526e-01
-7.02767968e-01 2.08087221e-01 6.44505620e-01 1.02588129e+00
5.51112950e-01 2.92382479e-01 -3.50396812e-01 8.92883480e-01
1.44733787e-01 6.27445936e-01 7.02955008e-01 -1.39774144e+00
1.80365920e-01 4.57825810e-01 -1.31759554e-01 -1.12569177e+00
-4.64904398e-01 -3.94249797e-01 -4.87820357e-01 4.58333522e-01
7.55010664e-01 1.04160331e-01 -1.16761518e+00 1.58608198e+00
-4.59716506e-02 -9.03073102e-02 9.03060362e-02 7.00431645e-01
1.00899923e+00 3.98997068e-01 8.55966732e-02 4.41815406e-02
1.11961246e+00 -7.60462642e-01 1.02699259e-02 -6.65639102e-01
1.69794157e-01 -9.20662999e-01 1.19913983e+00 4.51540977e-01
-9.89002585e-01 -9.39625978e-01 -9.07814622e-01 -2.32013136e-01
-3.65036577e-01 3.42586171e-03 3.45008880e-01 5.67591488e-01
-1.01584148e+00 3.71953517e-01 -5.05554736e-01 -5.56779563e-01
6.62934482e-01 2.60530561e-02 -3.57020408e-01 -2.26999238e-01
-5.30995011e-01 1.36177230e+00 3.94565135e-01 -1.59012467e-01
-1.05649805e+00 -8.39918435e-01 -8.05943549e-01 -1.88702777e-01
-3.59549522e-02 -7.54242480e-01 1.43805170e+00 -1.13261116e+00
-9.69972014e-01 1.23436522e+00 1.88483149e-02 -5.35867751e-01
3.26983362e-01 1.08413704e-01 -7.89488479e-02 2.62473494e-01
1.95621029e-01 1.12683153e+00 1.04750013e+00 -1.41003382e+00
-6.86026812e-01 -3.83814245e-01 -4.22154129e-01 1.24133341e-01
6.26580417e-02 -1.02223054e-01 -4.26425964e-01 -5.61985850e-01
-2.86686905e-02 -9.01475370e-01 2.54026558e-02 2.16777951e-01
-2.76379824e-01 -1.72610059e-01 7.54702926e-01 -4.30481195e-01
3.07970434e-01 -2.22616863e+00 -1.78870350e-01 1.50760531e-03
6.88702092e-02 3.69509727e-01 -3.52885336e-01 -1.19923770e-01
-1.76447913e-01 -5.81179149e-02 -1.29535958e-01 -7.26183802e-02
-1.47036910e-01 8.06846023e-02 -3.55293930e-01 3.29695970e-01
3.84131312e-01 9.77964938e-01 -7.68435717e-01 -3.60155702e-01
2.77249604e-01 2.39183620e-01 -3.22049826e-01 1.03293628e-01
-3.46513838e-01 2.85275340e-01 -1.24667101e-01 7.92360246e-01
2.77638048e-01 -3.12682867e-01 -3.43634099e-01 -3.47685933e-01
1.75095752e-01 -4.11853433e-01 -6.21280968e-01 1.26630652e+00
-9.01146010e-02 1.16728425e+00 -1.57803923e-01 -8.90982747e-01
1.02591121e+00 -1.58181071e-01 2.64477015e-01 -8.22157741e-01
1.93741888e-01 3.66814667e-03 1.92618757e-01 -6.94756329e-01
2.65287727e-01 1.76546156e-01 3.27753365e-01 2.52557337e-01
4.02792782e-01 -7.31619120e-01 2.63052285e-01 8.77902731e-02
8.68425667e-01 6.29088581e-02 1.67651832e-01 -3.43243510e-01
2.01221213e-01 1.95657477e-01 2.47699052e-01 1.03008890e+00
-6.43109441e-01 7.43722618e-01 1.79999411e-01 -3.79685044e-01
-1.02142131e+00 -1.26257873e+00 -2.99129725e-01 1.38475418e+00
2.63285130e-01 5.77952191e-02 -7.80273378e-01 -3.31596762e-01
9.65904593e-02 6.94316864e-01 -6.66305482e-01 -3.86606455e-01
-5.41770943e-02 -7.95168161e-01 6.07425869e-01 6.94898069e-01
6.37865007e-01 -1.15236139e+00 -9.07200694e-01 -2.06216767e-01
-6.20571384e-03 -1.34444463e+00 -3.40193242e-01 1.14579238e-01
-7.14517832e-01 -1.36739075e+00 -5.84575713e-01 -9.97025073e-01
6.44567847e-01 3.91546398e-01 1.15926468e+00 -6.14525657e-03
-8.77597630e-01 1.10516238e+00 5.07856347e-03 -8.63889217e-01
-4.57331568e-01 -3.69506240e-01 5.66910058e-02 -2.69953758e-01
3.30369025e-01 -1.18612438e-01 -4.99391854e-01 6.02890790e-01
-7.77036846e-01 -9.38810334e-02 5.71248829e-01 7.65078664e-01
7.12646186e-01 -8.76435265e-02 3.54320139e-01 -3.84969741e-01
7.87464499e-01 -1.68477863e-01 -6.38606191e-01 5.02712429e-01
-7.88132846e-01 -4.23393808e-02 2.06278607e-01 -4.32975531e-01
-9.31668222e-01 1.66627541e-01 1.15479872e-01 -3.92823189e-01
-4.23572809e-01 1.08528674e-01 1.17907889e-01 -5.06093204e-01
1.31279147e+00 4.82890040e-01 1.98117942e-01 -1.08406961e-01
5.16496599e-01 3.05820405e-01 1.04328728e+00 -6.20734036e-01
7.21996307e-01 2.43240431e-01 -1.08073195e-02 -9.91930842e-01
-8.51228237e-01 -5.07813215e-01 -5.09534061e-01 -3.11155558e-01
8.82193446e-01 -6.47168398e-01 -6.85487986e-01 7.84029603e-01
-1.10146451e+00 -4.87140894e-01 -3.12381655e-01 2.74671167e-01
-8.35692465e-01 3.78588855e-01 -3.26837718e-01 -4.67485726e-01
-2.64046878e-01 -1.29943049e+00 9.08991694e-01 5.65939963e-01
-2.21865579e-01 -8.92799377e-01 -7.32978284e-02 5.79606712e-01
2.46463925e-01 1.99053407e-01 9.13797021e-01 -6.47273540e-01
-5.36913693e-01 -1.43092245e-01 -5.22701442e-01 6.69027567e-01
-1.17229015e-01 4.45008636e-01 -1.30880976e+00 -2.41171345e-01
-1.30732924e-01 -7.29895651e-01 1.15371549e+00 5.86113870e-01
1.34024847e+00 1.32732853e-01 -1.94635957e-01 6.92928910e-01
1.21035433e+00 2.40275234e-01 4.52882677e-01 3.97821277e-01
4.62948978e-01 8.32883596e-01 3.85468602e-01 -1.88468665e-01
2.18858540e-01 3.09311420e-01 4.99058157e-01 -1.35903344e-01
-3.08684260e-01 4.21891324e-02 7.10272714e-02 -3.03885397e-02
-1.07499234e-01 -1.32120028e-01 -1.26088977e+00 4.69760507e-01
-1.52304602e+00 -1.03179240e+00 5.03585190e-02 1.87745881e+00
6.43147290e-01 3.17677438e-01 1.31861195e-01 -1.02260649e-01
6.59155786e-01 -3.12321275e-01 -9.67351735e-01 -2.53759533e-01
-2.36818731e-01 -1.19486183e-01 4.60857689e-01 2.83168644e-01
-1.00580704e+00 8.78730774e-01 7.97809219e+00 6.29075468e-01
-9.00414646e-01 -3.24756324e-01 8.54730368e-01 2.99978256e-01
1.52980193e-01 -4.41099077e-01 -6.93206847e-01 2.11519495e-01
5.98597765e-01 -1.15461133e-01 2.00817466e-01 1.00942945e+00
-1.64804876e-01 -2.33766690e-01 -1.32909024e+00 1.16800487e+00
4.96396422e-01 -1.31318271e+00 9.56834704e-02 -3.41014862e-01
7.38673270e-01 4.09341961e-01 5.13689935e-01 1.81806713e-01
4.38146204e-01 -1.24734819e+00 6.75678432e-01 5.62424481e-01
5.13911426e-01 -2.39594102e-01 3.92818063e-01 1.87104270e-01
-6.43818378e-01 -3.46343964e-01 -4.08986747e-01 1.80844039e-01
-4.13855374e-01 1.31942183e-01 -9.94169295e-01 -4.21857461e-02
1.03333974e+00 6.44354463e-01 -1.12805748e+00 1.38759184e+00
-2.62101796e-02 4.26721007e-01 -2.03311294e-01 -1.68906171e-02
-3.58014964e-02 2.25721881e-01 5.30185163e-01 1.29196334e+00
-7.39609450e-02 -5.20281233e-02 1.99740976e-01 8.88096750e-01
2.97273509e-02 -1.70708776e-01 -5.80180347e-01 1.69751927e-01
4.11118627e-01 1.11739492e+00 -8.89004230e-01 -4.94097173e-02
-2.80164003e-01 7.00781941e-01 1.12908386e-01 5.90333223e-01
-4.34602320e-01 -2.04288781e-01 6.47382021e-01 -6.26022220e-02
2.17572749e-01 -2.50790298e-01 -5.01648009e-01 -8.80115449e-01
-2.09762156e-01 -1.03040648e+00 3.44244838e-01 -1.35040188e+00
-1.66363287e+00 5.81352413e-01 2.83671796e-01 -8.35518420e-01
-2.44598269e-01 -1.22190702e+00 -5.98544478e-01 5.96162021e-01
-1.28431368e+00 -1.06670773e+00 -6.81907356e-01 8.19095314e-01
7.29090750e-01 -6.07373953e-01 5.83518505e-01 -3.89720798e-01
-4.21383083e-01 5.01415968e-01 -1.39717222e-03 3.03364992e-01
7.09246397e-01 -1.27717447e+00 4.12988096e-01 8.87825191e-01
3.11087698e-01 5.24363577e-01 8.70673656e-01 -2.73009598e-01
-1.08273184e+00 -9.88091409e-01 -4.45044180e-03 -9.65883613e-01
5.55072784e-01 7.34487101e-02 -8.08818519e-01 6.04305863e-01
4.37249392e-02 1.69347618e-02 5.40769517e-01 -3.39670926e-02
-8.30704510e-01 -2.04817355e-01 -1.01115489e+00 5.12185037e-01
7.53572226e-01 -5.35317242e-01 -7.66502321e-01 4.51501846e-01
2.70235866e-01 -3.57129961e-01 -5.54381967e-01 4.01309729e-01
4.66286302e-01 -9.99343395e-01 1.18831301e+00 -6.02962732e-01
2.76202172e-01 -3.08809191e-01 -3.63809347e-01 -1.26424313e+00
-4.19904262e-01 -3.34266961e-01 1.40522152e-01 9.06639993e-01
4.40219432e-01 -3.69944423e-01 4.52332765e-01 7.15620756e-01
-1.54825076e-01 -5.64521074e-01 -7.45064795e-01 -7.59091735e-01
6.14361018e-02 -5.27417183e-01 1.00946669e-02 5.43467760e-01
-4.02717829e-01 5.36442697e-01 1.94088593e-01 -7.15507567e-02
9.23446834e-01 1.15467578e-01 7.23463833e-01 -1.34056330e+00
-2.64709920e-01 -8.51834714e-01 -8.85142982e-01 -7.84659803e-01
3.72969359e-02 -7.66840816e-01 3.01095009e-01 -1.47861707e+00
1.26181215e-01 -1.14871778e-01 -1.74487710e-01 4.32139486e-01
-2.66862363e-01 5.32933474e-01 4.13387746e-01 3.61120224e-01
-5.52398443e-01 6.39755130e-02 1.31159270e+00 -5.31611562e-01
7.43171051e-02 1.19759299e-01 -9.94168520e-01 1.02392316e+00
8.55469942e-01 -7.10441321e-02 -6.78599477e-01 -6.20820701e-01
-4.91036214e-02 -4.53399956e-01 7.90134192e-01 -1.24454129e+00
3.12871367e-01 -2.90050536e-01 9.74842966e-01 -2.39505649e-01
3.80098253e-01 -5.53166449e-01 -4.56991732e-01 2.53481150e-01
-4.83838767e-01 9.29456428e-02 5.16862154e-01 6.62451923e-01
-1.39068052e-01 -2.55186856e-01 1.25434816e+00 -2.81604469e-01
-1.31572223e+00 2.99675792e-01 -4.06203628e-01 4.08078313e-01
1.05790567e+00 -6.64822519e-01 -7.16446459e-01 -2.69334853e-01
-7.76245415e-01 2.19446748e-01 4.37595606e-01 5.39987206e-01
9.26723957e-01 -7.94791639e-01 -6.89007580e-01 1.93394005e-01
4.87352967e-01 -8.45622271e-02 -8.13579094e-03 4.92120653e-01
-5.09186566e-01 1.28982261e-01 -6.32906973e-01 -1.00014329e+00
-1.28259182e+00 6.28973484e-01 8.05890024e-01 4.96607065e-01
-4.27627146e-01 1.14817810e+00 3.73107553e-01 -4.69409935e-02
3.16771597e-01 -4.28528666e-01 -1.51033357e-01 -2.10356027e-01
5.80812037e-01 2.03809887e-01 2.41631991e-03 -5.56149125e-01
-1.91235319e-01 7.76148200e-01 -1.36009902e-01 1.19345829e-01
9.00414705e-01 -5.50196357e-02 9.28617641e-02 4.59690869e-01
7.72943676e-01 -4.97338474e-01 -1.52908266e+00 -7.36926943e-02
1.88841671e-02 -3.19489062e-01 -5.68148792e-02 -1.04436374e+00
-7.06126869e-01 9.75714564e-01 8.86040032e-01 1.02551326e-01
1.16965020e+00 1.14767224e-01 2.53410786e-01 7.13824213e-01
1.23349339e-01 -1.13994384e+00 6.76562488e-01 6.04130328e-01
9.94704008e-01 -1.42478108e+00 1.64546333e-02 -1.54506534e-01
-8.19039404e-01 1.07202518e+00 1.07979369e+00 4.72312942e-02
5.75555265e-01 1.10251181e-01 3.68407428e-01 -1.93007514e-01
-5.61205804e-01 -3.53539765e-01 5.48731506e-01 1.39701200e+00
6.07417673e-02 -3.33343893e-01 5.18824756e-01 -9.60983261e-02
-2.36327842e-01 -2.33685493e-01 4.28263545e-01 5.94453037e-01
-8.33711267e-01 -5.31454623e-01 -4.50087041e-01 4.61432695e-01
-2.37611249e-01 1.92888977e-03 -5.70035160e-01 7.10257590e-01
1.89328417e-01 8.66270363e-01 -1.67639257e-04 -2.34151721e-01
3.15793127e-01 4.22673263e-02 8.12651277e-01 -6.88310444e-01
-3.12684804e-01 -2.69171715e-01 -3.11033905e-01 -3.59436691e-01
-4.63498920e-01 -5.68231106e-01 -9.25084531e-01 3.56867202e-02
-1.48024052e-01 -2.52537489e-01 6.66371167e-01 9.03120875e-01
2.08793595e-01 4.27294582e-01 9.13310647e-02 -8.08240354e-01
-5.91855526e-01 -8.91126573e-01 -3.61750335e-01 6.41170025e-01
3.95498633e-01 -6.78105414e-01 -1.27026141e-01 4.92986947e-01] | [9.891590118408203, 1.9254257678985596] |
beaaec90-36ed-476e-ac16-94bb378d2b14 | separation-of-line-drawings-based-on-split | null | null | http://openaccess.thecvf.com/content_cvpr_2014/html/Zou_Separation_of_Line_2014_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2014/papers/Zou_Separation_of_Line_2014_CVPR_paper.pdf | Separation of Line Drawings Based on Split Faces for 3D Object Reconstruction | Reconstructing 3D objects from single line drawings is often desirable in computer vision and graphics applications. If the line drawing of a complex 3D object is decomposed into primitives of simple shape, the object can be easily reconstructed. We propose an effective method to conduct the line drawing separation and turn a complex line drawing into parametric 3D models. This is achieved by recursively separating the line drawing using two types of split faces. Our experiments show that the proposed separation method can generate more basic and simple line drawings, and its combination with the example-based reconstruction can robustly recover wider range of complex parametric 3D objects than previous methods | ['Jianzhuang Liu', 'Heng Yang', 'Changqing Zou'] | 2014-06-01 | null | null | null | cvpr-2014-6 | ['3d-object-reconstruction'] | ['computer-vision'] | [ 1.99930951e-01 -2.17465754e-03 5.61198033e-02 -9.48248506e-02
-1.53532550e-01 -6.80226803e-01 6.23162031e-01 -5.07646203e-01
3.37244272e-01 5.15776217e-01 -1.74004808e-01 -4.25425380e-01
-5.50938211e-02 -8.76726806e-01 -4.29226339e-01 -1.91346228e-01
2.75600612e-01 9.79916811e-01 6.12139285e-01 1.01474263e-01
6.08169198e-01 1.41611063e+00 -1.30292308e+00 1.11585945e-01
9.17657375e-01 6.55328035e-01 5.06474525e-02 5.13809562e-01
-9.12984788e-01 -1.04404762e-02 -5.93367159e-01 -3.97244334e-01
5.14892578e-01 -3.70990068e-01 -4.06298608e-01 8.69584858e-01
1.95734471e-01 -7.01317012e-01 1.46906808e-01 8.08041871e-01
1.68864787e-01 -2.08271757e-01 1.36034477e+00 -1.07679164e+00
-3.83328080e-01 8.16933662e-02 -1.29382014e+00 -5.75237036e-01
6.28416896e-01 -4.00874794e-01 3.32275510e-01 -1.01024199e+00
7.10195601e-01 1.70455885e+00 8.19595695e-01 1.17216446e-01
-1.41601217e+00 -7.18537331e-01 -6.50569722e-02 -2.98655659e-01
-1.28023171e+00 -3.11911795e-02 1.50791621e+00 -4.88470703e-01
5.26064813e-01 4.47525382e-01 9.51594949e-01 4.00527179e-01
2.37392224e-02 9.38556850e-01 1.07412052e+00 -6.52783871e-01
-1.48693413e-01 5.42939119e-02 4.06068772e-01 9.16266203e-01
5.87222993e-01 -1.28105626e-01 2.16228038e-01 -4.19755965e-01
1.84684122e+00 2.00762928e-01 -3.79321188e-01 -8.29017162e-01
-1.33216298e+00 5.32715797e-01 1.83687299e-01 7.67637491e-02
-1.99846312e-01 -7.79604837e-02 2.12879539e-01 6.00895509e-02
3.20837706e-01 2.16643259e-01 9.45512801e-02 1.87615141e-01
-1.04799306e+00 1.72900975e-01 1.00043905e+00 1.48144293e+00
7.83012748e-01 1.62088200e-01 2.71639228e-01 9.96225417e-01
3.75776619e-01 5.17140985e-01 6.56719832e-03 -9.23222721e-01
3.20461512e-01 6.36214912e-01 2.83094585e-01 -1.14802074e+00
-3.51440400e-01 -4.46522087e-02 -8.62782538e-01 6.97333217e-01
-8.45755637e-03 1.89480573e-01 -7.94446290e-01 5.30663073e-01
4.63210166e-01 9.00455564e-02 -2.84939408e-01 5.11520207e-01
1.11200356e+00 8.91887963e-01 -6.62730277e-01 -2.60212660e-01
1.27152944e+00 -9.57990587e-01 -6.58898473e-01 3.84239793e-01
1.95934311e-01 -1.04071152e+00 8.59452128e-01 7.64749408e-01
-1.26138210e+00 -6.82042539e-01 -1.36172020e+00 -3.16377431e-01
1.99819565e-01 2.00893223e-01 5.64108610e-01 5.09879827e-01
-4.40340966e-01 3.27972353e-01 -4.20406342e-01 -3.70737864e-03
6.22609377e-01 2.53540128e-01 -5.90358317e-01 -1.80140749e-01
-8.75594839e-02 7.94727623e-01 3.95006865e-01 1.69852778e-01
4.32954729e-02 -6.23646140e-01 -5.82474470e-01 -1.44495666e-01
2.63634771e-01 -1.15305591e+00 1.03436136e+00 -2.53606111e-01
-1.78540242e+00 9.72955585e-01 -1.68850482e-01 1.35197759e-01
9.61687922e-01 -1.36069506e-01 -1.03417262e-01 2.44892120e-01
-1.81475177e-01 3.46994072e-01 1.21579969e+00 -1.74780083e+00
-2.46813267e-01 -3.39383572e-01 -1.58954486e-01 1.06652528e-01
5.56283534e-01 -3.20241034e-01 -5.85119486e-01 -5.55020273e-01
8.77435327e-01 -5.09506404e-01 -1.14963152e-01 5.82824051e-01
-7.98200309e-01 -2.33079806e-01 1.34488678e+00 -4.40032303e-01
8.08933139e-01 -1.95006669e+00 6.84777200e-02 5.74635565e-01
2.35522628e-01 2.19890416e-01 3.01407725e-02 5.11864066e-01
-3.72424871e-02 2.59425730e-01 -3.56918663e-01 -2.79683381e-01
-1.26454175e-01 8.19482580e-02 -5.22773147e-01 3.44931513e-01
1.12778775e-01 5.73603451e-01 -4.24170047e-01 -6.05995059e-01
3.62811565e-01 4.30733889e-01 -5.06960928e-01 1.11225292e-01
-1.68627381e-01 2.33910993e-01 -8.53017151e-01 6.40839696e-01
1.29517543e+00 3.74359012e-01 -1.95420697e-01 -4.92654145e-01
-1.60158679e-01 -4.45680588e-01 -1.52758265e+00 1.46250677e+00
-2.98650891e-01 3.79700869e-01 -5.93377277e-03 -7.98253477e-01
1.87109220e+00 1.70544773e-01 1.70829847e-01 -2.90438589e-02
-4.53772256e-03 3.23327810e-01 -2.59811491e-01 -3.72547209e-01
1.87246546e-01 -4.78147119e-01 9.49045718e-02 6.26286566e-01
-5.48728049e-01 -1.43660486e+00 -2.47915387e-01 -8.75042379e-02
5.00716507e-01 6.55726790e-01 5.68861604e-01 -2.05563396e-01
6.76635623e-01 4.23921794e-02 3.65164876e-01 2.33870775e-01
5.57862043e-01 7.75809050e-01 6.30409002e-01 -5.57677269e-01
-1.39463317e+00 -1.31038880e+00 -3.35230172e-01 -2.69905418e-01
3.69890004e-01 -4.65717465e-02 -4.76287544e-01 -4.39688146e-01
3.46075028e-01 7.68102348e-01 -2.49673650e-01 1.83502495e-01
-7.91756749e-01 -1.49419889e-01 2.38856971e-01 4.02844638e-01
8.73414814e-01 -6.36470616e-01 -4.88461584e-01 1.12994738e-01
4.93379951e-01 -7.86823928e-01 -5.03575683e-01 -3.50587249e-01
-1.37228811e+00 -1.13452148e+00 -9.91166949e-01 -1.11959577e+00
1.17885435e+00 8.58738005e-01 9.46952403e-01 4.76969212e-01
-1.58518925e-01 4.65692490e-01 -7.66453519e-02 -3.30363035e-01
-4.26881641e-01 -5.32329500e-01 -2.61244029e-01 -8.44769254e-02
-1.58673033e-01 -8.13434064e-01 -2.16885030e-01 4.24154103e-01
-9.71476853e-01 6.96597397e-01 4.75137234e-01 6.64835393e-01
9.49222624e-01 2.32745975e-01 3.09170246e-01 -1.10467482e+00
8.24193418e-01 4.20750603e-02 -7.92306066e-01 7.31971860e-02
8.25205445e-02 4.84527238e-02 6.28317058e-01 -4.70657229e-01
-1.25788450e+00 1.20321110e-01 -2.73402650e-02 -6.14602804e-01
-3.20434213e-01 6.08820170e-02 -3.23134333e-01 -1.75492063e-01
3.13883066e-01 3.42814118e-01 -9.11658444e-03 -9.64162052e-01
4.97167856e-01 4.65185493e-01 6.93408847e-01 -8.87810946e-01
1.25863528e+00 5.98757088e-01 5.62532067e-01 -1.32053673e+00
-1.15241624e-01 -2.16077402e-01 -1.34291387e+00 -9.89381745e-02
5.56744874e-01 -2.38001630e-01 -3.83593500e-01 4.56319362e-01
-1.75646651e+00 7.69810379e-02 -1.70072138e-01 2.34292075e-01
-8.61965477e-01 8.93750370e-01 -3.08320463e-01 -7.93278456e-01
-7.72268623e-02 -9.92997408e-01 1.14220357e+00 4.38408032e-02
-1.68333411e-01 -8.95393729e-01 7.55278319e-02 -3.91120004e-04
-2.71762222e-01 6.17897272e-01 1.45823109e+00 -2.40720794e-01
-7.75253534e-01 -5.20583153e-01 -4.11828876e-01 5.48461825e-02
2.25303978e-01 6.70803130e-01 -5.53430736e-01 2.01463252e-01
4.92380513e-03 2.04695389e-01 5.42862892e-01 8.06928501e-02
1.11228502e+00 -1.84870645e-01 -6.92563713e-01 7.66504109e-01
1.45479381e+00 4.98106360e-01 8.36242795e-01 8.78974497e-02
7.30490685e-01 4.77269053e-01 4.16484356e-01 2.14660704e-01
-3.11010610e-02 5.68982542e-01 4.73376550e-02 -2.19236732e-01
-2.60389477e-01 -4.54433620e-01 -2.86973625e-01 9.41532969e-01
-4.38665539e-01 2.03125119e-01 -5.03871381e-01 -1.32626548e-01
-1.49040163e+00 -9.73257899e-01 -7.69273818e-01 2.19548583e+00
5.58430731e-01 2.63720006e-01 1.31990090e-01 5.26894450e-01
7.08437622e-01 -1.81630477e-01 -3.43814790e-01 -7.56337643e-01
6.30530566e-02 -7.94736892e-02 -1.16996445e-01 5.27871013e-01
-5.43153286e-01 6.73514426e-01 7.11315107e+00 9.62182462e-01
-8.05822730e-01 -5.32662749e-01 8.50720853e-02 6.01125121e-01
-8.18674266e-01 9.37346369e-02 -6.72320783e-01 8.06105584e-02
-2.61891246e-01 -4.18517143e-01 -2.39450671e-02 5.50144851e-01
7.21679851e-02 -8.03305209e-02 -1.26747799e+00 1.40285015e+00
-3.96773517e-02 -1.34127653e+00 7.73329854e-01 -2.76072491e-02
7.12612152e-01 -1.01907611e+00 -1.67036444e-01 -4.97222543e-02
-6.73020855e-02 -8.54289114e-01 7.64280438e-01 1.00182652e+00
7.50118673e-01 -6.82781279e-01 2.27869615e-01 8.26685250e-01
-1.20071888e+00 5.32372117e-01 -6.27119303e-01 1.17224827e-01
1.94501296e-01 8.01759481e-01 -1.10283506e+00 9.68518794e-01
5.29916249e-02 5.11840224e-01 -1.57712385e-01 1.61399913e+00
-1.22077212e-01 7.52664432e-02 -5.36567092e-01 5.76047897e-02
8.06922317e-02 -1.10252023e+00 6.83290899e-01 1.09923637e+00
4.99906957e-01 6.09559894e-01 1.20760739e-01 1.28462815e+00
1.79038510e-01 2.97926277e-01 -1.04160154e+00 3.91651541e-01
6.32019460e-01 9.57977951e-01 -1.05453146e+00 -3.76420826e-01
-3.88848156e-01 1.02722812e+00 -2.59249546e-02 3.98067534e-01
-6.89939201e-01 -6.94507182e-01 -2.85258979e-01 1.65882170e-01
3.48869562e-01 -8.90762806e-01 -6.21855021e-01 -9.25117493e-01
8.80069751e-03 -2.57313073e-01 -8.24280083e-02 -1.34616017e+00
-1.00882614e+00 5.62474906e-01 4.18373257e-01 -1.65146232e+00
-6.61270395e-02 -6.62980318e-01 -9.52521026e-01 8.54238629e-01
-1.29862165e+00 -1.26006675e+00 -3.08145911e-01 4.83166993e-01
7.88962364e-01 2.48031914e-02 8.13420594e-01 -3.16399634e-01
-2.44034693e-01 9.65874717e-02 1.06688365e-01 -2.57841885e-01
1.40989974e-01 -9.56416547e-01 2.70832688e-01 3.53073716e-01
2.47341245e-02 4.41023558e-01 4.18282688e-01 -8.26641202e-01
-1.37851644e+00 -6.07676327e-01 2.23655403e-01 -9.78083536e-02
1.25368252e-01 -2.35055372e-01 -1.09065008e+00 6.10118210e-01
-7.40351081e-02 -4.11003888e-01 4.25446540e-01 -4.62599546e-01
-3.77117634e-01 1.40787780e-01 -1.59317648e+00 8.01235318e-01
1.15034318e+00 -7.49426931e-02 -1.12857127e+00 6.72479719e-02
5.75149417e-01 -4.04296339e-01 -7.45945036e-01 3.77384365e-01
8.64524961e-01 -1.14471972e+00 1.28352010e+00 -3.48935455e-01
1.05205454e-01 -6.33987963e-01 1.30447805e-01 -1.15847862e+00
-4.69724387e-01 -6.25085235e-01 1.44429326e-01 1.06327975e+00
1.68481171e-02 -6.81702435e-01 5.73770642e-01 4.22308534e-01
-2.26458907e-01 -6.28863394e-01 -6.77181482e-01 -8.50442946e-01
-9.43842381e-02 -9.92252380e-02 9.18998122e-01 8.33532512e-01
-5.01054227e-01 2.57275403e-01 -6.81886375e-02 -7.71488296e-03
9.56869960e-01 5.63236535e-01 1.21360242e+00 -1.84366870e+00
8.48025829e-02 -7.97951996e-01 -4.02562171e-01 -1.59390652e+00
6.87857866e-02 -8.23698640e-01 -3.78533751e-01 -1.85539091e+00
-7.23200068e-02 -7.26037860e-01 7.42645144e-01 -9.48709026e-02
3.59924883e-01 4.40729670e-02 1.48980975e-01 4.96866733e-01
1.55562028e-01 6.61889136e-01 1.92201746e+00 -4.92437147e-02
-3.00349265e-01 3.69560659e-01 -2.47388527e-01 1.16350043e+00
2.71597207e-01 -1.13107704e-01 -4.62613612e-01 -4.51088548e-01
-2.35662222e-01 3.22326720e-01 2.10419491e-01 -5.86823285e-01
-6.98957369e-02 -1.92154959e-01 6.70446217e-01 -1.50689697e+00
5.83614409e-01 -1.25847566e+00 6.91293895e-01 3.70258838e-01
9.28428918e-02 -2.81137712e-02 5.11860311e-01 3.94885540e-01
2.47812644e-01 -8.23116004e-01 6.89231157e-01 -3.43221933e-01
-2.46640965e-01 1.72934860e-01 -5.70534915e-02 -6.68729901e-01
1.06610131e+00 -1.09790945e+00 -9.92029812e-03 -9.94321406e-02
-6.68944716e-01 -4.74465154e-02 8.04923415e-01 -3.80368461e-03
1.15458798e+00 -1.82218516e+00 -7.16090739e-01 6.09240353e-01
-4.88437772e-01 6.30092084e-01 2.92610489e-02 2.97441453e-01
-1.30106175e+00 3.25205564e-01 -5.20363748e-01 -7.16759205e-01
-1.17243087e+00 6.14525557e-01 1.86952293e-01 2.65749842e-01
-1.30154395e+00 3.94593209e-01 5.20292222e-01 -3.47705781e-01
1.48929572e-02 -6.96347177e-01 -7.18852282e-02 -2.53658861e-01
3.40097517e-01 7.46157706e-01 -3.45452875e-01 -4.39251125e-01
9.51595902e-02 1.49069726e+00 1.79313764e-01 -1.10838801e-01
1.41035044e+00 1.44840211e-01 -3.02459419e-01 4.08486515e-01
1.17440665e+00 5.17741740e-01 -9.89902079e-01 -4.20372486e-02
-2.68463194e-01 -8.40869665e-01 -3.01975548e-01 -2.93545306e-01
-7.79745817e-01 9.80020344e-01 -3.14939022e-01 5.29124558e-01
7.68346310e-01 -1.87436193e-02 6.84868932e-01 3.59373391e-01
6.34845555e-01 -3.72030288e-01 8.28629956e-02 1.15343125e-03
1.56676149e+00 -4.15674537e-01 5.00093699e-01 -1.21036303e+00
-1.51745081e-01 1.97672117e+00 2.72761196e-01 -7.11768568e-01
5.55216849e-01 2.88957953e-01 -5.50113976e-01 -3.10023725e-01
-2.25391984e-01 4.40155834e-01 5.74307203e-01 8.14023376e-01
3.61091308e-02 -4.16309573e-02 -2.22916856e-01 7.50289083e-01
-3.15436959e-01 -1.44061610e-01 9.14990485e-01 7.76326001e-01
-6.82966828e-01 -1.09996939e+00 -9.89388227e-01 2.99592167e-01
3.64321619e-01 4.41938490e-01 -3.43267769e-01 1.05881548e+00
-2.69497037e-01 2.36949369e-01 3.21700871e-01 9.02099907e-02
8.83209765e-01 -1.13533260e-02 1.10192919e+00 -5.97392738e-01
3.46018404e-01 4.76651400e-01 5.82769848e-02 -1.39813602e-01
-4.01001036e-01 -5.43641984e-01 -1.40595222e+00 -1.98853508e-01
-5.63298643e-01 -4.52869050e-02 4.78380084e-01 6.12899899e-01
-3.93149331e-02 1.89320162e-01 9.00260806e-01 -1.29311359e+00
-3.55946928e-01 -4.17841792e-01 -9.30337250e-01 8.52839872e-02
2.34307617e-01 -9.88963008e-01 -2.62641311e-01 1.96167424e-01] | [8.780097961425781, -3.4324569702148438] |
6a560026-55f4-46d2-8406-871498dd2f01 | group-channel-pruning-and-spatial-attention | 2306.01526 | null | https://arxiv.org/abs/2306.01526v1 | https://arxiv.org/pdf/2306.01526v1.pdf | Group channel pruning and spatial attention distilling for object detection | Due to the over-parameterization of neural networks, many model compression methods based on pruning and quantization have emerged. They are remarkable in reducing the size, parameter number, and computational complexity of the model. However, most of the models compressed by such methods need the support of special hardware and software, which increases the deployment cost. Moreover, these methods are mainly used in classification tasks, and rarely directly used in detection tasks. To address these issues, for the object detection network we introduce a three-stage model compression method: dynamic sparse training, group channel pruning, and spatial attention distilling. Firstly, to select out the unimportant channels in the network and maintain a good balance between sparsity and accuracy, we put forward a dynamic sparse training method, which introduces a variable sparse rate, and the sparse rate will change with the training process of the network. Secondly, to reduce the effect of pruning on network accuracy, we propose a novel pruning method called group channel pruning. In particular, we divide the network into multiple groups according to the scales of the feature layer and the similarity of module structure in the network, and then we use different pruning thresholds to prune the channels in each group. Finally, to recover the accuracy of the pruned network, we use an improved knowledge distillation method for the pruned network. Especially, we extract spatial attention information from the feature maps of specific scales in each group as knowledge for distillation. In the experiments, we use YOLOv4 as the object detection network and PASCAL VOC as the training dataset. Our method reduces the parameters of the model by 64.7 % and the calculation by 34.9%. | ['Jiafeng Lu', 'Yongqing Chen', 'Zhuhua Hu', 'Yong Bai', 'Pu Li', 'Yun Chu'] | 2023-06-02 | null | null | null | null | ['quantization', 'model-compression'] | ['methodology', 'methodology'] | [ 0.2092469 -0.24338531 -0.13600956 -0.306323 0.2019245 -0.03241972
-0.05434289 -0.01305179 -0.78836715 0.3046437 -0.31109983 -0.07264562
-0.19400002 -1.1181128 -0.61668915 -0.827889 0.24756986 0.0262234
0.76877123 0.16317877 0.06153166 0.52743113 -1.4818175 0.30245847
0.92493385 1.4016051 0.71491635 0.1104269 -0.3107952 0.7261115
-0.73842233 -0.06930958 0.21122706 -0.5417975 -0.30935436 -0.10314039
0.04360507 -0.4535518 -0.7014022 1.4812484 0.47162774 0.09925044
0.23155871 -0.9641223 -0.03512308 0.8493476 -0.579693 0.25705844
-0.39561266 -0.02769327 0.42060962 -0.518328 0.36929098 1.1978296
0.4932863 0.49728003 -0.94636357 -1.2695125 0.32858425 0.6394431
-1.7016262 -0.5157864 0.91204804 -0.10545643 0.47204444 0.2315226
1.0067931 0.5402617 -0.26004577 0.7305781 0.6582558 -0.300582
0.21134622 0.23204647 0.09097533 0.85690427 0.40819013 -0.212781
-0.25816733 0.3738386 0.9398552 0.34801775 -0.42205176 -0.02015807
-0.8238916 0.65849984 0.7875857 0.24040087 -0.17478411 0.12389792
0.47597203 0.18855558 0.11732199 0.13646668 -0.49707776 0.14209712
-1.1492093 -0.01189194 0.5757999 0.8618182 0.92204523 -0.03194425
-0.07325939 1.049354 0.3202508 0.29290184 0.81237566 -0.69417053
0.6557769 0.99605453 -0.6262368 -1.4660212 -0.2706905 -0.646651
-1.1967565 -0.11594184 0.01049498 -0.16948406 -0.9843556 1.5972258
0.33165827 0.27509397 -0.13166866 0.80155826 0.8883755 0.7980815
-0.02366666 -0.53258145 1.2311887 -1.0823156 -0.6390241 -0.19645305
0.6814415 -0.48303324 0.8372756 0.4834248 -0.8274628 -0.7705596
-1.2004023 0.06796769 -0.22015038 0.5301749 0.5800952 0.25359738
-0.5221909 0.8438119 -0.80095464 0.08122377 0.80458575 0.4860638
0.01962868 -0.2914207 -1.1817187 0.46726844 0.9253826 0.23163044
-0.75216043 -0.3650577 -0.7327663 0.4733483 0.36080033 -0.19056933
0.84271693 -1.0902706 -1.193849 0.27968583 0.07261989 -0.4751724
0.24554317 0.07010227 -0.45516276 0.28450465 -0.23121405 0.7379257
0.8394588 -0.7118737 -0.8019482 -0.2806295 -0.05736941 0.24479078
-0.77716726 -0.18325035 -0.87663823 -0.5965994 0.5662286 -0.4816603
-0.2970198 0.04146387 -0.34032816 -0.10844701 1.2723247 -0.72512215
1.5139811 -2.613359 0.15790018 0.3616567 0.32307014 0.58019894
-0.06995145 -0.46687484 0.0753966 0.1403696 -0.20647761 -0.14460127
-0.58228517 0.3671065 -0.16521351 0.10064025 -0.02810559 0.40392202
-0.47292057 -0.94050044 0.09013923 0.29542124 -0.7679554 0.13673128
-0.13273521 0.07039972 -0.5879939 0.4329483 0.91296357 -0.00665811
-0.06194683 -0.5158166 -0.17451842 0.18503311 -1.3431364 1.4869868
-0.22428738 0.23788717 0.20814596 -1.1566169 1.126101 -0.11813458
0.2797273 -0.67926455 0.36072275 0.27192736 0.2632434 -0.45167747
0.02663356 0.21269663 0.37072298 0.03209986 0.0379475 0.06244766
0.28834367 0.01479827 0.8902377 -0.20619002 -0.1121543 0.09726679
0.5329622 -0.09484894 0.94103914 0.3612565 -0.05084009 0.3095476
0.5833969 -0.35004565 -0.6939794 -0.33471358 -0.25126338 0.78486943
0.5237857 -0.5398289 -0.89308405 -0.6113512 -0.14238107 0.31723067
-0.3054297 -0.6794124 -0.6685025 -0.8727379 0.45014682 0.4094438
1.2099823 -1.154396 -0.56512 0.10463649 0.10716756 -1.0638545
-0.13077903 0.28977838 -1.2126057 -1.0496055 -0.5649534 -1.1192619
0.8794982 0.31373093 0.2640662 0.48832306 -0.20471564 -0.4837169
-0.31192383 -0.4237976 0.15748589 0.43593735 -0.11640596 -0.06095982
0.35055405 -0.6860696 -0.56935996 0.26352334 -0.91051567 0.32689142
0.9009284 0.72336566 0.9027116 0.566775 0.1680437 -0.7055062
0.54468936 -0.2558483 -0.68731606 0.01456307 -0.6243265 -0.10440861
0.9473402 -0.74551773 -0.7345349 0.22966862 -0.07822374 -0.7766408
0.0437805 0.5596239 -0.41122004 -0.31116867 0.35730055 0.44781366
0.06341128 -0.6659652 -0.13074607 0.6404977 0.28254375 0.00938657
0.6737544 0.15785067 -0.0269891 -0.70924044 -0.7054239 -0.0109676
-0.38877043 0.01982975 0.75271213 -0.8175146 -0.5631486 0.6482904
-1.1754678 -0.24318132 -0.33848083 0.61562276 0.09303192 0.3644169
-0.74201155 -0.51996696 -0.32221484 -1.2075703 0.5307284 0.69488984
0.4114582 -0.28413877 -0.5078887 -0.1136821 0.43008286 -0.26792502
0.97183883 -0.53941697 -0.65668625 -0.18463543 -0.55888134 0.5958336
-0.14639828 -0.06117249 -0.84758204 -0.11651466 0.2959458 -0.16720963
1.2842147 0.3528945 1.9255853 -0.4454204 -0.53956187 1.079854
1.1915125 0.5115945 0.61518425 0.27054107 0.8853835 0.38656697
0.55047154 0.3150651 -0.06394436 0.43562502 0.5687325 -0.05148568
-0.21866472 -0.41744727 0.15931788 1.2607145 -0.07808124 0.19344907
-0.44594765 0.18402559 -1.6722791 -0.9235679 0.11990844 2.1680305
1.0292557 0.42700005 -0.3011056 0.36367175 0.8242272 0.18561871
-0.7098 0.0360612 -0.06360475 0.06984712 0.56273353 0.13684566
-0.9657641 0.9317438 5.2233706 1.394712 -1.3481954 0.02559497
0.67035913 -0.3339412 0.06346716 0.07082491 -1.2358967 0.7711893
0.46663612 0.14383748 0.6069333 1.1775217 0.03930423 -0.04274313
-0.6867882 1.1964648 -0.01444905 -1.1227953 0.04708018 -0.17986219
0.43997905 -0.06900646 -0.21835569 0.51507133 -0.33798012 -0.84846276
0.5634355 0.39423385 0.76968074 -0.85981643 0.87671363 0.69257957
-1.2838186 -0.3155092 -1.0493826 0.0638117 -0.1709961 0.92310286
-0.15412682 0.10796534 0.8642925 0.7170003 -0.6234571 1.2944992
-0.2664582 0.5468627 -0.6193012 -0.11440717 -0.06841613 -0.36648124
0.16827245 0.9659393 0.24252033 0.12736073 0.14412802 0.7293326
-0.28766385 0.15489304 -0.28297353 0.0837635 0.9063977 1.2740636
-0.7048005 -0.48356253 -0.1376571 0.89785284 0.35987917 0.2878716
-0.79600155 -0.95626384 0.13830814 0.05314006 0.471553 -0.05261533
-0.22625796 -0.9820279 0.20506941 -0.79919034 0.17730965 -0.414014
-0.5647165 0.5157748 -0.0545879 -1.1185344 0.2838185 -0.45463473
-0.5934973 0.8195569 -1.4400896 -0.6180248 -0.74483794 0.5532215
0.47280747 -0.26919827 0.4314864 0.69005793 -1.094397 0.6620235
0.01599371 0.21852793 0.42979553 -0.40808278 -0.18899462 0.72968346
-0.15627564 0.6606473 0.15685622 -0.507743 -0.97487533 -1.2912687
0.60510665 0.5839433 0.21535154 -0.37346023 -1.0715934 0.39448875
-0.43676916 0.14047515 0.30830187 -0.13151821 -0.1351487 -0.7318244
-1.089213 0.58606714 1.0014787 -0.219461 -0.32611108 0.34822345
1.0364488 -0.41366932 -0.5614144 0.44372463 0.33928218 -0.81319827
0.6851946 -0.10274088 0.51450145 -0.5165029 0.04725153 -1.0300118
-0.49191794 -0.0363217 -0.26261002 1.2092099 0.21438074 -0.59359384
0.9859312 0.09971384 -0.14735624 -1.1771718 -0.76676214 -0.46484807
-0.45414343 -0.2745617 0.7615625 0.8318223 -0.31935647 0.3027228
-0.1424574 0.0584297 0.33053967 0.05299824 0.46952802 -1.2265596
-0.42019373 -0.5570591 -0.44771913 -1.4971762 -0.22537822 -0.71220714
-0.12930717 -1.2966876 0.12377182 -0.77943766 -0.28007525 0.6482549
0.03806297 -0.03240579 0.2619192 0.64307415 -0.5594446 0.69839543
1.3638829 -0.12398325 -0.36043143 0.05233138 -0.529246 0.9421765
0.76884615 -0.6207926 -0.5395148 -0.6128907 -0.07370206 -0.20694382
0.19347481 -1.4258597 0.4890037 -0.04916174 0.7224471 -0.59277844
0.22833282 -1.0289325 -0.18847294 0.95627797 -0.1640758 -0.25793844
0.23179856 0.43843442 -0.35608137 -0.6946006 0.9318737 -0.06218833
-0.78697187 0.545666 0.05282281 -0.21938826 1.0385911 -0.34979922
-0.24202585 0.15100053 -0.54063386 0.45408022 0.1951121 0.11463763
0.7744988 -1.3114228 -0.2136393 0.7034674 -0.3970678 0.7245879
0.3769995 0.8812513 -0.7349774 0.09012001 -0.20304593 -0.45707402
-1.2580317 0.53062487 0.44334257 -0.11469021 -0.49064064 0.96024793
0.1861037 -0.06196479 0.566066 -0.3589396 -0.6027587 0.09624983
0.7161047 0.36115545 0.0110791 -0.2779181 -0.20342425 0.5774449
-0.36096588 0.32599777 1.1179152 0.18541515 -0.324589 0.12884213
1.2210871 -0.17709963 -1.1885513 -0.14688465 -0.4996544 -0.19955549
0.35740954 -0.34402832 -1.6830336 1.0196258 0.7204539 0.16165464
1.4735796 -0.2811013 0.8885242 0.5175379 0.17256463 -1.1753887
-0.16880824 0.5852304 0.5360047 -0.66186917 0.2226567 -0.72001183
-0.22589724 0.99503356 1.1153703 -0.29563126 0.8048389 0.28388605
-0.36390373 0.00650273 -0.44038314 0.14083056 -0.05233983 0.40089402
-0.06098896 -0.214978 -0.54701453 0.9865163 -0.21946761 0.06128697
0.08927123 0.7106929 -0.9948804 -0.734543 -0.12320961 0.83528346
-0.23940095 -0.26760325 -0.08859923 0.5290794 0.79676485 0.57806957
0.2643698 -0.8891805 0.3955287 -0.27448088 0.18586102 -0.6295482
-0.3901812 0.07555713 -0.37040716 -0.46806368 -0.04314033 -0.19267824
-1.4304245 -0.15800597 -0.7497859 0.20787114 0.597632 0.80104136
0.34836206 0.85112906 0.55198014 -0.61397636 -0.63885504 -1.047636
-0.6536465 0.12698981 -0.07985514 -0.54488057 -0.41488987 -0.27643114] | [8.54625415802002, 3.0406508445739746] |
7c11b262-c0af-4070-9af4-79c1718b046b | dermatologist-level-dermoscopy-skin-cancer | 1810.10348 | null | http://arxiv.org/abs/1810.10348v1 | http://arxiv.org/pdf/1810.10348v1.pdf | Dermatologist Level Dermoscopy Skin Cancer Classification Using Different Deep Learning Convolutional Neural Networks Algorithms | In this paper, the effectiveness and capability of convolutional neural
networks have been studied in the classification of 8 skin diseases. Different
pre-trained state-of-the-art architectures (DenseNet 201, ResNet 152, Inception
v3, InceptionResNet v2) were used and applied on 10135 dermoscopy skin images
in total (HAM10000: 10015, PH2: 120). The utilized dataset includes 8
diagnostic categories - melanoma, melanocytic nevi, basal cell carcinoma,
benign keratosis, actinic keratosis and intraepithelial carcinoma,
dermatofibroma, vascular lesions, and atypical nevi. The aim is to compare the
ability of deep learning with the performance of highly trained dermatologists.
Overall, the mean results show that all deep learning models outperformed
dermatologists (at least 11%). The best ROC AUC values for melanoma and basal
cell carcinoma are 94.40% (ResNet 152) and 99.30% (DenseNet 201) versus 82.26%
and 88.82% of dermatologists, respectively. Also, DenseNet 201 had the highest
macro and micro averaged AUC values for overall classification (98.16%, 98.79%,
respectively). | ['Amirreza Rezvantalab', 'Somayeh Karimijeshni', 'Habib Safigholi'] | 2018-10-21 | null | null | null | null | ['skin-cancer-classification'] | ['medical'] | [-2.76720710e-02 2.67403752e-01 -7.12500280e-03 8.95507708e-02
4.83108386e-02 -4.12876904e-01 5.90849876e-01 1.06130771e-01
-4.98224974e-01 8.88077974e-01 -1.69973165e-01 -3.42390656e-01
-4.82938498e-01 -7.45208204e-01 2.02912614e-02 -8.11267614e-01
3.79240699e-02 2.60346830e-01 -7.27669150e-03 -1.43976107e-01
1.39250189e-01 6.57797575e-01 -1.17843330e+00 4.22166467e-01
1.19764674e+00 1.08330333e+00 -3.92683238e-01 1.10945010e+00
2.05435697e-02 9.06632364e-01 -6.49240494e-01 -7.42574811e-01
2.41635293e-01 -1.20559186e-01 -6.66532993e-01 2.89129317e-02
7.57254183e-01 -4.38867927e-01 -2.91908950e-01 9.70246255e-01
6.70383632e-01 -4.07307208e-01 1.24260414e+00 -7.10609376e-01
-9.00119483e-01 1.75505981e-01 -9.92883742e-01 1.91617757e-01
-2.38276348e-02 3.14171344e-01 3.23982507e-01 -5.22811592e-01
8.64423573e-01 8.12076688e-01 1.00893879e+00 8.12514961e-01
-6.09908819e-01 -4.72984999e-01 -3.37597281e-01 3.05799454e-01
-1.39798450e+00 1.17085814e-01 -2.04556614e-01 -6.84029520e-01
9.49154556e-01 4.48858500e-01 8.71931434e-01 1.12922513e+00
7.54022658e-01 3.10014188e-01 1.58139360e+00 -3.14855993e-01
1.24209724e-01 5.71533203e-01 3.90650034e-01 8.01300645e-01
4.94007468e-01 1.48661882e-02 6.25566319e-02 -2.12050341e-02
9.92712796e-01 -8.02885219e-02 -3.03209368e-02 5.56538701e-01
-3.42215300e-01 4.84689176e-01 4.53257322e-01 3.89745831e-01
-7.53145337e-01 -1.81124151e-01 5.33178866e-01 4.48513240e-01
2.54586458e-01 3.28753173e-01 -1.39462382e-01 2.31740102e-01
-4.90087420e-01 -1.55499905e-01 7.88847148e-01 2.78869390e-01
2.70714372e-01 4.24104221e-02 -1.39492720e-01 1.05296504e+00
-2.47846022e-02 2.53726929e-01 6.66936815e-01 -3.87364328e-01
-1.23773813e-01 7.85031676e-01 -1.89232826e-01 -6.99423790e-01
-6.23335898e-01 -4.93333995e-01 -1.52988482e+00 4.25842345e-01
4.98239011e-01 -5.03587544e-01 -1.71445322e+00 8.32960427e-01
2.48616245e-02 -1.20897010e-01 1.90996066e-01 8.35066199e-01
1.16425014e+00 2.10190609e-01 4.21957076e-01 2.17887789e-01
1.30506420e+00 -7.59741247e-01 -5.64329565e-01 1.50588199e-01
2.90766478e-01 -8.38897705e-01 3.85240406e-01 8.60376418e-01
-1.14305246e+00 -6.36506498e-01 -8.83666813e-01 2.75923043e-01
-5.46612382e-01 4.84604329e-01 8.05741310e-01 7.46218085e-01
-1.39450979e+00 4.91289794e-01 -5.43849289e-01 -9.77621913e-01
4.94744152e-01 4.63187248e-01 -6.70313418e-01 3.51519361e-02
-1.04996169e+00 1.16361248e+00 2.57785082e-01 1.41229331e-01
-8.69354784e-01 -7.50300705e-01 -2.53205001e-01 -3.42841476e-01
-7.54533857e-02 -7.66818941e-01 9.14845705e-01 -1.18766344e+00
-1.29621327e+00 1.28996193e+00 4.64257687e-01 -6.77972496e-01
6.61551356e-01 7.54589913e-03 -7.80378580e-01 5.67745149e-01
-4.40505892e-01 6.03864491e-01 3.71565312e-01 -8.85390103e-01
-8.98070633e-01 -5.25801361e-01 1.97427154e-01 -4.69405809e-03
-3.17732126e-01 -1.95354238e-01 -6.45681918e-02 -1.37376472e-01
-3.68157178e-01 -7.92959452e-01 -3.57263267e-01 3.00210774e-01
-8.56063068e-01 -3.25732350e-01 4.41830426e-01 -9.87550139e-01
1.13887846e+00 -1.92184806e+00 -4.56727952e-01 5.19601345e-01
6.92793548e-01 1.05511737e+00 -8.75390172e-02 4.60974544e-01
-2.33259588e-01 3.57077211e-01 1.56374276e-01 2.95805007e-01
-4.60171402e-01 -4.73453924e-02 5.28413117e-01 3.87083560e-01
3.10504109e-01 8.40537012e-01 -5.08196592e-01 -4.90779072e-01
7.24699914e-01 7.85468221e-01 1.61238894e-01 -1.56394571e-01
2.75452793e-01 -1.94321856e-01 -1.78135321e-01 1.16205513e+00
9.64592993e-01 -5.15906870e-01 2.60359824e-01 -3.82961601e-01
8.65228102e-02 -5.85279882e-01 -9.60905373e-01 9.32810664e-01
-6.55976176e-01 8.18413913e-01 2.03885123e-01 -2.91201025e-01
8.17412496e-01 4.02134866e-01 2.52516210e-01 -5.29622078e-01
3.05477470e-01 3.79178613e-01 5.43481767e-01 -9.87798691e-01
-1.52616957e-02 -4.99034598e-02 7.79259682e-01 -1.00298285e-01
1.64776683e-01 4.45691705e-01 2.99492896e-01 5.04531041e-02
1.26309943e+00 -5.74497283e-01 6.72722638e-01 -2.33179927e-01
6.59754694e-01 -5.97458184e-02 -8.86278749e-02 5.78863263e-01
-3.86484027e-01 4.99899089e-01 7.34864831e-01 -7.04130650e-01
-1.02074933e+00 -1.36756551e+00 -4.57213968e-01 3.41956586e-01
-2.09761590e-01 1.48735166e-01 -9.28090811e-01 -6.76453531e-01
1.55507177e-01 1.79492041e-01 -1.07891750e+00 9.53706726e-02
-3.44405025e-02 -9.64478612e-01 8.09645116e-01 6.09385490e-01
1.01505172e+00 -7.89873362e-01 -2.47934699e-01 -1.44907653e-01
5.90843856e-01 -6.92331195e-01 3.90724331e-01 -1.32607043e-01
-4.78642255e-01 -1.59710884e+00 -1.08322322e+00 -7.12626338e-01
8.62063348e-01 -2.51181901e-01 8.79652083e-01 7.48415440e-02
-1.15788388e+00 1.14797503e-01 -1.63833916e-01 -4.11078095e-01
-4.76925254e-01 6.22794107e-02 -1.31560832e-01 9.52840969e-02
6.54569507e-01 -2.97926307e-01 -8.72841120e-01 -4.43518870e-02
-8.16902339e-01 -1.51664495e-01 1.29069507e+00 8.99817646e-01
4.11863655e-01 -1.39995292e-01 5.30372083e-01 -1.22875416e+00
8.20640504e-01 -5.97634792e-01 -1.22139923e-01 2.77608156e-01
-7.76514769e-01 -9.39421892e-01 6.99157298e-01 -3.74865770e-01
-9.69807208e-01 -9.44011733e-02 -3.85565192e-01 -4.34704691e-01
-6.40668929e-01 4.27006721e-01 6.75623000e-01 -4.29129064e-01
1.19720948e+00 7.82726705e-02 3.47611189e-01 -3.07037473e-01
-2.64328063e-01 1.14664483e+00 4.95721340e-01 2.20295757e-01
2.51510412e-01 3.67069572e-01 2.58003861e-01 -1.22560668e+00
-4.55376178e-01 -4.35196877e-01 -4.51768160e-01 -2.93196291e-01
1.04389274e+00 -8.21121037e-01 -5.68334103e-01 1.20723701e+00
-8.63230824e-01 -1.38385460e-01 -2.15329781e-01 3.04917902e-01
1.86246008e-01 4.83232349e-01 -9.65759873e-01 -9.94902134e-01
-8.84558022e-01 -6.11679494e-01 5.69705188e-01 8.02593529e-01
-2.56271482e-01 -1.47946668e+00 -4.69473302e-02 1.35341063e-01
6.35596156e-01 8.66566896e-01 8.63375425e-01 -7.81695604e-01
1.59951091e-01 -5.48243523e-01 -7.23110855e-01 8.72257590e-01
3.44096303e-01 7.51755953e-01 -1.16170144e+00 -9.86955091e-02
-6.18565023e-01 -4.62111652e-01 1.03013837e+00 7.98656821e-01
1.30422258e+00 -1.03255801e-01 -3.46269518e-01 7.28632808e-01
1.93695962e+00 5.81697226e-01 1.03297222e+00 4.44123745e-02
4.66591388e-01 5.20048797e-01 1.21187851e-01 2.80008167e-01
-9.30854157e-02 -2.88164586e-01 6.98013544e-01 -7.60880649e-01
-4.53433722e-01 1.40847042e-01 -2.74174452e-01 1.82998672e-01
-6.25093579e-01 -4.92543370e-01 -1.08857596e+00 7.04107404e-01
-1.04163611e+00 -5.97920597e-01 -4.51347589e-01 1.82264674e+00
4.87651139e-01 1.58608351e-02 1.65441737e-01 1.19544864e-01
9.86374795e-01 -2.75952727e-01 -7.30215728e-01 -8.15504491e-01
-1.25101611e-01 8.42916608e-01 6.75127566e-01 2.20635757e-01
-1.14174259e+00 6.02118194e-01 6.25539541e+00 9.68343914e-01
-1.39269292e+00 -2.37169713e-01 1.06136835e+00 1.53511778e-01
1.82259351e-01 -7.65965343e-01 -6.59293652e-01 4.56790745e-01
9.35545564e-01 1.47777095e-01 7.55020678e-02 6.40691638e-01
-2.39586085e-01 -3.19796413e-01 -5.99497676e-01 7.87238777e-01
-1.64763778e-01 -1.47895753e+00 1.27219617e-01 2.90637404e-01
9.89931524e-01 2.69594062e-02 4.30955082e-01 1.00640289e-01
2.21078798e-01 -1.53079593e+00 -5.51509917e-01 8.58102679e-01
1.53640592e+00 -7.17144132e-01 1.42886913e+00 -3.27270143e-02
-6.26074731e-01 -4.18419344e-03 -4.15926009e-01 8.55162665e-02
-4.86209333e-01 7.15157866e-01 -1.18392742e+00 6.04168713e-01
6.47407889e-01 6.70553148e-01 -7.17367887e-01 1.11386061e+00
-9.11391824e-02 5.76997101e-01 -2.86572546e-01 -5.34708202e-01
7.32708097e-01 7.32029527e-02 1.50464535e-01 1.32687545e+00
4.29020301e-02 -1.49277508e-01 -5.40911198e-01 5.47145784e-01
2.44687170e-01 2.93595612e-01 -4.18015301e-01 -2.06275880e-01
2.79493690e-01 1.80789411e+00 -3.68680507e-01 -2.22846165e-01
-1.51588395e-01 4.76111203e-01 -1.96543783e-01 3.55315745e-01
-6.46234453e-01 -8.11538577e-01 5.10192394e-01 3.43542248e-01
-1.64384797e-01 6.81709170e-01 -3.52084190e-01 -5.74546695e-01
-4.36257839e-01 -8.40092599e-01 5.85810721e-01 -7.05988884e-01
-1.83188975e+00 7.50777185e-01 -4.66856360e-01 -8.59804273e-01
-1.90896705e-01 -1.41467714e+00 -1.06121731e+00 1.17610872e+00
-1.45280385e+00 -1.40174448e+00 -1.03741205e+00 4.74216849e-01
2.29958370e-01 -6.85151219e-01 8.70839298e-01 -1.90111697e-02
-5.74379683e-01 8.68916273e-01 2.29204684e-01 2.40354612e-01
6.72858179e-01 -1.49977279e+00 -2.54730489e-02 2.38859549e-01
-9.68248844e-01 6.68061674e-01 -1.16266543e-02 -6.29944324e-01
-8.86253238e-01 -1.19008470e+00 4.28996563e-01 1.97165772e-01
7.13765502e-01 3.83700043e-01 -5.84136128e-01 1.16297431e-01
6.48502588e-01 -1.61056221e-01 1.15253007e+00 6.15277737e-02
-1.76501632e-01 -1.07129760e-01 -1.75558150e+00 5.55098772e-01
4.75053251e-01 -7.14543313e-02 6.01458140e-02 5.38717270e-01
-1.00543857e-01 -4.13386047e-01 -1.49679244e+00 4.55222636e-01
9.06250715e-01 -1.24205625e+00 6.59798801e-01 -7.37267971e-01
9.99442697e-01 2.69168109e-01 2.04966128e-01 -9.60106075e-01
-3.22463036e-01 -2.84972161e-01 -1.34863332e-01 7.84709752e-01
4.98761564e-01 -9.91203010e-01 1.03732646e+00 2.42539838e-01
-1.75240003e-02 -1.55843580e+00 -6.47087634e-01 -4.37047929e-01
3.51730943e-01 4.65423226e-01 8.57659131e-02 9.91404951e-01
-4.40621972e-01 9.67632011e-02 5.31903468e-02 -1.42119393e-01
4.57768708e-01 -5.71235478e-01 5.27282059e-01 -1.22588944e+00
-5.25773242e-02 -4.96045113e-01 -8.03396821e-01 -1.65912569e-01
-4.81245965e-01 -4.96708661e-01 -8.98075342e-01 -1.94737065e+00
2.64830649e-01 -3.04407567e-01 -6.07133329e-01 7.22695708e-01
-4.48615476e-02 3.81791800e-01 -8.64471197e-02 -1.26781151e-01
1.59195825e-01 -5.71617424e-01 1.60528696e+00 -1.17373578e-01
1.15315430e-01 2.06738263e-01 -7.77302980e-01 8.35423648e-01
1.13222969e+00 4.91921723e-01 -2.86588460e-01 -1.29320443e-01
-9.95862484e-02 -1.79812193e-01 4.73194599e-01 -1.27564573e+00
1.63170949e-01 -1.31001696e-01 9.36985791e-01 -5.70042908e-01
3.29045951e-01 -4.23978329e-01 1.59785733e-01 1.04578698e+00
-2.37201005e-01 -3.85616273e-01 2.17217162e-01 3.27192396e-01
-1.77334428e-01 -3.46901864e-01 1.38056028e+00 -4.30221349e-01
-9.47089136e-01 4.51313645e-01 -5.05427361e-01 -3.97360325e-01
1.52112591e+00 -6.82346106e-01 -8.75066817e-01 -1.82070926e-01
-9.83885884e-01 -3.77137810e-02 2.68955737e-01 -1.98474266e-02
7.10742891e-01 -8.29228401e-01 -9.55589890e-01 1.14159957e-01
4.52406593e-02 -2.75037020e-01 9.36211050e-01 9.89763975e-01
-1.21670377e+00 4.42241371e-01 -9.33028519e-01 -5.50123096e-01
-1.41729712e+00 -1.70836344e-01 8.79565418e-01 -3.11861426e-01
-2.67590079e-02 1.24167871e+00 -6.81319535e-02 -2.15560913e-01
2.31050357e-01 -1.42782658e-01 -4.00416762e-01 -8.54965597e-02
4.36920822e-01 9.45862889e-01 2.27365971e-01 6.78926408e-02
-2.65344650e-01 6.97026968e-01 -8.40337038e-01 5.89347064e-01
9.09548104e-01 5.31192899e-01 -2.39201084e-01 -4.37660664e-02
9.66508687e-01 -3.57990295e-01 -6.07218385e-01 3.76749933e-01
-6.21482074e-01 -2.04440892e-01 -1.02057075e-02 -1.81664681e+00
-1.10336876e+00 1.02694833e+00 1.16333616e+00 5.12889624e-01
1.38492084e+00 -6.46706820e-01 5.35170496e-01 1.15787014e-01
4.75298800e-02 -1.22980905e+00 -1.19558461e-01 1.64433360e-01
5.45537412e-01 -1.05791819e+00 8.45390335e-02 -5.59986889e-01
-7.67503798e-01 1.47833526e+00 1.20372355e+00 -6.31016195e-01
7.13448167e-01 1.93618163e-01 3.59245151e-01 -2.46750534e-01
-9.32390928e-01 -1.76278487e-01 3.00591737e-01 1.08797169e+00
5.42583585e-01 1.49377167e-01 -6.86596334e-01 2.80487031e-01
8.08053315e-02 1.23688206e-01 7.35739648e-01 4.63683486e-01
-3.00083011e-01 -5.03916144e-01 5.83455674e-02 1.26900637e+00
-8.91614735e-01 2.55944151e-02 -1.06815565e+00 1.54161501e+00
5.51854908e-01 6.40684664e-01 2.13891402e-01 -6.40803993e-01
1.19232476e-01 -4.36402321e-01 4.76775646e-01 -4.94742185e-01
-1.06711698e+00 -1.42604085e-02 4.05374348e-01 -1.33438289e-01
-9.27755758e-02 -1.37327597e-01 -7.69464016e-01 -5.33810735e-01
-3.13985527e-01 -2.02368960e-01 9.24692988e-01 4.54913974e-01
1.32178098e-01 5.69727898e-01 3.79157633e-01 1.80331022e-01
-7.02563822e-01 -1.40736747e+00 -9.85111296e-01 -9.90568288e-03
2.11188689e-01 -2.84422070e-01 -3.45200479e-01 -1.02866277e-01] | [15.714210510253906, -3.0189549922943115] |
