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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
43c83c7e-cad8-4dc4-8e5e-7323e6bebed9 | salient-mask-guided-vision-transformer-for | 2305.07102 | null | https://arxiv.org/abs/2305.07102v1 | https://arxiv.org/pdf/2305.07102v1.pdf | Salient Mask-Guided Vision Transformer for Fine-Grained Classification | Fine-grained visual classification (FGVC) is a challenging computer vision problem, where the task is to automatically recognise objects from subordinate categories. One of its main difficulties is capturing the most discriminative inter-class variances among visually similar classes. Recently, methods with Vision Transformer (ViT) have demonstrated noticeable achievements in FGVC, generally by employing the self-attention mechanism with additional resource-consuming techniques to distinguish potentially discriminative regions while disregarding the rest. However, such approaches may struggle to effectively focus on truly discriminative regions due to only relying on the inherent self-attention mechanism, resulting in the classification token likely aggregating global information from less-important background patches. Moreover, due to the immense lack of the datapoints, classifiers may fail to find the most helpful inter-class distinguishing features, since other unrelated but distinctive background regions may be falsely recognised as being valuable. To this end, we introduce a simple yet effective Salient Mask-Guided Vision Transformer (SM-ViT), where the discriminability of the standard ViT`s attention maps is boosted through salient masking of potentially discriminative foreground regions. Extensive experiments demonstrate that with the standard training procedure our SM-ViT achieves state-of-the-art performance on popular FGVC benchmarks among existing ViT-based approaches while requiring fewer resources and lower input image resolution. | ['Fahad Shahbaz Khan', 'Hisham Cholakkal', 'Aliakbar Abdurahimov', 'Muhammad Hamza Sharif', 'Dmitry Demidov'] | 2023-05-11 | null | null | null | null | ['fine-grained-image-classification'] | ['computer-vision'] | [ 5.54764688e-01 -2.31005296e-01 1.46387801e-01 -3.18058133e-02
-7.57707238e-01 -5.14259696e-01 7.97461331e-01 4.56874333e-02
-2.49845281e-01 6.77442431e-01 1.97501690e-03 -1.62448138e-02
2.82575339e-02 -6.16841257e-01 -6.28933847e-01 -1.09163678e+00
2.29726210e-01 1.27697140e-01 7.97872603e-01 -8.49928632e-02
3.71649563e-01 4.82576400e-01 -1.93410110e+00 5.01732707e-01
9.30365860e-01 1.35613668e+00 5.57487249e-01 5.56453288e-01
-1.49014145e-01 1.00936687e+00 -6.80394650e-01 -3.47874045e-01
2.76772022e-01 -5.17949581e-01 -6.74350083e-01 2.74220377e-01
1.04964101e+00 -1.86218351e-01 -1.21651024e-01 1.28350353e+00
3.52168351e-01 1.92155838e-01 6.46922350e-01 -1.11023271e+00
-6.36152804e-01 5.09024365e-03 -8.01882982e-01 7.35262096e-01
-1.05472721e-01 2.96016932e-01 9.66898561e-01 -1.04858232e+00
2.11563483e-01 1.19709206e+00 3.92355919e-01 4.84454811e-01
-1.40564346e+00 -4.29854780e-01 5.51201761e-01 6.53936625e-01
-1.40604877e+00 -5.22497952e-01 8.53543222e-01 -3.81584048e-01
7.55120814e-01 5.88715434e-01 4.74953651e-01 8.47060740e-01
1.86951771e-01 7.58655846e-01 1.22213233e+00 -2.82188714e-01
2.56094217e-01 1.67819038e-01 1.00066416e-01 6.18394554e-01
1.99639842e-01 1.66195370e-02 -3.34102303e-01 5.91054857e-02
6.44099593e-01 1.95352927e-01 -5.43269992e-01 -5.93229771e-01
-9.93952990e-01 6.93968475e-01 7.38264620e-01 3.81134957e-01
-4.40950125e-01 -2.17732880e-02 2.69364268e-01 1.59232896e-02
5.00201523e-01 2.24378780e-01 -3.71408790e-01 2.19720796e-01
-9.30154383e-01 -5.18212616e-02 1.60619080e-01 6.54993176e-01
9.24291134e-01 1.60858214e-01 -6.73702896e-01 8.75506878e-01
-1.86340027e-02 3.09930831e-01 5.38101435e-01 -6.75525963e-01
2.57544905e-01 7.06400156e-01 -1.03308506e-01 -1.11328173e+00
1.08896896e-01 -8.17673862e-01 -1.10100639e+00 4.32603151e-01
4.63402241e-01 3.07907224e-01 -1.23381090e+00 1.51007378e+00
2.25651309e-01 1.89642414e-01 -1.29051402e-01 1.05429053e+00
9.06002164e-01 4.84345436e-01 1.87870860e-01 -1.87969297e-01
1.39271367e+00 -1.11451733e+00 -2.36714780e-01 -6.23889387e-01
-3.78232412e-02 -6.24393523e-01 1.00630796e+00 1.29234791e-01
-9.48299289e-01 -9.66692805e-01 -7.93767810e-01 3.95451784e-02
-3.62443388e-01 3.58946919e-02 4.73136574e-01 5.52551389e-01
-1.00127244e+00 4.28566992e-01 -4.48557705e-01 -1.10326737e-01
8.35058868e-01 1.47006899e-01 -3.93929899e-01 -2.52285331e-01
-7.20247388e-01 6.83350086e-01 3.07697862e-01 2.32191041e-01
-1.18077469e+00 -6.38830125e-01 -8.11286449e-01 2.83677161e-01
4.76781875e-01 -4.35129344e-01 7.43479133e-01 -1.43200958e+00
-1.09128916e+00 7.92187452e-01 -4.06792581e-01 -4.16214943e-01
5.65505326e-01 -3.69576253e-02 -8.48763585e-02 3.28583777e-01
3.59250426e-01 5.76662540e-01 1.31453073e+00 -1.27037442e+00
-9.20880556e-01 -4.19719845e-01 2.81107314e-02 1.39846474e-01
-3.12145233e-01 -9.34209004e-02 -5.94806254e-01 -7.85456061e-01
-4.23337482e-02 -5.73555112e-01 -2.37721369e-01 8.80803913e-02
-2.95089394e-01 -3.06265086e-01 1.04548693e+00 -5.06226063e-01
8.77000213e-01 -2.34975410e+00 9.29453894e-02 -1.36412248e-01
4.80694741e-01 5.98391473e-01 -2.09151223e-01 -7.32405186e-02
1.91512108e-01 -1.21597089e-01 -1.80500686e-01 -1.06238648e-01
-3.79519969e-01 1.43315181e-01 -3.79166573e-01 5.25667191e-01
7.40202248e-01 9.71194506e-01 -8.78237486e-01 -6.10162973e-01
6.05293393e-01 3.90426725e-01 -1.76876605e-01 1.95107758e-01
-1.17078014e-01 3.09595585e-01 -3.75794113e-01 7.35690355e-01
8.11576724e-01 -3.12262744e-01 -8.32516402e-02 -3.79526258e-01
2.45604459e-02 -1.19342364e-01 -1.03909433e+00 1.07255852e+00
-2.07852319e-01 8.65729392e-01 5.12084104e-02 -1.30951583e+00
7.41701961e-01 -1.58726588e-01 5.88661171e-02 -9.26357329e-01
-6.64255524e-04 7.00541362e-02 4.53325287e-02 -3.41715485e-01
2.65614182e-01 -8.66901316e-03 2.51779586e-01 -9.45800245e-02
6.01509511e-02 1.45328835e-01 1.28139602e-02 8.32315758e-02
9.67611611e-01 4.34010243e-03 3.85879517e-01 -3.20451826e-01
8.41242611e-01 -7.09139109e-02 7.65402734e-01 8.39130878e-01
-6.70981824e-01 7.64146626e-01 1.15520820e-01 -2.79805690e-01
-5.84259689e-01 -1.17663705e+00 -1.33426785e-01 1.28366375e+00
4.92959559e-01 -7.78222606e-02 -4.82279480e-01 -1.06166327e+00
1.99306570e-02 4.64997143e-01 -1.03091300e+00 -2.06889912e-01
-4.50872868e-01 -6.08175159e-01 2.16724142e-01 6.91307604e-01
6.61242008e-01 -1.09521282e+00 -9.26636934e-01 1.07390217e-01
-2.89069504e-01 -9.42151427e-01 -3.80276591e-01 4.92643207e-01
-6.07755184e-01 -1.08125210e+00 -1.04245567e+00 -8.67677569e-01
7.50098050e-01 9.84124839e-01 1.18731284e+00 1.39883578e-01
-5.22642672e-01 4.41925554e-03 -2.13141546e-01 -3.02510053e-01
-6.96161529e-03 -4.09038544e-01 -3.24315906e-01 4.49742168e-01
3.33421648e-01 -1.83124661e-01 -8.39579701e-01 4.30758864e-01
-6.61907971e-01 1.36227226e-02 7.54098117e-01 1.22057903e+00
6.97060287e-01 2.56576538e-01 6.15444481e-01 -5.99955142e-01
-5.24582565e-02 -1.94223002e-01 -3.52087110e-01 3.56253415e-01
-2.11874306e-01 -8.46856162e-02 8.48985255e-01 -5.13170183e-01
-1.05314791e+00 -8.26656222e-02 1.80568948e-01 -6.16227090e-01
-2.66039252e-01 -1.77587807e-01 -3.37416351e-01 -4.13911462e-01
4.40789074e-01 7.21631169e-01 -2.95239627e-01 -3.32224429e-01
3.10109556e-01 4.75988030e-01 6.66295171e-01 -2.54069746e-01
7.27557182e-01 6.28064990e-01 -1.62777796e-01 -8.32185686e-01
-9.25674140e-01 -7.48394668e-01 -5.91947138e-01 -2.63766795e-01
8.33235741e-01 -8.83208573e-01 -2.46844053e-01 3.96520644e-01
-7.52930403e-01 -1.68151274e-01 -3.58334601e-01 2.29199305e-02
-2.81816065e-01 6.20838225e-01 -3.86047870e-01 -8.49481165e-01
-2.49156982e-01 -1.17149460e+00 1.29118347e+00 4.58279192e-01
1.22276135e-01 -7.11877763e-01 -4.21629012e-01 4.67755854e-01
5.46551824e-01 2.56898880e-01 9.30830657e-01 -3.47241372e-01
-6.39537096e-01 7.89027438e-02 -7.83903420e-01 4.84616965e-01
4.54679817e-01 -1.00839756e-01 -1.30391908e+00 -4.15278643e-01
-1.01439670e-01 -2.55806834e-01 1.28604400e+00 4.05555338e-01
1.24418890e+00 -1.67425379e-01 -4.21581358e-01 5.79700291e-01
1.61987126e+00 1.95756152e-01 4.74104702e-01 2.29663432e-01
9.05319095e-01 7.11976588e-01 6.38321042e-01 8.19291100e-02
1.73093490e-02 7.29736567e-01 5.73051751e-01 -4.55051601e-01
-4.55500245e-01 1.08035728e-01 2.73476958e-01 1.56387672e-01
-2.28810355e-01 -1.95962247e-02 -5.10967493e-01 8.30644071e-01
-1.82493854e+00 -1.07161689e+00 -1.90901488e-01 2.24560738e+00
6.34196222e-01 1.83891729e-01 5.31924702e-02 2.34717652e-01
8.06291699e-01 2.60382235e-01 -6.47095084e-01 -7.78253078e-02
-4.05753046e-01 2.68887788e-01 3.58991385e-01 1.24690693e-03
-1.37774801e+00 8.61900091e-01 5.22724104e+00 1.22418284e+00
-1.18502808e+00 2.35653352e-02 9.25001204e-01 7.92891830e-02
8.48976523e-02 -2.30228737e-01 -6.91853940e-01 6.38065875e-01
3.39540809e-01 8.31220299e-02 1.17302038e-01 8.37032378e-01
-7.39890486e-02 -2.80969322e-01 -8.91271055e-01 9.75544810e-01
2.17893481e-01 -1.23290706e+00 1.83190823e-01 -1.14062235e-01
7.44400382e-01 -1.22120447e-01 1.41752198e-01 2.76285350e-01
1.98283307e-02 -1.00378478e+00 8.57169449e-01 2.43746623e-01
7.97283173e-01 -7.14267671e-01 8.36489797e-01 2.08777890e-01
-1.46072531e+00 -3.38935256e-01 -7.19205737e-01 -4.99211550e-02
-2.88249820e-01 5.72344720e-01 -4.14762497e-01 6.05976045e-01
9.93517399e-01 6.87988639e-01 -8.91704142e-01 1.08837390e+00
-4.74179387e-02 5.24040520e-01 3.12969834e-02 1.12242617e-01
3.80799800e-01 1.75935268e-01 4.53289747e-01 1.27377069e+00
-2.39614155e-02 -7.08243847e-02 2.81797618e-01 7.49022245e-01
1.53905675e-01 -6.30902648e-02 -4.22243893e-01 2.83123970e-01
1.10332267e-02 1.25330913e+00 -1.01381636e+00 -5.25076568e-01
-4.75358516e-01 1.29440904e+00 3.75847220e-01 3.40346366e-01
-7.41746604e-01 -2.60162175e-01 7.26061046e-01 6.15132041e-02
9.56988573e-01 4.58765961e-02 -1.77959219e-01 -1.11489725e+00
1.70068499e-02 -9.14322793e-01 4.95830089e-01 -5.77027023e-01
-1.47856426e+00 7.78338790e-01 -3.87991637e-01 -1.23301041e+00
7.24717006e-02 -5.95171928e-01 -6.24076366e-01 9.10887420e-01
-1.80741441e+00 -1.26210344e+00 -6.60370350e-01 8.07276368e-01
1.01649606e+00 -3.98433879e-02 5.66320837e-01 9.64515433e-02
-5.09987235e-01 6.33534431e-01 1.65040806e-01 7.40111470e-02
6.82613313e-01 -1.43578315e+00 2.14241087e-01 1.19127357e+00
2.17688084e-01 3.54791015e-01 4.62247252e-01 -6.01880729e-01
-1.04832280e+00 -1.43164694e+00 5.78844786e-01 -3.51475775e-01
4.60931689e-01 -4.48840916e-01 -1.16814554e+00 9.74352956e-02
9.23069790e-02 4.40241128e-01 2.68766463e-01 -2.20247731e-01
-4.84883666e-01 -2.16565609e-01 -9.35257435e-01 4.41493660e-01
1.04768753e+00 -5.08449376e-01 -6.33073807e-01 2.72752140e-02
3.30936193e-01 1.75910056e-01 -2.46088743e-01 4.97524023e-01
3.37644368e-01 -1.21106493e+00 1.18008626e+00 -3.46030891e-01
3.58041853e-01 -6.34902477e-01 -2.55497485e-01 -1.14334118e+00
-8.10532093e-01 -1.35557815e-01 -1.01143263e-01 1.39967000e+00
-1.52132317e-01 -4.88046169e-01 6.44971907e-01 1.33775651e-01
3.76902968e-02 -6.18812203e-01 -9.82675374e-01 -7.40651667e-01
-2.46075079e-01 -1.10795327e-01 2.42264643e-01 7.85816014e-01
-7.18775511e-01 3.34283233e-01 -3.41071099e-01 2.29039699e-01
8.75821829e-01 5.49221694e-01 5.28434634e-01 -1.20712984e+00
-4.41661507e-01 -5.78251719e-01 -7.29592323e-01 -8.60820353e-01
-7.78422803e-02 -5.70717633e-01 4.37600017e-01 -1.42594135e+00
5.11637330e-01 -3.08336467e-01 -7.23393679e-01 4.52699006e-01
-8.08119595e-01 8.84193599e-01 4.10071939e-01 2.74809480e-01
-8.23066235e-01 5.88489950e-01 1.29906917e+00 -5.80355942e-01
9.03429911e-02 -1.05045810e-02 -8.18278193e-01 5.93497157e-01
4.70536739e-01 -1.52080700e-01 -4.46142524e-01 -1.68083146e-01
-5.99542618e-01 -3.84408295e-01 8.53974998e-01 -1.08561635e+00
5.76407649e-02 -2.25211352e-01 8.91633570e-01 -6.09662175e-01
2.75201589e-01 -5.76201499e-01 -1.44555897e-01 4.62965101e-01
-5.24436608e-02 -3.01116437e-01 3.15028071e-01 7.72231698e-01
-3.25800657e-01 8.40953216e-02 1.09260893e+00 -1.08016409e-01
-1.32047987e+00 1.51112452e-01 -4.77205992e-01 1.17275037e-01
1.18340766e+00 -5.10443628e-01 -4.64987487e-01 -1.07028857e-02
-4.30119872e-01 -3.89821567e-02 5.19839883e-01 2.87136346e-01
6.85285747e-01 -1.02590322e+00 -7.72920191e-01 4.42951262e-01
2.83630073e-01 -4.67473865e-02 6.50229037e-01 7.13582993e-01
-7.86828697e-02 3.61027926e-01 -5.50098598e-01 -9.04045403e-01
-1.49813211e+00 9.36746299e-01 3.45633388e-01 -1.50027439e-01
-8.27231467e-01 1.10787511e+00 1.11947072e+00 3.43812525e-01
2.67304838e-01 -3.07618618e-01 -3.69160771e-01 4.48972471e-02
7.54441321e-01 2.33371511e-01 1.73174277e-01 -7.51128495e-01
-7.01901197e-01 6.34062648e-01 -2.02757075e-01 5.60621917e-01
1.11305642e+00 -3.06023002e-01 1.46754652e-01 1.33253843e-01
1.03128529e+00 8.32709000e-02 -1.68511009e+00 -4.34795350e-01
-1.66729301e-01 -8.72946382e-01 2.42583886e-01 -7.06277311e-01
-1.20874608e+00 1.02539611e+00 8.21292043e-01 2.87172854e-01
1.51918864e+00 1.28095150e-01 4.14956897e-01 -8.94169286e-02
4.51104641e-01 -7.57068217e-01 1.93346247e-01 2.06202269e-01
7.66411960e-01 -1.44718564e+00 -1.12743035e-01 -4.94426787e-01
-6.54354811e-01 1.00736594e+00 7.94773877e-01 -1.69361159e-01
2.60760695e-01 -4.50832304e-03 1.19541906e-01 -3.60954483e-03
-6.46003962e-01 -6.67717457e-01 6.25956655e-01 9.74170029e-01
1.29618481e-01 -1.60349291e-02 1.16255954e-01 3.56425166e-01
3.76314819e-01 -5.38142979e-01 1.66289091e-01 6.47884250e-01
-5.23914158e-01 -6.79545999e-01 -5.06937802e-01 5.69184721e-01
-6.13829434e-01 -1.67446867e-01 -5.03459215e-01 5.02738595e-01
3.92318308e-01 1.06281292e+00 1.80692866e-01 -2.00103596e-01
1.19392239e-01 -3.88617098e-01 4.54620719e-01 -4.63205308e-01
-6.38603151e-01 2.53882974e-01 -2.13518336e-01 -7.03028321e-01
-4.93670225e-01 -5.86177230e-01 -7.04100430e-01 -5.35714114e-03
-3.35254848e-01 -1.69603284e-02 4.56805676e-02 8.75545621e-01
2.84740508e-01 8.04662704e-01 5.73536575e-01 -1.09795761e+00
-4.02189910e-01 -7.18637466e-01 -3.90916556e-01 6.02309704e-01
7.20118284e-01 -9.37378824e-01 -2.49222592e-01 -1.46634309e-02] | [9.668898582458496, 1.9131972789764404] |
70e773b0-6d76-44a4-817e-9f4886c6713f | ordered-neurons-integrating-tree-structures | 1810.09536 | null | https://arxiv.org/abs/1810.09536v6 | https://arxiv.org/pdf/1810.09536v6.pdf | Ordered Neurons: Integrating Tree Structures into Recurrent Neural Networks | Natural language is hierarchically structured: smaller units (e.g., phrases) are nested within larger units (e.g., clauses). When a larger constituent ends, all of the smaller constituents that are nested within it must also be closed. While the standard LSTM architecture allows different neurons to track information at different time scales, it does not have an explicit bias towards modeling a hierarchy of constituents. This paper proposes to add such an inductive bias by ordering the neurons; a vector of master input and forget gates ensures that when a given neuron is updated, all the neurons that follow it in the ordering are also updated. Our novel recurrent architecture, ordered neurons LSTM (ON-LSTM), achieves good performance on four different tasks: language modeling, unsupervised parsing, targeted syntactic evaluation, and logical inference. | ['Alessandro Sordoni', 'Yikang Shen', 'Shawn Tan', 'Aaron Courville'] | 2018-10-22 | ordered-neurons-integrating-tree-structures-1 | https://openreview.net/forum?id=B1l6qiR5F7 | https://openreview.net/pdf?id=B1l6qiR5F7 | iclr-2019-5 | ['constituency-grammar-induction'] | ['natural-language-processing'] | [ 2.19344541e-01 5.10271251e-01 -2.93645948e-01 -5.96200049e-01
-2.64397800e-01 -6.21636748e-01 2.08025128e-01 5.78623533e-01
-3.76789033e-01 8.84712696e-01 4.64488506e-01 -6.33024096e-01
2.94607401e-01 -1.17475355e+00 -9.14214373e-01 -5.44948757e-01
-2.77794987e-01 5.68874180e-01 4.55346256e-01 -6.68430552e-02
1.48748234e-01 4.44756687e-01 -1.12121737e+00 7.41746426e-01
4.64797437e-01 7.77745485e-01 2.02024534e-01 5.38450003e-01
-6.67551696e-01 1.54506695e+00 -5.87299764e-01 -1.44350752e-01
-4.04844314e-01 -4.40887094e-01 -1.12861550e+00 -2.84151405e-01
2.15284340e-02 -4.18091744e-01 -5.83850145e-02 9.70553994e-01
-7.63294473e-02 3.38486135e-01 2.35280797e-01 -6.78848088e-01
-9.18973088e-01 1.48401558e+00 -1.43885881e-01 3.02809149e-01
9.66563970e-02 -4.62688684e-01 1.50535703e+00 -5.46511233e-01
6.66366041e-01 1.21057642e+00 6.29024804e-01 6.46051764e-01
-1.40903282e+00 -4.32695955e-01 8.15715492e-01 -2.24705547e-01
-6.45081460e-01 -2.53113329e-01 4.69145805e-01 -1.24497525e-01
1.47175467e+00 6.98436275e-02 3.84956300e-01 8.39848518e-01
5.14611959e-01 8.76991689e-01 7.48223126e-01 -4.83591378e-01
3.03022146e-01 1.06239673e-02 9.36477125e-01 8.68383527e-01
2.05057383e-01 -2.24040046e-01 -4.25686032e-01 2.17514969e-02
5.30039549e-01 -2.52495985e-03 1.89736247e-01 -6.30167723e-02
-8.28669190e-01 9.18334305e-01 5.26128173e-01 7.81584620e-01
-3.57720405e-01 4.58459735e-01 4.53205496e-01 3.52461755e-01
2.42373362e-01 4.54168767e-01 -8.44340980e-01 1.27944723e-01
-7.89256573e-01 -1.77700698e-01 9.11367655e-01 9.47509229e-01
9.09233391e-01 9.35395434e-02 -2.32743904e-01 7.68333137e-01
1.70152202e-01 -6.03473671e-02 5.82424164e-01 -9.29919004e-01
4.05868024e-01 9.69432354e-01 -1.58665419e-01 -7.06503987e-01
-5.59851825e-01 -2.78989822e-01 -7.76094735e-01 -1.98922418e-02
1.02950342e-01 -2.21357763e-01 -1.07811165e+00 2.07274723e+00
-2.96729021e-02 -2.56053478e-01 1.98262602e-01 3.41540128e-01
1.08908165e+00 1.11746478e+00 4.45651203e-01 -2.90029675e-01
1.22925806e+00 -8.52226555e-01 -5.20761669e-01 -4.60152537e-01
6.96888268e-01 -2.22920313e-01 8.43688428e-01 2.45808642e-02
-1.45013678e+00 -4.76018012e-01 -8.19073915e-01 -4.94765371e-01
-4.99446124e-01 -2.76411891e-01 6.76020026e-01 7.55178854e-02
-1.31675839e+00 6.77418947e-01 -9.42976713e-01 -2.20900029e-01
-1.19228125e-01 5.34243166e-01 -2.07417220e-01 4.15881217e-01
-1.55283320e+00 7.96186626e-01 7.14616418e-01 2.77566284e-01
-6.14745617e-01 -2.88222104e-01 -1.03313696e+00 4.52152550e-01
1.15016341e-01 -6.36003256e-01 1.48795187e+00 -1.06445801e+00
-1.27099752e+00 8.00025642e-01 -6.87354565e-01 -8.80912185e-01
-1.83343425e-01 -1.64664418e-01 -1.80917740e-01 -2.27624908e-01
1.11595534e-01 9.81354117e-01 5.36416531e-01 -1.17031884e+00
-8.45259786e-01 -3.17294657e-01 3.77133161e-01 2.23466650e-01
-3.39916617e-01 2.90795892e-01 -4.41949487e-01 -4.91048932e-01
5.04278898e-01 -7.21470833e-01 -2.09110633e-01 -6.46715879e-01
-6.44660413e-01 -5.21571279e-01 4.49033260e-01 -4.62525517e-01
1.60262096e+00 -1.97176242e+00 3.91368747e-01 7.93725078e-04
3.10253084e-01 -9.44356397e-02 9.09100175e-02 3.14933866e-01
-1.95833474e-01 5.56340277e-01 -2.84596711e-01 -2.53974736e-01
-1.97052825e-02 6.86316848e-01 -5.11527538e-01 -7.25064576e-02
2.79503196e-01 1.00931811e+00 -8.28010082e-01 -4.94855493e-01
-1.79422155e-01 1.36543691e-01 -6.79331183e-01 -8.55652392e-02
-5.73770106e-01 -3.04827746e-02 -4.47677970e-01 4.89762396e-01
1.53681368e-01 -4.27577853e-01 3.42717648e-01 -3.21947448e-02
-6.51645482e-01 1.11014402e+00 -9.89274800e-01 9.30381894e-01
-4.51128900e-01 6.43618286e-01 5.26811853e-02 -8.50030780e-01
6.96932018e-01 3.87016982e-01 4.98497263e-02 -6.41429067e-01
4.54123393e-02 6.17519245e-02 4.55455929e-01 -7.47062340e-02
5.54712176e-01 -1.37076825e-01 -3.89858693e-01 6.81956589e-01
4.38379981e-02 2.01447174e-01 4.81952101e-01 3.39927435e-01
1.16087663e+00 -1.75176952e-02 2.85097092e-01 -2.31113002e-01
5.17691493e-01 -7.05650300e-02 9.11417067e-01 9.51740742e-01
1.81910321e-02 -1.27515960e-02 9.34111714e-01 -6.45070553e-01
-1.07558084e+00 -9.60517764e-01 -1.18910871e-01 1.90738261e+00
-3.02637666e-01 -4.22182322e-01 -7.16579676e-01 -4.34735179e-01
-6.96326718e-02 6.91467643e-01 -9.29383874e-01 1.87095866e-01
-1.04086339e+00 -3.97943825e-01 5.38150668e-01 9.20631289e-01
3.40051919e-01 -1.84635556e+00 -7.44608104e-01 6.53417051e-01
2.83435490e-02 -8.98663938e-01 -1.81846246e-01 9.84282374e-01
-1.27509856e+00 -7.54980326e-01 3.41013446e-02 -1.28889358e+00
9.55597579e-01 -2.81668782e-01 1.27264929e+00 1.35786623e-01
3.72447819e-01 -8.00521448e-02 -2.53289700e-01 -1.36177287e-01
-4.84719187e-01 4.02090251e-01 -3.04725379e-01 -3.04753095e-01
3.75395030e-01 -3.87436837e-01 1.25593171e-01 -1.68268621e-01
-7.83034980e-01 1.88351780e-01 4.89986748e-01 9.57828462e-01
6.78683221e-01 -2.48961542e-02 1.97446316e-01 -1.40384209e+00
6.65840745e-01 -2.66523570e-01 -4.98837948e-01 3.28116417e-01
-2.12229878e-01 5.65624416e-01 8.81356239e-01 -3.37577730e-01
-9.57191527e-01 -1.61154971e-01 -1.54877767e-01 1.33105084e-01
-9.03726816e-02 8.62344861e-01 1.16627328e-02 5.85041642e-01
1.69286951e-01 2.19365835e-01 -4.85705703e-01 -4.15424049e-01
2.20207736e-01 1.71152160e-01 3.94477427e-01 -5.99420130e-01
4.05168056e-01 5.99991158e-02 -2.36365214e-01 -5.16771495e-01
-9.13149178e-01 2.82851867e-02 -8.77161980e-01 5.06347679e-02
9.80807483e-01 -5.88099957e-01 -6.66668773e-01 2.72396713e-01
-1.51494825e+00 -5.75739682e-01 -3.53265673e-01 1.15138337e-01
-8.63957629e-02 -4.23107902e-03 -1.33931231e+00 -7.59691775e-01
-5.48773408e-01 -9.28066254e-01 5.61290145e-01 2.07590804e-01
-6.55790389e-01 -1.15168977e+00 -1.29787028e-01 -3.52838397e-01
2.51973122e-01 1.41800540e-02 1.59601009e+00 -8.61230433e-01
-6.05575502e-01 4.17800620e-02 2.90979743e-02 3.70808572e-01
1.87109366e-01 3.34417820e-01 -5.22985280e-01 -9.76334736e-02
9.22977850e-02 -3.75893682e-01 1.08720481e+00 5.90006292e-01
9.60963428e-01 -5.81532955e-01 -2.94344276e-01 2.92713195e-01
9.86179769e-01 6.54664993e-01 4.59390879e-01 4.90260333e-01
5.57057083e-01 6.62422657e-01 -9.75872576e-02 -5.37857115e-02
4.71820176e-01 -4.85732593e-02 2.55811900e-01 -1.30790576e-01
2.57991672e-01 -2.15778425e-01 6.66838706e-01 1.12455106e+00
1.72047347e-01 -3.46398503e-01 -9.62375820e-01 5.15780985e-01
-1.66523790e+00 -9.32177901e-01 -2.07487717e-01 1.85691583e+00
1.02583802e+00 7.23425567e-01 -4.78079319e-01 3.47156972e-02
8.22093129e-01 3.46081734e-01 -7.80602276e-01 -1.09959257e+00
-2.44670019e-01 4.92374718e-01 2.51363248e-01 8.38972449e-01
-9.39503193e-01 1.43876588e+00 6.93940067e+00 -1.73504408e-02
-9.81270134e-01 1.02953173e-01 8.53618145e-01 -2.76961662e-02
-5.40370286e-01 1.04010783e-01 -1.11557341e+00 1.34652093e-01
1.03353024e+00 1.26791611e-01 3.14174026e-01 6.98967516e-01
7.40301311e-02 -3.54412615e-01 -1.35241771e+00 1.69012785e-01
-1.74185127e-01 -1.43823957e+00 3.35394055e-01 -2.88576752e-01
6.69119537e-01 2.36852229e-01 -1.24620423e-01 5.23391843e-01
8.93946826e-01 -9.68777239e-01 7.62216449e-01 2.51338035e-01
2.31001437e-01 -6.84206426e-01 5.67643881e-01 3.10171247e-01
-1.25056326e+00 -1.52587041e-01 -3.35852444e-01 -4.29560870e-01
1.75926611e-01 4.07930374e-01 -4.72503036e-01 -5.82735427e-02
7.99924910e-01 4.68127906e-01 -3.19658667e-01 2.75967896e-01
-7.43381977e-01 7.31240928e-01 -1.48244455e-01 -4.25972074e-01
7.13386416e-01 -2.09748164e-01 2.76771098e-01 1.27819788e+00
-2.10765675e-01 1.60694197e-01 4.61149365e-02 9.35274601e-01
-3.81004393e-01 -1.46497741e-01 -4.11972225e-01 -1.35680243e-01
6.49697661e-01 9.86895740e-01 -7.08047748e-01 -6.64585829e-01
-4.75677133e-01 6.00597441e-01 8.17187071e-01 4.50018972e-01
-6.23325109e-01 -2.96814710e-01 5.73362291e-01 9.31486115e-03
4.09241259e-01 -2.81641930e-01 -5.09569287e-01 -7.77862906e-01
-9.86716300e-02 -5.11768818e-01 6.18457735e-01 -7.59920895e-01
-1.01623750e+00 9.22732711e-01 -3.22303444e-01 -3.87638062e-01
-5.56451380e-01 -6.99010789e-01 -7.22064257e-01 8.57636034e-01
-1.24348402e+00 -9.06740844e-01 3.91594350e-01 4.26375151e-01
4.34885353e-01 1.43047243e-01 1.08524811e+00 -3.91609073e-02
-8.28175604e-01 2.54356146e-01 -1.31038995e-02 7.43439496e-01
1.67080030e-01 -1.26848912e+00 5.14489532e-01 9.31914091e-01
1.08048379e-01 1.21962047e+00 4.23375398e-01 -6.47670865e-01
-9.77364898e-01 -1.06461489e+00 1.61903083e+00 -1.75239995e-01
9.70142126e-01 -6.80332124e-01 -1.17851555e+00 1.38359010e+00
3.25346977e-01 -3.57564956e-01 6.35312676e-01 4.81595993e-01
-3.73742014e-01 -1.21086955e-01 -5.97771883e-01 6.11618519e-01
7.28612959e-01 -6.41914606e-01 -1.03161597e+00 2.51733251e-02
1.01779890e+00 -3.31394464e-01 -6.42677009e-01 5.24573624e-01
4.35911983e-01 -8.53334785e-01 4.73545998e-01 -9.90199089e-01
5.58072031e-01 -1.39567256e-01 -8.89942721e-02 -1.00638151e+00
-6.84638500e-01 -3.28009039e-01 -3.80951941e-01 1.25093532e+00
1.03499055e+00 -7.39369154e-01 9.21502590e-01 8.80721211e-01
-3.74415904e-01 -8.56043637e-01 -8.25264037e-01 -3.78895909e-01
2.98841655e-01 -4.05330181e-01 4.44610059e-01 8.00782979e-01
1.31143764e-01 7.00610697e-01 -4.70607169e-03 8.17166641e-02
1.79351091e-01 4.26170886e-01 7.57967979e-02 -1.20365131e+00
-1.17216662e-01 -5.98754525e-01 1.56341165e-01 -1.28065670e+00
3.60155582e-01 -9.89114344e-01 1.31991133e-01 -1.93500197e+00
2.86831826e-01 -4.04004961e-01 -5.50464034e-01 9.85878408e-01
3.41746248e-02 -2.39640743e-01 5.32324426e-02 1.22675456e-01
-4.95795608e-01 2.83902399e-02 8.79984319e-01 -2.51579821e-01
-4.05389488e-01 -2.29854628e-01 -6.95654631e-01 9.69357729e-01
9.70558226e-01 -5.98102689e-01 1.61448941e-02 -8.32818508e-01
4.16440576e-01 2.04679854e-02 -1.34660006e-01 -7.63758063e-01
6.42836630e-01 1.09411255e-01 5.09200692e-01 -8.35632145e-01
-5.69027569e-03 -4.03954238e-01 3.32840830e-02 6.89053535e-01
-8.19873393e-01 5.64180911e-01 2.62700438e-01 2.44745556e-02
-3.52162540e-01 -4.74507958e-01 6.49933040e-01 -5.34034848e-01
-6.37381613e-01 5.13501652e-02 -8.93262684e-01 1.62972454e-02
4.64409798e-01 -3.03358108e-01 -2.91917324e-01 -5.06728217e-02
-1.03126276e+00 4.07150447e-01 1.03567109e-01 2.81509846e-01
5.50414860e-01 -9.87040102e-01 -3.49243492e-01 6.39540628e-02
-3.55790734e-01 3.10715467e-01 -1.61409050e-01 3.69130671e-01
-3.34232867e-01 8.66596162e-01 5.28793521e-02 -3.05483371e-01
-1.08036053e+00 3.72895807e-01 4.04086918e-01 -8.18745315e-01
-6.11810803e-01 1.11365509e+00 3.22804481e-01 -5.95587909e-01
4.60681945e-01 -1.11630142e+00 -4.39187407e-01 3.41738969e-01
3.60167414e-01 -1.40225083e-01 3.16320956e-02 -5.15055120e-01
-3.02800655e-01 2.45499954e-01 -3.88684064e-01 2.58621834e-02
1.44047952e+00 1.34317860e-01 -9.54209745e-01 1.24303079e+00
9.69783604e-01 -1.48671940e-01 -8.37125659e-01 -2.48321950e-01
5.26723981e-01 4.84640300e-01 -8.55662525e-02 -6.56731129e-01
-8.10292602e-01 7.79035747e-01 -2.52061397e-01 4.83736515e-01
8.80441785e-01 2.20940709e-01 6.90039635e-01 7.78378248e-01
4.14813608e-01 -1.08898771e+00 1.97456423e-02 1.42619693e+00
5.66442907e-01 -7.26434827e-01 -3.36676210e-01 -8.91448855e-02
-3.64517182e-01 1.26384926e+00 6.93222225e-01 -1.82390764e-01
4.13695008e-01 6.47343278e-01 -3.29312652e-01 -1.57212704e-01
-1.24734044e+00 -1.84844226e-01 -7.11526945e-02 -3.45099457e-02
8.91273022e-01 8.68075341e-02 -1.82926938e-01 6.69178843e-01
-2.39080280e-01 -1.93217248e-01 4.54043716e-01 1.10648084e+00
-8.65524590e-01 -1.19322908e+00 -1.93735957e-01 5.95100522e-01
-4.88594741e-01 -6.34122670e-01 -4.81689990e-01 8.74351084e-01
2.07042530e-01 6.89075112e-01 7.46418238e-01 -6.16231896e-02
3.08037490e-01 3.79015923e-01 2.82415390e-01 -1.02632129e+00
-9.65495110e-01 -5.82611701e-03 2.03865871e-01 -2.55086750e-01
-3.95957202e-01 -6.75436258e-01 -2.02067518e+00 -3.42654735e-01
-5.71890734e-02 3.75105351e-01 2.59620667e-01 8.57658505e-01
1.39199533e-02 9.78903472e-01 2.40245387e-01 -2.73382723e-01
-1.91229999e-01 -9.93124425e-01 -4.04498219e-01 1.05504453e-01
3.83895427e-01 -2.25280926e-01 -1.42709702e-01 6.36624768e-02] | [10.37945556640625, 9.291975021362305] |
2da586c2-2995-4b72-bd0f-d290684d94e9 | optimal-transport-minimization-crowd | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Lin_Optimal_Transport_Minimization_Crowd_Localization_on_Density_Maps_for_Semi-Supervised_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Lin_Optimal_Transport_Minimization_Crowd_Localization_on_Density_Maps_for_Semi-Supervised_CVPR_2023_paper.pdf | Optimal Transport Minimization: Crowd Localization on Density Maps for Semi-Supervised Counting | The accuracy of crowd counting in images has improved greatly in recent years due to the development of deep neural networks for predicting crowd density maps. However, most methods do not further explore the ability to localize people in the density map, with those few works adopting simple methods, like finding the local peaks in the density map. In this paper, we propose the optimal transport minimization (OT-M) algorithm for crowd localization with density maps. The objective of OT-M is to find a target point map that has the minimal Sinkhorn distance with the input density map, and we propose an iterative algorithm to compute the solution. We then apply OT-M to generate hard pseudo-labels (point maps) for semi-supervised counting, rather than the soft pseudo-labels (density maps) used in previous methods. Our hard pseudo-labels provide stronger supervision, and also enable the use of recent density-to-point loss functions for training. We also propose a confidence weighting strategy to give higher weight to the more reliable unlabeled data. Extensive experiments show that our methods achieve outstanding performance on both crowd localization and semi-supervised counting. Code is available at https://github.com/Elin24/OT-M. | ['Antoni B. Chan', 'Wei Lin'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['crowd-counting'] | ['computer-vision'] | [-3.58016461e-01 -7.83237964e-02 -7.79785169e-03 -6.07622445e-01
-6.23939633e-01 -1.64298669e-01 5.78893363e-01 2.65204817e-01
-8.29181194e-01 9.90205407e-01 2.99383819e-01 8.08291435e-02
2.38008156e-01 -8.91202748e-01 -6.49639606e-01 -5.88856578e-01
-1.14205413e-01 8.73795450e-01 5.41993201e-01 5.97628169e-02
2.04288885e-01 2.41626307e-01 -1.49113047e+00 -1.97952732e-01
1.15555251e+00 7.63987720e-01 3.51746321e-01 8.18510950e-01
-2.56618470e-01 9.98982608e-01 -7.81586170e-01 -7.06013858e-01
1.70441225e-01 -3.05410057e-01 -6.76400840e-01 2.81516127e-02
4.53966647e-01 -5.35461783e-01 -3.05148572e-01 1.09672439e+00
6.12990320e-01 1.64452896e-01 9.14212465e-01 -1.24358988e+00
-6.71010554e-01 2.74949044e-01 -9.33343887e-01 5.57664335e-01
4.12326366e-01 6.65869862e-02 5.35666943e-01 -1.02072358e+00
2.52952009e-01 1.28110397e+00 9.35644090e-01 5.47484338e-01
-7.54427075e-01 -5.48102856e-01 -6.21212833e-03 2.42932066e-02
-1.88104928e+00 -3.22656631e-01 4.28512067e-01 -5.79802990e-01
6.03400588e-01 -2.36432552e-01 6.85242295e-01 5.77427328e-01
-3.67210805e-01 8.38203549e-01 9.28394079e-01 -4.56007957e-01
4.30015445e-01 1.89404055e-01 -1.58114001e-01 1.08155274e+00
4.58126813e-01 -4.06660765e-01 -5.95122635e-01 -3.00858527e-01
9.37673926e-01 1.05478689e-01 -1.64862409e-01 -1.00115538e-01
-8.94601047e-01 1.10565674e+00 7.47671247e-01 1.54648185e-01
-1.86401933e-01 1.96337029e-01 2.04680815e-01 -2.89302021e-01
8.68229926e-01 6.84616268e-02 2.82948813e-03 -3.15166265e-01
-1.01087046e+00 4.01168644e-01 8.07582259e-01 9.23375309e-01
1.06427026e+00 -3.77417713e-01 -3.63558680e-01 7.80870616e-01
3.06297362e-01 8.54869604e-01 3.59266065e-02 -1.30577469e+00
7.34611154e-01 5.75888991e-01 4.35309023e-01 -9.94544506e-01
-2.95362294e-01 1.70809820e-01 -7.46799111e-01 -1.38259456e-02
9.03051794e-01 -3.97694319e-01 -7.70285428e-01 1.46855617e+00
4.07938510e-01 3.45336169e-01 -3.06055754e-01 9.86422479e-01
4.97917622e-01 5.12409568e-01 9.94090289e-02 1.79407597e-01
9.97842669e-01 -7.90433884e-01 -4.20995504e-01 -2.58050978e-01
7.89984822e-01 -2.38099024e-01 1.02552021e+00 -1.74849302e-01
-1.05232203e+00 -3.01422060e-01 -7.19948828e-01 -8.53340551e-02
-3.36773187e-01 -3.77258169e-03 6.39975667e-01 6.35774314e-01
-1.22904551e+00 4.50826943e-01 -7.85200775e-01 -4.65212017e-01
9.03060317e-01 3.70009929e-01 -8.98229852e-02 -1.13318995e-01
-1.05595791e+00 8.55998516e-01 1.09011315e-01 -8.12366977e-02
-5.97188413e-01 -4.78327245e-01 -1.12593663e+00 -3.55867177e-01
1.15790494e-01 -6.33905113e-01 1.18931246e+00 -6.27504110e-01
-1.12348104e+00 1.08058655e+00 -7.59650826e-01 -4.44591910e-01
6.66547954e-01 -5.14192879e-02 1.58880144e-01 3.84240240e-01
8.30707133e-01 1.23580050e+00 4.64879483e-01 -1.28624177e+00
-1.00124049e+00 -2.01507390e-01 -4.49614003e-02 2.28586078e-01
-3.12009394e-01 7.54880384e-02 -5.90033770e-01 -1.95202976e-01
-2.65612155e-01 -7.01938272e-01 -2.37747520e-01 1.79536268e-01
-3.04985881e-01 -5.07270873e-01 5.95618546e-01 -5.46600640e-01
1.01306093e+00 -1.78982151e+00 -2.59788841e-01 2.21007079e-01
5.78311145e-01 2.22058848e-01 1.88262880e-01 -7.46364892e-02
7.06936002e-01 -6.54387996e-02 -5.66992581e-01 -9.09001470e-01
-7.69189522e-02 3.60570818e-01 1.05685204e-01 7.38190830e-01
3.39237928e-01 9.98004198e-01 -1.21784282e+00 -1.04464817e+00
3.29009712e-01 6.04149580e-01 -4.52597022e-01 1.69454142e-01
-4.34879884e-02 6.29043221e-01 -2.39939794e-01 8.75506103e-01
8.52521956e-01 -5.17118752e-01 -2.42374212e-01 3.92774165e-01
-1.88700795e-01 -7.86200538e-02 -8.91254842e-01 1.28101420e+00
-2.10779309e-01 5.21556795e-01 -1.27977859e-02 -5.65543592e-01
9.04623032e-01 -5.28734401e-02 4.48237032e-01 -6.22404516e-01
2.11854100e-01 3.75128150e-01 -4.39930707e-01 -3.20705682e-01
7.31434941e-01 -1.16530828e-01 -1.23654887e-01 2.98497409e-01
9.20637976e-04 -5.38438447e-02 4.92904484e-01 2.21017987e-01
8.76758695e-01 -9.60634574e-02 9.84170809e-02 -1.40261278e-01
4.12614256e-01 -3.99140455e-02 4.92241532e-01 8.31590235e-01
-6.04670286e-01 8.41513097e-01 4.52315778e-01 -4.42192197e-01
-1.19743824e+00 -1.05050910e+00 -2.19050616e-01 9.39115226e-01
3.58664751e-01 -2.09978651e-02 -1.04418218e+00 -7.28021324e-01
2.02450499e-01 2.96879619e-01 -7.33158290e-01 3.88989449e-01
-8.90376508e-01 -8.35127890e-01 8.48254681e-01 7.48225868e-01
7.51025081e-01 -8.68130028e-01 -4.08647448e-01 -3.68881896e-02
-5.38305819e-01 -1.07875431e+00 -7.67486572e-01 7.33686909e-02
-6.36295259e-01 -9.43477035e-01 -1.31149077e+00 -8.32719028e-01
9.84257638e-01 3.75394464e-01 1.18305063e+00 3.98333967e-01
-5.29213101e-02 3.70141774e-01 -2.06423670e-01 -4.97312546e-01
2.98686437e-02 2.33242363e-01 1.87414080e-01 -1.23977587e-01
9.15882051e-01 -3.55961710e-01 -6.81858897e-01 5.08922875e-01
-4.83139843e-01 -4.41121221e-01 1.41474202e-01 4.78435367e-01
6.56048059e-01 -2.00324640e-01 5.90750933e-01 -5.44922352e-01
6.25881195e-01 -6.95027411e-01 -6.33949041e-01 -4.15749550e-02
-2.41007745e-01 -2.74097156e-02 3.83517683e-01 -3.35018486e-01
-7.22078502e-01 8.62191096e-02 -1.26384586e-01 -4.77465183e-01
-1.31399676e-01 -2.55657751e-02 5.32684773e-02 -2.28241518e-01
6.88342035e-01 3.53030451e-02 -9.20136049e-02 -1.81136668e-01
9.42798182e-02 6.29379749e-01 5.62124193e-01 -6.08908117e-01
7.64443815e-01 8.61434460e-01 -1.26320407e-01 -7.64885306e-01
-1.21342242e+00 -7.85127699e-01 -8.96297038e-01 -3.62594247e-01
1.07650483e+00 -1.01073718e+00 -7.52144754e-01 4.68340665e-01
-1.26768374e+00 -5.50967097e-01 -3.51772279e-01 3.18837076e-01
-4.30810004e-01 5.68288624e-01 -5.64186096e-01 -1.21626413e+00
-1.47714600e-01 -8.43362391e-01 1.49589467e+00 5.23513675e-01
-9.40930471e-02 -1.37401354e+00 2.86184430e-01 2.50670105e-01
3.27569366e-01 9.84321386e-02 6.85481653e-02 -2.98391879e-01
-5.63570559e-01 -2.12601557e-01 -5.54899752e-01 1.18755989e-01
-1.30211160e-01 -2.73105800e-01 -1.02331245e+00 -1.01979084e-01
-3.30480993e-01 -4.99555349e-01 1.06189632e+00 8.44428957e-01
1.11331844e+00 -2.40117744e-01 -5.02550542e-01 7.57050395e-01
1.17384827e+00 -3.13860625e-01 6.27956867e-01 3.86194676e-01
9.04353678e-01 5.45213163e-01 5.14551520e-01 6.23028159e-01
1.18059492e+00 5.52832067e-01 2.26652786e-01 -8.91406685e-02
-1.60034895e-02 -6.85492218e-01 1.59173682e-01 7.19857991e-01
-3.95003706e-01 -2.25511342e-01 -1.10504520e+00 6.74665749e-01
-1.95837808e+00 -1.03641498e+00 -1.19275972e-01 2.12429166e+00
6.82822347e-01 -9.11106355e-03 5.93347013e-01 -5.22041656e-02
1.06514001e+00 3.03438446e-03 -3.22404861e-01 2.37541571e-01
-2.62898058e-01 -1.03934640e-02 8.04068923e-01 8.43088448e-01
-1.35943496e+00 9.00224388e-01 6.28544760e+00 7.91760564e-01
-5.68238139e-01 3.28275740e-01 8.74333501e-01 -4.94054668e-02
1.12271383e-01 -3.03288817e-01 -1.26581013e+00 8.64269376e-01
6.68951452e-01 1.17082886e-01 1.89837024e-01 8.51292193e-01
2.50326306e-01 -6.34156883e-01 -7.64851213e-01 1.16660774e+00
1.86978206e-01 -1.12901795e+00 -2.32315004e-01 2.29697108e-01
8.70939434e-01 1.03718475e-01 -1.78530991e-01 2.88446128e-01
5.17591119e-01 -1.00089550e+00 8.88256609e-01 6.89750671e-01
7.84624279e-01 -8.62348318e-01 1.09482431e+00 6.01564288e-01
-1.41238844e+00 1.06935605e-01 -7.87284136e-01 -2.84800500e-01
4.53562021e-01 7.62094796e-01 -8.86122584e-01 -9.44224671e-02
8.18373919e-01 5.46115160e-01 -6.47416890e-01 1.27439559e+00
-3.51018608e-01 3.78564715e-01 -6.41669333e-01 -4.38215196e-01
1.74575925e-01 -2.01432407e-01 2.40494817e-01 1.51136243e+00
3.37456256e-01 -1.39825359e-01 4.23790634e-01 8.54431987e-01
-2.23510996e-01 -5.31399101e-02 -6.62608504e-01 4.61886078e-01
7.18773663e-01 9.67669427e-01 -9.60186899e-01 -4.20689970e-01
-2.13470042e-01 8.88460159e-01 8.64851058e-01 2.07741767e-01
-9.29092526e-01 -3.36594850e-01 2.93333173e-01 6.43916488e-01
2.60532618e-01 -3.45904887e-01 -2.58600533e-01 -1.02337408e+00
1.48612618e-01 -9.22195762e-02 2.86941618e-01 -4.93361801e-01
-1.63018394e+00 3.27581972e-01 2.72357017e-01 -9.48691487e-01
-1.31865323e-01 -3.30128223e-01 -8.22739244e-01 8.97645116e-01
-1.68210232e+00 -8.36567283e-01 -6.16018355e-01 6.29126906e-01
1.89083800e-01 3.67212445e-02 4.00000125e-01 4.72947389e-01
-2.64555782e-01 5.48558772e-01 -1.87990442e-01 4.90854084e-01
4.79486525e-01 -1.51045942e+00 3.45351189e-01 7.22268164e-01
-2.46036068e-01 3.18476468e-01 3.49755079e-01 -8.80465388e-01
-5.30248463e-01 -1.25876415e+00 1.09314489e+00 -9.19426203e-01
4.92957681e-01 -4.68216568e-01 -7.42423654e-01 4.68305051e-01
-1.77849516e-01 2.38125622e-01 4.48619515e-01 -1.21063925e-01
-8.54781419e-02 2.14948773e-01 -1.44896352e+00 2.71978021e-01
1.00082767e+00 -4.10984904e-01 -1.86461791e-01 5.92638791e-01
5.59295058e-01 -3.13832194e-01 -5.16790450e-01 -2.59639416e-02
-3.65060456e-02 -1.02939069e+00 9.32438016e-01 2.42238939e-01
3.28919172e-01 -5.33804357e-01 -3.81130725e-02 -1.14184868e+00
-2.58222461e-01 -2.75884271e-01 -2.43511081e-01 1.10122800e+00
3.61075997e-01 -6.19962633e-01 1.11310160e+00 4.99870241e-01
8.15108865e-02 -7.09552526e-01 -1.14385068e+00 -6.66607499e-01
9.78063494e-02 -1.09970734e-01 5.91854453e-01 7.67753839e-01
-1.12714015e-01 1.39759347e-01 -4.17984307e-01 1.49209440e-01
1.00093102e+00 -4.78121608e-01 9.62421834e-01 -1.11219716e+00
5.84009513e-02 -3.81990939e-01 -4.97449249e-01 -1.41190851e+00
1.15822449e-01 -7.77301550e-01 3.97233695e-01 -1.91123199e+00
4.22488511e-01 -6.98206723e-01 3.48064661e-01 2.57630348e-01
-5.55676341e-01 5.58319330e-01 1.51342973e-01 5.60727894e-01
-1.29307008e+00 7.06090987e-01 1.17167521e+00 -7.85577670e-03
-2.17667207e-01 7.75351152e-02 -5.25890887e-01 9.81172502e-01
1.01474166e+00 -6.19424284e-01 -2.65959534e-03 -5.75317025e-01
3.27648185e-02 -2.78173357e-01 4.89890158e-01 -1.35654330e+00
4.67379183e-01 5.36577180e-02 6.14339292e-01 -6.20674133e-01
4.91271704e-01 -2.84651428e-01 -5.87355971e-01 3.13737512e-01
6.34220093e-02 1.11241922e-01 -2.71574974e-01 5.91896594e-01
-8.32374915e-02 -5.34335673e-01 8.89945328e-01 -2.72384226e-01
-5.35841882e-01 4.56043690e-01 -1.07831389e-01 3.74811471e-01
9.47510302e-01 -3.93337876e-01 -4.12522376e-01 -5.75272441e-01
-3.82124990e-01 5.52391410e-01 7.33613372e-01 -1.80548787e-01
6.16792262e-01 -1.29322791e+00 -7.96847463e-01 -1.48130581e-01
3.51471119e-02 5.23914754e-01 -1.89541504e-01 8.93056214e-01
-8.57739747e-01 2.68284798e-01 1.58765152e-01 -8.79610777e-01
-7.14191556e-01 2.76651502e-01 5.06218791e-01 -2.34763637e-01
-5.67884505e-01 1.11887944e+00 5.35729751e-02 -6.13596082e-01
3.95824254e-01 -1.23987027e-01 -1.43870845e-01 -2.03514904e-01
8.75673592e-01 7.41427243e-01 -3.79090279e-01 -9.02449250e-01
-4.30240452e-01 6.04922950e-01 2.20582917e-01 -1.98733896e-01
1.22727680e+00 -2.07474798e-01 -7.45693513e-04 4.09067124e-01
1.07912493e+00 -3.17488089e-02 -1.65720463e+00 -2.74705768e-01
-3.59266177e-02 -7.60644138e-01 -2.55999148e-01 -2.04883233e-01
-9.57238972e-01 7.90555418e-01 4.44338620e-01 1.65559217e-01
5.87310791e-01 3.61708105e-01 7.72598028e-01 3.57400686e-01
6.50317788e-01 -1.12077188e+00 2.79597223e-01 8.08718324e-01
3.72270167e-01 -1.53425157e+00 -4.41558799e-03 -3.23908836e-01
-8.34646225e-01 6.05803072e-01 8.39384735e-01 -2.59909928e-01
7.03872919e-01 5.07931650e-01 2.46066060e-02 -3.96112800e-01
-6.14311658e-02 -6.20143592e-01 -1.34349540e-01 1.07681918e+00
2.97921985e-01 9.83446315e-02 1.95546508e-01 1.35786623e-01
-2.76632488e-01 1.64740101e-01 3.12119752e-01 9.47360337e-01
-9.18952107e-01 -6.54593647e-01 -7.60153472e-01 6.98590040e-01
-1.59510761e-01 2.92379819e-02 -2.74751484e-01 4.98966098e-01
4.27765220e-01 9.95875657e-01 4.69240725e-01 -1.94488958e-01
2.66802967e-01 -2.37335488e-01 5.12637198e-01 -4.63072598e-01
-1.88022077e-01 -2.96991378e-01 -1.09139815e-01 -1.19870476e-01
-7.30019987e-01 -6.67513549e-01 -1.38252616e+00 -8.44203711e-01
-3.37668180e-01 2.44133249e-01 3.65000546e-01 8.66674125e-01
-3.39545570e-02 -5.34760952e-02 4.53999132e-01 -1.47160387e+00
-1.28579229e-01 -1.10958505e+00 -5.76713026e-01 3.37853044e-01
1.38331562e-01 -9.70661998e-01 -4.40847218e-01 -1.44362196e-01] | [8.365384101867676, -0.34966716170310974] |
cdac7f2f-4083-4006-b8d7-9be0dd8bdb0b | online-nonnegative-matrix-factorization-with | 1608.00075 | null | http://arxiv.org/abs/1608.00075v2 | http://arxiv.org/pdf/1608.00075v2.pdf | Online Nonnegative Matrix Factorization with General Divergences | We develop a unified and systematic framework for performing online
nonnegative matrix factorization under a wide variety of important divergences.
The online nature of our algorithm makes it particularly amenable to
large-scale data. We prove that the sequence of learned dictionaries converges
almost surely to the set of critical points of the expected loss function. We
do so by leveraging the theory of stochastic approximations and projected
dynamical systems. This result substantially generalizes the previous results
obtained only for the squared-$\ell_2$ loss. Moreover, the novel techniques
involved in our analysis open new avenues for analyzing similar matrix
factorization problems. The computational efficiency and the quality of the
learned dictionary of our algorithm are verified empirically on both synthetic
and real datasets. In particular, on the tasks of topic learning, shadow
removal and image denoising, our algorithm achieves superior trade-offs between
the quality of learned dictionary and running time over the batch and other
online NMF algorithms. | ['Renbo Zhao', 'Vincent Y. F. Tan', 'Huan Xu'] | 2016-07-30 | null | null | null | null | ['shadow-removal'] | ['computer-vision'] | [-3.70411351e-02 -2.20700428e-01 1.37961814e-02 -8.19464326e-02
-8.90230536e-01 -7.48437107e-01 2.03644454e-01 -1.03669509e-01
-2.89732486e-01 4.39660102e-01 1.67738870e-01 -2.57711887e-01
-3.83787602e-01 -4.77586329e-01 -7.01822639e-01 -1.10982549e+00
-2.01659247e-01 2.08932668e-01 -2.55411416e-01 -2.80709326e-01
3.25604044e-02 2.89506704e-01 -1.29385495e+00 -3.54307383e-01
9.69688892e-01 1.05161154e+00 1.67212889e-01 6.25296712e-01
1.88456461e-01 5.02833903e-01 -1.62514895e-01 -4.76050556e-01
7.60655761e-01 -4.10162836e-01 -2.68496990e-01 4.95802134e-01
4.05854076e-01 -1.31055266e-01 -4.31827992e-01 1.38360596e+00
6.97814047e-01 2.69156367e-01 5.38040996e-01 -1.07071221e+00
-5.36984384e-01 3.55957508e-01 -7.76463747e-01 3.14713895e-01
1.34615013e-02 -1.09757118e-01 1.24807692e+00 -1.03926992e+00
5.23314595e-01 9.74124253e-01 9.42658365e-01 2.29608536e-01
-1.42016828e+00 -5.77271760e-01 1.91100627e-01 -2.16187030e-01
-1.21929133e+00 -5.27342618e-01 5.10710478e-01 -6.03147030e-01
3.83308381e-01 2.19676241e-01 5.44550538e-01 6.43920839e-01
2.20903188e-01 7.43833005e-01 1.03367305e+00 -5.87128162e-01
3.15403670e-01 1.42547622e-01 1.63156599e-01 8.85982513e-01
3.52109313e-01 5.91807887e-02 -6.81996286e-01 -4.68169063e-01
6.62213564e-01 2.05229372e-01 -3.59561265e-01 -6.92910552e-01
-1.06693482e+00 8.70165050e-01 -1.25723064e-01 1.01489395e-01
-3.26786935e-01 2.26714596e-01 3.90339941e-01 5.76364458e-01
8.12544823e-01 1.87095463e-01 -4.01273042e-01 2.46631727e-02
-9.42124546e-01 1.28468379e-01 9.14286971e-01 7.91215301e-01
5.87396920e-01 2.76512265e-01 1.11419730e-01 7.86158144e-01
5.15483283e-02 1.08271241e+00 1.11099690e-01 -1.04488099e+00
3.37211847e-01 -2.94993687e-02 3.07503492e-01 -1.23630548e+00
-2.11593255e-01 -6.89747810e-01 -1.03466117e+00 -9.11679342e-02
5.04034877e-01 -2.70613015e-01 -4.28922355e-01 1.93058157e+00
5.27631581e-01 2.57832915e-01 -1.92674920e-01 8.16675603e-01
-1.53246433e-01 6.30125821e-01 -4.10548151e-01 -8.15161526e-01
8.62051368e-01 -7.49797404e-01 -9.59974110e-01 3.27965021e-02
4.66129154e-01 -8.13946843e-01 1.07587111e+00 5.74323058e-01
-1.12727058e+00 -2.24804550e-01 -7.09003508e-01 2.31274873e-01
2.26983473e-01 2.99595874e-02 8.60213578e-01 6.63441002e-01
-1.06253266e+00 6.17585897e-01 -8.88435066e-01 -2.22378105e-01
1.79429516e-01 2.32501298e-01 -5.13634235e-02 -7.79407769e-02
-8.04114997e-01 3.12969297e-01 -2.56817877e-01 2.11081997e-01
-9.19938743e-01 -8.26503336e-01 -5.14018595e-01 -6.41735923e-03
3.57106477e-01 -7.62117684e-01 1.16066623e+00 -1.10710108e+00
-1.36728942e+00 7.57526338e-01 -3.59622657e-01 -5.74305892e-01
5.77007473e-01 -4.64135200e-01 -1.47325233e-01 1.25394091e-01
2.50368714e-01 -1.23577707e-01 1.27782810e+00 -8.44383061e-01
-4.05389667e-01 -4.49929088e-01 4.40213010e-02 8.35084692e-02
-8.75340104e-01 -1.91186413e-01 -3.33899349e-01 -1.02495790e+00
2.17689816e-02 -1.14774466e+00 -5.74181736e-01 1.81012303e-01
-5.19517027e-02 1.56243533e-01 3.28126878e-01 -6.97489977e-01
1.32208037e+00 -2.27599192e+00 4.84368861e-01 3.42839181e-01
2.55176693e-01 -1.16539948e-01 2.54765321e-02 6.30891204e-01
-9.82655734e-02 -1.69568941e-01 -3.79951000e-01 -4.75099593e-01
6.13662750e-02 1.02691121e-01 -8.54543507e-01 9.84491229e-01
-4.58572417e-01 5.56717753e-01 -8.07564676e-01 -3.84881132e-04
-1.10723607e-01 2.54667670e-01 -6.98827684e-01 -2.78299581e-02
-3.98157816e-03 2.47744218e-01 -3.17558140e-01 2.88976312e-01
6.00590110e-01 -4.63579953e-01 3.07256132e-01 -2.87342686e-02
1.37468904e-01 -1.38748422e-01 -1.51594639e+00 1.75374866e+00
-7.87713051e-01 6.19057000e-01 7.66876042e-01 -1.12404859e+00
3.37345034e-01 3.26918751e-01 7.06183255e-01 -4.06897485e-01
5.32290973e-02 2.16912046e-01 -3.92523140e-01 -2.09668905e-01
3.18470091e-01 -5.33463478e-01 -4.44297194e-02 9.01617467e-01
5.92869855e-02 1.86430484e-01 2.60865957e-01 6.46073580e-01
9.22702849e-01 -2.58397251e-01 2.51909554e-01 -7.90278316e-01
2.16746703e-01 -2.79253602e-01 5.47165513e-01 9.35804486e-01
-1.00385293e-01 2.82786816e-01 4.75995332e-01 -2.95381874e-01
-1.11254323e+00 -1.16375566e+00 -1.40540466e-01 1.16145313e+00
3.36221717e-02 -5.22587538e-01 -5.63045859e-01 -4.51772779e-01
2.69043326e-01 2.25707784e-01 -7.07139909e-01 -9.30877496e-03
-2.03797281e-01 -1.02238071e+00 4.69996892e-02 2.55657434e-01
9.69908573e-03 -3.53954166e-01 -8.93602371e-02 1.04746811e-01
-1.19955100e-01 -1.06570601e+00 -9.39633965e-01 5.59271239e-02
-1.16365135e+00 -1.02245498e+00 -5.68176985e-01 -4.80883956e-01
8.58517766e-01 7.73352921e-01 9.70762849e-01 -7.28748515e-02
-2.29179755e-01 8.42728496e-01 -1.63318679e-01 -2.69585133e-01
-1.45697773e-01 -1.44807935e-01 4.13084626e-01 4.38064337e-01
-1.59962639e-01 -8.78605068e-01 -6.54333055e-01 2.30211750e-01
-1.01149082e+00 -2.60988325e-01 2.67675072e-01 1.06252229e+00
8.33555520e-01 3.36152673e-01 4.88905072e-01 -9.90443051e-01
7.39463747e-01 -4.44702178e-01 -9.02148962e-01 1.56028003e-01
-7.59863853e-01 2.30732411e-01 8.12591076e-01 -4.03823942e-01
-7.99832463e-01 1.50407016e-01 1.87809199e-01 -4.76731718e-01
7.10658312e-01 3.29669982e-01 -5.04718386e-02 -8.92379209e-02
5.71735561e-01 3.08493495e-01 1.34927273e-01 -4.34357017e-01
6.61969960e-01 4.13420647e-01 4.72649604e-01 -6.41238213e-01
1.19475031e+00 1.12019229e+00 5.68086244e-02 -1.02585971e+00
-1.27241600e+00 -7.49413669e-01 -3.10004026e-01 -1.44857347e-01
3.32070023e-01 -1.08216572e+00 -6.99615538e-01 3.84667188e-01
-7.12475538e-01 -2.71959156e-01 -5.51726699e-01 2.75286496e-01
-7.39720345e-01 5.07218599e-01 -7.01583028e-01 -1.09843588e+00
-4.55765873e-01 -7.84562767e-01 1.04332900e+00 -2.27707252e-01
2.23092914e-01 -1.26392031e+00 4.48292524e-01 1.31897628e-01
2.59220153e-01 8.63629803e-02 6.24375701e-01 1.23638585e-02
-3.27788591e-01 -1.22409716e-01 -3.52638587e-03 5.89006484e-01
3.46714705e-02 -1.54173538e-01 -7.40814030e-01 -6.81744695e-01
5.36127269e-01 -5.54922707e-02 1.10523641e+00 5.96282601e-01
9.99467552e-01 -3.35794717e-01 -1.55490935e-01 8.09213281e-01
1.68102705e+00 -3.02828312e-01 3.26447010e-01 -1.10202366e-02
5.67735612e-01 3.39963555e-01 5.30991018e-01 8.25912416e-01
-4.31862734e-02 4.85266775e-01 1.63134575e-01 1.17418960e-01
2.77877837e-01 -1.86875835e-02 5.83296597e-01 1.20151877e+00
2.68241912e-01 -4.44472283e-02 -6.89765513e-01 8.03694069e-01
-1.95988345e+00 -9.95108962e-01 1.52841642e-01 2.47549319e+00
7.63736069e-01 -1.02443114e-01 3.40653926e-01 1.91307440e-01
6.59332514e-01 1.66770503e-01 -6.22264862e-01 -1.44068256e-01
-3.63019019e-01 3.22535604e-01 7.07052529e-01 8.13395560e-01
-1.13090098e+00 6.53300703e-01 6.66755962e+00 8.94619405e-01
-8.96417320e-01 3.78878415e-01 4.02799398e-01 -4.67415094e-01
-4.74047393e-01 -8.59560892e-02 -6.84005558e-01 3.18659753e-01
9.00825202e-01 -3.56058002e-01 8.49872351e-01 7.74575114e-01
3.26938272e-01 6.89599812e-02 -7.35688329e-01 1.08599460e+00
6.08022623e-02 -1.37826979e+00 -3.21035981e-01 1.64540216e-01
1.18706882e+00 -3.76890339e-02 3.84606957e-01 -1.46578714e-01
4.91980702e-01 -5.94960153e-01 7.33738005e-01 4.04091239e-01
7.00293183e-01 -8.90066922e-01 2.61514634e-01 2.97617257e-01
-1.00030696e+00 -4.18211520e-01 -3.57611477e-01 -2.12658957e-01
3.40808421e-01 1.28448665e+00 -1.44715771e-01 3.52050543e-01
4.98798072e-01 7.98368335e-01 -1.40743688e-01 8.52950394e-01
-3.76240015e-02 6.96443617e-01 -6.68037117e-01 1.85904965e-01
-7.61602744e-02 -6.10534608e-01 7.71427095e-01 9.82408583e-01
2.29617462e-01 4.40959409e-02 1.69429332e-01 3.59797388e-01
-2.81187922e-01 3.90792072e-01 -5.71146965e-01 -2.33605340e-01
2.61361927e-01 1.16830444e+00 -7.71421969e-01 -1.03894673e-01
-3.75269949e-01 1.02358091e+00 3.06318790e-01 4.44969982e-01
-6.91826284e-01 -2.18522117e-01 8.88074458e-01 1.67867839e-01
5.94457328e-01 -5.31628609e-01 -3.13106596e-01 -1.54720807e+00
2.87231207e-01 -1.04475737e+00 3.80372703e-01 -1.55281261e-01
-1.48785818e+00 2.00806186e-01 -2.66797543e-01 -1.05643809e+00
4.59227674e-02 -4.16091055e-01 -3.04453820e-01 3.81256431e-01
-1.06014383e+00 -5.62458992e-01 1.12821862e-01 6.29405260e-01
3.55798185e-01 -6.90875854e-03 6.11899853e-01 5.58574736e-01
-6.04830801e-01 4.60801989e-01 1.07599342e+00 -1.64041445e-01
5.17282367e-01 -1.14327276e+00 3.00673127e-01 1.20131850e+00
5.09673119e-01 6.47589207e-01 9.75592256e-01 -3.92812610e-01
-1.78179479e+00 -8.82646501e-01 4.40423369e-01 -3.70744795e-01
1.17101407e+00 -5.95958173e-01 -4.50274199e-01 5.46356261e-01
-5.45843914e-02 1.94212627e-02 8.60056043e-01 3.83993924e-01
-4.66188222e-01 -3.60345066e-01 -9.05664027e-01 5.13116419e-01
1.18005455e+00 -6.75113976e-01 5.05560525e-02 8.73340845e-01
6.17193341e-01 -2.21486077e-01 -6.88277364e-01 3.67205441e-02
6.17696226e-01 -7.83187151e-01 9.38408554e-01 -7.95051098e-01
2.92681843e-01 -4.00090590e-02 -5.35341501e-01 -1.12504625e+00
-1.52591377e-01 -1.20803165e+00 -4.27564293e-01 7.94966698e-01
1.98206559e-01 -5.84047377e-01 6.55011296e-01 2.84360737e-01
2.92966753e-01 -8.25218856e-01 -8.19468796e-01 -9.63169038e-01
1.19257428e-01 -6.13026977e-01 -3.75046432e-02 8.14001918e-01
-2.82620519e-01 1.54187843e-01 -6.71380162e-01 1.99972317e-01
1.11397052e+00 5.01726747e-01 6.86190367e-01 -1.09545171e+00
-8.40886533e-01 -1.59592599e-01 -8.59195516e-02 -1.26282358e+00
2.27822408e-01 -9.69809175e-01 -2.15411410e-01 -1.03711915e+00
3.36877435e-01 -4.38597411e-01 -1.97958395e-01 1.56183532e-02
-1.85413286e-01 2.00153291e-01 3.25055033e-01 3.13513398e-01
-7.75270581e-01 6.47166967e-01 9.70065653e-01 1.65222034e-01
-2.34123304e-01 1.42180711e-01 -8.27270210e-01 7.96511471e-01
4.16950554e-01 -6.80518508e-01 -4.83058453e-01 -6.27331495e-01
7.62727439e-01 9.06520262e-02 1.03642963e-01 -6.03389740e-01
1.75938219e-01 -1.44087151e-01 -1.56205758e-01 -2.72983074e-01
1.31827950e-01 -7.85041749e-01 8.62864628e-02 4.69054312e-01
-2.64155179e-01 7.89280012e-02 -6.31908178e-02 1.14250898e+00
3.60997580e-02 7.07988814e-02 9.35640395e-01 2.05255583e-01
-5.22763617e-02 4.87954974e-01 -3.17057073e-01 4.63515222e-01
8.90748858e-01 7.50171617e-02 2.27571726e-01 -8.61677945e-01
-9.66786921e-01 1.30670309e-01 4.31341201e-01 -2.46954113e-02
2.86792785e-01 -1.22721982e+00 -5.91505885e-01 2.34363765e-01
-3.30474705e-01 -3.49266469e-01 3.78194571e-01 1.12169921e+00
-3.07432204e-01 1.37264967e-01 3.86186302e-01 -4.72727418e-01
-1.03586388e+00 5.93221843e-01 3.16330701e-01 -4.12009448e-01
-5.46539485e-01 9.10167634e-01 3.39942724e-01 -2.19760314e-01
3.51622880e-01 1.73699893e-02 6.07144177e-01 1.37079075e-01
6.90149844e-01 5.96522450e-01 3.16267321e-03 -3.58268291e-01
-5.25875092e-02 5.09233415e-01 4.21192609e-02 -3.34722489e-01
1.59134328e+00 -5.30221760e-01 -3.63350064e-01 4.60106879e-01
1.28795564e+00 5.88308632e-01 -1.34124196e+00 -4.50289786e-01
-2.44162291e-01 -5.66590667e-01 3.64205167e-02 -1.78184062e-01
-1.34366500e+00 6.82887554e-01 6.55448735e-01 3.04682702e-01
1.33129478e+00 -2.41148785e-01 8.93376291e-01 4.68304724e-01
4.73837227e-01 -1.09129417e+00 1.18672505e-01 4.94127244e-01
6.50512040e-01 -1.00708270e+00 1.55645803e-01 -2.73894310e-01
-2.50386149e-01 8.10027421e-01 -1.25692889e-01 -4.52282369e-01
1.08605731e+00 4.30883646e-01 -8.35393369e-02 -7.04004839e-02
-8.04775000e-01 -4.48450446e-02 1.09696696e-02 2.97842115e-01
6.50266260e-02 9.01767835e-02 -4.51116145e-01 4.11816537e-01
-1.73293173e-01 -3.24870735e-01 5.58942080e-01 6.55193388e-01
-4.07179683e-01 -1.08328104e+00 -2.63569325e-01 4.30736989e-01
-6.88940823e-01 -2.75018096e-01 -3.70590203e-02 3.32395524e-01
-3.62340719e-01 7.79203296e-01 -2.71788865e-01 -1.39350414e-01
-4.97907512e-02 -5.06738611e-02 5.57216704e-01 -4.56559360e-01
-2.93383777e-01 2.41289869e-01 -3.52939188e-01 -6.05390429e-01
-4.14605230e-01 -9.64726865e-01 -7.95899391e-01 -6.24305546e-01
-6.60891593e-01 2.54167348e-01 4.92981076e-01 8.27529430e-01
4.23803031e-01 3.16963643e-02 9.97903764e-01 -5.82417369e-01
-1.00473297e+00 -3.77119929e-01 -9.42333460e-01 3.82482499e-01
3.85706991e-01 -4.23553228e-01 -7.69681394e-01 1.26892611e-01] | [7.071174144744873, 4.51841926574707] |
eb6672cf-ef1f-480f-a239-2e028d77d2bc | learnable-distribution-calibration-for-few | 2210.00232 | null | https://arxiv.org/abs/2210.00232v1 | https://arxiv.org/pdf/2210.00232v1.pdf | Learnable Distribution Calibration for Few-Shot Class-Incremental Learning | Few-shot class-incremental learning (FSCIL) faces challenges of memorizing old class distributions and estimating new class distributions given few training samples. In this study, we propose a learnable distribution calibration (LDC) approach, with the aim to systematically solve these two challenges using a unified framework. LDC is built upon a parameterized calibration unit (PCU), which initializes biased distributions for all classes based on classifier vectors (memory-free) and a single covariance matrix. The covariance matrix is shared by all classes, so that the memory costs are fixed. During base training, PCU is endowed with the ability to calibrate biased distributions by recurrently updating sampled features under the supervision of real distributions. During incremental learning, PCU recovers distributions for old classes to avoid `forgetting', as well as estimating distributions and augmenting samples for new classes to alleviate `over-fitting' caused by the biased distributions of few-shot samples. LDC is theoretically plausible by formatting a variational inference procedure. It improves FSCIL's flexibility as the training procedure requires no class similarity priori. Experiments on CUB200, CIFAR100, and mini-ImageNet datasets show that LDC outperforms the state-of-the-arts by 4.64%, 1.98%, and 3.97%, respectively. LDC's effectiveness is also validated on few-shot learning scenarios. | ['Qixiang Ye', 'Qi Tian', 'Ren Wang', 'Lingxi Xie', 'Boyu Yang', 'Binghao Liu'] | 2022-10-01 | null | null | null | null | ['few-shot-class-incremental-learning'] | ['methodology'] | [ 7.92799592e-02 1.83040395e-01 -3.49113524e-01 -3.72662276e-01
-9.46877480e-01 -2.65763909e-01 6.13720298e-01 5.30925393e-02
-4.94541496e-01 9.67248499e-01 -1.70108080e-01 1.79664776e-01
2.74948403e-02 -9.20410812e-01 -9.65222657e-01 -1.01969779e+00
6.81784302e-02 6.75685465e-01 2.60947734e-01 1.25032604e-01
-6.37781322e-02 2.07955595e-02 -1.83664787e+00 1.43824503e-01
7.52378345e-01 1.00871813e+00 3.91380161e-01 3.80328745e-01
-2.19571173e-01 7.55332232e-01 -5.18133998e-01 -5.74129522e-01
-5.23006134e-02 -3.41921598e-01 -3.41851681e-01 5.21419793e-02
2.78718293e-01 -4.39606339e-01 -1.44121885e-01 1.10671484e+00
4.74849552e-01 3.73493791e-01 8.86904240e-01 -1.35240042e+00
-8.66008222e-01 7.81294346e-01 -4.82207417e-01 5.13612963e-02
-2.93233752e-01 1.20308980e-01 7.12646782e-01 -1.12148714e+00
5.11627436e-01 1.18110120e+00 6.26187027e-01 9.31433260e-01
-1.27452230e+00 -7.06347466e-01 3.76021981e-01 3.12018394e-01
-1.52624369e+00 -5.05401611e-01 7.32729852e-01 -4.62474227e-01
5.13343453e-01 -1.22914366e-01 4.18369919e-01 1.44820857e+00
-1.22138364e-02 1.05136824e+00 7.27161288e-01 -4.16253924e-01
9.36915517e-01 4.96762276e-01 4.67212409e-01 4.22698587e-01
3.67916912e-01 6.30146787e-02 -5.65263867e-01 -2.40005702e-01
5.01014650e-01 2.60150105e-01 -2.09794864e-01 -7.45360613e-01
-7.33267426e-01 9.52591181e-01 1.30608484e-01 4.38071676e-02
-2.05113843e-01 9.83463079e-02 2.94696659e-01 2.11159717e-02
6.59873724e-01 -1.20767459e-01 -3.54881018e-01 -4.14211005e-02
-8.92564476e-01 1.67470753e-01 5.54783583e-01 1.31220269e+00
1.09679711e+00 2.70831913e-01 -5.01847804e-01 1.12842429e+00
1.80642545e-01 8.78987312e-01 6.93525553e-01 -8.70478392e-01
3.05524588e-01 2.37032726e-01 3.76066156e-02 -6.00726843e-01
-2.63064913e-02 -6.24018967e-01 -1.03451502e+00 -7.98652172e-02
2.47736186e-01 -1.64055392e-01 -1.05322480e+00 2.08437681e+00
3.26335669e-01 6.12040937e-01 8.35249647e-02 4.32549149e-01
5.59684455e-01 7.09091842e-01 1.07565351e-01 -3.08583140e-01
9.63263214e-01 -7.10722327e-01 -6.38194799e-01 -2.58899093e-01
2.79515177e-01 -9.90610942e-02 1.26990604e+00 3.22607517e-01
-8.18832219e-01 -7.67428815e-01 -1.07340443e+00 2.45014668e-01
-3.30869198e-01 5.65048903e-02 4.25845981e-01 8.69612813e-01
-7.27715194e-01 6.94908023e-01 -1.01003480e+00 5.85232563e-02
9.18562353e-01 -1.10076971e-01 -2.72341426e-02 -2.67570138e-01
-1.11529279e+00 5.81236422e-01 4.99548495e-01 -2.15513736e-01
-1.20445657e+00 -1.28474820e+00 -8.88835073e-01 3.13445359e-01
3.50435704e-01 -5.35651326e-01 1.28587008e+00 -7.71339715e-01
-1.66280448e+00 4.09388900e-01 -1.26492679e-01 -7.36524105e-01
5.93638301e-01 -2.06840187e-01 -1.55441478e-01 1.16321612e-02
7.16802031e-02 7.29097128e-01 1.15372968e+00 -1.23253798e+00
-5.22963762e-01 -2.50762671e-01 -2.82926589e-01 2.21984871e-02
-4.72290367e-01 -8.76323044e-01 -5.82913458e-01 -6.45292819e-01
-9.59473252e-02 -8.06962132e-01 1.19954757e-01 9.15526748e-02
-1.14723407e-01 -1.64672747e-01 7.77636528e-01 -1.96086943e-01
1.22363448e+00 -2.29690957e+00 7.44403452e-02 5.07099926e-02
6.46056309e-02 2.78300762e-01 -2.60040313e-01 -6.09787274e-03
6.80456534e-02 -3.48430485e-01 -6.11840367e-01 -7.43074179e-01
1.69042051e-01 5.21484792e-01 -6.14360690e-01 4.21415001e-01
2.25633070e-01 7.79147565e-01 -1.07611179e+00 -2.53030032e-01
3.12398404e-01 6.91239655e-01 -6.86908066e-01 1.13425203e-01
-3.15953225e-01 1.82636127e-01 1.16198912e-01 6.38704419e-01
9.03835058e-01 -3.12543750e-01 3.05994600e-01 1.25752063e-03
3.66281837e-01 -2.80887753e-01 -1.37867367e+00 1.65061617e+00
-4.86081719e-01 3.47345948e-01 -3.33109021e-01 -1.09688938e+00
9.60744441e-01 4.65194434e-02 9.92834866e-02 -4.93779331e-01
3.13692451e-01 -5.88677004e-02 -4.43450332e-01 -6.53805211e-02
1.39243916e-01 -2.79259115e-01 -1.90174486e-02 4.76775169e-01
6.84368372e-01 7.53978314e-03 9.29963738e-02 3.18517745e-01
6.34056926e-01 7.29907677e-02 1.73379466e-01 -1.78226233e-01
4.38085884e-01 -5.69977045e-01 7.71014869e-01 1.05158877e+00
-2.01002762e-01 5.44509053e-01 3.02623987e-01 -3.09845030e-01
-8.70914578e-01 -1.58559883e+00 -2.99485058e-01 1.12178087e+00
-3.07536554e-02 -2.16772020e-01 -7.95875013e-01 -5.61893821e-01
6.26285300e-02 1.25722826e+00 -8.60498130e-01 -5.62346458e-01
-1.35665119e-01 -8.65429640e-01 -3.80718289e-03 5.84076762e-01
4.88082767e-01 -7.51122296e-01 -6.34862542e-01 3.50115895e-01
-6.78071082e-02 -7.49340773e-01 -3.72956961e-01 2.40277454e-01
-8.30557168e-01 -1.13288248e+00 -9.34041560e-01 -5.58523417e-01
7.26325333e-01 3.02363634e-01 9.56534922e-01 -3.74352753e-01
-3.34900707e-01 5.82910359e-01 -2.69211084e-01 -4.98675466e-01
-1.75476387e-01 1.39045924e-01 2.36827925e-01 3.33705604e-01
3.76218110e-01 -7.68959463e-01 -3.25031549e-01 1.35559626e-02
-8.64791214e-01 -7.19344765e-02 4.19094503e-01 1.15196717e+00
8.01070750e-01 4.06776033e-02 7.70558059e-01 -1.37013865e+00
2.90014863e-01 -8.36428523e-01 -6.29292548e-01 3.92275691e-01
-7.35578775e-01 1.63660586e-01 5.71058512e-01 -9.08278346e-01
-1.36069870e+00 -1.36091501e-01 1.93397462e-01 -8.83981586e-01
1.17291212e-01 1.61415920e-01 -3.45661253e-01 3.32400888e-01
6.52064562e-01 4.25215006e-01 6.39492571e-02 -5.33912420e-01
6.76208377e-01 5.53672731e-01 7.82538831e-01 -8.67099762e-01
7.31122136e-01 4.24739748e-01 -4.16562527e-01 -7.89833069e-01
-1.25587606e+00 -2.30622977e-01 -7.38387883e-01 -1.95705280e-01
4.39963460e-01 -1.23650146e+00 -2.97111303e-01 8.00494432e-01
-7.73209810e-01 -5.47980905e-01 -7.49620736e-01 4.07324255e-01
-6.10289276e-01 1.82664003e-02 -4.24188823e-01 -9.00458574e-01
-2.57883728e-01 -7.85484314e-01 8.42016637e-01 3.17209691e-01
1.21223316e-01 -9.73628342e-01 2.06628010e-01 -7.86121860e-02
4.76817667e-01 -3.68009321e-02 9.68143880e-01 -5.02789974e-01
-4.41758364e-01 -9.61954817e-02 -6.54093921e-02 6.79404557e-01
1.61018312e-01 -2.17378423e-01 -1.43384874e+00 -5.51523626e-01
-1.41469305e-02 -4.96751428e-01 1.06725526e+00 4.22100663e-01
1.40516281e+00 -2.33390808e-01 -2.30816424e-01 6.48323357e-01
1.34563959e+00 1.37171164e-01 5.30214310e-01 -1.27325565e-01
4.59171027e-01 3.36889118e-01 4.66908514e-01 8.25397491e-01
3.77128005e-01 1.90909684e-01 1.77327156e-01 6.60891414e-01
-3.10922831e-01 -4.73444611e-01 1.82386607e-01 7.41144300e-01
2.76962876e-01 6.25306368e-02 -7.62866735e-01 5.26779056e-01
-1.85509026e+00 -1.17177832e+00 4.78881091e-01 2.36633420e+00
1.15902090e+00 2.58871853e-01 -1.22732520e-01 1.09367341e-01
8.24152946e-01 1.09423123e-01 -1.02954185e+00 2.71036595e-01
-3.43356170e-02 4.12628800e-01 2.69895494e-01 6.08635545e-01
-1.10557687e+00 7.03268528e-01 5.62264061e+00 1.03917432e+00
-8.37153435e-01 3.66420805e-01 8.49730313e-01 -3.58457267e-01
-2.58778423e-01 -2.13876724e-01 -1.19679832e+00 7.65057862e-01
1.09595418e+00 -2.42901593e-01 4.55826133e-01 1.19423902e+00
-5.32337725e-01 -1.35306954e-01 -1.07300365e+00 1.22649205e+00
3.05390686e-01 -1.55665338e+00 1.27066433e-01 -3.92407000e-01
8.60478163e-01 -1.14334464e-01 3.98528039e-01 1.07141936e+00
3.39410692e-01 -4.13944572e-01 8.92085612e-01 8.38097155e-01
9.52618480e-01 -9.99650955e-01 6.44757032e-01 5.95541894e-01
-9.60216761e-01 -2.50721306e-01 -8.04455221e-01 1.16141811e-01
-1.88828662e-01 8.94008934e-01 -6.23578012e-01 1.63902417e-01
6.54992819e-01 8.94608498e-01 -6.65167570e-01 8.37692022e-01
-1.76431865e-01 7.60751069e-01 -2.24552318e-01 1.20524682e-01
-1.03782564e-01 -1.30862921e-01 1.28339931e-01 8.88074338e-01
2.93152303e-01 6.21457770e-02 7.04459324e-02 9.43861842e-01
-2.02242792e-01 -2.38409936e-01 -1.66020051e-01 2.80893236e-01
1.01308167e+00 8.70437860e-01 -5.55762470e-01 -7.16993511e-01
-1.51297703e-01 9.48854029e-01 6.23446882e-01 4.51782227e-01
-1.04493773e+00 -3.76258254e-01 5.96038818e-01 -1.26954183e-01
5.96147537e-01 1.35435924e-01 1.18460404e-02 -1.37363935e+00
-1.65346682e-01 -4.07396913e-01 3.54672849e-01 -6.11516953e-01
-1.60512781e+00 4.23803180e-01 3.24220389e-01 -1.16200471e+00
-2.55362868e-01 -3.49269301e-01 -6.12792313e-01 6.24761999e-01
-1.61483228e+00 -8.59176993e-01 -3.38609338e-01 6.14965081e-01
6.21127367e-01 -3.80053848e-01 8.04283798e-01 3.11822534e-01
-7.31949627e-01 9.06953812e-01 4.96677607e-01 -8.94117728e-02
6.98387265e-01 -1.09107947e+00 1.97196171e-01 4.32447135e-01
6.36964440e-02 5.40357292e-01 5.04358590e-01 -5.16579926e-01
-1.07335234e+00 -1.43030202e+00 4.97304320e-01 -4.78071094e-01
5.81936717e-01 -6.30270004e-01 -1.28449595e+00 6.20977521e-01
-2.68914998e-01 3.85734707e-01 8.81634474e-01 -3.08243427e-02
-7.09511280e-01 -3.93317968e-01 -1.19961309e+00 2.28566378e-01
8.45696032e-01 -6.41956866e-01 -5.90223849e-01 2.29551136e-01
6.86953425e-01 -2.90441871e-01 -6.17242277e-01 -5.58021478e-03
4.30758387e-01 -7.72730172e-01 1.06726921e+00 -4.12753016e-01
6.85579260e-04 6.60415217e-02 -4.15306836e-01 -1.33129859e+00
-2.89746553e-01 -2.34205931e-01 -6.55730426e-01 1.60016716e+00
1.68362215e-01 -5.75354159e-01 7.48351216e-01 6.11410797e-01
5.96777014e-02 -6.54103041e-01 -9.57460403e-01 -1.00692880e+00
2.96490669e-01 -6.16636276e-01 7.18656242e-01 7.92765141e-01
-4.41743881e-01 3.89271826e-01 -3.92700553e-01 -1.93595551e-02
1.06072414e+00 2.00732164e-02 7.04275846e-01 -1.37117362e+00
-4.07997489e-01 -1.88920766e-01 2.34979279e-02 -7.10007966e-01
4.77434486e-01 -8.75216186e-01 1.09397560e-01 -9.29698825e-01
4.71390307e-01 -4.63729054e-01 -4.34198886e-01 5.31435072e-01
-3.34811956e-01 -4.11791801e-02 2.69266933e-01 2.00088337e-01
-7.29112566e-01 1.10547912e+00 7.65873313e-01 -3.20831507e-01
-1.72295645e-01 1.56205416e-01 -5.91801107e-01 6.18664563e-01
7.51707137e-01 -5.03324747e-01 -9.78702664e-01 -2.40598783e-01
-2.58656472e-01 -2.40429878e-01 3.39768559e-01 -1.12991714e+00
2.53562003e-01 -2.46150270e-02 6.03415549e-01 -7.43443489e-01
3.60033631e-01 -5.44899940e-01 4.37757634e-02 3.58728290e-01
-3.09675992e-01 -6.84258699e-01 1.64308637e-01 1.05663824e+00
5.68439029e-02 -3.31466496e-01 1.22738075e+00 -6.72029927e-02
-7.14443564e-01 4.34700340e-01 -1.91230431e-01 3.70980263e-01
1.09401155e+00 2.43613608e-02 -2.56590456e-01 -1.79249048e-01
-7.79631495e-01 2.36251593e-01 1.66183978e-01 2.98005879e-01
6.99599028e-01 -1.51383698e+00 -3.97302568e-01 5.64533174e-01
2.65576154e-01 2.73498803e-01 6.47070408e-01 3.06944102e-01
1.14123516e-01 3.06857731e-02 -7.33613968e-03 -6.46492243e-01
-7.42095292e-01 7.87285149e-01 1.31455243e-01 1.13049019e-02
-5.69044113e-01 1.06416428e+00 1.17090017e-01 -5.74271083e-01
6.36310816e-01 -2.26942301e-01 -4.20223102e-02 3.73518497e-01
9.32324111e-01 4.11095560e-01 -1.20228238e-01 -1.34174556e-01
-1.51161626e-01 2.36124203e-01 -2.64442503e-01 6.83550835e-02
1.45995319e+00 -2.27397472e-01 3.28543127e-01 9.79533494e-01
1.16831148e+00 -5.78715861e-01 -1.87380195e+00 -7.36717820e-01
-1.42785624e-01 -3.80134702e-01 3.59779671e-02 -7.62635946e-01
-1.10858405e+00 1.16945338e+00 7.48841643e-01 -4.13002074e-01
8.64553332e-01 -7.95462579e-02 6.14075541e-01 4.02643561e-01
7.15942442e-01 -1.15495789e+00 2.81782299e-01 5.59253097e-01
5.63362479e-01 -1.16542029e+00 -8.60175416e-02 6.33518025e-02
-7.86368489e-01 8.89083445e-01 7.48039424e-01 -1.12330712e-01
1.07123983e+00 3.80625948e-02 -3.68122399e-01 2.68805116e-01
-9.21641886e-01 -4.08879332e-02 1.45211861e-01 7.50858307e-01
-9.19817984e-02 6.39323145e-02 3.67974639e-01 1.11282682e+00
3.95948701e-02 1.50044747e-02 4.05364305e-01 8.22139204e-01
-6.71619296e-01 -8.73255551e-01 -1.55181378e-01 3.64481568e-01
2.94765264e-01 -4.40197513e-02 3.25910658e-01 6.92660093e-01
3.09882581e-01 5.66129684e-01 5.08839011e-01 -9.22808051e-02
5.68209440e-02 3.72649789e-01 4.75778699e-01 -7.93364882e-01
2.21014485e-01 6.05571419e-02 -4.92364287e-01 -3.61615181e-01
-1.67270094e-01 -6.84211850e-01 -1.07652414e+00 -1.93461150e-01
-2.74875522e-01 1.82594329e-01 3.64359528e-01 7.82082319e-01
3.87929708e-01 5.23415685e-01 6.71830893e-01 -9.24844086e-01
-1.06164443e+00 -9.35091853e-01 -8.90780866e-01 2.91735381e-01
3.05219382e-01 -1.09562445e+00 -5.15470743e-01 7.68646821e-02] | [9.856146812438965, 3.242680311203003] |
3aab6d5c-0b40-4490-a599-d2ad2fd38770 | cebed-a-benchmark-for-deep-data-driven-ofdm | 2306.13761 | null | https://arxiv.org/abs/2306.13761v1 | https://arxiv.org/pdf/2306.13761v1.pdf | CeBed: A Benchmark for Deep Data-Driven OFDM Channel Estimation | Deep learning has been extensively used in wireless communication problems, including channel estimation. Although several data-driven approaches exist, a fair and realistic comparison between them is difficult due to inconsistencies in the experimental conditions and the lack of a standardized experimental design. In addition, the performance of data-driven approaches is often compared based on empirical analysis. The lack of reproducibility and availability of standardized evaluation tools (e.g., datasets, codebases) hinder the development and progress of data-driven methods for channel estimation and wireless communication in general. In this work, we introduce an initiative to build benchmarks that unify several data-driven OFDM channel estimation approaches. Specifically, we present CeBed (a testbed for channel estimation) including different datasets covering various systems models and propagation conditions along with the implementation of ten deep and traditional baselines. This benchmark considers different practical aspects such as the robustness of the data-driven models, the number and the arrangement of pilots, and the number of receive antennas. This work offers a comprehensive and unified framework to help researchers evaluate and design data-driven channel estimation algorithms. | ['Greg Dudek', 'Steve Liu', 'Di wu', 'Amal Feriani'] | 2023-06-23 | null | null | null | null | ['experimental-design'] | ['methodology'] | [-1.12570867e-01 -5.62926054e-01 -4.48860899e-02 -4.37479138e-01
-9.05096412e-01 -1.85782224e-01 4.05782908e-01 9.49137881e-02
-4.18953657e-01 1.10117996e+00 3.63700241e-02 -7.76055515e-01
-1.85614452e-01 -7.23329723e-01 -4.95216370e-01 -1.09346676e+00
-9.17823017e-01 2.03982703e-02 -3.19870561e-01 -2.91809887e-01
3.67805399e-02 3.47225755e-01 -1.02520514e+00 -2.34880894e-01
2.77707607e-01 1.05492663e+00 1.81590363e-01 1.08780503e+00
2.74125785e-01 3.82200569e-01 -8.75383079e-01 -1.70027122e-01
6.19631335e-02 -6.46609604e-01 -3.47395837e-01 -1.33429676e-01
-4.69023399e-02 -6.40521407e-01 -6.22604430e-01 6.22201085e-01
1.33882463e+00 -4.78291243e-01 4.66559112e-01 -1.44270682e+00
1.74891844e-01 7.54146099e-01 -1.47657126e-01 1.08734109e-02
2.17537448e-01 -1.10870138e-01 7.68422008e-01 -6.77960098e-01
4.75913286e-01 1.06112278e+00 1.02409124e+00 1.70389459e-01
-1.15894353e+00 -9.45474386e-01 -2.45871633e-01 1.13759048e-01
-1.59522009e+00 -7.78738558e-01 2.35252589e-01 -2.91203827e-01
6.05713844e-01 7.95943886e-02 5.87563694e-01 1.47482979e+00
1.67558953e-01 5.69667995e-01 9.40666139e-01 -6.90124094e-01
4.97661412e-01 3.04653794e-01 -5.66636994e-02 4.23881292e-01
4.10254806e-01 4.33296919e-01 -3.90620261e-01 -3.25008243e-01
5.99019647e-01 -6.86461687e-01 -2.78427780e-01 -5.65384686e-01
-9.98838127e-01 6.12514734e-01 1.20729171e-01 1.96373716e-01
-1.65069923e-01 6.92328155e-01 5.44626713e-01 6.35401189e-01
-5.20921052e-02 3.25431198e-01 -4.30872232e-01 -4.24731404e-01
-1.08673394e+00 5.26395440e-01 1.17225718e+00 1.31110418e+00
4.98918980e-01 3.33065391e-01 -2.21907154e-01 3.75636160e-01
6.15053713e-01 7.01925397e-01 -1.23838641e-01 -8.16925168e-01
6.98157787e-01 -5.92194617e-01 -4.23026718e-02 -1.18482316e+00
-6.43910050e-01 -9.86860335e-01 -1.09041584e+00 -1.71915352e-01
5.81645548e-01 -1.02922928e+00 -3.16252321e-01 1.57475293e+00
-4.47290331e-01 1.52462021e-01 3.89664739e-01 7.65350699e-01
6.64764702e-01 4.73710209e-01 -1.86713994e-01 -1.05597973e-01
7.28802800e-01 -7.50931680e-01 -5.99711657e-01 1.09958194e-01
1.04005063e+00 -8.20758402e-01 4.21195090e-01 5.19201159e-01
-8.43823373e-01 -3.89877588e-01 -1.46183681e+00 5.32229245e-01
-2.16057003e-01 1.05396193e-02 6.57972217e-01 1.48462963e+00
-9.95000005e-01 4.05357182e-01 -5.26416898e-01 -4.33669806e-01
3.12574565e-01 5.12773097e-01 -8.07374045e-02 -3.16835418e-02
-1.49277437e+00 4.60706055e-01 3.66401881e-01 1.44537479e-01
-1.09408092e+00 -4.34253454e-01 -7.05005348e-01 2.33965635e-01
-6.82987943e-02 -2.20888779e-01 1.46435928e+00 -5.26060581e-01
-1.43542111e+00 3.10907215e-02 1.92210719e-01 -8.47984791e-01
5.60812414e-01 -1.10627346e-01 -7.11051404e-01 -3.93504620e-01
-3.35801840e-01 2.47870967e-01 4.98503804e-01 -1.15405953e+00
-5.93295693e-01 1.11308895e-01 -1.80326268e-01 -4.38903980e-02
-3.29170614e-01 -2.72913992e-01 -8.37620497e-01 -6.16079628e-01
-2.34839708e-01 -9.46870208e-01 -6.41481698e-01 -2.88916081e-01
-6.77421629e-01 8.50285709e-01 6.52937472e-01 -5.23291290e-01
1.56897330e+00 -2.28474617e+00 -3.06468546e-01 7.96944857e-01
9.83872712e-02 3.69064584e-02 -1.29159898e-01 8.59246671e-01
1.71833187e-01 1.25726193e-01 6.77316338e-02 -2.51368284e-01
2.02511892e-01 4.03531380e-02 1.41519174e-01 6.54756546e-01
-3.72351378e-01 4.09224600e-01 -7.89761305e-01 -1.77780449e-01
3.44412953e-01 5.81444263e-01 -7.78295517e-01 2.62273341e-01
2.72003889e-01 6.47159576e-01 -4.59549159e-01 4.85714674e-01
9.32670653e-01 -1.20745845e-01 4.76509064e-01 -1.12707302e-01
-2.26478681e-01 4.59836602e-01 -1.38190496e+00 1.39430010e+00
-9.67750788e-01 1.28728187e+00 1.54694304e-01 -1.08554935e+00
6.91225052e-01 6.84939146e-01 4.58300561e-01 -1.05451810e+00
3.79006118e-01 5.60232580e-01 3.25942695e-01 -2.32846811e-01
3.87505710e-01 5.28850734e-01 -2.74289191e-01 3.31817478e-01
1.91965371e-01 9.86935049e-02 1.90657064e-01 2.06690636e-02
1.00207543e+00 -3.10365081e-01 3.87781411e-01 -3.39867234e-01
4.75340396e-01 -3.46453220e-01 2.30558187e-01 1.20377600e+00
-1.84345543e-01 4.26693738e-01 6.83692992e-01 -2.43154213e-01
-1.12749708e+00 -8.24634612e-01 -5.57600319e-01 9.89799857e-01
3.05604756e-01 -6.46968246e-01 -6.82895005e-01 -1.29271895e-01
4.95918095e-02 5.48618972e-01 -4.28193748e-01 3.07498849e-03
-8.25928077e-02 -1.30866778e+00 1.10435081e+00 2.08556056e-01
7.99345016e-01 -4.92890656e-01 -3.05319339e-01 6.34724677e-01
2.71082763e-02 -1.29600489e+00 2.77478337e-01 7.11827755e-01
-3.29996586e-01 -6.54764354e-01 -8.22315753e-01 -5.03516853e-01
3.56686711e-01 1.37914822e-01 1.09400868e+00 9.81164277e-02
-7.21152052e-02 1.95599973e-01 -5.35085797e-01 -4.78784770e-01
-5.24731815e-01 4.48052764e-01 -1.13064624e-01 -3.81810844e-01
1.64685789e-02 -3.46875429e-01 -4.43472534e-01 4.25042063e-01
-2.69669056e-01 -2.56083757e-01 7.58871078e-01 8.31390500e-01
5.34168072e-02 3.56650710e-01 4.27906424e-01 -7.58876443e-01
7.96463907e-01 -7.80707598e-01 -6.82343781e-01 8.57411325e-02
-7.10106194e-01 2.43642062e-01 5.07131457e-01 1.82061344e-01
-7.87865043e-01 -3.89860958e-01 -7.44058490e-01 5.20042717e-01
-1.17510661e-01 7.51151800e-01 -3.38877350e-01 -4.92654473e-01
7.09319711e-01 5.17722704e-02 -1.85823128e-01 -2.40382910e-01
-7.75662661e-02 1.06666362e+00 2.08305735e-02 -6.18331075e-01
8.95903051e-01 4.69621979e-02 -4.12338413e-02 -1.28972757e+00
-4.14111465e-01 -2.79492795e-01 -3.72330278e-01 -6.11026138e-02
2.23890141e-01 -1.10801506e+00 -3.49188417e-01 6.06705725e-01
-9.61086333e-01 -6.85587585e-01 3.59805763e-01 7.41785944e-01
-4.65743184e-01 1.78457394e-01 -4.20481950e-01 -5.54245412e-01
-1.89105749e-01 -1.60644257e+00 8.41751456e-01 -1.77274987e-01
-9.31775048e-02 -1.25489378e+00 -3.35805789e-02 -2.00104803e-01
8.28443110e-01 2.93379992e-01 5.99656582e-01 -2.29740873e-01
-4.10334289e-01 -2.09369704e-01 -3.07452589e-01 2.06733331e-01
-2.05215111e-01 9.01337620e-03 -1.15739548e+00 -6.48521185e-01
-7.69741297e-01 -1.12591624e-01 4.23385710e-01 6.74947083e-01
1.39124668e+00 -2.03769971e-02 -4.71765220e-01 1.04323041e+00
1.44084227e+00 1.00120962e-01 9.33361948e-01 5.31817973e-01
1.31237909e-01 8.71074870e-02 4.56414223e-01 8.01059783e-01
3.12487602e-01 9.49630857e-01 4.80938584e-01 -4.52465355e-01
1.17608927e-01 2.55582392e-01 1.93301901e-01 5.59352636e-01
-1.15437259e-03 -7.85272539e-01 -9.07298386e-01 -9.26203746e-03
-1.29770267e+00 -8.51839602e-01 -4.87998486e-01 2.46154737e+00
3.85092139e-01 2.89557099e-01 -3.12983654e-02 5.57891726e-01
3.34587485e-01 -3.61548811e-02 -1.41848624e-01 -4.58627164e-01
-1.41553178e-01 7.18214139e-02 1.08583081e+00 3.91697735e-01
-1.27579665e+00 5.14704943e-01 7.14268875e+00 8.18984926e-01
-1.45928001e+00 -9.70187187e-02 6.36134923e-01 3.30859572e-01
1.19592108e-01 -2.06952184e-01 -7.92699873e-01 4.23476160e-01
1.34863472e+00 -1.86090112e-01 2.34225780e-01 7.36943603e-01
4.17312860e-01 -3.62791605e-02 -9.79310036e-01 1.01647854e+00
-3.80290031e-01 -1.49203646e+00 -3.75952601e-01 4.81908679e-01
6.94376945e-01 2.77063936e-01 1.66105315e-01 4.88593549e-01
-2.91559212e-02 -1.10534930e+00 7.52772272e-01 1.18618011e-01
7.93717444e-01 -7.32692361e-01 1.23773527e+00 -1.64753467e-01
-1.09728754e+00 -1.79166839e-01 9.17384997e-02 -2.84743220e-01
1.27002701e-01 6.91015720e-01 -8.82762253e-01 5.63316405e-01
5.82243741e-01 4.77247447e-01 -3.46018404e-01 1.47421885e+00
1.13927491e-01 9.47098494e-01 -1.11234881e-01 -1.73004776e-01
4.16610301e-01 3.38830858e-01 1.74264997e-01 1.71966398e+00
6.53245032e-01 -4.85812724e-01 1.70347244e-01 1.52001008e-01
1.93312466e-02 1.73873335e-01 -3.21515590e-01 1.73646182e-01
8.83531451e-01 1.05436134e+00 -2.62298614e-01 1.88587174e-01
-7.72522092e-01 5.11231363e-01 -3.51929396e-01 6.27663016e-01
-9.65893924e-01 -5.94960392e-01 8.47559512e-01 7.18193278e-02
8.35033208e-02 -8.72857153e-01 -1.63041934e-01 -5.50000310e-01
-3.56136829e-01 -9.10182416e-01 -6.46179989e-02 -9.65932235e-02
-6.58538938e-01 2.71637350e-01 -2.86343601e-02 -1.35078847e+00
-2.96617180e-01 -5.75521588e-01 -4.17404324e-01 7.19788194e-01
-1.66975510e+00 -3.74538809e-01 -6.73882961e-01 2.97688544e-01
1.19429812e-01 -4.40082252e-01 1.17788374e+00 1.02860355e+00
-7.27489889e-01 1.06932604e+00 7.41728246e-01 3.10127169e-01
7.02399373e-01 -9.89898562e-01 5.94224334e-01 5.08833885e-01
-1.56634629e-01 4.46871907e-01 1.02374530e+00 5.52826151e-02
-1.45652938e+00 -1.00477457e+00 2.69090891e-01 3.09880376e-01
4.66185480e-01 -7.12233901e-01 -3.81415814e-01 1.73049092e-01
2.07356006e-01 -2.85349935e-02 9.31444824e-01 3.43484730e-01
1.48743808e-01 -1.20465219e-01 -7.31096208e-01 6.03819370e-01
6.13507509e-01 -4.20514867e-02 8.83748949e-01 1.71604559e-01
2.64125280e-02 -5.60300171e-01 -8.73069227e-01 1.35004491e-01
9.81069088e-01 -1.28392303e+00 8.46203566e-01 8.07423666e-02
-6.41464964e-02 5.94341494e-02 -4.63260561e-01 -1.39249039e+00
-1.71123102e-01 -7.94913590e-01 -1.83962166e-01 9.25974846e-01
7.75779128e-01 -3.75641912e-01 8.90867472e-01 -1.51661813e-01
1.14198513e-01 -6.23893201e-01 -5.52483559e-01 -8.62980783e-01
3.26438099e-02 -8.57851744e-01 5.92188060e-01 6.29404545e-01
-2.51423746e-01 8.41173716e-03 -5.70776045e-01 4.38684255e-01
6.34424090e-01 -4.43021268e-01 1.28969240e+00 -1.02727675e+00
-3.82754445e-01 -3.69770169e-01 -7.61139452e-01 -1.04030252e+00
-1.08998172e-01 -4.50201273e-01 6.98625147e-02 -1.17563319e+00
-1.92893669e-01 -1.11326706e+00 -1.31425023e-01 3.58892418e-02
4.07849759e-01 1.35460421e-01 -1.52224615e-01 -1.42719716e-01
-4.20360506e-01 4.39794868e-01 8.62624586e-01 7.88211972e-02
-3.23440939e-01 2.83134937e-01 -5.80368936e-01 3.31241041e-01
1.41424847e+00 -1.82981804e-01 -3.50854754e-01 -5.01136780e-01
2.32350782e-01 1.40225619e-01 3.65451038e-01 -1.67961895e+00
1.48138538e-01 2.21856728e-01 4.36793059e-01 -1.30941957e-01
1.32816657e-01 -9.58160460e-01 2.09787846e-04 6.68362319e-01
-2.41359204e-01 -3.55204463e-01 2.24121049e-01 4.08470422e-01
-2.12241262e-01 6.87797442e-02 6.37489676e-01 3.60702872e-01
-9.70717490e-01 3.45285326e-01 -6.23432815e-01 1.28303394e-01
9.59414542e-01 -3.39709282e-01 -2.23006215e-02 -9.96379793e-01
-5.10163248e-01 1.97679758e-01 -2.94717122e-02 2.13991433e-01
1.48568317e-01 -1.15504682e+00 -7.60490656e-01 2.99733251e-01
2.79899538e-01 -5.90563476e-01 -1.78165939e-02 9.93858874e-01
-8.10216427e-01 7.00358272e-01 -2.73912132e-01 -5.05019069e-01
-1.11130357e+00 -1.93746299e-01 7.56569564e-01 -1.92737430e-01
-1.90396652e-01 6.59051538e-01 -2.93895036e-01 -4.02438700e-01
8.36376667e-01 -1.12336837e-01 1.21172704e-01 -2.22823977e-01
3.32935184e-01 3.91908109e-01 3.51948947e-01 -3.93916458e-01
-2.85198241e-01 9.17808414e-02 1.52016163e-01 -1.18112572e-01
9.97897744e-01 -2.97121763e-01 3.69352579e-01 2.51121982e-03
1.40488267e+00 -5.94727648e-03 -1.13066411e+00 -5.29538840e-02
-2.23043058e-02 -2.74977624e-01 5.41560471e-01 -5.02024055e-01
-1.00801373e+00 9.83614624e-01 1.04378295e+00 2.63617754e-01
9.39696491e-01 -6.38185322e-01 5.61771393e-01 5.68642795e-01
8.48676085e-01 -1.04202938e+00 -2.36795828e-01 7.07110763e-01
4.38463897e-01 -1.11231971e+00 3.46970186e-02 -5.36209464e-01
-2.85771415e-02 1.21402538e+00 1.68596640e-01 4.96622562e-01
1.05091703e+00 8.87513041e-01 2.32557207e-01 3.74631360e-02
-3.47527415e-01 -1.99099094e-01 -1.75327495e-01 8.26246917e-01
1.09325778e+00 1.05934180e-01 -1.91799492e-01 2.62488574e-01
-4.52789098e-01 -2.63993651e-01 5.91983974e-01 9.02164698e-01
-4.32896882e-01 -1.55319321e+00 -4.96045589e-01 4.24208850e-01
-3.07537794e-01 -1.39631554e-01 1.81009453e-02 1.13970423e+00
4.60635312e-02 1.01534426e+00 -2.12179884e-01 -4.87847686e-01
7.75284395e-02 -7.50698686e-01 3.41368943e-01 -1.99345484e-01
-5.67667070e-04 -1.31457269e-01 4.71363455e-01 -3.93778175e-01
-1.82586968e-01 -5.89350522e-01 -1.07204258e+00 -5.87551475e-01
-4.22554255e-01 4.02086765e-01 1.03036201e+00 8.64613652e-01
2.05354124e-01 8.41342151e-01 7.98562706e-01 -9.46929395e-01
-2.82157212e-01 -6.04918838e-01 -9.68926013e-01 -1.33712649e-01
5.23070335e-01 -6.22257650e-01 -1.41467631e-01 -2.87730813e-01] | [6.31269645690918, 1.4407579898834229] |
2ff18ab5-82ed-42f4-918d-36cbad403013 | the-muse-2022-multimodal-sentiment-analysis | 2207.05691 | null | https://arxiv.org/abs/2207.05691v2 | https://arxiv.org/pdf/2207.05691v2.pdf | The MuSe 2022 Multimodal Sentiment Analysis Challenge: Humor, Emotional Reactions, and Stress | The Multimodal Sentiment Analysis Challenge (MuSe) 2022 is dedicated to multimodal sentiment and emotion recognition. For this year's challenge, we feature three datasets: (i) the Passau Spontaneous Football Coach Humor (Passau-SFCH) dataset that contains audio-visual recordings of German football coaches, labelled for the presence of humour; (ii) the Hume-Reaction dataset in which reactions of individuals to emotional stimuli have been annotated with respect to seven emotional expression intensities, and (iii) the Ulm-Trier Social Stress Test (Ulm-TSST) dataset comprising of audio-visual data labelled with continuous emotion values (arousal and valence) of people in stressful dispositions. Using the introduced datasets, MuSe 2022 2022 addresses three contemporary affective computing problems: in the Humor Detection Sub-Challenge (MuSe-Humor), spontaneous humour has to be recognised; in the Emotional Reactions Sub-Challenge (MuSe-Reaction), seven fine-grained `in-the-wild' emotions have to be predicted; and in the Emotional Stress Sub-Challenge (MuSe-Stress), a continuous prediction of stressed emotion values is featured. The challenge is designed to attract different research communities, encouraging a fusion of their disciplines. Mainly, MuSe 2022 targets the communities of audio-visual emotion recognition, health informatics, and symbolic sentiment analysis. This baseline paper describes the datasets as well as the feature sets extracted from them. A recurrent neural network with LSTM cells is used to set competitive baseline results on the test partitions for each sub-challenge. We report an Area Under the Curve (AUC) of .8480 for MuSe-Humor; .2801 mean (from 7-classes) Pearson's Correlations Coefficient for MuSe-Reaction, as well as .4931 Concordance Correlation Coefficient (CCC) and .4761 for valence and arousal in MuSe-Stress, respectively. | ['Björn W. Schuller', 'Erik Cambria', 'Alan Cowen', 'Andreas König', 'Eva-Maria Meßner', 'Lukas Stappen', 'Niklas Müller', 'Alexander Kathan', 'Panagiotis Tzirakis', 'Alice Baird', 'Shahin Amiriparian', 'Lukas Christ'] | 2022-06-23 | null | null | null | null | ['humor-detection'] | ['natural-language-processing'] | [-9.22822654e-02 -1.62277073e-01 4.02430534e-01 -3.05348635e-01
-7.51334786e-01 -3.82865369e-01 2.43723348e-01 4.38759565e-01
-4.92409825e-01 4.02157664e-01 3.57959956e-01 5.35539329e-01
3.24800313e-02 -3.21122944e-01 -9.91424024e-02 -6.96628630e-01
-3.78610671e-01 -8.93735047e-03 -4.34207231e-01 -6.08059108e-01
4.12824284e-03 2.11235374e-01 -1.83814728e+00 9.05000269e-01
9.36163366e-02 1.60897934e+00 -5.50142169e-01 9.60341275e-01
3.92839670e-01 1.19601858e+00 -6.72870755e-01 -5.96402645e-01
-1.78957164e-01 -6.38195515e-01 -8.63049388e-01 -7.86339417e-02
1.42479822e-01 5.78667760e-01 1.29951119e-01 8.33780169e-01
9.95881557e-01 4.08813119e-01 5.58173537e-01 -1.52302909e+00
-1.68924674e-01 3.50008011e-01 -4.63767946e-01 -4.03861329e-02
8.67458105e-01 1.86995044e-02 1.02612340e+00 -1.05580890e+00
7.12392211e-01 8.85793090e-01 9.07463670e-01 5.84575534e-01
-9.99108136e-01 -4.90936816e-01 -3.98887455e-01 6.98398292e-01
-1.24974227e+00 -1.73359901e-01 8.59832525e-01 -7.08569765e-01
9.61199284e-01 4.68134195e-01 8.82659495e-01 1.56291819e+00
2.14769140e-01 7.80894876e-01 1.41208863e+00 -2.69249946e-01
3.85200083e-01 4.47980940e-01 6.23944849e-02 3.62933934e-01
-8.23613942e-01 -4.09845728e-03 -1.04399145e+00 -4.26404960e-02
-1.05203494e-01 -3.35437626e-01 -1.75018266e-01 2.39160165e-01
-1.01665747e+00 6.71689868e-01 2.79917866e-01 4.14459616e-01
-6.52581394e-01 -6.49545789e-02 1.41331637e+00 7.07716048e-01
4.41514045e-01 4.97178972e-01 -1.53132945e-01 -4.51540440e-01
-8.46027732e-01 3.33223045e-01 8.05676222e-01 3.57060730e-01
4.30848897e-01 2.92867333e-01 -2.50842303e-01 1.29614365e+00
-1.50299221e-01 2.97920018e-01 7.24154651e-01 -8.52637410e-01
-1.84166059e-01 6.61553323e-01 -1.94944724e-01 -1.37750733e+00
-1.06746554e+00 -2.82265306e-01 -1.09847593e+00 3.32217395e-01
8.01826268e-02 -4.51060057e-01 -2.97666848e-01 1.79774511e+00
1.28674611e-01 -1.59911215e-01 2.49057680e-01 1.17671657e+00
1.37588716e+00 6.03403926e-01 -2.52022129e-02 -5.62161565e-01
1.59468377e+00 -4.95453686e-01 -7.08054304e-01 -6.09026216e-02
5.31627476e-01 -6.65737450e-01 1.09814978e+00 9.98463809e-01
-1.13606966e+00 -5.75080574e-01 -9.06041086e-01 1.74000263e-01
-6.79572046e-01 -3.66780199e-02 6.59976378e-02 4.98629212e-01
-1.00062239e+00 3.78769308e-01 2.59637665e-02 -3.12750906e-01
3.79824787e-02 2.72057056e-01 -6.94131732e-01 5.56314051e-01
-1.72567713e+00 8.77765298e-01 -1.89795028e-02 3.38499844e-01
-7.11896062e-01 -4.84308273e-01 -9.32930887e-01 -4.14183401e-02
-6.16014339e-02 -1.91618741e-01 9.95266497e-01 -1.54313385e+00
-1.49245477e+00 1.44051635e+00 2.71937221e-01 -3.24427187e-01
2.83407897e-01 -1.59692422e-01 -7.55454004e-01 2.98600674e-01
-1.11408480e-01 5.41867673e-01 6.01006210e-01 -1.14005554e+00
-2.66473204e-01 -3.25868309e-01 -4.22552705e-01 1.46192178e-01
-5.37094295e-01 5.37533164e-01 5.48885241e-02 -4.64348197e-01
-3.52357298e-01 -8.61643136e-01 -2.71568149e-02 -8.46760213e-01
-4.47214961e-01 -2.23049387e-01 5.60469270e-01 -5.81697404e-01
1.36892867e+00 -2.42978930e+00 3.93845350e-01 3.53106529e-01
1.43355310e-01 -6.78163162e-03 -3.03603470e-01 6.79340005e-01
-6.31015539e-01 -2.78500766e-01 6.18815757e-02 -4.46954817e-01
4.06279594e-01 -2.10736424e-01 -2.58740425e-01 3.99852276e-01
8.96941051e-02 7.25143075e-01 -6.66117072e-01 -5.91679573e-01
-4.59780581e-02 5.98840892e-01 -3.61517489e-01 3.33938897e-01
2.63657391e-01 2.67438173e-01 1.43871993e-01 7.41193593e-01
1.17927745e-01 2.73267925e-01 2.39084009e-02 -3.35427284e-01
-1.63335383e-01 -3.27065527e-01 -9.01785374e-01 1.16821432e+00
-5.07945478e-01 8.37703168e-01 3.17105092e-02 -7.04913855e-01
1.47018898e+00 8.28744888e-01 8.61888587e-01 -8.42808545e-01
6.21251822e-01 8.16991776e-02 -3.17686260e-01 -9.25721526e-01
7.15847552e-01 -7.17796862e-01 -9.06435132e-01 2.23424852e-01
3.19549680e-01 -1.86010510e-01 1.54476762e-01 -1.06084786e-01
1.03411734e+00 -1.84083670e-01 2.28303567e-01 4.51774634e-02
6.99722409e-01 -3.83597493e-01 5.92203379e-01 1.71609670e-01
-5.70446134e-01 6.75471067e-01 9.23615992e-01 -5.45340002e-01
-7.06140697e-01 -6.69763505e-01 6.40661940e-02 1.44715261e+00
-3.61650556e-01 -5.27221203e-01 -7.43591845e-01 -2.55092472e-01
-3.84484023e-01 4.19621229e-01 -9.59647715e-01 -3.07682991e-01
-6.17124997e-02 -6.99494600e-01 8.62232029e-01 4.75446433e-01
2.13612035e-01 -1.73222363e+00 -1.00663161e+00 -2.45962553e-02
-5.88825822e-01 -1.09738958e+00 -4.24027108e-02 5.99705160e-01
-2.29687504e-02 -8.52020025e-01 -4.40080166e-01 -7.71367610e-01
-1.34194285e-01 -5.45780897e-01 1.20904195e+00 -4.38269347e-01
-3.96219701e-01 7.51409411e-01 -6.66020513e-01 -5.15867412e-01
-1.62005275e-01 -3.81632149e-01 2.04772487e-01 4.31823015e-01
6.15950465e-01 -6.30245566e-01 -2.57707745e-01 3.30474854e-01
-7.80667841e-01 -3.56185377e-01 4.87666689e-02 9.86209154e-01
5.42320549e-01 -3.14598650e-01 8.42971981e-01 -3.68730813e-01
8.98595631e-01 -5.96258283e-01 1.90251797e-01 -7.71529898e-02
-7.77295604e-02 -9.29117441e-01 6.84240937e-01 -6.36143506e-01
-7.37087727e-01 9.41106901e-02 -1.96643487e-01 -7.18966603e-01
-4.12026256e-01 6.60686195e-01 2.07141325e-01 1.78814739e-01
9.34804618e-01 -3.33446749e-02 -3.47168356e-01 -2.68596541e-02
-1.39812725e-02 8.54054034e-01 1.06471884e+00 -4.43905890e-01
1.03921436e-01 -1.42838096e-03 1.08615085e-01 -1.02379477e+00
-9.02798176e-01 -5.33299625e-01 -4.24233586e-01 -1.00533831e+00
1.16974747e+00 -9.24102962e-01 -1.46165872e+00 5.56811154e-01
-9.65411186e-01 -1.99434310e-01 -4.54906911e-01 4.49059218e-01
-9.66531992e-01 1.12773590e-01 -9.71431017e-01 -1.25877595e+00
-6.62833631e-01 -6.88493252e-01 8.76988232e-01 1.43974662e-01
-1.02969599e+00 -7.10491657e-01 5.70947111e-01 5.16599715e-01
-2.46405825e-02 9.66396093e-01 7.28828549e-01 -8.08554411e-01
8.52287531e-01 -4.48553711e-01 1.56532347e-01 7.07653463e-01
-4.50592399e-01 -5.06911278e-02 -1.32415664e+00 7.38781691e-02
1.53514072e-01 -1.26655853e+00 6.84745073e-01 1.64686456e-01
9.82252657e-01 -8.50863196e-03 6.14156425e-01 3.06237102e-01
9.14021134e-01 1.65483698e-01 7.76544511e-01 2.88419425e-01
3.10065687e-01 1.07757485e+00 7.73102641e-01 8.21829081e-01
2.14147940e-01 6.71164095e-01 5.88316381e-01 -2.33904645e-01
4.60511416e-01 1.31269604e-01 9.12328243e-01 1.06714678e+00
-3.01224917e-01 7.97476768e-02 -9.20968711e-01 5.21121979e-01
-1.73272538e+00 -1.24233067e+00 -5.97412527e-01 1.93091011e+00
7.51877189e-01 -2.27170527e-01 6.30707979e-01 7.87271261e-01
5.73889375e-01 1.85431227e-01 -8.88073295e-02 -1.46839499e+00
-5.10806680e-01 1.84906244e-01 -5.53062797e-01 3.06301653e-01
-1.15587854e+00 7.18111575e-01 5.33672857e+00 6.85460210e-01
-1.33794153e+00 4.96758185e-02 6.62623048e-01 -6.45580828e-01
1.83340237e-01 -6.66780055e-01 -1.79544583e-01 3.72220874e-01
1.29470420e+00 1.12341002e-01 2.54968733e-01 7.45085955e-01
1.92631170e-01 -5.47389649e-02 -8.56313467e-01 1.41802216e+00
6.14375412e-01 -5.63232124e-01 -5.85689664e-01 -3.78983587e-01
5.45516551e-01 -8.86543095e-02 2.09212080e-01 6.71658158e-01
-3.88987660e-01 -1.23638141e+00 7.24439204e-01 9.18018878e-01
9.59288299e-01 -1.18374300e+00 1.13947058e+00 -8.47266912e-02
-8.87389123e-01 -2.50360817e-01 4.81555387e-02 -1.74521402e-01
1.37362316e-01 6.47712231e-01 -4.35700238e-01 3.58181894e-01
1.11294746e+00 5.89705050e-01 -4.26759154e-01 5.84946632e-01
-1.00823738e-01 6.31092608e-01 3.75407562e-02 -2.34344169e-01
2.10513338e-01 -1.29772024e-02 5.55485964e-01 1.93400121e+00
2.50369698e-01 -3.04462519e-02 -1.05248220e-01 5.69426715e-01
7.98561722e-02 6.30927384e-01 -4.44446415e-01 -3.30603123e-02
-1.22788176e-01 1.83033001e+00 -4.32686418e-01 -1.32856816e-01
7.07135201e-02 9.71790671e-01 -9.92143750e-02 2.16305465e-01
-7.64478147e-01 -5.97415328e-01 4.17977333e-01 -3.32729608e-01
-1.13610595e-01 4.55136180e-01 -3.21526915e-01 -8.12808990e-01
-3.44571084e-01 -1.11959004e+00 5.71336687e-01 -1.16955984e+00
-1.47412932e+00 1.05121541e+00 -3.06866616e-01 -1.16651618e+00
-4.78476107e-01 -5.52303314e-01 -6.88871503e-01 5.74197650e-01
-7.06313729e-01 -1.12260652e+00 -3.77600014e-01 8.08611393e-01
2.61715829e-01 -2.31708989e-01 1.08372867e+00 5.02256393e-01
-5.76589704e-01 5.67906559e-01 -5.92510641e-01 8.71161893e-02
9.39387858e-01 -1.33456898e+00 -5.18447101e-01 4.82997410e-02
-1.04009457e-01 -1.62254050e-02 9.85217214e-01 -1.21019945e-01
-1.02544022e+00 -8.72659147e-01 1.28930557e+00 -2.44550675e-01
7.20288038e-01 -3.73865813e-01 -8.02976966e-01 2.38223642e-01
3.10827255e-01 -9.10205767e-02 1.23302901e+00 3.29612643e-01
-3.77317965e-01 -1.14780612e-01 -1.04190111e+00 3.84268045e-01
2.94630200e-01 -7.87735522e-01 -3.80403161e-01 1.57141417e-01
9.73629579e-02 -2.41800815e-01 -1.25197732e+00 5.32768846e-01
8.18008125e-01 -1.51220214e+00 6.06667101e-01 -8.03683519e-01
1.07046962e+00 -3.59131210e-02 -4.74523246e-01 -1.33129990e+00
-6.03638589e-02 -5.51057637e-01 1.81224108e-01 1.17846310e+00
5.70799187e-02 -3.86062041e-02 4.37984854e-01 3.82737592e-02
-2.11395845e-01 -1.08763993e+00 -1.06103063e+00 -2.41581559e-01
-1.37950674e-01 -9.13103342e-01 1.50481805e-01 1.33610332e+00
7.75375426e-01 6.44713521e-01 -7.38013148e-01 -4.37384099e-01
9.16059781e-03 -3.03049982e-02 6.85982704e-01 -1.05399430e+00
-1.59700021e-01 -6.19934201e-01 -8.54066253e-01 1.10242352e-01
3.14355463e-01 -8.15240979e-01 1.01302311e-01 -9.57861543e-01
2.78704524e-01 5.74502409e-01 -7.24749565e-01 7.20494866e-01
3.29771966e-01 8.31309736e-01 3.48710328e-01 -4.86284524e-01
-8.59295368e-01 5.79805791e-01 6.27372980e-01 5.62115721e-02
-3.20645690e-01 -2.20675260e-01 -4.92765576e-01 9.97323334e-01
6.45126045e-01 -1.31408110e-01 5.15894927e-02 6.07286394e-01
1.12688518e+00 4.33503240e-01 6.60935640e-01 -1.16238272e+00
-1.15597863e-02 -1.02205842e-03 4.66522217e-01 -6.32112443e-01
8.03169429e-01 -4.63074654e-01 -5.45759425e-02 2.08660781e-01
-5.66880703e-01 1.16758138e-01 2.69201577e-01 -7.71911368e-02
-6.31687999e-01 -1.44123793e-01 9.44542766e-01 1.39173329e-01
-5.30086279e-01 -3.32855374e-01 -7.28822291e-01 1.51214376e-01
9.15291965e-01 -2.73176935e-02 -1.75524130e-01 -8.66594613e-01
-1.15010405e+00 2.15261236e-01 -2.37407465e-03 4.60907191e-01
8.22333515e-01 -1.49186802e+00 -8.49799752e-01 -6.30510552e-03
3.35677326e-01 -7.10196197e-01 8.66316319e-01 1.41779792e+00
1.06280081e-01 -1.65087849e-01 -5.50368845e-01 -5.11191547e-01
-1.64979756e+00 4.50986832e-01 3.90327305e-01 -1.42509133e-01
-7.14549646e-02 8.76685441e-01 -7.49834254e-02 -4.51567680e-01
1.86758831e-01 3.00848514e-01 -6.85830355e-01 1.04072297e+00
6.19765162e-01 6.77427590e-01 7.33102337e-02 -1.25693369e+00
-2.83792257e-01 4.10090089e-01 5.08007526e-01 -1.90097168e-01
1.43178225e+00 1.10986441e-01 -4.05411422e-01 1.27029645e+00
1.26769686e+00 2.23668721e-02 -4.56863195e-01 1.91885591e-01
-4.95930016e-02 2.34530598e-01 -9.19461623e-02 -1.19495487e+00
-9.75619197e-01 1.13529539e+00 7.80851245e-01 4.84822690e-01
1.60255122e+00 -2.81313956e-01 6.67086720e-01 3.14380318e-01
-6.80960640e-02 -1.50012243e+00 3.61549705e-01 8.12007964e-01
1.33016348e+00 -8.20049942e-01 -3.46333683e-01 1.61178797e-01
-1.44733608e+00 1.14817119e+00 4.72189426e-01 9.45534259e-02
3.17230374e-01 1.54799595e-01 3.92629445e-01 -5.18456459e-01
-1.07606709e+00 -2.46029481e-01 5.16995728e-01 3.42551529e-01
6.47840559e-01 1.34050235e-01 -2.96131074e-01 1.35515749e+00
-6.86179221e-01 -2.95241419e-02 5.56810379e-01 3.41090143e-01
-2.76857555e-01 -4.83596444e-01 -6.09957695e-01 1.22063063e-01
-5.80366552e-01 2.18970388e-01 -1.09854245e+00 4.71928269e-01
5.58155656e-01 1.11624718e+00 -1.36352301e-01 -1.07961738e+00
5.96898437e-01 4.52595145e-01 1.37211800e-01 -3.16069812e-01
-1.50330961e+00 2.13389173e-01 3.81099939e-01 -7.37680256e-01
-4.91774291e-01 -6.23636603e-01 -1.27206302e+00 -1.40409529e-01
2.54833102e-01 3.27648997e-01 5.70310771e-01 6.66205287e-01
3.63664180e-02 5.81446707e-01 9.39057112e-01 -1.01561427e+00
-4.68176343e-02 -1.07910943e+00 -9.93764162e-01 8.36686909e-01
1.89499527e-01 -4.05194134e-01 -4.91165191e-01 8.95866901e-02] | [13.418718338012695, 5.04036283493042] |
5448be65-df8f-4329-8c18-8c125bbc73bc | an-approach-modality-reduction-and-face | 1312.1681 | null | https://arxiv.org/abs/1312.1681v1 | https://arxiv.org/pdf/1312.1681v1.pdf | An Approach: Modality Reduction and Face-Sketch Recognition | To recognize face sketch through face photo database is a challenging task for todays researchers. Because face photo images in training set and face sketch images in testing set have different modality. Difference between two face photos of difference person is smaller than the difference between same person in a face photo and face sketched. In this paper, for reduction of the modality between face photo and face sketch we first bring face photo and face sketch images in a new dimension using 2D Discrete Haar wavelet transform with scale 3 followed by a negative approach. After that, extract features from transformed images using Principal Component Analysis (PCA). Thereafter, we use SVM classifier and K-NN classifier for better classification. Our proposed method is experimentally verified by its robustness against faces that are captured in a good lighting condition and in a frontal pose. The experiment has been conducted with 100 male and female face images as training set and 100 male and female face sketch images as testing set collected from CUHK training and testing cropped photos and CUHK training and testing cropped sketches. | ['Sourav Pramanik', 'Dr. Debotosh Bhattacharjee'] | 2013-12-05 | null | null | null | null | ['sketch-recognition'] | ['computer-vision'] | [ 2.06148162e-01 -2.29164064e-01 3.22193131e-02 -4.24889982e-01
-9.37549323e-02 -7.27491081e-01 6.21576071e-01 -9.41037953e-01
-4.36089821e-02 5.41545093e-01 -9.64619592e-02 2.51775265e-01
6.73357174e-02 -8.21350098e-01 -5.06201804e-01 -6.31745577e-01
2.22816244e-01 2.72923797e-01 -2.40850642e-01 1.27676561e-01
1.48206487e-01 1.07810628e+00 -1.76851010e+00 3.96416903e-01
-1.88287392e-01 9.44428265e-01 -2.10777059e-01 6.39682472e-01
-8.54311064e-02 1.97349682e-01 -6.07534111e-01 -5.23770809e-01
9.02748346e-01 -5.69763958e-01 -2.85822511e-01 7.05988050e-01
8.67398381e-01 -5.13813794e-01 -4.91217107e-01 8.96901309e-01
5.32711565e-01 -7.64563214e-03 1.00572348e+00 -1.74071741e+00
-1.13706923e+00 4.94714268e-02 -8.78282905e-01 1.01094926e-03
4.41719413e-01 -7.29691312e-02 2.16532554e-02 -1.37663722e+00
6.56082392e-01 1.67669237e+00 5.06026804e-01 1.00376058e+00
-1.15038216e+00 -1.25322950e+00 -4.83039707e-01 2.52028048e-01
-1.64332998e+00 -7.06402719e-01 1.22271609e+00 -3.20169985e-01
6.15078747e-01 2.68489003e-01 5.70202470e-01 8.33282769e-01
-2.69592144e-02 3.03388804e-01 1.54147732e+00 -5.80545723e-01
-1.11544907e-01 6.60912871e-01 -2.16142923e-01 7.38928437e-01
-1.11657605e-01 1.91006854e-01 -4.63432968e-01 -3.58139068e-01
1.02780998e+00 4.39156443e-01 -1.35329768e-01 -6.12102356e-03
-7.34192312e-01 6.03156507e-01 -1.22210249e-01 3.06173027e-01
-3.89476895e-01 -3.83858949e-01 3.70254107e-02 7.18958855e-01
-1.82132214e-01 -4.66790229e-01 -1.61820471e-01 4.68461752e-01
-9.97579396e-01 4.61191982e-02 7.69278944e-01 1.06171012e+00
7.21188545e-01 2.80447192e-02 2.19695121e-01 1.24445796e+00
2.90599197e-01 1.10047674e+00 6.16009235e-01 -6.68619812e-01
1.47113204e-01 6.70873463e-01 -2.75397480e-01 -1.25551760e+00
2.90226012e-01 3.71220469e-01 -8.09679329e-01 6.93923414e-01
2.25465432e-01 1.02899194e-01 -9.19288278e-01 1.22729468e+00
3.64098698e-01 3.15353781e-01 3.64825368e-01 7.24129796e-01
9.84264970e-01 7.55453646e-01 -1.14007883e-01 -4.32581007e-01
1.54222524e+00 -8.17445159e-01 -6.68017149e-01 8.44291225e-02
-5.39955735e-01 -1.30321097e+00 1.11181605e+00 6.49950206e-01
-1.02682495e+00 -9.72089767e-01 -1.17455852e+00 -5.44797676e-03
-3.94672006e-01 9.68501866e-01 1.37815297e-01 1.00650001e+00
-7.50898004e-01 4.25854236e-01 -1.13802701e-01 -5.92700660e-01
5.52137077e-01 5.89950860e-01 -1.20937800e+00 -2.66437590e-01
-5.88533044e-01 6.15998030e-01 -1.93685349e-02 1.53044268e-01
-6.61261737e-01 -3.43894750e-01 -5.51330328e-01 8.33091605e-03
-1.46603733e-01 4.56763320e-02 5.24091363e-01 -1.40972626e+00
-1.56541896e+00 1.19007397e+00 -2.29122579e-01 1.96251646e-01
4.16820824e-01 2.84148574e-01 -9.31097984e-01 1.68509677e-01
-2.56200045e-01 6.54716730e-01 1.72845733e+00 -1.09248710e+00
-2.91309953e-01 -9.49413180e-01 -5.95137656e-01 1.27021521e-01
-5.56849957e-01 1.98416531e-01 -3.62373561e-01 -4.87887889e-01
3.03587615e-01 -9.94709551e-01 7.71970212e-01 2.85227060e-01
3.75920981e-02 -2.19647646e-01 1.65739322e+00 -8.81972253e-01
6.41861916e-01 -2.05562687e+00 -1.48761153e-01 4.48262542e-01
-3.67012829e-01 3.89913797e-01 -2.72851050e-01 6.07467830e-01
-3.87239873e-01 -1.70887992e-01 1.11095592e-01 -8.63200724e-02
-4.41702187e-01 2.79926777e-01 -4.75730896e-01 5.05168319e-01
6.63716972e-01 3.75608474e-01 -1.91450015e-01 -6.76912606e-01
1.67563409e-01 7.75992751e-01 -1.63432777e-01 2.80503035e-01
5.35340130e-01 3.47000062e-01 -3.01298231e-01 9.92692053e-01
1.37974954e+00 5.08593678e-01 1.49462104e-01 -5.76563001e-01
2.17414498e-01 -9.89601016e-01 -1.46902227e+00 9.98031378e-01
-2.93843061e-01 6.40119851e-01 8.57919529e-02 -7.64774621e-01
1.43623626e+00 5.26067257e-01 -2.06917599e-02 -4.08796608e-01
3.20474952e-02 -1.18714087e-01 -2.27766961e-01 -8.41713011e-01
-1.48100078e-01 -3.75512928e-01 6.32022798e-01 5.22632182e-01
2.74798900e-01 -2.38804698e-01 -2.61818506e-02 -3.49532634e-01
5.16439617e-01 -2.14866642e-02 2.77646959e-01 -4.99904640e-02
1.25468695e+00 -3.66564274e-01 2.15510234e-01 4.45523337e-02
-2.63846755e-01 4.67868686e-01 3.03998202e-01 -6.68900847e-01
-1.35390186e+00 -8.85907292e-01 -1.93209141e-01 7.04686761e-01
-3.14654529e-01 3.22604865e-01 -6.43036187e-01 -5.75358450e-01
3.33101988e-01 -1.30275205e-01 -8.49003792e-01 2.00917453e-01
-7.32972920e-01 -1.88143000e-01 6.81615949e-01 3.71747345e-01
6.83178425e-01 -1.29822516e+00 -1.19899042e-01 -3.94863307e-01
5.44329107e-01 -8.14763665e-01 -5.07232547e-01 -7.38267779e-01
-6.75288320e-01 -1.53046608e+00 -1.09325624e+00 -1.20858729e+00
1.08862889e+00 2.41627350e-01 4.98728156e-01 1.90209046e-01
-7.90170670e-01 6.02775693e-01 -5.04798740e-02 -6.29352808e-01
-3.08695197e-01 -7.92315543e-01 3.47940415e-01 2.51752973e-01
6.76751733e-01 -5.43049037e-01 -5.82924843e-01 3.95531803e-01
-8.25449824e-01 -4.44148362e-01 5.98219514e-01 1.07854342e+00
2.22148061e-01 2.91050911e-01 4.56571132e-01 -4.65458691e-01
5.78307033e-01 -1.55368984e-01 -5.76684833e-01 6.22236073e-01
-2.36101538e-01 -3.69875520e-01 7.30940461e-01 -5.52661479e-01
-1.12070918e+00 4.48171377e-01 1.01188093e-01 -7.11038053e-01
-4.41272557e-01 -3.13693613e-01 -2.13337749e-01 -4.14433867e-01
6.46501184e-01 5.19620419e-01 6.02456808e-01 -4.08282399e-01
2.39210408e-02 9.43889797e-01 6.39419556e-01 -3.66384208e-01
1.17994678e+00 7.74644911e-01 2.17150286e-01 -1.28984118e+00
2.22276554e-01 -1.58023447e-01 -9.19593394e-01 -4.28223819e-01
7.16170490e-01 -8.49855721e-01 -8.81897390e-01 5.11354446e-01
-9.16381240e-01 4.94451255e-01 3.96178246e-01 2.89810151e-01
-1.28338873e-01 4.60318685e-01 -3.91998351e-01 -1.25095677e+00
-5.25768936e-01 -9.42373931e-01 1.03796542e+00 4.80275512e-01
3.88625294e-01 -5.17205894e-01 -1.86918750e-01 3.57567698e-01
1.80220097e-01 1.17114566e-01 8.06568325e-01 -3.42758060e-01
-1.19794630e-01 -7.47347713e-01 -3.38733345e-01 8.20624709e-01
5.40718615e-01 5.19529283e-01 -1.33395100e+00 -3.63693178e-01
2.11790219e-01 -3.75176013e-01 4.64988559e-01 -9.95773301e-02
9.00566459e-01 -3.90415668e-01 -1.93711877e-01 2.73517966e-01
1.54044116e+00 4.86393452e-01 6.77291274e-01 -4.22315866e-01
2.51800388e-01 7.95538247e-01 4.38430458e-01 1.98095456e-01
-1.89758152e-01 4.26526308e-01 -3.55071127e-01 1.08906195e-01
-2.61949807e-01 -3.23598385e-01 3.11715007e-01 5.09970605e-01
-1.96076736e-01 2.21046299e-01 -5.58723569e-01 8.00564140e-02
-1.23904872e+00 -1.34883714e+00 2.92029381e-01 2.28777075e+00
4.65997845e-01 -5.82926154e-01 1.90954372e-01 6.99745119e-01
9.40176249e-01 -3.63717109e-01 -2.26921380e-01 -2.02508673e-01
-2.35014543e-01 4.69956428e-01 1.17596798e-01 2.79050767e-01
-8.86293411e-01 7.62115180e-01 5.45842266e+00 8.04759562e-01
-1.59463942e+00 -1.40223116e-01 7.01046050e-01 -4.74768924e-03
1.23246074e-01 -1.42635375e-01 -6.88678920e-01 4.30478692e-01
2.27904469e-01 -3.92701924e-02 7.77005196e-01 9.01855350e-01
-2.06802338e-01 -1.57544449e-01 -1.21321690e+00 1.59498453e+00
4.64275181e-01 -1.03770983e+00 2.05313727e-01 -8.11809152e-02
6.21855378e-01 -7.73397744e-01 2.52825975e-01 2.69726396e-01
-5.42938530e-01 -1.38847804e+00 2.33158186e-01 5.51272273e-01
1.06094599e+00 -7.03182697e-01 4.19671327e-01 2.46047944e-01
-1.19724584e+00 -2.33028814e-01 -9.44127679e-01 -9.78877172e-02
-4.53841180e-01 -1.72914997e-01 -7.73956716e-01 1.12932377e-01
6.49311006e-01 2.60311306e-01 -3.65146756e-01 3.67835701e-01
4.16818887e-01 2.00097606e-01 -2.88353235e-01 1.60480723e-01
-2.26669282e-01 -5.43364406e-01 1.61314070e-01 8.94919932e-01
6.16432190e-01 5.38760841e-01 -7.76999369e-02 6.87178016e-01
-1.36465907e-01 3.55759889e-01 -1.06757843e+00 -2.19597295e-01
7.95864701e-01 1.46719372e+00 -6.49031401e-01 -4.80105549e-01
-4.42399085e-01 1.24718845e+00 -1.83059081e-01 3.77286464e-01
-4.66083914e-01 -2.31853008e-01 2.02237025e-01 -1.14170730e-01
1.27464429e-01 9.94318277e-02 1.66169450e-01 -6.94051683e-01
1.70739412e-01 -8.48298192e-01 2.79151052e-01 -7.34524965e-01
-1.29630184e+00 7.52235413e-01 -5.04328683e-02 -1.22570360e+00
1.34259298e-01 -8.47393811e-01 -7.62469471e-01 1.25159597e+00
-1.23902452e+00 -1.55676675e+00 -6.00411534e-01 8.89122427e-01
7.07289457e-01 -9.14794803e-01 9.68208849e-01 3.58958662e-01
-3.75939935e-01 7.11447597e-01 -9.47534144e-02 3.74405056e-01
8.74181986e-01 -5.18114924e-01 1.31540550e-02 3.18583846e-01
1.33213967e-01 7.49833584e-01 2.73525029e-01 -6.31545126e-01
-1.76990354e+00 -7.00646222e-01 6.77917361e-01 -4.60044324e-01
-5.33463880e-02 -2.49535084e-01 -5.53186536e-01 5.56227088e-01
2.42086515e-01 3.61817867e-01 6.87591553e-01 -6.54118538e-01
-3.80574703e-01 -2.20803037e-01 -1.80690050e+00 3.05921406e-01
5.75066209e-01 -7.22691715e-01 -5.02130866e-01 3.40580702e-01
-3.23071003e-01 -1.34087950e-02 -9.59298730e-01 2.73052931e-01
1.33303225e+00 -9.49520707e-01 1.13136184e+00 -6.67573333e-01
1.96884826e-01 -3.47628891e-01 -3.17044199e-01 -5.81337214e-01
-1.83208787e-03 -2.34289110e-01 4.68135655e-01 1.42549372e+00
-3.34590040e-02 -4.93026853e-01 9.82651711e-01 5.06434500e-01
7.37540185e-01 -3.69553328e-01 -6.53273940e-01 -8.35848868e-01
2.08609030e-02 4.16303068e-01 6.72249377e-01 1.08928251e+00
-2.55085260e-01 8.42367411e-02 -5.99068940e-01 1.94645733e-01
7.61380851e-01 3.31994332e-02 1.03962183e+00 -1.34242570e+00
7.45591298e-02 2.42019929e-02 -5.52532732e-01 2.46047080e-02
2.93119371e-01 -5.95588923e-01 -6.42644525e-01 -7.88073957e-01
3.82757992e-01 -2.46968701e-01 3.18476588e-01 5.09143949e-01
3.36177379e-01 9.14678574e-01 2.84177512e-01 2.93922156e-01
6.99735165e-01 3.49590421e-01 1.00780272e+00 -2.48357430e-01
1.63658578e-02 -1.81920037e-01 2.09957048e-01 5.20301878e-01
4.90620196e-01 -2.71309912e-01 -6.20618701e-01 -5.53760584e-03
-5.93792856e-01 3.17193687e-01 4.78161186e-01 -1.06695175e+00
6.36291355e-02 -1.48228467e-01 1.26929665e+00 -4.63648766e-01
7.59850919e-01 -1.16041088e+00 4.91552621e-01 4.57569510e-01
-9.44823399e-02 2.67702788e-01 1.75225556e-01 2.79153109e-01
-2.21277177e-01 -2.22215354e-01 9.69830334e-01 -3.25836629e-01
-6.41591072e-01 2.94538528e-01 1.68283328e-01 -6.30213857e-01
1.40160954e+00 -7.39526510e-01 -2.89409280e-01 -2.85384476e-01
-8.44670594e-01 -3.45436573e-01 4.51940835e-01 5.28808177e-01
1.07947528e+00 -1.69984889e+00 -7.39371300e-01 1.06776953e+00
2.01930050e-02 -9.89315093e-01 4.94702041e-01 4.40249503e-01
-6.29438281e-01 4.14058149e-01 -1.12122369e+00 -2.38752425e-01
-2.24467087e+00 6.99456453e-01 6.71060458e-02 8.54892910e-01
-3.67767990e-01 9.40072536e-01 -2.18751840e-02 -3.79841477e-01
1.84460744e-01 1.50770783e-01 -3.41121048e-01 -1.44074932e-02
7.28859723e-01 3.87777954e-01 -1.33115992e-01 -1.09546113e+00
-3.38909477e-01 1.14877081e+00 8.19863826e-02 -1.92581266e-01
1.09556627e+00 1.95232242e-01 -3.28427553e-01 -3.84285264e-02
1.54702759e+00 3.48963320e-01 -8.82923841e-01 -8.82046372e-02
-4.27912533e-01 -1.04165769e+00 -3.27193499e-01 -6.33712411e-01
-1.10851562e+00 9.40662265e-01 1.24183440e+00 -9.82331112e-02
1.06190431e+00 -2.25959048e-01 4.51647013e-01 4.56467479e-01
5.76929450e-02 -1.23988199e+00 7.07103461e-02 -1.85792178e-01
1.40388441e+00 -1.43930435e+00 3.50506185e-03 -4.46406543e-01
-7.39234388e-01 1.57677567e+00 5.62612176e-01 -5.23192644e-01
9.19528186e-01 1.19415648e-01 2.39915460e-01 -2.02468947e-01
-5.81849039e-01 2.05988362e-01 3.86625677e-01 8.33454669e-01
3.79167944e-01 -1.41995132e-01 -2.40205362e-01 2.55845636e-01
-2.34999925e-01 3.63023221e-01 2.92378724e-01 9.86130834e-01
-2.51233757e-01 -1.18054450e+00 -9.20054555e-01 3.83181095e-01
-2.16127723e-01 3.10536891e-01 -6.05425298e-01 8.41322243e-01
2.39000827e-01 7.88096547e-01 1.76890612e-01 -3.70743752e-01
1.69566661e-01 5.80082357e-01 1.13257098e+00 -1.94769159e-01
-2.26721197e-01 -8.72710161e-03 -3.56465578e-01 -1.00129060e-01
-2.60201395e-01 -5.58779955e-01 -9.58491325e-01 -3.09899390e-01
-9.96727049e-02 -2.30128113e-02 9.98426914e-01 4.97634888e-01
-5.66729251e-03 -2.58538783e-01 9.63302374e-01 -8.62932324e-01
-7.34055996e-01 -9.67860937e-01 -7.57172585e-01 5.43754280e-01
1.69746920e-01 -6.22324288e-01 -2.21532986e-01 6.31264806e-01] | [13.142444610595703, 0.6790149807929993] |
f89dc484-9950-4df1-9687-cd7faecbfcbb | cafeboost-causal-feature-boost-to-eliminate | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Qiu_CafeBoost_Causal_Feature_Boost_To_Eliminate_Task-Induced_Bias_for_Class_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Qiu_CafeBoost_Causal_Feature_Boost_To_Eliminate_Task-Induced_Bias_for_Class_CVPR_2023_paper.pdf | CafeBoost: Causal Feature Boost To Eliminate Task-Induced Bias for Class Incremental Learning | Continual learning requires a model to incrementally learn a sequence of tasks and aims to predict well on all the learned tasks so far, which notoriously suffers from the catastrophic forgetting problem. In this paper, we find a new type of bias appearing in continual learning, coined as task-induced bias. We place continual learning into a causal framework, based on which we find the task-induced bias is reduced naturally by two underlying mechanisms in task and domain incremental learning. However, these mechanisms do not exist in class incremental learning (CIL), in which each task contains a unique subset of classes. To eliminate the task-induced bias in CIL, we devise a causal intervention operation so as to cut off the causal path that causes the task-induced bias, and then implement it as a causal debias module that transforms biased features into unbiased ones. In addition, we propose a training pipeline to incorporate the novel module into existing methods and jointly optimize the entire architecture. Our overall approach does not rely on data replay, and is simple and convenient to plug into existing methods. Extensive empirical study on CIFAR-100 and ImageNet shows that our approach can improve accuracy and reduce forgetting of well-established methods by a large margin. | ['Lili Pan', 'Qingbo Wu', 'Fanman Meng', 'Lanxiao Wang', 'Heqian Qiu', 'Haitao Wen', 'Hongliang Li', 'Benliu Qiu'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['class-incremental-learning', 'incremental-learning'] | ['computer-vision', 'methodology'] | [ 2.37132281e-01 1.59017250e-01 -2.69568533e-01 -5.99718213e-01
-2.88337588e-01 -2.51904666e-01 7.93306768e-01 -1.13393150e-01
-5.35256565e-01 1.12709165e+00 1.80239469e-01 -3.58911663e-01
-1.84274182e-01 -5.73934257e-01 -1.05445623e+00 -6.51766658e-01
1.24197990e-01 3.55740428e-01 5.63215852e-01 -2.04318285e-01
3.01536679e-01 8.05303156e-02 -1.39541960e+00 2.62077183e-01
1.06451380e+00 6.76720500e-01 2.29833990e-01 2.42874026e-01
-8.05753693e-02 1.17900062e+00 -6.34395540e-01 -4.17737901e-01
6.63505197e-02 -6.18533731e-01 -8.29819500e-01 -1.26758322e-01
5.99936903e-01 -4.66530621e-01 -1.11907385e-01 8.81620049e-01
4.37614679e-01 -6.99899485e-03 6.48832798e-01 -1.36720955e+00
-9.23186779e-01 1.01630592e+00 -7.05867231e-01 2.87983865e-01
-6.70870692e-02 1.70155779e-01 8.41123581e-01 -1.18229532e+00
4.26525116e-01 1.32799602e+00 9.38925683e-01 7.58192301e-01
-1.52824831e+00 -1.06591702e+00 7.26812065e-01 1.28267363e-01
-1.00036097e+00 -3.92945319e-01 9.88264084e-01 -5.21064937e-01
8.09780657e-01 -9.01045874e-02 7.23105788e-01 1.50743151e+00
4.50749189e-01 1.02338088e+00 1.14554119e+00 -5.06347656e-01
4.54222500e-01 3.32331151e-01 4.62190539e-01 6.59413517e-01
4.24424946e-01 2.88256556e-01 -9.68672156e-01 -6.12226687e-02
3.19575071e-01 3.68302822e-01 -1.49169147e-01 -5.43686986e-01
-1.09004450e+00 7.21315742e-01 5.76945364e-01 1.15201332e-01
-2.68697172e-01 3.13507557e-01 2.21933141e-01 6.41432881e-01
6.39843345e-01 3.73954296e-01 -8.20027649e-01 3.29890519e-01
-9.02566493e-01 4.05660331e-01 4.15389478e-01 7.00871646e-01
1.01451051e+00 1.06079154e-01 -3.62077862e-01 8.34872842e-01
-2.63296459e-02 2.87267238e-01 8.32645833e-01 -6.29907429e-01
1.90734997e-01 5.31246543e-01 -7.18659116e-03 -5.71626961e-01
-3.41014236e-01 -8.36903393e-01 -7.90017247e-01 1.63218588e-01
2.96045184e-01 -1.11003704e-01 -1.02372539e+00 2.10499191e+00
2.51809388e-01 2.01436237e-01 -2.68651128e-01 6.67429388e-01
3.07582229e-01 1.25019997e-01 3.23994607e-01 -4.50986415e-01
9.41203654e-01 -1.01541758e+00 -6.83350027e-01 -5.63454688e-01
4.49979395e-01 -3.51939499e-01 1.58609319e+00 4.75206077e-01
-7.46327162e-01 -5.74665189e-01 -1.27249336e+00 -3.59324552e-02
-2.96857685e-01 -2.77128577e-01 8.48565876e-01 4.64450538e-01
-8.08768690e-01 6.73823178e-01 -6.99044228e-01 -3.11157908e-02
7.65168011e-01 1.16925545e-01 6.02788478e-02 -1.09802596e-01
-1.38556373e+00 1.01581204e+00 4.38928455e-01 -3.98538679e-01
-1.31819117e+00 -1.19745791e+00 -4.60436195e-01 -5.46837561e-02
5.64233601e-01 -1.00545692e+00 1.43066537e+00 -1.26423776e+00
-1.26218522e+00 5.77099323e-01 -4.39450234e-01 -8.20733309e-01
6.36678338e-01 -6.24527812e-01 -7.88468048e-02 -5.88748097e-01
4.19558957e-02 7.29546428e-01 1.41058350e+00 -1.36630964e+00
-7.85820305e-01 -3.56791556e-01 4.57644761e-02 1.76404729e-01
-7.73232102e-01 -4.26292270e-01 -4.88518178e-02 -7.70611048e-01
-1.74536705e-02 -8.71044278e-01 2.44385134e-02 -2.29525015e-01
-1.63427100e-01 -4.68088895e-01 8.28180373e-01 -3.21773142e-01
1.27084875e+00 -2.09248829e+00 7.85947889e-02 -4.80160624e-01
5.48138559e-01 1.12277895e-01 -5.27927577e-02 -6.79446682e-02
-1.77176654e-01 -5.98496711e-03 -4.32156771e-01 -5.55533826e-01
-3.02732527e-01 1.86933860e-01 -8.99956763e-01 2.38826647e-01
2.20530063e-01 1.00429380e+00 -1.22808063e+00 -2.22896248e-01
-1.73626587e-01 2.02568009e-01 -7.40061939e-01 9.28564295e-02
-5.01989841e-01 2.52406955e-01 -3.65355648e-02 5.06339133e-01
6.94323957e-01 -1.95313185e-01 -2.56521106e-02 2.12464213e-01
-1.58747703e-01 5.67618251e-01 -7.34695077e-01 1.91233265e+00
-5.00784755e-01 4.42365080e-01 -5.06003320e-01 -9.76011395e-01
8.02977920e-01 7.85644129e-02 1.58591390e-01 -6.87601149e-01
-3.28384757e-01 2.90799528e-01 1.28125148e-02 -2.37269863e-01
3.78457189e-01 -3.94748807e-01 4.98851936e-04 5.85492671e-01
3.59795898e-01 1.59264989e-02 -8.79291221e-02 3.08160603e-01
1.17354667e+00 1.22461125e-01 2.82375574e-01 -3.20076615e-01
1.91427723e-01 1.36404261e-01 7.83604920e-01 1.10985065e+00
-2.46744990e-01 3.13403636e-01 4.42862511e-01 -7.80093253e-01
-8.23688805e-01 -1.09848821e+00 -4.25139219e-02 1.53360116e+00
-2.10577682e-01 -7.58319423e-02 -3.97361726e-01 -1.27845824e+00
3.69264066e-01 1.15953529e+00 -1.01451957e+00 -7.34307110e-01
-4.44687009e-01 -1.12150466e+00 2.55595356e-01 3.75027776e-01
5.17545521e-01 -9.40047562e-01 -4.69813794e-01 1.81302503e-01
-5.52531853e-02 -3.49452704e-01 -4.72384661e-01 7.07122982e-01
-1.21499622e+00 -9.23845351e-01 -5.63278377e-01 -5.32906950e-01
5.27322471e-01 5.33270299e-01 1.36839831e+00 -1.28830582e-01
-1.60019115e-01 8.48941728e-02 -6.71582520e-02 -9.90634084e-01
-1.38494477e-01 4.51086193e-01 2.22180933e-01 -5.09421788e-02
4.68598276e-01 -7.07992554e-01 -5.81495047e-01 1.94865718e-01
-6.77520335e-01 2.03027651e-02 7.32310951e-01 1.22738612e+00
3.11201751e-01 2.58495081e-02 1.16216874e+00 -1.25100613e+00
6.49086654e-01 -6.81497216e-01 -3.80920231e-01 1.18111044e-01
-1.29708207e+00 2.21619129e-01 6.26588881e-01 -7.87170470e-01
-1.37579560e+00 5.98561019e-02 2.12803781e-01 -5.22015393e-01
1.32887051e-01 3.46078664e-01 -1.25771552e-01 3.14066499e-01
9.81728852e-01 3.17426294e-01 -1.26152724e-01 -5.99691808e-01
4.88696069e-01 2.71891862e-01 3.34593058e-01 -3.62296104e-01
7.81316280e-01 6.09872162e-01 -2.89553791e-01 -2.56664127e-01
-1.48891878e+00 -3.98935117e-02 -6.13467157e-01 -1.63261175e-01
3.06626618e-01 -1.11494040e+00 -2.94649959e-01 6.27371430e-01
-1.20259905e+00 -6.48116052e-01 -5.62038124e-01 3.52430850e-01
-2.25050390e-01 -2.59090990e-01 -3.01901788e-01 -6.05963588e-01
-2.48606399e-01 -7.17787623e-01 6.63599491e-01 4.34217863e-02
-2.69474924e-01 -1.02408862e+00 1.92248911e-01 -1.58945754e-01
7.03982294e-01 -2.85874814e-01 1.10265708e+00 -6.11729681e-01
-4.43566591e-01 1.50895998e-01 -9.86270607e-02 5.36246479e-01
1.68919504e-01 -4.45071965e-01 -1.18321335e+00 -4.80589807e-01
2.96340227e-01 -5.19853115e-01 1.59610879e+00 3.04258466e-01
1.21863925e+00 -3.31693083e-01 -3.10992837e-01 5.19188821e-01
1.36030424e+00 1.28556162e-01 4.47970867e-01 3.70640308e-01
7.23322570e-01 5.14654517e-01 5.12898266e-01 2.31796414e-01
3.85094345e-01 2.31694832e-01 4.29872781e-01 9.18256193e-02
-5.60252964e-01 -6.40991807e-01 4.80857134e-01 6.96928263e-01
1.64323345e-01 1.30046144e-01 -8.78193319e-01 6.37688816e-01
-1.93115449e+00 -9.52851474e-01 -1.88701637e-02 2.28540635e+00
1.37713885e+00 4.12980646e-01 -1.78222150e-01 1.19391240e-01
4.63895172e-01 9.15766954e-02 -1.00848234e+00 -2.09514666e-02
7.01947091e-03 2.11016294e-02 4.34577495e-01 5.59614301e-01
-9.52261984e-01 1.19271493e+00 6.82734632e+00 7.36353457e-01
-1.14156485e+00 4.29362386e-01 7.89493084e-01 -4.47616458e-01
-4.80886042e-01 2.08897859e-01 -1.04933119e+00 4.27567184e-01
7.35342681e-01 -4.21824366e-01 4.74001914e-01 1.14995503e+00
-5.78582995e-02 -1.09157763e-01 -1.44853568e+00 5.99142909e-01
9.62435827e-02 -1.11134553e+00 2.57163405e-01 -2.59732515e-01
8.60367477e-01 -3.38724703e-02 2.72230268e-01 8.72595966e-01
7.16674745e-01 -7.94539750e-01 8.56469035e-01 5.65434992e-01
5.97778678e-01 -5.34917653e-01 3.19383502e-01 5.76780438e-01
-5.22656322e-01 -4.24185812e-01 -6.14155591e-01 -2.70008087e-01
-2.53048271e-01 1.24881721e+00 -9.20544505e-01 1.17595613e-01
8.15862358e-01 8.18682551e-01 -8.52284789e-01 8.30485463e-01
-5.65041363e-01 9.25739884e-01 1.61678225e-01 1.18088417e-01
-2.19271958e-01 3.71234059e-01 5.15798151e-01 1.00268447e+00
2.19153583e-01 -4.54346299e-01 -7.23827109e-02 8.76184046e-01
-4.13219720e-01 -2.33756751e-01 -7.14278519e-01 4.41398084e-01
6.14379287e-01 8.84375393e-01 -3.94036472e-01 -5.70019245e-01
-2.73740321e-01 1.13208616e+00 7.56739199e-01 3.46211195e-01
-9.66356039e-01 -1.72072366e-01 4.89582658e-01 2.30079472e-01
1.20938674e-01 -2.99484399e-03 -6.92810416e-01 -1.14337039e+00
-1.70271583e-02 -8.06659877e-01 3.05233985e-01 -6.26478553e-01
-1.39745414e+00 3.33662868e-01 -9.42893624e-02 -8.84452045e-01
-5.29195070e-02 -1.56112671e-01 -5.34425735e-01 7.96494603e-01
-1.83737350e+00 -8.75754416e-01 -3.84174258e-01 7.50916481e-01
8.53976905e-01 -2.58390278e-01 6.40843809e-01 2.47259945e-01
-6.29953682e-01 6.61620796e-01 -1.31795764e-01 -3.66667300e-01
1.37799716e+00 -1.38507795e+00 2.60490030e-01 7.53330112e-01
3.82942744e-02 1.01165843e+00 7.32953429e-01 -8.48288953e-01
-1.16833985e+00 -1.42930186e+00 1.13334739e+00 -6.59963191e-01
6.76579893e-01 -5.82840145e-01 -1.09512722e+00 9.45564687e-01
2.31238693e-01 -4.26069349e-02 4.21991169e-01 5.81512809e-01
-8.37485611e-01 -5.45210958e-01 -8.40154827e-01 4.52825665e-01
1.42669189e+00 -2.68960506e-01 -1.05229032e+00 3.59550297e-01
1.16002309e+00 -7.91582093e-02 -2.86502801e-02 3.14844459e-01
4.89745766e-01 -1.03582370e+00 7.40840971e-01 -8.08052361e-01
6.16413772e-01 -1.54509127e-01 1.87451512e-01 -1.82063329e+00
-5.21321595e-01 -3.86025399e-01 -4.13271159e-01 1.31035507e+00
5.14813066e-01 -9.89426017e-01 5.85836112e-01 2.49428898e-01
-1.56209543e-01 -5.51582873e-01 -7.81605065e-01 -9.94182050e-01
3.25887740e-01 -4.19171661e-01 6.12928987e-01 1.14802945e+00
-3.64268064e-01 8.13078165e-01 -5.64688563e-01 -1.55880719e-01
8.56611490e-01 5.62786357e-03 6.93931043e-01 -1.45628726e+00
-2.72401303e-01 -3.55585068e-01 4.06630516e-01 -9.82538581e-01
2.19926223e-01 -9.67455328e-01 1.18948422e-01 -1.13287807e+00
4.61107403e-01 -6.90398216e-01 -8.21412086e-01 8.82274747e-01
-5.75278759e-01 -2.32865866e-02 -6.19412437e-02 5.97615719e-01
-6.30257189e-01 7.94866621e-01 9.99191642e-01 -2.19769999e-01
-4.30983067e-01 -7.51015097e-02 -1.05515444e+00 8.04951131e-01
7.60193527e-01 -9.78919804e-01 -7.00908899e-01 -6.77865744e-01
5.15136719e-01 -5.41828632e-01 4.24469262e-01 -9.49997663e-01
2.47631907e-01 -1.19907372e-01 5.71299672e-01 -3.82033557e-01
-1.42282531e-01 -6.87700450e-01 -1.12252191e-01 5.93134284e-01
-6.04225159e-01 -2.09780812e-01 2.24639010e-03 8.17403376e-01
-1.84356961e-02 -7.01413676e-02 8.09063196e-01 -2.86865532e-01
-4.82549101e-01 1.62966847e-01 -1.54495686e-01 1.40255854e-01
9.43173587e-01 3.86415273e-01 -5.34932494e-01 4.09957767e-02
-4.79830235e-01 2.96644688e-01 2.45532468e-01 6.98448598e-01
5.32048881e-01 -1.37696064e+00 -6.97227538e-01 1.62352100e-01
6.41647056e-02 -2.86415312e-02 1.91685200e-01 8.64000618e-01
4.29369599e-01 3.23508024e-01 -5.56290038e-02 -5.10221422e-01
-8.01105797e-01 7.52394617e-01 2.29641080e-01 -4.98448253e-01
-3.53615552e-01 1.17061102e+00 3.92585218e-01 -3.38693082e-01
3.49319220e-01 -1.60558984e-01 2.27308087e-03 5.61542451e-01
5.88510334e-01 2.29537100e-01 2.66916364e-01 2.82310843e-01
-3.25170010e-01 3.89375724e-02 -5.30254722e-01 -2.18978487e-02
1.41612709e+00 -1.27526566e-01 -1.91229284e-01 1.08129859e+00
9.07300770e-01 -1.69137537e-01 -1.68156004e+00 -5.13325214e-01
1.44451737e-01 -3.92579913e-01 1.04654230e-01 -1.16969490e+00
-8.51161182e-01 9.59165752e-01 6.74967349e-01 1.74453910e-02
1.08465278e+00 -1.59386247e-01 5.60204029e-01 3.23883653e-01
5.17270744e-01 -1.13924682e+00 5.25455356e-01 8.84061873e-01
1.13362062e+00 -1.26779723e+00 1.14647649e-01 4.50158902e-02
-5.35862625e-01 7.29825199e-01 8.54993939e-01 -1.82301313e-01
6.66098416e-01 1.45547703e-01 -1.86226398e-01 -2.90754600e-03
-1.32947063e+00 -4.07245979e-02 3.29017192e-02 4.48865026e-01
4.29676026e-01 -1.75474241e-01 -4.80812013e-01 8.69329751e-01
-8.55683610e-02 3.38268816e-01 2.81493515e-01 9.17391002e-01
-6.03883982e-01 -9.31883752e-01 -1.97327375e-01 4.65919554e-01
-2.64323026e-01 -3.89041573e-01 -3.90992165e-01 6.28840029e-01
4.43535209e-01 5.18843234e-01 8.90596062e-02 -5.45612395e-01
3.46312463e-01 6.75552368e-01 3.29792053e-01 -6.74708903e-01
-5.22897482e-01 -3.24078053e-01 -2.87175089e-01 -4.12510037e-01
-6.96344003e-02 -6.12648189e-01 -9.62419748e-01 -1.65184766e-01
-1.21687844e-01 -1.55879870e-01 5.70461452e-01 8.20107162e-01
4.01841968e-01 8.89383972e-01 7.58788109e-01 -6.59818470e-01
-7.98450887e-01 -1.15538728e+00 -3.03578228e-01 3.91103059e-01
6.03319585e-01 -9.53994036e-01 -6.97869062e-01 1.81446224e-01] | [9.819576263427734, 3.43099308013916] |
456bb86b-1648-48b2-a507-070a8c27ee9c | exemplar-based-contrastive-self-supervised | 2202.02601 | null | https://arxiv.org/abs/2202.02601v1 | https://arxiv.org/pdf/2202.02601v1.pdf | Exemplar-Based Contrastive Self-Supervised Learning with Few-Shot Class Incremental Learning | Humans are capable of learning new concepts from only a few (labeled) exemplars, incrementally and continually. This happens within the context that we can differentiate among the exemplars, and between the exemplars and large amounts of other data (unlabeled and labeled). This suggests, in human learning, supervised learning of concepts based on exemplars takes place within the larger context of contrastive self-supervised learning (CSSL) based on unlabeled and labeled data. We discuss extending CSSL (1) to be based mainly on exemplars and only secondly on data augmentation, and (2) to apply to both unlabeled data (a large amount is available in general) and labeled data (a few exemplars can be obtained with valuable supervised knowledge). A major benefit of the extensions is that exemplar-based CSSL, with supervised finetuning, supports few-shot class incremental learning (CIL). Specifically, we discuss exemplar-based CSSL including: nearest-neighbor CSSL, neighborhood CSSL with supervised pretraining, and exemplar CSSL with supervised finetuning. We further discuss using exemplar-based CSSL to facilitate few-shot learning and, in particular, few-shot CIL. | ['Daniel T. Chang'] | 2022-02-05 | null | null | null | null | ['few-shot-class-incremental-learning'] | ['methodology'] | [ 2.84377396e-01 2.69865394e-01 -4.18630123e-01 -6.20784104e-01
-4.93361324e-01 -5.78433454e-01 7.65220821e-01 7.47362375e-01
-5.57260513e-01 9.32002783e-01 6.17750213e-02 8.82994309e-02
-4.64685053e-01 -9.15627658e-01 -4.55295205e-01 -5.33486426e-01
-3.27892274e-01 6.53993547e-01 4.48474109e-01 -4.51436371e-01
2.21575275e-01 6.47620380e-01 -2.08615088e+00 1.64274916e-01
7.56499112e-01 7.07321823e-01 5.47295548e-02 4.94101644e-01
-6.12048209e-01 7.31495857e-01 -3.89225215e-01 7.89339617e-02
2.22948730e-01 -4.46162671e-01 -9.38472867e-01 2.95086861e-01
5.08206189e-01 9.46431234e-02 5.64320087e-02 5.72407067e-01
2.23456562e-01 8.48091900e-01 7.47432649e-01 -1.38050628e+00
-6.99094176e-01 6.31788135e-01 -2.59998262e-01 1.34147257e-01
3.07383239e-01 7.47419521e-02 6.96684062e-01 -1.29433668e+00
8.06238711e-01 1.08471310e+00 7.87503719e-01 6.80806279e-01
-1.19364452e+00 -2.97452182e-01 3.15703839e-01 3.79489541e-01
-1.48849154e+00 -4.06686515e-01 6.81253016e-01 -1.75936252e-01
8.62315178e-01 6.86119497e-02 8.46821189e-01 8.78739834e-01
-5.20708442e-01 8.76377225e-01 1.06084549e+00 -1.03026390e+00
9.02736604e-01 3.88536453e-01 5.86523354e-01 5.42670071e-01
1.04908146e-01 3.70983094e-01 -3.77966434e-01 -6.45672251e-03
5.15505195e-01 5.77577233e-01 1.63939938e-01 -6.46377742e-01
-1.14784694e+00 8.72326851e-01 5.69335341e-01 6.19611502e-01
-2.25331262e-01 -1.55493870e-01 6.17361486e-01 5.37553906e-01
3.17635298e-01 6.35099411e-01 -5.15182912e-01 1.48926927e-02
-1.14968860e+00 -7.18548521e-02 6.54587805e-01 1.28873551e+00
1.30808330e+00 3.36432755e-01 -1.08007595e-01 1.03117967e+00
-2.52946794e-01 3.39745283e-01 1.16455448e+00 -7.78132796e-01
-2.80100107e-01 4.77890670e-01 -6.86007738e-02 -1.52969331e-01
-3.80979210e-01 -2.68084854e-01 -6.23498619e-01 1.65699393e-01
6.17078803e-02 -9.40088928e-02 -1.34738874e+00 1.55497169e+00
4.10073787e-01 5.22518039e-01 2.19013751e-01 3.15798342e-01
9.17273521e-01 5.32498360e-01 1.53839365e-01 -8.49806964e-01
7.47472703e-01 -9.33599472e-01 -4.83444303e-01 -2.56628990e-01
8.20355117e-01 -2.30307162e-01 1.36012864e+00 3.29286456e-01
-6.50113821e-01 -1.11370826e+00 -1.10319400e+00 1.84169710e-01
-1.04725707e+00 -6.12700939e-01 7.51919627e-01 5.61065793e-01
-9.11875486e-01 7.00877070e-01 -5.48699021e-01 -8.04540098e-01
3.68795365e-01 1.44871861e-01 -3.88128668e-01 -4.11109954e-01
-1.27984738e+00 8.15950990e-01 9.94468212e-01 -4.07008618e-01
-7.96372712e-01 -6.09805763e-01 -1.10163736e+00 -8.03117305e-02
6.93835080e-01 -2.24501878e-01 1.14057732e+00 -1.01837790e+00
-1.12925208e+00 6.53414011e-01 1.24157909e-02 -5.74996531e-01
4.81934175e-02 -2.09448133e-02 -7.75021911e-01 2.13231400e-01
1.63844541e-01 1.00772274e+00 8.34593236e-01 -1.35547030e+00
-7.50797451e-01 -2.23172337e-01 -5.52075505e-02 3.45098078e-01
-6.22263253e-01 -4.94122714e-01 4.71571274e-03 -5.76669931e-01
8.67469758e-02 -8.19101751e-01 -2.46983588e-01 -1.47624508e-01
-6.96706101e-02 -5.73932469e-01 1.14644241e+00 3.36449206e-01
1.07402575e+00 -2.16142273e+00 -3.00754189e-01 1.66232780e-01
-1.93107943e-03 4.66710389e-01 -4.46508735e-01 6.27150476e-01
-3.78457457e-01 -1.01922475e-01 -3.85483414e-01 -4.69374619e-02
-2.84256130e-01 6.41290545e-01 -3.42206478e-01 8.35630745e-02
9.21248868e-02 1.23399043e+00 -1.48599994e+00 -6.14851117e-01
4.14473861e-01 4.42373427e-03 -1.57600552e-01 1.12851657e-01
-1.55511528e-01 1.49595812e-01 -1.23454884e-01 7.94721186e-01
2.22609758e-01 -2.00376749e-01 -1.23016030e-01 -1.39410682e-02
-1.45662844e-01 -3.67174387e-01 -1.39057291e+00 1.81287634e+00
-4.48185295e-01 4.60536242e-01 -7.95054376e-01 -1.23172498e+00
1.02760541e+00 4.01370704e-01 3.93791676e-01 -1.58055171e-01
-1.28531978e-01 8.86996984e-02 -1.60206959e-01 -4.31148797e-01
3.72579008e-01 -7.11683750e-01 -7.45383427e-02 7.80633867e-01
7.14562654e-01 -5.00819683e-01 7.12716162e-01 4.35289770e-01
7.79115379e-01 7.68445134e-02 1.00187361e+00 -8.25807378e-02
2.94387758e-01 2.64291853e-01 3.13363731e-01 1.11431825e+00
-4.10184890e-01 5.10766387e-01 -2.80798733e-01 -6.21942818e-01
-1.01046073e+00 -1.34641600e+00 -2.88174331e-01 1.57027960e+00
3.14857848e-02 -3.22679162e-01 -2.66744703e-01 -1.09725523e+00
4.20230851e-02 1.16385818e+00 -9.62846220e-01 -4.07370031e-01
-3.10609967e-01 -4.83420789e-01 5.18474318e-02 7.73784578e-01
2.84231752e-01 -1.46963179e+00 -5.56610823e-01 3.08863640e-01
4.22729284e-01 -6.37734413e-01 -1.99953824e-01 8.32896531e-01
-1.05909956e+00 -9.67360020e-01 -5.38984060e-01 -9.72369015e-01
7.95669913e-01 4.55540746e-01 9.77576375e-01 1.94636628e-03
-5.95429063e-01 9.81491804e-01 -8.87613893e-01 -6.45488143e-01
-2.49180272e-01 -3.82319748e-01 5.97055674e-01 -1.27301991e-01
6.18174851e-01 -6.71895504e-01 -7.12105110e-02 1.33486226e-01
-1.05929327e+00 -2.27292880e-01 5.61773896e-01 1.07552886e+00
8.39822769e-01 2.11617902e-01 1.09875941e+00 -1.51041424e+00
4.07154024e-01 -6.76945627e-01 4.48798353e-04 5.57945848e-01
-8.45457375e-01 -3.22323516e-02 7.86404490e-01 -8.96048009e-01
-1.13396347e+00 2.13153958e-01 1.66262239e-01 -6.56100154e-01
-8.00134480e-01 5.14651477e-01 2.79655486e-01 1.66320875e-02
1.33204830e+00 4.41401899e-01 -1.40322104e-01 -2.07487494e-01
1.18585622e+00 7.25347042e-01 5.10317981e-01 -5.78228772e-01
8.20130825e-01 5.00710249e-01 -1.50913656e-01 -1.09065592e+00
-1.24353981e+00 -6.74991608e-01 -1.45600605e+00 -2.77408659e-01
4.05095696e-01 -6.41297281e-01 -1.86505225e-02 -9.46435481e-02
-3.89230818e-01 -3.89264077e-01 -1.60281599e+00 5.43801367e-01
-7.00683594e-01 3.07696640e-01 -4.08924431e-01 -9.08109844e-01
-6.65655062e-02 -4.50354278e-01 5.58457434e-01 3.84415925e-01
-3.46574754e-01 -1.15972376e+00 1.62319079e-01 -1.72757298e-01
2.91922510e-01 1.52178988e-01 7.72008836e-01 -1.40475547e+00
1.06478073e-01 -4.95362818e-01 2.30443463e-01 3.89028966e-01
5.59500039e-01 -1.68330386e-01 -1.01608419e+00 -5.91702461e-01
-5.04488200e-02 -9.85219955e-01 8.33929658e-01 2.54309624e-01
9.82118189e-01 -1.17481522e-01 -2.94775724e-01 7.92789683e-02
1.37377810e+00 3.07997763e-01 1.89046487e-01 -2.51677707e-02
6.01870418e-01 7.43194938e-01 9.55786705e-01 4.71920252e-01
-6.23044074e-02 2.88294610e-02 1.59517415e-02 6.12775143e-03
-2.59525508e-01 -2.02088922e-01 -1.03079572e-01 1.06604195e+00
-7.35861659e-02 2.22903430e-01 -8.89760196e-01 6.38052106e-01
-1.86977804e+00 -1.15691566e+00 1.90592945e-01 2.29633927e+00
9.39862132e-01 4.02829379e-01 2.73350030e-01 3.43980849e-01
8.51259649e-01 -1.52224764e-01 -9.70496178e-01 -3.74232292e-01
-1.62041515e-01 4.71766859e-01 7.56646914e-04 2.82737464e-01
-1.29037535e+00 1.14762151e+00 7.06445742e+00 1.03092158e+00
-8.33303034e-01 1.08808270e-02 4.73578066e-01 -2.19158694e-01
-6.49122968e-02 9.13218111e-02 -7.17064559e-01 1.97650984e-01
9.86967981e-01 -5.07878900e-01 3.14904392e-01 1.27007437e+00
-2.42416814e-01 -9.33330730e-02 -1.46015644e+00 8.00111532e-01
2.50723004e-01 -1.36196852e+00 3.04921776e-01 -4.46002185e-01
1.17442417e+00 -1.43890798e-01 -1.10945590e-01 1.07276595e+00
6.06172621e-01 -7.19436467e-01 2.14328140e-01 4.18932855e-01
1.08219361e+00 -7.85702586e-01 5.40874302e-01 5.61868250e-01
-1.31430387e+00 -4.86564904e-01 -5.89752674e-01 -6.01876564e-02
-3.31210345e-02 5.29771745e-01 -8.75148416e-01 3.57148826e-01
5.18001199e-01 1.04972601e+00 -6.81487679e-01 9.17699754e-01
-3.71719062e-01 4.82277483e-01 -1.85286365e-02 -1.66141003e-01
3.28576475e-01 7.37109035e-02 1.57842323e-01 1.25060081e+00
1.07614845e-01 4.38499242e-01 6.32154465e-01 4.69118297e-01
1.17462896e-01 2.24491119e-01 -6.75497591e-01 -3.27773020e-02
1.01709521e+00 1.27224529e+00 -9.95676994e-01 -1.03743219e+00
-5.52821100e-01 8.38873267e-01 4.36418802e-01 4.25457299e-01
-1.90758318e-01 -5.17113626e-01 1.64894070e-02 8.93414468e-02
3.51843238e-01 3.62653323e-02 8.70983973e-02 -9.24063981e-01
-6.44308925e-01 -4.93683487e-01 7.94477940e-01 -8.60961258e-01
-1.82122791e+00 4.91062105e-01 3.43757778e-01 -1.46467292e+00
-5.11214912e-01 -2.83042967e-01 -6.81827068e-01 3.52484524e-01
-1.32209015e+00 -1.09879673e+00 -3.34666938e-01 8.84997666e-01
1.04026473e+00 -3.46857995e-01 1.15129101e+00 -1.81572258e-01
-1.13033108e-01 4.40772891e-01 2.12897316e-01 -6.25964403e-02
9.18896377e-01 -1.49680817e+00 1.59747913e-01 4.94803607e-01
6.21908963e-01 8.40155780e-01 4.96587366e-01 -6.97198510e-01
-9.40801144e-01 -1.34750652e+00 6.19974971e-01 -1.91250384e-01
6.33606851e-01 -1.77908957e-01 -1.26818621e+00 7.08556414e-01
-1.82179376e-01 6.01445019e-01 1.16608620e+00 3.59292179e-01
-4.92868662e-01 -8.01821947e-02 -1.36439681e+00 4.73005861e-01
1.10751927e+00 -4.62850779e-01 -1.26692212e+00 5.74363351e-01
8.36539984e-01 1.73661545e-01 -7.60805547e-01 4.91124392e-01
3.24224293e-01 -8.11639488e-01 1.05122066e+00 -1.04146039e+00
-1.35104684e-02 -2.42365792e-01 -7.64495581e-02 -1.60068154e+00
-6.02234960e-01 -4.74557206e-02 -5.04566908e-01 9.67081547e-01
2.52273053e-01 -3.70024323e-01 8.00744176e-01 4.47654247e-01
-2.64375567e-01 -6.68363392e-01 -6.85141623e-01 -1.43722665e+00
4.04933281e-02 -3.92433107e-01 3.88945401e-01 1.41194379e+00
4.01464045e-01 4.46883529e-01 -2.14242205e-01 -4.77522790e-01
5.82273006e-01 4.02330279e-01 5.62936664e-01 -1.57990289e+00
-2.59977162e-01 -8.41109082e-02 -5.02599776e-01 -5.78210115e-01
-6.75452277e-02 -9.76360738e-01 2.04941347e-01 -1.39996433e+00
2.47139737e-01 -6.17155969e-01 -7.79911816e-01 8.94413888e-01
-4.00992304e-01 3.61398697e-01 1.40456170e-01 3.94559979e-01
-1.03483486e+00 4.63172942e-01 1.00199258e+00 -6.74176365e-02
-5.00892997e-01 -8.28073695e-02 -5.99085271e-01 7.47140110e-01
6.42068326e-01 -3.00045460e-01 -7.90406764e-01 2.90747881e-01
-1.89114094e-01 -2.94296920e-01 -5.84076345e-02 -1.25485408e+00
4.57639515e-01 -4.12280679e-01 5.99833369e-01 -4.09044772e-01
3.36484849e-01 -6.65322840e-01 -2.72575259e-01 4.64991212e-01
-6.86101556e-01 -5.99079609e-01 2.12377608e-02 7.72168219e-01
-2.38596439e-01 -7.23199010e-01 8.92423987e-01 -6.61585629e-01
-1.59125125e+00 4.65120435e-01 -2.99079806e-01 3.76050800e-01
1.27399695e+00 -6.16729617e-01 -8.81798789e-02 -1.12717398e-01
-1.41465914e+00 -8.93980265e-03 3.46686810e-01 3.73579264e-01
8.48657191e-01 -1.51458955e+00 -5.49347878e-01 3.37154478e-01
8.53909433e-01 -7.34956563e-02 2.71031171e-01 2.59520769e-01
1.12439722e-01 2.24621758e-01 -2.70506799e-01 -4.57832009e-01
-9.70995843e-01 1.35550356e+00 -2.04952940e-01 2.36588806e-01
-6.37733877e-01 1.04116607e+00 -1.08641542e-01 -5.77071428e-01
3.76082629e-01 5.49518466e-02 -4.66576278e-01 3.72950017e-01
7.33380377e-01 4.53719795e-01 -7.50934631e-02 -2.88396716e-01
-6.17458969e-02 3.57061476e-01 -2.78658926e-01 6.37934655e-02
1.44791186e+00 -1.26930267e-01 2.24095300e-01 1.40552378e+00
1.03238213e+00 -3.06497097e-01 -1.08897841e+00 -7.43353605e-01
2.06206620e-01 -2.83297092e-01 -3.09541762e-01 -8.70677829e-01
-3.62123013e-01 8.50607216e-01 7.00214863e-01 2.81850547e-01
9.90663350e-01 7.74877742e-02 4.28716630e-01 9.59811032e-01
5.63679039e-01 -1.42329860e+00 5.37559032e-01 5.10957479e-01
5.40171266e-01 -1.58897913e+00 1.59091786e-01 -1.22181781e-01
-6.96516275e-01 1.29206383e+00 8.03499818e-01 -7.42854103e-02
8.62831295e-01 1.69178005e-02 -3.31930779e-02 -1.43352196e-01
-7.17352629e-01 -5.40680587e-01 3.04774225e-01 8.41043830e-01
3.64350677e-01 2.01086700e-02 -1.70132369e-02 2.98766643e-01
9.77970064e-02 2.86255479e-02 4.05959785e-01 1.44253886e+00
-1.17841220e+00 -1.09622884e+00 -1.55166075e-01 8.13118398e-01
5.26441813e-01 -3.27911437e-01 -2.12944195e-01 8.16962004e-01
5.44081390e-01 8.96862686e-01 2.65675962e-01 -1.61778465e-01
2.63054132e-01 4.08072591e-01 4.47948515e-01 -1.43509221e+00
-2.83523440e-01 -2.16828197e-01 -3.03502560e-01 -8.32832307e-02
-5.73636055e-01 -5.42507648e-01 -1.49138999e+00 1.62342310e-01
-4.73477662e-01 4.50518459e-01 2.84461170e-01 1.15694141e+00
-4.96675260e-02 3.43812972e-01 8.66464615e-01 -8.20203483e-01
-3.18079680e-01 -1.12415171e+00 -9.76982832e-01 5.78590810e-01
1.06296264e-01 -7.38721371e-01 -4.59083825e-01 3.53995651e-01] | [10.016103744506836, 3.010403633117676] |
08c7bf55-a613-45fa-a0d3-7bd96f71876c | edgeface-efficient-face-recognition-model-for | 2307.01838 | null | https://arxiv.org/abs/2307.01838v1 | https://arxiv.org/pdf/2307.01838v1.pdf | EdgeFace: Efficient Face Recognition Model for Edge Devices | In this paper, we present EdgeFace, a lightweight and efficient face recognition network inspired by the hybrid architecture of EdgeNeXt. By effectively combining the strengths of both CNN and Transformer models, and a low rank linear layer, EdgeFace achieves excellent face recognition performance optimized for edge devices. The proposed EdgeFace network not only maintains low computational costs and compact storage, but also achieves high face recognition accuracy, making it suitable for deployment on edge devices. Extensive experiments on challenging benchmark face datasets demonstrate the effectiveness and efficiency of EdgeFace in comparison to state-of-the-art lightweight models and deep face recognition models. Our EdgeFace model with 1.77M parameters achieves state of the art results on LFW (99.73%), IJB-B (92.67%), and IJB-C (94.85%), outperforming other efficient models with larger computational complexities. The code to replicate the experiments will be made available publicly. | ['Sebastien Marcel', 'Ketan Kotwal', 'Hatef Otroshi Shahreza', 'Christophe Ecabert', 'Anjith George'] | 2023-07-04 | null | null | null | null | ['face-recognition'] | ['computer-vision'] | [-3.14756095e-01 -2.44815022e-01 -2.96858549e-01 -5.29187918e-01
-1.74975529e-01 5.46082258e-02 3.91292810e-01 -7.32879460e-01
6.32357597e-02 3.20395082e-01 5.67570329e-02 -3.39442849e-01
-2.70917505e-01 -6.75655782e-01 -5.95685959e-01 -4.19925719e-01
-3.00541371e-01 3.62687498e-01 -3.85579735e-01 1.39048472e-02
-1.39555514e-01 8.97675216e-01 -1.83137369e+00 4.52811956e-01
2.27464929e-01 1.73743558e+00 -2.32751131e-01 2.51728833e-01
-7.83715919e-02 6.19388521e-01 -2.52647847e-01 -7.91084230e-01
3.00927728e-01 1.33113965e-01 -4.54177111e-01 -2.07531080e-01
8.87105584e-01 -5.32811105e-01 -7.73645937e-01 7.06514478e-01
8.57641220e-01 -3.01186383e-01 3.60588789e-01 -1.46453571e+00
-9.27223861e-01 4.38616395e-01 -5.66064417e-01 2.09778085e-01
2.84371376e-01 -1.74881965e-01 5.30444384e-01 -1.68785703e+00
2.16772214e-01 1.29421222e+00 1.18129408e+00 7.14212298e-01
-8.59472036e-01 -1.34695840e+00 -7.28813335e-02 5.38155198e-01
-1.88654351e+00 -1.26228988e+00 4.80048358e-01 1.09530903e-01
1.42391956e+00 1.09852262e-01 5.09864748e-01 1.14578032e+00
1.23861536e-01 6.22826576e-01 8.73988807e-01 -1.89231798e-01
-1.87216893e-01 -2.57864267e-01 6.10549636e-02 1.27109921e+00
4.83301967e-01 3.16131145e-01 -1.12883282e+00 -1.30952179e-01
5.83239675e-01 3.06439936e-01 -9.98720899e-02 1.77296594e-01
-5.14976144e-01 2.71240145e-01 5.57093978e-01 7.94225931e-02
-4.94479150e-01 4.09578979e-01 2.89280534e-01 2.47199133e-01
4.48846281e-01 -5.22174418e-01 -2.00703040e-01 -2.02252105e-01
-1.02051604e+00 -4.07929361e-01 9.27423537e-01 1.21852016e+00
4.89254504e-01 4.37535405e-01 -2.73035437e-01 9.44705009e-01
6.81824267e-01 1.05408382e+00 1.25914961e-01 -5.53595006e-01
2.53587395e-01 5.45529962e-01 -3.50641400e-01 -1.18904305e+00
-4.09249544e-01 -5.87291360e-01 -1.19601882e+00 -2.84078736e-02
-2.11704209e-01 1.03146113e-01 -9.52438831e-01 1.39378846e+00
1.62314087e-01 8.09931934e-01 -1.01591907e-01 6.91531479e-01
1.48990011e+00 4.08161819e-01 6.66897669e-02 1.42675728e-01
1.56751382e+00 -1.02706206e+00 -6.31227732e-01 -2.07552329e-01
2.27799609e-01 -9.01259124e-01 6.88909709e-01 2.00258866e-01
-1.04187906e+00 -6.20741785e-01 -1.07501960e+00 -4.41436581e-02
-2.05464900e-01 7.67142177e-01 1.11931646e+00 1.20825791e+00
-1.45446146e+00 3.75382781e-01 -8.47675741e-01 -2.30845675e-01
1.07935274e+00 9.45645213e-01 -6.87957764e-01 -3.94477606e-01
-6.98568285e-01 4.77592140e-01 -2.70281047e-01 4.90357280e-01
-8.19227874e-01 -7.08812416e-01 -7.65758872e-01 3.47290903e-01
3.82612422e-02 -6.12975419e-01 9.56194937e-01 -5.62736034e-01
-1.68564916e+00 7.58565843e-01 -4.82305348e-01 -1.82752013e-01
1.48368388e-01 -3.20521563e-01 -9.59901571e-01 2.17711944e-02
-3.24880749e-01 5.80494523e-01 8.55331659e-01 -7.11665809e-01
-4.85571861e-01 -6.47202849e-01 -3.21331739e-01 -3.81701827e-01
-9.26861763e-01 2.10052371e-01 -8.51431727e-01 -3.84853154e-01
-1.03947155e-01 -9.81005013e-01 4.62609649e-01 3.82391036e-01
-1.79888740e-01 -3.97019774e-01 1.15302634e+00 -4.11892295e-01
1.18926990e+00 -2.19612980e+00 -5.90929747e-01 2.85576344e-01
2.07647890e-01 6.78147018e-01 -5.10044038e-01 1.53000608e-01
-5.16961180e-02 -1.43257320e-01 3.65481883e-01 -7.88493812e-01
-1.81166306e-02 1.69014391e-02 -3.57970931e-02 6.03371561e-01
7.90457055e-02 1.09085643e+00 -4.99789745e-01 -2.15537772e-01
4.99444306e-02 1.14202726e+00 -7.67734468e-01 3.95746566e-02
5.53616226e-01 -2.32061684e-01 -2.70864516e-01 1.23702300e+00
1.12768197e+00 -3.92633498e-01 3.03087592e-01 -6.68003440e-01
2.86036670e-01 4.74012643e-02 -1.16385877e+00 1.21951401e+00
-6.26603246e-01 6.91481531e-01 2.68413335e-01 -5.62128901e-01
1.03652024e+00 3.23281974e-01 4.51962948e-01 -7.09116697e-01
3.42311502e-01 2.84228802e-01 -2.33316049e-01 -2.25628957e-01
1.27539039e-01 2.81071156e-01 4.56914306e-01 3.18553567e-01
3.75922531e-01 8.55204463e-01 -1.24772452e-01 -1.97679460e-01
1.01583087e+00 -9.89282578e-02 -9.92909297e-02 -4.90185052e-01
6.38950169e-01 -9.08581316e-01 4.74510223e-01 7.52365649e-01
-3.97396125e-02 2.17368260e-01 -8.76633972e-02 -8.04283381e-01
-3.97945076e-01 -1.01414335e+00 -4.84694183e-01 8.04898381e-01
-1.86387245e-02 -8.73263896e-01 -6.10517025e-01 -4.21544105e-01
2.90173829e-01 -1.05215386e-01 -5.14201283e-01 -2.85964042e-01
-4.31596607e-01 -6.80442810e-01 9.17085946e-01 8.47757339e-01
1.12707460e+00 -7.84583509e-01 -1.96276754e-01 -1.69341177e-01
1.14827648e-01 -1.27088845e+00 -5.63356161e-01 -2.75113076e-01
-6.66334033e-01 -1.02462876e+00 -3.44102591e-01 -1.09255135e+00
7.34800041e-01 3.94577146e-01 1.19531107e+00 5.56150794e-01
-4.48202968e-01 3.39077652e-01 -1.00117438e-01 -2.15568274e-01
3.49552363e-01 4.77435775e-02 5.73278368e-01 2.68241793e-01
5.78581214e-01 -4.78907526e-01 -7.75382519e-01 3.99592131e-01
-4.89468992e-01 -2.19510987e-01 6.87326252e-01 1.16815436e+00
5.20227313e-01 -1.62358031e-01 4.70946521e-01 -7.46668577e-01
2.63145417e-01 -3.74964416e-01 -5.34624577e-01 4.25038844e-01
-9.50743318e-01 -2.85255939e-01 5.57341218e-01 -2.87177205e-01
-8.85802984e-01 1.14276066e-01 -4.67291266e-01 -6.27207220e-01
2.69344598e-01 2.62978226e-01 -4.29591298e-01 -7.88215935e-01
2.36276314e-01 9.89109874e-02 1.23281971e-01 -5.57089090e-01
8.89750421e-02 1.09262574e+00 6.29050732e-01 -4.14900661e-01
5.65232813e-01 5.20142853e-01 1.34950161e-01 -9.86439824e-01
-3.86078268e-01 -2.34463125e-01 -1.66892502e-02 -2.71717042e-01
6.80209026e-02 -1.28559768e+00 -1.48364449e+00 8.15845072e-01
-8.47499490e-01 -3.39383334e-02 1.64557025e-01 3.23212147e-01
-1.61161330e-02 1.11492679e-01 -9.42422688e-01 -6.95604146e-01
-9.15711999e-01 -1.06583774e+00 1.40904021e+00 4.64295775e-01
2.74063826e-01 -6.30930901e-01 -7.50811994e-01 3.58451933e-01
9.03816640e-01 -2.16377497e-01 2.83502966e-01 -3.64680588e-01
-6.08166814e-01 -3.38167965e-01 -5.17451942e-01 1.99473515e-01
2.65853554e-01 2.47530220e-03 -1.22782326e+00 -7.27219701e-01
-3.48445296e-01 -3.56278807e-01 9.05421793e-01 5.68920225e-02
1.35208702e+00 -3.73798400e-01 -6.45755827e-01 1.21249306e+00
1.22990108e+00 1.14061788e-03 9.74244416e-01 -1.04561061e-01
6.00555420e-01 5.79696335e-02 -5.46458643e-03 6.79694414e-01
3.80184531e-01 7.49072850e-01 2.09763721e-01 -1.53691053e-01
-4.23321128e-01 -2.16850773e-01 4.49519902e-01 8.25310826e-01
1.23847984e-02 -1.47500739e-01 -8.26311111e-01 2.41415337e-01
-1.39361703e+00 -9.63196874e-01 1.15006633e-01 1.76817274e+00
2.86611021e-01 -1.51522160e-01 1.18234772e-02 -4.80333492e-02
5.72669983e-01 -1.21608833e-02 -5.10671675e-01 -2.98022836e-01
-1.66687772e-01 7.47888625e-01 4.39704895e-01 1.81466728e-01
-1.03326762e+00 9.09200191e-01 6.81122255e+00 8.35592389e-01
-1.30242169e+00 1.93457931e-01 9.29755449e-01 -2.68292546e-01
1.79157972e-01 -4.39646423e-01 -1.32251191e+00 4.48116809e-01
1.13320243e+00 5.52656613e-02 6.98182225e-01 1.08415484e+00
-3.26077372e-01 5.40020883e-01 -1.08604133e+00 1.69237137e+00
4.59900618e-01 -1.55900085e+00 6.27734587e-02 1.06317393e-01
5.01289308e-01 2.74851024e-01 3.63690794e-01 2.53497481e-01
-5.93984164e-02 -1.37984312e+00 4.54594493e-01 5.26973188e-01
1.43898892e+00 -7.99793780e-01 8.49415123e-01 -2.16540873e-01
-1.57815397e+00 -2.23018140e-01 -5.64428329e-01 -3.45412716e-02
-2.66085088e-01 6.54087484e-01 -3.57119381e-01 3.42321306e-01
1.18147671e+00 8.83940876e-01 -5.28949082e-01 8.25927794e-01
1.02022007e-01 5.38829446e-01 -6.35954797e-01 1.62593499e-02
-2.40889564e-01 1.17513411e-01 -1.30026892e-01 1.31641769e+00
5.87044597e-01 2.10513592e-01 3.79789397e-02 6.09142005e-01
-7.50170350e-01 -2.00880319e-02 -6.12815320e-01 -2.98914220e-02
8.97675455e-01 1.53410995e+00 -4.69474345e-01 -1.87104419e-01
-6.51930571e-01 8.57160568e-01 4.31387037e-01 9.21875536e-02
-7.91489840e-01 -3.00184518e-01 1.08839822e+00 -6.15665838e-02
5.20499170e-01 8.10780674e-02 -5.35386279e-02 -9.42313910e-01
2.09113285e-01 -8.50860536e-01 2.92233169e-01 -5.22994518e-01
-1.33481312e+00 1.07719064e+00 -5.19444287e-01 -9.99025643e-01
1.26304135e-01 -1.14792359e+00 -4.05012459e-01 6.83550179e-01
-1.55218220e+00 -1.43301833e+00 -7.10006893e-01 6.42376840e-01
-3.96293998e-02 -6.68351471e-01 1.06088710e+00 9.02407169e-01
-9.26903844e-01 1.52604282e+00 1.04659505e-01 3.71968806e-01
4.98029321e-01 -3.36774677e-01 7.73404002e-01 6.32687747e-01
2.93256283e-01 8.59690249e-01 -9.75516587e-02 -4.24830496e-01
-2.20843530e+00 -1.23805356e+00 8.56887341e-01 -2.34249771e-01
2.39895299e-01 -4.94900852e-01 -5.13953805e-01 5.14441729e-01
-1.41597753e-02 7.23172665e-01 9.42974687e-01 2.29349360e-01
-7.84166455e-01 -8.18926573e-01 -1.32465863e+00 4.97206867e-01
1.76355362e+00 -8.47502232e-01 1.49225369e-01 2.79357880e-01
-8.94384086e-03 -3.73616457e-01 -1.03286803e+00 6.76634014e-01
1.04520488e+00 -9.47298527e-01 1.20517027e+00 -4.37395155e-01
-4.37749028e-02 -1.19560272e-01 -1.35873199e-01 -8.10203791e-01
-7.21481740e-01 -6.20995939e-01 -6.77067101e-01 1.28630018e+00
1.77964404e-01 -9.42364335e-01 1.07021189e+00 3.53680640e-01
-9.70875379e-03 -1.09870315e+00 -1.23562729e+00 -1.00951219e+00
-5.94822049e-01 -3.32489133e-01 9.98204648e-01 6.09579742e-01
-2.74981171e-01 1.16904743e-01 -5.09560943e-01 3.12629670e-01
6.73714936e-01 2.32473671e-01 5.50340056e-01 -1.19851553e+00
-8.41376483e-02 -3.47812802e-01 -8.42528522e-01 -1.04583430e+00
5.58320343e-01 -1.05177593e+00 -4.32757109e-01 -1.09233177e+00
2.51407385e-01 -5.39960504e-01 -3.60863239e-01 1.01741731e+00
1.54839769e-01 1.03987253e+00 2.07335994e-01 1.40165836e-01
-5.31907558e-01 6.66122735e-01 7.59862423e-01 -3.98167551e-01
3.03977460e-01 -3.46592039e-01 -7.80876696e-01 4.47864741e-01
6.18209898e-01 -9.24579203e-02 -2.51550972e-01 -7.26688206e-01
-2.85065323e-01 -1.99407712e-01 1.64785519e-01 -1.37239754e+00
6.47138238e-01 3.81516993e-01 9.06446338e-01 -1.32042617e-01
6.54366195e-01 -9.92120385e-01 4.75977004e-01 5.95470846e-01
3.44210386e-01 3.33892345e-01 5.66118836e-01 2.06082731e-01
-7.07690343e-02 4.52694833e-01 5.69031775e-01 5.11058927e-01
-8.04534674e-01 9.65470135e-01 2.92964503e-02 -4.12728786e-01
7.78778732e-01 -3.76546890e-01 -6.89547420e-01 -2.54487157e-01
-2.86246479e-01 1.00493198e-02 1.71023145e-01 7.82944679e-01
1.07297111e+00 -1.78718746e+00 -7.46064305e-01 9.83968914e-01
2.20415607e-01 -7.44698882e-01 2.09425300e-01 6.45726204e-01
-5.71328640e-01 5.86875737e-01 -4.19648468e-01 -4.71540153e-01
-1.45542812e+00 2.32386008e-01 4.90486115e-01 3.08059812e-01
-6.40033364e-01 1.02511108e+00 -3.37382019e-01 -1.51994407e-01
5.09104848e-01 2.16852129e-01 6.99811950e-02 -3.64460647e-01
1.08081853e+00 4.75667894e-01 5.43343127e-01 -9.47208047e-01
-8.85074735e-01 6.90046787e-01 -4.68340181e-02 7.80479252e-01
1.40558898e+00 2.42444336e-01 -3.03246945e-01 -4.42491472e-01
1.36536086e+00 -7.01547042e-02 -9.42698359e-01 -1.16327479e-01
-2.18471646e-01 -5.83028436e-01 2.88638443e-01 -7.41760910e-01
-1.73875725e+00 5.31680882e-01 1.28231108e+00 -4.95065540e-01
1.43033624e+00 -2.19463035e-01 1.00780594e+00 6.48894250e-01
8.62876356e-01 -7.40117490e-01 -5.11439703e-02 5.95867574e-01
6.63496912e-01 -1.09664083e+00 -6.33201674e-02 -7.22409844e-01
1.23832345e-01 1.05185068e+00 7.66686141e-01 6.85940906e-02
1.32531536e+00 8.62258852e-01 2.04153582e-02 -3.22148889e-01
-8.42803657e-01 -9.98449745e-04 3.75742495e-01 5.51783442e-01
5.57984591e-01 8.96737203e-02 6.24811044e-03 6.99535966e-01
-3.63422669e-02 3.00522566e-01 -2.40067869e-01 8.15654695e-01
7.40061402e-02 -1.16131020e+00 -1.74095422e-01 7.50759304e-01
-5.17582834e-01 -2.95980364e-01 -4.24337834e-01 4.96508092e-01
1.05492234e-01 1.06959224e+00 1.65096819e-01 -8.97281766e-01
3.49719495e-01 6.99264109e-02 5.65432549e-01 -1.37578398e-01
-6.12374723e-01 -4.63015437e-01 1.99059453e-02 -9.56779063e-01
-1.94302827e-01 -2.47107103e-01 -9.80036855e-01 -1.09377337e+00
-4.39885676e-01 -6.41659573e-02 7.89461911e-01 6.15441561e-01
1.25001287e+00 2.92589277e-01 6.52605653e-01 -8.58155847e-01
-4.27386850e-01 -8.37351978e-01 -5.82842112e-01 2.54119754e-01
1.38483629e-01 -8.94619048e-01 -8.48465133e-03 -2.33474985e-01] | [13.31177806854248, 0.7361772656440735] |
5f44fc15-baca-4eaf-add0-abd56802dc50 | bencoref-a-multi-domain-dataset-of-nominal | 2304.03682 | null | https://arxiv.org/abs/2304.03682v3 | https://arxiv.org/pdf/2304.03682v3.pdf | BenCoref: A Multi-Domain Dataset of Nominal Phrases and Pronominal Reference Annotations | Coreference Resolution is a well studied problem in NLP. While widely studied for English and other resource-rich languages, research on coreference resolution in Bengali largely remains unexplored due to the absence of relevant datasets. Bengali, being a low-resource language, exhibits greater morphological richness compared to English. In this article, we introduce a new dataset, BenCoref, comprising coreference annotations for Bengali texts gathered from four distinct domains. This relatively small dataset contains 5200 mention annotations forming 502 mention clusters within 48,569 tokens. We describe the process of creating this dataset and report performance of multiple models trained using BenCoref. We expect that our work provides some valuable insights on the variations in coreference phenomena across several domains in Bengali and encourages the development of additional resources for Bengali. Furthermore, we found poor crosslingual performance at zero-shot setting from English, highlighting the need for more language-specific resources for this task. | ['Nabeel Mohammed', 'Mohammad Mamun Or Rashid', 'Mojammel Hossain', 'Shadman Rohan'] | 2023-04-07 | null | null | null | null | ['coreference-resolution'] | ['natural-language-processing'] | [-8.53983015e-02 7.37060383e-02 -3.12316149e-01 -3.81151229e-01
-1.39135551e+00 -9.98700380e-01 7.98894942e-01 1.90536648e-01
-5.69056988e-01 9.61558104e-01 9.09068525e-01 -9.87491235e-02
-2.22970068e-01 -5.28202653e-01 -3.89213085e-01 -6.71164036e-01
1.24320671e-01 1.15108573e+00 2.01075196e-01 -5.85797489e-01
4.09203887e-01 5.62389374e-01 -1.10901237e+00 4.43030447e-01
8.73874128e-01 1.21928649e-02 3.94693226e-01 2.88810134e-01
-3.49137858e-02 6.60996974e-01 -7.11123228e-01 -7.82023191e-01
5.34462696e-03 -4.70715702e-01 -1.53544581e+00 -5.59515953e-01
6.39686942e-01 2.92345405e-01 -4.60089773e-01 9.80732918e-01
9.46883082e-01 3.97233367e-01 4.13401544e-01 -5.94140112e-01
-6.06830657e-01 1.27928555e+00 -4.76892233e-01 7.20333576e-01
5.48997045e-01 -4.14768457e-01 1.29403889e+00 -6.76143885e-01
1.23525739e+00 1.26160109e+00 5.69747150e-01 6.77795410e-01
-9.39969838e-01 -9.28109050e-01 -3.65494490e-01 6.13928914e-01
-1.31221855e+00 -8.66323411e-01 5.54346383e-01 1.20473152e-03
1.08342791e+00 4.57943171e-01 1.31584238e-02 1.10265088e+00
-4.02703732e-01 6.67234123e-01 1.06558466e+00 -6.60810113e-01
-2.30243891e-01 6.34511411e-02 2.32914947e-02 -1.50529342e-03
2.53034443e-01 -9.56087336e-02 -7.12639332e-01 -2.63774842e-01
3.74202847e-01 -5.23639679e-01 -5.18944860e-01 -2.45017305e-01
-1.11830294e+00 9.84567821e-01 1.39848128e-01 7.74817407e-01
-1.36678591e-01 -6.68574572e-01 7.17866600e-01 3.35357577e-01
2.54130036e-01 7.89175987e-01 -5.65708756e-01 -4.40091968e-01
-9.72029865e-01 3.02465528e-01 1.11644173e+00 1.07059276e+00
3.55933040e-01 -3.03916335e-01 1.66646525e-01 1.27695465e+00
-2.37695649e-01 4.24169242e-01 2.01382622e-01 -1.03383553e+00
8.37676227e-01 4.05387014e-01 1.42981067e-01 -8.84893477e-01
-4.43808377e-01 -1.41109908e-02 -4.36996937e-01 -5.36807239e-01
6.40081465e-01 -1.24692157e-01 -2.49699652e-01 1.81015635e+00
2.09592670e-01 -2.45526209e-01 5.44167578e-01 9.02415931e-01
1.26435483e+00 3.59159708e-01 1.10970058e-01 -3.77616614e-01
1.59656763e+00 -6.67866290e-01 -6.55351579e-01 2.45061982e-02
5.43664336e-01 -1.16568732e+00 7.36780107e-01 1.75700009e-01
-1.20111203e+00 -5.34054972e-02 -6.19046807e-01 -2.01441467e-01
-2.57274032e-01 -1.92686468e-01 5.30040085e-01 5.24748504e-01
-5.43127120e-01 3.49465042e-01 -6.30569756e-01 -9.77966309e-01
4.74806018e-02 2.91250855e-01 -7.87636578e-01 -3.10945600e-01
-1.60598505e+00 1.37438309e+00 5.83603084e-01 -2.79363871e-01
-5.37277877e-01 -7.33739674e-01 -8.33430231e-01 -1.89886168e-01
5.46584547e-01 1.87323000e-02 1.18053734e+00 -4.42179471e-01
-9.89581823e-01 1.34607506e+00 -2.95505047e-01 -4.20077026e-01
3.11464190e-01 -4.16215092e-01 -8.36809099e-01 1.61951423e-01
4.48980987e-01 3.76203746e-01 -7.38073364e-02 -1.08784997e+00
-9.55224156e-01 -3.60643417e-01 7.40796328e-02 4.17045861e-01
1.11380130e-01 7.14572132e-01 -4.14982557e-01 -6.42771125e-01
2.33856309e-02 -9.10181224e-01 1.67201847e-01 -1.27690005e+00
-9.06255394e-02 -5.30993879e-01 6.90076709e-01 -7.41256535e-01
1.24261427e+00 -2.01874614e+00 2.17331816e-02 -1.37929708e-01
-1.92693278e-01 3.97628158e-01 -6.42514452e-02 8.75800610e-01
-1.86929852e-01 1.37707829e-01 -2.03166068e-01 2.32125744e-01
-1.03085987e-01 3.11476558e-01 -3.16582382e-01 5.52190304e-01
1.39561802e-01 6.86911225e-01 -1.22009289e+00 -6.71099305e-01
3.16623338e-02 4.17895079e-01 -4.52304542e-01 -9.88354161e-03
2.76606411e-01 4.52883571e-01 -4.09704864e-01 7.07858503e-01
7.35730708e-01 4.73033875e-01 6.67573273e-01 -3.89775097e-01
-5.82297564e-01 8.71803463e-01 -1.07399845e+00 1.80087483e+00
-2.92242587e-01 7.38993406e-01 1.31340772e-01 -1.03496790e+00
9.62000012e-01 5.72251201e-01 3.81280392e-01 -8.05746138e-01
1.91490054e-01 4.03473258e-01 3.33980769e-01 -5.72469652e-01
1.01443732e+00 -2.24641845e-01 -6.17430687e-01 3.96054089e-01
1.98563367e-01 -1.46522298e-01 4.91653591e-01 2.62839168e-01
1.08760870e+00 -5.35783283e-02 4.08950806e-01 -5.33425570e-01
5.00891566e-01 5.52367568e-01 9.00483608e-01 7.38769054e-01
-2.83166885e-01 5.47700167e-01 2.08309799e-01 -1.52915254e-01
-9.86637712e-01 -7.41302669e-01 -7.19189942e-01 1.20821214e+00
-7.82577172e-02 -5.77966154e-01 -5.13390541e-01 -5.00448287e-01
-3.62439692e-01 8.05079222e-01 -4.78227586e-01 3.36447954e-01
-1.26445484e+00 -8.83231401e-01 1.17002928e+00 3.33637863e-01
5.76719046e-01 -1.27541924e+00 -4.83133286e-01 2.91081578e-01
-9.17242527e-01 -1.14596808e+00 -3.52661103e-01 3.51086050e-01
-5.13300240e-01 -1.44904184e+00 -5.25508702e-01 -1.08372188e+00
-7.04509988e-02 2.10457698e-01 1.36019862e+00 -3.09263945e-01
-6.49352521e-02 1.68383554e-01 -5.77696443e-01 -4.62005913e-01
-5.11457741e-01 4.38210785e-01 3.49263698e-02 -8.20408881e-01
1.25750470e+00 -2.48658627e-01 -4.27714549e-02 1.33791327e-01
-4.25446182e-01 -5.77997565e-01 3.28090906e-01 8.10570061e-01
3.82966965e-01 -1.79730624e-01 5.96854687e-01 -1.33017468e+00
6.13451421e-01 -6.29305661e-01 -2.08913177e-01 3.10540825e-01
-9.70021188e-02 -5.11862487e-02 3.09938073e-01 -2.14631319e-01
-1.58387256e+00 -3.74872625e-01 -1.33210987e-01 3.49916667e-01
-3.58390659e-01 4.56290692e-01 -4.67042297e-01 1.94318518e-01
6.16857231e-01 -2.10442558e-01 -5.57093322e-01 -7.14968562e-01
4.08474356e-01 1.00078821e+00 1.07895982e+00 -1.04779780e+00
3.86775553e-01 2.52212226e-01 -4.28827047e-01 -1.07909930e+00
-8.82346570e-01 -9.03228045e-01 -1.01429176e+00 3.29154313e-01
5.71584523e-01 -9.91219819e-01 -5.70449829e-01 1.23627279e-02
-1.03342474e+00 -1.59013811e-02 -6.03808202e-02 6.62949860e-01
-3.19177151e-01 4.18072760e-01 -7.21648097e-01 -4.49878305e-01
-5.33554912e-01 -9.09414172e-01 4.99592751e-01 3.07076871e-01
-7.84056842e-01 -9.24026191e-01 6.32811010e-01 6.70109570e-01
-2.02290639e-02 9.27607268e-02 8.75709951e-01 -1.23231471e+00
1.01047873e-01 7.38056228e-02 -1.36059538e-01 -3.09426844e-01
2.41775274e-01 -3.33183445e-02 -8.00376892e-01 -2.49161854e-01
-2.98938490e-02 -5.64206898e-01 6.81774557e-01 7.71875381e-02
1.66944623e-01 -1.05663322e-01 -4.03495252e-01 2.08403379e-01
1.26278925e+00 3.41831207e-01 5.44120491e-01 1.13391435e+00
4.24811065e-01 7.04982162e-01 1.00755727e+00 1.70667008e-01
6.79320276e-01 6.82229400e-01 -1.21385381e-01 1.89476132e-01
-1.50436789e-01 3.15603577e-02 1.36960521e-01 1.01392472e+00
-4.66383755e-01 1.66357141e-02 -1.14265132e+00 1.20374346e+00
-1.60831594e+00 -1.41842866e+00 -2.63521552e-01 2.09157801e+00
1.38835335e+00 -2.20022440e-01 3.79929245e-02 -1.03393555e-01
9.69921649e-01 6.68796822e-02 -9.17913541e-02 -5.49415886e-01
-6.15640283e-01 5.84550202e-01 3.35876703e-01 4.41596746e-01
-1.14823914e+00 1.29941046e+00 6.19074869e+00 5.78160942e-01
-8.15205038e-01 1.63595662e-01 -2.29276061e-01 -3.19486745e-02
-8.24647993e-02 2.60831922e-01 -9.81363356e-01 1.17001891e-01
1.06458139e+00 -2.42171273e-01 1.30093783e-01 3.90112519e-01
1.04095228e-01 8.53854269e-02 -9.45018709e-01 6.96646154e-01
2.33792171e-01 -1.20343769e+00 -1.80256128e-01 -1.97105467e-01
7.19208241e-01 5.69065511e-01 -4.61255312e-01 3.84261340e-01
7.74845660e-01 -8.70711207e-01 4.40898776e-01 -1.02842942e-01
6.96236253e-01 -1.19469416e+00 9.59113359e-01 1.29962221e-01
-9.86361444e-01 2.69528091e-01 -7.11828709e-01 2.45406196e-01
2.42062867e-01 -4.17074263e-02 -7.57405341e-01 8.66484821e-01
1.00292063e+00 5.55281758e-01 -3.15601051e-01 9.22020435e-01
-2.90156066e-01 7.58056939e-01 -1.49638668e-01 3.24567288e-01
3.31292897e-01 -2.81886924e-02 7.32945919e-01 1.80896199e+00
-7.74247805e-03 4.68014032e-01 6.46108687e-02 2.65327692e-01
-1.86114490e-01 3.18252683e-01 -3.76657903e-01 5.02160601e-02
1.24787033e+00 1.26431596e+00 -5.27881920e-01 -7.45134428e-02
-6.90623701e-01 6.88762188e-01 6.49654090e-01 3.36280800e-02
-6.23699903e-01 -6.26846671e-01 7.16671288e-01 -2.25639343e-01
1.89062580e-01 4.91047390e-02 1.72842234e-01 -1.19611454e+00
-5.40097892e-01 -1.39012361e+00 1.06998336e+00 -2.64164090e-01
-1.21520996e+00 6.59393489e-01 2.52231359e-01 -6.63325250e-01
-2.18599945e-01 -1.76889881e-01 -4.38300610e-01 8.73172700e-01
-1.37443984e+00 -1.12987280e+00 1.94280129e-02 7.29901552e-01
6.61109984e-01 -2.40113690e-01 1.18111980e+00 4.21530575e-01
-5.76514602e-01 7.36282051e-01 2.18314931e-01 6.44712627e-01
1.45398867e+00 -1.19707537e+00 3.83749753e-01 7.67414033e-01
3.44229937e-01 9.39107060e-01 9.69669640e-01 -6.13033891e-01
-1.05054915e+00 -6.55007899e-01 1.45051920e+00 -7.20606148e-01
8.05058360e-01 4.74611968e-02 -1.11619699e+00 8.38651180e-01
6.98303938e-01 -6.26753449e-01 1.00185406e+00 5.82881868e-01
-4.09193248e-01 4.12358880e-01 -1.25600863e+00 4.50519413e-01
1.06506205e+00 -6.29076838e-01 -1.61684382e+00 -6.46017566e-02
2.89529443e-01 -5.71651042e-01 -1.29525721e+00 4.02766913e-01
2.23030716e-01 -7.27720618e-01 7.64469326e-01 -7.22821474e-01
1.13565095e-01 -4.72438373e-02 -4.40234393e-01 -1.39468026e+00
-5.41537046e-01 -5.11007011e-01 1.89461946e-01 1.94757485e+00
4.11393195e-01 -3.36113632e-01 5.73634803e-01 4.81742352e-01
-2.59796202e-01 1.19374529e-01 -9.05659795e-01 -4.51099634e-01
6.67687535e-01 -2.25734249e-01 5.92153370e-01 1.61511552e+00
5.65561116e-01 6.40031517e-01 -2.17610404e-01 4.89637330e-02
5.50368309e-01 4.08921003e-01 5.15782833e-01 -1.23022926e+00
2.46179894e-01 -1.69149727e-01 -2.95218057e-03 -5.07148266e-01
6.01171017e-01 -1.05404341e+00 -1.43285329e-02 -1.23225701e+00
5.90165973e-01 -6.46236479e-01 2.68325489e-02 5.87832630e-01
-1.65639222e-01 3.30800205e-01 1.20523870e-01 4.57304984e-01
-4.26497012e-01 2.92245317e-02 8.48306835e-01 -9.77600217e-02
-1.66079819e-01 -3.70749444e-01 -9.71803784e-01 7.22159505e-01
7.73616970e-01 -6.01647258e-01 5.62507585e-02 -4.78802592e-01
-7.77500272e-02 -1.36355475e-01 -4.05716330e-01 -4.23023105e-01
3.47265124e-01 -4.17784005e-01 1.51199594e-01 -7.44055629e-01
1.18046487e-02 -4.50106949e-01 9.83461589e-02 7.03019127e-02
-1.40341192e-01 1.33031294e-01 3.26985329e-01 -3.28408293e-02
-4.79441196e-01 -3.70903730e-01 7.01017320e-01 -4.87229109e-01
-1.20649898e+00 -1.50547221e-01 -2.37694383e-01 9.07847047e-01
6.17281139e-01 -4.99570891e-02 -5.09467483e-01 7.70204291e-02
-5.15097678e-01 2.71660745e-01 5.62073350e-01 7.34588802e-01
1.07251391e-01 -1.13661873e+00 -1.24087811e+00 -2.97279745e-01
2.25550011e-01 -1.97228491e-01 1.85893118e-01 6.21483564e-01
-3.04355204e-01 7.54191935e-01 -3.65323514e-01 -1.47595599e-01
-1.62762856e+00 4.27814513e-01 -4.10136096e-02 -2.28680804e-01
-7.90369093e-01 6.66685939e-01 -1.75645009e-01 -8.13223183e-01
4.91260700e-02 5.66755116e-01 -7.84847856e-01 5.59180021e-01
5.69910347e-01 6.48425341e-01 1.64834782e-01 -1.32960522e+00
-8.15158606e-01 2.71198511e-01 -4.68182623e-01 -4.67526317e-02
1.49085760e+00 -3.29726160e-01 -2.81499445e-01 2.58015662e-01
7.19800889e-01 6.15273535e-01 -4.66229498e-01 -2.05289036e-01
6.12559557e-01 -3.75768930e-01 -3.06628525e-01 -7.60112464e-01
-7.34068751e-01 4.46494639e-01 1.67701364e-01 -1.56364009e-01
8.00915599e-01 4.12182748e-01 6.25135362e-01 4.15691853e-01
4.65134978e-01 -1.13894558e+00 -5.27716339e-01 1.07147360e+00
7.28276134e-01 -1.15679073e+00 -9.20638442e-02 -1.40999928e-01
-8.24850500e-01 7.94928491e-01 5.08107185e-01 1.22952461e-01
1.41197011e-01 4.06781673e-01 4.19134408e-01 -3.91625553e-01
-5.54833591e-01 -4.04453367e-01 -2.47704517e-03 8.42571497e-01
1.19626760e+00 7.03535601e-02 -8.43318462e-01 8.26376319e-01
-7.01319814e-01 -3.30428034e-01 6.81936443e-01 8.00997674e-01
-3.00051332e-01 -1.55204451e+00 -4.73577142e-01 9.64439139e-02
-1.19275820e+00 -4.92367893e-01 -6.96395159e-01 1.25058794e+00
3.27269104e-03 1.15214074e+00 1.13042995e-01 1.41403705e-01
5.22353172e-01 5.88612817e-03 6.96430624e-01 -7.16701806e-01
-7.43136525e-01 -1.08469293e-01 7.34087288e-01 -1.01533279e-01
-9.33567584e-01 -1.10726178e+00 -1.32151890e+00 -7.29649127e-01
-5.12627184e-01 9.25992548e-01 4.69628945e-02 9.84204829e-01
-4.66680825e-02 2.77865231e-02 1.04981951e-01 -4.82303381e-01
-1.26664758e-01 -1.14586627e+00 -6.06175900e-01 6.05298579e-01
-1.58688769e-01 -6.56092763e-01 -1.59304857e-01 -1.42907977e-01] | [9.327056884765625, 9.485779762268066] |
642b9df0-2aed-4b8f-ae93-bba4129775e6 | kerm-knowledge-enhanced-reasoning-for-vision | 2303.15796 | null | https://arxiv.org/abs/2303.15796v1 | https://arxiv.org/pdf/2303.15796v1.pdf | KERM: Knowledge Enhanced Reasoning for Vision-and-Language Navigation | Vision-and-language navigation (VLN) is the task to enable an embodied agent to navigate to a remote location following the natural language instruction in real scenes. Most of the previous approaches utilize the entire features or object-centric features to represent navigable candidates. However, these representations are not efficient enough for an agent to perform actions to arrive the target location. As knowledge provides crucial information which is complementary to visible content, in this paper, we propose a Knowledge Enhanced Reasoning Model (KERM) to leverage knowledge to improve agent navigation ability. Specifically, we first retrieve facts (i.e., knowledge described by language descriptions) for the navigation views based on local regions from the constructed knowledge base. The retrieved facts range from properties of a single object (e.g., color, shape) to relationships between objects (e.g., action, spatial position), providing crucial information for VLN. We further present the KERM which contains the purification, fact-aware interaction, and instruction-guided aggregation modules to integrate visual, history, instruction, and fact features. The proposed KERM can automatically select and gather crucial and relevant cues, obtaining more accurate action prediction. Experimental results on the REVERIE, R2R, and SOON datasets demonstrate the effectiveness of the proposed method. | ['Shuqiang Jiang', 'YaoWei Wang', 'Jiahao Yang', 'Zihan Wang', 'Xiangyang Li'] | 2023-03-28 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Li_KERM_Knowledge_Enhanced_Reasoning_for_Vision-and-Language_Navigation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Li_KERM_Knowledge_Enhanced_Reasoning_for_Vision-and-Language_Navigation_CVPR_2023_paper.pdf | cvpr-2023-1 | ['vision-and-language-navigation'] | ['robots'] | [-2.02339053e-01 -2.87211627e-01 -1.16869472e-01 -3.57705146e-01
-3.03066075e-01 -4.67081517e-01 7.52274215e-01 2.56695688e-01
-4.71053481e-01 6.90539360e-01 4.84297097e-01 -1.25659645e-01
-4.14726794e-01 -8.87887299e-01 -6.01141870e-01 -5.71093321e-01
3.37646855e-03 9.12368521e-02 6.82967126e-01 -4.92835552e-01
6.73748016e-01 5.49099386e-01 -1.98836553e+00 3.32089275e-01
1.07401633e+00 1.02219951e+00 7.83794701e-01 1.92751929e-01
-3.38445455e-01 1.24386680e+00 -2.63399482e-01 1.97405398e-01
-6.36283830e-02 -2.82626897e-01 -7.27596998e-01 -1.60580501e-01
1.38608977e-01 -6.74607217e-01 -3.35885048e-01 9.51290607e-01
1.99628830e-01 8.37242365e-01 5.55881143e-01 -1.17305303e+00
-7.18687713e-01 3.43768924e-01 -1.62616596e-01 2.73596495e-01
9.75724459e-01 3.84010047e-01 6.19176149e-01 -1.11390114e+00
6.76883042e-01 1.47436249e+00 -1.08927272e-01 2.82418042e-01
-4.79065210e-01 -3.13592136e-01 8.32893193e-01 8.67885351e-01
-1.26584518e+00 -4.31776911e-01 8.45221221e-01 -3.65464985e-01
8.44843090e-01 1.01390108e-01 9.71582472e-01 8.98407578e-01
9.61857364e-02 1.09783745e+00 1.10756814e+00 -2.89083272e-01
3.13954711e-01 1.39636308e-01 1.51649922e-01 1.19538975e+00
1.96801916e-01 2.35525712e-01 -1.04261613e+00 1.97740123e-01
1.01034760e+00 2.38372833e-01 -5.30078173e-01 -7.02988565e-01
-1.49326944e+00 5.56398332e-01 8.07393610e-01 2.30974749e-01
-6.41119182e-01 -2.46005617e-02 1.47205353e-01 1.75643601e-02
-2.86110967e-01 2.11304754e-01 -2.88578123e-01 -1.51623428e-01
-5.04584163e-02 -6.19173609e-02 3.83937865e-01 1.23354053e+00
9.54460144e-01 1.37684122e-02 -2.88480043e-01 5.13751388e-01
5.37363946e-01 5.13718963e-01 6.50910139e-01 -1.04634786e+00
4.59571779e-01 1.05299199e+00 2.45222554e-01 -1.29599988e+00
-5.72171628e-01 -3.51170078e-02 -3.82268578e-01 2.89541721e-01
6.74005374e-02 1.99483514e-01 -8.81138921e-01 1.56272960e+00
6.07787371e-01 -4.93852124e-02 3.23260069e-01 1.17335212e+00
1.34564328e+00 5.90225041e-01 2.33164728e-01 -2.77278628e-02
1.41569340e+00 -1.25233698e+00 -8.12133312e-01 -3.91776770e-01
5.77549100e-01 -2.68289298e-01 1.13761330e+00 3.01721275e-01
-6.33936286e-01 -7.77745664e-01 -7.78575361e-01 -1.72959626e-01
-8.27935159e-01 1.99753627e-01 8.30936074e-01 -2.00840598e-03
-8.90121639e-01 4.60972451e-02 -5.72797060e-01 -6.23014212e-01
2.06809089e-01 1.00842349e-01 -5.50722063e-01 -2.18630776e-01
-1.14758444e+00 1.04681182e+00 8.08216751e-01 2.68834740e-01
-1.20725942e+00 -2.60682255e-01 -1.09099340e+00 -9.07934532e-02
8.46072674e-01 -7.57709503e-01 8.39631140e-01 -5.87967753e-01
-1.53971040e+00 3.00102890e-01 -1.76822931e-01 -9.84034911e-02
1.21368676e-01 -2.61459380e-01 -4.41460818e-01 2.76375145e-01
1.66718617e-01 6.77479088e-01 3.98398966e-01 -1.69080460e+00
-1.14655840e+00 -3.58749866e-01 6.49300158e-01 8.13376069e-01
-1.41446233e-01 -4.66974169e-01 -6.91385686e-01 -2.03644723e-01
4.97486770e-01 -5.43257654e-01 -1.35405138e-01 9.01794434e-02
-2.32910678e-01 -3.27333212e-01 6.73966527e-01 -7.65773535e-01
1.01948428e+00 -1.93141532e+00 1.85661659e-01 3.26140016e-01
4.86216396e-02 1.48244444e-02 -1.51643366e-01 4.79888052e-01
3.31229180e-01 -3.39990139e-01 3.18245798e-01 2.80896962e-01
-1.12195216e-01 2.25873277e-01 -8.48556086e-02 -3.59801874e-02
-4.02165979e-01 9.55709040e-01 -1.20987070e+00 -6.00197315e-01
6.20981574e-01 4.80872989e-01 -3.93919796e-01 1.96066886e-01
-3.07978570e-01 7.01498389e-01 -1.09187317e+00 6.70284092e-01
2.46259332e-01 3.03304475e-02 -1.19181529e-01 -3.59052300e-01
-2.89364964e-01 1.43750831e-01 -1.22420478e+00 2.00786114e+00
-6.79028094e-01 3.00585777e-01 -1.54936060e-01 -4.52478439e-01
9.60376203e-01 -1.04247786e-01 1.24304302e-01 -1.08164608e+00
8.91488418e-02 1.27979210e-02 -1.91410914e-01 -9.51244533e-01
4.82930869e-01 5.98726451e-01 4.53860201e-02 1.08736277e-01
-1.46275416e-01 3.39309067e-01 1.62038222e-01 2.72489667e-01
9.36309218e-01 7.65825629e-01 7.29620576e-01 -4.24255803e-02
8.60611260e-01 2.25013956e-01 3.87135297e-01 7.42306411e-01
-3.22199136e-01 -1.68078557e-01 -1.30762249e-01 -5.27909398e-01
-1.00287408e-01 -9.96656418e-01 5.42966008e-01 1.26181293e+00
9.24470127e-01 -3.81328285e-01 -4.04766113e-01 -7.46027589e-01
-3.18240613e-01 1.02163780e+00 -6.79853082e-01 -3.82863790e-01
-6.13736689e-01 -4.15871479e-02 -2.95596451e-01 6.29846573e-01
9.63418126e-01 -1.49795902e+00 -1.19195318e+00 2.97953673e-02
-4.42214340e-01 -8.58031034e-01 -2.98501641e-01 3.47942598e-02
-6.39646292e-01 -1.14301562e+00 -1.09154373e-01 -9.85917568e-01
9.62570846e-01 7.32394695e-01 8.08512151e-01 3.74303848e-01
-2.83171330e-02 8.34289789e-01 -6.54305220e-01 5.00400690e-03
1.20673127e-01 -4.04886454e-01 2.75519304e-02 4.97165732e-02
3.19734722e-01 -3.99125725e-01 -9.76081192e-01 3.35737914e-01
-5.01809239e-01 4.74653572e-01 8.39090765e-01 5.79919577e-01
7.94774890e-01 2.73638695e-01 2.49743611e-01 -3.66664350e-01
6.75744355e-01 -3.77588600e-01 -3.89316738e-01 6.58942282e-01
-3.70139569e-01 1.31347567e-01 4.82843608e-01 -3.31061244e-01
-1.57835221e+00 1.50568217e-01 3.10146421e-01 -1.88103318e-01
-5.87959409e-01 8.57291460e-01 -5.72458208e-01 -8.85899067e-02
3.94256383e-01 7.20112383e-01 -4.39443797e-01 -2.61363357e-01
7.06552744e-01 1.65941581e-01 4.94009167e-01 -7.31609881e-01
4.82222468e-01 4.88068104e-01 -4.32574786e-02 -3.90700817e-01
-7.01381266e-01 -4.83952612e-01 -6.95336640e-01 -5.51486492e-01
8.02332222e-01 -8.21516216e-01 -1.11643457e+00 3.08942404e-02
-1.04804027e+00 -2.21952155e-01 -1.58688948e-01 6.82119727e-01
-7.21961915e-01 1.72116995e-01 -1.23673677e-01 -8.49596918e-01
9.66579281e-03 -1.10567594e+00 8.01714361e-01 6.25360966e-01
1.99656323e-01 -8.23576689e-01 -2.82491744e-01 3.63167256e-01
2.54633009e-01 5.99200279e-02 9.05941486e-01 -5.15894651e-01
-1.11812377e+00 2.25094240e-02 -2.48260036e-01 -5.49847364e-01
3.49274516e-01 -2.34702244e-01 -5.06236970e-01 3.77978757e-02
-1.51768520e-01 -2.16546059e-01 8.54916871e-01 1.10563517e-01
1.00595820e+00 -2.76791364e-01 -7.11765826e-01 3.15449148e-01
1.32669353e+00 9.86050308e-01 4.76191819e-01 6.00798011e-01
7.12119460e-01 6.87164783e-01 1.17909849e+00 5.15157700e-01
8.46115172e-01 6.72679126e-01 5.32093346e-01 2.51777649e-01
-1.63626984e-01 -5.05620003e-01 3.93359125e-01 6.19985402e-01
-3.79487127e-01 -6.42104074e-02 -7.73690104e-01 4.34364200e-01
-2.22634006e+00 -1.03886962e+00 2.97690574e-02 1.92857039e+00
5.90028703e-01 2.48454250e-02 -2.01669812e-01 -3.76773804e-01
4.07070398e-01 -5.01675010e-02 -8.15627158e-01 1.27268657e-01
1.27235204e-01 -5.20092130e-01 6.63037598e-02 5.51908731e-01
-8.09512615e-01 1.25275731e+00 4.95035410e+00 8.06405544e-01
-5.24223328e-01 -9.78351757e-02 -1.84263363e-02 7.11258873e-02
-3.56732756e-01 5.31025901e-02 -7.87152529e-01 1.23323843e-01
1.27330467e-01 -1.31119296e-01 7.03700006e-01 1.02156639e+00
3.47972453e-01 -7.25493789e-01 -1.02582455e+00 1.05638099e+00
2.86182940e-01 -1.01857901e+00 4.91490930e-01 -2.63417900e-01
4.20485884e-01 -5.36620080e-01 1.24024034e-01 5.45521915e-01
5.22472978e-01 -7.89302707e-01 7.53672719e-01 1.11871028e+00
1.91236690e-01 -7.06720829e-01 5.04168630e-01 5.69247305e-01
-1.59069312e+00 -2.69505650e-01 -2.44216517e-01 6.11939430e-02
1.14133902e-01 -1.07742563e-01 -6.52845025e-01 8.99899840e-01
8.74669671e-01 6.87396586e-01 -7.68539369e-01 1.00453866e+00
-5.49781621e-01 -9.11689848e-02 -1.15635239e-01 -3.56807321e-01
2.63491720e-01 -2.98246026e-01 3.93860161e-01 5.26449203e-01
3.98428738e-01 5.13247788e-01 4.57832992e-01 6.60522342e-01
4.89284545e-01 2.45262906e-01 -5.83078682e-01 4.85885024e-01
5.96103251e-01 1.15983272e+00 -9.01274264e-01 -3.00753593e-01
-5.34712195e-01 9.35449004e-01 6.30662560e-01 7.40290999e-01
-7.93977439e-01 -3.14174414e-01 4.47820455e-01 -2.28010774e-01
3.42346013e-01 -3.91135961e-01 1.71357885e-01 -9.07584071e-01
4.77548838e-02 -6.36789680e-01 4.16924089e-01 -1.19946682e+00
-8.29930305e-01 5.67054510e-01 1.05981745e-01 -1.36153603e+00
-4.52454053e-02 -5.97064555e-01 -5.41254401e-01 6.66834533e-01
-1.64084828e+00 -1.26886284e+00 -9.78642225e-01 9.67725098e-01
7.04412282e-01 -2.92208254e-01 6.78910375e-01 -1.62170723e-01
-4.13529366e-01 1.98837114e-03 -2.89707303e-01 -5.34550473e-02
2.77612209e-01 -9.31089222e-01 -2.06031427e-01 7.08298028e-01
2.17926234e-01 1.04773915e+00 4.05250341e-01 -8.62537324e-01
-1.52622330e+00 -8.04678023e-01 2.79702932e-01 -4.19803113e-01
2.79503942e-01 1.32288992e-01 -7.14111447e-01 7.95483470e-01
5.44531345e-02 -1.92950979e-01 5.31839192e-01 -9.18881372e-02
-2.64876366e-01 -7.38936067e-02 -8.62510860e-01 1.12867951e+00
1.46217167e+00 -3.41004342e-01 -8.17657471e-01 2.17179760e-01
7.34581470e-01 -4.37289059e-01 -3.50368649e-01 2.66111583e-01
5.73458612e-01 -1.12785912e+00 1.11346102e+00 -5.19748151e-01
1.57209203e-01 -8.52811635e-01 -3.08793664e-01 -1.33254349e+00
-4.94199097e-01 -3.99579965e-02 -2.79346108e-01 1.09490311e+00
2.15727404e-01 -4.65323985e-01 4.43688333e-01 5.50154388e-01
-3.03525150e-01 -6.49915457e-01 -8.24034810e-01 -5.58545589e-01
-7.97674239e-01 -2.58287609e-01 8.58781815e-01 6.76529706e-01
1.62677512e-01 2.50374615e-01 -2.02895612e-01 3.94946069e-01
2.57409960e-01 4.63847876e-01 8.16518009e-01 -1.20565617e+00
2.10225865e-01 -4.85547721e-01 -3.97507519e-01 -1.52364171e+00
1.15962453e-01 -7.61497080e-01 2.34967425e-01 -2.44696641e+00
1.66894555e-01 -3.37322682e-01 -4.63739187e-01 6.13322914e-01
-2.15747640e-01 -5.14719784e-01 1.40356809e-01 2.16749012e-01
-1.13799381e+00 8.34182143e-01 1.68477631e+00 -1.81202099e-01
-4.93744135e-01 -2.46790051e-01 -5.47267497e-01 9.37863171e-01
7.75395274e-01 1.03813335e-01 -8.17429960e-01 -4.70699757e-01
2.30224341e-01 2.12637469e-01 6.60687268e-01 -1.21935606e+00
6.40324950e-01 -5.83256841e-01 6.18754983e-01 -8.73664439e-01
6.78838730e-01 -1.25565422e+00 -9.56207514e-02 4.10097063e-01
-3.43919486e-01 -2.00465824e-02 1.07891403e-01 8.83576930e-01
-1.79029688e-01 -1.56700507e-01 9.82950628e-02 -4.00791496e-01
-1.83813429e+00 2.11967558e-01 -3.89146566e-01 -2.39286989e-01
1.24728703e+00 -3.93532544e-01 -6.18304968e-01 -4.80203062e-01
-5.98711789e-01 6.37738824e-01 2.64105052e-01 6.83296204e-01
1.25388610e+00 -1.56806493e+00 -1.32405967e-01 1.68656141e-01
5.10410130e-01 5.69143565e-03 5.04090428e-01 7.55166888e-01
-3.27579290e-01 5.17001748e-01 -4.64653224e-01 -3.25296968e-01
-1.10624993e+00 9.13034976e-01 2.22046643e-01 1.92779109e-01
-6.95788383e-01 7.60162830e-01 5.36462367e-01 -3.76794100e-01
3.26410174e-01 -3.19285184e-01 -1.11499465e+00 -8.51897076e-02
5.84988594e-01 3.01120311e-01 -3.58820230e-01 -7.58720517e-01
-5.37997842e-01 7.73610055e-01 9.45991352e-02 -4.76646125e-02
8.78810525e-01 -4.74922210e-01 -3.46934609e-02 4.83235478e-01
5.58269203e-01 2.32833456e-02 -1.25099683e+00 -4.20091003e-01
-1.74371183e-01 -5.11635363e-01 -5.91455400e-02 -1.07362354e+00
-1.01076543e+00 6.93304062e-01 5.67881584e-01 -1.48650646e-01
1.09165561e+00 7.39726052e-02 3.84553850e-01 7.73971617e-01
9.72046673e-01 -1.03253794e+00 4.84834969e-01 4.20753390e-01
1.11863899e+00 -1.25097167e+00 7.99534395e-02 -4.76691693e-01
-7.20735848e-01 1.13218522e+00 1.35204351e+00 5.23009419e-01
4.92381513e-01 -3.57194990e-01 -3.52104381e-02 -4.48286265e-01
-6.77630663e-01 -6.39888644e-01 4.73894626e-01 8.25854242e-01
-9.23958570e-02 -5.92783876e-02 -1.90044455e-02 6.03983819e-01
-9.44767371e-02 -1.74746126e-01 1.84514314e-01 1.15244305e+00
-9.52022791e-01 -3.48188430e-01 -3.08524072e-01 3.00988615e-01
5.95270932e-01 -4.79020402e-02 -2.01527700e-01 7.44960189e-01
3.52077752e-01 1.26528335e+00 -9.53754559e-02 -5.16871750e-01
5.40926039e-01 -6.21872954e-02 3.65676939e-01 -4.74963069e-01
-1.98048547e-01 -2.92833149e-01 -6.79463521e-02 -9.70416129e-01
-6.53706849e-01 -4.86821979e-01 -1.90268540e+00 -4.37160470e-02
-1.77348286e-01 -8.05082321e-02 3.86917770e-01 9.92414951e-01
3.26859981e-01 7.44187593e-01 2.38193244e-01 -6.68571711e-01
1.56399518e-01 -5.86739719e-01 -1.74643248e-01 3.33643228e-01
2.37957388e-01 -1.17689264e+00 -1.62894204e-01 -5.00431284e-02] | [4.496055603027344, 0.45615705847740173] |
1ede3539-6fdc-4572-aa53-04216483f3dc | confidence-driven-bounding-box-localization | 2303.01803 | null | https://arxiv.org/abs/2303.01803v1 | https://arxiv.org/pdf/2303.01803v1.pdf | Confidence-driven Bounding Box Localization for Small Object Detection | Despite advancements in generic object detection, there remains a performance gap in detecting small objects compared to normal-scale objects. We for the first time observe that existing bounding box regression methods tend to produce distorted gradients for small objects and result in less accurate localization. To address this issue, we present a novel Confidence-driven Bounding Box Localization (C-BBL) method to rectify the gradients. C-BBL quantizes continuous labels into grids and formulates two-hot ground truth labels. In prediction, the bounding box head generates a confidence distribution over the grids. Unlike the bounding box regression paradigms in conventional detectors, we introduce a classification-based localization objective through cross entropy between ground truth and predicted confidence distribution, generating confidence-driven gradients. Additionally, C-BBL describes a uncertainty loss based on distribution entropy in labels and predictions to further reduce the uncertainty in small object localization. The method is evaluated on multiple detectors using three object detection benchmarks and consistently improves baseline detectors, achieving state-of-the-art performance. We also demonstrate the generalizability of C-BBL to different label systems and effectiveness for high resolution detection, which validates its prospect as a general solution. | ['Xianbin Cao', 'Yanjing Li', 'Baochang Zhang', 'Huixin Sun'] | 2023-03-03 | null | null | null | null | ['small-object-detection'] | ['computer-vision'] | [ 1.34782314e-01 7.77818039e-02 -2.23425120e-01 -6.17111981e-01
-1.46001172e+00 -5.90829611e-01 4.73496884e-01 4.21638042e-01
-3.24956954e-01 4.08056498e-01 -1.13530912e-01 -4.96911779e-02
3.43583375e-01 -3.82791102e-01 -8.46939683e-01 -5.24557471e-01
-5.32333665e-02 4.79728997e-01 9.12058473e-01 2.85435230e-01
2.90929586e-01 5.74917674e-01 -1.23125303e+00 4.34599370e-01
5.46174228e-01 1.39374804e+00 1.59075335e-01 4.66107935e-01
2.17048198e-01 7.18607008e-01 -7.20087409e-01 -3.66925359e-01
4.24964875e-01 -7.76742920e-02 -2.61435151e-01 -1.70037821e-01
9.90416229e-01 -3.62555146e-01 1.52067952e-02 1.16307831e+00
3.33759636e-01 -5.13316169e-02 1.02979541e+00 -1.32391047e+00
-6.86564922e-01 4.18328583e-01 -9.32120025e-01 3.19400370e-01
1.06831543e-01 7.00461715e-02 1.17645073e+00 -1.29573286e+00
5.13374984e-01 1.49613833e+00 1.16203403e+00 5.05301595e-01
-1.64945722e+00 -7.63235450e-01 5.58431864e-01 -1.34498879e-01
-1.75052023e+00 4.01453711e-02 2.09446132e-01 -8.33257020e-01
9.36087489e-01 1.30601510e-01 3.26191097e-01 7.72585988e-01
4.94702041e-01 8.72715414e-01 1.19391561e+00 -4.30224121e-01
2.65666962e-01 5.20384133e-01 1.67006463e-01 7.40957618e-01
4.94998425e-01 2.56746858e-01 -4.59226012e-01 -1.30062178e-01
5.32688558e-01 -1.52335033e-01 6.62427247e-02 -6.76605999e-01
-1.15533423e+00 9.88176644e-01 9.87824440e-01 -2.89944351e-01
-5.57376258e-03 4.15080309e-01 2.94742167e-01 -5.02030551e-01
7.09563315e-01 4.46970582e-01 -1.71156675e-01 4.25126076e-01
-1.26584983e+00 4.90701944e-01 6.90639198e-01 1.11657560e+00
7.06739008e-01 -3.21282148e-01 -9.17833388e-01 5.97405076e-01
6.13204241e-01 5.76293349e-01 7.39666596e-02 -7.04741478e-01
2.82912314e-01 4.66024160e-01 2.94223666e-01 -9.61177289e-01
-5.46313882e-01 -5.73635101e-01 -2.92676598e-01 4.99566346e-01
4.24071223e-01 8.71966854e-02 -9.67551529e-01 1.69825816e+00
3.63245457e-01 9.87099186e-02 -4.76497084e-01 1.11686683e+00
5.43500185e-01 5.63818157e-01 2.48870209e-01 1.26899123e-01
1.37759423e+00 -1.17026269e+00 -3.54333878e-01 -4.60983366e-01
5.41501343e-01 -5.46453238e-01 8.85548353e-01 2.53572017e-01
-7.99322963e-01 -5.40877879e-01 -1.21252263e+00 -6.63002208e-02
-3.92283410e-01 4.22241807e-01 3.87222171e-01 6.39774442e-01
-9.57001507e-01 2.99652755e-01 -8.64483297e-01 -3.20037305e-01
7.30639935e-01 2.98930377e-01 -4.88244742e-03 2.23819047e-01
-6.02189302e-01 1.22234166e+00 5.58554471e-01 -8.66075158e-02
-8.53254795e-01 -8.49602282e-01 -7.96220601e-01 -1.71923246e-02
4.22827035e-01 -2.25619793e-01 1.37989426e+00 -4.34819907e-01
-8.47759187e-01 9.05137599e-01 -5.93509525e-02 -7.88136780e-01
6.85895085e-01 -2.73221761e-01 -1.15423806e-01 -3.42123955e-01
5.12272954e-01 1.31952918e+00 8.33482265e-01 -1.34583318e+00
-1.13872147e+00 -2.34221533e-01 -4.11670953e-01 9.63976607e-02
3.10262423e-02 6.07667118e-02 -5.45038521e-01 -5.74285030e-01
3.39293540e-01 -8.74915004e-01 -1.62824973e-01 2.91561902e-01
-5.42346060e-01 -3.71028125e-01 8.55013549e-01 -2.92369932e-01
1.27178907e+00 -2.09256768e+00 -3.36126745e-01 2.34684989e-01
4.40291971e-01 -3.55728045e-02 2.96565890e-02 -2.20679149e-01
3.04033160e-01 7.80464709e-02 -6.50624856e-02 -5.19226849e-01
2.50508457e-01 -1.69095799e-01 -6.52185023e-01 6.86785638e-01
8.10855865e-01 8.45296443e-01 -8.89741778e-01 -7.00610280e-01
1.92103073e-01 5.35285473e-01 -6.77694142e-01 -4.40901034e-02
-3.10970813e-01 1.92255422e-01 -3.58191490e-01 9.93594825e-01
7.87907720e-01 -4.17242646e-01 -2.92196214e-01 -3.44679058e-01
-1.68662503e-01 1.20854981e-01 -1.15650618e+00 1.10199416e+00
3.16655673e-02 8.70130956e-01 -1.25110447e-01 -5.87365270e-01
1.09599411e+00 -3.80519897e-01 2.50194639e-01 -3.29732478e-01
9.77907553e-02 3.05036843e-01 -2.10742623e-01 2.26265311e-01
6.47994697e-01 -1.28713856e-02 -2.85415143e-01 -9.52302886e-04
-7.25744516e-02 -4.50770855e-01 9.68918577e-02 1.84679374e-01
1.07421291e+00 2.40452632e-01 3.34238082e-01 -2.75843740e-01
2.01219246e-01 2.14771956e-01 6.28759265e-01 1.37574983e+00
-6.48616493e-01 8.59266758e-01 3.75805199e-01 -2.75605470e-01
-1.04973722e+00 -1.28165817e+00 -6.51296794e-01 1.35028362e+00
4.63601947e-01 -3.60189378e-01 -8.36752295e-01 -1.02804255e+00
5.57486475e-01 5.28236687e-01 -6.41336501e-01 -6.79954588e-02
-4.50752676e-01 -8.25772107e-01 4.52382356e-01 9.43136811e-01
3.24168861e-01 -7.28946984e-01 -6.65244222e-01 1.94372684e-01
1.61455199e-01 -1.18762910e+00 -6.60063148e-01 5.43638289e-01
-6.14105821e-01 -8.11008453e-01 -4.26627249e-01 -5.32619417e-01
6.18444622e-01 2.36423403e-01 1.25196028e+00 -2.71151692e-01
-8.01289141e-01 1.53450236e-01 -2.21878693e-01 -7.62501121e-01
-2.41698742e-01 -2.06926875e-02 1.29408851e-01 -3.08803529e-01
6.86998069e-01 3.26527804e-01 -6.93326890e-01 4.45698112e-01
-2.69986123e-01 -2.58120298e-01 7.34084368e-01 9.21845734e-01
1.02139902e+00 -3.06479394e-01 4.22891706e-01 -7.68656731e-01
4.57453132e-01 -4.02785420e-01 -1.17984211e+00 1.68177202e-01
-8.63741040e-01 7.66808018e-02 -6.10486930e-03 -5.57127237e-01
-7.87054062e-01 3.29774201e-01 2.12719932e-01 -3.05713683e-01
1.65314272e-01 -1.22236066e-01 3.07726562e-01 -3.39825243e-01
1.02216601e+00 -2.13531271e-01 -3.91564578e-01 -2.37765223e-01
5.34768820e-01 5.47549844e-01 6.60994530e-01 -4.23005611e-01
6.64959848e-01 6.62643254e-01 -3.94448526e-02 -2.47364342e-01
-1.34783292e+00 -8.25205386e-01 -5.90889931e-01 -2.61548579e-01
9.08645213e-01 -1.06315994e+00 -6.26321733e-01 -1.00443855e-01
-1.05514526e+00 -2.20064521e-01 -4.55102414e-01 4.60402280e-01
-5.79288006e-01 -4.71681803e-02 -5.36067665e-01 -1.07641387e+00
-1.42124221e-01 -1.24718678e+00 1.79169524e+00 3.42187226e-01
-2.37617299e-01 -5.53595841e-01 -1.52187226e-02 3.78963165e-02
1.71790853e-01 2.10780203e-01 4.87490922e-01 -6.66198492e-01
-8.61021876e-01 -5.45316279e-01 -8.80741298e-01 2.94611424e-01
-4.69628870e-01 -4.50839736e-02 -1.09876704e+00 -4.46407437e-01
-3.91256034e-01 -5.24055123e-01 1.28892040e+00 6.51784718e-01
1.07087517e+00 1.08739540e-01 -8.51065159e-01 5.08769274e-01
1.42433655e+00 -8.56225342e-02 1.26474664e-01 3.63773733e-01
5.15738845e-01 4.41360831e-01 1.04790545e+00 4.38168615e-01
9.07414258e-02 8.68478179e-01 4.09108818e-01 -2.20224723e-01
-3.84986460e-01 -5.01959741e-01 3.18690121e-01 -2.28107013e-02
5.15765488e-01 -1.04878083e-01 -8.91838789e-01 3.22388023e-01
-1.89267290e+00 -6.48291111e-01 -8.44832808e-02 2.07006073e+00
7.38759935e-01 4.36741620e-01 2.78035820e-01 -3.81032854e-01
8.20408821e-01 -1.31722808e-01 -8.92908692e-01 2.11903043e-02
-8.08148086e-02 -2.21014753e-01 9.76084948e-01 4.77742344e-01
-1.48661876e+00 1.17718697e+00 7.32907438e+00 9.09795403e-01
-9.99730825e-01 3.37863654e-01 7.23932743e-01 -2.81718820e-01
3.24775904e-01 -1.11855522e-01 -1.80635381e+00 3.87383699e-01
4.20069784e-01 2.32929274e-01 -1.11898728e-01 1.53210270e+00
-3.77474390e-02 -2.52799869e-01 -1.40427554e+00 8.15293074e-01
8.14379528e-02 -1.29826474e+00 -3.11594337e-01 -3.22381780e-02
8.89123380e-01 4.05933648e-01 3.33134621e-01 6.29161060e-01
4.59877312e-01 -9.02629852e-01 1.36328542e+00 1.92011729e-01
7.86601603e-01 -3.51469874e-01 6.74196422e-01 2.69878745e-01
-1.24798477e+00 -2.30964527e-01 -7.10421383e-01 2.09823802e-01
4.14828174e-02 5.76233089e-01 -1.37852550e+00 -2.96225011e-01
9.29920256e-01 3.69080871e-01 -8.43354404e-01 1.26165283e+00
-1.93816915e-01 4.70483631e-01 -5.03507793e-01 -3.34249914e-01
3.34155887e-01 1.33703858e-01 5.65229833e-01 1.87953651e+00
2.21191272e-02 -3.43849838e-01 5.65928340e-01 1.40623081e+00
-5.67126796e-02 -2.67421454e-03 -3.18570435e-01 4.00169075e-01
8.08482766e-01 1.43278730e+00 -1.16894507e+00 -2.90607184e-01
-3.32121640e-01 6.81921303e-01 6.17539108e-01 1.94051415e-01
-1.08100319e+00 -1.26572147e-01 3.60515893e-01 2.63318926e-01
4.28997219e-01 -3.73041555e-02 -5.80336630e-01 -8.19844365e-01
4.33601476e-02 -4.07062680e-01 2.63279676e-01 -7.18401194e-01
-1.56478691e+00 5.05567968e-01 1.67688981e-01 -9.86764312e-01
2.76120603e-02 -1.00914204e+00 -1.86612889e-01 7.15463698e-01
-1.42486823e+00 -1.20925164e+00 -4.47572887e-01 9.71869007e-02
5.57213247e-01 4.29814979e-02 4.69427019e-01 1.17610700e-01
-4.30385441e-01 6.96480989e-01 1.84343502e-01 4.28939387e-02
9.73029971e-01 -1.56831884e+00 5.50213277e-01 8.60364616e-01
2.44090363e-01 3.19886595e-01 7.26824701e-01 -8.45604956e-01
-8.17100823e-01 -1.48535168e+00 3.80658180e-01 -9.01839256e-01
6.41422391e-01 -9.07485783e-01 -8.17831635e-01 7.12985635e-01
-5.45619071e-01 5.81070185e-01 2.54148841e-01 4.78058644e-02
-8.00074935e-01 -2.53212694e-02 -1.26611197e+00 4.28712338e-01
8.28100204e-01 -3.98349196e-01 -3.38733464e-01 6.71336174e-01
6.37501121e-01 -5.71352482e-01 -4.81938601e-01 7.04718709e-01
5.69371045e-01 -1.00632358e+00 1.01451719e+00 -2.06581190e-01
-9.61395800e-02 -6.43471718e-01 -3.44210118e-01 -8.62518430e-01
-5.34160018e-01 -1.41155481e-01 -1.79698542e-01 9.50770319e-01
6.00954831e-01 -3.93566757e-01 9.87942874e-01 5.47794223e-01
-6.49552345e-02 -7.81114340e-01 -7.95852900e-01 -1.01378214e+00
5.01229651e-02 -5.34012675e-01 1.26668617e-01 3.26769590e-01
-1.59054190e-01 1.51246294e-01 1.00614410e-02 3.39435667e-01
8.28258395e-01 -9.63390321e-02 5.88555098e-01 -1.25653255e+00
-3.29346538e-01 -6.86880946e-01 -6.77586615e-01 -1.38782477e+00
4.20877300e-02 -7.89804459e-01 8.31291020e-01 -1.25896120e+00
5.11259139e-01 -8.53391230e-01 -2.79668450e-01 3.71982694e-01
-3.15824181e-01 8.58635247e-01 2.32788309e-01 4.21762943e-01
-1.07385075e+00 2.34884888e-01 7.34157801e-01 -1.58707201e-01
-9.02886838e-02 -7.82171413e-02 -5.33922613e-01 8.22557092e-01
3.39518905e-01 -6.82990909e-01 -1.14637949e-02 3.44642363e-02
5.10257818e-02 -6.01321399e-01 4.60494667e-01 -1.10677755e+00
3.15403908e-01 1.46293387e-01 6.86655700e-01 -9.89778042e-01
3.42894882e-01 -6.82780504e-01 -4.67700601e-01 3.93719077e-01
-3.90893102e-01 -3.08293700e-01 2.11491928e-01 8.61405492e-01
-3.52012664e-02 -1.94824427e-01 1.13457668e+00 2.46870160e-01
-8.31038177e-01 7.20527917e-02 -7.25541860e-02 1.50385007e-01
1.21392047e+00 -3.77891749e-01 -3.88815790e-01 2.18707044e-02
-4.95088279e-01 3.09354633e-01 3.12098086e-01 3.29260141e-01
4.51889694e-01 -1.32300341e+00 -6.92307651e-01 2.40296587e-01
4.14133519e-01 1.16116717e-01 -2.98163772e-01 8.27117682e-01
-5.11525154e-01 5.96743286e-01 3.51142526e-01 -1.16043305e+00
-1.13512695e+00 5.34406364e-01 4.98358488e-01 -1.81781471e-01
-6.14054143e-01 1.43801308e+00 6.58284664e-01 -2.14169204e-01
6.70531332e-01 -6.80343330e-01 2.35476896e-01 -1.23780422e-01
5.59843242e-01 3.31881285e-01 7.11481273e-02 -4.95348245e-01
-5.80294490e-01 5.50582230e-01 -4.66494888e-01 -2.37930529e-02
7.77898073e-01 -1.39568616e-02 3.54566485e-01 4.71043766e-01
8.74560416e-01 -1.90640867e-01 -1.97104526e+00 -1.37972042e-01
3.78588110e-01 -5.61421692e-01 2.19680414e-01 -9.36059296e-01
-4.40328687e-01 6.86545908e-01 9.59111989e-01 1.55719459e-01
6.80677593e-01 5.05084515e-01 2.33763441e-01 3.33445042e-01
3.12036365e-01 -1.18119669e+00 2.61841983e-01 4.11670744e-01
9.01281774e-01 -1.72099268e+00 2.65128583e-01 -6.38439655e-01
-6.66344285e-01 8.18327308e-01 9.28436279e-01 -2.58891910e-01
5.40388167e-01 6.98168933e-01 -2.49455753e-03 -7.95771405e-02
-4.75242049e-01 -2.89486229e-01 5.86277306e-01 7.22604752e-01
4.29256171e-01 1.48350134e-01 3.74810770e-02 5.37282526e-01
-1.51179861e-02 -3.60522091e-01 7.69576663e-03 7.64521360e-01
-8.52775514e-01 -5.33062220e-01 -7.73041368e-01 5.14576435e-01
-2.74285316e-01 -6.89762309e-02 -4.48445082e-01 1.01001620e+00
3.68360072e-01 8.59981060e-01 3.22671682e-01 -1.81832463e-01
3.31174016e-01 -1.25762403e-01 4.74616528e-01 -9.57356870e-01
-4.72985625e-01 2.61029065e-01 -3.47379386e-01 -7.32037067e-01
5.21097332e-02 -7.04397440e-01 -1.36903048e+00 1.79808751e-01
-1.07021701e+00 -1.61224872e-01 9.00273800e-01 5.49726844e-01
2.27742627e-01 4.72625196e-01 1.53913543e-01 -1.11278450e+00
-1.05659997e+00 -1.04743516e+00 -5.99637568e-01 2.96394408e-01
2.82475501e-01 -9.98599768e-01 -3.89753640e-01 -1.08446248e-01] | [9.18389892578125, 1.090420126914978] |
c9d4635c-a3ac-4cc3-bada-8b6d3d5ef2f5 | digging-deeper-into-egocentric-gaze | 1904.06090 | null | http://arxiv.org/abs/1904.06090v1 | http://arxiv.org/pdf/1904.06090v1.pdf | Digging Deeper into Egocentric Gaze Prediction | This paper digs deeper into factors that influence egocentric gaze. Instead
of training deep models for this purpose in a blind manner, we propose to
inspect factors that contribute to gaze guidance during daily tasks. Bottom-up
saliency and optical flow are assessed versus strong spatial prior baselines.
