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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
84520eeb-8eae-409a-a5b3-ddbdcf4f0cf4 | os-msl-one-stage-multimodal-sequential-link | 2207.01241 | null | https://arxiv.org/abs/2207.01241v1 | https://arxiv.org/pdf/2207.01241v1.pdf | OS-MSL: One Stage Multimodal Sequential Link Framework for Scene Segmentation and Classification | Scene segmentation and classification (SSC) serve as a critical step towards the field of video structuring analysis. Intuitively, jointly learning of these two tasks can promote each other by sharing common information. However, scene segmentation concerns more on the local difference between adjacent shots while classification needs the global representation of scene segments, which probably leads to the model dominated by one of the two tasks in the training phase. In this paper, from an alternate perspective to overcome the above challenges, we unite these two tasks into one task by a new form of predicting shots link: a link connects two adjacent shots, indicating that they belong to the same scene or category. To the end, we propose a general One Stage Multimodal Sequential Link Framework (OS-MSL) to both distinguish and leverage the two-fold semantics by reforming the two learning tasks into a unified one. Furthermore, we tailor a specific module called DiffCorrNet to explicitly extract the information of differences and correlations among shots. Extensive experiments on a brand-new large scale dataset collected from real-world applications, and MovieScenes are conducted. Both the results demonstrate the effectiveness of our proposed method against strong baselines. | ['Bo Ren', 'Deqiang Jiang', 'Xinghua Jiang', 'Zhuoxuan Jiang', 'Di Yin', 'Lingfeng Qiao', 'Ye Liu'] | 2022-07-04 | null | null | null | null | ['scene-segmentation'] | ['computer-vision'] | [ 3.31669152e-01 -9.93738472e-02 -5.05795598e-01 -4.74982142e-01
-5.97649992e-01 -3.56963813e-01 6.75331235e-01 2.37992063e-01
-2.55024731e-01 4.27341491e-01 5.35791993e-01 3.03238630e-02
-8.76217522e-03 -6.11330390e-01 -5.07491171e-01 -6.57932460e-01
1.54598266e-01 5.92488758e-02 8.11179996e-01 -7.92349577e-02
1.43134341e-01 -1.38167560e-01 -1.23167646e+00 4.59090292e-01
6.66329920e-01 1.11818945e+00 2.94019520e-01 2.92492360e-01
-3.69369507e-01 1.08659112e+00 3.67161185e-02 -4.97610867e-01
1.75776482e-01 -5.50423026e-01 -8.54802847e-01 4.97831672e-01
3.82213771e-01 -1.75941080e-01 -2.96442032e-01 1.20248365e+00
1.79561198e-01 3.16795945e-01 4.44578707e-01 -1.35128522e+00
-1.52815983e-01 8.09994936e-01 -1.01501167e+00 3.32613558e-01
3.30040157e-01 6.42794278e-03 1.28085816e+00 -7.33030021e-01
6.20041430e-01 1.18898618e+00 4.31638509e-01 1.12823136e-01
-1.10917318e+00 -4.58058178e-01 5.66941202e-01 3.78977507e-01
-1.16235614e+00 -4.46318597e-01 1.08924258e+00 -6.99427545e-01
2.02965409e-01 7.85146803e-02 5.71493030e-01 9.37556088e-01
-1.76910087e-01 1.24371755e+00 8.71947169e-01 1.45123191e-02
5.28586246e-02 6.43354505e-02 4.38319296e-01 6.11682355e-01
-5.00471778e-02 -3.95964354e-01 -6.19441450e-01 1.24405518e-01
5.08657336e-01 2.35180438e-01 -4.71895337e-01 -7.26988792e-01
-1.32891238e+00 6.36709869e-01 4.12517577e-01 3.23610157e-01
-1.78515285e-01 -2.10857131e-02 6.08667672e-01 2.19810501e-01
4.54663366e-01 -1.52301947e-02 -2.16624901e-01 -9.11859646e-02
-8.98807943e-01 -4.71562162e-05 5.75845003e-01 9.24578249e-01
7.72429764e-01 -3.10277969e-01 -4.43241447e-01 7.25272775e-01
5.73237598e-01 -1.59486860e-01 4.24892157e-01 -5.81924438e-01
7.25844741e-01 6.15072250e-01 -3.49693149e-02 -1.18267643e+00
-1.76192284e-01 -5.24291277e-01 -6.16912723e-01 -2.82367766e-01
2.30366737e-01 -6.08026981e-02 -8.10780942e-01 1.78349710e+00
6.08105540e-01 8.40530634e-01 -1.40489265e-01 1.09336782e+00
9.22013700e-01 7.92036831e-01 2.86540121e-01 -3.27525765e-01
1.39893448e+00 -1.54532599e+00 -7.42067754e-01 -2.31081173e-01
4.47860390e-01 -7.52660751e-01 9.48945940e-01 2.53492653e-01
-1.05328870e+00 -8.32063317e-01 -1.20804667e+00 -1.03891455e-01
-2.03255743e-01 -1.26833349e-01 5.84300041e-01 9.07292068e-02
-5.56925952e-01 5.51446080e-01 -7.20758677e-01 -4.86132026e-01
3.68055999e-01 -8.91985223e-02 -1.19227581e-01 -7.96461180e-02
-1.13947666e+00 3.87458652e-01 6.75906003e-01 9.29386765e-02
-7.39304781e-01 -4.23421234e-01 -8.27409625e-01 -2.77460292e-02
9.02500808e-01 -4.67883378e-01 1.09965563e+00 -9.47999239e-01
-1.10331833e+00 7.99955368e-01 -1.08088143e-01 -2.53506482e-01
5.91253042e-01 -4.34178710e-01 -4.59574014e-01 2.37322785e-02
3.75568777e-01 5.68394065e-01 7.18728602e-01 -1.44379961e+00
-1.01244652e+00 -2.42092833e-01 1.25569254e-01 3.66149157e-01
-3.40563953e-01 4.81557213e-02 -1.14795732e+00 -6.29296303e-01
1.46198526e-01 -4.74562079e-01 -4.96561490e-02 -1.56634331e-01
-6.22464478e-01 -2.12759167e-01 9.03966248e-01 -4.84914899e-01
1.49636829e+00 -2.28994203e+00 4.41747159e-01 1.43162739e-02
2.80576706e-01 1.87048480e-01 -1.79822549e-01 4.08899516e-01
-1.50582781e-02 -1.54431760e-01 -2.83011496e-01 -5.71758389e-01
-9.48616415e-02 1.51215822e-01 -3.76537055e-01 3.80476862e-01
5.94086237e-02 8.51153970e-01 -1.17857385e+00 -8.53639841e-01
4.02449034e-02 -2.36117728e-02 -3.71529400e-01 3.74617457e-01
-2.07967058e-01 6.27910376e-01 -5.47626734e-01 6.52130187e-01
5.09872377e-01 -2.42441982e-01 2.15718478e-01 -4.81063396e-01
2.16645766e-02 1.37333363e-01 -1.36872804e+00 2.08760595e+00
5.44070415e-02 3.14002305e-01 -1.79174282e-02 -1.20054698e+00
6.08384490e-01 2.15430573e-01 6.30066574e-01 -4.46562618e-01
2.06333250e-01 -8.19017589e-02 -1.39451399e-01 -8.33139956e-01
4.02232260e-01 -2.67333418e-01 -1.56865314e-01 3.16928864e-01
3.07514101e-01 3.38568121e-01 3.26909959e-01 3.27206790e-01
7.88977206e-01 4.56072539e-01 4.00105029e-01 -3.36826853e-02
5.59848249e-01 -1.83598071e-01 8.17247033e-01 6.11546814e-01
-4.48275208e-01 5.47032177e-01 7.23605752e-01 -7.23098293e-02
-5.67496538e-01 -1.02098453e+00 1.95701212e-01 1.22440743e+00
8.50571454e-01 -6.12237394e-01 -5.57359040e-01 -1.00075138e+00
-2.87375331e-01 3.55574220e-01 -4.59462434e-01 -3.70907709e-02
-4.37616140e-01 -5.94530106e-01 3.22606772e-01 6.14115536e-01
6.40498042e-01 -7.09416568e-01 -4.88569438e-01 -1.81706324e-02
-4.04957324e-01 -1.37247312e+00 -7.15626538e-01 -3.52632254e-02
-5.93772769e-01 -1.16503286e+00 -6.51903927e-01 -8.73259485e-01
2.80266076e-01 7.47436821e-01 9.66023207e-01 1.38737448e-02
1.27462402e-01 3.59737426e-01 -6.35533094e-01 1.68415662e-02
1.84165731e-01 3.49183567e-02 -2.29002923e-01 6.47999585e-01
3.29818904e-01 -6.43160403e-01 -6.39219820e-01 2.41964906e-01
-9.19058621e-01 5.41420162e-01 4.44977760e-01 5.67178905e-01
5.36905408e-01 6.17637858e-02 4.93924111e-01 -1.00985420e+00
4.26342189e-01 -9.68008816e-01 -2.32584566e-01 4.55298692e-01
-2.71772265e-01 -1.32299572e-01 4.23760504e-01 -2.05556676e-01
-1.27234244e+00 1.18365794e-01 1.16638966e-01 -5.09781599e-01
-2.67060220e-01 8.91858518e-01 -6.26996756e-01 5.08568764e-01
-4.64214832e-02 1.89174742e-01 -3.23608667e-01 -5.25536656e-01
5.78622162e-01 5.38926899e-01 7.88072944e-01 -3.48539025e-01
9.30142105e-01 5.22219539e-01 -2.01515138e-01 -5.67193866e-01
-1.18417323e+00 -1.00606108e+00 -8.39647889e-01 -3.80916566e-01
1.21490145e+00 -1.07216156e+00 -3.08388174e-01 4.25162733e-01
-1.00543940e+00 -6.99114501e-02 -1.09749861e-01 4.35624510e-01
-2.73483276e-01 7.40269899e-01 -4.88060266e-01 -5.30858457e-01
7.49191269e-02 -1.09341431e+00 1.04591632e+00 4.65194613e-01
-1.94951966e-02 -9.20596302e-01 1.84148252e-01 5.09025455e-01
-1.55278355e-01 2.36517012e-01 7.96620131e-01 -8.41164231e-01
-5.61384320e-01 -1.43239334e-01 -5.23309469e-01 2.28265002e-01
1.97749197e-01 -6.60958141e-02 -9.16500032e-01 -1.77716315e-01
1.07399590e-01 -3.37482125e-01 1.10301292e+00 6.81972727e-02
9.70464051e-01 -4.13088053e-02 -3.96599025e-01 5.13004303e-01
1.45944059e+00 1.91209927e-01 3.36277902e-01 1.93462238e-01
1.09181869e+00 7.90958405e-01 8.80985916e-01 4.92705405e-01
5.98701298e-01 8.23854268e-01 3.59270424e-01 -1.48658112e-01
-1.78028718e-01 -6.22395158e-01 3.95377010e-01 1.07584929e+00
-9.77096520e-03 -3.01439166e-01 -6.65565252e-01 4.21845168e-01
-2.31895494e+00 -9.58743215e-01 -2.59447932e-01 2.06103826e+00
6.49928212e-01 2.94387877e-01 3.25910032e-01 -9.86342132e-02
8.87752771e-01 6.97165191e-01 -4.46789891e-01 3.14998150e-01
1.17097117e-01 -4.44635570e-01 2.04939425e-01 2.73295343e-01
-1.44408643e+00 1.00730968e+00 4.83618355e+00 9.69427526e-01
-1.00084949e+00 1.43774480e-01 6.89507544e-01 -1.19105158e-02
-2.23403156e-01 3.09188664e-01 -5.99812806e-01 6.62429452e-01
3.91380996e-01 -3.53333876e-02 1.33746430e-01 7.37815619e-01
2.98736453e-01 -1.35240808e-01 -1.20021594e+00 9.70126510e-01
1.98475018e-01 -1.05290794e+00 1.38407439e-01 -2.39071026e-01
6.71824396e-01 -4.28618751e-02 -2.36018792e-01 3.63629520e-01
8.45149383e-02 -6.14209890e-01 1.03564513e+00 6.29914463e-01
3.66645932e-01 -5.12458205e-01 7.82502115e-01 3.83712322e-01
-1.77553368e+00 1.11825272e-01 -1.43017262e-01 -5.14918752e-02
6.04566336e-01 4.03890669e-01 -3.46119493e-01 9.06837821e-01
4.78893727e-01 1.26253557e+00 -6.55538023e-01 1.30511785e+00
-2.13345721e-01 5.48501611e-01 1.34963915e-01 3.67947698e-01
4.09072191e-01 -3.45889449e-01 5.97358167e-01 1.19379008e+00
3.19327265e-02 1.41671628e-01 6.98670208e-01 7.31637061e-01
-1.32295206e-01 1.26077190e-01 -3.98478806e-01 -1.20314928e-02
3.60501230e-01 1.45826089e+00 -1.08408427e+00 -4.66741502e-01
-7.59712398e-01 1.02529645e+00 2.10192531e-01 3.78965110e-01
-1.28590953e+00 -2.71846741e-01 4.47173804e-01 -6.42255321e-02
3.80098611e-01 -1.26597911e-01 -8.35500583e-02 -1.47751939e+00
5.05578099e-03 -6.04664981e-01 5.19390881e-01 -5.95735669e-01
-1.32125556e+00 3.96985054e-01 1.02585264e-01 -1.52233279e+00
1.72319159e-01 -2.88893878e-01 -9.70500827e-01 3.66555423e-01
-1.69588733e+00 -1.28332353e+00 -4.58085388e-01 6.75261319e-01
9.38508987e-01 1.57329410e-01 1.60023943e-01 4.92797524e-01
-9.00423944e-01 3.13960612e-01 -1.26273572e-01 2.77439326e-01
7.85617650e-01 -1.11805391e+00 3.98264602e-02 1.28742552e+00
4.30287480e-01 3.54819864e-01 4.96544272e-01 -7.84925461e-01
-1.11224163e+00 -1.18445730e+00 5.92968762e-01 -1.40974656e-01
8.07631791e-01 -2.88419098e-01 -9.10965323e-01 5.66972971e-01
3.71080011e-01 -1.90194026e-01 8.53389859e-01 3.81747968e-02
-3.00249040e-01 4.39407118e-02 -5.02530098e-01 5.64868927e-01
1.26468718e+00 -4.56132203e-01 -8.01554620e-01 -1.96143147e-02
8.95903885e-01 -1.56851679e-01 -6.98439002e-01 4.34125841e-01
2.22475946e-01 -1.04137826e+00 8.02092969e-01 -7.15290010e-01
8.40973496e-01 -5.43638051e-01 -3.76656592e-01 -1.01105964e+00
-2.64959902e-01 -5.63139379e-01 -2.05215707e-01 1.74179375e+00
1.26017824e-01 -1.06006861e-01 5.38174391e-01 3.65420759e-01
-2.84144223e-01 -7.71318555e-01 -5.76735675e-01 -5.51440537e-01
-5.23130000e-01 -6.03061497e-01 4.07021940e-01 1.08426285e+00
2.52558798e-01 8.50045323e-01 -6.32059753e-01 1.22700520e-01
5.07967889e-01 3.79055828e-01 8.00051987e-01 -1.21562696e+00
-4.79571462e-01 -5.03961623e-01 -3.82282883e-01 -1.60014045e+00
9.88466442e-02 -9.45338547e-01 2.61462599e-01 -1.49175215e+00
6.81836545e-01 -2.51021206e-01 -6.30691290e-01 3.94431114e-01
-4.35946405e-01 1.53379500e-01 3.74814123e-01 5.58120728e-01
-1.28818893e+00 6.33246362e-01 1.07423449e+00 -1.48696125e-01
-1.57770693e-01 -2.86337957e-02 -7.80174375e-01 1.09557295e+00
3.18066806e-01 -1.81974933e-01 -8.39457095e-01 -4.83929098e-01
3.77905592e-02 1.29631579e-01 3.13991666e-01 -1.01483297e+00
4.34648812e-01 -3.58632654e-01 -4.11021970e-02 -5.50659239e-01
1.89356759e-01 -8.01567852e-01 -1.16003610e-01 8.40435326e-02
-4.59369481e-01 -3.44897538e-01 -2.00864851e-01 1.02332890e+00
-5.48579931e-01 -2.11100519e-01 6.24871194e-01 -5.23148850e-02
-1.35590291e+00 4.37573045e-01 -6.61083236e-02 6.67898133e-02
1.32452655e+00 -2.11165115e-01 -2.46106207e-01 -3.05414885e-01
-7.22753644e-01 5.76776862e-01 2.70061225e-01 6.67536318e-01
5.52946210e-01 -1.32062197e+00 -4.10685062e-01 -9.24013406e-02
2.92623818e-01 4.58529703e-02 4.15862381e-01 1.21391618e+00
-1.16566055e-01 1.51580915e-01 -4.79572453e-02 -5.36687672e-01
-1.35220110e+00 6.97238624e-01 1.92412194e-02 -3.26508522e-01
-5.00004351e-01 9.16782141e-01 4.61888283e-01 -2.56652422e-02
4.78009433e-01 -5.25880679e-02 -6.24972939e-01 5.87734401e-01
3.61137688e-01 3.42500240e-01 -4.42084789e-01 -8.50377262e-01
-1.95036352e-01 5.31188071e-01 -1.54944971e-01 1.06986731e-01
1.23019958e+00 -6.22720540e-01 -1.35717109e-01 7.73476541e-01
1.20376623e+00 -1.04207054e-01 -1.45079029e+00 -5.03055274e-01
2.23372281e-01 -5.12254834e-01 -5.08149937e-02 -4.18677717e-01
-1.24670589e+00 8.55429411e-01 3.47187251e-01 2.88604558e-01
1.10223162e+00 3.23234051e-01 1.14342558e+00 -5.02549633e-02
1.78275019e-01 -1.12120891e+00 3.03144068e-01 2.58926749e-01
3.63841861e-01 -1.27326131e+00 -1.03846908e-01 -7.05719769e-01
-9.76426840e-01 9.25906777e-01 7.73922622e-01 1.31391004e-01
5.31364560e-01 -2.01099962e-01 -1.74122855e-01 -2.90429294e-01
-5.80313265e-01 -5.50724149e-01 5.59085786e-01 2.02896997e-01
3.79809827e-01 -7.53713623e-02 -3.78246188e-01 9.09990549e-01
3.56989861e-01 -1.39288185e-02 2.22844854e-01 9.50073063e-01
-6.02481484e-01 -1.03140450e+00 1.27624562e-02 2.31678888e-01
-3.51130456e-01 3.81302349e-02 -3.18172514e-01 5.55028975e-01
3.79437357e-01 9.53141510e-01 -1.05997458e-01 -7.29938924e-01
1.08764015e-01 7.08710775e-02 2.53116135e-02 -6.29519880e-01
-2.50001311e-01 5.20537555e-01 -4.20276709e-02 -6.36746645e-01
-7.71757782e-01 -7.09924698e-01 -1.10390365e+00 6.18495829e-02
-3.58310431e-01 5.58629222e-02 3.21354456e-02 1.12876582e+00
6.01482615e-02 6.77296937e-01 6.34881735e-01 -6.99069917e-01
-2.09403381e-01 -6.84644818e-01 -6.10883534e-01 7.45329022e-01
4.87049147e-02 -6.84244812e-01 -8.80251154e-02 2.67706484e-01] | [9.493596076965332, 0.6137098670005798] |
4fa06af9-3252-4948-91f5-8c41c57f08b1 | large-scale-evolution-of-convolutional-neural | 1703.05422 | null | http://arxiv.org/abs/1703.05422v1 | http://arxiv.org/pdf/1703.05422v1.pdf | Large Scale Evolution of Convolutional Neural Networks Using Volunteer Computing | This work presents a new algorithm called evolutionary exploration of
augmenting convolutional topologies (EXACT), which is capable of evolving the
structure of convolutional neural networks (CNNs). EXACT is in part modeled
after the neuroevolution of augmenting topologies (NEAT) algorithm, with
notable exceptions to allow it to scale to large scale distributed computing
environments and evolve networks with convolutional filters. In addition to
multithreaded and MPI versions, EXACT has been implemented as part of a BOINC
volunteer computing project, allowing large scale evolution. During a period of
two months, over 4,500 volunteered computers on the Citizen Science Grid
trained over 120,000 CNNs and evolved networks reaching 98.32% test data
accuracy on the MNIST handwritten digits dataset. These results are even
stronger as the backpropagation strategy used to train the CNNs was fairly
rudimentary (ReLU units, L2 regularization and Nesterov momentum) and these
were initial test runs done without refinement of the backpropagation
hyperparameters. Further, the EXACT evolutionary strategy is independent of the
method used to train the CNNs, so they could be further improved by advanced
techniques like elastic distortions, pretraining and dropout. The evolved
networks are also quite interesting, showing "organic" structures and
significant differences from standard human designed architectures. | ['Travis Desell'] | 2017-03-15 | null | null | null | null | ['l2-regularization'] | ['methodology'] | [-9.60047841e-02 3.64620358e-01 4.79623824e-01 -2.15845510e-01
5.87917030e-01 -5.98529577e-01 5.78473985e-01 -1.99029908e-01
-1.02701592e+00 9.74476755e-01 -2.55076557e-01 -4.77129906e-01
-1.57183677e-01 -7.93281555e-01 -9.32716668e-01 -7.09187031e-01
-4.02333260e-01 5.98095894e-01 4.19880450e-01 -6.44424200e-01
2.67677121e-02 8.66895139e-01 -1.64158547e+00 1.58700690e-01
7.03918219e-01 7.01161563e-01 2.49230061e-02 1.14165294e+00
6.25379160e-02 6.18495882e-01 -6.96926177e-01 -6.28136575e-01
4.17541325e-01 -3.26687217e-01 -7.67674804e-01 -3.45621139e-01
2.38373578e-01 2.13608995e-01 -2.97205538e-01 8.91270220e-01
6.49383307e-01 1.91768333e-01 2.29904443e-01 -1.24290359e+00
-6.16729736e-01 6.17816091e-01 1.99651774e-02 3.70106488e-01
-3.82105738e-01 4.08476770e-01 4.49546278e-01 -6.72775447e-01
7.43083715e-01 1.15070498e+00 1.19671845e+00 7.85293877e-01
-1.13625753e+00 -6.10795140e-01 -1.46387890e-01 4.75991605e-04
-1.18375766e+00 -2.06247821e-01 2.78641731e-01 -2.51596391e-01
1.49288344e+00 3.75997543e-01 1.34326887e+00 1.03500307e+00
2.72307962e-01 2.75327981e-01 9.97674525e-01 -5.96379340e-01
5.04599690e-01 4.89397615e-01 -6.75765648e-02 8.75443697e-01
3.70959848e-01 5.16610980e-01 -1.75671488e-01 -3.06252718e-01
7.85081983e-01 -4.24676478e-01 -1.27908245e-01 -2.14629710e-01
-9.53588545e-01 8.18449259e-01 6.54754281e-01 4.09728825e-01
-5.08514881e-01 3.78443390e-01 4.90865260e-01 6.77627206e-01
9.58977118e-02 1.15449893e+00 -7.67956495e-01 -1.99331656e-01
-7.97032237e-01 2.82684416e-01 1.07354903e+00 5.90962946e-01
7.72290647e-01 4.72122908e-01 1.56613290e-01 6.07326627e-01
-7.62624061e-03 -1.72673613e-01 7.72982836e-01 -1.08142066e+00
-7.95754418e-02 6.74654603e-01 -4.49918777e-01 -6.91766918e-01
-5.54669917e-01 -6.11993670e-01 -8.63435984e-01 7.70477891e-01
2.80552238e-01 -8.79645109e-01 -1.06846154e+00 1.44683838e+00
1.93502069e-01 1.08314827e-01 3.83056663e-02 5.71259677e-01
5.94588041e-01 4.41563606e-01 2.34046765e-02 3.32277507e-01
9.92570519e-01 -1.16276944e+00 -1.36252299e-01 5.91107719e-02
4.94529903e-01 -4.12072778e-01 6.44666970e-01 5.64438105e-01
-1.20531964e+00 -6.06212556e-01 -1.14088583e+00 4.96111065e-01
-7.09452450e-01 -2.91888237e-01 8.79125774e-01 1.18824875e+00
-1.72494900e+00 1.14602983e+00 -9.95106637e-01 -4.95563209e-01
6.19348466e-01 7.60963261e-01 -2.30130032e-01 3.13760042e-01
-1.06313264e+00 9.95655060e-01 8.83858442e-01 5.01929708e-02
-8.47058654e-01 -6.94765627e-01 -3.12209994e-01 2.86228303e-02
-1.00216299e-01 -8.15339744e-01 7.64200747e-01 -1.57978320e+00
-1.79065633e+00 6.25087321e-01 4.84009087e-01 -7.28140473e-01
5.28020859e-01 2.65831470e-01 -3.49294484e-01 -1.69057772e-01
-7.42701590e-01 1.19290006e+00 5.72182894e-01 -9.18426991e-01
-3.95452827e-01 -2.51511559e-02 -1.24838039e-01 7.35239265e-03
-7.54645288e-01 2.86550313e-01 -1.67487711e-01 -6.80395484e-01
-2.43763238e-01 -1.18218517e+00 -4.98379618e-01 9.43796784e-02
-2.63520721e-02 7.70363659e-02 7.25889921e-01 -5.95006526e-01
6.80331469e-01 -1.83410418e+00 4.07303274e-01 3.37677509e-01
2.04257414e-01 7.23500311e-01 -3.83562982e-01 1.74645856e-01
-3.68938833e-01 3.75141382e-01 -3.20569336e-01 -5.81763200e-02
-1.93556458e-01 4.55407262e-01 3.47943336e-01 1.15963981e-01
4.83772427e-01 9.67381120e-01 -6.10844254e-01 5.21914996e-02
-3.68122429e-01 7.37775445e-01 -8.14498544e-01 4.68677133e-02
-1.19391665e-01 4.02960598e-01 1.22299623e-02 4.08264518e-01
4.47517812e-01 7.50439391e-02 -1.63759142e-02 2.77041733e-01
-3.97695601e-01 -3.27552915e-01 -9.53272045e-01 1.54696441e+00
-1.36222884e-01 9.35090661e-01 1.22148111e-01 -9.45169270e-01
1.15907633e+00 2.92103857e-01 2.08628297e-01 -5.57885528e-01
1.60681993e-01 4.02909338e-01 5.79869032e-01 -4.69003260e-01
6.06411219e-01 8.19817558e-02 5.10902882e-01 1.61730766e-01
3.95272076e-01 -3.27915013e-01 2.25349367e-01 -1.22240335e-01
1.44264758e+00 2.91445076e-01 -2.80290008e-01 -4.32066411e-01
1.82809219e-01 4.14787263e-01 5.91816008e-01 5.76404095e-01
-1.18611053e-01 5.32552958e-01 4.87381816e-01 -9.78713155e-01
-1.70113695e+00 -5.34078836e-01 -2.08431214e-01 9.58225310e-01
-5.79954267e-01 -1.16040543e-01 -9.11681771e-01 -4.05401975e-01
-1.97065577e-01 5.35307229e-01 -8.16997230e-01 -3.00041318e-01
-7.40855098e-01 -1.13437343e+00 1.25435662e+00 4.55410749e-01
8.34224284e-01 -1.53522885e+00 -9.49085414e-01 3.91111970e-01
6.89601302e-01 -4.02525246e-01 2.69679576e-01 7.76491165e-01
-1.10056305e+00 -9.84708667e-01 -1.00746238e+00 -8.15896213e-01
6.74430430e-01 -5.23383141e-01 1.13737011e+00 5.24263263e-01
-7.27183640e-01 -6.20462596e-02 -2.51928747e-01 -6.11534417e-01
-5.83747804e-01 3.17484319e-01 -2.97994502e-02 -4.66647804e-01
-9.14894864e-02 -8.67502153e-01 -2.36045986e-01 1.09515935e-01
-8.89599979e-01 -8.58511850e-02 6.47224247e-01 1.19488883e+00
1.39888763e-01 2.16853738e-01 2.37251312e-01 -8.79197121e-01
8.75419438e-01 -4.94054109e-01 -5.81981719e-01 3.05778801e-01
-7.69866347e-01 6.57173172e-02 5.80422461e-01 -6.94589674e-01
-8.96582127e-01 2.76307552e-03 -2.94092834e-01 -4.28463608e-01
-2.86816388e-01 4.30565268e-01 3.82809192e-01 -6.01105094e-01
1.04242599e+00 -3.12541090e-02 1.22897662e-01 -3.11571747e-01
-3.20155285e-02 2.49649137e-01 4.39522684e-01 -6.33993387e-01
6.24371648e-01 2.89979000e-02 -2.01771319e-01 -8.22495759e-01
2.29204103e-01 3.96637648e-01 -5.22441924e-01 -6.38227537e-02
7.69167483e-01 -4.75340933e-01 -6.16118670e-01 8.40723991e-01
-1.08947527e+00 -9.41013813e-01 -6.36488795e-01 2.91596532e-01
-1.20597899e-01 -2.65656888e-01 -7.08158731e-01 -5.52802026e-01
-2.71566510e-01 -9.89239931e-01 1.11501917e-01 6.52437091e-01
-1.65991753e-01 -1.06449306e+00 3.36027741e-01 -2.39210963e-01
1.19909573e+00 5.10261476e-01 1.04745197e+00 -6.01136208e-01
-8.30782950e-02 7.45039573e-03 4.48621577e-03 6.57739520e-01
-4.14891571e-01 5.09290218e-01 -8.00683200e-01 -5.73839545e-01
-2.38314316e-01 -4.27411884e-01 8.54377210e-01 -3.01575810e-02
1.03124225e+00 -3.01626682e-01 -1.84515640e-01 1.04095745e+00
1.29648495e+00 4.40661490e-01 9.88148689e-01 8.92361581e-01
4.74681079e-01 6.29166484e-01 -3.82371932e-01 2.50308901e-01
-1.09122671e-01 3.36972892e-01 6.61521137e-01 -2.44733006e-01
-2.46684581e-01 3.74136120e-01 2.29711875e-01 7.33881891e-01
-6.52785063e-01 -4.18752395e-02 -1.28798473e+00 2.57953912e-01
-1.52279043e+00 -9.54970837e-01 -1.72985226e-01 1.83919013e+00
7.67297804e-01 1.63459420e-01 3.57644521e-02 -1.28084406e-01
6.89206958e-01 -4.43785220e-01 -8.01583350e-01 -1.00855577e+00
-5.53280413e-01 7.80698061e-01 6.47246957e-01 1.12250350e-01
-7.02788115e-01 7.70925224e-01 7.21499872e+00 3.30029458e-01
-1.26987898e+00 1.90990925e-01 7.98928499e-01 -3.18286091e-01
-7.73828700e-02 5.68568446e-02 -6.17072344e-01 4.62338597e-01
1.32861662e+00 3.30245234e-02 8.42894673e-01 7.56380200e-01
-2.96425581e-01 1.19238153e-01 -5.35405338e-01 3.66582096e-01
-2.48749495e-01 -1.74301481e+00 -3.04779828e-01 1.60387442e-01
1.17207277e+00 6.84511125e-01 -8.39298442e-02 5.48881173e-01
7.56472051e-01 -1.34156048e+00 8.84678304e-01 5.84908187e-01
6.77360833e-01 -1.16527557e+00 8.33932042e-01 3.37035894e-01
-6.92286551e-01 -4.47667301e-01 -7.35135794e-01 -9.32766348e-02
-2.88419962e-01 2.38882020e-01 -6.36403143e-01 1.45245999e-01
1.29130971e+00 1.00230835e-01 -1.11061037e+00 1.32182813e+00
7.31813908e-02 6.07291996e-01 -3.92171919e-01 -3.18500966e-01
4.26237553e-01 3.31561007e-02 5.51768363e-01 1.27978551e+00
2.77373701e-01 -1.23343773e-01 -4.99020845e-01 1.01065445e+00
-1.10360004e-01 -2.55736887e-01 -2.97775298e-01 -9.21748951e-02
4.59630340e-01 1.40086961e+00 -8.26504052e-01 -1.94238782e-01
7.90273547e-02 9.13958132e-01 5.67179918e-01 3.72316986e-01
-8.16563725e-01 -7.17772841e-01 5.80143452e-01 -9.40307975e-02
4.85435069e-01 -1.84836686e-01 -9.24387947e-02 -7.77440965e-01
-3.59679192e-01 -1.23143613e+00 -1.74144022e-02 -7.16190994e-01
-9.21032429e-01 1.28851879e+00 -3.19918901e-01 -4.98468846e-01
-2.29127519e-03 -8.59020650e-01 -1.02507877e+00 9.33631659e-01
-1.12245619e+00 -7.81626403e-01 -4.71657813e-01 5.50463200e-01
7.99563378e-02 -7.27802575e-01 1.06963301e+00 1.74838394e-01
-8.76578987e-01 6.38396084e-01 3.20454031e-01 3.65663730e-02
2.52786636e-01 -1.17910409e+00 7.11073518e-01 5.95946431e-01
-1.18358865e-01 7.34283805e-01 6.18007779e-01 -4.99177814e-01
-1.21151984e+00 -1.10396326e+00 6.15259767e-01 1.70028675e-02
5.34290195e-01 -1.55117929e-01 -1.03333783e+00 5.12386382e-01
7.78849900e-01 1.04146846e-01 2.98372537e-01 1.11313388e-02
-7.81989545e-02 2.19161958e-01 -1.30195105e+00 4.88902807e-01
1.08413148e+00 1.46601886e-01 -2.66945302e-01 2.07169324e-01
7.30824769e-01 -5.47270656e-01 -7.04273522e-01 1.36699453e-01
4.20402944e-01 -1.12841606e+00 6.24793887e-01 -9.47120488e-01
2.84740001e-01 2.13346654e-03 1.54020801e-01 -1.60607409e+00
-4.59880948e-01 -8.41560900e-01 2.38520894e-02 1.15428770e+00
6.89109385e-01 -1.04810929e+00 1.12389684e+00 8.15030456e-01
-4.05991465e-01 -8.88936818e-01 -9.24129725e-01 -8.69119704e-01
3.56411606e-01 -5.96893206e-02 8.24067116e-01 1.04263127e+00
-4.90125805e-01 -1.36348158e-01 1.07841395e-01 -3.31073940e-01
1.60552889e-01 -7.44340301e-01 5.64629614e-01 -1.28365147e+00
-8.15017939e-01 -9.58695650e-01 -6.09614432e-01 -2.21984312e-02
-2.43257638e-02 -9.00131762e-01 -2.47418135e-01 -9.29562569e-01
-1.19110636e-01 -8.93889666e-01 -1.68331146e-01 8.19201827e-01
2.78529882e-01 3.95465285e-01 1.26540959e-01 1.14476964e-01
1.78824160e-02 3.93826038e-01 8.82893860e-01 -6.67230636e-02
-3.04847986e-01 -1.33963615e-01 -5.37257195e-01 6.55746102e-01
1.01173294e+00 -7.12121308e-01 -1.12156600e-01 -7.39631355e-01
5.52772284e-01 -5.81704319e-01 5.67646265e-01 -1.43117309e+00
4.56294686e-01 2.31684953e-01 6.58400059e-01 1.10979840e-01
7.36146420e-02 -6.14741623e-01 8.54703069e-01 1.05966675e+00
-3.82345080e-01 7.02726901e-01 7.16145635e-01 1.74789891e-01
-6.11923076e-02 -5.93239009e-01 9.53769863e-01 -4.30799842e-01
-7.56413877e-01 1.18678503e-01 -5.92981279e-01 -1.39887393e-01
1.16746366e+00 -4.66674775e-01 -4.28020477e-01 2.63437301e-01
-7.55455554e-01 1.07776940e-01 6.72945023e-01 1.97345793e-01
4.11581457e-01 -1.10216582e+00 -8.28065157e-01 3.59702468e-01
-5.19177496e-01 6.96573406e-03 -9.91619900e-02 6.21305168e-01
-1.20583916e+00 2.11502723e-02 -8.91596735e-01 -3.74785662e-01
-1.15832627e+00 3.06998938e-01 8.95674825e-01 -1.42905936e-01
-6.38326585e-01 1.39133668e+00 -6.32071376e-01 -8.29742432e-01
3.16362649e-01 2.02659756e-01 -1.40144944e-01 -8.99267122e-02
4.03779030e-01 4.38191593e-01 4.30775046e-01 -6.80339560e-02
-2.74974525e-01 1.31458312e-01 5.22497781e-02 -1.69616506e-01
2.06139231e+00 5.74338794e-01 -4.11390781e-01 -8.36180747e-02
1.03021264e+00 -5.68543077e-01 -1.48615217e+00 1.31514966e-01
2.76435278e-02 1.53760254e-01 5.59296496e-02 -9.49902236e-01
-1.50642312e+00 7.92610884e-01 7.82363355e-01 4.79153618e-02
1.13157380e+00 -5.24754465e-01 3.85019124e-01 8.92879844e-01
2.90209800e-01 -1.09822214e+00 2.29913518e-01 8.33361804e-01
9.23082113e-01 -6.68518543e-01 -8.27787071e-02 4.19991672e-01
-5.39789438e-01 1.52685285e+00 8.19539845e-01 -5.37523448e-01
5.18498480e-01 7.06094682e-01 -1.01938218e-01 -3.45276058e-01
-1.02237391e+00 2.95027882e-01 -1.38012096e-01 8.20525765e-01
4.37727511e-01 -2.35720292e-01 -6.60653114e-02 2.69601010e-02
-4.09340024e-01 -3.64482701e-02 4.99427497e-01 1.11431515e+00
-3.86930168e-01 -1.10938346e+00 -3.47038031e-01 2.83426702e-01
-2.30368033e-01 -2.25391552e-01 -4.63598162e-01 1.02530479e+00
5.01657546e-01 3.25292945e-01 1.37955427e-01 -4.95169461e-01
3.33821684e-01 1.02637596e-01 5.32250106e-01 -3.94328684e-01
-1.61524427e+00 -5.82555532e-01 1.05276212e-01 -3.40696245e-01
-9.25177410e-02 -5.64214230e-01 -1.13280165e+00 -7.48613834e-01
-4.23392117e-01 1.50969073e-01 1.12241566e+00 5.09307981e-01
4.71057385e-01 8.55971217e-01 3.31645519e-01 -1.34312081e+00
-4.51895118e-01 -8.81866693e-01 -4.26018894e-01 -2.46800259e-02
-2.73744732e-01 -1.92336932e-01 -2.51185685e-01 -4.30806875e-02] | [8.281147956848145, 3.243543863296509] |
cea1cd7b-fb2c-4ae8-bbf8-a9083744e449 | learning-6-dof-fine-grained-grasp-detection | 2301.11564 | null | https://arxiv.org/abs/2301.11564v1 | https://arxiv.org/pdf/2301.11564v1.pdf | Learning 6-DoF Fine-grained Grasp Detection Based on Part Affordance Grounding | Robotic grasping is a fundamental ability for a robot to interact with the environment. Current methods focus on how to obtain a stable and reliable grasping pose in object wise, while little work has been studied on part (shape)-wise grasping which is related to fine-grained grasping and robotic affordance. Parts can be seen as atomic elements to compose an object, which contains rich semantic knowledge and a strong correlation with affordance. However, lacking a large part-wise 3D robotic dataset limits the development of part representation learning and downstream application. In this paper, we propose a new large Language-guided SHape grAsPing datasEt (named Lang-SHAPE) to learn 3D part-wise affordance and grasping ability. We design a novel two-stage fine-grained robotic grasping network (named PIONEER), including a novel 3D part language grounding model, and a part-aware grasp pose detection model. To evaluate the effectiveness, we perform multi-level difficulty part language grounding grasping experiments and deploy our proposed model on a real robot. Results show our method achieves satisfactory performance and efficiency in reference identification, affordance inference, and 3D part-aware grasping. Our dataset and code are available on our project website https://sites.google.com/view/lang-shape | ['Yue Zhang', 'Yu Zheng', 'Yi Ren', 'Penglei Sun', 'Yaoxian Song'] | 2023-01-27 | null | null | null | null | ['robotic-grasping'] | ['robots'] | [-2.29820967e-01 -2.22210675e-01 -1.94286227e-01 -5.56324601e-01
-4.40836966e-01 -8.24828923e-01 2.99518228e-01 -3.83900315e-01
1.76526740e-01 -9.44305584e-02 1.23373397e-01 5.91451637e-02
-4.08970952e-01 -8.37309599e-01 -1.20758498e+00 -5.90047896e-01
-2.20398858e-01 5.66413701e-01 2.84722596e-01 -3.63452822e-01
3.07707489e-01 8.44069779e-01 -1.62099874e+00 4.85113233e-01
5.75724721e-01 1.18694103e+00 1.05746710e+00 2.66576648e-01
-2.08971798e-01 3.51912349e-01 -2.11285129e-01 1.23666577e-01
6.62011087e-01 3.26794118e-01 -7.24431217e-01 6.53638244e-02
2.42043197e-01 -8.34201157e-01 -3.43076766e-01 1.07652175e+00
3.95404130e-01 -1.89972416e-01 6.90352321e-01 -1.46994853e+00
-1.01188290e+00 9.47822094e-01 -2.27752283e-01 -6.23239577e-01
2.36132935e-01 2.37243757e-01 7.06018448e-01 -8.54490459e-01
4.59919006e-01 1.96739960e+00 3.27193201e-01 7.73973286e-01
-5.04156768e-01 -5.75173378e-01 3.38393658e-01 -2.27872550e-01
-9.74379718e-01 -6.18299954e-02 8.25036168e-01 -4.60979521e-01
8.03737283e-01 -1.35794759e-01 4.58669126e-01 1.05440235e+00
1.49579525e-01 1.14522433e+00 7.21515119e-01 -1.14617892e-01
2.49264808e-03 -4.83334601e-01 7.59025738e-02 8.85802388e-01
4.18291181e-01 -1.87070563e-01 -1.19925618e-01 2.95225419e-02
1.44228101e+00 5.95973372e-01 1.34668037e-01 -8.11227798e-01
-1.61770856e+00 3.50625068e-01 9.88989472e-01 2.18102798e-01
-5.31264424e-01 6.37988687e-01 3.01632553e-01 2.37223268e-01
-7.77170956e-02 3.60680342e-01 -8.05342555e-01 1.12365678e-01
2.37242132e-01 5.65832496e-01 9.02115226e-01 1.97402227e+00
6.95163071e-01 -4.23478067e-01 -3.03914785e-01 8.67034912e-01
5.89274287e-01 9.49705780e-01 2.08586350e-01 -1.24472690e+00
4.97253180e-01 1.17600238e+00 2.46199220e-01 -6.40837371e-01
-4.04055715e-01 1.19966514e-01 -6.07490838e-01 1.00238129e-01
2.77075499e-01 3.77999067e-01 -1.03827500e+00 1.52519917e+00
3.37228209e-01 -6.43886209e-01 -1.19124196e-01 1.22495770e+00
9.30986106e-01 2.92659670e-01 1.68234229e-01 5.08786500e-01
1.63086033e+00 -1.15427446e+00 -2.84558207e-01 -7.47446064e-03
3.81236523e-01 -7.31349707e-01 1.18367207e+00 3.27032387e-01
-8.24539602e-01 -6.56172216e-01 -8.80610168e-01 -4.53577757e-01
-4.76247728e-01 6.14540517e-01 1.12436283e+00 2.51873285e-02
-5.26042819e-01 6.62532389e-01 -9.91996467e-01 -4.52990174e-01
6.61030531e-01 3.96791309e-01 -4.77900803e-01 -5.44665992e-01
-6.05640650e-01 8.59855533e-01 6.29501104e-01 2.47453496e-01
-1.12506998e+00 -2.68481582e-01 -7.82177866e-01 -1.53954506e-01
6.78844452e-01 -7.18029022e-01 1.49130666e+00 -9.72141102e-02
-1.30427659e+00 8.14436793e-01 2.57747233e-01 2.50840873e-01
3.74839157e-01 -5.65472126e-01 3.12823594e-01 1.23198412e-01
3.21329623e-01 8.39700818e-01 1.00382078e+00 -1.61127484e+00
-2.41817385e-01 -6.74956322e-01 5.02252579e-01 2.49659643e-03
3.56053226e-02 -1.51950404e-01 -3.93057227e-01 -6.31324351e-01
8.94346893e-01 -8.36531520e-01 3.61811966e-02 5.66046417e-01
-3.91842395e-01 -8.37316096e-01 7.88445175e-01 -4.13456976e-01
1.16692655e-01 -2.09222603e+00 3.43314886e-01 -1.88860357e-01
2.47616395e-02 -4.30705436e-02 -5.34212530e-01 6.98954225e-01
3.84582222e-01 -5.45843579e-02 -4.32599224e-02 1.06884941e-01
5.48031986e-01 4.33503538e-01 -5.44363260e-01 2.08926409e-01
3.37946892e-01 1.22594202e+00 -1.09204984e+00 -3.11429769e-01
1.58468187e-01 8.58654901e-02 -7.15580344e-01 6.43166482e-01
-6.56255841e-01 1.88612938e-01 -1.29493487e+00 1.53506231e+00
8.78433824e-01 -4.09400463e-02 1.46153979e-02 -7.29194999e-01
-1.10911414e-01 -1.18920714e-01 -9.03391004e-01 2.16668057e+00
-4.79632884e-01 -4.69405085e-01 4.16509777e-01 -7.98349261e-01
1.35169339e+00 -6.71113282e-02 4.72473383e-01 -2.49139562e-01
2.92102247e-01 6.43794119e-01 -6.03925856e-03 -9.04018283e-01
2.52156287e-01 2.75306225e-01 -1.35270268e-01 4.26527798e-01
1.24371886e-01 -8.34246159e-01 -8.23990479e-02 -3.23209912e-02
8.79452825e-01 9.77876604e-01 2.36077607e-02 -5.26266932e-01
4.10899185e-02 4.28164266e-02 8.63604173e-02 6.90977395e-01
2.75882594e-02 3.75841260e-01 1.45122394e-01 -3.90162468e-01
-1.12291896e+00 -1.19649315e+00 -1.90607551e-02 1.26182151e+00
7.69275606e-01 5.15750423e-03 -6.12716675e-01 -3.87275219e-01
7.21788108e-01 1.84518889e-01 -4.33182567e-01 -1.11665592e-01
-7.84377933e-01 -1.95978820e-01 2.06066653e-01 7.59897053e-01
7.03972876e-01 -1.60221314e+00 -7.50849962e-01 1.40272796e-01
-1.22188248e-01 -1.13443589e+00 -3.33381474e-01 1.26690298e-01
-9.97302353e-01 -1.30155337e+00 -8.46151948e-01 -1.16678274e+00
9.19113040e-01 7.46144235e-01 7.37622380e-01 2.40788415e-01
-4.69280213e-01 6.47356093e-01 -8.55669320e-01 -5.93626440e-01
-2.78567493e-01 1.33392647e-01 3.11612427e-01 -6.08404040e-01
2.88645685e-01 -4.63478029e-01 -6.21997654e-01 3.97336364e-01
-9.43326354e-01 -4.57642525e-02 1.35716403e+00 6.58228874e-01
5.27028799e-01 -3.83129507e-01 5.87618768e-01 7.36398250e-02
6.03242517e-01 -3.35487276e-01 -3.95380706e-01 4.78435367e-01
5.11573888e-02 1.18991211e-01 3.20379943e-01 -6.09009206e-01
-7.30693579e-01 1.08141616e-01 8.45619962e-02 -7.47001827e-01
-3.90745819e-01 1.60407290e-01 -5.46648324e-01 -1.39570087e-01
2.23408088e-01 2.13844299e-01 2.86780864e-01 -1.02855515e+00
5.37830174e-01 9.81016457e-01 4.57606375e-01 -1.31091988e+00
7.13596463e-01 3.26819152e-01 2.08192952e-02 -5.43226361e-01
-7.21628785e-01 -4.86756772e-01 -1.14208221e+00 -5.88458739e-02
6.93307161e-01 -9.41510379e-01 -1.23885083e+00 7.86365211e-01
-1.48003888e+00 -4.46703255e-01 -4.74919677e-02 4.07547921e-01
-1.02954769e+00 3.86119217e-01 -6.68282688e-01 -6.95872903e-01
-6.28014147e-01 -1.20493901e+00 1.84961736e+00 -1.56765893e-01
3.31377923e-01 4.73700576e-02 -6.58619463e-01 1.04407340e-01
2.52909601e-01 1.90221474e-01 9.29764628e-01 -3.03434372e-01
-1.06508636e+00 -1.93937272e-01 -7.79763997e-01 3.75625789e-01
5.24232805e-01 -3.34707081e-01 -3.95529360e-01 -4.12775576e-01
-9.76082757e-02 -6.99160695e-01 6.77763879e-01 8.29363763e-02
1.41125298e+00 -2.01195896e-01 -4.24061656e-01 2.95494318e-01
1.27485871e+00 5.64310327e-02 1.78570092e-01 1.78248540e-01
8.86995554e-01 7.93504834e-01 1.03827083e+00 5.18264294e-01
3.51810664e-01 4.15997058e-01 8.73843133e-01 6.22167230e-01
-2.63908088e-01 -5.03017724e-01 1.41192809e-01 7.80747414e-01
-4.92563218e-01 1.82534121e-02 -7.61290610e-01 4.34626609e-01
-1.85599637e+00 -6.63811684e-01 2.54928712e-02 1.51063073e+00
5.74891508e-01 -2.10051656e-01 -1.32129237e-01 -2.79090345e-01
5.11606932e-01 -3.63631845e-01 -9.40330148e-01 2.45422021e-01
1.32430777e-01 -2.69678652e-01 2.97631353e-01 -8.68079513e-02
-9.73915577e-01 1.23470378e+00 4.63180113e+00 6.96750522e-01
-8.11942756e-01 -1.73703194e-01 -3.13476324e-01 4.57879931e-01
-2.06063509e-01 -3.29439044e-01 -6.18248999e-01 6.53051734e-02
-4.08602178e-01 3.76760066e-01 6.63625777e-01 1.47174227e+00
-1.56072691e-01 1.53090984e-01 -1.51447690e+00 1.06126058e+00
-3.52697521e-02 -9.05423939e-01 4.46816117e-01 -1.91041708e-01
2.38761991e-01 8.40883031e-02 -4.54468012e-01 4.76596445e-01
2.99481690e-01 -7.61438131e-01 1.26814747e+00 6.92322016e-01
8.12680900e-01 -1.11981630e-01 4.42745119e-01 6.07380986e-01
-1.28822029e+00 -3.84164840e-01 -8.19975436e-01 -1.34845497e-03
-2.31159590e-02 1.45099089e-01 -4.73470211e-01 7.21074462e-01
9.78595078e-01 7.65405178e-01 -6.13487922e-02 8.59716296e-01
-3.15936416e-01 -2.43827790e-01 -2.85526723e-01 -5.23687780e-01
1.69717342e-01 8.38510096e-02 4.70350772e-01 7.94586301e-01
5.24637938e-01 5.31099677e-01 6.32583737e-01 1.29956925e+00
-6.28002509e-02 -1.73378050e-01 -5.00001967e-01 -3.54113162e-01
5.86701155e-01 1.09067357e+00 -6.38605058e-01 -1.75269097e-01
-1.26080438e-01 8.32180738e-01 3.83323491e-01 6.35536388e-02
-3.05818409e-01 -4.15838242e-01 6.52576149e-01 -2.39896446e-01
3.68447006e-01 -8.61754179e-01 -9.28264335e-02 -1.14469802e+00
3.55313599e-01 -6.89977407e-01 -2.13917732e-01 -1.08371651e+00
-1.76238990e+00 3.86774033e-01 3.05863053e-01 -1.25218213e+00
2.04084694e-01 -1.43775654e+00 -1.38832301e-01 7.68158078e-01
-1.37915051e+00 -1.82900369e+00 -7.66432881e-01 5.90868175e-01
6.98282957e-01 -1.02541916e-01 8.35575461e-01 -1.17096022e-01
6.08738288e-02 7.71967173e-02 -3.94243211e-01 1.95243567e-01
4.35960770e-01 -8.95228028e-01 3.63631248e-01 1.24439150e-01
-4.25077349e-01 1.02701020e+00 3.25800508e-01 -7.83158123e-01
-2.54494953e+00 -1.06359625e+00 1.37250572e-01 -6.18069232e-01
5.73139548e-01 -6.68059289e-01 -6.34749115e-01 6.88881576e-01
-4.92118537e-01 -4.37857554e-04 -2.18077406e-01 -1.47955552e-01
-5.77471375e-01 9.41910222e-02 -1.38523686e+00 5.17791629e-01
1.87150824e+00 -3.01964283e-01 -9.18197513e-01 6.34328842e-01
1.23355126e+00 -5.15839338e-01 -1.07301450e+00 1.01865923e+00
1.09279549e+00 -3.80124867e-01 1.15895212e+00 -5.54969132e-01
4.47599411e-01 -2.71986544e-01 -8.47648680e-01 -8.35275173e-01
-4.82482105e-01 -1.18344590e-01 -1.29850671e-01 9.00525331e-01
-3.16171706e-01 -4.61150467e-01 4.37257618e-01 2.06914335e-01
-6.23853803e-01 -8.30667257e-01 -4.58455384e-01 -1.27334487e+00
1.68487370e-01 -2.91538149e-01 1.20939314e+00 4.72717464e-01
-7.26128817e-02 -2.57214576e-01 1.12653479e-01 3.44840080e-01
7.43354619e-01 9.96907234e-01 9.50384557e-01 -1.31281936e+00
1.56080490e-02 -6.17719233e-01 -2.97561944e-01 -1.58654249e+00
1.40648320e-01 -1.07532465e+00 5.40444255e-01 -1.83923888e+00
3.73137206e-01 -9.76793706e-01 -8.22962684e-05 1.01950312e+00
2.71230906e-01 -1.52416840e-01 3.96325976e-01 6.03135526e-01
-3.76395434e-01 7.98766971e-01 2.06265426e+00 -4.28113878e-01
5.61062060e-02 -5.83807938e-02 -4.25021976e-01 5.11416852e-01
9.20099080e-01 -6.72533438e-02 -8.47005397e-02 -8.89862895e-01
-1.92798674e-01 -1.01710550e-01 7.58589149e-01 -6.48705184e-01
-1.00150838e-01 -3.77141953e-01 1.86597571e-01 -6.93697989e-01
3.80921394e-01 -9.92946506e-01 -3.49731654e-01 7.04148829e-01
-6.54895380e-02 -2.62596220e-01 -1.75352045e-03 4.53752667e-01
2.26310208e-01 -3.05845350e-01 2.28762284e-01 -7.10908055e-01
-1.13731575e+00 7.84073055e-01 3.43937695e-01 -4.64651138e-01
9.06402946e-01 1.81951635e-02 -2.76627392e-01 2.52273679e-01
-4.42055404e-01 3.14239025e-01 5.78063250e-01 1.13802445e+00
7.27592587e-01 -1.47479594e+00 -5.89196205e-01 1.97341189e-01
4.53259647e-01 4.52219069e-01 1.84959203e-01 2.34535083e-01
-6.20411217e-01 3.71883899e-01 -5.68276644e-01 -6.78712964e-01
-6.64093792e-01 7.95275807e-01 7.44157890e-03 6.93172097e-01
-7.01496243e-01 9.17057753e-01 1.88401371e-01 -1.27612674e+00
3.95436227e-01 -9.54492331e-01 1.58982068e-01 -6.32321656e-01
1.34454802e-01 3.23762625e-01 -3.81649047e-01 -3.32653165e-01
-4.38543200e-01 9.29606736e-01 2.58488327e-01 5.74049175e-01
1.45676947e+00 1.85426116e-01 -7.64011919e-01 3.37385327e-01
1.04325926e+00 -4.53878403e-01 -1.34508955e+00 -3.01234778e-02
-1.52017608e-01 -3.98316056e-01 -6.28882229e-01 -8.89235258e-01
-6.42926395e-01 8.77196670e-01 2.71230221e-01 3.52472207e-03
5.67324102e-01 7.03127205e-01 6.78158760e-01 1.18390763e+00
1.38996387e+00 -6.88811064e-01 3.84719223e-01 7.99258590e-01
1.82720876e+00 -1.45394528e+00 -1.91101283e-01 -8.34498405e-01
-1.27912000e-01 1.27342522e+00 9.04676020e-01 -4.44244981e-01
8.27696502e-01 1.52058601e-01 -1.14023484e-01 -4.74999636e-01
-1.83143795e-01 -7.78281391e-02 2.77854502e-01 6.60581231e-01
8.13077986e-02 4.15821582e-01 1.11081591e-02 9.32905316e-01
-2.78692156e-01 8.13973099e-02 -1.24688216e-01 1.30253685e+00
-8.39368701e-01 -1.01579905e+00 -8.28708485e-02 3.96822035e-01
2.08707035e-01 2.74700731e-01 -2.11496919e-01 6.56360984e-01
3.03706139e-01 7.03788221e-01 -2.00077057e-01 -3.29760611e-01
6.12616599e-01 -2.15041012e-01 1.11998236e+00 -7.79403210e-01
2.65092198e-02 -2.65375197e-01 -4.38556433e-01 -9.31458235e-01
-4.97591436e-01 -4.63857919e-01 -1.58080661e+00 1.35188356e-01
-2.28931606e-01 -3.05157959e-01 1.10106301e+00 9.07363534e-01
4.33523685e-01 4.24748093e-01 4.06376272e-01 -1.60374641e+00
-1.15447927e+00 -1.27204347e+00 -8.24100673e-01 3.89982611e-01
1.07482836e-01 -1.09352350e+00 5.95185198e-02 -2.65995771e-01] | [5.737176895141602, -0.8964018225669861] |
3054a947-647d-4916-bdad-ddf753c95c6d | multi-level-cross-modal-feature-alignment-via | 2306.06066 | null | https://arxiv.org/abs/2306.06066v1 | https://arxiv.org/pdf/2306.06066v1.pdf | Multi-level Cross-modal Feature Alignment via Contrastive Learning towards Zero-shot Classification of Remote Sensing Image Scenes | Zero-shot classification of image scenes which can recognize the image scenes that are not seen in the training stage holds great promise of lowering the dependence on large numbers of labeled samples. To address the zero-shot image scene classification, the cross-modal feature alignment methods have been proposed in recent years. These methods mainly focus on matching the visual features of each image scene with their corresponding semantic descriptors in the latent space. Less attention has been paid to the contrastive relationships between different image scenes and different semantic descriptors. In light of the challenge of large intra-class difference and inter-class similarity among image scenes and the potential noisy samples, these methods are susceptible to the influence of the instances which are far from these of the same classes and close to these of other classes. In this work, we propose a multi-level cross-modal feature alignment method via contrastive learning for zero-shot classification of remote sensing image scenes. While promoting the single-instance level positive alignment between each image scene with their corresponding semantic descriptors, the proposed method takes the cross-instance contrastive relationships into consideration,and learns to keep the visual and semantic features of different classes in the latent space apart from each other. Extensive experiments have been done to evaluate the performance of the proposed method. The results show that our proposed method outperforms state of the art methods for zero-shot remote sensing image scene classification. All the code and data are available at github https://github.com/masuqiang/MCFA-Pytorch | ['Zhigang Han', 'Wei Yang', 'Zheng Li', 'Suqiang Ma', 'Chun Liu'] | 2023-05-31 | null | null | null | null | ['scene-classification'] | ['computer-vision'] | [ 4.34920281e-01 -4.95372593e-01 -1.02274515e-01 -4.02056009e-01
-7.88589656e-01 -1.76100716e-01 5.38925290e-01 2.46397197e-01
-1.69888377e-01 2.59982646e-01 -6.79645687e-02 3.79124552e-01
-4.78193551e-01 -8.79356802e-01 -3.66255015e-01 -1.13886619e+00
1.06546655e-01 1.13090068e-01 3.27104062e-01 -6.16835877e-02
2.60215282e-01 1.69616833e-01 -2.15181303e+00 3.45092416e-01
6.34575605e-01 1.02684534e+00 5.95594227e-01 3.70649517e-01
-2.36407340e-01 7.60775447e-01 -3.02290320e-01 1.39379606e-01
3.09435874e-01 -4.55316007e-01 -5.60042500e-01 5.06338120e-01
8.52314055e-01 -8.55682045e-02 -1.23340718e-01 1.54734707e+00
5.45349300e-01 4.40339267e-01 6.84740782e-01 -1.43286610e+00
-6.41040027e-01 6.04271255e-02 -6.71436250e-01 3.70844930e-01
-3.92457657e-02 -1.08114593e-02 1.04491103e+00 -1.08052659e+00
5.33056259e-01 1.13127542e+00 2.93698370e-01 2.53084302e-01
-9.83549356e-01 -6.23571455e-01 2.92162508e-01 5.82031488e-01
-1.69520724e+00 -4.26400095e-01 9.24801350e-01 -8.12601209e-01
5.87609649e-01 4.01427895e-01 5.39251268e-01 6.96314216e-01
6.30079024e-03 6.44200742e-01 1.16236997e+00 -3.84097219e-01
3.71081740e-01 2.95641750e-01 4.84315544e-01 3.79277557e-01
-1.01885438e-01 -1.10717721e-01 -2.51244098e-01 -4.03833240e-02
2.58665353e-01 6.75017297e-01 -4.11604166e-01 -6.01313531e-01
-9.13414299e-01 8.92504752e-01 6.94033027e-01 5.00455976e-01
-2.31050566e-01 -3.77641439e-01 3.09437513e-01 2.12133408e-01
6.03730619e-01 9.45554227e-02 -2.94236898e-01 4.12843555e-01
-7.82008171e-01 2.04746220e-02 3.23119044e-01 8.58183801e-01
1.17304015e+00 -1.87869415e-01 -1.96038455e-01 1.15421939e+00
1.75783128e-01 6.45206749e-01 4.63670641e-01 -5.03949165e-01
5.01210809e-01 7.50892937e-01 -1.35605440e-01 -1.41433036e+00
-3.71444039e-02 -4.29592103e-01 -9.97590601e-01 8.55463147e-02
5.52747771e-02 4.58118528e-01 -1.07236040e+00 1.40396369e+00
4.47964072e-01 4.31650788e-01 3.33782107e-01 8.43629539e-01
1.14258611e+00 1.01665473e+00 1.30904064e-01 -3.27814102e-01
1.27533448e+00 -9.53942060e-01 -3.83107960e-01 -4.21930552e-01
4.86031443e-01 -7.96167374e-01 1.21559274e+00 -1.91128850e-01
-4.31713372e-01 -7.86131024e-01 -1.00346148e+00 2.28525475e-01
-6.63943172e-01 7.84792900e-02 3.74581844e-01 2.26059854e-01
-3.79393011e-01 3.99692684e-01 -4.91980106e-01 -6.19383395e-01
5.32232404e-01 -9.41792578e-02 -4.47002232e-01 -5.99467754e-01
-9.56856847e-01 4.95049030e-01 5.40631533e-01 1.97230682e-01
-1.00458884e+00 -6.66423798e-01 -9.37335074e-01 1.59938425e-01
4.91404802e-01 5.67903295e-02 7.35192895e-01 -1.15966678e+00
-7.28512108e-01 9.36987638e-01 -9.85973477e-02 4.03703004e-01
2.31748462e-01 1.06778935e-01 -5.89381695e-01 1.05411902e-01
4.77499843e-01 4.48418379e-01 7.07742453e-01 -1.32711923e+00
-6.56294227e-01 -5.77625155e-01 -8.78875703e-02 3.45876902e-01
-4.28389966e-01 -7.29311956e-03 -1.98357299e-01 -3.70165825e-01
5.75383544e-01 -8.04809153e-01 -8.48746598e-02 1.68913603e-01
-1.44440427e-01 -4.34298180e-02 1.13010490e+00 -1.37365326e-01
8.27243090e-01 -2.43853378e+00 -1.21401444e-01 3.93603444e-02
-1.64948136e-01 3.84411603e-01 -2.79720873e-01 4.52495992e-01
-1.99128047e-01 -1.10567518e-01 -3.32807720e-01 1.56054541e-01
-2.79866576e-01 3.33614200e-01 -2.77891874e-01 6.51135981e-01
1.41313910e-01 6.27000809e-01 -1.10665941e+00 -6.05717063e-01
5.23289084e-01 3.31193596e-01 -1.50300190e-01 4.08155084e-01
1.26668289e-01 4.75261122e-01 -4.60008204e-01 7.66338944e-01
8.39961469e-01 -3.62925678e-01 -6.74450072e-03 -4.23603296e-01
-2.92556752e-02 -2.84127742e-01 -1.43401456e+00 1.31036353e+00
-1.44618183e-01 3.02823842e-01 -3.28412443e-01 -1.27819431e+00
9.15005684e-01 2.40192056e-01 5.43771267e-01 -9.04448092e-01
8.46958682e-02 9.03179273e-02 -2.13042811e-01 -8.11365604e-01
3.51578206e-01 -3.63389224e-01 -7.14253709e-02 1.16419435e-01
1.95844948e-01 -1.94103897e-01 2.34515309e-01 5.94502054e-02
4.56808776e-01 -1.86833844e-01 5.45570135e-01 -2.92074025e-01
5.54293573e-01 4.74422285e-03 6.62366271e-01 8.11952770e-01
-3.81813735e-01 5.75585902e-01 -9.76744220e-02 -3.87493491e-01
-1.05777788e+00 -1.00051677e+00 -5.37129462e-01 1.08697879e+00
6.68919921e-01 -1.55599132e-01 -2.43198723e-01 -3.24803829e-01
-9.74043310e-02 3.79273355e-01 -7.03876019e-01 -1.50306240e-01
8.42751712e-02 -8.76970232e-01 -3.44219506e-02 2.41093948e-01
6.28205419e-01 -8.56600583e-01 -5.34839153e-01 -1.93255711e-02
-2.72182047e-01 -9.78552163e-01 -5.13116270e-02 2.35120505e-01
-7.01147676e-01 -1.14644229e+00 -6.08460307e-01 -9.82435882e-01
8.59075129e-01 9.82428014e-01 7.48551071e-01 2.36116927e-02
-5.50493479e-01 2.69877434e-01 -6.48615301e-01 -2.01612458e-01
-7.81871825e-02 -2.12433606e-01 -2.61252850e-01 4.60699320e-01
5.51404297e-01 -4.22127336e-01 -6.15479529e-01 5.75263858e-01
-1.09952497e+00 1.09994635e-01 3.15263987e-01 1.02064013e+00
7.54568040e-01 3.32688361e-01 1.60317078e-01 -8.59952450e-01
-3.21232006e-02 -7.56143093e-01 -4.21047568e-01 5.09043157e-01
-4.12216783e-01 -3.10981631e-01 4.36767817e-01 -3.43581796e-01
-9.59301472e-01 1.06595591e-01 3.11577260e-01 -5.84446609e-01
-4.28272039e-01 5.27888894e-01 -2.97534525e-01 5.11745326e-02
5.78974426e-01 3.72243851e-01 -1.77884981e-01 -3.55361193e-01
2.79716462e-01 9.77453709e-01 2.77752489e-01 -3.10819149e-01
7.69707203e-01 6.55120373e-01 -2.35084578e-01 -1.10132194e+00
-1.28577495e+00 -1.14094245e+00 -7.68088102e-01 -4.73646849e-01
8.75938952e-01 -1.24324155e+00 1.88434660e-01 7.44730115e-01
-5.68156362e-01 6.16964512e-02 -2.78556883e-01 4.30187643e-01
-2.25204989e-01 4.93374437e-01 -2.45491669e-01 -7.49888957e-01
-1.71706870e-01 -1.14546049e+00 1.17696881e+00 3.69330108e-01
2.66887516e-01 -7.33706295e-01 3.76875028e-02 3.70254874e-01
2.43242189e-01 1.49070561e-01 9.26542222e-01 -5.75928748e-01
-4.20555234e-01 -2.12647766e-01 -3.43076527e-01 4.89505798e-01
4.75555778e-01 6.48079962e-02 -1.12311113e+00 -4.71135795e-01
2.06629187e-01 -5.52578211e-01 7.09289849e-01 2.67578304e-01
8.93262327e-01 -1.50065407e-01 -2.53998190e-01 6.07786059e-01
1.92632794e+00 2.06126302e-01 6.08874500e-01 2.55111068e-01
9.05020714e-01 7.73926258e-01 1.09592021e+00 5.31024992e-01
1.76595241e-01 6.02574229e-01 4.36370254e-01 -2.08543107e-01
1.05693586e-01 -1.13127850e-01 9.46892425e-02 7.67225564e-01
7.89750963e-02 -1.87208161e-01 -8.39677036e-01 6.80561364e-01
-1.89982176e+00 -1.36315525e+00 -3.54304522e-01 2.06864977e+00
4.57138121e-01 -2.54661769e-01 -3.42337042e-01 2.09125847e-01
9.78376985e-01 6.05078995e-01 -4.30541605e-01 1.67158306e-01
-2.33817086e-01 -6.98465258e-02 2.74089038e-01 2.74918526e-01
-1.44479108e+00 8.60539317e-01 4.81618023e+00 1.02980232e+00
-1.29468787e+00 3.15187752e-01 6.29184484e-01 3.74570787e-02
6.67621046e-02 1.91261783e-01 -7.33259261e-01 4.95834559e-01
3.37148577e-01 -5.76553531e-02 1.24478810e-01 1.01716256e+00
8.72132089e-03 -3.07729125e-01 -7.93499649e-01 1.11648989e+00
3.41135472e-01 -1.00380147e+00 1.33961961e-01 -1.35212913e-01
1.04542625e+00 1.28881827e-01 -3.92531604e-02 2.15191662e-01
-1.51483819e-01 -7.20299363e-01 6.50478721e-01 5.53153992e-01
6.33833051e-01 -4.01240289e-01 8.12726378e-01 4.36638504e-01
-1.52583253e+00 -3.33243459e-01 -7.56935537e-01 -1.28252000e-01
-1.76893383e-01 5.59835911e-01 -3.85317355e-01 6.09961748e-01
7.78497994e-01 1.13201761e+00 -7.81701505e-01 9.78011966e-01
-1.71793008e-03 2.75160313e-01 6.08006455e-02 1.75969869e-01
3.00367802e-01 -2.95999050e-01 3.86967510e-01 8.39465499e-01
4.11641955e-01 1.71897918e-01 5.96077859e-01 7.28926063e-01
4.47829992e-01 3.13300878e-01 -7.22421706e-01 -2.98883468e-02
3.68112326e-01 1.27663183e+00 -7.49185681e-01 -6.02962554e-01
-4.97606903e-01 7.75840163e-01 3.52866612e-02 2.87325144e-01
-7.76670218e-01 -2.06667304e-01 5.27652383e-01 -1.09900041e-02
3.09875041e-01 5.08345962e-02 9.27249789e-02 -1.38389516e+00
1.21447198e-01 -5.69323540e-01 6.45271838e-01 -9.50648606e-01
-1.60124135e+00 5.46828985e-01 1.77829474e-01 -1.83734071e+00
2.13745758e-01 -3.32866967e-01 -5.58113158e-01 6.87750280e-01
-1.76170433e+00 -1.35146117e+00 -8.66333008e-01 8.32666695e-01
6.36681974e-01 -1.71589211e-01 9.21147108e-01 3.97278130e-01
-4.17833090e-01 1.90718949e-01 5.03407061e-01 1.55823171e-01
4.42479104e-01 -7.69829512e-01 -2.39997894e-01 9.64867055e-01
9.55896676e-02 3.59117299e-01 5.69496632e-01 -4.74017978e-01
-9.70898986e-01 -1.33003008e+00 6.46650076e-01 -2.21942794e-02
6.51842296e-01 -2.27531165e-01 -1.14810896e+00 2.67510116e-01
-2.08464026e-01 4.69021201e-01 8.48150671e-01 -1.23369016e-01
-5.21071434e-01 -4.57197458e-01 -8.99257421e-01 2.01593190e-01
9.05968904e-01 -8.51877153e-01 -5.95263839e-01 4.97423977e-01
2.04729021e-01 9.24338177e-02 -7.28383899e-01 5.11697650e-01
3.53961080e-01 -8.71470630e-01 8.59256268e-01 -4.03689176e-01
5.34725845e-01 -6.17471397e-01 -6.71033323e-01 -1.11785781e+00
-4.35483754e-01 4.24964190e-01 5.75998962e-01 1.28122437e+00
-6.89647645e-02 -3.96711081e-01 3.95450741e-01 2.92204142e-01
-4.47663814e-02 -3.59767348e-01 -8.46363962e-01 -8.75806093e-01
-1.96460366e-01 -1.12888932e-01 2.10267305e-01 1.31018019e+00
-2.76470512e-01 5.36756396e-01 -6.03830755e-01 4.51707214e-01
6.77658260e-01 7.18881667e-01 7.58820772e-01 -1.31530702e+00
-2.62904972e-01 -1.98893964e-01 -8.49868715e-01 -4.76696521e-01
-7.87818283e-02 -1.17785907e+00 2.42102116e-01 -1.47303092e+00
9.02835786e-01 -4.29288805e-01 -5.81934869e-01 6.58900738e-01
-3.19812119e-01 3.95363063e-01 3.53052616e-01 5.82611501e-01
-7.53654122e-01 7.41967976e-01 1.01970840e+00 -5.71315706e-01
3.95634845e-02 -1.90358430e-01 -3.85114342e-01 4.96883452e-01
6.38930142e-01 -7.18973160e-01 -5.74134290e-01 -2.65591264e-01
-1.65005356e-01 -1.92528740e-01 6.03199959e-01 -1.24965489e+00
2.20179439e-01 -4.79334325e-01 2.28009760e-01 -5.05329788e-01
3.29256237e-01 -9.96605158e-01 4.31104630e-01 4.27918017e-01
-1.54396892e-01 -3.44673663e-01 -1.50323212e-01 6.49692357e-01
-5.60728848e-01 -3.51234972e-01 1.21397400e+00 -4.47243392e-01
-1.28380895e+00 4.60562736e-01 -1.20656267e-01 1.14337252e-02
1.23197186e+00 -4.25157726e-01 -5.41084349e-01 -7.03433752e-02
-7.15223670e-01 1.52969718e-01 4.34284151e-01 5.52391589e-01
6.34982288e-01 -1.19241107e+00 -6.94297433e-01 3.86864007e-01
9.61648405e-01 -1.03171982e-01 9.39789414e-01 7.83632159e-01
-1.95337012e-01 1.01219311e-01 -5.58206558e-01 -9.66667891e-01
-1.57776904e+00 5.46611607e-01 4.88965094e-01 9.44679156e-02
-5.33319294e-01 7.65310764e-01 6.29991472e-01 -4.45379317e-01
3.24721970e-02 -2.37765796e-02 -3.52958024e-01 4.73301083e-01
5.62484026e-01 1.08572751e-01 3.02992333e-02 -1.13946331e+00
-4.31098729e-01 8.84718835e-01 -3.99905145e-02 4.88546818e-01
1.60095203e+00 -1.76826134e-01 -2.50757933e-01 9.15112436e-01
1.53218925e+00 -3.49341333e-01 -1.05980802e+00 -6.38492823e-01
-1.44075453e-01 -9.82505798e-01 1.65008277e-01 -5.00019610e-01
-1.22428501e+00 9.71702814e-01 1.19495463e+00 1.60914689e-01
1.12637591e+00 1.47627175e-01 2.90123224e-01 1.59406185e-01
3.75382364e-01 -1.10428691e+00 2.59347558e-01 4.50075358e-01
4.73086506e-01 -1.70262611e+00 1.26457751e-01 -4.02658343e-01
-6.79045618e-01 8.59126508e-01 5.74184895e-01 -1.61287159e-01
7.73017228e-01 -3.44212651e-01 2.61768430e-01 -5.55795789e-01
-4.33191240e-01 -5.06880879e-01 2.83322334e-01 4.15060997e-01
1.25109017e-01 2.22334832e-01 -1.33224964e-01 7.37537444e-02
3.45969439e-01 -3.00285935e-01 2.38942906e-01 1.03082812e+00
-6.89273655e-01 -8.57747078e-01 -2.15714827e-01 3.72292489e-01
-1.22864805e-01 -1.99058443e-01 -1.06006943e-01 5.78432441e-01
3.20299119e-01 1.02275324e+00 1.92565992e-01 -4.54763412e-01
2.23574191e-01 4.70918491e-02 1.59492865e-01 -7.52322376e-01
-1.87090605e-01 7.16480389e-02 -2.80056357e-01 -2.92308867e-01
-8.22261155e-01 -6.60604358e-01 -8.87982905e-01 9.59343985e-02
-4.60478663e-01 1.67112395e-01 4.63671833e-01 8.22270930e-01
2.25849882e-01 2.64968574e-01 9.91898239e-01 -9.09067810e-01
-5.09913683e-01 -8.94126832e-01 -9.49387848e-01 8.57812583e-01
2.76495427e-01 -8.42436433e-01 -6.95099473e-01 3.72778215e-02] | [9.793902397155762, 1.9657878875732422] |
e7b7c65d-9f1b-4d15-bd2b-fd30609756a2 | predicting-real-time-scientific-experiments | 2204.11718 | null | https://arxiv.org/abs/2204.11718v1 | https://arxiv.org/pdf/2204.11718v1.pdf | Predicting Real-time Scientific Experiments Using Transformer models and Reinforcement Learning | Life and physical sciences have always been quick to adopt the latest advances in machine learning to accelerate scientific discovery. Examples of this are cell segmentation or cancer detection. Nevertheless, these exceptional results are based on mining previously created datasets to discover patterns or trends. Recent advances in AI have been demonstrated in real-time scenarios like self-driving cars or playing video games. However, these new techniques have not seen widespread adoption in life or physical sciences because experimentation can be slow. To tackle this limitation, this work aims to adapt generative learning algorithms to model scientific experiments and accelerate their discovery using in-silico simulations. We particularly focused on real-time experiments, aiming to model how they react to user inputs. To achieve this, here we present an encoder-decoder architecture based on the Transformer model to simulate real-time scientific experimentation, predict its future behaviour and manipulate it on a step-by-step basis. As a proof of concept, this architecture was trained to map a set of mechanical inputs to the oscillations generated by a chemical reaction. The model was paired with a Reinforcement Learning controller to show how the simulated chemistry can be manipulated in real-time towards user-defined behaviours. Our results demonstrate how generative learning can model real-time scientific experimentation to track how it changes through time as the user manipulates it, and how the trained models can be paired with optimisation algorithms to discover new phenomena beyond the physical limitations of lab experimentation. This work paves the way towards building surrogate systems where physical experimentation interacts with machine learning on a step-by-step basis. | ['Juan Manuel Parrilla-Gutierrez'] | 2022-04-25 | null | null | null | null | ['scientific-results-extraction'] | ['natural-language-processing'] | [ 1.93698362e-01 9.71326381e-02 8.22768286e-02 9.94035453e-02
-1.02901213e-01 -5.76817572e-01 7.99255192e-01 7.95147941e-02
-3.75904739e-01 8.25406194e-01 -6.40038013e-01 -4.29829061e-01
-2.16846183e-01 -8.90842140e-01 -1.07069850e+00 -9.81296360e-01
-2.28556439e-01 5.96443832e-01 3.59767109e-01 -1.81535050e-01
2.96103239e-01 6.64988697e-01 -1.92210734e+00 2.01332539e-01
2.98072875e-01 2.15914950e-01 4.32702601e-01 1.03884208e+00
2.02020645e-01 4.16978061e-01 -5.67860305e-01 3.29828620e-01
-6.11044168e-02 -7.67412424e-01 -4.48959500e-01 -2.70358711e-01
-5.43526769e-01 4.56857622e-01 4.45431359e-02 5.52981734e-01
5.76543450e-01 -8.41917545e-02 6.15380168e-01 -9.23389018e-01
-4.44821715e-01 5.27366221e-01 8.93714093e-03 -6.31092712e-02
4.68391448e-01 8.06411326e-01 2.78851956e-01 -3.04471076e-01
9.55534697e-01 1.08863366e+00 3.89403462e-01 6.97885692e-01
-1.62762761e+00 -5.31550527e-01 -3.06069702e-01 2.15785384e-01
-9.87858713e-01 -1.81095570e-01 7.63089120e-01 -4.98698086e-01
8.35034132e-01 2.24688813e-01 1.16480768e+00 1.44748664e+00
8.22256029e-01 3.37937295e-01 1.33792222e+00 -5.07165670e-01
8.15962136e-01 2.64000624e-01 -5.99430859e-01 4.77658778e-01
1.30203113e-01 5.29949427e-01 -3.93955141e-01 2.36211568e-01
9.24355268e-01 -2.53464133e-01 -1.87702239e-01 -1.48455650e-01
-1.18727553e+00 4.93717104e-01 1.32370532e-01 8.58023763e-01
-3.20790619e-01 4.98802990e-01 1.31709322e-01 4.01991576e-01
-1.75759166e-01 1.31394362e+00 -7.38392889e-01 -3.23180854e-01
-9.09175754e-01 4.56809878e-01 8.47915530e-01 2.76313007e-01
6.41609609e-01 1.14921421e-01 1.65635526e-01 2.34678701e-01