0a691553-cd5c-4adb-bd07-9f3cf031b602 | context-enriched-molecule-representations | 2305.09481 | null | https://arxiv.org/abs/2305.09481v1 | https://arxiv.org/pdf/2305.09481v1.pdf | Context-enriched molecule representations improve few-shot drug discovery | A central task in computational drug discovery is to construct models from known active molecules to find further promising molecules for subsequent screening. However, typically only very few active molecules are known. Therefore, few-shot learning methods have the potential to improve the effectiveness of this critical phase of the drug discovery process. We introduce a new method for few-shot drug discovery. Its main idea is to enrich a molecule representation by knowledge about known context or reference molecules. Our novel concept for molecule representation enrichment is to associate molecules from both the support set and the query set with a large set of reference (context) molecules through a Modern Hopfield Network. Intuitively, this enrichment step is analogous to a human expert who would associate a given molecule with familiar molecules whose properties are known. The enrichment step reinforces and amplifies the covariance structure of the data, while simultaneously removing spurious correlations arising from the decoration of molecules. Our approach is compared with other few-shot methods for drug discovery on the FS-Mol benchmark dataset. On FS-Mol, our approach outperforms all compared methods and therefore sets a new state-of-the art for few-shot learning in drug discovery. An ablation study shows that the enrichment step of our method is the key to improve the predictive quality. In a domain shift experiment, we further demonstrate the robustness of our method. Code is available at https://github.com/ml-jku/MHNfs. | ['Günter Klambauer', 'Sepp Hochreiter', 'Friedrich Rippmann', 'Daniel Kuhn', 'Lukas Friedrich', 'Philipp Seidl', 'Johannes Schimunek'] | 2023-04-24 | null | null | null | null | ['drug-discovery'] | ['medical'] | [ 4.13531333e-01 -2.69918144e-02 -4.02271122e-01 -9.95348245e-02
-6.58636451e-01 -4.54932183e-01 5.56310952e-01 4.98401970e-01
-3.42778563e-01 1.24352837e+00 1.13068022e-01 -5.73638827e-02
-3.96756321e-01 -9.65313554e-01 -8.97426665e-01 -9.32692051e-01
-3.81425805e-02 5.56007743e-01 1.13950700e-01 -2.75125623e-01
2.41650999e-01 6.46003604e-01 -1.00763404e+00 2.17514634e-01
9.17794228e-01 5.78408062e-01 2.89022237e-01 1.21651836e-01
-9.42547545e-02 6.45279646e-01 -3.48513842e-01 -2.91854113e-01
6.33127689e-02 -5.68191409e-01 -8.95047307e-01 -4.26963329e-01
2.42374986e-01 -6.22965619e-02 -2.77185440e-01 1.29695106e+00
9.14408267e-01 5.13556957e-01 6.27725184e-01 -6.05484247e-01
-7.40932584e-01 5.67029119e-01 -2.55156755e-01 2.53689349e-01
3.00847262e-01 3.64290565e-01 9.11810517e-01 -1.33459949e+00
1.17846596e+00 8.14541757e-01 6.27832174e-01 6.94446921e-01
-1.48331869e+00 -8.20287049e-01 -1.20290488e-01 2.03148261e-01
-1.55206525e+00 -6.66099668e-01 5.38772166e-01 -5.29813588e-01
1.09264648e+00 -2.16333549e-02 6.30683124e-01 1.15078664e+00
1.80430725e-01 4.09656197e-01 5.91674447e-01 -2.11034089e-01
9.23595130e-01 1.82407856e-01 4.43127394e-01 5.47991753e-01
1.65913507e-01 3.78027439e-01 -6.17640913e-01 -4.50195432e-01
5.32994628e-01 3.74821514e-01 -5.38817406e-01 -5.37905753e-01
-8.60700130e-01 1.10813415e+00 7.01700747e-01 5.26219964e-01
-6.11529171e-01 1.30176842e-01 1.32028952e-01 1.49558470e-01
4.33191687e-01 1.24495459e+00 -3.27552170e-01 8.76523778e-02
-7.15769708e-01 1.55515015e-01 7.94428349e-01 7.08574116e-01
7.87352979e-01 -2.10364521e-01 -4.06216532e-01 5.33720613e-01
-8.06861967e-02 1.16657270e-02 2.46223837e-01 -6.24863029e-01
-1.90671071e-01 3.46427321e-01 1.54984966e-01 -4.59378004e-01
-3.11594754e-01 -6.12438023e-01 -6.64701939e-01 1.70356467e-01
3.67611438e-01 -3.17964077e-01 -1.02383733e+00 1.72571909e+00
2.31399804e-01 5.34797668e-01 2.44369566e-01 6.98476672e-01
9.58092988e-01 6.17149651e-01 5.02115548e-01 -4.64944988e-01
1.10267460e+00 -6.50335014e-01 -6.59866095e-01 2.15058953e-01
6.81050181e-01 -4.04660255e-01 6.91435099e-01 4.77716587e-02
-8.57035398e-01 -2.36588731e-01 -1.16205502e+00 3.42226297e-01
-7.54241943e-01 -4.90155518e-01 8.78241301e-01 3.26117933e-01
-5.60543001e-01 1.54029918e+00 -3.97347152e-01 -2.74616599e-01
1.04641151e+00 8.43014956e-01 -4.76005524e-01 -2.57649839e-01
-1.67202091e+00 8.94725502e-01 4.73979205e-01 -3.82544637e-01
-1.01616859e+00 -1.28641248e+00 -6.13243997e-01 4.29461688e-01
5.04920602e-01 -8.59354198e-01 1.05008936e+00 -8.36494982e-01
-1.57943273e+00 2.80228645e-01 -2.37980112e-01 -5.31610787e-01
-1.29822716e-01 -1.24493159e-01 -1.37862682e-01 2.05052584e-01
-1.85125478e-04 6.43186092e-01 6.10782862e-01 -7.61157036e-01
-1.32937893e-01 -4.04653162e-01 -1.47243738e-01 1.34907529e-01
-1.16555884e-01 -1.46938756e-01 -1.27075896e-01 -4.38068092e-01
-5.36111712e-01 -6.61160469e-01 -6.18552923e-01 -1.89452782e-01
-1.35067210e-01 -1.55521631e-01 4.00891662e-01 -1.60135761e-01
9.37183678e-01 -2.06649423e+00 4.17455256e-01 4.22283560e-01
4.52291220e-01 5.37958622e-01 -3.37565899e-01 6.91198707e-01
-7.22850144e-01 1.72251090e-01 -3.56322885e-01 3.72225732e-01
-4.69881564e-01 -2.42750525e-01 -3.13096821e-01 5.13041377e-01
3.86363298e-01 1.30505967e+00 -1.42638683e+00 1.90714505e-02
3.87516282e-02 6.99669242e-01 -5.05266309e-01 -1.34530783e-01
-5.49319625e-01 5.46654463e-01 -4.70243007e-01 6.25675499e-01
3.69668245e-01 -6.79130137e-01 3.67857032e-02 -9.94256437e-02
9.47443545e-02 1.17658518e-01 -9.14595902e-01 1.96385443e+00
1.31935611e-01 1.97216853e-01 -7.71477520e-01 -7.79924631e-01
6.39608502e-01 6.44030154e-01 5.73886514e-01 -6.02176070e-01
1.61459729e-01 2.11669743e-01 3.81162971e-01 -2.40364790e-01
-1.95193216e-01 -6.31568968e-01 5.44779897e-01 8.42750967e-02
3.97354245e-01 8.49710405e-02 2.08781883e-01 3.37103695e-01
1.42051959e+00 -6.05714396e-02 6.82228446e-01 -2.11919248e-01
1.07041247e-01 -7.02205300e-02 5.63016295e-01 6.32908881e-01
-1.42257199e-01 1.71212673e-01 4.25643593e-01 -4.22303528e-01
-1.03268313e+00 -9.28957582e-01 -3.86462301e-01 7.28437901e-01
-9.07074008e-03 -5.15033126e-01 -5.51473975e-01 -6.87602460e-01
1.27473911e-02 7.55309820e-01 -9.74550843e-01 -4.95940328e-01
3.93265449e-02 -9.14661288e-01 2.46524379e-01 2.31619298e-01
-1.14766753e-03 -1.13094366e+00 -1.35417551e-01 4.80532408e-01
4.48517889e-01 -3.64991307e-01 -3.33378613e-01 6.99119329e-01
-8.02875876e-01 -1.09752202e+00 -1.02796292e+00 -5.96694887e-01
5.46450973e-01 3.75927985e-02 8.64212811e-01 -2.23865941e-01
-5.00075281e-01 -4.02639002e-01 -2.32353032e-01 -5.29934704e-01
-1.83540866e-01 -1.59907728e-01 1.14629112e-01 -1.44774035e-01
6.84156895e-01 -8.02466452e-01 -6.10140324e-01 -1.73990577e-01
-6.24211490e-01 -3.23639691e-01 6.50244415e-01 1.23868442e+00
7.41939723e-01 -1.89627353e-02 9.35568273e-01 -1.21172822e+00
6.69322550e-01 -6.06216788e-01 -5.60869873e-01 8.54126215e-02
-5.11781216e-01 4.29773182e-01 5.65063357e-01 -6.49312496e-01
-6.19106054e-01 4.49577838e-01 -2.05162778e-01 -5.45524299e-01
-1.94125116e-01 6.73832417e-01 -2.15097874e-01 -3.01211059e-01
1.05220997e+00 1.43922595e-02 -9.29305479e-02 -4.45901036e-01
4.10564452e-01 2.74802297e-01 1.31689794e-02 -2.64447302e-01
3.70517641e-01 2.70174682e-01 2.99272269e-01 -8.65642011e-01
-6.06741965e-01 -6.83053613e-01 -6.01395667e-01 3.50792855e-01
6.13263607e-01 -7.90435314e-01 -9.21309412e-01 -1.98416393e-02
-1.15799963e+00 -3.10742259e-01 -7.36566365e-01 5.96840441e-01
-4.59521800e-01 7.92067200e-02 -5.92161059e-01 -3.89588177e-01
-5.30208528e-01 -1.08417583e+00 7.09146798e-01 2.84288734e-01
-3.81028324e-01 -8.68937254e-01 6.46264851e-01 -3.00045848e-01
1.72192678e-01 1.80089042e-01 1.12082183e+00 -1.10481763e+00
-4.95699316e-01 -1.84613094e-01 -8.89842138e-02 -2.74164021e-01
3.59879971e-01 -2.33828127e-01 -1.09685230e+00 -2.08873913e-01
-2.04248622e-01 -1.88360915e-01 1.21720612e+00 6.81175351e-01
8.64551425e-01 -1.64593965e-01 -6.74111664e-01 5.93864441e-01
1.46341777e+00 6.87267780e-01 8.80608499e-01 -1.41515940e-01
5.64466357e-01 4.81340766e-01 4.99160498e-01 3.99085492e-01
-6.37876868e-01 6.06475115e-01 1.59959748e-01 -1.88854128e-01
-3.63484584e-02 -1.39182180e-01 2.47791000e-02 1.35260478e-01
-1.59330800e-01 -7.85086527e-02 -8.95880103e-01 2.99642444e-01
-1.82583129e+00 -1.40868485e+00 3.28180224e-01 2.33253002e+00
1.34425759e+00 -6.23144880e-02 8.60084221e-02 -2.48677313e-01
6.48909152e-01 -3.07638049e-01 -9.58764791e-01 5.88252284e-02
6.00132048e-02 9.64796662e-01 2.87939996e-01 6.39840186e-01
-1.01860869e+00 1.21333313e+00 6.03334999e+00 1.12059879e+00
-1.08111131e+00 -1.25934556e-01 7.40019202e-01 -2.51070827e-01
-8.00437406e-02 1.06126070e-01 -9.39798594e-01 3.01603287e-01
1.21133113e+00 -4.97460842e-01 3.52589756e-01 8.57563317e-01
1.36939362e-01 1.44608857e-04 -1.47497892e+00 8.55631828e-01
-2.33075440e-01 -2.14293003e+00 1.26316980e-01 1.00637272e-01
8.17173362e-01 8.97844359e-02 2.55569406e-02 9.89149809e-02
3.32744539e-01 -1.37172055e+00 -2.07856581e-01 6.49121761e-01
9.14156139e-01 -9.59064066e-01 6.50603950e-01 4.91109304e-02
-7.93929994e-01 7.83005804e-02 -5.29875398e-01 5.08035794e-02
-1.84956700e-01 6.32302701e-01 -1.10916889e+00 4.03249800e-01
1.93772361e-01 1.16061652e+00 -4.31169063e-01 1.29101956e+00
-9.39057246e-02 2.40243807e-01 3.77956741e-02 -2.77896702e-01
1.96236327e-01 -8.48966017e-02 4.91189122e-01 8.53334427e-01
-9.04089306e-03 3.78757238e-01 -4.63045314e-02 1.19490457e+00
-3.93731147e-01 3.32873166e-01 -9.43669140e-01 -5.87913096e-01
4.42156404e-01 1.01278448e+00 -7.11080968e-01 -5.88300467e-01
-1.48799405e-01 8.86420906e-01 3.76138330e-01 6.39939249e-01
-6.18057549e-01 -8.00329685e-01 6.14789426e-01 1.09693401e-01
4.03208673e-01 4.24889892e-01 -6.24331720e-02 -8.43945384e-01
-6.59250259e-01 -9.79005396e-01 1.46360308e-01 -4.84745145e-01
-1.31944191e+00 4.93062288e-01 -4.20719892e-01 -9.80897903e-01
1.22120544e-01 -5.73883176e-01 -5.76443493e-01 1.02570629e+00
-1.22614968e+00 -5.35966396e-01 9.66561884e-02 5.46932101e-01
5.49852550e-01 -2.33327180e-01 1.26380563e+00 3.45069349e-01
-6.02413416e-01 2.50761956e-01 4.51475888e-01 -1.99410036e-01
9.52036202e-01 -1.05584085e+00 3.54582548e-01 3.24929208e-01
2.41980225e-01 1.07947969e+00 7.33586490e-01 -1.03021741e+00
-1.17821681e+00 -1.07207453e+00 5.54742098e-01 -4.35634434e-01
6.22626960e-01 -2.26601198e-01 -1.12572014e+00 1.87579319e-01
-4.33339961e-02 8.53139758e-02 1.27829111e+00 2.09991619e-01
-2.56170541e-01 2.01433867e-01 -1.01847708e+00 6.72242999e-01
8.17563534e-01 -4.51927960e-01 -5.21299601e-01 6.01661801e-01
9.02585387e-01 2.17787400e-02 -9.20535564e-01 4.19042885e-01
4.88862127e-01 -4.33208823e-01 1.06375611e+00 -1.18233800e+00
4.31307703e-01 -2.95342088e-01 -2.74454034e-03 -1.55774903e+00
-7.17634499e-01 -7.20729649e-01 -3.22810203e-01 5.11938691e-01
6.78975046e-01 -4.08688098e-01 8.15433145e-01 5.57115555e-01
1.18388228e-01 -1.13861096e+00 -6.82313681e-01 -7.81961739e-01
2.13182077e-01 2.34394193e-01 4.40154046e-01 1.14514959e+00
5.18464267e-01 1.05405867e+00 -2.71801144e-01 -8.91899318e-02
5.12218595e-01 1.81150749e-01 2.55963683e-01 -1.39590824e+00
-6.90125406e-01 -4.16675091e-01 -2.54756242e-01 -4.13750201e-01
-6.55837953e-02 -1.09496927e+00 -1.35757595e-01 -1.36041784e+00
6.37477160e-01 -1.09929964e-01 -5.41723847e-01 5.63404322e-01
-2.73829520e-01 -2.22568475e-02 -8.17429349e-02 1.02106772e-01
-5.52077174e-01 6.52912319e-01 1.15921843e+00 -4.55212623e-01
-6.74162745e-01 -5.22429997e-04 -8.52707267e-01 4.32848215e-01
8.58612001e-01 -5.45376241e-01 -5.31032622e-01 4.89104390e-01
2.54181117e-01 -5.40214330e-02 1.42967492e-01 -8.58097911e-01
3.07951927e-01 -2.51466006e-01 6.89799428e-01 -1.40872046e-01
3.82032454e-01 -4.99277264e-01 3.09974164e-01 8.64826620e-01
-3.21896642e-01 -6.75667882e-01 3.29756260e-01 7.62222826e-01
-4.12853658e-02 -3.88754994e-01 1.03997910e+00 -4.13168877e-01
-6.94917917e-01 7.06847131e-01 -1.83845371e-01 -5.38922139e-02
1.04784918e+00 -5.39499894e-02 -2.86981553e-01 -9.20487195e-02
-9.79782462e-01 -3.30267958e-02 4.09857303e-01 6.38046265e-02
7.88862109e-01 -1.14447176e+00 -4.01427031e-01 -4.79731057e-03
2.89076567e-01 -2.98894078e-01 9.11991596e-02 8.18890989e-01
-1.26302704e-01 5.94382882e-01 -1.79513097e-01 -1.09372415e-01
-1.23133540e+00 1.18849313e+00 4.73923981e-01 -2.01755881e-01
-4.61181313e-01 1.08176589e+00 5.10438561e-01 1.50360078e-01
2.81098425e-01 1.61365971e-01 -2.73155868e-01 1.70392111e-01
8.90036464e-01 1.35086805e-01 -2.67780498e-02 -1.23202480e-01
-4.27228570e-01 4.59425032e-01 -5.28821111e-01 1.83727756e-01
1.70311582e+00 5.80226123e-01 1.35237068e-01 3.68326932e-01
1.09350741e+00 -1.87515274e-01 -9.51243401e-01 -2.39033520e-01
1.48591012e-01 -2.74989396e-01 1.25866145e-01 -1.00108683e+00
-5.05709946e-01 8.61789227e-01 7.39688516e-01 -4.98337120e-01
7.58241713e-01 1.29488185e-01 3.55047584e-01 8.54226470e-01
1.86272398e-01 -8.63047302e-01 2.94758320e-01 5.24539411e-01
7.96888709e-01 -1.41434550e+00 1.39081493e-01 -3.53930533e-01
-4.24386293e-01 9.57721293e-01 2.93333858e-01 -8.69354680e-02
7.04250216e-01 5.98168187e-02 -5.83338499e-01 -7.94371247e-01
-7.79751062e-01 -3.56621265e-01 2.39654228e-01 4.93915349e-01
5.82613885e-01 -2.82502532e-01 -1.58705011e-01 6.27247036e-01
6.29722297e-01 3.50285441e-01 1.98656797e-01 8.75385940e-01
-6.06806755e-01 -1.26090240e+00 3.30377631e-02 3.70498776e-01
-4.62321430e-01 -8.72065306e-01 -7.34748363e-01 3.78990352e-01
9.51856449e-02 5.84297836e-01 -2.82946020e-01 -2.48427212e-01
9.90120694e-02 3.59355241e-01 5.42664230e-01 -1.08893764e+00
-5.01031399e-01 2.52294302e-01 -1.41552880e-01 -4.63969857e-01
-1.44377947e-01 -1.31005973e-01 -1.34071350e+00 -1.41343519e-01
-6.26392066e-01 5.18112302e-01 2.92036146e-01 7.39899337e-01
7.77482927e-01 6.48576379e-01 4.27738309e-01 -7.79987991e-01
-3.23608547e-01 -7.81554520e-01 -7.53810942e-01 2.37258092e-01
3.60043854e-01 -7.84154177e-01 -3.07289734e-02 1.71666201e-02] | [5.116528034210205, 5.784980297088623] |
8f2c5091-a021-447b-95a6-71c95725542d | generalizing-graph-neural-networks-beyond | 2006.11468 | null | https://arxiv.org/abs/2006.11468v2 | https://arxiv.org/pdf/2006.11468v2.pdf | Beyond Homophily in Graph Neural Networks: Current Limitations and Effective Designs | We investigate the representation power of graph neural networks in the semi-supervised node classification task under heterophily or low homophily, i.e., in networks where connected nodes may have different class labels and dissimilar features. Many popular GNNs fail to generalize to this setting, and are even outperformed by models that ignore the graph structure (e.g., multilayer perceptrons). Motivated by this limitation, we identify a set of key designs -- ego- and neighbor-embedding separation, higher-order neighborhoods, and combination of intermediate representations -- that boost learning from the graph structure under heterophily. We combine them into a graph neural network, H2GCN, which we use as the base method to empirically evaluate the effectiveness of the identified designs. Going beyond the traditional benchmarks with strong homophily, our empirical analysis shows that the identified designs increase the accuracy of GNNs by up to 40% and 27% over models without them on synthetic and real networks with heterophily, respectively, and yield competitive performance under homophily. | ['Mark Heimann', 'Jiong Zhu', 'Danai Koutra', 'Yujun Yan', 'Leman Akoglu', 'Lingxiao Zhao'] | 2020-06-20 | null | http://proceedings.neurips.cc/paper/2020/hash/58ae23d878a47004366189884c2f8440-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/58ae23d878a47004366189884c2f8440-Paper.pdf | neurips-2020-12 | ['node-classification-on-non-homophilic'] | ['graphs'] | [-1.89551301e-02 5.43344975e-01 -5.40357590e-01 -2.41570890e-01
3.76936138e-01 -5.24456978e-01 6.43267155e-01 2.97473937e-01
6.31574914e-02 4.73082095e-01 7.32460320e-02 -5.66281080e-01
-3.79018486e-01 -1.32672274e+00 -7.45723665e-01 -5.84161460e-01
-6.36026978e-01 5.88846922e-01 7.37711191e-02 -2.66356319e-01
-1.86798036e-01 4.98824358e-01 -1.23852527e+00 -3.04599218e-02
7.13906705e-01 6.71790361e-01 -3.80196244e-01 5.06495416e-01
1.21646002e-01 8.92867386e-01 -1.79639474e-01 -4.67615217e-01
5.45856833e-01 -3.84234875e-01 -7.54746258e-01 -2.25611210e-01
3.81807327e-01 -3.53837386e-02 -9.75663006e-01 1.11638892e+00
4.54148561e-01 -2.78412886e-02 8.12902689e-01 -1.57079041e+00
-9.61650431e-01 1.19134355e+00 -4.81706262e-01 9.20830816e-02
1.93745315e-01 4.93477359e-02 1.51022971e+00 -5.87307692e-01
9.35640275e-01 1.47044003e+00 1.16114271e+00 3.42236996e-01
-1.56198347e+00 -6.46469772e-01 2.16819629e-01 -2.88835634e-02
-1.42197311e+00 -2.40152314e-01 8.52701068e-01 -1.52653202e-01
9.46080804e-01 7.24686310e-02 5.89012504e-01 1.36959136e+00
1.27056375e-01 5.76913416e-01 8.88692379e-01 -1.17574506e-01
1.10821359e-01 1.24645710e-01 6.19161487e-01 8.71311009e-01
8.48046780e-01 1.45656019e-01 -1.37693673e-01 -1.97973922e-01
6.44278646e-01 3.77793074e-01 -2.88112015e-01 -6.43376350e-01
-9.10265803e-01 9.51946855e-01 1.16985452e+00 1.75029635e-01
-4.37918395e-01 1.57990009e-01 4.66012955e-01 8.70215535e-01
4.70738202e-01 8.12699914e-01 -2.51300074e-02 3.66999716e-01
-5.71859598e-01 -2.54297942e-01 1.25672305e+00 9.19781506e-01
1.07813752e+00 1.16811767e-01 1.17393099e-01 6.94659233e-01
9.87040624e-02 1.51763812e-01 3.85225356e-01 -4.74965781e-01
4.55151886e-01 1.22912490e+00 -5.51490545e-01 -1.52114320e+00
-6.96776509e-01 -9.64227498e-01 -1.32162774e+00 -2.58826375e-01
2.12238088e-01 -2.93886244e-01 -8.09601128e-01 1.85988045e+00
1.30538866e-01 4.50666398e-01 2.01069176e-01 5.00336051e-01
1.25426257e+00 4.39752400e-01 -2.47609839e-01 1.83798745e-01
8.67933631e-01 -1.28612435e+00 -4.29188877e-01 -1.33708045e-01
1.12630749e+00 1.45260110e-01 8.76827538e-01 -2.51388252e-01
-8.39933395e-01 -1.93583012e-01 -1.18094325e+00 3.45253944e-01
-7.83321798e-01 -3.82431358e-01 7.99571097e-01 7.50309825e-01
-1.43169224e+00 1.16713750e+00 -6.33783102e-01 -5.57045519e-01
4.54899698e-01 3.75010341e-01 -6.59576416e-01 -2.30733216e-01
-1.45367277e+00 5.92272758e-01 4.64430630e-01 -4.32685390e-02
-7.82864153e-01 -7.10017920e-01 -9.82462168e-01 4.60065782e-01
2.16279447e-01 -6.24002695e-01 4.76476073e-01 -1.07587194e+00
-9.34761405e-01 7.29095995e-01 2.41557211e-01 -7.94755638e-01
3.87065142e-01 2.21198618e-01 -3.90076131e-01 2.74593741e-01
-6.98526204e-02 6.83082700e-01 3.70831221e-01 -1.29333055e+00
-1.14237085e-01 -2.46552467e-01 3.20482880e-01 1.69539168e-01
-9.28506792e-01 -4.83901858e-01 -2.17842311e-01 -2.95302004e-01
2.15887517e-01 -1.19494390e+00 -2.43998304e-01 -2.68173605e-01
-9.57502067e-01 -2.52819687e-01 8.44309390e-01 -1.78520009e-01
1.18243504e+00 -1.85555005e+00 1.08857438e-01 7.85165370e-01
9.43853259e-01 9.22598094e-02 -4.30163056e-01 6.68293893e-01
-3.97592664e-01 4.44690824e-01 1.75530866e-01 -3.08681369e-01
-4.31559794e-02 3.69559705e-01 1.15404420e-01 5.08067131e-01
8.02585110e-02 1.26310515e+00 -1.04784274e+00 -2.41390139e-01
-7.02786371e-02 3.72897565e-01 -4.77299750e-01 -1.18933562e-02
2.20107377e-01 -1.96236745e-01 -3.05014402e-01 6.69868290e-01
5.04932046e-01 -1.07941830e+00 5.88986158e-01 1.16243223e-02
6.61672711e-01 3.56358826e-01 -1.01103997e+00 1.07912672e+00
-2.62306064e-01 7.09079742e-01 -8.69037677e-03 -1.01995158e+00
1.11596155e+00 1.20855793e-02 2.84728974e-01 -4.23947215e-01
8.65778700e-02 1.14946462e-01 3.80123436e-01 3.08001917e-02
2.95411348e-01 3.60548824e-01 2.83013880e-01 6.16353095e-01
1.13507949e-01 6.23567641e-01 2.15074465e-01 5.82436860e-01
1.70175731e+00 -6.77502394e-01 3.09701174e-01 -5.53025544e-01
2.36176345e-02 -3.32957387e-01 5.05443394e-01 1.02783811e+00
-2.10990056e-01 4.96425152e-01 1.06475830e+00 -6.85984373e-01
-7.67583311e-01 -8.64681959e-01 1.08109117e-01 1.14039600e+00
3.23876530e-01 -6.35795891e-01 -4.94658381e-01 -1.09215868e+00
4.13967013e-01 2.84168065e-01 -1.04420197e+00 -7.11358428e-01
-2.96732813e-01 -1.03569043e+00 7.20291793e-01 5.45723438e-01
3.13004524e-01 -9.17718351e-01 5.18144146e-02 -1.24077715e-01
3.28003556e-01 -1.08971620e+00 1.84420645e-01 4.04057264e-01
-9.41752374e-01 -1.27646446e+00 -5.19776821e-01 -8.62361789e-01
9.70992565e-01 4.65445161e-01 1.36698723e+00 6.35851264e-01
1.07915521e-01 2.82201767e-01 -4.32112426e-01 3.45215410e-01
-4.93730903e-01 6.49489045e-01 1.10039622e-01 -1.10342205e-02
1.16373837e-01 -1.18863022e+00 -6.88632548e-01 5.63617766e-01
-5.25970399e-01 3.08659244e-02 7.27830470e-01 9.82376993e-01
4.04979251e-02 1.58128113e-01 6.96378529e-01 -1.61307967e+00
7.11293936e-01 -1.08217669e+00 -9.10496190e-02 2.23596886e-01
-1.07502246e+00 5.62774949e-04 9.53136504e-01 -3.96523625e-01
-2.74361372e-01 -2.38180801e-01 2.34346703e-01 -5.75403392e-01
1.20679336e-02 7.94616580e-01 -1.27727926e-01 -5.09804904e-01
9.87889349e-01 -5.57781709e-03 9.71834064e-02 -2.67832160e-01
2.91080773e-01 3.29077423e-01 2.40251809e-01 -3.05739582e-01
9.74882007e-01 5.06821930e-01 3.00959855e-01 -6.88478529e-01
-4.68394130e-01 -2.23619029e-01 -5.40516615e-01 8.25685337e-02
1.98189571e-01 -9.79541242e-01 -5.06291151e-01 1.92202732e-01
-6.74995959e-01 -4.94788080e-01 -1.61643669e-01 2.43771300e-01
-4.45364229e-02 3.98748308e-01 -9.65661883e-01 -4.10565525e-01
-3.49091917e-01 -9.90392447e-01 5.03535330e-01 1.51392639e-01
-1.27812132e-01 -1.52142596e+00 -1.14523977e-01 -8.11585039e-02
4.89067882e-01 5.22924304e-01 1.27127564e+00 -1.26142061e+00
-3.22833568e-01 -3.95362258e-01 -5.71810484e-01 1.79616123e-01
1.53090060e-01 -5.23366705e-02 -8.72262120e-01 -6.00211263e-01
-8.16150904e-01 -3.19211334e-01 1.19603372e+00 2.61931837e-01
8.14879954e-01 -4.40280944e-01 -7.16713607e-01 9.27843511e-01
1.24848139e+00 -3.02203774e-01 5.52403629e-01 1.77806690e-01
1.07947433e+00 6.36052072e-01 -3.19776177e-01 9.25589055e-02
4.51596528e-01 2.50915676e-01 8.67078781e-01 -3.26949656e-01
-2.91822076e-01 -6.41268134e-01 1.17778718e-01 8.68491113e-01
-5.80607727e-02 -5.88378489e-01 -1.14920104e+00 4.98741388e-01
-1.84346342e+00 -5.18233180e-01 -9.39870998e-02 1.76834559e+00
3.16989392e-01 3.67614508e-01 2.23843873e-01 1.89855054e-01
1.09340823e+00 4.68652606e-01 -6.61693931e-01 -1.14472322e-01
-3.64007235e-01 -2.45261818e-01 6.86128557e-01 2.65330821e-01
-1.08814430e+00 1.01930857e+00 6.46335983e+00 5.43789566e-01
-9.69532073e-01 -2.28005543e-01 8.33595216e-01 2.92431831e-01
-5.34879923e-01 1.38849571e-01 -5.87279260e-01 2.09388763e-01
1.09362507e+00 -1.94883980e-02 4.03307438e-01 1.12709963e+00
-3.54318887e-01 7.02584624e-01 -1.43836212e+00 6.62549198e-01
-1.18203759e-02 -1.56246293e+00 2.76040673e-01 2.30009124e-01
9.93961453e-01 3.27045709e-01 5.38752601e-02 4.20244128e-01
9.43507016e-01 -1.27764046e+00 -3.01715620e-02 4.82527539e-03
6.21370137e-01 -5.39947629e-01 8.58464658e-01 1.27096415e-01
-1.16868293e+00 -3.01537931e-01 -4.18984890e-01 -2.62690455e-01
-3.47115189e-01 4.97246623e-01 -8.91525149e-01 7.14271665e-01
5.44316113e-01 1.29026580e+00 -8.86436343e-01 8.25200737e-01
-2.36577660e-01 8.60707343e-01 -3.50497782e-01 -2.41815031e-01
2.86683828e-01 -6.09302334e-02 4.86107975e-01 1.12088895e+00
6.96656033e-02 -4.30028945e-01 1.65236726e-01 8.91810000e-01
-6.39244437e-01 -1.76920295e-02 -1.09835255e+00 -3.25418353e-01
7.48944700e-01 1.38763726e+00 -1.08191025e+00 -1.47776172e-01
-2.83587128e-01 6.43074155e-01 6.58318639e-01 6.28876567e-01
-4.59771305e-01 -6.33240283e-01 4.98898983e-01 4.47177857e-01
8.14086422e-02 1.05488740e-01 -1.71882108e-01 -1.00184095e+00
-3.12674344e-01 -9.21020806e-01 5.57988107e-01 -4.19958144e-01
-1.62153137e+00 8.87875676e-01 -2.63889492e-01 -8.25433433e-01
-2.74893731e-01 -6.00965083e-01 -1.06287003e+00 4.26593751e-01
-1.53428626e+00 -1.18052197e+00 -4.51772511e-01 4.70373362e-01
-2.45597973e-01 -3.29002976e-01 6.78614259e-01 3.05183381e-01
-8.56286168e-01 9.36286449e-01 2.73686558e-01 6.75649762e-01
2.86776006e-01 -1.32565284e+00 8.03795576e-01 5.98494291e-01
3.12211633e-01 8.62270236e-01 4.14062560e-01 -7.17657268e-01
-1.54648006e+00 -1.36283004e+00 6.79713786e-01 -1.48389220e-01
9.09792304e-01 -7.21400201e-01 -8.86114299e-01 9.09146488e-01
-1.17583320e-01 5.73643684e-01 4.63827819e-01 6.37343168e-01
-7.80169427e-01 -1.90464601e-01 -1.00760448e+00 8.80827844e-01
1.68661797e+00 -6.27354145e-01 -1.27076238e-01 5.47989130e-01
9.58554149e-01 -2.56765466e-02 -1.02619958e+00 5.98157585e-01
3.87873799e-01 -1.02127278e+00 1.11803067e+00 -1.02589488e+00
5.42077780e-01 2.20177680e-01 -3.77385728e-02 -1.59235775e+00
-5.79872489e-01 -7.74096131e-01 -2.28390262e-01 1.00491011e+00
7.45980203e-01 -1.16772044e+00 1.17235446e+00 2.05175936e-01
4.76394594e-02 -8.16164911e-01 -6.89227998e-01 -9.32430625e-01
1.50805460e-02 1.80341303e-01 7.58340418e-01 1.47896242e+00
9.61646885e-02 6.62886143e-01 -2.50082791e-01 2.07912829e-02
4.57080901e-01 3.26800942e-01 8.16530049e-01 -1.74943328e+00
-1.43541172e-01 -6.27341568e-01 -8.42532575e-01 -9.09255624e-01
6.85802341e-01 -1.38771963e+00 -4.78412181e-01 -1.47734320e+00
2.63036340e-01 -6.06279671e-01 -5.58669508e-01 6.04478836e-01
-4.29842547e-02 2.34440327e-01 -6.97361007e-02 2.64659256e-01
-7.37072051e-01 5.23602784e-01 1.09292114e+00 -2.24734709e-01
-2.55511522e-01 3.07539925e-02 -9.50108111e-01 6.68397009e-01
6.65914536e-01 -4.24447894e-01 -6.08776748e-01 -8.64430368e-02
5.94288886e-01 -2.27056235e-01 4.42052811e-01 -1.04583681e+00
2.21849784e-01 2.69902498e-01 2.41904363e-01 -6.52725101e-02
1.34798348e-01 -7.82210410e-01 4.66181710e-02 6.28129303e-01
-6.48461223e-01 3.61183584e-01 -3.23723018e-01 9.62304294e-01
6.20036945e-03 1.42528370e-01 4.40342009e-01 -3.54542099e-02
-4.38692659e-01 4.11730230e-01 9.80071127e-02 1.30584940e-01
8.19867909e-01 -3.72462481e-01 -1.04545307e+00 -7.95311570e-01
-6.39499366e-01 3.91533285e-01 5.37743211e-01 3.41655672e-01
4.89924312e-01 -1.43302965e+00 -6.52461767e-01 2.77170241e-01
1.25051007e-01 -2.79368639e-01 3.22258323e-02 8.84845078e-01
-4.59796429e-01 7.92983100e-02 -3.14562678e-01 -6.78064942e-01
-1.04664946e+00 4.79996234e-01 5.18393695e-01 -4.14441586e-01
-9.47656333e-01 6.21294796e-01 3.65082711e-01 -8.62145185e-01
3.61659884e-01 -1.30530506e-01 -2.43421465e-01 -9.13499668e-02
-9.77293551e-02 4.45182920e-01 -3.26789543e-02 -5.61315596e-01
-2.38954782e-01 8.73187184e-02 -1.46001264e-01 6.77627921e-01
1.39617586e+00 1.00805342e-01 -1.25650480e-01 3.20128143e-01
1.40482593e+00 -3.14276457e-01 -9.03357804e-01 -4.40942049e-01
9.41621289e-02 -1.22173227e-01 -1.28523529e-01 -3.25582266e-01
-1.35804772e+00 8.37158442e-01 1.51050925e-01 8.33349824e-01
6.51769280e-01 -8.71477835e-03 5.17939746e-01 9.30916190e-01
1.37222275e-01 -7.61198521e-01 7.17211664e-02 6.63966715e-01
5.31816542e-01 -1.06877279e+00 5.71834557e-02 -6.90995038e-01
-2.77365834e-01 9.43532646e-01 7.46002555e-01 -4.46972042e-01
9.27576303e-01 -1.12476153e-02 -3.00626338e-01 -4.41066116e-01
-1.03928363e+00 -6.93865418e-02 8.74793231e-02 7.16209292e-01
9.40813720e-02 2.42671266e-01 2.48663321e-01 3.59133035e-01
-2.16854781e-01 -6.24852479e-01 5.53906918e-01 5.43068469e-01
-2.25847170e-01 -5.42372644e-01 3.15477848e-01 8.94265711e-01
-2.03885082e-02 -1.78775206e-01 -9.55086589e-01 1.04821289e+00
-4.11371648e-01 8.17665815e-01 9.80404988e-02 -9.84527171e-01
7.26069510e-02 -1.51946738e-01 -1.09266378e-01 -5.56199610e-01
-7.99498081e-01 -5.12706041e-01 3.80168408e-01 -4.83008385e-01
-9.96168107e-02 -2.93029789e-02 -9.27143216e-01 -6.83506131e-01
-4.36349392e-01 2.72453845e-01 1.91818610e-01 5.98466218e-01
6.48070991e-01 5.22390246e-01 8.03880751e-01 -6.89364314e-01
-6.95965886e-01 -1.02294946e+00 -8.27804208e-01 6.42403662e-01
1.83799222e-01 -7.30933249e-01 -9.31100547e-01 -8.50081503e-01] | [6.969155788421631, 6.172756195068359] |
5a50d6f7-334a-4e0c-9164-3ac0eced06b5 | vos-gan-adversarial-learning-of-visual | 1803.09092 | null | https://arxiv.org/abs/1803.09092v2 | https://arxiv.org/pdf/1803.09092v2.pdf | Adversarial Framework for Unsupervised Learning of Motion Dynamics in Videos | Human behavior understanding in videos is a complex, still unsolved problem and requires to accurately model motion at both the local (pixel-wise dense prediction) and global (aggregation of motion cues) levels. Current approaches based on supervised learning require large amounts of annotated data, whose scarce availability is one of the main limiting factors to the development of general solutions. Unsupervised learning can instead leverage the vast amount of videos available on the web and it is a promising solution for overcoming the existing limitations. In this paper, we propose an adversarial GAN-based framework that learns video representations and dynamics through a self-supervision mechanism in order to perform dense and global prediction in videos. Our approach synthesizes videos by 1) factorizing the process into the generation of static visual content and motion, 2) learning a suitable representation of a motion latent space in order to enforce spatio-temporal coherency of object trajectories, and 3) incorporating motion estimation and pixel-wise dense prediction into the training procedure. Self-supervision is enforced by using motion masks produced by the generator, as a co-product of its generation process, to supervise the discriminator network in performing dense prediction. Performance evaluation, carried out on standard benchmarks, shows that our approach is able to learn, in an unsupervised way, both local and global video dynamics. The learned representations, then, support the training of video object segmentation methods with sensibly less (about 50%) annotations, giving performance comparable to the state of the art. Furthermore, the proposed method achieves promising performance in generating realistic videos, outperforming state-of-the-art approaches especially on motion-related metrics. | ["P. D'Oro", 'D. Giordano', 'C. Spampinato', 'S. Palazzo', 'M. Shah'] | 2018-03-24 | null | null | null | null | ['unsupervised-video-object-segmentation'] | ['computer-vision'] | [ 3.39691758e-01 5.32030035e-03 -3.29083085e-01 -6.62923977e-02
-6.07233822e-01 -5.76125145e-01 9.11581337e-01 -1.63619965e-01
-4.17404741e-01 7.13568211e-01 6.87864050e-02 1.01169311e-01
4.60242964e-02 -8.18232775e-01 -1.01027250e+00 -9.12329912e-01
-4.93438989e-02 4.60400701e-01 5.47656894e-01 5.29595166e-02
1.72846615e-02 4.30035949e-01 -1.71890187e+00 3.65497231e-01
8.15215707e-01 9.93245125e-01 3.15189302e-01 8.12116861e-01
1.03517227e-01 1.12740207e+00 -2.85547405e-01 -4.34085205e-02
2.23455772e-01 -8.30175698e-01 -8.25580120e-01 5.05142510e-01
4.83555108e-01 -3.36251736e-01 -2.28539005e-01 8.07474852e-01
-2.08248291e-02 3.42517436e-01 7.39233792e-01 -1.08374083e+00
-1.88830167e-01 3.13296914e-01 -3.46583128e-01 1.88141525e-01
3.94442439e-01 6.27648056e-01 7.97953188e-01 -5.18904388e-01
1.09855151e+00 9.33648348e-01 3.57638747e-01 8.18694353e-01
-1.48348069e+00 -3.15944254e-01 2.12159455e-01 3.64850491e-01
-1.23235846e+00 -3.59527230e-01 8.65665913e-01 -8.27750683e-01
6.58549607e-01 1.52910724e-01 8.30862105e-01 1.55190313e+00
-1.13261468e-03 7.40191460e-01 1.02313685e+00 -1.96748808e-01
5.15277207e-01 1.23276941e-01 -5.58972299e-01 5.93486130e-01
-8.57027769e-02 1.54273331e-01 -3.93848985e-01 1.30343974e-01
1.01667547e+00 -9.04589891e-02 -3.64345819e-01 -5.36794722e-01
-1.11222184e+00 7.71195531e-01 4.07872260e-01 5.90420067e-01
-6.12165749e-01 3.38301390e-01 2.35471904e-01 3.69774438e-02
5.11914790e-01 1.17410451e-01 -1.69856980e-01 -2.28688583e-01
-1.60563004e+00 3.33233893e-01 6.72708213e-01 6.45366192e-01
7.99685895e-01 3.14844877e-01 -3.43562424e-01 3.59960616e-01
2.19946787e-01 4.14203465e-01 6.81506991e-01 -1.07967985e+00
4.18217361e-01 5.02396464e-01 2.53631711e-01 -1.12891245e+00
-1.04096726e-01 -2.20992535e-01 -9.01901424e-01 3.00362259e-01
6.64885640e-01 6.85238689e-02 -1.08056784e+00 1.76293349e+00
4.90597129e-01 6.88751459e-01 -2.96142101e-02 9.79543149e-01
2.19679430e-01 9.32881117e-01 2.73036152e-01 -3.35225075e-01
9.21625018e-01 -1.16596317e+00 -5.70872068e-01 -5.26146479e-02
4.62231904e-01 -4.76596504e-01 8.36263537e-01 3.82627994e-01
-1.21967995e+00 -9.09873128e-01 -7.77380168e-01 3.01086783e-01
-1.23935722e-01 1.22307181e-01 3.14883560e-01 5.15473664e-01
-1.08558106e+00 8.38662684e-01 -1.20468342e+00 -2.45071083e-01
5.01455247e-01 4.06463623e-01 -5.15718400e-01 -1.90160930e-01
-9.19092894e-01 5.91198027e-01 4.92450237e-01 1.23511776e-01
-1.28150117e+00 -5.72594345e-01 -9.69653785e-01 -9.79689974e-03
3.37003171e-01 -8.87735784e-01 7.16384530e-01 -1.44175518e+00
-1.59754801e+00 5.50810516e-01 -6.08060360e-02 -8.03771615e-01
1.08126771e+00 -1.00405671e-01 -1.32928148e-01 5.92161238e-01
5.44287963e-03 1.02208793e+00 1.29840624e+00 -1.27185845e+00
-6.29048944e-01 9.02416855e-02 -2.10232288e-02 -6.40869141e-02
-3.46334249e-01 -3.83059204e-01 -6.08708322e-01 -9.29277360e-01
-3.98302257e-01 -1.02082992e+00 -4.71653551e-01 -1.74493819e-01
-8.02761838e-02 -1.11093838e-02 1.05203915e+00 -9.29049969e-01
1.13381636e+00 -1.92952228e+00 6.07439101e-01 1.94195613e-01
-6.86671063e-02 4.19414669e-01 -1.31672055e-01 2.54378200e-01
5.84538169e-02 -6.42885640e-02 -5.13066411e-01 -6.45218909e-01
-2.35607699e-01 2.95846432e-01 -2.31705785e-01 5.44475436e-01
4.86237049e-01 8.87112498e-01 -1.10639358e+00 -3.83438587e-01
6.25762880e-01 6.62670255e-01 -6.71837389e-01 5.40350020e-01
-6.56817019e-01 1.09467828e+00 -3.91658306e-01 2.67463684e-01
4.70556766e-01 -2.42257655e-01 3.64790142e-01 -7.43327588e-02
8.52440968e-02 -6.22434961e-03 -1.17244887e+00 1.86723423e+00
-5.19845009e-01 6.25440478e-01 -1.92957804e-01 -1.24548399e+00
7.05366254e-01 4.65348393e-01 8.19132745e-01 -4.90589619e-01
8.20470154e-02 1.19816884e-01 -1.58670887e-01 -6.20524228e-01
1.51859477e-01 7.87814185e-02 1.05363600e-01 2.90912122e-01
3.93431515e-01 3.19566503e-02 4.79509950e-01 3.63532007e-02
1.17393672e+00 7.27354944e-01 7.15796351e-02 -8.67669703e-04
8.63336444e-01 -2.42239553e-02 4.99291122e-01 3.95094842e-01
1.23090096e-01 8.38318050e-01 4.78154659e-01 -4.84508157e-01
-1.23624229e+00 -7.56630361e-01 3.91796529e-01 6.14737451e-01
-1.22212023e-02 -1.91414908e-01 -1.19173956e+00 -1.05679882e+00
-3.32265675e-01 5.90068758e-01 -8.22602272e-01 -6.12298809e-02
-8.24061215e-01 -4.68133241e-01 1.68050319e-01 6.36889219e-01
3.66713464e-01 -1.39196134e+00 -7.97757864e-01 4.29808497e-01
-2.60556102e-01 -1.61995637e+00 -3.02539080e-01 -1.12028264e-01
-9.16180491e-01 -9.50184345e-01 -9.44396138e-01 -4.36594427e-01
6.99387968e-01 -3.89852980e-03 8.84868741e-01 6.66638836e-02
-3.20636362e-01 5.01576304e-01 -4.73205566e-01 3.82994205e-01
-6.51951969e-01 1.08925486e-02 -1.92039981e-01 5.94071805e-01
-1.59872994e-01 -7.61339307e-01 -8.20592761e-01 3.72961521e-01
-1.30701423e+00 1.44980550e-01 5.47297955e-01 8.29233706e-01
5.99648893e-01 -1.25004100e-02 2.90912390e-01 -9.09016013e-01
-2.18163699e-01 -6.03511512e-01 -7.89678335e-01 -4.04488854e-02
-3.64205152e-01 6.15412444e-02 7.49301672e-01 -6.08228385e-01
-1.08662403e+00 5.14227867e-01 -1.77274823e-01 -8.27340841e-01
-5.20812333e-01 4.76255491e-02 -2.36709416e-01 -7.62455687e-02
5.35205543e-01 4.25986081e-01 -4.01387289e-02 -2.34056652e-01
5.39791167e-01 3.56081687e-02 5.31184018e-01 -3.23389024e-01
1.03720152e+00 7.67145038e-01 3.61039527e-02 -7.61420012e-01
-4.90391135e-01 -4.87166137e-01 -8.56457770e-01 -5.94575405e-01
1.33124018e+00 -8.77903521e-01 -3.07050020e-01 3.68554592e-01
-1.08018053e+00 -6.80245280e-01 -5.05707324e-01 3.90641510e-01
-9.96138155e-01 2.52899498e-01 -5.17624676e-01 -7.10582852e-01
-1.01028429e-02 -1.25107729e+00 1.01424551e+00 1.22981906e-01
-2.95921504e-01 -1.24917507e+00 1.71613574e-01 6.35200739e-01
2.96351820e-01 7.12101698e-01 5.09874284e-01 -4.05202538e-01
-1.17573369e+00 -1.51939675e-01 1.60103932e-01 6.36771560e-01
3.12146917e-03 -9.88795422e-03 -9.16837931e-01 -1.65688142e-01
-6.80462942e-02 -2.17002362e-01 8.72900248e-01 3.92868936e-01
1.05272532e+00 -4.02130306e-01 -1.47986367e-01 5.63509285e-01
1.49715555e+00 -2.92073991e-02 8.77376258e-01 2.28834331e-01
9.71979082e-01 7.20843375e-01 7.35946238e-01 3.96226466e-01
1.33027449e-01 9.91689563e-01 5.68601727e-01 -5.71085587e-02
-4.27928448e-01 -3.19888115e-01 5.64974844e-01 5.38667083e-01
-5.96120119e-01 -3.00898761e-01 -6.86586976e-01 6.43178165e-01
-2.11250114e+00 -1.30843163e+00 -1.27166137e-01 2.22671723e+00
5.16933024e-01 1.11720063e-01 4.40705001e-01 2.16272488e-01
4.82343733e-01 3.01125050e-01 -2.96286434e-01 -4.81149815e-02
1.69415213e-02 2.42198318e-01 2.82665133e-01 3.95241469e-01
-1.28047717e+00 1.12456286e+00 5.04403877e+00 8.10367882e-01
-1.26632380e+00 3.42749953e-01 7.24935532e-01 -2.57165134e-02
-1.56552017e-01 -7.79474154e-02 -5.88090003e-01 7.01859474e-01
1.01460361e+00 4.00215417e-01 3.98903579e-01 8.27496231e-01
5.30715585e-01 -5.79086356e-02 -1.11603582e+00 7.73881912e-01
1.55335918e-01 -1.60700798e+00 2.10268021e-01 1.92402005e-01
1.10117888e+00 -3.17973942e-01 -2.75439844e-02 1.93499103e-02
-1.07297249e-01 -1.04415274e+00 1.01025438e+00 6.95288897e-01
4.75437909e-01 -6.51872814e-01 7.16998577e-01 5.26339412e-01
-1.24259317e+00 -4.44080532e-02 -1.52105484e-02 2.75833309e-01
4.63855833e-01 4.10799354e-01 -6.05660856e-01 6.95823014e-01
6.00019515e-01 9.59809065e-01 -3.30025256e-01 7.61302650e-01
-2.54910737e-01 8.22774589e-01 -1.56991705e-01 3.86744440e-01
5.29086113e-01 -3.63577873e-01 6.18162870e-01 1.32990992e+00
2.71977276e-01 -2.80505300e-01 1.86570749e-01 9.08642828e-01
9.43758190e-02 6.43975437e-02 -5.71973026e-01 -1.41530126e-01
-1.29715085e-01 1.25326455e+00 -1.07588565e+00 -4.28229451e-01
-2.80408353e-01 1.28151429e+00 1.77049890e-01 3.92881900e-01
-9.92274880e-01 3.72119635e-01 4.04418200e-01 4.43880081e-01
7.92432070e-01 -3.63872766e-01 2.23741621e-01 -1.14903390e+00
1.32026464e-01 -8.38803947e-01 3.09220046e-01 -3.54647666e-01
-9.05157804e-01 7.71218956e-01 -6.32445095e-03 -1.48297584e+00
-9.46337700e-01 -4.58687007e-01 -6.10222578e-01 5.50427735e-01
-1.31848395e+00 -1.36307549e+00 -5.16880929e-01 7.28669703e-01
7.95426786e-01 -8.61776695e-02 6.01495624e-01 4.05481905e-01
-3.72834474e-01 2.25872383e-01 -2.67590880e-01 -4.23730444e-03
3.19250703e-01 -1.12711167e+00 9.46420729e-02 9.84012425e-01
4.60845858e-01 3.85235287e-02 7.27977693e-01 -5.32215536e-01
-1.15064812e+00 -1.32071280e+00 5.03909349e-01 -3.72095078e-01
5.44609725e-01 -4.31507915e-01 -9.99550581e-01 4.72855717e-01
1.58901379e-01 3.90424669e-01 3.80152494e-01 -7.40094781e-01
1.80147924e-02 -5.74303381e-02 -9.21625853e-01 3.43417525e-01
9.47690010e-01 -2.86575258e-01 -2.48899832e-01 2.41774350e-01
4.29815710e-01 -2.73041129e-01 -7.55345583e-01 3.54012936e-01
4.41190630e-01 -1.19878757e+00 9.48144495e-01 -5.25517225e-01
7.37203479e-01 -4.10701334e-01 6.70615882e-02 -9.97005820e-01
-4.84179631e-02 -6.96298003e-01 -5.33329487e-01 1.30227005e+00
1.15906835e-01 -8.15759599e-02 1.08461905e+00 2.29219303e-01
1.40849888e-01 -7.51429975e-01 -7.72599101e-01 -6.99271977e-01
-3.23376179e-01 -3.87520283e-01 1.23891920e-01 8.69094133e-01
-6.50939405e-01 -2.71867141e-02 -7.46385038e-01 -2.26829760e-03
4.97214913e-01 4.04173434e-02 9.47869480e-01 -7.03945935e-01
-6.99923992e-01 -3.76702845e-01 -8.26121271e-01 -9.84203458e-01
4.78085041e-01 -6.29099607e-01 8.68217424e-02 -1.37416899e+00
-9.49242935e-02 -1.47330493e-01 -1.52333662e-01 1.76052392e-01
-1.32679297e-02 5.85239947e-01 3.39265138e-01 1.91596985e-01
-5.49727201e-01 5.54430008e-01 1.23573458e+00 -1.48979887e-01
-2.50563025e-01 9.24794674e-02 8.91576484e-02 8.31881642e-01
4.74743307e-01 -4.73127782e-01 -5.07634163e-01 -1.09184846e-01
-1.46703273e-01 1.72754839e-01 6.74188673e-01 -1.39858782e+00
9.27999318e-02 -2.19155595e-01 5.75404823e-01 -2.57967323e-01
5.38863838e-01 -9.69458044e-01 3.32466185e-01 5.99148750e-01
-2.45165959e-01 -1.60409629e-01 -4.37474111e-03 8.13447535e-01
-4.62512612e-01 -1.18458003e-01 8.14268887e-01 -1.71073943e-01
-8.35469902e-01 4.78592932e-01 -4.90891159e-01 -4.76818420e-02
1.26200104e+00 -2.07955852e-01 2.25714400e-01 -4.59717900e-01
-9.51995850e-01 -1.41837701e-01 5.28011382e-01 4.07499552e-01
4.85138208e-01 -1.14698458e+00 -4.48810577e-01 7.25433454e-02
-1.91905975e-01 -3.15307230e-02 4.72081900e-01 9.32840109e-01
-6.99020207e-01 2.26945743e-01 -3.89560103e-01 -8.39431226e-01
-9.38518822e-01 6.39942706e-01 2.10314840e-01 -5.54579973e-01
-7.78683603e-01 5.91608703e-01 2.17390016e-01 1.40694663e-01
7.47234151e-02 -1.76196471e-01 -3.17279816e-01 5.08733839e-02
5.02969503e-01 2.57128596e-01 -7.24439621e-02 -8.87566328e-01
-1.15290456e-01 6.62941277e-01 2.56467909e-01 -2.00891316e-01
1.30853152e+00 3.15861069e-02 1.53581738e-01 4.04219836e-01
1.14932704e+00 -4.28879671e-02 -1.83720899e+00 1.92122966e-01
5.97915724e-02 -4.45504814e-01 -1.54696226e-01 -3.69380802e-01
-1.23750544e+00 9.67931807e-01 5.21957219e-01 1.20489240e-01
1.20900762e+00 -3.07310037e-02 8.36502433e-01 -2.31570870e-01
4.47982043e-01 -8.92277062e-01 4.25371736e-01 1.96015518e-02
6.48204446e-01 -1.08247352e+00 -2.22249568e-01 -5.00526965e-01
-7.40210116e-01 1.09798491e+00 6.13469660e-01 -4.53824490e-01
3.35796416e-01 4.98772599e-02 -8.70828778e-02 2.11118788e-01
-7.55901158e-01 -2.27399155e-01 6.70943022e-01 4.75007027e-01
2.71679521e-01 -2.22073883e-01 -1.75889432e-01 2.39982694e-01
2.43105218e-01 1.79404225e-02 1.96265996e-01 6.91913247e-01
-6.53018504e-02 -1.34029245e+00 -2.52456248e-01 8.30921382e-02
-3.70457441e-01 2.54736543e-01 -2.16064230e-03 8.58145535e-01
4.63847518e-01 6.83374941e-01 1.25908941e-01 -2.19074205e-01
1.79189406e-02 6.09979546e-03 5.82816720e-01 -5.04865050e-01
-4.90650326e-01 2.35330179e-01 -1.37530327e-01 -1.07792389e+00
-8.17178130e-01 -7.71159649e-01 -1.05528796e+00 1.74120560e-01
1.34970903e-01 -4.07145843e-02 5.14344990e-01 1.12867761e+00
2.07572222e-01 6.25479460e-01 6.21861339e-01 -1.42940080e+00
-3.00255567e-01 -7.63238192e-01 -2.64428794e-01 9.32568431e-01
2.19330341e-01 -7.28904486e-01 -2.90970922e-01 7.83781648e-01] | [8.973320007324219, -0.09844426810741425] |
3c87761b-5173-4ffc-a58a-3ba8721d1caa | retrieval-oriented-masking-pre-training | 2210.15133 | null | https://arxiv.org/abs/2210.15133v1 | https://arxiv.org/pdf/2210.15133v1.pdf | Retrieval Oriented Masking Pre-training Language Model for Dense Passage Retrieval | Pre-trained language model (PTM) has been shown to yield powerful text representations for dense passage retrieval task. The Masked Language Modeling (MLM) is a major sub-task of the pre-training process. However, we found that the conventional random masking strategy tend to select a large number of tokens that have limited effect on the passage retrieval task (e,g. stop-words and punctuation). By noticing the term importance weight can provide valuable information for passage retrieval, we hereby propose alternative retrieval oriented masking (dubbed as ROM) strategy where more important tokens will have a higher probability of being masked out, to capture this straightforward yet essential information to facilitate the language model pre-training process. Notably, the proposed new token masking method will not change the architecture and learning objective of original PTM. Our experiments verify that the proposed ROM enables term importance information to help language model pre-training thus achieving better performance on multiple passage retrieval benchmarks. | ['Pengjun Xie', 'Guangwei Xu', 'Yanzhao Zhang', 'Dingkun Long'] | 2022-10-27 | null | null | null | null | ['passage-retrieval'] | ['natural-language-processing'] | [ 1.13134600e-01 -2.12126270e-01 -2.72920161e-01 7.17871636e-02
-1.17097712e+00 -3.76434475e-01 9.25622106e-01 6.07348800e-01
-7.59944320e-01 6.99909151e-01 4.67401743e-01 -3.80929619e-01
9.96098667e-02 -6.98864877e-01 -5.27735412e-01 -6.08605146e-01
-1.08909801e-01 1.63994655e-01 1.83648482e-01 -3.65131170e-01
3.35505217e-01 2.61468768e-01 -1.36610436e+00 6.21137440e-01
8.47012162e-01 6.66454673e-01 5.71744561e-01 4.09699202e-01
-4.29232001e-01 8.31732512e-01 -8.67224455e-01 -4.29776944e-02
8.09528604e-02 -3.30343157e-01 -8.58592033e-01 -3.34592015e-01
1.17499299e-01 -2.32594475e-01 -4.34439898e-01 7.55005658e-01
6.41225874e-01 6.94372237e-01 8.49333942e-01 -4.94976461e-01
-7.99858451e-01 1.04566288e+00 -5.58283865e-01 6.09783292e-01
2.91891605e-01 -1.69597507e-01 1.23573971e+00 -1.35644436e+00
3.06804806e-01 1.18791354e+00 4.29425329e-01 4.27800834e-01
-9.29995000e-01 -4.23708439e-01 3.45324159e-01 5.87606877e-02
-1.78507853e+00 -3.66276383e-01 7.29563892e-01 -6.25418276e-02
1.25954664e+00 4.97133136e-01 1.20723099e-01 8.41806889e-01
1.94348454e-01 1.13652885e+00 7.50148356e-01 -8.33721757e-01
-3.06411162e-02 4.07886207e-01 5.19410014e-01 4.50686097e-01
2.00203881e-01 -1.96810424e-01 -5.56003809e-01 -2.21772432e-01
3.40392113e-01 1.74384251e-01 -3.52976650e-01 1.82553917e-01
-1.02874923e+00 7.40796030e-01 4.88233179e-01 6.50990307e-01
-4.17444795e-01 -7.81036541e-02 4.46428746e-01 3.45387697e-01
3.31686825e-01 3.99750233e-01 -3.54142398e-01 1.01258412e-01
-1.20908904e+00 2.42104456e-01 3.00953954e-01 7.35614896e-01
8.10482800e-01 1.30496129e-01 -9.93151367e-01 9.86700773e-01
3.89382303e-01 2.56233811e-01 9.17035401e-01 -4.60064188e-02
7.50941992e-01 7.30703413e-01 1.10740721e-01 -7.64907002e-01
-3.11784744e-01 -7.50866592e-01 -8.89058411e-01 -3.10407847e-01
5.18448837e-02 1.47170380e-01 -1.13321614e+00 1.63351190e+00
-1.75604045e-01 -1.58063829e-01 1.34900436e-01 8.59479904e-01
9.15050149e-01 9.03220475e-01 4.69949186e-01 -1.90254450e-01
1.40671885e+00 -1.01850247e+00 -8.12570572e-01 -2.20204040e-01
9.01544809e-01 -1.10708487e+00 1.40110457e+00 1.00226253e-02
-1.18409503e+00 -6.91564083e-01 -9.79395449e-01 -3.18390101e-01
-5.89380503e-01 2.35758618e-01 5.55610776e-01 4.89713430e-01
-9.28525627e-01 4.33740109e-01 -5.45339823e-01 -2.10200638e-01
2.12871373e-01 2.62049913e-01 -4.00609337e-02 -7.83082321e-02
-1.61057162e+00 9.42263544e-01 5.61701775e-01 1.00412786e-01
-9.48093355e-01 -7.44215131e-01 -8.10294271e-01 5.16606390e-01
9.05931517e-02 -6.27333522e-01 1.00077105e+00 -6.14630699e-01
-1.13674295e+00 7.76965797e-01 -6.14090025e-01 -6.87853456e-01
2.91266471e-01 -4.90892917e-01 -3.40363652e-01 3.86614442e-01
-1.39468521e-01 7.32109249e-01 1.15483928e+00 -1.31534469e+00
-2.89457381e-01 1.72580823e-01 -1.16094366e-01 4.99276102e-01
-7.34033525e-01 2.00849950e-01 -7.01485097e-01 -1.04173970e+00
3.93853523e-02 -5.09884953e-01 -1.58665329e-01 -5.29648006e-01
-2.87980467e-01 -3.89447451e-01 4.81256992e-01 -6.88371420e-01
1.71220541e+00 -1.92869544e+00 -1.53006941e-01 1.51471496e-01
6.99553639e-02 5.36292076e-01 -2.88821906e-01 7.72196293e-01
9.04857442e-02 3.42259109e-01 -4.61995751e-02 -8.47699463e-01
3.23760584e-02 -8.71252269e-02 -5.70423126e-01 1.41770557e-01
1.70229733e-01 1.15210545e+00 -6.85376108e-01 -8.38327289e-01
2.07168832e-01 4.55247939e-01 -6.00130260e-01 2.53973305e-01
-2.07489535e-01 9.17418376e-02 -3.79836291e-01 5.99910319e-01
6.02603018e-01 -2.71278352e-01 -2.77626783e-01 1.69825763e-01
1.11116245e-01 8.50045145e-01 -8.56435359e-01 1.61511922e+00
-6.04964018e-01 4.97690022e-01 -2.26548851e-01 -6.48527503e-01
8.79532397e-01 5.50929844e-01 3.16564053e-01 -8.51167083e-01
1.50321484e-01 -1.28636450e-01 -4.87308986e-02 -1.79808706e-01
1.17670667e+00 -1.88421518e-01 -3.74949090e-02 5.22716165e-01
5.97855002e-02 4.70117867e-01 1.13197193e-01 5.39827883e-01
9.05142903e-01 -2.82015860e-01 1.91773087e-01 -2.21084118e-01
7.69670367e-01 -2.55670305e-02 2.62566775e-01 1.27025986e+00
7.72998575e-03 7.48835921e-01 -1.62626401e-01 1.80051014e-01
-7.89232790e-01 -9.49291348e-01 -7.00163189e-03 1.32160020e+00
-9.86239873e-03 -5.79819024e-01 -3.61819893e-01 -5.05213916e-01
-1.58970550e-01 5.97367644e-01 -3.94786358e-01 -6.65776312e-01
-6.50950611e-01 -8.89605701e-01 5.80722809e-01 4.56085294e-01
3.37863088e-01 -1.40512741e+00 -1.11389749e-01 2.53689110e-01
-3.03274781e-01 -9.07464743e-01 -6.76177502e-01 2.64962643e-01
-8.48068535e-01 -4.04197037e-01 -1.01596165e+00 -1.05338418e+00
7.99238920e-01 6.78715706e-01 1.22932172e+00 6.33579612e-01
-3.87743786e-02 3.17732960e-01 -8.14603031e-01 -1.69930011e-01
-3.55250776e-01 5.22543967e-01 -2.09221896e-02 -2.00814214e-02
6.84986889e-01 -3.38183135e-01 -7.50137269e-01 -2.54028291e-02
-1.17520356e+00 -1.02798559e-01 8.73747528e-01 9.23168480e-01
4.23911214e-01 2.11437479e-01 7.87287533e-01 -6.64824486e-01
9.69374418e-01 -4.41315562e-01 -1.90915301e-01 4.17045325e-01
-6.38034463e-01 8.38395357e-02 7.35009372e-01 -6.42394841e-01
-9.75191534e-01 -4.99475181e-01 -1.85557872e-01 -3.86958152e-01
1.13930769e-01 7.87360489e-01 5.51008172e-02 5.51630743e-02