Task-specific cues such as vanishing point, manipulation point, and hand
regions are analyzed as representatives of top-down information. We also look
into the contribution of these factors by investigating a simple recurrent
neural model for ego-centric gaze prediction. First, deep features are
extracted for all input video frames. Then, a gated recurrent unit is employed
to integrate information over time and to predict the next fixation. We also
propose an integrated model that combines the recurrent model with several
top-down and bottom-up cues. Extensive experiments over multiple datasets
reveal that (1) spatial biases are strong in egocentric videos, (2) bottom-up
saliency models perform poorly in predicting gaze and underperform spatial
biases, (3) deep features perform better compared to traditional features, (4)
as opposed to hand regions, the manipulation point is a strong influential cue
for gaze prediction, (5) combining the proposed recurrent model with bottom-up
cues, vanishing points and, in particular, manipulation point results in the
best gaze prediction accuracy over egocentric videos, (6) the knowledge
transfer works best for cases where the tasks or sequences are similar, and (7)
task and activity recognition can benefit from gaze prediction. Our findings
suggest that (1) there should be more emphasis on hand-object interaction and
(2) the egocentric vision community should consider larger datasets including
diverse stimuli and more subjects. | ['Hamed R. -Tavakoli', 'Esa Rahtu', 'Ali Borji', 'Juho Kannala'] | 2019-04-12 | null | null | null | null | ['eye-tracking'] | ['computer-vision'] | [ 1.15677357e-01 -2.29353651e-01 -3.06185901e-01 -2.15769187e-01
-2.71267798e-02 -2.00944200e-01 4.28243667e-01 -2.58219630e-01
-3.26606095e-01 4.54855084e-01 7.42912054e-01 -6.06547184e-02
-1.84868306e-01 -2.54179835e-01 -7.00960338e-01 -8.21906686e-01
6.86611831e-02 -5.06765425e-01 3.73795539e-01 -2.80407965e-01
9.64645267e-01 3.56716901e-01 -2.24418116e+00 6.12219051e-02
9.63399470e-01 1.04528940e+00 5.17097175e-01 5.85349560e-01
4.05627370e-01 9.09775198e-01 -1.65563166e-01 -4.15745080e-02
1.95256621e-01 -6.51293457e-01 -6.79387033e-01 -2.54208688e-02
9.17238474e-01 -5.24983466e-01 -1.79173589e-01 1.04134166e+00
6.84783220e-01 2.62243658e-01 6.63899660e-01 -1.10045755e+00
-6.54360056e-01 4.32405099e-02 -7.92840064e-01 7.13050067e-01
5.10725319e-01 6.74452841e-01 9.88679886e-01 -8.72921467e-01
6.56520784e-01 1.06325233e+00 2.54917055e-01 4.74294692e-01
-9.70479429e-01 -5.71738482e-01 6.85264111e-01 5.28967500e-01
-9.59447920e-01 -6.24932170e-01 8.29070389e-01 -6.87607586e-01
8.35531652e-01 7.70466551e-02 7.32290030e-01 1.32019937e+00
5.46627045e-01 8.34001958e-01 1.13747323e+00 -3.12618345e-01
-1.33455887e-01 -9.55044758e-03 2.01292887e-01 5.70820570e-01
2.75154710e-01 3.02753747e-01 -9.05610919e-01 2.98063964e-01
8.42758596e-01 2.94904143e-01 -7.81917393e-01 -4.64610755e-01
-1.64245892e+00 5.96654892e-01 8.75890374e-01 3.47849756e-01
-6.09268069e-01 1.06314823e-01 1.83886975e-01 -5.78817073e-03
2.44020328e-01 7.85932839e-01 -2.14889795e-01 -2.03902289e-01
-8.97406042e-01 1.76712692e-01 2.13980749e-01 7.61220872e-01
1.01746607e+00 1.01310641e-01 -6.35914803e-01 5.78296423e-01
4.21090633e-01 4.86365616e-01 5.67871988e-01 -1.02631068e+00
5.31138718e-01 4.89999473e-01 2.97045290e-01 -1.06750906e+00
-6.71478927e-01 -4.74144101e-01 -2.33926833e-01 3.46992463e-01
5.79512835e-01 -1.15970150e-01 -7.32394278e-01 1.95132613e+00
-4.15758006e-02 -1.29133001e-01 -5.33937395e-01 1.52144706e+00
6.40828133e-01 8.98928195e-02 1.85806096e-01 -4.19688612e-01
1.67758012e+00 -9.32158053e-01 -8.91265750e-01 -4.44570750e-01
6.22303605e-01 -7.27682412e-01 1.29734015e+00 1.41863137e-01
-1.28329062e+00 -7.51227498e-01 -8.24983954e-01 -2.19011486e-01
-2.15108812e-01 1.59809023e-01 6.11576617e-01 2.79009998e-01
-1.22341549e+00 3.60664725e-01 -5.73016405e-01 -7.58059919e-01
3.29952389e-01 3.16054851e-01 -7.93759823e-02 4.72471379e-02
-1.01275730e+00 1.04844594e+00 -2.50951480e-02 2.45680809e-01
-9.29810226e-01 -5.51472664e-01 -8.53791475e-01 1.63723052e-01
2.42892891e-01 -1.17640555e+00 1.05142343e+00 -1.24077499e+00
-1.38293290e+00 6.36013746e-01 -1.19370806e+00 -9.98080615e-03
1.62053928e-01 -4.44820017e-01 -1.20575339e-01 3.87824774e-01
1.41131371e-01 9.33859348e-01 1.14575410e+00 -1.20758998e+00
-7.58089423e-01 -5.79949200e-01 2.67658234e-01 5.58512509e-01
-2.53765613e-01 1.05590299e-01 -1.98265329e-01 -4.84105587e-01
9.87189338e-02 -8.46453369e-01 3.18571687e-01 -1.49849489e-01
-2.77511448e-01 -3.79116148e-01 6.67647541e-01 -4.59197670e-01
1.43354642e+00 -1.98703587e+00 2.40971103e-01 -2.19569176e-01
5.33717036e-01 2.02317685e-02 9.30453986e-02 1.67158648e-01
-3.32016110e-01 -8.24445393e-03 2.99008936e-01 -1.90999329e-01
-2.78202504e-01 -4.69428718e-01 -2.19977766e-01 5.65494001e-01
2.11791590e-01 1.06438339e+00 -1.00462222e+00 -2.10600272e-01
2.20575377e-01 3.21457505e-01 -8.17869306e-01 -2.98076365e-02
1.08988345e-01 6.06269538e-01 -3.21268618e-01 7.08619535e-01
4.34041560e-01 -5.22468567e-01 -3.91913712e-01 -2.59730548e-01
-4.49759692e-01 4.34304446e-01 -6.37217522e-01 1.61059535e+00
-2.18396157e-01 1.16548026e+00 -3.01329523e-01 -4.00286257e-01
4.78912443e-01 1.33011073e-01 9.18082148e-02 -9.09827173e-01
4.73904669e-01 -4.56101485e-02 3.77212912e-01 -9.02246475e-01
6.11516654e-01 2.48595655e-01 4.43030059e-01 5.25739908e-01
7.20455870e-02 4.55434740e-01 1.13852568e-01 3.03960238e-02
6.49424493e-01 5.39335549e-01 3.07099879e-01 -2.63833463e-01
4.86876369e-01 -3.04879606e-01 4.41989124e-01 5.45800209e-01
-7.53888547e-01 7.98806310e-01 7.62453139e-01 -1.95669293e-01
-5.81975520e-01 -5.24783373e-01 2.36235876e-02 1.56779850e+00
6.85495198e-01 -2.90268302e-01 -6.98227525e-01 -5.09039044e-01
-4.16110933e-01 7.24454701e-01 -9.85245347e-01 -2.70213783e-01
-3.97123665e-01 -2.66772687e-01 -1.41073033e-01 5.34839690e-01
5.52266300e-01 -1.15552151e+00 -1.00562668e+00 -3.56300861e-01
-4.26486105e-01 -7.42634654e-01 -8.06761086e-01 -2.12716579e-01
-9.55698252e-01 -1.28624082e+00 -9.53477800e-01 -6.36310101e-01
8.33325744e-01 1.11172426e+00 7.36614585e-01 -1.06611028e-01
1.30107105e-01 5.46847761e-01 -2.40634292e-01 -4.30014670e-01
5.37896037e-01 -1.31078940e-02 1.49530888e-01 1.60279527e-01
6.99017406e-01 -3.73671144e-01 -1.23076832e+00 5.73719621e-01
-2.09309340e-01 3.70790027e-02 7.24442482e-01 7.35553682e-01
8.31752643e-02 -6.14696085e-01 3.14644873e-01 -5.50396264e-01
5.32109201e-01 -3.90572071e-01 -3.35141391e-01 1.21118970e-01
-6.23763144e-01 -1.35445893e-01 -8.75839964e-03 -2.27890983e-01
-1.34493113e+00 -2.67583907e-01 4.46405381e-01 -6.84007764e-01
-3.12350184e-01 2.01651379e-01 6.39433935e-02 -1.19257502e-01
1.07418561e+00 1.55724183e-01 1.45625055e-01 -7.37869292e-02
-2.64331009e-02 4.16536868e-01 -1.27679016e-03 -1.01501375e-01
3.38057727e-01 6.18559062e-01 -7.81412721e-02 -7.17070162e-01
-9.26127732e-01 -4.47120368e-01 -8.99793684e-01 -5.54114819e-01
1.07883894e+00 -9.58694756e-01 -1.21293378e+00 4.13489193e-01
-1.05532646e+00 -1.28775939e-01 1.18453309e-01 9.11753058e-01
-6.84934318e-01 3.01401824e-01 -1.76739991e-01 -9.33068633e-01
-9.24733207e-02 -1.46434021e+00 1.10555553e+00 6.29275084e-01
-3.66479337e-01 -8.15533638e-01 -2.43862391e-01 3.68056744e-01
3.21689278e-01 -1.47167489e-01 5.41263580e-01 -2.92949915e-01
-9.41616058e-01 4.63417247e-02 -5.34540355e-01 -9.77306664e-02
2.30751440e-01 7.19244555e-02 -1.30663097e+00 -1.43881202e-01
1.09721467e-01 -1.04196951e-01 8.89387429e-01 9.34954166e-01
9.58333611e-01 -6.51987121e-02 -4.53213394e-01 6.99519336e-01
1.00405562e+00 1.17965288e-01 7.39614487e-01 4.16177273e-01
7.78110206e-01 1.08055556e+00 7.97754228e-01 1.34431794e-01
6.84526980e-01 6.15494549e-01 6.92142904e-01 -6.51693763e-03
-1.68302566e-01 -2.12156326e-01 6.04726434e-01 3.02410603e-01
-6.08313203e-01 -1.15570817e-02 -8.14114153e-01 5.94078839e-01
-1.79792857e+00 -1.31153703e+00 -2.84720391e-01 2.36479330e+00
4.06642109e-01 2.97969747e-02 2.60230035e-01 -1.24063618e-01
7.95165300e-01 2.87399799e-01 -7.76623487e-01 -1.98680207e-01
-2.20285617e-02 -4.14400458e-01 2.88040578e-01 2.03022078e-01
-9.18577373e-01 9.39017653e-01 6.16955376e+00 7.00319558e-02
-1.65068471e+00 7.43538663e-02 1.87361285e-01 -5.54384172e-01
-1.58139855e-01 6.91261515e-02 -9.56980646e-01 6.02995276e-01
5.26592791e-01 1.08760417e-01 3.90587956e-01 7.41611719e-01
6.14031196e-01 -5.05177081e-01 -1.09365213e+00 1.00500309e+00
4.87865061e-01 -9.13931549e-01 -1.64373845e-01 4.10736769e-01
5.57545602e-01 1.87515318e-01 3.76466453e-01 7.98642039e-02
-2.49681100e-01 -8.95667911e-01 7.32456923e-01 8.63863051e-01
4.94241059e-01 -3.06620508e-01 5.33508182e-01 2.82552123e-01
-8.26149881e-01 -4.25966889e-01 -2.69003212e-01 -2.03603655e-01
8.85124356e-02 -6.77549094e-03 -4.70904976e-01 3.07145238e-01
9.42975163e-01 1.17978716e+00 -8.18039536e-01 1.30469596e+00
-5.40954471e-01 4.02398527e-01 1.58427313e-01 -1.95104163e-02
2.74210628e-02 -2.07958594e-02 7.24627733e-01 8.55362773e-01
1.66161284e-01 -7.59846270e-02 -6.68059468e-01 9.15199101e-01
3.41457248e-01 -1.39123321e-01 -6.50674760e-01 2.85321444e-01
2.90935248e-01 9.99154091e-01 -3.99365753e-01 6.28513470e-02
-5.87872088e-01 7.43995488e-01 2.40124717e-01 1.11433864e+00
-6.76620126e-01 -4.90587831e-01 8.48654211e-01 2.65695900e-01
4.54382300e-01 -8.84711519e-02 -4.31958109e-01 -1.33274758e+00
-7.14587793e-02 -3.93836856e-01 1.57239050e-01 -1.45879841e+00
-8.46347690e-01 4.71486181e-01 -2.91752070e-01 -1.65048313e+00
-4.53385472e-01 -6.19046986e-01 -6.37382090e-01 1.26230311e+00
-1.96322548e+00 -8.73766959e-01 -8.24547589e-01 8.39717507e-01
5.94301045e-01 -1.05217479e-01 4.57160205e-01 -9.01371613e-02
-7.04122305e-01 4.51320171e-01 -4.12325263e-01 -1.01286665e-01
1.15418780e+00 -9.27078187e-01 -1.78039953e-01 1.13684237e+00
-1.26052260e-01 1.27436090e+00 6.47139251e-01 -4.80780214e-01
-1.21131384e+00 -5.57632744e-01 9.38584507e-01 -8.01158667e-01
4.05638963e-01 -1.91290066e-01 -8.03386211e-01 7.62521744e-01
3.84573787e-01 -1.09146684e-01 5.48083663e-01 4.53729749e-01
-1.55466661e-01 -7.34083727e-02 -5.80111980e-01 7.82271326e-01
1.21867085e+00 -6.61263227e-01 -7.97561765e-01 4.83463034e-02
3.78712922e-01 -2.75614530e-01 -3.77972335e-01 1.86591342e-01
7.61265397e-01 -1.39003193e+00 6.29146218e-01 -4.77347523e-01
5.99758148e-01 -3.55113149e-01 3.01752746e-01 -1.23060703e+00
-7.86779046e-01 -5.95029116e-01 -3.22720259e-01 8.17219436e-01
1.61077350e-01 -7.35293269e-01 3.55617523e-01 5.16803265e-01
-8.33424777e-02 -6.39143467e-01 -5.12288511e-01 -3.20366710e-01
-4.20614809e-01 -2.28944823e-01 2.98799187e-01 6.97813570e-01
1.02777019e-01 4.53592360e-01 -4.09560829e-01 -3.13732959e-02
2.80240148e-01 1.12020783e-01 8.74380231e-01 -1.28856993e+00
3.30300838e-01 -7.31308639e-01 -4.04271066e-01 -1.37508619e+00
5.77422231e-02 -3.68131250e-01 -9.27238688e-02 -1.47180927e+00
9.48632881e-02 1.20268427e-01 -4.63878632e-01 2.01537028e-01
-5.86430669e-01 1.27117813e-01 2.46977225e-01 7.13624477e-01
-4.43100750e-01 5.45954287e-01 1.81730819e+00 3.36977035e-01
-4.10070032e-01 1.49803832e-01 -1.18816638e+00 6.63243890e-01
3.34822804e-01 3.85954827e-02 -5.23935497e-01 -5.07992387e-01
4.16429818e-01 4.15287400e-03 5.33650339e-01 -7.84241736e-01
5.15888155e-01 -2.67542094e-01 4.90912884e-01 -4.87473309e-01
3.96550745e-01 -6.06968343e-01 -6.91325545e-01 2.51308978e-01
-9.26757231e-02 -2.39226982e-01 7.95363709e-02 6.91336095e-01
-2.35882729e-01 -8.63096714e-02 7.95058787e-01 1.84528064e-02
-8.49050641e-01 1.66325524e-01 -2.71147490e-01 -9.80073959e-02
1.01399231e+00 -8.83011699e-01 -7.19945788e-01 -5.74313045e-01
-6.76893234e-01 2.33841896e-01 7.49512792e-01 7.08761334e-01
3.10013235e-01 -9.32386518e-01 -3.23634565e-01 3.59983057e-01
3.67980242e-01 -3.15058261e-01 3.98981512e-01 1.65165484e+00
-1.47340387e-01 8.46858919e-01 -4.04860944e-01 -1.07591808e+00
-9.55544293e-01 6.67397141e-01 3.72393966e-01 6.04649425e-01
-3.02474290e-01 1.04285467e+00 1.03959823e+00 4.27995533e-01
1.57556161e-01 -3.79804760e-01 -8.32917690e-01 6.35707140e-01
6.98818147e-01 4.01068091e-01 -1.25726312e-01 -9.14810419e-01
-3.17603409e-01 9.58625436e-01 -7.80800059e-02 3.05057675e-01
1.11359525e+00 -6.45315945e-01 5.48200309e-02 6.13356829e-01
7.47622848e-01 1.58175323e-02 -1.62857938e+00 -1.24333844e-01
-3.40954572e-01 -8.50234210e-01 1.37660429e-01 -6.35527670e-01
-9.56978738e-01 1.33813834e+00 6.28526092e-01 -4.62417193e-02
1.28081715e+00 -1.22118153e-01 2.33981714e-01 2.98943510e-03
2.37010196e-01 -8.24436784e-01 2.02331364e-01 5.47764838e-01
9.52673018e-01 -1.49273896e+00 2.32743588e-03 -8.43756944e-02
-1.08254230e+00 9.19793844e-01 9.54001248e-01 -7.73722678e-02
5.48624635e-01 -7.79871166e-01 -1.35893285e-01 -4.08895254e-01
-8.68603647e-01 -7.05566227e-01 7.36458957e-01 4.81996804e-01
5.55324912e-01 -4.46143717e-01 -1.69690669e-01 5.19888163e-01
-1.92658007e-01 1.04535803e-01 3.39837313e-01 6.83439016e-01
-4.34737623e-01 -2.60108888e-01 -3.71732742e-01 5.45375824e-01
-2.87223876e-01 -1.36497915e-01 -1.95028424e-01 7.58719087e-01
4.51604426e-02 8.39915454e-01 2.03479707e-01 -3.58860105e-01
3.06855708e-01 1.46893084e-01 4.42858726e-01 -6.72299683e-01
-5.09615481e-01 2.20238343e-01 -2.55125135e-01 -9.30011392e-01
-6.79539382e-01 -9.27513242e-01 -6.94686472e-01 -1.41112268e-01
-4.42694962e-01 -3.22350174e-01 5.54514170e-01 9.15958285e-01
8.60739708e-01 6.51436746e-01 4.66850489e-01 -1.31352890e+00
9.38704517e-03 -1.22034991e+00 -5.91479123e-01 3.38389397e-01
6.72643542e-01 -1.31529748e+00 -7.58595884e-01 1.31060347e-01] | [13.983227729797363, 0.04935143142938614] |
d8b480f7-f86a-4e23-ab26-77e3727e7d23 | neural-distribution-learning-for-generalized | null | null | https://openreview.net/forum?id=SyG4RiR5Ym | https://openreview.net/pdf?id=SyG4RiR5Ym | Neural Distribution Learning for generalized time-to-event prediction | Predicting the time to the next event is an important task in various domains.
However, due to censoring and irregularly sampled sequences, time-to-event prediction has resulted in limited success only for particular tasks, architectures and data. Using recent advances in probabilistic programming and density networks, we make the case for a generalized parametric survival approach, sequentially predicting a distribution over the time to the next event.
Unlike previous work, the proposed method can use asynchronously sampled features for censored, discrete, and multivariate data.
Furthermore, it achieves good performance and near perfect calibration for probabilistic predictions without using rigid network-architectures, multitask approaches, complex learning schemes or non-trivial adaptations of cox-models.
We firmly establish that this can be achieved in the standard neural network framework by simply switching out the output layer and loss function. | ['Jung-Woo Ha', 'Jaegul Choo', 'Jaesung Huh', 'Adrian Kim', 'Egil Martinsson'] | 2018-09-27 | null | null | null | null | ['time-to-event-prediction'] | ['time-series'] | [ 1.73386574e-01 1.69591215e-02 -4.59311754e-01 -7.47982681e-01
-9.59849238e-01 -2.79901356e-01 6.45146370e-01 4.57228988e-01
-5.56147695e-01 1.10090292e+00 1.19580003e-02 -5.51081240e-01
-6.00088775e-01 -7.67108798e-01 -5.86132765e-01 -6.30171239e-01
-6.33069336e-01 1.05255115e+00 1.01431958e-01 4.46327209e-01
-7.09904954e-02 5.97967267e-01 -1.03823996e+00 -1.31147131e-01
4.51136172e-01 8.95996094e-01 -3.35503787e-01 6.89957738e-01
-5.54725118e-02 6.44812286e-01 -2.98739284e-01 -4.90742445e-01
-2.20231324e-01 -1.46941006e-01 -4.62514520e-01 -4.96407688e-01
4.88273539e-02 -3.74343008e-01 -2.92066842e-01 4.77962673e-01
4.71329540e-01 2.97295786e-02 1.18272567e+00 -1.37350261e+00
-4.53748703e-01 6.09107494e-01 -3.22020978e-01 9.27783623e-02
-4.08350155e-02 -3.55155975e-01 6.21129811e-01 -6.24554336e-01
2.38185406e-01 1.09836936e+00 1.17496109e+00 4.71286237e-01
-1.70461690e+00 -4.93683398e-01 -1.52271956e-01 -5.12150116e-02
-1.37015188e+00 -3.32079083e-01 6.31652951e-01 -6.20619893e-01
6.95756555e-01 -2.03625113e-02 2.61170566e-01 1.68535447e+00
6.70026958e-01 4.73965019e-01 9.10757363e-01 -2.22115338e-01
5.71264982e-01 -7.06648827e-02 1.52860701e-01 1.80459157e-01
1.44068629e-01 5.08393764e-01 -1.76906303e-01 -6.81970716e-01
8.08697820e-01 6.09306157e-01 1.50449976e-01 -5.30590355e-01
-1.04219854e+00 9.93975699e-01 -2.15754062e-02 1.11152023e-01
-5.51188171e-01 3.36175561e-01 4.53278393e-01 3.92672300e-01
5.88615417e-01 -1.68319002e-01 -7.32736886e-01 -2.02237099e-01
-1.38482618e+00 2.46082261e-01 1.15110254e+00 6.48980975e-01
3.53854328e-01 -4.71703261e-02 -4.32356983e-01 5.91738403e-01
1.51114494e-01 1.25827402e-01 2.76442111e-01 -9.19109344e-01
2.07559988e-01 1.24507926e-01 3.55763048e-01 -5.85801780e-01
-8.50758314e-01 -4.13020581e-01 -1.21267700e+00 1.92397818e-01
9.31653321e-01 -5.13359904e-01 -8.38421226e-01 1.88562179e+00
2.12186992e-01 2.69234926e-01 -2.13817552e-01 3.13329577e-01
8.35111737e-02 4.48265254e-01 5.16489744e-01 -1.45630360e-01
1.16392851e+00 -2.09047541e-01 -3.72030556e-01 -1.71026349e-01
4.01414692e-01 -3.29277039e-01 6.14683509e-01 4.13170040e-01
-8.27832162e-01 -1.67110175e-01 -5.96458852e-01 -3.70651856e-02
-2.37345770e-01 1.03384852e-02 7.44362831e-01 7.05981553e-01
-9.13533509e-01 9.23870265e-01 -1.30568528e+00 -4.70475197e-01
4.41629052e-01 4.49919969e-01 -1.29544362e-01 -1.36393622e-01
-1.08866584e+00 7.23125160e-01 5.07845044e-01 -8.84788632e-02
-1.03044808e+00 -8.84708762e-01 -5.79606295e-01 3.96176159e-01
4.25460897e-02 -8.36578548e-01 1.05021644e+00 -8.19582999e-01
-1.54181230e+00 5.14288664e-01 -9.08774734e-02 -8.00732017e-01
8.11066329e-01 8.16378742e-02 -2.50637859e-01 -2.89988279e-01
-1.25534713e-01 4.36173588e-01 7.39215016e-01 -5.27038693e-01
-4.89856541e-01 -4.20206159e-01 -3.86817575e-01 -2.85028696e-01
-2.32268244e-01 6.80331737e-02 -9.72849503e-02 -6.84427738e-01
-2.21570790e-01 -7.66152501e-01 -3.89403790e-01 1.39370337e-01
-1.55026793e-01 -5.22662640e-01 3.48558873e-01 -6.68684661e-01
9.98503923e-01 -2.19465518e+00 -6.52426779e-02 1.01272650e-01
-6.01044158e-03 -3.15416992e-01 1.28989130e-01 6.26400054e-01
-1.30018413e-01 -7.25295767e-02 -2.83929795e-01 -7.25896537e-01
6.52785972e-02 3.06031197e-01 -8.16950575e-02 5.86951315e-01
4.63085473e-02 6.83672428e-01 -6.48500144e-01 -5.58723927e-01
5.38015440e-02 7.72305250e-01 -3.36016893e-01 1.35914713e-01
-2.40565743e-02 4.64241266e-01 -3.60899210e-01 4.72781688e-01
3.59623581e-01 -4.87825304e-01 2.37679482e-01 4.80378568e-01
1.19889796e-01 -7.59890229e-02 -9.69764829e-01 1.58277106e+00
-4.68799859e-01 4.73854750e-01 -1.36378691e-01 -1.36795831e+00
9.22270834e-01 7.21690834e-01 5.57104588e-01 -2.01145396e-01
1.41926289e-01 2.03054339e-01 -3.72345388e-01 -1.52082756e-01
-2.89561246e-02 -6.73993945e-01 -2.56758332e-01 3.05445433e-01
2.27656662e-01 4.87253457e-01 -9.97779220e-02 -3.35783720e-01
1.25136399e+00 1.42494276e-01 2.86189497e-01 -1.10394605e-01
1.41816854e-01 -8.31530988e-02 5.82843423e-01 1.19471717e+00
-1.46382123e-01 7.65070796e-01 9.09052908e-01 -5.98807454e-01
-1.18752551e+00 -1.42362428e+00 -6.43856168e-01 1.34834695e+00
-6.28528476e-01 3.19948763e-01 -2.35874310e-01 -4.89911497e-01
1.98873594e-01 9.05030906e-01 -6.92462921e-01 -1.19160831e-01
-3.76868963e-01 -1.24330735e+00 6.97104692e-01 7.04401314e-01
-2.26304054e-01 -1.01847816e+00 -2.67719507e-01 7.02928841e-01
9.79031622e-02 -6.00720346e-01 -2.27113172e-01 6.63560510e-01
-1.22008789e+00 -8.78669202e-01 -1.11625636e+00 -7.20708728e-01
2.31821612e-01 -5.05148292e-01 1.21583390e+00 -4.71103430e-01
-6.27906471e-02 2.22983450e-01 2.16611132e-01 -4.07650799e-01
-4.16914344e-01 3.42437595e-01 -1.06204212e-01 4.73277047e-02
4.00131047e-01 -1.01921606e+00 -6.96596980e-01 1.93888932e-01
-8.76031995e-01 -3.19568306e-01 4.84460592e-01 1.04032648e+00
4.67178434e-01 -9.59997624e-02 1.14193237e+00 -8.53021502e-01
5.33092797e-01 -1.00545442e+00 -6.91542447e-01 2.20277861e-01
-8.07399094e-01 7.23814368e-02 8.07807565e-01 -6.24428809e-01
-9.31394100e-01 9.84797850e-02 -9.75644812e-02 -2.08160833e-01
-5.66438615e-01 7.36917853e-01 2.14316055e-01 4.22968268e-01
6.53024852e-01 2.93520927e-01 1.31413624e-01 -6.08008265e-01
5.02384044e-02 5.63319504e-01 5.11849582e-01 -5.30789316e-01
2.16698140e-01 6.62904561e-01 2.69513369e-01 -3.33615869e-01
-7.70674109e-01 -1.44289955e-01 -6.06691480e-01 2.68730909e-01
7.53910184e-01 -8.39609325e-01 -8.74182045e-01 4.32445437e-01
-1.03399026e+00 -4.51454878e-01 -3.06431502e-01 6.90070033e-01
-7.95381010e-01 2.16064334e-01 -7.89882183e-01 -9.07916307e-01
-3.65884721e-01 -5.44162273e-01 8.05077076e-01 5.24958186e-02
-1.75998285e-01 -1.48816717e+00 3.05949271e-01 -1.92397028e-01
6.51240110e-01 5.06057620e-01 1.25540888e+00 -1.06166327e+00
-9.43579748e-02 -7.85602033e-01 -1.10727265e-01 1.79024145e-01
1.55671081e-02 -1.55464053e-01 -6.81918144e-01 -3.26694727e-01
-1.02543160e-01 -3.41946870e-01 6.47169650e-01 9.52406049e-01
1.44713390e+00 -1.11784831e-01 -5.29244900e-01 6.65207982e-01
1.28201258e+00 2.13332564e-01 3.51807445e-01 1.79779768e-01
1.96696669e-01 6.32274330e-01 -1.65547859e-02 6.63938224e-01
6.42487228e-01 4.94840622e-01 3.17858368e-01 7.45764608e-03
2.36821994e-01 -1.77403882e-01 1.09897353e-01 -7.43848011e-02
1.69433966e-01 -2.68868804e-01 -1.03603137e+00 7.74855673e-01
-1.86994863e+00 -1.10844874e+00 -1.17270947e-01 2.61580229e+00
9.77666557e-01 3.27886075e-01 2.54145086e-01 -2.49402523e-01
6.42782390e-01 -9.01538506e-02 -5.96843898e-01 -2.59758592e-01
3.67416330e-02 2.11860254e-01 5.71553528e-01 2.49435470e-01
-1.18890202e+00 1.90563291e-01 7.42793512e+00 8.31584811e-01
-1.05047095e+00 3.07403177e-01 9.71305668e-01 -1.46315172e-01
1.20484442e-01 6.98099136e-02 -6.51775777e-01 5.42635620e-01
1.46372747e+00 -6.54041320e-02 2.53549248e-01 7.62320042e-01
2.65040576e-01 1.48831457e-01 -1.54611170e+00 5.48719823e-01
-4.02247369e-01 -1.08818030e+00 -4.60447758e-01 2.50323247e-02
4.20018524e-01 2.19034448e-01 9.41348635e-03 4.93590117e-01
6.69998467e-01 -1.25876641e+00 4.18370336e-01 9.28401053e-01
8.20889533e-01 -8.49908710e-01 8.50609720e-01 6.51301980e-01
-6.07733965e-01 -1.53465778e-01 -3.00699651e-01 -9.84432697e-02
3.60267609e-01 8.15071106e-01 -8.73325825e-01 2.96454877e-01
4.99442905e-01 3.14663082e-01 -2.88991421e-01 1.36732304e+00
1.83901519e-01 1.14212251e+00 -7.66844153e-01 1.34395808e-01
-6.51986301e-02 1.31217480e-01 1.27363071e-01 1.12440050e+00
6.00270033e-01 -5.03683805e-01 1.81079775e-01 6.21825516e-01
2.20464364e-01 -2.05803052e-01 -4.23544466e-01 3.86575699e-01
2.88899183e-01 1.08609486e+00 -6.22626245e-01 -1.97182208e-01
-5.24804771e-01 6.88189626e-01 5.02306879e-01 4.86256093e-01
-8.38301361e-01 -2.33466029e-01 3.29005904e-02 2.57775337e-01
4.25294369e-01 -3.10815871e-01 -3.60433161e-01 -9.66368020e-01
-2.28903294e-01 -2.38475353e-01 8.02667141e-01 -2.77008414e-01
-1.84318328e+00 3.25723976e-01 1.58880547e-01 -8.44887078e-01
-6.09836638e-01 -4.19344306e-01 -8.62694681e-01 9.55207348e-01
-1.39425957e+00 -1.01292562e+00 2.15221316e-01 7.15372264e-01
9.24076438e-02 1.25843912e-01 1.01085150e+00 3.77237082e-01
-5.06127119e-01 6.11388266e-01 6.88685060e-01 3.83089483e-02
8.09719741e-01 -1.48082519e+00 2.37756893e-01 2.74487138e-01
-1.55091032e-01 2.34864280e-01 6.80205286e-01 -5.36483943e-01
-1.00698435e+00 -1.07153225e+00 1.13230872e+00 -4.46312189e-01
9.82218266e-01 -5.05711973e-01 -9.79502618e-01 7.64114678e-01
-1.82222247e-01 4.58301231e-02 8.26050520e-01 5.97727299e-01
-1.05838165e-01 -2.96335399e-01 -1.09578741e+00 2.10163817e-01
5.32214403e-01 -3.66368562e-01 -4.87770021e-01 4.44261551e-01
2.56141901e-01 -1.42860770e-01 -1.10763609e+00 1.25580758e-01
7.10284352e-01 -7.41724551e-01 1.06674385e+00 -7.99068809e-01
3.31737936e-01 1.90764308e-01 1.06742539e-01 -9.83476758e-01
-4.34292704e-01 -5.41904569e-01 -2.27751300e-01 1.22017920e+00
5.18246472e-01 -6.78890526e-01 9.69459355e-01 9.17915225e-01
1.78391546e-01 -5.41343689e-01 -1.68640029e+00 -7.82546282e-01
4.99730468e-01 -5.77183425e-01 4.64578271e-01 6.80175841e-01
-1.99648559e-01 3.11376601e-02 -7.28161693e-01 2.35831127e-01
8.62667382e-01 7.23836049e-02 3.33017111e-01 -1.76901937e+00
-4.37307894e-01 -2.53913224e-01 -1.47821590e-01 -7.71683633e-01
2.08196342e-01 -7.67571211e-01 5.71161509e-04 -1.22038448e+00
2.15193242e-01 -6.67345822e-01 -5.87999225e-01 6.69117630e-01
-5.69125712e-02 -2.02833507e-02 -3.85429740e-01 1.97570190e-01
-3.45123023e-01 4.98715639e-01 4.82489347e-01 2.00167701e-01
-2.24906564e-01 6.86518967e-01 -3.67111564e-01 5.40875256e-01
8.26681018e-01 -9.20032322e-01 -2.10429937e-01 -2.40115479e-01
2.04363227e-01 9.28764701e-01 8.45293581e-01 -8.91908348e-01
3.63073289e-01 -1.99011892e-01 6.42464161e-01 -4.91569251e-01
2.40358919e-01 -9.98762786e-01 5.15726388e-01 3.72731835e-01
-3.32189411e-01 7.99841508e-02 2.02485528e-02 1.10259163e+00
-6.19237758e-02 -2.90254056e-01 5.06046832e-01 8.42670500e-02
-2.00643167e-02 3.96671623e-01 -5.27983069e-01 -1.38281554e-01
9.40153956e-01 -1.27693698e-01 1.45056278e-01 -3.46024066e-01
-1.32059765e+00 3.06998312e-01 1.62778944e-01 1.23183616e-01
1.77076980e-01 -1.19511592e+00 -9.00245965e-01 4.75910679e-02
-2.21836776e-01 -3.92973721e-02 3.83931428e-01 9.70740736e-01
-1.84153929e-01 3.13212454e-01 -1.03235440e-02 -6.18869066e-01
-6.37638628e-01 6.41254306e-01 4.44748700e-01 -6.64224923e-01
-5.38113296e-01 3.24812174e-01 2.32022908e-02 -4.22379851e-01
3.56354386e-01 -1.64164856e-01 -4.36583683e-02 5.63834161e-02
2.47897938e-01 4.39961523e-01 -4.22977880e-02 1.37738675e-01
-2.18748435e-01 -1.37448506e-02 -1.55111194e-01 -1.08494237e-01
1.47448039e+00 -1.01634093e-01 1.24510422e-01 7.99408853e-01
9.77185488e-01 -6.34500206e-01 -1.61306584e+00 -9.65452567e-02
2.81651855e-01 1.39936760e-01 -4.64958251e-02 -8.35202992e-01
-6.76838815e-01 9.42994595e-01 6.49604976e-01 4.12385404e-01
1.00992584e+00 7.44338855e-02 2.60183066e-01 1.69946536e-01
2.58762658e-01 -5.65324247e-01 -5.28978884e-01 4.51641172e-01
5.17483294e-01 -1.18630052e+00 -1.82535753e-01 1.83240622e-01
-2.18715504e-01 1.26224625e+00 3.39166028e-03 -2.76147664e-01
1.15135777e+00 2.88770974e-01 -1.96282089e-01 1.36102229e-01
-9.06432092e-01 4.70470518e-01 1.14105046e-01 7.76896358e-01
5.90829670e-01 1.04033291e-01 -3.24571937e-01 7.25383461e-01
9.16358009e-02 5.75126827e-01 4.04779494e-01 6.91040933e-01
-8.40188563e-02 -1.18732488e+00 -3.64006668e-01 7.86713302e-01
-9.74963605e-01 -1.72730282e-01 4.00730044e-01 7.85542428e-01
-3.15932453e-01 6.19633913e-01 3.37236673e-01 3.58202070e-01
9.04212669e-02 5.70772827e-01 5.36547154e-02 -3.21466357e-01
-3.81882250e-01 2.45698318e-02 -1.38667315e-01 -1.13792725e-01
-2.00756803e-01 -9.32881653e-01 -8.55831981e-01 -3.16509306e-01
-5.02063707e-02 -1.67208631e-03 5.03830612e-01 9.25392449e-01
5.40865064e-01 2.62102634e-01 4.88096505e-01 -8.30800116e-01
-9.85387266e-01 -1.02680492e+00 -8.14245939e-01 9.95733738e-02
3.96157861e-01 -5.10102153e-01 -3.72975200e-01 3.93379033e-02] | [7.753687381744385, 5.583449363708496] |
3cae8110-4ca7-47d6-ac67-e01ed4b44f2a | leveraging-a-semantically-annotated-corpus-to | null | null | https://aclanthology.org/W15-0101 | https://aclanthology.org/W15-0101.pdf | Leveraging a Semantically Annotated Corpus to Disambiguate Prepositional Phrase Attachment | null | ['Guy Emerson', 'Ann Copestake'] | 2015-04-01 | null | null | null | ws-2015-4 | ['prepositional-phrase-attachment'] | ['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] | [-7.437986373901367, 3.823320150375366] |
a5245720-42e3-445e-b7a0-6f6a48d1a473 | feature-representation-for-icu-mortality | 1512.05294 | null | http://arxiv.org/abs/1512.05294v2 | http://arxiv.org/pdf/1512.05294v2.pdf | Feature Representation for ICU Mortality | Good predictors of ICU Mortality have the potential to identify high-risk
patients earlier, improve ICU resource allocation, or create more accurate
population-level risk models. Machine learning practitioners typically make
choices about how to represent features in a particular model, but these
choices are seldom evaluated quantitatively. This study compares the
performance of different representations of clinical event data from MIMIC II
in a logistic regression model to predict 36-hour ICU mortality. The most
common representations are linear (normalized counts) and binary (yes/no).
These, along with a new representation termed "hill", are compared using both
L1 and L2 regularization. Results indicate that the introduced "hill"
representation outperforms both the binary and linear representations, the hill
representation thus has the potential to improve existing models of ICU
mortality. | ['Harini Suresh'] | 2015-12-16 | null | null | null | null | ['icu-mortality', 'l2-regularization'] | ['medical', 'methodology'] | [ 8.11955854e-02 -3.28158528e-01 -4.52038437e-01 -3.07856500e-01
-7.03067362e-01 -7.09494576e-02 1.67672843e-01 1.07570088e+00
-5.91745913e-01 9.12873089e-01 6.10880494e-01 -7.00550735e-01
-5.93156815e-01 -6.87085271e-01 -2.13655625e-02 -5.46835005e-01
-1.13325916e-01 6.48194075e-01 -2.94634372e-01 1.30174354e-01
1.67611465e-01 6.52783632e-01 -9.12794471e-01 3.30644488e-01
6.48065031e-01 8.30390036e-01 -2.37201229e-01 5.13917565e-01
3.95126753e-02 1.15443754e+00 -3.97356689e-01 2.25304425e-01
8.10398757e-02 -7.09141195e-01 -5.19245207e-01 -5.00004947e-01
-7.29028136e-02 -2.54370291e-02 -1.12304471e-01 2.04492927e-01
8.20882261e-01 5.39795905e-02 1.31904423e+00 -7.14627922e-01
-2.41684884e-01 4.60188568e-01 -2.50514537e-01 5.79340875e-01
1.50676504e-01 2.56883800e-02 5.32410920e-01 -6.06462121e-01
3.96257788e-02 1.15191460e+00 1.02287006e+00 3.76743376e-01
-1.38098848e+00 -6.19425476e-01 -3.06142986e-01 -1.99512728e-02
-1.31466424e+00 -3.88951087e-03 3.11824739e-01 -8.26467991e-01
1.01176655e+00 1.75988540e-01 3.38321596e-01 1.05198169e+00
5.74935913e-01 7.11204335e-02 1.26153994e+00 -4.74413902e-01
-5.72077371e-02 2.03439593e-01 5.55712938e-01 7.06426859e-01
6.88024521e-01 3.37334514e-01 -2.15079486e-01 -6.91852033e-01
9.63881791e-01 6.32946193e-01 -3.03285480e-01 -1.64455771e-01
-1.20218766e+00 1.18352938e+00 2.95988381e-01 3.35692137e-01
-7.66997993e-01 1.66234866e-01 6.66850924e-01 8.33299458e-02
2.01305717e-01 6.58630133e-01 -5.71043551e-01 -1.02786705e-01
-5.51034391e-01 -7.10581020e-02 6.60829484e-01 1.85603872e-01
4.79031831e-01 5.31332269e-02 -5.11683226e-01 1.02653313e+00
6.82205707e-02 1.00530542e-01 8.08569908e-01 -6.22632980e-01
2.72737145e-01 6.86408877e-01 -6.34515807e-02 -7.17905819e-01
-8.74315381e-01 -4.82582092e-01 -9.43863869e-01 2.35165656e-01
2.62452453e-01 -5.24142504e-01 -6.07157052e-01 1.29527819e+00
-4.89672899e-01 -6.76567629e-02 1.46975026e-01 4.65931773e-01
7.23187804e-01 4.14557874e-01 8.39770436e-01 -3.43444526e-01
1.14420784e+00 -4.82673109e-01 -5.04676640e-01 -1.16320282e-01
1.00443602e+00 -4.18045491e-01 5.76277196e-01 3.27576518e-01
-9.04985845e-01 -3.09673876e-01 -5.02109468e-01 3.43603939e-01
-2.54163831e-01 1.21228881e-01 6.31006241e-01 4.81539100e-01
-8.26805413e-01 7.88691163e-01 -5.93202531e-01 -4.36576992e-01
2.98137903e-01 4.46815044e-01 -2.97560036e-01 -8.86075869e-02
-1.02295935e+00 1.31164205e+00 3.18281949e-01 -4.60363746e-01
-5.13902605e-01 -6.75917208e-01 -9.02287364e-01 3.21998745e-01
-1.92558259e-01 -1.32548964e+00 7.33408391e-01 -7.06935406e-01
-9.17450607e-01 8.02679837e-01 -2.02373534e-01 -7.10883021e-01
4.13213789e-01 -3.99438947e-01 2.42550927e-03 1.99037030e-01
-1.27641976e-01 2.26089954e-01 3.97312760e-01 -9.68938887e-01
-4.46812391e-01 -2.98121721e-01 -3.34379435e-01 2.46690974e-01
1.02957888e-02 3.60604793e-01 5.73544204e-01 -8.53682578e-01
-2.60985434e-01 -7.15101957e-01 -6.20584488e-01 -2.35174343e-01
4.46211286e-02 -3.03950310e-01 5.29110253e-01 -5.44451058e-01
1.48131394e+00 -1.95558345e+00 -1.01580486e-01 1.29765674e-01
1.59458324e-01 9.76454690e-02 3.28416675e-01 4.83884871e-01
-5.19307554e-01 4.75097567e-01 -2.64015734e-01 -1.83050349e-01
-5.38095891e-01 2.16319054e-01 1.31264985e-01 5.28391063e-01
3.07319313e-01 7.75746465e-01 -7.12624907e-01 -5.57962537e-01
7.63264894e-01 6.59389377e-01 -5.23063302e-01 4.45375890e-01
5.06632149e-01 6.47315860e-01 -4.40511674e-01 1.66202575e-01
3.88979004e-03 -3.78278375e-01 1.64454952e-01 8.90490487e-02
-6.95321113e-02 2.62032896e-01 -5.03486216e-01 7.53800511e-01
-3.97243559e-01 4.04612809e-01 -5.77342868e-01 -1.16541064e+00
1.14739799e+00 5.48976243e-01 7.75608480e-01 -2.44256124e-01
5.01991510e-01 9.55561697e-02 7.78447688e-02 -4.25315380e-01
-4.18082714e-01 -8.34589779e-01 -1.96680445e-02 1.29250184e-01
-3.80435646e-01 9.49509144e-02 -1.42629340e-01 -2.16947824e-01
1.08673680e+00 -1.44521698e-01 9.01873827e-01 -4.08596367e-01
4.23706502e-01 -1.39042428e-02 5.24546623e-01 8.54962111e-01
-1.32982850e-01 9.31325376e-01 7.41628051e-01 -7.18275547e-01
-5.28647244e-01 -9.43322897e-01 -5.57708919e-01 7.41274655e-01
-4.16778088e-01 -6.88864291e-02 -1.66296929e-01 -5.18641293e-01
1.67075634e-01 9.51363862e-01 -8.54062915e-01 -5.73077440e-01
-7.01801181e-01 -1.20546854e+00 4.22823161e-01 8.52660835e-01
-2.94232100e-01 -9.71289456e-01 -1.09024918e+00 4.92055088e-01
1.54805761e-02 -4.39346701e-01 1.06560607e-02 7.94216514e-01
-1.43992734e+00 -1.38805997e+00 -8.41369510e-01 -6.83120549e-01
5.24543703e-01 -1.53374568e-01 1.21925485e+00 2.74697572e-01
-5.14991879e-01 2.99158603e-01 -5.02214134e-01 -5.57605088e-01
-5.08480430e-01 -1.86688349e-01 1.85651273e-01 -2.25078776e-01
5.02382994e-01 -3.75391960e-01 -8.97580206e-01 3.43135931e-02
-5.97564876e-01 -3.62624466e-01 8.09136391e-01 1.02207983e+00
5.27655423e-01 -4.63008732e-01 1.01254082e+00 -1.06656039e+00
8.14094245e-01 -9.55298901e-01 1.98317364e-01 -4.97454703e-02
-1.17227864e+00 -5.24805523e-02 8.26944172e-01 -2.62685210e-01
-4.53398585e-01 -1.65744126e-01 1.16582155e-01 -5.04324734e-01
-4.50908035e-01 5.45680642e-01 4.77154791e-01 1.51669189e-01
7.43386269e-01 -1.49817437e-01 -2.58981735e-02 -7.12677360e-01
-3.25858653e-01 4.83312011e-01 1.18177973e-01 -5.22678018e-01
4.18559276e-02 -7.09962249e-02 3.55025470e-01 -5.86480618e-01
-5.29586315e-01 -8.30831707e-01 -6.50364041e-01 2.69213349e-01
1.08579528e+00 -8.63367617e-01 -4.67015475e-01 -1.23628741e-02
-6.55763149e-01 -3.63970011e-01 -4.71045762e-01 9.47425961e-01
-6.00210667e-01 3.98120612e-01 -5.68003774e-01 -7.04653859e-01
-3.73885065e-01 -1.10301244e+00 6.34025872e-01 4.61201444e-02
-7.69306064e-01 -1.39447153e+00 2.89221019e-01 5.27494308e-03
5.24787605e-01 7.99527466e-01 1.74814534e+00 -1.08029389e+00
2.86512434e-01 -5.05945444e-01 -3.09833378e-01 5.25820613e-01
4.49579656e-01 -8.53475109e-02 -5.01827776e-01 -2.08854198e-01
-3.93694714e-02 -6.21592477e-02 8.86046588e-01 7.77460039e-01
1.18343747e+00 -2.85329878e-01 -2.86266088e-01 7.43747473e-01
1.49592483e+00 5.20514131e-01 4.96596605e-01 3.58804822e-01
4.93572414e-01 3.96339774e-01 1.14238761e-01 7.39898205e-01
2.48765558e-01 2.25565508e-01 2.87585348e-01 -3.51576984e-01
3.96693870e-02 1.98811352e-01 -3.65244150e-02 6.90185308e-01
-2.98256457e-01 -9.42393914e-02 -1.17804921e+00 3.41901302e-01
-1.60820210e+00 -5.62473238e-01 -3.55914861e-01 2.31805468e+00
5.17314911e-01 -7.36294389e-02 1.65093809e-01 1.20001115e-01
5.38652658e-01 -1.38171747e-01 -2.22962245e-01 -9.41874921e-01
-2.81537920e-01 5.14328122e-01 5.77503741e-01 1.30664110e-01
-1.17240846e+00 2.43680060e-01 7.64867449e+00 8.75463989e-03
-9.53764439e-01 -1.02335095e-01 1.04790723e+00 6.90209121e-02
3.35713625e-01 -7.12299952e-03 -4.49211895e-01 2.60458052e-01
1.50404584e+00 -2.39360526e-01 -7.33352005e-02 7.57219076e-01
5.38179040e-01 -7.59361088e-02 -1.24703133e+00 9.51839566e-01
8.04477558e-03 -1.17513406e+00 2.89009869e-01 -1.47034481e-01
4.73173589e-01 -2.03101560e-01 -2.37175226e-01 4.23790157e-01
2.58234978e-01 -1.51828873e+00 9.26258266e-02 9.30424154e-01
1.12403524e+00 -6.74524069e-01 1.32016778e+00 1.12825952e-01
-7.98025668e-01 -4.15729612e-01 -2.52820700e-01 -2.73926646e-01
1.37017727e-01 8.13458338e-02 -9.87227917e-01 2.74357289e-01
6.68695092e-01 7.00795412e-01 -6.58390582e-01 1.28996480e+00
2.73660213e-01 8.83099437e-01 -8.64247829e-02 1.75957546e-01
1.95525721e-01 -3.85359973e-02 2.55491197e-01 1.44697928e+00
4.09239411e-01 3.98912758e-01 1.88540235e-01 3.67660314e-01
1.85587257e-01 5.25595069e-01 -6.61919296e-01 2.34668449e-01
-3.40012065e-03 5.63475132e-01 -3.91738653e-01 -4.03754473e-01
-5.31011879e-01 4.33267415e-01 1.81046352e-01 1.20226897e-01
-3.67673814e-01 -1.06539711e-01 4.44331825e-01 5.70134521e-01
-2.91963756e-01 1.30468369e-01 -7.86875784e-01 -8.01453531e-01
-8.57843339e-01 -7.53030360e-01 9.75211799e-01 -6.01055801e-01
-1.39941251e+00 5.49235165e-01 2.13122621e-01 -1.15244269e+00
-3.32479894e-01 -7.93084502e-01 -7.28781402e-01 1.27435315e+00
-1.37751901e+00 -3.07820946e-01 -4.31212217e-01 4.90236789e-01
3.57619941e-01 -1.98371544e-01 1.43794656e+00 1.05347149e-01
-5.89697123e-01 3.43492448e-01 2.23396495e-01 1.96637690e-01
8.47434521e-01 -1.22655737e+00 -4.28233564e-01 -3.85850221e-02
-6.39817953e-01 7.44814634e-01 3.85777205e-01 -6.13921165e-01
-6.18969023e-01 -1.14518154e+00 7.22207487e-01 -5.00550687e-01
1.58263505e-01 7.02755094e-01 -1.17056561e+00 6.85388863e-01
-9.51332375e-02 -2.37942666e-01 1.33473241e+00 -1.25175059e-01
-2.00670287e-01 1.93984956e-01 -1.17249095e+00 3.22660238e-01
3.57808083e-01 -1.04483739e-01 -9.15028632e-01 2.52795428e-01
2.23473892e-01 2.72028387e-01 -1.35241640e+00 9.43506002e-01
4.77572858e-01 -1.04100478e+00 1.18144763e+00 -1.09881103e+00
3.62425685e-01 2.76900381e-01 9.97750685e-02 -1.22324610e+00
-7.22293735e-01 -1.91704407e-01 1.64702028e-01 6.03625476e-01
4.17561978e-01 -1.00179613e+00 4.81339842e-01 6.63487375e-01
3.24790962e-02 -1.15113389e+00 -8.98883700e-01 -4.81931478e-01
6.63541019e-01 7.61548756e-03 2.42614523e-01 9.24119174e-01
7.23252967e-02 4.14698452e-01 -3.04069430e-01 -3.53695303e-01
3.18717033e-01 -2.62149423e-01 2.41630077e-01 -1.68357825e+00
-5.26604168e-02 -1.00381982e+00 -4.00140017e-01 -7.45814592e-02
-6.49606436e-02 -9.42281425e-01 -3.34018707e-01 -1.99320698e+00
4.76218313e-01 -7.89797068e-01 -1.06218290e+00 4.78823125e-01
-5.55222631e-01 -1.27385870e-01 -5.58295026e-02 4.40447181e-01
1.21143341e-01 2.45272413e-01 6.54523611e-01 3.27022493e-01
-3.30826551e-01 1.66877404e-01 -6.51884913e-01 8.74370694e-01
1.11541367e+00 -8.11739743e-01 -4.17818092e-02 2.89515853e-02
-2.59049147e-01 5.32736063e-01 2.54854023e-01 -9.83638108e-01
-2.91893333e-01 -2.82562494e-01 7.35707343e-01 -2.04857625e-02
9.84370559e-02 -7.98526645e-01 1.29807889e-01 7.78943241e-01
-5.02269447e-01 6.15187824e-01 2.72248805e-01 3.60593557e-01
-2.19099909e-01 -2.75994927e-01 1.14348531e+00 -2.61772394e-01
-1.34780690e-01 1.40735731e-01 -6.60221457e-01 5.59545532e-02
9.99503195e-01 -1.77897900e-01 -1.85986802e-01 -4.09467191e-01
-9.25877452e-01 -2.53299368e-03 3.13748866e-01 2.40576521e-01
6.65473044e-01 -8.42457056e-01 -9.81269360e-01 -1.37097593e-02
1.53381512e-01 -2.84987986e-01 2.94767153e-02 1.15716290e+00
-9.90208864e-01 4.46020067e-01 -2.46811002e-01 -1.64234832e-01
-9.72557187e-01 7.32420921e-01 5.50905645e-01 -5.62956154e-01
-8.01677942e-01 6.79953098e-01 5.66783845e-01 1.33082762e-01
1.95498630e-01 -1.69370472e-01 -7.44901121e-01 1.19527644e-02
3.01557809e-01 7.16471374e-01 -3.61468583e-01 -7.44238019e-01
-5.07352233e-01 6.28800392e-01 1.74934149e-01 5.08471429e-01
1.22146106e+00 1.56583309e-01 -1.89601868e-01 8.74886215e-01
1.11302567e+00 -3.00817966e-01 -6.21850193e-01 1.14518397e-01
1.58297747e-01 -1.13388292e-01 -1.34250626e-01 -8.84745836e-01
-5.44280648e-01 1.14927340e+00 8.15468192e-01 -4.50474508e-02
1.04159451e+00 -2.19910130e-01 4.47949320e-01 2.45151326e-01
8.83772299e-02 -4.86452699e-01 -1.06787287e-01 4.51950848e-01
7.37757802e-01 -1.14666677e+00 5.51970415e-02 -1.77222431e-01
-9.05764401e-01 1.35468554e+00 1.85665980e-01 -4.99340773e-01
7.80819356e-01 4.54653837e-02 3.79946709e-01 8.11932981e-02
-7.67020524e-01 2.78055556e-02 3.01169187e-01 4.83928621e-01
8.89323175e-01 3.97565156e-01 -6.70420885e-01 7.98086941e-01
1.90303460e-01 1.81857139e-01 4.67775941e-01 8.11172783e-01
-3.98264408e-01 -1.08996391e+00 -2.89671361e-01 1.16817451e+00
-1.06997931e+00 -2.92726547e-01 -2.54218578e-01 6.98910177e-01
-2.00386755e-02 8.76198173e-01 -1.42499030e-01 -1.11286141e-01
5.96256018e-01 6.27331734e-01 2.80482857e-03 -8.02832901e-01
-1.10053611e+00 -9.51684415e-02 1.87848374e-01 -2.40994468e-01
-2.04164982e-01 -5.98228335e-01 -1.20525849e+00 2.94732992e-02
-6.79768855e-03 3.19368154e-01 3.20379943e-01 6.23328269e-01
3.89233351e-01 9.15608168e-01 4.91638273e-01 -5.69891751e-01
-7.15384841e-01 -1.14449584e+00 -4.40907359e-01 3.68487984e-01
3.92990410e-01 -7.79878378e-01 -7.28454769e-01 -6.64539263e-02] | [8.070352554321289, 6.0597310066223145] |
5f64f003-14be-4840-848f-7af4f383f0df | combining-noise-to-image-and-image-to-image | 1905.13456 | null | https://arxiv.org/abs/1905.13456v3 | https://arxiv.org/pdf/1905.13456v3.pdf | Combining Noise-to-Image and Image-to-Image GANs: Brain MR Image Augmentation for Tumor Detection | Convolutional Neural Networks (CNNs) achieve excellent computer-assisted diagnosis with sufficient annotated training data. However, most medical imaging datasets are small and fragmented. In this context, Generative Adversarial Networks (GANs) can synthesize realistic/diverse additional training images to fill the data lack in the real image distribution; researchers have improved classification by augmenting data with noise-to-image (e.g., random noise samples to diverse pathological images) or image-to-image GANs (e.g., a benign image to a malignant one). Yet, no research has reported results combining noise-to-image and image-to-image GANs for further performance boost. Therefore, to maximize the DA effect with the GAN combinations, we propose a two-step GAN-based DA that generates and refines brain Magnetic Resonance (MR) images with/without tumors separately: (i) Progressive Growing of GANs (PGGANs), multi-stage noise-to-image GAN for high-resolution MR image generation, first generates realistic/diverse 256 X 256 images; (ii) Multimodal UNsupervised Image-to-image Translation (MUNIT) that combines GANs/Variational AutoEncoders or SimGAN that uses a DA-focused GAN loss, further refines the texture/shape of the PGGAN-generated images similarly to the real ones. We thoroughly investigate CNN-based tumor classification results, also considering the influence of pre-training on ImageNet and discarding weird-looking GAN-generated images. The results show that, when combined with classic DA, our two-step GAN-based DA can significantly outperform the classic DA alone, in tumor detection (i.e., boosting sensitivity 93.67% to 97.48%) and also in other medical imaging tasks. | ['Yujiro Furukawa', 'Leonardo Rundo', 'Changhee Han', 'Yudai Nagano', 'Ryosuke Araki', 'Hideki Nakayama', 'Hideaki Hayashi', 'Giancarlo Mauri'] | 2019-05-31 | null | null | null | null | ['multimodal-unsupervised-image-to-image'] | ['computer-vision'] | [ 6.04375482e-01 2.64958590e-01 1.44660622e-01 -7.80097991e-02
-1.10513270e+00 -2.48351574e-01 4.96438444e-01 -4.55227494e-01
-4.13405657e-01 1.01533651e+00 1.08797848e-01 -2.46876642e-01
1.93368882e-01 -1.00459445e+00 -7.17026353e-01 -1.27860904e+00
3.61334175e-01 6.83568895e-01 -3.48423235e-02 -2.18850806e-01
-3.22891980e-01 5.00727773e-01 -1.18975377e+00 3.78523111e-01
1.02068830e+00 9.48709965e-01 4.39657480e-01 8.23347569e-01
-5.00647649e-02 9.92644072e-01 -7.44649112e-01 -2.49776125e-01
2.20452145e-01 -9.38121080e-01 -5.08526146e-01 7.44725466e-02
1.84521079e-02 -5.17937422e-01 -3.47914129e-01 1.10433590e+00
8.42308164e-01 -2.13570595e-01 7.31420994e-01 -1.16208923e+00
-1.02846885e+00 6.81361377e-01 -6.51823461e-01 1.64659917e-01
-7.72703066e-02 6.30594313e-01 6.98285848e-02 -4.79449421e-01
6.78370297e-01 9.28450763e-01 5.38775504e-01 9.71289217e-01
-1.30409610e+00 -6.66675389e-01 -5.40213585e-01 -2.78606545e-02
-1.14541447e+00 -3.99149582e-02 8.97922218e-01 -4.33768898e-01
3.95103008e-01 5.06695092e-01 8.65882277e-01 1.48994815e+00
4.46556956e-01 6.10903084e-01 1.59440506e+00 -3.23003769e-01
8.48110840e-02 2.36391574e-02 -4.06740546e-01 5.07424831e-01
2.02548862e-01 2.71660239e-01 -5.64370453e-02 1.14075042e-01
1.12432957e+00 1.13898546e-01 -4.04255092e-01 2.79793411e-01
-1.13513005e+00 7.83261836e-01 6.40808702e-01 6.14186168e-01
-7.97958136e-01 1.90404654e-01 3.42717230e-01 1.41646504e-01
3.65786791e-01 5.27340412e-01 -4.23287675e-02 1.45572990e-01
-1.05747306e+00 1.84500501e-01 1.85917586e-01 5.69629610e-01
5.33482313e-01 6.84105456e-01 -5.37266314e-01 8.68849099e-01
-1.67821169e-01 6.57532454e-01 1.06214786e+00 -6.30649090e-01
1.45005509e-01 4.44251776e-01 -4.38110352e-01 -6.29397273e-01
-4.49628055e-01 -9.56381738e-01 -1.66061831e+00 3.03812087e-01
1.76624238e-01 -2.44422168e-01 -1.39681840e+00 1.87818813e+00
2.30525404e-01 1.52471840e-01 1.64345190e-01 9.47177231e-01
1.12465084e+00 4.30925310e-01 1.60477042e-01 -3.13590109e-01
1.39762092e+00 -9.77178514e-01 -8.65219891e-01 -8.47359002e-02
5.11454225e-01 -5.96029222e-01 1.13362193e+00 3.22749048e-01
-1.32725000e+00 -5.52867889e-01 -8.56965423e-01 2.40445584e-01
-1.26137555e-01 1.91085652e-01 4.59018797e-01 8.81235421e-01
-1.32088482e+00 1.96532786e-01 -9.14887249e-01 7.81980231e-02
8.47673237e-01 2.12776095e-01 -3.90933603e-01 -3.12772274e-01
-1.20667052e+00 7.74468064e-01 3.39512795e-01 4.42701019e-02
-1.27807426e+00 -1.10943961e+00 -6.77150726e-01 -1.77525669e-01
1.01360321e-01 -1.29476643e+00 6.90925419e-01 -1.48025894e+00
-1.51532769e+00 8.72331321e-01 1.58682674e-01 -5.50786018e-01
8.48241031e-01 2.49374375e-01 -2.31072456e-01 3.06886882e-01
4.85429615e-02 1.00454032e+00 8.28869104e-01 -1.45770860e+00
-2.24922001e-02 -3.70421022e-01 -2.29295954e-01 1.25210971e-01
7.44131729e-02 -1.10698260e-01 -6.75504794e-03 -1.11359262e+00
1.49966013e-02 -9.14780080e-01 -3.96192044e-01 -2.91801751e-01
-6.10393226e-01 3.83393258e-01 6.91457808e-01 -1.03361797e+00
7.65275657e-01 -1.90146255e+00 8.07650238e-02 1.28306493e-01
4.79186088e-01 2.85172224e-01 -2.71866232e-01 -1.27373397e-01
-3.76432657e-01 2.87094355e-01 -6.38865590e-01 -3.86084855e-01
-4.89496708e-01 2.45596498e-01 3.04202706e-01 3.22780788e-01
2.29340062e-01 1.53952777e+00 -7.56558359e-01 -4.36264843e-01
2.76120722e-01 7.00156808e-01 -5.33359408e-01 3.48083317e-01
1.25566483e-01 1.19608963e+00 -2.31807768e-01 7.01990485e-01
9.08493698e-01 -1.51144892e-01 -1.24710292e-01 -4.31614488e-01
4.26390648e-01 -4.61373001e-01 -5.83023012e-01 1.48912740e+00
-4.59556699e-01 3.75940531e-01 1.83684267e-02 -1.13168991e+00
7.64510632e-01 3.24702114e-01 7.29684055e-01 -8.55044365e-01
2.96003193e-01 2.25147471e-01 2.54443675e-01 -4.91073102e-01
1.49519190e-01 -6.68483317e-01 2.65723288e-01 1.17121652e-01
5.19651771e-02 -2.38799438e-01 -1.46809295e-01 -8.84225499e-03
1.10103512e+00 -2.82475471e-01 -4.97111268e-02 -5.89717962e-02
4.59977448e-01 -6.29399419e-02 2.75385946e-01 8.15123320e-01
1.62476003e-02 1.17097461e+00 5.03463268e-01 -2.07477674e-01
-1.35298276e+00 -9.44379270e-01 2.64238268e-02 2.86065847e-01
-2.26542220e-01 3.02590251e-01 -1.22042882e+00 -6.41571462e-01
-5.66964805e-01 6.37867987e-01 -8.55521798e-01 -4.16246563e-01
-4.55684811e-01 -1.34923279e+00 8.97401750e-01 5.19967318e-01
7.31659830e-01 -1.30137157e+00 -3.31907302e-01 2.15398446e-01
-2.79610276e-01 -1.01199961e+00 -3.03433180e-01 1.25736356e-01
-8.34244847e-01 -9.01867807e-01 -1.29685152e+00 -6.75335884e-01
1.02713525e+00 -2.62368381e-01 1.13357556e+00 1.12821028e-01
-5.41276634e-01 2.04108655e-01 -5.52512527e-01 -1.83473229e-01
-1.03567696e+00 -1.27821401e-01 -2.88643569e-01 7.08162785e-02
-3.86660755e-01 -7.08271861e-01 -8.96993577e-01 2.15996996e-01
-1.33104515e+00 5.31025827e-01 1.14055884e+00 1.48793972e+00
7.69243777e-01 -2.94043534e-02 6.49476767e-01 -1.13584161e+00
5.65327644e-01 -4.83705610e-01 -1.14004061e-01 1.23700604e-01
-5.66350937e-01 -1.39039353e-01 7.01689005e-01 -8.09635878e-01
-9.73276079e-01 -1.21534295e-01 -6.58819258e-01 -8.29071403e-01
-3.21885705e-01 4.10745412e-01 -1.27121329e-01 -1.56967103e-01
9.11138654e-01 8.28356862e-01 3.55018586e-01 2.64334213e-03
2.87396014e-01 3.49097550e-01 6.67913198e-01 -4.02340323e-01
6.74983621e-01 4.22060370e-01 7.66220689e-02 -5.48241735e-01
-3.06456298e-01 2.16477066e-01 -1.98041826e-01 -1.97299004e-01
1.29506874e+00 -8.24665010e-01 -3.16007704e-01 9.16078508e-01
-8.33181560e-01 -6.54446781e-01 -6.68338776e-01 6.24835908e-01
-5.95838368e-01 7.65257776e-02 -9.56683695e-01 -4.48117018e-01
-7.96699584e-01 -1.71185398e+00 9.47906911e-01 3.05329710e-01
1.99061573e-01 -8.40672672e-01 -1.63120598e-01 6.53900743e-01
9.08480942e-01 8.95059943e-01 8.96900594e-01 -5.23970366e-01
-4.55643088e-01 -6.43314347e-02 -2.27509782e-01 7.45481253e-01
2.01244265e-01 -4.52550918e-01 -8.54441226e-01 -3.01471680e-01
2.52933711e-01 -4.08161551e-01 7.97038913e-01 9.23001468e-01
1.35835111e+00 -3.67895067e-01 2.62333769e-02 8.75212848e-01
1.38244891e+00 4.89079297e-01 1.31609416e+00 9.32277516e-02
9.43485141e-01 7.18834326e-02 1.06549105e-02 1.88565820e-01
4.10610326e-02 4.40540493e-01 4.49956924e-01 -7.65255928e-01
-7.28611588e-01 -1.65958092e-01 1.35058373e-01 7.18587577e-01
-3.00529748e-01 -2.70670325e-01 -5.65244317e-01 3.64912182e-01
-1.12041771e+00 -7.76161253e-01 -1.06475145e-01 1.80414152e+00
8.80334079e-01 -1.16216298e-02 -7.10652545e-02 3.89272161e-02
8.37313473e-01 -1.24569692e-01 -6.79314137e-01 1.27551109e-01
-4.73654836e-01 6.25275970e-01 5.12430251e-01 2.71174699e-01
-6.36317372e-01 6.48427963e-01 5.40818119e+00 1.45755100e+00
-1.48950779e+00 6.78160608e-01 1.38855720e+00 1.03269316e-01
-6.54555500e-01 -3.69564533e-01 -1.70366958e-01 6.53090000e-01
7.08365440e-01 3.12901825e-01 4.98076379e-01 6.09992623e-01
2.23535985e-01 -5.55840582e-02 -5.73392510e-01 1.16003883e+00
1.13225162e-01 -1.45588410e+00 4.01075482e-01 1.58251926e-01
9.75934863e-01 -2.11938992e-01 2.42892325e-01 2.69977599e-01
1.46919712e-01 -1.41163445e+00 5.67238152e-01 7.19122469e-01
1.43816602e+00 -7.02777982e-01 1.15226221e+00 2.07097992e-01
-6.24017000e-01 2.82186151e-01 -8.80188271e-02 6.10428393e-01
1.80048764e-01 7.88427830e-01 -8.25927734e-01 7.49510705e-01
4.53947723e-01 2.00341731e-01 -6.09750688e-01 4.60475683e-01
-6.16161674e-02 6.88467622e-01 -1.36898339e-01 2.71298528e-01
2.02952310e-01 -2.31291801e-01 5.78842103e-01 8.65788341e-01
4.31071550e-01 1.89628899e-01 -1.13621883e-01 1.39623857e+00
-1.54124647e-01 -1.70472544e-02 -4.61649865e-01 2.26391464e-01
1.78211674e-01 1.39212751e+00 -8.89643610e-01 -5.04686117e-01
6.36984482e-02 1.10874844e+00 -1.35501564e-01 4.23060596e-01
-9.26267922e-01 -7.13923275e-02 5.01151010e-02 4.10086155e-01
-6.74765408e-02 3.54920924e-01 -4.13923472e-01 -9.48068738e-01
-1.73569709e-01 -1.14995635e+00 1.97563231e-01 -1.10803139e+00
-1.22992015e+00 1.02655876e+00 -1.62555411e-01 -9.87727404e-01
-2.42409110e-01 -2.20961183e-01 -7.21860111e-01 9.72658515e-01
-1.32804465e+00 -1.57340884e+00 -6.68703496e-01 7.22598493e-01
4.80696470e-01 -2.71893829e-01 5.81760705e-01 5.03451645e-01
-3.46729517e-01 9.24469173e-01 -6.54243724e-03 2.34109640e-01
4.11627591e-01 -1.00826597e+00 1.46962199e-02 7.88644075e-01
-2.61254996e-01 2.46908844e-01 3.70831341e-01 -6.66846216e-01
-1.16109645e+00 -1.24572933e+00 1.87003575e-02 -9.78460386e-02
1.75899923e-01 -1.11972854e-01 -8.02077472e-01 5.68836629e-01
3.02510172e-01 1.23102449e-01 3.07730079e-01 -8.07986081e-01
2.33677343e-01 -1.22771621e-01 -1.84311008e+00 6.19318962e-01
7.77562022e-01 -1.66661665e-01 4.35887778e-04 2.66563475e-01
8.32162082e-01 -8.02036881e-01 -1.06391776e+00 6.68141127e-01
2.37058669e-01 -9.61050391e-01 1.03741467e+00 -3.82012695e-01
9.14438963e-01 -1.94188520e-01 -2.88741644e-02 -1.54304910e+00
-1.65430099e-01 -3.98356706e-01 4.60837811e-01 1.01229203e+00
2.72773236e-01 -8.11193943e-01 7.91774571e-01 4.44699615e-01
-4.77780819e-01 -1.00400710e+00 -9.24983442e-01 -5.05849063e-01
2.55984455e-01 -2.11077079e-01 6.91349208e-01 1.11534178e+00
-7.97560036e-01 -1.19557865e-01 -4.53293025e-01 -1.39913857e-01
5.08036613e-01 -6.10890567e-01 6.56730473e-01 -5.31386971e-01
-3.37205201e-01 -5.84014535e-01 -4.52789903e-01 -4.83935446e-01
-1.14076167e-01 -9.62640643e-01 -3.05195987e-01 -1.36139739e+00
3.24178338e-01 -6.11051440e-01 -1.47413656e-01 4.87285763e-01
-3.61875027e-01 6.83775306e-01 2.09280655e-01 1.32331520e-01
9.85208973e-02 5.41400611e-01 1.97527623e+00 -3.76768768e-01
2.22435147e-02 -2.60458618e-01 -7.71466017e-01 2.19961509e-01
7.19369113e-01 -5.12364626e-01 -3.42366189e-01 -3.91897932e-02
-1.85386047e-01 4.87836272e-01 6.26904845e-01 -1.15250731e+00
-6.80390373e-02 6.68574348e-02 5.93163967e-01 -4.08874750e-01
6.11086674e-02 -5.15769660e-01 8.32048118e-01 8.66014898e-01
-2.14437231e-01 1.34770595e-03 5.52516542e-02 2.00018376e-01
-3.68108392e-01 -2.38459691e-01 1.16004050e+00 -5.26257038e-01
-1.16631992e-01 3.70455176e-01 -3.11881214e-01 1.66324288e-01
9.60267425e-01 -1.95757613e-01 -2.73510844e-01 -4.60407436e-01
-9.04478133e-01 -2.57289916e-01 3.05871397e-01 -2.06706303e-04
6.31782770e-01 -1.42755508e+00 -1.12366152e+00 3.10850978e-01
-2.98248678e-01 3.17771167e-01 8.72700334e-01 1.27840030e+00
-8.62645388e-01 -1.13711402e-01 -4.30787474e-01 -7.91346490e-01
-9.44560647e-01 4.29729193e-01 6.12841129e-01 -8.24243367e-01
-6.82811975e-01 7.40612447e-01 5.46609402e-01 -3.77035826e-01
-2.65464902e-01 -3.02347630e-01 -1.08183302e-01 -2.58108377e-01
4.00728375e-01 1.21002972e-01 2.77787387e-01 -3.70745391e-01
8.86602327e-03 3.61145347e-01 -3.42586264e-03 -3.57684530e-02
1.25596654e+00 2.13533103e-01 -1.79823801e-01 -4.58056852e-02
1.17975938e+00 -2.03762844e-01 -1.05001462e+00 3.20619978e-02
-8.26239765e-01 -2.18023837e-01 1.58818379e-01 -1.05399215e+00
-1.87557364e+00 6.27071798e-01 1.07055712e+00 -7.11527169e-02
1.56640327e+00 -3.95777300e-02 9.28357542e-01 -4.49520618e-01
1.53222471e-01 -5.48021138e-01 2.61800706e-01 -5.67449033e-02
1.07123303e+00 -1.11975610e+00 -2.99824953e-01 -1.74969777e-01
-8.83823097e-01 8.94687653e-01 5.25430202e-01 -1.26814798e-01
3.09140027e-01 5.12156010e-01 1.63146302e-01 -1.27007499e-01
-3.33003521e-01 -1.41925476e-02 1.52320430e-01 8.69357109e-01
7.65965208e-02 3.33429903e-01 -2.52663672e-01 8.36153567e-01
-3.26310635e-01 1.23376146e-01 5.39412677e-01 4.78559822e-01
1.23408586e-01 -8.96989584e-01 -5.61269939e-01 7.57368743e-01
-5.06004930e-01 -4.32675064e-01 9.32723880e-02 7.66919732e-01
3.90893489e-01 5.78466594e-01 -7.52497911e-02 -4.63535011e-01
1.16073564e-01 -8.61625448e-02 4.93353575e-01 -1.99632078e-01
-8.95004034e-01 2.51439869e-01 -2.08479643e-01 -2.96318889e-01
-3.46300721e-01 -3.53060603e-01 -9.65666533e-01 -2.97969341e-01
-2.71094024e-01 -1.46600768e-01 6.93052590e-01 8.40871453e-01
2.41770223e-01 1.26053512e+00 5.75246692e-01 -7.81108916e-01
-3.02963018e-01 -1.26470065e+00 -5.39776385e-01 6.20802402e-01
1.79713219e-01 -4.22995716e-01 -4.79155838e-01 1.09345235e-01] | [14.055320739746094, -2.0179402828216553] |
012c9b68-6a87-4647-a249-fb62b59557ad | simplymime-a-control-at-our-fingertips | 2304.11377 | null | https://arxiv.org/abs/2304.11377v1 | https://arxiv.org/pdf/2304.11377v1.pdf | SimplyMime: A Control at Our Fingertips | The utilization of consumer electronics, such as televisions, set-top boxes, home theaters, and air conditioners, has become increasingly prevalent in modern society as technology continues to evolve. As new devices enter our homes each year, the accumulation of multiple infrared remote controls to operate them not only results in a waste of energy and resources, but also creates a cumbersome and cluttered environment for the user. This paper presents a novel system, named SimplyMime, which aims to eliminate the need for multiple remote controls for consumer electronics and provide the user with intuitive control without the need for additional devices. SimplyMime leverages a dynamic hand gesture recognition architecture, incorporating Artificial Intelligence and Human-Computer Interaction, to create a sophisticated system that enables users to interact with a vast majority of consumer electronics with ease. Additionally, SimplyMime has a security aspect where it can verify and authenticate the user utilising the palmprint, which ensures that only authorized users can control the devices. The performance of the proposed method for detecting and recognizing gestures in a stream of motion was thoroughly tested and validated using multiple benchmark datasets, resulting in commendable accuracy levels. One of the distinct advantages of the proposed method is its minimal computational power requirements, making it highly adaptable and reliable in a wide range of circumstances. The paper proposes incorporating this technology into all consumer electronic devices that currently require a secondary remote for operation, thus promoting a more efficient and sustainable living environment. | ['Anitha Subramanian', 'Saraju P. Mohanty', 'Athresh Kiran', 'Gaurav Reddy Tadkapally', 'Sibi Chakkaravarthy Sethuraman'] | 2023-04-22 | null | null | null | null | ['hand-gesture-recognition', 'hand-gesture-recognition-1', 'gesture-recognition'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 2.74308711e-01 -3.89485151e-01 -5.51067293e-02 -1.06443353e-01
3.23743224e-02 -7.78863430e-01 3.11823189e-01 -3.01906884e-01
-5.99935472e-01 3.51389617e-01 -2.27825776e-01 -2.37540349e-01
-2.48825103e-01 -8.31204772e-01 -2.97017768e-02 -7.15103447e-01
1.45084113e-01 -4.90889465e-03 3.13874424e-01 -5.81485890e-02
6.16612434e-02 7.66262770e-01 -1.69948065e+00 -1.57809570e-01
5.08081853e-01 1.27375507e+00 2.08512485e-01 5.95525265e-01
2.47365251e-01 1.92710251e-01 -5.66258132e-01 -6.16397224e-02
4.74330217e-01 -5.81984520e-02 -1.66416213e-01 -4.36553098e-02
-1.65376499e-01 -9.36609209e-01 -1.70792714e-01 6.52524471e-01
7.56660998e-01 1.93674967e-01 2.11254895e-01 -1.09293151e+00
-1.84164867e-01 5.17822541e-02 -5.70987523e-01 -3.12592834e-01
8.33363414e-01 4.48380709e-01 5.98858356e-01 -1.77374169e-01
3.58455747e-01 8.43354404e-01 4.23056811e-01 4.47654605e-01
-7.97846735e-01 -9.16966558e-01 -5.90476729e-02 2.10556820e-01
-1.49773490e+00 -4.41825479e-01 6.52944744e-01 -1.77464217e-01
1.02114987e+00 7.60253906e-01 8.03493619e-01 8.81433308e-01
6.96728975e-02 4.36568439e-01 9.20728445e-01 -6.63763463e-01
3.45982254e-01 2.20785886e-01 1.20886564e-01 6.21571243e-01
4.90490466e-01 -7.70763606e-02 -1.04917496e-01 -4.56460677e-02
7.54321158e-01 6.01610720e-01 -5.47771215e-01 -2.28781849e-01
-1.11727476e+00 1.65572390e-01 1.38621166e-01 4.74653870e-01
-5.80664873e-01 -1.09791629e-01 2.49724105e-01 -3.72339152e-02
-4.22861964e-01 -7.21176267e-02 -2.99463034e-01 -5.97884238e-01
-7.08849907e-01 -1.72945410e-01 8.63563538e-01 9.73858654e-01
8.27135816e-02 -3.07398945e-01 3.59927624e-01 4.62181687e-01
6.88060880e-01 6.78013563e-01 5.50743163e-01 -4.48253244e-01
4.07527149e-01 8.21513951e-01 4.19090658e-01 -1.01737952e+00
-4.25547749e-01 1.30089447e-01 -8.14744115e-01 5.99015892e-01
2.59975255e-01 -1.20696157e-01 -7.97545791e-01 1.36316395e+00
4.87299711e-01 -3.78434151e-01 -2.62482584e-01 8.54081273e-01
3.30097973e-01 5.17044961e-01 1.70966133e-01 -2.30998509e-02
1.56103063e+00 -3.61123711e-01 -9.03342247e-01 1.33842468e-01
3.66637902e-03 -8.54819596e-01 1.12221384e+00 7.22025037e-01
-4.36543465e-01 -4.95491624e-01 -1.43324447e+00 2.23486453e-01
-4.59685236e-01 3.49026233e-01 5.70550740e-01 1.17520630e+00
-6.75121248e-01 5.47356904e-01 -1.08684790e+00 -9.56002355e-01
2.11361870e-01 8.78394663e-01 -3.34545851e-01 2.06601024e-01
-6.88610673e-01 7.67141998e-01 2.63247877e-01 2.32990324e-01
-5.08448146e-02 -1.69705048e-01 -5.37515879e-01 -9.31723267e-02
9.19923410e-02 -2.38983110e-01 9.70714271e-01 -5.73281288e-01
-1.89581859e+00 3.94629538e-01 1.81150138e-01 2.25987300e-01
8.45165312e-01 -6.18947268e-01 -7.77060509e-01 2.12972865e-01
-4.23595995e-01 2.26415634e-01 7.32958436e-01 -7.98228741e-01
-6.60149753e-01 -4.42833275e-01 -8.25132951e-02 3.48980986e-02
-7.91386724e-01 1.16546854e-01 -4.78898853e-01 -4.25745517e-01
5.36257923e-02 -1.07682860e+00 3.37198764e-01 2.04522640e-01
-2.29318514e-01 -1.62908752e-02 1.43988347e+00 -8.12478423e-01
1.28048360e+00 -2.07598066e+00 -2.08951950e-01 5.32719791e-01
-1.84828401e-01 6.15043104e-01 3.24096441e-01 4.06366527e-01
1.93636924e-01 -2.11733252e-01 4.12732475e-02 6.08932823e-02
1.10488608e-01 9.16572586e-02 1.40102684e-01 5.49176633e-01
-1.56986460e-01 5.18064976e-01 -4.53291655e-01 -1.17321454e-01
8.49380672e-01 7.25853980e-01 -1.60633042e-01 1.81281313e-01
3.99680972e-01 3.87029052e-01 -5.92350602e-01 9.08122838e-01
4.44416672e-01 5.39933480e-02 3.30330789e-01 -2.19072238e-01
-3.67290705e-01 -1.18257374e-01 -1.62154543e+00 1.21400833e+00
-5.69776535e-01 3.08293968e-01 3.88299078e-01 -4.39779818e-01
7.95293450e-01 6.13618076e-01 4.99558687e-01 -5.65872371e-01
5.43092370e-01 1.72066763e-01 -3.46285254e-01 -8.44244897e-01
1.87122434e-01 2.27706760e-01 -2.97358204e-02 6.25607550e-01
-5.10441422e-01 2.46458337e-01 -1.80220991e-01 -2.27068320e-01
1.28541183e+00 6.15477085e-01 5.94285369e-01 8.04871917e-02
6.70621276e-01 -3.88623267e-01 2.45513916e-01 2.67750055e-01
-3.08378756e-01 4.66647409e-02 -6.12997711e-01 -2.48176858e-01
-7.54469335e-01 -1.07798970e+00 2.10756794e-01 9.78589892e-01
3.11350793e-01 -6.88267425e-02 -5.52520037e-01 -2.82558233e-01
-3.44832949e-02 3.27880442e-01 7.90427402e-02 1.28961861e-01
-6.63968027e-01 -3.05108398e-01 4.45733994e-01 5.45154214e-01
9.57284093e-01 -8.83868217e-01 -1.55284524e+00 1.02942526e-01
1.12784304e-01 -1.07251263e+00 -3.82701486e-01 -6.91163912e-02
-7.98165858e-01 -9.10903335e-01 -6.61051214e-01 -6.37465656e-01
6.45696938e-01 3.24539989e-01 2.93970257e-01 1.99320540e-01
-6.27965868e-01 5.96374214e-01 -6.09249353e-01 -4.30434614e-01
-2.21709795e-02 4.69249785e-02 3.34513366e-01 8.10092986e-02
5.47905266e-01 -7.41539776e-01 -8.88805091e-01 4.86324042e-01
-8.55938971e-01 -1.61503211e-01 6.52029514e-01 3.97796482e-01
-5.67273945e-02 3.23732525e-01 4.57827300e-01 -2.91269839e-01
4.02435094e-01 -1.94971547e-01 -6.57610357e-01 2.65217721e-01
-6.25455201e-01 -1.98573992e-01 7.19948351e-01 -5.63231885e-01
-1.09176874e+00 3.54167581e-01 1.63818318e-02 2.82338589e-01
-4.22506720e-01 -1.86265543e-01 -5.66956103e-01 -2.70617992e-01
2.20826000e-01 1.19657805e-02 -9.94490460e-03 -6.23131692e-01
2.39833459e-01 1.44366181e+00 6.60104036e-01 -1.11856781e-01
9.00092900e-01 5.34519792e-01 -1.35504484e-01 -1.22418213e+00
4.24504429e-01 -7.25870609e-01 -7.37437963e-01 -5.52159429e-01
7.46743619e-01 -5.35585344e-01 -1.43216896e+00 7.93043017e-01
-9.50069726e-01 1.25592425e-01 2.96445459e-01 5.69935381e-01
2.52945665e-02 5.26934505e-01 -2.74102926e-01 -1.26284194e+00
-6.63108528e-01 -9.18945134e-01 7.44440317e-01 5.77228010e-01
-9.11638916e-01 -5.36013484e-01 -5.25347054e-01 4.77411807e-01
6.33416116e-01 5.75476706e-01 8.14137936e-01 -2.08397061e-01
-6.66475952e-01 -7.78079152e-01 -1.54165868e-02 2.00223729e-01
8.25451672e-01 -3.78947072e-02 -1.01223266e+00 -4.47327912e-01
-5.96392937e-02 8.81184265e-02 3.28154229e-02 -2.14411423e-01
6.57593369e-01 -3.44147146e-01 -5.77845633e-01 1.90782607e-01
1.42372739e+00 9.50324833e-01 7.22651601e-01 4.88542110e-01
5.70308864e-01 4.93718594e-01 5.27878225e-01 5.72777689e-01
6.41257688e-02 7.43348420e-01 2.96063066e-01 -5.13548963e-02
3.43908638e-01 7.74262771e-02 4.22250032e-01 6.52799189e-01
-4.66914088e-01 -2.19646156e-01 -4.92528588e-01 1.88262895e-01
-1.55957162e+00 -7.82578409e-01 1.56864509e-01 2.42142630e+00
3.89661431e-01 -1.53131485e-01 2.09352300e-01 7.02825665e-01
8.50550056e-01 -3.20571631e-01 -6.42937899e-01 -1.46369785e-01
5.69178045e-01 7.89350271e-01 7.09476113e-01 3.15077342e-02
-1.12277508e+00 3.60521674e-01 5.69053888e+00 6.88767955e-02
-1.42285669e+00 -8.92851800e-02 -1.00093052e-01 -3.23763728e-01
3.42417985e-01 -4.96132761e-01 -3.54028642e-01 5.60390055e-01
5.51900983e-01 3.47437978e-01 5.10376751e-01 9.00542498e-01
5.03417253e-01 -4.05825406e-01 -1.22010255e+00 1.25652003e+00
-2.61937559e-01 -5.90976298e-01 -2.67990261e-01 9.11799595e-02
1.08779468e-01 -3.80388528e-01 -8.23564604e-02 -3.05900097e-01
-1.36542544e-01 -7.91976035e-01 8.05013478e-01 2.62974977e-01
7.50866413e-01 -7.47400999e-01 7.49975562e-01 2.86065578e-01
-1.33918452e+00 -3.83754730e-01 3.44712466e-01 -4.14935976e-01
3.25888902e-01 -6.78272569e-04 -7.29532719e-01 3.61019671e-01
9.49244201e-01 -1.19001992e-01 -1.66492034e-02 7.78138757e-01
-1.60046041e-01 1.26767173e-01 -7.31178701e-01 -5.99614263e-01
-3.04798305e-01 -2.00294659e-01 3.89905781e-01 1.00288236e+00
6.01772785e-01 4.12771463e-01 -1.05310760e-01 3.58488679e-01
1.91420451e-01 1.28478870e-01 -2.93091208e-01 -6.70438632e-02
5.67970216e-01 1.34814250e+00 -9.79389131e-01 -1.07888371e-01
-4.17256027e-01 1.46733260e+00 -5.11087358e-01 1.07754558e-01
-6.68296576e-01 -9.74598944e-01 6.81459904e-01 2.64459938e-01
2.73437142e-01 -6.13615513e-01 -1.37011573e-01 -6.85875952e-01
4.50413287e-01 -8.75348628e-01 1.37577787e-01 -4.47694838e-01
-7.98888922e-01 2.13842094e-01 -1.25234097e-01 -1.18515980e+00
-2.15456918e-01 -9.26934481e-01 -4.50493306e-01 7.37876117e-01
-9.18690205e-01 -1.11984217e+00 -7.57145584e-01 6.99683785e-01
3.39346677e-01 -4.99192774e-02 1.01192415e+00 6.64710462e-01
-4.70142007e-01 6.69086039e-01 8.29740986e-02 1.01526447e-01
5.79831600e-01 -6.83784187e-01 1.79553136e-01 8.96824956e-01
-3.11187327e-01 1.08548665e+00 4.77089673e-01 -5.66443563e-01
-1.93750107e+00 -5.11566699e-01 3.66615206e-01 -2.70541131e-01
2.54705906e-01 -4.84093159e-01 -3.04202795e-01 4.45189446e-01
-6.82459846e-02 -4.54168499e-01 6.46002471e-01 -3.90202731e-01
-2.39950512e-02 -5.31721234e-01 -1.67226493e+00 6.12126112e-01
9.91417408e-01 -4.78570044e-01 -4.39917505e-01 -1.13701992e-01
8.69045630e-02 3.41353589e-03 -7.21981049e-01 1.39749527e-01
1.42029214e+00 -8.26610088e-01 6.90777719e-01 6.67290092e-02
-3.99464816e-01 -5.81196129e-01 -8.60666856e-02 -6.48612440e-01
-8.46520662e-02 -9.80464816e-01 -5.60759194e-02 1.45623958e+00
1.35459993e-02 -1.12492836e+00 6.63066328e-01 1.30902183e+00
4.30602342e-01 -3.80596161e-01 -9.01691794e-01 -6.42814457e-01
-9.09119487e-01 -2.71574855e-01 8.09846520e-01 6.08325422e-01
4.13433582e-01 -1.51594669e-01 -2.58008063e-01 3.58776540e-01
4.98973161e-01 -3.40649962e-01 7.34341681e-01 -1.30138230e+00
-3.13139170e-01 -3.07746921e-02 -8.10190916e-01 -1.02960539e+00
-6.09626114e-01 -3.82409930e-01 6.82756770e-03 -1.35069084e+00
-1.69139907e-01 -3.83970469e-01 -1.45548701e-01 6.29127145e-01
3.30640435e-01 4.12604094e-01 1.53959394e-01 1.69571802e-01
-1.32406801e-01 6.18643174e-03 6.84394538e-01 -7.82250389e-02
-5.42985260e-01 1.70119360e-01 -3.67036790e-01 7.53166318e-01
7.65151799e-01 -2.80120261e-02 -4.12906438e-01 -1.07499860e-01
-1.36547238e-01 -4.85304594e-01 3.15573156e-01 -1.30440748e+00
3.04373801e-01 7.96085671e-02 8.02228212e-01 -1.97753429e-01
2.70779252e-01 -1.70183682e+00 5.37917852e-01 8.18186224e-01
4.03693080e-01 1.43205315e-01 -6.09727856e-03 2.04160869e-01
4.26203161e-01 1.73044845e-01 7.57236779e-01 6.61124960e-02
-8.09617817e-01 -2.96296477e-01 -5.62958419e-01 -1.15488696e+00
1.52397704e+00 -6.52429640e-01 -1.13538928e-01 -2.90727943e-01
-3.83192480e-01 -8.99663121e-02 5.65744817e-01 5.69535196e-01
3.92984837e-01 -1.14665353e+00 3.79511207e-01 4.21162337e-01
-1.59759998e-01 -2.31262356e-01 3.31610516e-02 3.95661801e-01
-9.29274321e-01 4.70748693e-01 -4.04601276e-01 -4.28728193e-01
-1.80441773e+00 2.09747419e-01 -2.11342245e-01 2.30774134e-01
-8.22504103e-01 3.45004886e-01 -6.92277908e-01 -6.65599853e-02
4.34739500e-01 -6.16141438e-01 -6.90378845e-02 -7.99824074e-02
8.07028234e-01 9.70544219e-01 1.22591235e-01 -4.88406986e-01
-6.17364347e-01 8.08616698e-01 8.92186984e-02 -2.22405598e-01
1.31070042e+00 -3.16565126e-01 -8.78896564e-03 8.41332227e-02
8.46346140e-01 2.05386765e-02 -9.83903050e-01 4.64994282e-01
3.34336124e-02 -4.89852488e-01 -2.77102459e-02 -9.92536247e-01
-7.92129159e-01 3.33860338e-01 1.33106661e+00 1.48353040e-01
1.43461454e+00 -4.97488916e-01 1.17930961e+00 4.90296453e-01
9.74428654e-01 -1.23973668e+00 -2.16484457e-01 -7.67655298e-02
4.79468197e-01 -9.38080013e-01 2.31315404e-01 -3.37380052e-01
-1.70524314e-01 1.24710727e+00 3.27768147e-01 2.34095976e-01
5.35277486e-01 4.92769301e-01 2.03057095e-01 1.03801303e-01
1.93068549e-01 1.42855823e-01 -1.36985302e-01 7.92246580e-01
2.79595494e-01 2.47400373e-01 -3.94139498e-01 6.31955206e-01
-9.00657922e-02 4.17967916e-01 6.00340515e-02 1.56895530e+00
-3.70380789e-01 -1.20627916e+00 -8.01160872e-01 3.09145182e-01
-5.02358019e-01 4.91387308e-01 -2.78769583e-01 8.75130951e-01
2.89558917e-01 1.34132469e+00 -1.78889692e-01 -5.82324862e-01
4.26989377e-01 2.37888575e-01 4.28996444e-01 -1.58094034e-01
-4.43606615e-01 -7.89963827e-02 -1.36856481e-01 -5.43327332e-01
-4.29545194e-01 -4.25244033e-01 -1.37502456e+00 -4.22558546e-01
-3.00238758e-01 -2.99691528e-01 1.07553685e+00 8.84721160e-01
4.45607752e-01 3.56090397e-01 5.08582592e-01 -1.00169969e+00
-6.06773555e-01 -8.40487957e-01 -5.34138381e-01 2.86225259e-01
1.52540848e-01 -6.94643021e-01 -5.72910495e-02 7.41286799e-02] | [6.532510280609131, -0.19912363588809967] |
26257370-1d9e-4855-a7d1-9c31e3184ecc | bridging-the-gap-in-multilingual-semantic | null | null | https://aclanthology.org/2020.coling-main.120 | https://aclanthology.org/2020.coling-main.120.pdf | Bridging the Gap in Multilingual Semantic Role Labeling: a Language-Agnostic Approach | Recent research indicates that taking advantage of complex syntactic features leads to favorable results in Semantic Role Labeling. Nonetheless, an analysis of the latest state-of-the-art multilingual systems reveals the difficulty of bridging the wide gap in performance between high-resource (e.g., English) and low-resource (e.g., German) settings. To overcome this issue, we propose a fully language-agnostic model that does away with morphological and syntactic features to achieve robustness across languages. Our approach outperforms the state of the art in all the languages of the CoNLL-2009 benchmark dataset, especially whenever a scarce amount of training data is available. Our objective is not to reject approaches that rely on syntax, rather to set a strong and consistent language-independent baseline for future innovations in Semantic Role Labeling. We release our model code and checkpoints at https://github.com/SapienzaNLP/multi-srl. | ['Roberto Navigli', 'Simone Conia'] | 2020-12-01 | null | null | null | coling-2020-8 | ['semantic-role-labeling'] | ['natural-language-processing'] | [-1.78949523e-03 -8.11440498e-02 -5.98958910e-01 -4.78490353e-01
-1.00587130e+00 -1.08016789e+00 8.92544448e-01 3.34141463e-01
-8.42726588e-01 9.19062197e-01 8.26183438e-01 -4.50004905e-01
2.57873051e-02 -2.58942574e-01 -4.35031325e-01 -2.33836249e-01
4.07073587e-01 4.89265591e-01 2.13523239e-01 -6.67532742e-01
1.29085734e-01 1.30706340e-01 -1.52860165e+00 3.22937876e-01
8.72331202e-01 2.03311607e-01 1.78768143e-01 1.66529372e-01
-1.81144595e-01 9.38454628e-01 -3.86783749e-01 -3.15565467e-01
1.28338397e-01 -2.97526419e-01 -1.12360907e+00 -2.41058022e-01
2.87578642e-01 1.87027231e-01 -8.39259550e-02 9.12889361e-01
5.91293871e-01 9.84289050e-02 2.26546809e-01 -1.16724145e+00
-8.19888413e-01 9.23015118e-01 -3.12965393e-01 2.79508859e-01
3.91717285e-01 1.63022391e-02 1.41452980e+00 -7.19820857e-01
1.13607407e+00 1.36125469e+00 5.44807851e-01 6.01008654e-01
-1.04236877e+00 -5.31326950e-01 5.04762471e-01 1.05643325e-01
-1.30252767e+00 -8.28997076e-01 7.21965015e-01 -3.09843391e-01
9.85945940e-01 1.13739990e-01 2.14557886e-01 1.06499279e+00
-5.43710709e-01 6.73206568e-01 1.43717933e+00 -7.71618783e-01
-1.76442817e-01 1.64882038e-02 1.49233624e-01 5.87386668e-01
4.13678765e-01 -3.71133506e-01 -6.59940600e-01 -1.37814611e-01
4.02149498e-01 -5.23485541e-01 -3.71534735e-01 -2.50833094e-01
-1.38345182e+00 6.61142111e-01 -1.96952093e-02 9.19314563e-01
-2.59852950e-02 4.59332205e-02 6.12993300e-01 4.19305414e-01
6.40043318e-01 8.83475125e-01 -8.74515295e-01 -4.24092859e-01
-8.38457406e-01 2.91926503e-01 7.70200968e-01 6.60938740e-01
4.68333781e-01 -1.23728827e-01 1.55156747e-01 1.19830525e+00
3.62360477e-01 1.37247667e-01 4.21175867e-01 -1.26363277e+00
4.92375582e-01 6.08019888e-01 2.24080876e-01 -5.04608095e-01
-3.17421257e-01 -2.67278105e-01 1.83312520e-02 -2.40410715e-01
8.83114100e-01 -5.51186912e-02 -5.83532214e-01 2.11648536e+00
3.35973054e-01 -3.46380293e-01 1.49553880e-01 9.20316696e-01
7.52459466e-01 2.80465275e-01 4.50714499e-01 1.29558280e-01
1.53875697e+00 -1.06528258e+00 -4.75897342e-01 -4.44302320e-01
9.34022427e-01 -8.25225770e-01 1.63871348e+00 1.26188099e-01
-9.43393350e-01 -2.89975166e-01 -9.47304487e-01 -4.19338673e-01
-5.15388787e-01 2.23659873e-01 9.01183307e-01 5.24245977e-01
-1.18419397e+00 3.16395164e-01 -6.75051332e-01 -8.48213375e-01
1.96920168e-02 -2.53883824e-02 -5.40270388e-01 -2.91940063e-01
-1.46300006e+00 1.05114567e+00 4.88816828e-01 -2.47932747e-01
-5.97681701e-01 -7.21745908e-01 -9.77336407e-01 -3.58887196e-01
6.60458267e-01 -5.01734316e-01 1.39348984e+00 -1.16473913e+00
-1.16530859e+00 1.37215924e+00 -3.22163194e-01 -1.98005110e-01
4.79482651e-01 -3.41192782e-01 -3.39626849e-01 7.21534416e-02
5.58410347e-01 7.17240632e-01 2.55437851e-01 -1.28406620e+00
-8.34069908e-01 -2.29997188e-01 6.12240434e-01 5.41618705e-01
-3.56398433e-01 6.71934247e-01 -4.39468294e-01 -5.44283211e-01
-1.26141727e-01 -1.03149724e+00 -1.62370235e-01 -3.00446838e-01
2.09656470e-02 -4.83409286e-01 3.29605490e-01 -8.20177972e-01
1.10357630e+00 -1.95964742e+00 -2.05263644e-02 -4.44081545e-01
-1.47128299e-01 1.25738040e-01 -8.76015499e-02 6.18792892e-01
-5.54842912e-02 5.42957723e-01 -2.91738540e-01 -3.83365154e-01
7.69960508e-03 3.64450395e-01 -3.16362716e-02 6.26817465e-01
9.27234516e-02 7.49364436e-01 -1.27027833e+00 -4.42245901e-01
6.22845776e-02 2.99054176e-01 -5.90204895e-01 -2.36238062e-01
-2.98448563e-01 7.86638796e-01 -4.02698845e-01 8.35830152e-01
1.66772828e-01 -2.68714428e-01 6.76524758e-01 1.09735742e-01
-5.37030578e-01 1.14169133e+00 -9.00828898e-01 2.19723344e+00
-6.32108927e-01 3.54384422e-01 2.47031927e-01 -8.46955419e-01
3.98929238e-01 3.91990542e-01 4.47335392e-01 -6.84865415e-01
2.86867991e-02 6.50101244e-01 3.61056805e-01 6.87898742e-03
5.29667497e-01 -7.06427321e-02 -3.79233807e-01 5.13910532e-01
1.52026908e-02 -6.82218671e-02 5.59294403e-01 1.13342501e-01
9.16500866e-01 4.04137820e-01 5.16666651e-01 -8.83332551e-01
7.85619974e-01 4.02723819e-01 7.27886498e-01 5.92950761e-01
-3.88673633e-01 4.61873680e-01 3.55968446e-01 -5.71096614e-02
-1.08994794e+00 -5.94194114e-01 -2.21495822e-01 1.59164441e+00
2.39245445e-02 -5.87646961e-01 -5.75447738e-01 -9.23001945e-01
-4.31054505e-03 7.54178345e-01 -3.50461662e-01 3.28658462e-01
-8.49893630e-01 -8.14496636e-01 8.61270130e-01 3.60385746e-01
3.79735619e-01 -1.00915217e+00 -3.73692542e-01 2.90716112e-01
-3.55919033e-01 -1.42358148e+00 -1.02844700e-01 2.53517684e-02
-4.00939792e-01 -1.14257324e+00 -4.12971258e-01 -6.45639181e-01
3.41181874e-01 2.42142737e-01 1.47613835e+00 2.29171082e-01
2.31113866e-01 4.57323134e-01 -7.18141437e-01 -3.38602304e-01
-4.59137231e-01 3.81019711e-01 -1.24611743e-01 -3.46485943e-01
3.47526044e-01 -3.27141076e-01 -3.70138466e-01 1.33970797e-01
-6.96725607e-01 -7.26086572e-02 1.67614236e-01 6.51337445e-01
5.43746710e-01 -1.70382813e-01 7.43441880e-01 -1.22112584e+00
5.82857311e-01 -5.74876547e-01 -4.01246756e-01 3.31662536e-01
-6.58191144e-01 4.37427536e-02 6.67382896e-01 6.59886748e-02
-1.23872733e+00 -2.95554966e-01 -3.69537264e-01 2.43678421e-01
-2.34879285e-01 6.94665551e-01 -3.10416967e-01 2.38454774e-01
5.89321136e-01 -9.50714275e-02 -2.82313347e-01 -7.83352554e-01
5.59749365e-01 5.85443437e-01 3.33074838e-01 -1.09833121e+00
6.23616159e-01 5.32249689e-01 -4.35273260e-01 -6.95000112e-01
-1.15881538e+00 -6.91990674e-01 -6.35784030e-01 1.82622865e-01
7.46608078e-01 -1.35859716e+00 -5.85919060e-02 4.22823340e-01
-1.02288628e+00 -6.47438169e-01 -2.95969635e-01 3.19196016e-01
-3.55476707e-01 2.53210068e-01 -7.11281657e-01 -4.07332033e-01
-2.13394240e-01 -1.03996348e+00 9.49804366e-01 1.52073652e-02
-5.50139725e-01 -1.18371749e+00 1.08374059e-01 7.81226397e-01
3.24104577e-01 9.79229361e-02 9.10992920e-01 -1.01099169e+00
-2.19927594e-01 1.53076217e-01 -1.29313096e-01 3.15053642e-01
3.21503609e-01 -1.80213451e-01 -8.59716654e-01 -2.56722987e-01
-1.61301777e-01 -6.27561569e-01 6.96564674e-01 3.25386301e-02
8.04345548e-01 -9.22422931e-02 4.20583822e-02 3.27037007e-01
1.33465040e+00 -5.56983352e-02 2.11137906e-01 8.05498779e-01
7.99706459e-01 7.73186147e-01 9.96732712e-01 -3.50947194e-02
1.21369219e+00 6.33804977e-01 4.09907028e-02 -1.08963631e-01
-3.28547835e-01 -4.39319074e-01 3.23256224e-01 8.57847869e-01
-9.51463878e-02 -2.90689379e-01 -1.31127620e+00 9.76585090e-01
-1.97935951e+00 -6.44272149e-01 -4.31704558e-02 2.03201079e+00
1.21215212e+00 1.24974318e-01 1.96548074e-01 -3.94868553e-02
6.06727719e-01 6.40818536e-01 -2.30776787e-01 -2.32100263e-01
-3.58098805e-01 1.69874296e-01 4.79570508e-01 6.71470344e-01
-1.17929542e+00 1.70432460e+00 5.72887945e+00 7.36418605e-01
-1.05828023e+00 6.46823227e-01 4.88140941e-01 4.30256426e-02
-5.06765664e-01 3.40063959e-01 -8.76150966e-01 3.71872663e-01
9.18326497e-01 -2.42060781e-01 5.14228463e-01 6.67940021e-01
1.76620841e-01 -1.15583889e-01 -8.65985394e-01 6.35661662e-01
1.53087780e-01 -1.10580432e+00 -1.98738426e-01 -2.10353687e-01
6.15544856e-01 6.00223660e-01 -2.16487363e-01 2.98250079e-01
7.18202353e-01 -9.41806257e-01 1.18132687e+00 -1.98959827e-01
9.73950803e-01 -4.17709738e-01 5.73445618e-01 3.48174721e-01
-1.09959292e+00 9.89550725e-02 -1.36543587e-01 -3.20482552e-01
1.13978595e-01 3.26515704e-01 -4.49892730e-01 6.58590019e-01
6.19209707e-01 9.50023651e-01 -6.95501387e-01 4.40870613e-01
-7.96742678e-01 8.58661771e-01 -2.47121155e-01 3.22734118e-01
3.36688310e-01 3.50134796e-03 5.35954297e-01 1.35310340e+00
-9.48493406e-02 -5.31011298e-02 4.85630125e-01 4.63073075e-01
-2.59054005e-01 5.63673675e-01 -6.43274486e-01 -4.00598973e-01
7.62110412e-01 1.25606394e+00 -8.97693157e-01 -2.49693021e-01
-7.33590543e-01 7.22876966e-01 5.65819323e-01 1.33243307e-01
-5.99665701e-01 1.15518190e-01 7.19500601e-01 2.52213627e-01
-1.78387061e-01 -5.13499260e-01 -3.16561848e-01 -1.40541708e+00
-1.86413378e-02 -1.37830687e+00 7.93340027e-01 -5.41847885e-01
-1.10989034e+00 4.16311830e-01 1.05977826e-01 -7.42942750e-01
-3.22507918e-01 -6.52146578e-01 -6.90810382e-02 9.17274356e-01
-1.97850800e+00 -1.60075963e+00 1.35838121e-01 5.56152999e-01
6.35055006e-01 1.89184025e-02 6.69162631e-01 4.66529876e-01
-3.57929677e-01 5.52643955e-01 -2.85679460e-01 2.12742433e-01
1.15433729e+00 -1.34996676e+00 4.48666841e-01 1.14114559e+00
3.25382501e-01 8.29233229e-01 7.23910630e-01 -5.64503610e-01
-1.12559307e+00 -7.78941512e-01 1.47458029e+00 -9.23835397e-01
1.05264664e+00 -5.12126744e-01 -8.23956907e-01 8.34236979e-01
2.63339788e-01 9.28729996e-02 7.26666152e-01 4.99589026e-01
-5.71976542e-01 2.78478146e-01 -9.08212364e-01 6.48047507e-01
1.50364518e+00 -8.21234941e-01 -7.20980108e-01 1.98906764e-01
7.67166317e-01 -3.15349281e-01 -8.32483709e-01 4.67821062e-01
5.52552998e-01 -7.14523554e-01 7.93316126e-01 -6.40687764e-01
4.16483164e-01 -4.45804954e-01 -4.54526842e-01 -1.31602061e+00
-2.01293945e-01 -5.18978357e-01 5.06505072e-01 1.53407896e+00
7.53743827e-01 -8.66465330e-01 3.61941218e-01 4.45399344e-01
-3.75462681e-01 -4.64842975e-01 -7.72519767e-01 -6.70667887e-01
3.83774757e-01 -3.84587675e-01 4.27412599e-01 1.41302431e+00
-1.09460182e-01 5.83950162e-01 -9.43226516e-02 -3.53336148e-02
2.08567813e-01 1.52041540e-01 5.92881203e-01 -1.26441944e+00
-2.52067477e-01 -3.08477342e-01 9.61122140e-02 -5.53523779e-01
7.17736423e-01 -1.26561475e+00 -1.25571087e-01 -1.80518460e+00
2.12363064e-01 -8.70870829e-01 -4.53388751e-01 9.72921848e-01
-2.97784090e-01 2.98275858e-01 3.38727742e-01 3.97314996e-01
-8.63624036e-01 2.35564396e-01 1.02972722e+00 2.83758610e-01
-1.99173301e-01 -6.14769876e-01 -1.41582966e+00 1.01074481e+00
9.86432016e-01 -6.02862716e-01 -3.28022957e-01 -7.53211379e-01
2.29142606e-01 -3.68374050e-01 1.56575084e-01 -6.40463114e-01
-1.09731734e-01 -3.08860064e-01 -4.79865037e-02 5.30056395e-02
4.97672856e-02 -4.73071128e-01 -1.26629323e-01 2.16072947e-01
-3.99035484e-01 1.81900337e-01 1.77961215e-01 9.24755782e-02
-3.48455191e-01 -3.14297020e-01 6.03144586e-01 -5.20464897e-01
-1.02469659e+00 -3.69036128e-03 -1.18960887e-01 6.80764794e-01
6.01641774e-01 8.22603926e-02 -8.13029587e-01 -1.79132074e-01
-2.92675614e-01 3.13903213e-01 1.03651690e+00 8.10832024e-01
-2.82650977e-01 -9.54728663e-01 -9.67853427e-01 -1.98201910e-01
4.50436473e-01 -2.47782588e-01 -1.05771244e-01 6.80942655e-01
-4.47161525e-01 7.57980764e-01 2.36330181e-02 -9.29936916e-02
-1.08653080e+00 8.88628736e-02 2.17521042e-01 -6.67955756e-01
-2.70759225e-01 7.18927145e-01 -3.97897996e-02 -8.37940633e-01
-1.85446814e-01 6.43950328e-02 -4.02337402e-01 2.22615048e-01
2.16629729e-01 5.19410782e-02 8.50704759e-02 -9.40767646e-01
-7.71763623e-01 1.46384582e-01 -1.66712143e-02 -5.53132057e-01
1.35361373e+00 -1.87851503e-01 -3.14872652e-01 6.57663047e-01
7.87770450e-01 6.78482831e-01 -9.43581581e-01 -2.28046060e-01
6.37131691e-01 -3.78124774e-01 -4.56708334e-02 -1.16146183e+00
-6.54200375e-01 4.59991992e-01 7.44025782e-02 -9.38224420e-02
8.29229414e-01 3.79775107e-01 4.93218869e-01 8.49069059e-02
6.30596817e-01 -1.28934956e+00 -2.75938421e-01 8.84043634e-01
8.91649127e-01 -1.28735197e+00 -1.98753346e-02 -4.97811317e-01
-8.28021944e-01 4.26156759e-01 5.34939349e-01 6.97682425e-02
5.10880828e-01 1.29166096e-01 4.06700969e-01 -2.33800545e-01
-7.79940724e-01 -7.36643791e-01 9.87945721e-02 3.83841634e-01
1.10575652e+00 1.81523129e-01 -1.01509356e+00 6.58646047e-01
-3.75334442e-01 -1.93811566e-01 3.01958382e-01 1.17809713e+00
-1.77403107e-01 -1.69489312e+00 -2.23756786e-02 3.89530547e-02
-1.04709113e+00 -5.75462162e-01 -5.69624007e-01 1.15676713e+00
1.45555109e-01 9.62052345e-01 -1.24516085e-01 1.44008398e-01
3.40650856e-01 4.63636011e-01 4.51983452e-01 -1.12945068e+00
-7.02905178e-01 -5.24506951e-03 8.51371706e-01 -5.35937130e-01
-9.41549599e-01 -1.02117884e+00 -1.23532796e+00 -1.08533442e-01
2.14726448e-01 1.03816085e-01 5.47014296e-01 8.94215882e-01
3.07435036e-01 2.08295301e-01 1.13174163e-01 -3.86186719e-01
-3.56301486e-01 -8.14550340e-01 -2.71818966e-01 6.20185792e-01
-1.81206921e-03 -6.54159725e-01 -2.88520277e-01 -7.15696663e-02] | [10.473494529724121, 9.606160163879395] |
bfb1b1df-01a7-4ac5-b090-6ecee2609c1e | multi-task-convolutional-neural-network-for | 1702.04710 | null | http://arxiv.org/abs/1702.04710v2 | http://arxiv.org/pdf/1702.04710v2.pdf | Multi-Task Convolutional Neural Network for Pose-Invariant Face Recognition | This paper explores multi-task learning (MTL) for face recognition. We answer
the questions of how and why MTL can improve the face recognition performance.
First, we propose a multi-task Convolutional Neural Network (CNN) for face
recognition where identity classification is the main task and pose,
illumination, and expression estimations are the side tasks. Second, we develop
a dynamic-weighting scheme to automatically assign the loss weight to each side
task, which is a crucial problem in MTL. Third, we propose a pose-directed
multi-task CNN by grouping different poses to learn pose-specific identity
features, simultaneously across all poses. Last but not least, we propose an
energy-based weight analysis method to explore how CNN-based MTL works. We
observe that the side tasks serve as regularizations to disentangle the
variations from the learnt identity features. Extensive experiments on the
entire Multi-PIE dataset demonstrate the effectiveness of the proposed
approach. To the best of our knowledge, this is the first work using all data
in Multi-PIE for face recognition. Our approach is also applicable to
in-the-wild datasets for pose-invariant face recognition and achieves
comparable or better performance than state of the art on LFW, CFP, and IJB-A
datasets. | ['Xiaoming Liu', 'Xi Yin'] | 2017-02-15 | null | null | null | null | ['robust-face-recognition'] | ['computer-vision'] | [ 1.09314092e-01 -5.23906112e-01 -1.84173912e-01 -6.89652622e-01
-9.06596541e-01 -4.73527461e-01 4.32954699e-01 -6.65561974e-01
-3.30260187e-01 5.63275397e-01 -2.57445201e-02 3.02992105e-01
-2.50653118e-01 -2.33259469e-01 -8.19565713e-01 -9.88663614e-01
2.02303737e-01 3.81215751e-01 -3.37025642e-01 -1.34131804e-01
1.45776300e-02 9.11327600e-01 -1.56915414e+00 2.65808225e-01
4.77631956e-01 1.32785547e+00 -2.03203797e-01 2.46968329e-01
2.00649068e-01 5.03195226e-01 -4.15309370e-01 -7.08760619e-01
3.97870153e-01 -3.43533158e-01 -6.49917960e-01 9.69504043e-02
1.13463759e+00 -9.38749909e-02 -6.20886832e-02 9.31163073e-01
8.01173389e-01 1.74370352e-02 5.88925660e-01 -1.52350950e+00
-5.48578262e-01 -7.93955922e-02 -9.39621329e-01 -5.50562800e-05
1.82847694e-01 -7.53808767e-02 8.99696112e-01 -1.32260382e+00
4.85031009e-01 1.48196781e+00 7.26856649e-01 8.43153715e-01
-1.19895720e+00 -1.02212524e+00 1.96622849e-01 3.80950600e-01
-1.67667627e+00 -8.34712744e-01 9.53132629e-01 -3.17258626e-01
5.33474505e-01 2.54769355e-01 1.11609131e-01 1.47097564e+00
2.34699678e-02 8.59955430e-01 1.06193340e+00 -3.64056557e-01
-1.81263432e-01 -1.57438278e-01 -1.62669882e-01 9.74399149e-01
7.76273236e-02 -6.64959475e-02 -9.32552040e-01 -2.35199794e-01
5.39540887e-01 -9.36359242e-02 -2.87043661e-01 -3.75704199e-01
-7.69666135e-01 6.76014006e-01 6.38570860e-02 1.65918306e-01
-2.04925835e-01 2.34389663e-01 4.99338329e-01 2.93019652e-01
8.33934307e-01 6.14622571e-02 -6.75956011e-01 1.16373412e-01
-8.52315068e-01 1.55603930e-01 7.07426190e-01 7.51613855e-01
7.84640133e-01 1.58081874e-01 -5.11561573e-01 1.25600326e+00
4.78473455e-01 5.11375189e-01 1.82863101e-01 -7.27439165e-01
2.82541335e-01 3.36800307e-01 -1.68185472e-01 -9.83821750e-01
-3.78591239e-01 -5.63283920e-01 -8.37190330e-01 1.40426710e-01
3.97100270e-01 -3.28380913e-01 -8.24577630e-01 2.14449215e+00
2.26117700e-01 4.76412445e-01 -2.83402026e-01 8.81364107e-01
8.14035594e-01 7.72896409e-02 1.38334721e-01 -1.61265850e-01
1.52566385e+00 -9.38076496e-01 -6.81204498e-01 -2.13546053e-01
1.97139129e-01 -7.74238765e-01 5.64000905e-01 2.51657784e-01
-7.42711246e-01 -7.02180982e-01 -8.87542963e-01 4.86720800e-02
-1.30396917e-01 7.28672683e-01 6.45979166e-01 8.44869792e-01
-1.09931409e+00 3.80915076e-01 -6.33611441e-01 -2.54828542e-01
8.64011049e-01 4.98063266e-01 -5.16619682e-01 -2.12538913e-01
-9.90756094e-01 7.85656810e-01 -2.16178149e-02 4.26169664e-01
-1.05383873e+00 -6.77686095e-01 -9.01597440e-01 -1.09006137e-01
4.71733302e-01 -4.67543155e-01 8.30050409e-01 -1.17535937e+00
-1.44857836e+00 9.60180700e-01 -4.64608252e-01 1.47959560e-01
2.47715801e-01 -2.56813318e-01 -4.93559539e-01 2.74766572e-02
4.93400618e-02 5.24452150e-01 1.14035213e+00 -1.37959349e+00
-1.49058655e-01 -8.44533920e-01 -1.82600722e-01 -2.08907090e-02
-6.49371266e-01 3.25189292e-01 -8.09501946e-01 -6.86910927e-01
-1.37124583e-01 -9.70655084e-01 2.51432747e-01 4.24851626e-01
-2.75652736e-01 -4.33401674e-01 9.25823152e-01 -6.34716511e-01
7.93183208e-01 -2.28011417e+00 4.11062658e-01 3.02709211e-02
2.61870511e-02 3.16291481e-01 -5.48603773e-01 6.75289184e-02
-1.98237598e-01 -7.57227512e-03 -3.52803022e-02 -7.63304949e-01
6.16875440e-02 1.81982517e-01 8.35792068e-03 6.42567039e-01
4.68090832e-01 1.06494784e+00 -3.97730410e-01 -4.58022654e-01
-1.84671551e-01 9.09947336e-01 -5.73366582e-01 9.63809565e-02
-3.95974666e-02 5.82405925e-01 -5.02888858e-01 9.99192297e-01
9.98841524e-01 -6.48991838e-02 9.62293223e-02 -8.05359364e-01
9.44381207e-02 -3.66993099e-01 -1.00143445e+00 1.77984595e+00
-4.61898386e-01 4.27528232e-01 3.63005787e-01 -1.12944067e+00
9.10529733e-01 4.06271398e-01 5.80871820e-01 -5.31934738e-01
1.72213763e-01 2.82045811e-01 -9.59013551e-02 -4.49512452e-01
-8.32412113e-03 -2.20317170e-01 1.15065180e-01 3.55651557e-01
5.80647707e-01 3.53018045e-01 -2.19383482e-02 -4.40618902e-01
6.17721379e-01 3.94303262e-01 2.16378450e-01 -2.96496838e-01
8.12067509e-01 -6.94380581e-01 9.80875850e-01 4.58393097e-01
-4.51132685e-01 4.91024137e-01 4.26273823e-01 -4.78931606e-01
-5.61627269e-01 -6.93220377e-01 -2.74082094e-01 1.45446777e+00
-1.68725237e-01 -2.51138527e-02 -6.40353262e-01 -8.02113652e-01
1.64570391e-01 1.33227035e-01 -8.41715157e-01 -4.43714932e-02
-5.99636853e-01 -1.15501916e+00 7.32179105e-01 5.55023730e-01
7.23555624e-01 -8.35207641e-01 1.27699614e-01 -1.10698454e-01
-1.17693588e-01 -1.23752904e+00 -8.18499267e-01 9.48046148e-03
-4.58254725e-01 -1.10524631e+00 -7.13927329e-01 -8.66969407e-01
7.28988230e-01 -4.99425530e-02 9.02922988e-01 -4.24804688e-02
-5.01064181e-01 4.29902673e-01 -1.01888239e-01 -3.34355623e-01
3.32960695e-01 -7.73200393e-02 1.23608708e-01 7.66949773e-01
3.48630726e-01 -5.97023010e-01 -6.01849198e-01 4.71412838e-01
-5.18815458e-01 -3.35711509e-01 6.18300617e-01 1.05682516e+00
5.49312472e-01 -2.65234083e-01 6.57077014e-01 -7.29879379e-01
3.93717706e-01 -2.24056900e-01 -3.65868092e-01 5.51898360e-01
-1.83458447e-01 1.84970479e-02 4.77559388e-01 -4.28518295e-01
-1.16407919e+00 3.69652063e-01 -3.03006202e-01 -6.35605872e-01
-1.21481359e-01 1.09170616e-01 -5.83078623e-01 -7.67043829e-01
3.49281400e-01 2.45748252e-01 4.77219559e-03 -5.33713222e-01
1.69301748e-01 3.60954463e-01 2.39797577e-01 -8.13917339e-01
8.53012502e-01 6.22543752e-01 3.09217542e-01 -8.49178493e-01
-1.22820818e+00 -4.01549041e-01 -7.42476523e-01 -2.80472785e-01
8.64856780e-01 -1.03586435e+00 -1.24127293e+00 7.53713191e-01
-1.15336251e+00 -1.07019022e-02 2.41594866e-01 1.89896122e-01
-3.52395266e-01 3.04345846e-01 -5.92633665e-01 -8.31897855e-01
-3.99256110e-01 -1.18430495e+00 1.38627636e+00 2.34411821e-01
2.24710092e-01 -1.00732112e+00 -8.49623233e-02 7.03150511e-01
4.22088981e-01 2.70794958e-01 5.32304227e-01 -7.36424267e-01
-4.04726058e-01 1.09322831e-01 -4.60528344e-01 3.23948056e-01
2.10326150e-01 -1.76329911e-01 -1.41044760e+00 -6.24942422e-01
7.21423626e-02 -8.03181887e-01 1.08728778e+00 1.81644261e-01
1.59551859e+00 -2.44208470e-01 -2.05089852e-01 1.04576433e+00
1.41977286e+00 6.79932386e-02 4.64155465e-01 -6.81311637e-02
9.03437674e-01 4.65902865e-01 4.01073664e-01 4.20659423e-01
3.43262523e-01 1.09874988e+00 2.69009292e-01 -1.73327789e-01
-1.92483962e-01 7.55848885e-02 4.04998571e-01 5.66969812e-01
-2.72780210e-01 -1.17102690e-01 -6.31576836e-01 2.22323090e-01
-1.67573690e+00 -9.54235435e-01 2.72536963e-01 1.89531624e+00
7.88706303e-01 -4.53635842e-01 1.41800120e-02 -2.78384954e-01
6.25247240e-01 4.47292715e-01 -6.18069470e-01 -7.24977180e-02
-2.10202366e-01 6.47459924e-01 1.97835311e-01 2.87149787e-01
-1.43866515e+00 8.61353457e-01 5.49324226e+00 1.15024853e+00
-1.31042254e+00 4.15941834e-01 7.82822192e-01 -1.87026799e-01
1.84574530e-01 -4.55202788e-01 -1.09981728e+00 2.10224763e-01
4.67927247e-01 1.39065757e-01 4.87378627e-01 8.38125169e-01
-7.26097077e-02 3.18131328e-01 -1.12015283e+00 1.12694228e+00
8.23274493e-01 -7.87397087e-01 4.46384475e-02 -9.07903090e-02
6.59592092e-01 -1.64031535e-01 3.00605357e-01 2.99304873e-01
-8.78992900e-02 -1.11679220e+00 4.97069061e-01 4.87924099e-01
9.52065706e-01 -7.97520697e-01 5.67288697e-01 -7.10144937e-02
-1.42478323e+00 -9.45391431e-02 -2.42442980e-01 3.29544395e-01
-1.42529249e-01 4.10657346e-01 -3.15779626e-01 7.12265790e-01
4.28660691e-01 7.22331941e-01 -7.23865747e-01 7.55317926e-01
-1.73013315e-01 3.72345626e-01 3.19132805e-02 1.95514128e-01
-5.58839701e-02 -6.01854473e-02 3.87208223e-01 1.13633394e+00
2.86453664e-01 -2.13365212e-01 3.09116483e-01 7.85806894e-01
-5.04747331e-01 8.28714371e-02 -3.06356311e-01 4.71531786e-02
2.32704014e-01 1.51947331e+00 -3.13379228e-01 1.83007374e-01
-5.56756675e-01 1.26104736e+00 5.71952105e-01 5.25181413e-01
-8.07666242e-01 -1.06874220e-01 1.03023291e+00 -3.96052718e-01
4.37164873e-01 -1.69813931e-01 -1.08333506e-01 -1.14825535e+00
2.69653559e-01 -9.33814287e-01 3.38173479e-01 -2.41940901e-01
-1.54377735e+00 5.76323032e-01 -2.78200030e-01 -9.54093933e-01
5.46909273e-02 -9.54863906e-01 -5.61853945e-01 9.94282007e-01
-1.85392284e+00 -1.70138848e+00 -2.53871262e-01 7.38812566e-01
5.44209123e-01 -5.03703177e-01 1.00353467e+00 7.17264295e-01
-9.89940166e-01 1.08441401e+00 -5.02579659e-03 4.25168097e-01
9.91029382e-01 -8.07422042e-01 1.18934616e-01 6.96158171e-01
2.51010090e-01 6.19859397e-01 2.06746802e-01 -4.05335546e-01
-1.68458056e+00 -1.24518001e+00 7.25133419e-01 -3.76742870e-01
3.62301469e-01 -6.71771049e-01 -8.49333405e-01 5.57030022e-01
-6.92832395e-02 4.53524619e-01 1.01221597e+00 3.37790132e-01
-7.26190865e-01 -4.80813712e-01 -1.09195006e+00 2.80624747e-01
1.12334538e+00 -8.16801488e-01 -1.80033389e-02 3.99849445e-01
2.50492007e-01 -3.13272595e-01 -8.24514508e-01 6.51673138e-01
8.97723258e-01 -7.16280878e-01 1.15934527e+00 -5.61388850e-01
3.39883804e-01 -2.13943779e-01 -1.84797749e-01 -1.32436371e+00
-3.59390199e-01 -3.71233493e-01 3.60395014e-02 1.44446790e+00
3.37908775e-01 -6.43159688e-01 6.56154454e-01 2.31302902e-01
1.56220049e-01 -9.27258730e-01 -1.26938927e+00 -8.23299229e-01
1.67674012e-02 -1.98789150e-01 4.05532509e-01 9.58233297e-01
-7.01830029e-01 9.16644484e-02 -7.56188691e-01 1.90124273e-01
1.04054356e+00 2.20175728e-01 4.47712302e-01 -1.23982203e+00
-2.16789722e-01 -3.05910975e-01 -3.04955155e-01 -7.40370691e-01
7.53005922e-01 -8.99965942e-01 -6.22678846e-02 -9.29786026e-01
6.00688159e-01 -1.96638897e-01 -4.46668416e-01 8.64992917e-01
-2.54978806e-01 6.96772993e-01 2.50392109e-01 3.33707444e-02
-6.42768204e-01 8.09574425e-01 1.38743448e+00 -1.95808858e-01
2.91045606e-01 -6.14587441e-02 -8.52537274e-01 5.55266857e-01
6.61467075e-01 -2.54526138e-01 -2.77228683e-01 -4.97134775e-01
-1.40821338e-01 -1.97673962e-01 5.63193619e-01 -7.81777620e-01
3.23659003e-01 -1.64067417e-01 7.19605744e-01 -1.50199950e-01
7.36727118e-01 -7.20746756e-01 -6.69919373e-03 2.08990887e-01
-1.08400248e-01 -2.92935550e-01 5.57760298e-01 3.91912729e-01
-1.60171807e-01 1.82910502e-01 9.74718928e-01 1.62229225e-01
-6.86619341e-01 8.00370753e-01 1.81186810e-01 1.30807415e-01
8.55378151e-01 4.48506698e-02 -2.08762407e-01 -9.06035230e-02
-5.54383755e-01 2.25827396e-01 5.86169772e-02 6.56388819e-01
5.12383342e-01 -1.64478862e+00 -9.73919272e-01 4.56210524e-01
3.45145375e-01 -5.37092984e-01 3.47089142e-01 8.38705659e-01
3.62210982e-02 4.01147813e-01 -3.95854950e-01 -4.54197168e-01
-1.61703944e+00 1.75665379e-01 5.52133560e-01 -7.27529079e-02
7.77372196e-02 1.25898278e+00 3.39235276e-01 -5.26287317e-01
2.31931090e-01 4.37059432e-01 -1.92982852e-01 3.91200125e-01
5.25743127e-01 1.12903774e-01 9.42974538e-02 -9.69766080e-01
-7.39462972e-01 1.00763452e+00 -2.48499751e-01 1.07972786e-01
1.51641846e+00 1.05407074e-01 -3.43480825e-01 1.96018070e-01
1.53816521e+00 -1.92199886e-01 -1.30594528e+00 -2.01702252e-01
-6.62221462e-02 -5.58560848e-01 -5.29034734e-02 -8.53697240e-01
-1.44902170e+00 8.63744915e-01 6.85103238e-01 -7.01810062e-01
1.39342785e+00 -4.74490225e-02 5.05039394e-01 2.62630105e-01
4.28800136e-01 -1.09127021e+00 3.79668713e-01 4.76848006e-01
1.26990950e+00 -1.45544565e+00 -7.67545998e-02 -5.56308866e-01
-3.82015973e-01 1.10162830e+00 9.40222561e-01 3.20293158e-01
7.20050395e-01 1.88649341e-01 -5.64334029e-03 -2.77594954e-01
-5.53258836e-01 -2.47469097e-01 6.71425462e-01 3.97323728e-01
6.33759081e-01 -7.44251758e-02 -1.73538715e-01 5.53202271e-01
1.79167792e-01 1.89057346e-02 -3.13007146e-01 6.89567804e-01
-3.17918882e-02 -1.44308841e+00 -3.81125748e-01 2.92751759e-01
-5.50517678e-01 1.06831364e-01 -5.57422996e-01 5.47481477e-01
3.80475491e-01 7.72334516e-01 -2.29799271e-01 -4.00783449e-01
3.11850458e-01 4.51319098e-01 8.49257827e-01 -4.64601219e-01
-5.41153371e-01 -8.09857473e-02 -2.88741682e-02 -7.09447205e-01
-4.87501740e-01 -7.40867138e-01 -5.05374670e-01 -4.16130908e-02
-1.34672895e-01 -2.41963357e-01 6.71711683e-01 9.46382642e-01
4.49790806e-01 3.96562994e-01 9.27698731e-01 -9.60288763e-01
-5.48129618e-01 -9.70035493e-01 -5.10371208e-01 5.54238439e-01
3.87432724e-01 -9.99547064e-01 -3.68907034e-01 -1.52365044e-01] | [13.296011924743652, 0.6888681650161743] |
de124c9e-ac7d-4dad-bea8-5a24ec85163a | enhancements-to-the-boun-treebank-reflecting | 2207.11782 | null | https://arxiv.org/abs/2207.11782v1 | https://arxiv.org/pdf/2207.11782v1.pdf | Enhancements to the BOUN Treebank Reflecting the Agglutinative Nature of Turkish | In this study, we aim to offer linguistically motivated solutions to resolve the issues of the lack of representation of null morphemes, highly productive derivational processes, and syncretic morphemes of Turkish in the BOUN Treebank without diverging from the Universal Dependencies framework. In order to tackle these issues, new annotation conventions were introduced by splitting certain lemmas and employing the MISC (miscellaneous) tab in the UD framework to denote derivation. Representational capabilities of the re-annotated treebank were tested on a LSTM-based dependency parser and an updated version of the BoAT Tool is introduced. | ['Balkız Öztürk', 'Tunga Güngör', 'Arzucan Özgür', 'Suzan Üsküdarlı', 'Şaziye Betül Özateş', 'Onur Güngör', 'Merve Gürbüz', 'Muhammet Şen', 'Salih Furkan Akkurt', 'Büşra Marşan'] | 2022-07-24 | null | null | null | null | ['miscellaneous'] | ['miscellaneous'] | [-1.53756425e-01 4.33820218e-01 1.26053259e-01 -4.13936257e-01
-3.65295529e-01 -5.32044113e-01 3.00333261e-01 2.53114671e-01
-6.98244333e-01 1.03693593e+00 3.33654076e-01 -8.33082139e-01
-1.65289730e-01 -5.27090073e-01 -1.34437025e-01 -3.75715584e-01
9.29721892e-02 1.83478311e-01 2.45414585e-01 -3.18701327e-01
5.85463829e-02 5.13534248e-01 -9.35225964e-01 3.77879411e-01
9.55912948e-01 2.50459582e-01 2.98163801e-01 2.98793346e-01
-2.99348265e-01 7.37999022e-01 -8.49538386e-01 -8.48779500e-01
-5.41869588e-02 -6.08688116e-01 -1.00215125e+00 -2.78830081e-01
1.09711602e-01 -1.34180591e-01 -4.64422330e-02 1.07207358e+00
1.28913775e-01 7.30842426e-02 3.95086110e-01 -5.40586829e-01
-8.12732875e-01 1.40546310e+00 -9.97565761e-02 6.23198628e-01
1.75943729e-02 -4.77164268e-01 1.08576405e+00 -4.79433745e-01
9.46205437e-01 1.34219611e+00 3.20210129e-01 6.16525114e-01
-1.10293877e+00 -2.96563119e-01 2.13893473e-01 1.56110093e-01
-1.06167829e+00 -5.06862938e-01 5.06321669e-01 -1.50642931e-01
1.78549731e+00 1.41078001e-02 4.37968910e-01 1.11055708e+00
3.91528964e-01 4.35738832e-01 9.35293198e-01 -9.85558808e-01
-1.73094515e-02 2.70696282e-01 4.33275014e-01 5.29920638e-01
6.10906303e-01 2.56215986e-02 -2.75340408e-01 2.77629673e-01
8.50063384e-01 -7.47273088e-01 -1.21650480e-01 1.54805958e-01
-1.01443398e+00 8.90434563e-01 2.24648416e-01 1.21357501e+00
-1.77997798e-01 2.28513088e-02 8.41916263e-01 3.61765653e-01
3.45968336e-01 4.88846928e-01 -9.15393829e-01 -4.26586829e-02
-3.54241282e-01 -3.10778528e-01 8.34855437e-01 6.68859541e-01
4.14573908e-01 7.17163324e-01 2.28005782e-01 8.66950750e-01
1.83793321e-01 6.36591390e-02 8.47147584e-01 -9.40504432e-01
3.01785707e-01 3.49445760e-01 -1.12108707e-01 -5.57750106e-01
-2.84141839e-01 -1.56993553e-01 -1.87646449e-01 -1.51467636e-01
5.78481615e-01 -6.39916956e-01 -9.18682456e-01 1.82968199e+00
2.06786737e-01 -5.56194961e-01 5.30511320e-01 3.94503057e-01
6.41412020e-01 6.14476264e-01 5.08548439e-01 -4.54239011e-01
1.02427125e+00 -6.68899059e-01 -1.18583310e+00 -2.01984569e-01
1.24472332e+00 -5.39073229e-01 8.59644413e-01 1.30007312e-01
-1.24447334e+00 -4.39839214e-01 -8.97700548e-01 -1.45182669e-01
-7.24013150e-01 7.88535625e-02 8.07295382e-01 8.79350364e-01
-1.06244051e+00 6.59818411e-01 -9.29159820e-01 -4.41473693e-01
-2.84600765e-01 7.55723044e-02 -6.64100111e-01 1.37180269e-01
-1.49190116e+00 1.56501102e+00 1.11502099e+00 6.12957835e-01
-6.44617006e-02 -1.37297884e-01 -1.32850111e+00 -2.52056360e-01
1.67001054e-01 -1.50552437e-01 1.22857130e+00 -1.21447587e+00
-1.43157315e+00 9.03177559e-01 1.29940152e-01 -5.38281620e-01
8.68038312e-02 -4.24877793e-01 -7.24245131e-01 -1.35703519e-01
1.00910693e-01 1.14266075e-01 2.55807102e-01 -7.60967553e-01
-9.36365962e-01 -4.35712278e-01 3.76754552e-02 1.03216588e-01
-1.41576260e-01 6.01468563e-01 -9.46211964e-02 -6.03170335e-01
3.12826745e-02 -6.43862605e-01 -9.95007232e-02 -9.48511899e-01
-1.97361514e-01 -4.70521390e-01 5.68226218e-01 -1.13887417e+00
1.57918322e+00 -2.02488589e+00 2.63974249e-01 -2.13012323e-01
-4.13104773e-01 5.43431222e-01 -8.43752399e-02 4.88206923e-01
-2.45211512e-01 4.22272086e-01 -2.36713558e-01 -4.37386446e-02
-9.34281647e-02 1.06707740e+00 7.87965655e-02 3.64499092e-01
3.73711139e-01 8.55949581e-01 -8.82858276e-01 -3.71012688e-01
3.20735633e-01 4.26116943e-01 -1.17580742e-01 -2.47217178e-01
-1.25085518e-01 2.68085718e-01 -6.60409033e-02 5.62697530e-01
2.41001815e-01 7.22627878e-01 8.70748699e-01 5.83018996e-02
-5.93592584e-01 1.07732785e+00 -9.85148489e-01 1.64295828e+00
-5.57767093e-01 3.71369362e-01 2.13937894e-01 -9.22200382e-01
6.44311249e-01 4.87295955e-01 -3.19999635e-01 -7.49854863e-01
3.23783547e-01 5.63160062e-01 7.49738753e-01 -3.80343765e-01
5.80603182e-01 -3.88318151e-01 -4.54392523e-01 7.42771849e-02
4.93649483e-01 4.19892222e-02 5.98724723e-01 -8.64406079e-02
7.75924563e-01 6.00399852e-01 8.04084957e-01 -5.28497517e-01
5.96685112e-01 -1.97174046e-02 9.03309941e-01 1.44198373e-01
-1.38804913e-01 -3.53724547e-02 6.63994014e-01 -2.82851756e-01
-6.78326011e-01 -9.88204300e-01 -3.88523668e-01 1.13551581e+00
-7.06718028e-01 -4.80076700e-01 -8.37117076e-01 -8.12587798e-01
-4.12669152e-01 1.36481571e+00 -5.59125602e-01 2.29591399e-01
-1.30731666e+00 -5.67340612e-01 8.06454420e-01 6.55689955e-01
6.02490865e-02 -1.52042711e+00 -9.88117099e-01 7.65011191e-01
1.63967498e-02 -1.28952003e+00 2.57887155e-01 7.49635279e-01
-7.55248785e-01 -1.04196656e+00 -1.64568633e-01 -9.30543959e-01
2.45842278e-01 -4.00110662e-01 8.34661305e-01 1.10174790e-01
1.57744884e-01 -2.37497687e-01 -6.93038821e-01 -4.06063944e-01
-8.03706288e-01 9.06356871e-02 -2.49826774e-01 -7.31969893e-01
5.05605102e-01 -3.96160632e-01 4.90089506e-01 -3.90078515e-01
-8.51633608e-01 -3.92211735e-01 2.43309692e-01 7.03032374e-01
1.54749617e-01 -1.28944054e-01 6.98158026e-01 -1.26699388e+00
6.31473184e-01 -3.16250890e-01 -5.48675478e-01 2.75110751e-02
-3.31894070e-01 2.56752700e-01 5.87432265e-01 -1.41093776e-01
-1.54358745e+00 -2.45435327e-01 -5.34630001e-01 1.42182916e-01
-4.01927203e-01 8.77346933e-01 -5.45813143e-01 4.42654580e-01
3.82137001e-01 -1.15718618e-01 -4.65355098e-01 -6.48202598e-01
5.52642345e-01 4.51139450e-01 6.25818193e-01 -7.90606737e-01
3.88681740e-01 -1.73832119e-01 -1.84591234e-01 -9.72987115e-01
-6.64089501e-01 3.32432650e-02 -1.32553196e+00 2.55315304e-01
9.15609956e-01 -7.14466095e-01 5.99304102e-02 4.39794511e-01
-1.73927176e+00 -4.38988090e-01 -4.49949414e-01 5.00049531e-01
-1.90938503e-01 6.17456734e-01 -1.29976881e+00 -6.72649682e-01
-2.28608012e-01 -8.75982463e-01 1.89060807e-01 -1.32029522e-02
-4.87353981e-01 -1.35707092e+00 5.35863787e-02 -1.62437528e-01
2.87580788e-01 4.30389762e-01 1.50115442e+00 -9.91618574e-01
1.62035108e-01 1.93347409e-01 -5.39989620e-02 6.47242010e-01
4.22479063e-01 2.54644006e-01 -7.13848531e-01 1.60477206e-01
1.18292347e-01 7.21439943e-02 6.21050656e-01 6.94708386e-03
-5.96831366e-02 -2.60608673e-01 4.73852493e-02 2.91897237e-01
1.50662005e+00 6.10058725e-01 4.49279785e-01 4.84149069e-01
5.23195386e-01 1.04855597e+00 4.04619217e-01 1.11948829e-02
4.46580529e-01 2.16240257e-01 1.86988860e-01 4.32182461e-01
-1.69291213e-01 -4.37813960e-02 6.62538052e-01 1.32650352e+00
3.63935456e-02 -2.00142369e-01 -9.16925430e-01 8.89897048e-01
-1.59276044e+00 -3.62855554e-01 -9.07478392e-01 1.92651463e+00
8.52051139e-01 4.84940916e-01 -3.19807708e-01 1.95403129e-01
7.83310056e-01 1.39028147e-01 1.99975356e-01 -1.30695546e+00
-2.06176654e-01 7.17192888e-01 2.67580390e-01 7.24689424e-01
-7.27115750e-01 1.74146593e+00 6.72461128e+00 3.87230486e-01
-1.14128864e+00 3.76888692e-01 -2.53731996e-01 3.55739802e-01
-2.40439862e-01 2.07281485e-01 -8.66335213e-01 2.72482663e-01
1.59481740e+00 -3.80149670e-02 -6.13489933e-02 4.99530226e-01
5.71092553e-02 -2.95767158e-01 -1.05005121e+00 2.38229871e-01
1.31562263e-01 -9.61000025e-01 2.14552417e-01 -6.46124259e-02
8.92894566e-02 1.49162456e-01 -4.89294827e-01 4.33834881e-01
5.01948237e-01 -6.91549838e-01 8.99011970e-01 -2.68367946e-01
5.84104002e-01 -7.99873114e-01 9.00233448e-01 1.17129333e-01
-1.05931628e+00 1.55726150e-01 -5.45890212e-01 -3.63338768e-01
4.65173483e-01 4.38443869e-01 -5.59116602e-01 8.39631319e-01
3.92980009e-01 4.07793850e-01 -4.95576888e-01 5.86554050e-01
-9.95474696e-01 6.86900616e-01 -2.43447453e-01 1.08179778e-01
5.09521902e-01 -4.14047956e-01 6.14775002e-01 1.70883012e+00
1.95928708e-01 -1.78084254e-01 -2.65986264e-01 4.94979262e-01
2.49652654e-01 2.92613298e-01 -5.99444628e-01 -8.50414485e-02
3.98682564e-01 1.19494700e+00 -9.49861348e-01 -1.53424829e-01
-5.41404247e-01 1.12272632e+00 7.01731384e-01 1.83022097e-01
-6.49722219e-01 -3.87726247e-01 4.68936622e-01 -5.23078367e-02
4.49667722e-01 -6.11129761e-01 -1.42827537e-03 -9.65199709e-01
-9.98042598e-02 -4.37243372e-01 5.62575698e-01 -5.44057846e-01
-9.51652527e-01 1.01362848e+00 2.77120143e-01 -1.59170181e-01
-6.50356412e-01 -9.36503112e-01 -6.46764219e-01 1.00133801e+00
-1.59710073e+00 -1.18806660e+00 6.92729473e-01 3.85023683e-01
5.75434625e-01 2.10167721e-01 1.22324550e+00 1.87892467e-01
-9.51341331e-01 2.40615636e-01 -1.87743127e-01 4.26375836e-01
5.28199971e-01 -1.44927883e+00 6.09707952e-01 1.27880037e+00
4.96658571e-02 7.41436005e-01 6.49769366e-01 -6.47230446e-01
-7.33279586e-01 -8.46130788e-01 1.60759437e+00 -2.58504719e-01
1.05929577e+00 -2.79144913e-01 -1.23821366e+00 1.25732124e+00
7.36951649e-01 -1.76959217e-01 8.30603838e-01 1.83962598e-01
-3.44006032e-01 2.78738737e-01 -9.42274451e-01 5.98540783e-01
8.52838695e-01 -2.97565818e-01 -1.44289041e+00 -2.10498840e-01
8.01758528e-01 -1.82863981e-01 -8.51927936e-01 2.06024334e-01
2.32445300e-01 -8.67793024e-01 1.91131473e-01 -6.86320126e-01
-2.62917429e-02 8.73218477e-02 -2.91415483e-01 -1.53426957e+00
-2.51660556e-01 -6.17881536e-01 9.04951617e-02 1.62381685e+00
6.16884649e-01 -7.58539379e-01 1.64817795e-01 6.09184265e-01
-6.67405188e-01 -1.57443851e-01 -1.33561039e+00 -7.12448955e-01
5.35414636e-01 -3.68848115e-01 1.14666916e-01 1.00392199e+00
5.11148632e-01 7.28387535e-01 5.04345894e-02 -2.35726044e-01
1.90250948e-01 -4.33468044e-01 5.40104285e-02 -1.46120942e+00
-1.95018083e-01 -3.31246734e-01 -2.16236841e-02 -1.71660781e-01
7.35932767e-01 -9.19852078e-01 3.69967967e-02 -1.60647857e+00
-6.30775511e-01 -2.89733946e-01 -3.76680255e-01 6.13079250e-01
1.86600015e-01 -2.41516858e-01 4.25883830e-01 -1.65217459e-01
-6.77556396e-02 2.53895402e-01 6.18402302e-01 4.97771591e-01
-4.11603123e-01 -6.20052338e-01 -7.57972658e-01 1.12338924e+00
8.50454509e-01 -6.93429887e-01 1.27154198e-02 -8.97575855e-01
2.04379365e-01 1.77307308e-01 -2.30242610e-01 -6.68810189e-01
-5.27346134e-02 -3.13010141e-02 1.81017697e-01 -2.74627298e-01
8.81997123e-02 -5.84735513e-01 7.09068552e-02 6.55541837e-01
5.18489964e-02 3.28310937e-01 6.81652844e-01 -1.82697475e-01
-2.08616257e-01 -7.80101299e-01 7.72914946e-01 -4.74905580e-01
-8.19732904e-01 -5.07699490e-01 -1.00323164e+00 -4.65374789e-04
9.31566894e-01 -3.25775087e-01 -2.46520773e-01 3.51893842e-01
-8.25190902e-01 -1.70335829e-01 2.42070511e-01 3.91622603e-01
2.23883480e-01 -8.17963243e-01 -6.39195979e-01 2.01147512e-01
-2.27287933e-01 -1.82449430e-01 4.41568941e-02 4.62961495e-01
-6.03198528e-01 7.04384446e-01 -6.38366222e-01 3.34457487e-01
-9.81674790e-01 4.29623306e-01 3.64588559e-01 -5.46902835e-01
-6.15106761e-01 5.77377439e-01 -2.19835252e-01 -5.77495456e-01
-2.04524517e-01 -6.08829379e-01 -4.96047229e-01 3.58480871e-01
1.15530089e-01 3.02355677e-01 4.31284875e-01 -1.12870812e+00
-4.44839150e-01 -1.35004878e-01 -1.16837360e-01 -3.04495960e-01
1.20350742e+00 -2.18501836e-01 -4.83506531e-01 7.50550628e-01
7.23761678e-01 5.45776248e-01 -4.33854431e-01 1.15819648e-01
5.44682562e-01 2.78917640e-01 -3.50509882e-02 -9.96443033e-01
-6.30141675e-01 8.18811834e-01 -1.50822535e-01 2.34584138e-01
5.53161502e-01 -1.52269363e-01 7.98792899e-01 2.91601926e-01
2.25193620e-01 -1.14990449e+00 -7.58222163e-01 1.17847538e+00
6.80129886e-01 -2.98942119e-01 -5.01612127e-01 -5.58922470e-01
-5.67809165e-01 1.35959506e+00 7.18737662e-01 -1.99501231e-01
4.59702104e-01 4.84593481e-01 3.32490981e-01 -9.61802825e-02
-6.30727768e-01 -4.76592094e-01 -3.02809179e-01 7.93991327e-01
1.07696080e+00 1.51453570e-01 -1.31501245e+00 7.07876861e-01
-3.48485678e-01 -3.75642061e-01 9.11123514e-01 9.99472320e-01
-2.50390172e-01 -1.44268262e+00 -2.05166236e-01 -4.44022156e-02
-1.13138044e+00 -4.57687378e-01 -3.92549455e-01 1.47345853e+00
3.76800895e-01 8.73096824e-01 2.99565822e-01 4.23559286e-02
3.60288262e-01 4.84553665e-01 7.25656688e-01 -1.03402114e+00
-9.74880219e-01 4.26508904e-01 7.25258529e-01 -2.58512259e-01
-6.44535959e-01 -5.89755893e-01 -1.44485545e+00 2.12760404e-01
-4.15795058e-01 5.10682583e-01 7.18090475e-01 1.23358500e+00
-1.33397266e-01 6.45145953e-01 -2.07183570e-01 -5.88232338e-01
-2.53446370e-01 -1.25271308e+00 -6.68409884e-01 -4.67267260e-02
-1.60428643e-01 -3.11181068e-01 -2.89579153e-01 -9.77913663e-02] | [10.398140907287598, 10.037576675415039] |
3f802eb3-5b13-42ff-ad04-b95419764984 | semantic-guided-single-image-reflection | 1907.11912 | null | https://arxiv.org/abs/1907.11912v3 | https://arxiv.org/pdf/1907.11912v3.pdf | Semantic Guided Single Image Reflection Removal | Reflection is common in images capturing scenes behind a glass window, which is not only a disturbance visually but also influence the performance of other computer vision algorithms. Single image reflection removal is an ill-posed problem because the color at each pixel needs to be separated into two values, i.e., the desired clear background and the reflection. To solve it, existing methods propose priors such as smoothness, color consistency. However, the low-level priors are not reliable in complex scenes, for instance, when capturing a real outdoor scene through a window, both the foreground and background contain both smooth and sharp area and a variety of color. In this paper, inspired by the fact that human can separate the two layers easily by recognizing the objects, we use the object semantic as guidance to force the same semantic object belong to the same layer. Extensive experiments on different datasets show that adding the semantic information offers a significant improvement to reflection separation. We also demonstrate the applications of the proposed method to other computer vision tasks. | ['ShaoDi You', 'Yunfei Liu', 'Feng Lu', 'Yu Li'] | 2019-07-27 | null | null | null | null | ['reflection-removal'] | ['computer-vision'] | [ 6.63683355e-01 -5.36008701e-02 3.11023057e-01 -3.24671537e-01
-2.00537339e-01 -3.03410292e-01 2.78076708e-01 -2.66764045e-01
-2.48243734e-01 5.67531765e-01 -1.54899359e-01 -2.83558015e-02
2.31029302e-01 -6.81349933e-01 -6.37723267e-01 -1.11314750e+00
6.38194025e-01 1.54874679e-02 7.96428621e-01 1.15084171e-03
3.35945129e-01 3.17585886e-01 -1.67113471e+00 4.84622478e-01
9.73715305e-01 9.36728358e-01 5.92353523e-01 2.25593194e-01
-3.38921756e-01 6.32574379e-01 -4.50662792e-01 -6.42013699e-02
5.58326542e-01 -4.13535118e-01 -2.99597651e-01 5.55416822e-01
5.23071766e-01 -3.99200112e-01 4.26549241e-02 1.59177482e+00
1.08867660e-02 1.82311401e-01 5.92854202e-01 -1.10653496e+00
-4.64167714e-01 8.35781842e-02 -1.20439029e+00 -4.83853854e-02
2.19569966e-01 1.31668122e-02 4.66608644e-01 -8.68312716e-01
1.84122890e-01 1.21255720e+00 4.34507400e-01 4.53467101e-01
-8.24841917e-01 -6.65681899e-01 7.07020223e-01 1.91638246e-01
-1.06941748e+00 -4.83233750e-01 1.10549307e+00 -3.01209867e-01
1.74535826e-01 3.87632668e-01 5.08856595e-01 6.66716933e-01
5.18245324e-02 8.21890175e-01 1.37048185e+00 -4.36956793e-01
1.76597670e-01 6.54648364e-01 2.53919303e-01 6.36502087e-01
7.18898833e-01 -3.39260250e-02 -1.83523297e-01 6.66304380e-02
6.15209460e-01 4.45847571e-01 -7.51920640e-01 -1.81633413e-01
-8.89537692e-01 3.32112759e-01 2.44288415e-01 9.86417830e-02
-1.75396979e-01 -2.53557265e-01 -3.00243288e-01 -1.35374978e-01
2.53242940e-01 -1.51192009e-01 -2.11240202e-01 5.57660460e-01
-8.34370673e-01 -1.14631958e-01 6.11351490e-01 7.28413522e-01
8.07038009e-01 -6.40687868e-02 1.56869367e-01 8.63876581e-01
7.47310281e-01 7.02163458e-01 5.83850667e-02 -6.54555678e-01
3.81334931e-01 6.88505709e-01 5.28870761e-01 -1.18453097e+00
-8.88869762e-02 -2.73171932e-01 -8.66555393e-01 6.66630030e-01
6.88893259e-01 -1.57770421e-02 -1.18915629e+00 1.35530615e+00
6.31355822e-01 3.67218018e-01 4.14521731e-02 1.30347586e+00
9.69191909e-01 8.55084717e-01 -3.00868690e-01 -3.50545645e-01
1.47037995e+00 -9.88417089e-01 -7.24325955e-01 -7.11751282e-01
-2.36277193e-01 -1.27645755e+00 8.54649484e-01 8.44095767e-01
-1.05616128e+00 -6.15621865e-01 -1.18932545e+00 1.18020087e-01
-1.19073540e-02 3.17972541e-01 5.59010923e-01 6.88281655e-01
-5.71589351e-01 1.11215025e-01 -7.33236432e-01 -2.85444736e-01
8.94699767e-02 -1.54272109e-01 -1.22497618e-01 -5.81646740e-01
-8.41122627e-01 6.70920134e-01 1.63169980e-01 7.18672931e-01
-7.15566039e-01 -2.59988546e-01 -4.76975381e-01 -1.48417994e-01
6.27316058e-01 -3.79533678e-01 6.49191439e-01 -1.30514920e+00
-1.40171874e+00 8.67364585e-01 -2.93195277e-01 1.84171468e-01
6.41596675e-01 -4.12729919e-01 -3.59491825e-01 1.34368882e-01
-1.41118020e-01 -6.34242594e-02 1.24788857e+00 -1.89998817e+00
-8.44972312e-01 -4.53265488e-01 1.24057263e-01 4.22116876e-01
1.74023658e-02 -9.11766198e-03 -8.78437936e-01 -2.39302576e-01
6.52996242e-01 -8.32105458e-01 -1.49815887e-01 1.85410798e-01
-5.13685644e-01 1.63657248e-01 9.69638407e-01 -7.81515300e-01
5.41249931e-01 -2.33252716e+00 -2.55040497e-01 2.71853149e-01
1.20958708e-01 2.37948269e-01 3.14307064e-02 -1.31295070e-01
7.16968924e-02 -2.04955652e-01 -3.99174839e-01 2.34860405e-02
-4.12831962e-01 1.05255879e-01 -1.49689287e-01 6.55916572e-01
8.30481276e-02 1.24965630e-01 -7.32789576e-01 -4.96759921e-01
1.78713396e-01 6.10312879e-01 -2.66634911e-01 3.34116578e-01
-5.22091538e-02 4.96711850e-01 -6.46445572e-01 5.57200432e-01
1.33482647e+00 3.02410364e-04 1.24059014e-01 -4.67204720e-01
-1.25881568e-01 -1.79154009e-01 -1.75541103e+00 1.03238082e+00
-1.79635376e-01 4.56489593e-01 4.49819654e-01 -9.11050200e-01
1.09811723e+00 -5.67882843e-02 3.97252977e-01 -6.12489820e-01
6.48617968e-02 7.46935681e-02 -7.95386881e-02 -6.83673382e-01
4.56173122e-01 -2.33404323e-01 5.41797280e-01 2.44198054e-01
-7.31153429e-01 -1.00060880e-01 7.23499730e-02 -1.00720830e-01
6.36321843e-01 2.23750129e-01 -8.48727524e-02 -1.96964353e-01
6.57723427e-01 -1.52515888e-01 1.08238184e+00 6.80341005e-01
5.95621504e-02 1.04367721e+00 3.58933657e-02 -1.71481818e-01
-4.92741913e-01 -1.21504581e+00 8.05136282e-03 7.16789246e-01
9.08109784e-01 1.38298273e-01 -6.45888805e-01 -3.04399192e-01
-2.66476244e-01 4.74366337e-01 -2.23596901e-01 -1.50327206e-01
-5.40318727e-01 -8.74858618e-01 -2.05776900e-01 1.38011664e-01
8.59207511e-01 -7.52021492e-01 -8.11669528e-01 -1.11135449e-02
-5.49395859e-01 -1.36802959e+00 -3.74046028e-01 -1.02800652e-01
-8.09216321e-01 -1.42391932e+00 -7.30062962e-01 -9.42199230e-01
1.01167691e+00 1.03341794e+00 9.81985927e-01 5.26180983e-01
-4.53869671e-01 2.42656827e-01 -3.65335435e-01 -4.55281019e-01
-8.79423767e-02 -6.70745254e-01 -2.57224530e-01 5.92292547e-01
2.27630943e-01 -2.26884440e-01 -7.44019330e-01 4.92620409e-01
-1.00477529e+00 3.99544805e-01 6.35973692e-01 5.35123527e-01
3.77061605e-01 6.66255236e-01 -2.42503107e-01 -8.68826926e-01
2.04335168e-01 -1.25170201e-02 -8.25811684e-01 4.47634697e-01
-2.13899449e-01 -2.60466218e-01 4.26872641e-01 -4.09925520e-01
-1.74528515e+00 2.58693308e-01 4.01076883e-01 -1.02079153e-01
-3.85128677e-01 9.45172086e-03 -4.69015300e-01 5.60489297e-02
1.58538237e-01 3.49274576e-01 -1.20371148e-01 -4.49627578e-01
1.92696407e-01 5.24543703e-01 3.48281890e-01 -4.09807265e-01
9.98831868e-01 9.88649786e-01 -6.57771528e-02 -1.27462149e+00
-8.85161757e-01 -6.27495110e-01 -3.86268526e-01 -2.93527693e-01
9.96582866e-01 -8.11813354e-01 -6.39284790e-01 5.82767427e-01
-1.19584370e+00 -1.10884212e-01 2.50931680e-01 6.05968297e-01
-8.26504230e-02 6.74781263e-01 -4.88939196e-01 -1.14342713e+00
-5.42106815e-02 -1.15733969e+00 8.43693972e-01 7.27044582e-01
5.64219773e-01 -7.17550576e-01 -2.63523757e-01 5.12228012e-01
1.32289767e-01 4.71639633e-02 7.57406712e-01 8.93621370e-02
-9.92394447e-01 7.50623345e-02 -5.60836315e-01 5.12015998e-01
2.88756192e-01 2.35672355e-01 -1.13744950e+00 -3.90455425e-02
5.15605211e-01 1.66732073e-01 1.11780763e+00 3.98995250e-01
7.13109672e-01 -5.10004023e-03 -3.44367445e-01 6.07230365e-01
1.67650855e+00 4.80543792e-01 9.45548654e-01 2.71954954e-01
7.56574631e-01 9.48287368e-01 8.62619519e-01 3.36083204e-01
1.11211397e-01 4.30858314e-01 5.62074363e-01 -6.06526077e-01
-1.80140823e-01 2.79900014e-01 4.79448527e-01 5.67573726e-01
-3.52462620e-01 -2.34985054e-01 -5.76550961e-01 1.58025324e-01
-1.81956756e+00 -8.91794384e-01 -5.72912633e-01 2.33567858e+00
6.78592205e-01 1.73603833e-01 -3.24359626e-01 1.88778430e-01
8.78538847e-01 4.43220958e-02 -4.13168997e-01 6.25692531e-02
-2.19991028e-01 -1.53141454e-01 5.29506028e-01 5.22189260e-01
-8.04105878e-01 7.38736212e-01 5.44997358e+00 5.25355220e-01
-1.25279176e+00 -2.13838905e-01 5.48987210e-01 2.23709971e-01
-3.32523167e-01 2.06966355e-01 -8.71020615e-01 4.39202398e-01
-1.77340478e-01 3.27735633e-01 3.59989643e-01 4.82020527e-01
3.02328497e-01 -6.84431016e-01 -8.20538044e-01 1.17430997e+00
2.69444048e-01 -4.93000060e-01 -2.26884857e-02 -2.62803674e-01
6.47865653e-01 -4.08019930e-01 7.86376521e-02 -3.22731853e-01
2.05848321e-01 -8.21905077e-01 6.21966124e-01 6.47797108e-01
1.99924588e-01 -4.01767492e-01 6.05489373e-01 3.55029166e-01
-1.22642708e+00 1.76578864e-01 -6.56941533e-01 2.44291183e-02
9.16747376e-02 9.45068836e-01 -4.84103322e-01 6.14259601e-01
8.59807432e-01 6.92705452e-01 -3.95601749e-01 1.26462519e+00
-3.92357320e-01 2.63367742e-01 -4.37051743e-01 3.36012810e-01
8.12489614e-02 -9.35351849e-01 4.96919006e-01 8.75326276e-01
9.63730216e-02 3.17652822e-01 3.67181391e-01 1.01072562e+00
4.79526550e-01 -1.19291611e-01 -3.62887323e-01 2.59634912e-01
-1.46052297e-02 1.32300651e+00 -1.03340721e+00 -2.02680826e-01
-7.84241199e-01 1.22613084e+00 -1.88752890e-01 9.13296521e-01
-9.04706419e-01 -3.76765966e-01 4.88651007e-01 2.28606984e-01
2.32854545e-01 -2.78755784e-01 -2.56245643e-01 -1.29600847e+00
3.13121527e-01 -7.05877423e-01 1.51781008e-01 -1.04544783e+00
-1.09822023e+00 3.83223504e-01 -2.49647841e-01 -1.46715248e+00
6.62218630e-01 -8.31895590e-01 -6.75016880e-01 8.42029512e-01
-1.97524452e+00 -8.93687606e-01 -8.20623279e-01 8.09707463e-01
6.17767990e-01 2.47935176e-01 2.24413902e-01 3.51116508e-01
-5.94841361e-01 2.15184335e-02 2.08061442e-01 5.40461689e-02
7.63380289e-01 -8.92255127e-01 -3.31295282e-01 1.21289730e+00
1.00347884e-01 6.04692519e-01 9.31913197e-01 -5.17033219e-01
-1.34342313e+00 -8.19794893e-01 1.84355974e-01 4.08766381e-02
1.09954968e-01 -2.76864201e-01 -1.01135921e+00 4.87412602e-01
1.49899706e-01 -1.87467068e-01 1.84233442e-01 -1.50135845e-01
-1.77938938e-01 -3.28392029e-01 -9.38494265e-01 6.20708525e-01
7.04843044e-01 -8.74773264e-02 -6.21325493e-01 2.30058864e-01
2.75876880e-01 -3.15743864e-01 -1.61473006e-01 3.74000669e-01
5.35008609e-01 -1.45554221e+00 1.14131200e+00 -1.38292089e-01
2.58039862e-01 -7.51664400e-01 -1.18635438e-01 -1.02454948e+00
-1.30671054e-01 -2.05954209e-01 4.38048095e-01 1.21853650e+00
2.45920822e-01 -7.63001084e-01 7.65616417e-01 7.26222873e-01
9.12222341e-02 -5.88446409e-02 -3.78809482e-01 -5.76756120e-01
-5.16201794e-01 -2.71117926e-01 2.60775238e-01 8.66995156e-01
-4.53720629e-01 1.60254031e-01 -5.36779284e-01 6.63354576e-01
1.01686287e+00 5.79873741e-01 8.48415136e-01 -1.44326115e+00
-3.49377424e-01 -2.79412895e-01 -1.14172079e-01 -1.25078487e+00
-3.23896885e-01 -4.18377578e-01 5.24199009e-01 -1.80645871e+00
4.93719965e-01 -6.16258204e-01 -3.41484368e-01 1.11465909e-01
-5.10763288e-01 1.69574022e-01 2.31842116e-01 8.95929784e-02
-4.73460495e-01 5.13591409e-01 1.36455965e+00 -2.73769408e-01
-2.63797104e-01 2.16675028e-01 -7.20857799e-01 1.31092763e+00
7.55820274e-01 -4.31078196e-01 -5.53020537e-01 -6.31420970e-01
8.20713714e-02 -1.91864982e-01 3.38726699e-01 -8.84704888e-01
1.19676895e-01 -5.78339994e-01 4.67450470e-01 -5.20745695e-01
5.49086928e-01 -1.35291255e+00 8.39833915e-02 4.16725308e-01
2.43579164e-01 -4.89134073e-01 5.68291880e-02 8.97577465e-01
-9.78279710e-02 -3.76039416e-01 1.10665965e+00 -5.08613944e-01
-8.19998741e-01 3.35724615e-02 -1.77918807e-01 -5.52044846e-02
9.92598951e-01 -5.97346067e-01 -4.55985785e-01 -3.30649495e-01
-4.35895801e-01 6.87160417e-02 4.59849894e-01 3.21616977e-01
9.53864813e-01 -8.41393709e-01 -5.97589552e-01 3.16764891e-01
-5.24475947e-02 2.76683778e-01 3.55294853e-01 9.11320329e-01
-5.73034465e-01 -1.83150128e-01 -1.59567118e-01 -6.75200522e-01
-1.73253882e+00 3.97493869e-01 3.14020336e-01 2.86810607e-01
-9.14207280e-01 7.49871254e-01 1.02248335e+00 7.11750090e-02
3.68999779e-01 -5.23296177e-01 -3.28120768e-01 -2.92400956e-01
7.30971634e-01 3.17307562e-01 -2.85422117e-01 -4.43917692e-01
-4.00946200e-01 1.03664660e+00 -4.29570079e-02 5.59095778e-02
1.04536068e+00 -4.02388126e-01 -2.74315983e-01 5.48466980e-01
7.87976921e-01 3.88317019e-01 -1.22559083e+00 -3.33673537e-01
-2.36279398e-01 -9.58125591e-01 6.93005929e-03 -4.91930097e-01
-1.48259699e+00 8.84455323e-01 5.62642753e-01 2.74829447e-01
1.26597977e+00 -2.50856429e-01 6.44282401e-01 1.92197919e-01
3.35152566e-01 -1.13939464e+00 2.14581639e-01 1.43006310e-01
6.90905869e-01 -1.46198261e+00 3.14474344e-01 -1.00711262e+00
-6.74998522e-01 1.24458492e+00 6.84331298e-01 -1.95525289e-01
5.74033737e-01 1.84876099e-01 4.14466888e-01 -5.41473515e-02
-3.09306204e-01 -4.27357137e-01 3.70158613e-01 6.23708248e-01
3.19211364e-01 -2.06506863e-01 -1.75908715e-01 4.08587307e-01
3.43425870e-01 -2.42176563e-01 7.58758843e-01 7.80489385e-01
-8.72860909e-01 -6.16777062e-01 -8.31974804e-01 8.60672593e-02
-4.51212227e-01 -7.04400148e-03 -2.04570577e-01 6.58018649e-01
2.05077678e-01 1.38746965e+00 -5.35736457e-02 3.28986831e-02
1.87486932e-01 -4.03006792e-01 5.33768356e-01 -5.55923462e-01
2.55712569e-02 5.15579700e-01 -1.50758862e-01 -4.65697497e-01
-6.94711506e-01 -3.92636299e-01 -1.41860747e+00 9.51837376e-02
-4.97450531e-01 -5.38553577e-03 6.98939383e-01 8.49236727e-01
-2.94260323e-01 4.97893840e-01 4.92934346e-01 -5.56311548e-01
-8.71277601e-02 -6.93627059e-01 -1.01329994e+00 6.80724144e-01
3.13761234e-01 -6.89509273e-01 -7.11590886e-01 3.18372965e-01] | [10.311537742614746, -2.682293653488159] |
1ec5b430-b157-4ce9-80fd-4fb0bae861ea | causality-compensated-attention-for | null | null | https://openreview.net/forum?id=8XqDnrmZQNF | https://openreview.net/pdf?id=8XqDnrmZQNF | Causality Compensated Attention for Contextual Biased Visual Recognition | Visual attention does not always capture the essential object representation desired for robust predictions. Attention modules tend to underline not only the target object but also the common co-occurring context that the module thinks helpful in the training. The problem is rooted in the confounding effect of the context leading to incorrect causalities between objects and predictions, which is further exacerbated by visual attention. In this paper, to learn causal object features robust for contextual bias, we propose a novel attention module named Interventional Dual Attention (IDA) for visual recognition. Specifically, IDA adopts two attention layers with multiple sampling intervention, which compensates the attention against the confounder context. Note that our method is model-agnostic and thus can be implemented on various backbones. Extensive experiments show our model obtains significant improvements in classification and detection with lower computation. In particular, we achieve the state-of-the-art results in multi-label classification on MS-COCO and PASCAL-VOC. | ['Thomas H. Li', 'Ge Li', 'Jingjia Huang', 'Ruyang Liu'] | 2023-02-25 | null | null | null | iclr-2023-2 | ['multi-label-image-classification'] | ['computer-vision'] | [ 1.90853477e-01 -1.10325634e-01 -2.65389681e-01 -4.37346488e-01
-3.93313974e-01 -2.05329716e-01 6.31311953e-01 2.55772084e-01
-1.59765616e-01 4.97162104e-01 1.77400887e-01 -1.69740275e-01
1.79532785e-02 -4.38877642e-01 -8.07657361e-01 -7.02261567e-01
3.59484255e-01 1.37762249e-01 1.06581263e-01 6.85516074e-02
3.39353710e-01 3.66325378e-01 -1.60481310e+00 7.44184315e-01
6.28080487e-01 1.09574687e+00 4.61148202e-01 5.35839140e-01
-1.88625738e-01 1.18196046e+00 -6.12822354e-01 -4.63336080e-01
-2.83271432e-01 -4.29473400e-01 -8.33859801e-01 2.39650220e-01
7.69358099e-01 -4.38257419e-02 -1.33072257e-01 1.04294956e+00
3.47532630e-01 -1.43664420e-01 9.64766860e-01 -1.45923042e+00
-1.10220933e+00 5.10987878e-01 -8.47813964e-01 3.15103740e-01
4.44555469e-02 2.81150967e-01 1.38063240e+00 -1.16632128e+00
2.27790117e-01 1.57833445e+00 4.56157058e-01 5.62090933e-01
-1.22108030e+00 -7.84741163e-01 6.63031101e-01 6.03872001e-01
-1.20745480e+00 -1.50726259e-01 6.51876628e-01 -6.39400184e-01
7.39821315e-01 5.04152775e-01 2.24015594e-01 1.41029775e+00
2.55347610e-01 9.08759296e-01 9.99076843e-01 -3.86448383e-01
-6.90855384e-02 2.79357731e-01 3.48276734e-01 5.72778583e-01
1.99618340e-01 8.14901292e-02 -3.82472903e-01 -1.22349486e-01
4.25385535e-01 2.61997938e-01 -2.32469678e-01 -3.14785428e-02
-1.11604548e+00 7.78224707e-01 7.72170722e-01 1.57620758e-01
-3.40087354e-01 3.80465537e-01 2.95352578e-01 -7.35366866e-02
4.02280539e-01 3.98359388e-01 -4.33732718e-01 5.99804103e-01
-5.65566361e-01 2.38530293e-01 1.63886726e-01 8.49610567e-01
5.29502988e-01 -8.54180977e-02 -6.31859124e-01 8.71115863e-01
5.96163571e-01 3.72443140e-01 4.33288604e-01 -5.08455098e-01
2.50840783e-01 7.71466911e-01 2.83269137e-02 -1.08151960e+00
-4.48931217e-01 -7.41126835e-01 -9.77750480e-01 1.69550896e-01
2.43690178e-01 2.93264389e-01 -1.06129003e+00 1.74219465e+00
3.58859628e-01 2.55800664e-01 -2.24326581e-01 9.80279744e-01
1.10367680e+00 3.75668973e-01 7.21958995e-01 -2.69080698e-01
1.54375625e+00 -1.20105636e+00 -7.47367799e-01 -4.76710290e-01
3.79093796e-01 -7.87937999e-01 1.12130785e+00 1.47956952e-01
-4.97218966e-01 -7.72540867e-01 -9.72870290e-01 -1.83814257e-01
-1.75905630e-01 3.41531128e-01 5.16968310e-01 2.10221112e-01
-3.50731701e-01 5.14965475e-01 -2.65657455e-01 -2.50421286e-01
6.76619530e-01 1.40519708e-01 -1.26839459e-01 -5.44136465e-02
-1.13256538e+00 8.13991547e-01 3.58332306e-01 1.34630755e-01
-1.16461957e+00 -7.62796521e-01 -5.14143229e-01 2.55263448e-01
5.08001268e-01 -5.00141263e-01 1.19200492e+00 -1.28319526e+00
-9.48784471e-01 7.75228381e-01 -2.00425163e-01 -1.91665962e-01
5.02701342e-01 -3.75178158e-01 -3.89443606e-01 -1.78336203e-01
2.73489594e-01 6.53962135e-01 1.02810812e+00 -1.42913043e+00
-7.42113709e-01 -4.67173129e-01 -1.29089011e-02 2.31545474e-02
-2.83895075e-01 2.03497171e-01 -3.51129174e-01 -8.55512679e-01
-7.90388733e-02 -8.36229384e-01 -2.74571866e-01 3.98473218e-02
-5.08168101e-01 -4.74703252e-01 8.73784423e-01 -3.64771545e-01
1.25652456e+00 -2.15315723e+00 1.14155367e-01 -2.64236093e-01
3.27175140e-01 1.05440453e-01 -1.38823777e-01 1.27010375e-01
-3.68582875e-01 3.08478713e-01 -5.29411584e-02 -3.28409165e-01
-2.63761997e-01 1.68980643e-01 -6.69098735e-01 3.79012972e-01
4.89564478e-01 1.02279472e+00 -1.05278790e+00 -6.53337777e-01
1.14858933e-02 3.05557460e-01 -4.78733301e-01 4.80708092e-01
-4.07064974e-01 5.17952442e-01 -3.69067043e-01 6.78359568e-01
4.59485620e-01 -7.70349920e-01 3.77307862e-01 -4.10362333e-01
5.32933651e-03 1.54316455e-01 -9.92712080e-01 1.17857361e+00
-2.06320763e-01 4.51519638e-01 -1.86849609e-01 -8.25734377e-01
6.82193220e-01 3.14212650e-01 2.32904144e-02 -5.73358119e-01
1.21012457e-01 2.36248486e-02 9.04536247e-02 -6.17313683e-01
3.11730623e-01 -1.49102837e-01 1.77519083e-01 2.11076930e-01
1.35358110e-01 5.05921245e-01 -1.60513893e-01 1.99507400e-01
5.75765908e-01 1.72424763e-01 5.62911689e-01 -3.73863250e-01
5.55075586e-01 -1.89366832e-01 8.98476243e-01 8.45659077e-01
-2.13576436e-01 5.81037343e-01 6.53488934e-01 -3.88328969e-01
-7.64526367e-01 -6.69498026e-01 -1.76130757e-01 1.38690841e+00
2.36667678e-01 -4.62919086e-01 -2.68330127e-01 -1.17518699e+00
1.26858160e-01 9.11876261e-01 -1.16221821e+00 -3.53551880e-02
-2.51554787e-01 -8.84045660e-01 2.73591846e-01 8.21734548e-01
1.85130253e-01 -1.05217886e+00 -4.63582754e-01 8.34337771e-02
-2.30144039e-01 -8.47431242e-01 -5.60837746e-01 3.08949947e-01
-7.32215285e-01 -1.33678353e+00 -5.60562432e-01 -4.85192060e-01
6.20279610e-01 4.52733457e-01 1.09319663e+00 2.58267045e-01
-3.77580225e-01 5.68181574e-02 -1.89396977e-01 -6.70235872e-01
-2.59548396e-01 -2.77423322e-01 -1.80454805e-01 4.00479227e-01
3.41378003e-01 -2.14365810e-01 -5.12883723e-01 4.19732660e-01
-6.08686805e-01 2.43906528e-01 5.85087180e-01 1.25205302e+00
6.04999721e-01 -4.48380530e-01 6.62624419e-01 -1.00003123e+00
2.20658049e-01 -7.50900745e-01 -3.51874441e-01 5.39455891e-01
-7.43965626e-01 9.38516706e-02 5.53007245e-01 -5.29278219e-01
-1.02097273e+00 7.47735649e-02 1.10367998e-01 -7.40105927e-01
-3.03715229e-01 3.42545331e-01 -3.37688863e-01 2.66672492e-01
6.72030210e-01 5.24533466e-02 -3.54449600e-01 -5.57065248e-01
3.27385068e-01 5.42092562e-01 2.03647062e-01 -3.44944239e-01
2.64548451e-01 3.73906881e-01 8.36989731e-02 -5.07584155e-01
-1.31349897e+00 -3.33955228e-01 -4.94429648e-01 -3.95087659e-01
1.00644720e+00 -8.52409124e-01 -9.05789137e-01 2.62648165e-01
-1.44739544e+00 -7.28560844e-03 1.48166463e-01 2.71516651e-01
-1.64247468e-01 2.05908149e-01 -5.05081594e-01 -8.55403662e-01
-2.02643022e-01 -1.22665954e+00 8.93556952e-01 1.22269273e-01
-1.52950108e-01 -7.06298351e-01 -1.37700006e-01 3.49648297e-01
1.25024721e-01 1.01504251e-02 9.76569176e-01 -6.83029175e-01
-6.14407778e-01 1.81346789e-01 -6.58682346e-01 1.23184815e-01
1.62916798e-02 5.95895611e-02 -1.46268344e+00 -1.59942567e-01
-1.63574263e-01 -4.62420166e-01 1.18500662e+00 1.90758154e-01
1.57104456e+00 -3.71384531e-01 -5.77837169e-01 2.63826281e-01
1.35604501e+00 3.94314714e-02 5.64814031e-01 8.69075730e-02
1.03203428e+00 6.99747801e-01 8.96262348e-01 2.84952402e-01
3.00324768e-01 8.76537740e-01 8.51089716e-01 -1.88236356e-01
-2.98502862e-01 -2.00711697e-01 4.07967940e-02 3.41417193e-01
4.39903624e-02 -3.08528572e-01 -8.81292701e-01 5.99764466e-01
-2.13268828e+00 -9.49365199e-01 -6.21228158e-01 1.92491269e+00
7.55266070e-01 -4.55989093e-02 -1.75522536e-01 -1.30923733e-01
7.59968638e-01 1.52251050e-01 -5.07459164e-01 -2.02463731e-01
-2.75733564e-02 -2.61451095e-01 3.25525939e-01 4.39059585e-01
-1.19163120e+00 7.28037775e-01 6.21558094e+00 9.46639001e-01
-1.23109591e+00 3.38992298e-01 9.65694964e-01 -5.72509505e-02
-2.43657351e-01 -1.40038356e-01 -1.02172935e+00 3.99756849e-01
6.48898125e-01 6.92265034e-02 -3.33342748e-03 1.02054381e+00
-1.60770282e-01 7.12883100e-02 -1.31041634e+00 7.58478880e-01
6.70516491e-02 -1.22416806e+00 2.16591731e-01 -3.90666351e-03
6.03384852e-01 -2.33588979e-01 7.84882829e-02 3.39529335e-01
1.60393521e-01 -1.20127356e+00 9.40329731e-01 5.38941979e-01
7.32288301e-01 -6.16487205e-01 6.19677961e-01 2.50937372e-01
-1.03643501e+00 -3.07666123e-01 -3.92792374e-01 -8.23602974e-02
-3.13568771e-01 6.02292657e-01 -8.49068165e-01 4.99010742e-01
7.97511339e-01 8.18158627e-01 -8.29222560e-01 7.35585570e-01
-5.89249730e-01 7.42582798e-01 2.88845420e-01 -1.78171918e-01
1.41541660e-01 4.48308438e-01 4.37935114e-01 1.34746850e+00
-4.90174629e-02 5.70382178e-02 1.47576213e-01 1.09065986e+00
-1.26685321e-01 2.43327349e-01 -4.47517008e-01 1.03141360e-01
1.85282260e-01 1.29838037e+00 -5.19751728e-01 -5.61186433e-01
-5.25699556e-01 8.94176722e-01 8.12522411e-01 3.41525108e-01
-1.14496422e+00 9.77533683e-02 6.98559642e-01 -7.33897164e-02
4.93110299e-01 2.31217757e-01 -4.78407025e-01 -1.12123334e+00
-2.68798262e-01 -7.99315631e-01 5.07587790e-01 -7.96359181e-01
-1.60309994e+00 4.56858218e-01 -3.01155508e-01 -1.08048522e+00
3.24249327e-01 -7.08587229e-01 -6.59871757e-01 1.01673436e+00
-1.58699715e+00 -1.27590036e+00 -4.17697370e-01 3.23309690e-01
6.79256916e-01 7.02667534e-02 8.47568095e-01 4.08977300e-01
-7.33505309e-01 6.99608624e-01 -2.71045715e-01 -9.53461379e-02
1.02542722e+00 -1.22958601e+00 1.21774897e-02 7.20836163e-01
1.07317455e-01 6.09919012e-01 6.54913604e-01 -7.61805773e-01
-9.20356572e-01 -1.40556645e+00 1.05281103e+00 -5.22759736e-01
5.60367405e-01 -2.69462287e-01 -1.17997038e+00 8.29670191e-01
3.29375058e-01 1.92031756e-01 8.60867262e-01 3.91298532e-01
-8.32109809e-01 -1.08574145e-01 -7.49610305e-01 5.74972153e-01
8.50165308e-01 -4.16115522e-01 -6.18165553e-01 2.91422307e-01
8.42813134e-01 -1.49868637e-01 -5.07039726e-01 5.56295455e-01
5.69691539e-01 -8.89001131e-01 8.73394310e-01 -1.07031298e+00
8.44694316e-01 -3.83216560e-01 -2.06658423e-01 -1.07881331e+00
-9.78355825e-01 1.17128849e-01 -2.01932088e-01 1.29543829e+00
5.01536846e-01 -3.14937294e-01 2.96137810e-01 4.42439735e-01
-5.53873815e-02 -8.50623131e-01 -7.17690945e-01 -4.66265440e-01
-3.50615606e-02 -5.65388560e-01 5.12806118e-01 1.26733971e+00
-2.55785018e-01 7.02399671e-01 -7.03210771e-01 4.30946648e-01
7.90636003e-01 4.34022635e-01 4.41660285e-01 -1.15490866e+00
-4.17057037e-01 -4.46049720e-01 -1.26765639e-01 -8.51733923e-01
1.12287894e-01 -8.88013124e-01 9.29599181e-02 -1.23059213e+00
7.00641334e-01 -3.73677790e-01 -7.50288963e-01 7.52193987e-01
-8.55488539e-01 1.76018074e-01 3.61438602e-01 3.90205860e-01
-7.74035871e-01 6.39091730e-01 1.24201417e+00 -3.75599891e-01
2.15151921e-01 -2.73913532e-01 -8.58188868e-01 8.01739216e-01
5.57607770e-01 -6.92498565e-01 -1.48459315e-01 -3.57547760e-01
1.78465456e-01 -2.01663658e-01 7.87185431e-01 -6.65686488e-01
1.17628731e-01 -4.97085094e-01 6.33530080e-01 -5.95914364e-01
3.19487900e-02 -7.98482299e-01 4.82228026e-03 5.62585950e-01
-6.68670177e-01 -2.12208256e-01 2.66080946e-01 8.96419585e-01
-1.05187841e-01 -2.20531464e-01 9.69161034e-01 -1.46251321e-01
-6.93862736e-01 1.41081512e-01 -1.78428695e-01 -1.70035064e-01
9.92544770e-01 3.01627904e-01 -5.98030925e-01 5.46392798e-02
-6.97353661e-01 4.13049966e-01 1.10188544e-01 6.65280461e-01
5.47215521e-01 -1.55715394e+00 -7.30664909e-01 -7.15298131e-02
4.72849190e-01 -3.88289899e-01 4.09773082e-01 8.56940150e-01
1.89814195e-01 5.34711182e-01 -1.73144210e-02 -6.72623634e-01
-1.47688079e+00 1.13321829e+00 3.75525892e-01 -2.34840140e-01
-2.63502300e-01 1.01223469e+00 9.31708992e-01 -5.99411801e-02
3.60848129e-01 -2.60896087e-01 -4.70665604e-01 2.16021910e-01
8.43448877e-01 1.78590700e-01 -2.34691262e-01 -5.41232586e-01
-5.31597853e-01 4.10968542e-01 -3.14200103e-01 4.93996978e-01
1.04914129e+00 -1.00948520e-01 -1.07604258e-01 6.93553686e-01
8.78110170e-01 -1.78333402e-01 -1.32362163e+00 -2.25919127e-01
-6.30392954e-02 -4.86820132e-01 1.10023573e-01 -1.16999257e+00
-1.08335674e+00 1.15228212e+00 6.30713344e-01 9.49901864e-02
9.15292382e-01 8.48768353e-02 1.39355779e-01 -2.71098278e-02
-8.37987512e-02 -7.33323395e-01 3.15025121e-01 4.15044904e-01
1.46121776e+00 -1.56850827e+00 1.08839445e-01 -5.09811103e-01
-7.08683252e-01 1.02771842e+00 1.04490840e+00 1.32574543e-01
4.80579168e-01 -1.53418481e-01 1.52396094e-02 -3.34189653e-01
-1.09816897e+00 -2.30452210e-01 5.69715023e-01 2.06181079e-01
6.79819822e-01 1.26002014e-01 -3.06075931e-01 8.79371464e-01
7.53548145e-01 -2.32662946e-01 1.48248121e-01 4.43593830e-01
-4.12875086e-01 -7.49419570e-01 -4.69494730e-01 3.27174544e-01
-6.60636425e-01 -3.16260546e-01 -4.79205072e-01 5.88728547e-01
3.92137557e-01 8.51176679e-01 -5.21608144e-02 -3.72478485e-01
2.61454910e-01 1.25146613e-01 3.06753904e-01 -6.31003499e-01
-7.82344103e-01 1.39665231e-01 -1.27203748e-01 -6.84337676e-01
-3.23492140e-01 -5.08389711e-01 -1.04247105e+00 1.34223700e-01
-4.96813983e-01 -1.88446507e-01 4.29103702e-01 9.15029049e-01
5.14516890e-01 9.90680277e-01 4.90913481e-01 -5.97279668e-01
-4.53314841e-01 -1.17614734e+00 -3.45407218e-01 6.42725408e-01
3.69514227e-01 -9.59894240e-01 -3.17246020e-01 -2.71959715e-02] | [9.89003849029541, 2.0389726161956787] |
b5c77183-c19c-4324-bce2-c967be3596f0 | visually-grounded-word-embeddings-and-richer | 1707.01009 | null | http://arxiv.org/abs/1707.01009v5 | http://arxiv.org/pdf/1707.01009v5.pdf | Visually Grounded Word Embeddings and Richer Visual Features for Improving Multimodal Neural Machine Translation | In Multimodal Neural Machine Translation (MNMT), a neural model generates a
translated sentence that describes an image, given the image itself and one
source descriptions in English. This is considered as the multimodal image
caption translation task. The images are processed with Convolutional Neural
Network (CNN) to extract visual features exploitable by the translation model.