4.75451320e-01 4.46206152e-01 8.78306091e-01 -8.41168761e-01
-5.92916191e-01 6.11589432e-01 -5.12858573e-03 -6.02866888e-01
-7.80992687e-01 -4.53689516e-01 -6.03056908e-01 6.68986499e-01
4.95964527e-01 -2.63923436e-01 -9.01733875e-01 1.62446392e+00
5.98672330e-01 3.07003438e-01 1.56347513e-01 8.49141479e-01
2.77261704e-01 8.91879857e-01 -3.01736873e-03 -3.81435215e-01
1.19908249e+00 -1.82869211e-01 -5.53124428e-01 2.51813650e-01
8.29787493e-01 -5.25475681e-01 9.70485866e-01 6.00083530e-01
-9.38697517e-01 -6.96167946e-01 -1.03327632e+00 4.09065604e-01
-8.32967877e-01 -1.67762086e-01 4.91973400e-01 6.17617667e-01
-9.64987040e-01 1.19413650e+00 -1.17665458e+00 -4.53787953e-01
1.40874103e-01 5.62363505e-01 4.19936255e-02 7.19042063e-01
-1.29349101e+00 8.92281413e-01 4.68189836e-01 -8.59846100e-02
-9.50220644e-01 -1.17044663e+00 -4.21963125e-01 -9.12604854e-02
1.61141574e-01 -9.79230285e-01 1.07623041e+00 -5.82805693e-01
-2.45738053e+00 7.47514725e-01 1.05485104e-01 -5.35485268e-01
4.94480729e-01 2.69217104e-01 -4.79927570e-01 -2.01894984e-01
-2.09264949e-01 6.08529031e-01 4.32666749e-01 -1.25088358e+00
-2.50169605e-01 -8.40590596e-02 -3.12329501e-01 -5.10075331e-01
9.98898745e-02 -3.21476042e-01 2.23325510e-02 -2.60079056e-01
-2.73479193e-01 -1.04661953e+00 -3.63185495e-01 -1.62428215e-01
-2.57263184e-01 7.16267526e-02 1.09133506e+00 1.83955699e-01
7.70535648e-01 -1.65622330e+00 3.17958832e-01 2.20742077e-01
-1.60026893e-01 5.20869792e-01 1.65435165e-01 8.08481991e-01
-2.34665096e-01 3.96926366e-02 -1.37360349e-01 8.30626711e-02
-1.28746524e-01 4.37885709e-02 -1.70820937e-01 2.14797586e-01
3.63326639e-01 1.12579095e+00 -1.03005445e+00 -3.71433608e-02
3.15378577e-01 6.47205234e-01 -3.06966364e-01 3.18518817e-01
-8.91646087e-01 1.18939829e+00 -3.80031615e-01 2.78288603e-01
6.91769198e-02 -1.45590112e-01 1.19589202e-01 8.07081386e-02
-5.55164397e-01 1.37971174e-02 -9.33683217e-01 1.42661965e+00
-4.78492945e-01 4.97596115e-01 -2.33048901e-01 -1.36270571e+00
1.18051279e+00 4.24077511e-01 5.49562991e-01 -1.00128758e+00
5.08912623e-01 3.72949511e-01 4.03460264e-01 -8.35184097e-01
-3.23512763e-01 -3.24128717e-01 4.95859794e-02 4.35685724e-01
-2.03542918e-01 -7.93882847e-01 3.48767728e-01 -3.36906463e-01
1.12683809e+00 6.64973497e-01 4.33400879e-03 -3.18772346e-01
6.84004068e-01 2.54514456e-01 1.73272993e-02 5.12590587e-01
3.27837497e-01 2.25150421e-01 4.58354503e-01 -5.61820388e-01
-1.23777366e+00 -7.57830739e-01 -5.63704073e-02 5.76563239e-01
-6.81240764e-03 -5.44066615e-02 -8.09376240e-01 7.86425918e-02
-8.56920555e-02 8.57666135e-01 -7.29228497e-01 -5.36593378e-01
-4.95271206e-01 -9.35039759e-01 2.64432162e-01 -7.46200457e-02
1.30940795e-01 -1.42561090e+00 -1.30917788e+00 6.44333303e-01
7.64372051e-01 -6.94947541e-01 3.23188156e-01 5.59707761e-01
-8.38217378e-01 -9.51471388e-01 -5.33586144e-01 -5.03248513e-01
3.79955024e-01 -3.31691325e-01 6.76905572e-01 -1.81077108e-01
-9.34645653e-01 2.56387979e-01 -1.07913226e-01 -6.94176674e-01
-1.12445545e+00 1.15176655e-01 2.53467381e-01 -2.43174240e-01
1.67467296e-02 -1.10189342e+00 -6.37701452e-01 3.30160022e-01
-1.13895273e+00 2.06914917e-01 6.43839180e-01 7.26965487e-01
7.22668409e-01 1.63813233e-01 6.83973491e-01 -8.24993849e-01
5.41816056e-01 -3.57878298e-01 -8.86739969e-01 7.33430833e-02
-7.43162453e-01 5.53560555e-01 1.02899659e+00 -8.44266534e-01
-6.44807160e-01 3.25454414e-01 -1.35266393e-01 -3.89711499e-01
-4.36183125e-01 6.16331220e-01 -8.38381499e-02 1.20899633e-01
7.06182182e-01 4.20663804e-01 2.12816879e-01 -1.67105705e-01
1.71473220e-01 4.06991601e-01 3.28412414e-01 -7.06819117e-01
6.64338648e-01 2.53262818e-01 6.53441668e-01 -7.89373159e-01
-1.97193503e-01 8.47557932e-03 -5.72411716e-01 -4.75023270e-01
6.71442509e-01 -3.12396735e-01 -1.48369408e+00 4.12154704e-01
-7.29395330e-01 -9.41572368e-01 -5.30617356e-01 4.95885342e-01
-9.61577296e-01 -3.22566181e-01 -3.58897954e-01 -9.75801408e-01
-1.96414486e-01 -1.11613464e+00 1.03762877e+00 6.87770128e-01
-5.58242440e-01 -1.01621723e+00 5.83713889e-01 -1.74675688e-01
5.69379628e-01 6.00504637e-01 8.49448621e-01 -3.06559116e-01
-6.13784552e-01 -2.47956604e-01 6.13650084e-01 -3.66072536e-01
3.66742052e-02 6.25513136e-01 -8.69116664e-01 -6.03870675e-02
6.61780164e-02 -9.58489329e-02 4.69761550e-01 3.44314039e-01
1.14462388e+00 -6.90024123e-02 -6.74893081e-01 4.42993522e-01
1.30500412e+00 6.70753717e-01 8.62970769e-01 4.75322425e-01
3.60259444e-01 6.03951037e-01 1.92323714e-01 9.30766165e-02
-1.35896027e-01 7.57889986e-01 2.82187968e-01 -6.56206103e-04
7.24746496e-05 -1.98823273e-01 4.20341372e-01 5.57455778e-01
-2.84362048e-01 -1.30375057e-01 -8.95433426e-01 3.04690748e-01
-1.54873681e+00 -1.18600953e+00 -3.19731802e-01 2.35024548e+00
1.05974758e+00 4.11570311e-01 2.27565214e-01 2.68037736e-01
3.09651583e-01 -5.65245807e-01 -6.90767467e-01 -8.32426608e-01
1.30196407e-01 5.53553641e-01 2.36357585e-01 2.18021452e-01
-6.15054250e-01 7.02730596e-01 5.75151014e+00 5.69970012e-01
-1.96958840e+00 -2.77138919e-01 7.02915490e-01 -7.51103237e-02
-9.94694382e-02 -2.23313663e-02 -6.26312971e-01 7.30434239e-01
1.68334234e+00 -2.13170037e-01 3.89128059e-01 5.21620750e-01
7.99870610e-01 -2.96679884e-01 -1.29189837e+00 5.34919739e-01
-5.80400348e-01 -1.63159263e+00 -1.34455517e-01 3.83176394e-02
4.23015028e-01 -3.19445461e-01 3.08261346e-03 2.37812370e-01
2.49847949e-01 -1.16298139e+00 5.23057640e-01 9.58824456e-01
6.42548025e-01 -3.98185700e-01 3.92935991e-01 6.28674984e-01
-6.79163218e-01 1.87650546e-01 -7.66802207e-02 -2.01451749e-01
1.56377196e-01 5.55661976e-01 -1.20171630e+00 3.10974121e-01
4.18681890e-01 3.93955529e-01 -2.30156258e-01 1.07894158e+00
-8.64987001e-02 9.02940392e-01 -5.05124629e-01 -7.66125143e-01
3.41128148e-02 -2.26094261e-01 4.55937535e-01 1.22065175e+00
5.67278922e-01 -2.12669298e-01 -4.84113008e-01 1.47061133e+00
3.96050721e-01 -1.93102583e-01 -6.56149030e-01 -3.17040116e-01
2.00166386e-02 1.12718761e+00 -1.07925987e+00 -8.85294974e-02
2.76977509e-01 5.73087811e-01 -2.80467033e-01 2.83252329e-01
-1.02257884e+00 -3.57357234e-01 4.66353685e-01 5.37020266e-01
3.16771179e-01 -3.41167003e-01 -1.28701761e-01 -7.24410057e-01
-4.48182613e-01 -4.32231694e-01 -3.15097272e-01 -9.54914272e-01
-6.65000618e-01 1.07069500e-01 -9.04130191e-02 -8.87941539e-01
-3.53673309e-01 -6.13727212e-01 -6.17383003e-01 6.72087371e-01
-1.03011155e+00 -7.32241750e-01 4.98279817e-02 1.03960715e-01
2.57578343e-01 4.84084710e-02 9.81733799e-01 -1.72392830e-01
-6.37254655e-01 1.47179276e-01 3.34335893e-01 -3.31105888e-01
4.28067803e-01 -1.14333355e+00 2.80244499e-01 3.64795059e-01
1.61975041e-01 3.70995402e-01 1.40525830e+00 -5.62429488e-01
-1.78381109e+00 -8.19860876e-01 4.59766895e-01 -2.62949497e-01
6.43861949e-01 -6.36597335e-01 -1.08624613e+00 8.79064947e-02
2.42434293e-02 -1.87632255e-02 2.96085298e-01 -4.40410942e-01
6.12126701e-02 -1.91049874e-01 -1.03463662e+00 8.59794974e-01
7.52769470e-01 -1.64446786e-01 -2.39284202e-01 3.51630121e-01
4.70987409e-01 -4.96982872e-01 -9.10269022e-01 3.27481151e-01
7.30063856e-01 -6.98438168e-01 7.72403836e-01 -5.59622049e-01
6.28177166e-01 -4.78697211e-01 6.81654632e-01 -1.41461766e+00
-7.33801946e-02 -1.08366156e+00 -1.12920143e-01 9.17325318e-01
5.91461420e-01 -6.98041320e-01 8.61053288e-01 3.09339434e-01
2.15156097e-02 -1.06965864e+00 -8.35675478e-01 -7.95296192e-01
1.74659416e-01 -2.33927101e-01 4.20509964e-01 7.60187864e-01
3.11679572e-01 2.99229711e-01 8.16819072e-02 7.14615509e-02
1.20308980e-01 7.36321062e-02 8.28542829e-01 -1.05931485e+00
-6.67683423e-01 -7.73178339e-01 -7.01932847e-01 -4.29810047e-01
-2.65161425e-01 -5.25002897e-01 1.23779446e-01 -1.12545919e+00
-2.86139458e-01 -3.19711775e-01 -2.98516471e-02 2.07116202e-01
-1.80521477e-02 -2.60549597e-02 -1.20419838e-01 1.04799256e-01
-1.13533832e-01 3.76642734e-01 1.35962713e+00 1.34613952e-02
-5.84489703e-01 3.23513672e-02 -1.90734699e-01 1.70634717e-01
9.43103969e-01 -5.55851460e-01 -3.19498867e-01 3.23805749e-01
5.74125230e-01 6.02249764e-02 3.52695763e-01 -1.51666570e+00
3.49240661e-01 -1.75201491e-01 4.96854931e-01 3.96150909e-02
8.70389044e-02 -7.29749203e-01 7.56935954e-01 8.96518886e-01
-4.24057156e-01 -3.01767915e-01 5.36352873e-01 4.78421628e-01
4.57942523e-02 -1.98302157e-02 8.06062043e-01 -1.37280554e-01
-1.63525209e-01 -1.66505888e-01 -9.55426276e-01 -3.99853617e-01
1.52421582e+00 -3.49539727e-01 -8.68073031e-02 -1.31791875e-01
-9.47630644e-01 1.26893222e-01 4.76692140e-01 1.71056151e-01
1.36601076e-01 -6.83481872e-01 -4.34484154e-01 3.16949457e-01
-2.14686796e-01 -8.49312264e-03 3.43623720e-02 7.59865105e-01
-6.48075223e-01 6.15755022e-01 -3.48873645e-01 -7.97581255e-01
-9.75387335e-01 8.90058041e-01 6.98818326e-01 -3.12588811e-01
-2.79581875e-01 5.60227454e-01 -2.08660260e-01 -2.97591418e-01
-3.34997416e-01 -5.20188212e-01 -1.39226601e-01 -2.85805255e-01
2.22285599e-01 -1.12527855e-01 8.27490985e-02 1.69881597e-01
-2.01263130e-01 8.06283653e-01 2.50262469e-01 -1.16809830e-01
1.69831681e+00 2.85853207e-01 1.34697959e-01 1.05564439e+00
8.34347665e-01 -2.52799928e-01 -1.33818901e+00 3.83966625e-01
1.04851551e-01 1.75466985e-01 -2.09510148e-01 -9.29762542e-01
-8.51335645e-01 7.91023374e-01 5.84357381e-01 4.62150604e-01
1.04818368e+00 4.19427715e-02 5.04279613e-01 3.94496083e-01
5.04021168e-01 -1.09198749e+00 8.84728283e-02 1.86670765e-01
6.65375233e-01 -7.78926492e-01 -1.13184571e-01 -1.20305412e-01
-9.90096033e-02 1.62699318e+00 1.42456338e-01 -2.22367972e-01
6.88689113e-01 6.71223342e-01 -5.16572073e-02 -2.70209402e-01
-9.86774206e-01 8.30924511e-03 -1.68470800e-01 5.87172151e-01
5.04473805e-01 -5.49169630e-02 -2.62023032e-01 2.15394616e-01
-2.50227600e-01 7.01344967e-01 7.13689387e-01 8.78356516e-01
-3.51513863e-01 -1.34015584e+00 -4.03071165e-01 2.03319654e-01
-2.14389518e-01 5.84291220e-01 -2.48520076e-01 8.58907402e-01
2.93640286e-01 4.97954041e-01 -3.25388648e-02 -3.63107383e-01
3.41724604e-01 3.65501910e-01 6.21430695e-01 -5.31700850e-01
-6.44311547e-01 -8.60899389e-02 -3.55414212e-01 -4.92536843e-01
-4.49615330e-01 -7.98613548e-01 -1.45161343e+00 -8.73349831e-02
-7.64794573e-02 1.81607783e-01 9.90901232e-01 8.91662419e-01
6.54174149e-01 1.05171871e+00 3.43699425e-01 -1.00828493e+00
-5.72633706e-02 -6.51542068e-01 -2.04199180e-01 1.51169166e-01
3.09897810e-02 -5.44745922e-01 -4.55678999e-01 4.56501842e-01] | [7.821029186248779, 3.070791721343994] |
468b8497-bfb2-46f0-b48e-eafd68db3b43 | corpus-creation-and-analysis-for-named-entity | null | null | https://aclanthology.org/P19-2025 | https://aclanthology.org/P19-2025.pdf | Corpus Creation and Analysis for Named Entity Recognition in Telugu-English Code-Mixed Social Media Data | Named Entity Recognition(NER) is one of the important tasks in Natural Language Processing(NLP) and also is a subtask of Information Extraction. In this paper we present our work on NER in Telugu-English code-mixed social media data. Code-Mixing, a progeny of multilingualism is a way in which multilingual people express themselves on social media by using linguistics units from different languages within a sentence or speech context. Entity Extraction from social media data such as tweets(twitter) is in general difficult due to its informal nature, code-mixed data further complicates the problem due to its informal, unstructured and incomplete information. We present a Telugu-English code-mixed corpus with the corresponding named entity tags. The named entities used to tag data are Person({`}Per{'}), Organization({`}Org{'}) and Location({`}Loc{'}). We experimented with the machine learning models Conditional Random Fields(CRFs), Decision Trees and BiLSTMs on our corpus which resulted in a F1-score of 0.96, 0.94 and 0.95 respectively. | ['Vamshi Krishna Srirangam', 'Manish Shrivastava', 'Appidi Abhinav Reddy', 'Vinay Singh'] | 2019-07-01 | null | null | null | acl-2019-7 | ['entity-extraction'] | ['natural-language-processing'] | [-4.80156451e-01 1.58994600e-01 1.15757458e-01 -4.05582935e-01
-7.97500253e-01 -7.71468759e-01 7.49067545e-01 6.19713366e-01
-7.96086133e-01 1.25112188e+00 5.08538961e-01 -5.54912090e-01
3.61059248e-01 -6.71016932e-01 -4.31898355e-01 -2.76400924e-01
-5.07154278e-02 3.23644876e-01 2.45676652e-01 -2.71099478e-01
3.59219581e-01 3.46764863e-01 -1.07906568e+00 2.17070192e-01
9.49607909e-01 3.35341901e-01 4.70248699e-01 7.13166177e-01
-8.43169689e-01 1.32432544e+00 -6.35651171e-01 -7.10218310e-01
-1.90234661e-01 -2.17272732e-02 -1.35579371e+00 -1.89931050e-01
7.20599899e-03 1.76518738e-01 3.45001295e-02 1.32455730e+00
2.29528800e-01 2.56828845e-01 7.54439652e-01 -9.50100422e-01
-5.87738574e-01 7.82747686e-01 -5.58288813e-01 1.88385606e-01
4.62009102e-01 -4.71962214e-01 8.33865881e-01 -8.54620457e-01
9.17577267e-01 8.78898680e-01 8.40244174e-01 3.83648008e-01
-8.29023182e-01 -6.53345287e-01 -2.75662273e-01 -3.18009883e-01
-1.64811242e+00 -3.89234602e-01 1.93197429e-01 -8.64056766e-01
1.08245409e+00 -9.96647179e-02 -3.40015744e-03 6.48743033e-01
3.20824206e-01 3.98662716e-01 1.25785553e+00 -7.34479010e-01
-1.24677017e-01 7.10698485e-01 3.55642170e-01 7.87345707e-01
-2.71851290e-02 -4.77194846e-01 -2.10321788e-02 -2.00039938e-01
4.87777203e-01 3.31825502e-02 2.80232728e-01 5.03702939e-01
-1.32187331e+00 9.53135967e-01 1.56214356e-01 1.00954175e+00
-3.73857766e-01 1.24694509e-02 4.85327065e-01 2.47348458e-01
3.85280848e-01 1.41196534e-01 -8.83643746e-01 -5.91075599e-01
-7.20254660e-01 -1.75394982e-01 1.10263813e+00 1.40360522e+00
9.79515791e-01 -2.02438906e-01 7.40647987e-02 1.28285873e+00
5.14281809e-01 3.98340672e-01 5.16854584e-01 -5.60663700e-01
6.68363452e-01 6.18487954e-01 3.47880423e-01 -9.41156924e-01
-4.32879299e-01 -2.04709381e-01 -6.61608279e-01 -3.56743753e-01
2.03151003e-01 -9.23147023e-01 -8.81332040e-01 1.62434065e+00
9.66938585e-02 -2.53755420e-01 3.73682052e-01 1.66959837e-01
1.04069245e+00 8.92076671e-01 6.98104858e-01 -1.20979674e-01
1.70807958e+00 -8.80329788e-01 -8.64968181e-01 -5.44522889e-02
9.21110451e-01 -1.35513568e+00 3.89326274e-01 -2.14343891e-01
-6.42820060e-01 -3.30201447e-01 -4.59256858e-01 -9.83777866e-02
-1.03772366e+00 3.65579516e-01 4.37458396e-01 8.02204847e-01
-9.18690801e-01 3.50453347e-01 -6.85987055e-01 -9.22321141e-01
1.09956324e-01 4.39432651e-01 -7.78056622e-01 2.68583000e-01
-1.11292648e+00 9.34455812e-01 5.77977300e-01 -2.36811697e-01
-3.37871522e-01 -4.67645526e-02 -1.00034857e+00 -2.15153456e-01
1.31657004e-01 -3.11654457e-03 1.27901995e+00 -8.24994028e-01
-9.54497993e-01 1.33807075e+00 -3.03395480e-01 -2.75292218e-01
1.63489178e-01 1.27800070e-02 -1.09924781e+00 -3.23190451e-01
7.94133544e-01 5.53733647e-01 8.52957070e-02 -1.02776229e+00
-1.19963729e+00 -2.60129154e-01 -1.66354459e-02 -1.44344136e-01
-4.13203761e-02 9.49268401e-01 6.90388978e-02 -6.73486233e-01
-1.14617348e-01 -1.02279782e+00 -2.58284062e-01 -9.10849214e-01
-5.94880223e-01 -5.13075888e-01 4.86355007e-01 -1.11540711e+00
1.57653832e+00 -2.01713443e+00 -4.26671714e-01 -5.72994240e-02
1.50316358e-01 2.19596446e-01 5.70952833e-01 9.81547832e-01
-4.28334624e-03 5.67027271e-01 -2.40109265e-01 -2.80382931e-01
3.64770517e-02 4.55027193e-01 2.08339036e-01 3.33693177e-01
1.58737913e-01 4.44306344e-01 -8.80961597e-01 -9.41345155e-01
-2.06659600e-01 2.93950647e-01 -8.62367153e-02 1.12437338e-01
1.72644988e-01 4.47920918e-01 -6.28276885e-01 5.22222817e-01
5.01963437e-01 4.56675664e-02 5.38595319e-02 2.78568864e-01
-1.01006305e+00 2.40811497e-01 -1.15847135e+00 1.52133811e+00
-6.62447989e-01 9.12390709e-01 1.94125921e-01 -6.01071298e-01
1.00797427e+00 7.77552605e-01 2.40032047e-01 -1.25750527e-01
2.78382540e-01 5.48165560e-01 -1.82331055e-01 -8.41549635e-01
8.28821361e-01 -1.89782709e-01 -6.09822512e-01 2.72375852e-01
2.44913995e-01 4.36314136e-01 5.33094525e-01 1.29801527e-01
1.07039142e+00 1.71158873e-02 7.76808798e-01 -5.15265644e-01
7.54860640e-01 2.19864890e-01 6.35314465e-01 4.26136822e-01
-4.41638887e-01 2.22568288e-01 3.24259609e-01 -2.21793458e-01
-1.30209517e+00 -6.64910436e-01 -4.66193765e-01 1.26144862e+00
-4.00087863e-01 -2.16350079e-01 -6.09587550e-01 -6.90307736e-01
-5.24912000e-01 7.82920361e-01 -3.88811231e-01 6.75547242e-01
-5.63505530e-01 -6.52352929e-01 7.60027826e-01 1.53761476e-01
7.19836831e-01 -1.26668596e+00 -2.71591339e-02 6.15276814e-01
-5.00417292e-01 -1.26602578e+00 -4.29409117e-01 4.48672414e-01
-3.89244944e-01 -7.12403655e-01 -7.16781735e-01 -1.41517675e+00
7.52894402e-01 -2.29429036e-01 1.27138233e+00 -3.24505717e-02
3.92124355e-02 -6.75863400e-02 -6.42826140e-01 -6.21250033e-01
-6.56528533e-01 3.30264896e-01 -2.16967285e-01 -6.51145950e-02
8.02057445e-01 -5.50537646e-01 -1.78437512e-02 2.10345536e-02
-7.76026905e-01 -3.22797187e-02 6.10749424e-01 6.55009270e-01
-5.64683694e-03 -9.73791108e-02 5.30534208e-01 -1.40856040e+00
6.22825146e-01 -9.51659977e-01 -3.00922722e-01 2.11588338e-01
-1.25285402e-01 8.41679797e-02 4.70155507e-01 -2.71491278e-02
-1.25752866e+00 2.39981055e-01 -5.42907417e-01 5.99837184e-01
-8.44049513e-01 6.78291619e-01 4.16222028e-02 2.29286656e-01
6.05780780e-01 1.06353186e-01 -6.98841214e-01 -5.47914326e-01
-5.29843569e-02 1.35634267e+00 3.83114457e-01 -5.68151295e-01
5.03230989e-01 6.51995912e-02 -4.41298157e-01 -9.02712464e-01
-7.29439259e-01 -1.00288248e+00 -9.95649517e-01 -6.71034902e-02
1.51012325e+00 -1.06760776e+00 -5.30370116e-01 3.30029309e-01
-1.46177697e+00 2.24204242e-01 2.74541140e-01 7.10174203e-01
-1.36827817e-02 2.45364249e-01 -1.04238307e+00 -1.00824153e+00
-2.08952844e-01 -7.78255701e-01 7.06165850e-01 4.99380738e-01
-2.57949710e-01 -1.23368466e+00 2.87075847e-01 3.20421666e-01
2.76380420e-01 5.56767285e-01 8.38304162e-01 -1.20963931e+00
7.86830336e-02 -1.72805116e-01 -4.56384629e-01 1.73389003e-01
2.25055367e-01 2.28089109e-01 -7.65737236e-01 1.18284479e-01
-3.40327978e-01 -9.51195732e-02 1.22105248e-01 -1.46261230e-01
1.32596716e-01 -4.96319294e-01 -3.96777511e-01 -2.26602957e-01
1.75509059e+00 5.20792425e-01 4.28097367e-01 4.46982592e-01
7.54797399e-01 6.80319011e-01 3.31284851e-01 5.10960221e-01
7.20015466e-01 2.89948881e-01 -1.61965594e-01 -6.06066696e-02
2.51063734e-01 -2.22723275e-01 3.62938851e-01 1.28396320e+00
-1.92626312e-01 6.92171510e-03 -1.42226732e+00 9.53889072e-01
-1.59593403e+00 -1.02039766e+00 -8.22643876e-01 1.90366662e+00
1.15767622e+00 -1.25053406e-01 2.32490972e-02 -2.74686158e-01
1.30273771e+00 -2.18749970e-01 4.61008161e-01 -7.15109289e-01
-8.87065828e-02 1.80273131e-01 8.29492629e-01 4.71641481e-01
-1.30626726e+00 1.04751956e+00 4.37929296e+00 8.12480688e-01
-7.63193011e-01 5.13606787e-01 3.07522953e-01 1.01406503e+00
1.81584075e-01 4.30824280e-01 -1.27221191e+00 6.42005801e-01
1.27259302e+00 -7.46855065e-02 1.55489489e-01 7.63120949e-01
2.37075821e-01 -3.07973772e-01 -7.27215946e-01 7.59381354e-01
-1.13160998e-01 -9.99040008e-01 -4.42778140e-01 1.96509406e-01
1.08156788e+00 4.93862391e-01 -6.68906450e-01 6.57613814e-01
9.74570990e-01 -8.21080387e-01 7.67556071e-01 3.65120709e-01
6.25684261e-01 -7.65961707e-01 1.03986669e+00 6.61188304e-01
-1.44746101e+00 1.06474757e-01 -3.23793024e-01 2.79836189e-02
3.15253019e-01 5.05540431e-01 -9.23024178e-01 5.35767496e-01
7.99735487e-01 4.37157243e-01 -4.41032946e-01 9.08878148e-01
-2.28584796e-01 5.23411036e-01 -2.57403970e-01 -5.03674448e-01
5.95635355e-01 -1.77111194e-01 3.31733882e-01 1.90608454e+00
3.03580523e-01 1.18013006e-02 3.47046733e-01 5.62544286e-01
-4.47514743e-01 6.26831234e-01 -6.86227322e-01 -1.97472587e-01
4.85473722e-01 1.35757458e+00 -9.77828681e-01 -4.63518769e-01
-6.90817475e-01 7.23612964e-01 3.43332887e-01 2.37368688e-01
-5.81138372e-01 -1.19312763e+00 -7.06706643e-02 1.28719032e-01
1.57018587e-01 -3.53786141e-01 1.89442933e-01 -1.08027434e+00
-1.09955572e-01 -3.84812504e-01 3.20197761e-01 -5.92168212e-01
-1.35334384e+00 1.01859224e+00 -2.60437459e-01 -1.13382161e+00
-1.74148425e-01 -3.55905801e-01 -3.71522039e-01 1.00630748e+00
-1.28920221e+00 -1.25948632e+00 -3.30349170e-02 6.22107744e-01
4.54261005e-01 -4.08342689e-01 9.95507777e-01 9.08975124e-01
-4.82082874e-01 1.78599730e-01 3.78339380e-01 1.07953858e+00
6.44710660e-01 -1.42742968e+00 3.91711891e-02 7.83737183e-01
4.71417643e-02 8.68420064e-01 6.78898633e-01 -7.27652550e-01
-6.79612339e-01 -1.02462924e+00 1.94365740e+00 -4.98578876e-01
1.12366760e+00 -3.51956725e-01 -4.36380982e-01 9.54438150e-01
7.82140374e-01 -1.62301332e-01 1.03466761e+00 6.01311289e-02
9.91120469e-03 2.86417633e-01 -1.37177229e+00 3.30355227e-01
5.56149542e-01 -6.38182759e-01 -8.55199516e-01 5.85180461e-01
5.29639900e-01 -6.53028414e-02 -1.22733879e+00 -2.05300465e-01
1.49929911e-01 -8.00737023e-01 1.39534444e-01 -4.80571449e-01
2.52539158e-01 -3.73222053e-01 -1.65709406e-01 -9.86266851e-01
-1.76666632e-01 -5.92407286e-01 8.21812630e-01 2.03236794e+00
9.33134437e-01 -6.25856161e-01 3.00414592e-01 7.07142591e-01
-1.88308254e-01 5.76620102e-02 -8.59789670e-01 -3.56493473e-01
1.58522978e-01 -3.37380856e-01 3.78090292e-01 1.40572941e+00
1.91819683e-01 5.75969517e-01 -2.91565448e-01 4.24297154e-02
3.33798558e-01 -4.09330070e-01 4.05965716e-01 -1.49309218e+00
1.37828767e-01 3.15920636e-03 -3.23726624e-01 -7.39021242e-01
2.28931263e-01 -8.23301613e-01 2.46318668e-01 -1.67580462e+00
2.83296734e-01 -7.65990317e-01 2.18581125e-01 6.71491325e-01
2.95182884e-01 -2.68126652e-02 -1.94947217e-02 3.16216558e-01
-7.01517105e-01 -2.82108542e-02 7.43905723e-01 1.66185170e-01
-1.02639318e-01 5.29395193e-02 -4.19277728e-01 7.18639195e-01
7.41869748e-01 -1.05534434e+00 3.11587363e-01 -2.99825221e-01
6.68546855e-01 3.08783174e-01 -2.46878192e-01 -7.61735976e-01
5.01243651e-01 -5.59792519e-02 3.74721512e-02 -3.95221144e-01
-1.97605565e-01 -9.37326968e-01 6.34426698e-02 2.87190676e-01
-1.83367670e-01 5.35601974e-01 -2.97745294e-03 2.11265013e-01
-5.51894307e-01 -8.26264620e-01 7.21555769e-01 -5.71179748e-01
-6.55772507e-01 8.44731778e-02 -9.38134849e-01 2.34312773e-01
1.04107118e+00 -5.46226650e-02 -6.78390265e-02 -1.03362873e-01
-9.67393816e-01 1.49406075e-01 2.97693491e-01 3.09898555e-01
-1.17167801e-01 -1.23024607e+00 -8.42066109e-01 -2.04511404e-01
2.54540533e-01 -1.76641479e-01 -3.17932875e-03 7.26188421e-01
-9.37475145e-01 5.36007881e-01 -3.21565837e-01 -1.05823651e-01
-1.23227108e+00 1.03203170e-01 9.01718289e-02 -4.61905926e-01
2.62712184e-02 6.37917161e-01 -3.81246418e-01 -8.75130892e-01
-9.30813029e-02 -3.86757613e-03 -9.47257102e-01 1.71776831e-01
1.16818689e-01 4.12543863e-01 -5.39692678e-02 -1.48916328e+00
-3.82583022e-01 3.85090590e-01 -2.44379919e-02 -1.74577862e-01
1.34715581e+00 -4.32146817e-01 -7.34738529e-01 7.88323998e-01
1.49184453e+00 5.93759537e-01 -1.95958704e-01 -2.10976839e-01
7.03783333e-01 -1.72094554e-01 -3.88447046e-01 -5.00588179e-01
-6.98287606e-01 5.48786819e-01 4.10550356e-01 4.87824142e-01
5.49549699e-01 1.78215668e-01 7.22779453e-01 3.39510441e-01
5.41439712e-01 -1.55655932e+00 -6.96699262e-01 1.02568614e+00
3.28454912e-01 -1.46090376e+00 -4.74785745e-01 -2.69880474e-01
-6.82173967e-01 1.13052547e+00 3.40584695e-01 9.22492519e-02
1.11465383e+00 3.35789770e-01 1.74743101e-01 -1.50123850e-01
-3.74398082e-01 -3.63980591e-01 -1.82443276e-01 4.70879197e-01
1.15822423e+00 2.63449728e-01 -8.70915651e-01 6.68408155e-01
-3.74556929e-01 -1.25130624e-01 8.22199225e-01 1.29557180e+00
-6.10483944e-01 -1.33208001e+00 -3.89739752e-01 4.56575662e-01
-1.48555779e+00 -2.92647868e-01 -1.26537839e-02 7.26398230e-01
6.27224684e-01 1.41752160e+00 -1.27181023e-01 -1.38930470e-01
7.92417303e-02 1.30333602e-01 -2.73213238e-01 -9.49130952e-01
-1.10527980e+00 -5.90565689e-02 6.08417273e-01 2.50535220e-01
-7.69749880e-01 -7.31321931e-01 -1.85810626e+00 -4.83874053e-01
-3.25144559e-01 8.80965412e-01 9.70175087e-01 1.11698389e+00
9.23268050e-02 9.37354565e-02 6.20837092e-01 -1.45404860e-01
7.20938966e-02 -1.31772959e+00 -7.98778474e-01 4.73701000e-01
3.66194397e-02 -3.72051895e-01 -2.52093405e-01 5.14209032e-01] | [9.802860260009766, 9.89149284362793] |
fc41f0ba-53ad-44ba-bee2-2b61649478ae | dehazenet-an-end-to-end-system-for-single | 1601.07661 | null | http://arxiv.org/abs/1601.07661v2 | http://arxiv.org/pdf/1601.07661v2.pdf | DehazeNet: An End-to-End System for Single Image Haze Removal | Single image haze removal is a challenging ill-posed problem. Existing
methods use various constraints/priors to get plausible dehazing solutions. The
key to achieve haze removal is to estimate a medium transmission map for an
input hazy image. In this paper, we propose a trainable end-to-end system
called DehazeNet, for medium transmission estimation. DehazeNet takes a hazy
image as input, and outputs its medium transmission map that is subsequently
used to recover a haze-free image via atmospheric scattering model. DehazeNet
adopts Convolutional Neural Networks (CNN) based deep architecture, whose
layers are specially designed to embody the established assumptions/priors in
image dehazing. Specifically, layers of Maxout units are used for feature
extraction, which can generate almost all haze-relevant features. We also
propose a novel nonlinear activation function in DehazeNet, called Bilateral
Rectified Linear Unit (BReLU), which is able to improve the quality of
recovered haze-free image. We establish connections between components of the
proposed DehazeNet and those used in existing methods. Experiments on benchmark
images show that DehazeNet achieves superior performance over existing methods,
yet keeps efficient and easy to use. | ['DaCheng Tao', 'Chunmei Qing', 'Xiangmin Xu', 'Kui Jia', 'Bolun Cai'] | 2016-01-28 | null | null | null | null | ['single-image-haze-removal'] | ['computer-vision'] | [ 2.83887655e-01 -2.93997109e-01 6.85624003e-01 -3.30509663e-01
-4.41195935e-01 5.86525188e-04 3.07697892e-01 -4.04841214e-01
-2.88426816e-01 4.56042886e-01 2.90907267e-02 -1.94476932e-01
-1.32067099e-01 -1.07312107e+00 -8.53251338e-01 -1.19804442e+00
-1.30705044e-01 -2.29202986e-01 3.34503382e-01 -5.18117607e-01
4.88906503e-02 2.46028215e-01 -1.40594399e+00 3.11626226e-01
1.25226653e+00 1.14059544e+00 4.09176737e-01 8.92550468e-01
2.55855173e-01 1.05523944e+00 -7.60258257e-01 -1.38238415e-01
5.31837642e-01 -2.52358466e-01 -3.36534709e-01 8.39658454e-02
7.36790001e-01 -8.19763958e-01 -6.40920699e-01 1.20293999e+00
4.36807722e-01 3.38199466e-01 8.84158671e-01 -1.05860531e+00
-1.04566729e+00 2.12363929e-01 -4.94551003e-01 3.47830832e-01
-3.77241492e-01 3.21426332e-01 2.92605132e-01 -1.11648345e+00
1.55886738e-02 1.11688328e+00 4.82892781e-01 3.81205887e-01
-5.63104331e-01 -1.02281201e+00 1.26569912e-01 2.29534864e-01
-1.77553308e+00 -5.45296133e-01 5.88171065e-01 -2.60631800e-01
7.02418447e-01 2.44554922e-01 4.75304097e-01 3.72492284e-01
7.01973677e-01 7.28272855e-01 8.33189368e-01 -2.92024016e-01
-6.56592324e-02 8.13819617e-02 -9.05942023e-02 6.54095292e-01
3.42678815e-01 3.49696241e-02 -3.14570338e-01 3.48541170e-01
7.41141438e-01 3.11299473e-01 -6.55373633e-01 3.03389549e-01
-9.77326632e-01 6.46901965e-01 1.01843905e+00 -8.44151825e-02
-4.46922213e-01 2.13932067e-01 -4.07087743e-01 5.30066311e-01
8.45498502e-01 3.66595566e-01 -3.38812172e-02 6.83005571e-01
-8.82593393e-01 3.81431639e-01 3.36772978e-01 1.10459661e+00
1.04337561e+00 4.42849725e-01 -2.67875075e-01 7.13111401e-01
5.85103691e-01 8.41176033e-01 1.40718520e-01 -7.48673439e-01
4.74701107e-01 1.84162617e-01 3.15033793e-01 -9.36182261e-01
-7.53101287e-03 -4.05094653e-01 -1.36456251e+00 6.93583727e-01
-2.62196034e-01 -2.92360187e-01 -1.40753603e+00 1.08953977e+00
2.76680529e-01 6.07338369e-01 3.44903469e-01 1.24275124e+00
1.06052160e+00 1.51021409e+00 -3.16162348e-01 1.69476159e-02
9.28208470e-01 -1.27822590e+00 -9.60154355e-01 -4.04276967e-01
1.05644643e-01 -7.97632992e-01 5.29680490e-01 5.87930024e-01
-1.14484584e+00 -4.21133757e-01 -1.40803564e+00 -3.03138018e-01
-6.42162502e-01 -2.06319451e-01 2.02861324e-01 3.80641669e-01
-1.29760516e+00 2.51671016e-01 -6.76374793e-01 4.28432733e-01
4.12737936e-01 3.69694263e-01 -5.42173088e-02 -4.00260538e-01
-1.54809499e+00 1.00076354e+00 5.18637836e-01 1.01406205e+00
-1.48178816e+00 -7.04072833e-01 -1.11709249e+00 1.34905696e-01
2.68617839e-01 -8.62177491e-01 9.75516498e-01 -9.01961088e-01
-1.55877876e+00 3.20770860e-01 1.66916251e-01 -4.86935288e-01
2.27233320e-01 -6.53526604e-01 -5.52900910e-01 2.67176777e-01
-3.00448388e-01 6.33672059e-01 1.57207167e+00 -1.54957664e+00
-6.43635392e-01 2.16416493e-01 5.96711934e-02 4.11850274e-01
-3.11993569e-01 -2.91672163e-02 -6.20550275e-01 -7.34237671e-01
-9.04868692e-02 -5.30405104e-01 -2.62877882e-01 2.41249532e-01
-5.13928711e-01 1.42895430e-01 9.94020224e-01 -8.32778096e-01
1.10109007e+00 -2.08553004e+00 1.27808616e-01 1.13208540e-01
7.24332273e-01 5.90228915e-01 -2.70859361e-01 2.09748641e-01
5.77400587e-02 -5.51035814e-02 -6.33778870e-01 -4.56808001e-01
-2.55593419e-01 -1.22352786e-01 -2.46439382e-01 7.82829344e-01
5.54337621e-01 6.65543199e-01 -6.38776779e-01 -3.22212726e-01
5.83650827e-01 9.87964332e-01 -3.81978810e-01 7.41447031e-01
-2.56489694e-01 2.13889748e-01 -3.03597867e-01 5.15240312e-01
1.51518285e+00 1.16938449e-01 -6.97950780e-01 -9.06519070e-02
-4.02495027e-01 -1.57323673e-01 -9.44687307e-01 1.04567254e+00
-7.82150865e-01 8.59784842e-01 3.86521280e-01 -5.70463955e-01
8.37903738e-01 2.53768086e-01 -1.33522928e-01 -6.71582580e-01
3.25941920e-01 1.23879306e-01 -2.63713032e-01 -7.97765791e-01
6.01955771e-01 -2.55070746e-01 5.47703087e-01 -5.26257902e-02
-1.49643287e-01 -6.63397193e-01 -3.32601368e-01 8.78037810e-02
6.37732685e-01 -2.07906559e-01 -1.21733911e-01 -3.12184870e-01
6.97308719e-01 -3.05992395e-01 3.44931841e-01 6.54374421e-01
4.61441837e-02 1.00344753e+00 -1.39432833e-01 -4.90192145e-01
-8.11022818e-01 -1.10405004e+00 -5.17294593e-02 7.27314949e-01
6.33506000e-01 1.98107213e-02 -7.82035053e-01 -1.22288167e-01
-4.07400221e-01 6.27625763e-01 -7.35502601e-01 -5.84021747e-01
-6.84532821e-01 -8.30746949e-01 1.67919755e-01 -7.51191974e-02
1.12709618e+00 -1.02547085e+00 -3.20233077e-01 1.06875442e-01
-2.77719885e-01 -1.11770391e+00 -4.80080038e-01 -2.96365861e-02
-4.79643106e-01 -6.56484425e-01 -9.37121451e-01 -9.75468814e-01
5.49035370e-01 9.83761668e-01 9.00040507e-01 5.13305366e-01
-1.21978290e-01 -2.28594646e-01 -4.02122080e-01 -9.93190110e-01
-2.19603345e-01 -1.72295436e-01 -1.97637677e-01 5.25711298e-01
7.99177960e-02 -5.32852650e-01 -1.11727285e+00 2.88611740e-01
-1.48646986e+00 1.99182227e-01 7.95202553e-01 5.93779087e-01
3.30019623e-01 7.18858004e-01 7.39968568e-02 -7.21782923e-01
3.27257782e-01 -8.88025045e-01 -7.47787774e-01 6.29330927e-04
-4.97520357e-01 -2.63694733e-01 5.84564805e-01 -2.27246255e-01
-1.41498864e+00 -4.04991388e-01 -2.06981808e-01 -8.92416835e-01
-4.96196486e-02 5.28402746e-01 -2.55110979e-01 -4.60765749e-01
5.65961003e-01 5.26532829e-01 -1.80621311e-01 -2.95169294e-01
3.39543730e-01 9.18921709e-01 8.28691900e-01 -6.98904023e-02
1.49650943e+00 7.24382997e-01 -2.31092021e-01 -1.14359844e+00
-9.99071658e-01 -5.41744351e-01 -4.25350338e-01 -2.94928253e-01
1.15526330e+00 -1.40047407e+00 -5.26078105e-01 1.01209295e+00
-1.12414420e+00 -6.18301332e-01 3.49975348e-01 4.85507220e-01
-9.12080482e-02 3.48563008e-02 -6.53379917e-01 -9.08726513e-01
-7.00790644e-01 -1.06574178e+00 1.15945411e+00 3.35858524e-01
8.62344682e-01 -9.87438321e-01 -1.71523362e-01 3.72289777e-01
8.77542436e-01 3.66609663e-01 5.97187221e-01 2.49286130e-01
-1.33592653e+00 -4.68919352e-02 -5.71019113e-01 8.18071485e-01
2.97811061e-01 1.68900061e-02 -1.22816527e+00 -4.82384056e-01
1.89004824e-01 -7.37089217e-02 1.48348546e+00 5.08703887e-01
1.20018923e+00 -5.56782424e-01 -2.48130839e-02 1.29527700e+00
1.59518766e+00 3.07371020e-02 1.07956898e+00 6.08162999e-01
9.59657788e-01 5.17977118e-01 5.92199147e-01 3.24978270e-02
2.22222060e-01 3.25260073e-01 1.07317746e+00 -5.89760363e-01
-3.22455108e-01 4.10863936e-01 4.68478501e-01 8.68071079e-01
-2.33750746e-01 -8.45753133e-01 -4.88996655e-01 5.64814806e-01
-1.60726595e+00 -6.84068918e-01 -1.67808414e-01 1.84426618e+00
6.45047843e-01 3.87169532e-02 -5.27778924e-01 -1.15979485e-01
5.59906006e-01 3.96834999e-01 -2.85989583e-01 -2.06052333e-01
-1.59776464e-01 1.80373102e-01 5.54809928e-01 7.64351130e-01
-1.28588462e+00 1.07456553e+00 5.69641304e+00 6.89807236e-01
-1.18369782e+00 1.14229461e-02 3.40644658e-01 -2.62729526e-02
-1.83081359e-01 -4.78390962e-01 -4.69860882e-01 4.52834874e-01
9.01351571e-01 5.21958470e-02 5.63707054e-01 1.82977304e-01
4.61124569e-01 3.25998710e-03 -6.63125336e-01 8.31915915e-01
1.88347191e-01 -1.28092825e+00 2.93359011e-01 -1.73229977e-01
1.01379704e+00 1.91318408e-01 5.12023687e-01 2.49304235e-01
3.60195488e-01 -1.27050638e+00 1.01195920e+00 6.54389441e-01
7.86345005e-01 -7.79850841e-01 1.10121691e+00 3.26878756e-01
-1.07563388e+00 9.06957164e-02 -8.23017299e-01 6.66393265e-02
-7.74716865e-03 7.31350064e-01 -7.34400034e-01 8.28325987e-01
9.80945528e-01 7.55284011e-01 -3.80919695e-01 1.27470517e+00
-4.83336210e-01 7.46677756e-01 -2.27473542e-01 5.31333983e-01
5.68106651e-01 -3.09790224e-01 4.55453157e-01 1.38593757e+00
4.16374415e-01 4.13533062e-01 -2.76242848e-02 8.98645401e-01
-2.28364199e-01 -3.85706216e-01 -4.79848504e-01 4.68897372e-01
1.29391640e-01 1.21207654e+00 -3.15296084e-01 -2.69436687e-01
-4.17012125e-01 1.06281650e+00 -3.38220894e-02 7.38404334e-01
-9.16669965e-01 -8.13627839e-01 8.92618001e-01 -7.06489310e-02
3.70744824e-01 -8.85897055e-02 2.79092282e-01 -1.06512928e+00
-1.14565730e-01 -8.49020541e-01 1.13845333e-01 -1.23457444e+00
-1.32702398e+00 7.72547543e-01 1.53284937e-01 -1.31971729e+00
4.15456265e-01 -7.60163963e-01 -1.05622816e+00 1.32780063e+00
-2.50465608e+00 -1.32779503e+00 -1.07819998e+00 8.28377903e-01
8.10383379e-01 2.39183486e-01 2.30974272e-01 4.50172037e-01
-7.33576894e-01 2.47223437e-01 2.32419670e-01 5.59069030e-02
7.90416479e-01 -1.16406155e+00 4.72923338e-01 1.31886876e+00
-4.13689733e-01 6.54362440e-01 9.13896799e-01 -5.24755538e-01
-1.25923848e+00 -1.88720667e+00 1.70335159e-01 -3.47784668e-01
3.41015279e-01 -4.30688769e-01 -1.30614221e+00 8.18165243e-01
6.46722853e-01 2.92044818e-01 1.07055016e-01 -6.77075326e-01
-3.43156219e-01 -2.78603256e-01 -9.70823884e-01 5.29115200e-01
5.35961926e-01 -3.26620549e-01 -4.97215688e-01 3.87544662e-01
1.21804667e+00 -6.52311683e-01 -5.19639671e-01 2.99721211e-01
2.45675996e-01 -9.32794869e-01 1.09437156e+00 -1.17822252e-01
6.89317703e-01 -6.42533720e-01 -4.11521345e-02 -1.64742446e+00
-4.76064473e-01 -8.92061949e-01 -1.91722542e-01 8.27179790e-01
3.60222578e-01 -6.72773719e-01 4.51230973e-01 1.05648085e-01
-6.10423028e-01 -6.81118965e-01 -3.81669819e-01 -6.87223494e-01
3.14162374e-01 -2.04171669e-02 8.81013155e-01 9.11886513e-01
-8.75295281e-01 8.91822726e-02 -9.35015678e-01 1.16855597e+00
9.42627430e-01 -2.86261942e-02 6.47458613e-01 -8.83057415e-01
2.13223130e-01 -9.44746807e-02 -9.18781236e-02 -1.03725863e+00
-1.52225308e-02 -5.37953556e-01 7.70717382e-01 -1.68854761e+00
-4.06044023e-03 -2.72944540e-01 -4.78685409e-01 3.02044898e-01
-5.59952557e-01 5.58683872e-01 7.67165050e-02 1.89953744e-01
-4.16268349e-01 9.83253062e-01 1.57561243e+00 -7.35664010e-01
-2.77355045e-01 1.19910471e-01 -5.02735674e-01 5.96383989e-01
8.41981590e-01 -4.45803076e-01 -4.53417927e-01 -1.16565049e+00
4.58929092e-02 -2.17033133e-01 4.36203510e-01 -1.08449030e+00
4.21509206e-01 -2.37869918e-01 5.80029488e-01 -5.97373903e-01
6.13647759e-01 -8.48410964e-01 5.41907968e-03 1.68928191e-01
-7.19233751e-02 -2.93794930e-01 2.60604084e-01 4.94580060e-01
-4.89930928e-01 -1.20437220e-01 9.24299240e-01 -1.24755152e-01
-7.39513993e-01 8.50981236e-01 -5.70649087e-01 -3.69484633e-01
7.80337214e-01 -1.18693113e-01 -5.85205257e-01 -5.86351514e-01
-2.31473342e-01 4.17400628e-01 2.09865391e-01 3.27973008e-01
1.33100557e+00 -9.19265032e-01 -1.15818179e+00 3.02825272e-01
1.48745194e-01 6.75629497e-01 3.87889862e-01 6.26737475e-01
-1.16762614e+00 -6.52039051e-02 3.52161154e-02 -3.28330249e-01
-1.06762981e+00 6.75014675e-01 7.07441628e-01 2.92528719e-01
-1.00448716e+00 1.14170754e+00 8.82552743e-01 -1.81929976e-01
-1.45728579e-02 -4.04779375e-01 -2.25144058e-01 -5.13227582e-01
9.67288375e-01 2.12500840e-01 2.53283322e-01 -4.95862991e-01
2.44947132e-02 5.50771654e-01 -1.06619135e-01 2.19501436e-01
1.56148195e+00 -4.50968325e-01 -4.38608587e-01 4.51391377e-02
1.16125059e+00 -7.99926966e-02 -1.54071093e+00 -3.92752469e-01
-7.13426471e-01 -6.68233037e-01 7.13137329e-01 -5.24797380e-01
-1.50351620e+00 1.12378120e+00 6.51858807e-01 4.75540832e-02
1.49052954e+00 -4.32567537e-01 1.15635073e+00 6.49188161e-01
3.49178463e-02 -2.96631336e-01 -4.03543003e-02 4.31499034e-01
9.69518006e-01 -1.24822998e+00 1.61160603e-01 -6.23541951e-01
-1.94624126e-01 9.79325056e-01 8.41648161e-01 -1.47684351e-01
8.90050530e-01 1.59691691e-01 3.89604688e-01 -5.32847404e-01
-5.45797765e-01 -8.43158439e-02 3.50552946e-01 5.30983627e-01
-1.41232312e-01 -2.09713459e-01 6.39674127e-01 1.11387424e-01
-3.53372619e-02 -2.54529774e-01 7.50723064e-01 7.72633731e-01
-9.19921160e-01 -2.50904500e-01 -7.55725741e-01 2.18029439e-01
-2.96294570e-01 -4.28137869e-01 5.59025407e-02 6.24711096e-01
4.54902023e-01 1.23593533e+00 8.00335780e-02 -2.76080042e-01
7.04791248e-02 -7.51550496e-01 2.09204242e-01 -5.60618579e-01
-3.97764921e-01 1.22943491e-01 -3.64996850e-01 -2.48224705e-01
-5.43600500e-01 1.44674316e-01 -9.19792712e-01 -4.36306715e-01
-5.69741428e-01 2.75637895e-01 3.86494279e-01 7.78961420e-01
-7.25046732e-03 9.44998384e-01 9.42116141e-01 -1.30073905e+00
7.87925869e-02 -1.09426999e+00 -7.67944455e-01 8.30510706e-02
1.31140828e+00 -6.10617220e-01 -5.55097103e-01 2.97780484e-01] | [10.939643859863281, -3.2091450691223145] |
29b7775d-18c9-44b4-8b25-911348d8286f | alp-data-augmentation-using-lexicalized-pcfgs | 2112.11916 | null | https://arxiv.org/abs/2112.11916v1 | https://arxiv.org/pdf/2112.11916v1.pdf | ALP: Data Augmentation using Lexicalized PCFGs for Few-Shot Text Classification | Data augmentation has been an important ingredient for boosting performances of learned models. Prior data augmentation methods for few-shot text classification have led to great performance boosts. However, they have not been designed to capture the intricate compositional structure of natural language. As a result, they fail to generate samples with plausible and diverse sentence structures. Motivated by this, we present the data Augmentation using Lexicalized Probabilistic context-free grammars (ALP) that generates augmented samples with diverse syntactic structures with plausible grammar. The lexicalized PCFG parse trees consider both the constituents and dependencies to produce a syntactic frame that maximizes a variety of word choices in a syntactically preservable manner without specific domain experts. Experiments on few-shot text classification tasks demonstrate that ALP enhances many state-of-the-art classification methods. As a second contribution, we delve into the train-val splitting methodologies when a data augmentation method comes into play. We argue empirically that the traditional splitting of training and validation sets is sub-optimal compared to our novel augmentation-based splitting strategies that further expand the training split with the same number of labeled data. Taken together, our contributions on the data augmentation strategies yield a strong training recipe for few-shot text classification tasks. | ['Yo-Sub Han', 'Jeong-Won Cha', 'Seong Joon Oh', 'Daecheol Woo', 'Hazel Kim'] | 2021-12-16 | null | null | null | null | ['few-shot-text-classification', 'semi-supervised-text-classification-1'] | ['natural-language-processing', 'natural-language-processing'] | [ 4.21292633e-01 4.51471925e-01 -4.62526441e-01 -6.96379364e-01
-8.25599074e-01 -2.70817757e-01 8.10363948e-01 3.15418184e-01
-3.23603123e-01 8.75021279e-01 4.99196857e-01 -4.08280581e-01
3.01771462e-01 -8.73426020e-01 -3.17932814e-01 -6.59851193e-01
4.05547768e-01 6.80024564e-01 3.22963148e-02 -6.99778438e-01
4.61085737e-02 1.74812246e-02 -1.98066747e+00 5.79238117e-01
8.49433780e-01 7.26577461e-01 1.38564911e-02 6.67288244e-01
-9.30646420e-01 1.01790988e+00 -5.27066469e-01 -5.92484593e-01
-1.12749888e-02 -7.91692615e-01 -9.03956831e-01 2.23725870e-01
3.23756099e-01 -1.66976020e-01 1.13163255e-01 8.19527805e-01
2.85618097e-01 3.44847411e-01 6.91986322e-01 -1.19662082e+00
-5.61384439e-01 1.10458779e+00 -2.66289473e-01 2.43342817e-01
1.15840204e-01 3.56197497e-03 1.30686533e+00 -1.15660977e+00
7.05536008e-01 1.20807874e+00 5.77135801e-01 1.05797839e+00
-1.44927990e+00 -4.44066703e-01 3.65068048e-01 -5.24572209e-02
-7.46879816e-01 -4.70537245e-01 8.38084280e-01 -2.85712361e-01
1.12059045e+00 2.31070369e-01 5.25956392e-01 1.51338291e+00
-1.43926516e-01 1.03702569e+00 9.43666220e-01 -1.12815726e+00
4.80869025e-01 2.40040645e-01 8.40850830e-01 5.31957686e-01
3.01155686e-01 -2.45153338e-01 -5.39960384e-01 -2.35014036e-01
-3.06270495e-02 -5.76731823e-02 9.91679057e-02 -2.58723110e-01
-7.54720211e-01 1.40183771e+00 -2.41790891e-01 5.32598436e-01
-1.50758550e-01 -2.24940460e-02 7.29161263e-01 2.32128546e-01
8.66725206e-01 5.69987059e-01 -5.09855866e-01 -1.58663705e-01
-8.24384749e-01 3.18747461e-01 9.03879642e-01 9.61190403e-01
6.80644095e-01 2.91790456e-01 -4.83521909e-01 9.58254039e-01
5.10808378e-02 1.02473460e-01 7.51668334e-01 -6.10322535e-01
5.56772351e-01 6.72597766e-01 -1.31108448e-01 -2.45482191e-01
-2.01407656e-01 -3.51454526e-01 -6.47240281e-01 1.40891075e-01
3.92934889e-01 -2.31354341e-01 -9.47881401e-01 1.83056974e+00
4.21610177e-01 -1.91580400e-01 2.34215677e-01 3.88286233e-01
6.72136307e-01 6.49482787e-01 5.45947313e-01 -3.62209290e-01
1.56772089e+00 -1.03262365e+00 -1.04031086e+00 -4.91202742e-01
1.26596749e+00 -5.38572490e-01 1.65721273e+00 2.07878083e-01
-9.26274478e-01 -4.99873757e-01 -1.18198133e+00 -1.47535890e-01
-4.86806840e-01 -2.90336907e-01 7.82361388e-01 1.10185313e+00
-4.64626461e-01 5.38513303e-01 -6.33143723e-01 -2.42698565e-01
6.16023779e-01 -6.62803352e-02 -1.42597944e-01 8.27882066e-03
-1.31881618e+00 8.81408453e-01 5.13431966e-01 -6.56193793e-01
-5.10569811e-01 -7.64010966e-01 -1.04481757e+00 1.52011111e-01
5.61399937e-01 -4.91883188e-01 1.63538957e+00 -9.09929991e-01
-1.40938723e+00 9.27857935e-01 -3.11893553e-01 -7.01256931e-01
3.30751300e-01 -2.39841178e-01 -8.15600380e-02 -1.77014798e-01
6.41039014e-02 7.16237545e-01 7.47423232e-01 -1.03292060e+00
-6.59319162e-01 -3.09195638e-01 -1.50163591e-01 2.52491891e-01
-7.11552620e-01 2.97898501e-01 2.18933180e-01 -8.91494215e-01
-3.73957872e-01 -5.06845713e-01 -3.46432954e-01 -3.70687246e-01
-1.98759571e-01 -4.99472976e-01 7.53571749e-01 -2.12670729e-01
1.24278927e+00 -1.82981145e+00 -1.17846802e-01 -3.01571876e-01
1.62415747e-02 3.51301610e-01 -2.08904833e-01 5.26247740e-01
-1.72710001e-01 4.04977471e-01 -3.28817964e-01 -8.67178380e-01
1.36694714e-01 4.50773895e-01 -8.04459751e-01 -7.00005442e-02
4.22748387e-01 9.81069982e-01 -9.27892566e-01 -5.18288791e-01
2.18068153e-01 2.08883643e-01 -5.83417773e-01 1.73511982e-01
-6.99709475e-01 1.00290090e-01 -4.82223451e-01 2.96730995e-01
2.39461780e-01 -2.06568241e-01 1.44222975e-01 2.75857270e-01
6.26189932e-02 5.99567235e-01 -8.96100879e-01 1.65379679e+00
-2.95211524e-01 3.99251521e-01 -3.20491552e-01 -1.17649949e+00
1.09278476e+00 3.22582364e-01 1.57320336e-01 -3.54638487e-01
5.13806224e-01 2.55993977e-02 -9.59219132e-03 -3.29378456e-01
6.46079838e-01 -6.99198961e-01 -2.69802630e-01 8.07175338e-01
4.31044012e-01 -2.38277510e-01 5.93299508e-01 3.38341147e-01
1.00581133e+00 2.60791361e-01 7.90943325e-01 -2.61280149e-01
3.61066550e-01 2.87516057e-01 3.87702584e-01 9.09666479e-01
-1.35486603e-01 5.53948045e-01 5.23110628e-01 -6.47634268e-01
-1.49328685e+00 -5.51075459e-01 -8.44951570e-02 1.73437321e+00
-5.11879683e-01 -5.77655375e-01 -9.29091454e-01 -9.30372596e-01
-4.05425668e-01 1.40853727e+00 -8.10759127e-01 -2.24613860e-01
-4.94852901e-01 -1.10990584e+00 4.75243151e-01 6.77492857e-01
8.35038349e-02 -1.15546417e+00 -5.65946758e-01 3.15106302e-01
-2.62880206e-01 -1.11363971e+00 -3.18161219e-01 7.15819418e-01
-9.52048779e-01 -9.00167942e-01 -3.02780390e-01 -8.48076284e-01
5.74298680e-01 2.20091015e-01 1.15095806e+00 4.67662588e-02
-3.03250700e-01 5.75900823e-02 -8.88369083e-01 -8.36642981e-01
-9.99051988e-01 2.31073588e-01 -6.85241148e-02 -1.96162656e-01
9.80798542e-01 -4.97254103e-01 7.57772615e-03 -3.84363811e-03
-9.71415043e-01 3.34003270e-01 2.14506239e-01 1.16869426e+00
3.07863116e-01 -4.05891329e-01 7.98946202e-01 -1.33163691e+00
8.95652354e-01 -3.69594961e-01 -2.32599720e-01 3.14642251e-01
-7.23319590e-01 4.17166322e-01 7.52419889e-01 -5.80818295e-01
-1.30959308e+00 1.48988394e-02 -3.32743108e-01 -1.30291983e-01
-4.73633766e-01 3.48330051e-01 -1.23157486e-01 4.71011132e-01
1.20920944e+00 2.59357035e-01 1.60742357e-01 -3.80713314e-01
7.28598833e-01 6.75389946e-01 1.51688635e-01 -8.10847282e-01
5.04167020e-01 3.87868971e-01 -2.58932143e-01 -8.87419581e-01
-1.40040493e+00 -3.10980082e-01 -8.64519298e-01 1.17505834e-01
7.93906450e-01 -6.27172410e-01 -7.46856108e-02 1.31583631e-01
-1.18111217e+00 -4.02936161e-01 -1.07304204e+00 1.96399093e-01
-5.67942560e-01 3.66484433e-01 -5.98577440e-01 -1.13492155e+00
-6.77105784e-01 -6.64795160e-01 9.49249983e-01 -9.07474160e-02
-5.77948332e-01 -8.74250650e-01 2.13646680e-01 2.44694456e-01
2.44404942e-01 8.20708498e-02 1.16184640e+00 -1.49255943e+00
1.22362137e-01 -3.02678913e-01 2.64936984e-01 4.00595486e-01
-1.18088732e-02 -1.51263639e-01 -1.37868369e+00 -6.50147675e-03
3.82198960e-01 -7.43519425e-01 7.26959586e-01 1.90205783e-01
9.39288497e-01 -2.95223147e-01 -1.15818493e-01 1.78831115e-01
1.03792953e+00 1.67216554e-01 5.23532391e-01 2.63797730e-01
4.78308797e-01 9.63333070e-01 6.16046846e-01 5.64293265e-01
-3.22542451e-02 3.35419923e-01 9.88977700e-02 1.73197433e-01
-1.84821516e-01 -3.94657701e-01 2.18310922e-01 7.31696308e-01
2.19732270e-01 -3.58620048e-01 -9.61279333e-01 4.07783180e-01
-1.71720839e+00 -9.83682096e-01 -2.23378465e-01 2.02580070e+00
1.14902174e+00 3.02424073e-01 1.22249201e-01 4.72784251e-01
7.35899031e-01 1.64261281e-01 -2.62686908e-01 -4.94618773e-01
-1.91077843e-01 5.79800963e-01 2.70393118e-02 5.84013939e-01
-1.01428068e+00 1.01656902e+00 6.34656906e+00 9.34117913e-01
-5.34350038e-01 3.05041820e-01 8.35099041e-01 -1.75796807e-01
-4.55018938e-01 7.00493306e-02 -1.32594359e+00 2.71305144e-01
1.18163359e+00 -2.71281630e-01 8.67691189e-02 1.12704432e+00
-1.43713197e-02 1.45755172e-01 -1.26805067e+00 6.44082129e-01
2.89479375e-01 -1.45491481e+00 3.03428441e-01 6.81795226e-03
7.62612045e-01 -1.22093998e-01 -2.24627718e-01 8.40951920e-01
5.58504045e-01 -9.05559480e-01 7.25236297e-01 1.92200281e-02
5.52787006e-01 -6.45403087e-01 7.36112833e-01 6.97765231e-01
-8.01989317e-01 -7.98226073e-02 -4.12867546e-01 -4.84714329e-01
2.44999602e-01 4.99375314e-01 -8.97886992e-01 1.71916261e-01
2.20014304e-01 1.54700652e-01 -5.18948734e-01 2.74692714e-01
-4.30734217e-01 7.41940916e-01 -8.08965191e-02 -4.94619220e-01
3.17870200e-01 5.72457090e-02 4.74614352e-01 1.18592048e+00
-6.32247627e-02 1.87279776e-01 2.33242109e-01 8.41179609e-01
-1.60163179e-01 4.65198725e-01 -7.01653242e-01 -2.45967433e-01
4.25860465e-01 1.21540940e+00 -7.22227156e-01 -7.58707285e-01
-6.14529848e-01 5.39696157e-01 5.79207361e-01 4.20498401e-02
-4.35910910e-01 -2.55112946e-01 4.85610843e-01 1.53777823e-01
7.46968836e-02 3.54622267e-02 -6.45680547e-01 -1.31308627e+00
-7.12454617e-02 -1.02838254e+00 5.29761672e-01 -3.90454769e-01
-1.63124573e+00 5.65049350e-01 1.39158025e-01 -9.70052838e-01
-5.55371821e-01 -6.84807897e-01 -7.07396746e-01 8.60778153e-01
-1.28733623e+00 -1.11471999e+00 -1.79264128e-01 1.03906423e-01
1.23193753e+00 -2.87886649e-01 1.32241929e+00 -1.36536136e-01
-5.59775412e-01 4.71961468e-01 -3.64157706e-02 1.21731132e-01
5.24494886e-01 -1.29748356e+00 8.63049448e-01 9.83433902e-01
2.53403455e-01 4.91825134e-01 8.64559591e-01 -7.37945497e-01
-1.02366936e+00 -9.85774338e-01 1.06367207e+00 -5.97154140e-01
7.44332552e-01 -6.76060796e-01 -1.25113249e+00 7.84010172e-01
2.89184928e-01 6.65728301e-02 1.17650652e+00 4.02099103e-01
-3.79861802e-01 2.88890749e-01 -9.58880067e-01 6.80729926e-01
8.73907566e-01 -2.72267669e-01 -1.20164311e+00 2.72774011e-01
8.28940094e-01 -5.81206381e-03 -3.03481579e-01 2.24605501e-01
2.61228383e-01 -6.43509626e-01 5.54958463e-01 -1.32198942e+00
8.18886936e-01 3.48380566e-01 -2.80083686e-01 -1.10869884e+00
-7.56772533e-02 -5.75955510e-01 -3.03860217e-01 1.38337243e+00
5.08845747e-01 -4.36861545e-01 1.09268212e+00 9.23439443e-01
-2.57596165e-01 -7.39180088e-01 -6.99023068e-01 -8.30450535e-01
3.83417010e-01 -7.38346159e-01 4.04150128e-01 1.09568751e+00
5.19950688e-01 1.03496027e+00 -4.03081805e-01 -6.63664997e-01
5.87360322e-01 -1.65667489e-01 8.61004233e-01 -1.46091056e+00
-1.79647282e-01 -5.65508425e-01 7.97878671e-03 -7.83823967e-01
2.90409297e-01 -1.04135501e+00 2.12434709e-01 -1.32522500e+00
3.63528073e-01 -3.02193940e-01 -2.40174565e-03 6.79537892e-01
-5.14922023e-01 -9.31792781e-02 8.57717767e-02 1.90792810e-02
-3.95574331e-01 8.06738913e-01 8.96742642e-01 3.19898268e-03
-2.96426803e-01 -1.77630723e-01 -9.45358574e-01 8.11452866e-01
9.67696607e-01 -6.50630713e-01 -6.77439988e-01 -7.44485259e-02
1.70903683e-01 -3.35752428e-01 -1.30153850e-01 -6.81255043e-01
-4.39508297e-02 -2.12346524e-01 1.43817708e-01 -3.86286557e-01
3.19927007e-01 -3.86271656e-01 -5.32705069e-01 3.95604551e-01
-8.18793416e-01 -3.41475129e-01 1.63759366e-01 5.11794567e-01
-2.82693580e-02 -8.36892724e-01 9.09725428e-01 -3.42325002e-01
-4.77031440e-01 7.45771080e-02 -5.23353994e-01 5.10554254e-01
1.01400328e+00 -1.69661865e-01 -3.64791691e-01 -1.40341327e-01
-7.23319829e-01 -2.85747368e-02 2.73783773e-01 5.70695221e-01
2.65890688e-01 -1.23691428e+00 -8.68441820e-01 2.42586270e-01
2.91622967e-01 -7.76357725e-02 6.39328063e-02 3.97238910e-01
-1.31437287e-01 3.41556400e-01 -5.96514456e-02 -2.87667900e-01
-1.31090224e+00 7.32156873e-01 1.89614724e-02 -5.88180363e-01
-7.71304607e-01 8.12507272e-01 1.52793333e-01 -4.20421064e-01
2.71860540e-01 -3.30389529e-01 -3.55891883e-01 3.41592640e-01
9.55561221e-01 1.10447377e-01 1.51629373e-01 -1.79932877e-01
1.06030077e-01 -9.05286223e-02 -4.18204159e-01 -2.67397612e-01
1.43524027e+00 3.61156017e-02 2.25240886e-01 7.61013985e-01
7.90025711e-01 -2.08104879e-01 -9.95043576e-01 -4.15216476e-01
3.02528381e-01 -3.51004362e-01 -6.58152997e-02 -5.61882913e-01
-2.95591235e-01 1.11081839e+00 1.20464876e-01 5.50327420e-01
6.63748085e-01 -1.70452278e-02 6.20737672e-01 4.89861429e-01
1.43944174e-01 -1.16299665e+00 2.77115613e-01 7.80684650e-01
6.10128522e-01 -1.36501372e+00 -1.74803823e-01 -5.67860782e-01
-8.79322767e-01 1.13795662e+00 7.08121300e-01 1.58850223e-01
3.02425086e-01 4.53971148e-01 -1.44899443e-01 -5.14264256e-02
-1.07658851e+00 -3.61403704e-01 1.36161357e-01 6.47722781e-01
7.65148759e-01 -1.75387740e-01 -4.74508315e-01 9.78233576e-01
-2.69968569e-01 -1.22752890e-01 5.29830873e-01 1.20367575e+00
-9.22823250e-01 -1.39119017e+00 -3.24167669e-01 5.59492350e-01
-4.70727652e-01 -3.84155393e-01 -3.23271185e-01 8.35886300e-01
-1.16468161e-01 9.52723384e-01 7.88013935e-02 -1.07213393e-01
1.47604316e-01 1.03040612e+00 4.06351686e-01 -1.31536520e+00
-4.73619401e-01 -8.23211893e-02 3.95354301e-01 1.87648740e-02
-1.25679359e-01 -5.72418571e-01 -1.25731826e+00 -7.67857283e-02
-3.55031937e-01 3.50638270e-01 6.33586228e-01 1.29640257e+00
6.97893351e-02 5.44732213e-01 4.03017342e-01 -6.98038459e-01
-9.74352002e-01 -1.38555491e+00 -3.13290060e-01 5.50714016e-01
-1.61758866e-02 -4.25067395e-01 -3.86147439e-01 1.63625151e-01] | [10.949853897094727, 8.229547500610352] |
a018c791-af65-4363-84dd-2687d7cc87f9 | virel-unsupervised-visual-relations-discovery | 2207.00590 | null | https://arxiv.org/abs/2207.00590v1 | https://arxiv.org/pdf/2207.00590v1.pdf | ViRel: Unsupervised Visual Relations Discovery with Graph-level Analogy | Visual relations form the basis of understanding our compositional world, as relationships between visual objects capture key information in a scene. It is then advantageous to learn relations automatically from the data, as learning with predefined labels cannot capture all possible relations. However, current relation learning methods typically require supervision, and are not designed to generalize to scenes with more complicated relational structures than those seen during training. Here, we introduce ViRel, a method for unsupervised discovery and learning of Visual Relations with graph-level analogy. In a setting where scenes within a task share the same underlying relational subgraph structure, our learning method of contrasting isomorphic and non-isomorphic graphs discovers the relations across tasks in an unsupervised manner. Once the relations are learned, ViRel can then retrieve the shared relational graph structure for each task by parsing the predicted relational structure. Using a dataset based on grid-world and the Abstract Reasoning Corpus, we show that our method achieves above 95% accuracy in relation classification, discovers the relation graph structure for most tasks, and further generalizes to unseen tasks with more complicated relational structures. | ['Jure Leskovec', 'Tailin Wu', 'Daniel Zeng'] | 2022-07-04 | null | null | null | null | ['relation-classification'] | ['natural-language-processing'] | [ 3.12783599e-01 5.60455024e-01 -3.36180270e-01 -4.52770561e-01
-1.47908390e-01 -8.51314425e-01 8.23097110e-01 5.35130322e-01
1.93703160e-01 2.74939746e-01 3.81823599e-01 -5.19163013e-01
-3.64811659e-01 -8.53036106e-01 -7.68016934e-01 -2.93841392e-01
-3.30242008e-01 7.48269975e-01 4.01522249e-01 -5.95017662e-03
-2.09389217e-02 5.66734612e-01 -1.53543425e+00 8.18970978e-01
3.58446956e-01 6.12818420e-01 1.42532438e-01 6.61367416e-01
-3.74902695e-01 1.53685582e+00 -3.65305692e-01 -6.49961114e-01
1.40975401e-01 -5.68828940e-01 -1.50448680e+00 2.07840279e-01
7.13293612e-01 1.53124422e-01 -5.98668754e-01 7.97559023e-01
-2.96323150e-01 1.93483904e-01 8.25642884e-01 -1.47319102e+00
-9.68101382e-01 8.14270139e-01 -3.92027736e-01 2.82286048e-01
5.54836929e-01 -3.85498762e-01 1.75425816e+00 -5.12830019e-01
1.11760664e+00 1.44546747e+00 3.68818134e-01 3.89964849e-01
-1.86568093e+00 -4.77590740e-01 4.60162610e-01 3.92640859e-01
-1.29119468e+00 -2.90663093e-01 7.35294342e-01 -7.49540091e-01
1.29696369e+00 3.44341218e-01 7.09225118e-01 8.62050891e-01
2.30717018e-01 4.53279823e-01 1.20040488e+00 -5.09278476e-01
-7.10287467e-02 -1.86943207e-02 3.40012938e-01 9.79388893e-01
2.32145563e-01 -7.87292868e-02 -8.74511480e-01 9.98670384e-02
7.61460483e-01 -6.09081574e-02 -3.19877625e-01 -9.06402051e-01
-1.39218509e+00 4.33036059e-01 8.87340009e-01 3.23699743e-01
2.00188041e-01 2.35107139e-01 2.74517059e-01 6.59469187e-01
4.54018831e-01 7.08866537e-01 -2.66338885e-01 4.05474424e-01
-6.59078658e-02 -8.86299536e-02 1.02378428e+00 1.32287538e+00
1.13376355e+00 -5.87105989e-01 9.43911001e-02 5.28832793e-01
2.23661080e-01 -6.11697026e-02 3.83494236e-02 -1.06662703e+00
5.01498520e-01 1.11313426e+00 -4.06319082e-01 -1.14499378e+00
-5.83560526e-01 -1.04175866e-01 -6.06409431e-01 1.53425902e-01
4.05932397e-01 5.19818842e-01 -7.77907848e-01 1.71055889e+00
2.00675935e-01 -1.95055783e-01 2.93435335e-01 6.03515625e-01
1.10522211e+00 2.74967998e-01 1.60390884e-01 -1.07447058e-01
1.53807437e+00 -7.46496677e-01 -5.16745329e-01 -5.37735164e-01
9.95077670e-01 -5.20148158e-01 1.22369778e+00 3.42934765e-02
-6.39904737e-01 -5.28521478e-01 -9.03475702e-01 -5.74054599e-01
-4.49269921e-01 -4.68827128e-01 1.06168735e+00 8.75995234e-02
-1.06921518e+00 3.70277464e-01 -4.82771695e-01 -8.55709076e-01
5.91103911e-01 1.66259691e-01 -6.96284711e-01 -5.93802258e-02
-7.37437427e-01 1.05013978e+00 5.81595778e-01 -1.43621370e-01
-7.39410341e-01 -4.04682875e-01 -1.00915134e+00 8.88560936e-02
7.39101112e-01 -6.78476691e-01 1.01243460e+00 -9.31398451e-01
-7.05246449e-01 1.48624802e+00 -3.05481017e-01 -1.78989872e-01
-3.29680480e-02 -6.83788359e-02 -2.37094462e-01 8.57910439e-02
1.43485010e-01 4.92780238e-01 4.24223781e-01 -1.71711445e+00
-3.63574266e-01 -3.52667302e-01 5.58626652e-01 4.19721276e-01
1.46756276e-01 -1.65887803e-01 -3.47733676e-01 -2.95349896e-01
6.83695316e-01 -9.17393148e-01 7.25947693e-02 -6.31194934e-02
-6.70604825e-01 -4.65531558e-01 6.63645446e-01 -1.90186664e-01
6.27428830e-01 -2.27474737e+00 3.23447198e-01 3.57549995e-01
8.24259043e-01 -5.47967792e-01 -1.60979796e-02 6.05072916e-01
-4.47271287e-01 5.80966212e-02 -5.86065017e-02 5.75046390e-02
-8.14767322e-04 6.47141516e-01 -4.93972600e-01 2.40483388e-01
7.76180848e-02 1.12863219e+00 -1.20529270e+00 -6.71154559e-01
-7.75663480e-02 -1.36863306e-01 -3.03525507e-01 4.98057693e-01
-4.28020239e-01 4.17167395e-01 -2.34266058e-01 3.38217229e-01
7.72643508e-03 -8.73333275e-01 8.68732870e-01 -3.64717066e-01
3.77604693e-01 5.83374441e-01 -8.92316341e-01 1.62301099e+00
-4.39595997e-01 9.73834455e-01 -5.32261431e-01 -1.11365104e+00
9.85108554e-01 2.07490310e-01 1.95343599e-01 -6.24965370e-01
-1.45644203e-01 -1.73619688e-01 3.22422177e-01 -7.52764940e-01
1.32487178e-01 -3.28035980e-01 3.46909626e-03 6.79087460e-01
2.67446250e-01 -2.59297013e-01 2.47800052e-01 7.73364305e-01
1.29576874e+00 2.32219204e-01 4.18244123e-01 -2.72263736e-01
1.46563232e-01 1.88524738e-01 2.18016520e-01 6.10855401e-01
2.34159186e-01 1.04728431e-01 1.19298887e+00 -7.71785319e-01
-8.99222910e-01 -1.50748682e+00 1.33299544e-01 1.29050922e+00
4.85466719e-01 -7.88405120e-01 -2.00169876e-01 -9.41753685e-01
-1.09460145e-01 4.50938821e-01 -6.69258952e-01 -2.03314722e-01
-4.05701041e-01 -1.38338640e-01 1.46823630e-01 4.99981433e-01
8.07475895e-02 -1.24078035e+00 -4.58482921e-01 -3.48051697e-01
-2.15251431e-01 -1.68700254e+00 3.80625986e-02 5.35755098e-01
-7.76744962e-01 -1.62595654e+00 3.72396290e-01 -1.23978174e+00
9.90657270e-01 2.70198286e-01 1.70088398e+00 3.90577465e-01
-3.36687475e-01 6.29475772e-01 -2.39900529e-01 2.60528959e-02
-4.80739415e-01 -7.33013749e-02 -2.92701095e-01 -1.13473460e-01
2.88641572e-01 -6.92951322e-01 -5.63015677e-02 1.78517982e-01
-5.05507708e-01 5.09424448e-01 3.36764008e-01 5.79950154e-01
6.71292484e-01 1.98477268e-01 8.00819248e-02 -1.49223983e+00
4.23372865e-01 -1.85653508e-01 -5.62202454e-01 6.64186120e-01
-4.93782878e-01 5.33214629e-01 4.59594786e-01 -2.93008596e-01
-1.15498042e+00 8.91413614e-02 7.12508142e-01 -2.37555027e-01
-2.42814273e-01 3.96118194e-01 -2.92371839e-01 1.77650958e-01
9.30132389e-01 -1.36233151e-01 -3.66238981e-01 -1.97337270e-01
7.62551367e-01 -3.81383183e-03 5.70294142e-01 -8.57379377e-01
9.22009051e-01 5.93466461e-01 5.41749477e-01 -4.84666049e-01
-1.31836152e+00 -3.59968871e-01 -1.09534454e+00 -9.75301564e-02
1.19726419e+00 -8.45295787e-01 -9.77391779e-01 -3.27982217e-01
-1.21964359e+00 -6.88408136e-01 -4.86544222e-01 3.73072267e-01
-6.28743887e-01 1.24881618e-01 -6.59106433e-01 -4.94224221e-01
3.52107793e-01 -7.90309906e-01 8.38585615e-01 -1.57016784e-01
-5.54590166e-01 -1.14327335e+00 1.72649190e-01 2.89464712e-01
-3.39499086e-01 2.68199265e-01 1.71421075e+00 -8.01582277e-01
-1.02452874e+00 3.58078271e-01 -6.43552840e-01 -2.09830642e-01
4.15847540e-01 -1.98511973e-01 -9.08487678e-01 8.11341628e-02
-3.39599401e-01 -6.91316605e-01 5.98393202e-01 -2.24356294e-01
1.11356413e+00 -2.30076760e-01 -6.26128614e-01 5.26818752e-01
1.17038846e+00 1.91786081e-01 4.59461004e-01 1.68667719e-01
1.25245667e+00 1.21889353e+00 1.64796934e-01 -2.57995993e-01
5.81468880e-01 4.40244555e-01 3.18743825e-01 -1.73056811e-01