4.64421898e-01 7.34543920e-01 -4.77466285e-01 -9.24372256e-01
2.12311029e-01 5.77515602e-01 8.38795751e-02 -6.29845977e-01
9.05489564e-01 3.40122819e-01 -2.14433208e-01 -8.25743556e-01
-6.33386493e-01 -1.02187216e+00 -5.56481302e-01 1.71549946e-01
4.65346128e-01 -1.12968981e+00 -4.46278930e-01 8.06340352e-02
-9.45980608e-01 -4.10733931e-02 -7.61072412e-02 5.30521631e-01
1.04540050e-01 6.75995946e-01 -7.37223387e-01 -1.09546673e+00
-8.71722281e-01 -7.47244418e-01 1.00798166e+00 2.04036951e-01
-2.67193496e-01 -9.40133274e-01 -1.71631992e-01 2.66270906e-01
6.47692323e-01 -5.17831802e-01 1.28994405e+00 -9.62104499e-01
-5.77385664e-01 -4.45631146e-01 -1.73068330e-01 3.49762499e-01
8.87754411e-02 -4.26423609e-01 -1.07963049e+00 -5.73037744e-01
2.87404265e-02 -2.59020180e-01 1.41701078e+00 1.63389996e-01
8.75147283e-01 -2.88972974e-01 -1.17438741e-01 1.55483484e-01
1.52442253e+00 -5.44054173e-02 7.46358335e-01 4.38267589e-01
5.99857509e-01 2.87860274e-01 6.01036191e-01 3.15627903e-01
2.05450952e-01 6.13644719e-01 -6.98627606e-02 -2.05083564e-01
-2.78123349e-01 -5.47393382e-01 6.14181280e-01 1.20099235e+00
4.49828386e-01 -3.90841395e-01 -7.38226771e-01 7.71634758e-01
-1.62935317e+00 -7.21944749e-01 1.55817822e-01 2.37199593e+00
1.13310933e+00 4.42389667e-01 -8.07630792e-02 2.48082981e-01
4.01296884e-01 2.38895908e-01 -3.25358622e-02 -2.90992200e-01
-2.90742487e-01 4.26515669e-01 3.20593804e-01 7.41136730e-01
-9.12482023e-01 1.20903575e+00 6.09585285e+00 1.26350009e+00
-8.94570827e-01 8.82696658e-02 1.65322587e-01 -1.53556630e-01
-5.40756345e-01 1.04292259e-02 -1.24084091e+00 3.15819561e-01
8.31448972e-01 -3.84139180e-01 1.98178291e-01 6.28190756e-01
4.64733928e-01 -1.96338728e-01 -8.51104558e-01 7.48880684e-01
1.44441739e-01 -1.21437514e+00 9.05773640e-01 -2.94334680e-01
5.95672488e-01 -1.27394825e-01 1.52843177e-01 8.87686729e-01
-1.46339491e-01 -1.02180970e+00 5.02393186e-01 5.05046010e-01
3.08707714e-01 -6.67228341e-01 6.51868522e-01 5.30765533e-01
-1.19014263e+00 1.29171208e-01 -7.29944944e-01 -2.25729682e-02
-3.66413663e-03 5.13655424e-01 -9.93334949e-01 5.99984586e-01
3.40641797e-01 3.94111603e-01 -8.75686228e-01 1.11079490e+00
-6.83085015e-03 8.36443007e-01 -1.77425906e-01 -2.07159191e-01
3.62251073e-01 9.37131234e-03 6.50463700e-01 1.57081831e+00
-9.23543721e-02 -2.64900833e-01 1.25592127e-01 8.29796731e-01
-3.27034324e-01 4.68498349e-01 -3.31496477e-01 -1.47486597e-01
4.39550400e-01 1.09779787e+00 -7.18074441e-01 -5.67600846e-01
-4.33632404e-01 8.95954609e-01 3.02202076e-01 6.46977484e-01
-5.35853148e-01 -3.32696050e-01 3.86208743e-01 -2.19125859e-03
4.50786412e-01 -2.71612287e-01 -3.66457343e-01 -1.27349210e+00
1.81908131e-01 -7.85132468e-01 4.56093997e-01 -5.04641235e-01
-1.02558911e+00 6.20160043e-01 -1.70550510e-01 -1.41347218e+00
-7.43666738e-02 -2.55776733e-01 -6.26111448e-01 1.35847485e+00
-2.13883018e+00 -1.09658229e+00 1.45977259e-01 6.24702156e-01
8.77715647e-01 -9.65423509e-02 8.90153646e-01 5.80357611e-01
-4.96487826e-01 9.72012401e-01 2.74705619e-01 3.48587126e-01
6.51551723e-01 -1.14410961e+00 8.12702402e-02 1.01581585e+00
3.79965872e-01 1.34594965e+00 5.36843359e-01 -6.13424838e-01
-1.32878125e+00 -1.08765697e+00 1.21616757e+00 -4.96208668e-01
4.00715768e-01 -4.75426435e-01 -1.30443656e+00 4.98730153e-01
2.46131584e-01 -4.99585927e-01 6.71729624e-01 1.44764423e-01
-1.45770252e-01 1.41742706e-01 -7.92484224e-01 9.09335077e-01
4.60621029e-01 -8.15678835e-01 -1.10720956e+00 2.60738403e-01
8.97461116e-01 -1.16013095e-01 -4.74937528e-01 3.54380667e-01
1.74911737e-01 -3.39453906e-01 1.25079834e+00 -6.79565251e-01
3.02647531e-01 -2.71501034e-01 -3.38997412e-03 -9.19688821e-01
-2.50225961e-01 -6.62374616e-01 -5.10695696e-01 1.53937411e+00
5.10570526e-01 -3.34999919e-01 5.74175239e-01 2.77157187e-01
-1.18063748e-01 -7.12931931e-01 -7.30977833e-01 -8.54641795e-01
3.25889438e-01 -2.95444787e-01 3.41772228e-01 7.63552964e-01
9.87108052e-02 3.97119969e-01 -4.60260361e-01 1.25393122e-01
2.91562229e-01 1.33236423e-01 4.80322301e-01 -8.41988325e-01
-9.96289775e-02 -5.66803217e-01 6.07157685e-02 -1.51002777e+00
3.15832570e-02 -1.29753971e+00 -2.04733144e-02 -1.54994082e+00
3.70080173e-01 -4.01125252e-01 -9.59661067e-01 3.51265043e-01
-8.43854129e-01 1.46453813e-01 9.58983377e-02 3.97072494e-01
-8.54825854e-01 8.53049576e-01 1.12768710e+00 -1.73584342e-01
-3.18188667e-01 -2.58685574e-02 -7.77164221e-01 3.21711630e-01
7.02869594e-01 -7.48700440e-01 -3.89124870e-01 -4.10813272e-01
1.30859941e-01 -1.35840982e-01 1.69258788e-01 -7.90081501e-01
2.83248037e-01 2.29528457e-01 3.89352292e-01 -9.42764699e-01
3.80249500e-01 -7.36891985e-01 -5.72921216e-01 2.71916717e-01
-6.75773203e-01 1.35594770e-01 5.06679058e-01 3.96627545e-01
-5.30696869e-01 -5.23624539e-01 3.50284845e-01 -2.74428487e-01
-7.61725187e-01 6.09928332e-02 -6.84565723e-01 3.67153659e-02
2.41074294e-01 -5.82409352e-02 -1.18160220e-02 -3.54289174e-01
-5.07046938e-01 4.27985519e-01 1.80688649e-01 6.60649836e-01
6.33281648e-01 -1.21642470e+00 -7.73618162e-01 6.91332072e-02
1.07078075e-01 -2.54880339e-01 3.16383421e-01 7.66197503e-01
-1.50880471e-01 9.09056306e-01 3.07464302e-01 -3.65106136e-01
-1.32942963e+00 8.01714242e-01 1.27514616e-01 -8.72578919e-01
-8.79394829e-01 8.04899216e-01 1.40503868e-01 -3.97905800e-03
5.07008553e-01 -4.69691664e-01 -5.58672488e-01 9.27745402e-02
8.24265778e-01 5.51354624e-02 4.26467657e-01 -4.99614924e-01
-2.47081503e-01 3.17675978e-01 -6.60331547e-01 -1.28082559e-01
1.04931700e+00 -4.56024379e-01 -3.82305048e-02 4.78482127e-01
1.31665325e+00 2.92835444e-01 -6.32926404e-01 -6.93973005e-01
3.38004798e-01 -2.68841743e-01 2.75245667e-01 -7.54589081e-01
-6.05357587e-01 9.84948397e-01 2.82941818e-01 7.16199651e-02
1.16300952e+00 -2.89134830e-01 1.07286525e+00 6.34263635e-01
2.61302203e-01 -1.00857127e+00 5.49332686e-02 6.31207108e-01
7.88235366e-01 -1.16224527e+00 5.48535474e-02 -1.37522221e-01
-5.50137341e-01 7.63998151e-01 4.49944258e-01 -2.10848767e-02
5.10398805e-01 -7.37060905e-02 2.19460592e-01 -1.06984064e-01
-8.18663597e-01 -4.21400189e-01 6.43335581e-01 1.87585026e-01
5.53191841e-01 -2.71063238e-01 -5.47439992e-01 6.80073380e-01
-2.02586912e-02 -2.59759575e-01 1.34952873e-01 1.12261486e+00
-5.73348999e-01 -1.12456191e+00 -3.88282359e-01 4.76418167e-01
-6.86416149e-01 -8.07287216e-01 -2.83812493e-01 6.28426492e-01
-3.74247044e-01 9.31761146e-01 -1.74078569e-01 -9.40260887e-02
1.03482597e-01 5.16363621e-01 1.67709708e-01 -8.61535668e-01
-1.05868983e+00 1.59158781e-01 -2.27080822e-01 -8.52821171e-02
-1.30905390e-01 -1.23950645e-01 -1.25208795e+00 -8.02075863e-02
-5.29584110e-01 5.89404881e-01 2.47226194e-01 1.12051749e+00
2.39840209e-01 6.51224375e-01 5.88545561e-01 -4.60678965e-01
-5.34313440e-01 -1.37679005e+00 -4.11480606e-01 5.28419793e-01
5.13090491e-01 -4.35659647e-01 -4.39665496e-01 -1.42285645e-01] | [11.337117195129395, 7.932877063751221] |
a615a2d9-d0c0-471a-88d1-6eb98014e398 | tracking-from-patterns-learning-corresponding | 2010.10051 | null | https://arxiv.org/abs/2010.10051v1 | https://arxiv.org/pdf/2010.10051v1.pdf | Tracking from Patterns: Learning Corresponding Patterns in Point Clouds for 3D Object Tracking | A robust 3D object tracker which continuously tracks surrounding objects and estimates their trajectories is key for self-driving vehicles. Most existing tracking methods employ a tracking-by-detection strategy, which usually requires complex pair-wise similarity computation and neglects the nature of continuous object motion. In this paper, we propose to directly learn 3D object correspondences from temporal point cloud data and infer the motion information from correspondence patterns. We modify the standard 3D object detector to process two lidar frames at the same time and predict bounding box pairs for the association and motion estimation tasks. We also equip our pipeline with a simple yet effective velocity smoothing module to estimate consistent object motion. Benifiting from the learned correspondences and motion refinement, our method exceeds the existing 3D tracking methods on both the KITTI and larger scale Nuscenes dataset. | ['Shaojie Shen', 'Peiliang Li', 'Jieqi Shi'] | 2020-10-20 | null | null | null | null | ['3d-object-tracking'] | ['computer-vision'] | [-3.75580072e-01 -5.18767059e-01 -2.76179582e-01 -3.13763380e-01
-4.64350313e-01 -8.56672347e-01 8.24798405e-01 1.64064635e-02
-5.31488895e-01 3.51651073e-01 -4.17744935e-01 -2.89626658e-01
1.94895267e-01 -6.77986443e-01 -5.55296838e-01 -4.15195614e-01
2.78820582e-02 9.04218674e-01 1.24001551e+00 -2.42380351e-02
2.25827247e-01 1.22712636e+00 -1.67959368e+00 -4.00369436e-01
6.13282144e-01 1.03253722e+00 3.50682259e-01 8.16511154e-01
-3.76951247e-01 3.08618903e-01 -1.00725405e-01 -1.54236719e-01
4.38073963e-01 1.01468273e-01 -5.60013689e-02 1.22596040e-01
9.94406462e-01 -5.12311101e-01 -4.97223586e-01 8.73679399e-01
1.61660045e-01 9.40505713e-02 7.29515433e-01 -1.58350420e+00
6.43696710e-02 -3.49138826e-01 -8.48517537e-01 4.10661668e-01
1.60756573e-01 3.38483691e-01 5.48832476e-01 -1.07600856e+00
8.69609594e-01 1.35105479e+00 1.10480273e+00 5.48860371e-01
-1.09867084e+00 -1.02806914e+00 2.21597090e-01 2.39686996e-01
-1.54714632e+00 -5.67998409e-01 7.33210385e-01 -7.83464253e-01
9.37233448e-01 7.25553334e-02 8.47626865e-01 6.22393966e-01
3.11431084e-02 6.17329597e-01 7.09054947e-01 1.23141361e-02
-1.90235972e-02 -3.96742150e-02 9.53739658e-02 6.21431351e-01
5.22258162e-01 5.05826950e-01 -3.77896816e-01 -1.44012114e-02
6.61559343e-01 1.80027351e-01 2.37245634e-01 -9.77206886e-01
-1.33749807e+00 5.87442815e-01 2.32090220e-01 -9.59701911e-02
-1.13949120e-01 5.09449124e-01 1.91346750e-01 -1.86385557e-01
5.20484507e-01 -3.76325190e-01 -3.40639740e-01 -2.86724195e-02
-1.18891358e+00 6.38571024e-01 4.20251817e-01 1.52544904e+00
9.90187883e-01 -5.20941429e-02 -1.71138018e-01 2.55618896e-02
5.91036320e-01 1.08195770e+00 -2.74879128e-01 -1.15071785e+00
3.23690623e-01 5.17816305e-01 5.41827679e-01 -9.29066837e-01
-3.65682691e-01 3.96528430e-02 -2.79033095e-01 7.28275359e-01
4.44775134e-01 1.04113623e-01 -8.81945670e-01 1.20791042e+00
1.04987860e+00 8.23215902e-01 -2.87556708e-01 8.34981799e-01
7.44110286e-01 3.35920632e-01 1.79861747e-02 -5.55502586e-02
1.11353076e+00 -6.81700528e-01 -7.08758414e-01 -2.15936631e-01
6.48174167e-01 -6.30741417e-01 1.89569220e-01 -3.62497985e-01
-9.88623440e-01 -8.12492669e-01 -9.64791894e-01 -1.84270397e-01
-2.83328354e-01 6.38256446e-02 3.65515798e-01 4.15483922e-01
-8.86843204e-01 5.06025434e-01 -1.29862463e+00 -3.68062109e-01
5.91532946e-01 3.80625337e-01 -3.95498246e-01 1.95899546e-01
-5.81045330e-01 1.28493762e+00 5.00029802e-01 9.05761123e-02
-6.50145829e-01 -8.85770857e-01 -9.68651712e-01 -5.82900763e-01
1.75056383e-01 -7.47390628e-01 1.15254009e+00 -4.62489687e-02
-1.21795881e+00 1.09242225e+00 -7.00913131e-01 -5.03029883e-01
8.69377434e-01 -4.51197922e-01 -3.38431656e-01 -1.61663339e-01
2.32046604e-01 9.67237294e-01 9.33369458e-01 -1.16444147e+00
-1.18829441e+00 -4.37255412e-01 -6.19726062e-01 5.64355142e-02
3.05237502e-01 -2.44735181e-02 -6.69566393e-01 -2.37631008e-01
3.48100364e-01 -1.12359822e+00 -2.32728437e-01 7.25382745e-01
4.99437451e-02 -4.43107694e-01 1.58592558e+00 -2.17734441e-01
7.68860757e-01 -2.14008975e+00 -8.35552812e-02 8.44290257e-02
3.81705791e-01 2.07141146e-01 3.53032649e-02 -3.86650898e-02
5.09689927e-01 -4.00515854e-01 -2.46221013e-02 -5.87006927e-01
9.77793857e-02 1.61505327e-01 -3.09506327e-01 8.72530758e-01
4.30401802e-01 1.07870901e+00 -1.06133091e+00 -7.22613811e-01
6.36806607e-01 5.96993804e-01 -3.49695325e-01 1.23124480e-01
-3.29018891e-01 5.75929046e-01 -5.60793757e-01 6.52364910e-01
1.05234087e+00 1.57137454e-01 -4.23671573e-01 -2.07499325e-01
-7.38422096e-01 1.45415828e-01 -1.31978524e+00 1.68553841e+00
8.84868428e-02 9.51571524e-01 7.19989613e-02 -4.58307326e-01
1.08718872e+00 8.33282024e-02 7.63003826e-01 -2.89449722e-01
1.98325694e-01 2.60816127e-01 -3.89404714e-01 -2.72951514e-01
6.76855326e-01 -9.40545872e-02 2.12327823e-01 6.40824065e-02
-2.44862020e-01 -4.99588579e-01 8.05424806e-03 9.38491374e-02
9.48225617e-01 6.76231503e-01 1.77717686e-01 -7.86167085e-02
4.43363011e-01 6.31666660e-01 6.59099817e-01 6.28164113e-01
-5.64981103e-01 3.70312005e-01 -3.87471408e-01 -4.16810274e-01
-1.21519578e+00 -1.30688298e+00 -3.83713216e-01 4.58479464e-01
7.39920557e-01 -2.68845856e-01 -7.91289806e-02 -7.08127379e-01
7.41863012e-01 4.46164042e-01 -3.36315989e-01 7.39054009e-02
-7.85043836e-01 1.26753375e-02 2.76053339e-01 7.93670535e-01
3.33426476e-01 -5.28826952e-01 -9.01544213e-01 2.05159307e-01
2.60444194e-01 -1.53930187e+00 -7.23831654e-01 -1.48054317e-01
-1.04391646e+00 -9.08519626e-01 -2.22659007e-01 -5.78912556e-01
3.79937649e-01 8.74163806e-01 7.92464375e-01 7.20902607e-02
-2.22107023e-01 2.60831028e-01 -3.64891291e-02 -6.97846234e-01
-3.10017467e-01 -2.66971171e-01 4.03338969e-01 -2.79821217e-01
8.99279654e-01 -3.71074021e-01 -4.30300564e-01 4.93723303e-01
3.37605625e-02 1.61858797e-01 3.09289575e-01 1.54595271e-01
8.81078899e-01 -1.24636605e-01 4.50698007e-03 -2.38746509e-01
-3.49091321e-01 -2.55962938e-01 -1.23513234e+00 -1.79826587e-01
-2.40197882e-01 -2.22800583e-01 -2.29209617e-01 -6.46713614e-01
-7.83185840e-01 7.91424930e-01 6.32611811e-02 -1.06989467e+00
-4.21622396e-01 -3.03581446e-01 -4.37995195e-02 -3.25720251e-01
3.03791910e-01 -1.09170131e-01 -6.94925860e-02 -4.35472250e-01
5.42429984e-01 3.08256775e-01 7.54332423e-01 -1.97392657e-01
1.68373573e+00 1.07016265e+00 2.98865765e-01 -7.29128897e-01
-7.72502959e-01 -9.29337859e-01 -1.41089237e+00 -5.56153178e-01
1.21282780e+00 -1.18531621e+00 -9.81501639e-01 2.08326682e-01
-1.48207843e+00 -1.51535735e-01 -2.50741154e-01 8.50809753e-01
-4.81806010e-01 2.44225264e-01 -2.98360288e-01 -9.55689192e-01
-5.37787052e-03 -8.11750591e-01 1.56171095e+00 1.83371425e-01
-1.97049573e-01 -8.04886043e-01 4.01550353e-01 -4.40193899e-02
5.10357097e-02 4.93318319e-01 7.40119368e-02 -3.43103223e-02
-1.10677183e+00 -4.33547795e-01 -4.18019027e-01 -3.71190101e-01
8.06710273e-02 2.70646483e-01 -9.60257471e-01 -2.06401363e-01
-2.37983316e-01 3.65004212e-01 8.16816568e-01 5.03478289e-01
6.10878170e-01 3.90108496e-01 -9.42148030e-01 7.74408281e-01
1.12915754e+00 9.37690213e-02 2.08633453e-01 3.08016866e-01
1.01407325e+00 5.90304732e-01 1.05869591e+00 1.23987958e-01
4.61596280e-01 1.03054190e+00 3.59793991e-01 4.01527099e-02
-3.85855168e-01 -4.21346217e-01 3.74908090e-01 5.90256751e-01
-1.11136630e-01 4.18466866e-01 -9.25417781e-01 6.17104888e-01
-2.02312732e+00 -1.17530823e+00 -7.60740995e-01 2.03172541e+00
5.09497702e-01 5.08414507e-01 3.32270473e-01 -2.44451851e-01
7.93644190e-01 -1.36254206e-01 -6.95866346e-01 3.70770514e-01
8.84103701e-02 -5.57160378e-02 9.46930289e-01 7.73074210e-01
-1.17535901e+00 1.31396902e+00 6.09088850e+00 5.02974689e-01
-7.45556593e-01 1.91489816e-01 -4.29828644e-01 -2.02579886e-01
3.70728672e-02 2.39153191e-01 -1.73703313e+00 2.52411038e-01
6.95055485e-01 -1.58039004e-01 -2.14200675e-01 8.35520625e-01
4.37482089e-01 -3.98283228e-02 -1.16411948e+00 1.00605309e+00
-2.90975422e-01 -1.53491688e+00 -3.42374891e-01 1.31243035e-01
5.99063933e-01 3.89130235e-01 -3.38996649e-01 7.01761618e-02
3.96407634e-01 -6.06086433e-01 1.23363745e+00 4.91151661e-01
5.35641313e-01 -5.28299510e-01 2.00326204e-01 6.01473808e-01
-1.98872519e+00 2.39869565e-01 -3.28354359e-01 -1.95264012e-01
6.26818776e-01 3.75532001e-01 -9.35874104e-01 3.09699565e-01
7.77177751e-01 1.20902240e+00 -6.18706882e-01 1.45446444e+00
5.46351448e-02 2.03630969e-01 -6.43056750e-01 1.09677300e-01
1.34340510e-01 -4.20543849e-01 1.02010286e+00 1.24625587e+00
4.93955970e-01 8.17428976e-02 5.09858310e-01 1.03856313e+00
3.56132388e-01 -4.31079119e-01 -8.99382412e-01 4.37312067e-01
8.82363737e-01 1.32950735e+00 -8.91873300e-01 -4.13292170e-01
-4.92569149e-01 5.74649036e-01 1.65271834e-01 6.07497431e-02
-1.08029437e+00 -9.77847353e-03 1.15241456e+00 4.68214661e-01
7.67795622e-01 -1.00581360e+00 -2.48429060e-01 -1.01411009e+00
4.33653928e-02 1.66839004e-01 2.28465535e-02 -7.80432463e-01
-1.04995823e+00 1.59930676e-01 3.08709741e-01 -1.65743101e+00
-5.65557480e-02 -5.22030115e-01 -5.78897655e-01 7.66880333e-01
-1.79696584e+00 -1.27924776e+00 -5.16105831e-01 6.65377676e-01
5.44791639e-01 2.71425843e-01 8.31615403e-02 3.57144147e-01
-2.65072197e-01 6.99457759e-03 -1.44707903e-01 1.21392787e-01
5.10121524e-01 -9.62113082e-01 8.43645692e-01 8.73536646e-01
4.19782326e-02 3.37557405e-01 7.53217876e-01 -1.17538726e+00
-1.54509878e+00 -1.42033052e+00 9.28084612e-01 -1.25418556e+00
9.03838396e-01 -6.35999620e-01 -9.87262249e-01 7.34122932e-01
-2.76272774e-01 4.22411174e-01 3.93437147e-02 -4.45520431e-01
-1.02642104e-01 -1.02132283e-01 -9.01384294e-01 4.69918311e-01
1.62228596e+00 -4.06203210e-01 -5.58965325e-01 8.99316818e-02
7.65889168e-01 -8.05098295e-01 -6.28312826e-01 5.35763562e-01
5.98083198e-01 -4.94571745e-01 1.25460982e+00 -5.01566708e-01
-5.57316959e-01 -1.09224796e+00 -2.44961586e-02 -5.77238083e-01
-3.83703083e-01 -5.29910624e-01 -5.55067837e-01 1.16131449e+00
-1.34258851e-01 -2.31786698e-01 1.05052793e+00 4.86646533e-01
-2.30096012e-01 3.44225392e-02 -1.12090206e+00 -1.16929162e+00
-2.52806097e-01 -7.45466769e-01 5.22450805e-01 6.81463480e-01
-6.83203340e-01 1.37678176e-01 -1.19955525e-01 5.56584418e-01
1.25757134e+00 3.02815139e-01 1.22452772e+00 -1.71719587e+00
2.98819214e-01 -4.33270782e-01 -7.97328651e-01 -1.42602086e+00
4.60098624e-01 -8.70073497e-01 3.93027395e-01 -1.16721451e+00
-1.10399686e-01 -6.37575924e-01 3.03237319e-01 1.11658946e-01
-1.03936054e-01 4.65227991e-01 2.36012101e-01 3.40950310e-01
-6.43167019e-01 5.16973495e-01 1.05681431e+00 -2.31174082e-02
-2.61152536e-01 2.11157978e-01 2.14805976e-01 8.12854469e-01
3.88017863e-01 -7.11839259e-01 -1.31733179e-01 -4.20357585e-01
-2.89520770e-01 -1.83373168e-01 8.51848900e-01 -9.87488270e-01
5.40502965e-01 -3.63082588e-01 5.77560484e-01 -1.56827152e+00
5.74939787e-01 -1.18999016e+00 3.97877991e-01 5.41579425e-01
1.62325576e-01 3.15386474e-01 4.65717733e-01 7.47520089e-01
2.89554864e-01 2.89988294e-02 8.17492008e-01 7.41382539e-02
-1.09999144e+00 8.24289501e-01 -2.84508526e-01 -1.69853404e-01
1.41211212e+00 -7.71206975e-01 4.46289293e-02 1.69289395e-01
-4.83704299e-01 4.55672055e-01 7.88611829e-01 6.48929775e-01
6.71643436e-01 -1.78092206e+00 -7.74456441e-01 2.11896122e-01
1.54963523e-01 3.37313861e-01 -1.22964971e-01 1.01038778e+00
-4.12226111e-01 4.15821761e-01 -6.61140084e-02 -1.37030196e+00
-1.52969551e+00 5.73092341e-01 3.76901090e-01 4.26619023e-01
-1.09415424e+00 5.40844440e-01 -1.78758595e-02 -3.55041742e-01
1.87146425e-01 -4.77742672e-01 -9.12640542e-02 -5.75981699e-02
4.96800721e-01 5.17829537e-01 -7.84800574e-02 -1.10170901e+00
-6.76016629e-01 1.16864920e+00 1.15700960e-01 -9.35893953e-02
1.04258442e+00 -3.23406667e-01 2.41410986e-01 5.80695093e-01
1.08168638e+00 -2.22329628e-02 -1.92779982e+00 -2.31310189e-01
3.88839960e-01 -8.49963427e-01 1.66534558e-01 1.75513417e-01
-8.06124568e-01 9.02944446e-01 7.95219302e-01 -1.85086414e-01
3.64501029e-01 2.41630554e-01 8.01406860e-01 1.64816618e-01
3.94488543e-01 -8.32792342e-01 -2.51430005e-01 7.06172526e-01
4.02245134e-01 -1.34693480e+00 1.59153014e-01 -5.50689697e-01
-2.04550520e-01 1.00039768e+00 7.13533938e-01 -1.94320306e-01
7.09046304e-01 6.53692961e-01 -2.41776258e-02 -2.07455814e-01
-5.88573933e-01 -6.42795444e-01 5.41571736e-01 9.24520314e-01
-2.03314610e-02 -2.22198352e-01 8.14437047e-02 -2.27939382e-01
-1.92746595e-02 -4.53420840e-02 2.86085554e-03 9.47050810e-01
-7.68651247e-01 -9.52667773e-01 -5.93814552e-01 6.57888055e-02
3.61884445e-01 2.84137458e-01 -1.90645918e-01 9.75455225e-01
3.16785246e-01 6.59863114e-01 5.45656800e-01 -3.02396476e-01
4.83152419e-01 -1.06625624e-01 6.99034810e-01 -4.83827531e-01
-1.04629032e-01 2.02967197e-01 -1.64528847e-01 -6.54810131e-01
-7.16507614e-01 -1.22188962e+00 -1.49833059e+00 -4.12522405e-01
-5.48535168e-01 -2.14427575e-01 7.91947544e-01 9.78210270e-01
3.58733565e-01 4.58322540e-02 4.04838473e-01 -1.55197716e+00
-2.62206942e-01 -5.28123140e-01 -2.37722576e-01 3.87400150e-01
6.52863860e-01 -1.14169526e+00 -4.15445194e-02 1.50886491e-01] | [6.739437580108643, -2.2889902591705322] |
51017c16-9c5e-4d32-81aa-1d31b415ccac | event-based-motion-segmentation-with-spatio | 2012.08730 | null | https://arxiv.org/abs/2012.08730v3 | https://arxiv.org/pdf/2012.08730v3.pdf | Event-based Motion Segmentation with Spatio-Temporal Graph Cuts | Identifying independently moving objects is an essential task for dynamic scene understanding. However, traditional cameras used in dynamic scenes may suffer from motion blur or exposure artifacts due to their sampling principle. By contrast, event-based cameras are novel bio-inspired sensors that offer advantages to overcome such limitations. They report pixelwise intensity changes asynchronously, which enables them to acquire visual information at exactly the same rate as the scene dynamics. We develop a method to identify independently moving objects acquired with an event-based camera, i.e., to solve the event-based motion segmentation problem. We cast the problem as an energy minimization one involving the fitting of multiple motion models. We jointly solve two subproblems, namely event cluster assignment (labeling) and motion model fitting, in an iterative manner by exploiting the structure of the input event data in the form of a spatio-temporal graph. Experiments on available datasets demonstrate the versatility of the method in scenes with different motion patterns and number of moving objects. The evaluation shows state-of-the-art results without having to predetermine the number of expected moving objects. We release the software and dataset under an open source licence to foster research in the emerging topic of event-based motion segmentation. | ['Shaojie Shen', 'SiQi Liu', 'Xiuyuan Lu', 'Guillermo Gallego', 'Yi Zhou'] | 2020-12-16 | null | null | null | null | ['motion-segmentation'] | ['computer-vision'] | [ 5.64885259e-01 -4.40714687e-01 1.36051759e-01 -9.20784175e-02
-2.95719266e-01 -7.26044536e-01 4.03343558e-01 -5.61451167e-02
-7.06320286e-01 5.36267042e-01 -1.98761269e-01 1.29795477e-01
-1.14674591e-01 -5.16723037e-01 -7.05937445e-01 -9.18537736e-01
-9.21519250e-02 1.82345510e-01 7.51583815e-01 4.72628057e-01
3.01505476e-01 5.54075062e-01 -1.62877238e+00 1.66740671e-01
5.75463474e-01 6.16214871e-01 6.04790866e-01 1.01504028e+00
2.28138324e-02 8.69380474e-01 -3.62430811e-01 1.77681386e-01
1.98384896e-01 -6.46276951e-01 -7.65235305e-01 6.19876027e-01
1.61603600e-01 -1.96335763e-01 -3.32314819e-01 9.12338555e-01
2.36056015e-01 4.53752398e-01 3.93794388e-01 -1.38197279e+00
4.80696224e-02 -7.68657327e-02 -7.02748835e-01 7.11995304e-01
4.22534585e-01 3.21447372e-01 5.22962093e-01 -5.86569488e-01
9.64709222e-01 8.47228706e-01 2.89053410e-01 5.40954649e-01
-1.24423683e+00 -1.39977828e-01 1.79913804e-01 5.20170033e-01
-1.26490808e+00 -5.00562191e-01 8.07741404e-01 -6.89119995e-01
7.98368931e-01 3.11536163e-01 7.50499249e-01 1.07905436e+00
1.90937087e-01 7.25537062e-01 8.79461825e-01 -1.60347447e-01
5.26288807e-01 -1.60369873e-01 3.63639817e-02 4.79559839e-01
2.82702297e-01 5.17463312e-03 -7.44740665e-01 -6.35198802e-02
8.74502778e-01 1.69771940e-01 -5.10828555e-01 -4.55377638e-01
-1.52707672e+00 3.62619579e-01 -7.17966706e-02 3.21649343e-01
-5.70112884e-01 3.82553071e-01 1.28640741e-01 -1.50075972e-01
1.98815405e-01 -1.02249463e-03 -2.25841925e-01 -2.16154188e-01
-9.09121990e-01 8.21749046e-02 7.61633992e-01 6.63162947e-01
6.86103582e-01 -1.48797512e-01 3.68664332e-05 3.02405626e-01
3.19419682e-01 2.78912216e-01 3.94412130e-01 -1.19078624e+00
1.22034632e-01 4.35538888e-01 3.43683153e-01 -1.09533083e+00
-3.62964720e-01 1.06850222e-01 -8.17721486e-01 1.69868156e-01
7.41757691e-01 -1.02339253e-01 -7.68065333e-01 1.61334813e+00
6.39002323e-01 8.07992518e-01 -1.78409904e-01 1.02396882e+00
4.34357285e-01 9.01856840e-01 1.24912582e-01 -6.87241077e-01
1.39562619e+00 -6.36668146e-01 -9.20094252e-01 -1.69315919e-01
1.56233609e-01 -5.27417481e-01 6.52937114e-01 4.68189508e-01
-1.06918573e+00 -4.85116661e-01 -8.24900985e-01 2.42938623e-01
-3.48693043e-01 -2.08284423e-01 5.71981668e-01 4.98911023e-01
-8.99753928e-01 4.19587046e-01 -1.24070477e+00 -5.64922333e-01
2.96062350e-01 4.05966192e-01 -2.42380321e-01 1.01007849e-01
-7.69973755e-01 5.50757349e-01 5.33611059e-01 2.02402860e-01
-1.07908034e+00 -5.39674580e-01 -6.92497849e-01 -2.15626806e-01
5.24558008e-01 -8.76609325e-01 9.35932100e-01 -8.79258096e-01
-1.37099099e+00 9.57561493e-01 -5.11872768e-01 -4.05151546e-01
6.31290555e-01 -3.95436510e-02 -2.17573062e-01 5.06615996e-01
1.00210100e-01 6.04289532e-01 7.63423204e-01 -1.13833296e+00
-8.94577622e-01 -2.20856458e-01 -2.08257034e-01 1.47801176e-01
-1.96156070e-01 3.52220833e-02 -8.60473156e-01 -4.13756728e-01
-4.75752689e-02 -9.12109256e-01 -3.91545743e-01 9.58136246e-02
-1.25398353e-01 4.39978614e-02 1.03297305e+00 -4.65114236e-01
1.18343937e+00 -2.13629389e+00 2.44832322e-01 -1.42043695e-01
1.81689516e-01 4.09044027e-02 1.37351707e-01 2.71026462e-01
1.42488629e-01 -1.66789576e-01 -6.04171276e-01 -2.25248650e-01
-5.87290108e-01 1.64764017e-01 1.93707217e-02 7.15187669e-01
1.31605193e-01 6.51719868e-01 -1.18783426e+00 -7.48311102e-01
7.12896526e-01 2.86621273e-01 -1.42623350e-01 2.23680034e-01
-3.19099069e-01 9.87287939e-01 -4.26005900e-01 4.69443768e-01
5.66463590e-01 -4.11985755e-01 1.09519452e-01 -1.84627786e-01
-3.59635949e-01 -3.26766342e-01 -1.58895993e+00 1.92531002e+00
-7.90967569e-02 9.27028716e-01 1.04680210e-01 -8.89096797e-01
4.26770002e-01 4.05343831e-01 1.03454590e+00 -3.55582267e-01
4.36795643e-03 -1.16790645e-01 -2.77414709e-01 -8.11041057e-01
4.57826436e-01 5.62714599e-02 1.40914828e-01 3.14935148e-01
-3.03281955e-02 1.74656093e-01 4.72361773e-01 1.19552203e-01
1.21294570e+00 2.53052801e-01 2.91031122e-01 6.78642914e-02
3.74176681e-01 1.90175489e-01 6.66529477e-01 7.85381198e-01
-3.18506867e-01 6.57596111e-01 4.54083942e-02 -4.54754740e-01
-8.85372996e-01 -1.12657988e+00 -6.32072389e-02 5.65307856e-01
5.84714472e-01 -1.04511879e-01 -7.61492968e-01 -3.06213200e-01
-3.25227588e-01 4.07131106e-01 -5.14775693e-01 1.83263317e-01
-7.06428528e-01 -1.18510222e+00 1.72224790e-01 4.00427997e-01
5.27913153e-01 -1.05613744e+00 -1.44625068e+00 5.38009167e-01
-4.82733071e-01 -1.49308312e+00 -3.97312224e-01 -4.32162583e-02
-8.84082496e-01 -1.27006686e+00 -5.77156186e-01 -4.44193691e-01
6.06558502e-01 5.05524755e-01 8.92432690e-01 -2.69190997e-01
-7.53918290e-01 9.03581798e-01 -8.92879441e-02 -2.72493124e-01
-6.26448393e-02 -4.23949629e-01 -1.49715498e-01 5.19717574e-01
2.42986247e-01 -5.87853730e-01 -9.47035611e-01 3.25370818e-01
-1.32857049e+00 2.62281537e-01 2.55684018e-01 3.43072027e-01
8.07385743e-01 2.73412406e-01 1.54191688e-01 -5.75200021e-01
-5.71305752e-02 -5.32550752e-01 -8.55838478e-01 2.26122841e-01
-2.08426230e-02 -2.51952738e-01 3.44580770e-01 -6.09984577e-01
-1.17942429e+00 7.74478793e-01 5.14822125e-01 -5.11070311e-01
-5.23832262e-01 1.48591146e-01 -1.94003358e-01 2.39374992e-02
2.27182388e-01 4.90238488e-01 -3.41876566e-01 -1.16669744e-01
4.08435732e-01 2.98271000e-01 8.90658081e-01 -1.95063099e-01
4.66757923e-01 1.23398280e+00 1.96531847e-01 -1.07005191e+00
-3.35513353e-01 -1.04773116e+00 -6.80991292e-01 -8.08952451e-01
1.26204932e+00 -7.95972168e-01 -8.41857374e-01 7.91213751e-01
-1.35709500e+00 -3.09004098e-01 -3.18232983e-01 5.82856357e-01
-7.35775888e-01 4.84425902e-01 -5.95134854e-01 -9.83769178e-01
2.70472374e-02 -7.76441753e-01 9.80651915e-01 4.48048919e-01
-1.11536399e-01 -1.09256029e+00 2.40143821e-01 3.26884896e-01
8.66918825e-03 8.12622070e-01 2.35061988e-01 -6.72563091e-02
-1.17700291e+00 -2.55755149e-02 1.03358917e-01 -1.40064046e-01
1.70183063e-01 1.65801868e-01 -1.03797412e+00 4.99740206e-02
4.58447665e-01 4.17116374e-01 7.76786387e-01 8.80293906e-01
1.02213776e+00 7.48745576e-02 -5.94446480e-01 5.41976929e-01
1.62240088e+00 4.85996425e-01 6.90304816e-01 2.62513876e-01
8.18782866e-01 8.21850419e-01 6.09801650e-01 6.70094907e-01
2.23836452e-01 7.97005713e-01 5.02528131e-01 5.40393628e-02
3.05963680e-02 4.65608798e-02 3.43061477e-01 4.04362947e-01
-1.17846146e-01 -5.40893197e-01 -9.34836984e-01 8.63406777e-01
-2.19770026e+00 -1.26352417e+00 -7.56456435e-01 2.29950976e+00
4.50246751e-01 -1.83746159e-01 2.33983561e-01 4.13164198e-02
1.01152098e+00 1.43126234e-01 -5.86200178e-01 2.51644790e-01
-2.72620797e-01 -3.27530086e-01 5.16032636e-01 3.27712327e-01
-1.14815128e+00 7.32330322e-01 5.84015322e+00 4.53852504e-01
-1.00077903e+00 1.76193520e-01 5.72064281e-01 -2.94819087e-01
1.38496608e-01 1.51238590e-01 -6.46589041e-01 7.23755360e-01
9.06358421e-01 -1.92942500e-01 3.87460291e-01 2.58072823e-01
7.85818100e-01 -6.31496787e-01 -1.09059703e+00 1.15653205e+00
-3.19279172e-02 -1.27504432e+00 -8.63382146e-02 6.01106621e-02
6.82491302e-01 -3.17900106e-02 -3.39289010e-01 -6.41582608e-01
2.45237667e-02 -6.19600236e-01 7.90413916e-01 7.67107666e-01
2.83055097e-01 -3.54593784e-01 2.83806086e-01 3.86849314e-01
-1.37136424e+00 7.31677115e-02 -6.55026808e-02 -1.43263459e-01
7.60604620e-01 7.18093872e-01 -4.43529546e-01 5.54651916e-01
7.11645424e-01 9.20956790e-01 -3.73044014e-01 1.37236643e+00
1.80121943e-01 4.92005259e-01 -4.26356286e-01 2.89211214e-01
-4.91565093e-02 -4.24725771e-01 8.91116619e-01 1.19471872e+00
2.84757972e-01 3.03277701e-01 7.19203651e-02 1.00084710e+00
2.26740062e-01 -3.61393809e-01 -5.76195896e-01 1.19191878e-01
2.13884443e-01 1.31488431e+00 -1.31675553e+00 -3.39548558e-01
-5.11559486e-01 1.47016191e+00 -3.04726716e-02 4.92019147e-01
-8.13316703e-01 -5.01338132e-02 3.45998675e-01 7.71695599e-02
3.88615280e-01 -4.21788335e-01 3.61164473e-02 -1.18384254e+00
1.10202990e-01 -3.22301835e-01 5.24641335e-01 -8.74072015e-01
-9.84121203e-01 1.24715202e-01 2.48760074e-01 -1.24371779e+00
-1.07598677e-01 -4.56563771e-01 -6.34143710e-01 4.03690636e-01
-1.24933207e+00 -7.51007199e-01 -6.15354002e-01 7.39843786e-01
7.83209741e-01 5.93649566e-01 2.98455119e-01 2.62832224e-01
-6.60981953e-01 -4.03805435e-01 1.71630248e-01 3.56365591e-02
4.60949302e-01 -1.13245666e+00 1.17047010e-02 1.27736735e+00
3.45341057e-01 -6.62556943e-03 7.69493222e-01 -6.10642731e-01
-1.62772036e+00 -1.09208286e+00 6.11253560e-01 -6.28725767e-01
5.46168149e-01 -3.95240009e-01 -9.15367007e-01 4.85900044e-01
1.10107988e-01 2.30295539e-01 5.37952065e-01 -8.01522851e-01
4.93761182e-01 6.26471709e-04 -8.68897676e-01 4.44107443e-01
1.20571995e+00 -1.98723212e-01 -3.42575341e-01 2.92599708e-01
4.50886846e-01 -3.72693270e-01 -5.57151914e-01 2.44929865e-01
3.39627951e-01 -9.55613196e-01 1.01195323e+00 -2.34046504e-01
1.03924923e-01 -8.90140057e-01 1.96325555e-01 -7.20357835e-01
-1.02733508e-01 -9.06204283e-01 -3.38320345e-01 1.24462867e+00
-1.23124979e-01 -3.79299760e-01 8.34794104e-01 5.25823891e-01
1.85056567e-01 -2.26249218e-01 -1.06053662e+00 -8.23375583e-01
-8.00402582e-01 -5.18334866e-01 -4.34713885e-02 9.42772925e-01
-2.44675785e-01 9.11802500e-02 -3.35638344e-01 4.51532871e-01
8.68783176e-01 1.34548396e-01 6.81857824e-01 -1.02879381e+00
-3.32060665e-01 -2.68427014e-01 -5.65818787e-01 -9.86425579e-01
-1.10571459e-02 -3.86806548e-01 3.13982427e-01 -1.40077579e+00
4.67689157e-01 6.53653443e-02 -1.86604396e-01 -5.53932823e-02
-3.49085003e-01 1.66691720e-01 7.44014084e-02 4.52235162e-01
-9.21731174e-01 1.84500515e-01 9.74823594e-01 4.23630588e-02
-4.23887700e-01 -1.15758516e-01 6.94273040e-02 8.29829156e-01
4.98244196e-01 -6.64720118e-01 -4.84684050e-01 -3.50250304e-01
-1.83226727e-02 3.59449208e-01 8.95969152e-01 -1.19452417e+00
7.24778056e-01 -4.69857931e-01 2.71355242e-01 -5.73786855e-01
2.84088880e-01 -9.95853543e-01 7.68727720e-01 4.93580848e-01
-9.88059789e-02 1.97760947e-02 3.68237883e-01 1.14124131e+00
-2.36516014e-01 -2.17495456e-01 6.63615108e-01 -2.66146243e-01
-1.13719082e+00 2.85953283e-01 -8.43679488e-01 2.40400154e-02
1.45068431e+00 -5.56422710e-01 -7.46094435e-02 -2.28717297e-01
-7.63379514e-01 1.77975759e-01 5.97292960e-01 2.36490682e-01
5.99788547e-01 -9.29613054e-01 -5.30950963e-01 -1.41853541e-01
-3.51627134e-02 4.72395234e-02 5.22833526e-01 1.00550568e+00
-5.27612507e-01 2.01931372e-01 -1.44781232e-01 -1.00137258e+00
-1.30624473e+00 7.23785162e-01 4.20479476e-01 -1.33549228e-01
-5.75682342e-01 5.38904965e-01 3.83567393e-01 3.80685031e-01
8.06553103e-03 -2.81256974e-01 -1.33863479e-01 8.12765658e-02
5.07238746e-01 6.92224383e-01 -2.18874156e-01 -5.43852150e-01
-5.56631386e-01 7.19954252e-01 3.07778507e-01 -3.44166309e-01
1.09009671e+00 -6.07021689e-01 -4.80825454e-02 7.14136183e-01
8.87527585e-01 -2.96528041e-01 -1.73026919e+00 8.54066312e-02
9.55480784e-02 -4.94178653e-01 -4.52867746e-02 -3.60560507e-01
-9.46984470e-01 7.70231366e-01 7.23622799e-01 1.64149970e-01
1.40391612e+00 3.69179994e-02 7.29706049e-01 -1.53148519e-02
3.22519064e-01 -1.00897872e+00 1.06115140e-01 2.90239248e-02
2.88055658e-01 -1.17266488e+00 -6.88907877e-02 -6.97973073e-01
-4.09790605e-01 1.03386796e+00 3.49618495e-01 9.36740413e-02
3.56900245e-01 3.09705734e-01 -2.01610252e-01 -1.91393971e-01
-7.19891310e-01 -3.46050829e-01 5.69255240e-02 6.33500636e-01
5.64591512e-02 -1.91976801e-01 -2.93591112e-01 1.63073689e-01
5.60114384e-01 2.67650753e-01 6.57264471e-01 1.09548604e+00
-2.07980081e-01 -7.82400548e-01 -5.29293120e-01 4.87655923e-02
-4.38244969e-01 2.18361363e-01 -3.01852763e-01 6.53553426e-01
2.13704452e-01 1.15786910e+00 2.28200972e-01 -4.50894088e-02
2.19799533e-01 -8.62690955e-02 5.62498033e-01 -3.72728050e-01
-3.28998446e-01 1.59293041e-01 -2.44088694e-01 -7.79222429e-01
-1.07032955e+00 -1.00673842e+00 -1.38703215e+00 5.59972189e-02
-2.33895883e-01 -1.70000836e-01 5.65135777e-01 8.09260488e-01
4.82337594e-01 5.09205937e-01 4.17405635e-01 -9.57645655e-01
2.55788267e-01 -4.22105193e-01 -4.63533610e-01 6.81848824e-01
2.77290583e-01 -3.97667497e-01 -4.31888491e-01 9.40789104e-01] | [8.66247844696045, -1.192142128944397] |
166def54-c13e-4aa4-a2b6-96b9abb4915f | image-to-image-translation-via-hierarchical | 2103.01456 | null | https://arxiv.org/abs/2103.01456v1 | https://arxiv.org/pdf/2103.01456v1.pdf | Image-to-image Translation via Hierarchical Style Disentanglement | Recently, image-to-image translation has made significant progress in achieving both multi-label (\ie, translation conditioned on different labels) and multi-style (\ie, generation with diverse styles) tasks. However, due to the unexplored independence and exclusiveness in the labels, existing endeavors are defeated by involving uncontrolled manipulations to the translation results. In this paper, we propose Hierarchical Style Disentanglement (HiSD) to address this issue. Specifically, we organize the labels into a hierarchical tree structure, in which independent tags, exclusive attributes, and disentangled styles are allocated from top to bottom. Correspondingly, a new translation process is designed to adapt the above structure, in which the styles are identified for controllable translations. Both qualitative and quantitative results on the CelebA-HQ dataset verify the ability of the proposed HiSD. We hope our method will serve as a solid baseline and provide fresh insights with the hierarchically organized annotations for future research in image-to-image translation. The code has been released at https://github.com/imlixinyang/HiSD. | ['Rongrong Ji', 'Yongjian Wu', 'Feiyue Huang', 'Xudong Mao', 'Xiaopeng Hong', 'Liujuan Cao', 'Jie Hu', 'Shengchuan Zhang', 'Xinyang Li'] | 2021-03-02 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Li_Image-to-Image_Translation_via_Hierarchical_Style_Disentanglement_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Li_Image-to-Image_Translation_via_Hierarchical_Style_Disentanglement_CVPR_2021_paper.pdf | cvpr-2021-1 | ['multimodal-unsupervised-image-to-image'] | ['computer-vision'] | [ 4.00005221e-01 -8.75223950e-02 -4.05321211e-01 -7.03773737e-01
-7.14156449e-01 -8.70672405e-01 6.31268382e-01 -3.61980796e-01
-2.17293520e-02 6.54458523e-01 4.52768356e-01 -1.87052593e-01
1.77657545e-01 -3.52092683e-01 -4.84601945e-01 -6.42695844e-01
4.28087771e-01 4.45200831e-01 -2.71597445e-01 7.67808259e-02
3.73466946e-02 8.03015977e-02 -1.27755082e+00 3.25889230e-01
8.71324956e-01 7.25986958e-01 5.93236871e-02 2.99759924e-01
1.02426782e-01 3.89791459e-01 -1.56766266e-01 -7.29492128e-01
3.06330740e-01 -6.14793658e-01 -7.70153642e-01 3.74237180e-01
4.68431085e-01 -2.60778576e-01 -1.16713166e-01 1.19141340e+00
4.59196568e-01 -2.57011622e-01 5.67909420e-01 -1.60121870e+00
-1.23290479e+00 4.92593855e-01 -7.19997883e-01 -3.59201729e-01
3.01137000e-01 -8.36144835e-02 1.32992864e+00 -1.20600116e+00
6.27536535e-01 1.21995068e+00 2.05568835e-01 7.40194917e-01
-1.51554072e+00 -1.01164985e+00 2.39265114e-01 1.12228980e-02
-1.41965592e+00 -5.81149578e-01 7.66974151e-01 -6.65440261e-01
1.98013455e-01 4.51479435e-01 4.71200407e-01 1.44518161e+00
-2.21049041e-01 1.07758760e+00 1.69789660e+00 -4.86360252e-01
-1.15885176e-02 3.85185689e-01 5.52507350e-03 7.57657588e-01
2.63840407e-01 1.13996834e-01 -7.57768214e-01 2.62891389e-02
8.14628720e-01 -1.63026273e-01 -1.49460524e-01 -5.44705629e-01
-1.67613363e+00 6.83539093e-01 2.22724155e-01 1.95376435e-03
9.78867039e-02 -1.64156288e-01 3.16931933e-01 2.83851832e-01
6.21275544e-01 4.67218876e-01 -3.29823464e-01 2.48826351e-02
-7.93758750e-01 1.09489426e-01 4.71057147e-01 1.45262122e+00
8.31028938e-01 -1.77439764e-01 -4.58671063e-01 9.29959178e-01
2.28970096e-01 6.18660092e-01 2.37684444e-01 -9.68404174e-01
3.83375496e-01 5.88059783e-01 1.84320971e-01 -9.57673311e-01
-2.30135456e-01 -3.82404447e-01 -1.05431402e+00 1.74209271e-02
2.12370679e-01 -5.18282466e-02 -8.42125118e-01 2.14358640e+00
2.02573821e-01 -1.13745093e-01 -2.48519793e-01 1.05229199e+00
6.55481637e-01 4.72969890e-01 7.56438002e-02 -1.79208796e-02
1.68445623e+00 -1.07155538e+00 -8.07185769e-01 -1.35965645e-01
4.86402065e-01 -1.01430643e+00 1.62368131e+00 1.59641132e-01
-7.89163053e-01 -6.16146982e-01 -9.12001014e-01 -2.61209935e-01
6.36891974e-03 5.75129509e-01 6.22641027e-01 5.42634249e-01
-1.04103553e+00 5.47050796e-02 -7.81202435e-01 -4.22779709e-01
3.68060321e-01 3.09462041e-01 -4.00668293e-01 -1.35081813e-01
-1.17563200e+00 6.08403206e-01 1.11387029e-01 7.17365146e-02
-6.65045738e-01 -3.40446174e-01 -7.44536579e-01 -2.84351796e-01
3.73032421e-01 -8.97529185e-01 1.11382365e+00 -1.10496652e+00
-1.45208549e+00 1.19409394e+00 -3.34791988e-01 2.53497839e-01
4.83100921e-01 -2.38569796e-01 -2.80612409e-01 -2.05415815e-01
3.71724457e-01 1.13209021e+00 7.41799593e-01 -1.61695158e+00
-5.38480878e-01 -4.76411045e-01 1.49677157e-01 4.44895923e-01
-7.06214309e-01 1.88347325e-01 -6.43236697e-01 -1.03056967e+00
3.32549363e-01 -1.44504189e+00 1.31533682e-01 1.10735498e-01
-6.94501162e-01 -8.16819966e-02 4.83203739e-01 -5.61836004e-01
1.15763509e+00 -2.16224360e+00 7.03579843e-01 -1.66537657e-01
3.53076547e-01 -1.12409569e-01 -2.97517836e-01 3.51790637e-01
6.47024736e-02 3.09418797e-01 -2.06203297e-01 -7.08594203e-01
1.94404498e-01 1.89770445e-01 -1.76884174e-01 2.59838074e-01
1.23886466e-01 9.28631723e-01 -8.37671995e-01 -6.93981469e-01
-9.59215462e-02 2.07339063e-01 -5.28862953e-01 3.68218154e-01
-8.56103152e-02 8.77681494e-01 -5.94771266e-01 8.62242043e-01
5.85900187e-01 -4.23438579e-01 1.97760388e-01 -5.33253312e-01
-1.12960905e-01 8.58326629e-02 -7.67870665e-01 1.93490422e+00
-3.01556170e-01 5.05650699e-01 -8.64378586e-02 -4.92837369e-01
8.97481918e-01 3.45699459e-01 4.12992090e-01 -5.39739072e-01
-8.77630059e-03 2.15276405e-01 -1.18529692e-01 -4.19643253e-01
5.31884789e-01 -2.19701871e-01 -3.48781347e-01 4.37768996e-01
-5.49530797e-02 -4.87361401e-02 1.50741845e-01 2.31616840e-01
5.67813098e-01 5.92863798e-01 1.37802392e-01 -3.67790729e-01
3.96562546e-01 2.62858416e-03 8.73145998e-01 4.55525726e-01
-4.23465490e-01 8.80507529e-01 4.53841120e-01 -2.76944131e-01
-1.26875591e+00 -1.11360741e+00 -5.62512912e-02 1.19018579e+00
2.95183986e-01 -5.30128121e-01 -7.34047949e-01 -6.93044603e-01
-2.76053697e-01 4.68192190e-01 -6.20338023e-01 1.01653142e-02
-3.26865375e-01 -7.16129899e-01 7.27215827e-01 4.41355795e-01
5.29558599e-01 -8.00579607e-01 -3.36554721e-02 -2.53153950e-01
-7.12756455e-01 -1.27303958e+00 -7.63932407e-01 -1.91095024e-01
-6.21814251e-01 -6.18540645e-01 -7.59231806e-01 -9.58320439e-01
8.43076289e-01 4.59662646e-01 1.13138902e+00 -2.31642514e-01
2.93795735e-01 1.18991192e-02 -5.72544694e-01 -1.75919667e-01
-3.34291220e-01 2.31926873e-01 2.07601190e-01 1.86174616e-01
3.09538662e-01 -5.87296128e-01 -6.28425539e-01 7.15370238e-01
-9.18381929e-01 8.68837059e-01 6.76936269e-01 9.43344355e-01
7.16405690e-01 -3.64481091e-01 4.08189923e-01 -9.66319323e-01
5.84600329e-01 -2.33939335e-01 -4.33783978e-01 4.53902125e-01
-6.91920638e-01 3.46613526e-02 3.72400373e-01 -4.63028431e-01
-1.04447126e+00 2.04242200e-01 2.07212374e-01 -3.79157990e-01
-2.55323410e-01 3.84428233e-01 -5.23843169e-01 1.33522987e-01
3.02685946e-01 1.45353973e-01 -1.67410478e-01 -4.29949969e-01
6.23893082e-01 7.82704532e-01 3.32190722e-01 -9.76814926e-01
7.98906326e-01 3.63283455e-01 -2.40454599e-01 -2.44525298e-01
-1.21852458e+00 -1.71689734e-01 -8.66930723e-01 -2.29393095e-01
9.89558280e-01 -1.21417534e+00 -2.39157334e-01 4.49284196e-01
-1.00312769e+00 -1.31623164e-01 3.77765261e-02 3.96962434e-01
-5.98937273e-01 3.84931654e-01 -7.44054794e-01 -5.01010716e-01
-1.48795381e-01 -1.42056358e+00 1.31992793e+00 1.20722175e-01
-5.40787280e-01 -8.13989341e-01 -3.68839428e-02 7.88698673e-01
2.17722192e-01 5.28133586e-02 1.01717222e+00 -4.08593923e-01
-5.77170372e-01 4.94657364e-03 -5.06411970e-01 2.25998506e-01
4.16762412e-01 -1.05002858e-01 -9.03704047e-01 -3.35382849e-01
-5.04882038e-02 -6.57200336e-01 5.47200739e-01 -5.54320253e-02
9.71002996e-01 -3.42175245e-01 -1.35624260e-01 6.34045184e-01
1.06732428e+00 -7.63185173e-02 3.15072596e-01 3.84744167e-01
1.13143110e+00 5.94805896e-01 5.03244698e-01 2.93873012e-01
6.41691625e-01 1.03949273e+00 7.30005279e-02 -3.88888568e-01
-3.36474627e-01 -4.97327060e-01 4.58418608e-01 1.26717556e+00
-7.39129782e-02 -3.05190027e-01 -7.33462155e-01 2.30516642e-01
-1.74489260e+00 -5.51378191e-01 -1.01323374e-01 2.11452746e+00
1.14449131e+00 -2.38154884e-02 2.43512504e-02 -2.30636761e-01
8.01718235e-01 2.05305681e-01 -4.79312420e-01 -9.61421281e-02
-2.72280782e-01 -3.46851826e-01 1.54336736e-01 3.89790595e-01
-1.16925478e+00 1.18181241e+00 5.68014908e+00 8.21541071e-01
-9.79090929e-01 2.45262951e-01 7.02238083e-01 1.16802463e-02
-5.22122979e-01 2.78537452e-01 -7.13415146e-01 4.55291808e-01
2.94097424e-01 -7.88745135e-02 5.17069697e-01 4.53420788e-01
2.89509445e-01 1.61958426e-01 -1.21398675e+00 8.73170257e-01
6.52483776e-02 -8.29813063e-01 4.28205073e-01 9.94503573e-02
8.92796934e-01 -2.66458333e-01 3.45646232e-01 1.52567059e-01
3.99485499e-01 -6.77432060e-01 9.52596605e-01 3.09741080e-01
1.17764044e+00 -3.70621353e-01 4.01943952e-01 2.50693023e-01
-1.05697429e+00 3.21556926e-01 -1.44743174e-01 -3.08629088e-02
2.27999344e-01 5.15654266e-01 -4.32617456e-01 7.76753247e-01
3.54566485e-01 8.88380408e-01 -7.03438044e-01 4.87578481e-01
-4.56269324e-01 5.86070299e-01 1.54685572e-01 1.57804862e-01
3.07883639e-02 -5.69054067e-01 5.25009930e-01 1.10761225e+00
3.66553903e-01 2.81206742e-02 4.52705145e-01 1.06881702e+00
-2.08603427e-01 8.31143558e-02 -5.17421603e-01 -1.69057205e-01
5.61710894e-01 1.41306520e+00 -8.14471602e-01 -3.50206822e-01
-6.03309393e-01 1.32118058e+00 4.17651653e-01 5.28018713e-01
-8.63282561e-01 5.00265993e-02 6.72183335e-01 -1.84732035e-01
-1.77775890e-01 -3.68822455e-01 -6.83280885e-01 -1.68572891e+00
2.13242080e-02 -1.08752382e+00 1.65883794e-01 -9.09534276e-01
-1.47163749e+00 7.60435283e-01 3.27202342e-02 -1.55195343e+00
2.42376029e-01 -4.31649685e-01 -7.27744028e-02 8.07717919e-01
-1.06091070e+00 -1.61402714e+00 -3.05051178e-01 3.91618639e-01
7.24920034e-01 -9.40901265e-02 9.89715159e-01 4.49038804e-01
-9.09784555e-01 8.61847997e-01 4.16194797e-02 1.71106637e-01
1.12117982e+00 -1.09941578e+00 2.33475357e-01 9.10045564e-01
4.45090145e-01 6.51336789e-01 7.00356007e-01 -6.20042026e-01
-1.31519246e+00 -1.10967934e+00 1.02295005e+00 -5.94074845e-01
8.05584908e-01 -7.87590683e-01 -6.91847563e-01 8.31756771e-01
3.79650205e-01 -2.61293948e-01 9.61627543e-01 2.55112916e-01
-5.93081534e-01 1.15400543e-02 -7.05453277e-01 8.74757707e-01
1.17402077e+00 -5.19479215e-01 -2.86026329e-01 3.31441820e-01
8.91411662e-01 -3.14930946e-01 -8.72771680e-01 4.65776563e-01
6.59517527e-01 -8.65797937e-01 6.34682178e-01 -5.14915764e-01
8.61677766e-01 -3.73051137e-01 -4.72299457e-01 -1.26820529e+00
-6.35058999e-01 -4.56646502e-01 2.27084503e-01 1.60160840e+00
5.33472359e-01 -3.64245921e-01 6.03299856e-01 6.67257786e-01
-1.34082541e-01 -7.70994484e-01 -5.98047793e-01 -8.11741948e-01
1.06379464e-02 -1.22462310e-01 5.40069759e-01 1.16700041e+00
-1.29743248e-01 7.65533090e-01 -8.49410951e-01 9.88634452e-02
6.47047698e-01 4.66556877e-01 7.26447940e-01 -9.18613255e-01
-2.94446200e-01 -4.61868912e-01 -2.12072283e-02 -1.24447513e+00
2.25978374e-01 -1.08667099e+00 -8.21660757e-02 -1.28481793e+00
7.27523088e-01 -7.02610910e-01 -3.21908176e-01 6.67547584e-01
-4.49815184e-01 7.06207633e-01 3.99323732e-01 7.71969974e-01
-6.84606314e-01 7.41050482e-01 1.63030016e+00 -1.49460420e-01
9.41487849e-02 -6.12865463e-02 -9.49489415e-01 6.09391570e-01
8.68791938e-01 -5.08440793e-01 -4.22634453e-01 -7.72689760e-01
3.47810566e-01 -6.52979836e-02 2.11266518e-01 -5.25168300e-01
-1.49816260e-01 -3.02707911e-01 2.35557169e-01 -2.57597446e-01
4.03441548e-01 -7.20908880e-01 2.56903470e-01 4.41144928e-02
-7.18425274e-01 1.99356899e-01 -1.67168647e-01 3.50633860e-01
-3.18704128e-01 1.71247229e-01 5.46319366e-01 7.04067349e-02
-4.94396031e-01 4.28247839e-01 -1.48597807e-02 9.05729458e-03
7.51581073e-01 -4.70414534e-02 -1.87443331e-01 -3.30169559e-01
-7.56795883e-01 2.36734986e-01 8.03148806e-01 6.39682531e-01
3.68559003e-01 -1.85165608e+00 -8.86485338e-01 2.10067928e-01
5.74610412e-01 -1.80394605e-01 -3.72744724e-02 8.78287911e-01
-3.84523235e-02 1.92893878e-01 -3.12098861e-01 -6.24635637e-01
-1.35199726e+00 5.42920232e-01 -1.60053089e-01 -1.70252636e-01
-3.98358315e-01 7.28903294e-01 5.98020315e-01 -5.17332256e-01
-6.86306693e-03 1.00964002e-01 3.08029242e-02 7.02719614e-02
1.02150932e-01 1.44519627e-01 -1.99778989e-01 -8.63673210e-01
-4.36207131e-02 6.16130054e-01 -1.78801611e-01 -3.50430429e-01
9.61204708e-01 -6.75392091e-01 -3.50342810e-01 7.34255195e-01
1.03769720e+00 1.80954784e-01 -1.27071428e+00 -4.04857486e-01
2.84071099e-02 -5.43269336e-01 -2.84997672e-01 -9.85814095e-01
-1.00708580e+00 7.90775359e-01 4.04957056e-01 -3.53031419e-02
1.28165543e+00 9.50531438e-02 6.87542140e-01 5.69994189e-02
4.51070487e-01 -6.68116391e-01 8.37778449e-02 4.11363423e-01
8.95865321e-01 -1.33407462e+00 -5.59707284e-02 -6.43171012e-01
-1.03724539e+00 6.91231489e-01 7.85527945e-01 2.74693489e-01
2.71386057e-01 1.78710312e-01 3.63745004e-01 2.24708468e-02
-6.56300366e-01 -1.01111598e-01 4.89579290e-01 3.91129613e-01
7.92944074e-01 5.36813796e-01 -3.81275207e-01 5.48035681e-01
-2.41016075e-01 6.23468868e-03 3.24631900e-01 6.46905780e-01
-6.67003617e-02 -1.54491282e+00 -2.56056964e-01 1.17141485e-01
-1.31297126e-01 -2.64012754e-01 -6.77735925e-01 6.21445775e-01
2.27007017e-01 9.77887034e-01 -2.64874667e-01 -7.28880525e-01
2.83460647e-01 6.67801648e-02 5.43680251e-01 -5.88864267e-01
-2.16291636e-01 2.67005354e-01 9.68220308e-02 -3.88213068e-01
-5.71086228e-01 -6.29343510e-01 -7.89287567e-01 -1.54786393e-01
-2.44616196e-01 1.51293809e-02 4.24610585e-01 7.37993002e-01
5.57429433e-01 2.36679971e-01 8.19553494e-01 -8.25981796e-01
-4.13586736e-01 -9.79583085e-01 -3.85578036e-01 7.12227881e-01
1.50034437e-02 -7.60096014e-01 -1.09245978e-01 4.45363045e-01] | [11.225132942199707, 0.1545376479625702] |
31df83f2-1532-4272-87ae-c0203aed05d9 | image-based-fashion-product-recommendation | 1805.08694 | null | http://arxiv.org/abs/1805.08694v2 | http://arxiv.org/pdf/1805.08694v2.pdf | Image Based Fashion Product Recommendation with Deep Learning | We develop a two-stage deep learning framework that recommends fashion images
based on other input images of similar style. For that purpose, a neural
network classifier is used as a data-driven, visually-aware feature extractor.