So far, the CNNs used are pre-trained on object detection and localization
task. We hypothesize that richer architecture, such as dense captioning models,
may be more suitable for MNMT and could lead to improved translations. We
extend this intuition to the word-embeddings, where we compute both linguistic
and visual representation for our corpus vocabulary. We combine and compare
different confi | ['Stéphane Dupont', 'Jean-Benoit Delbrouck', 'Omar Seddati'] | 2017-07-04 | null | null | null | null | ['dense-captioning'] | ['computer-vision'] | [ 5.39238989e-01 5.41193485e-01 -2.68300116e-01 -4.33006972e-01
-8.28435600e-01 -6.14876688e-01 1.07136583e+00 -2.75097936e-01
-4.87860799e-01 7.10433662e-01 3.49323839e-01 -2.34295964e-01
6.96932554e-01 -7.16988802e-01 -1.27200270e+00 -3.05716366e-01
4.58692789e-01 5.12461543e-01 -1.94095820e-01 -2.83152819e-01
-7.90280849e-02 1.74134165e-01 -1.05561721e+00 9.92431939e-01
5.58716953e-01 9.25538182e-01 3.01841319e-01 6.62781000e-01
-1.40700132e-01 6.21534705e-01 -4.01940674e-01 -1.02084267e+00
1.06158286e-01 -6.78271711e-01 -8.45016241e-01 2.03939393e-01
6.64673269e-01 -3.76710147e-01 -7.64929712e-01 1.16474962e+00
5.37447780e-02 -2.05202311e-01 8.90672803e-01 -1.30133688e+00
-1.40662801e+00 6.49379551e-01 -3.42148185e-01 -6.48329780e-02
4.66039628e-01 3.38496506e-01 8.23621690e-01 -1.26628458e+00
1.03404617e+00 1.43283546e+00 1.79699659e-01 7.56638706e-01
-1.44058120e+00 -1.20647445e-01 -3.15030426e-01 2.08091125e-01
-1.21752071e+00 -3.95753205e-01 6.35002017e-01 -2.61040419e-01
7.24541366e-01 2.45187059e-01 3.20419669e-01 1.74899185e+00
2.63441235e-01 1.11677599e+00 1.06525540e+00 -5.14402390e-01
-6.45369515e-02 5.77866316e-01 -5.22488952e-01 6.26902580e-01
1.05072230e-01 1.20655231e-01 -5.84069312e-01 2.05794722e-01
8.63423944e-01 -8.02267492e-02 -3.42360824e-01 -4.26416636e-01
-1.51992130e+00 1.10664713e+00 7.47267962e-01 1.88548163e-01
-4.53927308e-01 3.42294157e-01 4.08786684e-01 4.00725067e-01
3.28096241e-01 6.78484082e-01 -2.21360043e-01 2.03674033e-01
-8.92169178e-01 -1.40495330e-01 5.57649016e-01 1.06995070e+00
9.96436715e-01 -1.77097656e-02 -4.40131038e-01 6.25153542e-01
2.88061142e-01 9.19609725e-01 5.04332244e-01 -9.97119546e-01
6.78949535e-01 4.40051854e-01 1.90289784e-02 -9.10536230e-01
-1.73468925e-02 -2.36904413e-01 -8.20021987e-01 -1.27435729e-01
1.23866558e-01 -1.49158491e-02 -1.04143846e+00 1.84855878e+00
-2.48425111e-01 -1.34578317e-01 5.39724886e-01 1.29640317e+00
9.38106894e-01 7.35467017e-01 1.38837650e-01 9.62139964e-02
1.50299537e+00 -1.12217343e+00 -6.68497622e-01 -4.67011869e-01
3.87912244e-01 -7.47464955e-01 1.27396977e+00 -1.53642029e-01
-1.30744958e+00 -5.52468538e-01 -9.79807138e-01 -3.24682415e-01
-5.45534015e-01 2.82221675e-01 3.98703605e-01 2.40975007e-01
-1.24643099e+00 1.25317454e-01 -5.31373322e-01 -7.81524777e-01
5.65347195e-01 1.73605278e-01 -6.03597045e-01 -2.17909440e-01
-1.24670959e+00 1.28880978e+00 3.02308381e-01 1.60602912e-01
-1.21729732e+00 2.46293992e-02 -1.12407351e+00 3.46305370e-02
2.09865365e-02 -1.21140683e+00 1.27902865e+00 -1.83453834e+00
-1.25306857e+00 1.18963754e+00 -4.33571190e-01 -5.98683774e-01
2.98697859e-01 1.25184059e-01 -9.82442647e-02 6.20798886e-01
2.84310505e-02 1.57517660e+00 1.12089610e+00 -1.53723085e+00
-2.09205002e-01 -2.39156440e-01 1.58467025e-01 2.09546953e-01
-3.76551092e-01 7.93301128e-03 -4.78413343e-01 -5.04178047e-01
-2.99183816e-01 -1.00436103e+00 -8.86738971e-02 -5.40946312e-02
-4.19209182e-01 1.11350305e-01 7.48040438e-01 -7.24823713e-01
5.12020111e-01 -1.89000690e+00 7.08444297e-01 -1.27921298e-01
1.28027529e-01 1.31199092e-01 -8.10855389e-01 5.00420272e-01
6.90778494e-02 2.17828467e-01 -2.13987589e-01 -5.38683414e-01
5.50891310e-02 3.98860693e-01 -5.14987469e-01 2.59867519e-01
7.74165630e-01 1.85758388e+00 -6.41795099e-01 -4.84262347e-01
8.53263810e-02 6.56921804e-01 -1.46057293e-01 4.43724126e-01
-5.26257455e-01 4.79223520e-01 -2.44162902e-01 5.84127545e-01
6.59765899e-01 -2.54846513e-01 -1.18155248e-01 -5.09048223e-01
2.18688607e-01 -8.10614377e-02 -5.19562066e-02 1.94252992e+00
-6.39144480e-01 1.19142699e+00 -1.09716721e-01 -1.01789796e+00
7.41377115e-01 3.83159339e-01 -2.01082639e-02 -9.03208554e-01
3.79290968e-01 1.67505994e-01 -1.98297054e-01 -8.81649196e-01
6.20740592e-01 -2.34477714e-01 -1.54677287e-01 4.77673024e-01
2.61163712e-01 -1.66326895e-01 1.73275154e-02 2.56374925e-01
7.74253607e-01 3.28357100e-01 -1.27671465e-01 1.60494596e-01
3.17658752e-01 3.35579991e-01 -1.30498052e-01 6.16620004e-01
-3.14596817e-02 9.41714168e-01 4.33810264e-01 -5.38908124e-01
-1.51558816e+00 -1.21866894e+00 1.55299857e-01 9.05065298e-01
5.59010655e-02 1.41704991e-01 -9.11032081e-01 -6.39067233e-01
-3.65506500e-01 8.41432750e-01 -7.11453915e-01 -3.92588228e-01
-5.27952969e-01 -3.02177012e-01 7.04485893e-01 5.44193864e-01
5.26404142e-01 -1.28758729e+00 -7.35011756e-01 -4.81650159e-02
-5.42291582e-01 -1.45748019e+00 -6.56806707e-01 -1.22425472e-02
-7.75781453e-01 -6.94350660e-01 -1.06169105e+00 -1.01577067e+00
9.13372159e-01 9.11022872e-02 1.28204489e+00 -6.18859120e-02
-5.53027838e-02 6.99489534e-01 -4.99624521e-01 -3.38705063e-01
-8.08536410e-01 2.77331382e-01 -1.20302662e-01 2.64702410e-01
4.53096747e-01 -2.31087789e-01 -5.04866838e-01 7.09302025e-03
-1.17030406e+00 6.72982872e-01 1.26512873e+00 9.07536805e-01
3.03371400e-01 -1.02250075e+00 4.06902641e-01 -4.74264055e-01
6.41296804e-01 -3.16294700e-01 -1.98069543e-01 5.17423809e-01
-1.74149856e-01 2.82478660e-01 4.46608245e-01 -7.66232312e-01
-7.91097641e-01 2.38531917e-01 1.14375584e-01 -8.27703595e-01
-2.74517298e-01 5.55552304e-01 -4.46207710e-02 -4.40994427e-02
6.73940957e-01 6.01890624e-01 1.31668717e-01 1.82959461e-03
8.03730130e-01 6.28589749e-01 6.59891546e-01 -4.56484884e-01
9.56793308e-01 5.07858992e-01 -2.83146929e-02 -4.83628064e-01
-6.60754621e-01 2.01989323e-01 -7.86447227e-01 -1.72718778e-01
1.45039463e+00 -9.62136507e-01 -3.92768174e-01 -2.31329396e-01
-1.88398147e+00 -1.50241405e-01 -9.90282372e-03 4.08787757e-01
-8.86966705e-01 2.48255711e-02 -5.94991148e-01 -6.03606105e-01
-2.75003135e-01 -1.34253502e+00 1.48314786e+00 1.95012674e-01
-1.56995311e-01 -9.65237021e-01 -6.30329326e-02 4.52829570e-01
5.97600877e-01 2.48729050e-01 9.81239796e-01 -5.84483624e-01
-9.21825230e-01 -1.19256832e-01 -7.85491943e-01 3.01789284e-01
-2.88538903e-01 -4.57063079e-01 -1.06282389e+00 -2.04515278e-01
-1.32125020e-01 -6.40090704e-01 1.05521345e+00 2.01018080e-01
7.34741271e-01 -5.39036274e-01 -3.18033993e-01 4.88885403e-01
1.50250542e+00 -1.12379089e-01 8.24971676e-01 2.32137769e-01
7.06314027e-01 7.91036129e-01 2.87877470e-01 -3.17555040e-01
4.45468038e-01 6.95356190e-01 6.61907852e-01 -6.08757794e-01
-2.40564227e-01 -4.41548407e-01 6.83444440e-01 6.02664351e-01
9.15552974e-02 -4.04324383e-01 -8.90526414e-01 5.79054058e-01
-1.99059129e+00 -6.56599164e-01 6.53474480e-02 1.59077847e+00
6.93410516e-01 -3.33254278e-01 -2.08892703e-01 -7.00717032e-01
6.95729017e-01 -1.13693736e-01 -5.12378216e-01 -7.43737876e-01
-5.16950071e-01 2.89546382e-02 4.09971833e-01 3.90275657e-01
-9.23665464e-01 1.12970603e+00 6.74578524e+00 4.10479963e-01
-1.13207591e+00 5.14035583e-01 8.94534171e-01 1.54816791e-01
-4.95766699e-01 -6.97057694e-02 -2.18243510e-01 1.45484895e-01
1.24638975e+00 2.05417886e-01 5.96523702e-01 4.84340072e-01
1.39395937e-01 8.92729759e-02 -1.53825092e+00 1.06493640e+00
7.05732346e-01 -1.56678820e+00 6.89125061e-01 1.59425005e-01
8.71557891e-01 1.10169522e-01 4.77367908e-01 3.48654807e-01
3.47834290e-03 -1.32424462e+00 8.61780167e-01 7.87462354e-01
9.73540545e-01 -4.19849038e-01 1.04531431e+00 5.75065017e-02
-7.26433277e-01 1.74668983e-01 -5.11521757e-01 2.07322568e-01
2.04317003e-01 -7.56957456e-02 -9.25963283e-01 5.29522479e-01
3.66688490e-01 5.51646054e-01 -8.47488165e-01 4.99368072e-01
-2.48985872e-01 2.79097736e-01 2.01173186e-01 -8.23763758e-02
4.23366308e-01 -2.91593999e-01 4.98647958e-01 1.18128586e+00
3.61258447e-01 -4.57415044e-01 -2.46940516e-02 1.52604175e+00
-4.54900503e-01 3.79674919e-02 -9.81470585e-01 -4.06456113e-01
-4.47292775e-02 1.16253352e+00 -4.17150497e-01 -5.33195436e-01
-7.85974562e-01 1.53765881e+00 3.80316257e-01 8.86289179e-01
-1.05803359e+00 -6.49368018e-02 3.40127975e-01 -5.29769436e-03
1.33820087e-01 -1.59050152e-01 -4.21040952e-01 -1.49933529e+00
-1.62055448e-01 -8.86160493e-01 -1.54800966e-01 -1.43858159e+00
-1.24948883e+00 1.01179075e+00 7.72439837e-02 -1.17899680e+00
-3.66949290e-01 -9.05070543e-01 -4.52496797e-01 7.67213166e-01
-1.43489420e+00 -1.78155410e+00 -4.47984748e-02 3.48170727e-01
7.67699778e-01 -2.73219287e-01 8.41710985e-01 1.29274502e-01
-2.61274189e-01 5.79834104e-01 -3.16904277e-01 3.40802372e-01
7.89731801e-01 -9.42684114e-01 1.99767783e-01 5.86012423e-01
4.70569283e-01 5.84955335e-01 8.54403794e-01 -4.75455672e-01
-1.59826744e+00 -1.38850963e+00 9.61589336e-01 -8.18783998e-01
5.49157321e-01 -4.27064627e-01 -6.86301291e-01 8.67647827e-01
1.19157088e+00 -3.27814966e-01 4.49479610e-01 -5.07595062e-01
-3.71460080e-01 3.14671844e-01 -9.56315875e-01 8.68702412e-01
7.67926574e-01 -8.40435505e-01 -8.06784809e-01 3.83403629e-01
9.38471079e-01 -2.70026147e-01 -7.11589813e-01 1.58216536e-01
3.20919842e-01 -4.33673799e-01 9.69874501e-01 -8.47416639e-01
9.87521470e-01 -8.02144036e-02 -5.24675727e-01 -1.29246998e+00
1.47338107e-01 -2.63018608e-01 7.38419443e-02 8.90478194e-01
9.85575497e-01 -3.18482399e-01 4.74545717e-01 5.26998580e-01
-9.92237329e-02 -5.39631188e-01 -9.80712056e-01 -6.03782952e-01
2.94381529e-01 -2.13391930e-01 5.08110940e-01 8.64979804e-01
-1.06172182e-01 6.88911676e-01 -5.22728980e-01 -2.41414700e-02
3.38993728e-01 1.03067756e-01 7.89225221e-01 -7.30227590e-01
6.46667704e-02 -4.56737339e-01 -2.44176880e-01 -9.87928510e-01
4.94313747e-01 -1.32339227e+00 4.95525375e-02 -1.61226833e+00
7.46337652e-01 5.47297359e-01 3.27353626e-02 6.05944097e-01
1.78032085e-01 8.66937280e-01 4.00926352e-01 4.35861081e-01
-7.29906380e-01 8.48932266e-01 1.54299378e+00 -5.78182995e-01
2.97073036e-01 -6.33407772e-01 -5.70038676e-01 3.43664885e-01
5.86518168e-01 -4.16934043e-01 -7.15686083e-02 -9.79982376e-01
3.24119210e-01 1.55652016e-01 7.64991164e-01 -4.97831941e-01
4.91504651e-03 -1.07110478e-01 6.58393204e-01 -3.66936207e-01
5.23241162e-01 -8.50467145e-01 -2.81866901e-02 3.91527832e-01
-8.26664209e-01 2.03427628e-01 1.71319529e-01 4.33385342e-01
-5.46011269e-01 -1.42599806e-01 5.97708464e-01 -3.11036795e-01
-3.47517699e-01 2.76738793e-01 -4.81475413e-01 -4.33654934e-01
7.90087402e-01 -1.92279339e-01 -3.68687451e-01 -7.48902619e-01
-9.00446296e-01 5.19555919e-02 7.32680261e-01 7.29408145e-01
9.28676605e-01 -1.82299852e+00 -8.41019034e-01 2.84257829e-02
4.35561359e-01 -5.24672091e-01 -2.50679970e-01 9.48648393e-01
-4.38582152e-01 7.51406550e-01 -3.69635046e-01 -7.82543123e-01
-7.96560407e-01 8.08376074e-01 3.62428606e-01 1.38614044e-01
-3.30707371e-01 4.14746940e-01 4.38015431e-01 -3.80340695e-01
-1.49877250e-01 -3.53124380e-01 -5.40982410e-02 -2.57355273e-01
4.27743465e-01 -3.57974917e-01 -4.21589524e-01 -1.13338304e+00
-1.24724656e-01 4.82622653e-01 1.24109410e-01 -5.69329500e-01
9.38155890e-01 -4.34806287e-01 -2.39310354e-01 3.23979884e-01
1.41199934e+00 -5.21778822e-01 -1.03690982e+00 -3.67112190e-01
-1.10913716e-01 -1.86955154e-01 -1.85812980e-01 -7.82652020e-01
-9.35346425e-01 1.25054383e+00 6.80497706e-01 -2.07844615e-01
1.05113780e+00 4.53241765e-01 9.06347334e-01 4.97042269e-01
2.79689699e-01 -7.84108102e-01 5.19475937e-01 4.95349109e-01
1.30835772e+00 -1.65793121e+00 -5.49655437e-01 7.40637630e-02
-1.01859558e+00 1.34145999e+00 7.12766886e-01 -5.21700345e-02
-1.23241335e-01 -2.41402775e-01 1.52714103e-01 -1.27601489e-01
-8.38412106e-01 -3.24558824e-01 5.97728968e-01 6.85531139e-01
3.49129349e-01 1.73183337e-01 -2.19814107e-02 5.76217532e-01
-3.18867993e-03 -1.59453422e-01 5.55572808e-01 5.27716160e-01
-3.42334837e-01 -1.04104447e+00 -5.00650167e-01 1.04254022e-01
-8.50269198e-02 -2.21167892e-01 -9.14886475e-01 5.92391789e-01
1.58414513e-01 8.51745605e-01 1.41042560e-01 -4.01895255e-01
2.26499945e-01 2.27822766e-01 6.35292292e-01 -6.11629844e-01
-3.97170812e-01 1.33230789e-02 -1.01686820e-01 -5.19031167e-01
-5.78546405e-01 -4.70067620e-01 -1.00016677e+00 2.19791811e-02
9.00575072e-02 -8.16103593e-02 9.09549475e-01 9.61431026e-01
3.45477670e-01 3.53294939e-01 2.89921612e-01 -1.02304852e+00
-1.65476084e-01 -1.24373698e+00 1.07292987e-01 5.48561513e-01
2.55548686e-01 -2.96197236e-01 -1.22070827e-01 4.61176276e-01] | [11.418649673461914, 1.4732046127319336] |
feea9ea6-4aee-4dfc-a478-8088ca184948 | videollm-modeling-video-sequence-with-large | 2305.13292 | null | https://arxiv.org/abs/2305.13292v2 | https://arxiv.org/pdf/2305.13292v2.pdf | VideoLLM: Modeling Video Sequence with Large Language Models | With the exponential growth of video data, there is an urgent need for automated technology to analyze and comprehend video content. However, existing video understanding models are often task-specific and lack a comprehensive capability of handling diverse tasks. The success of large language models (LLMs) like GPT has demonstrated their impressive abilities in sequence causal reasoning. Building upon this insight, we propose a novel framework called VideoLLM that leverages the sequence reasoning capabilities of pre-trained LLMs from natural language processing (NLP) for video sequence understanding. VideoLLM incorporates a carefully designed Modality Encoder and Semantic Translator, which convert inputs from various modalities into a unified token sequence. This token sequence is then fed into a decoder-only LLM. Subsequently, with the aid of a simple task head, our VideoLLM yields an effective unified framework for different kinds of video understanding tasks. To evaluate the efficacy of VideoLLM, we conduct extensive experiments using multiple LLMs and fine-tuning methods. We evaluate our VideoLLM on eight tasks sourced from four different datasets. The experimental results demonstrate that the understanding and reasoning capabilities of LLMs can be effectively transferred to video understanding tasks. We release the code at https://github.com/cg1177/VideoLLM. | ['LiMin Wang', 'Tong Lu', 'Yu Qiao', 'Yali Wang', 'Yi Wang', 'Junting Pan', 'Yifei HUANG', 'Jilan Xu', 'Jiahao Wang', 'Yin-Dong Zheng', 'Guo Chen'] | 2023-05-22 | null | null | null | null | ['video-understanding'] | ['computer-vision'] | [ 3.51166308e-01 -1.25341907e-01 -3.88417363e-01 -4.19514716e-01
-7.44459212e-01 -6.28364623e-01 8.02034497e-01 -2.78325319e-01
-1.10720672e-01 4.78171945e-01 4.88877028e-01 -4.73407120e-01
2.72895008e-01 -4.72366631e-01 -1.09378934e+00 -1.46127537e-01
3.48395944e-01 8.87465402e-02 1.32604674e-01 1.58869792e-02
2.04225674e-01 -2.05787480e-01 -1.27835548e+00 8.89849603e-01
9.00678337e-01 8.17598343e-01 5.75891912e-01 8.70667756e-01
-2.20858514e-01 1.57738912e+00 -1.38666436e-01 -4.97930676e-01
4.25934233e-02 -6.08260751e-01 -1.09200478e+00 -2.85082962e-02
2.58854479e-01 -5.40691078e-01 -6.15090728e-01 7.60198236e-01
1.12760745e-01 2.02673867e-01 4.26155329e-01 -1.63079309e+00
-7.81420112e-01 7.88135529e-01 -3.10559958e-01 2.02899620e-01
8.61546040e-01 4.11183983e-01 1.00361037e+00 -8.47883403e-01
6.09026492e-01 1.56062937e+00 4.45060134e-01 6.94698215e-01
-7.01079369e-01 -7.14712739e-01 3.18301499e-01 6.92759693e-01
-1.08274686e+00 -5.49294174e-01 4.05044556e-01 -2.89069027e-01
1.18213713e+00 -8.23920146e-02 3.82470459e-01 1.42723453e+00
7.75801390e-02 1.24552882e+00 7.07213879e-01 -4.14200246e-01
-1.03060319e-03 -2.01979607e-01 7.34429583e-02 9.66711819e-01
-2.19889715e-01 -2.40576267e-01 -9.87918556e-01 1.83670551e-01
7.70856261e-01 3.83522399e-02 -4.08508539e-01 -4.08424400e-02
-1.36636114e+00 6.24948144e-01 1.31722227e-01 1.08494451e-02
-2.51988739e-01 6.51617587e-01 7.02163696e-01 3.89937758e-01
2.22148269e-01 1.75771758e-01 -4.54699159e-01 -5.63281536e-01
-6.80177689e-01 3.78908095e-04 9.48835611e-01 1.21818650e+00
4.17060524e-01 -2.70067990e-01 -1.70425460e-01 6.19119704e-01
4.42427784e-01 5.04840612e-01 4.17921454e-01 -1.37370646e+00
7.36413598e-01 5.65673590e-01 -1.76974580e-01 -9.48312104e-01
-7.49717280e-02 1.65532783e-01 -5.33792377e-01 -4.40515876e-01
1.88554525e-01 -9.85939652e-02 -8.13868940e-01 1.95729542e+00
-1.08299367e-02 6.21202171e-01 2.66689390e-01 8.78764749e-01
7.51657009e-01 9.59190428e-01 5.68148792e-01 5.46378009e-02
1.40402758e+00 -1.39549315e+00 -4.62722301e-01 -3.47286463e-01
7.87955523e-01 -5.37328720e-01 1.36363280e+00 2.66250968e-01
-1.03307045e+00 -5.74195445e-01 -7.63044536e-01 -2.92330086e-01
-5.53919375e-02 5.77843674e-02 5.93977332e-01 1.84685051e-01
-1.04368138e+00 1.25945672e-01 -8.97864401e-01 -6.09824359e-01
5.57339191e-01 -2.58484203e-02 -2.88643390e-01 -5.72659612e-01
-1.37738359e+00 6.76017880e-01 5.98425031e-01 4.88390289e-02
-1.23160601e+00 -9.89349127e-01 -1.05771923e+00 8.07477385e-02
7.26705074e-01 -1.23530424e+00 1.58739698e+00 -1.29789698e+00
-1.43081760e+00 5.74360371e-01 -5.85116625e-01 -4.90524232e-01
3.55873644e-01 -7.02731192e-01 -2.72813946e-01 6.40975773e-01
2.12746620e-01 8.62649143e-01 8.42320204e-01 -1.00531435e+00
-7.31898069e-01 1.99965417e-01 3.88676256e-01 3.57301772e-01
-2.66770571e-01 2.73081124e-01 -8.57475579e-01 -5.84460557e-01
-4.15892988e-01 -7.86627412e-01 4.46663685e-02 -9.88268703e-02
-2.27611020e-01 -1.82898283e-01 7.36917973e-01 -8.47376823e-01
1.16033077e+00 -1.99484527e+00 3.47058445e-01 -1.54497325e-01
1.59261659e-01 2.79754728e-01 -4.86606926e-01 5.99236429e-01
-1.94828939e-02 3.90555978e-01 -1.48369059e-01 -2.79312432e-01
-5.10062836e-03 3.92219514e-01 -4.81623083e-01 -2.22116679e-01
3.92488480e-01 1.33069873e+00 -1.13051045e+00 -5.64531803e-01
2.87864655e-01 2.56488085e-01 -7.35719025e-01 3.97053331e-01
-7.46199131e-01 5.09101212e-01 -7.92509496e-01 6.39087558e-01
1.00108489e-01 -7.60144591e-01 2.32484668e-01 -1.81224704e-01
2.59278864e-01 2.89913088e-01 -5.84838271e-01 2.22894764e+00
-6.30872190e-01 7.82873869e-01 -4.24624085e-01 -9.89980340e-01
2.68390834e-01 6.85261607e-01 2.44734734e-01 -6.26837611e-01
-9.28137545e-03 -1.51695916e-03 -2.36642540e-01 -1.04413605e+00
2.24796101e-01 -8.99959058e-02 -1.76704317e-01 7.18838871e-01
3.02146554e-01 1.43466249e-01 2.97003299e-01 8.15395355e-01
1.21134663e+00 4.43117619e-01 2.90731311e-01 1.86371371e-01
6.59113765e-01 1.31424204e-01 4.14135069e-01 9.31580424e-01
-9.87356305e-02 5.21627069e-01 5.30682504e-01 -3.30791682e-01
-9.73280668e-01 -9.26600933e-01 6.12023294e-01 1.28759229e+00
1.62937164e-01 -7.52197981e-01 -8.56939912e-01 -5.69862902e-01
-1.61300749e-01 7.94092834e-01 -3.30007821e-01 -2.87371725e-01
-4.76925105e-01 -3.87339979e-01 9.20233428e-01 7.78925478e-01
8.64958227e-01 -1.11265016e+00 -4.34094667e-01 -1.29892537e-02
-8.14069986e-01 -1.78330350e+00 -3.48961294e-01 -5.40152073e-01
-6.13966405e-01 -1.13991654e+00 -2.40551099e-01 -6.38647079e-01
4.54751462e-01 5.25822937e-01 1.35761523e+00 3.36747795e-01
-4.09753341e-03 9.02584195e-01 -7.35441387e-01 -2.74466157e-01
-6.76984608e-01 1.07180521e-01 2.20748764e-02 -9.95574817e-02
4.46737736e-01 -4.67483521e-01 -4.19471860e-01 2.24452332e-01
-9.36794817e-01 7.18237281e-01 5.11551082e-01 5.90376973e-01
2.08958164e-01 -1.60193279e-01 5.63003957e-01 -6.74289942e-01
4.96747196e-01 -7.92068958e-01 -1.65578455e-01 5.89714944e-01
-5.76576255e-02 1.30246311e-01 7.69353747e-01 -3.58918995e-01
-1.41043699e+00 -1.62077636e-01 -3.66822891e-02 -6.42177165e-01
-2.05720037e-01 7.29162574e-01 -3.81321877e-01 1.72773197e-01
1.98643833e-01 2.96995521e-01 -1.83864255e-02 -2.66061097e-01
7.03512013e-01 4.45187271e-01 7.28842556e-01 -9.23032641e-01
6.52968764e-01 3.59745771e-01 -5.31756580e-01 -6.44028604e-01
-1.10430658e+00 -2.92981982e-01 -3.10270309e-01 -4.30725574e-01
1.14096773e+00 -1.23164630e+00 -8.83531272e-01 4.84456033e-01
-1.28668988e+00 -6.73318386e-01 1.34059578e-01 3.72927457e-01
-8.20353627e-01 5.50385773e-01 -8.71451199e-01 -4.73352015e-01
-2.59913772e-01 -1.14445460e+00 9.59397614e-01 3.35290819e-01
-4.04540032e-01 -1.20323658e+00 -4.09033865e-01 9.28904593e-01
1.18680678e-01 -2.92831182e-01 1.14472806e+00 -5.33159554e-01
-8.68485212e-01 6.12574220e-02 -4.10135925e-01 2.90765733e-01
3.97712411e-03 -8.75865296e-02 -8.64053249e-01 1.20361559e-01
-1.10682219e-01 -7.64623344e-01 8.98138702e-01 2.26692989e-01
1.31371069e+00 -2.44546577e-01 -3.31414074e-01 6.06934786e-01
1.21393919e+00 1.89492300e-01 5.99745870e-01 3.89679581e-01
1.01987004e+00 2.55393982e-01 6.42187357e-01 2.72898853e-01
9.00576293e-01 3.34708005e-01 2.29065746e-01 2.14932397e-01
-1.52986541e-01 -7.70417094e-01 7.86541522e-01 1.04917133e+00
-2.05681056e-01 -6.22033060e-01 -1.14597058e+00 3.69808108e-01
-2.23081374e+00 -1.07129931e+00 1.43650129e-01 1.51057172e+00
7.82368302e-01 -7.58384541e-02 -2.82670200e-01 -2.74795443e-01
3.79415333e-01 1.03321724e-01 -6.12954259e-01 -3.41634810e-01
9.94407833e-02 -1.49791613e-01 2.38260716e-01 4.15148795e-01
-9.05369878e-01 1.36427414e+00 6.12515354e+00 9.69986558e-01
-7.73573756e-01 1.06743254e-01 4.68956977e-01 -3.70712578e-02
-4.15082783e-01 2.45845124e-01 -5.30634642e-01 4.44026172e-01
1.14636838e+00 -2.07539245e-01 5.45752466e-01 4.45778072e-01
5.72049260e-01 -2.24816278e-01 -1.27637911e+00 1.03950381e+00
2.70265102e-01 -1.41185498e+00 5.97062111e-01 -5.14357924e-01
5.14702380e-01 -1.13287158e-02 -2.31758311e-01 3.94195437e-01
3.43904674e-01 -1.06878591e+00 7.75923967e-01 5.83548844e-01
5.70638955e-01 -3.19166392e-01 3.95421118e-01 4.35854852e-01
-1.21309149e+00 -2.42911324e-01 -1.14855617e-01 -3.53693575e-01
6.48225546e-01 1.80153266e-01 -7.64169991e-01 7.04243958e-01
5.58794022e-01 1.23388124e+00 -3.43291193e-01 5.44455886e-01
-7.71855831e-01 8.75656128e-01 -6.03326596e-02 3.54365110e-01
1.76811978e-01 8.59721079e-02 3.41852039e-01 1.24474740e+00
1.29996642e-01 3.49270254e-01 2.40997136e-01 7.30430901e-01
-3.54031920e-01 -2.16507494e-01 -5.02978861e-01 -5.18338263e-01
4.83136952e-01 1.08223498e+00 -4.32103902e-01 -7.18428731e-01
-8.82656217e-01 1.12923336e+00 1.49046540e-01 5.69813013e-01
-1.27333522e+00 2.73474723e-01 8.18463147e-01 -2.48795748e-01
4.43951376e-02 -3.92125845e-01 -1.38953090e-01 -1.65819049e+00
-4.40131836e-02 -1.24228883e+00 5.41156769e-01 -1.36438394e+00
-1.16363919e+00 3.40042740e-01 2.01277599e-01 -8.90966475e-01
-3.46927285e-01 -6.28962159e-01 -5.25954187e-01 4.46148366e-01
-1.50517023e+00 -1.41307652e+00 -3.22813958e-01 7.16515660e-01
1.18536866e+00 -1.40204638e-01 5.41128933e-01 2.87736505e-01
-6.38014376e-01 3.68261278e-01 -4.68252063e-01 2.33852133e-01
6.76290751e-01 -8.05453598e-01 6.12372816e-01 1.06945312e+00
8.32177103e-02 6.90639198e-01 4.91029412e-01 -7.21464694e-01
-1.73411965e+00 -1.18273902e+00 7.74733901e-01 -6.57496333e-01
9.84712660e-01 -3.29590231e-01 -8.55394661e-01 1.14158428e+00
3.97033334e-01 -4.37220842e-01 7.82091439e-01 -2.12447777e-01
-5.84537506e-01 3.82807553e-01 -4.80112284e-01 8.54521573e-01
1.36520481e+00 -8.82488251e-01 -8.55820298e-01 2.83810318e-01
1.11972642e+00 -3.84069741e-01 -7.84268439e-01 2.51801074e-01
5.49716175e-01 -5.23480773e-01 9.92288589e-01 -1.01875234e+00
1.16151810e+00 -2.84277588e-01 -1.79432571e-01 -1.02235293e+00
-5.00098802e-02 -5.64985693e-01 -2.53999263e-01 1.11104953e+00
4.21612084e-01 -3.76138806e-01 3.87223631e-01 9.52055275e-01
-1.50137022e-01 -4.88556206e-01 -3.00855577e-01 -5.52965641e-01
-2.55094469e-01 -1.00275660e+00 2.65397131e-01 8.46319556e-01
2.42776066e-01 5.03135741e-01 -4.73308831e-01 2.67903626e-01
4.06123400e-01 1.90689817e-01 7.85061419e-01 -5.59405923e-01
-4.32784528e-01 -1.48875713e-01 -1.32313937e-01 -1.45723403e+00
5.99538386e-01 -9.47808683e-01 2.05066688e-02 -1.61724544e+00
7.21131742e-01 -3.05522960e-02 -2.40677521e-01 9.04246628e-01
-4.46909726e-01 1.04168914e-01 5.83370507e-01 3.36535931e-01
-1.03794336e+00 4.34888631e-01 1.21561646e+00 5.44373281e-02
1.24214582e-01 -4.89305168e-01 -7.62849927e-01 1.05327415e+00
7.61026800e-01 -2.19817504e-01 -7.61971295e-01 -9.42857563e-01
3.68314475e-01 1.91564053e-01 5.92842162e-01 -8.26231956e-01
2.99609333e-01 -3.33753735e-01 9.51613635e-02 -1.23154998e-01
2.70276010e-01 -4.87021476e-01 -3.05718984e-02 1.87926471e-01
-5.84719419e-01 5.19140661e-02 2.92236298e-01 5.32423079e-01
-3.76139522e-01 5.76300127e-03 1.52050108e-01 -2.92397648e-01
-1.57274425e+00 2.63712287e-01 -5.65443337e-01 2.40169004e-01
1.00192916e+00 -1.32991686e-01 -4.21330363e-01 -7.62416422e-01
-4.50922310e-01 7.16889918e-01 5.38233399e-01 9.01921868e-01
7.05470622e-01 -1.03043413e+00 -5.55578709e-01 -1.26679808e-01
1.10530429e-01 -8.61089751e-02 4.24529582e-01 8.91087532e-01
-4.51663703e-01 4.72209036e-01 -1.68610066e-01 -5.23605645e-01
-1.17629731e+00 5.74403048e-01 1.66693211e-01 -3.09633873e-02
-6.27760112e-01 8.14423859e-01 6.00953162e-01 -1.45797551e-01
-9.63454973e-03 -4.19244498e-01 1.27245396e-01 -3.70797247e-01
7.46749043e-01 1.02210939e-01 -5.10393202e-01 -4.08870012e-01
-2.94720471e-01 4.02187765e-01 3.52443382e-02 -1.93584755e-01
1.20475328e+00 -4.74688739e-01 -7.27064610e-02 3.73858839e-01
1.04965842e+00 -5.43548703e-01 -1.43709743e+00 -1.71866298e-01
8.17860365e-02 -2.37516657e-01 -3.77472430e-01 -7.17209637e-01
-9.56122696e-01 9.16852713e-01 -2.01290146e-01 -3.33276063e-01
1.44668889e+00 1.00787401e-01 1.06707108e+00 3.06689322e-01
3.73836756e-01 -8.53899360e-01 3.79435629e-01 7.89576530e-01
6.07828856e-01 -1.26214838e+00 -3.37719291e-01 -6.32180393e-01
-9.90527332e-01 1.09114850e+00 5.67429960e-01 2.23351374e-01
1.78749397e-01 2.49761268e-01 6.33102879e-02 -9.14101377e-02
-1.26439285e+00 -6.59136381e-03 6.95873722e-02 3.53169262e-01
3.87040675e-01 -1.87045932e-01 -7.37904459e-02 7.78581500e-01
1.19382657e-01 7.12974608e-01 4.36959863e-01 9.36581910e-01
-2.11448804e-01 -1.02446413e+00 -7.48822019e-02 1.30307570e-01
-3.00644934e-01 -3.99894208e-01 -2.11078331e-01 6.34180367e-01
-1.25387296e-01 1.07737696e+00 -8.61715972e-02 -4.66798872e-01
-8.43322426e-02 3.64958614e-01 4.58793670e-01 -5.15673101e-01
-2.59322584e-01 -3.90175670e-01 2.30965003e-01 -9.05614972e-01
-7.10183561e-01 -5.24678528e-01 -1.55338740e+00 -2.86138237e-01
2.09645763e-01 -5.86919934e-02 2.32573077e-01 1.39726245e+00
4.70405579e-01 5.31883001e-01 1.55515775e-01 -5.39515376e-01
-1.05536908e-01 -6.58660889e-01 1.33591041e-01 6.05899811e-01
-1.19309658e-02 -4.30861801e-01 3.45387124e-02 8.23895693e-01] | [10.415485382080078, 1.0253098011016846] |
c22a932a-5a80-48ee-ad1a-0a96bd9ea813 | manifold-aware-self-training-for-unsupervised | 2305.10808 | null | https://arxiv.org/abs/2305.10808v1 | https://arxiv.org/pdf/2305.10808v1.pdf | Manifold-Aware Self-Training for Unsupervised Domain Adaptation on Regressing 6D Object Pose | Domain gap between synthetic and real data in visual regression (\eg 6D pose estimation) is bridged in this paper via global feature alignment and local refinement on the coarse classification of discretized anchor classes in target space, which imposes a piece-wise target manifold regularization into domain-invariant representation learning. Specifically, our method incorporates an explicit self-supervised manifold regularization, revealing consistent cumulative target dependency across domains, to a self-training scheme (\eg the popular Self-Paced Self-Training) to encourage more discriminative transferable representations of regression tasks. Moreover, learning unified implicit neural functions to estimate relative direction and distance of targets to their nearest class bins aims to refine target classification predictions, which can gain robust performance against inconsistent feature scaling sensitive to UDA regressors. Experiment results on three public benchmarks of the challenging 6D pose estimation task can verify the effectiveness of our method, consistently achieving superior performance to the state-of-the-art for UDA on 6D pose estimation. | ['Kui Jia', 'YaoWei Wang', 'Zelin Xu', 'Ke Chen', 'Jiehong Lin', 'Yichen Zhang'] | 2023-05-18 | null | null | null | null | ['6d-pose-estimation-1', 'unsupervised-domain-adaptation'] | ['computer-vision', 'methodology'] | [ 1.18284464e-01 1.43079832e-01 -7.23970115e-01 -4.79926169e-01
-1.07104349e+00 -5.98875225e-01 7.59582460e-01 -5.05524836e-02
-4.90164645e-02 5.02376735e-01 3.40371966e-01 3.11525404e-01
-8.81771222e-02 -3.18115354e-01 -9.19261932e-01 -6.30653977e-01
-1.22178502e-01 6.16029501e-01 2.69253384e-02 -2.22208232e-01
1.31405756e-01 5.87349474e-01 -1.33654785e+00 1.87101085e-02
7.77720988e-01 1.22810543e+00 -1.36052087e-01 1.24222241e-01
3.26865047e-01 3.18350762e-01 -3.15694064e-01 -1.90204829e-01
6.06635869e-01 9.70029552e-03 -3.57792228e-01 3.79099935e-01
9.67660666e-01 -3.30035925e-01 -2.91661948e-01 8.92500281e-01
4.35476840e-01 4.14960720e-02 1.30474532e+00 -1.30544937e+00
-8.13070655e-01 -9.62509960e-02 -9.93206680e-01 2.15984718e-03
2.45572388e-01 3.10627460e-01 8.67647469e-01 -1.02169466e+00
8.54543686e-01 1.36508441e+00 8.39802504e-01 6.13495767e-01
-1.70303237e+00 -8.38053346e-01 4.11002994e-01 -9.75985155e-02
-1.52868772e+00 -5.47807157e-01 1.02924037e+00 -7.06310511e-01
7.50379860e-01 8.64476338e-02 2.39566952e-01 1.51701581e+00
8.82969946e-02 5.96178353e-01 8.94552112e-01 -1.13543488e-01
1.61173359e-01 1.02141932e-01 -1.89832747e-02 5.35004973e-01
2.38582522e-01 1.96832001e-01 -6.86523914e-01 -6.96831867e-02
1.10115039e+00 -1.29675269e-01 -2.24369943e-01 -1.53653646e+00
-1.24234796e+00 8.67127240e-01 7.72008538e-01 -1.69799149e-01
-1.92119509e-01 -4.44328263e-02 5.18721640e-01 1.81439236e-01
9.30142641e-01 3.99065167e-01 -6.87396765e-01 2.49190822e-01
-4.59954649e-01 3.15869004e-01 1.98437482e-01 1.21010888e+00
6.62616074e-01 1.81179866e-01 -4.81755853e-01 9.64279890e-01
4.13509429e-01 5.80208957e-01 2.69362062e-01 -7.70540833e-01
7.88219571e-01 8.11890662e-01 2.08125353e-01 -1.09070301e+00
-4.50012833e-01 -8.16105545e-01 -8.24204862e-01 4.03342545e-01
6.74368143e-01 2.12207481e-01 -1.03544676e+00 1.98745692e+00
6.53504491e-01 2.11346611e-01 -1.54366255e-01 1.21483517e+00
5.87062478e-01 3.29013765e-01 2.26087064e-01 -1.42292306e-01
9.88149285e-01 -6.95540667e-01 -2.63520658e-01 -3.33726764e-01
6.87485576e-01 -5.43757200e-01 9.86023128e-01 -7.19826203e-03
-5.77045739e-01 -8.13721240e-01 -1.08509219e+00 3.45143527e-02
-8.43349174e-02 3.64000529e-01 5.71224511e-01 5.41018605e-01
-6.15964592e-01 4.73949581e-01 -6.68318272e-01 -2.96469331e-01
7.18155444e-01 5.34656048e-01 -6.96926117e-01 5.21153733e-02
-7.88441241e-01 1.03702545e+00 1.79710165e-01 3.21811065e-02
-1.06339931e+00 -1.09965265e+00 -1.27529359e+00 -5.11448860e-01
4.30525035e-01 -6.65119648e-01 7.84590662e-01 -8.87280941e-01
-1.50166047e+00 1.31157374e+00 5.31645529e-02 -3.67965072e-01
5.97750127e-01 -3.44566464e-01 -9.06827077e-02 -1.35219619e-01
2.28044987e-01 9.58792686e-01 1.40376377e+00 -1.44351017e+00
-4.20088261e-01 -9.79959071e-01 -3.32921475e-01 5.27090847e-01
-4.55496103e-01 -5.89543462e-01 -4.50132281e-01 -7.99884200e-01
4.72724557e-01 -1.07041037e+00 -1.68475375e-01 2.71034658e-01
-2.27923185e-01 -4.02104676e-01 8.38548362e-01 -6.13042116e-01
8.44746113e-01 -2.15137410e+00 6.69318795e-01 2.52889961e-01
3.42331022e-01 -8.02153796e-02 -5.04219532e-01 -7.30442628e-02
-3.12102586e-01 -3.98383409e-01 -4.33940850e-02 -4.20390069e-01
-9.27573908e-03 1.62593201e-02 -4.28951621e-01 1.10467732e+00
6.01680338e-01 1.07118368e+00 -7.36420751e-01 -3.28117490e-01
3.04938167e-01 4.62938547e-01 -4.99255925e-01 2.92778671e-01
-3.18346143e-01 8.62979770e-01 -7.83303440e-01 8.01191688e-01
7.76775122e-01 -1.57861784e-01 -6.55256957e-02 -5.67820430e-01
9.72025692e-02 2.21211404e-01 -1.02843380e+00 2.19076896e+00
-2.57503271e-01 3.17300737e-01 -1.12560809e-01 -1.18016493e+00
1.36706150e+00 -1.57596022e-01 5.66239834e-01 -7.54707217e-01
1.37351342e-02 2.17860308e-03 -3.68141830e-01 -1.34161592e-01
2.05123052e-01 3.15426707e-01 -1.84255511e-01 -9.69759524e-02
1.92940414e-01 -1.45579368e-01 -4.55337912e-01 -1.14332400e-01
6.69613957e-01 8.97614300e-01 1.46490246e-01 -3.00112456e-01
4.37595069e-01 1.27672508e-01 6.84644222e-01 2.65615135e-01
-5.12621641e-01 6.54439509e-01 3.06241244e-01 -4.49630827e-01
-1.15311372e+00 -1.36238992e+00 -4.03009683e-01 1.04935825e+00
4.67487365e-01 -2.27608174e-01 -4.09987837e-01 -1.17360818e+00
4.27795708e-01 4.05059993e-01 -1.01389563e+00 -4.96841669e-01
-5.61513066e-01 -5.02809167e-01 2.07333565e-01 6.35387242e-01
2.20617682e-01 -4.55745369e-01 -2.05825657e-01 -2.62807477e-02
9.03454274e-02 -1.03333235e+00 -5.51306188e-01 3.35192055e-01
-1.10794771e+00 -9.06524062e-01 -8.69197547e-01 -8.52589190e-01
7.56350040e-01 2.14935511e-01 9.64384556e-01 -6.59216046e-01
-2.25951880e-01 5.86644948e-01 -1.78137586e-01 -1.47756651e-01
-1.43630803e-01 1.36626244e-01 2.84242749e-01 -8.25622976e-02
3.90542865e-01 -5.75884819e-01 -7.37515390e-01 7.00236976e-01
-2.07508981e-01 -1.72423956e-03 5.85821867e-01 8.26980352e-01
7.68590331e-01 -6.97023988e-01 4.94675100e-01 -5.41713417e-01
1.00538418e-01 -4.45645750e-01 -6.09672129e-01 4.65724841e-02
-5.87489843e-01 2.34452143e-01 2.52736002e-01 -7.03181207e-01
-7.88778961e-01 4.07008648e-01 2.99922436e-01 -9.33986127e-01
-2.34938115e-01 -3.75444330e-02 -1.94110051e-01 -3.52088571e-01
1.14625776e+00 1.47175655e-01 3.56689155e-01 -5.05983591e-01
4.69977766e-01 2.97595650e-01 4.55533832e-01 -6.39508247e-01
1.38192713e+00 2.89527565e-01 2.16941565e-01 -4.71834183e-01
-1.00451899e+00 -3.92887205e-01 -9.73789871e-01 -1.86042652e-01
7.83405542e-01 -1.35861993e+00 -4.29500788e-01 2.10969210e-01
-7.82983422e-01 -2.81381130e-01 -2.77393252e-01 4.07681942e-01
-9.39563632e-01 1.43125325e-01 -3.43682528e-01 -5.07670283e-01
-8.00432190e-02 -8.50975215e-01 1.64317417e+00 -3.51386294e-02
-3.19112659e-01 -8.06490898e-01 2.14936227e-01 3.80428135e-01
6.83092773e-02 7.33962655e-01 7.25030005e-01 -5.90410233e-01
-4.79883254e-01 -1.63183168e-01 -3.62961441e-01 2.57905841e-01
2.77364403e-01 -4.34135795e-01 -9.24953580e-01 -6.12568855e-01
-2.48574153e-01 -5.45426726e-01 8.03685188e-01 4.93966788e-01
1.15253782e+00 4.32450846e-02 -6.42594099e-01 8.13193619e-01
1.12496841e+00 -3.81712943e-01 3.60820591e-01 3.33197892e-01
8.80358040e-01 6.87847495e-01 1.24664366e+00 4.74060029e-01
3.81512403e-01 1.04643548e+00 7.44719863e-01 -7.68923461e-02
-2.05376655e-01 -4.29511875e-01 2.88031995e-01 2.00877205e-01
-1.75275683e-01 4.91004825e-01 -7.02599943e-01 2.81245619e-01
-1.78856206e+00 -5.55849671e-01 2.77617276e-02 2.46460938e+00
7.22047985e-01 2.19453782e-01 3.52673560e-01 -1.30047858e-01
7.45586812e-01 2.91938871e-01 -7.80144930e-01 2.32326597e-01
7.73233026e-02 -5.06599993e-03 6.84719145e-01 6.86684549e-02
-1.50368118e+00 1.15398574e+00 5.52585125e+00 1.17367256e+00
-1.16884959e+00 -9.56901163e-02 7.06426740e-01 4.03618552e-02
2.79600639e-02 -3.03609699e-01 -9.38995898e-01 1.19847685e-01
4.37040627e-01 1.10936321e-01 1.89727321e-01 1.16014159e+00
-1.45584652e-02 3.39653671e-01 -1.36333740e+00 1.16275227e+00
8.78501311e-02 -1.20548594e+00 1.10875212e-01 1.67790279e-01
6.53328419e-01 -1.52665183e-01 4.69040692e-01 5.79704463e-01
8.87369215e-02 -1.02796590e+00 7.18384683e-01 2.13272586e-01
1.15986919e+00 -8.08009326e-01 1.55681491e-01 2.54760623e-01
-1.25587046e+00 -1.33165613e-01 -5.77979028e-01 2.01086760e-01
-3.82266849e-01 1.07789256e-01 -1.10793006e+00 5.88727772e-01
7.23387301e-01 1.37610674e+00 -7.49515235e-01 6.74639225e-01
-1.19799227e-02 1.89412102e-01 3.56780831e-03 3.28565359e-01
1.19700089e-01 -2.48915821e-01 8.23171318e-01 8.40512455e-01
3.31779826e-03 -2.83741593e-01 2.52875894e-01 7.78317571e-01
1.90115392e-01 9.86551866e-02 -7.67793894e-01 3.14090043e-01
2.79975563e-01 1.26823175e+00 -3.45867515e-01 2.69064426e-01
-1.53032690e-01 1.01053751e+00 6.21334851e-01 4.67041701e-01
-1.11692989e+00 2.09608421e-01 1.14376152e+00 3.52581382e-01
4.95216876e-01 -4.59584594e-01 -4.59386438e-01 -1.02378559e+00
5.69925159e-02 -7.96040297e-01 4.61237311e-01 -4.57947850e-01
-1.60329843e+00 2.80446947e-01 -5.88652454e-02 -2.05213928e+00
-1.86378673e-01 -8.22857440e-01 -1.75845116e-01 8.04790080e-01
-1.38899541e+00 -1.69955933e+00 -2.71734178e-01 8.49749029e-01
5.43470919e-01 -2.92610615e-01 8.04263473e-01 1.07503720e-01
-4.05145049e-01 9.37828183e-01 -1.57445028e-01 1.43123314e-01
1.23595536e+00 -9.77623999e-01 2.52357543e-01 4.05687481e-01
1.03918105e-01 4.02438670e-01 6.41402304e-01 -6.93417013e-01
-1.65296376e+00 -1.42078829e+00 1.41828984e-01 -1.02871382e+00
6.09304428e-01 -7.18916953e-01 -7.96499252e-01 6.93048239e-01
-4.74113077e-01 3.28547001e-01 4.83925670e-01 2.61023015e-01
-9.17855263e-01 -2.98564404e-01 -1.07184935e+00 6.59356654e-01
1.48729980e+00 -4.14330631e-01 -6.10230744e-01 2.83374101e-01
8.04441273e-01 -6.97688401e-01 -9.90954697e-01 8.65354776e-01
5.40261447e-01 -5.62133014e-01 1.37663412e+00 -7.50648141e-01
3.26209217e-01 -4.13783431e-01 -2.85551488e-01 -1.25734282e+00
-5.32261968e-01 -2.61477023e-01 -3.56603563e-01 1.18996155e+00
2.38624662e-01 -3.58986288e-01 1.07801211e+00 3.62689018e-01
-1.09769680e-01 -6.40640974e-01 -1.25225997e+00 -8.17290485e-01
1.68514878e-01 -1.57449171e-01 1.80215120e-01 1.06437814e+00
-2.90279984e-01 3.93501133e-01 -5.08170366e-01 4.41928536e-01
9.03329313e-01 1.19022682e-01 1.31299698e+00 -1.51135623e+00
-7.03923404e-02 -3.16016316e-01 -7.83499479e-01 -1.32945430e+00
4.47716117e-01 -1.02950716e+00 -1.39238715e-01 -8.94451797e-01
2.09915023e-02 -5.39738119e-01 -3.37208956e-01 3.79724324e-01
-7.86309764e-02 4.92796063e-01 -4.54216115e-02 2.40377337e-01
-6.22137964e-01 1.04696071e+00 1.61164796e+00 -3.86417240e-01
-5.03554463e-01 6.11995533e-02 -6.89732671e-01 4.87942427e-01
4.22291517e-01 -5.01861811e-01 -3.63547862e-01 -2.04500616e-01
-1.43599004e-01 -1.11351587e-01 5.71667492e-01 -9.03981864e-01
-2.10679919e-01 -2.67143667e-01 1.01630199e+00 -5.41194677e-01
5.60299516e-01 -9.47680354e-01 -2.19139025e-01 3.68291587e-01
-2.91286230e-01 -5.51812708e-01 3.54171455e-01 9.71559525e-01
-3.14561091e-02 4.55494404e-01 9.19948995e-01 3.51386845e-01
-9.43422079e-01 6.75059795e-01 4.30471808e-01 1.78221360e-01
1.25519562e+00 -5.12893975e-01 -1.92226678e-01 -9.50651318e-02
-7.00848699e-01 3.21014374e-01 5.70218623e-01 8.76724780e-01
5.77116668e-01 -1.60495651e+00 -7.81851470e-01 3.53700191e-01
6.91904187e-01 9.09567997e-02 1.12452053e-01 8.70541990e-01
2.24102139e-02 2.01793194e-01 -3.64764541e-01 -1.27580035e+00
-1.16408837e+00 5.49952805e-01 3.83530110e-01 -2.02129990e-01
-6.11351728e-01 8.61265004e-01 4.96717125e-01 -5.53508520e-01
3.62623751e-01 -1.24752790e-01 -2.82700300e-01 5.74085005e-02
7.06398338e-02 1.32979184e-01 -1.55173630e-01 -7.89919853e-01
-5.94551563e-01 9.22345042e-01 -6.83043450e-02 3.86598051e-01
1.43353367e+00 -2.21605912e-01 4.21800017e-01 3.76190037e-01
1.24577487e+00 -9.45831388e-02 -2.04557419e+00 -3.25548559e-01
2.52072122e-02 -6.12259567e-01 -1.39050171e-01 -6.76677525e-01
-7.10760653e-01 4.66375142e-01 9.10102904e-01 -4.92120028e-01
8.89724731e-01 2.22026110e-01 2.59270132e-01 2.53403634e-01
5.17772198e-01 -1.09380102e+00 4.05742854e-01 4.09726322e-01
1.33416486e+00 -1.73282003e+00 2.66464567e-03 -6.79437935e-01
-7.85982192e-01 9.72198665e-01 1.07559752e+00 -6.84450090e-01
4.39364970e-01 -2.06945986e-01 3.80219556e-02 7.65007883e-02
-4.42655355e-01 -9.02905986e-02 8.92292500e-01 1.07843256e+00
1.89827636e-01 -8.17470327e-02 1.96289103e-02 5.78715444e-01
5.11559658e-02 -4.81557637e-01 -3.75538826e-01 5.29866815e-01
-3.10612708e-01 -9.42255974e-01 -5.04802763e-01 2.35316917e-01
1.71596706e-01 3.81282389e-01 -4.46851939e-01 1.13441837e+00
-2.89094131e-02 4.11858827e-01 9.28297788e-02 -4.22432840e-01
5.15263438e-01 -1.52814984e-02 7.59271502e-01 -6.32750452e-01
-2.42636025e-01 1.89033896e-01 -1.09601617e-01 -8.30574095e-01
-3.17757487e-01 -6.03455186e-01 -9.29737806e-01 2.34435767e-01
-6.50125667e-02 -3.78000915e-01 6.02709949e-01 7.32310236e-01
4.91747677e-01 3.37789804e-01 7.99750865e-01 -1.29391623e+00
-9.67858493e-01 -9.30718482e-01 -4.57562327e-01 6.95309043e-01
4.51061100e-01 -1.30489469e+00 -3.23639601e-01 -3.04907057e-02] | [7.642087936401367, -2.8075766563415527] |
509d2320-3cd9-43cf-940d-668ec2853653 | fast-non-local-neural-networks-with-spectral | null | null | https://doi.org/10.1145/3343031.3351029 | https://doi.org/10.1145/3343031.3351029 | Fast Non-Local Neural Networks with Spectral Residual Learning | Effectively modeling long-range spatial correlation is crucial in context-sensitive visual computing tasks, such as human pose estimation and video classification. Enlarging receptive field is popularly adopted in building such non-local deep networks. However, current solutions, including dilation convolution or self-attention based operators, mostly suffer from either low computational efficacy or insufficient receptive field. This paper proposes spectral residual learning (SRL), a novel network architectural design for achieving fully global receptive field. A neural block that implements SRL has three key components: a local-to-global transform that projects some ordinary local features into a spectral domain, compiled operations in the spectral domain, and a global-to-local transform that converts all data back to the original local format. We show its equivalence to conducting residual learning in some spectral domain and carefully re-formulate a variety of neural layers into their spectral forms, such as ReLU or convolutions. The benefits of SRL is three-fold: first, all operations have global receptive field, namely any update affects all image positions. This can extract richer context information in various vision tasks; Secondly, the local-to-global / global-to-local transforms in SRL are defined by bi-linear unitary matrices, which is both computation and parameter economic; Lastly, SRL is a generic formulation, here instantiated by Fourier transform and real orthogonal matrix. We conduct comprehensive evaluations on two challenging tasks, including human pose estimation from images and video classification. All experiments clearly show performance improvement by large margins in comparison with conventional non-local network designs. | ['Qi Tian', 'Lingxi Xie', 'Yadong Mu', 'Guiyu Tian', 'Lu Chi'] | 2019-10-15 | null | null | null | mm-19-proceedings-of-the-27th-acm | ['video-classification'] | ['computer-vision'] | [ 3.99835825e-01 -5.08179367e-01 -7.53381625e-02 -4.14279193e-01
-4.01272058e-01 -3.14619422e-01 3.60225052e-01 -6.52790070e-01
-5.54853916e-01 5.39485157e-01 4.63004470e-01 -2.56744176e-02
-2.26515159e-01 -6.18023813e-01 -8.20654809e-01 -9.34112728e-01
-8.40037614e-02 -4.03328061e-01 3.21493775e-01 -3.22599888e-01
1.43592998e-01 6.94061160e-01 -1.43761086e+00 4.40926194e-01
4.80494261e-01 1.25462079e+00 2.51659930e-01 6.82933629e-01
2.06690162e-01 8.11936617e-01 -5.00470579e-01 1.03818409e-01
4.62760329e-01 -2.17502445e-01 -6.66256726e-01 -1.36158869e-01
5.12640893e-01 -2.25225121e-01 -5.91321468e-01 1.07922900e+00
7.80371606e-01 3.36742640e-01 5.71654439e-01 -8.11243117e-01
-1.12557244e+00 1.35426506e-01 -7.45074928e-01 3.80225152e-01
2.86573976e-01 1.46932974e-01 7.81353891e-01 -1.21483839e+00
4.08176064e-01 1.42156255e+00 9.27998245e-01 4.10936266e-01
-1.03312337e+00 -5.99376678e-01 1.81654111e-01 2.13735133e-01
-1.67355263e+00 -1.63218871e-01 7.89419591e-01 -3.47478300e-01
1.13778126e+00 4.44640040e-01 5.76358557e-01 1.04729688e+00
4.84565049e-01 5.51144838e-01 1.26465714e+00 -3.02800357e-01
-1.07794695e-01 -3.99886370e-01 -1.10060960e-01 4.85926837e-01
-7.03508630e-02 2.62943268e-01 -7.93436229e-01 2.48047695e-01
1.25651646e+00 3.04065317e-01 -5.43314338e-01 -2.79251158e-01
-1.30224323e+00 5.85900247e-01 9.18005049e-01 2.81007677e-01
-3.33644241e-01 3.63718748e-01 3.52862179e-01 3.22841406e-01
3.37884694e-01 2.47939393e-01 -4.41885769e-01 3.52348626e-01
-7.48794973e-01 6.40536994e-02 1.77438349e-01 6.87430322e-01
7.53170431e-01 2.27070078e-01 -3.59281510e-01 9.49551046e-01
1.91282943e-01 5.76804459e-01 8.29907537e-01 -7.62231469e-01
3.55349779e-01 2.77796656e-01 -3.04124862e-01 -1.29724872e+00
-6.70032978e-01 -6.72483623e-01 -1.36518931e+00 1.43673852e-01
-4.28219587e-02 -1.30799875e-01 -9.28017616e-01 1.97301269e+00
1.30345985e-01 2.08391160e-01 -8.52340553e-03 1.26076686e+00
1.13840926e+00 8.32514882e-01 -6.00887761e-02 -3.09813917e-01
1.35898912e+00 -9.98549283e-01 -6.98653758e-01 -3.94440770e-01
2.02887908e-01 -7.75683641e-01 1.01234984e+00 1.52509868e-01
-8.59709263e-01 -1.23742557e+00 -1.20526278e+00 -4.09210116e-01
-4.24851716e-01 2.60498047e-01 8.44199002e-01 3.13116997e-01
-1.29220462e+00 4.87203032e-01 -4.66068834e-01 -2.55018443e-01
1.36446252e-01 5.31984627e-01 -4.42208380e-01 1.56229660e-01
-1.27259922e+00 6.23465955e-01 3.89864951e-01 3.37880224e-01
-4.27041620e-01 -4.96729374e-01 -1.02726793e+00 2.44827773e-02
3.77055198e-01 -5.88982582e-01 7.91806996e-01 -9.02897775e-01
-1.41346407e+00 6.96004629e-01 -1.45794958e-01 -3.49147618e-01
1.17496572e-01 -2.54548848e-01 -6.19621098e-01 -7.61208609e-02
-4.59336117e-03 8.15021873e-01 1.20201957e+00 -7.45516777e-01
-4.76740927e-01 -3.47164720e-01 2.56417748e-02 5.86303532e-01
-4.66827244e-01 7.00865639e-03 -5.27231753e-01 -1.18266797e+00
4.31908458e-01 -8.72488022e-01 -1.94927752e-01 -8.42277408e-02
-6.67248527e-03 -3.03938568e-01 9.41220343e-01 -6.94826424e-01
1.43214512e+00 -2.21572518e+00 2.48709112e-01 2.62416512e-01
2.66258121e-01 3.72749418e-01 -1.96185455e-01 9.45776701e-02
-5.92678428e-01 -2.05530733e-01 1.72662828e-02 -1.97941717e-02
-3.89337867e-01 -1.81545299e-02 -3.37191910e-01 4.95336115e-01
2.84539342e-01 1.09109867e+00 -6.64220035e-01 -2.67443150e-01
2.70136595e-01 7.20695138e-01 -5.63350618e-01 -7.04481527e-02
3.42056572e-01 4.49050456e-01 -1.95183381e-01 4.96041477e-01
7.83232391e-01 -1.71950504e-01 -6.88554421e-02 -7.29882360e-01
-2.28570983e-01 3.29604149e-02 -1.30534363e+00 1.79081178e+00
-4.34339553e-01 6.13050699e-01 -2.55315937e-02 -1.01485968e+00
1.01887488e+00 -9.43225697e-02 4.57025170e-01 -9.53573048e-01
3.62314694e-02 6.86828792e-02 -5.69823906e-02 -4.22032356e-01
5.19699216e-01 7.24454597e-02 -5.70490249e-02 1.53615758e-01
1.77151442e-01 3.88523310e-01 -6.50013462e-02 -2.74818450e-01
7.19112813e-01 3.25129896e-01 4.68191534e-01 -4.48203474e-01
7.52026320e-01 -6.90461934e-01 6.62243545e-01 4.44449723e-01
-1.49804816e-01 7.87972450e-01 1.68519557e-01 -7.76588023e-01
-7.53157079e-01 -9.53648925e-01 -1.18558131e-01 1.43391550e+00
2.83395320e-01 -3.92447978e-01 -6.13966286e-01 -4.13076937e-01
-2.16895938e-01 -8.83696377e-02 -6.69439375e-01 -4.58486080e-01
-8.20735097e-01 -6.18070006e-01 3.72457594e-01 8.66212845e-01
1.04741108e+00 -1.21520281e+00 -7.16081262e-01 -2.07689591e-02
-9.29794684e-02 -8.87155116e-01 -1.06238508e+00 3.02315801e-01
-6.98614299e-01 -8.13199759e-01 -8.59903038e-01 -1.06912816e+00
6.58356071e-01 5.39592028e-01 6.12662852e-01 -1.92724109e-01
-5.84633350e-01 1.64050028e-01 -1.10953487e-01 5.25941700e-02
4.37277764e-01 -8.67108032e-02 3.15998912e-01 2.34772891e-01
3.68921548e-01 -5.49959064e-01 -9.22416151e-01 4.11214918e-01
-8.36722493e-01 -9.25208554e-02 8.96734893e-01 1.03647733e+00
6.94039345e-01 2.54094880e-02 4.34243947e-01 -5.14758527e-01
7.89693475e-01 -1.58455297e-02 -1.69074267e-01 2.00499311e-01
-1.47410944e-01 2.39545614e-01 5.57101488e-01 -6.48131907e-01
-1.12476635e+00 1.24500096e-01 -1.97357729e-01 -5.19876599e-01
1.12518132e-01 5.08143783e-01 -6.02761144e-03 -3.26072633e-01
1.00762153e+00 3.06098521e-01 -2.21978929e-02 -2.04317644e-01
2.40619019e-01 5.24792016e-01 7.11294830e-01 -4.83035922e-01
7.05924988e-01 3.15553308e-01 1.62660733e-01 -9.16820645e-01
-7.07311153e-01 -5.20123005e-01 -6.54570639e-01 -1.76399410e-01
1.00613785e+00 -9.88046825e-01 -9.20990944e-01 6.54124916e-01
-9.26473320e-01 -1.57422036e-01 -2.79253274e-01 5.79924941e-01
-3.98356646e-01 3.57195556e-01 -5.70084214e-01 -5.51269829e-01
-3.74948978e-01 -1.16416883e+00 1.02794778e+00 3.53703052e-01
-8.26118235e-03 -7.98516512e-01 -2.36177921e-01 5.63895814e-02
5.96857965e-01 4.44533676e-02 5.37547946e-01 -5.29119074e-02
-4.36046332e-01 6.05345406e-02 -4.45959926e-01 6.30149364e-01
9.45675522e-02 -3.78311932e-01 -1.07537234e+00 -6.80178165e-01
1.31326258e-01 -4.47452009e-01 1.11615026e+00 6.71295643e-01
1.47182775e+00 -1.29934415e-01 -1.46425501e-01 1.20737147e+00
1.25666118e+00 2.39822581e-01 7.48072207e-01 1.77824959e-01
9.04348850e-01 4.13178504e-01 4.14769322e-01 1.90906718e-01
1.05088554e-01 6.13692343e-01 1.81648135e-01 -3.53080302e-01
-4.22131747e-01 -1.56115770e-01 4.62560028e-01 7.89314747e-01
-5.79801619e-01 2.92733550e-01 -5.58877051e-01 1.17377587e-01
-1.83726144e+00 -1.06238449e+00 3.27724606e-01 2.33180332e+00
7.06594706e-01 -7.45390495e-03 -6.14893213e-02 -2.13285796e-02
7.58210838e-01 5.45925975e-01 -5.89258790e-01 -3.00627351e-01
-3.63936961e-01 5.03405809e-01 6.04762375e-01 3.29190493e-01
-1.45582128e+00 9.76893961e-01 6.27722073e+00 1.04123688e+00
-1.50933814e+00 6.01572469e-02 7.09942043e-01 6.45913482e-02
1.31159276e-01 -3.12541723e-01 -7.18815923e-01 1.68351188e-01
3.87398809e-01 3.47127199e-01 6.48802161e-01 8.08894694e-01
-1.14942230e-01 1.05619416e-01 -9.50612247e-01 1.56843734e+00
2.82749951e-01 -1.23055673e+00 1.51463002e-01 5.95720075e-02
7.52409816e-01 -1.62839696e-01 4.28576410e-01 4.67997462e-01
-2.18047947e-01 -1.29044104e+00 6.98395431e-01 6.00145996e-01
1.27014947e+00 -9.71730709e-01 5.81476033e-01 7.41172358e-02
-1.68579793e+00 -3.03975493e-01 -6.74320579e-01 -1.73164830e-01
-2.72213697e-01 2.58814573e-01 -1.66757330e-01 3.72116864e-01
1.17823899e+00 7.82182634e-01 -5.62648654e-01 6.45241559e-01
-5.24077266e-02 9.77862254e-02 -1.50587454e-01 -2.77681779e-02
1.87259376e-01 -2.69682914e-01 3.39146882e-01 1.30288625e+00
2.90576428e-01 9.03647766e-02 2.43662044e-01 7.09000170e-01
-6.37235343e-02 1.39323240e-02 -5.85576534e-01 3.69561523e-01
2.63767093e-01 1.27266574e+00 -5.30342579e-01 -3.48540060e-02
-4.61890757e-01 1.17295229e+00 3.04515719e-01 6.90473258e-01
-9.21901166e-01 -6.70434535e-01 7.42867470e-01 -4.02645767e-02
2.69931048e-01 -3.03578258e-01 -1.17265843e-01 -9.66868699e-01
3.02170485e-01 -9.00471270e-01 2.27625445e-01 -7.99877346e-01
-9.97460604e-01 6.46902263e-01 -1.13131396e-01 -1.39883745e+00
-1.56001389e-01 -9.63873804e-01 -4.14560109e-01 9.92488027e-01
-1.29504502e+00 -1.35308778e+00 -4.92881179e-01 1.01263249e+00
5.70055902e-01 -2.29091868e-01 4.55214977e-01 3.82363409e-01
-3.30960423e-01 8.31345022e-01 -1.84534475e-01 2.56525218e-01
8.73927653e-01 -9.20824528e-01 2.83166170e-01 1.00038803e+00
2.12655906e-02 1.07324672e+00 2.07405359e-01 -5.18108845e-01
-1.31642997e+00 -1.12803078e+00 4.25715744e-01 -7.73484632e-02
3.87730598e-01 -3.49024594e-01 -6.61840141e-01 6.43995047e-01
1.92843173e-02 5.19933522e-01 2.74528533e-01 1.20263569e-01
-4.21804816e-01 -4.81385142e-01 -7.39463806e-01 7.42703915e-01
1.47007275e+00 -8.15545499e-01 -3.81897151e-01 2.06276447e-01
6.32977307e-01 -5.93315721e-01 -6.89331472e-01 7.96208739e-01
5.91731966e-01 -1.18258810e+00 1.36956716e+00 -3.59933197e-01
3.17402869e-01 -5.67141771e-01 -3.70534927e-01 -1.13806581e+00
-9.39835966e-01 -5.78350544e-01 -6.65553054e-03 8.67997527e-01
2.20283270e-02 -7.40973413e-01 4.90184605e-01 -2.90481523e-02
-3.14578593e-01 -9.27162349e-01 -8.77099097e-01 -6.60804987e-01
-1.49595797e-01 -3.49641144e-01 3.68652433e-01 7.71678269e-01
-3.83186966e-01 5.92444003e-01 -6.97145462e-01 1.62140459e-01
3.57781112e-01 -5.50954193e-02 5.50488293e-01 -9.39840496e-01
-4.14318472e-01 -5.14482856e-01 -6.58109248e-01 -1.54684401e+00
-1.36814341e-01 -6.71997190e-01 3.79765034e-02 -1.22455013e+00
3.01092982e-01 -2.49591529e-01 -4.38587070e-01 4.31885839e-01
-1.34820193e-01 5.34595966e-01 3.14597297e-03 3.06752264e-01
-4.14163977e-01 4.56684828e-01 1.51325369e+00 -1.11205876e-01
-3.50866467e-01 -6.88171834e-02 -5.92432261e-01 8.16042542e-01
5.78462124e-01 2.10698992e-01 -6.27524793e-01 -3.21130037e-01
2.17365116e-01 -1.98616818e-01 4.88375068e-01 -1.37863123e+00
3.94259334e-01 -1.79619685e-01 8.37774634e-01 -5.86000681e-01
5.02774119e-01 -5.90535879e-01 3.66275124e-02 4.11608487e-01
-2.99062431e-01 1.42364100e-01 5.43169677e-02 4.96521860e-01
-4.57111210e-01 1.52097374e-01 8.91882300e-01 -9.76962596e-02
-1.25694668e+00 4.72726136e-01 7.25516910e-03 -1.09537385e-01
1.05521500e+00 -4.58990335e-01 -2.16481179e-01 -3.83942544e-01
-7.50799656e-01 -1.88814595e-01 1.37350699e-02 5.30779660e-01
8.68500650e-01 -1.70915532e+00 -4.76860464e-01 6.43345416e-01
7.41980446e-04 1.17891282e-01 5.80525696e-01 9.81414139e-01
-6.34763777e-01 6.51786208e-01 -4.32946056e-01 -7.55411386e-01
-1.27683270e+00 7.54263103e-01 4.49985981e-01 -7.79989660e-02
-5.63895166e-01 1.19502842e+00 8.46917570e-01 -4.41214204e-01
3.34040225e-01 -5.11477470e-01 -3.01731288e-01 -4.14556116e-02
6.25497162e-01 2.18532160e-01 1.42724607e-02 -8.59456241e-01
-4.31885064e-01 1.05736542e+00 1.02523006e-01 2.17957437e-01
1.11549914e+00 -4.73896787e-02 -2.64858961e-01 8.71402249e-02
1.46539629e+00 -1.74010731e-02 -1.48194957e+00 -5.12183130e-01
-4.23318177e-01 -3.57004791e-01 1.27383277e-01 -4.36352074e-01
-1.27672172e+00 8.73573005e-01 8.25755537e-01 -7.39879981e-02
1.52372873e+00 -1.23772152e-01 5.01345158e-01 4.96055394e-01
3.36527765e-01 -1.30396354e+00 5.08214831e-01 7.98573613e-01
1.28025436e+00 -1.04330456e+00 1.15358792e-01 -4.33834940e-01
-5.00250101e-01 1.14346218e+00 9.05711055e-01 -2.17208147e-01
6.83197796e-01 2.04907224e-01 -2.47581899e-02 -7.14223385e-02
-3.37802738e-01 -3.12937468e-01 7.59327590e-01 6.52652860e-01
5.84445238e-01 -1.59212295e-02 1.63895544e-02 3.76260221e-01
-2.29798898e-01 -2.58991122e-01 -1.69066161e-01 8.64489734e-01
-4.10905033e-01 -6.89403236e-01 -5.50736189e-01 2.33280033e-01
-2.36572072e-01 -2.47081906e-01 -6.79179057e-02 6.85633302e-01
5.92211306e-01 5.52127600e-01 1.74092218e-01 -7.14751422e-01
1.88779205e-01 -2.28791460e-01 4.00646269e-01 -4.30721164e-01
-3.29660952e-01 4.11985368e-01 -3.69747937e-01 -7.53966928e-01
-5.56962252e-01 -3.19531798e-01 -1.21256673e+00 -2.37988755e-01
-9.37441066e-02 -4.23305750e-01 4.11968470e-01 7.67497361e-01
2.47239262e-01 9.04288411e-01 5.26122570e-01 -1.13681996e+00
-4.74480659e-01 -1.02188778e+00 -4.72225577e-01 4.47895288e-01
4.17562217e-01 -6.69196129e-01 -1.06908612e-01 3.22261900e-02] | [10.698599815368652, -1.6603575944900513] |
6b3489b6-6653-418b-96be-6c3c4d43e577 | domain-specific-pretraining-improves | 2302.09833 | null | https://arxiv.org/abs/2302.09833v2 | https://arxiv.org/pdf/2302.09833v2.pdf | Domain-Specific Pre-training Improves Confidence in Whole Slide Image Classification | Whole Slide Images (WSIs) or histopathology images are used in digital pathology. WSIs pose great challenges to deep learning models for clinical diagnosis, owing to their size and lack of pixel-level annotations. With the recent advancements in computational pathology, newer multiple-instance learning-based models have been proposed. Multiple-instance learning for WSIs necessitates creating patches and uses the encoding of these patches for diagnosis. These models use generic pre-trained models (ResNet-50 pre-trained on ImageNet) for patch encoding. The recently proposed KimiaNet, a DenseNet121 model pre-trained on TCGA slides, is a domain-specific pre-trained model. This paper shows the effect of domain-specific pre-training on WSI classification. To investigate the effect of domain-specific pre-training, we considered the current state-of-the-art multiple-instance learning models, 1) CLAM, an attention-based model, and 2) TransMIL, a self-attention-based model, and evaluated the models' confidence and predictive performance in detecting primary brain tumors - gliomas. Domain-specific pre-training improves the confidence of the models and also achieves a new state-of-the-art performance of WSI-based glioma subtype classification, showing a high clinical applicability in assisting glioma diagnosis. We will publicly share our code and experimental results at https://github.com/soham-chitnis10/WSI-domain-specific. | ['Ashwin Srinivasan', 'Lovekesh Vig', 'Shlomo Berkovsky', 'Antonio Di Ieva', 'Tanmay Tulsidas Verlekar', 'Tirtharaj Dash', 'Sidong Liu', 'Soham Rohit Chitnis'] | 2023-02-20 | null | null | null | null | ['whole-slide-images', 'multiple-instance-learning'] | ['computer-vision', 'methodology'] | [ 2.82437414e-01 3.54020059e-01 -2.50731468e-01 -2.83787817e-01
-1.33849287e+00 1.18368194e-01 4.84476298e-01 2.56262451e-01
-5.98559380e-01 6.59849405e-01 1.58412576e-01 -4.86734986e-01
-1.39751926e-01 -8.11124921e-01 -6.16570711e-01 -1.11742914e+00
-6.40028566e-02 6.75343812e-01 5.48268914e-01 -1.58726007e-01
3.65780815e-02 4.14831996e-01 -1.33251452e+00 8.81082118e-01
8.64095271e-01 1.16235328e+00 6.64719403e-01 7.88970113e-01
-3.41005504e-01 8.72796178e-01 -5.25622368e-01 -1.90573037e-01
-8.53304714e-02 -1.72731504e-01 -1.00330758e+00 -2.25939304e-01
3.77957821e-01 -1.56715438e-02 -4.81179595e-01 9.86576557e-01
7.22577393e-01 -4.69253898e-01 7.25008965e-01 -8.53911042e-01
-4.46458757e-01 4.66084838e-01 -5.01269937e-01 6.00002468e-01
-1.61339447e-01 3.29188496e-01 7.59220660e-01 -5.94270885e-01
7.08434284e-01 4.87020373e-01 9.46421266e-01 6.99784994e-01
-7.28653729e-01 -7.43176401e-01 -9.83013734e-02 6.15606964e-01
-1.59328949e+00 -1.52636945e-01 3.52993429e-01 -4.48903382e-01
1.20482922e+00 4.17375296e-01 8.30157936e-01 1.04945362e+00
6.54198050e-01 9.62872446e-01 1.36798251e+00 -4.62594211e-01
4.60181274e-02 -5.47948442e-02 2.52152085e-01 7.68863916e-01
1.47543577e-02 -5.56168333e-02 -9.46397930e-02 3.80410515e-02
6.99861825e-01 3.16813231e-01 -5.41308820e-01 3.07997376e-01
-1.18500507e+00 5.78452826e-01 8.42477500e-01 8.00154984e-01
-1.75302282e-01 1.52563706e-01 4.73812848e-01 1.13588937e-01
7.73671627e-01 3.24183702e-01 -5.26371777e-01 3.25450748e-01
-1.07385576e+00 -1.03379548e-01 3.50468755e-01 5.44758141e-01
3.98467571e-01 -3.65806758e-01 -5.22914588e-01 9.90866125e-01
3.44538875e-02 -1.50704198e-02 1.41097558e+00 9.49183479e-02
1.60109475e-01 9.00803447e-01 -5.16480267e-01 -2.57452101e-01
-9.92086589e-01 -1.07581496e+00 -9.79881465e-01 1.42606497e-01
2.92374045e-01 2.38305882e-01 -1.47692597e+00 1.37687385e+00
6.98603988e-02 7.66781807e-01 1.29402056e-01 6.90446675e-01
1.32373083e+00 3.53934377e-01 1.82946593e-01 1.36060029e-01
1.45746386e+00 -1.15734959e+00 -4.77807164e-01 -7.64857084e-02
1.17640924e+00 -3.99525940e-01 8.90127063e-01 1.81020752e-01
-8.24630082e-01 -2.19917521e-01 -9.23535883e-01 -1.74923744e-02
-8.08881700e-01 3.27394068e-01 4.08813089e-01 4.03228164e-01
-1.46494043e+00 3.66221309e-01 -9.87256169e-01 -5.10223269e-01
8.40596855e-01 5.71524262e-01 -6.31052613e-01 -1.69652984e-01
-9.83167410e-01 1.03291082e+00 3.00189763e-01 -2.16454178e-01
-1.23993194e+00 -1.27438021e+00 -4.19701099e-01 1.44124404e-01
-4.74580340e-02 -8.17099929e-01 1.24913251e+00 -1.05047309e+00
-1.26257110e+00 1.45551562e+00 -7.79051483e-02 -5.94093621e-01
4.87546086e-01 4.59989369e-01 -5.15353918e-01 1.85147196e-01
5.66355288e-02 7.60474384e-01 4.15433705e-01 -8.87255669e-01
-7.22509503e-01 -4.65252101e-01 -1.60968453e-02 1.54471353e-01
-3.04439485e-01 -3.34911644e-01 -4.44358408e-01 -5.79657853e-01
-6.83673918e-02 -8.14837635e-01 -4.73437637e-01 -7.87310451e-02
-4.92726564e-01 -2.79195964e-01 5.30607760e-01 -8.37940753e-01
8.47662032e-01 -2.09443426e+00 -1.06202014e-01 1.11383468e-01
3.17702770e-01 5.68898559e-01 -4.17665809e-01 -9.84071521e-04
-3.70538116e-01 2.52936780e-01 -2.29192108e-01 -4.36846256e-01
-4.20763284e-01 -4.99364641e-03 4.51065123e-01 4.87825632e-01
3.64562809e-01 9.53131974e-01 -7.49732971e-01 -4.34187502e-01
1.59732640e-01 4.97992158e-01 -3.90643060e-01 3.20086032e-01
-3.43966410e-02 3.74567389e-01 -1.02953196e-01 5.68943143e-01
5.83604932e-01 -4.20345038e-01 -2.11709529e-01 -2.70607919e-01
9.53633636e-02 1.55049860e-01 -4.71350938e-01 1.79604888e+00
-4.97836620e-01 5.55102348e-01 -1.16694346e-01 -1.23824048e+00
5.78473747e-01 4.96830434e-01 5.60516417e-01 -8.08938205e-01
2.27459684e-01 4.47831452e-01 2.34318197e-01 -7.50237703e-01
-1.58314884e-01 -1.27766311e-01 5.06535411e-01 -5.21927886e-02
3.82762015e-01 1.76399834e-02 2.23259568e-01 2.60855034e-02
1.64454234e+00 -4.61827785e-01 4.00754541e-01 -4.45073426e-01
5.95296741e-01 9.62299258e-02 4.64090466e-01 5.41675627e-01
-1.90166757e-01 9.99584079e-01 3.30849916e-01 -4.32953566e-01
-6.67759597e-01 -7.31640220e-01 -6.13711417e-01 6.92460656e-01
-2.64920950e-01 -9.74980593e-02 -5.99066615e-01 -7.66259134e-01
-1.50115997e-01 3.92281741e-01 -1.11990392e+00 -1.32857487e-01
-1.22732356e-01 -1.29001248e+00 5.27498603e-01 5.97503841e-01
4.64017898e-01 -8.13876510e-01 -3.29746962e-01 2.70654529e-01
-1.11885667e-01 -8.45725000e-01 -8.62024799e-02 4.70954418e-01
-7.99290717e-01 -1.25985861e+00 -1.27104616e+00 -1.01350737e+00
1.14476991e+00 2.56311856e-02 9.69836891e-01 3.86520267e-01
-7.67984271e-01 2.93301702e-01 -4.49923545e-01 -7.94781148e-01
-4.23811823e-01 3.17947924e-01 -6.44448757e-01 6.56104386e-02
4.42457318e-01 -3.50343585e-01 -8.22644055e-01 1.72628626e-01
-9.87950802e-01 4.00893122e-01 1.14127457e+00 1.13988161e+00
8.87213588e-01 -1.75097182e-01 5.60619414e-01 -1.13607895e+00
3.57798696e-01 -5.43634593e-01 -1.15718827e-01 3.63651931e-01
-1.93008408e-01 -2.53736973e-01 4.08747882e-01 -2.43836418e-01
-7.83854008e-01 -1.63125247e-01 -7.74844885e-01 -3.23527396e-01
-4.36642826e-01 8.57287467e-01 2.07708508e-01 -3.33272517e-01
8.41355324e-01 2.59421885e-01 2.99112201e-01 -1.51325271e-01
-4.52597558e-01 7.28971601e-01 2.53217131e-01 -5.26842251e-02
1.72654435e-01 4.82836545e-01 -9.93174464e-02 -6.68421328e-01
-8.95827115e-01 -9.44627821e-01 -3.69749904e-01 -1.17348038e-01
8.94130170e-01 -9.95377541e-01 -2.03271270e-01 8.61646116e-01
-8.39800119e-01 -8.45549762e-01 -2.37094238e-01 3.54801148e-01
-3.45232397e-01 -1.58692122e-01 -8.36133420e-01 -2.12945715e-01
-7.00473964e-01 -1.45007408e+00 1.08476281e+00 2.15355992e-01
9.67026651e-02 -1.13374209e+00 3.45706642e-01 3.27382028e-01
7.83713877e-01 1.71073049e-01 9.44089234e-01 -9.90634382e-01
-2.96035677e-01 -4.02010739e-01 -3.76375914e-01 1.86503500e-01
4.32016365e-02 -1.50060013e-01 -1.33608127e+00 -4.10384089e-01
-4.91476357e-01 -1.99996158e-01 1.27785027e+00 6.90761387e-01
1.60539114e+00 -5.78135140e-02 -8.62636387e-01 7.98984408e-01
1.73469186e+00 1.96408391e-01 8.65274727e-01 4.82972294e-01
4.43860888e-01 2.52394229e-01 1.00081414e-01 2.08681032e-01
4.72974211e-01 3.48733425e-01 6.91614330e-01 -6.44850075e-01
-5.95948040e-01 1.73684478e-01 -1.25573680e-01 6.60163939e-01
-2.73117423e-01 -2.08794087e-01 -1.31270719e+00 8.52738500e-01
-1.52192438e+00 -6.97096348e-01 3.32779028e-02 1.87586379e+00
9.81868386e-01 -2.00672764e-02 -3.06086421e-01 3.43675196e-01
4.82229084e-01 -1.93519264e-01 -4.14958298e-01 -5.58272749e-02
-3.50620784e-02 5.55078387e-01 5.46808541e-01 1.92161545e-01
-1.13153017e+00 7.00438857e-01 5.60605383e+00 1.28989708e+00
-1.52790403e+00 6.28036022e-01 1.17800391e+00 -1.34805977e-01
-6.89298436e-02 -5.73175371e-01 -8.44639480e-01 4.40764785e-01
9.79526281e-01 2.46449426e-01 -3.17361146e-01 7.49432802e-01
-2.53320131e-02 -1.39960796e-01 -8.56664002e-01 7.64847934e-01
8.78491476e-02 -1.82243824e+00 -5.35942279e-02 1.89288244e-01
7.44883060e-01 5.71305573e-01 1.58564836e-01 5.05959749e-01
7.75964931e-02 -9.87571478e-01 2.42872581e-01 5.34430981e-01
1.12640321e+00 -4.80352640e-01 1.51476216e+00 5.93527518e-02
-7.86812425e-01 -4.48051617e-02 -5.67524076e-01 3.90329123e-01
-3.46342713e-01 8.85485113e-01 -1.31316662e+00 5.25928438e-01
7.72652984e-01 7.96715558e-01 -9.01302159e-01 1.56692100e+00
-6.15162812e-02 9.36562181e-01 -2.11990058e-01 -1.11135095e-02
3.54185462e-01 4.36462313e-01 1.04764655e-01 1.55922115e+00
5.42918622e-01 1.02237269e-01 2.07937509e-02 5.49123168e-01
1.67652994e-01 1.88871458e-01 -9.24044773e-02 1.87730566e-01
-5.00640552e-03 1.52526915e+00 -8.30615342e-01 -4.41458583e-01
-4.73424852e-01 5.85481942e-01 4.08166707e-01 3.25044930e-01
-6.85562670e-01 -2.47102067e-01 5.22600532e-01 5.36699176e-01
2.00576290e-01 4.03339893e-01 -2.89802581e-01 -9.43148911e-01
-4.94821370e-01 -6.37018740e-01 5.94080210e-01 -7.37449586e-01
-1.36754096e+00 9.10007060e-01 -3.78348142e-01 -1.16442144e+00
1.86349630e-01 -9.19450104e-01 -1.04968119e+00 6.47318661e-01
-1.99082649e+00 -1.50218606e+00 -5.88853359e-01 6.91209972e-01
4.33732510e-01 -3.94036025e-01 1.12393522e+00 2.27072060e-01
-6.43271387e-01 9.70417559e-01 1.31737024e-01 2.09921643e-01
7.21138597e-01 -1.27322149e+00 -1.32267764e-02 3.54380667e-01
-2.71344334e-01 1.42588630e-01 2.91987568e-01 -2.93818861e-01
-7.58049190e-01 -1.39147890e+00 5.86918890e-01 -1.12696541e-02
6.98547900e-01 4.02425341e-02 -9.23094988e-01 6.76708221e-01
2.63045430e-01 2.83032119e-01 1.28728521e+00 -7.14227855e-02
-5.13806976e-02 -2.53584862e-01 -1.43922663e+00 3.03652138e-01
8.99359584e-01 -2.53748983e-01 -1.92249656e-01 6.00790322e-01
4.76980895e-01 -5.94122946e-01 -1.07831502e+00 6.87758923e-01
2.43826970e-01 -8.47314298e-01 8.25869799e-01 -3.83615762e-01
4.30250049e-01 -1.23827413e-01 1.76574439e-01 -1.68588221e+00
-6.80161774e-01 3.06847334e-01 4.41289544e-01 7.95718908e-01
8.48043561e-01 -9.57106113e-01 8.63832057e-01 3.43996912e-01
-8.30637693e-01 -1.45748150e+00 -1.05779481e+00 -5.04487336e-01
2.85450161e-01 -1.78287923e-01 6.02338910e-01 1.01392078e+00
1.21038541e-01 -2.80285478e-01 4.73207265e-01 2.47174546e-01
1.24147221e-01 -3.47034872e-01 2.33123064e-01 -9.51771915e-01
-3.43538463e-01 -7.83711791e-01 -1.09239924e+00 -2.84042895e-01
6.00051954e-02 -1.29679465e+00 -2.72384405e-01 -1.91656232e+00
3.53640527e-01 -5.59805155e-01 -8.43273461e-01 8.99194181e-01
-2.78272569e-01 4.34296250e-01 -1.61739916e-01 1.90690190e-01
-4.87850517e-01 2.26371393e-01 1.21300685e+00 -6.23646259e-01
2.28727043e-01 -6.85290480e-03 -4.98857468e-01 4.14320618e-01
8.87839675e-01 -2.84062237e-01 -2.33945370e-01 -3.28331530e-01
-1.98730648e-01 -1.46522254e-01 4.16849196e-01 -1.60350287e+00
3.77765149e-01 1.95505664e-01 5.41033804e-01 -5.97097099e-01
2.44868219e-01 -5.54579914e-01 2.25225054e-02 7.31478512e-01
-3.18746239e-01 -2.82408834e-01 4.68466580e-01 2.87674934e-01
-5.13618708e-01 -4.17324215e-01 9.01137769e-01 -4.37537819e-01
-7.35285997e-01 7.72222877e-01 -6.14508808e-01 -3.11932534e-01
1.14346826e+00 -2.44874716e-01 -6.60617948e-01 1.68725476e-01
-9.59925354e-01 -1.53488573e-02 1.85148403e-01 -4.62576328e-03
5.30707121e-01 -9.51608241e-01 -9.56407785e-01 1.54046819e-01
4.34982598e-01 3.34479600e-01 7.38120675e-01 1.32975471e+00
-7.08258510e-01 4.45360720e-01 -3.27934891e-01 -8.16415191e-01
-1.09826601e+00 2.60431230e-01 7.33199537e-01 -9.25840914e-01
-3.95313621e-01 1.44050539e+00 5.46342969e-01 -4.67020243e-01
1.64173558e-01 -7.13030398e-01 -3.70806575e-01 -2.79806912e-01
7.67469645e-01 -1.03499182e-01 7.20084012e-01 -2.18711928e-01
-2.80639559e-01 3.63535345e-01 -5.84509432e-01 2.80823588e-01
1.37420166e+00 4.99244511e-01 -1.03044987e-01 1.16515942e-01
1.33523285e+00 -6.52180731e-01 -8.80220413e-01 -3.02665442e-01
-2.14406520e-01 1.41411880e-02 5.16366124e-01 -1.24417651e+00
-1.45060551e+00 9.35370803e-01 1.09699762e+00 -1.74763009e-01
1.25829303e+00 9.36491229e-03 5.93512297e-01 -5.73026426e-02
3.76059711e-01 -6.96353555e-01 6.38479972e-03 3.35234016e-01
8.70063901e-01 -1.34417093e+00 -3.67778808e-01 -3.50107402e-01
-3.17166209e-01 9.78373051e-01 7.64836013e-01 -8.48376602e-02
9.19330597e-01 4.59601939e-01 1.86986521e-01 -1.84025794e-01
-9.06982720e-01 -4.37997937e-01 3.20556223e-01 7.49417126e-01
6.81958199e-01 4.29736316e-01 -2.02070430e-01 8.85611594e-01
-9.22919065e-02 2.91550010e-01 3.70614320e-01 8.14505219e-01
-3.94794762e-01 -9.03997838e-01 -7.45728686e-02 1.06777024e+00
-5.24100482e-01 -2.35223874e-01 -2.85049796e-01 7.44103074e-01
2.53650337e-01 4.21418786e-01 8.35322365e-02 -3.63142997e-01
1.27655968e-01 1.07215345e-02 3.76323849e-01 -6.85838640e-01
-7.98306823e-01 -1.11112364e-01 -9.17062163e-03 -2.86794096e-01
-3.60597521e-01 -3.16578388e-01 -1.23924708e+00 3.40230651e-02
-4.21336114e-01 1.41514456e-02 7.30345488e-01 8.85715723e-01
4.84133184e-01 1.06230104e+00 1.53597310e-01 -7.84986019e-01
-7.69562498e-02 -1.31079626e+00 -4.82165813e-01 2.28531078e-01
1.50701091e-01 -6.22624278e-01 -3.62345755e-01 -2.34576568e-01] | [15.083391189575195, -2.822807550430298] |
41d35dac-005b-4229-a751-03f588d74443 | orthogonal-transform-based-generative | 2206.01743 | null | https://arxiv.org/abs/2206.01743v1 | https://arxiv.org/pdf/2206.01743v1.pdf | Orthogonal Transform based Generative Adversarial Network for Image Dehazing | Image dehazing has become one of the crucial preprocessing steps for any computer vision task. Most of the dehazing methods try to estimate the transmission map along with the atmospheric light to get the dehazed image in the image domain. In this paper, we propose a novel end-to-end architecture that directly estimates dehazed image in Krawtchouk transform domain. For this a customized Krawtchouk Convolution Layer (KCL) in the architecture is added. KCL is constructed using Krawtchouk basis functions which converts the image from the spatial domain to the Krawtchouk transform domain. Another convolution layer is added at the end of the architecture named as Inverse Krawtchouk Convolution Layer (IKCL) which converts the image back to the spatial domain from the transform domain. It has been observed that the haze is mainly present in lower frequencies of hazy images, wherein the Krawtchouk transform helps to analyze the high and low frequencies of the images separately. We have divided our architecture into two branches, the upper branch deals with the higher frequencies while the lower branch deals with the lower frequencies of the image. The lower branch is made deeper in terms of the layers as compared to the upper branch to address the haze present in the lower frequencies. Using the proposed Orthogonal Transform based Generative Adversarial Network (OTGAN) architecture for image dehazing, we were able to achieve competitive results when compared to the present state-of-the-art methods. | ['Vijeta Khare', 'Manish Khare', 'Mantra Sanathra', 'Ahlad Kumar'] | 2022-06-03 | null | null | null | null | ['image-dehazing'] | ['computer-vision'] | [ 2.96238512e-01 1.84094638e-01 7.65825868e-01 3.77524160e-02
-2.09916279e-01 -2.08265349e-01 5.00484943e-01 -5.36432862e-01
-4.47531551e-01 5.77669322e-01 -2.59379875e-02 -2.48638630e-01
-6.93506822e-02 -1.27791786e+00 -7.40093887e-01 -1.14447379e+00
2.22062603e-01 -1.83599204e-01 5.23272872e-01 -5.05976379e-01
5.10090701e-02 2.07930699e-01 -1.61489749e+00 5.54650962e-01
7.98827529e-01 1.01202619e+00 2.43836101e-02 9.98034120e-01
5.16160689e-02 7.00248182e-01 -8.08532238e-01 -2.55161107e-01
4.86432254e-01 -6.79975092e-01 -4.19250786e-01 -7.91897625e-02
5.30592203e-01 -3.42535049e-01 -3.77281278e-01 1.28575063e+00
4.19628561e-01 1.74142495e-01 8.12650621e-01 -1.01987815e+00
-7.98970222e-01 1.97920665e-01 -6.09511614e-01 2.69159675e-01
-2.45153308e-01 -9.54042841e-03 3.77735645e-01 -7.75477409e-01
3.80904704e-01 9.01598215e-01 7.42931008e-01 2.73224860e-01
-8.21650803e-01 -7.33750343e-01 -3.70256424e-01 6.01704955e-01
-1.49337542e+00 -1.29256379e-02 1.01213992e+00 -3.73926103e-01
7.59887516e-01 2.03938231e-01 5.02222538e-01 4.48363304e-01
4.84563857e-01 2.21557066e-01 1.22371292e+00 -7.83141673e-01
-3.64227518e-02 1.97988391e-01 -8.99002776e-02 4.07026112e-01
1.41125787e-02 3.52359891e-01 -1.08085558e-01 4.12137359e-01
6.28912985e-01 2.52114013e-02 -3.83521020e-01 1.94517151e-01
-7.00167537e-01 9.08746064e-01 7.87318766e-01 6.46745741e-01
-5.10431468e-01 4.59035605e-01 -9.31638032e-02 3.77853632e-01
4.52213407e-01 1.70996711e-01 2.35960539e-02 4.84114796e-01
-1.23099697e+00 1.94975510e-01 4.27097291e-01 6.56185448e-01
1.04342365e+00 4.14985538e-01 -5.30732125e-02 5.36697447e-01
3.19944024e-01 5.00295937e-01 3.49332064e-01 -6.05468929e-01
3.11841935e-01 2.75316596e-01 -2.30921302e-02 -8.51010680e-01
-6.71125725e-02 -5.51038146e-01 -1.05553675e+00 8.16293895e-01
2.88185835e-01 -4.81112450e-01 -1.32054436e+00 1.15941930e+00
2.29039699e-01 5.63963950e-01 2.84086883e-01 9.67702925e-01
7.99167037e-01 1.20690310e+00 -2.67491221e-01 2.14751899e-01
1.52503192e+00 -9.99849617e-01 -7.22741723e-01 -2.66667772e-02
5.18695004e-02 -1.21891880e+00 5.73310018e-01 3.80867034e-01
-1.01899302e+00 -7.77103543e-01 -1.34550047e+00 -1.93289444e-01
-7.50972569e-01 1.68674901e-01 -1.75460912e-02 7.90325284e-01
-1.44758618e+00 3.97626996e-01 -4.46364343e-01 -1.60728227e-02
1.54022267e-02 1.94223747e-01 -2.36566931e-01 -8.09083506e-02
-1.57364357e+00 1.00122511e+00 5.49306273e-01 3.11663896e-01
-6.88077211e-01 -6.75840855e-01 -8.40579689e-01 1.86867222e-01
3.11096273e-02 -5.58385491e-01 5.55098057e-01 -1.12070632e+00
-1.52619243e+00 6.24543071e-01 1.06028095e-01 -6.85971916e-01
4.60886478e-01 -5.38881049e-02 -5.31215131e-01 3.86835337e-01
-3.46424669e-01 4.63126719e-01 1.71092737e+00 -1.19868886e+00
-6.64276540e-01 3.61591950e-02 -8.06815177e-02 3.07174772e-02
-1.24326363e-01 -3.37280869e-01 -1.33052722e-01 -1.20559549e+00
-2.49926709e-02 -7.14169204e-01 2.34705091e-01 -2.84647822e-01
-1.94690198e-01 2.21241027e-01 1.37957370e+00 -1.10179424e+00
1.08414948e+00 -2.43871951e+00 1.56141490e-01 2.46702239e-01
2.84255236e-01 5.80462217e-01 1.44044831e-01 5.48119068e-01
-4.97327954e-01 -2.69561231e-01 -3.32837701e-01 -2.15384528e-01
-2.94145674e-01 -5.31913564e-02 -3.74271363e-01 5.59088945e-01
3.51362109e-01 5.49524128e-01 -4.75518852e-01 -8.42272043e-02
5.78781009e-01 9.49734747e-01 -5.81639469e-01 2.79231995e-01
-4.53550406e-02 3.84540468e-01 -2.15899702e-02 6.91472553e-03
1.29819596e+00 4.46125567e-01 -6.42747998e-01 -4.49048162e-01
-4.15428221e-01 -1.92400023e-01 -1.00498593e+00 1.06781852e+00
-6.61022186e-01 9.74359751e-01 2.28761584e-01 -1.01833034e+00
8.44752729e-01 5.42045236e-01 1.55978873e-01 -4.60133791e-01
2.98112929e-01 1.32528335e-01 5.56820258e-02 -4.58574653e-01
4.28285807e-01 -5.82378864e-01 4.63214397e-01 1.66902825e-01
1.86855644e-01 -5.80860853e-01 -1.87988296e-01 -2.09918275e-01
7.14234829e-01 8.79445523e-02 1.48458242e-01 -2.28975102e-01
9.94843960e-01 -2.45612971e-02 -2.34719083e-01 3.42766672e-01
1.85584053e-02 8.34989607e-01 1.82579577e-01 -4.97221172e-01
-1.11457157e+00 -1.06592023e+00 -1.36160105e-02 7.38073945e-01
9.67506543e-02 6.05221540e-02 -1.11899567e+00 -2.27447987e-01
-3.33755732e-01 9.00842547e-01 -7.97825098e-01 -5.63645124e-01
-4.90601301e-01 -5.61603129e-01 6.61409914e-01 -1.29573107e-01
1.35845447e+00 -9.65117395e-01 -5.34202337e-01 2.22699381e-02
-7.78593794e-02 -1.11152232e+00 -4.63265538e-01 1.82581723e-01
-4.22880083e-01 -7.68703580e-01 -1.01529253e+00 -7.92162895e-01
4.55487698e-01 2.66979128e-01 6.31093681e-01 -9.76679102e-02
-2.98590750e-01 -6.09498732e-02 -6.31573319e-01 -8.90452087e-01
-5.13909996e-01 -1.48101762e-01 -5.88696897e-01 4.87418115e-01
3.62356454e-01 -7.89210618e-01 -7.70688713e-01 1.65439948e-01
-1.38129354e+00 -4.55608498e-03 5.99084914e-01 6.57089353e-01
2.18953669e-01 1.01583958e+00 5.80424964e-02 -7.63857007e-01
2.14401424e-01 -3.44764501e-01 -6.89211369e-01 -2.02813938e-01
-2.75470823e-01 1.57417148e-01 8.47651780e-01 -3.33522052e-01
-1.07563674e+00 -2.28717372e-01 -4.31308657e-01 -6.15167975e-01
-2.32769340e-01 8.61250907e-02 1.41740173e-01 -4.59388793e-01
7.00405896e-01 5.68449438e-01 -1.82722509e-02 -2.89824247e-01
2.22726762e-01 7.33473301e-01 7.44631350e-01 2.07061190e-02
1.41627502e+00 6.97567403e-01 8.86789411e-02 -1.22082818e+00
-5.41921973e-01 -4.56394821e-01 -4.04787183e-01 -3.78266424e-01
1.44392407e+00 -8.71404588e-01 -3.78507644e-01 7.29110301e-01
-1.10011721e+00 -2.05000371e-01 -2.07162961e-01 5.22644758e-01
-3.11077833e-01 1.85534224e-01 -5.16963959e-01 -8.79022658e-01
-4.20536935e-01 -1.16112065e+00 9.60615635e-01 3.19157720e-01
3.42194527e-01 -1.00723588e+00 -7.75430873e-02 2.53596127e-01
8.11021209e-01 3.78766805e-01 1.09216475e+00 -1.06252506e-02
-6.30104721e-01 -2.21528873e-01 -2.64194012e-01 8.40621769e-01
6.75711855e-02 -1.59055859e-01 -1.29763222e+00 -1.27516761e-01
5.23052812e-01 3.06013703e-01 1.16867721e+00 5.70575893e-01
7.18847692e-01 -6.16052076e-02 3.03503066e-01 9.55375850e-01
1.74782169e+00 2.69385636e-01 1.38420510e+00 5.75609684e-01
6.38207555e-01 5.33017695e-01 2.55123556e-01 -1.08689647e-02
-3.74878906e-02 6.19939625e-01 6.44510806e-01 -5.37464380e-01
-6.42065287e-01 5.28972894e-02 4.26883876e-01 2.75132746e-01
-3.36940140e-01 -3.91491324e-01 -4.70901579e-01 4.09805506e-01
-1.34010935e+00 -9.02406514e-01 -2.46805310e-01 2.02324581e+00
5.63184142e-01 2.84565147e-02 -1.76159024e-01 3.69001329e-01
7.33574986e-01 3.15931320e-01 -4.83622178e-02 -6.26598060e-01
8.09920505e-02 7.34342098e-01 8.05958807e-01 8.27930748e-01
-1.18461800e+00 7.68064618e-01 5.27007914e+00 1.03538728e+00
-1.30970335e+00 1.56856060e-01 2.46025890e-01 2.51890063e-01
-1.86803918e-02 -2.72926122e-01 -4.55706060e-01 5.69488943e-01
9.63440537e-01 2.28226691e-01 5.80061197e-01 3.52053195e-01
1.37667909e-01 -2.60419250e-01 -5.35647452e-01 8.69997561e-01
3.86043787e-02 -9.90769267e-01 1.61823124e-01 3.51067074e-02
6.81026042e-01 -1.58060923e-01 3.50693583e-01 2.99040601e-02
1.71256348e-01 -1.02272737e+00 8.25991213e-01 4.08141136e-01
8.71859133e-01 -1.06994569e+00 1.12552381e+00 3.85928005e-01
-1.24215698e+00 2.17349932e-01 -4.83811349e-01 4.38628793e-02
-8.32978189e-02 5.88342428e-01 -8.41439188e-01 7.31433988e-01
8.33710015e-01 2.47056171e-01 -2.06371337e-01 6.40696108e-01
-4.33146298e-01 5.22351623e-01 -2.23443821e-01 5.55889010e-01
5.84821880e-01 -5.47792315e-01 5.36830604e-01 1.13488352e+00
7.10882246e-01 1.08807534e-01 -4.48599368e-01 9.46074426e-01
-4.75081196e-03 -3.93411458e-01 -5.40349901e-01 1.58053741e-01
-1.49213046e-01 1.04405046e+00 -5.47599375e-01 -2.76267648e-01
-3.39024842e-01 1.15556240e+00 -3.45411271e-01 6.16124868e-01
-9.88428593e-01 -9.26411390e-01 5.92440844e-01 2.20183522e-01
6.60685480e-01 -1.43045232e-01 -2.47230045e-02 -8.22956741e-01
-2.16447085e-01 -5.58131218e-01 2.68301308e-01 -1.03976667e+00
-9.81322050e-01 7.85037339e-01 3.66121866e-02 -1.16691005e+00
-1.63788542e-01 -7.18641698e-01 -6.74446166e-01 1.67841029e+00
-2.06796145e+00 -1.24121702e+00 -6.56680346e-01 8.92465234e-01
5.17177820e-01 -3.60343000e-03 5.87256253e-01 3.88516605e-01
-1.84055924e-01 4.44552630e-01 1.41884208e-01 1.35466248e-01
6.56751394e-01 -1.21421945e+00 1.01323538e-01 1.29111040e+00
-2.15782568e-01 3.68961304e-01 9.66834664e-01 -3.91991884e-01
-7.70458996e-01 -1.22655702e+00 5.57684898e-01 -7.63744339e-02
3.74274492e-01 -2.92647004e-01 -9.03418362e-01 5.85115433e-01
7.05141664e-01 -7.93147646e-03 4.21546489e-01 -9.92615759e-01
-2.91347295e-01 -1.35781780e-01 -1.35990858e+00 4.59527016e-01
1.40945807e-01 -5.69406092e-01 -5.50731957e-01 2.50648838e-02
8.11255157e-01 -4.73955810e-01 -6.22441471e-01 1.33181006e-01
3.68949145e-01 -1.34791839e+00 1.13269651e+00 3.42845246e-02
7.92426407e-01 -7.88283110e-01 -1.27134949e-01 -1.39899087e+00
-2.37321466e-01 -6.00836813e-01 7.16041699e-02 9.18495238e-01
2.99574941e-01 -8.44836175e-01 7.58551002e-01 -6.57551065e-02
-1.96295246e-01 -2.62030661e-01 -7.43845880e-01 -3.79623562e-01
7.75639638e-02 -2.35574514e-01 4.36554015e-01 7.87292302e-01
-7.13295341e-01 1.51904136e-01 -6.08307719e-01 7.74979830e-01
6.96902275e-01 -5.69597259e-02 4.40760225e-01 -7.67790318e-01
-3.24361682e-01 -2.01489896e-01 -5.79344094e-01 -4.45177704e-01
-2.34249040e-01 -5.28912485e-01 3.91114429e-02 -1.45021152e+00
-3.96883041e-01 3.36643793e-02 -1.32014751e-01 1.13718674e-01
-1.54945999e-01 7.07611024e-01 2.70825237e-01 5.61686559e-03
5.28886020e-01 2.57220685e-01 1.36036515e+00 -2.79216826e-01
-9.14308354e-02 1.05766237e-01 -2.56753057e-01 6.07101917e-01
6.72348142e-01 -3.78137857e-01 -5.80151439e-01 -4.88853961e-01
-1.35782659e-01 -2.24613816e-01 4.91763651e-01 -1.38828897e+00
2.58993626e-01 3.22360724e-01 5.02184510e-01 -5.64871788e-01
4.89406765e-01 -1.25970328e+00 4.48915809e-01 6.11571550e-01
6.62976876e-02 -1.96965456e-01 1.85614675e-01 3.44818622e-01
-6.34920776e-01 -2.63315678e-01 1.20837855e+00 -1.29677117e-01
-6.40111804e-01 -2.20608083e-03 -5.21441519e-01 -4.48096216e-01
1.23232591e+00 -5.61986327e-01 -2.82637954e-01 -6.21815741e-01
-7.30295181e-01 -3.17617297e-01 2.88451374e-01 -2.50763297e-02
7.47747838e-01 -8.97018313e-01 -7.15454578e-01 5.88507354e-01
-3.04772913e-01 8.12917128e-02 5.52750468e-01 8.81754279e-01
-1.24367726e+00 3.16005230e-01 -3.30184013e-01 -1.90283492e-01
-1.22800112e+00 6.13761902e-01 6.85042322e-01 -1.72328919e-01
-6.88198805e-01 8.80595028e-01 5.18681228e-01 9.70911160e-02
-1.22453086e-01 -2.96567917e-01 -4.69083369e-01 -7.78561458e-02
6.67830527e-01 3.79505575e-01 2.14657292e-01 -8.55549395e-01
-2.10551675e-02 8.30928087e-01 2.81885624e-01 -2.28190899e-01
1.46437001e+00 -3.42164822e-02 -4.53341812e-01 -5.42557947e-02
1.48432148e+00 1.22313723e-01 -1.07945287e+00 -4.31467034e-02
-5.03804803e-01 -4.25917119e-01 6.46726787e-01 -5.59670210e-01
-1.33611131e+00 1.20145822e+00 8.56848538e-01 3.77027690e-01
1.71251082e+00 -5.02722323e-01 8.82498860e-01 -1.30075738e-01
1.60474062e-01 -6.02742553e-01 -1.10681474e-01 4.24483836e-01
6.89497471e-01 -6.52838528e-01 -1.26216933e-01 -6.61010027e-01
-3.42067897e-01 1.30272591e+00 1.41191766e-01 -5.42211056e-01
8.42187703e-01 3.23177963e-01 2.25569218e-01 -2.86246210e-01
-1.58300519e-01 -2.88531721e-01 4.27156508e-01 7.25097775e-01
2.38713101e-01 -2.81238645e-01 -4.36623096e-02 1.04702301e-01
-4.18406636e-01 -2.43362978e-01 6.29522562e-01 5.44026673e-01
-5.35344601e-01 -8.47607255e-01 -9.81661081e-01 3.65182497e-02
-7.11149096e-01 -4.25484926e-01 -1.49428532e-01 8.57139528e-01
7.40151823e-01 1.06030703e+00 2.51576602e-02 -3.60084236e-01
3.51664662e-01 9.07344446e-02 3.68549526e-01 -3.90208423e-01
-7.11531579e-01 3.09951484e-01 -2.79519647e-01 -2.19848484e-01
-3.84443551e-01 3.74932513e-02 -1.15952086e+00 -4.32522088e-01
-2.43205622e-01 2.98086494e-01 6.98599935e-01 7.99721420e-01
-9.50211212e-02 1.10322094e+00 5.33677161e-01 -1.03453195e+00
-4.14633155e-02 -1.11831546e+00 -8.59345973e-01 2.80385703e-01
9.38269913e-01 -4.60518122e-01 -4.76683497e-01 4.78602231e-01] | [10.909393310546875, -3.144644021987915] |
a6614559-ab78-4216-bf15-f469264b0cff | 4d-seismic-history-matching-incorporating | 1905.07469 | null | https://arxiv.org/abs/1905.07469v1 | https://arxiv.org/pdf/1905.07469v1.pdf | 4D Seismic History Matching Incorporating Unsupervised Learning | The work discussed and presented in this paper focuses on the history matching of reservoirs by integrating 4D seismic data into the inversion process using machine learning techniques. A new integrated scheme for the reconstruction of petrophysical properties with a modified Ensemble Smoother with Multiple Data Assimilation (ES-MDA) in a synthetic reservoir is proposed. The permeability field inside the reservoir is parametrised with an unsupervised learning approach, namely K-means with Singular Value Decomposition (K-SVD). This is combined with the Orthogonal Matching Pursuit (OMP) technique which is very typical for sparsity promoting regularisation schemes. Moreover, seismic attributes, in particular, acoustic impedance, are parametrised with the Discrete Cosine Transform (DCT). This novel combination of techniques from machine learning, sparsity regularisation, seismic imaging and history matching aims to address the ill-posedness of the inversion of historical production data efficiently using ES-MDA. In the numerical experiments provided, I demonstrate that these sparse representations of the petrophysical properties and the seismic attributes enables to obtain better production data matches to the true production data and to quantify the propagating waterfront better compared to more traditional methods that do not use comparable parametrisation techniques. | ['Clement Etienam'] | 2019-05-16 | null | null | null | null | ['seismic-imaging'] | ['miscellaneous'] | [ 1.47071198e-01 -4.32381555e-02 3.72009873e-01 8.20862055e-02
-6.35266840e-01 7.97628239e-02 7.88069367e-01 2.83466488e-01
-4.91749287e-01 6.89481974e-01 4.17846709e-01 -7.60558061e-03
-4.98201787e-01 -8.88071895e-01 -5.43340802e-01 -1.26768768e+00
-6.05917931e-01 5.49206257e-01 -1.33264601e-01 -6.38454854e-01
3.61004263e-01 8.42515469e-01 -1.67832267e+00 -4.81476746e-02
6.72179282e-01 8.11084390e-01 1.84929639e-01 4.52448428e-01
2.10953295e-01 7.22525299e-01 -3.34611908e-02 3.43021244e-01
3.07315886e-01 -2.18784496e-01 -3.99863362e-01 1.64275482e-01
1.56495988e-01 -2.09671661e-01 -1.32816240e-01 8.59157205e-01
3.79901081e-01 4.54110175e-01 8.94543946e-01 -6.49914145e-01
1.63118839e-01 3.59691769e-01 -6.15236461e-01 2.01031670e-01
2.84896851e-01 -1.05597533e-01 6.91193879e-01 -1.13382316e+00
3.46710980e-01 9.35326457e-01 9.30284619e-01 -3.15823555e-01
-1.60534310e+00 -1.31608397e-01 -6.64296627e-01 1.69840366e-01
-1.35439622e+00 -6.52074516e-01 1.15894723e+00 -8.13120484e-01
7.36306071e-01 3.87988925e-01 7.98749030e-01 4.62629050e-01
-2.33138219e-01 3.50053191e-01 1.29261482e+00 -6.73222482e-01
2.81802386e-01 2.03455999e-01 -2.18922704e-01 3.22015822e-01
2.06812233e-01 5.62569261e-01 -5.67370474e-01 -3.48293722e-01
7.92654991e-01 -9.02984589e-02 -2.61103570e-01 -2.33290374e-01
-1.28461421e+00 9.81699705e-01 4.73370254e-02 6.87511146e-01
-9.40881670e-01 1.27423601e-02 3.77097040e-01 3.24723750e-01
5.78014135e-01 5.77904344e-01 -5.80497645e-02 9.19919536e-02
-1.56386530e+00 5.10855436e-01 9.99669373e-01 4.04945761e-01
1.11020613e+00 9.02736723e-01 2.64756113e-01 4.91621733e-01
7.72086263e-01 1.11277997e+00 4.91671443e-01 -1.16342187e+00
2.46499509e-01 1.23931125e-01 2.13440433e-01 -1.17351198e+00
-1.96364090e-01 -4.32250530e-01 -1.13229275e+00 3.05305868e-01
4.04704422e-01 -1.21680513e-01 -3.39212328e-01 1.04390264e+00
4.41479385e-01 6.50982141e-01 6.02915823e-01 5.98789573e-01
4.77233171e-01 8.47298145e-01 -5.85381910e-02 -4.38939661e-01
1.10474646e+00 -3.22573811e-01 -6.77118003e-01 1.44534305e-01
4.75930631e-01 -7.48297274e-01 2.80379295e-01 4.42696333e-01
-1.06059694e+00 -4.12968785e-01 -1.09830701e+00 5.15220821e-01
-2.20866308e-01 -1.01519033e-01 2.91456759e-01 4.73783553e-01
-8.68475080e-01 1.27152634e+00 -9.27230895e-01 7.82922655e-02
-3.45973134e-01 2.20514406e-02 -4.43518668e-01 2.98219234e-01
-1.14697099e+00 1.10415018e+00 4.81303096e-01 5.00214815e-01
-8.30119848e-01 -7.51452744e-01 -9.60273921e-01 -3.17036882e-02
-2.38094658e-01 -2.51437575e-01 4.27177340e-01 -6.43370867e-01
-1.50107765e+00 4.54896510e-01 -7.21869841e-02 -4.92366821e-01
4.55470771e-01 9.71923545e-02 -3.49536628e-01 2.76098222e-01
-7.34184086e-02 -3.67678016e-01 1.19891119e+00 -1.39612603e+00
-1.25941053e-01 -2.34393254e-01 -8.14910173e-01 -5.12016239e-03
-1.74705669e-01 -2.24400014e-01 1.94003999e-01 -8.55180323e-01
6.41673505e-01 -5.48023462e-01 -5.63346803e-01 -6.16981506e-01
-1.55492220e-04 4.07614321e-01 5.44879973e-01 -1.26267576e+00
9.29510355e-01 -2.41341662e+00 5.93126237e-01 8.40715170e-01
-9.27783772e-02 -4.27425988e-02 2.82847792e-01 8.39820623e-01
-2.30076775e-01 -4.42729205e-01 -9.29808259e-01 -5.43292642e-01
-2.23139927e-01 5.49266696e-01 -2.54272789e-01 9.72917914e-01
2.90953238e-02 1.77315995e-01 -7.90448308e-01 -4.13799971e-01
3.97325933e-01 5.77158749e-01 -3.99960101e-01 1.99179873e-01
1.21631913e-01 9.34639692e-01 -3.63336474e-01 4.28048253e-01
9.42035258e-01 2.23364055e-01 3.81687954e-02 -2.63149351e-01
-7.05803514e-01 -3.36873263e-01 -1.71061409e+00 1.34507895e+00
-4.18063641e-01 3.09970558e-01 5.96716762e-01 -1.57199907e+00
1.28411293e+00 6.03878975e-01 7.20597506e-01 -5.59118807e-01
-1.57334462e-01 8.28670681e-01 -3.01710635e-01 -8.11026573e-01
4.57866400e-01 -5.39214075e-01 3.18193823e-01 2.35308290e-01
-1.25282556e-01 -4.25156891e-01 2.25164164e-02 -6.50038719e-02
5.72851539e-01 2.12067708e-01 1.32337108e-01 -7.17502654e-01
9.33241606e-01 -2.03782450e-02 2.83270657e-01 4.72765774e-01
4.46633816e-01 4.79378998e-01 2.54571289e-01 -1.86285466e-01
-1.42348468e+00 -5.26588976e-01 -5.36487520e-01 3.59082311e-01
-2.67823100e-01 1.67835981e-01 -1.74514592e-01 4.44614172e-01
2.85452515e-01 6.27323627e-01 -6.26623034e-01 2.09124476e-01
-7.22093344e-01 -1.06159508e+00 4.19788420e-01 1.63731843e-01
3.37073833e-01 -7.59415567e-01 -5.09378731e-01 5.85262358e-01
-5.37521876e-02 -8.53010118e-01 6.24581218e-01 2.01092616e-01
-1.50492895e+00 -6.03573024e-01 -8.58443379e-01 -4.80439872e-01
3.53731900e-01 -4.77340698e-01 7.80895710e-01 -1.85859710e-01
1.97398856e-01 4.12108421e-01 -4.42035049e-01 -1.57034338e-01
-8.42611194e-01 -4.68918890e-01 -2.27383673e-02 5.52007020e-01
-1.63119480e-01 -1.09725642e+00 -3.19976181e-01 6.36734720e-03
-8.51951361e-01 -2.70002484e-01 6.94447994e-01 7.67923832e-01
5.26577652e-01 7.67036080e-02 2.08680287e-01 -6.62096262e-01
2.73198873e-01 -8.42182279e-01 -7.11925328e-01 2.56582033e-02
-4.92117047e-01 3.09892088e-01 2.18586460e-01 -1.14196762e-01
-1.16442513e+00 -1.04002148e-01 -3.49545002e-01 -3.11144710e-01
-1.14623882e-01 1.05111408e+00 3.19817245e-01 -5.72942257e-01
7.96982408e-01 7.58769810e-01 3.65149617e-01 -8.78790677e-01
6.46956190e-02 4.94073987e-01 5.41380405e-01 -6.64218843e-01
9.59990501e-01 7.61725545e-01 5.47006369e-01 -1.70902288e+00
-2.20368057e-02 -9.35044408e-01 -5.15910625e-01 -3.57652962e-01
5.09554982e-01 -1.03305864e+00 -4.30259973e-01 7.33845055e-01
-8.78221393e-01 -1.28783718e-01 -6.84886515e-01 1.03156209e+00
-4.74528790e-01 8.65397394e-01 -4.55978632e-01 -1.21600735e+00
-2.66313434e-01 -8.94217789e-01 8.97627234e-01 -2.45740324e-01
2.55104244e-01 -1.18044150e+00 4.15198326e-01 1.75008506e-01
7.33153820e-01 7.86684155e-01 6.56761229e-01 -4.28552002e-01
-4.32047635e-01 -2.41483837e-01 1.89945891e-01 5.74955225e-01
-2.57616282e-01 -1.43537670e-01 -9.81986284e-01 -2.54651904e-01
5.86052120e-01 9.74194854e-02 9.03905690e-01 4.30083215e-01
2.39408389e-01 -2.53781885e-01 7.07573593e-02 7.92160571e-01
1.87178028e+00 -1.05264910e-01 6.95916593e-01 5.99784911e-01
3.93068492e-01 7.41607249e-01 4.78441387e-01 1.06455815e+00
-1.84789658e-01 3.46528351e-01 6.38763309e-01 1.21854447e-01
3.26099783e-01 3.20406228e-01 1.85253158e-01 1.04349279e+00
-9.35686171e-01 5.54645121e-01 -1.08435464e+00 7.00554073e-01
-1.58242130e+00 -1.15379059e+00 -4.49823052e-01 2.36473274e+00
4.80375886e-01 -1.68768927e-01 -2.39604071e-01 7.27494538e-01
3.52441370e-01 3.84114593e-01 1.32263824e-01 -2.36706331e-01
-3.81180912e-01 4.94674206e-01 9.06373322e-01 7.77848601e-01
-8.79093766e-01 1.04307933e-02 5.38162947e+00 5.28723598e-01
-1.06680846e+00 3.03091496e-01 -1.79408962e-04 7.43470609e-01
-5.70901871e-01 1.35360792e-01 -5.94659507e-01 3.47041965e-01
1.17365670e+00 1.78441927e-01 4.80029434e-01 4.38839555e-01
6.36492848e-01 -3.56312215e-01 -4.96205419e-01 7.88858175e-01
5.36137959e-03 -1.49534547e+00 -3.66809160e-01 5.33092357e-02
6.80391014e-01 2.00304329e-01 -1.81082696e-01 -3.85827990e-03
-3.45399678e-01 -5.90793192e-01 8.06335509e-01 1.26440716e+00
4.57793653e-01 -4.20426339e-01 7.76439607e-01 3.76469791e-01
-1.14766777e+00 -1.59747064e-01 -1.79585516e-02 -1.46368220e-01
6.70531809e-01 7.76134670e-01 -6.28345370e-01 9.56047654e-01
5.27863681e-01 1.01427436e+00 1.78703282e-03 1.09191585e+00
1.84141904e-01 7.24736392e-01 -8.24520946e-01 3.59053910e-01
3.78012568e-01 -8.87164414e-01 1.14862239e+00 9.59252298e-01
7.69872785e-01 2.11736307e-01 -1.90882593e-01 7.51083970e-01
6.90717936e-01 3.19191277e-01 -6.02900326e-01 -2.79503074e-02
1.78188831e-02 8.00937235e-01 -2.24914491e-01 -2.61050195e-01
-2.59315014e-01 1.80447519e-01 -5.80148399e-01 4.14806306e-01
-2.44107038e-01 -4.75050882e-02 2.36778408e-01 5.41720271e-01
5.30358255e-01 -4.80618477e-01 -1.19429529e-01 -9.19993460e-01
-2.43794858e-01 -6.03431702e-01 2.00617075e-01 -5.50263226e-01
-1.05245674e+00 3.00849676e-01 3.88732076e-01 -1.42429745e+00
-4.63207990e-01 -4.19656456e-01 -5.48272789e-01 1.30608773e+00
-1.88392448e+00 -1.00744462e+00 -1.63484290e-01 7.78917074e-01
-4.42695096e-02 -2.65373528e-01 8.97614062e-01 4.79906887e-01
-8.74911547e-02 -2.98526078e-01 1.03791428e+00 -8.47522914e-02
2.57437170e-01 -1.16026247e+00 -1.73977822e-01 7.67541945e-01
-3.51643890e-01 2.92045444e-01 1.26047766e+00 -7.58489370e-01
-1.48615897e+00 -5.90719342e-01 9.77358758e-01 4.26265866e-01
9.42138433e-01 4.46230829e-01 -1.13887250e+00 5.29334128e-01
1.45595387e-01 1.93494096e-01 5.47561646e-01 -2.93936431e-01
1.77627459e-01 -2.49820873e-01 -1.10235488e+00 -1.05770245e-01
-2.15456247e-01 -4.67131853e-01 -8.74913275e-01 3.38277638e-01
-9.24168751e-02 -1.92858160e-01 -1.58433676e+00 6.17400289e-01
4.25231993e-01 -8.99550855e-01 1.13792145e+00 -2.46376425e-01
2.66030848e-01 -3.72438341e-01 -4.93175477e-01 -9.79738891e-01
-6.69783801e-02 -5.66409528e-01 -1.89915195e-01 8.67517054e-01
2.76289463e-01 -7.39385724e-01 6.95592463e-01 1.78298160e-01
-1.67055093e-02 -2.71983266e-01 -1.12075889e+00 -5.76911926e-01
-1.66846754e-03 -3.34088475e-01 3.09193552e-01 1.14884794e+00
-1.00490943e-01 -2.25531563e-01 -6.46118164e-01 6.06269121e-01
1.20786524e+00 5.51116243e-02 3.92714024e-01 -1.39592040e+00
-6.44152105e-01 -1.27151638e-01 -4.20892298e-01 -4.43533778e-01
-2.57280409e-01 -5.41250885e-01 -8.82192627e-02 -1.01592696e+00
-5.70620954e-01 -7.77925253e-01 -1.27184182e-01 -2.05679834e-01
3.78406167e-01 4.06621024e-02 2.85061821e-03 6.20597363e-01
4.44690257e-01 7.26036906e-01 1.09829164e+00 -4.69384715e-02
-2.30987966e-01 4.45718616e-02 3.83689135e-01 6.05607271e-01
4.35857475e-01 -4.81753170e-01 6.29154369e-02 -2.40136042e-01
4.36443001e-01 8.10515702e-01 3.64104062e-01 -1.34620988e+00
2.72377312e-01 5.27231283e-02 1.02285706e-02 -5.78928769e-01
5.41123807e-01 -1.01763070e+00 7.01687694e-01 6.37839794e-01
1.44434944e-01 -2.19644681e-01 9.54140574e-02 6.87525928e-01
-7.39395797e-01 -8.13080311e-01 7.93506682e-01 -5.32792032e-01
-9.70202088e-01 -1.20005950e-01 -5.12615085e-01 -3.72488409e-01
6.93716168e-01 -4.27185446e-01 4.33151245e-01 -1.90654635e-01
-1.32973719e+00 -7.02485666e-02 8.69156569e-02 -7.44102597e-01
7.50727296e-01 -9.10142362e-01 -1.15593147e+00 6.59572124e-01
-1.09661743e-01 -2.95557518e-04 5.08992434e-01 1.46272588e+00
-1.11756253e+00 -8.60068500e-02 -2.01092392e-01 -7.07670271e-01
-8.49011719e-01 1.31303638e-01 5.00111401e-01 -4.04816329e-01
-6.11004829e-01 6.46383047e-01 -7.07318246e-01 -2.21747875e-01
-1.03990413e-01 1.40453726e-01 -9.46565807e-01 6.11321747e-01
2.05164135e-01 7.03613043e-01 1.94202542e-01 -1.17326128e+00
1.18356623e-01 6.73528373e-01 7.49776244e-01 -3.30015898e-01
1.87767959e+00 -5.52824698e-02 -5.90970933e-01 5.58518529e-01
1.17654109e+00 4.99694757e-02 -9.81739223e-01 -4.78784800e-01
1.67304963e-01 -4.00286168e-01 5.31835079e-01 -4.48763855e-02
-9.95809913e-01 7.60458946e-01 4.41141069e-01 3.17823678e-01
9.92233515e-01 -5.18135250e-01 3.87707561e-01 4.99523640e-01
1.15382053e-01 -9.10073817e-01 -3.92870516e-01 4.37562436e-01
1.06337523e+00 -1.03294194e+00 3.67415607e-01 -1.65945441e-01
-3.68409306e-01 1.39529109e+00 -4.26277190e-01 -3.67059290e-01
1.14721811e+00 2.77381241e-01 -2.32497588e-01 -2.89114028e-01
1.16973668e-01 -2.25805208e-01 -1.46692559e-01 1.75654456e-01
-2.71633752e-02 -2.45010033e-02 -2.83613771e-01 -7.52488226e-02
9.22952294e-02 -1.26651257e-01 3.77296120e-01 1.12729979e+00
-4.62044746e-01 -8.99106145e-01 -1.15491772e+00 3.22658628e-01
-2.06899688e-01 -2.22807929e-01 7.73985445e-01 6.34937584e-01
9.61117912e-03 4.48152602e-01 2.81444024e-02 6.07390441e-02
3.89406502e-01 1.33608535e-01 2.64105707e-01 -2.85504073e-01
-5.27041793e-01 4.49480563e-01 1.32772237e-01 -1.40913799e-01
-1.06806922e+00 -1.21056271e+00 -7.91870236e-01 -3.17131162e-01
-3.03518683e-01 7.07671463e-01 8.49928439e-01 9.99637961e-01
-2.90879607e-01 2.69287229e-01 7.51574695e-01 -1.59204793e+00
-7.48763919e-01 -1.10647500e+00 -1.42440951e+00 3.54879409e-01
6.61031842e-01 -7.06860006e-01 -8.43744397e-01 1.83236212e-01] | [6.856794834136963, 2.7442433834075928] |
5149174a-a322-44f2-adf0-93709c04c79b | hierarchical-compositional-representations | 2208.09424 | null | https://arxiv.org/abs/2208.09424v2 | https://arxiv.org/pdf/2208.09424v2.pdf | Hierarchical Compositional Representations for Few-shot Action Recognition | Recently action recognition has received more and more attention for its comprehensive and practical applications in intelligent surveillance and human-computer interaction. However, few-shot action recognition has not been well explored and remains challenging because of data scarcity. In this paper, we propose a novel hierarchical compositional representations (HCR) learning approach for few-shot action recognition. Specifically, we divide a complicated action into several sub-actions by carefully designed hierarchical clustering and further decompose the sub-actions into more fine-grained spatially attentional sub-actions (SAS-actions). Although there exist large differences between base classes and novel classes, they can share similar patterns in sub-actions or SAS-actions. Furthermore, we adopt the Earth Mover's Distance in the transportation problem to measure the similarity between video samples in terms of sub-action representations. It computes the optimal matching flows between sub-actions as distance metric, which is favorable for comparing fine-grained patterns. Extensive experiments show our method achieves the state-of-the-art results on HMDB51, UCF101 and Kinetics datasets. | ['Shiguang Shan', 'Xin Jin', 'Shuzhe Wu', 'Jie Zhang', 'Changzhen Li'] | 2022-08-19 | null | null | null | null | ['few-shot-action-recognition'] | ['computer-vision'] | [ 4.08801019e-01 -4.49818999e-01 -4.43303287e-01 -2.79128820e-01
-7.01017201e-01 -2.03294918e-01 5.35143673e-01 -1.17471755e-01
-2.20723644e-01 4.02784199e-01 7.98393726e-01 1.14788465e-01
-3.20618421e-01 -6.57978773e-01 -4.67251271e-01 -9.57931280e-01
-1.26249894e-01 7.76369078e-03 9.67908740e-01 1.42051643e-02
2.05707967e-01 5.09757519e-01 -1.83928049e+00 7.49740541e-01
6.36947989e-01 9.66339469e-01 4.91800718e-02 4.93786186e-01
-4.59003672e-02 1.29557693e+00 -5.80433965e-01 5.35959080e-02
3.34474713e-01 -8.86115551e-01 -9.25415814e-01 5.06651640e-01
2.15636924e-01 -3.30664724e-01 -5.96962094e-01 1.14503419e+00
1.97838008e-01 7.61627972e-01 7.42460012e-01 -1.29948497e+00
-6.83127463e-01 3.02927643e-01 -8.74792635e-01 5.37138522e-01
3.58030140e-01 8.65709335e-02 8.43722105e-01 -5.38241208e-01
4.62885499e-01 1.54202878e+00 5.58402061e-01 5.30587554e-01
-6.86943948e-01 -4.17505145e-01 5.00016391e-01 8.31660032e-01
-1.37113094e+00 -2.88622379e-01 5.28816164e-01 -4.93906170e-01
8.90672505e-01 2.35288501e-01 5.19602597e-01 8.45622003e-01
1.15390971e-01 1.10583448e+00 5.84928513e-01 -2.56522715e-01
4.06017959e-01 -4.33698058e-01 1.38391376e-01 5.10275781e-01
5.25419451e-02 -3.78290474e-01 -2.27825001e-01 -4.26823869e-02
6.56486630e-01 8.31591964e-01 -3.00575405e-01 -2.92775810e-01
-1.23436129e+00 6.62908435e-01 3.68721634e-01 6.49547279e-01
-5.80254316e-01 -9.61306617e-02 5.69243729e-01 -6.30731508e-02
2.43924931e-01 -1.07349023e-01 -2.46306926e-01 -3.81568313e-01
-6.95149720e-01 -9.13642813e-03 1.79209515e-01 8.42833221e-01
6.86759889e-01 -8.18160102e-02 -5.55518389e-01 9.27873850e-01
-7.76174217e-02 2.09884375e-01 7.09356606e-01 -8.64925444e-01
3.91970783e-01 9.09398675e-01 1.01715416e-01 -1.31411064e+00
-1.20603710e-01 1.68112263e-01 -1.05945241e+00 -1.35754019e-01
3.03564548e-01 1.29648089e-01 -9.64390635e-01 1.37225652e+00
4.91452187e-01 6.34539604e-01 9.28167552e-02 1.02003598e+00
7.35856295e-01 8.89460087e-01 2.91253150e-01 -3.43349844e-01
1.36948514e+00 -1.22734702e+00 -5.98787189e-01 1.48763219e-02
7.17130482e-01 -3.13014537e-01 8.52188349e-01 -3.36027443e-02
-8.07981431e-01 -8.12314510e-01 -7.85912097e-01 1.59277856e-01
-4.19169843e-01 1.75197273e-02 3.27095270e-01 3.73636872e-01
-6.10311985e-01 4.22195882e-01 -9.20356154e-01 -6.87883615e-01
7.21241474e-01 -7.33085498e-02 -3.91611069e-01 -4.47368234e-01
-1.12991178e+00 4.58701640e-01 3.78626257e-01 -1.38354257e-01
-1.00322855e+00 -4.52089816e-01 -8.58347833e-01 6.54152036e-02
6.97506011e-01 -1.79424927e-01 9.53921676e-01 -9.81497824e-01
-1.35900521e+00 5.68828344e-01 -2.61795789e-01 -3.74450922e-01
2.18540743e-01 5.78027740e-02 -7.76991129e-01 4.56366539e-01
2.48890221e-01 4.97257769e-01 9.01860893e-01 -6.33626580e-01
-1.07727194e+00 -5.05336344e-01 2.75023729e-01 2.35704556e-01
-3.10189784e-01 2.81974465e-01 -6.34493589e-01 -8.98339808e-01
1.82864249e-01 -7.82882750e-01 -2.14875594e-01 -5.05724177e-02
-2.62524728e-02 -5.19824624e-01 9.80678260e-01 -4.35983658e-01
1.40303481e+00 -2.42457175e+00 2.06953987e-01 -2.08677664e-01
-1.12244353e-01 4.50095117e-01 -2.70313501e-01 3.15421730e-01
-5.20715751e-02 -9.51563194e-02 -2.70658493e-01 1.98298156e-01
-1.79955706e-01 1.34691060e-01 -1.15335271e-01 5.25173962e-01
1.25871867e-01 8.03642988e-01 -9.70871806e-01 -6.09822154e-01
4.03726757e-01 1.95841640e-01 -3.25349689e-01 1.83166370e-01
-3.22262221e-03 2.68358856e-01 -7.62636364e-01 9.22635138e-01
5.09651721e-01 -1.79640651e-01 -6.21344335e-02 -2.35513702e-01
-9.90454033e-02 -1.10709734e-01 -1.15288711e+00 1.53635645e+00
1.08386442e-01 4.95940328e-01 -3.21533263e-01 -1.40970457e+00
7.35074461e-01 6.57888204e-02 7.77352095e-01 -7.20512629e-01
8.38583708e-02 -1.74133837e-01 -5.67763336e-02 -7.02577174e-01
3.33666861e-01 -2.15776414e-01 -1.19809575e-01 3.66541862e-01
-3.04937214e-01 4.79002625e-01 3.42405200e-01 7.03558978e-03
1.34400451e+00 8.75426084e-02 7.35747993e-01 -1.03145458e-01
7.94134140e-01 4.15467471e-02 9.48944569e-01 6.15730047e-01
-6.26155257e-01 7.73485422e-01 2.87136644e-01 -6.09420300e-01
-4.97339815e-01 -8.79143476e-01 2.70684779e-01 1.35650289e+00
4.35267746e-01 -3.91547114e-01 -7.15832531e-01 -1.03277457e+00
-1.63410023e-01 5.15528619e-01 -8.85271311e-01 -2.74724066e-01
-6.55163884e-01 -6.27792656e-01 4.77100313e-01 8.91040385e-01
8.43001008e-01 -1.33127260e+00 -9.39188480e-01 1.44861773e-01
-2.16800973e-01 -9.50701177e-01 -6.89525127e-01 -3.73823762e-01
-6.71543717e-01 -1.31047130e+00 -1.06207013e+00 -7.26957023e-01
6.51587665e-01 1.04380310e+00 5.85997820e-01 -5.86917102e-02
-5.59309304e-01 4.68261063e-01 -9.72970843e-01 -1.01739027e-01
4.31189947e-02 -2.86986560e-01 4.64113131e-02 5.13956487e-01
8.23898554e-01 -1.38835222e-01 -6.71302736e-01 8.39162707e-01
-1.11464477e+00 -2.36289948e-01 6.68195724e-01 4.03902054e-01
7.90216804e-01 4.01196659e-01 3.49650681e-01 -4.11675721e-01
1.96014822e-01 -5.87712407e-01 -1.85803965e-01 6.21469975e-01
2.26794556e-01 -2.60000557e-01 6.96752548e-01 -5.35816252e-01
-1.17305350e+00 1.27670895e-02 4.13608670e-01 -6.95699334e-01
-5.74196219e-01 2.46372521e-01 -4.87820148e-01 2.13853702e-01
3.06794107e-01 5.53079784e-01 -2.88293183e-01 -4.38740909e-01
3.15238386e-01 8.02864254e-01 3.16576421e-01 -3.08000952e-01
3.59435916e-01 8.22044671e-01 -1.44978017e-01 -1.17702103e+00
-9.01137531e-01 -7.85790145e-01 -8.18826556e-01 -3.16646427e-01
1.41274261e+00 -6.51020348e-01 -4.93141532e-01 6.19691670e-01
-8.06492090e-01 -7.91018382e-02 -3.38456869e-01 6.02505505e-01
-5.32386780e-01 7.71895289e-01 -2.33855829e-01 -7.55601764e-01
1.50617007e-02 -1.12613297e+00 1.09218645e+00 4.09217089e-01
-2.20122002e-02 -6.27283335e-01 1.10187002e-01 4.06920820e-01
6.09910488e-02 2.70042181e-01 6.32451713e-01 -6.05360746e-01
-4.71443564e-01 -2.41453737e-01 -2.60267287e-01 2.84870565e-01
5.07532775e-01 -8.98881704e-02 -4.51216608e-01 -1.87122256e-01
-8.51937458e-02 -1.98638156e-01 1.04574120e+00 5.18285334e-01
1.43298936e+00 -3.16221714e-01 -4.68061000e-01 3.31320018e-01
9.45466042e-01 7.31061935e-01 1.01406932e+00 2.09511191e-01
7.54159987e-01 5.59569955e-01 1.06181514e+00 5.86427748e-01
9.90530401e-02 7.63834059e-01 1.10148005e-01 3.01433742e-01
6.64992929e-02 -2.32029036e-02 4.78591770e-01 4.55623329e-01
-2.94383883e-01 -1.72916744e-02 -8.25771749e-01 5.24574339e-01
-2.26404262e+00 -1.63216281e+00 9.63392705e-02 2.10550714e+00
2.87659049e-01 9.25200176e-04 3.68203849e-01 1.29163526e-02
1.04275727e+00 5.92441380e-01 -6.67772532e-01 5.05354255e-02
4.81487662e-02 -2.57190466e-01 3.20634335e-01 -1.85147375e-02
-1.49547684e+00 8.23922992e-01 5.22958469e+00 1.20516324e+00
-6.20717525e-01 5.58456890e-02 6.85899436e-01 -6.79788068e-02
3.45990002e-01 -3.42657357e-01 -7.96163261e-01 6.65433347e-01
5.50097346e-01 3.66359428e-02 1.17059596e-01 1.00574183e+00
3.11670393e-01 -2.12821782e-01 -9.04769540e-01 1.11506164e+00
4.22499627e-01 -1.24865448e+00 3.99067491e-01 -2.73719221e-01
8.12122941e-01 -3.73735249e-01 -2.59880871e-01 4.66033936e-01
7.25334808e-02 -8.03689420e-01 3.92182082e-01 6.03553414e-01
5.95589161e-01 -7.91068017e-01 5.65721154e-01 3.20942938e-01
-1.81386650e+00 -3.17169964e-01 -6.80792391e-01 -3.78353447e-02
4.55988310e-02 7.86071867e-02 1.09625898e-01 6.09458268e-01
9.83964980e-01 1.35701776e+00 -5.06209671e-01 1.07104969e+00
2.18421489e-01 3.75774264e-01 1.14706799e-01 -5.34218289e-02
6.08311772e-01 -3.98364216e-01 2.65371293e-01 1.01402915e+00
2.24817678e-01 7.72858977e-01 4.96513098e-01 1.79681480e-01
1.60321131e-01 -1.27533255e-02 -4.82732058e-01 -2.08361626e-01
3.44949812e-01 8.82750332e-01 -8.98031414e-01 -6.81314290e-01
-7.19995022e-01 9.45665896e-01 1.42341778e-01 2.81082392e-01
-1.09895861e+00 -6.07968569e-01 9.76739824e-01 1.23449564e-01
6.35120273e-01 5.08789346e-02 3.14811498e-01 -1.29469979e+00
8.68998095e-03 -9.52756763e-01 8.54981005e-01 -4.71017957e-01
-1.25838220e+00 4.17658150e-01 3.46868187e-01 -1.83104396e+00
2.06799619e-02 -3.86219800e-01 -7.64863193e-01 1.63252667e-01
-1.06624532e+00 -9.45681989e-01 -5.70126653e-01 8.94003093e-01
1.09317315e+00 -2.96632975e-01 4.99893337e-01 3.35125327e-01
-8.22018027e-01 3.50055099e-01 2.01095060e-01 3.71184409e-01
3.88752699e-01 -6.44523025e-01 1.38689101e-01 9.11646903e-01
2.43668053e-02 3.04349929e-01 2.48455852e-01 -6.82391167e-01
-1.16153550e+00 -1.43795168e+00 4.41859037e-01 -2.06012219e-01
5.90480030e-01 1.04633808e-01 -1.00512445e+00 5.82613051e-01
-1.16645694e-01 2.15545177e-01 6.99940920e-01 -3.72366816e-01
-3.34214717e-01 -1.70699582e-01 -1.02952683e+00 5.90356290e-01
1.40951455e+00 -3.58336210e-01 -8.42472970e-01 2.68334121e-01
5.22557855e-01 2.06786275e-01 -4.94615376e-01 5.12222826e-01
4.86902207e-01 -1.16320276e+00 1.02123916e+00 -9.93007302e-01
2.82994628e-01 -6.49409056e-01 -3.79777491e-01 -9.24983799e-01
-7.01724410e-01 -2.75336474e-01 -6.18560798e-02 1.04127240e+00
-2.13141218e-01 -3.26689810e-01 8.28123033e-01 3.69718522e-01
-1.24663241e-01 -8.12560797e-01 -8.77921164e-01 -1.01663589e+00
-2.99997509e-01 -2.57003307e-01 5.62483907e-01 7.94162095e-01
9.07478575e-03 -7.70989582e-02 -5.90006649e-01 1.03094662e-02
5.60657561e-01 3.66995931e-01 7.26882994e-01 -9.99597967e-01
-2.47057140e-01 -4.32490796e-01 -8.55286598e-01 -1.25001085e+00
1.08970381e-01 -4.09229428e-01 2.24032015e-01 -1.68581629e+00
4.93186772e-01 3.51209193e-02 -5.71017027e-01 4.88084763e-01
-2.93979406e-01 2.62617528e-01 1.37151882e-01 4.40224469e-01
-1.33905852e+00 8.55214059e-01 9.51045334e-01 -3.35684776e-01
-2.48990253e-01 -8.00005570e-02 -2.82628506e-01 9.33157325e-01
7.38363266e-01 -3.16555083e-01 -5.28975129e-01 -1.75356999e-01
-6.41170979e-01 -1.49511863e-02 1.34769633e-01 -1.47151160e+00
2.00211093e-01 -6.16289854e-01 3.85702640e-01 -7.46836245e-01
2.07522765e-01 -9.17788446e-01 1.35492593e-01 5.22872269e-01
-4.26097929e-01 -2.20080897e-01 -1.26027092e-01 9.89728034e-01
-3.70358884e-01 -4.57071178e-02 9.77176309e-01 -2.46293128e-01
-1.23767340e+00 5.52530706e-01 -5.92732131e-01 2.18705133e-01
1.66492093e+00 -5.92888653e-01 -2.27683395e-01 -1.69112042e-01
-6.19913042e-01 1.87407643e-01 2.59007066e-01 7.65690327e-01
8.52584600e-01 -1.49670100e+00 -5.74244022e-01 8.57969373e-02
4.15424079e-01 -2.86337465e-01 8.72872412e-01 9.80195343e-01
-3.82853687e-01 3.93728048e-01 -3.96475703e-01 -4.78832692e-01
-1.69642866e+00 7.97678292e-01 1.30603164e-01 -6.42704442e-02
-8.49788785e-01 5.55807829e-01 5.10311544e-01 8.89291316e-02
4.24927384e-01 -2.88774729e-01 -4.80115384e-01 2.77223587e-01
1.14884341e+00 7.95013845e-01 -5.41837275e-01 -1.10296035e+00
-7.03163028e-01 8.33771944e-01 -1.25498116e-01 5.18847704e-01
1.27755070e+00 -8.47723261e-02 4.59173620e-02 4.10037249e-01
1.29698896e+00 -5.71454287e-01 -1.42489231e+00 -2.27994263e-01
3.93062867e-02 -8.54139924e-01 -3.52020890e-01 -1.48961082e-01
-1.05698395e+00 8.55987787e-01 4.64447170e-01 1.60298109e-01
1.23253727e+00 2.54386127e-01 8.55314493e-01 5.46389759e-01
4.44142848e-01 -1.17609143e+00 5.04683197e-01 4.58386838e-01
6.90758765e-01 -1.13493347e+00 1.59036200e-02 -2.30634570e-01
-9.29986537e-01 8.77360821e-01 6.24605656e-01 -1.31633639e-01
6.87165916e-01 -3.12940150e-01 -2.03683510e-01 -1.19104058e-01
-4.25795048e-01 -4.19117272e-01 4.46507990e-01 6.87321007e-01
-1.84449796e-02 1.15560345e-01 -1.47863835e-01 5.69957972e-01
6.21363580e-01 -1.96000747e-03 1.85664818e-01 1.16695821e+00
-8.43088865e-01 -4.63389784e-01 -2.91409940e-01 5.39997160e-01
-3.39331836e-01 2.45757356e-01 -2.12584451e-01 6.96721494e-01
1.90648347e-01 1.04038846e+00 2.26688996e-01 -6.11501873e-01
4.03192550e-01 2.23330017e-02 2.85176218e-01 -5.81354201e-01
-8.93325135e-02 -7.80783594e-02 -2.21692458e-01 -8.33891690e-01
-8.49761963e-01 -8.93380821e-01 -1.28833485e+00 -1.07490428e-01
8.38560686e-02 5.20008393e-02 -1.37749225e-01 1.04468513e+00
4.90697861e-01 4.82359439e-01 6.37360513e-01 -8.58910084e-01
-3.32228601e-01 -8.88336897e-01 -7.61257112e-01 7.83425689e-01
9.73530114e-03 -9.11484599e-01 -3.46721202e-01 2.61564583e-01] | [8.47197151184082, 0.750688374042511] |
0159fa85-143e-462c-a925-2879efa81364 | gan-based-deep-distributional-reinforcement | 1905.03929 | null | https://arxiv.org/abs/1905.03929v3 | https://arxiv.org/pdf/1905.03929v3.pdf | GAN-powered Deep Distributional Reinforcement Learning for Resource Management in Network Slicing | Network slicing is a key technology in 5G communications system. Its purpose is to dynamically and efficiently allocate resources for diversified services with distinct requirements over a common underlying physical infrastructure. Therein, demand-aware resource allocation is of significant importance to network slicing. In this paper, we consider a scenario that contains several slices in a radio access network with base stations that share the same physical resources (e.g., bandwidth or slots). We leverage deep reinforcement learning (DRL) to solve this problem by considering the varying service demands as the environment state and the allocated resources as the environment action. In order to reduce the effects of the annoying randomness and noise embedded in the received service level agreement (SLA) satisfaction ratio (SSR) and spectrum efficiency (SE), we primarily propose generative adversarial network-powered deep distributional Q network (GAN-DDQN) to learn the action-value distribution driven by minimizing the discrepancy between the estimated action-value distribution and the target action-value distribution. We put forward a reward-clipping mechanism to stabilize GAN-DDQN training against the effects of widely-spanning utility values. Moreover, we further develop Dueling GAN-DDQN, which uses a specially designed dueling generator, to learn the action-value distribution by estimating the state-value distribution and the action advantage function. Finally, we verify the performance of the proposed GAN-DDQN and Dueling GAN-DDQN algorithms through extensive simulations. | ['Yuxiu Hua', 'Xianfu Chen', 'Honggang Zhang', 'Zhifeng Zhao', 'Rongpeng Li'] | 2019-05-10 | null | null | null | null | ['distributional-reinforcement-learning'] | ['methodology'] | [-5.41859716e-02 1.98235616e-01 -3.07451814e-01 -2.21219867e-01
-6.11128747e-01 -4.66317654e-01 -3.26937027e-02 -8.36838484e-01
1.98600337e-01 1.31610107e+00 1.63073152e-01 -6.35049224e-01
-3.94811630e-01 -1.06511569e+00 -3.64168108e-01 -1.11082113e+00
-2.45008186e-01 3.72180998e-01 -3.99963379e-01 -1.55502751e-01
-2.40133137e-01 4.98148888e-01 -5.17022371e-01 -3.95154804e-01
8.10785592e-01 1.36574733e+00 1.52027249e-01 3.97392362e-01
2.90990043e-02 8.70318234e-01 -1.14933193e+00 -2.90302396e-01
6.63497567e-01 -9.75512207e-01 -2.70848393e-01 2.52560228e-01
-7.34872460e-01 -5.29027045e-01 -7.48366952e-01 9.43403065e-01
8.82060826e-01 8.03227797e-02 2.48662695e-01 -1.65514183e+00
-6.45969748e-01 7.08310783e-01 -6.00864828e-01 3.71966213e-01
-4.32095051e-01 5.14625371e-01 8.49770665e-01 -4.28986959e-02
1.95747331e-01 1.10954857e+00 1.14494994e-01 6.93201423e-01
-1.08075750e+00 -8.51140320e-01 1.65384397e-01 8.24360698e-02
-1.20857120e+00 -4.49376076e-01 7.61279225e-01 6.52427897e-02
2.99685806e-01 1.92324921e-01 6.72263861e-01 7.84462333e-01
1.74241453e-01 5.27296484e-01 4.65217531e-01 -1.93383887e-01
5.53744674e-01 -1.85801145e-02 -9.57108378e-01 2.03160703e-01
-2.30797902e-01 1.81696713e-01 1.47605762e-01 -1.86314180e-01
1.28822100e+00 1.65850837e-02 -4.18305159e-01 -2.51317859e-01
-7.14954197e-01 6.18057668e-01 4.89183158e-01 -6.38526380e-02
-8.62500548e-01 5.42764425e-01 2.15369239e-01 5.77810585e-01
2.46615261e-01 2.02138662e-01 -2.73824781e-01 -2.43073776e-01
-8.64731610e-01 3.27522680e-02 3.98416251e-01 1.38829303e+00
4.28065419e-01 8.00958574e-01 -7.98538387e-01 6.73128426e-01
1.51742741e-01 6.76823616e-01 2.81472683e-01 -1.12274790e+00
5.64074993e-01 -1.18461251e-01 4.12261039e-01 -6.21887028e-01
-3.04936707e-01 -1.23830295e+00 -1.03802598e+00 -1.69048324e-01
1.20957322e-01 -9.17154789e-01 -6.80691481e-01 2.14443469e+00
7.72012547e-02 4.33952421e-01 4.24178056e-02 9.48065996e-01
1.17326893e-01 8.01387608e-01 -1.56725362e-01 -6.63168669e-01
8.07820261e-01 -9.00574148e-01 -7.68759966e-01 1.50327995e-01
3.91785920e-01 -3.18794042e-01 8.04869354e-01 8.28863904e-02
-1.37392545e+00 -2.95690447e-01 -9.52134728e-01 7.59445786e-01
4.47764933e-01 5.93976788e-02 2.43500739e-01 1.03441226e+00
-1.00927866e+00 2.13561863e-01 -2.99859911e-01 1.04751803e-01
8.07504594e-01 4.85994399e-01 4.80200171e-01 1.85490493e-02
-1.40869844e+00 8.32606927e-02 1.15742616e-01 1.64689302e-01
-1.26699352e+00 -9.09816861e-01 -1.98173672e-01 5.78613460e-01
6.61403060e-01 -6.82781756e-01 1.29187667e+00 -1.01697683e+00
-1.98493135e+00 -2.12428626e-02 3.54356408e-01 -5.72297752e-01
5.38827538e-01 4.38081473e-01 -7.77653456e-01 -7.55499757e-04
-1.84997797e-01 1.34435609e-01 7.73175240e-01 -1.03692007e+00
-5.43084145e-01 -1.60589367e-01 2.69888997e-01 1.73736677e-01
-8.25430974e-02 -2.25047812e-01 -2.55490720e-01 -8.32775593e-01
-1.63093135e-01 -7.94219375e-01 -4.72173601e-01 -3.34481359e-01
-5.68456590e-01 1.05630696e-01 5.78429222e-01 -5.21835029e-01
1.36734009e+00 -2.14554763e+00 1.05212502e-01 5.68390071e-01
-3.85899432e-02 2.27556080e-01 -2.75852680e-01 3.85236025e-01
2.66222626e-01 3.47437024e-01 3.78437117e-02 5.93324490e-02
8.45638067e-02 4.13355172e-01 -4.29639339e-01 1.77102089e-01
-5.68606928e-02 7.98815668e-01 -9.78589356e-01 1.24826841e-01
1.31176203e-01 7.98448771e-02 -5.87774098e-01 4.99383658e-01
-4.75393951e-01 6.32534027e-01 -8.29391181e-01 5.60523152e-01
8.12269092e-01 -2.82423180e-02 3.36155117e-01 -1.27088308e-01
2.79694587e-01 -1.16112955e-01 -9.09297824e-01 1.27098143e+00
-9.98156488e-01 1.08351156e-01 1.93205342e-01 -1.22535050e+00
1.05973113e+00 4.93931293e-01 5.88921487e-01 -1.03589785e+00
3.41460139e-01 1.76483393e-01 3.46639842e-01 -1.14856951e-01
7.44198114e-02 -5.70677459e-01 -1.41768038e-01 5.56841433e-01
5.15762530e-02 2.57588662e-02 -2.98098534e-01 4.39051725e-02
1.10140777e+00 -3.19975108e-01 2.02321604e-01 -4.47279215e-02
5.90860665e-01 -8.91653121e-01 1.03347492e+00 4.63685989e-01
-4.79931176e-01 1.73838660e-01 1.12740159e+00 -2.30468456e-02
-1.10088360e+00 -1.16094649e+00 2.80255020e-01 7.81445861e-01
1.29766420e-01 4.12814528e-01 -5.39077342e-01 -6.29313111e-01
-7.43046999e-02 1.05392087e+00 -2.54851550e-01 -4.75101560e-01
-7.10735992e-02 -6.45124257e-01 4.97402489e-01 3.61455053e-01
7.18722582e-01 -9.22075868e-01 -1.02018222e-01 5.74872315e-01
8.35051760e-02 -1.11319065e+00 -7.16521800e-01 5.36010303e-02
-2.90755361e-01 -2.71142870e-01 -1.01545799e+00 -2.84737736e-01
3.51785392e-01 1.61662232e-02 1.05262363e+00 -3.56189817e-01
1.02668785e-01 2.44047418e-01 -8.72702822e-02 -1.25000104e-01
-4.36612219e-01 2.45046616e-01 -2.26370338e-03 3.49216670e-01
-2.87622273e-01 -1.12939596e+00 -8.35281610e-01 5.90926886e-01
-7.41029024e-01 -1.78972840e-01 4.77650523e-01 7.80159473e-01
5.45509219e-01 4.84558940e-01 1.52242064e+00 -6.46265566e-01
7.61921942e-01 -9.16364312e-01 -7.05635786e-01 2.89178938e-01
-2.77384788e-01 -1.69476476e-02 1.15181673e+00 -2.49267519e-02
-9.78977740e-01 -6.96245492e-01 -2.31992200e-01 -5.79552710e-01
4.79023129e-01 2.24895090e-01 -8.62915456e-01 1.97377518e-01
2.15802297e-01 2.21910939e-01 -4.09503579e-02 1.41854420e-01
4.83340085e-01 7.96444476e-01 1.92766249e-01 -7.10589051e-01
8.97911370e-01 2.21087769e-01 2.87983447e-01 -2.96001196e-01
-8.91947389e-01 5.99831752e-02 1.53907388e-01 -2.40862668e-01
1.72034264e-01 -9.15830791e-01 -7.81528294e-01 2.36003339e-01
-6.87878251e-01 -5.78599215e-01 -6.67855501e-01 2.58264750e-01
-1.00630057e+00 8.02944321e-03 -3.21219176e-01 -9.98776495e-01
-4.03287470e-01 -9.64339197e-01 4.72741365e-01 4.93990034e-01
5.56504965e-01 -7.43094146e-01 -2.19877556e-01 9.98607874e-02
8.21093917e-01 3.99944454e-01 9.00585294e-01 -1.14043288e-01
-7.31013715e-01 2.80965030e-01 -3.78695726e-01 5.99948287e-01
4.24128711e-01 -3.13772827e-01 -4.53093767e-01 -4.83721644e-01
-9.46941897e-02 -1.04448386e-01 1.57758757e-01 8.27920556e-01
1.64226842e+00 -4.71885175e-01 -5.86608946e-02 8.50529492e-01
1.50320172e+00 8.14903438e-01 9.81592774e-01 -7.55305216e-02
4.00311500e-01 8.42496231e-02 5.93575776e-01 1.14967012e+00
3.72866876e-02 7.78676033e-01 8.24995399e-01 6.63047880e-02
1.67849779e-01 -3.01854163e-01 2.08995968e-01 2.05764323e-01
1.45745561e-01 -1.01177216e+00 -2.35853657e-01 3.12610686e-01
-1.59970939e+00 -1.19758356e+00 6.40801489e-01 2.16646504e+00
5.04619002e-01 3.30789924e-01 3.71365130e-01 -4.30334732e-03
6.38784051e-01 9.46464762e-02 -1.07241118e+00 -3.00580353e-01
-6.11204393e-02 1.24394573e-01 8.25480580e-01 2.01151073e-01
-3.15833062e-01 6.64775014e-01 4.72408390e+00 1.27443087e+00
-1.18810511e+00 1.38812307e-02 1.08023870e+00 -2.92018026e-01
-6.24916196e-01 -3.05764288e-01 -1.40403792e-01 1.00999367e+00
1.16333294e+00 -4.88273919e-01 8.37117553e-01 7.29294360e-01
9.55084085e-01 2.52533346e-01 -6.77233934e-01 8.09683383e-01
-6.03034794e-01 -1.33804762e+00 -8.65048021e-02 3.01243484e-01
7.51084507e-01 -1.55446380e-01 2.88259387e-01 5.91314793e-01
3.63836050e-01 -9.72203016e-01 5.98796129e-01 5.47709346e-01
1.04260266e+00 -1.27218831e+00 7.88227975e-01 9.54463705e-02
-8.55017245e-01 -6.54527843e-01 -2.51164913e-01 2.42972583e-01
5.73625505e-01 8.48473907e-01 -7.73904979e-01 6.23933017e-01
-1.65278427e-02 2.75800139e-01 3.67363453e-01 9.70592320e-01
-3.42658430e-01 8.54835808e-01 9.49188098e-02 8.11355039e-02
3.15973163e-01 -3.85727048e-01 5.46920598e-01 4.72826272e-01
6.69531047e-01 1.80967480e-01 2.89276212e-01 9.69571710e-01
-4.91334349e-01 -1.40713617e-01 6.65398547e-03 -5.11196814e-02
1.01523900e+00 1.19318521e+00 -4.38463062e-01 -9.59101617e-02
-1.50814265e-01 7.18459845e-01 -2.19043374e-01 6.88010275e-01
-1.16114998e+00 -3.93236816e-01 1.12267387e+00 5.17942905e-02
3.56470257e-01 -1.00359760e-01 -9.15709287e-02 -5.84731042e-01
-9.22581926e-02 -6.36836767e-01 8.61288682e-02 -6.57892406e-01
-1.21933722e+00 4.27561402e-01 -5.38904130e-01 -1.35289574e+00
-3.33939224e-01 2.28050929e-02 -7.58944213e-01 1.09427285e+00
-1.67033076e+00 -7.80309081e-01 -7.34115317e-02 6.06936812e-01
3.64723772e-01 -5.87993801e-01 4.83915687e-01 5.17567158e-01
-8.06782722e-01 8.95730317e-01 4.42490548e-01 -3.99291143e-03
9.79773402e-02 -1.05969918e+00 1.05997436e-01 6.23219252e-01
-4.00614828e-01 -3.70327570e-02 5.70251942e-01 -7.83140399e-03
-1.05095875e+00 -1.20675516e+00 -3.19791287e-01 4.30042744e-01
5.67202389e-01 -2.04658478e-01 -4.40373600e-01 3.81221205e-01
5.26666641e-02 3.96585613e-01 6.34338081e-01 -4.76449817e-01
2.12171778e-01 -5.17330706e-01 -1.55388224e+00 6.01557672e-01
1.03307033e+00 -2.38368884e-01 5.99210262e-01 1.91053003e-01
1.06568480e+00 -4.40060586e-01 -5.46438813e-01 5.68925031e-02
1.64816052e-01 -8.90890777e-01 5.44464469e-01 -7.59889185e-01
1.70321420e-01 -1.90584034e-01 -4.05573457e-01 -1.83593011e+00
-3.70803267e-01 -1.10868227e+00 -2.61330426e-01 1.34695721e+00
1.24404892e-01 -5.44241011e-01 1.03024912e+00 2.57811457e-01
-2.50141770e-01 -1.09966803e+00 -1.15893090e+00 -1.00306153e+00
-1.01574771e-01 -1.51548564e-01 1.28142738e+00 4.11045581e-01
-3.66639495e-01 2.28534549e-01 -6.79043829e-01 2.81301349e-01
5.39637804e-01 -4.43706810e-02 7.27266133e-01 -4.38307673e-01
-6.37046814e-01 -3.49990845e-01 -3.60956192e-01 -1.23535955e+00
1.83141723e-01 -6.45452678e-01 -1.92057192e-01 -1.32512701e+00
-2.87387252e-01 -9.58931148e-01 -8.19720387e-01 6.02740645e-02
2.00405538e-01 -2.97782362e-01 2.38960221e-01 -1.98123068e-01
-3.16541880e-01 1.06582606e+00 1.62115741e+00 7.69474059e-02
-3.98885936e-01 8.02036166e-01 -9.77265954e-01 -1.39217470e-02
9.90024388e-01 -1.93916529e-01 -1.10222018e+00 -8.01146254e-02
1.24474064e-01 8.67790222e-01 8.65913108e-02 -9.68761563e-01
-3.76659691e-01 -6.48872972e-01 -9.20884833e-02 -2.99915552e-01
3.09811607e-02 -1.06123054e+00 1.68734461e-01 3.05298001e-01
-1.44256800e-01 -4.05405760e-01 -1.09412283e-01 8.76262963e-01
9.93692651e-02 1.90875411e-01 8.47067654e-01 1.23102419e-01
-4.07657355e-01 9.86538231e-01 -3.95685196e-01 3.98555487e-01
1.14057732e+00 4.69829701e-02 6.28326461e-02 -1.15734100e+00
-5.59295058e-01 6.97575331e-01 -9.45089683e-02 1.68742537e-01
3.71783167e-01 -1.44856560e+00 -6.24607742e-01 7.12916926e-02
-2.95593858e-01 -2.02692911e-01 7.47749031e-01 5.10667562e-01
-1.90131396e-01 3.35578591e-01 -3.02756071e-01 -1.68974638e-01
-4.61152762e-01 3.86437297e-01 8.69275570e-01 -4.22430664e-01
-1.30284876e-01 6.25080764e-01 -3.64457741e-02 -1.74357042e-01
1.63607091e-01 -2.60154661e-02 2.75476813e-01 -2.51851559e-01
2.58556366e-01 6.11492276e-01 -3.07958778e-02 2.18782928e-02
-3.21237333e-02 -2.64657229e-01 2.12511659e-01 -8.62673521e-02
1.11677527e+00 -5.89533210e-01 1.50576875e-01 -1.47988349e-01
9.49096859e-01 3.84121835e-02 -1.36882663e+00 -3.94911557e-01
-5.31800032e-01 -5.77177823e-01 3.72118205e-01 -8.14274251e-01
-1.69170356e+00 4.21990842e-01 6.24054670e-01 5.18942058e-01
1.49817848e+00 -1.35398924e-01 1.11406732e+00 -2.77362168e-01
5.77866077e-01 -9.99836922e-01 1.31519571e-01 1.49444044e-01
5.44659376e-01 -6.14590824e-01 -3.34771216e-01 -6.14077598e-02
-6.29226267e-01 9.30354536e-01 6.59696400e-01 1.41140699e-01
7.82765567e-01 1.50325730e-01 -2.70781219e-02 1.70492321e-01
-6.51262879e-01 -9.55041647e-02 -3.12317669e-01 6.53962970e-01
1.13362722e-01 6.34800434e-01 -2.49620713e-02 7.62362838e-01
-8.44704956e-02 5.67656010e-02 7.71713495e-01 3.22235376e-01
-2.19688267e-01 -1.18916702e+00 -2.36826017e-02 5.76740444e-01
-2.97165215e-01 2.21925043e-02 4.25054461e-01 4.07781690e-01
1.99897766e-01 7.96168268e-01 1.77702412e-01 -2.12232500e-01
1.94834560e-01 -4.87832367e-01 2.45291948e-01 -4.44617748e-01
2.83240136e-02 -7.76328668e-02 -1.21948365e-02 -3.19217891e-01
-1.25097126e-01 -3.10802490e-01 -1.21967602e+00 -4.62879747e-01
-1.78184122e-01 3.19070816e-01 2.23948717e-01 1.02614701e+00
4.66115952e-01 1.12697649e+00 1.66121876e+00 -2.65951574e-01
-9.86208200e-01 -5.80309272e-01 -1.12785566e+00 -2.63817102e-01
4.28249538e-01 -4.02837336e-01 -1.96633577e-01 -5.94582975e-01] | [5.9226393699646, 1.6820505857467651] |
996b1633-02e9-4e90-a9b3-58e2130c4fb2 | abaw-learning-from-synthetic-data-multi-task | 2207.01138 | null | https://arxiv.org/abs/2207.01138v2 | https://arxiv.org/pdf/2207.01138v2.pdf | ABAW: Learning from Synthetic Data & Multi-Task Learning Challenges | This paper describes the fourth Affective Behavior Analysis in-the-wild (ABAW) Competition, held in conjunction with European Conference on Computer Vision (ECCV), 2022. The 4th ABAW Competition is a continuation of the Competitions held at IEEE CVPR 2022, ICCV 2021, IEEE FG 2020 and IEEE CVPR 2017 Conferences, and aims at automatically analyzing affect. In the previous runs of this Competition, the Challenges targeted Valence-Arousal Estimation, Expression Classification and Action Unit Detection. This year the Competition encompasses two different Challenges: i) a Multi-Task-Learning one in which the goal is to learn at the same time (i.e., in a multi-task learning setting) all the three above mentioned tasks; and ii) a Learning from Synthetic Data one in which the goal is to learn to recognise the basic expressions from artificially generated data and generalise to real data. The Aff-Wild2 database is a large scale in-the-wild database and the first one that contains annotations for valence and arousal, expressions and action units. This database is the basis for the above Challenges. In more detail: i) s-Aff-Wild2 -- a static version of Aff-Wild2 database -- has been constructed and utilized for the purposes of the Multi-Task-Learning Challenge; and ii) some specific frames-images from the Aff-Wild2 database have been used in an expression manipulation manner for creating the synthetic dataset, which is the basis for the Learning from Synthetic Data Challenge. In this paper, at first we present the two Challenges, along with the utilized corpora, then we outline the evaluation metrics and finally present the baseline systems per Challenge, as well as their derived results. More information regarding the Competition can be found in the competition's website: https://ibug.doc.ic.ac.uk/resources/eccv-2023-4th-abaw/. | ['Dimitrios Kollias'] | 2022-07-03 | null | null | null | null | ['action-unit-detection'] | ['computer-vision'] | [ 2.42019311e-01 -9.06620175e-04 2.78071404e-01 -5.65130591e-01
-1.00032723e+00 -4.17827189e-01 6.33880079e-01 7.12399557e-02
-5.77076733e-01 6.72656775e-01 1.92737788e-01 5.16688883e-01
1.88047558e-01 -6.36156276e-02 -3.61998737e-01 -8.41077328e-01
-2.70747900e-01 3.19194287e-01 -2.11124599e-01 -4.73670006e-01
-2.52943456e-01 2.41051167e-01 -1.84038603e+00 8.14636588e-01
1.17403962e-01 1.35583651e+00 -2.71337956e-01 8.34508717e-01
3.07046831e-01 7.68719673e-01 -7.01764584e-01 -6.06507778e-01
2.41436549e-02 -5.97465336e-01 -8.89851093e-01 -4.94821332e-02
3.59712005e-01 2.61302501e-01 2.44431034e-01 8.27777505e-01
8.77687871e-01 1.88150287e-01 6.29925609e-01 -1.66954827e+00
-2.46670227e-02 1.77119225e-01 -5.28753996e-01 6.18162332e-03
6.30009472e-01 -6.69503212e-02 8.75193357e-01 -9.39988434e-01
7.26480603e-01 1.22452390e+00 6.81635976e-01 9.76467907e-01
-1.04871392e+00 -5.47722578e-01 9.24181100e-03 4.49143291e-01
-1.27051497e+00 -5.27724147e-01 8.59045386e-01 -5.48485398e-01
1.11805904e+00 4.74630773e-01 8.80695581e-01 1.92884469e+00
-1.69348672e-01 1.19077992e+00 1.56868660e+00 -5.16272783e-01
2.80447364e-01 2.17253476e-01 2.47210905e-01 3.20166975e-01
-5.83500862e-01 -1.11028692e-02 -5.31438649e-01 -8.16549584e-02
1.32957548e-01 -8.21720958e-01 -1.65726155e-01 -8.77273679e-02
-8.74871254e-01 7.50025511e-01 6.42938763e-02 6.15449548e-01
-4.89540637e-01 1.56547695e-01 1.27027261e+00 4.98070031e-01
6.45972371e-01 3.02854419e-01 -5.16902983e-01 -6.04613781e-01
-6.49560392e-01 5.07874250e-01 8.08610141e-01 5.70066929e-01
5.08822083e-01 1.75951242e-01 -4.56692576e-01 1.27246368e+00
1.59709513e-01 1.91742331e-01 6.23135149e-01 -1.04739070e+00
5.72416522e-02 2.93848366e-01 -7.08804652e-03 -6.47791445e-01
-6.41761422e-01 1.62894621e-01 -8.10191572e-01 3.56151134e-01
2.26762816e-01 -4.76693541e-01 -7.61167288e-01 2.12815213e+00
1.38614312e-01 -2.93239690e-02 2.12876365e-01 7.12109685e-01
1.05087125e+00 6.45943642e-01 2.82379955e-01 -6.73241079e-01
1.43645132e+00 -8.59963298e-01 -1.00631535e+00 -1.51333034e-01
5.96944213e-01 -5.94995737e-01 9.22801137e-01 6.61369503e-01
-1.07239616e+00 -6.46382153e-01 -1.15522242e+00 1.61653072e-01
-6.57028198e-01 1.60580426e-01 5.60112715e-01 6.31586850e-01
-1.33253491e+00 9.09693465e-02 -4.91000921e-01 -4.39308792e-01
3.25362891e-01 2.34389499e-01 -4.32247460e-01 4.10356462e-01
-1.55605316e+00 1.13601124e+00 2.40589291e-01 5.80794849e-02
-8.88109982e-01 -3.60688686e-01 -9.77987111e-01 -4.66080546e-01
3.54989991e-02 -1.45621970e-01 1.41647136e+00 -1.77542865e+00
-1.47653401e+00 1.64056456e+00 7.85804018e-02 -4.02295440e-01
5.97026944e-01 -1.68254793e-01 -6.91816032e-01 -3.79257984e-02
-1.55529678e-01 9.59965408e-01 6.89459860e-01 -1.29296863e+00
-3.76567751e-01 -4.84695166e-01 -4.05980796e-02 1.01846457e-01
-3.87527235e-03 3.05470079e-01 -5.42775810e-01 -5.22935867e-01
-8.64546418e-01 -1.09429860e+00 6.50966167e-02 -4.02848989e-01
-2.26040021e-01 -4.74737495e-01 9.01577771e-01 -3.77274245e-01
1.07699192e+00 -2.36799622e+00 3.27511579e-01 8.38804767e-02
5.89240864e-02 4.84734148e-01 -2.10667938e-01 3.56616288e-01
-5.83736956e-01 -2.24746570e-01 -1.47065863e-01 -6.43076539e-01
1.63057074e-01 2.12129489e-01 2.56471150e-02 5.07890701e-01
1.28977507e-01 8.68950367e-01 -6.79300666e-01 -3.07391614e-01
3.04747224e-01 5.66467583e-01 5.59768006e-02 2.33651802e-01
-1.03513248e-01 3.68153930e-01 -1.45005837e-01 3.97273988e-01
3.60544562e-01 4.26194131e-01 5.13609983e-02 -4.45953935e-01
-5.82356751e-02 -4.70800281e-01 -9.62538958e-01 1.66370702e+00
-6.05663419e-01 8.59832466e-01 2.77281910e-01 -1.19029951e+00
1.06622970e+00 7.56426930e-01 7.71888793e-01 -8.30021143e-01
4.56378043e-01 1.93162560e-01 -1.31149918e-01 -6.37970924e-01
1.42732039e-01 -3.02768618e-01 -5.54097712e-01 -8.15525725e-02
5.16375542e-01 -2.14461565e-01 4.91689175e-01 -4.21883576e-02
9.73756135e-01 3.61070633e-01 5.59302807e-01 -2.40459844e-01
8.73069882e-01 -3.58967543e-01 7.21175790e-01 1.65824816e-01
-7.57403970e-01 4.06343460e-01 7.73108006e-01 -6.14654064e-01
-8.59886289e-01 -8.63165438e-01 -1.68649495e-01 9.84996796e-01
-2.40833223e-01 -5.69751859e-01 -8.89096141e-01 -4.51795042e-01
-4.28702265e-01 8.56431663e-01 -1.20016456e+00 -1.57726899e-01
-1.98695138e-01 -7.71457672e-01 9.28674698e-01 4.12910908e-01
5.77488303e-01 -1.59368932e+00 -9.30358052e-01 -6.38128296e-02
-4.96446401e-01 -1.23303235e+00 -2.98643783e-02 4.43654895e-01
-5.91262244e-02 -1.02171290e+00 -5.64062238e-01 -5.86005807e-01
3.37109938e-02 -6.89976215e-01 1.21435821e+00 -3.70970100e-01
-6.48604333e-01 7.03190148e-01 -5.97528040e-01 -1.01088524e+00
-4.49200451e-01 -4.56677735e-01 -4.57868800e-02 4.26075995e-01
7.10506856e-01 -4.93465394e-01 -2.15571493e-01 8.54000077e-02
-7.03092396e-01 5.56122104e-04 3.57776821e-01 6.98944271e-01
7.63447642e-01 -6.34029448e-01 6.62099123e-01 -7.66991854e-01
7.16848612e-01 -9.11915004e-02 -2.42640555e-01 2.52557844e-01
-8.78787860e-02 -4.55782086e-01 4.08821464e-01 -2.68627614e-01
-1.04758644e+00 3.27064872e-01 -5.78168869e-01 -6.84323847e-01
-4.86784250e-01 6.86195418e-02 -4.02847528e-01 1.92538857e-01
7.30176330e-01 -6.87445030e-02 -2.23396093e-01 -4.76139672e-02
3.06824863e-01 8.11652184e-01 6.04898751e-01 -4.91041034e-01
-6.25287369e-03 1.94372624e-01 9.20892432e-02 -9.86454129e-01
-9.74366248e-01 -6.46327794e-01 -6.89403772e-01 -7.37503290e-01
1.41257799e+00 -1.00482941e+00 -6.19027436e-01 9.43466425e-01
-1.09598362e+00 -6.50412560e-01 -4.37218666e-01 1.10234417e-01
-9.77013826e-01 -1.59283206e-01 -6.08937800e-01 -1.07564175e+00
-5.23997068e-01 -1.02565920e+00 1.24726200e+00 1.54299632e-01
-7.42258966e-01 -1.15077293e+00 5.92190564e-01 4.97207522e-01
1.42964363e-01 1.05939758e+00 5.43995142e-01 -7.23570108e-01
7.69275904e-01 -1.28201127e-01 2.14228362e-01 8.70125175e-01
-1.88365340e-01 3.77673730e-02 -1.60460413e+00 -7.17831105e-02
5.83953820e-02 -1.20373213e+00 8.38866591e-01 2.98456162e-01
1.15680158e+00 1.58929721e-01 3.38994190e-02 3.80620807e-01
1.28330350e+00 3.26555043e-01 9.20157015e-01 1.14138998e-01
2.52105772e-01 7.70369470e-01 8.56868565e-01 4.89872694e-01
2.37251744e-01 1.02078390e+00 3.58938485e-01 -1.69402584e-01
2.14030892e-02 3.75484258e-01 5.61609030e-01 7.31619060e-01
-3.44256759e-01 -1.40875176e-01 -8.08483958e-01 5.97230375e-01
-1.91797090e+00 -1.10964799e+00 -2.79798269e-01 1.90383923e+00
8.35633337e-01 -1.33973300e-01 5.14662385e-01 2.46110812e-01
4.80165958e-01 3.63578886e-01 -2.54094958e-01 -1.20979118e+00
-3.55986029e-01 4.09587264e-01 -2.05269873e-01 5.21396637e-01
-1.46682906e+00 5.75589299e-01 4.96356630e+00 9.52324331e-01
-1.07048309e+00 3.70868623e-01 8.91050994e-01 -4.36308026e-01
2.39037544e-01 -5.95517099e-01 -5.60913324e-01 3.09445798e-01
1.32955348e+00 -6.63795918e-02 1.64860621e-01 7.61438549e-01
2.71933347e-01 -1.86106607e-01 -1.08525419e+00 1.41764212e+00
4.98238057e-01 -6.94964528e-01 -3.69554609e-01 -2.29137167e-01
4.51885998e-01 1.61411837e-01 -1.24693155e-01 8.17965508e-01
-1.59020215e-01 -1.02859807e+00 7.64979243e-01 3.69863629e-01
1.10411143e+00 -8.39321434e-01 9.16241109e-01 -3.70879211e-02
-1.03776503e+00 7.97545351e-03 8.26107487e-02 6.77227825e-02
1.69826716e-01 4.23931479e-01 5.99389672e-02 5.00314593e-01
9.15233850e-01 7.04161346e-01 -4.35018957e-01 5.08098304e-01
-2.77611703e-01 4.22286302e-01 5.67630418e-02 -2.12586597e-01
3.61195862e-01 -1.47964895e-01 4.76709723e-01 1.77958584e+00
-8.02143514e-02 -5.38847968e-02 1.50190890e-02 4.57645208e-01
-6.82462677e-02 1.72337249e-01 -4.84358788e-01 1.27163216e-01
-5.36859408e-02 1.86359107e+00 -3.68502438e-01 -3.57633591e-01
-3.51725817e-01 1.25638640e+00 2.89608598e-01 1.45498529e-01
-1.08856082e+00 -4.07675087e-01 7.54687726e-01 -3.72533411e-01
2.77589299e-02 4.63963032e-01 2.09342360e-01 -8.41466367e-01
-7.35449195e-02 -1.03402495e+00 4.66906339e-01 -1.02988756e+00
-1.24958646e+00 9.81451273e-01 3.48014712e-01 -8.94980371e-01
-5.23158550e-01 -7.96527624e-01 -5.73411047e-01 6.98312342e-01
-9.82603014e-01 -1.36633456e+00 -5.41305542e-01 7.92598546e-01
6.77679598e-01 -4.46469933e-02 1.30630422e+00 3.48850578e-01
-5.93132496e-01 5.64232767e-01 -2.46945575e-01 1.53950796e-01
9.29163635e-01 -1.46302986e+00 -5.18147647e-02 3.17994058e-01
1.68136522e-01 -2.43559256e-01 8.21791172e-01 -2.87823398e-02
-8.78986537e-01 -1.07050169e+00 7.33687341e-01 -4.08506930e-01
5.85815907e-01 -7.93142319e-01 -6.00032389e-01 6.39128625e-01
5.19508362e-01 6.76663145e-02 9.51683640e-01 -4.87140566e-02
-2.50078559e-01 -1.03363998e-01 -1.15664828e+00 2.84478277e-01
6.32165730e-01 -3.80937040e-01 -4.59124535e-01 3.40114743e-01
1.71697542e-01 -4.63333935e-01 -9.63098884e-01 5.18701077e-01
7.65131056e-01 -1.25194705e+00 6.26107693e-01 -4.09798652e-01
4.82765317e-01 4.38563563e-02 -4.71553653e-01 -1.57559645e+00
1.05215926e-02 -6.01389408e-01 -6.61221966e-02 1.42666185e+00
3.09087217e-01 -2.15063006e-01 3.46940547e-01 2.12572649e-01
-1.31636426e-01 -1.12731016e+00 -1.23318064e+00 -5.30835211e-01
-2.62594409e-02 -7.87236333e-01 -1.37002960e-01 8.77080202e-01
-4.50479090e-02 7.87251592e-01 -3.81320268e-01 -5.21056294e-01
4.47412997e-01 -1.87966079e-01 6.37249231e-01 -8.96766543e-01
-3.64267200e-01 -5.27795970e-01 -7.81938016e-01 -1.28821880e-01
3.69833201e-01 -6.28668368e-01 7.03657493e-02 -1.16256380e+00
1.38975635e-01 3.70675772e-01 -3.32469523e-01 6.83710158e-01
-5.38226959e-05 5.14966905e-01 3.30956727e-01 -2.59688467e-01
-7.38141418e-01 6.31083369e-01 7.59046078e-01 -1.23770192e-01
1.93307344e-02 -5.21969376e-03 -4.17421281e-01 8.13623488e-01
8.90532613e-01 8.72171205e-03 -4.86108989e-01 1.75770432e-01
1.52593106e-01 -2.65093476e-01 4.35874969e-01 -9.41300750e-01
-3.39869052e-01 2.04364464e-01 4.01364416e-01 -3.12860042e-01
9.43857491e-01 -5.66745281e-01 -1.67622659e-02 1.90444872e-01
-3.45860511e-01 -6.98454399e-03 7.12799668e-01 1.44754916e-01
-4.79281873e-01 -2.21260518e-01 1.18949580e+00 -1.38860941e-01
-1.12092423e+00 -1.83989964e-02 -5.27629435e-01 2.43771136e-01
1.63840520e+00 -1.46647841e-02 -4.34193239e-02 -4.96645600e-01
-1.10708177e+00 3.89240742e-01 4.68843020e-02 6.15970492e-01
4.39572424e-01 -1.38370860e+00 -9.48428869e-01 -6.87715858e-02
5.38208246e-01 -5.20479679e-01 4.94322747e-01 1.12739754e+00
-1.41043022e-01 1.72163785e-01 -5.69207966e-01 -5.50758541e-01
-1.84960353e+00 3.97711486e-01 6.81761146e-01 -4.50089753e-01
-1.40049383e-02 9.37854111e-01 9.33359489e-02 -3.30799520e-01
1.80798203e-01 1.65262938e-01 -5.25207818e-01 5.40449977e-01
6.03907466e-01 1.79924801e-01 1.69146433e-01 -9.96549964e-01
-5.59554696e-01 3.83315742e-01 2.15908542e-01 -3.00178528e-01
1.39980543e+00 -2.89468206e-02 -1.73842996e-01 1.21188498e+00
1.48265469e+00 -3.96553814e-01 -9.40255046e-01 2.73737490e-01
1.06459193e-01 -6.66122232e-03 -3.29928957e-02 -1.07682621e+00
-9.21308458e-01 1.00868189e+00 9.44065511e-01 9.57044363e-02
1.53691769e+00 -6.11350965e-03 3.43274415e-01 1.82780288e-02
2.03297928e-01 -1.49036920e+00 1.07329354e-01 4.10354793e-01
1.38219690e+00 -1.06335819e+00 -3.00827026e-01 -1.75193101e-01
-1.20987535e+00 1.12783289e+00 8.09499681e-01 5.92662431e-02
4.48355794e-01 4.84350532e-01 4.02138472e-01 -4.30777997e-01
-1.07359076e+00 -3.22066754e-01 1.96306810e-01 7.97911525e-01
6.95925474e-01 5.70501685e-02 -2.34980553e-01 7.78399467e-01
-7.85759315e-02 7.64441639e-02 3.99437040e-01 6.00510597e-01
8.23166221e-02 -1.06246138e+00 -8.19544420e-02 3.66076052e-01
-5.85405171e-01 3.26375008e-01 -1.00203228e+00 8.53383362e-01
5.29597402e-01 8.71480227e-01 -2.52708159e-02 -5.61871648e-01
7.90084600e-01 4.70011175e-01 7.23683715e-01 -3.73280078e-01
-9.48092639e-01 1.40215680e-01 7.23642647e-01 -8.95006776e-01
-9.47181284e-01 -9.41473126e-01 -1.03218913e+00 1.01300232e-01
1.55742258e-01 2.69737214e-01 7.52402365e-01 7.82715201e-01
-4.61894795e-02 7.11672366e-01 4.72389221e-01 -1.10410333e+00
-1.85054123e-01 -1.17940056e+00 -8.12798381e-01 8.69850516e-01
6.16476834e-02 -6.73268676e-01 -3.49221587e-01 1.63548589e-01] | [13.571793556213379, 2.1999659538269043] |
5b72a6b9-a400-4373-a3fa-a0f850988f53 | unsupervised-few-shot-learning-via-self-1 | 1912.12178 | null | https://arxiv.org/abs/1912.12178v1 | https://arxiv.org/pdf/1912.12178v1.pdf | Unsupervised Few-shot Learning via Self-supervised Training | Learning from limited exemplars (few-shot learning) is a fundamental, unsolved problem that has been laboriously explored in the machine learning community. However, current few-shot learners are mostly supervised and rely heavily on a large amount of labeled examples. Unsupervised learning is a more natural procedure for cognitive mammals and has produced promising results in many machine learning tasks. In the current study, we develop a method to learn an unsupervised few-shot learner via self-supervised training (UFLST), which can effectively generalize to novel but related classes. The proposed model consists of two alternate processes, progressive clustering and episodic training. The former generates pseudo-labeled training examples for constructing episodic tasks; and the later trains the few-shot learner using the generated episodic tasks which further optimizes the feature representations of data. The two processes facilitate with each other, and eventually produce a high quality few-shot learner. Using the benchmark dataset Omniglot and Mini-ImageNet, we show that our model outperforms other unsupervised few-shot learning methods. Using the benchmark dataset Market1501, we further demonstrate the feasibility of our model to a real-world application on person re-identification. | ['Zilong Ji', 'Si Wu', 'Xiaolong Zou', 'Tiejun Huang'] | 2019-12-20 | null | null | null | null | ['unsupervised-few-shot-learning', 'unsupervised-few-shot-image-classification'] | ['computer-vision', 'computer-vision'] | [ 2.86549360e-01 -7.48227686e-02 -2.80092120e-01 -4.96634245e-01
-5.12575626e-01 2.16280669e-02 7.36357272e-01 1.61139876e-01
-6.45043373e-01 9.21868503e-01 2.52833813e-01 4.23989981e-01
-2.83503681e-01 -1.01867306e+00 -5.22506773e-01 -6.25975192e-01
1.32691320e-02 7.45600879e-01 4.46604252e-01 -1.26911685e-01
1.22625433e-01 -7.23919347e-02 -2.01629996e+00 1.18393518e-01
1.07210624e+00 8.22730839e-01 3.48843664e-01 3.43579143e-01
-3.56889665e-01 8.42433274e-01 -4.48494524e-01 -3.74930292e-01
7.66335651e-02 -6.98614419e-01 -8.09607506e-01 4.14461762e-01
-7.60969287e-03 2.47839764e-02 -2.65463084e-01 1.01372409e+00
6.93396330e-01 9.18874383e-01 7.76797533e-01 -1.18400538e+00
-7.27978885e-01 6.04039490e-01 -2.97324568e-01 3.05875510e-01
1.28358021e-01 7.26343393e-02 6.94092751e-01 -1.18951190e+00
6.88428164e-01 9.92192566e-01 6.89970493e-01 8.42088997e-01
-1.01602185e+00 -6.10079706e-01 -1.02707997e-01 6.13807142e-01
-1.51816487e+00 -5.07124722e-01 7.38006055e-01 -2.97163367e-01
8.24911952e-01 -2.26255476e-01 8.05320263e-01 9.76544023e-01
-1.40885532e-01 8.32439244e-01 1.03310490e+00 -5.91178060e-01
7.23078609e-01 3.92839730e-01 4.41902101e-01 4.58365440e-01
3.09693236e-02 2.84200460e-01 -7.18596578e-01 9.19825435e-02
3.57059538e-01 5.95511913e-01 -7.89080411e-02 -4.62411821e-01
-7.93969572e-01 9.12527502e-01 2.78864473e-01 4.72482085e-01
-3.12326998e-01 -2.13837028e-01 3.96963418e-01 3.53836179e-01
4.57795173e-01 3.84465396e-01 1.45364031e-02 -1.62860408e-01
-9.18417215e-01 1.05932154e-01 6.11273408e-01 1.02920175e+00
1.05096436e+00 7.09467009e-02 -2.46769086e-01 1.31089520e+00
2.44862046e-02 2.28010312e-01 1.16638196e+00 -4.44959313e-01
5.07599749e-02 5.92329264e-01 -3.15652750e-02 -4.08572763e-01
-2.75231242e-01 -4.88280237e-01 -7.35393763e-01 -1.03800207e-01
1.41362667e-01 -2.33221844e-01 -1.08094931e+00 1.59675121e+00
2.61030793e-01 6.40766919e-01 2.02902988e-01 7.28416383e-01
8.60348344e-01 6.88737631e-01 3.73155355e-01 -4.59505618e-01
1.01667976e+00 -1.12302840e+00 -7.39995539e-01 -2.89775997e-01
6.81845069e-01 -2.01840147e-01 1.14237583e+00 2.05311522e-01
-8.37117732e-01 -9.19337869e-01 -1.13189399e+00 3.96009266e-01
-7.63610661e-01 -2.41390616e-01 6.11175418e-01 7.71821856e-01
-6.60096765e-01 8.53150427e-01 -4.74687517e-01 -8.01260114e-01
7.76047647e-01 2.43628964e-01 -1.65177107e-01 -3.16734135e-01
-1.42939663e+00 8.16603303e-01 9.10554767e-01 -5.35832167e-01
-9.20425177e-01 -5.53273797e-01 -9.27397907e-01 2.67695636e-01
6.73864186e-01 -4.90778446e-01 1.22808909e+00 -8.96244347e-01
-1.54038501e+00 7.39866316e-01 -7.46173561e-02 -7.98193038e-01
-3.87059711e-02 -1.40434682e-01 -7.16767788e-01 2.54781581e-02
2.30316311e-01 6.72404528e-01 1.01378155e+00 -1.01041234e+00
-8.27351987e-01 -3.86884898e-01 -3.77910912e-01 3.67197186e-01
-8.64863098e-01 -3.79826993e-01 -3.27824831e-01 -7.59607017e-01
-2.98524380e-01 -7.55959749e-01 -1.07848048e-01 -4.57178414e-01
8.66181925e-02 -5.71491539e-01 7.04456389e-01 -6.05870895e-02
1.16137373e+00 -2.16840410e+00 1.24049850e-01 4.84232046e-02
1.00492286e-02 5.48025310e-01 1.05287954e-02 1.44551039e-01
-1.15733348e-01 -3.53421450e-01 -4.58086729e-01 -3.10034931e-01
-7.54112191e-03 3.18574727e-01 -3.32523793e-01 1.20731115e-01
-5.05749509e-02 1.11370206e+00 -1.24397421e+00 -4.74933147e-01
3.41215819e-01 9.43068415e-02 -3.43864322e-01 3.69212121e-01
-2.68210489e-02 2.87999660e-01 -1.10235021e-01 5.10055482e-01
2.21627891e-01 -3.20512086e-01 2.14716047e-02 2.24916175e-01
4.81269928e-03 -3.05561662e-01 -1.26327193e+00 1.79550052e+00
-2.72625834e-01 3.67477655e-01 -9.45320845e-01 -1.29806447e+00
9.84718025e-01 1.82909608e-01 3.57352495e-01 -7.33868897e-01
2.93197691e-01 1.23773687e-01 -1.06226861e-01 -5.35533428e-01
3.50943804e-01 -5.95512211e-01 -9.05033574e-02 7.04496980e-01
8.98160040e-01 4.10176553e-02 4.80477095e-01 2.54644424e-01
8.09966445e-01 7.02790320e-02 5.22009730e-01 -6.72608688e-02
3.22084069e-01 7.59253977e-04 6.65912807e-01 1.02987993e+00
-3.76068562e-01 5.35829902e-01 -2.34260559e-01 -5.68705916e-01
-9.55236435e-01 -1.20458913e+00 -6.93254471e-02 1.48518181e+00
2.90363908e-01 -5.95466316e-01 -7.81296670e-01 -5.89123964e-01
-1.07452631e-01 9.86278474e-01 -7.26230145e-01 -6.41006529e-01
-1.07858360e-01 -9.42732453e-01 2.03575149e-01 7.26813495e-01
7.31181324e-01 -1.47688925e+00 -6.46845758e-01 3.26297015e-01
7.73794502e-02 -7.36311495e-01 -3.41966510e-01 3.71908486e-01
-8.09821844e-01 -9.59808826e-01 -9.68822837e-01 -1.01966894e+00
5.70206761e-01 4.22202766e-01 9.27264929e-01 4.65563685e-02
-5.75255692e-01 5.77301025e-01 -6.09790206e-01 -5.59783697e-01
5.91867417e-02 3.69232632e-02 5.51480055e-01 3.32433552e-01
9.02866840e-01 -7.02601433e-01 -2.97798723e-01 2.70500004e-01
-8.63661885e-01 -2.03426301e-01 4.81108516e-01 1.19943118e+00
7.12600708e-01 3.91291320e-01 1.12262583e+00 -1.27001071e+00
6.82215452e-01 -6.23081088e-01 -2.13608876e-01 3.48482013e-01
-7.33732402e-01 1.29912093e-01 6.63927138e-01 -7.82737970e-01
-1.37017131e+00 6.74106181e-02 4.38215658e-02 -5.24511158e-01
-2.63211519e-01 5.17058432e-01 -1.46180764e-01 7.23050311e-02
8.70013833e-01 6.53080046e-01 -2.21196637e-01 -4.14524317e-01
5.55899978e-01 7.17513680e-01 7.10983157e-01 -4.65423405e-01
7.73662031e-01 4.20637608e-01 -5.25670588e-01 -1.09313321e+00
-1.14814901e+00 -6.88279867e-01 -9.20563519e-01 -3.76940161e-01
7.91295469e-01 -7.17086971e-01 -2.48249918e-01 4.66704100e-01
-5.49710572e-01 -3.91968787e-01 -9.28011894e-01 5.84090292e-01
-7.22006738e-01 8.23790953e-02 -2.83225983e-01 -7.57916749e-01
-2.35137582e-01 -6.26166999e-01 4.93600875e-01 8.43998373e-01
-2.31702551e-01 -9.01263177e-01 3.65680099e-01 -1.88238770e-02
3.70642960e-01 -3.18909079e-01 7.03051627e-01 -1.01474249e+00
-1.15236074e-01 -2.73361802e-02 1.07516609e-01 3.02737299e-02
3.45580041e-01 -7.39028275e-01 -1.05477655e+00 -3.77894491e-01
2.59784430e-01 -8.26253951e-01 1.14236128e+00 1.84271201e-01
1.16669488e+00 2.69900467e-02 -3.79913151e-01 5.10032833e-01
1.22427094e+00 2.54558265e-01 6.52793229e-01 2.82224923e-01
4.30535495e-01 5.24726868e-01 6.83088779e-01 5.60743570e-01
6.60611838e-02 3.55227292e-01 -2.10662290e-01 3.99662197e-01
-1.46276325e-01 -3.07051271e-01 -3.14266570e-02 9.50129986e-01
-2.94877708e-01 2.51330823e-01 -8.62182558e-01 7.35171318e-01
-2.12531185e+00 -1.46211851e+00 6.45212770e-01 2.23268151e+00
8.14804375e-01 3.10070783e-01 3.42247456e-01 1.31869033e-01
9.48652208e-01 5.09128943e-02 -8.62919629e-01 4.67982143e-02
2.09493488e-02 5.23386598e-01 4.76432331e-02 -3.92555911e-03
-1.20645046e+00 1.19572794e+00 5.85636759e+00 9.10144985e-01
-6.70857012e-01 3.01066667e-01 3.78807247e-01 -2.73153335e-01
2.34848574e-01 -1.39442250e-01 -9.05248642e-01 5.67422152e-01
1.05460811e+00 -6.13371372e-01 4.09231544e-01 1.10102987e+00
-2.29950368e-01 -1.86675973e-02 -9.29909468e-01 1.22287405e+00
4.80012447e-01 -1.33760190e+00 1.97852030e-01 -3.05035979e-01
9.85604286e-01 -2.55796045e-01 7.62874410e-02 9.20037866e-01
3.28568220e-01 -8.75044167e-01 2.24354863e-01 6.72707081e-01
7.31856167e-01 -1.11452723e+00 5.81730068e-01 5.85003614e-01
-1.21221006e+00 -5.54796576e-01 -7.80617416e-01 -2.58146375e-02
2.88500100e-01 2.54753441e-01 -5.85816443e-01 3.34043652e-01
7.76460648e-01 1.07755673e+00 -6.61910355e-01 1.39986873e+00
-2.80482948e-01 5.51276684e-01 1.18460260e-01 -1.00711085e-01
9.50116068e-02 -7.89874569e-02 3.51586938e-01 8.77095580e-01
2.17705667e-01 5.10769963e-01 2.88934022e-01 6.60454333e-01
-1.91353559e-01 1.95976675e-01 -6.96986914e-01 -1.41846105e-01
6.04041755e-01 1.09495270e+00 -9.12304044e-01 -8.07452142e-01
-5.95708430e-01 1.05614591e+00 6.89208508e-01 2.15607300e-01
-5.99938571e-01 -6.89235806e-01 1.46021649e-01 6.83819428e-02
4.71801639e-01 1.78447306e-01 5.99205866e-02 -1.31900525e+00
-4.72300500e-01 -6.64752245e-01 7.28069127e-01 -5.98605216e-01
-1.63764894e+00 5.66411018e-01 3.99422683e-02 -1.45353985e+00
-4.78637040e-01 -2.68144637e-01 -8.92655194e-01 4.22659934e-01
-1.35753977e+00 -7.34523475e-01 -4.75924581e-01 9.34655786e-01
1.09282696e+00 -7.93911457e-01 1.02368391e+00 3.30852777e-01
-4.97736037e-01 5.47803879e-01 2.28565827e-01 1.44009918e-01
8.57925415e-01 -1.11944711e+00 2.60386646e-01 7.08400548e-01
3.83155227e-01 7.97863066e-01 5.35627544e-01 -8.27036679e-01
-9.98785913e-01 -1.33183086e+00 6.55899048e-01 -1.09667607e-01
4.54903215e-01 -2.60647267e-01 -1.06292367e+00 6.68778896e-01
1.21704221e-01 5.23273572e-02 1.12326789e+00 1.63790375e-01
-1.56397238e-01 -3.12547348e-02 -9.60610151e-01 5.68424523e-01
1.18214190e+00 -4.85747963e-01 -1.24440908e+00 4.34880912e-01
6.48839116e-01 1.98071837e-01 -5.83771825e-01 1.20442860e-01
3.51088554e-01 -8.84903550e-01 9.50513065e-01 -8.71650875e-01
1.12812229e-01 -3.70398797e-02 -1.40523363e-03 -1.74725890e+00
-5.03148615e-01 -4.70053732e-01 -3.78990889e-01 1.23540878e+00
3.56633998e-02 -4.59794551e-01 7.20152617e-01 4.92897332e-01
-9.18388814e-02 -5.62073052e-01 -5.99841475e-01 -1.15204322e+00
-2.13641360e-01 -2.43974358e-01 5.70362329e-01 1.09932327e+00
3.57333601e-01 6.01417065e-01 -6.14550412e-01 -4.32479799e-01
9.56947625e-01 1.57175422e-01 6.39929235e-01 -1.59783804e+00
-3.02285403e-01 -1.18261293e-01 -3.88320118e-01 -7.69011915e-01
2.83849061e-01 -1.06218767e+00 2.13149697e-01 -1.23249912e+00
5.61645269e-01 -1.88078582e-01 -4.92666423e-01 4.06724006e-01
-3.62389088e-01 2.49201834e-01 1.51159927e-01 3.23639750e-01
-1.11392522e+00 8.77261758e-01 7.69152999e-01 -2.41985187e-01
-5.58529913e-01 1.30165547e-01 -6.52582526e-01 8.04934204e-01
8.44420493e-01 -5.87345481e-01 -7.82521248e-01 8.27210695e-02
-3.64566952e-01 -4.46086913e-01 1.27029538e-01 -1.47026765e+00
5.93813837e-01 -3.99532393e-02 7.23200440e-01 -4.89317924e-01
4.61352140e-01 -5.72646797e-01 -1.37931094e-01 5.36161482e-01
-3.39943647e-01 -4.11297619e-01 -2.01857626e-01 8.20025742e-01
-1.44232363e-01 -6.39834702e-01 1.06960702e+00 -4.73844141e-01
-1.42699456e+00 6.01597309e-01 -2.22076133e-01 3.08572680e-01
1.29211891e+00 -3.24095905e-01 -1.66694567e-01 -3.10286820e-01
-1.14922869e+00 2.58507729e-01 2.49651194e-01 5.97144902e-01
9.06674683e-01 -1.73263562e+00 -4.49283868e-01 3.67066443e-01
4.51885462e-01 -4.20644999e-01 4.32769060e-01 4.48330671e-01
1.87436178e-01 2.76996851e-01 -5.50966859e-01 -4.64672625e-01
-9.43699598e-01 1.05835140e+00 4.74205688e-02 4.57056761e-02
-8.18021357e-01 6.87921643e-01 -1.50597095e-01 -4.01067853e-01
3.78195882e-01 5.22128165e-01 -6.29352391e-01 3.22192401e-01
1.01690936e+00 5.16522050e-01 -2.81339556e-01 -5.50676703e-01
-4.30127718e-02 2.39734352e-01 -3.52740645e-01 -1.04741074e-01
1.70621884e+00 -1.22661270e-01 3.67499352e-01 8.61469328e-01
1.03669357e+00 -7.41283774e-01 -1.14663899e+00 -7.18857408e-01
3.08906317e-01 -5.04136026e-01 -3.14400941e-01 -3.60429436e-01
-6.89810872e-01 8.67477894e-01 8.05267155e-01 6.94943145e-02
9.10519779e-01 1.12764187e-01 9.51830029e-01 7.07810104e-01
6.39675558e-01 -1.75888109e+00 5.53711772e-01 5.47218561e-01
2.17382938e-01 -1.30989039e+00 -6.93267211e-02 -6.08722605e-02
-8.08587313e-01 8.41597497e-01 8.72691751e-01 -2.62575626e-01
6.76714599e-01 -1.84299052e-01 -3.80209208e-01 -1.15017131e-01
-6.17583334e-01 -6.13726914e-01 2.32972324e-01 8.01417291e-01
5.08854054e-02 -1.50880203e-01 -3.31147790e-01 1.16250932e+00
-1.65082067e-02 3.25063616e-01 2.23367676e-01 1.15089989e+00
-9.66451764e-01 -1.05069268e+00 -1.13386475e-01 8.41781020e-01
3.88460159e-02 -1.07050501e-01 -1.63233742e-01 4.87797767e-01
2.35832080e-01 7.90020108e-01 1.44934967e-01 -3.60636353e-01
2.57856965e-01 5.12696326e-01 4.59208250e-01 -1.14716375e+00
-2.63372719e-01 -1.36385471e-01 -3.35772753e-01 -2.71593332e-01
-6.11368656e-01 -6.14467323e-01 -1.20927489e+00 2.27304567e-02
-2.71593243e-01 4.57306087e-01 1.60380706e-01 1.26925039e+00
2.03965470e-01 4.60756332e-01 6.06779635e-01 -8.19233298e-01
-4.92161632e-01 -1.00396764e+00 -9.34778035e-01 7.84668624e-01
-2.32699513e-01 -1.18008268e+00 -3.10587406e-01 2.56428301e-01] | [10.042914390563965, 3.125276565551758] |
d532d9da-a648-4523-857a-cd3129900b8b | q-based-equilibria | 2304.12647 | null | https://arxiv.org/abs/2304.12647v1 | https://arxiv.org/pdf/2304.12647v1.pdf | Q-based Equilibria | In dynamic environments, Q-learning is an adaptative rule that provides an estimate (a Q-value) of the continuation value associated with each alternative. A naive policy consists in always choosing the alternative with highest Q-value. We consider a family of Q-based policy rules that may systematically favor some alternatives over others, for example rules that incorporate a leniency bias that favors cooperation. In the spirit of Compte and Postlewaite [2018], we look for equilibrium biases (or Qb-equilibria) within this family of Q-based rules. We examine classic games under various monitoring technologies. | ['Olivier Compte'] | 2023-04-25 | null | null | null | null | ['q-learning'] | ['methodology'] | [-4.32957917e-01 3.57774407e-01 -6.97131336e-01 -1.09436050e-01
-2.76426256e-01 -8.25908244e-01 6.76413178e-01 2.21555665e-01
-1.03417814e+00 1.15527809e+00 2.73655057e-01 -6.02183342e-01
-4.67775077e-01 -9.24457729e-01 -4.84959453e-01 -6.10971868e-01
-4.01977450e-01 2.28528768e-01 4.58980910e-02 -4.22432780e-01
5.35524189e-01 9.34655499e-03 -1.24520171e+00 -1.74710646e-01
8.98296535e-01 9.33475077e-01 -2.22216964e-01 9.42579389e-01
3.92783016e-01 8.51124704e-01 -7.79717922e-01 -8.67096126e-01
8.38591814e-01 -3.92394215e-01 -4.74404931e-01 -7.56834522e-02
-1.49987534e-01 -6.98147357e-01 -1.67889237e-01 1.36711192e+00
5.61422050e-01 3.46321791e-01 6.90879464e-01 -1.19764352e+00
-6.15094602e-01 8.65699053e-01 -5.47998607e-01 6.25900686e-01
3.60978335e-01 6.74713671e-01 1.35570931e+00 -1.75754517e-01
6.37512922e-01 1.36580849e+00 4.93313193e-01 6.01380527e-01
-1.39783561e+00 -6.03711665e-01 4.41856503e-01 3.44595462e-02
-7.66948879e-01 -3.91310036e-01 4.96142179e-01 -3.08122128e-01
7.76011050e-01 3.74091305e-02 1.08835447e+00 9.17143226e-01
6.94031239e-01 7.33053446e-01 1.52469635e+00 -3.40107828e-01
6.76667154e-01 -9.92574394e-02 -8.71232748e-02 1.44318253e-01
8.18038225e-01 8.73533070e-01 -6.63692653e-01 -6.73591137e-01
8.12533319e-01 -8.61905143e-02 -2.93193781e-03 -5.17840445e-01
-6.61047220e-01 1.00250208e+00 1.32541925e-01 -3.56159270e-01
-9.79459047e-01 3.79764646e-01 2.57744163e-01 8.85527968e-01
3.07379931e-01 9.04245794e-01 -3.83924425e-01 -4.21743810e-01
-4.10569638e-01 7.64972568e-01 7.56680548e-01 5.63703358e-01
8.78512800e-01 8.66933018e-02 -5.47302842e-01 3.83963197e-01
2.85610795e-01 3.51966858e-01 3.66048038e-01 -1.86489367e+00
2.88912028e-01 -1.26451841e-02 9.22168076e-01 -4.38217938e-01
-3.01004916e-01 -6.88983679e-01 -2.53465593e-01 4.60828006e-01
4.42527264e-01 -8.77272844e-01 -2.17334196e-01 2.08910894e+00
4.06709760e-01 -1.75768808e-01 9.74798501e-02 6.24794722e-01
-1.75581738e-01 3.95027161e-01 1.06654331e-01 -8.84904623e-01
7.36937523e-01 -5.32909513e-01 -6.17127836e-01 -1.74812704e-01
3.23824614e-01 -1.65065840e-01 1.04588413e+00 6.14312828e-01
-1.13307118e+00 1.35291934e-01 -6.62639260e-01 9.28909302e-01
1.89926669e-01 -8.43266368e-01 6.72362268e-01 9.43105638e-01
-1.37999165e+00 7.86544204e-01 -5.20838439e-01 -4.24856991e-01
4.79626924e-01 2.64231741e-01 3.29349220e-01 5.74775934e-01
-1.09895182e+00 9.25404668e-01 1.15790069e-01 -3.15543681e-01
-1.27152979e+00 -4.18618143e-01 -7.97632858e-02 8.62716213e-02
9.98586833e-01 -6.22510076e-01 1.83327007e+00 -1.46414495e+00
-2.06239820e+00 4.81746882e-01 3.86137992e-01 -6.95564985e-01
7.95603454e-01 -1.65881306e-01 4.17217761e-02 -1.19736180e-01
3.46154183e-01 3.80632848e-01 9.23495531e-01 -1.02224565e+00
-7.56310701e-01 -2.45873854e-01 5.18760204e-01 6.89580142e-01
-4.00166899e-01 1.31286100e-01 6.51346087e-01 -5.78573167e-01
-5.48404157e-01 -9.48306382e-01 -7.29197621e-01 -2.46933669e-01
-3.03613693e-01 -3.64207536e-01 -2.83894688e-01 -9.26137641e-02
1.08395195e+00 -1.64054382e+00 -3.06605577e-01 2.74011523e-01
3.37145090e-01 -2.47265980e-01 -2.56525785e-01 4.20183927e-01
4.66762125e-01 3.63448441e-01 5.71528338e-02 1.59463435e-01