-2.83688635e-01 -5.00317693e-01 -4.77068238e-02 6.17464006e-01
-2.07050681e-01 -2.56793853e-02 -1.19514132e+00 3.31638962e-01
-1.89504087e+00 -9.33466196e-01 -4.28761631e-01 1.95060861e+00
9.75550175e-01 2.84498334e-01 -1.59898728e-01 -1.79492652e-01
5.59657276e-01 2.01870382e-01 -3.27565223e-01 -3.80210370e-01
-6.31727055e-02 2.83124954e-01 1.42388061e-01 5.96514940e-01
-9.11407769e-01 1.21591926e+00 6.59270239e+00 1.84149370e-01
-4.48180377e-01 -8.28559846e-02 4.99533772e-01 2.88860559e-01
-5.35767496e-01 5.36104739e-01 -3.01370412e-01 -4.52427238e-01
6.07588112e-01 -4.15182896e-02 7.21573710e-01 5.37006557e-01
-4.66229260e-01 -1.28446653e-01 -1.88313746e+00 9.99163389e-01
1.00023441e-01 -1.25880122e+00 1.54108495e-01 4.25550388e-03
5.22639036e-01 -2.99998134e-01 -2.82377392e-01 1.56717360e-01
8.57824445e-01 -1.33938968e+00 5.82821012e-01 4.38971549e-01
6.58903182e-01 -4.43578213e-01 2.78726816e-01 1.40839800e-01
-1.44869828e+00 2.20037460e-01 -3.34457964e-01 -3.87621999e-01
-1.69955745e-01 2.13803276e-01 -8.79370153e-01 4.50093120e-01
7.35325396e-01 9.83354211e-01 -9.15443480e-01 4.08971161e-01
-8.09691846e-01 1.75540045e-01 1.73857704e-01 1.39330566e-01
-2.47671485e-01 -2.52537161e-01 1.89604267e-01 9.67370629e-01
-2.17746034e-01 2.99894124e-01 3.24455082e-01 9.39189613e-01
-2.52010107e-01 -5.27469330e-02 -1.27387154e+00 -6.33449107e-02
3.40870976e-01 1.23893356e+00 -1.16419411e+00 -3.60732526e-01
-6.08517528e-01 9.26191688e-01 1.03360617e+00 6.25056744e-01
-4.53766823e-01 -3.58894020e-02 4.85038340e-01 1.44681990e-01
9.26215351e-02 -3.60724717e-01 -1.94355711e-01 -1.14953315e+00
1.77956242e-02 -7.72817373e-01 6.94301546e-01 -1.11345136e+00
-1.43760991e+00 6.75022960e-01 2.86952555e-01 -9.47220147e-01
-5.00431135e-02 -8.83976877e-01 -2.56527513e-01 6.68521464e-01
-9.93808091e-01 -1.30524957e+00 -4.93132025e-01 6.56466603e-01
2.33288407e-01 -2.74873618e-02 9.95759726e-01 -3.66695195e-01
-1.43268406e-01 1.17191393e-02 -5.49611151e-01 4.16819036e-01
5.23714960e-01 -1.65162051e+00 4.83086914e-01 5.93457282e-01
1.12862849e+00 7.58960009e-01 6.01806760e-01 -6.28291786e-01
-1.39493811e+00 -7.85611093e-01 1.08094442e+00 -7.99121976e-01
1.24057484e+00 -9.37981069e-01 -1.12864029e+00 1.36984730e+00
3.40095401e-01 3.95627201e-01 8.15537512e-01 8.45521033e-01
-1.27932632e+00 -4.98407185e-02 -4.86318350e-01 9.15472448e-01
1.86464453e+00 -1.32763517e+00 -9.10581648e-01 6.25544429e-01
8.78356218e-01 -4.72714633e-01 -9.36729670e-01 1.50208920e-01
3.04169208e-01 -9.37976420e-01 9.33192074e-01 -1.05502152e+00
6.21062398e-01 -3.52083445e-01 -3.37648600e-01 -1.12918079e+00
-4.73957837e-01 -3.17517400e-01 -6.41919971e-02 1.15725946e+00
7.02262163e-01 -6.05900526e-01 6.24891698e-01 5.93030274e-01
1.38041556e-01 -4.51338977e-01 -5.41568518e-01 -8.73904884e-01
-3.41921479e-01 -3.79275918e-01 3.46655309e-01 1.32140350e+00
3.34357828e-01 1.18663073e+00 3.22127044e-02 3.30711156e-01
6.57240570e-01 6.64611280e-01 9.57129180e-01 -1.45778120e+00
-5.90206504e-01 -2.45169505e-01 -6.86743438e-01 -8.36899936e-01
7.67725825e-01 -1.36823142e+00 -2.31815446e-02 -1.92854083e+00
5.95655382e-01 -4.93710458e-01 -7.51949027e-02 7.37265408e-01
-7.37923682e-02 1.65990531e-01 1.09436557e-01 5.15707076e-01
-7.72614777e-01 5.44816107e-02 1.43121397e+00 -3.53137523e-01
3.21529731e-02 -4.18995142e-01 -6.63651347e-01 8.00865591e-01
4.12378818e-01 -5.57981908e-01 -9.97308016e-01 -4.39151049e-01
7.29609728e-01 -1.08692944e-01 7.15054095e-01 -5.64638972e-01
1.38629466e-01 -3.52504373e-01 3.81139517e-01 -1.35634989e-01
1.46807030e-01 -1.06312263e+00 5.00179827e-01 1.65329754e-01
-7.35079110e-01 6.86662346e-02 -6.05241507e-02 7.12097049e-01
-3.75273079e-02 6.29473031e-02 2.52488762e-01 -1.77318916e-01
-9.64712203e-01 7.19807744e-02 5.08792549e-02 4.03234541e-01
9.91240680e-01 -7.37601891e-02 -8.89632761e-01 -2.46102840e-01
-1.16800094e+00 -5.41929938e-02 6.35949433e-01 4.30618316e-01
5.09128273e-01 -1.35184586e+00 -2.26234943e-01 -1.12906754e-01
6.98426366e-01 2.04059005e-01 -1.67677760e-01 4.85813707e-01
-4.39025521e-01 3.12293991e-02 -1.83867529e-01 -7.20062912e-01
-1.44992399e+00 1.06366801e+00 2.80945539e-01 -2.67010957e-01
-1.03331923e+00 8.99039686e-01 1.04985762e+00 -3.23671192e-01
5.85397333e-02 -4.66999114e-01 -2.51558065e-01 4.39237105e-03
1.92212850e-01 -2.47836307e-01 -8.09925869e-02 -7.43629694e-01
-3.71132255e-01 5.60312510e-01 -1.55391274e-02 -7.27029331e-03
1.16265130e+00 1.07217133e-02 -5.17834246e-01 1.05760944e+00
1.10487950e+00 -1.02541797e-01 -1.14949310e+00 -5.46771586e-01
5.48262537e-01 -5.73035896e-01 -5.08512795e-01 -3.97169173e-01
-8.33003998e-01 7.13618875e-01 -7.14334995e-02 7.05343723e-01
9.46009040e-01 1.06629765e+00 -2.43974235e-02 6.77782118e-01
4.45284158e-01 -4.26726162e-01 5.39508283e-01 5.30115724e-01
8.30669701e-01 -1.10955131e+00 3.11530948e-01 -1.01820719e+00
-7.57605433e-01 1.00304544e+00 8.33177090e-01 -1.64478868e-02
4.94339406e-01 2.24622473e-01 -2.34398581e-02 -9.55295026e-01
-1.05313659e+00 -5.17714322e-01 6.43033445e-01 8.65994632e-01
6.52133107e-01 1.91712081e-01 3.06243479e-01 -6.93183616e-02
-3.23795676e-01 -6.84743106e-01 3.01154733e-01 8.37074339e-01
-2.27515697e-01 -1.07824683e+00 -3.07611842e-02 4.67022359e-01
-1.02657611e-02 -2.35260382e-01 -8.98445487e-01 1.04479516e+00
1.11566305e-01 7.78702199e-01 4.72332180e-01 -3.20722640e-01
4.51320320e-01 7.55983666e-02 1.01903772e+00 -1.12634706e+00
-2.26079926e-01 -3.01408738e-01 3.98167998e-01 -7.46583462e-01
-7.68993676e-01 -3.56710166e-01 -1.51583183e+00 -1.18164115e-01
4.81523313e-02 8.65923986e-02 -2.04058677e-01 1.20328653e+00
3.37555185e-02 6.64298713e-01 1.77781954e-01 -2.66259909e-01
3.83466572e-01 -5.29123366e-01 -7.18568265e-01 1.00545371e+00
2.35510051e-01 -8.14213812e-01 -1.15350731e-01 6.73860192e-01] | [10.41714859008789, 1.6806588172912598] |
ff47498e-3cd5-4029-875a-9444f0d15324 | learning-foresightful-dense-visual-affordance | 2303.11057 | null | https://arxiv.org/abs/2303.11057v2 | https://arxiv.org/pdf/2303.11057v2.pdf | Learning Foresightful Dense Visual Affordance for Deformable Object Manipulation | Understanding and manipulating deformable objects (e.g., ropes and fabrics) is an essential yet challenging task with broad applications. Difficulties come from complex states and dynamics, diverse configurations and high-dimensional action space of deformable objects. Besides, the manipulation tasks usually require multiple steps to accomplish, and greedy policies may easily lead to local optimal states. Existing studies usually tackle this problem using reinforcement learning or imitating expert demonstrations, with limitations in modeling complex states or requiring hand-crafted expert policies. In this paper, we study deformable object manipulation using dense visual affordance, with generalization towards diverse states, and propose a novel kind of foresightful dense affordance, which avoids local optima by estimating states' values for long-term manipulation. We propose a framework for learning this representation, with novel designs such as multi-stage stable learning and efficient self-supervised data collection without experts. Experiments demonstrate the superiority of our proposed foresightful dense affordance. Project page: https://hyperplane-lab.github.io/DeformableAffordance | ['Hao Dong', 'Chuanruo Ning', 'Ruihai Wu'] | 2023-03-20 | null | null | null | null | ['deformable-object-manipulation'] | ['robots'] | [-1.73749745e-01 -2.20775455e-01 -5.84489226e-01 -9.87644568e-02
-2.34990656e-01 -7.22882986e-01 2.32875302e-01 -4.78362799e-01
-1.76611245e-01 8.75324190e-01 1.92702964e-01 4.01417352e-02
-4.69927073e-01 -3.59851897e-01 -8.45210671e-01 -7.69580901e-01
-1.97102770e-01 5.40058792e-01 1.75254792e-01 -2.89930522e-01
2.34855860e-01 6.32780850e-01 -1.37624872e+00 -1.64770588e-01
9.65118289e-01 7.82713890e-01 6.18171930e-01 7.15344608e-01
2.89203167e-01 5.17728209e-01 -1.78103000e-01 1.60646781e-01
4.54076350e-01 4.53484170e-02 -5.51704466e-01 2.63874322e-01
1.68716669e-01 -6.22180223e-01 -5.38182080e-01 1.05076730e+00
3.91456544e-01 5.61381042e-01 5.56989789e-01 -1.32449567e+00
-1.04070926e+00 3.30097675e-01 -2.91770637e-01 -2.22903565e-02
4.58948642e-01 9.20144975e-01 8.00882459e-01 -4.68058378e-01
6.80230141e-01 1.65096724e+00 1.97930932e-01 1.01103175e+00
-1.19558311e+00 -5.76170504e-01 6.86825156e-01 8.74551460e-02
-1.03063965e+00 -3.28906715e-01 7.59056211e-01 -6.03807569e-01
7.92985618e-01 2.94994503e-01 9.85004246e-01 1.54547203e+00
2.83271939e-01 1.12487674e+00 1.20353949e+00 8.89123678e-02
3.07868868e-01 -4.45309848e-01 -1.92341596e-01 1.01637638e+00
3.34047139e-01 3.84322107e-01 -4.13279459e-02 -1.75671577e-01
1.51491702e+00 5.08672297e-01 -3.44560087e-01 -1.07307732e+00
-1.59959400e+00 5.56958497e-01 5.75849533e-01 -1.60211727e-01
-4.51877207e-01 5.04318655e-01 2.25838885e-01 3.24104309e-01
6.48928285e-02 7.54331708e-01 -6.26608253e-01 -1.42527834e-01
-2.32692406e-01 6.92489684e-01 7.27867305e-01 1.23265624e+00
4.71050560e-01 1.25366002e-01 -2.48448983e-01 4.92118537e-01
3.94271195e-01 7.17014492e-01 2.91113734e-01 -1.21559310e+00
4.97086704e-01 4.46685702e-01 6.52433515e-01 -8.47456872e-01
-5.36378622e-01 -1.36267655e-02 -9.05320644e-01 4.51522559e-01
4.71022725e-01 -2.56182581e-01 -1.06740439e+00 1.83719802e+00
5.96580982e-01 5.66930994e-02 -2.12612927e-01 1.31271160e+00
2.36220196e-01 5.84339976e-01 1.06389232e-01 -8.41014236e-02
8.85145187e-01 -9.61369991e-01 -8.66594553e-01 7.17401598e-03
-9.49196797e-03 -3.42306226e-01 1.53496373e+00 3.13547581e-01
-1.11069095e+00 -4.40674931e-01 -7.01744378e-01 -8.09175074e-02
-3.00986264e-02 1.58261150e-01 1.04673958e+00 -1.18113384e-01
-5.64048707e-01 1.03089762e+00 -1.52314973e+00 -1.35992274e-01
4.92896080e-01 5.50558507e-01 -1.61827058e-01 2.32150465e-01
-8.24762344e-01 1.03052378e+00 4.64602798e-01 3.55773091e-01
-1.47938228e+00 -4.50624704e-01 -7.54722655e-01 -1.65831581e-01
1.00072622e+00 -8.53505969e-01 1.25995469e+00 -5.78916013e-01
-1.93853378e+00 2.89009511e-01 1.66573539e-01 1.92821383e-01
7.60147572e-01 -6.61330044e-01 1.36053368e-01 -1.31295875e-01
-7.37172291e-02 5.21353066e-01 1.20039225e+00 -1.26630831e+00
-2.56548021e-02 -4.14131254e-01 4.37795281e-01 2.93841749e-01
-2.92260915e-01 -3.90680611e-01 -1.50149897e-01 -8.01465869e-01
1.79246411e-01 -1.39807212e+00 -6.23222649e-01 5.72808862e-01
-5.28790414e-01 -4.81242388e-01 7.68328249e-01 -4.06253785e-01
1.06563032e+00 -1.78701770e+00 9.87553239e-01 -6.85527623e-02
2.35862046e-01 2.75284082e-01 -2.20076710e-01 3.66301268e-01
2.96147317e-01 -3.62750180e-02 4.19611447e-02 2.04714611e-01
3.47352326e-01 5.05439520e-01 -2.78454363e-01 4.29970294e-01
2.32156441e-01 1.18378186e+00 -1.29997289e+00 -2.76111901e-01
3.89645457e-01 2.69496202e-01 -7.16042578e-01 5.64861536e-01
-7.63743818e-01 9.17582214e-01 -1.28388476e+00 8.68239880e-01
3.09785634e-01 -3.77955705e-01 5.07611074e-02 -3.44188452e-01
-1.59296244e-01 -4.34611365e-02 -1.41484272e+00 2.01071954e+00
-4.68821675e-01 -2.72966232e-02 1.85716063e-01 -9.11951065e-01
7.61348605e-01 1.75035730e-01 3.49753618e-01 -1.11092888e-01
2.22044766e-01 9.56050530e-02 -1.30375370e-03 -1.16657960e+00
2.31065929e-01 2.84528290e-03 -6.49862960e-02 3.96775961e-01
5.91416135e-02 -5.48739493e-01 2.27589477e-02 -1.86293826e-01
9.51546371e-01 8.50278437e-01 2.36934960e-01 -2.83134967e-01
1.42441034e-01 -1.90248370e-01 5.56137145e-01 5.25074422e-01
-3.62147003e-01 1.79649711e-01 4.16985780e-01 -6.59078062e-01
-1.08573723e+00 -1.19876778e+00 1.31448731e-01 9.37352538e-01
5.63674688e-01 -6.85327351e-02 -3.80237699e-01 -4.31101799e-01
3.78114372e-01 4.24014151e-01 -6.53412163e-01 -3.91603321e-01
-7.08370209e-01 -3.49564314e-01 -2.10768998e-01 6.36998534e-01
3.30467880e-01 -1.44581389e+00 -8.87129545e-01 1.53181747e-01
5.41264564e-02 -8.32809865e-01 -5.86204648e-01 -1.13780990e-01
-1.02508605e+00 -1.00107598e+00 -8.76544535e-01 -6.24596477e-01
7.39677250e-01 1.02903999e-01 7.61340499e-01 2.40070783e-02
-6.12479329e-01 6.40629530e-01 -1.58556759e-01 -5.40531203e-02
-1.10075474e-01 1.29175231e-01 4.75871861e-01 -2.48940825e-01
-3.05092514e-01 -6.38667822e-01 -8.30867589e-01 4.38959301e-01
-6.54842854e-01 1.44939065e-01 7.36262381e-01 9.55464363e-01
6.63132250e-01 -3.92503113e-01 4.03438896e-01 -3.35856199e-01
6.62665606e-01 -3.80727649e-01 -7.05242753e-01 3.83053988e-01
-2.20194101e-01 4.03823644e-01 4.41348106e-01 -1.12559044e+00
-1.04919779e+00 1.43224642e-01 1.66028887e-01 -9.48951900e-01
-1.91290483e-01 2.73112543e-02 -9.91102159e-02 -1.59474865e-01
6.20945454e-01 -9.76007283e-02 2.80106336e-01 -4.41504717e-01
5.00535071e-01 3.99149656e-01 7.17388690e-02 -1.35426342e+00
9.14123535e-01 3.17985684e-01 -8.58049840e-02 -3.53661656e-01
-9.40728784e-01 -4.96576838e-02 -8.13366771e-01 -2.91677207e-01
7.67725587e-01 -6.48017824e-01 -1.25679660e+00 4.40531462e-01
-9.29576516e-01 -8.68008018e-01 -4.03759390e-01 5.98288536e-01
-1.01452100e+00 1.76985174e-01 -6.70339704e-01 -7.22818553e-01
-1.44376516e-01 -1.28372943e+00 1.21039510e+00 3.77418876e-01
-2.62727499e-01 -7.22582221e-01 8.40492919e-02 1.97838172e-01
4.28047121e-01 7.07851589e-01 5.60767591e-01 1.23482473e-01
-1.06257534e+00 3.32841910e-02 6.63686022e-02 2.66989946e-01
5.14184177e-01 -1.53509779e-02 -3.95828545e-01 -6.70511782e-01
-3.50752980e-01 -9.66417134e-01 5.00414073e-01 3.37187022e-01
1.65195131e+00 -5.07677734e-01 -4.28137034e-01 5.29688478e-01
1.13703907e+00 7.72875249e-02 2.29695827e-01 1.16305150e-01
9.93724406e-01 4.40465569e-01 8.65994811e-01 7.06774473e-01
2.56084383e-01 5.86284995e-01 6.65433347e-01 3.03737372e-01
1.08503073e-01 -3.30277830e-01 3.00550193e-01 6.46007478e-01
-6.38653457e-01 7.10324664e-03 -6.25128269e-01 4.56370384e-01
-2.07732630e+00 -8.64692628e-01 3.71324718e-01 1.93923140e+00
9.96884108e-01 -8.61874409e-03 5.92656471e-02 -4.29998279e-01
5.14288723e-01 1.73869893e-01 -1.37405431e+00 3.05558555e-02
4.70150024e-01 -8.04972649e-02 1.45831734e-01 4.15077388e-01
-1.06702697e+00 1.16398418e+00 5.51341009e+00 6.80042982e-01
-1.17348993e+00 -9.04409885e-02 -1.41625684e-02 -4.93926257e-01
-2.83913314e-01 -1.54026657e-01 -5.55378973e-01 4.54618216e-01
1.92177057e-01 1.08663244e-02 9.10708725e-01 8.54498267e-01
2.82774746e-01 1.67874664e-01 -1.14983189e+00 7.85463631e-01
-4.15517390e-01 -9.92503881e-01 -5.67783741e-03 -1.19027458e-01
8.65284920e-01 -1.53226584e-01 1.10760495e-01 4.66747046e-01
7.10349202e-01 -7.15587497e-01 8.45379353e-01 7.72226989e-01
7.03796327e-01 -3.55517149e-01 -1.48124397e-01 5.45844913e-01
-9.46134984e-01 -4.39910948e-01 -5.53374946e-01 -2.58896470e-01
2.05611318e-01 1.96381941e-01 -1.18738599e-01 1.46175742e-01
6.65283203e-01 9.26279902e-01 1.79592837e-02 7.70227611e-01
-3.63060504e-01 1.06431514e-01 -4.07334268e-01 -6.27598703e-01
1.76292866e-01 -3.28386277e-01 8.17736208e-01 5.41308224e-01
1.52874216e-01 3.32642287e-01 7.17724264e-01 1.08638346e+00
1.65159509e-01 -2.97003120e-01 -7.01421976e-01 -2.33739585e-01
1.61739007e-01 1.25799930e+00 -5.06201565e-01 -2.93270200e-01
2.63937600e-02 8.57509017e-01 6.60676479e-01 4.94919896e-01
-9.79873955e-01 -7.12067410e-02 9.63021278e-01 2.15356145e-02
2.83524394e-01 -8.71908069e-01 1.99217469e-01 -1.75662386e+00
1.75087765e-01 -8.34512889e-01 -4.37301882e-02 -8.07789326e-01
-1.26141882e+00 1.38868704e-01 2.39178896e-01 -1.35948896e+00
3.04656848e-02 -7.05926657e-01 -2.52058834e-01 4.32214171e-01
-1.34529042e+00 -1.14927828e+00 -4.52264398e-01 8.93159866e-01
6.69777989e-01 8.91767964e-02 6.72341228e-01 -3.62092778e-02
-5.81682861e-01 1.94463551e-01 1.90371387e-02 -1.20303653e-01
4.69189614e-01 -1.27199340e+00 8.96461979e-02 2.41140634e-01
-2.67491549e-01 6.96826339e-01 6.67141378e-01 -6.61967158e-01
-1.98527193e+00 -9.09136832e-01 -2.44147405e-01 -5.28016925e-01
9.73836422e-01 -4.45749491e-01 -8.58185112e-01 9.10883546e-01
-1.10049099e-01 3.74364108e-01 3.29373255e-02 9.40577313e-02
-6.23784661e-02 2.26138428e-01 -1.02771866e+00 9.85171080e-01
1.66982365e+00 -3.17442358e-01 -5.89807510e-01 6.25087380e-01
7.42005348e-01 -9.67455745e-01 -9.93481100e-01 5.41345298e-01
8.31039727e-01 -3.83173168e-01 9.88619745e-01 -1.22311985e+00
3.15663993e-01 -3.70113462e-01 8.16764161e-02 -1.46701121e+00
-5.60358226e-01 -9.49103713e-01 -6.63305938e-01 7.08618701e-01
-1.31568108e-02 -5.57518601e-01 5.65240085e-01 6.72206938e-01
-2.02807620e-01 -9.54241455e-01 -6.50184870e-01 -1.00939763e+00
1.29631637e-02 -1.38707915e-02 5.72767019e-01 8.72554779e-01
-1.91919133e-01 -4.44704965e-02 -6.23284340e-01 2.68628538e-01
5.95717907e-01 5.55390656e-01 8.85738552e-01 -1.03121150e+00
-4.31572467e-01 -2.55074263e-01 -2.37154379e-01 -1.38243353e+00
4.04499739e-01 -7.13076770e-01 2.48632357e-01 -1.45882761e+00
1.87067062e-01 -8.36488307e-01 -1.54581651e-01 7.36575067e-01
-2.25257769e-01 -3.70464951e-01 2.57681668e-01 3.07278067e-01
-8.28092277e-01 1.11331141e+00 2.34863400e+00 -1.75931543e-01
-2.60534763e-01 1.06138878e-01 -2.98936069e-01 6.97053790e-01
8.60263169e-01 -2.01859385e-01 -5.82539856e-01 -7.10766137e-01
1.48278251e-01 2.53944665e-01 7.36411989e-01 -7.20429480e-01
-1.31562978e-01 -9.15841520e-01 2.71313637e-01 -3.35279778e-02
4.54615861e-01 -7.44539380e-01 1.28507420e-01 7.28976429e-01
-5.09700060e-01 -8.16114768e-02 6.70829490e-02 9.75506544e-01
2.71704406e-01 1.24700174e-01 6.77227437e-01 -3.81127536e-01
-6.75550520e-01 8.52245927e-01 -1.51399776e-01 1.03782341e-01
1.26932967e+00 1.24989629e-01 -7.51121342e-02 -8.44639242e-02
-1.24112701e+00 5.96887589e-01 5.51036477e-01 8.30522776e-01
5.72758079e-01 -1.40327203e+00 -3.45344096e-01 1.54739305e-01
-1.36922374e-01 3.91137272e-01 3.72967899e-01 5.38511097e-01
-3.13026696e-01 -8.69674142e-03 -5.78532040e-01 -6.47462606e-01
-8.19020212e-01 6.96906090e-01 2.01552451e-01 -1.04622781e-01
-7.61772811e-01 7.14031875e-01 7.55568594e-02 -6.99581981e-01
1.54646620e-01 -8.68208230e-01 1.16077639e-01 -3.45684677e-01
8.18547308e-02 4.28938061e-01 -5.69512010e-01 -4.11586426e-02
-3.09029743e-02 7.91198730e-01 -8.71601980e-03 2.74508715e-01
1.53415668e+00 7.63894022e-02 1.54515374e-02 5.30142486e-01
8.83063734e-01 -6.04794860e-01 -2.02997684e+00 -1.32386982e-01
-2.75950253e-01 -6.39586687e-01 -2.34685704e-01 -6.33610368e-01
-9.33770657e-01 8.36179137e-01 4.12528276e-01 2.53542606e-02
6.45959437e-01 8.71490315e-02 7.42740273e-01 1.02562773e+00
7.87079871e-01 -1.10477197e+00 6.47809446e-01 4.91817653e-01
1.47700894e+00 -1.46218812e+00 4.16526981e-02 -1.92560002e-01
-6.57311141e-01 1.07336891e+00 1.00333667e+00 -6.59009039e-01
7.98678458e-01 1.82434604e-01 -2.18691170e-01 -1.91501036e-01
-8.67668211e-01 -2.52408758e-02 2.51647025e-01 4.81170088e-01
-1.24876514e-01 3.50930393e-01 7.02491477e-02 2.48395577e-01
1.18996888e-01 1.86388001e-01 2.20707849e-01 1.29905832e+00
-4.37920958e-01 -9.38526630e-01 -1.47028670e-01 4.62389708e-01
-3.68611962e-02 4.41640109e-01 4.69726622e-02 8.49823117e-01
-1.71465322e-01 2.43006378e-01 -1.73425987e-01 -8.93892627e-03
5.06799221e-01 -2.83822566e-01 1.03383899e+00 -6.44292951e-01
-1.68916345e-01 -6.91706911e-02 -3.45996678e-01 -1.05500150e+00
-4.23479259e-01 -9.36716020e-01 -1.20769131e+00 -1.73480332e-01
-4.52981859e-01 -3.67512017e-01 4.35014546e-01 8.00268233e-01
4.05691147e-01 5.09490788e-01 6.52196765e-01 -1.50390565e+00
-1.26818490e+00 -9.38235104e-01 -4.58270043e-01 4.86175179e-01
7.10819364e-01 -1.30309653e+00 -2.00115576e-01 1.76258441e-02] | [4.785405158996582, 0.6302896738052368] |
9e5b06e5-75f3-4d3f-b9b9-92d102c9c966 | unsupervised-single-image-deraining-with-self | 1811.08575 | null | http://arxiv.org/abs/1811.08575v1 | http://arxiv.org/pdf/1811.08575v1.pdf | Unsupervised Single Image Deraining with Self-supervised Constraints | Most existing single image deraining methods require learning supervised
models from a large set of paired synthetic training data, which limits their
generality, scalability and practicality in real-world multimedia applications.
Besides, due to lack of labeled-supervised constraints, directly applying
existing unsupervised frameworks to the image deraining task will suffer from
low-quality recovery. Therefore, we propose an Unsupervised Deraining
Generative Adversarial Network (UD-GAN) to tackle above problems by introducing
self-supervised constraints from the intrinsic statistics of unpaired rainy and
clean images. Specifically, we firstly design two collaboratively optimized
modules, namely Rain Guidance Module (RGM) and Background Guidance Module
(BGM), to take full advantage of rainy image characteristics: The RGM is
designed to discriminate real rainy images from fake rainy images which are
created based on outputs of the generator with BGM. Simultaneously, the BGM
exploits a hierarchical Gaussian-Blur gradient error to ensure background
consistency between rainy input and de-rained output. Secondly, a novel
luminance-adjusting adversarial loss is integrated into the clean image
discriminator considering the built-in luminance difference between real clean
images and derained images. Comprehensive experiment results on various
benchmarking datasets and different training settings show that UD-GAN
outperforms existing image deraining methods in both quantitative and
qualitative comparisons. | ['Jianxin Lin', 'Zhibo Chen', 'Xin Jin', 'Zhikai Chen', 'Wei Zhou'] | 2018-11-21 | null | null | null | null | ['single-image-deraining'] | ['computer-vision'] | [ 4.20892000e-01 -1.18304223e-01 3.37941319e-01 -4.48307097e-01
-7.46735871e-01 -3.36491913e-01 4.25987840e-01 -6.15850210e-01
-2.19665408e-01 9.95177865e-01 -1.78449541e-01 -2.77417183e-01
3.10434729e-01 -9.36941743e-01 -8.11428845e-01 -1.35439587e+00
2.74347365e-01 -1.40641168e-01 -1.02478296e-01 -1.36098295e-01
-1.86948106e-01 2.23978773e-01 -1.39160585e+00 -5.74561348e-03
1.64976668e+00 7.59142578e-01 4.35876787e-01 5.73435903e-01
-3.87271075e-03 1.03342652e+00 -6.81358576e-01 -2.54005790e-01
5.49931645e-01 -9.82384264e-01 -3.38102542e-02 2.87562013e-01
6.29614413e-01 -6.60876095e-01 -5.67416608e-01 1.38280439e+00
6.47155881e-01 6.42849952e-02 3.92076790e-01 -1.14663064e+00
-9.23239589e-01 1.45786807e-01 -5.26186824e-01 2.58196056e-01
-3.66183631e-02 7.09424317e-01 5.63687801e-01 -8.91918421e-01
3.84965152e-01 1.13195825e+00 4.16472524e-01 5.35770237e-01
-1.34652412e+00 -6.50713086e-01 1.64526209e-01 -6.27574744e-03
-1.21024024e+00 -4.48452055e-01 8.13232422e-01 -2.42102295e-01
1.13231272e-01 5.17768860e-01 4.32347149e-01 1.14412975e+00
6.58319145e-02 5.79311550e-01 1.74104178e+00 -1.72369540e-01
2.58353680e-01 4.18842107e-01 -3.13915074e-01 4.94662732e-01
4.91299123e-01 4.66488212e-01 -5.21818176e-02 2.26941600e-01
7.48645604e-01 9.61690173e-02 -8.47563386e-01 -1.85437202e-01
-8.12070012e-01 6.56250119e-01 7.45064020e-01 6.85681179e-02
-4.08468872e-01 -2.17255548e-01 -3.25509906e-01 4.99592483e-01
6.03429675e-01 2.39962682e-01 -1.60634108e-02 5.49020708e-01
-1.07515216e+00 2.27584764e-01 4.44777578e-01 7.34234750e-01
1.02389693e+00 6.31773651e-01 -3.01175624e-01 7.35960662e-01
3.74980122e-01 1.25207901e+00 3.44083071e-01 -7.35412478e-01
5.29416084e-01 3.47997457e-01 3.19449484e-01 -1.05279863e+00
1.30031854e-01 -4.78839844e-01 -1.18544519e+00 5.80931127e-01
2.41028458e-01 -1.80548623e-01 -1.15275836e+00 1.59241223e+00
3.79964650e-01 2.83042550e-01 4.11114782e-01 1.29254174e+00
9.07335401e-01 8.84197712e-01 -8.44214335e-02 -5.51175892e-01
7.94656694e-01 -8.98538411e-01 -7.42259145e-01 -4.29424673e-01
1.25258520e-01 -6.99190974e-01 1.15865505e+00 2.25295961e-01
-1.16487336e+00 -7.74371207e-01 -1.15794706e+00 2.01334625e-01
-1.07813463e-01 1.19297430e-01 2.17819214e-01 7.51822472e-01
-7.65243232e-01 1.96052551e-01 -4.06958610e-01 -5.32512590e-02
4.74566460e-01 -4.58744764e-02 -2.27729797e-01 -5.87599635e-01
-1.32714224e+00 6.70005322e-01 2.24979728e-01 8.13805282e-01
-1.31870091e+00 -6.18525505e-01 -7.90654838e-01 -2.60856152e-01
-1.78284496e-02 -7.63651967e-01 6.19688690e-01 -1.53580737e+00
-1.46014357e+00 8.39104652e-01 1.98163688e-01 -3.55332881e-01
6.31693780e-01 -2.35445216e-01 -6.30550742e-01 2.49506608e-01
-7.47166127e-02 2.72619992e-01 1.47052240e+00 -1.86463499e+00
-4.59249586e-01 -1.65220991e-01 2.52502430e-02 4.13544834e-01
1.91954687e-01 -4.50948298e-01 -2.06361011e-01 -9.62119460e-01
-5.69837317e-02 -7.34911382e-01 -2.68865854e-01 -6.14547580e-02
-5.80437005e-01 8.20072591e-01 1.05025148e+00 -9.62477088e-01
9.27185535e-01 -2.14250946e+00 4.72090878e-02 9.42331776e-02
2.28787944e-01 5.73575974e-01 -2.50997126e-01 -2.78620161e-02
-8.45486671e-02 -2.46035486e-01 -7.68539965e-01 -2.11178750e-01
-2.71567583e-01 4.50230896e-01 -5.12906432e-01 7.01413095e-01
4.67455178e-01 8.39973032e-01 -1.03820360e+00 -4.53160673e-01
3.19986016e-01 5.51918685e-01 -2.13279903e-01 8.70960712e-01
-1.91616446e-01 9.72965896e-01 -3.17628533e-01 6.99157953e-01
1.25522137e+00 1.54827699e-01 -2.07014848e-02 -2.03482464e-01
2.01768115e-01 -2.38923684e-01 -9.25034225e-01 1.03943145e+00
-5.73415041e-01 4.62776452e-01 3.31530482e-01 -9.34139848e-01
9.78162110e-01 3.97603884e-02 -4.97653671e-02 -1.00836957e+00
7.51890838e-02 1.58205241e-01 -1.52655616e-01 -7.32878625e-01
1.71384811e-01 -3.68736923e-01 3.18767220e-01 2.00834662e-01
-1.26729563e-01 -4.14719433e-01 -2.01249570e-01 2.37592265e-01
8.12515140e-01 1.15077637e-01 -2.37467229e-01 1.07658774e-01
7.25651205e-01 -3.72870743e-01 7.79391289e-01 7.43409097e-01
-1.83879018e-01 1.16041315e+00 7.52111077e-02 -2.29351744e-01
-8.79393637e-01 -1.28092492e+00 -6.58678487e-02 5.53952575e-01
5.15971005e-01 4.52225477e-01 -8.85657966e-01 -7.37555504e-01
-1.55490831e-01 8.11947405e-01 -6.87429368e-01 -1.40359640e-01
-5.03297508e-01 -1.27306437e+00 4.47755247e-01 4.62048501e-02
9.55827475e-01 -1.13296974e+00 -1.94649294e-01 -8.43772888e-02
-2.24572748e-01 -1.24999726e+00 -4.57149804e-01 4.12390716e-02
-6.40164614e-01 -1.07259691e+00 -7.41724908e-01 -5.94531417e-01
9.51031446e-01 6.06045663e-01 1.10016894e+00 1.63428172e-01
-3.81422698e-01 7.07258135e-02 -4.34865594e-01 -2.39083186e-01
-5.64787209e-01 -5.37234008e-01 -3.10462624e-01 5.16396880e-01
-1.06916666e-01 -7.88516939e-01 -9.69037712e-01 3.14114571e-01
-1.38237655e+00 1.54750824e-01 8.73514831e-01 1.03424430e+00
5.15793800e-01 1.60433486e-01 4.42439198e-01 -9.54170346e-01
2.15497181e-01 -5.95327199e-01 -7.04847693e-01 2.03289136e-01
-6.22586906e-01 -1.75642937e-01 7.18126833e-01 -2.49699146e-01
-1.46027863e+00 -8.70348811e-02 3.92115191e-02 -5.74549854e-01
-8.76638368e-02 9.50319543e-02 -7.69563437e-01 -1.00279674e-01
5.42712748e-01 7.37300158e-01 7.44600818e-02 -5.84797114e-02
6.44184291e-01 5.06672740e-01 8.73058319e-01 -2.53890604e-01
1.50069642e+00 6.08446121e-01 -3.10736656e-01 -6.56138003e-01
-9.06581819e-01 -5.25354967e-02 -2.10156158e-01 -2.87681937e-01
8.97249937e-01 -1.22265255e+00 -6.10363707e-02 9.05858219e-01
-8.28680396e-01 -6.44605458e-01 -3.46257657e-01 3.17423224e-01
-3.21701795e-01 4.25046831e-01 -5.05519986e-01 -9.70040202e-01
-5.54917574e-01 -1.03185618e+00 8.24171066e-01 5.70763230e-01
6.69507086e-01 -8.74221742e-01 -5.23764081e-02 5.58236539e-01
6.67794108e-01 5.53870320e-01 6.99642360e-01 1.35472845e-02
-8.99995863e-01 -8.94483104e-02 -3.94279391e-01 1.09660077e+00
3.19946676e-01 -1.74332038e-01 -1.02795386e+00 -4.60186064e-01
3.84123266e-01 -3.18255484e-01 9.97508287e-01 2.87154794e-01
7.73874998e-01 -5.32306910e-01 1.26562506e-01 8.98509085e-01
1.65459597e+00 -5.40234633e-02 1.26539338e+00 3.03832859e-01
8.91906798e-01 4.13207263e-01 7.85411239e-01 1.06492855e-01
1.47569060e-01 3.71736616e-01 7.75137722e-01 -5.84496796e-01
-3.70735824e-01 -1.25289217e-01 7.28764653e-01 6.44850969e-01
-3.87507677e-02 -4.16159868e-01 -2.24947318e-01 4.24384385e-01
-1.60905302e+00 -1.08948123e+00 -9.97611359e-02 2.44414997e+00
9.35846388e-01 -2.44522723e-03 -2.52399385e-01 -8.30295533e-02
6.55556619e-01 4.83408153e-01 -6.22306824e-01 1.52450740e-01
-7.31066465e-01 2.64259189e-01 6.08256161e-01 4.68070209e-01
-1.11830974e+00 7.07530975e-01 5.07863235e+00 6.94339335e-01
-1.24046099e+00 2.29287103e-01 9.54658926e-01 -3.28071974e-02
-5.16894817e-01 -3.09349056e-02 -4.31495637e-01 9.33485866e-01
5.99864483e-01 2.43113652e-01 5.15118241e-01 4.00596589e-01
6.39656305e-01 -2.03417808e-01 -4.82204080e-01 7.93270946e-01
1.73261151e-01 -8.02266419e-01 1.83752552e-01 -1.71738848e-01
1.03537667e+00 -5.36574423e-02 2.90041447e-01 2.15805620e-01
3.01445425e-01 -1.12940133e+00 5.27570307e-01 8.51892948e-01
7.99785972e-01 -5.08038044e-01 8.21975648e-01 1.84328645e-01
-7.11114407e-01 4.59943935e-02 -3.53366703e-01 3.75919044e-01
1.18585706e-01 1.11082566e+00 -2.90101767e-01 8.90832424e-01
6.84040368e-01 5.54277062e-01 -5.11748731e-01 7.18155921e-01
-6.97789431e-01 7.96252787e-01 -7.70077389e-03 8.05700302e-01
8.91564041e-02 -8.25646698e-01 7.33115613e-01 1.08765709e+00
1.40029535e-01 1.16459191e-01 -5.52661568e-02 1.07902932e+00
-3.35272737e-02 -2.27615401e-01 -5.66024601e-01 1.68965429e-01
3.23272496e-01 1.37748182e+00 -3.24644059e-01 -2.32572749e-01
-3.26360136e-01 1.27184081e+00 -1.35986090e-01 8.04054201e-01
-1.12347829e+00 -3.59479696e-01 6.95848107e-01 -3.90249956e-03
4.16959047e-01 4.13480550e-02 -9.32302326e-02 -1.32516062e+00
1.29907772e-01 -1.24430013e+00 -8.64090472e-02 -1.02976871e+00
-1.48363805e+00 8.08200121e-01 -4.24220175e-01 -1.49723625e+00
1.48333743e-01 -1.63596049e-01 -1.01973855e+00 1.14096868e+00
-2.07207537e+00 -1.21651709e+00 -1.04228699e+00 6.58928454e-01
4.11248058e-01 -7.69321695e-02 4.43015277e-01 4.77857053e-01
-9.10327673e-01 5.78822434e-01 4.60323006e-01 6.08517006e-02
9.69845891e-01 -1.24368298e+00 6.90188482e-02 1.30689836e+00
-5.57698607e-02 2.49805525e-01 8.84454310e-01 -4.26870286e-01
-1.38344646e+00 -1.74576223e+00 1.88064665e-01 -1.64672196e-01
2.72107124e-01 -2.00029239e-01 -1.20338535e+00 3.95610392e-01
1.98575631e-01 4.76194859e-01 1.41122937e-01 -6.91487670e-01
-5.10390818e-01 -4.85068798e-01 -1.33542919e+00 4.50458258e-01
6.30596638e-01 -3.78393859e-01 -3.58234942e-01 4.59235817e-01
7.64431417e-01 -4.65893567e-01 -5.43626308e-01 4.44008231e-01
1.17612667e-01 -1.38074052e+00 1.01531661e+00 -1.53350100e-01
7.00408816e-01 -7.54902899e-01 -1.74888894e-01 -1.51629925e+00
1.19391523e-01 -6.01341903e-01 -1.41039401e-01 1.39741957e+00
6.17743507e-02 -8.04337859e-01 4.73291218e-01 5.34119420e-02
-5.97383305e-02 -5.32026768e-01 -4.70184773e-01 -6.98270619e-01
-4.24522720e-02 9.05432701e-02 4.47293460e-01 1.04463828e+00
-8.86708677e-01 1.23125456e-01 -7.70688951e-01 7.10343480e-01
1.01144600e+00 3.83578181e-01 1.00564718e+00 -5.62376797e-01
-4.42651182e-01 -7.90224373e-02 -2.14510664e-01 -9.43073034e-01
-3.51422206e-02 -5.73337615e-01 5.49430609e-01 -1.22112489e+00
1.03420712e-01 -4.76845831e-01 -2.81546891e-01 1.74484402e-01
-7.12902188e-01 5.62218368e-01 2.23531634e-01 4.40242589e-01
-4.18428510e-01 9.49068964e-01 1.43932772e+00 -4.01792347e-01
-1.62716642e-01 -2.74518225e-02 -6.29183769e-01 5.89994907e-01
7.51440823e-01 -4.75449026e-01 -3.85284364e-01 -4.36327368e-01
-2.86263466e-01 -5.03739193e-02 6.72353804e-01 -1.06502187e+00
-1.95791870e-01 -3.45433027e-01 3.90859306e-01 -1.99436083e-01
7.31351450e-02 -6.16079390e-01 2.13094220e-01 3.16659063e-01
4.50368784e-03 -3.86090457e-01 -1.13767572e-01 8.53579402e-01
-5.37777960e-01 1.08672030e-01 1.22636271e+00 -6.20793626e-02
-5.14548302e-01 2.38567978e-01 -1.10916041e-01 8.71624053e-02
8.21205199e-01 -2.09557101e-01 -5.17729223e-01 -5.34525692e-01
-4.19477940e-01 1.79151818e-01 5.69926381e-01 2.11855873e-01
6.15028083e-01 -1.10106754e+00 -9.18440580e-01 3.95304769e-01
6.20705970e-02 2.23418757e-01 6.56829894e-01 1.00569272e+00
-5.69498599e-01 -2.80493677e-01 -1.17559172e-01 -3.17034990e-01
-9.84605908e-01 4.82221454e-01 4.93140519e-01 -1.76861122e-01
-6.68118238e-01 6.71746671e-01 7.14789867e-01 -3.00063044e-01
1.36831999e-01 -1.09071627e-01 1.57899752e-01 -3.46344680e-01
6.26111925e-01 6.90065026e-02 1.68922450e-02 -5.99670529e-01
1.28902882e-01 2.02680677e-01 1.13956779e-01 2.79956579e-01
1.20906210e+00 -5.15355706e-01 -1.58307627e-01 2.45156392e-01
9.45023954e-01 1.24398880e-01 -1.62415266e+00 -2.80336529e-01
-5.89231312e-01 -6.89894259e-01 2.12025762e-01 -8.13931346e-01
-1.66886067e+00 6.06390536e-01 1.01317978e+00 7.61870891e-02
1.66466808e+00 -5.24055183e-01 9.13967431e-01 6.38372600e-02
4.39536348e-02 -6.05458438e-01 5.89745641e-02 1.89152777e-01
7.94986129e-01 -1.45004761e+00 1.81243494e-02 -3.23987931e-01
-7.51280487e-01 6.65242076e-01 6.26789808e-01 -3.30306649e-01
3.47592771e-01 8.21540207e-02 5.50861895e-01 -8.32273718e-03
-2.35430807e-01 -3.52483481e-01 3.06192432e-02 7.17686951e-01
-7.43907541e-02 -1.10468611e-01 -1.75989643e-01 3.94860923e-01
2.00797945e-01 -1.57699183e-01 5.29472351e-01 7.46862650e-01
-4.90378499e-01 -8.31988513e-01 -7.25856125e-01 2.60584116e-01
-2.01962218e-01 -3.53049397e-01 -3.17971781e-02 6.67210698e-01
4.34691995e-01 1.01280642e+00 -1.19236097e-01 -2.33359367e-01
2.05371380e-01 -4.56551522e-01 4.17716444e-01 -2.74782836e-01
-3.71606618e-01 3.01865004e-02 -3.97602111e-01 -4.22277838e-01
-7.48012483e-01 -3.02729070e-01 -7.31158435e-01 -1.12786397e-01
-3.32291275e-01 1.67871311e-01 3.84999454e-01 8.15039933e-01
2.23448023e-01 4.46490556e-01 1.28115785e+00 -1.00648916e+00
-4.63227630e-01 -9.06919777e-01 -9.82057273e-01 6.20263457e-01
5.64329624e-01 -4.04843539e-01 -8.57568443e-01 4.09182400e-01] | [10.933990478515625, -3.170112371444702] |
2fa04522-cb10-4253-b046-bb577208113b | cross-domain-ner-using-cross-domain-language | null | null | https://aclanthology.org/P19-1236 | https://aclanthology.org/P19-1236.pdf | Cross-Domain NER using Cross-Domain Language Modeling | Due to limitation of labeled resources, cross-domain named entity recognition (NER) has been a challenging task. Most existing work considers a supervised setting, making use of labeled data for both the source and target domains. A disadvantage of such methods is that they cannot train for domains without NER data. To address this issue, we consider using cross-domain LM as a bridge cross-domains for NER domain adaptation, performing cross-domain and cross-task knowledge transfer by designing a novel parameter generation network. Results show that our method can effectively extract domain differences from cross-domain LM contrast, allowing unsupervised domain adaptation while also giving state-of-the-art results among supervised domain adaptation methods. | ['Yue Zhang', 'Chen Jia', 'Xiaobo Liang'] | 2019-07-01 | null | null | null | acl-2019-7 | ['cross-domain-named-entity-recognition'] | ['natural-language-processing'] | [ 2.26049691e-01 3.22669856e-02 -4.36861962e-01 -6.06868088e-01
-8.04726064e-01 -1.08868217e+00 5.99695385e-01 -1.02686144e-01
-8.10281694e-01 1.13839996e+00 -1.03831850e-02 -1.09026648e-01
2.15588212e-01 -7.11076260e-01 -5.69486976e-01 -2.75090456e-01
4.12962765e-01 8.15916359e-01 5.53162932e-01 -3.50989640e-01
-2.22283266e-02 5.34059227e-01 -8.62944901e-01 1.25109181e-01
1.22082067e+00 4.21804368e-01 1.29269332e-01 2.10316047e-01
-7.40983427e-01 2.65143573e-01 -8.46660316e-01 -6.05431139e-01
1.85814023e-01 -5.21079063e-01 -1.06653142e+00 -2.25687996e-01
3.22376162e-01 1.21436909e-01 8.92657265e-02 9.58938599e-01
7.56643653e-01 3.00207227e-01 8.17888796e-01 -1.09377837e+00
-9.36230600e-01 5.21435738e-01 -3.51672918e-01 5.97885549e-02
1.92955926e-01 -1.69318765e-01 6.45184040e-01 -6.35715127e-01
8.85484219e-01 9.28900480e-01 7.95467079e-01 1.22156870e+00
-1.45721829e+00 -8.79053473e-01 6.04602396e-02 -3.05098053e-02
-1.31148851e+00 -4.01206881e-01 9.54285622e-01 -2.37229481e-01
9.52696562e-01 -3.91639560e-01 -2.04112485e-01 1.42951262e+00
-4.25346375e-01 6.14765227e-01 1.16077685e+00 -8.62410843e-01
2.35067531e-01 4.82571840e-01 -1.78464934e-01 9.57026407e-02
9.72662643e-02 -3.62166427e-02 -1.85336351e-01 -1.49629340e-01
7.18281090e-01 -4.40767854e-01 -1.23650767e-01 -5.10391533e-01
-1.12353814e+00 7.45734215e-01 1.38261080e-01 8.41958046e-01
-1.11571640e-01 -6.08931482e-01 6.26790822e-01 4.62956935e-01
5.21812439e-01 7.74255812e-01 -1.18169081e+00 -4.13616784e-02
-7.72205710e-01 -1.05058186e-01 1.03538644e+00 1.04074430e+00
8.13301325e-01 2.03891415e-02 4.46057171e-02 1.30109453e+00
-6.48773089e-02 6.61495209e-01 7.35659003e-01 -4.79011416e-01
8.30370903e-01 7.16778100e-01 6.49825260e-02 -4.11263585e-01
-4.22601253e-01 -6.57165125e-02 -7.19451189e-01 -1.80704985e-02
5.88911235e-01 -4.80867505e-01 -9.60271537e-01 2.06213570e+00
3.85627419e-01 1.78554803e-01 6.32834315e-01 3.81837338e-01
7.85389900e-01 4.93153602e-01 6.84962630e-01 3.47230248e-02
1.27554035e+00 -9.13108885e-01 -5.93630135e-01 -5.18546641e-01
8.03206205e-01 -5.77646554e-01 1.20944369e+00 -4.52274531e-02
-5.83455622e-01 -7.23500252e-01 -8.15543175e-01 -1.33613488e-02
-8.31792116e-01 1.54442787e-01 6.02976717e-02 6.06464684e-01
-7.02631176e-01 3.65921289e-01 -5.39900184e-01 -8.31062436e-01
2.12374985e-01 3.26869279e-01 -7.74650514e-01 -1.04659826e-01
-1.62455010e+00 1.16636169e+00 8.75011742e-01 -5.99395096e-01
-3.87188256e-01 -8.84754837e-01 -8.82570922e-01 -1.41712446e-02
1.87643990e-01 -5.56953669e-01 1.20324922e+00 -1.25994492e+00
-1.68575776e+00 1.12000537e+00 1.62047714e-01 -2.35233694e-01
1.58714741e-01 -1.30172968e-01 -1.05449855e+00 -4.18328419e-02
2.05985934e-01 7.27155387e-01 5.20526350e-01 -1.25385106e+00
-6.15899980e-01 -3.20246309e-01 -9.37260315e-02 2.04967022e-01
-6.70169830e-01 1.36576802e-01 -2.58996814e-01 -5.77910602e-01
-4.31284159e-01 -8.59785378e-01 -1.98851392e-01 -6.23248518e-01
-1.29762247e-01 -3.81580234e-01 9.18346524e-01 -6.78774178e-01
1.18635046e+00 -2.15254450e+00 -1.32233262e-01 -1.34970978e-01
-3.12404007e-01 7.87523866e-01 -4.82468516e-01 4.25095439e-01
-1.87699556e-01 3.33674282e-01 -4.40299720e-01 -2.39996180e-01
-3.00328657e-02 3.00190300e-01 -1.75268903e-01 3.59881148e-02
6.21088147e-01 6.25440300e-01 -8.59813690e-01 -7.22217560e-01
1.23873400e-02 3.99816275e-01 -2.87651598e-01 4.37270999e-01
-1.93189159e-01 9.07542229e-01 -6.01470947e-01 2.70965248e-01
8.40218902e-01 -1.79246619e-01 5.12290418e-01 -8.56923610e-02
2.62506935e-03 3.03626716e-01 -1.06638479e+00 2.00953889e+00
-7.72761762e-01 3.22493672e-01 -2.60519356e-01 -1.01515698e+00
1.21063030e+00 5.19754469e-01 2.70042598e-01 -8.47173154e-01
-1.15293097e-02 3.28640908e-01 -2.80286282e-01 -3.09664965e-01
2.70356029e-01 -6.48728311e-01 -5.20314574e-01 1.97807908e-01
5.38386345e-01 -6.25462234e-02 1.35684147e-01 -2.36386821e-01
8.98446500e-01 4.16919976e-01 6.14944041e-01 7.22598517e-03
7.31195390e-01 3.48789841e-01 7.94066727e-01 3.84116977e-01
-3.36497843e-01 5.88161588e-01 1.84941560e-01 -5.89763783e-02
-1.11985719e+00 -1.02054429e+00 -2.60223508e-01 1.31295061e+00
-5.51510453e-02 8.40110257e-02 -8.75418246e-01 -1.38174903e+00
-1.42172977e-01 8.62273395e-01 -5.32214046e-01 -1.09289400e-01
-8.42732251e-01 -4.12037671e-01 1.04590833e+00 7.54785240e-01
7.08819807e-01 -1.22433913e+00 -1.06309265e-01 2.84407616e-01
-4.33442920e-01 -1.58055294e+00 -5.01976311e-01 3.96253824e-01
-1.03957868e+00 -8.96567404e-01 -9.94251251e-01 -1.19701099e+00
5.87310731e-01 -1.63082808e-01 1.43492019e+00 -5.78456283e-01
1.98198020e-01 2.28058189e-01 -5.59814572e-01 -2.44557679e-01
-7.75338531e-01 6.68250501e-01 1.34105250e-01 -2.41603583e-01
7.57415712e-01 -6.54142082e-01 -9.01675746e-02 6.40402675e-01
-1.08186913e+00 -4.54902470e-01 7.37831891e-01 9.89422381e-01
5.73360562e-01 -2.62325089e-02 8.93349648e-01 -1.31198382e+00
5.43579936e-01 -6.11780107e-01 -5.70217431e-01 5.60110033e-01
-4.58939523e-01 3.00627530e-01 1.11290598e+00 -8.12940300e-01
-1.63676822e+00 1.99908659e-01 -6.34176359e-02 -2.21337214e-01
-8.11825097e-01 2.32559144e-01 -7.07678437e-01 8.43436494e-02
1.01270747e+00 6.20665960e-02 -4.08309817e-01 -8.05007517e-01
4.62214559e-01 7.73007810e-01 5.57210088e-01 -8.03352773e-01
8.72856081e-01 2.14931265e-01 -3.62002045e-01 -5.82239747e-01
-9.29122031e-01 -7.03961790e-01 -1.35077381e+00 3.97884637e-01
1.01768839e+00 -9.63708937e-01 -5.65504469e-02 2.54865766e-01
-1.27149856e+00 -5.02019405e-01 -3.69536549e-01 4.54095930e-01
-3.70674789e-01 3.15412462e-01 -2.48171449e-01 -3.40287626e-01
-1.10583521e-01 -6.42076790e-01 7.27730334e-01 3.36404651e-01
-1.77009255e-01 -1.59332609e+00 7.51761317e-01 2.03058630e-01
3.66294742e-01 2.12390244e-01 9.53955054e-01 -1.45239210e+00
2.77095109e-01 -5.51789477e-02 -2.12880298e-01 6.90155983e-01
4.07474518e-01 -4.63672489e-01 -9.70974624e-01 -1.41718209e-01
-2.75657117e-01 -4.16779757e-01 3.62336040e-01 -1.04499668e-01
6.02651238e-01 5.17674498e-02 -5.13267100e-01 4.82535869e-01
1.54941714e+00 2.02358767e-01 5.79370797e-01 7.04922378e-01
6.57884002e-01 6.97727740e-01 7.06846595e-01 2.00589582e-01
4.56957608e-01 8.14687729e-01 -2.45154873e-01 -4.09570396e-01
-2.48143822e-01 -4.16220486e-01 2.97151029e-01 6.39803588e-01
2.15579689e-01 -3.65421057e-01 -1.31649578e+00 9.04340506e-01
-1.60611808e+00 -7.39373684e-01 1.92317978e-01 2.08431053e+00
1.26732659e+00 -5.26005328e-02 2.66716540e-01 -3.36367100e-01
1.01765621e+00 -2.21072271e-01 -7.04850197e-01 -3.09122890e-01
-2.49410957e-01 5.29435515e-01 6.87574685e-01 1.38496131e-01
-1.29787803e+00 1.34981322e+00 6.22286463e+00 7.79296100e-01
-1.00356948e+00 2.72015214e-01 8.87463465e-02 5.44852674e-01
7.99161270e-02 -9.22102481e-03 -1.05792224e+00 4.11116779e-01
1.17719352e+00 7.87075683e-02 -2.24038456e-02 1.01801503e+00
-2.47632489e-01 3.18923891e-01 -1.03551126e+00 6.46352232e-01
6.72108084e-02 -6.49036646e-01 3.60788479e-02 -1.41511306e-01
8.84094119e-01 6.43554851e-02 -2.61403590e-01 6.14529550e-01
8.17210436e-01 -7.25931227e-01 7.31974915e-02 1.61294281e-01
1.13535500e+00 -7.56923258e-01 7.84749568e-01 3.95111501e-01
-1.07045829e+00 2.75001585e-01 -4.57064211e-01 3.67117792e-01
1.91208407e-01 4.86363649e-01 -1.08872509e+00 7.60283887e-01
5.74124992e-01 5.06723464e-01 -5.11734307e-01 7.90545106e-01
-3.76010776e-01 7.12255120e-01 -2.67650813e-01 3.35692197e-01
-7.60460347e-02 -1.56060560e-02 3.23745430e-01 1.55512309e+00
3.53237808e-01 -1.30856574e-01 2.58882403e-01 6.58994198e-01
-3.27905536e-01 2.00750023e-01 -8.27382863e-01 -1.23285867e-01
7.98590541e-01 1.02127361e+00 -4.88191873e-01 -3.06521088e-01
-5.60233176e-01 1.39062369e+00 5.40358961e-01 3.44094574e-01
-7.80795097e-01 -8.04575801e-01 7.04547703e-01 -1.86610773e-01
5.04546344e-01 -2.20893964e-01 -3.25832307e-01 -1.40391278e+00
-1.07440643e-01 -8.40308726e-01 7.35713303e-01 -3.40684742e-01
-1.83413672e+00 7.87888646e-01 7.42758065e-02 -1.46540153e+00
-2.51377821e-01 -6.01627290e-01 -4.10077184e-01 8.60430241e-01
-1.97171676e+00 -1.40180421e+00 4.00900245e-02 8.52976859e-01
4.18315738e-01 -3.89644295e-01 1.08399022e+00 5.98086417e-01
-3.53601545e-01 9.45951819e-01 4.68802750e-01 5.62915266e-01
1.55693305e+00 -1.29609478e+00 6.91766918e-01 7.02687621e-01
4.07016613e-02 5.15870988e-01 2.96568900e-01 -6.40067101e-01
-6.94642186e-01 -1.18661535e+00 1.15882254e+00 -7.49404490e-01
5.77033997e-01 -3.77696604e-01 -1.17658389e+00 7.40228534e-01
2.64501303e-01 1.16354942e-01 1.09546077e+00 3.55644107e-01
-6.80242836e-01 -1.00682294e-02 -1.44149733e+00 4.24072057e-01
1.05241835e+00 -5.10210991e-01 -9.74009812e-01 -4.41482812e-02
7.17225969e-01 -2.79166311e-01 -1.17894614e+00 3.62008572e-01
1.90353498e-01 -5.77001750e-01 1.00931156e+00 -9.02427435e-01
1.66245580e-01 -2.68158764e-01 7.39363357e-02 -1.72916996e+00
-2.69206882e-01 -1.67172953e-01 3.92559409e-01 2.09167099e+00
7.14012444e-01 -6.99764550e-01 6.48710370e-01 5.58897495e-01
5.64231873e-02 3.00270498e-01 -7.54340827e-01 -1.32565510e+00
7.09699214e-01 -3.22688818e-02 7.04509020e-01 1.61674237e+00
-1.03351986e-02 7.30093777e-01 -2.25062773e-01 3.42545033e-01
3.29653621e-01 -1.89096540e-01 7.47678459e-01 -1.49311411e+00
-1.20535411e-01 -1.18019946e-01 -8.04159939e-02 -9.06171739e-01
7.41233528e-01 -8.67865026e-01 2.36378044e-01 -1.34280777e+00
-7.54785836e-02 -7.31657326e-01 -4.88978624e-01 6.76585853e-01
-1.85969412e-01 1.32729098e-01 2.19738767e-01 2.44932637e-01
-7.29361713e-01 3.97715330e-01 1.16160834e+00 9.61293206e-02
-3.56722444e-01 -1.89662308e-01 -6.16564691e-01 5.66605389e-01
1.02533793e+00 -8.28192472e-01 -3.23962778e-01 -5.83739042e-01
-2.79393971e-01 -1.32492289e-01 -1.66646197e-01 -1.15501642e+00
1.26844168e-01 -3.46047103e-01 3.61995846e-01 -1.21922053e-01
2.90122200e-02 -1.04440606e+00 -1.66584611e-01 -6.21892372e-03
-2.72630751e-01 -7.98089579e-02 5.91243744e-01 4.86143827e-01
-5.12320101e-01 -3.79112720e-01 1.10554469e+00 -6.27820268e-02
-1.13214052e+00 -6.92204246e-03 -1.66667327e-02 7.43481517e-01
9.45866823e-01 -2.24228293e-01 -2.95330524e-01 1.81437299e-01
-6.76501572e-01 1.42265528e-01 6.94521308e-01 4.08412695e-01
-2.81011779e-02 -1.42825377e+00 -7.47804999e-01 9.47337747e-02
4.96536583e-01 1.13991193e-01 2.67363071e-01 2.36611173e-01
5.09088784e-02 3.51640970e-01 -5.64706206e-01 -4.44365054e-01
-1.03585005e+00 5.65889955e-01 2.29786351e-01 -9.74652529e-01
-2.24991471e-01 6.53502941e-01 3.36280882e-01 -1.43806481e+00
-1.09689631e-01 1.91900730e-01 -4.83982950e-01 3.15204300e-02
2.92745352e-01 1.85095578e-01 8.03121850e-02 -5.10506749e-01
-4.90856826e-01 7.21681237e-01 4.91960393e-03 -1.47003949e-01
1.20261288e+00 -2.62731105e-01 2.87353396e-01 3.25843990e-01
1.22292173e+00 7.57538453e-02 -1.15200889e+00 -6.00840688e-01
4.68124062e-01 -1.55592680e-01 -4.44940835e-01 -1.36929774e+00
-6.40729964e-01 9.28124547e-01 6.25193179e-01 -1.03165120e-01
1.28638160e+00 6.72665760e-02 9.30260241e-01 4.79086131e-01
3.76448900e-01 -1.38143349e+00 -1.43237896e-02 9.77543771e-01
3.91270190e-01 -1.62352562e+00 -4.12336290e-01 -2.71760851e-01
-1.00863457e+00 1.08731759e+00 9.45209801e-01 6.64644912e-02
4.27555591e-01 1.68139473e-01 5.14753520e-01 2.81302541e-01
-2.86235780e-01 -4.35650945e-01 2.98413754e-01 1.16157448e+00
6.11956596e-01 -1.39212534e-01 -2.14977682e-01 8.06078255e-01
1.76762030e-01 1.14556298e-01 1.16522640e-01 8.80048573e-01
-1.60650745e-01 -1.99083555e+00 -3.37075740e-01 -2.32201979e-01
-5.72246253e-01 -1.74052358e-01 -5.32643914e-01 1.16943908e+00
1.68085173e-01 8.78162086e-01 -1.57699451e-01 -2.16157466e-01
8.78226042e-01 5.43138802e-01 4.11809087e-01 -8.52010429e-01
-7.01366723e-01 -3.48446071e-01 2.20159292e-01 3.24756913e-02
-5.60866654e-01 -3.98029983e-01 -1.34046960e+00 7.91666955e-02
-3.01349938e-01 3.83945137e-01 5.92042446e-01 9.22546625e-01
6.39334977e-01 3.04109722e-01 5.24425626e-01 -4.47410911e-01
-5.65345824e-01 -1.03337038e+00 -5.09250402e-01 8.93047094e-01
1.38254687e-01 -7.64492035e-01 -2.45141357e-01 3.96780550e-01] | [9.84460735321045, 9.538708686828613] |
de44a022-1734-490e-8c99-cc7b8bd3c5af | nbc-softmax-darkweb-author-fingerprinting-and | 2212.08184 | null | https://arxiv.org/abs/2212.08184v1 | https://arxiv.org/pdf/2212.08184v1.pdf | NBC-Softmax : Darkweb Author fingerprinting and migration tracking | Metric learning aims to learn distances from the data, which enhances the performance of similarity-based algorithms. An author style detection task is a metric learning problem, where learning style features with small intra-class variations and larger inter-class differences is of great importance to achieve better performance. Recently, metric learning based on softmax loss has been used successfully for style detection. While softmax loss can produce separable representations, its discriminative power is relatively poor. In this work, we propose NBC-Softmax, a contrastive loss based clustering technique for softmax loss, which is more intuitive and able to achieve superior performance. Our technique meets the criterion for larger number of samples, thus achieving block contrastiveness, which is proven to outperform pair-wise losses. It uses mini-batch sampling effectively and is scalable. Experiments on 4 darkweb social forums, with NBCSAuthor that uses the proposed NBC-Softmax for author and sybil detection, shows that our negative block contrastive approach constantly outperforms state-of-the-art methods using the same network architecture. Our code is publicly available at : https://github.com/gayanku/NBC-Softmax | ['Marius Portmann', 'Shekhar S. Chandra', 'Gayan K. Kulatilleke'] | 2022-12-15 | null | null | null | null | ['metric-learning', 'metric-learning'] | ['computer-vision', 'methodology'] | [-1.64124906e-01 -5.59950650e-01 -3.27446461e-01 -7.68262684e-01
-6.75954580e-01 -4.49944705e-01 5.18586397e-01 -6.19668420e-03
-5.53786874e-01 5.86251855e-01 -7.21750334e-02 6.73799217e-02
-1.85337976e-01 -6.12094283e-01 -2.88150728e-01 -5.31894147e-01
-1.56322733e-01 2.67024279e-01 2.10089698e-01 -1.76123619e-01
5.60951650e-01 2.53378540e-01 -1.35738838e+00 2.67261952e-01
7.81479418e-01 1.24209893e+00 2.36285150e-01 6.85640097e-01
-1.62257016e-01 7.77686596e-01 -6.52118921e-01 -6.75744057e-01
2.13469639e-01 -4.48173702e-01 -6.44100487e-01 -3.89615297e-01
6.17232561e-01 -3.48083168e-01 -4.84392226e-01 1.16258156e+00
7.10716963e-01 1.18330508e-01 7.15575516e-01 -1.51007652e+00
-7.83508956e-01 6.94377065e-01 -7.16006398e-01 2.23632336e-01
2.33926341e-01 -2.60623902e-01 1.41172719e+00 -6.83810174e-01
2.93759167e-01 1.55287480e+00 6.35543525e-01 5.53421438e-01
-1.06764698e+00 -1.08097649e+00 3.46714817e-02 6.65868044e-01
-1.61370325e+00 -2.09608227e-01 1.14908624e+00 -1.21004894e-01
4.89818603e-01 2.77353615e-01 2.70657599e-01 1.41479540e+00
-1.00279756e-01 1.15286422e+00 1.12777567e+00 -2.92145610e-01
2.47071192e-01 4.72407103e-01 1.59309328e-01 6.54713571e-01
-6.91507338e-03 -7.54631609e-02 -7.11317241e-01 -3.33705455e-01
6.35534465e-01 2.88192064e-01 -2.22521722e-01 -1.08199991e-01
-1.01792192e+00 1.01282597e+00 5.22176385e-01 5.23571491e-01
2.78455079e-01 1.68132886e-01 4.17652160e-01 7.60339618e-01
4.22812879e-01 2.44123608e-01 -3.28417301e-01 -4.26135629e-01
-1.27583814e+00 2.01692805e-01 8.56806338e-01 8.17437112e-01
3.75813454e-01 -1.85919449e-01 -2.86485136e-01 1.21840334e+00
5.39827287e-01 6.26727760e-01 7.22020328e-01 -1.00236642e+00
3.48917961e-01 5.63672602e-01 -2.47575194e-01 -1.17628396e+00
-2.35628113e-01 -5.38369656e-01 -9.08084333e-01 4.16132063e-01
5.32451808e-01 2.00911779e-02 -3.40012223e-01 1.91199076e+00
8.51823911e-02 4.39680278e-01 -4.94180441e-01 8.67641628e-01
7.26261020e-01 5.89197516e-01 -1.79655805e-01 3.94497402e-02
1.27873135e+00 -5.99939704e-01 -4.59148645e-01 1.66937366e-01
4.71231669e-01 -1.01083028e+00 1.18565583e+00 6.64110482e-01
-8.75514090e-01 -6.67876840e-01 -1.29122412e+00 3.45657878e-02
-5.40483534e-01 6.43281862e-02 3.63805085e-01 9.35697138e-01
-8.34340274e-01 9.12625432e-01 -5.30616522e-01 -4.30811435e-01
6.26281977e-01 2.59105355e-01 -4.13256362e-02 1.51553124e-01
-1.21291041e+00 4.90854591e-01 2.77840734e-01 -1.91967994e-01
-4.49230999e-01 -6.83857501e-01 -5.29630065e-01 2.97072858e-01
4.46856916e-02 -1.51453659e-01 1.06295455e+00 -1.05713940e+00
-1.72019351e+00 1.04341638e+00 3.29213478e-02 -4.36693996e-01
6.89627290e-01 3.45544070e-02 -5.37091017e-01 9.78686102e-03
-1.77708566e-01 5.93407512e-01 9.62491512e-01 -9.10066307e-01
-6.87684774e-01 -5.04463434e-01 -2.82149967e-02 -1.24608148e-02
-1.10317206e+00 3.02383810e-01 -3.49959671e-01 -1.09649861e+00
-2.33204275e-01 -7.06892729e-01 3.84597570e-01 5.16785741e-01
-1.72849476e-01 -6.66339338e-01 9.71695244e-01 -6.79216027e-01
1.55486858e+00 -2.11729765e+00 -1.67106539e-02 3.97729754e-01
3.84495109e-01 4.13520962e-01 -1.46068871e-01 7.35719129e-02
3.62523556e-01 -1.71989296e-02 -3.46889228e-01 -6.08862221e-01
5.06651402e-01 -2.24413082e-01 1.05382800e-01 5.50005853e-01
-1.23797998e-01 5.48864543e-01 -6.48682952e-01 -7.61431277e-01
1.52750403e-01 5.21392703e-01 -3.73040199e-01 3.24174166e-01
6.23639114e-02 -1.03815429e-01 -2.70658076e-01 7.49095559e-01
9.15185869e-01 -2.82590777e-01 3.30419280e-02 -8.87244567e-02
3.25505197e-01 -1.31348103e-01 -1.53468251e+00 1.88327444e+00
-5.15625834e-01 9.10613775e-01 1.23591140e-01 -1.08753252e+00
1.23970044e+00 -4.41171788e-02 3.58909339e-01 -8.99918020e-01
3.26095611e-01 2.87729144e-01 -1.83997348e-01 -1.60406485e-01
2.29936481e-01 1.33891389e-01 -4.36946414e-02 5.59487224e-01
1.48019448e-01 3.77269417e-01 3.55981261e-01 3.19530278e-01
9.29013073e-01 2.31156964e-02 4.81348298e-02 -3.07272881e-01
9.45228755e-01 -9.08113897e-01 7.63292074e-01 7.47670472e-01
-6.02764130e-01 5.85075021e-01 5.06584406e-01 -1.38988674e-01
-8.84413421e-01 -1.22129071e+00 -3.47511500e-01 1.29129195e+00
3.84269357e-01 -4.51947689e-01 -7.40583658e-01 -9.14008856e-01
2.20906734e-01 4.36588317e-01 -5.82016110e-01 -2.50730962e-01
-5.94267547e-01 -6.76810324e-01 7.07730949e-01 3.31797272e-01
6.82618558e-01 -1.00325549e+00 -1.14624590e-01 8.92939344e-02
-6.51658624e-02 -8.86984706e-01 -7.56491542e-01 -2.26480171e-01
-5.95934451e-01 -1.14333487e+00 -1.05530667e+00 -1.00527847e+00
1.96054369e-01 1.27934635e-01 1.15269995e+00 2.94270869e-02
-4.03442502e-01 1.56751454e-01 -3.89207155e-01 -2.05357850e-01
-1.16148099e-01 2.80304402e-01 2.91013196e-02 2.77237475e-01
7.72932112e-01 -6.58159196e-01 -8.23348105e-01 3.86148214e-01
-7.05640614e-01 -5.46599686e-01 4.20536339e-01 8.94366503e-01
1.57979161e-01 -1.06797114e-01 6.97486162e-01 -6.97343230e-01
7.47394681e-01 -4.26106393e-01 -3.18728477e-01 3.28524679e-01
-8.57957244e-01 -1.39071187e-02 7.79379904e-01 -5.03200173e-01
-9.82261777e-01 -5.59274673e-01 1.77354459e-02 -3.44900489e-01
-6.63991719e-02 -2.55707279e-02 -3.07965189e-01 -1.61137402e-01
6.82523489e-01 1.86925918e-01 1.41826212e-01 -7.60421932e-01
1.71284184e-01 1.28092122e+00 1.08046107e-01 -5.25164366e-01
6.62806630e-01 3.89987767e-01 -1.89594448e-01 -8.62120926e-01
-5.25490940e-01 -7.00429440e-01 -3.99788260e-01 -2.35169917e-01
5.63780546e-01 -7.16092527e-01 -1.04181480e+00 7.36118495e-01
-8.86397123e-01 -2.27012590e-01 5.48189282e-02 5.96657276e-01
-4.01823640e-01 7.07184553e-01 -7.81338274e-01 -6.81393087e-01
-6.48673654e-01 -9.18093860e-01 8.14995527e-01 4.32379752e-01
-3.22022617e-01 -1.18053830e+00 -9.05668922e-03 3.21962714e-01
6.38222516e-01 1.33352369e-01 4.82313484e-01 -8.05191100e-01
-1.86000764e-01 -1.09491214e-01 -6.05689764e-01 7.84521639e-01
1.07640900e-01 -1.05823658e-01 -9.73741770e-01 -6.25570476e-01
-2.63309896e-01 -4.96269673e-01 8.98206413e-01 1.67914122e-01
1.56684244e+00 -1.07540883e-01 -5.05928583e-02 6.84113801e-01
1.38828158e+00 -9.53940004e-02 3.41154903e-01 3.95124167e-01
7.65829027e-01 4.29035574e-01 4.04404372e-01 8.36527169e-01
3.03245574e-01 7.53985584e-01 3.44841540e-01 -7.09393695e-02
-1.32384539e-01 2.40857825e-01 4.53360230e-01 6.39388025e-01
2.44057924e-01 -3.62989753e-01 -4.21079099e-01 2.34890088e-01
-1.91436160e+00 -1.14628398e+00 7.49143735e-02 2.26878023e+00
9.69552875e-01 4.26519871e-01 5.37836373e-01 3.98211658e-01
9.15241182e-01 3.93989980e-01 -5.07861733e-01 -5.74975312e-01
-1.06350571e-01 2.59276837e-01 2.85151929e-01 4.52328205e-01
-1.18028247e+00 7.88765252e-01 4.58801508e+00 1.43454671e+00
-1.02086782e+00 1.91105217e-01 6.22416675e-01 -4.84357864e-01
6.74440851e-03 -5.05717874e-01 -6.78621888e-01 1.20861685e+00
7.99160957e-01 1.64064225e-02 6.01347864e-01 7.50331402e-01
2.65929759e-01 3.65937531e-01 -1.08206773e+00 1.56704032e+00
3.56347710e-01 -9.27913725e-01 -1.79883584e-01 -1.16186514e-01
5.39592087e-01 -4.79428880e-02 2.06564412e-01 4.47139770e-01
1.25993133e-01 -7.65012145e-01 6.73781812e-01 4.81371731e-01
6.94592178e-01 -1.05345297e+00 8.00956964e-01 -2.08763406e-02
-1.13216209e+00 -1.03361122e-01 -4.67590004e-01 6.04611374e-02
-1.05136231e-01 7.24718809e-01 -2.47191325e-01 1.03935599e-01
8.63877773e-01 1.16952074e+00 -7.40196466e-01 1.20468128e+00
-4.67936173e-02 8.25460315e-01 -3.15287560e-01 -5.09782135e-01
1.26345038e-01 -2.56487012e-01 5.46452641e-01 1.41994560e+00
2.67195404e-01 -5.85850120e-01 1.79695711e-01 9.23558116e-01
-5.89127660e-01 3.07816535e-01 -1.69093117e-01 2.52010554e-01
5.73017657e-01 1.31946015e+00 -5.86073875e-01 -3.96151304e-01
-4.22338367e-01 1.29723203e+00 3.53144944e-01 1.31819323e-01
-1.03407180e+00 -1.01390398e+00 8.50884736e-01 -1.10204779e-01
4.47204590e-01 -2.32170131e-02 -4.77501839e-01 -1.17667317e+00
1.78557992e-01 -8.50104153e-01 6.66540802e-01 -2.71245897e-01
-1.66870570e+00 2.72751302e-01 -4.10767168e-01 -1.47276783e+00
9.36925784e-02 -8.05433095e-01 -8.02209198e-01 7.93521047e-01
-1.49191189e+00 -1.18106866e+00 -3.81489754e-01 6.77795768e-01
5.84961593e-01 -5.54007828e-01 4.76756334e-01 7.07012475e-01
-5.45820713e-01 1.40879714e+00 6.89072728e-01 2.18881175e-01
1.38859046e+00 -1.62994862e+00 -2.63665132e-02 6.23379827e-01
1.10821664e-01 4.42055404e-01 4.93936360e-01 -2.59259671e-01
-8.81738245e-01 -9.82236743e-01 8.23295534e-01 -2.68047124e-01
5.62048554e-01 -5.71304321e-01 -9.05561686e-01 -2.44664308e-02
3.20839852e-01 -1.49989992e-01 9.78451192e-01 9.09923092e-02
-7.64725685e-01 -6.50043726e-01 -1.55501425e+00 4.76976424e-01
1.00996351e+00 -5.71594894e-01 -3.14832032e-01 3.00067723e-01
3.55076224e-01 1.83012515e-01 -1.08824825e+00 7.41230249e-02
8.72850478e-01 -1.04260027e+00 1.03696394e+00 -2.30744213e-01
3.29082191e-01 -1.25254482e-01 -2.48436794e-01 -9.92078483e-01
-3.80895972e-01 -5.64041078e-01 -5.74629247e-01 1.63611448e+00
2.40457952e-01 -5.27650535e-01 1.05001557e+00 1.77073896e-01
4.04922932e-01 -7.27347314e-01 -8.51343632e-01 -1.11234736e+00
3.59692425e-01 -2.27708682e-01 5.57120264e-01 1.04016078e+00
-6.54982356e-03 1.92794338e-01 -5.41986048e-01 -3.19689751e-01
1.12975299e+00 3.62121529e-04 4.57638294e-01 -1.51071215e+00
-5.26184142e-01 -9.94601905e-01 -5.10953605e-01 -9.33452964e-01
2.76588053e-01 -1.01060498e+00 -3.78982663e-01 -1.06923795e+00
3.57531697e-01 -3.15476984e-01 -8.88907969e-01 1.42252222e-01
-2.08646700e-01 6.17427349e-01 2.70446151e-01 2.19570085e-01
-9.94305491e-01 6.71007276e-01 9.14165258e-01 -5.14003098e-01
-2.46048227e-01 1.75050125e-01 -6.09684587e-01 5.29700160e-01
1.16388309e+00 -4.46752429e-01 -2.61327893e-01 -1.00678250e-01
8.96675065e-02 -5.52948236e-01 1.90380856e-01 -1.04354060e+00
1.25149265e-01 2.34916687e-01 5.01725316e-01 -3.44381630e-01
2.56152451e-01 -7.01066613e-01 -6.39460504e-01 5.92694104e-01
-6.06610835e-01 -1.57893747e-01 -2.82282948e-01 6.10846400e-01
-2.50103533e-01 -3.71871829e-01 9.62839007e-01 2.98865121e-02
-6.63394272e-01 4.24969971e-01 -5.91465980e-02 2.80847311e-01
7.14696169e-01 -2.41156682e-01 -1.37966941e-03 -3.41067284e-01
-3.98158818e-01 3.19664776e-01 3.13903302e-01 6.31635666e-01
5.26899219e-01 -1.56263149e+00 -9.56809103e-01 2.24280972e-02
2.47664019e-01 -6.99820459e-01 9.24330279e-02 6.51304781e-01
-4.19142663e-01 3.40217948e-02 -2.66162634e-01 -6.08594954e-01
-1.51581907e+00 2.06027418e-01 1.58785373e-01 -2.13917270e-01
-1.50436088e-01 9.34706807e-01 -2.26437792e-01 -6.32153988e-01
7.55113184e-01 1.87685415e-02 -2.35723853e-01 1.77262813e-01
7.47221410e-01 9.49479043e-01 1.44260784e-03 -3.40215236e-01
-3.75636011e-01 5.03892720e-01 -3.29468936e-01 1.95042808e-02
1.18919170e+00 -1.78299442e-01 -4.41260375e-02 2.74619311e-01
1.81504238e+00 -2.37281486e-01 -1.19029129e+00 -6.52640283e-01
-2.48185955e-02 -6.21204495e-01 2.70746082e-01 -7.65628457e-01
-1.04207993e+00 8.48091304e-01 1.28848290e+00 2.41681099e-01
9.33600724e-01 -2.02012822e-01 1.19911039e+00 3.63746494e-01
1.40195623e-01 -1.48337054e+00 3.55439425e-01 2.93300122e-01
6.09808922e-01 -1.75120735e+00 -1.04247205e-01 1.07168131e-01
-1.29806623e-01 1.17660940e+00 6.35786295e-01 -1.85094699e-01
1.00428677e+00 6.02115430e-02 -8.72434303e-02 1.61012262e-01
-1.33133218e-01 -4.29265089e-02 2.85511911e-01 3.79690915e-01