The latter then serves as input for similarity-based recommendations using a
ranking algorithm. Our approach is tested on the publicly available Fashion
dataset. Initialization strategies using transfer learning from larger product
databases are presented. Combined with more traditional content-based
recommendation systems, our framework can help to increase robustness and
performance, for example, by better matching a particular customer style. | ['Markus Haltmeier', 'Hessel Tuinhof', 'Clemens Pirker'] | 2018-05-06 | null | null | null | null | ['product-recommendation'] | ['miscellaneous'] | [ 1.97286457e-01 -3.61134380e-01 -3.64243716e-01 -7.84153938e-01
-3.28538626e-01 -7.09701121e-01 6.44387007e-01 3.18591714e-01
-2.79136389e-01 6.06877953e-02 3.78048509e-01 -2.64459044e-01
-2.33374283e-01 -9.43334281e-01 -6.91948414e-01 -3.27459812e-01
4.16194826e-01 3.80937427e-01 -5.76831587e-02 -4.06699866e-01
5.87178230e-01 5.88472247e-01 -1.71518958e+00 8.90479207e-01
3.86788011e-01 1.45064735e+00 1.91346154e-01 5.08740544e-01
-1.37881562e-01 5.30224264e-01 -3.82964551e-01 -6.18384004e-01
3.73248696e-01 -2.63797343e-01 -5.97972453e-01 3.03347021e-01
9.99160707e-01 -4.29928958e-01 -3.55031461e-01 8.38613093e-01
3.25557351e-01 2.99526602e-01 8.87848854e-01 -8.13315868e-01
-1.35044467e+00 7.44730651e-01 -3.01008672e-01 3.35798934e-02
3.34244162e-01 2.08479643e-01 1.06585109e+00 -8.89184654e-01
7.82839119e-01 1.28734767e+00 3.83198917e-01 2.28544086e-01
-1.35688519e+00 -3.04677933e-01 2.54380345e-01 3.25995773e-01
-9.05680358e-01 -1.83277190e-01 1.26471281e+00 -4.05402750e-01
6.86588764e-01 2.91185051e-01 8.33258510e-01 1.22262609e+00
2.52010822e-01 1.11811769e+00 1.05093491e+00 -1.49803028e-01
3.23014230e-01 4.63154793e-01 1.16591610e-01 6.25476778e-01
5.33114932e-02 1.86231315e-01 -2.91229248e-01 1.54915705e-01
8.06629658e-01 3.88855040e-01 2.06709117e-01 -6.82773471e-01
-8.28782499e-01 1.03518987e+00 9.22416747e-01 1.64521858e-01
-5.06642282e-01 -1.01924837e-01 4.41794455e-01 6.79925263e-01
5.62446892e-01 7.73133755e-01 -4.16036129e-01 4.87545371e-01
-1.20852840e+00 3.05426478e-01 5.98094821e-01 5.54917097e-01
4.61688519e-01 -1.35686724e-02 -4.96839345e-01 8.14639509e-01
5.72054923e-01 2.55358309e-01 4.12172616e-01 -6.54571831e-01
6.34007230e-02 6.88945651e-01 4.93906476e-02 -1.23731768e+00
-4.36213762e-01 -6.20799780e-01 -4.72787112e-01 5.52819371e-01
4.65495855e-01 2.39326552e-01 -1.00277340e+00 1.11026931e+00
2.37094797e-02 -7.03506768e-02 -7.63810892e-03 1.42365837e+00
1.21808136e+00 4.07977611e-01 1.04621753e-01 3.65067780e-01
1.55116105e+00 -9.97082353e-01 -2.72122264e-01 -3.54385795e-03
2.40832314e-01 -1.15604389e+00 1.18530667e+00 7.18548596e-01
-9.16597843e-01 -1.06226313e+00 -1.22839952e+00 -4.28702123e-02
-7.64168322e-01 5.86713076e-01 6.53517365e-01 6.56177998e-01
-1.08870304e+00 8.01914811e-01 -1.94011062e-01 -7.20055103e-01
4.76997823e-01 3.41555536e-01 -2.27060392e-01 -1.99565664e-01
-7.76865363e-01 7.90826261e-01 1.92108214e-01 9.38575789e-02
-5.87584257e-01 -6.37664258e-01 -8.87714088e-01 -8.73199850e-02
5.15935048e-02 -8.51216257e-01 8.74958396e-01 -1.41329038e+00
-1.74430597e+00 8.93847942e-01 3.62967849e-01 -5.10434091e-01
4.42998588e-01 -2.53164440e-01 -6.89332664e-01 7.63979629e-02
-2.56218851e-01 6.69152260e-01 1.20126235e+00 -1.44895923e+00
-4.41958368e-01 -3.73513341e-01 4.00185764e-01 -5.44470735e-02
-3.82921427e-01 3.05841900e-02 -5.21180749e-01 -9.83467877e-01
-1.20360792e-01 -9.70690608e-01 -1.76222697e-01 4.95741628e-02
-4.06899899e-01 -1.54292673e-01 7.33752966e-01 -7.17802882e-01
6.66662633e-01 -2.12771130e+00 1.84707284e-01 6.07485771e-01
1.74286053e-01 3.22300702e-01 -4.96174812e-01 2.87339389e-01
1.20925695e-01 -3.42224509e-01 4.64841068e-01 -1.52881920e-01
-3.50431092e-02 -2.13410825e-01 -1.44072637e-01 2.80848980e-01
4.38113481e-01 1.12312222e+00 -8.26785505e-01 -1.05859473e-01
7.20823050e-01 5.36577582e-01 -6.23502851e-01 3.26807171e-01
-3.01018357e-01 4.15288419e-01 -2.16125131e-01 5.68114579e-01
4.60591525e-01 -2.17432126e-01 1.99083149e-01 -8.84208083e-01
2.25523561e-01 2.22337879e-02 -1.04730356e+00 1.91118813e+00
-8.52205276e-01 5.91086388e-01 -3.52261037e-01 -1.01307261e+00
1.28052592e+00 -2.75529385e-01 3.09441864e-01 -9.05761302e-01
4.91167277e-01 -2.01303348e-01 -4.97083068e-02 -4.34124202e-01
8.70808661e-01 4.74919021e-01 -5.59263565e-02 4.27834988e-01
1.47737503e-01 1.86050966e-01 1.03105024e-01 1.86481416e-01
6.58380687e-01 4.85919952e-01 -3.68330665e-02 -2.19890937e-01
6.06780529e-01 1.37713626e-01 -5.31377364e-03 5.97732127e-01
1.97702482e-01 6.93320394e-01 -1.06605172e-01 -9.56125498e-01
-1.11744869e+00 -1.02015758e+00 -2.59164721e-02 1.38650417e+00
2.02913105e-01 -2.58736759e-01 -3.37499827e-01 -9.45223272e-01
3.35684299e-01 6.50984526e-01 -7.98891962e-01 -4.46596712e-01
-3.14947844e-01 -2.14562923e-01 -2.80107617e-01 7.65251875e-01
2.69504815e-01 -1.18862677e+00 -3.46034706e-01 3.16823512e-01
5.54806769e-01 -8.01144660e-01 -5.03844440e-01 2.58587778e-01
-8.46991897e-01 -8.66452098e-01 -9.71793294e-01 -9.55481529e-01
7.76243985e-01 3.31168741e-01 1.29241025e+00 8.04211572e-02
-2.23247990e-01 6.51387513e-01 -5.49825966e-01 -1.72460541e-01
-3.84681314e-01 -1.21625066e-01 -6.85924888e-02 2.77290910e-01
5.35609722e-01 -1.36070907e-01 -1.11263669e+00 2.93747544e-01
-6.35457337e-01 -1.80397719e-01 7.76559770e-01 8.33047986e-01
6.59931958e-01 -2.90138155e-01 5.10901332e-01 -9.64679182e-01
9.01088178e-01 -3.97156477e-01 -5.73568285e-01 2.92911261e-01
-7.09777355e-01 2.05330886e-02 8.61607432e-01 -5.72465837e-01
-8.57227802e-01 3.98728192e-01 -7.16012418e-02 -4.01918739e-01
-5.55480003e-01 3.87498260e-01 -1.21839587e-02 -1.64248154e-01
6.37045264e-01 -5.46855107e-02 -9.24467817e-02 -8.12317014e-01
1.00984704e+00 7.35778928e-01 4.62365031e-01 -1.24175258e-01
5.92029870e-01 2.37560049e-01 -2.60615528e-01 -3.84512752e-01
-7.28636503e-01 -3.60574484e-01 -8.88598979e-01 -4.55127805e-01
5.95802963e-01 -7.16974020e-01 -7.07207143e-01 6.89674094e-02
-7.45810986e-01 -2.16071069e-01 -2.93214589e-01 4.87924069e-01
-5.06984890e-01 7.45937526e-02 -7.15110064e-01 -4.50798094e-01
-4.72735316e-01 -1.13532317e+00 1.17128050e+00 4.03251171e-01
-3.81253690e-01 -9.79426205e-01 -5.20244017e-02 6.34801209e-01
5.87712586e-01 -4.22212481e-02 8.61357152e-01 -8.73960733e-01
-2.35189036e-01 -4.65515524e-01 -4.60868388e-01 2.52151400e-01
5.94978146e-02 5.15046716e-02 -9.12229598e-01 -2.85740137e-01
-5.58911145e-01 -2.18950540e-01 1.12947476e+00 5.52571237e-01
1.41052163e+00 -1.20981082e-01 -1.84710443e-01 5.68079889e-01
1.47775149e+00 1.51665881e-01 3.69243920e-01 4.77754503e-01
8.75627041e-01 5.66202819e-01 6.88341439e-01 2.71645278e-01
3.95736277e-01 8.69278073e-01 4.74112540e-01 -5.17674327e-01
-4.29591894e-01 -2.71546692e-01 6.01108335e-02 1.86315715e-01
-1.35549288e-02 -6.83478266e-02 -2.77780652e-01 2.62254089e-01
-1.99025226e+00 -8.41275871e-01 1.58491209e-01 1.99100995e+00
3.96302432e-01 1.05494559e-01 6.67265892e-01 -4.27829064e-02
3.12418878e-01 -8.35459232e-02 -8.19082975e-01 -7.81461716e-01
7.50982463e-02 1.38947025e-01 4.40648317e-01 -1.94883898e-01
-1.30623293e+00 9.18765426e-01 6.60478258e+00 4.45770621e-01
-1.42682874e+00 -2.29516044e-01 8.05432916e-01 -2.13197500e-01
-2.70054489e-01 -5.86611748e-01 -4.64345306e-01 3.63359541e-01
6.06194139e-01 2.54279613e-01 5.87897301e-01 1.17884111e+00
1.79548681e-01 1.74135581e-01 -1.17145348e+00 1.08921814e+00
4.28585291e-01 -1.62083805e+00 1.68988824e-01 -1.67502165e-02
6.56375349e-01 -2.86023498e-01 5.73714554e-01 3.11584055e-01
5.05999446e-01 -8.38051915e-01 8.80051255e-01 6.39657557e-01
5.12874007e-01 -8.60103846e-01 8.57311130e-01 -4.37303573e-01
-1.16571605e+00 -2.26700574e-01 -5.65386236e-01 1.19980045e-01
-1.18946927e-02 4.03397650e-01 -5.34861267e-01 5.96631944e-01
8.25830221e-01 1.08990157e+00 -8.71738672e-01 1.14390147e+00
1.25712920e-02 3.63289833e-01 1.41567618e-01 -7.02273399e-02
9.66668576e-02 -3.07411402e-01 1.04033746e-01 1.20773315e+00
3.54038514e-02 -3.62329602e-01 5.49551845e-01 7.96619117e-01
-5.27308881e-02 4.68534857e-01 -7.07347929e-01 6.76686466e-02
-2.20791455e-02 1.87933815e+00 -1.05208457e+00 -2.96520144e-01
-5.89020431e-01 1.26565742e+00 2.43121997e-01 4.02891606e-01
-6.43155754e-01 -1.33932963e-01 5.76415479e-01 9.85645279e-02
8.32359076e-01 -3.91397960e-02 -3.02634805e-01 -8.18034351e-01
-4.16719824e-01 -9.55907822e-01 4.68817562e-01 -7.59703755e-01
-1.86377764e+00 6.51413500e-01 -3.59483957e-01 -1.46428812e+00
-1.03423715e-01 -9.18407500e-01 -7.09166288e-01 6.02796197e-01
-1.29391241e+00 -1.73904991e+00 -2.80061632e-01 6.84130788e-01
6.40255392e-01 -6.77043259e-01 6.22278392e-01 2.26113275e-01
-4.64886963e-01 9.15820897e-01 8.38439837e-02 1.32451802e-01
8.54602277e-01 -1.35645700e+00 3.17184120e-01 4.85579908e-01
6.10728979e-01 6.09080374e-01 6.60973310e-01 -5.32397449e-01
-1.64360297e+00 -1.29076004e+00 1.16928525e-01 -3.72952431e-01
7.42778897e-01 -2.58945435e-01 -5.74550807e-01 3.69344771e-01
4.13133860e-01 -8.79460499e-02 9.34815526e-01 3.30253184e-01
-5.37815690e-01 -3.76843691e-01 -1.20588803e+00 5.65016866e-01
8.77750754e-01 -4.09265399e-01 -5.54238200e-01 2.30559096e-01
3.38945121e-01 -1.73856497e-01 -1.12032962e+00 7.78889582e-02
9.85394657e-01 -7.40130842e-01 1.25613403e+00 -1.02775562e+00
5.94596565e-01 -1.83163792e-01 -1.57294199e-01 -1.59173667e+00
-7.93357134e-01 -1.67103037e-01 -1.34381190e-01 1.06649363e+00
4.88810390e-01 -1.06131822e-01 8.01130950e-01 3.24973464e-01
4.69271503e-02 -6.26486063e-01 -1.58163548e-01 -6.79992199e-01
-2.25393608e-01 -3.58029157e-01 7.95807540e-01 7.08039641e-01
-1.37465015e-01 3.80397230e-01 -5.25264800e-01 -9.82929096e-02
4.61783111e-01 7.13824928e-01 7.86497593e-01 -1.30840313e+00
-3.29763383e-01 -5.58827341e-01 -5.40509462e-01 -8.65065753e-01
-2.70962110e-03 -1.16056240e+00 -2.06348866e-01 -1.69917047e+00
6.53205216e-02 -3.40083390e-01 -9.35643375e-01 3.83236617e-01
2.03036834e-02 9.73283708e-01 3.34547997e-01 -1.16808005e-01
-6.81611717e-01 2.12048203e-01 1.27354801e+00 -5.89437842e-01
-3.40237737e-01 5.13185672e-02 -9.55175340e-01 6.29401207e-01
9.34834421e-01 -1.04238935e-01 -3.71165931e-01 -2.92245358e-01
1.57233909e-01 -4.13290620e-01 4.39161628e-01 -9.68479395e-01
-9.87127423e-02 9.46687087e-02 1.14637327e+00 -7.13203251e-01
2.24258468e-01 -1.06808066e+00 4.25759517e-02 3.94116729e-01
-7.87059665e-01 6.11952171e-02 1.16411537e-01 5.73776007e-01
7.69057199e-02 -6.45808578e-02 4.01728779e-01 1.80200506e-02
-1.11414492e+00 1.72625348e-01 -2.77409017e-01 -8.40155602e-01
8.80648077e-01 -2.26999909e-01 -2.43416280e-01 -5.11673570e-01
-9.80904102e-01 -1.73300624e-01 5.00472784e-01 1.10464144e+00
7.49071598e-01 -1.66645813e+00 -5.80851912e-01 2.76011109e-01
6.49634659e-01 -1.03484130e+00 1.36551604e-01 4.47176218e-01
-1.80163190e-01 1.47792846e-01 -6.85190260e-01 -6.13321662e-01
-1.32497835e+00 9.59438980e-01 5.78633398e-02 1.47055397e-02
-6.17224038e-01 6.96962535e-01 -1.36639550e-01 -3.13809693e-01
2.21956700e-01 -3.24902207e-01 -8.82863700e-01 1.54062971e-01
7.37554610e-01 1.04708202e-01 1.92066491e-01 -7.04966009e-01
-2.23871410e-01 6.73606038e-01 -3.55891615e-01 2.47245193e-01
1.44691217e+00 -1.21581167e-01 4.35070932e-01 2.06026450e-01
1.10470343e+00 -1.12976141e-01 -1.27021492e+00 -4.15941745e-01
-1.20712142e-03 -5.99793673e-01 5.39045095e-01 -1.19576526e+00
-1.62400162e+00 4.84288603e-01 1.28235555e+00 2.40615845e-01
1.18948865e+00 1.08455479e-01 7.05711544e-01 2.82644570e-01
1.85628697e-01 -1.10278583e+00 2.49567017e-01 -1.40298799e-01
1.03859031e+00 -1.60760033e+00 6.17668703e-02 6.53968425e-03
-9.09251034e-01 1.44167447e+00 5.35016775e-01 -5.86026073e-01
6.71369076e-01 -1.18774585e-01 3.74110758e-01 -3.83208841e-01
-2.73744166e-01 -4.80636746e-01 1.10878241e+00 7.73780167e-01
7.18222201e-01 7.59332478e-02 -2.08749488e-01 7.03270197e-01
-4.64400016e-02 9.07508060e-02 -5.82606010e-02 5.64422369e-01
-2.85351425e-01 -1.22409356e+00 -1.91646442e-01 8.93224478e-01
1.44279404e-02 4.79750131e-04 -4.85628694e-01 3.33676517e-01
1.15348831e-01 1.06806660e+00 2.66711175e-01 -8.46142769e-01
4.98562276e-01 -2.86606669e-01 5.75424790e-01 -5.14234960e-01
-1.00811708e+00 3.60876918e-02 1.46012232e-01 -8.43921304e-01
-3.44264299e-01 -5.16118765e-01 -4.75548983e-01 -2.15960547e-01
2.18033958e-02 -4.25153852e-01 8.16391110e-01 8.50775421e-01
5.17525017e-01 6.13930583e-01 7.49764383e-01 -1.09162962e+00
-2.74128914e-01 -7.37762332e-01 -3.75108451e-01 9.88202333e-01
-1.95094626e-02 -6.27245367e-01 3.77430052e-01 1.42992988e-01] | [11.065834045410156, 0.12605157494544983] |
afb1c3d0-3b37-4715-9486-993eb37cb5e3 | experimental-study-on-reinforcement-learning | 2011.09246 | null | https://arxiv.org/abs/2011.09246v1 | https://arxiv.org/pdf/2011.09246v1.pdf | Experimental Study on Reinforcement Learning-based Control of an Acrobot | We present computational and experimental results on how artificial intelligence (AI) learns to control an Acrobot using reinforcement learning (RL). Thereby the experimental setup is designed as an embedded system, which is of interest for robotics and energy harvesting applications. Specifically, we study the control of angular velocity of the Acrobot, as well as control of its total energy, which is the sum of the kinetic and the potential energy. By this means the RL algorithm is designed to drive the angular velocity or the energy of the first pendulum of the Acrobot towards a desired value. With this, libration or full rotation of the unactuated pendulum of the Acrobot is achieved. Moreover, investigations of the Acrobot control are carried out, which lead to insights about the influence of the state space discretization, the episode length, the action space or the mass of the driven pendulum on the RL control. By further numerous simulations and experiments the effects of parameter variations are evaluated. | ['Daniel A. Duecker', 'Alexej Bespalko', 'Leo Dostal'] | 2020-11-18 | null | null | null | null | ['acrobot'] | ['playing-games'] | [ 1.63260117e-01 4.64061826e-01 -2.87647724e-01 4.60109532e-01
5.58645308e-01 -6.22283161e-01 7.36073077e-01 1.23889931e-01
-8.44111204e-01 1.13005090e+00 -2.78347939e-01 -1.45838344e-02
-2.71969855e-01 -6.94193542e-01 -5.47422588e-01 -1.36258221e+00
2.70284619e-02 2.19748318e-01 -3.47544342e-01 -3.23864698e-01
1.60797834e-01 5.91000617e-01 -1.36523402e+00 -1.07979882e+00
4.94798243e-01 8.51722240e-01 4.49127585e-01 6.18002117e-01
7.54268765e-01 3.82944882e-01 -6.09393477e-01 4.68549937e-01
1.29221842e-01 -5.11335611e-01 -3.97617638e-01 1.12896860e-01
-6.00391507e-01 4.10361774e-03 -1.68510318e-01 8.47950637e-01
6.04008794e-01 5.25855958e-01 7.37443328e-01 -1.04689097e+00
1.75957143e-01 5.71289003e-01 -1.94379479e-01 -7.90265650e-02
-1.04474805e-01 2.85600811e-01 5.97395003e-01 8.81272852e-02
8.19310427e-01 6.38941348e-01 -6.41311556e-02 6.72831237e-01
-9.56637561e-01 -1.31401673e-01 -3.61846536e-01 5.82598113e-02
-1.07393003e+00 -1.17942058e-01 7.28499830e-01 -2.61306524e-01
4.81081277e-01 3.85364331e-02 1.15580189e+00 7.06862271e-01
7.09972024e-01 3.25687110e-01 9.00747597e-01 -8.47479045e-01
8.97477210e-01 1.42722964e-01 -3.89170021e-01 5.29898822e-01
5.36391318e-01 3.65686625e-01 -6.50802627e-02 1.91811427e-01
5.50360918e-01 -3.75622720e-01 -3.48008633e-01 -9.14170206e-01
-8.64499092e-01 5.65690696e-01 7.81600118e-01 4.20359671e-01
-6.09430850e-01 3.39898378e-01 1.08671643e-01 1.43121839e-01
-3.21034819e-01 9.88761663e-01 -1.75833821e-01 -2.30706900e-01
-3.01529855e-01 2.46634662e-01 8.39394629e-01 3.47983062e-01
1.99404031e-01 1.19781218e-01 2.09195718e-01 2.52241671e-01
9.14515778e-02 5.66313922e-01 5.43227673e-01 -7.01704502e-01
4.92496304e-02 7.39156067e-01 6.46540403e-01 -4.80415374e-01
-6.74372137e-01 -1.58850893e-01 -6.07419074e-01 4.54802096e-01
3.59741479e-01 -1.10114706e+00 -4.58412021e-01 1.66412854e+00
5.55864096e-01 -7.59601295e-01 3.93594414e-01 1.19524384e+00
1.60595030e-01 7.65162468e-01 1.04411148e-01 -6.20506287e-01
1.29635215e+00 -2.97904730e-01 -7.89606631e-01 -2.82769531e-01
2.79136300e-01 -2.31704026e-01 5.99307775e-01 -2.67776512e-02
-1.24541223e+00 -2.88726628e-01 -1.35776544e+00 4.63070840e-01
-4.85531867e-01 4.19708401e-01 1.72881380e-01 1.86745107e-01
-3.12369913e-01 6.98406458e-01 -1.07509947e+00 -3.34751099e-01
-3.77332926e-01 1.67694882e-01 9.47001725e-02 4.43502486e-01
-9.64024305e-01 1.37420952e+00 6.12754762e-01 1.48377761e-01
-4.98596638e-01 -2.50287950e-01 -5.64029455e-01 1.71583831e-01
4.58619386e-01 -8.26373339e-01 1.05385005e+00 -6.12511992e-01
-1.91647053e+00 4.32488829e-01 5.21920264e-01 -6.28831446e-01
5.73397815e-01 1.27025619e-01 3.30411315e-01 1.52097240e-01
-3.20185810e-01 6.38635278e-01 7.52301097e-01 -7.57654369e-01
-1.03870869e-01 -4.23888028e-01 -1.48002014e-01 7.71280825e-01
-1.93818271e-01 -6.87205017e-01 4.35382128e-01 -2.88996011e-01
-8.92141163e-02 -1.42151868e+00 -2.39761934e-01 -3.51532727e-01
-2.21933737e-01 -4.00359035e-01 6.56980097e-01 2.33102813e-01
4.59826410e-01 -1.88095653e+00 1.05717945e+00 -9.43187699e-02
-4.09667879e-01 7.17342421e-02 4.37574118e-01 9.16371465e-01
2.76055872e-01 -3.22138846e-01 -2.93338746e-02 3.64461720e-01
2.20438931e-03 5.37730157e-02 -1.31929189e-01 4.42274570e-01
2.43829414e-02 6.24793291e-01 -8.55259299e-01 1.11564212e-02
2.86950469e-01 4.25869226e-01 -2.34115645e-01 3.74492139e-01
-3.99169385e-01 5.13353705e-01 -8.13044608e-01 2.00453788e-01
-1.01675438e-02 1.23958983e-01 5.79421639e-01 -1.55351758e-01
-7.79191375e-01 -8.77586566e-03 -6.58452511e-01 1.12679207e+00
-5.23171544e-01 4.36229527e-01 2.01580450e-01 -9.63034987e-01
1.22765481e+00 6.74093366e-02 7.95167208e-01 -7.71172464e-01
6.25522435e-01 3.88152823e-02 7.80889988e-02 -4.21911895e-01
4.70789284e-01 -1.34058848e-01 -3.02481763e-02 2.52177030e-01
1.23434532e-02 -7.95394897e-01 4.08997655e-01 -1.59329519e-01
9.01381314e-01 1.79616034e-01 4.66131091e-01 -3.98128480e-01
4.36249703e-01 2.19521746e-01 6.67338297e-02 1.33101195e-01
-7.58125540e-03 -4.03540611e-01 4.57800567e-01 -2.26154312e-01
-1.13190854e+00 -6.03722990e-01 -2.18591928e-01 6.58493280e-01
6.59070432e-01 4.33302224e-02 -7.90053189e-01 2.02748373e-01
1.82646066e-01 8.18457484e-01 -4.28186029e-01 -4.68281090e-01
-4.63337541e-01 -1.05223513e+00 -4.65690851e-01 8.40493739e-02
5.65146446e-01 -1.38690138e+00 -1.98243916e+00 2.98080474e-01
-2.59354189e-02 -1.01636255e+00 1.90054193e-01 5.94518244e-01
-6.16058707e-01 -1.06019461e+00 -5.25365293e-01 -2.64005870e-01
5.38790584e-01 -3.86591107e-01 4.67753083e-01 -6.43114224e-02
-5.09573042e-01 5.81464708e-01 -3.68006498e-01 -6.80307269e-01
-3.17611426e-01 3.22274119e-01 1.31674528e-01 -3.06451291e-01
-1.19954139e-01 -4.49775428e-01 -6.53848648e-01 3.11672986e-01
-7.06685483e-01 8.01287293e-02 4.07928705e-01 5.88220775e-01
3.72608781e-01 2.71146387e-01 5.69049716e-01 1.89952508e-01
6.82191551e-01 -2.06436992e-01 -1.15419376e+00 -9.01262835e-02
-5.23363531e-01 3.52388799e-01 5.42220414e-01 -6.11098886e-01
-7.09604263e-01 2.66965419e-01 4.25666511e-01 1.67708248e-01
2.38905951e-01 1.13002270e-01 -3.57264988e-02 8.32083896e-02
2.67556518e-01 3.25366795e-01 1.64766759e-01 -1.19062185e-01
2.62201041e-01 2.72871852e-01 3.31872135e-01 -4.66301829e-01
5.03325641e-01 4.08509113e-02 6.54744327e-01 -9.71440017e-01
-3.58295031e-02 1.17559269e-01 -4.05868500e-01 -5.41273177e-01
8.48025739e-01 -2.29139224e-01 -1.33638144e+00 4.92857635e-01
-6.91842914e-01 -7.23383069e-01 -5.60706317e-01 5.03171146e-01
-9.12742198e-01 -3.57199520e-01 -2.19103456e-01 -1.24632478e+00
-6.62328601e-01 -9.10475075e-01 6.39351904e-01 8.52046311e-01
-1.48888335e-01 -6.00278199e-01 2.85852581e-01 -8.44129175e-02
3.37614894e-01 6.99845195e-01 1.00995338e+00 1.57828704e-01
-3.78268600e-01 -1.00224793e-01 4.25283849e-01 -9.16448385e-02
-2.98982292e-01 -1.48090899e-01 -1.87021434e-01 -6.05824888e-01
-8.52375105e-03 -4.72382545e-01 3.36404920e-01 5.27420938e-01
2.47027799e-01 -4.32795823e-01 -5.87756276e-01 2.53680591e-02
1.62662816e+00 4.90894496e-01 1.32503033e-01 5.53216279e-01
-1.00146858e-02 6.06836021e-01 7.46201873e-01 8.48525107e-01
1.47237197e-01 9.69213784e-01 8.01204026e-01 4.18237031e-01
2.64551580e-01 -7.81625882e-02 4.39473838e-01 1.48282751e-01
-2.61648297e-01 -2.16037244e-01 -5.46891868e-01 3.38098705e-01
-1.49285710e+00 -5.24294317e-01 4.25507754e-01 2.45182014e+00
4.38500255e-01 -1.08608499e-01 2.46791095e-01 3.49858135e-01
5.86961746e-01 7.98056498e-02 -8.01445425e-01 -5.70795417e-01
3.00787210e-01 -2.52457470e-01 8.52680445e-01 4.74813521e-01
-4.12472129e-01 3.96937698e-01 5.32373285e+00 1.87585324e-01
-1.64573824e+00 -6.38475120e-01 7.22088441e-02 -3.31224650e-01
2.08149627e-01 -1.68067515e-01 -6.18187368e-01 7.79884934e-01
6.31025612e-01 -4.13657874e-01 9.99115765e-01 6.37413085e-01
3.55018497e-01 -7.23833203e-01 -6.56572759e-01 2.82632411e-01
-4.91845340e-01 -9.34564710e-01 -5.69654524e-01 1.56110808e-01
3.77296120e-01 -8.65108818e-02 -2.35992059e-01 2.22743437e-01
-1.27822101e-01 -2.86588997e-01 7.26510882e-01 6.39729500e-01
2.76831150e-01 -8.03674817e-01 5.97516477e-01 8.38278294e-01
-8.36547673e-01 -6.51695073e-01 -3.23818564e-01 -2.44984776e-01
2.22597450e-01 4.05780047e-01 -1.06014740e+00 3.25734913e-01
2.16033533e-01 -4.65118606e-03 -1.92875192e-01 4.96394068e-01
-4.23231363e-01 1.48720279e-01 -6.49626017e-01 -1.24634171e+00
2.42407739e-01 -6.69490814e-01 7.48460889e-01 2.04060525e-01
1.31959483e-01 1.47059932e-01 -1.02436334e-01 1.00808108e+00
6.72002509e-02 1.42386220e-02 -6.12793088e-01 -3.28683734e-01
4.37218040e-01 1.42711914e+00 -7.84040868e-01 8.49008858e-02
6.32252574e-01 3.85775477e-01 7.48164626e-03 2.22704515e-01
-7.55774200e-01 -5.70676029e-01 3.70160043e-01 4.36707474e-02
1.43046811e-01 -2.53865421e-01 1.82075411e-01 -6.88669205e-01
2.27916911e-02 7.66434819e-02 -1.19358793e-01 -8.14888775e-01
-3.66079599e-01 1.38858646e-01 1.18583791e-01 -7.24124670e-01
-6.70495808e-01 -2.17109755e-01 -5.17850876e-01 2.91122705e-01
-1.11687338e+00 -3.61646384e-01 -1.71939865e-01 3.14741805e-02
9.25609022e-02 -2.01638751e-02 5.46115339e-01 -4.20287162e-01
-6.81102157e-01 -1.64426133e-01 4.88788277e-01 -2.43307784e-01
4.65235263e-02 -1.10961401e+00 -3.41247320e-01 3.38275045e-01
-5.21447182e-01 1.55943274e-01 1.19569409e+00 -2.14750469e-01
-1.96879089e+00 -4.90296841e-01 4.64199424e-01 2.64717817e-01
4.29413974e-01 -2.43743226e-01 -1.85049176e-01 1.29315615e-01
3.04749966e-01 -8.59097093e-02 -1.79767773e-01 -5.02833247e-01
8.53180945e-01 -3.76794696e-01 -1.16834927e+00 6.46407247e-01
3.23052943e-01 1.86178610e-01 -2.23888859e-01 1.54552057e-01
3.41420859e-01 -4.56981063e-01 -7.74064541e-01 5.43864906e-01
7.80739605e-01 -4.06580269e-01 6.32335901e-01 -3.66272807e-01
1.44218013e-01 -4.08567563e-02 1.82453290e-01 -1.57797670e+00
1.54094130e-01 -5.83585739e-01 -3.55552763e-01 7.79864252e-01
-1.01731047e-01 -4.92314428e-01 6.02757335e-01 5.46428375e-02
4.05219913e-01 -9.08796310e-01 -1.04012156e+00 -4.34918910e-01
-5.08888885e-02 7.24928141e-01 1.48575991e-01 4.05491233e-01
6.26874089e-01 6.04220331e-01 1.16536655e-02 1.78364351e-01
3.75071943e-01 1.86335176e-01 5.43047667e-01 -8.71862292e-01
-8.11002329e-02 -2.93834001e-01 -1.60904288e-01 -5.40268719e-01
-6.05467223e-02 -4.05909836e-01 1.79566190e-01 -1.37393057e+00
-3.70254010e-01 -2.51378477e-01 -6.87866285e-02 1.24164678e-01
1.03281066e-01 -1.49799153e-01 3.73574853e-01 -1.00827649e-01
-5.60348295e-02 7.35807419e-01 1.40755177e+00 7.65088275e-02
-7.90739894e-01 3.89001280e-01 -5.48558123e-02 2.55499005e-01
9.96471345e-01 -3.86949688e-01 -5.40706277e-01 1.88220307e-01
3.28594238e-01 5.84636092e-01 1.78471327e-01 -1.01695013e+00
2.04584986e-01 -1.97034016e-01 1.52262747e-01 -2.05243483e-01
3.22296172e-01 -7.34871328e-01 2.08484143e-01 1.16051078e+00
-3.73695403e-01 -8.03467259e-03 -1.45582587e-01 4.91693825e-01
2.29802996e-01 -5.24628341e-01 1.00467026e+00 -4.20397110e-02
-1.29251674e-01 -3.03568184e-01 -8.17712665e-01 -3.50000679e-01
1.26913190e+00 3.87040138e-01 8.07213411e-03 -2.33336225e-01
-4.72542077e-01 3.09026986e-01 5.61858296e-01 1.58482447e-01
-1.76606491e-01 -8.97868216e-01 -1.85945570e-01 2.88024664e-01
-2.69889116e-01 -1.16867065e-01 -8.36139098e-02 7.58418679e-01
-4.11942124e-01 3.90468568e-01 -7.45002985e-01 -4.12783682e-01
-1.07770371e+00 6.01508737e-01 7.55880773e-01 -2.65070230e-01
-3.57702941e-01 -2.64164954e-02 -6.73844397e-01 -1.48786649e-01
1.32464692e-01 -3.10773075e-01 -2.12348163e-01 1.99615344e-01
-8.94972980e-02 3.91049504e-01 -1.60579965e-01 -6.26026988e-02
-1.95232406e-01 6.22307837e-01 6.89216554e-01 -1.34299815e-01
1.35966480e+00 -2.26185527e-02 2.96115708e-02 1.46175519e-01
4.90666866e-01 -3.28364849e-01 -1.37290311e+00 4.35972273e-01
-2.07813129e-01 2.77008384e-01 1.79340720e-01 -8.90327275e-01
-4.12193000e-01 4.92494643e-01 4.72155750e-01 5.46528697e-01
1.08005857e+00 -2.60755092e-01 1.43811926e-01 7.08541214e-01
4.40780193e-01 -1.23557091e+00 1.49155244e-01 2.72471428e-01
7.66359329e-01 -6.05969667e-01 1.67628139e-01 2.56538779e-01
-6.02574468e-01 1.29750991e+00 4.08786446e-01 -4.43570793e-01
3.28057736e-01 1.05587386e-01 -2.95202821e-01 3.11417758e-01
-6.34331524e-01 1.01057924e-02 -8.12711418e-02 1.51548758e-01
4.77589294e-02 3.11428756e-01 -8.82665396e-01 7.80556649e-02
-2.64365584e-01 6.99516833e-02 7.35314429e-01 9.21598375e-01
-7.70584166e-01 -7.40634918e-01 -3.27899069e-01 -4.33740735e-01
8.50286987e-03 5.98850608e-01 -4.00562406e-01 1.20233560e+00
1.72146618e-01 5.49542189e-01 1.78988397e-01 2.45251447e-01
3.81105423e-01 2.76121125e-02 6.83490813e-01 -6.01279996e-02
-1.00527495e-01 -1.06543310e-01 -1.60970852e-01 -5.50047727e-03
-3.23334426e-01 -4.75864351e-01 -1.41875613e+00 8.57932642e-02
-5.73603570e-01 5.61062872e-01 1.30003119e+00 9.32030916e-01
2.12548431e-02 3.52664083e-01 8.74349236e-01 -9.00650263e-01
-8.18049967e-01 -1.07108557e+00 -5.31034350e-01 -2.78634667e-01
4.38014090e-01 -6.71815038e-01 -4.26265657e-01 -3.88537169e-01] | [4.553380012512207, 2.0878703594207764] |
f5818613-896e-416e-b025-b283d6f52398 | on-the-benefits-of-biophysical-synapses | 2303.04944 | null | https://arxiv.org/abs/2303.04944v1 | https://arxiv.org/pdf/2303.04944v1.pdf | On the Benefits of Biophysical Synapses | The approximation capability of ANNs and their RNN instantiations, is strongly correlated with the number of parameters packed into these networks. However, the complexity barrier for human understanding, is arguably related to the number of neurons and synapses in the networks, and to the associated nonlinear transformations. In this paper we show that the use of biophysical synapses, as found in LTCs, have two main benefits. First, they allow to pack more parameters for a given number of neurons and synapses. Second, they allow to formulate the nonlinear-network transformation, as a linear system with state-dependent coefficients. Both increase interpretability, as for a given task, they allow to learn a system linear in its input features, that is smaller in size compared to the state of the art. We substantiate the above claims on various time-series prediction tasks, but we believe that our results are applicable to any feedforward or recurrent ANN. | ['Radu Grosu', 'Julian Lemmel'] | 2023-03-08 | null | null | null | null | ['time-series-prediction'] | ['time-series'] | [ 5.06264567e-01 2.93762892e-01 3.39321285e-01 1.04840770e-01
3.13828677e-01 -8.06819379e-01 6.29311085e-01 9.99612138e-02
-5.18794775e-01 8.91102910e-01 -1.83059067e-01 -4.95699316e-01
-4.23767865e-01 -6.31613791e-01 -7.87807226e-01 -8.67814422e-01
-1.11374870e-01 3.37228209e-01 2.60172367e-01 -3.04377884e-01
2.74512060e-02 8.94882917e-01 -1.53601384e+00 3.48714702e-02
6.77492678e-01 1.23847580e+00 2.04771891e-01 6.62801981e-01
2.51398683e-01 7.26316392e-01 -2.61628121e-01 -3.91999066e-01
2.42972404e-01 -2.85008460e-01 -5.93408942e-01 -4.95438784e-01
1.82215404e-02 1.23994239e-01 -4.14705932e-01 5.59057653e-01
2.43144512e-01 2.09129080e-01 9.85796630e-01 -9.34830964e-01
-4.89494026e-01 5.44951618e-01 2.04441085e-01 3.20334196e-01
-3.70182186e-01 1.35163188e-01 1.00311542e+00 -5.79602897e-01
2.80401647e-01 9.65018272e-01 9.22829390e-01 7.12832332e-01
-1.67520988e+00 -4.56941187e-01 1.11043319e-01 -6.23879619e-02
-1.16233087e+00 -5.70058167e-01 5.28557956e-01 -3.68282884e-01
1.33225298e+00 2.96249270e-01 1.05950499e+00 1.11070478e+00
3.65524203e-01 2.07834855e-01 9.92904603e-01 -5.27547657e-01
4.21315283e-01 1.07812881e-01 1.00377329e-01 3.98736447e-01
4.76686388e-01 1.47145599e-01 -3.50521892e-01 1.04463613e-02
1.03575206e+00 2.04424217e-01 -1.63423747e-01 -8.63304213e-02
-8.58843029e-01 6.39950991e-01 4.79981631e-01 7.94161916e-01
-5.68074048e-01 5.20121336e-01 3.23523819e-01 4.97790664e-01
3.24321628e-01 8.61636162e-01 -7.52579808e-01 -9.88913253e-02
-8.81033897e-01 2.35194176e-01 1.07086062e+00 4.79826570e-01
7.20410347e-01 2.15233266e-01 3.45205516e-01 5.78051448e-01
1.98434785e-01 5.38117945e-01 5.62096715e-01 -9.66046453e-01
1.94436148e-01 6.54663801e-01 -1.56598061e-01 -9.43379998e-01
-6.88959122e-01 -7.24114001e-01 -1.23218584e+00 1.11611009e-01
5.04891276e-01 -1.35642186e-01 -6.23814821e-01 1.99721086e+00
-2.28785008e-01 -2.27199383e-02 2.44772136e-01 5.52703738e-01
2.82892495e-01 7.43698716e-01 -1.17037378e-01 -4.01589632e-01
1.33294404e+00 -5.46709538e-01 -6.36443794e-01 -4.16533262e-01
4.90299314e-01 -2.83309221e-01 7.64058053e-01 3.78832459e-01
-1.18625009e+00 -4.22152638e-01 -8.89625490e-01 -1.32984698e-01
-7.36765862e-01 1.53411984e-01 6.89272881e-01 3.77193719e-01
-1.30029905e+00 9.14369643e-01 -1.03679073e+00 -4.79275674e-01
1.28061876e-01 7.96216547e-01 -2.71615759e-02 6.95874989e-01
-1.28589427e+00 1.29132903e+00 5.44032574e-01 4.31729615e-01
-2.86633849e-01 -6.59302473e-01 -3.43551397e-01 3.78079712e-01
-8.14087614e-02 -8.79776239e-01 1.04426706e+00 -8.88579309e-01
-1.67397034e+00 6.05984807e-01 -1.88383609e-01 -1.00242567e+00
2.31556371e-01 1.06207304e-01 -2.84829259e-01 -1.64421741e-02
-7.02580214e-01 6.92182243e-01 6.72376335e-01 -1.07560098e+00
-6.00746274e-02 -1.54951245e-01 6.64991811e-02 -5.46678044e-02
-3.46691370e-01 -3.94732833e-01 2.52017200e-01 -6.00665808e-01
2.71804780e-01 -1.21978712e+00 -4.32588488e-01 1.04307838e-01
-1.02725528e-01 1.32850325e-02 6.57965064e-01 -4.11203265e-01
1.27354574e+00 -1.97675514e+00 2.38198191e-01 4.35190529e-01
2.49857724e-01 5.27494550e-01 -5.20891957e-02 7.46613860e-01
-2.36209273e-01 1.89464614e-01 -2.81378984e-01 -1.37503162e-01
-6.10102639e-02 5.67934155e-01 -6.60478413e-01 1.57960728e-01
3.65704954e-01 1.27739716e+00 -4.21172112e-01 -7.59594142e-02
2.62144879e-02 6.70188963e-01 -3.66118848e-01 -5.75561523e-02
-1.48979843e-01 2.89826632e-01 -5.30998826e-01 -1.81335986e-01
2.52719820e-01 -3.08630586e-01 1.94231153e-01 -1.77580014e-01
-4.30491716e-01 6.60911381e-01 -8.49008918e-01 1.21989381e+00
-6.58692002e-01 9.66988564e-01 -2.34081745e-01 -1.27432668e+00
9.11914170e-01 4.91217673e-01 3.64514768e-01 -6.06612265e-01
9.92909372e-02 4.23880756e-01 4.52077061e-01 -1.30688801e-01
6.72322349e-04 -2.17034668e-01 2.84022897e-01 5.73873401e-01
1.99280769e-01 -2.03656971e-01 2.55802840e-01 -2.73390889e-01
1.07179010e+00 -5.51433489e-03 3.45262736e-01 -4.84108001e-01
4.49235499e-01 -5.70102513e-01 1.49639979e-01 6.52144611e-01
4.08422232e-01 1.50112778e-01 7.91082263e-01 -5.71671546e-01
-1.29778552e+00 -8.63016248e-01 -3.83909404e-01 6.01181030e-01
-5.57427824e-01 -7.54758641e-02 -8.35610032e-01 2.51373857e-01
-1.34598151e-01 4.06655014e-01 -5.63523650e-01 -3.19723755e-01
-7.46977746e-01 -7.00278223e-01 8.62046957e-01 4.25316393e-01
7.54911602e-02 -1.21328187e+00 -1.21260381e+00 4.58452046e-01
2.26758927e-01 -1.02668047e+00 4.66787964e-02 9.18005228e-01
-1.51356995e+00 -8.07230175e-01 -4.30261225e-01 -6.14246249e-01
6.76437736e-01 -3.16345721e-01 8.63196969e-01 7.69811869e-02
-7.80626461e-02 1.70251638e-01 7.98065513e-02 -5.42783022e-01
-4.73001719e-01 3.61121088e-01 2.83235013e-01 -1.87571675e-01
1.34783769e-02 -1.29563284e+00 -5.22494674e-01 1.10619031e-02
-1.11211693e+00 1.55224264e-01 9.18448150e-01 8.27127993e-01
4.48740482e-01 -4.56624068e-02 6.02410495e-01 -9.33524609e-01
7.21844494e-01 -3.26243900e-02 -5.57478786e-01 3.32044959e-01
-9.78023529e-01 4.95823503e-01 9.95192766e-01 -5.22745848e-01
-5.74074924e-01 6.63947761e-02 -8.96840170e-02 -6.05320074e-02
8.96999687e-02 6.68055832e-01 3.94067585e-01 -2.34366030e-01
7.17704952e-01 5.24270177e-01 1.64114282e-01 -4.40421104e-01
3.59662086e-01 1.45733669e-01 3.24573219e-01 -4.86613423e-01
7.68713713e-01 1.35424986e-01 5.75987637e-01 -8.70311201e-01
-4.39994276e-01 -2.21289769e-02 -9.27734017e-01 -3.08205932e-02
5.60772955e-01 -5.27244568e-01 -1.17514396e+00 4.05714780e-01
-1.39341056e+00 -4.31704223e-01 -6.19770646e-01 3.37907672e-01
-7.96299517e-01 -1.16915993e-01 -9.04011846e-01 -1.21065807e+00
-3.07692319e-01 -7.68044293e-01 5.22218049e-01 2.28790969e-01
-3.11117530e-01 -1.39784002e+00 2.99356692e-02 -3.98656964e-01
6.82423055e-01 1.25883490e-01 1.21793067e+00 -6.69865072e-01
-3.52570891e-01 -1.64672434e-01 -4.72296774e-02 4.06998575e-01
-2.83565335e-02 2.36559644e-01 -1.09405768e+00 -1.72298655e-01
4.03476089e-01 -1.58946723e-01 8.17224145e-01 4.16978419e-01
1.05005860e+00 -8.01089406e-01 -2.89016694e-01 2.82186270e-01
1.27972913e+00 2.22451970e-01 5.50792217e-01 2.67857313e-02
4.04311687e-01 6.09660029e-01 -3.03527534e-01 1.31916165e-01
-1.87581219e-02 4.22922164e-01 2.99330235e-01 -2.08471105e-01
1.89904451e-01 -2.57267624e-01 2.45098770e-01 1.47827113e+00
-6.61236584e-01 -3.09151798e-01 -9.00888622e-01 1.06961615e-01
-1.99503112e+00 -1.00378919e+00 -8.01803619e-02 2.12724710e+00
8.52520525e-01 2.31437564e-01 8.38474259e-02 4.66960281e-01
2.75008708e-01 -9.94582698e-02 -8.32044780e-01 -5.78357816e-01
-4.78632241e-01 4.77777958e-01 5.91793418e-01 2.29292959e-01
-5.01351118e-01 5.10876477e-01 7.55547953e+00 6.03714705e-01
-1.28509057e+00 -1.16585366e-01 6.80542827e-01 -2.12407500e-01
-1.87881723e-01 5.44849737e-03 -5.59691489e-01 3.52982879e-01
1.69116604e+00 -1.02089113e-02 9.81611073e-01 2.17801198e-01
3.10790181e-01 1.87622935e-01 -1.41598165e+00 8.94108713e-01
-3.62195760e-01 -1.29935801e+00 7.67375231e-02 2.81032115e-01
4.67927337e-01 4.63297032e-02 2.13668406e-01 1.32273495e-01
-1.34432614e-01 -1.25668812e+00 7.34964967e-01 9.04339433e-01
3.45472246e-01 -5.08472025e-01 5.14240086e-01 5.73631644e-01
-9.30554271e-01 -1.60884887e-01 -5.45408666e-01 -5.52263260e-01
-1.04671858e-01 6.12558782e-01 -4.05648142e-01 8.84256512e-02
4.16764200e-01 6.33714080e-01 -5.78273892e-01 6.66010499e-01
1.14497226e-02 7.79887676e-01 -7.23707020e-01 -4.41320717e-01
2.32025594e-01 -5.21469712e-01 1.70200467e-01 1.17243457e+00
4.38033313e-01 1.79305702e-01 -5.74350119e-01 1.12244725e+00
1.55250743e-01 -1.63283706e-01 -5.62907517e-01 -3.35855514e-01
4.80115443e-01 1.18585563e+00 -6.67391717e-01 -5.68036772e-02
-1.66394621e-01 5.99871397e-01 6.34448707e-01 6.11349404e-01
-5.53055584e-01 -2.66014278e-01 3.91619474e-01 1.75865009e-01
2.86489785e-01 -3.61093014e-01 -3.48092735e-01 -9.16610956e-01
2.17667907e-01 -8.08239222e-01 -1.90074444e-01 -6.55111432e-01
-1.12188947e+00 5.30222535e-01 -2.44416431e-01 -1.02402449e+00
-4.19783890e-01 -1.18837821e+00 -5.20196676e-01 9.03833687e-01
-1.30587244e+00 -8.49565566e-01 4.06686693e-01 2.75500029e-01
-6.66315109e-02 1.69419080e-01 1.15897381e+00 1.94956765e-01
-5.31107962e-01 3.49484324e-01 3.03108841e-01 -1.71432450e-01
7.32085556e-02 -9.90204692e-01 4.11053002e-01 5.11629581e-01
2.18730599e-01 1.00816011e+00 8.22028339e-01 2.36703660e-02
-1.41126442e+00 -7.25720882e-01 1.08906198e+00 -4.01814491e-01
9.30307150e-01 -7.15930998e-01 -1.16320825e+00 7.31747091e-01
6.87574372e-02 1.41573446e-02 4.93662119e-01 2.44885787e-01
-1.72666863e-01 -2.48637632e-01 -7.59566903e-01 7.38824964e-01
1.14756334e+00 -7.77632594e-01 -4.68696505e-01 1.58094421e-01
5.75146496e-01 -2.35172316e-01 -1.05525887e+00 1.20187409e-01
9.46987331e-01 -9.34313715e-01 9.52771008e-01 -6.13308072e-01
2.68622965e-01 -1.12482697e-01 1.05804011e-01 -1.42821145e+00
-5.30430019e-01 -5.07591009e-01 -4.67189938e-01 1.00560629e+00
8.50655079e-01 -1.29545176e+00 4.75860626e-01 8.07339370e-01
4.41693440e-02 -1.14002812e+00 -1.15802288e+00 -7.69498646e-01
2.85228580e-01 -1.39382929e-01 3.46237510e-01 5.45708299e-01
9.72020328e-02 5.93933105e-01 -1.17780469e-01 -6.34445027e-02
6.08301312e-02 -2.04691872e-01 1.55852303e-01 -1.84096444e+00
-4.34255660e-01 -6.79406226e-01 -3.88191432e-01 -1.17785203e+00
1.05079159e-01 -6.85441792e-01 -1.59101233e-01 -1.06635809e+00
-1.82578757e-01 -7.10983336e-01 -5.28564215e-01 6.05177939e-01
3.38508695e-01 1.35617152e-01 2.68037915e-01 6.04810417e-01
-1.93601940e-03 2.43116572e-01 1.04451168e+00 1.43642336e-01
-2.10393146e-01 9.11018401e-02 -4.47575539e-01 6.48098528e-01
8.20284605e-01 -5.47799528e-01 -4.36202019e-01 -3.40598315e-01
8.00947845e-01 7.08670542e-02 4.68759060e-01 -1.15860856e+00
4.31588918e-01 1.04189888e-01 5.04003882e-01 -3.38129908e-01
6.32854462e-01 -1.01419699e+00 4.70055908e-01 9.38364565e-01
-5.76880395e-01 3.90173823e-01 2.97348887e-01 4.46353316e-01
-1.21447220e-01 -4.82417256e-01 6.24369085e-01 -8.11449364e-02
-1.54727213e-02 1.39119714e-01 -7.52379775e-01 -1.80418581e-01
5.95237017e-01 -3.79090607e-01 -1.92802116e-01 -2.51547128e-01
-7.43609965e-01 -2.29115769e-01 1.83130458e-01 2.77706124e-02
2.45341241e-01 -1.09715533e+00 -2.22229183e-01 2.79674232e-01
-2.96177268e-01 -8.83543491e-02 -1.14499271e-01 9.86683309e-01
-4.24577713e-01 9.06939447e-01 -3.34934801e-01 -3.67444009e-01
-7.50709355e-01 4.68883127e-01 5.78504562e-01 -4.49252605e-01
-4.35569733e-01 3.80440295e-01 1.29656494e-01 -2.65444905e-01
-4.18824777e-02 -9.60157096e-01 -1.74278855e-01 2.54826155e-02
6.36547878e-02 3.05415392e-01 1.21321559e-01 -2.58952618e-01
-2.54455656e-01 6.60402477e-01 2.30578482e-01 -8.95881131e-02
1.42228901e+00 -3.77462506e-02 -3.62305015e-01 9.85956788e-01
8.62503767e-01 -3.96592885e-01 -1.25904858e+00 -2.24292934e-01
5.09288982e-02 4.44440067e-01 -9.24339741e-02 -7.49802649e-01
-8.53241146e-01 1.00144303e+00 5.39421022e-01 8.08194041e-01
1.10024798e+00 -2.84654498e-01 4.87631500e-01 1.04412985e+00
1.30272761e-01 -8.97576153e-01 -3.71078700e-01 6.99244976e-01
8.21216583e-01 -5.98042846e-01 -8.38481784e-02 -2.52608716e-01
-1.34599909e-01 1.41591239e+00 1.87377289e-01 -3.83002669e-01
5.18857718e-01 4.64044958e-01 -1.26074925e-01 4.57224362e-02
-1.11849487e+00 6.87199533e-02 3.26544225e-01 5.28354049e-01
7.13389277e-01 -1.34012476e-01 -4.28407162e-01 5.00667572e-01
-4.49691445e-01 2.38312483e-02 3.38855267e-01 4.49717343e-01
-5.45161664e-01 -1.04267311e+00 1.02792747e-01 5.52837193e-01
-2.43507743e-01 -2.68335015e-01 -3.39126617e-01 7.79758692e-01
6.13105372e-02 6.91696584e-01 2.03394830e-01 6.84949511e-04
4.54530120e-01 2.96055704e-01 6.86428249e-01 -2.92852402e-01
-8.36228311e-01 -2.20558047e-01 -2.30878547e-01 -1.06637731e-01
-4.02656764e-01 -6.03099883e-01 -1.24008775e+00 -2.77441859e-01
-2.89535969e-01 1.86066166e-01 6.79855943e-01 1.06550443e+00
5.12277007e-01 6.55470550e-01 2.38797128e-01 -9.34324682e-01
-6.84737682e-01 -9.77372646e-01 -6.26166880e-01 1.20137952e-01
4.51345116e-01 -4.71447706e-01 -3.84189874e-01 9.04092044e-02] | [7.8981122970581055, 3.3096773624420166] |
f3329718-d068-4e61-b227-01c42c29f391 | an-incremental-phase-mapping-approach-for-x | 2211.04011 | null | https://arxiv.org/abs/2211.04011v1 | https://arxiv.org/pdf/2211.04011v1.pdf | An Incremental Phase Mapping Approach for X-ray Diffraction Patterns using Binary Peak Representations | Despite the huge advancement in knowledge discovery and data mining techniques, the X-ray diffraction (XRD) analysis process has mostly remained untouched and still involves manual investigation, comparison, and verification. Due to the large volume of XRD samples from high-throughput XRD experiments, it has become impossible for domain scientists to process them manually. Recently, they have started leveraging standard clustering techniques, to reduce the XRD pattern representations requiring manual efforts for labeling and verification. Nevertheless, these standard clustering techniques do not handle problem-specific aspects such as peak shifting, adjacent peaks, background noise, and mixed phases; hence, resulting in incorrect composition-phase diagrams that complicate further steps. Here, we leverage data mining techniques along with domain expertise to handle these issues. In this paper, we introduce an incremental phase mapping approach based on binary peak representations using a new threshold based fuzzy dissimilarity measure. The proposed approach first applies an incremental phase computation algorithm on discrete binary peak representation of XRD samples, followed by hierarchical clustering or manual merging of similar pure phases to obtain the final composition-phase diagram. We evaluate our method on the composition space of two ternary alloy systems- Co-Ni-Ta and Co-Ti-Ta. Our results are verified by domain scientists and closely resembles the manually computed ground-truth composition-phase diagrams. The proposed approach takes us closer towards achieving the goal of complete end-to-end automated XRD analysis. | ['Ankit Agrawal', 'Michael Bedzyk', 'Yip-Wah Chung', 'Alok Choudhary', 'Wei-keng Liao', 'Denis T. Keane', 'Justin Liao', 'Ruifeng Zhang', 'K. V. L. V. Narayanachari', 'Dipendra Jha'] | 2022-11-08 | null | null | null | null | ['x-ray-diffraction'] | ['miscellaneous'] | [ 3.84803772e-01 1.88141428e-02 -2.16688484e-01 -1.65469259e-01
-7.47261107e-01 -2.29658052e-01 4.10340160e-01 7.48778284e-01
-2.31255040e-01 6.83390737e-01 -3.88954103e-01 -4.16035563e-01
-6.03558660e-01 -9.67567682e-01 -1.88620344e-01 -8.65410209e-01
2.74978757e-01 1.09501266e+00 3.75901669e-01 1.78487763e-01
6.15638256e-01 5.14518917e-01 -1.90766966e+00 3.56332392e-01
1.02821863e+00 1.08190846e+00 2.54670203e-01 3.51278752e-01
-5.09284139e-01 7.50501513e-01 -3.40036124e-01 1.12420954e-01
1.75726980e-01 -4.82289165e-01 -6.82905138e-01 2.71941692e-01
-9.61802825e-02 1.27258927e-01 3.09658408e-01 1.10791719e+00
4.27609161e-02 8.11707154e-02 9.06969547e-01 -1.27592444e+00
-5.49057782e-01 5.57774901e-01 -8.87777925e-01 -1.91406012e-01
5.02680898e-01 -6.79339394e-02 1.02945209e+00 -6.17947996e-01
7.27437615e-01 6.83294475e-01 4.60987359e-01 1.14201941e-02
-1.14890766e+00 -6.27279282e-01 -1.85370594e-01 6.85676455e-01
-1.48299372e+00 -6.89476058e-02 9.99096274e-01 -5.27801275e-01
1.02337086e+00 2.65415192e-01 6.20144665e-01 4.31058049e-01
2.15935446e-02 1.67994842e-01 1.55141330e+00 -6.42454028e-01
7.29828954e-01 -7.42851570e-02 3.24200213e-01 3.42000902e-01
5.03643215e-01 -2.69738466e-01 -5.98867312e-02 -6.42410219e-02
7.99558237e-02 1.81261197e-01 -9.61521491e-02 -6.50157452e-01
-6.94365799e-01 4.90082443e-01 9.21455473e-02 6.83055162e-01
-6.27593577e-01 -4.28989768e-01 1.22073842e-02 2.00724810e-01
6.56174570e-02 5.19764721e-01 -1.58051953e-01 -4.95008603e-02
-1.43422866e+00 4.72594619e-01 6.31867170e-01 8.95336330e-01
1.12957573e+00 -3.72793168e-01 5.07591903e-01 5.82022846e-01
3.65979165e-01 3.38980734e-01 5.12051105e-01 -7.38432050e-01
1.50845319e-01 1.02105391e+00 1.70382708e-01 -9.97929335e-01
-3.61056685e-01 7.93718249e-02 -6.41852975e-01 3.12617838e-01
2.96021730e-01 3.91509891e-01 -1.05935824e+00 9.22057331e-01
4.29544121e-01 -3.70188177e-01 -5.90586588e-02 8.10240746e-01
3.55927587e-01 3.44034225e-01 1.77511033e-02 -5.99855959e-01
1.27874911e+00 -5.84575832e-01 -6.60546839e-01 3.33457023e-01
3.86245042e-01 -8.00668836e-01 1.00473750e+00 7.49930918e-01
-9.55127656e-01 -3.24808747e-01 -1.45736957e+00 3.08051944e-01
-4.51197654e-01 -2.46013299e-01 3.59031826e-01 6.57184541e-01
-5.04861414e-01 8.85558128e-01 -8.74774575e-01 -3.45228344e-01
5.22538960e-01 5.68718672e-01 -3.14152837e-01 -2.61517495e-01
-5.76449752e-01 5.75384080e-01 3.77429515e-01 3.88649898e-03
-1.28179118e-01 -7.47288644e-01 -3.90834004e-01 -1.67303190e-01
7.93791652e-01 -1.39666066e-01 1.19893253e+00 -6.14544749e-01
-1.23309565e+00 5.43298125e-01 -1.15694113e-01 -1.45010054e-01
1.85296476e-01 2.44796962e-01 -6.78601682e-01 2.10229710e-01
2.42486224e-01 2.86959596e-02 4.88669157e-01 -1.45003128e+00
-8.64104688e-01 -5.06463885e-01 -4.51885700e-01 -1.10243030e-01
3.04369386e-02 -2.15630040e-01 -4.50666249e-02 -9.61507037e-02
6.28762245e-01 -6.22966886e-01 -2.55521536e-01 -4.72743183e-01
-2.98930854e-01 -2.14917943e-01 9.37625051e-01 -4.05942589e-01
1.32416606e+00 -1.84809446e+00 -1.26357689e-01 9.49795663e-01
4.28775996e-01 -6.71926811e-02 6.01208568e-01 4.77120101e-01
-3.02596122e-01 -5.99386636e-03 -7.42960751e-01 2.43111685e-01
1.26558810e-01 1.58452407e-01 2.11401522e-01 4.57231492e-01
1.12447269e-01 3.31899345e-01 -8.64099681e-01 -5.25301337e-01
2.45654196e-01 -3.59751657e-02 -4.56677943e-01 2.72386502e-02
-6.03443444e-01 9.52317268e-02 -3.91552001e-01 1.01354444e+00
9.55820024e-01 -2.21198171e-01 4.99578238e-01 -5.49739361e-01
-4.75820303e-01 1.26603350e-01 -1.43761575e+00 1.30859029e+00
1.50237992e-01 1.09201305e-01 -1.52199809e-02 -1.37631714e+00
1.05784452e+00 1.91588193e-01 9.39554572e-01 -1.09758162e+00
2.85042137e-01 4.67986524e-01 3.33305746e-01 -4.95987356e-01
4.87124741e-01 -2.38096073e-01 8.45505893e-02 9.25572813e-01
-2.90859461e-01 -1.47457510e-01 4.42458004e-01 -8.37939009e-02
1.03918862e+00 9.87970680e-02 4.47039157e-01 -3.90499532e-01
8.02054584e-01 7.07460225e-01 2.95069247e-01 1.32972151e-01
5.71378991e-02 6.46786869e-01 3.42459947e-01 -2.62916416e-01
-1.46158004e+00 -9.97777700e-01 -4.10806425e-02 3.78806114e-01
1.34805575e-01 -5.63011229e-01 -7.40389347e-01 -2.28181973e-01
-8.95159543e-02 5.13437212e-01 -3.34760845e-01 -1.90358851e-02
-5.24188757e-01 -8.99708807e-01 1.10687204e-01 2.79118210e-01
3.16011459e-01 -8.94819558e-01 -5.92841566e-01 3.80687803e-01
1.95701897e-01 -7.97845066e-01 1.28818512e-01 3.73334169e-01
-6.24456167e-01 -1.60933137e+00 -2.67791271e-01 -6.35955274e-01
7.28944719e-01 1.16000161e-01 6.86843276e-01 1.84571594e-01
-7.10284293e-01 1.53577045e-01 -4.78298455e-01 -1.67652205e-01
-5.81215322e-01 -1.16971880e-02 -2.99837552e-02 -1.93666860e-01
7.13129520e-01 -6.75462723e-01 -5.50160289e-01 3.05470765e-01
-9.59338784e-01 -7.91065991e-02 5.87969363e-01 4.51210350e-01
9.54207003e-01 8.97648692e-01 3.44007015e-01 -1.05542815e+00
6.95468843e-01 -4.25014913e-01 -5.52706361e-01 4.66807634e-01
-1.21930528e+00 1.98820174e-01 7.40920842e-01 -1.62424788e-01
-9.06694651e-01 1.53680993e-02 -2.57777069e-02 -2.98391908e-01
-3.55818748e-01 6.73485160e-01 -3.62148166e-01 1.46092907e-01
4.86523062e-01 1.50841773e-02 -3.42004038e-02 -3.71784091e-01
6.53808638e-02 1.07009292e+00 6.72050834e-01 -6.58712983e-01
7.96011508e-01 3.44827503e-01 1.57322764e-01 -7.50598431e-01
-2.39974499e-01 -4.57718909e-01 -7.24591792e-01 -1.63638860e-01
9.03678119e-01 -3.85679811e-01 -8.75756264e-01 1.85403556e-01
-4.93003368e-01 8.41585323e-02 -6.00506291e-02 2.96349436e-01
-2.94391602e-01 7.25238681e-01 -3.10579658e-01 -7.08835542e-01
-3.13194871e-01 -1.32927632e+00 6.51180565e-01 1.37102932e-01
-5.40014029e-01 -4.90234852e-01 4.57670167e-02 5.41216254e-01
2.36731574e-01 2.23486543e-01 1.45532906e+00 -6.79373503e-01
-5.30321360e-01 -7.54221454e-02 -1.67008147e-01 4.35692146e-02
6.61157846e-01 4.30955619e-01 -6.77149355e-01 1.64094150e-01
9.01159570e-02 7.93012008e-02 2.53563315e-01 5.67038171e-02
9.27889109e-01 2.27328509e-01 -3.65399957e-01 -3.48358564e-02
1.35592937e+00 7.80203760e-01 6.62815392e-01 6.24471426e-01
7.18010902e-01 5.96636355e-01 1.04481971e+00 4.62004870e-01
3.56375366e-01 4.94644046e-01 3.87203805e-02 1.96572632e-01
1.72932893e-01 4.62394953e-02 -1.27083987e-01 8.34391952e-01
-3.09048891e-01 4.26222235e-02 -1.23274529e+00 3.12560111e-01
-1.67098010e+00 -8.46405864e-01 -3.52865070e-01 2.00413227e+00
7.72650063e-01 3.29866499e-01 2.55638778e-01 9.39499974e-01
7.49956906e-01 -4.30289477e-01 -3.91557336e-01 -3.03057879e-01
1.06970720e-01 7.71119177e-01 5.02913117e-01 2.97351599e-01
-7.42794037e-01 4.87227321e-01 5.57170582e+00 5.72458625e-01
-1.22938299e+00 -2.70750940e-01 2.65529186e-01 7.49989226e-02
-3.91086191e-01 2.45133653e-01 -3.22273642e-01 5.76228917e-01
6.99336171e-01 -3.22308272e-01 4.48138833e-01 7.08920002e-01
1.57088503e-01 -5.95574021e-01 -1.07694733e+00 9.74544227e-01
-1.76310048e-01 -1.32017624e+00 -1.33718118e-01 1.60307631e-01
3.81464034e-01 -4.79450822e-01 -3.39965701e-01 -1.68258116e-01
2.88034678e-01 -8.40007484e-01 6.09732509e-01 3.70085448e-01
5.48375428e-01 -9.01578665e-01 5.63491881e-01 4.64782007e-02
-1.14681971e+00 -6.70239106e-02 -1.05917677e-01 5.11858389e-02
-3.00269537e-02 6.38792038e-01 -1.02799404e+00 9.68489110e-01
6.98390663e-01 3.94552052e-01 -3.60106617e-01 7.94947445e-01
2.98545718e-01 2.92399347e-01 -3.33559126e-01 1.49281934e-01
7.57899731e-02 -6.20482922e-01 -1.25688091e-01 5.70191801e-01
4.61571008e-01 9.99349579e-02 4.21946682e-02 8.68099570e-01
1.97792277e-01 1.82440639e-01 -2.05481842e-01 -2.81823963e-01
6.48930132e-01 1.04684997e+00 -1.38838363e+00 -3.09913039e-01
-4.54267263e-01 2.98158288e-01 -2.71612052e-02 -4.09302041e-02
-5.71742356e-01 -2.12898955e-01 2.69791782e-01 6.22575462e-01
4.91237044e-01 -3.08769286e-01 -5.62649250e-01 -5.63522875e-01
2.02826057e-02 -1.10858631e+00 3.18135291e-01 -6.27937615e-01
-1.09622431e+00 4.18165833e-01 1.23674273e-01 -1.00730312e+00
-3.85502189e-01 -4.22870219e-01 -4.23566908e-01 4.31250721e-01
-1.07632291e+00 -7.11663008e-01 -3.56973201e-01 4.58318800e-01
2.20658123e-01 1.99306607e-02 5.31785905e-01 6.47888422e-01
-5.12106299e-01 2.55014896e-01 1.09930020e-02 -4.38884079e-01