4.23740149e-01 5.49299605e-02 -1.31353736e-01 5.92234313e-01
-1.58686638e-01 8.28580976e-01 -1.10585177e+00 -4.57092002e-02
-2.92661995e-01 -5.59758067e-01 -9.97957587e-01 3.79192591e-01
-2.55299836e-01 3.08394492e-01 -5.45209587e-01 4.41662669e-01
2.79192060e-01 -6.82410449e-02 4.96206880e-01 8.49270403e-01
-4.12822455e-01 5.51511288e-01 -1.26518297e+00 8.32585394e-01
-1.18318602e-01 2.03550622e-01 2.22856000e-01 -8.65563512e-01
5.22329032e-01 2.79794335e-01 4.78465170e-01 -5.28092682e-01
3.12483460e-01 1.37626678e-01 4.24595714e-01 -2.57901549e-01
4.72103536e-01 -4.33746099e-01 -7.37371072e-02 1.12977040e+00
-1.16519697e-01 -8.31009075e-02 1.30774453e-01 1.83256909e-01
1.15298212e+00 6.78420886e-02 9.34007287e-01 -7.55098104e-01
-1.49465233e-01 -1.35115951e-01 1.00028217e+00 1.70050728e+00
-8.98870647e-01 -2.70782024e-01 9.17258978e-01 -1.34026781e-01
-9.11395013e-01 -1.14732051e+00 2.18835160e-01 1.46878016e+00
8.13880339e-02 -1.31526127e-01 -4.50782031e-01 -7.21821010e-01
5.31350672e-01 6.42705381e-01 -8.12667251e-01 -3.17660004e-01
-1.38313994e-01 -8.09792519e-01 1.64054319e-01 2.66344726e-01
5.75380266e-01 -1.20100260e+00 -1.10155213e+00 3.06937367e-01
9.62032005e-02 -2.02001244e-01 -6.13429785e-01 1.69235960e-01
-1.08455002e+00 -8.13586771e-01 -4.49457496e-01 2.31027931e-01
3.05070788e-01 6.16090484e-02 1.22806251e+00 6.21259958e-02
6.10729218e-01 6.59151912e-01 -3.94364148e-01 -7.74725318e-01
-4.60863650e-01 -7.73019940e-02 6.62722230e-01 -4.23042364e-02
3.00455689e-01 -3.83791029e-01 -9.82257903e-01 5.34092523e-02
-4.88285661e-01 -7.20254660e-01 2.15292796e-01 7.45213747e-01
2.91709363e-01 -1.65630892e-01 9.47479010e-01 -9.83395875e-01
1.29414237e+00 -8.33541751e-01 -1.03495014e+00 2.40364596e-01
-1.10841060e+00 -4.79706973e-02 6.41703904e-01 -5.73643029e-01
-1.30741763e+00 -4.15264338e-01 3.45665365e-01 1.65608600e-01
1.88046828e-01 3.99156392e-01 1.41162634e-01 8.84388387e-02
9.57660258e-01 2.80147506e-04 -4.45717499e-02 -2.56499887e-01
5.70966959e-01 7.43980408e-01 2.70429343e-01 -9.51200902e-01
6.02854609e-01 2.86942929e-01 -2.66858846e-01 -3.08896512e-01
-6.53123617e-01 1.60919473e-01 -4.15454283e-02 -6.00680828e-01
3.25771868e-01 -7.62328744e-01 -1.12735200e+00 3.42092842e-01
-6.44595265e-01 -7.05342472e-01 -7.54571557e-01 4.36153084e-01
-9.38492656e-01 1.98747814e-01 -4.45336580e-01 -1.41822410e+00
-2.32698973e-02 -7.42194057e-01 9.29998904e-02 7.11332083e-01
-3.11741263e-01 -7.18966722e-01 4.73416150e-01 -1.03906706e-01
6.11903012e-01 -1.73919499e-01 3.57978433e-01 -5.59096575e-01
-4.74836797e-01 1.20683946e-01 4.45619524e-01 -1.89732775e-01
4.88725724e-03 1.99710205e-01 -7.50654042e-01 -7.91301966e-01
1.20121881e-01 -4.00968105e-01 4.97561008e-01 1.01181364e+00
6.57218516e-01 -8.97737682e-01 -6.74783736e-02 2.92797863e-01
1.26297092e+00 6.66357160e-01 8.05217475e-02 8.37146282e-01
-3.00573587e-01 5.84189117e-01 7.62417495e-01 1.03024173e+00
4.12859529e-01 5.09512663e-01 5.91726661e-01 5.16068280e-01
6.38332069e-01 -5.63969016e-01 6.77714884e-01 -2.84464080e-02
-1.84742317e-01 -3.55375409e-01 -5.20986617e-01 5.67919135e-01
-2.02124286e+00 -1.25418580e+00 6.15295529e-01 2.41339803e+00
9.77321565e-01 4.75582421e-01 1.10094893e+00 -2.38352269e-01
7.37124383e-01 1.15673043e-01 -1.22837389e+00 -7.67166495e-01
-1.97449178e-01 -1.02398358e-01 7.58172631e-01 6.44043863e-01
-6.88945830e-01 8.56919646e-01 8.19131756e+00 3.73963863e-01
-8.15363705e-01 2.47477934e-01 8.22679281e-01 -5.73147953e-01
-4.57691491e-01 1.91782579e-01 -5.45786142e-01 4.63048786e-01
1.12791526e+00 -9.42272067e-01 8.46435070e-01 7.60164797e-01
5.43390453e-01 -5.38311362e-01 -8.09254587e-01 3.75580251e-01
-6.37766600e-01 -1.16607881e+00 -4.51071322e-01 2.34895378e-01
1.04660523e+00 2.14917466e-01 2.01128468e-01 3.72184098e-01
1.68132961e+00 -7.01458216e-01 1.20916402e+00 3.99389952e-01
4.35365111e-01 -7.35446870e-01 4.55583453e-01 5.89674354e-01
-5.20864189e-01 -7.81525016e-01 -3.79982889e-01 -8.57279301e-01
-1.81500182e-01 3.29185486e-01 -3.24214399e-01 -1.80109795e-02
9.29205418e-01 4.49782312e-01 -1.70150712e-01 1.08487022e+00
-5.16825080e-01 8.91931415e-01 -1.67616516e-01 -3.85662079e-01
1.42145097e-01 -5.09227097e-01 6.01788342e-01 7.25077152e-01
1.13264143e-01 3.02964419e-01 2.04444200e-01 7.96169579e-01
-1.98123083e-01 1.72742475e-02 -6.21330261e-01 9.20457914e-02
1.08166838e+00 8.82989943e-01 -6.63545251e-01 -3.56256813e-01
-2.20383242e-01 4.02970165e-01 3.46523702e-01 4.80653197e-01
-2.46125937e-01 -8.54220018e-02 1.07868516e+00 -1.32679448e-01
7.73729682e-02 1.17299937e-01 -3.49523693e-01 -1.15130377e+00
-2.30390638e-01 -1.02285004e+00 7.89617062e-01 -4.40943629e-01
-1.52503037e+00 -7.44824708e-02 1.55608952e-01 -1.04143345e+00
-4.22694117e-01 -2.15435937e-01 -8.75794768e-01 4.14182365e-01
-1.26337481e+00 -3.92480940e-02 4.72797841e-01 2.34498337e-01
1.90268725e-01 -1.39489740e-01 1.67664155e-01 -4.16550398e-01
-6.26463473e-01 4.04916614e-01 6.18674874e-01 -3.15167248e-01
5.69233119e-01 -1.53442419e+00 2.46491387e-01 9.06741142e-01
-1.47915795e-01 6.36685252e-01 1.02192128e+00 -6.09200835e-01
-1.12340224e+00 -6.44700766e-01 3.09153110e-01 -6.20906174e-01
7.54296660e-01 3.99136171e-02 -4.69066441e-01 7.96989441e-01
3.56419951e-01 -3.15774798e-01 4.30784643e-01 7.00975358e-02
-5.91188744e-02 -2.63759285e-01 -1.20469034e+00 1.01888955e+00
1.22393119e+00 -3.36938679e-01 -4.15907413e-01 2.41476316e-02
6.97305381e-01 2.75378879e-02 -6.03163183e-01 -1.36593923e-01
5.87969184e-01 -1.33323419e+00 6.22995257e-01 -8.68126452e-01
8.00874382e-02 -3.90899293e-02 -6.42645881e-02 -1.89662850e+00
-6.56124651e-01 -1.42692542e+00 -1.88469607e-02 8.66401255e-01
1.81818977e-01 -1.00432801e+00 8.61457229e-01 6.55125439e-01
4.54120636e-01 -4.76138592e-01 -1.23172915e+00 -1.17913055e+00
6.59301043e-01 -1.40357211e-01 8.98039937e-01 7.16172993e-01
4.01009887e-01 -1.17703341e-01 -2.81565845e-01 -2.04664901e-01
1.08749461e+00 9.07265916e-02 6.75509274e-01 -1.04183388e+00
-6.39739335e-01 -8.79978299e-01 1.94961533e-01 -1.01766670e+00
7.46497288e-02 -4.23187494e-01 8.74949843e-02 -9.75172997e-01
3.73065948e-01 -3.50205690e-01 -6.90940619e-01 2.05198556e-01
-4.36502099e-01 -1.39487490e-01 3.71923923e-01 3.46599877e-01
-8.74070048e-01 5.99460423e-01 8.92009437e-01 5.48812822e-02
-4.95930940e-01 2.30762690e-01 -1.32063043e+00 4.76813376e-01
9.14918005e-01 -7.76346266e-01 -3.19553852e-01 -5.10893716e-03
6.60510838e-01 5.73483944e-01 1.83424726e-01 -2.59685487e-01
2.61682589e-02 -1.19454384e+00 -1.44381091e-01 1.38763487e-01
-3.77049774e-01 -1.84727386e-01 1.51646659e-01 9.77900326e-01
-7.23375261e-01 4.04185027e-01 -3.60905319e-01 7.60582864e-01
3.07939857e-01 -3.82445306e-01 8.83611381e-01 -4.97590512e-01
-1.31402716e-01 1.84228092e-01 -1.03727460e+00 3.85400683e-01
8.93288136e-01 -1.79334059e-01 -4.56426114e-01 -1.15838087e+00
-3.41534525e-01 5.01175702e-01 8.04683983e-01 2.11933665e-02
1.67335659e-01 -1.05075610e+00 -7.11296320e-01 -2.69154429e-01
-1.06980652e-02 -6.28476977e-01 -2.37226725e-01 5.96153498e-01
-2.79018003e-02 9.11803320e-02 -5.89622498e-01 7.39087015e-02
-6.00646496e-01 3.56884062e-01 9.86625075e-01 -3.29094201e-01
-1.50677398e-01 8.08172584e-01 6.07791217e-03 -3.12705636e-01
-4.89794277e-02 5.87608553e-02 -1.23715580e-01 9.10443217e-02
3.14412624e-01 7.03998148e-01 -4.02299702e-01 -4.72539254e-02
-2.56777018e-01 4.18756641e-02 2.45546967e-01 -9.29888904e-01
1.24517846e+00 -5.12791097e-01 5.13683483e-02 4.70955193e-01
9.95001271e-02 -1.30394220e-01 -1.86878490e+00 -4.50855076e-01
4.30220455e-01 -7.44151056e-01 5.47473170e-02 -9.46896851e-01
-8.02612424e-01 2.60310650e-01 5.16481578e-01 6.31612241e-01
1.03096724e+00 -4.05876517e-01 -5.22556305e-02 5.22603810e-01
8.18246305e-01 -1.59500384e+00 6.56352565e-02 2.58807749e-01
4.51868683e-01 -1.11628604e+00 2.05867365e-01 5.62430441e-01
-8.87366533e-01 6.58986151e-01 6.64880157e-01 -2.68488944e-01
8.04499388e-01 -8.05035010e-02 2.51428783e-01 5.63321216e-03
-1.39510167e+00 -3.99851769e-01 -4.96867955e-01 8.02884102e-01
8.93776268e-02 5.57891428e-01 -8.78410518e-01 6.77626967e-01
-2.35796973e-01 -6.86242953e-02 1.27430940e+00 8.84856224e-01
-6.64268255e-01 -1.05791628e+00 -5.25739610e-01 1.04535329e+00
-5.47466218e-01 2.32836202e-01 -5.89443505e-01 3.40033472e-01
-2.33225420e-01 1.37822282e+00 1.97173372e-01 -2.00889349e-01
2.92953670e-01 -1.21076062e-01 3.53101701e-01 -5.61969340e-01
-9.92369950e-01 1.18464261e-01 2.77969893e-02 -6.52711511e-01
-5.75857997e-01 -1.00601554e+00 -7.27326691e-01 -5.93018711e-01
-2.85238653e-01 2.83067375e-01 8.37043021e-03 7.52256334e-01
1.18493073e-01 -1.22025877e-01 1.06144106e+00 -4.21258211e-01
-1.41412437e+00 -7.98257053e-01 -1.01907253e+00 3.26686502e-01
4.14464384e-01 -8.63079369e-01 -4.65333551e-01 -6.11263573e-01] | [4.211745262145996, 2.723013162612915] |
ee8a5ad2-3d07-4b1d-83f4-be83d8a0aada | human-instance-segmentation-and-tracking-via | 2203.16966 | null | https://arxiv.org/abs/2203.16966v1 | https://arxiv.org/pdf/2203.16966v1.pdf | Human Instance Segmentation and Tracking via Data Association and Single-stage Detector | Human video instance segmentation plays an important role in computer understanding of human activities and is widely used in video processing, video surveillance, and human modeling in virtual reality. Most current VIS methods are based on Mask-RCNN framework, where the target appearance and motion information for data matching will increase computational cost and have an impact on segmentation real-time performance; on the other hand, the existing datasets for VIS focus less on all the people appearing in the video. In this paper, to solve the problems, we develop a new method for human video instance segmentation based on single-stage detector. To tracking the instance across the video, we have adopted data association strategy for matching the same instance in the video sequence, where we jointly learn target instance appearances and their affinities in a pair of video frames in an end-to-end fashion. We have also adopted the centroid sampling strategy for enhancing the embedding extraction ability of instance, which is to bias the instance position to the inside of each instance mask with heavy overlap condition. As a result, even there exists a sudden change in the character activity, the instance position will not move out of the mask, so that the problem that the same instance is represented by two different instances can be alleviated. Finally, we collect PVIS dataset by assembling several video instance segmentation datasets to fill the gap of the current lack of datasets dedicated to human video segmentation. Extensive simulations based on such dataset has been conduct. Simulation results verify the effectiveness and efficiency of the proposed work. | ['Mingbo Zhao', 'Lu Cheng'] | 2022-03-31 | null | null | null | null | ['human-instance-segmentation', 'video-instance-segmentation'] | ['computer-vision', 'computer-vision'] | [ 2.58310407e-01 -2.16399938e-01 -1.25840411e-01 -2.09678203e-01
1.89071804e-01 -2.89007664e-01 2.12731138e-01 7.57873207e-02
-5.54810047e-01 4.62604493e-01 -4.56804410e-02 4.29053187e-01
-3.73536460e-02 -6.59618199e-01 -4.50463712e-01 -7.59508014e-01
4.10305001e-02 2.19843969e-01 5.78329146e-01 -3.54423225e-02
1.31937161e-01 3.05764019e-01 -1.34109998e+00 7.98154026e-02
5.02853572e-01 8.09348524e-01 3.55507851e-01 2.90377617e-01
-2.24297777e-01 6.53551042e-01 -7.34599233e-01 -2.81161398e-01
5.16500950e-01 -3.56973946e-01 -4.81180310e-01 4.41356182e-01
3.47279757e-01 -4.52156484e-01 -7.40547180e-01 1.12438452e+00
3.78914446e-01 4.25182790e-01 4.76380408e-01 -1.37282550e+00
-1.46706000e-01 4.25331861e-01 -1.07086277e+00 6.43650115e-01
1.33468643e-01 1.03363179e-01 5.14723539e-01 -5.01089752e-01
6.71671093e-01 1.14519656e+00 4.71097946e-01 6.13578498e-01
-7.23304331e-01 -6.44158959e-01 5.27678967e-01 4.83209074e-01
-1.51548362e+00 -1.26583710e-01 1.02651441e+00 -4.80710864e-01
2.43319348e-01 2.82058150e-01 1.08512855e+00 7.35890687e-01
-1.20075770e-01 1.04664421e+00 6.51374340e-01 -1.48531288e-01
-5.81589863e-02 3.07420552e-01 3.48831236e-01 6.69721067e-01
2.02442318e-01 -1.56213924e-01 -1.34157479e-01 1.55668885e-01
6.88756764e-01 4.54647154e-01 -5.45108020e-01 -3.36834908e-01
-1.15023577e+00 4.81764287e-01 3.48092169e-01 5.05310535e-01
-3.37040901e-01 -4.71947938e-02 5.02632439e-01 -1.21543050e-01
1.03878044e-01 -1.58786893e-01 -2.55654663e-01 1.90003961e-02
-1.07252944e+00 2.57588059e-01 4.43062514e-01 8.91702175e-01
5.78353226e-01 -1.27461299e-01 -3.58928055e-01 7.13731766e-01
4.63813283e-02 3.25390309e-01 4.22614574e-01 -5.25331676e-01
5.49596250e-01 8.86419237e-01 1.04937553e-01 -1.59085774e+00
-2.36844808e-01 -4.06630576e-01 -7.96492517e-01 -1.26154512e-01
4.23037946e-01 -2.82153994e-01 -8.94036531e-01 1.66849339e+00
6.41306877e-01 8.23744595e-01 -9.61871222e-02 1.19724560e+00
7.14183688e-01 9.01381075e-01 2.15036169e-01 -4.23279107e-01
1.50462604e+00 -9.19284523e-01 -8.19510162e-01 -1.53168246e-01
2.73036808e-01 -6.63989902e-01 4.55886036e-01 2.43262306e-01
-8.22294593e-01 -9.09277201e-01 -1.06415200e+00 3.83466035e-01
-2.41133392e-01 2.86809474e-01 3.00537199e-01 5.08767068e-01
-4.81057942e-01 1.76239163e-01 -5.84480226e-01 -4.42701876e-01
4.15715426e-01 1.90145388e-01 -2.70653546e-01 -7.04277605e-02
-1.05919302e+00 3.33319336e-01 7.27063417e-01 5.68405330e-01
-6.94601417e-01 -4.16867286e-01 -4.91339862e-01 -1.72461212e-01
7.23691940e-01 -3.95639747e-01 8.00788879e-01 -1.43534720e+00
-8.97869825e-01 5.11073172e-01 -1.26231432e-01 -4.25065845e-01
7.65868485e-01 -1.22046933e-01 -5.14220178e-01 1.08590379e-01
4.25710343e-02 6.20539606e-01 8.54891002e-01 -1.17972171e+00
-9.89014983e-01 -5.43497205e-01 2.30979659e-02 4.96795058e-01
-4.57743734e-01 2.21852604e-02 -1.14403200e+00 -5.36492705e-01
2.03687362e-02 -9.26536620e-01 -1.85646400e-01 -1.49037912e-01
-2.10638940e-01 -1.34153262e-01 1.19026756e+00 -9.61339355e-01
1.51206255e+00 -2.31143069e+00 2.90226817e-01 2.96610534e-01
1.47050366e-01 5.62700212e-01 1.30850077e-01 9.26618427e-02
1.39146119e-01 -1.41330317e-01 -4.14264947e-01 1.24494508e-02
-4.92383242e-01 9.62651521e-02 8.23434815e-02 5.04698813e-01
3.52109037e-02 4.59371746e-01 -6.95301414e-01 -8.47824633e-01
3.56216192e-01 3.57130349e-01 -3.47797543e-01 2.86929041e-01
2.02885084e-02 3.32030326e-01 -5.19029856e-01 5.08109212e-01
8.55047464e-01 1.64411470e-01 1.03219278e-01 -5.90260386e-01
-3.11068222e-02 -7.30853736e-01 -1.69058263e+00 1.47712743e+00
4.71119173e-02 5.70768356e-01 1.39127644e-02 -1.16748857e+00
6.84532464e-01 2.63970762e-01 6.74307406e-01 -5.64491689e-01
2.30977327e-01 -2.49812111e-01 1.15450114e-01 -9.41170394e-01
4.44087654e-01 4.09866422e-01 2.33299300e-01 -3.65886316e-02
-3.08066696e-01 5.03132224e-01 5.36749959e-01 1.02616884e-01
6.82387233e-01 1.56425402e-01 1.07120894e-01 6.22956939e-02
8.41962159e-01 5.35811856e-02 8.01925719e-01 5.49629569e-01
-4.53817546e-01 4.80635852e-01 2.24599406e-01 -4.16391104e-01
-9.04078305e-01 -6.89132512e-01 9.85236838e-03 6.51368797e-01
7.40985394e-01 -2.84425020e-02 -1.00017214e+00 -6.59607053e-01
-2.56424159e-01 2.46562943e-01 -5.22230804e-01 4.68633175e-02
-9.83217418e-01 -8.77075016e-01 4.45018113e-01 3.65022212e-01
9.22873735e-01 -1.15460753e+00 -7.68113196e-01 3.21840882e-01
-2.25372523e-01 -1.28907979e+00 -6.10623419e-01 -3.63961995e-01
-7.03268707e-01 -1.37889862e+00 -8.56392562e-01 -9.73123372e-01
7.89951205e-01 4.88993824e-01 5.30366778e-01 2.66320020e-01
-5.47747135e-01 2.90772080e-01 -3.37676197e-01 -2.79397815e-02
-8.47528651e-02 -1.30270272e-01 -1.05707981e-01 3.96600842e-01
4.12367225e-01 -3.09387408e-02 -8.25409412e-01 4.36699480e-01
-1.01386571e+00 4.39281277e-02 3.03949952e-01 5.23755074e-01
2.68042237e-01 4.63236600e-01 3.91105235e-01 -6.88865662e-01
4.59087104e-01 -4.65484768e-01 -4.62209672e-01 3.16467106e-01
-7.56971315e-02 -5.31066716e-01 5.14105082e-01 -6.09042048e-01
-1.01412129e+00 1.48420081e-01 2.60864884e-01 -6.19575381e-01
-2.43841946e-01 2.33872116e-01 -3.87836009e-01 2.21350491e-01
4.51055169e-02 4.30761039e-01 4.85586450e-02 -1.14584595e-01
1.60542026e-01 7.47439265e-01 4.16240424e-01 -1.28170133e-01
6.37271821e-01 4.12806541e-01 -1.33318961e-01 -1.05125117e+00
-1.78801224e-01 -6.85587883e-01 -5.15983045e-01 -5.85363388e-01
1.27171016e+00 -9.06326234e-01 -5.85052311e-01 5.29849350e-01
-1.12614262e+00 2.51198441e-01 2.41702244e-01 4.90382016e-01
-4.46263999e-02 7.13338315e-01 -4.43232179e-01 -1.04867649e+00
-1.27411604e-01 -1.19647932e+00 7.91416645e-01 6.29174352e-01
-3.58308491e-04 -8.21866274e-01 -2.13876605e-01 3.89303207e-01
-1.01898685e-01 2.59828240e-01 6.45159125e-01 -5.71605444e-01
-7.68376291e-01 -3.59209597e-01 -2.50438184e-01 4.06065643e-01
1.69186160e-01 2.50395060e-01 -5.98188221e-01 -2.19902366e-01
3.28290761e-02 3.34158659e-01 6.99113071e-01 3.90552849e-01
1.04980505e+00 -2.07066298e-01 -6.92109585e-01 3.18474054e-01
1.34804928e+00 6.32945716e-01 6.86432481e-01 1.78609684e-01
1.02536786e+00 7.59401083e-01 1.01034403e+00 2.47728109e-01
-3.13874222e-02 8.66719782e-01 2.69354463e-01 -1.38419598e-01
-2.25365888e-02 -8.13637376e-02 2.22322538e-01 6.34746909e-01
-2.10115299e-01 -4.06539798e-01 -6.18574381e-01 6.65997922e-01
-2.00137973e+00 -1.20357049e+00 -1.45924538e-01 2.29312062e+00
3.71450007e-01 1.13340110e-01 3.44123989e-01 1.84989125e-01
1.36872089e+00 2.50291169e-01 -3.42607856e-01 1.84509486e-01
2.63188547e-03 -3.31854731e-01 4.05922353e-01 1.96450129e-01
-1.17628765e+00 7.30801105e-01 4.66666842e+00 9.93566930e-01
-1.07414722e+00 -1.21748643e-02 6.92828774e-01 -1.99611232e-01
2.82570213e-01 -1.51078701e-01 -9.05239880e-01 8.75616908e-01
1.54860064e-01 2.33215079e-01 2.92634457e-01 6.27316236e-01
4.96269137e-01 -2.66442716e-01 -8.45642686e-01 1.18501818e+00
1.88081324e-01 -1.02946758e+00 8.69560018e-02 -1.18180655e-01
4.67816085e-01 -5.20670772e-01 -1.63522169e-01 9.52248871e-02
-5.27831078e-01 -6.47495568e-01 5.86933672e-01 5.26780248e-01
2.70522416e-01 -8.54380071e-01 7.91091383e-01 4.60514486e-01
-1.42300189e+00 -2.63429433e-01 -4.08959180e-01 1.53591439e-01
3.48369122e-01 2.56380826e-01 -6.75486267e-01 5.33786178e-01
6.42164052e-01 6.84944808e-01 -4.31557745e-01 1.25626159e+00
2.46294081e-01 3.77958596e-01 -2.68274277e-01 -4.88239937e-02
2.91030586e-01 -5.28018534e-01 6.96664631e-01 1.19124329e+00
1.72262505e-01 2.86199361e-01 5.68401158e-01 6.15669370e-01
1.06633544e-01 3.32166523e-01 -6.83200955e-01 1.98123485e-01
5.01875997e-01 1.28928566e+00 -8.97433221e-01 -4.37578470e-01
-4.71568406e-01 1.12997627e+00 -9.33690295e-02 4.39965189e-01
-1.23736250e+00 -3.27415138e-01 2.82751411e-01 2.45293543e-01
4.49131340e-01 1.43340463e-02 1.95000172e-01 -8.63942683e-01
2.61176676e-01 -8.28926861e-01 4.22408491e-01 -4.86100107e-01
-9.27161515e-01 4.96829003e-01 2.67203808e-01 -1.30185473e+00
1.25853568e-01 -3.38599741e-01 -6.32749856e-01 5.16307056e-01
-9.11862135e-01 -9.29199517e-01 -5.99464059e-01 7.37527788e-01
8.86614025e-01 -2.02200264e-01 6.18513934e-02 7.50302970e-01
-1.10983074e+00 4.78432626e-01 -2.35140502e-01 5.23311615e-01
2.97030240e-01 -5.81039906e-01 -9.99781936e-02 1.12239778e+00
2.06684873e-01 5.95169425e-01 6.52832747e-01 -9.87210631e-01
-1.04318976e+00 -1.05462027e+00 2.62231231e-01 -1.30153835e-01
2.44967893e-01 -1.25104845e-01 -7.97805250e-01 4.21128243e-01
2.82323539e-01 -3.84298414e-02 3.12320709e-01 -3.84781748e-01
2.12571025e-01 -2.59178579e-01 -1.23930013e+00 7.90191770e-01
1.09656382e+00 -7.93256089e-02 -5.32044709e-01 1.87456235e-01
5.08898616e-01 -2.01490402e-01 -5.59572816e-01 4.84957755e-01
4.22035038e-01 -9.23108220e-01 9.15744066e-01 -4.96085793e-01
1.99035957e-01 -8.17793489e-01 -8.96871928e-03 -8.20377648e-01
-1.82271510e-01 -2.82747716e-01 -1.09962210e-01 1.44282854e+00
-6.47298694e-02 -1.83796987e-01 8.21033001e-01 6.61360323e-01
2.16066495e-01 -7.25887775e-01 -8.32687438e-01 -4.34323668e-01
-6.96395338e-01 -1.87782887e-02 4.99748796e-01 8.42632771e-01
-3.27307582e-01 1.32457688e-01 -4.77500528e-01 3.67821574e-01
6.48653626e-01 -1.08153693e-01 8.82016838e-01 -9.82166886e-01
-3.27664882e-01 -1.89451665e-01 -7.56404161e-01 -1.13775718e+00
-1.09117165e-01 -4.72741812e-01 -8.69035795e-02 -1.27202511e+00
5.14651000e-01 -3.94060731e-01 -3.58763844e-01 -4.44857124e-03
-4.71407801e-01 2.41707221e-01 3.46556753e-01 2.34040648e-01
-5.52580893e-01 2.74003476e-01 1.10438240e+00 -3.13922226e-01
-4.05229300e-01 8.45435541e-03 -9.16628018e-02 7.09267974e-01
7.41481006e-01 -3.47747087e-01 -6.55471444e-01 -2.25136086e-01
-3.66050482e-01 3.15471530e-01 4.98361409e-01 -1.22471166e+00
2.79719085e-01 -1.70577049e-01 5.55439711e-01 -5.68856299e-01
3.10823292e-01 -1.24481225e+00 3.75527918e-01 5.77406168e-01
9.55404434e-03 1.41463622e-01 1.41902730e-01 8.73443782e-01
-2.33353168e-01 -4.72248882e-01 6.54995561e-01 -3.01467806e-01
-1.15499675e+00 4.89480108e-01 -2.64812142e-01 5.30478992e-02
1.54949248e+00 -7.69290566e-01 -1.86073575e-02 -8.10547620e-02
-4.74596441e-01 2.81074017e-01 2.44937956e-01 5.87756991e-01
5.42398334e-01 -1.02185404e+00 -5.25827110e-01 1.22713789e-01
6.94119707e-02 -1.12993017e-01 6.98654294e-01 8.31272662e-01
-6.05374157e-01 3.42184640e-02 -2.29256302e-01 -6.74027383e-01
-1.56566417e+00 6.06908023e-01 3.70326132e-01 1.43202305e-01
-6.11779869e-01 5.89610696e-01 3.15663725e-01 1.32992178e-01
4.19520050e-01 -4.25333194e-02 -6.30422652e-01 3.19979608e-01
5.19928873e-01 4.90361243e-01 -4.47703749e-01 -9.51926947e-01
-3.49261254e-01 7.24312067e-01 -3.12932640e-01 2.87598390e-02
8.79430175e-01 -2.36070842e-01 7.55284028e-03 4.17505503e-01
1.08599818e+00 1.04267344e-01 -1.30746353e+00 6.23043627e-03
-2.58442193e-01 -7.30832279e-01 -9.89772081e-02 -3.72634143e-01
-1.40209043e+00 5.63665390e-01 9.76106763e-01 1.06609568e-01
9.88151610e-01 -4.31766927e-01 1.02881670e+00 1.15838408e-01
3.54940414e-01 -1.24535120e+00 -2.23271828e-03 -7.07470849e-02
2.36494303e-01 -9.87044811e-01 2.38186289e-02 -5.46535850e-01
-8.98008525e-01 8.75315428e-01 9.99714255e-01 -1.45299822e-01
4.19265330e-01 -4.75987867e-02 5.22839874e-02 -5.80272861e-02
-8.45106170e-02 -9.13762152e-02 9.90231037e-02 4.92522627e-01
1.09567299e-01 -8.85754749e-02 -5.44668138e-01 3.51995021e-01
3.54905784e-01 6.58273995e-02 4.77540374e-01 8.26772511e-01
-5.08972049e-01 -8.51825297e-01 -4.33445841e-01 4.32179481e-01
-5.06207287e-01 2.18013853e-01 -7.94755369e-02 9.44700658e-01
5.24829924e-01 8.02875161e-01 1.45699218e-01 -2.85542518e-01
2.32780278e-01 -2.88688485e-02 4.82674569e-01 -1.14049248e-01
-5.83639085e-01 6.88600168e-02 -9.51649472e-02 -2.06941366e-01
-6.95881128e-01 -7.24066436e-01 -1.33304048e+00 -1.56198859e-01
-3.06001127e-01 9.32585001e-02 2.23563313e-01 7.71287858e-01
8.34725350e-02 5.22869349e-01 4.37278569e-01 -5.58722496e-01
-1.91627011e-01 -7.93342590e-01 -4.85298187e-01 6.94734573e-01
6.77622408e-02 -6.51347637e-01 -1.12835646e-01 9.00965929e-02] | [9.011638641357422, -0.23204058408737183] |
717e9604-a86d-4ac0-88d9-f59cdb80f7e0 | language-features-matter-effective-language | 1908.06327 | null | https://arxiv.org/abs/1908.06327v1 | https://arxiv.org/pdf/1908.06327v1.pdf | Language Features Matter: Effective Language Representations for Vision-Language Tasks | Shouldn't language and vision features be treated equally in vision-language (VL) tasks? Many VL approaches treat the language component as an afterthought, using simple language models that are either built upon fixed word embeddings trained on text-only data or are learned from scratch. We believe that language features deserve more attention, and conduct experiments which compare different word embeddings, language models, and embedding augmentation steps on five common VL tasks: image-sentence retrieval, image captioning, visual question answering, phrase grounding, and text-to-clip retrieval. Our experiments provide some striking results; an average embedding language model outperforms an LSTM on retrieval-style tasks; state-of-the-art representations such as BERT perform relatively poorly on vision-language tasks. From this comprehensive set of experiments we propose a set of best practices for incorporating the language component of VL tasks. To further elevate language features, we also show that knowledge in vision-language problems can be transferred across tasks to gain performance with multi-task training. This multi-task training is applied to a new Graph Oriented Vision-Language Embedding (GrOVLE), which we adapt from Word2Vec using WordNet and an original visual-language graph built from Visual Genome, providing a ready-to-use vision-language embedding: http://ai.bu.edu/grovle. | ['Andrea Burns', 'Reuben Tan', 'Kate Saenko', 'Stan Sclaroff', 'Bryan A. Plummer'] | 2019-08-17 | language-features-matter-effective-language-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Burns_Language_Features_Matter_Effective_Language_Representations_for_Vision-Language_Tasks_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Burns_Language_Features_Matter_Effective_Language_Representations_for_Vision-Language_Tasks_ICCV_2019_paper.pdf | iccv-2019-10 | ['phrase-grounding'] | ['natural-language-processing'] | [ 1.24053657e-01 -6.01007976e-02 -2.37689644e-01 -3.48110080e-01
-7.65326738e-01 -3.63764673e-01 8.08565915e-01 9.46231335e-02
-7.25554883e-01 2.81910837e-01 5.35067856e-01 -5.45671046e-01
2.80832559e-01 -5.75143933e-01 -7.16334820e-01 -2.97703683e-01
1.84326544e-01 2.04088449e-01 1.95132852e-01 -2.64976054e-01
1.47550508e-01 3.41318458e-01 -1.22956908e+00 3.84129405e-01
2.05886602e-01 6.76933706e-01 4.56166536e-01 1.10094583e+00
-5.65988421e-01 1.07321465e+00 -3.50098580e-01 -4.24797088e-01
6.86649308e-02 -1.93347827e-01 -7.82544672e-01 2.78145552e-01
9.71745729e-01 -4.07718331e-01 -6.95654631e-01 7.20148146e-01
5.43544590e-01 2.58360687e-03 7.70481408e-01 -1.36332023e+00
-1.79657543e+00 9.91934612e-02 -4.60307896e-01 1.80036828e-01
3.97105157e-01 5.30013204e-01 1.31596947e+00 -1.19806206e+00
7.12047040e-01 1.60690486e+00 5.31529725e-01 8.15684676e-01
-1.28205764e+00 -6.61182776e-02 3.51117790e-01 2.69864261e-01
-1.21001256e+00 -4.06775445e-01 5.54444134e-01 -7.31344283e-01
1.49044991e+00 -8.66181254e-02 6.09038651e-01 1.23861241e+00
3.50072920e-01 8.81147325e-01 8.91139805e-01 -6.34453654e-01
-2.73591995e-01 2.92580307e-01 3.64589870e-01 1.20032132e+00
3.40938330e-01 -4.79653738e-02 -5.16127110e-01 -9.42532867e-02
7.53703117e-01 -4.34674621e-02 -4.46428329e-01 -4.39498156e-01
-1.36029458e+00 1.25836051e+00 5.80959201e-01 2.50276595e-01
-1.79939538e-01 7.39343107e-01 6.07369125e-01 5.69224060e-01
5.34775674e-01 4.01748389e-01 -3.12517792e-01 4.16261107e-01
-5.81067145e-01 8.24021176e-02 6.84434831e-01 8.88050616e-01
9.64897633e-01 1.68473646e-01 -5.82514226e-01 8.84178221e-01
6.47257388e-01 7.19983876e-01 5.00066638e-01 -7.10900903e-01
1.02268398e-01 2.11447105e-01 -2.90449262e-01 -8.22546899e-01
-1.21816635e-01 1.57785844e-02 -5.80141008e-01 1.70841321e-01
8.79479796e-02 4.77274954e-02 -1.43606889e+00 1.71070313e+00
-3.21925223e-01 -3.27560678e-02 3.01418543e-01 7.16425300e-01
1.44118500e+00 7.86722898e-01 3.12550843e-01 1.94682255e-01
1.58460414e+00 -1.42356348e+00 -6.35900199e-01 -6.95424914e-01
7.06963718e-01 -7.82281220e-01 1.32158875e+00 1.46116232e-02
-8.98878574e-01 -8.38391483e-01 -8.34588110e-01 -7.23833025e-01
-7.04747200e-01 1.36054605e-01 6.50077105e-01 1.75267667e-01
-1.83362830e+00 -1.36180952e-01 -4.87832963e-01 -8.52997899e-01
3.26839089e-01 6.16366742e-03 -5.31792462e-01 -4.74603921e-01
-9.40971553e-01 1.30496466e+00 1.78853303e-01 -1.00852855e-01
-1.02474225e+00 -3.82683605e-01 -1.38638258e+00 -2.22172663e-01
3.11611235e-01 -1.40190399e+00 1.05688798e+00 -1.01533222e+00
-1.04406869e+00 1.32139409e+00 -3.52168769e-01 -5.08333385e-01
2.40628086e-02 -1.21517844e-01 -2.09939048e-01 4.14355129e-01
1.70689464e-01 1.24526775e+00 1.21291792e+00 -1.36576593e+00
-6.70460239e-02 -7.05444142e-02 2.13620394e-01 1.14163078e-01
-3.52966756e-01 3.91225629e-02 -9.16493535e-01 -5.69781601e-01
-4.86444503e-01 -7.99394667e-01 -2.02479482e-01 6.31361842e-01
-1.62095651e-01 -5.75065017e-01 5.91055095e-01 -6.26106977e-01
8.28102708e-01 -2.07914352e+00 1.55165404e-01 -3.07943255e-01
5.64408958e-01 2.82930672e-01 -1.03112686e+00 6.33430183e-01
-2.52084564e-02 2.74296105e-01 3.03442590e-02 -5.99432647e-01
-2.81710345e-02 5.74111879e-01 -3.47882807e-01 3.23244691e-01
5.84728837e-01 1.56401539e+00 -8.24251413e-01 -6.19960666e-01
4.21827883e-01 6.98181629e-01 -4.93070006e-01 2.70479649e-01
-5.11849999e-01 -2.22115591e-01 -3.75808120e-01 6.24768496e-01
1.60070106e-01 -6.05583370e-01 -3.06071937e-01 -3.16034883e-01
9.68124494e-02 -1.56026527e-01 -3.63435328e-01 1.93921351e+00
-6.10025048e-01 1.15324104e+00 -1.27292573e-01 -1.04675400e+00
7.25177526e-01 1.79963186e-01 1.41719133e-01 -6.55731797e-01
-3.76448110e-02 -1.42045125e-01 -8.48669410e-02 -8.93850744e-01
5.86653352e-01 4.22911420e-02 1.60734490e-01 3.07280630e-01
5.77545702e-01 -4.63664591e-01 1.58898830e-01 5.99588692e-01
1.00546205e+00 -5.15448079e-02 2.70601690e-01 -3.69062051e-02
4.64424521e-01 1.52101830e-01 -2.69766420e-01 9.60602462e-01
-2.12531447e-01 6.91095471e-01 2.33543128e-01 -3.64700466e-01
-1.07014525e+00 -1.14267516e+00 1.64116740e-01 1.35114050e+00
-2.23808721e-01 -6.42140627e-01 -1.99122027e-01 -5.52280605e-01
2.18174964e-01 7.59092569e-01 -6.80891335e-01 -1.96220025e-01
-1.06519811e-01 -3.10502797e-01 4.58629161e-01 4.74264860e-01
4.87610400e-02 -1.34756649e+00 -3.67798239e-01 -9.56152659e-03
2.43027866e-01 -1.51244032e+00 -6.32177770e-01 2.21313257e-03
-4.65377152e-01 -7.97670484e-01 -1.13567841e+00 -1.23407805e+00
6.26959383e-01 7.11120725e-01 1.47401428e+00 3.88293087e-01
-5.74217856e-01 1.51357079e+00 -3.70500326e-01 -5.97968459e-01
-2.81308711e-01 -3.11910421e-01 -1.37169451e-01 -2.63456255e-01
5.84419727e-01 -3.55628058e-02 -4.90262836e-01 -3.87684077e-01
-1.06887388e+00 5.20076305e-02 7.98934042e-01 1.06684518e+00
5.68771660e-01 -8.27852964e-01 3.87203366e-01 -3.72239590e-01
1.04834545e+00 -2.68523604e-01 -3.05255264e-01 6.60273671e-01
-4.78354484e-01 2.34395832e-01 1.99052930e-01 -6.41136467e-01
-2.87622422e-01 -1.41579434e-01 -2.82222062e-01 -8.77319098e-01
-1.13662556e-02 6.73633516e-01 3.75582457e-01 -3.91745180e-01
5.27036428e-01 6.39371991e-01 3.93827766e-01 -1.92582190e-01
1.11492491e+00 5.04688144e-01 3.45813245e-01 -3.41378152e-01
8.27332437e-01 3.12774926e-01 -2.24440306e-01 -1.40745926e+00
-8.31263959e-01 -6.85089588e-01 -3.36178392e-01 -2.28029508e-02
1.37968695e+00 -1.24845111e+00 -4.51757640e-01 1.02506801e-01
-1.57302845e+00 -4.32943463e-01 -4.81286883e-01 3.25575173e-01
-4.80380237e-01 6.74923718e-01 -4.78817672e-01 -5.81978619e-01
-5.06085336e-01 -1.18441725e+00 1.38168204e+00 -7.93424398e-02
2.24104151e-02 -1.44301283e+00 1.22700736e-01 3.48479033e-01
5.59768140e-01 -1.23004600e-01 1.01798952e+00 -5.70405662e-01
-5.96611261e-01 6.48741573e-02 -6.50970340e-01 7.14899898e-01
-2.16121376e-02 -2.56780088e-02 -9.49463546e-01 -5.26915252e-01
-2.76176512e-01 -9.70515251e-01 1.48277593e+00 5.06974995e-01
1.12741113e+00 -3.02351657e-02 -1.73073143e-01 6.01146698e-01
1.75237978e+00 -3.84070814e-01 4.96456921e-01 1.91396266e-01
9.94281530e-01 4.47023928e-01 -3.34605351e-02 -1.08345792e-01
7.51347721e-01 3.90138626e-01 2.92343825e-01 -5.14212012e-01
-6.90753341e-01 -2.82938898e-01 6.77221239e-01 1.04264534e+00
3.27975720e-01 -4.43584234e-01 -9.82445657e-01 8.97704303e-01
-1.67285168e+00 -7.47892141e-01 4.74984869e-02 1.73636746e+00
6.08056009e-01 -5.85922552e-03 -1.57676324e-01 -6.46005750e-01
1.69759229e-01 5.65848708e-01 -4.87396687e-01 -7.30688691e-01
-1.57903790e-01 2.72609144e-01 5.45240700e-01 8.36320221e-01
-9.13287878e-01 1.36414874e+00 6.65360355e+00 6.08779609e-01
-1.27866149e+00 2.86133796e-01 3.26844245e-01 1.27416492e-01
-6.70915127e-01 -6.32298663e-02 -6.33259475e-01 -1.67271882e-01
7.36313939e-01 -2.99423665e-01 2.42952719e-01 7.38738716e-01
2.08155643e-02 -3.21511738e-02 -1.35747886e+00 1.34256649e+00
8.40008259e-01 -1.49761355e+00 7.37208128e-01 3.89557029e-03
3.82994771e-01 6.40613198e-01 1.75762340e-01 5.16946554e-01
3.38634223e-01 -1.36130524e+00 4.20575321e-01 6.30875111e-01
1.06758583e+00 5.76807521e-02 4.24651891e-01 -3.53977866e-02
-1.14055169e+00 2.67213285e-01 -6.32721841e-01 6.89431205e-02
1.64921299e-01 2.14606643e-01 -7.78392196e-01 4.39755172e-01
3.93682092e-01 9.37485278e-01 -1.01443577e+00 6.62594438e-01
-1.86511889e-01 3.65054071e-01 1.11638345e-02 -2.46679127e-01
6.07456505e-01 6.36981279e-02 3.54590535e-01 1.42227495e+00
2.95154694e-02 -2.90409178e-01 5.03182173e-01 9.26960170e-01
-6.88168928e-02 1.53251216e-01 -1.36446190e+00 -4.56076771e-01
-1.56775385e-01 1.17098975e+00 -2.66746134e-01 -4.31970686e-01
-1.17074668e+00 1.15653932e+00 4.42255884e-01 7.83239245e-01
-5.41510642e-01 -2.26948187e-01 7.41885960e-01 -2.37395883e-01
3.99838597e-01 -5.89940429e-01 2.06913799e-01 -1.43793416e+00
-2.04275712e-01 -6.19483232e-01 9.03168246e-02 -1.29431808e+00
-1.90090573e+00 6.87334776e-01 -6.22080676e-02 -8.29552531e-01
-2.74201453e-01 -1.09525275e+00 -5.01221657e-01 9.47475910e-01
-2.06445432e+00 -1.62784684e+00 -2.77904093e-01 8.91722143e-01
9.04292345e-01 -3.80306721e-01 9.17213082e-01 -6.03093430e-02
-1.59947693e-01 5.33540785e-01 -2.51343578e-01 3.46824378e-01
8.30116987e-01 -1.11541724e+00 5.50194681e-01 6.68560326e-01
7.65318513e-01 6.11816645e-01 3.98753285e-01 -3.35436225e-01
-1.78678751e+00 -1.24484038e+00 1.03575242e+00 -7.65479028e-01
1.06252956e+00 -4.84481096e-01 -8.02784562e-01 9.77581680e-01
8.05412889e-01 2.05664262e-01 6.47850931e-01 -5.79589866e-02
-5.33333659e-01 2.52258569e-01 -6.08354807e-01 8.13481569e-01
8.48195553e-01 -1.15457737e+00 -1.00565326e+00 7.78077722e-01
1.22560227e+00 1.41625017e-01 -5.17784715e-01 5.32165635e-03
3.95111084e-01 -4.24477935e-01 1.37992215e+00 -7.91312397e-01
4.65057969e-01 -1.08270742e-01 -3.05764407e-01 -1.22188270e+00
-4.34144139e-01 -2.35368505e-01 1.51589930e-01 8.56751442e-01
3.18308920e-01 -6.97242081e-01 2.02713072e-01 1.51359051e-01
-3.14808562e-02 -7.29320347e-01 -6.09390557e-01 -7.52688646e-01
2.89026082e-01 -5.84693432e-01 -2.12909728e-01 7.95890331e-01
-4.97232854e-01 9.44545448e-01 -3.33070159e-01 -9.34008285e-02
5.96731722e-01 -1.04088634e-01 8.36734951e-01 -9.36834872e-01
-3.11838508e-01 -3.45746875e-01 -5.35110772e-01 -1.22405541e+00
5.25616169e-01 -1.21857190e+00 -2.30338518e-02 -2.29910946e+00
3.05969685e-01 2.64482975e-01 -2.59420604e-01 7.12791920e-01
-1.44455567e-01 3.53815109e-01 5.09825170e-01 8.87660980e-02
-7.27957368e-01 6.49943948e-01 1.37522662e+00 -5.84763110e-01
1.86649397e-01 -6.47191882e-01 -7.65291750e-01 6.49005830e-01
5.44411719e-01 -1.16375417e-01 -6.61902130e-01 -1.01598859e+00
1.05789872e-02 -1.33575201e-01 7.06975281e-01 -6.66863561e-01
2.29984015e-01 -1.84531156e-02 2.08991170e-01 -3.13016862e-01
5.80283821e-01 -5.83506942e-01 -5.60158849e-01 4.03260201e-01
-5.37179470e-01 2.65525967e-01 4.08634543e-01 6.35976970e-01
-2.61370003e-01 -1.26971111e-01 4.50991601e-01 -4.02778447e-01
-1.20201945e+00 5.26949525e-01 -5.35320520e-01 1.21824294e-01
8.07050645e-01 -2.50450850e-01 -4.98455822e-01 -5.41023076e-01
-8.63758385e-01 4.61498439e-01 4.90871668e-01 7.31253266e-01
1.15133452e+00 -1.24168384e+00 -9.35029507e-01 1.50776938e-01
7.28979886e-01 -3.79994243e-01 -1.20948114e-01 5.77454150e-01
-5.37026942e-01 6.73863053e-01 -4.91732545e-02 -7.61942029e-01
-1.28930128e+00 8.41004491e-01 1.67392954e-01 -1.14038035e-01
-6.86330676e-01 1.01825011e+00 4.85478193e-01 -1.76683113e-01
3.56512010e-01 -5.55548072e-01 -2.41874501e-01 2.26779468e-02
4.41152126e-01 -4.62426752e-01 -3.79815161e-01 -6.01293087e-01
-3.69638681e-01 9.49685633e-01 -1.20167159e-01 -1.78032786e-01
1.19300294e+00 -1.81103081e-01 -1.50734946e-01 6.48505628e-01
1.40621781e+00 -2.73211122e-01 -7.03815639e-01 -4.45019841e-01
-9.44051966e-02 -2.20301613e-01 2.88529426e-01 -6.18647933e-01
-9.54828143e-01 1.30119550e+00 6.07269585e-01 8.14743415e-02
9.35526073e-01 3.74492854e-01 6.40467405e-01 7.26056039e-01
1.20383836e-01 -6.73323810e-01 5.75182199e-01 6.89985693e-01
1.24736404e+00 -1.53378594e+00 7.39763752e-02 9.22379829e-03
-8.44629407e-01 1.11612284e+00 5.92333853e-01 -3.19420308e-01
7.58750975e-01 -7.05728903e-02 2.37288341e-01 -4.45400089e-01
-1.06561542e+00 -8.35088909e-01 6.36387467e-01 8.09805334e-01
5.01019061e-01 -2.44575739e-01 -7.39994422e-02 -4.68212143e-02
2.91591525e-01 1.17119797e-01 5.46186566e-01 8.02843332e-01
-4.35907632e-01 -1.00623703e+00 -7.24670291e-02 5.27717650e-01
-1.49365425e-01 -6.19075000e-01 -5.65436602e-01 7.97167242e-01
-1.08578436e-01 8.00676644e-01 1.17909968e-01 -2.32521191e-01
1.33532450e-01 2.00290129e-01 6.61027670e-01 -1.11947632e+00
-3.31499577e-01 -1.37720481e-01 9.86099765e-02 -6.14255965e-01
-5.24933398e-01 -5.44881336e-02 -8.47982645e-01 2.02772375e-02
-2.94878799e-02 -2.61494428e-01 6.31599486e-01 7.83370078e-01
3.22918355e-01 5.91632605e-01 1.47671448e-02 -8.26188564e-01
-4.17121977e-01 -8.19768369e-01 -3.15820515e-01 4.04010147e-01
6.78444922e-01 -3.83077323e-01 -3.65866274e-01 2.24445716e-01] | [10.819847106933594, 1.7985570430755615] |
70acb4f5-d072-40c7-8bf7-af965bf2fb4e | retrieve-program-repeat-complex-knowledge | 2010.15875 | null | https://arxiv.org/abs/2010.15875v1 | https://arxiv.org/pdf/2010.15875v1.pdf | Retrieve, Program, Repeat: Complex Knowledge Base Question Answering via Alternate Meta-learning | A compelling approach to complex question answering is to convert the question to a sequence of actions, which can then be executed on the knowledge base to yield the answer, aka the programmer-interpreter approach. Use similar training questions to the test question, meta-learning enables the programmer to adapt to unseen questions to tackle potential distributional biases quickly. However, this comes at the cost of manually labeling similar questions to learn a retrieval model, which is tedious and expensive. In this paper, we present a novel method that automatically learns a retrieval model alternately with the programmer from weak supervision, i.e., the system's performance with respect to the produced answers. To the best of our knowledge, this is the first attempt to train the retrieval model with the programmer jointly. Our system leads to state-of-the-art performance on a large-scale task for complex question answering over knowledge bases. We have released our code at https://github.com/DevinJake/MARL. | ['Wei Wu', 'Guilin Qi', 'Gholamreza Haffari', 'Yuan-Fang Li', 'Yuncheng Hua'] | 2020-10-29 | null | null | null | null | ['knowledge-base-question-answering'] | ['natural-language-processing'] | [-1.46234185e-01 1.25271976e-01 1.47380844e-01 -5.28450131e-01
-1.37327230e+00 -9.19707596e-01 3.64868462e-01 2.65153021e-01
-4.07256931e-01 5.46450794e-01 -1.61369368e-02 -6.58915222e-01
7.61514083e-02 -6.94826007e-01 -7.88050234e-01 -2.02740729e-01
4.08653438e-01 7.75876224e-01 5.27855575e-01 -3.50265175e-01
4.00843292e-01 8.39133561e-03 -1.59228384e+00 4.76821452e-01
9.90757525e-01 5.41951120e-01 2.27771148e-01 1.03486872e+00
-4.51809943e-01 1.15968776e+00 -7.57941484e-01 -7.65197217e-01
7.14231133e-02 -4.67686832e-01 -1.34793103e+00 -3.06199729e-01
7.93501854e-01 -4.93530929e-01 -1.09295826e-02 1.03271806e+00
3.27296764e-01 6.94272816e-02 4.02678251e-01 -9.60727394e-01
-9.12133336e-01 4.90578145e-01 -1.46776512e-01 3.32446486e-01
6.10795677e-01 2.05048934e-01 1.15331054e+00 -8.18317533e-01
4.36550319e-01 1.08928311e+00 4.28835005e-01 6.53995216e-01
-1.22193456e+00 -3.75805497e-01 1.83187481e-02 1.52343407e-01
-1.14343965e+00 -5.44879138e-01 5.78459859e-01 -3.91442716e-01
1.16652048e+00 3.39472502e-01 6.79536164e-02 6.14524782e-01
-4.30748947e-02 7.65843332e-01 1.02677202e+00 -6.20043933e-01
2.04354867e-01 3.80175799e-01 5.57726383e-01 9.29655790e-01
-1.60654746e-02 -2.55661577e-01 -1.51406154e-01 -5.43207884e-01
2.19477847e-01 -2.49521419e-01 -2.74572313e-01 -2.29751512e-01
-7.23729312e-01 6.83082163e-01 3.19643348e-01 2.74002105e-01
-2.22960800e-01 2.68551409e-01 2.23570436e-01 8.15098345e-01
2.53689259e-01 8.83961678e-01 -7.46241748e-01 -2.49035418e-01
-5.78697026e-01 4.70397830e-01 1.51424181e+00 9.18606520e-01
1.06301045e+00 -4.16235745e-01 -2.12075680e-01 8.06351840e-01
2.71403670e-01 4.11750525e-01 4.18537468e-01 -1.22826469e+00
5.17939270e-01 6.34073913e-01 2.15183198e-01 -5.59789479e-01
5.65105937e-02 -2.46981338e-01 -7.86367208e-02 -3.78273204e-02
7.79295981e-01 -1.81666404e-01 -6.80391490e-01 1.66279817e+00
4.74378586e-01 -1.34453624e-01 1.01349734e-01 7.95426369e-01
6.78212464e-01 7.11406887e-01 2.25518458e-02 2.80549228e-01
1.41752207e+00 -1.26236033e+00 -4.00765836e-01 -3.61410201e-01
8.82264674e-01 -7.48090744e-01 1.38948035e+00 2.56803393e-01
-1.30586720e+00 -3.71184856e-01 -7.34895051e-01 -3.63917589e-01
-1.74114704e-01 -4.79137488e-02 3.68767083e-01 3.26563805e-01
-1.23295176e+00 3.83557320e-01 -7.57388711e-01 -4.32597071e-01
8.97414088e-02 1.97903618e-01 1.34421661e-01 -2.61934668e-01
-1.05046904e+00 1.05601835e+00 1.62625670e-01 -2.03514248e-01
-9.40199435e-01 -7.28653848e-01 -6.96352303e-01 4.80032563e-02
6.58604264e-01 -8.86060774e-01 1.99682617e+00 -1.06426322e+00
-1.64109278e+00 8.40354443e-01 -5.09597480e-01 -1.56436682e-01
1.45013213e-01 -4.31714892e-01 6.19753860e-02 6.00157753e-02
2.36482218e-01 3.52038443e-01 6.63214922e-01 -1.17108226e+00
-5.41434228e-01 -4.94743317e-01 5.84774137e-01 1.99912652e-01
-2.23628685e-01 2.30877459e-01 -7.14045227e-01 -2.46321410e-01
-2.67989665e-01 -8.58426869e-01 -9.34127532e-03 -1.46420643e-01
-4.51448001e-03 -7.59630740e-01 3.12533468e-01 -8.41848850e-01
1.35136843e+00 -2.15859294e+00 1.64228320e-01 8.72295424e-02
2.08125278e-01 3.30465555e-01 -3.17418128e-01 4.65022296e-01
1.68028131e-01 9.26408097e-02 -4.44313228e-01 -1.52615547e-01
8.85584950e-02 2.68573225e-01 -4.93565887e-01 3.09387706e-02
2.58880168e-01 1.13915753e+00 -1.05785036e+00 -3.99716139e-01
-4.52079505e-01 -2.50204243e-02 -7.92016089e-01 8.48158598e-01
-9.70201969e-01 9.59250852e-02 -6.91831589e-01 4.83959079e-01
2.38312289e-01 -3.74725848e-01 1.30811915e-01 2.29130432e-01
1.35256261e-01 6.84691012e-01 -8.46969068e-01 1.83164442e+00
-7.31063426e-01 2.85834461e-01 2.80004174e-01 -8.14813673e-01
8.17948639e-01 2.79703766e-01 -3.92113566e-01 -5.58949649e-01
-2.16443747e-01 5.17963111e-01 6.61995308e-03 -9.71272111e-01
3.90008003e-01 -2.31473669e-01 -1.29852206e-01 7.45209396e-01
2.80024055e-02 -4.21630651e-01 1.51065528e-01 2.32331544e-01
1.21426690e+00 3.12971056e-01 7.83732161e-02 -2.00659901e-01
5.33126533e-01 2.18168065e-01 1.82344452e-01 9.55116749e-01
2.19375983e-01 3.37281048e-01 5.33057749e-01 -3.87887806e-01
-7.96242535e-01 -1.09538674e+00 2.45747730e-01 1.53300905e+00
-2.25496024e-01 -4.05647397e-01 -9.16945755e-01 -8.13772678e-01
-1.39025906e-02 1.02349532e+00 -3.02763820e-01 -8.93526375e-02
-9.34451640e-01 -4.66034748e-02 6.76426709e-01 5.44227183e-01
2.91261971e-01 -1.09880114e+00 -5.50008416e-01 1.98683664e-01
-1.20425180e-01 -6.08256519e-01 -7.45409667e-01 1.26169696e-01
-8.37099612e-01 -1.00950420e+00 -4.15106267e-01 -1.00257635e+00
7.11468101e-01 -3.64100710e-02 1.60889494e+00 5.43578863e-01
-1.25072345e-01 5.50754249e-01 -1.82352513e-01 -1.82362139e-01
-4.91679758e-01 4.05981302e-01 -6.35248363e-01 -3.84715647e-01
6.23393476e-01 -5.11145651e-01 -4.77993369e-01 2.07538992e-01
-1.06811881e+00 -2.67702192e-01 4.73980755e-01 7.82625198e-01
4.50295329e-01 -2.55615383e-01 6.14623249e-01 -1.35015416e+00
8.55965734e-01 -6.79586887e-01 -8.43938708e-01 5.43704748e-01
-4.45602417e-01 5.53116381e-01 8.21557462e-01 -2.90776312e-01
-1.18412232e+00 -1.20553926e-01 -3.32235068e-01 -2.12919295e-01
-2.54816055e-01 6.25815332e-01 6.20755926e-02 1.68208793e-01
9.51701999e-01 1.73675239e-01 -1.47292092e-01 -6.67251706e-01
5.69449246e-01 8.40399265e-01 5.59622526e-01 -1.14451647e+00
7.60636687e-01 5.81185445e-02 -6.01559460e-01 -4.74746555e-01
-1.23799682e+00 -4.98430192e-01 -2.70948261e-01 2.67090350e-01
6.41041756e-01 -6.47030175e-01 -4.30613786e-01 2.38097817e-01
-1.37204599e+00 -6.28450871e-01 -1.76308766e-01 -5.53187951e-02
-3.77739131e-01 1.13698438e-01 -6.94425344e-01 -5.76421738e-01
-4.49363291e-01 -1.04404414e+00 8.92399609e-01 2.27819666e-01
-5.24164975e-01 -9.73097384e-01 4.63679522e-01 7.27662444e-01
6.83991373e-01 -2.87739724e-01 1.19785094e+00 -9.12937701e-01
-8.57504725e-01 -1.60788894e-01 -1.06960408e-01 3.77953410e-01
-1.54085055e-01 -2.24443302e-01 -9.50489998e-01 -2.79260188e-01
3.90068978e-01 -8.31257999e-01 5.85310876e-01 -3.58868837e-01
1.29243147e+00 -4.93744045e-01 -4.80771773e-02 3.95335138e-01
1.32247245e+00 -8.51543173e-02 5.60937166e-01 2.54630923e-01
5.79609394e-01 5.88978112e-01 4.03974831e-01 -7.45377019e-02
8.44150424e-01 3.79375130e-01 3.89967002e-02 3.79957855e-01
-8.18234980e-02 -3.79008651e-01 3.82374257e-01 8.93813550e-01
4.42569941e-01 -9.97770727e-02 -1.24605393e+00 8.24022949e-01
-1.83812594e+00 -6.59991145e-01 5.64699024e-02 2.11265516e+00
1.23423374e+00 -1.00238457e-01 -7.04524666e-02 -4.80125993e-01
4.85056669e-01 3.09164878e-02 -6.24300182e-01 -7.00733125e-01
6.46272957e-01 4.88092512e-01 3.56918829e-03 8.81158829e-01
-6.44440293e-01 9.19228494e-01 5.66997957e+00 4.70598817e-01
-8.88203204e-01 3.59264046e-01 3.12884569e-01 -2.02513896e-02
-7.42630541e-01 4.11169410e-01 -7.56424367e-01 2.36200750e-01
1.26325381e+00 -1.57574028e-01 7.87024975e-01 8.57239664e-01
-4.68949556e-01 -1.03838541e-01 -1.43714476e+00 6.03474557e-01
3.56837541e-01 -9.44970250e-01 7.12318905e-03 -3.56601000e-01
5.93146384e-01 1.78271785e-01 -1.72682554e-01 6.95706725e-01
4.87587392e-01 -9.13046300e-01 5.01398146e-01 5.83752155e-01
4.58365262e-01 -5.37495315e-01 6.36345029e-01 6.33367717e-01
-7.86467254e-01 -7.18955994e-02 -3.16160381e-01 -1.25885889e-01
-3.74719799e-02 2.46181458e-01 -7.73887873e-01 2.44985476e-01
6.40388668e-01 2.41351008e-01 -6.69303477e-01 8.36669266e-01
-5.21898627e-01 8.05248618e-01 -3.08171153e-01 -2.88896173e-01
9.58493575e-02 9.92550850e-02 3.48804921e-01 1.06755304e+00
1.74741805e-01 3.21801454e-01 3.35341632e-01 1.06374621e+00
-4.21491712e-01 8.34318697e-02 -5.50517917e-01 -1.08958893e-01
3.39029491e-01 1.06157279e+00 7.55069545e-03 -5.43650508e-01
-6.08864903e-01 9.09230113e-01 1.06278646e+00 4.83252048e-01
-5.37038922e-01 -6.74003541e-01 2.90988952e-01 2.83108294e-01
3.55024874e-01 -6.98721856e-02 -6.81711063e-02 -1.37936532e+00
6.14942014e-01 -1.27611983e+00 6.94609761e-01 -7.70697415e-01
-1.42865360e+00 5.91081798e-01 4.12486540e-03 -5.18794537e-01
-4.27593350e-01 -3.97273481e-01 -6.66691422e-01 1.14923108e+00
-1.75333953e+00 -6.07531607e-01 -8.87306333e-02 4.00289059e-01
4.80308354e-01 7.35825375e-02 9.63418245e-01 2.13409200e-01
-3.17052692e-01 6.40587389e-01 3.46086510e-02 2.11550027e-01
7.72453487e-01 -1.49907458e+00 4.26349580e-01 6.14110768e-01
7.08758980e-02 1.05318236e+00 4.86694932e-01 -4.58446324e-01
-1.77379537e+00 -8.63636374e-01 1.29606640e+00 -9.64237511e-01
1.04545677e+00 -2.83098310e-01 -1.46190584e+00 8.74107778e-01
4.71207142e-01 -2.73042828e-01 7.60926008e-01 1.89825013e-01
-7.62832642e-01 -1.62958786e-01 -9.04911995e-01 5.57746649e-01
5.59810638e-01 -9.49475586e-01 -1.21362627e+00 3.34687889e-01
7.52505660e-01 -5.23842275e-01 -9.17940319e-01 1.73004568e-02
3.14494550e-01 -6.25149786e-01 6.32680237e-01 -6.19925797e-01
5.13288796e-01 -5.87090313e-01 -1.33994922e-01 -1.07490051e+00
-1.27554849e-01 -5.38673759e-01 -1.73860148e-01 1.34725034e+00
8.10239494e-01 -8.65003943e-01 5.25288701e-01 1.12448359e+00
-7.55024776e-02 -9.39267218e-01 -7.02958822e-01 -5.05007446e-01
3.76943469e-01 -2.60066390e-01 5.78542233e-01 6.74308717e-01
7.53440987e-03 7.47421563e-01 1.15885675e-01 2.73101300e-01
3.37571293e-01 5.49791992e-01 9.13833261e-01 -8.28670025e-01
-8.87737632e-01 -1.64949745e-01 2.30854526e-01 -1.40674317e+00
3.98949921e-01 -1.17256045e+00 4.17746872e-01 -1.51071656e+00
2.67759800e-01 -4.46802825e-01 -9.48538035e-02 6.13889873e-01
-5.14568567e-01 -1.78540394e-01 1.70439109e-03 3.05967420e-01
-8.47926915e-01 2.81369001e-01 1.04685521e+00 -8.64944831e-02
-4.62279767e-02 2.95616017e-04 -1.12512279e+00 5.52122414e-01
1.03345120e+00 -8.07456493e-01 -5.22467673e-01 -1.13086855e+00
4.30404842e-01 2.75817901e-01 2.85495251e-01 -5.74614406e-01
5.39548695e-01 -1.40810357e-02 -1.37057304e-01 -7.99108520e-02
6.13880306e-02 -5.59786499e-01 -2.14044377e-01 3.67218316e-01
-6.64373040e-01 4.59790230e-01 3.79801363e-01 3.80470693e-01
-2.85377949e-01 -8.27196181e-01 5.38940191e-01 -3.62360835e-01
-4.02646571e-01 3.00041921e-02 -2.96341419e-01 7.54806161e-01
7.07078636e-01 3.93848300e-01 -6.38488233e-01 -3.89172107e-01
-3.32779288e-01 6.92155540e-01 6.88210905e-01 2.50221044e-01
4.00278360e-01 -8.57394874e-01 -7.93818593e-01 -8.36283937e-02
2.44095951e-01 2.87186593e-01 -2.25202829e-01 3.98416281e-01
-5.92876315e-01 2.84570783e-01 4.53805566e-01 -4.23786074e-01
-1.09356022e+00 5.66137791e-01 5.45770526e-01 -4.17688638e-01
-3.62128168e-01 8.63000453e-01 6.94571882e-02 -8.79171610e-01
1.66615695e-01 -2.20181048e-01 5.47466837e-02 -2.02741399e-01
6.19034171e-01 -4.57001738e-02 8.96632820e-02 -2.64221411e-02
-2.85184115e-01 4.65268940e-01 -5.06122172e-01 -2.25045815e-01
1.14190447e+00 9.99763012e-02 -6.65544450e-01 4.96971905e-01
1.36407590e+00 -3.01942229e-02 -8.17226112e-01 -5.62132955e-01
3.34789246e-01 -4.90386754e-01 -2.28811607e-01 -1.04409277e+00
-6.25868797e-01 7.80759037e-01 6.09063730e-02 1.14323728e-01
9.31876719e-01 4.40917969e-01 8.72870088e-01 1.14867139e+00
3.06702107e-01 -8.13315451e-01 1.72953457e-01 7.87510991e-01
8.72446835e-01 -1.15315855e+00 -2.68901229e-01 -4.50045727e-02
-5.24781585e-01 9.54668403e-01 9.34450209e-01 -9.15692449e-02
4.15313780e-01 1.53949827e-01 4.10529315e-01 -3.64894331e-01
-1.15633774e+00 2.83073112e-02 6.14316389e-02 3.39662045e-01
6.91611409e-01 -5.84217608e-02 -3.10195033e-02 4.48973417e-01
-2.70258427e-01 1.98714763e-01 3.24174047e-01 1.26181161e+00
-4.71907616e-01 -1.42192101e+00 -2.80957282e-01 5.62880099e-01
-6.01707101e-01 -3.86130363e-01 -4.52865750e-01 4.23926324e-01
-3.36651295e-01 1.02145934e+00 -8.09478015e-02 1.01986192e-01
6.35087013e-01 5.23392677e-01 4.93024230e-01 -1.15204251e+00
-1.06623757e+00 -4.72290397e-01 4.24709857e-01 -4.31725383e-01
-5.26938960e-02 -4.12748635e-01 -1.28664732e+00 -3.16195451e-02
-2.17625529e-01 7.40651131e-01 4.00772244e-01 8.56726289e-01
5.41044235e-01 1.87179983e-01 4.39390033e-01 -4.15951423e-02
-1.18597901e+00 -7.48779178e-01 2.53497604e-02 5.19172788e-01
3.54955077e-01 4.77931574e-02 -4.08630639e-01 -3.57088260e-02] | [11.12990951538086, 8.049180030822754] |
87999a1f-1f62-48a4-b6f2-53bfe363387a | image-harmonization-datasets-hcoco-hadobe5k | 1908.10526 | null | https://arxiv.org/abs/1908.10526v4 | https://arxiv.org/pdf/1908.10526v4.pdf | Image Harmonization Dataset iHarmony4: HCOCO, HAdobe5k, HFlickr, and Hday2night | Image composition is an important operation in image processing, but the inconsistency between foreground and background significantly degrades the quality of composite image. Image harmonization, which aims to make the foreground compatible with the background, is a promising yet challenging task. However, the lack of high-quality public dataset for image harmonization, which significantly hinders the development of image harmonization techniques. Therefore, we contribute an image harmonization dataset iHarmony4 by generating synthesized composite images based on existing COCO (resp., Adobe5k, day2night) dataset, leading to our HCOCO (resp., HAdobe5k, Hday2night) sub-dataset. To enrich the diversity of our dataset, we also generate synthesized composite images based on our collected Flick images, leading to our HFlickr sub-dataset. The image harmonization dataset iHarmony4 is released at https://github.com/bcmi/Image_Harmonization_Datasets. | ['Wenyan Cong', 'Zhixin Ling', 'Weiyuan Li', 'Li Niu', 'Jianfu Zhang', 'Liu Liu', 'Liqing Zhang'] | 2019-08-28 | null | null | null | null | ['image-harmonization'] | ['computer-vision'] | [ 3.43350261e-01 -4.62814569e-01 1.96221948e-01 -8.63328110e-03
-4.88769352e-01 -8.08034241e-01 6.02044702e-01 7.09582046e-02
-2.11608961e-01 5.73479533e-01 -2.05305386e-02 -1.28360212e-01
1.07671760e-01 -8.82273197e-01 -6.87961280e-01 -8.04130971e-01
5.20513833e-01 -2.70816386e-01 3.01670104e-01 -2.60222465e-01
1.80043459e-01 2.59335607e-01 -1.76802254e+00 3.82667273e-01
1.07091832e+00 9.42084014e-01 3.68342042e-01 8.00008893e-01
1.23703986e-01 4.06268597e-01 -5.66278994e-01 -5.24927974e-01
4.92945194e-01 -7.01927245e-01 -4.60995108e-01 2.07053885e-01
5.81890583e-01 8.87902454e-02 -2.35012606e-01 1.29958785e+00
6.96134627e-01 2.54749525e-02 2.60503471e-01 -1.72139561e+00
-7.84403563e-01 1.93707675e-01 -4.83640254e-01 1.02087893e-01
-2.42700819e-02 4.39959824e-01 7.05426216e-01 -6.90777481e-01
7.97585130e-01 9.22681212e-01 2.84280032e-01 2.01088995e-01
-1.24891317e+00 -8.54617417e-01 -1.95674345e-01 4.00369078e-01
-1.53365862e+00 -5.97202241e-01 6.90034926e-01 -1.92723140e-01
1.22882172e-01 7.70357966e-01 9.02331769e-01 7.75342703e-01
-9.57243592e-02 5.18547356e-01 1.31657231e+00 -3.83660585e-01
-1.82388816e-02 -4.12556864e-02 -5.28868616e-01 1.16981611e-01
1.10150680e-01 -1.65648431e-01 -4.56347793e-01 2.90929168e-01
5.79081476e-01 -1.89522818e-01 -6.81868017e-01 -8.39562118e-02
-1.37509024e+00 3.34912300e-01 5.18660724e-01 2.20669895e-01
-1.50255665e-01 -1.73866913e-01 4.10735682e-02 2.96482891e-01
3.72200876e-01 2.64506400e-01 -1.77622989e-01 3.63666750e-02
-9.16935802e-01 4.25307393e-01 2.16299936e-01 1.07189023e+00
8.60435665e-01 -1.16953418e-01 -1.87164307e-01 1.05762756e+00
-1.42501174e-02 7.25392342e-01 1.90994114e-01 -1.22042847e+00
3.36910576e-01 6.88495457e-01 2.09481239e-01 -1.19232786e+00
-7.17721507e-02 -5.14826365e-02 -9.97908950e-01 2.05250770e-01
2.14116201e-01 -3.15615945e-02 -6.33080125e-01 1.45585501e+00
5.05327702e-01 6.81289136e-02 -4.34925314e-04 9.58605587e-01
1.17742038e+00 8.20152938e-01 -3.02586943e-01 -2.50399053e-01
1.32389009e+00 -1.20282137e+00 -6.23100460e-01 2.50802729e-02
-4.71708411e-03 -1.28388834e+00 1.20667744e+00 5.17277062e-01
-1.04011977e+00 -7.84380317e-01 -9.17517900e-01 -1.29355401e-01
-4.36109602e-01 -1.57703236e-02 2.22858146e-01 5.45068681e-01
-1.13482869e+00 6.38226718e-02 -1.29448652e-01 -3.96730751e-01
2.32852653e-01 -2.86054522e-01 -4.32731748e-01 -2.17535138e-01
-1.01163375e+00 4.87127095e-01 5.56825697e-01 1.79706216e-01
-6.58106267e-01 -8.01607072e-01 -6.99122012e-01 -4.64886218e-01
4.71923143e-01 -3.26728404e-01 8.55042279e-01 -9.97878730e-01
-1.20867574e+00 1.13476992e+00 7.52273276e-02 -1.82324246e-01
7.09448218e-01 1.05446562e-01 -7.41435349e-01 1.80408046e-01
1.35201961e-01 1.04050887e+00 7.81727791e-01 -1.61932039e+00
-7.88503647e-01 -1.82733729e-01 4.71544042e-02 1.90286532e-01
-1.29181311e-01 -1.30168140e-01 -1.09535646e+00 -8.81172121e-01
-2.46025354e-01 -1.04271007e+00 3.31954136e-02 -9.93445218e-02
-6.24675274e-01 3.26434314e-01 9.24422443e-01 -8.45370293e-01
1.31843555e+00 -2.53695703e+00 -1.29942121e-02 4.57298160e-02
1.53398186e-01 2.01285630e-01 -4.72873718e-01 3.68292093e-01
-2.21898183e-01 2.01519072e-01 -4.27288830e-01 -1.04124010e-01
-9.46652815e-02 5.91847207e-03 -9.04453695e-02 4.14276034e-01
2.07480967e-01 7.77578115e-01 -7.87015378e-01 -6.74637675e-01
4.35848773e-01 4.84214127e-01 -5.07672668e-01 2.06737652e-01
-2.88206100e-01 7.30982542e-01 9.76475775e-02 7.24300385e-01
1.07760847e+00 2.09116507e-02 1.03867888e-01 -7.12479949e-01
-4.72305298e-01 -5.46591103e-01 -1.20596409e+00 1.59048390e+00
-3.02267551e-01 8.87430549e-01 3.14239189e-02 -5.09174407e-01
9.86456931e-01 9.47160199e-02 6.34643018e-01 -1.20361423e+00
6.37351573e-02 2.05177039e-01 -1.54516488e-01 -2.03087226e-01
7.97769368e-01 3.26763928e-01 -1.02576412e-01 2.00089261e-01
-4.42141026e-01 -6.42679393e-01 6.90434635e-01 1.80204138e-01
5.97869694e-01 1.28232032e-01 1.18447974e-01 -7.73342028e-02
4.87948745e-01 2.16544330e-01 5.65403283e-01 4.11168277e-01
-3.33610952e-01 1.24775600e+00 3.84781957e-01 -3.58928323e-01
-1.31415784e+00 -1.17424762e+00 -1.86516434e-01 6.33701086e-01
6.19455576e-01 -6.01224005e-01 -9.36532974e-01 -2.18085814e-02
-2.20638216e-01 3.97090316e-01 -4.74489272e-01 8.99676457e-02
-3.00033391e-01 -8.77337992e-01 6.11923575e-01 -7.29429275e-02
1.18424070e+00 -1.23469973e+00 -5.76370716e-01 -3.65446135e-02
-7.01661825e-01 -1.14709044e+00 -7.33847558e-01 -3.88443112e-01
-6.25435486e-02 -1.43993354e+00 -8.35333347e-01 -6.30638599e-01
5.62977016e-01 7.09726512e-01 1.30445981e+00 2.18949765e-01
-6.38213515e-01 1.77999765e-01 -6.68003380e-01 -3.93953383e-01
-4.90094453e-01 -3.24046731e-01 -2.77683139e-01 2.70706505e-01
-4.43071499e-02 -3.72571945e-01 -1.11714125e+00 6.40637577e-01
-1.52396858e+00 6.32941246e-01 2.12911665e-01 4.34946179e-01
9.71307874e-01 5.51316261e-01 1.28252327e-01 -5.60227931e-01
4.52872068e-01 -1.69130385e-01 -8.70795131e-01 4.86285836e-01
-3.18374813e-01 -5.67478716e-01 3.66847783e-01 -1.63916752e-01
-1.21615505e+00 -4.59994376e-02 2.42120087e-01 -2.94128716e-01
-6.44361973e-02 1.75813630e-01 -5.29578269e-01 7.95125030e-03
5.42780459e-01 2.40148693e-01 -5.54253981e-02 -3.64764392e-01
5.71940362e-01 7.49983251e-01 1.10654008e+00 -2.06697896e-01
7.41317391e-01 4.81447965e-01 -2.87097037e-01 -8.85544956e-01
-4.52257901e-01 -2.22138360e-01 -4.15417880e-01 -5.11902690e-01
1.05020654e+00 -9.90199924e-01 -3.50284547e-01 8.84475350e-01
-8.90114307e-01 -7.68893898e-01 -5.04083671e-02 -9.34007019e-02
-2.97927946e-01 3.64727795e-01 -2.11505949e-01 -4.28758264e-01
-3.52874905e-01 -1.19388926e+00 8.76999557e-01 5.03511727e-01
-1.59799680e-02 -4.52965200e-01 7.32735246e-02 6.15394831e-01
3.81504655e-01 6.44106030e-01 5.74943662e-01 3.29137087e-01
-4.93658334e-01 7.40336776e-02 -6.20178163e-01 3.71376336e-01
3.62531513e-01 6.95341170e-01 -8.23213577e-01 -9.59142298e-02
-4.06385154e-01 3.98958698e-02 6.85034156e-01 3.36828858e-01
1.28598976e+00 -1.60040259e-01 9.99660566e-02 6.80209935e-01
1.35666466e+00 2.91908771e-01 1.16091311e+00 7.15669870e-01
6.79811716e-01 5.75439692e-01 8.73637915e-01 6.01756811e-01
6.41864598e-01 9.60417926e-01 4.36248243e-01 -5.69679320e-01
-5.86942554e-01 8.92097317e-03 1.59279272e-01 7.39933431e-01
-5.49412258e-02 -5.26114702e-01 -7.99727559e-01 4.72868800e-01
-1.81387317e+00 -7.77065158e-01 -5.34636915e-01 2.13579559e+00
9.76436794e-01 -3.91055524e-01 1.31688640e-01 1.94853783e-01
1.08690035e+00 2.53629357e-01 -4.15391237e-01 7.80571776e-04
-7.64459729e-01 3.75234224e-02 3.38014394e-01 3.11299026e-01
-1.16754282e+00 8.10029209e-01 5.08559608e+00 9.98708844e-01
-1.34311414e+00 8.84245336e-02 8.77996027e-01 -2.35951841e-01
-3.36652696e-01 -1.00595199e-01 -2.30350897e-01 1.00052810e+00
4.92093593e-01 -3.21840286e-01 6.50672138e-01 3.53305578e-01
4.12306160e-01 -4.98612970e-01 -4.04412985e-01 1.39177644e+00
1.89473495e-01 -1.32442784e+00 -5.95055968e-02 -1.09669946e-01
1.14944267e+00 -1.60636216e-01 2.31961876e-01 -1.66981786e-01
1.10596567e-01 -8.33242893e-01 9.18206811e-01 3.33116591e-01
1.03575850e+00 -7.26415813e-01 5.11809230e-01 -1.36335373e-01
-1.43646753e+00 3.63706827e-01 -2.01311424e-01 5.50229669e-01
2.16742516e-01 7.02021360e-01 -3.01575124e-01 8.21555316e-01
1.18111527e+00 7.26709545e-01 -1.14369631e+00 1.22117007e+00
-2.17388853e-01 1.41370103e-01 -1.28738536e-02 6.15292251e-01
-2.48761952e-01 -6.06073916e-01 4.13282275e-01 1.03405762e+00
4.46704000e-01 -1.41326711e-01 5.03145345e-02 7.43687570e-01
-1.19761333e-01 2.73543626e-01 -3.46086949e-01 -1.43569157e-01
5.50422192e-01 1.46846437e+00 -8.01549196e-01 -2.82164931e-01
-2.00402439e-01 1.18295753e+00 -3.38450342e-01 3.33154321e-01
-1.13404691e+00 -3.84219199e-01 6.37647808e-01 -3.61928414e-03
2.55300943e-02 7.67535120e-02 -1.16051629e-01 -1.21288526e+00
1.49308935e-01 -1.40572500e+00 3.93303245e-01 -1.17951381e+00
-1.02297008e+00 6.40886843e-01 -2.07454246e-02 -1.48853779e+00
3.09389681e-01 -1.55833378e-01 -6.37557685e-01 7.06720889e-01
-1.34873474e+00 -1.35911691e+00 -1.06395817e+00 7.97402501e-01
3.88925403e-01 4.63456176e-02 4.95169878e-01 6.75716043e-01
-7.90282965e-01 3.27642769e-01 9.08279419e-02 -1.30290985e-01
1.11286867e+00 -8.97627890e-01 4.48733807e-01 1.04948330e+00
9.73336548e-02 6.14176318e-02 7.79834330e-01 -5.19208431e-01
-1.24621403e+00 -1.33356953e+00 4.23376858e-01 -1.91423014e-01
3.65349561e-01 -1.75975800e-01 -8.60697269e-01 1.14680730e-01
4.60692227e-01 1.96164414e-01 4.72964019e-01 -7.58440912e-01
-1.71007097e-01 -3.51457745e-01 -1.01955414e+00 9.51434314e-01
8.80712390e-01 -3.77889961e-01 1.84703618e-01 1.68863550e-01
7.71884203e-01 -4.51280057e-01 -9.78648484e-01 4.55301195e-01
5.72137058e-01 -1.28109241e+00 8.77056539e-01 3.22700918e-01
6.62241161e-01 -9.79955196e-01 -2.59484887e-01 -1.27902544e+00
-5.34043722e-02 -5.99629164e-01 4.38977569e-01 1.60379148e+00
2.74229527e-01 -4.62636322e-01 3.09536934e-01 3.66244555e-01
2.47662123e-02 -1.65905982e-01 -4.20862824e-01 -5.74078560e-01
-2.75908053e-01 -4.07881171e-01 8.93259823e-01 9.61534798e-01
-4.97709513e-01 -2.50332922e-01 -5.06924152e-01 -1.05012413e-02
5.49990952e-01 4.84546423e-01 1.40090597e+00 -6.90250099e-01
-1.71210274e-01 -4.26579952e-01 -2.46751592e-01 -2.66883373e-01
-2.77719378e-01 -6.40634418e-01 4.60277498e-02 -1.46391416e+00
3.54542971e-01 -4.15447891e-01 -7.40820915e-02 4.72412348e-01
-3.19922715e-01 1.32692146e+00 4.88529325e-01 1.48158655e-01
-6.87126637e-01 4.17358518e-01 1.43281353e+00 -2.54882991e-01
-2.09443957e-01 -4.96553451e-01 -7.09121466e-01 4.31658924e-01
1.01582885e+00 -1.68637350e-01 -2.13514641e-01 -2.60288328e-01
1.79817572e-01 -2.80872703e-01 5.35109460e-01 -9.96434212e-01
-8.90965089e-02 -3.86659473e-01 5.02830029e-01 -7.29315281e-01
1.79540873e-01 -5.72991312e-01 1.04050338e+00 4.46705759e-01
7.88785145e-02 2.38334194e-01 3.53242874e-01 2.11997271e-01
-5.82730651e-01 1.51601002e-01 9.76109207e-01 -1.31638318e-01
-1.11640465e+00 3.79311323e-01 -4.25552838e-02 2.04153746e-01
1.08413470e+00 -2.31697872e-01 -5.99139035e-01 -3.71634066e-01
-3.22615296e-01 2.18442872e-01 1.08372867e+00 5.43604314e-01
4.91561443e-01 -1.52457321e+00 -7.69446850e-01 1.92206353e-01
5.35765529e-01 1.68797702e-01 7.02538610e-01 8.62142742e-01
-1.00722003e+00 -1.92876428e-01 -6.01025760e-01 -4.84140664e-01
-1.63211632e+00 4.21236873e-01 -6.85086893e-03 1.82987511e-01
-6.15720093e-01 4.81284916e-01 3.37584764e-01 -2.43738338e-01
-1.62156001e-01 1.89270396e-02 -6.41579106e-02 2.31328517e-01
6.87662303e-01 3.95354658e-01 9.56600532e-02 -8.93924832e-01
-2.44154826e-01 5.22309363e-01 4.56897229e-01 -1.35316059e-01
1.16862535e+00 -5.05374312e-01 -5.65789521e-01 8.99304822e-02
1.11498201e+00 2.38847077e-01 -1.09543419e+00 3.85215320e-02
-2.87724644e-01 -8.86488736e-01 -8.65733698e-02 -6.47799969e-01
-1.35823035e+00 4.15813625e-01 8.81056607e-01 2.91976810e-01
1.62470758e+00 -3.35261852e-01 7.18778551e-01 -2.14736238e-01
1.80863723e-01 -1.07524872e+00 1.90857708e-01 1.53491542e-01
1.17268848e+00 -1.38194251e+00 7.96153098e-02 -4.45999563e-01
-8.39829028e-01 7.57499456e-01 6.61026359e-01 2.13808253e-01
3.13569605e-01 1.41422078e-01 5.23318648e-01 4.02935334e-02
-3.30784202e-01 -3.14978451e-01 3.00476462e-01 6.20829463e-01
3.45141739e-01 2.10010394e-01 -1.49104208e-01 2.04788074e-01
-4.73762184e-01 -1.00991070e-01 5.27728498e-01 6.26608074e-01
-1.25222683e-01 -1.14754260e+00 -6.62284255e-01 5.72791211e-02
-2.04039097e-01 -2.34027505e-01 -4.83532667e-01 6.62451267e-01
5.76192498e-01 1.20639741e+00 1.26614809e-01 -2.80786663e-01
2.55009472e-01 -6.03577018e-01 3.93865407e-01 -1.44953147e-01
-3.96700740e-01 2.67821312e-01 -9.05378163e-02 -7.15621352e-01
-6.74684882e-01 -3.23566467e-01 -8.68713856e-01 -5.94944894e-01
5.07132635e-02 -1.26652405e-01 6.21903300e-01 3.65645140e-01
3.83361667e-01 5.06846964e-01 6.32178068e-01 -9.71039474e-01
5.25191367e-01 -6.45275414e-01 -6.52592063e-01 7.38409400e-01
-1.80325620e-02 -4.42052722e-01 -1.20499514e-01 5.99537313e-01] | [11.21517276763916, -1.2293883562088013] |
aedadfd0-d8ea-46a6-a2d8-055950e8f746 | monoise-a-multi-lingual-and-easy-to-use | null | null | https://aclanthology.org/P19-3032 | https://aclanthology.org/P19-3032.pdf | MoNoise: A Multi-lingual and Easy-to-use Lexical Normalization Tool | In this paper, we introduce and demonstrate the online demo as well as the command line interface of a lexical normalization system (MoNoise) for a variety of languages. We further improve this model by using features from the original word for every normalization candidate. For comparison with future work, we propose the bundling of seven datasets in six languages to form a new benchmark, together with a novel evaluation metric which is particularly suitable for cross-dataset comparisons. MoNoise reaches a new state-of-art performance for six out of seven of these datasets. Furthermore, we allow the user to tune the {`}aggressiveness{'} of the normalization, and show how the model can be made more efficient with only a small loss in performance. The online demo can be found on: http://www.robvandergoot.com/monoise and the corresponding code on: https://bitbucket.org/robvanderg/monoise/ | ['Rob van der Goot'] | 2019-07-01 | null | null | null | acl-2019-7 | ['lexical-normalization'] | ['natural-language-processing'] | [-1.91923231e-01 -2.03705937e-01 -2.72546083e-01 -4.47029382e-01
-8.80163670e-01 -7.32174993e-01 9.85546887e-01 5.10895252e-01
-9.34731245e-01 5.66582084e-01 2.62173057e-01 -2.46158317e-01
-1.51184663e-01 -6.08174086e-01 -3.67884845e-01 -1.93713397e-01
1.67878792e-01 5.05308092e-01 3.94116759e-01 -4.81802821e-01
2.85668522e-01 3.35358739e-01 -1.44759572e+00 3.03615153e-01
6.01719737e-01 7.27705598e-01 1.18316017e-01 6.27139628e-01
-1.26094550e-01 1.48820028e-01 -5.45400798e-01 -8.18432152e-01
4.04786974e-01 -1.30414113e-01 -8.47051978e-01 -4.31050688e-01
2.91466087e-01 -3.12785991e-02 4.45535965e-02 1.07033122e+00
8.10729921e-01 1.08757205e-01 3.80085588e-01 -1.16225564e+00
-2.98450679e-01 8.38543653e-01 -3.55476558e-01 3.07426572e-01
4.37662065e-01 1.44670665e-01 1.19887900e+00 -1.10482919e+00
7.49404550e-01 1.08972716e+00 4.99538779e-01 4.06313539e-01
-1.13875413e+00 -7.61390328e-01 -2.63406392e-02 2.16621339e-01
-1.55519640e+00 -7.79541016e-01 3.83181512e-01 -2.38089472e-01
1.24943161e+00 4.37446773e-01 3.09807003e-01 1.00996411e+00
-1.67379007e-01 6.42575324e-01 9.65216815e-01 -6.37126088e-01
-8.35189223e-02 1.41034395e-01 3.21229577e-01 5.58595598e-01
2.08380267e-01 -3.16072442e-02 -5.10391295e-01 -1.35142654e-01
4.63081330e-01 -3.73638391e-01 -1.01106361e-01 -3.53284776e-01
-1.35010660e+00 8.20269465e-01 2.68718123e-01 4.02057111e-01
-1.39020309e-01 -1.44296139e-03 6.61695421e-01 3.27394754e-01
4.66259480e-01 4.73046064e-01 -5.00185311e-01 -5.33249795e-01
-9.25794601e-01 3.80264997e-01 9.52234983e-01 8.49838495e-01
6.41179919e-01 -3.18924278e-01 2.25799326e-02 1.11413956e+00
-7.38005713e-03 3.00364882e-01 5.75985372e-01 -8.62120092e-01
5.43341815e-01 5.39344966e-01 -2.10345201e-02 -6.07428193e-01
-7.08270729e-01 -4.28418934e-01 -5.48904896e-01 -9.37838405e-02
6.23663485e-01 -5.38339093e-02 -5.88259399e-01 1.61278617e+00
3.80502045e-01 -3.04108262e-01 -6.57980591e-02 8.05378914e-01
9.65718150e-01 5.21876633e-01 -6.35149842e-03 -4.53470722e-02
1.56252813e+00 -9.11129296e-01 -5.71954966e-01 -2.23095953e-01
9.21480954e-01 -1.15558982e+00 1.30835283e+00 3.84766459e-01
-1.07011485e+00 -3.68545562e-01 -1.13672006e+00 -1.68584555e-01
-6.65259957e-01 1.52762413e-01 6.46695673e-01 6.21167958e-01
-8.97398531e-01 5.92790544e-01 -1.03355253e+00 -7.84844160e-01
6.86951131e-02 4.32539195e-01 -4.87244785e-01 5.34623750e-02
-1.34251320e+00 1.07268298e+00 7.29001701e-01 -1.61364615e-01
-1.69185594e-01 -6.02484584e-01 -7.37231374e-01 -3.18654031e-01
5.83434880e-01 -4.67602640e-01 1.29916096e+00 -7.30102718e-01
-1.49132514e+00 1.15920258e+00 1.31668225e-01 -2.94620365e-01
6.86171651e-01 -4.06199664e-01 -5.62257826e-01 -2.88761854e-01
1.43930867e-01 5.69228530e-01 2.42220148e-01 -6.84639692e-01
-5.71719944e-01 -1.33714572e-01 2.30597690e-01 1.00441396e-01
-3.36907834e-01 6.20378494e-01 -8.71922910e-01 -8.17105472e-01
-3.10512871e-01 -7.67050683e-01 -3.18557248e-02 -4.44052100e-01
-4.09626484e-01 -2.23114893e-01 9.62804854e-02 -7.38422275e-01
1.66959405e+00 -2.19396639e+00 1.85745776e-01 1.45989642e-01
4.24279682e-02 4.22520280e-01 -2.63593227e-01 6.45553291e-01
-2.58082896e-01 3.51481199e-01 -3.08973044e-01 -3.18370789e-01
1.82501197e-01 1.00815274e-01 2.61160314e-01 5.23286521e-01
1.30995274e-01 8.03840280e-01 -8.40579152e-01 -5.73228449e-02
1.64479777e-01 4.28774863e-01 -5.75332165e-01 -1.09632701e-01
1.35430079e-02 1.16313726e-01 -8.46125260e-02 3.75781775e-01
5.83852410e-01 1.14288284e-02 2.57726818e-01 -9.46369246e-02
-2.59612650e-01 7.90322244e-01 -1.55889440e+00 1.75417960e+00
-5.25132298e-01 3.61344099e-01 4.84492863e-03 -5.89774251e-01
7.92067707e-01 5.80365472e-02 1.20523810e-01 -7.93287873e-01
2.66008437e-01 4.31989223e-01 1.56574845e-01 -1.59285754e-01
7.75376797e-01 1.07746519e-01 -3.02060217e-01 6.04994714e-01
1.25357106e-01 -1.86056606e-02 7.48375595e-01 2.11167216e-01
8.68418992e-01 1.11300461e-01 7.07958877e-01 -4.15471852e-01
6.38816059e-01 -1.91114902e-01 3.98205251e-01 5.51813602e-01
-1.62668079e-02 5.30894101e-01 4.62770671e-01 -1.78048417e-01
-1.15956128e+00 -9.02847946e-01 -3.18752617e-01 1.31303620e+00
-1.91883758e-01 -1.09622550e+00 -7.26617932e-01 -5.26940048e-01
3.45046259e-02 9.39938545e-01 -5.00755668e-01 1.06374197e-01
-5.48949122e-01 -8.92760336e-01 7.65574217e-01 3.70856404e-01
2.51408219e-01 -9.71651316e-01 -4.35177565e-01 4.20421660e-02
-2.09350854e-01 -9.59401250e-01 -4.93620068e-01 3.61803263e-01
-4.98759478e-01 -9.07690406e-01 -4.10447568e-01 -3.23773474e-01
1.91081733e-01 -8.48156959e-02 1.12575972e+00 2.26238936e-01
-2.02512562e-01 7.22234547e-02 -6.48584604e-01 -3.77513081e-01
-6.25451803e-01 6.89838946e-01 1.23004019e-01 -2.83482462e-01
6.23263240e-01 -3.86186600e-01 -4.26263332e-01 4.47563082e-01
-9.75073874e-01 7.96055943e-02 2.04972893e-01 6.99908674e-01
3.94886225e-01 -2.14501217e-01 3.72056872e-01 -1.02729762e+00
7.15069532e-01 -6.06298864e-01 -7.19492316e-01 1.39646493e-02
-6.82419837e-01 8.79900977e-02 5.17780066e-01 -6.90095648e-02
-5.92401147e-01 -2.49399826e-01 -5.63440740e-01 1.34707078e-01
-1.94481090e-01 6.82390630e-01 -2.35664442e-01 3.85983586e-01
8.14657629e-01 -2.10914627e-01 -3.78369987e-02 -9.06969190e-01
7.24953711e-01 8.99970591e-01 5.30989170e-01 -6.63213909e-01
7.03866661e-01 1.67717993e-01 -4.67914492e-01 -5.54590762e-01
-6.08096719e-01 -7.59890199e-01 -8.42273533e-01 8.89921784e-02
4.11146104e-01 -7.52741694e-01 -4.14185643e-01 3.23245287e-01
-1.05584478e+00 -3.54540467e-01 -1.94934711e-01 4.45980787e-01
-3.56750101e-01 3.37577641e-01 -5.48513889e-01 -3.07477534e-01
-3.37996453e-01 -1.04134929e+00 8.41439188e-01 -1.32200019e-02
-6.42925382e-01 -9.69568789e-01 5.96477836e-02 3.19010496e-01
4.13142115e-01 7.89213404e-02 7.54085124e-01 -1.22148502e+00
6.89333212e-03 -4.09492671e-01 -1.23917289e-01 4.97534961e-01
3.95035866e-04 4.13394384e-02 -7.58552969e-01 -4.14171427e-01
-2.24518985e-01 -8.91312584e-02 7.32885122e-01 -6.82939664e-02
6.02468610e-01 -4.06242907e-02 -5.09239025e-02 6.78457856e-01
1.33243346e+00 -2.66301006e-01 6.09745622e-01 7.85988808e-01
5.23392022e-01 5.61670184e-01 6.67564332e-01 5.15990317e-01
3.62878263e-01 1.07898295e+00 1.65130004e-01 -3.39676021e-03
-1.47087917e-01 -5.62748984e-02 4.28742051e-01 1.15775168e+00
-1.96139991e-01 -1.48934335e-01 -1.13602960e+00 5.04728198e-01
-1.81430280e+00 -9.08175468e-01 -3.37245345e-01 2.44471812e+00
8.44378114e-01 3.14065367e-01 5.47252655e-01 1.43497258e-01
5.39983213e-01 2.08738506e-01 -1.93247288e-01 -7.59245634e-01
-2.00144529e-01 4.81255442e-01 8.16956103e-01 8.57765377e-01
-1.10289407e+00 1.29301250e+00 5.62949944e+00 1.09852290e+00
-1.12645924e+00 4.12391156e-01 2.11561054e-01 -3.89103889e-01
-1.33954182e-01 1.34145737e-01 -9.42170858e-01 3.18162262e-01
1.26192153e+00 -5.43725312e-01 6.06675506e-01 4.92876619e-01
1.82423145e-01 -1.72283053e-01 -9.65363443e-01 8.83554697e-01
4.03751358e-02 -1.01309216e+00 -1.29687220e-01 -4.78541628e-02
1.87564388e-01 4.71979707e-01 -2.84800857e-01 3.51904660e-01
2.36346096e-01 -8.33765149e-01 9.17022228e-01 2.10836977e-01
8.69461715e-01 -8.46364975e-01 7.50831127e-01 2.77922362e-01
-1.13417733e+00 1.63483992e-01 -1.64834619e-01 -2.07558855e-01
1.35593012e-01 4.67828393e-01 -6.25052214e-01 1.00794101e+00
7.36384630e-01 6.58378661e-01 -1.01093960e+00 1.04337180e+00
-4.57765847e-01 5.87526023e-01 -5.59030294e-01 -4.07138281e-02
3.75271030e-02 -4.23331046e-03 7.38856971e-01 1.59581304e+00
1.78694561e-01 -2.83962041e-01 1.17489241e-01 4.50805515e-01
-1.25676423e-01 7.22356081e-01 -1.65371925e-01 -2.75773015e-02
5.11936724e-01 1.42686617e+00 -7.86919832e-01 -3.31655920e-01
-3.86895210e-01 8.23490143e-01 5.41717529e-01 9.53520983e-02
-8.62190306e-01 -7.32173502e-01 5.86226404e-01 2.21278667e-01
3.55582118e-01 -1.60432786e-01 -2.30627522e-01 -1.37637794e+00
1.38113037e-01 -1.03614450e+00 6.15719497e-01 -5.17527938e-01
-1.19172966e+00 7.22110569e-01 3.69937718e-01 -9.77688134e-01
-3.66549253e-01 -7.13920653e-01 -1.82125419e-01 9.88116682e-01
-1.26388812e+00 -9.47726250e-01 -1.03596345e-01 4.88949358e-01
3.49450082e-01 -1.31268263e-01 9.15097356e-01 6.55063391e-01
-6.16074324e-01 8.97091746e-01 1.87867120e-01 9.54397023e-02
1.08345711e+00 -1.24404705e+00 8.72043908e-01 8.78240943e-01
2.42258862e-01 7.72327483e-01 7.89574146e-01 -5.15098333e-01
-1.04113722e+00 -8.07578683e-01 1.22028327e+00 -5.79634309e-01
1.18359089e+00 -7.47361541e-01 -8.04768205e-01 7.17229426e-01
3.30467194e-01 -2.46368110e-01 7.72155523e-01 4.95673627e-01
-5.89589357e-01 -1.56073915e-02 -8.00616205e-01 6.73201382e-01
1.10749960e+00 -2.81092942e-01 -5.09580314e-01 3.86173576e-01
4.57435071e-01 -5.07677913e-01 -1.01860464e+00 2.78015465e-01
6.54669762e-01 -1.03495800e+00 7.10991323e-01 -3.60700577e-01
2.28438869e-01 -4.17341769e-01 -4.42472577e-01 -1.22608662e+00
-3.04501116e-01 -6.83260500e-01 2.75417775e-01 1.51845610e+00
7.70218253e-01 -9.56287920e-01 2.48785704e-01 2.77857512e-01
6.73187384e-03 -6.92539930e-01 -9.16294098e-01 -9.66690719e-01
2.70607322e-01 -8.64321113e-01 5.97378790e-01 9.26849186e-01
1.42021328e-01 4.09907371e-01 -1.91566095e-01 -8.17040652e-02
2.65376925e-01 -2.05821663e-01 9.29830372e-01 -9.83474791e-01
-4.00807142e-01 -6.97943091e-01 -4.92031604e-01 -7.52267003e-01
-4.43140231e-02 -1.55917192e+00 -2.95177698e-01 -1.39892507e+00
2.19233230e-01 -3.19301933e-01 -3.45852971e-01 7.26698101e-01
-1.87106237e-01 4.83866006e-01 5.66522181e-01 2.51976609e-01
-5.31423450e-01 2.02720568e-01 6.15325987e-01 1.97075084e-01
-2.04372346e-01 -1.11293226e-01 -8.42626572e-01 6.04835629e-01
1.03188872e+00 -6.75534546e-01 6.65387511e-02 -4.49748337e-01
3.01388115e-01 -4.92343277e-01 1.75086439e-01 -7.93944538e-01
1.39646515e-01 1.70267135e-01 9.90501717e-02 -1.87531397e-01
1.81631178e-01 -3.75727475e-01 2.41926283e-01 2.69506991e-01
-1.90735430e-01 5.13520956e-01 5.49574077e-01 -4.82712612e-02
-1.70299202e-01 -3.40661108e-01 8.35209787e-01 -5.76471351e-03
-6.39629126e-01 -6.05368242e-02 -1.70907438e-01 2.64232278e-01
7.32364774e-01 1.92183126e-02 -4.69202161e-01 -1.26408949e-01
-6.24370456e-01 2.38612875e-01 7.00852871e-01 6.37150407e-01
8.10911134e-02 -1.21335649e+00 -8.03204417e-01 1.07166313e-01
4.03668404e-01 -4.76701766e-01 -1.23961717e-01 9.56941545e-01
-7.01328576e-01 4.22259122e-01 -1.97054997e-01 -2.60005653e-01
-1.39813900e+00 4.37611073e-01 2.45901957e-01 -4.66652334e-01
-3.75344574e-01 6.58835649e-01 -2.33761549e-01 -7.99373507e-01
3.00060716e-02 -7.98729509e-02 -1.27979904e-01 2.21593037e-01
5.08871853e-01 3.72506142e-01 4.42562640e-01 -9.14222956e-01
-7.11421967e-01 2.36989096e-01 -2.77604371e-01 -3.23344737e-01
1.48381233e+00 -9.45092961e-02 -2.73401797e-01 6.18388057e-01
1.10982239e+00 4.46234852e-01 -6.04244471e-01 -3.19245726e-01
3.63691181e-01 -3.60296488e-01 -1.08262263e-01 -1.01158834e+00
-8.12296987e-01 5.72065413e-01 4.46802825e-01 7.46256411e-02
1.08895302e+00 -6.31141942e-03 4.68296796e-01 1.61334977e-01
2.30209157e-01 -1.10611773e+00 -5.18503666e-01 6.24494255e-01
9.81167793e-01 -1.02785027e+00 3.54503334e-01 -3.94163698e-01
-5.49159169e-01 9.76957083e-01 2.66600192e-01 -7.21228942e-02
7.02109277e-01 3.71769398e-01 2.45847255e-01 -1.02545127e-01
-7.31110096e-01 -3.61992598e-01 1.95350245e-01 3.19918662e-01
8.09475899e-01 2.29161069e-01 -8.57353091e-01 6.33454740e-01
-6.26222134e-01 -1.88648641e-01 4.26485628e-01 8.67912114e-01
-1.45173416e-01 -1.78406882e+00 -2.53998160e-01 3.83794010e-01
-6.41721070e-01 -4.99283373e-01 -3.67658079e-01 1.09199953e+00
2.01537292e-02 7.76176810e-01 -1.31646484e-01 -3.41765136e-01
6.23893321e-01 2.84103692e-01 3.62926692e-01 -7.08430648e-01
-8.83436084e-01 1.84864506e-01 4.90178078e-01 -6.75189853e-01
-3.84588212e-01 -8.51404727e-01 -1.05484629e+00 -6.22796834e-01
-1.32963419e-01 8.52861851e-02 7.10220516e-01 7.66463161e-01
3.21185887e-01 2.94403523e-01 1.81374088e-01 -7.89426267e-01
-4.41638499e-01 -1.10756290e+00 -4.01608884e-01 4.81525183e-01
-1.76018685e-01 -7.12882876e-01 -4.63100940e-01 -1.59663677e-01] | [10.181941032409668, 9.808755874633789] |
0fc9956f-ee11-467c-93a9-fe773fc2f304 | aspect-extraction-from-product-reviews-using | null | null | https://aclanthology.org/E17-2107 | https://aclanthology.org/E17-2107.pdf | Aspect Extraction from Product Reviews Using Category Hierarchy Information | Aspect extraction abstracts the common properties of objects from corpora discussing them, such as reviews of products. Recent work on aspect extraction is leveraging the hierarchical relationship between products and their categories. However, such effort focuses on the aspects of child categories but ignores those from parent categories. Hence, we propose an LDA-based generative topic model inducing the two-layer categorical information (CAT-LDA), to balance the aspects of both a parent category and its child categories. Our hypothesis is that child categories inherit aspects from parent categories, controlled by the hierarchy between them. Experimental results on 5 categories of Amazon.com products show that both common aspects of parent category and the individual aspects of sub-categories can be extracted to align well with the common sense. We further evaluate the manually extracted aspects of 16 products, resulting in an average hit rate of 79.10{\%}. | ['Yinfei Yang', 'Cen Chen', 'Minghui Qiu', 'Forrest Bao'] | 2017-04-01 | null | null | null | eacl-2017-4 | ['aspect-extraction'] | ['natural-language-processing'] | [-2.19419584e-01 5.15638828e-01 -6.89327359e-01 -6.95913553e-01
-5.27233958e-01 -8.06384325e-01 8.56195867e-01 4.25600976e-01
9.83311310e-02 3.16912919e-01 6.26269639e-01 -2.51352578e-01
2.85526067e-01 -9.09015298e-01 -3.57233703e-01 -6.94665492e-01
1.26666844e-01 4.37554091e-01 3.82825106e-01 6.86724409e-02
8.05648416e-02 -6.26764074e-02 -1.65047204e+00 5.61870456e-01
5.76037049e-01 8.62643659e-01 -1.57334581e-01 4.05159183e-02
-7.77454078e-01 3.36813539e-01 -7.67320812e-01 -8.03280771e-01
1.60986669e-02 -2.78224558e-01 -5.07526338e-01 5.93650460e-01
2.45722786e-01 6.65952265e-02 1.43513754e-01 1.01372933e+00
-8.06318149e-02 -1.64640561e-01 1.12879276e+00 -1.54395938e+00
-6.93753004e-01 1.01547754e+00 -8.75135422e-01 -1.66208163e-01
2.17881799e-01 -1.76244989e-01 1.88013279e+00 -7.94776678e-01
6.11076713e-01 1.05924618e+00 4.25607234e-01 6.95346668e-02
-1.24543011e+00 -6.43597841e-01 7.07883358e-01 -1.35206506e-01
-1.25337660e+00 -3.52987722e-02 7.89160907e-01 -5.59344351e-01
1.17986321e+00 5.81898093e-02 1.03369129e+00 7.04549015e-01
2.53307641e-01 8.73571157e-01 1.00370073e+00 -1.86118215e-01
3.80905360e-01 7.86998451e-01 8.03646922e-01 -9.51461028e-03
9.54652071e-01 -4.61785525e-01 -4.87777203e-01 -3.82871389e-01
1.22639239e-01 -2.64480591e-01 1.37461439e-01 -5.87819993e-01
-6.95999444e-01 1.07650161e+00 -3.74069288e-02 4.53747779e-01
-4.59240764e-01 -2.03290880e-01 1.76709384e-01 -1.58811405e-01
6.92206740e-01 3.15462440e-01 -9.41801846e-01 4.13435735e-02
-5.96283674e-01 3.49398375e-01 1.17902672e+00 1.87440813e+00
1.00801861e+00 -3.39449733e-01 -1.11235708e-01 8.57857883e-01
5.82230330e-01 4.20748115e-01 2.42945045e-01 -5.56240320e-01
1.76894516e-01 1.19717598e+00 -5.91555163e-02 -1.01541114e+00
-2.04059467e-01 -6.47140086e-01 -3.16179335e-01 -2.75267512e-01
-3.16726714e-01 -7.00338855e-02 -9.51461971e-01 1.64341140e+00
6.44269943e-01 -4.34432596e-01 -4.16667946e-03 8.60869884e-02
1.11311221e+00 5.52052379e-01 5.04766166e-01 -2.72060364e-01
1.89500141e+00 -7.93722093e-01 -9.48815703e-01 -4.63345677e-01
5.68974972e-01 -9.77230430e-01 1.04623187e+00 1.35047466e-01
-6.39834523e-01 -3.37611109e-01 -1.07608628e+00 5.52093536e-02
-6.17602289e-01 2.45589331e-01 1.16122150e+00 5.56363404e-01
-5.69238722e-01 -5.42768426e-02 -7.39058137e-01 -3.00101191e-01
4.34910208e-01 1.65761217e-01 -2.28771314e-01 1.79851517e-01
-8.13763380e-01 4.34947312e-01 4.84948397e-01 -5.58630645e-01
-5.13446569e-01 -8.66064429e-01 -1.01214135e+00 1.91980213e-01
5.45206249e-01 -7.03916550e-01 1.19584000e+00 -6.43721104e-01
-7.75241554e-01 8.89412105e-01 -4.96678382e-01 -1.50281200e-02
-4.82819796e-01 -4.62464541e-01 -6.75689340e-01 -8.88843983e-02
4.98866171e-01 6.45058036e-01 7.91186571e-01 -1.48958206e+00
-1.20991266e+00 -4.16160226e-01 3.86957705e-01 1.70318753e-01
-5.42055726e-01 2.84428112e-02 -6.83106542e-01 -5.61781526e-01
2.16052905e-01 -1.21586192e+00 -1.08358338e-01 -6.12072349e-01
-9.44111884e-01 -6.80054486e-01 8.01177979e-01 -1.51602298e-01
1.53688669e+00 -2.19326019e+00 -2.80403048e-01 6.49095699e-02
4.39133555e-01 -1.90817207e-01 5.77185117e-02 3.56823385e-01
-8.53364095e-02 2.17067465e-01 -4.60094176e-02 -1.51485234e-01
3.61005902e-01 2.82599181e-01 -5.05433500e-01 8.62949267e-02
2.10034832e-01 6.80437684e-01 -6.28931046e-01 -5.76985538e-01
-3.27143192e-01 3.12751859e-01 -6.53725505e-01 1.70835257e-01
-4.47044969e-01 -3.50652277e-01 -6.32867217e-01 7.04805195e-01
8.10941637e-01 -2.31130272e-01 2.53172487e-01 -5.17030418e-01
7.40804449e-02 1.05891013e+00 -1.07757843e+00 1.13163495e+00
-4.48976666e-01 3.83842647e-01 -2.20487922e-01 -3.34799260e-01
9.27419066e-01 1.78172588e-01 5.13816118e-01 3.32628191e-02
1.03534825e-01 -1.58012062e-01 1.40824258e-01 -1.74423113e-01
6.96889699e-01 -3.03223222e-01 -4.62417185e-01 7.24406362e-01
3.93185526e-01 -6.63598716e-01 6.61629260e-01 5.03799617e-01
6.58298969e-01 1.97984785e-01 7.89293706e-01 -5.14349759e-01
1.19238801e-01 1.48505628e-01 5.99493086e-01 4.55942392e-01
1.25416875e-01 4.91510957e-01 1.01157641e+00 -2.59989440e-01
-9.30078030e-01 -1.05268705e+00 -1.96208999e-01 1.05038536e+00
9.30482596e-02 -1.26822722e+00 -5.01955509e-01 -1.12155282e+00
-7.20179901e-02 1.24950850e+00 -8.21922362e-01 -7.47336000e-02
-2.34526604e-01 -9.70033467e-01 -1.28731757e-01 5.07967472e-01
1.69721410e-01 -7.83578575e-01 -4.16992664e-01 1.47206202e-01
-9.87840518e-02 -1.40820432e+00 -3.78081232e-01 2.22124159e-01
-4.20472234e-01 -1.09932792e+00 -1.18775904e-01 -5.73982954e-01
5.31048834e-01 2.59325176e-01 1.76211989e+00 -2.54501551e-01
2.55054176e-01 2.40584552e-01 -4.29098248e-01 -7.18162000e-01
-3.03037405e-01 3.77184540e-01 -1.60393238e-01 -1.19962908e-01
1.38288558e+00 -6.92539573e-01 -2.27736741e-01 4.98958409e-01
-8.18862200e-01 -1.13035075e-01 6.18924320e-01 2.60647327e-01
7.51189768e-01 4.97902274e-01 1.01889826e-01 -1.33256483e+00
3.48557860e-01 -6.43120170e-01 -3.42623144e-01 6.42008483e-02
-8.19512844e-01 -1.68933347e-01 -4.19627726e-02 -6.64925158e-01
-1.16421568e+00 -1.41872928e-01 7.48469085e-02 2.89653987e-01
-3.19196701e-01 6.59998298e-01 -7.40807533e-01 9.71910477e-01
2.39431381e-01 -5.31076677e-02 -5.96081793e-01 -3.47347766e-01
5.35975158e-01 6.85593247e-01 -6.39547184e-02 -5.35158277e-01
6.54441237e-01 4.18146998e-01 -2.29684815e-01 -8.52143526e-01
-1.46971178e+00 -7.63948619e-01 -5.61785281e-01 2.40129575e-01
6.92900777e-01 -1.18920004e+00 -5.78043424e-02 1.74474776e-01
-1.13793683e+00 3.97929072e-01 -5.68452656e-01 4.35753763e-01
-1.94542021e-01 -6.68507069e-03 -5.25678515e-01 -4.03220862e-01
-1.05162457e-01 -1.10499132e+00 1.30528212e+00 3.46224010e-01
-9.13147807e-01 -8.70051682e-01 9.87653658e-02 1.73257682e-02
1.17555834e-01 -5.94696738e-02 1.32542241e+00 -1.29173076e+00
-5.06204784e-01 -3.83180648e-01 7.35482387e-03 2.41145834e-01
6.83813512e-01 1.27059609e-01 -9.67939138e-01 6.96349423e-03
2.35396653e-01 2.89544523e-01 6.25220180e-01 3.58350247e-01
4.35840130e-01 -5.41720808e-01 -7.82472074e-01 1.03185721e-01
1.26834583e+00 2.91009933e-01 4.13682312e-01 1.60035074e-01
8.49361181e-01 9.88389611e-01 7.73228526e-01 2.55867720e-01
6.10473394e-01 7.85460413e-01 -5.17647602e-02 8.37175176e-02
-1.72516957e-01 -3.89692843e-01 3.75199318e-01 1.09593487e+00
4.07706946e-01 -1.82734132e-01 -6.21386528e-01 8.78981829e-01
-1.20408726e+00 -5.85743010e-01 -4.23674196e-01 1.90513563e+00
1.00936353e+00 5.25188684e-01 1.90740690e-01 -1.87712833e-01
6.29842520e-01 2.44212553e-01 -3.42120856e-01 -3.37023586e-01
-1.50638163e-01 -1.80028856e-01 -7.01309294e-02 2.63291478e-01
-1.31134415e+00 9.24799025e-01 6.37177563e+00 9.07756090e-01
-3.49761695e-01 3.18381488e-02 5.27400315e-01 9.87994447e-02
-9.67192590e-01 4.99173850e-01 -1.62344503e+00 2.79794753e-01
6.39425874e-01 -4.85238850e-01 -6.09166920e-01 1.37352252e+00
-2.60781497e-01 -1.48045883e-01 -1.16278422e+00 2.56956100e-01
3.02432477e-01 -6.93259060e-01 5.15402555e-01 4.55710262e-01
1.08310437e+00 -3.42028886e-01 7.05103111e-03 5.13220549e-01
6.41975880e-01 -5.92455685e-01 6.90693557e-01 -1.37904882e-01
4.26379889e-01 -7.87045062e-01 9.49926674e-01 -5.78396730e-02
-1.39464903e+00 3.79337877e-01 -4.85513568e-01 1.86586529e-01
1.87089622e-01 1.25682020e+00 -7.60792315e-01 4.65885371e-01
7.66123712e-01 9.36674297e-01 -5.92830420e-01 6.41692102e-01
-7.70562947e-01 7.28645742e-01 -2.12451681e-01 -1.28339052e-01
1.01322755e-01 -4.04207289e-01 7.29196966e-01 1.25754774e+00
1.52970985e-01 -3.85345705e-02 1.50221810e-01 9.82418418e-01
1.05628155e-01 2.58686453e-01 -7.51211643e-01 -3.75489950e-01
3.62410814e-01 1.42409718e+00 -9.70210195e-01 -7.17690110e-01
-8.34070444e-01 2.06815675e-01 -6.26649708e-02 2.54120171e-01
-7.86604702e-01 -5.35644948e-01 1.09988785e+00 1.09674014e-01
9.60011840e-01 -9.63597894e-02 -6.05727196e-01 -1.27732062e+00
2.75171220e-01 -7.91628718e-01 4.95574743e-01 -5.06963491e-01
-1.34761202e+00 1.02495432e+00 4.11441594e-01 -1.36231411e+00
-2.83605307e-01 -3.50760728e-01 -4.75655019e-01 5.49760342e-01
-1.25244999e+00 -1.46185780e+00 6.89364821e-02 2.44191736e-02
6.35348201e-01 -8.71237665e-02 8.17594707e-01 -5.12470081e-02
-1.87670052e-01 4.19487506e-01 -5.97046435e-01 -1.25638202e-01
4.22425598e-01 -1.43673551e+00 4.43181753e-01 7.20443189e-01
4.88765627e-01 1.19935596e+00 1.06940675e+00 -8.63721192e-01
-6.85792625e-01 -9.60216761e-01 1.49934101e+00 -7.93241084e-01
8.92949879e-01 -5.87835073e-01 -8.89026105e-01 9.40661550e-01
6.52211845e-01 -4.93001789e-01 1.33415926e+00 8.89311910e-01
-9.11807477e-01 2.68123252e-03 -6.92270517e-01 6.01245403e-01
7.27416158e-01 -4.32443947e-01 -9.82474923e-01 1.16082609e-01
1.19620848e+00 1.10732831e-01 -8.33862126e-01 4.73995388e-01
4.40539598e-01 -7.99354374e-01 6.61823213e-01 -6.73946440e-01
6.25411749e-01 -5.17181337e-01 -3.15871805e-01 -1.31703198e+00
-1.83542386e-01 -1.89380497e-01 -2.33895317e-01 2.07818174e+00
9.23920035e-01 -3.33532363e-01 6.32078230e-01 7.76344717e-01
4.80506709e-03 -7.66519845e-01 -4.32016194e-01 -5.93314171e-01
6.46024495e-02 -4.89800185e-01 9.52963173e-01 7.06110239e-01
1.61945537e-01 1.05465317e+00 1.51135474e-01 7.22463876e-02
3.77898991e-01 7.91689336e-01 6.63840890e-01 -1.15807819e+00
-4.62970614e-01 -2.92454481e-01 -4.24581349e-01 -9.32540059e-01
1.64118707e-01 -5.18284559e-01 -3.39149199e-02 -1.60023177e+00
7.44644344e-01 -4.46314275e-01 -8.11168924e-02 6.78210497e-01
-4.03210372e-01 3.10763344e-02 -1.16912276e-01 1.41750097e-01
-6.35196090e-01 6.55589223e-01 9.72120106e-01 -3.78506690e-01
-3.40499103e-01 4.17482436e-01 -1.43084514e+00 1.07152295e+00
7.20327973e-01 -7.40672231e-01 -5.31863034e-01 4.49565612e-02
4.28986400e-01 -8.68019044e-01 -4.38132256e-01 -4.56828684e-01
-1.25842303e-01 -6.24970421e-02 -3.10814194e-02 -8.35396051e-01
2.23181665e-01 -9.13502991e-01 1.38554007e-01 4.99029271e-02
-4.05038029e-01 -6.89518675e-02 6.23961873e-02 4.29894269e-01
-4.84812528e-01 -1.33957043e-01 3.52313608e-01 -1.08578980e-01
-3.86556894e-01 1.83360964e-01 -3.96702379e-01 1.35950640e-01
9.17482615e-01 1.88416451e-01 -2.49303803e-01 -4.59213436e-01
-7.89748907e-01 -5.80773614e-02 4.50833470e-01 6.32217050e-01
1.50105029e-01 -1.34533322e+00 -5.07119060e-01 8.46860781e-02
6.86754227e-01 -6.08016439e-02 -1.84646741e-01 6.36974335e-01
2.78031588e-01 5.52021623e-01 2.68801659e-01 -5.15973270e-01
-1.25245810e+00 8.16687226e-01 -3.13058883e-01 -4.26211119e-01
-4.43557918e-01 5.88116288e-01 8.24201584e-01 -3.73286813e-01
2.80951224e-02 -4.58222330e-01 -6.02077365e-01 5.13985276e-01
5.82480967e-01 -9.39605162e-02 7.59928823e-02 -7.45091915e-01
-3.97890329e-01 6.30638063e-01 -4.77584928e-01 -1.19480155e-01
1.41742933e+00 -2.48982102e-01 -4.05929148e-01 7.07330704e-01
1.29884291e+00 4.02356446e-01 -8.86735320e-01 -2.33931854e-01
3.13493490e-01 -3.02012533e-01 -2.56627709e-01 -6.05513811e-01
-1.11991978e+00 4.89621073e-01 -2.13485956e-01 5.58349550e-01
9.93861914e-01 7.44958520e-01 5.17159939e-01 -5.58194146e-02
4.27481145e-01 -7.10337639e-01 -2.47888207e-01 3.62444490e-01
6.62366688e-01 -8.29587877e-01 4.52984959e-01 -1.32266176e+00
-8.93852293e-01 5.67517281e-01 6.54859841e-01 -1.26360878e-02
1.11092126e+00 4.77066755e-01 1.38849527e-01 -6.16895497e-01
-1.11977673e+00 -3.40710372e-01 6.16639376e-01 5.50053656e-01
7.77868569e-01 3.27620387e-01 -6.15503788e-01 1.08164036e+00
-5.06084800e-01 -5.71014047e-01 5.42831838e-01 9.07977998e-01
-2.45076001e-01 -1.17993796e+00 3.18037897e-01 6.58349514e-01
-8.27208638e-01 -3.92940342e-01 -5.96964180e-01 9.91056979e-01
2.95112967e-01 1.07481170e+00 1.09143399e-01 -3.09264600e-01
3.72977018e-01 -3.93678620e-03 -2.37193517e-02 -1.08367121e+00
-3.93214524e-01 5.67922592e-01 4.41235244e-01 -3.39286774e-01
-6.26259983e-01 -9.03590262e-01 -7.34446466e-01 2.40849376e-01
-5.62530100e-01 5.74084520e-01 6.78170562e-01 9.89587486e-01
4.65053350e-01 4.04171199e-01 5.76138437e-01 -9.38885733e-02
-1.57041803e-01 -1.12939441e+00 -9.01074350e-01 3.75318885e-01
-9.51835420e-03 -7.84617305e-01 -6.31180644e-01 3.24956238e-01] | [11.35513973236084, 6.695321559906006] |
e26b7b43-4057-49e9-a70c-ff6aad6da0d5 | sepico-semantic-guided-pixel-contrast-for | 2204.08808 | null | https://arxiv.org/abs/2204.08808v2 | https://arxiv.org/pdf/2204.08808v2.pdf | SePiCo: Semantic-Guided Pixel Contrast for Domain Adaptive Semantic Segmentation | Domain adaptive semantic segmentation attempts to make satisfactory dense predictions on an unlabeled target domain by utilizing the supervised model trained on a labeled source domain. In this work, we propose Semantic-Guided Pixel Contrast (SePiCo), a novel one-stage adaptation framework that highlights the semantic concepts of individual pixels to promote learning of class-discriminative and class-balanced pixel representations across domains, eventually boosting the performance of self-training methods. Specifically, to explore proper semantic concepts, we first investigate a centroid-aware pixel contrast that employs the category centroids of the entire source domain or a single source image to guide the learning of discriminative features. Considering the possible lack of category diversity in semantic concepts, we then blaze a trail of distributional perspective to involve a sufficient quantity of instances, namely distribution-aware pixel contrast, in which we approximate the true distribution of each semantic category from the statistics of labeled source data. Moreover, such an optimization objective can derive a closed-form upper bound by implicitly involving an infinite number of (dis)similar pairs, making it computationally efficient. Extensive experiments show that SePiCo not only helps stabilize training but also yields discriminative representations, making significant progress on both synthetic-to-real and daytime-to-nighttime adaptation scenarios. | ['Guoren Wang', 'Gao Huang', 'Chi Harold Liu', 'Mingjia Li', 'Shuang Li', 'Binhui Xie'] | 2022-04-19 | null | null | null | null | ['synthetic-to-real-translation'] | ['computer-vision'] | [ 4.50609267e-01 5.54439239e-03 -3.64467174e-01 -6.97498441e-01
-8.08024526e-01 -6.27570808e-01 5.91931164e-01 8.19438249e-02
-2.30973125e-01 7.54632831e-01 -5.23843020e-02 -5.28822318e-02
-8.29564333e-02 -7.17930734e-01 -6.37522340e-01 -9.56285059e-01
2.76442915e-01 4.06727582e-01 3.24124277e-01 -1.25296356e-03
1.40953764e-01 3.11811537e-01 -1.55196953e+00 1.62132546e-01
1.19732642e+00 1.19772613e+00 4.84666765e-01 8.12832490e-02
-3.85357589e-01 4.22176957e-01 -5.73844850e-01 -1.81791842e-01
4.39640492e-01 -6.20799363e-01 -5.51038921e-01 5.46518326e-01
4.97825921e-01 4.48572114e-02 1.53992847e-01 1.27026713e+00
2.86831111e-01 3.27544302e-01 9.14767087e-01 -1.20302594e+00
-6.33400023e-01 3.45751882e-01 -7.60856748e-01 1.67883605e-01
-1.18982926e-01 1.72984138e-01 1.00212908e+00 -7.61936069e-01
3.96133244e-01 9.98685956e-01 6.06491685e-01 2.33840346e-01
-1.09630561e+00 -6.72495484e-01 5.31048357e-01 2.05992445e-01
-1.51146114e+00 -1.65789649e-01 1.13782215e+00 -1.83895856e-01
4.26130235e-01 4.63985279e-02 5.22040904e-01 8.99794459e-01
-4.32950079e-01 8.47001076e-01 1.24013078e+00 -5.14915764e-01
4.58570421e-01 5.30454457e-01 1.74531043e-02 3.62377673e-01
1.17349848e-01 -1.36074170e-01 -3.91820520e-01 7.71216899e-02
5.47734022e-01 -4.30808403e-02 -1.81113645e-01 -7.52502084e-01
-9.46891367e-01 9.01308954e-01 6.36431336e-01 2.20142439e-01
-2.59331912e-01 -2.80369580e-01 1.73001215e-01 -2.65886225e-02
7.35963702e-01 2.54687101e-01 -5.64256847e-01 2.39528835e-01
-1.09745932e+00 1.57493532e-01 4.11914051e-01 9.56480980e-01
1.21941817e+00 4.45557982e-02 -1.56387448e-01 1.19637489e+00
1.83943793e-01 5.36057115e-01 5.13928831e-01 -9.07750249e-01
4.38829899e-01 6.96693003e-01 2.36152764e-02 -7.44113028e-01
-2.36244604e-01 -7.95480013e-01 -5.22369385e-01 -7.30368346e-02
6.09511256e-01 -7.80919716e-02 -1.08137798e+00 1.95787311e+00
5.55706561e-01 2.16748744e-01 1.07379459e-01 9.29000854e-01
2.43281305e-01 5.49011827e-01 5.19581318e-01 -1.81274965e-01
1.00356603e+00 -1.02922976e+00 -2.40134850e-01 -5.42262614e-01
4.78497922e-01 -3.66983056e-01 1.24433482e+00 7.49234930e-02
-6.40922725e-01 -7.66828120e-01 -9.30782974e-01 1.89760730e-01
-6.03077948e-01 1.03693575e-01 5.81614017e-01 7.17449069e-01
-7.49828517e-01 5.38361311e-01 -5.28247952e-01 -4.16750789e-01
6.74440265e-01 -1.50006525e-02 1.55256866e-02 -3.07783514e-01
-9.99443412e-01 6.10332787e-01 7.44747043e-01 -2.56749630e-01
-8.90806735e-01 -8.05645585e-01 -8.65831435e-01 3.19772065e-02
3.64912599e-01 -4.79578823e-01 9.13524747e-01 -1.61609006e+00
-1.10485423e+00 9.27444696e-01 -2.41645306e-01 -3.19949597e-01
3.21752816e-01 4.93669510e-02 -4.63257194e-01 3.47990125e-01
4.19652969e-01 1.07838309e+00 9.66585100e-01 -1.66666281e+00
-8.39807272e-01 -4.78557259e-01 -1.15818426e-01 5.80894232e-01
-5.37209630e-01 -3.97653013e-01 -4.44652319e-01 -6.88005924e-01
2.95284003e-01 -8.04249704e-01 -2.57122397e-01 1.25909612e-01
-3.94844651e-01 -2.44209900e-01 5.47093153e-01 -4.61249560e-01
8.38537633e-01 -2.41444445e+00 -2.00871304e-01 3.47880423e-01
-2.15205625e-01 1.09850295e-01 -2.35120192e-01 -5.19111343e-02
-2.46088523e-02 -2.65151322e-01 -7.00188816e-01 -2.30591297e-01
-1.53108612e-01 3.61900210e-01 -2.60064602e-01 3.91351104e-01
5.02326787e-01 7.81271160e-01 -1.00242972e+00 -5.87075949e-01
2.88951725e-01 1.90535694e-01 -5.66132545e-01 1.18130885e-01
-4.38696325e-01 5.06477237e-01 -5.87388515e-01 8.29027236e-01
9.43036020e-01 -2.01241329e-01 2.00984344e-01 -1.16859145e-01
1.43889770e-01 9.92743745e-02 -1.09642828e+00 1.79701662e+00
-4.36172247e-01 3.72691631e-01 -1.48081020e-01 -1.42436254e+00
1.10339475e+00 -2.93609321e-01 5.56396127e-01 -1.02261472e+00
2.44098380e-02 2.54204273e-01 -3.64202023e-01 -2.00806007e-01
4.28542554e-01 -3.85461420e-01 -1.12140246e-01 2.03723758e-01
2.08275080e-01 -1.68489248e-01 7.24578351e-02 -1.62822872e-01
5.15095890e-01 3.69837165e-01 2.27824822e-01 -4.00060058e-01
4.00684178e-01 3.59224170e-01 4.90643919e-01 7.16176689e-01
-3.43198627e-01 8.36413741e-01 1.51016876e-01 -4.51797480e-03
-9.46154654e-01 -1.38767648e+00 -4.18673933e-01 1.24403906e+00
5.06810546e-01 2.42945179e-01 -9.74111438e-01 -8.66052389e-01
2.27665398e-02 1.04040790e+00 -5.71267903e-01 -3.20710510e-01
-2.25612670e-01 -9.71554816e-01 1.20046645e-01 6.47626996e-01
7.01721489e-01 -9.36214507e-01 -4.32981759e-01 9.66141522e-02
-1.02369703e-01 -1.05552018e+00 -2.86657304e-01 6.41517103e-01
-8.43159437e-01 -1.06141746e+00 -9.75526690e-01 -8.26729357e-01
8.55832994e-01 5.54872870e-01 1.03927207e+00 -3.87979746e-01
-2.11074725e-02 3.07209849e-01 -4.59744245e-01 -2.85935491e-01
-4.88407202e-02 8.48312397e-04 -2.06891969e-01 1.51572362e-01
6.64765775e-01 -6.10504508e-01 -5.83214164e-01 3.14652234e-01
-8.33662212e-01 -6.54566810e-02 5.39191663e-01 8.21958482e-01
8.56703281e-01 2.21503019e-01 6.39683962e-01 -9.40777361e-01
1.87002540e-01 -7.18117237e-01 -4.23851371e-01 3.97145271e-01
-6.14416301e-01 -2.31042933e-02 6.76945448e-01 -5.19065142e-01
-1.37115133e+00 2.74789006e-01 3.03785242e-02 -4.94048893e-01
-3.47084463e-01 2.91994661e-01 -4.25193071e-01 1.64218754e-01
8.13806653e-01 6.29501104e-01 -1.77219421e-01 -3.64236176e-01
7.21333921e-01 6.51420176e-01 7.72682309e-01 -7.28045046e-01
9.05902803e-01 6.06751442e-01 -3.42370361e-01 -7.45436728e-01
-1.25341475e+00 -8.38344753e-01 -7.54381120e-01 3.27325496e-03
7.27946281e-01 -1.14433384e+00 1.08556733e-01 3.66560280e-01
-6.69433653e-01 -5.12785673e-01 -5.95860720e-01 2.54402161e-01
-4.99997020e-01 3.31625849e-01 2.09935736e-02 -7.38371849e-01
2.53274500e-01 -7.73425579e-01 1.20900965e+00 4.99396741e-01
-5.66314673e-04 -1.16291916e+00 -7.60635063e-02 3.27674985e-01
1.53634399e-01 1.16288900e-01 8.20290804e-01 -8.72514129e-01
-3.94893080e-01 7.27579519e-02 -6.54280007e-01 7.51160383e-01
2.67337501e-01 -3.99830580e-01 -1.24596119e+00 -1.24934450e-01
7.78866410e-02 -3.06353927e-01 9.57311153e-01 3.59649956e-01
1.48635888e+00 -1.80649552e-02 -3.28591973e-01 7.09084213e-01
1.61191654e+00 1.70394376e-01 4.01240349e-01 3.30883443e-01
7.50963509e-01 8.70702386e-01 8.69280040e-01 5.22535741e-01
3.57853889e-01 5.19377112e-01 2.73435563e-01 -2.88909614e-01
-2.87020385e-01 -4.62933838e-01 1.39200121e-01 2.41283700e-01
3.50926131e-01 -1.52021423e-01 -6.87462986e-01 9.28373516e-01
-1.51389921e+00 -6.05966389e-01 1.15061752e-01 2.21568346e+00
8.54828835e-01 8.92361179e-02 1.71915054e-01 3.86842974e-02
8.79588962e-01 1.38125122e-01 -9.09180224e-01 -8.58712122e-02
-2.72361279e-01 2.12585613e-01 7.18030870e-01 1.42059058e-01
-1.26746714e+00 9.91041124e-01 6.07047319e+00 1.31908035e+00
-1.07302022e+00 2.31829718e-01 8.75461817e-01 1.60603821e-01
-5.79870760e-01 -2.48042420e-02 -7.63073623e-01 6.78970516e-01
7.54367173e-01 -6.76060095e-02 2.64339626e-01 1.17095244e+00
1.46658003e-01 -3.58747363e-01 -7.87782848e-01 8.11223805e-01
4.81168851e-02 -9.09094155e-01 1.77488133e-01 9.04406328e-03
1.01873791e+00 -8.84613469e-02 1.14159547e-01 2.88323373e-01
3.03136110e-01 -6.43085182e-01 7.79300630e-01 2.12276369e-01
7.68035352e-01 -7.77333081e-01 4.68503356e-01 3.40866923e-01
-1.09284568e+00 -2.74874091e-01 -6.26078129e-01 1.52812511e-01
-6.81774318e-02 7.21422851e-01 -7.22524643e-01 5.99591970e-01
6.14556253e-01 8.29663575e-01 -7.14814961e-01 9.57732201e-01
-1.89424410e-01 6.99794829e-01 -3.41085047e-01 2.34587878e-01
5.30147433e-01 -3.36361080e-01 1.86202765e-01 1.19842529e+00
3.55225474e-01 -2.82913097e-03 2.86147326e-01 9.62750256e-01
9.80147868e-02 1.09791212e-01 -3.91689211e-01 5.79284914e-02
5.33693373e-01 1.15875363e+00 -9.81833518e-01 -3.35544080e-01
-4.69159842e-01 1.15584970e+00 2.66178012e-01 5.46161592e-01
-8.91658306e-01 -1.03222199e-01 6.67860925e-01 7.43771046e-02
4.70239013e-01 3.18662524e-02 -7.32938349e-01 -8.68356407e-01
3.84306498e-02 -4.96360600e-01 4.42694455e-01 -8.08407605e-01
-1.50861275e+00 3.48910838e-01 1.96908519e-01 -1.45103514e+00
-1.34634092e-01 -5.23412585e-01 -6.04530156e-01 8.24738920e-01
-2.05640244e+00 -1.19417167e+00 -5.28692126e-01 7.53410935e-01
6.08086169e-01 -1.23163968e-01 5.94730973e-01 2.53464997e-01
-5.11390150e-01 6.11778259e-01 4.73600715e-01 -1.14507012e-01
7.83296049e-01 -1.32132614e+00 -4.24375199e-02 7.92721510e-01
2.27733970e-01 3.19143772e-01 5.62103391e-01 -3.91704142e-01
-5.20815253e-01 -1.39962399e+00 5.12570798e-01 -2.84665942e-01
5.43143630e-01 -7.19474629e-02 -1.06604552e+00 3.48643064e-01
-2.30560943e-01 9.81663167e-02 6.32794201e-01 3.90671827e-02
-3.15211862e-01 -2.93906868e-01 -1.17597222e+00 4.23663974e-01
1.12914634e+00 -5.50110579e-01 -5.47550023e-01 4.16813403e-01
6.39589965e-01 -9.71131772e-02 -3.92029017e-01 3.93490762e-01
9.32633802e-02 -1.02386880e+00 9.69088733e-01 -2.01270476e-01
4.13202673e-01 -3.84678066e-01 -3.32594365e-01 -1.39809060e+00
-1.51601225e-01 1.73779517e-01 2.00174689e-01 1.42370415e+00
3.52145761e-01 -4.41978693e-01 9.96606410e-01 4.01268065e-01
-3.93146068e-01 -5.49110055e-01 -7.14692235e-01 -9.24944818e-01
1.51680544e-01 -5.36731541e-01 4.36667174e-01 1.01982796e+00
-4.29810375e-01 -1.36681674e-02 2.46912353e-02 2.59885609e-01
6.23459697e-01 2.95788258e-01 4.91414249e-01 -1.02462864e+00
-3.15317065e-01 -3.24670762e-01 -2.27586851e-01 -1.25524974e+00
3.67108881e-01 -8.86370838e-01 3.30634147e-01 -1.31744659e+00
3.80440623e-01 -8.53043675e-01 -6.51172340e-01 4.43883717e-01
-4.47803110e-01 3.93196642e-01 -7.30130496e-03 3.10993224e-01
-7.70317018e-01 7.26093888e-01 1.24838471e+00 -2.94119537e-01
-1.19583942e-01 1.30969390e-01 -1.09216952e+00 6.25444829e-01
7.28517234e-01 -4.59695369e-01 -7.52746642e-01 -1.32585987e-01
-3.84709656e-01 -5.15839577e-01 5.13784170e-01 -1.10810840e+00
-6.57984689e-02 -3.81772429e-01 6.06068492e-01 -4.54913139e-01
2.74563789e-01 -8.69103014e-01 -2.08928376e-01 -1.79523509e-02
-3.26479644e-01 -7.60368764e-01 1.30751804e-01 6.75601184e-01
-4.33808476e-01 -4.59488124e-01 1.04540133e+00 -2.34051019e-01
-1.29806423e+00 1.16192520e-01 6.49952739e-02 2.70414054e-01
1.10480464e+00 -6.68707430e-01 -8.34323873e-04 -1.03457011e-01
-5.03871202e-01 2.09511176e-01 6.14471138e-01 2.56945848e-01
3.06208372e-01 -1.20877683e+00 -3.89814109e-01 1.85267791e-01
4.75028723e-01 2.24678978e-01 3.95914674e-01 4.78602946e-01
-1.38867438e-01 2.76565909e-01 -2.17932299e-01 -8.09986591e-01
-6.82442069e-01 5.62621236e-01 3.54779035e-01 7.03159953e-03
-3.71951789e-01 9.21987712e-01 7.96743631e-01 -4.80574459e-01
6.28945753e-02 -2.32340842e-02 -1.88939050e-01 2.96458662e-01
3.14551204e-01 7.41364658e-02 -8.89515951e-02 -6.02933943e-01
-3.07685912e-01 6.08582258e-01 1.48029849e-01 1.54179409e-01
1.17472827e+00 -4.77568388e-01 2.51209408e-01 3.00736010e-01
1.25353062e+00 -1.38976812e-01 -2.01411700e+00 -5.12599707e-01
5.04375175e-02 -5.37791252e-01 4.03691754e-02 -9.97832656e-01
-1.13657939e+00 7.84890652e-01 8.33279312e-01 -2.47771256e-02
1.49288547e+00 1.60716444e-01 6.50903225e-01 5.83462790e-02
2.76102930e-01 -1.34316671e+00 1.79791212e-01 2.78280288e-01
3.30587447e-01 -1.51271069e+00 -1.86339810e-01 -6.03993654e-01
-9.84723508e-01 8.75523210e-01 6.36738658e-01 -2.94291470e-02
4.86271411e-01 -2.09660232e-01 1.71634182e-01 3.08584794e-03
-5.60037009e-02 -6.91381454e-01 4.49299932e-01 1.07619381e+00
1.39734834e-01 2.00695738e-01 2.37854999e-02 5.90479016e-01
6.15853854e-02 -1.86934337e-01 6.06543105e-03 5.85309029e-01
-6.70038998e-01 -1.05069244e+00 -2.55881637e-01 3.48721683e-01
-7.88356513e-02 -2.59831846e-01 -2.45474085e-01 6.98104799e-01
4.70087349e-01 8.35120022e-01 3.99943233e-01 -1.25564128e-01
2.55550802e-01 2.46281445e-01 3.26158166e-01 -7.21410632e-01
-2.19223797e-02 8.14512223e-02 -1.61465079e-01 -3.87075871e-01
-5.62984645e-01 -7.61187851e-01 -1.22188652e+00 7.79903457e-02
-1.43603697e-01 2.87350062e-02 6.27130389e-01 1.11576581e+00
1.20157257e-01 4.13481802e-01 8.81949365e-01 -8.32966924e-01
-5.14187634e-01 -8.76678348e-01 -9.16856349e-01 6.18458390e-01
3.10442038e-02 -9.56189215e-01 -4.17571217e-01 1.12333715e-01] | [9.67320442199707, 1.3594319820404053] |
1739466f-cbc6-45d7-81ff-27435dd4e36c | weakly-supervised-video-salient-object-1 | 2207.07269 | null | https://arxiv.org/abs/2207.07269v1 | https://arxiv.org/pdf/2207.07269v1.pdf | Weakly Supervised Video Salient Object Detection via Point Supervision | Video salient object detection models trained on pixel-wise dense annotation have achieved excellent performance, yet obtaining pixel-by-pixel annotated datasets is laborious. Several works attempt to use scribble annotations to mitigate this problem, but point supervision as a more labor-saving annotation method (even the most labor-saving method among manual annotation methods for dense prediction), has not been explored. In this paper, we propose a strong baseline model based on point supervision. To infer saliency maps with temporal information, we mine inter-frame complementary information from short-term and long-term perspectives, respectively. Specifically, we propose a hybrid token attention module, which mixes optical flow and image information from orthogonal directions, adaptively highlighting critical optical flow information (channel dimension) and critical token information (spatial dimension). To exploit long-term cues, we develop the Long-term Cross-Frame Attention module (LCFA), which assists the current frame in inferring salient objects based on multi-frame tokens. Furthermore, we label two point-supervised datasets, P-DAVIS and P-DAVSOD, by relabeling the DAVIS and the DAVSOD dataset. Experiments on the six benchmark datasets illustrate our method outperforms the previous state-of-the-art weakly supervised methods and even is comparable with some fully supervised approaches. Source code and datasets are available. | ['Wenqiang Zhang', 'Qianyu Guo', 'Yan Wang', 'Wei zhang', 'Haozhe Xing', 'Shuyong Gao'] | 2022-07-15 | null | null | null | null | ['video-salient-object-detection'] | ['computer-vision'] | [ 2.83668548e-01 3.69956270e-02 -5.41129589e-01 -3.30316246e-01
-6.60366952e-01 -2.14102551e-01 4.04125333e-01 5.93686737e-02
-4.42219853e-01 9.26301718e-01 2.49231979e-01 6.07851967e-02
2.82192051e-01 -4.18553352e-01 -8.33014965e-01 -7.39192545e-01
5.42593971e-02 -1.45527780e-01 8.69482517e-01 -2.00870708e-02
4.69272643e-01 1.03708334e-01 -1.61003423e+00 3.69510740e-01
7.96489060e-01 1.11357927e+00 6.41765118e-01 4.13166672e-01
-1.36705518e-01 1.24191308e+00 -1.04651213e-01 -2.02809036e-01
2.59244740e-01 -3.46002251e-01 -9.04262364e-01 2.51536995e-01
6.75044894e-01 -5.38483799e-01 -1.89761057e-01 9.80590463e-01
3.13715935e-01 1.79692343e-01 2.41522223e-01 -1.31455219e+00
-6.26470089e-01 5.07182300e-01 -1.06843030e+00 6.24443531e-01
1.70937806e-01 2.55600870e-01 1.08674443e+00 -1.19266582e+00
6.55658066e-01 1.08889031e+00 4.68870997e-01 4.05846387e-01
-1.03709054e+00 -5.28945565e-01 4.74560916e-01 4.34546918e-01
-1.18108213e+00 -3.73362422e-01 1.13632274e+00 -4.56737518e-01
7.43311226e-01 -8.77236351e-02 7.49609530e-01 1.01220059e+00
5.32976612e-02 1.28831768e+00 1.00048220e+00 -3.02164763e-01
4.95230928e-02 9.53413993e-02 -2.74645921e-04 9.48967457e-01
-2.26486847e-02 2.83020735e-01 -9.21509147e-01 2.22248346e-01
9.66685593e-01 2.42630988e-01 -3.64515215e-01 -5.37178993e-01
-1.53327465e+00 5.96471131e-01 6.73566699e-01 1.77004129e-01
-3.42967212e-01 1.91885263e-01 4.02215391e-01 -2.71561384e-01
7.74573863e-01 2.00728312e-01 -5.15015125e-01 -1.59609318e-01
-1.17035174e+00 1.74513664e-02 1.92464471e-01 1.07735384e+00
1.14475381e+00 2.14709155e-02 -5.03972173e-01 6.18012071e-01
3.67944270e-01 1.96690068e-01 4.04604197e-01 -9.89203513e-01
3.83893788e-01 5.89007080e-01 3.21048141e-01 -9.72512424e-01
-3.46985787e-01 -4.58801061e-01 -5.82021236e-01 1.49506226e-01
3.54115158e-01 -6.18908070e-02 -8.06741238e-01 1.47497010e+00
3.98658812e-01 7.01549470e-01 -1.00869119e-01 1.46890211e+00
8.10179234e-01 4.99511957e-01 3.36224645e-01 -3.12811524e-01
1.12745464e+00 -1.64425933e+00 -6.78784072e-01 -2.62516052e-01
5.53894103e-01 -7.35682428e-01 1.15571010e+00 7.93354288e-02
-1.19635856e+00 -7.76342750e-01 -8.51701498e-01 -4.66099054e-01
-2.28976503e-01 3.21835846e-01 7.08809257e-01 1.51745856e-01
-9.49256003e-01 2.71809310e-01 -9.53944683e-01 -2.14031622e-01
7.81119287e-01 4.73603830e-02 -2.35197231e-01 -3.59736383e-03
-1.14469218e+00 6.67822421e-01 4.43057686e-01 1.42800108e-01
-1.12918413e+00 -7.72351921e-01 -1.06490684e+00 5.02215326e-02
4.94134575e-01 -5.45269310e-01 9.19193029e-01 -1.16564071e+00
-1.32615244e+00 7.53304899e-01 -5.09218574e-01 -5.57716310e-01
4.89978969e-01 -3.85957539e-01 -1.15834847e-02 5.05566657e-01
3.95665348e-01 1.28336477e+00 9.43646610e-01 -1.37000346e+00
-1.05016887e+00 -1.23031281e-01 2.94865638e-01 2.66827255e-01
-4.51776892e-01 9.67916995e-02 -5.51403046e-01 -7.31962979e-01
-2.80752014e-02 -6.86619282e-01 -2.24619731e-01 3.83310854e-01
-4.62692320e-01 -2.95013279e-01 1.21805441e+00 -4.70529914e-01
1.08509219e+00 -2.10244632e+00 7.96657056e-03 -4.14742202e-01
3.36305112e-01 4.02828574e-01 2.27556583e-02 -1.33273631e-01
5.23416214e-02 -1.82171062e-01 -3.58381718e-01 -7.73797452e-01
-2.82529026e-01 1.18828267e-01 -3.87201339e-01 4.69739497e-01
6.18670881e-01 1.01437426e+00 -1.37496340e+00 -1.02216899e+00
6.34627104e-01 4.54583347e-01 -5.75171173e-01 1.89958632e-01
-2.80560404e-01 5.70969820e-01 -4.81798798e-01 9.32938695e-01
5.42400777e-01 -4.67209578e-01 -4.07056361e-01 -3.80231023e-01
-4.56171930e-01 -4.78808582e-03 -8.45747948e-01 2.07901907e+00
-2.72448838e-01 8.50127637e-01 -2.11769491e-01 -1.00984073e+00
8.19819868e-01 1.67214617e-01 6.99954510e-01 -4.95063722e-01
9.52145830e-02 9.78888422e-02 -4.33732092e-01 -6.64865434e-01
7.38474011e-01 1.61944509e-01 2.66004592e-01 1.48838133e-01
2.04363272e-01 2.92409062e-01 3.02652121e-01 3.56352150e-01
8.07335317e-01 7.44661808e-01 1.09300971e-01 -2.51416028e-01
6.89583957e-01 6.33878559e-02 8.89678717e-01 5.32120764e-01
-8.68155420e-01 9.73565698e-01 4.49229419e-01 -5.64808011e-01
-8.86644661e-01 -7.28899658e-01 -1.34892419e-01 1.21914423e+00
8.26574206e-01 -4.45865512e-01 -5.59709668e-01 -9.62745547e-01
-2.26501912e-01 2.40588978e-01 -7.34228909e-01 5.22092870e-03
-5.35711050e-01 -4.30265933e-01 1.45806253e-01 7.48310208e-01
9.26282108e-01 -1.13781965e+00 -1.00424278e+00 1.46085054e-01
-5.13908803e-01 -1.42269039e+00 -7.05913901e-01 2.56284140e-02
-5.87909758e-01 -8.90608847e-01 -8.96731615e-01 -8.96638215e-01
5.80009639e-01 7.76247919e-01 1.04536998e+00 1.84762254e-02
-2.54721701e-01 1.37005523e-01 -5.08430123e-01 -3.09883326e-01
5.19863069e-01 -3.92023847e-02 -1.26622081e-01 2.80534148e-01
4.15074170e-01 -3.49097282e-01 -9.17320728e-01 4.39157039e-01
-7.05679297e-01 5.03623307e-01 6.10795617e-01 9.25425410e-01
9.01099205e-01 -4.78685051e-01 5.29225290e-01 -7.60780990e-01
-2.17282519e-01 -3.87938917e-01 -4.23661619e-01 7.17578903e-02
-3.50091010e-01 -8.46194550e-02 3.27365637e-01 -1.77022249e-01
-1.28161705e+00 4.63634610e-01 1.49044871e-01 -9.77187455e-01
-2.26727366e-01 2.66932905e-01 -4.08842713e-02 -6.14841953e-02
3.88144940e-01 2.49496549e-01 -3.30457807e-01 -2.90469408e-01
4.20913219e-01 4.26023155e-01 7.69243896e-01 -4.48630691e-01
5.68877459e-01 8.98112118e-01 -1.58257008e-01 -6.52094781e-01
-1.40943408e+00 -8.85269821e-01 -9.95567560e-01 -4.24758703e-01
1.20511472e+00 -1.15104187e+00 -3.40599179e-01 2.92303532e-01
-1.31353176e+00 -3.39947462e-01 -3.36014479e-01 5.23844242e-01
-7.17930079e-01 3.96883756e-01 -4.35118943e-01 -7.75791705e-01
-1.66225731e-01 -1.15299129e+00 1.54483342e+00 4.04039204e-01
8.72243568e-02 -9.75853860e-01 -1.66163027e-01 4.26115543e-01
2.50763655e-01 2.55385458e-01 1.68974370e-01 -1.65886372e-01
-1.00415552e+00 1.79707393e-01 -6.96965396e-01 1.62525535e-01
-5.42610399e-02 -1.03750832e-01 -1.07618868e+00 -1.50954723e-01
-3.15402359e-01 -4.15049523e-01 1.17953134e+00 5.79842150e-01
1.43323338e+00 -4.48097773e-02 -5.13752341e-01 7.41820514e-01
1.29708457e+00 -2.37456486e-01 5.50194800e-01 4.99389201e-01
1.13189805e+00 5.77746689e-01 1.16150153e+00 4.60167050e-01
5.57411909e-01 6.98940396e-01 6.92915022e-01 -4.60088968e-01
-4.26988482e-01 -3.70545417e-01 3.80565375e-01 2.54407018e-01
-2.42585659e-01 -5.67973070e-02 -5.76101422e-01 8.45775843e-01
-2.16836572e+00 -1.11480105e+00 -1.50712520e-01 1.82550013e+00
8.01054716e-01 2.02635124e-01 1.61491588e-01 7.57659553e-03
7.48158753e-01 4.25004870e-01 -5.47986209e-01 1.74908578e-01
-3.43823195e-01 -2.73509026e-01 4.93942142e-01 4.36911106e-01
-1.46852016e+00 1.22932923e+00 5.34608078e+00 7.02884197e-01
-1.23892224e+00 3.16941142e-01 8.15051675e-01 -1.88335538e-01
-4.04315516e-02 2.20037788e-01 -8.93893182e-01 6.42416537e-01
4.31555986e-01 2.71966875e-01 -2.24992752e-01 9.60592151e-01
5.76262772e-01 -3.36880863e-01 -9.94721830e-01 1.03255272e+00
3.04426640e-01 -1.46863639e+00 -1.05453797e-01 -1.57870024e-01
1.01216304e+00 1.28767803e-01 -5.02269389e-03 8.79751667e-02
-3.22993174e-02 -5.72686613e-01 9.20574963e-01 5.68008006e-01
5.76872945e-01 -5.04537642e-01 7.71620810e-01 1.29642367e-01
-1.49210227e+00 -1.26050502e-01 -3.83837253e-01 -2.41947975e-02
4.71344829e-01 7.10751116e-01 -3.43142033e-01 4.17424470e-01
1.02754915e+00 1.63426554e+00 -6.53138876e-01 1.19841921e+00
-2.35765681e-01 6.03894770e-01 -7.18583912e-02 2.61289716e-01
4.98999655e-01 7.73619907e-03 5.52093387e-01 1.30447185e+00
5.00478558e-02 1.86843026e-05 4.87606943e-01 9.09520328e-01
7.80852139e-02 -1.02819636e-01 -3.06303471e-01 3.78040940e-01
2.58620590e-01 1.46453381e+00 -8.94150496e-01 -5.25843441e-01
-6.62574589e-01 1.08643997e+00 3.56158495e-01 4.62917268e-01
-1.03347886e+00 -2.78275281e-01 5.55978298e-01 1.28508925e-01
4.32939678e-01 -1.24519683e-01 -3.14947397e-01 -1.35945427e+00
8.90987739e-02 -2.27984071e-01 3.93890858e-01 -1.06041694e+00
-1.03495884e+00 3.89611870e-01 -8.87290835e-02 -1.60144424e+00
7.38888159e-02 -3.37290943e-01 -6.04953468e-01 6.86516881e-01
-2.31542063e+00 -1.41201186e+00 -6.05968714e-01 7.40381002e-01
9.84876275e-01 5.92065975e-02 2.78566092e-01 4.15834367e-01
-6.34057641e-01 4.04880673e-01 -2.63595164e-01 2.31508583e-01
9.03752029e-01 -1.18257618e+00 -1.44748077e-01 1.09867048e+00
1.29054457e-01 3.15741867e-01 4.57848430e-01 -4.45995092e-01
-1.03887618e+00 -1.40331519e+00 7.73983419e-01 -4.88252401e-01
5.58744371e-01 -1.46087736e-01 -9.47309136e-01 6.72537684e-01
2.36826837e-01 7.54940152e-01 2.80409455e-01 -2.58317858e-01
-1.16953798e-01 2.50052963e-03 -7.63317347e-01 3.29580396e-01
1.14996231e+00 -4.66795236e-01 -4.17568535e-01 2.94985652e-01
1.00103045e+00 -4.46224004e-01 -5.98073840e-01 3.65759194e-01
3.24711591e-01 -1.17306697e+00 1.05983257e+00 -1.97982401e-01
8.87624860e-01 -7.21872807e-01 1.10816918e-01 -9.03953910e-01
-1.99497998e-01 -5.55270672e-01 -4.70036209e-01 1.39888716e+00
1.56795785e-01 -2.09674269e-01 9.39709187e-01 3.76051277e-01
-5.20817161e-01 -9.41922665e-01 -7.15757549e-01 -4.85767990e-01
-4.23980951e-01 -3.18993837e-01 2.36265242e-01 9.94036853e-01
9.68552306e-02 2.90656656e-01 -6.74381137e-01 1.15477026e-01
6.55817151e-01 2.38306656e-01 6.16559446e-01 -9.84430850e-01
-8.87067914e-02 -4.38925356e-01 -5.15724719e-01 -1.40750885e+00
3.08155894e-01 -6.43910110e-01 2.37370104e-01 -1.35042334e+00
3.56936216e-01 -4.02153790e-01 -4.93702263e-01 6.94350064e-01
-6.00375473e-01 5.47827899e-01 1.74197704e-01 4.41348523e-01
-1.10466790e+00 7.93453693e-01 1.40745842e+00 -2.08874986e-01
-2.44969785e-01 -3.16957355e-01 -4.86137509e-01 7.96388268e-01
5.21172762e-01 -2.25485981e-01 -4.67015475e-01 -5.12332857e-01
-2.06516713e-01 4.43398068e-03 6.52324736e-01 -1.06101501e+00
3.33092034e-01 -2.88433105e-01 4.66518641e-01 -8.61217797e-01
4.16620821e-01 -6.81606472e-01 -6.21239424e-01 1.25469506e-01
-5.87869644e-01 -8.14653039e-02 -4.32193503e-02 9.15039897e-01
-5.53011417e-01 1.57472044e-02 8.93109500e-01 -7.80910477e-02
-1.24548817e+00 7.78567255e-01 -4.15664241e-02 2.97013875e-02
1.13451993e+00 -2.48327374e-01 -3.52456689e-01 -2.54792739e-02
-4.55004543e-01 4.11032557e-01 4.14459258e-01 6.06179655e-01
7.58478940e-01 -1.26360631e+00 -6.16253138e-01 1.51132718e-01
3.63479286e-01 2.80985713e-01 3.64078134e-01 1.19855082e+00
-3.62883449e-01 4.08051312e-01 -3.14750850e-01 -1.04343200e+00
-9.89495099e-01 7.50338614e-01 1.60938293e-01 5.43777645e-02
-7.73710310e-01 1.03888857e+00 4.86247510e-01 2.42136925e-01
2.65372366e-01 -3.03085804e-01 -4.90450084e-01 7.75089338e-02
6.62268519e-01 1.74870804e-01 -2.84588963e-01 -9.46274936e-01
-2.78186470e-01 7.62311339e-01 8.75315722e-03 1.04303375e-01
1.17122936e+00 -6.27749801e-01 -5.54078259e-02 4.15290505e-01
1.23467588e+00 -3.72042775e-01 -2.03702354e+00 -5.28560400e-01
-1.61369264e-01 -9.39392567e-01 2.93473005e-01 -3.94253761e-01
-1.29509175e+00 1.21879625e+00 6.72566712e-01 -3.84868607e-02
1.14931118e+00 -2.52081864e-02 9.06688511e-01 -9.34681222e-02
2.74385720e-01 -1.11447442e+00 3.41317862e-01 3.90385121e-01
6.08672976e-01 -1.81519055e+00 -1.13903373e-01 -5.86939275e-01
-8.56718898e-01 8.95113349e-01 1.03141117e+00 -8.88609216e-02
5.72351694e-01 -1.27954766e-01 -4.40560728e-02 1.13714403e-02
-6.75192535e-01 -5.66335440e-01 2.67010927e-01 6.22216642e-01
4.84880775e-01 -2.64711142e-01 -2.36964732e-01 3.81286383e-01
2.57949650e-01 2.36349240e-01 5.24447441e-01 8.89441073e-01
-6.02461278e-01 -6.82924628e-01 -1.64282918e-01 2.36060083e-01
-3.85593832e-01 -2.35215068e-01 -2.24765744e-02 5.27516186e-01
3.85766894e-01 9.07341301e-01 1.88183695e-01 -1.94553062e-01
-1.18549667e-01 -2.44050443e-01 1.29212096e-01 -4.61963296e-01
-1.80056289e-01 1.15811333e-01 -1.38576671e-01 -7.56802619e-01
-1.12054348e+00 -7.18273938e-01 -1.38548279e+00 1.05877779e-01
-3.22031736e-01 -1.23001542e-02 3.03628206e-01 1.04654896e+00
3.51601928e-01 5.74330211e-01 5.40053844e-01 -1.52870083e+00
1.69165999e-01 -7.97513843e-01 -3.95231485e-01 5.31780601e-01
6.38740361e-01 -1.06187379e+00 -2.69645303e-01 5.23076773e-01] | [9.610840797424316, -0.33181050419807434] |
5132b0d3-5236-4f1d-b2e6-51b90bf53fe5 | evaluating-the-covid-19-identification-resnet | 2107.14549 | null | https://arxiv.org/abs/2107.14549v1 | https://arxiv.org/pdf/2107.14549v1.pdf | Evaluating the COVID-19 Identification ResNet (CIdeR) on the INTERSPEECH COVID-19 from Audio Challenges | We report on cross-running the recent COVID-19 Identification ResNet (CIdeR) on the two Interspeech 2021 COVID-19 diagnosis from cough and speech audio challenges: ComParE and DiCOVA. CIdeR is an end-to-end deep learning neural network originally designed to classify whether an individual is COVID-positive or COVID-negative based on coughing and breathing audio recordings from a published crowdsourced dataset. In the current study, we demonstrate the potential of CIdeR at binary COVID-19 diagnosis from both the COVID-19 Cough and Speech Sub-Challenges of INTERSPEECH 2021, ComParE and DiCOVA. CIdeR achieves significant improvements over several baselines. | ['Björn W. Schuller', 'Lyn Jones', 'Panagiotis Tzirakis', 'Alexander Gaskell', 'Harry Coppock', 'Alican Akman'] | 2021-07-30 | null | null | null | null | ['covid-19-detection'] | ['medical'] | [-1.07580878e-01 -4.31674808e-01 -9.48545411e-02 -1.81102410e-01
-9.84610558e-01 -6.96430385e-01 1.06473155e-01 1.07692853e-01
-7.23658919e-01 6.78238451e-01 2.24027902e-01 -4.39323708e-02
-8.04054439e-02 -3.33163053e-01 -3.37947726e-01 -2.38050118e-01
-2.45016366e-01 1.15981364e+00 -4.15288061e-01 1.36467263e-01
-7.11259186e-01 7.41000101e-02 -1.36844146e+00 2.18315512e-01
2.76578814e-01 8.63385379e-01 -2.22737223e-01 1.43639410e+00
5.72542489e-01 3.73403609e-01 -9.97769415e-01 -5.58745265e-01
1.66876048e-01 -2.27186605e-02 -8.97352636e-01 -9.24540699e-01
8.22204590e-01 -8.16332698e-01 -2.34425709e-01 5.22836328e-01
1.43797469e+00 -2.90065706e-01 6.79003775e-01 -1.54300821e+00
-6.34090900e-01 7.24274397e-01 -1.89880297e-01 6.39597595e-01
5.37233055e-01 5.54233253e-01 1.09868860e+00 -8.27503383e-01
3.04492384e-01 1.32800591e+00 1.59470105e+00 9.15624559e-01
-1.18623984e+00 -1.42595971e+00 -5.88464439e-01 1.20599698e-02
-1.39112306e+00 6.72584400e-02 7.62352273e-02 -8.91075671e-01
1.34078658e+00 1.46106467e-01 5.18687606e-01 2.05919313e+00
-3.49341005e-01 6.84078455e-01 6.43762887e-01 4.55083668e-01
-8.12423006e-02 -1.53653815e-01 4.68678147e-01 1.45026132e-01
2.70168126e-01 5.07906318e-01 -5.13235092e-01 -7.75487959e-01
1.42479941e-01 -1.24978051e-01 9.37488228e-02 5.74207544e-01
-1.20139468e+00 7.84907103e-01 8.42893645e-02 -2.01031826e-02
-3.89780581e-01 4.14890617e-01 8.84025693e-01 5.51092848e-02
4.11309332e-01 6.88683391e-01 -5.27819455e-01 -4.44922715e-01
-8.87556016e-01 7.41402984e-01 9.30314600e-01 4.29873228e-01
1.34643450e-01 1.77292556e-01 -7.63236105e-01 1.16296256e+00
7.23604634e-02 1.37262630e+00 4.95063990e-01 -9.92676735e-01
6.58600390e-01 -9.35453624e-02 2.99313784e-01 -6.03265882e-01
-8.58163893e-01 -2.52620816e-01 -7.72799551e-01 -4.05439883e-01
-9.26114023e-02 -9.46227193e-01 -8.71450603e-01 1.90510750e+00
8.17686692e-03 7.04937160e-01 4.65876423e-02 5.80380559e-01
1.72225082e+00 1.74556732e-01 2.64283478e-01 2.31464967e-01
1.28055501e+00 -3.80312324e-01 -7.75300741e-01 -5.75820059e-02
3.92595172e-01 -6.05392218e-01 5.59779763e-01 4.11462009e-01
-8.09201419e-01 -7.61705816e-01 -8.43796194e-01 3.86207581e-01
-4.24200654e-01 9.18361098e-02 -1.29771948e-01 6.84864283e-01
-1.25767112e+00 3.94397229e-01 -3.76617104e-01 -9.26007796e-03
2.91932970e-01 6.40563428e-01 -2.80617893e-01 4.37852293e-01
-1.68847275e+00 3.64285588e-01 4.62610871e-02 5.92595451e-02
-1.22887051e+00 -9.59116817e-01 -7.57104754e-01 -2.68428653e-01
-2.22892180e-01 -8.11683893e-01 1.76249802e+00 -6.43051341e-02
-4.92132843e-01 1.18681979e+00 -7.14070722e-02 -1.02450752e+00
5.64449966e-01 -5.21676183e-01 -7.70430028e-01 2.48047322e-01
4.44104135e-01 6.87900305e-01 9.07903492e-01 -1.05176544e+00
-5.83146632e-01 5.15208878e-02 -5.68798184e-01 -1.50528997e-01
-1.47090450e-01 3.91583443e-01 8.73401947e-03 -6.74418211e-01
-1.12437713e+00 -1.21727955e+00 2.83746183e-01 -4.70358491e-01
-9.44090128e-01 -5.83900332e-01 1.12370431e+00 -8.32081854e-01
1.03469491e+00 -2.13572693e+00 -2.68995404e-01 -1.23108461e-01
7.45158851e-01 1.22490585e+00 -5.08505225e-01 5.02650499e-01
-3.51192802e-01 4.23313081e-01 -9.76346582e-02 -1.00189602e+00
6.94735721e-03 2.65106827e-01 -4.05915231e-01 4.21782702e-01
4.07573730e-01 8.54304850e-01 -9.91389871e-01 -5.46997309e-01
2.82774925e-01 5.71804583e-01 -5.14775157e-01 7.63061523e-01
1.75032631e-01 4.73110050e-01 4.59334999e-02 7.68553555e-01
7.99351811e-01 -6.93267882e-02 -4.99639064e-01 -2.73394324e-02
1.84219256e-01 1.02744857e-03 -5.42607784e-01 8.02297831e-01
-3.68588477e-01 8.37924838e-01 4.18802708e-01 -7.38323987e-01
6.39152348e-01 9.30658400e-01 7.96721578e-01 -2.04674035e-01
2.63023227e-01 3.50243509e-01 -7.87771717e-02 -8.80073249e-01
5.14931798e-01 3.36722583e-02 -3.48920077e-01 4.04377937e-01
2.02577531e-01 7.07952306e-02 -8.04740265e-02 5.03482856e-03
1.22889924e+00 -6.99868619e-01 -1.05610766e-01 2.05076665e-01
1.45958781e-01 -3.37417245e-01 5.39584637e-01 1.24999428e+00
-7.49227881e-01 1.18335462e+00 7.56725371e-02 -2.63133466e-01
-8.38621080e-01 -1.55324149e+00 -3.80823314e-01 9.38355446e-01
-4.02888775e-01 -1.48809597e-01 -5.71206033e-01 -5.90818346e-01
4.93585080e-01 2.86027253e-01 -9.23204303e-01 1.27082631e-01
-7.14837909e-01 -5.64825177e-01 1.83431506e+00 9.05872285e-01
5.42355597e-01 -1.34715736e+00 -1.53246015e-01 2.36985713e-01
-6.21708989e-01 -1.50614095e+00 -6.64696813e-01 3.30907315e-01
1.81814089e-01 -1.18195975e+00 -1.11695588e+00 -9.24129248e-01
-4.40149754e-01 -8.09542313e-02 1.22198009e+00 -1.68214049e-02
-8.47098589e-01 6.92959249e-01 -3.05941224e-01 -9.70652759e-01
-6.12618983e-01 2.91153520e-01 7.12049484e-01 -2.89136916e-01
8.81233454e-01 -2.45713010e-01 -7.03093529e-01 2.53069788e-01
-3.97761822e-01 -6.69339955e-01 -2.47215539e-01 8.42609823e-01
2.40989998e-01 -7.45022774e-01 1.14257669e+00 -2.90722162e-01
1.06273556e+00 -7.64980078e-01 -1.18040949e-01 -3.22224498e-01
-3.64517897e-01 -7.31306851e-01 4.39265013e-01 -3.88097048e-01
-1.29332528e-01 -1.31352633e-01 -7.16456890e-01 -9.71289217e-01
-3.04369777e-01 8.23022276e-02 7.90059149e-01 5.99779963e-01
6.30945802e-01 -5.14170192e-02 -3.04814856e-02 -4.44831431e-01
-1.07229203e-01 1.44928861e+00 1.31946659e+00 -2.67432541e-01
8.64475191e-01 2.74133176e-01 -5.62860548e-01 -8.02609861e-01
-7.80414283e-01 -1.15520346e+00 -3.96715760e-01 4.96946946e-02
1.82616222e+00 -1.46195924e+00 -1.34650970e+00 9.93710637e-01
-1.34647512e+00 -2.41828978e-01 -1.96031302e-01 5.34196317e-01
-1.61142007e-01 2.37198874e-01 -5.84429443e-01 -1.05477738e+00
-1.20914388e+00 -1.03660715e+00 1.57293177e+00 -1.28738885e-03
-1.00034201e+00 -5.81937313e-01 8.36556554e-01 5.35234034e-01
3.22047979e-01 3.56918603e-01 3.00177574e-01 -1.55224097e+00
3.84690076e-01 -4.86984670e-01 -3.24194461e-01 8.82150531e-01
1.87218189e-01 3.29122730e-02 -1.61774898e+00 -4.21739191e-01
-5.25793314e-01 -9.01751816e-01 8.70584249e-01 3.88791710e-01
1.23923624e+00 -3.56145889e-01 -3.04786325e-01 6.11000955e-01
8.77465665e-01 1.80035055e-01 -9.89065170e-02 -1.43340081e-01
8.74301970e-01 3.60020041e-01 2.70409822e-01 6.67329609e-01
4.62464690e-01 5.05473495e-01 4.92562085e-01 -2.74532080e-01
-2.14031279e-01 -1.17923476e-01 3.14135365e-02 7.20487118e-01
1.49296820e-01 -4.85154510e-01 -1.32681835e+00 1.27261984e+00
-1.41489697e+00 -1.09894919e+00 3.37968469e-02 1.93269408e+00
8.56322646e-01 -5.25132239e-01 7.68402755e-01 3.00998557e-02
1.03970563e+00 2.24382058e-01 -6.76633358e-01 -7.17147708e-01
-9.24056992e-02 5.27322769e-01 3.89359057e-01 1.90337583e-01
-1.43804252e+00 2.70425677e-01 7.13055992e+00 4.96492475e-01
-9.33368027e-01 2.46907488e-01 3.25010836e-01 -5.34342229e-01
3.00737232e-01 -1.13087535e+00 -9.86435771e-01 6.97524369e-01
1.30882835e+00 1.57892182e-01 4.62891787e-01 6.48910105e-01
3.23700719e-02 5.82775235e-01 -1.38713801e+00 1.39049566e+00
-1.26488376e-02 -1.07746875e+00 -4.96880949e-01 -1.04237236e-01
6.52811110e-01 1.03747940e+00 5.44611275e-01 6.13689899e-01
7.61132717e-01 -1.52399766e+00 3.40205073e-01 -1.54848620e-01
1.59154105e+00 -7.39406645e-01 1.16366553e+00 -5.92502169e-02
-1.30475664e+00 -2.72361755e-01 1.89796552e-01 4.90243584e-01
1.47373259e-01 1.69987172e-01 -1.71456778e+00 1.87284738e-01
1.25525403e+00 7.35244751e-01 -2.32470736e-01 8.43749762e-01
2.15786740e-01 1.02343714e+00 -3.24920893e-01 -5.45567535e-02
2.73049444e-01 6.90298259e-01 8.65027964e-01 1.90260792e+00
7.64966384e-02 -4.26127970e-01 1.36644058e-02 8.90185118e-01
-4.28879976e-01 -6.83008373e-01 -8.31345975e-01 -4.08719033e-01
8.20583582e-01 8.46298099e-01 4.37792867e-01 -4.17997688e-01
4.63224091e-02 6.25750840e-01 -1.55438796e-01 2.83544987e-01
-9.97472346e-01 -4.54320610e-01 1.27447045e+00 -2.47349516e-01
5.76598287e-01 3.33234489e-01 6.60905316e-02 -6.72411501e-01
-2.74314612e-01 -1.24903536e+00 4.36138451e-01 -6.02092445e-01
-2.02221060e+00 6.12521410e-01 1.49682447e-01 -1.25608063e+00
-9.56631780e-01 -4.97740239e-01 -7.61306345e-01 1.09771371e+00
-1.26453531e+00 -8.40048730e-01 -3.52242887e-01 7.08933771e-01
3.94228071e-01 -3.74809384e-01 1.12438071e+00 5.68421483e-01
-5.43321908e-01 1.18375301e+00 6.24262914e-02 7.01517940e-01
1.16999316e+00 -1.37475669e+00 1.01631010e+00 2.63781339e-01
-3.18621427e-01 3.77730578e-01 7.29801834e-01 -8.27639520e-01
-8.34526539e-01 -1.71257377e+00 9.29847181e-01 -9.45577741e-01
5.80260038e-01 -5.02175331e-01 -6.91465497e-01 4.79908973e-01
2.38171145e-01 -3.84463035e-02 7.97366083e-01 2.92373030e-03
-6.21075451e-01 -2.94932667e-02 -1.44299018e+00 1.01503134e-01
8.37313414e-01 -1.26437342e+00 -7.66124785e-01 4.25941676e-01
1.23736024e+00 -2.57469505e-01 -1.08420956e+00 7.37240911e-01
7.93156445e-01 -5.34061909e-01 1.50752974e+00 -6.08571529e-01
3.21222603e-01 2.51252383e-01 -1.12039112e-01 -1.32077754e+00
5.94816282e-02 -9.89042819e-01 1.45702004e-01 1.24198389e+00
2.86135048e-01 -8.21517110e-01 5.15810192e-01 1.22153416e-01
1.73463598e-01 -3.19763035e-01 -1.30908835e+00 -9.53702927e-01
2.83095360e-01 -6.64293408e-01 8.18907201e-01 8.96104813e-01
-4.32558924e-01 2.65247136e-01 -8.70391905e-01 1.97311923e-01
4.70731974e-01 -4.23043579e-01 9.06527817e-01 -1.44995010e+00
-5.51208794e-01 -1.61073029e-01 -7.53960013e-02 -2.20466956e-01
1.24755546e-01 -5.63510001e-01 3.49205226e-01 -1.43364930e+00
3.29862714e-01 -2.27382347e-01 -3.03448766e-01 5.99278390e-01
-1.85296193e-01 4.98127818e-01 2.29150936e-01 8.74115154e-02
-3.57393950e-01 6.75380230e-02 5.87599993e-01 -5.24768770e-01
-1.04909956e-01 2.52166986e-01 -2.88254023e-01 6.25757217e-01
9.17001367e-01 -6.59088016e-01 -1.54937312e-01 -2.76900291e-01
3.10104191e-02 1.52913079e-01 6.06571436e-01 -9.28315461e-01
-1.20056555e-01 5.11989534e-01 1.07207432e-01 -1.26761842e+00
8.82732630e-01 -3.66127342e-01 -2.32946530e-01 7.17639565e-01
-6.09247863e-01 4.27571774e-01 4.86385703e-01 7.18151987e-01
-3.64822112e-02 -5.83327003e-03 5.21977544e-01 8.66478533e-02
-7.82220811e-02 4.45526749e-01 -6.98953807e-01 8.58996570e-01
3.92692626e-01 1.57295987e-01 -8.88792872e-01 -5.18260300e-01
-7.20804095e-01 5.01104772e-01 -3.33539784e-01 7.94562638e-01
6.01682484e-01 -1.07083488e+00 -1.27192068e+00 2.75668889e-01
3.33833605e-01 1.39676677e-02 2.20036238e-01 7.64003694e-01
-4.47193414e-01 6.39954686e-01 1.27033144e-01 -1.00140905e+00
-1.87883139e+00 4.18731809e-01 5.44019520e-01 -4.10209268e-01
-2.47319892e-01 1.22371459e+00 2.31600776e-01 -1.03225362e+00
7.39043176e-01 -4.03664917e-01 -2.18217745e-01 1.64627582e-01
6.21823430e-01 6.28106833e-01 -1.13654464e-01 -7.10357308e-01
-8.06272388e-01 4.13861632e-01 5.46384305e-02 -2.80913655e-02
9.26183105e-01 2.42155343e-01 1.90993920e-01 2.40686074e-01
1.63583624e+00 -1.97124854e-01 -4.46101367e-01 -1.02648750e-01
-5.64586937e-01 1.21695414e-01 -4.66763854e-01 -1.11284626e+00
-7.03125656e-01 1.12160647e+00 1.15851462e+00 1.84029073e-01
7.83391595e-01 1.94231030e-02 1.41438735e+00 6.64672613e-01
-2.40531757e-01 -8.63721490e-01 9.50920284e-02 6.09751284e-01
8.80422950e-01 -1.44093537e+00 -6.30670965e-01 1.75163969e-01
-6.72557652e-01 3.86381686e-01 5.64264894e-01 -1.67795226e-01
7.04098940e-01 5.36973476e-01 3.08380276e-01 -3.78013343e-01
-5.72541893e-01 -1.38277322e-01 1.91169679e-01 1.23471737e+00
1.50776878e-01 5.85358918e-01 3.11851859e-01 6.99380755e-01
-5.23887336e-01 5.19329496e-02 2.11922392e-01 2.56083101e-01
2.05758840e-01 -5.54650009e-01 -3.28382939e-01 5.73742092e-01
-1.02174509e+00 -1.96266472e-01 -1.03397250e+00 7.98894942e-01
5.24970829e-01 1.27567720e+00 2.50931889e-01 -1.00428796e+00
1.38935030e-01 -1.06814727e-01 -2.22100869e-01 -6.95660889e-01
-1.31528771e+00 -2.43000031e-01 7.44274318e-01 -3.62253755e-01
-4.64563757e-01 -5.87904394e-01 -1.07563829e+00 -8.71016160e-02
1.20731063e-01 -1.22874968e-01 3.98605883e-01 7.37603188e-01
2.93912113e-01 5.77966332e-01 6.69720113e-01 -6.84329569e-01
-9.23254132e-01 -1.13829851e+00 -3.64558518e-01 2.42242217e-01
1.35005403e+00 -4.31800067e-01 -6.40320897e-01 -4.21023428e-01] | [14.568501472473145, 4.104015350341797] |
88876aad-d424-4cc3-9d3c-c76a1d6fe975 | automatic-sleep-stage-classification-with | 2008.09416 | null | https://arxiv.org/abs/2008.09416v1 | https://arxiv.org/pdf/2008.09416v1.pdf | Automatic sleep stage classification with deep residual networks in a mixed-cohort setting | Study Objectives: Sleep stage scoring is performed manually by sleep experts and is prone to subjective interpretation of scoring rules with low intra- and interscorer reliability. Many automatic systems rely on few small-scale databases for developing models, and generalizability to new datasets is thus unknown. We investigated a novel deep neural network to assess the generalizability of several large-scale cohorts. Methods: A deep neural network model was developed using 15684 polysomnography studies from five different cohorts. We applied four different scenarios: 1) impact of varying time-scales in the model; 2) performance of a single cohort on other cohorts of smaller, greater or equal size relative to the performance of other cohorts on a single cohort; 3) varying the fraction of mixed-cohort training data compared to using single-origin data; and 4) comparing models trained on combinations of data from 2, 3, and 4 cohorts. Results: Overall classification accuracy improved with increasing fractions of training data (0.25$\%$: 0.782 $\pm$ 0.097, 95$\%$ CI [0.777-0.787]; 100$\%$: 0.869 $\pm$ 0.064, 95$\%$ CI [0.864-0.872]), and with increasing number of data sources (2: 0.788 $\pm$ 0.102, 95$\%$ CI [0.787-0.790]; 3: 0.808 $\pm$ 0.092, 95$\%$ CI [0.807-0.810]; 4: 0.821 $\pm$ 0.085, 95$\%$ CI [0.819-0.823]). Different cohorts show varying levels of generalization to other cohorts. Conclusions: Automatic sleep stage scoring systems based on deep learning algorithms should consider as much data as possible from as many sources available to ensure proper generalization. Public datasets for benchmarking should be made available for future research. | ['Helge B. D. Sorensen', 'Poul Jennum', 'Emmanuel Mignot', 'Alexander Neergaard Olesen'] | 2020-08-21 | null | null | null | null | ['automatic-sleep-stage-classification'] | ['medical'] | [-2.47752100e-01 -8.65070745e-02 -2.58346111e-01 -4.17677939e-01
-7.22617090e-01 -4.40961003e-01 -6.07495308e-02 4.31237400e-01
-7.25095093e-01 9.68131185e-01 7.99920186e-02 -3.64222080e-01
-3.73984605e-01 -7.62445092e-01 -5.33069789e-01 -2.31162727e-01
-5.61273515e-01 3.32473785e-01 1.08664289e-01 -7.50222653e-02
-1.36075675e-01 1.34099290e-01 -1.28990996e+00 -5.62632605e-02
9.00166333e-01 1.14372647e+00 -8.61524567e-02 2.76419491e-01
3.30429465e-01 2.06958771e-01 -1.00687802e+00 -4.33107130e-02
4.27403241e-01 -4.17023093e-01 -2.21294582e-01 -2.30908766e-01
2.75544584e-01 -9.26210210e-02 1.43994585e-01 1.04504573e+00
7.03743219e-01 1.50644481e-01 5.70007861e-01 -1.20144200e+00
-4.29374903e-01 3.94408911e-01 -4.47670698e-01 4.90940720e-01
1.31827027e-01 2.36047730e-01 7.63315678e-01 -2.57734895e-01
2.71260679e-01 2.77296007e-01 9.77610707e-01 6.89546108e-01
-1.29408431e+00 -1.43586588e+00 5.96761443e-02 -1.54780447e-01
-1.53467178e+00 -3.46740603e-01 3.81573468e-01 -5.56095362e-01
9.71061707e-01 1.89960733e-01 1.02377939e+00 8.28500867e-01
4.21945065e-01 -1.36014953e-01 1.25290143e+00 1.16576096e-02
1.61905676e-01 3.43688965e-01 2.34147668e-01 7.27992177e-01
5.01606822e-01 1.19299062e-01 -3.80822778e-01 -3.83800089e-01
4.64951187e-01 2.10855249e-02 2.86099594e-02 2.35374540e-01
-7.74491608e-01 8.26901019e-01 3.92514110e-01 3.37710708e-01
-1.45767525e-01 -2.47572646e-01 4.09298152e-01 3.85377109e-01
4.85203415e-01 5.43876350e-01 -5.34363270e-01 -2.74256736e-01
-1.25738525e+00 2.27620542e-01 6.97112620e-01 6.10244751e-01
5.57672203e-01 8.35838988e-02 3.17140311e-01 1.23821425e+00
1.65827110e-01 5.52198112e-01 7.90160596e-01 -1.13703203e+00
4.98447955e-01 7.90054440e-01 1.82860345e-01 -8.00763667e-01
-1.23953247e+00 -6.91026747e-01 -9.82703447e-01 1.61664099e-01
3.21637541e-01 -2.34916806e-01 -5.98782599e-01 2.08484936e+00
-3.18719864e-01 -3.28639477e-01 -7.38295764e-02 5.72694540e-01
9.01229203e-01 3.13141018e-01 1.82424515e-01 -4.20089364e-01
1.39468825e+00 -3.75280172e-01 -2.71760881e-01 -4.73849177e-01
2.55895555e-01 -4.88078475e-01 1.13771367e+00 5.65351605e-01
-1.34679246e+00 -7.14633644e-01 -1.19272101e+00 2.93983370e-01
-2.81523466e-01 1.75010964e-01 1.52020246e-01 1.00158751e+00
-1.33854115e+00 6.75979495e-01 -1.10607123e+00 -3.85177583e-01
5.54520547e-01 9.22835767e-01 -1.10758267e-01 1.16559453e-01
-1.16582286e+00 7.37646580e-01 1.20099269e-01 -5.65170357e-03
-6.47800028e-01 -7.70618737e-01 -6.12762034e-01 9.61355641e-02
-1.78042248e-01 -6.59961164e-01 9.33020830e-01 -9.52211380e-01
-7.75511682e-01 8.61178219e-01 -2.18253776e-01 -3.64700109e-01
1.19918533e-01 1.95612423e-02 -8.96492541e-01 -2.16413867e-02
4.87849414e-01 5.60921550e-01 1.92685261e-01 -1.01459897e+00
-5.93206882e-01 -7.61199176e-01 1.03335619e-01 1.00237273e-01
-2.30810747e-01 2.12814540e-01 -4.80886959e-02 -6.17336333e-01
-7.08483756e-02 -1.16030395e+00 -2.46584550e-01 -1.97538346e-01
-5.38914539e-02 -1.39423549e-01 5.96367866e-02 -7.84850538e-01
1.42982149e+00 -1.91925502e+00 -4.04032111e-01 1.16707981e-01
5.94122469e-01 4.09206927e-01 3.21463257e-01 2.76605755e-01
-5.36644533e-02 3.21806520e-01 -3.86965215e-01 -4.56741184e-01
-1.69040740e-01 -1.16940960e-01 4.04837877e-01 4.34020579e-01
-1.27428427e-01 4.71628755e-01 -6.19640648e-01 -2.24603832e-01
1.58288181e-01 1.59934998e-01 -5.55662632e-01 -2.28931140e-02
4.41959441e-01 1.02297992e-01 -1.32920682e-01 5.86091280e-01
3.45581800e-01 -3.89013350e-01 1.15096852e-01 -4.83596772e-02
-1.97991908e-01 2.88589597e-01 -9.44319010e-01 1.48428512e+00
-4.20055598e-01 6.40010953e-01 7.14956550e-03 -9.49082136e-01
1.25824594e+00 2.75845021e-01 6.90500021e-01 -6.61302626e-01
1.50930047e-01 3.07130873e-01 6.93706632e-01 -2.12227568e-01
2.59261549e-01 -8.80251288e-01 -2.28751749e-01 4.47884768e-01
3.66553143e-02 -3.84277515e-02 1.97656095e-01 -4.87621687e-02
1.19693553e+00 -5.44496536e-01 2.86619037e-01 -7.02305555e-01
2.04674363e-01 -2.45389670e-01 8.93090844e-01 7.45521665e-01
-6.44932091e-01 6.27295315e-01 5.59937179e-01 -3.90458703e-01
-6.64902747e-01 -1.35136354e+00 -6.63587272e-01 7.91565180e-01
2.73204949e-02 -5.64270914e-01 -6.60122156e-01 -4.69810277e-01
2.36093737e-02 7.83018768e-01 -7.10068166e-01 -5.67010283e-01
-2.49698609e-02 -1.13806558e+00 8.11293542e-01 8.64421129e-01
5.70496500e-01 -1.06835794e+00 -7.36673117e-01 8.74807462e-02
-1.12788633e-01 -7.89377630e-01 -3.11901718e-01 3.64965051e-01
-9.56725001e-01 -9.36158657e-01 -5.45478880e-01 -3.74879062e-01
4.58387554e-01 -1.89591825e-01 1.31367719e+00 3.00599724e-01
-1.06690284e-02 2.68029898e-01 -1.97918907e-01 -5.89390039e-01
-1.78025171e-01 4.14700024e-02 6.19408667e-01 -3.44276547e-01
6.78496122e-01 -7.13666439e-01 -1.01837397e+00 4.93202627e-01
-6.47095025e-01 -4.57524180e-01 4.02337670e-01 7.67844677e-01
4.25164759e-01 -6.18101843e-02 8.40569675e-01 -6.72281623e-01
5.46564519e-01 -7.42023766e-01 -3.95324707e-01 -1.53061002e-01
-1.21110594e+00 -5.56539237e-01 8.83293509e-01 -4.08673346e-01
-5.05764902e-01 -4.02376741e-01 -2.11596772e-01 -4.02956128e-01
-1.87349856e-01 4.41221744e-01 1.39421284e-01 5.82799017e-01
8.73331726e-01 3.53966877e-02 8.41032118e-02 -3.40739608e-01
-3.78735870e-01 6.58979356e-01 2.87781119e-01 -4.84545767e-01
5.74763000e-01 3.28169197e-01 -2.75074273e-01 -6.49176300e-01
-5.38861156e-01 -3.02163184e-01 -4.43557590e-01 -1.09895188e-02
9.90263104e-01 -1.04460001e+00 -5.92749476e-01 2.68955886e-01
-1.97143063e-01 -8.39099586e-01 -1.88128397e-01 8.75061870e-01
-1.44466326e-01 9.34058726e-02 -3.84531647e-01 -3.40862036e-01
-7.68335283e-01 -1.23566329e+00 5.85738719e-01 4.58572894e-01
-8.49742889e-01 -1.02841187e+00 2.33019143e-01 6.84889913e-01
4.06959802e-01 2.86727875e-01 6.41011000e-01 -8.69400024e-01
2.03458890e-01 -3.75676185e-01 -2.43823975e-01 7.54720688e-01
5.23883283e-01 -9.47656706e-02 -8.82836282e-01 -5.57626069e-01
3.55774760e-02 -2.42115870e-01 4.62079793e-01 6.79225564e-01
1.21034718e+00 -1.32570729e-01 -2.73477614e-01 4.84607249e-01
1.23150945e+00 8.20459425e-01 5.32837152e-01 4.85180587e-01
2.44545072e-01 3.03006768e-01 1.26365677e-01 2.75851697e-01
4.66182560e-01 5.14424026e-01 -1.77482255e-02 6.26970232e-02
1.99887067e-01 1.31687626e-01 2.67558843e-01 5.12818396e-01
-2.00711176e-01 1.20680489e-01 -9.91141140e-01 6.36414289e-01
-1.20513129e+00 -8.38027120e-01 1.21177390e-01 2.38088179e+00
6.24450386e-01 6.37924194e-01 6.16292477e-01 9.79241952e-02
6.60184562e-01 3.06861196e-02 -7.12379754e-01 -6.44183576e-01
2.24399492e-01 6.39226437e-01 2.53800720e-01 9.78004932e-03
-6.23812914e-01 2.10328385e-01 4.24434233e+00 3.56763244e-01
-1.13240135e+00 2.22011298e-01 7.40835309e-01 -7.83750415e-01
2.99937218e-01 -1.21034808e-01 -5.85326731e-01 1.14005458e+00
1.42999887e+00 -5.39362073e-01 2.32787147e-01 5.57126939e-01
3.71740013e-01 -3.32885742e-01 -9.54524279e-01 9.61569071e-01
1.84628144e-01 -1.10136664e+00 -6.52553678e-01 2.03007832e-01
7.31251121e-01 5.09549439e-01 7.82853458e-03 5.23416936e-01
1.31864205e-01 -1.08168304e+00 7.37525225e-01 2.41952032e-01
1.28049135e+00 -4.81582403e-01 9.29667592e-01 3.59898657e-02
-1.09240091e+00 -2.06625387e-01 -1.53436959e-01 -4.39099729e-01
3.57933827e-02 6.25892341e-01 -6.38347208e-01 3.09113234e-01
1.39797688e+00 4.91271198e-01 -5.93516052e-01 7.80844629e-01
7.17883930e-02 7.85595655e-01 -4.25164640e-01 -1.15163699e-01
-6.35837168e-02 -4.13557850e-02 2.30783448e-02 8.98457646e-01
4.13501471e-01 2.27170408e-01 -1.72618553e-01 8.75104010e-01
-2.27138266e-01 4.08361293e-02 -5.16147852e-01 1.33652776e-01
4.45131570e-01 1.08243501e+00 -8.62262726e-01 -2.97579974e-01
-7.76029825e-01 2.68138319e-01 1.61164418e-01 1.38298213e-01
-7.44901180e-01 -5.47777236e-01 6.41336322e-01 6.27778351e-01
-1.16826020e-01 1.82503030e-01 -5.95024943e-01 -1.06335115e+00
1.68225938e-03 -6.94079638e-01 8.02312195e-01 -8.59106958e-01
-1.35265303e+00 6.60297692e-01 2.06309915e-01 -1.35511124e+00
7.54622519e-02 -3.36274475e-01 -7.06647992e-01 1.15540373e+00
-6.98545933e-01 -4.49946940e-01 -3.82963181e-01 3.50810736e-01
2.99839795e-01 -2.83676803e-01 7.78370380e-01 3.88057828e-01
-6.60152733e-01 9.01056826e-01 1.56293422e-01 1.77190468e-01
7.52014697e-01 -1.05845058e+00 -1.72538400e-01 3.69661897e-01
-3.63304198e-01 1.22079647e+00 2.90509667e-02 -5.46290994e-01
-6.87695980e-01 -1.06129122e+00 7.68324435e-01 -6.08636618e-01
4.83746916e-01 -4.51315492e-02 -8.93570602e-01 7.35801339e-01
1.29439253e-02 -2.83369899e-01 1.29369187e+00 6.49779975e-01
6.46016095e-03 -7.08855450e-01 -1.42445886e+00 3.83724242e-01
7.06920624e-01 -4.20800954e-01 -5.57085752e-01 -5.62115014e-02
2.14698225e-01 -3.69940311e-01 -1.43916655e+00 3.99161339e-01
9.59635675e-01 -1.17508924e+00 7.30451822e-01 -2.15208083e-01
4.58423883e-01 -1.34627834e-01 -9.74590704e-02 -1.05860054e+00
-2.67084092e-01 -2.54211068e-01 4.47237730e-01 1.02320790e+00
8.27901363e-01 -9.79810774e-01 7.28714287e-01 1.11361587e+00
-7.11879432e-01 -9.14789498e-01 -1.13997853e+00 -6.54553652e-01
5.31024516e-01 -5.45718253e-01 5.28736711e-01 9.10137296e-01
1.95734367e-01 2.27499858e-01 -1.80030074e-02 1.45695001e-01
2.69747019e-01 -1.15374774e-01 3.89956295e-01 -1.19130301e+00
-1.92990005e-01 -5.10440767e-01 -4.33322281e-01 -8.86563212e-02
-2.24036500e-01 -9.88702893e-01 -4.22095180e-01 -1.51526237e+00
5.74459553e-01 -7.16098785e-01 -8.91585648e-01 5.96864760e-01
-1.10603042e-01 4.40451592e-01 -3.98571678e-02 2.90517688e-01
-1.75176844e-01 1.14683717e-01 5.76601088e-01 5.21940328e-02
-5.29577196e-01 2.60903001e-01 -1.00577259e+00 6.75201654e-01
1.39741576e+00 -5.59184372e-01 -6.34623468e-01 -1.67890206e-01
2.74996638e-01 6.96694255e-02 3.41391295e-01 -1.33440065e+00
-9.95000750e-02 2.96271648e-02 7.37304807e-01 -3.01700592e-01
3.93679440e-01 -3.33840817e-01 4.41211551e-01 8.05338800e-01
-1.07411489e-01 6.24019504e-01 6.03639364e-01 1.72865868e-01
2.05499932e-01 -1.48219779e-01 7.84128428e-01 -3.09877843e-01
-1.13138445e-01 1.31960353e-03 -5.24599373e-01 3.31376314e-01
5.62578619e-01 -3.67944181e-01 -3.62034589e-01 -3.07523876e-01
-9.76370931e-01 1.06611267e-01 6.11619294e-01 4.91728961e-01
2.64197946e-01 -1.03937006e+00 -2.88048595e-01 1.90685734e-01
1.24190174e-01 -1.97557107e-01 2.91719466e-01 1.07467830e+00
-2.98917413e-01 1.74054459e-01 -3.50759149e-01 -3.08242351e-01
-1.11779892e+00 3.36220711e-01 5.68133414e-01 -9.90041420e-02
-1.98411241e-01 7.29970574e-01 2.53128052e-01 -4.40928429e-01
-1.21085308e-01 -6.22451007e-01 1.31043822e-01 1.46112233e-01
2.77921170e-01 7.18027174e-01 2.66180664e-01 -4.77275521e-01
-5.81276715e-01 4.94664639e-01 2.57729907e-02 -5.57695366e-02
1.20725608e+00 -4.13363799e-03 -1.21077649e-01 7.32296526e-01
1.16173303e+00 7.99634978e-02 -8.39751542e-01 3.17518800e-01
-4.52275306e-01 9.75540094e-03 -2.94561416e-01 -1.14536822e+00
-1.10705554e+00 6.47409081e-01 1.06719041e+00 2.07512841e-01
1.30132854e+00 6.18359596e-02 5.98710954e-01 -9.15870145e-02
3.22900712e-01 -1.14483404e+00 -2.23070964e-01 -1.05897762e-01
5.43906331e-01 -1.14196694e+00 2.16349185e-01 5.24101794e-01
-6.12079322e-01 8.40172529e-01 7.25643098e-01 -1.77722666e-02
9.31115806e-01 -1.37321293e-01 2.15931252e-01 -4.17235702e-01
-5.41245699e-01 2.56480455e-01 2.67947763e-01 3.43755007e-01
4.48796391e-01 3.09158713e-01 -9.17634070e-01 1.39817441e+00
-6.59187973e-01 3.26826245e-01 5.04939318e-01 7.27042615e-01
4.44174372e-02 -1.01353121e+00 -1.05375230e-01 1.24259925e+00
-6.94025815e-01 -8.01496431e-02 -1.71192184e-01 1.00236380e+00
6.41005754e-01 1.23347414e+00 3.59959751e-01 -6.04584455e-01
4.94663388e-01 3.04680943e-01 9.38608944e-02 -7.51556039e-01
-7.29198039e-01 -3.39590609e-02 3.13988149e-01 -3.09147775e-01
-3.94298345e-01 -9.90095198e-01 -1.27165449e+00 -7.65817463e-02
-2.16198325e-01 4.22414571e-01 3.25864583e-01 6.37002707e-01
4.72240090e-01 4.21684384e-01 3.16818297e-01 -5.28890669e-01
-1.30640879e-01 -1.20265102e+00 -8.81287456e-01 3.14198673e-01
2.45434582e-01 -8.61544430e-01 -6.32771969e-01 -1.58862323e-01] | [13.457074165344238, 3.49280047416687] |
3301d5d8-9aed-42ef-bdce-af33b46910f2 | cx-db8-a-queryable-extractive-summarizer-and | 2012.03942 | null | https://arxiv.org/abs/2012.03942v1 | https://arxiv.org/pdf/2012.03942v1.pdf | CX DB8: A queryable extractive summarizer and semantic search engine | Competitive Debate's increasingly technical nature has left competitors looking for tools to accelerate evidence production. We find that the unique type of extractive summarization performed by competitive debaters - summarization with a bias towards a particular target meaning - can be performed using the latest innovations in unsupervised pre-trained text vectorization models. We introduce CX_DB8, a queryable word-level extractive summarizer and evidence creation framework, which allows for rapid, biasable summarization of arbitarily sized texts. CX_DB8s usage of the embedding framework Flair means that as the underlying models improve, CX_DB8 will also improve. We observe that CX_DB8 also functions as a semantic search engine, and has application as a supplement to traditional "find" functionality in programs and webpages. CX_DB8 is currently used by competitive debaters and is made available to the public at https://github.com/Hellisotherpeople/CX_DB8 | ['Allen Roush'] | 2020-12-07 | null | null | null | null | ['query-based-extractive-summarization', 'extractive-document-summarization'] | ['natural-language-processing', 'natural-language-processing'] | [ 9.07159820e-02 5.70988178e-01 -7.26850569e-01 -2.41480153e-02
-1.26878703e+00 -9.24887061e-01 8.30353498e-01 6.96307659e-01
-5.65044820e-01 9.18645918e-01 1.25698745e+00 -5.21474898e-01
5.97418360e-02 -6.21560574e-01 -5.50466359e-01 -3.64723831e-01
4.50810939e-01 4.65457261e-01 1.09898515e-01 -5.58030605e-01
9.30032790e-01 4.31684740e-02 -1.32106268e+00 4.82635498e-01
8.89973402e-01 2.38290951e-01 1.95148230e-01 1.08672678e+00
-5.08146048e-01 7.12469935e-01 -9.71550167e-01 -5.28799772e-01
-1.31462529e-01 -4.30415034e-01 -1.19939947e+00 -6.74271941e-01
6.32694304e-01 -1.23868473e-02 -4.54484552e-01 1.02954781e+00
6.70963109e-01 -9.40556154e-02 4.66334581e-01 -9.03209388e-01
-6.33267760e-01 1.37085104e+00 -4.24551994e-01 7.22028792e-01
5.15490651e-01 1.52278692e-01 1.66843867e+00 -6.62832618e-01
1.20264649e+00 1.02953100e+00 4.93393749e-01 5.84939301e-01
-1.16875899e+00 -5.24603784e-01 -1.18751079e-01 2.37368956e-01
-7.29562879e-01 -4.72287208e-01 9.01892662e-01 -2.03972310e-01
1.16657209e+00 7.92653024e-01 8.77731740e-01 1.43759203e+00
2.99531907e-01 1.25125742e+00 7.14583516e-01 -5.88671327e-01
2.35759497e-01 3.06097388e-01 5.49420893e-01 5.13552845e-01
7.42013335e-01 -5.95687509e-01 -1.11119735e+00 -4.45115298e-01
1.05127860e-02 -5.61509788e-01 -5.33259690e-01 1.92160755e-01
-1.32679772e+00 1.09945405e+00 1.76093597e-02 5.34390509e-01
-4.00918424e-01 3.98792744e-01 9.43264544e-01 4.76177871e-01
9.33926463e-01 1.20245123e+00 -4.93697971e-01 -7.96571970e-01
-1.31570661e+00 6.93285286e-01 1.19685340e+00 6.88006163e-01
3.43977094e-01 -1.25746831e-01 -3.46486300e-01 8.92334044e-01
7.52015486e-02 2.86915630e-01 8.39598715e-01 -1.27349353e+00
6.14925086e-01 6.03625655e-01 -7.34615475e-02 -9.27434087e-01
-5.07404134e-02 -7.24749148e-01 -2.75021881e-01 -1.76060647e-01
-1.00646064e-01 -2.16880858e-01 -2.87748933e-01 1.34609044e+00
2.02263430e-01 -3.93018186e-01 1.38377905e-01 3.76908153e-01
1.29756987e+00 1.03258038e+00 -7.85847381e-02 -3.96712013e-02
1.31359088e+00 -1.02638221e+00 -8.92500341e-01 -2.40882054e-01
9.19833899e-01 -1.03880572e+00 1.20430517e+00 4.28046703e-01
-1.30766165e+00 -6.19449690e-02 -1.49479568e+00 -4.49836731e-01
-5.12353837e-01 -4.48634923e-02 5.52307963e-01 5.11227608e-01
-1.22808576e+00 7.23819196e-01 -5.15028477e-01 -5.09780407e-01
5.58392525e-01 -1.88012615e-01 -2.92048544e-01 1.21801160e-01
-9.77077305e-01 1.12836313e+00 3.82138819e-01 -4.82211590e-01
-5.40090561e-01 -1.02239025e+00 -8.38776350e-01 1.53156310e-01
6.02318943e-01 -8.38208020e-01 1.59490848e+00 -6.34314239e-01
-1.32887042e+00 9.98012245e-01 -1.70363382e-01 -5.43530822e-01
3.21991444e-01 -6.86575532e-01 -1.61645994e-01 3.30839217e-01
4.31452096e-01 4.51813698e-01 5.64465940e-01 -9.07557487e-01
-4.49341774e-01 3.54445493e-03 -1.01055369e-01 2.36806020e-01
-5.82530081e-01 2.64289856e-01 -1.33163601e-01 -7.83876896e-01
-1.64252132e-01 -5.71308196e-01 -4.94441483e-03 -3.60841930e-01
-7.12000251e-01 -6.47632360e-01 6.11426294e-01 -9.16887820e-01
1.56292105e+00 -1.76316404e+00 3.83228660e-01 -3.33307721e-02
4.50159967e-01 3.27910513e-01 -1.08354636e-01 1.07594538e+00
2.01000199e-01 6.73129201e-01 -1.76062539e-01 -1.61818340e-02
3.51706594e-01 -9.82791036e-02 -5.49943388e-01 2.42263645e-01
2.03937620e-01 1.07583988e+00 -1.20998406e+00 -6.54545784e-01
-3.71316969e-02 3.38592976e-02 -5.56552410e-01 -1.01313613e-01
-4.69659984e-01 -1.96182236e-01 -6.39935613e-01 5.03645837e-01
2.20864698e-01 -3.62280130e-01 2.02585652e-01 2.04347894e-02
-4.57918674e-01 7.83554971e-01 -5.97269118e-01 2.05495691e+00
-3.38438243e-01 1.07921147e+00 -6.06126804e-03 -8.43258739e-01
6.99215055e-01 2.18488872e-01 2.20846221e-01 -2.85253137e-01
2.33286858e-01 3.09265643e-01 -2.20704272e-01 -3.37257117e-01
1.30496442e+00 1.62707686e-01 -3.35816056e-01 7.62805402e-01
2.84566551e-01 -6.80478036e-01 6.82801425e-01 1.02955616e+00
1.41796875e+00 3.99205042e-03 4.38352197e-01 -6.47440553e-01
2.23766536e-01 6.23395622e-01 2.33465508e-01 8.27656448e-01
1.57994956e-01 3.28393489e-01 4.52467918e-01 -1.72642261e-01
-1.24624217e+00 -8.92841220e-01 -5.70175461e-02 1.19533956e+00
-8.76747295e-02 -1.10245073e+00 -6.05853975e-01 -6.42689407e-01
5.50648719e-02 1.38523817e+00 -5.43462873e-01 -2.28050843e-01
-6.11336887e-01 -5.43629825e-01 6.84792042e-01 2.89266884e-01
1.93323597e-01 -1.15237164e+00 -6.93338215e-01 3.07938665e-01
-4.31857646e-01 -4.52877909e-01 -5.21619320e-01 2.75675833e-01
-7.98961222e-01 -8.81315053e-01 -7.11351514e-01 -4.44655120e-01
2.13493183e-01 1.73122376e-01 1.37580907e+00 1.52019635e-01
-3.49824250e-01 6.41679406e-01 -5.47345161e-01 -7.88003743e-01
-8.48842204e-01 6.42026186e-01 -3.17566037e-01 -6.90438688e-01
3.29014599e-01 -5.42337000e-01 -7.14090884e-01 -5.18335462e-01
-1.16444969e+00 5.93166985e-02 4.74364132e-01 7.22689509e-01
8.90989602e-02 -4.91841525e-01 7.19107032e-01 -1.18241990e+00
1.46829164e+00 -5.09455681e-01 -1.47537008e-01 1.74125239e-01
-7.86978245e-01 3.05137604e-01 2.06103742e-01 -6.13961853e-02
-1.03963172e+00 -8.43439400e-01 -3.83001029e-01 2.51267195e-01
3.90911371e-01 9.49823201e-01 3.25280041e-01 5.46731591e-01
1.08994567e+00 2.14544103e-01 4.06521261e-02 -5.63734233e-01
8.78278911e-01 6.23856604e-01 4.39617962e-01 -5.27452230e-01
7.53009319e-01 4.24676836e-01 -5.49585879e-01 -1.05391121e+00
-7.80289531e-01 -7.91494906e-01 1.05422139e-01 -1.78465575e-01
6.80534720e-01 -7.08599985e-01 -2.63729513e-01 -1.53969809e-01
-1.32689655e+00 -2.38304466e-01 -8.59296858e-01 2.33794853e-01
-3.42307925e-01 4.11237985e-01 -7.30182767e-01 -2.69202292e-01
-1.04137027e+00 -5.53606927e-01 9.49827790e-01 4.68546659e-01
-1.06689429e+00 -9.76624668e-01 5.86628914e-01 5.15629590e-01
4.40285295e-01 5.34255989e-02 8.85966659e-01 -1.06725705e+00
-1.55855462e-01 -3.75031710e-01 9.91515666e-02 4.29733425e-01
-1.76797986e-01 5.23564279e-01 -5.68901181e-01 -5.81888184e-02
-8.15824121e-02 -4.84495372e-01 1.02963400e+00 3.55298162e-01
8.57538342e-01 -7.57668078e-01 -2.44198948e-01 -5.97806424e-02
1.15391827e+00 -2.80804753e-01 4.63405401e-01 7.35415101e-01
2.11884692e-01 4.82860893e-01 3.99855614e-01 3.88295680e-01
2.45589986e-01 2.35833183e-01 -3.83506007e-02 2.98913330e-01
-3.89387339e-01 -4.32234526e-01 5.05769432e-01 1.73271453e+00
3.21502872e-02 -3.54539216e-01 -8.17664921e-01 7.57002294e-01
-1.78124022e+00 -1.29087031e+00 -8.78193900e-02 1.69557118e+00
1.28986335e+00 3.10846359e-01 -1.21930115e-01 -7.19588324e-02
4.31327641e-01 6.52139008e-01 -3.14602137e-01 -9.44262862e-01
-2.49890476e-01 6.85685039e-01 5.51864088e-01 4.68884587e-01
-5.00666857e-01 8.38769138e-01 6.29810143e+00 1.33240473e+00
-9.25916255e-01 2.83851445e-01 3.83466244e-01 -3.91631603e-01
-9.39701796e-01 1.76732913e-01 -7.69981265e-01 4.47719246e-01
1.05769837e+00 -1.11118734e+00 -1.96720406e-01 1.03845704e+00
1.00482866e-01 -1.98565528e-01 -8.10022295e-01 5.97147703e-01
2.40961254e-01 -2.23602796e+00 8.99963006e-02 -1.43557146e-01
7.83584893e-01 3.04670841e-01 -1.42373424e-02 4.45390701e-01
6.31864727e-01 -7.46566415e-01 7.61891663e-01 2.74262786e-01
4.44485068e-01 -5.82141697e-01 6.70035005e-01 2.35652342e-01
-5.19276381e-01 1.75057739e-01 -3.84540081e-01 1.69126898e-01
3.12077701e-01 6.30460739e-01 -7.73798525e-01 6.14097297e-01
4.40268666e-01 7.54479587e-01 -5.99592268e-01 9.91639674e-01
-4.69543159e-01 1.03757834e+00 -9.18075368e-02 -6.47670686e-01
3.51484507e-01 5.25471941e-02 1.23416877e+00 1.62253237e+00
2.47610584e-01 -1.60971373e-01 -2.08034128e-01 6.45699680e-01
-5.31491339e-01 4.38600749e-01 -4.62539643e-01 -5.33047855e-01
6.56401932e-01 1.29009008e+00 -6.79446340e-01 -8.44898224e-01
8.93871933e-02 7.80447185e-01 1.09551176e-01 7.13183731e-02
-6.13701284e-01 -6.39159024e-01 2.41067469e-01 -2.11098175e-02
1.65849492e-01 4.18872871e-02 -2.92501748e-01 -1.35387170e+00
-2.59554219e-02 -1.28878677e+00 3.92823219e-01 -8.35067093e-01
-1.04107916e+00 4.68927413e-01 9.91878808e-02 -6.42125666e-01
-1.79464310e-01 -2.07371235e-01 -9.42699432e-01 4.36206639e-01
-1.10731554e+00 -7.08262444e-01 7.91339427e-02 -6.60925135e-02
1.18294775e+00 -1.32339180e-01 6.87233210e-01 -1.65186584e-01
-3.22143912e-01 3.67079765e-01 3.88564050e-01 -1.73389599e-01
6.83168590e-01 -1.41903627e+00 4.06564862e-01 6.93976939e-01
6.40624106e-01 8.67972016e-01 1.42324698e+00 -6.39897883e-01
-1.37852490e+00 -6.68902755e-01 1.33079517e+00 -7.01844096e-01
1.08431363e+00 -1.16947167e-01 -6.97743773e-01 5.11499465e-01
1.01153588e+00 -9.46131110e-01 8.52370262e-01 3.87400508e-01