5.66568136e-01 1.56850174e-01 -5.97332001e-01 5.74250281e-01
-1.34460360e-01 -2.15350002e-01 1.62286356e-01 7.59577096e-01
-3.58287603e-01 -1.33528721e+00 -2.90169448e-01 5.60537696e-01
-6.05570495e-01 -1.30432010e-01 -4.44972277e-01 4.40426111e-01
1.59859404e-01 1.04057467e+00 9.49461758e-02 -4.73114669e-01
9.62905437e-02 -1.32396564e-01 4.67348754e-01 -6.91626668e-02
-6.23711586e-01 -2.47612670e-01 -2.56439209e-01 -6.02419138e-01
-3.27792436e-01 -6.14113748e-01 -7.90377557e-01 -8.71693790e-01
-2.26151362e-01 2.59478986e-02 6.02473676e-01 6.13144994e-01
2.81419992e-01 2.38581061e-01 1.26216412e+00 -4.88481343e-01
-8.34424853e-01 -1.24771404e+00 -7.93377459e-01 7.82823324e-01
1.07741624e-01 -5.51164985e-01 -6.65556431e-01 -9.37958062e-02] | [9.489439964294434, 3.27461314201355] |
5068df03-747f-4a0d-9a8f-3c762dbe73f4 | unsupervised-representation-learning-for-gaze | 1911.06939 | null | https://arxiv.org/abs/1911.06939v4 | https://arxiv.org/pdf/1911.06939v4.pdf | Unsupervised Representation Learning for Gaze Estimation | Although automatic gaze estimation is very important to a large variety of application areas, it is difficult to train accurate and robust gaze models, in great part due to the difficulty in collecting large and diverse data (annotating 3D gaze is expensive and existing datasets use different setups). To address this issue, our main contribution in this paper is to propose an effective approach to learn a low dimensional gaze representation without gaze annotations, which to the best of our best knowledge, is the first work to do so. The main idea is to rely on a gaze redirection network and use the gaze representation difference of the input and target images (of the redirection network) as the redirection variable. A redirection loss in image domain allows the joint training of both the redirection network and the gaze representation network. In addition, we propose a warping field regularization which not only provides an explicit physical meaning to the gaze representations but also avoids redirection distortions. Promising results on few-shot gaze estimation (competitive results can be achieved with as few as <= 100 calibration samples), cross-dataset gaze estimation, gaze network pretraining, and another task (head pose estimation) demonstrate the validity of our framework. | ['Jean-Marc Odobez', 'Yu Yu'] | 2019-11-16 | unsupervised-representation-learning-for-gaze-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Yu_Unsupervised_Representation_Learning_for_Gaze_Estimation_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Yu_Unsupervised_Representation_Learning_for_Gaze_Estimation_CVPR_2020_paper.pdf | cvpr-2020-6 | ['gaze-redirection', 'head-pose-estimation'] | ['computer-vision', 'computer-vision'] | [ 1.88389763e-01 2.22223356e-01 -1.18090948e-02 -5.64590335e-01
-9.92150530e-02 -2.36662447e-01 5.02559185e-01 -5.06677389e-01
-3.78549814e-01 4.95293230e-01 9.84103829e-02 3.71110067e-02
-1.36527255e-01 -1.24156617e-01 -7.78758764e-01 -8.36119950e-01
2.98085481e-01 4.86048348e-02 1.97103918e-01 -2.53387332e-01
5.27330756e-01 2.63953775e-01 -2.10084152e+00 -2.81088233e-01
9.53318179e-01 1.06496894e+00 3.68598163e-01 4.24489200e-01
1.15318164e-01 6.35702968e-01 -5.08024335e-01 -1.20946467e-01
6.15812168e-02 -6.34365141e-01 -8.39285493e-01 -7.79614449e-02
7.96562910e-01 -2.11301133e-01 2.74530202e-01 9.39720094e-01
6.45721674e-01 3.87682766e-01 7.63952076e-01 -1.47210467e+00
-7.32431233e-01 2.31126472e-01 -9.65631962e-01 3.02444369e-01
5.05888045e-01 7.90134743e-02 8.28484356e-01 -6.47429943e-01
5.81847250e-01 1.06517696e+00 3.85523498e-01 1.02149510e+00
-1.17517900e+00 -9.57950652e-01 4.11077231e-01 4.28090096e-01
-1.36499500e+00 -6.28775120e-01 9.44749355e-01 -4.47218895e-01
5.35316110e-01 1.72220409e-01 4.03624266e-01 1.39828932e+00
-2.92051118e-02 7.36243486e-01 1.27079201e+00 -6.50047600e-01
-8.83541815e-03 2.86452234e-01 1.99636087e-01 5.47002196e-01
-2.79427338e-02 -1.10839307e-01 -7.58406758e-01 3.41165036e-01
3.56556088e-01 -1.43556856e-02 -7.68457472e-01 -5.87914765e-01
-8.45144153e-01 6.65883005e-01 8.55086446e-01 2.03257799e-01
-2.59375811e-01 1.97073501e-02 9.40585807e-02 2.28626668e-01
6.93131685e-01 3.27352077e-01 -1.24339819e-01 -1.93137914e-01
-8.55604529e-01 4.53126170e-02 6.28924191e-01 6.91021323e-01
7.08081782e-01 -1.64275140e-01 6.70037568e-02 7.39590108e-01
6.85177445e-01 6.66417956e-01 6.08067870e-01 -5.61649799e-01
4.29847747e-01 4.56458271e-01 -9.27311480e-02 -9.48151886e-01
-5.46779037e-01 5.97267319e-03 -5.64434528e-01 5.77480793e-01
5.88006318e-01 -1.76943079e-01 -7.83756435e-01 2.09019160e+00
4.05959755e-01 3.27183157e-01 -2.30343387e-01 1.29543126e+00
8.04676294e-01 3.60023648e-01 1.19020775e-01 -3.41205090e-01
1.36332011e+00 -7.96801805e-01 -9.49865758e-01 -3.10006917e-01
5.44283628e-01 -6.81256115e-01 1.31884384e+00 2.71276057e-01
-9.88877714e-01 -3.89001876e-01 -9.17955339e-01 -2.78200746e-01
-2.90150404e-01 -1.68881752e-02 5.88633895e-01 6.96341634e-01
-1.20094705e+00 1.79943696e-01 -7.28310406e-01 -7.15692580e-01
4.87794459e-01 6.58462048e-01 -4.13786441e-01 7.52365068e-02
-9.27171469e-01 1.11925745e+00 -1.10000417e-01 2.19936252e-01
-3.64243358e-01 -5.79132795e-01 -8.57182264e-01 3.31709906e-02
5.39098501e-01 -6.19215667e-01 9.00179803e-01 -1.43050623e+00
-1.65151381e+00 7.95237839e-01 -5.04349530e-01 -1.16564490e-01
1.96889147e-01 -3.27258021e-01 -1.44099683e-01 -1.69353068e-01
-1.48801640e-01 8.35511148e-01 1.25451565e+00 -1.17358577e+00
-4.50290203e-01 -7.01453149e-01 7.39939883e-02 4.34632540e-01
-4.14490610e-01 2.22727776e-01 -4.44218129e-01 -2.19704360e-01
-1.18751913e-01 -1.24837935e+00 4.09803033e-01 1.71306394e-02
-3.62237900e-01 -4.07676250e-01 9.50554192e-01 -5.22948384e-01
1.29530942e+00 -2.22628069e+00 5.14779985e-01 -4.94436659e-02
4.60049987e-01 2.04951331e-01 1.75135420e-03 -2.06109762e-01
-3.56026649e-01 -3.66547219e-02 -5.85042983e-02 -6.93689167e-01
-2.06557214e-02 -7.65058920e-02 -2.28969991e-01 6.06038749e-01
1.64218962e-01 9.15355682e-01 -6.30814373e-01 -3.34219962e-01
1.37050495e-01 8.15411866e-01 -5.15702903e-01 2.21792534e-01
9.03290808e-02 5.21799922e-01 -1.86840832e-01 2.80519724e-01
7.00354993e-01 -3.91506940e-01 -1.78440899e-01 -2.42278442e-01
-1.35363057e-01 1.92511797e-01 -9.17956650e-01 1.66441929e+00
-3.65467101e-01 1.05467272e+00 -9.91417915e-02 -5.51769912e-01
6.36877596e-01 1.17999844e-01 2.98144430e-01 -8.22876871e-01
6.22549951e-01 -1.19235128e-01 1.19241565e-01 -7.45613694e-01
4.76450652e-01 -1.15736544e-01 1.82238191e-01 8.71562779e-01
2.10189164e-01 2.93391377e-01 3.78210284e-02 -1.43032596e-01
4.43161577e-01 3.87461036e-01 1.53070301e-01 -1.43893734e-01
6.30132139e-01 -6.09673619e-01 2.16620877e-01 3.41626853e-02
-3.35115939e-01 7.86749303e-01 7.49554157e-01 -1.83037519e-01
-5.50112367e-01 -4.37819123e-01 -4.99332920e-02 1.45310187e+00
2.87380457e-01 -1.51433334e-01 -1.18443465e+00 -9.20018554e-01
-2.94123679e-01 6.75754070e-01 -1.25079155e+00 -2.51329213e-01
-5.04430294e-01 -6.54128313e-01 2.83658743e-01 4.24236447e-01
2.14090243e-01 -1.06731939e+00 -6.62036300e-01 -5.47840536e-01
-1.30033523e-01 -7.86941826e-01 -6.14006341e-01 6.10409863e-03
-5.39774597e-01 -1.30454028e+00 -8.20284724e-01 -5.19240975e-01
8.33958864e-01 4.74810362e-01 7.11286068e-01 -7.87610635e-02
1.29264757e-01 1.82770982e-01 -1.69950828e-01 -7.36968815e-01
2.73234770e-02 2.56661952e-01 1.92561805e-01 2.29396433e-01
6.64173484e-01 -3.67257148e-01 -4.87087667e-01 4.93281037e-01
-6.30284727e-01 -1.57296166e-01 3.40473384e-01 6.80114686e-01
2.25896284e-01 -6.62732005e-01 2.89990008e-01 -8.16910625e-01
7.14155972e-01 -3.59556794e-01 -7.33509839e-01 2.16712892e-01
-7.40649581e-01 2.53489345e-01 1.41096547e-01 -5.58152914e-01
-1.08167303e+00 -1.75797924e-01 8.02159403e-03 -6.90694690e-01
-2.90421873e-01 1.87329412e-01 -1.06217869e-01 -2.85843819e-01
9.58593309e-01 -2.31452808e-01 3.93964589e-01 -4.50886607e-01
2.60205090e-01 5.69290221e-01 -3.57723795e-02 2.67357789e-02
5.68434238e-01 4.66364175e-01 1.40907973e-01 -7.48566389e-01
-9.65353608e-01 -2.48578370e-01 -8.55890453e-01 -2.72818536e-01
9.78761256e-01 -6.21466815e-01 -1.14037836e+00 4.98077065e-01
-8.91537666e-01 -2.75142282e-01 -1.90682393e-02 5.32801628e-01
-3.86282712e-01 1.50341272e-01 1.00562215e-01 -7.69616663e-01
-3.08074206e-01 -1.30089355e+00 1.10544801e+00 4.37738925e-01
-1.95082754e-01 -1.02504504e+00 5.46828769e-02 3.06332707e-01
5.24270117e-01 1.53590232e-01 4.77981955e-01 -3.08770478e-01
-5.35143793e-01 1.18320845e-01 -4.17994767e-01 2.75842160e-01
2.52950132e-01 1.07226156e-01 -1.40843749e+00 -1.72196999e-01
1.40474141e-01 -5.01574397e-01 8.23837042e-01 6.29902124e-01
9.71938252e-01 1.71638072e-01 -3.68464768e-01 7.45588660e-01
1.12726355e+00 -2.30697975e-01 6.89838350e-01 1.78770974e-01
1.00901151e+00 9.43092704e-01 5.83511293e-01 1.04172349e-01
5.80786288e-01 8.98637414e-01 6.38085842e-01 4.52455133e-02
-2.45201081e-01 3.68825742e-03 2.72363544e-01 4.74413335e-01
-4.18232352e-01 -3.47943336e-01 -8.81170332e-01 2.89276451e-01
-1.60283589e+00 -1.03829682e+00 -1.91606402e-01 2.39782071e+00
5.86388230e-01 -1.44356370e-01 2.09397003e-01 -7.27593759e-03
5.31971395e-01 2.28138074e-01 -5.54742754e-01 -3.21690559e-01
1.94087818e-01 1.19164132e-01 1.82500169e-01 4.00945038e-01
-1.00363648e+00 7.95641840e-01 5.68867540e+00 3.06760132e-01
-1.87740850e+00 1.81969061e-01 1.91809416e-01 -4.59339023e-01
4.76494245e-02 -2.62394577e-01 -8.81484926e-01 6.82493508e-01
9.29747164e-01 1.56191468e-01 4.49781060e-01 7.44241238e-01
1.40081748e-01 -1.97405592e-01 -1.13461328e+00 1.26776791e+00
7.52406240e-01 -7.25296974e-01 -6.06383026e-01 2.81601965e-01
3.50578189e-01 1.33839577e-01 4.57381159e-01 1.22625940e-01
-3.94554585e-01 -1.06165695e+00 6.14737153e-01 8.02798986e-01
9.56124008e-01 -5.80717862e-01 6.35463893e-01 2.06342876e-01
-7.06449032e-01 -1.36804029e-01 -1.97253823e-01 -4.78390045e-02
7.97955394e-02 -1.60132959e-01 -7.76924551e-01 1.81076840e-01
8.54284465e-01 8.05492640e-01 -9.32774782e-01 9.70318913e-01
-3.86681139e-01 4.81615007e-01 -3.59471381e-01 -1.71067759e-01
-1.19922608e-01 -1.88430063e-02 3.68034720e-01 5.93449056e-01
1.47370696e-01 -1.12553000e-01 -5.22903502e-01 7.87555218e-01
-1.46387175e-01 7.50850663e-02 -6.02482975e-01 2.82242000e-01
1.17620774e-01 1.12174904e+00 -4.09339219e-01 3.04179877e-01
-3.66299987e-01 7.81431615e-01 5.94309747e-01 5.31471014e-01
-7.56428719e-01 -3.93941492e-01 8.15851927e-01 2.89474547e-01
2.59889305e-01 4.24457528e-02 -1.79990292e-01 -1.26411307e+00
3.31554882e-04 -6.39208019e-01 8.60306025e-02 -1.07768166e+00
-1.07763970e+00 6.37160897e-01 6.88139945e-02 -1.27857208e+00
-6.16401017e-01 -5.71952045e-01 -4.88409042e-01 1.24186087e+00
-1.79289699e+00 -1.24750006e+00 -8.02850664e-01 8.01114917e-01
3.09437901e-01 -4.59414124e-02 7.98026025e-01 2.62455255e-01
-7.46737897e-01 9.42088127e-01 -9.80975479e-02 -2.38538295e-01
1.16733897e+00 -1.16442668e+00 9.97955725e-02 6.75497472e-01
-1.16480559e-01 7.81242669e-01 7.74184406e-01 -1.31283045e-01
-1.21293080e+00 -4.90155995e-01 8.46876681e-01 -7.99554229e-01
4.89414781e-01 -5.70122719e-01 -1.12996185e+00 6.86287522e-01
3.95447791e-01 4.94161248e-02 7.26657629e-01 5.50214410e-01
-2.90960997e-01 -1.09571733e-01 -9.14481699e-01 5.15847147e-01
8.75284076e-01 -4.47776526e-01 -5.80485106e-01 1.20041892e-03
3.13951343e-01 -6.13132298e-01 -3.81288320e-01 5.52487969e-02
6.44339085e-01 -1.07075655e+00 6.82378054e-01 -3.87482345e-01
1.77233309e-01 -2.28267968e-01 3.04353893e-01 -1.20000780e+00
4.37387191e-02 -4.28231120e-01 -2.96666473e-01 1.21464336e+00
1.84770972e-01 -6.51938558e-01 7.03239560e-01 7.12389410e-01
2.55327493e-01 -6.16280377e-01 -6.82557046e-01 -3.60962272e-01
-1.86066255e-01 -3.24826211e-01 6.30024850e-01 8.92781556e-01
-4.48362418e-02 7.52490401e-01 -5.39374828e-01 1.02716975e-01
5.09402275e-01 -5.94011620e-02 9.78992581e-01 -1.58303022e+00
1.12608671e-01 -4.41367328e-01 -4.77457017e-01 -1.01804149e+00
4.81053233e-01 -3.72767985e-01 9.75562185e-02 -1.12684083e+00
-6.52518049e-02 -5.13388157e-01 -2.85985678e-01 4.49403375e-01
-2.58727014e-01 3.56107175e-01 3.21180731e-01 3.46072048e-01
-5.33550203e-01 5.25040507e-01 1.37667549e+00 1.96142033e-01
-3.94346446e-01 2.33500138e-01 -9.70329225e-01 5.85564554e-01
4.59814578e-01 -3.62674922e-01 -7.55517006e-01 -6.06465161e-01
3.70643586e-01 -3.75369340e-01 4.74328130e-01 -8.48877728e-01
3.63642305e-01 -1.74506065e-02 1.50351599e-01 -3.21562886e-01
5.92825294e-01 -8.11338603e-01 -2.52885163e-01 -1.38362586e-01
-2.97407061e-01 -2.23874241e-01 1.60005599e-01 5.17112970e-01
-1.14411391e-01 -3.85176659e-01 7.92639375e-01 3.50071371e-01
-6.31930530e-01 1.57430589e-01 2.87617087e-01 -5.06267883e-02
1.03809190e+00 -4.28040683e-01 -3.39637607e-01 -3.66892070e-01
-4.96044904e-01 1.61026537e-01 8.01276684e-01 8.43112946e-01
3.19685608e-01 -1.02678430e+00 -2.98262775e-01 5.31989157e-01
1.82173595e-01 8.37300718e-02 2.30808839e-01 1.22376418e+00
-6.59844652e-02 4.16708916e-01 -3.89227629e-01 -8.90434384e-01
-1.58523560e+00 7.47108221e-01 3.62008691e-01 2.73517758e-01
-2.24399492e-01 1.15142047e+00 4.74326432e-01 -1.43732846e-01
2.79989958e-01 -1.90994695e-01 -8.42628896e-01 4.58679020e-01
8.56296718e-01 2.51490116e-01 3.57646607e-02 -1.25441360e+00
-3.79773915e-01 9.41666961e-01 -1.72817811e-01 7.99492821e-02
1.29398286e+00 -5.66545904e-01 -6.82313554e-03 5.89218080e-01
1.30542624e+00 -9.26899835e-02 -1.33992052e+00 -3.11588664e-02
-3.16054881e-01 -6.12905979e-01 2.23401651e-01 -6.53424382e-01
-1.07261646e+00 1.28531647e+00 1.01059878e+00 1.29217625e-01
1.26395392e+00 1.20049134e-01 3.59364629e-01 2.58891195e-01
1.60823315e-01 -8.38543117e-01 2.53805495e-03 4.08872813e-01
9.16834354e-01 -1.57824838e+00 -1.06230810e-01 -2.46694908e-01
-9.25575078e-01 7.99413741e-01 7.64566302e-01 -1.72484983e-02
6.91301823e-01 -2.79571593e-01 2.33840153e-01 -3.10475528e-01
-5.79773843e-01 -4.14462298e-01 6.70019150e-01 7.02178836e-01
6.16478086e-01 -4.66540366e-01 -1.69444859e-01 1.95888653e-01
-1.30330279e-01 1.08189724e-01 4.02006626e-01 7.19973564e-01
-2.08125949e-01 -9.89215374e-01 -4.37256545e-01 3.10533345e-01
-4.71793979e-01 1.09666690e-01 -2.55735338e-01 8.53829205e-01
2.47936696e-03 7.68273234e-01 1.21836469e-01 -2.35787392e-01
3.95389706e-01 2.51451671e-01 6.15783036e-01 -5.31202257e-01
-4.90759254e-01 2.09063757e-02 -3.96629691e-01 -6.46951556e-01
-8.12759042e-01 -7.69278467e-01 -9.12232280e-01 -2.50228256e-01
-8.12084436e-01 -1.74711913e-01 9.22792912e-01 1.02919519e+00
4.75353718e-01 4.62577909e-01 4.72576827e-01 -1.19895589e+00
-3.44291359e-01 -1.22340405e+00 -5.95219731e-01 5.85243404e-01
5.70508897e-01 -1.22817981e+00 -4.54554230e-01 1.85468167e-01] | [14.111705780029297, 0.05675153061747551] |
03ec8d2e-fe09-40f0-9af0-bea69440e55b | cross-lingual-transfer-of-cognitive | 2302.12695 | null | https://arxiv.org/abs/2302.12695v2 | https://arxiv.org/pdf/2302.12695v2.pdf | Cross-Lingual Transfer of Cognitive Processing Complexity | When humans read a text, their eye movements are influenced by the structural complexity of the input sentences. This cognitive phenomenon holds across languages and recent studies indicate that multilingual language models utilize structural similarities between languages to facilitate cross-lingual transfer. We use sentence-level eye-tracking patterns as a cognitive indicator for structural complexity and show that the multilingual model XLM-RoBERTa can successfully predict varied patterns for 13 typologically diverse languages, despite being fine-tuned only on English data. We quantify the sensitivity of the model to structural complexity and distinguish a range of complexity characteristics. Our results indicate that the model develops a meaningful bias towards sentence length but also integrates cross-lingual differences. We conduct a control experiment with randomized word order and find that the model seems to additionally capture more complex structural information. | ['Lisa Beinborn', 'Nora Hollenstein', 'Charlotte Pouw'] | 2023-02-24 | null | null | null | null | ['cross-lingual-transfer'] | ['natural-language-processing'] | [-3.42751980e-01 -3.42464745e-01 -1.96894333e-01 -3.60217571e-01
-4.69662428e-01 -8.58137608e-01 8.89859140e-01 5.09580731e-01
-1.00906837e+00 3.28719795e-01 4.74077761e-01 -7.98112929e-01
-1.22048780e-01 -3.67152900e-01 -4.70111370e-01 -2.04943009e-02
2.20275268e-01 3.32389891e-01 1.51115015e-01 -4.96630996e-01
8.27888370e-01 3.05581599e-01 -1.42478859e+00 3.37268025e-01
1.33586335e+00 -5.24210893e-02 6.94096923e-01 5.63961208e-01
-2.09822744e-01 3.85927886e-01 -5.75064898e-01 -4.55688417e-01
2.44925961e-01 -4.66573060e-01 -7.38342762e-01 -1.10860020e-01
8.14972460e-01 1.66270033e-01 -2.26883128e-01 9.69667375e-01
4.96394306e-01 -1.22129500e-01 7.10748494e-01 -5.06342709e-01
-1.27923965e+00 6.55539751e-01 -3.81642252e-01 8.92101943e-01
9.65291977e-01 5.01413524e-01 1.14105690e+00 -8.68522227e-01
8.60208511e-01 1.51444173e+00 5.86402655e-01 2.67531782e-01
-1.44081628e+00 -4.72492933e-01 4.49812263e-01 3.69374990e-01
-1.19764745e+00 -5.58071792e-01 5.35625637e-01 -7.44448900e-01
1.42930746e+00 1.16572350e-01 9.60714877e-01 8.56732488e-01
8.12665701e-01 1.92243591e-01 2.05208135e+00 -8.68734181e-01
-2.66315967e-01 3.52416486e-01 4.85775024e-01 6.16471648e-01
4.40970898e-01 4.14027125e-01 -9.85678256e-01 4.33063358e-01
2.68170446e-01 -5.50253749e-01 -2.73370832e-01 2.74954140e-01
-1.35352814e+00 7.42827833e-01 1.61026001e-01 8.64278018e-01
4.56529669e-02 -5.90027630e-01 2.70038664e-01 9.45875466e-01
4.47892398e-01 8.03367734e-01 -4.74102467e-01 -1.44523636e-01
-7.92572200e-01 1.60884216e-01 6.99636638e-01 7.54866123e-01
5.64055443e-01 2.11250596e-02 -2.64112681e-01 1.02737260e+00
3.92290831e-01 7.84214675e-01 7.69084930e-01 -5.70145130e-01
7.88612068e-01 4.20958906e-01 -2.87325889e-01 -7.12027609e-01
-7.74972200e-01 -1.54848889e-01 2.35140696e-02 3.63171160e-01
7.52657115e-01 7.97886848e-02 -4.46582913e-01 1.93189263e+00
-1.09723985e-01 -1.16242039e+00 -3.24237287e-01 5.41391969e-01
3.77828568e-01 3.90578330e-01 5.47654927e-01 -4.83328611e-01
1.45552862e+00 -5.36566913e-01 -6.00496531e-01 -5.93640029e-01
1.09061301e+00 -9.97769237e-01 1.85318017e+00 2.18514100e-01
-1.33159721e+00 -7.16909170e-01 -1.01518142e+00 -4.43049997e-01
-5.69975078e-01 -2.13902712e-01 4.06752378e-01 8.89060736e-01
-1.57458985e+00 3.89367282e-01 -2.93693006e-01 -6.78199172e-01
-1.45204470e-01 2.71058176e-02 -4.42384958e-01 -3.78538072e-02
-1.13784921e+00 1.66826606e+00 4.52418685e-01 -2.36964270e-01
-1.42557546e-01 -4.00884241e-01 -8.56743157e-01 -1.79590359e-01
-2.67823815e-01 -6.51098192e-01 1.09530675e+00 -1.28584814e+00
-1.39053082e+00 1.48785627e+00 -4.93775636e-01 9.25339609e-02
9.89433601e-02 2.27759890e-02 -5.70857286e-01 -4.73226756e-02
2.84985304e-01 5.94014823e-01 6.04789674e-01 -7.07293093e-01
-4.68207330e-01 -4.09367383e-01 6.00086041e-02 5.58976531e-01
-3.22686017e-01 6.49397016e-01 1.07652381e-01 -8.88620496e-01
-1.80705234e-01 -9.64993238e-01 3.11414003e-01 -1.95066586e-01
5.66740111e-02 -4.15312737e-01 -1.98681086e-01 -9.43182468e-01
1.62023330e+00 -1.99029541e+00 2.22144678e-01 8.79948959e-02
2.24745423e-01 -2.40695789e-01 -4.92947161e-01 5.78067422e-01
-2.06024554e-02 3.91380131e-01 -1.41697392e-01 -7.21478388e-02
3.41311395e-01 -2.39641115e-01 3.63060504e-01 4.78540659e-01
1.49007579e-02 1.28645515e+00 -6.70619190e-01 -4.95350480e-01
-2.05270097e-01 6.36152104e-02 -8.05064559e-01 -1.97256476e-01
8.26016515e-02 3.01018655e-01 3.70119303e-01 2.76182532e-01
4.80197102e-01 -2.84559820e-02 1.91693768e-01 3.22264433e-01
-6.51384950e-01 6.17440224e-01 -3.95644665e-01 1.53027046e+00
-6.21916592e-01 1.21700430e+00 -2.79547751e-01 -3.18324894e-01
5.51984906e-01 -5.92754930e-02 -5.07007301e-01 -1.40324509e+00
1.75379366e-01 1.98656797e-01 1.18617070e+00 -6.70645416e-01
4.00999665e-01 -4.63153869e-01 -7.03629479e-02 7.02562749e-01
-8.30698162e-02 -1.55786574e-01 6.42019510e-01 3.07061840e-02
5.37715554e-01 -2.31775478e-01 4.93815035e-01 -9.79288280e-01
5.06516635e-01 -1.47233650e-01 2.49921203e-01 7.46439695e-01
-2.78534085e-01 -6.44653440e-02 3.06083590e-01 -1.71212211e-01
-1.20708632e+00 -1.07092571e+00 -4.84451860e-01 1.55861628e+00
-2.47540012e-01 -3.50174099e-01 -8.99293184e-01 -1.60872757e-01
-1.78230226e-01 1.30997729e+00 -6.03940785e-01 -2.42089018e-01
-5.14499128e-01 -5.50423086e-01 2.50637680e-01 3.91563237e-01
1.92182511e-02 -1.22518229e+00 -5.44570267e-01 -9.85943154e-02
-5.16604371e-02 -1.15434277e+00 -8.85372818e-01 -3.25232655e-01
-8.83131742e-01 -8.17106187e-01 -4.62432384e-01 -1.01353133e+00
5.72969317e-01 2.29301289e-01 1.31227148e+00 2.62910962e-01
-1.82006091e-01 4.97317314e-01 -2.05530897e-01 -4.65142757e-01
-6.64174080e-01 2.34002382e-01 3.11204106e-01 -7.10434914e-01
7.96287000e-01 -2.34808758e-01 -2.19977304e-01 2.30424441e-02
-6.27889335e-01 1.10803641e-01 6.48930252e-01 3.62850457e-01
-8.58555809e-02 -5.17668188e-01 3.52476329e-01 -6.82674348e-01
1.21413445e+00 -2.30872080e-01 -6.16982877e-01 3.31275135e-01
-9.01097000e-01 2.39138544e-01 2.75289327e-01 -5.56995153e-01
-1.04786527e+00 -6.29736364e-01 2.83996928e-02 5.43347776e-01
-3.22203994e-01 7.61000037e-01 8.77218544e-02 -2.05337122e-01
7.97546387e-01 3.48327816e-01 -6.20458275e-02 -2.67048687e-01
1.90607786e-01 3.82468253e-01 -1.02304285e-02 -7.14994252e-01
8.60079408e-01 -1.46737307e-01 -3.92657310e-01 -1.27525914e+00
-6.75944626e-01 3.00158318e-02 -9.41551507e-01 -2.43022338e-01
9.83798921e-01 -1.07908309e+00 -5.90237379e-01 4.94654477e-01
-1.27579892e+00 -5.61525166e-01 1.47929356e-01 8.34185064e-01
-3.65090489e-01 3.58610511e-01 -8.09420705e-01 -3.84010911e-01
-4.74554189e-02 -9.58579421e-01 6.24643028e-01 -5.80171943e-02
-6.85951173e-01 -1.45608521e+00 3.42145801e-01 8.95713419e-02
4.04864669e-01 -3.72819811e-01 1.34804201e+00 -6.72846854e-01
-4.40218121e-01 1.84700295e-01 -1.56877249e-01 -1.42651349e-01
4.87431325e-02 7.14599267e-02 -4.63157594e-01 -2.47954950e-01
1.76686287e-01 -1.12004757e-01 6.84401035e-01 3.94026220e-01
2.92996764e-01 -2.56695479e-01 1.34365618e-01 4.64330524e-01
1.22004414e+00 2.61688605e-03 4.77183431e-01 4.78102803e-01
6.31554246e-01 8.78484428e-01 1.02004349e-01 -1.80307195e-01
8.85217607e-01 6.21607244e-01 -6.62026346e-01 3.90763313e-01
-1.38928488e-01 -1.05335355e-01 9.96650755e-01 1.52442789e+00
1.86887812e-02 -2.50255894e-02 -1.16983390e+00 4.20645952e-01
-8.81087959e-01 -1.00305629e+00 -2.52562881e-01 2.27634597e+00
1.20288348e+00 4.41832781e-01 3.58799696e-01 7.44006634e-02
4.78318989e-01 1.52164787e-01 -1.43637404e-01 -9.51422393e-01
-4.28545952e-01 5.64722978e-02 2.63679236e-01 1.00968361e+00
-4.04846579e-01 1.34618878e+00 7.99332714e+00 4.36080277e-01
-9.42804694e-01 3.90716456e-02 1.47702172e-01 -5.68414517e-02
-5.75141013e-01 -2.18588844e-01 -7.82210350e-01 4.35243815e-01
1.35913324e+00 -6.75806880e-01 6.40437365e-01 6.36054426e-02
3.93842369e-01 -5.08768678e-01 -1.35313046e+00 7.01779723e-01
3.91031027e-01 -6.86816156e-01 7.66488612e-02 -7.98839182e-02
6.58492267e-01 1.07204542e-01 1.60248488e-01 2.78516144e-01
3.55685875e-02 -1.00957489e+00 9.45909560e-01 7.29468524e-01
8.66988659e-01 -4.37358409e-01 5.47764190e-02 4.68339920e-01
-9.14227605e-01 -8.07329193e-02 -3.01188469e-01 -7.39879012e-01
1.19298428e-01 4.83924262e-02 -3.40835482e-01 -2.63935447e-01
4.42386359e-01 4.65261221e-01 -1.30781102e+00 8.46742392e-01
-3.78299028e-01 3.81428689e-01 1.38190147e-02 -2.13259801e-01
-8.67192075e-02 -2.11992607e-01 5.38186073e-01 1.34124815e+00
2.56754249e-01 -1.06230006e-01 -2.09703386e-01 6.95801973e-01
2.33942226e-01 7.66290426e-01 -8.66980076e-01 -2.62023509e-01
4.24469560e-01 6.78620994e-01 -7.46355712e-01 -8.12471136e-02
-9.51500237e-01 1.00241196e+00 7.97371686e-01 4.59502429e-01
-3.73525470e-01 -2.47923687e-01 3.74828219e-01 2.89477587e-01
4.77584787e-02 -8.39563489e-01 -8.06081414e-01 -1.29495907e+00
9.49547887e-02 -1.05010128e+00 -2.10193489e-02 -9.21964288e-01
-1.19480395e+00 4.78287309e-01 -7.84259010e-03 -7.40397096e-01
-1.14448711e-01 -9.96306419e-01 -3.47029567e-01 1.30005574e+00
-1.32585001e+00 -9.13303196e-01 5.27660623e-02 5.77405512e-01
8.02342057e-01 -2.14432701e-01 5.90175331e-01 1.69717558e-02
-3.46973807e-01 8.62674534e-01 -1.89488798e-01 -5.78184538e-02
1.00826573e+00 -1.31912422e+00 6.25010967e-01 8.88041854e-01
4.40343380e-01 1.11152279e+00 7.20411062e-01 -8.43076348e-01
-9.64383900e-01 -4.63559866e-01 1.75987089e+00 -9.39153552e-01
9.23967779e-01 -4.31515783e-01 -8.25477064e-01 7.65532613e-01
9.05116379e-01 -7.96107650e-01 8.65022242e-01 4.14841175e-01
-6.95404947e-01 3.33842367e-01 -6.28401279e-01 9.81225431e-01
1.43569255e+00 -1.11346757e+00 -1.27128279e+00 1.26372099e-01
8.06353688e-01 -1.59975663e-02 -7.50012338e-01 -1.19124398e-01
5.02178848e-01 -1.16110122e+00 5.60732901e-01 -5.82923055e-01
4.81156796e-01 -3.62537242e-02 1.32355958e-01 -1.74971199e+00
-9.85405087e-01 -3.47537279e-01 4.15811092e-01 8.01307917e-01
7.18124509e-01 -8.71228337e-01 -1.78545505e-01 5.17228723e-01
-5.24132550e-02 -5.37799820e-02 -8.80780160e-01 -9.03599858e-01
7.18836665e-01 -3.11662227e-01 7.94918463e-03 1.01320219e+00
3.34148914e-01 6.72551930e-01 3.64379704e-01 -6.25673309e-02
3.68787944e-01 -2.35171199e-01 2.64554292e-01 -1.63134611e+00
-6.91222176e-02 -1.09973300e+00 -1.18546516e-01 -6.73856437e-01
6.14874482e-01 -1.21497369e+00 -2.88199931e-01 -1.15896189e+00
2.61381090e-01 7.59824291e-02 3.06962486e-02 2.91298069e-02
-4.69838649e-01 -1.92598179e-02 6.59180224e-01 1.56428665e-01
-4.43177015e-01 2.54487693e-01 1.35766184e+00 2.11864397e-01
-2.66696066e-01 -2.95686394e-01 -1.07191801e+00 7.87197828e-01
7.75237679e-01 -3.86363983e-01 -3.77680361e-01 -8.79705489e-01
7.26423860e-01 -2.74560153e-01 1.59358039e-01 -9.40795362e-01
1.57883346e-01 -3.05515647e-01 3.82472396e-01 -5.98307960e-02
-2.49849007e-01 -5.55319130e-01 -3.73902470e-01 5.00865281e-01
-4.99882042e-01 8.92050266e-01 6.59942985e-01 1.56854421e-01
7.96251521e-02 -2.42587060e-01 7.65290797e-01 -2.42386125e-02
-5.29580414e-01 -1.63548172e-01 -8.16244304e-01 6.29896462e-01
6.96096241e-01 -5.04881501e-01 -2.95176655e-01 -1.07248858e-01
-6.34032428e-01 -5.07726148e-02 7.62835383e-01 7.74390876e-01
6.33542513e-05 -1.30005038e+00 -9.09387708e-01 1.32015526e-01
2.85967618e-01 -1.08096993e+00 3.89650278e-02 9.18540835e-01
-4.73066270e-01 7.87883341e-01 -3.78746510e-01 -4.18223083e-01
-1.22141266e+00 8.11996341e-01 2.23189846e-01 6.23424165e-02
-1.35028869e-01 6.22094929e-01 2.60462254e-01 -3.46371621e-01
-2.17217237e-01 -3.76085609e-01 -4.08225447e-01 5.20246744e-01
6.27091289e-01 3.00398648e-01 -2.45533615e-01 -9.92300034e-01
-2.22965166e-01 1.11552060e+00 -3.08182925e-01 -4.17504668e-01
8.03514242e-01 -6.18966520e-01 -3.59193802e-01 1.23194480e+00
9.52429295e-01 5.72435558e-01 -7.73931980e-01 -4.06662643e-01
3.87596369e-01 -3.48552883e-01 -2.35881925e-01 -8.09413135e-01
-3.92075688e-01 8.03861082e-01 5.83469987e-01 -1.26892239e-01
7.89175749e-01 2.65016854e-02 1.08452968e-01 6.07571304e-01
2.58288622e-01 -1.49329555e+00 -9.51635465e-02 9.75945473e-01
1.12383580e+00 -1.05753791e+00 1.17815044e-02 -4.03163992e-02
-6.71542883e-01 8.78437996e-01 8.14242482e-01 4.80300859e-02
5.19783854e-01 3.62576917e-02 1.92279771e-01 -1.04086228e-01
-9.40418005e-01 -3.80369365e-01 5.89247465e-01 6.90287828e-01
1.00393379e+00 -3.87240164e-02 -1.07408297e+00 4.03610140e-01
-8.84424984e-01 -3.72360021e-01 5.10870159e-01 8.11712801e-01
-6.56661749e-01 -1.15735066e+00 -5.46110570e-01 3.44914854e-01
-2.77843982e-01 -8.46585810e-01 -4.19924587e-01 1.01702249e+00
-3.98515873e-02 6.54168129e-01 3.77582908e-01 -1.37826065e-02
4.64041203e-01 5.15025854e-01 8.93511415e-01 -8.28653693e-01
-8.28513682e-01 -3.48749347e-02 -3.48302685e-02 -2.44663075e-01
-2.91814476e-01 -1.02569687e+00 -8.30485225e-01 -5.12352765e-01
1.52112339e-02 -1.91432014e-01 3.84409487e-01 1.10821474e+00
1.31618723e-01 2.79484421e-01 2.82603949e-01 -6.29884958e-01
-2.46002942e-01 -1.29394460e+00 -4.96609360e-01 3.91193569e-01
2.77202249e-01 -5.38493395e-01 -3.86511743e-01 1.97365999e-01] | [10.80453109741211, 9.92039680480957] |
5d3683da-66be-46ce-8257-c4fd46a0e74d | p2m-detrack-processing-in-pixel-in-memory-for | 2205.14285 | null | https://arxiv.org/abs/2205.14285v1 | https://arxiv.org/pdf/2205.14285v1.pdf | P2M-DeTrack: Processing-in-Pixel-in-Memory for Energy-efficient and Real-Time Multi-Object Detection and Tracking | Today's high resolution, high frame rate cameras in autonomous vehicles generate a large volume of data that needs to be transferred and processed by a downstream processor or machine learning (ML) accelerator to enable intelligent computing tasks, such as multi-object detection and tracking. The massive amount of data transfer incurs significant energy, latency, and bandwidth bottlenecks, which hinders real-time processing. To mitigate this problem, we propose an algorithm-hardware co-design framework called Processing-in-Pixel-in-Memory-based object Detection and Tracking (P2M-DeTrack). P2M-DeTrack is based on a custom faster R-CNN-based model that is distributed partly inside the pixel array (front-end) and partly in a separate FPGA/ASIC (back-end). The proposed front-end in-pixel processing down-samples the input feature maps significantly with judiciously optimized strided convolution and pooling. Compared to a conventional baseline design that transfers frames of RGB pixels to the back-end, the resulting P2M-DeTrack designs reduce the data bandwidth between sensor and back-end by up to 24x. The designs also reduce the sensor and total energy (obtained from in-house circuit simulations at Globalfoundries 22nm technology node) per frame by 5.7x and 1.14x, respectively. Lastly, they reduce the sensing and total frame latency by an estimated 1.7x and 3x, respectively. We evaluate our approach on the multi-object object detection (tracking) task of the large-scale BDD100K dataset and observe only a 0.5% reduction in the mean average precision (0.8% reduction in the identification F1 score) compared to the state-of-the-art. | ['Peter A. Beerel', 'Akhilesh R. Jaiswal', 'Ajey P. Jacob', 'Wael Abd-Almageed', 'Andrew Schmidt', 'Ravi T. Lakkireddy', 'Shunlin Lu', 'Mulin Tian', 'Zixu Wang', 'Zeyu Liu', 'Joe Mathai', 'Zihan Yin', 'Souvik Kundu', 'Gourav Datta'] | 2022-05-28 | null | null | null | null | ['total-energy'] | ['miscellaneous'] | [ 5.09704828e-01 -1.63487177e-02 -4.56152521e-02 -2.85421550e-01
-5.85621893e-01 -4.11030233e-01 4.42142367e-01 1.26887545e-01
-9.30497408e-01 1.23150043e-01 -6.43481016e-01 -4.85479116e-01
5.22903621e-01 -7.44662941e-01 -1.05384302e+00 -6.89830065e-01
3.36011261e-01 -1.83466107e-01 6.22848690e-01 2.36649841e-01
7.92461485e-02 4.39300776e-01 -1.70166314e+00 2.49681681e-01
3.49384010e-01 1.82370830e+00 2.64997661e-01 7.12577403e-01
2.96125352e-01 6.26726925e-01 -5.32291293e-01 -2.97507942e-01
4.95309353e-01 9.07279551e-02 2.40386352e-02 -3.55749786e-01
7.83492684e-01 -8.10614467e-01 -5.50088882e-01 1.11310029e+00
4.78028089e-01 -4.50896055e-01 1.74634635e-01 -1.21617770e+00
-1.71562314e-01 2.74058282e-01 -9.94870722e-01 1.42779469e-01
-4.61354464e-01 6.02729142e-01 1.04146183e-01 -7.69300163e-01
1.41844109e-01 1.17229784e+00 7.62870371e-01 6.45246148e-01
-1.05892861e+00 -1.21606183e+00 -3.99244249e-01 1.20788336e-01
-1.27046931e+00 -7.17051268e-01 4.45557535e-01 -1.95036337e-01
1.31828570e+00 3.13495323e-02 6.83709979e-01 6.21066093e-01
6.06958330e-01 3.41108054e-01 8.53861272e-01 -1.58015847e-01
3.25912058e-01 9.44861863e-03 3.65947276e-01 7.22187400e-01
9.29639280e-01 2.22560048e-01 -5.61941445e-01 1.68484390e-01
9.17056501e-01 1.45820498e-01 8.05200711e-02 3.36751074e-01
-1.05771577e+00 4.82824206e-01 5.59733272e-01 -3.78722310e-01
-2.79917985e-01 9.73201692e-01 5.28478205e-01 3.18835899e-02
-1.32567972e-01 -5.24377413e-02 -3.94486159e-01 -7.80307129e-02
-9.65337336e-01 -7.85192661e-03 3.09983820e-01 1.17238557e+00
8.67596805e-01 1.58060968e-01 -7.01088086e-02 8.38846862e-02
7.02348590e-01 1.02169311e+00 4.01307404e-01 -8.09131265e-01
6.19925439e-01 8.34516108e-01 1.52168766e-01 -7.43454337e-01
-6.30290866e-01 -1.40404418e-01 -8.60294402e-01 6.37852013e-01
3.65169704e-01 -3.43179107e-01 -1.05376840e+00 1.36092341e+00
3.83409619e-01 8.84836763e-02 2.59664774e-01 1.07338023e+00
8.06102216e-01 8.77575576e-01 1.92286059e-01 -6.30321652e-02
1.83225107e+00 -8.90960515e-01 -5.54015040e-01 -4.91811216e-01
2.58719295e-01 -6.30333424e-01 6.55526042e-01 1.33655146e-01
-9.71628547e-01 -9.79529560e-01 -1.69032609e+00 -3.79619628e-01
-1.90752894e-01 6.89831257e-01 3.52117270e-01 8.54573250e-01
-8.81094694e-01 2.81832963e-01 -1.19219327e+00 -1.95218846e-01
6.53667033e-01 1.00282300e+00 4.19665761e-02 1.77280009e-01
-5.66096544e-01 5.83545804e-01 1.76360667e-01 1.82753757e-01
-8.33791018e-01 -1.13429952e+00 -4.78624314e-01 6.01750147e-03
2.14072660e-01 -6.67284846e-01 1.25267887e+00 -9.75361586e-01
-1.72464311e+00 5.16324520e-01 -5.24965636e-02 -9.83815014e-01
2.79715508e-01 -1.78394839e-01 -4.82169271e-01 1.92123204e-01
-3.74474078e-01 8.94010425e-01 1.06579483e+00 -5.67711473e-01
-9.17232752e-01 -5.47968090e-01 -1.58589348e-01 -2.97846317e-01
-4.22220826e-01 5.44671789e-02 -6.95717692e-01 -1.89911619e-01
-5.95231727e-02 -1.04763341e+00 -3.56675774e-01 4.83604759e-01
5.43457344e-02 1.87810943e-01 1.44888175e+00 -1.81057200e-01
8.65252316e-01 -2.38504958e+00 -7.70764947e-01 -1.57098010e-01
2.81309068e-01 7.29815304e-01 1.18386596e-01 -3.53887200e-01
2.72728801e-01 -3.06043297e-01 1.91498145e-01 -3.48719865e-01
-3.72457832e-01 -9.61929336e-02 -1.73204675e-01 7.85924554e-01
3.47650439e-01 9.66037095e-01 -3.31751049e-01 -1.15047023e-01
3.93216491e-01 6.39192522e-01 -3.87181491e-01 -2.60692447e-01
-5.23014441e-02 1.10700481e-01 -3.42859924e-01 5.79440713e-01
1.02427757e+00 -3.46464604e-01 1.76908687e-01 -8.55443954e-01
-4.70764965e-01 1.97459489e-01 -9.88844991e-01 1.47281456e+00
-3.47446948e-01 1.06221092e+00 2.98196554e-01 -4.17159826e-01
9.23532784e-01 -1.48311421e-01 2.10623145e-01 -1.01716828e+00
6.68608606e-01 2.73586482e-01 -5.10629714e-02 -2.32007667e-01
8.38469088e-01 1.07748426e-01 -3.96188080e-01 -5.93767688e-02
-2.02823982e-01 6.04782522e-01 -2.41184190e-01 -1.43151343e-01
1.26010668e+00 -8.84141997e-02 -1.49522955e-02 -4.90802467e-01
4.27062243e-01 3.63665283e-01 7.27970243e-01 4.15970176e-01
-2.24057674e-01 -8.06464627e-03 1.47733837e-01 -4.01251078e-01
-1.13686466e+00 -7.63276637e-01 -1.27407610e-01 6.35679126e-01
6.70172036e-01 -3.30234349e-01 -7.90184021e-01 3.57587337e-02
2.35300586e-01 3.98477614e-01 -6.69447854e-02 -2.84370214e-01
-8.16808701e-01 -7.88000822e-01 8.28504264e-01 9.18507516e-01
9.40336645e-01 -3.80288959e-01 -1.81472683e+00 4.21734452e-01
6.68372810e-01 -1.51850438e+00 -2.07493082e-01 3.43663722e-01
-8.72000039e-01 -7.46436417e-01 -7.37541728e-03 -5.58694959e-01
8.07702959e-01 3.26236486e-01 6.43463314e-01 -4.14796054e-01
-5.90091169e-01 -2.58523207e-02 9.45169851e-02 -6.54050887e-01
-1.80529207e-01 -4.22473282e-01 1.03140801e-01 1.41577139e-01
6.87329590e-01 -4.23874371e-02 -9.70095813e-01 2.47755036e-01
-5.61105728e-01 2.90912807e-01 7.95400679e-01 6.02063417e-01
7.81350970e-01 -7.43941665e-02 2.44062304e-01 -4.04533982e-01
-4.10179436e-01 -7.25579336e-02 -1.72921145e+00 -2.29195833e-01
-6.55279279e-01 -5.90694696e-02 7.51234591e-01 -6.27955019e-01
-9.05976534e-01 7.90953577e-01 2.68963665e-01 -6.67941988e-01
3.09472352e-01 -3.05297315e-01 -2.00087994e-01 -5.68824708e-01
3.35443288e-01 -9.61267129e-02 1.49449393e-01 -9.79988426e-02
9.03030038e-02 7.86934972e-01 8.68391573e-01 2.73077548e-01
6.53341353e-01 7.10281730e-01 3.27629834e-01 -9.37453866e-01
8.75011310e-02 -1.11982390e-01 4.38309573e-02 -4.62592989e-01
1.05507958e+00 -1.76295555e+00 -1.39260077e+00 8.00822020e-01
-1.27516866e+00 -3.50067258e-01 -6.71894029e-02 6.19531214e-01
-7.95095935e-02 -1.58864960e-01 -5.31231284e-01 -6.74682975e-01
-8.51903617e-01 -1.36471808e+00 1.09865892e+00 7.90685296e-01
1.45839736e-01 -5.65647930e-02 -8.83352935e-01 3.85290205e-01
6.35036707e-01 1.59580171e-01 5.46048284e-01 1.05754621e-01
-1.23046410e+00 -1.73359022e-01 -7.35182524e-01 1.46903381e-01
-3.55413884e-01 -6.96573704e-02 -1.15932906e+00 -3.08877736e-01
5.91127202e-02 -4.04620692e-02 9.73537683e-01 5.01394570e-01
9.36199307e-01 -4.59042232e-04 -7.30625510e-01 8.29314768e-01
1.97269654e+00 2.84166873e-01 6.36134624e-01 -1.34434268e-01
8.03276837e-01 -2.31529236e-01 7.17992604e-01 4.78586227e-01
2.92281181e-01 8.50260973e-01 6.40867651e-01 2.25770865e-02
-3.20074528e-01 -1.92517117e-02 8.73486221e-01 4.12412047e-01
4.25487220e-01 -3.07676673e-01 -7.76291728e-01 3.11985403e-01
-1.68737996e+00 -5.83039403e-01 -6.49393678e-01 2.14072466e+00
4.72057968e-01 2.90944099e-01 -2.07848102e-01 -1.15465224e-01
7.55589724e-01 -1.74231291e-01 -1.00929189e+00 -2.92162836e-01
8.66611674e-02 2.45731607e-01 1.74220288e+00 1.14192761e-01
-9.90828872e-01 7.20819831e-01 4.62334538e+00 8.83965433e-01
-1.68567836e+00 2.58505404e-01 4.73097026e-01 -6.59431934e-01
3.86573076e-01 -1.53793812e-01 -1.60505176e+00 5.75110793e-01
1.61926603e+00 1.03953540e-01 1.12712570e-01 1.05504823e+00
1.65750965e-01 -4.93372768e-01 -1.24143684e+00 1.42753029e+00
-1.18223444e-01 -1.61154211e+00 -4.03152883e-01 2.22631305e-01
3.68712366e-01 3.14466774e-01 1.25898376e-01 -2.69312020e-02
-1.45175204e-01 -6.20642960e-01 1.06446016e+00 1.74446627e-01
1.23047602e+00 -8.26373279e-01 4.43760931e-01 1.61871433e-01
-1.38165128e+00 -4.33222502e-01 -5.86506486e-01 -1.29935190e-01
1.03707574e-01 9.13399518e-01 -6.51649952e-01 1.75042078e-03
1.00757992e+00 1.88543111e-01 -4.44700897e-01 5.29366910e-01
2.37306595e-01 4.61217284e-01 -7.09743977e-01 -3.92694086e-01
1.25284865e-01 1.56873494e-01 2.58372754e-01 1.13885748e+00
2.72878408e-01 2.57611483e-01 -2.76174515e-01 7.65367627e-01
-2.56995350e-01 -6.98737860e-01 -5.33850230e-02 2.83208400e-01
5.62684357e-01 1.65900433e+00 -8.10579002e-01 -5.37762463e-01
-7.16716886e-01 9.21621859e-01 -3.50093603e-01 -2.82304108e-01
-1.34249270e+00 -5.24071336e-01 1.03684509e+00 2.71250457e-01
6.91999018e-01 -4.21712101e-01 -5.92669666e-01 -4.88138527e-01
2.91113079e-01 -3.00102890e-01 -2.35029280e-01 -3.57430220e-01
-5.43916166e-01 3.25658828e-01 -5.11719346e-01 -1.11368859e+00
2.79356986e-01 -9.67910409e-01 -1.08311832e-01 7.76855469e-01
-1.67234886e+00 -8.30217242e-01 -8.54501247e-01 3.46051306e-01
3.78447443e-01 3.24711949e-02 4.13991153e-01 7.65974283e-01
-9.28714514e-01 8.89444053e-01 2.57133812e-01 1.65592030e-01
4.30607289e-01 -4.73295361e-01 8.09611738e-01 9.91561413e-01
-7.09599078e-01 3.21918517e-01 2.59191453e-01 -5.49268663e-01
-2.56581426e+00 -1.81937587e+00 2.54383981e-01 -3.37266065e-02
3.92673254e-01 -6.37564540e-01 -4.97030407e-01 3.01005363e-01
1.32961258e-01 6.92360580e-01 2.97165602e-01 -1.02563262e+00
-1.76759928e-01 -7.74613798e-01 -1.37755454e+00 4.93224263e-01
7.83509076e-01 -2.29712814e-01 1.82764366e-01 -1.04087099e-01
6.88839197e-01 -5.35148203e-01 -6.29262269e-01 4.27632928e-01
8.39729905e-01 -6.62387073e-01 6.47905469e-01 3.12473536e-01
1.79121837e-01 -8.66572857e-01 -2.97063053e-01 -3.32621157e-01
-1.36821389e-01 -8.94645929e-01 -2.15514824e-01 1.15514398e+00
3.55319172e-01 -5.83659947e-01 1.03220725e+00 5.44377148e-01
-2.50779331e-01 -6.11836970e-01 -1.24556625e+00 -6.89342856e-01
-5.93290389e-01 -5.36376297e-01 4.74041790e-01 9.89897773e-02
-1.61008552e-01 3.96690607e-01 -5.09873107e-02 6.77229047e-01
7.35677123e-01 -1.83115564e-02 8.03198695e-01 -9.26483095e-01
-1.83617428e-01 -1.13412760e-01 -8.40039253e-01 -1.36231840e+00
-3.88649166e-01 -3.96487266e-01 1.38647109e-01 -8.52586865e-01
-1.03747316e-01 -3.87286395e-01 -5.54635674e-02 4.19598579e-01
1.08981773e-01 5.50020576e-01 2.98146099e-01 -9.04441029e-02
-6.06835425e-01 1.73647314e-01 6.10939026e-01 -8.75270516e-02
-9.81553793e-02 -4.71567333e-01 -4.00004029e-01 6.42409682e-01
6.14993393e-01 -4.56879765e-01 -1.12329759e-01 -6.34759367e-01
3.11179878e-03 1.35704651e-01 7.58954048e-01 -1.77017581e+00
6.89538836e-01 2.31900960e-01 8.22609365e-01 -8.48627925e-01
6.22320831e-01 -1.14603388e+00 4.93054897e-01 8.96319985e-01
2.85079867e-01 8.13210011e-02 7.75829017e-01 4.42998409e-01
1.72499940e-01 2.76972711e-01 1.15144157e+00 3.38360012e-01
-9.97168601e-01 3.31682526e-02 -4.86111730e-01 -4.59385425e-01
1.54116750e+00 -5.45750320e-01 -7.63382792e-01 3.86167407e-01
1.04720168e-01 1.11347474e-01 4.84592825e-01 1.83425844e-01
5.72454929e-01 -1.15266120e+00 -2.84716815e-01 6.36004329e-01
-9.07804966e-02 1.45557046e-01 4.28438753e-01 8.08007658e-01
-7.76866257e-01 6.68819308e-01 -4.05897200e-01 -9.24365222e-01
-1.42931521e+00 4.37422693e-01 1.36901945e-01 2.44705662e-01
-7.36680686e-01 8.72039139e-01 -1.44331142e-01 6.38068318e-01
5.93047701e-02 -7.72727370e-01 4.16399598e-01 -6.37696832e-02
8.24848533e-01 7.31783330e-01 1.84083730e-01 -4.66646761e-01
-5.92511415e-01 7.93927968e-01 -1.27213478e-01 2.53063291e-01
1.11960340e+00 1.39234245e-01 1.64186478e-01 -2.99762264e-02
1.35167563e+00 -4.05054539e-01 -1.76037014e+00 4.45410609e-02
-2.46707246e-01 -1.60795048e-01 5.76322675e-01 -5.37171662e-01
-1.38710737e+00 5.54581344e-01 1.28001356e+00 -2.47900203e-01
1.51865292e+00 -1.42400935e-01 1.07264292e+00 2.85806447e-01
4.30892646e-01 -1.14892113e+00 -3.02346855e-01 2.11193532e-01
1.85137525e-01 -1.00401521e+00 2.43631199e-01 -5.29561579e-01
-2.45033458e-01 1.12552774e+00 9.50309157e-01 -1.19717546e-01
4.92959291e-01 1.06852973e+00 -4.62419242e-02 -3.43360868e-03
-7.26571321e-01 2.04723448e-01 -2.91045129e-01 3.44801813e-01
-1.66101620e-01 1.98245287e-01 1.24263223e-02 6.90988660e-01
1.23110123e-01 3.16457361e-01 5.12175322e-01 9.65762556e-01
-4.68692690e-01 -5.48512816e-01 -4.71427828e-01 5.24475574e-01
-2.41683379e-01 1.31785870e-01 2.21283272e-01 4.05646026e-01
5.71705341e-01 9.43004131e-01 7.84932315e-01 -7.26922929e-01
3.05462748e-01 -4.11712646e-01 3.41545820e-01 -1.47021070e-01
-7.94527650e-01 4.38312721e-03 -1.11782804e-01 -9.72853839e-01
-2.78725475e-01 -4.10303354e-01 -1.64196706e+00 -3.72498810e-01
-4.33412224e-01 -7.38666177e-01 1.31536293e+00 5.35802543e-01
8.71538579e-01 7.89871633e-01 3.77111167e-01 -9.27035630e-01
-4.53602433e-01 -5.17309844e-01 -2.34760627e-01 -5.11061728e-01
5.51301479e-01 -3.78859729e-01 4.81024794e-02 1.31097481e-01] | [8.298600196838379, 2.407804012298584] |
3085ae14-8faf-4d26-a3f2-412a8a2e06b5 | arigan-synthetic-arabidopsis-plants-using | 1709.00938 | null | http://arxiv.org/abs/1709.00938v1 | http://arxiv.org/pdf/1709.00938v1.pdf | ARIGAN: Synthetic Arabidopsis Plants using Generative Adversarial Network | In recent years, there has been an increasing interest in image-based plant
phenotyping, applying state-of-the-art machine learning approaches to tackle
challenging problems, such as leaf segmentation (a multi-instance problem) and
counting. Most of these algorithms need labelled data to learn a model for the
task at hand. Despite the recent release of a few plant phenotyping datasets,
large annotated plant image datasets for the purpose of training deep learning
algorithms are lacking. One common approach to alleviate the lack of training
data is dataset augmentation. Herein, we propose an alternative solution to
dataset augmentation for plant phenotyping, creating artificial images of
plants using generative neural networks. We propose the Arabidopsis Rosette
Image Generator (through) Adversarial Network: a deep convolutional network
that is able to generate synthetic rosette-shaped plants, inspired by DCGAN (a
recent adversarial network model using convolutional layers). Specifically, we
trained the network using A1, A2, and A4 of the CVPPP 2017 LCC dataset,
containing Arabidopsis Thaliana plants. We show that our model is able to
generate realistic 128x128 colour images of plants. We train our network
conditioning on leaf count, such that it is possible to generate plants with a
given number of leaves suitable, among others, for training regression based
models. We propose a new Ax dataset of artificial plants images, obtained by
our ARIGAN. We evaluate this new dataset using a state-of-the-art leaf counting
algorithm, showing that the testing error is reduced when Ax is used as part of
the training data. | ['Sotirios A. Tsaftaris', 'Hanno Scharr', 'Mario Valerio Giuffrida'] | 2017-09-04 | null | null | null | null | ['plant-phenotyping'] | ['computer-vision'] | [ 8.63938749e-01 3.87176365e-01 2.72553325e-01 1.91529244e-02
-4.30552065e-01 -1.11927772e+00 5.03033280e-01 -3.85420211e-02
-1.08051542e-02 7.55041063e-01 -7.84257770e-01 -7.20565200e-01
1.15282319e-01 -1.36282837e+00 -1.05629528e+00 -8.73044491e-01
2.33122244e-01 7.94394791e-01 8.64110067e-02 -2.03164481e-02
-2.93276370e-01 8.82279336e-01 -1.50153053e+00 1.72780737e-01
9.59509611e-01 9.89223719e-01 3.07885259e-01 1.14682043e+00
-1.64770097e-01 5.94694674e-01 -9.37852085e-01 -1.46812648e-01
5.02923608e-01 -6.78818583e-01 -9.52486992e-01 4.02570933e-01
2.87920892e-01 -1.51880577e-01 1.92320153e-01 8.03674102e-01
5.21931648e-01 -5.62318563e-01 5.16694248e-01 -1.60869384e+00
-8.18237424e-01 8.85550022e-01 -5.46773732e-01 -4.76574719e-01
-2.96081811e-01 5.23308277e-01 6.00368917e-01 -3.49684328e-01
5.71324706e-01 9.65318561e-01 6.64720237e-01 6.60757303e-01
-1.78598070e+00 -2.54678637e-01 -3.42399716e-01 -2.59475648e-01
-1.13145542e+00 2.22079813e-01 4.93271559e-01 -3.08847517e-01
5.77451438e-02 4.48979944e-01 7.54418910e-01 1.14164221e+00
-4.44069356e-01 8.27284575e-01 1.16993725e+00 -5.86521864e-01
4.15742695e-01 -1.61677986e-01 -5.35219789e-01 6.17844939e-01
3.96863818e-01 1.98989898e-01 5.06455362e-01 3.65242222e-03
9.77502882e-01 -2.38071576e-01 -2.29154035e-01 -5.15590847e-01
-9.36964333e-01 9.12948489e-01 5.88877082e-01 3.97963822e-01
-3.62209707e-01 2.23820448e-01 2.48334408e-01 -1.35419205e-01
7.11035058e-02 6.66742921e-01 -7.74036467e-01 3.60335618e-01
-9.41667914e-01 4.40052688e-01 1.01175308e+00 9.40396011e-01
6.32202566e-01 3.24259579e-01 -4.13739085e-01 5.24236023e-01
-1.28475100e-01 6.65941119e-01 1.66718543e-01 -1.12917447e+00
-2.20995784e-01 7.69614518e-01 5.97527809e-02 -5.39025426e-01
-2.96023488e-01 -2.29581863e-01 -1.14540386e+00 7.66286731e-01
9.77116823e-01 -3.12042683e-01 -1.12256956e+00 1.79278696e+00
2.30291426e-01 1.71430409e-01 2.14448810e-01 4.64204222e-01
7.42364645e-01 6.41190767e-01 1.64351776e-01 1.91523880e-01
1.13365269e+00 -4.88565296e-01 -2.29831159e-01 -8.80497321e-02
7.60924697e-01 -5.56360066e-01 9.32620585e-01 3.83763790e-01
-1.06862044e+00 -6.78267539e-01 -8.70799661e-01 2.09539175e-01
-7.64706492e-01 7.08392203e-01 7.82231450e-01 1.00430727e+00
-9.38875318e-01 8.90698075e-01 -5.40961981e-01 -3.68717819e-01
9.79838014e-01 4.26619470e-01 -3.56703490e-01 -1.22929931e-01
-5.95730186e-01 6.06142938e-01 7.28742659e-01 2.82015532e-01
-1.37978899e+00 -9.91074622e-01 -5.92940211e-01 4.70157474e-01
4.16543961e-01 -5.44412494e-01 1.02010655e+00 -1.04788232e+00
-1.73008406e+00 1.22010505e+00 6.02527976e-01 -6.44674361e-01
4.80749339e-01 3.49881083e-01 -2.16254406e-02 -4.56439555e-02
-1.21654170e-02 1.21387041e+00 8.42656374e-01 -1.62625670e+00
-4.36388373e-01 -3.68910402e-01 6.58746809e-02 -6.70124352e-01
4.69995551e-02 -3.84988606e-01 4.99782264e-02 -3.61430168e-01
-3.62779945e-01 -1.28316677e+00 -4.06498671e-01 3.59592468e-01
-8.45121682e-01 3.13687265e-01 1.28792012e+00 -5.62498569e-01
1.82052776e-01 -1.70808983e+00 1.59440041e-01 6.15296178e-02
-1.22032855e-02 9.88071084e-01 -7.20480978e-01 3.32815722e-02
-4.74632293e-01 5.40169537e-01 -9.45833921e-01 -1.71705842e-01
-2.08443195e-01 4.91767943e-01 -2.74963230e-01 8.73948336e-02
8.32925677e-01 1.13435984e+00 -7.41417646e-01 -1.25277147e-01
4.16378230e-01 6.17846012e-01 -2.52458513e-01 4.48822975e-01
-6.86511397e-01 9.18451130e-01 -1.79619372e-01 7.45587170e-01
1.23702407e+00 -9.00900811e-02 2.16209814e-01 8.28338563e-02
-1.88124403e-01 -5.14880240e-01 -7.42986917e-01 1.19654000e+00
-4.59328294e-01 3.23192328e-01 7.22429380e-02 -1.23032594e+00
1.14871180e+00 2.22311556e-01 6.08321309e-01 -1.21981382e-01
1.86988115e-01 2.03529984e-01 2.22023368e-01 1.87808394e-01
-1.54566765e-01 3.78655136e-01 2.22044259e-01 -1.46191567e-02
3.13533932e-01 -8.31043541e-01 7.06887782e-01 -5.60938008e-02
1.35660338e+00 4.69501585e-01 5.58147989e-02 -4.95136566e-02
7.08131611e-01 1.83401361e-01 3.89735430e-01 6.38989031e-01
-5.86314052e-02 9.67146158e-01 8.78441870e-01 -4.09177959e-01
-1.42695522e+00 -6.01961732e-01 -2.78694350e-02 4.84490991e-01
-6.09410524e-01 -6.91083297e-02 -1.12258732e+00 -8.03929806e-01
8.34106002e-03 7.16104567e-01 -7.36437201e-01 -9.14078206e-02
-4.20135319e-01 -1.12755764e+00 1.10699046e+00 3.73295784e-01
8.22434664e-01 -1.71460927e+00 -8.76779079e-01 1.55704647e-01
1.10242106e-01 -1.44436324e+00 4.59496826e-01 6.68251276e-01
-6.92312896e-01 -1.27474272e+00 -1.03519177e+00 -6.56215370e-01
6.95363522e-01 -1.68659642e-01 1.40928352e+00 3.10020894e-01
-8.52872729e-01 1.04773872e-01 -4.10107255e-01 -6.94839358e-01
-9.40571487e-01 4.27006274e-01 -6.88123524e-01 -2.19436988e-01
-1.63814932e-01 -7.67122388e-01 -4.03936058e-01 -7.14194868e-03
-1.34655786e+00 5.31094670e-02 6.64219260e-01 1.08045161e+00
6.70882821e-01 -7.27579594e-02 3.96401465e-01 -1.30203021e+00
7.61914626e-02 -2.06545636e-01 -1.45393634e+00 4.31473643e-01
-1.31603569e-01 -1.22190863e-01 1.19227779e+00 -2.34335601e-01
-8.05090785e-01 6.97579682e-01 -2.66397804e-01 -1.55863628e-01
-7.53380060e-01 4.12766933e-01 -6.51513219e-01 -4.43162113e-01
7.20767915e-01 1.06972821e-01 5.00902273e-02 -1.23319894e-01
7.00504839e-01 7.70460516e-02 8.32420170e-01 -5.60567260e-01
1.14223218e+00 4.54540253e-01 7.43439198e-01 -7.73389459e-01
-9.24600780e-01 5.14977090e-02 -8.63548100e-01 1.30015770e-02
7.23873854e-01 -3.89238864e-01 -1.07231724e+00 7.89248228e-01
-1.21711349e+00 -9.08677340e-01 -9.20864940e-01 -1.94087133e-01
-8.76260161e-01 1.64894328e-01 -2.12554961e-01 -6.85095549e-01
-3.60490352e-01 -9.87998426e-01 1.15734088e+00 4.42497462e-01
4.02428746e-01 -8.84799540e-01 -5.16171269e-02 -9.96100157e-02
4.32009727e-01 9.71884370e-01 8.80005181e-01 -3.57533842e-01
-6.77271724e-01 -4.58958685e-01 -5.97806990e-01 5.74239194e-01
1.28791779e-01 6.09636426e-01 -1.15176690e+00 -7.39792585e-02
-2.42007285e-01 -6.64676428e-01 7.63667226e-01 6.60757720e-01
1.77503133e+00 -6.17579855e-02 -1.45121619e-01 7.43005037e-01
1.63449216e+00 -2.13825610e-02 1.18861544e+00 7.98139423e-02
8.01109970e-01 5.37101388e-01 2.23212138e-01 3.51797938e-01
-2.45895579e-01 3.66070479e-01 1.08373487e+00 -5.52287757e-01
-9.99583676e-02 3.96739170e-02 -3.28877866e-01 -1.43530354e-01
-2.20991939e-01 -6.90412939e-01 -8.33662331e-01 5.56046665e-01
-1.53355169e+00 -9.10183609e-01 -4.19350564e-01 2.12931752e+00
8.48841786e-01 -6.58256188e-02 4.45370376e-02 5.36261797e-01
6.32833958e-01 -1.62803203e-01 -6.55827641e-01 -2.51710892e-01
-5.38962305e-01 8.51799130e-01 7.95094728e-01 1.28734261e-01
-1.30309832e+00 1.19667339e+00 5.68395948e+00 4.81404781e-01
-1.13502169e+00 -1.97260410e-01 1.08864379e+00 7.67339945e-01
-1.10204704e-02 3.63263488e-01 -6.74547732e-01 2.06140012e-01
9.11957860e-01 2.91741461e-01 4.46037322e-01 9.51382697e-01
-1.37472495e-01 -1.07188687e-01 -7.35589802e-01 3.73025417e-01
-1.29414290e-01 -1.39172983e+00 1.59477264e-01 6.95811585e-02
7.15692282e-01 -4.45610851e-01 1.89539101e-02 3.24596912e-01
7.65878558e-01 -1.29133487e+00 7.95535296e-02 2.68512815e-01
9.88496900e-01 -5.64451396e-01 6.12094998e-01 3.98028791e-01
-9.78172898e-01 1.60393286e-02 -6.81473076e-01 4.96130139e-01
-9.69052613e-02 9.06237602e-01 -9.41358328e-01 6.33268893e-01
4.57217544e-01 2.31410056e-01 -1.00286365e+00 9.61024702e-01
-4.02356029e-01 8.32367599e-01 -4.25755262e-01 4.32312101e-01
1.00818142e-01 -2.77221620e-01 1.03547752e-01 6.69125795e-01
5.94778061e-01 -1.95804298e-01 1.52615830e-01 1.47435486e+00
-2.94159442e-01 -2.55249590e-01 -8.04742754e-01 -4.80448246e-01
1.13653608e-01 1.88820291e+00 -1.03623927e+00 -1.46583468e-01
6.78869337e-02 1.21124423e+00 1.45893306e-01 -1.06618397e-01
-7.81695366e-01 -3.19188416e-01 -3.58629599e-02 -1.12696737e-01
6.44992173e-01 1.99486036e-02 -2.19136685e-01 -5.99332273e-01
-4.61004227e-01 -1.01403129e+00 -2.45628245e-02 -9.22242761e-01
-9.99214411e-01 4.48636234e-01 -4.25564826e-01 -6.69367671e-01
-2.50048220e-01 -1.07441533e+00 -7.93603897e-01 1.02160668e+00
-1.40839493e+00 -1.79935920e+00 -9.58609879e-01 1.86355010e-01
1.04507588e-01 -1.92407608e-01 1.55008006e+00 6.55562133e-02
-3.49853456e-01 2.40430862e-01 -1.68266632e-02 2.37455860e-01
3.83272141e-01 -1.56684840e+00 8.02507997e-01 9.17354345e-01
1.07072681e-01 -1.51617616e-01 4.00737971e-01 -4.64412838e-01
-1.20132995e+00 -1.58748984e+00 4.37560022e-01 -3.90632927e-01
2.43753627e-01 -1.43115535e-01 -7.97492504e-01 7.65809000e-01
-8.53137858e-03 5.82825184e-01 2.47041166e-01 -6.09987974e-01
-6.65280819e-02 4.81314659e-02 -1.57829833e+00 5.34195006e-01
6.75640523e-01 -2.96539590e-02 5.12655675e-01 6.07422769e-01
6.04166389e-01 -3.33220243e-01 -8.33928525e-01 7.15941787e-01
2.57385582e-01 -8.59095633e-01 1.12202406e+00 -4.90586966e-01
6.64831877e-01 -3.91073436e-01 2.95939017e-02 -1.59573102e+00
-1.81621894e-01 -6.09393179e-01 3.91700640e-02 1.54783332e+00
3.29436064e-01 -2.11767465e-01 9.61553752e-01 1.53594345e-01
2.27346737e-02 -3.19632739e-01 -2.99906582e-01 -7.53085017e-01
4.48745340e-01 1.21883124e-01 9.22029793e-01 7.64334857e-01
-9.16635811e-01 2.88005378e-02 -2.91342884e-01 2.04397261e-01
3.43846291e-01 1.57995537e-01 1.08899629e+00 -1.60645854e+00
-2.30111390e-01 -2.81167001e-01 -3.93054098e-01 -4.08079803e-01
3.31273764e-01 -8.97693515e-01 9.11442637e-02 -1.24110770e+00
-3.00152078e-02 -5.78919828e-01 4.82407719e-01 7.50165343e-01
-1.24735162e-01 3.21594298e-01 3.19004059e-01 -3.06182384e-01
2.85188556e-01 2.49565586e-01 1.48966455e+00 -4.32946444e-01
4.25538309e-02 2.83737779e-01 -5.24295032e-01 5.99959373e-01
1.15809178e+00 -4.94339317e-01 -1.61858067e-01 -2.92826146e-01
-3.96523662e-02 -1.72578022e-01 1.03783429e+00 -1.41779387e+00
-3.38358641e-01 8.18562657e-02 7.85641909e-01 -5.47889531e-01
3.17784995e-01 -8.52339566e-01 3.31031770e-01 7.61944473e-01
-1.53336942e-01 -2.22146690e-01 4.86492932e-01 -6.14209101e-02
1.28859997e-01 -7.47670054e-01 1.09891450e+00 -3.95795554e-01
-3.96541208e-01 5.14056981e-01 -2.73601025e-01 -4.05490845e-02
1.04959691e+00 4.36928794e-02 -4.46544945e-01 -1.90856352e-01
-5.91217339e-01 -1.58050463e-01 2.88339794e-01 9.38742980e-03
4.76933718e-02 -1.00951922e+00 -9.60878849e-01 1.21078670e-01
1.51994452e-01 3.52637887e-01 -6.06595427e-02 2.53128827e-01
-1.09450221e+00 2.70096511e-01 -7.59786129e-01 -7.42666662e-01
-9.15593207e-01 6.79166615e-01 3.80645216e-01 -6.74814045e-01
-2.84075797e-01 5.70329130e-01 1.47181422e-01 -1.12806964e+00
-1.15950555e-01 -4.67494011e-01 -2.98659265e-01 -3.19369256e-01
2.14174792e-01 1.58260614e-02 1.33926168e-01 -1.68504849e-01
1.62814572e-01 1.38681352e-01 5.40184200e-01 1.99853510e-01
1.47778535e+00 6.99510753e-01 -1.29744291e-01 1.22739248e-01
5.74211240e-01 -3.73709083e-01 -1.17848432e+00 4.61888790e-01
8.81212130e-02 -3.27738702e-01 -2.26032197e-01 -1.11803079e+00
-1.41822481e+00 8.49663854e-01 9.38164473e-01 7.51094460e-01
1.19188631e+00 -5.15472054e-01 5.89291394e-01 4.23765838e-01
1.32727802e-01 -5.63384056e-01 -2.98635453e-01 5.57556987e-01
9.43653941e-01 -1.36300731e+00 -5.04465878e-01 -7.51661897e-01
-1.81676209e-01 1.12338960e+00 6.07480764e-01 -3.30202639e-01
3.98265958e-01 7.48389184e-01 -1.49518335e-02 1.12133518e-01
-3.50093126e-01 -6.39617443e-01 -4.54137266e-01 1.32119203e+00
3.66930038e-01 1.71605393e-01 -5.33695109e-02 3.60974759e-01
-6.83794320e-02 2.50717312e-01 7.55254209e-01 8.89643431e-01
4.25433889e-02 -1.55680752e+00 -4.08569634e-01 2.04618245e-01
-4.21602219e-01 -1.18165433e-01 -7.37750053e-01 9.95117009e-01
3.77963781e-01 5.43437600e-01 -1.49352744e-01 2.75001049e-01
2.38889247e-01 3.54411490e-02 6.58904910e-01 -6.46441698e-01
-6.51928723e-01 -5.84285975e-01 -3.04363072e-01 -2.34089896e-01
-4.12214875e-01 -4.16577876e-01 -8.79456341e-01 -3.81751120e-01
-4.20054913e-01 -3.44427317e-01 9.20080364e-01 7.16022253e-01
2.74632394e-01 9.89922285e-01 5.87829649e-01 -1.03358459e+00
-3.91536862e-01 -8.99060190e-01 -6.80451810e-01 4.35692906e-01
-1.70195952e-01 -2.83072829e-01 -1.53870180e-01 4.40204412e-01] | [9.125652313232422, -1.5299009084701538] |
a6f46d92-4767-45c2-9460-cbd5c4c42089 | varitex-variational-neural-face-textures | 2104.05988 | null | https://arxiv.org/abs/2104.05988v3 | https://arxiv.org/pdf/2104.05988v3.pdf | VariTex: Variational Neural Face Textures | Deep generative models can synthesize photorealistic images of human faces with novel identities. However, a key challenge to the wide applicability of such techniques is to provide independent control over semantically meaningful parameters: appearance, head pose, face shape, and facial expressions. In this paper, we propose VariTex - to the best of our knowledge the first method that learns a variational latent feature space of neural face textures, which allows sampling of novel identities. We combine this generative model with a parametric face model and gain explicit control over head pose and facial expressions. To generate complete images of human heads, we propose an additive decoder that adds plausible details such as hair. A novel training scheme enforces a pose-independent latent space and in consequence, allows learning a one-to-many mapping between latent codes and pose-conditioned exterior regions. The resulting method can generate geometrically consistent images of novel identities under fine-grained control over head pose, face shape, and facial expressions. This facilitates a broad range of downstream tasks, like sampling novel identities, changing the head pose, expression transfer, and more. Code and models are available for research on https://mcbuehler.github.io/VariTex. | ['Otmar Hilliges', 'Thabo Beeler', 'Gengyan Li', 'Abhimitra Meka', 'Marcel C. Bühler'] | 2021-04-13 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Buhler_VariTex_Variational_Neural_Face_Textures_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Buhler_VariTex_Variational_Neural_Face_Textures_ICCV_2021_paper.pdf | iccv-2021-1 | ['face-model'] | ['computer-vision'] | [ 1.99728951e-01 4.81642485e-01 1.61306322e-01 -6.66240692e-01
-7.06821978e-01 -6.82728350e-01 8.40177059e-01 -7.19893456e-01
1.16172470e-02 6.10250592e-01 4.78898019e-01 4.25776899e-01
2.90754348e-01 -5.84719777e-01 -1.01993454e+00 -7.85145283e-01
3.28151733e-01 5.25274038e-01 -3.42054218e-01 -1.72342002e-01
-4.69120257e-02 7.82040834e-01 -1.84043646e+00 1.32366136e-01
3.14098120e-01 7.68560648e-01 -9.60020721e-02 6.48477495e-01
3.69107753e-01 4.05999631e-01 -2.66967595e-01 -6.33224428e-01
3.30180407e-01 -6.24656022e-01 -4.64190185e-01 3.87282819e-01
8.55637670e-01 -5.58351159e-01 -2.07440346e-01 7.79409051e-01
5.99244893e-01 7.62649551e-02 9.31574821e-01 -1.03598416e+00
-1.04829323e+00 2.27587670e-01 -5.67265928e-01 -4.04252708e-01
3.72584134e-01 2.84644753e-01 7.32348144e-01 -1.02674246e+00
8.95947874e-01 1.51475048e+00 4.69827384e-01 1.13260782e+00
-1.58746231e+00 -8.38111341e-01 9.28409994e-02 -2.96199769e-01
-1.49739540e+00 -1.07709265e+00 7.16215432e-01 -5.25861621e-01