5.35782099e-01 -1.22099733e+00 -1.88062459e-01 6.59426272e-01
-1.89839974e-01 3.82176965e-01 8.78375828e-01 -9.12457645e-01
-1.54964972e+00 -7.65533566e-01 6.01542354e-01 6.18832335e-02
8.43880653e-01 -2.76739538e-01 -1.04415023e+00 1.96471557e-01
1.53531179e-01 -4.64429349e-01 6.98419988e-01 -2.04474658e-01
-4.66929488e-02 -2.06025213e-01 -1.29047811e+00 5.41214108e-01
6.90401733e-01 -3.67966503e-01 -6.13221228e-01 4.24837321e-02
5.72514348e-02 3.30753848e-02 -1.21395016e+00 4.39619958e-01
5.08153379e-01 -9.65557039e-01 7.13403344e-01 -5.50798327e-02
3.01911384e-01 -7.33341813e-01 -2.39967108e-01 -6.78947866e-01
-2.18215927e-01 -1.85056821e-01 4.42441791e-01 1.09230435e+00
3.28205854e-01 -2.27870136e-01 1.06050789e+00 7.58968294e-01
-3.27820241e-01 -6.12549484e-01 -6.95780814e-01 -5.30109346e-01
-3.31968248e-01 -5.09465814e-01 8.49272251e-01 9.98009920e-01
1.97270408e-01 1.60992280e-01 1.07839480e-01 1.33744404e-01
7.63017237e-01 5.87324679e-01 5.16844928e-01 -1.61882269e+00
-2.53059328e-01 -5.42326510e-01 -4.19324696e-01 -7.77684152e-02
-2.19311237e-01 -8.13188493e-01 5.43234609e-02 -1.51515162e+00
7.60027468e-02 -7.51116991e-01 -1.40780002e-01 3.47135693e-01
1.98229536e-01 1.04651131e-01 -1.93342283e-01 4.94104743e-01
-6.40083492e-01 2.00418934e-01 1.05541718e+00 -1.96854472e-01
-1.64092913e-01 -3.29556048e-01 -6.90295696e-01 5.16839683e-01
6.56052947e-01 -5.66835821e-01 -5.49019933e-01 3.14626545e-01
2.94977188e-01 -3.36932987e-02 -2.17740461e-01 -1.29860628e+00
6.03554606e-01 -3.25969577e-01 1.69939175e-01 -1.00566733e+00
-1.61669970e-01 -1.25077426e+00 8.96043003e-01 4.46362495e-01
2.89652109e-01 8.56877416e-02 -1.62247539e-01 2.22926676e-01
-3.38306844e-01 -2.77788848e-01 5.43997943e-01 -2.25122511e-01
-6.43467009e-01 8.17473419e-03 -6.50091887e-01 -4.74301308e-01
1.45965874e+00 -7.41283119e-01 -6.45687580e-02 1.69970036e-01
-8.94254327e-01 4.36391644e-02 9.19895351e-01 9.66793224e-02
3.30273509e-01 -1.13527536e+00 -1.06633388e-01 3.57223630e-01
2.36187413e-01 3.09006155e-01 2.19656765e-01 8.21129382e-01
-8.78331661e-01 2.62987673e-01 -3.65629196e-01 -8.09123099e-01
-1.36385572e+00 7.41258204e-01 -5.37623838e-02 -2.55387455e-01
-5.33451200e-01 -4.31371741e-02 -3.69740486e-01 -1.82810441e-01
-1.27393112e-01 -6.59230471e-01 -1.22748539e-01 2.38163561e-01
2.34635130e-01 5.09862900e-01 6.22008085e-01 -3.21224779e-01
-3.95440549e-01 6.65581763e-01 -2.06316724e-01 1.50392234e-01
1.56883502e+00 5.03635220e-02 -2.50765413e-01 3.68560642e-01
8.19051206e-01 -1.27047256e-01 -6.42892957e-01 4.13985737e-02
7.30997086e-01 -1.68139249e-01 -9.69070122e-02 -4.63793755e-01
-8.18877220e-01 3.78647804e-01 5.21189690e-01 3.53778213e-01
1.13686144e+00 1.15390845e-01 6.82779849e-01 5.25878668e-01
2.63877809e-01 -1.45667064e+00 -2.42788076e-01 3.21133435e-02
2.06725091e-01 -9.42799628e-01 4.73843575e-01 -5.39144397e-01
-2.38656163e-01 1.16154635e+00 3.38429332e-01 7.76614174e-02
6.46352649e-01 6.82912469e-01 2.97322404e-03 -5.37454188e-01
-5.99677384e-01 -3.55596729e-02 -9.11930948e-03 5.10334790e-01
3.79272580e-01 2.08476186e-02 -3.74896795e-01 5.81386268e-01
-4.31195527e-01 1.83051288e-01 3.99330616e-01 1.49712706e+00
-7.51915276e-01 -1.64421773e+00 -7.27363288e-01 6.91274405e-01
-1.67387679e-01 2.13322565e-01 -4.22727883e-01 9.89934444e-01
1.54445380e-01 8.69738936e-01 2.82381743e-01 -4.54017222e-01
5.02200007e-01 1.46854788e-01 7.19437599e-01 -2.78794587e-01
-1.87157273e-01 1.90758601e-01 -8.55676308e-02 -2.98245609e-01
-8.16834688e-01 -8.18953633e-01 -1.71781540e+00 -4.45464373e-01
-2.72955358e-01 4.78049159e-01 7.80967593e-01 1.08162701e+00
6.03961796e-02 5.11069596e-01 6.05865240e-01 -5.04131317e-01
-6.23538382e-02 -7.23064065e-01 -7.48276293e-01 5.78984559e-01
-1.49919748e-01 -1.01760888e+00 -1.38828009e-01 -5.56572787e-02] | [5.246114730834961, 5.231610298156738] |
4071887b-bea1-446e-8aaa-4a75d0ced8c4 | a-vae-based-bayesian-bidirectional-lstm-for | 2103.12969 | null | https://arxiv.org/abs/2103.12969v2 | https://arxiv.org/pdf/2103.12969v2.pdf | A VAE-Bayesian Deep Learning Scheme for Solar Generation Forecasting based on Dimensionality Reduction | The advancement of distributed generation technologies in modern power systems has led to a widespread integration of renewable power generation at customer side. However, the intermittent nature of renewable energy poses new challenges to the network operational planning with underlying uncertainties. This paper proposes a novel Bayesian probabilistic technique for forecasting renewable solar generation by addressing data and model uncertainties by integrating bidirectional long short-term memory (BiLSTM) neural networks while compressing the weight parameters using variational autoencoder (VAE). Existing Bayesian deep learning methods suffer from high computational complexities as they require to draw a large number of samples from weight parameters expressed in the form of probability distributions. The proposed method can deal with uncertainty present in model and data in a more computationally efficient manner by reducing the dimensionality of model parameters. The proposed method is evaluated using quantile loss, reconstruction error, and deterministic forecasting evaluation metrics such as root-mean square error. It is inferred from the numerical results that VAE-Bayesian BiLSTM outperforms other probabilistic and deterministic deep learning methods for solar power forecasting in terms of accuracy and computational efficiency for different sizes of the dataset. | ['Adnan Anwar', 'Md. Enamul Haque', 'Md. Apel Mahmud', 'Shama Naz Islam', 'Devinder Kaur'] | 2021-03-24 | null | null | null | null | ['probabilistic-deep-learning'] | ['computer-vision'] | [-5.48451066e-01 -1.77330717e-01 -3.44446898e-02 -3.72683823e-01
-8.90327692e-01 -4.80353385e-01 6.17196918e-01 -5.95082939e-02
1.29170671e-01 1.49362683e+00 1.32397652e-01 -3.39633286e-01
-6.17489100e-01 -1.32811797e+00 -8.01751316e-01 -1.09240103e+00
5.04373722e-02 8.07396054e-01 -3.39098603e-01 1.85745984e-01
9.27333310e-02 4.85528439e-01 -1.28098941e+00 -2.93390095e-01
1.05157292e+00 1.45938456e+00 1.91780522e-01 4.27838922e-01
-2.72065811e-02 4.76437062e-01 -8.81550848e-01 -1.10408485e-01
8.38602521e-03 8.15095976e-02 -1.01423986e-01 -6.57944918e-01
-3.66053641e-01 -7.21536219e-01 -1.79276109e-01 8.74070704e-01
1.01256394e+00 3.68153453e-01 1.11632419e+00 -1.28205550e+00
-5.53215086e-01 9.09814954e-01 -2.35058203e-01 3.04553479e-01
-2.54398465e-01 -1.66261449e-01 7.12766051e-01 -8.50480556e-01
-3.37373346e-01 9.53874886e-01 7.84648955e-01 3.38810571e-02
-1.08129930e+00 -6.24737859e-01 -2.71126002e-01 4.49052095e-01
-1.41591692e+00 -7.90499374e-02 8.83147717e-01 -4.83576715e-01
1.26588237e+00 -2.62527049e-01 5.88249326e-01 1.12805367e+00
5.71636975e-01 3.30409378e-01 9.77056921e-01 -2.42479950e-01
7.80981541e-01 -1.31140813e-01 -1.29759863e-01 4.65640500e-02
5.12505233e-01 6.99322820e-01 -2.70779341e-01 -2.06383660e-01
2.99691349e-01 -6.72471300e-02 7.73191676e-02 7.60899559e-02
-6.06612444e-01 1.02473855e+00 3.51266265e-01 3.65206033e-01
-7.65837431e-01 6.61109209e-01 3.44114631e-01 -8.17894116e-02
6.71681762e-01 -4.08838779e-01 -6.57272279e-01 -1.59265637e-01
-1.36097097e+00 -1.91560853e-02 9.90144730e-01 7.05856562e-01
3.77249956e-01 9.72310126e-01 -1.61191285e-01 5.79072118e-01
7.65075982e-01 1.21000886e+00 4.13878679e-01 -7.53679752e-01
1.62220284e-01 -1.65194035e-01 3.68321270e-01 -7.92091131e-01
-3.14366043e-01 -7.01608658e-01 -1.31941092e+00 2.31197685e-01
-1.38540380e-02 -6.73703611e-01 -7.89464355e-01 1.51317167e+00
1.07392840e-01 1.15227632e-01 1.80123523e-01 3.74224693e-01
4.91461605e-01 1.34073603e+00 6.27630278e-02 -2.57192433e-01
6.92128420e-01 -3.47385049e-01 -9.14200604e-01 3.78324270e-01
4.82197478e-02 -4.20343250e-01 8.86567086e-02 4.41445202e-01
-9.04949129e-01 -4.54611748e-01 -1.13387442e+00 9.85902846e-02
-6.75425291e-01 2.53060997e-01 1.62075415e-01 9.18038905e-01
-9.23731804e-01 8.51388097e-01 -9.18925345e-01 1.38873026e-01
4.99314755e-01 1.24057524e-01 4.74888802e-01 2.96650976e-01
-1.29703283e+00 1.14219952e+00 6.40562952e-01 7.40798473e-01
-1.11823618e+00 -9.39252555e-01 -5.95488727e-01 6.76087916e-01
6.87612295e-02 -9.28890944e-01 1.05944920e+00 -3.66020173e-01
-1.89976871e+00 -6.36347413e-01 -2.50416286e-02 -7.70582616e-01
3.70814383e-01 -5.57289049e-02 -4.42707509e-01 -3.73957306e-01
-3.91626984e-01 2.21637458e-01 1.03052044e+00 -1.03162372e+00
-5.58825314e-01 -2.53827810e-01 -4.92049664e-01 -1.41332313e-01
-1.77978575e-01 -7.33259082e-01 8.81516933e-01 -6.81749046e-01
-2.13884071e-01 -4.11681414e-01 8.96184221e-02 -3.61863196e-01
-3.00577193e-01 -7.18027115e-01 7.12857306e-01 -1.22172129e+00
8.53840768e-01 -1.46831369e+00 -2.18023006e-02 4.58548754e-01
-4.31590885e-01 6.24166653e-02 2.60322034e-01 7.13988245e-01
8.33014399e-02 1.80984706e-01 -1.53182909e-01 -2.84942865e-01
6.15064621e-01 6.65296853e-01 -7.79208422e-01 1.64528206e-01
-1.64053038e-01 8.58050764e-01 -7.06828535e-01 -1.40037999e-01
7.06974030e-01 9.53301013e-01 2.03018308e-01 -9.24175754e-02
-6.49998605e-01 -5.10294773e-02 -2.73347259e-01 4.28591877e-01
6.82276070e-01 -4.04897332e-02 2.47431686e-03 -3.62938762e-01
3.96716595e-02 1.02058113e-01 -1.08988571e+00 1.32727218e+00
-7.60662615e-01 6.61430001e-01 -4.04229522e-01 -1.28893399e+00
9.05259609e-01 3.97145212e-01 3.40366215e-01 -3.90061617e-01
1.73594251e-01 2.55530924e-01 -3.59682441e-01 -3.35717164e-02
4.41396028e-01 -3.61809731e-01 2.40891159e-01 4.29559112e-01
2.93866277e-01 -2.50313193e-01 1.91277996e-01 -1.21042654e-01
3.92008066e-01 1.38194829e-01 -5.94003387e-02 -3.09464872e-01
1.40580028e-01 -7.06078708e-01 6.00778699e-01 6.75252259e-01
3.25453579e-01 1.79614853e-02 6.37212619e-02 -2.62669116e-01
-1.06497931e+00 -1.35912132e+00 -3.06621701e-01 4.38452721e-01
-5.38762450e-01 1.55757546e-01 -2.38918856e-01 -2.53127009e-01
3.63423526e-01 1.65444338e+00 -3.09059113e-01 1.54428616e-01
8.12098980e-02 -1.04216981e+00 2.46179566e-01 8.45145583e-01
3.28568906e-01 -6.78272963e-01 -2.93027312e-01 3.88629168e-01
-6.88753128e-02 -6.46928072e-01 1.67825922e-01 1.54296964e-01
-9.32038069e-01 -6.28269434e-01 -9.31587160e-01 -9.20515805e-02
1.96085885e-01 -4.58321929e-01 1.13455975e+00 -5.53804219e-01
-7.26918057e-02 4.91085589e-01 -4.64600101e-02 -7.25112975e-01
-2.54011601e-01 1.01746261e-01 1.31988972e-01 -4.53947276e-01
5.69208302e-02 -9.28036034e-01 -4.57519472e-01 -1.72200322e-01
-6.44022107e-01 -4.20344234e-01 3.70770067e-01 9.60827827e-01
5.60313702e-01 9.23843324e-01 1.22440886e+00 1.77008305e-02
7.44156241e-01 -8.50830019e-01 -1.37279856e+00 4.68751758e-01
-1.08062875e+00 1.70474485e-01 5.03739476e-01 1.19077880e-03
-1.22664177e+00 -6.34775043e-01 4.46125641e-02 -2.82164216e-01
1.12852104e-01 1.00277257e+00 5.36991805e-02 1.94417626e-01
7.23862834e-03 5.00177920e-01 -3.31206650e-01 -5.03910184e-01
3.58574480e-01 5.28136849e-01 1.64549991e-01 -7.13554084e-01
7.76925206e-01 1.22748502e-01 4.53449517e-01 -8.42515349e-01
-7.07725823e-01 3.95223647e-01 -2.21334577e-01 -3.40668559e-01
3.56849551e-01 -9.59520757e-01 -1.01319051e+00 7.54624307e-01
-1.12781227e+00 -1.82388976e-01 -4.73924756e-01 6.24107122e-01
-3.75878751e-01 2.36063480e-01 -3.64629388e-01 -1.31181824e+00
-7.90191650e-01 -6.93672657e-01 6.30688250e-01 3.71839821e-01
4.28195149e-01 -1.44511187e+00 7.17425942e-02 1.99710727e-02
8.92082691e-01 3.22255850e-01 1.20312786e+00 -3.68267506e-01
-5.10848522e-01 -3.46182972e-01 -1.90202802e-01 8.80814791e-01
-5.82843199e-02 3.32225174e-01 -8.18244100e-01 -3.25154454e-01
8.34237412e-02 -1.13489449e-01 5.15263438e-01 1.11161041e+00
1.20904434e+00 -3.36687356e-01 -8.26877058e-02 1.77902147e-01
1.76471961e+00 2.25454673e-01 4.82788265e-01 -2.46668637e-01
2.54568011e-01 2.00930834e-01 -2.06199102e-02 9.00011122e-01
6.48039997e-01 -1.71653360e-01 6.40755653e-01 5.94174862e-01
4.34176475e-01 -3.53002846e-01 3.18907142e-01 1.01270878e+00
1.21585671e-02 -7.23941505e-01 -7.43402719e-01 7.85394788e-01
-1.74550593e+00 -1.23655057e+00 3.68700862e-01 2.21079946e+00
6.21824205e-01 -3.39464843e-02 -2.12054551e-01 2.45640561e-01
6.53077066e-01 1.22887440e-01 -7.81155944e-01 -3.58875751e-01
-1.97351292e-01 1.89521655e-01 7.39634573e-01 5.65832198e-01
-5.08780420e-01 8.79837945e-02 6.00726557e+00 1.19500661e+00
-8.93676400e-01 1.87107116e-01 7.39795446e-01 -2.06020817e-01
-5.93175113e-01 -1.67175397e-01 -8.81649792e-01 1.09206247e+00
1.37448037e+00 -3.80510032e-01 5.26190758e-01 5.61996877e-01
4.62896049e-01 -5.67960024e-01 -9.00431693e-01 8.77753317e-01
-1.84973031e-01 -1.56000018e+00 -4.26730700e-03 2.70358399e-02
1.10477400e+00 5.01436949e-01 1.10822245e-01 4.14014339e-01
5.43698072e-01 -1.22647047e+00 5.74645936e-01 1.47358537e+00
1.95775419e-01 -1.05306411e+00 1.17487299e+00 6.02186859e-01
-8.67968976e-01 -3.27907294e-01 -5.37368715e-01 -7.72329420e-02
5.56663871e-01 1.44051516e+00 -5.23498297e-01 9.04665709e-01
9.59102035e-01 5.20566642e-01 2.41493210e-01 9.23089683e-01
-2.03314558e-01 8.67252707e-01 -1.07960510e+00 -3.59429389e-01
-9.40708816e-02 -4.66936588e-01 4.30042982e-01 4.66598243e-01
1.05687296e+00 -2.55073041e-01 -1.14903770e-01 1.24993813e+00
-1.03925236e-01 -6.63240194e-01 -2.78600186e-01 -3.16690117e-01
1.00841856e+00 9.68560278e-01 -7.32802525e-02 -2.30885133e-01
-5.32236509e-02 1.76549301e-01 -2.90358931e-01 5.85042000e-01
-8.82223368e-01 -1.64609179e-01 2.92855315e-02 -3.21472526e-01
7.38806844e-01 -3.29181224e-01 -4.22118634e-01 -7.63634324e-01
1.06215343e-01 -8.67230527e-04 3.52375746e-01 -1.17941427e+00
-1.86205089e+00 1.82072148e-01 3.02571028e-01 -5.80628753e-01
-7.85546720e-01 -5.55103421e-01 -7.38397300e-01 1.35097408e+00
-1.91842115e+00 -1.14959550e+00 -2.84327623e-02 3.43786567e-01
3.70259792e-01 -4.33491260e-01 8.60482216e-01 9.95376632e-02
-6.26964569e-01 2.26874784e-01 1.10686183e+00 -2.26806968e-01
3.05405669e-02 -1.32678092e+00 -1.58768833e-01 6.16992891e-01
-1.51946947e-01 -3.17677706e-02 7.58321881e-01 -5.80510855e-01
-1.10587668e+00 -8.66137445e-01 6.91465795e-01 4.31513116e-02
6.99313879e-01 7.90955648e-02 -4.97471541e-01 4.93971229e-01
6.94306135e-01 -6.13656566e-02 5.92942476e-01 -2.10455954e-02
7.74108097e-02 -4.72716630e-01 -1.51712167e+00 -8.31311122e-02
-1.12235270e-01 -4.90881771e-01 -4.70835686e-01 3.53705496e-01
2.22649679e-01 5.87971285e-02 -1.56098449e+00 5.57474971e-01
5.21758914e-01 -6.35721505e-01 7.70358324e-01 -1.94919199e-01
1.63004398e-01 -2.81486392e-01 -3.91910970e-01 -1.73256791e+00
3.00020464e-02 -5.88980556e-01 -9.24937785e-01 1.48694468e+00
2.95864612e-01 -9.48481798e-01 5.09636879e-01 6.83240354e-01
9.75309610e-02 -6.78676069e-01 -1.58198297e+00 -8.26092422e-01
6.00188196e-01 -4.52472329e-01 8.08998048e-01 4.54087019e-01
-5.28130949e-01 -2.57782675e-02 -2.47613341e-01 5.06141305e-01
1.13244510e+00 1.00170143e-01 8.65347311e-02 -1.37145090e+00
4.26416397e-02 -2.73517907e-01 2.93646362e-02 -6.12186134e-01
2.85758942e-01 -4.91304964e-01 2.25897655e-02 -1.93549132e+00
-1.77998170e-01 -3.66544306e-01 -5.96305728e-01 3.02335948e-01
4.57317531e-01 -3.31530273e-01 -1.12207651e-01 -1.95572525e-01
7.83458799e-02 1.49734151e+00 3.29416931e-01 -3.45304579e-01
3.99294198e-01 3.68770301e-01 6.29042983e-02 6.13406837e-01
1.14088798e+00 -4.12189037e-01 -5.86009204e-01 -6.33926272e-01
7.34128773e-01 3.99932712e-01 5.58177412e-01 -1.12300539e+00
4.24653679e-01 5.70295304e-02 8.28390241e-01 -1.50403035e+00
3.34420204e-01 -9.82960641e-01 4.24633831e-01 3.17675531e-01
1.91669688e-01 2.24957094e-02 3.16442251e-01 9.23859656e-01
3.77188139e-02 -4.59409505e-01 4.56379771e-01 2.89034486e-01
-3.05551559e-01 2.04215020e-01 -7.04181850e-01 -3.86973798e-01
7.04252541e-01 7.71586001e-02 -4.35539395e-01 -5.50244212e-01
-8.68756950e-01 5.15701890e-01 -3.22422981e-01 1.34515017e-01
4.78216916e-01 -1.30967355e+00 -7.25839198e-01 -1.34281844e-01
-7.30576038e-01 2.10936237e-02 5.39130449e-01 4.33664322e-01
-2.45030269e-01 6.03587091e-01 1.28373146e-01 -5.51726222e-01
-2.70920366e-01 1.09519571e-01 7.22618461e-01 -2.05409423e-01
-1.54690087e-01 5.92871904e-01 -8.30256522e-01 -3.84683311e-01
3.59118134e-01 -5.39446414e-01 2.46429560e-03 4.52303976e-01
6.70234412e-02 1.15795386e+00 8.52715373e-02 -1.80927083e-01
3.91280986e-02 2.53205985e-01 4.61123139e-01 -1.57546058e-01
1.59437633e+00 -4.02450502e-01 -2.83315033e-01 8.16927016e-01
6.84059978e-01 -5.35699129e-01 -1.29988313e+00 -2.20642865e-01
-1.37659046e-03 5.77522255e-02 8.56162965e-01 -1.45389199e+00
-1.24342287e+00 1.11888683e+00 1.06993747e+00 3.55802059e-01
8.26731861e-01 -5.33305228e-01 8.57196450e-01 6.00580096e-01
4.47789133e-01 -1.48833084e+00 -6.12547576e-01 5.07470548e-01
9.58462715e-01 -1.22763550e+00 1.06809184e-01 6.29030406e-01
1.06903501e-01 1.20286238e+00 -1.96461640e-02 -1.69347525e-01
1.38803041e+00 3.65926653e-01 -4.01401490e-01 3.70481849e-01
-7.31959939e-01 2.84535676e-01 1.90929040e-01 6.43539369e-01
-1.01459689e-01 3.20464492e-01 -1.67244181e-01 9.21494067e-01
-5.13308495e-02 2.90301919e-01 4.16547954e-01 5.52438915e-01
-3.19016069e-01 -7.71403551e-01 -2.95123398e-01 7.18121767e-01
-2.73868829e-01 -2.75855005e-01 6.55568480e-01 2.76346654e-01
-1.06307983e-01 1.10027182e+00 1.38723195e-01 2.33288467e-01
-1.75779611e-01 2.12573022e-01 4.19077724e-01 2.02560844e-03
-6.54951204e-03 4.88610789e-02 -1.78069204e-01 -2.74849795e-02
-3.26414794e-01 -5.76238811e-01 -9.12792027e-01 -5.66564023e-01
-6.22248054e-01 2.84858048e-01 1.24094546e+00 1.21749711e+00
3.98943603e-01 7.98284709e-01 6.75652027e-01 -1.24538004e+00
-1.33834934e+00 -1.38944387e+00 -8.12146783e-01 -5.82856119e-01
1.73677266e-01 -7.63507187e-01 -5.54908216e-01 -3.75792474e-01] | [6.138530254364014, 2.864084243774414] |
62e74e1f-83dd-42e1-9016-965ecd5fa7b4 | proceedings-of-the-1st-international-workshop-6 | 2301.10062 | null | https://arxiv.org/abs/2301.10062v1 | https://arxiv.org/pdf/2301.10062v1.pdf | Proceedings of the 1st International Workshop on Reading Music Systems | The International Workshop on Reading Music Systems (WoRMS) is a workshop that tries to connect researchers who develop systems for reading music, such as in the field of Optical Music Recognition, with other researchers and practitioners that could benefit from such systems, like librarians or musicologists. The relevant topics of interest for the workshop include, but are not limited to: Music reading systems; Optical music recognition; Datasets and performance evaluation; Image processing on music scores; Writer identification; Authoring, editing, storing and presentation systems for music scores; Multi-modal systems; Novel input-methods for music to produce written music; Web-based Music Information Retrieval services; Applications and projects; Use-cases related to written music. These are the proceedings of the 1st International Workshop on Reading Music Systems, held in Paris on the 20th of September 2018. | ['Alexander Pacha', 'Jan Hajič jr.', 'Jorge Calvo-Zaragoza'] | 2022-12-01 | null | null | null | null | ['music-information-retrieval'] | ['music'] | [ 4.46038187e-01 -3.97618592e-01 -1.21915758e-01 3.33743125e-01
-8.45007360e-01 -1.07128263e+00 2.12449685e-01 1.53348586e-02
-8.22027549e-02 -2.90078163e-01 2.84925699e-01 8.56997594e-02
-7.91854978e-01 -3.56102675e-01 -9.86525938e-02 -2.50565201e-01
2.58106828e-01 5.12305439e-01 6.78936318e-02 -7.15578422e-02
1.04726481e+00 6.28624797e-01 -2.09611082e+00 4.67480093e-01
3.03411394e-01 7.52818108e-01 1.79544657e-01 1.18366921e+00
-3.65177959e-01 4.80659127e-01 -1.01312149e+00 -3.47882837e-01
5.06836951e-01 -7.11495578e-01 -7.27051139e-01 -2.56600440e-01
7.73150027e-01 -4.87182029e-02 2.75116134e-02 1.30049181e+00
8.84418726e-01 1.28395155e-01 5.16647041e-01 -9.07973766e-01
-7.70529568e-01 1.31088984e+00 -6.73296392e-01 -4.94024158e-03
8.97866249e-01 -1.80358768e-01 1.08516169e+00 -8.11588526e-01
5.67784488e-01 1.08313823e+00 7.14743435e-01 3.66786152e-01
-9.96981025e-01 -8.49536955e-01 -6.37691855e-01 4.00188744e-01
-1.43194115e+00 -7.53771424e-01 6.75038993e-01 -4.84079450e-01
6.63817167e-01 7.74138868e-01 9.37005043e-01 7.11289823e-01
-5.52122951e-01 1.03968763e+00 7.81927049e-01 -9.82185304e-01
-7.13456273e-02 1.01933680e-01 6.18972890e-02 3.10485780e-01
-1.34461075e-01 -1.11352652e-01 -1.30404413e+00 -2.34405901e-02
1.23756802e+00 -6.02731466e-01 -1.23207480e-01 4.42967527e-02
-1.79279101e+00 2.40954250e-01 -1.93687826e-01 5.75870454e-01
-2.55750477e-01 -4.81960326e-02 5.03971279e-01 7.21336365e-01
-1.68612167e-01 1.00949800e+00 -1.43447980e-01 -8.19332659e-01
-1.38190639e+00 4.54882681e-01 9.22800124e-01 1.01407921e+00
-5.79396412e-02 2.11326755e-03 -2.59644151e-01 1.44923019e+00
3.30430031e-01 7.49872088e-01 7.11997151e-01 -1.20759082e+00
1.99691907e-01 4.95939255e-01 2.31772009e-02 -5.51597774e-01
-3.28011274e-01 -1.16610892e-01 -3.95691633e-01 3.16958189e-01
4.84302223e-01 3.08944196e-01 -4.66622978e-01 9.57491159e-01
-1.46584123e-01 2.45026767e-01 -1.84602663e-01 1.15355003e+00
1.38689661e+00 5.98976016e-01 -5.71525931e-01 -1.04529366e-01
1.50201380e+00 -8.74493122e-01 -6.28844738e-01 6.37736976e-01
-8.31316924e-04 -1.88998163e+00 1.08448982e+00 1.18071449e+00
-1.71100831e+00 -7.59093583e-01 -9.32947218e-01 -2.08660051e-01
-1.60855338e-01 5.43969095e-01 4.24096972e-01 7.38736868e-01
-1.04307795e+00 8.58848274e-01 -2.40342945e-01 -6.96983635e-01
-2.41835620e-02 4.49433923e-01 2.21442744e-01 5.60067773e-01
-5.48299015e-01 5.65947771e-01 2.88421195e-02 -4.19465095e-01
-2.36968115e-01 -8.64376187e-01 1.60312355e-01 -7.39141479e-02
2.61176527e-02 -5.68993270e-01 1.54652429e+00 -1.12052131e+00
-2.06357932e+00 1.41370356e+00 1.06931649e-01 6.16931468e-02
2.25025341e-01 -4.68981355e-01 -7.90148139e-01 -3.37144881e-02
6.11571185e-02 4.45216328e-01 7.69117296e-01 -6.93433821e-01
-5.89677930e-01 -4.20685083e-01 -3.05065960e-01 3.12716722e-01
-2.58171529e-01 6.30950093e-01 -9.36889529e-01 -9.61065710e-01
2.17920795e-01 -1.01184928e+00 6.57590628e-01 -2.10545868e-01
-7.72475719e-01 -2.06461981e-01 5.86936533e-01 -5.69621921e-01
1.29639506e+00 -2.45310831e+00 3.18978518e-01 6.53061807e-01
-4.58351403e-01 5.39905369e-01 -3.52246314e-01 5.95978141e-01
1.86389431e-01 -2.60408223e-01 5.28123140e-01 1.04674153e-01
3.71014565e-01 -5.70223629e-01 -4.75057155e-01 -5.26445098e-02
-8.02726448e-01 5.72031379e-01 -6.99719429e-01 -3.63873005e-01
3.88330370e-01 1.55873552e-01 -4.08245251e-02 -1.05067819e-01
7.08963946e-02 2.15098932e-01 -2.28616729e-01 8.28536153e-01
3.78835469e-01 5.88211082e-02 -1.22749463e-01 -1.31591439e-01
-5.10635197e-01 3.66066337e-01 -1.98723245e+00 2.21199226e+00
-8.72877911e-02 9.53603506e-01 4.32339609e-02 -1.52362198e-01
1.21233582e+00 5.57205081e-01 6.84763968e-01 -5.16705990e-01
-1.03090681e-01 6.43427551e-01 -1.57443747e-01 -4.82313633e-01
1.04598749e+00 2.22988605e-01 3.50157738e-01 1.00429142e+00
-4.02345955e-02 -3.29084098e-01 5.84025800e-01 2.94868797e-02
8.29154849e-01 3.27316940e-01 -1.64868645e-02 -9.59004536e-02
5.82980454e-01 -6.86722770e-02 -7.04183728e-02 8.52574825e-01
2.76876986e-02 7.30460584e-01 -1.29847571e-01 -5.57511896e-02
-1.03192079e+00 -1.29532409e+00 -3.24570894e-01 1.57923734e+00
1.34905860e-01 -8.65232110e-01 -8.03590775e-01 2.19079062e-01
-8.11002031e-02 -2.15056110e-02 1.52172834e-01 2.64088362e-01
-2.53930122e-01 4.08442393e-02 1.19052470e+00 4.04281527e-01
2.39603594e-01 -1.72125971e+00 -3.32586080e-01 1.47050485e-01
1.07435822e-01 -4.34649527e-01 -7.32474089e-01 5.28945215e-03
-8.13533425e-01 -1.20085037e+00 -1.11968672e+00 -1.11485994e+00
-1.00992844e-01 2.05095828e-01 1.14624763e+00 1.50431663e-01
-5.42996585e-01 7.32215345e-01 -4.02770370e-01 -1.10721660e+00
-3.27184558e-01 3.90461385e-01 1.19342051e-01 -2.45146573e-01
8.44520152e-01 -6.58856392e-01 -5.94983816e-01 3.41376454e-01
-6.71139717e-01 1.69180915e-01 5.40989935e-01 1.96933627e-01
6.29587293e-01 -5.51963270e-01 6.32747293e-01 -6.19344354e-01
8.86272252e-01 3.02986860e-01 -6.22484803e-01 4.15066779e-01
-5.45388222e-01 -3.82856727e-01 1.68502152e-01 -5.80015838e-01
-6.18865132e-01 4.97312732e-02 7.83913434e-02 -1.11332864e-01
-2.41154149e-01 1.76758561e-02 7.09972382e-02 -1.23304278e-01
9.72413540e-01 3.02631497e-01 -2.46238530e-01 -9.98991311e-01
3.55397880e-01 1.32536781e+00 1.11587656e+00 -5.51641583e-01
7.94857562e-01 2.47893885e-01 9.79007855e-02 -1.04142022e+00
-5.32294929e-01 -1.05461788e+00 -7.43143678e-01 -5.55080712e-01
5.13567746e-01 -6.32525384e-01 -9.18384790e-01 6.56860471e-01
-7.46702492e-01 1.98656437e-03 -9.14002478e-01 6.39904380e-01
-7.55745888e-01 6.55244738e-02 -8.02788377e-01 -7.84173906e-01
-6.90841794e-01 -8.48194718e-01 1.37341905e+00 5.55012047e-01
-9.26310778e-01 -5.80349445e-01 5.57397127e-01 8.21417451e-01
4.25456494e-01 -4.63485569e-01 6.15646482e-01 -2.80045360e-01
-5.47411382e-01 -1.25464350e-01 -1.70693244e-03 1.47811353e-01
-1.88277975e-01 4.00944173e-01 -1.26113772e+00 1.23136111e-01
-9.03312743e-01 -3.02085966e-01 6.75407410e-01 5.28248966e-01
1.03974020e+00 3.92991275e-01 7.09685907e-02 5.73640108e-01
8.86527896e-01 3.29594165e-01 8.58272374e-01 6.34476423e-01
6.13622546e-01 2.29479209e-01 3.90769184e-01 7.08327055e-01
-2.66022414e-01 1.23090506e+00 -3.22881602e-02 1.35717273e-01
-4.83735472e-01 -2.45251313e-01 6.46401584e-01 1.27345145e+00
-8.14866722e-01 2.57462054e-01 -4.19089854e-01 2.95084924e-01
-1.70917809e+00 -1.24288285e+00 -7.35189319e-01 2.52622104e+00
6.52724385e-01 -5.44873238e-01 7.05150962e-01 8.81236076e-01
6.79193556e-01 -1.37723953e-01 -2.50789523e-01 -6.88239098e-01
-1.68131024e-01 8.28312278e-01 9.55930501e-02 -1.29820213e-01
-1.00000012e+00 8.89747322e-01 6.59972620e+00 9.46683288e-01
-1.15570700e+00 -1.57500982e-01 -5.27213573e-01 -6.08199596e-01
-8.98467451e-02 -2.19770923e-01 -6.14428163e-01 2.45087072e-02
8.53438854e-01 -2.19203562e-01 1.13936627e+00 4.66137081e-01
3.22005957e-01 7.20357746e-02 -9.52234745e-01 1.47880745e+00
2.72099316e-01 -1.53055894e+00 1.45588920e-01 2.74533238e-02
1.13766682e+00 1.22315019e-01 5.88162057e-02 -1.43201113e-01
-1.00461461e-01 -9.66198981e-01 1.15938127e+00 1.02781761e+00
8.65797341e-01 -7.69567728e-01 3.02298695e-01 -8.12957138e-02
-1.27977586e+00 9.73695591e-02 -2.64552653e-01 5.82586154e-02
-1.08311689e-02 3.31557602e-01 -2.14345559e-01 2.77216583e-01
8.13145399e-01 9.45238709e-01 -7.85018086e-01 1.86082101e+00
-7.05777109e-02 5.47668159e-01 -8.09237733e-02 -1.65753305e-01
-5.27636945e-01 -2.53507048e-01 9.20991063e-01 1.26120567e+00
8.91476810e-01 -2.55807132e-01 -1.85985103e-01 6.77828908e-01
-9.90371332e-02 7.55404890e-01 -2.44239137e-01 -3.62202436e-01
4.66115206e-01 1.31334424e+00 -7.29907095e-01 -2.06568956e-01
3.95864155e-03 1.00643909e+00 -6.30243421e-01 9.36213732e-02
-2.31731668e-01 -6.89931154e-01 3.31270218e-01 3.13462168e-01
-1.09821491e-01 4.16094214e-02 -7.28642642e-01 -8.33253145e-01
-1.05758928e-01 -1.27903414e+00 2.92796820e-01 -1.21198285e+00
-1.20315766e+00 -5.17791836e-03 -4.64732468e-01 -1.60122299e+00
-9.74662453e-02 -7.18683302e-01 -6.24609888e-01 7.79969633e-01
-4.81675327e-01 -1.23061240e+00 -1.98125303e-01 7.70479023e-01
3.54331046e-01 -9.85285759e-01 1.42223096e+00 6.49305105e-01
3.44294198e-02 7.61842608e-01 5.22705913e-01 6.79727793e-02
1.19877195e+00 -1.27752149e+00 -5.14403507e-02 1.03609264e-01
1.37377870e+00 5.43739736e-01 1.39375493e-01 -1.28005207e-01
-1.73594022e+00 -4.31694567e-01 1.09724081e+00 -4.43041116e-01
4.29397672e-01 2.82134637e-02 -3.40553880e-01 4.29335475e-01
2.29089096e-01 -1.05252075e+00 1.07442236e+00 6.32851124e-01
-3.75502050e-01 -2.95850068e-01 -6.71938479e-01 3.55938166e-01
1.05318534e+00 -8.44006598e-01 -3.72827262e-01 4.08088923e-01
-2.97793746e-01 -5.70857525e-01 -8.70200276e-01 -3.31442803e-01
1.23032486e+00 -7.76634455e-01 9.79146361e-01 -6.66478500e-02
3.35749418e-01 -7.51154780e-01 9.42149982e-02 -8.91459286e-01
-4.13259745e-01 -1.08814800e+00 3.01935583e-01 1.46468091e+00
4.36955184e-01 2.46267334e-01 5.18457890e-01 -2.73016721e-01
-2.75156647e-01 2.14219496e-01 -6.12485647e-01 -6.70875251e-01
-4.08585578e-01 -9.04918134e-01 5.72015584e-01 6.98342204e-01
5.08371413e-01 4.19262618e-01 -4.49220657e-01 -2.16178313e-01
5.11505604e-01 3.50814819e-01 1.31602955e+00 -1.62989485e+00
-8.36372852e-01 -1.00952899e+00 -7.73190677e-01 -8.12911153e-01
-4.51278061e-01 -1.05376279e+00 -2.47418851e-01 -1.53482580e+00
4.09865886e-01 1.28000662e-01 -4.87185776e-01 3.37004423e-01
3.79812837e-01 9.16980028e-01 6.59588993e-01 7.47856736e-01
-9.08352077e-01 -5.37228584e-01 1.48460114e+00 -1.26907274e-01
-6.17366254e-01 3.40328515e-01 -7.44538903e-01 6.03663862e-01
4.35452610e-01 -1.72575280e-01 -2.51879901e-01 -2.47985795e-01
8.30973506e-01 -5.21251075e-02 3.23315918e-01 -1.38530028e+00
4.49639142e-01 4.19140980e-03 3.03713858e-01 -8.31612825e-01
2.47775465e-01 -3.85524809e-01 6.11062706e-01 -7.21018240e-02
-6.13796532e-01 -6.77444739e-03 9.69053730e-02 -1.75698921e-01
-4.20577377e-01 -6.75828755e-01 5.49597979e-01 -9.46600810e-02
-5.93828022e-01 -3.55734318e-01 -5.11127532e-01 -3.00067842e-01
6.43312156e-01 -4.01621431e-01 -2.78303713e-01 -5.61610043e-01
-1.06324601e+00 -3.53804648e-01 4.33996439e-01 6.77669704e-01
3.55567336e-01 -1.40467203e+00 -7.67027736e-01 2.78125525e-01
3.35883677e-01 -7.79847562e-01 1.04462370e-01 7.52424419e-01
-7.76384711e-01 2.47037441e-01 -3.34809244e-01 -4.16679978e-01
-1.88642585e+00 3.25365849e-02 -2.07304806e-02 -1.03050135e-02
-6.67671204e-01 7.71139443e-01 -5.59164941e-01 -3.24780077e-01
6.16711318e-01 2.50592232e-02 -4.22503501e-01 1.40832126e-01
1.10048640e+00 8.60298991e-01 1.15176365e-01 -4.70669538e-01
-1.08953903e-03 1.04105699e+00 3.22912067e-01 -6.60415888e-01
1.24388397e+00 5.54630347e-02 -1.41252548e-01 1.05300403e+00
3.38879198e-01 3.05147886e-01 -2.49040812e-01 -1.10619381e-01
1.95119411e-01 -5.56467831e-01 4.01167944e-02 -1.29808104e+00
-9.12840724e-01 7.62849689e-01 1.04897499e+00 3.54701787e-01
1.19503474e+00 -3.08798160e-02 7.29626894e-01 4.46586549e-01
3.10277134e-01 -1.71174788e+00 1.75351072e-02 6.65050268e-01
9.75926042e-01 -5.62564313e-01 -7.28155747e-02 5.75634986e-02
-2.41228446e-01 1.75874460e+00 -4.46173213e-02 4.10040803e-02
4.90843922e-01 2.70008236e-01 6.96758687e-01 -3.63030851e-01
-2.19067901e-01 -2.15871096e-01 9.18395102e-01 8.05767179e-01
1.13170397e+00 3.67189825e-01 -3.11186314e-01 6.75307035e-01
-1.00012672e+00 4.16858643e-01 4.22215790e-01 4.89426821e-01
-3.06553215e-01 -1.44681919e+00 -9.65085924e-01 4.71109301e-01
-5.57313621e-01 -1.48986101e-01 -1.17009366e+00 1.32688671e-01
3.35106045e-01 9.03687418e-01 1.49021745e-02 -4.50435936e-01
6.86372459e-01 2.49214485e-01 8.84483933e-01 -6.62450135e-01
-1.32113588e+00 5.80755651e-01 7.63579085e-02 -3.61062288e-01
-6.85401022e-01 -8.60777020e-01 -1.23737633e+00 -5.82937956e-01
-3.21098089e-01 1.28627703e-01 1.19760764e+00 5.16445756e-01
2.59942949e-01 4.78926271e-01 1.47800371e-01 -1.07006669e+00
-1.85430497e-01 -1.29357779e+00 -1.46056604e+00 5.02368093e-01
-2.65031725e-01 8.33512172e-02 -6.77113757e-02 4.54406381e-01] | [16.00710105895996, 5.238487720489502] |
dc8ea22e-1dc9-4965-b0f9-6dfc7add1e00 | a-simple-and-efficient-deep-scanpath | 2112.04610 | null | https://arxiv.org/abs/2112.04610v1 | https://arxiv.org/pdf/2112.04610v1.pdf | A Simple and efficient deep Scanpath Prediction | Visual scanpath is the sequence of fixation points that the human gaze travels while observing an image, and its prediction helps in modeling the visual attention of an image. To this end several models were proposed in the literature using complex deep learning architectures and frameworks. Here, we explore the efficiency of using common deep learning architectures, in a simple fully convolutional regressive manner. We experiment how well these models can predict the scanpaths on 2 datasets. We compare with other models using different metrics and show competitive results that sometimes surpass previous complex architectures. We also compare the different leveraged backbone architectures based on their performances on the experiment to deduce which ones are the most suitable for the task. | ['Aladine Chetouani', 'Mohamed Amine Kerkouri'] | 2021-12-08 | null | null | null | null | ['scanpath-prediction'] | ['computer-vision'] | [-2.84984298e-02 -5.59697896e-02 -3.39404494e-01 -4.41995680e-01
2.44743258e-01 -3.98615330e-01 5.89452505e-01 -2.53694475e-01
-5.68556964e-01 4.41794902e-01 6.17877766e-02 -5.14868736e-01
-2.54399538e-01 -3.41474205e-01 -9.17467296e-01 -5.96919835e-01
-1.14336796e-01 2.97555774e-01 4.24796551e-01 -2.58982837e-01
7.74040818e-01 5.34332216e-01 -1.65661931e+00 3.09326619e-01
2.67516166e-01 8.64220619e-01 1.59657076e-01 7.46415317e-01
1.48143142e-01 1.31195474e+00 -3.30411136e-01 -3.94040972e-01
3.35044354e-01 -2.65943497e-01 -9.36984301e-01 -2.52090782e-01
7.91803539e-01 -2.29117170e-01 -3.60830545e-01 7.82743156e-01
1.19140409e-01 -3.46642509e-02 6.13503575e-01 -1.32530475e+00
-7.24543869e-01 4.23581630e-01 -8.18128347e-01 9.43632483e-01
2.36261219e-01 3.36276203e-01 1.11757672e+00 -8.96836042e-01
6.23796642e-01 1.15892124e+00 5.41028082e-01 5.01757026e-01
-1.08255923e+00 -6.52379632e-01 2.38673091e-01 9.01553452e-01
-1.10955131e+00 -3.73822123e-01 5.84038258e-01 -5.21263421e-01
9.38570321e-01 3.72640528e-02 5.73317170e-01 1.31898046e+00
5.03988206e-01 7.88434684e-01 1.22573781e+00 -4.85851377e-01
-1.75000533e-01 -1.17074281e-01 3.46726537e-01 8.45940232e-01
1.62424609e-01 1.92541882e-01 -6.45563424e-01 4.55894649e-01
6.44517899e-01 1.15344010e-01 -3.93154323e-01 -5.37359118e-01
-1.07629108e+00 8.31438005e-01 9.22068417e-01 2.75490433e-01
-4.16149348e-01 2.51689017e-01 6.85824007e-02 3.47500890e-01
2.81200439e-01 4.17985708e-01 -4.98098582e-01 1.02416359e-01
-1.11456537e+00 2.11698219e-01 6.16525531e-01 8.64642024e-01
8.37449849e-01 -1.60563424e-01 -2.93569505e-01 3.13580632e-01
4.32498753e-01 1.52282283e-01 3.59921217e-01 -8.52861404e-01
1.70073450e-01 6.65873230e-01 1.20555377e-02 -8.30792844e-01
-7.06677020e-01 -6.36458397e-01 -4.84572262e-01 5.92088878e-01
6.11733615e-01 -7.82052428e-02 -9.76975679e-01 1.56749570e+00
-1.64564803e-01 1.37420267e-01 -4.47614729e-01 9.86642063e-01
5.80256462e-01 3.85502160e-01 1.22609340e-01 5.16741350e-02
1.25062239e+00 -1.51804435e+00 -5.72096765e-01 -3.57503772e-01
3.04444075e-01 -7.59419262e-01 1.29887617e+00 5.83119810e-01
-1.26979053e+00 -7.54039943e-01 -1.10399902e+00 -2.60970682e-01
-4.27500993e-01 1.39098853e-01 4.76045132e-01 4.86164600e-01
-1.61921108e+00 5.67123830e-01 -7.17251241e-01 -4.83359903e-01
7.13204443e-01 3.69496077e-01 -7.47223198e-02 4.82141748e-02
-8.46674860e-01 1.15016878e+00 2.21691728e-01 -6.97980775e-03
-1.17888439e+00 -5.77150643e-01 -4.22132850e-01 4.45545733e-01
3.17520738e-01 -8.13140512e-01 1.40391743e+00 -1.31746089e+00
-1.19955468e+00 1.07656729e+00 -2.98221558e-01 -8.27654600e-01
4.41312224e-01 -4.04753774e-01 -8.53666365e-02 6.17937446e-02
-3.16221982e-01 1.02569342e+00 1.14021361e+00 -9.81493711e-01
-7.71972418e-01 -2.86303014e-01 6.52309418e-01 3.66759524e-02
-1.29439682e-01 9.56864804e-02 -4.29687113e-01 -2.04087079e-01
-4.61507112e-01 -1.07694805e+00 -1.79220870e-01 6.25393540e-02
-4.28538740e-01 -3.96160096e-01 6.36458695e-01 -2.58929789e-01
1.27531374e+00 -1.84299517e+00 2.68243253e-01 -1.66275769e-01
6.11574769e-01 3.27254385e-01 -1.62730321e-01 3.24458539e-01
-1.98852077e-01 1.30825862e-01 2.22559214e-01 -3.70697379e-01
-1.74084917e-01 5.85021712e-02 -1.73185393e-01 4.93001103e-01
2.13227749e-01 9.44632530e-01 -4.91408855e-01 -3.95431042e-01
8.37572515e-02 4.29641604e-01 -6.30725443e-01 3.32268268e-01
-2.75044233e-01 6.04253948e-01 -6.15476146e-02 3.87899250e-01
5.87806106e-01 -6.97726905e-01 -1.04692936e-01 -1.08222134e-01
-3.09838653e-01 1.60547510e-01 -3.52279693e-01 1.57286823e+00
-2.65155703e-01 1.27676082e+00 -4.09476787e-01 -8.34004164e-01
7.22601354e-01 -2.65355140e-01 8.57750028e-02 -1.01517534e+00
2.82097369e-01 -1.35145217e-01 5.28801024e-01 -6.67742074e-01
4.94110435e-01 2.07111269e-01 5.01849413e-01 5.38318634e-01
2.32800707e-01 7.83711612e-01 2.44762331e-01 5.56167737e-02
9.67088521e-01 1.84261069e-01 4.03605849e-01 -4.54679966e-01
6.65395379e-01 -6.63467720e-02 -1.09345414e-01 7.99071908e-01
-3.54041785e-01 5.85270345e-01 9.28979099e-01 -9.21033680e-01
-1.02906299e+00 -6.98545218e-01 1.86077893e-01 1.61652839e+00
2.44391352e-01 -1.47135288e-01 -9.20621037e-01 -6.98543370e-01
-3.84336799e-01 5.76455534e-01 -1.24085474e+00 6.87479973e-02
-6.35030091e-01 -3.63169432e-01 4.90456551e-01 6.49551034e-01
1.99526086e-01 -1.39073801e+00 -1.24972403e+00 -4.01308000e-01
1.31319063e-02 -1.02034903e+00 -1.65340602e-01 2.86205173e-01
-6.99142873e-01 -1.46573055e+00 -7.82092094e-01 -6.29894853e-01
3.23875129e-01 5.39529920e-01 1.55869019e+00 4.39162821e-01
-4.89327051e-02 1.83710605e-01 -3.00509930e-01 -7.15591192e-01
7.80997649e-02 6.36870265e-01 -2.44647264e-01 4.17116210e-02
9.21378076e-01 -3.09767753e-01 -1.03562152e+00 1.81279853e-01
-7.10134566e-01 1.24092899e-01 8.02963018e-01 4.76793140e-01
2.95510441e-01 -5.74956119e-01 4.35111485e-03 -1.06379426e+00
4.62507188e-01 -7.67941236e-01 -6.94590926e-01 1.50149375e-01
-8.14805627e-01 2.50595093e-01 3.50702018e-01 -1.23203047e-01
-7.15327501e-01 -1.48715287e-01 -5.95493466e-02 -7.95108140e-01
-3.36117834e-01 4.48071599e-01 5.46988726e-01 -3.95201027e-01
7.17225134e-01 1.25803247e-01 -2.95575503e-02 -3.99275035e-01
2.03373656e-01 1.56764686e-01 1.94124982e-01 9.19950157e-02
6.50143862e-01 6.18464708e-01 2.33053356e-01 -5.60946047e-01
-1.12259436e+00 -3.56673121e-01 -1.01252031e+00 -3.20345551e-01
9.71970260e-01 -7.12092400e-01 -1.12699354e+00 3.06794584e-01
-1.30087674e+00 -4.30212885e-01 1.55886650e-01 3.22183698e-01
-5.63657105e-01 3.43475416e-02 -2.10784271e-01 -5.43424785e-01
-1.28260046e-01 -1.17271149e+00 7.79097199e-01 5.06510615e-01
-1.36437044e-01 -1.09865701e+00 3.13465297e-01 1.43443704e-01
7.38068223e-01 2.37326533e-01 8.89962077e-01 -6.87530935e-01
-8.91376317e-01 3.24349731e-01 -5.59076130e-01 1.34010091e-02
-4.35641736e-01 1.34310797e-01 -1.14101696e+00 -3.41433853e-01
-1.38500407e-02 -2.46166587e-01 1.14169192e+00 7.25039542e-01
1.50959039e+00 -3.78483422e-02 -4.30762470e-01 8.76542449e-01
1.68441188e+00 3.30637321e-02 7.84324169e-01 7.14662850e-01
6.91328883e-01 6.99686766e-01 4.28328961e-01 -1.20604120e-03
4.10527170e-01 6.81675434e-01 1.15306616e+00 -1.27179191e-01
-1.42470390e-01 4.30518296e-03 3.44403923e-01 3.76340896e-01
-3.95372897e-01 -4.83671546e-01 -1.10279441e+00 5.83304048e-01
-1.83903575e+00 -9.00165021e-01 -4.22406584e-01 1.77887428e+00
1.38543040e-01 3.13370615e-01 4.69719768e-01 -1.35133952e-01
5.13023913e-01 4.98693615e-01 -6.86003029e-01 -6.71487749e-01
7.23512620e-02 4.33825105e-01 6.24924064e-01 2.87306726e-01
-9.66104507e-01 8.93972218e-01 7.38805866e+00 1.47394896e-01
-1.32851624e+00 2.69003898e-01 5.68697155e-01 -4.09564823e-01
1.55508846e-01 -5.78160472e-02 -8.16987634e-01 3.60688239e-01
1.25464427e+00 1.14768572e-01 4.01159257e-01 7.20810533e-01
1.56454533e-01 -4.33089510e-02 -1.44380260e+00 9.76479948e-01
3.14715892e-01 -1.35598576e+00 4.96599376e-02 1.52684599e-01
5.43841541e-01 5.84950268e-01 4.87016082e-01 2.85146207e-01
1.74201906e-01 -1.50781989e+00 7.49493718e-01 8.16016316e-01
5.82855403e-01 -4.70104188e-01 5.60128987e-01 3.15918028e-01
-7.09664643e-01 -4.86191571e-01 -4.04679775e-01 -4.28977370e-01
-5.73903508e-02 -9.52894613e-02 -7.38232374e-01 1.60800710e-01
9.60715890e-01 8.10584545e-01 -1.24233222e+00 1.31929910e+00
-3.58911872e-01 5.99117458e-01 2.28504613e-01 -1.62921295e-01
7.49988258e-01 1.37314886e-01 2.68417627e-01 1.15821803e+00
2.67111689e-01 -3.60309690e-01 -3.09818089e-01 1.07352197e+00
-1.98940963e-01 -2.35478580e-02 -6.20776176e-01 2.52044618e-01
1.69993639e-01 1.15326107e+00 -5.81859767e-01 3.60516533e-02
-7.55397439e-01 6.36478841e-01 6.15307271e-01 4.01982576e-01
-1.12350595e+00 -1.58939913e-01 5.20956039e-01 4.08321023e-01
5.56519866e-01 -2.34428663e-02 -2.08484486e-01 -1.01523173e+00
-1.03303909e-01 -6.76457942e-01 3.26812297e-01 -1.21757436e+00
-1.02486134e+00 8.80569100e-01 -9.94908810e-02 -1.10957980e+00
-2.58863330e-01 -1.00248861e+00 -7.32990384e-01 8.94865632e-01
-2.10534000e+00 -1.13470447e+00 -7.48216569e-01 9.13197041e-01
8.56188655e-01 -2.13402927e-01 4.65235651e-01 8.39611590e-02
-5.35676956e-01 4.81895506e-01 -3.31263572e-01 -2.94145066e-02
6.97479129e-01 -1.34999287e+00 5.88406801e-01 8.92416716e-01
4.06707764e-01 5.91730654e-01 6.73680305e-01 -6.86306879e-02
-9.38048303e-01 -7.34094203e-01 8.59189272e-01 -7.21408665e-01
6.02594018e-01 -1.33125246e-01 -9.93242443e-01 9.95030046e-01
9.91602421e-01 -5.97631931e-02 5.13134718e-01 3.96956682e-01
-4.69633579e-01 9.33371857e-02 -6.10452771e-01 6.02127016e-01
1.05368865e+00 -3.12443674e-01 -6.27990901e-01 1.79616502e-03
5.30201793e-01 -3.15891653e-01 -2.81803638e-01 1.11296982e-01
7.58435071e-01 -1.73025715e+00 8.01925302e-01 -9.14929867e-01
7.96413481e-01 -6.76821843e-02 2.88449585e-01 -1.16214323e+00
-7.97637701e-01 -3.82000029e-01 -2.98616529e-01 5.13670564e-01
3.45341235e-01 -3.11598629e-01 1.01215088e+00 7.92527944e-02
1.37933746e-01 -9.43324029e-01 -4.52711761e-01 -2.66720057e-01
7.36743882e-02 -1.48957046e-02 6.06247663e-01 6.01265132e-01
-4.97871220e-01 5.54366052e-01 -4.84857082e-01 -2.83076372e-02
4.06551301e-01 1.68905035e-02 8.21534514e-01 -1.63897598e+00
-8.49037338e-03 -7.65417635e-01 -5.14480770e-01 -1.10896730e+00
2.24732906e-01 -5.83612263e-01 -2.36585632e-01 -1.07647920e+00
3.76318216e-01 9.66709629e-02 -7.83940256e-01 4.08432990e-01
-1.26100361e-01 3.54868650e-01 3.44388247e-01 5.88939905e-01
-8.59072983e-01 -1.67871788e-02 1.13758814e+00 -3.21639813e-02
8.40439126e-02 6.69098645e-02 -7.42960095e-01 7.96866894e-01
8.21428657e-01 -4.11678612e-01 -5.23239434e-01 -9.60546494e-01
6.40870452e-01 -2.47570649e-01 4.67150241e-01 -1.07883894e+00
4.96288925e-01 7.92473033e-02 4.12615657e-01 -5.98376513e-01
6.53348118e-02 -8.81848037e-01 -2.94181585e-01 5.83257794e-01
-6.18515134e-01 3.97232264e-01 1.04189865e-01 5.25153160e-01
-2.55750030e-01 -4.06215370e-01 9.63759363e-01 -1.41546816e-01
-9.29906607e-01 5.39059997e-01 -2.04656839e-01 -1.34846553e-01
1.09564745e+00 -3.23709786e-01 -3.83558869e-01 -3.39083552e-01
-6.71442807e-01 3.69972736e-02 5.00711739e-01 7.59110034e-01
4.69168693e-01 -9.23439503e-01 -4.75989789e-01 1.92418799e-01
1.91996276e-01 -4.57154453e-01 6.50293976e-02 1.29506743e+00
-8.12882483e-01 8.34054112e-01 -9.11399484e-01 -8.52556646e-01
-1.23248053e+00 9.19809222e-01 3.96048844e-01 -3.29159200e-01
-2.41754651e-01 7.86813855e-01 5.01271844e-01 1.61591217e-01
3.02754939e-01 -2.58996278e-01 -9.00073349e-01 1.89423159e-01
5.86607039e-01 3.16728950e-01 1.14897668e-01 -6.86542690e-01
-2.94960737e-01 5.55791557e-01 -3.50269347e-01 3.62662107e-01
1.42433310e+00 -4.71372426e-01 1.38981612e-02 5.24321795e-01
1.10103166e+00 -3.90316308e-01 -1.50582874e+00 -2.14107290e-01
1.45739585e-01 -4.35974389e-01 8.76062438e-02 -5.76596558e-01
-1.38214791e+00 1.40076029e+00 8.28809202e-01 4.80673164e-01
1.24587595e+00 -2.01434776e-01 4.12289828e-01 2.75029302e-01
1.50013730e-01 -5.74690700e-01 3.27994049e-01 5.81013322e-01
6.28686488e-01 -1.32543778e+00 -2.22660124e-01 3.70977260e-03
-7.30447471e-01 1.38896894e+00 7.21090555e-01 -5.60068250e-01
6.43397272e-01 -1.39661089e-01 5.09896874e-02 -7.68571377e-01
-1.00611818e+00 -4.64113295e-01 3.73092532e-01 5.67231774e-01
5.81882894e-01 -3.03412676e-01 -1.89599589e-01 1.37057543e-01
-8.59879423e-03 1.83719739e-01 7.72665858e-01 4.73153859e-01
-4.52746123e-01 -7.79369891e-01 -1.52473480e-01 3.07706565e-01
-6.40670300e-01 -1.37295812e-01 -5.27930379e-01 1.05715549e+00
4.60758358e-02 6.58332527e-01 1.76906586e-01 -2.86443174e-01
1.43827155e-01 -2.76323594e-02 5.76680243e-01 -5.35308719e-01
-6.84293568e-01 -3.28029722e-01 -4.55648780e-01 -9.51500535e-01
-9.01558459e-01 -4.70631778e-01 -5.09560943e-01 -4.55014259e-01
-1.23724930e-01 -3.43819648e-01 4.60274845e-01 9.33566093e-01
2.40856528e-01 6.30127490e-01 5.74762225e-01 -1.03886008e+00
-2.85234392e-01 -1.09544396e+00 -2.74906248e-01 3.34238380e-01
6.23373330e-01 -8.77604008e-01 -1.82030380e-01 1.45631716e-01] | [10.066036224365234, 1.4815974235534668] |
b7b9e658-da68-463e-8362-d22169aa6d59 | spike-flownet-event-based-optical-flow | 2003.06696 | null | https://arxiv.org/abs/2003.06696v3 | https://arxiv.org/pdf/2003.06696v3.pdf | Spike-FlowNet: Event-based Optical Flow Estimation with Energy-Efficient Hybrid Neural Networks | Event-based cameras display great potential for a variety of tasks such as high-speed motion detection and navigation in low-light environments where conventional frame-based cameras suffer critically. This is attributed to their high temporal resolution, high dynamic range, and low-power consumption. However, conventional computer vision methods as well as deep Analog Neural Networks (ANNs) are not suited to work well with the asynchronous and discrete nature of event camera outputs. Spiking Neural Networks (SNNs) serve as ideal paradigms to handle event camera outputs, but deep SNNs suffer in terms of performance due to the spike vanishing phenomenon. To overcome these issues, we present Spike-FlowNet, a deep hybrid neural network architecture integrating SNNs and ANNs for efficiently estimating optical flow from sparse event camera outputs without sacrificing the performance. The network is end-to-end trained with self-supervised learning on Multi-Vehicle Stereo Event Camera (MVSEC) dataset. Spike-FlowNet outperforms its corresponding ANN-based method in terms of the optical flow prediction capability while providing significant computational efficiency. | ['Kaushik Roy', 'Adarsh Kumar Kosta', 'Kenneth Chaney', 'Kostas Daniilidis', 'Alex Zihao Zhu', 'Chankyu Lee'] | 2020-03-14 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/6736_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123740358.pdf | eccv-2020-8 | ['motion-detection', 'event-based-optical-flow'] | ['computer-vision', 'computer-vision'] | [ 1.29454479e-01 -5.67454755e-01 2.31981218e-01 -1.21985644e-01
-8.78179669e-02 -3.80276144e-01 5.29820919e-01 -1.84296936e-01
-7.64634907e-01 6.81358993e-01 -3.40599529e-02 6.30548820e-02
2.45037571e-01 -5.63600421e-01 -6.91408396e-01 -6.83896363e-01
2.92107940e-01 -2.13701218e-01 6.42620981e-01 1.12813219e-01
3.64287436e-01 5.58004081e-01 -1.73706579e+00 1.02595270e-01
5.51732779e-01 1.24116683e+00 2.71681875e-01 7.45001137e-01
-4.84892987e-02 1.30014932e+00 -3.34761202e-01 -1.53953984e-01
1.77042484e-01 -4.82830942e-01 4.77205329e-02 -3.88695806e-01
6.05936587e-01 -7.10814953e-01 -1.10449338e+00 9.05722022e-01
4.16330278e-01 9.77501646e-02 2.96296805e-01 -1.60972202e+00
-3.02580684e-01 -1.55620694e-01 -3.10014695e-01 7.85647929e-01
2.97666728e-01 6.58546865e-01 5.50461590e-01 -9.86453533e-01
5.89072108e-01 1.03028774e+00 7.91868687e-01 7.65449703e-01
-1.15809989e+00 -7.36514509e-01 -1.69838116e-01 5.07470429e-01
-1.01144767e+00 -7.89474189e-01 6.43192232e-01 -5.00557125e-01
1.36888146e+00 -4.50238466e-01 8.17174733e-01 1.41808856e+00
6.78864956e-01 7.22848892e-01 7.70636201e-01 3.54958832e-01
6.49401546e-01 -4.24208820e-01 -1.90053046e-01 6.33184552e-01
3.96394938e-01 3.21777850e-01 -1.06841505e+00 6.66617602e-02
1.14472258e+00 3.47597629e-01 -4.09956098e-01 -1.76035538e-01
-1.55152047e+00 5.00379682e-01 5.61147630e-01 -1.37678042e-01
-4.70468611e-01 7.72011578e-01 5.64278066e-01 1.42586023e-01
-1.16935872e-01 1.29688576e-01 -1.39272436e-01 -4.07722890e-01
-9.45496440e-01 1.64319292e-01 6.45949841e-01 7.91312575e-01
6.35929644e-01 7.18671083e-01 -2.28742901e-02 1.69012412e-01
1.93761125e-01 5.89252055e-01 6.86587453e-01 -1.42306364e+00
2.86529988e-01 5.09925783e-01 1.40283659e-01 -1.03325522e+00
-4.25096542e-01 -4.55391370e-02 -1.13736141e+00 7.21866608e-01
5.97171068e-01 -5.22706397e-02 -6.47281408e-01 1.73854160e+00
-1.66854680e-01 5.68050981e-01 2.72202432e-01 1.17422092e+00
8.38776410e-01 8.06310773e-01 -2.21771393e-02 -2.78851897e-01
9.95324969e-01 -7.56699443e-01 -7.06002533e-01 -6.97425723e-01
1.81552544e-01 -2.72418916e-01 7.63401926e-01 2.70010948e-01
-1.03436112e+00 -5.34551144e-01 -1.13760722e+00 -2.36438185e-01
-2.45393768e-01 8.70103091e-02 5.05580842e-01 2.90915817e-01
-1.08405709e+00 4.51536447e-01 -1.29884088e+00 -4.36934352e-01
6.27373993e-01 5.66140950e-01 -3.61113936e-01 1.97141677e-01
-7.57919014e-01 6.36801004e-01 2.25934729e-01 1.90967724e-01
-9.20995355e-01 -6.47816479e-01 -9.28356409e-01 1.91768304e-01
-1.17929414e-01 -8.99522722e-01 1.15735340e+00 -9.89596605e-01
-1.78486872e+00 5.51842749e-01 -3.87872159e-01 -8.92230451e-01
4.89170521e-01 -2.59678178e-02 -1.72841206e-01 5.67683399e-01
5.66267408e-02 1.06779134e+00 7.74195194e-01 -5.55877686e-01
-6.89003170e-01 -2.10853502e-01 -1.24377429e-01 -9.64306481e-03
-4.09632176e-01 -1.75346956e-01 2.93916650e-02 -3.47201288e-01
5.23488559e-02 -8.50975931e-01 -2.53001064e-01 6.13370240e-01
2.29283854e-01 1.32888600e-01 1.19779718e+00 -4.20029052e-02
8.43991637e-01 -2.06686544e+00 -2.51799256e-01 -5.60477376e-01
2.85316616e-01 5.30771077e-01 -5.45747541e-02 1.42480165e-01
2.56122261e-01 -5.99581838e-01 -9.90820080e-02 -1.38347775e-01
-4.48553503e-01 2.27142319e-01 -4.21442538e-01 6.06905937e-01
5.08080542e-01 1.02663136e+00 -1.01100993e+00 -4.33950514e-01
7.91171789e-01 5.96562862e-01 -4.52980578e-01 4.35692631e-02
-1.12568907e-01 6.55391455e-01 -1.77934840e-01 5.77240348e-01
4.45939928e-01 -4.20401245e-01 -2.04644173e-01 -2.84007400e-01
-3.06037217e-01 -6.99997619e-02 -1.02033567e+00 1.75316203e+00
-2.79750764e-01 1.30092216e+00 -7.62964636e-02 -7.70213544e-01