-3.76885980e-01 2.80497149e-02 -7.59884775e-01 8.94448400e-01
7.86737740e-01 -6.25192344e-01 -1.32135236e+00 6.75655067e-01
1.02897727e+00 -1.71776623e-01 -7.16453731e-01 8.44567269e-03
4.60426003e-01 -5.03358483e-01 6.24087632e-01 -7.27548897e-01
1.13197303e+00 8.65263790e-02 8.37702304e-02 -1.49458635e+00
-2.84307487e-02 -1.03531349e+00 -2.06041917e-01 1.24551821e+00
8.72504234e-01 -6.45595074e-01 7.09732592e-01 3.05085540e-01
-4.96751785e-01 -6.18672132e-01 -9.12808836e-01 -5.69437087e-01
4.03647453e-01 -3.71430814e-01 9.25886780e-02 1.00606918e+00
6.07957482e-01 5.84619761e-01 6.40652180e-02 -6.35812283e-01
4.17921782e-01 1.16522238e-01 8.36809099e-01 -9.48282778e-01
-3.11330706e-01 -8.81124675e-01 -5.77801578e-02 -1.14644086e+00
-5.03774956e-02 -1.22170651e+00 -3.51382233e-02 -1.79512036e+00
4.31680560e-01 1.03549331e-01 -1.60099164e-01 2.42560640e-01
-1.45785958e-01 1.11703992e-01 9.52578261e-02 1.55130222e-01
-6.75668359e-01 5.25060534e-01 8.86882126e-01 -4.28138018e-01
-1.38484597e-01 -3.92876893e-01 -1.30828619e+00 5.58970809e-01
1.04469514e+00 -6.84906602e-01 -2.63838917e-01 -1.20840348e-01
8.14206183e-01 -3.75463784e-01 6.71017990e-02 -6.55101657e-01
3.35903287e-01 1.48135632e-01 -1.31129816e-01 -6.70228243e-01
1.09189384e-01 2.12825518e-02 -3.96805443e-02 5.00439167e-01
-8.46322834e-01 3.14968288e-01 3.30603331e-01 3.52526516e-01
-2.04145029e-01 -7.32642949e-01 1.21286198e-01 -3.27271581e-01
-4.27084953e-01 -2.83769697e-01 -7.73927391e-01 5.22864044e-01
5.14064848e-01 -1.61365807e-01 -8.24767172e-01 -4.88371760e-01
-3.46338630e-01 1.71274915e-01 3.17335218e-01 4.08870429e-01
5.84332228e-01 -9.80955720e-01 -1.19123089e+00 -6.01813912e-01
1.34672135e-01 -3.33613187e-01 9.04620811e-02 7.07024574e-01
-7.44003654e-01 5.27139544e-01 4.52484824e-02 -1.79778501e-01
-1.46355855e+00 1.63080916e-01 -5.41495740e-01 -4.78438616e-01
-8.37101638e-01 1.00833070e+00 -3.80370170e-01 -1.99365899e-01
-5.40837273e-02 -2.50940949e-01 -7.00922310e-02 4.04254287e-01
5.90431571e-01 4.97207195e-01 -1.04539245e-01 -1.81067020e-01
-3.52679119e-02 -6.02992624e-02 -4.98909026e-01 -5.03043711e-01
1.78693235e+00 8.41890555e-03 -4.15452540e-01 7.18354464e-01
1.12775207e+00 5.44106960e-01 -4.49690610e-01 -1.79571402e-03
3.36882442e-01 -8.29272792e-02 1.70144901e-01 -8.94995630e-01
-5.38519144e-01 4.70567882e-01 -2.37075433e-01 2.70977616e-01
4.66743290e-01 2.94381648e-01 7.25944936e-01 4.81655747e-01
-1.89940073e-02 -1.41694319e+00 1.89003512e-01 3.57422143e-01
1.20830405e+00 -8.41958702e-01 5.56406319e-01 -4.04558145e-02
-6.24700129e-01 1.06869733e+00 4.10822406e-03 -1.74614057e-01
2.94041395e-01 -6.18829951e-03 -1.33623733e-04 -6.83846474e-01
-1.09252667e+00 2.48316914e-01 2.57224560e-01 9.28416923e-02
6.54014826e-01 1.97340967e-03 -1.01130247e+00 3.80278289e-01
-7.11186409e-01 -2.12438926e-01 8.55839550e-01 1.03782547e+00
-7.50066757e-01 -1.15696728e+00 -1.07608534e-01 8.01470578e-01
-7.15291560e-01 -5.67068636e-01 -7.33036935e-01 9.47238505e-01
-4.23274785e-01 8.83249998e-01 -2.59918213e-01 -9.31028053e-02
3.75598520e-02 2.78044492e-01 2.69103229e-01 -7.72058785e-01
-9.96024549e-01 -3.22219193e-01 8.40124369e-01 -6.02942705e-01
-2.25816935e-01 -6.43100917e-01 -9.89085674e-01 -6.06423736e-01
-4.07523334e-01 8.37727547e-01 9.50042069e-01 5.57948947e-01
4.32348043e-01 4.64729279e-01 1.08586736e-01 -5.81994593e-01
-7.47692287e-01 -1.12756431e+00 -2.45811164e-01 1.94591686e-01
5.51810209e-03 -7.82435164e-02 -5.65677345e-01 -7.68002644e-02] | [12.334686279296875, 9.51749324798584] |
d1d6ea8a-5934-4687-a743-8819e5ecfe3a | model-based-clustering-with-hidden-markov | 1312.7024 | null | http://arxiv.org/abs/1312.7024v1 | http://arxiv.org/pdf/1312.7024v1.pdf | Model-based clustering with Hidden Markov Model regression for time series with regime changes | This paper introduces a novel model-based clustering approach for clustering
time series which present changes in regime. It consists of a mixture of
polynomial regressions governed by hidden Markov chains. The underlying hidden
process for each cluster activates successively several polynomial regimes
during time. The parameter estimation is performed by the maximum likelihood
method through a dedicated Expectation-Maximization (EM) algorithm. The
proposed approach is evaluated using simulated time series and real-world time
series issued from a railway diagnosis application. Comparisons with existing
approaches for time series clustering, including the stand EM for Gaussian
mixtures, $K$-means clustering, the standard mixture of regression models and
mixture of Hidden Markov Models, demonstrate the effectiveness of the proposed
approach. | ['Gérard Govaert', 'Allou Samé', 'Faicel Chamroukhi', 'Patrice Aknin'] | 2013-12-25 | null | null | null | null | ['time-series-clustering'] | ['time-series'] | [-1.07246563e-01 -3.29066366e-01 6.31469935e-02 -2.40822554e-01
-6.11383080e-01 -1.55332223e-01 5.24367273e-01 -2.64022369e-02
-2.47702703e-01 4.16104883e-01 -3.20202738e-01 -3.91537398e-01
-5.75345159e-01 -5.38288593e-01 7.65178353e-02 -1.34465981e+00
-4.56136853e-01 1.06884396e+00 2.08821639e-01 1.07695453e-01
1.90860227e-01 7.45705664e-01 -1.30212379e+00 -1.12959124e-01
5.74014485e-01 4.88342285e-01 1.49222419e-01 1.03539157e+00
-7.63414651e-02 9.47237968e-01 -8.94018352e-01 2.77226001e-01
-2.14882359e-01 -3.13854218e-01 -4.34107184e-01 5.73376656e-01
-1.17039287e+00 4.10776585e-01 -3.43341410e-01 5.73435307e-01
2.21983925e-01 6.40879095e-01 1.17108107e+00 -1.64053714e+00
-1.15324497e-01 6.60719752e-01 -6.84368372e-01 3.61410499e-01
3.46928723e-02 3.78834307e-02 6.13854229e-02 -3.78396630e-01
3.84135276e-01 1.20607460e+00 7.66858995e-01 1.63784996e-01
-1.67032480e+00 -5.24007261e-01 -1.47785038e-01 3.20167780e-01
-1.83596802e+00 -7.01377615e-02 8.10311794e-01 -6.86622202e-01
1.25440526e+00 2.04514071e-01 5.03584623e-01 6.89291477e-01
4.94812787e-01 5.83770394e-01 1.44412017e+00 -4.40214932e-01
6.56947434e-01 -2.88378000e-02 4.24406528e-01 -9.01887566e-02
-1.79631099e-01 -6.97885305e-02 3.63162272e-02 -7.03503370e-01
6.85424507e-01 3.14689308e-01 1.09103888e-01 3.69453430e-02
-1.06320667e+00 6.37255907e-01 -5.11052430e-01 6.07049465e-01
-6.97561383e-01 1.35030895e-01 1.90514103e-01 3.91579270e-01
4.12609100e-01 -3.48667115e-01 -5.29893696e-01 -2.79799342e-01
-1.41755426e+00 6.17468171e-02 9.15795267e-01 1.14190924e+00
4.33542848e-01 4.23711091e-01 3.04037243e-01 4.35330242e-01
8.31440687e-01 7.38763988e-01 4.18930680e-01 -9.69178677e-01
-1.42270863e-01 2.29149967e-01 2.18628943e-01 -6.27525568e-01
-4.31488305e-01 -2.00325787e-01 -8.91909719e-01 -8.66256002e-03
8.00221413e-02 -3.53452623e-01 -1.00685835e+00 1.22464979e+00
4.71342117e-01 4.97686803e-01 1.99418217e-01 4.28544432e-02
1.93924859e-01 1.08506191e+00 8.29694942e-02 -1.28259051e+00
9.78907406e-01 -4.94183093e-01 -1.38960564e+00 9.87525582e-01
1.82746783e-01 -1.01400602e+00 -1.28680646e-01 6.27460003e-01
-1.08507597e+00 -6.48917973e-01 -2.84440696e-01 9.73177791e-01
-2.23282591e-01 -1.91601127e-01 1.76365450e-01 5.49675584e-01
-1.11066556e+00 6.30528390e-01 -1.62576091e+00 -4.80645925e-01
-3.64812851e-01 7.22265661e-01 7.81417787e-02 4.29984361e-01
-8.66695344e-01 6.43560231e-01 4.33631241e-01 4.22912359e-01
-9.91229594e-01 -2.14427650e-01 -4.16612923e-01 -2.75323123e-01
-1.40333056e-01 -3.39009225e-01 1.27987623e+00 -5.27451992e-01
-1.69207501e+00 5.45219064e-01 -5.30714512e-01 -2.56702989e-01
3.93216312e-01 1.17583819e-01 -1.05475354e+00 2.92153627e-01
-2.14711577e-01 -1.55753255e-01 1.22538126e+00 -1.29629982e+00
-4.67468381e-01 -1.25429720e-01 -9.61092830e-01 -2.85493195e-01
3.94824982e-01 5.20573437e-01 -2.51690090e-01 -5.51477909e-01
5.04974782e-01 -1.15327358e+00 -9.71431494e-01 -1.12395871e+00
-3.29412341e-01 -4.81020927e-01 1.11302948e+00 -7.71697462e-01
1.66772127e+00 -2.10660219e+00 1.90856740e-01 7.24754333e-01
-8.62489864e-02 -6.51230276e-01 8.69669497e-01 1.21874380e+00
-4.43866819e-01 -1.27894565e-01 -4.69999403e-01 -4.99446392e-01
-1.79521684e-02 4.52721745e-01 -1.86628953e-01 9.25387263e-01
-2.13106275e-01 6.01511961e-03 -6.12835884e-01 -7.69071758e-01
4.65049565e-01 5.75784147e-01 6.33941144e-02 2.25363657e-01
7.19595030e-02 7.32578278e-01 -2.00485095e-01 4.65746313e-01
5.49819291e-01 -9.40003172e-02 1.72527149e-01 1.29632115e-01
-4.83668596e-01 -5.85136771e-01 -1.50493538e+00 1.05733621e+00
2.66922712e-01 5.48272789e-01 -1.00563012e-01 -1.18644214e+00
1.06790161e+00 1.16929245e+00 1.11748683e+00 3.31536502e-01
3.37283224e-01 1.31184403e-02 -8.05300698e-02 -5.94120622e-01
1.48041129e-01 -4.16736096e-01 -2.58322090e-01 6.41451180e-01
9.08417180e-02 -9.56022143e-02 -4.92141023e-02 3.48743163e-02
1.09936035e+00 1.27879798e-01 2.27967843e-01 -3.27996254e-01
4.30796504e-01 4.37471062e-01 5.71079552e-01 5.57005525e-01
-2.96886981e-01 7.73330405e-02 7.57124275e-02 6.40041828e-02
-1.15909755e+00 -1.16944528e+00 -2.69061536e-01 5.34205675e-01
-3.57603222e-01 -1.94346577e-01 -5.75775266e-01 1.63380444e-01
-4.84269410e-01 8.38692546e-01 -4.95573908e-01 6.43738285e-02
-5.19593298e-01 -1.28878057e+00 2.74180502e-01 1.96319267e-01
1.80357069e-01 -1.03170490e+00 -4.33620036e-01 7.22251832e-01
-1.70961142e-01 -1.01052535e+00 2.93634444e-01 3.75211596e-01
-1.42179751e+00 -5.58397472e-01 -5.86194634e-01 -6.34929061e-01
6.38202965e-01 8.61664116e-03 8.36097598e-01 -4.26631004e-01
-2.88600296e-01 7.83316076e-01 -3.41380864e-01 -3.50965321e-01
-7.87059486e-01 -3.50953430e-01 1.54254749e-01 2.70133644e-01
7.53034532e-01 -9.05787706e-01 -4.24485952e-01 3.71144861e-01
-7.55015612e-01 -5.10381997e-01 3.64480823e-01 4.50631201e-01
7.39279211e-01 1.45024884e+00 8.55683312e-02 -5.26360810e-01
5.19816458e-01 -9.60960329e-01 -5.76056063e-01 -2.78101508e-02
-6.22485518e-01 -3.27514976e-01 1.93201527e-01 -7.79484928e-01
-1.25465167e+00 1.72322914e-01 3.25112492e-01 -8.43858361e-01
-8.29885781e-01 5.29961944e-01 4.68957610e-02 5.75312436e-01
1.88113183e-01 5.28542876e-01 1.04573324e-01 -5.27677000e-01
1.14411168e-01 8.79256427e-01 5.24946928e-01 -4.18772489e-01
8.72716606e-01 8.57712805e-01 3.14050503e-02 -1.17381978e+00
1.23081759e-01 -1.10926962e+00 -9.43509459e-01 -7.82510221e-01
1.27521229e+00 -9.25195813e-01 -6.65008903e-01 9.78484690e-01
-7.52456129e-01 -2.38571495e-01 -1.39845565e-01 9.15169239e-01
-9.02708769e-01 4.13261533e-01 -1.18055785e+00 -1.55712187e+00
2.60084886e-02 -7.68192470e-01 7.12226331e-01 1.27889708e-01
-4.68196720e-01 -1.59745181e+00 5.53993285e-01 1.28681794e-01
1.43737406e-01 6.76656544e-01 8.25172722e-01 -5.65165520e-01
-1.52844876e-01 -5.45982659e-01 7.14149415e-01 8.93053189e-02
5.63017465e-02 9.55955863e-01 -6.69363916e-01 -3.44994545e-01
6.08130395e-01 7.41043329e-01 7.41526261e-02 8.57375324e-01
5.04382074e-01 1.34725168e-01 -7.03205109e-01 -9.58230793e-02
1.66190112e+00 7.78093934e-01 6.65332556e-01 2.32742801e-01
4.00207460e-01 6.05224252e-01 6.11511648e-01 8.79397869e-01
2.84851849e-01 1.53916642e-01 -1.11552492e-01 -1.81375578e-01
9.04074728e-01 5.22507966e-01 4.08434689e-01 1.56651056e+00
-6.81568027e-01 6.73011169e-02 -1.33967781e+00 6.68771267e-01
-2.06414461e+00 -1.66009593e+00 -1.00448561e+00 1.94987297e+00
5.41960657e-01 3.25903594e-01 5.48002601e-01 7.51596570e-01
1.08603823e+00 -4.76152986e-01 -1.35870323e-01 -4.48005021e-01
-1.97975500e-03 2.70863205e-01 5.31648636e-01 2.90754169e-01
-1.00173795e+00 6.38022363e-01 7.72001553e+00 7.60820687e-01
-6.82778001e-01 2.98566490e-01 5.50085455e-02 3.00325066e-01
1.89923033e-01 3.55453223e-01 -5.13153970e-01 4.44076270e-01
1.68884563e+00 -2.27923930e-01 5.32901697e-02 4.88942087e-01
1.04675555e+00 -2.19659209e-01 -5.09947002e-01 8.49501133e-01
-3.11432093e-01 -5.03809512e-01 -3.14947814e-01 3.62663597e-01
8.68468463e-01 -3.65220644e-02 -1.42733455e-01 3.14206630e-01
7.30212331e-01 -8.10881495e-01 4.76988375e-01 1.05383646e+00
-5.93059584e-02 -1.00698268e+00 5.17851591e-01 5.39671361e-01
-1.40337110e+00 -1.71035811e-01 -7.02637881e-02 -5.73491566e-02
5.69380045e-01 7.09808528e-01 -6.43556118e-01 6.63810730e-01
8.58455896e-01 4.94787961e-01 3.10796835e-02 9.95113015e-01
2.22057223e-01 1.18846083e+00 -4.99939293e-01 2.00117692e-01
1.99125454e-01 -5.09531796e-01 7.96995401e-01 1.33037472e+00
3.67090583e-01 3.20898503e-01 2.48184636e-01 5.64271927e-01
1.24790466e+00 -3.65574509e-02 -3.67959946e-01 3.95428799e-02
6.06528759e-01 1.02973831e+00 -1.36581004e+00 -7.34156728e-01
-2.72214234e-01 6.46907866e-01 -6.85819149e-01 6.87083840e-01
-1.06378376e+00 -1.38336807e-01 -1.67925265e-02 6.26714453e-02
5.02482176e-01 -7.12122560e-01 1.56964332e-01 -7.89629936e-01
-7.07501769e-01 -5.74076772e-01 6.58064485e-01 -8.23121428e-01
-1.43834782e+00 7.72641897e-01 8.65043283e-01 -1.63267946e+00
-5.51438451e-01 -1.80314630e-02 -7.93713331e-01 8.08995962e-01
-6.09525442e-01 -8.59101236e-01 7.59589225e-02 1.35674131e+00
6.89799488e-01 -1.92749485e-01 7.25271642e-01 1.51264772e-01
-7.58416653e-01 -3.04915845e-01 6.98468208e-01 -8.45495686e-02
2.40545526e-01 -1.39561164e+00 -1.23182490e-01 7.24046350e-01
-6.32056236e-01 7.77294040e-01 1.48027968e+00 -7.80810595e-01
-1.02578235e+00 -8.01873147e-01 7.60121107e-01 -2.81460881e-01
9.27396595e-01 7.56877288e-02 -9.55565989e-01 8.45027983e-01
4.96052265e-01 -7.18762219e-01 1.33831012e+00 -1.85301766e-01
4.88973826e-01 2.91209936e-01 -9.79223430e-01 2.60009080e-01
8.98467004e-02 -4.35939550e-01 -8.29217494e-01 4.12549138e-01
2.95334607e-01 3.60114872e-01 -1.70339203e+00 3.69615555e-01
1.55781209e-01 -5.33041775e-01 5.82731843e-01 -4.30402398e-01
-2.89121687e-01 -5.84537089e-01 -2.56087124e-01 -9.46595907e-01
-7.36505389e-01 -1.15036702e+00 -3.55047017e-01 1.31615794e+00
2.83324331e-01 -4.15479928e-01 4.47588474e-01 6.43639028e-01
5.89052849e-02 -1.58735231e-01 -1.03466570e+00 -9.01783407e-01
-1.99185595e-01 -6.55091524e-01 4.47341293e-01 1.40870345e+00
1.38510227e-01 -3.02063301e-02 -2.84633934e-01 5.67153573e-01
1.28582466e+00 6.98799193e-02 6.75182879e-01 -1.38157272e+00
-5.03379822e-01 -2.27575451e-01 -5.29669344e-01 -4.45657462e-01
1.92426741e-02 -1.36728555e-01 2.00806797e-01 -1.40031612e+00
2.65405811e-02 -2.37877771e-01 -4.14061427e-01 -2.49016687e-01
1.09962352e-01 -3.26207429e-01 -2.55701900e-01 9.79340494e-01
-4.88533288e-01 2.23677740e-01 4.51503664e-01 2.82023996e-01
-4.66137230e-01 6.63065195e-01 4.65941340e-01 7.32361913e-01
9.62930262e-01 -8.92505765e-01 -4.24578398e-01 4.27427500e-01
-3.30100566e-01 8.24218035e-01 4.95957434e-02 -1.14566505e+00
4.66668814e-01 -2.82467604e-01 1.17513806e-01 -1.38624561e+00
2.94111371e-01 -1.30946517e+00 1.28716004e+00 7.22281814e-01
1.69458345e-01 6.61499441e-01 -1.24721967e-01 8.65315795e-01
-4.37692225e-01 4.07178663e-02 7.49602437e-01 -1.30181789e-01
-5.39489210e-01 1.17162399e-01 -1.55083334e+00 -8.73240411e-01
1.51428902e+00 -5.23382604e-01 3.52248549e-01 -5.04736602e-01
-1.83607304e+00 1.15439489e-01 1.93145961e-01 -1.11012988e-01
4.46446180e-01 -1.12988460e+00 -5.57906568e-01 -1.41814560e-01
-3.49475443e-01 -2.31262147e-01 5.05157292e-01 1.53877902e+00
-4.55541223e-01 3.25354673e-02 -1.44476056e-01 -1.21005094e+00
-1.38226008e+00 9.85466599e-01 4.87523794e-01 -4.10141319e-01
-5.10287225e-01 1.90036088e-01 -5.63317358e-01 -2.20561564e-01
-2.38415319e-02 -1.44759402e-01 -5.02898157e-01 6.40222803e-02
1.43271893e-01 9.26079154e-01 -4.10058975e-01 -1.02933264e+00
-1.05864175e-01 5.91497481e-01 3.42693686e-01 -7.66620100e-01
1.36050761e+00 -5.98091185e-01 -4.40123349e-01 1.49846840e+00
8.74091744e-01 -5.21808863e-01 -9.00051892e-01 -7.01855049e-02
3.91447425e-01 1.77420467e-01 -1.21060394e-01 3.23714106e-04
-6.32661045e-01 5.45580506e-01 7.12577224e-01 7.44431973e-01
1.30695355e+00 -1.55003360e-02 3.25558066e-01 1.03649765e-01
3.66604030e-01 -1.42989361e+00 -5.10720789e-01 -7.47280642e-02
1.76053390e-01 -5.83495855e-01 -1.23518594e-01 -2.45129317e-01
-4.97911513e-01 1.09317398e+00 -2.54976433e-02 -3.60282868e-01
1.63979590e+00 6.10188007e-01 2.58086234e-01 -3.72165322e-01
-1.18468451e+00 -1.19886085e-01 -3.36521089e-01 6.07371688e-01
2.13567227e-01 2.89356381e-01 -1.31192341e-01 3.41839641e-01
-8.44374672e-03 -1.06106393e-01 4.58506733e-01 1.42843473e+00
-4.57677484e-01 -6.26616001e-01 -1.12590277e+00 6.98837712e-02
-6.50817633e-01 3.98222446e-01 1.02584556e-01 9.50110555e-01
1.46473125e-02 1.54951859e+00 2.55737036e-01 -2.09451467e-01
1.57479987e-01 3.00202489e-01 1.16225220e-01 -2.69987464e-01
-5.25002539e-01 1.27304780e+00 -3.40862244e-01 -1.47175327e-01
-1.02791786e+00 -1.34076965e+00 -1.35098565e+00 -4.28822607e-01
-6.05879545e-01 7.74098217e-01 8.35673571e-01 8.65205884e-01
-2.66591311e-01 6.52117193e-01 1.06509542e+00 -6.62235379e-01
-3.09516847e-01 -1.22516286e+00 -1.00975525e+00 2.09330589e-01
-1.44717529e-01 -3.82198155e-01 -7.13174582e-01 8.12358141e-01] | [7.154639720916748, 3.644670248031616] |
c861debf-401e-4f71-8300-9ba3b4b2a153 | high-accuracy-phishing-detection-based-on | 2004.03960 | null | https://arxiv.org/abs/2004.03960v1 | https://arxiv.org/pdf/2004.03960v1.pdf | High Accuracy Phishing Detection Based on Convolutional Neural Networks | The persistent growth in phishing and the rising volume of phishing websites has led to individuals and organizations worldwide becoming increasingly exposed to various cyber-attacks. Consequently, more effective phishing detection is required for improved cyber defence. Hence, in this paper we present a deep learning-based approach to enable high accuracy detection of phishing sites. The proposed approach utilizes convolutional neural networks (CNN) for high accuracy classification to distinguish genuine sites from phishing sites. We evaluate the models using a dataset obtained from 6,157 genuine and 4,898 phishing websites. Based on the results of extensive experiments, our CNN based models proved to be highly effective in detecting unknown phishing sites. Furthermore, the CNN based approach performed better than traditional machine learning classifiers evaluated on the same dataset, reaching 98.2% phishing detection rate with an F1-score of 0.976. The method presented in this paper compares favourably to the state-of-the art in deep learning based phishing website detection. | ['Suleiman Y. Yerima', 'Mohammed K. Alzaylaee'] | 2020-04-08 | null | null | null | null | ['phishing-website-detection'] | ['adversarial'] | [-1.56249434e-01 -2.62306005e-01 2.04520300e-01 -5.03828973e-02
-4.85843092e-01 -7.90986896e-01 8.80971909e-01 3.93959165e-01
-4.32133764e-01 4.97476161e-01 -3.37729305e-01 -5.50829828e-01
1.36124805e-01 -1.03963399e+00 -4.06504452e-01 -6.61765575e-01
-1.21547095e-01 2.99292296e-01 2.61324167e-01 -3.13292682e-01
9.35785353e-01 4.56912845e-01 -1.17273355e+00 3.39796126e-01
9.35249746e-01 1.08838046e+00 -1.17074922e-01 8.01987648e-01
1.75828993e-01 3.53089094e-01 -8.70870650e-01 -7.16904759e-01
5.66255093e-01 -2.25154273e-02 -6.04636908e-01 -4.66191500e-01
7.12513983e-01 -5.63048184e-01 -5.83774030e-01 1.22185981e+00
4.20538664e-01 -2.12588653e-01 6.31757975e-01 -8.38954926e-01
-9.74102497e-01 1.60505921e-02 -4.35437411e-01 5.04038632e-01
1.80908322e-01 1.80359349e-01 1.09608722e+00 -8.39022040e-01
4.75890994e-01 7.58604944e-01 7.71037340e-01 1.75284624e-01
-1.03488398e+00 -5.51258981e-01 -6.98744237e-01 6.77311420e-01
-1.08721924e+00 4.67893519e-02 4.74437982e-01 -4.96708691e-01
8.47322702e-01 -5.28326258e-02 4.68683273e-01 1.19711173e+00
6.30961299e-01 5.57205498e-01 1.25710559e+00 -3.74177486e-01
2.54049599e-01 4.75131541e-01 6.64055467e-01 4.47451591e-01
9.48211432e-01 3.96238238e-01 1.83715478e-01 -4.78024095e-01
6.58831656e-01 -4.07937448e-03 4.53179069e-02 -2.33295411e-01
-1.06631303e+00 1.47089362e+00 7.19664216e-01 4.50182855e-01
-6.41408503e-01 -2.26105243e-01 9.96856391e-01 3.61597776e-01
4.47907150e-01 4.96218801e-01 -5.18652856e-01 1.23219647e-01
-6.76710665e-01 3.22318524e-01 1.09858453e+00 1.87832713e-01
5.12124658e-01 -2.53257044e-02 4.43965316e-01 7.85255134e-01
2.94262856e-01 3.37345690e-01 2.09883824e-01 -3.12830359e-01
3.06600630e-01 5.64483762e-01 1.15958294e-02 -1.36707830e+00
-1.36716083e-01 -4.54197586e-01 -7.31354237e-01 4.83358830e-01
5.98632872e-01 1.12882843e-02 -8.69637430e-01 8.98536205e-01
8.65722522e-02 5.01929149e-02 -2.41008595e-01 6.63284600e-01
5.46682000e-01 7.93181658e-01 2.31742069e-01 5.05008876e-01
1.56399405e+00 -6.27017319e-01 -3.25630158e-01 -7.22088218e-02
5.66083074e-01 -7.14616299e-01 8.21908653e-01 5.39658010e-01
-4.20473129e-01 -3.46995384e-01 -1.14201081e+00 3.68220866e-01
-1.01684821e+00 -1.29814267e-01 5.60464501e-01 1.41033220e+00
-9.46367145e-01 3.76253277e-01 -1.32195994e-01 -6.96861923e-01
6.40110314e-01 3.30666125e-01 -5.64966857e-01 -2.82127142e-01
-1.64985752e+00 1.20986724e+00 7.76918709e-01 -1.33081913e-01
-1.03543675e+00 -2.71404207e-01 -3.74592751e-01 5.20110011e-01
1.38822123e-01 -1.57625332e-01 9.88881588e-01 -7.10843325e-01
-9.49316204e-01 9.05955374e-01 5.86487591e-01 -6.67808831e-01
1.97379708e-01 -1.27578110e-01 -5.84371090e-01 6.21050596e-02
-6.22302629e-02 7.11903255e-03 6.94499433e-01 -1.24022460e+00
-5.97126365e-01 -6.93906665e-01 1.39941603e-01 -3.96371007e-01
-5.26435792e-01 1.53160170e-01 3.21743876e-01 -1.98567331e-01
-2.30724782e-01 -9.47584689e-01 -1.26644686e-01 -5.07921696e-01
-3.79567832e-01 -2.37719104e-01 1.02981293e+00 -1.30963218e+00
9.31958675e-01 -1.67127788e+00 -6.00744247e-01 4.71940100e-01
5.04892707e-01 1.30563116e+00 -2.24071406e-02 6.31890357e-01
-1.88300923e-01 2.29795143e-01 6.42221048e-02 3.37080866e-01
-8.58335793e-02 -1.73905209e-01 -3.03308871e-02 6.06661141e-01
-3.50939818e-02 8.10160220e-01 -7.67338693e-01 -2.18488891e-02
5.89784443e-01 6.87050045e-01 -1.56779587e-01 2.30037257e-01
3.04765463e-01 -1.93697214e-01 -3.59770656e-01 8.73539746e-01
1.08077562e+00 -2.88424492e-01 6.30525127e-02 1.10077314e-01
2.09704608e-01 2.89345860e-01 -5.43609917e-01 8.29953492e-01
-5.62911391e-01 1.04241920e+00 -1.58060919e-02 -8.37580740e-01
1.21860611e+00 2.89292693e-01 -4.64603566e-02 -8.10880005e-01
5.00136137e-01 4.94729400e-01 2.38993943e-01 -5.34016907e-01
3.29404473e-01 1.57289177e-01 2.20514089e-01 5.27372062e-01
7.77709410e-02 7.37281024e-01 4.37195972e-02 3.41409236e-01
1.23384714e+00 -5.25298059e-01 6.27836287e-01 -3.65521699e-01
9.73651588e-01 1.26831187e-02 -1.49949282e-01 8.07238519e-01
-7.40342975e-01 2.43645579e-01 4.98157054e-01 -1.00871360e+00
-1.54755926e+00 -9.59623337e-01 -3.19188505e-01 8.31875145e-01
-9.03744772e-02 4.48418558e-02 -1.02537346e+00 -1.22720242e+00
-2.64575006e-03 5.83179533e-01 -8.19652438e-01 2.91005280e-02
-5.34389973e-01 -8.38526487e-01 7.07675993e-01 3.05344552e-01
8.82359326e-01 -1.30071545e+00 -1.99647829e-01 2.97290474e-01
-1.26431957e-01 -6.10847235e-01 1.32068112e-01 -1.12183265e-01
-5.34372807e-01 -1.52430058e+00 -8.96489084e-01 -7.34332204e-01
3.79543275e-01 5.86921513e-01 7.97136784e-01 8.75559598e-02
-6.42786562e-01 -1.91881835e-01 -4.23167199e-01 -1.99749559e-01
-6.99594855e-01 5.60204685e-01 -2.59500861e-01 -6.07864223e-02
9.82196331e-01 -1.14143141e-01 -5.97310722e-01 2.63677686e-01
-8.18716109e-01 -8.13107371e-01 8.75024855e-01 9.19230282e-01
-6.47156596e-01 2.12983847e-01 1.03643954e+00 -1.01645958e+00
9.71964359e-01 -9.67899024e-01 -6.44371212e-01 8.81533250e-02
-1.01025343e+00 -2.51132518e-01 5.04650176e-01 -4.09810245e-01
-1.00721323e+00 -1.86478004e-01 -3.62748206e-01 2.21957564e-01
-6.55072272e-01 5.56028008e-01 -8.66554119e-03 -5.87499082e-01
1.14707792e+00 2.29532927e-01 1.15709648e-01 -5.44296384e-01
9.75102708e-02 1.06061554e+00 1.79474682e-01 8.39626342e-02
8.14956009e-01 2.00204775e-01 -1.61715195e-01 -8.98001790e-01
-2.84561515e-01 -1.01335514e+00 -7.68190622e-01 -3.26108783e-01
8.06875765e-01 -4.50855583e-01 -1.12547338e+00 1.01204038e+00
-1.39885581e+00 4.07665044e-01 8.16406906e-01 1.11910701e-01
1.63211271e-01 9.81442332e-01 -9.31440949e-01 -8.70649576e-01
-6.46385312e-01 -7.73731351e-01 5.95248640e-01 8.87177512e-02
-5.69300866e-03 -1.51724505e+00 5.46669602e-01 6.85185611e-01
8.87610495e-01 3.76193702e-01 6.60042048e-01 -1.43408489e+00
-3.70476246e-01 -1.20062923e+00 -9.52547014e-01 5.68507612e-01
1.23797273e-02 -1.68207020e-01 -1.02958834e+00 -3.45973939e-01
5.97504899e-02 -1.80372030e-01 9.13166344e-01 1.77729681e-01
5.29890120e-01 -3.68240893e-01 -1.60914406e-01 7.48625994e-02
1.84617436e+00 4.79308873e-01 1.16406763e+00 1.18605220e+00
6.59459591e-01 7.83299983e-01 4.81664956e-01 3.53035927e-01
-7.68955573e-02 2.98385650e-01 1.02274430e+00 2.46387087e-02
3.99561703e-01 1.87166721e-01 2.12740585e-01 9.02831331e-02
-4.57235537e-02 -2.09498480e-01 -1.17680681e+00 4.64576334e-01
-1.49740088e+00 -1.11763108e+00 -7.04120636e-01 2.29783034e+00
-1.07607059e-02 7.74822682e-02 3.19747776e-01 2.70108849e-01
1.25687027e+00 2.21537426e-01 -2.58495629e-01 -5.89954913e-01
2.24279955e-01 1.63587317e-01 8.58839154e-01 2.68268585e-01
-1.50166976e+00 6.76213324e-01 6.04583549e+00 6.59648061e-01
-1.00797307e+00 1.25402749e-01 4.85798299e-01 5.14881194e-01
2.77647287e-01 -2.48083800e-01 -6.57183707e-01 7.97787964e-01
1.17451012e+00 -2.20626453e-03 3.19069654e-01 1.20986497e+00
-8.63759965e-02 5.45606529e-03 -2.87262321e-01 7.67312646e-01
3.40588629e-01 -1.31022561e+00 -1.30480364e-01 7.20651567e-01
6.36803448e-01 1.65679157e-01 4.12926599e-02 2.90256649e-01
2.62435883e-01 -9.67124045e-01 4.19151820e-02 -2.50183433e-01
2.53675193e-01 -8.93772900e-01 1.55501282e+00 2.46078849e-01
-7.54577637e-01 -3.75487059e-01 -5.01385331e-01 2.07043067e-01
1.33391261e-01 7.77218878e-01 -1.40328205e+00 3.85530800e-01
6.32133007e-01 3.58499944e-01 -7.32099771e-01 1.34675109e+00
-1.12752996e-01 6.60235822e-01 3.20119038e-02 -3.60369623e-01
6.03006005e-01 1.48078287e-02 2.64276236e-01 1.39509594e+00
1.28816441e-01 -4.23577040e-01 -2.74618626e-01 6.89235151e-01
3.03332627e-01 1.47902519e-01 -7.88426220e-01 -2.87587941e-01
1.54516265e-01 1.61364007e+00 -6.66466236e-01 -5.52951515e-01
-4.55804110e-01 8.40339601e-01 8.66896063e-02 7.36283138e-02
-9.50125158e-01 -8.42089593e-01 6.09911263e-01 3.74911457e-01
1.34809569e-01 -2.65954077e-01 -6.48644447e-01 -1.20154953e+00
-3.93417031e-01 -9.03307557e-01 5.30325651e-01 -2.94911593e-01
-1.67615664e+00 7.86789179e-01 -4.90145266e-01 -9.07138646e-01
-1.05217412e-01 -1.06413162e+00 -7.25464523e-01 9.91778731e-01
-1.39257538e+00 -1.23770452e+00 -6.41866922e-01 3.64709556e-01
4.34040368e-01 -5.08230805e-01 6.55754983e-01 5.00438035e-01
-3.45495820e-01 1.45163238e-01 2.31398210e-01 4.61629540e-01
6.39710307e-01 -1.18176389e+00 1.10495889e+00 7.94155598e-01
-3.26472670e-01 8.51331890e-01 4.28363591e-01 -8.61853421e-01
-9.02726054e-01 -7.25480080e-01 1.07032692e+00 -5.37603438e-01
8.92551899e-01 -2.42473513e-01 -1.19440353e+00 4.31629390e-01
3.91967893e-01 -5.58200121e-01 7.25012541e-01 1.66007489e-01
-9.58809495e-01 3.36464643e-01 -1.52778530e+00 1.62662402e-01
1.59607723e-01 -4.90028858e-01 -5.08736670e-01 3.54927301e-01
-3.92866991e-02 4.08671498e-01 -7.20501542e-01 -2.08401475e-02
7.81402826e-01 -1.18334579e+00 1.19001460e+00 -6.85733199e-01
4.29721951e-01 8.94689262e-02 4.54228595e-02 -1.12160277e+00
-7.38818407e-01 1.76384762e-01 -8.66897628e-02 8.14037144e-01
1.41749084e-01 -1.04562616e+00 1.22935069e+00 3.92579705e-01
3.64634693e-01 -2.76270419e-01 -5.99414110e-01 -7.66666114e-01
1.74708948e-01 3.18327546e-01 1.04157254e-01 1.44129622e+00
5.15273511e-02 2.51076043e-01 -5.53214610e-01 9.91196036e-02
1.16852653e+00 -5.73285401e-01 6.35342836e-01 -1.73444057e+00
-1.11619115e-01 -4.57603574e-01 -7.68583059e-01 -4.26334888e-01
-6.23608790e-02 -5.00005007e-01 -2.85328388e-01 -1.51887274e+00
3.33898187e-01 -2.72685617e-01 -4.29617882e-01 4.82316613e-01
-2.37132937e-01 6.34338796e-01 9.94501710e-02 3.29184741e-01
-1.01013817e-01 -1.51907980e-01 7.00442314e-01 -1.73630103e-01
1.78094208e-01 2.17820793e-01 -5.07834733e-01 7.63406381e-02
1.34664392e+00 -2.63441205e-01 7.55485743e-02 2.59543270e-01
1.78109080e-01 -3.35794598e-01 4.01774019e-01 -9.02025104e-01
-3.38421538e-02 1.35608211e-01 5.11535227e-01 -5.45476079e-01
5.09096347e-02 -6.38099074e-01 -2.07694754e-01 1.13386834e+00
-1.82338804e-01 -1.83761880e-01 -4.87249121e-02 4.96932060e-01
-6.09125495e-02 -5.13189077e-01 1.14670730e+00 -2.78224975e-01
-9.29307461e-01 1.48341849e-01 -7.02814221e-01 -6.39841735e-01
1.33861244e+00 -2.74464697e-01 -7.97745705e-01 -4.31784213e-01
-4.62290317e-01 -4.78707045e-01 3.85082722e-01 5.61527789e-01
5.16558111e-01 -1.02697158e+00 -8.22260618e-01 2.25793049e-01
1.69590369e-01 -1.00074208e+00 3.08456719e-01 3.05191398e-01
-1.29928517e+00 9.62295830e-01 -1.00294960e+00 -2.15794876e-01
-1.41556001e+00 8.93581331e-01 -2.01644599e-02 -3.14889312e-01
-2.00366527e-01 4.02345747e-01 1.31526858e-01 -6.59111261e-01
-1.00539044e-01 6.58311069e-01 -4.43493396e-01 -7.46280625e-02
7.71728218e-01 1.06624222e+00 3.70831490e-01 -5.94287515e-01
-2.64994025e-01 -4.34728758e-03 -6.51550710e-01 8.98870453e-02
1.16315150e+00 2.72092968e-01 -2.49755219e-01 -3.78745466e-01
1.47097325e+00 -4.15540248e-01 -5.91013134e-01 1.48416832e-01
3.63287270e-01 -9.97361600e-01 4.60502990e-02 -1.10478556e+00
-8.19420695e-01 1.06993186e+00 7.05496192e-01 1.05737138e+00
5.06991148e-01 -2.45726883e-01 1.06984282e+00 3.81865740e-01
5.30778885e-01 -7.29389250e-01 2.36882761e-01 5.75221121e-01
3.78290087e-01 -1.66221011e+00 -2.73061574e-01 -5.42575479e-01
-3.37534815e-01 1.41367769e+00 6.20942891e-01 -6.33731008e-01
4.92958754e-01 -2.41243705e-01 2.38990068e-01 -4.78477478e-01
-1.84012979e-01 1.21193878e-01 3.83827314e-02 1.03860795e+00
3.27402323e-01 1.25505924e-01 -5.67930996e-01 -2.73439914e-01
3.16365421e-01 -3.49540532e-01 6.45654678e-01 7.22127259e-01
-9.22513008e-01 -1.16929185e+00 -6.63538814e-01 5.24478674e-01
-7.40060449e-01 -2.90274352e-01 -7.84245372e-01 9.53530610e-01
-3.36023360e-01 8.39634001e-01 -2.64696687e-01 -5.25751054e-01
-1.93260670e-01 8.31164494e-02 -5.17574139e-02 -3.08400482e-01
-9.75492120e-01 -3.78908604e-01 1.46214783e-01 -1.26349792e-01
2.75706410e-01 -5.44906020e-01 -4.01370615e-01 -1.07078445e+00
-7.55565643e-01 -8.76785889e-02 1.40927112e+00 5.68728387e-01
3.47127289e-01 -4.76369113e-02 8.67593706e-01 -8.05816531e-01
-9.78932142e-01 -9.99194205e-01 -6.97219551e-01 5.01978099e-01
1.38193220e-01 -5.32932937e-01 -6.40096486e-01 -2.94586331e-01] | [7.810333251953125, 9.98774242401123] |
9053a599-0189-4607-a714-42e92e50e8e0 | efficiently-measuring-the-cognitive-ability | 2306.10512 | null | https://arxiv.org/abs/2306.10512v1 | https://arxiv.org/pdf/2306.10512v1.pdf | Efficiently Measuring the Cognitive Ability of LLMs: An Adaptive Testing Perspective | Large language models (LLMs), like ChatGPT, have shown some human-like cognitive abilities. For comparing these abilities of different models, several benchmarks (i.e. sets of standard test questions) from different fields (e.g., Literature, Biology and Psychology) are often adopted and the test results under traditional metrics such as accuracy, recall and F1, are reported. However, such way for evaluating LLMs can be inefficient and inaccurate from the cognitive science perspective. Inspired by Computerized Adaptive Testing (CAT) used in psychometrics, we propose an adaptive testing framework for LLM evaluation. Rather than using a standard test set and simply reporting accuracy, this approach dynamically adjusts the characteristics of the test questions, such as difficulty, based on the model's performance. This allows for a more accurate estimation of the model's abilities, using fewer questions. More importantly, it allows LLMs to be compared with humans easily, which is essential for NLP models that aim for human-level ability. Our diagnostic reports have found that ChatGPT often behaves like a ``careless student'', prone to slip and occasionally guessing the questions. We conduct a fine-grained diagnosis and rank the latest 6 instruction-tuned LLMs from three aspects of Subject Knowledge, Mathematical Reasoning, and Programming, where GPT4 can outperform other models significantly and reach the cognitive ability of middle-level students. Different tests for different models using efficient adaptive testing -- we believe this has the potential to become a new norm in evaluating large language models. | ['Enhong Chen', 'Shijin Wang', 'Qingyang Mao', 'Zheng Zhang', 'Guanhao Zhao', 'Zhenya Huang', 'Rui Lv', 'Weizhe Huang', 'Yuting Ning', 'Qi Liu', 'Yan Zhuang'] | 2023-06-18 | null | null | null | null | ['mathematical-reasoning'] | ['natural-language-processing'] | [-4.60682720e-01 4.36508991e-02 -1.73782527e-01 -2.38005802e-01
-6.18491769e-01 -7.68496215e-01 1.90228581e-01 5.88925064e-01
-5.16895533e-01 5.18983424e-01 -7.77177289e-02 -8.68763089e-01
-6.78661346e-01 -1.07972038e+00 -5.80881417e-01 -1.47850558e-01
2.62388259e-01 7.98624933e-01 5.55425763e-01 -3.26022923e-01
6.41300499e-01 3.30735236e-01 -1.68188453e+00 1.25515193e-01
1.72081792e+00 5.80845118e-01 4.33681518e-01 6.10480130e-01
-4.87777621e-01 9.62997556e-01 -7.73814917e-01 -8.58814538e-01
-3.57213855e-01 -4.20630485e-01 -9.34237123e-01 -5.12403846e-01
5.02692044e-01 -2.76648670e-01 1.14782333e-01 1.13019264e+00
3.30306709e-01 1.97979867e-01 5.30562699e-01 -9.66379046e-01
-1.08287764e+00 7.67933905e-01 -1.36375595e-02 1.26425758e-01
7.69329309e-01 2.86655158e-01 7.38955140e-01 -5.49823761e-01
3.10919404e-01 1.42539561e+00 7.93913662e-01 3.24484795e-01
-1.14849353e+00 -7.48731673e-01 7.36739710e-02 4.96962339e-01
-1.28650606e+00 1.33774236e-01 1.22477755e-01 -5.63233197e-01
6.10132098e-01 2.62358069e-01 1.00320399e+00 7.75169253e-01
4.67338502e-01 4.87669349e-01 1.53810096e+00 -5.11692405e-01
2.21267223e-01 5.68245769e-01 5.54319203e-01 6.93765819e-01
6.03735209e-01 -9.75060686e-02 -4.77702975e-01 6.87008351e-02
5.88530719e-01 -2.13368326e-01 -9.59885642e-02 -3.76901496e-03
-1.00815785e+00 7.45284736e-01 1.20450538e-02 6.11442327e-01
-1.19360507e-01 -3.30572933e-01 -7.91258737e-03 6.61388814e-01
-5.63632436e-02 8.45015824e-01 -6.45602942e-01 -5.73251903e-01
-7.86940098e-01 4.93994981e-01 9.63172376e-01 8.22624326e-01
4.38685805e-01 -9.95637402e-02 -5.53330541e-01 9.73484457e-01
4.32974726e-01 5.52693367e-01 1.09050417e+00 -1.02435863e+00
3.38226885e-01 1.19547486e+00 -1.90357015e-01 -9.42017972e-01
-3.02769005e-01 -6.65641963e-01 -3.32372040e-01 3.69796574e-01
5.83337545e-01 1.59302503e-01 -7.33263552e-01 1.85118306e+00
1.16951227e-01 -1.79182664e-01 -2.06483141e-01 5.61553121e-01
7.91378736e-01 3.33332330e-01 4.71263200e-01 -8.78835693e-02
1.43053627e+00 -7.53848672e-01 -5.80716491e-01 -2.54393160e-01
1.08575821e+00 -6.75680637e-01 1.76652443e+00 6.98172212e-01
-1.47991633e+00 -8.08710635e-01 -9.77741241e-01 -9.79542583e-02
-6.09221935e-01 -3.87257218e-01 4.55905348e-01 1.15082490e+00
-1.25457048e+00 5.95734000e-01 -3.52381051e-01 -1.21972784e-01
-3.47577631e-02 2.67918050e-01 -1.27433851e-01 -3.25425297e-01
-1.19145525e+00 1.22731066e+00 5.31931341e-01 -4.16409284e-01
-6.19741023e-01 -7.79832542e-01 -6.21945083e-01 5.04540205e-01
2.39223644e-01 -6.89878285e-01 1.32922232e+00 -5.88602602e-01
-1.50150514e+00 7.18635142e-01 2.70045958e-02 -3.78640443e-02
3.99604350e-01 -8.56668353e-02 -2.96112329e-01 -1.05820999e-01
1.85370252e-01 2.53843576e-01 1.84874520e-01 -5.97697973e-01
-4.19821113e-01 -2.75227964e-01 2.73172498e-01 1.80590168e-01
-5.60364366e-01 3.53579484e-02 -1.81414291e-01 -4.08081055e-01
2.24148825e-01 -6.84682608e-01 -7.03199883e-04 -3.08409095e-01
6.60657808e-02 -5.61270177e-01 1.72183350e-01 -6.81966364e-01
1.68596184e+00 -1.78030801e+00 -1.91102609e-01 4.30576622e-01
4.90453511e-01 3.66585582e-01 -2.12507933e-01 1.37017071e-01
1.25014380e-01 6.11256897e-01 1.53173476e-01 8.68975520e-02
3.55810553e-01 -1.90826476e-01 1.60148382e-01 -1.59860834e-01
-8.09565559e-02 1.07560503e+00 -9.04026031e-01 -6.93982899e-01
-3.94589491e-02 -4.83394116e-02 -8.97268534e-01 3.07144850e-01
-1.13245308e-01 2.56217986e-01 -2.88583606e-01 5.79241693e-01
3.81097674e-01 -2.54437476e-01 -2.12159790e-02 6.69974744e-01
-1.52803078e-01 5.31150401e-01 -1.06816435e+00 1.00507724e+00
-4.63216722e-01 4.40908283e-01 -3.97890389e-01 -8.18690062e-01
1.04591203e+00 2.00465724e-01 -2.61494935e-01 -1.09587181e+00
-1.08532801e-01 3.26969743e-01 6.58969998e-01 -6.75614238e-01
3.06373358e-01 5.84671795e-02 -5.23619987e-02 4.56839979e-01
1.26171038e-01 -4.42048281e-01 4.06640291e-01 2.49093100e-02
1.13444996e+00 -3.15368325e-02 3.81913573e-01 -3.76161397e-01
6.79119706e-01 -7.05459574e-03 3.57763171e-01 1.15536380e+00
-1.06020525e-01 -7.75575936e-02 4.38400805e-01 -2.72373948e-02
-7.96048045e-01 -1.04846454e+00 -1.31642357e-01 1.43851280e+00
-3.63128424e-01 -4.08328801e-01 -8.93181741e-01 -3.28660458e-01
-6.67394400e-02 1.17564547e+00 -2.48991668e-01 -5.23580611e-01
-3.72843206e-01 -5.33119321e-01 5.41825235e-01 3.78106982e-01
5.65001309e-01 -1.06979239e+00 -2.69898921e-01 1.31468713e-01
-1.85435593e-01 -8.18953097e-01 -6.41527474e-02 -1.70225844e-01
-9.53295231e-01 -8.66032243e-01 -4.28701282e-01 -7.97594607e-01
4.67665672e-01 -2.73265596e-02 1.40697622e+00 8.46631706e-01
2.26034060e-01 5.28649509e-01 -3.27261716e-01 -4.91845280e-01
-6.20590866e-01 7.88084418e-02 8.96542594e-02 -8.24414790e-01
4.61248726e-01 -3.88130963e-01 -3.61528039e-01 4.46566880e-01
-7.32532799e-01 -1.15864016e-01 9.99809980e-01 5.68180084e-01
2.51438022e-01 8.82873386e-02 6.17092550e-01 -8.67719531e-01
1.17758405e+00 -4.19514686e-01 -4.63035733e-01 5.71188986e-01
-1.16128755e+00 1.88779496e-02 5.58480501e-01 -8.97774279e-01
-8.41336012e-01 -9.58594739e-01 -2.81760365e-01 1.20284192e-01
-1.61573380e-01 9.02884722e-01 4.33922336e-02 -4.98491287e-01
7.65718222e-01 3.71956289e-01 -2.35822186e-01 -3.28772902e-01
-2.15126127e-01 4.28568572e-01 3.47640634e-01 -1.07684648e+00
9.32306826e-01 -6.57421768e-01 -2.36748755e-01 -5.04021823e-01
-8.84947777e-01 -1.19931959e-02 -2.88874984e-01 -2.68619567e-01
4.10616785e-01 -4.75923538e-01 -1.22018433e+00 2.92795062e-01
-8.23621809e-01 -5.60670793e-01 -8.14380422e-02 7.95788527e-01
-3.50887597e-01 2.62680799e-01 -7.24723339e-01 -8.06394637e-01
-4.73579802e-02 -1.35821271e+00 4.73548800e-01 5.84878147e-01
-5.93165159e-01 -1.22540939e+00 5.76914065e-02 7.11377144e-01
8.01979005e-01 -4.40180689e-01 1.67842460e+00 -1.12436068e+00
-4.38833416e-01 -2.74487853e-01 6.14174716e-02 3.02609235e-01
-5.26450932e-01 -5.77973798e-02 -6.23419344e-01 -4.69775386e-02
2.26466149e-01 -4.44196194e-01 2.29581818e-01 2.12332770e-01
1.31196713e+00 -4.06581849e-01 -3.23847830e-02 3.23382080e-01
1.18327987e+00 3.25598180e-01 6.67762220e-01 6.48297608e-01
2.44121000e-01 5.78496516e-01 5.53170562e-01 -2.05236286e-01
7.09768593e-01 3.54433954e-01 -1.31982222e-01 4.93474334e-01
1.48404287e-02 -4.33376878e-01 6.44713044e-01 1.31673586e+00
-3.75320762e-02 -1.19433545e-01 -1.35879409e+00 2.15348348e-01
-1.27045739e+00 -6.97401106e-01 -1.33397609e-01 2.54127860e+00
1.00108683e+00 4.73987222e-01 -5.57873510e-02 1.75120637e-01
3.93647879e-01 -5.82868874e-01 -3.99350822e-01 -7.87406206e-01
7.30639547e-02 5.23258626e-01 -2.36969311e-02 5.25103748e-01
-1.70819819e-01 8.51717770e-01 6.97611284e+00 1.03204489e+00
-8.88359189e-01 1.24084368e-01 6.69324458e-01 2.77308136e-01
-6.17908061e-01 -1.27059948e-02 -1.02389550e+00 5.95536947e-01
1.65713429e+00 -5.24608135e-01 3.79151523e-01 7.11188436e-01
1.43918350e-01 -2.89106756e-01 -9.54435825e-01 6.18753612e-01
-9.62365493e-02 -8.90151680e-01 2.32708260e-01 7.41044730e-02
6.15751088e-01 -5.50066769e-01 2.07751408e-01 1.24859345e+00
3.88296932e-01 -1.22316933e+00 8.65894616e-01 8.14142108e-01
5.25183022e-01 -5.08655965e-01 9.84549582e-01 7.63541281e-01
-6.94778085e-01 -2.67201513e-01 -4.01333421e-01 -4.53450710e-01
-5.85587204e-01 3.58543068e-01 -9.13845778e-01 -6.56791450e-03
7.31590629e-01 -2.86912680e-01 -1.15815854e+00 1.34931374e+00
-3.73365849e-01 8.09560359e-01 -1.52840363e-02 -6.31178916e-01
-1.30375177e-01 -1.16609201e-01 1.36287019e-01 8.90260696e-01
6.86456263e-01 1.99097082e-01 1.60926744e-01 1.06356835e+00
1.05046473e-01 5.59090257e-01 -3.51960778e-01 -2.37311497e-01
8.54079723e-01 9.50882554e-01 -6.45203829e-01 -4.31441844e-01
-4.04090971e-01 4.76490915e-01 4.84052926e-01 8.95044729e-02
-6.04127288e-01 -2.24664271e-01 2.94121116e-01 3.53214622e-01
-3.77365232e-01 -2.88362145e-01 -5.37624359e-01 -8.27854753e-01
-1.13974094e-01 -1.17916739e+00 2.35715285e-01 -9.71527874e-01
-1.15356517e+00 2.05430105e-01 1.38835013e-01 -9.62234974e-01
-3.25662553e-01 -9.98664975e-01 -6.68206394e-01 9.48355794e-01
-1.00284696e+00 -6.67896509e-01 -4.88488346e-01 4.15168524e-01
2.79669046e-01 -3.94130908e-02 8.00371826e-01 1.71110213e-01
-4.63847607e-01 9.84487295e-01 -1.55334473e-01 1.23467572e-01
7.25702703e-01 -1.36174524e+00 9.45504829e-02 5.77046216e-01
-1.61324069e-01 9.62652266e-01 4.79499578e-01 -6.70201421e-01
-1.16195679e+00 -5.59232414e-01 1.00016904e+00 -5.96712649e-01
8.00872087e-01 1.70975132e-03 -1.27672613e+00 5.41748166e-01
-1.42488509e-01 -6.97342336e-01 9.73238170e-01 2.65145034e-01
-7.68255070e-02 2.33110450e-02 -1.12463832e+00 7.47585177e-01
9.73918974e-01 -3.80989730e-01 -8.92545223e-01 2.27869481e-01
8.49065423e-01 -3.59944403e-01 -1.22617829e+00 1.70882612e-01
6.52916729e-01 -1.09440935e+00 9.27579403e-01 -5.86798310e-01
3.31562698e-01 2.22021908e-01 1.80765018e-01 -1.38173115e+00
-6.08000100e-01 -9.31829661e-02 -1.73093081e-02 1.23962235e+00
5.54982007e-01 -9.78303254e-01 4.03735936e-01 9.29152489e-01
-2.49996021e-01 -8.86936903e-01 -5.78823507e-01 -8.31420124e-01
4.07328635e-01 -6.24749720e-01 6.77081287e-01 1.03929663e+00
1.24685215e-02 1.40432388e-01 5.12872458e-01 8.45293775e-02
3.18789333e-01 -2.45523974e-01 5.60592115e-01 -1.86302483e+00
-2.18692318e-01 -9.89021361e-01 -4.19222414e-01 -7.86360085e-01
2.00879425e-01 -8.76890540e-01 -2.05219775e-01 -1.24861479e+00
2.31963411e-01 -4.81009036e-01 -1.28249720e-01 4.01876926e-01
-4.20978367e-01 -7.64609501e-02 8.75381157e-02 -1.68266639e-01
-5.56602240e-01 2.54396368e-02 1.38956428e+00 9.00727734e-02
-1.44607797e-01 -6.05272017e-02 -1.09556746e+00 9.75553513e-01
8.85139346e-01 -4.66889679e-01 -4.05980825e-01 -3.24622810e-01
5.94816029e-01 1.11411594e-01 2.11475179e-01 -1.33264160e+00
5.12957692e-01 -5.28664768e-01 6.01876676e-01 2.76275668e-02
-7.74182230e-02 -3.96483958e-01 -1.63138062e-01 6.10259354e-01
-3.32003027e-01 6.21999323e-01 3.83879989e-01 -1.28589153e-01
-1.40755296e-01 -9.25137043e-01 5.66368639e-01 -3.43849719e-01
-6.70510232e-01 2.82496810e-02 -3.90160948e-01 2.78261214e-01
7.34497666e-01 -3.03239256e-01 -5.92532158e-01 -4.63445663e-01
-6.03294015e-01 3.15448701e-01 4.48287874e-01 2.83541709e-01
6.04326904e-01 -1.04460168e+00 -7.34344482e-01 1.17541425e-01
-5.17248474e-02 -1.87201947e-01 3.16946596e-01 1.06962276e+00
-6.61605239e-01 6.87211931e-01 -1.90180436e-01 -3.81165028e-01
-1.11462808e+00 3.67750883e-01 2.59153545e-01 -5.52206874e-01
-7.20591098e-02 6.54657483e-01 -2.69238576e-02 -6.90462291e-01
5.96395060e-02 -5.96423447e-01 -3.85141015e-01 -3.81924100e-02
7.04836071e-01 3.84826660e-01 1.55345589e-01 -6.89949393e-02
4.87607941e-02 5.32689631e-01 -6.12941347e-02 -4.90509421e-02
9.71487164e-01 5.44291697e-02 -3.10401350e-01 7.45831966e-01
5.88197052e-01 3.86951387e-01 -2.76810884e-01 -4.75933887e-02
3.80192548e-01 -2.89123148e-01 -3.51764075e-02 -1.22261906e+00
-3.44643354e-01 6.58001244e-01 5.71422756e-01 2.23763287e-01
9.10627067e-01 -1.50301665e-01 1.28948838e-01 7.41691530e-01
4.73186910e-01 -1.02114701e+00 1.87818870e-01 7.70407319e-01
6.55661225e-01 -1.01322436e+00 -2.67074317e-01 -9.42991450e-02
-2.15361878e-01 1.07710099e+00 1.05911171e+00 2.53001720e-01
2.35867575e-01 1.09073348e-01 -1.55940667e-01 4.18973789e-02
-9.15325642e-01 -7.77780861e-02 5.37933230e-01 3.78187805e-01
9.29388285e-01 1.49814859e-01 -6.64042830e-01 1.02459264e+00
-9.74272311e-01 2.69492716e-02 5.55919588e-01 5.98624051e-01
-8.68899345e-01 -1.08527720e+00 -8.53322327e-01 6.58760488e-01
-2.66165376e-01 -3.17533642e-01 -2.78421789e-01 1.00386286e+00
2.49715835e-01 8.20278525e-01 -1.11469604e-01 -4.01815563e-01
4.58878905e-01 6.97669148e-01 5.66239655e-01 -7.30022192e-01
-7.07275808e-01 -6.08588457e-01 -2.21468613e-01 -2.64870167e-01
1.35541767e-01 -4.49868739e-01 -1.02847981e+00 -8.59360576e-01
-2.25662500e-01 4.50824291e-01 4.31961834e-01 1.16562200e+00
2.23347219e-03 5.56473434e-01 -4.27426919e-02 -2.05389246e-01
-8.73583138e-01 -1.28778303e+00 -2.80421704e-01 2.34470055e-01
-1.72944456e-01 -7.78003991e-01 -3.77684981e-01 -6.42892838e-01] | [10.02295207977295, 7.435790538787842] |
13b96357-1fd4-4d4e-83cd-3ff145298fe3 | s4nd-single-shot-single-scale-lung-nodule | 1805.02279 | null | http://arxiv.org/abs/1805.02279v2 | http://arxiv.org/pdf/1805.02279v2.pdf | S4ND: Single-Shot Single-Scale Lung Nodule Detection | The state of the art lung nodule detection studies rely on computationally
expensive multi-stage frameworks to detect nodules from CT scans. To address
this computational challenge and provide better performance, in this paper we
propose S4ND, a new deep learning based method for lung nodule detection. Our
approach uses a single feed forward pass of a single network for detection and
provides better performance when compared to the current literature. The whole
detection pipeline is designed as a single $3D$ Convolutional Neural Network
(CNN) with dense connections, trained in an end-to-end manner. S4ND does not
require any further post-processing or user guidance to refine detection
results. Experimentally, we compared our network with the current
state-of-the-art object detection network (SSD) in computer vision as well as
the state-of-the-art published method for lung nodule detection (3D DCNN). We
used publically available $888$ CT scans from LUNA challenge dataset and showed
that the proposed method outperforms the current literature both in terms of
efficiency and accuracy by achieving an average FROC-score of $0.897$. We also
provide an in-depth analysis of our proposed network to shed light on the
unclear paradigms of tiny object detection. | ['Naji Khosravan', 'Ulas Bagci'] | 2018-05-06 | null | null | null | null | ['lung-nodule-detection'] | ['medical'] | [ 1.09174192e-01 2.96814024e-01 -8.19047764e-02 -4.89709750e-02
-8.63968432e-01 -2.37900466e-01 3.19503605e-01 4.73998711e-02
-6.98189735e-01 -1.26733929e-01 -2.57296443e-01 -7.36359656e-01
7.97930285e-02 -7.38266885e-01 -7.57376790e-01 -4.69925225e-01
-5.57213053e-02 6.06954753e-01 1.02402902e+00 4.80690673e-02
-3.28243405e-01 6.81493759e-01 -1.03882647e+00 4.93197441e-01
3.14533412e-01 1.08897495e+00 3.84074330e-01 1.00223088e+00
2.80241907e-01 8.13330531e-01 -8.84299204e-02 -4.80557680e-01
6.30379140e-01 -2.67320722e-01 -8.27424049e-01 -9.51428264e-02
5.48233151e-01 -7.37916589e-01 -6.35568619e-01 7.10932970e-01
8.59646618e-01 -5.85311174e-01 4.72393453e-01 -5.74561954e-01
-3.07870597e-01 7.79381633e-01 -5.51104665e-01 7.96848595e-01
-3.49454194e-01 4.92511302e-01 6.93532050e-01 -9.98449922e-01
3.48607272e-01 7.76560843e-01 1.14370763e+00 7.59213507e-01
-7.40429223e-01 -6.72461331e-01 -4.40904468e-01 -1.59030035e-01
-1.22184741e+00 7.89234787e-02 1.89021692e-01 -3.01812947e-01
9.64854002e-01 2.97350854e-01 9.55912352e-01 7.12525547e-01
3.42467666e-01 8.32542121e-01 6.18076026e-01 -4.62780118e-01
-5.25301695e-02 1.01534098e-01 -2.88676172e-01 1.53159678e+00
8.43524337e-01 4.24903899e-01 6.95601404e-02 -9.05364677e-02
1.19712162e+00 1.51047528e-01 1.40100211e-01 -6.45407200e-01
-1.40366566e+00 9.89750266e-01 1.29951918e+00 4.83206064e-01
-4.01618987e-01 7.12037265e-01 5.11157811e-01 -3.67208332e-01
1.83732495e-01 3.82327408e-01 -1.65768359e-02 4.02530313e-01
-1.00451064e+00 7.27404281e-02 7.33972549e-01 5.41754186e-01
1.16806580e-02 -8.62245932e-02 -8.06889236e-01 4.25999671e-01
5.15230358e-01 3.73986900e-01 5.80079615e-01 -4.33036387e-01
1.48243532e-01 7.26767123e-01 -1.05832025e-01 -4.25118774e-01
-7.77137160e-01 -9.57006276e-01 -8.27925980e-01 3.61725599e-01
4.19265985e-01 -1.33918360e-01 -1.33962119e+00 1.12803817e+00
1.83104798e-01 2.43259713e-01 -2.39315987e-01 1.11831260e+00
1.47664177e+00 1.98225696e-02 1.06527880e-01 4.22048420e-01
1.70368195e+00 -1.42284226e+00 3.85317951e-02 -1.90401927e-01
7.85234213e-01 -7.94330597e-01 7.11140692e-01 -9.48221684e-02
-1.45097494e+00 -5.41160405e-01 -1.28644347e+00 4.31107171e-02
-1.19913686e-02 7.26485372e-01 6.34331226e-01 1.04554856e+00
-1.17488682e+00 3.63840163e-01 -1.45653069e+00 -6.63394094e-01
9.88397598e-01 7.48228848e-01 2.03117758e-01 -4.09416780e-02
-7.11881936e-01 1.13774562e+00 3.38004440e-01 -2.78404802e-02
-1.43287206e+00 -1.14414060e+00 -3.28333199e-01 1.63854241e-01
6.68997824e-01 -1.12075412e+00 2.06782222e+00 -3.99781346e-01
-1.05646062e+00 1.17382145e+00 4.48107928e-01 -9.56039608e-01
9.39750910e-01 8.29392076e-02 7.19184503e-02 2.87336648e-01
-1.11114293e-01 8.14981043e-01 6.36559606e-01 -6.15045428e-01
-7.69360125e-01 1.32226914e-01 -5.52422553e-02 3.61446366e-02
-9.68851149e-02 -8.76901373e-02 -7.10444331e-01 -3.05483848e-01
-1.69060472e-02 -1.06212616e+00 -6.42853558e-01 7.23269761e-01
-3.97682577e-01 -3.30792964e-01 8.82307827e-01 -9.69665721e-02
8.76093984e-01 -1.59430468e+00 -5.09543598e-01 1.13620207e-01
7.38955200e-01 6.72533691e-01 9.99937728e-02 -2.07281873e-01
-4.21980470e-02 9.63496417e-02 -2.29746938e-01 -1.93544447e-01
-1.39207810e-01 -9.47785228e-02 5.90541005e-01 5.40107310e-01
3.46503347e-01 1.38145936e+00 -8.27329457e-01 -8.34278882e-01
6.39452219e-01 4.93048638e-01 -5.67777157e-01 1.96170360e-01
3.07676550e-02 -1.17391258e-01 -4.61120188e-01 8.32365215e-01
5.39010823e-01 -6.38138950e-01 -8.34068283e-02 -4.61038560e-01
-2.11809784e-01 4.29714359e-02 -8.07305694e-01 1.74803400e+00
-3.95074517e-01 5.13456821e-01 5.55539653e-02 -7.91992545e-01
3.91066253e-01 4.95614260e-01 6.89835727e-01 -4.04504955e-01
4.89229411e-01 6.76002324e-01 6.40541196e-01 -4.47066188e-01
-1.33967504e-01 -4.43677045e-02 3.20222497e-01 5.12753606e-01
1.03025258e-01 -2.69450307e-01 3.22377354e-01 2.06659198e-01
1.73795569e+00 -4.69191134e-01 5.27560234e-01 -4.25204933e-01
3.81422132e-01 3.81841600e-01 -2.60076076e-01 1.24234068e+00
-6.01149261e-01 7.33997345e-01 2.00801268e-01 -6.50114238e-01
-1.14395380e+00 -1.09272516e+00 -3.37618828e-01 8.87874603e-01
-2.08666772e-01 1.37412874e-03 -8.14246297e-01 -1.22387290e+00
-2.99626365e-02 5.05165420e-02 -8.91140223e-01 6.65266216e-02
-8.90789628e-01 -1.03788948e+00 9.64631796e-01 8.40213895e-01
6.68928683e-01 -1.23342502e+00 -1.33267510e+00 3.54475826e-01
2.09914953e-01 -9.47130859e-01 -3.80527318e-01 5.70233226e-01
-8.66105378e-01 -1.35150146e+00 -1.07127678e+00 -1.00275314e+00
6.67325616e-01 4.40832645e-01 1.45972776e+00 3.75525564e-01
-1.21383142e+00 1.85263827e-01 2.59256251e-02 -6.37062311e-01
-5.89436531e-01 3.21876705e-01 -5.07125080e-01 -7.31796443e-01
3.98579776e-01 9.05728266e-02 -1.19003260e+00 2.36049101e-01
-9.83762085e-01 3.06517743e-02 1.50955272e+00 7.81241298e-01
4.11054552e-01 -2.34398454e-01 1.53204411e-01 -9.91553128e-01
2.87335694e-01 -3.24968040e-01 -6.53023601e-01 3.00898161e-02
-5.47497571e-01 3.08135841e-02 1.35406509e-01 -9.66974571e-02
-7.35659540e-01 6.29195273e-01 -2.27536187e-01 -5.07157922e-01
3.53547782e-02 3.25195312e-01 8.89141142e-01 -5.74033082e-01
1.15555251e+00 7.60344416e-02 1.42854974e-02 -1.60176381e-01
1.87586203e-01 3.23634028e-01 5.44509649e-01 1.46848142e-01
9.30386066e-01 7.72636473e-01 3.15169364e-01 -4.37155396e-01
-1.13214517e+00 -9.34650123e-01 -8.38120699e-01 -2.78689265e-01
9.16816175e-01 -9.09610510e-01 -6.05320573e-01 1.59841314e-01
-8.63554120e-01 -2.42071599e-01 -4.90200788e-01 5.48793852e-01
-4.06619191e-01 1.37675777e-01 -8.34446430e-01 -2.72970468e-01
-8.77613962e-01 -1.12229300e+00 1.01404452e+00 1.57231823e-01
3.74916010e-02 -6.95377111e-01 8.05826262e-02 9.79999974e-02
1.03064299e+00 1.62118569e-01 5.36707282e-01 -6.13757074e-01
-6.72810316e-01 -6.00685775e-01 -8.75581503e-01 -5.32572195e-02
1.05547078e-01 -1.87011525e-01 -7.47424424e-01 -4.56389427e-01
-1.50773525e-01 -4.40061390e-01 1.40170598e+00 8.27926874e-01
1.43123996e+00 2.16215640e-01 -6.91207170e-01 4.84817415e-01
1.60243404e+00 -2.38073438e-01 3.24521035e-01 3.30015182e-01
5.43538451e-01 -1.83032155e-01 8.17479715e-02 2.30413243e-01
-2.62317479e-01 2.03286886e-01 1.03742075e+00 -7.77811825e-01
-8.26062918e-01 1.37418777e-01 -2.53355175e-01 1.34285316e-01
-2.98839927e-01 -1.26901686e-01 -1.26872182e+00 5.88688672e-01
-1.58838987e+00 -7.60380089e-01 -1.69336691e-01 1.69360292e+00
4.92176563e-01 5.61043084e-01 2.26698026e-01 -1.36231571e-01
5.33591330e-01 -5.91566451e-02 -5.79144955e-01 -4.44671512e-03
5.67837477e-01 7.20821261e-01 9.21593845e-01 -2.01766263e-03
-1.71237242e+00 4.73486871e-01 6.65793943e+00 7.87902832e-01
-1.26880264e+00 3.83240044e-01 4.89575803e-01 -4.09669280e-01
3.80916208e-01 -6.90047204e-01 -6.78381085e-01 -2.24749878e-01
7.03388393e-01 1.08083434e-01 -2.36028567e-01 1.22268093e+00
-1.98789593e-02 -8.50007907e-02 -1.19747198e+00 8.13535213e-01
2.34380048e-02 -1.73395729e+00 -1.89938009e-01 -2.42073238e-02
6.16578639e-01 8.80202830e-01 1.42032489e-01 4.21305984e-01
3.43214005e-01 -1.24771130e+00 2.49201819e-01 1.07125361e-02
8.14146876e-01 -4.41117644e-01 1.14757156e+00 2.78901547e-01
-1.48988378e+00 4.06584777e-02 -4.73432064e-01 4.58383590e-01
-2.11599126e-01 5.76540828e-01 -1.67471516e+00 1.69444159e-01
8.69844615e-01 4.44535196e-01 -1.09665620e+00 1.47458196e+00
1.88622087e-01 7.50899136e-01 -5.25113404e-01 -2.75766402e-01
7.49097407e-01 6.45226836e-01 2.92712808e-01 1.78799903e+00
4.95403439e-01 7.07815513e-02 2.49845132e-01 9.95610178e-01
-4.29613054e-01 -4.21102718e-02 -2.56232083e-01 6.52372837e-02
-2.92108282e-02 1.61952889e+00 -1.13937128e+00 -5.00704348e-01
-5.73004603e-01 6.90371931e-01 8.01770389e-02 -4.38251644e-01
-1.07752001e+00 -3.45500670e-02 -1.95291579e-01 5.58978558e-01
7.14673460e-01 2.49195680e-01 -1.31670192e-01 -5.94209254e-01
-1.22682460e-01 -2.53713101e-01 5.97624004e-01 -3.81360531e-01
-1.25966728e+00 6.00978136e-01 -3.00147265e-01 -1.32114363e+00
-5.95695227e-02 -9.55858648e-01 -7.74332881e-01 5.85646331e-01
-1.66844523e+00 -1.31652069e+00 -6.51202798e-01 4.80735719e-01
7.10124612e-01 1.71883833e-02 7.49457479e-01 3.06417853e-01
-3.57197613e-01 6.02448821e-01 -1.75528795e-01 4.91925597e-01
5.54267645e-01 -1.30810297e+00 4.70189065e-01 7.14254558e-01
-1.29497737e-01 -8.73430297e-02 1.82586566e-01 -4.36438531e-01
-1.44272971e+00 -1.56737530e+00 4.28961575e-01 -4.75884378e-01
4.90256757e-01 1.34079927e-03 -5.00581861e-01 6.03287220e-01
2.14168832e-01 9.30552602e-01 3.09703320e-01 -3.58576089e-01
-3.38505954e-03 8.84555131e-02 -1.32093966e+00 3.22489232e-01
8.80046606e-01 9.47243571e-02 -3.25525254e-01 5.48319340e-01
5.77142000e-01 -9.14714694e-01 -5.48338175e-01 7.01828957e-01
5.64975381e-01 -1.21100903e+00 1.31046462e+00 -1.14568546e-01
5.06561399e-01 -1.51630357e-01 3.40484798e-01 -6.46835744e-01
-7.34269142e-01 1.40564859e-01 -1.71607897e-01 -6.16428591e-02
6.94925845e-01 -2.75828373e-02 1.51985860e+00 1.64125025e-01
-6.13281965e-01 -1.16441894e+00 -9.87497389e-01 -5.25775254e-01
1.10605396e-01 -3.24813157e-01 -4.95463274e-02 3.23469579e-01
-5.00454009e-01 -2.41500199e-01 1.63759559e-01 1.13274835e-01
7.94971406e-01 2.60812812e-03 2.90078580e-01 -1.07973349e+00
-3.67204070e-01 -7.73384213e-01 -3.23899686e-01 -7.03591645e-01
-7.94976592e-01 -1.23246086e+00 2.57571399e-01 -1.74569952e+00
7.72867858e-01 -1.84533775e-01 -5.84450662e-01 4.52782422e-01
-7.44021535e-02 7.89050162e-01 4.78532501e-02 3.09354782e-01
-7.89021611e-01 -9.98109803e-02 1.52100849e+00 -2.25146845e-01
5.05124889e-02 3.19810331e-01 -4.11005944e-01 8.70540261e-01
7.19052970e-01 -7.58638620e-01 4.87733595e-02 -2.85024077e-01
-2.77627319e-01 -5.02398387e-02 7.03558445e-01 -1.60975122e+00
4.54798609e-01 3.65912080e-01 8.07131827e-01 -1.11700249e+00
2.84303576e-01 -8.64668667e-01 -3.84381294e-01 1.73852575e+00
-2.04578102e-01 -4.82755005e-01 2.80396819e-01 5.44200122e-01
1.93801880e-01 -2.53216952e-01 1.16408491e+00 -6.29999518e-01
-4.53071415e-01 5.90265989e-01 -5.47444284e-01 -3.73401552e-01
1.16031826e+00 -2.19558105e-01 -8.53808150e-02 2.83352107e-01
-7.86035538e-01 3.49543020e-02 -1.00951552e-01 4.73898835e-02
4.84944910e-01 -1.19416893e+00 -1.12857080e+00 1.13807850e-01
1.54184252e-01 2.79836386e-01 1.45794200e-02 1.05705202e+00
-1.08126044e+00 9.16527271e-01 6.24366850e-02 -1.10776782e+00
-1.25075042e+00 3.15521926e-01 1.03099179e+00 -8.45245302e-01
-9.33823705e-01 1.15939033e+00 7.96292797e-02 -3.27630728e-01
4.40604359e-01 -8.80473018e-01 5.79377636e-02 -4.29298460e-01
1.68981060e-01 2.87020564e-01 5.74488521e-01 2.34328598e-01
-5.80952406e-01 3.20267856e-01 -3.13350290e-01 2.66282380e-01
1.08315241e+00 5.52884221e-01 3.94815862e-01 -1.19277477e-01
9.12213027e-01 -4.95847404e-01 -9.97864008e-01 -3.70247096e-01
-9.78034735e-02 -6.10910095e-02 3.73483866e-01 -1.03370893e+00
-1.32501435e+00 6.02832258e-01 1.26662731e+00 1.76700369e-01
8.93485785e-01 5.03745258e-01 7.24265873e-01 6.90072000e-01
-2.21941620e-01 -5.60192645e-01 5.24237514e-01 2.87442327e-01
6.82987392e-01 -1.88684928e+00 2.58052498e-01 -6.58847511e-01
-9.67190936e-02 1.31270301e+00 9.75322902e-01 -5.63781261e-01
9.25764084e-01 4.30509508e-01 1.84639283e-02 -7.56463051e-01
-6.20453954e-01 -4.83734757e-01 5.65702617e-01 1.87833533e-01
8.22778106e-01 4.74535227e-02 -2.06276566e-01 3.13014567e-01
9.25332904e-02 1.94309101e-01 3.92667443e-01 1.22686291e+00
-8.37407589e-01 -5.70797682e-01 -3.65441173e-01 9.98768151e-01
-6.09121859e-01 -1.89660490e-02 -3.19151133e-01 1.27047968e+00
2.07405806e-01 4.65744227e-01 -3.34164873e-03 2.49770701e-01
3.82119656e-01 -2.98888624e-01 4.83474910e-01 -1.01596534e+00
-1.07165754e+00 1.45767376e-01 -1.33957431e-01 -5.91840506e-01
-4.24537569e-01 -4.33501095e-01 -1.21274769e+00 -1.50854260e-01
-5.03816307e-01 -3.37832302e-01 6.66987896e-01 6.38837457e-01
-1.44994795e-01 1.18270016e+00 2.52449095e-01 -9.38189447e-01
-6.61235988e-01 -1.02458584e+00 -2.26535484e-01 3.86527404e-02
2.00052112e-01 -2.71958828e-01 -1.99122682e-01 -2.80675977e-01] | [15.35279369354248, -2.1580348014831543] |
e00ee139-19ea-4d3e-8036-dc2d53f0f4de | balanced-chamfer-distance-as-a-comprehensive | null | null | http://proceedings.neurips.cc/paper/2021/hash/f3bd5ad57c8389a8a1a541a76be463bf-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/f3bd5ad57c8389a8a1a541a76be463bf-Paper.pdf | Balanced Chamfer Distance as a Comprehensive Metric for Point Cloud Completion | Chamfer Distance (CD) and Earth Mover’s Distance (EMD) are two broadly adopted metrics for measuring the similarity between two point sets. However, CD is usually insensitive to mismatched local density, and EMD is usually dominated by global distribution while overlooks the fidelity of detailed structures. Besides, their unbounded value range induces a heavy influence from the outliers. These defects prevent them from providing a consistent evaluation. To tackle these problems, we propose a new similarity measure named Density-aware Chamfer Distance (DCD). It is derived from CD and benefits from several desirable properties: 1) it can detect disparity of density distributions and is thus a more intensive measure of similarity compared to CD; 2) it is stricter with detailed structures and significantly more computationally efficient than EMD; 3) the bounded value range encourages a more stable and reasonable evaluation over the whole test set. We adopt DCD to evaluate the point cloud completion task, where experimental results show that DCD pays attention to both the overall structure and local geometric details and provides a more reliable evaluation even when CD and EMD contradict each other. We can also use DCD as the training loss, which outperforms the same model trained with CD loss on all three metrics. In addition, we propose a novel point discriminator module that estimates the priority for another guided down-sampling step, and it achieves noticeable improvements under DCD together with competitive results for both CD and EMD. We hope our work could pave the way for a more comprehensive and practical point cloud similarity evaluation. Our code will be available at https://github.com/wutong16/DensityawareChamfer_Distance. | ['Dahua Lin', 'Ziwei Liu', 'Tai Wang', 'Junzhe Zhang', 'Liang Pan', 'Tong Wu'] | 2021-12-01 | null | https://openreview.net/forum?id=B46BjXrLidN | https://openreview.net/pdf?id=B46BjXrLidN | neurips-2021-12 | ['point-cloud-completion'] | ['computer-vision'] | [-4.04434413e-01 -4.19805348e-01 -2.25631848e-01 -3.26858163e-01
-9.78195608e-01 -3.53549898e-01 7.11189508e-01 4.11387116e-01
-3.07162136e-01 4.81722564e-01 1.00758839e-02 -4.37681302e-02
-3.05839449e-01 -1.02118969e+00 -5.81584275e-01 -6.06043518e-01
-5.47427796e-02 6.27602935e-01 6.96628809e-01 -1.56735733e-01
2.48106331e-01 8.55707288e-01 -1.64436650e+00 -1.60674557e-01
1.20049787e+00 1.12586069e+00 3.96881074e-01 1.15319202e-02
-3.15588042e-02 4.33053076e-01 -4.45670277e-01 -3.46756279e-01
2.25102276e-01 -1.95012137e-01 -6.64118111e-01 -2.96346456e-01
4.01415497e-01 -3.12898159e-01 -1.30191162e-01 1.03649044e+00
5.94652057e-01 1.79505765e-01 8.84736359e-01 -1.31905687e+00
-3.83958787e-01 1.64773241e-01 -8.79778028e-01 3.89788710e-02
4.80833769e-01 1.16091490e-01 8.68022561e-01 -1.04886699e+00
4.38882113e-01 1.13176656e+00 8.61238062e-01 2.18179896e-01
-9.52016890e-01 -6.08067632e-01 7.07235187e-02 1.63136780e-01
-1.81846786e+00 -2.28694975e-02 8.98544490e-01 -3.97736669e-01
3.94231796e-01 4.42411453e-01 5.98738015e-01 7.14351296e-01
-8.97582173e-02 9.08129215e-01 7.43640900e-01 -1.03158429e-01
4.25159127e-01 -1.40592963e-01 -2.36794412e-01 5.55665791e-01
3.04408461e-01 1.08443491e-01 -7.89920986e-02 -2.22403899e-01
7.94081867e-01 3.00653636e-01 -5.79799414e-01 -7.99295723e-01
-1.25931287e+00 6.64765358e-01 8.11285019e-01 4.53584552e-01
-2.19725430e-01 -3.94841693e-02 2.21858561e-01 1.43797360e-02
4.57731068e-01 1.75089121e-01 -1.05836960e-02 -3.35044026e-01
-9.84017909e-01 3.98731023e-01 4.80334580e-01 1.01722944e+00
8.57188880e-01 -4.40064281e-01 -1.77615181e-01 9.75625217e-01
3.46258193e-01 6.96757793e-01 2.57848024e-01 -7.70622492e-01
4.91552800e-01 7.41436601e-01 7.99718052e-02 -1.30432379e+00
-3.52832645e-01 -2.03553900e-01 -9.92582381e-01 3.09986800e-01
4.40102994e-01 3.18772256e-01 -5.71184754e-01 1.44977248e+00
3.87325883e-01 2.19650224e-01 -4.36139107e-01 1.11607552e+00
8.40586364e-01 4.57339644e-01 -1.65947735e-01 1.45532668e-01
1.04224277e+00 -6.47767425e-01 -3.82289767e-01 -1.08882589e-02
7.14510262e-01 -7.80448854e-01 1.15503514e+00 2.55562305e-01
-1.17052114e+00 -4.69988644e-01 -1.16375554e+00 -1.56742521e-02
-1.12924479e-01 -5.69648631e-02 4.07253474e-01 3.35770667e-01
-9.23725963e-01 8.57474685e-01 -1.05025625e+00 -3.64193678e-01
4.90553021e-01 1.67744502e-01 -1.85066849e-01 -2.52920508e-01
-8.68218660e-01 8.34437966e-01 1.01107091e-01 -6.26494065e-02
-5.09033442e-01 -7.57327795e-01 -9.49788749e-01 2.08121352e-02
7.91864619e-02 -5.94253302e-01 8.90834808e-01 -3.12364399e-01
-1.24116969e+00 7.88250923e-01 -1.26773164e-01 -7.39184543e-02
8.27471256e-01 -3.69270854e-02 -2.47877985e-01 1.00022770e-01
2.20245585e-01 6.19635046e-01 5.07719338e-01 -1.38512850e+00
-5.61283946e-01 -4.91163015e-01 -1.74035966e-01 1.91667810e-01
-2.14883551e-01 -3.18405956e-01 -6.49579763e-01 -6.64642274e-01
3.78385931e-01 -8.10577452e-01 8.77549034e-03 4.33208346e-01
-3.31471354e-01 -3.07864219e-01 7.75898457e-01 -2.52046287e-01
1.35884476e+00 -2.27206254e+00 -7.03033060e-02 4.44230288e-01
3.01956892e-01 2.55939752e-01 -8.54385421e-02 4.45959091e-01
1.57326207e-01 -2.24175397e-02 -5.67828774e-01 -2.95734495e-01
2.50884388e-02 2.67774034e-02 -2.96758134e-02 6.65984690e-01
1.67135432e-01 7.60414183e-01 -1.09481633e+00 -5.45102835e-01
4.88068730e-01 7.07126737e-01 -3.86175245e-01 -1.93312578e-02
1.91882849e-01 3.78003031e-01 -4.95043248e-01 7.25366414e-01
1.28671873e+00 -2.45400414e-01 -3.17161500e-01 -2.33421877e-01
4.17200997e-02 -1.16145629e-02 -1.31637621e+00 1.62772691e+00
-3.97451699e-01 3.75165761e-01 -9.99469236e-02 -6.18775368e-01
1.36160815e+00 -1.72473025e-02 5.72475255e-01 -7.22393215e-01
1.31452680e-01 4.59309816e-01 -2.34412298e-01 -1.78528070e-01
4.46834713e-01 -1.96612582e-01 1.32694140e-01 1.65338576e-01
-3.51005286e-01 -4.22332197e-01 -3.27533409e-02 1.09830990e-01
9.93036807e-01 5.37353987e-03 2.57181883e-01 -3.77472728e-01
5.79845607e-01 -3.18298519e-01 5.62743962e-01 3.78373653e-01
-3.20269525e-01 1.00659585e+00 1.62185326e-01 -2.52693802e-01
-8.18898678e-01 -1.34115446e+00 -5.60186505e-01 3.60502332e-01
9.75399017e-01 -4.15301234e-01 -4.88125086e-01 -7.45220423e-01
2.94081688e-01 5.50257385e-01 -3.92055839e-01 -1.55830711e-01
-5.18327475e-01 -5.11477053e-01 2.30972618e-01 6.81542695e-01
5.69032073e-01 -4.99432623e-01 -2.24033728e-01 -1.62765920e-01
-3.20068389e-01 -8.37790430e-01 -6.92443907e-01 -3.16134572e-01
-1.09594512e+00 -1.24104190e+00 -1.10210657e+00 -7.93946028e-01
6.94499791e-01 5.40079653e-01 1.16899443e+00 3.10116410e-01
3.53717953e-02 1.87312484e-01 -5.09640098e-01 -3.30813885e-01
-2.06199557e-01 8.08837265e-02 -2.30423696e-02 -1.03724785e-01
5.22418916e-01 -6.60181344e-01 -9.03695107e-01 8.97392929e-01
-9.18029010e-01 -2.55714685e-01 4.42899287e-01 5.53651989e-01
7.77120113e-01 2.88162343e-02 2.54511893e-01 -4.21729565e-01
4.52592492e-01 -4.92097318e-01 -5.15932143e-01 -2.78026587e-03
-6.25052094e-01 -6.68423548e-02 5.32684386e-01 -1.92251235e-01
-7.36842215e-01 -2.31823370e-01 -3.25875014e-01 -5.72516918e-01
-1.03990786e-01 8.64773318e-02 -2.88845539e-01 -2.68480659e-01
5.17872393e-01 6.48815930e-02 1.14564970e-01 -5.98144114e-01
-2.14025564e-02 6.41733170e-01 4.22646582e-01 -6.44315302e-01
9.92250860e-01 8.48212242e-01 -1.05313668e-02 -7.37858295e-01
-4.29764688e-01 -7.18925059e-01 -4.90091950e-01 -2.14021042e-01
4.58885223e-01 -8.36120248e-01 -7.21656799e-01 4.15573746e-01
-9.63611066e-01 -1.30430469e-02 -2.57027835e-01 7.27127075e-01
-4.89210963e-01 7.66740501e-01 -4.47563142e-01 -6.35172963e-01
-2.10889801e-01 -1.12137687e+00 1.28426301e+00 1.36362195e-01
-1.12504378e-01 -1.12995386e+00 1.83151230e-01 7.22138733e-02
3.19082886e-01 4.33211118e-01 6.59884870e-01 -3.83862466e-01
-5.85997760e-01 -4.11290169e-01 -2.87147224e-01 3.80185723e-01
1.92833722e-01 2.01997414e-01 -8.18316698e-01 -4.53143865e-01
3.97420228e-02 3.39677818e-02 6.78543627e-01 3.35720062e-01
1.26312947e+00 8.23045373e-02 -4.87196535e-01 6.76071882e-01
1.52532232e+00 8.97771269e-02 9.77346122e-01 5.21931648e-01
6.22891784e-01 4.08257544e-01 9.39315200e-01 4.99751478e-01
6.21814251e-01 9.18865383e-01 7.08934307e-01 -2.46583149e-01
-9.01495740e-02 -4.80057925e-01 1.64060727e-01 1.04304278e+00
-2.45093450e-01 -1.95551872e-01 -1.05774474e+00 5.89988649e-01
-1.81075823e+00 -9.02511537e-01 -3.89229387e-01 2.59366369e+00
5.35891950e-01 8.09763223e-02 2.38167748e-01 2.71812767e-01
8.47411871e-01 1.09159373e-01 -3.88138622e-01 1.75041463e-02
-1.24153204e-01 9.22713131e-02 3.66636992e-01 5.08002818e-01
-9.74466503e-01 4.83254135e-01 5.30758858e+00 1.37202632e+00
-9.28766251e-01 -4.53185253e-02 5.05670607e-01 -4.43131141e-02
-4.62655216e-01 -7.12961257e-02 -5.48422396e-01 9.67052996e-01
3.16659927e-01 -1.19963795e-01 1.98177658e-02 8.82100940e-01
9.34443027e-02 -2.51903087e-01 -1.09770215e+00 1.30604005e+00
-1.01107255e-01 -1.10247302e+00 5.22922352e-03 1.24679640e-01
6.38902605e-01 2.10078478e-01 -7.17028156e-02 1.51087359e-01
2.18049288e-02 -8.74354959e-01 7.81875193e-01 4.98002201e-01
7.73938298e-01 -9.07531381e-01 9.41955566e-01 4.81944472e-01
-1.45695472e+00 3.43217492e-01 -5.58235824e-01 1.71432823e-01
1.82985857e-01 8.55392516e-01 -4.32056755e-01 7.80254543e-01
7.28118122e-01 8.65528405e-01 -4.91153061e-01 1.55176187e+00
-9.31364447e-02 1.87341213e-01 -4.82092530e-01 -8.33331887e-03
1.66451126e-01 -4.28847283e-01 6.96026504e-01 1.01935112e+00
5.64374804e-01 -3.15104514e-01 1.04772136e-01 8.54341924e-01
4.07915004e-02 1.65154278e-01 -5.73104620e-01 5.60166657e-01
8.97018373e-01 1.08372045e+00 -7.53460109e-01 -1.24698363e-01
-3.29441935e-01 7.11982906e-01 3.36753100e-01 1.10789530e-01
-1.05742371e+00 -4.90691960e-01 7.04948664e-01 5.97964704e-01
2.91130424e-01 -4.75674383e-02 -3.84033948e-01 -1.09421885e+00
4.03458089e-01 -3.50423366e-01 2.86045671e-01 -7.10715353e-01
-1.60483909e+00 6.94083333e-01 7.59409890e-02 -2.03366685e+00
3.43417563e-02 -3.87434125e-01 -7.71542013e-01 7.25030482e-01
-1.55085516e+00 -7.26774812e-01 -7.14277744e-01 5.69891274e-01
1.58260763e-01 2.70803809e-01 4.88530844e-01 6.33144379e-01
-3.84485960e-01 6.84311867e-01 2.24815369e-01 -1.70081966e-02
7.56467760e-01 -1.15389323e+00 2.55373716e-01 5.84363580e-01
-8.31211805e-02 4.39818531e-01 4.17033225e-01 -5.56298077e-01
-9.84124601e-01 -1.05709064e+00 6.26777709e-01 -5.96051574e-01
3.44303250e-01 -8.25560912e-02 -1.11972761e+00 2.63297737e-01
-3.62422287e-01 4.91829254e-02 3.93328846e-01 -5.57283759e-02
-1.21678248e-01 -1.98027521e-01 -1.33763456e+00 3.46943319e-01
1.07104886e+00 -3.51525426e-01 -4.09907907e-01 2.15408638e-01
3.32385033e-01 -3.97445053e-01 -1.26297927e+00 7.37660706e-01
2.38207862e-01 -1.38075030e+00 9.93943155e-01 1.93081275e-01
3.41646671e-01 -5.77983499e-01 -1.10410392e-01 -1.19962585e+00
-4.19731349e-01 -3.27102900e-01 -1.18631221e-01 1.37462008e+00
1.71949938e-01 -6.71615005e-01 8.39108825e-01 4.17022049e-01
-1.38052717e-01 -1.18838489e+00 -9.64010596e-01 -1.35515881e+00
1.88124225e-01 -4.98974174e-01 1.01809013e+00 1.02292478e+00
2.81502269e-02 -2.53341515e-02 -2.75489576e-02 7.93378651e-02
5.48523784e-01 1.86990902e-01 7.70929337e-01 -1.35087395e+00
1.23453595e-01 -6.07236803e-01 -8.68964553e-01 -1.32714212e+00
-3.00065577e-01 -1.00957787e+00 -7.06422627e-02 -1.66643965e+00
1.74075678e-01 -8.44006896e-01 -1.27777994e-01 5.31919934e-02
-1.99910834e-01 3.35056484e-01 1.64983734e-01 6.62871540e-01
-5.35218179e-01 9.82828438e-01 1.47219992e+00 -6.92940280e-02
-1.05017878e-01 1.51024804e-01 -4.96076524e-01 8.54270160e-01
7.77804077e-01 -3.46657991e-01 -2.52359480e-01 -3.95698279e-01
-7.30023608e-02 -2.19241828e-01 3.94141853e-01 -1.24031758e+00
3.04600745e-01 -1.50077455e-02 3.09268385e-01 -7.19515383e-01
3.37267190e-01 -9.54545379e-01 9.20859799e-02 3.22764784e-01
3.07267934e-01 8.30445290e-02 2.66585331e-02 5.36817253e-01
-4.93911177e-01 -1.18790954e-01 1.06561160e+00 2.03804821e-01
-6.09184682e-01 6.46978498e-01 2.65811592e-01 2.06278801e-01
1.09557092e+00 -6.59526110e-01 -2.16676608e-01 -4.94222850e-01
-3.08696330e-01 3.24818760e-01 1.10543132e+00 4.03059989e-01
7.94856012e-01 -1.74081719e+00 -7.09834218e-01 1.54912040e-01
3.76340449e-01 3.89289200e-01 1.18792027e-01 1.17186451e+00
-8.36937189e-01 6.76979348e-02 1.27213940e-01 -1.08081663e+00
-9.89054382e-01 5.01071036e-01 3.04225266e-01 -1.55831724e-02
-7.60657609e-01 7.25306273e-01 3.01922411e-01 -5.07748127e-01
3.09851587e-01 -4.07416105e-01 1.07802562e-01 -1.36318430e-01
5.18831432e-01 7.90506005e-01 2.96135634e-01 -7.51841962e-01
-6.68393254e-01 8.67933512e-01 -1.75364893e-02 2.98003912e-01
1.16426682e+00 -5.76988980e-02 5.60004525e-02 2.13009417e-01
1.36961377e+00 9.97239947e-02 -1.39842880e+00 -1.34581700e-01
3.57527696e-02 -9.30634260e-01 6.61207810e-02 -4.65884030e-01
-1.15803862e+00 8.04265082e-01 6.66163504e-01 2.33950481e-01
1.08890378e+00 2.40194842e-01 9.48231816e-01 -1.06237359e-01
6.27253652e-01 -9.08215106e-01 -1.18613243e-04 3.25166047e-01
9.44024682e-01 -1.21804404e+00 1.60920411e-01 -5.64493656e-01
-5.69525301e-01 9.41869020e-01 5.89474142e-01 -2.45988384e-01
7.34410882e-01 3.03478926e-01 -1.07084274e-01 -3.36392939e-01
-2.37616286e-01 -2.01958910e-01 4.27665591e-01 6.47972107e-01
3.04776192e-01 -5.89893237e-02 -2.76534557e-01 2.05193281e-01
-2.24738404e-01 -8.33680555e-02 9.42067504e-02 7.61952460e-01
-4.45203811e-01 -8.83455396e-01 -3.87742072e-01 4.40424472e-01
3.34870480e-02 1.53299347e-01 -2.17034236e-01 9.82330263e-01
8.75587314e-02 7.10099041e-01 2.70473272e-01 -4.70608979e-01
6.14029408e-01 -6.12109601e-01 4.05909300e-01 -2.22183943e-01
-1.10505380e-01 -1.49330765e-01 -3.68650764e-01 -8.11427236e-01
-3.46808970e-01 -5.75697958e-01 -1.26989305e+00 -8.11288834e-01
-5.81819296e-01 2.31820047e-01 4.96638507e-01 5.42411864e-01
4.64814126e-01 7.29690492e-02 8.39761317e-01 -9.69720781e-01
-5.27959347e-01 -7.97754645e-01 -7.03752995e-01 5.88718414e-01
1.77794904e-01 -1.00849545e+00 -6.32136285e-01 -5.15178680e-01] | [7.88663387298584, -3.1680209636688232] |
d61c90b1-180f-4f57-a068-e00e4f97a878 | alzheimers-disease-diagnostics-by-adaptation | 1607.00455 | null | http://arxiv.org/abs/1607.00455v1 | http://arxiv.org/pdf/1607.00455v1.pdf | Alzheimer's Disease Diagnostics by Adaptation of 3D Convolutional Network | Early diagnosis, playing an important role in preventing progress and
treating the Alzheimer\{'}s disease (AD), is based on classification of
features extracted from brain images. The features have to accurately capture
main AD-related variations of anatomical brain structures, such as, e.g.,
ventricles size, hippocampus shape, cortical thickness, and brain volume. This
paper proposed to predict the AD with a deep 3D convolutional neural network
(3D-CNN), which can learn generic features capturing AD biomarkers and adapt to
different domain datasets. The 3D-CNN is built upon a 3D convolutional
autoencoder, which is pre-trained to capture anatomical shape variations in
structural brain MRI scans. Fully connected upper layers of the 3D-CNN are then
fine-tuned for each task-specific AD classification. Experiments on the
CADDementia MRI dataset with no skull-stripping preprocessing have shown our
3D-CNN outperforms several conventional classifiers by accuracy. Abilities of
the 3D-CNN to generalize the features learnt and adapt to other domains have
been validated on the ADNI dataset. | ['Ayman El-Baz', 'Ehsan Hosseini-Asl', 'Robert Keynto'] | 2016-07-02 | null | null | null | null | ['skull-stripping'] | ['medical'] | [-2.93543041e-01 1.42059952e-01 3.43639523e-01 -9.86131310e-01
-3.35074186e-01 -7.36355269e-03 3.23056847e-01 -9.97347161e-02
-4.53832746e-01 6.22972250e-01 1.31577954e-01 -8.26854482e-02
-2.64462739e-01 -7.53058612e-01 -5.57775676e-01 -4.43196595e-01
-7.73529589e-01 9.74235177e-01 5.03259003e-01 -5.73883541e-02
-1.75773725e-01 7.78561831e-01 -1.29149079e+00 6.01064205e-01
4.96838272e-01 1.47073710e+00 3.06764692e-01 4.50028688e-01
-9.51915532e-02 2.87790328e-01 -4.22422409e-01 -6.90408200e-02
3.29090476e-01 -4.64827195e-02 -7.53909588e-01 1.33083314e-01
4.77102518e-01 -7.65073240e-01 -2.72505760e-01 9.00178492e-01
6.39734387e-01 -5.92214823e-01 8.67254674e-01 -8.53549898e-01
-8.24826419e-01 3.71416926e-01 -6.09620363e-02 9.57449555e-01
-1.62885144e-01 2.33054757e-02 6.63243711e-01 -7.57128954e-01
4.40953285e-01 1.40722847e+00 1.09385252e+00 7.86495864e-01
-9.39492404e-01 -5.69122374e-01 -2.73553003e-02 5.15440106e-01
-1.04880607e+00 -9.05841962e-02 4.86973882e-01 -9.46402252e-01
9.12084877e-01 -4.39142771e-02 1.14999831e+00 1.14518189e+00
7.30413318e-01 4.62445527e-01 1.37102687e+00 -6.71275780e-02
4.03163403e-01 -2.64817744e-01 4.90525454e-01 7.67921746e-01
4.44063574e-01 9.17439684e-02 2.58437634e-01 -2.18010560e-01
9.16192710e-01 2.64590114e-01 -9.92650688e-02 -2.62960643e-01
-1.05035746e+00 9.41918075e-01 8.41322303e-01 3.77076745e-01
-5.40097356e-01 -1.46337643e-01 5.45131624e-01 3.28573942e-01
6.18338287e-01 9.76257697e-02 -9.58934963e-01 5.60409725e-01
-7.79809713e-01 4.57672983e-01 4.03756440e-01 4.61451977e-01
3.78208995e-01 -9.45140049e-02 -1.47902399e-01 1.05660009e+00
6.05535269e-01 4.10341114e-01 1.25699961e+00 -5.77460706e-01
1.69774219e-01 1.11168802e+00 -5.69508255e-01 -3.76299173e-01
-1.16422510e+00 -4.49452668e-01 -1.02539647e+00 4.88177896e-01
2.86449730e-01 -2.73237735e-01 -1.21992910e+00 1.23595440e+00
2.34950289e-01 -3.44801158e-01 -3.09551120e-01 1.04844117e+00
9.99606669e-01 8.06541182e-03 1.23000413e-01 1.49573609e-01
1.47947299e+00 -4.55550045e-01 -2.70902932e-01 -4.37546909e-01
6.00725174e-01 -7.02629611e-02 7.23938942e-01 1.35266185e-01
-7.33748913e-01 -4.68276948e-01 -1.03257215e+00 -8.59500915e-02
-6.47694290e-01 2.16089055e-01 4.39670950e-01 6.33358061e-01
-1.33460093e+00 5.21865606e-01 -1.04798591e+00 -4.92036253e-01
1.20729959e+00 6.05290413e-01 -6.28388107e-01 -6.08126782e-02
-1.11185908e+00 1.14261973e+00 5.40591896e-01 3.09662670e-02
-9.87631619e-01 -8.78194034e-01 -5.05370319e-01 -1.01981139e-04
-4.84564334e-01 -9.91157472e-01 1.12484992e+00 -1.08421671e+00
-1.23721468e+00 1.20463419e+00 4.14225519e-01 -8.41362655e-01
3.70649070e-01 -1.25950828e-01 -6.19262815e-01 3.27535868e-01
1.36089161e-01 7.58934557e-01 7.30134904e-01 -7.61919975e-01
-3.11364293e-01 -1.24643290e+00 -3.31634253e-01 -3.27294290e-01
-3.52477819e-01 5.76870777e-02 4.33620632e-01 -5.55652142e-01
3.20222855e-01 -9.06129956e-01 -4.13718045e-01 4.43666220e-01
-2.32307926e-01 -9.85502228e-02 1.00718355e+00 -9.68367279e-01
5.58283865e-01 -1.83163893e+00 -1.94813088e-01 1.27427995e-01
7.49821961e-01 4.06886160e-01 7.80848488e-02 -5.01041651e-01
-4.78628993e-01 -2.53677461e-02 -3.97908211e-01 7.39633963e-02
-1.03287771e-01 2.72587895e-01 3.88273060e-01 4.63692784e-01
3.99261326e-01 8.97581100e-01 -3.73125643e-01 -2.62122810e-01
-1.24040060e-02 6.69899046e-01 -6.63008749e-01 2.36156508e-01
-2.28049651e-01 3.89613122e-01 -6.47289574e-01 4.23536599e-01
6.50703490e-01 -2.84354776e-01 -5.49956225e-02 -9.65895727e-02
1.84028044e-01 1.88683365e-02 -4.95566934e-01 1.22838128e+00
-1.90185681e-01 4.82372791e-01 -4.02864330e-02 -1.45440042e+00
1.05012286e+00 2.54785329e-01 5.42286813e-01 -8.84593606e-01
5.49274623e-01 3.78177464e-01 4.56855357e-01 -7.07037568e-01
-7.77199507e-01 -4.31126207e-02 3.55963349e-01 2.69532412e-01
3.58414620e-01 4.56466347e-01 -2.06521675e-02 -3.93371493e-01
1.14941537e+00 -3.92875284e-01 3.60388041e-01 -6.12888336e-01
6.87184215e-01 -1.63102791e-01 4.25186068e-01 2.24896550e-01
-4.75720972e-01 6.13037944e-01 5.25018394e-01 -1.20136452e+00
-1.15855753e+00 -1.13543093e+00 -6.40769780e-01 6.88914597e-01
-6.04083240e-01 7.61922896e-02 -8.58294964e-01 -8.30698252e-01
2.46449903e-01 1.18112974e-01 -9.64841962e-01 -2.96821386e-01
-6.04350626e-01 -1.22691917e+00 4.38723713e-01 8.78579080e-01
7.66036630e-01 -8.66692364e-01 -8.62571299e-01 2.06946567e-01
4.96879488e-01 -9.39587653e-01 -3.57373543e-02 4.43262666e-01
-1.56929708e+00 -1.52088988e+00 -1.10417807e+00 -1.11054468e+00
7.24590421e-01 -4.76511896e-01 1.04213798e+00 -6.58145994e-02
-6.67523980e-01 3.01991016e-01 -2.20756307e-01 -6.63682997e-01
-3.25756878e-01 1.72348976e-01 1.37592763e-01 7.18101114e-02
7.22961426e-01 -1.07735610e+00 -8.48745048e-01 4.28719074e-02
-5.32081664e-01 -5.04921734e-01 9.32735443e-01 6.22534096e-01
7.74228334e-01 -1.13294967e-01 6.82670295e-01 -7.11229563e-01
7.30298936e-01 -4.54019159e-01 -1.32820740e-01 -1.45606510e-02
-6.74098432e-01 -1.93141447e-03 4.71091390e-01 -2.85335273e-01
-6.45304441e-01 1.30050346e-01 -5.80815077e-01 -1.84144363e-01
-6.49048507e-01 2.19877064e-01 -4.83431935e-01 -1.69470366e-02
7.01114237e-01 1.40955120e-01 3.39490354e-01 -8.03524435e-01
-1.37500642e-02 7.79351175e-01 2.89940983e-01 -1.98395148e-01
3.62380117e-01 3.15359771e-01 9.06233862e-02 -7.43374407e-01
-9.00407314e-01 -7.53286630e-02 -1.31186473e+00 -1.33392941e-02
1.14074624e+00 -8.12266588e-01 -9.46105123e-02 7.33859897e-01
-9.03077960e-01 -5.08785009e-01 -1.20392188e-01 5.76259255e-01
-4.07801270e-01 -1.85705330e-02 -6.27227426e-01 -4.35664430e-02
-8.03126872e-01 -1.23594308e+00 8.87977779e-01 2.13912129e-02
-1.26929075e-01 -1.24051809e+00 2.38242209e-01 1.37275323e-01
6.21519744e-01 2.39849761e-01 1.46013367e+00 -1.25660849e+00
2.14251056e-02 -2.39202052e-01 -5.16471744e-01 6.80864811e-01
1.03390120e-01 -5.36395788e-01 -8.34143519e-01 -1.06529884e-01
-9.81461816e-03 -1.99267343e-01 9.45463061e-01 7.69261539e-01
1.37555504e+00 -4.35790837e-01 -3.49224418e-01 5.84353685e-01
1.07652855e+00 2.30130836e-01 5.75137436e-01 7.57580519e-01
4.09111172e-01 1.42587245e-01 -3.00866783e-01 2.86383569e-01
4.44450468e-01 4.40336704e-01 7.07778990e-01 -3.82806621e-02
-4.38975096e-01 6.17816329e-01 2.38649085e-01 5.46232343e-01
-2.48738557e-01 5.24708152e-01 -9.99753833e-01 3.85491073e-01
-1.25324035e+00 -5.98974466e-01 -2.17916444e-01 1.54326367e+00
8.53671372e-01 2.30361804e-01 4.03874755e-01 1.36001974e-01
6.34614408e-01 -1.67090982e-01 -6.65304363e-01 -3.06656182e-01
-8.53955597e-02 6.93892479e-01 2.51636297e-01 -8.31256434e-02
-1.26367354e+00 3.80508363e-01 6.37517691e+00 -1.48076579e-01
-1.33564186e+00 4.65835243e-01 6.42606378e-01 5.22762351e-02
3.05171371e-01 -7.62031972e-01 -6.74016118e-01 4.30365652e-01
9.89872277e-01 1.16458483e-01 1.26710668e-01 1.08770502e+00
5.57310842e-02 5.58529139e-01 -1.17909503e+00 7.51153231e-01
9.73744988e-02 -9.93479311e-01 2.80285239e-01 1.52925506e-01
3.49372804e-01 3.88328344e-01 5.21208942e-02 3.98222357e-01
-3.72381764e-03 -1.20681310e+00 3.90947074e-01 8.09229970e-01
6.79272830e-01 -5.05483806e-01 9.94263887e-01 -4.96572554e-02
-7.75865376e-01 -4.77248937e-01 -4.23469156e-01 4.55629118e-02
-1.40431374e-01 9.28311884e-01 -1.13837326e+00 -9.99813452e-02
1.28594506e+00 7.76359141e-01 -1.03244030e+00 1.27778113e+00
-3.37440544e-03 5.82232475e-01 -9.62860212e-02 1.94730446e-01
1.30058080e-01 5.71922138e-02 3.16395432e-01 1.25241709e+00
5.86178303e-01 1.34086683e-01 9.96802598e-02 9.67741668e-01
3.16840112e-02 1.43850461e-01 -3.17890346e-01 7.93942511e-02
-1.99196026e-01 1.05329239e+00 -5.36732852e-01 -1.92211047e-01
-3.65502536e-01 8.42760503e-01 1.30574986e-01 6.55212402e-02
-3.27182025e-01 -9.01066065e-02 7.69380510e-01 5.84035277e-01
5.24510682e-01 -1.29715979e-01 -4.74509418e-01 -8.41471910e-01
-8.30541551e-02 -6.63824677e-01 4.97781545e-01 -8.19729507e-01
-1.79871428e+00 9.62432027e-01 -2.05840275e-01 -7.13572145e-01
6.92561343e-02 -1.39858735e+00 -7.05073893e-01 7.19779670e-01
-1.40563524e+00 -1.20911217e+00 -4.68306184e-01 9.79123116e-01
4.10111219e-01 -8.51611435e-01 1.09725952e+00 4.56363082e-01
-5.47692239e-01 2.42756963e-01 -7.55907893e-02 5.52139878e-01
5.63326240e-01 -1.40301764e+00 2.40556866e-01 1.69305578e-01
-4.04197246e-01 4.24489588e-01 7.66787529e-02 -6.29334569e-01
-7.12639868e-01 -1.82271016e+00 7.79291928e-01 -2.02449948e-01
5.58303297e-01 -7.30393752e-02 -8.56681049e-01 9.67153668e-01
-1.34466320e-01 6.12092793e-01 8.48637998e-01 -1.31104469e-01
-4.88296539e-01 -4.66499865e-01 -1.37235880e+00 1.06585689e-01
1.13100946e+00 -2.01804563e-01 -1.14706302e+00 5.54897308e-01
3.26975524e-01 -1.65626720e-01 -1.21317375e+00 3.43195677e-01
6.47940040e-01 -8.93519104e-01 1.37238741e+00 -1.02448368e+00
4.57467467e-01 3.01580876e-01 -1.86466560e-01 -1.48030412e+00
-6.39542460e-01 4.59989935e-01 -2.62474597e-01 6.62415087e-01
2.41990268e-01 -7.29201496e-01 4.89940256e-01 3.89721364e-01
-4.44565833e-01 -8.98381591e-01 -1.13749909e+00 -7.63462603e-01
6.60989940e-01 -1.71955958e-01 7.61202157e-01 6.78596318e-01
-4.96108055e-01 2.32354909e-01 3.78176898e-01 3.61603886e-01
5.92080891e-01 -3.52637082e-01 1.27341188e-02 -2.27763247e+00
2.34424248e-01 -5.88012695e-01 -1.03131461e+00 -2.95011789e-01
3.15667331e-01 -1.45013082e+00 -4.97506738e-01 -1.56588781e+00
2.06448853e-01 -2.94095367e-01 -5.20576835e-01 6.28091991e-01
1.78863183e-01 2.54027158e-01 -2.01598689e-01 -2.99731195e-02
-2.46455103e-01 5.44967771e-01 1.43455553e+00 -5.27861536e-01
-3.00244868e-01 1.46192878e-01 -4.73843694e-01 9.32500362e-01
1.09218812e+00 -4.07662392e-01 -2.82510102e-01 -3.45033586e-01
-2.77832091e-01 -5.47079682e-01 9.08835649e-01 -1.46328938e+00
-2.90216982e-01 3.89030963e-01 1.35412967e+00 -3.30225319e-01
9.52045694e-02 -9.13353682e-01 -2.03696147e-01 6.92044914e-01
-2.48299599e-01 1.01190157e-01 1.34313896e-01 2.09816471e-01
9.82401222e-02 -5.05411439e-02 9.73559678e-01 -5.95600843e-01
-5.44796348e-01 7.63934076e-01 -5.31078160e-01 4.51890267e-02
8.51770282e-01 -2.06667006e-01 6.45942101e-03 2.46045455e-01
-1.27516377e+00 -6.05092868e-02 -1.30893871e-01 3.28370988e-01
6.60549164e-01 -1.47747362e+00 -7.99787641e-01 3.36558193e-01
-1.33896200e-03 7.96659850e-03 1.16001971e-01 8.56179059e-01
-6.28745139e-01 6.11600995e-01 -1.10770309e+00 -6.84612274e-01
-1.10126483e+00 3.42455864e-01 1.03858888e+00 -1.69980854e-01
-1.11331344e+00 7.46000767e-01 6.19201697e-02 -4.64205474e-01
2.91030675e-01 -7.33446479e-01 -6.86327875e-01 2.24451609e-02
7.98482060e-01 2.30227962e-01 5.59036374e-01 -4.79247719e-01
-5.87732971e-01 4.15563464e-01 -4.18417752e-01 4.25482303e-01
2.09862781e+00 6.32979199e-02 -2.92694241e-01 1.08613290e-01
1.34075630e+00 -9.37198222e-01 -1.05549848e+00 -8.12401697e-02
1.88196391e-01 2.83971190e-01 4.90994304e-01 -1.08023882e+00
-1.71741772e+00 1.04772425e+00 1.65034437e+00 -2.26025339e-02
9.39410031e-01 2.34376043e-01 1.07233310e+00 5.11840463e-01
2.64669389e-01 -8.68977904e-01 9.00277123e-03 4.78792369e-01
1.35608637e+00 -1.03574109e+00 -9.06608105e-02 1.87749043e-01
-2.13011906e-01 1.86294484e+00 6.96159899e-01 -4.98273849e-01
1.28018153e+00 9.33741182e-02 1.81803823e-01 -5.51236808e-01
-1.97230726e-01 -9.02605504e-02 4.30644900e-01 1.01649761e+00
2.31754974e-01 1.74624458e-01 -2.29544431e-01 1.62417424e+00
-2.88997144e-01 3.44161719e-01 1.35768384e-01 5.95368028e-01
-6.79917276e-01 -9.77084875e-01 -3.85564625e-01 1.02008498e+00
-3.75292063e-01 1.75008789e-01 -6.67745948e-01 8.67122650e-01
7.85870016e-01 6.40787780e-02 3.71396482e-01 -1.67399302e-01
4.80585575e-01 5.19356608e-01 6.28558993e-01 -6.02856398e-01
-4.52549279e-01 -4.27048385e-01 -1.23668924e-01 -5.35257936e-01
-3.72191072e-01 -7.69531906e-01 -1.29332817e+00 2.26950288e-01
3.60953897e-01 -5.09900093e-01 4.43106353e-01 1.18177748e+00
4.53874141e-01 7.87753522e-01 2.04790473e-01 -9.75722969e-01
-5.69424212e-01 -1.22050214e+00 -8.84009242e-01 1.83008134e-01
5.18941283e-01 -9.49686766e-01 -1.22449599e-01 2.53058046e-01] | [14.195489883422852, -1.7570874691009521] |
5647069d-c24b-4b88-ab80-fceea2d35ae1 | fsgan-subject-agnostic-face-swapping-and | 1908.05932 | null | https://arxiv.org/abs/1908.05932v1 | https://arxiv.org/pdf/1908.05932v1.pdf | FSGAN: Subject Agnostic Face Swapping and Reenactment | We present Face Swapping GAN (FSGAN) for face swapping and reenactment. Unlike previous work, FSGAN is subject agnostic and can be applied to pairs of faces without requiring training on those faces. To this end, we describe a number of technical contributions. We derive a novel recurrent neural network (RNN)-based approach for face reenactment which adjusts for both pose and expression variations and can be applied to a single image or a video sequence. For video sequences, we introduce continuous interpolation of the face views based on reenactment, Delaunay Triangulation, and barycentric coordinates. Occluded face regions are handled by a face completion network. Finally, we use a face blending network for seamless blending of the two faces while preserving target skin color and lighting conditions. This network uses a novel Poisson blending loss which combines Poisson optimization with perceptual loss. We compare our approach to existing state-of-the-art systems and show our results to be both qualitatively and quantitatively superior. | ['Yuval Nirkin', 'Yosi Keller', 'Tal Hassner'] | 2019-08-16 | fsgan-subject-agnostic-face-swapping-and-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Nirkin_FSGAN_Subject_Agnostic_Face_Swapping_and_Reenactment_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Nirkin_FSGAN_Subject_Agnostic_Face_Swapping_and_Reenactment_ICCV_2019_paper.pdf | iccv-2019-10 | ['face-reenactment', 'facial-inpainting'] | ['computer-vision', 'computer-vision'] | [ 4.73199457e-01 2.45579168e-01 2.41775736e-01 -5.91031194e-01
-7.67055631e-01 -6.12253964e-01 4.92502421e-01 -6.91407442e-01
-1.37686178e-01 6.06805801e-01 4.10253443e-02 1.00854971e-01
3.07915181e-01 -4.50071305e-01 -8.88569355e-01 -5.70071340e-01
1.79788828e-01 2.03309029e-01 -2.69410789e-01 -2.25176662e-01
-2.15319350e-01 9.22292292e-01 -1.58840263e+00 3.24513018e-01
5.04545569e-01 9.81446028e-01 -3.14336509e-01 6.89985991e-01
2.47351080e-01 5.45301437e-01 -4.04640883e-01 -7.00454831e-01
6.91517293e-01 -6.26225352e-01 -5.94583929e-01 3.20389003e-01
1.19260991e+00 -6.14597797e-01 -2.60845065e-01 8.06208193e-01
7.20532656e-01 1.49136513e-01 4.81028378e-01 -1.65560853e+00
-7.61884391e-01 4.23715226e-02 -8.86177719e-01 -4.46587622e-01
4.26695824e-01 -6.77525178e-02 5.54879427e-01 -1.14355445e+00
7.49147177e-01 1.57858300e+00 8.75833392e-01 1.07354319e+00
-1.48910260e+00 -8.99558246e-01 6.75353259e-02 -1.65432617e-01
-1.38321769e+00 -8.23103547e-01 9.08852160e-01 -2.99741983e-01
4.81517673e-01 2.95526057e-01 6.40600145e-01 1.11739206e+00
6.63883984e-03 5.33815086e-01 1.14205682e+00 -4.41713095e-01
-3.98119017e-02 2.41141841e-02 -5.99182129e-01 9.15923536e-01
-3.08844179e-01 -5.31709343e-02 -5.00883222e-01 -2.25423679e-01
1.13178301e+00 1.01304390e-01 -4.81646717e-01 -6.65300906e-01
-6.82176590e-01 7.14396596e-01 4.40604538e-01 -3.67220305e-02
-2.37481222e-01 3.84657681e-01 2.48963803e-01 3.34932089e-01
6.14223242e-01 2.48063095e-02 -2.08645776e-01 3.45556945e-01
-1.15199232e+00 3.05251747e-01 6.12110913e-01 7.66100705e-01
7.19640076e-01 2.74309754e-01 -1.50176123e-01 1.09219742e+00
3.03419828e-01 4.72623616e-01 6.19483478e-02 -1.51687205e+00
-1.02813959e-01 1.43238708e-01 -9.17503908e-02 -9.98472810e-01
-9.84406099e-02 5.05548269e-02 -9.40821767e-01 6.92252338e-01
8.10520053e-02 -2.36974955e-01 -1.06146979e+00 2.15738678e+00
6.12782955e-01 4.37063187e-01 -8.41509327e-02 5.83038449e-01
8.78304899e-01 3.41816425e-01 -2.59356588e-01 -2.74702609e-01
1.16189086e+00 -9.92584884e-01 -7.32555568e-01 7.16261938e-02
-7.87563697e-02 -8.87286365e-01 6.84673190e-01 2.45121300e-01
-1.51930308e+00 -3.39678347e-01 -8.29840362e-01 -3.30966681e-01
-1.08501635e-01 5.61673455e-02 1.68973997e-01 8.10810804e-01
-1.80647194e+00 7.51467228e-01 -7.41765499e-01 -4.47343230e-01
6.07865453e-01 7.21885622e-01 -7.20613778e-01 2.92465966e-02
-7.34196544e-01 6.00940168e-01 -3.26566935e-01 2.49812499e-01
-8.62314284e-01 -9.25072670e-01 -1.09002006e+00 -1.72251940e-01
2.00376064e-01 -8.92346680e-01 1.26312518e+00 -1.53737235e+00
-1.91290641e+00 1.12102628e+00 -4.09672916e-01 -1.92305297e-01
7.78971434e-01 -9.98352468e-02 -1.85142756e-01 1.60978988e-01
2.76601519e-02 1.15886843e+00 1.36190903e+00 -1.51902223e+00
-2.25460455e-01 -2.83387333e-01 -1.98882878e-01 3.13570261e-01
-7.73394480e-02 2.04344794e-01 -7.26487756e-01 -9.12347436e-01
-8.09338689e-02 -9.49687719e-01 1.32618636e-01 6.86414599e-01
-3.80327165e-01 2.33237863e-01 1.18885350e+00 -9.71603692e-01
5.01713872e-01 -2.25483799e+00 3.20979804e-01 1.94594234e-01
2.76922673e-01 2.04099476e-01 -5.57944238e-01 1.89764217e-01
-4.85813826e-01 -1.09121561e-01 -4.78322506e-01 -1.23189330e+00
-1.60282701e-01 1.10614233e-01 -2.48616084e-01 6.76168084e-01
2.62449324e-01 8.26642573e-01 -5.43198347e-01 -2.06251860e-01
2.62617141e-01 1.13145971e+00 -7.91996419e-01 4.35550213e-01
-4.46005464e-02 5.86562812e-01 3.84534806e-01 7.74171352e-01
1.22842157e+00 2.88997203e-01 2.57020593e-01 -3.48069549e-01
7.59772584e-02 -2.61661828e-01 -9.63404894e-01 1.62073457e+00
-6.32795691e-01 5.93447208e-01 6.18103921e-01 -3.26066822e-01
7.47769177e-01 3.92088860e-01 4.87417787e-01 -2.89633483e-01
1.40232027e-01 1.67521462e-01 -5.38114130e-01 -1.52587667e-01
2.72229761e-01 -1.70720130e-01 5.22920728e-01 5.34560919e-01
1.53347254e-01 -3.86157721e-01 -2.29255453e-01 -3.15890871e-02
6.41689062e-01 3.36366028e-01 1.59584209e-02 -1.30744398e-01
4.73852307e-01 -9.28617060e-01 6.42803669e-01 2.10413620e-01
-3.60151008e-02 1.14790058e+00 6.32616818e-01 -4.08551961e-01
-1.03222644e+00 -1.18826783e+00 2.28911266e-02 1.00008225e+00
-1.55157924e-01 -1.04327157e-01 -1.21221852e+00 -6.13309801e-01
-1.43360853e-01 3.76433790e-01 -8.69100213e-01 1.29560485e-01
-6.66526258e-01 -3.76994759e-01 4.88344491e-01 4.61330384e-01
6.14560068e-01 -9.85261321e-01 -2.79375553e-01 -1.26005039e-01
-1.26821533e-01 -9.24226284e-01 -9.74530339e-01 -2.77350545e-01
-5.93166649e-01 -1.10170722e+00 -1.05206096e+00 -1.01293790e+00
9.88466680e-01 1.86056003e-01 9.56887960e-01 2.28488326e-01
-2.87652850e-01 6.15597010e-01 2.51402892e-02 -8.26566815e-02
-3.66948515e-01 -1.64299339e-01 2.48551250e-01 4.61041838e-01
-2.24100366e-01 -9.30509865e-01 -6.52422309e-01 2.41564810e-01
-1.01825106e+00 7.79455304e-02 4.31847423e-02 7.78203487e-01
4.72528934e-01 -3.32401007e-01 2.65986115e-01 -9.15347278e-01
3.70469809e-01 -9.15602371e-02 -5.72936475e-01 2.60884225e-01
-2.13917360e-01 -3.23673964e-01 4.28483754e-01 -3.25878471e-01
-1.09316504e+00 3.24442536e-01 -2.68809468e-01 -9.65060353e-01
2.08201274e-01 -4.73304242e-01 -3.18933189e-01 -6.67706907e-01
3.51067275e-01 -1.44405499e-01 3.52745414e-01 -2.81179994e-01
6.01443708e-01 3.38004202e-01 8.56515944e-01 -4.26037967e-01
8.79142225e-01 8.78960729e-01 -3.32650319e-02 -7.98148870e-01
-4.14875627e-01 1.09794214e-01 -7.00696886e-01 -2.33505011e-01
7.78131127e-01 -7.24183440e-01 -9.54506040e-01 8.53351533e-01
-1.40590692e+00 -4.56557393e-01 -3.66101444e-01 -4.14998792e-02
-7.03253984e-01 2.98762590e-01 -5.45780897e-01 -7.49368966e-01
-4.16864306e-01 -1.07207108e+00 1.25324631e+00 2.95291036e-01
-1.01407163e-01 -1.00857520e+00 1.34459466e-01 2.11078420e-01
4.46113795e-01 6.51210964e-01 4.48935956e-01 -5.86770102e-02
-4.63064164e-01 1.04356550e-01 -2.66274810e-01 4.83745694e-01
4.54741627e-01 3.04986686e-01 -1.34852648e+00 -6.77510679e-01
1.37073219e-01 -2.99207747e-01 8.66496265e-01 5.69820464e-01
1.18938887e+00 -6.35716617e-01 -2.35368818e-01 1.07629883e+00
1.12553489e+00 1.74307793e-01 8.58653903e-01 -1.10801570e-01
9.30406690e-01 9.22841549e-01 -7.03116357e-02 2.31782496e-01
2.88189709e-01 6.52492642e-01 4.81009454e-01 -5.44598222e-01
-3.43456745e-01 -3.58816892e-01 2.77511865e-01 3.23004991e-01
2.47331150e-02 -3.09318572e-01 -3.39289516e-01 5.19997597e-01
-1.66713738e+00 -7.24218309e-01 4.01597589e-01 2.18771744e+00
5.95065892e-01 -5.88037014e-01 1.83315173e-01 -2.07301080e-01
1.00361991e+00 1.07065231e-01 -6.42516375e-01 -4.55702901e-01
-1.84704304e-01 3.95163476e-01 3.01979095e-01 9.02779698e-01
-9.40335214e-01 9.98522818e-01 6.90518618e+00 6.05087876e-01
-1.15612423e+00 2.78508753e-01 1.00479734e+00 -3.53728712e-01
-4.58116829e-01 -3.08632046e-01 -3.30311149e-01 1.51820853e-01
3.53849679e-01 3.34239483e-01 8.44325900e-01 4.83715266e-01
-6.39026836e-02 1.77902579e-01 -1.21027195e+00 1.21756959e+00
6.30854964e-01 -1.31934321e+00 -2.27515516e-03 -1.19868413e-01
9.71450090e-01 -2.44075269e-01 5.62480628e-01 -2.05024824e-01
3.84677500e-01 -1.37458801e+00 6.65524125e-01 5.11991441e-01
1.29890966e+00 -9.13522840e-01 3.74285519e-01 -5.55085182e-01
-1.11495459e+00 1.89326420e-01 1.33716129e-02 4.03587341e-01
2.47763067e-01 2.14920789e-01 -5.62726378e-01 3.88427764e-01
7.34675646e-01 6.91383362e-01 -1.68639451e-01 5.39259076e-01
-1.84775248e-01 -7.43623674e-02 -3.64283681e-01 6.42255962e-01
-2.04614222e-01 -2.59823292e-01 5.91895401e-01 7.67057300e-01
4.01594609e-01 -1.36977192e-02 -1.39421657e-01 9.18939710e-01
-6.02861404e-01 -8.75039212e-03 -6.25321984e-01 4.23990726e-01
4.36787993e-01 1.47835112e+00 -5.75299084e-01 9.02049243e-03
-2.29567289e-01 1.50269413e+00 3.78498107e-01 4.85219985e-01
-6.13646865e-01 -2.44317710e-01 9.70559716e-01 1.86821312e-01
4.70771581e-01 9.84380171e-02 2.07005143e-02 -1.01845574e+00
1.66076273e-01 -8.83899927e-01 2.01897934e-01 -9.00466502e-01
-1.35877705e+00 9.15968657e-01 -7.15874955e-02 -8.07535887e-01
-3.03027987e-01 -4.80403066e-01 -8.10300708e-01 8.07307959e-01
-1.48081040e+00 -1.57172072e+00 -2.85693735e-01 7.06056297e-01
3.51096153e-01 -2.30487928e-01 6.64053202e-01 2.85655588e-01
-5.57677686e-01 9.74605143e-01 -9.48702395e-02 9.49368104e-02
8.51258636e-01 -8.79739642e-01 7.63038278e-01 6.93195581e-01
-1.53618380e-01 6.31415546e-01 5.08636475e-01 -6.20959461e-01
-1.11395216e+00 -1.40261292e+00 6.36996746e-01 -2.70264059e-01
6.40100092e-02 -6.14368141e-01 -8.02364945e-01 1.06581450e+00
5.06886244e-01 1.74395353e-01 4.61559981e-01 -2.43373439e-01
-5.86050510e-01 -2.62698144e-01 -1.70889866e+00 8.80961180e-01
9.97474730e-01 -5.61624169e-01 8.12112689e-02 2.07083419e-01
5.78837812e-01 -6.87242806e-01 -5.04088044e-01 3.84769171e-01
8.03513527e-01 -1.18397677e+00 8.89247358e-01 -2.11274073e-01
2.53797621e-01 -4.38517749e-01 -1.54631153e-01 -1.30047810e+00
-5.74615188e-02 -1.30630875e+00 2.73199618e-01 1.47445071e+00
8.91329795e-02 -7.59019971e-01 7.25231409e-01 7.35112786e-01
1.60505965e-01 -7.72462845e-01 -1.10871577e+00 -4.96100157e-01
3.59758660e-02 -8.72865319e-02 7.68954635e-01 9.19997394e-01
-3.96485001e-01 3.35671529e-02 -7.83052266e-01 -1.53064970e-02
7.08620191e-01 -1.68135881e-01 7.55042791e-01 -7.78867602e-01
-6.48513958e-02 -2.33116582e-01 -2.67713606e-01 -6.80261910e-01
5.47029376e-01 -7.30474114e-01 -3.81857380e-02 -1.10692191e+00
1.34309083e-01 -2.06316561e-01 1.77973181e-01 7.83920288e-01
2.65641481e-01 9.91274059e-01 2.61572063e-01 -1.19575016e-01
4.30176109e-02 7.09742665e-01 1.08944643e+00 3.80977094e-02
-2.55868942e-01 -2.11904198e-01 -7.16778278e-01 7.87125528e-01
6.43736899e-01 -2.98954546e-01 -2.94383705e-01 -4.48359758e-01
3.08743771e-02 4.28866223e-02 5.78725517e-01 -7.77519584e-01
7.59660527e-02 5.11236116e-02 5.70968211e-01 -2.16623887e-01
7.73175776e-01 -8.45131040e-01 5.52949607e-01 3.23220938e-01
-2.67011255e-01 1.73938319e-01 3.58318001e-01 3.41738135e-01
1.72178909e-01 3.58631015e-02 1.33149838e+00 -3.69255408e-03
-5.42565295e-03 6.33105636e-01 -1.41170055e-01 -9.51239616e-02
1.12412477e+00 -3.17251831e-01 8.94786119e-02 -7.69400179e-01
-5.27756095e-01 5.76888435e-02 8.65746737e-01 3.74646395e-01
8.70035350e-01 -1.58481979e+00 -9.58700001e-01 7.35750914e-01
-3.68702710e-01 -1.09220974e-01 4.13011521e-01 5.99841535e-01
-7.65111208e-01 -1.86462164e-01 -4.22518045e-01 -4.60947961e-01
-1.71938109e+00 3.23244512e-01 7.66449332e-01 3.16217333e-01
-4.28982675e-01 1.13746953e+00 5.17749667e-01 -7.40545273e-01
1.25698090e-01 6.26433045e-02 3.95922922e-02 -9.40384194e-02
5.62261522e-01 3.54937166e-01 -2.74574962e-02 -8.68315995e-01
-3.97720009e-01 8.81638765e-01 -2.56385803e-01 -3.94328803e-01
1.36807191e+00 -1.76252589e-01 -5.13486207e-01 1.53081656e-01
1.42184043e+00 1.11295313e-01 -1.66014254e+00 -8.89078751e-02
-9.08376813e-01 -7.74343014e-01 -2.21784502e-01 -5.31093717e-01
-1.64320779e+00 6.24885142e-01 6.74822271e-01 -2.62320220e-01
1.41605961e+00 -1.56356886e-01 8.56902540e-01 -1.13379329e-01
-1.16568431e-01 -8.77314806e-01 1.75662816e-01 3.18782002e-01
1.14354455e+00 -9.10337031e-01 -5.36490083e-02 -6.01928830e-01
-4.73200113e-01 1.09515691e+00 5.62881589e-01 -2.65045077e-01
6.61753893e-01 4.07636940e-01 1.94457859e-01 -7.20371120e-03
-4.10399735e-01 3.33929092e-01 3.40062678e-01 8.95224869e-01
3.73844951e-01 -2.61175036e-01 2.70268261e-01 -2.68521626e-02
-3.33080947e-01 -1.54572293e-01 4.75709707e-01 7.20605671e-01
1.71654761e-01 -1.13887596e+00 -4.88780677e-01 2.16102704e-01
-4.57214653e-01 -1.17537081e-01 -4.43174601e-01 5.34462869e-01
2.55976617e-01 6.42896473e-01 4.01123822e-01 -1.16165549e-01
1.19779654e-01 -1.44722149e-01 9.30054903e-01 -4.25123245e-01
-5.37201166e-01 1.81590527e-01 -2.54729539e-01 -6.58734560e-01
-5.99324107e-01 -6.85478270e-01 -8.97255361e-01 -6.07087076e-01
-1.12278841e-01 -3.59306276e-01 5.16782105e-01 6.57801151e-01
4.85398054e-01 2.81313956e-01 9.34939444e-01 -1.20970273e+00
5.33955246e-02 -4.44754779e-01 -7.31758893e-01 1.95851013e-01
8.14829350e-01 -4.58704352e-01 -3.31633806e-01 2.80164093e-01] | [12.709162712097168, -0.17633309960365295] |
f4800640-50c1-4778-922b-f08b313fddea | 3d-graph-contrastive-learning-for-molecular | 2208.06360 | null | https://arxiv.org/abs/2208.06360v2 | https://arxiv.org/pdf/2208.06360v2.pdf | 3D Graph Contrastive Learning for Molecular Property Prediction | Self-supervised learning (SSL) is a method that learns the data representation by utilizing supervision inherent in the data. This learning method is in the spotlight in the drug field, lacking annotated data due to time-consuming and expensive experiments. SSL using enormous unlabeled data has shown excellent performance for molecular property prediction, but a few issues exist. (1) Existing SSL models are large-scale; there is a limitation to implementing SSL where the computing resource is insufficient. (2) In most cases, they do not utilize 3D structural information for molecular representation learning. The activity of a drug is closely related to the structure of the drug molecule. Nevertheless, most current models do not use 3D information or use it partially. (3) Previous models that apply contrastive learning to molecules use the augmentation of permuting atoms and bonds. Therefore, molecules having different characteristics can be in the same positive samples. We propose a novel contrastive learning framework, small-scale 3D Graph Contrastive Learning (3DGCL) for molecular property prediction, to solve the above problems. 3DGCL learns the molecular representation by reflecting the molecule's structure through the pre-training process that does not change the semantics of the drug. Using only 1,128 samples for pre-train data and 1 million model parameters, we achieved the state-of-the-art or comparable performance in four regression benchmark datasets. Extensive experiments demonstrate that 3D structural information based on chemical knowledge is essential to molecular representation learning for property prediction. | ['Sunyoung Kwon', 'Kisung Moon'] | 2022-05-31 | null | null | null | null | ['molecular-property-prediction'] | ['miscellaneous'] | [ 5.43706775e-01 -1.53748840e-01 -1.16370916e+00 -3.12208951e-01
-4.44964260e-01 -5.12263119e-01 3.03950638e-01 7.26351678e-01
-2.06179217e-01 1.16521096e+00 2.06047408e-02 -5.47024310e-01
-1.20243207e-01 -8.94691348e-01 -8.60361636e-01 -9.32186007e-01
-2.30544642e-01 4.01791334e-01 2.68518776e-01 -3.18811178e-01
3.06539059e-01 7.57915258e-01 -1.17546999e+00 3.58972520e-01
9.99302745e-01 7.42208421e-01 3.33723515e-01 1.65061921e-01
-3.19288045e-01 8.48462939e-01 -3.44549179e-01 1.08933978e-01
2.07359388e-01 -6.69552922e-01 -8.71212661e-01 -1.29786178e-01
3.71541202e-01 9.54770669e-02 -2.49229029e-01 1.00506771e+00
4.58095402e-01 4.00366411e-02 9.05519187e-01 -1.03378630e+00
-6.31075263e-01 3.13154548e-01 -6.25543714e-01 4.94434685e-02
6.01573527e-01 3.23894136e-02 1.14640045e+00 -9.57051218e-01
7.79985964e-01 1.05058897e+00 4.89357471e-01 6.49844170e-01
-1.22021806e+00 -7.61875093e-01 1.99676454e-01 4.18081522e-01
-1.41329408e+00 -7.29586259e-02 9.00864601e-01 -4.75270391e-01
1.11223042e+00 2.09063187e-01 6.92441881e-01 9.84326184e-01
2.69575983e-01 5.24425924e-01 1.10229588e+00 -3.12058002e-01
4.39224988e-01 6.96165338e-02 2.72025138e-01 9.40264642e-01
3.34085792e-01 1.24100350e-01 -6.04619861e-01 -4.99495238e-01
4.68655288e-01 3.54476005e-01 -3.57465118e-01 -5.98691702e-01
-8.58452380e-01 1.00660741e+00 7.07528293e-01 1.13776125e-01
-1.67796239e-01 -3.69176835e-01 3.70689929e-01 4.51001078e-01
4.65896666e-01 6.37713969e-01 -8.01412106e-01 3.15436214e-01
-4.92716283e-01 -7.04347193e-02 6.88817978e-01 7.85318196e-01
1.02059412e+00 -1.56832654e-02 2.60079890e-01 5.26697278e-01
3.63882303e-01 2.29743540e-01 5.02650678e-01 -3.89378853e-02
4.28701252e-01 1.08396220e+00 -3.68111253e-01 -8.35412681e-01
-5.85334659e-01 -2.46305540e-01 -8.86155605e-01 1.69011950e-01
1.29846275e-01 1.70779780e-01 -1.12294555e+00 1.56612027e+00
3.68549496e-01 3.81796390e-01 2.94874161e-01 5.76436162e-01
1.22873425e+00 7.43208289e-01 4.55133140e-01 -5.06927013e-01
8.56892288e-01 -8.86880934e-01 -4.95505929e-01 5.38726002e-02
1.12384260e+00 -5.78683913e-01 9.08781469e-01 3.23856503e-01
-4.64758664e-01 -4.76671398e-01 -1.40125024e+00 1.80498332e-01
-7.57356763e-01 -2.37935334e-01 1.07651412e+00 7.51807868e-01
-3.83084774e-01 1.01646686e+00 -5.11088073e-01 -5.96052855e-02
7.51854062e-01 8.16576064e-01 -7.59736419e-01 -3.52464497e-01
-1.34414577e+00 8.41224253e-01 6.00044668e-01 -2.80134976e-01
-7.89296985e-01 -8.95695269e-01 -1.03460348e+00 -2.02899843e-01
5.55153728e-01 -3.32086027e-01 5.24590135e-01 -7.59944260e-01
-1.38891971e+00 5.37853718e-01 -1.47072300e-01 -1.46403015e-01
1.81570113e-01 1.88546702e-01 -5.28017521e-01 1.28546089e-01
-5.64437471e-02 3.67017925e-01 6.53953195e-01 -1.19001055e+00
-2.32759148e-01 -5.76591671e-01 -3.49624828e-02 1.50916159e-01
-3.32286954e-01 -5.54571152e-01 -1.54037654e-01 -6.18952632e-01
1.09989561e-01 -8.73565972e-01 -4.45614219e-01 -1.09304912e-01
-4.57918286e-01 -4.17391181e-01 9.04758811e-01 -1.47159591e-01
1.12164032e+00 -1.81611335e+00 1.45217925e-01 4.68898833e-01
3.05018812e-01 6.05940163e-01 -4.34426814e-01 7.11027026e-01
-6.32197857e-01 2.70800829e-01 -3.20557475e-01 2.63012528e-01
-5.79889596e-01 9.84425023e-02 -1.17663383e-01 4.63405311e-01
3.13570589e-01 7.96152949e-01 -1.28568733e+00 -4.87910748e-01
7.01755062e-02 4.77425694e-01 -5.73093355e-01 1.70110777e-01
-3.76634419e-01 5.14373899e-01 -6.96385622e-01 9.04686809e-01
7.93747067e-01 -5.71980715e-01 5.86923182e-01 -4.30979997e-01
1.67724326e-01 4.05059755e-01 -9.56565261e-01 1.69714630e+00
-5.34761436e-02 -4.22419198e-02 -7.89626420e-01 -1.13300884e+00
9.56599236e-01 2.40637869e-01 7.62308300e-01 -6.70723140e-01
-2.85019219e-01 3.45397145e-01 1.25249580e-01 -4.20944452e-01
-9.77085307e-02 -2.37038419e-01 4.81879830e-01 2.01350346e-01
1.39142379e-01 1.32756615e-02 1.52401671e-01 9.75300968e-02
1.08751774e+00 3.39794010e-01 7.96189070e-01 -7.57001787e-02
7.13041365e-01 9.61856619e-02 7.65457094e-01 3.98876101e-01
1.76679403e-01 1.94945633e-01 5.04816949e-01 -5.53779602e-01
-6.89887583e-01 -7.36814022e-01 -2.90438324e-01 9.22101736e-01
3.04430366e-01 -7.72845566e-01 -2.71120042e-01 -1.13802814e+00
1.81904212e-02 3.04679662e-01 -6.95999086e-01 -4.52594638e-01
-3.32772166e-01 -1.05559945e+00 8.41357186e-02 2.55854279e-01
3.68079662e-01 -1.01045144e+00 2.85189182e-01 3.20525944e-01
4.84330297e-01 -7.16661632e-01 -4.06698853e-01 5.67901731e-01
-1.12201941e+00 -1.58103001e+00 -3.76577079e-01 -9.56009448e-01
9.53901649e-01 4.81229514e-01 8.46653819e-01 2.04485491e-01
-3.13889384e-01 -1.82051927e-01 -3.62653673e-01 -4.22970325e-01
-4.05493021e-01 8.45267028e-02 1.30755231e-01 -1.24244183e-01
4.77883309e-01 -5.62261224e-01 -5.84472477e-01 2.74406761e-01
-7.72715867e-01 -1.28573969e-01 7.60512948e-01 1.13981164e+00
9.52942967e-01 1.52174607e-01 8.01029384e-01 -1.60585535e+00
4.06809807e-01 -5.31461775e-01 -4.00604755e-01 3.55674773e-01
-9.68443036e-01 2.46727139e-01 8.26764107e-01 -4.86240417e-01
-6.32801235e-01 4.75536257e-01 -1.48771495e-01 -1.84979632e-01
-3.38096879e-02 8.51893127e-01 -3.80053222e-01 -3.85750294e-01
7.33561218e-01 2.58068293e-01 1.01841047e-01 -5.31414807e-01
-4.12497707e-02 5.89851201e-01 -2.22072273e-01 -4.62190539e-01
6.12701416e-01 1.53003156e-01 6.73894823e-01 -8.14866424e-01
-1.07311106e+00 -6.13527954e-01 -8.11438084e-01 3.67514461e-01
4.41225111e-01 -1.03313148e+00 -9.08718288e-01 2.20620170e-01
-7.24984348e-01 -1.88208386e-01 -1.15791775e-01 7.17860818e-01
-3.27874184e-01 7.05786824e-01 -4.05082643e-01 -3.52255613e-01
-3.97395521e-01 -1.23975813e+00 6.42064154e-01 -1.12138249e-01
1.02868602e-02 -1.10161209e+00 2.37077981e-01 4.41472024e-01
-8.06073397e-02 2.94182241e-01 1.55390847e+00 -9.03285861e-01
-4.31510866e-01 -2.49312669e-01 1.36174010e-02 2.70338923e-01
8.54633987e-01 -2.25419924e-01 -8.54843676e-01 -4.49915618e-01
-2.92380631e-01 -5.69572747e-01 8.75173211e-01 1.25694603e-01
1.31733537e+00 -3.29103351e-01 -4.68332440e-01 5.28855383e-01
1.42142737e+00 3.49754065e-01 6.02167964e-01 2.46262662e-02
1.07593954e+00 5.24627566e-01 8.17335606e-01 2.27111354e-01
-2.04498209e-02 5.80529094e-01 4.70581204e-01 -4.36454207e-01
-6.26347736e-02 -6.74195111e-01 2.77201325e-01 6.79084897e-01
-3.89407337e-01 -1.55197665e-01 -7.23448336e-01 -1.09759904e-01
-1.85307145e+00 -9.55375195e-01 -2.25509867e-01 2.39120078e+00
1.27760768e+00 1.58391863e-01 3.41216512e-02 1.09670721e-01
4.76889610e-01 7.73616135e-02 -1.07117951e+00 -7.82708824e-02
-1.56393915e-01 4.31352794e-01 6.58096731e-01 4.38436568e-01
-1.02761519e+00 1.12751830e+00 6.21549225e+00 1.10101402e+00
-1.28072166e+00 -3.67633969e-01 6.62692249e-01 4.82915550e-01
-3.16621065e-01 3.63581777e-02 -9.17127192e-01 3.50048929e-01
8.51756394e-01 -2.28275992e-02 7.88359269e-02 8.69643331e-01
1.63455933e-01 9.41688269e-02 -1.25588155e+00 1.03580749e+00
1.36639804e-01 -1.73465800e+00 6.16039395e-01 1.93111554e-01
7.51778305e-01 -6.76227063e-02 -8.97179125e-04 1.00911677e-01
-2.89560761e-02 -1.26604784e+00 -1.66610286e-01 1.75840586e-01
9.65147138e-01 -8.35935533e-01 7.00492680e-01 2.62398660e-01
-1.17927825e+00 3.03994477e-01 -5.08631051e-01 1.13572598e-01
-4.62295592e-01 3.24343294e-01 -1.31431293e+00 8.11744452e-01
1.87245294e-01 1.40284085e+00 -6.49890482e-01 9.53002512e-01
-2.95571148e-01 6.18748665e-01 2.63923649e-02 -2.16476217e-01
2.08919838e-01 -4.79238331e-01 9.48960036e-02 9.21527386e-01
-7.55639151e-02 3.48727889e-02 4.66321528e-01 4.23414707e-01
-4.24410045e-01 5.74135184e-01 -8.52008343e-01 -3.31892401e-01
2.99962550e-01 8.68031919e-01 -5.65050066e-01 -2.24596977e-01
-6.92130208e-01 7.92889535e-01 2.61886746e-01 2.85476774e-01
-4.83290195e-01 -2.57686675e-01 4.48136300e-01 2.30227649e-01
1.48324713e-01 -4.72521894e-02 7.60328770e-02 -8.58428776e-01
-3.11981440e-01 -9.86541569e-01 5.42460322e-01 -3.53546023e-01
-1.55667877e+00 4.91062582e-01 -3.38992625e-01 -1.56447148e+00
1.66431308e-01 -1.00556362e+00 -1.08609132e-01 6.55871570e-01
-1.98214078e+00 -1.03904235e+00 -4.54706475e-02 6.74274981e-01
3.76070559e-01 -5.11118650e-01 1.29375744e+00 3.21468294e-01
-5.53543687e-01 5.14796376e-01 1.53413415e-01 -1.20386697e-01
9.18756008e-01 -1.26471162e+00 2.38887463e-02 -9.72840488e-02
1.75277337e-01 7.56951809e-01 3.09839875e-01 -9.03447151e-01
-1.64021134e+00 -1.19743645e+00 6.82851672e-01 -2.99387902e-01
6.20913446e-01 -3.01444113e-01 -1.15485847e+00 2.76064575e-01
-3.47699910e-01 3.87969732e-01 1.56036770e+00 1.93145499e-01
-6.49633467e-01 4.26157862e-02 -1.15246093e+00 3.64700675e-01
1.18402874e+00 -5.90722024e-01 -2.90770054e-01 8.66137087e-01
7.28239596e-01 -1.59393162e-01 -1.10795283e+00 5.91425121e-01
2.94842929e-01 -4.69390571e-01 1.01477623e+00 -1.26476693e+00
2.87344962e-01 -4.62357908e-01 -6.96984539e-03 -1.13954222e+00
-2.97849029e-01 -4.75066632e-01 -2.70341992e-01 6.25156939e-01
8.20310712e-01 -6.43261790e-01 1.13220763e+00 3.36195320e-01
-3.48839574e-02 -1.09446120e+00 -6.77565634e-01 -8.76699507e-01
1.14726350e-02 1.61589891e-01 5.60760379e-01 1.37231958e+00
1.87479734e-01 9.55170393e-01 -4.97287899e-01 3.12601030e-02
4.19918925e-01 3.22437078e-01 5.72641194e-01 -1.49441624e+00
-3.29725295e-01 -2.33983826e-02 -7.02781081e-01 -1.04659367e+00
2.76273489e-01 -1.26245773e+00 -5.72803974e-01 -1.37149310e+00
5.74890256e-01 -4.56563771e-01 -5.41778803e-01 8.34709287e-01
-2.48516664e-01 3.86313573e-02 -2.29470983e-01 4.24445778e-01
-5.88248551e-01 6.43296242e-01 1.62298429e+00 -6.42081380e-01
-4.97696370e-01 1.16944283e-01 -6.68662429e-01 4.56483305e-01
7.77706981e-01 -4.97187138e-01 -7.44498074e-01 3.36540669e-01
2.18425021e-01 -1.29700050e-01 -1.48718670e-01 -5.64907849e-01
1.67189464e-01 -5.05576015e-01 5.60972512e-01 -5.79513669e-01
3.82098734e-01 -9.90229428e-01 3.00930198e-02 7.72584796e-01
-2.36694142e-01 -3.90772939e-01 6.32158145e-02 1.01782858e+00
-3.20068419e-01 -4.50336248e-01 7.53351808e-01 -1.84692189e-01
-7.41018176e-01 8.63303781e-01 2.46125869e-02 -2.59393334e-01
1.03124976e+00 -3.94415259e-01 -1.90808102e-01 9.54762697e-02
-5.82189560e-01 -2.32843328e-02 3.96465361e-01 2.04379827e-01
8.05270016e-01 -1.29149663e+00 -3.92360508e-01 2.28448659e-01
5.48004329e-01 1.65063348e-02 -7.83588365e-03 3.76846015e-01
-5.01126289e-01 4.91747618e-01 -1.17861897e-01 -3.17026168e-01
-1.48878133e+00 8.90980959e-01 2.01180905e-01 -3.07197481e-01
-3.06395411e-01 4.95582461e-01 8.49141330e-02 -4.53337133e-01
9.17185843e-02 1.02104619e-01 -4.74503219e-01 2.18735725e-01
4.47030962e-01 1.38112660e-02 2.75593758e-01 -3.81614655e-01
-4.40825820e-01 6.92652404e-01 -6.38181925e-01 8.23346853e-01
1.77787805e+00 6.19056284e-01 -7.39716217e-02 3.58582377e-01
1.42553234e+00 -8.00019577e-02 -9.96856987e-01 -4.51139331e-01
9.55406204e-02 -3.46512496e-01 8.38939250e-02 -7.02786922e-01
-8.36032033e-01 6.85069442e-01 6.42199159e-01 -2.74807990e-01
7.22816348e-01 -1.50335655e-01 4.95914638e-01 9.45412636e-01
4.29017782e-01 -1.08624041e+00 3.04734498e-01 3.86722356e-01
6.30986989e-01 -1.63739228e+00 7.13603973e-01 -9.30204034e-01
-5.20267248e-01 1.20719683e+00 5.30420780e-01 1.96148202e-01
8.25073004e-01 -2.72251427e-01 -2.24373206e-01 -5.53420484e-01
-5.82139313e-01 -9.28351134e-02 4.69930470e-01 6.41918421e-01
7.97765970e-01 -4.02401984e-02 -3.28889310e-01 4.15718816e-02
4.38307285e-01 -1.23628780e-01 1.37323275e-01 1.18491018e+00
-4.17969733e-01 -1.75709474e+00 1.32805824e-01 5.30004203e-01
-1.94429561e-01 -3.08786333e-01 -8.18415523e-01 6.82149053e-01
3.92966270e-02 8.03004503e-01 -5.33384264e-01 -4.26214129e-01
2.61179388e-01 -5.99473640e-02 6.51715457e-01 -1.06322849e+00
-2.37108245e-01 -1.31814510e-01 -1.82915539e-01 -3.40833485e-01
-7.13402092e-01 -1.85523584e-01 -1.62988710e+00 -1.20833002e-01
-4.65311378e-01 3.72221589e-01 4.93759394e-01 7.39099801e-01
5.62794685e-01 3.68266761e-01 1.11526811e+00 -3.74136358e-01
-4.08146203e-01 -7.41364419e-01 -7.51775920e-01 4.50911969e-01
2.89429307e-01 -6.91464543e-01 -1.99671730e-01 4.01756801e-02] | [5.131664752960205, 5.887186050415039] |
3edcd9f0-6f99-4567-82da-a04feddc93d6 | shape-interaction-matrix-revisited-and | 1509.02649 | null | http://arxiv.org/abs/1509.02649v2 | http://arxiv.org/pdf/1509.02649v2.pdf | Shape Interaction Matrix Revisited and Robustified: Efficient Subspace Clustering with Corrupted and Incomplete Data | The Shape Interaction Matrix (SIM) is one of the earliest approaches to
performing subspace clustering (i.e., separating points drawn from a union of
subspaces). In this paper, we revisit the SIM and reveal its connections to
several recent subspace clustering methods. Our analysis lets us derive a
simple, yet effective algorithm to robustify the SIM and make it applicable to
realistic scenarios where the data is corrupted by noise. We justify our method
by intuitive examples and the matrix perturbation theory. We then show how this
approach can be extended to handle missing data, thus yielding an efficient and
general subspace clustering algorithm. We demonstrate the benefits of our
approach over state-of-the-art subspace clustering methods on several
challenging motion segmentation and face clustering problems, where the data
includes corrupted and missing measurements. | ['Mathieu Salzmann', 'Pan Ji', 'Hongdong Li'] | 2015-09-09 | shape-interaction-matrix-revisited-and-1 | http://openaccess.thecvf.com/content_iccv_2015/html/Ji_Shape_Interaction_Matrix_ICCV_2015_paper.html | http://openaccess.thecvf.com/content_iccv_2015/papers/Ji_Shape_Interaction_Matrix_ICCV_2015_paper.pdf | iccv-2015-12 | ['face-clustering'] | ['computer-vision'] | [ 3.32566768e-01 -2.50257879e-01 4.19002026e-02 -5.57407886e-02
-8.46030295e-01 -9.12634134e-01 4.47076648e-01 -4.32375044e-01
-4.17916290e-02 2.96305299e-01 2.63863713e-01 -1.67282671e-01
-4.77769464e-01 -5.20889228e-03 -5.65138102e-01 -1.15673614e+00
-7.69979805e-02 5.77158153e-01 6.10076934e-02 6.25105053e-02
-9.00918338e-03 6.22885942e-01 -1.48174238e+00 -3.27590220e-02
7.71926522e-01 3.09267312e-01 -9.73509476e-02 5.19049168e-01
1.61905974e-01 1.33118644e-01 -2.87526041e-01 -2.72456149e-04
4.00575489e-01 -5.41046977e-01 -8.39944303e-01 6.81702912e-01
3.66940707e-01 -2.77461037e-02 -2.87793517e-01 9.77798104e-01
4.61704433e-01 4.45795089e-01 9.01310623e-01 -1.43296897e+00
-9.11685601e-02 4.22851264e-01 -7.24667072e-01 -6.01309240e-02
3.48762572e-01 -3.12020302e-01 5.31477690e-01 -9.56450462e-01
8.40385675e-01 1.38572931e+00 7.15934336e-01 6.76184058e-01
-1.84663963e+00 -2.85557777e-01 8.84132385e-02 1.63157806e-01
-1.41321254e+00 -8.37746263e-01 9.06749249e-01 -7.13966548e-01
4.21164364e-01 5.07994354e-01 2.87584305e-01 1.03101277e+00
-5.40833533e-01 1.07383883e+00 1.01653063e+00 -4.91970807e-01
4.76241201e-01 -2.44507551e-01 4.69946146e-01 1.67367518e-01
2.83952147e-01 -1.07299469e-01 -2.32696578e-01 -6.95095778e-01
5.90109229e-01 -1.47622585e-01 -5.18129885e-01 -1.30064559e+00
-1.32165074e+00 7.31313109e-01 -1.89177737e-01 2.92117238e-01
-5.17170951e-02 -2.84503084e-02 1.48999527e-01 1.78594086e-02
2.62895942e-01 -5.28703071e-02 -1.18465662e-01 1.53049966e-02
-1.35478091e+00 2.52932519e-01 7.66977608e-01 9.88914251e-01
4.73094761e-01 1.17500626e-01 1.67283341e-01 7.76970923e-01
2.97377527e-01 6.31422997e-01 9.96604115e-02 -1.50075567e+00
3.43344435e-02 1.05327591e-01 2.10280940e-01 -8.08906436e-01
-3.60726625e-01 -1.39099523e-01 -8.34265113e-01 1.67227201e-02
6.76549315e-01 -4.80818935e-02 -6.88278675e-01 1.85520828e+00
5.36422849e-01 6.26765490e-01 5.93441613e-02 8.34174275e-01
4.92803067e-01 3.40337276e-01 -4.68401521e-01 -8.22407007e-01
6.83923304e-01 -5.94249368e-01 -7.37517118e-01 3.33601832e-01
4.19760704e-01 -7.61238098e-01 5.89149296e-01 5.34813106e-01
-1.03336811e+00 -4.09203470e-01 -7.07999468e-01 2.26663247e-01
5.19009233e-02 5.23831882e-03 4.51353550e-01 8.25085580e-01
-1.28388309e+00 8.55330408e-01 -1.20935500e+00 -8.94177198e-01
1.60282776e-01 5.62237978e-01 -3.65741611e-01 -1.20440856e-01
-3.85768712e-01 3.04020226e-01 1.35772392e-01 7.53942952e-02
-7.00946689e-01 -3.56482744e-01 -6.76161408e-01 -3.06040883e-01
5.14937997e-01 -7.11468160e-01 1.10815525e+00 -7.62104392e-01
-1.23790193e+00 5.98280072e-01 -7.01619387e-01 -1.04304731e-01
4.00544018e-01 -1.85193717e-01 -3.43873173e-01 3.15955937e-01
4.54993872e-03 3.34647566e-01 1.10847223e+00 -1.68328154e+00
-3.96307968e-02 -5.38027644e-01 -6.54052138e-01 1.52651981e-01
-2.26414561e-01 1.57536909e-01 -7.67112315e-01 -6.24304056e-01
6.05156541e-01 -1.27463210e+00 -4.49600965e-01 -3.66552562e-01
-6.91976249e-01 9.32501704e-02 9.97091174e-01 -4.85128313e-01
1.30995953e+00 -2.37829876e+00 8.75209093e-01 6.43434167e-01
2.44180143e-01 -4.92442921e-02 -5.00279963e-02 6.42635822e-01
-4.48256105e-01 6.37590373e-03 -7.91175663e-01 -4.52258766e-01
1.70396008e-02 2.64697969e-01 -3.70111734e-01 9.09993589e-01
-2.71967649e-01 5.53715944e-01 -8.03684771e-01 -4.07850564e-01
1.92167833e-01 4.26758140e-01 -4.70783830e-01 -1.24331690e-01
2.64031440e-01 8.28660965e-01 -1.67511269e-01 5.80492973e-01
9.38548803e-01 -3.52478810e-02 4.32290465e-01 -1.83324978e-01
9.27924179e-03 -3.67087752e-01 -1.82214963e+00 1.60796022e+00
4.17357594e-01 5.43555319e-01 7.52296865e-01 -1.07157350e+00
3.81164551e-01 3.51833403e-01 1.16318858e+00 3.48171413e-01
7.27673322e-02 1.42373562e-01 -8.51454511e-02 -2.04342768e-01
4.86191288e-02 -4.49721999e-02 1.17710307e-02 5.96317708e-01
9.49108228e-02 4.90015820e-02 1.97894096e-01 4.57745105e-01
8.91162634e-01 1.91374242e-01 2.84770787e-01 -5.69687903e-01
6.77702367e-01 -5.66037856e-02 6.24801755e-01 7.14835227e-01
-4.22611684e-01 7.85306454e-01 2.57094651e-01 1.31756514e-01
-1.01364100e+00 -1.34334397e+00 -2.51881748e-01 6.59839869e-01
3.02968472e-02 -5.86914361e-01 -1.21564388e+00 -3.74435812e-01
2.40658205e-02 4.17575300e-01 -5.27275562e-01 1.24123842e-01
-5.57015479e-01 -9.60593641e-01 3.01911354e-01 5.68453014e-01
1.34489477e-01 -4.07980025e-01 -5.55323847e-02 -1.22164413e-02
-3.77485484e-01 -1.15266788e+00 -5.67253470e-01 -1.28952235e-01
-1.24682331e+00 -1.19437504e+00 -6.30799174e-01 -5.55461168e-01
7.27571845e-01 7.73376048e-01 8.93422484e-01 -2.40591001e-02
-5.27102910e-02 1.23026836e+00 -1.84565216e-01 8.09496548e-03
-4.02673662e-01 -3.67949903e-01 7.36489058e-01 2.83118218e-01
5.61263680e-01 -7.41144955e-01 -2.38272414e-01 4.70358104e-01
-1.07200968e+00 -1.71368599e-01 8.33610073e-02 8.17435384e-01
7.43187189e-01 1.71399295e-01 2.06717998e-02 -8.58368278e-01
4.51979697e-01 -4.64143753e-01 -4.77972656e-01 1.10456049e-01
-4.09918547e-01 -7.42394477e-02 4.08840239e-01 -4.78052467e-01
-8.42967927e-01 7.83478796e-01 3.37886542e-01 -7.52557337e-01
-4.46128041e-01 1.28337368e-01 -5.45551777e-01 -2.40479231e-01
4.53842342e-01 2.75819629e-01 2.06668824e-01 -7.75068581e-01
7.91681290e-01 5.53664863e-01 8.22414696e-01 -6.79713309e-01
1.16353691e+00 1.08426034e+00 2.80541241e-01 -1.20567656e+00
-2.63176918e-01 -1.05041087e+00 -1.41495216e+00 -2.16063380e-01
6.87210917e-01 -7.10199594e-01 -7.22195327e-01 3.97628963e-01
-8.05566072e-01 -1.26843274e-01 -2.54279584e-01 3.95779818e-01
-1.05870891e+00 1.04925144e+00 -5.79225838e-01 -1.10585594e+00
2.63655275e-01 -1.03631902e+00 1.19127464e+00 -1.95201725e-01
-2.40517378e-01 -1.05619049e+00 2.61687368e-01 4.01330024e-01
-1.19046964e-01 4.31974232e-01 6.55636072e-01 -6.58673704e-01
-3.25263351e-01 1.92507267e-01 2.74568945e-01 1.93372458e-01
6.08576424e-02 2.36040294e-01 -9.79435146e-01 -6.31406128e-01
3.62002492e-01 2.98709363e-01 9.11158562e-01 6.68562591e-01
9.52644765e-01 -1.00997187e-01 -7.03965783e-01 7.58543134e-01
1.29758012e+00 -4.51696627e-02 4.24501687e-01 -1.83837432e-02
9.45001066e-01 8.17046762e-01 3.28468412e-01 2.84401506e-01
-1.48903774e-02 8.90227020e-01 -1.77507102e-02 -2.59727743e-02
1.03432700e-01 1.56317130e-01 4.78079647e-01 1.10384357e+00
-1.21176273e-01 1.89196855e-01 -8.86613667e-01 6.35104656e-01
-2.34035659e+00 -1.24592388e+00 -5.06311178e-01 2.35915732e+00
3.99153620e-01 -5.14346123e-01 7.65486240e-01 4.35589910e-01
8.90345812e-01 -1.33485705e-01 -5.56400955e-01 1.03534512e-01
-3.88689518e-01 -6.40367866e-02 3.93413514e-01 5.35559297e-01
-1.50029540e+00 8.93627584e-01 7.64053392e+00 7.65656531e-01
-4.13040459e-01 -8.27849843e-03 1.76747814e-01 -1.56752303e-01
-1.86004594e-01 7.30187669e-02 -4.19503927e-01 2.40062356e-01
8.71418715e-01 -8.89451653e-02 8.20507348e-01 6.77612185e-01
3.76749039e-01 6.84379861e-02 -1.35806108e+00 1.29263163e+00
1.49371088e-01 -8.31269920e-01 -1.56033561e-01 2.48408183e-01
8.60197365e-01 -2.30472073e-01 7.34970942e-02 -2.21189603e-01
3.42636675e-01 -9.37683880e-01 3.14231008e-01 3.67187858e-01
5.81264019e-01 -6.38255179e-01 2.35169783e-01 4.39089507e-01
-1.16479766e+00 -1.32658988e-01 -2.50336736e-01 8.47765952e-02
2.54501611e-01 4.55624133e-01 -5.12231171e-01 7.74649620e-01
7.21943080e-01 9.17545080e-01 -4.66854036e-01 1.07626522e+00
3.12914759e-01 7.75894523e-01 -5.93522727e-01 4.85598832e-01
-1.13431185e-01 -8.73853207e-01 1.06779110e+00 1.28319991e+00
3.54807734e-01 3.44004720e-01 2.10938826e-01 7.49872863e-01
3.55045438e-01 5.83816133e-02 -8.67462158e-01 1.87255621e-01
4.40002918e-01 1.15680695e+00 -9.82151389e-01 -2.88255334e-01
-4.63402987e-01 1.08071816e+00 9.85581428e-03 8.01050961e-01
-4.35073674e-01 1.72485098e-01 8.29789639e-01 -5.48167005e-02
4.82448131e-01 -5.66048145e-01 -3.61814231e-01 -1.38571191e+00
-1.11819401e-01 -1.03393662e+00 4.54960287e-01 -4.43926334e-01
-1.35703993e+00 5.48832938e-02 4.40440387e-01 -1.49121225e+00
-4.07591552e-01 -6.63690209e-01 -4.57805604e-01 4.27974820e-01
-6.71228349e-01 -8.85945559e-01 -5.41039109e-02 9.38461959e-01
2.38922834e-01 -1.23142645e-01 7.29233563e-01 1.64958000e-01
-8.25472832e-01 2.52879739e-01 8.59622180e-01 3.18328701e-02
6.48879230e-01 -1.37742054e+00 2.53520697e-01 1.29330981e+00
4.28717464e-01 1.04397142e+00 9.49722767e-01 -5.58999121e-01
-1.71702552e+00 -7.87594914e-01 1.74331948e-01 -9.16066825e-01
7.21601784e-01 -6.04543149e-01 -9.65178490e-01 8.45231414e-01
-7.25772083e-02 -2.63443172e-01 1.09615481e+00 1.33802667e-01
-2.52041847e-01 1.78636119e-01 -1.06480730e+00 7.67218709e-01
1.31757367e+00 -3.31595629e-01 -5.60422599e-01 3.23343843e-01
2.36707270e-01 -1.27010226e-01 -6.83518529e-01 5.27256846e-01
4.09079134e-01 -1.10992455e+00 1.24129570e+00 -7.63528943e-01
-3.72246981e-01 -6.60245419e-01 -5.23380280e-01 -1.18626249e+00
-6.03757560e-01 -1.09205508e+00 -4.89209890e-01 1.17190146e+00
-5.48123531e-02 -3.02285463e-01 8.39571536e-01 6.76319838e-01
1.25218168e-01 -3.31120133e-01 -1.08074510e+00 -1.04124689e+00
3.38770479e-01 -5.46370089e-01 3.57091993e-01 1.16364360e+00
2.09181547e-01 1.82909682e-01 -4.15648133e-01 2.06401855e-01
1.20037866e+00 2.59372920e-01 9.47178066e-01 -1.48088896e+00
-4.33900267e-01 -4.38869923e-01 -3.55137646e-01 -1.13945889e+00
3.16358089e-01 -7.44092882e-01 -5.40172160e-02 -1.13593245e+00
5.40780842e-01 -3.37099694e-02 1.18484423e-02 3.65972109e-02
-1.65552512e-01 2.31501207e-01 4.58680719e-01 6.10761106e-01
-6.74055874e-01 4.02900189e-01 8.19040596e-01 -2.21584295e-03
-3.04026961e-01 1.28932923e-01 -6.58596873e-01 9.30874288e-01
5.72587192e-01 -2.29587436e-01 -5.47018111e-01 5.82598113e-02
-3.65805417e-01 -8.74572694e-02 3.66770267e-01 -1.00240052e+00
3.44278902e-01 -1.10481538e-01 4.10225540e-01 -6.55088544e-01
3.96468729e-01 -1.05399358e+00 5.29210627e-01 2.38034889e-01
1.41344890e-01 -2.57847756e-01 1.90881714e-01 9.05256450e-01
-4.33785282e-02 -7.94763584e-03 7.11004794e-01 1.16567947e-01
-5.98810792e-01 1.21253565e-01 -3.85373175e-01 -8.12912434e-02
1.09487820e+00 -5.08875012e-01 6.09032102e-02 -6.11442983e-01
-1.25026977e+00 1.57527611e-01 8.45261931e-01 1.17122911e-01
4.33407128e-01 -1.45759928e+00 -6.08208060e-01 1.83583498e-01
-5.09943925e-02 -2.78403193e-01 2.75569975e-01 1.15998006e+00
-1.67830750e-01 3.34308684e-01 1.59802645e-01 -1.12557077e+00
-1.68132102e+00 1.12737751e+00 1.80295438e-01 3.50782752e-01
-4.96829003e-01 4.31429654e-01 2.89943457e-01 -4.38595414e-01
2.81146407e-01 -2.50563323e-02 -1.16570517e-01 -1.39523387e-01
4.09689307e-01 9.18547630e-01 -2.68056333e-01 -1.22439671e+00
-5.74707925e-01 1.06679869e+00 4.31125611e-01 -4.76122946e-01
1.14909160e+00 -4.34823364e-01 -3.74005526e-01 7.41964579e-01
1.05906153e+00 2.68265873e-01 -1.21775973e+00 -8.29306990e-02
1.67861789e-01 -3.35661322e-01 -3.68894339e-01 -1.00992948e-01
-8.10742676e-01 7.74484754e-01 5.17084897e-01 2.66692549e-01
1.42576325e+00 1.48632124e-01 4.85049039e-01 3.38432401e-01
3.78888339e-01 -1.13308656e+00 -3.40661556e-01 1.18443981e-01
6.07313037e-01 -1.13237023e+00 1.20092869e-01 -9.53501463e-01
-4.63019907e-01 1.01785207e+00 1.66440681e-01 -7.06365332e-02
8.76225829e-01 2.67469436e-01 4.50648414e-03 -7.98055455e-02
-3.56565773e-01 -3.35260779e-01 4.27685380e-01 9.04092133e-01
3.12531501e-01 1.06487490e-01 -2.26303741e-01 5.00972986e-01
-1.34149902e-02 -3.83918911e-01 5.90814173e-01 7.62490690e-01
-3.21891040e-01 -1.28271735e+00 -9.96502638e-01 2.07092896e-01
-3.26799631e-01 1.77697584e-01 -8.62907529e-01 6.74951196e-01
-8.89655724e-02 1.25222790e+00 -2.06909582e-01 -2.72885889e-01
1.23742051e-01 3.90551537e-01 5.29148638e-01 -4.33226019e-01
-3.31838354e-02 8.75810146e-01 -1.96159944e-01 -8.47733974e-01
-8.77757072e-01 -1.41483927e+00 -1.07082212e+00 -4.01122957e-01
-2.16164991e-01 2.98918247e-01 2.57532775e-01 7.66124487e-01
4.04816896e-01 1.00686967e-01 6.08835280e-01 -9.68612015e-01
-5.21645844e-01 -6.63588703e-01 -7.98033893e-01 7.00075328e-01
3.82717609e-01 -7.32526064e-01 -5.08700073e-01 4.60264891e-01] | [7.70885705947876, 4.428779125213623] |
393acf08-c687-4eec-8d96-57a2f0979ecd | analysing-affective-behavior-in-the-second | 2106.15318 | null | https://arxiv.org/abs/2106.15318v2 | https://arxiv.org/pdf/2106.15318v2.pdf | Analysing Affective Behavior in the second ABAW2 Competition | The Affective Behavior Analysis in-the-wild (ABAW2) 2021 Competition is the second -- following the first very successful ABAW Competition held in conjunction with IEEE FG 2020- Competition that aims at automatically analyzing affect. ABAW2 is split into three Challenges, each one addressing one of the three main behavior tasks of valence-arousal estimation, basic expression classification and action unit detection. All three Challenges are based on a common benchmark database, Aff-Wild2, which is a large scale in-the-wild database and the first one to be annotated for all these three tasks. In this paper, we describe this Competition, to be held in conjunction with ICCV 2021. We present the three Challenges, with the utilized Competition corpora. We outline the evaluation metrics and present the baseline system with its results. More information regarding the Competition is provided in the Competition site: https://ibug.doc.ic.ac.uk/resources/iccv-2021-2nd-abaw. | ['Stefanos Zafeiriou', 'Elnar Hajiyev', 'Irene Kotsia', 'Dimitrios Kollias'] | 2021-06-14 | null | null | null | null | ['action-unit-detection'] | ['computer-vision'] | [-3.57963261e-03 5.32251894e-02 -8.28009769e-02 -8.25196207e-01
-1.00079024e+00 -5.53563058e-01 4.63559896e-01 1.87563151e-01
-5.20851076e-01 7.51944423e-01 3.99062246e-01 3.80967498e-01
1.94255412e-01 3.52720777e-03 4.39141802e-02 -4.35365230e-01
-4.14082617e-01 2.11369216e-01 -2.46198937e-01 -5.98733366e-01
-3.82150449e-02 1.31185120e-02 -1.59513557e+00 8.45532477e-01
1.39096230e-01 1.47584152e+00 -4.91033733e-01 6.94559216e-01
2.45014250e-01 8.82655799e-01 -5.67735910e-01 -7.45073080e-01
-1.18662193e-01 -6.02027059e-01 -1.05478442e+00 -4.03350979e-01
4.67606485e-02 2.00431123e-01 2.47007385e-01 8.66402864e-01
7.45278001e-01 2.68250734e-01 5.11330783e-01 -1.74040723e+00
-3.69560719e-02 3.59159410e-01 -3.65073502e-01 1.89170569e-01
7.34930038e-01 -8.61173347e-02 1.14582574e+00 -9.34865117e-01
6.62518144e-01 1.18928564e+00 7.62567639e-01 9.93499756e-01
-9.43005145e-01 -5.36153018e-01 1.84138983e-01 3.93935353e-01
-1.27121878e+00 -5.57651341e-01 9.26913559e-01 -3.95218670e-01
1.31815207e+00 6.13695145e-01 8.29540491e-01 1.61474621e+00
-1.84447408e-01 1.17465997e+00 1.52306795e+00 -2.38247782e-01
3.59764814e-01 1.17502481e-01 6.60385728e-01 2.88783222e-01
-5.06919086e-01 -2.93370932e-01 -8.55124831e-01 -1.91723436e-01
-2.17671305e-01 -8.59471858e-01 -2.54556209e-01 -2.34215483e-02
-7.67067194e-01 7.71788955e-01 8.53609517e-02 4.49693441e-01
-5.31484902e-01 1.22190773e-01 1.16465223e+00 3.66464704e-01
7.73899257e-01 5.13859332e-01 -5.40788114e-01 -9.93695855e-01
-6.79628432e-01 4.70111966e-01 8.17950368e-01 5.80052853e-01
2.90565431e-01 -8.55294392e-02 -4.42408890e-01 1.24989498e+00
2.49573812e-01 1.15054078e-01 6.70213044e-01 -1.03769064e+00
1.55528590e-01 4.69936341e-01 1.46724492e-01 -5.47793686e-01
-8.20058048e-01 1.36689395e-01 -6.77549601e-01 2.60473583e-02
-8.76518413e-02 -4.51204777e-01 -5.15180469e-01 2.08967495e+00
1.93953350e-01 -2.23343372e-01 1.47937359e-02 6.88904405e-01
1.11619723e+00 6.35981381e-01 3.65727365e-01 -6.79688871e-01
1.44420719e+00 -9.90971982e-01 -1.27088690e+00 -2.34683976e-01
7.00021684e-01 -6.16040409e-01 9.44333732e-01 6.98407471e-01
-1.37905192e+00 -4.10206050e-01 -1.29703629e+00 -6.17542192e-02
-7.63028204e-01 -2.20962882e-01 8.22786629e-01 8.60110223e-01
-1.29919994e+00 2.01917872e-01 -5.66135883e-01 -6.00167096e-01
2.70775944e-01 3.00170124e-01 -3.10800284e-01 6.01469159e-01
-1.68964994e+00 1.18204820e+00 7.72461221e-02 4.37173508e-02
-5.35982609e-01 -5.16614258e-01 -8.26756656e-01 -4.41432178e-01
1.11041479e-02 6.93251789e-02 1.73639584e+00 -1.32244492e+00
-1.59543693e+00 1.52020192e+00 -5.37358075e-02 -3.68502319e-01
3.56044203e-01 -2.43892267e-01 -9.79944050e-01 -1.59078345e-01
-1.27313063e-01 8.30201209e-01 1.73894525e-01 -9.03156757e-01
-3.26255411e-01 -5.51020265e-01 -1.40264377e-01 1.59009680e-01
-2.08806112e-01 4.49777037e-01 -7.15499818e-01 -3.73688579e-01
-7.66204298e-01 -1.04303813e+00 5.76058254e-02 -5.64395249e-01
-2.85063565e-01 -5.97979844e-01 5.29886127e-01 -2.91123122e-01
1.59810805e+00 -2.31534290e+00 6.06863573e-02 1.32402629e-01
7.30849057e-02 3.89624536e-01 -2.28610277e-01 6.82640135e-01
-5.31073809e-01 -6.70652464e-02 -8.84920731e-02 -7.02474356e-01
4.60793704e-01 1.30192470e-02 -5.59337530e-03 5.17736137e-01
1.46756351e-01 1.02732396e+00 -7.15417206e-01 -8.07965398e-02
3.22126150e-02 3.84021312e-01 -5.67134432e-02 4.16587204e-01
-2.40532104e-02 1.23306304e-01 -2.42721081e-01 4.73322660e-01
5.73472977e-01 4.68788952e-01 1.93904310e-01 -3.08151275e-01
-1.48615122e-01 1.62180871e-01 -8.81818414e-01 1.84252226e+00
-4.49759245e-01 6.47456467e-01 3.58293384e-01 -1.00537515e+00
1.06488669e+00 6.02984905e-01 8.22499275e-01 -1.06064582e+00
4.16664660e-01 3.05731297e-01 -1.66821495e-01 -6.19915485e-01
3.78330052e-01 -1.05018631e-01 -6.88430429e-01 1.30822256e-01
5.59014499e-01 -1.94079638e-01 6.71142280e-01 1.66686758e-01
1.07185245e+00 2.15461954e-01 7.84768999e-01 -5.22479177e-01
7.39586651e-01 -5.25803804e-01 7.68191814e-01 2.21973330e-01
-8.91022146e-01 3.78959030e-01 8.99897993e-01 -5.79208791e-01
-4.80665207e-01 -7.90253699e-01 -1.04517646e-01 1.10553050e+00
-9.08633769e-02 -1.06887817e+00 -8.43430638e-01 -5.37315965e-01
-3.85054588e-01 9.19666350e-01 -1.31370401e+00 -2.62723505e-01
7.35968575e-02 -8.65577936e-01 7.65264630e-01 5.10071337e-01
5.93384504e-01 -1.48739803e+00 -9.25524235e-01 5.15779946e-03
-5.55722892e-01 -1.02717030e+00 -1.68152586e-01 7.01712012e-01
-1.24318376e-01 -1.04616678e+00 -2.65560806e-01 -3.76880527e-01
-5.83429113e-02 -7.12945104e-01 1.53906655e+00 -1.84554517e-01
-4.83187854e-01 7.68136203e-01 -6.79754376e-01 -9.37876582e-01
-3.06288660e-01 -9.74889323e-02 -4.16070484e-02 2.17314243e-01
1.02223122e+00 -4.41389740e-01 -3.96037787e-01 1.41810343e-01
-5.48054218e-01 -1.17483005e-01 1.47476882e-01 4.84660894e-01
4.96688843e-01 -7.20495284e-01 6.91672921e-01 -9.64738667e-01
8.82171094e-01 -3.81075054e-01 -1.72709703e-01 1.11942738e-01
-3.66371453e-01 -4.88488019e-01 1.71764836e-01 5.06851599e-02
-9.66882467e-01 1.19729020e-01 -7.84623444e-01 -5.54665662e-02
-3.64085078e-01 2.76779503e-01 -4.22816873e-01 3.81667763e-01
6.06785178e-01 -2.15153441e-01 -4.54121679e-01 -1.38333783e-01
1.51734307e-01 8.92365098e-01 6.33715153e-01 -4.66098875e-01
-7.49343857e-02 7.29361176e-02 -1.49081230e-01 -6.71581566e-01
-1.16635120e+00 -8.44258010e-01 -6.27746820e-01 -7.90638089e-01
1.26473105e+00 -7.97864318e-01 -8.46638680e-01 8.22726369e-01
-9.99485731e-01 -5.96767604e-01 -3.83469254e-01 1.47601128e-01
-7.40939558e-01 -2.65463829e-01 -6.99136376e-01 -1.23692811e+00
-6.65185869e-01 -9.31910157e-01 1.15886033e+00 1.14698857e-01
-1.01599264e+00 -8.14570665e-01 7.47230351e-01 4.32411969e-01
4.58811849e-01 6.27286315e-01 4.47036922e-01 -8.78719211e-01
1.03853083e+00 -2.48490006e-01 2.97633767e-01 6.38229012e-01
-2.10039854e-01 -7.97697380e-02 -1.52364397e+00 -2.66791750e-02
-2.93136686e-02 -1.29481363e+00 8.14532399e-01 6.84306547e-02
1.14650834e+00 3.46691430e-01 1.63552135e-01 3.05233210e-01
1.09167469e+00 5.03418684e-01 8.11671197e-01 1.52452111e-01
1.77830663e-02 6.28121555e-01 9.23883379e-01 7.16557026e-01
2.33896077e-01 7.67353237e-01 4.75763738e-01 -2.01391041e-01
1.94800496e-01 2.63312072e-01 5.73670328e-01 9.52217937e-01
-3.10333639e-01 -2.30169445e-01 -7.90916026e-01 5.53673387e-01
-1.90933359e+00 -1.07614183e+00 -4.22132790e-01 1.97208655e+00
9.11668837e-01 -7.00404644e-02 6.20853841e-01 1.13745913e-01
2.52017379e-01 3.68878305e-01 -2.00381562e-01 -1.31883192e+00
-6.48162067e-02 4.89010125e-01 -4.10559177e-01 6.52898908e-01
-1.31485963e+00 4.78438675e-01 5.46803904e+00 8.30733478e-01
-9.73758817e-01 3.67188364e-01 1.14454317e+00 -6.01309001e-01
1.42158747e-01 -6.05629683e-01 -5.27044177e-01 4.95788276e-01
1.42517960e+00 -1.47600755e-01 1.56199545e-01 7.61673748e-01
3.64478767e-01 -2.92689502e-01 -1.22513068e+00 1.32592332e+00
2.97225028e-01 -6.00004911e-01 -5.05266786e-01 -2.77856916e-01
4.35177594e-01 1.96062222e-01 -4.20309193e-02 9.62917864e-01
-9.69481096e-02 -9.98292327e-01 7.16014981e-01 2.82673150e-01
8.62449288e-01 -9.86552715e-01 9.38524306e-01 -3.10548931e-01
-1.05562258e+00 2.97401585e-02 4.67530936e-02 -1.72532365e-01
6.91481531e-02 6.02538824e-01 2.50609875e-01 2.82469869e-01
1.18047857e+00 7.35068679e-01 -5.14233947e-01 4.60251868e-01
-2.42696330e-01 7.15602458e-01 -4.21786048e-02 -4.32937831e-01
3.42166603e-01 -1.77493498e-01 2.03488722e-01 1.79470980e+00
-5.33521958e-02 2.85000086e-01 5.11158481e-02 3.63333702e-01
-1.98157102e-01 3.50567251e-01 -3.37573171e-01 2.42640451e-02
2.22733513e-01 1.96293557e+00 -3.07840437e-01 -3.08585584e-01
-4.25606400e-01 1.22813308e+00 4.25104886e-01 -2.13468634e-02
-1.27459705e+00 -4.00878042e-01 9.28589463e-01 -3.24093699e-01
-1.44657165e-01 4.45156276e-01 2.41667330e-01 -8.32349002e-01
-1.17166691e-01 -9.08907413e-01 5.86141109e-01 -1.01683915e+00
-1.39381766e+00 1.01944625e+00 1.21551946e-01 -8.88646424e-01
-2.69463301e-01 -7.73216367e-01 -6.62477851e-01 4.74662960e-01
-1.06024611e+00 -1.28031170e+00 -5.19284129e-01 7.86663234e-01
6.45527244e-01 1.82894394e-01 1.36353862e+00 4.30067152e-01
-8.91268432e-01 6.29649401e-01 -2.68081188e-01 -5.11188768e-02
1.06603158e+00 -1.50050342e+00 3.77935022e-02 3.75608087e-01
-1.09016344e-01 6.20089807e-02 6.96485281e-01 -2.04205085e-02
-9.82126057e-01 -9.80654240e-01 9.90575016e-01 -4.52520937e-01
8.33853245e-01 -1.06128442e+00 -5.58411896e-01 6.44982100e-01
8.18476200e-01 -4.25245687e-02 1.21654713e+00 3.72187614e-01
-2.85428047e-01 -1.24119885e-01 -1.03497255e+00 3.59571368e-01
7.91800439e-01 -2.40500569e-01 -3.35878998e-01 3.33571225e-01
2.13997141e-01 -3.55700612e-01 -9.97409940e-01 5.40344954e-01
6.43523455e-01 -1.32986879e+00 4.73887205e-01 -6.28871918e-01
5.54275274e-01 1.15987211e-01 -4.57956940e-01 -1.57143438e+00
-1.57090396e-01 -8.56563866e-01 -5.87672852e-02 1.67376232e+00
5.14760375e-01 -3.34229618e-01 4.52621758e-01 5.97427309e-01
-1.52503610e-01 -1.10986924e+00 -1.04646790e+00 -5.34760237e-01
1.83505043e-02 -7.90476143e-01 3.94643135e-02 1.02242172e+00
6.92393720e-01 8.46209526e-01 -3.43822062e-01 -6.45215988e-01
1.37170210e-01 -4.88423556e-02 4.83487129e-01 -9.20748651e-01
-3.20716321e-01 -5.89444458e-01 -6.63231909e-01 -2.49666423e-01
1.36816248e-01 -6.35263264e-01 1.97573259e-01 -1.09880781e+00
3.70708913e-01 6.28995419e-01 -6.29249215e-01 8.96631062e-01
-1.18445523e-01 7.13411570e-01 1.02546982e-01 -3.91557455e-01
-1.10667527e+00 6.94127262e-01 4.75962996e-01 -6.95723519e-02
1.02954663e-01 -1.78580612e-01 -6.35207117e-01 7.58437335e-01
1.22616959e+00 6.10040054e-02 -6.23690605e-01 2.72761375e-01
2.65247971e-01 -4.45218891e-01 1.88119877e-02 -1.01157463e+00
-3.12089652e-01 7.42470250e-02 2.78182328e-01 -5.28272271e-01
8.64874423e-01 -4.42484796e-01 -1.48313582e-01 7.37972334e-02
-6.53160632e-01 1.65461406e-01 6.87100351e-01 -6.52246699e-02
-4.15748715e-01 -2.45264918e-01 1.08979952e+00 -2.76093036e-02
-1.06758213e+00 -4.55576591e-02 -7.75041282e-01 3.69545013e-01
1.44602895e+00 3.50391746e-01 -2.55770802e-01 -5.89845955e-01
-9.49029684e-01 5.26803195e-01 -1.03238888e-01 5.74220002e-01
2.28314489e-01 -1.43263280e+00 -8.36254001e-01 -1.67247638e-01
6.90226257e-01 -8.32160056e-01 5.50639391e-01 1.23905826e+00
-4.24639918e-02 3.59076500e-01 -4.50502306e-01 -1.91074014e-01
-1.75537634e+00 5.23161769e-01 5.51849425e-01 -6.77975953e-01
1.46099299e-01 1.04152155e+00 4.24476564e-02 -3.90982300e-01
1.33161142e-01 8.28977302e-02 -3.79477143e-01 3.45425397e-01
6.36105359e-01 2.99187869e-01 4.22450095e-01 -6.26107931e-01
-7.67019093e-01 6.53135255e-02 1.58642605e-01 -3.18909854e-01
1.44869971e+00 -1.16621062e-01 -3.84337217e-01 1.18971395e+00
1.52225041e+00 -2.53500968e-01 -6.65805042e-01 3.55697960e-01
2.13702455e-01 1.11069620e-01 6.44180328e-02 -1.28834844e+00
-9.66627657e-01 8.80714536e-01 7.86182284e-01 2.64265478e-01
1.44463933e+00 -7.84060657e-02 4.62318301e-01 1.27116412e-01
4.44789678e-02 -1.69186330e+00 4.72496590e-03 4.08696622e-01
1.27467859e+00 -9.61639702e-01 -1.06528834e-01 -2.64010280e-01
-1.13694620e+00 9.18255210e-01 1.03445756e+00 7.91539997e-02
6.21543050e-01 4.72510219e-01 3.84648144e-01 -6.00189447e-01
-1.23868990e+00 -2.60084569e-01 2.28142589e-01 5.32269359e-01
1.16014183e+00 1.02110729e-01 -5.50911129e-01 1.10995436e+00
-1.10413276e-01 -5.32995537e-02 3.65306556e-01 8.21030617e-01
2.83476319e-02 -1.11587811e+00 4.79938015e-02 4.87136573e-01
-5.49132288e-01 1.82433948e-01 -1.25605440e+00 8.78515899e-01
2.73311198e-01 1.13137615e+00 3.19500640e-02 -6.19204998e-01
7.97999322e-01 6.13661945e-01 4.74320173e-01 -3.40997577e-01
-1.01577175e+00 8.77596289e-02 7.81090140e-01 -1.29816175e+00
-7.93253541e-01 -9.54353809e-01 -1.14244258e+00 -1.98963016e-01
2.71571040e-01 5.45662403e-01 6.49133623e-01 5.94267547e-01
1.31106913e-01 5.77823937e-01 5.54978669e-01 -8.09716046e-01
-8.04451406e-02 -1.11527944e+00 -9.41263855e-01 8.43688607e-01
-1.11755401e-01 -5.98208606e-01 -2.23152786e-01 -3.93920988e-02] | [13.57007884979248, 2.2381234169006348] |
afcc236e-b1a9-4c69-b67c-f10e9fe53046 | diffsinger-diffusion-acoustic-model-for | 2105.02446 | null | https://arxiv.org/abs/2105.02446v6 | https://arxiv.org/pdf/2105.02446v6.pdf | DiffSinger: Singing Voice Synthesis via Shallow Diffusion Mechanism | Singing voice synthesis (SVS) systems are built to synthesize high-quality and expressive singing voice, in which the acoustic model generates the acoustic features (e.g., mel-spectrogram) given a music score. Previous singing acoustic models adopt a simple loss (e.g., L1 and L2) or generative adversarial network (GAN) to reconstruct the acoustic features, while they suffer from over-smoothing and unstable training issues respectively, which hinder the naturalness of synthesized singing. In this work, we propose DiffSinger, an acoustic model for SVS based on the diffusion probabilistic model. DiffSinger is a parameterized Markov chain that iteratively converts the noise into mel-spectrogram conditioned on the music score. By implicitly optimizing variational bound, DiffSinger can be stably trained and generate realistic outputs. To further improve the voice quality and speed up inference, we introduce a shallow diffusion mechanism to make better use of the prior knowledge learned by the simple loss. Specifically, DiffSinger starts generation at a shallow step smaller than the total number of diffusion steps, according to the intersection of the diffusion trajectories of the ground-truth mel-spectrogram and the one predicted by a simple mel-spectrogram decoder. Besides, we propose boundary prediction methods to locate the intersection and determine the shallow step adaptively. The evaluations conducted on a Chinese singing dataset demonstrate that DiffSinger outperforms state-of-the-art SVS work. Extensional experiments also prove the generalization of our methods on text-to-speech task (DiffSpeech). Audio samples: https://diffsinger.github.io. Codes: https://github.com/MoonInTheRiver/DiffSinger. The old title of this work: "Diffsinger: Diffusion acoustic model for singing voice synthesis". | ['Zhou Zhao', 'Feiyang Chen', 'Yi Ren', 'Chengxi Li', 'Jinglin Liu'] | 2021-05-06 | null | null | null | null | ['singing-voice-synthesis'] | ['speech'] | [-9.94959027e-02 6.57729730e-02 7.13031786e-03 1.86624229e-01
-1.22407091e+00 -5.90697169e-01 2.28502989e-01 -7.53268838e-01
1.73075527e-01 6.07766688e-01 3.63707781e-01 -2.20647842e-01
2.21613213e-01 -6.94138348e-01 -8.01672757e-01 -1.01149678e+00
3.24603468e-01 2.60474563e-01 7.17888996e-02 -1.85379028e-01
-2.09687322e-01 -1.85521729e-02 -1.37937582e+00 2.09637925e-01
9.01502192e-01 7.66216457e-01 5.74194372e-01 1.08191860e+00
-2.44340114e-02 4.57220525e-01 -8.09193313e-01 -6.65414631e-02
2.91605175e-01 -1.18772209e+00 -2.78074920e-01 -3.92351896e-01
3.66676897e-02 -3.82227749e-01 -4.51135397e-01 1.09466934e+00
7.83422470e-01 2.95001030e-01 6.62236512e-01 -1.06843197e+00
-7.31597066e-01 1.04268694e+00 -4.45246585e-02 -1.20628193e-01
1.87848449e-01 6.59579635e-01 1.11665392e+00 -9.48398173e-01
3.93626362e-01 1.12725413e+00 7.00869739e-01 1.00155234e+00
-1.03569853e+00 -1.09172106e+00 -3.07937980e-01 6.27530664e-02
-1.41238928e+00 -7.05304205e-01 9.76192653e-01 -4.06876147e-01
4.41341341e-01 5.04526734e-01 6.44399703e-01 1.26945090e+00
-1.56292468e-01 8.38319004e-01 8.18241179e-01 -3.87824029e-01
2.36945823e-01 -1.89919099e-01 -4.85870510e-01 5.77755034e-01
-5.73024690e-01 4.01121318e-01 -7.41367936e-01 -1.10934578e-01
8.33343565e-01 -4.88865584e-01 -4.89970177e-01 3.16311300e-01
-1.04894853e+00 6.63652956e-01 9.93123744e-03 3.96717012e-01
-3.36453676e-01 2.43485570e-01 2.27098733e-01 3.46448600e-01
3.19574237e-01 3.13808888e-01 -1.15092739e-01 -4.19894874e-01
-1.31586456e+00 4.19399530e-01 8.41240644e-01 7.85812974e-01
4.26517844e-01 7.68904805e-01 -5.49880743e-01 1.02357543e+00
4.35395569e-01 8.47293258e-01 6.64540529e-01 -1.09984291e+00
4.88271356e-01 -3.06985021e-01 -6.03172444e-02 -4.50455964e-01
2.10380226e-01 -4.59286839e-01 -8.79656553e-01 8.69394243e-02
3.33549529e-01 -5.74173629e-01 -7.88642764e-01 1.91535151e+00
2.86552221e-01 8.01153898e-01 4.00866307e-02 1.05513012e+00
7.16380119e-01 1.23833120e+00 -3.18885267e-01 -4.66386914e-01
7.43824840e-01 -1.22175550e+00 -9.57229316e-01 1.94930926e-01
1.62117094e-01 -1.01921749e+00 1.53887343e+00 5.05448043e-01
-1.30956423e+00 -8.41732860e-01 -1.05119765e+00 9.57399458e-02
2.06102431e-01 2.36644730e-01 -5.00176474e-02 7.30239689e-01
-9.32689965e-01 9.31112111e-01 -8.35145473e-01 4.57943410e-01
4.68706377e-02 6.23982511e-02 2.01916009e-01 3.89233261e-01
-1.29085231e+00 2.31824368e-01 1.10446617e-01 1.30916506e-01
-1.44014990e+00 -1.01094234e+00 -5.45698643e-01 -5.03928252e-02
1.93544611e-01 -7.76911020e-01 1.59586012e+00 -9.09013152e-01
-2.35864782e+00 1.90581128e-01 -2.27556854e-01 -4.66936737e-01
5.63887358e-01 -3.44083726e-01 -4.05496657e-01 -4.35685441e-02
-8.24148357e-02 3.08950245e-01 1.26591921e+00 -1.16698730e+00
-2.88282216e-01 1.70568615e-01 -4.38130289e-01 3.10775727e-01
-3.58635485e-01 -1.01707943e-01 -3.80389452e-01 -1.05339491e+00
-1.82113260e-01 -8.84367108e-01 -8.24073255e-02 -5.09597898e-01
-7.78465867e-01 -8.91362056e-02 7.64809489e-01 -9.75038767e-01
1.70826864e+00 -2.24243498e+00 3.83643895e-01 2.32275110e-02
-2.22970888e-01 5.70517302e-01 -1.42771855e-01 5.35148382e-01
1.37312457e-01 2.19478920e-01 -5.16254663e-01 -6.78040206e-01
2.17684761e-01 7.76575729e-02 -7.16354012e-01 1.32038906e-01
-7.58381337e-02 7.71202683e-01 -8.89385939e-01 -3.88534307e-01
9.73381326e-02 5.46451867e-01 -8.40723872e-01 7.01961875e-01
-3.88138980e-01 8.27340066e-01 -2.07068473e-01 4.49999690e-01
3.54363769e-01 4.31756496e-01 -1.18051834e-01 -7.47008398e-02
-1.91307470e-01 6.00234687e-01 -1.25590444e+00 1.80764616e+00
-7.93265045e-01 2.92666525e-01 3.12548846e-01 -5.73048294e-01
1.03857386e+00 6.12401485e-01 1.59160584e-01 -1.47646025e-01
-5.09979799e-02 4.45898533e-01 1.41616762e-02 -3.82024437e-01
3.76361549e-01 -6.16185129e-01 1.52570605e-01 2.68765807e-01
1.43178195e-01 -7.09103942e-01 -1.99574053e-01 -1.63685098e-01
7.36923397e-01 3.87390822e-01 -6.97628409e-02 1.41577736e-01
5.70807934e-01 -4.26506013e-01 7.18938768e-01 5.76699018e-01
1.08257227e-01 1.01938677e+00 2.56391436e-01 3.36269259e-01
-1.10018587e+00 -1.27881598e+00 1.44653916e-01 8.53588104e-01
-2.40820169e-01 -5.17762899e-01 -1.32595587e+00 -3.76690954e-01
-2.90542960e-01 1.06735241e+00 -1.35141283e-01 -2.30210960e-01
-7.18929231e-01 -3.51187199e-01 1.08910048e+00 2.65010566e-01
2.79498369e-01 -1.40665472e+00 1.60808519e-01 4.30055231e-01
-2.48580933e-01 -7.30353296e-01 -1.21199107e+00 -2.55181938e-01
-6.55269504e-01 -4.81390387e-01 -9.41731691e-01 -7.60279775e-01
5.76428548e-02 -3.14147681e-01 7.37539947e-01 -1.22338064e-01
1.45957932e-01 1.14373110e-01 -2.32460126e-01 -1.91184595e-01
-1.09257364e+00 3.53178792e-02 3.73251647e-01 1.99340820e-01
-2.77827799e-01 -9.09327209e-01 -6.85852349e-01 2.52578467e-01
-8.69940341e-01 1.18336402e-01 2.80824751e-01 1.09619200e+00
9.04805422e-01 -1.27471900e-02 8.58562708e-01 -6.37668312e-01
8.88684273e-01 -4.62997079e-01 -4.79052871e-01 -1.01270728e-01
-5.05094349e-01 -7.37636909e-02 1.19483280e+00 -7.83782244e-01
-1.01342690e+00 -1.45144805e-01 -8.42900395e-01 -9.51957107e-01
-2.49874704e-02 2.46425033e-01 -3.31785649e-01 4.79576170e-01
5.00133872e-01 5.60575485e-01 3.12266983e-02 -7.26413310e-01
4.92221534e-01 9.62127864e-01 6.77277565e-01 -6.34119749e-01
1.06316721e+00 5.59729859e-02 -3.94772649e-01 -9.72371519e-01
-5.92434585e-01 -9.68496278e-02 -1.72093034e-01 -1.53660089e-01
7.14455128e-01 -7.76396394e-01 -5.97093642e-01 7.35151052e-01
-1.05670214e+00 -8.63215864e-01 -7.07955301e-01 6.58140898e-01
-8.56180668e-01 4.39551920e-01 -9.62229788e-01 -1.10760581e+00
-6.51352167e-01 -1.07662356e+00 8.40203226e-01 1.83861852e-01
-8.40695351e-02 -7.39286542e-01 3.77295345e-01 4.59727049e-01
3.26954782e-01 -2.86775045e-02 6.98324621e-01 -3.05498481e-01
-5.24713635e-01 2.87397087e-01 5.11730671e-01 1.02353299e+00
9.35839564e-02 2.04151973e-01 -1.14901268e+00 -1.38855055e-01
9.08039883e-02 -1.49644762e-01 8.02782536e-01 5.82428813e-01
1.10670102e+00 -5.93015730e-01 1.24191545e-01 7.96213686e-01
9.25207615e-01 3.71633887e-01 6.65455759e-01 -3.24332088e-01
7.63518572e-01 2.19647497e-01 6.98426604e-01 4.39359844e-01
4.15014215e-02 6.66796267e-01 1.05943426e-01 1.11061759e-01
-7.92927384e-01 -9.97514307e-01 8.88952255e-01 1.82713091e+00
-1.09839730e-01 -2.32658148e-01 -4.45143044e-01 5.84269643e-01
-1.46551430e+00 -9.92218971e-01 4.98061478e-02 2.37277842e+00
1.25549579e+00 -1.60387177e-02 1.85700223e-01 2.97191262e-01
6.74536049e-01 4.17725533e-01 -6.23437822e-01 -3.89716148e-01
5.35595566e-02 5.58172464e-01 1.51686341e-01 8.75954151e-01
-5.79349041e-01 1.15879798e+00 5.18579102e+00 1.51403952e+00
-1.17458212e+00 3.50224793e-01 3.57862562e-01 -1.89983323e-01
-7.25171745e-01 -4.65829000e-02 -9.08881426e-01 7.15398610e-01
1.09760678e+00 -1.37256652e-01 1.03858054e+00 5.50222576e-01
6.70855463e-01 5.94252229e-01 -8.11937451e-01 8.01819980e-01
-6.38036281e-02 -1.31385696e+00 5.06364182e-02 -1.83669642e-01
7.45117486e-01 -1.70020476e-01 4.03653443e-01 4.53023851e-01
1.11297958e-01 -1.08662832e+00 1.09011364e+00 6.45408869e-01
1.25256121e+00 -6.78753912e-01 3.53901148e-01 6.74302936e-01
-1.28788853e+00 5.66370711e-02 -7.62829557e-02 1.53021917e-01
4.60810483e-01 4.07421917e-01 -1.07542419e+00 5.57999313e-01
3.71730983e-01 3.52079272e-01 2.54968762e-01 8.07027280e-01
-4.98819649e-01 1.59248257e+00 -3.10858995e-01 -7.60039091e-02
7.87124410e-02 -4.37365443e-01 1.01187921e+00 1.16759121e+00
7.80219018e-01 4.80920859e-02 -1.76416621e-01 1.33837521e+00
-1.68536052e-01 2.55261511e-01 -4.27718669e-01 -3.17137957e-01
7.19514132e-01 8.26441884e-01 9.97179151e-02 -1.45218000e-01
-8.57786834e-03 1.02803218e+00 -1.12844303e-01 5.23728728e-01
-9.34455812e-01 -4.27134454e-01 6.88228548e-01 1.86443582e-01
4.47810769e-01 -2.88173944e-01 -1.47329554e-01 -9.35959280e-01
-1.46037981e-01 -1.06502891e+00 -1.08035304e-01 -6.89879715e-01
-1.08421886e+00 6.79173112e-01 -3.40413988e-01 -1.30757654e+00
-6.12706542e-01 -1.29790559e-01 -9.81640995e-01 1.19501066e+00
-1.34186804e+00 -9.60790813e-01 2.14782298e-01 5.01647353e-01
9.37830925e-01 -3.07537258e-01 7.82859802e-01 3.27813774e-01
-5.30723929e-01 7.41316557e-01 2.27906480e-01 1.77745186e-02
6.32181585e-01 -1.27023745e+00 5.73914766e-01 6.86316848e-01
4.42870945e-01 3.38783860e-01 8.74170303e-01 -5.45605421e-01
-1.19677866e+00 -9.61462677e-01 6.97572589e-01 -2.08624691e-01
7.08106697e-01 -3.52878779e-01 -9.83015478e-01 2.59578079e-01
2.73657650e-01 -3.13830048e-01 5.62938929e-01 -3.45330715e-01
-3.49859819e-02 -1.61624804e-01 -8.91332567e-01 7.73380518e-01
9.66476381e-01 -7.02296376e-01 -5.51188827e-01 7.36150369e-02
1.13564682e+00 -5.73107481e-01 -8.51394236e-01 2.36642718e-01
4.55109298e-01 -8.99059117e-01 8.11404765e-01 -2.41499320e-01
4.86060470e-01 -4.99103189e-01 -1.56070113e-01 -1.56810522e+00
7.12950202e-03 -1.42262948e+00 -3.43274146e-01 1.60845435e+00
5.57380795e-01 -3.86402369e-01 5.79536676e-01 -1.53494179e-01
-5.55954635e-01 -6.77962244e-01 -9.96133387e-01 -9.06075895e-01
3.79010677e-01 -6.22982562e-01 7.49620497e-01 6.35749578e-01
-2.95589268e-01 3.41088563e-01 -6.81821287e-01 2.50951856e-01
5.66968858e-01 6.78979605e-02 8.49067152e-01 -5.74330807e-01
-9.60706413e-01 -3.59760255e-01 4.02030468e-01 -1.46756542e+00
1.51616260e-01 -8.18541646e-01 3.94169778e-01 -1.18990481e+00
-3.91405910e-01 -5.88494658e-01 -3.15635204e-01 1.65315762e-01
-2.17377722e-01 2.01711029e-01 4.54041004e-01 2.16613695e-01
4.32039686e-02 9.73969877e-01 1.72828722e+00 -8.07574484e-03
-6.55767858e-01 4.83988374e-01 -2.19263762e-01 6.68880761e-01
9.82485652e-01 -5.36563694e-01 -4.20106351e-01 -4.63414863e-02
-1.86729327e-01 6.79496169e-01 1.70738325e-01 -9.67581332e-01
2.17068836e-01 -1.29906088e-01 -2.80115604e-01 -4.71534014e-01
6.71895921e-01 -1.88104495e-01 2.97350615e-01 3.95249337e-01
-3.77911329e-01 -4.92585570e-01 5.89766689e-02 4.79526848e-01
-3.55548799e-01 -4.07198817e-01 6.98779523e-01 8.38971809e-02
-5.70233464e-02 3.41707736e-01 -2.57288039e-01 2.86580592e-01
5.79924822e-01 1.05772719e-01 1.24715269e-01 -6.27636969e-01
-7.55698800e-01 -2.70069361e-01 5.45277707e-02 3.98107886e-01
6.42891288e-01 -1.45207238e+00 -8.96803737e-01 3.49961251e-01
-4.08670217e-01 6.89185411e-02 4.55603212e-01 6.27372861e-01
-2.88918823e-01 9.50285792e-02 3.31241846e-01 -2.70040959e-01
-1.02826297e+00 2.55313307e-01 5.08141458e-01 -9.63168219e-02
-6.24377251e-01 8.83665562e-01 2.53915220e-01 -6.05356693e-01
2.47686744e-01 -4.18593287e-01 2.23490484e-02 -2.81185508e-01
3.47952366e-01 4.51150984e-01 -2.85632730e-01 -5.18271685e-01
1.78172410e-01 4.68592167e-01 5.08479714e-01 -5.66302121e-01
9.77059245e-01 -1.14355989e-01 1.52922601e-01 7.35975385e-01
9.87385035e-01 7.68214881e-01 -1.47400594e+00 5.58463372e-02
-5.21611571e-01 -2.46501550e-01 3.96852382e-02 -7.37585843e-01
-9.63403046e-01 1.05538690e+00 3.76766145e-01 1.76424891e-01
1.08526790e+00 -1.20555483e-01 1.46430421e+00 -1.65419653e-01
1.97764626e-03 -1.00848866e+00 1.27077773e-02 4.34132874e-01
1.11487627e+00 -5.98519266e-01 -6.95087552e-01 -2.29874328e-01
-7.98531950e-01 9.22751904e-01 3.17782611e-01 -2.43287891e-01
6.33475959e-01 4.33990955e-01 2.66115606e-01 4.28910792e-01
-6.80387080e-01 -1.98702943e-02 3.06730002e-01 5.60526133e-01
2.91328490e-01 3.90727580e-01 -3.04822206e-01 1.16052771e+00
-6.90503597e-01 -4.03475799e-02 2.67687500e-01 8.01592600e-03
-3.09358329e-01 -1.26598763e+00 -5.41824937e-01 -2.33215448e-02
-5.75874269e-01 -5.55898726e-01 -2.10446447e-01 2.22126901e-01
2.07688585e-01 1.07588470e+00 -2.62646109e-01 -6.57768786e-01
3.55264962e-01 2.81486362e-01 2.12850854e-01 -5.70554316e-01
-7.13743031e-01 4.30085748e-01 8.27473775e-02 -2.51560628e-01
1.11153290e-01 -6.24914587e-01 -1.40523088e+00 -1.78806469e-01
-3.97721261e-01 4.80948895e-01 5.67753196e-01 6.47108495e-01
1.36994228e-01 8.33926618e-01 1.02628672e+00 -6.88104212e-01
-1.09858525e+00 -1.14059579e+00 -8.08801353e-01 1.66845068e-01
4.61862534e-01 -1.53482214e-01 -6.87419653e-01 1.59408867e-01] | [15.480462074279785, 6.177434921264648] |
c3383294-54b7-43e7-80d2-8488f5c2f13d | building-advanced-dialogue-managers-for-goal | 1806.00780 | null | http://arxiv.org/abs/1806.00780v1 | http://arxiv.org/pdf/1806.00780v1.pdf | Building Advanced Dialogue Managers for Goal-Oriented Dialogue Systems | Goal-Oriented (GO) Dialogue Systems, colloquially known as goal oriented
chatbots, help users achieve a predefined goal (e.g. book a movie ticket)
within a closed domain. A first step is to understand the user's goal by using
natural language understanding techniques. Once the goal is known, the bot must
manage a dialogue to achieve that goal, which is conducted with respect to a
learnt policy. The success of the dialogue system depends on the quality of the
policy, which is in turn reliant on the availability of high-quality training
data for the policy learning method, for instance Deep Reinforcement Learning.