3.47235978e-01 2.63532102e-01 6.03828251e-01 1.44649839e+00
-7.11144879e-03 4.63358164e-01 1.13507724e+00 -4.97185737e-01
1.53786585e-01 7.10275471e-02 -6.15722954e-01 9.17430639e-01
-1.34606123e-01 1.18151747e-01 -6.66704476e-01 -1.81833521e-01
1.27473247e+00 -1.46229267e-01 -2.37195984e-01 -6.76499486e-01
-8.04498494e-01 7.65199721e-01 2.15176105e-01 -6.81984276e-02
-1.27899140e-01 5.02260804e-01 2.47957800e-02 -1.47015434e-02
4.74978089e-01 4.03607130e-01 -2.93911248e-01 2.41168395e-01
-7.24558294e-01 5.56264281e-01 5.84584892e-01 1.08724189e+00
8.63682449e-01 2.42248416e-01 -3.41325909e-01 8.68452072e-01
3.87149841e-01 7.52402544e-01 1.31203920e-01 -1.50100946e+00
-2.27013126e-01 -3.57963145e-02 5.25209773e-03 -7.79237986e-01
-5.86335398e-02 -1.78318843e-01 -4.79376286e-01 3.38915974e-01
3.45407724e-01 -3.24725181e-01 -1.07992578e+00 2.38493109e+00
5.75413823e-01 -3.69413607e-02 -3.68404716e-01 8.20355952e-01
7.42482126e-01 3.47044766e-01 8.38931426e-02 -7.63834640e-02
1.32513356e+00 -4.54317778e-01 -5.86064458e-01 -2.09939227e-01
-2.57634949e-02 -8.53501320e-01 1.02674615e+00 7.36879632e-02
-1.42701590e+00 -2.91516423e-01 -5.70723176e-01 -4.44366455e-01
1.05602564e-02 2.36192212e-01 7.01755047e-01 5.99675894e-01
-1.41769946e+00 3.69956762e-01 -7.93122470e-01 -5.31028152e-01
5.71569204e-01 5.64317226e-01 -5.39725661e-01 1.27537698e-01
-7.88353860e-01 7.40960121e-01 -3.09529006e-01 4.49611284e-02
-9.33567822e-01 -6.91212952e-01 -1.06534016e+00 -4.03922386e-02
9.82179046e-02 -1.08394265e+00 1.51407349e+00 -1.32499397e+00
-1.91167414e+00 1.33375418e+00 -4.19113636e-01 1.23179346e-01
6.15589142e-01 6.54491708e-02 1.37298033e-01 5.68102933e-02
9.97209847e-02 1.29927254e+00 1.32721794e+00 -1.28576732e+00
1.35553554e-01 -6.18867517e-01 -2.32808739e-01 2.79722720e-01
-1.26023352e-01 2.03347400e-01 -5.62355638e-01 -6.76499426e-01
-1.98139861e-01 -1.13072324e+00 1.05109535e-01 6.04111373e-01
-2.44029716e-01 9.19850543e-02 4.72823799e-01 -5.97220182e-01
5.28778315e-01 -2.23723841e+00 5.02937675e-01 1.15448058e-01
2.28738874e-01 -2.40855843e-01 -1.66226611e-01 1.22737788e-01
-8.26771185e-02 -5.15204854e-02 -1.87002927e-01 -8.15350592e-01
1.66711241e-01 1.60519317e-01 -2.88412660e-01 5.87602615e-01
3.55644107e-01 1.14253712e+00 -4.95160520e-01 -3.17090958e-01
-1.89100541e-02 9.98581588e-01 -8.76892447e-01 1.87930241e-01
-4.46456254e-01 9.38065350e-01 -2.99091965e-01 4.94739085e-01
7.25774288e-01 -3.28945667e-02 2.34976009e-01 -1.36349559e-01
-1.04977908e-02 -7.31639042e-02 -7.15402007e-01 1.68605936e+00
-4.72600281e-01 5.18817127e-01 4.79712158e-01 -3.10104072e-01
8.29474092e-01 3.76551062e-01 1.85505360e-01 -3.71641487e-01
4.65354174e-01 4.23664488e-02 -3.26509327e-01 -2.75083452e-01
1.24797292e-01 -5.33809960e-01 4.56650592e-02 3.64203960e-01
2.53185570e-01 -4.41479683e-01 -2.15777367e-01 -1.38161769e-02
5.02699316e-01 5.30225873e-01 1.63485426e-02 -3.29741597e-01
1.28601894e-01 -6.34656012e-01 3.48483026e-01 2.95261413e-01
1.27606913e-01 8.93965721e-01 5.22857726e-01 -2.10877120e-01
-1.32863224e+00 -1.24368048e+00 -3.18939030e-01 1.29010105e+00
-3.34018439e-01 2.31832229e-02 -9.16441083e-01 -1.17642649e-01
9.73179862e-02 5.89371026e-01 -9.54060912e-01 -1.22070111e-01
-6.80156589e-01 -3.18166822e-01 4.85068381e-01 3.45105290e-01
1.25369743e-01 -1.14801407e+00 -1.68792814e-01 -2.35358447e-01
5.06628528e-02 -1.03290975e+00 -7.76691437e-01 -3.05175781e-01
-5.13962746e-01 -6.05324507e-01 -9.27393675e-01 -8.12489808e-01
1.02762914e+00 -3.52137864e-01 8.83377910e-01 -2.61696339e-01
-4.49555933e-01 5.18958330e-01 1.36029571e-01 -2.57997811e-01
-4.17842358e-01 -8.19414183e-02 8.86702463e-02 2.98963338e-01
-1.23494916e-01 -8.41960013e-01 -5.70855856e-01 1.21287242e-01
-8.11969876e-01 2.53986567e-01 3.26826513e-01 5.76451659e-01
5.05104184e-01 -7.07206964e-01 2.63929844e-01 -8.28437030e-01
3.95257115e-01 -2.56352723e-01 -6.25159979e-01 1.59641474e-01
-2.84214765e-02 3.03651452e-01 4.34104413e-01 -4.43538934e-01
-1.31484354e+00 2.06230119e-01 -1.78629220e-01 -5.33466518e-01
-2.49903932e-01 -1.90200925e-01 -4.62125033e-01 -1.24812484e-01
6.45428956e-01 6.87727556e-02 2.37289235e-01 -3.65955710e-01
8.19336534e-01 3.47317517e-01 7.66479731e-01 -9.32232559e-01
7.91764140e-01 7.39040673e-01 1.01134762e-01 -8.16893935e-01
-6.00171506e-01 1.59482107e-01 -8.37221205e-01 -1.04649693e-01
1.07749891e+00 -9.68291998e-01 -6.79597855e-01 6.51742339e-01
-1.04412436e+00 -6.22571528e-01 -3.09727222e-01 1.42810896e-01
-8.68025422e-01 4.85459566e-02 -5.54591060e-01 -5.50292730e-01
-2.60000914e-01 -1.06683397e+00 1.58025038e+00 8.95978957e-02
-5.21628022e-01 -7.27623940e-01 -1.47919923e-01 2.74956822e-01
3.61531556e-01 4.03911948e-01 6.49376929e-01 1.21445060e-01
-8.41623127e-01 1.76979691e-01 -7.12488890e-02 7.37862363e-02
2.05344871e-01 2.45086089e-01 -1.24241662e+00 -3.27189416e-01
-1.52900144e-01 -5.31790853e-01 6.14426613e-01 5.35483479e-01
1.11669922e+00 -5.18504202e-01 -2.23632172e-01 1.07087207e+00
9.41403389e-01 -8.62238333e-02 5.29782593e-01 -2.93181628e-01
6.88561499e-01 6.76024377e-01 -6.00126125e-02 5.71898937e-01
4.07369137e-01 9.40968931e-01 1.05206454e-02 -8.08139332e-03
-3.29758763e-01 -4.98981148e-01 3.67801040e-01 2.67278463e-01
-8.65370333e-02 6.07496537e-02 -5.44005692e-01 3.88936520e-01
-1.34565055e+00 -9.59114671e-01 5.92348635e-01 2.02751827e+00
1.12880015e+00 -3.50212991e-01 8.35970417e-02 -4.91269171e-01
7.04494059e-01 -2.08126735e-02 -7.30308473e-01 -4.13231611e-01
2.48209164e-02 5.86592674e-01 1.75074026e-01 7.30172694e-01
-9.02219653e-01 1.16691434e+00 6.16906500e+00 5.34899950e-01
-1.35183060e+00 5.37822321e-02 7.31541872e-01 -3.85864943e-01
-7.73239315e-01 -1.89411491e-01 -9.68734145e-01 1.40509725e-01
5.93027294e-01 -4.70065139e-02 7.71861315e-01 6.11288249e-01
1.32148549e-01 3.65499943e-01 -1.19741631e+00 8.78587484e-01
4.10436928e-01 -1.18827081e+00 8.74809697e-02 1.44256875e-01
8.49194288e-01 -3.15688461e-01 6.02184534e-01 3.33455689e-02
4.23693299e-01 -1.11490083e+00 1.11036909e+00 7.07585335e-01
1.49315536e+00 -5.25281847e-01 -1.73442200e-01 1.17695229e-02
-7.29861617e-01 9.95957032e-02 -8.46140236e-02 8.12591314e-02
5.35092726e-02 8.26331154e-02 -6.54073298e-01 -2.34096840e-01
5.30635774e-01 3.13068479e-01 -3.19473654e-01 3.85310769e-01
-4.72798616e-01 2.19019607e-01 -3.87912095e-01 2.65776932e-01
-2.82962024e-01 -1.66699942e-02 3.91910344e-01 8.29391897e-01
3.69721681e-01 2.70866901e-01 -1.65062949e-01 1.32574916e+00
-4.26915973e-01 3.69870812e-02 -8.09763014e-01 1.02215715e-01
5.13886213e-01 1.11236143e+00 -3.59624296e-01 1.15153501e-02
-6.87006935e-02 1.15004730e+00 4.92167652e-01 5.63658595e-01
-7.49691546e-01 2.05616012e-01 9.75031435e-01 5.02703428e-01
3.61606717e-01 -2.69941628e-01 -2.97474951e-01 -1.24394274e+00
-1.87929701e-02 -8.32011759e-01 -1.10101931e-01 -9.77597713e-01
-1.12974751e+00 5.30153871e-01 2.38564417e-01 -5.37361860e-01
-5.19269466e-01 -4.95085537e-01 -4.39298213e-01 9.57738280e-01
-9.62756813e-01 -1.72267509e+00 -2.35100940e-01 5.43158650e-01
3.47757459e-01 -3.10835913e-02 9.77431238e-01 4.86373268e-02
-4.45505112e-01 9.00174856e-01 -1.18298747e-01 9.20909792e-02
8.90191317e-01 -8.48877490e-01 5.51497817e-01 4.41593379e-01
-2.47902554e-02 8.92586470e-01 7.06045508e-01 -4.87078309e-01
-1.33152139e+00 -1.01591170e+00 7.89647818e-01 -7.46263266e-01
2.04233900e-01 -9.07494366e-01 -5.59238613e-01 1.05916512e+00
1.93991393e-01 2.27498747e-02 6.93794847e-01 1.70714874e-02
-5.91801107e-01 -1.36483293e-02 -1.22588956e+00 8.93040001e-01
1.17949617e+00 -7.72050738e-01 7.55842961e-03 2.10534751e-01
5.48592865e-01 -5.41956961e-01 -7.38337576e-01 3.39202315e-01
9.63314235e-01 -8.55506241e-01 9.71932948e-01 -3.98308873e-01
3.95332605e-01 -1.41947508e-01 -1.86636880e-01 -1.11809587e+00
-4.33661878e-01 -9.30147707e-01 1.91237777e-01 1.24298418e+00
2.78605193e-01 -6.32880569e-01 6.76966429e-01 1.00007832e+00
-1.81678683e-02 -7.58478105e-01 -7.52360523e-01 -4.16056246e-01
3.59379649e-01 -6.09603943e-03 9.08175230e-01 7.33038425e-01
-3.40514958e-01 2.39356831e-01 -4.87109244e-01 -2.15431415e-02
6.46695852e-01 5.38960733e-02 8.99044871e-01 -9.06037033e-01
-4.61858869e-01 -4.40252721e-01 -2.25489214e-01 -9.10685956e-01
5.52021325e-01 -8.53773534e-01 -4.65751626e-02 -1.08623648e+00
2.39043251e-01 -3.08632165e-01 3.79057705e-01 6.13920271e-01
7.28745311e-02 5.39242625e-01 1.81085363e-01 7.55218491e-02
-1.04271129e-01 7.87530899e-01 1.57071948e+00 3.07731867e-01
-2.46319082e-02 -2.39592701e-01 -9.16344285e-01 6.90782547e-01
7.85901368e-01 -9.70583186e-02 -5.43182373e-01 -7.12499738e-01
5.42860990e-03 1.38396591e-01 6.05561137e-01 -4.76654887e-01
-7.85546452e-02 -2.98576772e-01 7.48978138e-01 1.50469825e-01
8.02476466e-01 -3.61093640e-01 4.66456801e-01 -2.12471578e-02
-4.54539120e-01 -2.01433793e-01 2.73777992e-01 2.14246348e-01
2.32481092e-01 2.14196235e-01 1.08909905e+00 -3.12754303e-01
-3.17805022e-01 5.52634776e-01 -2.10968643e-01 4.67934385e-02
8.76005709e-01 -2.32211396e-01 -3.11044399e-02 -6.69365346e-01
-1.08438981e+00 -2.66238809e-01 9.24205661e-01 4.57018673e-01
4.40878600e-01 -1.45846474e+00 -8.92852008e-01 8.14853847e-01
-4.10854854e-02 2.22751368e-02 2.33302623e-01 4.95802999e-01
-3.90611738e-01 8.31366256e-02 -3.86393070e-01 -4.60827917e-01
-1.31905174e+00 2.51739413e-01 5.94529390e-01 4.43266451e-01
-3.52080077e-01 1.25803959e+00 8.40094864e-01 -6.20024800e-01
-6.22146502e-02 8.71638581e-03 1.97600707e-01 -9.80511382e-02
2.84326583e-01 -7.12262318e-02 -3.08312923e-01 -9.79785919e-01
-1.17704667e-01 8.07321250e-01 6.55381382e-02 -3.07324588e-01
1.19259953e+00 -2.36545101e-01 -2.16501430e-01 2.09478050e-01
1.20347726e+00 1.81408435e-01 -1.69111979e+00 3.22316103e-02
-7.18364596e-01 -6.41939640e-01 -3.09502661e-01 -6.33634984e-01
-1.08705056e+00 8.67166519e-01 3.72299969e-01 -6.21262789e-01
1.02011704e+00 3.52552563e-01 5.50071478e-01 -9.56991091e-02
4.66262639e-01 -6.35345459e-01 1.74166828e-01 3.71757030e-01
1.27186847e+00 -9.28412378e-01 -2.10022584e-01 -4.29863989e-01
-6.12664521e-01 8.21677685e-01 6.77714348e-01 -8.21352601e-02
6.97039306e-01 4.08069402e-01 8.59118849e-02 -2.27334291e-01
-7.77080178e-01 2.24389490e-02 4.25761968e-01 6.94211602e-01
6.85687125e-01 2.15092674e-01 1.10190906e-01 2.74370223e-01
-6.68468118e-01 -3.53977531e-02 2.19867200e-01 4.94228929e-01
-1.85077861e-01 -1.21426392e+00 -4.02088583e-01 1.55343473e-01
-4.27113026e-01 2.95153842e-03 -4.21071649e-01 4.09258276e-01
3.39834988e-01 4.08910036e-01 1.44314602e-01 5.86469285e-02
1.05559267e-01 2.91395068e-01 1.15474951e+00 -9.18781757e-01
-1.42453045e-01 3.00188601e-01 -7.74720982e-02 -3.58222395e-01
-1.09184898e-01 -9.75889862e-01 -1.10686338e+00 -5.07787645e-01
-8.24820846e-02 -3.09420168e-01 4.90442336e-01 6.65655613e-01
5.36643386e-01 6.61663413e-02 5.03395557e-01 -1.21893418e+00
-4.70607162e-01 -7.18178749e-01 -6.13649964e-01 5.84677339e-01
3.38372380e-01 -6.92931175e-01 -2.64824331e-01 3.65571469e-01] | [12.717199325561523, -0.29717379808425903] |
bf903e0c-c1f0-476f-9985-ca4e800157b8 | parameter-sensitivity-of-deep-feature-based | 2208.10743 | null | https://arxiv.org/abs/2208.10743v1 | https://arxiv.org/pdf/2208.10743v1.pdf | Parameter Sensitivity of Deep-Feature based Evaluation Metrics for Audio Textures | Standard evaluation metrics such as the Inception score and Fr\'echet Audio Distance provide a general audio quality distance metric between the synthesized audio and reference clean audio. However, the sensitivity of these metrics to variations in the statistical parameters that define an audio texture is not well studied. In this work, we provide a systematic study of the sensitivity of some of the existing audio quality evaluation metrics to parameter variations in audio textures. Furthermore, we also study three more potentially parameter-sensitive metrics for audio texture synthesis, (a) a Gram matrix based distance, (b) an Accumulated Gram metric using a summarized version of the Gram matrices, and (c) a cochlear-model based statistical features metric. These metrics use deep features that summarize the statistics of any given audio texture, thus being inherently sensitive to variations in the statistical parameters that define an audio texture. We study and evaluate the sensitivity of existing standard metrics as well as Gram matrix and cochlear-model based metrics to control-parameter variations in audio textures across a wide range of texture and parameter types, and validate with subjective evaluation. We find that each of the metrics is sensitive to different sets of texture-parameter types. This is the first step towards investigating objective metrics for assessing parameter sensitivity in audio textures. | ['Lonce Wyse', 'Zhuoyao Li', 'Purnima Kamath', 'Zequn Gong', 'Yize Wei', 'Chitralekha Gupta'] | 2022-08-23 | null | null | null | null | ['texture-synthesis'] | ['computer-vision'] | [ 2.29821101e-01 -5.67366421e-01 4.59355593e-01 -2.74438530e-01
-1.44046462e+00 -8.56345057e-01 4.97225761e-01 5.86996675e-01
-1.93073407e-01 3.86574209e-01 5.04140258e-01 6.49430379e-02
-6.16509378e-01 -5.34942687e-01 -2.31815204e-01 -8.33307683e-01
-4.95664895e-01 -5.12751043e-02 5.02346992e-01 -3.13464075e-01
4.50050622e-01 3.59570712e-01 -2.03702736e+00 4.89767313e-01
3.25366914e-01 1.51200485e+00 -9.57729816e-02 1.06206846e+00
3.08328658e-01 2.66720831e-01 -9.94301677e-01 1.18832916e-01
-3.57712768e-02 -5.80315471e-01 -5.24277031e-01 -7.37426579e-02
5.42296767e-01 3.08691204e-01 4.33379374e-02 1.11896503e+00
9.68078494e-01 9.92086306e-02 7.74112999e-01 -8.54626536e-01
-1.79751366e-01 6.67875230e-01 1.07898971e-03 5.33822596e-01
8.64884019e-01 -5.74042946e-02 1.07649028e+00 -6.82211161e-01
3.41989964e-01 1.35402393e+00 8.95547509e-01 -7.37779811e-02
-1.12330854e+00 -4.73450840e-01 -3.84204388e-01 1.74369663e-01
-1.55990958e+00 -7.69361079e-01 5.65474331e-01 -7.05296814e-01
5.71537673e-01 6.98455513e-01 3.45734566e-01 6.58188999e-01
5.27583420e-01 2.02388287e-01 1.50651562e+00 -7.30632663e-01
3.51531118e-01 -8.30390956e-03 -4.94946726e-02 2.09115550e-01
-2.09241539e-01 3.15473229e-01 -7.42585123e-01 -3.67374092e-01
5.40599048e-01 -9.59123194e-01 -3.63824546e-01 -1.68104693e-01
-1.19082439e+00 4.94653404e-01 -9.79282111e-02 6.59955442e-01
2.76755899e-01 3.37109476e-01 5.68621278e-01 7.54816592e-01
3.21865350e-01 7.56688297e-01 -3.11605781e-01 -6.56613708e-01
-7.52023339e-01 3.63983989e-01 8.96572053e-01 5.16629994e-01
5.47750175e-01 3.58043611e-01 -3.65556121e-01 1.22884834e+00
2.78177619e-01 6.63332164e-01 5.46385109e-01 -9.44908917e-01
2.55596787e-01 -1.68638602e-01 -1.05553411e-01 -1.26396596e+00
-4.08597916e-01 -3.91703188e-01 -4.02951866e-01 2.23488614e-01
6.10835373e-01 2.42792070e-02 -5.45200109e-01 1.61085546e+00
9.24454033e-02 -9.82956663e-02 -1.98680177e-01 7.88904428e-01
8.57240081e-01 3.66190046e-01 -4.55337286e-01 5.43148406e-02
1.18045104e+00 -2.47438729e-01 -6.18846059e-01 2.42013231e-01
3.45072865e-01 -1.49502480e+00 1.40531790e+00 6.01896107e-01
-1.09179771e+00 -6.86417043e-01 -1.27560961e+00 3.53618920e-01
-2.85485476e-01 -1.42322108e-01 -1.10411728e-02 1.17634606e+00
-1.19092643e+00 6.52024984e-01 -3.70305121e-01 -3.53691071e-01
-3.51356298e-01 3.86041284e-01 -2.65721023e-01 2.74207860e-01
-1.08405674e+00 5.69448233e-01 -3.36305909e-02 -1.55209646e-01
-9.00867283e-01 -6.52909696e-01 -6.79450333e-01 5.94537295e-02
-4.23304364e-02 -6.96819574e-02 1.43789947e+00 -5.41774392e-01
-1.79306376e+00 4.72866654e-01 2.17850327e-01 -6.50758073e-02
2.98311025e-01 3.37980986e-02 -9.80084300e-01 9.19820294e-02
-1.81629702e-01 2.15021282e-01 9.50964630e-01 -1.28068173e+00
-5.13888896e-01 1.76433176e-02 -1.27498537e-01 4.67007235e-02
-1.09375268e-01 3.39703321e-01 -1.55624598e-01 -1.05537331e+00
1.54792964e-01 -8.48236084e-01 2.25782424e-01 -2.12884650e-01
-2.86317170e-01 2.96200514e-01 4.03439790e-01 -4.68948781e-01
1.49774146e+00 -2.53600931e+00 -1.20436363e-01 5.45488060e-01
-1.59440473e-01 -1.43944994e-01 -4.11647886e-01 6.45418286e-01
-3.42009775e-02 1.45331547e-01 2.95101805e-03 1.59152403e-01
2.64671117e-01 -2.07668200e-01 -2.37757280e-01 3.91986758e-01
-1.00608811e-01 2.02415347e-01 -7.28607118e-01 -2.16581866e-01
1.58605680e-01 6.32628560e-01 -6.56138539e-01 -1.25421599e-01
1.73991054e-01 3.82761240e-01 -1.01316497e-01 5.12029707e-01
2.80412227e-01 4.42864984e-01 -2.23226905e-01 -4.76743847e-01
-5.86263016e-02 5.07828832e-01 -1.48008835e+00 1.19337213e+00
-6.39354229e-01 1.02801943e+00 -8.00899342e-02 -5.85870922e-01
1.20796025e+00 5.00388026e-01 3.48511487e-01 -7.94573784e-01
1.43888280e-01 5.72311342e-01 3.61050248e-01 -3.74168366e-01
5.21597385e-01 -2.11618051e-01 -2.87532747e-01 4.43514615e-01
2.21679404e-01 -7.22567201e-01 1.37506351e-01 -3.02472621e-01
1.02344286e+00 -3.97373378e-01 1.35465950e-01 -7.05109000e-01
5.63491762e-01 -8.23513746e-01 1.88511312e-01 8.41369629e-01
-1.78522408e-01 1.06544268e+00 5.57565629e-01 7.95721337e-02
-9.10563767e-01 -1.21304274e+00 -6.26020849e-01 1.06741285e+00
-8.95310640e-02 -7.28334546e-01 -7.80510843e-01 9.42659825e-02
5.75902052e-02 1.40235856e-01 -6.36981964e-01 -1.46704108e-01
-1.38377294e-01 -6.76853776e-01 1.02128148e+00 2.22098321e-01
1.45331115e-01 -6.79287195e-01 -4.49289948e-01 2.72981256e-01
-1.27616867e-01 -9.16287959e-01 -4.80078608e-01 3.16433698e-01
-5.54594398e-01 -8.17821503e-01 -3.54309469e-01 -4.94749576e-01
-9.23852548e-02 -3.07340354e-01 1.13520658e+00 -2.89055288e-01
-3.31470162e-01 6.88589692e-01 -7.00200260e-01 -4.86915201e-01
-6.93003774e-01 -1.32269219e-01 4.06940997e-01 3.18874896e-01
-3.57332826e-01 -7.41737962e-01 -5.25184155e-01 7.42905676e-01
-9.89475787e-01 -8.56030464e-01 1.87035665e-01 4.64697480e-01
7.52926409e-01 3.89051706e-01 4.78976697e-01 -2.13035166e-01
1.06298733e+00 -1.92193940e-01 -2.02366486e-01 -2.03686692e-02
-1.89392269e-01 5.08634038e-02 3.41370523e-01 -5.27563989e-01
-4.21857387e-01 -4.99599993e-01 -2.33837560e-01 -1.62787875e-03
-1.71618368e-02 6.75866783e-01 -2.54376441e-01 -3.84159744e-01
9.56035972e-01 -6.00940175e-02 -2.30197653e-01 -6.12103105e-01
-2.48152204e-02 7.77315855e-01 5.56928694e-01 -7.98497856e-01
6.27350450e-01 2.58298784e-01 1.31968921e-03 -1.21154428e+00
-4.87487793e-01 -4.39043254e-01 -2.82952368e-01 -3.09733063e-01
5.02242982e-01 -5.97763538e-01 -4.29078609e-01 5.70561349e-01
-5.61688304e-01 -3.76965642e-01 -2.84346640e-01 7.36410320e-01
-7.91393936e-01 3.03266734e-01 -5.44608474e-01 -7.77086020e-01
1.54834092e-01 -1.41400003e+00 1.41664505e+00 -1.24235660e-01
-5.69503963e-01 -1.08817947e+00 2.88182020e-01 -1.01302125e-01
6.35166705e-01 3.24475527e-01 9.24380779e-01 -2.54606217e-01
-1.00346558e-01 -2.69947678e-01 6.95701987e-02 4.50005919e-01
4.30038661e-01 3.20456624e-01 -1.15209067e+00 -4.33382541e-01
1.51379675e-01 -2.83745825e-02 5.68723202e-01 5.43931425e-01
9.86619234e-01 -2.57210910e-01 2.97643960e-01 5.65175772e-01
1.26341808e+00 3.94720018e-01 6.85591161e-01 3.16888839e-01
1.98180422e-01 4.76977259e-01 6.07655108e-01 4.83493090e-01
-2.19690591e-01 1.07517934e+00 1.33514598e-01 6.42941073e-02
-2.81089008e-01 4.10878882e-02 6.24900877e-01 1.42075169e+00
1.32008279e-02 -2.29693949e-01 -8.01350355e-01 3.42593342e-01
-1.06555104e+00 -7.26398289e-01 2.93809809e-02 2.32073665e+00
9.51527178e-01 1.91768423e-01 3.84898484e-01 9.41961467e-01
5.16690016e-01 1.44030243e-01 1.01238407e-01 -7.18146920e-01
-2.99754322e-01 6.37215972e-01 2.44817451e-01 6.39004886e-01
-1.05295122e+00 3.94543082e-01 7.56550646e+00 1.08951986e+00
-1.33297312e+00 -8.32721442e-02 2.96089083e-01 -8.53770375e-02
-1.73234627e-01 -1.40961885e-01 -5.21136284e-01 4.54847425e-01
1.15359461e+00 -2.18697280e-01 4.35532749e-01 3.61470133e-01
2.58430779e-01 -3.07424843e-01 -1.12187743e+00 1.08061874e+00
1.12215266e-01 -9.31206703e-01 9.19316802e-03 9.61442590e-02
5.70461571e-01 -6.67013079e-02 5.44009149e-01 -2.00091735e-01
-2.41564035e-01 -1.12891495e+00 1.08655775e+00 5.44397175e-01
1.05313492e+00 -6.37693048e-01 6.88796639e-01 -5.29387832e-01
-1.42267549e+00 5.36040477e-02 -1.53075740e-01 7.55120292e-02
-5.81362890e-03 8.87398899e-01 -5.54679275e-01 3.65355015e-01
1.09749699e+00 2.26424426e-01 -7.61882603e-01 1.58513713e+00
2.54264742e-01 8.25433552e-01 -4.19107139e-01 -2.19590981e-02
4.55213524e-02 2.48772144e-01 8.12056839e-01 1.55359721e+00
7.54222274e-01 -4.34660971e-01 -1.84914485e-01 4.20366317e-01
5.52597165e-01 3.52686077e-01 -3.36272955e-01 -1.59767177e-02
5.96604526e-01 9.28683698e-01 -7.52170622e-01 1.45237893e-01
2.41017025e-02 2.85785973e-01 -5.93082130e-01 3.57066274e-01
-5.78399479e-01 -7.78787494e-01 9.58295524e-01 3.00696522e-01
3.62718493e-01 -3.11577439e-01 -1.97450727e-01 -6.63323462e-01
-6.42685443e-02 -1.31963015e+00 5.25661446e-02 -8.85033429e-01
-1.23553455e+00 8.04641902e-01 6.01206757e-02 -1.68187535e+00
-1.73688307e-01 -5.63248575e-01 -5.06529272e-01 7.49648869e-01
-9.25420284e-01 -3.92469496e-01 -1.24031596e-01 4.90702987e-01
3.33656780e-02 -3.47016782e-01 9.26347792e-01 4.84311819e-01
-8.32196996e-02 9.09929395e-01 3.02518338e-01 -2.53323346e-01
8.38153362e-01 -1.23298252e+00 1.18246309e-01 3.73216689e-01
3.81786406e-01 5.59075296e-01 1.22359908e+00 -7.33570084e-02
-1.11098993e+00 -7.56771147e-01 5.46332479e-01 -5.29448450e-01
9.28846955e-01 -4.71294373e-01 -8.48469257e-01 8.75498131e-02
-1.68372065e-01 -2.19843030e-01 9.50202107e-01 1.57439038e-01
-4.32659209e-01 -4.38639164e-01 -8.69150639e-01 4.48433608e-01
6.64091706e-01 -1.11044061e+00 -3.50321412e-01 1.64409950e-01
4.56110924e-01 -3.91930848e-01 -1.44139111e+00 4.94256407e-01
8.84005845e-01 -1.19764912e+00 9.74571764e-01 1.53653203e-02
8.15304518e-02 -6.24710858e-01 -8.13959181e-01 -1.51885200e+00
-2.82578111e-01 -9.37341332e-01 3.61986369e-01 1.36147797e+00
6.16744578e-01 -5.69872200e-01 7.57473260e-02 -2.04820558e-01
-1.83689505e-01 -6.16301179e-01 -1.13335395e+00 -1.10020626e+00
7.24896714e-02 -8.00524533e-01 8.30537140e-01 7.29161382e-01
2.25245148e-01 1.63413495e-01 8.20049942e-02 8.81684572e-02
2.89711446e-01 -3.24973792e-01 6.72819018e-01 -1.12415469e+00
-5.22846758e-01 -7.15550065e-01 -9.71647084e-01 -4.01369125e-01
-5.32501459e-01 -7.49223351e-01 1.82197675e-01 -9.27446187e-01
-3.28604996e-01 -7.11699605e-01 -6.03665411e-01 -5.17900549e-02
1.60436183e-01 6.18590593e-01 7.11093843e-02 -1.22326113e-01
-3.05349529e-01 3.57896239e-01 1.01777339e+00 -1.18466362e-01
-2.18958706e-01 2.99351010e-03 -5.37186086e-01 6.20024502e-01
7.37323165e-01 -4.07760024e-01 -4.15302306e-01 -2.60675788e-01
4.18225795e-01 -4.91887890e-02 3.37999254e-01 -1.62446618e+00
-1.42677471e-01 9.42160338e-02 3.45833451e-02 -1.10548884e-01
2.58533776e-01 -3.99129421e-01 2.02400714e-01 5.62357418e-02
-3.65215868e-01 2.67015249e-01 4.47984189e-01 3.88761967e-01
-7.41007030e-01 -6.05016239e-02 9.60847616e-01 3.62220615e-01
-4.36193615e-01 -1.77182227e-01 -6.10707760e-01 1.93260506e-01
2.58852810e-01 -5.59926033e-01 -4.71977107e-02 -6.65201962e-01
-8.62461567e-01 -5.71824074e-01 3.77075881e-01 5.42921960e-01
2.74885356e-01 -1.46142495e+00 -6.47588015e-01 3.09909552e-01
2.57214218e-01 -6.35347009e-01 6.64204806e-02 9.79399085e-01
-5.95088542e-01 3.96006644e-01 -2.39758849e-01 -1.01063097e+00
-1.48215795e+00 6.76686391e-02 7.04955459e-01 9.53149125e-02
-5.32023422e-02 9.72834826e-01 1.46778062e-01 -3.09551626e-01
3.29810947e-01 -9.52581882e-01 -9.76966918e-02 1.68849185e-01
4.64406967e-01 3.67643476e-01 6.34561121e-01 -8.50489318e-01
-3.59207809e-01 9.84831095e-01 5.20860910e-01 -7.04389691e-01
8.98349583e-01 -2.51161069e-01 -5.06006479e-02 1.02412915e+00
1.37486172e+00 3.99640948e-01 -7.28277385e-01 2.01963522e-02
-2.20654458e-02 -4.92467433e-01 7.43650049e-02 -8.12707722e-01
-8.73390436e-01 7.96700478e-01 1.05912519e+00 5.88700593e-01
1.34394562e+00 -1.42019644e-01 3.87773186e-01 4.53187786e-02
4.43284094e-01 -1.13417041e+00 3.43098998e-01 6.23732448e-01
1.14958549e+00 -5.04605532e-01 -1.04948923e-01 -3.74851942e-01
-3.20967287e-01 1.09398139e+00 -7.46133849e-02 -1.26639884e-02
1.03449404e+00 6.31806374e-01 5.52589118e-01 -6.47223890e-02
-5.23087144e-01 -1.11829534e-01 6.71291173e-01 8.20287943e-01
7.40969718e-01 1.50900707e-01 -5.74657060e-02 3.15113902e-01
-1.06701517e+00 -3.77792746e-01 4.49313611e-01 6.00473940e-01
-5.58512688e-01 -1.16992176e+00 -6.73281610e-01 3.72647405e-01
-7.01895833e-01 -8.10557902e-02 -4.57523882e-01 4.98160630e-01
1.39964810e-02 1.30738854e+00 -1.03755051e-03 -7.95459270e-01
5.70174396e-01 -3.91820185e-02 7.44653463e-01 -5.69396138e-01
-7.18153179e-01 3.09709907e-01 2.22020656e-01 -4.75023419e-01
-3.99088502e-01 -9.53865469e-01 -8.07501256e-01 -2.09521562e-01
-4.82891351e-01 1.71211332e-01 6.42050087e-01 8.21052372e-01
5.38175479e-02 6.20008647e-01 6.30691230e-01 -8.95804644e-01
-3.00305873e-01 -1.08300519e+00 -1.01632714e+00 2.94248134e-01
5.88588595e-01 -7.67461836e-01 -6.66371942e-01 -1.95061043e-02] | [15.450896263122559, 5.5339508056640625] |
542ad8cc-bf39-4ba9-b591-79454afbc265 | d-d-learning-human-dynamics-from-dynamic | 2209.08790 | null | https://arxiv.org/abs/2209.08790v1 | https://arxiv.org/pdf/2209.08790v1.pdf | D&D: Learning Human Dynamics from Dynamic Camera | 3D human pose estimation from a monocular video has recently seen significant improvements. However, most state-of-the-art methods are kinematics-based, which are prone to physically implausible motions with pronounced artifacts. Current dynamics-based methods can predict physically plausible motion but are restricted to simple scenarios with static camera view. In this work, we present D&D (Learning Human Dynamics from Dynamic Camera), which leverages the laws of physics to reconstruct 3D human motion from the in-the-wild videos with a moving camera. D&D introduces inertial force control (IFC) to explain the 3D human motion in the non-inertial local frame by considering the inertial forces of the dynamic camera. To learn the ground contact with limited annotations, we develop probabilistic contact torque (PCT), which is computed by differentiable sampling from contact probabilities and used to generate motions. The contact state can be weakly supervised by encouraging the model to generate correct motions. Furthermore, we propose an attentive PD controller that adjusts target pose states using temporal information to obtain smooth and accurate pose control. Our approach is entirely neural-based and runs without offline optimization or simulation in physics engines. Experiments on large-scale 3D human motion benchmarks demonstrate the effectiveness of D&D, where we exhibit superior performance against both state-of-the-art kinematics-based and dynamics-based methods. Code is available at https://github.com/Jeffsjtu/DnD | ['Cewu Lu', 'Gang Yu', 'Gang Liu', 'Chao Xu', 'Siyuan Bian', 'Jiefeng Li'] | 2022-09-19 | null | null | null | null | ['3d-human-pose-estimation', 'human-dynamics'] | ['computer-vision', 'computer-vision'] | [-2.74273425e-01 5.30823395e-02 -2.55025983e-01 1.44871697e-01
-3.99388015e-01 -4.31728274e-01 5.95312893e-01 -6.00114465e-01
-2.32340157e-01 7.58724570e-01 3.87444019e-01 4.09423606e-03
2.00934038e-01 -4.48656112e-01 -1.19834149e+00 -5.49032032e-01
-5.84859513e-02 6.21695876e-01 2.91033566e-01 -4.11552846e-01
-1.87888339e-01 4.05630916e-01 -1.20060325e+00 -1.97693005e-01
5.77061176e-01 4.74158347e-01 3.21130842e-01 1.15782642e+00
6.58666193e-01 8.24392736e-01 -3.45431447e-01 -4.28912602e-02
4.86443400e-01 -5.31932831e-01 -6.04075193e-01 1.80892408e-01
4.01051253e-01 -6.21453881e-01 -1.01400769e+00 6.43449485e-01
7.25845575e-01 5.01717985e-01 3.66108745e-01 -1.23978126e+00
-4.59808260e-01 -8.09562113e-03 -3.96241844e-01 1.28984926e-02
1.07534027e+00 8.41392636e-01 5.39234579e-01 -7.61845350e-01
9.95435536e-01 1.49237061e+00 7.52642751e-01 1.03283525e+00
-1.03875923e+00 -1.84844121e-01 2.82340616e-01 2.05552444e-01
-1.24386048e+00 -1.08507544e-01 7.75641739e-01 -5.17891347e-01
1.04641616e+00 1.49186835e-01 1.17934608e+00 1.62996662e+00
6.80056930e-01 9.97693837e-01 6.72840118e-01 -2.19690368e-01
1.35120735e-01 -5.75137615e-01 -3.44157130e-01 9.71978068e-01
2.22760171e-01 5.00764608e-01 -6.63827121e-01 -1.02652922e-01
1.26247549e+00 8.93518031e-02 -5.67284644e-01 -7.79561043e-01
-1.56745410e+00 4.36466396e-01 5.20549476e-01 -3.33758593e-01
-5.44494748e-01 6.62593246e-01 1.14123702e-01 -2.80659646e-01
2.44331509e-01 2.75069356e-01 -6.27017736e-01 -3.71461034e-01
-4.03418750e-01 1.05165827e+00 7.12049901e-01 1.08710372e+00
1.06821761e-01 1.40287831e-01 -1.17040463e-01 1.29133835e-01
2.29235470e-01 7.77503967e-01 2.26086378e-01 -1.52538431e+00
5.26782572e-01 2.67392308e-01 5.97517848e-01 -9.62846696e-01
-6.56563997e-01 -1.83106944e-01 -6.17804408e-01 2.92709947e-01
5.23814619e-01 -3.51660341e-01 -8.79001141e-01 1.72501028e+00
7.70709038e-01 4.88055855e-01 -1.25236973e-01 1.52778220e+00
3.23556632e-01 7.09049404e-01 -2.08250582e-01 1.84904277e-01
1.00216639e+00 -1.03854477e+00 -5.97749949e-01 -3.58703047e-01
4.67466325e-01 -3.98148030e-01 1.11619294e+00 4.63393718e-01
-1.28661323e+00 -7.54080534e-01 -8.14397871e-01 -1.54797226e-01
2.30008826e-01 9.43546519e-02 2.93231964e-01 1.00091957e-01
-7.33370781e-01 8.91187191e-01 -1.62790489e+00 -4.66364771e-01
-1.15529627e-01 3.32398951e-01 -4.28651333e-01 1.20529467e-02
-1.17220676e+00 1.14583158e+00 2.48410299e-01 2.28769854e-01
-1.19394124e+00 -7.71980822e-01 -1.13163149e+00 -4.92935747e-01
7.44361401e-01 -1.58104181e+00 1.28176665e+00 -4.11989748e-01
-2.04480267e+00 4.14916337e-01 -1.23430066e-01 -4.60192710e-01
1.04673696e+00 -1.06529057e+00 1.39437243e-01 4.99218673e-01
-5.61097786e-02 9.70984042e-01 6.86269522e-01 -1.40843439e+00
-3.63252223e-01 -2.65262406e-02 5.65510616e-02 6.39850497e-01
3.53015214e-01 -5.52894592e-01 -8.14750850e-01 -7.43767142e-01
9.71181970e-03 -1.60420096e+00 -4.04564917e-01 3.39259207e-01
-5.48821449e-01 1.18280761e-01 5.98843515e-01 -7.85125554e-01
9.03286636e-01 -1.57454789e+00 8.58204186e-01 -2.32567698e-01
3.31070870e-02 4.80079614e-02 2.76326984e-01 2.60970652e-01
2.79269844e-01 -2.89613008e-01 -1.29755750e-01 -3.96490872e-01
7.27572069e-02 3.36265385e-01 -2.88877428e-01 6.12504601e-01
1.78872555e-01 1.10731447e+00 -1.09213018e+00 -3.08537781e-01
5.95966518e-01 7.94448078e-01 -9.00472462e-01 4.36643690e-01
-5.07236779e-01 1.00240171e+00 -5.73682308e-01 4.99490887e-01
1.53104782e-01 -2.33873188e-01 1.54434100e-01 -2.25706458e-01
-3.65084000e-02 1.25297353e-01 -1.23319840e+00 2.10106921e+00
-2.17092514e-01 1.47440076e-01 -5.16418330e-02 -5.35283566e-01
2.51912177e-01 3.60645562e-01 5.51174760e-01 -6.93142191e-02
3.22207510e-01 -1.12428097e-02 -2.00998038e-01 -6.07700825e-01
4.73098308e-01 8.38288516e-02 -1.20498404e-01 -2.42778193e-02
-1.31760672e-01 -4.13913190e-01 -2.60597408e-01 7.05969408e-02
1.21702564e+00 1.20562458e+00 2.01915070e-01 9.47605446e-03
2.83505857e-01 4.22179401e-01 6.97307408e-01 5.77055693e-01
-3.02745014e-01 9.13053393e-01 -7.30340881e-03 -4.26521838e-01
-1.28505826e+00 -1.38173926e+00 4.37557697e-01 4.89333898e-01
5.80465317e-01 -2.98677713e-01 -9.22665894e-01 -3.33597690e-01
1.62465513e-01 4.32525843e-01 -4.43581671e-01 -3.52318734e-01
-1.14732802e+00 -3.97160023e-01 3.14046204e-01 7.37385035e-01
3.57952535e-01 -8.44660282e-01 -1.10833478e+00 2.45930418e-01
-4.40018058e-01 -1.37000930e+00 -7.80087888e-01 -3.14892322e-01
-8.37046146e-01 -1.14765501e+00 -1.01345026e+00 -4.15181577e-01
6.06039166e-01 1.22805394e-01 9.01885569e-01 -2.78542470e-02
-4.38368469e-01 7.78237700e-01 -1.70189843e-01 -5.45541756e-02
-2.26660624e-01 -1.50531754e-01 6.70640171e-01 -3.93163592e-01
-1.79662198e-01 -3.92266363e-01 -9.04098988e-01 5.12505233e-01
-4.22392368e-01 4.15813535e-01 3.23228329e-01 8.44880581e-01
7.95419574e-01 -2.20555097e-01 -1.65199682e-01 -2.73897290e-01
9.36557874e-02 -2.38925979e-01 -3.98955286e-01 -2.55010068e-01
8.63871817e-03 -1.49292583e-02 6.15216553e-01 -7.43761241e-01
-1.11898935e+00 5.47739148e-01 1.40059618e-02 -9.53144073e-01
-2.55783588e-01 7.96385109e-02 -8.83524418e-02 2.16705754e-01
6.56398714e-01 1.26616791e-01 1.72332730e-02 -2.68154442e-01
4.29126322e-01 -1.72179818e-01 1.04519463e+00 -8.57400358e-01
1.00411880e+00 7.45903552e-01 2.21831620e-01 -6.15972400e-01
-8.02237093e-01 -3.99279982e-01 -8.76174271e-01 -4.63651627e-01
1.14090550e+00 -1.16688228e+00 -1.07297206e+00 6.69738829e-01
-1.15263236e+00 -9.03714895e-01 -1.81211248e-01 8.78440499e-01
-1.12076139e+00 4.21039581e-01 -8.91026318e-01 -7.58178294e-01
-4.94214892e-02 -1.16199982e+00 1.39840508e+00 8.39079693e-02
-5.43299794e-01 -8.77193809e-01 2.44610503e-01 2.91655630e-01
-2.94031292e-01 8.25230896e-01 2.75392711e-01 2.27399677e-01
-6.49992049e-01 -1.61605924e-01 5.25113404e-01 -6.94718435e-02
-1.03913210e-01 -9.11421180e-02 -4.67210889e-01 -3.29720676e-01
-1.05755784e-01 -2.94554383e-01 4.26352859e-01 7.78869450e-01
7.59998739e-01 -3.56608897e-01 -5.32389700e-01 6.39661133e-01
1.04420221e+00 -1.92332476e-01 5.34783900e-01 3.74507964e-01
1.17299902e+00 6.62323117e-01 7.63600886e-01 6.04568720e-01
6.55892491e-01 1.02933574e+00 4.58785355e-01 2.45631367e-01
-2.15476692e-01 -7.43541300e-01 7.18110323e-01 7.01745033e-01
-4.73321706e-01 -2.33057901e-01 -9.54688370e-01 4.20820028e-01
-2.17230368e+00 -9.56103563e-01 -1.52703017e-01 1.99856269e+00
7.58383393e-01 3.28963429e-01 3.26290399e-01 -9.87850428e-02
6.40031457e-01 -2.39190180e-02 -7.90604234e-01 3.41892242e-01
1.78852066e-01 -5.11273555e-02 5.71268916e-01 7.70800591e-01
-9.11950409e-01 1.04553843e+00 5.33359194e+00 2.65787989e-01
-1.00131857e+00 -1.09568862e-02 7.95046911e-02 -5.11932492e-01
7.40639493e-02 -9.00281244e-04 -9.30116296e-01 3.32569569e-01
7.45635033e-01 5.88810183e-02 2.99741358e-01 8.15376282e-01
7.69423366e-01 -1.93146437e-01 -1.26721084e+00 8.95433486e-01
-4.43917699e-02 -1.22891510e+00 4.45416346e-02 7.77243376e-02
8.79485965e-01 -1.13220572e-01 -1.14132471e-01 6.92840889e-02
4.74552900e-01 -6.71224594e-01 1.25458145e+00 9.21324551e-01
3.51369023e-01 -6.14995718e-01 3.56220990e-01 7.52023160e-01
-1.23649836e+00 7.83303306e-02 -1.40556544e-01 -3.36868763e-01
7.97254920e-01 2.74128020e-01 -3.71570826e-01 6.52955294e-01
7.64335573e-01 9.20939982e-01 -1.06358908e-01 7.08020449e-01
-4.95649278e-01 3.54818702e-01 -5.61021745e-01 1.36596989e-02
1.69206023e-01 -9.81266201e-02 1.01166034e+00 7.99237251e-01
3.09223562e-01 4.85817879e-01 7.64826655e-01 8.12903941e-01
2.69789636e-01 -4.81378496e-01 -6.05990946e-01 4.98101711e-01
9.39919874e-02 8.70929003e-01 -3.93116295e-01 -3.36469769e-01
6.21274412e-02 1.36686695e+00 9.32670981e-02 2.08717719e-01
-1.11271155e+00 1.37383014e-01 7.54092395e-01 3.77325594e-01
1.55095980e-01 -8.79758716e-01 -1.96075663e-02 -1.47164273e+00
3.47729802e-01 -6.93385839e-01 1.33875772e-01 -1.07328725e+00
-9.67043638e-01 1.85730696e-01 5.15341222e-01 -1.36646974e+00
-5.75547278e-01 -6.29348755e-01 -3.30940008e-01 6.39284730e-01
-1.03000045e+00 -9.84857917e-01 -4.58881676e-01 6.57710493e-01
7.22581208e-01 3.28035563e-01 4.49732035e-01 -7.54269958e-02
-3.25541854e-01 9.37229767e-02 -4.21384066e-01 2.51419917e-02
6.84959114e-01 -1.29966831e+00 8.96672368e-01 7.98321009e-01
-2.16173247e-01 5.40706336e-01 1.12520945e+00 -1.05891085e+00
-1.94973791e+00 -1.04162979e+00 5.27374744e-01 -1.11090720e+00
5.75713575e-01 -3.10034961e-01 -6.58594072e-01 7.31927335e-01
-2.87391335e-01 3.48312825e-01 -1.22987643e-01 -7.07614124e-01
1.70743600e-01 4.38514739e-01 -8.53256285e-01 1.07928371e+00
1.59675729e+00 -2.80791879e-01 -6.71160638e-01 4.25530076e-01
8.45773160e-01 -1.23368132e+00 -6.80117548e-01 4.03278738e-01
7.33628571e-01 -6.11967146e-01 1.23193669e+00 -6.08901143e-01
5.15449286e-01 -6.22953176e-01 -4.50898670e-02 -1.23757076e+00
-1.52511463e-01 -1.09550655e+00 -7.50007987e-01 4.03082460e-01
-1.18904658e-01 -2.42184252e-01 1.03404105e+00 5.49025953e-01
-2.33259633e-01 -8.67748559e-01 -8.58135164e-01 -1.12606573e+00
-1.94541495e-02 -4.84731674e-01 1.29748672e-01 5.21071851e-01
-9.22308788e-02 -1.02531351e-01 -7.97592103e-01 6.11288965e-01
7.44900107e-01 -3.27039748e-01 1.31972802e+00 -6.37628496e-01
-7.46296585e-01 -2.28812844e-02 -3.22748393e-01 -1.56506062e+00
3.95570964e-01 -3.22993040e-01 5.06637692e-01 -1.50828552e+00
-8.04162621e-02 8.49554762e-02 4.48174447e-01 1.69127375e-01
-4.46584314e-01 5.97962029e-02 4.41053331e-01 3.95663708e-01
-3.70838135e-01 7.96798527e-01 1.53681374e+00 1.95899069e-01
-4.12570655e-01 -5.14172390e-03 9.94337499e-02 1.12753570e+00
5.04661202e-01 -4.23611432e-01 -4.13653910e-01 -4.34499532e-01
-5.15115671e-02 3.72158468e-01 1.10674345e+00 -1.30952430e+00
2.43326589e-01 -3.24947774e-01 5.16119480e-01 -6.17766798e-01
6.27491415e-01 -6.08573079e-01 6.07816339e-01 9.73215938e-01
-2.78309882e-01 3.50623846e-01 4.26003449e-02 8.64660144e-01
2.90166616e-01 3.87869656e-01 6.44506097e-01 -3.57036710e-01
-6.45898938e-01 5.24518609e-01 -1.93403661e-01 1.98036313e-01
1.00320566e+00 -2.61312097e-01 -7.59026408e-02 -5.51537514e-01
-9.28633869e-01 2.33102888e-01 6.99556828e-01 5.69289327e-01
6.26178622e-01 -1.36441934e+00 -5.37504494e-01 -3.01876813e-01
-1.35440320e-01 4.39738333e-01 3.34442914e-01 7.46750534e-01
-9.74582434e-01 3.51999432e-01 -8.35187808e-02 -9.45586026e-01
-8.63870740e-01 5.39884865e-01 4.82392371e-01 -1.68228716e-01
-1.18080747e+00 5.55024385e-01 2.69353628e-01 -5.35297275e-01
1.02810502e-01 -4.41816658e-01 4.00636494e-01 -8.58886123e-01
2.19007835e-01 6.62986696e-01 -3.39879185e-01 -8.53214622e-01
-2.98633069e-01 8.49763811e-01 3.54918927e-01 -5.21816730e-01
1.08704329e+00 -1.88657358e-01 5.85976064e-01 3.57136011e-01
8.59842002e-01 -2.04534009e-01 -2.28983712e+00 1.25553906e-01
-3.65002364e-01 -4.57991689e-01 -2.85207599e-01 -5.89315951e-01
-7.13815689e-01 6.84354484e-01 2.69914657e-01 -5.61039388e-01
6.52114570e-01 -8.36407766e-02 1.25424087e+00 5.03381073e-01
7.53816843e-01 -1.13025570e+00 3.17358732e-01 5.99860728e-01
1.06655765e+00 -9.66275036e-01 1.74038306e-01 -5.60393572e-01
-7.66102493e-01 1.01974273e+00 9.39677417e-01 -5.74796140e-01
6.02741718e-01 2.19936639e-01 -8.15545619e-02 -2.00601500e-02
-7.81593919e-01 6.15869053e-02 4.35029387e-01 7.03394055e-01
7.52567947e-02 -6.40971493e-03 1.21809877e-01 3.53751987e-01
-3.45647186e-01 1.61753997e-01 5.35889745e-01 1.28500307e+00
-2.04853863e-01 -8.22519302e-01 -6.19985640e-01 -3.05896401e-01
-3.10034424e-01 3.28605622e-01 -3.15472424e-01 1.10747278e+00
-2.52072718e-02 6.06562436e-01 -6.59672692e-02 -3.96847188e-01
6.55393541e-01 -1.49146706e-01 8.62702429e-01 -5.51910996e-01
-4.44526553e-01 1.79881826e-01 7.59977047e-05 -1.04618227e+00
-4.25206959e-01 -8.44264865e-01 -1.65137529e+00 -4.88875538e-01
-1.09710038e-01 -3.25016975e-01 2.40386665e-01 9.39003348e-01
4.60065961e-01 4.64349121e-01 1.12130150e-01 -1.76822698e+00
-6.49599731e-01 -5.80059409e-01 -1.40645921e-01 5.91887057e-01
4.34592426e-01 -1.03780365e+00 -2.17084408e-01 4.67222691e-01] | [7.108145713806152, -0.5382227301597595] |
d70a963c-cc48-4636-ac16-5f9626bcdb92 | named-entity-and-relation-extraction-with | 2212.01612 | null | https://arxiv.org/abs/2212.01612v1 | https://arxiv.org/pdf/2212.01612v1.pdf | Named Entity and Relation Extraction with Multi-Modal Retrieval | Multi-modal named entity recognition (NER) and relation extraction (RE) aim to leverage relevant image information to improve the performance of NER and RE. Most existing efforts largely focused on directly extracting potentially useful information from images (such as pixel-level features, identified objects, and associated captions). However, such extraction processes may not be knowledge aware, resulting in information that may not be highly relevant. In this paper, we propose a novel Multi-modal Retrieval based framework (MoRe). MoRe contains a text retrieval module and an image-based retrieval module, which retrieve related knowledge of the input text and image in the knowledge corpus respectively. Next, the retrieval results are sent to the textual and visual models respectively for predictions. Finally, a Mixture of Experts (MoE) module combines the predictions from the two models to make the final decision. Our experiments show that both our textual model and visual model can achieve state-of-the-art performance on four multi-modal NER datasets and one multi-modal RE dataset. With MoE, the model performance can be further improved and our analysis demonstrates the benefits of integrating both textual and visual cues for such tasks. | ['Wei Lu', 'Kewei Tu', 'Pengjun Xie', 'Yong Jiang', 'Jiong Cai', 'Xinyu Wang'] | 2022-12-03 | null | null | null | null | ['multi-modal-named-entity-recognition'] | ['natural-language-processing'] | [ 1.73987731e-01 -7.48479292e-02 -7.70085454e-02 -2.03727946e-01
-1.43112957e+00 -6.50232971e-01 7.46189058e-01 1.37148172e-01
-5.93821943e-01 5.74710011e-01 2.66195029e-01 2.83310339e-02
2.25690622e-02 -7.11829603e-01 -6.76429212e-01 -5.15431166e-01
4.93748039e-01 2.30076835e-01 5.89335740e-01 -3.05739902e-02
3.32658172e-01 5.36797881e-01 -1.62720609e+00 6.05503559e-01
6.84520125e-01 1.17140937e+00 5.10300815e-01 6.74891412e-01
-3.70723188e-01 9.61950183e-01 -2.85394311e-01 -4.77182746e-01
-7.92746618e-02 3.34220529e-02 -8.68447542e-01 -2.76459605e-02
3.82516354e-01 -3.15837353e-01 -3.87885153e-01 8.77904713e-01
6.06253386e-01 3.75351042e-01 1.03222346e+00 -1.04482436e+00
-8.15034628e-01 1.67507693e-01 -6.51336014e-01 1.70466498e-01
2.72216469e-01 -8.70035496e-03 8.94686282e-01 -1.25872529e+00
7.84480035e-01 1.10393262e+00 1.60237387e-01 3.19961250e-01
-6.33800924e-01 -5.09998083e-01 7.80989602e-02 3.14329475e-01
-1.42821968e+00 -6.18894398e-01 7.81609178e-01 -4.12360549e-01
8.08180511e-01 1.21960230e-01 1.93163857e-01 7.84223258e-01
6.71491784e-04 1.13925588e+00 1.06090915e+00 -5.52740693e-01
-1.57343596e-01 4.77165490e-01 6.90688100e-03 7.23834991e-01
-1.02609739e-01 -1.53546743e-02 -6.72479033e-01 8.09393525e-02
5.90478420e-01 -2.09971648e-02 -2.09201828e-01 -2.06225947e-01
-1.23454261e+00 4.47797686e-01 6.79632485e-01 4.30696785e-01
-7.71840513e-01 6.50679395e-02 2.02751309e-01 -2.02096671e-01
3.44316572e-01 2.37377658e-01 -2.36920565e-01 3.05515349e-01
-1.10521889e+00 -3.18200290e-01 4.92786169e-01 8.81801128e-01
9.98006344e-01 -3.71625870e-01 -5.57321489e-01 1.05567527e+00
4.66872841e-01 6.96567357e-01 3.49877924e-01 -6.59250557e-01
6.01321936e-01 8.38773310e-01 2.54484892e-01 -1.14256501e+00
-2.88933575e-01 -1.38344526e-01 -6.67498469e-01 -1.67895541e-01
2.38550585e-02 -4.56622243e-02 -1.23952651e+00 1.27597666e+00
4.10150170e-01 1.09462969e-01 3.36225599e-01 1.07835412e+00
1.27090645e+00 8.88933599e-01 6.14136219e-01 -2.04488039e-02
1.64064932e+00 -1.19259512e+00 -6.55290782e-01 -5.32160878e-01
4.11204815e-01 -1.15359974e+00 7.50670195e-01 -8.70053247e-02
-8.64820957e-01 -4.78184819e-01 -8.11171472e-01 -2.50488967e-01
-7.05426633e-01 6.50644183e-01 2.82618076e-01 1.15399085e-01
-9.57320631e-01 5.19644618e-02 -4.84920293e-01 -6.67506397e-01
4.34538007e-01 8.91384333e-02 -5.29840291e-01 -4.08267587e-01
-1.04532528e+00 9.76372242e-01 6.33205116e-01 2.11317286e-01
-9.74648893e-01 -2.78519988e-01 -7.92831302e-01 -2.01652222e-03
5.54783583e-01 -8.06353629e-01 9.12905395e-01 -7.68981874e-01
-1.07542872e+00 9.32160735e-01 -3.33372116e-01 -5.97107150e-02
2.26648659e-01 -2.65671402e-01 -4.27906632e-01 6.29228473e-01
2.50599205e-01 1.05006349e+00 7.53003955e-01 -1.67619884e+00
-8.53075206e-01 -3.79812598e-01 1.16283089e-01 5.59786201e-01
-4.00156349e-01 2.37974063e-01 -1.05912542e+00 -5.60558736e-01
-1.03988811e-01 -8.01211536e-01 -3.41099910e-02 -4.07032371e-02
-6.05014741e-01 -2.34605014e-01 7.94289827e-01 -8.00556660e-01
1.05163109e+00 -2.08187079e+00 2.37024408e-02 3.74472700e-02
-2.06858423e-02 2.06741974e-01 -4.16637510e-01 5.04099727e-01
1.58096179e-01 2.48493776e-01 1.92859262e-01 -1.54557809e-01
-2.07010105e-01 -3.80530804e-02 -1.87182248e-01 3.62102352e-02
3.55127424e-01 1.30338895e+00 -6.95405662e-01 -1.13751960e+00
5.04572034e-01 6.09224856e-01 2.22124040e-01 3.73572111e-01
-2.11091667e-01 2.21426025e-01 -7.73856461e-01 8.98635447e-01
3.93651366e-01 -5.22388995e-01 -1.33380666e-01 -8.02192152e-01
-7.94620514e-02 -2.71331400e-01 -1.11655557e+00 1.49188721e+00
-5.93382478e-01 6.00977004e-01 -2.06994101e-01 -8.09391320e-01
7.02638745e-01 4.27271724e-01 3.15229028e-01 -8.62085700e-01
1.51157543e-01 -6.27793297e-02 -6.38604820e-01 -8.29950213e-01
7.54920721e-01 -1.60899945e-02 1.03356376e-01 3.46515208e-01
4.15896103e-02 2.61672646e-01 1.77512407e-01 3.64485443e-01
8.91654253e-01 3.31499577e-01 2.33480960e-01 3.15711915e-01
6.06813729e-01 3.27452958e-01 2.06944227e-01 7.84302711e-01
-1.46795243e-01 6.06909275e-01 1.12994611e-01 -3.90597470e-02
-9.59100783e-01 -7.93921888e-01 8.09346959e-02 1.18121064e+00
4.29569244e-01 -2.90086240e-01 -3.89861405e-01 -8.44626188e-01
-4.32923526e-01 5.71646452e-01 -6.07720196e-01 5.14171049e-02
-2.09725544e-01 -6.01759017e-01 4.90438700e-01 6.54014051e-01
7.87284911e-01 -1.25983310e+00 -3.94847423e-01 -9.62346271e-02
-5.15610754e-01 -1.46547890e+00 -3.54484528e-01 -5.38741760e-02
-7.42000818e-01 -1.05883920e+00 -1.02425838e+00 -9.78757203e-01
8.21412683e-01 4.70023990e-01 8.63396168e-01 -2.47920733e-02
-3.02146554e-01 7.90947378e-01 -6.48372948e-01 -2.68917769e-01
-9.60369855e-02 -2.15218607e-02 -4.22940910e-01 1.49254858e-01
3.33295673e-01 -1.06561929e-01 -8.06716025e-01 3.61561179e-01
-1.11985290e+00 3.18780273e-01 1.25428176e+00 7.13928401e-01
8.91070485e-01 -3.77845690e-02 3.39019120e-01 -6.24935985e-01
2.57138938e-01 -4.04904574e-01 -1.75927565e-01 8.20304871e-01
-3.73149157e-01 2.11682275e-01 2.64513671e-01 -4.18756038e-01
-1.48820007e+00 3.97301048e-01 1.18883096e-01 -4.81153727e-01
-3.86487484e-01 7.79945493e-01 -2.08295599e-01 7.90406540e-02
5.58497727e-01 4.42997038e-01 -6.29904151e-01 -4.60517555e-01
5.33654094e-01 9.72341895e-01 5.59630036e-01 -5.18989027e-01
7.55243003e-01 4.24924940e-01 -1.64270580e-01 -7.57651985e-01
-1.22530174e+00 -8.82492483e-01 -7.16536701e-01 -4.67884779e-01
1.17862737e+00 -1.20476818e+00 -2.18638375e-01 1.63224608e-01
-1.24461412e+00 1.84288099e-01 1.06764711e-01 4.15707529e-01
-3.46522242e-01 3.35578471e-01 -4.52371508e-01 -9.05823410e-01
-5.96592188e-01 -1.00552380e+00 1.42200410e+00 6.08712256e-01
4.12128180e-01 -6.62127078e-01 -1.61685452e-01 8.00342500e-01
2.45055705e-01 2.24478869e-03 9.23424959e-01 -7.28343844e-01
-8.39831054e-01 -3.20501238e-01 -8.31015527e-01 7.99444020e-02
-2.15405971e-01 -1.70903623e-01 -1.20882165e+00 1.77523538e-01
-5.89888215e-01 -4.26540077e-01 9.80875552e-01 1.33290172e-01
7.44206727e-01 -2.86981147e-02 -6.92937613e-01 1.44917145e-01
1.77590835e+00 1.93216130e-01 7.81205952e-01 4.62489069e-01
8.78649473e-01 7.47315288e-01 1.04511416e+00 2.32737496e-01
5.40892243e-01 6.61375165e-01 3.42785597e-01 -2.32118085e-01
-2.97599643e-01 -2.98552841e-01 4.04254466e-01 6.51729822e-01
-4.29816358e-02 -4.62847829e-01 -1.02231574e+00 8.31027865e-01
-2.10333729e+00 -8.86989713e-01 9.69687253e-02 1.73460031e+00
7.11345613e-01 -2.92297244e-01 -1.74696892e-01 -3.45673025e-01
9.64080691e-01 2.23327056e-02 -6.11453176e-01 8.54656175e-02
-2.01959282e-01 -1.49674490e-01 4.52827692e-01 3.52244452e-02
-1.35809207e+00 1.20524240e+00 5.48521185e+00 1.08396816e+00
-8.05038571e-01 9.39644650e-02 5.77374518e-01 2.15609878e-01
-1.97027206e-01 6.42524362e-02 -9.24843431e-01 -6.48058727e-02
8.31659138e-01 -7.28750660e-04 7.72398412e-02 7.09993958e-01
5.41078448e-02 -3.78733307e-01 -8.40561211e-01 1.19928086e+00
3.93261671e-01 -1.29358149e+00 4.41244215e-01 -8.74306038e-02
6.48521721e-01 6.89937174e-02 -1.89584240e-01 2.28899524e-01
1.11192830e-01 -9.22957063e-01 5.86981416e-01 1.14402902e+00
7.57700920e-01 -8.39842975e-01 9.13425386e-01 3.87893736e-01
-1.43862796e+00 6.11448847e-02 -2.92681932e-01 6.21937037e-01
2.09494025e-01 4.58015919e-01 -7.35833526e-01 8.94139946e-01
7.18161225e-01 6.84741437e-01 -9.10534501e-01 1.15462434e+00
-3.09477597e-01 1.76916838e-01 -2.89419264e-01 -5.28084785e-02
8.35042596e-02 2.29775116e-01 3.85912478e-01 1.27043355e+00
1.74901545e-01 3.40173930e-01 1.86182633e-01 6.64777696e-01
-2.34147117e-01 5.45003057e-01 -5.01310408e-01 -2.53265649e-01
4.08403635e-01 1.74555123e+00 -8.83212447e-01 -3.76207620e-01
-5.85738242e-01 1.13464773e+00 4.79274392e-01 5.55684865e-01
-6.38831854e-01 -5.36064386e-01 -9.96505916e-02 -1.48147807e-01
5.99172294e-01 -7.96183869e-02 1.00604529e-02 -1.28415847e+00
2.02710237e-02 -4.85725611e-01 5.30638039e-01 -1.49262202e+00
-1.34490120e+00 7.42891848e-01 -8.48017260e-02 -1.24262142e+00
-7.24349245e-02 -5.78264654e-01 -2.14493930e-01 7.55775452e-01
-1.82513750e+00 -1.76183081e+00 -4.09492761e-01 6.21824205e-01
5.79199851e-01 3.58620025e-02 7.31168449e-01 3.73002768e-01
-7.27778494e-01 2.37424850e-01 3.89792994e-02 4.25875783e-01
9.97718096e-01 -9.98726487e-01 -3.83783489e-01 9.09667075e-01
3.79231602e-01 3.45857084e-01 1.25709012e-01 -7.22406685e-01
-1.45825517e+00 -1.26835394e+00 7.69730508e-01 -4.57292318e-01
4.12870318e-01 7.70062581e-02 -5.62911510e-01 5.26734352e-01
1.43575564e-01 -8.18456803e-03 6.48030937e-01 -1.74054742e-01
-2.31152371e-01 1.00783072e-01 -8.48451316e-01 6.11722469e-01
5.87589860e-01 -6.89402103e-01 -6.78214908e-01 1.54279977e-01
5.27318418e-01 -2.12979525e-01 -1.02912056e+00 3.85003626e-01
5.04130781e-01 -5.20553827e-01 1.15156245e+00 -4.32121664e-01
6.61644101e-01 -5.56968093e-01 -4.17818874e-01 -8.92935932e-01
1.92577131e-02 1.68046013e-01 -1.90032888e-02 1.66686428e+00
5.41702330e-01 1.22184820e-01 3.86022925e-01 6.39262438e-01
1.90305322e-01 -5.26470542e-01 -6.25294447e-01 -3.31848860e-01
-2.85204470e-01 -6.04353964e-01 1.37627110e-01 7.80348778e-01
-2.98726499e-01 8.50753188e-01 -4.26533580e-01 4.28054422e-01
3.80994380e-01 3.77473712e-01 6.18284941e-01 -1.09540391e+00
1.58421546e-01 -6.29455373e-02 -2.27302387e-01 -1.07447755e+00
1.10587038e-01 -9.55813289e-01 1.91445485e-01 -2.15003324e+00
7.76521087e-01 -2.22017318e-01 -4.84906614e-01 7.79521465e-01
-2.88368881e-01 5.75335860e-01 3.62309992e-01 5.60095847e-01
-1.22027719e+00 5.63410759e-01 1.24724638e+00 -2.40812629e-01
-8.14808831e-02 -5.45199454e-01 -6.78424776e-01 6.00885928e-01
3.91372353e-01 -5.10510743e-01 -7.73606822e-02 -4.22421873e-01
2.68970460e-01 2.37731770e-01 6.39386475e-01 -8.42831373e-01
6.34803116e-01 -1.49922311e-01 9.12847400e-01 -8.82007122e-01
3.38624865e-01 -8.63805473e-01 -7.31835142e-02 -2.68118441e-01
-2.98818320e-01 -2.57488489e-01 2.84810275e-01 7.92811871e-01
-3.87622744e-01 -2.34134421e-01 5.25442719e-01 -1.31657630e-01
-1.20986009e+00 3.80618840e-01 -1.31673008e-01 -4.95042801e-02
1.07978284e+00 -1.39171004e-01 -5.68802774e-01 -3.78684342e-01
-8.49464834e-01 4.20469701e-01 2.40433335e-01 6.04341388e-01
9.23615336e-01 -1.33622479e+00 -5.80946624e-01 -3.66383642e-01
6.10770345e-01 -9.83362198e-02 5.74563324e-01 9.19790506e-01
-2.25008637e-01 5.59377015e-01 3.95421498e-02 -4.38670903e-01
-1.31314301e+00 7.46890068e-01 2.19174936e-01 -2.70302266e-01
-3.69472414e-01 4.25811321e-01 2.69805133e-01 -1.75623864e-01
4.45720479e-02 2.76869118e-01 -7.80859530e-01 3.49616230e-01
5.60088634e-01 1.48376659e-01 -1.61239594e-01 -1.05184448e+00
-4.42369819e-01 9.39297318e-01 -1.60364568e-01 -2.92226553e-01
1.27347004e+00 -5.31136990e-01 1.87136768e-03 3.29778701e-01
1.11270130e+00 -1.84739903e-01 -8.83061171e-01 -5.21894336e-01
6.81419903e-03 -1.34582818e-01 3.35713863e-01 -1.16401875e+00
-1.09747767e+00 9.12133634e-01 6.75976396e-01 -2.62509108e-01
1.35497177e+00 4.11715239e-01 8.40418816e-01 5.91414869e-01
1.77719370e-01 -1.12698209e+00 2.10561678e-01 3.13244134e-01
7.09175408e-01 -1.50916350e+00 -1.91975441e-02 -4.78559464e-01
-8.90864491e-01 1.10434914e+00 6.34673655e-01 3.16197872e-01
5.61207652e-01 -1.15908206e-01 2.16270864e-01 -3.37661505e-01
-7.87311733e-01 -8.96804810e-01 9.49481249e-01 5.95112324e-01
2.64031708e-01 -2.96206921e-01 -1.30685210e-01 6.49906337e-01
4.13728952e-01 6.14523813e-02 1.25140160e-01 9.28573310e-01
-5.67825973e-01 -9.37990487e-01 -4.92335767e-01 4.84003425e-01
-4.93318021e-01 -3.18985194e-01 -3.58925521e-01 6.29695892e-01
-6.09843656e-02 1.14412010e+00 -1.64722621e-01 -3.51235241e-01
3.73705447e-01 1.81317776e-01 1.63813353e-01 -5.37665009e-01
-3.64629626e-01 2.52805263e-01 2.23400503e-01 -5.34567475e-01
-8.06519210e-01 -2.21701771e-01 -1.22184002e+00 2.87490785e-01
-6.32210970e-01 1.50966337e-02 7.25606203e-01 1.15271235e+00
4.75289762e-01 4.62674409e-01 3.44690263e-01 -6.38879657e-01
2.97762975e-02 -7.86057472e-01 -2.84486800e-01 4.78356451e-01
5.64526906e-03 -5.29307902e-01 -1.23386666e-01 3.72441590e-01] | [10.795315742492676, 1.4694178104400635] |
69cca4a6-04bc-4b28-8792-2bffcf9fd6bb | sketching-image-gist-human-mimetic | 2007.08760 | null | https://arxiv.org/abs/2007.08760v1 | https://arxiv.org/pdf/2007.08760v1.pdf | Sketching Image Gist: Human-Mimetic Hierarchical Scene Graph Generation | Scene graph aims to faithfully reveal humans' perception of image content. When humans analyze a scene, they usually prefer to describe image gist first, namely major objects and key relations in a scene graph. This humans' inherent perceptive habit implies that there exists a hierarchical structure about humans' preference during the scene parsing procedure. Therefore, we argue that a desirable scene graph should be also hierarchically constructed, and introduce a new scheme for modeling scene graph. Concretely, a scene is represented by a human-mimetic Hierarchical Entity Tree (HET) consisting of a series of image regions. To generate a scene graph based on HET, we parse HET with a Hybrid Long Short-Term Memory (Hybrid-LSTM) which specifically encodes hierarchy and siblings context to capture the structured information embedded in HET. To further prioritize key relations in the scene graph, we devise a Relation Ranking Module (RRM) to dynamically adjust their rankings by learning to capture humans' subjective perceptive habits from objective entity saliency and size. Experiments indicate that our method not only achieves state-of-the-art performances for scene graph generation, but also is expert in mining image-specific relations which play a great role in serving downstream tasks. | ['Xilin Chen', 'Wenbin Wang', 'Ruiping Wang', 'Shiguang Shan'] | 2020-07-17 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/1834_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123580222.pdf | eccv-2020-8 | ['scene-parsing'] | ['computer-vision'] | [ 5.56350768e-01 4.85265464e-01 -8.27980638e-02 -6.55124784e-01
-1.27846420e-01 -4.90900576e-02 4.92068201e-01 3.91095221e-01
-2.02429146e-01 1.83665663e-01 5.27186036e-01 -2.71919161e-01
-1.64167471e-02 -1.07786059e+00 -8.13635647e-01 -3.49846661e-01
6.48180246e-02 1.48777172e-01 5.79525948e-01 -3.59059989e-01
6.08051181e-01 1.30689695e-01 -1.71455419e+00 6.74745619e-01
9.03978586e-01 9.15673494e-01 1.06872571e+00 4.41180438e-01
-3.18100154e-01 1.50682259e+00 -3.95810664e-01 -5.71318328e-01
-1.87653393e-01 -6.25208437e-01 -1.22671473e+00 5.35631061e-01
2.23991498e-01 -9.59503874e-02 -2.52489358e-01 1.50302708e+00
2.73636758e-01 2.15215340e-01 3.75927955e-01 -1.12857926e+00
-9.93733823e-01 1.01133525e+00 -5.67059159e-01 3.92430246e-01
5.62365294e-01 4.28958498e-02 1.24554384e+00 -5.86268663e-01
7.78039873e-01 1.48080838e+00 1.42853901e-01 3.33812237e-01
-8.82998228e-01 -1.71335936e-01 5.92359841e-01 5.74136615e-01
-1.20591068e+00 -2.29923770e-01 1.12848127e+00 -3.16048443e-01
8.95109236e-01 5.62451422e-01 7.94742942e-01 8.10824633e-01
4.50200558e-01 8.93698454e-01 9.97538626e-01 -3.74839872e-01
1.13625050e-01 2.49437943e-01 1.04057677e-01 9.25817847e-01
-3.14889699e-02 -3.88621897e-01 -6.55316055e-01 3.90609741e-01
9.05029476e-01 -1.00899987e-01 -9.13504288e-02 -5.16996264e-01
-1.37992454e+00 6.85391068e-01 1.01298749e+00 2.90860772e-01
-5.90172112e-01 -1.03859112e-01 2.67130256e-01 -1.22659847e-01
1.57705024e-01 6.60968065e-01 -2.16050491e-01 4.54951495e-01
-3.91886592e-01 9.31442007e-02 3.93232733e-01 1.13895035e+00
8.65311742e-01 -1.40019044e-01 -4.89658028e-01 8.63026619e-01
3.82093668e-01 8.22903439e-02 6.62178397e-01 -9.06890631e-01
4.41071928e-01 1.21941268e+00 -2.96241045e-01 -1.61080956e+00
-3.90403867e-01 -4.12215948e-01 -1.04220474e+00 -3.84378791e-01
-2.96852648e-01 5.17928600e-01 -1.07829130e+00 1.75816762e+00
2.72673190e-01 -7.90904239e-02 4.74806726e-02 1.10279679e+00
1.21520948e+00 6.71590447e-01 6.20699823e-01 -2.39055067e-01
1.78235912e+00 -9.21725869e-01 -6.22392774e-01 -6.58248484e-01
2.61555463e-01 -3.86569023e-01 1.52815962e+00 2.58638132e-02
-1.01184571e+00 -8.50302279e-01 -8.99398744e-01 -4.70618546e-01
-5.54417133e-01 -1.51538253e-01 7.94850349e-01 4.88590710e-02
-1.17423546e+00 1.71383828e-01 -2.18720362e-01 -6.61869645e-01
4.10783917e-01 1.30574971e-01 -5.69131225e-02 2.14340523e-01
-1.48204672e+00 7.89292336e-01 8.53295922e-01 2.89101422e-01
-7.95325637e-01 -1.87295735e-01 -1.31449938e+00 1.31175041e-01
5.57161689e-01 -9.98436809e-01 1.04035795e+00 -1.05843163e+00
-9.10860121e-01 1.26707339e+00 -3.08605641e-01 -5.52373469e-01
-9.06084031e-02 9.20115635e-02 -3.96795660e-01 3.32893968e-01
4.10255730e-01 1.05079556e+00 8.82293522e-01 -1.61795402e+00
-7.77956426e-01 -2.41851628e-01 4.49709535e-01 7.66537786e-01
-1.81376010e-01 2.66070247e-01 -6.37348890e-01 -5.27776062e-01
5.43969989e-01 -5.09049475e-01 -6.82433903e-01 -6.00988209e-01
-8.14632416e-01 -3.89025658e-01 4.78939831e-01 -5.07212043e-01
1.52364075e+00 -2.03900766e+00 2.34422013e-02 8.46616253e-02
5.15890360e-01 -1.78717449e-01 -2.60909367e-03 1.59720197e-01
-1.61938384e-01 2.00959027e-01 -1.96875930e-01 1.23054296e-01
-1.59354270e-01 1.77608892e-01 -4.02643621e-01 -2.99284998e-02
5.60572371e-02 1.26175177e+00 -1.11454642e+00 -1.04085207e+00
3.76027465e-01 -1.82607323e-02 -5.15399218e-01 3.77178043e-01
-4.91914868e-01 1.87236428e-01 -6.99679554e-01 4.26325858e-01
3.74924809e-01 -8.38610888e-01 2.35589042e-01 -5.33509374e-01
-5.17035201e-02 3.53075534e-01 -8.33004475e-01 1.65665722e+00
-1.13056123e-01 3.71038705e-01 -4.56679583e-01 -8.94101083e-01
8.25062990e-01 -1.23472646e-01 2.25152016e-01 -1.05252182e+00
2.09854841e-01 -3.85808617e-01 -2.67063566e-02 -8.31511319e-01
7.07547307e-01 1.96591523e-02 -2.87717074e-01 3.64083424e-02
-1.63700297e-01 -1.99163228e-01 1.26285657e-01 5.74212253e-01
7.99784064e-01 5.55602415e-03 5.13390183e-01 -5.10012865e-01
6.75994515e-01 1.79123297e-01 4.26102906e-01 7.08822012e-01
-2.32517347e-01 5.14968693e-01 6.65361106e-01 -6.40819728e-01
-8.43665957e-01 -1.07326150e+00 2.81726480e-01 1.41086197e+00
8.94060552e-01 -5.37594438e-01 -8.80972385e-01 -4.62952018e-01
-5.62974811e-01 9.49448824e-01 -7.03418553e-01 -3.50867182e-01
-5.61768293e-01 -6.07696176e-01 -1.21503636e-01 3.38822752e-01
9.25581098e-01 -1.78716493e+00 -1.03621888e+00 1.22859024e-01
-4.13543433e-01 -1.26399827e+00 -5.10002315e-01 4.89051454e-02
-5.70172787e-01 -9.78771210e-01 -1.27622873e-01 -1.26171207e+00
9.87058461e-01 5.71164370e-01 1.44895041e+00 5.33795543e-02
-1.47181954e-02 1.18138164e-01 -4.09787267e-01 -2.12021068e-01
-2.51471907e-01 3.23645249e-02 -2.44226277e-01 8.53514969e-02
3.35882753e-01 -3.56178820e-01 -7.12959826e-01 1.65906578e-01
-9.23600554e-01 9.37258065e-01 6.54962361e-01 3.84678423e-01
7.62440979e-01 5.58847785e-01 6.88652769e-02 -1.20695114e+00
6.63567007e-01 -2.36110255e-01 -3.35914969e-01 5.29233932e-01
-4.44381833e-01 1.65110782e-01 5.20773470e-01 -3.93249318e-02
-1.45780170e+00 1.89468935e-01 8.46315995e-02 -3.09953392e-01