8.73136640e-01 2.75566071e-01 6.31973922e-01 -1.11040342e+00
3.36934865e-01 3.37680399e-01 3.99733782e-02 -5.71480870e-01
4.57592219e-01 6.88834935e-02 1.21551774e-01 2.03078791e-01
1.93535447e-01 3.20209086e-01 2.03154251e-01 1.82657734e-01
1.45746469e+00 2.26018682e-01 2.58511722e-01 -2.78334832e-04
4.77640718e-01 -3.77488211e-02 8.22905779e-01 4.98709738e-01
-7.15932667e-01 7.15959430e-01 2.87231416e-01 -7.77877748e-01
-1.10383844e+00 -1.10772073e+00 5.18690683e-02 5.42810977e-01
5.76744378e-01 -8.39097276e-02 -3.97242486e-01 1.27950728e-01
-1.42540976e-01 2.47485533e-01 -2.32120708e-01 -1.89017519e-01
-6.63945675e-01 -4.79088515e-01 7.59403408e-01 7.29460478e-01
7.63045013e-01 -1.19474328e+00 -1.53246546e+00 6.84016883e-01
-1.82681069e-01 -1.90143621e+00 -1.96733236e-01 1.84434816e-01
-9.28435147e-01 -9.77878094e-01 -5.64150035e-01 -7.64584661e-01
4.50203747e-01 4.90394920e-01 7.24997818e-01 -4.86061692e-01
-4.03542310e-01 2.75527000e-01 1.56632245e-01 -1.25685215e-01
-1.79251418e-01 -2.87561655e-01 1.87640414e-01 2.21157044e-01
5.63605249e-01 -8.23488057e-01 -1.03812099e+00 1.94655448e-01
-9.59190905e-01 4.65488136e-02 3.93161714e-01 8.74979556e-01
4.83424902e-01 -4.64186162e-01 5.93372703e-01 -1.14980519e-01
1.27908736e-01 -2.97866970e-01 -1.03796101e+00 -1.88693851e-01
-2.12620050e-01 -2.00013369e-02 1.13087416e+00 -5.11325479e-01
-1.01306129e+00 4.76183593e-01 1.28417566e-01 -7.31833816e-01
-1.89323246e-01 4.07148860e-02 4.10077870e-01 -3.83444428e-01
7.45112240e-01 3.31845731e-01 1.69144541e-01 2.90795416e-01
-2.18828812e-01 3.87531251e-01 1.09767127e+00 3.79084833e-02
3.65786433e-01 1.04794550e+00 4.21404570e-01 -8.07383120e-01
-4.43163216e-01 -4.33289438e-01 -3.96150589e-01 -6.01054609e-01
9.82650161e-01 -1.34900749e+00 -1.13606322e+00 8.92874360e-01
-1.50465512e+00 -2.11893439e-01 -6.67630360e-02 8.65381300e-01
-7.86234438e-01 2.77359873e-01 -9.11628306e-01 -8.63080680e-01
-2.49106556e-01 -1.10466397e+00 9.21707094e-01 6.78983212e-01
-1.73157956e-02 -8.54931056e-01 -7.00760111e-02 -3.52876224e-02
5.46747446e-01 4.64101017e-01 2.76656181e-01 -6.98943064e-02
-1.04160845e+00 -1.75034091e-01 -5.51092982e-01 1.09428652e-01
-1.92257404e-01 2.83013657e-02 -1.21908510e+00 -1.97850928e-01
2.08994397e-03 -3.94218832e-01 9.07814920e-01 7.42035747e-01
9.32159007e-01 5.82534112e-02 -2.13618755e-01 9.16956007e-01
1.86782169e+00 2.02408209e-01 7.42745459e-01 3.28902841e-01
5.17301679e-01 2.09891558e-01 4.98443693e-02 6.01230621e-01
3.27992886e-01 3.05326164e-01 8.42687845e-01 2.36580148e-01
-1.54584840e-01 -1.52182683e-01 7.54889369e-01 5.93767226e-01
7.90213495e-02 -3.96197587e-01 -8.36743712e-01 6.82528555e-01
-2.07113624e+00 -1.29248273e+00 -3.50488126e-01 2.12925267e+00
2.86432296e-01 3.14244211e-01 4.98549342e-02 6.45683929e-02
6.95662796e-01 1.76456288e-01 -8.92497838e-01 -2.67166823e-01
-5.10075986e-01 -8.55434760e-02 5.99750102e-01 -1.12768605e-01
-8.19405138e-01 6.87634408e-01 5.57679415e+00 2.44521931e-01
-1.40982652e+00 -1.31267319e-02 2.89585948e-01 -4.20316756e-01
2.87698776e-01 -1.20427266e-01 -8.18747222e-01 7.68094778e-01
1.25762701e+00 -4.81742695e-02 4.37020242e-01 5.30306220e-01
4.95276600e-01 -4.44867879e-01 -1.10767841e+00 1.55339503e+00
2.15417091e-02 -1.76372266e+00 -4.18046899e-02 -8.56730267e-02
6.76752448e-01 4.64994222e-01 -1.84263721e-01 -1.39046893e-01
-7.35151693e-02 -7.77046978e-01 7.00677633e-01 4.29180264e-01
7.85759866e-01 -3.71112943e-01 6.51260674e-01 3.21025074e-01
-1.20136678e+00 -3.84244263e-01 -4.84682947e-01 -4.82081741e-01
4.92575437e-01 4.08481121e-01 -1.69852212e-01 -1.17285743e-01
9.01205778e-01 1.13518894e+00 -2.36998051e-01 1.34465110e+00
3.34565461e-01 2.91602969e-01 -5.23036599e-01 -4.62172292e-02
3.76971155e-01 -2.30423406e-01 6.14952147e-01 1.10534728e+00
5.20945907e-01 -2.32703309e-03 -1.10526033e-01 1.04258406e+00
-5.99298514e-02 -5.83748519e-01 -9.37159777e-01 -1.44989183e-02
4.74310338e-01 1.30020916e+00 -9.46724415e-01 -1.93893507e-01
-8.75033915e-01 1.11497700e+00 3.84215452e-02 4.52764541e-01
-8.10695767e-01 -5.58217466e-01 9.12712216e-01 5.09145707e-02
3.72778773e-01 -4.78126287e-01 -3.24199706e-01 -1.31804371e+00
1.12475157e-01 -2.32833371e-01 -8.48489581e-04 -1.10311460e+00
-1.04314637e+00 5.89734316e-01 -6.11557245e-01 -1.79726231e+00
-3.73086065e-01 -8.58999968e-01 -5.97792506e-01 3.54416221e-01
-1.75846541e+00 -8.26550663e-01 -7.90507972e-01 8.14187527e-01
5.87855875e-01 -2.36316890e-01 5.28189540e-01 3.93025100e-01
-5.49400806e-01 1.61425993e-01 1.54874012e-01 2.99256742e-01
6.76698864e-01 -8.11060727e-01 4.17693466e-01 1.22852468e+00
-7.00312406e-02 9.81150419e-02 5.37258148e-01 -1.82089165e-01
-1.84053957e+00 -1.26884222e+00 7.16135144e-01 -2.21801773e-01
8.25896144e-01 -2.69267470e-01 -6.86205924e-01 5.43606281e-01
2.76587695e-01 7.51002610e-01 1.70843259e-01 -9.32281375e-01
-2.30274305e-01 -3.76094162e-01 -8.76659930e-01 5.87800264e-01
9.65027094e-01 -6.76182866e-01 -3.79856139e-01 -7.14432746e-02
3.39704067e-01 -4.06552762e-01 -4.23528463e-01 2.02109560e-01
5.68504810e-01 -1.39923525e+00 8.93311262e-01 -1.43367752e-01
6.32600427e-01 -5.94781101e-01 -1.30978292e-02 -9.69615638e-01
-1.88710496e-01 -7.98151910e-01 -3.48908961e-01 8.01750124e-01
-7.89203122e-02 -8.50051105e-01 9.52611446e-01 3.18525910e-01
-1.24417402e-01 -3.35748374e-01 -1.17707670e+00 -8.16287398e-01
-4.12197381e-01 -3.06708425e-01 1.35384491e-02 4.12966996e-01
-8.64116028e-02 4.16290998e-01 -3.19897324e-01 7.15041533e-02
9.12784159e-01 -1.44807459e-03 5.32268345e-01 -1.18839967e+00
3.82547155e-02 -4.99583602e-01 -1.05219424e+00 -1.13422132e+00
1.76916078e-01 -4.91298527e-01 1.54761821e-01 -1.28029418e+00
-3.84728089e-02 3.15611869e-01 -1.69028610e-01 1.31757870e-01
2.22668096e-01 6.22195482e-01 1.62638396e-01 3.37522328e-01
-8.22107553e-01 5.17103791e-01 8.71555328e-01 1.55773282e-01
-2.21140444e-01 -3.24937850e-01 -3.29693481e-02 7.31385410e-01
5.68810821e-01 -3.23434979e-01 -3.21031183e-01 -6.92264020e-01
2.96237916e-01 4.86637563e-01 9.67392325e-01 -1.65991890e+00
1.04422438e+00 7.01167583e-02 5.91506779e-01 -5.34299493e-01
4.56381857e-01 -8.29188585e-01 8.38177353e-02 6.70554817e-01
-1.76993653e-01 4.47745383e-01 2.47899950e-01 8.67667913e-01
-5.35258114e-01 1.82495594e-01 1.00925159e+00 -1.33273244e-01
-1.18287492e+00 5.13391018e-01 -1.00089455e+00 2.09511727e-01
1.01709914e+00 -8.75616193e-01 -6.79209173e-01 -3.28740597e-01
-2.00758249e-01 -2.16501970e-02 5.41761220e-01 2.83767760e-01
8.79178703e-01 -1.20714653e+00 -3.30627233e-01 5.14124572e-01
1.44438460e-01 -1.42585197e-02 1.43541485e-01 9.85450149e-01
-8.83066714e-01 5.45822978e-01 -9.53807116e-01 -1.02556515e+00
-6.00225210e-01 3.06831688e-01 4.18123007e-01 2.77325302e-01
-8.48686457e-01 4.99623537e-01 1.90658286e-01 2.07146078e-01
2.69504339e-01 -3.78317714e-01 7.72445053e-02 -1.51822031e-01
5.60870409e-01 5.98552406e-01 -5.59390374e-02 -4.77672786e-01
-4.75139737e-01 5.71236074e-01 3.75253230e-01 5.55723049e-02
1.24985206e+00 -3.45560193e-01 1.68173686e-01 4.01217073e-01
1.21836698e+00 -7.69572854e-01 -2.06664205e+00 -9.32217613e-02
-1.08879924e-01 -2.17786431e-01 2.61785775e-01 -1.96386784e-01
-1.27235305e+00 1.12802660e+00 6.16860032e-01 -1.03670456e-01
1.22917819e+00 -5.63541353e-01 1.29552913e+00 5.85242212e-01
3.82417709e-01 -9.80005682e-01 2.86208481e-01 6.39533341e-01
1.86471209e-01 -1.42242062e+00 -3.59942436e-01 9.35839042e-02
-3.83103877e-01 1.60432589e+00 8.20399463e-01 -6.13650858e-01
4.48037237e-01 6.52390540e-01 -5.68564981e-02 9.57551673e-02
-1.15008044e+00 -1.75982863e-01 -4.06686068e-01 6.02599442e-01
-5.99063076e-02 -5.30599415e-01 1.96512446e-01 1.09162264e-01
2.00390965e-01 6.14486396e-01 8.48473907e-01 8.47302973e-01
-3.87232035e-01 -2.62492150e-01 -7.90443048e-02 3.77294272e-01
-4.19338912e-01 -2.92825382e-02 4.40593772e-02 4.44443047e-01
-1.89521581e-01 8.85729611e-01 5.22665262e-01 -2.46103689e-01
1.55719295e-01 -2.44850233e-01 4.68453556e-01 3.62382419e-02
-5.40052176e-01 -1.98052719e-01 -4.33927059e-01 -9.14487660e-01
-7.60172248e-01 -5.48212409e-01 -1.40786672e+00 -5.34958482e-01
1.15288861e-01 -5.34315825e-01 5.98491132e-01 8.80995989e-01
5.79096854e-01 3.39001030e-01 4.80578870e-01 -9.56802666e-01
-1.83825254e-01 -4.07599777e-01 -2.94761211e-01 3.70041788e-01
7.26105869e-01 -4.37891841e-01 -5.41251123e-01 4.30521846e-01] | [8.665432929992676, -1.1606724262237549] |
9da51c36-2be5-4a6b-99ac-ede16d9bb2a3 | all-in-1-short-text-classification-with-one | 1710.09589 | null | http://arxiv.org/abs/1710.09589v1 | http://arxiv.org/pdf/1710.09589v1.pdf | ALL-IN-1: Short Text Classification with One Model for All Languages | We present ALL-IN-1, a simple model for multilingual text classification that
does not require any parallel data. It is based on a traditional Support Vector
Machine classifier exploiting multilingual word embeddings and character
n-grams. Our model is simple, easily extendable yet very effective, overall
ranking 1st (out of 12 teams) in the IJCNLP 2017 shared task on customer
feedback analysis in four languages: English, French, Japanese and Spanish. | ['Barbara Plank'] | 2017-10-26 | null | null | null | null | ['multilingual-word-embeddings', 'multilingual-text-classification'] | ['methodology', 'miscellaneous'] | [-6.46601260e-01 -3.33499402e-01 -7.01109052e-01 -4.22266245e-01
-9.76532161e-01 -9.08014357e-01 7.56743252e-01 9.43792522e-01
-1.04385781e+00 8.53858888e-01 5.71303606e-01 -1.05537117e+00
1.65602267e-01 -1.39712244e-01 -3.00354809e-01 -1.36485711e-01
1.84118718e-01 9.39291239e-01 -7.09674880e-02 -8.56588364e-01
2.26755694e-01 8.75124186e-02 -8.14409316e-01 5.46939671e-01
8.18870187e-01 5.79648077e-01 -1.23397291e-01 1.06838977e+00
-2.75093526e-01 6.95748389e-01 -1.59408614e-01 -8.67186010e-01
-7.75237978e-02 1.51819095e-01 -1.07621217e+00 -3.31866592e-01
5.63098669e-01 5.09694964e-02 -1.22067235e-01 6.24049664e-01
6.32060528e-01 -1.44038513e-01 8.96696627e-01 -8.64226222e-01
-1.19726455e+00 1.07894015e+00 -4.94846255e-01 6.25412226e-01
8.82330775e-01 -3.54148954e-01 1.79396176e+00 -1.43296432e+00
8.70000124e-01 1.42926490e+00 9.26230311e-01 9.98758301e-02
-1.36647856e+00 -5.17234623e-01 3.78661364e-01 2.32581109e-01
-8.57307196e-01 -1.53092474e-01 2.62273103e-01 -7.26321042e-01
1.67660081e+00 1.91840246e-01 4.51581657e-01 1.35000265e+00
4.74360764e-01 1.17645264e+00 1.24694097e+00 -1.04641712e+00
-3.56722504e-01 7.79769003e-01 7.79970706e-01 7.45799720e-01
5.80109050e-03 -2.95489520e-01 -6.61724210e-01 -4.98860836e-01
-9.48310494e-02 -2.37379223e-01 -3.15425217e-01 -2.50880867e-01
-1.24291527e+00 1.39771509e+00 5.03122993e-02 5.70615053e-01
-2.99461409e-02 -4.75598663e-01 7.40828216e-01 1.02627826e+00
9.89408493e-01 4.92383689e-01 -1.26382351e+00 -1.00662969e-01
-6.45280361e-01 1.96358114e-01 9.35698211e-01 5.82813323e-01
5.76416492e-01 -1.75445452e-01 2.16982812e-01 1.25018716e+00
3.67098242e-01 4.04535860e-01 1.04479420e+00 -7.53027573e-02
6.92529678e-01 5.94675124e-01 -8.62552822e-02 -7.15011358e-01
-7.77768612e-01 -3.61547530e-01 -2.66157746e-01 -1.83691889e-01
6.30504116e-02 -2.34678537e-01 -1.93499282e-01 9.34804916e-01
-5.52888438e-02 -6.46011591e-01 2.41402611e-01 6.52923763e-01
9.06530738e-01 5.41114450e-01 -2.36327842e-01 1.16147690e-01
1.29586387e+00 -1.58813131e+00 -5.03192067e-01 -2.39602730e-01
1.29407084e+00 -1.40495634e+00 1.16447043e+00 7.57685363e-01
-5.53189874e-01 -3.59905154e-01 -1.09223223e+00 -4.53693181e-01
-8.59992564e-01 2.42915526e-01 7.68481672e-01 7.95696735e-01
-1.16324258e+00 2.07174972e-01 -2.46661112e-01 -7.26132810e-01
-1.34948924e-01 1.63232714e-01 -9.13529515e-01 -3.64887685e-01
-1.24277198e+00 1.25727141e+00 3.46310288e-01 -4.40771341e-01
-2.50411838e-01 -6.74159348e-01 -1.05075634e+00 -5.39859414e-01
-3.51463675e-01 -1.08584382e-01 1.52448654e+00 -9.58931625e-01
-1.33020842e+00 1.14189374e+00 -2.64630497e-01 -3.10371578e-01
4.12570983e-01 -6.08733892e-01 -1.02793992e+00 -4.19840246e-01
3.16238433e-01 4.20702130e-01 4.62945133e-01 -6.93988323e-01
-8.19555283e-01 -2.87993193e-01 -3.29948843e-01 2.13815510e-01
-7.69206226e-01 7.11764574e-01 5.11354953e-02 -6.74893379e-01
-2.40970448e-01 -8.30273509e-01 -1.82462677e-01 -6.48801804e-01
-2.10850134e-01 -8.14173222e-01 4.28421676e-01 -1.00443745e+00
1.20422447e+00 -1.78500748e+00 7.42617026e-02 -1.77580595e-01
3.29857655e-02 8.46952423e-02 -4.35235381e-01 8.93803239e-01
-3.58753711e-01 3.07147801e-01 4.75178182e-01 -4.43312317e-01
1.84126049e-01 8.38497356e-02 -3.58567178e-01 6.83976054e-01
3.39204490e-01 9.00445759e-01 -1.06484377e+00 -3.42172533e-01
7.65514094e-03 2.47527987e-01 -5.56820393e-01 -1.17860585e-01
1.31866440e-01 8.85205492e-02 -1.99849214e-02 6.91549659e-01
4.60263640e-01 3.08755487e-01 5.54308891e-01 5.67200072e-02
-4.74198461e-01 7.78115213e-01 -8.11093986e-01 1.69469857e+00
-7.31391490e-01 1.01048613e+00 -4.70303707e-02 -9.01935041e-01
9.09443736e-01 6.07126772e-01 1.54351816e-01 -8.33860159e-01
1.63001433e-01 8.33953261e-01 -4.60974798e-02 -3.71912599e-01
5.79314828e-01 4.87573780e-02 -6.07069790e-01 8.57477307e-01
5.94439089e-01 6.88531920e-02 4.98564124e-01 3.93392742e-01
7.40993381e-01 -1.13088310e-01 6.42072856e-01 -6.88505411e-01
4.94536161e-01 1.01834111e-01 1.71431124e-01 5.15470147e-01
-2.96240687e-01 2.62362331e-01 3.50237250e-01 -8.89522851e-01
-1.08335841e+00 -7.95990646e-01 -4.91294622e-01 2.08255243e+00
-7.04558671e-01 -7.30935454e-01 -4.67332304e-02 -1.14125216e+00
4.24322873e-01 6.08270109e-01 -4.72384900e-01 4.57847059e-01
-4.19567138e-01 -8.03086996e-01 5.15663445e-01 3.13148469e-01
-5.42449534e-01 -8.94272029e-01 1.92957580e-01 4.47877705e-01
-1.88333526e-01 -1.00788629e+00 -7.10726917e-01 7.74784625e-01
-7.22074270e-01 -1.06605589e+00 -5.51234782e-01 -1.35785472e+00
2.42719993e-01 6.13832697e-02 1.67449093e+00 -3.62676471e-01
-2.96360672e-01 2.83506125e-01 -7.65777349e-01 -5.50585687e-01
-3.79752904e-01 5.66230297e-01 5.06131351e-01 -2.74755269e-01
1.00062096e+00 2.51002871e-02 1.37237608e-02 3.46098170e-02
-2.91930646e-01 -5.52780628e-01 5.17021656e-01 1.10042489e+00
7.65900835e-02 -4.40476000e-01 6.87666953e-01 -1.10116303e+00
1.24117970e+00 -6.10932767e-01 -2.69326959e-02 2.96407968e-01
-9.02613580e-01 -4.69809212e-02 7.60169566e-01 -3.65246117e-01
-3.91008437e-01 -2.59271324e-01 -4.65126991e-01 5.01604617e-01
2.44008861e-02 7.42626131e-01 4.68664736e-01 1.59709737e-01
9.29238617e-01 -1.08196018e-02 -3.96302938e-01 -8.83953869e-01
7.63617992e-01 1.17807209e+00 3.17814648e-02 -4.10470933e-01
3.76180619e-01 -1.21425271e-01 -9.58018661e-01 -7.25606799e-01
-8.59071612e-01 -1.09069681e+00 -1.04098666e+00 1.64389625e-01
6.58325911e-01 -1.17093849e+00 -2.80470192e-01 1.86004579e-01
-1.08042657e+00 5.23353256e-02 2.49466524e-01 7.68169880e-01
1.03815898e-01 1.81797072e-01 -1.27482510e+00 -6.09636724e-01
-5.52567422e-01 -8.74850214e-01 7.63714373e-01 -4.50522870e-01
-7.28493452e-01 -1.61275756e+00 6.39073372e-01 3.75267684e-01
4.27489370e-01 -5.03088951e-01 1.15631354e+00 -1.26734614e+00
6.08558238e-01 -5.11031628e-01 -3.71234417e-02 7.22353697e-01
-5.75532392e-02 1.53767678e-03 -8.18380892e-01 -6.65723383e-01
-3.59718978e-01 -1.03867447e+00 1.14087534e+00 -7.44845122e-02
4.36068863e-01 -3.28152359e-01 -3.73854116e-02 -7.41123334e-02
1.36916649e+00 -3.67478073e-01 -1.64353877e-01 6.90723956e-01
6.74100041e-01 6.19335473e-01 3.08717012e-01 2.29628205e-01
8.69012237e-01 4.43362743e-01 7.68442452e-02 -8.43253210e-02
3.17094117e-01 -2.30823368e-01 8.51767838e-01 1.86701620e+00
4.25426453e-01 -1.58378810e-01 -1.06683588e+00 7.43493497e-01
-1.74153614e+00 -4.22179997e-01 -6.58356905e-01 1.81841910e+00
1.03814089e+00 1.03107855e-01 2.36855030e-01 2.50725985e-01
1.03908315e-01 1.11018375e-01 6.73975721e-02 -1.42350233e+00
-4.41045374e-01 6.42013490e-01 6.01259112e-01 7.67719567e-01
-1.18822932e+00 9.51438665e-01 7.23585176e+00 5.18246531e-01
-8.15234840e-01 6.61675394e-01 3.11289012e-01 -1.34873912e-02
-4.76604432e-01 -8.68609920e-02 -1.11177790e+00 1.24243766e-01
1.19647312e+00 -8.13724026e-02 6.30432740e-02 7.71381974e-01
-3.72914314e-01 1.10112324e-01 -1.25104880e+00 7.16354728e-01
5.16888499e-01 -1.16355383e+00 2.05969781e-01 -2.55201489e-01
8.75398517e-01 7.05347061e-01 7.29824901e-02 8.82439256e-01
7.19301939e-01 -1.14051831e+00 7.07038283e-01 7.46240169e-02
4.85392511e-01 -9.49317098e-01 1.04185283e+00 4.55264360e-01
-8.20639551e-01 -1.70097828e-01 -3.39384079e-01 -1.55405462e-01
1.16586350e-01 5.02877176e-01 -8.52659345e-01 4.93983358e-01
8.31559479e-01 1.06608558e+00 -7.23195553e-01 4.47764605e-01
-2.15861633e-01 8.05463910e-01 9.30903628e-02 -1.76042289e-01
4.82405335e-01 1.97233051e-01 2.97045588e-01 1.98493493e+00
2.75053382e-02 -8.86881649e-01 7.11563945e-01 -3.10440272e-01
-1.91048652e-01 1.10290408e+00 -6.22664571e-01 -1.08209730e-03
1.78210884e-02 1.30596352e+00 -2.11541086e-01 -1.50299907e-01
-8.46293807e-01 1.01255369e+00 8.19373071e-01 -5.29214274e-04
-2.09972635e-01 -5.29059529e-01 2.72723615e-01 -8.33803117e-02
2.33518228e-01 -1.71268210e-01 -9.90328658e-03 -1.34836030e+00
-2.07721237e-02 -1.27583086e+00 6.01204395e-01 -3.06215078e-01
-1.76855898e+00 8.63717198e-01 -7.95696437e-01 -8.86674047e-01
-2.69788355e-01 -1.38057649e+00 -1.01884797e-01 1.08342516e+00
-1.90712667e+00 -1.27214193e+00 5.54480374e-01 5.63147962e-01
6.04445994e-01 -8.00337851e-01 1.44070899e+00 6.33058548e-01
-5.92621267e-01 9.49105680e-01 5.58607697e-01 1.86933890e-01
1.45134985e+00 -1.67147899e+00 3.46609473e-01 1.77025825e-01
4.86673176e-01 5.45058429e-01 3.86476666e-01 -1.78847998e-01
-1.22181320e+00 -6.82091951e-01 2.16403818e+00 -9.87922251e-01
1.51543486e+00 -6.99425638e-01 -4.14403528e-01 9.32127774e-01
8.53931606e-01 -2.15330884e-01 1.35927701e+00 9.23192561e-01
-9.67904449e-01 4.17216122e-03 -6.75023496e-01 3.19101423e-01
1.91451862e-01 -8.80537629e-01 -7.91089475e-01 7.90170014e-01
7.46353626e-01 1.17039084e-01 -1.00009978e+00 -2.62618251e-02
6.70181096e-01 -6.44234717e-01 6.62924111e-01 -1.08634698e+00
4.65857446e-01 3.30258191e-01 -4.75775778e-01 -1.82161772e+00
-4.70443726e-01 -3.42013568e-01 3.10230702e-01 1.02372313e+00
1.14426017e+00 -9.85896587e-01 1.52661696e-01 -4.15637285e-01
-1.19202636e-01 -6.94772124e-01 -9.85647202e-01 -7.37722576e-01
7.73055792e-01 -7.01052368e-01 7.57829323e-02 1.47323942e+00
6.31746292e-01 9.74493384e-01 -2.55173326e-01 -5.05638123e-01
1.71200484e-01 -2.02814266e-01 5.69239080e-01 -1.59454858e+00
-3.90324980e-01 -8.11348975e-01 -3.35588038e-01 -1.00126135e+00
4.14777160e-01 -1.64508700e+00 -3.81334573e-01 -1.56132495e+00
2.50785023e-01 -2.57771224e-01 -6.16751313e-01 4.44873810e-01
-1.01161525e-01 3.22345436e-01 -8.35605338e-02 1.15739286e-01
-8.40225220e-01 8.40783715e-02 6.11085057e-01 -3.99672061e-01
3.82779360e-01 -2.32560605e-01 -8.89998138e-01 4.26889986e-01
5.76972246e-01 -6.51622355e-01 1.08662471e-01 -7.80800521e-01
7.62467563e-01 -4.25053000e-01 -2.20056757e-01 -2.43347347e-01
-1.15963146e-01 2.74784476e-01 2.89321482e-01 -4.32660013e-01
-2.10216254e-01 -5.76245606e-01 -7.07738400e-01 3.26854974e-01
-4.88990963e-01 8.62462044e-01 2.29730099e-01 4.35390294e-01
-4.33528692e-01 -4.55609709e-01 3.19596559e-01 2.31868271e-02
-6.25939786e-01 -4.40363847e-02 -8.01334202e-01 8.17066282e-02
6.24599099e-01 5.16336262e-01 -4.25816029e-01 -3.29480506e-02
-6.71798289e-01 5.17139435e-01 1.07301772e-01 9.82361019e-01
1.97075829e-01 -1.47261298e+00 -1.33238602e+00 2.78317332e-01
7.99065709e-01 -1.07072425e+00 -2.85246402e-01 9.37723696e-01
-5.51537931e-01 1.05224407e+00 8.71700570e-02 -3.26952010e-01
-1.18051267e+00 3.81338745e-01 -3.61604725e-05 -8.96168292e-01
-1.17300741e-01 1.15953660e+00 -6.14267170e-01 -1.54828143e+00
8.53677988e-02 -2.74273753e-01 -5.77528119e-01 7.07357109e-01
6.11858308e-01 2.13694111e-01 5.92730522e-01 -6.99263453e-01
-6.50798261e-01 3.34451467e-01 -5.71877778e-01 -4.00081605e-01
1.32563925e+00 9.20174737e-03 -5.82705915e-01 1.27681887e+00
1.61718357e+00 3.74472260e-01 -6.47869557e-02 -4.16681260e-01
5.30184388e-01 -2.69563925e-02 1.30663544e-01 -8.91555786e-01
-4.84350353e-01 9.77743149e-01 5.58420420e-01 3.18936616e-01
2.38862216e-01 1.16851386e-02 7.96174765e-01 5.82922280e-01
3.97184342e-01 -1.44221318e+00 -2.14326270e-02 1.06102991e+00
8.76503885e-01 -1.54553151e+00 -8.70338529e-02 3.56945008e-01
-6.84149861e-01 1.43490934e+00 2.39961952e-01 -1.08764805e-01
1.01856041e+00 1.30184054e-01 6.39623106e-01 5.05203754e-02
-1.14162135e+00 7.13732913e-02 4.96695846e-01 4.76938725e-01
1.36187077e+00 4.37826574e-01 -6.82371736e-01 6.13641798e-01
-4.83728021e-01 -4.74870652e-01 2.56117374e-01 8.96149933e-01
-3.46922666e-01 -1.55929077e+00 1.42160291e-02 5.72597563e-01
-8.79117727e-01 -6.84800327e-01 -4.65670258e-01 6.12292469e-01
-1.01674370e-01 1.19345212e+00 -2.45520957e-02 -4.52058375e-01
4.54220802e-01 3.90919805e-01 2.15437442e-01 -8.04252267e-01
-9.14601624e-01 -1.12088673e-01 5.53530276e-01 -1.19657457e-01
-2.15787292e-01 -9.24649239e-01 -6.42110169e-01 -4.08065379e-01
-3.74861002e-01 2.36015528e-01 9.32725966e-01 8.82780015e-01
4.38307747e-02 1.69386521e-01 7.50190794e-01 -3.89997423e-01
-5.66110492e-01 -1.46846414e+00 -5.72432995e-01 1.84045285e-01
3.54609460e-01 -2.99025059e-01 -5.31020761e-01 -2.75165796e-01] | [10.794553756713867, 9.83547306060791] |
4b3fb386-e3f0-4722-9276-eb3faadef1ed | trustguard-gnn-based-robust-and-explainable | 2306.13339 | null | https://arxiv.org/abs/2306.13339v1 | https://arxiv.org/pdf/2306.13339v1.pdf | TrustGuard: GNN-based Robust and Explainable Trust Evaluation with Dynamicity Support | Trust evaluation assesses trust relationships between entities and facilitates decision-making. Machine Learning (ML) shows great potential for trust evaluation owing to its learning capabilities. In recent years, Graph Neural Networks (GNNs), as a new ML paradigm, have demonstrated superiority in dealing with graph data. This has motivated researchers to explore their use in trust evaluation, as trust relationships among entities can be modeled as a graph. However, current trust evaluation methods that employ GNNs fail to fully satisfy the dynamicity nature of trust, overlook the adverse effects of attacks on trust evaluation, and cannot provide convincing explanations on evaluation results. To address these problems, in this paper, we propose TrustGuard, a GNN-based accurate trust evaluation model that supports trust dynamicity, is robust against typical attacks, and provides explanations through visualization. Specifically, TrustGuard is designed with a layered architecture that contains a snapshot input layer, a spatial aggregation layer, a temporal aggregation layer, and a prediction layer. Among them, the spatial aggregation layer can be plugged into a defense mechanism for a robust aggregation of local trust relationships, and the temporal aggregation layer applies an attention mechanism for effective learning of temporal patterns. Extensive experiments on two real-world datasets show that TrustGuard outperforms state-of-the-art GNN-based trust evaluation models with respect to trust prediction across single-timeslot and multi-timeslot, even in the presence of attacks. In particular, TrustGuard can explain its evaluation results by visualizing both spatial and temporal views. | ['Witold Pedrycz', 'Elisa Bertino', 'Jiahe Lan', 'Zheng Yan', 'Jie Wang'] | 2023-06-23 | null | null | null | null | ['decision-making'] | ['reasoning'] | [-1.03584170e+00 -1.63206100e-01 -3.05671275e-01 -4.59657878e-01
4.55063015e-01 -2.71611452e-01 4.27922219e-01 5.60358942e-01
6.63459301e-02 4.19881701e-01 -8.94372091e-02 -6.96260870e-01
-1.46227181e-01 -9.07888949e-01 -3.78818899e-01 -3.21732879e-01
-7.43442833e-01 -4.41620499e-02 5.27821541e-01 -3.99250329e-01
1.57699153e-01 3.85568440e-01 -9.99488294e-01 -3.28026824e-02
1.15952623e+00 1.23016357e+00 -6.71727598e-01 3.34018111e-01
3.78952950e-01 1.12802136e+00 -6.93134129e-01 -6.25037551e-01
-1.34299174e-01 -7.88799077e-02 -3.53270531e-01 -5.65459549e-01
-2.09403560e-01 -4.32658553e-01 -4.53974694e-01 9.08934057e-01
2.09554434e-01 -2.50609726e-01 6.95216954e-02 -2.25173998e+00
-9.72804427e-01 8.73124421e-01 -4.39802170e-01 1.38370320e-01
4.22529966e-01 1.41855851e-01 1.13692212e+00 -7.99228013e-01
2.14553356e-01 1.18622565e+00 7.24275291e-01 2.66233802e-01
-1.02414894e+00 -8.65479112e-01 6.44621730e-01 6.18643105e-01
-1.02115273e+00 -3.22792795e-03 9.59774017e-01 4.91732694e-02
1.07770860e+00 3.34664971e-01 1.09026361e+00 1.11274493e+00
7.16703117e-01 8.76702666e-01 1.16345763e+00 7.87937492e-02
4.12878364e-01 2.74332762e-01 3.32910448e-01 8.15080047e-01
6.13184988e-01 6.64467454e-01 -8.24675083e-01 -5.68553805e-01
5.68931818e-01 1.74517840e-01 -2.72755355e-01 -6.15823030e-01
-9.00352776e-01 8.07300091e-01 1.11281729e+00 1.17972665e-01
-1.12257570e-01 2.00038403e-01 7.12407291e-01 7.82916963e-01
6.14659309e-01 3.54583532e-01 -5.31228364e-01 4.97189611e-02
-6.78619266e-01 -2.91675270e-01 1.18941486e+00 5.53526044e-01
3.30485374e-01 5.25692523e-01 1.56825557e-02 8.84090960e-02
8.53826404e-01 3.15940678e-01 1.26318276e-01 -4.54313189e-01
6.67566136e-02 1.11372089e+00 -1.85668170e-02 -1.64463246e+00
-5.64759076e-01 -6.76688492e-01 -1.14649522e+00 5.86005926e-01
-1.64383128e-01 -6.74792007e-02 -5.33391654e-01 1.61090815e+00
5.74855387e-01 6.96733832e-01 2.48552695e-01 8.99033964e-01
7.80500412e-01 2.98055291e-01 -2.85616200e-02 -2.46553287e-01
9.25470769e-01 -1.02455914e+00 -8.48196745e-01 1.15765572e-01
5.32030165e-01 -2.73579478e-01 8.06694031e-01 4.93159354e-01
-7.29397535e-01 -3.17952603e-01 -1.57388055e+00 5.16906857e-01
-5.29209971e-01 -2.52338558e-01 1.04876149e+00 7.28136480e-01
-1.31549931e+00 8.04603040e-01 -9.30271626e-01 -8.51586983e-02
-2.51183733e-02 3.91333729e-01 -2.51556963e-01 1.63457781e-01
-1.83751595e+00 1.05260968e+00 6.80536404e-02 6.53566658e-01
-1.06753552e+00 -3.54026854e-01 -8.72461915e-01 2.29184464e-01
5.52727461e-01 -5.40958405e-01 9.19215083e-01 -7.86278486e-01
-1.16144764e+00 -9.69976187e-02 6.14398301e-01 -4.97469634e-01
4.68579203e-01 -1.28138140e-01 -1.16420245e+00 -3.94542180e-02
-2.72101283e-01 -2.97355264e-01 9.01442409e-01 -1.38084412e+00
-4.00764853e-01 -2.35994905e-01 4.63928729e-01 -1.25561282e-01
-5.22023380e-01 1.12355940e-01 5.73315145e-03 -4.70792979e-01
1.41797200e-01 -7.28254735e-01 -2.60002792e-01 5.87364063e-02
-4.89236593e-01 -7.60482624e-02 1.27492416e+00 -2.39077941e-01
1.74633706e+00 -1.84411466e+00 -2.95463949e-01 8.48796904e-01
7.40684330e-01 3.38686705e-01 1.13436714e-01 8.15552056e-01
-1.15579545e-01 3.96593302e-01 3.42674077e-01 -2.47911200e-01
1.77273750e-01 2.57689327e-01 -1.58817858e-01 4.94467705e-01
-5.74455820e-02 8.99277091e-01 -1.14513767e+00 -3.42432767e-01
-1.31679520e-01 2.95573652e-01 -1.60589665e-01 3.53230208e-01
-1.02932394e-01 -1.34848967e-01 -5.27095914e-01 9.44199979e-01
6.81591809e-01 -5.28271437e-01 6.15889490e-01 -3.12533766e-01
-1.58579592e-02 -4.00002263e-02 -1.29262280e+00 8.65168154e-01
-9.60428119e-02 3.01742315e-01 1.89691842e-01 -4.98196810e-01
1.12875473e+00 6.13225698e-01 4.24696624e-01 -5.62522829e-01
1.10604897e-01 7.74755850e-02 6.02856874e-02 -1.34522676e-01
3.61630946e-01 4.36473906e-01 -8.75851214e-02 1.10992289e+00
-1.45671025e-01 2.45428562e-01 -1.61506217e-02 6.85837328e-01
1.25119448e+00 -2.49095559e-01 1.92139208e-01 6.79906607e-02
3.58293921e-01 -3.37523162e-01 7.16283560e-01 5.45945942e-01
-4.64215398e-01 -2.09439412e-01 1.03002834e+00 -9.24598873e-01
-4.71633822e-01 -7.39443064e-01 2.80048192e-01 7.84380078e-01
4.40305412e-01 -9.13883269e-01 -2.52880484e-01 -1.17609358e+00
4.48503077e-01 5.98302662e-01 -9.97051299e-01 -8.74069691e-01
-4.71589305e-02 -4.39210176e-01 6.04922712e-01 6.58636093e-01
6.67187989e-01 -1.09228730e+00 -4.73962545e-01 1.60861582e-01
1.87299415e-01 -7.21753597e-01 -3.70180726e-01 1.10382542e-01
-7.40126073e-01 -1.44352698e+00 3.84279758e-01 -1.93327636e-01
7.49174595e-01 4.57382858e-01 1.32348883e+00 9.98495936e-01
2.50116527e-01 5.50222814e-01 -5.02503633e-01 -3.26004893e-01
-3.80604386e-01 -3.90548646e-01 6.02461755e-01 1.23775773e-01
2.81579159e-02 -7.78500795e-01 -5.18085539e-01 7.67061591e-01
-7.71152914e-01 -1.72382310e-01 4.33390886e-01 9.25832093e-01
-4.13752161e-02 1.86772361e-01 6.21706009e-01 -8.51358473e-01
1.34832752e+00 -7.21580744e-01 -6.21573508e-01 8.18568707e-01
-1.73593855e+00 -9.92915258e-02 5.38259327e-01 -3.54436755e-01
-6.04325414e-01 -7.41377831e-01 6.14418983e-01 -7.64993131e-01
6.38595521e-01 1.13232386e+00 2.25714501e-02 -4.46581125e-01
6.37035608e-01 -1.16882615e-01 1.22191742e-01 1.23346396e-01
2.70414710e-01 2.50299931e-01 7.63127804e-02 -6.01072729e-01
1.01103854e+00 1.90736622e-01 5.05958758e-02 1.22429848e-01
-4.45900053e-01 2.45990660e-02 -5.63792229e-01 -6.66295648e-01
1.97892934e-01 -7.06498921e-01 -1.18106318e+00 4.17647898e-01
-8.56377780e-01 -1.49844855e-01 9.59164277e-02 2.49564499e-01
-3.54708685e-03 3.67498815e-01 -8.42527211e-01 -9.25320089e-01
-6.78877950e-01 -8.10763299e-01 1.02770858e-01 2.99150199e-01
-3.95237468e-02 -1.30130243e+00 3.40073854e-02 -7.01587349e-02
6.32526815e-01 5.41304767e-01 7.28685617e-01 -6.95253789e-01
-6.27305627e-01 -4.34746206e-01 -1.98872089e-01 4.11765963e-01
8.75351951e-02 7.59638727e-01 -5.32850564e-01 -6.23710811e-01
-7.44070485e-02 -4.14577007e-01 3.50564986e-01 -9.37716514e-02
6.19299114e-01 -6.54821992e-01 -2.84114748e-01 3.09694171e-01
1.11612010e+00 -1.43764969e-02 2.44378656e-01 3.33082169e-01
6.10201955e-01 1.94478691e-01 8.76029074e-01 7.22772539e-01
8.86197269e-01 2.97549784e-01 1.25376964e+00 -3.81859303e-01
5.79858363e-01 -3.39829594e-01 4.12776947e-01 9.14349139e-01
-9.88721848e-02 -1.41497120e-01 -8.76251042e-01 1.43162161e-01
-2.45604277e+00 -8.94060552e-01 -2.68215239e-01 2.07952356e+00
5.47620118e-01 5.17885327e-01 7.47479573e-02 2.80305564e-01
3.31705898e-01 3.06203514e-01 -7.91747928e-01 -5.93169749e-01
9.10734758e-03 -3.58839273e-01 2.90299989e-02 1.59070730e-01
-7.25131392e-01 7.39818215e-01 6.21435308e+00 1.77775770e-01
-1.26444471e+00 -4.89069894e-02 4.28720236e-01 1.24942154e-01
-4.92178380e-01 1.96639895e-01 -7.91930780e-02 3.32060963e-01
9.87286568e-01 -3.71506125e-01 1.63330525e-01 9.38832343e-01
2.32707933e-01 4.11233269e-02 -1.06247032e+00 4.66526926e-01
-2.10835844e-01 -1.09731495e+00 -1.59869105e-01 -2.24418133e-01
4.92279917e-01 -1.45524248e-01 1.63138613e-01 3.06537628e-01
9.80314255e-01 -9.65771198e-01 5.70482194e-01 6.37518227e-01
2.84212083e-01 -1.00209558e+00 9.83335197e-01 1.43234059e-01
-1.24610960e+00 -3.47455174e-01 -1.66653499e-01 -1.75972313e-01
-4.88872826e-02 6.01976633e-01 -7.91738510e-01 1.01355577e+00
8.97260427e-01 8.66272271e-01 -7.25648105e-01 8.77569258e-01
-7.91412532e-01 5.54018915e-01 -1.02588512e-01 -2.63703406e-01
2.63432294e-01 -1.47510961e-01 1.69456184e-01 6.74219489e-01
1.15394317e-01 -4.27073464e-02 3.79788160e-01 5.48632443e-01
7.76493102e-02 -1.52746275e-01 -5.31495214e-01 -5.76155223e-02
7.02740371e-01 1.48595345e+00 -4.42220718e-01 -2.69020647e-01
-4.97468531e-01 6.32116973e-01 5.52150667e-01 5.44798076e-01
-1.00095379e+00 -1.34303913e-01 8.37255776e-01 -4.41354737e-02
-7.28224441e-02 -3.38159323e-01 -1.94769561e-01 -1.12877393e+00
4.91483584e-02 -1.10117483e+00 6.84489191e-01 -9.17526901e-01
-1.24110746e+00 8.97108018e-01 -3.27218592e-01 -1.38534236e+00
-7.87266120e-02 -1.43762276e-01 -1.07151878e+00 4.75785911e-01
-1.53615642e+00 -1.57327318e+00 -4.04409170e-01 9.82087493e-01
-3.92261118e-01 -1.29931539e-01 1.08494782e+00 -6.03759065e-02
-8.54068935e-01 7.80106962e-01 -2.31146768e-01 1.06444411e-01
6.54838383e-01 -1.35085237e+00 4.39328462e-01 1.00790167e+00
2.16090545e-01 8.05048585e-01 7.16844976e-01 -8.97262752e-01
-1.52233171e+00 -6.95033133e-01 7.26685584e-01 -3.40874851e-01
1.19786251e+00 -3.07392508e-01 -1.14373934e+00 4.96702343e-01
2.64194191e-01 5.53841412e-01 7.16355920e-01 4.59230870e-01
-8.92973125e-01 -5.18756092e-01 -1.23906529e+00 7.32636392e-01
7.42559373e-01 -6.99812889e-01 4.30932827e-02 2.06476703e-01
9.46382105e-01 -3.27133119e-01 -1.21513700e+00 4.83947366e-01
6.45066142e-01 -1.49287534e+00 6.41333938e-01 -7.82121480e-01
-2.11588535e-02 -5.15997529e-01 3.40029597e-01 -1.63189900e+00
-4.91095781e-01 -7.25688696e-01 -8.52047384e-01 1.02585363e+00
4.76798147e-01 -9.47928786e-01 5.26885509e-01 9.70690191e-01
2.05813721e-02 -1.10470510e+00 -7.33507216e-01 -6.32997692e-01
-8.46357703e-01 -3.87975484e-01 1.02123618e+00 1.33301723e+00
6.36525273e-01 8.28085691e-02 -7.44731247e-01 5.45104265e-01
5.45930862e-01 1.19244374e-01 6.62529230e-01 -1.42680693e+00
-1.67579070e-01 -3.14900994e-01 -3.52128386e-01 -2.65976042e-01
6.27649724e-02 -3.38347524e-01 -6.38284624e-01 -1.61197829e+00
-3.13880175e-01 -6.92597210e-01 -9.48666632e-01 8.56153429e-01
-1.80378094e-01 -2.50288904e-01 1.60614982e-01 3.03365916e-01
-8.23171973e-01 5.15565753e-01 1.18384504e+00 -1.79052755e-01
2.14320309e-02 4.60894018e-01 -7.49781549e-01 4.12517905e-01
7.96976507e-01 -4.77637529e-01 -7.38767326e-01 -2.26522878e-01
7.03462303e-01 4.12924200e-01 3.62058342e-01 -9.88108337e-01
7.90342152e-01 -2.86664426e-01 2.19248712e-01 -3.95965278e-01
1.03386872e-01 -1.27788365e+00 3.55893642e-01 7.28060782e-01
8.97997841e-02 8.48410070e-01 2.08383612e-02 7.97093928e-01
-3.93558532e-01 3.52745771e-01 1.62779540e-01 2.07489371e-01
-5.37669539e-01 6.72807455e-01 -4.49282862e-02 -5.94375968e-01
1.08688998e+00 -1.34901986e-01 -7.08085537e-01 -7.78375566e-01
-6.19430125e-01 8.70318532e-01 4.60926801e-01 5.86537957e-01
1.00552905e+00 -1.58719397e+00 -4.19029564e-01 3.07578415e-01
1.32383779e-01 -5.14062583e-01 1.23637013e-01 9.57811475e-01
-9.36121047e-02 4.22830246e-02 -2.78408647e-01 -2.74432778e-01
-1.34388983e+00 7.62701988e-01 3.77732038e-01 -6.61808431e-01
-4.57046330e-02 7.06223965e-01 -4.14784312e-01 -3.80901188e-01
4.48506981e-01 -3.49076360e-01 -2.77656704e-01 -4.63881008e-02
3.43486547e-01 3.45333755e-01 3.43601741e-02 -2.01151922e-01
-4.17898923e-01 -4.02775630e-02 -2.69624498e-02 3.05520087e-01
1.24695790e+00 4.16079387e-02 -4.89383399e-01 4.65676725e-01
4.76867229e-01 -1.71486720e-01 -1.16666925e+00 -2.49831334e-01
1.93099976e-01 -4.03471202e-01 2.07506735e-02 -1.24389672e+00
-1.45499504e+00 6.03536546e-01 2.83272803e-01 9.39136744e-01
1.01600683e+00 -6.28671408e-01 4.86409068e-01 1.59704641e-01
6.17468178e-01 -1.07233512e+00 5.35120845e-01 5.57388425e-01
8.37081075e-01 -1.33446443e+00 2.51457900e-01 -2.39537597e-01
-7.82357216e-01 1.15743947e+00 1.20442975e+00 2.53147501e-02
1.08934629e+00 2.37016246e-01 3.25901896e-01 -6.00700080e-01
-1.45495629e+00 3.56391758e-01 2.63466567e-01 5.16912997e-01
2.45982036e-02 2.77207583e-01 -1.51383802e-02 6.81279540e-01
2.10791767e-01 -2.17942134e-01 5.42792141e-01 9.75080967e-01
-1.90435231e-01 -1.18092859e+00 2.33991910e-02 3.40420425e-01
-9.23208222e-02 1.95120205e-03 -4.66470301e-01 8.51006150e-01
-2.71733880e-01 1.20583320e+00 -4.63329971e-01 -1.23162925e+00
5.17482758e-01 -4.40536290e-01 -2.84051478e-01 -2.51659513e-01
-1.18036664e+00 -4.50103253e-01 2.07444653e-01 -9.61050451e-01
-1.81048989e-01 -1.56891450e-01 -1.23343551e+00 -8.78077269e-01
-6.46945894e-01 7.36716270e-01 6.36325181e-01 5.56450605e-01
4.47529465e-01 4.84010398e-01 1.13406909e+00 -2.05580130e-01
-7.32039154e-01 -8.25940430e-01 -8.02006543e-01 -2.29654387e-02
3.11451584e-01 -7.54047930e-01 -5.67588449e-01 -7.68886685e-01] | [7.111410617828369, 6.420343399047852] |
ec7ab246-d86a-4833-a8cb-c2bcf9c0abb3 | multi-image-summarization-textual-summary | 2006.08686 | null | https://arxiv.org/abs/2006.08686v1 | https://arxiv.org/pdf/2006.08686v1.pdf | Multi-Image Summarization: Textual Summary from a Set of Cohesive Images | Multi-sentence summarization is a well studied problem in NLP, while generating image descriptions for a single image is a well studied problem in Computer Vision. However, for applications such as image cluster labeling or web page summarization, summarizing a set of images is also a useful and challenging task. This paper proposes the new task of multi-image summarization, which aims to generate a concise and descriptive textual summary given a coherent set of input images. We propose a model that extends the image-captioning Transformer-based architecture for single image to multi-image. A dense average image feature aggregation network allows the model to focus on a coherent subset of attributes across the input images. We explore various input representations to the Transformer network and empirically show that aggregated image features are superior to individual image embeddings. We additionally show that the performance of the model is further improved by pretraining the model parameters on a single-image captioning task, which appears to be particularly effective in eliminating hallucinations in the output. | ['Sebastian Goodman', 'Kazoo Sone', 'Radu Soricut', 'Nicholas Trieu', 'Pradyumna Narayana'] | 2020-06-15 | null | null | null | null | ['abstractive-sentence-summarization'] | ['natural-language-processing'] | [ 7.55552411e-01 5.29234052e-01 6.93319887e-02 -4.87943500e-01
-1.29343486e+00 -3.81861597e-01 6.67539716e-01 3.73162031e-01
-2.16618150e-01 5.39399922e-01 5.75268149e-01 1.97459459e-01
2.23575369e-01 -3.84381175e-01 -9.74096119e-01 -6.12080097e-01
2.21722543e-01 5.33072650e-01 -2.89919470e-02 6.18367083e-03
1.79992810e-01 2.76233196e-01 -1.61675131e+00 6.60447657e-01
8.73517632e-01 6.81325436e-01 5.67076206e-01 9.63270068e-01
-7.77164921e-02 8.15691054e-01 -9.11656618e-01 -4.18758005e-01
1.74783898e-04 -9.15250838e-01 -8.57603252e-01 7.60875642e-01
1.05239320e+00 -4.57320303e-01 -1.85262024e-01 1.09984875e+00
5.37607253e-01 2.91867882e-01 9.09771025e-01 -1.26326871e+00
-9.16440129e-01 5.19463956e-01 -5.49406707e-01 -3.98111492e-02
3.36254865e-01 6.50158674e-02 1.10678434e+00 -8.34581435e-01
8.24769020e-01 1.40938985e+00 2.53566831e-01 5.43971360e-01
-1.41155756e+00 -1.20099358e-01 2.00425029e-01 2.23114744e-01
-1.11063385e+00 -5.68143785e-01 6.47209406e-01 1.02265831e-02
9.22475696e-01 4.88488078e-01 5.29799998e-01 8.16405118e-01
2.09979221e-01 1.16917360e+00 7.82388628e-01 -3.26278687e-01
4.42289934e-02 3.50536257e-01 -5.43958955e-02 6.46817803e-01
3.62998128e-01 -8.38401437e-01 -3.94222558e-01 8.02597255e-02
5.11623442e-01 1.06371671e-01 -2.68643469e-01 -3.30549717e-01
-1.23689091e+00 9.40615773e-01 5.85201204e-01 2.07226992e-01
-6.05506599e-01 3.28772306e-01 3.50507915e-01 5.21574132e-02
6.56817496e-01 7.99866498e-01 1.01882972e-01 4.22451347e-01
-1.20829940e+00 4.35490549e-01 6.61658108e-01 1.00908995e+00
8.56071949e-01 1.12561695e-01 -5.10122836e-01 1.09617853e+00
-2.82413036e-01 5.54067492e-01 4.22327310e-01 -1.33252048e+00
3.54550391e-01 4.50253755e-01 6.05700351e-02 -9.13355827e-01
-1.64524898e-01 -3.99578243e-01 -1.11754966e+00 -1.09682418e-01
-1.00914657e-01 4.47796732e-02 -1.06025672e+00 1.73106682e+00
-5.56774475e-02 -1.49552628e-01 3.23280632e-01 7.54399836e-01
1.10476339e+00 1.08848965e+00 1.09434985e-01 -5.04592717e-01
1.39217186e+00 -1.42199636e+00 -6.83734000e-01 -5.18131614e-01
2.98484147e-01 -6.38433874e-01 1.01549673e+00 1.64369598e-01
-1.50549436e+00 -4.98986065e-01 -9.65765119e-01 -6.11403644e-01
-1.68270990e-01 1.75680175e-01 3.41983259e-01 4.07730751e-02
-1.41575289e+00 3.98814410e-01 -3.09052140e-01 -7.90951848e-01
6.56750977e-01 2.60738254e-01 -6.46158099e-01 -4.27018106e-01
-6.09689236e-01 1.10022795e+00 5.84407151e-01 -3.10192585e-01
-7.72624731e-01 -4.41877991e-01 -1.06771791e+00 3.77834082e-01
1.79770902e-01 -1.22363567e+00 1.37385774e+00 -1.13389766e+00
-1.07627714e+00 9.09164667e-01 -4.10117835e-01 -6.22579396e-01
2.69999176e-01 -1.01792783e-01 1.33625120e-01 7.63535917e-01
4.18635339e-01 1.48030972e+00 1.00728238e+00 -1.65249097e+00
-5.31415880e-01 -1.85138345e-01 -1.10094607e-01 7.62503684e-01
-5.09583116e-01 -1.79318741e-01 -6.13647461e-01 -4.91127014e-01
-2.91026354e-01 -8.29618096e-01 -5.26589572e-01 -3.44454944e-01
-7.45253563e-01 -2.29057476e-01 7.01959670e-01 -7.14834750e-01
8.88247669e-01 -1.91088045e+00 4.18846101e-01 -2.58383721e-01
1.40931755e-01 7.70753697e-02 -5.48174441e-01 5.57399511e-01
-1.10170521e-01 1.13402508e-01 -4.65657622e-01 -8.78079295e-01
-2.46247530e-01 1.86589375e-01 -4.74027544e-01 1.25911430e-01
4.33956534e-01 1.17891335e+00 -8.78762364e-01 -7.70345032e-01
2.67767668e-01 2.59520292e-01 -5.10022402e-01 3.91074151e-01
-3.68921906e-01 1.86966911e-01 -2.84465224e-01 1.45314232e-01
5.11388838e-01 -4.00976002e-01 -1.72598630e-01 -3.83685678e-01
1.81099355e-01 -2.58349001e-01 -5.99408150e-01 1.87464130e+00
-5.20336926e-01 7.30454028e-01 -1.92046657e-01 -1.05364645e+00
5.24406672e-01 1.85703427e-01 5.06885827e-01 -6.44023716e-01
-8.74327272e-02 -1.25005290e-01 -3.05879861e-01 -5.92291355e-01
1.13974595e+00 -1.26494065e-01 -4.19393569e-01 7.82730579e-01
1.47114873e-01 -7.13309050e-01 4.90448028e-01 6.40216291e-01
8.82581294e-01 -2.15006739e-01 4.23016131e-01 5.88420779e-02
3.42218220e-01 3.44612956e-01 1.02146231e-02 9.18689489e-01
1.62977561e-01 1.28961480e+00 5.49310267e-01 -2.21647397e-01
-1.38208306e+00 -1.01705384e+00 3.16269964e-01 9.37290430e-01
2.05208689e-01 -4.08489048e-01 -1.03268433e+00 -4.58658844e-01
-2.84309506e-01 1.03023529e+00 -5.13819158e-01 -2.75230378e-01
-4.02927637e-01 -7.79393375e-01 2.79439896e-01 2.76328146e-01
6.20721757e-01 -1.28100169e+00 -6.28196418e-01 7.38193691e-02
-7.04219460e-01 -1.28002882e+00 -7.34476328e-01 -2.55134348e-02
-6.81932032e-01 -6.48651183e-01 -1.03333294e+00 -1.11893368e+00
9.94574130e-01 6.92191243e-01 1.14782250e+00 -1.18585430e-01
-4.05984849e-01 7.50210464e-01 -2.12867588e-01 -4.61485535e-01
-7.23471701e-01 2.56104529e-01 -2.43639708e-01 1.22316539e-01
-2.41940290e-01 -3.59991252e-01 -5.06422639e-01 -3.58609259e-01
-1.41713548e+00 6.03074074e-01 1.08797133e+00 7.13637173e-01
6.95950806e-01 -1.78952381e-01 7.66940117e-01 -1.00911403e+00
1.00375974e+00 -3.72348934e-01 1.48962751e-01 4.73139644e-01
-1.51289433e-01 2.13716075e-01 7.53963649e-01 -3.45091730e-01
-1.09165680e+00 3.67343932e-01 -2.08958127e-02 -4.92883503e-01
-1.39889985e-01 3.05698723e-01 -1.13188006e-01 2.62064397e-01
5.20406723e-01 6.36251867e-01 3.42724532e-01 -7.59329796e-02
8.82331073e-01 6.07236087e-01 9.23865020e-01 -1.11962162e-01
5.50675690e-01 5.62672794e-01 -1.21561855e-01 -1.27959073e+00
-1.08315527e+00 -5.36507666e-01 -4.19597089e-01 -1.69418111e-01
1.10451877e+00 -1.00610280e+00 -3.52589846e-01 9.80023891e-02
-1.50067854e+00 3.68225090e-02 -4.91675735e-01 -3.17042843e-02
-1.03381038e+00 5.62702417e-01 -3.23508084e-01 -4.07512963e-01
-7.18318999e-01 -1.11478353e+00 1.41867614e+00 2.99095064e-01
-3.68325621e-01 -7.77822018e-01 -8.77327546e-02 3.82599115e-01
2.36544490e-01 3.08803588e-01 9.34190929e-01 -6.93266094e-01
-7.64209092e-01 -1.74519837e-01 -4.06886220e-01 5.74082673e-01
3.38960662e-02 -3.08624953e-01 -8.37158978e-01 -2.35054031e-01
1.40974134e-01 -6.26653731e-01 1.38138592e+00 5.70365131e-01
1.32626104e+00 -5.83843052e-01 -1.98694497e-01 3.05405289e-01
1.44074237e+00 -2.23842099e-01 7.06550598e-01 -2.11594570e-02
7.82391548e-01 8.58158171e-01 3.44919771e-01 1.21004403e-01
5.64574242e-01 3.37235183e-01 5.44914067e-01 -4.89630759e-01
-4.77585733e-01 -2.90210605e-01 3.34225863e-01 7.05998361e-01
3.24139625e-01 -5.34407794e-01 -3.50043416e-01 7.96673715e-01
-1.97406721e+00 -1.38535202e+00 -5.43245394e-03 1.84882045e+00
7.48095572e-01 -4.23807919e-01 5.72555363e-02 -3.47274363e-01
8.64927649e-01 3.94314438e-01 -6.46229327e-01 -6.30249798e-01
-3.55052471e-01 -1.21376768e-01 3.62268418e-01 3.97931516e-01
-1.02904189e+00 1.08101106e+00 6.39194632e+00 8.61143768e-01
-7.50275254e-01 1.75383575e-02 1.02537835e+00 -2.34189257e-01
-3.94783318e-01 -2.51891017e-01 -4.60706383e-01 1.63743675e-01
7.19515800e-01 -4.38798130e-01 1.79993406e-01 7.43555725e-01
2.40618169e-01 -4.88563001e-01 -1.17304242e+00 1.22381663e+00
9.19977069e-01 -1.56126714e+00 1.02792144e+00 4.51535471e-02
1.11215937e+00 -2.36570314e-01 3.30125540e-01 8.33781213e-02
1.02750123e-01 -1.17930019e+00 6.57193780e-01 3.86008352e-01
8.62946391e-01 -7.75174916e-01 5.49220800e-01 3.02569747e-01
-6.43730462e-01 -1.82862114e-02 -5.02345860e-01 2.69845277e-01
3.59251291e-01 4.31679010e-01 -9.90360618e-01 5.86426735e-01
3.64792615e-01 8.25175643e-01 -9.82682168e-01 1.06099820e+00
-3.36716212e-02 3.50388698e-02 -2.68425513e-02 -3.08126975e-02
3.55883718e-01 -1.69205606e-01 5.68084896e-01 1.28178012e+00
3.82952929e-01 1.41259599e-02 1.98068574e-01 7.18179762e-01
-3.74599606e-01 1.85436621e-01 -9.61827159e-01 -7.60794654e-02
1.30156241e-03 1.55471647e+00 -7.99541116e-01 -7.75860131e-01
-2.32175514e-01 1.52353919e+00 3.36346060e-01 4.09442067e-01
-5.74389935e-01 -3.20966393e-01 2.72850364e-01 -1.20842874e-01
3.96793097e-01 1.00247458e-01 -2.45104998e-01 -1.17436504e+00
9.43592284e-03 -8.09315979e-01 2.29613364e-01 -1.29988730e+00
-9.95095074e-01 7.88703024e-01 2.66660005e-01 -1.14592409e+00
-6.28282309e-01 -1.08690381e-01 -8.05816948e-01 5.54885685e-01
-1.42762184e+00 -1.39351308e+00 -4.52747375e-01 4.02172685e-01
1.11257577e+00 -1.63471457e-02 8.54800582e-01 -1.67035937e-01
-4.22366351e-01 2.08829582e-01 1.65708318e-01 -2.52177268e-01
8.27779233e-01 -1.29386508e+00 2.46557459e-01 9.44622695e-01
2.47581929e-01 3.46278012e-01 1.15022910e+00 -4.63478178e-01
-1.08386576e+00 -1.52074742e+00 9.24966455e-01 -2.12905660e-01
1.42659023e-01 -2.24020541e-01 -7.24882662e-01 7.12687850e-01
1.11010110e+00 -4.36609179e-01 5.33016145e-01 -4.77058500e-01
-8.50828290e-02 -1.03841841e-01 -9.82213616e-01 8.76639247e-01
6.33371830e-01 -2.59984523e-01 -6.65076971e-01 7.00009763e-01
7.92586684e-01 -1.19101450e-01 -4.68649238e-01 -6.31878078e-02
1.66450769e-01 -8.90559256e-01 1.02393758e+00 -3.98276687e-01
1.01340151e+00 -1.99358627e-01 -7.74807576e-03 -1.75472188e+00
-2.35888466e-01 -4.92426187e-01 3.42498690e-01 1.26653659e+00
2.77230501e-01 -2.10602000e-01 5.32953799e-01 3.70907128e-01
-2.93018579e-01 -4.52922672e-01 -5.15938461e-01 -4.05070543e-01
-9.63635668e-02 4.33465168e-02 1.90089837e-01 4.26929235e-01
-1.97773471e-01 9.57063496e-01 -5.35606325e-01 -6.43980205e-02
8.06870282e-01 3.36717635e-01 8.97273064e-01 -8.49772573e-01
3.91010493e-02 -4.07610476e-01 -2.13619158e-01 -9.96336520e-01
3.10724139e-01 -1.11420572e+00 2.91383445e-01 -2.30455709e+00
8.29263747e-01 2.48705983e-01 1.94328517e-01 2.63929039e-01
-3.04156870e-01 5.83186746e-01 5.77888489e-01 1.84586734e-01
-1.14310062e+00 5.93194902e-01 1.38671696e+00 -4.59719688e-01
6.66674459e-03 -1.88226193e-01 -1.19338822e+00 4.06537980e-01
7.00875998e-01 -3.02652836e-01 -6.85193539e-01 -5.40849626e-01
-6.42263368e-02 1.74733967e-01 4.29528028e-01 -9.32860792e-01
3.90760690e-01 1.10334791e-01 5.69465697e-01 -7.74603367e-01
6.15808070e-01 -3.91497374e-01 -9.01362672e-03 2.50703484e-01
-6.87662184e-01 2.19267502e-01 1.35849625e-01 6.47670686e-01
-4.37699288e-01 -3.39489967e-01 8.44160199e-01 -5.69949269e-01
-6.36834979e-01 2.83369660e-01 -2.73279369e-01 9.78422537e-02
1.02994955e+00 -3.60934585e-01 -3.21281433e-01 -9.20178354e-01
-5.81970334e-01 3.89595717e-01 6.49399400e-01 3.93906265e-01
1.01670647e+00 -1.21668172e+00 -1.07302666e+00 -1.51228666e-01
2.47169062e-01 2.34957531e-01 4.58853096e-01 5.57928622e-01
-4.79693234e-01 4.48356330e-01 -3.47484708e-01 -7.54364491e-01
-1.43595660e+00 4.80790019e-01 -4.16060165e-02 -2.44217336e-01
-7.12400377e-01 5.71274877e-01 7.05138266e-01 8.80903155e-02
3.59738469e-02 -2.31062278e-01 -3.52228045e-01 3.40305626e-01
7.02639341e-01 -1.91154439e-04 -1.35525659e-01 -8.51808131e-01
9.54789147e-02 5.63021541e-01 -4.94679481e-01 -2.77180344e-01
1.60902643e+00 -4.51992482e-01 -3.01796526e-01 3.47731709e-01
1.48373938e+00 -5.01494825e-01 -1.11221087e+00 -8.53915662e-02
-2.37141103e-01 -6.36326894e-02 -2.36231521e-01 -5.31482875e-01
-7.48131931e-01 8.02766562e-01 4.10000868e-02 1.66453972e-01
1.20319963e+00 3.37567717e-01 8.21606815e-01 5.12056530e-01
-1.79851219e-01 -9.39534605e-01 6.92235827e-01 2.26548731e-01
1.36302578e+00 -1.32306600e+00 2.20505252e-01 -2.18406722e-01
-1.20721865e+00 9.68511760e-01 4.25343633e-01 -2.90783316e-01
-1.99572727e-01 -2.98301309e-01 -1.67948127e-01 -1.23772658e-01
-9.42169011e-01 -3.11017454e-01 2.71017402e-01 5.62265337e-01
8.71232376e-02 -1.09564051e-01 -6.34429371e-03 1.60974205e-01
-2.06742421e-01 -3.92652780e-01 1.15462482e+00 4.78081137e-01
-6.86969995e-01 -6.50804639e-01 -3.72743189e-01 7.96950519e-01
-4.41818774e-01 -1.75992548e-01 -5.59321046e-01 5.37208796e-01
-2.85010219e-01 8.61001194e-01 3.33071083e-01 1.02221228e-01
2.69503146e-01 6.40110224e-02 5.39614737e-01 -1.01188600e+00
-4.57776457e-01 -3.70367505e-02 2.22275704e-02 -3.16941142e-01
-6.41095579e-01 -6.71589255e-01 -1.15050745e+00 7.60841519e-02
-1.64618399e-02 1.38739526e-01 7.65739441e-01 7.63071716e-01
4.41545606e-01 6.52741075e-01 5.47437608e-01 -1.15947664e+00
-4.03374314e-01 -9.26036596e-01 -5.01645923e-01 7.74709642e-01
4.96349633e-01 -9.23175812e-02 -2.00019076e-01 5.33545494e-01] | [11.047393798828125, 0.7834670543670654] |
15d9574e-d856-43ae-944c-f8f435c4cbcf | two-pass-discourse-segmentation-with-pairing | 1407.8215 | null | http://arxiv.org/abs/1407.8215v1 | http://arxiv.org/pdf/1407.8215v1.pdf | Two-pass Discourse Segmentation with Pairing and Global Features | Previous attempts at RST-style discourse segmentation typically adopt
features centered on a single token to predict whether to insert a boundary
before that token. In contrast, we develop a discourse segmenter utilizing a
set of pairing features, which are centered on a pair of adjacent tokens in the
sentence, by equally taking into account the information from both tokens.