Due to the domain specificity, the amount of available data is typically too
low to allow the training of good dialogue policies. In this master thesis we
introduce a transfer learning method to mitigate the effects of the low
in-domain data availability. Our transfer learning based approach improves the
bot's success rate by $20\%$ in relative terms for distant domains and we more
than double it for close domains, compared to the model without transfer
learning. Moreover, the transfer learning chatbots learn the policy up to 5 to
10 times faster. Finally, as the transfer learning approach is complementary to
additional processing such as warm-starting, we show that their joint
application gives the best outcomes. | ['Vladimir Ilievski'] | 2018-06-03 | null | null | null | null | ['goal-oriented-dialogue-systems'] | ['natural-language-processing'] | [-3.12900305e-01 5.68439305e-01 9.30574909e-03 -2.27202371e-01
-8.11754227e-01 -7.65000045e-01 5.72859883e-01 2.73775250e-01
-4.46282595e-01 1.22034740e+00 4.51520719e-02 -3.79308343e-01
4.84212041e-02 -8.18347156e-01 -3.58493656e-01 -4.63354886e-01
1.50096834e-01 9.39666867e-01 4.38351691e-01 -9.09539521e-01
5.20491242e-01 2.17301980e-01 -9.66637135e-01 -6.71211630e-02
7.61325002e-01 6.29324019e-01 3.43861192e-01 9.88675594e-01
-3.09850335e-01 1.11837304e+00 -7.78735399e-01 -1.17548235e-01
9.28112119e-02 -6.77375674e-01 -1.50113392e+00 2.10694689e-02
-1.71388596e-01 -6.19098961e-01 -1.17387893e-02 7.71566927e-01
2.84177035e-01 6.06851816e-01 5.06971240e-01 -1.15095115e+00
2.27050553e-03 4.02633846e-01 3.41877039e-03 -1.05116501e-01
4.88079369e-01 5.04920602e-01 1.19932449e+00 -1.65570751e-01
6.35438740e-01 1.15659225e+00 3.36758167e-01 9.49877501e-01
-1.13003004e+00 -1.20934248e-01 8.09811726e-02 -5.83134331e-02
-5.88736653e-01 -4.58628863e-01 5.11380136e-01 -3.94625068e-01
9.06939089e-01 -1.91808298e-01 3.81533653e-01 8.77030015e-01
-1.15023002e-01 5.14707446e-01 1.09681010e+00 -5.81555605e-01
4.76312965e-01 5.27964830e-01 8.01462773e-03 6.65368676e-01
-3.94644737e-01 -1.14103884e-01 -3.47286910e-01 -1.11891940e-01
7.97734737e-01 -4.73364443e-01 2.44317763e-02 -1.33412600e-01
-7.12322772e-01 1.01473451e+00 3.63420755e-01 6.47546113e-01
-4.96243715e-01 2.95199137e-02 5.75773239e-01 7.44828224e-01
3.35093260e-01 9.28508341e-01 -5.66806257e-01 -7.82194674e-01
-5.26739419e-01 2.68059731e-01 1.47239685e+00 7.15610683e-01
9.20642138e-01 -4.07511778e-02 -3.41325700e-02 7.87253439e-01
1.97255671e-01 1.10978350e-01 2.03336731e-01 -1.22808158e+00
5.36488533e-01 5.55144668e-01 6.37382269e-01 -5.86963475e-01
-2.44200572e-01 1.63870752e-01 -2.64284492e-01 4.77863699e-01
1.18524468e+00 -6.27091110e-01 -4.22417700e-01 1.75820971e+00
4.95366335e-01 -2.52366662e-01 2.79438943e-01 7.72212327e-01
3.75975281e-01 7.39551187e-01 2.23240256e-01 -2.46425346e-01
1.17369473e+00 -7.89952040e-01 -5.28728127e-01 -3.51395577e-01
8.86746287e-01 -5.86997747e-01 1.23217630e+00 4.05544043e-01
-9.80249107e-01 -2.48713359e-01 -5.42917311e-01 9.28336009e-02
-2.67986983e-01 -3.78649801e-01 3.52240682e-01 5.52827179e-01
-9.29233611e-01 9.58143532e-01 -6.89640999e-01 -6.53694153e-01
9.80618149e-02 4.46418315e-01 -3.61778438e-01 5.95014133e-02
-1.24225008e+00 1.24823618e+00 4.16545540e-01 -3.41989785e-01
-1.01179457e+00 -3.94783348e-01 -6.12744808e-01 2.00893357e-01
6.60073578e-01 -2.33069927e-01 1.72881830e+00 -9.65616941e-01
-2.19366264e+00 7.32491255e-01 1.40317038e-01 -4.91881132e-01
6.19275451e-01 6.14371151e-03 4.25870538e-01 2.72360682e-01
6.33228943e-02 4.21171129e-01 7.24374473e-01 -1.08568704e+00
-8.10188830e-01 -3.13008428e-01 8.41510117e-01 4.65025187e-01
-3.53489757e-01 2.12346718e-01 -1.59025192e-01 2.26665780e-01
-5.13362408e-01 -9.37953234e-01 -3.14979047e-01 -3.03930044e-01
3.98888178e-02 -6.88689709e-01 7.14630485e-01 -8.02713633e-01
8.16230536e-01 -1.81121361e+00 1.85438842e-01 3.19030359e-02
2.88824499e-01 5.15630782e-01 -7.88730085e-02 8.95591736e-01
3.16641748e-01 -1.41037256e-02 9.04683098e-02 -2.39725307e-01
1.69561338e-02 3.39791447e-01 4.66113389e-02 1.25315368e-01
1.17962271e-01 5.69331944e-01 -1.08484864e+00 -3.78518701e-01
2.30794862e-01 -1.36502504e-01 -5.72661221e-01 7.48527586e-01
-7.59530008e-01 9.19116437e-01 -6.89308524e-01 -2.42695864e-02
-5.35889454e-02 -2.08877355e-01 4.03069764e-01 5.59613705e-01
-1.36223156e-02 8.55553150e-01 -8.09193313e-01 1.50033259e+00
-9.30333614e-01 5.48070788e-01 4.64447647e-01 -1.27294016e+00
1.05229557e+00 6.49705350e-01 5.28920054e-01 -4.30212498e-01
1.11643292e-01 1.78977266e-01 2.26412669e-01 -4.05023247e-01
3.13413531e-01 -4.09257352e-01 -8.87020316e-04 7.36184061e-01
1.18373312e-01 -3.94506663e-01 2.05348700e-01 3.80465984e-01
1.12510204e+00 1.82239905e-01 2.82611907e-01 -1.69699907e-01
5.09898484e-01 4.29078966e-01 3.85596395e-01 7.35783696e-01
-2.76169628e-01 -2.87389159e-02 1.08901298e+00 -3.19613725e-01
-1.12309337e+00 -2.89109170e-01 4.40107018e-01 1.41936505e+00
-1.00326896e-01 -3.55126798e-01 -1.03465676e+00 -9.85960186e-01
-2.57386178e-01 8.14457595e-01 -3.43416601e-01 -1.41769722e-01
-7.02669322e-01 -1.56667560e-01 3.84015203e-01 1.74481019e-01
7.72340357e-01 -1.46113181e+00 -6.25363469e-01 5.44785142e-01
-2.76900589e-01 -1.02609503e+00 -1.70432299e-01 2.41842985e-01
-7.87843704e-01 -1.22030663e+00 -5.35868049e-01 -5.12721181e-01
2.62315243e-01 -1.11763388e-01 1.00467193e+00 3.16065490e-01
1.37205720e-01 4.53584373e-01 -5.83972216e-01 -3.33440989e-01
-9.83108401e-01 3.48463029e-01 2.94180308e-02 -3.51448119e-01
3.36760193e-01 -6.02871835e-01 -3.18208903e-01 2.17637196e-01
-3.52060348e-01 -1.47936270e-01 1.71951234e-01 1.07811153e+00
-3.13172340e-01 1.15096062e-01 8.01586211e-01 -8.51555228e-01
1.18722188e+00 -3.90340060e-01 -8.48305881e-01 -6.36006072e-02
-5.49051523e-01 2.09690064e-01 8.86484981e-01 -3.81273866e-01
-9.90899503e-01 -8.64648968e-02 -2.07964823e-01 4.70779836e-02
-3.38540673e-01 4.51543421e-01 7.06428383e-03 6.54893741e-02
9.55048859e-01 7.27167428e-02 5.11816800e-01 -4.55029547e-01
2.56472707e-01 7.64352202e-01 1.24080889e-02 -9.68367875e-01
5.77886462e-01 -7.39653260e-02 -3.24960113e-01 -8.95317912e-01
-6.75604880e-01 -4.68255162e-01 -5.97046375e-01 -3.16980302e-01
7.75291502e-01 -5.04820049e-01 -1.17859638e+00 3.67472619e-01
-1.19178545e+00 -1.09235084e+00 -2.52784789e-01 2.84196764e-01
-9.13339078e-01 2.09090561e-01 -6.22392237e-01 -1.14908469e+00
-1.60150677e-01 -9.96861458e-01 5.47861397e-01 3.96337777e-01
-2.77786583e-01 -1.10154283e+00 3.33848931e-02 4.36479539e-01
4.88712311e-01 -1.27326265e-01 8.20539892e-01 -1.13230824e+00
-3.47685188e-01 -1.83763757e-01 -2.87189614e-02 6.22416914e-01
2.73188978e-01 -2.98673600e-01 -7.59650409e-01 -1.86470911e-01
1.18077293e-01 -7.22813189e-01 1.39203906e-01 5.63962832e-02
3.79293948e-01 -5.00811398e-01 -1.35346642e-02 -2.99368858e-01
9.56267416e-01 3.64401579e-01 5.59172273e-01 4.77092683e-01
3.22051406e-01 9.74280059e-01 1.05369043e+00 4.89308834e-01
3.79936218e-01 8.16357851e-01 3.92998755e-01 1.96434498e-01
2.94890046e-01 -2.12399170e-01 3.33759815e-01 1.83513373e-01
-1.64972886e-01 5.32517545e-02 -9.29264367e-01 5.56090236e-01
-2.12617421e+00 -1.00528812e+00 1.82581872e-01 2.24892116e+00
1.14208663e+00 2.85100043e-01 7.97076225e-01 -5.09639159e-02
4.93020922e-01 -1.57803610e-01 -5.58233678e-01 -6.99745774e-01
7.62634635e-01 1.26484752e-01 2.95765370e-01 8.87572408e-01
-7.49745131e-01 1.42187631e+00 5.50490761e+00 5.30802846e-01
-1.05262876e+00 1.58713862e-01 4.97881770e-01 2.93040276e-01
3.14417332e-01 2.31588364e-01 -7.48979688e-01 3.97601008e-01
1.28041899e+00 -1.11297540e-01 9.53346312e-01 1.02445889e+00
5.83120525e-01 -4.56598908e-01 -1.00713277e+00 4.90620613e-01
-5.41844845e-01 -1.07911718e+00 -7.73945272e-01 3.21790248e-01
1.67813182e-01 -1.26754910e-01 -5.50246835e-01 7.08096862e-01
9.23056543e-01 -9.04485822e-01 1.23251677e-01 2.19572917e-01
3.77353877e-01 -7.95499980e-01 6.18190348e-01 1.09488547e+00
-6.27145588e-01 -1.48228124e-01 -3.50281328e-01 -4.00956810e-01
6.35592341e-02 1.20050430e-01 -1.45663178e+00 2.53792167e-01
3.31732184e-01 2.63354480e-01 1.42856732e-01 5.93025208e-01
-4.50462669e-01 8.51695359e-01 -4.70365435e-01 -5.01392901e-01
5.88259161e-01 -4.30995226e-01 4.77122396e-01 7.54535019e-01
-2.23560169e-01 4.94651794e-01 4.45780158e-01 7.99823642e-01
8.40779096e-02 1.13967240e-01 -7.28994370e-01 -2.48766571e-01
3.60383749e-01 1.21427178e+00 -3.70010078e-01 -3.30281675e-01
-1.24187581e-01 9.28031087e-01 5.52338839e-01 4.03491780e-02
-4.37218517e-01 -2.84916788e-01 6.02901697e-01 3.87677252e-02
8.70304108e-02 -3.54684740e-01 -1.31175116e-01 -8.26473415e-01
-1.62464917e-01 -1.27280176e+00 1.72328249e-01 -4.69649792e-01
-1.01421940e+00 4.24856424e-01 -1.16162241e-01 -8.95893991e-01
-9.36995268e-01 -5.55676579e-01 -6.98369265e-01 1.20065331e+00
-1.41227245e+00 -7.45782197e-01 -2.48780865e-02 5.66408873e-01
5.85176587e-01 -3.52262318e-01 1.05816770e+00 -5.71772382e-02
-3.00003231e-01 2.24154890e-01 -1.13861918e-01 2.85063326e-01
8.28337252e-01 -1.59609461e+00 6.22371994e-02 2.36124650e-01
-1.06725305e-01 5.70159137e-01 9.04620409e-01 -5.00331283e-01
-1.16855407e+00 -6.51864827e-01 8.45562935e-01 -4.80611444e-01
7.19193876e-01 -2.60018766e-01 -9.34203207e-01 5.77046752e-01
3.42123479e-01 -4.29744393e-01 3.72816503e-01 4.53997970e-01
-4.64601219e-02 -2.02900339e-02 -1.39154625e+00 4.79897469e-01
3.70590270e-01 -3.87881517e-01 -6.61798179e-01 6.22931063e-01
5.91231644e-01 -4.05184507e-01 -8.90467763e-01 -3.66123497e-01
2.18483552e-01 -9.28460777e-01 4.33275431e-01 -1.11252439e+00
4.35149819e-01 1.38337150e-01 1.42437056e-01 -1.55942333e+00
-9.68368351e-03 -1.03941250e+00 2.21984297e-01 1.22754204e+00
3.73165131e-01 -6.61758006e-01 9.42484319e-01 1.21637368e+00
2.51369119e-01 -3.74601096e-01 -7.44776011e-01 -7.45357096e-01
4.15427595e-01 -9.56488922e-02 3.47472280e-01 7.78185129e-01
6.18284106e-01 8.10134888e-01 -4.75755960e-01 -6.08531125e-02
2.05014527e-01 -3.84460688e-02 1.21457779e+00 -1.21737337e+00
-4.94842649e-01 -1.71806201e-01 -1.90409627e-02 -1.22320402e+00
2.05100000e-01 -3.83640796e-01 3.02362233e-01 -1.51417243e+00
-3.89675766e-01 -6.86836481e-01 1.94830284e-01 7.57281005e-01
-1.32058173e-01 -3.46132606e-01 3.92337412e-01 1.05442740e-01
-5.64429998e-01 3.40723991e-01 1.22627950e+00 2.38101214e-01
-7.03348696e-01 4.18898165e-01 -6.51829422e-01 6.49662197e-01
1.20390427e+00 -3.94211560e-01 -4.33562666e-01 1.14307165e-01
-1.08140111e-01 6.60288334e-01 5.93928136e-02 -6.92001224e-01
2.60753930e-01 -5.73019087e-01 -3.26040596e-01 2.56662220e-01
4.55835581e-01 -7.55148828e-01 -6.61059260e-01 4.81098741e-01
-4.89287466e-01 -3.04511964e-01 7.93126374e-02 3.71083885e-01
-8.46563205e-02 -5.77878654e-01 9.69257414e-01 -5.97692370e-01
-5.68763852e-01 -1.00998536e-01 -6.98838830e-01 1.96690187e-01
1.04382551e+00 -4.01694216e-02 -1.18660472e-01 -1.19511366e+00
-8.91701341e-01 5.45047045e-01 2.28594750e-01 1.38521612e-01
1.09207049e-01 -6.19914591e-01 -6.29369617e-01 -3.43517303e-01
-1.44390300e-01 8.81552137e-03 -2.69464940e-01 6.98806763e-01
-3.70073229e-01 3.48184377e-01 -2.80879408e-01 -3.00369561e-01
-1.28969634e+00 2.47493863e-01 6.45838857e-01 -6.62830651e-01
-5.97843885e-01 5.06184697e-01 -2.05066085e-01 -6.20336950e-01
2.76964873e-01 1.95551395e-01 -3.74714136e-01 -3.48360762e-02
4.70312715e-01 1.89812303e-01 -7.51335407e-03 -7.32104629e-02
-7.07073063e-02 -1.40459910e-01 -2.09995970e-01 -5.63060164e-01
1.49438667e+00 -3.38023826e-02 -7.34526366e-02 3.27199101e-01
7.25434601e-01 -2.13742241e-01 -1.38672638e+00 -2.73365587e-01
2.41198123e-01 -4.00868624e-01 -2.37499923e-01 -9.67447102e-01
-4.46314067e-01 7.90461779e-01 9.83323082e-02 7.02556372e-01
7.42724240e-01 -2.15320960e-01 6.85045838e-01 7.76646495e-01
6.07535601e-01 -1.41004157e+00 3.92974406e-01 7.96191633e-01
6.01911187e-01 -1.45732880e+00 -4.15059716e-01 -1.78847104e-01
-9.82285678e-01 1.15527606e+00 7.46002018e-01 -2.70170420e-01
1.06994890e-01 2.05228515e-02 2.44472131e-01 -2.01451257e-01
-8.51712644e-01 -3.57051671e-01 -2.77193397e-01 8.35571170e-01
2.76236951e-01 -1.33576229e-01 -3.19045842e-01 1.71787202e-01
-1.21307202e-01 -4.69205808e-03 6.14250600e-01 8.18051040e-01
-9.06573057e-01 -1.50314653e+00 -2.99508154e-01 2.16492921e-01
-3.45291555e-01 1.32169381e-01 -6.77586854e-01 8.11173856e-01
-5.16713142e-01 1.46121836e+00 -2.42676318e-01 -1.75262918e-03
3.19479495e-01 4.42827016e-01 3.09646666e-01 -1.08473670e+00
-9.49611127e-01 -7.53059685e-02 6.04724109e-01 -5.16628623e-01
-1.27395988e-01 -4.00658727e-01 -1.37298763e+00 -4.57075387e-01
-2.61089504e-01 5.78149319e-01 6.56061649e-01 1.26855791e+00
7.18061849e-02 1.54418051e-01 8.31084311e-01 -5.79848647e-01
-1.04049993e+00 -1.20145774e+00 -4.15328920e-01 4.17209834e-01
2.56236881e-01 -4.19366121e-01 -1.56073987e-01 -2.03384325e-01] | [13.034567832946777, 8.044500350952148] |
d923fd28-c43a-468e-8129-ea2e5fe2ae83 | image-super-resolution-improved-by-edge | null | null | https://ieeexplore.ieee.org/document/8914550 | https://ieeexplore.ieee.org/document/8914550 | Image Super-Resolution Improved by Edge Information | As well as in other knowledge domains, deep learning techniques have revolutionized the development of image super-resolution approaches. State-of-the-art algorithms for this problem have employed convolutional neural networks in residual architectures with a number of layers and generic loss functions, such as L1 and Peak Signal-to-Noise Ratio (PSNR). These frameworks (architectures + loss functions) are generic and do not address the main characteristics of an image for human visual perception (luminance, contrast, and structure) resulting in better images, however, with noise mainly at the edges. In this work, we present an edge enhanced super-resolution (EESR) method using a novel residual neural network with focus on image edges and a mix of loss functions that use PSNR, L1, Multiple-Scale Structural Similarity (MS-SSIM), and a new loss function based on the pencil sketch technique. As main contribution, the proposed framework aims to leverage the limits of image super-resolution and presents an improvement of the results in terms of the SSIM metric and achieving competitive results for the PSNR metric. | ['Helio Pedrini', 'Eldrey Galindo'] | 2019-10-06 | null | null | null | smc-2019-10 | ['ms-ssim'] | ['computer-vision'] | [ 6.01974905e-01 -1.85983330e-01 6.31682798e-02 -6.11825064e-02
-5.59165061e-01 -1.53148636e-01 4.62003767e-01 -2.12193653e-01
-4.06283289e-01 8.05772722e-01 3.03320646e-01 2.55669296e-01
-4.35685456e-01 -8.75759363e-01 -6.41212583e-01 -5.61975062e-01
-1.93176776e-01 -5.60270429e-01 5.59055805e-01 -5.47487736e-01
3.85831684e-01 8.12071323e-01 -1.80257130e+00 5.49678326e-01
6.80103958e-01 1.23854816e+00 2.19707876e-01 7.22936928e-01
1.04137801e-01 1.08158159e+00 -4.85462964e-01 -4.44498658e-01
3.54868770e-01 -4.01474535e-01 -6.12718642e-01 -2.51723975e-01
7.79259384e-01 -5.25788784e-01 -6.13065243e-01 1.09389162e+00
7.32600391e-01 2.54874766e-01 2.86500543e-01 -5.93418002e-01
-9.12724078e-01 4.75924700e-01 -6.69101715e-01 6.78039253e-01
1.98385850e-01 -2.90890455e-01 7.88066924e-01 -9.23860908e-01
6.77260041e-01 1.30669510e+00 9.51754808e-01 2.16599271e-01
-1.21906102e+00 -4.20593500e-01 -3.09723318e-01 7.19933093e-01
-1.28719723e+00 -2.91517735e-01 8.25471997e-01 4.09817509e-02
8.46367002e-01 1.48985788e-01 5.58344238e-02 9.83507037e-01
4.41038668e-01 4.01791364e-01 1.41929853e+00 -4.35616493e-01
8.55655223e-02 1.91360235e-01 -2.54214913e-01 3.27118635e-01
2.27592900e-01 4.10602927e-01 -6.00711584e-01 5.08151889e-01
1.39472854e+00 -9.69485790e-02 -3.65488410e-01 -2.64941037e-01
-8.18247020e-01 6.68625891e-01 6.76008582e-01 7.78152287e-01
-3.72260183e-01 7.94829130e-02 3.27807426e-01 2.99048662e-01
3.44633639e-01 1.92455038e-01 -2.48431906e-01 1.81278035e-01
-1.28658879e+00 1.67200476e-01 3.92422378e-01 5.46943009e-01
5.35433531e-01 2.87310809e-01 -4.88727540e-01 9.49421465e-01
-2.36957580e-01 8.14481452e-02 4.37634528e-01 -1.34653604e+00
1.52278781e-01 3.30046743e-01 1.02724925e-01 -1.35760558e+00
-4.70909297e-01 -9.80311453e-01 -1.19076443e+00 8.36388826e-01
-2.34821555e-03 2.61252463e-01 -7.22292364e-01 1.68070090e+00
-4.41088788e-02 2.10200429e-01 1.66993245e-01 1.09055853e+00
9.59361672e-01 3.40746135e-01 -2.33380139e-01 -1.54886201e-01
1.19556403e+00 -9.49979067e-01 -8.25985730e-01 1.04139457e-02
-2.00692311e-01 -8.29483628e-01 8.19372416e-01 4.52806145e-01
-1.79070711e+00 -1.21579552e+00 -1.33709443e+00 -4.70193565e-01
-3.08977693e-01 6.30300343e-02 1.63595229e-01 6.26648307e-01
-1.58866894e+00 1.14430809e+00 -3.47568184e-01 -1.99450418e-01
3.60202461e-01 2.50412345e-01 -3.11434746e-01 -1.29540801e-01
-1.29853606e+00 9.72542167e-01 3.36969912e-01 -4.83692326e-02
-5.03851831e-01 -8.18474591e-01 -7.11312175e-01 4.29356575e-01
2.40097240e-01 -5.58525383e-01 5.55274725e-01 -1.11727607e+00
-1.57451952e+00 6.91593170e-01 2.70340979e-01 -6.53171003e-01
6.99220240e-01 -2.74982423e-01 -6.40791655e-01 4.61363077e-01
-2.06769183e-01 3.85996103e-01 1.08351898e+00 -1.48011732e+00
-6.72781110e-01 -3.51227552e-01 3.71109769e-02 7.96665903e-03
-1.16798744e-01 2.36507103e-01 -1.11261576e-01 -9.14961994e-01
-3.39486115e-02 -3.07658404e-01 4.54834430e-03 1.18407823e-01
2.05025166e-01 4.12431598e-01 8.24502289e-01 -1.14111102e+00
1.25705326e+00 -2.24745560e+00 1.74237013e-01 -8.33993778e-02
2.88397133e-01 6.16758227e-01 -3.40910226e-01 3.05182412e-02
-3.41236204e-01 1.03882235e-02 -3.77192199e-01 -2.24032879e-01
-4.13450986e-01 -3.90888983e-03 3.23754437e-02 3.36209387e-01
2.52751559e-01 6.64988697e-01 -7.10859060e-01 -4.25995559e-01
5.67902505e-01 1.26479077e+00 -3.52162689e-01 4.52186503e-02
4.98710871e-01 2.75362581e-01 1.22361682e-01 2.62612104e-01
1.19495833e+00 -5.18615805e-02 -1.44976690e-01 -7.44874120e-01
-4.33363616e-01 -3.37103814e-01 -1.46615684e+00 1.53233874e+00
-7.15592325e-01 7.55962074e-01 3.79137218e-01 -7.77553380e-01
1.03016543e+00 1.26289785e-01 5.21250367e-01 -1.10317755e+00
-3.69963720e-02 3.30936372e-01 -3.17434520e-01 -3.51538897e-01
6.37744308e-01 -1.51020742e-03 6.76071525e-01 -2.36388654e-01
1.52342141e-01 2.58263767e-01 6.53569922e-02 -6.99868202e-02
9.59572375e-01 2.19043300e-01 2.46105179e-01 -1.24936417e-01
1.04204881e+00 -5.99732518e-01 2.18604282e-01 6.58197641e-01
-2.42692873e-01 1.16643512e+00 2.12701768e-01 -3.95011663e-01
-1.58250260e+00 -1.05753827e+00 -2.06528023e-01 8.60015750e-01
2.08917946e-01 1.15778156e-01 -7.87920356e-01 7.95235038e-02
-3.78149778e-01 5.53609967e-01 -5.91602385e-01 -7.35740438e-02
-8.01657796e-01 -4.73584980e-01 3.74736667e-01 5.93537629e-01
1.11301112e+00 -1.06140101e+00 -8.60949278e-01 1.42802373e-01
-1.67753607e-01 -1.51322210e+00 -3.14159781e-01 -1.64839216e-02
-9.10509348e-01 -8.66294682e-01 -1.17886782e+00 -7.28701949e-01
2.09084824e-01 2.86001593e-01 1.11297762e+00 -2.05158219e-01
-6.20296597e-01 1.64231375e-01 -4.13119644e-01 2.56338805e-01
-3.00753206e-01 -1.86039120e-01 -4.33336884e-01 1.97060898e-01
-2.47904584e-01 -9.08910930e-01 -1.03902590e+00 1.41502887e-01
-1.20890737e+00 -7.30022192e-02 9.94467378e-01 7.92995751e-01
4.62492704e-01 4.43733245e-01 3.63943756e-01 -4.76274371e-01
5.89027822e-01 -1.12330124e-01 -3.98273289e-01 1.31988049e-01
-7.77721405e-01 1.14785284e-01 7.76180506e-01 -3.06388140e-01
-1.39899528e+00 -4.13919240e-01 -1.18391104e-01 -5.57141662e-01
-1.11270495e-01 6.79203868e-02 1.24585263e-01 -5.35412371e-01
8.30432177e-01 4.23069447e-01 -3.57622690e-02 -5.67532122e-01
1.97517827e-01 4.35408890e-01 8.72154176e-01 1.89012010e-02
5.99784672e-01 6.70843720e-01 5.19400656e-01 -7.72255778e-01
-6.21390998e-01 -3.53053331e-01 -5.68455696e-01 -1.20997205e-01
7.15429962e-01 -8.62899482e-01 -6.14028752e-01 4.08993870e-01
-9.70008492e-01 9.33902413e-02 -3.66418004e-01 3.97509933e-01
-6.32251620e-01 5.87750912e-01 -7.59443939e-01 -6.90677464e-01
-6.17363214e-01 -1.07730508e+00 7.88885832e-01 4.56812292e-01
4.14343476e-01 -7.60230839e-01 -2.23764971e-01 3.22946727e-01
1.33821547e+00 6.02981031e-01 7.58880913e-01 2.59588137e-02
-5.81745982e-01 1.15201361e-01 -9.51006413e-01 8.82413268e-01
-2.28823461e-02 -5.16454458e-01 -1.00269735e+00 -4.18278843e-01
2.15297833e-01 4.32639197e-02 1.00234044e+00 6.91677332e-01
1.01867974e+00 -2.37385094e-01 2.99053788e-01 9.44889724e-01
2.03765583e+00 -7.12579563e-02 1.22945333e+00 5.94021261e-01
3.84423316e-01 4.98312443e-01 1.04346834e-01 4.61919844e-01
-6.00067861e-02 8.79678011e-01 5.46921670e-01 -4.25337762e-01
-9.39868271e-01 1.77790880e-01 2.85025179e-01 2.94937611e-01
-6.37136817e-01 1.33688837e-01 -1.88256472e-01 5.21148324e-01
-1.53078365e+00 -1.05229867e+00 1.62420180e-02 2.31832218e+00
6.06206894e-01 5.76222166e-02 -1.00785092e-01 3.51642787e-01
8.56415212e-01 2.45000213e-01 -5.81418216e-01 -5.17939150e-01
-7.20936060e-01 6.94988072e-01 9.10813570e-01 5.67691267e-01
-9.38212693e-01 6.61100805e-01 6.10962439e+00 9.81213391e-01
-1.18111265e+00 1.80185422e-01 6.77717924e-01 8.69702995e-02
-2.32676156e-02 -4.39891040e-01 -4.40837562e-01 2.17821956e-01
8.78790736e-01 -2.81975195e-02 8.59501421e-01 6.24590695e-01
1.93536043e-01 8.40479732e-02 -6.32647872e-01 1.21198356e+00
3.73164058e-01 -1.30548346e+00 2.09097579e-01 -1.93857327e-01
7.72911012e-01 -1.28205165e-01 5.44304729e-01 3.07887383e-02
-3.55825067e-01 -1.35632622e+00 5.87763667e-01 6.26966238e-01
1.07166922e+00 -9.58743811e-01 1.03152370e+00 -3.50806504e-01
-1.23520839e+00 -4.20355439e-01 -4.32274133e-01 2.70645797e-01
-1.50257915e-01 3.93602312e-01 -3.02114692e-02 7.94934332e-01
1.09834468e+00 5.26976466e-01 -5.63739061e-01 1.00964189e+00
1.73151776e-01 1.93940699e-02 8.09029862e-02 5.21109998e-01
1.56325713e-01 -2.37357300e-02 7.67543495e-01 1.37379289e+00
4.87412184e-01 -1.57733150e-02 -3.32045794e-01 1.06135046e+00
-3.37157622e-02 8.38191211e-02 -1.09244972e-01 4.45825785e-01
1.00670151e-01 1.33212674e+00 -5.61523616e-01 6.98280483e-02
-4.99802977e-01 1.20984757e+00 -9.20330435e-02 5.24864554e-01
-6.13190770e-01 -5.87511659e-01 4.59911913e-01 5.09297490e-01
5.50495207e-01 -9.19837505e-02 -3.19841474e-01 -7.56634057e-01
1.34668276e-01 -9.57754850e-01 1.81006029e-01 -8.30290496e-01
-9.83697295e-01 8.66703033e-01 -1.56236917e-01 -1.09859514e+00
-3.05979494e-02 -6.72806561e-01 -1.87419862e-01 1.01843047e+00
-2.14990854e+00 -1.23500633e+00 -7.13138819e-01 7.47700989e-01
6.48776770e-01 -1.31135374e-01 3.08769703e-01 5.94179869e-01
-4.60367799e-02 7.49949753e-01 3.15821558e-01 -1.18080657e-02
6.90082312e-01 -1.05935097e+00 2.50243843e-01 9.05171633e-01
-4.76711333e-01 3.59211951e-01 8.89687240e-01 -1.82213068e-01
-1.04088080e+00 -8.53879452e-01 3.46890390e-01 1.08691148e-01
3.49125981e-01 1.19980477e-01 -9.01871145e-01 2.01516494e-01
2.28670374e-01 1.96812406e-01 -6.14337139e-02 -4.78297889e-01
-3.32881093e-01 -3.38557333e-01 -1.68539548e+00 4.50277478e-01
9.20302391e-01 -4.32356596e-01 -1.62924215e-01 -2.95906007e-01
6.51140451e-01 -2.78185904e-01 -1.10301006e+00 6.83402002e-01
5.69479048e-01 -1.75346982e+00 1.54913247e+00 7.61634065e-03
6.60919547e-01 -3.67498964e-01 -3.46901119e-01 -9.25587475e-01
-8.03525507e-01 -3.23831052e-01 -2.69975185e-01 9.28293824e-01
7.22362846e-02 -3.17200243e-01 5.08595407e-01 1.75046504e-01
-2.45207902e-02 -6.70813322e-01 -9.64703977e-01 -6.72295809e-01
-2.08136961e-01 6.41041547e-02 2.34394163e-01 4.81202483e-01
-6.30093396e-01 6.13538213e-02 -7.03691959e-01 1.32735476e-01
1.06665051e+00 -2.49925211e-01 1.82459101e-01 -1.12016344e+00
-1.49499327e-01 -6.37484729e-01 -6.97928965e-01 -5.84783435e-01
-3.10334474e-01 -4.36165124e-01 -3.77885073e-01 -1.55438638e+00
2.88786411e-01 1.64751619e-01 -6.34045780e-01 -9.22843888e-02
4.58879098e-02 5.56470573e-01 1.69073164e-01 4.12857682e-02
-3.74441236e-01 3.04134667e-01 1.32068503e+00 7.93201625e-02
-2.13164449e-01 -2.60367095e-01 -5.73942125e-01 5.72899103e-01
6.50674343e-01 1.05673127e-01 -1.83748424e-01 -3.40632081e-01
2.48110548e-01 5.81513904e-02 5.68894207e-01 -1.48404980e+00
2.74511844e-01 2.66927630e-01 8.59482884e-01 -4.23406422e-01
4.90106016e-01 -8.44371855e-01 2.38280237e-01 4.25195813e-01
-5.39590240e-01 -7.33883977e-02 1.87190920e-01 3.92638654e-01
-3.44050050e-01 -3.86779010e-02 1.53410196e+00 -3.32499206e-01
-9.16158617e-01 -1.98613163e-02 1.82976618e-01 -1.75801203e-01
8.71976852e-01 -6.50251865e-01 -3.13048631e-01 -3.68991077e-01
-6.81655109e-01 -4.43688691e-01 5.15733600e-01 3.72902602e-01
9.71510112e-01 -9.93432403e-01 -1.22873127e+00 1.34177446e-01
-3.43474269e-01 -4.61632848e-01 6.96847200e-01 7.79486418e-01
-7.06184983e-01 3.42325360e-01 -9.29976404e-01 -2.16135815e-01
-1.29933691e+00 8.13737214e-01 6.75129294e-01 -4.48890179e-01
-8.14862728e-01 6.44466817e-01 1.41141236e-01 1.65682793e-01
3.60577762e-01 -1.13136873e-01 -5.62197030e-01 -1.69159964e-01
9.66562867e-01 8.84055316e-01 7.63434693e-02 -7.63150871e-01
-1.86839566e-01 8.31851900e-01 -1.50402859e-01 3.78337391e-02
1.58177722e+00 -4.99463618e-01 -1.09856874e-01 3.52723077e-02
1.27858114e+00 -1.24922819e-01 -1.29378545e+00 -3.87535542e-01
-2.01870427e-01 -6.69353902e-01 6.06177449e-01 -9.76708949e-01
-1.36426699e+00 5.94589412e-01 1.46352351e+00 4.19499688e-02
1.63098478e+00 -3.74409020e-01 9.75517988e-01 -2.73615092e-01
2.33285740e-01 -1.13955700e+00 3.21884185e-01 2.35759959e-01
1.04837751e+00 -1.13551974e+00 6.16798140e-02 -3.27779919e-01
-3.92521650e-01 1.28811896e+00 2.49778986e-01 -3.45868230e-01
4.66186464e-01 4.14223611e-01 -2.20312178e-01 1.85669303e-01
-2.24120289e-01 -3.59309167e-01 3.05889755e-01 7.18286097e-01
6.00515485e-01 -4.02110428e-01 -4.71926361e-01 4.05957401e-01
1.46992877e-01 3.69993329e-01 5.20598412e-01 4.45200682e-01
-4.87181246e-01 -6.71398759e-01 -5.53812385e-01 2.37158164e-02
-9.55267727e-01 -2.64930546e-01 8.76031891e-02 5.31525135e-01
3.77601832e-01 9.57177699e-01 -1.71132982e-01 -2.60109365e-01
4.81295019e-01 -6.41321242e-01 6.15769029e-01 9.55676064e-02
-5.23842871e-01 -1.94166407e-01 -2.47950912e-01 -7.55988836e-01
-6.79799557e-01 -7.37262592e-02 -8.59194815e-01 -4.04116035e-01
-5.87381087e-02 -3.21178138e-01 7.91100085e-01 4.95923877e-01
3.41621339e-01 8.76556695e-01 5.80859959e-01 -9.92446005e-01
-2.99637258e-01 -9.53980744e-01 -7.29378998e-01 5.44931650e-01
6.82295203e-01 -4.80934799e-01 -4.36827421e-01 5.12168650e-03] | [11.101686477661133, -2.0299758911132812] |
4b3a39db-171d-44b3-aa00-c966e2c4c78f | pan-towards-efficient-and-accurate-end-to-end | 2105.00405 | null | https://arxiv.org/abs/2105.00405v4 | https://arxiv.org/pdf/2105.00405v4.pdf | PAN++: Towards Efficient and Accurate End-to-End Spotting of Arbitrarily-Shaped Text | Scene text detection and recognition have been well explored in the past few years. Despite the progress, efficient and accurate end-to-end spotting of arbitrarily-shaped text remains challenging. In this work, we propose an end-to-end text spotting framework, termed PAN++, which can efficiently detect and recognize text of arbitrary shapes in natural scenes. PAN++ is based on the kernel representation that reformulates a text line as a text kernel (central region) surrounded by peripheral pixels. By systematically comparing with existing scene text representations, we show that our kernel representation can not only describe arbitrarily-shaped text but also well distinguish adjacent text. Moreover, as a pixel-based representation, the kernel representation can be predicted by a single fully convolutional network, which is very friendly to real-time applications. Taking the advantages of the kernel representation, we design a series of components as follows: 1) a computationally efficient feature enhancement network composed of stacked Feature Pyramid Enhancement Modules (FPEMs); 2) a lightweight detection head cooperating with Pixel Aggregation (PA); and 3) an efficient attention-based recognition head with Masked RoI. Benefiting from the kernel representation and the tailored components, our method achieves high inference speed while maintaining competitive accuracy. Extensive experiments show the superiority of our method. For example, the proposed PAN++ achieves an end-to-end text spotting F-measure of 64.9 at 29.2 FPS on the Total-Text dataset, which significantly outperforms the previous best method. Code will be available at: https://git.io/PAN. | ['Chunhua Shen', 'Tong Lu', 'Zhibo Yang', 'Ding Liang', 'Xuebo Liu', 'Xiang Li', 'Enze Xie', 'Wenhai Wang'] | 2021-05-02 | null | null | null | null | ['text-spotting', 'scene-text-detection'] | ['computer-vision', 'computer-vision'] | [ 5.34732342e-01 -4.42647010e-01 1.25257492e-01 -2.16288149e-01
-6.43610537e-01 -2.09023178e-01 5.11044323e-01 -1.87932495e-02
-2.59285927e-01 -2.92957146e-02 3.36733125e-02 -1.67737603e-01
3.32344890e-01 -6.97732508e-01 -6.35667324e-01 -6.73265040e-01
5.46439171e-01 1.81937963e-01 6.73917294e-01 1.94335684e-01
5.69167078e-01 3.84914637e-01 -1.59062779e+00 5.49132466e-01
1.15328574e+00 1.27482927e+00 5.72047174e-01 8.86234641e-01
-2.87574619e-01 8.42289865e-01 -3.32841247e-01 -2.80578256e-01
1.05759658e-01 -2.36108750e-01 -2.69247621e-01 4.47309464e-01
6.79308772e-01 -6.02824628e-01 -5.25700510e-01 7.70916104e-01
6.96016073e-01 3.51936854e-02 5.08071423e-01 -7.88906157e-01
-7.39629030e-01 5.07258654e-01 -9.43195462e-01 -5.36575802e-02
1.91191256e-01 2.64369458e-01 9.01565909e-01 -1.30541635e+00
1.93482980e-01 9.24906433e-01 7.03761518e-01 2.52733320e-01
-9.45802033e-01 -4.50815260e-01 1.07301950e-01 8.13190341e-02
-1.52757955e+00 -4.61161762e-01 7.20962107e-01 -2.80049115e-01
9.37642395e-01 4.75398064e-01 4.09854621e-01 6.81998074e-01
1.14831686e-01 1.53470683e+00 7.24165082e-01 -5.44869602e-01
-4.45298962e-02 7.09413737e-02 1.75296530e-01 9.52778101e-01
8.35443214e-02 -2.54832655e-01 -6.29405856e-01 1.72853008e-01
6.98084295e-01 3.14716130e-01 -4.53680396e-01 -2.01413915e-01
-1.41726863e+00 4.08869714e-01 3.81404877e-01 2.23898977e-01
-2.87620455e-01 2.84084439e-01 4.69005913e-01 -1.88330114e-01
3.75398368e-01 -1.41014785e-01 -4.72314134e-02 -6.04009032e-02
-1.36405134e+00 -4.42140177e-02 3.72814000e-01 8.29336643e-01
5.98955274e-01 1.70153424e-01 -5.59553266e-01 1.05952287e+00
1.74195111e-01 8.43337715e-01 7.34477282e-01 -7.65062943e-02
6.39922738e-01 9.42032695e-01 -5.84844612e-02 -9.33357894e-01
-3.71438205e-01 -1.99755490e-01 -9.49679136e-01 6.45747315e-03
2.80553639e-01 -3.11951172e-02 -9.60917890e-01 9.55889106e-01
3.20676357e-01 1.03678823e-01 -1.64376453e-01 9.84124839e-01
8.65403831e-01 8.10173333e-01 -1.69039428e-01 3.03786933e-01
1.59883964e+00 -1.33946836e+00 -6.31458461e-01 -2.60835826e-01
7.01419473e-01 -1.00706983e+00 1.22161031e+00 4.00551319e-01
-1.03736413e+00 -5.92739522e-01 -9.72177207e-01 -4.10674810e-01
-3.73200417e-01 1.02361214e+00 3.78033519e-01 7.79369354e-01
-1.07904053e+00 1.65122524e-01 -8.19219351e-01 -5.33950686e-01
5.29697776e-01 4.31458175e-01 -7.77844489e-02 9.46645960e-02
-5.66241205e-01 2.65492290e-01 4.67786521e-01 2.50014186e-01
-5.05766451e-01 -4.48609024e-01 -7.44833946e-01 3.81643981e-01
4.93959695e-01 -5.27431726e-01 1.07781875e+00 -1.05894673e+00
-1.67522407e+00 7.67204523e-01 -2.83171773e-01 -2.82695234e-01
7.08765388e-01 -4.07945573e-01 -1.76965162e-01 3.24468225e-01
-9.70325693e-02 6.04145467e-01 1.26801634e+00 -7.95659542e-01
-1.00512171e+00 -2.97971517e-01 -4.75469887e-01 3.91708612e-01
-6.88265979e-01 3.23889405e-01 -9.47655201e-01 -1.01150501e+00
4.13125530e-02 -5.37732899e-01 1.62714317e-01 4.33088183e-01
-8.03561926e-01 -1.98620319e-01 1.36443400e+00 -7.45520711e-01
1.36671388e+00 -2.23037386e+00 -2.51661330e-01 -1.27848387e-01
3.52291375e-01 4.57374245e-01 3.89011055e-02 3.28117907e-01
6.92225397e-02 -1.18064821e-01 -1.29243881e-01 -5.04210949e-01
1.90918356e-01 -6.03377283e-01 -4.98946875e-01 6.35314941e-01
1.27450690e-01 1.13978910e+00 -3.52600783e-01 -6.95310235e-01
8.34155977e-01 5.46770573e-01 -1.91460550e-01 1.39805660e-01
-2.49824107e-01 -2.58937269e-01 -5.78819156e-01 1.00610375e+00
8.22849810e-01 -4.84804213e-01 -2.27536678e-01 -2.35255346e-01
-2.16414645e-01 -2.35019594e-01 -1.13552761e+00 1.29011405e+00
-1.07471265e-01 9.26070392e-01 1.56910896e-01 -7.71953940e-01
1.05905735e+00 -4.16478366e-02 2.39246011e-01 -6.35200739e-01
4.31261033e-01 4.52676192e-02 -5.31537056e-01 -3.93353134e-01
8.20971012e-01 3.38685572e-01 3.84616368e-02 5.87532163e-01
-3.72237086e-01 1.65910944e-01 2.63291113e-02 1.51769280e-01
1.01468742e+00 2.26204228e-02 5.03668981e-03 -1.03320263e-01
7.44756043e-01 -1.04568243e-01 1.11547649e-01 7.93404460e-01
-1.72909200e-01 7.87067771e-01 1.64965868e-01 -4.98318702e-01
-9.05733824e-01 -8.46244693e-01 -2.58547425e-01 1.18118322e+00
3.19270998e-01 -4.47039813e-01 -8.02672029e-01 -5.89035630e-01
3.18386666e-02 4.46977973e-01 -5.59064209e-01 2.28342295e-01
-5.48452199e-01 -7.32035816e-01 7.51601815e-01 7.74861276e-01
1.07055509e+00 -9.85490143e-01 -7.86556005e-01 -1.11457989e-01
-7.96354115e-02 -1.23310292e+00 -9.55951273e-01 1.12313569e-01
-6.59376323e-01 -7.94217527e-01 -1.06843364e+00 -9.55272734e-01
6.99406385e-01 7.16542125e-01 5.02182424e-01 1.12066977e-01
-5.37427247e-01 4.17273968e-01 -4.72715914e-01 -1.60849914e-01
4.23266506e-03 -1.50093839e-01 -3.43374074e-01 5.66577137e-01
2.70147026e-01 -1.58998147e-02 -7.69549310e-01 4.42189068e-01
-1.04609787e+00 6.88199759e-01 8.32299888e-01 7.83165693e-01
5.31201899e-01 2.47691646e-02 9.40936431e-02 -5.80085278e-01
3.36747438e-01 3.16578150e-02 -7.56174266e-01 4.00605083e-01
-3.02315652e-01 -2.17445165e-01 8.57903957e-01 -3.99870783e-01
-1.12139916e+00 4.88681227e-01 8.92861113e-02 -4.09938574e-01
-2.37630114e-01 1.38255671e-01 -4.00598273e-02 -2.39331827e-01
4.54693407e-01 9.57657337e-01 -1.59083784e-01 -3.40416253e-01
3.43685448e-01 1.25870466e+00 5.08363247e-01 -4.16187853e-01
6.28258228e-01 7.14994252e-01 -4.20804352e-01 -1.26462173e+00
-4.27958548e-01 -6.16057158e-01 -5.44997036e-01 -2.79459357e-01
8.10325503e-01 -9.48132753e-01 -7.83526123e-01 8.64114285e-01
-8.83845925e-01 -6.13675773e-01 9.08495486e-02 2.54970819e-01
-5.28662860e-01 8.96750927e-01 -6.78541839e-01 -8.98833871e-01
-9.59301233e-01 -9.53072548e-01 1.69292462e+00 3.84765714e-01
1.84925631e-01 -7.00357199e-01 -3.74981731e-01 5.18126428e-01
4.89078015e-01 6.60936756e-04 5.09839356e-01 -4.53167200e-01
-7.31954098e-01 -4.79608506e-01 -8.16941261e-01 1.23152718e-01
-3.29155959e-02 1.04998954e-01 -1.10830307e+00 -2.36748442e-01
-2.93860883e-01 -2.10318863e-01 1.18304014e+00 4.24533606e-01
1.31968474e+00 8.64198580e-02 -5.48497260e-01 7.17786551e-01
1.41575730e+00 2.29372010e-02 5.87254167e-01 3.02564234e-01
9.31065857e-01 2.80000180e-01 4.74698424e-01 6.42471910e-01
9.21861753e-02 6.70601845e-01 6.77523538e-02 -2.85745829e-01
-3.29704911e-01 -1.04699336e-01 5.51660299e-01 4.64877456e-01
4.06988829e-01 -4.79205012e-01 -9.61349308e-01 3.68786901e-01
-2.02085376e+00 -8.11118603e-01 -3.91402334e-01 2.11931562e+00
4.44375813e-01 3.50722717e-03 1.61006063e-01 1.39224470e-01
1.14387834e+00 2.54062235e-01 -7.16225505e-01 -2.00200140e-01
-3.08088571e-01 -1.53825432e-02 4.59618837e-01 1.41891375e-01
-1.44190729e+00 1.18078625e+00 5.40340662e+00 1.33271861e+00
-1.35422957e+00 -2.46755227e-01 7.01344550e-01 -1.07797096e-02
3.66922200e-01 -2.51173079e-01 -9.76118863e-01 6.42140448e-01
5.02032399e-01 -1.78272903e-01 2.35419661e-01 8.28958869e-01
2.53026247e-01 -1.64322317e-01 -6.65594280e-01 1.16502380e+00
4.11492109e-01 -1.26410043e+00 1.80485159e-01 -2.45129779e-01
4.90553409e-01 3.77008654e-02 1.69438303e-01 6.33124635e-02
-5.55174202e-02 -8.41319978e-01 9.13528800e-01 3.38265449e-01
1.15334427e+00 -5.45698404e-01 4.01815623e-01 4.10118490e-01
-1.59332657e+00 -1.63712412e-01 -4.39160109e-01 2.53838062e-01
-1.39889553e-01 8.06211770e-01 -7.47311652e-01 4.79076713e-01
6.18038356e-01 9.15364504e-01 -6.56421304e-01 1.19804907e+00
-1.26961902e-01 6.52315974e-01 -4.58875597e-01 -4.65807140e-01
2.44288266e-01 -8.66013672e-03 3.11900556e-01 1.68595636e+00
3.03369969e-01 -1.38908654e-01 1.68676898e-01 9.23016846e-01
-1.22989528e-01 4.22587872e-01 -3.01143259e-01 -5.17866947e-02
2.04097211e-01 1.48075855e+00 -1.20825088e+00 -4.78504568e-01
-5.67763448e-01 1.57843888e+00 2.95210153e-01 2.57268399e-01
-9.67718780e-01 -9.68903542e-01 3.55667509e-02 -5.20727672e-02
9.17368412e-01 4.40366045e-02 -4.79836822e-01 -1.32319510e+00
4.55159545e-01 -6.26878738e-01 2.02116773e-01 -1.03920567e+00
-1.03699493e+00 4.29344356e-01 -7.01448619e-01 -1.18757963e+00
5.82613051e-01 -8.90092432e-01 -8.03692222e-01 7.76764035e-01
-1.42494166e+00 -1.51168299e+00 -7.34509706e-01 5.97028732e-01
8.30694318e-01 6.74149096e-02 4.73981887e-01 1.24502487e-01
-1.07214642e+00 8.56236219e-01 5.30083954e-01 4.94768858e-01
7.97355831e-01 -1.14886630e+00 5.55592060e-01 1.01047122e+00
9.78429765e-02 1.83482230e-01 1.85563385e-01 -7.35810101e-01
-1.64636064e+00 -1.34745598e+00 6.09738052e-01 -3.17997128e-01
5.79416633e-01 -5.87997913e-01 -9.22590911e-01 5.54473877e-01
1.10578917e-01 -4.37952094e-02 3.70632082e-01 -3.67974937e-01
-2.90141493e-01 -1.41451955e-01 -8.49641025e-01 7.60155976e-01
7.25197971e-01 -3.10642809e-01 -3.05284232e-01 5.18287778e-01
3.35467964e-01 -5.50479293e-01 -3.51808488e-01 9.46834385e-02
6.23789668e-01 -1.13372028e+00 7.97520041e-01 1.19235545e-01
3.57355833e-01 -4.26361650e-01 -2.15649027e-02 -4.76781487e-01
-3.47645849e-01 -6.22655451e-01 -2.03706294e-01 1.15659499e+00
1.27602458e-01 -5.67893565e-01 1.03755701e+00 3.88855368e-01
-2.61537045e-01 -8.62101793e-01 -7.90207863e-01 -6.38901353e-01
-1.82312772e-01 -4.87352103e-01 2.21069932e-01 6.41363740e-01
9.67573076e-02 2.18069330e-01 -4.81181026e-01 1.76323757e-01
4.82461959e-01 5.07492483e-01 7.90946305e-01 -9.58489418e-01
-1.89127207e-01 -8.37714314e-01 -4.25475031e-01 -1.65240932e+00
-1.17179222e-01 -8.12986851e-01 2.35303387e-01 -1.51567197e+00
5.11600196e-01 -1.43637300e-01 -7.45607987e-02 6.22789145e-01
-5.09149313e-01 2.71163702e-01 2.04162225e-01 2.03365162e-01
-1.00991142e+00 7.22024024e-01 9.94119823e-01 -1.81468099e-01
-8.43309164e-02 -2.06714913e-01 -5.15237510e-01 6.72264457e-01
7.91384637e-01 -2.41113380e-02 8.82165134e-02 -3.76929581e-01
-5.41369133e-02 -1.25930697e-01 3.60733122e-01 -1.19205880e+00
6.18812144e-01 2.17144623e-01 7.00551271e-01 -1.01851916e+00
2.22342908e-01 -7.65737474e-01 -3.60867083e-01 3.78636569e-01
-2.09146589e-01 -2.65478671e-01 3.87722224e-01 6.29059196e-01
-1.75621565e-02 -1.17349133e-01 8.19639623e-01 4.03031141e-01
-8.66214991e-01 1.37435868e-01 -4.59275514e-01 -3.43543589e-01
1.20061469e+00 -6.10345185e-01 -5.53420722e-01 -1.22477517e-01
-1.33606210e-01 3.24618250e-01 6.36142612e-01 2.79245406e-01
9.14466023e-01 -1.01578772e+00 -7.95033395e-01 4.17330712e-01
2.24039286e-01 -9.99366120e-02 5.56366563e-01 1.06057692e+00
-8.55140150e-01 6.71678543e-01 1.65040970e-01 -7.21145868e-01
-1.44352472e+00 5.78648925e-01 2.83130825e-01 -1.35298848e-01
-1.06317461e+00 7.07198918e-01 4.57441688e-01 -1.89873159e-01
3.71121407e-01 -4.05629277e-01 1.31331980e-01 -3.97240430e-01
8.84677827e-01 3.44742924e-01 -7.96665400e-02 -5.63302755e-01
-1.91818222e-01 9.90757883e-01 -3.13667446e-01 2.29243532e-01
8.99041653e-01 -1.59288734e-01 6.59407079e-02 1.48700774e-01
8.37862194e-01 7.12247267e-02 -1.32944560e+00 -4.23109114e-01
-2.89705545e-01 -5.83292127e-01 3.39664549e-01 -8.74914646e-01
-1.01465464e+00 9.45751131e-01 5.75442255e-01 1.14986055e-01
1.38225758e+00 -2.16652483e-01 1.01881695e+00 4.78517354e-01
-7.28912950e-02 -1.01682198e+00 1.55709803e-01 3.79874676e-01
6.21149957e-01 -1.21116388e+00 -7.03591704e-02 -5.16385555e-01
-6.62643790e-01 1.44482267e+00 5.57847857e-01 -8.62327293e-02
2.31425911e-01 4.58031774e-01 5.28878253e-03 -1.00356586e-01
-5.50525308e-01 -3.44344109e-01 3.48504454e-01 2.01403394e-01
4.00511622e-01 -3.05489134e-02 1.44238666e-01 5.28947830e-01
2.09809393e-01 1.34677151e-02 2.85672128e-01 8.89097631e-01
-8.03594887e-01 -6.00553334e-01 -6.79595709e-01 7.76356995e-01
-3.51375490e-01 -3.37046295e-01 -5.26012778e-01 5.60114145e-01
-2.48482078e-01 8.08207691e-01 1.04553938e-01 -4.18348342e-01
2.79716909e-01 6.01596721e-02 2.86633354e-02 -3.46079588e-01
-6.51663482e-01 3.22955579e-01 -3.73487383e-01 -4.99421000e-01
9.15592462e-02 -4.64112163e-01 -1.35107529e+00 -4.60389167e-01
-7.38240063e-01 -2.56224483e-01 5.07533550e-01 6.31872118e-01
5.66162646e-01 4.40304220e-01 5.66758156e-01 -9.98366535e-01
-2.88464218e-01 -9.13848698e-01 -7.51918912e-01 1.09677128e-01
2.65853196e-01 -2.46301234e-01 -2.73699254e-01 2.87644237e-01] | [12.040014266967773, 2.181122064590454] |
0cc5d1cb-884a-496f-8cb0-7b77cc3ee081 | siamese-masked-autoencoders | 2305.14344 | null | https://arxiv.org/abs/2305.14344v1 | https://arxiv.org/pdf/2305.14344v1.pdf | Siamese Masked Autoencoders | Establishing correspondence between images or scenes is a significant challenge in computer vision, especially given occlusions, viewpoint changes, and varying object appearances. In this paper, we present Siamese Masked Autoencoders (SiamMAE), a simple extension of Masked Autoencoders (MAE) for learning visual correspondence from videos. SiamMAE operates on pairs of randomly sampled video frames and asymmetrically masks them. These frames are processed independently by an encoder network, and a decoder composed of a sequence of cross-attention layers is tasked with predicting the missing patches in the future frame. By masking a large fraction ($95\%$) of patches in the future frame while leaving the past frame unchanged, SiamMAE encourages the network to focus on object motion and learn object-centric representations. Despite its conceptual simplicity, features learned via SiamMAE outperform state-of-the-art self-supervised methods on video object segmentation, pose keypoint propagation, and semantic part propagation tasks. SiamMAE achieves competitive results without relying on data augmentation, handcrafted tracking-based pretext tasks, or other techniques to prevent representational collapse. | ['Li Fei-Fei', 'Jia Deng', 'Jiajun Wu', 'Agrim Gupta'] | 2023-05-23 | null | null | null | null | ['video-object-segmentation', 'video-semantic-segmentation'] | ['computer-vision', 'computer-vision'] | [ 2.78716356e-01 2.34390974e-01 -1.85530782e-01 -3.38953376e-01
-3.63395661e-01 -2.64342457e-01 5.19034207e-01 -3.11779886e-01
-4.59308535e-01 4.27710146e-01 1.15120821e-01 9.30428579e-02
3.38770360e-01 -5.87398410e-01 -1.48029768e+00 -5.52000403e-01
-8.78716186e-02 4.95309711e-01 5.60274303e-01 -1.04448654e-01
-1.33958414e-01 3.73459399e-01 -1.45773399e+00 5.65085351e-01
3.82162780e-01 9.69482422e-01 3.51853639e-01 7.60319412e-01
-3.76449227e-02 1.18013573e+00 -3.87600541e-01 -4.63998795e-01
4.10891742e-01 -6.21747434e-01 -9.62053537e-01 5.65924466e-01
9.86657262e-01 -7.38704562e-01 -8.60853851e-01 9.48725879e-01
-1.14777088e-01 3.33313763e-01 5.98225713e-01 -1.51003420e+00
-8.99775267e-01 6.50965571e-01 -6.40986204e-01 3.41121733e-01
7.55864605e-02 2.82011420e-01 8.72803509e-01 -1.05092216e+00
8.96467328e-01 1.29190421e+00 6.44761324e-01 6.70894921e-01
-1.28088593e+00 -4.64063227e-01 5.87567985e-01 2.85799772e-01
-1.19345343e+00 -4.64484602e-01 8.94334733e-01 -6.09217227e-01
8.99255991e-01 -2.33401265e-02 7.39468157e-01 1.08438694e+00
4.48985351e-03 1.10567832e+00 4.12503809e-01 -2.15811282e-01
1.76712185e-01 -8.02511424e-02 -1.44014254e-01 8.94097567e-01
-5.80742732e-02 1.82965979e-01 -4.98684198e-01 2.07442686e-01
1.03786325e+00 2.87491709e-01 -4.30924803e-01 -6.12139404e-01
-1.39636922e+00 7.44285464e-01 7.15474546e-01 1.10300384e-01
-5.50873518e-01 3.76218319e-01 5.54653294e-02 2.09904328e-01
3.96303564e-01 2.99250275e-01 -5.47319233e-01 3.89580458e-01
-1.20264757e+00 2.62959003e-01 3.84740561e-01 1.05463290e+00
9.40585434e-01 3.41580629e-01 -2.05986127e-01 4.94123816e-01
3.04464787e-01 2.69020587e-01 2.91719913e-01 -1.30231786e+00
2.37235561e-01 5.12510598e-01 1.70735121e-01 -1.01789653e+00
-1.02311410e-01 -2.73591816e-01 -6.43692732e-01 3.23866218e-01
4.59887296e-01 8.55348259e-03 -1.28925145e+00 1.72668278e+00
2.86593199e-01 3.84313166e-01 1.17287431e-02 1.16854692e+00
7.90075123e-01 8.41579318e-01 5.15962765e-02 5.46735749e-02
1.05645263e+00 -1.45943177e+00 -6.33072376e-01 -5.93820512e-01
-3.94407026e-02 -7.31936455e-01 6.08840227e-01 1.06363960e-01
-1.47061265e+00 -9.19035137e-01 -9.14464474e-01 -4.22920197e-01
-2.51664251e-01 8.41289535e-02 4.95810121e-01 -3.98898721e-02
-1.07770324e+00 6.90399170e-01 -1.20102704e+00 -2.69442767e-01
8.41837645e-01 3.73857290e-01 -4.25992608e-01 -2.53567308e-01
-7.60716259e-01 6.19237721e-01 2.74988860e-01 2.40750417e-01
-1.22796750e+00 -7.32150912e-01 -1.21929014e+00 4.48925532e-02
4.19284791e-01 -8.21284890e-01 1.16056907e+00 -1.54110467e+00
-1.29419637e+00 8.84564996e-01 -1.93699419e-01 -7.73402929e-01
6.89876020e-01 -3.87539327e-01 -3.17008257e-01 4.18150216e-01
2.21747175e-01 1.58333325e+00 1.31371677e+00 -1.22852862e+00
-6.87076449e-01 -1.36191294e-01 3.25654633e-02 9.53641087e-02
4.41641659e-02 -2.07835506e-03 -9.29666936e-01 -8.47878695e-01
5.59985787e-02 -9.47158337e-01 -2.75722593e-01 5.55529594e-01
-2.68592954e-01 -8.89048055e-02 1.19598567e+00 -7.54881620e-01
6.34373426e-01 -2.22930574e+00 3.77935708e-01 -2.51808554e-01
2.88515061e-01 2.00408921e-01 -4.21687275e-01 -9.91085544e-02
-1.65727258e-01 -2.80309021e-01 -5.11714399e-01 -4.51284766e-01
-2.65302658e-01 4.23901856e-01 -2.60540813e-01 6.16906285e-01
6.84089363e-01 1.09175003e+00 -9.25405264e-01 -4.63392198e-01
4.67814893e-01 5.56914508e-01 -7.84764230e-01 4.48147953e-01
-7.10296750e-01 4.33005869e-01 4.90014963e-02 6.51086688e-01
5.92613757e-01 -5.62332630e-01 -1.93337083e-01 -3.95057678e-01
6.64064288e-02 -2.59407535e-02 -1.01072097e+00 1.89185727e+00
-1.09179929e-01 1.02468300e+00 1.01965830e-01 -1.10818779e+00
5.47405660e-01 3.53322402e-02 6.32199049e-01 -4.82361436e-01
3.01478773e-01 -2.04310089e-01 -1.79191530e-02 -5.28762817e-01
3.87896776e-01 3.36288303e-01 3.48958820e-01 3.71965468e-01
4.22657520e-01 7.42396563e-02 3.04684103e-01 3.00588429e-01
7.20000148e-01 4.87434804e-01 -4.22005057e-02 2.97957622e-02
3.68916124e-01 1.99090242e-01 6.60238504e-01 6.50597811e-01
-3.23285788e-01 1.09799957e+00 2.17811361e-01 -7.02314258e-01
-1.30065548e+00 -9.90179598e-01 2.38292679e-01 1.16669047e+00
3.87685299e-01 -1.17945231e-01 -7.12756217e-01 -8.67568135e-01
1.13143235e-01 4.67095494e-01 -9.48819101e-01 -1.66420326e-01
-8.10631096e-01 -3.02646160e-01 7.24028870e-02 9.10933733e-01
5.80808163e-01 -1.18531787e+00 -5.66650093e-01 3.23882729e-01
-2.09582478e-01 -1.43290019e+00 -9.00460064e-01 1.63894355e-01
-6.88254416e-01 -1.20216203e+00 -9.36689436e-01 -1.09299052e+00
8.91916990e-01 4.64632660e-01 1.24821401e+00 -9.00390744e-03
-3.39169085e-01 3.79566878e-01 -2.48721704e-01 -1.10227160e-01
-4.20376331e-01 -2.69119740e-01 -2.26681098e-01 2.85757214e-01
3.61609370e-01 -3.20061415e-01 -8.26596260e-01 3.83939803e-01
-9.40208197e-01 8.46278816e-02 5.76840997e-01 8.74886930e-01
8.00117075e-01 -1.62878767e-01 4.01641279e-02 -7.19549060e-01
-2.86151618e-01 -4.18972492e-01 -6.33960247e-01 1.65846437e-01
-1.03527904e-01 9.97338369e-02 4.45942849e-01 -6.00796044e-01
-7.48071253e-01 4.25673306e-01 5.86445779e-02 -1.26796055e+00
-2.40711138e-01 -1.38461679e-01 -9.26682651e-02 5.19258492e-02
4.70852375e-01 3.36048901e-01 3.19206715e-01 -3.61989290e-01
5.98889410e-01 1.90219432e-01 1.08086360e+00 -1.21970855e-01
7.33402014e-01 7.67047763e-01 -4.93309319e-01 -5.55253386e-01
-8.70646179e-01 -3.37164938e-01 -8.71212244e-01 -2.64976084e-01
1.44283795e+00 -1.06907940e+00 -4.18679595e-01 4.14182037e-01
-1.30565023e+00 -5.03107548e-01 -5.92173576e-01 4.85996485e-01
-5.98859012e-01 4.28812057e-02 -5.91753662e-01 -2.43441209e-01
-9.94334146e-02 -1.09370804e+00 1.08218777e+00 3.59746069e-01
-2.05120459e-01 -7.19120383e-01 -3.37048769e-01 4.67998713e-01
2.33515412e-01 1.16350025e-01 5.85942328e-01 -4.06650841e-01
-9.63805318e-01 -4.23871763e-02 -2.22324207e-01 5.31541526e-01
1.35747924e-01 2.13336766e-01 -8.43061447e-01 -3.30928653e-01
-2.74770886e-01 -2.69898534e-01 1.11039507e+00 6.99615657e-01
1.36185968e+00 -5.64691186e-01 -3.54199171e-01 8.20818007e-01
1.04698098e+00 2.19168171e-01 4.79980916e-01 1.64366513e-01
8.41973841e-01 5.63013256e-01 2.90808737e-01 1.49716094e-01
4.05326247e-01 4.76812273e-01 7.38974273e-01 -3.78453374e-01
-5.90046525e-01 -3.46651256e-01 3.06329370e-01 3.46251518e-01
1.61905587e-01 -2.20042378e-01 -5.63338757e-01 7.06746221e-01
-1.84159803e+00 -1.07824922e+00 9.92253274e-02 1.82492113e+00
6.84108078e-01 1.04016416e-01 2.18725011e-01 -2.43193209e-01
9.75828469e-01 2.99762636e-01 -8.48175824e-01 9.46085900e-02
-2.87059158e-01 7.58654252e-02 4.47522670e-01 5.96873403e-01
-1.45999694e+00 1.20376444e+00 5.99570656e+00 2.36654207e-01
-1.16194975e+00 1.04088090e-01 7.77741373e-01 -1.46893442e-01
-3.18007357e-02 -3.26428488e-02 -5.23051143e-01 4.48951751e-01
3.82099569e-01 3.39705467e-01 3.87368023e-01 9.36640382e-01
-1.26602113e-01 4.03638743e-02 -1.41102433e+00 8.51907492e-01
3.68727922e-01 -1.64955616e+00 3.37648317e-02 -3.25476021e-01
1.01000690e+00 1.79472357e-01 1.49141341e-01 1.84993982e-01
4.27009165e-01 -9.36646044e-01 9.66144204e-01 4.02263165e-01
5.21046758e-01 -4.17799801e-01 5.58899522e-01 -1.24418912e-02
-9.42713380e-01 -6.16506524e-02 -3.35151732e-01 6.78387880e-02
2.47698471e-01 1.45811737e-01 -5.95308959e-01 1.16901591e-01
9.42100406e-01 1.01840711e+00 -5.36582172e-01 9.74031270e-01
-1.05479278e-01 4.70571131e-01 -2.73437411e-01 3.89055938e-01
4.19064701e-01 6.54768618e-03 4.77564454e-01 1.14334285e+00
-1.58805072e-01 1.04400717e-01 3.24170291e-01 1.21804118e+00
-2.82598078e-01 -4.66081142e-01 -3.37888747e-01 -4.88665588e-02
2.89938271e-01 9.21380639e-01 -7.87880301e-01 -6.29780233e-01
-7.40120530e-01 1.51577699e+00 6.98240325e-02 6.21784508e-01
-9.11494970e-01 -6.41491786e-02 7.87043095e-01 2.89464593e-01
8.63939822e-01 -2.10407659e-01 1.13932312e-01 -1.18098259e+00
-1.02784671e-01 -7.62773097e-01 4.55290228e-01 -9.13473368e-01
-1.31213415e+00 7.23929465e-01 -1.58153325e-01 -1.36065161e+00
-3.65311712e-01 -4.80413765e-01 -6.44548297e-01 4.41192478e-01
-1.43435860e+00 -1.07747471e+00 -4.41367984e-01 7.90888131e-01
1.03050995e+00 -1.08276680e-01 3.98216516e-01 2.48087525e-01
-6.90143943e-01 4.50722307e-01 -1.65127918e-01 5.78450084e-01
5.07476032e-01 -1.09590662e+00 6.32829010e-01 1.05888581e+00
4.91582394e-01 2.78655678e-01 6.87868178e-01 -5.78569770e-01
-1.16373920e+00 -1.53934991e+00 6.74222589e-01 -3.27909797e-01
4.80766386e-01 -4.33341652e-01 -1.07687628e+00 1.02610636e+00
4.46073025e-01 7.59091794e-01 2.34650880e-01 -4.41664129e-01
-3.53797317e-01 -9.03552473e-02 -7.98043430e-01 5.95362902e-01
1.08212471e+00 -4.77651060e-01 -7.53544629e-01 4.67494726e-01
9.14397061e-01 -6.60501659e-01 -5.66564322e-01 3.22292536e-01
3.79616529e-01 -8.60539138e-01 1.13955307e+00 -9.42358077e-01
6.76857531e-01 -4.26020026e-01 -9.02188942e-02 -9.20786262e-01
-3.32232058e-01 -5.69655120e-01 -3.78267467e-01 8.65101635e-01
1.37231275e-01 -3.57671501e-03 1.18537784e+00 6.18930280e-01
-1.97141305e-01 -6.31590366e-01 -4.86107945e-01 -6.33613527e-01
-8.35778117e-02 -1.91987634e-01 3.65956187e-01 8.96706164e-01
-7.14033902e-01 1.68546572e-01 -5.38474262e-01 3.79637301e-01
7.06300199e-01 1.90465420e-01 8.53154898e-01 -7.65232444e-01
-4.51323211e-01 -4.14775968e-01 -5.27256429e-01 -1.45772862e+00
2.65555412e-01 -5.37243426e-01 1.87705189e-01 -1.32166028e+00
-2.29843222e-02 -7.12237973e-03 -1.89215645e-01 6.50462925e-01
-2.53554821e-01 5.05712688e-01 2.88354754e-01 3.15059036e-01
-7.86451459e-01 6.97826087e-01 1.27444196e+00 -5.34705520e-01
-2.60799736e-01 -1.40109751e-02 -2.82769471e-01 8.75275493e-01
4.02862310e-01 -4.77874607e-01 -3.22089344e-01 -8.50120962e-01
-3.18681389e-01 3.54747511e-02 7.41565049e-01 -9.30650771e-01
4.22769368e-01 -8.00801292e-02 8.88717175e-01 -7.78402507e-01
3.96128833e-01 -8.48301888e-01 3.46262120e-02 3.88112545e-01
-4.08937067e-01 2.23037183e-01 7.01190308e-02 7.77099490e-01
-4.28082258e-01 3.17060724e-02 8.16383183e-01 -2.86452591e-01
-1.09472013e+00 5.53810835e-01 -3.37635279e-01 7.47434199e-02
1.13370693e+00 -4.32072729e-01 -3.42528671e-02 -2.60088861e-01
-9.96301234e-01 2.81292558e-01 4.92698997e-01 7.30591714e-01
7.88630784e-01 -1.05811751e+00 -6.54716134e-01 3.79806966e-01
-5.34848720e-02 4.60866749e-01 4.63662148e-01 6.38312340e-01
-5.72223425e-01 2.44953111e-01 -4.03689384e-01 -8.34208727e-01
-1.03497934e+00 7.72067487e-01 4.79108959e-01 4.05366957e-01
-8.61929238e-01 1.32961214e+00 4.46400911e-01 -1.33424858e-02
5.22089303e-01 -3.33439827e-01 5.95756900e-03 -6.71510771e-02
5.09019673e-01 9.63546187e-02 -9.17172730e-02 -7.35395253e-01
-3.77198070e-01 4.93951678e-01 -4.33639079e-01 1.85714532e-02
1.38558972e+00 -1.51734456e-01 -6.51824242e-03 1.60128042e-01
1.36800528e+00 -3.43905807e-01 -2.13075876e+00 -4.89022553e-01
-3.93846899e-01 -5.16026318e-01 9.65146348e-02 -5.78404009e-01
-1.59928977e+00 7.75441825e-01 5.02190888e-01 -3.63008946e-01
1.04281461e+00 3.91332358e-01 8.59508753e-01 1.73612177e-01
-6.32319376e-02 -8.35444212e-01 4.67418402e-01 2.42157772e-01
8.33598912e-01 -1.52287745e+00 -1.27358317e-01 -2.37058431e-01
-8.42032909e-01 1.06294608e+00 9.82293904e-01 -3.44483197e-01
6.46637738e-01 2.37249419e-01 7.74261430e-02 -1.43626496e-01
-7.90667534e-01 -3.63026232e-01 3.62145156e-01 6.81016862e-01
5.26638031e-02 -3.81365597e-01 3.44464719e-01 3.77161980e-01
-2.33553983e-02 -2.00844750e-01 3.65227699e-01 9.21454370e-01
-2.92141497e-01 -7.19325066e-01 -2.86660016e-01 2.77527183e-01
-2.93667197e-01 -1.93309300e-02 -3.74499977e-01 8.79993379e-01
3.54255706e-01 7.17134118e-01 7.36750126e-01 -1.77397400e-01
-4.23604343e-03 -9.12914872e-02 5.01495481e-01 -4.22404855e-01
-3.55994284e-01 2.37680167e-01 -4.12958443e-01 -7.78232276e-01
-6.09011412e-01 -8.32148790e-01 -1.29899430e+00 -6.96266219e-02
-1.66256234e-01 -1.27340958e-01 4.42470431e-01 9.25963581e-01
4.42825317e-01 7.22891629e-01 4.58483070e-01 -1.00460970e+00
-1.18184969e-01 -6.89179897e-01 -1.56634718e-01 6.14270926e-01
8.00864041e-01 -5.69869816e-01 -8.74234065e-02 6.97321117e-01] | [9.166678428649902, -0.01861005276441574] |
62361ac0-4123-487a-ba0c-505646833619 | label-inference-attack-against-split-learning | 2301.07284 | null | https://arxiv.org/abs/2301.07284v2 | https://arxiv.org/pdf/2301.07284v2.pdf | Label Inference Attack against Split Learning under Regression Setting | As a crucial building block in vertical Federated Learning (vFL), Split Learning (SL) has demonstrated its practice in the two-party model training collaboration, where one party holds the features of data samples and another party holds the corresponding labels. Such method is claimed to be private considering the shared information is only the embedding vectors and gradients instead of private raw data and labels. However, some recent works have shown that the private labels could be leaked by the gradients. These existing attack only works under the classification setting where the private labels are discrete. In this work, we step further to study the leakage in the scenario of the regression model, where the private labels are continuous numbers (instead of discrete labels in classification). This makes previous attacks harder to infer the continuous labels due to the unbounded output range. To address the limitation, we propose a novel learning-based attack that integrates gradient information and extra learning regularization objectives in aspects of model training properties, which can infer the labels under regression settings effectively. The comprehensive experiments on various datasets and models have demonstrated the effectiveness of our proposed attack. We hope our work can pave the way for future analyses that make the vFL framework more secure. | ['Jiankai Sun', 'Taiqing Wang', 'Tianyi Liu', 'Yuanshun Yao', 'Xin Yang', 'Shangyu Xie'] | 2023-01-18 | null | null | null | null | ['inference-attack'] | ['adversarial'] | [ 7.30352178e-02 6.00805394e-02 -5.56888878e-01 -3.06471854e-01
-1.01270270e+00 -1.14989769e+00 3.62868011e-01 1.59082174e-01
-2.72468895e-01 7.90325046e-01 -2.20501199e-01 -7.57963955e-01
-7.06589743e-02 -9.15751815e-01 -9.03339565e-01 -1.09070432e+00
-3.24798971e-01 -1.63292177e-02 -6.51605874e-02 9.36905667e-02
1.15199611e-02 5.70861220e-01 -1.03180420e+00 4.26335990e-01
3.69480044e-01 1.13378322e+00 -7.83820927e-01 4.22220796e-01
8.67082626e-02 7.16583014e-01 -4.31495696e-01 -9.25959289e-01
9.28324699e-01 -9.39821377e-02 -7.66705334e-01 -4.45911109e-01
4.91678596e-01 -4.75378692e-01 -3.38435173e-01 1.26281154e+00
4.37550813e-01 -3.51928383e-01 3.08883041e-01 -1.73224747e+00
-4.37550157e-01 8.11540961e-01 -5.17234147e-01 -2.79793411e-01
-1.96686894e-01 1.92879379e-01 1.10425031e+00 -4.88420218e-01
4.43490416e-01 1.08405292e+00 6.33063495e-01 8.51261079e-01
-1.07363927e+00 -1.23666227e+00 1.75392777e-01 2.04370350e-01
-1.27767360e+00 -2.06901431e-01 7.99605072e-01 -3.74773771e-01
2.92158455e-01 6.94769263e-01 1.09147444e-01 1.32920337e+00
6.66282373e-03 8.22909296e-01 1.53679717e+00 -3.19813251e-01
2.03452051e-01 8.25521827e-01 4.56294835e-01 6.63496792e-01
4.83443022e-01 3.95234644e-01 -4.31642950e-01 -8.96644890e-01
7.58675039e-02 2.43465632e-01 -4.72137064e-01 -5.75196862e-01
-8.32326472e-01 1.22155344e+00 3.01369697e-01 -2.98446100e-02
3.36261362e-01 1.57065809e-01 7.50944555e-01 5.74357092e-01
4.58654046e-01 1.13409467e-01 -8.74388456e-01 3.60163063e-01
-6.40360475e-01 4.29488644e-02 9.36345816e-01 7.84004271e-01
7.67382681e-01 -3.17346841e-01 -7.30479583e-02 1.04881771e-01
3.16904575e-01 3.89693975e-01 1.79948986e-01 -8.28223526e-01
9.07118499e-01 3.03737789e-01 2.72176057e-01 -9.02538776e-01
6.50024861e-02 -3.75898093e-01 -7.74406850e-01 3.34426701e-01
8.15088034e-01 -5.17068624e-01 -3.31928372e-01 2.13142037e+00
5.84170640e-01 1.75786123e-01 3.96328896e-01 6.68037474e-01
2.52974272e-01 3.76621842e-01 -1.07868508e-01 -4.05553170e-02
1.33523965e+00 -1.05830634e+00 -6.56858146e-01 2.96099335e-01
1.16208363e+00 -4.53223348e-01 9.95066881e-01 3.75762045e-01
-6.73967183e-01 8.37094150e-03 -1.36024833e+00 -5.37093282e-02
-6.51461363e-01 6.19593542e-03 8.88998628e-01 1.34263229e+00
-6.32324159e-01 6.19328856e-01 -8.46511185e-01 9.00101438e-02
8.38838458e-01 5.27022183e-01 -6.78219676e-01 1.59124702e-01
-1.48497021e+00 3.35253716e-01 4.80409637e-02 4.05511670e-02
-8.91420841e-01 -7.12499797e-01 -7.43527353e-01 1.75153911e-01
2.88316458e-01 -2.39791363e-01 9.31442142e-01 -5.60123682e-01
-1.28331792e+00 9.05877233e-01 1.85557604e-01 -6.58858657e-01
9.58092034e-01 1.80840313e-01 -1.95915699e-01 6.91690147e-02
-2.29143679e-01 -6.29779175e-02 8.88966739e-01 -1.33140337e+00
-7.38956153e-01 -8.70328665e-01 3.70549589e-01 -4.02028739e-01
-8.58753502e-01 -7.60717988e-02 2.76778340e-01 -1.98007956e-01
-2.75962830e-01 -9.37341332e-01 -1.17291078e-01 1.26468241e-01
-6.66558981e-01 -9.87065211e-02 1.10082924e+00 -5.87952852e-01
1.16377425e+00 -2.48303819e+00 -3.63987744e-01 4.60609019e-01
2.72614866e-01 2.78352022e-01 2.06701800e-01 5.25402427e-01
6.24603629e-02 6.18460536e-01 -1.07286014e-01 -5.80303073e-01
2.85762012e-01 -6.75949976e-02 -9.53523099e-01 1.05942762e+00
-1.23259209e-01 6.57943130e-01 -5.77617288e-01 -5.41409850e-01
-2.80198663e-01 4.79461819e-01 -4.16854769e-01 5.19548059e-02
9.55413133e-02 5.31420946e-01 -6.72894299e-01 5.60766876e-01
1.22189724e+00 -2.08715335e-01 3.85784775e-01 -1.37803704e-01
1.84500352e-01 2.22982958e-01 -1.22975194e+00 1.30831409e+00
-5.13918996e-01 2.17480123e-01 4.20322835e-01 -7.59760261e-01
6.38365805e-01 4.56005752e-01 2.02581301e-01 1.41696176e-02
1.58448875e-01 1.05241224e-01 -2.67886817e-01 -3.76651257e-01
-8.99569169e-02 -2.54911333e-01 -2.31824860e-01 7.84429729e-01
-2.68871695e-01 2.84145325e-01 -5.97308457e-01 2.51141876e-01
9.86584365e-01 3.91865931e-02 -1.32976279e-01 -1.77118145e-02
7.17816949e-01 -4.03138399e-01 7.22203791e-01 7.17936099e-01
-4.11854386e-01 1.07427754e-01 8.64012480e-01 -4.39467460e-01
-7.71684587e-01 -6.70398235e-01 -3.95705104e-01 9.55259919e-01
1.73237070e-01 -7.65050352e-01 -6.16405308e-01 -1.53908920e+00
3.36706221e-01 4.42608386e-01 -7.84012198e-01 -2.07366720e-01
-3.45411688e-01 -6.66952968e-01 1.08292890e+00 3.58500302e-01
3.68473709e-01 -3.76573920e-01 -3.50461394e-01 -4.63938385e-01
4.31998111e-02 -1.00234413e+00 -4.68310386e-01 4.35153484e-01
-7.31726885e-01 -1.30640483e+00 -1.93597749e-02 -5.22857010e-01
8.03109586e-01 -1.92747582e-02 2.51485705e-01 1.49892986e-01
-1.89401433e-02 1.27283394e-01 -1.84320321e-03 -3.64412963e-01
-2.92497486e-01 1.67662665e-01 -9.20948535e-02 5.57381392e-01
3.01897973e-01 -4.36979383e-01 -5.78519225e-01 2.89387673e-01
-9.17414725e-01 -2.01482221e-01 6.11927873e-03 1.06971860e+00
2.17164531e-01 1.51616246e-01 4.56527531e-01 -1.52470744e+00
3.86526406e-01 -6.11455679e-01 -8.22301149e-01 5.18974781e-01
-6.57657862e-01 1.24955431e-01 1.08705080e+00 -5.35608888e-01
-8.00940037e-01 -8.35986212e-02 7.88568333e-03 -4.52575594e-01
4.75168452e-02 -3.45997997e-02 -4.95855600e-01 -3.98670316e-01
4.52645600e-01 -1.21461429e-01 -4.20156214e-03 -6.25596464e-01
3.59641433e-01 1.01838195e+00 -1.65750027e-01 -7.98295498e-01
1.01109898e+00 9.24533248e-01 4.56149280e-01 6.65257797e-02
-8.14150333e-01 -3.97006050e-02 -1.53648540e-01 3.65347415e-01
4.68550444e-01 -7.24337637e-01 -1.39411962e+00 3.78142565e-01
-8.72213960e-01 -1.88428923e-01 -3.20129007e-01 4.61179912e-01
-3.31670076e-01 7.44892836e-01 -9.41814959e-01 -1.03102708e+00
-4.73071963e-01 -1.41647077e+00 8.10411096e-01 -2.87444368e-02
3.20593894e-01 -1.15192270e+00 -2.75009930e-01 5.68316042e-01
3.65711540e-01 3.93288612e-01 1.12616599e+00 -1.06862140e+00
-5.42783260e-01 -6.63212180e-01 -8.46001282e-02 5.77665865e-01
4.04560752e-02 -1.17756702e-01 -1.35967863e+00 -8.16875756e-01
5.08214712e-01 -7.22305536e-01 8.36602628e-01 -3.99890363e-01
1.55144477e+00 -7.28243887e-01 -2.79059500e-01 1.00451314e+00
1.53889298e+00 -3.07721227e-01 2.65093327e-01 1.71902686e-01
6.91086233e-01 7.51762450e-01 4.30595934e-01 5.61317861e-01
8.82838666e-02 4.81034905e-01 6.42459691e-01 -1.16267845e-01
2.92897135e-01 -5.33068061e-01 3.84155452e-01 4.37564075e-01
3.62573951e-01 5.04543260e-02 -4.28127140e-01 8.62832516e-02
-1.65551686e+00 -8.60630333e-01 -4.77645248e-02 2.43225646e+00
1.02576065e+00 -5.83964474e-02 -2.69122452e-01 1.83947891e-01
6.49076462e-01 2.21768409e-01 -5.37254393e-01 -6.20001078e-01
-1.28333285e-01 2.40789041e-01 1.03486848e+00 5.94925523e-01
-1.21420670e+00 6.42646492e-01 5.27398109e+00 1.03129208e+00
-1.28508258e+00 5.76811373e-01 7.94448495e-01 -9.22356918e-02
-4.00115848e-01 4.01507616e-01 -1.01495838e+00 5.20514727e-01
8.75318885e-01 -9.29795429e-02 4.23985064e-01 9.66139495e-01
-3.75786662e-01 4.70617175e-01 -1.29801738e+00 7.64535248e-01
-2.46546507e-01 -1.15841258e+00 -2.14846939e-01 4.62485015e-01
5.53270102e-01 -2.45451972e-01 5.12083113e-01 3.72666389e-01
4.16877508e-01 -9.90060747e-01 4.37020034e-01 1.33347541e-01
1.05281317e+00 -9.55756128e-01 6.53163016e-01 6.50154173e-01
-8.11163902e-01 -5.26047528e-01 -4.63286400e-01 8.91418010e-02
-3.89581025e-01 2.30503395e-01 -7.09957421e-01 8.22698534e-01
5.10479331e-01 4.07607019e-01 -4.82538968e-01 4.52312171e-01
-4.21462893e-01 9.65571463e-01 -3.82770032e-01 2.94633150e-01
1.59773797e-01 -1.82107210e-01 2.03445941e-01 1.05003834e+00
5.92082506e-03 -3.49753261e-01 1.77072346e-01 7.67693579e-01
-5.08437276e-01 3.92483532e-01 -6.71460867e-01 1.31235152e-01
3.81140202e-01 1.32681787e+00 -3.34973812e-01 -1.22011691e-01
-4.83139217e-01 6.39127791e-01 4.74902928e-01 4.33608919e-01
-7.97633469e-01 -3.21386963e-01 7.91411459e-01 -1.48844883e-01
1.69743568e-01 4.22515161e-02 -4.81059909e-01 -1.45686305e+00
1.47707418e-01 -9.16668475e-01 7.33901083e-01 2.31267095e-01
-1.64317131e+00 3.24257463e-01 -2.34231010e-01 -1.12782598e+00
1.24954112e-01 -4.18772906e-01 -3.77822936e-01 1.00342178e+00
-1.70601690e+00 -1.49019611e+00 2.95324594e-01 9.17097747e-01
-1.55578151e-01 -1.20008074e-01 1.13901806e+00 2.72032022e-01
-6.30831420e-01 1.41442037e+00 4.48138058e-01 4.17916417e-01
8.49829316e-01 -9.82088506e-01 3.08097620e-02 7.36498058e-01
2.03886077e-01 8.36650968e-01 3.08047205e-01 -4.64315087e-01
-1.67029417e+00 -9.33180034e-01 7.23655641e-01 -4.91797984e-01
7.03254104e-01 -9.40396607e-01 -6.98498547e-01 1.02451789e+00
6.23756973e-03 5.80183923e-01 1.19677091e+00 -5.52730411e-02
-9.80320930e-01 -3.86123151e-01 -1.72074556e+00 2.37695798e-01
6.69046938e-01 -8.08290064e-01 1.01265684e-01 5.69902241e-01
8.76733780e-01 -2.44943500e-01 -8.94830525e-01 4.65511113e-01
5.99298537e-01 -8.22822154e-01 6.31028235e-01 -1.07132113e+00
-7.42958784e-02 -2.45346695e-01 -5.05715728e-01 -7.93727040e-01
3.12565446e-01 -7.58271039e-01 -3.39232564e-01 1.52109194e+00
4.75646824e-01 -1.28301072e+00 8.99995148e-01 8.33205163e-01
7.37491429e-01 -9.18977976e-01 -1.24348164e+00 -6.24794722e-01
4.45038676e-01 -4.08019274e-01 1.07885969e+00 1.17225444e+00
1.19440164e-02 -3.08352649e-01 -7.26495981e-01 4.66260672e-01
8.35968018e-01 3.22996974e-01 8.29173028e-01 -9.64650035e-01
-5.38629711e-01 1.29854918e-01 -3.88700634e-01 -5.94188452e-01
6.64902985e-01 -1.16923463e+00 -7.50570774e-01 -6.09229207e-01
2.30296388e-01 -1.02158296e+00 -5.89467287e-01 7.26759851e-01
7.55854696e-02 1.69654638e-01 4.65023518e-01 2.10372508e-01
-2.10866630e-01 3.14578831e-01 1.12327754e+00 -4.36639011e-01
1.50588676e-01 4.94039148e-01 -9.19582665e-01 3.30315083e-01
9.51173246e-01 -7.65359044e-01 -4.31339443e-01 -8.28591082e-03
3.55196714e-01 -6.00087121e-02 3.76312643e-01 -4.84841228e-01
3.00244659e-01 -3.32413428e-02 1.45548984e-01 -2.00222373e-01
2.73059100e-01 -1.42316782e+00 9.61531773e-02 5.30897319e-01
-7.12423444e-01 -4.16389257e-01 -1.52389109e-01 6.43503249e-01
2.54979096e-02 -2.25134954e-01 6.53542221e-01 1.44016325e-01
1.56376064e-01 4.99456376e-01 2.94863731e-01 -8.88001248e-02
1.38730800e+00 1.45796344e-01 -5.84854186e-01 -2.50408202e-01
-7.70329118e-01 2.89828032e-01 6.03659868e-01 -1.32448766e-02
3.07860017e-01 -1.17829049e+00 -5.97619414e-01 5.68775773e-01
-8.88428017e-02 -1.96391016e-01 1.88861966e-01 9.52729285e-01
-1.56828701e-01 3.13344926e-01 7.37425163e-02 -1.82791308e-01
-1.47153509e+00 1.03946221e+00 3.82044673e-01 -5.27780294e-01
-6.17763758e-01 6.91495299e-01 2.25708067e-01 -6.05050683e-01
6.52290702e-01 -3.63372937e-02 1.71797723e-01 6.45271093e-02
5.27775705e-01 4.17938471e-01 9.63893607e-02 -4.57681447e-01
-2.03412294e-01 4.17026222e-01 -2.05325991e-01 -1.22706043e-02
1.10138774e+00 -1.09152488e-01 -2.86668748e-01 3.49182248e-01
1.69705558e+00 6.31148517e-01 -1.23386955e+00 -2.66930252e-01
-1.06249355e-01 -7.84798503e-01 -2.65213639e-01 -6.34837627e-01
-1.27462721e+00 1.26493692e+00 6.96315646e-01 2.85249203e-01
9.69468534e-01 -1.86221734e-01 9.45608795e-01 6.26727045e-02
8.05384219e-01 -7.40226567e-01 -5.35970390e-01 -4.12545465e-02
2.54907846e-01 -1.03252959e+00 -2.80944202e-02 -5.57190716e-01
-5.52588105e-01 1.03338873e+00 4.17579591e-01 1.68305799e-01
8.03810120e-01 4.74975020e-01 2.17008635e-01 1.83551982e-01
-8.56686354e-01 5.27073026e-01 -2.81084210e-01 3.95969987e-01
-8.26564208e-02 2.50374317e-01 -3.20743471e-01 8.96835685e-01
-4.30740230e-02 -2.23930955e-01 4.18187529e-01 9.47696924e-01
2.80384064e-01 -1.72583258e+00 -5.21813691e-01 2.23838657e-01
-1.17947948e+00 7.22683594e-02 -2.17336610e-01 5.68811893e-01
3.30679536e-01 9.27382946e-01 -5.66259980e-01 -4.41036910e-01
-4.30570990e-02 4.16238308e-01 2.23908797e-01 -2.81400174e-01
-9.82505977e-01 -3.17637980e-01 -1.27108946e-01 -4.60651934e-01
1.27758458e-02 -3.06097120e-01 -1.21740186e+00 -5.26888609e-01
-6.88771904e-01 4.96682674e-01 8.91717434e-01 4.73923743e-01
2.72086143e-01 -1.66610673e-01 1.43815255e+00 -3.18347454e-01
-1.59401059e+00 -5.29660761e-01 -9.45135295e-01 6.59936547e-01
5.87909281e-01 -3.02628636e-01 -9.94947493e-01 -2.63977706e-01] | [5.823479175567627, 6.782739162445068] |
c1e6916b-dc24-4590-bcbe-7a9b387b80a9 | sgaligner-3d-scene-alignment-with-scene | 2304.14880 | null | https://arxiv.org/abs/2304.14880v1 | https://arxiv.org/pdf/2304.14880v1.pdf | SGAligner : 3D Scene Alignment with Scene Graphs | Building 3D scene graphs has recently emerged as a topic in scene representation for several embodied AI applications to represent the world in a structured and rich manner. With their increased use in solving downstream tasks (eg, navigation and room rearrangement), can we leverage and recycle them for creating 3D maps of environments, a pivotal step in agent operation? We focus on the fundamental problem of aligning pairs of 3D scene graphs whose overlap can range from zero to partial and can contain arbitrary changes. We propose SGAligner, the first method for aligning pairs of 3D scene graphs that is robust to in-the-wild scenarios (ie, unknown overlap -- if any -- and changes in the environment). We get inspired by multi-modality knowledge graphs and use contrastive learning to learn a joint, multi-modal embedding space. We evaluate on the 3RScan dataset and further showcase that our method can be used for estimating the transformation between pairs of 3D scenes. Since benchmarks for these tasks are missing, we create them on this dataset. The code, benchmark, and trained models are available on the project website. | ['Iro Armeni', 'Daniel Barath', 'Marc Pollefeys', 'Ondrej Miksik', 'Sayan Deb Sarkar'] | 2023-04-28 | null | null | null | null | ['point-cloud-registration', '3d-scene-graph-alignment'] | ['computer-vision', 'computer-vision'] | [ 3.26393515e-01 8.38311091e-02 3.21248919e-01 -3.78930002e-01
-4.49788302e-01 -9.84647334e-01 9.90901887e-01 1.87375590e-01
-3.52199197e-01 4.54208493e-01 6.29818201e-01 -2.23215058e-01
-1.40501663e-01 -6.02283239e-01 -9.07995462e-01 -4.37433273e-01
-1.79177120e-01 6.64543748e-01 2.47372672e-01 -5.26584744e-01
1.72144383e-01 6.05177879e-01 -1.65244091e+00 1.47815078e-01
2.55662054e-01 5.55883110e-01 4.17678654e-01 6.97666466e-01
-5.12347259e-02 7.52497196e-01 -3.78139079e-01 3.23112532e-02
4.67203081e-01 -2.08192140e-01 -8.01941872e-01 1.32559419e-01
5.59685230e-01 -1.05722345e-01 -6.58590555e-01 8.28345239e-01
5.78160465e-01 5.54397881e-01 7.45909929e-01 -1.37045300e+00
-5.01716256e-01 3.12080055e-01 -2.04547346e-01 5.31223975e-03
9.48475599e-01 -2.99142748e-02 8.27746212e-01 -6.64766788e-01
1.12088275e+00 1.29159546e+00 5.51108837e-01 4.47229624e-01
-1.30135369e+00 -1.21699773e-01 3.35229099e-01 2.88293689e-01
-1.18555534e+00 -6.26322865e-01 8.06329608e-01 -5.23366928e-01
1.29361653e+00 3.93021494e-01 8.49081039e-01 1.29733121e+00
1.38409778e-01 5.76522648e-01 9.98190820e-01 -5.14448225e-01
3.23567212e-01 -9.77971554e-02 -1.21999390e-01 7.47007072e-01
-3.71163781e-03 2.92725358e-02 -6.77355170e-01 2.14764237e-01
7.76004851e-01 -6.46683201e-02 -2.05234274e-01 -1.18276763e+00
-1.63060415e+00 5.46214163e-01 6.67258501e-01 2.71414846e-01
-1.80845559e-01 3.97245973e-01 1.13094233e-01 2.25206092e-01
2.15372667e-01 6.52422726e-01 -1.29623845e-01 -1.97974712e-01
-3.18360507e-01 6.51333988e-01 6.74395740e-01 1.24483705e+00
7.95269489e-01 -1.87792376e-01 -1.92542579e-02 4.40655887e-01
5.17403632e-02 3.93991977e-01 2.67486662e-01 -1.08170402e+00
4.92101640e-01 6.01386666e-01 1.53756320e-01 -1.00118601e+00
-7.92967379e-01 -5.53502627e-02 -5.11282265e-01 2.57068783e-01
3.03885698e-01 2.10833907e-01 -1.08811378e+00 1.97473228e+00
5.81395268e-01 2.44959503e-01 1.25021175e-01 7.68398523e-01
9.99403358e-01 3.74898881e-01 -1.25097707e-01 3.85749042e-01
1.16425192e+00 -8.85884583e-01 -5.36023617e-01 -6.24123454e-01
6.98517621e-01 -4.76164639e-01 1.14028466e+00 -1.42019361e-01
-9.12193894e-01 -2.68332392e-01 -1.05099165e+00 -5.57816684e-01
-8.80204201e-01 -3.25299263e-01 8.29470575e-01 1.65242329e-01
-1.34054530e+00 4.37673748e-01 -8.93861651e-01 -9.82696533e-01
1.96233228e-01 2.05964178e-01 -1.05609715e+00 -1.22367792e-01
-7.85258532e-01 1.39698935e+00 4.40662265e-01 1.16828559e-02
-8.52080941e-01 -4.79432821e-01 -1.36119592e+00 -4.71288323e-01
3.59234333e-01 -1.01401627e+00 9.44501281e-01 -3.96093816e-01
-1.45869863e+00 1.07002974e+00 -2.49054860e-02 -9.38689634e-02
4.71969217e-01 -1.25877276e-01 -1.00511633e-01 -2.21147746e-01
2.41604045e-01 6.86065853e-01 6.46513164e-01 -1.34118664e+00
-3.26028258e-01 -6.82312608e-01 5.16038954e-01 7.34181583e-01
1.84263438e-01 -4.23593402e-01 -3.61304849e-01 -2.53832221e-01
4.97716963e-01 -1.12109113e+00 -4.13226187e-01 6.23834170e-02
-6.11483634e-01 2.03423858e-01 6.79483175e-01 -6.86499476e-01
5.14389992e-01 -2.11322522e+00 7.40436614e-01 1.92955375e-01
9.16549042e-02 -3.54479700e-01 -2.07235530e-01 4.67428714e-01
-1.23478994e-01 -1.50186718e-01 -3.90896320e-01 -5.72427094e-01
3.96624565e-01 3.34584832e-01 -1.20185159e-01 6.28388762e-01
-7.30303228e-02 1.03799093e+00 -1.13914371e+00 -1.36279494e-01
5.95110953e-01 6.24094307e-01 -5.59256494e-01 3.14441711e-01
-1.27268389e-01 7.36858726e-01 -3.62871438e-01 2.64931589e-01
3.21690112e-01 1.20438352e-01 3.75604853e-02 -4.28130031e-01
-8.19438919e-02 3.72521579e-01 -1.25714552e+00 2.68748045e+00
-5.13242364e-01 6.28391147e-01 -2.44176015e-01 -9.36146080e-01
6.66977346e-01 -3.35530639e-02 3.73787045e-01 -6.77892804e-01
7.31799901e-02 -1.84285507e-01 -3.15376163e-01 -5.61730027e-01
7.98935056e-01 5.42974211e-02 -6.19260371e-01 3.33655089e-01
1.65348247e-01 -1.00298023e+00 1.01303332e-01 1.51724547e-01
1.47286403e+00 7.33821332e-01 3.88911575e-01 -1.45118013e-01
8.26330408e-02 6.94687068e-02 -7.74512589e-02 5.80820441e-01
-1.39799982e-01 6.36005163e-01 3.46125722e-01 -4.19201314e-01
-1.15634024e+00 -1.40566730e+00 2.38334522e-01 8.45437646e-01
4.72288966e-01 -4.25165832e-01 -5.26540160e-01 -5.17738998e-01
-2.28294488e-02 1.02589953e+00 -7.52849519e-01 -2.28342310e-01
-4.56255108e-01 -2.40281388e-01 2.02963486e-01 4.57915425e-01
2.74592280e-01 -8.63230407e-01 -9.75437462e-01 -2.22146735e-02
-2.77372479e-01 -1.37150133e+00 -2.81776816e-01 4.78128910e-01
-4.88167048e-01 -1.06867063e+00 -2.88542628e-01 -8.19068730e-01
7.47541606e-01 4.39313620e-01 1.20865738e+00 -4.78817225e-01
-2.57911503e-01 1.12100828e+00 -3.47304136e-01 -2.52369702e-01
-2.69321501e-01 -1.12373888e-01 7.41855800e-02 -3.96783859e-01
-6.19356483e-02 -8.99557948e-01 -2.27105781e-01 5.50560988e-02
-8.05034697e-01 5.46840787e-01 1.36940673e-01 4.48215902e-01
6.56691194e-01 -2.78641790e-01 -1.42301163e-02 -7.85042942e-01
5.54661572e-01 -3.10000271e-01 -4.72694457e-01 2.73019522e-01
1.72581896e-01 4.26507115e-01 1.41111583e-01 -1.74510390e-01
-8.19950283e-01 2.03199938e-01 -2.02907361e-02 -4.19156998e-01
-4.79800016e-01 4.56889153e-01 -4.54796225e-01 5.60385548e-03
5.57590783e-01 -9.61102694e-02 -2.18931645e-01 -1.84612080e-01
1.01414335e+00 1.13568582e-01 7.19466031e-01 -5.30914485e-01
8.90920877e-01 6.65610075e-01 3.45374852e-01 -7.53609061e-01
-6.03481650e-01 -3.59740585e-01 -9.74018753e-01 -1.76583916e-01
1.16115677e+00 -9.05002952e-01 -4.63967651e-01 3.44678611e-01
-1.17441988e+00 -8.58301342e-01 -4.66997415e-01 3.62238556e-01
-1.04258108e+00 1.53011322e-01 1.58554316e-02 -4.43252385e-01
2.62014329e-01 -1.09479499e+00 1.32424104e+00 4.04479168e-02
-3.94621223e-01 -1.17328858e+00 3.93379271e-01 5.48173934e-02
1.62324950e-01 7.78037250e-01 1.02316272e+00 -4.40684319e-01
-5.83463848e-01 -1.06270783e-01 1.59275785e-01 -2.74957150e-01
4.01836753e-01 -3.39661628e-01 -1.01342404e+00 -1.86324239e-01
-2.79766977e-01 -1.43387124e-01 6.21196687e-01 1.29324242e-01
8.96227896e-01 -4.43965755e-02 -3.96631300e-01 8.03106964e-01
1.20609629e+00 -1.53789707e-02 6.33849561e-01 6.16733611e-01
9.97138023e-01 8.15162182e-01 3.68625045e-01 2.87886977e-01
8.60765696e-01 1.03669620e+00 7.25697041e-01 1.20622568e-01
-3.29150349e-01 -5.46862006e-01 2.87461311e-01 7.77333379e-01
-1.36934698e-01 -3.58511001e-01 -7.99044013e-01 5.14881313e-01
-1.87005615e+00 -1.01581180e+00 2.64765352e-01 2.22161889e+00
4.23366517e-01 -1.44010056e-02 -1.06763341e-01 -1.67888060e-01
5.56810558e-01 5.31583846e-01 -6.33955657e-01 -2.57907629e-01
-2.21952587e-01 2.58048713e-01 4.20249879e-01 6.67838633e-01
-1.26035976e+00 1.14228606e+00 5.86602640e+00 9.54856426e-02
-8.44785273e-01 2.33136006e-02 3.34330276e-02 -1.41428635e-01
-5.31983018e-01 1.30449772e-01 -3.17548633e-01 -5.05401008e-03
5.20150721e-01 5.59911132e-02 7.78714478e-01 6.07897460e-01
-5.09693213e-02 -1.20288968e-01 -1.54138446e+00 1.23669171e+00
2.76650548e-01 -1.21031523e+00 2.13298597e-03 -1.46586150e-01
5.94783723e-01 2.02635229e-01 -1.11987658e-01 2.74521679e-01
6.71403050e-01 -9.37235057e-01 1.19407904e+00 5.42681634e-01
6.57447219e-01 -3.80697787e-01 2.69673824e-01 1.80458412e-01
-1.28270745e+00 2.13580325e-01 -1.24363750e-01 -2.23022085e-02
4.37030166e-01 1.18254237e-01 -8.79541218e-01 7.52481163e-01
7.21814215e-01 9.73540068e-01 -6.82223797e-01 9.83130395e-01
-3.85474235e-01 -3.30144614e-01 -4.87330467e-01 8.33280087e-02
3.60700451e-02 -2.91677862e-01 8.18449199e-01 7.16481864e-01
6.39385164e-01 -9.80394781e-02 8.57561678e-02 8.69635582e-01
1.41652659e-01 -2.89397568e-01 -1.21785021e+00 6.85473382e-02
4.13022041e-01 1.09271526e+00 -7.84416020e-01 4.94179688e-02
-1.41183674e-01 1.24659026e+00 5.84863305e-01 4.55074340e-01
-8.63645017e-01 -2.86685169e-01 7.35183477e-01 -3.69579270e-02
2.94577569e-01 -7.46538639e-01 -1.16947018e-01 -1.15788257e+00
-1.48620950e-02 -5.48368394e-01 1.17296018e-01 -1.17894006e+00
-1.03013194e+00 5.16652703e-01 3.29686731e-01 -1.08136666e+00
-3.15228313e-01 -6.00727856e-01 -4.08189505e-01 5.02467394e-01
-1.39129162e+00 -1.53942442e+00 -7.17462003e-01 8.60057533e-01
2.82946289e-01 1.63872063e-01 1.04233336e+00 -1.09276418e-02
-1.88745037e-01 1.65619597e-01 -1.03446960e-01 -1.15466528e-01
5.63345492e-01 -1.42045236e+00 6.81573927e-01 8.28450978e-01
5.20620942e-01 3.26402813e-01 9.33867872e-01 -4.49905157e-01
-1.73029864e+00 -1.05651367e+00 5.83657384e-01 -1.02090967e+00
5.99187911e-01 -8.96182358e-01 -4.53200758e-01 1.16165423e+00
1.32557750e-01 -1.22623026e-01 4.95755285e-01 1.08620569e-01
-4.73457724e-01 1.48335725e-01 -9.86662626e-01 1.01548028e+00
1.80736458e+00 -6.21201992e-01 -6.80327058e-01 4.83520389e-01
8.43260407e-01 -1.02014756e+00 -6.44952238e-01 2.63004810e-01
4.74559277e-01 -9.82231736e-01 1.18991113e+00 -6.40574872e-01
1.50625169e-01 -4.34292346e-01 -6.60868287e-01 -1.74029517e+00
-2.71992594e-01 -5.73956609e-01 2.37120669e-02 8.95301640e-01
1.68539256e-01 -5.83765686e-01 6.32782042e-01 7.82568514e-01
-4.70164359e-01 -1.93000540e-01 -1.10740626e+00 -5.72180748e-01
-3.06652039e-01 -5.87965608e-01 8.18886936e-01 7.50983834e-01
-7.39386082e-02 3.12576860e-01 -1.20742783e-01 3.16186637e-01
5.90519190e-01 1.01856247e-01 1.27200937e+00 -9.34302926e-01
-8.78858641e-02 -3.13560426e-01 -9.37960565e-01 -9.86287653e-01
4.86270994e-01 -1.09722733e+00 2.65692830e-01 -1.96900022e+00
-1.57252669e-01 -5.48154712e-01 -5.09399734e-03 6.21337533e-01
1.52881056e-01 -9.34277549e-02 3.43679935e-01 -1.13199174e-01
-6.23566091e-01 8.11656058e-01 1.16389477e+00 -2.06619740e-01
-3.12073559e-01 -4.67763454e-01 -4.18102026e-01 6.23995066e-01
6.02232873e-01 -1.09427258e-01 -5.80500066e-01 -7.46051013e-01
3.55172336e-01 -3.26987624e-01 6.17862880e-01 -1.15011811e+00
1.32157430e-01 -2.33364746e-01 1.05123177e-01 -4.70195591e-01
9.60696161e-01 -1.11779368e+00 5.17927945e-01 -2.46998686e-02
-2.64299780e-01 3.37716550e-01 4.35636491e-01 5.64048707e-01
1.39770955e-01 1.69333965e-02 1.84974045e-01 -4.04587507e-01
-1.27623248e+00 2.92757124e-01 -3.97691578e-02 -1.18086794e-02
1.14898312e+00 -4.10623759e-01 -4.32045698e-01 -5.01102507e-01
-7.36330926e-01 6.40429482e-02 9.26524282e-01 7.05030382e-01
6.78726912e-01 -1.54086101e+00 -4.07798827e-01 2.66617209e-01
3.65286827e-01 3.76011431e-01 2.16057509e-01 6.05591178e-01
-6.42701328e-01 2.57707983e-01 -5.45824528e-01 -6.27596974e-01
-1.04920483e+00 4.44216460e-01 3.97639722e-01 -6.57400787e-02
-6.55809999e-01 8.33663821e-01 3.25449407e-01 -1.00843787e+00
1.02155201e-01 -5.37710369e-01 -1.33577868e-01 -1.41829148e-01
1.55468747e-01 1.85519114e-01 1.04558453e-01 -9.56626534e-01
-5.34808338e-01 7.24929154e-01 3.92400652e-01 -3.18227291e-01
1.70739496e+00 -3.31221014e-01 -2.03263149e-01 8.56769979e-01
1.14236236e+00 -1.28919616e-01 -1.23382950e+00 -4.09979112e-02
-2.25310683e-01 -5.93053937e-01 -1.27512723e-01 -5.02154350e-01
-5.42676628e-01 6.93194211e-01 6.71954215e-01 6.32251650e-02
8.04314733e-01 3.73653769e-01 2.49252424e-01 5.73012114e-01
7.77164400e-01 -7.65352547e-01 1.47520453e-01 7.12710261e-01
1.20535171e+00 -1.03479326e+00 9.62311476e-02 -2.05374017e-01
-5.62299728e-01 8.54719222e-01 4.82891709e-01 -1.60040334e-03
4.56066906e-01 3.36019769e-02 -1.46607742e-01 -3.27867180e-01
-4.08833832e-01 -3.80489051e-01 1.88180938e-01 1.15494585e+00
1.32033722e-02 1.74420401e-01 6.10131323e-01 7.68644810e-02
-5.51990151e-01 -4.22608316e-01 5.19499421e-01 1.17382777e+00
-1.67051584e-01 -1.03045976e+00 -1.24925070e-01 2.10369959e-01
4.28497970e-01 1.17792748e-01 -3.72974366e-01 8.88870478e-01
7.61182010e-02 7.55758524e-01 1.39278710e-01 -4.68631476e-01
7.28306353e-01 -9.96814743e-02 1.15294719e+00 -7.34329224e-01
-9.40731913e-02 -4.00696665e-01 2.94658363e-01 -9.48278666e-01
-7.88689971e-01 -8.31966162e-01 -1.20074153e+00 -1.46538094e-01
-3.70461829e-02 -3.42055738e-01 9.27242041e-01 8.76876771e-01
4.51606423e-01 5.99135995e-01 1.24801271e-01 -1.52772045e+00
7.12014213e-02 -6.18058205e-01 -5.09898007e-01 7.79630721e-01
4.16294396e-01 -1.14177597e+00 -2.91585386e-01 8.00404251e-02] | [4.687088489532471, 0.44569772481918335] |
691803b4-7838-4561-9066-6ebe6bce4dd1 | hybridization-of-filter-and-wrapper | 2210.16496 | null | https://arxiv.org/abs/2210.16496v1 | https://arxiv.org/pdf/2210.16496v1.pdf | Hybridization of filter and wrapper approaches for the dimensionality reduction and classification of hyperspectral images | The high dimensionality of hyperspectral images often imposes a heavy computational burden for image processing. Therefore, dimensionality reduction is often an essential step in order to remove the irrelevant, noisy and redundant bands. And consequently, increase the classification accuracy. However, identification of useful bands from hundreds or even thousands of related bands is a nontrivial task. This paper aims at identifying a small set of bands, for improving computational speed and prediction accuracy. Hence, we have proposed a hybrid algorithm through band selection for dimensionality reduction of hyperspectral images. The proposed approach combines mutual information gain (MIG), Minimum Redundancy Maximum Relevance (mRMR) and Error probability of Fano with Support Vector Machine Bands Elimination (SVM-PF). The proposed approach is compared to an effective reproduced filters approach based on mutual information. Experimental results on HSI AVIRIS 92AV3C have shown that the proposed approach outperforms the reproduced filters. Keywords - Hyperspectral images, Classification, band Selection, filter, wrapper, mutual information, information gain. | ['Chafik Nacir', 'Ahmed Hammouch', 'Elkebir Sarhrouni', 'Maria Merzouqi', 'Asma Elmaizi'] | 2022-10-29 | null | null | null | null | ['classification-of-hyperspectral-images'] | ['computer-vision'] | [ 7.73956478e-01 -4.69929338e-01 3.17868501e-01 -1.95885733e-01
-3.35590571e-01 -5.08146942e-01 2.99672693e-01 1.01073347e-01
-1.37658954e-01 9.59770143e-01 -1.66941926e-01 -1.01952836e-01
-1.14286375e+00 -9.41833675e-01 2.44372822e-02 -1.06563711e+00
-2.67409801e-01 5.67746796e-02 -1.12707488e-01 -5.29857464e-02
4.88355279e-01 6.72559619e-01 -2.28485084e+00 2.30168566e-01
1.35705054e+00 1.04055452e+00 6.01163328e-01 5.34318388e-01
-3.57616805e-02 2.70190686e-01 -2.29525700e-01 2.34453976e-01
5.79192638e-01 -6.38303399e-01 -6.56388760e-01 5.08316517e-01
1.64228603e-02 -1.31072253e-01 1.90489426e-01 1.46420276e+00
5.08773148e-01 5.70455730e-01 9.35566485e-01 -7.84574449e-01
-1.71972752e-01 4.50714111e-01 -7.85330951e-01 2.07963362e-01
-2.12647930e-01 -2.14165956e-01 8.41114283e-01 -8.12173843e-01
3.14280659e-01 7.29594231e-01 5.45636117e-01 -2.12261975e-01
-1.22516048e+00 -5.08457720e-01 -2.90078849e-01 4.44337696e-01
-1.56932390e+00 -9.27068964e-02 6.39585912e-01 -6.10938907e-01
7.23401189e-01 8.72616529e-01 7.76424825e-01 -1.62413374e-01
-1.71503752e-01 2.26552442e-01 1.45482612e+00 -6.73880816e-01
1.54709920e-01 3.73297811e-01 6.97179973e-01 4.14263457e-01
6.93980515e-01 2.42986634e-01 -2.45499581e-01 -2.49357954e-01
9.06016678e-02 1.25732599e-02 -7.15084314e-01 -1.10690385e-01
-5.58029890e-01 7.78269053e-01 4.30536985e-01 4.58152235e-01
-5.58473170e-01 -8.66323113e-01 -3.33491303e-02 3.31566572e-01
5.42368770e-01 5.03158569e-01 -4.18427736e-01 4.21216786e-01
-8.84904683e-01 -2.51307953e-02 5.86785257e-01 5.28185248e-01
1.04970109e+00 -8.83311927e-02 4.49275784e-02 1.02958131e+00
2.95763016e-01 6.86176062e-01 4.16227818e-01 -2.45764360e-01
2.95567866e-02 9.35564876e-01 1.67271852e-01 -1.34277678e+00
-6.53011501e-01 -6.21388972e-01 -9.06310022e-01 3.67242545e-01
-2.06154585e-01 -1.74283549e-01 -8.07178736e-01 9.79187191e-01
2.81329811e-01 -2.71333396e-01 2.39526287e-01 1.10772610e+00
5.79659820e-01 8.41895103e-01 -5.81879765e-02 -5.76410115e-01
1.12211895e+00 -4.35701668e-01 -6.49086893e-01 -1.01121046e-01
3.09003234e-01 -1.05305946e+00 5.33800900e-01 5.73623836e-01
-5.00146449e-01 -3.85766834e-01 -1.08018410e+00 6.92655623e-01
-5.35792410e-01 6.49698973e-01 8.22630107e-01 7.87335932e-01
-4.59657460e-01 7.98080802e-01 -4.04277712e-01 -3.80130440e-01
3.44510198e-01 4.15340394e-01 -4.77166772e-01 -6.13812183e-04
-9.21068311e-01 7.22595036e-01 9.66467738e-01 1.59774721e-01
1.90835997e-01 -4.24501568e-01 -4.58309054e-01 7.06912354e-02
7.42614791e-02 -8.77323896e-02 3.90944064e-01 -1.07946074e+00
-1.19355428e+00 3.77866060e-01 -7.96223655e-02 -4.02685165e-01
2.53170985e-03 -6.45680726e-02 -5.60222328e-01 3.37083042e-01
-2.43001282e-01 1.06313573e-02 7.40491092e-01 -1.08660996e+00
-9.11861062e-01 -7.29561090e-01 -5.08587539e-01 4.12369430e-01
-5.90440571e-01 -2.29991168e-01 5.40594757e-01 -4.07754242e-01
8.45131278e-01 -8.67044449e-01 -8.60439613e-04 -4.65067178e-01
-3.49779397e-01 2.41352767e-01 1.05565023e+00 -8.63194644e-01
1.05738533e+00 -2.20801544e+00 -4.23528850e-02 7.59067595e-01
-2.27304608e-01 3.69542360e-01 8.01003650e-02 3.36033404e-01
-4.16035324e-01 6.03687353e-02 -5.37409067e-01 6.11424685e-01
-4.17004555e-01 -1.76135138e-01 4.79359590e-02 6.15386188e-01
1.52593285e-01 -2.26785749e-01 -5.46759248e-01 -1.10602356e-01
5.01629949e-01 5.51577568e-01 -2.21552685e-01 -4.78928015e-02
2.77437985e-01 2.50701368e-01 -2.23765999e-01 6.53221905e-01
1.18086648e+00 -1.45570979e-01 2.20672414e-02 -8.40011775e-01
-6.25584245e-01 -3.89749408e-01 -1.57654715e+00 1.02109218e+00
-1.39754057e-01 6.00654542e-01 -7.56711373e-03 -1.21244669e+00
1.08380985e+00 1.82702363e-01 5.91576636e-01 -1.78184703e-01
2.01857984e-01 3.29208463e-01 7.85622746e-02 -5.61600387e-01
3.57455313e-01 -6.28451928e-02 8.98144007e-01 8.25086161e-02
-8.93867761e-02 -8.08990821e-02 3.32414478e-01 -3.14087003e-01
3.99877429e-01 -1.86628222e-01 6.73995674e-01 -5.11939406e-01
8.24111283e-01 3.43229145e-01 3.03278923e-01 4.19616848e-01
-2.54617780e-02 2.69176871e-01 -2.29508895e-02 -1.62949860e-01
-9.36056852e-01 -4.24662948e-01 -6.16198301e-01 4.92955148e-01
2.27901712e-01 2.20736325e-01 -3.32494617e-01 -1.86814874e-01
6.82024732e-02 7.53035724e-01 -2.66439587e-01 -1.47443414e-01
2.91977197e-01 -1.42846417e+00 1.29340002e-02 -3.18362951e-01
8.68157625e-01 -3.38520706e-01 -5.52936852e-01 2.68365294e-01
-6.59243613e-02 -4.22454178e-01 2.46374816e-01 3.42270732e-01
-1.04698825e+00 -1.15137136e+00 -3.12049687e-01 -3.59472245e-01
6.89061582e-01 8.50673079e-01 5.20919442e-01 -1.43861681e-01
-8.60936999e-01 -1.96927086e-01 -7.72163332e-01 -5.64041615e-01
-7.92908669e-02 -2.64922321e-01 -2.18557820e-01 2.04698443e-01
5.23064554e-01 -4.75425392e-01 -6.27905786e-01 3.33874613e-01
-9.23760116e-01 -1.02983685e-02 8.22235346e-01 1.03830349e+00
6.89284265e-01 1.09941351e+00 3.92021209e-01 -6.86760724e-01
3.84597093e-01 -1.87245920e-01 -1.01525342e+00 2.06470579e-01
-7.48934150e-01 -6.30896091e-02 3.81939173e-01 1.78873241e-01
-1.29329038e+00 4.04990196e-01 2.11444110e-01 2.93950200e-01
-3.12439978e-01 7.39188135e-01 -1.32655039e-01 -6.50241673e-01
1.00837302e+00 4.92486209e-01 1.65799856e-01 -5.96759260e-01
1.78566221e-02 1.05711901e+00 9.31230038e-02 1.36804909e-01
7.86160767e-01 3.94654334e-01 3.68921965e-01 -1.45353460e+00
-6.47317410e-01 -9.12050188e-01 -6.16078794e-01 -2.33899325e-01
5.37373066e-01 -8.95481646e-01 -4.40050006e-01 3.52482885e-01
-8.01701963e-01 5.86854815e-01 1.00501709e-01 9.55738187e-01
-5.34105338e-02 4.02119428e-01 4.76902351e-02 -1.14061797e+00
-5.62898874e-01 -9.21370983e-01 4.62993324e-01 4.79809612e-01
3.12530607e-01 -5.65392733e-01 -3.50526154e-01 2.53584266e-01
3.71233433e-01 4.40417796e-01 8.15419137e-01 -6.00226343e-01
-3.25181037e-01 -3.55366945e-01 -6.42123222e-01 6.27267659e-01
6.42403364e-01 2.21141204e-01 -8.74961078e-01 7.55777117e-03
1.00604586e-01 1.21202953e-01 1.01535022e+00 5.90729237e-01
9.31506693e-01 -2.62070537e-01 -2.27733403e-01 7.15836227e-01
1.96585846e+00 6.09707475e-01 6.26911223e-01 5.87944508e-01
2.82811642e-01 6.32176042e-01 1.01226091e+00 9.09007788e-01
-5.18235147e-01 -1.43703043e-01 6.06253445e-01 -6.97976798e-02
2.60571063e-01 6.56615734e-01 -2.63769120e-01 5.24408340e-01
-5.61072648e-01 -1.39692128e-01 -7.46680140e-01 1.71852604e-01
-1.52702034e+00 -1.25103641e+00 -6.42474473e-01 2.27744937e+00
4.55194771e-01 -4.32978123e-01 -2.24115491e-01 8.68429780e-01
1.07296193e+00 -9.87279639e-02 -2.42770374e-01 5.95863909e-02
-5.00246644e-01 1.09547079e-01 9.47910011e-01 4.69715625e-01
-1.54261494e+00 3.71596694e-01 4.76330566e+00 6.48552120e-01
-1.05491841e+00 -1.53184935e-01 2.91804612e-01 3.52869898e-01
9.37792212e-02 7.48676807e-03 -6.20752990e-01 3.05324048e-01
5.10231793e-01 -2.18850970e-01 6.51786387e-01 7.24950314e-01
3.41215819e-01 -5.59944868e-01 -1.45382881e-01 1.09086001e+00
1.90401766e-02 -1.03134120e+00 2.41374820e-01 7.92744160e-02
9.36610460e-01 -2.68778116e-01 -9.54611450e-02 -4.80785400e-01
-1.45847410e-01 -5.05028248e-01 1.72146529e-01 8.00823450e-01
1.87195241e-01 -1.19468725e+00 9.62776840e-01 3.43828112e-01
-1.13657987e+00 -4.23094362e-01 -7.57896364e-01 -5.52140512e-02
-4.51575726e-01 1.14316511e+00 -7.39881098e-01 1.02174675e+00
5.89320600e-01 5.96354604e-01 -4.91687745e-01 1.56125283e+00
1.95054621e-01 3.48178744e-01 -3.76775444e-01 -9.01293755e-02
2.28542909e-01 -7.59890139e-01 7.09065676e-01 9.60094690e-01
7.63878822e-01 6.60150766e-01 -7.77103379e-02 5.23446381e-01
6.02198720e-01 7.09137499e-01 -4.91064578e-01 -2.81347781e-01
4.65163618e-01 1.23275983e+00 -8.48334789e-01 -2.41687402e-01
-2.07078412e-01 7.71695256e-01 -5.39424062e-01 1.41098844e-02
-4.69551951e-01 -8.11234355e-01 5.58207572e-01 -2.39040881e-01
1.78082943e-01 -2.96380129e-02 -2.67358035e-01 -7.47253239e-01
-3.22648168e-01 -8.06654811e-01 5.55709064e-01 -6.72408700e-01
-1.12976420e+00 6.62372708e-01 -1.77334726e-01 -1.72677052e+00
3.23648870e-01 -7.31690526e-01 -8.42895545e-03 1.23267376e+00
-1.51834404e+00 -7.18755245e-01 -7.95042276e-01 4.79126424e-01
3.91687542e-01 -3.12794745e-01 9.52057540e-01 3.23686272e-01
-5.72777987e-01 -2.01899737e-01 7.06905842e-01 -3.78899038e-01
3.93663436e-01 -9.97360289e-01 -7.61873186e-01 9.26094770e-01
-4.04466093e-01 2.31754139e-01 8.32924485e-01 -7.37132490e-01
-1.22540605e+00 -1.21302414e+00 3.96411270e-01 6.13502741e-01
4.18859452e-01 5.96970797e-01 -8.85380208e-01 -7.05094412e-02
2.39747087e-03 -3.19668949e-01 1.14221287e+00 -3.61449450e-01
2.15720102e-01 -4.14904505e-01 -1.25439465e+00 2.85810739e-01
5.30804574e-01 -1.17472120e-01 -1.74209476e-01 9.28829372e-01
7.26225004e-02 1.97984949e-01 -1.02295125e+00 5.66483498e-01
5.26236534e-01 -1.19359899e+00 9.26770091e-01 -1.94733173e-01
6.24868758e-02 -6.03437543e-01 -3.24525505e-01 -1.42045438e+00
-6.70992136e-01 -7.22623393e-02 6.12569034e-01 1.13741159e+00
5.00444770e-01 -6.51397109e-01 5.59993505e-01 4.12654579e-01
6.72580823e-02 4.77097444e-02 -4.81164724e-01 -8.16382051e-01
-7.60264814e-01 5.35098165e-02 7.01180160e-01 1.12833536e+00
-2.02654377e-02 -4.19616234e-03 -4.01138395e-01 8.14094961e-01
9.66316462e-01 3.96914154e-01 3.00448358e-01 -1.89946628e+00
-2.08630115e-01 -4.90734726e-01 -7.17842519e-01 1.78398332e-03
-2.84222811e-01 -7.55414665e-01 -8.65841582e-02 -1.67280483e+00
-1.72842834e-02 -1.76350668e-01 -2.66991615e-01 2.41994947e-01
-2.49379292e-01 2.47127101e-01 1.32763870e-02 3.20766926e-01
2.98087209e-01 3.12487155e-01 1.00427842e+00 -3.29303414e-01
-6.14270210e-01 2.27387741e-01 -5.86100996e-01 7.20685542e-01
1.04249835e+00 -3.91652167e-01 -4.89273757e-01 -3.26704159e-02
6.89977258e-02 -3.47862467e-02 7.03172088e-02 -1.33344817e+00
1.46917731e-01 -3.64822686e-01 7.30542064e-01 -8.23242784e-01
2.14250296e-01 -1.15549970e+00 6.35407090e-01 6.60925865e-01
-1.18199131e-02 -5.55489004e-01 1.66180968e-01 3.85766566e-01
-3.78634661e-01 -6.73406482e-01 9.20845091e-01 -2.97709227e-01
-9.92871821e-01 -1.00438762e-02 -3.89652014e-01 -9.93284941e-01
1.08191419e+00 -4.79375213e-01 -2.89986879e-01 1.23163916e-01
-9.19424653e-01 -2.19002038e-01 4.04408053e-02 -2.04262704e-01
5.44338346e-01 -9.37442899e-01 -8.60960186e-01 2.73227036e-01
2.85624832e-01 -5.14994502e-01 4.97668296e-01 7.41451979e-01
-8.51258695e-01 2.88837343e-01 -4.44427818e-01 -4.77609903e-01
-1.74594748e+00 1.22791499e-01 1.86525300e-01 6.78072795e-02
-1.30055532e-01 8.69739890e-01 -3.82800579e-01 -4.97589782e-02
-3.06488842e-01 2.71491148e-02 -8.26650739e-01 4.72507000e-01
6.45230353e-01 6.94391489e-01 4.41543579e-01 -7.34758198e-01
-1.98361948e-01 7.12315202e-01 3.40194702e-01 2.91230589e-01
1.41542315e+00 -2.47158110e-01 -6.68707728e-01 1.14156567e-02
9.86287236e-01 -1.54413909e-01 -6.69171512e-01 -1.53755192e-02
-6.32197559e-02 -1.09100139e+00 5.45045257e-01 -8.10102165e-01
-7.32919633e-01 5.25268614e-01 1.26127410e+00 5.90755224e-01
1.71628642e+00 -7.09623873e-01 -6.18439093e-02 8.32686484e-01
1.30468979e-01 -1.15551698e+00 -7.59806156e-01 2.38833874e-01
7.70947218e-01 -1.42835081e+00 4.29033190e-01 -8.37949157e-01
-3.53246808e-01 1.45442677e+00 3.13623309e-01 2.27178469e-01
1.04881847e+00 -2.21390724e-01 -2.19163537e-01 -3.17609141e-04
-1.57967985e-01 -6.86880827e-01 3.38577718e-01 5.99235475e-01
5.18811226e-01 2.56337494e-01 -8.95320058e-01 1.86570540e-01
-1.49091452e-01 -1.47053659e-01 3.90728086e-01 9.63752866e-01
-1.18983817e+00 -6.84641242e-01 -8.71778905e-01 1.03558862e+00
-1.90548986e-01 -2.17390105e-01 -2.85724998e-01 6.11186028e-01
1.92444935e-01 1.22376668e+00 -2.01957300e-01 -5.45181096e-01
2.58946687e-01 1.06551945e-01 1.98114038e-01 -3.56170982e-01
-1.06654555e-01 2.10850433e-01 8.21284670e-03 4.25770320e-02
-8.19095314e-01 -5.96733749e-01 -8.37675989e-01 7.47213662e-02
-1.12159014e+00 5.25111377e-01 1.20043504e+00 7.83689439e-01
2.21039832e-01 3.48098874e-01 9.88908172e-01 -8.10610712e-01
-5.73000789e-01 -1.04378617e+00 -1.22311294e+00 1.63493738e-01
1.71185911e-01 -8.10048401e-01 -6.64262533e-01 2.25912943e-01] | [9.755059242248535, -1.8168561458587646] |
6741f150-f03c-4afc-b58f-dc7b9f279989 | a-case-study-on-the-impact-of-dynamic-time | 2010.05270 | null | https://arxiv.org/abs/2010.05270v1 | https://arxiv.org/pdf/2010.05270v1.pdf | A Case-Study on the Impact of Dynamic Time Warping in Time Series Regression | It is well understood that Dynamic Time Warping (DTW) is effective in revealing similarities between time series that do not align perfectly. In this paper, we illustrate this on spectroscopy time-series data. We show that DTW is effective in improving accuracy on a regression task when only a single wavelength is considered. When combined with k-Nearest Neighbour, DTW has the added advantage that it can reveal similarities and differences between samples at the level of the time-series. However, in the problem, we consider here data is available across a spectrum of wavelengths. If aggregate statistics (means, variances) are used across many wavelengths the benefits of DTW are no longer apparent. We present this as another example of a situation where big data trumps sophisticated models in Machine Learning. | ['Pádraig Cunningham', 'Vivek Mahato'] | 2020-10-11 | null | null | null | null | ['time-series-regression'] | ['time-series'] | [ 4.25623059e-01 -7.23364830e-01 1.11108549e-01 -2.30587482e-01
-5.76106489e-01 -8.97305906e-01 7.51471817e-01 2.57680446e-01
-5.36214054e-01 7.64014602e-01 2.61205614e-01 -3.14949900e-01
-8.93228829e-01 -5.70570648e-01 -1.92163169e-01 -1.10980976e+00
-6.48290038e-01 9.69221517e-02 2.23894671e-01 -3.94716352e-01
4.30620998e-01 9.85124052e-01 -1.23318338e+00 3.07417754e-02
5.31450748e-01 6.91633224e-01 -1.21326618e-01 6.33547604e-01
-9.70878005e-02 3.73455286e-01 -4.85938102e-01 2.07052007e-02
5.28391659e-01 -3.90636921e-01 -6.36070192e-01 -3.59204561e-01
1.56436488e-01 2.44831786e-01 -1.54708713e-01 7.30087161e-01
5.23232400e-01 3.31642121e-01 4.79589909e-01 -1.04927957e+00
-3.60568821e-01 2.19028115e-01 -8.29893112e-01 5.81848264e-01
2.42983133e-01 3.42637330e-01 9.07223463e-01 -2.65294582e-01
7.09224582e-01 9.80681419e-01 1.07554579e+00 -8.46436322e-02
-1.74904621e+00 -3.22710872e-01 -2.86196947e-01 4.52128112e-01
-1.16314161e+00 -2.14886501e-01 9.99371707e-01 -5.56451738e-01
1.15251899e+00 5.33512652e-01 6.07822716e-01 7.13760972e-01
6.12550676e-01 -1.47549763e-01 1.66717720e+00 -5.60595036e-01
7.68270269e-02 -4.43911493e-01 2.94227839e-01 -5.46743236e-02
2.99051646e-02 6.56812251e-01 -6.40872061e-01 -4.10787612e-01
3.83498996e-01 1.95375368e-01 -8.86461213e-02 -2.25048989e-01
-1.48132968e+00 6.90083206e-01 2.27314234e-01 7.06968367e-01
-3.41817200e-01 -2.37849265e-01 4.41261441e-01 1.01160455e+00
7.53189981e-01 1.00165784e+00 -6.04901671e-01 -2.68873692e-01
-1.07277238e+00 2.40747526e-01 6.31458223e-01 1.10399827e-01
7.08878398e-01 -1.92071736e-01 1.71958610e-01 5.81069767e-01
-3.50904346e-01 4.11675066e-01 6.82427704e-01 -9.55678165e-01
5.67403510e-02 2.73144096e-01 1.08968087e-01 -9.05384123e-01
-9.66112673e-01 -2.50577182e-01 -5.77241659e-01 4.69862550e-01
9.00786996e-01 -1.16689511e-01 -6.00934684e-01 1.50828898e+00
1.59303933e-01 2.68838912e-01 1.20193437e-01 6.86014950e-01
2.04365477e-01 6.69133186e-01 -3.17108482e-02 -9.11747038e-01
1.14752328e+00 -2.94750154e-01 -6.89840794e-01 -2.17279326e-02
4.51109380e-01 -1.04629087e+00 9.16050851e-01 4.25248891e-01
-7.85283566e-01 -4.19572860e-01 -9.66937125e-01 3.29305567e-02
-6.90511405e-01 -8.91630709e-01 7.50713706e-01 3.81628692e-01
-9.95406032e-01 1.24756193e+00 -7.47866690e-01 -6.17107749e-01
-2.37494141e-01 1.83883756e-01 -4.40160394e-01 1.24482892e-01
-1.00460815e+00 9.60981905e-01 2.02692062e-01 -1.16190732e-01
2.87056807e-03 -1.01038527e+00 -3.23308140e-01 -1.88815027e-01
2.17507467e-01 1.54922707e-02 9.27520871e-01 -7.08406508e-01
-1.25536966e+00 5.83912313e-01 -2.67867111e-02 -2.63683617e-01
5.10646820e-01 2.96565384e-01 -9.59817708e-01 1.09451726e-01
-2.49464154e-01 -4.47962694e-02 7.24189162e-01 -5.71347177e-01
-2.42625132e-01 -6.50817871e-01 -3.93481702e-01 -5.39288282e-01
-4.04139096e-03 3.31802696e-01 4.08733368e-01 -5.03420293e-01
3.76908004e-01 -9.88273263e-01 -2.17885926e-01 -2.37571329e-01
-6.29596710e-02 -7.74265826e-02 8.28398526e-01 -5.81199825e-01
9.89420474e-01 -2.13690686e+00 5.39277168e-03 5.54279983e-01
2.21288800e-01 -7.31904581e-02 -2.51225889e-01 1.16465533e+00
-6.94561481e-01 -9.51995200e-04 -3.56515318e-01 3.70648235e-01
-2.74167120e-01 4.25405800e-02 -4.66498077e-01 7.79510736e-01
1.93642229e-01 6.16236567e-01 -7.64899313e-01 5.58713600e-02
1.67058200e-01 2.75674134e-01 -1.18320741e-01 -3.24855477e-01
2.06702322e-01 5.85609078e-01 1.02195125e-02 2.57166117e-01
6.15377486e-01 -8.29001591e-02 5.86876795e-02 -1.95602819e-01
-8.99449527e-01 1.37268044e-02 -9.32790160e-01 1.50533938e+00
-4.11155727e-03 9.73757982e-01 -4.37546730e-01 -1.12652004e+00
9.36539054e-01 1.91622704e-01 9.63870108e-01 -9.62531090e-01
-1.90071583e-01 3.22243720e-01 6.45632982e-01 -6.42564058e-01
1.55160666e-01 -5.57848215e-01 2.25093380e-01 5.94494581e-01
-1.63597822e-01 -2.27891415e-01 2.13624537e-01 -3.19728315e-01
1.33383274e+00 -9.61740911e-02 5.27604818e-01 -4.85822707e-01
2.67326415e-01 3.00042778e-01 2.18681723e-01 5.32136142e-01
-1.10384963e-01 3.69049072e-01 6.03013277e-01 -5.24468601e-01
-1.53567314e+00 -8.86818230e-01 -4.90905434e-01 7.40717053e-01
-2.85656303e-01 -4.74861443e-01 -7.28099123e-02 8.87070894e-02
3.13812673e-01 4.11359102e-01 -6.04874134e-01 -9.89408121e-02
-5.21449387e-01 -1.26091731e+00 1.51375920e-01 5.08512259e-02
2.15660527e-01 -6.63519382e-01 -8.62390757e-01 2.88234204e-01
1.79142341e-01 -8.64690483e-01 -1.71145275e-01 3.43471229e-01
-1.20858884e+00 -9.48834181e-01 -5.69657743e-01 4.54791263e-02
5.72282337e-02 6.53828263e-01 8.22743118e-01 -3.82342070e-01
-7.55536258e-01 2.36576051e-01 -2.90751070e-01 -5.98982632e-01
-2.18345419e-01 -2.66559035e-01 2.01023687e-02 -8.55031833e-02
5.36097527e-01 -1.09502947e+00 -4.21971768e-01 2.28609756e-01
-6.42000675e-01 -3.31736356e-01 2.45163590e-01 6.97346568e-01
3.20223331e-01 2.88237095e-01 4.08803910e-01 -6.28428757e-01
6.16595447e-01 -4.42235887e-01 -6.36200845e-01 1.83004439e-01
-7.37910569e-01 8.07231441e-02 6.00320280e-01 -6.54831350e-01
-6.66515827e-01 -3.90049726e-01 4.67071086e-01 -1.54540673e-01
6.24675155e-02 8.01780522e-01 6.47553384e-01 -3.91520143e-01
1.00847089e+00 1.78153381e-01 4.51176316e-01 -6.70441747e-01
6.69721514e-02 4.50433612e-01 3.89793277e-01 -5.10874629e-01
7.56423593e-01 6.59315348e-01 6.34190321e-01 -1.07040656e+00
-3.15718710e-01 -6.74166024e-01 -9.31670845e-01 -2.67997295e-01
4.83913481e-01 -6.05467379e-01 -7.84040987e-01 3.49026382e-01
-6.91521645e-01 -2.28028402e-01 -5.32374144e-01 7.91950464e-01
-6.13685668e-01 3.42988789e-01 -1.33021072e-01 -8.26703608e-01
2.08198458e-01 -6.51520669e-01 5.66310525e-01 2.99062971e-02
-3.19493651e-01 -1.12508821e+00 3.89173537e-01 -1.28245682e-01
6.51803017e-01 8.83795261e-01 8.95878494e-01 -6.87829077e-01
-1.83808386e-01 -3.93184245e-01 -1.19538326e-02 -1.02345735e-01
4.17943627e-01 4.23573703e-01 -1.18630195e+00 -4.17764604e-01
3.06080014e-01 2.46796235e-01 6.39917612e-01 6.03760421e-01
1.02711213e+00 1.61108058e-02 -1.19222336e-01 4.73858833e-01
1.54818714e+00 4.42594975e-01 5.32685220e-01 4.65960443e-01
3.57081264e-01 1.17957878e+00 3.68774205e-01 3.97000313e-01
-1.25518054e-01 6.70480072e-01 -5.41476011e-02 -1.17749199e-01
1.57147691e-01 3.33863169e-01 3.10210079e-01 6.12808704e-01
-6.46169424e-01 1.86925530e-01 -1.09641004e+00 3.01088154e-01
-1.65469980e+00 -1.68572104e+00 -5.83516240e-01 2.34162998e+00
5.81665993e-01 6.33259565e-02 6.81011856e-01 4.88197237e-01
5.83718598e-01 4.93803382e-01 -6.09237492e-01 -4.24723864e-01
-3.08333784e-01 2.33937085e-01 7.90833175e-01 3.74809444e-01
-7.33700514e-01 1.33577064e-01 7.49074411e+00 5.42342961e-01
-1.40821397e+00 1.39977559e-01 7.48231262e-02 -4.97107118e-01
-1.22497715e-01 1.37603685e-01 9.08555612e-02 4.66525853e-01
1.34265161e+00 -6.02114320e-01 7.36213923e-01 7.21402988e-02
6.49647057e-01 -3.06444049e-01 -9.71159756e-01 9.93455648e-01
-3.78269464e-01 -1.06339407e+00 -6.50730371e-01 2.61219412e-01
4.76452500e-01 2.86051065e-01 5.28309774e-03 -1.57771274e-01
1.47452084e-02 -8.78938437e-01 3.22446018e-01 7.57212758e-01
5.52527726e-01 -7.29513347e-01 4.18825805e-01 2.43984446e-01
-8.81924212e-01 -2.37121031e-01 -4.33253407e-01 -5.29673815e-01
3.26089934e-02 8.39757383e-01 -4.57532525e-01 7.20768511e-01
5.20355821e-01 7.76086450e-01 -6.25232697e-01 1.22406995e+00
5.28532982e-01 3.69194865e-01 -4.54375505e-01 1.44506514e-01
3.07733357e-01 -7.00335681e-01 6.06356919e-01 8.98042440e-01
6.79578185e-01 4.28362697e-01 -8.55478197e-02 4.47788477e-01
8.22416306e-01 5.19439466e-02 -7.93264747e-01 -4.27815586e-01
3.57674122e-01 1.15686071e+00 -7.77205288e-01 1.20714910e-01
-7.64881790e-01 5.22211075e-01 -2.08716050e-01 3.65350544e-01
-2.28897825e-01 -4.38384831e-01 9.18342233e-01 8.89900550e-02
-1.08808845e-01 -6.68759406e-01 -6.10732809e-02 -8.51802409e-01
-1.39535174e-01 -6.65928602e-01 6.47512138e-01 -1.05337119e+00
-1.53811014e+00 8.34355876e-02 1.97300538e-01 -1.29166234e+00
-1.94630340e-01 -8.27931643e-01 -6.84375286e-01 1.13191938e+00
-1.43585432e+00 -7.01341152e-01 -1.64615229e-01 5.96499741e-01
4.77858707e-02 2.78942548e-02 6.62311912e-01 1.40870232e-02
-3.44595850e-01 -1.52638972e-01 8.05704534e-01 -4.47912872e-01
8.71470571e-01 -1.35273445e+00 4.16435421e-01 7.08345294e-01
7.35952482e-02 6.78932488e-01 1.31328714e+00 -6.27384961e-01
-1.39273787e+00 -4.32117462e-01 8.46840978e-01 -2.43393570e-01
1.28639400e+00 -2.15229001e-02 -1.12141621e+00 3.16426277e-01
1.09744109e-01 -2.30404455e-02 1.15531325e+00 4.37667280e-01
-3.84060532e-01 -1.34668544e-01 -1.17528379e+00 3.79351616e-01
1.03582406e+00 -9.30211484e-01 -5.17049372e-01 6.83283627e-01
4.10156816e-01 -9.37406532e-03 -1.34197009e+00 1.50158361e-01
7.81403601e-01 -1.18682396e+00 1.20403194e+00 -8.52739632e-01
1.24686278e-01 -2.59581983e-01 -1.85431436e-01 -1.47693992e+00
-4.36070055e-01 -9.95093048e-01 4.78461266e-01 9.23747182e-01
1.21096939e-01 -9.85239148e-01 1.88271150e-01 6.02117896e-01
1.69954389e-01 -1.83966294e-01 -1.04604971e+00 -1.40822136e+00
1.09274708e-01 -3.50384712e-01 7.75578976e-01 1.25247192e+00
3.89858902e-01 -1.16504654e-01 -2.81246096e-01 1.25801831e-01
7.06495702e-01 3.80220860e-01 3.54889274e-01 -1.73453414e+00
-2.13116422e-01 -3.82370412e-01 -5.71839213e-01 7.38049671e-02
-1.74658075e-01 -6.79673791e-01 -4.10434574e-01 -7.91507185e-01
-3.05624842e-03 -1.22832634e-01 -4.94325191e-01 1.78861767e-01
2.21511796e-01 1.88492805e-01 2.69777179e-01 5.56284964e-01
3.49607319e-01 -1.59714609e-01 1.30353880e+00 2.68866383e-02
-2.87858546e-01 -1.25235051e-01 -3.84469330e-01 3.37380171e-01
9.37582672e-01 -3.71711016e-01 -3.52050036e-01 1.49223208e-01
1.59661755e-01 2.67844677e-01 5.00156283e-01 -1.04823220e+00
2.02700213e-01 -5.25678396e-01 3.66994262e-01 -4.46366817e-01
3.47703576e-01 -8.47701430e-01 7.67945230e-01 5.76389432e-01
-3.50546569e-01 5.44165134e-01 2.59326994e-01 4.59414482e-01
-3.07541527e-02 -1.91863179e-01 7.80878603e-01 -1.63896844e-01
-7.75930166e-01 -1.86857190e-02 -2.34394297e-01 -2.46680349e-01
9.10944402e-01 -1.63756087e-01 -6.53816283e-01 -1.41183123e-01
-7.53390551e-01 -3.93123962e-02 5.70870221e-01 1.71598122e-01
4.72467020e-02 -1.32760966e+00 -5.85325480e-01 1.63700059e-01
2.31783852e-01 -8.29904795e-01 4.57769275e-01 1.31440508e+00
-2.80489266e-01 4.36694413e-01 -5.70185721e-01 -5.58304191e-01
-1.36518943e+00 8.65475118e-01 3.01961690e-01 1.27313122e-01
-6.95185781e-01 4.66550142e-01 -3.15011144e-01 -1.06748722e-01
-6.08312547e-01 -3.23002934e-01 3.56924459e-02 6.23015702e-01
7.73001730e-01 5.05848825e-01 1.58912003e-01 -3.93061221e-01
-2.91408479e-01 9.63188469e-01 6.47966191e-02 -1.93460271e-01
1.68223858e+00 -1.72595397e-01 -4.36827600e-01 1.06754124e+00
1.26144755e+00 9.94464755e-02 -1.24172187e+00 -2.99107045e-01
2.39360854e-01 -6.33383870e-01 1.65322367e-02 -6.84571087e-01
-5.48547387e-01 8.61172438e-01 7.21184134e-01 1.03421521e+00
1.30167055e+00 -1.86907277e-01 3.36235285e-01 4.10551131e-01
2.33334288e-01 -1.03730273e+00 -2.54377186e-01 2.77278453e-01
8.63004446e-01 -1.06964350e+00 3.61260921e-01 -1.03024229e-01
-2.58275479e-01 1.67472899e+00 -1.64044917e-01 -8.80240574e-02
4.89102632e-01 8.44143108e-02 1.40681602e-02 -3.97162050e-01
-6.63901627e-01 -4.02992994e-01 1.03220321e-01 6.51300311e-01
4.07164842e-01 1.15387030e-02 -4.51821297e-01 -1.79949746e-01
-4.09611076e-01 -1.38350576e-01 7.18888879e-01 8.93117368e-01
-6.37789190e-01 -1.23785734e+00 -5.82037807e-01 4.97061610e-01
-2.29516238e-01 6.73635677e-02 -4.46239144e-01 1.04040301e+00
-1.77009583e-01 9.01228905e-01 1.76940069e-01 -3.96327615e-01
3.09442669e-01 5.78490317e-01 6.22504830e-01 -8.02361295e-02
-4.82093036e-01 2.07592055e-01 -3.45461583e-03 -7.60067105e-01
-8.68821740e-01 -1.07504404e+00 -7.27515876e-01 -7.80608356e-01
-3.03761754e-02 1.36069387e-01 6.89242184e-01 9.14468110e-01
1.59405932e-01 5.17301142e-01 9.10230100e-01 -9.05266285e-01
-6.09677553e-01 -6.82765126e-01 -1.08654475e+00 2.72635400e-01
8.97280157e-01 -5.75704753e-01 -6.42457962e-01 -3.58648002e-02] | [7.255536079406738, 3.3409440517425537] |
c6d29958-54bd-438f-9df5-295e9197d972 | monitored-distillation-for-positive-congruent | 2203.16034 | null | https://arxiv.org/abs/2203.16034v2 | https://arxiv.org/pdf/2203.16034v2.pdf | Monitored Distillation for Positive Congruent Depth Completion | We propose a method to infer a dense depth map from a single image, its calibration, and the associated sparse point cloud. In order to leverage existing models (teachers) that produce putative depth maps, we propose an adaptive knowledge distillation approach that yields a positive congruent training process, wherein a student model avoids learning the error modes of the teachers. In the absence of ground truth for model selection and training, our method, termed Monitored Distillation, allows a student to exploit a blind ensemble of teachers by selectively learning from predictions that best minimize the reconstruction error for a given image. Monitored Distillation yields a distilled depth map and a confidence map, or ``monitor'', for how well a prediction from a particular teacher fits the observed image. The monitor adaptively weights the distilled depth where if all of the teachers exhibit high residuals, the standard unsupervised image reconstruction loss takes over as the supervisory signal. On indoor scenes (VOID), we outperform blind ensembling baselines by 17.53% and unsupervised methods by 24.25%; we boast a 79% model size reduction while maintaining comparable performance to the best supervised method. For outdoors (KITTI), we tie for 5th overall on the benchmark despite not using ground truth. Code available at: https://github.com/alexklwong/mondi-python. | ['Alex Wong', 'Byung-Woo Hong', 'Allison Chen', 'Parth Agrawal', 'Tian Yu Liu'] | 2022-03-30 | null | null | null | null | ['depth-completion'] | ['computer-vision'] | [ 2.52682716e-01 6.22119904e-01 -1.04109876e-01 -4.02346581e-01
-1.29511750e+00 -5.53995311e-01 5.74084997e-01 1.00598842e-01
-4.64709878e-01 5.64164221e-01 5.00100628e-02 -1.25640616e-01
-2.38126982e-02 -5.95229864e-01 -1.13132322e+00 -8.84764910e-01
2.75629282e-01 6.53636932e-01 2.84049153e-01 4.89532292e-01
3.85888338e-01 1.66812524e-01 -1.63504112e+00 5.13371266e-02
1.01166546e+00 9.56784904e-01 4.05395180e-01 1.01502728e+00
2.23881647e-01 1.01877344e+00 -3.26332599e-01 -5.11727870e-01
4.50840473e-01 -1.88762292e-01 -8.11110735e-01 2.00790405e-01
1.25418210e+00 -5.11995077e-01 -2.73428917e-01 1.13134503e+00
5.16497672e-01 1.40943021e-01 7.34109640e-01 -9.23597217e-01
-3.60502690e-01 3.16112190e-01 -6.56569898e-01 -5.41851074e-02
2.02882767e-01 3.69464517e-01 8.64579916e-01 -1.12313604e+00
3.28294814e-01 8.51724505e-01 8.15990269e-01 4.34239000e-01
-1.56974304e+00 -7.40160286e-01 2.30540201e-01 2.51019672e-02
-1.44645047e+00 -6.82924867e-01 4.01973933e-01 -7.42345393e-01
8.27213645e-01 -6.85941195e-03 6.91491902e-01 8.54514599e-01
-2.59765029e-01 7.12181807e-01 1.05536270e+00 -4.20000643e-01
3.86615306e-01 3.80952746e-01 -1.42566428e-01 8.83643806e-01
-1.95412152e-03 2.83559442e-01 -1.02910781e+00 1.25626698e-01
8.35041463e-01 -3.03707898e-01 -3.90064389e-01 -6.99047804e-01
-1.21665335e+00 5.95746756e-01 7.83651173e-01 -5.32627106e-01
-2.40863562e-01 1.82909101e-01 -3.96691322e-01 5.86720817e-02
4.15725738e-01 5.57220876e-01 -4.36073273e-01 -1.09725013e-01
-1.28109574e+00 1.88311726e-01 6.24741137e-01 9.56886411e-01
1.40777516e+00 -1.36430204e-01 1.44465894e-01 6.20950103e-01
4.47753906e-01 5.65379798e-01 4.55000579e-01 -1.47431862e+00
3.89041096e-01 5.13077736e-01 1.33712828e-01 -7.15927720e-01
-6.67518675e-02 -4.90317404e-01 -4.42131132e-01 4.59910005e-01
4.80449706e-01 -4.68658879e-02 -1.27554750e+00 1.57073271e+00
6.05012894e-01 6.72787011e-01 1.14646941e-01 9.66454864e-01
7.47745633e-01 3.47228169e-01 -3.14038128e-01 1.37155041e-01
7.96503127e-01 -1.18440604e+00 1.51790259e-02 -8.11123550e-01
4.74450052e-01 -6.67198181e-01 1.05358410e+00 8.81998599e-01
-1.21143603e+00 -4.54539120e-01 -1.09784961e+00 -9.90802273e-02
-9.96166747e-03 1.73159286e-01 5.48972487e-01 2.80195653e-01
-1.41621900e+00 7.28198946e-01 -1.10273206e+00 -2.07930773e-01
5.10045409e-01 4.43521023e-01 -3.75341594e-01 -8.70295018e-02
-4.39683467e-01 6.25076056e-01 2.44396910e-01 -1.62582636e-01
-1.54255772e+00 -1.01862037e+00 -7.70474792e-01 -3.21195692e-01
1.40633911e-01 -6.02917612e-01 1.56791806e+00 -9.49117780e-01
-1.35548937e+00 1.14624786e+00 -1.71071336e-01 -5.97455144e-01
6.46658599e-01 -4.42110717e-01 2.52461433e-01 3.74123573e-01
3.19869310e-01 1.23774862e+00 1.00550723e+00 -1.48713946e+00
-6.98858202e-01 -3.33996654e-01 -1.19228236e-01 4.67700213e-01
-1.66520551e-01 -6.44061744e-01 -7.37884164e-01 -3.93146187e-01
8.36836934e-01 -1.04500020e+00 -2.55174965e-01 2.83803165e-01
-4.08125520e-01 3.74434561e-01 2.05254585e-01 -5.89527488e-01
8.39455128e-01 -2.03803539e+00 2.61366427e-01 2.66552657e-01
5.12011528e-01 -2.60378152e-01 1.77320793e-01 4.26750742e-02
-1.07416115e-03 -2.23258212e-01 -5.77608764e-01 -8.92141342e-01
-2.29564056e-01 3.07036251e-01 -3.02273899e-01 6.39044881e-01
6.62033111e-02 4.85656410e-01 -1.01815546e+00 -6.19902253e-01
2.91694403e-01 4.17881191e-01 -1.04691303e+00 4.51586068e-01
-2.36484706e-01 7.63552129e-01 1.97039824e-02 6.65708542e-01
5.99939764e-01 -5.50024629e-01 1.03855934e-02 5.57144433e-02
-1.17233224e-01 5.62758327e-01 -1.41632855e+00 2.04020166e+00
-3.38875175e-01 6.91358626e-01 3.30386788e-01 -6.74916565e-01
8.40453923e-01 2.15987325e-01 1.81089208e-01 -3.18370849e-01
-2.55050361e-01 2.12251231e-01 -2.75826603e-01 -2.63883114e-01
3.87377560e-01 -7.89242685e-02 4.05688912e-01 2.78955102e-01
3.93933684e-01 -6.78227603e-01 -4.08625513e-01 5.85258603e-01
1.31103969e+00 4.11742270e-01 -1.14758141e-01 -3.47718708e-02
9.61932447e-03 5.38005605e-02 5.42830765e-01 7.94725001e-01
-1.78661600e-01 1.03982759e+00 2.20838025e-01 -7.28502199e-02
-7.54737377e-01 -1.46595597e+00 -1.26581356e-01 1.00541592e+00
1.21010184e-01 -2.47285008e-01 -7.00257778e-01 -7.18440115e-01
1.60196632e-01 5.49672127e-01 -6.17297947e-01 9.00212303e-03
-8.23927745e-02 -3.47967505e-01 4.93360609e-01 4.96756703e-01
4.45060164e-01 -6.88436747e-01 -3.40735167e-01 -1.05420254e-01
-1.62060052e-01 -1.00009286e+00 -2.48382762e-01 6.24522865e-01
-1.12391925e+00 -1.00132155e+00 -3.97337407e-01 -7.69833446e-01
9.95007396e-01 2.95538425e-01 1.34306836e+00 -5.41468710e-02
6.36002123e-02 6.04613960e-01 1.23085026e-02 -2.54085779e-01
-7.13672787e-02 5.15632853e-02 1.43579930e-01 -2.33456239e-01
2.22929850e-01 -8.60641062e-01 -7.90240884e-01 1.10619210e-01
-4.94821459e-01 3.40625405e-01 5.11576235e-01 7.39303887e-01
1.01555789e+00 -3.46858650e-01 1.79242581e-01 -8.54159653e-01
-2.15212718e-01 -4.56831276e-01 -8.98271263e-01 -8.18408057e-02
-9.53367829e-01 2.02389121e-01 1.80931091e-01 -3.58849347e-01
-9.23869491e-01 4.69568491e-01 5.02576791e-02 -8.24855804e-01
-2.30735749e-01 7.74206519e-02 -6.07735142e-02 -3.01742733e-01
1.02024496e+00 1.51244685e-01 -9.75224897e-02 -6.20139778e-01
2.79511809e-01 3.35747182e-01 1.04793155e+00 -7.90167987e-01
9.58939075e-01 4.92148548e-01 -1.95297852e-01 -4.20771003e-01
-1.20423496e+00 -6.01129830e-01 -7.71669745e-01 -1.40584722e-01
8.40725601e-01 -1.59435940e+00 -5.84304094e-01 4.30663079e-01
-8.70609820e-01 -8.33914936e-01 -2.60888755e-01 5.99869788e-01
-6.22838914e-01 1.01675883e-01 -3.47250730e-01 -7.28906453e-01
1.48903169e-02 -1.10066080e+00 1.05185354e+00 2.70066351e-01
-1.46066159e-01 -9.63616490e-01 2.08703518e-01 6.80948496e-01
-8.34429264e-02 -7.37690777e-02 3.04663002e-01 -3.95427823e-01
-9.69839871e-01 -3.35350856e-02 -1.07028760e-01 4.56621528e-01
-1.68002248e-01 -1.40773347e-02 -1.64415967e+00 -3.00464451e-01
-1.05903573e-01 -6.78539217e-01 1.00011420e+00 2.47390196e-01
1.13770211e+00 -4.15149987e-01 -1.67158827e-01 1.05424321e+00
1.47391427e+00 -3.82259369e-01 3.44480187e-01 5.30936003e-01
9.92482364e-01 5.79800606e-01 4.37105834e-01 3.23947966e-01
6.49474382e-01 2.23685414e-01 8.50161195e-01 6.79407194e-02
-1.27382815e-01 -8.03762197e-01 3.51226717e-01 7.70548582e-01
1.49389163e-01 1.82715163e-01 -1.17503762e+00 9.06039357e-01
-1.58365667e+00 -5.37965059e-01 3.30920666e-02 2.57836843e+00
1.18413842e+00 3.34626734e-01 -1.56218961e-01 3.51305329e-03
3.10582161e-01 1.58958063e-02 -7.70682275e-01 1.06212355e-01
7.35123158e-02 2.64569074e-01 8.46843898e-01 9.69222426e-01
-1.02157927e+00 1.00013936e+00 5.62176037e+00 3.96112859e-01
-8.99782896e-01 8.48597735e-02 5.38297772e-01 -4.04930532e-01
-2.53792048e-01 2.45825112e-01 -9.43018317e-01 3.54682893e-01
9.00987923e-01 3.30766588e-01 6.53701484e-01 9.49828565e-01
1.22433491e-01 -5.02353609e-01 -1.36297917e+00 9.25923765e-01
6.18980452e-02 -1.21789014e+00 -2.69140691e-01 1.87037095e-01
1.23471951e+00 4.00089025e-01 2.39983588e-01 8.75306427e-02
8.25431466e-01 -1.05543911e+00 8.85135174e-01 6.16182625e-01
5.92930973e-01 -4.32415545e-01 2.69665837e-01 5.79110861e-01
-7.75206685e-01 1.34852782e-01 -2.85068423e-01 -2.00958386e-01
-4.86617118e-01 6.90846622e-01 -1.31684101e+00 1.63338795e-01
9.97169435e-01 9.05058503e-01 -8.91854525e-01 1.11926484e+00
-8.09132338e-01 9.69164550e-01 -5.30189514e-01 5.93584239e-01
2.17488050e-01 -1.04557708e-01 4.21162277e-01 9.64060247e-01
2.14917466e-01 -9.45040509e-02 1.80028677e-01 7.74028897e-01
-1.90067798e-01 -3.73844981e-01 -4.45125490e-01 4.26888853e-01
8.14174891e-01 1.05655062e+00 -3.55674416e-01 -3.84672552e-01
-1.77480415e-01 1.09331083e+00 6.33085549e-01 3.98202002e-01
-6.12620354e-01 9.62834433e-02 7.74488211e-01 1.89762071e-01
3.66038978e-01 -1.13106119e-02 -6.60638928e-01 -1.08637810e+00
-7.76990596e-03 -8.44551563e-01 3.33246976e-01 -1.13915324e+00
-1.09671223e+00 1.28920600e-01 -2.28396460e-01 -1.26350987e+00
-2.95293421e-01 -4.10960257e-01 -5.91289163e-01 9.99458909e-01
-1.49157250e+00 -8.62604022e-01 -6.13207221e-01 4.44931895e-01
3.67469430e-01 1.00580350e-01 5.82026958e-01 4.75138053e-03
-4.99068111e-01 6.42284036e-01 6.55604675e-02 3.45803201e-02
1.07295477e+00 -1.77510738e+00 1.70866206e-01 9.14677322e-01
4.04788226e-01 4.88505930e-01 7.85783589e-01 -5.20673215e-01
-1.11318934e+00 -1.18317878e+00 7.15476334e-01 -8.85282099e-01
6.37849987e-01 -1.70327142e-01 -9.79937792e-01 9.60184097e-01
-1.45515859e-01 1.30813181e-01 3.85124296e-01 1.20495018e-02
-5.11962950e-01 -1.71751261e-01 -1.21079087e+00 3.42152148e-01
8.30299020e-01 -7.04920471e-01 -3.88981253e-01 3.42814773e-01
7.51599848e-01 -7.12871373e-01 -8.04882109e-01 1.70557871e-01
4.94822294e-01 -1.33508277e+00 1.02822161e+00 -5.09028602e-03
6.64923549e-01 -5.06029010e-01 -2.90134370e-01 -1.29091799e+00
-9.56258625e-02 -3.42564344e-01 -2.52912611e-01 1.16956139e+00
4.88454849e-01 -4.34566081e-01 1.31979644e+00 6.24213874e-01
-2.43431881e-01 -7.58441031e-01 -7.54073918e-01 -4.77698922e-01
1.35150999e-01 -6.86285615e-01 4.51343656e-01 1.02899027e+00
-4.35195386e-01 1.78275645e-01 -8.63538533e-02 8.36740553e-01
1.05815732e+00 -1.82924122e-01 9.99121249e-01 -1.08461618e+00
-8.05711091e-01 -2.37071946e-01 -4.08271521e-01 -1.42340374e+00
9.32855811e-03 -9.25306082e-01 4.81454730e-01 -1.56665349e+00
7.65687004e-02 -5.47543645e-01 -4.66705672e-02 7.35702634e-01
-3.13709915e-01 4.33579534e-01 -2.64469236e-02 3.95542324e-01
-6.38350964e-01 3.63634020e-01 9.93578196e-01 -1.31144494e-01
-2.37269416e-01 9.13889892e-03 -6.73274934e-01 9.67521667e-01
6.36587143e-01 -5.57404220e-01 -5.35884559e-01 -7.90365160e-01
1.99251175e-01 -4.18254808e-02 8.55691552e-01 -1.37866843e+00
5.47207594e-01 -3.51909325e-02 7.52755225e-01 -5.06020427e-01
6.01428807e-01 -6.90726280e-01 2.63075158e-02 1.93492904e-01
-3.27535361e-01 -2.56895274e-01 2.24699095e-01 5.43422878e-01
-4.35736924e-02 -2.46874794e-01 9.37137723e-01 -1.16784841e-01
-7.19394982e-01 1.89011708e-01 -4.51087626e-03 1.61171094e-01
6.56842828e-01 -3.92887563e-01 -2.51859397e-01 -4.73111600e-01
-7.71617174e-01 4.11959678e-01 9.85265136e-01 -7.23104328e-02
6.71609461e-01 -9.85153913e-01 -5.02715170e-01 4.09516543e-01
1.50565341e-01 9.14495885e-01 -1.81265455e-02 7.12047756e-01
-4.31264520e-01 4.52434225e-03 2.96137363e-01 -9.27819252e-01
-9.26915407e-01 7.85751790e-02 5.95903993e-01 8.72359029e-04
-5.11648357e-01 1.64162862e+00 3.90205890e-01 -7.58997738e-01
5.88713109e-01 -3.24463546e-01 3.04197669e-01 -1.01907939e-01
3.97693962e-01 2.97211349e-01 2.90762365e-01 -3.93311918e-01
-2.64326632e-01 4.66060638e-01 1.51378304e-01 -4.86141533e-01
1.19342184e+00 -1.50136262e-01 9.84457508e-02 5.93394637e-01
9.64059889e-01 1.78523451e-01 -2.15931535e+00 -3.56623590e-01
-1.50091842e-01 -6.41551673e-01 4.24393088e-01 -1.02274907e+00
-1.02448511e+00 9.26915407e-01 5.54726481e-01 -2.83312589e-01
9.61005628e-01 1.36494219e-01 1.36343330e-01 3.87711823e-01
3.16607535e-01 -7.84810603e-01 2.27081776e-01 3.26187909e-01
5.42237580e-01 -1.63946116e+00 2.53440559e-01 -1.18119463e-01
-6.94491863e-01 5.80203176e-01 1.02401793e+00 -1.12526998e-01
5.06009519e-01 2.19426230e-01 2.73683161e-01 -2.07271591e-01
-9.03751731e-01 -1.91970736e-01 2.71033645e-01 7.15376139e-01
2.25259602e-01 1.98268350e-02 8.12373042e-01 2.89162040e-01
-6.97231770e-01 -2.02385098e-01 4.01587546e-01 7.67512083e-01
-7.29901135e-01 -4.90808517e-01 -7.08048403e-01 5.49704909e-01
-7.28824660e-02 -2.86936074e-01 -4.39219356e-01 3.79110426e-01
3.21014076e-01 9.63116586e-01 2.90573150e-01 -5.52266955e-01
8.26625451e-02 1.12509534e-01 4.44158137e-01 -9.47631299e-01
-5.57959199e-01 4.10958081e-02 -2.18788251e-01 -5.41381478e-01
-2.55254954e-01 -6.69217885e-01 -1.36629868e+00 -2.72291511e-01
-2.38163665e-01 1.43001422e-01 5.96804738e-01 7.43811905e-01
1.46053359e-01 2.71844834e-01 5.61827064e-01 -1.04917395e+00
-3.92363012e-01 -9.92739201e-01 -4.08740252e-01 -6.28729314e-02
6.35294139e-01 -6.13491237e-01 -7.54789710e-01 3.04081619e-01] | [8.59000015258789, -2.699378490447998] |
a3c95671-bcb2-4d85-8fa4-2a8f1323bbc8 | norm-guided-latent-space-exploration-for-text | 2306.08687 | null | https://arxiv.org/abs/2306.08687v1 | https://arxiv.org/pdf/2306.08687v1.pdf | Norm-guided latent space exploration for text-to-image generation | Text-to-image diffusion models show great potential in synthesizing a large variety of concepts in new compositions and scenarios. However, their latent seed space is still not well understood and has been shown to have an impact in generating new and rare concepts. Specifically, simple operations like interpolation and centroid finding work poorly with the standard Euclidean and spherical metrics in the latent space. This paper makes the observation that current training procedures make diffusion models biased toward inputs with a narrow range of norm values. This has strong implications for methods that rely on seed manipulation for image generation that can be further applied to few-shot and long-tail learning tasks. To address this issue, we propose a novel method for interpolating between two seeds and demonstrate that it defines a new non-Euclidean metric that takes into account a norm-based prior on seeds. We describe a simple yet efficient algorithm for approximating this metric and use it to further define centroids in the latent seed space. We show that our new interpolation and centroid evaluation techniques significantly enhance the generation of rare concept images. This further leads to state-of-the-art performance on few-shot and long-tail benchmarks, improving prior approach in terms of generation speed, image quality, and semantic content. | ['Gal Chechik', 'Haggai Maron', 'Nir Darshan', 'Rami Ben-Ari', 'Dvir Samuel'] | 2023-06-14 | null | null | null | null | ['long-tail-learning'] | ['methodology'] | [ 4.34030056e-01 -1.68838263e-01 -1.09417036e-01 -2.46471420e-01
-6.58270180e-01 -5.75024068e-01 1.13947809e+00 2.15919957e-01
-4.49335843e-01 6.46031857e-01 2.81416118e-01 7.19577745e-02
-4.46309485e-02 -1.02895069e+00 -6.13611460e-01 -8.38924885e-01
3.95338759e-02 4.79334384e-01 4.10935879e-01 -4.64835852e-01
4.70886171e-01 2.15353310e-01 -1.74394941e+00 1.34574577e-01
8.56005311e-01 7.04534829e-01 2.09324867e-01 7.28954315e-01
-1.78903505e-01 4.56121802e-01 -8.21898878e-01 -5.16006887e-01
4.62647408e-01 -8.45520318e-01 -6.28140986e-01 1.16183609e-01
4.27576900e-01 -2.48301223e-01 4.09196764e-02 1.00358617e+00
6.85514033e-01 4.56216574e-01 1.04267359e+00 -1.21094644e+00
-1.03626990e+00 6.40360475e-01 -6.67087197e-01 2.16463983e-01
1.95189744e-01 2.44588837e-01 1.00166428e+00 -9.96815383e-01
9.96215880e-01 1.07603550e+00 7.06028223e-01 6.21954381e-01
-1.27158046e+00 -5.10238588e-01 -3.35512608e-02 1.37329102e-01
-1.50314331e+00 -3.55965376e-01 4.51677442e-01 -4.46970999e-01
6.78884745e-01 1.29109487e-01 5.56518912e-01 1.19278061e+00
5.42091159e-03 6.69965208e-01 1.12065732e+00 -6.51538908e-01
5.20687521e-01 2.06552431e-01 -3.95157814e-01 5.25154769e-01
2.40218803e-01 4.36734001e-04 -6.03456259e-01 2.13950276e-02
8.48236024e-01 -1.25781447e-01 -2.11475909e-01 -5.37015975e-01
-1.34485495e+00 1.24138701e+00 3.02853197e-01 4.78740573e-01
-2.28946388e-01 3.74301225e-01 2.61940598e-01 4.85907793e-02
6.98339045e-01 7.21920431e-01 2.00723596e-02 -3.62410784e-01
-1.44622386e+00 3.72523665e-01 6.68059707e-01 9.02145743e-01
6.05099738e-01 6.01126812e-02 -6.80675685e-01 1.00227058e+00
-1.58243239e-01 3.70710522e-01 5.57150126e-01 -1.06910491e+00
1.04212455e-01 1.67276919e-01 9.99166146e-02 -8.39755118e-01
1.46533653e-01 -5.24653733e-01 -7.31202900e-01 2.99700737e-01
4.27552015e-01 8.93455073e-02 -1.17326808e+00 1.76033115e+00
2.22786933e-01 4.79778111e-01 -1.04164429e-01 7.26125956e-01
3.82116437e-01 6.26558244e-01 -1.25080541e-01 -9.50730890e-02
1.01643598e+00 -1.15229452e+00 -4.85526770e-01 1.52847068e-02
4.11776930e-01 -8.98883939e-01 1.40457499e+00 3.52916658e-01
-1.42078304e+00 -5.11025667e-01 -1.07626009e+00 -8.71612206e-02
-5.45554996e-01 -2.21934974e-01 7.06903934e-01 9.51844633e-01
-1.19791245e+00 9.05311942e-01 -5.29570341e-01 -4.69813138e-01
6.38164580e-01 1.98603664e-02 8.95911828e-02 -2.46893972e-01
-1.01915026e+00 8.49210680e-01 2.64184862e-01 -4.58781362e-01
-9.69955862e-01 -8.22771370e-01 -7.15569615e-01 1.76469479e-02
3.41384470e-01 -9.31905031e-01 1.15449178e+00 -7.41620839e-01
-1.58194768e+00 6.94348097e-01 4.38027680e-02 -6.56453729e-01
7.28571773e-01 -1.01604775e-01 -1.06435850e-01 3.10080290e-01
3.69718164e-01 1.11512208e+00 1.17469227e+00 -1.29411578e+00
-5.61144590e-01 1.30234361e-01 -2.14088745e-02 1.92204461e-01
-5.49953043e-01 -1.83346659e-01 -4.72757310e-01 -1.11226714e+00
-1.64939791e-01 -8.07070792e-01 -3.12439531e-01 1.58625960e-01
-1.28314093e-01 -1.16708450e-01 4.41597223e-01 -1.58515960e-01
1.06543148e+00 -1.92008221e+00 1.26038447e-01 2.65654594e-01
1.81934774e-01 1.85849532e-01 -3.02055597e-01 5.19489408e-01
1.77416027e-01 3.72802883e-01 -5.06448209e-01 -4.88726765e-01
9.38335201e-04 1.88868344e-01 -4.68483120e-01 2.23230094e-01
2.27116331e-01 9.35023904e-01 -1.21011281e+00 -4.23103839e-01
1.49235561e-01 6.07693076e-01 -6.99070811e-01 -1.17614090e-01
-3.31228882e-01 2.33963072e-01 5.56791015e-02 3.30474555e-01
4.64656383e-01 -2.91316450e-01 -3.95749003e-01 9.27871615e-02
6.48847073e-02 -2.33208135e-01 -1.11759639e+00 2.00144458e+00
-4.46739346e-01 6.63985193e-01 -5.19855142e-01 -6.19306505e-01
8.72660220e-01 -2.03334317e-02 4.97356772e-01 -4.06898350e-01
9.00424719e-02 2.26818398e-01 -9.27099660e-02 -8.37832540e-02
8.52616131e-01 -4.63259071e-01 2.84991324e-01 6.11489058e-01
1.95064738e-01 -6.24586284e-01 7.56420493e-01 4.81278986e-01
1.08947265e+00 9.24075097e-02 -1.48386788e-03 -2.33180895e-01
2.67653108e-01 -1.21700615e-01 6.37987480e-02 1.02431488e+00
9.70801339e-02 1.07867885e+00 2.33010724e-01 -1.24351820e-02
-1.36805248e+00 -1.28804421e+00 -5.33127598e-02 1.09958684e+00
1.50228724e-01 -5.67122698e-01 -8.48265886e-01 -4.56815094e-01
-1.19394787e-01 1.03149784e+00 -6.79225206e-01 -4.34951663e-01
-2.35429108e-01 -7.03157365e-01 5.70242405e-01 4.41072077e-01
2.73330003e-01 -8.58793497e-01 -5.11266232e-01 2.58908033e-01
-8.10297281e-02 -1.06404328e+00 -7.26148486e-01 4.60419897e-03
-8.38747323e-01 -7.73849189e-01 -1.51053417e+00 -5.59256136e-01
8.15170109e-01 4.21338558e-01 1.17941689e+00 3.16592082e-02
-3.87002528e-01 5.40874779e-01 -5.96085608e-01 -3.14270884e-01
-5.42846143e-01 -1.51109798e-02 -1.40976340e-01 1.24102114e-02
1.53785929e-01 -4.58836228e-01 -7.20323324e-01 2.58274585e-01
-1.26883900e+00 1.41141534e-01 5.39692461e-01 8.30495000e-01
4.53455091e-01 9.84024033e-02 5.45095384e-01 -7.83057034e-01
1.07973385e+00 -4.89303261e-01 -2.88071722e-01 2.02084556e-01
-8.12291145e-01 3.55689436e-01 4.34423923e-01 -5.56334615e-01
-1.00784397e+00 -1.67673692e-01 3.66002619e-02 -4.93206233e-01
1.22653179e-01 2.58682847e-01 4.17020798e-01 3.48127484e-02
1.04208446e+00 2.59416670e-01 1.73317213e-02 -2.01768652e-01
1.03413069e+00 2.93930531e-01 5.57571054e-01 -6.74645722e-01
9.64372277e-01 8.50669205e-01 6.70905262e-02 -7.99569011e-01
-7.26333201e-01 -5.61928034e-01 -6.30505204e-01 -7.22313002e-02
8.18268597e-01 -7.44527638e-01 -8.48100185e-02 4.40123439e-01
-9.01480258e-01 -4.50622737e-01 -8.62130046e-01 2.14124590e-01
-8.22701097e-01 5.23495197e-01 -5.33576548e-01 -5.69190443e-01
-3.42154592e-01 -1.03901112e+00 1.06567013e+00 4.07580473e-02
-3.00100684e-01 -1.05320692e+00 1.54640600e-01 3.48154083e-02
6.08601809e-01 2.27486625e-01 7.30968535e-01 -2.70434171e-01
-6.64359152e-01 3.64531316e-02 -1.88392356e-01 4.95340496e-01
-1.43072789e-03 -2.93285325e-02 -6.91245079e-01 -3.28116924e-01
-1.12901576e-01 -2.49265447e-01 1.30778050e+00 4.40909088e-01
9.48272586e-01 -1.25664502e-01 -1.34599626e-01 5.71568549e-01
1.50719631e+00 -8.17077234e-02 7.12469876e-01 1.96706459e-01
5.17421603e-01 3.83007556e-01 5.10732293e-01 6.59676194e-01
1.14845186e-01 5.72479367e-01 1.68512523e-01 3.50731462e-02
-4.92925525e-01 -2.53037691e-01 3.55613500e-01 7.56487846e-01
-9.94786918e-02 -2.97532707e-01 -7.83161700e-01 8.38879049e-01
-1.70675707e+00 -1.26353288e+00 7.59061873e-02 2.27901983e+00
1.30574358e+00 2.46302605e-01 1.15187950e-01 1.43115923e-01
7.26566434e-01 1.66000798e-01 -3.63606095e-01 -3.71284962e-01
-2.23940313e-01 7.05479920e-01 5.60831130e-01 5.00944555e-01
-8.23661864e-01 1.11681879e+00 6.46646261e+00 1.23171616e+00
-9.21634197e-01 3.19821090e-01 6.54283345e-01 -2.64302284e-01
-6.21549010e-01 6.93938648e-03 -7.65086114e-01 5.00016689e-01
7.49568224e-01 -4.22143161e-01 4.97607350e-01 8.12699854e-01
1.13003284e-01 -2.81548798e-01 -1.02443647e+00 1.05181515e+00
4.10582334e-01 -1.75021076e+00 2.61815280e-01 2.88275443e-02
1.41783082e+00 -1.09727286e-01 3.00722510e-01 1.48096532e-01
4.24466074e-01 -1.10889530e+00 9.08598602e-01 5.25891781e-01
1.00590515e+00 -8.36306572e-01 3.51961613e-01 1.12324744e-01
-9.81918871e-01 1.95108950e-02 -5.19214034e-01 1.03805453e-01
2.47452483e-01 7.26549745e-01 -9.67085183e-01 1.57103583e-01
3.34659696e-01 5.70971251e-01 -7.08450496e-01 1.25740182e+00
-2.21900046e-01 4.18206841e-01 -2.10069001e-01 1.49903051e-03
4.20837194e-01 -1.43944308e-01 4.70089734e-01 1.22578990e+00
6.70965910e-01 -1.79315001e-01 -8.43859613e-02 1.19627047e+00
-1.44353464e-01 1.29534036e-01 -5.59933960e-01 -5.15495092e-02
3.37620437e-01 1.02026474e+00 -1.30341113e+00 -5.76414883e-01
-2.09118515e-01 1.30171239e+00 -6.96338564e-02 3.67702007e-01
-9.71809745e-01 -4.98926461e-01 4.76126850e-01 2.82869279e-01
5.45635521e-01 -4.23451483e-01 -3.85792464e-01 -1.06154859e+00
-1.19396538e-01 -8.13223302e-01 1.36870205e-01 -7.78021753e-01
-1.37505805e+00 4.69657481e-01 2.25685209e-01 -1.03300440e+00
-3.58253598e-01 -3.22591752e-01 -6.55529737e-01 7.08200455e-01
-1.50649989e+00 -9.63623047e-01 -5.13200223e-01 5.70216537e-01
8.17953289e-01 -2.03209698e-01 6.52747691e-01 1.65743768e-01
-1.54966041e-02 4.48423505e-01 2.98811018e-01 -2.44937643e-01
8.80846739e-01 -1.34744370e+00 7.34794199e-01 9.74279225e-01
5.46969056e-01 6.31451368e-01 8.05304408e-01 -6.90522254e-01
-8.44013155e-01 -9.65370774e-01 5.59857845e-01 -5.15708029e-01
5.36185622e-01 -3.04433525e-01 -5.63301802e-01 2.09412009e-01
2.71543294e-01 -9.44772735e-02 5.60613573e-01 -1.24279745e-01
-3.41567636e-01 1.64875448e-01 -1.06132960e+00 9.04169023e-01
1.20872486e+00 -4.04935986e-01 -2.63033956e-01 4.53087747e-01
6.95594847e-01 -2.53974199e-01 -6.47154212e-01 1.56315625e-01
3.46502036e-01 -1.14951360e+00 1.20459962e+00 -3.18061918e-01
7.37282395e-01 -1.66174933e-01 1.94708575e-02 -1.46039867e+00
-2.69401222e-01 -9.63084280e-01 -1.82486534e-01 1.13213766e+00
3.00112545e-01 -2.72286028e-01 8.97607386e-01 3.80075127e-01
-3.89758572e-02 -9.57307041e-01 -6.31313324e-01 -1.07631171e+00
2.55213052e-01 -4.87988174e-01 4.72776860e-01 7.19190419e-01
-2.77994752e-01 2.84931242e-01 -3.49660814e-01 -5.34519434e-01
6.34884298e-01 -2.11119533e-01 8.02920520e-01 -9.98423517e-01
-1.59259871e-01 -7.78214514e-01 -3.54842484e-01 -1.17008829e+00
-2.18406841e-01 -9.77386236e-01 3.84259112e-02 -1.63902771e+00
1.27328455e-01 -5.94913840e-01 -1.85618907e-01 2.40107775e-02
-2.46662855e-01 5.55283904e-01 3.49722266e-01 3.03582489e-01
-4.93529856e-01 7.50637531e-01 1.38345420e+00 -1.37217775e-01
-1.54780284e-01 -1.70110434e-01 -7.21214175e-01 4.77918655e-01
8.84024978e-01 -5.03201425e-01 -7.33831465e-01 -2.95317978e-01
3.51088434e-01 -6.00578368e-01 3.08122933e-01 -1.31346822e+00
2.50049263e-01 -1.49048626e-01 2.90738136e-01 -3.23666304e-01
4.62948352e-01 -4.88776565e-01 -1.71968699e-01 2.92459905e-01
-5.06517708e-01 -2.64420025e-02 -1.77887708e-01 7.18582213e-01
-2.13573631e-02 -6.00681126e-01 8.67367566e-01 -3.80847871e-01
-7.34399736e-01 3.05334300e-01 -1.78074390e-01 5.25870144e-01
1.25387526e+00 -5.53598583e-01 -1.63045958e-01 -6.48378909e-01
-5.51310837e-01 -1.55826330e-01 6.88677907e-01 6.33708894e-01
7.43048489e-01 -1.46669054e+00 -8.00775468e-01 1.38366297e-01
1.51538610e-01 -1.78739011e-01 9.18475969e-04 6.95958912e-01
-6.52278721e-01 1.16252169e-01 -2.29776770e-01 -6.26865029e-01
-8.92336309e-01 6.04184330e-01 -2.47813836e-02 -1.56936496e-01
-5.95434070e-01 1.25096273e+00 8.09022635e-02 1.26870722e-01
2.12580144e-01 -4.02801871e-01 5.39059937e-02 2.97055334e-01
4.72859204e-01 4.59435582e-01 -2.19170362e-01 -5.71084857e-01
1.05064549e-01 5.46003819e-01 -9.19864103e-02 -5.23951411e-01
1.32947540e+00 -3.52694429e-02 1.11084320e-01 3.68815869e-01
1.15967977e+00 1.06503099e-01 -1.37079561e+00 -1.50704741e-01
3.39293517e-02 -7.59343922e-01 -7.62858242e-02 -6.61617041e-01
-9.33706641e-01 9.07006741e-01 6.06944382e-01 2.17736751e-01
8.66006374e-01 -5.19225411e-02 1.07185829e+00 1.17524303e-01
2.83074886e-01 -1.46890938e+00 7.59565830e-01 4.06497478e-01
9.47151065e-01 -1.08429074e+00 1.60261437e-01 -2.54924268e-01
-7.21464396e-01 9.89615321e-01 3.02985251e-01 -9.42697674e-02
5.41351199e-01 7.74486065e-02 -8.89667124e-02 -6.42637387e-02
-6.12723708e-01 -4.40039963e-01 2.63462901e-01 7.96098590e-01
3.43856514e-01 -9.75975990e-02 -3.64947200e-01 -1.24197314e-02
-3.31943661e-01 8.45399052e-02 7.05835044e-01 8.12282681e-01
-5.76197982e-01 -1.33972239e+00 -2.34929949e-01 4.84283060e-01
-3.96278769e-01 -4.72232044e-01 -3.82025018e-02 4.83111471e-01
3.64258111e-01 8.14309776e-01 -1.44041963e-02 -2.22971421e-02
-6.56789169e-02 6.88156039e-02 6.55124664e-01 -8.96763206e-01
-3.93115044e-01 -9.70815495e-02 -2.36911789e-01 -3.00588012e-01
-3.81178707e-01 -5.61091185e-01 -1.06002676e+00 -2.77211457e-01
-5.31372070e-01 8.33177492e-02 7.50433743e-01 6.96494460e-01
3.81638616e-01 4.01506633e-01 4.54706967e-01 -9.32642817e-01
-7.48812199e-01 -8.35166276e-01 -4.32731658e-01 7.81614125e-01
-4.43282984e-02 -8.00544500e-01 -3.23621333e-01 3.05585802e-01] | [11.226889610290527, -0.18234364688396454] |
ebf51e98-12bb-4050-b5e6-e3851cdcdfc1 | neural-representations-reveal-distinct-modes | 2212.00771 | null | https://arxiv.org/abs/2212.00771v1 | https://arxiv.org/pdf/2212.00771v1.pdf | Neural Representations Reveal Distinct Modes of Class Fitting in Residual Convolutional Networks | We leverage probabilistic models of neural representations to investigate how residual networks fit classes. To this end, we estimate class-conditional density models for representations learned by deep ResNets. We then use these models to characterize distributions of representations across learned classes. Surprisingly, we find that classes in the investigated models are not fitted in an uniform way. On the contrary: we uncover two groups of classes that are fitted with markedly different distributions of representations. These distinct modes of class-fitting are evident only in the deeper layers of the investigated models, indicating that they are not related to low-level image features. We show that the uncovered structure in neural representations correlate with memorization of training examples and adversarial robustness. Finally, we compare class-conditional distributions of neural representations between memorized and typical examples. This allows us to uncover where in the network structure class labels arise for memorized and standard inputs. | ['Marcin Kurdziel', 'Michał Jamroż'] | 2022-12-01 | null | null | null | null | ['memorization'] | ['natural-language-processing'] | [ 4.34689641e-01 3.03272277e-01 -2.07137186e-02 -3.16961288e-01
-5.20858347e-01 -8.77771020e-01 1.11828971e+00 1.72900334e-01
-3.26561570e-01 8.04857314e-01 2.25900769e-01 -6.75929710e-02
-5.57870448e-01 -8.32500756e-01 -1.00762546e+00 -8.85095596e-01
-4.32097949e-02 1.23514093e-01 2.34454617e-01 -1.74511120e-01
4.06717658e-01 6.52824581e-01 -1.77157128e+00 6.87062204e-01
6.14701927e-01 9.08252239e-01 -1.77142069e-01 3.45761806e-01
-5.78838848e-02 9.28426921e-01 -7.61134624e-01 -2.89408892e-01
2.00541839e-01 -2.52374828e-01 -8.21663618e-01 -5.79568669e-02
7.49283254e-01 1.55319080e-01 -4.64861453e-01 1.14606857e+00
8.80089682e-03 1.20026782e-01 1.54824030e+00 -8.17355037e-01
-7.98512101e-01 6.27615929e-01 -4.37248349e-01 3.37104231e-01
2.07510874e-01 3.12026255e-02 9.63627160e-01 -7.59947717e-01
6.94959521e-01 1.14045823e+00 6.73124433e-01 5.63696682e-01
-1.73711884e+00 -5.76705396e-01 1.30499303e-01 6.25752052e-03
-1.28334725e+00 -5.20449579e-01 6.70139849e-01 -8.49698365e-01
6.04615390e-01 7.16513544e-02 3.59023362e-01 1.34903085e+00
4.77093995e-01 4.81076449e-01 1.44528008e+00 -4.28934157e-01
2.66479164e-01 4.37711746e-01 7.06418931e-01 5.31226158e-01
4.19308871e-01 1.76514924e-01 -4.16997552e-01 -4.90249135e-02
5.68418264e-01 2.06044823e-01 -3.98900598e-01 -4.41994786e-01
-8.82411838e-01 8.06957245e-01 6.47985697e-01 6.15670681e-01
-2.52244443e-01 2.61447281e-01 2.50074238e-01 1.60371795e-01
4.99232471e-01 3.48663598e-01 -4.74200130e-01 4.87816334e-01
-8.90008330e-01 -4.21900488e-02 4.89268988e-01 8.18278134e-01
1.18474102e+00 2.30684802e-01 -2.26424739e-01 8.63199055e-01
-3.35815847e-02 1.14417873e-01 6.73491836e-01 -6.83535397e-01
1.80036023e-01 3.61934692e-01 -3.12331498e-01 -9.36385095e-01
-2.88642317e-01 -7.99600363e-01 -9.70543861e-01 3.76909047e-01
6.42174006e-01 3.18049788e-01 -1.01813018e+00 1.97977269e+00
-4.09660667e-01 -2.54521589e-03 2.14828596e-01 1.66062281e-01
7.98754275e-01 3.51623684e-01 2.73300320e-01 3.33443470e-02
1.31571388e+00 -4.21603203e-01 -3.50920588e-01 -1.36333033e-01
5.19189715e-01 -1.66784793e-01 9.05852079e-01 1.51157022e-01
-1.05520737e+00 -7.28592515e-01 -9.95319545e-01 2.99720168e-01
-5.49963117e-01 -2.20129136e-02 4.33723569e-01 7.79746115e-01
-9.72881973e-01 1.21596372e+00 -4.52682018e-01 -2.34378576e-01
7.50466228e-01 2.64663368e-01 -5.57627439e-01 4.06223796e-02
-9.91891503e-01 8.05352569e-01 4.12024945e-01 -2.23569840e-01
-1.16076291e+00 -5.73303223e-01 -9.14509535e-01 1.51624411e-01
-2.31917888e-01 -4.84824240e-01 8.22141349e-01 -1.18192768e+00
-1.04952097e+00 1.26499844e+00 -3.41785140e-02 -5.49795330e-01
2.19105005e-01 4.44879904e-02 -2.27368876e-01 9.05289799e-02
-2.47311860e-01 5.55721045e-01 1.25030506e+00 -1.57692444e+00
-1.95578244e-02 -2.50098884e-01 -2.96476157e-03 -5.03934383e-01
-2.59988397e-01 -3.28162372e-01 3.45953584e-01 -7.69811511e-01
3.18266451e-01 -8.00305307e-01 -3.41568305e-03 -2.52483934e-01
-7.35346079e-01 -1.34724855e-01 1.15852773e-01 -1.65212274e-01
7.91708946e-01 -2.16965103e+00 5.45423590e-02 4.66663629e-01
2.88494885e-01 9.80426837e-03 -1.46780148e-01 2.24916026e-01
-6.23071074e-01 3.64316553e-01 -4.21234518e-01 -3.63018572e-01
2.56105781e-01 3.36824954e-01 -9.02314007e-01 8.40222239e-01
3.80293190e-01 7.96969771e-01 -3.56225938e-01 -5.69109470e-02
1.13698967e-01 4.73087430e-01 -6.32554054e-01 1.73134416e-01
-3.51167656e-02 4.88328397e-01 -9.61329713e-02 3.49186957e-01
9.65447605e-01 -8.58994722e-02 7.03337863e-02 -1.29921868e-01
1.18476667e-01 4.22635078e-01 -6.64909720e-01 1.31083131e+00
-2.57515192e-01 7.16794670e-01 -3.74832153e-01 -1.41726041e+00
8.81204963e-01 -2.64728535e-03 -2.13504896e-01 -5.42335451e-01
1.88118517e-01 7.20221549e-02 1.67411551e-01 -1.69753745e-01
4.69529510e-01 -3.48753780e-01 -1.08266205e-01 5.55995822e-01
7.29852080e-01 8.80541876e-02 -3.09000909e-01 -5.90426065e-02
9.27031517e-01 6.24650717e-02 1.96843430e-01 -7.00809658e-01
2.48104692e-01 -5.22010505e-01 2.72646338e-01 1.14735901e+00
1.78637709e-02 7.14289427e-01 7.30643690e-01 -3.91618699e-01
-7.46852279e-01 -1.34425139e+00 -6.66693091e-01 1.19651425e+00
-1.19955510e-01 -1.47093356e-01 -8.03040385e-01 -5.27896047e-01
-1.28227070e-01 6.49014413e-01 -1.16446376e+00 -6.52988374e-01
-2.07166344e-01 -7.08338618e-01 6.27395570e-01 4.69397992e-01
7.75372833e-02 -1.37073398e+00 -5.21199644e-01 -2.73089737e-01
3.83090645e-01 -8.20863545e-01 3.28024447e-01 7.29573190e-01
-8.94768715e-01 -1.04968071e+00 -7.28482544e-01 -6.40245497e-01
8.48877132e-01 -3.11459322e-02 1.43378830e+00 1.19106390e-01
-2.72427917e-01 5.44371009e-01 -7.98071027e-02 -5.13045788e-02
-4.89011109e-01 1.99080080e-01 2.62261152e-01 -1.99415833e-02
-8.66128504e-02 -8.70944917e-01 -4.79276270e-01 2.73203284e-01
-1.14902043e+00 -5.88780284e-01 5.13723195e-01 6.70888960e-01
4.03794229e-01 2.88637578e-02 5.49817741e-01 -1.03271306e+00
5.06436527e-01 -8.82104993e-01 -3.20280224e-01 2.28289023e-01
-2.58625835e-01 5.02349198e-01 6.85154140e-01 -7.13511944e-01
-7.96363533e-01 -1.51915416e-01 -1.34552687e-01 -2.61105418e-01
-7.91841328e-01 2.53521532e-01 8.05438459e-02 -1.12575799e-01
9.60456014e-01 4.01946008e-01 -3.51424575e-01 -4.12322789e-01
4.10289228e-01 1.70471027e-01 5.09296954e-01 -8.78623009e-01
9.15438533e-01 4.06402886e-01 -1.38362944e-01 -8.24886501e-01
-1.01649070e+00 9.34727117e-03 -5.50213933e-01 7.35441595e-02
8.90541613e-01 -6.44179821e-01 -5.29188991e-01 3.48888844e-01
-1.13501859e+00 -3.75101835e-01 -6.54502988e-01 2.88706809e-01
-7.63270020e-01 1.88159734e-01 -6.59555614e-01 -9.84613419e-01
1.44268751e-01 -1.16940117e+00 6.48213744e-01 2.16270536e-01
-3.59318227e-01 -1.08372509e+00 2.22842619e-01 -2.00251058e-01
3.47087741e-01 1.71381325e-01 1.27400577e+00 -9.91577148e-01
-4.03263718e-01 -2.09547013e-01 -2.62456596e-01 2.74343729e-01
8.63574892e-02 9.59572271e-02 -1.56087899e+00 -1.70957133e-01
1.34284168e-01 -4.89250898e-01 1.65574741e+00 5.14166534e-01
1.42529333e+00 -3.29129189e-01 -2.46619925e-01 7.52813101e-01
1.28568292e+00 -8.35232511e-02 9.33879495e-01 3.01048994e-01
2.86622614e-01 1.08123589e+00 -8.69578123e-02 1.39947861e-01
-2.22893208e-01 4.61356342e-01 5.07606745e-01 2.80016392e-01
6.83024824e-02 -3.96928340e-01 3.98922890e-01 3.53570193e-01
-1.09199442e-01 -8.73160586e-02 -7.84384966e-01 6.39190376e-01
-1.48757648e+00 -1.16839039e+00 -1.36276865e-02 2.12477350e+00
7.88901925e-01 5.19460022e-01 -3.21385413e-02 2.10677341e-01
8.28049302e-01 2.21168905e-01 -3.52299899e-01 -2.80582398e-01
-3.01954865e-01 7.11887956e-01 3.22378397e-01 2.74371177e-01
-9.73002315e-01 8.12909782e-01 7.61499357e+00 1.04819870e+00
-8.04336488e-01 -6.15938120e-02 7.75751889e-01 1.82860777e-01
-6.15961850e-01 8.50667991e-03 -6.40715361e-01 2.85868526e-01
9.64280188e-01 3.08667094e-01 1.96664155e-01 7.28802562e-01
-6.66805744e-01 1.06261268e-01 -1.23588336e+00 7.22795546e-01
-1.96580306e-01 -1.38939190e+00 1.93825573e-01 1.24214172e-01
8.55183840e-01 -2.07320094e-01 6.35215998e-01 4.66898948e-01
3.29015672e-01 -1.45655560e+00 8.52390647e-01 1.06240535e+00
7.35315084e-01 -8.46874237e-01 4.47319835e-01 3.15576881e-01
-7.96512127e-01 -1.43442020e-01 -8.90027881e-01 7.88530428e-03
-4.65531915e-01 6.04908466e-01 -4.21822816e-01 1.40732542e-01
6.58296824e-01 6.96156144e-01 -7.81063855e-01 7.77483225e-01
-1.72049031e-01 7.06139386e-01 1.81479514e-01 3.70302349e-01
-1.36738688e-01 -7.61922123e-03 2.45610923e-01 1.26072407e+00
2.33301580e-01 -3.83242905e-01 -3.43815297e-01 1.31493270e+00
-1.95260629e-01 -3.13632548e-01 -9.20313597e-01 1.29417926e-01
2.47585163e-01 1.20739341e+00 -8.94584715e-01 -1.41233459e-01
9.92725715e-02 8.33127320e-01 7.22109318e-01 4.02747303e-01
-7.35350370e-01 -3.26750278e-01 7.32956171e-01 1.05171397e-01
3.33031356e-01 -1.97715178e-01 -2.69869417e-01 -1.16429436e+00
-2.08348319e-01 -4.75893706e-01 8.40349570e-02 -5.83043694e-01
-1.64854562e+00 8.69931877e-01 1.44173920e-01 -1.09207273e+00
-3.02865386e-01 -8.14448535e-01 -9.19993997e-01 6.95472896e-01
-1.52737021e+00 -7.79996037e-01 -1.17097758e-01 8.34646463e-01
7.37247765e-02 -3.83331507e-01 1.02959418e+00 -1.81786343e-01
-5.46337068e-01 8.82802606e-01 2.23780245e-01 1.64644659e-01
4.90433753e-01 -1.20840323e+00 3.25953573e-01 4.69105333e-01
5.23577213e-01 1.11800265e+00 6.92892075e-01 -2.35074073e-01
-6.96180701e-01 -1.00939667e+00 4.35251594e-01 -6.12250209e-01
6.50354803e-01 -5.29031873e-01 -1.24820113e+00 5.99372029e-01
2.82579184e-01 3.29692602e-01 9.19051349e-01 2.24045232e-01
-9.71592486e-01 2.29011819e-01 -1.22504485e+00 3.18279415e-01
1.03351402e+00 -1.10176134e+00 -9.03418005e-01 2.42387787e-01
6.21606588e-01 3.34459186e-01 -6.39373779e-01 2.71319896e-01
4.01860237e-01 -1.38212526e+00 1.06784987e+00 -9.59010303e-01
6.16095781e-01 -8.24971404e-03 -3.98334265e-01 -1.43549681e+00
-4.68674809e-01 -2.26762995e-01 1.45108327e-01 1.32541776e+00
6.11292064e-01 -8.18670869e-01 6.59666359e-01 5.33758044e-01
-1.04430139e-01 -4.78501767e-01 -9.45932746e-01 -7.67052174e-01
6.38643086e-01 -5.40016234e-01 3.54687274e-01 1.00699484e+00
-4.71682139e-02 3.71972402e-03 -8.62864405e-02 -1.32589757e-01
6.65356696e-01 8.86732638e-02 2.67278671e-01 -1.38478649e+00
-3.42457294e-01 -5.55546105e-01 -6.87886417e-01 -7.21363246e-01
8.74197841e-01 -1.27338171e+00 -7.03965202e-02 -1.05730677e+00
3.95749867e-01 -6.29761934e-01 -6.69532001e-01 3.94615889e-01
6.75610900e-02 5.68182766e-01 3.89592424e-02 4.18414205e-01
-5.11350453e-01 5.66631496e-01 6.38523519e-01 -1.13815881e-01
2.03471392e-01 5.71557693e-02 -7.98157156e-01 9.99413013e-01
1.07546830e+00 -6.38931453e-01 -2.75508702e-01 -2.35400811e-01
4.13030446e-01 -3.47931802e-01 7.30692625e-01 -1.23914039e+00
-7.10702538e-02 1.88542113e-01 7.64261365e-01 -3.94232810e-01
2.14228138e-01 -7.00778961e-01 -4.65254812e-03 3.67665470e-01
-7.94406652e-01 -3.36120605e-01 2.78133452e-01 6.61492467e-01
-5.18424809e-02 -6.83589518e-01 1.03564227e+00 -1.68635562e-01
-1.51981786e-01 1.02302574e-01 -8.87301743e-01 1.46638215e-01
7.55223811e-01 -3.47665340e-01 -5.10659516e-01 -3.06114793e-01
-1.17627144e+00 -5.63466787e-01 3.64075899e-01 2.39379308e-03
5.78234971e-01 -1.36429155e+00 -4.67948943e-01 2.74728477e-01
1.35526493e-01 -2.88421184e-01 3.12538266e-01 4.50316876e-01
-6.30294811e-03 1.70467764e-01 -5.67087591e-01 -6.77278936e-01
-6.87250078e-01 6.37168288e-01 4.72843885e-01 -2.00818628e-01
-1.46707356e-01 9.38558221e-01 7.30927527e-01 -5.46263099e-01
1.03261732e-01 -5.22262394e-01 -4.86299664e-01 3.72638822e-01
4.70950514e-01 6.35715649e-02 -8.54162425e-02 -6.55466676e-01
-2.34875798e-01 7.17461169e-01 -3.35372329e-01 2.12575555e-01
1.40739810e+00 2.35751957e-01 -9.85693485e-02 6.69420779e-01
1.30932903e+00 6.48351312e-02 -1.31416810e+00 -1.35730997e-01
1.96677178e-01 -3.58281471e-02 -5.59133232e-01 -3.01493526e-01
-1.14022183e+00 1.14590800e+00 6.08671904e-01 4.76007372e-01
8.25179815e-01 3.70036453e-01 3.63385342e-02 3.36123347e-01
3.82629305e-01 -5.44926882e-01 1.94740042e-01 8.03936422e-01
9.47413802e-01 -8.90069485e-01 -1.16229013e-01 -7.20168352e-02
-2.57393777e-01 1.17202747e+00 3.10566306e-01 -7.88045704e-01
7.46637344e-01 -8.28239485e-04 -4.76072550e-01 -2.18097642e-01
-6.02669239e-01 -2.21470952e-01 5.28793752e-01 8.85786533e-01
4.30563211e-01 -4.28621024e-02 2.47414231e-01 7.36344993e-01
-4.49920177e-01 -7.32531130e-01 5.25816262e-01 5.62896788e-01
-4.88345623e-01 -8.57894838e-01 -2.42297739e-01 3.71500582e-01
-5.20131111e-01 -2.27739692e-01 -4.74760205e-01 6.31457448e-01
1.15261592e-01 6.21270061e-01 1.91890061e-01 -6.08450949e-01
2.09145382e-01 3.03760618e-01 7.40260303e-01 -7.71964967e-01
-5.94966412e-01 -4.27160114e-01 -2.73410499e-01 -2.77388155e-01
-3.46431673e-01 -3.32284629e-01 -9.06084776e-01 -1.32179528e-03
-3.15058142e-01 4.38075028e-02 3.79698694e-01 8.82913351e-01
4.51769046e-02 6.58001006e-01 4.28661585e-01 -1.30505204e+00
-6.21016145e-01 -9.41396356e-01 -7.57205546e-01 5.78549206e-01
5.50152838e-01 -7.50851095e-01 -7.73273587e-01 -1.12704381e-01] | [9.556265830993652, 2.697054147720337] |
5ada126c-240f-429f-8f2e-02dec58583fd | learned-tree-search-for-long-horizon-social | 2304.01428 | null | https://arxiv.org/abs/2304.01428v1 | https://arxiv.org/pdf/2304.01428v1.pdf | Learned Tree Search for Long-Horizon Social Robot Navigation in Shared Airspace | The fast-growing demand for fully autonomous aerial operations in shared spaces necessitates developing trustworthy agents that can safely and seamlessly navigate in crowded, dynamic spaces. In this work, we propose Social Robot Tree Search (SoRTS), an algorithm for the safe navigation of mobile robots in social domains. SoRTS aims to augment existing socially-aware trajectory prediction policies with a Monte Carlo Tree Search planner for improved downstream navigation of mobile robots. To evaluate the performance of our method, we choose the use case of social navigation for general aviation. To aid this evaluation, within this work, we also introduce X-PlaneROS, a high-fidelity aerial simulator, to enable more research in full-scale aerial autonomy. By conducting a user study based on the assessments of 26 FAA certified pilots, we show that SoRTS performs comparably to a competent human pilot, significantly outperforming our baseline algorithm. We further complement these results with self-play experiments in scenarios with increasing complexity. | ['Jean Oh', 'Sebastian Scherer', 'Ian Higgins', 'Rohan Baijal', 'Joao P. A. Dantas', 'Jay Patrikar', 'Ingrid Navarro'] | 2023-04-04 | null | null | null | null | ['trajectory-prediction', 'social-navigation', 'robot-navigation'] | ['computer-vision', 'robots', 'robots'] | [-2.12921008e-01 3.85339737e-01 3.25472683e-01 -2.81123519e-01
-2.77453274e-01 -8.86959195e-01 4.83184457e-01 -2.48145938e-01
-6.36201024e-01 1.12836754e+00 6.63295612e-02 -6.33649051e-01
-5.11056244e-01 -4.86307353e-01 -3.58346760e-01 -3.44051003e-01
-7.56231725e-01 3.90958607e-01 4.68625605e-01 -8.91599715e-01
-3.84662785e-02 4.26438451e-01 -1.77030551e+00 -1.93598732e-01
1.01496947e+00 4.61956620e-01 1.03452906e-01 8.94916356e-01
1.17941141e+00 8.37317467e-01 -3.82442415e-01 -2.41142735e-01
6.62278593e-01 -2.93890566e-01 -9.27740157e-01 -2.73811549e-01
-1.27046481e-01 -6.96234822e-01 -1.45006031e-01 6.42610550e-01
6.61373556e-01 7.63001084e-01 4.84024525e-01 -1.87106335e+00
2.09196895e-01 5.78114808e-01 2.01498136e-01 -1.54689983e-01
7.52847731e-01 6.25039697e-01 8.84548962e-01 -3.31148416e-01
7.31042087e-01 1.12091029e+00 8.20819676e-01 3.42434615e-01
-6.65683210e-01 -6.65416896e-01 9.05646086e-02 -1.16005465e-01
-1.46055794e+00 -4.48806494e-01 -1.41691893e-01 -4.47329074e-01
9.95655656e-01 6.50729910e-02 8.52428913e-01 1.09493101e+00
4.39234048e-01 5.51780939e-01 9.96522367e-01 2.18801588e-01
6.49306118e-01 -2.29014784e-01 -5.98586321e-01 1.01773369e+00
4.99356657e-01 5.02682626e-01 -4.38759774e-01 -5.06835938e-01
1.19155400e-01 -5.16170382e-01 -3.54137957e-01 -8.65826130e-01
-1.35697699e+00 7.71721542e-01 6.85277462e-01 -1.12509385e-01
-5.27927756e-01 2.79488683e-01 -2.37926617e-02 3.24873567e-01
5.40091358e-02 9.35295999e-01 -3.01349998e-01 -5.74183047e-01
-5.96368790e-01 9.35099483e-01 1.30514145e+00 1.28977287e+00
2.76048094e-01 5.62374443e-02 2.31068060e-01 1.69628277e-01
6.43753529e-01 7.01140583e-01 2.56112665e-01 -1.70550537e+00
3.02681983e-01 3.66036773e-01 1.04948640e+00 -1.13823521e+00
-7.15437710e-01 -2.47875527e-01 -1.37712481e-02 9.71297503e-01
3.21847141e-01 -7.89366603e-01 -3.43863398e-01 1.39255738e+00
6.80368125e-01 -1.36520162e-01 5.94898343e-01 1.05038655e+00
2.59483159e-01 3.90790522e-01 -7.36640245e-02 1.94593504e-01
7.75409997e-01 -1.17452633e+00 -3.49567324e-01 -2.89656937e-01
1.19122493e+00 -2.24731877e-01 8.01435888e-01 4.03737485e-01
-7.92024314e-01 3.95439826e-02 -1.26996183e+00 4.59601611e-01
-2.53723055e-01 -1.01440050e-01 3.95230681e-01 6.84977710e-01
-1.35711944e+00 9.14615154e-01 -9.31681991e-01 -9.64435816e-01
4.24822778e-01 6.31823957e-01 -4.86975878e-01 3.82261956e-03
-9.64292884e-01 1.08142507e+00 -6.65505007e-02 -2.39090815e-01
-1.39353502e+00 1.14061413e-02 -9.31533277e-01 -4.19054687e-01
5.30154645e-01 -8.79065216e-01 1.59528041e+00 -2.39186570e-01
-1.57596231e+00 4.77710634e-01 3.21759909e-01 -8.00767779e-01
6.66831076e-01 -3.69320720e-01 8.66682157e-02 -1.73523463e-02
7.36091256e-01 8.80769372e-01 4.80955780e-01 -1.47629249e+00
-1.19220090e+00 -1.18901497e-02 5.76145709e-01 9.01915669e-01
1.55600965e-01 -2.12316096e-01 4.26227540e-01 -1.54695898e-01
-2.62049109e-01 -1.56759810e+00 -7.89012134e-01 6.33414974e-03
-1.07567675e-01 1.43332034e-01 4.84249443e-01 -2.97009110e-01
7.11532056e-01 -2.07213473e+00 2.80504614e-01 2.23935381e-01
1.10048205e-01 2.11588275e-02 2.75801285e-04 6.61502302e-01
6.72071099e-01 6.25933930e-02 -3.57932627e-01 -4.22813922e-01
3.99896443e-01 3.22323948e-01 -2.50229239e-01 5.42971969e-01
-4.61434156e-01 6.13880992e-01 -1.43648314e+00 -2.63360649e-01
-1.22852981e-01 -1.64323851e-01 -7.63589442e-01 -4.14439179e-02
6.67365044e-02 4.89567727e-01 -5.44976950e-01 7.97869086e-01
7.69380704e-02 2.59951472e-01 1.24072962e-01 7.26472199e-01
-4.00482863e-01 -1.33198023e-01 -8.84378791e-01 1.73106515e+00
-1.95915729e-01 4.92196023e-01 7.76785135e-01 7.26991966e-02
4.86457676e-01 2.93326024e-02 2.76939362e-01 -1.95947234e-02
4.44093347e-01 4.68747914e-01 6.91845715e-02 -3.73038113e-01
1.05741763e+00 2.80119758e-02 -4.07251745e-01 6.24284923e-01
-2.59674311e-01 -9.07909274e-01 -9.70349386e-02 2.24103197e-01
1.57821393e+00 3.94821644e-01 2.66172618e-01 -4.88859683e-01
7.54080042e-02 7.76877880e-01 2.44820997e-01 9.74522650e-01
-1.02339637e+00 1.01613611e-01 7.22852722e-02 -2.35201031e-01
-4.11788940e-01 -8.28220725e-01 3.20580840e-01 1.12060547e+00
8.34142625e-01 -6.00709736e-01 -9.10477817e-01 -8.57285082e-01
1.55722395e-01 1.09965968e+00 -3.84315908e-01 9.79657620e-02
-2.34743208e-01 -3.48751038e-01 8.01489651e-01 1.83575209e-02
4.92299646e-01 -7.18294382e-01 -1.48990607e+00 6.30351203e-03
-3.68677408e-01 -1.16137707e+00 -3.87186818e-02 3.83185148e-02
-3.26578200e-01 -1.13222158e+00 -3.39730352e-01 -5.66286922e-01
4.78641838e-01 7.58743525e-01 5.32516241e-01 2.95099437e-01
2.92019919e-02 1.00450754e+00 -7.17937171e-01 -5.92509031e-01
-4.19933140e-01 2.09241763e-01 8.55583072e-01 -7.06951499e-01
-1.66307330e-01 -4.68088448e-01 -6.27704680e-01 9.11766112e-01
-3.05362880e-01 -1.68211728e-01 1.19368508e-01 5.30443013e-01
-1.33272052e-01 4.94872749e-01 3.51328075e-01 -7.79613182e-02
9.55102980e-01 -8.10540855e-01 -6.74564183e-01 -2.12138548e-01
-5.48227370e-01 -3.93376380e-01 3.83988887e-01 1.98271140e-01
-5.94113946e-01 1.44361064e-01 2.30427817e-01 2.50907242e-01
-1.69451952e-01 4.38184857e-01 3.28163117e-01 -6.51652098e-01
1.00318682e+00 -3.83263052e-01 2.35681295e-01 4.84649748e-01
3.77991498e-01 7.90487826e-01 2.42637977e-01 -3.18431973e-01
1.01111543e+00 5.01271605e-01 8.07923079e-02 -8.15008700e-01
-2.07527503e-01 -5.73639087e-02 -5.51973343e-01 -5.35799205e-01
6.97638035e-01 -1.13822806e+00 -7.93420136e-01 2.54502922e-01
-7.77790964e-01 -1.22843564e+00 -4.58788306e-01 5.80643952e-01
-9.64768529e-01 2.56658435e-01 -9.24578235e-02 -1.12160003e+00
2.79442668e-01 -1.26298261e+00 8.68798375e-01 9.66119394e-02
-5.86249173e-01 -6.80214047e-01 3.11567098e-01 5.49150169e-01
4.72342491e-01 3.05841953e-01 7.23074898e-02 -7.01593399e-01
-8.48017991e-01 -1.44278929e-01 3.33149046e-01 -2.67666399e-01
-2.14854822e-01 -2.82635987e-01 -6.10555708e-01 -6.01951003e-01
-2.56224275e-01 -5.99984705e-01 7.07884729e-02 5.11676036e-02
-1.14419006e-01 -2.13121638e-01 -7.53236473e-01 3.25321823e-01
6.27468288e-01 2.67802835e-01 2.18559608e-01 7.73524404e-01
2.39292204e-01 1.17908680e+00 1.68863583e+00 7.24680543e-01
9.67317522e-01 3.88640970e-01 8.65926385e-01 3.18856210e-01
5.11851728e-01 -5.02643228e-01 6.01419270e-01 2.52689540e-01
-2.72959918e-01 -8.06024313e-01 -1.15278447e+00 6.84241116e-01
-2.15722752e+00 -7.24904299e-01 3.64132226e-02 1.87656248e+00
3.25768744e-03 -1.40798792e-01 3.04337740e-01 -9.43746790e-02
3.08633894e-01 -3.19793150e-02 -3.31540734e-01 -7.86907822e-02
1.85384154e-01 -4.64384168e-01 9.99170005e-01 8.44120860e-01
-1.09162998e+00 1.41120160e+00 6.80178404e+00 6.03657186e-01
-2.58109093e-01 -9.39573422e-02 9.54830367e-03 -3.19488972e-01
7.97053128e-02 1.96754113e-01 -5.72850645e-01 -7.41090700e-02
9.42038774e-01 -3.41610491e-01 8.57818007e-01 1.08189404e+00
2.85115123e-01 -7.33427107e-01 -6.83970809e-01 4.49279398e-01
-5.58137670e-02 -7.26690471e-01 -5.90735912e-01 2.85742074e-01
7.90024400e-01 4.40134048e-01 7.77670741e-02 4.19631958e-01
1.33014750e+00 -1.16340697e+00 1.05180550e+00 2.33562380e-01
2.76661366e-01 -7.42599845e-01 9.62313652e-01 8.54205847e-01
-9.60957646e-01 -5.94246268e-01 -2.54424959e-01 -5.37266195e-01
6.08640671e-01 -3.58576417e-01 -1.31687856e+00 6.10571146e-01
1.00394702e+00 5.97126365e-01 -3.61800849e-01 1.11947656e+00
-2.96516061e-01 2.67701954e-01 -5.04838347e-01 -6.51753664e-01
7.01475024e-01 -1.64721280e-01 1.10416865e+00 3.03906202e-01
7.45234907e-01 4.65412289e-01 3.27114135e-01 3.93999219e-01
4.45102274e-01 -6.41536489e-02 -1.56193554e+00 2.05997452e-02
4.74309921e-01 1.20580935e+00 -8.07651401e-01 2.47066826e-01
3.67801368e-01 9.78167236e-01 -1.31884381e-01 2.35179275e-01
-9.46074069e-01 -5.90372562e-01 9.23144281e-01 5.66252321e-02
3.55251953e-02 -8.05985034e-01 1.22350119e-01 -5.08856833e-01
-3.03655297e-01 -8.72202992e-01 -2.24626392e-01 -1.17477524e+00
-6.71025336e-01 1.03893650e+00 1.65861025e-01 -1.60026670e+00
-4.67108876e-01 -4.22071725e-01 -2.20626175e-01 3.98591340e-01
-1.27155769e+00 -1.19094062e+00 -5.78884125e-01 3.18744242e-01
2.38771781e-01 -5.80534637e-01 8.09042692e-01 -4.15013283e-01
-1.11508265e-01 6.33946881e-02 -1.99518114e-01 -6.65701628e-01
7.08665729e-01 -8.87422383e-01 7.11543500e-01 8.64879608e-01
-4.58054066e-01 5.56995153e-01 9.47068989e-01 -1.32877660e+00
-1.18414521e+00 -9.48430896e-01 3.82893980e-01 -1.06432736e+00
7.66088665e-01 -1.36278719e-01 1.17479339e-02 7.60214865e-01
2.16079235e-01 -5.03193557e-01 6.98513806e-01 -1.78612798e-01
6.85910463e-01 4.92596924e-01 -1.51397085e+00 1.19894290e+00
1.59621000e+00 -9.14627090e-02 -3.87333602e-01 3.90645981e-01
1.01311111e+00 -6.11643314e-01 -5.62063158e-01 5.21534324e-01
6.79643869e-01 -1.23027945e+00 6.11503959e-01 -2.42435992e-01
-8.69425852e-03 -8.06081057e-01 -4.79317576e-01 -1.67621434e+00
-2.56354272e-01 -1.16626501e+00 5.84363163e-01 5.49637020e-01
4.74471033e-01 -8.28148842e-01 7.20454037e-01 9.04645503e-01
-3.96498352e-01 -4.50203598e-01 -1.10797358e+00 -9.17909086e-01
-1.52706295e-01 -2.85282433e-01 6.52994514e-01 5.55999875e-01
4.22907770e-01 -7.52490535e-02 -3.62288475e-01 6.65527105e-01
4.38608974e-01 -6.41558170e-01 1.24442804e+00 -1.09364045e+00
5.17209582e-02 -5.00897579e-02 -5.41690350e-01 -9.10874724e-01
5.15144706e-01 -6.43348336e-01 7.61849165e-01 -1.58278537e+00
-7.15312839e-01 -7.84320712e-01 6.64876819e-01 3.38911116e-01
4.11214650e-01 4.92053851e-03 1.63427249e-01 5.97147569e-02
-1.09221113e+00 9.65732098e-01 1.10842133e+00 2.09190458e-01
-3.44768226e-01 3.24889898e-01 -7.08274245e-01 1.04381454e+00
9.16757882e-01 -4.13819045e-01 -7.34300971e-01 -1.88138977e-01
3.18181813e-01 1.79453209e-01 5.11563897e-01 -1.77016938e+00
8.49913955e-01 -2.61476457e-01 -3.53451759e-01 -3.05151492e-01
6.43036366e-01 -1.03453910e+00 1.05818324e-01 7.13095188e-01
-1.60166219e-01 3.04039985e-01 2.11225942e-01 9.34405625e-01
2.27636516e-01 -4.67744231e-01 2.92404711e-01 -8.86306763e-02
-5.65674961e-01 -2.60893963e-02 -1.44015098e+00 -7.50545710e-02
1.76166248e+00 -4.68846381e-01 -3.04003924e-01 -9.50659275e-01
-4.27010089e-01 9.11705911e-01 1.23990488e+00 1.56676903e-01
6.73439682e-01 -9.09289658e-01 -3.95260572e-01 1.04568094e-01
1.67029634e-01 -6.14328347e-02 9.02328864e-02 5.99428594e-01
-1.07285774e+00 1.89631164e-01 -3.22692096e-01 -2.12551326e-01
-1.21318769e+00 4.70818803e-02 4.10944760e-01 2.61700422e-01
-3.06556731e-01 8.98612499e-01 -2.29150757e-01 -1.01415849e+00
1.59751624e-01 -4.98812228e-01 -2.04939291e-01 -3.93206358e-01
1.74956232e-01 6.48837507e-01 -3.78602535e-01 -9.42174196e-01
-3.84946555e-01 2.75112003e-01 6.86297715e-01 -9.43545401e-01
1.00150764e+00 -5.93823254e-01 2.16378614e-01 2.73935109e-01
9.55907553e-02 3.25329155e-01 -1.47491241e+00 5.64372540e-01
-1.38136402e-01 -4.77281779e-01 7.77786598e-03 -7.71308959e-01
-8.31965506e-02 1.73822999e-01 4.96088445e-01 2.82550991e-01
4.65152711e-01 -4.78371680e-01 7.36857057e-01 1.05943727e+00
1.41909993e+00 -1.16108131e+00 -7.89658949e-02 7.53824174e-01
8.09659541e-01 -1.25438702e+00 9.05655771e-02 -1.50332928e-01
-1.30765033e+00 6.70451701e-01 6.67280257e-01 1.90704491e-03
4.34116215e-01 -2.60030963e-02 1.82924837e-01 -2.94158041e-01
-7.01472938e-01 -4.13973421e-01 -3.84749681e-01 1.18575478e+00
-4.37628627e-01 1.44637674e-01 3.36354613e-01 6.98498428e-01
-8.98544014e-01 -1.22327395e-02 9.92299020e-01 1.54465258e+00
-8.49805176e-01 -7.86294401e-01 -3.51755977e-01 8.71089399e-02
1.60324782e-01 1.53388128e-01 -7.28269458e-01 9.22456682e-01
9.97607708e-02 1.66535783e+00 -3.97769213e-01 -9.28977370e-01
4.80308175e-01 -3.58944803e-01 1.39322311e-01 -5.77999353e-01
-7.14399874e-01 -7.67631054e-01 7.17356443e-01 -8.94212425e-01
-4.38396543e-01 -9.61944878e-01 -1.36622787e+00 -4.94459510e-01
-2.38532200e-01 3.82568210e-01 6.40747547e-01 6.92763686e-01
4.68248010e-01 2.09487990e-01 6.21675968e-01 -1.11753106e+00
-5.73207319e-01 -4.38127160e-01 -6.35000944e-01 -3.90138119e-01
1.58475637e-01 -1.08550906e+00 -6.94694698e-01 -5.07402003e-01] | [4.74060583114624, 1.0471702814102173] |
5ded187b-0f47-4e0a-b2da-a31b51d5fed5 | 6d-camera-relocalization-in-ambiguous-scenes | 2004.04807 | null | https://arxiv.org/abs/2004.04807v2 | https://arxiv.org/pdf/2004.04807v2.pdf | 6D Camera Relocalization in Ambiguous Scenes via Continuous Multimodal Inference | We present a multimodal camera relocalization framework that captures ambiguities and uncertainties with continuous mixture models defined on the manifold of camera poses. In highly ambiguous environments, which can easily arise due to symmetries and repetitive structures in the scene, computing one plausible solution (what most state-of-the-art methods currently regress) may not be sufficient. Instead we predict multiple camera pose hypotheses as well as the respective uncertainty for each prediction. Towards this aim, we use Bingham distributions, to model the orientation of the camera pose, and a multivariate Gaussian to model the position, with an end-to-end deep neural network. By incorporating a Winner-Takes-All training scheme, we finally obtain a mixture model that is well suited for explaining ambiguities in the scene, yet does not suffer from mode collapse, a common problem with mixture density networks. We introduce a new dataset specifically designed to foster camera localization research in ambiguous environments and exhaustively evaluate our method on synthetic as well as real data on both ambiguous scenes and on non-ambiguous benchmark datasets. We plan to release our code and dataset under $\href{https://multimodal3dvision.github.io}{multimodal3dvision.github.io}$. | ['Tolga Birdal', 'Mai Bui', 'Haowen Deng', 'Slobodan Ilic', 'Nassir Navab', 'Leonidas Guibas', 'Shadi Albarqouni'] | 2020-04-09 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/2942_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123630137.pdf | eccv-2020-8 | ['camera-localization', 'camera-relocalization'] | ['computer-vision', 'computer-vision'] | [ 4.26043123e-02 -7.27500692e-02 1.81007579e-01 -4.22217309e-01
-9.71878469e-01 -9.12530899e-01 7.24849761e-01 -1.57004625e-01
-3.84673834e-01 5.68910301e-01 3.21980193e-02 -1.71030283e-01
-3.15970600e-01 -2.79420912e-01 -9.61184800e-01 -7.25306988e-01
1.94698691e-01 8.51589799e-01 9.77548771e-03 3.55630443e-02
1.92207217e-01 4.44200128e-01 -1.49641418e+00 1.22186698e-01
7.97617435e-01 9.07522082e-01 5.11339128e-01 7.33480036e-01
2.25230366e-01 5.26900828e-01 -2.43616343e-01 -5.93140006e-01
3.04788172e-01 7.52042234e-02 -5.19844294e-01 4.16230440e-01
7.20955491e-01 -4.69545275e-02 -3.72786760e-01 1.21457613e+00
3.25236857e-01 1.55760124e-01 8.31443071e-01 -1.18112886e+00
-2.93766677e-01 2.48598188e-01 -5.18710434e-01 -5.91251254e-02
3.87295455e-01 6.53824583e-02 8.74958158e-01 -9.36653376e-01
4.87796813e-01 1.09571493e+00 7.06316471e-01 2.69014835e-01
-1.46982789e+00 -2.54480779e-01 2.52653450e-01 7.79964104e-02
-1.65598190e+00 -5.01236320e-01 7.31130540e-01 -4.54159737e-01
4.88544732e-01 2.32130527e-01 2.88812906e-01 1.53661847e+00
6.43868372e-02 7.70190001e-01 9.38591003e-01 -1.74512118e-01
1.70679227e-01 4.58131544e-02 -2.42238238e-01 4.26836580e-01
5.63169755e-02 -5.60430512e-02 -3.81113887e-01 -7.90731758e-02
7.81289220e-01 -1.19628571e-03 -4.31607276e-01 -9.73050535e-01
-1.38519144e+00 6.06412709e-01 3.48533630e-01 -4.28568982e-02
-2.05309972e-01 8.91323164e-02 -1.35493129e-01 -1.51792094e-01
1.59658194e-01 2.59250313e-01 -3.41951370e-01 -1.83956474e-01
-8.60905290e-01 3.50014508e-01 7.25492954e-01 1.03423190e+00
8.03377807e-01 -2.54622459e-01 2.64443994e-01 9.16326761e-01
5.59591532e-01 5.55686116e-01 1.83936834e-01 -1.25086260e+00
4.93944824e-01 2.29971886e-01 2.92411566e-01 -9.50502813e-01
-5.40072858e-01 -6.16801918e-01 -9.02135611e-01 1.43162578e-01
6.56503677e-01 -1.41260922e-01 -9.22773600e-01 1.69855404e+00
1.43119395e-01 1.52257323e-01 -2.09675021e-02 1.17717838e+00
5.30752420e-01 4.81755406e-01 -3.68170857e-01 1.58061862e-01
1.20287263e+00 -7.24788547e-01 -3.10410708e-01 -7.29100108e-01
2.85798281e-01 -1.06435943e+00 7.33946919e-01 7.01187015e-01
-9.03012872e-01 -5.12867928e-01 -9.54188049e-01 7.32498476e-03
-2.16594636e-01 4.91991192e-01 4.21271980e-01 4.53340143e-01
-1.17701173e+00 3.54130268e-01 -1.07577336e+00 -3.84890199e-01
4.44805324e-02 3.85024697e-01 -5.68093538e-01 -3.06660354e-01
-7.02089250e-01 9.41436589e-01 4.12444234e-01 3.66197944e-01
-8.23723972e-01 -2.37400457e-01 -1.06847906e+00 -1.27172947e-01
6.40170097e-01 -7.00026333e-01 1.15979660e+00 -7.93230295e-01
-1.18096328e+00 6.64005876e-01 -1.97458148e-01 -2.23176122e-01
6.55969501e-01 -2.20968708e-01 -2.08494499e-01 -8.48755762e-02
6.20047562e-02 1.07299256e+00 9.38935220e-01 -1.68933129e+00
-4.91440326e-01 -3.65188628e-01 1.21806264e-01 3.14911067e-01
2.54248321e-01 -4.06306803e-01 -8.57205927e-01 -4.27483618e-01
5.61029494e-01 -1.34157562e+00 -3.84373307e-01 -2.36740321e-01
-7.04014301e-01 2.25770563e-01 4.86999512e-01 -6.09551311e-01
7.88038015e-01 -2.12140465e+00 5.03813326e-01 2.40607023e-01
8.67105946e-02 -7.18111992e-02 -1.00434422e-01 2.52622187e-01
-1.06148779e-01 -1.38055682e-01 -3.79231155e-01 -1.03393269e+00
2.65106618e-01 2.48581916e-01 -1.84095040e-01 7.47972488e-01
8.83950591e-02 5.84144115e-01 -5.65978289e-01 -1.80315077e-01
5.36377788e-01 6.95494056e-01 -4.91510093e-01 7.71745220e-02
-3.44535708e-01 5.30528545e-01 -9.33048651e-02 7.42727280e-01
1.00585389e+00 -1.45564437e-01 1.07518777e-01 -3.02879691e-01
1.15036964e-05 -7.55951852e-02 -1.67450535e+00 2.06737947e+00
-3.23903322e-01 6.23207331e-01 2.24187002e-01 -7.18960047e-01
8.67636085e-01 7.31900334e-03 2.49793932e-01 -2.17216954e-01
1.34221271e-01 2.65893728e-01 -2.49586552e-01 -2.67090976e-01
7.48424411e-01 4.83635515e-02 -2.43996501e-01 -4.84184511e-02
3.63610148e-01 -1.71710566e-01 1.11987129e-01 1.25137806e-01
6.84066534e-01 4.27729905e-01 -3.35067837e-03 -2.05622584e-01
4.00568366e-01 -2.33787030e-01 4.03681755e-01 7.77593017e-01
2.76937820e-02 1.26964092e+00 6.00768566e-01 -3.46089482e-01
-9.90113378e-01 -1.27239132e+00 -1.33910537e-01 5.94405413e-01
3.63856286e-01 -3.25054824e-01 -7.68068671e-01 -5.09352982e-01
-2.75628984e-01 7.52157927e-01 -4.24242973e-01 1.34947166e-01
-3.77994478e-02 -8.60789478e-01 3.12979162e-01 2.63792992e-01
2.60076553e-01 -5.41413188e-01 -3.14868927e-01 -3.05462331e-02
-5.60201228e-01 -1.39532089e+00 -4.08570021e-01 2.62877226e-01
-4.09767687e-01 -1.09784746e+00 -6.50082588e-01 -4.47630942e-01
6.12571299e-01 3.57384861e-01 9.29841280e-01 -4.27486271e-01
-1.84721231e-01 7.17274904e-01 -3.26167047e-02 -8.74985307e-02
-2.12039500e-01 -5.20637333e-02 2.27509513e-01 2.90283352e-01
2.29344666e-01 -5.90658903e-01 -5.93934953e-01 4.40860569e-01
-1.08110356e+00 5.29613812e-03 6.19093537e-01 5.91435313e-01
6.23530746e-01 -5.84204942e-02 -1.24669634e-01 -2.79155314e-01
3.28024536e-01 -5.48063457e-01 -9.15440500e-01 1.54653758e-01
-1.84374198e-01 7.72171617e-02 3.30091774e-01 -3.81194651e-01
-9.73674536e-01 4.77555037e-01 -1.07996456e-01 -7.46800423e-01
-6.95024729e-01 4.28931117e-01 -4.56586778e-01 7.47797266e-02
5.48166454e-01 2.35389426e-01 3.80034558e-03 -5.78448892e-01
4.58817542e-01 5.05176544e-01 8.44807386e-01 -6.45191133e-01
6.57026112e-01 6.16335154e-01 6.37229681e-02 -8.86983693e-01
-6.62043214e-01 -5.18044174e-01 -8.78023863e-01 -2.36925587e-01
8.94143939e-01 -1.20309162e+00 -6.06497884e-01 4.47624385e-01
-1.37898004e+00 -7.41557032e-02 1.30042017e-01 7.27897406e-01
-6.65803075e-01 5.31022847e-01 -1.43501177e-01 -8.05914998e-01
3.97877693e-01 -1.56308031e+00 1.37706769e+00 2.16020182e-01
-1.21994384e-01 -9.73056436e-01 1.32783040e-01 5.29547334e-01
1.77566990e-01 1.92215577e-01 4.45633948e-01 -4.58379775e-01
-9.22578514e-01 -2.87336469e-01 -9.23032314e-02 3.56371254e-01
-2.70196855e-01 7.02154711e-02 -1.19902015e+00 -1.95515841e-01
-1.89770907e-01 -1.31220192e-01 9.62724328e-01 5.21422029e-01
9.50746000e-01 9.22391787e-02 -2.85424411e-01 7.48565853e-01
1.42513227e+00 -1.28866687e-01 5.76412678e-01 3.37522984e-01
8.03022265e-01 6.58649087e-01 3.70662957e-01 4.75173593e-01
7.70973206e-01 8.70463133e-01 9.81100023e-01 1.34403750e-01
3.20172191e-01 -3.02062392e-01 3.05414319e-01 1.72427207e-01
3.81329395e-02 -6.43160224e-01 -1.09453535e+00 4.57830012e-01
-1.94917631e+00 -7.26427078e-01 -2.37025678e-01 2.36595631e+00
1.45848036e-01 1.42179847e-01 -1.58004463e-02 -2.01883703e-01
7.09475994e-01 2.06206113e-01 -2.86549181e-01 7.95692801e-02
-4.07605588e-01 -4.33696151e-01 5.26773572e-01 7.69230723e-01
-1.34053016e+00 7.86416173e-01 4.80032158e+00 8.19157422e-01
-1.00694227e+00 -1.20425925e-01 5.82183778e-01 -7.54059255e-02
-3.30634058e-01 2.68156677e-01 -8.44147205e-01 3.83747131e-01
6.23648584e-01 4.47192192e-01 5.94788194e-01 7.58752584e-01
6.31185994e-02 -2.92323679e-01 -1.13805389e+00 1.31114745e+00
3.10308635e-01 -1.08883214e+00 -1.73185408e-01 3.23479891e-01
5.12985408e-01 3.52150798e-01 2.67079651e-01 7.99670368e-02
1.11161664e-01 -9.92168605e-01 1.06659937e+00 7.30639279e-01
4.57338274e-01 -6.38131559e-01 7.79351592e-01 6.87423825e-01
-8.95068169e-01 -1.08051926e-01 -4.18240428e-01 1.87896281e-01
3.87181520e-01 5.51957965e-01 -6.82600498e-01 6.70915663e-01
7.20286965e-01 6.48070455e-01 -8.52500558e-01 1.19265115e+00
-2.05840886e-01 2.47554872e-02 -6.50143981e-01 3.13588828e-01
2.31618911e-01 -4.33995306e-01 8.63372445e-01 1.09140086e+00
6.99989080e-01 -3.27923119e-01 1.58880442e-01 9.52091694e-01
1.42604157e-01 -2.92869985e-01 -7.23151922e-01 2.98401952e-01
3.30190659e-01 1.33753324e+00 -7.87890851e-01 1.52584404e-01
-4.00245905e-01 1.05006206e+00 2.87596792e-01 4.30057138e-01
-8.99810672e-01 1.30195603e-01 6.10551238e-01 4.55706380e-02
4.37773108e-01 -5.52328348e-01 -3.96731272e-02 -1.40295219e+00
2.97737986e-01 -8.88615131e-01 2.16109455e-01 -1.06139779e+00
-1.21194911e+00 7.02351451e-01 2.25384817e-01 -1.51876450e+00
-4.55134779e-01 -7.67761052e-01 -2.87958384e-01 7.52617538e-01
-1.11420763e+00 -1.29627681e+00 -2.60671377e-01 5.10369956e-01
3.78815413e-01 -8.15076195e-03 6.22564495e-01 2.46773243e-01
-4.72711653e-01 2.42857426e-01 1.83600530e-01 -1.56595379e-01
8.70659351e-01 -1.39905739e+00 1.13027111e-01 8.97992611e-01
4.21519727e-01 6.19630337e-01 1.03353310e+00 -2.35283077e-01
-1.50271726e+00 -9.07606065e-01 5.89923978e-01 -8.16453338e-01
6.69977129e-01 -6.77126884e-01 -7.15819836e-01 7.70503283e-01
1.81321561e-01 3.30369473e-02 4.44821626e-01 7.77282417e-02
-2.25793451e-01 5.86179039e-03 -8.31452787e-01 5.65658987e-01
8.32136512e-01 -4.94917750e-01 -3.01758111e-01 1.76328555e-01
4.26896065e-01 -7.24715114e-01 -5.84851503e-01 4.48038310e-01
4.50515002e-01 -1.40147817e+00 9.79706168e-01 -3.30675393e-05
2.78374553e-01 -5.36711991e-01 -6.56970859e-01 -1.35191214e+00
-1.73315927e-01 -5.30649841e-01 1.81602478e-01 1.29514301e+00
5.87813437e-01 -4.67582524e-01 6.80112600e-01 6.54940605e-01
-2.28473991e-01 -4.73560482e-01 -1.15732777e+00 -5.57608783e-01
-1.44016862e-01 -8.45785201e-01 4.08153355e-01 6.30683422e-01
-4.17284727e-01 2.70512521e-01 -5.39311886e-01 7.86343873e-01
7.94787049e-01 -9.69371293e-03 8.89826238e-01 -1.02195883e+00
-4.89154607e-01 -4.88300383e-01 -5.15278339e-01 -1.29557991e+00
3.36319804e-01 -5.21559715e-01 1.42823920e-01 -1.38549030e+00
7.08954781e-02 -2.17533574e-01 1.76128939e-01 7.89737925e-02
8.24206471e-02 2.12641686e-01 2.64171451e-01 9.12786797e-02
-8.37358713e-01 5.98937929e-01 7.63235986e-01 -1.02612272e-01
-7.71362334e-02 2.38738850e-01 -6.01150751e-01 9.61241841e-01
7.78419018e-01 -2.06253693e-01 -3.96422356e-01 -6.66318238e-01
4.10536289e-01 2.55951941e-01 8.72012973e-01 -1.04129434e+00
3.36429715e-01 -9.68941767e-03 4.84672129e-01 -7.33094215e-01
8.72367561e-01 -1.07320619e+00 4.40397024e-01 -1.56247556e-01
-6.99131340e-02 1.10368133e-01 1.95208669e-01 7.12472439e-01
-3.05644095e-01 -2.66949266e-01 6.55097902e-01 -1.85573012e-01
-6.97914124e-01 2.55135596e-01 -3.79802912e-01 -2.02944130e-01
8.25927913e-01 -1.08035281e-01 -2.39704698e-01 -6.44325614e-01
-1.01273584e+00 2.71860421e-01 8.11732233e-01 6.66803181e-01
4.71623629e-01 -1.16361403e+00 -4.84913081e-01 2.05100358e-01
3.23318601e-01 3.93555135e-01 5.09886444e-01 9.27043378e-01
-5.10042667e-01 4.26142991e-01 1.16583198e-01 -1.10479283e+00
-9.59252894e-01 4.98357296e-01 6.03809059e-01 3.86144742e-02
-9.71691087e-02 7.90264964e-01 2.74188399e-01 -8.82713795e-01
3.13069403e-01 -2.65029132e-01 -1.16810296e-02 -5.13870083e-03
1.66163504e-01 1.48111999e-01 -2.05724686e-02 -9.94203687e-01
-4.10355747e-01 5.84032655e-01 2.62593329e-01 -3.31314981e-01
1.02351141e+00 -6.23513162e-01 -7.52093941e-02 3.88053447e-01
1.04034042e+00 1.50046140e-01 -1.61781120e+00 -1.46897286e-01
-1.27550721e-01 -3.84051442e-01 -1.89123638e-02 -7.31529236e-01
-7.37399340e-01 9.64868784e-01 6.79852128e-01 8.82523805e-02
9.16079342e-01 1.65756240e-01 1.05192453e-01 3.17815930e-01
3.61591160e-01 -8.42918515e-01 -1.67679965e-01 5.89498401e-01
9.24730539e-01 -1.50854516e+00 -2.08784863e-01 -3.14451635e-01
-8.26553345e-01 1.12596810e+00 5.10513902e-01 5.97072244e-02
5.91166079e-01 2.60550261e-01 1.48000389e-01 7.04612881e-02
-5.37318945e-01 -1.82294309e-01 3.83616805e-01 5.64942837e-01
1.21088497e-01 1.27841726e-01 3.50638926e-01 5.94510436e-01
-1.77479833e-01 -4.30121601e-01 6.72530591e-01 6.55822039e-01
-2.20498458e-01 -1.07817864e+00 -7.63441563e-01 -7.71420151e-02
-1.67685464e-01 5.36212884e-03 -4.23812807e-01 7.60371566e-01
4.19606686e-01 1.04571545e+00 9.58757922e-02 -3.45671445e-01
2.46741176e-01 -7.36214668e-02 5.00709593e-01 -2.42680967e-01
1.51309147e-01 5.18909574e-01 3.86938006e-02 -6.07479095e-01
-2.58642972e-01 -8.26050282e-01 -8.17989886e-01 -1.34583831e-01
-1.18427165e-01 -3.44654545e-02 9.47407603e-01 8.84540617e-01
3.50127518e-01 2.43492275e-01 3.54376674e-01 -1.46767747e+00
-4.25265312e-01 -9.73355711e-01 -6.46587014e-01 1.62118047e-01
5.46713889e-01 -7.90259004e-01 -5.19273877e-01 -1.83318853e-02] | [7.746932506561279, -2.1897871494293213] |
041e5e15-42e7-4d3d-a01e-aa55e88e6725 | preserving-fine-grain-feature-information-in | 2208.03684 | null | https://arxiv.org/abs/2208.03684v1 | https://arxiv.org/pdf/2208.03684v1.pdf | Preserving Fine-Grain Feature Information in Classification via Entropic Regularization | Labeling a classification dataset implies to define classes and associated coarse labels, that may approximate a smoother and more complicated ground truth. For example, natural images may contain multiple objects, only one of which is labeled in many vision datasets, or classes may result from the discretization of a regression problem. Using cross-entropy to train classification models on such coarse labels is likely to roughly cut through the feature space, potentially disregarding the most meaningful such features, in particular losing information on the underlying fine-grain task. In this paper we are interested in the problem of solving fine-grain classification or regression, using a model trained on coarse-grain labels only. We show that standard cross-entropy can lead to overfitting to coarse-related features. We introduce an entropy-based regularization to promote more diversity in the feature space of trained models, and empirically demonstrate the efficacy of this methodology to reach better performance on the fine-grain problems. Our results are supported through theoretical developments and empirical validation. | ['Vincent Gripon', 'Lucas Drumetz', 'Raphael Baena'] | 2022-08-07 | null | null | null | null | ['fine-grained-image-classification'] | ['computer-vision'] | [ 4.50011760e-01 3.59214813e-01 -1.95379332e-01 -6.20848715e-01
-8.42788339e-01 -4.29048330e-01 6.06044233e-01 2.70980150e-01
-1.59724340e-01 1.03511441e+00 -3.52313928e-02 2.65975714e-01
-3.93269747e-01 -9.21237290e-01 -5.77512622e-01 -8.74665737e-01
-2.09170617e-02 6.27199471e-01 5.96935600e-02 3.42717201e-01
2.15670958e-01 4.03600574e-01 -1.79837394e+00 3.87782872e-01
9.51788664e-01 1.39950395e+00 -2.82639265e-02 3.81375402e-01
3.18765454e-02 7.50402570e-01 -4.57495660e-01 -3.76975507e-01
3.71926546e-01 -1.77962378e-01 -1.08693397e+00 3.44422609e-01
7.12020338e-01 2.28753164e-01 2.74747729e-01 1.30913651e+00
-2.63143152e-01 2.07477823e-01 1.23359668e+00 -1.10355520e+00
-5.64842045e-01 4.63563472e-01 -5.52412271e-01 -3.26209515e-01
-3.16708609e-02 -1.71376362e-01 1.15153515e+00 -5.52685320e-01
4.03655171e-01 1.40619552e+00 9.24937367e-01 3.44009697e-01
-1.65125120e+00 -4.10724312e-01 -3.75264995e-02 -1.81344151e-01
-1.60370362e+00 -1.08561598e-01 6.78582728e-01 -8.52335274e-01
4.40557957e-01 3.44140023e-01 4.28656071e-01 8.10225904e-01
3.42953771e-01 2.25552663e-01 1.60494518e+00 -4.39547449e-01
4.15249884e-01 4.09832895e-01 6.55300438e-01 8.08872700e-01
5.21472394e-01 2.91173041e-01 -6.31272271e-02 -3.03348780e-01
5.76559246e-01 1.32925333e-02 -2.05846831e-01 -5.88080227e-01
-1.16243076e+00 1.21892035e+00 5.91539919e-01 2.27980718e-01
-2.77123302e-01 -4.35453355e-02 1.96610987e-01 2.57748932e-01
8.96709740e-01 6.53415561e-01 -5.76990962e-01 2.87248880e-01
-1.14600563e+00 1.72062635e-01 9.09805834e-01 6.46228433e-01
1.23449671e+00 -3.51813644e-01 -2.14116156e-01 7.86336005e-01
1.87391028e-01 2.30080649e-01 3.66154164e-01 -1.05891788e+00
9.43087488e-02 7.08204269e-01 6.87276348e-02 -7.16099203e-01
-6.45385265e-01 -7.27886260e-01 -1.07157731e+00 5.03364325e-01
8.13820004e-01 1.18143082e-01 -9.88900125e-01 1.61110234e+00
1.61784261e-01 -1.39789045e-01 -2.61743754e-01 7.34822631e-01
3.32953513e-01 5.57591200e-01 3.90653521e-01 -2.32884467e-01
1.24736989e+00 -7.11328089e-01 -3.46368760e-01 1.72294956e-02
7.64316559e-01 -4.90946084e-01 1.29038906e+00 4.62705463e-01
-7.03725219e-01 -7.71423161e-01 -1.00918090e+00 -7.14140432e-03
-5.26135981e-01 2.44319245e-01 5.61964869e-01 5.42903125e-01
-6.50435090e-01 1.09848547e+00 -6.40302539e-01 -2.22381547e-01
5.81375360e-01 3.47325385e-01 -2.79345512e-01 8.64119232e-02
-1.09898841e+00 1.00454652e+00 5.97640872e-01 -1.30926281e-01
-2.89090484e-01 -7.46398747e-01 -8.34223926e-01 2.06136361e-01
2.00189486e-01 -5.75807869e-01 8.18879247e-01 -1.05238426e+00
-1.05129743e+00 1.10686517e+00 1.24129474e-01 -4.84574139e-01
4.53983009e-01 1.57160059e-01 5.08408919e-02 -1.68405905e-01
2.01209888e-01 7.80387700e-01 1.07330775e+00 -1.23883963e+00
-5.21975100e-01 -3.86074662e-01 3.73288020e-02 -2.74732322e-01
-9.05634686e-02 -4.62237328e-01 3.62629265e-01 -7.00924993e-01
1.32839188e-01 -1.00072289e+00 -3.40258658e-01 -1.17658891e-01
-5.48862338e-01 -2.87177503e-01 3.04375052e-01 -2.06571996e-01
1.07391572e+00 -2.11317825e+00 3.17740999e-02 3.47044975e-01
3.30448776e-01 -1.52846992e-01 4.05799672e-02 3.01754288e-02
-2.89539360e-02 3.91684383e-01 -5.29997945e-01 -7.24797621e-02