-2.37889573e-01 5.66946566e-01 -4.66931641e-01 3.19891840e-01
5.76060653e-01 4.63179320e-01 -4.96072769e-01 8.35433453e-02
6.44244254e-01 3.94790024e-01 7.05474854e-01 -4.86672044e-01
5.77618778e-01 3.05755168e-01 -9.38136056e-02 -6.75894320e-01
-1.44749963e+00 -4.01311755e-01 -6.46712363e-01 -3.81383598e-01
1.26359308e+00 -9.48117793e-01 -7.45257676e-01 2.61677027e-01
-1.24185443e+00 -3.03137898e-01 -2.02512041e-01 3.97654213e-02
-5.18722236e-01 1.39288515e-01 -6.48104668e-01 -5.93130410e-01
-1.56075045e-01 -1.04174471e+00 1.19583488e+00 4.73388314e-01
-1.86025143e-01 -7.84993649e-01 -3.96423727e-01 3.86469007e-01
4.94559146e-02 2.69986004e-01 1.05336893e+00 -1.59553677e-01
-9.02860701e-01 3.97996426e-01 -7.85341322e-01 -1.20455749e-01
1.21834382e-01 -2.47988865e-01 -9.06759143e-01 1.72145665e-01
1.72681585e-01 -1.72493607e-01 8.06947410e-01 4.37558442e-01
1.51304555e+00 -6.45123839e-01 -2.80576080e-01 5.13192475e-01
1.43340302e+00 3.64505619e-01 8.12052608e-01 4.99935418e-01
1.26986825e+00 1.07778275e+00 6.05095387e-01 1.79023415e-01
1.04563856e+00 2.74202108e-01 4.10155952e-01 -3.83611143e-01
-9.76174474e-02 -8.72865498e-01 1.53154239e-01 8.87146890e-01
1.58591971e-01 -2.09011093e-01 -7.69910693e-01 4.57423627e-01
-1.87023556e+00 -1.00520563e+00 -3.81978631e-01 1.67302060e+00
9.49133813e-01 4.40864176e-01 -8.86906013e-02 -1.80020750e-01
9.92927194e-01 5.49100339e-01 -5.01659572e-01 -4.01978046e-01
-2.89786339e-01 -3.58578742e-01 3.93002719e-01 4.34770584e-01
-1.17286456e+00 1.65183544e+00 5.89892769e+00 7.54433095e-01
-1.01564229e+00 -1.74338534e-01 1.04642081e+00 6.31272674e-01
-5.86937547e-01 9.05363262e-02 -5.56596220e-01 3.14095855e-01
6.00259602e-01 -2.78629273e-01 3.18406999e-01 9.74841595e-01
2.15422407e-01 -3.31150591e-01 -8.06641638e-01 1.17628741e+00
2.06874147e-01 -1.26810050e+00 7.00063229e-01 -5.48356622e-02
6.00420117e-01 -4.55715507e-01 4.15320545e-02 2.65853703e-01
3.02073926e-01 -9.30838346e-01 8.94262791e-01 4.33558166e-01
6.22407198e-01 -4.50288534e-01 3.68331045e-01 3.24136287e-01
-1.50650847e+00 4.88332026e-02 -6.34449482e-01 -1.65181369e-01
8.50038901e-02 4.81181681e-01 -7.80330777e-01 3.93305421e-01
7.83405125e-01 7.24405646e-01 -1.11488795e+00 7.33264506e-01
-5.30360699e-01 2.12639824e-01 2.79064417e-01 -1.38832256e-01
2.59334177e-01 -1.59061596e-01 3.23290616e-01 1.08946276e+00
-6.04905151e-02 4.98031139e-01 1.33058786e-01 1.04036331e+00
-3.54585759e-02 2.34902352e-01 -7.09554017e-01 7.48862401e-02
2.92606235e-01 1.32867134e+00 -1.24274802e+00 -5.76970994e-01
-1.88084140e-01 1.17602956e+00 3.77601892e-01 4.00027364e-01
-9.07610714e-01 -3.68511640e-02 9.97309238e-02 2.02427790e-01
3.49749289e-02 1.33808171e-02 -4.51636910e-01 -1.10280764e+00
-8.08742642e-02 -7.37326086e-01 6.07815325e-01 -1.40283179e+00
-1.03942132e+00 1.00231493e+00 9.24606472e-02 -1.10372603e+00
-3.09486389e-02 -3.35978150e-01 -4.29688662e-01 5.25838494e-01
-1.30072963e+00 -1.23069358e+00 -5.22529483e-01 6.20759666e-01
1.00521576e+00 2.81988978e-01 4.54928458e-01 -1.27306089e-01
-5.86100161e-01 -4.26859073e-02 -1.04469645e+00 -7.82347750e-03
2.01892719e-01 -1.43083560e+00 5.96370459e-01 1.06533432e+00
3.92313451e-01 7.88002133e-01 9.65525627e-01 -8.19888353e-01
-1.12633419e+00 -1.13855183e+00 1.13992023e+00 -3.95097196e-01
4.90812778e-01 -4.80926961e-01 -8.53078008e-01 5.86008430e-01
4.99022692e-01 -3.14659387e-01 3.08556229e-01 -6.08555786e-02
-1.02916628e-01 -8.35835263e-02 -7.70322680e-01 1.02377832e+00
1.51717031e+00 -6.72289670e-01 -1.02708745e+00 3.02592605e-01
1.18123782e+00 -4.07148600e-01 -3.78861725e-01 6.37923419e-01
1.31235644e-01 -1.18279362e+00 9.82660472e-01 -4.37185258e-01
6.53602898e-01 -5.58964491e-01 -1.33292034e-01 -9.78697002e-01
-8.73347342e-01 -4.01792347e-01 5.90331964e-02 1.19277644e+00
2.73406237e-01 -2.68449813e-01 7.72711396e-01 6.77052557e-01
-1.43395141e-01 -6.62841618e-01 -2.37822622e-01 -1.91279516e-01
-7.67275572e-01 -3.96965981e-01 6.29866242e-01 1.04941881e+00
-3.51534784e-02 9.60618973e-01 -2.48310953e-01 3.85106772e-01
5.78185976e-01 4.46583390e-01 4.40280944e-01 -9.81909156e-01
2.67759413e-02 -4.25399542e-01 -2.87351102e-01 -1.14434493e+00
1.57124355e-01 -7.14742780e-01 3.02478939e-01 -1.92342687e+00
5.62614501e-01 -1.30828738e-01 -6.50364399e-01 2.87405044e-01
-4.89832014e-01 2.17665043e-02 1.78094342e-01 1.89761415e-01
-1.23154104e+00 5.29548824e-01 1.62322295e+00 -1.44030109e-01
-2.64783084e-01 -2.96369553e-01 -1.27301264e+00 1.01225019e+00
6.88829660e-01 -7.87900090e-02 -8.19895267e-01 -4.09842610e-01
5.32554805e-01 3.46457027e-02 4.47243989e-01 -9.02348459e-01
4.01745439e-01 -5.42574108e-01 5.09514391e-01 -6.61947668e-01
1.96403475e-03 -5.85360646e-01 7.41927475e-02 2.60894924e-01
-5.41251242e-01 1.80298895e-01 -2.26213366e-01 3.17181706e-01
-3.47175092e-01 1.67156473e-01 6.09895647e-01 -5.30843854e-01
-1.54698002e+00 3.39864433e-01 -1.74869493e-01 -8.25355053e-02
8.59436572e-01 -3.62922460e-01 -3.27635199e-01 -3.46892148e-01
-5.90355158e-01 3.25634122e-01 3.67517322e-01 5.50926507e-01
1.02550113e+00 -1.20672870e+00 -3.43908489e-01 1.94252506e-01
3.93092364e-01 1.35285467e-01 4.62081343e-01 3.58432740e-01
-5.89976370e-01 3.18488270e-01 -2.37542242e-01 -5.93985319e-01
-1.12929034e+00 1.04703856e+00 9.31994691e-02 -1.18312776e-01
-6.47241533e-01 1.11698830e+00 9.97876883e-01 -3.49861570e-02
1.61796231e-02 -6.03983283e-01 -7.81893849e-01 -6.47117151e-03
2.86609083e-01 -1.31097227e-01 -4.46354955e-01 -1.00672078e+00
-2.59510666e-01 4.58230525e-01 3.31806690e-02 5.02757430e-02
1.00910997e+00 -6.09476924e-01 -5.44753730e-01 5.72956264e-01
8.93714726e-01 -1.76157638e-01 -1.08635461e+00 -8.97642747e-02
2.55858690e-01 -4.68105882e-01 -1.33259818e-01 -5.10631382e-01
-8.71611357e-01 5.75768590e-01 3.15537930e-01 6.58917308e-01
1.44216108e+00 5.47174990e-01 7.00848162e-01 2.65451223e-01
4.46612805e-01 -8.60161483e-01 3.21091354e-01 3.43978524e-01
1.10810328e+00 -1.17798269e+00 1.42963892e-02 -8.11614215e-01
-1.18723559e+00 7.14670539e-01 1.01422870e+00 -1.06785484e-01
3.20039332e-01 -3.26338977e-01 -3.85538973e-02 -6.95106924e-01
-7.34790027e-01 -7.20561862e-01 6.38514817e-01 3.57610613e-01
3.25226635e-01 1.69174343e-01 -1.13891229e-01 5.02998471e-01
-6.17486119e-01 -5.27981639e-01 2.49543652e-01 5.73941469e-01
-8.47408354e-01 -6.70883834e-01 -8.12414065e-02 5.38688421e-01
-1.95569307e-01 -3.77471626e-01 -5.21515131e-01 3.87772590e-01
3.73980731e-01 9.66161788e-01 6.09990694e-02 -5.99698961e-01
3.72565508e-01 -3.94843727e-01 3.13250095e-01 -9.21750069e-01
-2.73337364e-01 3.69456299e-02 -8.07835981e-02 -8.39833856e-01
-6.54888570e-01 -3.93822670e-01 -1.48875308e+00 1.27240971e-01
6.56140000e-02 1.02128461e-01 1.40653268e-01 9.18778777e-01
1.74236923e-01 8.84530485e-01 6.16429925e-01 -4.72540498e-01
3.58357847e-01 -7.03729510e-01 -4.60870087e-01 6.79752469e-01
5.06554218e-03 -4.84069228e-01 2.88690571e-02 5.54306328e-01] | [10.311270713806152, 1.4774476289749146] |
c59f98b9-d9b3-4274-9987-d36207dc1c69 | analysis-of-mesh-based-motion-compensation-in | 2302.01594 | null | https://arxiv.org/abs/2302.01594v1 | https://arxiv.org/pdf/2302.01594v1.pdf | Analysis of mesh-based motion compensation in wavelet lifting of dynamical 3-D+t CT data | Factorized in the lifting structure, the wavelet transform can easily be extended by arbitrary compensation methods. Thereby, the transform can be adapted to displacements in the signal without losing the ability of perfect reconstruction. This leads to an improvement of scalability. In temporal direction of dynamic medical 3-D+t volumes from Computed Tomography, displacement is mainly given by expansion and compression of tissue. We show that these smooth movements can be well compensated with a mesh-based method. We compare the properties of triangle and quadrilateral meshes. We also show that with a mesh-based compensation approach coding results are comparable to the common slice wise coding with JPEG 2000 while a scalable representation in temporal direction can be achieved. | ['André Kaup', 'Jürgen Seiler', 'Thomas Richter', 'Wolfgang Schnurrer'] | 2023-02-03 | null | null | null | null | ['motion-compensation'] | ['computer-vision'] | [ 5.41549444e-01 2.85826087e-01 -2.40462422e-02 -4.30365205e-02
-6.34814441e-01 -1.95860520e-01 2.94037730e-01 -3.86203676e-02
-3.06792587e-01 8.01616251e-01 4.49313164e-01 5.52272648e-02
-8.29913765e-02 -8.95917594e-01 -5.91563702e-01 -6.09771848e-01
-4.98298585e-01 3.32405090e-01 7.91355252e-01 -4.39104795e-01
3.86272967e-02 4.65603113e-01 -9.52733994e-01 6.93556011e-01
4.46514100e-01 1.05771804e+00 7.23195523e-02 9.32532012e-01
-1.15839429e-02 7.68222928e-01 -9.77552161e-02 -7.90706351e-02
4.27287132e-01 -4.09872383e-01 -1.00027466e+00 2.64088631e-01
3.64582747e-01 -5.55956304e-01 -1.74442098e-01 8.56303632e-01
5.86889982e-01 -1.20980613e-01 3.03260297e-01 -7.65222669e-01
-3.84664312e-02 3.61768574e-01 -7.90566206e-01 -2.39094393e-03
5.84930778e-01 -4.31223363e-01 4.88232523e-01 -1.01311195e+00
1.24677134e+00 1.17497671e+00 1.22887981e+00 3.47282052e-01
-1.57413292e+00 -1.22085132e-01 -4.27450269e-01 1.97255775e-01
-1.19451749e+00 -1.29520744e-01 8.97023618e-01 -2.75181621e-01
9.25534666e-01 7.55319297e-01 1.03482139e+00 3.00792873e-01
6.77286386e-01 2.65264928e-01 1.14328063e+00 -3.39866132e-01
3.59263085e-03 -3.43088150e-01 -2.76205152e-01 7.34111190e-01
4.36095381e-03 1.21406049e-01 -4.18852061e-01 -2.81309366e-01
1.21109641e+00 3.38834450e-02 -7.91610479e-01 -4.75479841e-01
-1.57322240e+00 5.75928986e-01 4.59191203e-01 7.47595310e-01
-4.86087441e-01 6.61129773e-01 8.54847848e-01 4.58978653e-01
6.03724480e-01 -2.67105624e-02 -3.76864135e-01 3.40113975e-02
-1.56029570e+00 2.25602105e-01 5.38978338e-01 8.82930458e-01
4.52436537e-01 3.89772832e-01 -7.61487409e-02 3.91630739e-01
1.00876532e-01 3.03445280e-01 6.77615404e-01 -1.48105681e+00
-3.57471704e-02 7.65924454e-02 -9.92094129e-02 -9.56166029e-01
-5.33390641e-01 -3.78287762e-01 -1.26668012e+00 4.00714368e-01
3.28090847e-01 2.48507112e-01 -7.06289172e-01 1.33869231e+00
5.02703547e-01 3.63537639e-01 -2.05294609e-01 5.88408470e-01
5.41308403e-01 3.88348758e-01 -3.59143764e-01 -7.14251578e-01
1.52323020e+00 -5.56703031e-01 -1.05790579e+00 4.18990195e-01
4.58802164e-01 -1.15259075e+00 5.41163027e-01 6.27903104e-01
-1.67427886e+00 -6.41033113e-01 -1.00264430e+00 -2.97001898e-01
2.45908618e-01 -5.24773657e-01 2.13931888e-01 4.82907236e-01
-1.31852865e+00 1.06570542e+00 -1.15610552e+00 2.25241128e-02
2.33185425e-01 3.25168103e-01 -4.48402941e-01 -6.53367788e-02
-1.02457416e+00 1.04980731e+00 8.27161819e-02 -2.81698167e-01
-3.14987510e-01 -1.12927294e+00 -5.46461225e-01 -7.48080760e-02
-1.37315214e-01 -7.65450537e-01 9.85343337e-01 -7.12862909e-01
-1.15347970e+00 8.21282566e-01 -2.50617772e-01 -6.96556926e-01
1.00695646e+00 2.90982157e-01 -1.44683540e-01 7.24612653e-01
-9.23617482e-02 4.74050522e-01 1.16305912e+00 -1.23301196e+00
-1.40994772e-01 -1.03537552e-01 -3.97145659e-01 -8.74583200e-02
-1.35952353e-01 -2.14797065e-01 -1.03606917e-01 -1.11076474e+00
6.84112906e-01 -6.92333639e-01 -4.65876788e-01 7.24099517e-01
6.18683808e-02 7.44379163e-01 8.11681211e-01 -1.16795754e+00
1.32673585e+00 -2.12915111e+00 2.73959339e-01 2.81640589e-01
3.94891143e-01 -3.01194221e-01 2.87582785e-01 3.85673285e-01
-3.15424591e-01 -2.73680110e-02 -4.22369480e-01 -2.01298892e-01
-4.23640966e-01 5.24942756e-01 -2.06332415e-01 7.91821957e-01
-2.33507603e-01 6.37562513e-01 -7.76877344e-01 -7.77511120e-01
1.60668328e-01 9.07313943e-01 -9.12559867e-01 -4.72148746e-01
4.60926950e-01 6.71336532e-01 -2.57548928e-01 4.67783540e-01
1.10789084e+00 -1.33548966e-02 2.60075569e-01 -6.33375645e-01
-1.72631234e-01 4.05528508e-02 -1.17462111e+00 1.78705692e+00
-5.76397359e-01 3.98266256e-01 6.45504534e-01 -1.04767728e+00
5.87141395e-01 9.52519059e-01 1.00776899e+00 -5.27764738e-01
-1.22522436e-01 6.42455697e-01 -5.88773265e-02 -2.60945588e-01
4.49367553e-01 -7.39095688e-01 3.06754410e-01 2.68379766e-02
-1.84451863e-01 -6.23274982e-01 -3.92611511e-02 -5.25629334e-02
1.08811677e+00 -4.61163744e-02 3.88310760e-01 -7.24645674e-01
6.46978855e-01 1.26026243e-01 4.77481663e-01 8.29741284e-02
1.10106729e-02 8.71409655e-01 1.96300194e-01 -6.65604174e-01
-1.41591358e+00 -9.51014042e-01 -7.02663541e-01 4.22694176e-01
-1.92145512e-01 -3.57047915e-01 -6.51994646e-01 -3.17584425e-02
-8.54034945e-02 2.72042543e-01 -6.37752712e-01 2.61427909e-01
-1.20013011e+00 -6.54036701e-01 3.33684564e-01 6.35364115e-01
5.48960388e-01 -4.62728709e-01 -1.09485722e+00 6.68841660e-01
-3.11265051e-01 -1.12587583e+00 -5.46160638e-01 8.31435025e-02
-1.77777028e+00 -8.64605248e-01 -1.08925176e+00 -8.26206267e-01
6.15237355e-01 1.66557938e-01 1.18955612e+00 5.05994141e-01
-2.43895411e-01 3.62436295e-01 -1.68216214e-01 9.31249931e-02
-5.83037317e-01 -5.08297682e-01 -1.50445431e-01 -2.04727724e-01
-6.50591552e-01 -1.07554352e+00 -8.98128986e-01 2.47020259e-01
-1.13713622e+00 3.02185655e-01 -9.85042080e-02 1.02260029e+00
9.05505300e-01 9.61163566e-02 -1.95698123e-02 -6.50867522e-01
2.99052328e-01 4.16786224e-02 -1.93643361e-01 -1.68675646e-01
-5.53941727e-01 6.16106531e-03 5.00232637e-01 -3.89174789e-01
-7.06500649e-01 1.77836195e-01 -5.48133492e-01 -4.67203140e-01
5.47270596e-01 2.95085669e-01 6.02104068e-01 -6.94063842e-01
5.65648139e-01 4.00719792e-01 1.31303936e-01 -4.04671490e-01
1.98737144e-01 -7.06212744e-02 5.21226287e-01 -3.98224711e-01
6.29767776e-01 1.05344117e+00 6.22072637e-01 -9.03241038e-01
1.90016463e-01 -2.63208359e-01 -7.58141875e-01 -3.69561553e-01
7.60664344e-01 -5.86752951e-01 -4.23744649e-01 1.31833687e-01
-1.26388896e+00 -1.87521383e-01 -9.93200719e-01 5.95403790e-01
-9.16839600e-01 7.47108817e-01 -1.11674094e+00 -3.92888010e-01
-3.61816436e-01 -1.09561646e+00 1.09617710e+00 -7.02445865e-01
-4.01783794e-01 -1.18850374e+00 8.84840861e-02 -1.44594282e-01
8.22882473e-01 7.71749377e-01 9.33319092e-01 1.56719685e-01
-4.64199960e-01 -4.95591834e-02 1.88625768e-01 1.70939043e-01
-8.69636536e-02 -2.09715694e-01 -5.77249646e-01 -4.48436916e-01
7.34350264e-01 1.30908579e-01 6.62718296e-01 7.11640477e-01
9.33320940e-01 -3.91945422e-01 -2.33053327e-01 8.92959535e-01
1.68497407e+00 -9.45058763e-02 8.09598744e-01 1.48946777e-01
1.84725061e-01 3.79991204e-01 1.30064070e-01 5.53374827e-01
-1.62682399e-01 8.74115527e-01 2.97335953e-01 -2.30470896e-01
-6.44036353e-01 9.25806984e-02 -5.06686652e-03 1.26253355e+00
-7.91430473e-01 4.85916674e-01 -7.27113128e-01 4.12547231e-01
-1.43999195e+00 -1.04135191e+00 -6.06820822e-01 2.03387260e+00
9.99282300e-01 4.96765114e-02 -7.36747161e-02 9.21955347e-01
4.19624209e-01 3.27390544e-02 -8.83884355e-02 -5.81087947e-01
1.18898265e-01 4.91652220e-01 8.86165738e-01 9.15473580e-01
-6.19219124e-01 1.13376737e-01 7.96281862e+00 9.64838266e-01
-1.14744282e+00 7.26727903e-01 3.39134812e-01 6.13439009e-02
-5.41592717e-01 -2.63397545e-01 -6.22505508e-02 2.92580038e-01
7.13154197e-01 -3.47146749e-01 1.79312423e-01 4.95775729e-01
3.21950912e-01 9.53715220e-02 -7.38473356e-01 7.31022775e-01
-3.17215413e-01 -1.51260459e+00 -2.30871409e-01 2.17940673e-01
5.62533736e-01 -2.79811949e-01 -1.68914512e-01 -2.73505270e-01
-3.64353865e-01 -7.92906046e-01 7.41755903e-01 3.80527318e-01
1.22049999e+00 -4.87370610e-01 4.29892898e-01 2.78220147e-01
-1.59315395e+00 3.72801691e-01 -4.41477865e-01 7.84544945e-02
6.08760834e-01 7.64846921e-01 -5.80327511e-01 7.33352542e-01
5.64912021e-01 5.45368254e-01 1.05138190e-01 9.54699159e-01
3.38483751e-01 3.27823073e-01 -3.89586866e-01 8.41331780e-01
3.28497924e-02 -2.06074640e-01 7.26560771e-01 1.16735184e+00
8.05253804e-01 2.44977996e-01 2.76619550e-02 2.19437912e-01
2.77241260e-01 2.22727448e-01 -4.39094335e-01 7.50683606e-01
-1.26270398e-01 8.37247670e-01 -9.52542722e-01 -5.07316709e-01
-4.94457752e-01 1.25652754e+00 -4.02133971e-01 7.00758174e-02
-7.30025470e-01 1.24307178e-01 1.66018248e-01 7.93695152e-01
5.04151165e-01 -4.08614486e-01 -3.98169607e-01 -1.05553806e+00
-1.17250130e-01 -6.39063299e-01 1.89018682e-01 -6.39580607e-01
-8.76691759e-01 8.10360193e-01 2.35547245e-01 -1.55515981e+00
-2.23616228e-01 -4.40696925e-01 -2.23047793e-01 6.73475623e-01
-1.43723786e+00 -8.72338831e-01 -1.17319956e-01 9.36078429e-01
3.91383737e-01 5.16834140e-01 9.73801315e-01 5.65744817e-01
6.88929379e-01 2.84553707e-01 1.73594028e-01 -2.13288918e-01
4.74875957e-01 -1.09934235e+00 1.16739228e-01 4.88589346e-01
-3.29256773e-01 1.15714557e-01 9.76101220e-01 -7.19038010e-01
-1.20244944e+00 -8.24660182e-01 9.21375394e-01 2.43362978e-01
5.27205229e-01 9.85708162e-02 -1.12745190e+00 6.88797057e-01
3.83308738e-01 5.45409143e-01 3.73136282e-01 -7.98355341e-01
-1.21033564e-02 -3.98187488e-02 -1.79965758e+00 1.12191327e-01
8.73374343e-01 -2.93094337e-01 -5.55919647e-01 5.58758616e-01
5.97590804e-01 -8.58157933e-01 -1.41603374e+00 4.00044650e-01
7.26980925e-01 -1.16406143e+00 1.51515543e+00 -1.39384046e-01
4.42121148e-01 -1.20585792e-01 -3.23787600e-01 -9.53939080e-01
-4.56495553e-01 -7.13463843e-01 -1.09206468e-01 5.28685927e-01
-1.53164938e-01 -6.58369064e-01 6.33128047e-01 3.10947001e-01
-2.05334648e-01 -6.74375892e-01 -1.75089645e+00 -8.00194681e-01
2.41276368e-01 -3.74263823e-01 2.57781267e-01 1.04021919e+00
2.52332777e-01 -5.01001179e-01 -5.62556088e-01 -3.86310399e-01
6.74205959e-01 -6.55342340e-02 -7.54205808e-02 -9.77689564e-01
-5.29481947e-01 -2.36387461e-01 -6.90376997e-01 -1.02055514e+00
-5.76737702e-01 -1.00842261e+00 -2.71092176e-01 -1.45557570e+00
-1.55263901e-01 -2.75951803e-01 -2.62052715e-02 1.27496123e-01
3.72431606e-01 8.35513830e-01 2.87970126e-01 3.76037061e-01
3.98382455e-01 1.78190976e-01 2.03563190e+00 9.25619826e-02
1.67537913e-01 -2.86538810e-01 1.00560166e-01 9.28933322e-01
4.79265511e-01 -5.37124157e-01 -2.50441968e-01 -3.45866352e-01
1.56670436e-01 9.02528644e-01 2.34639093e-01 -1.11660492e+00
-2.61136517e-02 2.17031643e-01 4.73957300e-01 -4.06188518e-01
5.62019169e-01 -1.24281526e+00 6.56290710e-01 1.26667535e+00
-2.23156556e-01 4.09282774e-01 2.19783887e-01 3.64822000e-01
-3.58542800e-01 -3.00621003e-01 1.07163322e+00 -4.00518954e-01
-8.08373466e-02 1.55556500e-01 -5.41221440e-01 -1.25004575e-01
8.51328909e-01 -5.18231034e-01 3.55227798e-01 -4.43272501e-01
-1.23486900e+00 -2.87213326e-01 5.93199193e-01 -6.22310579e-01
7.80039608e-01 -1.66169477e+00 -8.37709427e-01 3.29527497e-01
-7.87637413e-01 -1.24856152e-01 3.13770980e-01 1.42723846e+00
-1.22576880e+00 1.07597828e-01 -4.24893200e-01 -6.95712745e-01
-1.41467440e+00 5.70779443e-01 3.83330047e-01 -7.24314749e-01
-1.04350936e+00 5.28307378e-01 -1.31249636e-01 1.55385330e-01
-1.38782367e-01 -7.18083441e-01 1.17202401e-01 2.85938978e-02
5.96195936e-01 6.46032810e-01 2.92712033e-01 -7.25422442e-01
-3.70817959e-01 9.64500606e-01 5.54008782e-01 -3.16024005e-01
1.40995991e+00 -1.15956604e-01 -2.99025714e-01 3.15060392e-02
1.26391530e+00 4.29027855e-01 -9.70115304e-01 8.28466192e-02
-3.90604198e-01 -6.42757654e-01 2.18852416e-01 -2.75384635e-01
-1.47538018e+00 9.59697902e-01 5.84501743e-01 5.28199136e-01
1.52730346e+00 -4.55108643e-01 1.22241187e+00 -3.82892430e-01
6.71578467e-01 -6.65497184e-01 -2.24197552e-01 1.99319068e-02
1.35756087e+00 -5.19438922e-01 4.37449127e-01 -1.06582320e+00
-1.48314491e-01 1.52671731e+00 -3.74385297e-01 -5.30580580e-01
1.00157392e+00 9.14952338e-01 -3.27214748e-01 4.88899415e-03
-4.74147171e-01 2.27458000e-01 2.76138604e-01 5.33679962e-01
7.17070520e-01 -4.76878509e-02 -1.00782502e+00 -9.90080908e-02
-1.18163787e-01 2.56698221e-01 4.95638192e-01 1.15160000e+00
-4.33131069e-01 -1.30312026e+00 -8.25444639e-01 1.60165876e-01
-7.62249827e-01 -1.01462930e-01 5.11484206e-01 9.29300547e-01
1.10992491e-01 2.86084652e-01 -3.16067822e-02 9.44959447e-02
3.63824815e-01 1.69849232e-01 9.88370061e-01 -2.30625406e-01
-6.54748201e-01 7.01659501e-01 8.44613761e-02 -9.46851432e-01
-7.58553743e-01 -6.50310040e-01 -1.44522047e+00 -3.98191333e-01
1.20576791e-01 -7.94160664e-02 3.66522878e-01 3.17196757e-01
-3.88997644e-02 8.39850068e-01 3.73739779e-01 -1.19267559e+00
-4.11365390e-01 -6.00997388e-01 -7.42530882e-01 6.46094799e-01
4.08653527e-01 -5.07882774e-01 -2.52400070e-01 4.84831542e-01] | [11.559479713439941, -2.3497023582458496] |
03209bf0-b5bc-41c6-b0d4-bea6ca379f0b | from-heuristics-to-language-models-a-journey | null | null | https://ceur-ws.org/Vol-3320/paper6.pdf | https://ceur-ws.org/Vol-3320/paper6.pdf | From Heuristics to Language Models: A Journey Through the Universe of Semantic Table Interpretation with DAGOBAH | This paper presents DAGOBAH SL 2022, a semantic table interpretation system that has been continuously improved over the last four years when participating in the SemTab challenge. This year, we have improved the lookup coverage using external resources and we have integrated language models for better understanding the table headers. We have also implemented various system optimizations that lead to a reduction in execution time of about 30%. In this paper, we also show the relevance of using deep learning-based approaches for resolving certain ambiguities and we discuss the limitations of existing corpora and systems for maturing further this research field. | ['Jixiong Liu and Raphaël Troncy', 'Thomas Labbé', 'Yoan Chabot', 'Viet-Phi Huynh'] | 2022-10-25 | null | null | null | semtab-iswc-2022-10 | ['column-type-annotation', 'cell-entity-annotation'] | ['natural-language-processing', 'natural-language-processing'] | [ 4.07197289e-02 5.73639631e-01 -3.60644341e-01 -6.58397138e-01
-1.07624280e+00 -5.66357315e-01 5.68040073e-01 7.75718272e-01
-4.79813039e-01 1.06104589e+00 5.94883204e-01 -6.31357074e-01
-5.32375909e-02 -9.13982749e-01 -7.94406295e-01 3.39516133e-01
8.64822492e-02 1.26911747e+00 4.14008826e-01 -7.13117301e-01
3.25572193e-01 3.16800624e-01 -1.47228825e+00 9.41543341e-01
4.70135391e-01 1.03018510e+00 -1.63720384e-01 3.97150964e-01
-5.50221741e-01 1.01991558e+00 -7.91807353e-01 -7.68080652e-01
2.80057222e-01 -9.41747874e-02 -1.52802658e+00 -4.33654875e-01
6.17099106e-01 -3.18096161e-01 -1.92241326e-01 7.93497026e-01
1.79868877e-01 3.69957350e-02 6.76278546e-02 -1.20202386e+00
-3.22968543e-01 1.21280849e+00 -6.38050288e-02 3.82727504e-01
4.10976589e-01 -3.78293723e-01 1.35434234e+00 -5.52312076e-01
9.76098657e-01 1.46042788e+00 6.68342948e-01 3.84730011e-01
-9.28381860e-01 -5.06385744e-01 7.52966180e-02 7.94726908e-01
-1.48449397e+00 -9.19055402e-01 3.46076787e-01 1.04922742e-01
2.05520868e+00 4.43606764e-01 2.31268749e-01 6.25848353e-01
2.49499172e-01 8.38686764e-01 8.05240750e-01 -8.33897412e-01
9.11812484e-02 1.83447272e-01 1.03381269e-01 6.98227465e-01
6.95455849e-01 -5.78140259e-01 -6.70413435e-01 1.06991485e-01
5.65676630e-01 -9.96785104e-01 4.48402613e-01 -2.22892493e-01
-1.25273132e+00 6.70817018e-01 2.07707018e-01 5.35892546e-01
-1.05815381e-01 2.35333964e-02 8.85276556e-01 1.89425811e-01
1.33428082e-01 8.63366842e-01 -8.45331311e-01 -4.65003818e-01
-7.59612381e-01 5.59129119e-01 1.42343032e+00 1.30312717e+00
4.99126166e-01 -1.98107108e-01 6.52682632e-02 6.79544330e-01
1.07929535e-01 4.51545119e-02 1.75878540e-01 -1.38395154e+00
9.68927979e-01 6.25815690e-01 3.08055192e-01 -9.97941375e-01
-4.66975540e-01 -1.62400529e-01 -2.21049204e-01 -2.07169667e-01
7.14536846e-01 2.07884505e-01 -5.15309453e-01 1.21733868e+00
-5.98816015e-02 -4.97486293e-01 3.86077136e-01 5.46362042e-01
6.14616990e-01 4.09081548e-01 2.11292520e-01 1.05764762e-01
1.53376925e+00 -9.50752914e-01 -1.19883847e+00 -3.90938431e-01
9.00520623e-01 -8.58989835e-01 9.95734632e-01 6.85995877e-01
-1.50191879e+00 -3.76108497e-01 -1.46596134e+00 -6.71069503e-01
-9.26802576e-01 -1.46160275e-01 9.90722001e-01 6.70779824e-01
-1.16730869e+00 6.33400679e-01 -1.13837588e+00 -7.80842900e-01
2.32037008e-01 5.83542585e-01 -2.58993328e-01 2.51951426e-01
-1.57773924e+00 1.18971324e+00 9.73916650e-01 -1.41924480e-02
-1.97660662e-02 -4.77053642e-01 -8.82516265e-01 1.39302254e-01
9.07805264e-01 -4.92127568e-01 1.52938771e+00 -1.00256510e-01
-1.08121264e+00 9.86046135e-01 -4.26179260e-01 -9.72463429e-01
3.48154634e-01 -3.90543431e-01 -8.28557432e-01 -7.16810897e-02
3.85731876e-01 7.89131045e-01 -5.24735808e-01 -7.59285748e-01
-1.00432169e+00 -3.62636298e-01 3.76738369e-01 3.24877739e-01
2.62648817e-02 3.82590145e-01 -7.88673162e-01 -5.57374835e-01
1.83588564e-01 -5.30881643e-01 -7.04694167e-02 -4.05140072e-01
-3.03222477e-01 -1.13043994e-01 4.60036159e-01 -1.16912818e+00
1.39239955e+00 -1.66379166e+00 -4.17440869e-02 6.24867640e-02
-7.18834251e-02 1.71931416e-01 2.59169877e-01 6.71728134e-01
2.45807454e-01 4.77364272e-01 -1.24129690e-01 -2.88009912e-01
2.38102645e-01 5.88996112e-01 -3.97968322e-01 -2.42925838e-01
3.84676367e-01 9.06462729e-01 -5.99497795e-01 -6.67367160e-01
2.15901718e-01 1.24859124e-01 -6.33914948e-01 -1.89361632e-01
-7.07574308e-01 1.08914068e-02 -1.30652413e-01 6.86650097e-01
4.68481004e-01 -1.93655655e-01 7.51223326e-01 -1.96302846e-01
-1.39738798e-01 1.23896348e+00 -1.46174431e+00 1.98633802e+00
-3.62647891e-01 5.38748920e-01 -1.56211585e-01 -8.01904917e-01
6.81108356e-01 1.02861449e-01 2.52696246e-01 -1.11754870e+00
3.25967446e-02 5.37885129e-01 -1.45733077e-02 -2.35186324e-01
1.07676041e+00 1.02373771e-01 -2.14770451e-01 1.99169666e-01
7.65274018e-02 1.50091708e-01 6.73210859e-01 1.86331138e-01
7.79466629e-01 3.34999830e-01 6.39077604e-01 -5.03642321e-01
8.60250294e-01 6.81922257e-01 4.67393219e-01 7.62285769e-01
-5.93868308e-02 3.11552167e-01 5.84735811e-01 -7.93698609e-01
-1.17998099e+00 -7.21669137e-01 -3.19992900e-01 9.53922689e-01
9.68240574e-02 -9.75444853e-01 -7.94021964e-01 -6.33819401e-01
3.58057069e-03 1.18302536e+00 -3.57983202e-01 1.78494930e-01
-1.25030959e+00 -6.80952132e-01 1.04709172e+00 1.13712633e+00
7.74057746e-01 -1.10325229e+00 -3.88812274e-01 6.21010125e-01
-5.42414069e-01 -1.55692875e+00 1.96874395e-01 5.09614527e-01
-7.67851412e-01 -9.71457839e-01 4.82438862e-01 -7.15890288e-01
-6.52452186e-02 -3.55309457e-01 1.63236201e+00 1.18544944e-01
-1.11137204e-01 -8.16301480e-02 -2.22484812e-01 -5.82016706e-01
-6.98104799e-01 6.85933948e-01 -1.02912821e-01 -8.53112817e-01
8.48100364e-01 3.05540357e-02 1.41115129e-01 1.25488386e-01
-6.36051953e-01 1.81962803e-01 2.06819221e-01 4.96914864e-01
5.60435355e-01 6.76802397e-02 5.35286367e-01 -1.38020825e+00
3.37245345e-01 -2.02354968e-01 -7.76005805e-01 3.43780994e-01
-9.74382460e-01 5.03218472e-01 4.80636716e-01 5.08968055e-01
-1.32038510e+00 -3.15334767e-01 -5.63932002e-01 3.76302838e-01
-2.10979655e-01 6.71756327e-01 -5.03792763e-01 2.33603001e-01
4.83920366e-01 -2.39950135e-01 -7.01985583e-02 -6.53711736e-01
2.70366848e-01 6.33998513e-01 6.30317688e-01 -8.31299484e-01
3.68919969e-01 4.03239518e-01 2.10667644e-02 -2.56153256e-01
-8.94436777e-01 -2.21661851e-01 -8.19962680e-01 3.82478207e-01
7.70037830e-01 -9.53253925e-01 -6.44445181e-01 1.93064392e-01
-1.02211750e+00 -2.09248871e-01 -2.54004836e-01 3.51884440e-02
-6.26231134e-01 1.94852859e-01 -9.10784006e-01 -3.50433469e-01
-2.03591511e-01 -1.21291840e+00 1.10728121e+00 -7.41806328e-02
-7.67084301e-01 -1.27530611e+00 -3.26684386e-01 7.66240418e-01
5.55902719e-01 -9.53488648e-02 1.15425134e+00 -1.08419299e+00
-8.56594086e-01 1.39142096e-01 -3.55657250e-01 -7.98340961e-02
1.56971171e-01 -4.14667219e-01 -7.30168998e-01 -3.71645428e-02
-3.12699139e-01 -2.46231794e-01 4.42249477e-01 -1.43472835e-01
1.11565912e+00 -3.13785076e-01 -4.34727490e-01 4.82462883e-01
1.40713477e+00 5.78472674e-01 7.11786509e-01 1.19389093e+00
5.39034665e-01 4.98631150e-01 9.83201087e-01 1.65647175e-02
8.65937412e-01 8.62556100e-01 1.07203580e-01 1.64156064e-01
-2.94651598e-01 -4.80200022e-01 -2.06909657e-01 7.58135498e-01
1.22542173e-01 -2.70540059e-01 -1.14114809e+00 4.58091348e-01
-1.89982748e+00 -4.48882997e-01 -7.00122043e-02 1.85404944e+00
8.64434481e-01 5.45570016e-01 -1.00681752e-01 1.95781514e-01
3.79172266e-01 7.96556324e-02 -3.19488019e-01 -8.28499794e-01
-2.95834184e-01 5.65976918e-01 9.21913207e-01 9.69025671e-01
-1.07271516e+00 1.49478960e+00 7.70830393e+00 4.95987117e-01
-7.56250560e-01 -2.36835033e-01 5.71879506e-01 2.33522758e-01
-2.15417325e-01 1.98986143e-01 -1.32246101e+00 -8.09768960e-02
1.34547377e+00 -4.51006562e-01 6.04076505e-01 7.63363898e-01
-2.39689872e-01 -1.83961675e-01 -1.24493527e+00 5.23965597e-01
1.89988583e-01 -1.85598588e+00 2.52644747e-01 -1.67433601e-02
2.32490227e-02 -5.20064905e-02 -3.87196541e-01 5.36091805e-01
2.46733293e-01 -9.29251969e-01 8.65850568e-01 1.06185846e-01
6.72198296e-01 -8.74433875e-01 1.06224680e+00 -1.36589780e-01
-9.79269564e-01 1.30933210e-01 -3.95162702e-01 -4.59377348e-01
4.08607535e-03 -9.98667255e-02 -1.20472229e+00 8.90894294e-01
9.19259071e-01 6.31414354e-01 -8.73084784e-01 4.29496884e-01
5.01526445e-02 3.35336953e-01 -3.10396105e-01 1.36278123e-01
1.41616672e-01 -6.98646680e-02 4.45149392e-01 1.25784755e+00
-1.14279881e-01 -2.04588503e-01 3.78081053e-02 7.61302948e-01
-1.16640203e-01 8.21181312e-02 -3.04433882e-01 -1.71677694e-01
6.00262463e-01 7.99864829e-01 -1.03465211e+00 -7.13960767e-01
-4.40681368e-01 7.22292900e-01 2.38347366e-01 -1.04850963e-01
-5.75616181e-01 -3.59200299e-01 6.73991978e-01 1.03868693e-01
2.38523513e-01 -3.18616986e-01 -7.97868371e-01 -1.20396960e+00
2.33868256e-01 -1.24839711e+00 8.07057440e-01 -8.07877600e-01
-7.29972780e-01 6.86901987e-01 4.33925450e-01 -5.08090854e-01
-6.12138629e-01 -7.00252771e-01 2.18527406e-01 9.27665830e-01
-1.22651541e+00 -7.52033532e-01 1.52250350e-01 1.07826486e-01
6.66899502e-01 -2.17923447e-01 1.24309313e+00 6.41090393e-01
-4.07158166e-01 7.98452795e-01 -1.74564987e-01 2.77158916e-01
7.50929177e-01 -1.48421884e+00 1.14469254e+00 9.24036562e-01
3.61678451e-01 9.79323149e-01 9.44321334e-01 -6.23218179e-01
-1.49110019e+00 -7.59108007e-01 1.58044159e+00 -7.01644599e-01
9.13461924e-01 -5.85485458e-01 -9.99718368e-01 1.40014958e+00
5.06155849e-01 -5.10909617e-01 4.35657144e-01 4.25738901e-01
-3.43977422e-01 -2.58763701e-01 -1.24208212e+00 5.34464180e-01
1.00550914e+00 -4.66729194e-01 -9.32161331e-01 2.71166831e-01
9.06126440e-01 -9.22896326e-01 -1.08359325e+00 5.52886605e-01
4.54344571e-01 -8.68134916e-01 8.31923723e-01 -8.51853073e-01
1.71768337e-01 -3.77627879e-01 -5.16713917e-01 -8.44541192e-01
-6.79584667e-02 -2.82707244e-01 -1.28231317e-01 1.30937564e+00
7.27536798e-01 -6.14587903e-01 1.08040464e+00 6.36409104e-01
-2.84360111e-01 -4.69270706e-01 -7.50442326e-01 -5.87548196e-01
2.56521195e-01 -7.15062320e-01 9.64343488e-01 9.00803208e-01
1.71269163e-01 3.86298686e-01 -9.63016823e-02 -2.04546172e-02
1.16056547e-01 -4.78009842e-02 5.19864738e-01 -1.02070796e+00
-1.74334988e-01 -2.26596922e-01 -3.58232379e-01 -8.52713585e-01
9.43218023e-02 -1.02471197e+00 -2.15814680e-01 -1.75615919e+00
-7.95881748e-02 -2.90423721e-01 -4.03184891e-01 7.79120505e-01
7.09178597e-02 3.82771902e-02 1.41026527e-01 5.71396351e-02
-7.51340330e-01 -2.06493378e-01 6.83466494e-01 -5.29112816e-02
1.27837613e-01 -5.97686768e-01 -1.30492520e+00 4.18351859e-01
8.20791125e-01 -2.69897103e-01 -1.58854142e-01 -6.92564309e-01
4.40297812e-01 -2.20253821e-02 -2.98048466e-01 -1.20895898e+00
2.46189207e-01 1.52573949e-02 4.55127090e-01 -8.13902259e-01
2.78450012e-01 -7.86398828e-01 7.00565353e-02 4.53252286e-01
-6.14395797e-01 5.22607327e-01 8.58413696e-01 -1.48219302e-01
-4.14653391e-01 3.94795043e-03 3.11683655e-01 -2.81909049e-01
-1.20996225e+00 -3.44785631e-01 -5.45351684e-01 2.33571202e-01
5.39915204e-01 6.58662319e-02 -4.92618293e-01 -1.14586480e-01
-6.76931560e-01 3.52560729e-01 5.36183417e-01 7.05267727e-01
1.05403662e-02 -1.03801930e+00 -3.17579687e-01 2.75878340e-01
1.93615839e-01 -1.23094156e-01 -3.16601485e-01 4.20326382e-01
-1.21284211e+00 1.24615431e+00 -6.80709064e-01 -1.77831680e-01
-1.12029111e+00 5.04412055e-01 1.77886158e-01 -5.45433462e-01
-6.52969480e-01 5.06766260e-01 -7.28478193e-01 -6.24323964e-01
3.92199844e-01 -5.79570770e-01 -2.33936161e-01 2.10594246e-03
4.89891618e-01 4.41995949e-01 9.33945477e-01 -3.94814938e-01
-6.78675652e-01 -3.63284424e-02 -3.81351024e-01 -1.99540585e-01
1.20387578e+00 -1.19382702e-01 -3.06108952e-01 6.22375846e-01
7.96801150e-01 1.17412023e-01 -4.35557961e-01 -1.68687850e-01
8.73119593e-01 -1.80406362e-01 -1.33237287e-01 -1.42786360e+00
-6.68248415e-01 5.55213392e-01 1.70010954e-01 9.85059589e-02
1.00758326e+00 -1.95478305e-01 1.04170001e+00 9.06781793e-01
6.51206434e-01 -1.20822704e+00 -6.98041022e-01 8.74397159e-01
5.37485123e-01 -1.15020442e+00 7.86723420e-02 -7.38026023e-01
-4.33204770e-01 1.17388606e+00 6.47514880e-01 3.11538309e-01
1.44806474e-01 7.12179244e-01 2.61479944e-01 -1.61290348e-01
-9.89744723e-01 -1.49321139e-01 -3.81892920e-02 4.17056739e-01
7.34466255e-01 7.02997893e-02 -5.46705484e-01 4.38091218e-01
-7.08522797e-01 4.84900214e-02 5.25972962e-01 1.27581203e+00
-3.20781529e-01 -1.84503174e+00 -2.49760792e-01 2.91622907e-01
-6.90538406e-01 -2.92768776e-01 -6.25442863e-01 1.43942344e+00
1.77604356e-03 8.54249179e-01 3.06033492e-01 -3.47872555e-01
5.83953083e-01 5.55243134e-01 5.98980129e-01 -7.50102818e-01
-6.71543121e-01 -1.89615399e-01 1.00602674e+00 -7.22209513e-01
-1.15535595e-01 -8.27537417e-01 -1.59167111e+00 -4.34961498e-01
3.04612279e-01 4.40210789e-01 6.36568367e-01 9.51275527e-01
4.68663782e-01 6.12885177e-01 -4.29570585e-01 -1.31962728e-02
-3.51734430e-01 -8.56871903e-01 -2.08152771e-01 -7.30748028e-02
-2.92956054e-01 -5.96973181e-01 7.13905245e-02 1.00069903e-01] | [9.558639526367188, 7.931718826293945] |
0d9ebb87-eae4-4a50-a0fb-e28edb48bd8e | learning-how-to-interact-with-a-complex | 2204.10374 | null | https://arxiv.org/abs/2204.10374v1 | https://arxiv.org/pdf/2204.10374v1.pdf | Learning how to Interact with a Complex Interface using Hierarchical Reinforcement Learning | Hierarchical Reinforcement Learning (HRL) allows interactive agents to decompose complex problems into a hierarchy of sub-tasks. Higher-level tasks can invoke the solutions of lower-level tasks as if they were primitive actions. In this work, we study the utility of hierarchical decompositions for learning an appropriate way to interact with a complex interface. Specifically, we train HRL agents that can interface with applications in a simulated Android device. We introduce a Hierarchical Distributed Deep Reinforcement Learning architecture that learns (1) subtasks corresponding to simple finger gestures, and (2) how to combine these gestures to solve several Android tasks. Our approach relies on goal conditioning and can be used more generally to convert any base RL agent into an HRL agent. We use the AndroidEnv environment to evaluate our approach. For the experiments, the HRL agent uses a distributed version of the popular DQN algorithm to train different components of the hierarchy. While the native action space is completely intractable for simple DQN agents, our architecture can be used to establish an effective way to interact with different tasks, significantly improving the performance of the same DQN agent over different levels of abstraction. | ['Doina Precup', 'Philippe Hamel', 'Tyler Jackson', 'Zafarali Ahmed', 'Daniel Toyama', 'Anita Gergely', 'Amelia Glaese', 'Gheorghe Comanici'] | 2022-04-21 | null | null | null | null | ['hierarchical-reinforcement-learning'] | ['methodology'] | [ 8.14598352e-02 3.96738231e-01 1.00961290e-02 -3.87020111e-02
-5.98238349e-01 -8.47218812e-01 4.93268162e-01 -3.76993239e-01
-4.12156135e-01 6.69676244e-01 -4.12972234e-02 -4.54323769e-01
-2.60517150e-01 -7.27357090e-01 -7.66383708e-01 -6.99157298e-01
-1.63895980e-01 7.61937082e-01 4.64760512e-01 -4.65909034e-01
1.58782434e-02 4.49013501e-01 -1.89124489e+00 5.52157998e-01
6.99978769e-01 6.65011823e-01 3.54697347e-01 1.14243460e+00
6.10989295e-02 1.11825800e+00 -7.70906389e-01 9.38002020e-02
4.19226050e-01 -5.08138537e-01 -1.14534283e+00 -7.77649879e-02
2.42539570e-01 -7.41768897e-01 2.69288570e-01 6.29629433e-01
4.75520581e-01 2.40200952e-01 5.32317460e-01 -1.46483243e+00
9.19465721e-02 8.27187300e-01 -5.17258123e-02 -4.83303428e-01
5.81714690e-01 4.39490110e-01 1.02909088e+00 -5.92932152e-03
4.35798317e-01 1.57171464e+00 3.84316236e-01 1.02962923e+00
-1.37526047e+00 -5.60106277e-01 3.90766352e-01 3.46370861e-02
-8.25336337e-01 -9.76919383e-02 3.92437369e-01 -3.74107867e-01
9.96710956e-01 2.19982103e-01 7.26391256e-01 1.41011798e+00
-8.12512711e-02 8.82639706e-01 1.32080555e+00 -2.83670664e-01
4.30832297e-01 -3.18832606e-01 1.96367219e-01 8.47916663e-01
-2.40334012e-02 2.53891826e-01 -4.63486254e-01 -5.05278826e-01
1.11457264e+00 -9.21194553e-02 1.05414994e-01 -5.55359721e-01
-9.37847853e-01 7.60572314e-01 3.34745675e-01 2.35162109e-01
-4.59551930e-01 4.62161809e-01 1.40456423e-01 5.85831761e-01
-3.62714291e-01 9.16476071e-01 -7.27407992e-01 -2.95544744e-01
-2.43354619e-01 5.71494877e-01 1.32604361e+00 7.02005088e-01
7.81926811e-01 -2.32485875e-01 -1.90749004e-01 5.58772087e-01
2.53532439e-01 1.79774985e-01 4.29004759e-01 -1.88626921e+00
1.08406991e-01 4.92826134e-01 4.16316241e-01 -3.89858395e-01
-6.32843256e-01 3.04556191e-02 -3.30697149e-01 1.04389000e+00
5.92588902e-01 -7.00568378e-01 -6.56617820e-01 1.93335164e+00
5.21049678e-01 2.84929983e-02 2.04959512e-01 6.92781806e-01
4.66393858e-01 5.66563845e-01 3.04922521e-01 5.19543365e-02
1.22334719e+00 -1.05247581e+00 -4.23205554e-01 -7.30008399e-03
7.45995879e-01 -2.29875028e-01 1.42583466e+00 7.23005295e-01
-1.21592546e+00 -6.91727519e-01 -9.66016054e-01 2.42252275e-02
-3.46892774e-01 -1.01230919e-01 8.46105278e-01 5.34753382e-01
-1.41705132e+00 7.21890330e-01 -9.43453372e-01 -1.62036955e-01
9.24057886e-02 8.91442418e-01 1.08810514e-01 2.74540752e-01
-1.06019461e+00 7.70648718e-01 3.62326413e-01 -2.87680447e-01
-1.59350550e+00 -2.67015725e-01 -7.25099266e-01 1.41800493e-01
8.47283840e-01 -8.64686668e-01 1.91344774e+00 -8.57575774e-01
-2.23264670e+00 3.98804367e-01 2.65802741e-01 -1.74017459e-01
4.07302141e-01 -3.04935902e-01 3.79234344e-01 3.13562900e-02
1.25027150e-01 8.89240623e-01 9.60611582e-01 -1.38487697e+00
-9.02851582e-01 -5.43378055e-01 1.10720420e+00 5.74764609e-01
7.79974163e-02 -2.13774979e-01 -4.03939098e-01 -3.73380840e-01
-4.58191305e-01 -1.36841667e+00 -4.23252583e-01 -3.43834162e-01
-1.92965448e-01 -6.31989062e-01 6.32081866e-01 -3.51581424e-01
7.96911478e-01 -1.95646238e+00 8.53532255e-01 1.32585824e-01
4.42399651e-01 -4.26776595e-02 -3.92977566e-01 3.09503853e-01
3.47516984e-01 -1.04938354e-02 -1.08171262e-01 -2.25478262e-01
3.95662785e-01 5.08253217e-01 1.11794591e-01 -2.02001393e-01
-2.32073724e-01 9.59577262e-01 -8.95095170e-01 -2.22003892e-01
-3.39612411e-03 2.40555406e-01 -8.26797068e-01 6.12243533e-01
-8.88505220e-01 6.52629793e-01 -6.53435647e-01 2.22279787e-01
-1.38939172e-02 -6.68829083e-02 6.97023153e-01 1.74601242e-01
1.04436856e-02 4.75784868e-01 -1.29634917e+00 1.82919908e+00
-7.96724677e-01 2.75002301e-01 4.91513938e-01 -6.56585515e-01
3.96080077e-01 3.29280972e-01 5.40811718e-01 -6.72371864e-01
-1.24575622e-01 -2.86323223e-02 3.26779515e-01 -5.78490078e-01
4.50449297e-03 1.09525934e-01 -4.03111130e-01 9.21019316e-01
1.44094020e-01 -3.57242167e-01 1.13849625e-01 -3.43610421e-02
1.41085243e+00 7.41048098e-01 1.04674041e-01 4.93678153e-02
3.49401891e-01 -9.03197229e-02 2.08363056e-01 1.22123182e+00
1.84348851e-01 -1.80340901e-01 8.41543376e-01 -3.72405440e-01
-6.00931406e-01 -1.03223181e+00 4.97970611e-01 2.04009128e+00
1.07721798e-01 -4.52004135e-01 -1.07491195e+00 -9.51137602e-01
-1.79927483e-01 5.39479017e-01 -5.67960680e-01 1.18411250e-01
-9.02738214e-01 -4.47151363e-01 7.45142043e-01 4.06141549e-01
7.36117661e-01 -1.79627860e+00 -1.36644781e+00 1.69809744e-01
-4.00374122e-02 -8.68528664e-01 -1.26277000e-01 5.34548461e-01
-7.88025439e-01 -1.28690326e+00 -5.64024806e-01 -8.72386932e-01
2.69360214e-01 -9.46612060e-02 1.24077094e+00 3.80122274e-01
-3.21332961e-02 8.63998234e-01 -4.11689073e-01 -2.82966584e-01
-5.47706187e-01 5.04781902e-01 -4.08388563e-02 -2.67549634e-01
-7.05469996e-02 -7.44456768e-01 -4.17924255e-01 3.40508997e-01
-8.72534215e-01 3.11807599e-02 5.16764104e-01 6.62535310e-01
1.35442898e-01 2.15004489e-01 2.24544168e-01 -9.61282253e-01
1.26850021e+00 -2.56980024e-02 -7.81997859e-01 1.22072279e-01
-2.38486230e-01 5.63873708e-01 6.23259246e-01 -6.63989067e-01
-1.05435503e+00 4.92747128e-01 -2.63052553e-01 1.68777004e-01
-5.39732337e-01 1.52123585e-01 -3.75240773e-01 -8.20888430e-02
7.26006687e-01 5.36933588e-03 -9.75443274e-02 -4.79453355e-01
6.28077149e-01 5.51480353e-01 2.75276989e-01 -1.32534277e+00
5.17124236e-01 6.96324706e-02 -6.91885576e-02 -6.85074866e-01
-4.55276996e-01 8.96984562e-02 -5.32415032e-01 -5.34428619e-02
1.06836677e+00 -4.04419422e-01 -1.80834675e+00 4.85773146e-01
-1.12390828e+00 -1.68298256e+00 -2.78412253e-01 -1.22324847e-01
-1.06492400e+00 5.64606860e-03 -7.83690333e-01 -6.31949902e-01
1.26429096e-01 -1.66564226e+00 1.42071342e+00 2.44309366e-01
-4.62047100e-01 -7.35820353e-01 2.80957103e-01 1.98144168e-01
2.67333776e-01 4.77307141e-02 1.25038230e+00 -4.60378468e-01
-4.36553955e-01 3.96988511e-01 2.04620913e-01 2.24416498e-02
7.15783685e-02 -2.97397435e-01 -8.17033529e-01 -1.76262230e-01
9.69530940e-02 -9.35573518e-01 3.49988729e-01 2.11550251e-01
1.23498976e+00 -4.95786428e-01 -2.08358765e-01 2.71752030e-01
1.03716171e+00 5.69487214e-01 6.85078859e-01 4.32398021e-01
6.85725152e-01 6.91201866e-01 4.60080028e-01 2.37936288e-01
5.13965011e-01 1.17585921e+00 5.69659948e-01 -2.77934503e-02
4.58136536e-02 -1.99584588e-02 6.28032744e-01 -5.39996065e-02
-4.93630499e-01 2.40551263e-01 -8.19007635e-01 -1.45819306e-01
-2.25286841e+00 -8.46264422e-01 2.76694953e-01 1.91745853e+00
1.09193087e+00 8.60918462e-02 7.87501812e-01 -1.69880956e-01
2.22465694e-01 -7.71663412e-02 -7.91867495e-01 -4.96447444e-01
6.15797698e-01 4.74271119e-01 -1.45396262e-01 9.87119317e-01
-1.01125872e+00 1.27095103e+00 6.55249977e+00 5.21963060e-01
-9.03747499e-01 8.20495337e-02 2.03145355e-01 3.01830540e-03
1.63613424e-01 -2.51998484e-01 -8.46071541e-01 9.20358449e-02
7.58327723e-01 4.87260520e-01 1.14781272e+00 7.67975688e-01
-1.78500395e-02 -1.32277966e-01 -1.59081900e+00 8.64599168e-01
-3.72837007e-01 -9.55885589e-01 7.34137297e-02 1.91131577e-01
3.92544925e-01 -2.20212072e-01 2.62210472e-03 1.09191275e+00
1.20154881e+00 -1.23092580e+00 3.14935118e-01 2.78217047e-01
5.60228229e-01 -5.16648054e-01 2.70685405e-01 7.88524687e-01
-9.12898362e-01 -3.78134161e-01 1.25712633e-01 -4.83289689e-01
-2.98068941e-01 -6.49450660e-01 -6.48525536e-01 1.00473493e-01
7.79470861e-01 2.20519751e-01 -5.32989025e-01 4.89800990e-01
-5.82441807e-01 1.33427992e-01 -3.95764023e-01 -2.90375799e-01
3.90698731e-01 -1.67517945e-01 1.87659279e-01 8.36771667e-01
-1.16403572e-01 3.88908893e-01 5.10870695e-01 5.80851972e-01
5.65458722e-02 -2.77315199e-01 -6.89677000e-01 2.37276390e-01
7.37194195e-02 1.22219884e+00 -5.35648584e-01 -3.97228211e-01
-2.31774494e-01 1.26192534e+00 7.77292371e-01 7.09489524e-01
-8.20326269e-01 -3.22153449e-01 9.34188366e-01 -3.91443595e-02
2.47596011e-01 -4.38013196e-01 -4.01550569e-02 -1.06525970e+00
-2.65221030e-01 -1.68667674e+00 3.48135680e-01 -8.60413373e-01
-8.43199372e-01 6.10443234e-01 1.36539012e-01 -6.88957274e-01
-7.90087461e-01 -8.36605132e-01 -4.59951699e-01 5.34419179e-01
-9.14807737e-01 -8.60060155e-01 -4.89868373e-01 9.93511915e-01
7.77508140e-01 -1.27193555e-01 1.31674027e+00 -2.75240332e-01
-4.30671453e-01 2.01748714e-01 -5.11952937e-01 1.20184608e-01
2.95039445e-01 -1.67526340e+00 9.86879766e-02 2.65719801e-01
1.55931637e-01 6.12404048e-01 4.11613107e-01 -6.03151917e-01
-1.16373050e+00 -5.22455335e-01 3.76879424e-02 -6.10811055e-01
4.77335095e-01 -5.76902807e-01 -5.47565222e-01 8.65694761e-01
5.14966547e-01 -6.47202909e-01 5.69228351e-01 3.97930115e-01
-2.32012108e-01 -7.13041723e-02 -7.93781579e-01 1.11888587e+00
1.17362201e+00 -4.28382665e-01 -6.54550254e-01 2.09917873e-01
7.62996554e-01 -5.01035869e-01 -6.72216237e-01 2.36372858e-01
8.15655887e-01 -1.00326455e+00 9.36115086e-01 -1.00737214e+00
3.59061033e-01 -2.72694498e-01 -1.01266000e-02 -1.38679886e+00
-2.76418328e-01 -9.19891179e-01 -3.78799975e-01 6.39762342e-01
3.22561860e-01 -4.28105146e-01 7.86835074e-01 7.28670061e-01
2.03435585e-01 -6.68220878e-01 -5.34086347e-01 -5.39603114e-01
1.16409525e-01 -4.31027949e-01 6.07434690e-01 4.69078690e-01
1.69089824e-01 7.79041588e-01 -2.11793095e-01 2.39052325e-01
5.19046903e-01 9.18778107e-02 1.17303336e+00 -1.32979763e+00
-1.15651953e+00 -6.70784414e-01 3.41234148e-01 -1.34459090e+00
3.98181677e-01 -6.81489646e-01 3.14489663e-01 -1.42379034e+00
-1.33218557e-01 -6.67895257e-01 2.94713061e-02 9.61937606e-01
7.52404332e-02 1.40375532e-02 4.74792212e-01 1.23311952e-01
-9.98306751e-01 1.80962682e-01 1.32115352e+00 -2.59550273e-01
-7.80045867e-01 3.81975919e-01 -6.44839227e-01 9.40694034e-01
9.78917658e-01 -4.12258983e-01 -6.50275052e-01 -4.88943160e-01
4.57608074e-01 3.36359680e-01 4.62870091e-01 -1.06861925e+00
1.63256034e-01 -2.34443083e-01 3.01636726e-01 1.60343990e-01
2.85644323e-01 -1.13333905e+00 -6.74070418e-02 6.90282404e-01
-6.66323662e-01 -2.44059712e-02 2.15425566e-01 8.69490802e-02
2.56530732e-01 -3.63921791e-01 3.33130181e-01 -3.79126549e-01
-4.89576757e-01 -1.46295324e-01 -9.71425295e-01 -2.54258007e-01
1.05026793e+00 3.50207873e-02 -5.89522570e-02 -6.96558416e-01
-1.38043499e+00 3.24623138e-01 3.97470057e-01 2.90907651e-01
3.07064027e-01 -1.01168776e+00 -1.80436879e-01 7.79538825e-02
-3.37411553e-01 1.88932553e-01 -3.03602934e-01 4.00607079e-01
-4.72224325e-01 1.69384345e-01 -5.40622771e-01 -3.59208703e-01
-1.30571604e+00 5.97277105e-01 6.75624728e-01 -6.55805290e-01
-3.58069032e-01 3.78959119e-01 3.90553474e-01 -7.98816144e-01
6.99649692e-01 -4.94842827e-01 -3.92408401e-01 -3.83156464e-02
4.23354387e-01 3.64425540e-01 -2.84161270e-01 3.73613760e-02
-1.32129371e-01 5.67164242e-01 3.59921008e-02 -6.29473686e-01
1.25532269e+00 2.14131877e-01 -1.29475072e-01 3.38655144e-01
5.99262953e-01 -2.70923644e-01 -1.57912230e+00 2.70467609e-01
1.36791795e-01 2.99294949e-01 -4.54100043e-01 -1.08875966e+00
-5.80789447e-01 7.27532327e-01 4.88716930e-01 5.27511477e-01
1.12966168e+00 9.99935269e-02 4.36107814e-01 9.65006173e-01
8.43505442e-01 -9.32152092e-01 5.33142447e-01 7.40406275e-01
1.00501263e+00 -9.90996599e-01 -3.22149664e-01 -1.02337942e-01
-5.87548435e-01 1.19591296e+00 8.77609134e-01 -2.65362650e-01
3.29029351e-01 5.39185882e-01 3.57116722e-02 -2.62305290e-01
-7.01233029e-01 -5.00783265e-01 -4.60507460e-02 9.15102482e-01
2.46138930e-01 1.03775291e-02 -6.68161437e-02 5.96025765e-01
-1.36431053e-01 9.61386934e-02 3.13421249e-01 9.61816967e-01
-2.27980435e-01 -1.56974411e+00 -2.77014494e-01 1.72439769e-01
-2.19199508e-01 1.82266131e-01 -6.84036851e-01 9.46665823e-01
2.93378204e-01 8.78931880e-01 -2.61346251e-01 -4.03689682e-01
4.14500952e-01 1.00133188e-01 9.62529838e-01 -1.00674093e+00
-8.75406742e-01 6.06630445e-02 9.13382918e-02 -1.16004026e+00
-3.84963959e-01 -5.96013367e-01 -1.49251795e+00 1.18957043e-01
5.29830456e-01 -3.29750367e-02 2.96989799e-01 1.08543682e+00
1.72657683e-01 6.63071513e-01 2.53867805e-01 -1.52194381e+00
-6.50062323e-01 -6.88823044e-01 -3.71563524e-01 3.23398530e-01
3.97862554e-01 -5.63505888e-01 5.68326935e-02 -7.23335519e-02] | [4.109830379486084, 1.4150550365447998] |
8f7d9151-d0a3-4c3e-b60d-b89bf185dded | structured-context-transformer-for-generic | 2206.02985 | null | https://arxiv.org/abs/2206.02985v1 | https://arxiv.org/pdf/2206.02985v1.pdf | Structured Context Transformer for Generic Event Boundary Detection | Generic Event Boundary Detection (GEBD) aims to detect moments where humans naturally perceive as event boundaries. In this paper, we present Structured Context Transformer (or SC-Transformer) to solve the GEBD task, which can be trained in an end-to-end fashion. Specifically, we use the backbone convolutional neural network (CNN) to extract the features of each video frame. To capture temporal context information of each frame, we design the structure context transformer (SC-Transformer) by re-partitioning input frame sequence. Note that, the overall computation complexity of SC-Transformer is linear to the video length. After that, the group similarities are computed to capture the differences between frames. Then, a lightweight fully convolutional network is used to determine the event boundaries based on the grouped similarity maps. To remedy the ambiguities of boundary annotations, the Gaussian kernel is adopted to preprocess the ground-truth event boundaries to further boost the accuracy. Extensive experiments conducted on the challenging Kinetics-GEBD and TAPOS datasets demonstrate the effectiveness of the proposed method compared to the state-of-the-art methods. | ['Longyin Wen', 'Tiejian Luo', 'Libo Zhang', 'YuFei Wang', 'Dexiang Hong', 'Xinyao Wang', 'CongCong Li'] | 2022-06-07 | null | null | null | null | ['boundary-detection'] | ['computer-vision'] | [ 2.27696270e-01 -2.95253485e-01 1.92071378e-01 -3.58823568e-01
-6.14403844e-01 -3.23694855e-01 4.14041936e-01 3.28594655e-01
-6.15949571e-01 3.18373799e-01 1.53052971e-01 9.69215017e-03
9.18819010e-02 -4.97428507e-01 -6.60118401e-01 -6.42385364e-01
-2.04284742e-01 -1.36511445e-01 5.81311345e-01 2.03977615e-01
3.28488618e-01 3.24277788e-01 -1.36944139e+00 3.70496064e-01
6.10999882e-01 1.44376969e+00 2.79478997e-01 4.97308046e-01
1.35143176e-01 9.13102865e-01 -4.71716762e-01 6.75570518e-02
9.19955373e-02 -5.07805407e-01 -7.09750712e-01 2.22533569e-01
3.08086663e-01 -3.80067140e-01 -5.43894529e-01 8.08405578e-01
5.73411465e-01 5.51995873e-01 3.54398519e-01 -1.07376432e+00
-1.48826808e-01 4.25819270e-02 -5.69391310e-01 8.70330691e-01
3.03377360e-01 -3.29294545e-03 8.82499516e-01 -1.09475613e+00
6.26551390e-01 1.10686672e+00 5.51418602e-01 1.59489930e-01
-9.13018882e-01 -4.13095027e-01 4.83353913e-01 6.41392171e-01
-1.50659108e+00 -2.65856564e-01 8.05487692e-01 -4.95153666e-01
8.03504646e-01 -4.96046208e-02 6.87442183e-01 9.71596122e-01
1.78812042e-01 7.54627168e-01 5.53721070e-01 -1.79219171e-01
3.15267920e-01 -7.38767385e-01 3.44142094e-02 6.37559533e-01
-1.97774142e-01 -1.32853109e-02 -6.89203560e-01 6.65532723e-02
7.31927574e-01 2.61509031e-01 -3.73860896e-01 -1.69985935e-01
-1.44781756e+00 5.31975925e-01 3.48361105e-01 -3.02165723e-03
-5.92209816e-01 1.18802056e-01 1.01070046e+00 -1.64729562e-02
5.14363766e-01 -2.03581735e-01 -2.74931103e-01 -2.55953312e-01
-9.99308407e-01 4.80188787e-01 4.89167362e-01 7.52029002e-01
5.70167661e-01 -2.64935911e-01 -6.64517224e-01 4.82582331e-01
1.00413918e-01 -2.19726056e-01 4.32117879e-01 -9.46042061e-01
4.13212031e-01 6.74386263e-01 3.00105512e-01 -1.02542233e+00
-5.08121669e-01 -1.67753279e-01 -6.64858818e-01 -1.59000069e-01
4.89879251e-01 -1.08472936e-01 -7.85560191e-01 1.59746075e+00
6.41740143e-01 9.25015926e-01 -1.17703542e-01 1.35902560e+00
7.51314223e-01 7.22676039e-01 2.02947095e-01 -2.39987418e-01
1.60424173e+00 -9.96801138e-01 -7.30221868e-01 -8.48131701e-02
3.82211804e-01 -6.04818285e-01 8.93812895e-01 2.15145573e-01
-9.86433148e-01 -7.76430428e-01 -1.02065122e+00 -1.37138054e-01
-5.22655174e-02 3.90794486e-01 2.68967360e-01 -5.39696701e-02
-6.80878997e-01 5.31202972e-01 -1.30310142e+00 -1.71481252e-01
3.93737763e-01 8.64983350e-02 -1.56804353e-01 1.63140073e-01
-1.28538096e+00 3.14309567e-01 8.04366469e-01 4.33488637e-01
-1.26319635e+00 -6.48856819e-01 -1.05642509e+00 1.79655701e-01
7.29737699e-01 -4.27040875e-01 1.44403410e+00 -8.43771100e-01
-1.34612417e+00 6.51838481e-01 -5.09702921e-01 -5.80427110e-01
4.52691048e-01 -4.27075297e-01 -1.14703305e-01 6.00922823e-01
1.14985190e-01 4.96775806e-01 7.36259818e-01 -6.51723146e-01
-9.63702083e-01 -8.65103230e-02 7.35851526e-02 2.83566505e-01
7.70390630e-02 3.77530336e-01 -8.01416457e-01 -8.78287613e-01
2.30172053e-01 -7.98425734e-01 -1.33265421e-01 -9.79578942e-02
-2.59567231e-01 -5.65979421e-01 1.02771795e+00 -8.47762406e-01
1.41727030e+00 -2.44720292e+00 1.17770962e-01 -7.35019371e-02
3.54639292e-01 1.68991089e-01 2.10738257e-01 3.07897776e-01
-1.90762818e-01 -3.44874203e-01 -2.08627731e-01 -3.62941444e-01
-1.76436707e-01 1.02347732e-01 -1.68013766e-01 6.03301048e-01
4.09133762e-01 7.25668371e-01 -8.65347266e-01 -5.73188424e-01
2.69291103e-01 3.61587077e-01 -3.32624406e-01 3.22143614e-01
-3.45990866e-01 5.46098471e-01 -3.79186660e-01 4.22779739e-01
4.70637262e-01 -2.90654957e-01 3.63836326e-02 -3.28027010e-01
-2.83920676e-01 3.71017992e-01 -1.24646330e+00 1.85906279e+00
-3.87018546e-02 6.63494587e-01 -1.08793333e-01 -1.15748465e+00
7.22814679e-01 4.60164189e-01 6.42976224e-01 -4.20370758e-01
3.55260819e-01 -8.18803832e-02 -8.96204785e-02 -6.87626600e-01
3.70093316e-01 1.81995586e-01 -1.11706465e-01 3.15305144e-01
-4.55379076e-02 7.19614208e-01 3.90428483e-01 5.50190285e-02
9.84002590e-01 3.49331677e-01 4.30175632e-01 -8.91767666e-02
8.02227318e-01 -1.70225859e-01 1.10664129e+00 4.72066760e-01
-5.87798297e-01 7.56113768e-01 6.57460511e-01 -8.69417310e-01
-8.35215449e-01 -1.12142313e+00 3.78653824e-01 9.37264144e-01
4.91864890e-01 -5.76840639e-01 -1.12457764e+00 -7.21381545e-01
-4.20599461e-01 1.86508283e-01 -6.71145618e-01 -1.65799364e-01
-8.31712842e-01 -4.70232219e-01 3.75638962e-01 8.76016915e-01
9.26310062e-01 -1.20480609e+00 -1.07109332e+00 5.07950306e-01
-5.49723685e-01 -1.30535555e+00 -9.71211135e-01 -1.67494506e-01
-5.76816559e-01 -1.18148243e+00 -4.93821830e-01 -9.35402155e-01
3.97903502e-01 3.55096847e-01 6.72542095e-01 1.57943368e-02
-3.34348351e-01 -1.08440571e-01 -5.17263293e-01 -9.13944691e-02
2.81605035e-01 -7.53931403e-02 -1.18693151e-01 3.53471726e-01
5.48354328e-01 -4.93173361e-01 -9.99994099e-01 4.44922805e-01
-8.48733127e-01 1.75560877e-01 2.03612849e-01 8.29667330e-01
9.33110893e-01 1.85954168e-01 4.50993240e-01 -4.24596637e-01
5.10513246e-01 -3.84024918e-01 -6.65791452e-01 1.86351389e-01
-1.30819589e-01 -3.22973251e-01 4.17038471e-01 -5.73756814e-01
-1.02101147e+00 1.47086278e-01 2.26994097e-01 -6.75433218e-01
-3.52439374e-01 8.23793948e-01 -1.82955340e-01 3.55153859e-01
2.89023429e-01 3.59928697e-01 -3.69995445e-01 -3.89958948e-01
1.14265829e-01 5.09891868e-01 9.99113083e-01 -5.51124275e-01
1.98539197e-01 5.48229158e-01 -2.00144678e-01 -5.43864965e-01
-1.13335431e+00 -6.41054630e-01 -6.09881163e-01 -3.71584207e-01
1.28048706e+00 -9.87881184e-01 -8.22192550e-01 7.45785654e-01
-1.35092139e+00 -3.66077363e-01 1.56234950e-02 6.39004529e-01
-6.67036116e-01 4.48940575e-01 -8.14388275e-01 -5.80480456e-01
-3.99935156e-01 -1.11170816e+00 1.18057883e+00 5.59718668e-01
-1.81276768e-01 -7.06111848e-01 -1.84039790e-02 1.33601516e-01
-1.55887559e-01 5.10054529e-01 5.88123500e-01 -7.31589079e-01
-5.54068446e-01 2.93582622e-02 -4.19466287e-01 6.30679056e-02
1.17676355e-01 -1.37369603e-01 -6.82746232e-01 -2.12555692e-01
1.87742077e-02 4.40782271e-02 7.06452489e-01 5.64316213e-01
1.33470690e+00 -4.01945673e-02 -2.94673771e-01 6.40113890e-01
9.60924625e-01 4.52692389e-01 5.74984610e-01 4.55062717e-01
6.71899796e-01 3.59185934e-01 8.96052957e-01 7.66819835e-01
5.42794347e-01 6.16656601e-01 1.76974773e-01 -1.76921841e-02
1.06588610e-01 -1.72637179e-01 2.82256991e-01 3.94634306e-01
-2.35807733e-04 -2.45174751e-01 -7.81548798e-01 6.69106603e-01
-2.15080166e+00 -9.55696821e-01 4.11909670e-02 2.05631113e+00
5.36500096e-01 3.52934569e-01 1.33283809e-01 5.16728945e-02
1.21026409e+00 2.38656223e-01 -6.26718581e-01 -6.67067096e-02
1.86045423e-01 -4.15631235e-02 7.06496537e-02 1.30929589e-01
-1.39119899e+00 9.42093015e-01 5.43054819e+00 7.95178950e-01
-1.08841920e+00 1.67946309e-01 5.53514779e-01 -1.63895264e-01
5.59260607e-01 7.82391280e-02 -6.14184320e-01 6.57305241e-01
7.58267105e-01 -6.24636337e-02 3.83002847e-01 5.91627121e-01
8.05229783e-01 -2.27648571e-01 -1.10606897e+00 1.09825122e+00
-4.37249243e-03 -1.21721923e+00 -3.00538450e-01 -3.00951779e-01
4.94070143e-01 -9.64268520e-02 -4.04586166e-01 2.03268036e-01
-2.01795980e-01 -5.36003947e-01 1.05243111e+00 5.69050014e-01
5.11892796e-01 -9.21630085e-01 6.37613595e-01 2.47277573e-01
-1.76697445e+00 -3.94397713e-02 -2.96018958e-01 -1.25748202e-01
4.42038834e-01 6.30105972e-01 -5.99358797e-01 7.15502620e-01
1.01209402e+00 8.82278979e-01 -2.83477366e-01 1.12826836e+00
-3.18015069e-01 6.27469480e-01 -3.16316038e-01 4.54823524e-01
3.62342685e-01 -5.85497543e-02 5.46933234e-01 1.19181073e+00
3.98674190e-01 2.19844118e-01 5.65165818e-01 8.05295527e-01
-1.56395156e-02 -2.31591791e-01 -3.08206677e-02 -3.12570222e-02
5.76337576e-01 1.30436587e+00 -1.00837421e+00 -4.91533011e-01
-4.89902169e-01 1.31417561e+00 3.85175616e-01 4.57887173e-01
-1.04000962e+00 -6.10685587e-01 6.14821613e-01 -1.03924178e-01
6.83052123e-01 -4.08234149e-01 5.00000939e-02 -1.03399467e+00
2.41416216e-01 -6.69762433e-01 7.31384695e-01 -8.11695278e-01
-1.15490806e+00 4.38502282e-01 1.00101031e-01 -1.17106736e+00
3.80926859e-03 -3.09248745e-01 -1.00881779e+00 7.38476574e-01
-1.33695686e+00 -9.89132345e-01 -6.65925443e-01 6.54183269e-01
8.69125009e-01 2.82000929e-01 1.30160630e-01 2.77001530e-01
-9.40398097e-01 3.49944413e-01 -2.77740926e-01 4.34040964e-01
6.69254839e-01 -1.02308893e+00 5.53230107e-01 1.10795999e+00
-3.16924453e-01 5.01108646e-01 4.46014792e-01 -8.38666141e-01
-8.55628192e-01 -1.37451339e+00 9.05212939e-01 7.93741196e-02
4.37652677e-01 -3.23975295e-01 -9.58065987e-01 7.28826761e-01
1.34258838e-02 3.95902961e-01 5.36897063e-01 -2.81691879e-01
-9.85012501e-02 -2.68219113e-02 -7.06605792e-01 6.76224530e-01
1.18045604e+00 -5.03221154e-01 -7.74805844e-01 -1.34827616e-02
1.03323650e+00 -7.97761917e-01 -8.25500071e-01 2.55492389e-01
2.44649366e-01 -9.29301322e-01 8.40551734e-01 -3.23799670e-01
3.78766537e-01 -7.06185400e-01 7.67350048e-02 -9.86291707e-01
-2.19392762e-01 -7.64462173e-01 -2.63571709e-01 1.22735488e+00
-2.25830972e-01 -3.21651369e-01 5.82534730e-01 3.66401196e-01
-3.99152815e-01 -7.90208578e-01 -1.04090941e+00 -6.12998545e-01
-6.14664912e-01 -3.33551079e-01 6.03851199e-01 7.08670020e-01
4.51305397e-02 2.49489665e-01 -2.15160936e-01 3.42165530e-01
3.84574592e-01 2.22227603e-01 4.54766303e-01 -1.02556193e+00
-7.04096109e-02 -1.03536047e-01 -4.39289063e-01 -1.17341185e+00
1.73181266e-01 -5.45179009e-01 2.61112601e-01 -1.29026294e+00
1.20518863e-01 4.27782722e-02 -6.52671933e-01 4.00782496e-01
-5.71852982e-01 9.19563547e-02 1.05905116e-01 1.41488045e-01
-1.05030775e+00 6.23945475e-01 1.11159599e+00 1.15966491e-01
-3.28028381e-01 -2.53636658e-01 -2.49644086e-01 7.73580372e-01
6.90861523e-01 -3.99392188e-01 -4.80203748e-01 -1.29708573e-01
-9.58130658e-02 2.24327192e-01 7.12453306e-01 -1.19275117e+00
2.46186987e-01 -1.62266940e-01 3.41937214e-01 -9.97493982e-01
1.36605039e-01 -5.66911280e-01 3.63235772e-02 2.82876253e-01
-3.96318644e-01 1.84201837e-01 1.20988853e-01 6.80671334e-01
-3.74781936e-01 -1.07956529e-01 4.71915483e-01 -5.70614897e-02
-8.20517957e-01 5.90093136e-01 -6.36397958e-01 1.60383284e-01
1.13833034e+00 -1.15948662e-01 -5.16018420e-02 -1.21986091e-01
-6.60355449e-01 3.14555019e-01 2.33077869e-01 2.25002959e-01
7.44825006e-01 -1.44870007e+00 -3.87006193e-01 1.03599224e-02
1.08190663e-01 4.48082954e-01 5.57351172e-01 1.17102432e+00
-5.50009847e-01 1.67331368e-01 -1.50148235e-02 -7.33153224e-01
-1.15922785e+00 7.42249727e-01 3.88309598e-01 -2.24546596e-01
-1.07035041e+00 7.14942336e-01 4.72173989e-01 1.92985997e-01
2.80978829e-01 -5.79928935e-01 -2.84546971e-01 9.79822800e-02
9.07179356e-01 3.12779754e-01 -5.24253920e-02 -4.21942681e-01
-4.55192715e-01 3.01679641e-01 -2.55802602e-01 -1.54270614e-02