Moreover, we propose a novel set of global features, which encode
characteristics of the segmentation as a whole, once we have an initial
segmentation. We show that both the pairing and global features are useful on
their own, and their combination achieved an $F_1$ of 92.6% of identifying
in-sentence discourse boundaries, which is a 17.8% error-rate reduction over
the state-of-the-art performance, approaching 95% of human performance. In
addition, similar improvement is observed across different classification
frameworks. | ['Graeme Hirst', 'Vanessa Wei Feng'] | 2014-07-30 | null | null | null | null | ['discourse-segmentation'] | ['natural-language-processing'] | [ 4.68869984e-01 5.87770522e-01 -3.64675790e-01 -3.66997898e-01
-1.03011084e+00 -8.15852106e-01 8.59575987e-01 7.56003976e-01
-4.55806285e-01 6.45845652e-01 3.47536922e-01 -2.23885491e-01
2.81528085e-01 -7.10802257e-01 -5.67919493e-01 -4.40470695e-01
3.43053713e-02 3.60754371e-01 6.34290874e-01 -4.09494698e-01
5.42499185e-01 3.77527513e-02 -1.37988842e+00 7.08118916e-01
1.08859205e+00 9.22236621e-01 9.33762565e-02 7.89100468e-01
-4.45378870e-01 7.96882570e-01 -1.07415712e+00 -2.28762373e-01
-1.03596233e-01 -5.71470976e-01 -1.48368251e+00 3.84782106e-02
3.03319663e-01 -7.29036406e-02 -9.85118672e-02 7.84081459e-01
1.16411932e-01 9.57750604e-02 6.08879149e-01 -6.94491982e-01
-3.08017373e-01 1.02478552e+00 -5.17088234e-01 2.99141020e-01
8.31703782e-01 -1.48546472e-01 1.30600333e+00 -6.16091907e-01
8.00527930e-01 9.79742229e-01 5.65159559e-01 3.24707657e-01
-1.11343074e+00 -7.12908134e-02 4.25800830e-01 1.55266702e-01
-9.87424493e-01 -3.37411582e-01 6.41057551e-01 -7.47870028e-01
1.14842200e+00 5.88355005e-01 4.68634784e-01 4.80174661e-01
2.99355220e-02 1.03549075e+00 1.15484214e+00 -7.82286465e-01
3.41105312e-02 -4.98338379e-02 6.08765662e-01 4.09768373e-01
-3.32774937e-01 -4.61978137e-01 -4.91168976e-01 8.27470422e-03
1.52275279e-01 -6.24758661e-01 -2.06007391e-01 2.49616846e-01
-1.21712637e+00 8.27582777e-01 4.54402536e-01 8.02852809e-01
-5.84408902e-02 -3.23314369e-01 6.53544605e-01 1.84242979e-01
5.06638348e-01 7.13984549e-01 -2.17946306e-01 -6.32362604e-01
-1.03608191e+00 4.67602432e-01 8.21032465e-01 7.79418766e-01
6.73758507e-01 -4.55874950e-01 -5.22901773e-01 6.63967073e-01
1.12119429e-02 -1.24800794e-01 3.84460956e-01 -9.84938622e-01
7.30079472e-01 1.01811874e+00 8.76204893e-02 -9.06402349e-01
-5.97171962e-01 -1.45209029e-01 -4.76542562e-01 -7.39396214e-02
8.00961733e-01 -2.21673578e-01 -7.19455302e-01 1.77458620e+00
3.29593688e-01 -1.71347469e-01 -6.98328465e-02 7.01760054e-01
8.02316308e-01 6.47931337e-01 -4.13515531e-02 -3.55930895e-01
1.52018595e+00 -1.12351871e+00 -8.38798761e-01 -2.27512017e-01
8.97321165e-01 -9.65743661e-01 7.87331998e-01 3.23650777e-01
-1.36472154e+00 -6.34865522e-01 -1.18149102e+00 -2.76393801e-01
-3.67319942e-01 5.13172597e-02 2.95626879e-01 7.01045752e-01
-1.05574274e+00 8.03636968e-01 -6.66998804e-01 3.88343073e-02
2.38961298e-02 3.27896386e-01 -5.56124896e-02 3.31421316e-01
-1.29603136e+00 8.89569342e-01 3.75192136e-01 -7.65901059e-02
-2.65190512e-01 -5.04504621e-01 -8.71292830e-01 1.39954567e-01
5.14174044e-01 -2.52229571e-01 1.39187384e+00 -9.38917518e-01
-1.53214264e+00 1.10813904e+00 -6.10347271e-01 -5.13439357e-01
7.77890444e-01 -2.64427155e-01 -2.20412105e-01 2.22098172e-01
2.41491184e-01 6.41248405e-01 4.35837746e-01 -1.08210385e+00
-7.65756786e-01 -1.70292526e-01 4.91845489e-01 3.42064589e-01
-2.02003300e-01 2.48546824e-01 -3.18108648e-01 -5.44273436e-01
2.61479467e-01 -7.55952477e-01 -1.28359020e-01 -4.39148068e-01
-7.33414233e-01 -6.72842562e-01 6.72271729e-01 -8.80605161e-01
1.75405848e+00 -1.90142858e+00 2.73495197e-01 3.92171182e-03
3.64294529e-01 5.58864057e-01 7.45798945e-02 5.09013414e-01
-4.92651127e-02 5.79644024e-01 -4.42511916e-01 -3.86202723e-01
-6.20734207e-02 -2.81885445e-01 -9.09417346e-02 2.79499620e-01
3.95053178e-01 7.49700487e-01 -9.23555017e-01 -4.73205298e-01
4.14846800e-02 -8.40788409e-02 -3.98293674e-01 1.12562202e-01
-3.12247545e-01 3.40722471e-01 -4.20336455e-01 1.60726488e-01
4.09016639e-01 -2.34172344e-01 5.58753252e-01 2.45369822e-01
-3.80427003e-01 9.08716023e-01 -1.01726711e+00 1.60070217e+00
-7.48320147e-02 8.29603016e-01 2.71390062e-02 -1.09725416e+00
9.68760133e-01 5.06442070e-01 4.25259560e-01 -4.61198151e-01
3.73582751e-01 2.33040899e-01 1.93752065e-01 -2.98449695e-01
9.06639278e-01 1.40154317e-01 -4.36286837e-01 5.38718641e-01
-2.30978578e-01 1.24922626e-01 7.42674053e-01 1.37690037e-01
1.16441107e+00 -1.10545985e-01 3.85495573e-01 -3.45835209e-01
8.66602302e-01 1.11779384e-02 5.93022704e-01 7.85283506e-01
-2.93883204e-01 9.35732067e-01 1.03600621e+00 -9.80566218e-02
-9.01058972e-01 -7.35154390e-01 -1.67851269e-01 1.05486917e+00
2.15396419e-01 -7.37861395e-01 -1.30889475e+00 -9.08708215e-01
-1.80391878e-01 6.74340069e-01 -7.75954187e-01 1.88433304e-01
-1.10447931e+00 -4.27511483e-01 4.81463999e-01 6.39232635e-01
4.63707954e-01 -9.13012683e-01 -6.99400187e-01 2.32556716e-01
-4.42630589e-01 -1.04359055e+00 -4.34311509e-01 8.27293023e-02
-5.29271543e-01 -1.18185401e+00 -5.83315969e-01 -8.37123692e-01
3.70807469e-01 6.75490350e-02 1.01564920e+00 3.49338084e-01
1.21246241e-01 -2.09715545e-01 -6.30803645e-01 -3.01279370e-02
-6.65205181e-01 5.86137235e-01 -3.96991760e-01 -2.75186241e-01
2.38553867e-01 -4.96276170e-02 -3.13090414e-01 2.10188001e-01
-5.92976570e-01 2.56362334e-02 -6.52771071e-02 9.23614442e-01
2.04366535e-01 -4.03544605e-01 7.82865524e-01 -1.14727449e+00
6.55053258e-01 -4.31286842e-01 -3.95408757e-02 2.69856066e-01
-2.20471203e-01 -1.79297984e-01 6.48527920e-01 -4.42485437e-02
-1.02830040e+00 -2.95734823e-01 -4.50552732e-01 2.51725793e-01
-2.63772786e-01 6.43876612e-01 -7.05225021e-02 4.68756586e-01
5.02065122e-01 -2.69961357e-01 -3.95057164e-02 -5.13107240e-01
2.75681049e-01 8.51264775e-01 3.86047870e-01 -5.80491126e-01
2.83262700e-01 -4.18996951e-03 -3.90315831e-01 -8.28066885e-01
-8.34663749e-01 -7.89807737e-01 -1.07498622e+00 -1.90961733e-01
1.03819263e+00 -5.77587128e-01 -6.77485645e-01 6.20314240e-01
-1.33605599e+00 -2.77869016e-01 -3.86823297e-01 6.41522631e-02
-5.32123208e-01 5.97854435e-01 -7.55718827e-01 -6.18883550e-01
-3.20205763e-02 -1.27360427e+00 8.85299861e-01 5.30520797e-01
-7.36019194e-01 -1.00239182e+00 -5.93867041e-02 6.05830789e-01
1.45429417e-01 3.55053127e-01 9.11763966e-01 -1.14192379e+00
-2.67502427e-01 -8.68768021e-02 -1.83650404e-01 2.41051152e-01
2.59933591e-01 1.95098341e-01 -9.88090873e-01 -2.63762683e-01
-5.61934114e-02 -1.14276722e-01 8.64366114e-01 2.03762472e-01
8.27013373e-01 -1.14135072e-01 -4.40450847e-01 2.96027884e-02
9.35619891e-01 4.60009843e-01 6.09886765e-01 5.23544073e-01
4.84096229e-01 6.19090080e-01 7.83220351e-01 4.07960981e-01
4.50539678e-01 7.90811181e-01 8.79661366e-02 -4.26441059e-02
-2.19332606e-01 2.43715867e-02 2.38362923e-01 8.37230146e-01
-7.62551799e-02 -4.17389303e-01 -1.24124599e+00 6.82511210e-01
-1.86038244e+00 -8.43832374e-01 -3.25095713e-01 2.00417638e+00
1.15580034e+00 6.06910527e-01 3.97880763e-01 4.71280873e-01
9.46259558e-01 5.69611549e-01 -2.52975643e-01 -7.09535062e-01
-1.24002330e-01 1.99922025e-01 1.75421834e-01 6.66086316e-01
-1.41830838e+00 1.03087950e+00 6.91975403e+00 8.10368478e-01
-1.05144846e+00 -1.52853966e-01 8.30997527e-01 2.85460353e-01
-4.48408991e-01 1.17520064e-01 -8.37730765e-01 4.50407147e-01
1.05293071e+00 -2.95044214e-01 -5.72792329e-02 5.69571257e-01
-4.50919531e-02 -4.50387090e-01 -1.09628761e+00 1.75636902e-01
8.43622014e-02 -1.53377461e+00 -2.89316922e-01 -9.64107960e-02
7.92986095e-01 -3.53072166e-01 -1.50531664e-01 2.16443017e-01
1.89183652e-02 -1.00439095e+00 9.51693356e-01 2.92326540e-01
6.00219190e-01 -6.22850657e-01 7.07755625e-01 5.03499866e-01
-1.16773307e+00 4.63358313e-02 1.59549788e-01 -4.80634987e-01
2.47335538e-01 2.87878305e-01 -1.02033353e+00 8.33985388e-01
3.66367429e-01 4.79161948e-01 -5.76905787e-01 9.12624598e-01
-3.14496517e-01 7.23438323e-01 -8.84210616e-02 -4.23123091e-02
5.41871607e-01 5.99175543e-02 6.60046041e-01 1.50129247e+00
-5.25515229e-02 1.12552427e-01 5.09609878e-01 6.59673750e-01
-6.52069151e-02 1.67602763e-01 -1.70097947e-01 1.58618987e-01
5.49333394e-01 1.09134758e+00 -9.14738595e-01 -5.51886559e-01
-2.36806884e-01 5.66399932e-01 5.09979427e-01 -4.61843833e-02
-7.42248893e-01 -7.03239918e-01 6.24184012e-01 -1.92422848e-02
3.63570392e-01 -1.80156797e-01 -6.39239848e-01 -7.94436097e-01
1.93007931e-01 -7.68734038e-01 1.87948644e-01 -3.49824503e-02
-9.92239594e-01 6.04621947e-01 -3.99157815e-02 -1.17664731e+00
-5.14129996e-01 -4.51305032e-01 -9.79964316e-01 8.32535684e-01
-1.42921448e+00 -8.63787830e-01 -7.10373223e-02 -1.15026645e-01
6.52647018e-01 1.83577836e-01 7.65539110e-01 -7.89111406e-02
-6.86768293e-01 7.81785369e-01 1.87096685e-01 6.23275697e-01
6.32786214e-01 -1.54043841e+00 6.63422763e-01 8.66526008e-01
-8.71724561e-02 6.61850810e-01 6.45129621e-01 -5.44190347e-01
-6.65563166e-01 -6.28036261e-01 1.37025595e+00 -2.68034548e-01
5.45731366e-01 -3.33695740e-01 -1.20480871e+00 6.10327661e-01
6.15473211e-01 -4.64885980e-01 7.02831864e-01 4.97049034e-01
-2.53681391e-01 4.57243741e-01 -9.38462675e-01 4.07569021e-01
7.51832604e-01 -4.60149169e-01 -1.05277395e+00 1.77364662e-01
7.62179911e-01 -8.62834394e-01 -1.19591951e+00 3.85329276e-01
3.22056890e-01 -1.26395977e+00 7.80098319e-01 -5.40603280e-01
6.40119970e-01 -1.32197872e-01 3.54857892e-02 -1.02272904e+00
-1.77571207e-01 -7.88425803e-01 1.99906304e-01 1.56431282e+00
6.45562887e-01 -4.32747185e-01 7.57548511e-01 6.12854719e-01
-6.76611304e-01 -9.93158519e-01 -1.11665869e+00 -6.44306123e-01
5.39668560e-01 -8.46922472e-02 4.67654049e-01 9.65382397e-01
8.54489267e-01 3.96346867e-01 7.29550347e-02 -4.38353181e-01
-1.24237940e-01 2.28159651e-01 4.32920545e-01 -1.09927678e+00
-3.86404842e-02 -7.86619902e-01 -1.28040716e-01 -1.36247075e+00
4.50068504e-01 -9.14358318e-01 2.79379547e-01 -1.49411118e+00
1.67480230e-01 -6.79044127e-01 -2.51356930e-01 3.49032283e-01
-6.47605002e-01 6.83573112e-02 6.45807683e-01 1.48024946e-01
-6.09925985e-01 3.41201872e-01 1.19822371e+00 -1.93618119e-01
-4.52958703e-01 5.26224747e-02 -7.08090544e-01 1.04847717e+00
8.52255046e-01 -2.52527446e-01 8.56744722e-02 -3.99682730e-01
-2.36186936e-01 2.88039863e-01 -1.55692235e-01 -9.13020909e-01
2.58808345e-01 -1.95949316e-01 1.75224051e-01 -6.90651715e-01
4.63445276e-01 1.45081021e-02 -3.75306040e-01 2.90724844e-01
-6.39402628e-01 6.90653473e-02 1.32288650e-01 3.00977707e-01
-5.48422217e-01 -5.68474412e-01 4.97031480e-01 -1.07932523e-01
-5.32259643e-01 -4.62026685e-01 -5.20953953e-01 3.24077189e-01
1.19534385e+00 -3.49200368e-01 -6.41362250e-01 -4.19606231e-02
-7.52599895e-01 2.79001206e-01 5.37069261e-01 4.33568984e-01
2.02943370e-01 -9.18064356e-01 -6.11891389e-01 -6.45982474e-02
-2.36970335e-01 1.62850544e-01 1.29247025e-01 7.78188527e-01
-4.28232849e-01 6.08135998e-01 -5.89274475e-03 -7.50483096e-01
-1.51010752e+00 3.68615799e-02 3.43063753e-03 -8.22694242e-01
-4.70038146e-01 8.86299729e-01 -9.94925722e-02 -2.29212031e-01
-5.25696436e-03 -3.49799961e-01 -5.77291906e-01 7.71376014e-01
4.37518001e-01 4.33358371e-01 1.68930039e-01 -8.91835809e-01
-5.07215381e-01 5.82913995e-01 -3.77456754e-01 -2.25395769e-01
9.79975939e-01 -1.73597395e-01 -1.13090165e-01 8.12116444e-01
1.09428322e+00 1.61026388e-01 -1.18336189e+00 -2.87753046e-01
5.41843355e-01 -4.86699581e-01 -3.65083694e-01 -7.38811076e-01
-5.64490139e-01 9.10382211e-01 6.48084038e-04 7.83351719e-01
8.09194922e-01 9.25627723e-02 8.31424892e-01 8.20935071e-02
-2.03058347e-02 -1.18691301e+00 1.39649183e-01 1.04221201e+00
6.59551024e-01 -1.07136440e+00 1.84245538e-02 -7.66046941e-01
-6.92570984e-01 1.21569574e+00 4.87588108e-01 -1.87394276e-01
2.39686295e-01 1.23159744e-01 7.97261521e-02 1.21763594e-01
-5.49181342e-01 -1.46295249e-01 3.78735811e-01 3.02471250e-01
9.44085717e-01 1.31963313e-01 -6.66048765e-01 5.74876785e-01
-3.89210433e-01 -3.43873590e-01 5.97470820e-01 1.14214480e+00
-7.60726929e-01 -1.46194315e+00 -2.41886869e-01 4.26906377e-01
-5.06158292e-01 1.25908896e-01 -4.70673829e-01 8.69452417e-01
4.01682593e-02 1.27529943e+00 3.31040889e-01 -4.16107863e-01
4.56410676e-01 3.19735408e-01 2.93798923e-01 -9.18043077e-01
-1.29582870e+00 9.32904705e-02 5.31625390e-01 -2.50667125e-01
-5.32272398e-01 -1.11055517e+00 -1.55020761e+00 -2.42851347e-01
-4.66784477e-01 2.32434511e-01 2.89994270e-01 1.22590566e+00
9.58796516e-02 7.63324976e-01 7.14466929e-01 -8.10889423e-01
-4.55903560e-01 -1.17955840e+00 -3.16775233e-01 4.97905910e-01
3.80712062e-01 -6.23669326e-01 -1.36112392e-01 5.42263165e-02] | [10.782076835632324, 9.480657577514648] |
764f5ee0-310f-4d73-af88-5fefe8bfecbb | text-summarization-with-oracle-expectation | 2209.12714 | null | https://arxiv.org/abs/2209.12714v1 | https://arxiv.org/pdf/2209.12714v1.pdf | Text Summarization with Oracle Expectation | Extractive summarization produces summaries by identifying and concatenating the most important sentences in a document. Since most summarization datasets do not come with gold labels indicating whether document sentences are summary-worthy, different labeling algorithms have been proposed to extrapolate oracle extracts for model training. In this work, we identify two flaws with the widely used greedy labeling approach: it delivers suboptimal and deterministic oracles. To alleviate both issues, we propose a simple yet effective labeling algorithm that creates soft, expectation-based sentence labels. We define a new learning objective for extractive summarization which incorporates learning signals from multiple oracle summaries and prove it is equivalent to estimating the oracle expectation for each document sentence. Without any architectural modifications, the proposed labeling scheme achieves superior performance on a variety of summarization benchmarks across domains and languages, in both supervised and zero-shot settings. | ['Mirella Lapata', 'Yumo Xu'] | 2022-09-26 | null | null | null | null | ['extractive-summarization'] | ['natural-language-processing'] | [ 6.13944888e-01 4.65171933e-01 -5.52293897e-01 -5.97232640e-01
-1.61735332e+00 -8.52776945e-01 6.07167482e-01 4.78940278e-01
-1.97965905e-01 9.43194985e-01 7.83003509e-01 -2.80296445e-01
1.06115706e-01 -4.78303462e-01 -6.53329074e-01 -4.46297169e-01
3.14261079e-01 4.91533697e-01 -5.33587970e-02 2.32736170e-01
5.23323178e-01 2.03729291e-02 -1.05167019e+00 4.06331748e-01
1.15986621e+00 6.02588952e-01 -5.52454442e-02 1.18118203e+00
-2.37983480e-01 9.93526995e-01 -1.03935456e+00 -6.99200034e-01
1.01156533e-01 -9.63125229e-01 -9.55681086e-01 3.66833061e-01
6.76259637e-01 -5.09052753e-01 -2.36698240e-01 1.04142034e+00
6.86212242e-01 4.27829958e-02 9.21823800e-01 -8.47495973e-01
-4.99255657e-01 1.21605837e+00 -4.01725948e-01 1.14446506e-01
4.08519059e-01 -9.10767689e-02 1.53611958e+00 -6.27686441e-01
5.50396383e-01 8.92828584e-01 5.88814199e-01 5.13099968e-01
-1.19409752e+00 8.20294097e-02 5.19709066e-02 -1.58280998e-01
-8.94699037e-01 -7.42725730e-01 6.99608028e-01 -1.26910478e-01
1.02511334e+00 8.06443930e-01 2.04633281e-01 1.20466578e+00
1.10468253e-01 1.38279450e+00 7.86322892e-01 -5.59977412e-01
4.80832815e-01 2.28577569e-01 5.87532103e-01 4.91414070e-01
6.15072429e-01 -7.15166807e-01 -4.59514380e-01 -4.13218230e-01
-9.17423889e-02 -2.68704712e-01 -2.21949548e-01 1.07359879e-01
-1.05729020e+00 7.95307815e-01 -1.93113402e-01 6.58490732e-02
-4.40128982e-01 1.65409774e-01 7.77312636e-01 2.11943597e-01
8.54000330e-01 6.95664704e-01 -1.78645983e-01 -1.44383878e-01
-1.47137558e+00 1.83097303e-01 1.18031669e+00 1.10233510e+00
6.26883328e-01 1.19415060e-01 -8.91710043e-01 8.38499725e-01
-5.67015260e-02 4.77372468e-01 5.35745740e-01 -1.14703417e+00
7.20360577e-01 3.94314855e-01 1.46902323e-01 -8.14043522e-01
-2.25249037e-01 -7.31670380e-01 -7.41960645e-01 -5.97552538e-01
-1.10549359e-02 -4.87026453e-01 -6.74004734e-01 1.56962276e+00
-3.76398079e-02 4.39622924e-02 4.45254892e-01 4.38585430e-01
1.01129556e+00 8.20325196e-01 -2.44762525e-01 -5.72564125e-01
9.05072331e-01 -1.34617746e+00 -8.84698510e-01 -4.74590272e-01
6.25781298e-01 -5.94192922e-01 1.08835804e+00 3.46036434e-01
-1.27464497e+00 -1.94348320e-01 -1.34023440e+00 -2.41717294e-01
1.31519496e-01 4.76224571e-01 5.03302753e-01 7.84036398e-01
-9.66355681e-01 6.76060975e-01 -8.19648087e-01 -4.00104463e-01
2.21067280e-01 1.10238843e-01 3.08306068e-02 1.61187708e-01
-6.60818100e-01 6.87775791e-01 5.41850805e-01 -1.57842755e-01
-9.57831860e-01 -1.89066485e-01 -8.06498766e-01 4.37673479e-01
4.42142338e-01 -9.66589749e-01 1.77753699e+00 -6.77801549e-01
-1.54321861e+00 6.75364494e-01 -5.30352592e-01 -8.80399644e-01
5.22297442e-01 -4.40289229e-01 7.62338564e-02 1.67984053e-01
3.58666658e-01 3.13884526e-01 6.90936863e-01 -1.15207481e+00
-5.88626146e-01 -4.02215533e-02 -5.62976636e-02 2.80188859e-01
-5.72624803e-01 -2.77246069e-02 -3.25215429e-01 -6.50693417e-01
1.70738958e-02 -5.01004338e-01 -4.35128719e-01 -8.93845558e-01
-1.16810095e+00 -3.31955314e-01 3.46140534e-01 -7.09146380e-01
1.55423844e+00 -1.66998076e+00 -9.75070894e-03 -2.48793170e-01
1.75671577e-01 1.81123048e-01 -1.61198884e-01 7.83894300e-01
4.66241032e-01 2.50856310e-01 -6.23055995e-01 -9.38276172e-01
3.67059350e-01 1.42482623e-01 -8.09917510e-01 1.58263981e-01
1.53918147e-01 1.02006114e+00 -1.08120608e+00 -6.64640307e-01
-1.77444130e-01 -7.60416538e-02 -5.51256835e-01 3.73856753e-01
-5.00791788e-01 7.97535256e-02 -4.58242625e-01 4.15352434e-01
4.16045666e-01 -2.68964767e-01 2.48811826e-01 1.66506350e-01
5.90808839e-02 7.18626142e-01 -9.12309945e-01 1.76157522e+00
-2.24813223e-01 7.26934016e-01 -3.45706373e-01 -1.15646374e+00
8.48675132e-01 3.54923576e-01 1.45422727e-01 -5.26902154e-02
8.47645625e-02 2.92410374e-01 -5.93054593e-01 -6.36024654e-01
1.04540229e+00 -1.40580460e-01 -4.87486988e-01 9.16893482e-01
4.14596558e-01 -2.90186435e-01 4.63229626e-01 4.94398683e-01
1.41020286e+00 -1.02367982e-01 5.80422461e-01 -3.07297539e-02
4.59146321e-01 4.52764966e-02 5.39065540e-01 1.38197148e+00
1.01727135e-01 9.65083778e-01 9.24036324e-01 9.23124924e-02
-9.27034974e-01 -1.01001132e+00 2.56167799e-01 1.12922406e+00
-1.97695255e-01 -7.21386731e-01 -1.04790103e+00 -8.50571275e-01
-4.01811123e-01 1.24532795e+00 -1.91674173e-01 -3.18014324e-01
-4.30718452e-01 -7.13562131e-01 8.00148487e-01 4.74723935e-01
2.96199143e-01 -9.21368957e-01 -6.35355294e-01 2.55888879e-01
-5.52318275e-01 -1.09029102e+00 -7.96281219e-01 3.08056355e-01
-9.78340030e-01 -5.96914530e-01 -6.71722353e-01 -7.47415185e-01
6.08210504e-01 2.14718416e-01 1.10768712e+00 -2.94939935e-01
1.48116782e-01 5.10365486e-01 -3.75045508e-01 -4.62831736e-01
-8.40674758e-01 7.17610836e-01 -1.33677036e-01 8.11039358e-02
1.60084039e-01 -3.29253942e-01 -2.75099963e-01 -4.26306307e-01
-8.83206904e-01 1.01815991e-01 8.64482939e-01 7.70638645e-01
4.80143934e-01 -2.11458221e-01 1.11478877e+00 -1.33139336e+00
1.21585214e+00 -4.01933223e-01 -1.51994005e-01 6.25319004e-01
-5.72484851e-01 3.96300465e-01 8.53109241e-01 -7.27955922e-02
-1.15953851e+00 -1.29783779e-01 -2.45192364e-01 1.82089716e-01
-1.23867691e-01 8.20198119e-01 -2.41802931e-02 6.61616206e-01
7.53311694e-01 5.74730694e-01 -3.59593004e-01 -4.52210814e-01
5.24924278e-01 1.08991957e+00 8.72772038e-01 -5.18575728e-01
4.37670171e-01 2.72800565e-01 -3.30412537e-01 -8.86310935e-01
-1.58212292e+00 -6.91585779e-01 -4.34097260e-01 1.09252869e-03
5.98556101e-01 -6.57346249e-01 -1.34039178e-01 2.41997495e-01
-1.35195196e+00 -1.15192950e-01 -6.49362028e-01 3.53571355e-01
-8.16862345e-01 8.67896259e-01 -5.56288242e-01 -9.36932504e-01
-1.05415726e+00 -6.83216155e-01 1.19974148e+00 3.09373021e-01
-5.07518589e-01 -8.73676300e-01 2.53761679e-01 3.21853459e-01
2.92862564e-01 1.27830610e-01 7.93051362e-01 -1.18814349e+00
-3.85303169e-01 -5.29070020e-01 3.08023058e-02 8.37207794e-01
1.89305797e-01 -1.24050109e-02 -1.05631971e+00 -2.39797831e-01
3.89021099e-01 -4.20727432e-01 1.17686486e+00 6.14962459e-01
8.50741923e-01 -9.15145278e-01 -6.06024452e-02 3.61647755e-01
1.21429217e+00 -1.42806947e-01 2.72487193e-01 5.78977773e-03
3.56774271e-01 4.91103023e-01 5.01053810e-01 5.50507665e-01
2.66606390e-01 5.12566417e-02 -4.05816473e-02 2.43865728e-01
3.26488689e-02 -5.17691910e-01 6.90042853e-01 1.25453150e+00
2.79217154e-01 -8.00484717e-01 -4.84847039e-01 7.04213440e-01
-2.07042718e+00 -1.00267518e+00 2.75060106e-02 2.24378800e+00
1.09714270e+00 3.34899068e-01 -1.54713821e-02 -2.86902152e-02
5.21141648e-01 4.87230659e-01 -5.87811828e-01 -5.64043343e-01
-3.67065459e-01 1.56621277e-01 4.77874577e-01 5.49332619e-01
-1.15230668e+00 8.81757259e-01 7.02911711e+00 6.99922979e-01
-9.19369161e-01 1.38758183e-01 6.49613798e-01 -2.30892688e-01
-5.43827236e-01 1.54985890e-01 -7.41095483e-01 3.00570041e-01
1.14983654e+00 -7.13685691e-01 2.55520865e-02 8.48550797e-01
2.99244076e-01 -4.68533710e-02 -1.34494996e+00 5.57202995e-01
5.51153600e-01 -1.52757132e+00 1.85868427e-01 -4.48039889e-01
1.20762551e+00 3.88489068e-02 -1.24794900e-01 3.60299259e-01
3.09869677e-01 -7.62449026e-01 7.13130951e-01 4.88520801e-01
6.13019288e-01 -5.96687376e-01 9.22532082e-01 7.09086359e-01
-5.05976558e-01 8.63362774e-02 -4.49256271e-01 -6.21156543e-02
3.34016621e-01 5.26606798e-01 -1.23701966e+00 7.23379672e-01
1.10076619e-02 6.36254966e-01 -7.20205843e-01 1.10846341e+00
-5.89574456e-01 1.13802373e+00 -1.17190801e-01 -2.55917639e-01
3.60583484e-01 -2.72002280e-01 8.45160127e-01 1.67832637e+00
3.04523557e-01 -1.68017700e-01 2.39704177e-01 7.55785108e-01
-4.82627034e-01 1.92693874e-01 -5.51109433e-01 -2.34526455e-01
4.47963417e-01 1.22186792e+00 -8.26798201e-01 -7.73207724e-01
-2.58711457e-01 1.10919166e+00 2.35369593e-01 4.65833664e-01
-6.41001403e-01 -5.94827235e-01 -7.93308690e-02 -2.40151942e-01
1.96158603e-01 -1.09758399e-01 -7.10417747e-01 -1.35358894e+00
2.81615347e-01 -8.03049386e-01 4.08976227e-01 -4.54041183e-01
-1.06437731e+00 4.80912626e-01 -1.99771370e-03 -1.02424896e+00
-5.35520613e-01 -9.37024597e-03 -9.46769118e-01 4.12938923e-01
-1.25780797e+00 -8.59413743e-01 -3.70355062e-02 -2.29535267e-01
1.01442337e+00 -6.21419102e-02 7.71485806e-01 -1.02316178e-01
-7.53780603e-01 6.98188901e-01 4.13572073e-01 -6.68093860e-02
8.22237968e-01 -1.71205485e+00 6.24953508e-01 1.20486295e+00
4.43439871e-01 5.52576184e-01 1.19294548e+00 -4.80706900e-01
-1.19630754e+00 -1.24039960e+00 1.24601793e+00 -5.13669193e-01
3.48036766e-01 -2.78124392e-01 -7.32900620e-01 1.00025344e+00
6.34245098e-01 -8.03581715e-01 8.16358507e-01 3.63155715e-02
-4.32741083e-02 1.86072197e-02 -7.95669019e-01 6.24123275e-01
7.96604156e-01 -4.43068653e-01 -1.02785575e+00 7.29011774e-01
9.05994594e-01 -2.47101247e-01 -4.82742041e-01 3.12603921e-01
1.90640762e-01 -8.03754866e-01 5.30076921e-01 -7.68041253e-01
7.32216895e-01 -3.76820080e-02 -1.77947760e-01 -1.35068893e+00
1.06067054e-01 -9.96640563e-01 -4.08249617e-01 1.56321132e+00
6.30986154e-01 -6.36573851e-01 8.01469564e-01 4.17616069e-01
-6.08001947e-01 -7.67147422e-01 -6.88367963e-01 -7.91164815e-01
-1.30807891e-01 -1.93915784e-01 3.56776625e-01 4.85573262e-01
1.84075788e-01 1.03016257e+00 -5.70584714e-01 -7.34620541e-02
6.84614658e-01 4.06305134e-01 9.97120798e-01 -8.53250086e-01
-3.73328418e-01 -6.36067748e-01 1.33947968e-01 -1.63377726e+00
3.58364135e-01 -1.11892653e+00 3.61025870e-01 -2.08768344e+00
6.78656042e-01 1.20827191e-01 -1.28913060e-01 2.53937155e-01
-4.23504025e-01 -1.73051223e-01 -7.65509158e-02 1.19858243e-01
-1.17778790e+00 6.01665556e-01 7.90056288e-01 -4.57614392e-01
-1.81726068e-01 4.15358573e-01 -1.15497720e+00 6.93809032e-01
7.97656298e-01 -6.24068439e-01 -4.97686386e-01 -4.79893208e-01
1.70917735e-01 1.06550448e-01 -1.11180812e-01 -8.33601475e-01
5.48843443e-01 3.81993018e-02 -1.80547848e-01 -1.08769608e+00
-3.33907157e-02 -5.65529242e-02 -2.04778522e-01 2.74246693e-01
-1.01518595e+00 -2.75264621e-01 -2.09986329e-01 7.61545479e-01
-2.81623632e-01 -9.20282543e-01 4.15394038e-01 -9.05084983e-02
-1.79614797e-01 -1.16697758e-01 -4.25352275e-01 3.85855168e-01
6.80899084e-01 -6.54111579e-02 -2.85521895e-01 -6.63277447e-01
-3.62156868e-01 3.38445187e-01 2.97612935e-01 7.93371424e-02
4.20919418e-01 -8.79777253e-01 -1.19536090e+00 -1.50392026e-01
-4.62116599e-02 1.73020810e-01 7.40193352e-02 7.80162513e-01
-4.93698210e-01 7.44388342e-01 3.53496045e-01 -5.28268278e-01
-1.16002870e+00 2.87879705e-01 -1.49680808e-01 -6.47675693e-01
-5.27473509e-01 7.07423151e-01 -1.63947955e-01 -4.30757940e-01
3.53672653e-01 -5.32012641e-01 -4.30622324e-02 2.06930503e-01
4.49290484e-01 6.49110079e-01 1.57181114e-01 -2.46702164e-01
1.90766186e-01 5.18029593e-02 -2.27788895e-01 -2.47240081e-01
1.26966786e+00 -2.37218603e-01 -2.20733553e-01 8.25114131e-01
1.09840357e+00 2.53668189e-01 -1.01174712e+00 -2.83216685e-01
4.75469887e-01 -3.81040014e-02 -2.57979512e-01 -5.34177363e-01
-3.75147402e-01 6.67571366e-01 -2.60574371e-01 5.53560317e-01
1.01359475e+00 1.15140103e-01 1.01743817e+00 8.04975271e-01
3.19974311e-02 -1.20943785e+00 7.24142119e-02 5.81839681e-01
7.44993627e-01 -1.10292184e+00 3.03908229e-01 -1.24339141e-01
-8.56157362e-01 9.45108533e-01 2.91314095e-01 -1.35297403e-01
-8.31115022e-02 5.30518554e-02 -1.02083527e-01 -1.15642950e-01
-9.51661408e-01 7.32315332e-02 1.97436705e-01 1.76687822e-01
4.23323601e-01 1.59243360e-01 -7.28219092e-01 9.14935470e-01
-4.30816025e-01 -2.43076324e-01 9.83608127e-01 8.47235262e-01
-7.47697055e-01 -8.27266932e-01 -3.03591732e-02 8.97211254e-01
-7.26740301e-01 -1.77804798e-01 -6.67946041e-01 1.78686216e-01
-6.39275670e-01 1.18221104e+00 -1.59400672e-01 -1.51180759e-01
1.41268477e-01 2.99376160e-01 2.56401360e-01 -1.14058399e+00
-5.00385523e-01 3.59855127e-03 4.34724599e-01 5.89092486e-02
-3.44060212e-01 -7.35107422e-01 -1.11418951e+00 -1.93575621e-02
-5.87926567e-01 5.32231569e-01 4.53649968e-01 9.32044387e-01
3.65773171e-01 5.53001165e-01 7.96428740e-01 -6.79233730e-01
-1.41731668e+00 -1.23231554e+00 -4.12656963e-01 2.02418074e-01
4.14309144e-01 2.07531303e-01 -4.71988440e-01 1.66276738e-01] | [12.434913635253906, 9.426795959472656] |
91326238-beb7-433b-91e2-d213077aa383 | overcoming-catastrophic-forgetting-by-xai | 2211.14177 | null | https://arxiv.org/abs/2211.14177v1 | https://arxiv.org/pdf/2211.14177v1.pdf | Overcoming Catastrophic Forgetting by XAI | Explaining the behaviors of deep neural networks, usually considered as black boxes, is critical especially when they are now being adopted over diverse aspects of human life. Taking the advantages of interpretable machine learning (interpretable ML), this work proposes a novel tool called Catastrophic Forgetting Dissector (or CFD) to explain catastrophic forgetting in continual learning settings. We also introduce a new method called Critical Freezing based on the observations of our tool. Experiments on ResNet articulate how catastrophic forgetting happens, particularly showing which components of this famous network are forgetting. Our new continual learning algorithm defeats various recent techniques by a significant margin, proving the capability of the investigation. Critical freezing not only attacks catastrophic forgetting but also exposes explainability. | ['Giang Nguyen'] | 2022-11-25 | null | null | null | null | ['interpretable-machine-learning'] | ['methodology'] | [ 7.03343451e-02 4.77139771e-01 5.92335537e-02 -1.80824891e-01
2.17750296e-01 -4.89624113e-01 4.81711239e-01 -1.10851778e-02
-2.93652087e-01 1.12712920e+00 -1.51010230e-01 -5.48780799e-01
-2.77045012e-01 -5.32308638e-01 -1.08091044e+00 -6.92484796e-01
-1.56075045e-01 4.53560084e-01 2.63831824e-01 -2.60845304e-01
1.12175062e-01 6.54712200e-01 -1.70142424e+00 2.22610220e-01
8.10240328e-01 4.46372390e-01 1.05774932e-01 8.02551866e-01
-6.96656182e-02 1.03210592e+00 -8.08369815e-01 -5.40709794e-01
2.62967069e-02 -2.91311204e-01 -9.41827893e-01 -1.37356430e-01
3.03343117e-01 -1.22009002e-01 -2.67934084e-01 8.22831571e-01
-1.88011646e-01 5.59531245e-03 3.33652735e-01 -1.75833130e+00
-7.52451181e-01 1.10455859e+00 -2.48933434e-01 4.68090296e-01
-1.98366299e-01 2.13578850e-01 6.32757902e-01 -7.19972014e-01
4.08230543e-01 1.22275269e+00 9.37964499e-01 1.00672507e+00
-1.01557386e+00 -8.59495640e-01 3.31985116e-01 4.26447630e-01
-1.23579288e+00 -2.33359843e-01 7.06690967e-01 -2.46646926e-01
1.02686965e+00 4.81897116e-01 9.87077415e-01 1.33099425e+00
9.73185480e-01 6.04004502e-01 8.65333676e-01 -3.35684538e-01
5.26309252e-01 6.87176809e-02 5.92213809e-01 1.01455414e+00
7.21045315e-01 1.55800790e-01 -7.41556823e-01 -3.40971313e-02
5.27601600e-01 4.12846267e-01 -2.76807696e-01 -6.71269093e-03
-8.98431480e-01 5.99998176e-01 2.25995287e-01 3.55617583e-01
5.29380105e-02 6.23254359e-01 1.79330856e-01 5.35774052e-01
5.27262211e-01 3.76586288e-01 -9.12396491e-01 -1.92814693e-02
-6.59642220e-01 -4.87089455e-02 6.84700191e-01 8.70823026e-01
6.71376288e-01 4.63803738e-01 4.40287948e-01 -3.04294918e-02
2.78761759e-02 3.29637319e-01 5.94350398e-01 -5.39434016e-01
-1.92630529e-01 8.52028072e-01 -1.06225029e-01 -6.63358688e-01
-6.29093468e-01 -8.48246813e-01 -1.17702329e+00 3.43144566e-01
4.01383750e-02 1.13388099e-01 -1.00866306e+00 1.85882533e+00
-4.26874980e-02 3.87358159e-01 -1.46576285e-01 5.14941990e-01
2.39424556e-01 3.04578602e-01 3.68070394e-01 -3.08341384e-01
9.14266586e-01 -8.05984974e-01 -7.28646457e-01 -1.50504217e-01
5.54723620e-01 -6.23060949e-02 1.32852268e+00 7.96140373e-01
-8.33876133e-01 -4.71074969e-01 -1.47732747e+00 3.55856866e-02
-5.81978440e-01 -3.37336630e-01 1.00736701e+00 8.42542350e-01
-1.27669871e+00 8.15660357e-01 -1.11858034e+00 -5.45175001e-02
4.70279127e-01 5.48524082e-01 -2.77614206e-01 3.44932526e-01
-1.08882678e+00 1.03346860e+00 5.26804209e-01 -1.17292225e-01
-1.17992520e+00 -9.73456562e-01 -4.43682015e-01 2.81785309e-01
2.90370941e-01 -1.02408457e+00 1.05304337e+00 -1.14726758e+00
-1.09626067e+00 5.00085413e-01 -2.88416505e-01 -1.09253168e+00
8.85338604e-01 -5.81928551e-01 -6.13154650e-01 -1.50001973e-01
-5.11916637e-01 5.56196690e-01 1.32829225e+00 -1.26623631e+00
-3.05932701e-01 -2.15130910e-01 3.26787323e-01 -4.59347069e-01
-6.15700603e-01 -7.57732570e-01 2.76040435e-01 -7.16151893e-01
-3.44736651e-02 -1.03321075e+00 -1.84731528e-01 4.23147529e-02
-6.04328930e-01 -5.85153848e-02 1.23151827e+00 -3.63660097e-01
1.15866983e+00 -2.10592079e+00 1.63179770e-01 -1.22895025e-01
9.85214114e-01 1.66062310e-01 7.47931078e-02 7.55550265e-02
-6.22517347e-01 5.93400359e-01 -4.18594122e-01 -6.95860147e-01
-2.20802009e-01 5.18677831e-01 -8.24208081e-01 5.48150897e-01
1.77458555e-01 9.21164453e-01 -7.43020833e-01 9.44956616e-02
2.06125304e-01 5.94024062e-01 -3.74540538e-01 -2.59386718e-01
-2.97667503e-01 4.00523782e-01 1.60298705e-01 4.50265080e-01
6.92092717e-01 -1.82172149e-01 -8.54895189e-02 1.24701172e-01
2.18515798e-01 -3.74172032e-01 -6.34798050e-01 1.28290987e+00
-4.33550179e-01 8.00277948e-01 -8.91983807e-01 -5.43194294e-01
7.63005197e-01 1.17615886e-01 -1.60528407e-01 -2.54469901e-01
9.10479799e-02 3.43618840e-02 -5.03297038e-02 -1.87720448e-01
4.93445635e-01 -3.70021731e-01 9.09516066e-02 6.22929752e-01
2.31823474e-01 2.70417511e-01 -2.14052722e-01 5.32830000e-01
1.49608624e+00 -1.75488219e-01 4.86046135e-01 -6.25644326e-01
3.91568482e-01 -2.19166249e-01 4.05988127e-01 1.04709852e+00
-1.71133757e-01 4.41165835e-01 5.02361298e-01 -1.21243429e+00
-1.07079053e+00 -1.19814157e+00 2.46379301e-01 6.99383497e-01
1.94213346e-01 -4.17482376e-01 -7.56103516e-01 -1.02116537e+00
-5.72712682e-02 1.16716635e+00 -1.28660262e+00 -8.61026466e-01
-2.61021048e-01 -5.61569333e-01 5.90183139e-01 4.92245108e-01
4.47844654e-01 -9.40467656e-01 -9.46274579e-01 -1.27543703e-01
1.73770949e-01 -8.16455483e-01 -4.10342515e-02 5.74646711e-01
-1.26172292e+00 -1.19022775e+00 -1.96314737e-01 -2.49835119e-01
8.20076346e-01 4.55854923e-01 1.14460719e+00 8.65723670e-01
-3.45995605e-01 4.47951615e-01 -1.72123075e-01 -6.44390285e-01
-7.81620443e-01 4.42456305e-01 6.99081600e-01 -3.00128043e-01
2.42606282e-01 -1.03921306e+00 -3.04705173e-01 1.41742704e-02
-1.26402235e+00 3.53464305e-01 4.99807298e-01 6.14125788e-01
2.78929561e-01 4.87713099e-01 7.51489758e-01 -1.16428280e+00
5.45761883e-01 -5.58076382e-01 -3.67069036e-01 3.91085148e-01
-9.59331930e-01 5.27317822e-01 1.23291540e+00 -7.27631032e-01
-7.82627523e-01 -2.12677449e-01 9.52210054e-02 -7.43760169e-01
-3.62552814e-02 -1.39293697e-04 -5.19381948e-02 -1.25744447e-01
7.83644199e-01 3.13248217e-01 -3.14205468e-01 -4.15152639e-01
2.38050640e-01 -1.01621248e-01 5.84602892e-01 -8.89639109e-02
1.25743544e+00 5.79872966e-01 1.56513318e-01 -6.52560711e-01
-9.42302883e-01 2.98467636e-01 -7.06909657e-01 -5.04234493e-01
2.42225274e-01 -5.09910107e-01 -9.63784158e-01 4.76094097e-01
-1.36152089e+00 -1.92558944e-01 -6.20088935e-01 9.77991596e-02
-3.52257937e-01 2.92710582e-04 -4.89613324e-01 -8.69656801e-01
-4.08404410e-01 -7.06481159e-01 5.90651631e-01 2.03367114e-01
-3.64228040e-01 -1.30663288e+00 -1.06240653e-01 -5.90764843e-02
5.15082181e-01 2.16532990e-01 1.29737723e+00 -7.85095394e-01
-5.24510145e-01 3.29157598e-02 2.63158709e-01 2.96272218e-01
5.36923222e-02 1.28254384e-01 -1.26742518e+00 -3.68852973e-01
2.94054657e-01 8.11678246e-02 1.02143037e+00 -8.71792901e-03
1.37561476e+00 -6.01573706e-01 -1.93094000e-01 5.54052770e-01
1.54323864e+00 3.88727561e-02 6.92780435e-01 4.03539956e-01
5.96136689e-01 2.33835325e-01 1.21502109e-01 5.86425662e-01
-8.04175064e-02 -2.98860073e-02 9.00344670e-01 4.49708709e-03
-1.50798768e-01 -4.25721914e-01 3.63909066e-01 8.98941517e-01
2.69066412e-02 -1.85904682e-01 -9.76306915e-01 2.36442059e-01
-1.69748724e+00 -8.02885056e-01 -2.13664815e-01 1.94398057e+00
6.72070324e-01 6.62697315e-01 -3.38110775e-01 5.86166739e-01
5.65997303e-01 -1.97923630e-01 -1.03051794e+00 -4.97250795e-01
-2.93171048e-01 2.99096555e-01 4.32806224e-01 6.46948338e-01
-6.48539841e-01 9.08838868e-01 6.65956116e+00 7.12601721e-01
-1.06531525e+00 3.48920792e-01 8.17158163e-01 2.91170627e-02
-5.41046441e-01 1.09494060e-01 -7.44542301e-01 3.47415090e-01
1.20896018e+00 -4.41572398e-01 5.39315760e-01 9.58562732e-01
-3.26028392e-02 -9.83021483e-02 -1.21356606e+00 6.51873648e-01
2.15178132e-01 -1.40527260e+00 5.99528134e-01 -3.76757741e-01
5.08666873e-01 -3.51076156e-01 4.65152353e-01 3.48277301e-01
3.09867620e-01 -1.14706755e+00 1.06728649e+00 7.58238554e-01
5.34642279e-01 -1.03643262e+00 6.53372288e-01 6.14510775e-01
-7.54328489e-01 -1.93485260e-01 -5.13869882e-01 -3.24156046e-01
-2.49578148e-01 9.98599827e-01 -9.33968604e-01 2.79273868e-01
7.35229909e-01 6.00112557e-01 -1.15315282e+00 8.63246918e-01
-4.75653410e-01 8.07857215e-01 -7.74863362e-02 1.21294007e-01
-2.21746638e-01 3.62350821e-01 7.10909665e-01 8.31459939e-01
3.77603978e-01 -2.62648374e-01 -6.60287440e-01 9.86339390e-01
-1.35689586e-01 -5.71095228e-01 -6.40226066e-01 2.03753915e-02
3.62792194e-01 1.00233972e+00 -1.11856472e+00 -2.92948008e-01
3.79906893e-01 1.06813335e+00 3.56412441e-01 1.20674737e-01
-1.16866708e+00 1.43478857e-02 4.50212121e-01 1.05715126e-01
-5.91815915e-03 -3.47893745e-01 -5.98701239e-01 -1.21018565e+00
-5.95764816e-02 -5.91982186e-01 1.78842694e-02 -7.89259851e-01
-1.02688956e+00 7.90733397e-01 -6.53501973e-02 -1.08543479e+00
2.83462435e-01 -3.73005658e-01 -5.32652497e-01 1.57584310e-01
-1.20117211e+00 -1.01057231e+00 -4.22575891e-01 6.07541561e-01
8.43319356e-01 -1.97487608e-01 1.00737858e+00 -3.88539284e-02
-5.74340641e-01 6.99853361e-01 -2.21410781e-01 -7.30242372e-01
3.35496157e-01 -1.17717552e+00 7.42179692e-01 9.18740928e-01
2.71797001e-01 9.88707900e-01 1.32959878e+00 -6.52296841e-01
-1.37749493e+00 -1.18773353e+00 7.58204103e-01 -6.09552741e-01
7.24477768e-01 -6.19236887e-01 -1.28185570e+00 9.63128030e-01
-1.05907058e-03 -2.87499607e-01 6.08036339e-01 1.07738994e-01
-5.74055016e-01 -1.85972288e-01 -1.20756173e+00 8.19938064e-01
1.24980617e+00 -2.49433339e-01 -7.71410227e-01 3.64247650e-01
1.40531385e+00 -1.63274109e-02 -1.77405983e-01 2.30877206e-01
4.38178301e-01 -1.18831861e+00 5.68143368e-01 -8.13916504e-01
3.76733214e-01 -8.07910562e-02 2.45063305e-01 -1.33271444e+00
-7.03458339e-02 -8.74315977e-01 -7.10261822e-01 9.62053418e-01
2.68710196e-01 -9.65363860e-01 7.49080062e-01 4.49280977e-01
-1.73842654e-01 -5.99719524e-01 -1.10029256e+00 -1.08995223e+00
4.71620001e-02 -6.29057705e-01 6.63195372e-01 9.45026577e-01
-2.11106107e-01 1.37670666e-01 -4.98451442e-01 2.04802111e-01
7.08575070e-01 -5.12799501e-01 4.03808355e-01 -1.54557717e+00
-4.03219879e-01 -1.93013191e-01 -5.10062099e-01 -3.60676795e-01
3.61017019e-01 -8.21261942e-01 -3.60859215e-01 -7.76859283e-01
3.98264885e-01 -1.02862805e-01 -4.50061888e-01 8.74001324e-01
-7.03800991e-02 1.30768821e-01 1.92032441e-01 3.62892449e-01
-6.93823278e-01 6.72723293e-01 7.97892511e-01 -2.99093455e-01
-7.24096820e-02 1.96018089e-02 -8.32762361e-01 1.07514155e+00
1.03784263e+00 -8.97851527e-01 -6.49918675e-01 -3.05480897e-01
7.09220588e-01 -3.55640322e-01 5.70119143e-01 -1.51265371e+00
2.97221929e-01 1.41328089e-02 3.51768494e-01 -3.45262110e-01
1.00389279e-01 -1.01234114e+00 7.20066249e-01 1.11746371e+00
-4.87651914e-01 4.81854647e-01 6.15776777e-01 7.62426853e-01
3.07317168e-01 -2.06235707e-01 8.31725180e-01 -3.41775082e-02
-6.81444407e-01 1.96313024e-01 -5.15959084e-01 -2.07305521e-01
1.26533341e+00 -1.80989597e-02 -7.44030476e-01 -2.74353027e-01
-9.69741940e-01 -8.13090205e-02 4.77486670e-01 5.76916873e-01
9.55120564e-01 -8.82034123e-01 -1.98437810e-01 3.15356433e-01
-2.14424834e-01 -4.65561509e-01 5.75271428e-01 6.94782555e-01
-6.61881626e-01 3.42851728e-01 -3.19325775e-01 -3.31014246e-01
-1.21091425e+00 6.94549263e-01 4.08253163e-01 -1.70065999e-01
-1.05930674e+00 8.58890414e-01 1.94161147e-01 -3.07144169e-02
1.71249211e-01 -4.38547134e-01 1.49471477e-01 -3.34188372e-01
7.57681251e-01 2.98032105e-01 2.23323628e-01 -2.86157504e-02
-3.36834103e-01 1.22113153e-01 -5.01838684e-01 1.24112569e-01
1.44675350e+00 -8.55679736e-02 -1.06764019e-01 9.96976137e-01
7.15451956e-01 -2.10744023e-01 -1.32037437e+00 2.70901322e-01
-3.67147513e-02 -4.03728597e-02 -3.20317149e-01 -8.92961204e-01
-9.51779604e-01 1.04455185e+00 7.15808451e-01 4.53728229e-01
1.10679173e+00 -2.62482822e-01 6.57692134e-01 7.02231884e-01
8.65601361e-01 -4.98852283e-01 1.46304414e-01 5.91904879e-01
8.58054101e-01 -7.69938946e-01 1.25216516e-02 -7.88122863e-02
-3.47116083e-01 1.34619677e+00 5.95921576e-01 -1.83951318e-01
5.63792586e-01 6.31697834e-01 -3.22372645e-01 3.86966579e-02
-1.26182902e+00 5.09715021e-01 -3.10064018e-01 4.74632055e-01
-1.92371443e-01 -6.99692369e-02 -1.65194571e-01 9.09653127e-01
-5.64084232e-01 1.38998941e-01 1.05882263e+00 7.85409451e-01
-7.11129069e-01 -6.86515987e-01 -1.33852512e-01 2.81024218e-01
-2.17311934e-01 -1.60111472e-01 -5.34519970e-01 1.01847422e+00
2.45230824e-01 7.11788952e-01 -2.49589562e-01 -9.82356429e-01
-8.50227773e-02 2.45357007e-01 3.09403360e-01 -3.26937288e-01
-7.29189217e-01 -8.07099521e-01 -5.37853181e-01 -4.35675681e-01
-1.14645541e-01 -1.23760208e-01 -1.39612675e+00 -9.66419816e-01
-2.61062771e-01 3.82315032e-02 5.19252002e-01 1.23896146e+00
3.43908757e-01 9.07372534e-01 3.88083249e-01 -5.30934274e-01
-4.12895054e-01 -7.16341734e-01 -6.82633281e-01 6.89077899e-02
6.53641224e-01 -7.07440436e-01 -1.06095707e+00 1.47856563e-01] | [9.777450561523438, 3.4199485778808594] |
d81a8dd8-6798-4797-8f14-fb823ad9130b | balanced-filtering-via-non-disclosive-proxies | 2306.15083 | null | https://arxiv.org/abs/2306.15083v2 | https://arxiv.org/pdf/2306.15083v2.pdf | Balanced Filtering via Non-Disclosive Proxies | We study the problem of non-disclosively collecting a sample of data that is balanced with respect to sensitive groups when group membership is unavailable or prohibited from use at collection time. Specifically, our collection mechanism does not reveal significantly more about group membership of any individual sample than can be ascertained from base rates alone. To do this, we adopt a fairness pipeline perspective, in which a learner can use a small set of labeled data to train a proxy function that can later be used for this filtering task. We then associate the range of the proxy function with sampling probabilities; given a new candidate, we classify it using our proxy function, and then select it for our sample with probability proportional to the sampling probability corresponding to its proxy classification. Importantly, we require that the proxy classification itself not reveal significant information about the sensitive group membership of any individual sample (i.e., it should be sufficiently non-disclosive). We show that under modest algorithmic assumptions, we find such a proxy in a sample- and oracle-efficient manner. Finally, we experimentally evaluate our algorithm and analyze generalization properties. | ['Aaron Roth', 'Michael Kearns', 'Emily Diana', 'Siqi Deng'] | 2023-06-26 | null | null | null | null | ['fairness', 'fairness'] | ['computer-vision', 'miscellaneous'] | [ 4.03479666e-01 3.44005018e-01 -8.60528886e-01 -6.67238235e-01
-1.20260990e+00 -1.33551741e+00 2.13435784e-01 6.31638169e-01
-7.56492078e-01 1.09810233e+00 5.67819551e-02 -4.26061630e-01
-8.30889121e-02 -1.05793011e+00 -9.87486839e-01 -7.31384575e-01
-5.56977727e-02 6.11674309e-01 -1.11801215e-01 5.66903532e-01
2.43642226e-01 5.52819431e-01 -1.00338447e+00 3.10266346e-01
6.66042268e-01 9.37510490e-01 -5.75683951e-01 4.76213425e-01
4.73789483e-01 4.21285212e-01 -7.71844625e-01 -6.27478838e-01
7.10426569e-01 -4.12590623e-01 -9.14838850e-01 4.92958799e-02
4.52187508e-01 -8.40846956e-01 -1.55544514e-02 1.33138371e+00
1.90417632e-01 1.21901050e-01 5.01104832e-01 -1.16917527e+00
-3.70647311e-01 9.24846590e-01 -3.28860402e-01 -2.43478012e-03
3.09810460e-01 4.43284661e-02 1.31526399e+00 -1.45030081e-01
7.23880410e-01 8.47696543e-01 2.58396745e-01 6.62630916e-01
-1.56561780e+00 -1.01646376e+00 1.68528762e-02 -5.19106150e-01
-1.12415481e+00 -7.36407876e-01 4.63004470e-01 -1.74711868e-01
-1.57202542e-01 1.07622302e+00 3.08220565e-01 7.37859786e-01
-3.24973971e-01 6.02385044e-01 1.37552655e+00 -1.11452579e-01
6.92108512e-01 6.21476412e-01 3.73182833e-01 4.43422854e-01
7.88244307e-01 1.45474046e-01 -5.29409945e-01 -9.26937282e-01
1.03675425e-01 2.04267412e-01 -4.66575533e-01 -4.87991512e-01
-6.88385963e-01 9.08418238e-01 2.78295010e-01 -1.82893157e-01
-1.06483392e-01 -5.40851662e-03 1.31958842e-01 6.53636515e-01
6.11148357e-01 4.68745351e-01 -3.67906898e-01 2.53892720e-01
-8.76950502e-01 4.91548300e-01 1.00451934e+00 8.37217569e-01
8.32200885e-01 -7.08179533e-01 -4.41653281e-01 2.28101969e-01
-2.37486914e-01 4.57136720e-01 -1.79069508e-02 -1.11887896e+00
6.32486224e-01 2.59900123e-01 7.46390641e-01 -5.72614074e-01
4.38739397e-02 -2.42224857e-01 -4.52453047e-01 -1.91982444e-02
7.96302617e-01 -4.11935449e-01 -3.43105823e-01 2.18181014e+00
5.10158420e-01 -2.71728784e-01 -7.34417588e-02 9.19257104e-01
2.70127179e-03 5.60847759e-01 -3.49811986e-02 -6.56478941e-01
1.34959269e+00 -1.68690950e-01 -3.03037822e-01 -2.31113266e-02
6.98238373e-01 -1.41008914e-01 8.97419035e-01 2.56627440e-01
-1.12503886e+00 2.70376712e-01 -1.04711819e+00 1.86873555e-01
-4.97835167e-02 -3.30886990e-01 6.66002929e-01 1.18203402e+00
-5.31320691e-01 5.99234343e-01 -5.36981165e-01 3.14062312e-02
8.17984879e-01 4.94950742e-01 -5.20352483e-01 -1.40520602e-01
-9.38246071e-01 8.56259391e-02 2.72185236e-01 -5.37002087e-01
-8.88649523e-01 -9.35921073e-01 -4.81738538e-01 2.93600649e-01
7.29299307e-01 -4.87007648e-01 1.20954204e+00 -8.98556650e-01
-9.75026488e-01 9.58833992e-01 -2.90671557e-01 -6.40309632e-01
9.90800798e-01 1.57422513e-01 1.45005640e-02 2.79845387e-01
2.84723073e-01 4.16879021e-02 6.79802001e-01 -1.29433167e+00
-8.94473970e-01 -6.87911808e-01 3.04722100e-01 1.52869508e-01
-6.27843201e-01 7.73848966e-02 9.85227376e-02 -1.77429542e-01
1.62658840e-01 -9.30727422e-01 -1.97258189e-01 1.92892060e-01
-6.80925071e-01 -5.47288742e-04 6.01109922e-01 -3.15576553e-01
1.07003355e+00 -2.27051854e+00 -6.45313919e-01 7.77500331e-01
5.40047824e-01 1.03835143e-01 1.55366272e-01 1.61146864e-01
2.81920880e-01 6.44365013e-01 -4.90260482e-01 -5.72583258e-01
1.16609940e-02 -1.62822574e-01 -6.78527832e-01 9.46896911e-01
-3.66311073e-01 4.33777124e-01 -8.96462023e-01 -4.70628776e-02
-3.68953556e-01 -3.79354209e-01 -9.40730155e-01 3.70355636e-01
-2.74423063e-01 1.13473885e-01 -7.31537282e-01 4.68775898e-01
1.06230450e+00 -1.25230923e-01 5.42270184e-01 1.77196026e-01
1.60737842e-01 4.86329257e-01 -1.19583821e+00 7.75579751e-01
-7.88455270e-03 2.48064309e-01 4.22487617e-01 -8.16853046e-01
5.87968171e-01 2.19061166e-01 2.61782467e-01 -3.39778662e-02
-1.95564963e-02 2.98272073e-01 -2.38019571e-01 -1.13233089e-01
4.94906753e-01 -4.20684874e-01 -2.74089187e-01 1.14419496e+00
-3.57253283e-01 4.14293736e-01 -3.75371724e-01 7.23421499e-02
1.16429818e+00 -7.26523638e-01 1.36920616e-01 -3.28987688e-01
2.23674789e-01 -2.91910917e-01 7.79684782e-01 1.24153101e+00
-2.97968030e-01 4.43912446e-01 8.81250739e-01 -1.04405798e-01
-9.21335340e-01 -1.08599782e+00 -1.68473154e-01 1.02011788e+00
6.21191002e-02 -1.36687189e-01 -7.80653238e-01 -1.15601170e+00
4.74709004e-01 5.94017386e-01 -6.76863909e-01 -1.58305690e-01
-4.03372273e-02 -6.96604311e-01 2.50288785e-01 -1.46064445e-01
5.78423679e-01 -5.56445897e-01 -4.87213284e-01 -9.55254361e-02
5.39930686e-02 -6.31130576e-01 -1.02583122e+00 2.57152587e-01
-7.58963883e-01 -1.21919727e+00 -5.09379208e-02 -3.95131588e-01
1.09737384e+00 2.58450031e-01 5.71665645e-01 9.71885622e-02
3.44550908e-01 1.40086621e-01 1.13223910e-01 -3.59812617e-01
-4.08064425e-01 2.01932535e-01 -1.50204867e-01 4.66074675e-01
5.09502411e-01 -4.59143788e-01 -6.62482202e-01 -2.71801837e-02
-9.25608933e-01 -6.30162239e-01 -2.33926363e-02 5.14613509e-01
4.43073452e-01 1.51360616e-01 7.07814097e-01 -1.81555557e+00
8.96782219e-01 -6.66641593e-01 -1.08220565e+00 2.99449295e-01
-5.37290752e-01 -1.10253364e-01 8.23566914e-01 -4.60531265e-01
-7.12660015e-01 -3.92258428e-02 5.40809453e-01 -9.88039225e-02
1.66137516e-01 3.14030260e-01 -5.67750275e-01 -1.43614769e-01
6.24081016e-01 -1.92580856e-02 1.58808917e-01 -3.77232373e-01
2.74377823e-01 9.01853323e-01 5.80762446e-01 -9.39948976e-01
8.26054096e-01 6.13844693e-01 -2.08298992e-02 -1.82995796e-01
-1.02725685e+00 -2.39318758e-01 -1.58615395e-01 2.56875992e-01
2.01232120e-01 -6.71925008e-01 -1.31448078e+00 1.46283969e-01
-7.07283080e-01 -1.80417374e-01 -4.89582777e-01 3.61929566e-01
-1.68819353e-01 2.38105014e-01 -2.39000618e-01 -1.31846368e+00
-9.99619663e-02 -7.68542051e-01 5.86574137e-01 1.42948572e-02
-2.60969490e-01 -7.32966840e-01 -3.57785672e-01 5.15705347e-01
1.06992699e-01 4.10162121e-01 8.18371713e-01 -1.31884050e+00
-9.84800220e-01 -8.83989871e-01 -8.58846772e-03 3.11903179e-01
2.09042996e-01 -3.42691332e-01 -1.08485091e+00 -6.84860408e-01
3.55074227e-01 -5.88470340e-01 6.93689883e-01 1.47417784e-01
1.77957439e+00 -1.18553460e+00 -3.29545408e-01 6.50372326e-01
1.15801883e+00 6.78850785e-02 5.78601211e-02 1.05225056e-01
2.83420056e-01 5.42273045e-01 4.43860769e-01 7.81453252e-01
2.33318686e-01 3.17738086e-01 7.94315934e-02 3.28171492e-01
7.18090832e-01 -6.10161543e-01 1.32567316e-01 -2.02899054e-01
4.96716380e-01 -1.42608494e-01 -2.58777052e-01 4.43558544e-01
-1.46779180e+00 -8.81116986e-01 5.17622948e-01 3.23409414e+00
1.45263255e+00 2.21840531e-01 3.42760384e-01 -3.52130942e-02
8.09240282e-01 1.09012291e-01 -8.01388144e-01 -3.81455511e-01
-1.05178609e-01 2.12578267e-01 1.27440345e+00 8.02186489e-01
-9.99846101e-01 4.39701825e-01 6.09360647e+00 5.74320257e-01
-9.35156286e-01 1.40415594e-01 1.31566417e+00 -4.13572729e-01
-9.50531960e-01 3.66806865e-01 -6.46191657e-01 7.31746316e-01
1.02624846e+00 -7.82074332e-01 6.67725444e-01 7.31343389e-01
-1.16106451e-01 -2.06493139e-01 -1.68161416e+00 4.75073904e-01
-2.67105699e-01 -1.35150087e+00 -1.61087364e-01 5.66512585e-01
7.90379584e-01 -5.31239986e-01 2.32755601e-01 1.23463556e-01
5.92380285e-01 -9.01195467e-01 8.07078063e-01 2.65126407e-01
7.55597532e-01 -1.01039696e+00 4.62743431e-01 6.77684546e-01
-4.16358441e-01 -4.48276460e-01 -4.88386393e-01 4.52939980e-02
-3.90011013e-01 8.87767076e-01 -8.23608339e-01 2.01742247e-01
2.58368731e-01 -1.70452341e-01 -1.36344105e-01 1.00088394e+00
2.15520263e-02 9.42837477e-01 -7.16932476e-01 -1.06731452e-01
-1.26080722e-01 -3.17751408e-01 5.47478557e-01 6.77265704e-01
2.52087831e-01 4.05222416e-01 3.63920450e-01 1.00047195e+00
-1.05905652e+00 1.45686176e-02 -5.45572937e-01 -1.00992387e-03
1.35912883e+00 9.10430968e-01 -2.00332880e-01 -4.57125008e-01
2.04496235e-02 6.01312399e-01 3.26117992e-01 3.49188238e-01
-4.15155202e-01 -2.32879117e-01 7.35594213e-01 2.09390894e-01
5.03869131e-02 4.32170510e-01 -3.86984736e-01 -1.03366530e+00
4.58732098e-01 -7.67164350e-01 8.41921449e-01 -1.60668463e-01
-1.38689935e+00 -9.32906494e-02 -3.61785898e-03 -8.04589570e-01
-7.65725896e-02 -3.72679755e-02 -5.10251582e-01 1.11965311e+00
-1.19759989e+00 -5.53859472e-01 2.79729843e-01 4.95061368e-01
-4.70833570e-01 1.03611529e-01 8.28247488e-01 -5.72340786e-02
-3.79417151e-01 1.23752093e+00 2.23337919e-01 7.14510977e-02
6.33361816e-01 -1.26480556e+00 -5.57894073e-02 8.47680807e-01
-2.38947272e-01 8.80307257e-01 4.67376918e-01 -6.31483793e-01
-1.34805179e+00 -1.33392191e+00 1.07166171e+00 -6.64163291e-01
3.67352933e-01 -7.37010598e-01 -8.83969128e-01 9.31513906e-01
-4.85332638e-01 3.92694950e-01 9.08502221e-01 2.95665830e-01
-7.10993171e-01 -6.06889546e-01 -1.93462467e+00 4.20594931e-01
7.81011939e-01 -8.71805072e-01 -1.83755934e-01 4.33894932e-01
7.74583042e-01 -3.57990026e-01 -7.31018364e-01 3.00706979e-02
4.14573699e-01 -7.66222298e-01 4.74680662e-01 -8.92469466e-01
-1.33522451e-01 -1.54215232e-01 -1.84492052e-01 -9.44545388e-01
-1.17855541e-01 -1.11786437e+00 -7.33573735e-03 1.42855990e+00
5.92376173e-01 -1.09432399e+00 1.21274400e+00 1.43438745e+00
6.05505347e-01 -4.38421011e-01 -1.19248855e+00 -6.12863243e-01
2.65261650e-01 5.30551560e-02 1.10930836e+00 1.09471178e+00
-4.24451120e-02 -1.85620576e-01 -4.59150285e-01 5.32373846e-01
9.35607314e-01 4.92506444e-01 7.30636418e-01 -1.13199532e+00
-1.96752682e-01 -6.81684446e-03 1.42670736e-01 -6.65496409e-01
2.24971071e-01 -8.79631758e-01 -1.90152109e-01 -6.08283699e-01
5.40729761e-01 -8.63345444e-01 -3.09610069e-01 5.63550174e-01
-4.07553464e-01 -3.34587283e-02 2.32216433e-01 3.19920927e-01
-3.84580106e-01 1.75115764e-02 8.34210634e-01 -2.62500327e-02
-2.27782533e-01 5.67269742e-01 -1.32995605e+00 -1.94436070e-02
8.92379284e-01 -7.76090503e-01 -4.09382939e-01 3.00021410e-01
1.17334045e-01 2.92246044e-01 3.45014036e-01 -5.36761224e-01
2.53763020e-01 -4.66296554e-01 2.34819010e-01 -8.02235156e-02
-1.41758978e-01 -8.38855147e-01 1.80459067e-01 5.12173772e-01
-1.05098641e+00 -5.15155852e-01 -3.92692775e-01 5.11161506e-01
3.72530729e-01 -3.89688343e-01 8.31456125e-01 7.86508396e-02
4.33274686e-01 6.41597807e-01 -1.48298889e-01 2.57782340e-01
8.86386216e-01 3.04660760e-02 -5.79759538e-01 -5.06355941e-01
-2.83948481e-01 3.77184540e-01 9.16304171e-01 -3.05913806e-01
2.11759329e-01 -1.14702141e+00 -6.00888193e-01 3.86380315e-01
2.51232564e-01 6.26445040e-02 -2.38621652e-01 4.02123302e-01
1.38979837e-01 1.62338674e-01 1.71113104e-01 -6.59772307e-02
-1.18579960e+00 6.87978208e-01 3.28473687e-01 2.07536593e-02
-1.62135929e-01 7.14325070e-01 2.08196953e-01 -3.65330011e-01
3.96334827e-01 -2.23687023e-01 4.08669233e-01 1.29092159e-02
6.76951408e-01 1.19813927e-01 1.07554970e-02 5.02093993e-02
-2.14748263e-01 -2.92464048e-01 -2.86360085e-01 -4.54339832e-01
1.19113231e+00 -6.64837360e-02 -2.88261116e-01 2.87390709e-01
1.54306436e+00 5.90516627e-01 -1.40567875e+00 -4.22392517e-01
-1.21250719e-01 -1.11951315e+00 -2.33988732e-01 -7.85537899e-01
-1.27293432e+00 2.74904400e-01 2.48450398e-01 5.83803594e-01
9.52468455e-01 -1.03251107e-01 4.61664706e-01 4.63589430e-01
4.61850375e-01 -8.82485211e-01 -6.69740081e-01 -4.16342676e-01
3.61427486e-01 -1.02018702e+00 2.02654019e-01 -3.34607571e-01
-2.49515459e-01 5.74824214e-01 2.29959935e-01 2.33916454e-02
4.90147799e-01 -8.67307559e-02 -3.62838805e-01 1.14071131e-01
-8.94449115e-01 4.77908164e-01 -2.03876086e-02 5.22787094e-01
-3.97274196e-02 5.51329732e-01 -3.32985163e-01 7.63945699e-01
-4.15355951e-01 -5.36176609e-03 8.02900434e-01 8.83290231e-01
-5.06390870e-01 -1.15581989e+00 -3.57205421e-01 8.51678610e-01
-9.32061017e-01 2.28832290e-02 -5.50230622e-01 3.49941969e-01
-1.37891904e-01 8.09736907e-01 4.68370095e-02 -6.23875186e-02
4.93889116e-02 5.30628003e-02 1.26431257e-01 -6.35897279e-01
-6.42092466e-01 -3.59445214e-01 2.85218149e-01 -4.23037499e-01
-3.85759324e-02 -8.94191384e-01 -7.15618312e-01 -9.00592089e-01
-2.20171347e-01 6.83406234e-01 2.44674325e-01 7.47208476e-01
1.49778828e-01 -5.41025698e-01 1.44461620e+00 9.66728330e-02
-1.15193200e+00 -3.66868019e-01 -8.68093491e-01 4.95084107e-01
8.06593716e-01 9.09432471e-02 -7.22482800e-01 -2.02566743e-01] | [5.990772247314453, 6.8665337562561035] |
7d4b611f-31b4-4b93-b706-462182904b88 | building-a-knowledge-based-dialogue-system | null | null | https://aclanthology.org/2022.sigdial-1.25 | https://aclanthology.org/2022.sigdial-1.25.pdf | Building a Knowledge-Based Dialogue System with Text Infilling | In recent years, generation-based dialogue systems using state-of-the-art (SoTA) transformer-based models have demonstrated impressive performance in simulating human-like conversations. To improve the coherence and knowledge utilization capabilities of dialogue systems, knowledge-based dialogue systems integrate retrieved graph knowledge into transformer-based models. However, knowledge-based dialog systems sometimes generate responses without using the retrieved knowledge.In this work, we propose a method in which the knowledge-based dialogue system can constantly utilize the retrieved knowledge using text infilling . Text infilling is the task of predicting missing spans of a sentence or paragraph. We utilize this text infilling to enable dialog systems to fill incomplete responses with the retrieved knowledge. Our proposed dialogue system has been proven to generate significantly more correct responses than baseline dialogue systems. | ['Yasuo Ariki', 'Tetsuya Takiguchi', 'Qiang Xue'] | null | null | null | null | sigdial-acl-2022-9 | ['text-infilling'] | ['natural-language-processing'] | [-1.82565693e-02 7.53884494e-01 -2.56052073e-02 -1.67199001e-01
-6.48750186e-01 -5.64285338e-01 8.23661864e-01 4.95542623e-02
-5.16611785e-02 1.27299523e+00 7.74438143e-01 -5.24970233e-01
2.18290105e-01 -1.19887578e+00 6.51510106e-03 6.65567145e-02
5.02547085e-01 8.94063890e-01 3.99243802e-01 -9.90728915e-01
5.07390618e-01 1.08690504e-02 -9.38585639e-01 7.66144753e-01
1.17672157e+00 2.63353437e-01 3.19994628e-01 9.80197310e-01
-9.61206198e-01 1.28234792e+00 -1.01454020e+00 -5.50022125e-01
-3.24643135e-01 -8.91728401e-01 -1.54712629e+00 4.09669504e-02
-3.02375178e-03 -5.15054286e-01 -6.26687169e-01 4.83218938e-01
4.83883739e-01 3.21933359e-01 5.22533178e-01 -1.00789201e+00
-6.46722615e-01 1.11353028e+00 1.71155959e-01 -3.21049280e-02
1.12596953e+00 2.78663248e-01 8.33092630e-01 -1.01533103e+00
7.69494116e-01 1.69478822e+00 4.85835254e-01 9.25463021e-01
-9.35604811e-01 -2.36404225e-01 3.57542969e-02 2.92351693e-01
-9.44280803e-01 -4.97388512e-01 7.63956368e-01 -1.86881855e-01
1.66623664e+00 2.88770914e-01 8.83290946e-01 9.38012719e-01
2.48196006e-01 8.02942216e-01 1.03269720e+00 -7.40365148e-01
-3.43667381e-02 3.76037478e-01 1.42683208e-01 1.18063605e+00
-4.76406217e-01 -3.21072459e-01 -1.05633283e+00 -3.20514977e-01
7.33651578e-01 -5.42450070e-01 -2.30769470e-01 2.06406847e-01
-1.02434075e+00 1.13837612e+00 1.02700524e-01 2.53218949e-01
-3.05675328e-01 -3.57622147e-01 4.90070224e-01 6.32719219e-01
5.45425177e-01 7.71238983e-01 -2.65190303e-01 -4.84412402e-01
-5.99811435e-01 3.15245181e-01 1.71586692e+00 1.18134475e+00
5.76574206e-01 1.91252708e-01 -7.43351519e-01 1.17274010e+00
2.65496731e-01 3.12825590e-01 5.78157902e-01 -8.95591378e-01
7.35772371e-01 1.12337196e+00 2.07384169e-01 -7.83815801e-01
-3.26968312e-01 2.01722473e-01 -7.14037597e-01 -3.83278668e-01
5.04866362e-01 -4.65155363e-01 -4.11038309e-01 1.48227119e+00
3.01470846e-01 -3.74283463e-01 5.51780343e-01 5.81079900e-01
9.99860466e-01 7.23215282e-01 4.61889058e-02 -4.12486404e-01
1.19198632e+00 -1.07275510e+00 -1.05200791e+00 -3.98458447e-03
8.40561092e-01 -8.61391604e-01 1.11573482e+00 1.85673654e-01
-1.21798944e+00 -4.73618746e-01 -6.92643464e-01 -8.67059827e-02
-2.61051476e-01 -7.94119164e-02 4.93187279e-01 6.89924657e-01
-1.27120364e+00 3.30306113e-01 -1.92939609e-01 -5.47020078e-01
-2.60425389e-01 1.15039244e-01 -3.88333411e-03 1.14199050e-01
-1.61463082e+00 1.44287920e+00 1.94227397e-01 2.68875211e-02
-6.56785190e-01 -4.72499013e-01 -9.31777477e-01 6.71273097e-02
4.86435592e-01 -1.09909904e+00 1.86312544e+00 -1.31541491e-01
-2.20092964e+00 2.72734523e-01 -3.49348962e-01 -4.71998602e-01
4.98741060e-01 -1.47629544e-01 -7.92220011e-02 1.25415161e-01
-1.85536563e-01 6.01836264e-01 4.17261481e-01 -9.68081474e-01
-3.38181436e-01 1.71802863e-02 3.68307918e-01 6.66010082e-01
-3.40805382e-01 -3.78764868e-01 -1.75870329e-01 -2.73796976e-01
-1.60277963e-01 -6.90696478e-01 -1.03689536e-01 -5.18560588e-01
-7.14823306e-01 -7.26692080e-01 8.46460640e-01 -5.33651114e-01
1.47844827e+00 -1.08165288e+00 8.10176507e-02 -7.69066811e-02
1.62921667e-01 4.47234601e-01 -2.64618635e-01 1.35259175e+00
7.54076481e-01 4.24556732e-02 2.07941279e-01 -1.64981991e-01
2.20798299e-01 3.05888623e-01 -5.57682633e-01 -4.90346879e-01
7.51194209e-02 1.10942328e+00 -1.07847047e+00 -7.92715132e-01
5.44730425e-01 -3.87270302e-02 -5.03536105e-01 6.49259925e-01
-7.61634350e-01 4.76286322e-01 -6.58802390e-01 2.10590959e-01
1.55843124e-01 -1.50744200e-01 6.78805888e-01 -3.74492258e-02
6.02395982e-02 8.31290781e-01 -4.94605392e-01 1.81148267e+00
-8.77788544e-01 5.36645472e-01 -1.35515392e-01 -4.17494357e-01
1.12136519e+00 6.02187872e-01 -7.76763931e-02 -6.76897705e-01
-1.88867897e-01 -1.74029335e-01 9.17850994e-03 -6.60785735e-01
1.23724461e+00 3.59622724e-02 -1.37979209e-01 9.61065054e-01
1.82472289e-01 -8.39822233e-01 4.02695924e-01 7.98256814e-01
1.19223320e+00 -3.13428044e-02 3.94172430e-01 9.23890397e-02
7.61233687e-01 4.54558760e-01 -3.20524499e-02 9.80815530e-01
1.45660296e-01 -1.92917347e-01 4.65901345e-01 -1.41488135e-01
-9.51990545e-01 -7.37393141e-01 6.13282561e-01 1.21982312e+00
-2.50483572e-01 -6.50791883e-01 -9.41802144e-01 -7.73007393e-01
-1.58703014e-01 1.10275328e+00 -3.75789776e-02 -1.88669831e-01
-6.96749568e-01 -7.35528395e-02 1.04070520e+00 4.49816257e-01
8.41835558e-01 -1.43675315e+00 -2.57538170e-01 7.15171874e-01
-7.62929678e-01 -9.40787554e-01 -5.71991563e-01 -4.79105264e-01
-8.44992697e-01 -1.17779660e+00 -5.16428292e-01 -6.56230927e-01
5.22606432e-01 2.71841198e-01 1.34340048e+00 3.35597605e-01
1.69617325e-01 6.97379589e-01 -6.09867454e-01 -8.66609067e-02
-1.21077180e+00 2.69803822e-01 -1.42335117e-01 -6.52762294e-01
2.00249687e-01 -1.82141960e-01 -3.90804440e-01 5.55376470e-01
-4.96330500e-01 5.68865716e-01 1.40962288e-01 1.20169187e+00
-2.06553638e-01 -3.28055173e-01 1.22894514e+00 -1.00893366e+00
1.84826398e+00 -2.04549536e-01 -2.34226301e-01 7.47287631e-01
-5.68244278e-01 3.27996969e-01 8.32493007e-01 -2.52281010e-01
-1.62396693e+00 -4.07569557e-01 -2.32949674e-01 1.20335244e-01
1.39125273e-01 7.66823053e-01 3.18550199e-01 2.76160035e-02
8.21418703e-01 4.67570394e-01 -1.56530924e-02 -1.54061005e-01
7.46801078e-01 9.01047707e-01 2.49515042e-01 -9.39730346e-01
2.22162589e-01 -4.17681783e-01 -3.54972154e-01 -8.06100965e-01
-7.02168167e-01 -4.42069292e-01 -5.38424432e-01 -5.52780390e-01
4.85483736e-01 -5.44967949e-01 -1.01722324e+00 3.46844703e-01
-1.63028657e+00 -6.86727941e-01 -2.80918151e-01 1.43392473e-01
-5.26045501e-01 5.89286327e-01 -1.00520253e+00 -1.27004421e+00
-7.27000535e-01 -8.09411466e-01 6.80907071e-01 3.54115635e-01
-6.45701349e-01 -1.25363362e+00 2.48582169e-01 6.08220100e-01
5.80170333e-01 -4.68453646e-01 1.13539362e+00 -8.90676379e-01
-4.40651089e-01 -5.60161052e-03 -8.41707587e-02 -6.59189671e-02
3.26787829e-01 -4.38776910e-01 -7.75921643e-01 5.31095490e-02
-8.54051635e-02 -7.57514715e-01 3.67777765e-01 -2.62095004e-01
6.11697197e-01 -8.25569391e-01 -1.01634279e-01 -4.08378124e-01
6.02176726e-01 2.87669271e-01 5.12504041e-01 -2.49747336e-01
3.63398731e-01 7.74276257e-01 6.85335279e-01 6.35015368e-01
8.12982798e-01 6.97701335e-01 -1.46449208e-01 3.52192253e-01
-3.30261528e-01 -7.27635562e-01 2.57208884e-01 1.42372763e+00
3.30265790e-01 -6.98881805e-01 -8.02846372e-01 5.68026185e-01
-2.07886815e+00 -8.56501997e-01 -1.20611489e-01 1.84620035e+00
1.46623957e+00 -3.83844934e-02 -1.01103134e-01 -2.92106241e-01
6.42110765e-01 5.64775467e-02 -4.26032394e-01 -8.38746428e-01
1.33504719e-01 3.08845609e-01 -1.69966444e-01 1.02469862e+00
-2.34306246e-01 1.45302892e+00 6.92075300e+00 7.19127655e-01
-6.44524872e-01 -1.23206496e-01 1.33245751e-01 3.98253441e-01
-2.48164251e-01 -5.10608777e-02 -8.25232267e-01 5.34613058e-02
1.06283367e+00 -7.90644228e-01 4.64194864e-01 4.71367598e-01
4.82355118e-01 -3.97320777e-01 -1.06373298e+00 3.28384489e-01
4.16865200e-02 -1.58744633e+00 4.52341318e-01 -2.53884584e-01
4.77612466e-01 -6.25722349e-01 -3.62429500e-01 7.52843142e-01
9.89218175e-01 -8.39087844e-01 5.97602762e-02 7.57402599e-01
6.70624733e-01 -3.86440486e-01 6.61081910e-01 7.16853619e-01
-1.09225237e+00 2.20296755e-01 -3.56343985e-01 -3.02588105e-01
4.08458650e-01 9.48012844e-02 -2.03948927e+00 8.04103851e-01
-2.46319786e-01 1.39886528e-01 -3.16889554e-01 6.81142628e-01
-5.48262298e-01 6.29669726e-01 -3.75121436e-03 -5.46454430e-01
2.26535395e-01 -8.81385133e-02 4.39297020e-01 1.34842110e+00
3.35838377e-01 5.07081568e-01 4.82530951e-01 8.01710784e-01
-2.71716088e-01 2.24043384e-01 -8.90345037e-01 -2.11981326e-01
8.09991121e-01 1.15601456e+00 -1.96579546e-01 -6.63539886e-01
-1.82140246e-01 9.14291859e-01 4.15611088e-01 1.63292333e-01
-2.59006441e-01 -2.95123547e-01 -5.77911772e-02 -2.05278650e-01
-2.74073392e-01 -3.66930455e-01 -1.09001495e-01 -9.19834495e-01
-3.35157037e-01 -1.01912057e+00 2.41869763e-01 -1.13248098e+00
-1.33285797e+00 7.17028379e-01 4.82921191e-02 -8.19672942e-01
-1.09465039e+00 -1.46414235e-01 -7.59628475e-01 1.07539868e+00
-1.06543159e+00 -1.08968318e+00 -2.41976336e-01 6.50259972e-01
1.29301775e+00 -3.70728314e-01 1.44068229e+00 -3.34702462e-01
-1.71369374e-01 3.85647327e-01 -4.07535046e-01 1.88358456e-01
8.03545356e-01 -1.40409160e+00 4.76281583e-01 2.35255703e-01
-5.83283082e-02 8.16593647e-01 6.42160714e-01 -1.03381276e+00
-1.49814391e+00 -8.57125044e-01 1.15998971e+00 -3.04642707e-01
7.30534613e-01 -1.64444670e-01 -1.04846442e+00 3.31932127e-01
8.25210214e-01 -9.25126553e-01 6.75745904e-01 2.02185854e-01
-1.99695632e-01 3.95760000e-01 -1.13356042e+00 7.46025980e-01
8.24068546e-01 -7.86569893e-01 -1.15186501e+00 5.59856772e-01
8.10834050e-01 -7.07338214e-01 -8.18860948e-01 -1.24101683e-01
3.81711155e-01 -7.24174619e-01 6.13710463e-01 -5.44663668e-01
2.86946684e-01 1.24072051e-02 3.63699824e-01 -1.81585693e+00
2.28114277e-01 -1.05494535e+00 -2.91479915e-01 1.09003735e+00
5.64544141e-01 -5.68705678e-01 8.22562099e-01 9.11001563e-01
-4.18619812e-01 -3.71401370e-01 -8.06219816e-01 -4.36074197e-01
1.27751037e-01 -2.15068925e-02 5.36235452e-01 8.10837448e-01
1.18877733e+00 1.02202082e+00 -6.05709493e-01 -6.08171344e-01
4.24564779e-02 7.09639639e-02 1.07568777e+00 -9.95141566e-01
-1.69157088e-01 -1.77648112e-01 5.15793145e-01 -1.56592572e+00
3.41807693e-01 -7.10535765e-01 1.75193742e-01 -1.97563791e+00
-3.83951925e-02 -3.29504132e-01 5.60496271e-01 5.35760224e-01
-2.81118602e-01 -3.69672090e-01 1.69314817e-01 5.49015142e-02
-7.50669837e-01 8.19660544e-01 1.77558088e+00 -3.29334944e-01
-5.91840267e-01 2.85303202e-02 -4.26624179e-01 4.57420021e-01
8.85102451e-01 -1.12875730e-01 -7.59418547e-01 -2.01222539e-01
1.70017958e-01 1.02263021e+00 7.86372577e-04 -5.34634113e-01
9.92999554e-01 -3.07246387e-01 8.26041307e-03 -7.90283620e-01
6.95755303e-01 -1.39581054e-01 -1.67882621e-01 3.84729087e-01
-8.89354467e-01 2.16109619e-01 2.65399396e-01 4.12102461e-01
-2.64489025e-01 -3.82749200e-01 1.17611527e-01 -5.66918671e-01
-4.97564554e-01 -2.28872076e-01 -1.08471084e+00 1.52530700e-01
4.61556196e-01 -4.38381210e-02 -8.47332239e-01 -9.98135865e-01
-5.35885930e-01 5.02355933e-01 3.08254957e-02 4.20423895e-01
1.12769878e+00 -9.46037710e-01 -6.78608716e-01 -1.47261426e-01
-2.28787325e-02 -4.20143545e-01 2.45518982e-01 4.86608893e-01
-3.62516761e-01 8.98947001e-01 -1.19305581e-01 -1.37903243e-01
-1.38303208e+00 -1.41355759e-02 2.96995133e-01 -8.72719169e-01
-5.10311604e-01 4.84703213e-01 -2.06712574e-01 -7.84499347e-01
5.75312041e-02 -3.73843312e-02 -4.81505126e-01 -1.29377738e-01
4.35282916e-01 3.52647781e-01 1.35438750e-02 -4.03953940e-02
1.64501950e-01 -6.85749799e-02 -2.41137892e-01 -6.49094582e-01
5.98786056e-01 -3.45456511e-01 -3.17002274e-02 3.43136609e-01
3.30933720e-01 -3.70711484e-03 -5.23113251e-01 -6.51304185e-01
4.14781831e-03 -3.53908330e-01 -3.44031334e-01 -1.26010251e+00
-3.56583565e-01 6.91020429e-01 -3.87696445e-01 5.64173281e-01
5.85647702e-01 -4.21359152e-01 9.29419458e-01 1.12803757e+00
5.82060039e-01 -1.26554203e+00 4.98979747e-01 1.24623573e+00
1.09882879e+00 -1.02903771e+00 -8.95091146e-02 -5.54891825e-01
-1.08658826e+00 1.39922988e+00 9.58951175e-01 4.40709323e-01
5.95221296e-02 1.38965771e-01 1.92963764e-01 -2.59455562e-01
-1.54779279e+00 -1.85875535e-01 1.76365703e-01 6.33397102e-01
6.21502161e-01 -1.15205012e-02 -4.85464960e-01 3.23035538e-01
-5.13669908e-01 1.26214504e-01 8.56920898e-01 9.11335826e-01
-7.81625032e-01 -1.39684749e+00 -1.89078227e-01 4.75683838e-01
1.46343866e-02 -4.12700772e-01 -1.06978106e+00 5.78493237e-01
-9.22440171e-01 1.81893182e+00 -3.30042064e-01 -4.90653187e-01
5.09658813e-01 7.04958081e-01 6.05421543e-01 -9.77801681e-01
-1.32710958e+00 -2.80848682e-01 1.07227814e+00 -1.54279619e-01
-2.29532957e-01 7.47253746e-02 -1.26883054e+00 -5.99153757e-01
-5.99917293e-01 7.65040159e-01 3.24043989e-01 9.60445821e-01
2.53168881e-01 5.44387937e-01 5.26995897e-01 -2.26451874e-01
-7.63733029e-01 -1.58584750e+00 -1.21837378e-01 -2.23749205e-02
-2.91748464e-01 -3.00937593e-01 9.57959071e-02 -1.54145062e-01] | [12.571342468261719, 8.14500904083252] |
bb01c0a5-8724-46c7-acb0-719e87c67dc6 | counterfactually-comparing-abstaining | 2305.10564 | null | https://arxiv.org/abs/2305.10564v1 | https://arxiv.org/pdf/2305.10564v1.pdf | Counterfactually Comparing Abstaining Classifiers | Abstaining classifiers have the option to abstain from making predictions on inputs that they are unsure about. These classifiers are becoming increasingly popular in high-stake decision-making problems, as they can withhold uncertain predictions to improve their reliability and safety. When evaluating black-box abstaining classifier(s), however, we lack a principled approach that accounts for what the classifier would have predicted on its abstentions. These missing predictions are crucial when, e.g., a radiologist is unsure of their diagnosis or when a driver is inattentive in a self-driving car. In this paper, we introduce a novel approach and perspective to the problem of evaluating and comparing abstaining classifiers by treating abstentions as missing data. Our evaluation approach is centered around defining the counterfactual score of an abstaining classifier, defined as the expected performance of the classifier had it not been allowed to abstain. We specify the conditions under which the counterfactual score is identifiable: if the abstentions are stochastic, and if the evaluation data is independent of the training data (ensuring that the predictions are missing at random), then the score is identifiable. Note that, if abstentions are deterministic, then the score is unidentifiable because the classifier can perform arbitrarily poorly on its abstentions. Leveraging tools from observational causal inference, we then develop nonparametric and doubly robust methods to efficiently estimate this quantity under identification. Our approach is examined in both simulated and real data experiments. | ['Aaditya Ramdas', 'Aditya Gangrade', 'Yo Joong Choe'] | 2023-05-17 | null | null | null | null | ['causal-inference', 'causal-inference'] | ['knowledge-base', 'miscellaneous'] | [ 3.86818349e-01 7.63638198e-01 -7.50643492e-01 -4.03185397e-01
-7.52613783e-01 -5.67166388e-01 6.38992786e-01 -4.54351269e-02
-5.52636385e-01 1.20402896e+00 8.87960866e-02 -1.09089243e+00
-1.72197700e-01 -6.05635941e-01 -9.88742590e-01 -7.93906152e-01
1.12991728e-01 1.96076229e-01 -2.04058215e-01 2.38767192e-01
5.52479997e-02 4.19068038e-01 -1.57285547e+00 9.20912325e-02
1.06056774e+00 8.61090541e-01 -3.75543505e-01 5.11052012e-01
5.05367219e-01 1.14739156e+00 -4.68052775e-01 -6.12849951e-01
3.63953114e-01 -4.66110170e-01 -5.44325292e-01 1.70643963e-02
3.22654486e-01 -4.26394045e-01 -2.54121065e-01 1.13631558e+00
-8.50175247e-02 7.01104850e-02 9.30881977e-01 -1.60716677e+00
-4.33963627e-01 6.78721786e-01 -2.43571550e-01 6.65404573e-02
1.92899004e-01 5.08343995e-01 9.50276017e-01 -2.54629314e-01
4.60471839e-01 1.02250504e+00 4.45616126e-01 5.79338133e-01
-1.52174127e+00 -9.40604866e-01 3.57186049e-01 2.39061788e-01
-1.19665146e+00 -6.24220669e-01 5.54699242e-01 -6.93825483e-01
3.40122372e-01 5.53166270e-01 5.27363658e-01 1.43178022e+00
6.65217996e-01 5.32597542e-01 1.37696707e+00 -3.49910378e-01
7.33311057e-01 2.32054353e-01 1.47850454e-01 2.53463924e-01
7.98365891e-01 1.07963264e+00 -3.32783252e-01 -5.33249915e-01
9.29669514e-02 1.56759486e-01 -8.79791006e-02 -3.38578641e-01
-8.91815305e-01 9.50366914e-01 7.04471245e-02 -3.15809518e-01
-5.50630033e-01 4.71252091e-02 -5.32932905e-03 4.47324634e-01
2.58804142e-01 1.96009710e-01 -2.53585666e-01 1.55415863e-01
-8.00192237e-01 6.08261883e-01 7.23666608e-01 6.11407995e-01
3.86564881e-01 -8.00032690e-02 -5.12053132e-01 2.45363966e-01
-8.68974999e-02 5.28462052e-01 3.74562204e-01 -1.15918291e+00
2.03128770e-01 2.86605448e-01 7.72094548e-01 -3.02158177e-01
-3.91198955e-02 -1.96399570e-01 -5.19742072e-01 8.12865913e-01
6.38220191e-01 -4.23943639e-01 -8.49255562e-01 2.02623439e+00
1.36718124e-01 9.96721089e-02 7.46208951e-02 7.25289702e-01
-1.12178378e-01 1.14342891e-01 5.37526608e-01 -6.91334605e-01
1.03436124e+00 -4.08113450e-02 -7.58639634e-01 -2.39251405e-01
7.86168635e-01 -2.61287421e-01 8.34193528e-01 4.40313935e-01
-9.35142994e-01 -1.95399061e-01 -1.16303694e+00 3.60651463e-01
1.83247194e-01 -1.73162952e-01 6.02361321e-01 7.96972871e-01
-5.10298491e-01 6.57404721e-01 -8.66985023e-01 1.98972940e-01
5.80564439e-01 2.56078660e-01 -2.43948549e-01 -3.35929766e-02
-1.34903896e+00 9.80022013e-01 2.21546084e-01 2.43289955e-02
-1.14055824e+00 -6.87700570e-01 -5.90366900e-01 3.06547970e-01
8.26111972e-01 -7.68531382e-01 1.49885523e+00 -1.12855911e+00
-9.48325098e-01 7.24940956e-01 -1.90649360e-01 -9.68476832e-01
1.15528810e+00 8.09982568e-02 -6.38417840e-01 -4.68838066e-01
4.34110463e-01 5.02087176e-02 9.08545494e-01 -8.59744787e-01
-8.69517684e-01 -6.68958724e-01 -1.34703731e-02 1.74219638e-01
3.80536884e-01 -3.09904963e-01 6.88571811e-01 -4.20685798e-01
-2.11202517e-01 -1.19240224e+00 -3.35349232e-01 1.78338930e-01
-7.61879683e-01 -1.15249820e-01 6.12937212e-01 -1.27306268e-01
9.16497827e-01 -2.20749807e+00 -6.96651995e-01 7.77461529e-02
1.77161649e-01 -2.13261917e-02 5.07856071e-01 -6.81488365e-02
-3.95603448e-01 2.79371679e-01 -4.84037906e-01 1.11391135e-01
-2.69055204e-03 2.71299660e-01 -4.79452014e-01 8.16396058e-01
2.34976277e-01 6.07995391e-01 -5.41552007e-01 -2.70738125e-01
1.42704278e-01 -2.05056176e-01 -4.91446912e-01 -2.43233480e-02
5.92501136e-03 2.96947598e-01 -5.79218745e-01 1.59655422e-01
6.43533885e-01 3.48018743e-02 2.28847519e-01 6.55885756e-01
-2.40740895e-01 2.37537652e-01 -1.19141030e+00 6.83169782e-01
-3.51590633e-01 3.96158904e-01 -7.22523779e-02 -1.20594776e+00
3.82781595e-01 2.59300828e-01 -3.56656983e-02 -3.13294232e-01
2.07267508e-01 3.23644370e-01 3.53186876e-01 -3.95401448e-01
3.53885256e-02 -9.95671451e-01 -2.09492713e-01 4.28416401e-01
-4.32646900e-01 5.72834089e-02 -3.11065763e-01 1.37028590e-01
1.19913471e+00 -3.40303153e-01 6.67743623e-01 -3.45897049e-01
2.28089601e-01 2.01860428e-01 9.80030358e-01 1.16270423e+00
-4.39389318e-01 2.39139646e-01 8.91540110e-01 -1.82959661e-01
-7.28530705e-01 -1.37903357e+00 -5.80163896e-01 4.11258161e-01
-1.73807129e-01 5.01979351e-01 -3.88933271e-01 -9.09965396e-01
4.87303942e-01 1.52978516e+00 -1.01065588e+00 -6.73402250e-01
-2.93297712e-02 -6.40410364e-01 1.96663201e-01 3.55820239e-01
9.90220979e-02 -5.47296524e-01 -6.66039109e-01 -8.36255103e-02
1.54915033e-03 -5.11924505e-01 -3.45479161e-01 1.78632051e-01
-7.19618380e-01 -1.48169243e+00 -3.10966104e-01 1.14516774e-02
6.21579707e-01 8.01715031e-02 7.84396112e-01 -2.82580201e-02
1.24369904e-01 4.57419865e-02 1.12260647e-01 -8.89918029e-01
-5.52086413e-01 -4.63994175e-01 2.41283551e-01 1.16303027e-01
6.13628447e-01 -2.81638056e-01 -6.76312029e-01 9.49772522e-02
-6.54598653e-01 -2.75829703e-01 4.80083555e-01 1.06724012e+00
5.01388013e-01 9.62331742e-02 9.29460883e-01 -1.56170583e+00
6.25578463e-01 -8.46338630e-01 -8.45135927e-01 2.89753735e-01
-1.06042194e+00 1.75980210e-01 6.47333503e-01 -7.30557323e-01
-1.41468239e+00 2.12817609e-01 2.36697957e-01 -2.86583781e-01
-4.15579736e-01 3.46599579e-01 -3.03026289e-01 4.48306262e-01
9.02808487e-01 -2.02437729e-01 2.25142807e-01 -1.88954204e-01
1.61952868e-01 7.47537136e-01 4.77992654e-01 -5.61316371e-01
6.31386697e-01 5.70375741e-01 2.28919417e-01 -4.79339242e-01
-1.13161600e+00 -3.71278543e-03 -1.05711304e-01 -9.06987488e-02
5.53858161e-01 -8.99566948e-01 -1.13157904e+00 2.53439900e-02
-6.24428093e-01 -1.68168321e-01 -8.99945259e-01 9.51159179e-01
-8.35910082e-01 -7.04276860e-02 2.39555031e-01 -1.44011939e+00
4.84408498e-01 -1.15543842e+00 5.20217419e-01 -2.29635462e-02
-4.55460548e-01 -7.79744983e-01 -1.68548495e-01 4.98791873e-01
8.21119696e-02 4.65069145e-01 1.11455405e+00 -1.07463181e+00
-6.38781011e-01 -7.65571594e-01 3.23469996e-01 3.63273591e-01
-7.24357069e-02 -1.76251516e-01 -9.64366794e-01 -2.85144486e-02
3.74947429e-01 -2.16116786e-01 7.39933252e-01 8.26034188e-01
1.38069308e+00 -9.41429913e-01 -4.56979573e-01 4.90064397e-02
9.13929522e-01 4.23466682e-01 3.77828479e-01 2.91970000e-02
-7.60455653e-02 7.27286577e-01 6.30924046e-01 3.66533250e-01
1.94451332e-01 5.08607149e-01 4.85927522e-01 2.49056309e-01
4.13860857e-01 -6.26947463e-01 3.21995229e-01 -4.04730320e-01
2.03474656e-01 -9.85182375e-02 -4.88024890e-01 6.10606611e-01
-1.81145966e+00 -1.32340181e+00 -3.12723935e-01 3.01098609e+00
7.68501222e-01 4.02502090e-01 1.35474369e-01 1.53965861e-01
6.86224818e-01 -2.38509655e-01 -1.11396658e+00 -6.47640347e-01
-3.20445858e-02 -2.87738703e-02 8.15965414e-01 7.45343208e-01
-7.46167839e-01 2.87338048e-01 6.38954496e+00 6.51505828e-01
-6.03753030e-01 1.21590681e-01 1.19937062e+00 -4.74684715e-01
-4.99676585e-01 5.10214150e-01 -7.32708216e-01 6.53229475e-01
1.28412664e+00 -9.02033806e-01 6.10905550e-02 9.70102727e-01
4.96301055e-01 -6.57687128e-01 -1.61354303e+00 2.85985172e-01
-3.54400754e-01 -1.19486940e+00 -2.43860200e-01 3.52559090e-01
6.77443326e-01 -3.66049439e-01 3.31909806e-01 4.80128944e-01
9.38591480e-01 -1.14524603e+00 1.12857234e+00 7.43899763e-01
8.22277129e-01 -9.24018204e-01 8.53963673e-01 1.02992976e+00
-1.46022990e-01 -4.27121431e-01 -1.74258761e-02 -5.36566079e-01
-1.38666108e-02 8.71143460e-01 -7.05636084e-01 1.71786547e-01
1.80759355e-01 9.11871567e-02 -2.65906043e-02 7.09625840e-01
-5.20666063e-01 9.02353764e-01 -2.02143714e-01 7.48447552e-02
-1.12861283e-01 -4.17619050e-02 5.91318369e-01 4.64173168e-01
1.86835751e-01 4.31194186e-01 -5.62433014e-03 1.20660961e+00
-1.10520326e-01 -3.43702942e-01 -1.04673636e+00 2.02512249e-01
5.69093168e-01 6.51380777e-01 -1.48726612e-01 -2.40701377e-01
-3.22160751e-01 2.71786302e-01 -5.91248684e-02 5.09092033e-01
-6.52422965e-01 -3.89637649e-02 7.41796613e-01 3.85192633e-01
-3.30293745e-01 3.56053084e-01 -7.10753858e-01 -1.01705968e+00
1.44979089e-01 -4.89914477e-01 6.25511527e-01 -5.90601742e-01
-1.30983269e+00 -3.89176570e-02 1.97388023e-01 -1.03765810e+00
-6.36949301e-01 -4.28212106e-01 -7.44474411e-01 1.14904165e+00
-1.28670859e+00 -5.79905808e-01 4.38800037e-01 4.03357476e-01
1.98176324e-01 1.57552257e-01 4.56953228e-01 -3.64031941e-01
-5.12527823e-01 6.18247986e-01 2.51263678e-02 -1.61875874e-01
3.53848249e-01 -1.10040963e+00 -1.99985087e-01 9.47174788e-01
-1.50323242e-01 5.14720559e-01 1.00361443e+00 -8.35805118e-01
-9.41580594e-01 -1.06422627e+00 9.37151670e-01 -6.30894840e-01
6.98067486e-01 -2.51922578e-01 -8.41398180e-01 9.94427323e-01
-5.69571257e-01 1.54873475e-01 7.01488674e-01 4.84712392e-01
-1.02436230e-01 5.39700389e-02 -1.51801229e+00 7.89930105e-01
7.59655595e-01 -3.88352871e-01 -8.92632306e-01 4.14433479e-01
3.92069280e-01 -6.98338598e-02 -3.54226440e-01 2.87409037e-01
5.88531971e-01 -9.81245399e-01 6.27412498e-01 -1.15425396e+00
3.43358994e-01 -3.93436737e-02 -8.93988311e-02 -1.36241126e+00
-2.04519272e-01 -4.89142060e-01 -2.75074709e-02 6.94477141e-01
7.78481960e-01 -1.03952968e+00 8.89812231e-01 1.34584165e+00
1.10590957e-01 -5.89091361e-01 -1.52574158e+00 -1.04019380e+00
5.78925610e-01 -6.44833803e-01 6.09462500e-01 8.79396319e-01
1.64265245e-01 1.71783477e-01 -3.15321594e-01 1.97702363e-01
9.11434412e-01 1.24084398e-01 3.66071433e-01 -1.49439633e+00
-1.27996743e-01 -1.66736200e-01 -2.46838152e-01 -3.99203062e-01
6.36708915e-01 -8.07000101e-01 1.18706096e-02 -8.16723168e-01
5.62875807e-01 -3.16721022e-01 -2.10283861e-01 5.29679418e-01
-4.15051460e-01 -5.73856115e-01 1.95100680e-02 1.95831642e-01
7.43788928e-02 5.89925528e-01 1.05145001e+00 -1.92801822e-02
-8.50237831e-02 7.40437686e-01 -1.17040014e+00 8.56202245e-01
7.34591067e-01 -8.36428940e-01 -5.51957428e-01 4.04156089e-01
-2.16000024e-02 7.29592204e-01 9.63087618e-01 -5.40606380e-01
1.68206528e-01 -7.33488619e-01 3.39203835e-01 -2.84064919e-01
7.52249956e-02 -9.04914200e-01 5.46954274e-01 8.17509174e-01
-8.66879344e-01 -3.65696400e-01 -4.89453487e-02 9.98323798e-01
8.94238725e-02 -3.33265185e-01 8.57167900e-01 5.43685979e-04
1.72429252e-02 1.80672273e-01 -7.59337485e-01 2.30650663e-01
1.27902710e+00 -1.09057784e-01 -4.18451577e-01 -5.78844905e-01
-1.02035511e+00 2.64608324e-01 5.38715303e-01 1.44580305e-01
4.49804395e-01 -1.25147259e+00 -8.30299795e-01 1.03306241e-01
1.95672572e-01 -3.30284655e-01 4.17395979e-01 8.63735616e-01
3.12179953e-01 5.17726421e-01 2.32748866e-01 -1.63992316e-01
-7.89744377e-01 8.45893681e-01 5.08099973e-01 -1.20301075e-01
-4.48029250e-01 3.74724060e-01 6.78231776e-01 -9.13666002e-03
-1.67191803e-01 -1.52003303e-01 2.16493219e-01 -2.26944223e-01
4.85616475e-01 4.86155570e-01 1.49090052e-01 -4.45725024e-01
-1.18210755e-01 -5.35283923e-01 -3.02776664e-01 -3.76454562e-01
9.39052939e-01 -5.53689525e-02 4.45091069e-01 7.00353563e-01
6.78047836e-01 -5.90155460e-02 -1.46630359e+00 2.42951643e-02
-2.73773670e-02 -7.28036761e-01 2.20443696e-01 -1.04281521e+00
-5.99591851e-01 5.97111881e-01 4.10701692e-01 2.48317301e-01
7.39628911e-01 -1.17904313e-01 3.15712914e-02 1.53741390e-01
4.57588613e-01 -9.76444304e-01 -7.48766959e-01 -1.86027527e-01
9.25115526e-01 -1.33510351e+00 1.12075666e-02 -2.26032689e-01
-8.07191312e-01 3.87438655e-01 2.24076211e-01 -5.99375293e-02
6.89028323e-01 9.63308364e-02 -2.13684201e-01 8.40534717e-02
-1.17987597e+00 1.00937486e-01 2.13854507e-01 5.39155543e-01
1.83063522e-01 8.09621930e-01 -6.10999465e-01 9.17489588e-01
-3.88765067e-01 2.75236815e-01 1.03815198e+00 7.46730626e-01
-2.91960031e-01 -6.97054863e-01 -4.76606190e-01 1.13473821e+00
-6.17519796e-01 8.43593553e-02 -3.44427556e-01 9.09866512e-01
1.40153974e-01 1.17233026e+00 5.38816815e-03 -7.46828329e-04
5.86398125e-01 3.26051086e-01 6.41489169e-03 -6.74445212e-01
4.47276235e-02 -4.04714853e-01 1.98752820e-01 -4.69407529e-01
-1.02400057e-01 -1.28376889e+00 -8.68796110e-01 -4.24349815e-01
-4.26759064e-01 2.43078128e-01 2.49489009e-01 1.16486776e+00
1.00223310e-01 2.35680863e-01 9.28358257e-01 -5.42844236e-02
-1.37899911e+00 -6.52770996e-01 -9.42017376e-01 2.40780309e-01
6.28042758e-01 -8.55014443e-01 -9.41525340e-01 -1.96731493e-01] | [8.537556648254395, 5.456897258758545] |
f9cc362b-8757-4784-97f4-14de22c12a1d | effective-spectral-unmixing-via-robust | 1409.0685 | null | http://arxiv.org/abs/1409.0685v4 | http://arxiv.org/pdf/1409.0685v4.pdf | Effective Spectral Unmixing via Robust Representation and Learning-based Sparsity | Hyperspectral unmixing (HU) plays a fundamental role in a wide range of
hyperspectral applications. It is still challenging due to the common presence
of outlier channels and the large solution space. To address the above two
issues, we propose a novel model by emphasizing both robust representation and
learning-based sparsity. Specifically, we apply the $\ell_{2,1}$-norm to
measure the representation error, preventing outlier channels from dominating
our objective. In this way, the side effects of outlier channels are greatly
relieved. Besides, we observe that the mixed level of each pixel varies over
image grids. Based on this observation, we exploit a learning-based sparsity
method to simultaneously learn the HU results and a sparse guidance map. Via
this guidance map, the sparsity constraint in the $\ell_{p}\!\left(\!0\!<\!