1.87988341e-01 2.19447225e-01 -2.42967784e-01 6.68639600e-01
5.10246217e-01 7.97584593e-01 -5.64394057e-01 -7.33412087e-01
2.07490504e-01 4.03236121e-01 -4.75025415e-01 -4.67206575e-02
-2.80459046e-01 3.48890424e-01 -6.83388352e-01 3.27170938e-01
6.40696645e-01 -6.43030047e-01 -3.49737972e-01 -4.64554936e-01
7.32720271e-02 8.91995728e-02 -1.05195975e+00 1.32050204e+00
-5.43443441e-01 4.82459188e-01 -2.71168858e-01 -1.21489811e+00
8.66511643e-01 -1.01537578e-01 3.96496415e-01 -3.14789116e-01
2.80812889e-01 9.80714262e-02 -5.29774278e-02 -1.46604717e-01
2.67165780e-01 -5.89148045e-01 -3.94774884e-01 3.69703501e-01
-6.51079463e-03 -3.38817179e-01 1.33502379e-01 -3.14286113e-01
6.63473010e-01 -2.41499133e-02 6.41394854e-01 -6.40344322e-01
4.62151378e-01 1.17837228e-01 5.03222227e-01 8.57648969e-01
-1.63077861e-01 6.26187921e-01 5.38971603e-01 -4.34961677e-01
-1.08818722e+00 -8.96207333e-01 -9.28106427e-01 8.84657383e-01
4.11556102e-02 -9.53435823e-02 -1.00008678e+00 -9.10879016e-01
8.68470073e-02 6.30215704e-01 -9.43088830e-01 -3.49418759e-01
-6.70925677e-02 -9.17091250e-01 1.97486892e-01 1.53334588e-01
3.97318155e-01 -1.04161370e+00 -6.86567008e-01 8.76528844e-02
-1.65924802e-01 -8.02379608e-01 -1.88747257e-01 6.55736387e-01
-9.64889884e-01 -1.03676009e+00 -5.61001003e-01 -5.68240583e-01
7.64014184e-01 -2.89123833e-01 1.35484588e+00 -3.68158408e-02
-3.36841732e-01 8.79142247e-03 -3.92711997e-01 -1.18860282e-01
-4.86953616e-01 3.35106730e-01 -6.32737204e-02 1.66117877e-01
5.55877149e-01 -3.67981642e-01 -1.28570423e-01 2.22043008e-01
-7.28878677e-01 -2.92286463e-02 6.77789569e-01 1.15315223e+00
6.75463974e-01 3.16228658e-01 7.02014565e-01 -1.27169943e+00
3.38978559e-01 -3.42432886e-01 -6.65044129e-01 3.69216353e-01
-7.94800520e-01 6.07432008e-01 7.83273995e-01 -5.19311011e-01
-8.55744183e-01 2.28646487e-01 1.05399206e-01 -2.84338564e-01
-5.61246812e-01 3.12480479e-01 -8.02305713e-02 -3.88136744e-01
1.01120210e+00 -1.02464952e-01 -3.27670813e-01 -5.53353608e-01
2.85143942e-01 5.40197313e-01 3.63694094e-02 -8.22378814e-01
7.12579906e-01 2.98615754e-01 4.00742918e-01 -7.14238286e-01
-1.55893433e+00 -2.82901317e-01 -8.80242765e-01 1.23102278e-01
9.52682734e-01 -5.93952060e-01 -4.52712178e-01 6.45486712e-02
-7.83974171e-01 -3.27565104e-01 -7.86192179e-01 4.68912333e-01
-8.22054625e-01 1.31032318e-01 -4.79726911e-01 -6.14503264e-01
1.79212987e-01 -1.08405817e+00 1.19794929e+00 3.10266912e-01
-3.81142378e-01 -1.13029146e+00 9.45895240e-02 1.63266554e-01
1.75689742e-01 2.28381395e-01 1.07234657e+00 -6.66710496e-01
-4.71143603e-01 -2.79970586e-01 -4.58184093e-01 4.93189216e-01
2.17374101e-01 -8.03966168e-03 -1.26734006e+00 -1.99195519e-01
2.07343400e-01 -7.20109642e-01 1.26040912e+00 6.38875961e-01
1.45207357e+00 -2.84976840e-01 -2.32312649e-01 7.01870739e-01
1.57632220e+00 -2.23454997e-01 3.15623015e-01 1.59790248e-01
6.02285206e-01 1.00131214e+00 8.12524080e-01 4.53631610e-01
-4.69185710e-02 5.58582127e-01 2.19766200e-01 -1.34286597e-01
-3.65992337e-02 -6.57073334e-02 -1.13990478e-01 3.61235112e-01
-8.60991552e-02 1.61033943e-02 -5.61189711e-01 2.33250529e-01
-1.47403526e+00 -8.00178707e-01 -1.02996133e-01 2.21197367e+00
1.02988875e+00 1.69270113e-01 5.17190620e-02 2.76571065e-01
6.77250683e-01 3.29991058e-02 -4.94797945e-01 -2.83493042e-01
2.94581410e-02 3.33907455e-01 6.03013575e-01 5.71488440e-01
-1.48730004e+00 7.52678037e-01 6.06174374e+00 1.02643049e+00
-1.02481139e+00 -3.51395048e-02 1.04501724e+00 1.19312800e-01
-9.71604735e-02 3.67033482e-02 -8.45403790e-01 4.66658235e-01
7.77355611e-01 3.10974568e-02 3.04916948e-01 8.93909693e-01
-1.68648705e-01 -1.42648816e-01 -1.41480684e+00 8.09520960e-01
-2.97248989e-01 -1.18388033e+00 -4.95423144e-03 2.81579167e-01
9.45161462e-01 -2.37607986e-01 4.96186689e-02 3.03092659e-01
3.73774379e-01 -1.21188235e+00 6.56067669e-01 6.01870358e-01
8.81914914e-01 -6.75748885e-01 6.48312509e-01 5.17385721e-01
-1.01567376e+00 1.74641493e-03 -7.04461575e-01 -8.46273229e-02
-3.68142664e-01 9.95253503e-01 -4.24019963e-01 2.27388665e-01
5.48763275e-01 6.05523109e-01 -6.43758416e-01 9.09074962e-01
2.23843098e-01 5.06260157e-01 -2.44389251e-01 -3.45989683e-04
2.29473710e-01 -3.52589220e-01 2.00266242e-02 1.19981170e+00
1.27807617e-01 1.54318199e-01 4.22857076e-01 9.99062419e-01
-6.92388276e-03 2.40852777e-02 -5.69589496e-01 1.23908997e-01
1.95568338e-01 1.43335700e+00 -9.09288108e-01 -4.19288754e-01
-5.20043075e-01 6.52480543e-01 5.31268775e-01 4.02731389e-01
-7.11394966e-01 -3.44498515e-01 4.93248165e-01 2.44566575e-01
2.21906632e-01 1.61365181e-01 -6.51965261e-01 -1.35386562e+00
-1.76367596e-01 -6.61486626e-01 2.55522639e-01 -4.47455972e-01
-1.81758356e+00 6.96359158e-01 -6.28720177e-03 -1.33566344e+00
-2.69717127e-01 -6.89226627e-01 -1.94949478e-01 9.53855276e-01
-1.44561970e+00 -1.06295037e+00 -2.23182425e-01 4.41141188e-01
3.31430793e-01 3.18095088e-01 9.25163925e-01 1.78593658e-02
-1.98529229e-01 4.65778112e-01 2.13128835e-01 -2.68779788e-02
5.80420911e-01 -1.48579168e+00 -6.59596100e-02 3.13497871e-01
1.67738080e-01 3.14014435e-01 7.57153988e-01 -5.75375497e-01
-5.46425819e-01 -1.31104004e+00 8.81687641e-01 -4.59178895e-01
6.80653036e-01 -2.99113065e-01 -1.08144188e+00 6.11264884e-01
-3.76193136e-01 3.64384472e-01 5.82488239e-01 2.46719405e-01
-4.48525906e-01 -6.09096512e-02 -1.49089122e+00 1.38339341e-01
7.27046251e-01 -6.06209576e-01 -6.66315138e-01 2.72329628e-01
4.05366838e-01 1.41920164e-01 -1.12943065e+00 4.67793107e-01
5.41390419e-01 -9.81710255e-01 9.23442304e-01 -6.39259815e-01
4.02125984e-01 -1.58941224e-01 -3.35963786e-01 -1.44927120e+00
-6.11844182e-01 6.56470703e-03 2.10993826e-01 1.14492202e+00
3.30948502e-01 -5.10822058e-01 9.61665452e-01 6.13855004e-01
1.44499913e-01 -7.48631179e-01 -8.06854427e-01 -9.15664315e-01
4.30551589e-01 -1.90795153e-01 4.16065276e-01 9.99638319e-01
-5.38683087e-02 3.64987016e-01 -1.86649323e-01 2.70400289e-02
1.08569932e+00 4.90346014e-01 2.63714284e-01 -2.03960896e+00
-1.93978161e-01 -5.07463455e-01 -5.00783861e-01 -6.42420411e-01
4.35463607e-01 -9.59986150e-01 3.20399255e-01 -1.11900079e+00
4.22812611e-01 -8.85285020e-01 -2.68554926e-01 4.45561677e-01
2.48630196e-02 5.46549618e-01 -4.23095375e-03 2.59180784e-01
-4.35056627e-01 5.26987433e-01 1.43511963e+00 -3.25126946e-01
9.14443731e-02 1.59902960e-01 -6.08097017e-01 1.03688967e+00
6.89810216e-01 -7.85868764e-01 -3.53227943e-01 2.80338526e-01
1.38285935e-01 -6.74320906e-02 5.33922076e-01 -1.06050241e+00
-2.57948458e-01 -3.81517082e-01 7.13613212e-01 -2.95507550e-01
1.95295155e-01 -1.00758791e+00 6.58615902e-02 3.51971507e-01
-6.67536974e-01 -7.79486656e-01 -1.34659618e-01 3.01778853e-01
-3.29417706e-01 -4.88022387e-01 1.24103415e+00 -1.27303049e-01
-3.60818535e-01 3.55887115e-01 -6.09367900e-02 2.82439232e-01
9.56402361e-01 -1.91140175e-01 -1.49498716e-01 -3.00996751e-03
-1.02479970e+00 -1.47544414e-01 7.29525924e-01 -1.35867193e-01
1.31988078e-01 -1.29345310e+00 -6.12050951e-01 3.59049320e-01
1.86169937e-01 -1.25524312e-01 -1.63926017e-02 5.81512392e-01
-2.30029389e-01 5.65490663e-01 -3.08879137e-01 -7.00038075e-01
-9.60289180e-01 6.65021300e-01 4.27964568e-01 -6.30851984e-01
-3.65756363e-01 7.54243255e-01 7.78020322e-01 -3.71698499e-01
1.35050252e-01 -8.16087723e-01 -2.91668087e-01 1.90002218e-01
2.12385878e-01 2.80169874e-01 -8.41798857e-02 -5.66314936e-01
-2.78664023e-01 1.02706850e+00 -6.26757294e-02 2.79364228e-01
1.04713690e+00 -2.10136876e-01 -5.32707982e-02 6.99238360e-01
1.34585404e+00 -1.71656802e-01 -1.61884367e+00 -3.09704632e-01
2.66757071e-01 -3.67396355e-01 1.24332272e-01 -7.84109592e-01
-8.53349924e-01 1.03309512e+00 6.35336041e-01 7.33570337e-01
1.12507498e+00 3.63359153e-01 2.86434591e-01 4.16480184e-01
4.89422619e-01 -9.62799430e-01 -2.68912107e-01 4.16632146e-01
6.32559597e-01 -1.52550256e+00 6.62324280e-02 -6.04619682e-01
-4.01333481e-01 1.28479421e+00 3.58985484e-01 -5.48755765e-01
8.95795643e-01 1.35696054e-01 -2.14258239e-01 -2.02487215e-01
-6.52684629e-01 -1.93456560e-01 6.35991991e-01 5.33896267e-01
4.91787642e-01 3.64914030e-01 -2.28214130e-01 7.27233708e-01
-3.15440536e-01 -7.51547934e-03 2.85519600e-01 3.74929547e-01
-6.51234210e-01 -1.05262196e+00 -3.95339251e-01 9.80849624e-01
-5.30521154e-01 -1.28440946e-01 -2.93767065e-01 8.38975608e-01
5.70562482e-01 7.33558178e-01 2.73334503e-01 -1.66802064e-01
8.07361826e-02 1.79285258e-01 6.72585666e-01 -7.14314520e-01
-3.50173235e-01 1.30335484e-02 -8.00813362e-02 -4.84300166e-01
-5.00407755e-01 -7.44569719e-01 -1.10769475e+00 -2.93949246e-01
-3.73441368e-01 2.55545884e-01 3.92932832e-01 1.13695168e+00
-1.29140377e-01 4.51439351e-01 6.72578990e-01 -9.36098635e-01
-8.91191363e-01 -1.10606670e+00 -1.02075374e+00 4.63308275e-01
6.92817748e-01 -9.40668643e-01 -7.14483976e-01 2.72806495e-01] | [9.509827613830566, 2.9193203449249268] |
e72b27aa-15e6-4b3f-976e-f03522e1f19e | vrkitchen2-0-indoorkit-a-tutorial-for | 2206.11887 | null | https://arxiv.org/abs/2206.11887v1 | https://arxiv.org/pdf/2206.11887v1.pdf | VRKitchen2.0-IndoorKit: A Tutorial for Augmented Indoor Scene Building in Omniverse | With the recent progress of simulations by 3D modeling software and game engines, many researchers have focused on Embodied AI tasks in the virtual environment. However, the research community lacks a platform that can easily serve both indoor scene synthesis and model benchmarking with various algorithms. Meanwhile, computer graphics-related tasks need a toolkit for implementing advanced synthesizing techniques. To facilitate the study of indoor scene building methods and their potential robotics applications, we introduce INDOORKIT: a built-in toolkit for NVIDIA OMNIVERSE that provides flexible pipelines for indoor scene building, scene randomizing, and animation controls. Besides, combining Python coding in the animation software INDOORKIT assists researchers in creating real-time training and controlling avatars and robotics. The source code for this toolkit is available at https://github.com/realvcla/VRKitchen2.0-Tutorial, and the tutorial along with the toolkit is available at https://vrkitchen20-tutorial.readthedocs.io/en/ | ['Song-Chun Zhu', 'Wensi Ai', 'Xiaofeng Gao', 'Steven Gong', 'Yizhou Zhao'] | 2022-06-23 | null | null | null | null | ['indoor-scene-synthesis'] | ['computer-vision'] | [-5.43386877e-01 -4.20562446e-01 4.17519271e-01 -2.04849571e-01
-3.24064493e-02 -7.20338106e-01 5.00939310e-01 -5.09118557e-01
-7.25684986e-02 3.68862599e-01 7.22867325e-02 -5.92363477e-01
4.59187269e-01 -1.03555727e+00 -5.16205132e-01 -6.09109044e-01
1.07327841e-01 1.83274701e-01 2.68217534e-01 -3.85851413e-01
8.36522803e-02 3.72036427e-01 -2.03941941e+00 -8.10908079e-02
8.71561289e-01 4.69124675e-01 1.04692316e+00 1.11564207e+00
-9.81461033e-02 8.52129638e-01 -3.90195906e-01 1.56059861e-01
2.42376342e-01 -5.12577951e-01 -4.61376399e-01 -3.42355758e-01
-4.97266352e-02 -4.47048366e-01 -5.52051246e-01 8.66895616e-01
9.18769360e-01 4.27950323e-01 1.12787217e-01 -1.47436523e+00
-3.70421290e-01 1.20791808e-01 -6.72019261e-04 -1.28828868e-01
8.89841557e-01 6.70003653e-01 4.17848617e-01 -7.99296558e-01
7.44830191e-01 1.04876971e+00 4.28693444e-01 8.76649499e-01
-6.88927591e-01 -8.70093107e-01 -6.65591359e-02 1.77762046e-01
-1.66076136e+00 -5.50690114e-01 7.44726896e-01 -4.63496894e-01
9.14004147e-01 6.53259695e-01 1.25758529e+00 1.33850241e+00
2.04468239e-03 7.72207081e-01 9.29176271e-01 -2.17422977e-01
4.79809105e-01 1.25978395e-01 -3.96508351e-02 8.95063579e-01
-1.91840172e-01 2.19638333e-01 -1.85925126e-01 1.81581080e-01
1.50331807e+00 -2.16872126e-01 -1.26368448e-01 -5.68966925e-01
-1.29718947e+00 5.00042319e-01 5.43298364e-01 1.40130162e-01
-1.58488154e-01 6.32726550e-01 3.41431379e-01 -1.43220916e-01
5.75711541e-02 4.71384287e-01 -3.66756022e-01 -5.14658272e-01
-3.00395519e-01 8.16361427e-01 6.53690159e-01 1.29613614e+00
4.27352965e-01 3.19763035e-01 3.57851416e-01 7.08889186e-01
3.51527005e-01 2.72036910e-01 3.19969803e-01 -1.58661413e+00
-4.49377894e-02 4.89407152e-01 1.82551265e-01 -7.49637127e-01
-5.58848500e-01 6.63511828e-02 -4.98936594e-01 6.00592554e-01
2.24177420e-01 -4.74703729e-01 -6.78571641e-01 1.28677261e+00
1.01224566e+00 4.39863145e-01 -1.83813035e-01 1.11629677e+00
1.49631441e+00 8.35071325e-01 -1.53896203e-02 3.74560714e-01
1.16851330e+00 -1.32386243e+00 -5.95367968e-01 6.77656680e-02
8.71868789e-01 -6.08837664e-01 1.65384305e+00 1.52903005e-01
-9.29436207e-01 -6.72235906e-01 -9.51728702e-01 -2.79291213e-01
-4.97327238e-01 -1.20899051e-01 1.17379677e+00 5.23130894e-01
-1.26626658e+00 4.14467931e-01 -1.27020454e+00 -6.63348258e-01
2.02900395e-01 2.07522780e-01 -1.36466250e-01 6.19515553e-02
-8.48104954e-01 8.48073781e-01 2.71301448e-01 -1.61257476e-01
-1.06778955e+00 -7.36214936e-01 -1.01227283e+00 -4.29436833e-01
2.86490887e-01 -1.11950088e+00 1.63037777e+00 -5.12109041e-01
-2.08589697e+00 7.64384508e-01 1.25526309e-01 1.70094490e-01
6.43227339e-01 -2.33447269e-01 -1.34807229e-01 -2.67543018e-01
1.30023852e-01 7.13913620e-01 5.38676046e-02 -1.29711199e+00
-4.23492700e-01 -2.15023920e-01 3.83953691e-01 6.98022068e-01
3.99471909e-01 2.27507412e-01 -6.23491943e-01 -2.47151911e-01
-1.97227910e-01 -7.88967013e-01 -6.58833504e-01 2.59326428e-01
-4.18796957e-01 2.33016312e-02 9.20737088e-01 -2.48075306e-01
1.01117206e+00 -2.11789441e+00 -6.09729588e-02 -1.69598654e-01
2.47754417e-02 1.12494484e-01 1.13596804e-01 5.01777291e-01
4.41296622e-02 -4.30557765e-02 2.76430964e-01 -2.58295476e-01
2.48105645e-01 5.48833646e-02 5.96542237e-03 2.15590477e-01
-5.51106751e-01 6.93081677e-01 -1.11151695e+00 -4.50490206e-01
1.15962946e+00 6.58621728e-01 -6.98700368e-01 3.71875226e-01
-2.08642527e-01 1.00763738e+00 -6.75385237e-01 4.93015885e-01
7.37649441e-01 -1.21450782e-01 1.28472790e-01 3.78345221e-01
-6.50567472e-01 3.73020679e-01 -1.52982223e+00 2.27265453e+00
-7.03911901e-01 6.72186434e-01 2.97373384e-01 -3.05702150e-01
6.92689180e-01 9.82484743e-02 2.54794568e-01 -5.10818243e-01
3.94812405e-01 1.37555180e-02 -2.73405522e-01 -6.94797337e-01
7.78660476e-01 6.63509667e-01 -8.37087482e-02 2.02804536e-01
-2.77045280e-01 -8.96175623e-01 6.79523647e-02 1.90315753e-01
1.11678278e+00 1.00805950e+00 1.91331640e-01 -2.15193644e-01
1.52432457e-01 3.72315943e-01 1.88470423e-01 4.71093416e-01
-1.14243016e-01 4.04587448e-01 -1.26514494e-01 -4.79170173e-01
-1.23000050e+00 -1.24029291e+00 -2.05944851e-01 1.26063263e+00
5.55768371e-01 -8.52269053e-01 -9.54265714e-01 2.90089756e-01
-2.90713876e-01 1.04727054e+00 -2.94785619e-01 3.30568016e-01
-3.64980131e-01 -2.99725205e-01 3.38973135e-01 4.65337425e-01
7.69046485e-01 -1.28660119e+00 -1.20777273e+00 -4.13112715e-02
-4.82934453e-02 -9.63121057e-01 9.69412178e-03 -6.18269704e-02
-4.90774274e-01 -8.27077925e-01 -3.33445936e-01 -8.64476204e-01
4.60955232e-01 7.29329646e-01 9.16179001e-01 4.61093366e-01
-4.79842216e-01 5.87104261e-01 -4.07371193e-01 -4.83476847e-01
-1.53356701e-01 -1.27047986e-01 1.12736352e-01 -9.94427562e-01
-1.85564056e-01 -8.14166427e-01 -6.87267542e-01 3.45566243e-01
-4.69353497e-01 1.24931359e+00 -4.07299727e-01 2.64031380e-01
4.63977307e-01 -1.13190740e-01 -1.39186382e-01 -4.04831856e-01
4.24330056e-01 -3.80545735e-01 -9.53190684e-01 -1.83026657e-01
7.75365531e-02 -4.95002598e-01 6.68546438e-01 -2.54450500e-01
-1.00908458e+00 2.34499365e-01 -6.20272398e-01 -1.76065579e-01
-6.53254807e-01 5.82776219e-02 -5.13382077e-01 -5.55452183e-02
6.25515282e-01 2.17516050e-01 -2.82517642e-01 -1.98607102e-01
4.94305700e-01 5.34903228e-01 5.44619083e-01 -6.53174162e-01
6.36418164e-01 3.13468605e-01 -4.02579337e-01 -1.02437389e+00
-2.35377878e-01 -1.59478143e-01 -5.32119811e-01 -6.79809690e-01
9.56723213e-01 -9.43012118e-01 -1.02011156e+00 6.95155799e-01
-1.04468739e+00 -1.25819814e+00 -2.07248673e-01 4.07861292e-01
-8.16997766e-01 -8.91239122e-02 -5.74328601e-01 -8.17686915e-01
-4.95526753e-02 -1.50789809e+00 9.22950923e-01 7.47074485e-01
-3.62271219e-01 -9.10003901e-01 3.43449205e-01 1.78154752e-01
4.05907780e-01 4.33547914e-01 3.23800445e-01 7.76577815e-02
-7.46348977e-01 9.94006321e-02 1.65963337e-01 -2.83635855e-01
-1.03903741e-01 5.67991257e-01 -1.14907134e+00 2.24299461e-01
-4.55531508e-01 -1.93187311e-01 -4.90163565e-02 6.01755977e-01
1.54788542e+00 1.31657764e-01 -4.94766831e-01 1.12570083e+00
1.30659890e+00 5.55500805e-01 8.48570764e-01 7.09425330e-01
1.00192356e+00 3.30385476e-01 6.05954528e-01 6.78053439e-01
8.83689940e-01 8.64666522e-01 5.34086466e-01 -1.24697961e-01
1.32684978e-02 -4.10890073e-01 9.20897052e-02 9.71482575e-01
-4.32855606e-01 -5.58585376e-02 -9.61903632e-01 1.86199903e-01
-1.88046455e+00 -1.07394540e+00 -5.63279510e-01 1.95349038e+00
5.09815514e-01 -3.64773214e-01 1.22617275e-01 -9.38435420e-02
5.30647039e-01 2.00842828e-01 -5.01464248e-01 -5.49430907e-01
3.03097963e-01 1.46907434e-01 1.51474550e-01 5.07658958e-01
-1.11679757e+00 1.50955629e+00 5.64066219e+00 7.22865760e-01
-1.09689486e+00 4.75384891e-02 1.99641615e-01 -2.37447619e-02
-1.55786857e-01 5.89877740e-02 -5.34886837e-01 5.00810146e-01
8.34117115e-01 -4.08477277e-01 8.11714053e-01 1.42576110e+00
7.36235261e-01 -3.55756938e-01 -7.47864842e-01 1.10176158e+00
-5.92981637e-01 -1.59017956e+00 -4.56647813e-01 -6.41380772e-02
6.05509996e-01 1.77010536e-01 -3.89125422e-02 3.88400793e-01
9.05077100e-01 -8.42824399e-01 1.00629902e+00 1.85313836e-01
7.17342854e-01 -6.02608502e-01 2.20043123e-01 4.19055313e-01
-1.40002632e+00 3.19053859e-01 -5.06197214e-01 -5.91077507e-01
1.89369187e-01 -1.63078383e-01 -7.84729898e-01 2.93419540e-01
1.03487337e+00 6.10421598e-01 -4.32932138e-01 1.31991005e+00
-3.91706258e-01 1.44282401e-01 -2.11573422e-01 -3.88581336e-01
-9.08451900e-02 -5.62553585e-01 3.53280097e-01 1.10326159e+00
2.64100373e-01 4.15356785e-01 1.74604893e-01 7.41306126e-01
2.74590254e-01 1.70600504e-01 -9.72090602e-01 2.05186263e-01
7.07135737e-01 1.51332438e+00 -9.81147349e-01 -2.10664734e-01
-8.64308551e-02 1.02977729e+00 5.84272966e-02 1.80971205e-01
-1.42965591e+00 -5.76873124e-01 1.13005555e+00 8.91239345e-02
-1.19986221e-01 -9.65074837e-01 -4.79337424e-01 -1.20098364e+00
-4.49030757e-01 -7.23334908e-01 -3.17095816e-01 -1.56843209e+00
-4.41834182e-01 5.83688855e-01 1.46864027e-01 -1.13232446e+00
-1.65426195e-01 -5.69034457e-01 -1.01154006e+00 3.61521661e-01
-7.32239068e-01 -1.30604708e+00 -1.02796233e+00 7.35641539e-01
8.25250506e-01 2.35064864e-01 1.12249231e+00 1.94603801e-01
-8.30711246e-01 -3.89435468e-03 2.08985582e-01 -5.04871309e-02
2.14451805e-01 -1.16187227e+00 1.00845003e+00 7.36468077e-01
-2.61671156e-01 4.39291239e-01 9.59266484e-01 -4.66040045e-01
-1.77073193e+00 -9.62497115e-01 1.28617063e-01 -6.57431364e-01
7.16320813e-01 -7.60425031e-01 -5.12052417e-01 7.82037199e-01
3.75028253e-01 -4.10645679e-02 5.18496871e-01 -1.60910800e-01
2.62963384e-01 2.86211669e-01 -9.52188611e-01 1.30744457e+00
1.62732661e+00 -2.84479320e-01 5.47420382e-02 4.87198889e-01
9.11888003e-01 -1.23288560e+00 -5.10638952e-01 -1.05496533e-01
8.23757827e-01 -1.15084136e+00 1.04628241e+00 -6.91908225e-03
3.86412323e-01 -6.96712971e-01 -3.02281708e-01 -1.18633664e+00
-3.44657928e-01 -5.80138624e-01 1.77494466e-01 1.03354633e+00
1.62880234e-02 -4.54000473e-01 7.06793308e-01 8.03572714e-01
-5.93676746e-01 -5.97929120e-01 -5.26592553e-01 -4.67538059e-01
-2.41333157e-01 -1.03684700e+00 7.26699054e-01 9.02852595e-01
8.31663162e-02 3.72984521e-02 -1.15398243e-01 3.16333681e-01
3.12965780e-01 -1.07822016e-01 1.52961397e+00 -7.97453403e-01
-2.14933753e-01 -4.69092011e-01 -3.49570543e-01 -1.01895201e+00
-7.54642859e-02 -6.21831715e-01 6.05922006e-02 -2.07163191e+00
-1.27091855e-01 -6.46794796e-01 6.04648471e-01 2.74734885e-01
9.84907597e-02 1.99816860e-02 3.38364959e-01 1.98888150e-03
-7.27245033e-01 6.82869971e-01 1.51321721e+00 4.13282245e-01
-5.30169904e-01 -4.05539908e-02 -4.82739121e-01 1.02215636e+00
1.10657310e+00 -1.18942529e-01 -4.95928019e-01 -6.06990159e-01
-1.63768455e-01 -1.21258438e-01 7.83455729e-01 -1.58413100e+00
3.56979012e-01 -4.58898067e-01 4.57799762e-01 -4.59457040e-01
7.57453024e-01 -6.61740899e-01 7.94275701e-01 5.11980295e-01
8.34596753e-02 3.50202501e-01 4.10092175e-01 -1.85889527e-01
5.59872806e-01 2.22332235e-02 4.87377256e-01 -5.81597269e-01
-1.12598372e+00 2.32873395e-01 -7.78272867e-01 -2.11468041e-01
1.42253220e+00 -5.05321980e-01 -3.79200757e-01 -5.16213357e-01
-6.44966185e-01 4.01079327e-01 1.13584340e+00 4.03645992e-01
6.46903455e-01 -1.30658543e+00 -1.83489680e-01 1.69760168e-01
-1.65011644e-01 3.33138645e-01 5.42880058e-01 3.48407030e-01
-1.53513956e+00 1.33275837e-01 -5.23763895e-01 -5.38182020e-01
-1.34212065e+00 4.16077316e-01 4.10773277e-01 3.41365188e-01
-7.98353195e-01 1.02648389e+00 4.81876582e-01 -1.04052579e+00
1.26603946e-01 -4.52605397e-01 2.88407262e-02 -8.87514353e-01
5.73426723e-01 6.12036288e-01 -3.00488740e-01 -4.82909054e-01
-4.49003458e-01 4.22579378e-01 5.81603944e-01 -1.68462843e-01
1.29790294e+00 -1.84356511e-01 1.79399848e-01 4.91288573e-01
7.54483044e-01 6.52483664e-03 -1.32030141e+00 5.44764459e-01
-4.71793026e-01 -6.30027652e-01 -1.32851735e-01 -4.57195461e-01
-6.49338722e-01 6.95307851e-01 6.70141339e-01 -1.24383353e-01
9.32931721e-01 -1.01317249e-01 4.79726911e-01 2.20432580e-01
1.06501520e+00 -9.06650066e-01 -1.53616413e-01 6.94295883e-01
9.31413770e-01 -8.60984325e-01 -3.30454074e-02 -4.90482509e-01
-8.19249988e-01 8.81857991e-01 1.08787525e+00 -4.31955069e-01
6.61451697e-01 7.21106052e-01 2.92578667e-01 -2.36167908e-01
-5.75133264e-01 -1.22292303e-01 -3.43067795e-01 1.03301632e+00
4.88918662e-01 3.65018189e-01 8.77746493e-02 5.70642591e-01
-1.02947915e+00 1.02058761e-01 7.76405156e-01 1.06547403e+00
-3.19597244e-01 -9.99287307e-01 -3.76832962e-01 -4.96996082e-02
-8.68861377e-03 2.79930513e-02 -1.52508710e-02 9.49574471e-01
2.41413996e-01 8.05295646e-01 1.27727866e-01 -8.41370642e-01
3.23072672e-01 -4.39435691e-01 6.03098154e-01 -6.63923204e-01
-7.18469620e-01 -1.62611708e-01 1.51238531e-01 -8.97170305e-01
-5.60785122e-02 -4.80580777e-01 -1.60476863e+00 -8.96158874e-01
-1.68298274e-01 9.76999477e-02 1.09397840e+00 3.21121812e-01
5.90553641e-01 6.36113226e-01 4.20941174e-01 -1.63954282e+00
4.41407114e-01 -5.85735559e-01 -5.25559545e-01 -2.22968876e-01
-1.68924794e-01 -8.17750335e-01 2.98576280e-02 -5.58801815e-02] | [4.525620460510254, 0.6779904365539551] |
78e7ea69-0fcf-4796-9686-d4cc99f9a04e | but-fit-at-semeval-2019-task-7-determining | 1902.10126 | null | http://arxiv.org/abs/1902.10126v2 | http://arxiv.org/pdf/1902.10126v2.pdf | BUT-FIT at SemEval-2019 Task 7: Determining the Rumour Stance with Pre-Trained Deep Bidirectional Transformers | This paper describes our system submitted to SemEval 2019 Task 7: RumourEval
2019: Determining Rumour Veracity and Support for Rumours, Subtask A (Gorrell
et al., 2019). The challenge focused on classifying whether posts from Twitter
and Reddit support, deny, query, or comment a hidden rumour, truthfulness of
which is the topic of an underlying discussion thread. We formulate the problem
as a stance classification, determining the rumour stance of a post with
respect to the previous thread post and the source thread post. The recent BERT
architecture was employed to build an end-to-end system which has reached the
F1 score of 61.67% on the provided test data. It finished at the 2nd place in
the competition, without any hand-crafted features, only 0.2% behind the
winner. | ['Lukáš Burget', 'Martin Fajcik', 'Pavel Smrz'] | 2019-02-25 | but-fit-at-semeval-2019-task-7-determining-1 | https://aclanthology.org/S19-2192 | https://aclanthology.org/S19-2192.pdf | semeval-2019-6 | ['rumour-detection'] | ['natural-language-processing'] | [-1.04397759e-01 6.17261171e-01 -3.97987038e-01 -2.11032256e-01
-5.86977124e-01 -3.04182708e-01 1.31356299e+00 5.47354400e-01
-2.47003227e-01 9.45101559e-01 6.15438938e-01 -4.90559280e-01
4.53883737e-01 -3.32175404e-01 -5.06226480e-01 -2.78504610e-01
-1.10850379e-01 5.36270618e-01 3.94546926e-01 -6.94502294e-01
1.03818882e+00 -2.13973194e-01 -1.23329115e+00 9.80290890e-01
2.92821854e-01 9.58543122e-01 -5.53487062e-01 7.11974025e-01
1.03956483e-01 1.90379143e+00 -9.21132147e-01 -5.19953728e-01
-4.14322823e-01 -5.37041187e-01 -1.28189993e+00 3.60682383e-02
6.50341868e-01 -1.01914115e-01 -1.21372357e-01 8.69783521e-01
1.86777353e-01 -2.13135123e-01 5.03949642e-01 -1.26485527e+00
-5.30018270e-01 1.09722078e+00 -6.23171031e-01 8.85438025e-01
4.29254144e-01 -1.08521000e-01 1.14502895e+00 -1.30611849e+00
8.67537200e-01 1.01897788e+00 7.94245780e-01 2.32600719e-01
-9.43423092e-01 -5.31044483e-01 -2.41903380e-01 4.11718339e-01
-9.20244098e-01 -7.05427051e-01 5.28933048e-01 -7.99692750e-01
7.39159524e-01 4.25893128e-01 3.62993360e-01 1.66176724e+00
5.07301271e-01 7.61624932e-01 1.66087151e+00 1.36380970e-01
2.38037989e-01 4.46505040e-01 6.87924564e-01 5.03227711e-01
1.69828176e-01 -3.28728259e-01 -1.02727163e+00 -6.88911796e-01
-1.67932317e-01 -4.60266322e-01 -3.64125192e-01 3.14893454e-01
-1.09453261e+00 1.17018843e+00 6.65115118e-01 2.94574648e-01
-6.69298351e-01 -4.32125539e-01 7.95439839e-01 7.32812226e-01
1.04522443e+00 6.39377713e-01 7.37065589e-03 -1.57054201e-01
-1.41051054e+00 6.89169347e-01 1.46895194e+00 3.67968053e-01
3.95669997e-01 -1.81871831e-01 -8.03830177e-02 5.41650712e-01
3.87976095e-02 1.50867954e-01 7.16885924e-01 -6.04420722e-01
3.71522397e-01 4.23602521e-01 5.72060883e-01 -1.34247804e+00
-5.99232435e-01 -7.03881800e-01 -6.15688026e-01 -1.39451884e-02
2.08554596e-01 -3.19187760e-01 -4.64962691e-01 1.00596142e+00
1.12245478e-01 1.83889106e-01 2.90229730e-02 1.10740256e+00
1.11696863e+00 6.60853326e-01 -4.14577156e-01 -6.06978893e-01
1.51370311e+00 -1.24692810e+00 -7.07506061e-01 -3.81951004e-01
4.37821150e-01 -1.29121959e+00 6.51134968e-01 5.89550495e-01
-1.17178941e+00 -2.52289195e-02 -1.39470780e+00 3.35157998e-02
-6.40207976e-02 -3.15890431e-01 -1.99623033e-02 1.20048836e-01
-8.74953985e-01 6.28805280e-01 -3.57492357e-01 -1.86930239e-01
1.38657242e-01 -4.88328874e-01 1.36404783e-01 1.87353298e-01
-1.59383965e+00 1.16167021e+00 6.11257404e-02 -8.68872255e-02
-1.19153881e+00 -6.70983195e-01 -1.71685860e-01 -3.67447585e-01
4.42606390e-01 -3.65454495e-01 1.68503976e+00 -1.08009386e+00
-1.27690172e+00 1.38673997e+00 -3.31672654e-02 -1.03005803e+00
1.17509043e+00 -3.44009399e-01 -6.39714897e-01 -2.14106500e-01
4.89126682e-01 -3.54759097e-01 1.23599243e+00 -8.56701076e-01
-6.47007704e-01 -2.63098329e-01 -1.98955446e-01 -2.14202419e-01
2.87353992e-01 4.69564855e-01 4.13291514e-01 -3.63372296e-01
8.36607590e-02 -9.91992414e-01 3.78194392e-01 -8.67801726e-01
-8.43610704e-01 -4.98080820e-01 8.04839015e-01 -6.43783748e-01
1.15336025e+00 -1.80069458e+00 -2.10079148e-01 -6.37663752e-02
6.02833807e-01 2.19459221e-01 4.02460426e-01 7.66313612e-01
4.85464856e-02 1.86077416e-01 8.14853087e-02 -2.99153358e-01
8.29330608e-02 -3.76930028e-01 -1.01988900e+00 9.58044827e-01
-1.55382782e-01 4.64576900e-01 -9.13815260e-01 -1.61718905e-01
-6.31736636e-01 9.46868807e-02 -1.87691972e-01 2.46147603e-01
-3.00086409e-01 1.38734162e-01 -3.52702677e-01 2.65330672e-01
6.12841189e-01 -6.59016490e-01 -4.92630526e-02 1.17596611e-01
-5.31572819e-01 1.24827099e+00 -4.65783924e-01 8.06365609e-01
-1.73435792e-01 1.07433760e+00 3.11559737e-01 -2.67063171e-01
1.06248224e+00 5.12975812e-01 1.10353656e-01 -4.73223090e-01
3.82425606e-01 6.36717319e-01 -2.13861406e-01 -4.14275199e-01
1.01344323e+00 -2.28283912e-01 -2.93497413e-01 9.54010010e-01
-5.99885464e-01 5.54354079e-02 -4.76239398e-02 6.33383870e-01
1.14962375e+00 -5.35715640e-01 5.28836250e-01 -5.81714749e-01
5.86822987e-01 3.92078727e-01 3.76963802e-02 8.42393935e-01
-2.92254061e-01 5.48364937e-01 9.50918138e-01 -6.54424548e-01
-9.44369316e-01 -4.74765658e-01 -8.17780346e-02 1.30311537e+00
-1.35729641e-01 -4.35975522e-01 -4.42899853e-01 -6.53607905e-01
1.13723874e-01 1.13475883e+00 -1.09444833e+00 2.46871352e-01
-4.20128554e-01 -7.59645820e-01 6.43579662e-01 -3.15565377e-01
3.67979020e-01 -1.06837022e+00 -6.62188292e-01 3.26736271e-01
-6.99500740e-01 -1.11747730e+00 -3.17365289e-01 -7.77386427e-02
-4.71710622e-01 -1.17458797e+00 -5.09535789e-01 -2.81173527e-01
-3.84982713e-02 2.37489894e-01 1.33959687e+00 4.65554178e-01
3.36723864e-01 -5.72818160e-01 -3.69375646e-01 -2.81896740e-01
-7.98380792e-01 1.69262737e-01 2.88054440e-03 9.84310731e-02
1.51983961e-01 -4.77174371e-01 -4.86783862e-01 5.17591596e-01
-4.26606029e-01 3.23753729e-02 1.55230254e-01 6.95598543e-01
-7.50515684e-02 -4.34803694e-01 9.45607305e-01 -1.36031067e+00
1.04846442e+00 -1.11278129e+00 -9.49189812e-02 -5.11445105e-01
-9.55407560e-01 -4.23484504e-01 2.97509015e-01 5.33324443e-02
-5.94679594e-01 -1.06067145e+00 -9.69718620e-02 8.25611502e-02
2.79950768e-01 7.65714288e-01 7.15174496e-01 4.73428041e-01
1.26468635e+00 9.47298631e-02 6.29693642e-02 -5.24907887e-01
1.76933587e-01 1.06738305e+00 5.75400949e-01 -1.43288355e-02
2.65103787e-01 4.62796628e-01 -5.05300462e-01 -7.65203476e-01
-1.75874233e+00 -7.73699999e-01 -2.06163805e-02 -1.34340313e-04
2.37835199e-01 -9.77160394e-01 -7.44221032e-01 4.60472494e-01
-1.40269125e+00 -3.95075744e-03 3.59670818e-01 4.95108515e-02
-4.12301749e-01 8.93391743e-02 -1.01266575e+00 -6.52928412e-01
-7.74846971e-01 -5.68002880e-01 2.05224976e-01 -3.12691420e-01
-7.49760926e-01 -6.32016182e-01 4.00605083e-01 9.21754360e-01
6.62380755e-01 4.23974067e-01 4.52553928e-01 -1.31393683e+00
-2.52512731e-02 -4.05097336e-01 -3.87201548e-01 1.23049676e-01
-2.53682047e-01 -9.10735130e-02 -1.03218269e+00 -5.01027226e-01
3.38597596e-01 -8.02214384e-01 1.11563325e+00 -2.16617480e-01
4.61075217e-01 -1.11159003e+00 3.43629308e-02 -1.58253536e-01
8.76045883e-01 -8.95607710e-01 3.26791763e-01 7.93909907e-01
1.83599144e-01 6.43648028e-01 5.45619726e-01 7.94616044e-01
5.01551569e-01 6.06814623e-01 6.91196144e-01 5.36940753e-01
-1.34056315e-01 -4.10708904e-01 7.97954679e-01 8.27180207e-01
1.53572068e-01 -3.32850754e-01 -9.24635828e-01 6.14789605e-01
-1.85105538e+00 -1.04407251e+00 -9.87847507e-01 1.72885811e+00
9.82668817e-01 7.82708287e-01 4.05200481e-01 5.33586033e-02
5.27881026e-01 6.84822738e-01 -2.02039048e-01 -8.32282126e-01
-1.22862190e-01 -5.39988697e-01 2.01422021e-01 8.26879680e-01
-9.15367484e-01 7.46042550e-01 5.67096996e+00 2.99728185e-01
-1.22368658e+00 4.57319945e-01 3.46810758e-01 -2.45479122e-01
-1.64656460e-01 -4.46115099e-02 -8.00145268e-01 5.44356883e-01
1.44660127e+00 -5.01309276e-01 9.00458843e-02 6.72752023e-01
5.36348760e-01 -1.32885247e-01 -8.80933046e-01 2.18747810e-01
5.98126054e-01 -1.54673779e+00 -2.98138589e-01 -2.76820306e-02
7.95428872e-01 8.27589095e-01 7.83248320e-02 5.15720367e-01
2.79305339e-01 -1.01587737e+00 1.18609321e+00 2.81586468e-01
3.19605134e-02 -5.21765947e-01 1.08184111e+00 9.80025291e-01
1.47905648e-01 2.77904600e-01 -2.64416546e-01 -4.38601106e-01
2.55087256e-01 8.99720013e-01 -1.70891392e+00 1.70654401e-01
4.31333512e-01 8.28122795e-01 -5.88826716e-01 6.78047836e-01
-5.23259878e-01 1.19635582e+00 1.10048234e-01 -3.21685463e-01
6.08881712e-01 6.27975821e-01 1.16873252e+00 1.34213209e+00
-1.94231704e-01 -1.58924848e-01 2.14938506e-01 7.24205911e-01
-5.80172002e-01 1.05389632e-01 -6.29450679e-02 -5.58919162e-02
2.93029666e-01 1.03848016e+00 -2.16521695e-01 -5.89099944e-01
1.12791680e-01 7.83935070e-01 2.53337443e-01 -8.23985487e-02
-6.93386734e-01 5.78541532e-02 1.66550934e-01 6.23969018e-01
1.49998680e-01 3.76473218e-01 -3.34333241e-01 -1.04464340e+00
-2.94138610e-01 -1.23030305e+00 3.79544258e-01 -6.92169905e-01
-1.51210248e+00 1.06117606e+00 -5.17449081e-01 -7.28707850e-01
-4.86747354e-01 -1.40362889e-01 -8.42124164e-01 9.47732985e-01
-1.77436447e+00 -9.94181275e-01 -2.92677611e-01 2.19515890e-01
5.96204996e-01 -4.50010188e-02 5.54557681e-01 -4.98237200e-02
-4.10610378e-01 2.44509846e-01 -6.29116818e-02 -8.64579082e-02
1.16529858e+00 -1.04061735e+00 5.09143412e-01 3.59521896e-01
-2.33358011e-01 4.23091054e-01 1.81086195e+00 -6.07416332e-01
-8.57005596e-01 -9.96895194e-01 1.75206172e+00 -8.68169785e-01
1.49380338e+00 5.12164459e-02 -1.30196989e+00 6.78572893e-01
5.99587977e-01 -2.00800970e-01 4.65748489e-01 4.77215499e-01
-6.69191539e-01 5.49062788e-01 -8.83383751e-01 1.21886648e-01
2.75121063e-01 -5.05088270e-01 -1.16987479e+00 6.72616959e-01
5.63756049e-01 -6.20243669e-01 -5.78265965e-01 -3.66197862e-02
3.70626628e-01 -1.17567730e+00 3.41638774e-01 -9.86449838e-01
1.24340630e+00 -3.00483815e-02 1.68314613e-02 -1.25641465e+00
-1.99159443e-01 -8.16922843e-01 -2.56053329e-01 8.61134052e-01
7.44040310e-01 -4.16449308e-01 7.30785787e-01 -1.24408744e-01
-1.25101075e-01 -6.55549049e-01 -7.17374861e-01 -4.09149915e-01
1.18916295e-01 -5.66539653e-02 4.30560522e-02 1.29947603e+00
4.55958426e-01 8.68050158e-01 -7.79157162e-01 -3.48366462e-02
6.72468066e-01 6.70748293e-01 7.49965429e-01 -1.25112784e+00
1.62032265e-02 -6.81936026e-01 4.26780246e-02 -7.91695178e-01
5.94146699e-02 -1.10275102e+00 1.67084798e-01 -1.21511769e+00
3.11400473e-01 1.10825576e-01 -3.25280309e-01 1.98721245e-01
-8.87414515e-02 3.21474493e-01 1.28292115e-02 9.94648397e-01
-7.91241825e-01 3.10760885e-01 1.03521430e+00 -1.59157753e-01
1.07519962e-01 4.94277626e-01 -8.73902440e-01 9.07633603e-01
7.72214532e-01 -7.85851300e-01 1.10199109e-01 1.02244154e-01
7.13853121e-01 3.55462939e-01 6.95823193e-01 -3.94344956e-01
3.68841261e-01 -4.56623062e-02 -4.08586711e-01 -8.67750645e-01
1.62810758e-01 -2.17951108e-02 -2.55317956e-01 5.41146636e-01
-7.58799672e-01 2.47334242e-01 -1.44117951e-01 7.17799723e-01
-1.78666085e-01 -2.27795213e-01 8.38651597e-01 -3.19784991e-02
-3.53433728e-01 -2.10146338e-01 -6.53556347e-01 5.79376936e-01
5.89379132e-01 4.49690759e-01 -1.01198554e+00 -5.88484466e-01
-8.23247135e-01 1.52049348e-01 5.04158556e-01 5.92738986e-01
4.45225149e-01 -6.92907810e-01 -1.77335989e+00 -3.84786099e-01
2.44280025e-01 -8.13830614e-01 -5.40295281e-02 1.39065742e+00
-3.43864888e-01 5.31935036e-01 -8.84160027e-02 -2.82196730e-01
-1.26638889e+00 3.03387582e-01 1.89684287e-01 -3.20100397e-01
-8.11646163e-01 7.49460816e-01 -5.41456461e-01 1.95555110e-02
-9.44381505e-02 1.42418027e-01 -3.75909030e-01 5.25349140e-01
1.08504748e+00 7.42882669e-01 5.48907220e-01 -7.77599514e-01
-3.51183295e-01 -6.02375507e-01 -5.87495565e-01 -9.89866853e-02
1.08878160e+00 -3.23768109e-01 -5.11636317e-01 9.54040885e-01
1.12815714e+00 2.63594121e-01 -4.67302352e-01 -6.80087745e-01
4.91792530e-01 -3.81877005e-01 4.50520307e-01 -1.26329625e+00
-4.36757982e-01 5.40421605e-01 -2.02726752e-01 8.94697726e-01
-1.30652264e-01 6.23122118e-02 9.87973273e-01 2.84176409e-01
1.30856082e-01 -8.26105952e-01 1.61808535e-01 1.07277715e+00
1.58198249e+00 -1.33470345e+00 4.65372086e-01 -1.06856391e-01
-8.73662651e-01 9.29902911e-01 2.21418723e-01 -3.04420650e-01
4.95671570e-01 -2.64790326e-01 2.61873752e-01 -7.53715396e-01
-1.45797491e+00 5.02266884e-01 2.55154252e-01 -3.03937525e-01
7.32353270e-01 3.97852182e-01 -6.84648395e-01 5.78348756e-01
-6.94913745e-01 1.01265185e-01 1.25967860e+00 6.15671337e-01
-9.63997841e-01 -1.11396834e-01 -2.76542187e-01 2.97497779e-01
-9.59074259e-01 -8.51877555e-02 -9.11699355e-01 5.69644213e-01
-3.10466588e-01 1.20994020e+00 -1.87217623e-01 -4.50900167e-01
2.53169388e-01 5.22867180e-02 -2.23317370e-01 -5.94565392e-01
-1.07168531e+00 -3.20447206e-01 8.85199428e-01 -4.98394459e-01
-2.01515108e-01 -8.42860758e-01 -9.93571162e-01 -1.15212691e+00
-2.15444028e-01 7.59190202e-01 5.39172292e-01 9.68572795e-01
1.98106289e-01 2.33962595e-01 8.16085637e-01 -3.54763389e-01
-1.17734957e+00 -1.54694831e+00 -4.32193130e-01 3.60600173e-01
9.93078589e-01 -3.99055064e-01 -7.85381734e-01 -1.93559989e-01] | [8.230732917785645, 10.116455078125] |
07d63608-3e56-45ae-b73d-4066dcb97851 | face-sketch-synthesis-via-semantic-driven | 2106.15121 | null | https://arxiv.org/abs/2106.15121v1 | https://arxiv.org/pdf/2106.15121v1.pdf | Face Sketch Synthesis via Semantic-Driven Generative Adversarial Network | Face sketch synthesis has made significant progress with the development of deep neural networks in these years. The delicate depiction of sketch portraits facilitates a wide range of applications like digital entertainment and law enforcement. However, accurate and realistic face sketch generation is still a challenging task due to the illumination variations and complex backgrounds in the real scenes. To tackle these challenges, we propose a novel Semantic-Driven Generative Adversarial Network (SDGAN) which embeds global structure-level style injection and local class-level knowledge re-weighting. Specifically, we conduct facial saliency detection on the input face photos to provide overall facial texture structure, which could be used as a global type of prior information. In addition, we exploit face parsing layouts as the semantic-level spatial prior to enforce globally structural style injection in the generator of SDGAN. Furthermore, to enhance the realistic effect of the details, we propose a novel Adaptive Re-weighting Loss (ARLoss) which dedicates to balance the contributions of different semantic classes. Experimentally, our extensive experiments on CUFS and CUFSF datasets show that our proposed algorithm achieves state-of-the-art performance. | ['Caifeng Shan', 'Qi Li', 'Xiaoxiao Dong', 'Weining Wang', 'Muyi Sun', 'Xingqun Qi'] | 2021-06-29 | null | null | null | null | ['face-sketch-synthesis', 'face-parsing'] | ['computer-vision', 'computer-vision'] | [ 2.41858318e-01 -3.71102691e-02 5.02092466e-02 -4.88100886e-01
-2.30176806e-01 -3.56372714e-01 6.45285964e-01 -6.47893190e-01
1.73023283e-01 5.76402366e-01 1.78129539e-01 1.42173573e-01
1.16209224e-01 -9.11741972e-01 -7.80510187e-01 -6.02686524e-01
4.95608330e-01 -5.98548912e-02 2.65795648e-01 -3.86755139e-01
1.08013488e-01 6.80215955e-01 -1.34964907e+00 2.56473899e-01
1.03827560e+00 1.15442836e+00 1.68423966e-01 1.23688318e-01
-4.03934956e-01 6.40914977e-01 -5.78721702e-01 -8.51803899e-01
3.13338190e-01 -6.12214446e-01 -1.83519244e-01 2.75238305e-01
6.15990758e-01 -4.52731311e-01 -5.29187560e-01 1.33269882e+00
4.92098629e-01 -6.56497851e-02 4.89409626e-01 -1.35725558e+00
-7.38584697e-01 1.20238163e-01 -8.50509226e-01 -1.17328100e-01
9.17475596e-02 4.48267281e-01 5.94360888e-01 -9.11062956e-01
6.32835984e-01 1.63850510e+00 3.17859501e-01 7.44856238e-01
-9.99501169e-01 -1.14065325e+00 3.26296657e-01 1.42500132e-01
-1.21873164e+00 -5.24149656e-01 1.45747662e+00 -7.68716484e-02
5.93636967e-02 8.76448303e-02 5.79811573e-01 1.22833514e+00
-1.23281620e-01 8.42341065e-01 1.04662466e+00 -2.09279969e-01
6.12666160e-02 1.64026722e-01 -7.23948717e-01 9.69772398e-01
1.04812473e-01 6.50555044e-02 -4.85171884e-01 6.36531860e-02
1.32740009e+00 2.38715395e-01 -2.96543568e-01 -3.21621805e-01
-7.57048965e-01 5.38914263e-01 5.93169630e-01 -3.49587798e-02
-2.01369077e-01 1.40792564e-01 2.00444698e-01 -2.22392147e-03
5.36303520e-01 1.82666510e-01 -4.20943880e-03 3.27433437e-01
-9.71143484e-01 4.00653094e-01 2.63570487e-01 1.03111923e+00
6.58658087e-01 4.60972250e-01 -4.48672324e-01 1.08286691e+00
3.44434083e-01 6.84661388e-01 1.40255943e-01 -8.78074408e-01
4.30901557e-01 7.34397769e-01 -6.76247030e-02 -1.31181312e+00
1.56346083e-01 -5.02264202e-01 -9.91722167e-01 2.93551862e-01
2.41328254e-01 2.49606147e-02 -9.93997514e-01 1.85012627e+00
3.81811500e-01 4.49466258e-01 -2.93795377e-01 1.13586032e+00
7.94385254e-01 5.84818065e-01 2.40850791e-01 6.75117597e-02
1.28437781e+00 -9.75436032e-01 -8.04788411e-01 -3.11438769e-01
-2.42781252e-01 -8.79639208e-01 1.26303136e+00 9.91936848e-02
-1.09759927e+00 -6.88343585e-01 -1.08034432e+00 -1.40515834e-01
7.52655566e-02 3.58211339e-01 6.23875260e-01 6.39650285e-01
-6.82351291e-01 3.71241897e-01 -5.36462784e-01 2.88283266e-03
1.04535413e+00 1.44467026e-01 -2.23028272e-01 -2.18543530e-01
-1.07411599e+00 3.08781624e-01 -8.29001069e-02 2.54836977e-01
-9.03596163e-01 -8.14969540e-01 -7.38208532e-01 3.17376375e-01
5.15519023e-01 -6.26521051e-01 7.23196566e-01 -1.37224412e+00
-1.68096817e+00 7.07833767e-01 -9.87689197e-02 3.50598484e-01
7.27084816e-01 2.60707401e-02 -4.02240664e-01 8.90683383e-02
-1.76701490e-02 6.79027259e-01 1.22230446e+00 -1.36324930e+00
-3.02330673e-01 -4.77998435e-01 1.47316858e-01 1.56967536e-01
-4.68707979e-01 1.01141408e-01 -6.54904842e-01 -1.20735550e+00
-1.32704392e-01 -7.07980156e-01 -4.19196412e-02 5.90489984e-01
-3.32114428e-01 -2.87393131e-03 9.89610672e-01 -7.02106476e-01
1.00686634e+00 -2.28178763e+00 2.47075051e-01 1.65921062e-01
8.93386900e-02 3.82275432e-01 -4.02122229e-01 1.00621544e-01
6.78196177e-02 1.28324190e-03 -3.14295799e-01 -4.41889763e-01
-6.95431009e-02 1.22246025e-02 -4.56875324e-01 2.36390233e-02
7.19160438e-01 9.34139371e-01 -7.74862170e-01 -3.56257290e-01
1.49157261e-02 7.32274592e-01 -6.12260044e-01 3.91982317e-01
-5.05874515e-01 4.85741794e-01 -7.24238694e-01 7.72945285e-01
1.10584259e+00 -4.04338278e-02 9.52516347e-02 -3.64601791e-01
2.92373925e-01 -1.68665081e-01 -9.69761014e-01 1.87811518e+00
-3.29536706e-01 2.30058536e-01 2.47405857e-01 -7.22957551e-01
1.07995713e+00 -8.32509845e-02 -3.08365636e-02 -7.80770183e-01
1.84158161e-01 1.79199517e-01 -1.34149492e-01 -2.41781890e-01
1.20742604e-01 -2.28623196e-01 2.24939167e-01 2.03830257e-01
-7.81959668e-02 -1.67071491e-01 -3.10159057e-01 1.39483184e-01
6.86947942e-01 3.07914257e-01 -2.26031974e-01 -2.91210711e-01
7.67272532e-01 -5.93487501e-01 8.42888355e-01 -2.63825823e-02
-5.12001440e-02 9.30746973e-01 7.08871067e-01 -4.69131023e-01
-1.09700537e+00 -1.00219810e+00 1.96547434e-01 8.89864266e-01
2.54750550e-01 -1.01746768e-01 -1.10797751e+00 -8.70303690e-01
-8.85898769e-02 3.26595813e-01 -6.47253573e-01 -3.61785293e-01
-6.37765229e-01 -4.99181062e-01 4.97944146e-01 6.44101918e-01
9.76529121e-01 -1.25422633e+00 -5.33630140e-02 -2.37904191e-02
7.98784643e-02 -1.16917300e+00 -8.22645187e-01 -6.60693884e-01
-4.74764377e-01 -9.38824415e-01 -1.05551720e+00 -8.07819664e-01
9.00044918e-01 1.92860693e-01 8.01165223e-01 3.06770712e-01
-4.12581295e-01 -9.74445045e-02 -1.05930440e-01 -3.69661659e-01
-1.13098159e-01 -5.11652231e-02 -2.49102041e-01 5.40037513e-01
1.91176916e-03 -7.62938559e-01 -9.39244032e-01 3.41393918e-01
-1.10355711e+00 2.18611702e-01 5.04131079e-01 9.22975659e-01
3.54062915e-01 7.27489963e-02 6.67691708e-01 -8.89730036e-01
4.95867163e-01 -3.06947082e-01 -6.27737820e-01 3.58923793e-01
-2.11182907e-01 -3.60598080e-02 7.73319364e-01 -4.41148967e-01
-1.48875856e+00 -7.73429573e-02 -3.48819673e-01 -7.84882843e-01
-6.84442744e-02 -7.57521763e-02 -9.06978607e-01 -2.16296881e-01
2.43768722e-01 3.47509205e-01 1.31172106e-01 -3.51198196e-01
2.56256968e-01 3.60622317e-01 4.76224571e-01 -8.23338330e-01
9.67042923e-01 4.75916326e-01 1.39061689e-01 -7.19877720e-01
-7.90280759e-01 4.11850184e-01 -2.03265905e-01 -1.55220225e-01
6.85062766e-01 -8.75090480e-01 -5.03949761e-01 7.50128090e-01
-1.18971014e+00 -2.72413313e-01 -1.04866866e-02 -1.34961501e-01
-3.21423382e-01 3.38651896e-01 -3.92977595e-01 -6.38741314e-01
-3.58584106e-01 -1.33177710e+00 1.21469152e+00 5.44042110e-01
2.85277992e-01 -7.16453075e-01 -4.40698236e-01 3.40065151e-01
5.71257353e-01 4.08180267e-01 9.75703180e-01 6.81175515e-02
-9.63047206e-01 2.01927409e-01 -7.09004045e-01 5.06540179e-01
3.24377328e-01 4.76063192e-02 -9.88145709e-01 -9.89275649e-02
-1.88298598e-01 -2.52010614e-01 7.57589817e-01 6.04849830e-02
1.56329632e+00 -3.34293872e-01 -1.14952557e-01 8.50191832e-01
1.21893132e+00 1.69401214e-01 9.20256436e-01 -2.22412705e-01
1.11146903e+00 7.90020704e-01 4.68488097e-01 4.04973507e-01
2.06129938e-01 7.81440675e-01 3.29105675e-01 -2.95042276e-01
-5.64962506e-01 -6.30629063e-01 2.03440376e-02 3.27065855e-01
-1.20690234e-01 -1.47693679e-01 -4.22927648e-01 2.87923485e-01
-1.58554244e+00 -8.00219893e-01 3.82988721e-01 1.95204103e+00
7.85974205e-01 2.82106493e-02 -4.60476093e-02 -5.47415614e-02
7.47349799e-01 3.57941061e-01 -6.60405099e-01 1.07524954e-02
-2.11077452e-01 4.07077819e-01 3.74048650e-02 1.54408202e-01
-8.00285220e-01 1.23591280e+00 4.87016201e+00 1.35097814e+00
-1.40848756e+00 5.37743792e-02 1.01108003e+00 1.97189171e-02
-7.17804492e-01 -2.94399530e-01 -6.09107256e-01 9.06243980e-01
5.44890761e-02 1.51252136e-01 6.35258377e-01 7.19548941e-01
8.31380710e-02 2.65576601e-01 -6.37725770e-01 1.06590509e+00
1.35377899e-01 -1.29086757e+00 3.89020473e-01 -8.29382539e-02
8.90511930e-01 -6.42846584e-01 4.21615690e-01 6.48371950e-02
-1.58742145e-02 -1.01869249e+00 7.51137674e-01 5.55188596e-01
1.27525973e+00 -1.05324507e+00 3.67701799e-01 -2.14328896e-02
-1.16999304e+00 1.71674564e-02 -3.91391009e-01 2.81206399e-01
1.99430976e-02 4.70325530e-01 -2.82702386e-01 3.64015013e-01
4.56308454e-01 6.00351691e-01 -4.47180420e-01 6.01396263e-01
-4.49978679e-01 3.51591378e-01 -2.09378675e-01 2.31819525e-01
9.66117904e-02 -2.13687882e-01 4.23789412e-01 8.43614876e-01
2.89954007e-01 2.68131942e-01 1.25987912e-02 1.29014361e+00
-4.77812022e-01 8.40624794e-02 -4.19900209e-01 -1.07465148e-01
5.20922422e-01 1.25984681e+00 -6.20295703e-01 -6.44240528e-02
-3.29847991e-01 1.24809885e+00 3.23347658e-01 4.01225150e-01
-8.40571165e-01 -4.20848012e-01 9.76652265e-01 5.07113636e-01
2.44034410e-01 5.28754629e-02 -2.39547238e-01 -1.23982656e+00
1.60478622e-01 -8.77593935e-01 -2.60004163e-01 -6.75984383e-01
-1.24198687e+00 7.00134456e-01 -3.23807418e-01 -9.85489070e-01
2.28698015e-01 -3.87099266e-01 -9.48718071e-01 9.81807053e-01
-1.73515999e+00 -1.57298481e+00 -5.97590625e-01 5.42094290e-01
7.34076083e-01 -4.36006725e-01 3.76471072e-01 5.62849164e-01
-7.16509461e-01 9.87047434e-01 -3.02387536e-01 2.54691392e-01
8.80047321e-01 -7.90739059e-01 4.93199438e-01 8.41602504e-01
-2.72360206e-01 5.60780883e-01 4.24880087e-01 -7.45237410e-01
-1.32852829e+00 -1.34637892e+00 3.39056969e-01 -2.71754488e-02
2.21582502e-01 -5.39858818e-01 -9.51061487e-01 2.87611872e-01
8.48339051e-02 2.91855723e-01 1.85144976e-01 -3.98217916e-01
-5.15574694e-01 -4.51091975e-01 -1.32425082e+00 7.22777247e-01
1.31167722e+00 -5.01962364e-01 -5.26402891e-02 5.07040471e-02
6.16044462e-01 -3.67155403e-01 -4.79530692e-01 6.71967447e-01
8.18648458e-01 -1.03096235e+00 9.50201869e-01 -3.52829665e-01
7.85655558e-01 -4.19909894e-01 1.21112011e-01 -1.19193089e+00
-2.02320442e-01 -5.70115566e-01 1.75577715e-01 1.55178952e+00
-1.13634035e-01 -4.53740060e-01 9.84580696e-01 5.01308382e-01
-5.58197945e-02 -8.60638797e-01 -5.42290270e-01 -4.47563499e-01
-1.36789128e-01 9.17164832e-02 1.03870773e+00 8.42501521e-01
-6.93269730e-01 9.65370834e-02 -5.65069854e-01 7.08575733e-03
7.44060218e-01 5.31129353e-02 8.01145494e-01 -1.00133288e+00
-9.38687995e-02 -5.63779891e-01 -4.44527805e-01 -9.34183300e-01
3.79754037e-01 -5.18868208e-01 -2.74810135e-01 -1.06830084e+00
1.13922358e-01 -7.76971400e-01 -1.79780915e-01 3.76183093e-01
-4.44614589e-01 4.52853292e-01 2.81256914e-01 -6.70521408e-02
-2.52565145e-01 1.03200734e+00 1.79032087e+00 -2.06175745e-01
1.74877226e-01 -1.84791103e-01 -7.97426939e-01 6.15884125e-01
5.41727841e-01 -1.84256509e-01 -7.43551075e-01 -5.16619325e-01
-2.54335534e-02 1.97850745e-02 5.13234258e-01 -8.31231833e-01
-1.76449716e-02 -3.25446665e-01 6.81216776e-01 -2.40482375e-01
5.06086051e-01 -7.74716437e-01 1.42823741e-01 1.32092774e-01
-1.23975992e-01 -2.92280078e-01 2.14739770e-01 5.75672567e-01
-3.32116932e-01 2.52991676e-01 1.10716140e+00 -1.13282360e-01
-5.31539142e-01 6.48499012e-01 4.65214491e-01 -1.10152867e-02
8.61399353e-01 -1.12842053e-01 -2.17862606e-01 -2.64047742e-01
-2.04399794e-01 6.53860122e-02 7.31035829e-01 6.74067438e-01
7.95180380e-01 -1.53080463e+00 -7.52587438e-01 7.86018968e-01
9.00404528e-03 1.37118638e-01 6.01724923e-01 2.41743982e-01
-5.29733777e-01 5.13311774e-02 -5.52368224e-01 -2.62294799e-01
-9.95308042e-01 3.65769953e-01 1.48391128e-01 8.96003395e-02
-4.61800277e-01 9.84112978e-01 9.14068758e-01 -2.83453226e-01
1.66617140e-01 6.05039634e-02 -7.08863512e-02 -3.06420058e-01
7.22452581e-01 1.52847841e-01 -1.39244288e-01 -5.14186084e-01
-1.83687359e-01 6.06297851e-01 -1.71650916e-01 1.87956974e-01
1.14151561e+00 1.96694564e-02 -1.19974710e-01 -2.67816991e-01
9.83082533e-01 1.94570348e-01 -1.67274177e+00 -1.21087104e-01
-4.99167979e-01 -8.16846073e-01 1.57236885e-02 -8.27421129e-01
-1.59926045e+00 1.07780564e+00 4.97370958e-01 -5.11641085e-01
1.29528904e+00 -2.69020677e-01 1.02585363e+00 -1.25210464e-01
3.97796273e-01 -7.12487221e-01 3.27380776e-01 8.78221020e-02
1.08546996e+00 -1.03356528e+00 -2.50110954e-01 -8.02974164e-01
-6.27764344e-01 9.40566957e-01 9.09249127e-01 -3.54496777e-01
5.40225506e-01 1.75693691e-01 -1.23708867e-01 -2.01416444e-02
-3.90405387e-01 1.76100224e-01 4.60400462e-01 4.14030850e-01
1.98985785e-01 -5.90337552e-02 -1.28622502e-01 7.59916782e-01
1.13285065e-01 1.17667019e-01 1.33377880e-01 5.55905044e-01
-1.97040185e-01 -1.25321126e+00 -2.05519482e-01 2.59650201e-01
-4.95984763e-01 -5.91793843e-02 -2.22722918e-01 3.89860719e-01
2.59314448e-01 4.70013738e-01 -1.07279927e-01 -2.73871899e-01
3.38073283e-01 -2.06300959e-01 6.60809457e-01 -4.60052878e-01
-2.35850036e-01 2.42872179e-01 -2.99087971e-01 -5.19975424e-01
-1.76795840e-01 -2.24953473e-01 -9.71031308e-01 -3.61922354e-01
-5.17512485e-02 -1.63998052e-01 6.01591468e-01 6.83475018e-01
5.18844724e-01 6.80324137e-01 8.22759509e-01 -9.00025845e-01
-3.32186222e-01 -8.29369664e-01 -5.77427685e-01 4.66577381e-01
1.52843103e-01 -9.68007505e-01 -1.60806760e-01 -1.86144635e-02] | [12.536463737487793, -0.15038596093654633] |
92f73e67-9a26-4767-9258-7bf7b6697ce6 | scalable-modular-synthetic-data-generation | 2211.05335 | null | https://arxiv.org/abs/2211.05335v2 | https://arxiv.org/pdf/2211.05335v2.pdf | Scalable Modular Synthetic Data Generation for Advancing Aerial Autonomy | One major barrier to advancing aerial autonomy has been collecting large-scale aerial datasets for training machine learning models. Due to costly and time-consuming real-world data collection through deploying drones, there has been an increasing shift towards using synthetic data for training models in drone applications. However, to increase widespread generalization and transferring models to real-world, increasing the diversity of simulation environments to train a model over all the varieties and augmenting the training data, has been proved to be essential. Current synthetic aerial data generation tools either lack data augmentation or rely heavily on manual workload or real samples for configuring and generating diverse realistic simulation scenes for data collection. These dependencies limit scalability of the data generation workflow. Accordingly, there is a major challenge in balancing generalizability and scalability in synthetic data generation. To address these gaps, we introduce a scalable Aerial Synthetic Data Augmentation (ASDA) framework tailored to aerial autonomy applications. ASDA extends a central data collection engine with two scriptable pipelines that automatically perform scene and data augmentations to generate diverse aerial datasets for different training tasks. ASDA improves data generation workflow efficiency by providing a unified prompt-based interface over integrated pipelines for flexible control. The procedural generative approach of our data augmentation is performant and adaptable to different simulation environments, training tasks and data collection needs. We demonstrate the effectiveness of our method in automatically generating diverse datasets and show its potential for downstream performance optimization. | ['Sakshi Mishra', 'Praveen Palanisamy', 'Mehrnaz Sabet'] | 2022-11-10 | null | null | null | null | ['synthetic-data-generation', 'synthetic-data-generation'] | ['medical', 'miscellaneous'] | [ 2.89957136e-01 -1.47544548e-01 3.49414527e-01 -3.45123023e-01
-3.58637035e-01 -1.20245433e+00 5.98070323e-01 1.35116756e-01
-2.31070310e-01 5.95168710e-01 -7.61222187e-03 -4.42530841e-01
-9.67381522e-02 -1.05784023e+00 -6.28121555e-01 -1.05317466e-01
-2.24165723e-01 7.09923565e-01 2.56993681e-01 -6.23856306e-01
-2.12508053e-01 9.53166246e-01 -2.21993804e+00 1.80597097e-01
1.02704060e+00 5.47450185e-01 2.68275231e-01 1.11993980e+00
3.18700522e-01 3.70153993e-01 -9.49438035e-01 2.64041930e-01
8.22395742e-01 -2.62914181e-01 -6.09715462e-01 6.70767054e-02
5.69160819e-01 -4.87699509e-01 2.24833246e-02 3.57709467e-01
6.46919727e-01 2.92845577e-01 1.24723628e-01 -1.72695291e+00
1.97433367e-01 7.16324672e-02 1.25419408e-01 -2.37631556e-02
7.46820047e-02 9.41303194e-01 3.61614197e-01 -3.37912410e-01
5.06311715e-01 7.33510494e-01 8.62538218e-01 5.20778120e-01
-1.34664297e+00 -6.50434673e-01 -3.74991377e-03 -6.96521580e-01
-1.26025617e+00 -4.46871340e-01 1.88722342e-01 -7.80505538e-01
9.80854750e-01 5.44013500e-01 1.41715038e+00 9.37757552e-01
-1.17789052e-01 2.72388965e-01 8.30225706e-01 -3.64376195e-02
3.52800071e-01 1.42611787e-01 -4.97184873e-01 6.76182151e-01
3.66774738e-01 3.77258182e-01 -3.08608532e-01 -1.44132495e-01
7.25667477e-01 -2.10015967e-01 -2.52418071e-01 -4.37166035e-01
-1.33595085e+00 4.42751467e-01 3.73075545e-01 -3.33658755e-01
-3.10532033e-01 2.19646424e-01 4.29881781e-01 2.64509588e-01
1.91983342e-01 1.22518933e+00 -8.36317480e-01 -2.78144479e-01
-1.11095059e+00 1.03708923e+00 7.23262191e-01 1.36219478e+00
7.48741210e-01 4.34467584e-01 -4.34649037e-03 5.12908876e-01
-5.15542924e-02 6.69502616e-01 3.38417292e-01 -1.02758765e+00
3.81195933e-01 1.06578934e+00 1.70579836e-01 -6.88650727e-01
-5.51407635e-01 -5.13031065e-01 -3.55716944e-01 7.31874406e-01
1.78502604e-01 -6.01360977e-01 -1.12519324e+00 1.62001061e+00
7.44288325e-01 -2.08204001e-01 2.20509142e-01 8.69156659e-01
7.57432878e-01 6.81424439e-01 1.30975991e-01 4.89252657e-01
6.91188037e-01 -6.19054735e-01 -6.67539984e-02 -4.72895026e-01
9.08812821e-01 -6.35973275e-01 1.45486689e+00 2.58340567e-01
-9.13887799e-01 -7.64311075e-01 -1.31841767e+00 8.12719017e-02
-7.00979114e-01 2.67815650e-01 8.48096490e-01 6.20054126e-01
-1.00443149e+00 7.12369382e-01 -9.40147698e-01 -5.86657465e-01
4.64026779e-01 5.96852243e-01 -3.85709643e-01 2.76667122e-02
-9.16741252e-01 5.92465818e-01 5.25786638e-01 -5.14320806e-02
-1.24743462e+00 -1.29373586e+00 -1.00005007e+00 -2.76156485e-01
2.84821212e-01 -1.06654322e+00 1.43348598e+00 -7.25215077e-01
-1.29248703e+00 4.55848783e-01 6.52376533e-01 -3.92954439e-01
6.56144321e-01 -7.31436312e-02 -2.23257184e-01 -3.66068035e-01
1.49480939e-01 1.38867426e+00 6.06818855e-01 -1.30962157e+00
-6.00841880e-01 -6.45606369e-02 3.45756620e-01 3.74094039e-01
-9.33964625e-02 -3.22947741e-01 1.30740814e-02 -3.82265359e-01
-2.30858088e-01 -1.16907334e+00 -4.79875624e-01 4.37990949e-02
2.50126477e-02 5.52374840e-01 1.26486647e+00 -3.81683975e-01
1.06360567e+00 -1.93820095e+00 2.52290294e-02 4.61140163e-02
-3.53279971e-02 5.48411846e-01 -1.68485194e-01 7.39941835e-01
-4.81693000e-02 2.39218414e-01 -4.28519338e-01 -6.54681399e-02
-9.40103307e-02 3.10627222e-01 -5.30036151e-01 -6.78008795e-02
5.05136907e-01 7.05341101e-01 -9.92238760e-01 -3.92080277e-01
4.81723279e-01 3.75023991e-01 -7.66380847e-01 6.06986225e-01
-6.72933459e-01 6.91579342e-01 -1.32506654e-01 7.44333804e-01
4.91362780e-01 1.35552183e-01 -3.94688621e-02 -6.20517060e-02
-5.74601948e-01 2.63506055e-01 -1.17639089e+00 1.85954726e+00
-7.36368001e-01 5.80310881e-01 1.45819634e-01 -1.50609672e-01
8.92102420e-01 1.02053462e-02 3.84440809e-01 -6.32896125e-02
6.63892925e-02 -2.33301278e-02 6.47598580e-02 -4.96999055e-01
8.95537078e-01 8.08180869e-02 -2.72498459e-01 3.32747132e-01
3.24548855e-02 -1.41931844e+00 4.62922215e-01 2.74389535e-01
1.00677192e+00 9.20393586e-01 1.76896360e-02 -3.26538533e-01
-1.66640013e-01 9.93735075e-01 3.59440356e-01 2.39865929e-01
1.36800990e-01 5.65018177e-01 2.32538730e-01 -7.78611362e-01
-1.29362273e+00 -8.29756975e-01 5.05440757e-02 9.15929079e-01
-2.77585566e-01 -7.95118630e-01 -8.86121392e-01 -5.47395587e-01
1.21387653e-01 8.65681708e-01 -5.11224568e-01 -2.31902122e-01
-1.94611445e-01 -8.36117923e-01 8.10382783e-01 4.37871009e-01
6.04348540e-01 -9.03451741e-01 -1.45388997e+00 4.60315160e-02
1.04043908e-01 -1.00288379e+00 -3.59712206e-02 2.51890659e-01
-8.32495868e-01 -9.69486475e-01 2.16829777e-02 -1.40037730e-01
6.55710340e-01 3.52752805e-01 1.26064575e+00 2.64071435e-01
-8.56946468e-01 5.68479776e-01 -3.74946147e-01 -9.28390682e-01
-6.74451053e-01 4.11018938e-01 8.79110247e-02 -7.63807535e-01
-3.21503103e-01 -5.90063870e-01 -4.92137313e-01 5.20575106e-01
-1.46338892e+00 4.39834148e-01 3.71604830e-01 4.15194660e-01
5.60209394e-01 2.59982597e-04 3.33019584e-01 -5.64183116e-01
6.47109926e-01 -4.29249346e-01 -1.10158908e+00 1.09339431e-02
-4.00984287e-01 -6.76908046e-02 9.23366129e-01 -2.24484801e-01
-6.17699683e-01 5.60479939e-01 6.68370277e-02 -4.24807698e-01
-5.11127532e-01 4.81196493e-01 -2.58785516e-01 -1.70043975e-01
1.02736402e+00 -1.34258464e-01 2.26884633e-01 1.50421143e-01
2.56548792e-01 4.11310017e-01 3.06878477e-01 -6.86528444e-01
1.03332007e+00 1.97898045e-01 1.21428698e-01 -1.06874120e+00
-4.91231680e-01 9.24524963e-02 -9.13988888e-01 -4.27441150e-01
6.60519361e-01 -1.01765919e+00 -3.24153364e-01 4.02160227e-01
-6.08620882e-01 -1.21405149e+00 -8.09592009e-01 2.99167067e-01
-4.23038334e-01 -1.64972484e-01 3.54489952e-01 -4.85672146e-01
-2.47177228e-01 -1.36370003e+00 1.23154676e+00 1.69365242e-01
-4.49520797e-01 -6.24419093e-01 2.88561553e-01 5.19294590e-02
6.76761031e-01 8.03747833e-01 5.64278185e-01 -1.63969725e-01
-8.79673302e-01 -5.06170213e-01 1.41594246e-01 9.75300744e-02
2.17458412e-01 6.57453895e-01 -8.59321654e-01 -3.38794351e-01
-6.79586172e-01 -7.92579114e-01 2.47877374e-01 -3.46850008e-02
1.15225470e+00 -7.62480572e-02 -2.92472482e-01 8.40928376e-01
1.23140788e+00 -7.85200968e-02 4.71875191e-01 2.89265603e-01
5.14634907e-01 6.42777801e-01 1.07178771e+00 5.49880743e-01
3.20735455e-01 5.97248971e-01 6.86250567e-01 -4.20169324e-01
3.76744904e-02 -4.19609725e-01 1.10666364e-01 6.02214821e-02
-1.15530141e-01 -3.72651756e-01 -1.24088597e+00 3.80661875e-01
-1.49797797e+00 -8.86361361e-01 -1.98981494e-01 2.46732426e+00
5.35933316e-01 -1.77159876e-01 3.11981529e-01 6.24302328e-02
5.62483743e-02 1.37869408e-03 -6.63544893e-01 -2.14135647e-01
1.47605687e-01 3.02215219e-01 6.80358708e-01 2.71519423e-01
-1.11890864e+00 1.20332253e+00 6.30852890e+00 1.40225127e-01
-1.21680117e+00 -5.37306607e-01 2.77723074e-01 -4.60792363e-01
-1.37408033e-01 1.68941751e-01 -8.54560614e-01 1.79094777e-01
1.03584611e+00 -1.72486171e-01 6.06719434e-01 1.06454051e+00
5.00924170e-01 -2.88832009e-01 -1.01333070e+00 6.13728046e-01
-1.70345053e-01 -1.42204452e+00 1.45114839e-01 2.35502318e-01
6.69679344e-01 1.93621725e-01 -1.28059685e-01 3.74594718e-01
8.55678439e-01 -1.08129799e+00 6.08228624e-01 2.60013074e-01
1.09066236e+00 -7.77993739e-01 3.74415815e-01 3.11123401e-01
-1.27302074e+00 7.09409788e-02 -1.35523766e-01 -2.54993021e-01
2.34079100e-02 1.48623779e-01 -1.55736232e+00 4.71301883e-01
8.10642481e-01 3.71079981e-01 -9.47249115e-01 9.71959114e-01
-1.53474147e-02 4.91619587e-01 -6.03230476e-01 2.00315654e-01
1.61713108e-01 -1.07618667e-01 4.45018649e-01 9.38730717e-01
3.90439868e-01 5.26898429e-02 4.67894971e-01 7.53642380e-01
3.20470124e-01 -2.05397576e-01 -1.18598771e+00 -3.21592748e-01
5.32796383e-01 1.49223375e+00 -6.64661765e-01 -1.46310672e-01
7.63213634e-02 6.90504730e-01 6.10927083e-02 -7.19991997e-02
-1.01317036e+00 -4.34530437e-01 1.10993004e+00 5.26012063e-01
-2.22504824e-01 -7.93996513e-01 -2.30149165e-01 -6.87585354e-01
-3.14487457e-01 -1.22947896e+00 1.50753856e-01 -9.72455680e-01
-5.04279196e-01 9.04835999e-01 4.38036412e-01 -1.64020014e+00
-2.53911287e-01 -3.68423402e-01 -4.30841357e-01 8.44997466e-01
-9.41036344e-01 -1.58711076e+00 -1.27427924e+00 2.85446048e-01
6.37772799e-01 -2.54292876e-01 1.04574478e+00 -1.16964635e-02
-4.99345690e-01 1.16028726e-01 -6.57000840e-01 -3.10423523e-01
5.70239663e-01 -1.20566881e+00 1.05827534e+00 9.76209700e-01
-1.92184657e-01 4.52162147e-01 7.96219051e-01 -9.18725610e-01
-1.47235441e+00 -1.62192929e+00 -6.61694854e-02 -7.45473385e-01
5.00815570e-01 -6.00212097e-01 -5.64583302e-01 7.41759300e-01
-3.01293638e-02 -5.02410568e-02 7.69845784e-01 -3.88236701e-01
2.10630208e-01 -2.56272703e-01 -1.09749782e+00 6.59973383e-01
1.18279648e+00 1.72665343e-02 -5.59984182e-04 4.19205755e-01
7.92548716e-01 -7.06502974e-01 -9.70233679e-01 8.73144686e-01
3.84041995e-01 -1.02034497e+00 7.88301229e-01 -6.88055456e-01
4.38984275e-01 -8.49974096e-01 -8.20536688e-02 -1.68286872e+00
-3.31769772e-02 -7.01335669e-01 6.47675395e-01 1.06208658e+00
7.25790918e-01 -4.32056487e-01 7.05976963e-01 9.77956533e-01
-6.91771805e-01 -5.07292926e-01 -1.30255431e-01 -5.09221435e-01
-2.61206657e-01 -5.13069808e-01 1.17437530e+00 8.93718421e-01
-3.56150180e-01 1.88593164e-01 -6.21315278e-02 5.05540848e-01
6.35465682e-02 -1.03125917e-02 1.83531737e+00 -1.25093281e+00
-3.31938803e-01 4.44401875e-02 -4.68243957e-01 -5.77146232e-01
-3.00315410e-01 -7.22629726e-01 5.53613678e-02 -1.75284994e+00
-6.24237597e-01 -1.02086699e+00 7.65175343e-01 7.49926209e-01
4.11764644e-02 5.88277765e-02 -1.73890796e-02 -1.33318454e-02
-6.12642430e-02 5.66003442e-01 1.06403589e+00 1.98891848e-01
-6.58519864e-01 7.91816562e-02 -3.60230595e-01 5.15470266e-01
1.00081563e+00 -2.32617736e-01 -9.78617728e-01 -6.82757318e-01
2.86749154e-01 -3.21821302e-01 5.33914506e-01 -1.89137661e+00
-6.02806695e-02 -3.38858336e-01 5.42873919e-01 -6.33016050e-01
4.78079557e-01 -8.96955431e-01 5.78348339e-01 3.76254469e-01
-6.42740205e-02 4.74813432e-01 9.89260912e-01 1.33650050e-01
5.10725342e-02 1.21517994e-01 7.23349512e-01 -4.02482629e-01
-5.27971566e-01 4.61064994e-01 -5.45238614e-01 1.50643915e-01
1.37513125e+00 -3.65895033e-01 -2.94627339e-01 -1.05833314e-01
-4.51279789e-01 4.03146684e-01 1.05580080e+00 4.46500123e-01
3.94723862e-01 -7.43609250e-01 -5.32344580e-01 8.29953313e-01
4.14235711e-01 8.08374465e-01 2.21101806e-01 3.06048155e-01
-1.05366611e+00 3.32852788e-02 -3.57670844e-01 -5.79884946e-01
-1.38076639e+00 3.44567686e-01 5.46167254e-01 1.13753982e-01
-3.30294162e-01 6.92443073e-01 -1.68878332e-01 -9.45006490e-01
-3.43526125e-01 -7.80225456e-01 2.97656655e-01 -1.59282997e-01
2.55912095e-01 3.37418675e-01 4.50726986e-01 -3.07180703e-01
-6.85017975e-03 7.02548400e-02 4.35959190e-01 -2.14238286e-01
1.32205236e+00 4.95280206e-01 1.88083977e-01 2.66107172e-01
2.56276041e-01 1.52538978e-02 -1.60578609e+00 6.71023786e-01
-5.19367278e-01 -6.73674047e-01 6.57945946e-02 -1.06132042e+00
-7.33474016e-01 6.34647310e-01 3.96643877e-01 1.63617924e-01
1.07212985e+00 -4.19739097e-01 4.22975659e-01 5.73112965e-01
5.87184548e-01 -1.01702213e+00 -2.20286727e-01 2.63474613e-01
1.21838105e+00 -1.05385995e+00 2.68417299e-01 -3.96278799e-01
-8.99609625e-01 9.30209577e-01 1.14446115e+00 1.57623723e-01
2.43611917e-01 7.78097987e-01 3.61033350e-01 -2.93888181e-01
-8.19088578e-01 -1.91977859e-01 -6.88743964e-02 9.29478526e-01
1.93547353e-01 4.18707728e-03 4.36822772e-01 6.18543178e-02
-8.11133504e-01 1.91283688e-01 7.17539370e-01 1.24856126e+00
-3.22933435e-01 -1.34506476e+00 -2.80870855e-01 4.95929033e-01
3.07516366e-01 -4.01693396e-02 -6.56101704e-01 1.25832987e+00
3.69447082e-01 8.07789445e-01 -3.52510437e-02 -5.41358590e-01
6.09858394e-01 -1.30661979e-01 3.32056522e-01 -9.79630530e-01
-8.23586226e-01 -4.43654299e-01 3.48746747e-01 -4.90627110e-01
-9.38146114e-02 -6.64770663e-01 -1.03481865e+00 -1.68377474e-01
-1.51383638e-01 1.33449242e-01 1.09888315e+00 5.61431885e-01
9.09026384e-01 7.02766538e-01 3.46962154e-01 -1.23219669e+00
-3.01151246e-01 -8.68062794e-01 5.77357486e-02 -2.30972879e-02
-1.62468031e-02 -6.32113457e-01 -8.52178484e-02 2.61673093e-01] | [7.7163004875183105, -1.1988612413406372] |
4c78acbe-ca67-4e4c-8aa1-e29a13e34a7f | hierachial-protein-function-prediction-with | 2007.12804 | null | https://arxiv.org/abs/2007.12804v1 | https://arxiv.org/pdf/2007.12804v1.pdf | Hierachial Protein Function Prediction with Tails-GNNs | Protein function prediction may be framed as predicting subgraphs (with certain closure properties) of a directed acyclic graph describing the hierarchy of protein functions. Graph neural networks (GNNs), with their built-in inductive bias for relational data, are hence naturally suited for this task. However, in contrast with most GNN applications, the graph is not related to the input, but to the label space. Accordingly, we propose Tail-GNNs, neural networks which naturally compose with the output space of any neural network for multi-task prediction, to provide relationally-reinforced labels. For protein function prediction, we combine a Tail-GNN with a dilated convolutional network which learns representations of the protein sequence, making significant improvement in F_1 score and demonstrating the ability of Tail-GNNs to learn useful representations of labels and exploit them in real-world problem solving. | ['Mladen Nikolić', 'Jovana Kovačević', 'Petar Veličković', 'Stefan Spalević'] | 2020-07-24 | null | null | null | null | ['protein-function-prediction'] | ['medical'] | [ 6.17036760e-01 7.72546828e-01 -1.40017658e-01 -4.79227930e-01
-5.57946414e-02 -5.88785529e-01 3.00201684e-01 4.01566476e-01
-9.71498042e-02 6.83946431e-01 3.09174061e-01 -6.56575084e-01
-3.37941915e-01 -1.03700650e+00 -1.17674494e+00 -6.11327946e-01
-4.15279567e-01 6.71044171e-01 1.17700100e-01 -2.79534906e-01
-1.34122387e-01 6.38227403e-01 -1.18365145e+00 6.61410451e-01
4.47931021e-01 8.92368317e-01 1.89845696e-01 6.25071168e-01
-2.54850015e-02 1.40943754e+00 -3.59536380e-01 -3.23377728e-01
-5.57534471e-02 -3.07098299e-01 -1.16833353e+00 -2.67868549e-01
2.06376642e-01 1.76682904e-01 -5.85279405e-01 6.44863546e-01
2.11150512e-01 2.03021705e-01 8.18698347e-01 -1.12311363e+00
-9.85388577e-01 6.49849892e-01 -4.73366715e-02 1.98886141e-01
4.90642279e-01 -8.26068595e-03 1.75170481e+00 -6.71495676e-01
9.81598616e-01 1.31984568e+00 7.81405687e-01 5.34575999e-01
-1.74446452e+00 -1.73505127e-01 1.52784333e-01 9.84620079e-02
-9.44367051e-01 -4.06970419e-02 5.45596361e-01 -3.59954566e-01
1.40233529e+00 7.85207674e-02 5.27399898e-01 1.33006108e+00
2.79509366e-01 6.90512657e-01 5.97537458e-01 -5.33735417e-02
-8.70901644e-02 -4.08717662e-01 2.92145878e-01 1.03109348e+00
-5.42204604e-02 2.13029869e-02 -5.03665209e-01 -2.73205966e-01
5.43022633e-01 2.66128063e-01 -3.61275077e-01 -7.74392545e-01
-1.10548234e+00 8.83061767e-01 1.05583525e+00 1.99370757e-01
-2.02598348e-01 5.60780942e-01 5.60159683e-01 6.55071080e-01
5.87892890e-01 8.63042593e-01 -8.03738832e-01 5.10932982e-01
-3.04295391e-01 1.96734473e-01 9.09392774e-01 8.49108636e-01
8.66620064e-01 -2.42489472e-01 -2.58210778e-01 6.95279717e-01
2.10797980e-01 -7.20185414e-02 2.58608639e-01 -4.03236121e-01
1.48469210e-01 1.06340754e+00 -4.73985583e-01 -9.76824522e-01
-9.36838925e-01 -5.13063490e-01 -8.17335844e-01 -1.70192018e-01
4.85980421e-01 1.46114692e-01 -7.17187643e-01 1.97467041e+00
1.21558927e-01 2.08428577e-01 1.22095540e-01 4.91477042e-01
9.89974558e-01 5.97611129e-01 3.78537774e-01 1.48570761e-01
1.14576542e+00 -8.06594610e-01 -3.15727562e-01 -9.33132172e-02
1.24486804e+00 1.65471062e-02 9.85311270e-01 1.69654474e-01
-7.18377173e-01 -6.11379564e-01 -9.07617331e-01 -4.56299424e-01
-7.02542841e-01 -3.47683996e-01 8.91507924e-01 7.29589611e-02
-1.33695757e+00 1.13341045e+00 -4.41409677e-01 -3.97471875e-01
6.75505519e-01 7.90592909e-01 -5.87642431e-01 -2.57451445e-01
-1.28128839e+00 1.04955232e+00 6.64513826e-01 8.30097273e-02
-1.00281274e+00 -7.50980914e-01 -1.00440967e+00 3.02674741e-01
3.64461601e-01 -6.19369745e-01 8.57126415e-01 -7.29144871e-01
-9.14532423e-01 1.03644407e+00 8.97717550e-02 -6.38235807e-01
4.40417714e-02 8.42933729e-02 -2.27271885e-01 1.48929268e-01
-2.36231506e-01 8.52101505e-01 4.60587114e-01 -1.00193572e+00
-2.53932744e-01 -3.07931304e-01 5.04581988e-01 3.74812819e-02
-2.76613116e-01 -3.24672431e-01 1.59758821e-01 -4.59319711e-01
3.84459458e-02 -8.54878843e-01 -5.33617795e-01 -1.03393555e-01
-6.44876122e-01 -6.32842541e-01 9.10712659e-01 -4.02310640e-01
9.37882781e-01 -2.05672741e+00 5.08226991e-01 3.57047439e-01
8.59442115e-01 4.97143716e-02 -5.87100089e-01 5.76548219e-01
-5.78109503e-01 5.97105250e-02 -1.82964519e-01 1.26473784e-01
-1.23225689e-01 5.02806425e-01 -1.56828538e-01 6.46408856e-01
6.60808146e-01 1.47064579e+00 -1.02620375e+00 -3.45829651e-02
-1.11200362e-01 3.62865776e-01 -5.65910578e-01 4.19121176e-01
-9.05017495e-01 5.21106422e-01 -5.24940550e-01 4.95138735e-01
1.45712808e-01 -9.86594439e-01 7.08406091e-01 -2.45062351e-01
5.77770293e-01 4.57250476e-01 -4.14994836e-01 1.53809023e+00
-2.91083038e-01 4.55615163e-01 -4.37479109e-01 -1.49045074e+00
1.10692036e+00 3.79108548e-01 7.35564232e-01 -6.17187738e-01
-1.50572792e-01 -3.79306115e-02 2.14354604e-01 -3.98289084e-01
3.33578996e-02 -5.36916256e-02 9.55922827e-02 5.02396643e-01
5.16231954e-01 3.35396856e-01 1.20318823e-01 3.41222733e-01
1.60459864e+00 4.96699989e-01 4.23759282e-01 -4.01069731e-01
5.14096320e-01 -3.32963169e-02 5.04697025e-01 5.69767058e-01
1.29523382e-01 3.40015858e-01 1.12299263e+00 -9.29935098e-01
-1.11408234e+00 -8.31084430e-01 -2.92413663e-02 1.73910940e+00
-1.07504733e-01 -6.03176534e-01 -4.55515027e-01 -1.13482845e+00
4.29777831e-01 1.52575433e-01 -7.02126980e-01 -5.67950845e-01
-7.69531667e-01 -5.83677411e-01 5.97477257e-01 5.12197077e-01
-2.74492174e-01 -1.55586445e+00 -1.70615301e-01 3.19754392e-01
1.11001059e-01 -1.07006061e+00 -2.13835582e-01 9.50735271e-01
-9.32123721e-01 -1.59969294e+00 -3.23788762e-01 -1.17915797e+00
7.68602908e-01 -8.66123289e-02 1.62781107e+00 4.18532550e-01
-1.42900363e-01 1.03064798e-01 -2.29691252e-01 -6.47624508e-02
-5.33124626e-01 3.01134765e-01 -2.36807004e-01 -8.28961059e-02
4.39716339e-01 -9.85268533e-01 -4.36330616e-01 2.45594636e-01
-1.02703190e+00 9.57183614e-02 3.96084100e-01 1.03220403e+00
5.85615337e-01 -2.98138618e-01 8.64841223e-01 -1.72013128e+00
5.38486481e-01 -6.83412015e-01 -4.73068506e-01 4.11566824e-01
-4.29975420e-01 5.35080850e-01 1.05346978e+00 -3.58273923e-01
-3.26117009e-01 2.85179496e-01 -1.46201134e-01 -1.74823746e-01
-9.65284854e-02 7.88926661e-01 -4.18303460e-02 -2.68317491e-01
9.44005609e-01 -6.36292296e-03 4.27363962e-02 -4.48324025e-01
6.79756761e-01 1.56771362e-01 3.70619506e-01 -6.34741306e-01
4.36212987e-01 8.19174126e-02 7.10605800e-01 -3.90716046e-01
-1.00869608e+00 -3.43214393e-01 -8.38845313e-01 1.03063799e-01
8.98548484e-01 -7.94537961e-01 -1.40176201e+00 -3.45239080e-02
-1.03586590e+00 -4.21205342e-01 -2.09028199e-01 1.56036671e-02
-8.55807304e-01 2.10684001e-01 -8.79086316e-01 -3.08449030e-01
-2.35533953e-01 -1.06655490e+00 8.92479956e-01 -2.87865728e-01
-2.23325595e-01 -1.37132752e+00 -7.53271878e-02 1.21353842e-01
2.54040331e-01 5.70653975e-01 1.81454086e+00 -1.29764676e+00
-5.40846765e-01 2.08374932e-01 -4.57326442e-01 1.13878421e-01
-9.82259586e-03 -5.56997597e-01 -1.01327693e+00 -5.60451448e-01
-4.50057417e-01 -8.65257442e-01 9.84055459e-01 3.28961343e-01
1.42368853e+00 -3.55978608e-01 -6.19211912e-01 6.83206618e-01
1.38374865e+00 -2.03736052e-01 3.88733089e-01 4.23014723e-02
1.06969523e+00 7.36311495e-01 -3.51327695e-02 1.01598859e-01
2.71533400e-01 4.88408655e-01 8.96301985e-01 -3.52965444e-01
-2.35213950e-01 -4.67592329e-01 2.59014696e-01 5.53022146e-01
2.10871287e-02 -4.87662494e-01 -8.84682655e-01 1.68851718e-01
-2.00107646e+00 -5.05171716e-01 -2.49813512e-01 1.65321970e+00
8.52837026e-01 1.98022164e-02 5.27240485e-02 -2.54531384e-01
6.20717883e-01 3.50506872e-01 -1.00121808e+00 -2.98026025e-01
-2.53693879e-01 4.87380415e-01 3.60782444e-01 2.28103891e-01
-1.15411747e+00 9.29440498e-01 6.83942366e+00 4.46103126e-01
-7.10440814e-01 -2.06226110e-01 8.03237975e-01 2.44408861e-01
-4.51197267e-01 -8.10519531e-02 -5.26035070e-01 -2.83188503e-02
1.37145662e+00 1.39095917e-01 5.20811319e-01 7.83088148e-01
-2.15681776e-01 5.72551012e-01 -1.72972143e+00 5.54307103e-01
-2.57561922e-01 -1.56648135e+00 2.21498191e-01 2.99130261e-01
4.09993768e-01 2.92872399e-01 -1.03576399e-01 4.44404215e-01
9.51200128e-01 -1.59017360e+00 1.46707550e-01 5.43724477e-01
9.39039469e-01 -7.21518934e-01 5.36603153e-01 2.50298470e-01
-1.10752285e+00 -1.90931708e-01 -6.79883599e-01 2.37514693e-02
-1.89879805e-01 5.30066550e-01 -1.07205844e+00 5.78012705e-01
3.73030156e-01 1.28772354e+00 -5.91699839e-01 4.99110341e-01
-2.19078690e-01 3.99234325e-01 3.08728535e-02 -1.53486252e-01
3.59937787e-01 -2.14714393e-01 8.45949948e-02 1.24337077e+00
-8.23422149e-02 -7.88812190e-02 4.79123205e-01 8.79742801e-01
-7.00339675e-01 9.28148031e-02 -1.13936698e+00 -4.82890308e-01
-1.28847361e-01 1.14866674e+00 -6.27621412e-01 -1.49115697e-01
-5.42465806e-01 7.51536131e-01 1.16243446e+00 4.82667089e-01
-5.75667024e-01 -2.57736951e-01 7.63241351e-01 9.97871235e-02
3.44902039e-01 1.28370533e-02 -4.12804224e-02 -8.77197444e-01
-3.99535000e-01 -9.60537672e-01 6.00329161e-01 -7.00147390e-01
-1.52032137e+00 5.62340617e-01 -5.48064530e-01 -7.56341338e-01
-9.25242677e-02 -1.24533260e+00 -1.04546793e-01 7.92512119e-01
-1.42865551e+00 -1.38640022e+00 2.21374363e-01 5.74619651e-01
8.53272006e-02 -2.42035151e-01 1.28892910e+00 1.66161992e-02
-2.61748374e-01 2.83293068e-01 1.58951551e-01 1.99216858e-01
2.97007203e-01 -1.52155888e+00 6.53360009e-01 3.31532031e-01
3.79940987e-01 7.44838357e-01 3.67077649e-01 -4.94346589e-01
-1.45661509e+00 -1.34212911e+00 1.07108772e+00 -6.52840257e-01
7.97489762e-01 -7.27559507e-01 -1.08989227e+00 1.00838768e+00
-1.16192818e-01 6.24880493e-01 9.27522659e-01 5.68311334e-01
-7.82083273e-01 1.38702869e-01 -9.30520296e-01 3.80876362e-01
1.68902409e+00 -8.71978760e-01 -3.40971977e-01 8.82115781e-01
1.18685687e+00 -2.26974383e-01 -1.31254852e+00 5.49456835e-01
3.56941879e-01 -6.52606547e-01 1.13644826e+00 -1.54678583e+00
6.69974566e-01 -3.53062861e-02 -1.69299647e-01 -1.27378631e+00
-7.35825360e-01 -3.57987285e-01 -2.79828697e-01 5.81243277e-01
7.30198503e-01 -4.70043302e-01 9.65972960e-01 1.22061484e-02
-3.46571416e-01 -1.23054934e+00 -5.51607728e-01 -3.16195518e-01
-2.83261426e-02 -1.16380125e-01 4.40494806e-01 9.43634033e-01
1.80652693e-01 8.46283913e-01 -3.86681765e-01 -2.14556921e-02
2.78952241e-01 2.23360479e-01 3.05910617e-01 -1.78608239e+00
-5.00159621e-01 -1.61513269e-01 -4.76746529e-01 -1.22341383e+00
6.34193599e-01 -1.67664909e+00 -1.17983371e-01 -1.51125705e+00
2.39463583e-01 -3.37784767e-01 -7.50753045e-01 8.73987913e-01
-5.35470136e-02 2.01731429e-01 -4.93209399e-02 1.17105767e-01
-7.62045801e-01 3.91345024e-01 1.36991000e+00 -3.07857007e-01
1.14234351e-01 3.51453573e-03 -8.70369494e-01 3.74966443e-01
4.34132069e-01 -5.60102701e-01 -6.24535918e-01 -2.67115593e-01
7.31350124e-01 1.65914789e-01 1.15228131e-01 -6.42877519e-01
8.91941711e-02 4.25180756e-02 4.58138764e-01 -1.50529340e-01
1.64401442e-01 -8.07226241e-01 8.06891695e-02 5.34426272e-01
-9.13835764e-01 2.40741730e-01 -1.66526154e-01 9.58719969e-01
-8.19897130e-02 -4.74867932e-02 5.29704094e-01 -3.95852089e-01
-5.08834422e-01 5.89740038e-01 -1.48322254e-01 -6.72905222e-02
7.07468092e-01 -6.71968833e-02 -5.12408197e-01 -1.52055994e-01
-1.17153418e+00 3.26316148e-01 1.03225373e-01 2.35516638e-01
6.01366282e-01 -1.12572813e+00 -4.53986734e-01 1.84926704e-01
3.59096318e-01 9.46962088e-02 -1.69892341e-01 4.57454383e-01
-4.67349648e-01 5.91177583e-01 -2.58044630e-01 -4.38319802e-01
-9.84718859e-01 9.75677669e-01 3.72177213e-01 -7.75868356e-01
-7.48826087e-01 9.99643385e-01 5.75096190e-01 -9.34695482e-01
2.48957023e-01 -4.02174085e-01 -5.47235131e-01 -1.42192975e-01
1.82274468e-02 -1.30783945e-01 2.00942338e-01 -3.78951341e-01
-2.23963320e-01 7.83400685e-02 -1.08242370e-01 7.04792023e-01
1.73646939e+00 4.14167494e-01 -6.39528632e-01 2.56180435e-01
1.43153310e+00 -6.32108510e-01 -1.07840526e+00 -5.73653638e-01
6.61303103e-01 2.19129369e-01 -4.35020328e-01 -7.56073177e-01
-9.48851347e-01 7.18298137e-01 4.65437621e-02 3.46005112e-01
8.00811231e-01 3.25297028e-01 5.21821380e-01 9.78593290e-01
2.24492326e-01 -4.83804196e-01 3.82772833e-01 7.84482419e-01
7.12451220e-01 -1.07973051e+00 -1.35292858e-01 -4.49384809e-01
-2.95609504e-01 1.38924551e+00 5.89093626e-01 -2.02688351e-01
6.79383695e-01 9.56877246e-02 -2.64281154e-01 -7.64306664e-01
-1.06497777e+00 -3.40210468e-01 2.63489962e-01 9.54487562e-01
7.53068566e-01 -7.32015520e-02 -3.87333194e-03 3.07872415e-01
1.30827799e-01 -1.91602007e-01 1.30897328e-01 5.42031527e-01
-4.30716127e-01 -1.24981070e+00 2.53026426e-01 8.22394371e-01
-2.75516301e-01 -3.50266457e-01 -4.85058576e-01 4.19486016e-01
1.36576183e-02 7.57729888e-01 -1.15136497e-01 -5.48083842e-01
2.90591508e-01 3.49475652e-01 3.26704562e-01 -1.03908968e+00
-8.34849596e-01 -3.85646075e-01 2.44560331e-01 -6.85979247e-01
-3.72831851e-01 -1.03749759e-01 -1.35925519e+00 -2.26684958e-01
-1.25258982e-01 8.60821754e-02 1.66738912e-01 9.10778642e-01
4.53344613e-01 8.67312849e-01 4.80482459e-01 -3.48200113e-01
-4.70364392e-01 -9.54816461e-01 -7.29237497e-01 7.92800188e-01
3.56355757e-01 -5.16276240e-01 1.33711338e-01 6.56506838e-03] | [6.8167877197265625, 6.280070781707764] |
e9328f16-78df-444c-8b9c-392e276e8f01 | mastering-atari-go-chess-and-shogi-by | 1911.08265 | null | https://arxiv.org/abs/1911.08265v2 | https://arxiv.org/pdf/1911.08265v2.pdf | Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model | Constructing agents with planning capabilities has long been one of the main challenges in the pursuit of artificial intelligence. Tree-based planning methods have enjoyed huge success in challenging domains, such as chess and Go, where a perfect simulator is available. However, in real-world problems the dynamics governing the environment are often complex and unknown. In this work we present the MuZero algorithm which, by combining a tree-based search with a learned model, achieves superhuman performance in a range of challenging and visually complex domains, without any knowledge of their underlying dynamics. MuZero learns a model that, when applied iteratively, predicts the quantities most directly relevant to planning: the reward, the action-selection policy, and the value function. When evaluated on 57 different Atari games - the canonical video game environment for testing AI techniques, in which model-based planning approaches have historically struggled - our new algorithm achieved a new state of the art. When evaluated on Go, chess and shogi, without any knowledge of the game rules, MuZero matched the superhuman performance of the AlphaZero algorithm that was supplied with the game rules. | ['Laurent SIfre', 'Edward Lockhart', 'Thore Graepel', 'Karen Simonyan', 'Ioannis Antonoglou', 'David Silver', 'Timothy Lillicrap', 'Thomas Hubert', 'Simon Schmitt', 'Demis Hassabis', 'Julian Schrittwieser', 'Arthur Guez'] | 2019-11-19 | null | null | null | null | ['game-of-go', 'game-of-shogi', 'game-of-chess'] | ['playing-games', 'playing-games', 'playing-games'] | [-3.26898247e-02 2.48847842e-01 2.03029085e-02 2.07366183e-01
-3.30555499e-01 -6.92248881e-01 7.29035378e-01 4.03922871e-02
-5.47093689e-01 1.12850308e+00 -1.28893673e-01 -3.50825220e-01
-4.32071239e-01 -7.49454319e-01 -4.41531867e-01 -5.96161485e-01
-6.10427737e-01 1.04387224e+00 6.77021205e-01 -8.82531166e-01
4.68544036e-01 1.69257224e-01 -1.47653973e+00 -1.57921985e-01
6.56656384e-01 6.83562279e-01 3.85339826e-01 9.22436476e-01
4.58509415e-01 1.08700740e+00 -3.98275882e-01 3.23495120e-02
4.79725480e-01 -5.99956155e-01 -1.04741979e+00 -1.46637276e-01
-2.26529747e-01 -1.75622523e-01 -3.54727685e-01 7.48134434e-01
3.03975523e-01 4.77797925e-01 4.46973503e-01 -1.24517608e+00
1.32852152e-01 5.03333688e-01 -3.06660123e-02 2.26339072e-01
5.69296062e-01 6.52016878e-01 8.89889300e-01 3.75732295e-02
9.47599173e-01 1.11347604e+00 3.86866391e-01 6.21485114e-01
-1.10291672e+00 -2.84475327e-01 1.18729636e-01 3.35678726e-01
-1.33640635e+00 -1.61529899e-01 3.06330711e-01 -4.21222806e-01
1.13674212e+00 -6.30180761e-02 1.01709616e+00 8.77650201e-01
6.28164351e-01 7.06367731e-01 1.11103606e+00 -3.90743285e-01
6.89677060e-01 -2.75899708e-01 -4.99740541e-01 7.90701628e-01
-1.71501320e-02 8.31071436e-01 -4.24851000e-01 -1.35890141e-01
1.14095235e+00 -6.34051025e-01 -1.24906391e-01 -7.64362395e-01
-1.31510603e+00 7.67627120e-01 3.47556800e-01 2.51771599e-01
-4.71119404e-01 3.30482870e-01 1.28758535e-01 4.35456514e-01
-1.74298465e-01 1.00432515e+00 -4.80473936e-01 -5.54283798e-01
-5.73405981e-01 8.93881440e-01 1.05399680e+00 7.22466707e-01
3.71435344e-01 1.98261425e-01 2.24465892e-01 1.11470476e-01
1.10666603e-01 2.06766129e-01 4.24995542e-01 -1.25290835e+00
4.22518611e-01 5.90375066e-01 6.08139038e-01 -9.20394838e-01
-8.10597241e-01 -3.37354720e-01 -1.63977772e-01 9.49658692e-01
7.08437800e-01 -3.29061419e-01 -7.89624512e-01 1.61149514e+00
5.27853847e-01 1.76348239e-01 3.50864083e-01 1.01391220e+00
2.07055911e-01 6.74614191e-01 -5.87712089e-03 -1.25934049e-01
9.43309724e-01 -9.95272100e-01 -1.26338914e-01 -5.47637165e-01
7.01130271e-01 -1.71834677e-01 1.03231096e+00 5.50542176e-01
-1.15494740e+00 -1.90390691e-01 -1.07730675e+00 4.84355271e-01
-2.59339307e-02 -7.14280903e-01 8.28538179e-01 3.62432569e-01
-1.13250291e+00 8.21086526e-01 -1.09382224e+00 -4.38392639e-01
3.62961516e-02 6.44990623e-01 -3.85227382e-01 -3.64649259e-02
-8.99046063e-01 1.41145372e+00 6.50753140e-01 -2.36497477e-01
-1.50379193e+00 -7.11535141e-02 -6.82514250e-01 1.67855680e-01
1.03986824e+00 -6.17123067e-01 1.75623870e+00 -7.12468505e-01
-1.78790963e+00 4.15815592e-01 4.05236691e-01 -6.12292230e-01
5.82413137e-01 1.31181180e-01 8.01305026e-02 -1.37673825e-01
1.45831540e-01 5.51307440e-01 4.13748890e-01 -1.10740852e+00
-7.91419089e-01 -2.01048270e-01 5.97703338e-01 6.76720560e-01
5.60028672e-01 -1.92220017e-01 -5.81897348e-02 -1.63141251e-01
-3.64779122e-02 -1.12094855e+00 -7.90954709e-01 -4.39529300e-01
3.71139101e-03 -1.87096283e-01 2.40986377e-01 -1.29536316e-01
9.67597306e-01 -1.77693212e+00 6.35303259e-01 2.37635434e-01
7.53339827e-02 4.22250330e-02 -8.61355290e-02 6.84494078e-01
3.28526855e-01 -1.65313855e-01 -1.78938434e-01 3.63604695e-01
1.65689498e-01 4.42927659e-01 9.77003053e-02 2.53634155e-01
-1.33869871e-01 8.47359896e-01 -1.25163782e+00 -4.17813003e-01
5.28757796e-02 -7.97000378e-02 -7.03927577e-01 1.32695049e-01
-7.29381800e-01 4.98142958e-01 -6.99279249e-01 4.20440644e-01
-1.88641369e-01 -7.78652355e-02 4.39850479e-01 7.36844242e-01
-3.28743607e-01 5.03666848e-02 -1.14902592e+00 1.57841814e+00
-1.92188770e-01 4.61158574e-01 1.00150138e-01 -6.34879291e-01
6.76133156e-01 2.77138501e-01 4.11663562e-01 -9.04782832e-01
3.55989486e-01 3.15578431e-02 4.90902126e-01 -5.24464071e-01
4.69603479e-01 -3.92930955e-01 -1.78080931e-01 4.00314629e-01
-2.10191980e-01 -7.29104221e-01 4.33489293e-01 -6.94308877e-02
1.61546218e+00 4.75099117e-01 6.04452491e-01 -3.02942902e-01
2.42121473e-01 8.87086570e-01 4.29107606e-01 8.98997366e-01
-2.58925647e-01 1.90260425e-01 6.76515818e-01 -8.48056376e-01
-9.28820312e-01 -7.94190347e-01 3.86218846e-01 1.00693691e+00
3.46784264e-01 -4.76758122e-01 -8.03383946e-01 -2.95607179e-01
-3.32132101e-01 9.87601340e-01 -7.81583965e-01 -5.17527796e-02
-7.62190998e-01 -3.71259451e-01 2.85176575e-01 2.58757532e-01
2.85442829e-01 -1.46611834e+00 -1.43902075e+00 5.98143756e-01
8.60554352e-03 -8.41586351e-01 -1.65942714e-01 2.27972388e-01
-7.51812875e-01 -1.27287316e+00 -1.70828730e-01 -6.53256357e-01
3.37657660e-01 -1.00866802e-01 1.15027058e+00 1.90394282e-01
-2.14936629e-01 3.83980393e-01 -2.91381359e-01 -4.86144990e-01
-4.54921395e-01 7.40530938e-02 -6.29489794e-02 -7.19833910e-01
-2.67149806e-01 -6.33988023e-01 -1.84599161e-01 4.84414876e-01
-6.11735821e-01 2.52001941e-01 3.66549075e-01 9.20177221e-01
3.24245065e-01 4.07414317e-01 1.05703749e-01 -5.51421046e-01
7.16583908e-01 -3.04843694e-01 -1.07794213e+00 7.67050534e-02
-3.61643910e-01 2.32281268e-01 5.35040617e-01 -4.36033547e-01
-7.96947837e-01 2.24364107e-03 2.84375638e-01 2.89185718e-02
-1.52539939e-01 6.60822868e-01 -1.57765225e-02 -1.59882694e-01
8.71193051e-01 6.58753142e-02 -5.75195663e-02 -1.48864076e-01
6.02119789e-02 -1.06070034e-01 5.16572297e-01 -8.80759060e-01
5.97350538e-01 -8.36446695e-03 2.69432127e-01 -6.09741569e-01
-4.10128891e-01 -3.03759295e-02 -4.37141627e-01 -3.15267265e-01
8.05908799e-01 -4.51755434e-01 -1.05530357e+00 4.50348973e-01
-8.69096637e-01 -1.08674598e+00 -4.55635130e-01 4.44447488e-01
-1.19115770e+00 -1.64817572e-01 -3.06162417e-01 -9.10451949e-01
2.67930508e-01 -1.27949095e+00 5.32633185e-01 3.32323015e-01
-2.61843026e-01 -8.35548043e-01 4.96654570e-01 -1.82409212e-02
4.50267196e-01 4.32615906e-01 1.07346761e+00 -5.32682300e-01
-6.71346247e-01 1.86420605e-02 2.56745636e-01 -4.95740116e-01
-2.20154196e-01 -2.41844952e-01 -2.77335435e-01 -2.97765225e-01
-1.34278297e-01 -4.05172318e-01 -6.00455189e-03 3.67317677e-01
4.17778492e-01 -2.04938188e-01 -3.24318111e-01 1.77817032e-01
1.30863059e+00 7.92718172e-01 5.52030146e-01 9.04101014e-01
1.00821689e-01 5.26139855e-01 7.18561530e-01 5.57759404e-01
6.77227974e-01 1.00393605e+00 7.93703854e-01 4.04723883e-01
3.17159563e-01 -3.86867642e-01 2.32184559e-01 2.17259586e-01
-6.26684010e-01 -2.73741245e-01 -1.35599637e+00 4.56950486e-01
-2.11188245e+00 -1.12424278e+00 3.75400424e-01 2.12362957e+00
6.52310014e-01 4.55717981e-01 5.01161098e-01 1.55983064e-02
2.60112345e-01 -8.22397843e-02 -8.21060777e-01 -4.15812254e-01
2.77023256e-01 2.39137277e-01 3.87399912e-01 8.06686163e-01
-9.20245051e-01 1.31640887e+00 7.04045153e+00 4.41916764e-01
-7.18406677e-01 -2.03111947e-01 2.02553198e-01 -7.51466602e-02
1.89851269e-01 1.93775699e-01 -5.00268221e-01 1.16626278e-01
9.59715307e-01 -5.07874310e-01 1.01155651e+00 9.94914114e-01
3.70675117e-01 -6.97191179e-01 -1.08274901e+00 5.04897654e-01
-2.22792238e-01 -1.23764706e+00 -5.39922357e-01 2.60804206e-01
5.55950344e-01 -5.25111444e-02 -1.23998761e-01 6.38179421e-01
1.04276383e+00 -1.43004477e+00 9.78598773e-01 3.28263611e-01
2.18868822e-01 -8.14734340e-01 7.03459561e-01 8.34690809e-01
-1.03391623e+00 -2.63921112e-01 -1.88343823e-01 -6.57027960e-01
1.89592317e-01 -4.86328185e-01 -1.08783376e+00 3.10061663e-01
4.88450974e-01 2.67002493e-01 -2.44677991e-01 1.34369099e+00
-4.10431385e-01 3.57173294e-01 -4.57826823e-01 -4.08188581e-01
6.84838951e-01 -1.74434781e-01 6.67578638e-01 5.57846606e-01
5.81104793e-02 7.92349577e-01 6.22898817e-01 4.83444780e-01
6.08873606e-01 -1.35056123e-01 -7.34290957e-01 -8.90017524e-02
2.45819256e-01 6.46039069e-01 -1.00630796e+00 5.72383928e-04
1.83341146e-01 3.80037218e-01 3.92152548e-01 1.01684138e-01
-8.67010295e-01 -2.06188153e-04 7.22964108e-01 1.54275194e-01
2.63478220e-01 -6.28779233e-01 -9.97627825e-02 -8.93917263e-01
-3.80637527e-01 -1.11259007e+00 2.65065610e-01 -8.34672809e-01
-4.67362106e-01 7.79096782e-01 5.00963569e-01 -9.96860504e-01
-8.02413702e-01 -4.50674295e-01 -5.62690616e-01 5.73550522e-01
-9.51189399e-01 -7.26976812e-01 -9.87472832e-02 5.39209247e-01
6.94656551e-01 -2.87616491e-01 8.68045211e-01 -3.37826699e-01
-3.88864726e-01 -2.76830234e-02 -1.40313283e-01 -3.21132988e-01
-2.96824519e-02 -1.08105826e+00 5.66461086e-01 5.99481523e-01
-1.24016412e-01 6.82574362e-02 1.09241188e+00 -6.16692722e-01
-1.44013572e+00 -3.60938549e-01 3.62282693e-01 -3.75766635e-01
8.44182312e-01 1.67652220e-02 -5.50716460e-01 7.46966779e-01
6.42263591e-02 -3.67778689e-01 5.40399216e-02 -9.65865925e-02
2.50688225e-01 3.60727757e-01 -1.00913322e+00 7.95353293e-01
1.19037950e+00 7.72856846e-02 -6.80765629e-01 2.66027242e-01
4.89527047e-01 -9.45147336e-01 -4.14997280e-01 7.87094459e-02
6.33842647e-01 -1.07175624e+00 7.39644825e-01 -1.02131879e+00
3.18925679e-01 -4.38658208e-01 -2.21821845e-01 -1.54571748e+00
-4.63269532e-01 -7.71568418e-01 6.10218309e-02 4.15274680e-01
3.76865149e-01 -4.56063747e-01 9.98729587e-01 8.56173933e-01
-1.45522561e-02 -7.54725695e-01 -9.38871980e-01 -8.94867718e-01
1.53822480e-02 -3.86655569e-01 6.58478200e-01 4.46596682e-01
2.98624665e-01 2.87062138e-01 -4.03271466e-01 1.45330921e-01
5.05635798e-01 -3.22595611e-02 9.25530434e-01 -1.12185383e+00
-6.51929975e-01 -6.04034781e-01 -5.97800672e-01 -7.67734051e-01
1.11816535e-02 -4.36249584e-01 3.44730377e-01 -1.59939897e+00
-2.37673640e-01 -5.59272110e-01 1.79998711e-01 6.45608246e-01
2.93820143e-01 -3.95368695e-01 5.64530849e-01 8.68489519e-02
-6.91578984e-01 4.29895937e-01 1.50726271e+00 -6.84284745e-03
-5.20129442e-01 4.75994237e-02 -4.45834458e-01 8.79229307e-01
8.88433397e-01 -5.15500784e-01 -7.30086565e-01 -4.77129608e-01
3.42728406e-01 6.59014225e-01 2.07370237e-01 -1.44015479e+00
5.08051217e-01 -9.01503623e-01 -1.09300531e-01 -3.77710201e-02
5.11122227e-01 -8.32596838e-01 4.61532116e-01 8.32263887e-01
-2.68086880e-01 4.76547509e-01 2.85026193e-01 5.07151902e-01
-4.70039956e-02 -3.15472752e-01 5.29983342e-01 -5.40823102e-01
-1.11209786e+00 8.23064744e-02 -7.19921470e-01 3.33170146e-01
1.48071659e+00 -2.57162064e-01 -1.10717274e-01 -6.29775584e-01
-9.00397718e-01 5.35503149e-01 7.08759129e-01 1.32874250e-01
5.06730080e-01 -7.77401984e-01 -6.56343579e-01 7.47950301e-02
-2.24616334e-01 1.28626883e-01 -1.07013891e-02 4.98518169e-01
-8.07246447e-01 2.32213184e-01 -7.37820685e-01 -3.04790497e-01
-1.04365575e+00 4.03566808e-01 5.61436117e-01 -6.16658211e-01
-7.06907451e-01 6.25327945e-01 1.73944578e-01 -3.35337132e-01
4.59903218e-02 -2.39915118e-01 -3.02807450e-01 -4.31280345e-01
5.27236104e-01 2.43297145e-01 -2.46341646e-01 -4.46301699e-01
-3.55307907e-01 3.27411562e-01 2.73955375e-01 -5.38783848e-01
1.64468241e+00 1.87888592e-01 2.76939064e-01 2.42423341e-01
-4.64921677e-03 -4.47701514e-01 -1.50722969e+00 1.34447262e-01
2.15043172e-01 -4.76828456e-01 2.23904382e-02 -1.07583106e+00
-6.06546819e-01 5.17249167e-01 1.13557898e-01 2.36060053e-01
7.95415044e-01 -1.95135415e-01 3.55675399e-01 7.14301407e-01
1.31962848e+00 -8.09536338e-01 4.03851867e-02 1.02260935e+00
7.76837707e-01 -1.02301478e+00 -2.28481181e-02 5.77038974e-02
-1.00332487e+00 1.04529810e+00 8.64847839e-01 -2.41195008e-01
1.90936103e-01 3.32598507e-01 -1.44073322e-01 -2.21062645e-01
-1.18455958e+00 -2.99661130e-01 -1.60758182e-01 9.15973663e-01
-2.07520977e-01 1.26607925e-01 2.09180228e-02 4.38990206e-01
-6.14872158e-01 7.56750628e-02 7.88647592e-01 1.27388346e+00
-7.47662067e-01 -1.00138044e+00 -4.71229732e-01 5.25836274e-02
2.34199479e-01 2.42984205e-01 -3.55324298e-01 1.33556116e+00
4.92014140e-02 8.40072513e-01 -2.67064244e-01 -3.80201161e-01
5.08459151e-01 -1.01332545e-01 8.36255670e-01 -6.54837847e-01
-6.82368636e-01 -2.71195441e-01 3.95330638e-01 -8.64038348e-01
-1.90081626e-01 -7.52231061e-01 -1.41830468e+00 -5.79378963e-01
6.14875415e-03 4.02319431e-01 5.10912716e-01 1.12193501e+00
-6.51133880e-02 2.86294609e-01 1.47687137e-01 -1.36799610e+00
-6.47796869e-01 -4.15133655e-01 -5.82176924e-01 -1.35659799e-01
9.98189822e-02 -9.70620811e-01 8.76245499e-02 -1.94022223e-01] | [3.907331943511963, 1.4090818166732788] |
a8279d6e-bd05-4d39-bbbc-e6cd01cecaf5 | sentence-to-label-generation-framework-for | 2306.15978 | null | https://arxiv.org/abs/2306.15978v1 | https://arxiv.org/pdf/2306.15978v1.pdf | Sentence-to-Label Generation Framework for Multi-task Learning of Japanese Sentence Classification and Named Entity Recognition | Information extraction(IE) is a crucial subfield within natural language processing. In this study, we introduce a Sentence Classification and Named Entity Recognition Multi-task (SCNM) approach that combines Sentence Classification (SC) and Named Entity Recognition (NER). We develop a Sentence-to-Label Generation (SLG) framework for SCNM and construct a Wikipedia dataset containing both SC and NER. Using a format converter, we unify input formats and employ a generative model to generate SC-labels, NER-labels, and associated text segments. We propose a Constraint Mechanism (CM) to improve generated format accuracy. Our results show SC accuracy increased by 1.13 points and NER by 1.06 points in SCNM compared to standalone tasks, with CM raising format accuracy from 63.61 to 100. The findings indicate mutual reinforcement effects between SC and NER, and integration enhances both tasks' performance. | ['Tatsunori Mori', 'Qinghao Zhang', 'Chengguang Gan'] | 2023-06-28 | null | null | null | null | ['multi-task-learning', 'sentence-classification', 'named-entity-recognition-ner', 'cg'] | ['methodology', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 5.18340707e-01 3.77327353e-01 3.98371480e-02 -6.07765973e-01
-1.28545511e+00 -7.40674615e-01 6.85518503e-01 3.47952634e-01
-6.62657201e-01 1.06062663e+00 4.68605220e-01 -2.27003798e-01
1.41189590e-01 -7.77767658e-01 -5.50029874e-01 1.73888162e-01
4.83381689e-01 2.81702310e-01 -1.11058675e-01 -2.10526884e-01
3.85509729e-01 3.90879922e-02 -1.24772108e+00 9.18337226e-01
1.38602173e+00 6.90242887e-01 1.20922111e-01 8.20708752e-01
-7.80401528e-01 9.38212574e-01 -1.05210495e+00 -8.02486241e-01
-4.44528647e-02 -3.79434049e-01 -1.13842094e+00 -1.00369185e-01
2.19244286e-01 3.79513264e-01 2.18401551e-01 8.74720991e-01
5.53144097e-01 1.62598923e-01 9.01551127e-01 -1.21724331e+00
-9.17424500e-01 1.01842546e+00 -2.10940212e-01 -2.81639189e-01
7.48865187e-01 -2.67798990e-01 8.74077559e-01 -9.93811965e-01
1.09672570e+00 9.27411675e-01 8.44093263e-01 8.55209589e-01
-1.14937115e+00 -5.03302336e-01 -2.09762514e-01 -3.60784858e-01
-1.30482197e+00 -4.90999073e-01 3.62978578e-01 -4.66173977e-01
1.41202748e+00 4.12255496e-01 1.04337946e-01 1.15227175e+00
4.95411932e-01 7.27590024e-01 1.24277329e+00 -8.94903779e-01
1.57091454e-01 3.85830641e-01 3.41017008e-01 2.35343143e-01
4.50820208e-01 -5.08407950e-01 -6.19675338e-01 8.99108648e-02
2.47093946e-01 -3.08349431e-01 1.29075885e-01 5.72418213e-01
-1.12164676e+00 7.36476719e-01 2.43400149e-02 4.65638220e-01
-4.55324888e-01 -3.48692507e-01 6.04152322e-01 2.81456083e-01
6.57039344e-01 1.11809182e+00 -5.42977095e-01 -2.80467480e-01
-1.15636444e+00 8.45862702e-02 1.19616854e+00 1.54819608e+00
4.14695531e-01 -1.80888101e-01 -7.61492431e-01 9.86168385e-01
1.61851734e-01 5.25273323e-01 4.37670946e-01 -6.52662992e-01
9.82168496e-01 8.69294047e-01 7.25401267e-02 -9.10092711e-01
-6.78092420e-01 -5.38157165e-01 -8.75653684e-01 -3.65595847e-01
2.49396656e-02 -5.46131849e-01 -7.00170219e-01 1.63932157e+00
-4.50332202e-02 -5.65338790e-01 3.96470279e-01 2.53224730e-01
1.07642901e+00 7.16988623e-01 6.22261345e-01 -4.99715842e-02
1.55345249e+00 -9.05419827e-01 -8.63797903e-01 -3.87187690e-01
1.18748379e+00 -7.75124729e-01 8.56868207e-01 1.51534304e-01
-8.99352670e-01 -6.29051149e-01 -7.74243653e-01 -1.87760070e-01
-6.82197213e-01 5.51676452e-01 3.18773270e-01 6.36297584e-01
-1.05774379e+00 4.27158624e-01 -3.56382370e-01 -4.04486805e-01
1.77851364e-01 6.71412349e-02 -4.25203264e-01 1.79440647e-01
-1.26720166e+00 1.06226063e+00 7.03215659e-01 -2.96710253e-01
-1.37224048e-01 -6.22459769e-01 -1.09527266e+00 -9.40983668e-02
1.08055890e-01 -8.56591165e-01 1.37019372e+00 -5.68733454e-01
-1.18137395e+00 8.99664223e-01 -2.64390528e-01 -3.29654604e-01
1.79392576e-01 -3.41400027e-01 -7.76114702e-01 5.15428558e-02
5.64641893e-01 8.85696411e-01 2.14978978e-01 -1.19625449e+00
-7.33124256e-01 -3.76770236e-02 -1.94914430e-01 3.96798551e-01
-3.08261603e-01 6.11747563e-01 8.01684931e-02 -6.53735995e-01
-2.35645816e-01 -5.69238067e-01 -1.83385864e-01 -7.35499740e-01
-7.78167903e-01 -4.77160096e-01 5.71910106e-02 -1.07347167e+00
1.80586267e+00 -1.84402788e+00 -1.19909488e-01 -1.56181715e-02
1.84041262e-01 2.20090374e-01 -3.09304744e-01 6.93387210e-01
8.30516145e-02 7.38794267e-01 -3.97858709e-01 -6.45422637e-01
1.82869256e-01 -2.44104713e-02 -4.42412794e-02 -4.06264246e-01
5.95599234e-01 1.26331675e+00 -7.80657232e-01 -7.38530576e-01
-2.86258042e-01 1.70604065e-01 -2.70502299e-01 2.28514358e-01
-1.63866952e-01 1.78156823e-01 -1.57401070e-01 5.13807058e-01
5.60137451e-01 -2.14752004e-01 2.07533389e-02 2.74201557e-02
-4.39594924e-01 6.23586118e-01 -1.29871988e+00 1.61254382e+00
-7.04614699e-01 4.44210440e-01 -2.12788925e-01 -2.56372929e-01
1.22462749e+00 4.33116853e-01 4.49218899e-02 -5.70397854e-01
5.72507456e-02 2.44202763e-01 -5.14185667e-01 -6.49974525e-01
1.11202621e+00 -6.51098639e-02 -7.93096006e-01 4.23250169e-01
4.52701002e-01 2.23767478e-02 5.92458963e-01 4.04709101e-01
1.26440465e+00 4.62895893e-02 6.21212304e-01 -7.30372593e-02
4.05680001e-01 3.64563614e-01 6.53049111e-01 8.55184317e-01
2.38510221e-01 4.20088291e-01 4.73682582e-01 4.44860831e-02
-1.21342099e+00 -7.04001188e-01 -1.40543178e-01 1.08056867e+00
-3.38525116e-01 -7.18525767e-01 -1.01746035e+00 -8.32169890e-01
-2.50582844e-01 1.30360293e+00 -2.36924112e-01 -7.57797435e-02
-3.85304272e-01 -6.50107861e-01 1.02925777e+00 6.37042284e-01
4.91054833e-01 -1.51977944e+00 -3.63061130e-01 4.09826398e-01
-4.98241991e-01 -1.20589340e+00 -5.26740074e-01 2.31240988e-01
-4.69808787e-01 -6.29047751e-01 -4.08391029e-01 -9.30075049e-01
7.35463977e-01 -1.65575385e-01 1.30391824e+00 -2.93172449e-01
-6.66218325e-02 1.62283584e-01 -7.25241601e-01 -4.05914754e-01
-1.04759634e+00 6.40547514e-01 -3.25051658e-02 -2.75094330e-01
4.72485721e-01 -4.85581607e-02 9.36024562e-02 -8.13674182e-02
-1.12785530e+00 5.54558039e-01 8.31619501e-01 6.75637960e-01
3.22948962e-01 -2.54340589e-01 8.99368286e-01 -1.40981507e+00
1.04245520e+00 -5.67813873e-01 -1.03823759e-01 7.18308747e-01
-9.41883802e-01 1.74801797e-01 6.98786974e-01 -1.47994384e-01
-1.38922405e+00 1.94166467e-01 -3.75045598e-01 4.25005227e-01
-6.96786225e-01 7.53490925e-01 -3.10710609e-01 2.99900413e-01
7.09598422e-01 2.84537882e-01 -3.55136067e-01 -5.60150921e-01
4.54742670e-01 1.32287979e+00 5.59112430e-01 -6.15753412e-01
4.64963019e-01 -3.90750587e-01 -3.75301540e-01 -7.10016191e-01
-1.09796405e+00 -4.74759877e-01 -8.52169871e-01 -3.01300019e-01
1.00039184e+00 -9.26494241e-01 -2.56232381e-01 2.81832010e-01
-1.57578099e+00 -7.31358901e-02 -3.36950332e-01 2.56207436e-01
-2.12300614e-01 2.60957152e-01 -8.75383198e-01 -7.74408400e-01
-8.78392816e-01 -6.56103909e-01 1.20020211e+00 4.37456518e-01
-6.30602837e-01 -8.19749534e-01 4.29570675e-02 5.67239344e-01
4.70195740e-01 2.35602245e-01 6.89066172e-01 -1.15776300e+00
1.03436736e-02 -1.06362827e-01 -3.09285522e-01 1.60227105e-01
-1.81155488e-01 -3.95318717e-02 -6.67394996e-01 5.75603023e-02
-1.60867676e-01 -1.53586283e-01 5.05917132e-01 -1.67726763e-02
5.46129584e-01 -5.49421191e-01 -2.66028970e-01 3.31280798e-01
1.50750983e+00 3.12411338e-01 6.86026573e-01 5.50528347e-01
6.53383493e-01 7.43194759e-01 4.69562978e-01 4.15661752e-01
8.18334937e-01 3.85641307e-01 -3.38953704e-01 -1.14797130e-02
-2.69919813e-01 -5.51441073e-01 4.33366030e-01 1.26130843e+00
4.42308486e-01 -5.06189466e-01 -1.06107187e+00 3.62185001e-01
-1.47734916e+00 -8.23370099e-01 -6.15443468e-01 1.63477731e+00
1.31430602e+00 1.42075464e-01 -2.66962409e-01 -6.56348020e-02
9.03815925e-01 -3.55798930e-01 -1.19648874e-01 -6.20950162e-01
-3.23426157e-01 2.89988756e-01 4.60570693e-01 3.54791313e-01
-1.24993241e+00 9.67090905e-01 6.10194397e+00 7.67152309e-01
-5.35926640e-01 7.39893019e-02 4.15557534e-01 3.94788384e-01
-4.27697182e-01 -2.05640420e-02 -1.40779567e+00 5.44485092e-01
1.50561094e+00 -4.03088182e-01 2.01328285e-02 8.47705126e-01
-5.20243421e-02 3.29214670e-02 -8.33228350e-01 8.13938022e-01
4.25272703e-01 -1.32658482e+00 3.61186415e-01 -1.27277896e-01
7.60329783e-01 -1.81876972e-01 -5.93622267e-01 6.43369973e-01
3.54386538e-01 -8.39502215e-01 8.86399806e-01 9.59966779e-01
1.09508729e+00 -5.37521958e-01 1.08147752e+00 4.94245350e-01
-1.20568645e+00 5.83725795e-02 -1.13674767e-01 3.14598307e-02
4.84276474e-01 6.81984901e-01 -9.49596405e-01 9.84264076e-01
4.21959519e-01 3.78715187e-01 -1.00079262e+00 9.68652606e-01
-4.35820103e-01 4.21372831e-01 -6.96992874e-02 -3.40211689e-01
-1.50514394e-01 -6.01598024e-02 4.17215228e-01 1.96172667e+00
2.22933516e-01 2.62738839e-02 1.66060612e-01 9.66428816e-01
-4.66139793e-01 4.91724908e-01 -6.28472388e-01 -1.52853414e-01
9.55792129e-01 1.62439978e+00 -8.06196034e-01 -5.45710981e-01
-3.65295172e-01 1.10493875e+00 3.98033679e-01 5.34215942e-02
-5.85798502e-01 -1.00867736e+00 4.01290953e-02 -3.57988566e-01
-2.39950474e-02 -2.13161305e-01 -9.17437613e-01 -1.36055231e+00
1.45257726e-01 -7.70153582e-01 3.70225221e-01 -7.75556386e-01
-1.47663367e+00 9.25540447e-01 -1.14942074e-01 -1.19002140e+00
-3.96054119e-01 -2.86963552e-01 -4.23538208e-01 7.63064504e-01
-1.22419047e+00 -1.27255952e+00 -1.71470836e-01 -3.91759053e-02
6.82002664e-01 -1.79062020e-02 1.09680545e+00 4.75586534e-01
-7.73850262e-01 8.57576489e-01 5.62814809e-02 6.38643920e-01
7.05467820e-01 -1.48599291e+00 6.06506765e-01 1.08369410e+00
1.73754871e-01 6.86311960e-01 1.86418131e-01 -1.07185400e+00
-8.46742034e-01 -1.52688682e+00 1.94626069e+00 -6.64357722e-01
3.59348476e-01 -6.51333153e-01 -7.24453449e-01 3.69278699e-01
3.97313356e-01 -7.51775384e-01 1.06490147e+00 -6.32826611e-02
-2.90656447e-01 3.17890793e-01 -1.23719406e+00 4.04354900e-01
9.46940660e-01 -5.74863315e-01 -9.36665297e-01 1.40422285e-01
1.08171427e+00 -3.69732291e-01 -1.26153111e+00 1.45674571e-01
5.61848730e-02 -3.13738018e-01 3.62332731e-01 -6.70809031e-01
6.60285711e-01 -3.29895824e-01 -1.05286725e-01 -1.32194293e+00
-4.01246071e-01 -4.15904522e-01 8.27369764e-02 2.12585449e+00
1.16283965e+00 -3.88494015e-01 1.65190652e-01 9.76202071e-01
-5.17448545e-01 -4.68959779e-01 -7.43646979e-01 -8.30630779e-01
2.06121453e-03 -5.74602962e-01 6.05208457e-01 9.96457517e-01
3.52542371e-01 8.77855897e-01 -6.47616908e-02 -8.13806057e-02
3.78156662e-01 -1.76083922e-01 3.41721863e-01 -1.29763782e+00
1.51730314e-01 -2.54811972e-01 5.22634387e-02 -5.83231091e-01
1.60042554e-01 -1.52737081e+00 1.40757412e-01 -2.04767418e+00
3.71145993e-01 -2.06408679e-01 6.15242459e-02 8.73753786e-01
-2.08796218e-01 -8.00953284e-02 5.21750212e-01 2.17512280e-01
-1.06532884e+00 3.41613531e-01 6.51697755e-01 1.85701758e-01
-2.41552502e-01 -2.39717081e-01 -1.13512802e+00 2.65714079e-01
1.02646995e+00 -8.09914768e-01 1.42579421e-01 -4.01984304e-01
4.35374290e-01 4.65712771e-02 -1.09530888e-01 -1.12238169e+00
3.95553499e-01 4.29366939e-02 5.41438460e-01 -5.06025851e-01
-3.01721305e-01 -2.71366775e-01 1.87636092e-01 1.68657407e-01
-9.49152529e-01 1.77666664e-01 1.72988340e-01 8.00762400e-02
-1.41787380e-01 -5.77946007e-01 3.70804042e-01 -4.00276482e-02
-6.63381219e-01 -2.60029435e-01 -6.39382184e-01 4.38054323e-01
9.27466810e-01 -1.13382123e-01 -5.86231112e-01 -7.00575709e-02
-5.64977169e-01 2.64873862e-01 2.19888777e-01 6.26697958e-01
6.20380700e-01 -1.31067598e+00 -9.14673209e-01 2.17154250e-01
2.31915832e-01 -9.56941843e-02 3.25516053e-02 5.30291259e-01
-2.74872959e-01 6.46567822e-01 -7.63456002e-02 -4.83038612e-02
-1.01401842e+00 1.71732232e-01 -1.30841643e-01 -6.26265824e-01
-3.65706861e-01 5.44058621e-01 -7.25366890e-01 -9.10299182e-01
-1.74787529e-02 -2.17143327e-01 -6.16877615e-01 2.19052836e-01
4.64846969e-01 4.92136300e-01 1.62835822e-01 -5.57695746e-01
-2.11829513e-01 1.08030871e-01 -6.90101646e-03 -2.65361339e-01
1.32699347e+00 -2.09156394e-01 -3.39030117e-01 4.24754858e-01
1.09323382e+00 -1.89645216e-02 -6.55490398e-01 -8.11659098e-02
8.60320568e-01 1.78617671e-01 -3.04648012e-01 -1.23266256e+00
-2.89983064e-01 3.73665482e-01 1.14527211e-01 1.61303848e-01
9.87222135e-01 1.12610236e-01 9.80172515e-01 3.58281106e-01
4.22694027e-01 -1.37644029e+00 -4.20134425e-01 8.77736986e-01
7.28260577e-01 -1.05579638e+00 -3.14150274e-01 -4.98233080e-01
-1.09078133e+00 1.14837408e+00 7.79011726e-01 3.25648308e-01
2.19173759e-01 5.20980656e-01 1.95959620e-02 -1.07619621e-01
-7.91890621e-01 -5.08249737e-02 4.95243579e-01 5.12821376e-01
8.35759163e-01 1.49654552e-01 -7.19267488e-01 1.33975065e+00
-3.47644240e-01 1.10984363e-01 5.59557736e-01 1.12715960e+00
-4.38701689e-01 -1.33853757e+00 -2.74307847e-01 8.77840996e-01
-5.31009853e-01 -4.96225864e-01 -4.59099442e-01 3.98847640e-01
1.25258744e-01 1.34155631e+00 -8.37244466e-02 -7.05393493e-01
5.82442582e-01 5.71394682e-01 -1.85692146e-01 -1.07066834e+00
-1.09679842e+00 -3.04804534e-01 6.12726510e-01 -1.09694257e-01
-3.47750753e-01 -6.22408450e-01 -1.44310904e+00 1.08137287e-01
-3.39379132e-01 5.29776990e-01 8.07735264e-01 8.91951740e-01
9.62115169e-01 6.16261780e-01 6.09098911e-01 -2.48057082e-01
-5.25961339e-01 -1.28415954e+00 -3.57406408e-01 4.64373559e-01
-3.10380280e-01 -1.21677138e-01 -2.28048295e-01 3.09961915e-01] | [9.737373352050781, 9.44758415222168] |
1903758d-3b77-4a80-acbd-05c722e0ce2e | sparse2dense-from-direct-sparse-odometry-to | 1903.09199 | null | http://arxiv.org/abs/1903.09199v1 | http://arxiv.org/pdf/1903.09199v1.pdf | Sparse2Dense: From direct sparse odometry to dense 3D reconstruction | In this paper, we proposed a new deep learning based dense monocular SLAM
method. Compared to existing methods, the proposed framework constructs a dense
3D model via a sparse to dense mapping using learned surface normals. With
single view learned depth estimation as prior for monocular visual odometry, we
obtain both accurate positioning and high quality depth reconstruction. The
depth and normal are predicted by a single network trained in a tightly coupled
manner.Experimental results show that our method significantly improves the
performance of visual tracking and depth prediction in comparison to the
state-of-the-art in deep monocular dense SLAM. | ['Patric Jensfelt', 'Jiexiong Tang', 'John Folkesson'] | 2019-03-21 | null | null | null | null | ['monocular-visual-odometry'] | ['robots'] | [-2.27304801e-01 -2.02931929e-02 -3.72879058e-01 -5.11086702e-01
-4.03368920e-01 -2.86763638e-01 7.17284322e-01 -3.42286885e-01
-1.72128215e-01 8.48458827e-01 5.29451333e-02 -6.66091510e-04
3.00135523e-01 -8.12267601e-01 -9.83058572e-01 -3.48197073e-01
2.57047921e-01 8.88247669e-01 2.57289290e-01 1.82536766e-01
3.82604837e-01 6.32818401e-01 -1.44212484e+00 -3.83184791e-01
7.55631983e-01 1.09921885e+00 5.29108167e-01 5.12227118e-01
-2.77152169e-03 1.14103580e+00 -6.98537305e-02 2.36528680e-01
4.82136935e-01 -1.20450586e-01 -4.39080566e-01 4.21469510e-02
1.13302517e+00 -6.84768140e-01 -9.04243588e-01 1.09935439e+00
4.41239089e-01 -1.24055512e-01 2.92291760e-01 -9.69231427e-01
-3.30685556e-01 -3.04649651e-01 -5.05346477e-01 -3.05470109e-01
8.54520023e-01 -1.59817785e-01 6.10554099e-01 -1.15932631e+00
1.06168866e+00 1.27084231e+00 7.99417973e-01 3.52910876e-01
-8.83841455e-01 -6.67455018e-01 1.77751467e-01 1.20041817e-01
-1.49190569e+00 -5.04184008e-01 6.67409718e-01 -5.68827510e-01
1.19546497e+00 -4.52154189e-01 9.29985523e-01 8.21067214e-01
6.99416876e-01 5.13625026e-01 9.99008894e-01 -2.33509213e-01
1.81602120e-01 -2.14364573e-01 -3.70994896e-01 9.76138055e-01
5.59470594e-01 6.74503684e-01 -9.53036368e-01 5.21011390e-02
1.21479380e+00 3.99908125e-01 -3.37684929e-01 -1.25371194e+00
-1.25295007e+00 7.72112310e-01 9.06436205e-01 -1.84857428e-01
-3.25985640e-01 6.73361719e-01 -9.18468460e-02 2.00659350e-01
5.54215372e-01 3.27806957e-02 -2.58361906e-01 -9.58710387e-02
-8.41304302e-01 1.89140350e-01 6.62544906e-01 1.29976666e+00
1.42047322e+00 2.73935020e-01 5.70053816e-01 1.98215663e-01
8.17755282e-01 1.01789784e+00 3.08483988e-01 -1.28791463e+00
3.83607745e-01 5.81362426e-01 4.40055609e-01 -9.59302366e-01
-4.92376626e-01 -4.93884802e-01 -6.64657474e-01 5.25540769e-01
-1.10730022e-01 1.49710374e-02 -1.06281424e+00 1.28946173e+00
3.86054605e-01 4.90063280e-01 8.97841007e-02 9.78963315e-01
7.84495652e-01 4.17774618e-01 -8.00212264e-01 8.34269673e-02
4.22859192e-01 -1.09527349e+00 -7.36489475e-01 -9.12708342e-01
3.47331077e-01 -8.21494699e-01 3.09506834e-01 3.63646895e-01
-8.20941329e-01 -5.87524891e-01 -1.24030352e+00 -4.13371503e-01
1.81171566e-01 -1.70227718e-02 8.87886286e-01 1.70821890e-01
-1.41079390e+00 3.69714558e-01 -1.20161903e+00 -5.60497403e-01
1.58495292e-01 4.56689000e-01 -5.29766500e-01 -5.95807970e-01
-8.20171714e-01 1.14284861e+00 2.85530508e-01 -8.95916596e-02
-1.24336720e+00 -4.51024383e-01 -1.57752514e+00 -5.07577538e-01
-1.33134509e-02 -1.39077640e+00 1.27908731e+00 -1.85256958e-01
-1.75883460e+00 1.09396160e+00 -6.99102402e-01 -6.88742518e-01
4.56829309e-01 -7.59362161e-01 3.65328193e-01 1.25768214e-01
4.17652845e-01 9.93160963e-01 5.48587441e-01 -1.56953967e+00
-7.62485087e-01 -6.31137967e-01 -3.35834338e-03 5.62117815e-01
7.38566101e-01 -7.37040222e-01 -5.35946131e-01 3.72624516e-01
1.13569915e+00 -1.04134715e+00 -3.43616813e-01 5.13673007e-01
-1.61099985e-01 4.24066424e-01 8.99655819e-01 -2.59225011e-01
4.29531544e-01 -1.58495653e+00 3.97637069e-01 -1.59551457e-01
5.78535259e-01 -3.20450068e-01 4.92742985e-01 3.34785461e-01
6.17199242e-01 -7.39364862e-01 3.12176377e-01 -1.06416166e+00
1.23837944e-02 4.97377396e-01 -2.56506532e-01 1.10757148e+00
-4.89366114e-01 1.00360382e+00 -1.11172259e+00 -1.93936750e-01
8.68171453e-01 6.61697626e-01 -5.06076992e-01 4.21684384e-01
-1.69146672e-01 9.37740743e-01 -1.56904817e-01 9.32122111e-01
8.47862482e-01 -4.86517251e-01 1.50403455e-01 -7.31601566e-02
-2.85683870e-01 5.45547962e-01 -8.82125318e-01 2.69697738e+00
-6.92842722e-01 8.30560267e-01 7.60142058e-02 -3.58977169e-01
1.35530233e+00 -6.62785619e-02 3.76059324e-01 -7.92050302e-01
1.43120274e-01 3.43540013e-01 -5.53891540e-01 1.17337711e-01
7.53368437e-01 -3.53460938e-01 8.15801248e-02 -1.29363596e-01
1.02635600e-01 -6.51241302e-01 -6.78831100e-01 2.71888375e-01
9.95530784e-01 7.99342692e-01 6.19449854e-01 -7.63208652e-03
3.51938546e-01 1.10953145e-01 7.65497983e-01 5.67131877e-01
-1.12784542e-01 5.69388747e-01 -1.62613451e-01 -6.57290220e-01
-1.11320496e+00 -1.25843644e+00 -9.73311961e-02 6.97881356e-02
9.81028795e-01 -2.58188933e-01 7.62121081e-02 -6.61433190e-02
4.34353560e-01 6.51956350e-02 -3.85079682e-01 2.36553252e-01
-4.32002425e-01 -3.22465296e-03 1.57009929e-01 5.20655751e-01
7.45893896e-01 -5.03437877e-01 -5.97386777e-01 1.51596889e-01
-3.10936850e-02 -1.53213978e+00 -7.37235835e-03 2.28703901e-01
-1.20170617e+00 -9.55990016e-01 -4.64493781e-01 -8.65359187e-01
4.90805656e-01 7.77228892e-01 9.96493101e-01 -2.87472069e-01
4.58809212e-02 3.44704598e-01 -1.32842176e-02 -1.88979000e-01
2.34121131e-03 -5.82763515e-02 2.62575537e-01 -4.64817792e-01
4.33108538e-01 -6.87579274e-01 -7.24774420e-01 2.17064500e-01
-8.29450786e-02 3.38514954e-01 4.01335627e-01 7.19897747e-01
9.76919115e-01 -6.49437964e-01 -6.38201758e-02 -5.50565779e-01
-3.84894341e-01 -3.81120443e-01 -1.35152006e+00 -5.46065688e-01
-8.24297249e-01 -9.66488644e-02 8.22052956e-02 1.26966223e-01
-8.61995220e-01 6.14121079e-01 -1.99728221e-01 -7.52960980e-01
-1.73197195e-01 3.09068799e-01 -5.06537780e-03 -8.38336229e-01
4.30516005e-01 3.72893363e-01 1.48147061e-01 -4.82734561e-01
2.23613739e-01 2.59624481e-01 6.90816641e-01 -1.84518173e-01
1.08084095e+00 1.15695810e+00 3.38042676e-01 -5.37813485e-01
-9.05904174e-01 -7.29659200e-01 -1.12303174e+00 -8.66117701e-02
8.20911765e-01 -1.85681891e+00 -6.70555830e-01 6.34034991e-01
-1.27929282e+00 -3.66998553e-01 6.50282875e-02 9.16913509e-01
-8.13265026e-01 4.86227185e-01 -6.45313442e-01 -6.18591070e-01
-2.22479969e-01 -1.04178131e+00 1.69098663e+00 9.94139016e-02
-2.49950998e-02 -1.32711887e+00 6.61179185e-01 2.07211897e-02
1.55286148e-01 5.66081643e-01 3.15303020e-02 4.18653190e-01
-1.53329003e+00 -1.42440155e-01 -3.58223736e-01 -2.49675244e-01
1.75147787e-01 -6.16037369e-01 -1.05422044e+00 -6.02245450e-01
8.80143121e-02 -3.37426394e-01 7.76356161e-01 6.19529605e-01
1.55672655e-01 2.41575986e-01 -6.73356473e-01 1.56846321e+00
1.93956530e+00 2.48128116e-01 6.86638951e-01 7.57130027e-01
1.35298157e+00 -4.54354682e-04 8.69471669e-01 5.19867837e-01
8.19570601e-01 7.23865449e-01 1.03678083e+00 -5.94571456e-02
9.80425030e-02 -7.30700314e-01 1.87026218e-01 6.80494368e-01
3.70657235e-01 2.73741126e-01 -8.97894800e-01 6.11000657e-01
-2.01340985e+00 -6.31525695e-01 -7.06181079e-02 2.01678443e+00
3.42343718e-01 1.57752767e-01 -5.27940512e-01 -1.84600830e-01
1.80571377e-01 5.35584509e-01 -6.99989021e-01 1.22881301e-01
-1.37081459e-01 1.06583118e-01 9.42258239e-01 1.23920357e+00
-9.35621440e-01 1.46164584e+00 6.35257101e+00 -6.99784383e-02
-1.35748911e+00 2.67195731e-01 -5.55370092e-01 -1.73473448e-01
-4.95569646e-01 2.54748851e-01 -1.34664154e+00 -3.26534826e-03
5.11761963e-01 1.92990527e-02 7.24523515e-02 1.07554579e+00
4.46006320e-02 -4.89174128e-01 -1.08421695e+00 1.60346735e+00
2.69912034e-01 -1.58783543e+00 -2.05739662e-01 4.89431947e-01
1.29850841e+00 8.16729188e-01 -2.24502131e-01 -7.02498183e-02
5.65282643e-01 -9.26150918e-01 7.64140964e-01 3.47245991e-01
8.20496321e-01 -4.95017260e-01 7.50907540e-01 6.22485340e-01
-1.42945421e+00 2.92170465e-01 -8.29512537e-01 -7.69331038e-01
5.02890110e-01 7.20442533e-01 -8.54414999e-01 7.64244795e-01
4.97530043e-01 1.59725416e+00 -1.74246907e-01 1.19893003e+00
-6.22580349e-01 -2.20421582e-01 -2.67663956e-01 3.47654849e-01
3.15011680e-01 5.07033656e-05 5.06642520e-01 5.13685524e-01
4.64748710e-01 -2.91144103e-01 4.51458156e-01 7.78561771e-01
4.86010388e-02 -3.69858891e-01 -1.38445604e+00 6.75880671e-01
5.92176557e-01 7.60524273e-01 -6.28871173e-02 -2.44814143e-01
-5.04288495e-01 1.16884518e+00 4.32160407e-01 1.65950775e-01
-4.53065872e-01 1.05185091e-01 1.04203367e+00 1.85937211e-02
3.17216992e-01 -7.81127393e-01 -4.92833674e-01 -1.56870997e+00
-1.26477294e-02 -2.24996999e-01 -2.48969078e-01 -1.08282018e+00
-9.48450506e-01 6.56750977e-01 -3.97508204e-01 -1.55285478e+00
-5.99891424e-01 -5.91968715e-01 -9.28032771e-02 1.11418951e+00
-2.19584441e+00 -1.23973536e+00 -1.03855848e+00 6.31967604e-01
3.64033729e-01 -4.41128276e-02 6.59120202e-01 2.46583477e-01
3.65336865e-01 -8.53646919e-02 6.75681680e-02 -1.58135235e-01
8.36244524e-01 -1.23628402e+00 9.34627175e-01 8.55381489e-01
2.03598574e-01 4.82854068e-01 5.81405640e-01 -1.02679968e+00
-1.69700122e+00 -1.07433319e+00 9.66606021e-01 -6.38541341e-01
3.49894017e-01 -3.27715248e-01 -5.70536494e-01 1.21219885e+00
5.94406761e-02 1.96341112e-01 -6.31120871e-04 -4.04481478e-02
-7.34698102e-02 -5.38415611e-02 -1.05174005e+00 3.74718383e-02
1.34903312e+00 -8.93523335e-01 -5.31826496e-01 1.59555316e-01
8.58884335e-01 -1.37861466e+00 -6.40559554e-01 4.85528797e-01
7.73991644e-01 -1.26546836e+00 1.04608035e+00 2.31829390e-01
-1.22883907e-02 -5.18209577e-01 -5.34719110e-01 -1.15180731e+00
-3.14335793e-01 -5.00462234e-01 -4.80773002e-01 2.64604568e-01
-1.76877782e-01 -7.01641619e-01 1.35638011e+00 -1.69136077e-01
-3.89143646e-01 -4.78579730e-01 -9.93428767e-01 -7.32022583e-01
-3.77343208e-01 -2.81074286e-01 3.26225519e-01 7.27318108e-01
-4.53038931e-01 4.80999678e-01 -8.76261473e-01 7.34833002e-01
1.16422319e+00 4.24931824e-01 1.32942295e+00 -1.52760363e+00
-1.42509177e-01 3.27645212e-01 -1.05178392e+00 -2.11889553e+00
4.01231706e-01 -6.51421726e-01 2.35760599e-01 -2.05116320e+00
-1.01629227e-01 -2.09893689e-01 3.00652266e-01 -1.04559434e-03
2.53804773e-01 3.16034138e-01 7.16400938e-03 4.45344746e-01
-6.40918911e-01 8.28509808e-01 1.27789509e+00 1.38132662e-01
-1.37301609e-01 -6.50526136e-02 -1.32447323e-02 9.15125906e-01
3.69012147e-01 -4.47732985e-01 -5.01764357e-01 -9.52914178e-01
3.24240625e-01 4.71021324e-01 5.31182706e-01 -1.35414886e+00
4.98970598e-01 -6.60836250e-02 5.65900445e-01 -1.39631605e+00
8.11995387e-01 -8.94499242e-01 1.69250563e-01 7.29338706e-01
4.37927395e-01 1.05249710e-01 6.18581362e-02 7.54248738e-01
-4.39622253e-01 2.72558391e-01 7.42404401e-01 -3.48785192e-01
-1.40454578e+00 7.31516957e-01 1.33983359e-01 -2.38903791e-01
7.28066325e-01 -3.58783573e-01 -3.57532620e-01 -7.13883281e-01
-5.40752172e-01 1.31093293e-01 1.29934156e+00 3.97103161e-01
1.05762351e+00 -1.60941112e+00 -1.63852960e-01 4.69120383e-01
2.33133897e-01 5.76147974e-01 4.04078960e-02 5.94706237e-01
-1.10021329e+00 9.15495992e-01 -4.22901392e-01 -1.31881666e+00
-9.56102669e-01 2.52273917e-01 5.89091241e-01 -1.01037547e-01
-1.00862312e+00 7.82785475e-01 5.23193896e-01 -8.26862276e-01
3.93934935e-01 -4.33981180e-01 3.02498430e-01 -6.34130716e-01
3.14245790e-01 1.97118610e-01 -1.79197162e-01 -7.86824465e-01
-6.00285888e-01 1.15197074e+00 3.73479605e-01 -1.26946554e-01
1.16095209e+00 -7.94806659e-01 -4.08371016e-02 6.25168085e-01
1.19982970e+00 3.87198627e-02 -1.77544856e+00 -4.59834754e-01
-3.60033363e-01 -9.73413408e-01 1.90923437e-01 -2.88825423e-01
-7.51725316e-01 1.00048363e+00 6.21594906e-01 -7.59493291e-01
5.93698442e-01 -1.09790219e-02 7.86954880e-01 4.82886851e-01
1.33389139e+00 -3.80027354e-01 7.39261955e-02 1.06153369e+00
5.67734897e-01 -1.69157064e+00 2.65448034e-01 -4.86038148e-01
-1.94996268e-01 8.48835826e-01 8.94140184e-01 -4.96520013e-01
5.39084971e-01 3.65323633e-01 3.93074632e-01 -2.93002903e-01
-6.56727433e-01 -2.58300871e-01 1.70023173e-01 8.47103417e-01
1.11920185e-01 -2.44349539e-01 4.25740600e-01 -5.25573909e-01
-1.31402895e-01 1.99994639e-01 3.98758918e-01 1.20985413e+00
-9.23046649e-01 -1.00317097e+00 -3.82903010e-01 -1.91173062e-01
2.38978669e-01 -9.00571495e-02 -2.65754014e-01 9.31164980e-01
-7.22255111e-02 4.46333855e-01 1.68813556e-01 -5.00924647e-01
6.75609484e-02 -2.95202672e-01 1.07286274e+00 -1.02597976e+00
8.66315290e-02 1.94919556e-02 -6.53905794e-02 -1.04114258e+00
-3.68172258e-01 -4.17086661e-01 -1.36286676e+00 -4.71698433e-01
-1.59644429e-02 -3.03376645e-01 9.26200628e-01 1.11500037e+00
3.48107368e-01 1.74475044e-01 4.14017141e-01 -1.48579526e+00
-2.21654266e-01 -7.73882508e-01 -6.86559916e-01 -1.50311038e-01
8.81504774e-01 -1.01964629e+00 -3.06824952e-01 -3.27467293e-01] | [8.031820297241211, -2.2456493377685547] |
c9e0fec2-496c-438a-a8ad-001f1138a913 | shallow-semantic-reasoning-from-an-incomplete | null | null | https://aclanthology.org/W16-0512 | https://aclanthology.org/W16-0512.pdf | Shallow Semantic Reasoning from an Incomplete Gold Standard for Learner Language | null | ['Markus Dickinson', 'Levi King'] | 2016-06-01 | null | null | null | ws-2016-6 | ['grammatical-error-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] | [-7.238481044769287, 3.7889959812164307] |
78884a1a-611f-4e6f-bf57-902431a30765 | openhps-an-open-source-hybrid-positioning | 2101.05198 | null | https://arxiv.org/abs/2101.05198v1 | https://arxiv.org/pdf/2101.05198v1.pdf | OpenHPS: An Open Source Hybrid Positioning System | Positioning systems and frameworks use various techniques to determine the position of an object. Some of the existing solutions combine different sensory data at the time of positioning in order to compute more accurate positions by reducing the error introduced by the used individual positioning techniques. We present OpenHPS, a generic hybrid positioning system implemented in TypeScript, that can not only reduce the error during tracking by fusing different sensory data based on different algorithms, but also also make use of combined tracking techniques when calibrating or training the system. In addition to a detailed discussion of the architecture, features and implementation of the extensible open source OpenHPS framework, we illustrate the use of our solution in a demonstrator application fusing different positioning techniques. While OpenHPS offers a number of positioning techniques, future extensions might integrate new positioning methods or algorithms and support additional levels of abstraction including symbolic locations. | ['Beat Signer', 'Maxim Van de Wynckel'] | 2020-12-29 | null | null | null | null | ['hybrid-positioning'] | ['computer-vision'] | [-2.08638132e-01 -2.64375303e-02 1.86486647e-01 -3.39109778e-01
-4.26897943e-01 -1.08627045e+00 5.06937087e-01 1.81561753e-01
-4.04279858e-01 5.00382364e-01 -2.00437099e-01 -3.34443599e-01
-5.07128358e-01 -8.43888044e-01 -8.31666231e-01 -4.57664609e-01
2.78551560e-02 6.53643608e-01 9.55257297e-01 -5.13594925e-01
5.09724140e-01 6.89320743e-01 -1.82280958e+00 -1.64633542e-01
3.66028368e-01 9.11576867e-01 4.72266614e-01 7.98845053e-01
1.26250237e-02 1.54229537e-01 -9.91953135e-01 2.86173910e-01
4.33895618e-01 1.25795305e-01 -2.41803691e-01 -7.69444644e-01
4.37267542e-01 -5.50472923e-02 3.79779398e-01 6.92710042e-01
8.32150340e-01 8.78603458e-02 -6.46356866e-02 -1.38735139e+00
-3.00751962e-02 6.48818970e-01 3.09021566e-02 -1.57727882e-01
9.71145451e-01 -2.49157056e-01 1.55822456e-01 -2.58366793e-01
4.73865837e-01 9.21195507e-01 1.67417586e+00 2.81702846e-01
-1.26545429e+00 -5.52339017e-01 -2.91393638e-01 -1.19995594e-01
-1.57670581e+00 -5.08827806e-01 4.82169092e-01 -3.36846411e-01
7.22965240e-01 6.98657691e-01 4.71972495e-01 8.41376603e-01
2.71310776e-01 1.04712501e-01 9.92661238e-01 -6.19881928e-01
3.36435705e-01 2.22703859e-01 1.25984147e-01 2.57122576e-01
4.25516605e-01 2.19939053e-01 -4.36012924e-01 -3.24135840e-01
8.22642982e-01 -1.13504879e-01 7.41951987e-02 -7.08729923e-01
-1.63900948e+00 2.15161636e-01 5.36923051e-01 5.47524571e-01
-1.15143783e-01 6.38722837e-01 1.10870734e-01 1.69792160e-01
-2.75866985e-01 5.49064815e-01 -8.28675210e-01 -4.89556879e-01
-1.07626855e+00 6.55166030e-01 8.20807219e-01 1.39644468e+00
8.33097994e-01 -1.67040840e-01 1.02121890e-01 2.61120766e-01
7.19085872e-01 7.93079138e-01 3.94797176e-01 -1.16573656e+00
3.48671436e-01 5.35357833e-01 6.06577873e-01 -1.14616144e+00
-8.93849671e-01 -1.71359599e-01 -5.84305776e-03 5.22429764e-01
4.81644720e-01 -4.17887658e-01 -5.64714730e-01 1.46537387e+00
6.23164594e-01 1.58093460e-02 -1.11406587e-01 5.63816726e-01
5.70642889e-01 3.30267519e-01 -2.39910841e-01 3.28179836e-01
9.96960282e-01 -5.19610882e-01 -6.86392367e-01 -1.22036435e-01
6.74166262e-01 -1.00900578e+00 5.07360995e-01 3.48381191e-01
-8.49253654e-01 -9.55680966e-01 -1.56854057e+00 -1.18278526e-01
-9.47525799e-01 3.01214550e-02 3.17557037e-01 1.04188418e+00
-1.31948519e+00 9.80072081e-01 -1.03884447e+00 -4.63972688e-01
-5.81040621e-01 1.00065327e+00 -1.01499103e-01 5.73325515e-01
-1.12548447e+00 1.29002213e+00 5.01153588e-01 2.13376284e-01
2.08194450e-01 -8.00724745e-01 -7.26996601e-01 -4.02852535e-01
2.48652741e-01 -7.56289601e-01 1.43239081e+00 -1.83336720e-01
-1.85353076e+00 3.19324136e-01 3.06205213e-01 -5.53063035e-01
5.48085570e-01 -3.20417255e-01 -7.01310098e-01 -2.95387477e-01
3.25657338e-01 6.22861028e-01 3.73934358e-01 -9.19012427e-01
-7.01470852e-01 -2.69673854e-01 1.14602245e-01 -2.39887387e-01
4.32561994e-01 -3.57982188e-01 -2.92524040e-01 -2.80051321e-01
5.09809315e-01 -1.26169920e+00 -3.69318813e-01 1.19448669e-01
-1.00341141e-01 2.70294219e-01 1.05706155e+00 -3.10996354e-01
1.07160997e+00 -2.01362157e+00 2.82803569e-02 4.57508057e-01
-2.36502528e-01 2.64342666e-01 4.88990039e-01 6.13158643e-01
3.49652410e-01 -3.91997725e-01 4.18745279e-01 -1.00867525e-01
5.01103520e-01 3.50701660e-01 -1.44469649e-01 4.09806609e-01
-6.48627818e-01 5.55751145e-01 -8.01645756e-01 -9.20891464e-02
9.07518327e-01 6.11865759e-01 -2.06685916e-01 -4.58654344e-01
-1.67669237e-01 5.89896739e-01 -4.30769205e-01 4.61162031e-01
7.27423191e-01 3.84886295e-01 2.17371255e-01 -1.91889435e-01
-7.98013806e-01 2.80974030e-01 -1.99328303e+00 1.96827066e+00
-5.22647440e-01 2.94684678e-01 1.19285770e-01 -2.69409716e-01
1.07122648e+00 1.47481054e-01 4.79092419e-01 -6.90896988e-01
2.85798341e-01 6.00339234e-01 -2.42506281e-01 -3.58565629e-01
1.04275751e+00 3.63083482e-01 -6.66261196e-01 1.89057618e-01
-3.08386292e-02 7.70369619e-02 5.14700226e-02 -2.38738313e-01
9.74700987e-01 1.08076251e+00 4.91123646e-01 -2.44207233e-01
7.13119566e-01 1.67002842e-01 2.67328143e-01 9.02917504e-01
-4.68256278e-03 4.71912622e-01 -1.36413053e-01 -3.01111460e-01
-9.08262968e-01 -9.15894270e-01 -4.25038278e-01 1.06218576e+00
5.56657374e-01 -7.60814011e-01 -4.91785705e-01 -1.01700865e-01
5.12579143e-01 5.31583250e-01 -3.82604539e-01 3.02679688e-01
-6.70763612e-01 -6.60687685e-02 1.02057040e+00 8.42542708e-01
4.51797217e-01 -5.64053774e-01 -1.18407309e+00 4.01977539e-01
2.19431669e-01 -8.11962783e-01 7.65001476e-02 3.40698451e-01
-8.88080359e-01 -8.79151464e-01 -3.74111682e-01 -4.18927699e-01
4.28066045e-01 1.74456120e-01 7.22132325e-01 2.01775998e-01
2.54043400e-01 6.43451869e-01 -4.22877103e-01 -7.04469323e-01
-3.59182209e-01 3.15520227e-01 1.63173243e-01 -7.14979708e-01
2.34174374e-02 -5.73877573e-01 -3.87739092e-01 6.68288410e-01
-3.24563295e-01 -1.21874779e-01 3.74915302e-01 6.21597469e-02
4.76046175e-01 -1.72766283e-01 2.03317657e-01 -5.24475396e-01
2.47443780e-01 -3.38400632e-01 -1.29384017e+00 1.00980684e-01
-4.98798400e-01 2.44177371e-01 3.38956177e-01 -2.63371944e-01
-6.58868611e-01 4.69005227e-01 -4.54509974e-01 1.48669491e-02
-2.79884547e-01 3.84537011e-01 4.36984487e-02 -8.21530938e-01
7.69977868e-01 -1.70551553e-01 7.59650767e-02 -6.82517111e-01
4.53824759e-01 5.83946705e-01 7.40766168e-01 -6.08272254e-01
7.31926084e-01 1.47197261e-01 2.93502271e-01 -4.25545692e-01
-1.24046020e-01 -5.63516617e-01 -9.32335317e-01 -4.04375792e-01
3.21033627e-01 -4.77013737e-01 -1.30481708e+00 3.20721388e-01
-9.81253326e-01 -1.96433514e-01 -2.06829920e-01 2.24665046e-01
-5.48970759e-01 1.42965704e-01 1.42697617e-01 -6.90013528e-01
4.34114933e-02 -1.17073524e+00 1.37827110e+00 3.30019414e-01
-4.34379190e-01 -9.46912169e-01 6.01591945e-01 -5.42059429e-02
7.21326828e-01 6.57577038e-01 -1.36688158e-01 -4.31759983e-01
-5.68322420e-01 -8.16509545e-01 5.67780316e-01 -3.32877338e-01
-1.40685529e-01 9.65441912e-02 -7.12188482e-01 -1.15461752e-01
-2.62613058e-01 4.89330441e-01 -2.42117703e-01 2.07620516e-01
3.23431671e-01 1.73256382e-01 -7.91446686e-01 3.67327958e-01
1.63825285e+00 3.61968875e-01 8.02884340e-01 1.09299564e+00
5.67385733e-01 1.50645256e-01 1.15228832e+00 -1.32154385e-02
4.15387958e-01 1.58131886e+00 4.18269783e-01 4.59837854e-01
7.47122839e-02 -1.24260284e-01 3.20770860e-01 4.32796031e-01
-1.98350176e-01 3.02405864e-01 -9.36518192e-01 7.29097947e-02
-2.10067081e+00 -9.62534964e-01 -7.98309863e-01 2.43853521e+00
4.34727013e-01 3.47486213e-02 3.54897171e-01 5.14254868e-01
7.38983333e-01 -3.73406261e-02 -4.64481637e-02 -5.19283891e-01
4.91524965e-01 1.37731016e-01 1.16867280e+00 5.00513673e-01
-9.28323567e-01 5.84145844e-01 7.18724203e+00 9.16353092e-02
-1.30403638e+00 4.78182547e-02 -9.16647434e-01 8.71573314e-02
3.90600204e-01 7.61839002e-02 -1.19707298e+00 7.64573038e-01
1.30125582e+00 -4.41617109e-02 8.18806142e-02 1.28081381e+00
-1.68479532e-01 -5.53982735e-01 -1.07459843e+00 7.17922270e-01
-1.22555114e-01 -1.51822722e+00 -9.39173341e-01 -4.08036597e-02
3.89876097e-01 5.68909906e-02 -1.77352086e-01 2.80269504e-01
1.89326834e-02 -8.35073069e-02 1.42569017e+00 5.35985529e-01
2.09577948e-01 -4.62762535e-01 9.55249608e-01 4.44247276e-01
-1.40959477e+00 8.66531432e-02 -1.80326626e-01 -4.89110619e-01
2.30639800e-01 6.81458712e-02 -8.34820390e-01 9.76216972e-01
7.15716004e-01 1.38545096e-01 -9.10430014e-01 1.45432496e+00
-1.46824285e-01 -3.10697228e-01 -9.25700843e-01 -2.32279494e-01
-7.93515146e-02 1.27826914e-01 3.85368377e-01 7.99252212e-01
6.52574360e-01 -4.56781834e-01 -4.44461666e-02 5.63319504e-01
7.81828105e-01 -3.35459888e-01 -5.73918700e-01 7.31598258e-01
8.44006062e-01 1.16279697e+00 -8.15261066e-01 -2.12842554e-01
-6.56643957e-02 3.67696017e-01 -4.13432479e-01 -2.35889286e-01
-1.14511573e+00 -6.42512381e-01 2.25551337e-01 4.12075937e-01
4.51092482e-01 -6.08022332e-01 -3.84679824e-01 -5.13074100e-01
4.82077226e-02 -2.52423495e-01 -6.31179586e-02 -1.20216632e+00
-1.95017010e-01 3.04776132e-01 3.87188733e-01 -1.73435128e+00
-7.76899397e-01 -6.10878229e-01 -1.13747038e-01 8.49225998e-01
-7.20532656e-01 -1.07966125e+00 -4.18584257e-01 2.75268674e-01
-6.68460876e-02 3.29080284e-01 1.00854254e+00 7.45853961e-01
-1.15600624e-03 3.22907329e-01 1.80804923e-01 -2.80100822e-01
8.76025558e-01 -1.31016076e+00 5.67677855e-01 6.73131168e-01
-8.62858891e-02 1.10909855e+00 1.18713760e+00 -7.22683370e-01
-1.46650112e+00 -6.36907637e-01 6.34541273e-01 -8.55463147e-01
5.94947219e-01 -4.68375415e-01 -4.84170794e-01 1.10952187e+00
-1.15015591e-02 -6.79108053e-02 3.69356155e-01 2.27445900e-01
1.10351019e-01 -4.30570424e-01 -1.49027872e+00 4.20816660e-01
7.41139293e-01 -2.14438036e-01 -9.27317619e-01 -4.23328765e-02
3.86092871e-01 -1.38183546e+00 -1.03549182e+00 2.33961284e-01
1.07551479e+00 -1.25469780e+00 1.22028685e+00 4.85685915e-01
-5.95130920e-01 -1.24859965e+00 -7.71628246e-02 -1.00267899e+00
-2.76261151e-01 -6.09468579e-01 5.21497708e-03 1.36407089e+00
2.70281583e-01 -9.66012299e-01 5.47889292e-01 5.01341105e-01
-4.32819694e-01 -1.55518159e-01 -9.95090008e-01 -8.44885945e-01
-6.22534633e-01 -6.48461580e-01 1.01283121e+00 5.52592158e-01
6.42925501e-02 1.54888062e-02 -3.33170742e-01 7.53670454e-01
3.55393291e-01 -1.16848484e-01 1.32893991e+00 -1.39078331e+00
-2.61655748e-01 1.94997676e-02 -1.15717638e+00 -8.04696560e-01
-5.02583265e-01 -2.33875275e-01 3.41435015e-01 -1.43096685e+00
-9.92906690e-01 -7.84066677e-01 1.36300221e-01 4.92658257e-01
5.82200944e-01 4.81144935e-01 2.13214889e-01 7.49838278e-02
-7.48791218e-01 -4.41929936e-01 3.36059123e-01 3.13249677e-01
-2.91696429e-01 4.54895705e-01 -3.70849401e-01 7.81680048e-01
5.36766112e-01 -6.49014950e-01 -3.16145867e-01 -5.29079854e-01
6.87297761e-01 -6.59255907e-02 6.75973833e-01 -1.92434907e+00
9.15580869e-01 2.47285053e-01 6.38282537e-01 -1.26431656e+00
3.00716072e-01 -1.47515166e+00 1.14086533e+00 6.43168867e-01
2.81349957e-01 3.14057350e-01 5.65584779e-01 2.80840397e-01
2.21583232e-01 -3.72206420e-01 2.88277596e-01 5.01697809e-02
-1.00009143e+00 -5.71239054e-01 -3.13291669e-01 -9.93859529e-01
1.40761864e+00 -8.09542775e-01 -4.66490030e-01 2.28238627e-01
-9.35769320e-01 1.25500306e-01 8.36600244e-01 6.40562832e-01
-1.08146317e-01 -1.56971955e+00 5.61603248e-01 2.55102366e-01
6.23168945e-02 -4.05804813e-02 4.55205217e-02 9.53827620e-01
-1.05955625e+00 7.85104811e-01 -5.45506120e-01 -8.83790791e-01
-1.24393868e+00 5.29919326e-01 3.56612504e-01 2.98518419e-01
-5.45070171e-01 1.59963757e-01 -8.83270979e-01 -8.20568323e-01
2.47172073e-01 -6.79398537e-01 -1.94309670e-02 -1.15099750e-01
5.67579865e-01 6.85763836e-01 6.76433682e-01 -7.09789097e-01
-8.57599556e-01 1.11486995e+00 5.45241475e-01 -3.85721534e-01
9.96910989e-01 -3.65622938e-01 1.66408122e-01 6.78933859e-01
5.46056092e-01 4.90947932e-01 -9.40348089e-01 3.63086969e-01
4.06294286e-01 -4.27357346e-01 -1.96189880e-01 -9.00431335e-01
-5.10054588e-01 2.03537598e-01 8.95780027e-01 5.02543151e-01
6.34020209e-01 -2.70939410e-01 3.86817276e-01 5.01110494e-01
1.21482086e+00 -1.16121829e+00 -7.16444612e-01 4.21386331e-01
6.56015456e-01 -5.99084377e-01 3.00195128e-01 -2.27978170e-01
-2.30936129e-02 1.29743493e+00 2.22071052e-01 -3.37007493e-01
4.92840171e-01 9.43107903e-01 2.67524213e-01 -6.65890053e-02
9.15723294e-02 -6.40425310e-02 4.81471345e-02 1.09913540e+00
3.32824886e-01 -1.80953294e-01 -4.48195279e-01 5.52019715e-01
-5.38043559e-01 2.03944609e-01 2.81057775e-01 1.69170058e+00
-5.33577859e-01 -1.66158688e+00 -1.17651832e+00 -1.60812438e-01
-1.66163817e-01 5.32349586e-01 -3.02993000e-01 1.15521395e+00
7.74168551e-01 7.04491377e-01 -2.63570491e-02 -6.50499463e-01
8.45861733e-01 -1.68947894e-02 8.41733038e-01 -4.35310096e-01
-1.01590049e+00 -5.70312403e-02 1.95519105e-01 -7.94654608e-01
-6.40028000e-01 -8.52799594e-01 -1.43568707e+00 -6.25392795e-01
-5.66094160e-01 5.01544885e-02 1.23625624e+00 7.72403717e-01
6.76055551e-01 7.87764013e-01 9.35540870e-02 -1.52898419e+00
-2.63171375e-01 -6.66362524e-01 -3.65816206e-01 -3.92267287e-01
2.28787705e-01 -1.08604360e+00 -5.63458018e-02 -2.81871200e-01] | [7.312210559844971, -2.0051138401031494] |
367d9b92-4470-481f-a6f3-ab4692d3a84c | forecasting-directional-movements-of-stock | 2004.10178 | null | https://arxiv.org/abs/2004.10178v2 | https://arxiv.org/pdf/2004.10178v2.pdf | Forecasting directional movements of stock prices for intraday trading using LSTM and random forests | We employ both random forests and LSTM networks (more precisely CuDNNLSTM) as training methodologies to analyze their effectiveness in forecasting out-of-sample directional movements of constituent stocks of the S&P 500 from January 1993 till December 2018 for intraday trading. We introduce a multi-feature setting consisting not only of the returns with respect to the closing prices, but also with respect to the opening prices and intraday returns. As trading strategy, we use Krauss et al. (2017) and Fischer & Krauss (2018) as benchmark. On each trading day, we buy the 10 stocks with the highest probability and sell short the 10 stocks with the lowest probability to outperform the market in terms of intraday returns -- all with equal monetary weight. Our empirical results show that the multi-feature setting provides a daily return, prior to transaction costs, of 0.64% using LSTM networks, and 0.54% using random forests. Hence we outperform the single-feature setting in Fischer & Krauss (2018) and Krauss et al. (2017) consisting only of the daily returns with respect to the closing prices, having corresponding daily returns of 0.41% and of 0.39% with respect to LSTM and random forests, respectively. | ['Jajati Keshari Sahoo', 'Ariel Neufeld', 'Pushpendu Ghosh'] | 2020-04-21 | null | null | null | null | ['stock-market-prediction'] | ['time-series'] | [-5.79204738e-01 2.85822153e-02 -5.60278535e-01 -3.42774838e-02
-5.13815224e-01 -9.00398016e-01 1.02102280e+00 -1.27841532e-01
-5.18400609e-01 1.20619833e+00 -1.78533923e-02 -7.38201082e-01
-2.03170523e-01 -1.29653776e+00 -9.39523876e-01 -4.63505238e-01
-4.20119107e-01 3.12547684e-01 -7.85648748e-02 -6.19322294e-03
2.25898206e-01 2.92612046e-01 -9.81453180e-01 8.47959593e-02
5.41078568e-01 2.02837753e+00 -1.59663752e-01 1.26474157e-01
-3.86852259e-03 8.02774906e-01 -2.49273136e-01 -7.75878549e-01
9.43748057e-01 2.23932728e-01 -1.57535911e-01 -2.16534615e-01
1.17796414e-01 -8.25659513e-01 -2.90525883e-01 8.60749245e-01
-8.64695981e-02 1.64272517e-01 5.25955498e-01 -9.55733895e-01
-7.78476000e-01 1.12455606e+00 -5.55203259e-01 4.73681152e-01
-4.28680956e-01 6.60930127e-02 1.51139677e+00 -8.51352990e-01
3.61405075e-01 6.76951408e-01 5.26326835e-01 3.66393887e-02
-1.11624646e+00 -8.10051203e-01 2.38428429e-01 -1.08765557e-01
-4.91558164e-01 -1.86994955e-01 4.18080539e-01 -5.26210606e-01
1.01743233e+00 5.51423579e-02 8.65192294e-01 1.00346184e+00
7.69242227e-01 5.89312315e-01 1.27138948e+00 -1.33634672e-01
-1.09442703e-01 -2.12328658e-02 2.45593965e-01 1.14461116e-01
4.82631683e-01 5.62348723e-01 -2.76826739e-01 -2.80606687e-01
1.15194452e+00 5.09989262e-01 1.14003822e-01 1.17614612e-01
-1.22027540e+00 1.18846810e+00 7.47305453e-01 3.23577166e-01
-9.38951850e-01 1.72615677e-01 3.45126063e-01 6.87301040e-01
6.26607895e-01 3.05155963e-01 -9.49834347e-01 3.15862475e-03
-1.22591913e+00 5.94064534e-01 9.34326530e-01 6.02180779e-01
6.25068247e-01 5.08743405e-01 -2.71317184e-01 4.99936730e-01
-2.65435074e-02 6.99172020e-01 7.56382942e-01 -8.29434514e-01
7.46603668e-01 1.81795552e-01 4.76803601e-01 -7.30419278e-01
-1.94469243e-01 -9.42536950e-01 -8.39649141e-01 1.71137955e-02
7.06291139e-01 -3.51561308e-01 -8.29363465e-01 1.69636357e+00
-2.50723004e-01 -2.32272921e-03 4.01548743e-02 5.08163035e-01
-4.39812504e-02 6.89198732e-01 -6.81940988e-02 -4.75769460e-01
1.25884545e+00 -1.09298253e+00 -2.03846216e-01 -1.29241288e-01
2.32188940e-01 -5.29460013e-01 5.85384548e-01 4.56709206e-01
-1.09485781e+00 -2.52112091e-01 -6.34404302e-01 6.96771085e-01
-3.24783266e-01 2.69313931e-01 6.09367371e-01 -6.07631728e-02
-7.50209451e-01 1.10489511e+00 -7.53160000e-01 5.31920671e-01
1.52929023e-01 1.42200559e-01 2.38077730e-01 6.83211207e-01
-1.58517528e+00 1.08353436e+00 1.05991140e-01 3.45369786e-01
-7.59218633e-01 -7.18524158e-01 -2.99253941e-01 1.41743630e-01
3.92873764e-01 -4.22824085e-01 1.38836539e+00 -8.84170651e-01
-1.31507885e+00 1.51279032e-01 4.26544070e-01 -1.75450647e+00
8.82691979e-01 -5.88206053e-01 -3.93518269e-01 -3.03231061e-01
2.17122316e-01 2.94391304e-01 5.66837728e-01 -6.44583046e-01
-1.09767354e+00 -4.11547750e-01 1.42842010e-01 -1.65316507e-01
-3.59959640e-02 -2.12949231e-01 5.29101849e-01 -1.16743982e+00
-3.01781837e-02 -9.35523570e-01 -4.41460870e-02 -7.17068195e-01
-6.02536738e-01 6.73594102e-02 3.52935731e-01 -1.09299147e+00
1.20196676e+00 -1.60139859e+00 -5.75444281e-01 4.22279835e-01
1.04454093e-01 -1.95328847e-01 3.38751078e-01 2.39953756e-01
-1.75748080e-01 -1.21187363e-02 -2.88665056e-01 4.33093533e-02
2.50820518e-01 1.22292805e-02 -1.11754358e+00 2.54734069e-01
1.82140514e-01 1.20678484e+00 -3.01684350e-01 3.70955050e-01
1.29392445e-01 -1.53776780e-01 -2.50764638e-02 -8.94028991e-02
-2.10344300e-01 -7.96758085e-02 -1.68097585e-01 8.05389881e-01
5.27675927e-01 1.57344952e-01 -3.13930325e-02 2.38107309e-01
-6.78163409e-01 8.50007534e-01 -8.17811668e-01 4.23047513e-01
-5.36095679e-01 4.52248931e-01 -3.30940224e-02 -8.34922016e-01
9.68340337e-01 1.57052070e-01 1.49389729e-01 -7.74884701e-01
-2.34626114e-01 6.85443163e-01 2.14664683e-01 4.13803518e-01
3.89126509e-01 -5.26310802e-01 -1.62864119e-01 7.36459911e-01
-1.85385332e-01 3.39760870e-01 2.21434951e-01 -3.11991245e-01
7.00650215e-01 9.14098620e-02 2.55908221e-01 -4.47750181e-01
-8.87509137e-02 -3.39888126e-01 6.50120080e-01 7.21988678e-01
1.99245334e-01 4.01990637e-02 8.59122396e-01 -5.46604812e-01
-9.43128943e-01 -1.24139297e+00 -2.75333673e-01 1.02785361e+00
-8.91891420e-01 5.74236512e-01 -3.51679981e-01 -6.59079075e-01
7.60622144e-01 9.20146346e-01 -8.81378293e-01 4.16991949e-01
-3.92016381e-01 -8.89985919e-01 3.31485003e-01 5.84140837e-01
7.80456185e-01 -1.26035452e+00 -9.26498771e-01 4.10521030e-01
3.04974884e-01 -5.88291824e-01 -4.28149730e-01 7.67466903e-01
-1.03566957e+00 -7.15345204e-01 -9.18804944e-01 -2.53144711e-01
1.29072428e-01 -2.87939072e-01 1.10433733e+00 -6.39484048e-01
5.90297043e-01 -2.30026782e-01 3.18894349e-02 -5.85714757e-01
2.40427762e-01 1.33407459e-01 -1.47388145e-01 2.12739155e-01
1.16147041e-01 -6.00934803e-01 -5.94730556e-01 1.79122120e-01
-6.71323061e-01 -2.26961836e-01 6.45507336e-01 1.06079352e+00
5.60868382e-01 -1.23781607e-01 8.77576172e-01 -8.38816106e-01
3.13213289e-01 -6.02542758e-01 -1.01443386e+00 2.09310189e-01
-1.08868337e+00 2.76340842e-02 5.83618164e-01 -4.84816432e-01
-9.27710772e-01 -4.34331566e-01 3.61392736e-01 -4.04602945e-01
2.55446166e-01 8.19601119e-01 3.92974257e-01 2.79319048e-01
-2.06726521e-01 4.11698580e-01 -1.78299248e-01 -6.01401865e-01
-2.19480204e-03 3.48891258e-01 4.54759389e-01 -2.68852353e-01
7.43199170e-01 2.88110375e-01 -2.43252814e-01 -1.15639053e-01
-9.73604321e-01 2.36328676e-01 -3.89820248e-01 4.49501686e-02
3.03209722e-01 -1.09037721e+00 -4.52456087e-01 7.78302312e-01
-7.52926230e-01 -4.80541438e-01 -5.75293958e-01 8.03591669e-01
-5.52594781e-01 -2.20114887e-01 -1.22167826e+00 -1.10939240e+00
-6.12195909e-01 -7.74816096e-01 5.48922777e-01 -5.19102514e-02
-2.26344049e-01 -1.02215052e+00 -1.07377693e-01 -4.86905798e-02
6.60330355e-01 6.53926373e-01 1.02677810e+00 -1.18285084e+00
-4.72112328e-01 -3.30149800e-01 -2.23771259e-01 5.41022718e-01
2.19273657e-01 1.87586963e-01 -7.07310140e-01 -1.39844313e-01
4.08811063e-01 -3.27070989e-02 1.28015566e+00 9.17004585e-01
5.54078698e-01 -8.28013718e-01 2.73734868e-01 3.48201871e-01
1.37338996e+00 5.59765458e-01 3.48039210e-01 1.02115762e+00
1.94609269e-01 8.16857994e-01 5.14412940e-01 6.69462860e-01
3.55182081e-01 3.30187112e-01 4.77817774e-01 3.44258636e-01
5.76732755e-01 -5.46779931e-01 5.91189563e-01 6.57476962e-01
-2.30062440e-01 3.29553485e-01 -6.60133719e-01 6.30568147e-01
-1.70645905e+00 -1.17042553e+00 -3.63845825e-02 2.50283337e+00
7.80125916e-01 7.10434318e-01 6.18981421e-01 -1.02695180e-02
6.49819314e-01 5.38082600e-01 -6.18949115e-01 -3.47225815e-01
-3.39354932e-01 2.23543614e-01 1.17612994e+00 3.16765606e-01
-1.08577406e+00 5.79846382e-01 5.17816544e+00 6.30706608e-01
-1.25618851e+00 -1.65255815e-01 1.18407440e+00 -4.34466004e-01
-4.77138311e-01 1.05609730e-01 -1.06188011e+00 1.02049208e+00
1.34415007e+00 -2.88143128e-01 6.40946329e-01 8.79498720e-01
2.50518285e-02 -1.33632019e-01 -7.66139150e-01 2.83137739e-01
-6.40113771e-01 -1.41585660e+00 9.19213817e-02 6.30184293e-01
7.23149598e-01 2.94443637e-01 5.60176134e-01 4.42753434e-01
4.23393130e-01 -8.98857951e-01 1.32510257e+00 7.12147474e-01
3.38566095e-01 -1.04468429e+00 9.77300167e-01 3.29711378e-01
-1.08702028e+00 -3.99052382e-01 -2.84134090e-01 -4.36585546e-01
7.14282095e-02 1.09538352e+00 -4.18163151e-01 4.95470285e-01
8.95843625e-01 6.36323214e-01 9.97093171e-02 4.54489231e-01
-1.53264299e-01 7.47225523e-01 -6.01541221e-01 3.19343805e-02
8.47743750e-01 -6.22627437e-01 6.64507523e-02 6.88676476e-01
7.19877481e-01 -4.50098008e-01 -2.76966184e-01 8.10287714e-01
-4.06744063e-01 -1.68320909e-01 -6.92267358e-01 -8.54585022e-02
4.27519798e-01 7.90847957e-01 -3.72940838e-01 -3.24736267e-01
-5.49512506e-01 1.63298503e-01 5.00350408e-02 4.00848210e-01
-7.22432375e-01 -3.29339296e-01 4.85169947e-01 4.03516322e-01
7.22733855e-01 -1.61258355e-01 -5.60690999e-01 -1.08141875e+00
3.72488737e-01 -5.39724469e-01 2.46711612e-01 -1.47115782e-01
-1.35054982e+00 3.40712875e-01 -1.21308245e-01 -1.35672212e+00
-6.71576083e-01 -6.35089755e-01 -9.03385818e-01 1.02122962e+00
-1.98894370e+00 -6.96996510e-01 6.94501877e-01 3.00639600e-01
4.62182820e-01 -3.17348868e-01 2.59697527e-01 -2.63009518e-01
-5.21218479e-01 2.62327969e-01 7.65643239e-01 4.09990132e-01
2.30087847e-01 -1.43558848e+00 8.37726355e-01 5.25764585e-01
3.41428071e-01 7.91411877e-01 2.42526412e-01 -7.48489916e-01
-8.17378819e-01 -1.20934606e+00 1.01455712e+00 -1.00367010e-01
1.37158334e+00 -9.78890881e-02 -7.72151291e-01 1.32219160e+00
3.61178070e-01 -1.61357611e-01 3.70078593e-01 -9.41462442e-02
-2.15602428e-01 -4.31495428e-01 -1.17128837e+00 3.12069923e-01
1.99178919e-01 -3.50572586e-01 -7.85918951e-01 7.88247064e-02
6.19695723e-01 -6.06855415e-02 -1.34719694e+00 4.71726120e-01
9.68220413e-01 -1.24058759e+00 6.68026924e-01 -4.43469614e-01
4.09717530e-01 2.93945372e-01 -2.77264446e-01 -1.44506645e+00
-2.20925435e-01 -4.37727392e-01 -2.06378371e-01 1.23913860e+00
8.51379216e-01 -1.43232715e+00 5.13161302e-01 3.84052187e-01
2.82504469e-01 -1.12741613e+00 -1.37178695e+00 -9.01621640e-01
6.01803482e-01 -3.56360883e-01 8.89931202e-01 6.73447967e-01
-5.09820879e-01 -9.49072167e-02 -4.41032678e-01 -4.27874178e-01
6.90816104e-01 8.96209896e-01 3.88796091e-01 -1.32382894e+00
-3.03638130e-01 -6.87147975e-01 3.64849299e-01 -7.58906782e-01
2.33921409e-01 -5.63537478e-01 -4.94715035e-01 -1.02322388e+00
2.51088687e-03 -1.47801772e-01 -8.43440294e-01 4.05727953e-01
4.47553754e-01 -5.22645526e-02 2.97935963e-01 4.90633935e-01
2.24501416e-01 3.70617300e-01 8.98878872e-01 -1.45523967e-02
-1.09671652e-01 5.98586082e-01 -6.35945022e-01 7.90476263e-01
8.72988522e-01 -1.11417577e-01 2.50320494e-01 -1.20509394e-01
1.68743625e-01 3.30085278e-01 5.89220107e-01 -4.69961196e-01
-1.90412045e-01 -2.99172163e-01 7.60052025e-01 -8.23005497e-01
-5.40043078e-02 -4.70069051e-01 1.65394023e-01 8.76521587e-01
-4.89813864e-01 5.45476377e-01 -2.24082209e-02 5.70242941e-01
-4.63947922e-01 -6.51051998e-02 4.94065642e-01 -3.66873235e-01
-1.05563655e-01 1.25118837e-01 -2.54694760e-01 -1.54475551e-02
1.05497289e+00 -2.42683794e-02 -2.88146317e-01 -7.19449997e-01
-9.61588562e-01 3.54771346e-01 2.22857162e-01 5.90105802e-02
3.04359078e-01 -1.33175516e+00 -7.15627909e-01 1.46187514e-01
-4.62089688e-01 -3.74431849e-01 -9.53564495e-02 9.88245845e-01
-1.28456235e-01 1.05565643e+00 -2.21814096e-01 1.17928959e-01
-1.71469688e-01 1.06787696e-01 3.97199541e-01 -9.73107040e-01
-1.88018575e-01 7.40316510e-01 2.36313537e-01 -2.50946879e-01
-5.25105298e-02 -8.39968801e-01 9.96624157e-02 8.75634372e-01
6.60351440e-02 6.08582795e-01 1.50849417e-01 -3.18780720e-01
-1.64706230e-01 2.78146029e-01 -2.55659167e-02 -2.81048238e-01
1.83027232e+00 1.06586434e-01 -1.33798540e-01 9.48390543e-01
8.75126958e-01 -2.71080844e-02 -1.57402575e+00 -3.04389060e-01
5.06023645e-01 -2.26163998e-01 -4.90385406e-02 -9.50540304e-01
-1.50071347e+00 7.09453583e-01 -1.96006405e-03 8.24838758e-01
7.99698591e-01 -3.89639199e-01 1.16628492e+00 1.60980165e-01
6.54121637e-01 -1.13578022e+00 -6.98789358e-01 4.98653322e-01
9.05293047e-01 -8.01474392e-01 -6.97717369e-02 5.02450109e-01
-8.69528174e-01 9.69458759e-01 -2.61231959e-02 -8.23968589e-01
8.37330163e-01 1.54387891e-01 -1.75310776e-01 2.35606641e-01
-1.11688423e+00 1.30622178e-01 1.62575155e-01 -1.30956704e-02
2.40656510e-01 5.72260082e-01 -2.81678408e-01 9.74282920e-01
-6.87478840e-01 -3.71600300e-01 4.43640053e-01 6.37769580e-01
-3.07178825e-01 -7.00189412e-01 -3.42260271e-01 1.38892341e+00
-1.11106598e+00 -4.44520056e-01 -2.87700832e-01 1.10069072e+00
-3.45070809e-01 4.19751793e-01 6.06220782e-01 -2.40026981e-01
4.49544042e-01 3.97075504e-01 -1.02304667e-01 -4.38049674e-01
-9.97649908e-01 4.86945212e-01 1.17527142e-01 -1.86835855e-01
-1.01374098e-04 -1.08592737e+00 -9.06378746e-01 -5.53028166e-01
-5.89395344e-01 2.64889687e-01 2.05062971e-01 8.74057531e-01
4.22852188e-02 5.17997518e-02 1.18197489e+00 -8.56785476e-01
-1.51796257e+00 -1.33503497e+00 -1.26781213e+00 -2.64443576e-01
6.39561355e-01 -6.05234206e-01 -8.30553651e-01 -3.77480626e-01] | [4.668651580810547, 4.038506984710693] |
5ff54e8c-8f15-408a-b351-9777920600f7 | timedial-temporal-commonsense-reasoning-in | 2106.04571 | null | https://arxiv.org/abs/2106.04571v1 | https://arxiv.org/pdf/2106.04571v1.pdf | TIMEDIAL: Temporal Commonsense Reasoning in Dialog | Everyday conversations require understanding everyday events, which in turn, requires understanding temporal commonsense concepts interwoven with those events. Despite recent progress with massive pre-trained language models (LMs) such as T5 and GPT-3, their capability of temporal reasoning in dialogs remains largely under-explored. In this paper, we present the first study to investigate pre-trained LMs for their temporal reasoning capabilities in dialogs by introducing a new task and a crowd-sourced English challenge set, TIMEDIAL. We formulate TIME-DIAL as a multiple-choice cloze task with over 1.1K carefully curated dialogs. Empirical results demonstrate that even the best performing models struggle on this task compared to humans, with 23 absolute points of gap in accuracy. Furthermore, our analysis reveals that the models fail to reason about dialog context correctly; instead, they rely on shallow cues based on existing temporal patterns in context, motivating future research for modeling temporal concepts in text and robust contextual reasoning about them. The dataset is publicly available at: https://github.com/google-research-datasets/timedial. | ['Manaal Faruqui', 'Yejin Choi', 'Luheng He', 'Shyam Upadhyay', 'Aditya Gupta', 'Lianhui Qin'] | 2021-06-08 | null | https://aclanthology.org/2021.acl-long.549 | https://aclanthology.org/2021.acl-long.549.pdf | acl-2021-5 | ['timedial'] | ['natural-language-processing'] | [-2.58718014e-01 2.17854545e-01 -2.50036925e-01 -5.09674788e-01
-6.85485780e-01 -9.86044645e-01 1.17004263e+00 1.23333961e-01
-3.82997185e-01 6.60764337e-01 8.43846381e-01 -5.41049659e-01
1.57034937e-02 -2.67627776e-01 -1.29300794e-02 -6.22573197e-02
-8.84501711e-02 7.81150043e-01 1.96012378e-01 -6.90936446e-01
1.71461344e-01 -6.75189570e-02 -8.84545505e-01 5.65845430e-01
7.49679983e-01 7.46093392e-01 -3.83018106e-02 8.24345887e-01
-3.20858389e-01 1.62077248e+00 -5.05302489e-01 -6.62090242e-01
-9.16981250e-02 -5.06694555e-01 -1.44127417e+00 -1.93449154e-01
1.00946680e-01 -4.64778334e-01 -5.77128232e-01 4.19949055e-01
7.16980398e-02 5.09176850e-01 4.45084333e-01 -1.44086874e+00
-7.58605421e-01 8.67731869e-01 2.42295086e-01 4.37542617e-01
7.99388766e-01 5.54809630e-01 1.44587302e+00 -6.49703562e-01
7.86056995e-01 1.45623481e+00 8.39753926e-01 8.63983810e-01
-1.10835707e+00 -2.89589673e-01 3.95208180e-01 3.01846623e-01
-8.27618659e-01 -6.72080457e-01 6.50409698e-01 -5.22713780e-01
1.32061863e+00 3.46359491e-01 4.21354592e-01 1.64555252e+00
5.28332451e-03 8.57971609e-01 1.42983127e+00 -1.57243609e-01
7.61153102e-02 -7.69194365e-02 4.76783693e-01 5.75045645e-01
-7.44859397e-01 3.09477933e-02 -8.83608341e-01 -2.49003261e-01
3.64888340e-01 -8.24591964e-02 -1.72146916e-01 3.31151396e-01
-1.43616164e+00 8.37248504e-01 2.95084178e-01 6.03828311e-01
-2.10051000e-01 1.27652278e-02 5.02476275e-01 4.41087663e-01
4.64788407e-01 6.22855723e-01 -7.56719351e-01 -9.17220294e-01
-6.23238266e-01 4.58240777e-01 1.29629087e+00 8.85267556e-01
1.45383328e-01 -4.46607500e-01 -4.08396274e-01 8.46673131e-01
1.19068801e-01 2.35002890e-01 4.63823408e-01 -1.35938597e+00
7.39637136e-01 6.15991294e-01 4.19365406e-01 -7.38238394e-01
-6.45815194e-01 3.70707124e-01 -3.09088916e-01 -5.11115909e-01
1.14833653e+00 -1.44057319e-01 -4.48833108e-01 2.18720984e+00
1.39557391e-01 1.14296012e-01 2.53890812e-01 6.47297382e-01
7.56802619e-01 6.12198353e-01 3.72657984e-01 -2.08011582e-01
1.62599719e+00 -1.06936085e+00 -1.00380957e+00 -5.11050403e-01
7.47241020e-01 -6.55586898e-01 1.58744991e+00 2.71343887e-01
-1.07398582e+00 -1.83051020e-01 -6.26811922e-01 -4.69640791e-01
-5.41650355e-01 -1.59479469e-01 9.17242348e-01 1.94336772e-01
-9.88469183e-01 5.20018280e-01 -7.65716136e-01 -8.37977052e-01
-3.29514630e-02 -2.28557587e-01 1.24386081e-03 1.30905733e-01
-1.70332813e+00 1.32564831e+00 3.04602887e-02 1.29031569e-01
-9.42891836e-01 -7.66628206e-01 -8.01827490e-01 -2.00714633e-01
6.10924125e-01 -2.75533617e-01 2.18197012e+00 -1.34257615e-01
-1.70407069e+00 9.82994139e-01 -3.38661104e-01 -5.68553865e-01
7.36971855e-01 -3.98569882e-01 -5.14625013e-01 5.66887185e-02
3.77177745e-01 5.41719198e-01 2.61245519e-01 -8.58977497e-01
-5.59892952e-01 2.88791314e-04 5.86786091e-01 7.72164464e-02
-1.13493972e-01 3.38218302e-01 -2.00550124e-01 -5.14344215e-01
2.87600569e-02 -1.16604376e+00 -8.92425701e-02 -1.55726850e-01
-3.84243995e-01 -6.98849499e-01 6.48704588e-01 -7.89078712e-01
1.41323566e+00 -1.90326726e+00 1.11973897e-01 -4.96348292e-01
2.99326718e-01 -2.43396848e-01 2.47726083e-01 1.03558707e+00
2.61745661e-01 2.03625992e-01 -3.37794930e-01 -7.52887368e-01
4.54951853e-01 3.35745662e-01 -8.18326533e-01 1.29134431e-01
1.45330742e-01 1.07829595e+00 -1.20438457e+00 -5.24302840e-01
3.38392168e-01 -1.54684884e-02 -4.02637988e-01 3.56757849e-01
-9.66922939e-01 6.96075678e-01 -2.96959758e-01 4.37938690e-01
5.89376502e-03 -4.87718225e-01 3.66535783e-01 1.69762835e-01
-1.45361781e-01 1.06747937e+00 -4.07599270e-01 1.88858259e+00
-6.13305211e-01 8.83889854e-01 -7.22116679e-02 -2.67979294e-01
4.52308327e-01 6.75158978e-01 2.74978936e-01 -6.67437613e-01
1.60512388e-01 1.28927520e-02 -4.72741062e-03 -6.83639765e-01
6.26794219e-01 -2.70272702e-01 -5.36917806e-01 7.74038434e-01
-5.45269139e-02 -3.57419193e-01 2.35058099e-01 6.64507687e-01
1.07676411e+00 -1.37796044e-01 1.44595161e-01 -1.18225381e-01
3.63683879e-01 3.20932508e-01 4.48794991e-01 6.63139999e-01
-7.67998517e-01 1.72624022e-01 7.68345594e-01 -4.96824086e-01
-7.08433986e-01 -8.47209215e-01 8.36236700e-02 1.50886190e+00
8.03190395e-02 -8.15999627e-01 -5.02287686e-01 -6.50521100e-01
-1.57075897e-01 1.26985919e+00 -6.43527389e-01 1.96402520e-01
-6.86911881e-01 -3.83350909e-01 9.37535584e-01 6.13619506e-01
5.77785552e-01 -1.13592017e+00 -6.78500116e-01 3.16716373e-01
-8.06321204e-01 -1.71344745e+00 -5.55887699e-01 6.10552169e-02
-6.69727564e-01 -1.05251634e+00 -2.97601879e-01 -4.28938001e-01
-1.60189226e-01 8.78729522e-02 1.40502048e+00 5.25313579e-02
6.43709525e-02 6.45983338e-01 -4.16688412e-01 -2.26445660e-01
-5.07780671e-01 -1.28456056e-01 -1.10615203e-02 -5.09977460e-01
6.97339714e-01 -6.10744119e-01 -4.35297281e-01 4.36782092e-01
-3.49184334e-01 1.12716667e-01 -1.82800844e-01 7.89035618e-01
-4.46167856e-01 -2.39299685e-01 4.32428718e-01 -6.53883934e-01
1.08760667e+00 -7.25221157e-01 -3.02480996e-01 2.92160273e-01
-4.81454790e-01 5.73005266e-02 4.20147210e-01 -5.34132600e-01
-1.49866343e+00 -5.65604150e-01 1.96016952e-01 -9.45741385e-02
-1.48343667e-01 5.57809889e-01 4.47024375e-01 6.52107716e-01
7.50530481e-01 4.17602025e-02 -1.69944227e-01 -2.62307853e-01
7.16792345e-01 3.29773247e-01 7.75918663e-01 -1.20953774e+00
5.10100305e-01 4.77987051e-01 -5.17320573e-01 -7.09287882e-01
-1.19777286e+00 -4.11878049e-01 -5.96363544e-01 -3.75842392e-01
1.23406875e+00 -9.65368032e-01 -1.05996549e+00 4.32809293e-01
-1.44553244e+00 -1.12259412e+00 1.43493086e-01 3.18774790e-01
-6.04763150e-01 3.26229900e-01 -1.30232000e+00 -1.05715847e+00
-1.05022959e-01 -8.31496239e-01 8.75210941e-01 5.57345226e-02
-1.00237560e+00 -1.38418031e+00 7.96652585e-02 8.07126880e-01
5.08950651e-01 1.25764668e-01 9.00110483e-01 -8.59573364e-01
-5.00783324e-01 1.57984555e-01 2.93385182e-02 -1.97537914e-01
4.81995083e-02 -1.26610950e-01 -1.18341839e+00 1.12675257e-01
1.94440335e-01 -8.25475216e-01 4.87135231e-01 -3.87548283e-02
5.89682758e-01 -4.55339491e-01 -2.23876730e-01 -8.16940516e-02
6.02686703e-01 3.65113318e-01 1.81502238e-01 2.09891871e-01
2.96718061e-01 8.98585618e-01 8.71403754e-01 5.97133458e-01
1.11011732e+00 6.18806362e-01 6.67568520e-02 4.46296364e-01
6.82575628e-02 -3.67026150e-01 4.17523235e-01 7.69178450e-01
1.64825141e-01 -3.89534920e-01 -1.46102071e+00 8.12440813e-01
-2.10422468e+00 -1.12680459e+00 -1.44901007e-01 1.54410136e+00
1.33348429e+00 2.61737257e-01 1.41654164e-01 -2.64433622e-01
3.92502040e-01 5.19682646e-01 -4.51147288e-01 -3.26373041e-01
-5.92668764e-02 -1.80944458e-01 -1.70250356e-01 9.90077794e-01
-9.90330637e-01 1.33571827e+00 6.01881361e+00 3.38499635e-01
-8.48064542e-01 2.90289223e-01 5.85016847e-01 -1.94718130e-03
-2.33618379e-01 3.40434492e-01 -5.43783128e-01 3.13662708e-01
1.16460836e+00 -3.30857605e-01 6.98283076e-01 6.73115849e-01
3.10048789e-01 -1.67325482e-01 -1.54122663e+00 7.88749039e-01
-2.42968306e-01 -1.08640373e+00 -5.50893188e-01 -3.32708538e-01
5.02038956e-01 -1.18701033e-01 7.51094427e-03 8.52548838e-01
8.97263110e-01 -1.10974634e+00 8.23709249e-01 3.78035098e-01
4.21676278e-01 -2.05612406e-02 3.02085698e-01 5.16770303e-01
-1.09969449e+00 -1.96391597e-01 2.46288046e-01 -4.97274071e-01
4.96326119e-01 5.95698357e-02 -1.12189400e+00 1.20841749e-01
6.56783700e-01 6.76226437e-01 -4.29813862e-01 1.19729392e-01
-3.76961559e-01 7.17633545e-01 -2.66289115e-01 -1.57789275e-01
4.86373305e-01 -4.06589992e-02 5.60390711e-01 1.29052842e+00
-8.46593082e-02 6.65649533e-01 2.80045509e-01 1.03088176e+00
9.72168241e-03 -4.00472552e-01 -5.58718204e-01 -5.14722466e-01
8.56371403e-01 9.74029005e-01 -6.02569342e-01 -4.62107331e-01
-3.77150148e-01 1.00605059e+00 4.56071258e-01 3.24095130e-01
-1.05593312e+00 2.96841264e-01 7.56259978e-01 -2.69897521e-01
-2.43212342e-01 -6.95879936e-01 -3.81173879e-01 -1.16516066e+00
7.27967098e-02 -9.64922726e-01 7.69587934e-01 -1.08620751e+00
-1.69893324e+00 6.06611550e-01 3.08814347e-01 -7.75943518e-01
-5.58945179e-01 -6.24486685e-01 -8.20631087e-01 8.08043659e-01
-1.16760790e+00 -1.16781580e+00 -3.84915978e-01 7.24077582e-01
9.65355575e-01 1.25194535e-01 8.42654705e-01 -8.94618481e-02
-3.94893169e-01 2.86226749e-01 -5.14811277e-01 2.24826574e-01
1.10191476e+00 -1.46496499e+00 5.42530656e-01 6.27395391e-01
-9.33858231e-02 1.12600279e+00 1.03252375e+00 -6.68195367e-01
-1.22149658e+00 -6.22128546e-01 1.56691980e+00 -1.31577325e+00
1.30388546e+00 -5.14321566e-01 -8.88350487e-01 1.10119510e+00
8.32165956e-01 -4.47179079e-01 5.99654913e-01 5.69538295e-01
-7.39681602e-01 5.00434875e-01 -1.01306951e+00 1.08052588e+00
1.34008062e+00 -1.24214160e+00 -1.19250906e+00 6.77079976e-01
1.15985644e+00 -7.09192753e-01 -9.91942883e-01 1.54187366e-01
4.07338560e-01 -8.70303392e-01 8.18068922e-01 -8.97893667e-01
6.76597476e-01 9.68155116e-02 -3.22372109e-01 -1.16966915e+00
9.75429639e-02 -1.01394629e+00 -2.32179180e-01 1.38609314e+00
4.69905913e-01 -4.75964248e-01 4.52869743e-01 1.26774549e+00
-8.27096552e-02 -4.23409522e-01 -7.78066754e-01 -7.22550631e-01
2.18889341e-01 -7.02752769e-01 3.05685014e-01 1.45393324e+00
7.61392713e-01 6.59316242e-01 -3.86420608e-01 -2.49831043e-02
7.11562485e-02 5.27677499e-02 6.02728009e-01 -1.13960159e+00
-1.57766938e-01 -5.59757769e-01 3.92814249e-01 -1.20530295e+00
5.34515858e-01 -6.56927943e-01 1.26472920e-01 -1.41503775e+00
-1.01514935e-01 -5.33293009e-01 3.17502201e-01 6.99133217e-01
-3.48990232e-01 -4.08947200e-01 2.54578590e-01 2.45360926e-01
-7.88166106e-01 6.36862755e-01 9.58458900e-01 -5.50471582e-02
-3.71919870e-01 -1.80159211e-01 -5.69138825e-01 5.97462296e-01
1.03772199e+00 -1.00911766e-01 -7.64583111e-01 -4.12211388e-01
8.99684653e-02 6.24249995e-01 5.55689335e-01 -7.02030540e-01
5.50096095e-01 -6.32508636e-01 -3.18825901e-01 -5.67182362e-01
8.27709675e-01 -5.72533607e-01 -1.79432973e-01 2.41978824e-01
-9.11271870e-01 4.60328281e-01 4.06571090e-01 5.63893616e-01
-1.90339774e-01 1.43513024e-01 3.55152607e-01 -3.45915973e-01
-9.11811769e-01 -7.21765012e-02 -5.36077976e-01 5.75338602e-01
7.97321379e-01 2.04505593e-01 -8.25282574e-01 -9.55733299e-01
-8.66384447e-01 6.74246550e-01 1.66116387e-01 6.94534838e-01
1.53946057e-01 -1.11986005e+00 -5.20614028e-01 -6.31455481e-01
2.98237711e-01 -2.93229938e-01 4.13673609e-01 9.97721970e-01
-3.98076288e-02 7.32259452e-01 1.70482188e-01 -5.29248297e-01
-1.01683462e+00 5.99351108e-01 3.40266615e-01 -4.16663468e-01
-6.19816422e-01 9.82052326e-01 -1.49194691e-02 -6.81002676e-01
3.13180149e-01 -7.31500208e-01 1.05439266e-02 1.52222171e-01
4.12499219e-01 1.56133547e-01 -2.67825246e-01 -3.58425081e-01
-4.76325512e-01 6.86313584e-03 -3.63422148e-02 -7.55528152e-01
1.01896298e+00 -3.39952946e-01 -1.84109792e-01 1.01967669e+00
9.06102777e-01 -1.03892526e-02 -1.18156028e+00 -6.34156048e-01
6.02707863e-01 -1.43850431e-01 -5.30940533e-01 -1.18238509e+00
-1.75831214e-01 7.59304941e-01 -1.15678504e-01 4.82112259e-01
6.99782968e-01 2.03142226e-01 8.31789374e-01 6.24625146e-01
4.05918956e-01 -1.08396721e+00 5.46404183e-01 1.28449476e+00
1.09805644e+00 -1.42543495e+00 -4.06681061e-01 -3.53127718e-01
-1.18343604e+00 1.02194667e+00 9.13345098e-01 3.52366000e-01
5.20466983e-01 1.30482852e-01 4.04762000e-01 -5.10280013e-01
-1.54398012e+00 -1.53126732e-01 -3.97973880e-02 1.70801386e-01
7.74520397e-01 9.89642888e-02 -7.90996552e-02 7.85770595e-01
-5.92336416e-01 -1.57377347e-01 3.57417285e-01 9.74801779e-01
-3.60865667e-02 -8.03148746e-01 -1.33295432e-01 1.22315316e-02
-1.77256599e-01 -3.21748495e-01 -8.08979750e-01 7.83236802e-01
-4.37598258e-01 1.56430852e+00 -1.10799134e-01 -4.38288897e-01
2.25581691e-01 6.84496105e-01 1.28690884e-01 -4.61049199e-01
-7.90845275e-01 -5.70972025e-01 7.42948711e-01 -7.54434943e-01
-3.54259223e-01 -8.65165889e-01 -1.30096948e+00 -6.22780442e-01
2.55594820e-01 2.59829730e-01 1.87383115e-01 1.13050294e+00
1.33680299e-01 3.72337699e-01 2.27593437e-01 -4.50744033e-01
-7.11130023e-01 -1.17294526e+00 8.41220990e-02 6.21098816e-01
3.53246659e-01 -6.33251548e-01 -5.37471533e-01 9.54391733e-02] | [12.649280548095703, 8.01131534576416] |
fe8e379f-1c73-46af-ab98-360b45fbc792 | schema-encoding-for-transferable-dialogue-1 | 2210.02351 | null | https://arxiv.org/abs/2210.02351v1 | https://arxiv.org/pdf/2210.02351v1.pdf | Schema Encoding for Transferable Dialogue State Tracking | Dialogue state tracking (DST) is an essential sub-task for task-oriented dialogue systems. Recent work has focused on deep neural models for DST. However, the neural models require a large dataset for training. Furthermore, applying them to another domain needs a new dataset because the neural models are generally trained to imitate the given dataset. In this paper, we propose Schema Encoding for Transferable Dialogue State Tracking (SETDST), which is a neural DST method for effective transfer to new domains. Transferable DST could assist developments of dialogue systems even with few dataset on target domains. We use a schema encoder not just to imitate the dataset but to comprehend the schema of the dataset. We aim to transfer the model to new domains by encoding new schemas and using them for DST on multi-domain settings. As a result, SET-DST improved the joint accuracy by 1.46 points on MultiWOZ 2.1. | ['Gary Geunbae Lee', 'Hyunmin Jeon'] | 2022-10-05 | null | https://aclanthology.org/2022.coling-1.28 | https://aclanthology.org/2022.coling-1.28.pdf | coling-2022-10 | ['dialogue-state-tracking', 'task-oriented-dialogue-systems'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.93420488e-01 6.63556635e-01 6.77247904e-03 -8.07454050e-01
-4.67199147e-01 -7.92332411e-01 6.30789697e-01 -1.50957450e-01
-4.26076382e-01 9.97267723e-01 8.14908743e-02 -2.00829431e-01
2.81338662e-01 -8.15284133e-01 -6.29546404e-01 -9.58294049e-02
2.50745118e-01 1.08039963e+00 4.24432516e-01 -1.05099392e+00
9.63467881e-02 2.04789937e-02 -8.31590831e-01 7.14532614e-01
1.17551577e+00 6.75375998e-01 4.85785455e-01 6.20799422e-01
-4.51879263e-01 6.62925661e-01 -1.07111967e+00 -4.88274276e-01
-4.02397513e-02 -8.89717698e-01 -1.40094304e+00 -6.63014725e-02
3.42420399e-01 -7.32536912e-01 -4.13982421e-01 9.62025940e-01
4.66451049e-01 2.76109397e-01 6.21135890e-01 -1.22705865e+00
-5.53498983e-01 7.80777812e-01 8.52945596e-02 -2.32754305e-01
6.78899825e-01 2.34506220e-01 6.00568950e-01 -5.34949839e-01
9.06245708e-01 1.35001326e+00 5.51452577e-01 1.55538452e+00
-9.46839333e-01 -5.30760229e-01 -1.22802675e-01 3.81999128e-02
-5.25877476e-01 -6.71191812e-01 8.00494611e-01 1.33600971e-03
1.06417704e+00 7.77356327e-02 4.72506136e-01 1.54117823e+00
-1.42268687e-01 1.17255652e+00 1.20136070e+00 -4.29647744e-01
1.57626867e-01 5.79694390e-01 1.48206010e-01 6.62486076e-01
-1.60518542e-01 -1.80268124e-01 -6.91118419e-01 2.45443836e-01
8.89314175e-01 -6.17380619e-01 -3.46755646e-02 -2.88268000e-01
-1.23533750e+00 1.01926327e+00 2.36453876e-01 1.92015037e-01
7.76710659e-02 -2.56712198e-01 1.06923699e+00 7.04616785e-01
4.28272843e-01 7.96883047e-01 -8.37441623e-01 -4.43100303e-01
-3.27823460e-01 2.56895572e-01 1.01958644e+00 1.18158913e+00
4.53323215e-01 2.10695982e-01 -4.31071341e-01 1.23363459e+00
4.31793369e-02 5.30513346e-01 7.35995591e-01 -9.35599744e-01
9.15070534e-01 8.33128870e-01 -1.45398095e-01 -3.80372256e-01
-4.65030074e-01 1.97658837e-01 -9.38769639e-01 -1.52797252e-01
6.80867970e-01 -5.86620390e-01 -7.01059997e-01 1.96493518e+00
3.45060974e-01 -1.82905748e-01 7.89510310e-01 6.51178122e-01
1.13190138e+00 9.68618512e-01 6.27273787e-03 2.32320353e-02
1.20532095e+00 -1.06787026e+00 -1.06248558e+00 -4.15411294e-01
1.05064511e+00 -2.46915519e-01 1.31890666e+00 2.46321023e-01
-1.10139358e+00 -6.21343851e-01 -9.28507507e-01 -2.88279325e-01
-5.98874748e-01 -2.80138105e-02 3.70301515e-01 6.30294085e-01
-1.06920302e+00 4.95157361e-01 -4.74448681e-01 -7.87494183e-01
2.58009553e-01 2.77804166e-01 -4.00430292e-01 1.32712394e-01
-1.85202694e+00 1.32220542e+00 8.94209325e-01 -1.65728658e-01
-9.24566329e-01 -2.49038801e-01 -1.15173686e+00 1.18658483e-01
3.41155469e-01 -5.16813397e-01 1.87987494e+00 -9.55155671e-01
-2.30364823e+00 8.74249101e-01 6.36746883e-02 -7.57879615e-01
4.11107183e-01 -1.17611490e-01 -3.07374328e-01 8.93193334e-02
4.52196561e-02 7.57232964e-01 5.72369695e-01 -9.77459610e-01
-4.51923311e-01 -1.62321106e-01 3.32457542e-01 4.53789175e-01
-4.94267106e-01 -2.34868914e-01 -6.92391023e-02 -3.85543168e-01
-2.46224627e-01 -9.39595997e-01 6.23476319e-02 -1.81604743e-01
-4.77084756e-01 -5.39884508e-01 1.04909885e+00 -7.32904732e-01
9.12951052e-01 -1.87358117e+00 3.29181761e-01 -4.75073665e-01
2.77573392e-02 8.74209046e-01 -3.01882088e-01 5.40027142e-01
1.81749001e-01 -2.72920907e-01 -3.62211049e-01 -1.24226332e-01
2.12385356e-01 4.12250847e-01 -2.66798079e-01 -2.54280299e-01
5.56812465e-01 9.20783877e-01 -1.01648521e+00 -2.12277502e-01
1.52385324e-01 -1.83572784e-01 -5.57290018e-01 7.80872107e-01
-6.16359830e-01 5.20984054e-01 -2.35932961e-01 -1.26563162e-01
6.18523180e-01 -1.34739488e-01 4.30503815e-01 -1.94073707e-01
1.54360056e-01 6.15716875e-01 -6.51497066e-01 2.46433568e+00
-4.88218635e-01 7.17364669e-01 -1.62408382e-01 -1.09763336e+00
1.40437293e+00 5.32142699e-01 7.87844062e-02 -8.49270523e-01
1.74301788e-01 9.51201916e-02 5.18810898e-02 -6.58357680e-01
7.21860886e-01 -3.89527321e-01 -6.10446692e-01 8.28944266e-01
6.99051380e-01 -4.55995500e-01 -2.91254781e-02 2.41682947e-01
7.60454297e-01 5.30849956e-02 1.84365928e-01 -4.35767733e-02
5.35349071e-01 3.43387395e-01 2.65762568e-01 5.42770445e-01
-3.49783570e-01 2.40191087e-01 5.48464417e-01 -3.83551091e-01
-1.06632674e+00 -6.70679629e-01 9.25963297e-02 1.33454871e+00
9.97106507e-02 -1.89426437e-01 -1.14045298e+00 -1.30661559e+00
-7.06801191e-02 9.80575144e-01 -5.95683455e-01 -7.01997876e-01
-5.04651666e-01 -2.38250181e-01 1.00467086e+00 3.77688676e-01
1.36003351e+00 -1.20628846e+00 -5.34743607e-01 3.85216296e-01
-3.58209044e-01 -1.13140464e+00 -4.34536129e-01 2.25149736e-01
-8.67207527e-01 -9.32752073e-01 -9.22251701e-01 -9.22526419e-01
3.20115179e-01 1.03549518e-01 8.49000931e-01 -3.27879697e-01
3.24924737e-01 2.89277732e-01 -4.14220363e-01 -2.59317338e-01
-1.15185463e+00 6.17984056e-01 1.06085598e-01 -1.84590563e-01
3.16652566e-01 1.10912971e-01 -3.23968045e-02 4.03929412e-01
-7.94653654e-01 5.55964530e-01 1.83937699e-01 1.26958704e+00
-4.07676280e-01 -3.59425157e-01 1.09296679e+00 -1.41218925e+00
1.32876039e+00 -4.34868813e-01 -3.93938392e-01 5.89342535e-01
-3.44371229e-01 4.32440192e-01 5.65160751e-01 -4.49270695e-01
-1.81109667e+00 -1.59622431e-01 -2.63294667e-01 -5.71477413e-02
-2.95428306e-01 5.79910517e-01 -2.44364813e-01 2.65774906e-01
8.46245229e-01 3.64754498e-01 4.17320132e-01 -4.32654500e-01
5.10255873e-01 1.12338674e+00 5.12082398e-01 -8.93723249e-01
3.60748917e-01 -2.18521982e-01 -4.98574406e-01 -5.69733083e-01
-7.79618144e-01 -7.76735228e-03 -9.81588066e-01 -1.19020157e-01
9.30165291e-01 -7.08003521e-01 -6.03492558e-01 8.45463991e-01
-1.57775915e+00 -1.07330787e+00 -2.55755365e-01 -4.45817411e-02
-7.09820092e-01 2.55594432e-01 -7.26483524e-01 -6.19021237e-01
-5.58854878e-01 -1.07636380e+00 9.44708705e-01 3.64347458e-01
-3.96254718e-01 -1.44376779e+00 4.17123556e-01 4.16414976e-01
5.94536245e-01 -8.17026943e-02 7.37116516e-01 -1.26766229e+00
9.38317180e-03 1.49246156e-01 -2.59689301e-01 5.18375099e-01
4.36550856e-01 -3.73288602e-01 -1.17833245e+00 -4.02438849e-01
8.99666846e-02 -1.02040982e+00 2.90585995e-01 -1.47109672e-01
6.78925335e-01 -4.75178421e-01 -4.01088409e-02 2.55155504e-01
7.75285363e-01 6.15190268e-01 4.17928308e-01 4.83561039e-01
3.84987921e-01 9.69222009e-01 9.72760022e-01 3.07847172e-01
7.57450104e-01 9.76703584e-01 9.37620818e-04 -4.81230207e-02
-2.38324165e-01 -4.56149012e-01 7.33552575e-01 9.72752035e-01
3.86924624e-01 -1.70874804e-01 -1.11107254e+00 3.64798903e-01
-1.84905851e+00 -6.79010153e-01 1.24913514e-01 1.74947929e+00
1.50654340e+00 1.31654635e-01 1.06424332e-01 -2.25511014e-01
7.93335974e-01 -5.68531230e-02 -9.85800326e-01 -7.94642210e-01
-3.30728805e-03 1.69193700e-01 -1.62188023e-01 3.24485004e-01
-8.99792790e-01 1.48175800e+00 5.87656641e+00 6.31905735e-01
-1.19952476e+00 9.97004062e-02 2.93998390e-01 4.92656678e-01
3.07464693e-02 -1.96020707e-01 -8.50677490e-01 3.84298772e-01
1.14998972e+00 -5.88958561e-01 2.37487674e-01 8.58449459e-01
-2.26766422e-01 -1.32901464e-02 -1.20116115e+00 6.03322566e-01
-2.59284233e-03 -1.11793530e+00 2.75088191e-01 -3.64070237e-01
4.95483756e-01 -3.22897404e-01 2.39355620e-02 9.18104053e-01
4.67880726e-01 -8.14376116e-01 2.74372637e-01 2.31669083e-01
1.21922684e+00 -4.84947920e-01 8.18751156e-01 4.87201482e-01
-6.13658011e-01 2.61336565e-01 -4.00225282e-01 4.11441550e-02
2.26793885e-02 -3.07040930e-01 -1.83265996e+00 3.73994797e-01
3.26947242e-01 8.90415311e-01 -5.74395657e-01 5.52783847e-01
-1.80379391e-01 4.45867419e-01 -9.20662880e-02 -2.71558225e-01
2.15518147e-01 -1.10528857e-01 2.06196636e-01 1.20571363e+00
1.98905140e-01 1.01462878e-01 -2.57892579e-01 8.76669049e-01
-1.40152469e-01 -1.66044325e-01 -1.11817420e+00 -1.51141226e-01
4.87514704e-01 9.51319277e-01 -4.94083613e-02 -6.84531510e-01
-1.85537830e-01 1.39210677e+00 4.72275406e-01 1.21710502e-01
-5.36392927e-01 -6.68960750e-01 6.00079834e-01 -3.77424538e-01
-2.11562216e-01 -1.07567891e-01 -1.75497070e-01 -1.45649028e+00
-4.16519076e-01 -1.13356555e+00 3.79497617e-01 -9.48790848e-01
-1.29935586e+00 8.56853843e-01 1.20456316e-01 -1.19822252e+00
-6.11105442e-01 -5.95433295e-01 -5.20924628e-01 8.66115630e-01
-1.35965872e+00 -1.16349709e+00 -2.62146145e-01 9.40546453e-01
1.20535886e+00 -6.49989069e-01 1.29516101e+00 -1.49330739e-02
-6.13756180e-01 9.23317373e-01 8.04924518e-02 5.90799093e-01
1.19436884e+00 -1.49295306e+00 8.74161065e-01 3.06335390e-01
-3.64688069e-01 6.87541187e-01 5.64498544e-01 -6.69588983e-01
-1.25483489e+00 -1.12280202e+00 6.30437076e-01 -6.68544173e-01
4.90055501e-01 -6.39753580e-01 -1.39801812e+00 8.11491430e-01
8.26682687e-01 -6.28744960e-01 6.67556703e-01 1.86785892e-01
-2.05473915e-01 8.98721963e-02 -1.44510984e+00 4.47448343e-01
9.35981214e-01 -7.18751967e-01 -1.10054004e+00 2.70935595e-01
9.63746667e-01 -8.49362552e-01 -1.13614845e+00 2.77271599e-01
3.05661678e-01 -7.70908475e-01 6.59927189e-01 -1.10023177e+00
3.59693438e-01 2.15358719e-01 1.85366705e-01 -1.83662736e+00
1.81887999e-01 -6.53339505e-01 -1.31256282e-01 1.34845555e+00
4.19258386e-01 -7.55025983e-01 6.33469701e-01 8.55080247e-01
-3.36259514e-01 -1.84086233e-01 -9.01843250e-01 -8.41266870e-01
4.11881983e-01 1.03742704e-01 8.13172340e-01 1.47858298e+00
6.94064200e-01 1.09301043e+00 -3.79427314e-01 -2.85792857e-01
2.76474029e-01 -3.87421958e-02 1.08540821e+00 -1.16868174e+00
1.20584536e-02 -1.82810813e-01 1.53303161e-01 -1.47989726e+00
3.37893337e-01 -8.58161271e-01 1.49334669e-01 -1.37037742e+00
-1.76999435e-01 -6.98122680e-01 2.14277133e-01 7.63521612e-01
-2.23776162e-01 -5.55648983e-01 3.34482938e-01 -5.65565489e-02
-7.34948814e-01 9.68533397e-01 1.72638798e+00 -4.29051518e-01
-2.72283733e-01 1.31377026e-01 -4.67666864e-01 2.29746819e-01
1.20315135e+00 -2.20177323e-01 -8.16824496e-01 -5.29251933e-01
-1.38720661e-01 6.59878135e-01 -4.23699394e-02 -9.22283173e-01
4.08154130e-01 -2.02723816e-01 2.38968637e-02 -3.33664089e-01
6.29281163e-01 -5.99964917e-01 -3.94848824e-01 4.97719020e-01
-7.58886874e-01 -2.50485957e-01 7.20571458e-01 2.29257420e-01
-3.69193763e-01 -6.22181654e-01 7.42122054e-01 -2.73165941e-01
-9.99497116e-01 -1.69052318e-01 -4.73256975e-01 2.37935737e-01
9.04027939e-01 -2.78942157e-02 -8.47528338e-01 -6.16086125e-01
-7.07507014e-01 4.36653644e-01 3.26017231e-01 6.03995621e-01
7.54683912e-01 -1.39899123e+00 -5.75682044e-01 2.60740310e-01
2.50334144e-01 2.77650714e-01 1.90533921e-01 2.37047270e-01
-4.39024478e-01 6.30378485e-01 -9.32353497e-01 -5.11062086e-01
-1.16455042e+00 2.52916545e-01 4.93646801e-01 -2.67102867e-01
-2.99613625e-01 7.61525929e-01 1.56519637e-01 -1.23037732e+00
7.34996721e-02 -3.34400594e-01 -4.80023474e-01 9.03591439e-02
2.64971882e-01 4.27592173e-02 -4.43316177e-02 -3.17575842e-01
2.14874938e-01 1.56492725e-01 -5.32573998e-01 -2.88008749e-01
1.30306852e+00 -1.01069093e-01 8.29254910e-02 5.98522305e-01
1.06615174e+00 -7.80758262e-01 -1.38146126e+00 -6.23047531e-01
1.84118256e-01 -1.07297271e-01 -4.69525546e-01 -1.20648336e+00
-6.69484377e-01 1.30253887e+00 4.73718017e-01 4.14545417e-01
8.28577578e-01 -2.58241594e-01 1.22838342e+00 9.03432369e-01
6.14705741e-01 -1.18067515e+00 5.30394971e-01 1.22954977e+00
9.92873669e-01 -1.43840170e+00 -5.10895073e-01 -9.86427143e-02
-1.31684649e+00 1.39333689e+00 1.19286346e+00 2.68361479e-01
4.02869135e-02 -2.14253604e-01 1.66623175e-01 -2.39025801e-01
-8.85643125e-01 -8.33481085e-03 3.84342819e-02 8.93887937e-01
2.51169682e-01 -1.82270929e-01 2.67735839e-01 7.84087181e-01
-1.91155002e-01 1.86543122e-01 6.88558698e-01 9.60639238e-01
-4.69549805e-01 -1.43976319e+00 7.70081058e-02 2.71084756e-01
3.44739318e-01 7.33910650e-02 -7.93542325e-01 6.75998151e-01
-4.14522141e-01 7.65449941e-01 -9.76944789e-02 -6.20663047e-01
7.33919144e-01 4.34734166e-01 5.35572171e-01 -1.09823024e+00
-8.39626074e-01 -5.25724769e-01 6.42790914e-01 -1.54356048e-01
-2.49718145e-01 -4.36376244e-01 -1.19926965e+00 -2.27471292e-01
-2.54216135e-01 4.51083213e-01 4.97588784e-01 8.17025363e-01
3.96046877e-01 4.76100743e-01 6.17240012e-01 -4.19519991e-01
-7.26682663e-01 -1.50232685e+00 -3.79577547e-01 4.79061335e-01
2.51013458e-01 -6.62027359e-01 1.04966648e-01 1.42365545e-01] | [12.81657600402832, 7.995871543884277] |
c5a629e0-8f67-41c0-ba3e-68e74b644380 | efficient-cnns-via-passive-filter-pruning | 2304.02319 | null | https://arxiv.org/abs/2304.02319v1 | https://arxiv.org/pdf/2304.02319v1.pdf | Efficient CNNs via Passive Filter Pruning | Convolutional neural networks (CNNs) have shown state-of-the-art performance in various applications. However, CNNs are resource-hungry due to their requirement of high computational complexity and memory storage. Recent efforts toward achieving computational efficiency in CNNs involve filter pruning methods that eliminate some of the filters in CNNs based on the \enquote{importance} of the filters. The majority of existing filter pruning methods are either "active", which use a dataset and generate feature maps to quantify filter importance, or "passive", which compute filter importance using entry-wise norm of the filters without involving data. Under a high pruning ratio where large number of filters are to be pruned from the network, the entry-wise norm methods eliminate relatively smaller norm filters without considering the significance of the filters in producing the node output, resulting in degradation in the performance. To address this, we present a passive filter pruning method where the filters are pruned based on their contribution in producing output by considering the operator norm of the filters. The proposed pruning method generalizes better across various CNNs compared to that of the entry-wise norm-based pruning methods. In comparison to the existing active filter pruning methods, the proposed pruning method is at least 4.5 times faster in computing filter importance and is able to achieve similar performance compared to that of the active filter pruning methods. The efficacy of the proposed pruning method is evaluated on audio scene classification and image classification using various CNNs architecture such as VGGish, DCASE21_Net, VGG-16 and ResNet-50. | ['Mark D. Plumbley', 'Arshdeep Singh'] | 2023-04-05 | null | null | null | null | ['scene-classification'] | ['computer-vision'] | [ 2.59558707e-01 1.95408612e-01 5.51709235e-01 -3.52706492e-01
1.31002069e-01 -2.59314567e-01 1.31137684e-01 2.95997024e-01
-8.53246391e-01 7.58943856e-01 5.82211800e-02 -2.88040459e-01
-4.87842262e-01 -1.11104262e+00 -5.35613418e-01 -4.38376069e-01
1.00891732e-01 -3.73569846e-01 8.24014544e-01 -3.00291747e-01
3.25897932e-01 5.53511798e-01 -2.06243801e+00 5.25783241e-01
5.09363532e-01 1.39903307e+00 2.21267208e-01 5.75512350e-01
-1.21935025e-01 4.93708014e-01 -7.23595142e-01 -2.76769549e-01
5.80922544e-01 -1.78682357e-01 -6.39727175e-01 -4.13863420e-01
5.11375427e-01 -2.84203440e-02 -2.22646028e-01 1.31532156e+00
4.81767923e-01 3.63129318e-01 3.82075787e-01 -8.93980205e-01
-5.04914448e-02 6.05416417e-01 -2.89274901e-01 6.70086920e-01
-4.60592546e-02 -2.51980811e-01 8.99852812e-01 -8.46048415e-01
4.81030762e-01 1.19249856e+00 9.94106889e-01 4.14569259e-01
-1.13058054e+00 -9.78507817e-01 2.67938763e-01 1.05404794e-01
-1.48556018e+00 -3.20512444e-01 4.50563341e-01 -4.52657163e-01
1.36518872e+00 3.89657438e-01 8.06374431e-01 1.52467534e-01
1.25372171e-01 1.53204188e-01 6.89248323e-01 -5.01880407e-01
2.25149453e-01 6.33168817e-02 4.21902150e-01 8.18387330e-01
4.00245368e-01 -4.23385622e-03 -6.96631551e-01 -3.63700911e-02
6.26881838e-01 -1.74673110e-01 -3.41331661e-01 2.85463333e-01
-6.31681144e-01 7.63645709e-01 4.47682828e-01 3.05472970e-01
-4.05286759e-01 2.93157130e-01 5.45002043e-01 4.64425355e-01
3.18316996e-01 2.87177145e-01 -6.28575444e-01 1.02570154e-01
-1.30865824e+00 4.48377371e-01 6.31150305e-01 8.55619073e-01
8.04944396e-01 4.30514455e-01 -2.05407977e-01 7.47076392e-01
2.46580765e-01 -1.54571727e-01 3.56840849e-01 -6.00580454e-01
5.33454597e-01 1.16025019e+00 -3.26206475e-01 -1.07227910e+00
-3.62789899e-01 -6.71586573e-01 -9.10246730e-01 8.06634724e-01
2.00569868e-01 -3.45780194e-01 -1.20939434e+00 1.64942944e+00
5.76949082e-02 7.52738258e-03 8.18367023e-03 5.19231141e-01
1.09842086e+00 6.11767709e-01 1.56096771e-01 -2.70269424e-01
1.24753678e+00 -5.37840009e-01 -5.99626243e-01 2.10923523e-01
3.65293980e-01 -1.08643508e+00 7.44923234e-01 5.69161892e-01
-1.08521616e+00 -9.29593205e-01 -1.56253374e+00 1.13790981e-01
-6.85813487e-01 4.89523739e-01 3.73113722e-01 8.10885727e-01
-1.10708725e+00 9.01650667e-01 -6.44860268e-01 -2.86686480e-01
6.85985625e-01 7.88201034e-01 -2.42827252e-01 5.25031805e-01
-1.02113771e+00 7.06076562e-01 8.39214981e-01 6.73439726e-02
-5.82065165e-01 -1.05759633e+00 -4.15827572e-01 4.71367568e-01
1.90483257e-01 -3.22612196e-01 9.58147943e-01 -9.43230093e-01
-1.18659759e+00 2.73352921e-01 1.71891466e-01 -7.62243867e-01
4.67444867e-01 -2.70532727e-01 -4.38329101e-01 1.49773493e-01
-3.80268693e-01 9.73351777e-01 6.33496463e-01 -5.93377829e-01
-1.10462403e+00 1.03808179e-01 4.33749527e-01 1.16813086e-01
-6.28228784e-01 -1.39290541e-01 -1.42492682e-01 -7.64834583e-01
4.06616151e-01 -5.07370353e-01 -2.87062258e-01 2.31420785e-01
-3.41669232e-01 -1.87877744e-01 1.14589000e+00 -2.12617993e-01
1.54767096e+00 -2.20030403e+00 -2.43318245e-01 2.41916597e-01
3.98948878e-01 7.46885180e-01 6.51786923e-02 2.09890589e-01
-1.97070867e-01 2.87468255e-01 9.45115928e-03 -3.93778831e-03
-6.10315919e-01 2.17568353e-01 -8.03320706e-02 1.53102323e-01
4.33319807e-01 4.66708466e-02 -5.95985115e-01 -3.90486330e-01
2.24106714e-01 6.48656726e-01 -8.71462822e-01 -8.98194388e-02
3.27509381e-02 -1.31529480e-01 -2.44441479e-01 3.21013302e-01
7.43904054e-01 3.75701964e-01 -1.84946001e-01 -7.68897414e-01
-4.67945188e-01 1.69221118e-01 -1.67903495e+00 1.04810536e+00
-1.02286600e-01 7.63777375e-01 9.16969180e-02 -1.17389274e+00
1.07553256e+00 5.90789318e-01 3.89117956e-01 -2.77566344e-01
4.23578054e-01 2.16074288e-01 4.66796577e-01 -1.69962630e-01
5.63129961e-01 6.75029680e-02 5.42613089e-01 -5.66772968e-02
4.73488778e-01 1.36835679e-01 7.22483337e-01 2.72478890e-02
1.32122445e+00 -1.89195246e-01 2.64793396e-01 -7.97084749e-01
9.72927272e-01 -1.59460247e-01 8.42895389e-01 6.74282730e-01
-3.47105302e-02 4.89327490e-01 5.60522616e-01 -5.20032644e-01
-1.00343466e+00 -8.28614235e-01 -2.50118494e-01 8.77387285e-01
-2.28127405e-01 -7.35005796e-01 -7.99183190e-01 -3.42820555e-01
-1.33635044e-01 2.23185748e-01 -5.67525029e-01 -1.82376429e-01
-5.13972223e-01 -6.92042351e-01 9.46636081e-01 5.19664645e-01
1.00413072e+00 -1.00840342e+00 -1.11197054e+00 3.71645749e-01
3.49128544e-01 -8.15878808e-01 6.93191886e-02 6.38194501e-01
-1.24661648e+00 -1.29285562e+00 -2.72254884e-01 -6.72275126e-01
9.28253055e-01 1.24432147e-01 7.53296852e-01 4.16710317e-01
-5.53478479e-01 -2.22440153e-01 -5.01622140e-01 -8.70237172e-01
2.18059421e-02 1.89225525e-01 -8.57744515e-02 -1.98930904e-01
2.31405213e-01 -7.61215270e-01 -6.24230444e-01 9.11315084e-02
-9.24987197e-01 3.76844034e-02 5.81621945e-01 8.10226083e-01
7.06326425e-01 6.62996113e-01 4.98054087e-01 -1.19573545e+00
9.54457521e-01 -1.79602906e-01 -7.12260067e-01 -4.27475534e-02
-5.49459875e-01 1.14027053e-01 8.77245009e-01 -5.83544195e-01
-9.55745339e-01 2.05220893e-01 -2.56398618e-01 -8.16358700e-02
8.89958516e-02 4.02960181e-01 9.75479186e-02 -3.22186589e-01
8.27313483e-01 -2.03742623e-01 -3.88953269e-01 -6.37275279e-01
-1.26897261e-01 8.55315030e-02 3.26429695e-01 -3.42997789e-01
6.35996044e-01 2.24548370e-01 3.11437458e-01 -8.84160221e-01
-4.01885450e-01 -3.34794879e-01 -5.72706819e-01 -3.72548044e-01
7.07828879e-01 -5.71594834e-01 -6.63209975e-01 2.63444573e-01
-1.18445742e+00 2.38009155e-01 -3.97128761e-01 6.23818994e-01
-4.30843420e-02 -1.49519727e-01 -3.77030432e-01 -8.87567043e-01
-7.09463000e-01 -1.25097406e+00 3.69308054e-01 5.02436161e-01
-2.37763777e-01 -4.14916933e-01 -5.31614304e-01 -4.02718127e-01
6.44528389e-01 3.77423286e-01 1.24827087e+00 -4.97741878e-01
-4.54342246e-01 -2.52090365e-01 -4.20422465e-01 8.20375621e-01
-7.98556283e-02 2.28460193e-01 -8.95915270e-01 -4.35905755e-02
-5.73688522e-02 3.85044850e-02 1.12948823e+00 4.24188167e-01
1.05842185e+00 -2.16782302e-01 -8.78907740e-02 5.38221180e-01
1.91696668e+00 4.77032900e-01 5.99378049e-01 1.20837741e-01
2.55159616e-01 3.58372569e-01 3.19978386e-01 4.55086410e-01
-5.66656768e-01 1.26347080e-01 6.15795970e-01 -1.30639032e-01
-4.06815916e-01 1.45816922e-01 -1.00097962e-01 8.46920192e-01
-4.77172911e-01 -1.18006177e-01 -5.56800365e-01 4.23807442e-01
-1.71554685e+00 -8.24146092e-01 -3.76433820e-01 2.07411981e+00
7.04098821e-01 8.01951349e-01 -2.58161277e-01 9.69554782e-01
5.75558364e-01 -1.67208612e-02 -1.94157407e-01 -8.31969261e-01
-1.33344373e-02 9.25270259e-01 7.19503522e-01 2.90442199e-01
-8.82096708e-01 7.23182499e-01 5.98784065e+00 7.82589078e-01
-1.07554996e+00 2.25508004e-01 2.92270660e-01 -3.12764615e-01
3.13545227e-01 1.83399096e-02 -1.05785549e+00 1.99790075e-01
5.87197304e-01 1.36991993e-01 1.77844137e-01 8.40491474e-01
2.45677620e-01 -4.60670829e-01 -9.08899724e-01 7.40064859e-01
-3.86777550e-01 -1.55532050e+00 1.55116782e-01 -1.35070518e-01
6.27300143e-01 -9.51521099e-02 -3.21489424e-01 1.58515334e-01
-5.04442453e-02 -9.49088335e-01 1.08598554e+00 5.68861604e-01
5.03000438e-01 -9.86020267e-01 9.72636282e-01 1.16747372e-01
-1.61094141e+00 -2.56250769e-01 -8.68288219e-01 -5.11421263e-01
-1.22541592e-01 7.08048403e-01 -5.08292496e-01 1.99995384e-01
1.08640599e+00 1.61036238e-01 -5.84488034e-01 1.33094406e+00
1.07193790e-01 6.24666095e-01 -4.89490151e-01 -2.21397996e-01
2.80168504e-01 7.18216971e-02 5.28602600e-01 1.44380927e+00
4.11727041e-01 1.85211137e-01 -2.14652896e-01 8.23417187e-01
-1.77919716e-01 3.11797619e-01 -3.12304914e-01 1.14294738e-01
6.27788663e-01 1.15261698e+00 -1.09731686e+00 -4.17507291e-01
-4.10518944e-01 2.12157771e-01 2.16948912e-01 2.56412596e-01
-6.19727552e-01 -1.00337410e+00 7.57606566e-01 3.21864247e-01
4.66035366e-01 -8.49116296e-02 -3.87755305e-01 -3.42855722e-01