1.09627879e+00 -3.32348585e-01 -1.67732500e-02 2.41096526e-01
1.04449868e+00 -4.86172169e-01 -1.99665916e+00 -3.22544664e-01
1.14574894e-01 -2.46429950e-01 2.62212548e-02 -3.07024568e-01
-1.03831661e+00 8.87100816e-01 6.06014252e-01 -1.60601944e-01
1.35568857e+00 -2.27504641e-01 1.14047682e+00 1.71664461e-01
1.88835966e-03 -9.88868356e-01 9.67896655e-02 5.13864815e-01
5.67903161e-01 -8.92251492e-01 -2.73732364e-01 -4.01241362e-01
-5.46888471e-01 1.12187636e+00 6.94062948e-01 -3.00155491e-01
4.79718477e-01 -2.70429999e-02 -1.22895494e-01 -3.02332133e-01
-6.18894994e-01 -2.26100788e-01 2.08248243e-01 2.86245227e-01
2.50015825e-01 -2.46908799e-01 -4.98807907e-01 8.40797663e-01
3.00722122e-01 5.14986813e-01 2.26979762e-01 1.18953764e+00
-4.07358706e-01 -7.45854974e-01 -3.37835580e-01 8.67550373e-02
-5.12125611e-01 2.26008948e-02 -1.43412083e-01 4.93213207e-01
3.67923081e-01 9.81582284e-01 3.82988065e-01 -4.23430294e-01
3.62313390e-01 1.85946345e-01 3.03805977e-01 -3.07848662e-01
-5.46740949e-01 4.28881735e-01 -1.02523319e-01 -8.37409317e-01
-5.14681160e-01 -7.68322170e-01 -1.72132170e+00 -5.19114770e-02
-7.80959502e-02 8.89419764e-02 4.07217294e-02 1.11492562e+00
5.08564651e-01 7.79942393e-01 4.65942442e-01 -1.07357252e+00
-9.06722695e-02 -9.22206402e-01 -3.00438076e-01 4.90991503e-01
3.27554107e-01 -1.01193452e+00 -7.27904439e-02 5.38660884e-01] | [8.660748481750488, 0.42408740520477295] |
456d2fee-ec23-49ee-b9bf-8d60b7c01e08 | translations-as-additional-contexts-for | 1806.05516 | null | http://arxiv.org/abs/1806.05516v1 | http://arxiv.org/pdf/1806.05516v1.pdf | Translations as Additional Contexts for Sentence Classification | In sentence classification tasks, additional contexts, such as the
neighboring sentences, may improve the accuracy of the classifier. However,
such contexts are domain-dependent and thus cannot be used for another
classification task with an inappropriate domain. In contrast, we propose the
use of translated sentences as context that is always available regardless of
the domain. We find that naive feature expansion of translations gains only
marginal improvements and may decrease the performance of the classifier, due
to possible inaccurate translations thus producing noisy sentence vectors. To
this end, we present multiple context fixing attachment (MCFA), a series of
modules attached to multiple sentence vectors to fix the noise in the vectors
using the other sentence vectors as context. We show that our method performs
competitively compared to previous models, achieving best classification
performance on multiple data sets. We are the first to use translations as
domain-free contexts for sentence classification. | ['Seung-won Hwang', 'Kyungjae Lee', 'Jinyeong Yeo', 'Reinald Kim Amplayo'] | 2018-06-14 | null | null | null | null | ['subjectivity-analysis'] | ['natural-language-processing'] | [ 4.46582615e-01 -1.07777961e-01 -1.48372516e-01 -7.68770218e-01
-9.59773481e-01 -7.02492476e-01 5.13622224e-01 2.96284378e-01
-5.84938109e-01 1.04788971e+00 3.12682509e-01 -5.06005704e-01
2.44985193e-01 -5.06965160e-01 -3.80869597e-01 -5.61515093e-01
4.24791932e-01 1.51372656e-01 4.97475117e-01 -5.63612044e-01
4.35412109e-01 1.54967561e-01 -1.41468799e+00 7.60849416e-01
6.92219734e-01 5.73791265e-01 1.88473761e-01 3.90353978e-01
-4.05404270e-01 5.64535379e-01 -9.97787535e-01 -4.92356509e-01
-1.65003613e-02 -4.66547132e-01 -8.80790949e-01 -5.22545241e-02
4.63981986e-01 4.01073359e-02 1.76352322e-01 9.04682755e-01
3.24329585e-01 9.25350934e-02 6.09001160e-01 -8.14020157e-01
-6.69599116e-01 8.84888589e-01 -2.31429189e-01 3.73747528e-01
6.44943953e-01 -4.24772710e-01 1.07555246e+00 -1.32500041e+00
8.63922536e-01 1.01061130e+00 8.04511845e-01 7.22493708e-01
-1.33144677e+00 -3.05762410e-01 4.98286426e-01 2.34553292e-01
-1.27130556e+00 -5.82092643e-01 9.96775031e-01 -1.92908511e-01
1.54096615e+00 5.91571689e-01 3.87665302e-01 1.58902979e+00
4.56858337e-01 5.77636302e-01 1.00407481e+00 -9.11611199e-01
3.82217139e-01 3.67480576e-01 5.19422054e-01 1.04332663e-01
1.79041624e-02 -4.90194499e-01 -4.30317909e-01 -3.86493325e-01
7.39607364e-02 -1.15777843e-01 -1.73363164e-01 -1.10266201e-01
-9.18268204e-01 1.03727090e+00 2.81857789e-01 7.38944948e-01
-2.56564826e-01 -2.07374915e-01 4.14824605e-01 6.38441563e-01
8.00175846e-01 5.21273375e-01 -8.18606794e-01 -1.58738330e-01
-8.34380746e-01 2.35332534e-01 5.82665861e-01 8.77349794e-01
5.18728256e-01 -3.17124397e-01 -3.21660727e-01 1.17946005e+00
-7.80612528e-02 1.86367884e-01 6.33955777e-01 -5.34398556e-01
8.01666260e-01 5.83791316e-01 1.54162701e-02 -8.59359026e-01
-5.06796420e-01 -5.94304144e-01 -4.15355861e-01 -2.96559304e-01
2.40916431e-01 -1.47044286e-01 -6.97765410e-01 1.93802679e+00
9.38091725e-02 -1.15770310e-01 3.01473230e-01 7.43171275e-01
6.28127813e-01 3.87208253e-01 9.79353637e-02 -4.34685230e-01
1.23503518e+00 -8.35331142e-01 -7.26940572e-01 -6.62706196e-01
1.02032387e+00 -1.00283039e+00 1.33750582e+00 1.06772967e-01
-7.53979087e-01 -3.59159708e-01 -1.09128869e+00 2.63966192e-02
-5.31099319e-01 -1.36367315e-02 5.75646818e-01 6.17196798e-01
-8.58764589e-01 4.48577017e-01 -6.42461002e-01 -5.46739638e-01
3.42364311e-02 3.90479445e-01 -4.98713583e-01 -8.99652317e-02
-1.40533864e+00 1.19818962e+00 2.43278202e-02 -1.34063616e-01
1.63467199e-01 -2.41937250e-01 -8.76231253e-01 -6.46743923e-02
1.31369894e-02 -6.43648386e-01 1.19540155e+00 -1.31266797e+00
-1.05401468e+00 6.38251245e-01 -4.69625980e-01 -2.91861922e-01
2.27128014e-01 -3.01211257e-03 -4.27136868e-01 -7.79680237e-02
2.45098442e-01 4.91198361e-01 7.13958919e-01 -9.60195780e-01
-4.46725577e-01 -3.25718910e-01 5.51230907e-02 2.73358852e-01
-6.45673454e-01 2.13369414e-01 -1.18558720e-01 -7.06465006e-01
2.76460290e-01 -8.93021643e-01 -2.94973582e-01 -5.05482912e-01
-1.81314126e-01 -2.50409335e-01 7.45419562e-01 -4.08063531e-01
1.52584791e+00 -2.14678359e+00 1.38186395e-01 3.03510968e-02
-2.92628795e-01 1.63737684e-01 -3.31145167e-01 5.31693101e-01
-1.30883664e-01 3.49334240e-01 -4.62790847e-01 -5.65564096e-01
-3.82023633e-01 5.34493387e-01 -1.28508016e-01 1.61506370e-01
5.45282841e-01 6.75018668e-01 -8.58422041e-01 -5.45528054e-01
-9.73256305e-02 4.05437678e-01 -5.03844380e-01 -3.37380856e-01
-5.72074531e-03 1.51756838e-01 -3.40731382e-01 2.91848749e-01
6.49337709e-01 -1.35985473e-02 4.93295372e-01 2.84822673e-01
1.26909748e-01 8.40319574e-01 -9.67337191e-01 1.62499416e+00
-6.16915822e-01 5.13356209e-01 -6.37369826e-02 -8.79844487e-01
9.35220182e-01 2.68014580e-01 1.30315632e-01 -4.99365300e-01
3.82844247e-02 3.03659230e-01 1.33869186e-01 -4.20581132e-01
6.17998719e-01 -2.72895932e-01 -3.18523526e-01 1.99799255e-01
-4.14056107e-02 7.31900930e-02 2.17012927e-01 1.53145567e-01
1.22062874e+00 -1.03951141e-01 4.13290203e-01 -2.25625381e-01
5.80224693e-01 1.72307005e-03 6.74931884e-01 6.97472811e-01
-1.62454784e-01 7.39261031e-01 3.36207032e-01 -3.24931026e-01
-9.65766370e-01 -7.90856302e-01 -6.02503538e-01 1.13439989e+00
-1.04847141e-01 -6.71249866e-01 -6.87414885e-01 -1.05143631e+00
-7.11142793e-02 9.37444866e-01 -5.52600563e-01 -1.80022836e-01
-7.43949711e-01 -9.01309609e-01 2.73682982e-01 6.50283575e-01
-1.02794431e-01 -8.38756800e-01 -1.70966983e-01 4.32988852e-01
-5.06502569e-01 -9.55670774e-01 -4.51129943e-01 7.12887228e-01
-8.90350163e-01 -6.97147369e-01 -3.09164345e-01 -9.85512733e-01
7.18733966e-01 3.37665737e-01 1.14446151e+00 2.64790088e-01
4.17352319e-01 -4.81897034e-02 -7.94857979e-01 -1.91151723e-01
-6.19471490e-01 4.90240872e-01 1.59864411e-01 -2.32399449e-01
8.35845053e-01 -5.43451309e-01 -2.08283812e-01 2.06028357e-01
-8.35847080e-01 -3.19899678e-01 2.94026554e-01 1.34215891e+00
2.99868286e-01 -2.47584954e-01 9.76792276e-01 -1.16066730e+00
9.51342225e-01 -5.06092846e-01 1.07325623e-02 1.91572487e-01
-5.45846939e-01 1.32026048e-02 8.57564330e-01 -5.82725048e-01
-7.78023958e-01 1.71844006e-01 -4.79915172e-01 -1.99679807e-02
-1.84324637e-01 5.87036371e-01 -1.08372241e-01 3.06037720e-02
1.04494655e+00 4.85231318e-02 -1.50457770e-01 -5.69897890e-01
-7.74287879e-02 1.06719387e+00 -5.32795824e-02 -2.70076364e-01
3.77791375e-01 -1.37665212e-01 -1.69220343e-01 -7.70622790e-01
-8.17579091e-01 -4.71816152e-01 -8.34666252e-01 9.99474823e-02
4.27581757e-01 -6.87121153e-01 2.27518782e-01 -1.09764047e-01
-1.32691622e+00 1.94524273e-01 -3.30604352e-02 4.11554992e-01
-1.47630587e-01 5.17734945e-01 -6.45574391e-01 -7.62677610e-01
-2.04086781e-01 -1.17351770e+00 1.06478047e+00 -1.70741856e-01
-7.17948794e-01 -1.00519860e+00 -1.08398765e-01 2.99769014e-01
4.61628258e-01 -1.30544290e-01 8.87297690e-01 -1.05375051e+00
1.09231241e-01 -3.55332971e-01 3.58757049e-01 4.86143231e-01
3.92488927e-01 -1.18302375e-01 -1.23362148e+00 -2.54330546e-01
1.81052193e-01 -8.65428969e-02 9.35531437e-01 9.28004906e-02
7.13541031e-01 -2.66623318e-01 -4.46720719e-01 3.76139134e-02
1.17992568e+00 1.34928212e-01 4.75252748e-01 3.96927834e-01
2.76737958e-01 5.87624073e-01 7.03775346e-01 1.27438799e-01
1.06168531e-01 8.25019777e-01 1.96872186e-02 1.01122029e-01
-1.50298262e-02 -1.73726901e-01 4.40990895e-01 9.65167761e-01
3.73145461e-01 -3.68017703e-01 -8.31805766e-01 5.39899945e-01
-1.75540590e+00 -1.00917792e+00 -3.52710426e-01 2.06491995e+00
9.64352250e-01 4.62062746e-01 -8.37433562e-02 3.30885291e-01
5.94731271e-01 -4.23168503e-02 -1.46920979e-01 -7.56456017e-01
-3.65640670e-01 2.17446417e-01 2.47139797e-01 7.86629915e-01
-1.20526695e+00 1.04500186e+00 7.17692852e+00 7.15830624e-01
-1.11589658e+00 2.73966908e-01 3.81923079e-01 -1.48354471e-01
-6.30680323e-01 1.92830175e-01 -8.77333462e-01 5.98210931e-01
9.56595838e-01 2.64521837e-01 1.47757739e-01 9.13681388e-01
8.63332152e-02 -6.88443631e-02 -1.16650236e+00 6.10805094e-01
2.37613007e-01 -1.14646649e+00 1.86338257e-02 -3.09314787e-01
4.93010014e-01 -2.07286533e-02 -1.55633435e-01 4.11542803e-01
-1.51135102e-01 -5.26798844e-01 6.05876386e-01 1.44323930e-01
4.52912241e-01 -6.33862853e-01 1.20273173e+00 5.17312586e-01
-6.67247772e-01 -1.63901746e-01 -5.24734497e-01 -4.20518994e-01
-2.98205912e-02 6.28511369e-01 -9.74456787e-01 4.44550604e-01
4.27256495e-01 5.69592178e-01 -9.33869600e-01 5.35607934e-01
-1.38800845e-01 6.70338392e-01 -3.47315431e-01 -4.81012940e-01
8.35821927e-02 1.39924005e-01 6.11590445e-01 1.46657002e+00
8.22336599e-02 -1.92825377e-01 1.80203557e-01 2.85957187e-01
5.71123837e-03 4.90092844e-01 -8.73596013e-01 4.17275041e-01
6.25346959e-01 9.57807004e-01 -6.49338007e-01 -3.03630024e-01
-7.23120928e-01 1.07191777e+00 5.38829505e-01 3.02837491e-01
-5.02169907e-01 -4.54718679e-01 7.30828047e-01 -4.06744145e-02
2.19075292e-01 -1.10435434e-01 -6.07266963e-01 -1.53641558e+00
5.95020056e-01 -7.27614939e-01 3.43565345e-01 -5.22176385e-01
-1.69895971e+00 8.01007509e-01 -1.12999320e-01 -1.37564313e+00
-4.26125735e-01 -5.38829625e-01 -3.72842193e-01 9.20408070e-01
-1.42713976e+00 -1.04558909e+00 2.77315825e-01 2.98607081e-01
6.89748824e-01 -1.62174329e-01 1.20174122e+00 2.52331853e-01
-4.55061406e-01 8.92129958e-01 1.85264528e-01 1.80173874e-01
1.03031003e+00 -1.14674759e+00 3.61161888e-01 9.66175675e-01
3.59828591e-01 1.04450524e+00 8.55301917e-01 -6.42694056e-01
-9.60970640e-01 -8.85525107e-01 1.69622302e+00 -8.34892809e-01
6.11466169e-01 -8.26238036e-01 -1.07636952e+00 5.74991465e-01
2.11455256e-01 -2.15610981e-01 1.07201910e+00 5.71982324e-01
-4.18246448e-01 -6.15294725e-02 -1.18335676e+00 5.96777380e-01
1.04564750e+00 -5.76982796e-01 -1.00345647e+00 4.58371609e-01
8.84705186e-01 -1.65939510e-01 -6.02294505e-01 1.91374674e-01
2.60849357e-01 -6.62179768e-01 8.06678534e-01 -8.99691403e-01
3.88778687e-01 -1.94170862e-01 -6.95540428e-01 -1.45233512e+00
-5.13841867e-01 4.67838906e-02 2.59776384e-01 1.38031578e+00
1.06498790e+00 -7.48398364e-01 5.87471068e-01 6.11506760e-01
-2.90429294e-01 -8.58383358e-01 -1.21011937e+00 -8.19840610e-01
5.10461450e-01 -6.92003787e-01 6.23212278e-01 1.11619294e+00
4.99872595e-01 9.04694796e-01 -1.03918724e-01 -5.56670763e-02
-4.00812477e-02 1.56889111e-01 2.62299120e-01 -1.04604483e+00
-3.15704167e-01 -2.87736118e-01 -5.23094893e-01 -8.92191947e-01
4.06967938e-01 -1.12206423e+00 -7.54638296e-03 -1.41743803e+00
1.34721458e-01 -3.98297817e-01 -2.74242491e-01 7.55954742e-01
-4.05927956e-01 2.89734811e-01 1.36762485e-01 1.53177276e-01
-3.59585643e-01 4.13431615e-01 9.51601744e-01 -2.21457526e-01
-1.95956647e-01 3.93583141e-02 -9.56471205e-01 5.15808821e-01
8.98121834e-01 -6.93972170e-01 -3.08166802e-01 -5.69681585e-01
2.04446375e-01 -2.77290821e-01 -1.02417491e-01 -7.53885329e-01
3.44425589e-02 -2.14005515e-01 7.05320537e-01 -4.02481049e-01
4.65687275e-01 -8.88350904e-01 -2.00658619e-01 1.99979410e-01
-5.35709083e-01 2.59988248e-01 7.18915015e-02 4.23526317e-01
-3.60905647e-01 -5.75776815e-01 5.19417226e-01 -8.18480998e-02
-3.34687620e-01 -4.60608929e-01 -7.35190332e-01 -1.27681553e-01
7.39522219e-01 -1.14191405e-01 -3.02758932e-01 -1.31108001e-01
-7.25688815e-01 -3.78954038e-02 6.46600187e-01 6.43815756e-01
5.75669825e-01 -1.31633270e+00 -6.10893965e-01 3.43245506e-01
3.58238250e-01 -4.14441794e-01 -5.49451858e-02 5.91602027e-01
-3.92606594e-02 5.12356102e-01 2.54931629e-01 -6.52779043e-01
-1.54988205e+00 6.46283686e-01 1.13760620e-01 -9.74966213e-02
-3.83186996e-01 6.90804303e-01 -4.35987383e-01 -6.30471945e-01
3.75432521e-02 -5.10694921e-01 -3.72059643e-01 2.00595513e-01
4.80191916e-01 -2.91424304e-01 7.51030505e-01 -6.41125858e-01
-8.59326780e-01 4.37593907e-01 -2.62040198e-01 -1.12971351e-01
1.19951749e+00 -2.76047260e-01 -4.09415457e-03 6.77040815e-01
1.26261759e+00 1.65620387e-01 -5.80709398e-01 -6.73808306e-02
3.06069344e-01 -4.84410644e-01 -1.45389035e-01 -8.45326662e-01
-5.03630579e-01 6.78148270e-01 2.88591504e-01 2.97986120e-01
1.26529336e+00 -7.70220459e-02 5.60487926e-01 5.72717607e-01
5.72373211e-01 -1.06718779e+00 -3.32093745e-01 7.64046729e-01
7.46307731e-01 -1.33322859e+00 -1.05570719e-01 -5.76073706e-01
-7.35194802e-01 1.14504385e+00 5.21410823e-01 1.45831913e-01
5.12341261e-01 3.52005571e-01 4.03303921e-01 3.22048396e-01
-1.13323796e+00 -1.54224932e-02 3.33741456e-02 6.48187697e-01
8.23503137e-01 5.26001900e-02 -9.39938009e-01 8.25294495e-01
-2.59901077e-01 -3.22521716e-01 5.68347692e-01 1.32380748e+00
-3.68049800e-01 -1.80603588e+00 -3.62535149e-01 6.12855077e-01
-6.22565269e-01 -3.57015699e-01 -8.60398531e-01 5.08722544e-01
7.29430020e-02 1.30910385e+00 -1.24518760e-02 -5.63352644e-01
3.95654321e-01 7.75920928e-01 3.12201917e-01 -9.20420229e-01
-1.04515409e+00 8.61317618e-04 5.90458274e-01 -1.36298046e-01
-2.71727830e-01 -9.45793688e-01 -9.78470027e-01 -1.37839801e-04
-5.44452429e-01 3.83148700e-01 6.68603659e-01 1.04270172e+00
5.45236051e-01 2.24294856e-01 7.90828526e-01 -4.07630414e-01
-7.86782503e-01 -1.26294935e+00 -1.76513046e-01 5.63525438e-01
4.14673030e-01 -6.87589765e-01 -4.80879605e-01 -1.01522058e-01] | [10.95880126953125, 9.070988655090332] |
8baf7bbd-5272-4f24-b379-425d3f9a77f6 | perturbation-analysis-of-randomized-svd-and | 2203.10262 | null | https://arxiv.org/abs/2203.10262v2 | https://arxiv.org/pdf/2203.10262v2.pdf | Perturbation Analysis of Randomized SVD and its Applications to High-dimensional Statistics | Randomized singular value decomposition (RSVD) is a class of computationally efficient algorithms for computing the truncated SVD of large data matrices. Given a $n \times n$ symmetric matrix $\mathbf{M}$, the prototypical RSVD algorithm outputs an approximation of the $k$ leading singular vectors of $\mathbf{M}$ by computing the SVD of $\mathbf{M}^{g} \mathbf{G}$; here $g \geq 1$ is an integer and $\mathbf{G} \in \mathbb{R}^{n \times k}$ is a random Gaussian sketching matrix. In this paper we study the statistical properties of RSVD under a general "signal-plus-noise" framework, i.e., the observed matrix $\hat{\mathbf{M}}$ is assumed to be an additive perturbation of some true but unknown signal matrix $\mathbf{M}$. We first derive upper bounds for the $\ell_2$ (spectral norm) and $\ell_{2\to\infty}$ (maximum row-wise $\ell_2$ norm) distances between the approximate singular vectors of $\hat{\mathbf{M}}$ and the true singular vectors of the signal matrix $\mathbf{M}$. These upper bounds depend on the signal-to-noise ratio (SNR) and the number of power iterations $g$. A phase transition phenomenon is observed in which a smaller SNR requires larger values of $g$ to guarantee convergence of the $\ell_2$ and $\ell_{2\to\infty}$ distances. We also show that the thresholds for $g$ where these phase transitions occur are sharp whenever the noise matrices satisfy a certain trace growth condition. Finally, we derive normal approximations for the row-wise fluctuations of the approximate singular vectors and the entrywise fluctuations of the approximate matrix. We illustrate our theoretical results by deriving nearly-optimal performance guarantees for RSVD when applied to three statistical inference problems, namely, community detection, matrix completion, and principal component analysis with missing data. | ['Minh Tang', 'Yichi Zhang'] | 2022-03-19 | null | null | null | null | ['matrix-completion'] | ['methodology'] | [ 2.72451222e-01 -9.30652302e-03 3.77075762e-01 6.89472705e-02
-9.29916024e-01 -6.04301512e-01 -2.37568513e-01 -1.03570186e-02
-2.99384445e-01 5.50981462e-01 -1.31444007e-01 -6.27528667e-01
-8.88292134e-01 -7.38162041e-01 -7.31987774e-01 -1.11489832e+00
-1.00704288e+00 -6.48904294e-02 -3.31914663e-01 -3.15776616e-01
-1.54806273e-02 3.42623413e-01 -1.33234274e+00 -2.31044114e-01
5.66372395e-01 1.40849757e+00 -5.50119169e-02 8.89903009e-01
1.26234084e-01 4.78026599e-01 -5.67698300e-01 -4.31608826e-01
7.37973094e-01 -7.13258862e-01 -2.74744213e-01 1.08821951e-01
-9.62313451e-03 3.52683775e-02 -7.03150690e-01 1.82084084e+00
4.88879621e-01 2.31535375e-01 7.21979320e-01 -1.09531212e+00
-3.97112489e-01 7.53411889e-01 -1.05575168e+00 2.69666225e-01
3.95825565e-01 5.27650453e-02 8.79219353e-01 -1.07657146e+00
2.94999212e-01 8.44829202e-01 8.53817225e-01 -6.55867625e-03
-1.39237094e+00 -1.01652014e+00 -1.56364322e-01 -1.84689954e-01
-2.09182906e+00 -5.25575936e-01 6.57481670e-01 -7.14109182e-01
5.16198814e-01 6.52753472e-01 1.69156268e-01 2.54630834e-01
-2.42676616e-01 2.63222724e-01 9.56337869e-01 -5.36636770e-01
3.20175022e-01 -1.21427916e-01 1.86711505e-01 9.06977117e-01
7.15581775e-01 -8.82519856e-02 -3.27554584e-01 -4.70192313e-01
8.00789416e-01 -9.64007080e-02 -6.11830950e-01 -3.21183428e-02
-1.05090642e+00 7.57868946e-01 -1.75584793e-01 2.97650725e-01
-5.15013218e-01 3.50633919e-01 7.51749277e-02 4.23342854e-01
3.61719057e-02 4.32449989e-02 -1.20923527e-01 -2.84509975e-02
-1.01599026e+00 1.30975753e-01 7.34972954e-01 1.09068263e+00
7.42070556e-01 3.08344871e-01 -4.06494066e-02 8.47199380e-01
1.58818975e-01 1.23804224e+00 -3.61903220e-01 -1.35968029e+00
8.74735653e-01 2.66928762e-01 3.37950706e-01 -1.18156397e+00
-2.63839722e-01 -5.57955384e-01 -1.40087759e+00 -8.73789787e-02
8.74851286e-01 -3.18001568e-01 -2.89417744e-01 2.00744843e+00
-3.20226736e-02 1.17223755e-01 2.55810749e-02 7.68520176e-01
3.11047137e-01 7.01458573e-01 -5.06310701e-01 -9.75049615e-01
1.12071443e+00 2.91683435e-01 -4.71183389e-01 -1.83851898e-01
2.32753545e-01 -9.12709236e-01 8.43502760e-01 4.86817211e-01
-1.34676635e+00 -1.54963404e-01 -1.11572826e+00 7.03826070e-01
3.87012571e-01 1.68321002e-02 7.52930790e-02 9.54610705e-01
-8.26090276e-01 4.88506824e-01 -5.16495705e-01 2.40958631e-01
7.48244822e-02 4.10385609e-01 -2.98317313e-01 -3.31486613e-01
-9.39690650e-01 -4.82178926e-02 -2.68643588e-01 4.73687112e-01
-5.48597515e-01 -4.95130658e-01 -5.36979735e-01 1.00227557e-01
3.98219168e-01 -9.46859121e-02 5.41128874e-01 -3.24825585e-01
-6.17545664e-01 7.55799532e-01 -5.55486530e-02 -1.70193702e-01
1.76101178e-01 4.13872033e-01 -5.11457801e-01 3.31273347e-01
3.06618035e-01 -5.29103637e-01 9.24788713e-01 -1.17558122e+00
-2.75344014e-01 -7.71682918e-01 -3.14827681e-01 -1.72277376e-01
-2.47497782e-01 1.01680115e-01 -3.37946564e-01 -8.11905563e-01
8.85378599e-01 -8.74864995e-01 -4.58544403e-01 -4.59867150e-01
-4.57240909e-01 3.95047516e-01 1.13274075e-01 -6.96380436e-01
1.70757580e+00 -2.48378873e+00 2.36850426e-01 1.14190745e+00
4.42229539e-01 -1.09205462e-01 3.82438526e-02 6.67726636e-01
-1.79990202e-01 -4.13453393e-02 -3.90331477e-01 1.05210081e-01
-1.29720597e-02 -3.50878090e-02 -1.59666240e-01 1.01504338e+00
-7.07886040e-01 3.54342647e-02 -7.83702910e-01 9.54165235e-02
-1.76594764e-01 3.03487539e-01 -4.49573725e-01 -6.67922273e-02
1.25388935e-01 5.69736175e-02 -3.55698138e-01 4.59260017e-01
9.94678259e-01 -3.31945449e-01 4.01914537e-01 -1.60583287e-01
-1.03070617e-01 -4.33756679e-01 -2.13656425e+00 1.01576269e+00
1.94413453e-01 3.05336356e-01 1.01404226e+00 -1.45980203e+00
7.44163096e-01 2.52410114e-01 9.21589792e-01 -4.27547425e-01
3.75449181e-01 3.32008392e-01 -7.15470910e-02 -2.49578729e-01
1.97689697e-01 -3.79336447e-01 -2.50446647e-01 5.50952911e-01
-2.18039379e-01 -7.70418197e-02 4.33007807e-01 4.54630584e-01
1.45033336e+00 -6.49430454e-01 2.15771496e-01 -5.77317894e-01
5.42102635e-01 -4.86709774e-01 6.91501617e-01 1.02404487e+00
-2.25587696e-01 3.66999239e-01 8.54934156e-01 1.71114311e-01
-9.78877962e-01 -1.22782028e+00 -1.22567765e-01 8.67067814e-01
1.33592367e-01 -3.97811979e-01 -9.04659688e-01 3.21959443e-02
-3.29236984e-02 6.06262326e-01 -3.07971865e-01 -1.98909305e-02
-3.78102213e-01 -1.20973885e+00 7.51404464e-01 1.85039937e-01
4.57975119e-01 -2.62435287e-01 -1.44870698e-01 2.12479919e-01
-3.19755942e-01 -1.19100809e+00 -6.85831487e-01 2.00669721e-01
-6.67561114e-01 -1.11378837e+00 -5.18988550e-01 -3.63497168e-01
9.34280574e-01 4.89294082e-01 5.49462378e-01 -7.31871799e-02
-2.88627714e-01 6.76835895e-01 -1.32168636e-01 -2.05259025e-01
-2.50311196e-01 -7.01932132e-01 5.57951629e-01 3.44686776e-01
-2.89688609e-03 -9.60593998e-01 -7.84934163e-01 4.29010123e-01
-1.01538539e+00 -3.78981441e-01 3.10764432e-01 8.10188115e-01
7.73146927e-01 4.96035129e-01 1.06056325e-01 -2.38894030e-01
7.67000794e-01 -3.55375618e-01 -7.67622471e-01 -7.97450636e-03
-5.32343388e-01 6.91164136e-02 7.82599330e-01 -2.55953908e-01
-2.83723325e-01 -4.67729196e-02 4.95499410e-02 -6.58474922e-01
4.53182936e-01 7.31673419e-01 -8.91394764e-02 -7.16046570e-03
9.80493546e-01 6.17663741e-01 7.79992761e-03 -5.00247478e-01
3.47589433e-01 6.01954579e-01 7.21445024e-01 -7.35720754e-01
9.38225269e-01 5.41733563e-01 4.69717979e-01 -1.21630418e+00
-5.18523514e-01 -4.83688444e-01 -6.97402135e-02 -1.25888914e-01
2.37255454e-01 -8.37712407e-01 -1.23686647e+00 1.83446199e-01
-4.80173916e-01 -2.19365768e-02 -4.03920323e-01 7.24106491e-01
-6.25727355e-01 6.59251034e-01 -5.44595897e-01 -1.31001949e+00
-2.93671012e-01 -1.01467025e+00 5.43039858e-01 -2.87562579e-01
-6.64976537e-02 -3.05514574e-01 -2.29704261e-01 2.56307065e-01
1.68315783e-01 2.30032846e-01 9.96882081e-01 -2.67812312e-01
-3.18683803e-01 -6.20000303e-01 -4.01703835e-01 8.02472889e-01
-6.72691166e-02 -3.81387711e-01 -4.18523341e-01 -7.00322926e-01
4.78737921e-01 4.83343810e-01 3.23580950e-01 5.84782839e-01
9.16134775e-01 -8.68819475e-01 -2.23748803e-01 4.77925211e-01
1.40174127e+00 2.00901374e-01 5.25987387e-01 -3.10574055e-01
2.21504390e-01 1.72911882e-01 1.50271848e-01 1.13669658e+00
-1.28650203e-01 3.00955683e-01 2.38807321e-01 4.60374326e-01
2.46498898e-01 2.18205705e-01 5.17498910e-01 1.04254007e+00
-1.46496788e-01 3.02146357e-02 -8.31116080e-01 4.86522228e-01
-1.44821680e+00 -1.20376658e+00 -2.68432885e-01 2.81263494e+00
6.37986302e-01 7.70665631e-02 1.30239815e-01 5.63866436e-01
8.73056233e-01 2.45479848e-02 -3.38534236e-01 9.38583352e-03
-3.10488522e-01 6.74462974e-01 9.05155718e-01 6.27390087e-01
-7.26529598e-01 1.01771601e-01 4.35947800e+00 1.03337491e+00
-7.65931010e-01 -8.06959346e-03 3.61775845e-01 -3.83720368e-01
-2.69189298e-01 -6.71428815e-02 -4.42837059e-01 6.63353503e-01
8.87734294e-01 -2.53933102e-01 8.34749758e-01 6.39482260e-01
3.60589445e-01 -2.44895548e-01 -7.56153584e-01 1.40341318e+00
-1.49649784e-01 -1.18576205e+00 -3.96512300e-01 3.77775878e-01
6.51381373e-01 -3.15944701e-01 2.25966662e-01 -3.54778655e-02
4.96171027e-01 -8.60619605e-01 5.76547682e-01 3.49288493e-01
1.27190280e+00 -7.76663780e-01 3.09783995e-01 4.52460974e-01
-1.58880782e+00 -4.50936347e-01 -2.74561226e-01 -9.41404551e-02
7.76503012e-02 1.13911796e+00 -3.80374610e-01 5.47885120e-01
8.73537838e-01 -1.37429923e-01 1.70047998e-01 6.67410076e-01
1.51543751e-01 8.13108325e-01 -6.87199295e-01 6.35028332e-02
1.22255504e-01 -7.91923702e-01 1.02308118e+00 9.87878919e-01
9.06881392e-01 9.34633911e-01 2.83585668e-01 5.54571569e-01
-2.46056974e-01 3.64432871e-01 -4.87376153e-01 -1.99165732e-01
7.32368052e-01 6.81275427e-01 -7.81940937e-01 -1.06099635e-01
-3.01582038e-01 7.71392882e-01 -1.74765438e-01 6.67777777e-01
-5.02095163e-01 -8.92104149e-01 1.03429759e+00 4.34788525e-01
3.98937196e-01 -6.42609000e-01 -4.60765809e-01 -9.83854234e-01
3.30847234e-01 -1.05410731e+00 5.72267234e-01 -3.39546978e-01
-1.18016410e+00 2.58909971e-01 -9.15390924e-02 -1.25531185e+00
-1.15188017e-01 -3.57140362e-01 8.34679604e-02 9.48587716e-01
-3.45193356e-01 -2.73597926e-01 1.41622543e-01 9.24713552e-01
-3.95194888e-01 -3.02903712e-01 6.89769149e-01 3.92468452e-01
-5.55454731e-01 7.04631746e-01 1.00619018e+00 5.02236128e-01
-6.68458119e-02 -5.91067135e-01 -1.95360035e-01 1.21403420e+00
-1.68632879e-03 9.41268802e-01 1.14259291e+00 -5.07133305e-01
-1.96947265e+00 -6.67770326e-01 4.60677266e-01 2.03913692e-02
6.12921417e-01 -4.64479774e-01 -4.76934016e-01 6.37754321e-01
-2.78063834e-01 2.25095406e-01 8.09445798e-01 -2.18610257e-01
-4.21105295e-01 -3.17280352e-01 -1.43441904e+00 6.81728601e-01
1.05906522e+00 -5.64695418e-01 9.63010564e-02 2.65523642e-01
1.98805720e-01 -3.44764799e-01 -1.03697526e+00 4.95481968e-01
3.13582271e-01 -1.02503455e+00 1.18395865e+00 -4.06018913e-01
-2.63838232e-01 -5.78622222e-01 -1.04730690e+00 -6.79019690e-01
-3.54147315e-01 -1.23051417e+00 -1.35998353e-01 7.05560803e-01
2.41978854e-01 -4.40794080e-01 7.81236887e-01 5.72898626e-01
1.86485171e-01 -4.29130197e-01 -1.51597738e+00 -9.25122321e-01
-5.39326370e-02 -8.67956817e-01 1.80835187e-01 7.93928146e-01
2.69350350e-01 -8.35007709e-03 -4.07503933e-01 5.49936056e-01
1.15833521e+00 -9.74951535e-02 5.02330720e-01 -1.03409421e+00
-5.98966300e-01 -4.16680545e-01 -3.41108739e-01 -9.89901662e-01
-4.05880868e-01 -9.14969504e-01 -1.68561861e-01 -1.01334202e+00
2.05212802e-01 -6.96708977e-01 -3.81465316e-01 1.15257807e-01
2.16818973e-01 1.44700557e-01 1.90746069e-01 7.34883547e-02
-2.87099510e-01 1.83794215e-01 8.05537879e-01 -5.21047227e-02
-3.18837106e-01 9.75733921e-02 -8.96885514e-01 6.72851205e-01
2.59550601e-01 -3.90351653e-01 -3.11589152e-01 -3.78333479e-02
7.55993068e-01 8.20226431e-01 5.69128655e-02 -1.04067791e+00
1.87008813e-01 -1.25904575e-01 7.42071569e-02 -3.43744189e-01
3.88925463e-01 -6.30202174e-01 5.57067156e-01 3.00128877e-01
3.85107123e-03 -2.41658762e-02 -6.17294721e-02 6.97688937e-01
2.24074692e-01 -2.29529887e-01 8.22926879e-01 -6.17683716e-02
1.49871662e-01 2.25468263e-01 -5.43503165e-01 1.49877653e-01
9.57680583e-01 -2.42566869e-01 1.61080033e-01 -9.09614265e-01
-7.57238328e-01 -2.14076564e-02 -1.38703687e-02 -3.46776932e-01
5.68464518e-01 -1.22839057e+00 -9.29694951e-01 4.13319767e-01
-1.14263006e-01 -2.18911082e-01 4.90443796e-01 1.04411006e+00
-2.69473046e-01 1.25785276e-01 4.80957925e-01 -5.39192259e-01
-9.84719992e-01 5.95585942e-01 2.63458073e-01 -3.32135148e-02
1.55908242e-02 1.18131351e+00 -2.50344366e-01 -8.19324329e-02
2.86753327e-01 -1.16519555e-01 5.67078829e-01 1.68385226e-02
6.21214390e-01 9.92076516e-01 1.06113829e-01 -8.04864764e-01
-2.76919931e-01 4.82596695e-01 4.11437839e-01 -4.42958564e-01
1.15485632e+00 -4.17772442e-01 -6.36989892e-01 -7.80291995e-03
1.46130359e+00 4.97419983e-01 -8.96194875e-01 -3.68227243e-01
-2.98975378e-01 -6.11821234e-01 -3.85410815e-01 -2.00196385e-01
-1.09242260e+00 6.80515587e-01 7.00314283e-01 3.34637702e-01
1.13804686e+00 2.39517242e-02 4.58757848e-01 3.65549833e-01
7.55160332e-01 -1.24703896e+00 -7.90988877e-02 4.26081151e-01
7.42620945e-01 -4.88812774e-01 1.14184946e-01 -3.65289986e-01
-7.86060765e-02 8.74839425e-01 -1.89116389e-01 -2.13448075e-03
9.41199541e-01 3.16560119e-01 -2.23365262e-01 -2.60339707e-01
6.91578910e-02 -5.96123263e-02 -5.12783602e-02 2.35068008e-01
2.16586903e-01 2.99606919e-01 -5.20766497e-01 1.05202889e+00
-2.83927351e-01 -1.52836278e-01 5.42034209e-01 7.53535748e-01
-5.83559453e-01 -8.86487722e-01 -1.06335568e+00 7.32351482e-01
-5.42841733e-01 -2.18086362e-01 1.94921762e-01 3.60304862e-01
4.36729454e-02 1.21331215e+00 -4.30326462e-01 -4.48097885e-01
4.03695971e-01 1.30003572e-01 6.08594716e-01 -1.20215200e-01
-2.58414447e-02 3.45534950e-01 -5.66288680e-02 -3.71152967e-01
1.28767595e-01 -8.62976611e-01 -1.48700809e+00 -7.72494853e-01
6.27075061e-02 4.58479583e-01 3.52584898e-01 6.14335537e-01
2.53143311e-01 -1.89659074e-01 9.87657487e-01 -3.18854809e-01
-8.97027075e-01 -6.61170185e-01 -1.29588008e+00 4.02056515e-01
3.80826443e-02 -2.30333120e-01 -6.15591049e-01 -6.15782402e-02] | [6.677084445953369, 4.715038776397705] |
5e977496-8aec-4f78-b62b-f7bd8a3a3fa8 | automatic-flare-spot-artifact-detection-and | 2103.04384 | null | https://arxiv.org/abs/2103.04384v1 | https://arxiv.org/pdf/2103.04384v1.pdf | Automatic Flare Spot Artifact Detection and Removal in Photographs | Flare spot is one type of flare artifact caused by a number of conditions, frequently provoked by one or more high-luminance sources within or close to the camera field of view. When light rays coming from a high-luminance source reach the front element of a camera, it can produce intra-reflections within camera elements that emerge at the film plane forming non-image information or flare on the captured image. Even though preventive mechanisms are used, artifacts can appear. In this paper, we propose a robust computational method to automatically detect and remove flare spot artifacts. Our contribution is threefold: firstly, we propose a characterization which is based on intrinsic properties that a flare spot is likely to satisfy; secondly, we define a new confidence measure able to select flare spots among the candidates; and, finally, a method to accurately determine the flare region is given. Then, the detected artifacts are removed by using exemplar-based inpainting. We show that our algorithm achieve top-tier quantitative and qualitative performance. | ['Coloma Ballester', 'Patricia Vitoria'] | 2021-03-07 | null | null | null | null | ['flare-removal'] | ['computer-vision'] | [ 8.87781918e-01 -4.39599782e-01 3.23419034e-01 -8.75370055e-02
-6.69943810e-01 -6.49093747e-01 3.33579302e-01 -5.54818548e-02
7.92188570e-02 7.39247620e-01 9.67253894e-02 4.07245338e-01
-9.38643590e-02 -5.63977361e-01 -8.38585913e-01 -6.41708136e-01
3.22865665e-01 -1.84754163e-01 5.48055887e-01 3.53461891e-01
6.69456780e-01 8.15796793e-01 -1.78409135e+00 2.57727444e-01
7.04014063e-01 9.64697599e-01 4.29615408e-01 6.19793713e-01
9.56556797e-02 9.85415637e-01 -8.73519182e-01 -1.20553225e-01
2.08348051e-01 -9.15385604e-01 -2.49754548e-01 6.51558876e-01
6.61387682e-01 -3.50641668e-01 2.27279842e-01 1.32441199e+00
4.29555401e-02 -7.82610476e-02 6.94313288e-01 -9.64431345e-01
-2.18724385e-01 -2.81744096e-02 -8.38162541e-01 5.19701280e-02
8.90125334e-01 2.69825846e-01 2.67578483e-01 -1.05804360e+00
8.58916640e-01 6.50956094e-01 6.13365591e-01 3.72505456e-01
-1.07050848e+00 -3.42086405e-01 -4.08239245e-01 1.26232877e-01
-1.25534177e+00 -5.48555851e-01 1.21290767e+00 -4.47390527e-01
5.59381545e-01 5.92134833e-01 6.02041364e-01 6.37820899e-01
5.94837606e-01 2.55709648e-01 1.52230632e+00 -7.21002102e-01
3.61990392e-01 1.41971350e-01 1.22457556e-01 6.02240205e-01
3.53194386e-01 1.19764671e-01 -5.85725784e-01 -3.35095346e-01
7.77346194e-01 2.19075549e-02 -9.06986713e-01 -1.53807342e-01
-7.19220817e-01 2.96620399e-01 3.62171903e-02 5.03531754e-01
-3.56114388e-01 9.07889754e-02 -1.71095312e-01 1.18679322e-01
3.02511275e-01 6.86788440e-01 -6.12277165e-02 1.13826822e-02
-1.05023038e+00 -1.02112517e-02 3.72995913e-01 6.44149363e-01
8.49758387e-01 -3.03526789e-01 -2.07615625e-02 5.82529724e-01
1.12307712e-01 7.27054536e-01 -8.58232379e-02 -9.60760057e-01
6.04378320e-02 4.54926908e-01 4.53342706e-01 -1.17981422e+00
-2.07914665e-01 -1.42429173e-02 -5.99123120e-01 8.43122602e-01
3.47714752e-01 -1.83094721e-02 -5.90589643e-01 1.26841557e+00
9.77756828e-02 3.41752917e-01 -2.93236017e-01 1.15981352e+00
4.74086434e-01 7.03999102e-01 -6.31326735e-01 -1.13544571e+00
1.06555700e+00 -6.36965454e-01 -1.34305251e+00 -7.74632469e-02
-8.60810876e-02 -1.23046553e+00 9.70022321e-01 9.72883523e-01
-1.53582275e+00 -4.27333206e-01 -1.00764620e+00 1.83034256e-01
1.96106434e-01 6.28685802e-02 1.72176376e-01 3.57210219e-01
-8.45895231e-01 5.35011172e-01 -3.82177383e-01 -1.42268166e-01
-2.31808387e-02 -1.51097164e-01 -8.68965238e-02 -1.75869346e-01
-5.78666151e-01 9.21683908e-01 -3.73591185e-01 6.57771304e-02
-5.28637290e-01 -5.27393937e-01 -5.55971682e-01 -1.27509832e-01
3.51036608e-01 -4.47959721e-01 8.32648456e-01 -1.38014412e+00
-1.45071352e+00 7.69202590e-01 -4.70990658e-01 -1.90077975e-01
5.59149086e-01 -2.44378626e-01 -5.83055198e-01 5.99885941e-01
-4.37374152e-02 -2.31556773e-01 1.27329695e+00 -1.80513811e+00
-5.20135522e-01 -1.81562394e-01 -3.37908179e-01 2.92020943e-02
1.68609649e-01 4.64535445e-01 -5.67059457e-01 -5.46369791e-01
1.12828635e-01 -5.13725042e-01 -2.75019314e-02 2.36992612e-01
-4.09623712e-01 2.36573175e-01 9.39761221e-01 -3.41060519e-01
1.58509874e+00 -2.34983492e+00 -8.55464414e-02 1.82248354e-01
1.34848699e-01 2.49058932e-01 2.99413890e-01 5.55758595e-01
3.05236913e-02 -1.00740984e-01 -7.08907187e-01 -5.12374006e-03
-5.74965715e-01 -1.26017585e-01 -4.54812318e-01 5.93706906e-01
2.41057366e-01 1.95896149e-01 -7.27409065e-01 -3.70489210e-01
4.15638506e-01 2.55751729e-01 -5.59890196e-02 3.30273956e-01
-1.19496427e-01 3.42826635e-01 -2.91088670e-01 5.75914085e-01
1.03512907e+00 1.63694486e-01 -2.86523461e-01 -3.71472001e-01
-6.22287273e-01 -3.81935984e-01 -1.35800409e+00 1.26666307e+00
-2.45607316e-01 6.22878432e-01 2.07669556e-01 -8.78817514e-02
9.80181158e-01 1.49213031e-01 2.26072803e-01 -4.33939904e-01
1.57345369e-01 4.72347111e-01 -6.69887602e-01 -9.18972194e-01
4.17377770e-01 -1.99122831e-01 2.44058043e-01 3.82858872e-01
-5.74136436e-01 -4.02562141e-01 2.79032320e-01 -1.58396870e-01
1.35357046e+00 -1.09296795e-02 3.03518742e-01 -2.30235875e-01
6.67141855e-01 -6.80263191e-02 4.01185453e-01 7.08576739e-01
-9.30026844e-02 1.02658057e+00 2.07422853e-01 -4.04314458e-01
-6.35307670e-01 -9.39057350e-01 -3.98537844e-01 9.02746096e-02
5.56061506e-01 -2.47590125e-01 -8.65244925e-01 -3.71986389e-01
-2.21621305e-01 6.67024612e-01 -4.13621306e-01 -6.97673783e-02
-4.78311956e-01 -4.35228050e-01 -1.53297395e-01 1.33861616e-01
5.11702776e-01 -1.22875023e+00 -1.12414968e+00 7.61795416e-02
-1.96342468e-01 -8.03212285e-01 -4.63818073e-01 5.33573069e-02
-7.49437034e-01 -1.47037470e+00 -5.50170600e-01 -5.16406059e-01
8.81412208e-01 6.50134742e-01 1.16582537e+00 1.83551759e-01
-6.21376753e-01 3.98271650e-01 -3.86597782e-01 -2.08673537e-01
-4.26013470e-01 -1.10481036e+00 -2.07223073e-01 3.97900820e-01
2.21293628e-01 -4.27568018e-01 -6.92385018e-01 1.54758111e-01
-1.04717600e+00 -2.92881066e-03 2.91845948e-01 3.47706705e-01
8.19737971e-01 4.54185635e-01 1.10090256e-01 -7.07492530e-01
5.02799869e-01 1.09203868e-01 -7.38806069e-01 -7.89627898e-03
-2.11783677e-01 -3.41897935e-01 6.93644345e-01 -2.65714854e-01
-1.20584273e+00 1.93984181e-01 2.88485974e-01 -6.02820992e-01
-4.57274675e-01 8.01956356e-02 -2.13950709e-01 -1.62400067e-01
8.36319029e-01 9.21853557e-02 -4.09123093e-01 -4.12077904e-01
3.21165323e-02 4.71642643e-01 7.22806156e-01 4.68442738e-02
8.72038484e-01 8.82149458e-01 1.82898819e-01 -1.07367194e+00
-9.08334911e-01 -4.58590567e-01 -3.76122385e-01 -8.16670299e-01
9.90826249e-01 -4.90502179e-01 -3.54297459e-01 7.48373687e-01
-1.31768286e+00 -1.50959641e-01 -2.82123685e-01 3.94539207e-01
-5.15151381e-01 4.55976635e-01 -4.16554302e-01 -1.17075849e+00
2.21573506e-02 -8.01035762e-01 9.93502140e-01 4.33353692e-01
-4.55855280e-02 -6.29274189e-01 2.62930363e-01 2.93284684e-01
9.92832556e-02 5.01749337e-01 6.05258405e-01 4.39377487e-01
-8.89957249e-01 -1.30569234e-01 -8.36762637e-02 4.37369436e-01
4.44025844e-01 5.28467953e-01 -1.11556733e+00 -1.29893348e-01
6.51384711e-01 6.03913330e-03 7.28077173e-01 6.27653003e-01
1.03285265e+00 -6.20328858e-02 -3.44298601e-01 6.99126780e-01
2.03236127e+00 4.67403024e-01 9.63314235e-01 2.27347225e-01
2.92693377e-01 5.71692586e-01 7.53151834e-01 4.46039498e-01
-5.70904195e-01 7.28972077e-01 4.05186087e-01 -4.41179544e-01
-3.73074591e-01 7.83707052e-02 3.29771429e-01 5.02216399e-01
-2.70759501e-03 -4.99884963e-01 -3.64533961e-01 3.76752645e-01
-1.53634155e+00 -1.08514297e+00 -1.04378021e+00 2.46561217e+00
6.65520132e-01 8.46299604e-02 -1.15154915e-01 3.71736079e-01
8.93698394e-01 -4.38541323e-02 -2.37012431e-01 -2.64877915e-01
-3.92984658e-01 1.23177685e-01 4.44077492e-01 8.37365091e-01
-7.69370496e-01 5.10884404e-01 6.46982813e+00 3.70645553e-01
-1.22280836e+00 -1.13472842e-01 1.43938407e-01 -8.38712081e-02
-4.47509080e-01 -1.14584766e-01 -4.60929275e-01 6.56095982e-01
3.70092928e-01 -2.92930193e-02 2.64194787e-01 2.45871261e-01
5.43506920e-01 -8.14951122e-01 -8.39246333e-01 1.07497549e+00
6.01519644e-01 -8.75100791e-01 -2.12146014e-01 -2.87884921e-01
7.96518803e-01 -6.16639078e-01 -7.92857707e-02 -8.49811077e-01
-3.10825676e-01 -6.40414357e-01 6.46947980e-01 1.03206682e+00
9.15566921e-01 -8.17796171e-01 3.26350838e-01 1.62547961e-01
-8.67668033e-01 1.99760094e-01 -3.84707510e-01 -1.52837321e-01
4.76156384e-01 1.14121199e+00 -3.41390729e-01 3.83194596e-01
5.38698733e-01 5.00648379e-01 -5.63977838e-01 1.39579511e+00
-5.96565664e-01 5.26717782e-01 -1.88003972e-01 2.58590847e-01
-1.80655792e-01 -4.34312284e-01 8.26961398e-01 1.02208507e+00
5.46469271e-01 3.15993160e-01 -2.41939545e-01 1.06535268e+00
3.07225049e-01 -7.04886392e-02 -8.70507658e-01 4.99891222e-01
1.53286278e-01 1.28923631e+00 -8.03285658e-01 -1.13662280e-01
-5.97479701e-01 1.37913847e+00 -3.54125768e-01 3.69727910e-01
-7.86985338e-01 -4.96520549e-01 2.46876329e-01 4.09418523e-01
-7.11075217e-02 2.44634196e-01 -3.19204718e-01 -1.10501683e+00
3.83929372e-01 -5.51785588e-01 -1.14516214e-01 -1.46102297e+00
-1.15956771e+00 5.65448344e-01 -2.41506398e-01 -1.51027691e+00
2.75516901e-02 -4.93677229e-01 -8.20425391e-01 8.04383576e-01
-1.50782561e+00 -7.29492903e-01 -5.38907170e-01 7.42276371e-01
6.16277516e-01 3.75998437e-01 6.89080536e-01 1.99128658e-01
-3.52545202e-01 -8.56488943e-02 2.16333807e-01 -3.66278380e-01
9.13393795e-01 -1.05210066e+00 -3.07915032e-01 1.46294844e+00
6.87609464e-02 1.47503465e-01 1.18433619e+00 -7.87796915e-01
-1.39799631e+00 -1.02007949e+00 1.06544602e+00 -4.64717388e-01
3.49756688e-01 -2.84028530e-01 -1.13367772e+00 2.86732405e-01
2.81530321e-01 -4.10596579e-02 2.61552513e-01 -5.16452014e-01
1.53537542e-02 -2.13179708e-01 -1.24790132e+00 3.69773477e-01
7.54869223e-01 -2.77463198e-01 -6.86831355e-01 3.69104654e-01
1.42800614e-01 -2.27740988e-01 -3.42275143e-01 4.20469016e-01
3.10388833e-01 -1.72178900e+00 5.69031954e-01 3.13229442e-01
6.45713151e-01 -7.16651201e-01 1.83294147e-01 -1.10361540e+00
-2.76603162e-01 -9.36631560e-01 1.12491280e-01 1.19426179e+00
1.35115072e-01 -2.59738773e-01 4.23317254e-01 1.56701818e-01
-2.79654622e-01 -3.16742957e-01 -6.14065707e-01 -6.15599394e-01
-5.72853863e-01 -2.86655426e-01 1.20437823e-01 7.79747248e-01
-8.82900804e-02 2.13069528e-01 -3.81933153e-01 4.35365945e-01
8.89129341e-01 4.09931451e-01 6.02863193e-01 -1.17778122e+00
-3.13779563e-01 -2.18541190e-01 -9.67766568e-02 -6.72618568e-01
-3.17912459e-01 -2.65191585e-01 4.02681977e-01 -1.49541891e+00
1.79536402e-01 7.19852969e-02 -5.66236265e-02 -8.51655304e-02
-1.37995571e-01 5.86627007e-01 1.69001475e-01 4.26755279e-01
-3.79067004e-01 6.18421696e-02 1.17015028e+00 3.83976489e-01
-3.09736043e-01 -2.27993876e-02 -2.05867022e-01 1.23418319e+00
4.32025790e-01 -3.36108476e-01 -3.68304491e-01 -3.28759462e-01
4.34890896e-01 4.85370010e-01 4.45576251e-01 -1.33479726e+00
4.97563407e-02 -2.54006118e-01 4.79606181e-01 -7.04176605e-01
2.53124297e-01 -1.12628162e+00 5.09446681e-01 3.00854564e-01
1.43181151e-02 -1.04138948e-01 -1.75515860e-02 5.39118707e-01
-2.61437118e-01 -7.57044017e-01 1.30473745e+00 -9.29674134e-02
-2.98492074e-01 -2.96038628e-01 -6.99229062e-01 -2.75539517e-01
1.17337704e+00 -3.22806150e-01 -4.14361596e-01 -2.23371267e-01
-3.48014325e-01 -3.66568923e-01 1.08544695e+00 -1.22535275e-02
8.16724956e-01 -1.12543893e+00 -3.95220429e-01 4.88236785e-01
1.73079893e-01 -3.50565225e-01 2.36437261e-01 8.05579782e-01
-8.58465731e-01 1.29950605e-02 9.00996570e-03 -4.23192203e-01
-1.60193014e+00 7.30100513e-01 2.34099612e-01 1.31704807e-01
-7.91722178e-01 7.76603639e-01 4.05440271e-01 8.06512356e-01
7.14739636e-02 -3.55232567e-01 -1.18665710e-01 -2.17904136e-01
9.77015615e-01 3.79938424e-01 1.15350358e-01 -4.41088021e-01
-2.69027323e-01 1.20870149e+00 4.20613617e-01 -1.19936824e-01
9.59214687e-01 -4.03273344e-01 -2.59003490e-01 7.28765666e-01
8.17047417e-01 6.77842081e-01 -1.32323492e+00 1.77228436e-01
-3.67976010e-01 -9.27390754e-01 2.71727704e-02 -9.68703330e-01
-9.28471684e-01 7.20188856e-01 5.73296666e-01 4.99933958e-01
1.88721716e+00 -6.23636208e-02 6.48870289e-01 -1.72185898e-01
4.18985337e-01 -9.07972217e-01 1.80626169e-01 -1.02215856e-01
1.00647151e+00 -7.05310702e-01 -1.97897591e-02 -7.32063413e-01
-3.10150027e-01 1.24874818e+00 3.48753214e-01 -3.01193565e-01
3.04008216e-01 7.29341030e-01 2.38788307e-01 -4.18158799e-01
-7.28406072e-01 -1.59698263e-01 8.80075172e-02 5.89202404e-01
2.98508137e-01 -4.65165347e-01 -7.40697742e-01 1.43487379e-01
2.73931235e-01 2.13904932e-01 1.07678401e+00 9.47936893e-01
-9.02058482e-01 -6.67472124e-01 -8.91971707e-01 2.35694453e-01
-5.38916469e-01 1.64006516e-01 -7.19721198e-01 4.12028611e-01
3.96648228e-01 1.21924424e+00 5.93014136e-02 1.04564294e-01
5.34836292e-01 -1.35987118e-01 7.53927588e-01 -3.70720595e-01
-5.22130787e-01 6.73463225e-01 -2.23220423e-01 -7.87645459e-01
-7.41607130e-01 -6.42733037e-01 -9.83976305e-01 2.39365026e-01
-5.55903912e-01 7.20591173e-02 3.96309793e-01 6.06910288e-01
1.30640054e-02 3.20013493e-01 9.02751088e-01 -5.29263794e-01
1.80263698e-01 -7.05415428e-01 -1.05385423e+00 6.69772029e-01
5.91990471e-01 -5.37790418e-01 -1.09418428e+00 5.06540835e-01] | [10.645915031433105, -2.7268576622009277] |
e48a7579-2efa-4fcb-8245-261f5aebf75f | prediction-of-the-position-of-external | 2106.01100 | null | https://arxiv.org/abs/2106.01100v6 | https://arxiv.org/pdf/2106.01100v6.pdf | Prediction of the Position of External Markers Using a Recurrent Neural Network Trained With Unbiased Online Recurrent Optimization for Safe Lung Cancer Radiotherapy | During lung radiotherapy, the position of infrared reflective objects on the chest can be recorded to estimate the tumor location. However, radiotherapy systems have a latency inherent to robot control limitations that impedes the radiation delivery precision. Prediction with online learning of recurrent neural networks (RNN) allows for adaptation to non-stationary respiratory signals, but classical methods such as RTRL and truncated BPTT are respectively slow and biased. This study investigates the capabilities of unbiased online recurrent optimization (UORO) to forecast respiratory motion and enhance safety in lung radiotherapy. We used 9 observation records of the 3D position of 3 external markers on the chest and abdomen of healthy individuals breathing during intervals from 73s to 222s. The sampling frequency was 10Hz, and the amplitudes of the recorded trajectories range from 6mm to 40mm in the superior-inferior direction. We forecast the 3D location of each marker simultaneously with a horizon value between 0.1s and 2.0s, using an RNN trained with UORO. We compare its performance with an RNN trained with RTRL, LMS, and offline linear regression. We provide closed-form expressions for quantities involved in the loss gradient calculation in UORO, thereby making its implementation efficient. Training and cross-validation were performed during the first minute of each sequence. On average over the horizon values considered and the 9 sequences, UORO achieves the lowest root-mean-square (RMS) error and maximum error among the compared algorithms. These errors are respectively equal to 1.3mm and 8.8mm, and the prediction time per time step was lower than 2.8ms (Dell Intel core i9-9900K 3.60 GHz). Linear regression has the lowest RMS error for the horizon values 0.1s and 0.2s, followed by LMS for horizon values between 0.3s and 0.5s, and UORO for horizon values greater than 0.6s. | ['Ritu Bhusal Chhatkuli', 'Kazuyuki Demachi', 'Hiroyuki Takahashi', 'Mitsuru Uesaka', 'Michel Pohl'] | 2021-06-02 | null | null | null | null | ['respiratory-motion-forecasting'] | ['medical'] | [ 1.74044415e-01 3.14176917e-01 -2.94785172e-01 2.27910623e-01
-9.82517719e-01 -2.57518142e-01 1.78361684e-01 1.36462152e-01
-8.50602031e-01 7.56704032e-01 6.20891564e-02 -4.94480431e-01
-3.80112410e-01 -3.40333909e-01 -6.02023959e-01 -1.08637297e+00
-1.41961172e-01 2.61667371e-01 1.68665290e-01 2.29882775e-03
-1.52391583e-01 8.33498895e-01 -9.75737751e-01 -3.84423472e-02
4.22488242e-01 1.12616992e+00 4.33654398e-01 1.05607092e+00
4.10736889e-01 1.03017664e+00 -6.14108443e-01 3.38618815e-01
-2.15956215e-02 -2.03002498e-01 -5.13959944e-01 -3.89835268e-01
-4.23094302e-01 -2.88939148e-01 -6.43388450e-01 5.65239429e-01
8.93853486e-01 2.60526001e-01 6.55617714e-01 -7.74569750e-01
2.58980215e-01 3.01838517e-01 -3.82343441e-01 1.50103167e-01
2.67343283e-01 -5.11302501e-02 1.76709548e-01 -4.98821467e-01
5.00473499e-01 2.40920588e-01 1.35917044e+00 8.35581124e-01
-6.10666454e-01 -2.88903266e-01 -4.89736915e-01 1.13842800e-01
-1.22398508e+00 -3.24357182e-01 3.14277738e-01 -4.08770531e-01
9.94900048e-01 6.69307470e-01 5.81023216e-01 1.18139410e+00
1.09213626e+00 1.69737831e-01 6.20894611e-01 -4.49745655e-01
2.09874362e-01 2.94987410e-01 -3.28495383e-01 4.89324570e-01
-1.13852089e-02 4.09745663e-01 -1.09228410e-01 -1.00195751e-01
6.92686260e-01 3.19061399e-01 -4.69282448e-01 2.96837449e-01
-1.29895341e+00 6.50296986e-01 7.05096662e-01 5.05054891e-01
-6.35113180e-01 4.34374571e-01 5.74280202e-01 -1.16546080e-01
2.82971293e-01 1.51189148e-01 -2.23522559e-01 -4.27997291e-01
-5.39039254e-01 -2.71041542e-01 5.90824127e-01 8.82805169e-01
-2.44393483e-01 -2.30716448e-02 -2.07566470e-01 7.82224715e-01
3.31128031e-01 5.61189711e-01 1.06577587e+00 -7.11239934e-01
3.32984835e-01 -4.49911468e-02 2.61829734e-01 -6.43459797e-01
-1.34555554e+00 -6.51618958e-01 -1.03767538e+00 -4.18949090e-02
5.45594752e-01 -4.88991708e-01 -5.94452858e-01 1.24104905e+00
5.78755975e-01 4.06352011e-03 2.67604172e-01 9.87230480e-01
9.57616389e-01 5.88056505e-01 -2.79770285e-01 -8.60269904e-01
1.18021405e+00 -9.36563671e-01 -7.78781891e-01 4.08554729e-03
1.11295569e+00 -7.50726759e-01 7.05880761e-01 1.84644744e-01
-1.00536895e+00 -2.77130753e-01 -7.45334685e-01 3.90882343e-01
2.88393587e-01 1.95349008e-01 -1.23942560e-02 5.91341734e-01
-1.02689910e+00 9.23029125e-01 -1.22940910e+00 -1.47126064e-01
-8.54280591e-02 5.92304051e-01 1.28833055e-01 2.99532980e-01
-1.18614316e+00 1.07037711e+00 4.97905873e-02 6.34420097e-01
-6.09022975e-01 -9.21886683e-01 -3.90658289e-01 -4.02568847e-01
1.66672602e-01 -6.11908674e-01 1.83913374e+00 -6.61653340e-01
-2.02025270e+00 3.05126756e-01 4.93388176e-02 -6.67368352e-01
8.00003529e-01 -4.67531681e-02 -3.03152204e-01 1.36522889e-01
-2.82730848e-01 1.37877762e-01 5.11494100e-01 -7.18193591e-01
-3.10116917e-01 -1.23237513e-01 -3.91804159e-01 3.85912150e-01
-2.45500997e-01 -1.17787793e-02 -1.47966087e-01 -4.01448399e-01
3.48225266e-01 -1.37371266e+00 -4.86234814e-01 -5.52134775e-02
-3.73989403e-01 1.73089266e-01 3.67397815e-01 -6.16389632e-01
9.34365273e-01 -1.84884655e+00 -1.79667100e-01 5.64393513e-02
-1.87410206e-01 -1.88438877e-01 5.15986919e-01 3.00606042e-01
-1.49838477e-01 -3.69568348e-01 -1.14677459e-01 -1.51844293e-01
-5.27505517e-01 2.23791748e-02 -1.94589701e-03 1.08283055e+00
-8.08625340e-01 8.75454485e-01 -8.84104073e-01 -2.83668905e-01
1.30634829e-01 7.80454934e-01 2.33312771e-02 5.31714000e-02
1.10379890e-01 6.48843944e-01 -4.21744823e-01 2.94213355e-01
1.30141616e-01 -9.64937508e-02 -8.94524679e-02 -2.81365156e-01
-3.80001217e-01 2.61477053e-01 -7.22090662e-01 1.52264512e+00
-9.51930463e-01 9.18772638e-01 -1.68923974e-01 -4.33702826e-01
7.51053870e-01 7.68358350e-01 1.18609023e+00 -8.29050183e-01
3.73510242e-01 2.70867378e-01 -1.29139528e-01 -6.76135600e-01
4.56172913e-01 -3.75102699e-01 3.72557305e-02 3.74474615e-01
-6.16029918e-01 -1.16535634e-01 -3.70892435e-01 -1.68473691e-01
1.30564797e+00 -1.34768248e-01 2.18649417e-01 7.27464035e-02
4.85521346e-01 8.82913917e-03 3.47942770e-01 6.87451124e-01
-1.38721287e-01 6.04845941e-01 6.31691292e-02 -2.95479417e-01
-6.68647051e-01 -8.67305458e-01 -3.86634827e-01 6.45190597e-01
1.73726439e-01 1.19754365e-02 -4.42095160e-01 -5.13017476e-01
-3.28620911e-01 9.34431732e-01 -2.63033569e-01 -2.74225980e-01
-8.61592591e-01 -9.41152751e-01 4.53060538e-01 7.11219609e-01
-4.95071942e-03 -1.18876863e+00 -1.18018246e+00 3.05549145e-01
-3.75862986e-01 -1.25813389e+00 -4.27844018e-01 3.19882751e-01
-1.26146209e+00 -8.17448139e-01 -9.39974427e-01 -3.61686856e-01
8.00067186e-01 7.22276941e-02 7.86276996e-01 -2.17388451e-01
-5.34837127e-01 6.05656445e-01 -2.16752902e-01 -3.99185449e-01
-6.20948374e-01 -2.69747898e-02 1.64697528e-01 -5.52978098e-01
-5.31470656e-01 -2.17168808e-01 -5.58198750e-01 4.76036310e-01
-4.22135651e-01 -4.97989170e-02 4.92657602e-01 8.63018572e-01
6.73645198e-01 1.53780729e-01 3.82021219e-01 -6.19593024e-01
2.35941425e-01 -5.03327191e-01 -8.24131191e-01 -4.15967442e-02
-4.17778730e-01 -1.42190680e-01 9.91160631e-01 -5.70792377e-01
-7.86583722e-01 2.65510857e-01 -1.08517654e-01 -4.85528201e-01
1.09498985e-01 2.90332049e-01 6.90424919e-01 -2.05808595e-01
1.09709513e+00 1.04912482e-01 1.49281397e-01 1.14479475e-01
1.44765778e-02 5.75815618e-01 5.98449767e-01 7.25625306e-02
4.95211840e-01 4.46121931e-01 3.58696461e-01 -1.05354500e+00
-6.44636035e-01 -4.36187029e-01 -3.18294317e-01 -6.94439173e-01
8.04419041e-01 -6.72764122e-01 -1.22292042e+00 3.85710806e-01
-9.89378035e-01 -6.81494713e-01 -5.45027435e-01 1.11599147e+00
-8.56063545e-01 2.16708988e-01 -9.77816522e-01 -1.05264807e+00
-9.99207497e-01 -9.02495086e-01 7.89093435e-01 3.32826644e-01
-2.71164477e-01 -1.04991162e+00 -1.29720196e-01 4.76100706e-02
4.66314167e-01 3.74079168e-01 2.12644473e-01 -3.24698120e-01
-1.68664247e-01 -5.62218904e-01 2.71104574e-01 -1.55740798e-01
1.13139890e-01 -1.92499712e-01 -9.45894063e-01 -5.82308948e-01
6.92659318e-01 8.60381350e-02 2.25431204e-01 8.92890513e-01
1.18608332e+00 -4.10370141e-01 -7.16230094e-01 4.39320654e-01
1.40150905e+00 6.22647166e-01 4.14124161e-01 4.53731239e-01
5.54371834e-01 2.78549165e-01 8.28499019e-01 4.90340710e-01
7.07470626e-02 8.17930043e-01 7.43841469e-01 1.09691516e-01
1.77436277e-01 3.17050904e-01 5.80260575e-01 9.85816777e-01
-8.93138275e-02 -2.82126218e-01 -1.02042019e+00 1.88222840e-01
-1.46845281e+00 -6.55644417e-01 -4.30575192e-01 2.71230459e+00
5.67281008e-01 2.39677981e-01 -1.06457444e-02 1.58280388e-01
5.04931033e-01 -1.05086133e-01 -5.60852766e-01 -4.67746258e-01
5.69066942e-01 -3.08034033e-01 1.13111269e+00 5.33088505e-01
-8.09247792e-01 -5.18628098e-02 5.31253433e+00 4.88080084e-01
-1.43941104e+00 1.18067965e-01 4.36798006e-01 -5.84052920e-01
1.12201653e-01 -4.69614834e-01 -8.33598137e-01 5.57967186e-01
1.73280537e+00 8.32134783e-02 3.56597602e-01 8.12978446e-01
4.38161135e-01 -4.16188151e-01 -8.44889998e-01 9.77142155e-01
-2.42174175e-02 -9.00566876e-01 -1.14648736e+00 3.11401952e-02
5.30742347e-01 5.87664962e-01 -8.75131637e-02 8.04903284e-02
-3.58088344e-01 -1.02392411e+00 4.46998328e-01 8.67104590e-01
9.06561553e-01 -8.32732558e-01 9.65890944e-01 9.81890440e-01
-9.23132062e-01 -3.41459394e-01 -2.70546228e-01 5.03771305e-01
2.70168453e-01 7.81554997e-01 -1.61259902e+00 7.81157494e-01
5.23913324e-01 4.63809222e-01 1.53737918e-01 1.01100171e+00
-1.08927436e-01 3.64558697e-01 -8.09216559e-01 -4.71206874e-01
1.21292338e-01 7.04834536e-02 6.23542190e-01 9.51883614e-01
5.76273501e-01 1.74620703e-01 -1.97590753e-01 9.35357511e-02
2.73888439e-01 -7.42004812e-02 -4.19115186e-01 4.09116626e-01
4.63029474e-01 1.14523625e+00 -6.00888073e-01 1.41063377e-01
-5.56171760e-02 7.28032351e-01 -2.69107938e-01 -1.12561393e-04
-1.34233141e+00 -4.81517196e-01 1.66447312e-02 2.85480171e-01
-2.84491032e-02 -1.54891893e-01 -2.29493544e-01 -3.67237628e-01
1.58783108e-01 -3.88331890e-01 4.20148194e-01 -7.98940241e-01
-5.36327362e-01 8.93821955e-01 -6.69586882e-02 -1.50399387e+00
-8.32198918e-01 -3.99061859e-01 -5.83662212e-01 6.50269449e-01
-1.08156371e+00 -2.43115753e-01 -4.69224006e-01 1.72400862e-01
4.83370334e-01 2.40048811e-01 9.19875622e-01 2.41978765e-01
-7.35058665e-01 5.55349410e-01 5.42877674e-01 -2.44890839e-01
3.16404104e-01 -8.26734602e-01 2.71613989e-02 2.41882682e-01
-2.77096272e-01 1.19105265e-01 8.01828206e-01 -4.58752155e-01
-1.61235332e+00 -1.21614110e+00 7.84862339e-01 -2.27702007e-01
3.95006388e-01 7.64372498e-02 -6.65373564e-01 4.28319961e-01
-2.61651039e-01 4.17849362e-01 1.89494818e-01 -4.59934652e-01
5.64416945e-01 -1.77544087e-01 -1.22160852e+00 5.56863427e-01
5.18248022e-01 -4.93623883e-01 6.56894073e-02 9.17093992e-01
5.41087031e-01 -1.31060755e+00 -1.25632238e+00 6.46702290e-01
6.46598995e-01 -6.31485999e-01 1.11268759e+00 1.31738394e-01
-9.53420401e-02 -6.22468516e-02 2.34837398e-01 -1.13397646e+00
-7.77971372e-02 -7.40713179e-01 -1.34437665e-01 3.08412969e-01
5.96262038e-01 -9.19119954e-01 1.04530013e+00 3.79834533e-01
-3.54533166e-01 -1.47209609e+00 -1.34453821e+00 -8.33279848e-01
-1.53916091e-01 -7.08848536e-01 -2.19336569e-01 3.50090563e-01
9.76886004e-02 -4.35727537e-02 -9.21895951e-02 3.66858959e-01
4.25846845e-01 -1.69214606e-01 1.66899949e-01 -5.83635628e-01
-3.08915585e-01 -2.28517845e-01 -2.25724161e-01 -9.64362860e-01
-1.02372497e-01 -8.55910301e-01 4.13044989e-01 -1.55567157e+00
-1.70666292e-01 -4.31434155e-01 -1.91705599e-01 1.85855910e-01
2.83753633e-01 -1.53913990e-01 -1.16392702e-01 1.56587541e-01
-1.75434595e-03 4.74433750e-01 1.00660968e+00 1.37326822e-01
-4.24376667e-01 7.05208540e-01 1.96261257e-01 9.08065379e-01
1.00371361e+00 -8.57084394e-01 -4.40628976e-01 -9.07408744e-02
1.57977521e-01 9.51168120e-01 2.42137909e-01 -1.24304616e+00
6.47718370e-01 1.31347045e-01 4.70565557e-01 -9.41090226e-01
6.15091443e-01 -1.04603994e+00 5.46821237e-01 1.03387523e+00
-1.98921263e-01 4.88570184e-02 3.65261614e-01 4.56579924e-01
3.08779806e-01 -5.73778450e-01 8.12430143e-01 2.72432156e-03
7.32450262e-02 1.00274809e-01 -7.08333373e-01 -1.62368894e-01
1.13749957e+00 -3.45936596e-01 2.51744762e-02 -5.03961086e-01
-7.80526400e-01 -4.05774228e-02 -3.28154862e-02 9.91687179e-02
7.25320458e-01 -1.28806841e+00 -4.12115514e-01 -1.15263060e-01
-5.10119736e-01 2.40591705e-01 4.62932974e-01 1.51244497e+00
-5.42880177e-01 5.67476809e-01 5.01154304e-01 -1.00442612e+00
-1.34745324e+00 3.30425650e-01 1.04597831e+00 -3.65250349e-01
-9.23562348e-01 8.26324642e-01 -2.24200994e-01 -1.97224051e-01
3.75397027e-01 -5.11669755e-01 -3.77164960e-01 -6.07694462e-02
3.00406992e-01 7.51609862e-01 4.01749372e-01 -4.81355637e-01
-4.18654680e-01 7.71994293e-01 9.26622823e-02 -2.50523966e-02
1.09985828e+00 1.12904496e-02 2.39543363e-01 6.92922771e-01
1.33111405e+00 -1.47662237e-01 -1.04064500e+00 9.99581367e-02
9.72032547e-02 5.95479198e-02 1.57030508e-01 -6.77516878e-01
-8.61076057e-01 3.84422213e-01 9.51376915e-01 4.87605147e-02
1.31048870e+00 -3.00063998e-01 9.10446882e-01 2.64009684e-01
3.35601926e-01 -9.75237012e-01 3.71138714e-02 6.42775118e-01
8.93186629e-01 -7.01529264e-01 2.68397450e-01 -2.83866465e-01
-6.23281181e-01 1.38672400e+00 3.05641770e-01 1.83525905e-01
4.18648899e-01 3.78176630e-01 2.31434420e-01 1.03145242e-01
-8.77216041e-01 5.12223125e-01 2.49580517e-01 -9.41190496e-02
5.65035820e-01 1.31878937e-02 -1.66890517e-01 2.80043751e-01
-2.06152648e-01 1.00999534e-01 6.70310915e-01 9.77592885e-01
-3.08630377e-01 -4.63676244e-01 -7.51511335e-01 2.02240482e-01
-5.68546355e-01 3.48552436e-01 5.52448094e-01 7.70723641e-01
-4.94460076e-01 7.76948869e-01 6.43496066e-02 -2.36220956e-01
6.64000034e-01 -3.29779088e-01 4.79090095e-01 -1.29305631e-01
-8.76163900e-01 1.02746129e-01 1.17665641e-01 -8.15821826e-01
-9.96955931e-02 -6.94931805e-01 -1.85867250e+00 1.07781470e-01
-6.12891674e-01 2.01246411e-01 1.17231202e+00 6.75166607e-01
-6.88542202e-02 9.48560417e-01 1.06562650e+00 -8.16061676e-01
-8.55000138e-01 -8.87270868e-01 -2.48172432e-01 -4.61451054e-01
5.46799719e-01 -2.76092649e-01 -6.30010903e-01 -5.21242082e-01] | [13.760318756103516, -2.649731159210205] |
3c5b0d03-d99e-4160-adc5-6c5d7fdcb072 | entropy-driven-mixed-precision-quantization | null | null | https://openreview.net/forum?id=E28hy5isRzC | https://openreview.net/pdf?id=E28hy5isRzC | Entropy-Driven Mixed-Precision Quantization for Deep Network Design | Deploying deep convolutional neural networks on Internet-of-Things (IoT) devices is challenging due to the limited computational resources, such as limited SRAM memory and Flash storage. Previous works re-design a small network for IoT devices, and then compress the network size by mixed-precision quantization. This two-stage procedure cannot optimize the architecture and the corresponding quantization jointly, leading to sub-optimal tiny deep models. In this work, we propose a one-stage solution that optimizes both jointly and automatically. The key idea of our approach is to cast the joint architecture design and quantization as an Entropy Maximization process. Particularly, our algorithm automatically designs a tiny deep model such that: 1) Its representation capacity measured by entropy is maximized under the given computational budget; 2) Each layer is assigned with a proper quantization precision; 3) The overall design loop can be done on CPU, and no GPU is required. More impressively, our method can directly search high-expressiveness architecture for IoT devices within less than half a CPU hour. Extensive experiments on three widely adopted benchmarks, ImageNet, VWW and WIDER FACE, demonstrate that our method can achieve the state-of-the-art performance in the tiny deep model regime. Code and pre-trained models are available at https://github.com/alibaba/lightweight-neural-architecture-search. | ['Xiuyu Sun', 'Hao Li', 'Hesen Chen', 'Ming Lin', 'Junyan Wang', 'Ce Ge', 'Zhenhong Sun'] | 2022-11-28 | null | null | null | conference-on-neural-information-processing-1 | ['face-detection'] | ['computer-vision'] | [-1.30739212e-01 1.87360853e-01 -5.01457810e-01 -4.09848869e-01
-3.34017962e-01 -3.57106924e-01 1.91650525e-01 -2.23803014e-01
-5.47255218e-01 4.40339565e-01 -1.19063139e-01 -5.79607189e-01
-2.86948290e-02 -9.05384541e-01 -7.62165666e-01 -5.88838160e-01
3.66969198e-01 3.89358968e-01 -1.25884816e-01 1.84017539e-01
-1.59918100e-01 2.79147238e-01 -1.52782536e+00 -3.58684659e-02
5.74551523e-01 1.81896687e+00 2.70109385e-01 5.76510310e-01
-4.15669046e-02 7.52604604e-01 -2.23109871e-01 -7.18592227e-01