p\!\leq\!1\right)$-norm is adaptively imposed according to the learnt mixed
level of each pixel. Compared with state-of-the-art methods, our model is
better suited to the real situation, thus expected to achieve better HU
results. The resulted objective is highly non-convex and non-smooth, and so it
is hard to optimize. As a profound theoretical contribution, we propose an
efficient algorithm to solve it. Meanwhile, the convergence proof and the
computational complexity analysis are systematically provided. Extensive
evaluations verify that our method is highly promising for the HU task---it
achieves very accurate guidance maps and much better HU results compared with
state-of-the-art methods. | ['Chunhong Pan', 'Ying Wang', 'Feiyun Zhu', 'Gaofeng Meng', 'Bin Fan'] | 2014-09-02 | null | null | null | null | ['hyperspectral-unmixing'] | ['computer-vision'] | [ 3.57224196e-01 -2.26036206e-01 -4.47170250e-02 2.97980495e-02
-7.91567862e-01 -9.14685950e-02 9.69711766e-02 -8.69836584e-02
-1.01724371e-01 7.48306274e-01 -1.09412381e-02 -2.22020783e-02
-4.89286363e-01 -9.25570726e-01 -6.38021588e-01 -1.32735527e+00
-6.71318695e-02 -1.50157824e-01 -2.51000047e-01 -1.83751255e-01
1.68536708e-03 1.90633550e-01 -1.56430078e+00 -1.93833590e-01
1.44704950e+00 1.31288099e+00 3.38611156e-01 -1.59265503e-01
1.65850505e-01 5.19790113e-01 -1.29391640e-01 1.10326916e-01
5.34271240e-01 -7.08253860e-01 -4.40386795e-02 4.75699902e-01
4.50292557e-01 -9.03773084e-02 -2.07087830e-01 1.81022978e+00
4.13601249e-01 3.36337656e-01 5.36706269e-01 -9.10737455e-01
-3.88099909e-01 3.12437415e-01 -1.16611886e+00 -1.56657577e-01
-5.70261888e-02 2.08071768e-01 1.10387158e+00 -8.59483421e-01
2.80447543e-01 5.56680143e-01 5.78871965e-01 -6.51184171e-02
-1.23726940e+00 -8.03581119e-01 3.63035947e-01 1.69126853e-01
-1.68455875e+00 -1.54411018e-01 9.45469260e-01 -4.80856568e-01
2.32046753e-01 3.98295820e-01 6.76423907e-01 4.68305618e-01
-1.89676285e-01 6.66633785e-01 1.31679165e+00 -3.05447310e-01
2.76333392e-01 -8.98366943e-02 -8.01417381e-02 6.70259535e-01
2.27353811e-01 5.97376227e-02 -3.97139937e-01 3.72605920e-02
8.67728114e-01 1.19490154e-01 -9.76382375e-01 -2.85801113e-01
-1.06829298e+00 7.66022265e-01 6.81974888e-01 3.95710051e-01
-5.16742885e-01 -2.20077723e-01 2.33928096e-02 6.35762140e-02
5.56682825e-01 2.79874086e-01 -5.88455424e-02 3.49651635e-01
-1.19729471e+00 1.64527759e-01 3.71766776e-01 7.03712463e-01
1.13388872e+00 2.37042442e-01 -3.03608268e-01 9.92682755e-01
3.59806359e-01 6.45657599e-01 1.57298490e-01 -8.29709291e-01
4.62087452e-01 5.75595975e-01 1.41944975e-01 -1.27766073e+00
-3.10103714e-01 -9.83774245e-01 -1.40004957e+00 2.66961038e-01
4.31443423e-01 -8.04356262e-02 -6.48881614e-01 1.71037471e+00
2.04647332e-01 4.84353870e-01 -1.18538231e-01 1.06204808e+00
5.09382188e-01 9.11394119e-01 -2.53348440e-01 -8.69829953e-01
1.07776809e+00 -7.59055436e-01 -8.38520408e-01 -2.00394630e-01
2.69933641e-01 -7.18246460e-01 1.00875139e+00 5.00639021e-01
-9.39922154e-01 -3.94254953e-01 -1.05794978e+00 3.87834042e-01
8.25110823e-02 4.08936203e-01 4.98745143e-01 4.15483117e-01
-5.70697367e-01 3.94989848e-01 -5.26377201e-01 -3.30148055e-03
2.47231767e-01 1.30317628e-01 -1.26405701e-01 -3.85114342e-01
-8.86052191e-01 3.40893745e-01 4.33108270e-01 4.83103573e-01
-3.21431100e-01 -7.54055381e-01 -8.97693515e-01 1.25774026e-01
6.31085813e-01 -3.80377591e-01 7.22185671e-01 -7.21211314e-01
-1.37189996e+00 4.78255540e-01 -3.56068730e-01 -5.57009913e-02
4.90216583e-01 1.89435463e-02 -3.90136063e-01 1.77965075e-01
1.72277182e-01 1.51515573e-01 1.10941422e+00 -1.48819137e+00
-6.34587646e-01 -5.25935411e-01 -4.64602321e-01 3.07133943e-01
-5.96931756e-01 -4.54437345e-01 -4.45160836e-01 -8.78757119e-01
6.38727248e-01 -7.40454972e-01 -3.52011085e-01 -4.85667922e-02
-3.96635950e-01 2.34536618e-01 6.68931782e-01 -6.63466334e-01
1.58117115e+00 -2.33387923e+00 2.02849075e-01 5.49518526e-01
2.01328471e-01 3.21364582e-01 2.70407796e-02 1.60596684e-01
-1.49417698e-01 -1.77714735e-01 -8.76860201e-01 -3.53959613e-02
-1.75340563e-01 2.62263231e-02 -1.85629904e-01 7.46500015e-01
-1.11159511e-01 5.15023768e-01 -1.02646375e+00 5.12669049e-02
4.15386230e-01 5.08893430e-01 -4.24722284e-01 1.73381671e-01
-1.71672016e-01 6.78641856e-01 -4.22732890e-01 6.88624084e-01
1.11689448e+00 -1.94037974e-01 1.69944122e-01 -4.64544505e-01
-5.56489646e-01 -3.53222489e-01 -1.68481183e+00 1.54209244e+00
-2.92675287e-01 1.82884663e-01 5.85296512e-01 -1.26210093e+00
1.04608130e+00 1.99302748e-01 7.17934310e-01 -7.49653041e-01
5.18507361e-02 4.70286250e-01 -1.10767670e-01 -3.92340511e-01
1.22836702e-01 -3.16369116e-01 2.74432808e-01 9.98149514e-02
-3.36209297e-01 -3.31986487e-01 2.27552876e-01 -2.16132537e-01
7.17640519e-01 -1.56504139e-01 4.38891828e-01 -5.73996723e-01
7.95904040e-01 -2.86375821e-01 8.27693105e-01 6.86858594e-01
-1.66360557e-01 5.81567883e-01 4.05041099e-01 -1.38364896e-01
-6.80738568e-01 -7.62387991e-01 -4.22010332e-01 6.27729356e-01
4.46570307e-01 -1.77728176e-01 -6.76434278e-01 -2.38780677e-01
-6.89110532e-02 5.44882536e-01 -4.00182784e-01 -2.26884931e-02
-4.16838676e-01 -1.32823217e+00 -8.34067166e-02 1.99206620e-01
8.15797389e-01 -7.95847118e-01 -1.96846515e-01 2.74938613e-01
-2.60097682e-01 -8.92263830e-01 -3.86087835e-01 1.51551574e-01
-7.57312298e-01 -1.01363981e+00 -9.10228372e-01 -6.13887906e-01
8.26335073e-01 6.09371245e-01 6.95573747e-01 -1.39854684e-01
-1.36646539e-01 -4.02647294e-02 -4.01149452e-01 -2.35318810e-01
8.72696787e-02 -2.60207295e-01 -9.76232067e-02 4.97688562e-01
8.76976848e-02 -7.18791068e-01 -7.78141022e-01 3.96716863e-01
-1.19073963e+00 -6.62144795e-02 6.19176090e-01 1.08433223e+00
1.01593518e+00 6.73863888e-01 4.00419325e-01 -7.83874869e-01
3.43604684e-01 -4.72629845e-01 -9.40508664e-01 1.82625547e-01
-6.87820733e-01 -3.33601147e-01 8.24342132e-01 -1.48913220e-01
-9.55454946e-01 1.34194806e-01 -7.89233372e-02 -4.99510467e-01
1.98936611e-02 7.28018045e-01 -4.13511932e-01 -3.18394542e-01
4.73926783e-01 4.03032482e-01 -6.16259053e-02 -4.45262164e-01
2.87570417e-01 4.88574773e-01 6.48396969e-01 -5.32544792e-01
1.04015088e+00 7.75565982e-01 1.49104252e-01 -1.15467453e+00
-9.69306588e-01 -7.32213676e-01 -2.26107165e-01 -2.01386660e-01
5.22720993e-01 -1.00446868e+00 -5.18520176e-01 6.09860778e-01
-7.28360057e-01 -2.85974711e-01 -3.85382771e-01 5.94330013e-01
-2.34250709e-01 6.67973518e-01 -3.87845248e-01 -9.09653306e-01
-3.35731328e-01 -1.24711359e+00 8.16424370e-01 3.77002835e-01
3.74976873e-01 -7.96344638e-01 -1.47035599e-01 2.64269263e-01
3.92038703e-01 3.58152807e-01 8.31306696e-01 1.64641827e-01
-5.69929779e-01 1.43357925e-02 -6.37957096e-01 5.47262013e-01
1.15828432e-01 -1.18717633e-01 -8.67985427e-01 -4.33000743e-01
3.71258318e-01 4.45290934e-03 1.06907749e+00 6.61681294e-01
1.41811657e+00 -3.16670120e-01 -5.11862375e-02 1.02334356e+00
1.82984102e+00 6.14978895e-02 7.58231461e-01 2.89524823e-01
6.34709418e-01 5.45066118e-01 5.39919794e-01 8.53810728e-01
2.64971349e-02 5.59362292e-01 7.88260579e-01 -3.15211207e-01
1.62270591e-01 -1.55114075e-02 1.60836875e-01 7.82627046e-01
-1.65610641e-01 -1.64652213e-01 -5.23157597e-01 4.86464947e-01
-1.97243810e+00 -9.41706359e-01 -4.66481626e-01 2.56621742e+00
6.07558906e-01 -2.38232628e-01 -2.60916114e-01 4.01829183e-01
8.05569828e-01 6.35837734e-01 -4.62677658e-01 3.33794773e-01
-4.38869566e-01 3.38329345e-01 5.37678123e-01 4.12516981e-01
-1.10863054e+00 5.64544678e-01 5.32143259e+00 1.07066441e+00
-1.32234728e+00 -6.44723624e-02 4.84332442e-01 2.89461911e-02
-4.26370263e-01 -3.90442200e-02 -5.05169988e-01 7.01849401e-01
5.94243407e-02 -2.48814207e-02 6.03464782e-01 5.05991042e-01
5.11690676e-01 -2.75778234e-01 -5.12001157e-01 1.25835907e+00
2.16461316e-01 -1.05760574e+00 -2.06447750e-01 2.20209807e-01
1.05042493e+00 -3.16657387e-02 1.61422268e-01 1.85944070e-03
-6.47755265e-02 -8.88647437e-01 6.15825117e-01 3.04688841e-01
9.43684161e-01 -6.94043756e-01 6.63763106e-01 4.93860692e-01
-1.43318987e+00 -2.89243072e-01 -4.94914681e-01 -1.05553828e-01
2.25796744e-01 1.42232621e+00 5.31797260e-02 1.01820481e+00
5.99314153e-01 9.44698513e-01 -2.46890932e-02 1.26729119e+00
-5.95712900e-01 5.31242549e-01 -3.74082863e-01 4.43340987e-01
2.47497484e-01 -9.71027553e-01 7.00797081e-01 9.15157020e-01
8.19048464e-01 4.24752563e-01 5.73091567e-01 8.17312241e-01
-1.44151941e-01 5.31361699e-01 -5.21878839e-01 1.97014734e-01
1.97596669e-01 1.28065085e+00 -4.90353912e-01 -4.56920117e-02
-5.17398298e-01 6.72544003e-01 -9.04755760e-03 6.63305700e-01
-5.61185300e-01 -2.43341342e-01 6.85085833e-01 1.23115875e-01
2.98727155e-01 -1.26411811e-01 -4.19275314e-01 -1.43583512e+00
3.01757753e-01 -9.52680409e-01 4.73229170e-01 -5.28333127e-01
-1.28993523e+00 4.39603657e-01 -3.38469535e-01 -1.52145267e+00
2.53815204e-01 -6.04392767e-01 -6.71193004e-01 8.71537030e-01
-2.06684446e+00 -7.72180319e-01 -6.85341001e-01 5.65091074e-01
2.82722175e-01 8.67325589e-02 6.16332531e-01 4.91919428e-01
-9.49984014e-01 3.17340910e-01 5.28088510e-01 -2.07157820e-01
4.56755131e-01 -8.79831374e-01 -6.53758764e-01 1.23448479e+00
-8.58186856e-02 3.53398412e-01 5.67116857e-01 -4.73374873e-01
-1.27359343e+00 -1.04653287e+00 3.51213753e-01 4.40638453e-01
7.91234851e-01 1.14887759e-01 -1.09423327e+00 3.00091684e-01
-1.98727213e-02 2.77193159e-01 7.42412508e-01 -2.33491629e-01
-2.79233456e-01 -3.95289391e-01 -1.08062851e+00 3.92696738e-01
8.72082472e-01 -3.17108750e-01 4.21869196e-02 3.69633436e-01
2.77872503e-01 -3.60002160e-01 -9.12082851e-01 8.64816070e-01
2.92723864e-01 -1.20200419e+00 8.12243700e-01 1.35190263e-01
3.31085354e-01 -7.15008497e-01 -3.52347881e-01 -1.36630309e+00
-4.94206548e-01 -6.54274702e-01 -2.01202184e-01 9.45399940e-01
2.70464987e-01 -8.69339585e-01 6.86669052e-01 1.84363544e-01
-2.59631902e-01 -8.34495962e-01 -9.31947827e-01 -7.40977883e-01
-2.10393846e-01 -3.17030728e-01 4.07127321e-01 1.11292064e+00
-2.82186400e-02 6.11664318e-02 -5.24810791e-01 6.73105955e-01
9.85995412e-01 4.09517020e-01 5.06380618e-01 -1.15099049e+00
-4.34022874e-01 -6.33559108e-01 -6.30127862e-02 -1.13809609e+00
-1.48543082e-02 -6.47221625e-01 1.70279428e-01 -1.37927604e+00
2.23175883e-01 -6.31440163e-01 -4.97917145e-01 3.61755222e-01
-3.95116538e-01 3.15115958e-01 1.61567047e-01 4.10737485e-01
-1.39289856e-01 7.89013386e-01 1.42654705e+00 -3.41442138e-01
-2.97862589e-01 -1.23190526e-02 -7.42661119e-01 7.09524989e-01
6.21052146e-01 -1.84928760e-01 -2.96585053e-01 -4.21397746e-01
2.71886736e-01 -4.08300012e-02 1.53204396e-01 -1.04316843e+00
1.22155458e-01 -3.89313012e-01 1.43315569e-01 -4.08299029e-01
2.61637956e-01 -9.38212931e-01 9.35259685e-02 2.17957199e-01
2.33362183e-01 -6.05356455e-01 -1.80189192e-01 5.13236761e-01
-5.30574203e-01 -2.89085656e-01 1.09365869e+00 -2.20577851e-01
-6.38952315e-01 5.79588056e-01 3.62651646e-02 -1.48045465e-01
1.01029313e+00 -1.53147474e-01 -1.40054867e-01 -5.19207895e-01
-4.54896569e-01 3.33958894e-01 6.11748397e-01 -3.49745572e-01
4.23042595e-01 -1.16210735e+00 -8.36485147e-01 4.46468681e-01
2.46442199e-01 2.91794598e-01 4.88959163e-01 1.16880047e+00
-3.31957817e-01 -2.87909433e-02 1.03199132e-01 -6.40930831e-01
-6.58363521e-01 3.02195966e-01 3.36391330e-01 -3.43359590e-01
-5.36630273e-01 8.23898852e-01 3.81280124e-01 -1.75883114e-01
1.68933272e-01 -1.73462316e-01 -1.08847931e-01 1.58344239e-01
7.65119314e-01 5.02373397e-01 2.05182955e-02 -7.10587084e-01
4.01697233e-02 7.75264144e-01 4.11860287e-01 1.25926197e-01
1.38575411e+00 -1.20617427e-01 -5.03104866e-01 4.06237505e-02
1.16525257e+00 2.59189606e-01 -1.46153772e+00 -3.12398940e-01
-4.70289052e-01 -7.59344697e-01 2.73487121e-01 -3.66155595e-01
-1.35368979e+00 7.64107585e-01 5.06588101e-01 2.90748864e-01
1.51851237e+00 -4.27292526e-01 7.76744306e-01 2.34214477e-02
3.48875642e-01 -1.13071585e+00 -4.16071638e-02 3.68755996e-01
7.28555501e-01 -1.30360758e+00 3.05919588e-01 -9.14756536e-01
-2.41091996e-01 9.35475767e-01 4.32646155e-01 9.77344140e-02
5.97818553e-01 -2.43992433e-02 6.99598566e-02 -1.44567460e-01
1.25208810e-01 -4.27597880e-01 2.94759482e-01 3.47396433e-01
2.98600256e-01 2.03429967e-01 -4.40244526e-01 3.02170813e-01
1.30610585e-01 -1.38213232e-01 2.78448999e-01 5.67480445e-01
-5.20544052e-01 -9.64167774e-01 -7.27954566e-01 5.27965844e-01
-2.33101889e-01 -1.75333187e-01 2.88615823e-01 5.19745409e-01
3.18774223e-01 9.80182230e-01 -3.68664056e-01 -5.32307997e-02
2.99709290e-01 -1.33534357e-01 2.76532173e-01 -4.52634305e-01
-7.30304942e-02 5.64014494e-01 -4.33359742e-01 -4.49740827e-01
-5.44515789e-01 -6.29975319e-01 -1.10911596e+00 -2.04066653e-02
-5.61719716e-01 2.32794762e-01 4.30370659e-01 8.40791404e-01
4.43969704e-02 3.29005212e-01 9.21922624e-01 -9.43850636e-01
-6.33093834e-01 -7.90823817e-01 -1.25233030e+00 4.24571663e-01
2.01884910e-01 -6.12640500e-01 -7.06367970e-01 -1.33918926e-01] | [10.29332160949707, -2.0657832622528076] |
7822d5eb-f1f7-458e-9fff-6cd73ae189a8 | albert-a-lite-bert-for-self-supervised | 1909.11942 | null | https://arxiv.org/abs/1909.11942v6 | https://arxiv.org/pdf/1909.11942v6.pdf | ALBERT: A Lite BERT for Self-supervised Learning of Language Representations | Increasing model size when pretraining natural language representations often results in improved performance on downstream tasks. However, at some point further model increases become harder due to GPU/TPU memory limitations and longer training times. To address these problems, we present two parameter-reduction techniques to lower memory consumption and increase the training speed of BERT. Comprehensive empirical evidence shows that our proposed methods lead to models that scale much better compared to the original BERT. We also use a self-supervised loss that focuses on modeling inter-sentence coherence, and show it consistently helps downstream tasks with multi-sentence inputs. As a result, our best model establishes new state-of-the-art results on the GLUE, RACE, and \squad benchmarks while having fewer parameters compared to BERT-large. The code and the pretrained models are available at https://github.com/google-research/ALBERT. | ['Zhenzhong Lan', 'Sebastian Goodman', 'Mingda Chen', 'Kevin Gimpel', 'Radu Soricut', 'Piyush Sharma'] | 2019-09-26 | null | https://openreview.net/forum?id=H1eA7AEtvS | https://openreview.net/pdf?id=H1eA7AEtvS | iclr-2020-1 | ['multi-task-language-understanding', 'multimodal-intent-recognition', 'linguistic-acceptability'] | ['methodology', 'miscellaneous', 'natural-language-processing'] | [-0.1512984 -0.11824317 -0.2719647 -0.58973986 -1.1566037 -0.3328084
0.4144712 0.24489816 -0.4869962 0.76392305 0.58904785 -0.40296945
0.0963416 -0.6527961 -0.6141716 -0.49708918 -0.19847432 0.4986565
0.07836869 -0.49946976 0.39386997 0.04238322 -1.1027365 0.6366613
0.83609146 0.5944886 0.17971106 0.8183478 0.00779242 0.9579501
-0.53410923 -0.5122891 0.12627533 -0.29789907 -1.0542109 -0.40997323
0.44296348 -0.2242881 -0.35676152 0.6729175 0.66982645 0.3724521
0.33829463 -0.9317845 -0.6618224 1.113521 -0.54975045 0.6895684
-0.01117727 0.24714291 1.3414512 -0.86217225 0.3391482 1.2027242
1.0147597 0.5566154 -1.5250257 -0.6363026 0.40607092 0.3007062
-1.2304587 -0.6609275 0.36379334 -0.02130398 1.6091765 0.12924398
0.45930988 1.340435 0.42073524 0.8394923 0.97204506 -0.37778154
-0.08373049 -0.2584782 0.78990006 0.66037077 0.13627598 0.07632883
-0.8085156 -0.05115192 0.48042557 -0.15552816 -0.15237786 0.26467538
-0.9329784 0.8589166 0.6054048 0.44958723 -0.09001082 0.6050275
0.7781783 0.52706695 0.88558525 0.590717 -0.6830705 -0.66632915
-1.0960736 0.1527764 0.8586634 0.9221437 0.6374497 -0.09081911
-0.30659482 0.914765 0.03330167 0.07709912 0.62278056 -1.1145583
0.91338414 0.43889147 -0.37930182 -0.6400542 -0.5158648 -0.79499984
-0.89761454 -0.17501599 0.44458717 -0.16341826 -0.671753 1.8727545
-0.23484345 0.2870967 0.09866513 0.7051082 0.8472667 0.8017568
0.17262241 0.02998592 1.102937 -1.6434433 -0.5380478 -0.69128156
1.2929463 -0.7411583 1.3880715 0.32714298 -1.2968106 -0.4989541
-0.9191888 -0.56158465 -0.14285521 0.24401964 0.7614032 0.3011024
-1.3405235 0.93487954 -1.0768163 -0.29543862 0.4518245 0.379983
-0.05854537 -0.03265387 -1.1655627 1.0205669 0.18232216 0.03801287
-0.65699434 -0.87318254 -0.608739 0.3651291 -0.03999531 -0.80840707
1.5297537 -0.7938307 -1.5228677 0.80702436 -0.34516492 -0.8058225
0.34141198 -0.8450398 0.03671631 -0.06471632 0.05395522 0.7691915
0.40430853 -0.7666337 -0.1570248 -0.13348483 0.1715839 0.27702957
-0.4932829 0.0231565 -0.5043382 -0.52525014 -0.22212408 -0.79536456
-0.4810182 -0.38728392 -0.32669264 -0.3215033 0.31609935 -0.5435037
1.4849156 -2.1515918 0.14843 -0.28298467 0.17773601 0.2915479
-0.45144436 0.71012276 -0.1283022 0.29147306 -0.15683177 -0.91666085
-0.37645826 0.19566186 -0.30915985 0.24012794 0.2584802 0.99371254
-0.85805017 -0.5048696 -0.10920751 0.20732172 -0.8972153 0.20563976
-0.09077848 0.38815576 -0.25005952 0.1318033 0.36010692 -0.61778873
0.04435508 0.16006109 0.04167491 0.99915665 -0.47795483 2.096258
-0.8556745 0.80765593 -0.04192162 -0.9522539 0.765678 0.13823232
0.13413385 -0.6706258 0.14591695 0.05182409 0.15830643 -0.31611672
0.5757667 0.11400971 0.0313205 0.5009519 0.0586872 -0.14757173
0.4421214 0.38668752 1.3340584 -0.06366038 0.09062251 -0.4045233
0.23910719 0.11534392 0.5096511 0.68573207 -0.21175757 0.48738772
0.6725305 -0.3905656 -0.9886306 -0.9259293 -0.01635232 1.3774161
-0.12379856 -0.8662505 -0.6600454 -0.49805778 -0.07805489 0.96104836
-0.50190425 -0.4569826 -0.8431492 -0.79999393 0.7870896 0.89603716
0.4669285 -0.8529502 -0.32844654 0.260041 -0.30926767 -0.93690795
-0.5671465 0.29732358 -1.3878524 -0.6854861 -0.46746704 -0.7448433
0.5489246 0.3178002 1.6234094 0.3159616 -0.07992528 -0.20015316
-0.518286 -0.05001807 -0.34419462 0.6238102 -0.22824472 -0.6803509
0.29514167 -0.65749955 -0.6279416 0.03148662 -0.45764813 0.30868992
0.57503474 1.1180506 0.17437665 -0.39630547 0.58947545 -1.048182
0.9817272 -0.55303365 -0.16842401 0.03433494 -0.9067366 0.20687613
0.8274977 -0.3041089 -0.967758 -0.4445148 -0.34534875 -0.20740856
0.01709078 0.6540901 0.55732805 0.46469763 1.0337586 -0.10037208
-0.02757321 -0.68709517 0.5330314 0.4753755 0.09997794 -0.65017354
0.29498002 0.2001155 -0.26215896 -0.42073965 -1.2609571 -0.56781477
-0.582303 0.20399971 0.5634564 -1.1753018 -0.3994947 0.2688012
-1.2853701 -0.8527511 -0.27095032 0.36703828 -0.47583646 0.20728703
-1.2140833 -0.48605597 -0.8885721 -0.9882542 1.1741173 0.02055802
-0.5186911 -1.1190653 0.21687678 0.5220646 0.50928277 -0.29886734
0.84628254 -0.46954376 -0.29820696 0.1014711 -0.25547844 0.50851065
-0.16332774 -0.06537703 -0.82184255 -0.45887417 -0.18861087 -0.6826246
1.3625755 0.33747926 1.2602335 -0.3533613 -0.24879472 0.8002693
1.2780272 -0.38758498 0.4001138 0.31966412 0.48267126 0.47791302
0.61813265 0.4120471 0.3623514 0.6680471 0.14872845 0.01984198
-0.30760586 -0.13860731 0.66566914 1.2280037 0.03066569 -0.368707
-1.1767386 0.4914497 -2.0136182 -0.9135738 -0.4861664 1.6840986
1.1786596 0.29826578 -0.02947224 -0.04031354 0.36359918 0.20209399
-0.36667177 -0.90541077 0.04899255 0.47594184 0.4951921 0.69382447
-0.9308527 1.446927 6.7801332 0.9011176 -1.0671499 0.4196042
1.041647 -0.55894303 -0.1293124 0.03346762 -1.0621082 0.35595283
1.4885892 -0.34971562 0.48697907 0.8275378 0.46544144 -0.05264209
-1.3171837 0.8174368 -0.13319623 -1.5269417 -0.1481778 -0.2416033
0.8163615 0.5915167 -0.10642359 0.67648274 0.5257991 -1.0910573
0.5760059 0.34278667 0.6407821 -0.7039421 0.89833105 0.46468002
-0.98471993 -0.17308421 -0.50876987 -0.52262115 0.10923102 0.80733263
-0.6981719 0.23841298 0.7156673 0.9384922 -0.82987994 0.859949
-0.4743118 1.052073 -0.26182285 -0.18757382 0.3825857 -0.05089922
0.23303708 1.4847721 0.06839909 -0.06019684 0.07093322 0.90531385
-0.30361563 0.08147099 -0.6218192 -0.05852802 0.5778602 1.0514948
-0.39759952 -0.4389702 -0.30287397 1.0357181 1.0754939 0.29954135
-0.9403919 -0.03728566 0.8080385 -0.09333731 -0.08361173 -0.67164296
-0.8157392 -1.156709 -0.12890647 -0.8200675 0.3692867 -0.46805036
-1.4763381 0.6083555 -0.15221888 -0.89856327 -0.3612076 -0.4040229
-0.9102012 0.70687056 -1.4540257 -0.8621187 -0.17486952 0.32370448
0.9174674 0.08876765 0.9716625 0.29951265 -0.75317305 0.86581147
0.24583884 0.20946537 0.74715984 -1.1195779 0.9032945 0.84835815
0.14375643 0.81103104 0.42593023 -0.3960302 -1.1113856 -1.0620548
1.2135992 -0.4180396 0.9568541 -0.48522028 -0.7396403 0.86975014
0.41316742 -0.11795077 0.7371705 0.72624636 -0.29258174 -0.07151873
-0.5690455 0.6842915 1.3313397 -0.43706122 -0.6096521 0.60788727
0.94716424 -0.42363244 -0.74788356 0.22341363 0.12021884 -1.0635302
0.87425995 -0.7578731 0.93793863 0.42655283 0.0994322 -1.511204
-0.55572546 -0.6068593 -0.1291952 1.1668832 0.7471848 -0.61745983
0.8484212 0.5832959 -0.46116874 -1.2161417 -1.0195407 -0.963888
0.5503802 -0.40075997 0.15916738 0.7842416 0.18230665 0.7238276
-0.19463585 -0.24039966 0.27371466 0.23003533 0.6168035 -0.78448594
-0.5220216 -0.7125981 0.02753801 -1.3772836 0.29790354 -1.1758729
0.04085881 -1.5163285 0.20550326 -0.6409345 -0.24225968 0.67833734
-0.43932125 0.1446088 0.23733726 0.37656516 -0.46940768 0.7050468
1.1580329 -0.07284696 -0.24090566 -0.14946622 -0.762697 0.52993244
1.3554369 -0.56829476 -0.3857781 -1.1735768 0.3036854 -0.17933144
0.13699004 -1.032706 0.2005677 0.18037717 -0.01853295 -0.27696198
0.5059238 -0.2525659 -0.48531762 0.5348224 -0.8506049 0.40378872
0.3281251 0.4068086 -0.21693811 -0.24212366 0.8865682 -0.13546495
-0.39044046 0.05576682 -0.37331277 0.34646928 0.70173806 0.24101044
-0.6341946 -0.42616135 -0.63334876 0.4956173 0.1705804 0.40268996
0.45971337 -1.0452485 -0.9149597 0.00956383 -0.18312609 0.06272625
0.08727044 1.119255 -0.79429376 0.38526362 0.12766095 -0.5239559
-1.3146118 0.23714896 0.22958542 -0.830675 -0.72709805 1.5591162
-0.05216001 -0.35356385 0.2676466 -0.6224415 -0.01578615 -0.0348029
0.33979937 0.4830583 0.19339813 -0.1605465 -0.34045118 0.2176241
-0.34838623 -0.02616931 1.5293646 0.05654691 -0.07715803 0.44864005
1.3937267 -0.21147923 -1.1052884 -0.25740072 0.14132519 -0.35117936
0.19449003 -0.7216827 -0.95481735 1.1684505 0.31611267 -0.08305684
1.0328872 -0.02930101 1.0346212 0.6759088 0.20934124 -1.169772
0.30468258 1.0977913 0.9421824 -1.1637133 0.02336543 -0.38812563
-0.6129457 1.0835813 0.8314005 -0.5373155 0.6413853 0.5976138
0.09205104 -0.05836774 -1.4021075 -0.0198368 -0.03802437 0.09499411
0.9865981 -0.0181218 -0.58499336 0.54468185 -0.6037879 -0.05862571
0.34864783 0.76310146 -0.30125284 -1.275868 0.08529104 0.5443325
-0.48509827 -0.76775056 -0.4338902 0.60721004 -0.28763717 1.240473
0.1870775 -0.43206084 0.10142715 0.20381679 0.33254844 -0.92720217
-1.1269349 -0.2301069 0.5783629 -0.8792595 -0.11678832 -0.390776
-1.4704849 -0.77342486 -0.26206326 0.1833444 0.34024087 0.73727125
0.73180723 0.67340714 0.42473635 -0.8972859 -0.94291186 -1.274084
-0.12217676 0.40560755 0.11225888 -0.2522647 -0.4344934 -0.09811296] | [11.151619911193848, 8.559986114501953] |
8f689fa9-1922-4e3c-93f2-b05f8829717b | self-calibrating-anomaly-and-change-detection | 2209.02379 | null | https://arxiv.org/abs/2209.02379v1 | https://arxiv.org/pdf/2209.02379v1.pdf | Self-Calibrating Anomaly and Change Detection for Autonomous Inspection Robots | Automatic detection of visual anomalies and changes in the environment has been a topic of recurrent attention in the fields of machine learning and computer vision over the past decades. A visual anomaly or change detection algorithm identifies regions of an image that differ from a reference image or dataset. The majority of existing approaches focus on anomaly or fault detection in a specific class of images or environments, while general purpose visual anomaly detection algorithms are more scarce in the literature. In this paper, we propose a comprehensive deep learning framework for detecting anomalies and changes in a priori unknown environments after a reference dataset is gathered, and without need for retraining the model. We use the SuperPoint and SuperGlue feature extraction and matching methods to detect anomalies based on reference images taken from a similar location and with partial overlapping of the field of view. We also introduce a self-calibrating method for the proposed model in order to address the problem of sensitivity to feature matching thresholds and environmental conditions. To evaluate the proposed framework, we have used a ground robot system for the purpose of reference and query data collection. We show that high accuracy can be obtained using the proposed method. We also show that the calibration process enhances changes and foreign object detection performance | ['Tomi Westerlund', 'Jorge Peña Queralta', 'Sahar Salimpour'] | 2022-08-26 | null | null | null | null | ['fault-detection'] | ['miscellaneous'] | [ 3.22085559e-01 -2.18818232e-01 5.54070652e-01 -3.61454576e-01
-2.73731858e-01 -4.07945603e-01 6.57603383e-01 5.36391497e-01
-2.65423089e-01 3.15925747e-01 -3.72711301e-01 8.63890499e-02
-5.80186769e-02 -6.09831870e-01 -1.06512046e+00 -6.59758389e-01
-1.32268086e-01 9.78223607e-02 7.56738544e-01 -2.80560076e-01
4.75737244e-01 1.01174426e+00 -2.13577056e+00 -5.14770038e-02
5.71503043e-01 1.03247535e+00 2.00366601e-01 5.64778924e-01
1.29979163e-01 3.74836802e-01 -8.17723215e-01 1.40615851e-01
5.96003175e-01 -2.01177858e-02 -5.66634893e-01 5.62795103e-01
8.05695117e-01 -4.42981720e-01 -3.86727482e-01 1.12208891e+00
4.25994605e-01 8.76016766e-02 6.40214980e-01 -1.42907274e+00
-4.19788092e-01 -1.03646919e-01 -6.84877396e-01 7.77538240e-01
2.14535221e-01 2.35878125e-01 4.21137363e-01 -8.34698915e-01
5.02615511e-01 9.52302575e-01 5.53064585e-01 5.13475463e-02
-9.03662980e-01 -3.45784456e-01 4.61955070e-01 6.90420926e-01
-1.27952623e+00 -3.46234232e-01 1.08267593e+00 -7.37138629e-01
9.82378364e-01 -8.57879221e-02 4.19136375e-01 8.75170767e-01
5.07251501e-01 3.98498029e-01 6.06974781e-01 -5.92051446e-01
3.13997477e-01 -5.03984913e-02 9.24862400e-02 5.58333337e-01
6.38567626e-01 1.50558695e-01 -3.62976670e-01 6.37061447e-02
4.73091006e-01 3.84361386e-01 -2.91478425e-01 -7.75740683e-01
-1.07456911e+00 5.91719329e-01 4.96944219e-01 5.49842834e-01
-5.86351156e-01 1.47515433e-02 3.80982518e-01 4.83183891e-01
2.67040193e-01 3.71546924e-01 -2.81258106e-01 2.81465888e-01
-7.38641858e-01 1.30746543e-01 1.45828724e-01 6.51685178e-01
8.40830147e-01 4.87715721e-01 1.67081803e-01 6.92348540e-01
2.72508532e-01 6.75556779e-01 6.08189583e-01 -4.89185035e-01
6.30524829e-02 7.46799827e-01 3.36949497e-01 -1.53966093e+00
-5.36663175e-01 -4.35723096e-01 -5.93918085e-01 5.36219299e-01
2.39200726e-01 1.80029720e-01 -1.06232953e+00 1.30652618e+00
4.72276866e-01 2.18988076e-01 1.99792147e-01 7.88674057e-01
5.01973450e-01 4.83969867e-01 -4.98832375e-01 -1.08217210e-01
7.96101034e-01 -5.65253437e-01 -5.66600680e-01 -5.03228307e-01
6.18346751e-01 -5.69026589e-01 8.41058135e-01 4.85228598e-01
-4.86274630e-01 -6.09290361e-01 -1.36448848e+00 3.96422684e-01
-6.53465211e-01 -6.59170328e-03 4.32211608e-02 2.04743430e-01
-9.19139683e-01 3.84674013e-01 -8.36614609e-01 -8.44388127e-01
-2.38531432e-03 1.19115144e-01 -6.81665182e-01 -1.20442852e-01
-8.12891901e-01 9.12089646e-01 4.23873782e-01 3.31928581e-01
-1.12336421e+00 -3.28907400e-01 -9.80738699e-01 -1.82691157e-01
2.74710029e-01 -8.96839891e-03 9.31013227e-01 -1.20720339e+00
-8.42596769e-01 9.44550216e-01 -5.47202416e-02 -6.51602805e-01
3.92099470e-01 -4.52041447e-01 -8.68662953e-01 1.02438070e-01
1.38520211e-01 1.57748550e-01 1.02653730e+00 -1.35199547e+00
-9.03724670e-01 -6.42927051e-01 -2.25034207e-01 -1.03477672e-01
1.07158218e-02 -1.62371770e-01 -2.38299295e-01 -5.44139504e-01
6.12264931e-01 -6.41385257e-01 -2.45011803e-02 -1.27301723e-01
-2.59304732e-01 1.70124322e-01 1.25288045e+00 -7.27630377e-01
8.63130808e-01 -2.20721889e+00 -3.10879141e-01 4.77967799e-01
-2.01559618e-01 1.91853091e-01 -5.69325266e-03 3.43316853e-01
-3.14085901e-01 -3.88447821e-01 -4.63028818e-01 2.19792843e-01
-4.04020637e-01 3.24984372e-01 -4.32934076e-01 9.19690967e-01
2.74399698e-01 1.69047758e-01 -5.78408599e-01 -4.10455465e-02
4.83181864e-01 7.03357831e-02 -2.92080253e-01 3.49568635e-01
-4.67044860e-02 5.44862330e-01 -1.30957320e-01 8.40015054e-01
6.38384640e-01 2.82335401e-01 -3.99342209e-01 -1.96392596e-01
-1.76557094e-01 -2.67738312e-01 -1.46958554e+00 1.39270771e+00
-1.35084823e-01 8.33941877e-01 -1.91450983e-01 -1.29399240e+00
1.25594676e+00 5.21273166e-02 5.23458898e-01 -1.02972758e+00
8.99060164e-03 2.88491189e-01 1.13464817e-01 -8.60975504e-01
6.56666458e-01 4.01131064e-01 1.21267453e-01 1.84472248e-01
-6.19793907e-02 6.86936453e-02 1.49191275e-01 -1.15545310e-01
1.26829576e+00 -4.26954851e-02 3.41550142e-01 1.97343137e-02
8.30637813e-01 1.28831670e-01 4.91957247e-01 8.96415472e-01
-4.06820148e-01 6.47474289e-01 1.12929054e-01 -8.03651273e-01
-1.19444084e+00 -9.80892777e-01 -2.26422980e-01 7.60550499e-01
5.36623180e-01 3.23519677e-01 -3.20026100e-01 -5.21085978e-01
2.02777877e-01 6.22279346e-01 -5.43592632e-01 -3.44021618e-01
-5.46612620e-01 -7.39971995e-01 2.55748004e-01 3.49232942e-01
6.82888031e-01 -1.13571417e+00 -1.12248695e+00 7.01946393e-02
4.42262366e-02 -9.57256854e-01 3.53400111e-01 2.37112820e-01
-7.56429255e-01 -1.27456999e+00 -2.74154156e-01 -7.31648743e-01
8.29150856e-01 2.94954360e-01 8.90766561e-01 -4.56488915e-02
-5.59772491e-01 8.81793320e-01 -4.57081854e-01 -5.38187146e-01
-3.87206674e-01 -3.33944261e-01 2.62061208e-01 4.08765525e-01
3.25157046e-01 -5.03639460e-01 -4.48287815e-01 4.45795506e-01
-1.05160010e+00 -6.15790606e-01 4.46506709e-01 7.01156497e-01
5.71452796e-01 7.98869431e-02 3.41858983e-01 -4.61514294e-01
1.85915858e-01 -5.90440094e-01 -9.16767299e-01 9.29804146e-02
-5.79297364e-01 2.98711695e-02 1.98617443e-01 -1.55105501e-01
-8.60439420e-01 2.11678788e-01 7.68668652e-02 -5.18286765e-01
-7.62718320e-01 2.32431948e-01 -2.30535552e-01 -1.28407300e-01
8.90410483e-01 2.43284792e-01 -1.05487324e-01 -3.91976178e-01
1.00850694e-01 6.94121003e-01 9.87951279e-01 -1.50829226e-01
1.01902282e+00 7.44995117e-01 -8.00623223e-02 -9.45960343e-01
-3.99626672e-01 -7.88296163e-01 -8.32306862e-01 -5.37141681e-01
6.28648996e-01 -7.94189095e-01 -1.87917743e-02 8.90534103e-01
-1.05852318e+00 -6.35057613e-02 -2.19390497e-01 3.45126957e-01
-3.30926567e-01 5.70551574e-01 4.32979576e-02 -8.29933286e-01
-1.63301706e-01 -9.26318884e-01 1.09882784e+00 1.71499789e-01
1.98714063e-01 -8.17875803e-01 2.08760202e-01 -2.71435142e-01
3.58258426e-01 7.21804798e-01 6.05980814e-01 -8.28895450e-01
-5.53067207e-01 -5.19978583e-01 -2.03031837e-03 3.68527085e-01
4.57574964e-01 8.53401981e-03 -1.12005639e+00 -4.69827712e-01
1.92413077e-01 1.37573943e-01 7.35192478e-01 1.85710236e-01
9.50289786e-01 -2.12274864e-02 -4.69766438e-01 5.66364944e-01
1.60028684e+00 5.40653110e-01 5.00346541e-01 1.04036450e+00
6.34302020e-01 5.48758507e-01 8.75773609e-01 5.71930051e-01
-9.07684397e-03 6.08387828e-01 1.06907344e+00 -1.21215381e-01
3.49058092e-01 2.07050532e-01 3.13111573e-01 1.17443077e-01
2.82731801e-01 -9.29297060e-02 -1.29923487e+00 1.17362440e+00
-1.79587173e+00 -9.70322669e-01 -7.89203644e-02 2.24739742e+00
-8.44280273e-02 1.14828221e-01 -1.66294381e-01 4.17505533e-01
1.03620267e+00 1.46405756e-01 -8.66647899e-01 -2.07121134e-01
-2.66289055e-01 -3.68389964e-01 6.75743282e-01 3.22436064e-01
-1.34207082e+00 6.55455768e-01 6.07336378e+00 1.03309862e-02
-1.44179463e+00 -1.19582511e-01 2.14100033e-01 1.97426707e-01
1.93779320e-01 -3.18241537e-01 -3.94694656e-01 3.25347781e-01
5.57855964e-01 1.57144517e-01 7.01531302e-03 1.00192773e+00
9.11089107e-02 -3.77329022e-01 -1.12970173e+00 1.17112279e+00
5.74901044e-01 -7.84953535e-01 -1.23814374e-01 -1.90636143e-01
5.94106615e-01 2.35128790e-01 -4.14512865e-02 -4.59287502e-02
-2.65950169e-02 -6.65894806e-01 7.14112520e-01 7.98447251e-01
1.82455152e-01 -6.56200051e-01 1.00329530e+00 1.55582339e-01
-8.63251388e-01 -4.03197736e-01 -4.28224444e-01 -8.87240320e-02
-6.16484322e-02 4.42806900e-01 -1.00406480e+00 5.72347283e-01
1.19835579e+00 7.16162860e-01 -1.16598809e+00 1.41846693e+00
8.33618939e-02 4.58123326e-01 -4.47634071e-01 4.96926308e-01
1.43986121e-01 -9.19260010e-02 8.81411433e-01 8.23345721e-01
6.73769593e-01 -4.04129356e-01 1.32985815e-01 5.81413507e-01
4.46158677e-01 -1.23264775e-01 -1.19672620e+00 4.75056589e-01
2.76106983e-01 1.00643790e+00 -7.13683009e-01 -6.50973618e-02
-5.02176523e-01 1.01811481e+00 -2.98627857e-02 4.57985878e-01
-6.06688201e-01 -4.04785991e-01 6.06611490e-01 1.84278458e-01
5.30547321e-01 -2.18339160e-01 -3.36815324e-03 -9.69583273e-01
3.72341543e-01 -6.36819482e-01 5.47382712e-01 -9.56074536e-01
-1.17692077e+00 5.69095969e-01 1.18615150e-01 -1.45503306e+00
-4.45774406e-01 -6.52948260e-01 -6.24957919e-01 5.16310513e-01
-1.45162916e+00 -8.67552817e-01 -8.61900747e-01 7.46632755e-01
4.93784428e-01 -5.35643995e-01 2.98757732e-01 1.51748627e-01
-6.14948034e-01 3.33569795e-01 3.86449188e-01 1.37732849e-01
8.07147205e-01 -1.11367309e+00 4.15810198e-01 1.74541616e+00
5.95504045e-02 4.87818085e-02 1.07769060e+00 -6.94970429e-01
-9.62925076e-01 -1.38215363e+00 1.82174057e-01 -4.23366457e-01
4.59018648e-01 -1.22360341e-01 -1.33869064e+00 8.33872497e-01
1.75048411e-03 2.93502510e-01 7.10331723e-02 -4.18065190e-01
-1.35349482e-01 -4.79790837e-01 -1.36754143e+00 3.21446419e-01
7.29917765e-01 -3.64460200e-01 -7.66767383e-01 1.25063688e-01
4.82001573e-01 -5.18960059e-01 -3.41177404e-01 6.20422721e-01
1.10498063e-01 -1.23367941e+00 6.81066155e-01 -2.46106565e-01
-1.65921435e-01 -8.61625910e-01 -4.21959817e-01 -1.21997046e+00
-2.36272886e-01 9.67993343e-04 4.90099676e-02 1.28877091e+00
-1.46543486e-02 -7.44929612e-01 4.76000518e-01 1.64356843e-01
-4.70199764e-01 -2.11422481e-02 -8.46423507e-01 -8.17286193e-01
-3.46341819e-01 -5.55639684e-01 4.71377224e-01 7.50121653e-01
-5.73854327e-01 -1.15280598e-01 -2.06972823e-01 1.04513395e+00
4.47129726e-01 9.95788425e-02 1.04499805e+00 -1.40808713e+00
7.77406842e-02 -1.72329741e-03 -1.21938384e+00 -2.90782154e-01
-5.39621934e-02 -5.32079101e-01 3.58780116e-01 -1.48756635e+00
-2.85656869e-01 -1.18236989e-01 -3.97586942e-01 4.05603290e-01
9.79607329e-02 2.31322289e-01 -2.74164885e-01 2.53974289e-01
-4.74712998e-01 5.21879971e-01 3.31519753e-01 -3.46649975e-01
-1.20788455e-01 -1.65181100e-01 5.67366835e-03 8.57291043e-01
8.43345702e-01 -3.67768884e-01 -2.83716083e-01 -5.55916786e-01
-1.10098338e-02 -5.04754424e-01 6.31069422e-01 -1.59633327e+00
6.28879815e-02 3.99921909e-02 5.92260599e-01 -7.97122419e-01
-1.05083115e-01 -1.24240685e+00 1.18161835e-01 4.36176121e-01
5.76876663e-02 6.61631346e-01 3.90272915e-01 8.36103439e-01
-4.58993167e-01 -3.95343810e-01 8.45611811e-01 1.17698507e-02
-1.51312435e+00 1.28688097e-01 -5.52637279e-01 -2.73834765e-01
1.36807179e+00 -4.13692683e-01 -2.94763505e-01 -1.40262157e-01
-4.91356373e-01 9.48522314e-02 7.94719100e-01 8.07440162e-01
8.42810571e-01 -1.20283210e+00 -4.72170562e-01 5.52541375e-01
7.73223519e-01 9.86343473e-02 3.19481939e-01 5.67057312e-01
-9.04998660e-01 1.22746803e-01 -6.28278077e-01 -1.14255869e+00
-1.15342689e+00 8.77291501e-01 7.86413193e-01 3.76391560e-01
-8.26998651e-01 2.66170710e-01 1.51239946e-01 -2.17866316e-01
3.48134339e-01 -4.93048996e-01 -3.95129740e-01 -2.26185337e-01
4.88710970e-01 4.03418154e-01 4.91855979e-01 -9.96014416e-01
-4.07265544e-01 6.56016052e-01 4.66092713e-02 8.57301429e-02
1.14866233e+00 -3.64001960e-01 6.08893149e-02 7.15297163e-01
7.29725838e-01 -1.40151411e-01 -1.21911097e+00 -4.72878784e-01
2.80432254e-01 -6.63241744e-01 1.82125829e-02 -5.40110886e-01
-1.05687296e+00 8.09955359e-01 1.56320083e+00 2.72804946e-01
1.26836944e+00 8.14429577e-03 2.99571037e-01 5.80952466e-01
3.69452864e-01 -1.07767701e+00 2.24269107e-01 4.79833335e-01
1.07358992e+00 -1.49649739e+00 -2.24108785e-01 4.28050198e-02
-4.14168268e-01 1.06397009e+00 9.96577978e-01 -2.92281479e-01
4.36238915e-01 -2.09994726e-02 2.95210928e-01 -3.43012601e-01
-2.89517105e-01 -4.27535832e-01 1.00259900e-01 1.04255366e+00
-1.58953115e-01 -3.95550340e-01 2.76981294e-01 7.31543452e-02
7.58934543e-02 -5.18500209e-01 8.36969554e-01 1.10002315e+00
-8.15758884e-01 -4.50808406e-01 -9.55261111e-01 3.36971104e-01
-3.23564053e-01 3.91553432e-01 -3.17857116e-01 8.40702236e-01
3.95620793e-01 9.03605402e-01 6.96928024e-01 -2.75900632e-01
5.75244606e-01 1.66390106e-01 1.41859144e-01 -2.88731098e-01
-9.90610272e-02 -2.41986826e-01 -3.07337701e-01 -6.80656135e-01
-4.42017496e-01 -7.56236672e-01 -1.07895875e+00 3.08174789e-01
-2.29972333e-01 -1.06405906e-01 6.36460543e-01 9.59198356e-01
3.13268334e-01 5.78916788e-01 5.80870330e-01 -8.50811362e-01
-6.60636798e-02 -9.53835428e-01 -7.14783013e-01 6.39376879e-01
8.96808982e-01 -1.01210248e+00 -5.11208177e-01 2.05816254e-01] | [7.68773078918457, 2.007030487060547] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.