5.69344163e-02 -5.20164728e-01 2.56589741e-01 -4.86744583e-01
-5.59479117e-01 1.11128962e+00 1.87378258e-01 -1.12930691e+00
3.64598066e-01 -8.54422808e-01 -6.43943787e-01 1.02295434e+00
-1.15321684e+00 -6.55045927e-01 -4.72576261e-01 6.92215323e-01
6.58375800e-01 -4.41274673e-01 6.78070903e-01 6.68936014e-01
-4.09173220e-01 6.87646806e-01 -3.96653801e-01 1.48910254e-01
1.83047116e-01 -8.04350734e-01 -1.71605572e-01 1.06408250e+00
-8.78937766e-02 8.45522881e-01 7.06753731e-01 -5.39138198e-01
-8.04134965e-01 -1.12531996e+00 7.28166878e-01 5.37550807e-01
1.63297415e-01 -2.60544956e-01 -6.58898473e-01 2.00065345e-01
1.65303305e-01 1.36443228e-01 4.18465763e-01 9.47741326e-03
-2.59551615e-01 -5.57402968e-01 -1.36851060e+00 4.34770048e-01
1.04983914e+00 -9.98646617e-02 -5.85872948e-01 -1.39694288e-01
5.05524457e-01 -1.46432430e-01 -7.07616031e-01 7.70745039e-01
6.18322372e-01 -1.25171757e+00 7.88143456e-01 -3.57980669e-01
2.58084446e-01 -4.96800035e-01 -1.55875668e-01 -8.75702977e-01
-5.14760435e-01 -3.00285518e-01 7.15887398e-02 1.23909068e+00
6.21459126e-01 -5.42991459e-01 8.49829853e-01 1.48990929e-01
-3.62579823e-01 -9.36092794e-01 -1.00006986e+00 -5.78067780e-01
-3.37084919e-01 -4.97705728e-01 4.98489708e-01 4.54899311e-01
-5.11520267e-01 1.54355645e-01 1.10479243e-01 1.86050273e-02
2.90097296e-01 -6.31134391e-01 3.14907849e-01 -1.73612201e+00
-8.73657092e-02 -7.21986294e-01 -8.27776492e-01 -4.52116966e-01
-5.07119358e-01 -6.55704319e-01 -1.19358942e-01 -1.43961775e+00
-2.51203120e-01 -4.56490010e-01 -4.46702629e-01 7.14740336e-01
2.35451430e-01 5.53251266e-01 1.64510131e-01 -1.18531115e-01
-4.11741510e-02 -3.98885608e-02 9.56909180e-01 -2.28476319e-02
-3.74577940e-01 -8.66369903e-03 -5.16720235e-01 1.00930703e+00
8.24265361e-01 -6.74591064e-01 -8.13181102e-01 -4.32041258e-01
4.22640413e-01 -5.38991094e-01 1.38587832e-01 -1.68418276e+00
5.29179990e-01 1.53802112e-01 6.15716934e-01 -7.93898106e-01
3.47263813e-01 -8.43150556e-01 3.14444959e-01 6.97368085e-01
-3.40322137e-01 3.62492055e-01 5.78969598e-01 3.95533383e-01
-4.17720616e-01 -6.59113526e-01 1.02947462e+00 -2.30491906e-01
-7.23194659e-01 3.29624908e-03 -6.80271864e-01 -1.33506268e-01
6.98793113e-01 -6.95663214e-01 -1.74812868e-01 2.54569948e-01
-7.96146750e-01 -3.19514751e-01 -4.13767129e-01 9.95916352e-02
6.92080081e-01 -1.08873236e+00 -5.01821101e-01 2.66174465e-01
-2.64433980e-01 1.09560438e-01 1.50267249e-02 6.90703273e-01
-9.49943483e-01 4.36000109e-01 -6.78954244e-01 -2.40650788e-01
-1.67600083e+00 1.62810665e-02 2.91714579e-01 -1.44084796e-01
-5.95715940e-01 1.27612615e+00 -5.67855425e-02 1.60002574e-01
6.50577903e-01 -8.21473718e-01 -4.85535443e-01 1.93264708e-01
4.46316063e-01 6.39336824e-01 5.32855570e-01 -3.50090712e-01
-2.78937221e-01 4.62008566e-01 -7.36673502e-03 -1.22721810e-02
1.42026770e+00 4.13405746e-01 -1.24201588e-01 -4.11422066e-02
1.02356088e+00 -6.75130710e-02 -9.07814085e-01 1.27261179e-02
1.98613070e-02 -2.90164500e-01 3.61626327e-01 -6.37538850e-01
-1.51633132e+00 8.74861956e-01 8.43104780e-01 9.32622179e-02
1.62018561e+00 -7.07382858e-01 4.74749893e-01 4.20975268e-01
7.65225813e-02 -1.47484326e+00 -1.41555175e-01 6.20597422e-01
7.60079682e-01 -6.39382899e-01 3.53007048e-01 -7.37274051e-01
1.44268587e-01 1.49684596e+00 8.61794829e-01 -5.24309099e-01
1.26008093e+00 6.27237618e-01 -2.73577064e-01 -1.50357381e-01
-7.72918820e-01 -2.97139615e-01 3.15760761e-01 4.45736796e-01
6.68690205e-01 -1.71897367e-01 -1.01331735e+00 7.03120053e-01
-4.73132968e-01 -5.14559336e-02 3.20896268e-01 1.05725777e+00
-7.36879110e-01 -1.01714408e+00 -1.11528769e-01 8.60944450e-01
-6.37975454e-01 -3.66010755e-01 -2.06091151e-01 9.52589154e-01
1.06837702e+00 1.05386472e+00 1.44799709e-01 -7.00004935e-01
6.23950005e-01 5.84446676e-02 3.73279899e-01 -7.44554996e-01
-1.21453130e+00 -9.18508843e-02 1.59750119e-01 -3.20476681e-01
-5.09653747e-01 -2.09349349e-01 -1.37901068e+00 -2.02367842e-01
-6.27729833e-01 1.46787539e-01 6.83616340e-01 8.33484709e-01
1.92697138e-01 1.05109096e+00 -4.10387106e-02 -3.89727980e-01
-2.58197099e-01 -1.00891840e+00 -3.72242272e-01 8.10452253e-02
-3.81059125e-02 -8.87531042e-01 -1.95850909e-01 1.34119913e-01] | [8.54523754119873, 3.0131213665008545] |
39153f62-f768-4192-b53f-c038ba46c751 | neural-decipherment-via-minimum-cost-flow | 1906.06718 | null | https://arxiv.org/abs/1906.06718v1 | https://arxiv.org/pdf/1906.06718v1.pdf | Neural Decipherment via Minimum-Cost Flow: from Ugaritic to Linear B | In this paper we propose a novel neural approach for automatic decipherment of lost languages. To compensate for the lack of strong supervision signal, our model design is informed by patterns in language change documented in historical linguistics. The model utilizes an expressive sequence-to-sequence model to capture character-level correspondences between cognates. To effectively train the model in an unsupervised manner, we innovate the training procedure by formalizing it as a minimum-cost flow problem. When applied to the decipherment of Ugaritic, we achieve a 5.5% absolute improvement over state-of-the-art results. We also report the first automatic results in deciphering Linear B, a syllabic language related to ancient Greek, where our model correctly translates 67.3% of cognates. | ['Regina Barzilay', 'Jiaming Luo', 'Yuan Cao'] | 2019-06-16 | neural-decipherment-via-minimum-cost-flow-1 | https://aclanthology.org/P19-1303 | https://aclanthology.org/P19-1303.pdf | acl-2019-7 | ['decipherment'] | ['natural-language-processing'] | [ 4.27806914e-01 2.57438481e-01 7.67911300e-02 -3.75812948e-01
-6.42321110e-01 -6.48448527e-01 4.67928648e-01 -5.71130514e-02
-8.28886628e-01 8.03963959e-01 4.12975043e-01 -7.13288307e-01
2.67346501e-01 -6.92013562e-01 -9.26665306e-01 -2.18398854e-01
7.08020627e-02 4.66553688e-01 -2.23201513e-01 -5.86805642e-01
3.51678044e-01 3.50059927e-01 -1.37541449e+00 6.73194349e-01
9.70451891e-01 2.30681002e-01 2.95327306e-01 7.07343102e-01
-1.32272527e-01 5.13767779e-01 -7.10097492e-01 -6.12719476e-01
1.39191315e-01 -6.77474320e-01 -1.01799464e+00 -2.25825906e-01
5.64573348e-01 -2.72326529e-01 -2.59191424e-01 7.73492157e-01
4.89381105e-01 -4.75435138e-01 6.83270931e-01 -2.78104812e-01
-9.17046070e-01 1.04475856e+00 -4.25684415e-02 3.90192479e-01
3.53641152e-01 -4.14379453e-03 1.17599416e+00 -9.63366091e-01
8.60058904e-01 1.12728667e+00 8.07345927e-01 6.13349557e-01
-1.23361671e+00 -2.77953357e-01 -1.91493690e-01 4.07762825e-01
-1.35769105e+00 -8.27307403e-01 6.05754495e-01 -4.17867303e-01
1.29007792e+00 2.23129883e-01 9.14015055e-01 8.94527555e-01
5.44166677e-02 9.94826913e-01 1.15100586e+00 -9.42338526e-01
8.38913172e-02 -1.65900588e-01 7.24220350e-02 1.03718090e+00
9.55215469e-02 2.34360501e-01 -7.37645864e-01 2.78217167e-01
5.07390082e-01 -7.80006230e-01 -3.30610543e-01 2.40657926e-01
-1.07646668e+00 6.75003707e-01 2.22245842e-01 6.48809075e-01
-1.64843008e-01 1.31007701e-01 4.64340955e-01 4.09902781e-01
3.39308500e-01 5.40514231e-01 -4.59460884e-01 -4.27968800e-01
-1.03784275e+00 -2.82124653e-02 8.62070918e-01 6.38690889e-01
5.45902371e-01 3.57324660e-01 5.01633942e-01 8.96472991e-01
5.55807054e-02 1.44205689e-01 9.03590739e-01 -7.36170053e-01
4.76883322e-01 4.47385907e-01 -4.04604018e-01 -4.84732687e-01
-2.50360310e-01 -5.48343599e-01 -5.17901123e-01 -6.84373593e-03
6.29040003e-01 1.40642777e-01 -7.77475059e-01 1.67322385e+00
-8.70706066e-02 -2.62271583e-01 2.26609126e-01 6.49673820e-01
4.34717476e-01 5.89959800e-01 -2.27765754e-01 -2.63505787e-01
1.24173272e+00 -5.84251404e-01 -5.47922134e-01 -3.21647227e-01
8.37498724e-01 -6.11628711e-01 1.40529537e+00 5.38171232e-01
-1.23829389e+00 -5.58937252e-01 -1.41833210e+00 -2.50990450e-01
-3.37602019e-01 3.08707297e-01 2.57808059e-01 7.06806064e-01
-1.08628011e+00 9.05932784e-01 -6.26018643e-01 -4.74119216e-01
1.18570872e-01 1.93000764e-01 -4.37812328e-01 3.66591573e-01
-1.24625707e+00 1.07473326e+00 8.48755062e-01 2.76779264e-01
-7.37046540e-01 -5.63177109e-01 -6.91984177e-01 -1.89217776e-02
4.63926308e-02 -5.57600439e-01 1.33862829e+00 -1.23518658e+00
-1.71341920e+00 1.48386073e+00 -2.16401964e-01 -7.29871035e-01
5.77184796e-01 -3.71964097e-01 -4.73210573e-01 1.90721318e-01
-1.43556058e-01 4.89116818e-01 7.07295716e-01 -1.01183379e+00
-5.24925172e-01 -1.99500799e-01 -2.77882397e-01 -1.12735346e-01
-4.19977814e-01 1.71809673e-01 -9.24071446e-02 -8.74946296e-01
-5.36892638e-02 -7.11814404e-01 2.35080928e-01 -2.57019997e-01
-3.39613140e-01 -2.50026099e-02 3.78044218e-01 -1.34381080e+00
1.36411870e+00 -1.94353199e+00 3.02195758e-01 -4.20075320e-02
2.23886073e-01 4.22057718e-01 -1.21563487e-01 5.15213192e-01
-1.70610070e-01 2.33767822e-01 -8.81479919e-01 -5.40303648e-01
7.45884106e-02 5.61699927e-01 -3.69377464e-01 4.95139122e-01
3.81035358e-01 1.10913610e+00 -8.86594415e-01 -3.63116801e-01
-2.67835885e-01 2.26913124e-01 -4.97354478e-01 7.45410249e-02
-1.87594518e-01 1.59090698e-01 2.64011443e-01 6.57526612e-01
4.71654296e-01 8.17584768e-02 5.89685321e-01 2.66821504e-01
-2.71175295e-01 7.54423261e-01 -5.20033836e-01 1.94608152e+00
-4.65907544e-01 8.80239010e-01 -1.42752469e-01 -1.09325361e+00
9.29826319e-01 2.65337706e-01 -2.08203137e-01 -6.71040237e-01
3.42100635e-02 5.72975218e-01 4.34476972e-01 -4.79340166e-01
9.06076670e-01 -5.03454506e-01 -2.31286556e-01 6.12538576e-01
1.49579957e-01 -1.53130363e-03 2.30848923e-01 5.12087308e-02
9.29513812e-01 3.68068039e-01 7.77817428e-01 -4.62598115e-01
6.37583673e-01 2.28950866e-02 5.48494995e-01 5.92090964e-01
-2.07956973e-02 7.01265574e-01 3.67166221e-01 -9.03782964e-01
-1.37024975e+00 -1.10451126e+00 -4.23103683e-02 9.87959862e-01
-5.20605743e-01 -6.61888480e-01 -9.54568386e-01 -6.24725878e-01
-1.33883476e-01 9.34626997e-01 -5.30774713e-01 -2.28208616e-01
-1.37437057e+00 -6.14800632e-01 1.30384135e+00 3.26255232e-01
3.31259787e-01 -1.28456652e+00 -6.53985739e-01 4.91145194e-01
-2.89768636e-01 -9.84156132e-01 -2.58922279e-01 3.21037650e-01
-7.88483262e-01 -8.49121213e-01 -4.99937832e-01 -1.04663241e+00
4.21437114e-01 -4.39789772e-01 1.31157601e+00 2.68169701e-01
-2.93249458e-01 -1.79463714e-01 -3.08745831e-01 -2.54931897e-01
-1.03761768e+00 5.28020501e-01 1.44508734e-01 -2.06965879e-01
3.70510519e-01 -6.90407217e-01 -1.03357717e-01 -1.86220258e-01
-9.29023385e-01 3.47908936e-03 5.51578760e-01 9.00582194e-01
2.20235690e-01 -5.51886976e-01 4.81858045e-01 -9.31777775e-01
6.70878828e-01 -1.73115358e-01 -4.73554820e-01 2.22811118e-01
-6.60004973e-01 4.59591955e-01 9.98109579e-01 -3.19589972e-01
-8.57932210e-01 1.92018718e-01 -4.17025805e-01 3.13335419e-01
-8.09467286e-02 6.56921089e-01 -1.03422850e-01 1.87492505e-01
8.71368587e-01 6.40484810e-01 -7.05463588e-02 -7.88254082e-01
6.66629434e-01 9.52709138e-01 1.06801152e+00 -7.56870210e-01
6.08539224e-01 2.39108950e-01 -3.65692288e-01 -1.01459885e+00
-4.71114039e-01 -4.59068231e-02 -1.04998040e+00 -6.47489056e-02
6.12721086e-01 -7.88599372e-01 -5.86509824e-01 4.71481174e-01
-1.51058793e+00 -3.71817887e-01 -2.52965420e-01 1.84397712e-01
-8.09027135e-01 5.37069619e-01 -9.30305183e-01 -4.79956925e-01
-5.14415026e-01 -7.23700464e-01 8.54406834e-01 -2.83974946e-01
-6.24758840e-01 -8.70597184e-01 4.17424619e-01 1.45288855e-01
1.50282919e-01 1.07126139e-01 1.34930682e+00 -5.14651060e-01
-4.68074173e-01 1.81187972e-01 1.41047001e-01 4.91050631e-01
3.87321375e-02 -1.46401271e-01 -9.63400364e-01 -1.62445322e-01
-9.39759314e-02 -1.77227721e-01 9.76542711e-01 -3.13230872e-01
7.19318748e-01 -4.92622882e-01 1.32211924e-01 8.24657083e-01
1.17494607e+00 1.25271305e-02 7.67640293e-01 6.68015957e-01
5.31238079e-01 7.60567784e-01 6.04272671e-02 2.54978269e-01
4.20211524e-01 4.38248962e-01 2.13783622e-01 1.53413579e-01
-3.89047891e-01 -6.03226364e-01 7.01201022e-01 1.47541046e+00
-1.57334730e-01 -1.15134366e-01 -1.27675259e+00 8.99123847e-01
-1.64694500e+00 -8.35935533e-01 -1.27959892e-01 1.95506907e+00
1.20733178e+00 8.07630718e-02 8.30213130e-02 4.62212741e-01
5.78488708e-01 5.30211963e-02 -1.22784860e-01 -8.14940155e-01
-4.40535665e-01 5.50260484e-01 2.99593270e-01 8.63917112e-01
-5.83047032e-01 1.59162962e+00 7.34048414e+00 5.03897727e-01
-1.11783433e+00 -5.92232049e-02 2.74293065e-01 -3.10535021e-02
-4.23622876e-01 1.83502689e-01 -6.57421172e-01 4.54648286e-01
1.18751323e+00 -4.90888245e-02 5.95477998e-01 3.13311338e-01
1.01141848e-01 2.11867109e-01 -1.35077107e+00 7.70095468e-01
4.40101445e-01 -1.31984890e+00 6.05462939e-02 -8.91425274e-03
3.07733119e-01 -5.26268035e-02 -1.86856672e-01 1.74461484e-01
1.44969910e-01 -1.13039219e+00 1.14832783e+00 5.41140795e-01
9.07188714e-01 -7.88296938e-01 4.04182047e-01 5.20666063e-01
-7.42156982e-01 1.51666433e-01 -3.26725841e-01 -4.25229758e-01
3.26014668e-01 4.23634470e-01 -1.24260437e+00 3.09791654e-01
2.80931085e-01 7.96048939e-01 -6.29103363e-01 7.20217109e-01
-8.73487294e-01 9.92341578e-01 -1.60190448e-01 -1.75906330e-01
9.22560617e-02 -1.56063497e-01 7.33598650e-01 1.69218767e+00
3.04813951e-01 -8.27678740e-02 -4.08987641e-01 7.56670356e-01
-1.94790870e-01 2.72135526e-01 -6.40535772e-01 -1.27122805e-01
2.51581132e-01 7.48877347e-01 -4.12974119e-01 -4.98416632e-01
-1.13819996e-02 1.42975903e+00 8.46929073e-01 1.20475136e-01
-4.40281332e-01 -5.83109140e-01 4.08795863e-01 2.00770110e-01
5.05578279e-01 -5.73370516e-01 -3.62119466e-01 -1.28495336e+00
3.42983872e-01 -1.10046756e+00 1.21487722e-01 -5.70358276e-01
-1.01927543e+00 8.17498088e-01 -2.39274859e-01 -9.26403761e-01
-6.12158954e-01 -8.67516220e-01 -4.83184308e-01 7.90202737e-01
-1.39144278e+00 -1.16626120e+00 2.41742492e-01 2.85977751e-01
3.90895575e-01 -4.61881846e-01 9.99870777e-01 2.96398431e-01
-3.29644978e-01 7.45276630e-01 1.63071454e-01 2.78683662e-01
5.67175150e-01 -1.38158596e+00 9.76025224e-01 1.23078823e+00
4.77361321e-01 8.53572130e-01 6.49420857e-01 -5.10433555e-01
-1.12121677e+00 -6.73504889e-01 1.72285271e+00 -1.89084411e-01
8.75147700e-01 -8.39917600e-01 -9.51891124e-01 7.96541274e-01
3.36087853e-01 -4.46925223e-01 6.66837454e-01 8.42456967e-02
-6.64122999e-01 2.42862869e-02 -7.82036424e-01 7.22079694e-01
1.38088906e+00 -1.01355588e+00 -1.23302090e+00 7.78890178e-02
7.61840940e-01 -1.60754308e-01 -6.39805436e-01 -1.03102319e-01
7.33759582e-01 -9.54968512e-01 6.78485990e-01 -6.77800894e-01
6.19159460e-01 -2.36078575e-01 -1.88649237e-01 -1.41050339e+00
-2.05639929e-01 -9.15947080e-01 -1.74799696e-01 1.02696776e+00
5.78372121e-01 -5.44957221e-01 8.17983568e-01 -1.31874979e-01
-4.15382475e-01 -2.98494637e-01 -1.20853019e+00 -8.62832546e-01
4.07742769e-01 -4.05889690e-01 4.05554831e-01 9.97665405e-01
3.35896224e-01 5.67372203e-01 -6.48344517e-01 -2.54665047e-01
3.79385799e-01 -1.65817700e-02 3.73945892e-01 -1.05864382e+00
-6.96478546e-01 -5.04045308e-01 -5.63343942e-01 -1.11960888e+00
5.02743065e-01 -1.47402263e+00 1.89568743e-01 -9.49905753e-01
-2.37324536e-01 3.65839712e-02 2.13031676e-02 6.40252292e-01
2.79300660e-02 2.94523597e-01 3.81704211e-01 2.77395219e-01
-2.38830447e-02 5.71655750e-01 5.79857111e-01 -1.43311322e-01
-1.16447411e-01 -3.71468544e-01 -6.92068756e-01 7.57629871e-01
1.12592614e+00 -6.44114852e-01 1.86398163e-01 -8.90993297e-01
4.70022023e-01 -3.08236152e-01 1.04221694e-01 -7.61565983e-01
1.10184975e-01 2.76409090e-01 2.40761563e-01 -4.56702948e-01
1.52512297e-01 -4.40528780e-01 -1.65259317e-01 9.00449753e-01
-2.65155017e-01 5.50428927e-01 1.61932468e-01 4.00633328e-02
-3.75102103e-01 -4.46334749e-01 6.37942731e-01 -3.55226457e-01
-7.93217063e-01 -3.87722164e-01 -8.40897560e-01 9.16877240e-02
5.83820879e-01 -2.37244666e-01 -3.77013177e-01 -1.05507366e-01
-6.17199183e-01 -3.19571555e-01 4.43055421e-01 3.51911396e-01
6.73606873e-01 -1.11413872e+00 -1.13511944e+00 3.02304447e-01
2.00769618e-01 -4.90087062e-01 -4.60643589e-01 2.79267877e-01
-1.02749610e+00 3.59300405e-01 -3.72671902e-01 -3.26741666e-01
-1.23154151e+00 3.79572421e-01 3.29480946e-01 -1.52075455e-01
-6.71358109e-01 5.69411099e-01 -2.82180429e-01 -7.43283987e-01
-1.08936615e-01 -2.31687754e-01 2.41690353e-02 -1.75493322e-02
4.42487419e-01 3.88496161e-01 2.63315380e-01 -7.68485129e-01
-3.94835025e-01 5.89061379e-01 -2.96214998e-01 -4.23694849e-01
1.48948109e+00 -3.14430483e-02 -3.87562931e-01 5.90801001e-01
1.31835425e+00 3.52160931e-01 -8.21974754e-01 -2.25285426e-01
5.32473087e-01 -1.64562821e-01 -3.48386168e-01 -9.07774568e-01
-5.09757996e-01 9.77056444e-01 3.22832763e-01 -2.27212340e-01
1.01153922e+00 -1.24003515e-01 9.69190836e-01 7.15891182e-01
2.05348715e-01 -1.29342556e+00 -1.19241975e-01 1.14332521e+00
8.95437717e-01 -8.22528958e-01 -1.61136135e-01 1.53335556e-02
-3.53690147e-01 1.42920876e+00 1.82685211e-01 -2.08280325e-01
1.28882378e-01 3.42607617e-01 2.11179897e-01 6.08485416e-02
-6.43453896e-01 -1.18953809e-01 2.01329693e-01 4.78794605e-01
5.44006288e-01 1.86969399e-01 -8.92238259e-01 4.93070662e-01
-9.52046156e-01 -1.45761237e-01 6.80185199e-01 9.68731165e-01
-5.02928257e-01 -1.37304473e+00 -2.59913772e-01 -1.83413208e-01
-5.25336862e-01 -6.50652945e-01 -7.64335454e-01 8.78975034e-01
2.44758412e-01 5.65049171e-01 1.78821281e-01 -4.09004837e-01
3.21837097e-01 4.17888016e-01 5.79707503e-01 -3.48777801e-01
-9.69320297e-01 -1.04305565e-01 4.10053492e-01 -2.32514039e-01
-9.07208174e-02 -5.90321958e-01 -1.26752007e+00 -4.29454744e-01
1.82556182e-01 -1.86036304e-02 5.64865291e-01 9.90739226e-01
1.57404989e-01 4.14117783e-01 2.42596939e-01 -3.38027805e-01
-5.96087039e-01 -9.66825724e-01 -3.60776514e-01 2.88908541e-01
4.10902619e-01 1.46237850e-01 -2.90123701e-01 3.11912626e-01] | [10.842726707458496, 10.02560806274414] |
b5e8a39f-e884-4b34-8d8b-6199621f411a | a-computational-framework-of-human-values-for | 2305.02748 | null | https://arxiv.org/abs/2305.02748v1 | https://arxiv.org/pdf/2305.02748v1.pdf | A computational framework of human values for ethical AI | In the diverse array of work investigating the nature of human values from psychology, philosophy and social sciences, there is a clear consensus that values guide behaviour. More recently, a recognition that values provide a means to engineer ethical AI has emerged. Indeed, Stuart Russell proposed shifting AI's focus away from simply ``intelligence'' towards intelligence ``provably aligned with human values''. This challenge -- the value alignment problem -- with others including an AI's learning of human values, aggregating individual values to groups, and designing computational mechanisms to reason over values, has energised a sustained research effort. Despite this, no formal, computational definition of values has yet been proposed. We address this through a formal conceptual framework rooted in the social sciences, that provides a foundation for the systematic, integrated and interdisciplinary investigation into how human values can support designing ethical AI. | ["Mark d'Inverno", 'Nardine Osman'] | 2023-05-04 | null | null | null | null | ['philosophy'] | ['miscellaneous'] | [ 4.86140132e-01 5.63430250e-01 -4.43433464e-01 -4.32684034e-01
1.51737064e-01 -3.29184592e-01 7.51223743e-01 5.01531780e-01
-5.55400014e-01 4.07332003e-01 9.98918056e-01 -3.41790348e-01
-7.92159677e-01 -5.10474682e-01 -2.19491601e-01 -5.97411335e-01
5.20272136e-01 1.95797950e-01 -2.89476454e-01 -4.78307188e-01
7.07641602e-01 1.49651647e-01 -1.49563646e+00 -1.06543504e-01
6.78992748e-01 6.41856790e-01 -5.28530538e-01 2.10108757e-01
1.15663745e-01 1.37201917e+00 -4.18107450e-01 -7.44263351e-01
-1.13190658e-01 -9.70546007e-01 -1.01272535e+00 -1.60576046e-01
-3.03814054e-01 -3.29995066e-01 4.30883825e-01 8.80764782e-01
4.65361536e-01 -2.54867613e-01 5.51864982e-01 -1.30383980e+00
-1.29069507e+00 1.07690442e+00 -1.21828457e-02 -2.24645644e-01
4.07113791e-01 5.35363674e-01 1.16978836e+00 8.30997825e-02
7.37308323e-01 1.12372506e+00 6.82210386e-01 7.09097147e-01
-1.14191711e+00 -3.04822594e-01 -2.54258901e-01 1.62675336e-01
-8.98706555e-01 -5.01495898e-01 7.79672563e-01 -8.96938026e-01
8.07703137e-01 5.55869699e-01 1.60443914e+00 9.47735012e-01
9.25016999e-02 1.66637748e-01 1.21755004e+00 -6.02878511e-01
3.19376826e-01 3.11439365e-01 -4.38163169e-02 -2.36640200e-01
7.79849589e-01 2.72745490e-01 -6.29560530e-01 -1.06722206e-01
3.43303978e-01 -3.93184096e-01 1.17319897e-01 -2.67272323e-01
-1.54606283e+00 6.61187530e-01 1.57972500e-01 8.08406234e-01
-7.39141524e-01 4.52646315e-01 4.25071895e-01 2.53934324e-01
1.02871135e-01 1.17614281e+00 -3.77016962e-02 -9.35972273e-01
-4.04452354e-01 4.18284237e-01 8.66809845e-01 3.58484745e-01
1.45517856e-01 -5.61620630e-02 2.15448603e-01 3.40665817e-01
8.60237837e-01 4.31347817e-01 -1.05003975e-01 -1.49990880e+00
-4.34080034e-01 8.86168301e-01 3.32239300e-01 -1.31962764e+00
-1.97336107e-01 -2.97339499e-01 -1.46836638e-01 2.03143910e-01
5.27151644e-01 -2.47988522e-01 -1.19785674e-01 1.78920758e+00
3.80789906e-01 -6.29157484e-01 2.84817278e-01 1.11967576e+00
3.14452440e-01 1.16114996e-01 3.20134461e-01 -1.17055327e-01
1.15187466e+00 -1.18760288e-01 -6.33622348e-01 -1.92516640e-01
7.65684903e-01 -5.39372802e-01 9.90236342e-01 3.85556817e-01
-1.41651547e+00 3.20995092e-01 -1.00365162e+00 -1.99662149e-01
-4.87393700e-02 -7.42222726e-01 1.00573266e+00 1.04143846e+00
-1.28623784e+00 5.13601601e-01 -4.39386904e-01 -7.86375701e-01
6.10122919e-01 3.81926894e-02 1.37528151e-01 3.44992638e-01
-1.03662336e+00 1.21360075e+00 3.11054289e-01 1.91468775e-01
-4.72234577e-01 -4.12628084e-01 -3.52114320e-01 -3.27016771e-01
4.76962626e-01 -9.06137645e-01 9.82548535e-01 -1.78729177e+00
-1.14888465e+00 1.05366576e+00 5.61629593e-01 -4.40640539e-01
4.11733955e-01 -1.33327052e-01 -3.45177144e-01 -8.42758715e-02
-6.75985664e-02 5.21617174e-01 7.63035119e-02 -1.56346118e+00
-3.29584718e-01 -3.22906315e-01 1.80097014e-01 1.75647721e-01
-5.25529981e-01 8.02416980e-01 8.32293451e-01 -3.09336990e-01
-1.57882124e-02 -7.63752520e-01 -2.03446910e-01 -1.50053948e-01
6.26401184e-03 -3.36642861e-01 -6.89507574e-02 -1.74607113e-01
1.40679622e+00 -2.10060406e+00 1.23786263e-01 3.33568335e-01
6.45082951e-01 -1.23428412e-01 4.52801317e-01 9.55386043e-01
3.14557642e-01 3.51124674e-01 -1.70147821e-01 7.39296019e-01
6.52978420e-01 3.47672373e-01 -2.00083345e-01 3.50796431e-01
5.88880032e-02 7.47092247e-01 -1.22758234e+00 -4.14415509e-01
2.57721394e-02 8.18703413e-01 -9.04493093e-01 -1.54802144e-01
5.66678382e-02 2.64773369e-01 -5.32427073e-01 4.65383261e-01
6.60791844e-02 -3.22423339e-01 6.50111079e-01 3.86391163e-01
-6.41918778e-01 3.23486149e-01 -7.75535345e-01 8.84404182e-01
2.75385648e-01 7.46330023e-01 6.64931955e-03 -8.43421459e-01
1.00035048e+00 3.56582314e-01 6.18351400e-01 -5.60712159e-01
7.24485576e-01 3.42497855e-01 8.98973703e-01 -6.38729930e-01
3.87502313e-01 -4.42940086e-01 -4.92136329e-02 9.15760219e-01
-4.53373015e-01 -4.51476723e-01 -1.84212580e-01 -7.59393200e-02
8.79661679e-01 7.13833272e-01 3.31469446e-01 -7.43739545e-01
3.61940920e-01 3.88914347e-01 9.06952262e-01 2.06793696e-01
-5.79976439e-01 -5.76281026e-02 7.42549360e-01 -5.00996828e-01
-1.02979279e+00 -7.15749979e-01 -2.43770391e-01 9.69301701e-01
2.86752582e-01 -2.21019506e-01 -1.10331690e+00 8.50156546e-02
2.46875938e-02 1.26145101e+00 -4.88503963e-01 -1.74503461e-01
-4.60997462e-01 -3.36880684e-01 3.88814479e-01 1.27247497e-01
2.41621092e-01 -1.38038564e+00 -1.76054752e+00 2.89110154e-01
1.62897319e-01 -5.80629826e-01 4.56956834e-01 -1.11273952e-01
-4.52805936e-01 -1.00922787e+00 -9.40787271e-02 -3.21834445e-01
5.86471796e-01 -1.06683612e-01 1.12234902e+00 5.40908575e-01
9.53469723e-02 8.13938618e-01 -4.26902056e-01 -9.03697252e-01
-8.06760490e-01 -5.15895307e-01 -1.63958177e-01 -1.87965900e-01
8.12267601e-01 -6.69288576e-01 -6.51849926e-01 3.07900086e-02
-1.06781173e+00 3.26005489e-01 5.11359811e-01 2.41731659e-01
-6.53336346e-02 -1.75540671e-01 9.65337574e-01 -7.60174870e-01
8.74487340e-01 -7.04235077e-01 2.15136871e-01 2.07295641e-01
-1.07361543e+00 -2.19493330e-01 1.75224096e-01 -1.83712646e-01
-1.00439167e+00 -9.00739133e-01 4.33828905e-02 6.91437423e-01
-2.47176677e-01 5.46109438e-01 -1.31290108e-01 5.45615889e-02
8.25600505e-01 -3.30340385e-01 4.97112691e-01 1.54611751e-01
4.84623313e-01 6.52008295e-01 3.40996772e-01 -1.00823724e+00
5.59547067e-01 2.87891924e-01 -8.71148258e-02 -8.78203809e-01
-4.28914428e-01 2.06707194e-01 -3.99499863e-01 -9.10749137e-01
1.13766253e+00 -5.21077394e-01 -1.05824018e+00 1.07445821e-01
-7.54538655e-01 -2.74805456e-01 -6.33085608e-01 8.37808788e-01
-9.10603762e-01 1.51816025e-01 -6.22115806e-02 -1.12658775e+00
-9.97092202e-02 -9.84111607e-01 2.06491604e-01 4.42067415e-01
-1.23598456e+00 -1.07020319e+00 -1.03380308e-01 8.77377450e-01
7.99061358e-01 8.75236988e-01 1.03701293e+00 -6.80917084e-01
-5.38402438e-01 1.29680023e-01 1.47393256e-01 3.91764462e-01
1.18042335e-01 1.28874347e-01 -6.13522947e-01 4.01291579e-01
5.31616271e-01 -5.95519423e-01 -2.39347175e-01 -2.58407984e-02
1.90684021e-01 -6.72925651e-01 9.55712274e-02 7.08524138e-02
1.36380863e+00 7.45881498e-01 7.15718031e-01 9.23250139e-01
3.98876369e-01 1.35805249e+00 6.20532393e-01 6.18212223e-01
9.58446980e-01 1.56974927e-01 1.26683176e-01 2.10495189e-01
5.22145391e-01 2.26156920e-01 7.90045261e-02 6.70331240e-01
-5.10057390e-01 3.36870372e-01 -1.45211744e+00 6.10144794e-01
-1.90840602e+00 -1.23614633e+00 -2.78413922e-01 1.95309699e+00
1.14660656e+00 3.23648542e-01 3.20852876e-01 1.62598595e-01
3.46430838e-01 1.83583796e-01 -3.71544212e-01 -1.11181045e+00
1.58299595e-01 -2.99748182e-01 5.17905988e-02 2.81084657e-01
-2.05403656e-01 8.14209521e-01 6.40305042e+00 -1.62055880e-01
-6.42631769e-01 -1.73069999e-01 5.98595083e-01 -5.27535379e-02
-1.31571233e+00 3.06212157e-01 -1.64815411e-01 3.25391859e-01
8.44058216e-01 -8.53849769e-01 5.47126114e-01 5.19275010e-01
4.19641852e-01 -2.86918908e-01 -9.87421274e-01 3.02926600e-01
1.47765085e-01 -8.84742796e-01 -3.97254318e-01 4.53639716e-01
5.39090753e-01 -6.48651302e-01 4.48107161e-02 -4.65629697e-01
7.49747992e-01 -1.15492499e+00 1.23462248e+00 5.42114317e-01
1.43532589e-01 -7.82106042e-01 5.37552297e-01 2.06844643e-01
-2.29489222e-01 -1.68923557e-01 -2.91891769e-02 -6.87119544e-01
2.01756105e-01 2.17691943e-01 -3.24472129e-01 -1.55655295e-01
4.67544824e-01 4.94393110e-01 -2.45880768e-01 4.04865116e-01
-4.74044010e-02 5.25635123e-01 -2.43093427e-02 -1.82768792e-01
2.74262011e-01 -5.50983071e-01 5.26815236e-01 7.78855026e-01
-2.58000288e-02 3.39275330e-01 -5.86737931e-01 1.18187177e+00
7.34354556e-01 1.56040505e-01 -7.24978507e-01 -9.06268001e-01
6.05141342e-01 1.02919888e+00 -9.71318781e-01 1.47482499e-01
-4.63472128e-01 3.90538690e-03 -3.72776210e-01 1.02495566e-01
-8.54698300e-01 -9.52850655e-02 9.04784322e-01 2.97502995e-01
-5.08487284e-01 -2.34762177e-01 -7.79743016e-01 -4.90649521e-01
-3.45454603e-01 -1.36100721e+00 -5.39444759e-02 -4.02484477e-01
-8.23213518e-01 9.31198522e-02 2.60332942e-01 -4.01794255e-01
-3.21664810e-01 -3.02162886e-01 -1.31746948e-01 4.97076124e-01
-7.47806847e-01 -1.38654006e+00 -1.33317441e-01 -8.90033916e-02
-4.53310251e-01 1.66800693e-01 6.28216863e-01 -3.79594505e-01
-1.87820986e-01 3.03919554e-01 -1.78292885e-01 -4.43575442e-01
4.34192508e-01 -8.55395675e-01 1.51175231e-01 6.48054838e-01
-5.97680569e-01 8.70267391e-01 9.24792945e-01 -6.58134937e-01
-1.74434185e+00 1.64386109e-01 1.31383717e+00 -6.97222888e-01
1.00292516e+00 7.21229911e-02 -7.77167380e-01 6.08924389e-01
7.03075409e-01 -8.87920082e-01 1.35387003e+00 -4.54827622e-02
1.98151730e-02 -7.33580068e-02 -1.49099147e+00 9.82581735e-01
1.45767367e+00 -1.59701914e-01 -8.53384793e-01 -2.70656705e-01
7.91896880e-01 1.59561664e-01 -1.46370423e+00 1.50159746e-01
9.19483721e-01 -1.44407821e+00 9.63499129e-01 -2.12594524e-01
6.46980226e-01 -9.83105451e-02 -2.15920284e-01 -5.33762097e-01
-4.72069949e-01 -8.48613739e-01 5.94416499e-01 1.60360527e+00
-3.66927274e-02 -1.17403972e+00 6.43114984e-01 1.51470089e+00
-3.04172486e-01 -8.62546980e-01 -6.44786716e-01 -3.38278338e-02
1.98342562e-01 -4.43055809e-01 1.16473734e+00 1.45366132e+00
4.50051874e-01 -5.51898638e-03 -6.56622350e-02 -5.29301345e-01
8.31673145e-01 -4.36971903e-01 7.94055939e-01 -1.55209196e+00
-8.88070911e-02 -8.92734647e-01 -6.14849031e-01 1.16265282e-01
-2.29739293e-01 -7.57823467e-01 -1.12661839e-01 -1.93929529e+00
6.48506135e-02 -3.44078481e-01 -1.15344785e-01 2.60932356e-01
2.34754547e-01 -2.55850434e-01 5.47072351e-01 1.75575346e-01
-3.68168741e-01 1.10139646e-01 1.20239735e+00 5.79646826e-01
-2.80255616e-01 -9.77966130e-01 -1.98846996e+00 9.80479002e-01
9.66642857e-01 -3.67204040e-01 -6.64496601e-01 -2.51656085e-01
1.27297068e+00 -1.74612120e-01 3.10434580e-01 -9.74094331e-01
4.66375612e-02 -1.27601647e+00 1.04708038e-01 2.36087158e-01
-2.18933493e-01 -1.12323856e+00 9.64866400e-01 6.43995643e-01
-6.46937251e-01 -1.19767889e-01 -1.01927496e-01 -3.88051599e-01
2.66903609e-01 -3.17019701e-01 6.30511940e-01 -4.35677692e-02
-4.83803779e-01 -6.61336064e-01 -3.16975415e-01 4.39083040e-01
1.29637170e+00 -9.14520621e-01 -6.53955817e-01 -1.21451318e-01
-3.01586747e-01 2.32861176e-01 1.27705967e+00 2.29916304e-01
4.99523491e-01 -1.22165680e+00 -5.50296068e-01 1.04776071e-02
-1.66531780e-03 -1.68224424e-01 1.06120318e-01 7.86052167e-01
-7.88182795e-01 1.76517278e-01 -4.62134153e-01 -2.22648568e-02
-7.39129066e-01 2.36778364e-01 3.11364621e-01 6.81308746e-01
-6.52314425e-01 6.61862612e-01 -7.68794166e-03 8.60203952e-02
-1.10485844e-01 2.27256909e-01 -3.57974559e-01 4.34028417e-01
2.72803962e-01 7.16033578e-01 -1.14237273e+00 -1.23108768e+00
-4.31465864e-01 5.92504978e-01 4.71067756e-01 -5.62258482e-01
1.61204386e+00 -1.26999602e-01 -7.52697408e-01 6.34359598e-01
6.02347374e-01 -1.50175273e-01 -9.77997661e-01 4.51180696e-01
4.68478620e-01 -8.82835925e-01 -3.11739594e-01 -1.02764785e+00
-1.81949973e-01 4.12774354e-01 3.02767068e-01 6.33039653e-01
1.04881573e+00 -8.17520395e-02 3.33578169e-01 -6.18458434e-04
4.44086045e-01 -1.66881537e+00 5.10685407e-02 4.17105816e-02
1.07611442e+00 -6.34542823e-01 1.25725597e-01 8.52907542e-03
-1.27267385e+00 8.69189560e-01 5.76851785e-01 1.26097262e-01
5.33951044e-01 1.07438415e-01 2.60038555e-01 -1.76333085e-01
-8.44853401e-01 -2.06804365e-01 -9.95107144e-02 8.73091400e-01
7.98606336e-01 2.18653679e-01 -1.06072485e+00 5.78146160e-01
-6.51858211e-01 5.55784106e-01 7.49454618e-01 9.85451043e-01
-6.66115165e-01 -1.20544517e+00 -4.67186689e-01 1.17037907e-01
-8.80229950e-01 2.01738775e-01 -1.05780017e+00 8.27085674e-01
5.15874028e-01 1.17746317e+00 5.03401086e-02 -3.14338863e-01
2.24222168e-01 -5.19732907e-02 3.24144453e-01 -1.73656732e-01
-1.34631574e+00 1.04024075e-01 4.88302141e-01 -5.02501011e-01
-8.03901196e-01 -9.42496359e-01 -1.40925491e+00 -9.47832227e-01
1.96842447e-01 2.78271914e-01 6.13552034e-01 6.68664813e-01
3.60709161e-01 2.89727896e-01 2.61370987e-01 -2.00252041e-01
-5.10704756e-01 -2.78039902e-01 -3.62882763e-01 4.04396951e-01
-6.68781102e-02 -4.09834445e-01 -4.95917886e-01 -2.57428978e-02] | [9.069456100463867, 6.343607425689697] |
19fe5e4f-be69-4acb-acac-0e7a321fb9fa | bert-based-clinical-knowledge-extraction-for | 2304.10996 | null | https://arxiv.org/abs/2304.10996v1 | https://arxiv.org/pdf/2304.10996v1.pdf | BERT Based Clinical Knowledge Extraction for Biomedical Knowledge Graph Construction and Analysis | Background : Knowledge is evolving over time, often as a result of new discoveries or changes in the adopted methods of reasoning. Also, new facts or evidence may become available, leading to new understandings of complex phenomena. This is particularly true in the biomedical field, where scientists and physicians are constantly striving to find new methods of diagnosis, treatment and eventually cure. Knowledge Graphs (KGs) offer a real way of organizing and retrieving the massive and growing amount of biomedical knowledge. Objective : We propose an end-to-end approach for knowledge extraction and analysis from biomedical clinical notes using the Bidirectional Encoder Representations from Transformers (BERT) model and Conditional Random Field (CRF) layer. Methods : The approach is based on knowledge graphs, which can effectively process abstract biomedical concepts such as relationships and interactions between medical entities. Besides offering an intuitive way to visualize these concepts, KGs can solve more complex knowledge retrieval problems by simplifying them into simpler representations or by transforming the problems into representations from different perspectives. We created a biomedical Knowledge Graph using using Natural Language Processing models for named entity recognition and relation extraction. The generated biomedical knowledge graphs (KGs) are then used for question answering. Results : The proposed framework can successfully extract relevant structured information with high accuracy (90.7% for Named-entity recognition (NER), 88% for relation extraction (RE)), according to experimental findings based on real-world 505 patient biomedical unstructured clinical notes. Conclusions : In this paper, we propose a novel end-to-end system for the construction of a biomedical knowledge graph from clinical textual using a variation of BERT models. | ['Bouchra El Asri', 'Zineb Elkaimbillah', 'Siham Yousfi', 'Mounia Mikram', 'Maryem Rhanoui', 'Ayoub Harnoune'] | 2023-04-21 | null | null | null | null | ['graph-construction', 'clinical-knowledge', 'named-entity-recognition-ner'] | ['graphs', 'miscellaneous', 'natural-language-processing'] | [ 2.68280953e-01 5.54728210e-01 -9.36243534e-02 -2.50001550e-01
-3.93938750e-01 -2.89029241e-01 3.91459763e-01 1.11801434e+00
-4.13436890e-01 9.75184977e-01 4.17987108e-01 -4.62177336e-01
-5.38154066e-01 -1.22266459e+00 -3.38267505e-01 -4.00938541e-01
-2.83330590e-01 7.61729658e-01 -1.86551129e-04 -4.76018339e-02
-1.06054712e-02 7.11866021e-01 -1.19729507e+00 4.07861441e-01
1.05507612e+00 7.47016788e-01 1.88601479e-01 4.99613881e-01
-5.38195431e-01 1.15360820e+00 -7.00994790e-01 -5.57918191e-01
-2.94169635e-01 -3.06793153e-01 -1.31800365e+00 -1.58687159e-01
-4.83976334e-01 2.77408391e-01 -8.48790035e-02 9.69401121e-01
4.70779032e-01 1.49043491e-02 7.30811000e-01 -7.77657866e-01
-6.59759045e-01 6.49555385e-01 -2.33028457e-01 6.54414250e-03
5.83530009e-01 -3.35793644e-01 7.84538269e-01 -6.13458693e-01
9.69232738e-01 1.12962139e+00 4.12570894e-01 4.56370741e-01
-1.00744271e+00 -4.44697142e-01 -1.46465704e-01 4.88280207e-01
-1.49516952e+00 -1.97027028e-01 3.99950862e-01 -6.03986204e-01
1.05905712e+00 2.82315582e-01 6.58901632e-01 7.06126451e-01
4.97910619e-01 4.99771118e-01 8.00343335e-01 -4.50884253e-01
1.93346053e-01 3.53007376e-01 4.26069200e-01 8.84656727e-01
5.68769276e-01 -3.82626742e-01 -3.86718005e-01 -3.63475561e-01
4.50070053e-01 2.20291421e-01 -4.29390520e-01 5.02677076e-03
-1.26353216e+00 5.40917456e-01 5.48643768e-01 7.51542628e-01
-7.03440309e-01 -3.87896061e-01 4.39187318e-01 -4.61498611e-02
1.46797135e-01 4.98140782e-01 -4.81172830e-01 1.01175919e-01
-7.44765222e-01 -9.51635689e-02 9.98672009e-01 9.82282460e-01
4.53155607e-01 -2.96074182e-01 -2.89320558e-01 6.13275528e-01
3.53711337e-01 3.64715308e-01 8.38723838e-01 -1.07218839e-01
2.69253612e-01 1.00450456e+00 -2.34829843e-01 -1.33445048e+00
-6.92338467e-01 -5.70123017e-01 -1.17451739e+00 -3.79552007e-01
2.77158096e-02 8.42482299e-02 -1.02117991e+00 1.43494630e+00
3.56863558e-01 -1.04813986e-02 5.32545447e-01 5.00937462e-01
1.34723687e+00 5.37322342e-01 6.04630172e-01 -3.57061714e-01
2.09273672e+00 -4.02902067e-01 -1.15403414e+00 1.86855838e-01
7.57394552e-01 -5.10196745e-01 2.26575926e-01 3.17589521e-01
-7.88310528e-01 -2.44455591e-01 -7.30441272e-01 -1.51708141e-01
-7.81015873e-01 1.81088805e-01 7.29903936e-01 3.41925055e-01
-8.37237239e-01 3.93468529e-01 -9.04450417e-01 -4.85864818e-01
6.02971494e-01 4.93370295e-01 -6.71430349e-01 -1.91163003e-01
-1.33717525e+00 1.00589120e+00 8.47847283e-01 2.87206382e-01
-2.91700095e-01 -7.29376018e-01 -9.35356498e-01 3.38140666e-01
3.94179553e-01 -1.15247166e+00 7.02635229e-01 -4.41351056e-01
-1.04465556e+00 7.79003203e-01 -1.53573632e-01 -4.43735421e-01
-4.64735068e-02 -1.43865362e-01 -7.89855659e-01 3.06910485e-01
1.57964341e-02 2.92193294e-01 1.26138330e-01 -9.47751284e-01
-2.85238802e-01 -5.77997446e-01 -1.91386700e-01 5.22096306e-02
-3.08046162e-01 -4.28888723e-02 -2.49489814e-01 -5.13075769e-01
8.79783276e-03 -5.38748324e-01 -3.43375891e-01 -3.64244640e-01
-6.21290803e-01 -2.59333491e-01 3.59771580e-01 -1.04164970e+00
1.33663309e+00 -1.79791451e+00 3.23666751e-01 4.18371975e-01
6.48574948e-01 5.67541301e-01 3.13058972e-01 6.42171502e-01
-2.78008312e-01 3.51760447e-01 -3.83243889e-01 8.13738480e-02
-4.02339131e-01 2.21699119e-01 -6.01036064e-02 -3.47109959e-02
2.26811215e-01 9.68720973e-01 -1.07408762e+00 -9.21310246e-01
1.23040326e-01 7.07312047e-01 -3.12608123e-01 2.19588667e-01
-1.91593423e-01 4.40874368e-01 -7.96169341e-01 6.21813297e-01
3.50124538e-01 -5.06777048e-01 6.39107168e-01 -3.60652715e-01
2.62727112e-01 1.17667846e-01 -8.67462337e-01 1.48153484e+00
-4.37607467e-01 2.38710448e-01 -3.72215927e-01 -1.14608562e+00
1.09845662e+00 7.34545231e-01 5.76911509e-01 -3.73305768e-01
2.34277695e-01 7.65580609e-02 -2.10091889e-01 -7.96514034e-01
1.37419358e-01 -4.35640633e-01 2.54146997e-02 -1.85856119e-01
2.99677283e-01 1.97713256e-01 2.24280700e-01 3.88209373e-01
1.27041948e+00 -8.55319649e-02 1.01224208e+00 -3.97953093e-02
9.43899930e-01 9.05812383e-02 5.96903384e-01 1.96479037e-01
2.79744565e-01 -1.92704089e-02 4.90711689e-01 -4.22386646e-01
-3.64517361e-01 -1.09206080e+00 -2.17078507e-01 8.73535872e-02
-3.02545667e-01 -6.99712694e-01 -4.48735088e-01 -6.48157299e-01
-1.59185439e-01 5.42495489e-01 -4.67269272e-01 -2.77943850e-01
-3.09663951e-01 -6.21469676e-01 5.59627593e-01 3.55318666e-01
3.83313686e-01 -1.31153738e+00 -5.68829715e-01 4.62350219e-01
-3.36335838e-01 -1.25415432e+00 1.92189530e-01 1.02169834e-01
-1.17604673e+00 -1.32525480e+00 -6.15284145e-01 -7.94294298e-01
8.97525728e-01 -3.30531508e-01 1.11977136e+00 1.39200892e-02
-6.96410358e-01 3.59931231e-01 -4.47751373e-01 -5.34347177e-01
-6.91595674e-01 2.69157272e-02 -2.21551076e-01 -3.82685438e-02
4.48217928e-01 -5.50152361e-01 -5.68188071e-01 -1.83106080e-01
-1.38291419e+00 2.02812687e-01 8.64119530e-01 7.37510562e-01
6.63393557e-01 2.99585819e-01 7.86511719e-01 -1.29415858e+00
7.72688866e-01 -7.55313754e-01 -3.19964975e-01 5.87252259e-01
-7.79348195e-01 3.78861934e-01 8.59883785e-01 3.50194052e-02
-1.06229508e+00 5.81215182e-03 -3.04111928e-01 -2.02214137e-01
-3.16213489e-01 1.09966385e+00 -1.66856050e-01 4.67997313e-01
6.07594371e-01 1.24860644e-01 -1.74833193e-01 -5.16134381e-01
4.07888025e-01 7.75739908e-01 4.40315336e-01 -4.84815538e-01
3.57507288e-01 2.56433874e-01 3.75148445e-01 -7.94768333e-01
-5.74097693e-01 -5.83013952e-01 -6.57465160e-01 -2.82336865e-02
1.19937432e+00 -6.55967236e-01 -8.83863688e-01 -6.50711656e-02
-1.12009323e+00 4.12392795e-01 -3.34317952e-01 6.24499500e-01
-1.53915644e-01 4.87817645e-01 -5.69722116e-01 -5.36363065e-01
-7.45381117e-01 -9.19347763e-01 9.15365160e-01 2.77907044e-01
-2.51422793e-01 -1.22695160e+00 1.22435749e-01 3.27205986e-01
1.14594139e-01 5.25571704e-01 1.61608779e+00 -8.89205575e-01
-2.58344531e-01 -2.33265340e-01 -1.78252727e-01 5.29498123e-02
5.77069163e-01 -2.74760246e-01 -6.51926041e-01 9.60392207e-02
-2.32729122e-01 1.44284293e-01 6.28320754e-01 3.01299617e-02
1.03282118e+00 -5.76112092e-01 -7.52734244e-01 2.04293221e-01
1.51964140e+00 4.86176759e-01 9.36553180e-01 -8.63861665e-03
6.95329905e-01 6.76897049e-01 2.86777496e-01 3.37340117e-01
4.15780813e-01 4.87063468e-01 8.53024572e-02 -1.16016649e-01
-3.42313386e-02 -1.79848880e-01 -6.60457388e-02 1.10487926e+00
-2.78159380e-01 -1.45232260e-01 -1.25362897e+00 6.20306253e-01
-1.71141088e+00 -8.38109791e-01 -1.88835621e-01 2.14943385e+00
1.17771363e+00 -1.45291030e-01 -3.08876872e-01 2.21099406e-01
6.32949531e-01 -4.59559560e-01 -1.39042452e-01 -3.25909168e-01
2.08844274e-01 6.40485048e-01 2.57190645e-01 3.34698439e-01
-6.66388035e-01 7.39562809e-01 4.92808914e+00 6.33437335e-01
-9.68776584e-01 -1.63797677e-01 4.14331794e-01 4.43484843e-01
-2.67276585e-01 -1.08342916e-01 -6.64520442e-01 2.55788583e-02
1.09597361e+00 -3.94194752e-01 -6.43338710e-02 5.24637759e-01
-8.15443471e-02 -2.66481154e-02 -1.00717819e+00 1.01774621e+00
-1.10042617e-01 -1.53815114e+00 4.35316443e-01 -1.22349802e-02
2.75592387e-01 -4.72420305e-01 -4.59170312e-01 1.15606897e-01
3.37344050e-01 -1.03272808e+00 8.39401856e-02 1.17112684e+00
7.67626703e-01 -5.70889533e-01 1.13060939e+00 2.30672300e-01
-1.28033078e+00 1.31161153e-01 -1.18534595e-01 4.31744754e-01
1.00540146e-01 1.03789759e+00 -1.44290006e+00 1.40022683e+00
4.46050704e-01 6.96342468e-01 -5.49948156e-01 1.03344893e+00
-2.19436318e-01 5.30176401e-01 -9.07371417e-02 -8.31025764e-02
-1.82388246e-01 5.69303818e-02 2.91430473e-01 1.48088205e+00
2.87931651e-01 4.41696227e-01 -9.62034389e-02 7.93758273e-01
-1.14534058e-01 5.89116156e-01 -6.93369925e-01 -5.12482882e-01
2.07225725e-01 1.35017645e+00 -1.06427896e+00 -5.94393313e-01
-1.56732067e-01 6.23485088e-01 2.82871693e-01 2.92105317e-01
-5.50681710e-01 -5.33586860e-01 1.44074515e-01 7.55164549e-02
-3.32220830e-03 4.68750112e-02 6.95009083e-02 -1.16848338e+00
-1.27091989e-01 -9.02406454e-01 7.33395576e-01 -6.24675810e-01
-1.24791157e+00 8.95682514e-01 1.44541606e-01 -1.07153118e+00
-1.98309675e-01 -6.41677976e-01 3.42747904e-02 7.77129591e-01
-1.52174628e+00 -9.66161549e-01 -2.57767767e-01 6.85141087e-01
6.25031739e-02 -9.28305686e-02 1.44181895e+00 5.15301287e-01
-2.64103681e-01 9.88784581e-02 -1.53310612e-01 3.17478985e-01
5.39586008e-01 -1.22273505e+00 -1.64626479e-01 3.94117415e-01
2.15154678e-01 1.04487371e+00 3.28697324e-01 -8.24254274e-01
-1.16798377e+00 -1.07054865e+00 1.40552032e+00 -8.62146989e-02
3.65765840e-01 -1.09161474e-01 -9.67941642e-01 5.40822744e-01
-1.44187110e-02 -6.89172819e-02 1.20657766e+00 1.45812392e-01
-1.85807288e-01 -1.02418758e-01 -1.32820857e+00 4.55441803e-01
7.70272076e-01 -4.50316191e-01 -8.25455070e-01 4.54127371e-01
6.88891113e-01 -1.24756210e-01 -1.33590448e+00 5.19581199e-01
3.51501077e-01 -4.95419174e-01 8.74166787e-01 -1.00440872e+00
3.49092931e-01 -4.02790785e-01 9.58937332e-02 -1.16773331e+00
-2.09848762e-01 -3.86518687e-01 -1.32326245e-01 1.06360805e+00
5.33989966e-01 -6.68745399e-01 4.41042334e-01 5.66895008e-01
4.46769921e-03 -9.84264433e-01 -7.94868469e-01 -2.43970469e-01
-3.62967253e-01 -1.99985832e-01 5.75105846e-01 1.18789113e+00
4.87317204e-01 5.92515528e-01 8.41160789e-02 2.83954471e-01
2.41156876e-01 3.98482382e-02 2.36295491e-01 -1.58939469e+00
-3.07326317e-01 -1.40633211e-01 -1.04445672e+00 -4.09180522e-01
-1.13916986e-01 -1.30454791e+00 -2.82634467e-01 -2.30594587e+00
2.41935700e-01 -3.50562662e-01 -4.80739087e-01 8.29949260e-01
-6.16468154e-02 -2.03078464e-01 -1.35661345e-02 7.71227926e-02
-3.45241725e-01 2.80287951e-01 1.11665535e+00 -2.27103010e-01
-2.26124138e-01 -9.68669951e-02 -6.62456155e-01 5.26456356e-01
6.46940768e-01 -5.82089901e-01 -4.05605733e-01 6.14675134e-02
5.73716521e-01 5.03548503e-01 2.88606375e-01 -9.28870857e-01
3.19564760e-01 5.50163053e-02 3.23177963e-01 -3.73044014e-01
-5.72756911e-03 -1.00325537e+00 4.62739736e-01 6.60739839e-01
-2.08115831e-01 3.79635207e-02 3.21841329e-01 5.46091318e-01
-4.29064572e-01 -1.68107733e-01 3.89373600e-01 -9.71218720e-02
-4.67321604e-01 2.46427879e-02 -3.76112908e-01 -1.12934880e-01
9.66955304e-01 6.70546889e-02 -2.92975247e-01 -1.91557854e-01
-1.36237419e+00 8.95475000e-02 -2.40214154e-01 3.56571048e-01
8.80432248e-01 -1.05396390e+00 -5.19573748e-01 1.15453452e-02
2.08279312e-01 1.42563626e-01 2.94319212e-01 8.31836462e-01
-7.53832519e-01 6.96422458e-01 -1.43172339e-01 -2.90089279e-01
-1.42289758e+00 7.38286316e-01 2.01490000e-01 -9.55859900e-01
-6.52240336e-01 4.88262802e-01 2.02748612e-01 -2.94838905e-01
7.20372191e-03 -5.86603820e-01 -8.22583437e-01 1.49687544e-01
4.14350569e-01 1.62200809e-01 3.23530018e-01 -5.30739188e-01
-5.83239853e-01 3.63085002e-01 -1.88602313e-01 2.03848779e-01
1.47806072e+00 2.74113983e-01 -4.40857321e-01 3.15076619e-01
9.88474905e-01 1.23404279e-01 9.85065475e-03 -1.79112136e-01
3.02899837e-01 5.46673834e-02 -1.26584202e-01 -1.05461574e+00
-1.01740313e+00 7.15037584e-01 4.10279512e-01 1.85969815e-01
1.25205183e+00 1.23243287e-01 6.08848691e-01 6.18496001e-01
4.51667041e-01 -5.55398107e-01 -3.51850390e-01 1.61428735e-01
7.38322496e-01 -6.82719171e-01 3.08389723e-01 -7.46809125e-01
-5.11478305e-01 1.33741510e+00 -7.48726055e-02 4.53646392e-01
9.49858010e-01 2.09454998e-01 -1.16088070e-01 -6.61475897e-01
-5.43086588e-01 -1.90298110e-01 4.10512298e-01 4.94875669e-01
6.50067568e-01 1.80488065e-01 -5.56611598e-01 8.55592728e-01
-8.18745121e-02 4.55123395e-01 1.56544492e-01 9.36952949e-01
-1.80185437e-01 -1.36014795e+00 -2.13648051e-01 5.54300070e-01
-6.64691806e-01 -3.49719942e-01 -3.18646193e-01 5.80126643e-01
2.96476185e-01 9.57074404e-01 -4.73535597e-01 -2.81016976e-01
3.91029835e-01 4.51798618e-01 3.89295846e-01 -8.15700173e-01
-5.95131934e-01 -2.77818024e-01 3.04452032e-01 -2.66787589e-01
-6.38677657e-01 -2.90992975e-01 -1.81882620e+00 2.71381378e-01
-4.13253605e-01 6.01821363e-01 5.74741006e-01 9.69012618e-01
7.42324769e-01 1.13477015e+00 1.76832035e-01 3.98172438e-02
-3.83119285e-02 -8.33127916e-01 -4.87547308e-01 3.93423587e-01
-1.48552880e-02 -5.23202598e-01 1.79922864e-01 5.36435902e-01] | [8.469648361206055, 8.58399486541748] |
9779b6ee-beb1-4dce-bcd9-e4854dd1a51b | text2mesh-text-driven-neural-stylization-for | 2112.03221 | null | https://arxiv.org/abs/2112.03221v1 | https://arxiv.org/pdf/2112.03221v1.pdf | Text2Mesh: Text-Driven Neural Stylization for Meshes | In this work, we develop intuitive controls for editing the style of 3D objects. Our framework, Text2Mesh, stylizes a 3D mesh by predicting color and local geometric details which conform to a target text prompt. We consider a disentangled representation of a 3D object using a fixed mesh input (content) coupled with a learned neural network, which we term neural style field network. In order to modify style, we obtain a similarity score between a text prompt (describing style) and a stylized mesh by harnessing the representational power of CLIP. Text2Mesh requires neither a pre-trained generative model nor a specialized 3D mesh dataset. It can handle low-quality meshes (non-manifold, boundaries, etc.) with arbitrary genus, and does not require UV parameterization. We demonstrate the ability of our technique to synthesize a myriad of styles over a wide variety of 3D meshes. | ['Rana Hanocka', 'Sagie Benaim', 'Richard Liu', 'Roi Bar-On', 'Oscar Michel'] | 2021-12-06 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Michel_Text2Mesh_Text-Driven_Neural_Stylization_for_Meshes_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Michel_Text2Mesh_Text-Driven_Neural_Stylization_for_Meshes_CVPR_2022_paper.pdf | cvpr-2022-1 | ['neural-stylization'] | ['computer-vision'] | [ 3.08022410e-01 3.10243219e-01 2.10152328e-01 -1.32551834e-01
-4.10413235e-01 -1.02941716e+00 7.68365562e-01 -2.15812460e-01
4.02391225e-01 4.72576767e-01 1.28309116e-01 -8.94931033e-02
2.66982168e-01 -9.96126294e-01 -9.36838329e-01 -2.44929761e-01
1.10793017e-01 7.53338814e-01 -4.31104563e-02 -3.15950811e-01
2.44726971e-01 1.07866669e+00 -1.36147451e+00 8.38433653e-02
6.98672771e-01 8.77741754e-01 -1.38125733e-01 9.72890437e-01
-2.27032557e-01 1.45366415e-01 -5.71185827e-01 -4.66363668e-01
5.19745350e-01 -3.09359431e-01 -6.50117815e-01 3.02065700e-01
8.62762272e-01 -3.10837150e-01 -1.99908867e-01 8.57891619e-01
3.29384148e-01 -2.06747755e-01 9.68450367e-01 -1.03954160e+00
-9.66994643e-01 1.73891410e-01 -4.68139797e-01 -4.22999829e-01
5.13728976e-01 3.18646669e-01 7.23604918e-01 -8.63939583e-01
1.21581173e+00 1.62013614e+00 5.80592930e-01 7.55365193e-01
-1.93975866e+00 -4.94336545e-01 -1.01807907e-01 -8.07332873e-01
-1.02575147e+00 -2.06709266e-01 1.16028595e+00 -7.95463681e-01
3.74851406e-01 2.85250217e-01 1.00246501e+00 1.36730468e+00
4.85880196e-01 3.55400324e-01 1.25244713e+00 -2.39811122e-01
2.23522693e-01 -5.23427017e-02 -3.64109099e-01 9.10951674e-01
-2.86710709e-02 1.73425481e-01 -4.84904647e-01 -3.55949432e-01
1.65653932e+00 -2.25321442e-01 -1.19253211e-01 -9.08297598e-01
-1.31635451e+00 6.19309306e-01 2.44753301e-01 -1.14826307e-01
-7.96594918e-02 4.20559883e-01 1.06039874e-01 4.42025572e-01
7.74747133e-01 1.03361964e+00 -3.90679747e-01 -1.10043675e-01
-7.26842403e-01 4.77234662e-01 8.85595202e-01 1.26462841e+00
9.26964998e-01 2.06161618e-01 -8.60228240e-02 4.62771326e-01
7.67353177e-02 7.88366497e-01 -4.98466134e-01 -1.46169960e+00
1.98460624e-01 7.62071490e-01 2.29793876e-01 -1.06402683e+00
-1.11495115e-01 -1.46346083e-02 -9.65580881e-01 1.08489358e+00
3.18548262e-01 -1.17300451e-01 -9.69203293e-01 1.71929252e+00
4.32673365e-01 -1.05372414e-01 -4.23925072e-01 7.23421454e-01
6.94289446e-01 5.23577988e-01 -1.45512179e-01 3.58856440e-01
9.61855888e-01 -3.04984629e-01 -3.79726470e-01 3.08452159e-01
1.77896798e-01 -1.01032102e+00 1.24998999e+00 2.83929020e-01
-1.72796178e+00 -5.91494858e-01 -1.11055160e+00 -4.61993039e-01
-3.65776986e-01 -5.65225743e-02 4.26232010e-01 2.59103239e-01
-1.25470376e+00 1.05289137e+00 -7.27182388e-01 -2.96191692e-01
4.38424259e-01 3.15978289e-01 -3.97385627e-01 4.11430120e-01
-8.27944994e-01 8.15517485e-01 -1.89172238e-01 -3.94221723e-01
-7.49053895e-01 -1.13463461e+00 -7.67055511e-01 -1.08081020e-01
8.88487697e-02 -1.34567809e+00 1.03222346e+00 -9.65177715e-01
-2.07473493e+00 1.35906971e+00 8.60471502e-02 3.81976217e-01
7.48951793e-01 -2.88220137e-01 -2.57883072e-01 1.25559941e-01
-1.44924715e-01 8.87949705e-01 1.26219904e+00 -1.78124642e+00
1.34579130e-02 -1.19216174e-01 3.49967331e-01 8.25973675e-02
2.58621424e-01 -3.76132160e-01 -3.94642562e-01 -1.16995013e+00
4.74890806e-02 -9.77524281e-01 -1.23554416e-01 7.51229525e-01
-7.87503242e-01 5.72229996e-02 9.88292336e-01 -2.36294806e-01
6.89882517e-01 -2.20654798e+00 8.25874865e-01 5.32986879e-01
5.29165149e-01 -1.90675274e-01 -3.20058614e-01 3.81641150e-01
-8.71661007e-02 4.78767812e-01 -2.13178262e-01 -4.91595507e-01
2.46774718e-01 1.25432655e-01 -4.22941446e-01 2.68641084e-01
6.75511420e-01 7.21402526e-01 -9.15963173e-01 -4.43847239e-01
2.41575509e-01 6.36163890e-01 -1.04329777e+00 4.20979023e-01
-8.34543943e-01 8.44992220e-01 -4.81593907e-01 4.89648819e-01
6.70236051e-01 -3.23474526e-01 -4.24794368e-02 -3.87476355e-01
-1.10077269e-01 9.67043117e-02 -1.24825418e+00 2.25821447e+00
-5.57738781e-01 4.89273638e-01 4.16688800e-01 -7.91524798e-02
1.17608607e+00 1.87232122e-01 4.26403522e-01 -5.04948311e-02
2.20690683e-01 1.53888553e-01 -7.24863529e-01 -3.24087501e-01
5.13284266e-01 -1.52377009e-01 -2.06247211e-01 7.66675830e-01
5.88833317e-02 -1.04112947e+00 -2.07805127e-01 2.57854760e-01
9.66212451e-01 6.58241987e-01 -9.45804641e-02 -6.28496230e-01
4.63935025e-02 -2.16597974e-01 5.22291511e-02 4.56137300e-01
5.11585534e-01 7.78528392e-01 7.58580029e-01 -5.01087368e-01
-1.60220909e+00 -1.60046089e+00 -1.10213272e-01 7.10627496e-01
2.30766535e-01 -3.02544504e-01 -6.26816869e-01 -4.39329624e-01
3.28944892e-01 5.15282691e-01 -9.52395976e-01 4.07276377e-02
-7.77547717e-01 2.23589078e-01 4.07229513e-01 2.29937673e-01
5.67284562e-02 -7.62690067e-01 -4.08993065e-01 -2.96743568e-02
5.32921195e-01 -7.95143843e-01 -9.14080143e-01 -8.24173614e-02
-9.99752939e-01 -1.01830447e+00 -6.08787358e-01 -7.81524181e-01
8.97187114e-01 -1.66823521e-01 1.64970672e+00 1.47766888e-01
-1.28659472e-01 5.58516383e-01 9.16049555e-02 -1.04380049e-01
-9.37653840e-01 1.98976606e-01 -1.48679689e-01 -1.23464040e-01
-4.82921243e-01 -1.08300436e+00 -5.89615822e-01 1.45900056e-01
-1.03341806e+00 6.43921077e-01 1.70969337e-01 4.72109556e-01
7.36214936e-01 -5.78102708e-01 8.06418210e-02 -8.90410423e-01
5.45679212e-01 -1.88497201e-01 -6.18856966e-01 6.72512427e-02
-1.49724796e-01 4.14684445e-01 5.28503954e-01 -6.37594759e-01
-8.54869187e-01 4.07871269e-02 8.55392888e-02 -6.86677158e-01
-3.27286303e-01 -1.56048641e-01 -3.44211996e-01 -1.58677340e-01
7.88349032e-01 -3.69938195e-01 7.99212903e-02 -7.41142571e-01
9.03904080e-01 6.43389001e-02 5.63147128e-01 -1.15548706e+00
1.27918935e+00 5.53192437e-01 4.92920071e-01 -5.45519114e-01
-6.34301364e-01 5.59053540e-01 -9.76863086e-01 -1.86158150e-01
9.69076872e-01 -5.76765716e-01 -7.59907365e-01 2.75442749e-01
-1.47594368e+00 -6.82047963e-01 -6.18841410e-01 -9.93849486e-02
-9.60652232e-01 -6.20990582e-02 -6.85043752e-01 -2.76280373e-01
-2.11927921e-01 -1.11415219e+00 1.46183264e+00 -1.30251437e-01
-7.00279772e-01 -1.06972194e+00 1.89907849e-01 -3.43364775e-01
4.85359311e-01 1.06841040e+00 1.50342321e+00 2.45012254e-01
-8.70109618e-01 6.32128343e-02 -1.69364750e-01 -1.99480746e-02
3.72019798e-01 6.09399259e-01 -8.05520475e-01 -6.97460771e-02
-3.53346467e-01 -3.34697872e-01 2.91708291e-01 5.99477626e-02
1.44664693e+00 -3.35737377e-01 -1.06949650e-01 1.03187299e+00
1.28293502e+00 -1.04252242e-01 3.05209756e-01 -1.75211638e-01
1.10765743e+00 4.54719245e-01 -3.38997036e-01 2.88005382e-01
1.15039028e-01 6.55217350e-01 3.48318875e-01 -3.19980472e-01
-3.43055040e-01 -5.79831839e-01 1.79886818e-02 6.95996106e-01
-2.43851930e-01 -1.46560773e-01 -5.96119583e-01 5.78911789e-02
-1.21171510e+00 -6.66953087e-01 4.20811325e-02 1.89549088e+00
1.07868111e+00 2.49116331e-01 -1.16350930e-02 -9.26538333e-02
4.93481636e-01 2.51818031e-01 -8.56104374e-01 -6.72328472e-01
-4.85943407e-02 6.83150291e-01 5.37199900e-02 7.36397624e-01
-8.51388514e-01 9.17774737e-01 7.10053635e+00 4.75406021e-01
-1.29059589e+00 -3.39868158e-01 3.56428832e-01 -2.05179259e-01
-1.01862764e+00 -1.69665828e-01 -1.48026571e-01 1.46520734e-01
2.56240457e-01 -1.45340458e-01 7.12002754e-01 6.55147374e-01
8.36935937e-02 4.37289238e-01 -1.66468096e+00 9.21824753e-01
-1.37986481e-01 -1.81727815e+00 6.55415893e-01 8.25141557e-03
8.24791908e-01 -4.93811429e-01 3.29280823e-01 -1.13374360e-01
7.37858713e-01 -1.18489182e+00 1.02358913e+00 8.14478517e-01
1.55067480e+00 -5.89804769e-01 -2.87297994e-01 -1.66028246e-01
-9.20468330e-01 8.53301525e-01 1.78775609e-01 1.43435493e-01
9.63983685e-02 4.50056970e-01 -2.77652323e-01 3.25521708e-01
3.68436933e-01 7.70602942e-01 -3.74086380e-01 5.36659300e-01
-1.60123482e-01 2.50150144e-01 -2.56072283e-01 2.02061698e-01
-2.64591333e-02 -3.17815006e-01 9.15317953e-01 9.78113174e-01
3.31909955e-01 2.46448055e-01 1.78556040e-01 1.47522855e+00
-4.77483153e-01 -4.39485498e-02 -7.45181024e-01 -4.75094318e-02
3.15192699e-01 1.01404083e+00 -6.07631743e-01 -2.80029267e-01
9.55355614e-02 1.23159480e+00 3.23379785e-01 5.83320439e-01
-4.78416562e-01 -2.39859357e-01 9.69275653e-01 3.03793341e-01
4.04458232e-02 -6.13261402e-01 -9.42005157e-01 -1.21741736e+00
-2.74304479e-01 -7.27356195e-01 -3.42538476e-01 -1.26495314e+00
-1.44426084e+00 5.40315270e-01 -7.48389661e-02 -1.20742464e+00
1.05511457e-01 -7.20229268e-01 -7.06943929e-01 1.03688705e+00
-9.13092315e-01 -1.14584816e+00 -2.53498971e-01 4.62886274e-01
2.23938048e-01 1.42790183e-01 1.08132887e+00 -1.27686635e-01
-1.25070006e-01 3.70987713e-01 -2.91321456e-01 -5.37521914e-02
6.50699556e-01 -1.61178958e+00 8.48898768e-01 2.71439135e-01
-1.19864501e-01 4.66652840e-01 9.00747716e-01 -7.03409612e-01
-1.73554409e+00 -1.07854903e+00 4.46504921e-01 -8.03267181e-01
4.99374360e-01 -5.63121915e-01 -7.86217213e-01 6.75353527e-01
2.66231745e-01 -2.20918462e-01 5.55439234e-01 5.65381488e-03
-5.92971265e-01 2.56123602e-01 -1.24734545e+00 1.04933012e+00
1.43347228e+00 -7.23711252e-01 -4.93380189e-01 6.87559992e-02
8.39329422e-01 -9.39194024e-01 -1.34014571e+00 1.94927365e-01
6.25063002e-01 -7.04867661e-01 9.00651872e-01 -8.31361175e-01
8.55827153e-01 -3.63261908e-01 1.17035247e-02 -1.59539533e+00
-5.66919386e-01 -1.02361667e+00 -2.18184814e-01 8.35806847e-01
2.62588471e-01 -1.08545147e-01 7.05629945e-01 6.83375716e-01
-2.17230275e-01 -8.06129634e-01 -6.01458251e-01 -4.71109986e-01
3.83716345e-01 -1.82117343e-01 1.01541054e+00 1.01814032e+00
-3.20412904e-01 1.93128169e-01 -2.09647924e-01 1.19829312e-01
7.84678340e-01 4.40593451e-01 1.07648313e+00 -1.47302210e+00
-1.49943337e-01 -7.17809975e-01 -1.61859661e-01 -1.08678830e+00
1.07299633e-01 -1.09845328e+00 -3.92673761e-01 -1.22490513e+00
-2.17947379e-01 -6.94959939e-01 3.53747010e-01 2.33521998e-01
1.03643477e-01 2.84303546e-01 2.93029547e-01 1.04963049e-01
-8.44568200e-03 7.88182199e-01 2.02887869e+00 -1.22212820e-01
-2.90147781e-01 -3.25197816e-01 -5.23688912e-01 8.89865160e-01
6.86226904e-01 -2.68166453e-01 -2.03784108e-01 -8.19114149e-01
5.52315593e-01 2.59038150e-01 6.30719423e-01 -6.44475043e-01
-3.95295650e-01 -3.13141346e-01 6.88249886e-01 -3.85996044e-01
4.10131514e-01 -7.61099279e-01 6.46355748e-01 3.73208039e-02
-6.38681054e-01 2.83319086e-01 2.87505865e-01 3.31768662e-01
3.23155195e-01 2.33296439e-01 9.05006766e-01 -1.32876888e-01
-2.18220670e-02 6.25875235e-01 -2.38933433e-02 1.94716558e-01
7.19274282e-01 -1.92894876e-01 -6.46662787e-02 -1.34616002e-01
-9.75803256e-01 -1.03830345e-01 1.44694901e+00 5.51038027e-01
4.64012921e-01 -1.83649361e+00 -4.63186741e-01 5.46008825e-01
-2.11685654e-02 2.98462003e-01 -8.39142948e-02 -5.93859628e-02
-8.18339467e-01 -3.63543779e-01 -4.07399237e-01 -5.90228438e-01
-7.09153950e-01 4.46426630e-01 4.21913028e-01 2.40440723e-02
-8.32476199e-01 5.68741560e-01 1.67552263e-01 -5.01083851e-01
1.06539177e-02 -6.18937790e-01 4.37950432e-01 -3.76939654e-01
1.34366617e-01 2.25697845e-01 -3.31539452e-01 -3.13741893e-01
1.26965910e-01 9.74308431e-01 4.23666358e-01 -2.62982279e-01
1.27035618e+00 1.74250498e-01 -1.74137071e-01 6.47803783e-01
1.25339806e+00 2.90472806e-01 -1.72416842e+00 1.13793975e-02
-4.45734233e-01 -4.21000183e-01 -1.81704432e-01 -7.49165893e-01
-1.04593349e+00 8.23539793e-01 3.05283040e-01 2.84131348e-01
5.99050403e-01 -8.11737403e-03 7.96514511e-01 9.12440568e-02
4.36104506e-01 -7.18092799e-01 3.74093294e-01 4.98608470e-01
1.46239090e+00 -7.91082680e-01 -1.71034098e-01 -4.54206407e-01
-2.42236778e-01 1.36363161e+00 4.72639769e-01 -7.03452289e-01
9.69473600e-01 4.12687421e-01 6.26740456e-02 -5.33704579e-01
-5.24839282e-01 3.13892126e-01 6.46628976e-01 4.40275311e-01
2.77929723e-01 1.85573921e-01 3.37161779e-01 -1.72892492e-02
-6.49040341e-01 -4.04422730e-03 4.43785071e-01 6.80934668e-01
-1.72043160e-01 -1.16040838e+00 -3.27733159e-01 3.06833059e-01
-5.00888415e-02 2.53092021e-01 -5.50833225e-01 6.60895050e-01
2.98300445e-01 4.21689093e-01 2.05404431e-01 -4.23861146e-01
5.69175839e-01 4.33837883e-02 8.99879336e-01 -6.52564764e-01
-4.67633158e-01 -2.93347929e-02 -1.30695507e-01 -8.19295049e-01
-4.11759883e-01 -3.88802439e-01 -8.99952173e-01 -6.45425200e-01
3.37464839e-01 -2.69757628e-01 5.09753406e-01 4.65806276e-01
7.02312350e-01 3.44679117e-01 7.39381731e-01 -1.53411496e+00
-1.18815966e-01 -6.19898975e-01 -6.07166171e-01 8.20198894e-01
4.25722837e-01 -8.57736170e-01 -2.99567699e-01 4.15802091e-01] | [9.066080093383789, -3.5486972332000732] |
bb7d8c64-8492-4eb0-9128-c718d3e79ad2 | deep-shape-analysis-on-abdominal-organs-for | 1808.01946 | null | http://arxiv.org/abs/1808.01946v1 | http://arxiv.org/pdf/1808.01946v1.pdf | Deep Shape Analysis on Abdominal Organs for Diabetes Prediction | Morphological analysis of organs based on images is a key task in medical
imaging computing. Several approaches have been proposed for the quantitative
assessment of morphological changes, and they have been widely used for the
analysis of the effects of aging, disease and other factors in organ
morphology. In this work, we propose a deep neural network for predicting
diabetes on abdominal shapes. The network directly operates on raw point clouds
without requiring mesh processing or shape alignment. Instead of relying on
hand-crafted shape descriptors, an optimal representation is learned in the
end-to-end training stage of the network. For comparison, we extend the
state-of-the-art shape descriptor BrainPrint to the AbdomenPrint. Our results
demonstrate that the network learns shape representations that better separates
healthy and diabetic individuals than traditional representations. | ['Sergios Gatidis', 'Benjamin Gutierrez-Becker', 'Christopher Schlett Fabian Bamberg', 'Annette Peters', 'Daniel Gutmann', 'Christian Wachinger'] | 2018-08-06 | null | null | null | null | ['diabetes-prediction'] | ['medical'] | [-5.41228950e-02 -1.27692893e-01 -6.73585236e-02 -6.97056293e-01
-3.22983056e-01 -1.96659669e-01 2.10007623e-01 8.75173390e-01
-4.57737625e-01 1.58866018e-01 -3.25612316e-04 -7.34513253e-02
-1.52916312e-01 -1.06280363e+00 -5.15099704e-01 -5.27864695e-01
-2.54354179e-01 7.46290684e-01 -2.22368374e-01 -2.52697486e-02
-8.77014175e-03 8.16083252e-01 -9.43475783e-01 1.41384095e-01
7.86496282e-01 1.21910107e+00 -2.97841281e-01 3.72075528e-01
-2.11909294e-01 -1.57834147e-03 -2.45022327e-01 -4.74195987e-01
2.40589768e-01 -1.75158039e-01 -3.30233604e-01 9.41134393e-02
6.58046067e-01 -2.66650915e-01 -3.83100510e-01 1.06855881e+00
7.81496167e-01 -2.03743681e-01 9.71342444e-01 -6.95335507e-01
-8.27954829e-01 3.49508703e-01 -4.64264631e-01 1.67020306e-01
-1.00649120e-02 -1.37805969e-01 5.75488687e-01 -8.31473768e-01
2.34966159e-01 1.23669267e+00 1.09545302e+00 2.06078038e-01
-1.25896299e+00 -6.34833694e-01 -1.88839272e-01 -2.21348837e-01
-1.32832456e+00 -3.01199168e-01 8.24615002e-01 -6.63025558e-01
3.13973755e-01 8.38745683e-02 8.64897907e-01 7.19235241e-01
4.93020028e-01 5.17551184e-01 7.09118724e-01 -3.56653661e-01
6.27254173e-02 -3.58796388e-01 -6.03689291e-02 1.01589751e+00
5.81699848e-01 1.04858227e-01 1.03417441e-01 -2.88560539e-01
1.28814948e+00 3.79602671e-01 2.77505498e-02 -6.63689673e-01
-1.05732012e+00 8.44949186e-01 6.04549885e-01 3.08296055e-01
-5.43987036e-01 2.07244903e-01 5.46329796e-01 6.11916482e-02
6.79082930e-01 3.34490746e-01 -4.30124521e-01 3.96514475e-01
-1.00432289e+00 1.13777973e-01 4.22031343e-01 3.63330215e-01
2.09465846e-01 -3.06796134e-02 -2.01607332e-01 1.01517034e+00
3.35412711e-01 5.01118481e-01 4.98477101e-01 -6.50753081e-01
4.41335961e-02 1.04135633e+00 -3.87671560e-01 -1.15035975e+00
-8.52467060e-01 -4.87044722e-01 -1.26785088e+00 3.52207065e-01
5.98850250e-01 2.07571834e-02 -1.05011690e+00 1.33462250e+00
4.08717036e-01 -1.13680087e-01 -3.05828243e-01 7.94989824e-01
1.05453110e+00 8.81872792e-03 2.10769534e-01 2.94806510e-01
1.25855160e+00 -5.31820357e-01 -2.98418552e-01 2.11242333e-01
3.31933349e-01 -4.33018029e-01 6.68743312e-01 3.79595459e-01
-1.16891754e+00 -4.91501153e-01 -8.91726971e-01 -1.44903079e-01
-4.13393170e-01 5.51079035e-01 6.53565228e-01 4.91908908e-01
-8.43414187e-01 9.90389824e-01 -1.15771508e+00 -4.47931647e-01
7.82111883e-01 5.74282944e-01 -4.03600544e-01 -3.21257929e-03
-6.79982901e-01 8.17758143e-01 1.68627515e-01 1.19551703e-01
-3.77171248e-01 -8.42405319e-01 -9.04242337e-01 1.57984763e-01
-2.00898841e-01 -9.56052184e-01 7.84900248e-01 -8.42846751e-01
-1.39918458e+00 1.16066909e+00 2.74560183e-01 -2.72382379e-01
6.13496780e-01 -1.32202849e-01 -2.14932233e-01 3.67378145e-01
-2.64222115e-01 5.00612915e-01 7.14594603e-01 -1.09580481e+00
5.81036694e-02 -8.08661401e-01 -2.54813462e-01 -1.15705334e-01
-3.67324680e-01 5.98888025e-02 -3.07927102e-01 -9.56589580e-01
3.15609872e-01 -9.18687642e-01 -4.07916665e-01 8.78857136e-01
-1.81662127e-01 -1.65561721e-01 3.08681279e-01 -8.57830584e-01
7.43178666e-01 -2.14596868e+00 6.84058294e-02 5.02932847e-01
4.75131929e-01 2.98014283e-01 -2.02919051e-01 9.80928093e-02
-6.59445003e-02 1.95383012e-01 -1.10338315e-01 -2.54901767e-01
-1.39418453e-01 7.56617188e-02 1.96385175e-01 8.53060722e-01
1.54730767e-01 9.25786257e-01 -6.76340520e-01 -6.35738075e-01
2.75263667e-01 7.61640251e-01 -4.66200113e-01 -5.33552095e-02
1.58620700e-01 5.21503747e-01 -4.67783332e-01 8.22356820e-01
6.77493393e-01 -3.14373940e-01 2.85643488e-01 -3.46290171e-01
4.82743047e-02 -1.64736256e-01 -7.44158685e-01 1.77801764e+00
-3.24807197e-01 1.99137226e-01 1.15109980e-01 -1.09312713e+00
1.11748290e+00 4.61639971e-01 9.33529615e-01 -6.81995511e-01
5.68240106e-01 2.00605154e-01 1.80577725e-01 -3.31514508e-01
-2.82157212e-01 -2.69899011e-01 1.59522772e-01 1.69842601e-01
8.99980068e-02 9.29913223e-02 1.34240374e-01 -5.16821623e-01
7.84489155e-01 -2.83226911e-02 4.97178793e-01 -3.55873674e-01
6.11011147e-01 -2.45921686e-01 5.02942383e-01 1.35087863e-01
-2.11079165e-01 7.74783254e-01 2.95919627e-01 -1.06123519e+00
-9.92104948e-01 -1.21704948e+00 -5.03485262e-01 9.19200122e-01
1.11268777e-02 -6.69105202e-02 -4.53239977e-01 -6.29946351e-01
5.92863083e-01 1.21151447e-01 -9.12654459e-01 -1.86968714e-01
-6.81777179e-01 -7.50698626e-01 7.13511527e-01 9.20456767e-01
2.59750336e-01 -9.52222824e-01 -5.59994221e-01 1.75686002e-01
1.94992110e-01 -7.09798634e-01 -4.18754667e-01 -2.97328178e-02
-1.38471138e+00 -1.16455352e+00 -1.02167094e+00 -7.50324667e-01
1.22581089e+00 -2.51002699e-01 1.07422280e+00 4.08624887e-01
-6.24911249e-01 3.59461546e-01 -1.27134964e-01 -7.02685893e-01
-3.40476841e-01 -2.27727871e-02 3.87338251e-02 2.47141849e-02
1.06418490e-01 -7.15335011e-01 -9.31090653e-01 1.22320622e-01
-7.89220154e-01 -2.34604076e-01 9.97786403e-01 6.33710086e-01
9.18219149e-01 -2.19855785e-01 3.59395117e-01 -6.27911985e-01
5.35649836e-01 -1.87755391e-01 -4.25427169e-01 3.20160389e-01
-5.23866057e-01 1.52194068e-01 5.29874861e-01 -5.35312891e-01
-4.23229665e-01 2.33034834e-01 -2.47610673e-01 -7.09277153e-01
-3.17863017e-01 6.39561117e-01 2.73123533e-01 -4.92233574e-01
5.10609090e-01 8.00540000e-02 4.16665614e-01 -8.13156426e-01
1.13884229e-02 2.30136603e-01 5.54996312e-01 -6.21854365e-01
5.73276639e-01 5.28896987e-01 4.71516639e-01 -7.83481121e-01
-6.40183330e-01 -4.27855909e-01 -1.16243780e+00 -1.53178675e-02
9.74796891e-01 -6.92457438e-01 -6.33405507e-01 4.05251712e-01
-9.09474671e-01 -2.09327176e-01 -2.41262719e-01 5.44710577e-01
-4.90024209e-01 4.58546281e-01 -6.00324690e-01 -3.43407065e-01
-8.86056185e-01 -8.96917760e-01 1.09484446e+00 1.20284051e-01
-1.63322404e-01 -1.29188526e+00 3.04904342e-01 2.05282737e-02
4.70957816e-01 8.73520076e-01 1.41019928e+00 -8.02378595e-01
1.03194930e-01 -4.71284211e-01 -3.66097063e-01 2.99724936e-01
1.99158728e-01 3.00658792e-02 -4.57061768e-01 -4.51214820e-01
-3.01303446e-01 -2.81854451e-01 8.33157837e-01 7.66958475e-01
1.45639849e+00 -2.21726790e-01 -3.37961465e-01 9.72268939e-01
1.48271894e+00 7.48524722e-03 4.14780855e-01 9.68096778e-02
5.59612334e-01 4.21611190e-01 8.70957151e-02 4.34915781e-01
2.94577330e-01 3.26383024e-01 5.18838823e-01 -6.70288801e-01
-2.01255396e-01 8.56071189e-02 -2.19479710e-01 6.30657971e-01
-3.79168957e-01 2.97900856e-01 -1.12669063e+00 3.48938048e-01
-1.47013116e+00 -5.47238648e-01 -9.50306579e-02 2.29348707e+00
7.68399537e-01 -6.90572634e-02 1.73379838e-01 -9.96934101e-02
5.83606660e-01 -2.00996563e-01 -6.56518579e-01 -2.35899299e-01
3.53086323e-01 5.55346131e-01 3.81472826e-01 -1.62649438e-01
-1.35520899e+00 3.18354219e-01 6.86107302e+00 1.48231462e-01
-1.47956550e+00 -2.33517587e-01 6.44460499e-01 1.72873005e-01
2.05470562e-01 -6.67837560e-01 -2.93866873e-01 2.46570989e-01
4.87129301e-01 1.08217731e-01 9.88068432e-02 7.86777198e-01
6.85850754e-02 2.05882952e-01 -1.31043315e+00 1.03976226e+00
6.38642013e-02 -9.72515941e-01 1.71253700e-02 1.45282164e-01
3.19790453e-01 -2.60970574e-02 9.37650725e-02 1.42130747e-01
-1.35159969e-01 -1.30147898e+00 5.45093536e-01 8.04920793e-01
1.05297637e+00 -6.63389266e-01 8.49663317e-01 5.56469075e-02
-1.24749887e+00 1.78393666e-02 -4.95186776e-01 1.50605395e-01
-2.24966615e-01 6.42292738e-01 -8.54661107e-01 4.60177392e-01
4.69662845e-01 5.36693096e-01 -8.34670246e-01 1.50207400e+00
1.07827842e-01 3.93652350e-01 -2.29612753e-01 3.31555992e-01
-2.11095288e-01 -4.27444845e-01 1.36829004e-01 1.23743975e+00
4.53831851e-01 9.86533761e-02 5.44125319e-01 9.94744480e-01
-8.39020237e-02 4.15578693e-01 -4.99150127e-01 -2.10642323e-01
-1.11738101e-01 1.42692888e+00 -8.47700477e-01 -3.15810114e-01
-3.59538019e-01 5.40738285e-01 1.39248431e-01 -1.53652709e-02
-4.96783942e-01 -2.50915587e-01 6.04990244e-01 3.90744656e-01
2.20690548e-01 -3.33456308e-01 -6.21015310e-01 -1.03531420e+00
-2.70464450e-01 -6.80876553e-01 5.11638999e-01 -3.74859899e-01
-1.76271832e+00 2.09808439e-01 -1.56624645e-01 -1.11031449e+00
1.44565940e-01 -8.31189811e-01 -9.29131269e-01 7.73295283e-01
-1.46783686e+00 -1.38398814e+00 -4.46049809e-01 4.58443373e-01
1.50307834e-01 5.06260083e-04 1.17100990e+00 4.10289079e-01
-4.83395547e-01 6.41820550e-01 1.02485299e-01 8.28708887e-01
6.90382838e-01 -1.35676074e+00 1.84256151e-01 3.17558616e-01
-1.38469800e-01 7.19291449e-01 2.59069860e-01 -7.89250493e-01
-1.20170200e+00 -1.18149781e+00 6.47523701e-01 -3.03547308e-02
2.48621240e-01 1.07147023e-01 -8.54994655e-01 3.68794173e-01
-9.18813050e-02 5.08087814e-01 9.93795037e-01 1.67264730e-01
-2.43161470e-01 -1.89302593e-01 -1.25306582e+00 2.22993270e-01
7.80863523e-01 7.06763240e-03 -5.62462687e-01 9.80937555e-02
1.52702527e-02 -4.74569738e-01 -1.28464758e+00 7.20517457e-01
9.74700570e-01 -6.22572422e-01 1.43571544e+00 -7.40492284e-01
4.28687423e-01 1.36254989e-02 1.98701963e-01 -1.25582325e+00
-6.90240622e-01 1.13250397e-01 -8.17179456e-02 7.38375962e-01
6.31484538e-02 -3.66763562e-01 7.95550466e-01 4.60430890e-01
-9.81355235e-02 -1.09689581e+00 -7.21070707e-01 -6.59954071e-01
5.00511587e-01 1.55395910e-01 4.59228814e-01 1.02044427e+00
-3.11759084e-01 -7.63796940e-02 2.90568024e-01 1.70983687e-01
7.28016615e-01 3.21185261e-01 5.47831237e-01 -2.06040764e+00
2.26561967e-02 -7.10062921e-01 -5.74686825e-01 -2.24299803e-01
-4.14919518e-02 -1.31908822e+00 -1.91779643e-01 -1.81110442e+00
2.55070239e-01 -4.71486360e-01 -6.17307246e-01 6.93134069e-01
-3.81110571e-02 3.18497747e-01 1.42978206e-01 3.88231501e-02
-2.61626571e-01 4.21664923e-01 1.47520518e+00 -3.75289619e-01
-1.34803966e-01 1.55951977e-01 -4.99094546e-01 9.58741069e-01
1.00240326e+00 -4.15525079e-01 -1.81002878e-02 -3.80405694e-01
1.38607308e-01 -5.65067865e-02 5.30784070e-01 -1.12747514e+00
-1.28204033e-01 -7.72352442e-02 1.20096791e+00 -6.36162162e-01
1.08412638e-01 -8.56574297e-01 -4.66180369e-02 8.23077023e-01
-2.96769977e-01 2.23764718e-01 1.57490343e-01 3.49498630e-01
-8.64639040e-03 -1.54657766e-01 1.02408051e+00 -3.33825529e-01
4.62453626e-02 6.50705755e-01 -8.94556493e-02 -4.74117622e-02
7.64254808e-01 -9.68432352e-02 1.17090285e-01 -4.20908965e-02
-8.95487487e-01 2.12877300e-02 4.62562501e-01 1.27213225e-01
7.97002733e-01 -1.54597950e+00 -7.52281249e-01 1.69950932e-01
2.47003771e-02 1.13915861e-01 -3.19896378e-02 1.18288028e+00
-8.70383978e-01 1.87266573e-01 -6.18618250e-01 -5.70934474e-01
-1.32780969e+00 5.61991870e-01 5.11739373e-01 -5.34968853e-01
-7.53284633e-01 5.75273454e-01 2.82862693e-01 -5.59706032e-01
2.63359159e-01 -8.30765367e-01 -4.03764933e-01 8.43720883e-02
3.24889094e-01 4.97008741e-01 1.79282308e-01 -4.74865884e-01
-2.38203734e-01 8.73090088e-01 2.43118450e-01 4.38468009e-01
1.69423676e+00 2.31528088e-01 -2.76723355e-01 2.58387327e-01
1.15016580e+00 -6.37798980e-02 -7.76914597e-01 -2.50936687e-01
-1.69590577e-01 -1.94677204e-01 1.37917608e-01 -8.94345939e-01
-1.35893571e+00 1.13120711e+00 9.22532737e-01 -8.81152824e-02
1.02411890e+00 -2.24286273e-01 9.84224796e-01 3.91000420e-01
1.76897749e-01 -8.23552847e-01 2.17988640e-01 2.67705500e-01
1.07924938e+00 -1.21716940e+00 4.19632971e-01 -3.45469058e-01
-2.87201107e-01 1.74948716e+00 3.58484775e-01 -4.26948696e-01
7.95820534e-01 2.18855590e-01 1.92280278e-01 -3.50166172e-01
-1.19397670e-01 -1.52576998e-01 8.36771071e-01 6.09409392e-01
6.51324034e-01 1.65181324e-01 -4.59182739e-01 7.37292945e-01
1.97702218e-02 1.45867959e-01 -2.91556045e-02 8.25682521e-01
-3.05330783e-01 -1.16723704e+00 -4.75012064e-01 7.56766498e-01
-8.11600804e-01 1.01384096e-01 -5.08299768e-01 7.39943564e-01
3.31949025e-01 2.15302229e-01 1.59809649e-01 -4.34897579e-02
3.11623633e-01 2.16704696e-01 7.66408205e-01 -6.14797235e-01
-6.86094582e-01 9.21136886e-02 -2.38628268e-01 -4.32115763e-01
-2.34143332e-01 -7.41985440e-01 -1.42217422e+00 4.12043370e-02
-5.74808903e-02 -4.86643672e-01 7.55562961e-01 6.43352151e-01
3.67312491e-01 5.48912227e-01 3.31061393e-01 -1.02290642e+00
-6.09319508e-01 -9.75951493e-01 -5.39809525e-01 6.48675382e-01
2.22226888e-01 -8.00597847e-01 2.38156900e-01 1.48397222e-01] | [14.211413383483887, -2.4599101543426514] |
1779cdc9-7d11-4eee-b7e0-1ce38b831e01 | dialogue-based-relation-extraction | 2004.08056 | null | https://arxiv.org/abs/2004.08056v1 | https://arxiv.org/pdf/2004.08056v1.pdf | Dialogue-Based Relation Extraction | We present the first human-annotated dialogue-based relation extraction (RE) dataset DialogRE, aiming to support the prediction of relation(s) between two arguments that appear in a dialogue. We further offer DialogRE as a platform for studying cross-sentence RE as most facts span multiple sentences. We argue that speaker-related information plays a critical role in the proposed task, based on an analysis of similarities and differences between dialogue-based and traditional RE tasks. Considering the timeliness of communication in a dialogue, we design a new metric to evaluate the performance of RE methods in a conversational setting and investigate the performance of several representative RE methods on DialogRE. Experimental results demonstrate that a speaker-aware extension on the best-performing model leads to gains in both the standard and conversational evaluation settings. DialogRE is available at https://dataset.org/dialogre/. | ['Kai Sun', 'Dian Yu', 'Dong Yu', 'Claire Cardie'] | 2020-04-17 | dialogue-based-relation-extraction-1 | https://aclanthology.org/2020.acl-main.444 | https://aclanthology.org/2020.acl-main.444.pdf | acl-2020-6 | ['dialog-relation-extraction'] | ['natural-language-processing'] | [ 4.53137197e-02 5.21383822e-01 -2.07602337e-01 -5.03430068e-01
-8.97839010e-01 -7.44733334e-01 1.35379553e+00 4.24842715e-01
-4.60303128e-01 7.78149247e-01 8.87341797e-01 -4.41722035e-01
-9.37032476e-02 -3.84845704e-01 -1.00913003e-01 -1.28621869e-02
3.52718458e-02 6.92512155e-01 2.99124748e-01 -9.48488176e-01
1.64811879e-01 1.72057316e-01 -1.02310336e+00 6.25482857e-01
5.76895535e-01 7.12066650e-01 -8.40748399e-02 1.00347698e+00
-2.12964877e-01 1.12843013e+00 -9.97178495e-01 -7.46316552e-01
-9.78340879e-02 -6.99236035e-01 -1.79983044e+00 4.13600393e-02
1.61070302e-01 -1.79130852e-01 -1.52740955e-01 5.52909553e-01
4.41104621e-01 2.64606416e-01 5.32652259e-01 -1.09706736e+00
2.14352310e-02 1.24256587e+00 8.75390619e-02 5.02183259e-01
1.06921947e+00 -7.37304538e-02 1.52179921e+00 -7.50318766e-01
8.55221450e-01 1.49658680e+00 4.65643823e-01 7.55748808e-01
-1.20098150e+00 -4.92448807e-02 5.91103248e-02 1.04418628e-01
-7.55557835e-01 -9.73730624e-01 8.47561717e-01 -3.15709800e-01
1.17501283e+00 6.70225024e-01 3.31886768e-01 1.44392085e+00
-1.70166358e-01 7.74101913e-01 1.24729526e+00 -5.89349270e-01
-1.01413853e-01 2.99364507e-01 6.17765188e-01 3.88221622e-01
-5.49130082e-01 -4.11160290e-01 -7.97781706e-01 -4.16444272e-01
5.27799763e-02 -8.24538350e-01 -5.12078106e-01 9.05711278e-02
-1.04937458e+00 1.09053290e+00 9.75947157e-02 5.69896340e-01
-8.66914988e-02 -5.40279984e-01 8.28210950e-01 6.08690321e-01
7.32115686e-01 8.54358017e-01 -6.45566702e-01 -5.97369134e-01
-3.70070994e-01 4.08338845e-01 1.63347030e+00 7.00911522e-01
1.68067396e-01 -7.58098245e-01 -4.82065618e-01 1.19782758e+00
-3.98754664e-02 -1.44772381e-01 4.63256419e-01 -9.48151886e-01
8.44856799e-01 7.30061531e-01 1.16929322e-01 -6.89887643e-01
-6.79078341e-01 1.66910905e-02 -5.61358333e-01 -5.36113977e-01
8.47237706e-01 -4.24505770e-01 2.18851194e-01 1.72016346e+00
5.17551601e-01 -3.47442746e-01 5.48657537e-01 5.07258832e-01
1.42597926e+00 4.68137234e-01 -1.24707222e-01 -5.51974416e-01
1.59081388e+00 -1.00916564e+00 -7.91946650e-01 -2.00583249e-01
8.46135795e-01 -8.92436385e-01 1.17113924e+00 9.80763957e-02
-9.73132670e-01 -2.11740240e-01 -7.12684989e-01 -1.80347279e-01
-5.43014742e-02 -1.94423571e-02 3.74599040e-01 4.66388285e-01
-6.75091505e-01 4.93119687e-01 -3.16466361e-01 -4.55251187e-01
-1.39372960e-01 3.14784087e-02 -2.50573307e-01 4.51672435e-01
-1.43573928e+00 1.18400526e+00 2.67673866e-03 6.46739081e-02
-3.40892762e-01 -3.36393356e-01 -8.83349419e-01 -2.91902851e-02
7.68851876e-01 -3.10073525e-01 1.89971006e+00 -5.45135438e-01
-1.90633476e+00 9.92513001e-01 -2.82192916e-01 -7.45756865e-01
7.74696290e-01 -3.83687645e-01 -2.16082364e-01 1.21736132e-01
-1.35822654e-01 2.26401091e-01 1.95711032e-01 -1.04571450e+00
-5.69667935e-01 7.36973211e-02 5.40953934e-01 2.48785183e-01
-1.44339070e-01 3.91972482e-01 2.60064099e-02 -2.55254596e-01
-3.67333651e-01 -1.00173020e+00 4.51752171e-02 -7.62684107e-01
-9.22766387e-01 -9.13631380e-01 5.78848004e-01 -6.01173937e-01
1.37093079e+00 -1.81537390e+00 2.33609140e-01 -1.85488060e-01
3.47611964e-01 3.20709407e-01 -6.64025545e-02 1.15136242e+00
1.18202008e-01 7.86714181e-02 -1.09514743e-01 -5.51414251e-01
5.93731776e-02 2.24752165e-03 -3.48013431e-01 1.20055653e-01
8.80419556e-03 9.61695850e-01 -7.63204277e-01 -4.35095459e-01
8.12506974e-02 1.29118070e-01 -2.24468783e-01 5.81762850e-01
-4.38439846e-01 7.88227499e-01 -4.23861444e-01 -6.90944046e-02
8.29909369e-02 -3.12550098e-01 5.98707855e-01 -1.01476267e-01
-1.31224960e-01 1.36719489e+00 -6.23602509e-01 1.49580514e+00
-8.38536263e-01 8.46245348e-01 5.54601997e-02 -7.07247078e-01
8.61285746e-01 6.57903612e-01 1.58824578e-01 -4.62762654e-01
3.03347409e-01 1.77298170e-02 4.04500484e-01 -5.17118812e-01
7.89468825e-01 3.29994373e-02 -2.84598172e-01 7.02018678e-01
3.99154127e-02 -1.69434622e-01 4.12849396e-01 5.16114712e-01
1.16296828e+00 -3.45796376e-01 6.97959542e-01 -2.47957870e-01
9.54180121e-01 2.42818967e-02 2.55477309e-01 5.60675919e-01
-2.43613809e-01 9.19015557e-02 8.87472391e-01 -1.51071265e-01
-6.22473419e-01 -3.89657438e-01 -1.13641746e-01 1.17478347e+00
3.47999744e-02 -8.19726050e-01 -6.59788072e-01 -1.12728238e+00
-3.52393061e-01 9.18903112e-01 -4.45811599e-01 2.25329638e-01
-7.86092281e-01 -5.04186511e-01 6.76537156e-01 2.31938940e-02
4.84282166e-01 -1.02776945e+00 -5.20077109e-01 2.75198400e-01
-8.16664994e-01 -1.67169058e+00 -3.82969499e-01 7.66000524e-02
-3.58817130e-01 -1.43291926e+00 -3.32140148e-01 -4.92271334e-01
-9.50821787e-02 5.27179651e-02 1.30995750e+00 1.07375227e-01
8.66594017e-02 5.45426309e-01 -7.03781962e-01 -8.62696096e-02
-1.17079127e+00 4.30504531e-01 -3.17800015e-01 -1.78614050e-01
3.21570963e-01 -3.39193434e-01 -3.81092161e-01 4.51939017e-01
-2.47545958e-01 4.44265753e-02 -8.51790085e-02 9.50635076e-01
-5.15174091e-01 -3.60148370e-01 8.44929576e-01 -1.24475372e+00
1.45446169e+00 -3.84778172e-01 -2.67562836e-01 3.61383915e-01
-4.62266803e-01 1.45467490e-01 4.58346725e-01 -2.60769725e-01
-1.21465027e+00 -4.18348849e-01 -2.85205305e-01 5.36355913e-01
-2.68238366e-01 5.40974617e-01 3.18135954e-02 4.10095543e-01
7.34265149e-01 -1.27419606e-01 1.25899673e-01 -5.16484261e-01
4.69999105e-01 9.00161564e-01 2.34838352e-01 -7.65298128e-01
3.20699573e-01 -1.26063213e-01 -1.74713701e-01 -9.15574789e-01
-1.17827153e+00 -7.39177465e-01 -6.12384021e-01 -1.93035141e-01
6.83411717e-01 -6.82161093e-01 -1.19539356e+00 1.73788238e-02
-1.48333049e+00 -5.83148301e-01 -2.23057851e-01 3.14586163e-01
-4.65441734e-01 5.08457661e-01 -9.35197294e-01 -9.66461718e-01
-4.57045317e-01 -9.88790572e-01 7.71955192e-01 1.04728118e-02
-9.99125659e-01 -1.23135686e+00 4.19349253e-01 8.24869990e-01
8.16702321e-02 -4.27862182e-02 7.45129287e-01 -1.51728988e+00
1.69957936e-01 9.08132270e-02 -2.70711184e-02 2.36899570e-01
2.46463716e-01 -1.37604818e-01 -1.09670269e+00 1.34869009e-01
3.37369554e-02 -6.91402197e-01 5.69101393e-01 -1.91264257e-01
3.83445650e-01 -7.10485995e-01 -1.27233326e-01 -4.08755243e-01
6.86941624e-01 -9.36626736e-03 2.89358020e-01 2.83756673e-01
2.97910541e-01 1.25968480e+00 7.54339159e-01 3.92661870e-01
8.21914494e-01 1.07625365e+00 1.69317983e-03 1.95238471e-01
-6.75958395e-02 -1.67156607e-01 4.19710696e-01 8.62371147e-01
7.84672946e-02 -6.80438995e-01 -9.39280689e-01 4.75686729e-01
-1.95989954e+00 -9.17360783e-01 -4.25828338e-01 1.89905465e+00
1.34684324e+00 2.13894159e-01 7.41053462e-01 2.42201537e-01
6.72339082e-01 3.59661430e-01 1.98814154e-01 -7.57104576e-01
-4.46709581e-02 1.49587274e-01 -1.01222768e-01 9.94283915e-01
-1.11326969e+00 9.05041337e-01 5.89002085e+00 6.53218925e-01
-8.29411805e-01 1.77378595e-01 5.74927926e-01 2.89480239e-01
-3.89918536e-02 5.86754759e-04 -8.96982312e-01 1.41097367e-01
1.26120353e+00 -3.04087520e-01 1.28606960e-01 5.51475406e-01
1.88958511e-01 -2.22822353e-01 -1.53380108e+00 4.59241062e-01
-5.70731685e-02 -1.22620189e+00 -2.73258150e-01 -4.98087937e-03
2.08848447e-01 -1.44377217e-01 -4.64868844e-01 3.12541157e-01
4.94564235e-01 -8.61130118e-01 3.01989138e-01 3.85535508e-02
2.25501671e-01 -4.16661710e-01 9.53204036e-01 4.88082379e-01
-9.69349504e-01 1.59395382e-01 2.17500880e-01 -2.07684427e-01
3.95907223e-01 3.44920933e-01 -1.56308091e+00 6.65595651e-01
1.47691756e-01 5.02047181e-01 -4.46171075e-01 4.43596423e-01
-4.96523887e-01 7.65550077e-01 -2.08376616e-01 -3.51671070e-01
8.87108594e-02 7.29971193e-03 9.45987761e-01 1.44912469e+00
-3.06644440e-01 3.22030842e-01 3.49190295e-01 3.77667755e-01
-2.01975033e-01 4.59755272e-01 -4.27770197e-01 -2.65276600e-02
6.87995255e-01 1.37427890e+00 -4.59863961e-01 -2.54063129e-01
-2.88866639e-01 8.20828080e-01 4.00193155e-01 -7.70743266e-02
-4.58317429e-01 8.38510506e-03 3.42877328e-01 -6.38800710e-02
7.70995021e-02 -1.65418684e-01 1.00865133e-01 -1.02101731e+00
4.69638631e-02 -1.18387759e+00 6.17069602e-01 -1.73748970e-01
-1.32239985e+00 1.04767537e+00 1.46137521e-01 -1.02131462e+00
-7.78881133e-01 -3.19944859e-01 -7.57988989e-01 5.78465760e-01
-1.31349051e+00 -1.04987884e+00 -5.96954450e-02 3.77877653e-01
9.99139905e-01 -9.83481631e-02 1.09587908e+00 -3.97372879e-02
-6.35385633e-01 6.17891490e-01 -4.97344553e-01 3.81095886e-01
7.96262801e-01 -1.41551077e+00 3.93119484e-01 4.28003013e-01
3.70381325e-01 4.92908359e-01 9.07265902e-01 -3.94445449e-01
-1.01814699e+00 -5.15047312e-01 1.52933228e+00 -7.48224735e-01
9.45503592e-01 -4.55231130e-01 -7.84112573e-01 5.56764662e-01
7.04694569e-01 -5.70838571e-01 8.86827409e-01 8.33636999e-01
-3.68221253e-01 2.73311704e-01 -9.23955560e-01 5.29870868e-01
9.27604020e-01 -7.29891717e-01 -1.04599714e+00 4.70533252e-01
6.36120796e-01 -6.36340976e-01 -1.13163137e+00 3.13878715e-01
3.43570203e-01 -1.06551290e+00 7.73874760e-01 -6.32391512e-01
7.06348896e-01 4.07937795e-01 3.58108692e-02 -1.37868416e+00
3.21369529e-01 -1.13687265e+00 -5.87540790e-02 1.66897988e+00
7.51205981e-01 -7.18167186e-01 3.34669739e-01 7.35990405e-01
8.70072842e-02 -7.26202548e-01 -9.13412929e-01 -6.22722030e-01
6.65591955e-02 -1.12048782e-01 4.04038668e-01 9.54284608e-01
5.98707080e-01 1.37249041e+00 -3.14746380e-01 -2.11882398e-01
2.98664775e-02 2.81338573e-01 1.04594374e+00 -1.34168398e+00
-3.77213031e-01 -4.50268328e-01 1.32521177e-02 -1.22616303e+00
6.02605879e-01 -6.92500472e-01 1.84673756e-01 -1.33264661e+00
1.90379601e-02 -3.61601859e-01 3.63404155e-01 1.81505248e-01
-3.94224048e-01 -2.17112809e-01 2.58448541e-01 1.70314997e-01
-6.92775488e-01 7.10729659e-01 1.14600027e+00 2.36940086e-02
-4.68725413e-01 4.45562780e-01 -6.22794867e-01 6.51199579e-01
8.62844348e-01 -2.81028539e-01 -3.35500181e-01 1.48074061e-01
6.47854134e-02 4.12867725e-01 1.91023141e-01 -3.09598416e-01
1.91637129e-01 -1.48049504e-01 -5.48105299e-01 -2.17222050e-01
4.99698639e-01 -3.38628888e-01 -5.24477541e-01 2.29865715e-01
-9.90414917e-01 3.46361659e-02 -2.79185455e-02 2.69332349e-01
-4.60287720e-01 -3.01332295e-01 4.39592659e-01 4.66792621e-02
-2.93417662e-01 -2.83845067e-01 -4.25761878e-01 6.92589819e-01
5.74662328e-01 3.59232366e-01 -5.70288360e-01 -8.79778743e-01
-6.92003429e-01 2.04883590e-01 -1.47355244e-01 4.61186737e-01
1.00913443e-01 -6.54833496e-01 -9.96340752e-01 -5.61435699e-01
1.58707023e-01 -2.06818178e-01 -1.41139746e-01 9.10620213e-01
-1.20468736e-01 5.10546863e-01 3.00461054e-01 -4.07829195e-01
-1.98728335e+00 -2.19271369e-02 2.53220707e-01 -8.67857695e-01
-6.58489347e-01 1.03006458e+00 -1.03445053e-01 -6.02923214e-01
2.82692313e-01 -1.96033686e-01 -6.40879273e-01 2.86475867e-01
4.87021625e-01 3.57502073e-01 1.17299460e-01 -6.72535896e-01
-3.34742665e-01 -1.03068084e-01 -2.84336120e-01 -4.92920697e-01
1.17073369e+00 -4.32494700e-01 -1.80991933e-01 7.97365844e-01
8.89254510e-01 6.02535307e-01 -8.01605165e-01 -5.45232534e-01
3.99030298e-01 -7.88088813e-02 -3.43317002e-01 -9.73926604e-01
-3.64023000e-01 5.48694670e-01 -1.25798866e-01 9.11332726e-01
6.06365979e-01 3.53834808e-01 8.59101355e-01 6.10749125e-01
6.92192614e-02 -8.69726062e-01 9.99153778e-02 9.75261688e-01
1.14497161e+00 -1.44541359e+00 1.56911105e-01 -9.67397034e-01
-9.97484565e-01 1.10271406e+00 5.33612728e-01 3.04384053e-01
3.97582322e-01 9.00081396e-02 3.31095964e-01 -4.46834058e-01
-1.16288733e+00 -2.79170960e-01 3.84723485e-01 1.64763033e-01
9.85293806e-01 4.07977179e-02 -6.84684992e-01 5.78714848e-01
-6.91785991e-01 -5.74874341e-01 6.60440385e-01 7.64214098e-01
-9.10350084e-02 -1.57542527e+00 1.50738910e-01 8.64995196e-02
-4.94959414e-01 -7.83220306e-02 -1.24739885e+00 8.84208739e-01
-5.89023888e-01 1.67946506e+00 -2.16959089e-01 -3.07830602e-01
3.64261806e-01 2.83028692e-01 4.62065339e-01 -9.28721428e-01
-1.31388140e+00 -2.07221389e-01 1.45602310e+00 -3.81619424e-01
-6.72757268e-01 -5.68868577e-01 -1.01493418e+00 -3.28287095e-01
-5.14716387e-01 7.35407233e-01 4.49219346e-01 1.29936814e+00
2.27567017e-01 3.26918155e-01 9.23424959e-01 -3.44536453e-01
-6.90464556e-01 -1.44108212e+00 4.88780923e-02 5.67583442e-01
3.72920662e-01 -4.95680392e-01 -4.43971336e-01 -2.24801779e-01] | [12.483976364135742, 8.051324844360352] |
3bdaaa81-e7ca-4ade-aaf6-47a17c988e51 | deep-learning-based-end-to-end-diagnosis | 2002.05536 | null | https://arxiv.org/abs/2002.05536v2 | https://arxiv.org/pdf/2002.05536v2.pdf | Deep Learning-based End-to-end Diagnosis System for Avascular Necrosis of Femoral Head | As the first diagnostic imaging modality of avascular necrosis of the femoral head (AVNFH), accurately staging AVNFH from a plain radiograph is critical yet challenging for orthopedists. Thus, we propose a deep learning-based AVNFH diagnosis system (AVN-net). The proposed AVN-net reads plain radiographs of the pelvis, conducts diagnosis, and visualizes results automatically. Deep convolutional neural networks are trained to provide an end-to-end diagnosis solution, covering tasks of femoral head detection, exam-view identification, side classification, AVNFH diagnosis, and key clinical notes generation. AVN-net is able to obtain state-of-the-art testing AUC of 0.97 (95% CI: 0.97-0.98) in AVNFH detection and significantly greater F1 scores than less-to-moderately experienced orthopedists in all diagnostic tests (p<0.01). Furthermore, two real-world pilot studies were conducted for diagnosis support and education assistance, respectively, to assess the utility of AVN-net. The experimental results are promising. With the AVN-net diagnosis as a reference, the diagnostic accuracy and consistency of all orthopedists considerably improved while requiring only 1/4 of the time. Students self-studying the AVNFH diagnosis using AVN-net can learn better and faster than the control group. To the best of our knowledge, this study is the first research on the prospective use of a deep learning-based diagnosis system for AVNFH by conducting two pilot studies representing real-world application scenarios. We have demonstrated that the proposed AVN-net achieves expert-level AVNFH diagnosis performance, provides efficient support in clinical decision-making, and effectively passes clinical experience to students. | ['Yang Li', 'Yan Li', 'Hua Tian'] | 2020-02-12 | null | null | null | null | ['head-detection'] | ['computer-vision'] | [-2.88544953e-01 3.75687480e-01 -4.92778957e-01 -3.25693160e-01
-1.04536533e+00 -6.27546534e-02 -2.27381274e-01 1.94043964e-01
-2.73689091e-01 6.64568663e-01 4.70446348e-02 -7.66836226e-01
-4.78985548e-01 -8.20348859e-01 -6.40206873e-01 -4.74796385e-01
-2.29211658e-01 1.08221185e+00 1.22253388e-01 1.44764721e-01
1.36279285e-01 7.12568700e-01 -1.49287164e+00 4.36124653e-01
8.91868830e-01 9.89665627e-01 2.85169631e-01 7.80605435e-01
3.17499101e-01 1.21538591e+00 -6.19757950e-01 -2.06022412e-01
1.20502412e-01 -1.76348463e-01 -9.66789246e-01 -2.28820555e-02
6.72429740e-01 -9.29493904e-01 -5.36649406e-01 4.23980892e-01
1.10610425e+00 -1.57939091e-01 7.18286097e-01 -8.94346714e-01
-5.26902258e-01 5.20282805e-01 -5.25879800e-01 2.70168602e-01
1.26289442e-01 5.73162138e-01 5.90227246e-01 -7.43554831e-01
8.84940863e-01 7.25089967e-01 1.00563204e+00 6.01209819e-01
-8.20164263e-01 -7.42334366e-01 -6.20816648e-01 4.07822669e-01
-1.14555967e+00 3.69975656e-01 4.38179933e-02 -6.67567134e-01
7.25218236e-01 2.19437942e-01 1.39980066e+00 1.07932806e+00
1.10226345e+00 9.71118569e-01 9.22366321e-01 -1.33776441e-01
2.76030958e-01 -3.35907310e-01 1.82463840e-01 9.14941788e-01
3.79976630e-01 7.61979967e-02 -3.08207959e-01 -1.06145917e-02
1.19632483e+00 1.62733138e-01 -2.83616990e-01 -3.31938982e-01
-1.02752090e+00 8.85987520e-01 6.73406601e-01 -1.83027059e-01
-6.35013700e-01 1.56631738e-01 7.16421604e-01 1.27800137e-01
-1.42426770e-02 3.26931715e-01 -7.52808526e-02 -2.81960219e-01
-8.99649084e-01 4.26069528e-01 6.51534379e-01 4.31847751e-01
-4.65824425e-01 -3.73187028e-02 -3.34261268e-01 1.00898063e+00
9.21981409e-02 6.46324456e-01 9.95229006e-01 -1.14072788e+00
-7.64544681e-02 4.62170452e-01 -5.15284359e-01 -5.15878379e-01
-8.45995665e-01 -8.90028536e-01 -7.65382886e-01 6.19247377e-01
3.59566182e-01 -2.76712447e-01 -1.41133153e+00 1.17486572e+00
2.44824126e-01 -2.61623770e-01 -2.30016619e-01 1.23104072e+00
1.12101436e+00 1.89547822e-01 -3.50091383e-02 -3.49585600e-02
1.77182138e+00 -9.57922995e-01 -5.51058531e-01 -6.29868954e-02
1.09767687e+00 -5.60919821e-01 1.06171882e+00 7.65793443e-01
-1.13178122e+00 -4.27459508e-01 -9.65821981e-01 7.55754486e-02
2.65454739e-01 4.76155013e-01 8.51291895e-01 5.05246520e-01
-9.55485761e-01 4.89598215e-01 -1.12406373e+00 -2.84433335e-01
6.39310539e-01 5.52512348e-01 -5.87310851e-01 -5.06185532e-01
-1.13157701e+00 1.13995433e+00 1.03407674e-01 4.19590957e-02
-1.25585473e+00 -1.15451205e+00 -7.94424891e-01 2.51578055e-02
3.10586542e-01 -1.09142900e+00 1.89014840e+00 -2.79663056e-01
-1.13399374e+00 1.02041781e+00 4.53533292e-01 -5.11925340e-01
8.31140935e-01 -2.27675810e-01 -2.76303709e-01 4.12297934e-01
4.09017831e-01 6.73670948e-01 2.91325003e-01 -8.82114470e-01
-8.59048605e-01 -4.30768311e-01 -3.25176954e-01 2.89393008e-01
3.68191657e-04 -5.63187301e-01 -1.42813921e-01 -5.19702375e-01
1.41270131e-01 -9.86584604e-01 -2.33219132e-01 5.36446214e-01
-2.41240785e-01 -2.47079238e-01 6.48266196e-01 -8.76964211e-01
1.13255644e+00 -2.02643251e+00 -5.89796424e-01 2.72242814e-01
9.79911566e-01 4.18150514e-01 4.28222626e-01 -2.22234684e-03
-4.29903477e-01 -4.31001745e-02 7.63277858e-02 2.66136497e-01
-3.94482672e-01 3.36703211e-01 4.96759862e-01 6.36226416e-01
-1.53921604e-01 8.57508004e-01 -7.53323495e-01 -6.85980439e-01
3.68964612e-01 3.66797894e-01 -5.83709657e-01 9.01156962e-02
3.98687214e-01 1.31277189e-01 -8.56212154e-02 7.92237699e-01
3.65452021e-01 -6.35616958e-01 1.81590691e-01 -6.40530372e-03
2.27553487e-01 -4.38378565e-02 -1.00461352e+00 1.50327480e+00
-5.06042302e-01 7.51575172e-01 -1.13874540e-01 -6.99292123e-01
6.76547408e-01 5.41290700e-01 6.90283358e-01 -7.73812115e-01
3.70397121e-01 7.67165124e-01 6.39232278e-01 -8.72973621e-01
9.39389095e-02 -2.96978027e-01 1.93321809e-01 3.60284626e-01
3.28126997e-01 1.73501968e-01 1.61812618e-01 2.15948895e-01
1.32748497e+00 -3.93017352e-01 5.16950488e-01 -1.15927316e-01
8.94872844e-02 1.88812837e-01 3.52454901e-01 1.06016862e+00
-2.68183351e-01 9.03728604e-01 5.70466995e-01 -5.98654330e-01
-8.68901074e-01 -1.31025267e+00 -4.38652545e-01 5.72780669e-01
-1.72002390e-01 -7.66325146e-02 -6.27979517e-01 -7.11800516e-01
3.27832252e-01 4.08714980e-01 -8.78147840e-01 -3.29886436e-01
-5.04475415e-01 -2.38863751e-01 5.67314982e-01 1.14978528e+00
2.60769278e-01 -1.07441354e+00 -8.67475927e-01 2.48533592e-01
-3.42221297e-02 -5.37709653e-01 -1.69961438e-01 2.33378291e-01
-9.72780228e-01 -1.17143857e+00 -1.38816857e+00 -1.15407729e+00
3.87092769e-01 -1.55731291e-01 9.69091535e-01 1.67398885e-01
-8.51234853e-01 1.86490923e-01 -2.07266156e-02 -4.47368652e-01
-3.80657583e-01 2.68172264e-01 -1.38920665e-01 -6.63334012e-01
8.55928436e-02 -3.39961201e-01 -1.00037956e+00 6.67874739e-02
-5.01590371e-01 2.07487419e-01 1.16439378e+00 1.11226380e+00
7.24513531e-01 -3.24384868e-01 4.01398122e-01 -1.08708489e+00
6.96244538e-01 -5.46598196e-01 1.01990358e-03 4.15440723e-02
-1.02066243e+00 -3.50798458e-01 3.45832080e-01 -3.76375347e-01
-7.98274696e-01 -1.57725811e-01 -5.27030051e-01 -4.56344843e-01
-2.56477803e-01 8.66523147e-01 7.66285479e-01 -1.12735592e-01
9.14489925e-01 -6.01839721e-02 5.12516081e-01 -9.15878937e-02
-2.73253441e-01 7.02040732e-01 9.81142759e-01 -3.68182957e-01
1.80385932e-01 1.07603580e-01 3.52877229e-01 -4.14420575e-01
-7.12421954e-01 -5.36883771e-01 -2.65076846e-01 -5.91845930e-01
7.23627448e-01 -9.62479711e-01 -8.94209146e-01 3.75042856e-01
-6.30337954e-01 -2.54772276e-01 -5.49746990e-01 9.80644166e-01
-4.96390343e-01 1.95236787e-01 -1.09388816e+00 -3.48667443e-01
-7.54907906e-01 -1.63112938e+00 8.66760433e-01 2.96547204e-01
-5.79928219e-01 -9.06881988e-01 1.65640544e-02 4.61406499e-01
2.67703325e-01 2.33816922e-01 9.88330543e-01 -8.46232414e-01
-9.04303715e-02 -6.15252376e-01 -2.29418993e-01 2.05465928e-01
-1.72083199e-01 -1.04112467e-02 -6.89232349e-01 -5.90193756e-02
-1.96291730e-01 -4.76580948e-01 5.99059165e-01 1.07912874e+00
1.31654000e+00 2.22493827e-01 -3.28878880e-01 6.87355399e-01
1.19370246e+00 4.40087885e-01 6.36169195e-01 6.34655714e-01
6.95870519e-01 1.00560471e-01 7.09624410e-01 6.08775437e-01
3.89360398e-01 1.21662043e-01 5.25222838e-01 -5.96169949e-01
-3.79787743e-01 1.10841691e-01 -2.19869986e-01 6.06429279e-01
-1.26716912e-01 -1.22004561e-01 -1.54204655e+00 7.11312532e-01
-1.47580123e+00 -4.85151321e-01 -4.74500269e-01 1.76535439e+00
6.51194572e-01 3.31871063e-01 2.85882745e-02 2.61730611e-01
5.44261634e-01 -4.08405095e-01 -5.76330066e-01 -3.86779606e-01
4.73484218e-01 7.03640282e-01 5.52444100e-01 -4.55510281e-02
-8.74126077e-01 7.59952366e-02 6.21679068e+00 8.68524730e-01
-1.48700070e+00 7.44621009e-02 6.92508578e-01 -2.27093041e-01
7.02501312e-02 -4.89771426e-01 -4.53636885e-01 2.65814841e-01
8.31172049e-01 -4.80518453e-02 -4.27623034e-01 1.17842948e+00
1.08242244e-01 -2.63716638e-01 -1.02184594e+00 8.82473469e-01
-1.20876446e-01 -1.69462502e+00 -5.22554740e-02 2.93099344e-01
4.70774561e-01 -1.51824042e-01 1.59571663e-01 5.97055435e-01
2.21304312e-01 -1.11681759e+00 3.56610298e-01 3.32744837e-01
1.18173230e+00 -7.71480858e-01 1.52120614e+00 -7.42304325e-02
-7.73315191e-01 -1.32372081e-01 1.50004506e-01 -7.11529108e-04
1.31194502e-01 4.85985190e-01 -1.38942826e+00 3.51367265e-01
7.18696117e-01 3.53869855e-01 -5.37064075e-01 1.50526834e+00
-2.14327365e-01 9.03909683e-01 -1.05769187e-01 2.05998808e-01
2.24476695e-01 6.33454800e-01 1.45381600e-01 8.61903250e-01
4.13902372e-01 2.30415389e-01 2.45172516e-01 3.92420262e-01
-8.23364109e-02 2.68866152e-01 -2.60242611e-01 2.11866990e-01
1.98558763e-01 1.22751796e+00 -8.82747769e-01 -4.47618157e-01
-2.92215109e-01 4.03233111e-01 -1.14968732e-01 -6.56188801e-02
-8.33802462e-01 -3.70970339e-01 1.79835483e-01 6.43038213e-01
-1.72933698e-01 5.14564931e-01 -7.45111406e-01 -4.78437036e-01
-1.40732080e-01 -9.04537499e-01 7.05786228e-01 -1.05367351e+00
-1.21741974e+00 2.10311934e-01 -3.03558260e-01 -1.42904210e+00
-5.21529913e-01 -8.68521750e-01 -4.99378175e-01 7.26421416e-01
-9.52040374e-01 -8.83297503e-01 -7.26989806e-01 2.94792593e-01
5.73806167e-01 -1.52659148e-01 8.10229182e-01 2.95627981e-01
-3.29047620e-01 7.92189717e-01 1.20462418e-01 4.90456045e-01
9.13064718e-01 -1.40608060e+00 -1.26137644e-01 3.42124254e-01
-8.62452090e-01 5.63184083e-01 2.72855878e-01 -9.18350816e-01
-1.10620415e+00 -8.73275101e-01 5.28639317e-01 1.43739343e-01
3.73762310e-01 4.27024722e-01 -8.06615114e-01 6.00406945e-01
-8.85883719e-02 2.31577784e-01 1.11151278e+00 -9.32469815e-02
7.95134380e-02 2.24735551e-02 -1.28514373e+00 4.45104778e-01
7.61220217e-01 -1.58591345e-01 -4.31291610e-01 3.85514051e-01
4.33195919e-01 -1.24526119e+00 -1.53007019e+00 8.88279855e-01
1.15482354e+00 -6.79897547e-01 8.14534962e-01 -4.82560158e-01
1.25676131e+00 3.07744831e-01 2.85285413e-01 -1.01320529e+00
-3.45267773e-01 2.81636626e-01 -5.25989421e-02 4.88510549e-01
3.76828194e-01 -4.94522542e-01 9.92350340e-01 2.68377483e-01
-7.21021354e-01 -1.40380216e+00 -7.31738269e-01 -3.54236931e-01
1.22899167e-01 -3.89696866e-01 1.51349008e-02 9.59158659e-01
-1.86977848e-01 -1.52518079e-01 9.40654650e-02 1.03080645e-01
5.11532128e-01 -1.10928997e-01 5.34455240e-01 -1.37650001e+00
-5.29137850e-01 -5.13809144e-01 -8.07780981e-01 -4.29726154e-01
-3.92600060e-01 -9.26212847e-01 -1.00957096e-01 -2.14295697e+00
2.76052773e-01 -1.61545202e-01 -1.88651741e-01 6.21266484e-01
-1.54857889e-01 2.38857821e-01 -6.78617135e-02 1.81152046e-01
-4.99732792e-02 4.68228990e-03 1.84266949e+00 -1.16606755e-02
1.75328538e-01 1.91167891e-01 -5.37764847e-01 6.17695510e-01
7.05084145e-01 -4.24663514e-01 -5.92359960e-01 -2.17659712e-01
-1.30564511e-01 6.64597750e-01 3.73707354e-01 -1.44419503e+00
1.74732178e-01 2.90059261e-02 8.94328296e-01 -8.78760874e-01
-3.73451039e-02 -4.10163373e-01 8.85307193e-02 1.14210689e+00
-2.84735948e-01 1.27951428e-01 2.26626053e-01 1.00160763e-02
-2.61376411e-01 -3.19422781e-01 7.59820402e-01 -4.53018993e-01
-8.24702024e-01 2.65547901e-01 -1.11958945e+00 1.62515625e-01
1.15947247e+00 -4.25228626e-01 -4.41512734e-01 -4.65046853e-01
-1.07348585e+00 3.42567354e-01 -3.98106836e-02 2.61320751e-02
8.72649670e-01 -1.16700661e+00 -7.01470077e-01 1.81942105e-01
4.43716012e-02 4.31495070e-01 8.38552773e-01 1.28412938e+00
-1.48187494e+00 3.81391644e-01 -6.02610588e-01 -1.06980622e+00
-1.30210888e+00 3.30387317e-02 5.30536950e-01 -5.72672904e-01
-1.03544497e+00 8.91896486e-01 1.69757575e-01 -7.03804433e-01
5.40576279e-01 -2.86629885e-01 -6.75372360e-03 -2.56095151e-03
4.30734187e-01 3.27567309e-01 5.58800161e-01 5.60292043e-02
-1.97962791e-01 1.68013230e-01 -5.91819584e-01 1.15988754e-01
1.14744747e+00 2.93519706e-01 3.94716144e-01 1.97422922e-01
9.53764081e-01 -4.80454624e-01 -6.22626483e-01 8.63407031e-02
-5.44952750e-01 -1.64985925e-01 3.37733686e-01 -1.19592500e+00
-1.35339355e+00 1.02623200e+00 1.29139555e+00 -3.52116644e-01
1.02802062e+00 -5.81735894e-02 1.00008881e+00 9.95341912e-02
1.49013966e-01 -8.49849641e-01 1.41376927e-01 2.59751379e-01
5.90018570e-01 -1.05775630e+00 8.04947540e-02 -2.48605892e-01
-6.71063006e-01 1.40133655e+00 1.09791505e+00 -9.75564942e-02
4.36885118e-01 4.75812644e-01 4.66674894e-01 -5.94675839e-01
-5.88194370e-01 1.64061174e-01 1.14602201e-01 6.29559219e-01
6.10829234e-01 4.15068060e-01 -5.90267956e-01 6.76883876e-01
-3.93539339e-01 5.77489138e-01 7.13205218e-01 1.21787250e+00
-4.67761308e-01 -5.22056639e-01 -4.25078720e-01 1.16384506e+00
-6.42558277e-01 1.47890866e-01 -1.34868296e-02 1.41780448e+00
6.18731119e-02 9.72839817e-02 2.26491280e-02 -4.77697521e-01
5.26712477e-01 -4.13352996e-02 2.76810139e-01 -6.74446881e-01
-8.16324115e-01 4.30753455e-02 7.61862518e-03 -4.40073609e-01
2.41604745e-01 -5.91840029e-01 -1.60874021e+00 -6.24389291e-01
-1.88569456e-01 -4.41498794e-02 5.23700237e-01 8.96070778e-01
3.96192893e-02 1.30597734e+00 7.75049105e-02 -4.73680258e-01
-6.59639299e-01 -1.01120055e+00 -8.68876457e-01 1.66697621e-01
1.22481503e-01 -8.88319135e-01 8.88379514e-02 -3.17729294e-01] | [15.158407211303711, -2.1124579906463623] |
e4a025e5-b168-4792-8a97-1d9113d0f152 | vehicle-position-estimation-with-aerial | 2004.08206 | null | https://arxiv.org/abs/2004.08206v2 | https://arxiv.org/pdf/2004.08206v2.pdf | Vehicle Position Estimation with Aerial Imagery from Unmanned Aerial Vehicles | The availability of real-world data is a key element for novel developments in the fields of automotive and traffic research. Aerial imagery has the major advantage of recording multiple objects simultaneously and overcomes limitations such as occlusions. However, there are only few data sets available. This work describes a process to estimate a precise vehicle position from aerial imagery. A robust object detection is crucial for reliable results, hence the state-of-the-art deep neural network Mask-RCNN is applied for that purpose. Two training data sets are employed: The first one is optimized for detecting the test vehicle, while the second one consists of randomly selected images recorded on public roads. To reduce errors, several aspects are accounted for, such as the drone movement and the perspective projection from a photograph. The estimated position is comapared with a reference system installed in the test vehicle. It is shown, that a mean accuracy of 20 cm can be achieved with flight altitudes up to 100 m, Full-HD resolution and a frame-by-frame detection. A reliable position estimation is the basis for further data processing, such as obtaining additional vehicle state variables. The source code, training weights, labeled data and example videos are made publicly available. This supports researchers to create new traffic data sets with specific local conditions. | ['Eduardo Sánchez Morales', 'Friedrich Kruber', 'Samarjit Chakraborty', 'Michael Botsch'] | 2020-04-17 | null | null | null | null | ['robust-object-detection', 'drone-based-object-tracking', 'traffic-classification'] | ['computer-vision', 'computer-vision', 'miscellaneous'] | [ 1.54846027e-01 -3.40242684e-01 -4.69721071e-02 -4.14620131e-01
-4.61207092e-01 -2.67648220e-01 4.52755392e-01 -2.46595025e-01
-6.18769407e-01 7.58969188e-01 -6.14457428e-01 -2.40924045e-01
-9.35318917e-02 -7.41657555e-01 -7.21381307e-01 -9.23961043e-01
-2.74354190e-01 4.38460022e-01 3.96118551e-01 -2.06661314e-01
1.63935125e-01 1.16104221e+00 -2.09611368e+00 -1.18388124e-01
5.24443805e-01 1.11461508e+00 5.95983982e-01 5.73457062e-01
4.44301456e-01 2.81487674e-01 -7.47014880e-01 -1.32708520e-01
5.16032457e-01 1.16792090e-01 -9.13341343e-02 3.38003814e-01
7.16257989e-01 -6.64653897e-01 -3.66718620e-01 9.51985121e-01
3.73905718e-01 1.52420700e-01 3.57979298e-01 -1.24111700e+00
4.37744930e-02 -4.16608416e-02 -3.26417238e-01 2.62139976e-01
-6.44439384e-02 3.31068039e-01 1.86604798e-01 -6.23552024e-01
6.64501369e-01 6.92062497e-01 4.24962938e-01 1.76484108e-01
-7.89554179e-01 -7.80110121e-01 -1.41638264e-01 6.93237245e-01
-1.90893567e+00 -4.96499330e-01 7.48123884e-01 -4.58886534e-01
8.88097763e-01 3.47915232e-01 7.84862578e-01 7.93914855e-01
3.44137073e-01 3.01395625e-01 8.58390093e-01 -2.74044901e-01
2.56704409e-02 5.44052362e-01 -1.19276322e-01 4.15899962e-01
4.53762203e-01 6.85115695e-01 1.79042518e-02 3.73311758e-01
6.60938978e-01 -3.91417146e-02 -3.42564583e-01 -4.79795605e-01
-9.68733490e-01 6.50737286e-01 4.20768678e-01 2.98327446e-01
-4.19904798e-01 -6.11876361e-02 1.10276066e-01 2.18439117e-01
3.29919577e-01 2.40170956e-01 -3.28778446e-01 -3.23208384e-02
-1.08279908e+00 2.85135448e-01 3.35725456e-01 9.61920142e-01
8.95426571e-01 5.12888670e-01 3.59100223e-01 3.43682140e-01
1.52395189e-01 9.66601372e-01 2.33203709e-01 -8.69734764e-01
4.12382275e-01 4.65257555e-01 2.52568930e-01 -1.26540792e+00
-4.01561707e-01 -4.97045964e-01 -4.96363044e-01 7.42031395e-01
1.79607660e-01 -3.56956631e-01 -7.26121485e-01 1.15880215e+00
4.99887824e-01 2.36512810e-01 1.02874607e-01 1.23264670e+00
7.00042009e-01 8.97944689e-01 -4.36321914e-01 -2.58716553e-01
1.02649319e+00 -5.66251099e-01 -7.93783069e-01 -3.19430649e-01
4.19090331e-01 -6.71487629e-01 2.18189731e-01 6.24572039e-01
-7.98864603e-01 -9.06446218e-01 -1.27343857e+00 4.39406544e-01
-6.13061547e-01 7.08098531e-01 2.62274832e-01 5.67133129e-01
-8.85596335e-01 4.03054088e-01 -7.40104139e-01 -2.78177053e-01
1.93715096e-01 5.80002546e-01 -5.37196577e-01 1.03579918e-02
-1.21363008e+00 1.26636922e+00 2.32182503e-01 5.75356245e-01
-9.92157459e-01 -2.99091071e-01 -8.10691714e-01 -1.46363959e-01
4.68876302e-01 -9.48012173e-02 8.74198496e-01 -9.27798927e-01
-1.38359106e+00 7.08376467e-01 9.80987400e-02 -8.09714139e-01
6.72863603e-01 -1.80960715e-01 -8.46590817e-01 3.07280511e-01
-2.27381904e-02 6.69958472e-01 1.05634809e+00 -1.19240165e+00
-8.64062369e-01 -4.06585842e-01 -1.03401713e-01 7.00084940e-02
-6.42862990e-02 5.48234694e-02 -4.54465866e-01 -4.91838939e-02
-3.57870549e-01 -1.04203308e+00 -6.22203685e-02 -9.38588604e-02
-1.69211760e-01 3.09845388e-01 1.34841621e+00 -6.87471688e-01
9.37645018e-01 -1.94224501e+00 -5.71073480e-02 1.32262245e-01
-7.92765990e-02 7.10590959e-01 1.57245994e-01 2.23960966e-01
-2.87760407e-01 -4.48276609e-01 2.24056747e-02 -3.70922089e-02
-5.29897273e-01 -1.11520797e-01 -4.35880907e-02 8.76913249e-01
2.09885731e-01 4.02982801e-01 -3.62970203e-01 -2.19481274e-01
9.38487589e-01 6.71348870e-01 -1.26932412e-01 1.02677740e-01
1.12849809e-01 3.85815054e-01 -1.67062387e-01 5.45791686e-01
9.57893908e-01 4.53954339e-01 -8.64630118e-02 -3.70811492e-01
-9.33337808e-01 -1.75048038e-01 -1.50164807e+00 9.89245951e-01
-3.18771511e-01 1.28244781e+00 3.25981319e-01 -8.90729070e-01
1.06673384e+00 3.86656702e-01 4.70754653e-01 -6.74112141e-01
5.81370115e-01 -7.81141818e-02 4.57981862e-02 -7.42494524e-01
7.68223166e-01 2.19998151e-01 4.04746205e-01 -2.64078736e-01
-1.80286348e-01 -1.82599738e-01 5.88773191e-01 -1.75875798e-01
5.48068047e-01 -5.28023206e-02 1.52283087e-01 -1.40774041e-01
6.30195022e-01 3.31060439e-01 3.21140140e-01 1.76683173e-01
-1.44817665e-01 4.55370039e-01 1.99870765e-01 -6.72819793e-01
-1.04070652e+00 -4.55968380e-01 -4.88570631e-01 1.43621728e-01
3.70329052e-01 2.83988193e-03 -7.78346777e-01 -2.07110584e-01
3.24632414e-02 7.13265240e-01 -4.98535156e-01 -7.57737905e-02
-6.24984145e-01 -5.20149350e-01 1.02859907e-01 3.66282165e-01
6.82886362e-01 -8.75144064e-01 -1.14867699e+00 1.39457792e-01
-7.74043426e-02 -1.34819233e+00 1.72998562e-01 -6.03424162e-02
-6.71511590e-01 -1.16140401e+00 -4.50793386e-01 -3.30476999e-01
6.21899188e-01 6.49441183e-01 7.87767649e-01 7.50370175e-02
-4.79175687e-01 1.09841786e-01 -8.27577785e-02 -5.32398224e-01
-3.57145041e-01 -2.56204307e-01 2.66183376e-01 9.60964784e-02
3.29132140e-01 -6.37349114e-02 -3.86894971e-01 7.25860536e-01
-5.09136260e-01 -5.86984083e-02 8.17533672e-01 2.64539599e-01
5.64549685e-01 4.31623101e-01 5.19209057e-02 -3.16752464e-01
-6.69137612e-02 -2.31565252e-01 -1.45209265e+00 -3.17528516e-01
-3.89632702e-01 -5.19345641e-01 5.75126827e-01 -3.77713561e-01
-9.34131682e-01 4.03307050e-01 -2.45569184e-01 -6.84667706e-01
-7.43619502e-01 2.37695307e-01 -1.87542632e-01 -2.64917284e-01
5.84186256e-01 -2.29026526e-02 2.11375475e-01 -5.05870096e-02
-1.35996222e-01 8.13593626e-01 5.81951797e-01 9.71797556e-02
8.35080802e-01 6.41530573e-01 9.74133685e-02 -1.64697015e+00
-2.86562026e-01 -5.11795640e-01 -8.46488714e-01 -8.57921839e-01
9.36641753e-01 -1.03118753e+00 -5.98308980e-01 3.87107909e-01
-1.00982523e+00 -1.92990780e-01 -1.06767334e-01 8.39219332e-01
-3.32572669e-01 8.70949849e-02 5.22070862e-02 -9.39713478e-01
1.78573523e-02 -1.30270720e+00 9.48241591e-01 2.76481599e-01
2.16348782e-01 -6.90023065e-01 -2.15234652e-01 4.40805137e-01
3.39785784e-01 1.99304953e-01 1.37962192e-01 -1.40664294e-01
-9.93377209e-01 -8.70515525e-01 6.30342588e-02 5.10086834e-01
-2.65736699e-01 4.85497862e-01 -8.73626888e-01 -3.44937384e-01
4.36366536e-02 1.06926203e-01 5.73866427e-01 7.34479606e-01
8.61872673e-01 1.30847096e-01 -5.44201672e-01 4.61579949e-01
1.49318290e+00 4.52740282e-01 8.49100471e-01 3.19624931e-01
4.74465281e-01 5.46741903e-01 1.12671244e+00 2.12466821e-01
1.84906237e-02 1.01880980e+00 9.29308414e-01 -2.12637708e-01
4.49616946e-02 3.43563080e-01 1.79911315e-01 3.17268252e-01
-4.18029189e-01 -2.88001269e-01 -8.11927736e-01 4.27832216e-01
-1.46579158e+00 -1.21953070e+00 -5.18014789e-01 2.51195431e+00
-1.25942841e-01 2.84825563e-01 3.78725193e-02 3.92241657e-01
7.86936283e-01 9.31608006e-02 -2.36326754e-01 -1.76846925e-02
9.03498828e-02 -3.90490502e-01 1.02114046e+00 5.65198839e-01
-1.39535022e+00 6.94612145e-01 6.01228809e+00 4.33215559e-01
-1.58396912e+00 -1.43165156e-01 1.35396928e-01 -1.87311009e-01
5.10149777e-01 -2.55206972e-01 -1.33501947e+00 4.59754676e-01
1.27577591e+00 1.58821255e-01 2.56575763e-01 1.06777585e+00
5.26197016e-01 -4.76411134e-01 -5.32854140e-01 9.38436627e-01
4.48879331e-01 -1.26971090e+00 -3.26405674e-01 3.14959109e-01
4.84454364e-01 2.77430177e-01 -1.17134042e-01 1.17175519e-01
-2.41541296e-01 -6.38490498e-01 8.05587888e-01 5.62410235e-01
9.23441648e-01 -1.17122602e+00 1.11309397e+00 3.93490255e-01
-1.28509223e+00 -1.94612026e-01 -6.33693337e-01 -6.70681968e-02
3.31955940e-01 3.82081836e-01 -9.28214192e-01 6.40939772e-01
6.04750037e-01 6.26124978e-01 -6.38507843e-01 1.11910582e+00
-9.33165550e-02 2.68151671e-01 -4.46330816e-01 -7.07509145e-02
2.07763985e-01 -3.88315499e-01 6.64509892e-01 8.83250535e-01
5.98384321e-01 -9.81474891e-02 6.01326637e-02 7.02192605e-01
4.15441900e-01 -2.39100859e-01 -9.26860809e-01 1.57282159e-01
2.47923642e-01 1.61022067e+00 -5.95784545e-01 -1.37980163e-01
-4.60471213e-01 4.87925053e-01 -1.22306369e-01 2.40295887e-01
-1.15367758e+00 -4.23573256e-01 7.47922480e-01 6.60192013e-01
5.96118450e-01 -5.42363703e-01 3.16031784e-01 -7.12570190e-01
-1.47494748e-01 -5.95392942e-01 -2.31792539e-01 -9.54927742e-01
-2.18898639e-01 7.92618930e-01 4.58337843e-01 -1.60009444e+00
-4.36374009e-01 -1.04440057e+00 -3.85480970e-01 6.75386727e-01
-1.46570218e+00 -9.97129738e-01 -6.16719246e-01 2.74262905e-01
5.64224303e-01 -4.35763299e-01 4.98234302e-01 6.57244146e-01
-9.20339465e-01 -2.00968515e-02 1.62782013e-01 1.71048697e-02
3.56710821e-01 -7.41592646e-01 1.56348780e-01 1.07859468e+00
-3.71871479e-02 1.41020298e-01 9.53612089e-01 -6.59440577e-01
-1.44861639e+00 -1.15093136e+00 5.99175632e-01 -3.85649532e-01
5.25051594e-01 -1.60096005e-01 -6.21086895e-01 6.45263374e-01
1.97001010e-01 8.92262012e-02 2.07179934e-01 -4.59641367e-01
6.79402351e-01 -5.95642209e-01 -9.44722533e-01 3.46663147e-01
4.13036793e-01 -4.58026640e-02 -2.17264771e-01 2.24279270e-01
1.51223660e-01 -7.08586991e-01 -5.24686515e-01 4.15768296e-01
5.24100959e-01 -1.22468758e+00 9.68481839e-01 -2.46779788e-02
1.75480300e-03 -5.71330130e-01 6.64693192e-02 -1.16348946e+00
-1.09330155e-01 -1.41080290e-01 1.01050399e-02 8.88024449e-01
3.00019234e-01 -5.57603538e-01 8.14862967e-01 3.66680056e-01
-2.93879569e-01 -4.94296819e-01 -1.02502561e+00 -8.48941863e-01
-5.29828966e-01 -6.39374733e-01 3.88350308e-01 4.30349261e-01
-6.56720698e-01 2.69999266e-01 -5.65248787e-01 7.36965537e-01
5.12075841e-01 7.94189274e-02 1.08134103e+00 -1.48381734e+00
3.08716059e-01 4.19895723e-02 -8.45117927e-01 -7.72065461e-01
1.14947803e-01 -2.17732146e-01 -6.45424500e-02 -1.28581393e+00
-5.63229024e-01 -1.52191207e-01 3.01197380e-01 -1.68581516e-01
3.88745815e-01 3.69702190e-01 7.55024627e-02 -4.73045297e-02
-2.49272823e-01 2.56506234e-01 9.50139284e-01 2.83224266e-02
7.36350268e-02 3.52078319e-01 -1.78631031e-04 7.92959690e-01
8.35701466e-01 -3.54992807e-01 -3.12355518e-01 -3.77787292e-01
-3.97008285e-02 8.41882452e-02 6.68721795e-01 -1.65731263e+00
2.98343569e-01 2.00583152e-02 6.13015890e-01 -1.24682367e+00
9.00783598e-01 -1.46218944e+00 5.09372175e-01 4.69447106e-01
2.11561874e-01 1.76591933e-01 3.57367128e-01 3.79365414e-01
-3.52236450e-01 -4.06866401e-01 9.45541441e-01 -1.19573750e-01
-1.06732690e+00 2.52282470e-01 -6.05746329e-01 -4.52035129e-01
1.61855304e+00 -5.44977546e-01 -6.83197901e-02 -3.97140324e-01
-3.96011829e-01 1.38785224e-03 4.41087335e-01 3.95310849e-01
5.88784397e-01 -1.22342610e+00 -6.09448075e-01 7.10934103e-01
-6.23899139e-02 -1.07384890e-01 4.94400710e-01 9.73016083e-01
-7.55894125e-01 8.84172380e-01 -4.43214297e-01 -7.84780502e-01
-1.64333344e+00 6.92637324e-01 4.43342149e-01 3.15361977e-01
-4.91536379e-01 3.16399604e-01 -2.78615355e-01 -5.37359081e-02
5.55157661e-02 -3.69315356e-01 -6.04328275e-01 1.90785944e-01
9.39528048e-01 5.64492285e-01 3.73683959e-01 -1.22539937e+00
-2.26118833e-01 7.93685675e-01 2.28412539e-01 1.37957513e-01
1.20102680e+00 -2.64217317e-01 3.35870594e-01 3.48250955e-01
1.03507042e+00 -1.71574056e-02 -1.47966945e+00 2.80232847e-01
-6.56183958e-01 -8.13439548e-01 5.21353245e-01 -3.57266217e-01
-1.29141998e+00 9.41751361e-01 1.03691816e+00 3.76210719e-01
9.46034670e-01 -4.22858506e-01 2.72067517e-01 4.86436009e-01
3.46276224e-01 -1.11225736e+00 -5.48830688e-01 3.48540723e-01
9.00869071e-01 -1.50340283e+00 2.17684358e-01 -3.76795411e-01
-6.16124868e-01 1.27458405e+00 7.27176666e-01 -5.40996306e-02
3.73976201e-01 5.22641659e-01 2.24271461e-01 -1.23493955e-01
-4.94522572e-01 -4.87620592e-01 1.92838416e-01 6.36762798e-01
7.23316669e-02 -3.93987596e-02 1.00896180e-01 -1.16398580e-01
-1.14849947e-01 -7.79878572e-02 5.03777623e-01 8.02354276e-01
-5.75481594e-01 -5.16840816e-01 -6.71526194e-01 3.01487476e-01
-1.56020433e-01 2.88253307e-01 -8.81125629e-02 1.44774485e+00
4.09288108e-01 9.12158966e-01 4.05052811e-01 -4.16925073e-01
5.64494967e-01 -3.66751403e-01 3.30923229e-01 -2.19022706e-01
-2.63111740e-01 -5.81314713e-02 2.53652304e-01 -4.67555344e-01
-5.46782911e-01 -7.24262297e-01 -9.20314014e-01 -3.54539424e-01
-6.88888431e-01 9.07120407e-02 1.23129725e+00 8.05157125e-01
2.24329188e-01 5.55479884e-01 6.59523308e-01 -1.52396798e+00
-1.36871591e-01 -7.86548495e-01 -5.24740934e-01 -2.29745552e-01
4.48660880e-01 -8.99669945e-01 -4.70493406e-01 5.32874987e-02] | [8.04917049407959, -1.2697261571884155] |
1ab37c0a-8b8b-4768-8590-1cbdd2fdf692 | time-aware-prompting-for-text-generation | 2211.02162 | null | https://arxiv.org/abs/2211.02162v1 | https://arxiv.org/pdf/2211.02162v1.pdf | Time-aware Prompting for Text Generation | In this paper, we study the effects of incorporating timestamps, such as document creation dates, into generation systems. Two types of time-aware prompts are investigated: (1) textual prompts that encode document timestamps in natural language sentences; and (2) linear prompts that convert timestamps into continuous vectors. To explore extrapolation to future data points, we further introduce a new data-to-text generation dataset, TempWikiBio, containing more than 4 millions of chronologically ordered revisions of biographical articles from English Wikipedia, each paired with structured personal profiles. Through data-to-text generation on TempWikiBio, text-to-text generation on the content transfer dataset, and summarization on XSum, we show that linear prompts on encoder and textual prompts improve the generation quality on all datasets. Despite having less performance drop when testing on data drawn from a later time, linear prompts focus more on non-temporal information and are less sensitive to the given timestamps, according to human evaluations and sensitivity analyses. Meanwhile, textual prompts establish the association between the given timestamps and the output dates, yielding more factual temporal information in the output. | ['Lu Wang', 'Shuyang Cao'] | 2022-11-03 | null | null | null | null | ['data-to-text-generation'] | ['natural-language-processing'] | [ 1.12353235e-01 6.91525489e-02 -2.45728642e-01 -4.23114955e-01
-6.72462523e-01 -7.79957294e-01 1.33751845e+00 6.81401134e-01
-5.99210620e-01 9.42836046e-01 1.03710854e+00 -8.05083364e-02
-4.98228483e-02 -8.84020329e-01 -7.89074659e-01 -4.68513779e-02
-3.40232462e-01 4.58876520e-01 2.27708608e-01 -5.02458692e-01
4.54177350e-01 -3.73166949e-01 -1.39731455e+00 6.56175137e-01
1.04257667e+00 8.09218466e-01 1.45963684e-01 8.80477726e-01
-2.55739480e-01 8.29797983e-01 -1.09064472e+00 -5.84704041e-01
1.49426177e-01 -3.04670364e-01 -7.02930272e-01 -4.93244320e-01
3.90464693e-01 -3.97799313e-01 -4.32112247e-01 5.36315262e-01
5.73963523e-01 -1.25584006e-01 7.12579012e-01 -1.32082164e+00
-1.00332499e+00 1.44562495e+00 -3.34084868e-01 4.79956895e-01
9.22565818e-01 2.20094636e-01 1.07513094e+00 -4.75655347e-01
1.20265639e+00 1.05288076e+00 7.98400640e-01 2.48124450e-01
-9.60746765e-01 -4.38915104e-01 -2.31446251e-02 2.08955809e-01
-8.24238956e-01 -4.19160664e-01 5.16599476e-01 -4.77486283e-01
1.00019407e+00 4.25616443e-01 4.91377264e-01 1.85542428e+00
5.35663545e-01 6.64699316e-01 5.94639599e-01 -2.69631326e-01
1.76325902e-01 -1.60781115e-01 3.05346191e-01 1.83643252e-01
2.36717701e-01 1.03378743e-01 -1.29240978e+00 -4.59564894e-01
1.78203523e-01 -2.27224350e-01 -2.87922889e-01 6.60082579e-01
-1.79620147e+00 7.66316414e-01 8.70792642e-02 2.82822847e-01
-7.15463281e-01 2.26358324e-01 7.88177907e-01 3.21444422e-01
7.76427746e-01 8.44026566e-01 -4.37461466e-01 -7.97958434e-01
-1.29487765e+00 5.58950245e-01 1.06443858e+00 1.34047306e+00
3.05007190e-01 -1.87421083e-01 -8.78928065e-01 4.57998335e-01
-1.09079368e-01 6.47067428e-01 9.61418450e-01 -6.44995332e-01
1.07713616e+00 4.38342184e-01 4.20717150e-01 -1.07966149e+00
-4.36294854e-01 -1.29374966e-01 -5.32322407e-01 -8.16739440e-01
3.61897290e-01 -5.70948541e-01 -6.91902220e-01 1.69520319e+00
-5.20940088e-02 -3.40785719e-02 7.47971460e-02 5.43217242e-01
8.78346741e-01 9.58005011e-01 -1.38963073e-01 -3.39672595e-01
1.48280835e+00 -4.75238025e-01 -8.86586189e-01 -1.61535934e-01
6.09454453e-01 -8.63386273e-01 1.16705155e+00 2.27689907e-01
-9.98053193e-01 -3.69287819e-01 -9.72354591e-01 -2.91489273e-01
-4.75907892e-01 6.66417331e-02 4.59223270e-01 1.96299136e-01
-1.03326285e+00 8.92172873e-01 -7.22374380e-01 -6.06528163e-01
-1.46642536e-01 -3.56865138e-01 -1.32447764e-01 2.83372253e-01
-1.85810673e+00 5.82266569e-01 4.33847219e-01 -3.23779494e-01
-4.13270503e-01 -1.20598459e+00 -6.68551385e-01 -2.91570313e-02
1.38324931e-01 -9.21620131e-01 1.61753345e+00 -2.05569342e-01
-1.21864176e+00 2.97259539e-01 -2.86731750e-01 -8.00198376e-01
7.96629727e-01 -4.23520625e-01 -4.87372369e-01 1.39333576e-01
3.51105571e-01 4.52379644e-01 5.78558624e-01 -7.74008393e-01
-7.98600018e-01 1.68370809e-02 -1.01115577e-01 -3.01426668e-02
-7.79707551e-01 6.45983815e-02 -3.85318905e-01 -8.92596900e-01
-4.14205492e-01 -7.15534627e-01 5.30496575e-02 -4.71823305e-01
-9.37866509e-01 -4.13478553e-01 4.00501549e-01 -1.00715137e+00
1.83726943e+00 -1.77114522e+00 -5.48628718e-03 -1.84439540e-01
7.40007758e-02 -2.88191646e-01 -1.98209301e-01 1.20413184e+00
3.33188087e-01 3.52300733e-01 6.16321005e-02 -4.51142460e-01
2.02242136e-01 -1.33214593e-01 -8.30506265e-01 -4.91601489e-02
-6.67629540e-02 1.02973378e+00 -1.26045573e+00 -4.85015988e-01
-4.08357173e-01 1.40882969e-01 -4.45627332e-01 4.92078699e-02
-5.67072690e-01 -6.38936684e-02 -3.40826541e-01 3.17046553e-01
6.04474656e-02 -3.10561299e-01 -2.11421520e-01 -2.11624682e-01
-3.58626485e-01 7.99189806e-01 -6.01783752e-01 1.75774288e+00
-3.50605637e-01 9.00583267e-01 -8.08290660e-01 1.77786484e-01
7.67583072e-01 5.51764786e-01 5.39889634e-01 -9.27206099e-01
-2.78025299e-01 -1.66277364e-01 -4.36639160e-01 -5.61956823e-01
1.57160473e+00 4.01742011e-01 -5.35009861e-01 9.13445234e-01
-1.60854399e-01 -4.14999500e-02 9.72912133e-01 8.52881312e-01
1.34049320e+00 -8.54362622e-02 3.85981873e-02 5.77464886e-02
-2.62107760e-01 9.38844606e-02 3.06390256e-01 9.00037169e-01
3.78385186e-01 8.00369740e-01 7.96355426e-01 -2.13101134e-01
-1.40290523e+00 -9.44157362e-01 1.99185833e-02 9.66209233e-01
-1.45701006e-01 -1.30059969e+00 -5.10830462e-01 -5.03887177e-01
3.27065349e-01 1.30848324e+00 -8.19756985e-01 -2.65227377e-01
-3.50305170e-01 -7.39589393e-01 9.59747434e-01 5.55299878e-01
2.05758587e-02 -9.35215473e-01 -8.69065702e-01 5.26449382e-01
-8.39570880e-01 -9.62677658e-01 -9.24906373e-01 -1.67419106e-01
-7.35066116e-01 -8.01997960e-01 -8.13358545e-01 9.39333141e-02
3.93312365e-01 -1.22045562e-01 1.15424407e+00 -3.11167359e-01
-1.09221716e-03 5.32475948e-01 -8.19155216e-01 -6.66152596e-01
-5.08406818e-01 5.01426280e-01 1.69705510e-01 -2.46712714e-01
3.96643532e-03 -5.51512122e-01 -6.01990759e-01 -5.34123331e-02
-1.01563573e+00 3.10178906e-01 2.02322423e-01 6.80659711e-01
-1.70744702e-01 -2.78533816e-01 7.15176642e-01 -1.11430824e+00
1.30528307e+00 -8.71818125e-01 -2.16403693e-01 2.19912291e-01
-8.36458802e-01 2.48702779e-01 8.64911318e-01 -5.13132632e-01
-1.13773119e+00 -6.16394997e-01 3.78034413e-01 1.34913668e-01
3.78045738e-01 1.02665126e+00 5.29937387e-01 1.05155146e+00
1.16880739e+00 4.49809045e-01 -3.53710234e-01 -4.05733317e-01
6.45458639e-01 7.15979457e-01 8.93636525e-01 -8.09424937e-01
7.79295444e-01 2.20284149e-01 -6.87006533e-01 -5.66012025e-01
-5.70918918e-01 -2.74465531e-01 -2.80502439e-01 -1.70812249e-01
4.47187811e-01 -8.17475080e-01 -3.92599165e-01 5.25518060e-01
-1.46515322e+00 -3.41851413e-01 -3.29540551e-01 2.46446714e-01
-4.94818419e-01 4.30725843e-01 -8.72031331e-01 -3.81778181e-01
-7.89705157e-01 -4.35056835e-01 1.04673374e+00 1.35910630e-01
-9.17788327e-01 -9.69348967e-01 2.33547330e-01 -6.50881156e-02
4.69161868e-01 3.53036523e-01 9.48212266e-01 -8.04236531e-01
-2.83749759e-01 -3.04232836e-01 -2.05923598e-02 -1.19031057e-01
3.12036991e-01 5.86422622e-01 -7.36588717e-01 -2.18564734e-01
-3.63437414e-01 -7.98748061e-03 6.45735204e-01 1.06515788e-01
7.94423521e-01 -9.46908534e-01 -3.34787548e-01 3.87271404e-01
8.93438399e-01 2.04529405e-01 5.44819355e-01 4.69201595e-01
3.89610767e-01 5.28016269e-01 7.12946236e-01 1.19996345e+00
8.86581481e-01 5.10104120e-01 1.22208223e-01 2.80567616e-01
6.25959635e-02 -6.53907180e-01 6.02465332e-01 1.05926645e+00
-7.49252290e-02 -8.46448481e-01 -8.48286808e-01 6.74614787e-01
-1.76936221e+00 -1.28320932e+00 -2.66057283e-01 2.10103345e+00
1.46215940e+00 2.53810197e-01 3.95938456e-02 2.22829208e-02
5.71936071e-01 4.81400669e-01 -3.24366570e-01 -2.30459988e-01
-2.80294180e-01 -3.03746432e-01 5.62968910e-01 1.67308643e-01
-6.45228803e-01 6.91456676e-01 6.44456482e+00 5.16521275e-01
-1.30886090e+00 -4.09069695e-02 2.09865674e-01 -5.69609106e-01
-7.96933830e-01 1.19448770e-02 -8.82272363e-01 1.36901271e+00
1.35305429e+00 -1.31122851e+00 3.09646606e-01 4.89544332e-01
5.33651292e-01 -2.78243780e-01 -1.61517859e+00 6.32748187e-01
3.26946229e-02 -1.69061863e+00 2.88489550e-01 -7.81110525e-02
8.28835726e-01 1.27085811e-02 -4.14000042e-02 3.04113209e-01
5.25261104e-01 -5.14186740e-01 1.29413581e+00 8.16471338e-01
7.92065382e-01 -4.73816395e-01 5.79510391e-01 3.64445716e-01
-8.20794940e-01 8.16170573e-02 -3.48530740e-01 -1.79560948e-02
6.14683270e-01 1.24500883e+00 -1.34689617e+00 9.50207651e-01
5.38924217e-01 1.02217984e+00 -9.64973927e-01 9.19080377e-01
-2.05172017e-01 7.55252540e-01 -3.80149007e-01 -2.43291408e-01
4.24078554e-02 2.19496816e-01 6.11918807e-01 1.54891086e+00
7.41457224e-01 -1.82564035e-01 -3.32944661e-01 7.10073709e-01
-2.99560338e-01 -1.87350705e-01 -5.68091154e-01 -4.17185336e-01
1.05615735e+00 1.09185314e+00 -4.75657791e-01 -6.30884290e-01
-1.57169718e-02 9.85646129e-01 1.26948297e-01 5.03979743e-01
-9.28403497e-01 -5.90331137e-01 3.63572091e-01 3.10307890e-01
-1.08911777e-02 -3.28536272e-01 -2.97246277e-01 -1.17194867e+00
2.30384246e-01 -6.67057455e-01 4.37968582e-01 -1.05960035e+00
-1.26338112e+00 6.35351062e-01 2.70474374e-01 -1.14463794e+00
-8.59567046e-01 1.46095932e-01 -8.18722069e-01 7.86253989e-01
-1.09875941e+00 -5.78338563e-01 -3.81398112e-01 2.50928909e-01
4.40752149e-01 4.50635329e-02 5.53725064e-01 3.09041023e-01
-1.95001557e-01 6.61376238e-01 2.67499238e-01 1.58411041e-02
1.28464973e+00 -1.33849311e+00 1.21726418e+00 9.82639015e-01
1.68013498e-02 1.07026160e+00 1.10144258e+00 -1.18036425e+00
-1.47030151e+00 -1.22772241e+00 1.55130029e+00 -7.35552192e-01
8.06310713e-01 -4.33462799e-01 -7.07150519e-01 8.05320144e-01
6.02856100e-01 -6.71043992e-01 4.43135381e-01 2.70069510e-01
-2.76401341e-01 1.34931892e-01 -7.27343500e-01 7.95490801e-01
1.02279389e+00 -6.50592446e-01 -8.21496904e-01 5.31078815e-01
1.20373440e+00 -6.18971288e-01 -1.15046811e+00 -2.17411965e-02
4.50351208e-01 -5.20018399e-01 6.52534425e-01 -6.63394034e-01
1.20645571e+00 -1.45925730e-01 2.05686525e-01 -1.84829080e+00
-1.63772851e-01 -9.62477267e-01 -3.94177407e-01 1.71770275e+00
6.32812440e-01 -5.69881618e-01 3.65533769e-01 5.73440552e-01
-2.28308469e-01 -4.02277470e-01 -9.14395392e-01 -9.28841531e-01
-3.09220046e-01 -3.60192299e-01 9.96887684e-01 1.05579424e+00
3.55495155e-01 2.79958993e-01 -5.30486882e-01 -2.60507852e-01
1.84064314e-01 -8.18949714e-02 7.94328094e-01 -1.12286413e+00
2.99400929e-03 -2.59917855e-01 6.92207590e-02 -8.20194483e-01
-3.86445194e-01 -9.89281297e-01 9.97689888e-02 -1.77374101e+00
1.32201269e-01 -1.91200510e-01 1.87910005e-01 4.57267433e-01
-3.07354182e-01 -2.64358044e-01 2.22241655e-01 1.95572525e-01
-5.42789161e-01 7.30628908e-01 9.50650156e-01 -6.37123510e-02
-1.69834927e-01 -3.20339561e-01 -7.52726018e-01 1.87951058e-01
4.40547019e-01 -6.61474288e-01 -4.74561304e-01 -6.43944919e-01
7.62517810e-01 3.77705455e-01 1.92024603e-01 -8.75879228e-01
6.39509201e-01 -1.60268813e-01 2.80309707e-01 -8.78547251e-01
8.90246555e-02 -7.45013505e-02 3.13002199e-01 2.53967494e-01
-7.94868529e-01 5.80243409e-01 7.43334442e-02 5.08527040e-01
1.39475185e-02 -1.46868229e-01 -1.31892830e-01 1.47655681e-01
-3.64557058e-01 1.90562159e-01 -6.32044196e-01 4.78724867e-01
6.42792404e-01 -1.32675931e-01 -1.14300632e+00 -6.65032566e-01
-2.13845298e-01 3.29384208e-01 5.21473885e-01 8.68237317e-01
3.28503132e-01 -1.29299212e+00 -9.14587021e-01 -1.73321620e-01
4.03988212e-01 -1.84356138e-01 6.31811544e-02 5.57910025e-01
-2.12059915e-01 6.11242414e-01 -1.55064166e-01 -4.26000237e-01
-8.93228710e-01 2.57238567e-01 -4.50260103e-01 -2.87251085e-01
-6.67246819e-01 6.32590175e-01 -3.05719942e-01 -1.70882300e-01
7.74856433e-02 -8.44705105e-01 -1.07910000e-02 7.07784712e-01
7.51528740e-01 6.03972733e-01 9.50562805e-02 5.32379607e-03
8.06611590e-03 -1.38030916e-01 -4.10240740e-01 -4.94213104e-01
1.35167980e+00 -1.20031044e-01 4.96124737e-02 7.60963142e-01
9.50704157e-01 1.80242419e-01 -1.23663831e+00 -3.76076490e-01
4.81135756e-01 -2.64179349e-01 -3.62031490e-01 -1.05455828e+00
-6.66089594e-01 2.97654390e-01 -1.32163912e-01 4.81478542e-01
6.37360275e-01 -1.64088607e-01 1.16559446e+00 3.05102378e-01
5.02114534e-01 -1.19593441e+00 2.26351172e-01 8.85815084e-01
1.11711395e+00 -7.85501301e-01 1.65973783e-01 1.30751058e-01
-6.84764147e-01 1.05634320e+00 3.95727992e-01 5.01881659e-01
1.28329664e-01 2.19932899e-01 -3.91921541e-03 4.70674224e-02
-1.49629855e+00 4.00154501e-01 1.85722187e-01 4.38526928e-01
6.46529138e-01 1.59913823e-02 -5.74861109e-01 9.89523053e-01
-1.01920378e+00 1.40325725e-01 1.09324384e+00 8.51123095e-01
3.90796252e-02 -8.67835939e-01 -3.19126576e-01 8.93819511e-01
-2.27603152e-01 -4.21556503e-01 -4.42209303e-01 5.02957344e-01
-3.74407977e-01 9.96294737e-01 2.88017273e-01 -4.91269320e-01
3.20777386e-01 8.87782797e-02 6.50954619e-02 -6.15077734e-01
-8.91515672e-01 -5.73139966e-01 5.88302612e-01 -4.35689688e-01
3.40419859e-01 -1.03516674e+00 -1.27479088e+00 -7.05004096e-01
-2.60971170e-02 3.74621689e-01 8.53150964e-01 6.28332973e-01
7.38018274e-01 7.76403904e-01 6.04016304e-01 -7.47722447e-01
-7.81696677e-01 -1.35697973e+00 -2.35760301e-01 5.51694870e-01
1.35723278e-01 -1.05316617e-01 -3.19982409e-01 4.51079011e-01] | [12.245514869689941, 9.102269172668457] |
92d57a44-3114-43c6-a837-e1af1cfe6c80 | siamese-object-tracking-for-unmanned-aerial | 2205.04281 | null | https://arxiv.org/abs/2205.04281v2 | https://arxiv.org/pdf/2205.04281v2.pdf | Siamese Object Tracking for Unmanned Aerial Vehicle: A Review and Comprehensive Analysis | Unmanned aerial vehicle (UAV)-based visual object tracking has enabled a wide range of applications and attracted increasing attention in the field of intelligent transportation systems because of its versatility and effectiveness. As an emerging force in the revolutionary trend of deep learning, Siamese networks shine in UAV-based object tracking with their promising balance of accuracy, robustness, and speed. Thanks to the development of embedded processors and the gradual optimization of deep neural networks, Siamese trackers receive extensive research and realize preliminary combinations with UAVs. However, due to the UAV's limited onboard computational resources and the complex real-world circumstances, aerial tracking with Siamese networks still faces severe obstacles in many aspects. To further explore the deployment of Siamese networks in UAV-based tracking, this work presents a comprehensive review of leading-edge Siamese trackers, along with an exhaustive UAV-specific analysis based on the evaluation using a typical UAV onboard processor. Then, the onboard tests are conducted to validate the feasibility and efficacy of representative Siamese trackers in real-world UAV deployment. Furthermore, to better promote the development of the tracking community, this work analyzes the limitations of existing Siamese trackers and conducts additional experiments represented by low-illumination evaluations. In the end, prospects for the development of Siamese tracking for UAV-based intelligent transportation systems are deeply discussed. The unified framework of leading-edge Siamese trackers, i.e., code library, and the results of their experimental evaluations are available at https://github.com/vision4robotics/SiameseTracking4UAV . | ['Geng Lu', 'Bowen Li', 'Ziang Cao', 'Junjie Ye', 'Guangze Zheng', 'Kunhan Lu', 'Changhong Fu'] | 2022-05-09 | null | null | null | null | ['visual-object-tracking'] | ['computer-vision'] | [-5.66432714e-01 -6.76765859e-01 -2.94498920e-01 3.40262562e-01
7.29738027e-02 -7.37169325e-01 3.73033375e-01 -3.86124969e-01
-4.11750555e-01 4.90741789e-01 -6.71454430e-01 -1.50418580e-01
-1.17059849e-01 -4.51396465e-01 -6.23487234e-01 -8.09274077e-01
-4.30840403e-01 1.79960579e-01 4.03863758e-01 -3.75304490e-01
-3.55911851e-01 9.99425888e-01 -1.53251636e+00 -6.92719460e-01
5.93940079e-01 1.08402610e+00 1.48331016e-01 6.88396871e-01
5.52385390e-01 2.62369215e-01 -6.94472492e-01 -2.96980143e-01
5.14226675e-01 1.32354140e-01 4.46811952e-02 -3.14652026e-02
8.32708359e-01 -3.75352025e-01 -5.78319848e-01 1.10826337e+00
3.94061267e-01 2.51420975e-01 3.03372651e-01 -2.02752399e+00
-5.34593403e-01 4.79063466e-02 -4.53847110e-01 2.56980032e-01
-1.05089461e-02 5.32168329e-01 6.57229364e-01 -7.60800660e-01
4.71879721e-01 7.14920580e-01 1.04565763e+00 3.27945679e-01
-5.39483190e-01 -8.38667333e-01 3.05774808e-01 1.54433683e-01
-1.48083508e+00 -2.07680628e-01 6.31046593e-01 -4.67080772e-01
8.54757309e-01 2.42490452e-02 1.25247657e+00 1.03614378e+00
5.14737844e-01 8.69228125e-01 3.36941898e-01 -3.03583834e-02
-1.66837294e-02 -8.71923864e-02 4.61806022e-02 1.25757635e+00
9.21007752e-01 8.45327735e-01 -2.50822008e-01 1.12784140e-01
5.77459991e-01 -1.45838391e-02 -3.24184150e-01 -8.53109658e-01
-1.36482263e+00 5.14566302e-01 9.51000869e-01 5.92511594e-02
-3.55047554e-01 5.80072582e-01 3.31817001e-01 2.07555816e-02
1.36229292e-01 3.53559732e-01 -2.79570341e-01 5.59083298e-02
-9.71529126e-01 2.77202815e-01 4.97303635e-01 1.43576419e+00
2.71967649e-01 8.87855291e-01 -1.78724919e-02 1.64754689e-01
6.41397715e-01 1.18415785e+00 2.95517221e-02 -7.64486670e-01
2.23011672e-02 4.52521503e-01 5.50546825e-01 -1.24329209e+00
-5.03360748e-01 -8.20411742e-01 -6.01816177e-01 5.43124557e-01
3.44995290e-01 -5.48167109e-01 -5.68614304e-01 1.48087358e+00
6.55534565e-01 2.17261627e-01 4.06559929e-02 1.25207412e+00
8.34691107e-01 4.36323404e-01 1.76606532e-02 1.18488789e-01
1.31006193e+00 -1.19068718e+00 -5.52237511e-01 -1.33695588e-01
5.39629579e-01 -6.05734825e-01 4.47878301e-01 3.94241810e-02
-4.66490597e-01 -7.65291512e-01 -1.24452436e+00 4.96602803e-01
-5.92104077e-01 8.47661257e-01 5.41426957e-01 8.15011501e-01
-1.03428304e+00 9.70061570e-02 -1.16333961e+00 -7.56234348e-01
5.10983527e-01 5.47787964e-01 -1.00509435e-01 1.69050395e-01
-1.01818657e+00 9.16117072e-01 1.57996222e-01 3.60313147e-01
-1.24496150e+00 -5.51384270e-01 -7.25255907e-01 -4.86753928e-03
5.35160065e-01 -9.13337946e-01 1.27573061e+00 -5.72358847e-01
-1.33356130e+00 3.64156365e-01 3.51565212e-01 -7.67575741e-01
4.89828616e-01 -2.00165838e-01 -4.60960627e-01 3.38097066e-02
1.59364119e-01 8.74952257e-01 9.78353620e-01 -9.09188390e-01
-1.06007946e+00 -2.44431183e-01 7.97574371e-02 7.77185857e-02
-4.88471001e-01 -1.07755974e-01 -4.19767082e-01 -5.07865906e-01
-4.58190501e-01 -1.29728210e+00 6.11765310e-02 5.76339424e-01
-7.45533779e-02 -2.32130036e-01 1.42644823e+00 1.02833524e-01
9.43532288e-01 -2.05661726e+00 -1.50477141e-01 -3.06047440e-01
4.54457521e-01 8.96127224e-01 -2.66527593e-01 3.09813410e-01
2.84499019e-01 -4.93485332e-01 2.42208883e-01 -5.25008105e-02
4.21881527e-02 -1.88240752e-01 -1.99999079e-01 7.66927958e-01
-4.31280211e-02 9.68920350e-01 -1.01873374e+00 -3.06399703e-01
5.16116381e-01 6.08690977e-01 -5.16361818e-02 1.64171718e-02
-1.09748311e-01 7.26123825e-02 -6.00450456e-01 1.16765356e+00
6.02994502e-01 -1.27711535e-01 -3.89258772e-01 -5.79734623e-01
-7.33385503e-01 -6.50158405e-01 -8.27545345e-01 1.09402430e+00
-2.34197211e-02 1.12106299e+00 4.69204903e-01 -5.57312191e-01
8.76331687e-01 1.40685156e-01 5.85542142e-01 -1.51039839e-01
5.56935370e-01 2.20664844e-01 1.33561954e-01 -1.93634108e-01
8.11023951e-01 3.50836456e-01 1.10386163e-01 -8.39371458e-02
8.52439851e-02 -8.09019133e-02 4.85901445e-01 1.75221354e-01
7.76223660e-01 3.45985442e-01 9.71137881e-02 -1.94563866e-01
4.51889724e-01 7.80881941e-01 3.60942394e-01 4.16498721e-01
-8.37289810e-01 -8.98906440e-02 -2.94431061e-01 -6.94764495e-01
-5.98953962e-01 -7.20295846e-01 -5.30089624e-02 6.38946235e-01
6.56810880e-01 -2.23681182e-01 -5.73953688e-01 -7.01074839e-01
3.36198062e-01 2.67667055e-01 -4.49924409e-01 -2.10490540e-01
-2.98295051e-01 -6.55547142e-01 9.25907135e-01 7.27701962e-01
7.58091688e-01 -6.90541565e-01 -1.32616282e+00 7.83617422e-02
1.94508433e-01 -1.35040963e+00 -4.69206750e-01 -1.34547800e-01
-6.20143354e-01 -1.38849783e+00 -8.30983460e-01 -7.01112568e-01
3.78183693e-01 1.15162742e+00 5.14657617e-01 6.77464977e-02
-3.44547361e-01 9.15708423e-01 -2.00493604e-01 -7.09924638e-01
-2.14767158e-02 -8.65344610e-03 6.59602523e-01 -3.03535819e-01
3.25770617e-01 1.71485856e-01 -4.24751133e-01 6.25799298e-01
-3.74618143e-01 -4.93869126e-01 4.31793958e-01 6.82829320e-01
4.06708032e-01 2.27361605e-01 4.61754426e-02 1.45773664e-01
2.84403354e-01 -3.14182043e-01 -1.68588090e+00 2.30652869e-01
-6.30394936e-01 -6.25581086e-01 7.13733673e-01 -5.01173913e-01
-5.25097668e-01 1.81153476e-01 3.19417566e-01 -1.01831877e+00
-4.61774990e-02 4.01184738e-01 1.91311553e-01 -9.03292477e-01
4.55785483e-01 6.87446818e-02 1.66513726e-01 2.43386999e-01
1.19127654e-01 3.47725391e-01 4.39538479e-01 -2.29225442e-01
1.13605297e+00 5.11391580e-01 2.24110559e-01 -1.26650155e+00
-5.84013879e-01 -2.58918017e-01 -3.52110207e-01 -9.52431500e-01
8.39719534e-01 -1.01831996e+00 -1.02589226e+00 6.32367373e-01
-1.15575898e+00 -3.02363783e-01 -2.34739631e-01 7.90708065e-01
-2.16396287e-01 2.50483185e-01 -1.89165562e-01 -8.50602925e-01
-2.86652952e-01 -1.37977040e+00 9.29067075e-01 7.13164508e-01
2.16547906e-01 -1.03138101e+00 5.11931488e-03 2.17493236e-01
5.27536631e-01 2.34933794e-01 -3.49269807e-02 -2.33994827e-01
-1.06712425e+00 -6.04150414e-01 -2.34101057e-01 1.35415103e-02
-4.17149179e-02 6.05590641e-01 -5.64395308e-01 -7.85982430e-01
-5.26883423e-01 5.90658486e-02 6.48891628e-01 6.78452134e-01
2.65386224e-01 2.49681696e-01 -8.04006100e-01 8.60855222e-01
1.30876493e+00 5.18092036e-01 -1.97906882e-01 5.34820914e-01
7.00797081e-01 3.83115672e-02 1.17873716e+00 3.55256319e-01
3.82938236e-01 8.24631214e-01 1.15602422e+00 -1.06246047e-01
-3.99168402e-01 -1.49283722e-01 6.05317712e-01 4.99661893e-01
-2.98905134e-01 -7.94604123e-01 -8.15643489e-01 4.21458721e-01
-1.51035511e+00 -8.75719488e-01 -4.13190067e-01 2.08144021e+00
-4.74257082e-01 -1.32327989e-01 3.46081764e-01 -4.39978272e-01
9.26354766e-01 3.77460599e-01 -7.18916535e-01 2.87177354e-01
-1.17836066e-01 -4.77281868e-01 8.56012821e-01 6.32634461e-02
-1.55201340e+00 1.31872046e+00 5.75685978e+00 5.74206531e-01
-1.27055717e+00 -8.60855132e-02 -3.41624796e-01 1.47055730e-01
7.20385015e-01 -2.67521590e-01 -1.39663053e+00 4.24690336e-01
7.61342525e-01 -2.45880917e-01 3.29518229e-01 1.23646796e+00
8.21180642e-02 -1.01589918e-01 -5.76206207e-01 8.81405473e-01
6.96179122e-02 -1.38261497e+00 -1.47910684e-01 1.80922717e-01
6.24195218e-01 6.50739610e-01 3.50871444e-01 1.68071166e-01
3.25354308e-01 -2.85012871e-01 8.80963385e-01 2.02913105e-01
7.37122297e-01 -6.14612222e-01 9.04703379e-01 6.80494085e-02
-1.70536292e+00 -2.59380221e-01 -5.51381588e-01 9.98032540e-02
2.46657491e-01 1.82037789e-03 -5.85225523e-01 6.57049537e-01
9.13746774e-01 1.06946647e+00 -3.89992088e-01 1.55783272e+00
-2.22876489e-01 3.44289422e-01 -5.28993666e-01 -4.59793270e-01
7.41181970e-01 -2.22522601e-01 1.05434012e+00 9.16809440e-01
6.89409077e-01 -7.76295885e-02 4.32425290e-01 7.56305993e-01
8.58631060e-02 -4.08805609e-01 -9.91443992e-01 -3.34312260e-01
6.41406238e-01 1.62547517e+00 -8.13414335e-01 -2.17549086e-01
-4.64386165e-01 3.80576611e-01 -1.29398346e-01 4.00869519e-01
-1.35772610e+00 -5.09161294e-01 9.84975934e-01 -4.37355489e-02
4.30223554e-01 -7.61991858e-01 3.43591481e-01 -1.02855337e+00
-4.00929302e-01 -5.36818802e-01 6.04683906e-02 -9.09661293e-01
-7.96961367e-01 8.75131249e-01 -4.47638147e-02 -1.91834795e+00
2.11454719e-01 -9.97239888e-01 -5.34041703e-01 3.04124594e-01
-1.48845160e+00 -1.45305991e+00 -9.18281436e-01 3.94254923e-01
5.31487286e-01 -7.29442358e-01 4.31492925e-01 2.97556192e-01
-9.46970522e-01 4.94855851e-01 5.37704639e-02 1.89392388e-01
5.49431503e-01 -5.74937642e-01 4.38021243e-01 1.26652217e+00
2.78249443e-01 4.85805899e-01 6.30160689e-01 -7.28625357e-01
-1.96410286e+00 -1.31378818e+00 2.69416958e-01 -3.47131163e-01
1.05915821e+00 4.21497561e-02 -3.16964388e-01 7.98635244e-01
2.88334966e-01 3.09013635e-01 2.09681794e-01 -4.45503473e-01
1.78229466e-01 -3.22624207e-01 -8.40495527e-01 7.06644297e-01
9.13699806e-01 1.09916247e-01 -5.80103397e-02 4.33146983e-01
5.44116318e-01 -6.12175703e-01 -7.46940911e-01 4.37560856e-01
8.10938776e-01 -7.35900879e-01 9.32110786e-01 -3.89852673e-01
-4.21185583e-01 -1.01696551e+00 -5.97900227e-02 -1.34755957e+00
-5.50126195e-01 -5.47510982e-01 -2.70789444e-01 8.88440430e-01
-2.66342852e-02 -8.30914795e-01 9.03415501e-01 8.34323689e-02
-5.52284539e-01 -4.23791409e-01 -7.69909501e-01 -1.35706711e+00
-4.97639745e-01 -9.26937759e-02 4.66193318e-01 6.38215721e-01
-5.36721647e-01 1.49833664e-01 -4.58316863e-01 7.87332416e-01
9.44201291e-01 1.21122718e-01 1.07053792e+00 -1.37797153e+00
3.76226902e-01 -4.25084233e-01 -7.13730335e-01 -1.14900887e+00
1.05315171e-01 -6.83933675e-01 1.44030964e-02 -1.26893389e+00
-4.60511923e-01 -3.46287042e-01 -5.52150281e-03 3.26446593e-01
1.49218872e-01 3.44289005e-01 4.49525744e-01 1.38489813e-01
-8.98344398e-01 8.57846916e-01 1.23138237e+00 -4.51486200e-01
8.76050536e-03 3.00581962e-01 -6.51194826e-02 5.43528318e-01
7.68275797e-01 -3.93479258e-01 -3.33301485e-01 -6.70161963e-01
-2.20730782e-01 -3.08836475e-02 6.43422365e-01 -1.60632622e+00
8.25932384e-01 1.16521329e-01 3.62208605e-01 -7.38410532e-01
5.04362047e-01 -1.32108784e+00 1.66790634e-01 8.74073744e-01
4.35425073e-01 4.78991419e-01 5.35557806e-01 5.11718631e-01
-3.72775167e-01 4.36952226e-02 8.81231904e-01 1.92384705e-01
-1.13023353e+00 7.37590075e-01 -7.48617291e-01 -1.74523547e-01
1.55406237e+00 -5.66691220e-01 -5.73842883e-01 -6.43940866e-02
-2.89973915e-01 5.07369339e-01 6.39119446e-01 4.99721736e-01
6.72937751e-01 -1.24873841e+00 -2.29983196e-01 3.36020291e-01
4.37441319e-02 -4.48147714e-01 1.20311342e-01 1.19857776e+00
-6.30333841e-01 7.84463108e-01 -5.61935425e-01 -8.80479872e-01
-1.36484945e+00 6.91776633e-01 4.48531181e-01 3.30263793e-01
-3.86219889e-01 8.26725364e-01 -1.99090596e-02 -1.10853925e-01
3.16370517e-01 -3.23730886e-01 -1.74769200e-03 -2.97946874e-02
2.14906543e-01 7.08697736e-01 -1.09424397e-01 -8.84688377e-01
-6.59733951e-01 9.08860207e-01 1.88586235e-01 5.31282783e-01
9.65294778e-01 2.46603768e-02 2.81136334e-01 -1.62942350e-01
5.10067403e-01 -2.48438269e-01 -1.56812787e+00 3.64809692e-01
-2.17032298e-01 -3.38859707e-01 3.28746080e-01 -5.11949241e-01
-1.49177659e+00 6.92107081e-01 8.68173659e-01 1.92417577e-01
8.25940788e-01 -3.90301377e-01 7.89822280e-01 6.05644763e-01
8.25568199e-01 -7.30508804e-01 -7.14173093e-02 6.85890853e-01
4.65039283e-01 -1.26490462e+00 1.08235791e-01 -2.27422640e-01
-4.52300936e-01 1.13829029e+00 9.60240602e-01 -2.27237031e-01
4.78145897e-01 4.53173131e-01 2.77424663e-01 -3.59298110e-01
-5.07761359e-01 -4.30053443e-01 1.46876693e-01 8.96530509e-01
7.86410570e-02 -3.47409807e-02 5.03108390e-02 -1.91739742e-02
1.04807019e-01 1.88656393e-02 3.25413257e-01 8.10560644e-01
-4.77243692e-01 -6.49705350e-01 -5.65259874e-01 1.04287595e-01
6.92088827e-02 2.59945631e-01 -2.19476372e-01 1.39358246e+00
1.36713460e-01 8.93241465e-01 -2.10148573e-01 -5.23393571e-01
4.99446422e-01 -5.71159899e-01 3.55321229e-01 -2.77193468e-02
-5.16535282e-01 -1.58740580e-01 -8.49710628e-02 -5.98057270e-01
-5.45097172e-01 -6.70816898e-01 -9.74365711e-01 -3.09865922e-01
-6.61046922e-01 2.42226139e-01 8.27971399e-01 5.15505731e-01
5.87045491e-01 5.77003419e-01 2.04765826e-01 -1.14687741e+00
-5.02382100e-01 -4.76267457e-01 -4.59670246e-01 -6.63111210e-01
4.23499316e-01 -1.36735690e+00 -2.26960525e-01 -2.09505573e-01] | [6.569521427154541, -2.01816725730896] |
b73af49d-c88c-4bdd-b8b5-c9c8991dbcb8 | using-pre-trained-transformer-for-better-lay | null | null | https://aclanthology.org/2020.sdp-1.38 | https://aclanthology.org/2020.sdp-1.38.pdf | Using Pre-Trained Transformer for Better Lay Summarization | In this paper, we tack lay summarization tasks, which aim to automatically produce lay summaries for scientific papers, to participate in the first CL-LaySumm 2020 in SDP workshop at EMNLP 2020. We present our approach of using Pre-training with Extracted Gap-sentences for Abstractive Summarization (PEGASUS; Zhang et al., 2019b) to produce the lay summary and combining those with the extractive summarization model using Bidirectional Encoder Representations from Transformers (BERT; Devlin et al., 2018) and readability metrics that measure the readability of the sentence to further improve the quality of the summary. Our model achieves a remarkable performance on ROUGE metrics, demonstrating the produced summary is more readable while it summarizes the main points of the document. | ['Seungwon Kim'] | null | null | null | null | emnlp-sdp-2020-11 | ['lay-summarization'] | ['natural-language-processing'] | [ 2.83330411e-01 8.84836435e-01 -1.85738593e-01 1.36347534e-02
-1.28283477e+00 -6.25641525e-01 7.62624204e-01 5.07891476e-01
-1.19592749e-01 1.35970259e+00 1.25770557e+00 -1.89803436e-01
-2.66252905e-01 -6.46356463e-01 -8.80800903e-01 -1.34737656e-01
3.29897553e-01 2.92686403e-01 -3.16734165e-01 -7.31720105e-02
7.70746708e-01 2.06071287e-01 -1.06650054e+00 5.46386421e-01
1.45146894e+00 6.08941376e-01 4.02471751e-01 1.36061251e+00
-8.35967138e-02 1.02304816e+00 -1.20132756e+00 -6.36138141e-01
-3.52568895e-01 -8.07617664e-01 -1.27335155e+00 -8.62316489e-02
7.00097919e-01 -3.91506255e-01 -4.25915956e-01 9.75100040e-01
5.11108518e-01 5.16044348e-02 1.01868570e+00 -7.80934036e-01
-8.70564401e-01 1.42067766e+00 -4.63507861e-01 3.69461954e-01
5.26063025e-01 -2.29712129e-01 1.50407887e+00 -6.70821369e-01
8.81727636e-01 1.01478636e+00 3.69302094e-01 7.42053449e-01
-9.39771891e-01 -1.84252709e-01 -1.38131365e-01 3.37491274e-01
-5.50469458e-01 -6.46119177e-01 6.45691097e-01 -1.60647016e-02
1.16957521e+00 6.49174094e-01 7.42890656e-01 1.25843298e+00
4.50736731e-01 1.09860873e+00 4.40944999e-01 -3.50593239e-01
1.62276492e-01 -3.21270943e-01 4.36850756e-01 3.05518270e-01
8.83963287e-01 -7.35526860e-01 -8.55786264e-01 1.87184885e-01
8.06144178e-02 -4.13621396e-01 -3.56024981e-01 3.61097872e-01
-1.44416356e+00 5.91861844e-01 2.72693485e-01 3.63762796e-01
-7.84357071e-01 3.00766621e-02 7.68146157e-01 2.11413041e-01
8.10117543e-01 1.02944517e+00 -1.37358516e-01 -3.57223958e-01
-1.67434537e+00 6.99461460e-01 1.01111531e+00 1.15217006e+00
8.57885554e-02 -1.17775977e-01 -1.09028375e+00 7.06150830e-01
-8.72984603e-02 4.84033376e-01 6.37616277e-01 -1.20931375e+00
9.31364834e-01 4.63361561e-01 5.89706451e-02 -6.30275607e-01
-1.49632588e-01 -6.15731597e-01 -1.08691859e+00 -5.39476097e-01
-3.51144195e-01 -4.12665039e-01 -5.57550788e-01 1.23640513e+00
-3.28585505e-01 -9.36935097e-02 5.48562765e-01 2.95111239e-01
1.72501552e+00 1.22007561e+00 -1.55777097e-01 -6.85223043e-01
1.08681595e+00 -1.34160781e+00 -1.17642987e+00 -2.47727092e-02
6.06103837e-01 -5.44142544e-01 5.56769133e-01 3.46661896e-01
-1.91190517e+00 -4.74447638e-01 -1.53413188e+00 -4.03916240e-01
3.80484089e-02 6.29251659e-01 1.06152073e-01 -7.64624476e-02
-1.28908825e+00 1.17087758e+00 -6.43395245e-01 -2.99699724e-01
6.73521161e-01 -1.64738461e-01 -1.52643651e-01 1.79902107e-01
-1.04221010e+00 9.24983621e-01 5.73829114e-01 -1.46071091e-01
-7.52480745e-01 -1.01917779e+00 -7.54730344e-01 4.54579860e-01
2.89698653e-02 -9.60158288e-01 1.38052392e+00 -3.48353028e-01
-1.61972427e+00 6.72955751e-01 -2.98689604e-01 -8.90935719e-01
5.52397728e-01 -6.77595615e-01 -2.12612599e-02 3.62950176e-01
3.18800241e-01 6.54587805e-01 2.94112951e-01 -8.40122938e-01
-5.87379813e-01 -1.14157498e-01 -1.96109295e-01 3.44897538e-01
-3.50496262e-01 -5.05415238e-02 -6.89541698e-02 -5.48061907e-01
-2.64564663e-01 -2.57797092e-01 7.64984861e-02 -9.07138944e-01
-9.99733627e-01 -6.49985790e-01 3.90406013e-01 -1.37453759e+00
1.62867665e+00 -1.66096592e+00 6.30027115e-01 -5.08611441e-01
4.76504058e-01 4.51974273e-01 -4.44690645e-01 1.01645637e+00
1.59054533e-01 4.09754336e-01 -3.51124555e-01 -5.32603920e-01
1.14345022e-01 -3.65165114e-01 -8.05577040e-01 -3.01652122e-02
4.43634510e-01 1.29071176e+00 -1.13960516e+00 -5.71424842e-01
-1.31554559e-01 2.85726856e-03 -3.12834293e-01 5.37302554e-01
-2.93757558e-01 2.91326642e-01 -2.77131289e-01 -1.23126190e-02
4.82548594e-01 -3.14960554e-02 -1.40653625e-01 -1.03916898e-01
-3.63157660e-01 1.03351128e+00 -4.35459375e-01 1.83317626e+00
-4.39828336e-01 1.18574023e+00 -3.72523755e-01 -8.90388012e-01
9.46518421e-01 4.57963258e-01 2.10292533e-01 -3.56130898e-01
-5.74457808e-04 2.48678654e-01 -2.31361181e-01 -4.62710828e-01
1.12251854e+00 4.36440259e-01 -5.89274429e-02 5.49833953e-01
4.65767652e-01 -6.95549011e-01 1.00637591e+00 8.28188658e-01
1.21783614e+00 1.18712969e-01 6.19814396e-01 -3.72360885e-01
7.09907770e-01 -1.16536573e-01 1.81914970e-01 8.73551548e-01
9.98089388e-02 9.15210485e-01 1.00225496e+00 1.83884893e-02
-1.46166408e+00 -9.45263505e-01 9.87510160e-02 4.44364518e-01
-4.18746054e-01 -9.22877908e-01 -1.15133178e+00 -7.46619225e-01
-3.55811715e-01 1.40007985e+00 -5.24445593e-01 -3.02165985e-01
-6.47405565e-01 -2.86439389e-01 6.82105899e-01 3.12126875e-01
6.10708892e-01 -1.33131826e+00 -5.53309262e-01 2.54928678e-01
-5.87952912e-01 -8.31415534e-01 -4.95161414e-01 -2.25821305e-02
-9.64513302e-01 -6.79664314e-01 -1.18960214e+00 -6.04223371e-01
3.21147382e-01 -4.17150892e-02 1.12583077e+00 -3.47755611e-01
-4.05859435e-04 -7.58348182e-02 -5.38608670e-01 -8.34826887e-01
-8.41187835e-01 6.15616024e-01 -4.21263129e-01 -5.79083979e-01
-1.32152319e-01 -4.10762221e-01 -4.27662224e-01 -8.28889906e-01
-7.24541128e-01 6.13906860e-01 6.91081882e-01 6.95955932e-01
3.69331032e-01 -4.79957253e-01 1.13490188e+00 -8.07997406e-01
1.34787321e+00 -3.37712735e-01 7.16406479e-02 4.73944068e-01
-4.74888176e-01 2.40399778e-01 7.37841487e-01 4.24849475e-03
-1.04915798e+00 -9.31835055e-01 -3.64818096e-01 2.37316832e-01
3.04997236e-01 7.87957788e-01 -7.97127113e-02 8.01651001e-01
4.51931417e-01 5.95428824e-01 -2.40172893e-01 -3.72162700e-01
4.75418657e-01 1.02525914e+00 6.87252641e-01 -2.84062833e-01
2.85076886e-01 -1.98514327e-01 -1.05594091e-01 -7.88325429e-01
-1.55289626e+00 -1.51320651e-01 -5.91899633e-01 -2.25631148e-01
7.47493684e-01 -7.65859425e-01 -3.36424142e-01 -2.03781445e-02
-1.75776088e+00 8.73115063e-02 -7.35722601e-01 3.09910595e-01
-7.92480350e-01 6.51219606e-01 -7.38730133e-01 -5.79928815e-01
-1.41896307e+00 -6.16563499e-01 1.22704697e+00 4.98304188e-01
-6.74278796e-01 -7.22156405e-01 4.30793285e-01 3.70956868e-01
2.32540578e-01 2.28617311e-01 6.98505998e-01 -8.73869181e-01
-6.13066405e-02 -1.86947644e-01 -3.21579725e-01 6.96144819e-01
6.09997511e-02 1.99084714e-01 -7.52647579e-01 -1.20498002e-01
-2.01859668e-01 -2.93874472e-01 1.29220569e+00 6.57168984e-01
1.31769454e+00 -9.81658101e-01 4.35870737e-02 3.60763699e-01
1.00573754e+00 -2.32087895e-01 8.99528205e-01 1.95988163e-01
5.64420700e-01 5.67306101e-01 4.54999655e-01 4.63131368e-01
3.19621354e-01 5.33447266e-02 -3.90970856e-02 4.13671523e-01
-6.95764422e-01 -5.09588838e-01 5.22880733e-01 1.77346480e+00
-1.62757084e-01 -7.37916529e-01 -4.93463248e-01 6.32405996e-01
-1.91835380e+00 -1.21424663e+00 -3.11696053e-01 1.77804852e+00
1.17125261e+00 3.71576840e-04 -1.67761594e-01 3.70347351e-02
5.15240908e-01 5.06341577e-01 -2.08263084e-01 -9.54286635e-01
-2.81171441e-01 2.52285033e-01 2.07763627e-01 4.32325482e-01
-7.20785916e-01 7.74214923e-01 5.83100557e+00 9.27281678e-01
-6.22804523e-01 -7.57142380e-02 5.62472761e-01 -2.15997621e-01
-6.21774435e-01 -2.61077791e-01 -7.32357085e-01 4.57888097e-01
1.48852170e+00 -1.04795611e+00 2.53015868e-02 2.10058585e-01
3.62238705e-01 -4.26401123e-02 -1.21966040e+00 5.96011519e-01
5.40998340e-01 -1.81021929e+00 6.13697708e-01 -1.90794587e-01
1.11299527e+00 -1.60032526e-01 -3.43415082e-01 3.87260675e-01
2.34388381e-01 -1.04874778e+00 7.88410187e-01 9.37126994e-01
6.98810697e-01 -8.94615531e-01 9.98356581e-01 4.85060334e-01
-4.45857286e-01 2.94079334e-01 -7.13580310e-01 -2.29269713e-02
1.90608934e-01 1.00560594e+00 -7.39328623e-01 1.15851808e+00
2.08409771e-01 1.27749074e+00 -5.94620287e-01 9.53789592e-01
-5.78971505e-01 9.33248937e-01 3.52650762e-01 -5.93179941e-01
1.77193224e-01 -1.68606609e-01 9.98272181e-01 1.49568963e+00
6.63059473e-01 -1.62084833e-01 -5.58439493e-01 1.07249677e+00
-7.92277753e-01 2.84571052e-01 -4.81062382e-01 -5.93738019e-01
3.76215994e-01 1.19921517e+00 -2.73509353e-01 -7.64970481e-01
2.22456455e-01 1.09213686e+00 4.42206502e-01 1.93802133e-01
-5.15090823e-01 -7.85980344e-01 -6.32368922e-02 -4.10828382e-01
1.95278868e-01 8.86832848e-02 -7.51723111e-01 -1.31875432e+00
2.37003937e-01 -7.21703947e-01 1.82605125e-02 -9.03883398e-01
-9.90306318e-01 6.01240754e-01 -1.52892753e-01 -9.10546660e-01
-1.62716821e-01 3.19360308e-02 -1.07516587e+00 8.26490521e-01
-1.33204031e+00 -1.13484168e+00 -3.45397592e-02 -4.94573206e-01
1.06628287e+00 -1.33661523e-01 5.88440597e-01 -3.10848057e-01
-6.70328021e-01 2.95586020e-01 4.22729552e-01 -1.25397205e-01
6.68642819e-01 -1.47132444e+00 7.69133508e-01 9.26479936e-01
-1.44856330e-02 5.35277247e-01 9.99000430e-01 -8.38924050e-01
-1.12523198e+00 -1.31479204e+00 1.75118208e+00 -3.55539024e-01
6.57272875e-01 -9.74382088e-02 -6.23259008e-01 4.30858433e-01
9.67503667e-01 -9.32303131e-01 3.50006610e-01 -1.22588195e-01
6.04300201e-02 4.13416214e-02 -7.27626681e-01 5.70213139e-01
8.82116914e-01 -9.58502144e-02 -1.28881633e+00 4.35634136e-01
1.21548045e+00 -3.15744638e-01 -8.57964098e-01 3.75728548e-01
2.79237002e-01 -5.70919633e-01 6.21160984e-01 -7.16924489e-01
1.71559441e+00 2.18099385e-01 3.13854277e-01 -1.89002299e+00
-3.50939393e-01 -5.21233380e-01 -4.89577293e-01 1.69037831e+00
4.97288078e-01 -1.15319945e-01 3.58248025e-01 -6.14461675e-02
-5.71798027e-01 -7.69487739e-01 -7.37784147e-01 -5.67560196e-01
6.08353436e-01 2.35968247e-01 6.04270101e-01 1.73281386e-01
6.06678545e-01 8.61250043e-01 -2.40844935e-01 -5.83464622e-01
6.69545889e-01 -1.26727894e-02 5.42777240e-01 -1.18262148e+00
1.34123996e-01 -8.07803690e-01 8.76243412e-02 -1.05644035e+00
4.40664977e-01 -1.13993216e+00 1.28657028e-01 -2.58697772e+00
8.00878942e-01 6.75491214e-01 -2.54235268e-02 -1.14074107e-02
-5.31213939e-01 -2.55133510e-01 2.82100707e-01 5.99634610e-02
-1.03908956e+00 8.27445328e-01 1.52867377e+00 -5.87489724e-01
-3.89746055e-02 -1.74742136e-02 -1.18544042e+00 2.01599766e-02
6.10961914e-01 -1.09593436e-01 -1.74810290e-01 -4.13736790e-01
4.15764004e-01 3.54308188e-01 -4.05051671e-02 -1.01310623e+00
2.52536505e-01 1.64429113e-01 2.75770813e-01 -9.15768087e-01
-2.43056551e-01 1.47325039e-01 -3.33559662e-01 5.14607370e-01
-1.28930378e+00 -1.28501784e-02 2.64253110e-01 3.08007330e-01
-2.77600080e-01 -6.40122712e-01 2.98254102e-01 5.22189215e-02
1.34896412e-01 -3.79745476e-02 -4.50300425e-01 3.75737578e-01
6.62135124e-01 2.59173572e-01 -7.29157031e-01 -4.99751210e-01
-2.56090581e-01 3.53976697e-01 3.23844980e-03 1.55611858e-01
6.36650443e-01 -1.01231897e+00 -1.70357144e+00 -4.14795160e-01
-2.10228950e-01 1.11789502e-01 2.91794449e-01 6.39626741e-01
-7.97758698e-01 9.84500647e-01 -2.04551369e-01 5.77089284e-03
-1.13173652e+00 2.13741019e-01 -1.90538883e-01 -8.28104019e-01
-8.64041388e-01 5.95075548e-01 -1.36558563e-01 -1.15366347e-01
1.31252825e-01 -3.90685946e-01 -6.62037969e-01 2.27430671e-01
9.92768586e-01 9.25413072e-01 2.99804598e-01 -9.37650576e-02
4.25589755e-02 -2.40547420e-03 -1.72717303e-01 -2.18826622e-01
1.80363178e+00 -1.78447589e-01 -4.85207558e-01 5.39619982e-01
1.16657817e+00 1.25501320e-01 -8.83049488e-01 1.55108586e-01
1.62739947e-01 2.83639044e-01 2.14700624e-01 -9.74010110e-01
-4.85665679e-01 9.66520727e-01 -4.30918992e-01 2.45512083e-01
7.73808062e-01 8.19667894e-03 1.09985781e+00 4.96168345e-01
-2.83449560e-01 -1.12638748e+00 -8.08472335e-02 7.51752377e-01
1.55114126e+00 -7.84244776e-01 4.74008381e-01 -7.29448795e-02
-6.46363378e-01 1.40483427e+00 8.67862850e-02 -3.07478070e-01
-1.86515808e-01 -1.35152694e-02 -6.49133384e-01 -3.36062014e-02
-1.11949313e+00 1.89959601e-01 7.15935230e-01 3.59934032e-01
7.07462192e-01 2.66825080e-01 -1.06818771e+00 9.11740422e-01
-7.61402726e-01 -3.15166102e-03 1.10806191e+00 4.38125521e-01
-6.94853306e-01 -6.46302700e-01 8.87909234e-02 9.34841037e-01
-4.85093355e-01 -2.92046994e-01 -8.37514997e-01 5.91891147e-02
-6.25723302e-01 1.11452401e+00 1.01196125e-01 -1.11984015e-01
3.96725327e-01 5.87855093e-02 6.02283001e-01 -8.53102863e-01
-7.51871586e-01 -2.41937593e-01 4.84616637e-01 -1.63649693e-01
-2.14183241e-01 -8.02808702e-01 -9.91324604e-01 -3.81469488e-01
-8.94829538e-03 6.10362947e-01 7.09582090e-01 8.40903282e-01
5.22036135e-01 1.22909248e+00 5.34339547e-01 -7.72820890e-01
-7.67751217e-01 -1.67656147e+00 -4.44494545e-01 2.54729599e-01
5.00668526e-01 1.51196674e-01 -3.54899287e-01 2.05271080e-01] | [12.549968719482422, 9.576931953430176] |
d1478abc-8886-48d8-a1b0-4bbbd5a09152 | neuralreg-an-end-to-end-approach-to-referring | 1805.08093 | null | http://arxiv.org/abs/1805.08093v1 | http://arxiv.org/pdf/1805.08093v1.pdf | NeuralREG: An end-to-end approach to referring expression generation | Traditionally, Referring Expression Generation (REG) models first decide on
the form and then on the content of references to discourse entities in text,
typically relying on features such as salience and grammatical function. In
this paper, we present a new approach (NeuralREG), relying on deep neural
networks, which makes decisions about form and content in one go without
explicit feature extraction. Using a delexicalized version of the WebNLG
corpus, we show that the neural model substantially improves over two strong
baselines. Data and models are publicly available. | ['Sander Wubben', 'Ákos Kádár', 'Thiago Castro Ferreira', 'Emiel Krahmer', 'Diego Moussallem'] | 2018-05-21 | neuralreg-an-end-to-end-approach-to-referring-1 | https://aclanthology.org/P18-1182 | https://aclanthology.org/P18-1182.pdf | acl-2018-7 | ['referring-expression-generation'] | ['computer-vision'] | [ 2.93631524e-01 8.34944010e-01 -3.37703496e-01 -5.36839008e-01
-9.40438569e-01 -7.91456044e-01 1.22668874e+00 2.43815169e-01
-5.71343839e-01 9.20162380e-01 8.19644809e-01 -3.40647936e-01
3.86236459e-01 -9.77695286e-01 -6.93714738e-01 -1.75548673e-01
2.97819316e-01 4.94874984e-01 -3.08289519e-03 -6.55049384e-01
4.44271028e-01 2.21429780e-01 -1.15543127e+00 5.39647758e-01
5.97027659e-01 8.96987975e-01 -2.09975645e-01 5.63694954e-01
-4.03614640e-01 1.62160313e+00 -8.77302289e-01 -7.28774309e-01
-3.08730841e-01 -5.62920094e-01 -1.35029161e+00 -3.31566244e-01
4.53350902e-01 -4.77592289e-01 -4.54445064e-01 9.79103923e-01
4.06264067e-01 4.77862835e-01 7.76081145e-01 -7.26340950e-01
-1.06197917e+00 1.20662260e+00 -6.45931885e-02 4.55967784e-01
5.47677457e-01 -2.37420291e-01 1.49015236e+00 -1.00121033e+00
1.10712612e+00 1.49128735e+00 3.85417372e-01 6.40857756e-01
-1.05305803e+00 -7.31245354e-02 3.36312026e-01 3.36350739e-01
-1.12715125e+00 -7.60322690e-01 9.22196269e-01 -2.56700784e-01
1.26278007e+00 1.69759631e-01 2.51900345e-01 1.23548913e+00
-1.07650876e-01 1.01082003e+00 7.78133333e-01 -6.75509572e-01
1.86135069e-01 -5.26233673e-01 3.08943421e-01 4.66738880e-01
-1.93334907e-01 -1.21761849e-02 -4.48311329e-01 -1.24789394e-01
3.53886455e-01 -7.55295992e-01 -5.26149929e-01 1.19595267e-01
-1.14116347e+00 1.01612496e+00 5.86124957e-01 6.09986901e-01
-5.09349525e-01 3.28880101e-01 4.82359231e-01 1.21251836e-01
4.70395535e-01 6.40527964e-01 -1.16785549e-01 -3.86232942e-01
-6.41825438e-01 7.61603534e-01 8.28857541e-01 9.66567039e-01
4.61689681e-01 -5.01984581e-02 -5.81610322e-01 5.78493655e-01
2.29087070e-01 2.86798179e-01 4.20993805e-01 -1.31946743e+00
6.85460746e-01 6.39817238e-01 2.45203033e-01 -1.13697588e+00
-4.11493331e-01 1.03254199e-01 -2.72397935e-01 -2.86735982e-01
3.49145174e-01 -4.41827267e-01 -5.99200010e-01 2.15930557e+00
1.57016292e-01 -4.30323869e-01 3.12161326e-01 8.64293754e-01
1.38984203e+00 7.47170210e-01 3.56214076e-01 -7.84170851e-02
1.37025034e+00 -7.94278502e-01 -1.20837522e+00 -1.61950290e-01
8.44223559e-01 -4.25525367e-01 9.30478752e-01 3.39149944e-02
-1.55390155e+00 -8.39380100e-02 -8.94434690e-01 -8.40230465e-01
-5.22252142e-01 1.15464963e-01 6.87560141e-01 9.06135589e-02
-1.28379846e+00 6.53902233e-01 -6.83858514e-01 -1.97363973e-01
3.37701559e-01 2.24821270e-01 -3.20590287e-01 2.38360912e-01
-1.59113967e+00 1.26518810e+00 4.99419272e-01 2.97114551e-01
-3.36649805e-01 -3.05815399e-01 -1.20578027e+00 3.66448388e-02
3.08648676e-01 -7.23730743e-01 1.77124429e+00 -1.16471362e+00
-1.72991109e+00 1.29787004e+00 -3.15404803e-01 -7.61491120e-01
4.10103321e-01 -4.79486048e-01 -2.24525109e-01 2.17360377e-01
3.08686048e-02 7.30327249e-01 4.26148921e-01 -9.59920347e-01
-4.46439445e-01 -2.66222686e-01 7.76395738e-01 2.25102931e-01
2.52962381e-01 4.75973129e-01 -2.70727817e-02 -6.25281453e-01
-8.36782977e-02 -5.68459094e-01 1.72426328e-01 -2.13229686e-01
-7.07928717e-01 -8.96133900e-01 5.08602679e-01 -7.58175790e-01
1.38530612e+00 -1.84532070e+00 4.77838606e-01 -2.93429494e-01
2.70982534e-01 6.44399449e-02 -1.66347608e-01 3.38981837e-01
-1.31164372e-01 4.07139093e-01 -3.61649543e-01 -1.42055556e-01
3.08989167e-01 2.07206830e-01 -4.20234740e-01 3.15891892e-01
6.44762218e-01 1.29886174e+00 -9.39444482e-01 -3.37396801e-01
-1.32322505e-01 3.68596792e-01 -5.17549872e-01 3.27874005e-01
-5.27230322e-01 6.28685532e-03 -4.29030389e-01 3.68940264e-01
1.68638751e-01 -3.02227914e-01 2.68899232e-01 -2.84213036e-01
-1.75081074e-01 1.14299953e+00 -5.05065739e-01 1.65349233e+00
-5.72054505e-01 9.11627054e-01 1.26187891e-01 -9.92290854e-01
8.02639604e-01 5.44519663e-01 -1.94765478e-01 -6.84393704e-01
6.41741097e-01 1.22017689e-01 8.73802006e-02 -4.06144857e-01
9.47765827e-01 -7.15563148e-02 -5.53450942e-01 6.05773926e-01
8.74305144e-02 -1.48459122e-01 5.01022220e-01 3.79058361e-01
1.13465583e+00 5.88704407e-01 6.43688560e-01 -2.16337860e-01
6.54618263e-01 4.53161485e-02 4.04340357e-01 5.60198307e-01
-1.39083639e-01 5.66780984e-01 1.02535081e+00 -3.15482646e-01
-8.18587005e-01 -7.88429677e-01 -1.02548435e-01 1.23820055e+00
-1.09123208e-01 -5.07705808e-01 -9.33722317e-01 -7.33719707e-01
-2.88437337e-01 1.44304693e+00 -1.05707705e+00 -1.25415120e-02
-1.35898566e+00 -3.07329625e-01 6.51677370e-01 6.83161139e-01
3.11347336e-01 -1.57237732e+00 -6.90824270e-01 2.62283474e-01
-4.17601258e-01 -1.34069395e+00 -2.09523290e-01 9.72949862e-02
-3.86987984e-01 -1.09939790e+00 -4.76697206e-01 -5.01691222e-01
4.59461689e-01 -3.80616158e-01 1.61517167e+00 2.47546017e-01
3.28279525e-01 -1.15861651e-02 -5.43603241e-01 -3.77166718e-01
-5.08619428e-01 4.65717465e-01 -5.09404182e-01 -4.14934129e-01
4.50809687e-01 -4.99341041e-01 -1.01396300e-01 -4.51692194e-01
-6.37190938e-01 1.06946059e-01 8.60449076e-02 8.72819424e-01
4.18995500e-01 -9.33146119e-01 6.19335592e-01 -1.20049548e+00
9.20054257e-01 -5.40308952e-01 -4.58597958e-01 2.79539108e-01
-1.38112023e-01 2.96637237e-01 5.28315067e-01 9.25078765e-02
-1.23335826e+00 -3.97736222e-01 -3.37670177e-01 -9.28044133e-03
-6.93171471e-02 6.88524723e-01 -2.71700293e-01 5.72818398e-01
6.83887124e-01 -2.25740194e-01 -4.89646047e-01 -3.12131435e-01
8.49897921e-01 4.58384842e-01 8.82632017e-01 -9.77811635e-01
3.45060796e-01 1.41531497e-01 -1.11229636e-01 -3.98337066e-01
-1.19266093e+00 2.17408746e-01 -7.24020898e-01 1.05676197e-01
9.17741835e-01 -7.82978415e-01 -5.36566079e-01 1.10460155e-01
-1.87166846e+00 -5.52041650e-01 -4.68555659e-01 1.77786067e-01
-7.78364956e-01 8.04907307e-02 -8.08767378e-01 -6.39044106e-01
-4.00367260e-01 -5.41721165e-01 1.13448751e+00 2.28540495e-01
-6.60941541e-01 -1.15596962e+00 -1.39125111e-02 -9.53157805e-03
4.80647147e-01 6.15234375e-01 1.00140750e+00 -1.06915200e+00
-3.73727173e-01 -5.19409180e-02 -4.55787271e-01 -8.86283368e-02
8.78318250e-02 1.94474652e-01 -9.32674229e-01 2.73783416e-01
-1.49630159e-01 -6.23745143e-01 8.28545630e-01 1.88752994e-01
9.09923613e-01 -6.40393555e-01 -1.41872332e-01 4.19945180e-01
9.87824023e-01 7.57032484e-02 7.40768850e-01 4.65387374e-01
5.95443785e-01 8.25355113e-01 4.74646032e-01 2.93053865e-01
8.50821435e-01 6.59579456e-01 4.53655243e-01 2.54639059e-01
-3.72529954e-01 -1.82045594e-01 3.59139889e-01 5.23947120e-01
-3.00755531e-01 -6.97274089e-01 -9.50517952e-01 7.99915433e-01
-1.86855304e+00 -1.08770287e+00 3.41904610e-02 1.39980161e+00
1.24198771e+00 -6.58452418e-03 -1.22314379e-01 -2.31785133e-01
7.66264617e-01 5.83794773e-01 -3.00145507e-01 -7.18421340e-01
-5.54108500e-01 4.56125677e-01 -1.52313665e-01 7.27072954e-01
-1.28901434e+00 1.33418190e+00 7.24799538e+00 2.67399728e-01
-1.00969970e+00 2.61792750e-03 5.69868803e-01 -4.18409184e-02
-4.94085073e-01 -8.21769908e-02 -7.31650770e-01 1.33920936e-02
1.16242695e+00 -4.33071911e-01 3.57973099e-01 7.28632689e-01
2.18860973e-02 1.29270740e-03 -1.39162171e+00 5.49034655e-01
3.30922663e-01 -1.51513302e+00 1.56756505e-01 -3.18149835e-01
4.31759685e-01 4.75104675e-02 -2.26471290e-01 4.14525747e-01
7.68405378e-01 -1.19186878e+00 9.95532990e-01 5.21852434e-01
5.81138432e-01 -6.44339085e-01 9.58204329e-01 -4.21160012e-02
-6.56833589e-01 3.00031632e-01 -5.31020425e-02 -2.72978634e-01
4.77122575e-01 2.20672209e-02 -5.99591017e-01 4.86293167e-01
2.47936010e-01 6.98021650e-01 -3.74018729e-01 2.21656978e-01
-9.62969005e-01 5.00498533e-01 -2.39879057e-01 -2.46939778e-01
3.64925593e-01 2.83023983e-01 6.76983058e-01 1.38007843e+00
1.30181849e-01 5.44900954e-01 -8.86127129e-02 1.47944808e+00
-4.93956417e-01 1.42379686e-01 -6.27207816e-01 -2.57614285e-01
6.58497334e-01 1.19883811e+00 -2.55500793e-01 -4.57603246e-01
-2.89517343e-01 7.33237684e-01 7.51345992e-01 3.79567027e-01
-7.65962064e-01 -5.52900255e-01 3.26990724e-01 -2.96714436e-02
4.90833104e-01 -5.57897948e-02 -1.12037607e-01 -1.13195789e+00
9.70285833e-02 -6.27084255e-01 3.44718158e-01 -9.48262453e-01
-1.34269357e+00 8.91488791e-01 1.17921293e-01 -6.60683155e-01
-1.01321661e+00 -7.37578511e-01 -8.00272226e-01 9.47550058e-01
-1.67352891e+00 -1.22611606e+00 -2.02917438e-02 3.09370071e-01
3.16622585e-01 9.53542516e-02 1.00054550e+00 -1.02117516e-01
-5.05004585e-01 5.31916380e-01 -4.89660501e-01 7.24960446e-01
4.81265813e-01 -1.41840041e+00 7.23984480e-01 7.17434466e-01
1.97977379e-01 8.35289776e-01 7.49976695e-01 -3.16733062e-01
-1.02640247e+00 -9.16058362e-01 1.54071665e+00 -5.78087032e-01
8.52387786e-01 -2.69596763e-02 -9.58239913e-01 1.00951207e+00
7.48019278e-01 8.67832378e-02 4.66912925e-01 3.48799765e-01
-4.10274863e-01 5.79326034e-01 -1.07273006e+00 5.44664621e-01
1.04565001e+00 -7.22018301e-01 -1.28362453e+00 1.42993227e-01
8.42233300e-01 -8.86916280e-01 -7.34366357e-01 3.44419479e-01
5.30130751e-02 -7.83572257e-01 4.94825155e-01 -1.14095318e+00
8.25308383e-01 -1.37374371e-01 -5.43476224e-01 -1.15024638e+00
-9.28654075e-02 -7.97225118e-01 -5.95140338e-01 1.32678807e+00
5.32003939e-01 -2.64685482e-01 3.85025024e-01 9.97437656e-01
-3.50446284e-01 -6.83050036e-01 -1.13258386e+00 -2.80665725e-01
5.94892204e-01 -2.26746991e-01 7.21331835e-01 9.83806968e-01
4.26600128e-01 8.57490003e-01 1.62348032e-01 -4.36159432e-01
-2.80618258e-02 2.51990288e-01 4.09025103e-01 -1.14880633e+00
2.41413508e-02 -6.29215598e-01 -1.56720191e-01 -1.04110599e+00
8.93602610e-01 -9.80709851e-01 4.03933711e-02 -1.66682494e+00
-9.51857194e-02 5.11912256e-02 6.00385899e-03 6.73010111e-01
-2.75104612e-01 -2.48330943e-02 2.48131841e-01 -9.82487798e-02
-7.90007174e-01 7.44481027e-01 1.01081574e+00 -1.24952435e-01
-8.00147131e-02 -5.94298363e-01 -8.15309167e-01 9.20378327e-01
6.80209517e-01 -1.43110245e-01 2.18563136e-02 -7.24545538e-01
5.18420041e-01 1.31002456e-01 4.25991952e-01 -4.15299267e-01
-4.16251346e-02 -1.32205263e-01 3.71290371e-02 -5.50088763e-01
3.54086578e-01 -1.28882200e-01 -5.25568187e-01 -2.13695183e-01
-9.76053596e-01 5.24483740e-01 2.49197572e-01 1.12427466e-01
-4.29571658e-01 -4.50372934e-01 2.80971587e-01 -1.71740785e-01
-5.51152945e-01 -2.39392787e-01 -3.68222982e-01 5.89007437e-01
5.21315575e-01 2.36252770e-01 -8.11050296e-01 -5.94762504e-01
-8.20746243e-01 1.37228705e-02 3.84114176e-01 3.97180647e-01
4.03244734e-01 -1.46582282e+00 -9.26939905e-01 -4.98087764e-01
-4.09030989e-02 2.51901478e-01 -1.74208105e-01 6.55627191e-01
-3.41904402e-01 5.79972208e-01 5.65476418e-02 -1.14272535e-01
-8.25308084e-01 5.13486266e-01 5.01253307e-01 -2.97983974e-01
-5.10416806e-01 8.46316576e-01 1.22994199e-01 -3.89246792e-01
-1.16602726e-01 -4.49519515e-01 -7.23191023e-01 1.89027920e-01
7.58430779e-01 -2.86056250e-02 1.85237937e-02 -1.23506391e+00
-2.81737357e-01 1.66009262e-01 -2.80009449e-01 -2.92621613e-01
1.27604759e+00 -1.38642907e-01 -4.31681156e-01 6.37208402e-01
1.04859269e+00 1.19259603e-01 -8.93831849e-01 -3.92552108e-01
4.29171741e-01 -7.05670267e-02 2.17139557e-01 -7.72921979e-01
-8.70859325e-01 8.55058968e-01 -1.51574284e-01 2.34433562e-01
8.10072064e-01 2.20710650e-01 7.24080265e-01 8.22689354e-01
1.10013358e-01 -1.09500468e+00 -2.52462387e-01 1.04772353e+00
1.26094222e+00 -8.76285195e-01 -1.56170130e-01 -3.01033497e-01
-5.56555867e-01 1.30564272e+00 7.82430589e-01 -3.89411032e-01
2.82004952e-01 2.60019779e-01 1.13556549e-01 -4.66445029e-01
-1.01807499e+00 -2.21783042e-01 1.76176250e-01 5.74689090e-01
1.23134899e+00 -7.00076595e-02 -5.69715798e-01 9.68341231e-01
-6.85027659e-01 -1.53235421e-01 6.56996310e-01 8.10266554e-01
-2.14763463e-01 -1.10931420e+00 -1.73663363e-01 1.95302099e-01
-7.42569506e-01 -4.56042588e-01 -8.30792427e-01 8.85861218e-01
-1.84324354e-01 8.35731506e-01 4.66468602e-01 2.02269237e-02
5.32133996e-01 1.63978577e-01 5.98964334e-01 -6.80727601e-01
-7.60829866e-01 -3.40436965e-01 1.00425649e+00 -5.63538074e-01
-8.65913689e-01 -8.53145957e-01 -1.66512728e+00 -3.07919800e-01
-2.93589264e-01 8.66464302e-02 2.25987837e-01 1.14540339e+00
3.62417608e-01 6.78721130e-01 1.81039572e-01 -8.46145093e-01
-4.63975102e-01 -1.10606563e+00 -1.07249506e-01 5.68302631e-01
2.64838874e-01 -6.81586385e-01 -1.46789312e-01 5.75714633e-02] | [10.846491813659668, 9.199474334716797] |
66ee6379-9b6e-4f2b-a1ea-a2bc73842028 | scene-aware-learning-network-for-radar-object | 2107.01469 | null | https://arxiv.org/abs/2107.01469v1 | https://arxiv.org/pdf/2107.01469v1.pdf | Scene-aware Learning Network for Radar Object Detection | Object detection is essential to safe autonomous or assisted driving. Previous works usually utilize RGB images or LiDAR point clouds to identify and localize multiple objects in self-driving. However, cameras tend to fail in bad driving conditions, e.g. bad weather or weak lighting, while LiDAR scanners are too expensive to get widely deployed in commercial applications. Radar has been drawing more and more attention due to its robustness and low cost. In this paper, we propose a scene-aware radar learning framework for accurate and robust object detection. First, the learning framework contains branches conditioning on the scene category of the radar sequence; with each branch optimized for a specific type of scene. Second, three different 3D autoencoder-based architectures are proposed for radar object detection and ensemble learning is performed over the different architectures to further boost the final performance. Third, we propose novel scene-aware sequence mix augmentation (SceneMix) and scene-specific post-processing to generate more robust detection results. In the ROD2021 Challenge, we achieved a final result of average precision of 75.0% and an average recall of 81.0%. Moreover, in the parking lot scene, our framework ranks first with an average precision of 97.8% and an average recall of 98.6%, which demonstrates the effectiveness of our framework. | ['Alberto Sangiovanni Vincentelli', 'Kurt Keutzer', 'Xiangyu Yue', 'Zangwei Zheng'] | 2021-07-03 | null | null | null | null | ['robust-object-detection', 'radar-object-detection'] | ['computer-vision', 'robots'] | [ 2.58387387e-01 -6.02050543e-01 1.46113887e-01 -5.07019401e-01
-7.61357188e-01 -3.17022920e-01 7.23155737e-01 -4.31693951e-03
-5.61017692e-01 3.51801187e-01 -3.20518345e-01 -1.81734413e-01
1.57598421e-01 -8.65107119e-01 -6.28334999e-01 -6.65419757e-01
1.50291532e-01 2.37869903e-01 4.14347947e-01 -2.33609155e-01
1.74264267e-01 6.29502237e-01 -1.97755396e+00 -3.47576179e-02
1.02035975e+00 1.19113326e+00 5.99335730e-01 7.35538304e-01
-4.56488468e-02 6.78814769e-01 -6.60826623e-01 -2.74392098e-01
5.23530781e-01 -3.16621922e-02 7.75604323e-02 2.06578940e-01
8.08661997e-01 -5.62602282e-01 -4.38030630e-01 1.08639789e+00
5.88082194e-01 1.74821913e-02 4.87163156e-01 -1.29498088e+00
-1.96218133e-01 1.48821995e-01 -7.67578363e-01 1.41184837e-01
-1.75935477e-01 5.13838768e-01 7.71326303e-01 -8.09986413e-01
5.87579533e-02 1.04231453e+00 3.15215617e-01 3.06387603e-01
-1.06546319e+00 -9.54102933e-01 6.62683025e-02 6.34100318e-01
-1.54181480e+00 -4.22790915e-01 7.48605013e-01 -6.14507258e-01
6.96106076e-01 9.26985219e-02 4.89998221e-01 9.53609347e-01
1.89986259e-01 7.67615616e-01 1.21260560e+00 1.54413655e-02
9.42907259e-02 1.17067941e-01 1.54909849e-01 5.76405585e-01
5.48704922e-01 3.81663114e-01 -3.92455906e-01 4.91360605e-01
2.97689855e-01 1.64188251e-01 2.11521462e-02 -3.28268558e-01
-1.17265606e+00 8.78980875e-01 5.76643884e-01 -3.53996009e-02
-4.45559233e-01 7.33745694e-02 3.22310150e-01 -1.00952402e-01
2.60662287e-01 1.77838489e-01 -1.29229501e-01 1.43384069e-01
-9.23970103e-01 3.17157418e-01 2.15916067e-01 9.14903224e-01
8.07343900e-01 4.22048062e-01 -5.71898110e-02 8.61467063e-01
1.83622152e-01 1.01563263e+00 6.55702278e-02 -7.26156592e-01
3.84619743e-01 6.60620809e-01 1.70063019e-01 -1.01303971e+00
-4.78925407e-01 -8.65718842e-01 -8.75082374e-01 4.59614873e-01
1.44223467e-01 2.61100885e-02 -1.13380098e+00 1.43925834e+00
2.51964957e-01 4.36043620e-01 2.27508351e-01 1.20475698e+00
8.63131881e-01 6.75396323e-01 5.91036938e-02 4.97840047e-02
1.46428788e+00 -8.47975850e-01 -5.40070891e-01 -7.08618164e-01
3.67583007e-01 -7.58424222e-01 6.92474663e-01 4.50057745e-01
-4.41670001e-01 -9.17046607e-01 -1.40315008e+00 2.68113554e-01
-2.74504304e-01 4.97166961e-01 4.74781096e-01 6.73817635e-01
-6.24367356e-01 4.86966595e-02 -5.60060680e-01 -2.67044991e-01
4.53743964e-01 -4.61330987e-04 -1.44432962e-01 -4.37530816e-01
-9.41457391e-01 1.08940160e+00 5.16239703e-01 2.62883425e-01
-1.09615993e+00 -6.69329166e-01 -7.75563955e-01 -1.86774433e-01
4.79178905e-01 -4.93202567e-01 9.85986710e-01 -4.21046168e-01
-1.13523579e+00 7.80717552e-01 6.55986145e-02 -9.60363328e-01
4.51261938e-01 -5.64810514e-01 -8.16679001e-01 1.26919121e-01
1.30475894e-01 8.79087985e-01 1.02479982e+00 -1.30227339e+00
-1.07418895e+00 -3.50439936e-01 -8.66260380e-02 1.43735722e-01
-6.43079281e-02 -2.05929577e-01 -5.49301803e-01 -3.29667926e-01
8.87328684e-02 -9.00318503e-01 -2.20127568e-01 -3.69446009e-01
-2.13532299e-01 -1.75524000e-02 9.55836833e-01 -5.59726715e-01
7.98830628e-01 -2.23949885e+00 -3.22110265e-01 -1.39246941e-01
1.64178893e-01 4.44331229e-01 -1.62980914e-01 -4.87287752e-02
1.08634062e-01 -4.31684554e-01 -4.18856680e-01 -1.30755976e-01
-1.23550463e-02 8.32017958e-02 -3.02865207e-01 4.66958135e-01
5.53523898e-01 6.93616927e-01 -5.65213263e-01 -3.80489558e-01
6.73078060e-01 4.79872763e-01 -2.94705957e-01 1.96598381e-01
-8.14582556e-02 3.95967185e-01 -2.73186564e-01 8.06949973e-01
9.60345984e-01 1.83980122e-01 -2.66009331e-01 -4.09588277e-01
-3.39947641e-01 7.82398880e-02 -1.30876791e+00 1.22492623e+00
-4.74882245e-01 7.93917418e-01 1.02282651e-01 -8.93404663e-01
1.38050413e+00 -3.02830666e-01 3.56808841e-01 -1.09832203e+00
1.21190742e-01 1.46430850e-01 2.16604620e-01 -4.25952077e-01
7.57709384e-01 -1.44906208e-01 -1.13864563e-01 -4.77023609e-02
-2.62722939e-01 -3.00883859e-01 2.71072626e-01 1.28347188e-01
1.07432938e+00 8.83192057e-04 1.34260915e-02 1.49546891e-01
6.56268001e-01 2.26935789e-01 5.73105574e-01 7.63912201e-01
-3.41294229e-01 5.43746948e-01 -2.00593904e-01 -4.10265744e-01
-9.32297766e-01 -9.68146503e-01 -2.55996644e-01 7.77631402e-01
4.06773448e-01 -1.16594195e-01 -4.60406125e-01 -4.00825858e-01
1.17939062e-01 1.00326252e+00 -1.24345534e-01 -3.22949857e-01
-5.57112515e-01 -7.84107625e-01 4.06471670e-01 6.02439106e-01
8.86080503e-01 -8.76072288e-01 -9.71606016e-01 2.11014122e-01
-2.29559883e-01 -1.61735642e+00 5.32982051e-02 -3.57483211e-03
-6.74312592e-01 -9.30548668e-01 -3.65599155e-01 -3.44063908e-01
2.42470935e-01 7.23531127e-01 8.92523468e-01 -2.97911614e-01
-5.45508265e-01 2.97587097e-01 -2.40337178e-01 -6.57934308e-01
-9.92691666e-02 -1.37624264e-01 2.31610551e-01 3.81740510e-01
4.34031308e-01 -2.09861144e-01 -5.74728668e-01 2.87530124e-01
-6.85575604e-01 -6.03480004e-02 1.08204544e+00 5.39167583e-01
3.44446003e-01 1.22044913e-01 3.79538357e-01 -2.74844527e-01
1.04473673e-01 -1.89412862e-01 -8.60263348e-01 -2.30847552e-01
-4.22455221e-01 -2.42200896e-01 3.49519283e-01 -1.22823440e-01
-8.70496690e-01 4.34155375e-01 -1.74508661e-01 -6.37883186e-01
-4.22273338e-01 1.75892457e-01 -2.03022897e-01 2.03365423e-02
6.11476362e-01 4.37829733e-01 5.71040027e-02 -3.11785460e-01
2.40725070e-01 6.78516984e-01 7.74158061e-01 -9.14948955e-02
1.21358132e+00 4.81276155e-01 8.74933973e-03 -1.09882033e+00
-9.18106377e-01 -7.00008214e-01 -4.37959075e-01 -4.95472968e-01
8.95420730e-01 -1.47125220e+00 -7.00422585e-01 5.29649198e-01
-9.27608490e-01 -1.60654172e-01 5.01056984e-02 8.35918307e-01
-4.05810952e-01 2.57545292e-01 -1.65769849e-02 -1.09306002e+00
-3.21048886e-01 -1.16152477e+00 1.22023606e+00 2.03548357e-01
1.81401044e-01 -2.67043203e-01 -2.45766744e-01 6.13647819e-01
4.71070468e-01 8.88603181e-02 3.31493676e-01 -4.24433261e-01
-9.61685598e-01 -4.68602002e-01 -5.96325755e-01 5.75513899e-01
-1.87689185e-01 -9.44051594e-02 -1.11030734e+00 -1.85351342e-01
-2.15748787e-01 -1.21255860e-01 1.23547864e+00 2.71879643e-01
9.61577237e-01 2.28544921e-01 -3.48061115e-01 4.66508895e-01
1.47430694e+00 2.37665430e-01 5.91968298e-01 4.19710189e-01
6.93048596e-01 5.82934022e-01 1.01362228e+00 3.63733798e-01
3.54288667e-01 8.47838402e-01 8.07963192e-01 7.58017972e-02
-2.30003670e-01 4.97251377e-02 6.78580165e-01 3.22402388e-01
1.29410267e-01 1.51945092e-02 -9.78041351e-01 6.78703904e-01
-1.61729538e+00 -1.02907610e+00 -3.44891012e-01 2.24568343e+00
2.42856130e-01 4.91738111e-01 1.80246189e-01 1.96507066e-01
8.16247821e-01 2.64402241e-01 -5.63325346e-01 4.79925573e-02
-2.01840460e-01 3.05681415e-02 7.99232721e-01 3.14465582e-01
-1.44335663e+00 9.20679271e-01 5.09477043e+00 7.96322703e-01
-1.24393785e+00 -4.22995575e-02 1.21773653e-01 -8.65328088e-02
1.74501032e-01 -1.82961658e-01 -1.22385311e+00 5.23661077e-01
9.07587528e-01 9.66950059e-02 -2.81900470e-03 9.88434732e-01
1.69423237e-01 -1.88936293e-01 -5.84294915e-01 1.14670587e+00
2.16813102e-01 -1.12874925e+00 -1.68736592e-01 1.35084596e-02
3.42913985e-01 3.38865429e-01 -1.55979386e-02 6.16448939e-01
3.21519136e-01 -8.09865415e-01 8.20033252e-01 5.08035243e-01
5.18111169e-01 -9.96892691e-01 7.81072676e-01 5.79650283e-01
-1.20534909e+00 -2.59065002e-01 -3.85183930e-01 -7.24355653e-02
3.21782321e-01 9.85998392e-01 -9.98520255e-01 6.08493090e-01
8.66661727e-01 6.96876943e-01 -8.08849454e-01 1.22559154e+00
-2.97014058e-01 4.77216572e-01 -4.28631783e-01 -4.74298485e-02
3.42337191e-01 -2.63632655e-01 7.89650917e-01 1.34556830e+00
4.31814253e-01 1.83145702e-02 5.13067603e-01 6.87545776e-01
1.59885898e-01 -2.26461411e-01 -7.94416070e-01 1.81240022e-01
3.42939436e-01 1.77892768e+00 -3.97989571e-01 -3.02833974e-01
-3.04226071e-01 6.30175591e-01 -7.74506852e-02 1.48329407e-01
-1.21308494e+00 -4.12618935e-01 9.20281649e-01 9.70237479e-02
6.04740560e-01 -5.40938020e-01 -2.25994080e-01 -8.66557002e-01
7.02784676e-03 -6.84971750e-01 6.45725355e-02 -7.97313809e-01
-1.04247284e+00 5.09679914e-01 -2.44447440e-01 -1.47042990e+00
4.39907797e-02 -7.65721858e-01 -4.03474897e-01 6.47827089e-01
-1.80265641e+00 -1.11811328e+00 -6.84630573e-01 3.78569484e-01
6.28276408e-01 -4.06081110e-01 3.70336652e-01 6.16826117e-01
-8.35738659e-01 3.17335844e-01 -1.84967920e-01 1.49913415e-01
7.34957457e-01 -1.01653254e+00 2.66344070e-01 1.20299220e+00
1.05278216e-01 1.84964508e-01 7.13827610e-01 -5.64180493e-01
-1.69829559e+00 -1.64160335e+00 5.72965562e-01 -2.49779835e-01
5.27837396e-01 -3.92009020e-01 -7.58356452e-01 3.93985093e-01
2.21178159e-02 -5.01566082e-02 2.47429356e-01 -1.02738760e-01
-4.53963012e-01 -6.24950767e-01 -9.08461213e-01 5.75115860e-01
8.59080613e-01 -2.61271358e-01 -3.72181654e-01 3.13075602e-01
5.33355236e-01 -4.12760168e-01 -4.75025237e-01 8.28992128e-01
4.20730919e-01 -1.04552853e+00 1.11267710e+00 -9.20885578e-02
2.41508335e-01 -9.21602607e-01 -5.57782471e-01 -9.30587709e-01
-4.59766656e-01 -5.35112843e-02 -1.30785108e-01 1.16230226e+00
2.28575796e-01 -5.02296865e-01 6.81622505e-01 -1.27613932e-01
-3.83018106e-01 -2.81657338e-01 -6.94811583e-01 -8.99364591e-01
-5.04308999e-01 -9.27717209e-01 3.77867520e-01 5.48121095e-01
-8.11204374e-01 7.08266675e-01 -4.95862126e-01 5.85707009e-01
9.10499096e-01 4.99684155e-01 1.17490160e+00 -1.36660218e+00
-8.21186695e-03 -4.34741825e-01 -7.03691065e-01 -1.00054562e+00
4.82091457e-02 -7.99843788e-01 4.18335229e-01 -1.48280799e+00
9.44142267e-02 -3.70795697e-01 -1.77143753e-01 1.75659567e-01
-2.71866888e-01 5.12310982e-01 3.81395549e-01 5.52762412e-02
-7.46365786e-01 6.65855706e-01 7.54872859e-01 -4.14637357e-01
5.21018840e-02 1.51608437e-01 -6.13438547e-01 6.64333344e-01
9.68831480e-01 -1.94193393e-01 1.81038100e-02 -3.84607643e-01
-9.17148590e-02 -2.93746799e-01 7.27871656e-01 -1.49083817e+00
2.64738679e-01 3.74470791e-03 5.44116020e-01 -1.34411573e+00
5.83488822e-01 -8.76537144e-01 -5.28696142e-02 6.50964200e-01
1.94087520e-01 -2.72875607e-01 4.05171722e-01 7.24173069e-01
-2.23840147e-01 1.36425709e-02 9.78197813e-01 1.58060044e-01
-1.34178102e+00 2.19756275e-01 -3.70682508e-01 -3.30375671e-01
1.20233953e+00 -3.21309298e-01 -1.54849902e-01 -2.57026434e-01
-2.42099956e-01 4.16540265e-01 1.09175988e-01 7.54354239e-01
7.85861373e-01 -1.24851871e+00 -1.09620357e+00 3.58195990e-01
4.72897470e-01 -1.40211545e-03 4.12613004e-01 8.53812754e-01
-1.91714153e-01 4.51010138e-01 -2.74628371e-01 -1.18089056e+00
-1.18203044e+00 5.12287021e-01 8.65097940e-02 1.09628960e-01
-6.43404365e-01 5.62810540e-01 1.93080008e-01 -4.00918335e-01
8.04973915e-02 1.32754957e-02 -2.89727747e-01 1.68562263e-01
6.10323787e-01 4.26056713e-01 2.78243572e-01 -8.33635330e-01
-5.37149131e-01 7.21862614e-01 -8.85781646e-02 6.17059395e-02
1.21754885e+00 5.03479764e-02 3.14611822e-01 2.68881023e-01
8.26707721e-01 -8.41865316e-02 -1.34738505e+00 -2.89165705e-01
-1.25044569e-01 -3.77764523e-01 3.48496348e-01 -7.55441070e-01
-1.13780260e+00 1.04339945e+00 9.28964496e-01 2.29548775e-02
1.14718926e+00 -2.14362383e-01 6.99633002e-01 4.45425332e-01
2.50734597e-01 -1.00244796e+00 -5.61332889e-02 7.09316075e-01
7.18679011e-01 -1.51634204e+00 8.47485587e-02 -2.79483527e-01
-7.91693687e-01 8.13951850e-01 6.32187665e-01 -5.32467067e-02
1.20698772e-01 4.08572018e-01 9.17942673e-02 -1.01605833e-01
-5.55792034e-01 -9.14312363e-01 3.69957477e-01 5.80898285e-01
-6.36829715e-03 1.45919874e-01 8.45615342e-02 3.38891000e-01
-1.99243098e-01 -2.35286370e-01 1.77990913e-01 6.62800431e-01
-9.17621493e-01 -6.21433079e-01 -5.27005255e-01 4.89244401e-01
-7.03451857e-02 1.39720244e-02 -1.63760200e-01 7.45304585e-01
2.70889670e-01 1.09885585e+00 2.39038870e-01 -7.14917183e-01
6.15698338e-01 7.71745667e-02 3.36309642e-01 -3.74293298e-01
-1.19374052e-01 9.68309417e-02 -3.98148596e-03 -5.37029207e-01
-4.56956506e-01 -7.07232475e-01 -1.05617714e+00 -2.16975689e-01
-3.59085649e-01 -2.98832744e-01 9.24372375e-01 8.16942811e-01
4.77315009e-01 7.23842323e-01 7.07592368e-01 -7.52202630e-01
-4.37650681e-01 -9.05170619e-01 -4.80541080e-01 6.40186816e-02
3.49130601e-01 -8.10226202e-01 -1.35928854e-01 -3.81780118e-02] | [7.944762706756592, -1.3919123411178589] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.