5.70026338e-01 -4.10489216e-02 -6.68979287e-01 -1.93343908e-01
3.78131866e-01 -7.03568280e-01 -6.52354300e-01 1.27424622e+00
6.66747510e-01 -2.13089257e-01 2.14112863e-01 -1.06223452e+00
-5.53064167e-01 1.06353891e+00 -2.55587935e-01 1.03424288e-01
-5.48061907e-01 4.02620047e-01 1.03174138e+00 -5.51469207e-01
1.19920403e-01 1.10842955e+00 5.43520451e-01 7.37728417e-01
-9.15852964e-01 -9.99321997e-01 2.53715403e-02 6.83825389e-02
-1.66733909e+00 -8.09416473e-01 6.36185825e-01 -6.15499020e-02
1.30105042e+00 5.39925173e-02 8.68038237e-01 8.64295661e-01
1.34573266e-01 4.39007491e-01 3.48410308e-01 -2.57834271e-02
6.24186456e-01 -1.05729960e-01 -1.86005160e-01 9.41325665e-01
5.06775975e-01 -9.40674469e-02 -1.81176096e-01 8.72262344e-02
7.07891107e-01 2.38376707e-01 2.48744547e-01 7.29401037e-02
-9.28227723e-01 7.88675487e-01 5.64503968e-01 3.48841280e-01
-5.43092787e-01 9.01627004e-01 6.14935338e-01 2.13018849e-01
1.80564418e-01 1.46253377e-01 -6.97460592e-01 -3.67467195e-01
-1.02823293e+00 -1.55091926e-01 8.60569477e-01 9.85702515e-01
7.80257225e-01 1.70800924e-01 8.89170915e-02 5.98404527e-01
3.73003632e-01 4.65624899e-01 6.65570080e-01 -1.05537212e+00
6.03055418e-01 5.63404858e-01 -3.83168638e-01 -7.93485165e-01
-3.12693208e-01 -3.61634672e-01 -1.34815216e+00 -2.63177723e-01
-5.05227149e-02 -2.71907628e-01 -8.08609366e-01 1.76274943e+00
3.11181337e-01 1.67652845e-01 4.73117605e-02 6.01660728e-01
9.21202540e-01 6.19807899e-01 1.19314782e-01 2.11592130e-02
1.62327647e+00 -1.01532006e+00 -5.33367872e-01 -4.73609984e-01
7.10003972e-01 -4.48364466e-01 9.82134402e-01 1.45014137e-01
-1.31295705e+00 -5.65512121e-01 -1.24263442e+00 -2.14557812e-01
-2.23468289e-01 3.92961264e-01 7.26374269e-01 8.67962360e-01
-1.16190541e+00 6.27330124e-01 -1.03164542e+00 -4.61726915e-03
9.86843526e-01 8.59435320e-01 4.74446267e-02 8.81754011e-02
-8.95490050e-01 3.51755828e-01 8.57234240e-01 -8.18654057e-03
-7.87091851e-01 -6.69023156e-01 -6.05008662e-01 4.99149621e-01
2.40649045e-01 -9.85189319e-01 1.31532228e+00 -8.62564206e-01
-1.72805607e+00 5.37656367e-01 -3.00732385e-02 -7.88235605e-01
3.16722780e-01 4.88595888e-02 -2.54643798e-01 1.04502164e-01
-4.09853429e-01 1.04558599e+00 1.01797974e+00 -6.02936029e-01
-5.21714985e-01 -2.73988187e-01 2.40076348e-01 -2.35125944e-01
-1.20199633e+00 -2.02716142e-01 -7.20390677e-01 -4.56941187e-01
-1.12728707e-01 -1.05852413e+00 -3.19650054e-01 2.32157484e-01
-4.15426850e-01 -3.93293977e-01 7.40410328e-01 -3.35017323e-01
1.56631756e+00 -2.10664082e+00 -1.30625144e-01 1.57326579e-01
5.73369741e-01 5.37188649e-01 -9.94113013e-02 -3.39285195e-01
2.34107912e-01 5.57578385e-01 1.59814611e-01 -6.00802243e-01
3.51422042e-01 2.67652541e-01 -6.76954091e-02 1.91291243e-01
-3.66781205e-02 1.38489437e+00 -4.75881457e-01 -6.23801529e-01
2.16137439e-01 7.79930055e-01 -7.82508373e-01 -1.06069759e-01
-2.30534732e-01 -8.38528425e-02 -5.30852854e-01 7.74063170e-01
6.74777150e-01 -6.96489394e-01 3.24877381e-01 -5.94436049e-01
5.71461059e-02 4.21185225e-01 -9.95161772e-01 1.71647346e+00
-8.17680180e-01 5.25007546e-01 -7.74239227e-02 -8.69876564e-01
8.67313206e-01 2.20818058e-01 4.46272284e-01 -8.87989044e-01
6.59638822e-01 4.89906758e-01 -1.22564413e-01 -2.42107093e-01
1.74395204e-01 1.92053244e-01 1.43962288e-02 4.68838871e-01
1.24368735e-01 7.94837028e-02 1.64175585e-01 -2.08451331e-01
1.16446805e+00 -5.51643372e-01 2.26314858e-01 -1.94232687e-01
2.04553694e-01 -6.30153775e-01 5.97471833e-01 5.18723547e-01
-2.07876608e-01 1.17427744e-01 4.31822062e-01 -8.84927034e-01
-1.28784621e+00 -6.55459344e-01 -6.81992471e-02 9.74716187e-01
-9.51184705e-02 -6.31984174e-01 -1.03246140e+00 -4.88231629e-01
-2.61656940e-01 3.18005145e-01 -4.44805562e-01 -2.94955492e-01
-7.80543029e-01 -5.30640602e-01 7.62722194e-01 6.38933361e-01
9.62023973e-01 -1.00993514e+00 -1.21172512e+00 1.09466434e-01
3.83215845e-02 -1.27883518e+00 -6.16010904e-01 3.93151760e-01
-1.12315619e+00 -5.49869239e-01 -3.87158990e-01 -9.81375933e-01
5.78687251e-01 -1.12698488e-01 1.10875440e+00 4.14520055e-01
-2.02854276e-01 -3.15193951e-01 4.80424538e-02 -1.34085476e-01
-1.19417943e-01 5.25344491e-01 2.33767461e-02 -2.81375259e-01
3.86240840e-01 -7.30236650e-01 -9.59937453e-01 8.92690793e-02
-9.61936891e-01 8.55522901e-02 8.43594432e-01 6.55654371e-01
9.05598938e-01 3.38168293e-01 4.10045594e-01 -2.77015835e-01
3.84801030e-01 -3.97044897e-01 -8.75883460e-01 1.19914874e-01
-6.69388890e-01 3.11834335e-01 8.25175345e-01 -6.73847556e-01
-5.15312374e-01 3.31058294e-01 -4.04783040e-01 -8.08019638e-01
1.23165049e-01 3.02424699e-01 -3.73541653e-01 -9.85545442e-02
2.79780835e-01 3.55725028e-02 -2.20148981e-01 -3.85398060e-01
3.64093751e-01 7.50681341e-01 2.37520754e-01 -3.82421315e-01
5.01854181e-01 3.15307885e-01 2.55777780e-02 -6.62370205e-01
-5.35088718e-01 8.31240788e-02 -4.23466861e-01 1.30489573e-01
7.72247434e-01 -1.10959923e+00 -1.19625461e+00 2.58234173e-01
-1.07931376e+00 -6.07207477e-01 -4.19147789e-01 3.27446193e-01
-2.49856740e-01 -1.91972069e-02 -5.73121011e-01 -5.71043193e-01
-1.09576654e+00 -1.53194535e+00 1.04259098e+00 3.04815590e-01
-3.39415520e-02 -9.13913846e-01 -3.74199957e-01 2.02850178e-01
6.87678814e-01 -7.55335689e-02 9.10421848e-01 -3.68404806e-01
-7.46510029e-01 -1.24147348e-01 -4.47600573e-01 5.60290873e-01
-1.01455115e-01 -2.43797362e-01 -8.96099031e-01 -3.56906474e-01
7.67957373e-03 -5.08862853e-01 9.27708447e-01 3.85096878e-01
1.93360734e+00 -8.23746920e-01 -1.22287832e-01 1.22071433e+00
1.54079735e+00 8.80890191e-02 6.85150385e-01 1.09056726e-01
8.07387173e-01 4.69368733e-02 4.36525047e-02 7.20561624e-01
4.58716094e-01 5.58572590e-01 7.18268335e-01 2.27052197e-01
-1.84361875e-01 -2.18848631e-01 1.40566319e-01 1.19120169e+00
9.91432667e-02 -3.97717804e-01 -8.66809964e-01 4.87812519e-01
-1.69464111e+00 -6.69741511e-01 4.77869719e-01 1.92252946e+00
8.67780089e-01 2.48827383e-01 -1.02323316e-01 1.53339192e-01
4.60209161e-01 2.00376898e-01 -1.09287262e+00 -4.83800054e-01
1.57818496e-01 4.90112960e-01 8.25885057e-01 1.18917644e-01
-1.05121577e+00 8.26346993e-01 5.43049383e+00 1.16206324e+00
-1.36604369e+00 4.16412443e-01 1.16155267e+00 -5.45810938e-01
-2.75696397e-01 -2.50657558e-01 -1.01103270e+00 7.12600172e-01
1.41386461e+00 -1.29009960e-02 7.60950744e-01 1.26048708e+00
-5.20478487e-02 5.90478241e-01 -1.04184198e+00 1.47908115e+00
-3.70954394e-01 -1.63985586e+00 4.22279723e-02 4.78958666e-01
7.47657359e-01 3.91013473e-01 3.68713200e-01 2.00361267e-01
1.00548618e-01 -1.11453581e+00 1.05143332e+00 2.87354458e-02
1.28654540e+00 -9.41164732e-01 7.14912951e-01 1.78884625e-01
-1.59366882e+00 -5.43995738e-01 -7.30937064e-01 1.32533729e-01
-1.26967251e-01 8.22959006e-01 -4.31530058e-01 -1.05554417e-01
9.24910486e-01 5.06900012e-01 -4.02112484e-01 6.88462079e-01
-8.48789737e-02 4.63501096e-01 -6.74853563e-01 -4.54271317e-01
2.59542823e-01 1.14400156e-01 -5.47157787e-02 9.79516685e-01
6.13172710e-01 5.66539019e-02 -6.54494613e-02 1.00367045e+00
-8.04906964e-01 -1.28259331e-01 -2.26942852e-01 -3.77974242e-01
9.81391728e-01 1.27560019e+00 -7.80320644e-01 -4.07923043e-01
-2.46601358e-01 7.91676819e-01 3.24346691e-01 3.50350365e-02
-1.09564638e+00 -4.07309622e-01 8.64272177e-01 -1.06127560e-01
7.58758187e-01 7.24601895e-02 -7.37986267e-01 -1.07823801e+00
3.08664858e-01 -7.72089660e-01 -7.13844085e-03 -3.33186477e-01
-7.64073789e-01 8.69036436e-01 -3.70531738e-01 -9.57897186e-01
-1.21824473e-01 -5.21094620e-01 -4.01040554e-01 3.74447227e-01
-1.51997435e+00 -1.09790969e+00 -4.21475679e-01 4.89772439e-01
4.12200987e-01 -2.14592710e-01 7.12572634e-01 6.78776503e-01
-8.82962286e-01 1.09025311e+00 2.44527415e-04 -1.60791334e-02
-8.10069516e-02 -7.21445084e-01 7.72278488e-01 4.94905472e-01
-3.43508087e-02 5.76949656e-01 3.13383430e-01 -3.23887557e-01
-1.70949090e+00 -1.44213521e+00 7.83652544e-01 9.62975323e-02
6.61404014e-01 -4.74478036e-01 -5.37379324e-01 5.11406362e-01
9.09830406e-02 4.52538788e-01 5.29685378e-01 -3.37184608e-01
-3.72694016e-01 -5.16160786e-01 -1.29211891e+00 4.93250996e-01
1.27777100e+00 -4.01427716e-01 1.94152966e-01 4.43121850e-01
1.09014261e+00 -2.56121933e-01 -1.01187289e+00 3.40824217e-01
8.20300221e-01 -7.20576108e-01 1.07939303e+00 -1.74695656e-01
4.37147707e-01 4.83822189e-02 -3.67850929e-01 -8.19430470e-01
-7.78909326e-02 -8.15505624e-01 -8.00218642e-01 1.04990709e+00
4.05086040e-01 -7.26269960e-01 8.62361729e-01 6.70279324e-01
6.38085976e-03 -1.37667823e+00 -1.17885101e+00 -7.26206064e-01
-1.87439639e-02 -5.60828567e-01 1.22004735e+00 3.68803412e-01
-3.60836208e-01 2.04810664e-01 -2.47157589e-01 -1.26268879e-01
5.29517174e-01 -1.70285046e-01 3.99304450e-01 -1.23263776e+00
-3.46418709e-01 -7.01363385e-01 -4.45363641e-01 -1.34551930e+00
1.99297085e-01 -6.23828173e-01 -4.26353477e-02 -1.26071668e+00
1.43021017e-01 -7.84132063e-01 -3.85519296e-01 8.43098402e-01
3.99634510e-01 4.00546610e-01 2.71370828e-01 1.58335239e-01
-7.83068776e-01 6.35414183e-01 9.81268823e-01 -2.72958606e-01
-4.59467806e-02 -2.13958845e-01 -8.48235071e-01 6.82932854e-01
1.17696798e+00 -5.05813241e-01 -6.00889742e-01 -9.09775317e-01
5.23747087e-01 -2.93601185e-01 3.00686955e-01 -1.22334135e+00
3.24531376e-01 1.71544522e-01 2.06411347e-01 -2.62661159e-01
3.79746675e-01 -9.30895448e-01 2.30270609e-01 7.87299097e-01
-1.74715996e-01 1.80875614e-01 2.36470416e-01 2.71355301e-01
1.32798612e-01 -2.49023601e-01 9.45230961e-01 -7.94201866e-02
-5.28972626e-01 8.04568768e-01 2.02977702e-01 2.98673287e-04
8.76970887e-01 -6.68110279e-03 -4.09054428e-01 -1.20583870e-01
-3.22226197e-01 -6.02927580e-02 3.71309310e-01 2.45692298e-01
6.43596947e-01 -1.53312886e+00 -2.73196250e-01 2.66312450e-01
-2.46963069e-01 1.53336748e-01 1.91605106e-01 7.01094389e-01
-6.69148386e-01 6.36169195e-01 -4.05034125e-02 -4.56691742e-01
-1.01618290e+00 5.03027737e-01 4.43146318e-01 -4.40366805e-01
-5.62590480e-01 1.04397893e+00 5.10553867e-02 -1.24173969e-01
5.28283119e-01 -6.43017232e-01 3.31067115e-01 -2.26447612e-01
7.35550702e-01 3.34518224e-01 2.21343234e-01 -5.32928765e-01
-4.04945344e-01 6.32902086e-01 3.03365998e-02 2.13606462e-01
1.41502523e+00 -6.73698336e-02 -1.72334626e-01 -1.67664513e-01
1.73200405e+00 -6.95775867e-01 -1.38130987e+00 -9.21729356e-02
-5.42086184e-01 -1.21513776e-01 6.02044284e-01 -2.72429526e-01
-1.73691809e+00 9.90132391e-01 9.11837161e-01 1.74817115e-01
1.44944119e+00 5.27223535e-02 1.39868605e+00 7.29767621e-01
3.10652971e-01 -1.09267867e+00 2.05013886e-01 4.74927068e-01
3.48481089e-01 -1.05594134e+00 -6.07651360e-02 2.14436483e-02
-1.52321815e-01 1.14161599e+00 5.65911472e-01 8.19648728e-02
9.68683481e-01 5.99951863e-01 -3.89646947e-01 -1.63930655e-01
-9.46156979e-01 6.37404248e-02 8.02300274e-02 7.98403248e-02
1.07669897e-01 2.57378995e-01 1.40346199e-01 7.97912240e-01
-6.01959348e-01 2.55333453e-01 -7.67755955e-02 8.74616802e-01
-4.07391429e-01 -7.87456632e-01 1.90697193e-01 6.37262762e-01
-5.81193984e-01 -3.05757940e-01 5.73628321e-02 2.04152942e-01
3.45166534e-01 8.60573173e-01 4.13323194e-01 -8.89404356e-01
2.88146660e-02 -2.48880208e-01 2.85832584e-01 -2.05981717e-01
-5.22107124e-01 -5.14997244e-02 -1.90904617e-01 -8.44077110e-01
-6.16967529e-02 -1.25473633e-01 -1.22409630e+00 -8.16360831e-01
-4.53237325e-01 -2.72249192e-01 1.08686686e+00 8.25918376e-01
7.68552244e-01 6.39452219e-01 3.78627211e-01 -9.05371130e-01
-6.89064085e-01 -7.79886544e-01 -1.16686761e-01 -1.37246221e-01
3.28087151e-01 -2.68929452e-01 -2.31795281e-01 -4.05564234e-02] | [8.580158233642578, 2.992003917694092] |
f29d5120-8d51-4ebe-8e2c-caed3846042c | mitigating-the-hubness-problem-for-zero-shot | 1907.06371 | null | https://arxiv.org/abs/1907.06371v1 | https://arxiv.org/pdf/1907.06371v1.pdf | Mitigating the Hubness Problem for Zero-Shot Learning of 3D Objects | The development of advanced 3D sensors has enabled many objects to be captured in the wild at a large scale, and a 3D object recognition system may therefore encounter many objects for which the system has received no training. Zero-Shot Learning (ZSL) approaches can assist such systems in recognizing previously unseen objects. Applying ZSL to 3D point cloud objects is an emerging topic in the area of 3D vision, however, a significant problem that ZSL often suffers from is the so-called hubness problem, which is when a model is biased to predict only a few particular labels for most of the test instances. We observe that this hubness problem is even more severe for 3D recognition than for 2D recognition. One reason for this is that in 2D one can use pre-trained networks trained on large datasets like ImageNet, which produces high-quality features. However, in the 3D case there are no such large-scale, labelled datasets available for pre-training which means that the extracted 3D features are of poorer quality which, in turn, exacerbates the hubness problem. In this paper, we therefore propose a loss to specifically address the hubness problem. Our proposed method is effective for both Zero-Shot and Generalized Zero-Shot Learning, and we perform extensive evaluations on the challenging datasets ModelNet40, ModelNet10, McGill and SHREC2015. A new state-of-the-art result for both zero-shot tasks in the 3D case is established. | ['Lars Petersson', 'Shafin Rahman', 'Ali Cheraghian', 'Dylan Campbell'] | 2019-07-15 | null | null | null | null | ['3d-object-recognition'] | ['computer-vision'] | [ 5.60909696e-02 -7.79480338e-02 2.71928255e-02 -1.94673210e-01
-6.57953441e-01 -2.98687041e-01 7.49947965e-01 6.15361072e-02
-1.86254129e-01 2.58041948e-01 -3.29131424e-01 6.73021004e-02
-1.29648104e-01 -7.94270992e-01 -6.10184073e-01 -6.97716951e-01
-9.10279378e-02 7.09975183e-01 6.63185418e-01 -2.42025293e-02
1.36129573e-01 8.70281875e-01 -2.10712337e+00 -1.21558398e-01
4.42943931e-01 1.28955710e+00 4.12845939e-01 5.20557761e-01
-2.85955608e-01 5.03036320e-01 -5.67282736e-01 -1.70450360e-01
5.66351235e-01 -8.64319131e-03 -6.08766854e-01 3.98600668e-01
7.58333862e-01 -2.17620775e-01 -9.05435532e-02 1.13334239e+00
5.72780728e-01 2.96276420e-01 9.36544955e-01 -1.46099365e+00
-3.90993774e-01 -2.43312836e-01 -5.17177701e-01 1.29325598e-01
2.00246245e-01 3.46014172e-01 8.89584601e-01 -1.30701840e+00
6.62481606e-01 1.25131106e+00 5.62183797e-01 5.44661105e-01
-8.04628611e-01 -4.40044761e-01 -1.62719414e-01 4.27248329e-01
-1.34560132e+00 -2.80593634e-01 8.62497211e-01 -6.79336429e-01
1.20328772e+00 5.08107655e-02 5.53012431e-01 1.17854726e+00
-8.17777216e-02 7.08061993e-01 9.81912076e-01 -3.63829374e-01
5.74810743e-01 9.15173739e-02 9.00110751e-02 3.44163507e-01
2.85026193e-01 2.19502538e-01 -2.70754695e-01 8.95180702e-02
5.53264976e-01 4.02757615e-01 -3.09074614e-02 -9.43768263e-01
-9.72332358e-01 6.77319586e-01 7.37154603e-01 3.17211926e-01
-3.16068143e-01 -2.73242742e-01 2.76290387e-01 2.97753900e-01
5.61490178e-01 4.89000589e-01 -2.62128264e-01 -3.43596376e-02
-7.41872191e-01 2.39048302e-01 7.39974141e-01 1.06104982e+00
9.79015589e-01 -1.38410792e-01 -1.41917421e-02 1.04822123e+00
1.90112591e-01 6.13632679e-01 3.52770984e-01 -6.76687598e-01
3.03238988e-01 8.66584897e-01 1.49606809e-01 -8.95084143e-01
-3.61837655e-01 -3.38909745e-01 -1.05939019e+00 6.16503358e-01
3.21348786e-01 3.54589671e-01 -1.43725789e+00 1.39692855e+00
4.59660619e-01 4.02191699e-01 3.16532850e-02 1.24088228e+00
8.84808540e-01 5.29079914e-01 -2.46889517e-01 3.60662490e-02
9.83070672e-01 -7.48889506e-01 -2.05359057e-01 -4.30129707e-01
4.72982526e-01 -5.03507137e-01 9.78738964e-01 8.65800083e-02
-5.74900448e-01 -5.45965791e-01 -1.12716436e+00 2.49039933e-01
-7.29175150e-01 -5.02743781e-01 2.96448559e-01 3.46662939e-01
-6.01471663e-01 5.95841289e-01 -5.42728901e-01 -8.43156517e-01
7.76311874e-01 1.14633769e-01 -5.75878680e-01 -6.32325292e-01
-1.01906216e+00 1.31658256e+00 3.00065815e-01 -2.02857003e-01
-1.16371834e+00 -6.78810716e-01 -8.90595078e-01 5.34124374e-02
5.13177156e-01 -2.71711320e-01 1.04084921e+00 -3.77433509e-01
-1.06912673e+00 1.22510540e+00 2.63939142e-01 -2.37555608e-01
5.15337586e-01 -3.02024037e-02 -1.74177840e-01 9.29624029e-03
1.71752334e-01 6.75968647e-01 1.02479315e+00 -1.44177449e+00
-5.63056767e-01 -7.33466268e-01 7.52353519e-02 1.24641232e-01
-6.95327967e-02 -2.27056473e-01 -3.83503497e-01 -2.03726903e-01
2.32646763e-01 -9.49067056e-01 -3.26058984e-01 2.34286815e-01
-1.90602988e-01 -4.55334842e-01 1.00170171e+00 1.52259499e-01
3.43421727e-01 -2.22300315e+00 4.61327890e-03 5.19858040e-02
2.34972030e-01 5.95733225e-01 -1.84855729e-01 1.01696402e-01
2.19583008e-02 -6.46708757e-02 -3.10173273e-01 -2.99001932e-01
-9.08836871e-02 4.97983217e-01 -1.62837416e-01 4.67313111e-01
4.79112506e-01 8.00919414e-01 -1.00658548e+00 -4.64816809e-01
6.40309989e-01 4.29868281e-01 -2.38654643e-01 4.58479643e-01
-7.54809603e-02 3.16532373e-01 -2.29684949e-01 9.47005033e-01
6.96567655e-01 -5.57460368e-01 -5.11797249e-01 -6.29340485e-02
-2.20521614e-02 -1.38018236e-01 -1.25903141e+00 1.67812908e+00
-3.69841278e-01 3.96083325e-01 -4.34286594e-01 -1.10610247e+00
1.08009839e+00 2.42418244e-01 5.55416644e-01 -5.74986935e-01
2.33705193e-01 2.85839021e-01 -1.97016254e-01 -5.63286424e-01
1.69684827e-01 -4.82425839e-01 -3.28719383e-03 1.94957927e-01
4.29022461e-01 -4.38632369e-01 4.61864993e-02 -1.22478623e-02
1.36048269e+00 -1.55736923e-01 2.86773741e-01 1.00510694e-01
2.47736454e-01 8.96056071e-02 5.41297376e-01 9.43740368e-01
-5.86376846e-01 8.72724831e-01 1.47600546e-01 -5.13795555e-01
-1.13858926e+00 -9.97242033e-01 -1.93452418e-01 5.60967505e-01
4.61562812e-01 -6.96219876e-02 -2.64950782e-01 -7.98372865e-01
2.45036602e-01 7.03467548e-01 -4.99078333e-01 -4.71093208e-01
2.53826808e-02 -3.01474154e-01 2.30150402e-01 3.29750001e-01
4.63048279e-01 -1.02379251e+00 -1.05364990e+00 9.55729261e-02
1.52336672e-01 -1.21836102e+00 2.03414753e-01 4.31117654e-01
-9.73569274e-01 -1.16846049e+00 -8.89258564e-01 -6.34522676e-01
5.91481805e-01 8.33835363e-01 1.10459840e+00 -1.49869502e-01
-5.02955616e-01 3.78373146e-01 -6.35021627e-01 -7.96757698e-01
-7.31172785e-02 -1.27191141e-01 2.23434970e-01 -1.30631989e-02
7.96382785e-01 -4.31681544e-01 -3.35855365e-01 5.04806519e-01
-7.75099099e-01 -2.22005665e-01 6.62124038e-01 8.06756675e-01
5.69164455e-01 4.73953523e-02 6.52322650e-01 -6.09115303e-01
3.72652002e-02 -4.80511159e-01 -5.18976808e-01 1.45597979e-01
-6.27476692e-01 -2.29060113e-01 3.38694960e-01 -6.00900114e-01
-6.69985712e-01 1.06551953e-01 9.55733377e-03 -1.27202082e+00
-6.07351482e-01 1.51506409e-01 -2.87646055e-01 -2.15303972e-01
7.58046448e-01 -5.34328399e-03 2.53159497e-02 -7.26953387e-01
1.17654786e-01 7.99502313e-01 2.93170422e-01 -1.20236494e-01
1.05568945e+00 4.28035975e-01 2.84083784e-01 -1.21614814e+00
-1.10392702e+00 -8.65841150e-01 -7.49678075e-01 -4.91902143e-01
7.94610560e-01 -8.32537591e-01 -1.89733759e-01 5.90317070e-01
-1.12936091e+00 -2.97782514e-02 -6.58342242e-01 4.37720895e-01
-6.45041347e-01 8.04543868e-02 -9.82731394e-03 -1.00371599e+00
-6.41982853e-02 -1.05907285e+00 1.16863167e+00 1.79393664e-01
-1.44562786e-02 -4.94058937e-01 1.36457473e-01 1.44616768e-01
2.57293046e-01 2.76615828e-01 9.75310087e-01 -8.13304722e-01
-7.03017890e-01 -6.19111657e-01 -4.03614670e-01 4.58435953e-01
6.11413606e-02 -3.41727048e-01 -1.11111534e+00 -2.43371770e-01
2.10866109e-01 -6.92618430e-01 7.37851143e-01 -3.33311572e-03
8.30159426e-01 2.34091505e-01 -2.12320209e-01 3.59127522e-01
1.50842881e+00 4.28608358e-02 5.36959171e-01 1.27800092e-01
7.44209111e-01 5.59383392e-01 9.82860506e-01 4.01496530e-01
1.25333726e-01 6.92302823e-01 8.64786744e-01 1.26578249e-02
-1.26717180e-01 -7.96581581e-02 -8.12008753e-02 6.13705516e-01
-5.47995493e-02 -1.96970716e-01 -1.23631012e+00 8.29332829e-01
-1.77727675e+00 -9.33806360e-01 1.16194800e-01 2.42526603e+00
4.37554002e-01 3.12582582e-01 -1.45731315e-01 3.81333232e-01
7.23274052e-01 1.64679214e-01 -6.74610794e-01 5.91143332e-02
-1.26468344e-02 9.56514254e-02 4.13866818e-01 -3.81190912e-03
-1.20407891e+00 7.73241103e-01 5.21494293e+00 6.91921175e-01
-1.16374600e+00 2.06844851e-01 3.05079699e-01 -1.76360905e-01
3.27403784e-01 -3.73306610e-02 -8.84516120e-01 4.01461214e-01
5.74684262e-01 -9.57259387e-02 -3.83993872e-02 1.23065138e+00
-1.92236751e-01 -7.55276531e-02 -1.31023443e+00 1.52046871e+00
2.54691601e-01 -1.22200286e+00 3.29838283e-02 8.49749446e-02
7.09966063e-01 4.88218963e-01 -1.67806581e-01 5.08851647e-01
-1.14974827e-01 -8.72235954e-01 3.87336165e-01 3.96568030e-01
7.77983487e-01 -5.73743463e-01 7.13582098e-01 7.69845486e-01
-9.10757005e-01 -1.81716941e-02 -9.23905313e-01 -6.25096709e-02
5.53040579e-02 8.71821940e-01 -9.12416339e-01 4.00306940e-01
9.08234954e-01 8.70356679e-01 -5.81860602e-01 1.54602373e+00
-1.11169890e-01 1.36981562e-01 -4.34837729e-01 -1.15961976e-01
2.66534865e-01 2.37160221e-01 8.72609317e-01 6.81022465e-01
4.61057395e-01 1.93053752e-01 2.03137234e-01 7.61240602e-01
-8.60035792e-02 -1.96599036e-01 -1.21115136e+00 1.43319324e-01
2.54101604e-01 1.25066066e+00 -6.27188742e-01 -2.82471597e-01
-6.61877811e-01 7.63920546e-01 2.72124588e-01 8.92008170e-02
-3.58931303e-01 -2.87519425e-01 7.14117050e-01 9.82686356e-02
5.46170890e-01 -2.48149171e-01 -6.20689504e-02 -1.17438078e+00
-8.80497172e-02 -5.64092040e-01 1.27016842e-01 -8.32744479e-01
-1.79515982e+00 3.54519576e-01 -4.79827188e-02 -1.63078511e+00
-1.81189418e-01 -7.96120465e-01 -5.67796409e-01 6.46592677e-01
-1.60966158e+00 -9.21537399e-01 -5.76450229e-01 6.12016141e-01
7.22075343e-01 -1.38478041e-01 7.62854517e-01 4.04744655e-01
-1.99666783e-01 7.66241923e-02 -1.09919056e-01 -7.26345703e-02
6.75820470e-01 -1.12903142e+00 2.61208624e-01 5.94827056e-01
3.63790393e-01 -7.37133808e-03 6.11350417e-01 -6.37580574e-01
-1.51650405e+00 -1.29706204e+00 9.19920266e-01 -7.44098604e-01
4.12569046e-01 -4.08588797e-01 -1.13818574e+00 2.46772274e-01
-5.39659500e-01 5.10323942e-01 4.77591515e-01 1.29093109e-02
-4.10506248e-01 -2.30540484e-02 -1.22865236e+00 2.23397732e-01
1.25259376e+00 -4.91679907e-01 -8.40178847e-01 4.59590703e-01
6.03678107e-01 -1.67928383e-01 -8.30181539e-01 6.79880619e-01
2.67333686e-01 -9.84730899e-01 9.73664403e-01 -5.81749082e-01
3.61479729e-01 -2.46633932e-01 -4.58067447e-01 -1.35306621e+00
-2.91958898e-01 1.30957127e-01 -4.29619491e-01 1.02486861e+00
1.80844858e-01 -2.23782569e-01 8.64152372e-01 5.90068936e-01
-3.75062451e-02 -6.18867934e-01 -1.16562545e+00 -1.35124540e+00
-1.30711541e-01 -6.17958486e-01 5.65662742e-01 9.81147647e-01
-3.99031818e-01 5.96448779e-01 -2.85567969e-01 5.03980443e-02
1.05118120e+00 2.66389340e-01 9.91152883e-01 -1.76612735e+00
8.24392810e-02 -1.52129859e-01 -1.14665079e+00 -8.30150366e-01
1.44121582e-02 -9.17004466e-01 3.72292280e-01 -1.42913783e+00
1.31539240e-01 -5.70971131e-01 -4.00894612e-01 4.23889250e-01
1.02289058e-01 3.04785877e-01 4.96221066e-01 3.37600768e-01
-7.37164199e-01 7.40716636e-01 1.27617931e+00 -4.22101617e-01
-1.78956017e-02 1.27204776e-01 -1.31260008e-01 6.58002079e-01
4.64144677e-01 -5.36285579e-01 -2.66033739e-01 -1.75335079e-01
-1.16060406e-01 -2.93015122e-01 7.50616014e-01 -1.30545318e+00
3.77905786e-01 -6.30484596e-02 4.21712101e-01 -8.83727312e-01
6.25222921e-01 -1.28428626e+00 -6.10510968e-02 2.35670462e-01
1.52300432e-01 -5.68767548e-01 -7.04019740e-02 6.99932039e-01
-2.05070302e-01 -3.20466280e-01 1.13276243e+00 -3.71407807e-01
-1.16001928e+00 6.89855576e-01 -4.85483930e-03 1.11116856e-01
1.23031223e+00 -5.50548971e-01 -1.76717460e-01 -1.09575763e-01
-5.52758038e-01 2.22499922e-01 6.31729245e-01 7.92549789e-01
9.22885478e-01 -1.47816789e+00 -4.69860017e-01 4.24310803e-01
7.51735270e-01 4.75486547e-01 2.48471439e-01 7.92318821e-01
-7.21991658e-02 2.82368124e-01 -3.90324473e-01 -1.10255253e+00
-9.64196444e-01 7.87954926e-01 2.66734093e-01 4.57200296e-02
-8.66510391e-01 8.88392746e-01 1.58036813e-01 -3.50867838e-01
4.29043770e-01 3.79143911e-03 -1.03293501e-01 2.18369916e-01
4.98149633e-01 3.42255831e-01 2.73227632e-01 -6.98187888e-01
-5.93606055e-01 7.08902001e-01 -1.43172434e-02 2.58677781e-01
1.54904437e+00 4.27187830e-02 2.06126153e-01 8.86665225e-01
1.20738995e+00 -8.90568912e-01 -1.35643327e+00 -5.40290833e-01
6.78758770e-02 -7.93936074e-01 1.36482209e-01 -5.50586104e-01
-9.28779721e-01 1.17064536e+00 9.17940497e-01 2.94326276e-01
8.34376276e-01 3.69861811e-01 6.54832065e-01 6.03450656e-01
6.85394168e-01 -1.10937166e+00 1.85000360e-01 8.50546122e-01
7.62876511e-01 -1.82578385e+00 -2.76003212e-01 -1.30945995e-01
-4.71715897e-01 7.52120316e-01 7.46453881e-01 -3.33556056e-01
7.95840800e-01 -1.11579254e-01 -7.19911680e-02 -4.90990371e-01
-6.64683282e-01 -5.79987466e-01 2.55215049e-01 8.07335198e-01
-2.97996163e-01 -1.40853316e-01 4.06044722e-01 3.03070515e-01
2.24057034e-01 -1.07865207e-01 2.80950427e-01 1.09229112e+00
-6.93819702e-01 -7.73030877e-01 -4.07944411e-01 6.17205620e-01
6.59018904e-02 3.33493173e-01 -6.16720796e-01 6.42987132e-01
2.91190743e-01 9.27503884e-01 9.89937559e-02 -5.10801554e-01
5.94642699e-01 2.06003442e-01 6.22052610e-01 -8.93962443e-01
-1.16475493e-01 -3.27394664e-01 -2.85888195e-01 -5.97040236e-01
-4.89484519e-01 -4.20975327e-01 -9.76207793e-01 -3.16776708e-02
-4.97045338e-01 -2.12677971e-01 8.70529592e-01 9.25367296e-01
3.51848543e-01 2.92620391e-01 7.04783559e-01 -1.29918587e+00
-7.19725251e-01 -9.29377198e-01 -8.06957424e-01 6.82161570e-01
3.91465634e-01 -1.39662671e+00 -5.86353183e-01 -3.26609820e-01] | [8.014104843139648, -3.216585874557495] |
85fe8ea8-613c-4e76-a220-a764e823fc89 | flat-and-nested-negation-and-uncertainty | null | null | https://openreview.net/forum?id=cA5yHxUxYyz | https://openreview.net/pdf?id=cA5yHxUxYyz | Flat and Nested Negation and Uncertainty Detection with PubMed BERT | Negation and uncertainty detection is an oft-studied challenge in biomedical NLP. Annotation style for the task has not been standardized and as such, the existing datasets not only vary in domain but require various algorithmic designs due to their structural differences. We present a new negation detection dataset in two versions from clinical publications. We further developed two BERT-based models to evaluate on each dataset version. Both models treat the task as a token-level multi-class classification task, one of which is capable of assigning more than one label per token in the case of recursive nesting. Our models achieve F1 scores of 76% and 72% on the development and test sets, respectively. | ['Anonymous'] | 2021-12-17 | null | null | null | acl-arr-december-2022-12 | ['negation-detection'] | ['natural-language-processing'] | [ 2.23598942e-01 2.68253982e-01 -4.29257929e-01 -7.29586184e-01
-8.17628980e-01 -7.16830909e-01 2.52903074e-01 6.96152985e-01
-6.03966653e-01 1.16424131e+00 3.68715227e-02 -2.86393017e-01
-3.00347321e-02 -4.31907952e-01 -3.38386923e-01 -4.73616749e-01
1.37919888e-01 6.59301102e-01 1.69670701e-01 9.77976918e-02
-9.16099399e-02 5.40748890e-03 -1.06104815e+00 8.03738415e-01
8.96966517e-01 8.73417735e-01 -4.27356005e-01 5.24510682e-01
-2.13118464e-01 7.12203205e-01 -7.89319158e-01 -7.04240143e-01
-2.55964577e-01 -3.25273424e-01 -9.53281403e-01 -4.87994045e-01
1.05355725e-01 2.41753563e-01 1.73724949e-01 1.25704908e+00
6.47140026e-01 -3.60435903e-01 1.01768982e+00 -1.07846165e+00
-3.79567921e-01 1.16525280e+00 -5.05649567e-01 2.60315448e-01
4.84206259e-01 -2.37121299e-01 1.23799384e+00 -7.76600361e-01
9.03365374e-01 1.10621917e+00 7.91229188e-01 7.25236773e-01
-1.22657526e+00 -6.36630118e-01 -1.24164157e-01 4.41894680e-02
-1.43471158e+00 -2.18153775e-01 3.62525791e-01 -5.59776008e-01
1.09449887e+00 1.62250578e-01 2.98019081e-01 1.15000594e+00
6.28459990e-01 6.84837461e-01 1.12236130e+00 -2.18070239e-01
3.64941716e-01 1.09904282e-01 2.47011423e-01 5.76216638e-01
5.02603054e-01 -2.02403858e-01 -4.15977865e-01 -4.19176072e-01
1.14374310e-01 -3.18056911e-01 -2.23670542e-01 1.87835634e-01
-1.19024205e+00 7.98642635e-01 -3.88636626e-02 5.33326924e-01
-2.02278331e-01 -1.35190710e-01 8.73972356e-01 2.76731491e-01
4.17968184e-01 5.07722497e-01 -1.05031538e+00 -2.23313853e-01
-9.63351905e-01 9.39711854e-02 1.01294756e+00 9.36419964e-01
2.37741143e-01 -3.78358066e-01 -4.49231774e-01 7.65715599e-01
3.37284297e-01 3.62925790e-02 7.08430648e-01 -6.07328832e-01
3.73035371e-01 6.32762134e-01 -1.79872319e-01 -7.16852248e-01
-8.83483768e-01 -4.62696373e-01 -9.76067543e-01 -2.84199893e-01
4.68866646e-01 -2.89822131e-01 -8.10812354e-01 1.86151481e+00
2.85194278e-01 1.64281711e-01 3.53830755e-01 4.30275500e-01
1.28608167e+00 1.55856147e-01 4.67807293e-01 -3.61907065e-01
1.80049515e+00 -7.14229763e-01 -1.37176609e+00 -8.68104119e-03
1.07557869e+00 -7.38744974e-01 5.43138206e-01 7.81510532e-01
-9.01780665e-01 -9.13736038e-03 -9.80712414e-01 -1.65274918e-01
-6.31469488e-01 1.51112244e-01 5.97996294e-01 9.30924535e-01
-7.06379294e-01 4.38935697e-01 -5.89157999e-01 -3.29808071e-02
5.47363162e-01 2.52513796e-01 -5.62180042e-01 8.17188099e-02
-1.77040017e+00 1.01644623e+00 7.03433335e-01 2.43231773e-01
-5.25059342e-01 -8.23391676e-01 -9.58483636e-01 -7.92119801e-02
3.68551254e-01 -6.57304108e-01 1.21958351e+00 -4.24574852e-01
-1.12705016e+00 1.26507926e+00 -9.29846615e-02 -3.34265590e-01
7.41936743e-01 -3.89342345e-02 -7.59793878e-01 -1.52066499e-01
4.56996053e-01 5.80243468e-01 3.53089631e-01 -7.76239395e-01
-7.33202398e-01 -1.52186662e-01 -1.42590076e-01 -9.47668962e-03
4.28693481e-02 1.45027444e-01 -2.31864393e-01 -7.51829207e-01
-3.01567223e-02 -6.19550347e-01 -1.50407016e-01 -8.76359046e-02
-7.33070433e-01 -4.85877156e-01 4.51114237e-01 -3.83542150e-01
1.44072580e+00 -2.10612392e+00 -9.90181714e-02 -8.70223902e-03
3.06740105e-01 1.60086542e-01 2.03365296e-01 -2.80413404e-02
-3.76325518e-01 6.91927016e-01 -3.71208340e-01 -1.82365611e-01
-4.36283275e-02 3.92137319e-01 9.56785530e-02 2.72394031e-01
5.42793095e-01 6.42401218e-01 -1.27578282e+00 -8.68500352e-01
-3.70531887e-01 1.81580350e-01 -5.39116740e-01 -2.42503703e-01
-3.46641928e-01 2.57834703e-01 -4.28338408e-01 1.06076670e+00
6.25273287e-01 -8.49650428e-02 4.55343246e-01 -1.94744855e-01
2.09475085e-01 3.39769810e-01 -1.26112235e+00 1.71027946e+00
5.80759197e-02 3.50690223e-02 1.42390002e-02 -8.24813604e-01
4.29072976e-01 7.27791786e-01 3.39392424e-01 -1.60335943e-01
1.70569971e-01 4.60716754e-01 2.65338719e-01 -6.50345087e-01
2.29726210e-01 -4.15201306e-01 -5.05571961e-01 -3.30678150e-02
1.16354071e-01 -5.29049113e-02 5.07449687e-01 7.43024983e-03
1.30883300e+00 -3.44761349e-02 8.12436044e-01 -3.10632616e-01
5.90817809e-01 -2.15617884e-02 1.29750907e+00 7.29210973e-01
-5.00337124e-01 5.79050303e-01 1.17835212e+00 -3.37087452e-01
-4.24391270e-01 -7.96591520e-01 -8.66399467e-01 7.69306004e-01
-4.92124557e-01 -4.47181821e-01 -4.21053022e-01 -1.16935527e+00
1.88345730e-01 6.16534352e-01 -7.69403160e-01 -6.59010410e-02
-3.34074914e-01 -1.22128320e+00 1.14473796e+00 3.49749744e-01
1.17653154e-01 -1.12746429e+00 -3.47539365e-01 3.43913227e-01
-3.01472485e-01 -1.49715126e+00 -3.19534391e-01 6.81343019e-01
-7.86255121e-01 -1.33748436e+00 -2.93215990e-01 -8.97987306e-01
5.13768077e-01 -8.31369460e-01 1.35770631e+00 -1.43768206e-01
-2.63261795e-01 -1.74963430e-01 -3.27613920e-01 -7.26051271e-01
-4.03299540e-01 1.14828914e-01 1.88794315e-01 -4.22300935e-01
6.79419458e-01 -7.69388601e-02 -3.01488161e-01 1.07608497e-01
-1.24567330e+00 -3.24740589e-01 6.54708743e-01 1.14171696e+00
8.41940820e-01 4.12479676e-02 1.00558138e+00 -1.47762108e+00
7.47233331e-01 -6.82170570e-01 -1.57253399e-01 2.97951192e-01
-6.03877485e-01 5.57700917e-02 4.01279300e-01 -1.86512217e-01
-7.22336411e-01 1.97160169e-01 -3.37413639e-01 1.18853293e-01
-2.49517366e-01 1.08125520e+00 -2.13078022e-01 4.22183216e-01
5.68899989e-01 -3.42617065e-01 -3.39754105e-01 -2.19746351e-01
2.68286288e-01 7.39883423e-01 3.78234476e-01 -5.13853192e-01
1.30785704e-01 6.18970096e-02 2.36276276e-02 -5.31418383e-01
-1.07486737e+00 -2.64246345e-01 -4.81275201e-01 2.72828609e-01
9.95106578e-01 -7.54178047e-01 -6.44510567e-01 4.22891825e-01
-1.20033634e+00 9.15947482e-02 -1.60915509e-01 5.02378821e-01
-1.73800830e-02 4.89032447e-01 -8.07065606e-01 -4.06067848e-01
-4.82689708e-01 -1.21791637e+00 8.85748863e-01 -1.03740199e-02
-5.10855019e-01 -1.09868276e+00 1.61000133e-01 1.27626538e-01
2.85422187e-02 5.42408705e-01 1.35207927e+00 -1.27199280e+00
2.72088557e-01 -3.43940407e-01 -1.36355693e-02 2.94881105e-01
2.13056356e-01 -3.61412056e-02 -9.42490101e-01 -1.12796016e-02
1.66094407e-01 -5.59449852e-01 7.73107469e-01 2.91860193e-01
1.20623863e+00 5.91724366e-02 -6.28629923e-01 6.37825057e-02
1.23361945e+00 6.31035119e-02 5.02614439e-01 1.60704225e-01
2.98806489e-01 4.07101899e-01 7.14071929e-01 2.43175536e-01
1.55227527e-01 4.88770545e-01 3.97237986e-01 1.40732467e-01
2.18674615e-01 1.07361615e-01 1.64747879e-01 5.08403540e-01
3.88738096e-01 -3.28386724e-01 -1.19412208e+00 5.48569977e-01
-1.77422273e+00 -6.29624546e-01 -3.31168860e-01 1.77035451e+00
1.48702896e+00 3.81776571e-01 -3.49187374e-01 1.89556330e-01
5.55503547e-01 -2.89538294e-01 -3.09597045e-01 -5.26334405e-01
-3.43886137e-01 3.00007194e-01 1.28634036e-01 2.76415080e-01
-1.24806154e+00 7.48894811e-01 7.06047010e+00 8.03441226e-01
-7.08937943e-01 2.51741558e-02 8.03006589e-01 3.26951332e-02
-1.91067219e-01 -2.36287832e-01 -1.03953671e+00 5.28817296e-01
9.33111787e-01 9.90462825e-02 -4.07076925e-01 5.42783976e-01
6.52109012e-02 -2.18567207e-01 -1.49859285e+00 7.47664034e-01
-5.58628105e-02 -9.72944915e-01 1.52256534e-01 -2.39694640e-01
7.46192098e-01 -7.08796754e-02 -9.66930389e-02 5.23449361e-01
5.00235483e-02 -1.29519427e+00 5.24025381e-01 4.30506647e-01
8.51836324e-01 -4.73425210e-01 1.47966945e+00 2.74101526e-01
-9.02146041e-01 1.97941393e-01 -1.16802484e-01 1.86415970e-01
2.15946645e-01 1.10679722e+00 -8.11640382e-01 8.34806383e-01
7.22445011e-01 5.77408135e-01 -4.37384367e-01 1.01224458e+00
-4.96841699e-01 5.59055030e-01 -3.46549720e-01 3.79531197e-02
8.88647288e-02 2.48439029e-01 3.36100042e-01 1.62890160e+00
-2.22452879e-02 5.83525561e-02 2.87305862e-01 7.82647014e-01
-3.51473957e-01 2.59845585e-01 -3.81077111e-01 -1.47791356e-01
5.17582297e-01 1.36373448e+00 -5.98682106e-01 -2.02054113e-01
-4.00185943e-01 5.70350945e-01 2.31607690e-01 -1.18096806e-02
-9.45166886e-01 -7.22660422e-01 3.67166489e-01 -3.36630493e-01
1.29470825e-01 3.28567147e-01 -3.82577181e-01 -1.12601161e+00
-6.00941628e-02 -9.10436809e-01 8.41080606e-01 -3.08756620e-01
-1.58193016e+00 5.34468889e-01 -1.50811628e-01 -9.07124579e-01
-2.18711630e-01 -9.31084275e-01 -2.14702561e-01 7.82534003e-01
-1.69579947e+00 -9.74194884e-01 1.40678123e-01 1.64744824e-01
1.10916533e-01 7.45002087e-03 1.32233655e+00 6.32029414e-01
-6.96104169e-01 9.20229554e-01 -5.77395037e-03 4.67827618e-01
1.19492137e+00 -1.58110809e+00 -1.56429052e-01 3.87291193e-01
-2.63839662e-01 7.37875462e-01 6.15494549e-01 -8.01599503e-01
-7.25342810e-01 -1.10456884e+00 1.58318639e+00 -6.70936286e-01
6.16198361e-01 -2.32156351e-01 -9.82080758e-01 7.39097178e-01
-5.17646642e-03 1.88141033e-01 1.19061899e+00 3.67098659e-01
-3.38251948e-01 2.65017122e-01 -1.65609503e+00 1.82281896e-01
7.07685828e-01 -3.09098095e-01 -8.77280474e-01 3.82349312e-01
5.29472947e-01 -8.40431750e-01 -1.20470846e+00 7.65969753e-01
2.59810507e-01 -4.93702531e-01 4.73854810e-01 -8.68253171e-01
5.24393976e-01 -4.73609537e-01 -5.64196520e-02 -1.02669156e+00
-8.42576995e-02 -3.37271094e-01 -9.19691324e-02 1.23478758e+00
1.13038623e+00 -5.75729668e-01 6.12729132e-01 3.83879066e-01
-2.29691163e-01 -9.40397441e-01 -1.16488886e+00 -5.34996450e-01
2.24055588e-01 -5.04633784e-01 2.71247804e-01 1.15820289e+00
4.24393147e-01 5.32233000e-01 8.19992125e-02 1.68189883e-01
1.65788665e-01 -2.81071723e-01 -1.33459806e-01 -1.42803788e+00
-3.40969265e-02 -4.06447530e-01 -4.81476843e-01 -2.48005569e-01
2.53573954e-01 -1.23420012e+00 5.12438379e-02 -1.40392685e+00
4.95323449e-01 -2.39597857e-01 -4.94227082e-01 9.86253619e-01
-2.28245422e-01 1.31694406e-01 -5.95120490e-01 -1.08139694e-01
-6.72643483e-01 1.18051969e-01 7.22784340e-01 -2.39212379e-01
-5.96975684e-02 -8.40504915e-02 -7.41908014e-01 8.32747400e-01
6.01341546e-01 -9.57694709e-01 -2.26323321e-01 -3.27315927e-01
7.84107685e-01 -9.36501399e-02 -2.77265813e-02 -7.56570637e-01
1.85980588e-01 8.66312475e-04 3.33186835e-01 -6.59187853e-01
-1.08515099e-01 -6.31855071e-01 4.59717773e-02 7.46081471e-01
-4.90295678e-01 7.70594701e-02 3.78554344e-01 6.97849095e-01
-3.04533154e-01 -5.70059180e-01 7.19590247e-01 -1.85573608e-01
-3.07023346e-01 -4.69905995e-02 -4.56678331e-01 4.37184721e-01
1.01094818e+00 1.33970708e-01 -3.36603850e-01 1.82968974e-01
-1.20873797e+00 5.48731685e-01 -8.45137089e-02 3.57875556e-01
4.96514946e-01 -1.01564658e+00 -7.91379094e-01 -1.75894901e-01
3.62483889e-01 2.91116923e-01 1.24905882e-02 9.38430429e-01
-3.78650874e-01 6.48193359e-01 1.22519780e-03 -5.98916650e-01
-1.36024201e+00 3.23929399e-01 4.85540658e-01 -7.59097159e-01
-1.09887190e-01 8.10127556e-01 -2.13422418e-01 -7.22653270e-01
3.37386608e-01 -7.41213441e-01 -5.55671751e-01 3.37250710e-01
4.23184335e-01 1.99352294e-01 4.16564375e-01 -2.33564958e-01
-6.93530858e-01 2.49877363e-01 -4.77865160e-01 1.77547410e-01
1.07284844e+00 3.78792644e-01 -5.86252630e-01 5.74922681e-01
9.17637825e-01 -1.28043085e-01 -1.75528541e-01 -1.86576366e-01
4.52688366e-01 3.17698807e-01 -8.93363655e-02 -1.39914584e+00
-7.45690107e-01 4.79161680e-01 4.08466995e-01 -9.84436949e-04
7.54354656e-01 -1.37575582e-01 3.26115608e-01 4.12400097e-01
1.98000520e-01 -1.24991906e+00 -2.43133381e-01 8.11075866e-01
6.63770616e-01 -1.29581666e+00 5.40937223e-02 -6.80540383e-01
-5.86228907e-01 9.93247569e-01 5.30255973e-01 5.33893943e-01
7.53753006e-01 4.12253767e-01 1.20119162e-01 -4.04923975e-01
-8.40756059e-01 1.89877316e-01 2.61673808e-01 4.01375026e-01
1.01961279e+00 6.99011460e-02 -9.52315032e-01 1.33710647e+00
9.87999979e-03 2.52704531e-01 4.45397168e-01 1.11265850e+00
-1.09260514e-01 -1.29435408e+00 -3.44563909e-02 7.26858616e-01
-1.15164638e+00 -1.63383737e-01 -3.81854206e-01 6.54628277e-01
4.98254389e-01 9.92328227e-01 -5.36631167e-01 -1.05145715e-01
4.22484607e-01 4.09509838e-01 1.33011147e-01 -1.03929794e+00
-9.04415548e-01 -2.62404740e-01 4.60610121e-01 -3.00794333e-01
-4.54142362e-01 -7.53259361e-01 -1.63848293e+00 1.61984250e-01
-5.45290828e-01 4.69976962e-01 3.45163763e-01 1.08043027e+00
2.80916274e-01 8.43899190e-01 -1.78378355e-02 1.29827574e-01
-7.32255518e-01 -1.09702992e+00 -6.60075545e-01 3.16484749e-01
1.12613551e-01 -6.38576508e-01 -3.53415489e-01 -9.26472098e-02] | [8.560747146606445, 8.760224342346191] |
2bd69e96-e3e6-4bec-a39f-7f45c169dc1f | paradiseo-from-a-modular-framework-for | 2105.00420 | null | https://arxiv.org/abs/2105.00420v1 | https://arxiv.org/pdf/2105.00420v1.pdf | Paradiseo: From a Modular Framework for Evolutionary Computation to the Automated Design of Metaheuristics ---22 Years of Paradiseo--- | The success of metaheuristic optimization methods has led to the development of a large variety of algorithm paradigms. However, no algorithm clearly dominates all its competitors on all problems. Instead, the underlying variety of landscapes of optimization problems calls for a variety of algorithms to solve them efficiently. It is thus of prior importance to have access to mature and flexible software frameworks which allow for an efficient exploration of the algorithm design space. Such frameworks should be flexible enough to accommodate any kind of metaheuristics, and open enough to connect with higher-level optimization, monitoring and evaluation softwares. This article summarizes the features of the ParadisEO framework, a comprehensive C++ free software which targets the development of modular metaheuristics. ParadisEO provides a highly modular architecture, a large set of components, speed of execution and automated algorithm design features, which are key to modern approaches to metaheuristics development. | ['Jan Gmys', 'Benjamin Bouvier', 'Alexandre Quemy', 'Juan J. Merelo', 'Marc Schoenauer', 'Sébastien Verel', 'Arnaud Liefooghe', 'Johann Dreo'] | 2021-05-02 | null | null | null | null | ['metaheuristic-optimization'] | ['methodology'] | [ 1.32877558e-01 -4.73724365e-01 -8.05016086e-02 -1.55884698e-01
-4.63410854e-01 -9.87152457e-01 5.12364745e-01 2.28896379e-01
-2.64935106e-01 7.03238010e-01 -4.69758451e-01 -5.88370025e-01
-7.14642406e-01 -1.15449357e+00 -1.03763208e-01 -8.42432261e-01
-3.37119520e-01 6.32359028e-01 2.54761666e-01 -6.18183792e-01
6.34214818e-01 8.22361588e-01 -2.35414815e+00 -3.81191052e-03
9.15340483e-01 8.28041255e-01 1.50722042e-01 8.50105703e-01
-4.24215823e-01 -1.54000312e-01 -6.54467702e-01 -2.03423813e-01
1.17816180e-01 -3.00257504e-01 -7.22367883e-01 -1.78380325e-01
-2.78526098e-01 6.77344918e-01 3.35910112e-01 7.73553967e-01
6.67612791e-01 3.14404428e-01 3.50696743e-01 -1.34286392e+00
-7.48310760e-02 2.01130778e-01 -4.60237801e-01 2.00904995e-01
5.21062434e-01 2.86360413e-01 1.06525147e+00 -2.49262407e-01
6.44248903e-01 1.04497075e+00 4.58345324e-01 2.74947464e-01
-1.14920497e+00 -1.70712024e-01 -2.59675719e-02 9.80198309e-02
-1.11212766e+00 -3.57012630e-01 3.72662216e-01 5.63293882e-02
1.38095450e+00 1.27301478e+00 9.89018559e-01 5.60898840e-01
3.23470980e-01 6.67567253e-01 9.87126648e-01 -6.45091355e-01
6.35589957e-01 -7.19507318e-03 -9.96719375e-02 4.27498132e-01
5.20096421e-01 3.43834996e-01 -3.52708548e-01 -4.34351176e-01
1.34745210e-01 -1.60884261e-01 -6.99842349e-02 -8.94349337e-01
-9.27878976e-01 8.17910016e-01 1.29420519e-01 4.87717360e-01
-3.18020344e-01 1.06941700e-01 3.71394932e-01 2.28495598e-01
-9.87143964e-02 1.03125131e+00 -5.08108795e-01 -7.54967093e-01
-7.05507636e-01 7.05938578e-01 1.05398452e+00 6.64701164e-01
5.88925838e-01 -1.32927820e-01 2.77708888e-01 6.87521100e-01
3.09761971e-01 3.08053941e-01 1.59500808e-01 -1.09476864e+00
2.05324665e-01 1.02885222e+00 2.55983174e-01 -1.10303223e+00
-6.55105293e-01 -8.48086894e-01 -3.32389712e-01 6.13076210e-01
1.37003332e-01 -4.21368219e-02 -4.74747926e-01 1.26900756e+00
8.07438731e-01 -6.92845941e-01 -8.40326920e-02 6.36817157e-01
2.56697774e-01 6.96007192e-01 -3.26424927e-01 -3.54753435e-01
1.05536926e+00 -8.51287484e-01 -5.02731264e-01 -1.70461789e-01
5.81839383e-01 -1.02778590e+00 9.04674590e-01 7.04909623e-01
-1.36195362e+00 -1.52430981e-02 -1.13413870e+00 3.74341071e-01
-1.02016878e+00 -6.31305158e-01 1.12897658e+00 1.31986928e+00
-1.05237508e+00 5.40926635e-01 -8.31318498e-01 -5.04792452e-01
1.34797290e-01 8.18633676e-01 9.75045189e-02 -1.33655695e-02
-7.00805724e-01 9.65103686e-01 8.23964000e-01 1.83180511e-01
-9.28891301e-02 -4.63451296e-01 -7.39359021e-01 1.01293862e-01
3.73283267e-01 -8.87400985e-01 8.18994284e-01 -9.59944606e-01
-1.48534918e+00 7.46677816e-01 1.00103579e-01 4.79087234e-02
4.14884478e-01 3.35930973e-01 -4.34704721e-01 -2.90360123e-01
-1.63853556e-01 3.01851988e-01 4.27646250e-01 -1.04239655e+00
-1.01979375e+00 -2.81604588e-01 1.16224941e-02 3.56434941e-01
-2.89047539e-01 1.36822149e-01 -4.01333570e-01 -2.97764331e-01
-2.24012241e-01 -7.33398855e-01 -6.35289490e-01 -4.25044090e-01
-1.23918680e-02 1.07851423e-01 7.07415283e-01 1.10743165e-01
1.59416509e+00 -1.97829318e+00 3.08228046e-01 6.34118736e-01
-1.47178307e-01 1.93494931e-01 -1.99866489e-01 8.95095646e-01
-4.79488960e-03 1.78324714e-01 -3.32380503e-01 3.13542962e-01
5.72583437e-01 2.98854291e-01 1.23918936e-01 5.24304211e-01
-1.09668933e-01 8.21271956e-01 -1.09421194e+00 -7.92056695e-02
6.21944606e-01 2.34042227e-01 -6.60585344e-01 -1.35450825e-01
-3.73209417e-01 -1.05342820e-01 -7.21463799e-01 1.01282418e+00
8.15806746e-01 2.26210803e-02 4.94272113e-01 5.26166439e-01
-9.75278199e-01 4.81868535e-02 -1.47792768e+00 1.59579146e+00
-3.60563189e-01 3.43033522e-01 5.10764003e-01 -8.24545979e-01
6.97568059e-01 6.10823520e-02 7.80056536e-01 -6.68581188e-01
3.08919728e-01 6.46418452e-01 1.12008667e-02 -3.11443210e-01
5.57585359e-01 3.62511188e-01 -1.00478373e-01 6.01172864e-01
-3.88141245e-01 -4.35689807e-01 8.92927051e-01 -5.43968737e-01
1.24992681e+00 1.39684975e-01 4.54154551e-01 -5.16483963e-01
6.56510353e-01 5.52936554e-01 3.95038575e-01 5.09817362e-01
4.64753620e-02 7.36354738e-02 2.15978026e-01 -8.46813262e-01
-8.91468108e-01 -7.92988420e-01 -4.00363594e-01 1.20454872e+00
1.25922769e-01 -5.81558287e-01 -6.95362628e-01 -1.34599343e-01
-9.22631025e-02 6.78682148e-01 -6.08452022e-01 9.33290496e-02
-3.37069571e-01 -1.43099940e+00 1.95254490e-01 -2.05380246e-01
1.85509369e-01 -1.11703420e+00 -1.34842813e+00 4.91935343e-01
2.20420882e-01 -3.06380630e-01 9.05166194e-02 4.09672558e-01
-9.82556462e-01 -1.17511725e+00 -2.15138167e-01 -5.67712843e-01
3.87623042e-01 2.91313797e-01 1.27722561e+00 3.40790123e-01
-9.32169080e-01 4.62549508e-01 -3.52272540e-01 -4.75502431e-01
-2.91347325e-01 3.99198741e-01 -4.53529567e-01 -3.71886432e-01
2.25877121e-01 -7.26611018e-01 -4.35920596e-01 3.81707251e-01
-1.06438816e+00 -2.65786201e-01 5.10894120e-01 7.19797075e-01
6.60219729e-01 4.44351912e-01 3.37737203e-02 -3.67264718e-01
8.04397523e-01 -5.06577969e-01 -1.06803000e+00 5.11584699e-01
-7.64064431e-01 9.34583023e-02 3.50861728e-01 7.44186938e-02
-5.46914697e-01 -2.56687105e-02 -2.54279613e-01 3.79112661e-01
-3.31918567e-01 5.59378803e-01 -4.16056275e-01 -5.16282737e-01
4.95704651e-01 7.77391866e-02 -2.70024538e-02 -3.28848392e-01
3.31776947e-01 4.75043327e-01 5.47233783e-02 -7.67201841e-01
5.83519518e-01 3.26665431e-01 1.65862486e-01 -9.71170723e-01
-1.50659708e-02 -5.15458107e-01 -1.56606510e-01 -1.28424555e-01
5.02556026e-01 1.57822311e-01 -8.95734310e-01 2.99204290e-01
-9.31147933e-01 -1.85153365e-01 -3.49018008e-01 -1.74341455e-01
-5.28986454e-01 1.10534139e-01 3.60019952e-01 -8.69386733e-01
-2.95540750e-01 -1.54348242e+00 7.16251731e-01 6.04216039e-01
-4.26921248e-01 -1.15137529e+00 5.45783639e-01 1.45275906e-01
6.76540613e-01 7.80512512e-01 8.98694336e-01 -3.99126440e-01
-6.74718142e-01 -3.38858485e-01 2.65915394e-01 -4.58443493e-01
6.96995929e-02 7.42069483e-01 -6.49307609e-01 -3.18665236e-01
-1.97810367e-01 1.78540424e-02 2.58447468e-01 5.24327338e-01
8.51768315e-01 -1.11715093e-01 -6.78343058e-01 8.95539165e-01
1.61805117e+00 5.59687734e-01 7.52641797e-01 1.26513100e+00
-1.41162932e-01 6.99204147e-01 6.97907150e-01 5.51550567e-01
2.11539328e-01 7.63166726e-01 8.70774329e-01 -8.33976716e-02
5.44066012e-01 5.50532460e-01 1.10025577e-01 4.51978564e-01
-1.88388616e-01 -5.16689241e-01 -1.15383899e+00 3.25282633e-01
-1.78608537e+00 -1.09272051e+00 -3.13133925e-01 2.29076147e+00
3.03462952e-01 -4.87200432e-02 3.22977692e-01 4.51948136e-01
6.26535118e-01 3.99463810e-02 -4.59600568e-01 -1.10716212e+00
-2.80358363e-02 3.11537862e-01 3.17554802e-01 3.49027514e-01
-9.29459870e-01 5.99296987e-01 7.36869240e+00 8.45020056e-01
-9.46503520e-01 -2.24740818e-01 2.47174755e-01 -3.32967550e-01
-5.37772536e-01 -4.13892046e-02 -4.20676768e-01 3.58646184e-01
1.01098299e+00 -4.22754675e-01 8.04051220e-01 6.94315910e-01
1.94295898e-01 -3.32513452e-01 -5.72521508e-01 6.84775889e-01
-3.27776074e-01 -1.69888020e+00 -2.94277877e-01 3.36633593e-01
8.77322555e-01 1.02946676e-01 5.22323996e-02 7.02047870e-02
2.99274117e-01 -1.23302710e+00 5.56104779e-01 3.30641642e-02
1.78969696e-01 -1.30078137e+00 6.55799031e-01 3.58072035e-02
-1.10442388e+00 -2.86112994e-01 -5.37917502e-02 -2.76546776e-02
1.26566529e-01 3.24699521e-01 -1.58061728e-01 8.72274458e-01
9.95779991e-01 2.11413771e-01 -5.63208580e-01 1.80167031e+00
4.31303322e-01 -6.97025433e-02 -6.13299072e-01 -5.14973879e-01
5.19596159e-01 -4.39903170e-01 7.51636922e-01 1.05195212e+00
2.18529403e-01 -3.42257768e-01 4.35882397e-02 4.96960908e-01
5.93488157e-01 4.95002925e-01 -4.78159666e-01 -2.82819718e-01
4.69795436e-01 1.35100150e+00 -1.09871149e+00 3.65069330e-01
-3.05972874e-01 4.81687248e-01 8.54429528e-02 1.83958188e-01
-8.31371248e-01 -7.93643355e-01 8.50624025e-01 -6.42266348e-02
2.04615057e-01 -2.30702609e-01 -5.53438425e-01 -5.87643087e-01
-4.68286984e-02 -1.39211285e+00 6.96398437e-01 -2.31334135e-01
-7.48830676e-01 6.44481897e-01 -2.53449947e-01 -7.74628520e-01
-2.09117979e-01 -7.81614602e-01 -7.37268209e-01 5.66835165e-01
-1.47147369e+00 -7.75657117e-01 -2.08260521e-01 1.84534416e-01
2.55723745e-01 -1.75351337e-01 1.15487683e+00 1.16122395e-01
-6.95516944e-01 9.40979198e-02 6.31384015e-01 -7.49852479e-01
2.08800882e-01 -1.26019835e+00 2.23538801e-01 6.70571804e-01
-1.83637559e-01 7.64320314e-01 1.08095586e+00 -1.08545497e-01
-2.03534746e+00 -4.73421067e-01 3.88388962e-01 -4.68804717e-01
7.68295467e-01 -1.08989082e-01 -3.39882970e-01 1.15585461e-01
2.35800430e-01 -5.07014275e-01 8.75257730e-01 4.03874665e-01
2.47466043e-01 -1.64506048e-01 -1.08011067e+00 7.05745995e-01
7.74239957e-01 2.33065143e-01 -2.46598199e-01 2.02884629e-01
-8.12361911e-02 -5.07533252e-01 -7.56481588e-01 3.30458999e-01
6.38933003e-01 -1.51564503e+00 1.17305064e+00 -3.13259274e-01
-7.00771138e-02 -4.53304410e-01 -5.77646270e-02 -9.59277689e-01
-1.91523865e-01 -1.29415047e+00 -1.57325730e-01 1.02676904e+00
5.94821095e-01 -9.10289168e-01 7.58252561e-01 7.50472426e-01
-2.43056536e-01 -1.01319742e+00 -1.00614250e+00 -7.98935354e-01
-5.49274273e-02 -4.77372855e-01 1.17599380e+00 6.32939756e-01
1.36398807e-01 -1.73668951e-01 3.63944381e-01 3.93869542e-02
5.00568748e-01 5.91727197e-01 8.04868400e-01 -1.28176796e+00
-3.52890611e-01 -1.29116809e+00 -5.04249215e-01 -2.95530558e-01
-4.30203617e-01 -6.82338536e-01 -3.41913462e-01 -1.61855328e+00
-2.20062941e-01 -7.09364831e-01 -2.42574856e-01 2.09672540e-01
1.04953259e-01 2.25220069e-01 -2.75205523e-01 -1.71953514e-01
-6.88632667e-01 4.04232234e-01 1.03888750e+00 -4.67808731e-02
-5.55539727e-01 1.01036631e-01 -6.56375885e-01 5.89830160e-01
8.70905757e-01 -5.11844158e-01 -3.79758358e-01 -4.12190884e-01
7.94572592e-01 -3.60146195e-01 1.45109043e-01 -1.12839115e+00
3.36494409e-02 -7.55868256e-01 1.36478215e-01 -3.71598572e-01
1.25518858e-01 -9.97085214e-01 5.82072377e-01 6.21785343e-01
1.75670773e-01 5.15515208e-01 4.85176861e-01 1.29333809e-01
-1.88676685e-01 -3.24877888e-01 7.95730770e-01 -2.03998178e-01
-6.89578056e-01 8.14777315e-02 -5.89541852e-01 2.10219949e-01
1.61879098e+00 -8.19876432e-01 -1.92936182e-01 2.83699721e-01
-3.65529835e-01 4.37244505e-01 1.13007569e+00 4.55952913e-01
9.27873701e-02 -7.61765182e-01 -2.19223872e-01 2.29433268e-01
8.97084028e-02 -2.11635977e-01 1.15748651e-01 5.30251205e-01
-1.07147288e+00 4.43696260e-01 -5.09470046e-01 -3.99875671e-01
-1.16670418e+00 6.45159245e-01 5.90306818e-01 -2.87582815e-01
-2.31954768e-01 6.88638687e-01 -5.52745283e-01 -6.24355614e-01
4.31857631e-02 1.33771211e-01 1.08676843e-01 7.54012465e-02
6.97100878e-01 7.48488963e-01 3.64115864e-01 -2.21193612e-01
-7.65960515e-01 5.13307333e-01 2.64456332e-01 -1.54843375e-01
1.83498418e+00 1.08000576e-01 -6.13983393e-01 -1.25736728e-01
6.24982953e-01 1.47586420e-01 -7.73282468e-01 9.49192286e-01
3.46635193e-01 -4.83685523e-01 2.08298534e-01 -1.02332878e+00
-8.35884273e-01 5.88297904e-01 5.65959692e-01 5.61100841e-01
1.42487729e+00 -6.47099197e-01 2.96674520e-01 3.83636147e-01
7.21087337e-01 -1.20526016e+00 -4.77019072e-01 3.90077561e-01
7.30038047e-01 -5.75609684e-01 2.25908846e-01 -1.66061789e-01
-1.68118011e-02 1.40666389e+00 2.18492508e-01 1.53314725e-01
2.94592738e-01 6.32304966e-01 6.48571029e-02 -4.37091708e-01
-7.19212413e-01 -3.80224347e-01 1.81612521e-01 8.14181745e-01
3.60389382e-01 -2.52368838e-01 -5.58507562e-01 1.41630560e-01
-2.06708953e-01 -1.17502220e-01 1.17324784e-01 1.37036383e+00
-4.92035598e-01 -1.86474061e+00 -8.22366178e-01 5.77615090e-02
-3.49759430e-01 1.32384509e-01 -4.95932847e-01 1.02938759e+00
1.95194080e-01 1.04951441e+00 -3.15299213e-01 7.56965876e-02
2.76811779e-01 -1.16720267e-01 5.58658004e-01 -2.18287572e-01
-1.13546968e+00 -1.48157135e-01 2.03097299e-01 -7.15126276e-01
-8.54937211e-02 -8.15106392e-01 -9.89605904e-01 -7.37315893e-01
-3.28699738e-01 7.34368563e-01 9.63518977e-01 6.89341426e-01
5.74988782e-01 4.91387099e-01 4.09811318e-01 -1.07931566e+00
-1.13543428e-01 7.46800601e-02 -5.73542714e-01 -3.48418415e-01
-9.19398069e-02 -7.53463209e-01 2.58986074e-02 -5.44738352e-01] | [5.786349296569824, 3.6255075931549072] |
a345486b-1512-463d-9594-5a1319942a1e | class-balanced-pixel-level-self-labeling-for | 2203.09744 | null | https://arxiv.org/abs/2203.09744v1 | https://arxiv.org/pdf/2203.09744v1.pdf | Class-Balanced Pixel-Level Self-Labeling for Domain Adaptive Semantic Segmentation | Domain adaptive semantic segmentation aims to learn a model with the supervision of source domain data, and produce satisfactory dense predictions on unlabeled target domain. One popular solution to this challenging task is self-training, which selects high-scoring predictions on target samples as pseudo labels for training. However, the produced pseudo labels often contain much noise because the model is biased to source domain as well as majority categories. To address the above issues, we propose to directly explore the intrinsic pixel distributions of target domain data, instead of heavily relying on the source domain. Specifically, we simultaneously cluster pixels and rectify pseudo labels with the obtained cluster assignments. This process is done in an online fashion so that pseudo labels could co-evolve with the segmentation model without extra training rounds. To overcome the class imbalance problem on long-tailed categories, we employ a distribution alignment technique to enforce the marginal class distribution of cluster assignments to be close to that of pseudo labels. The proposed method, namely Class-balanced Pixel-level Self-Labeling (CPSL), improves the segmentation performance on target domain over state-of-the-arts by a large margin, especially on long-tailed categories. | ['Lei Zhang', 'Xu Jia', 'Yabin Zhang', 'Chenhang He', 'Shuai Li', 'Ruihuang Li'] | 2022-03-18 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Li_Class-Balanced_Pixel-Level_Self-Labeling_for_Domain_Adaptive_Semantic_Segmentation_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Li_Class-Balanced_Pixel-Level_Self-Labeling_for_Domain_Adaptive_Semantic_Segmentation_CVPR_2022_paper.pdf | cvpr-2022-1 | ['synthetic-to-real-translation'] | ['computer-vision'] | [ 6.69440031e-01 4.23137963e-01 -5.51256657e-01 -7.33106434e-01
-1.01435912e+00 -6.00572169e-01 2.72075921e-01 -9.17084292e-02
-2.83066571e-01 7.46336579e-01 -8.13945681e-02 1.85365960e-01
2.76068032e-01 -6.03544056e-01 -7.38865137e-01 -1.01456511e+00
7.04602420e-01 7.73293197e-01 5.78290522e-01 2.63748407e-01
2.27451831e-01 2.47531082e-03 -1.45092750e+00 2.35023096e-01
1.38236988e+00 1.01463473e+00 3.62029910e-01 1.26504466e-01
-4.53767806e-01 5.84502935e-01 -3.93396646e-01 -2.53174394e-01
3.63563210e-01 -6.07270896e-01 -5.72859228e-01 7.17353463e-01
5.02501190e-01 7.25919530e-02 9.66506824e-02 1.53417802e+00
9.03237686e-02 8.37647449e-03 1.02171075e+00 -1.25272882e+00
-5.00407040e-01 5.26517093e-01 -1.10942829e+00 -1.26461804e-01
-4.52965260e-01 2.20039904e-01 9.97999072e-01 -6.88969135e-01
6.35449767e-01 9.94943380e-01 5.10706961e-01 6.81751847e-01
-1.47459245e+00 -9.17405546e-01 3.97469789e-01 9.40755978e-02
-1.42079508e+00 -2.47062534e-01 9.87338126e-01 -4.48288441e-01
-1.00617528e-01 -7.83992112e-02 3.66655141e-01 1.00395048e+00
-1.93156034e-01 9.97648716e-01 1.19611013e+00 -3.56545120e-01
5.43412209e-01 4.31481838e-01 2.34694391e-01 3.58945429e-01
1.61029637e-01 -2.60566026e-01 -4.30325627e-01 4.12825197e-02
5.14186621e-01 -1.02896348e-01 -9.82860178e-02 -7.03784168e-01
-1.08687878e+00 7.53910601e-01 3.60455185e-01 1.46214992e-01
-3.75080585e-01 -1.00853309e-01 1.18679471e-01 -1.35912806e-01
8.25211406e-01 1.01947248e-01 -6.79972947e-01 3.30329239e-01
-1.10971236e+00 8.28194469e-02 3.19924951e-01 9.51504648e-01
1.11988211e+00 -1.47079170e-01 -3.06210637e-01 1.08294439e+00
4.44832653e-01 4.55706984e-01 5.68611205e-01 -1.07223201e+00
4.66816157e-01 7.94547498e-01 -5.65235689e-02 -6.06394649e-01
-8.71333405e-02 -8.65820587e-01 -7.95630753e-01 2.33154044e-01
8.20939720e-01 -2.49985024e-01 -1.49784505e+00 1.98851752e+00
7.46041298e-01 2.47965842e-01 8.38151127e-02 9.27785039e-01
4.76496458e-01 7.09798813e-01 3.71724695e-01 -2.68683225e-01
9.65583026e-01 -1.22568786e+00 -4.82062936e-01 -8.04291785e-01
4.85104561e-01 -5.36321282e-01 1.25045931e+00 2.18203127e-01
-7.10249484e-01 -6.90938175e-01 -7.82656968e-01 1.73288882e-01
7.99649805e-02 1.52192175e-01 2.66646981e-01 4.23970968e-01
-8.11802745e-01 4.21095312e-01 -8.28005850e-01 -1.34652421e-01
9.82582510e-01 1.46800563e-01 -6.48164675e-02 -2.05633566e-01
-8.77487600e-01 3.73374850e-01 7.35714674e-01 -3.58146906e-01
-8.15350175e-01 -7.58502722e-01 -6.88984573e-01 -3.89664918e-02
5.66360295e-01 -2.80809939e-01 1.25255346e+00 -1.69647706e+00
-1.26901472e+00 1.11935985e+00 -2.90569544e-01 -3.47997606e-01
6.35012090e-01 1.06317632e-01 -1.53266519e-01 -1.24702873e-02
5.13931632e-01 1.17198896e+00 1.04390454e+00 -1.60082829e+00
-9.88809645e-01 -6.17613912e-01 -6.11223221e-01 5.46050549e-01
-3.17740798e-01 -6.59837306e-01 -7.23940194e-01 -5.80195904e-01
5.34725666e-01 -1.02686119e+00 -4.16821212e-01 1.32419989e-01
-6.31215334e-01 -1.05112307e-01 7.10793316e-01 -4.40765768e-01
7.72146404e-01 -2.17435980e+00 3.60019766e-02 2.39566818e-01
1.39768124e-01 1.07862897e-01 -1.39146885e-02 -4.47803706e-01
-5.73606603e-02 -3.38311464e-01 -7.74778306e-01 -3.46485645e-01
-1.39592826e-01 2.36231700e-01 -2.22455159e-01 5.07227957e-01
1.56193972e-01 7.25479782e-01 -8.04025471e-01 -8.11131179e-01
1.93410926e-02 1.66956767e-01 -4.54175651e-01 2.19235018e-01
-7.92050421e-01 8.35705578e-01 -5.60911179e-01 6.54194951e-01
9.59837019e-01 -5.03064990e-01 1.21905878e-01 -1.02814741e-01
2.05468550e-01 -1.51983008e-01 -1.12545919e+00 1.67730153e+00
-8.05030614e-02 4.09287214e-01 1.78725988e-01 -1.15293491e+00
1.07824481e+00 -7.74418637e-02 5.38776994e-01 -7.14998722e-01
1.35810718e-01 4.00110394e-01 -9.83907506e-02 -1.33727953e-01
8.66000876e-02 -4.46945220e-01 9.77879241e-02 3.61433238e-01
-2.23489739e-02 -1.47278666e-01 -7.12672323e-02 -1.11869082e-01
5.32414854e-01 2.79529244e-01 8.82337540e-02 -2.13467479e-01
3.52121264e-01 5.07440269e-01 9.88318801e-01 4.19129133e-01
-3.74360740e-01 8.37395191e-01 5.18260181e-01 1.19543225e-01
-1.09496486e+00 -1.06982756e+00 -2.07124978e-01 1.12706065e+00
3.81943017e-01 3.63652408e-01 -1.22636604e+00 -1.06994128e+00
-7.87949935e-03 8.37167919e-01 -5.60014725e-01 -2.86451817e-01
-3.04075927e-01 -9.90164638e-01 7.92619660e-02 4.74154353e-01
6.18241727e-01 -1.02800238e+00 -1.92702338e-01 1.86240166e-01
-3.06118608e-01 -1.04049456e+00 -6.42633855e-01 5.54534793e-01
-9.92296338e-01 -1.00864482e+00 -1.11193287e+00 -1.06477916e+00
1.00985193e+00 2.77703732e-01 9.40647840e-01 -4.89894927e-01
-7.44325574e-04 -2.68900692e-01 -2.86176890e-01 -2.08486229e-01
-4.72810596e-01 2.28137836e-01 -3.26477438e-01 3.55761558e-01
5.14564812e-01 -4.76002038e-01 -7.01792121e-01 5.44046819e-01
-8.64157557e-01 3.84236872e-01 6.89577281e-01 8.89276683e-01
1.20142615e+00 3.34991336e-01 7.65610933e-01 -1.36370659e+00
-8.52047354e-02 -6.73800111e-01 -6.00391507e-01 2.06891641e-01
-7.22460568e-01 6.12265244e-02 5.76633275e-01 -6.26739383e-01
-1.25423300e+00 6.97203934e-01 -1.14984326e-02 -6.09418154e-01
-3.93813521e-01 5.50184138e-02 -4.96542484e-01 1.69044435e-01
7.48068213e-01 3.41116816e-01 -1.65301654e-02 -4.63043332e-01
4.79068965e-01 7.75749564e-01 7.79510915e-01 -3.83547008e-01
8.22362483e-01 6.18683696e-01 -2.73821890e-01 -3.64774466e-01
-1.30292308e+00 -5.86936593e-01 -6.70563519e-01 -6.81311339e-02
1.00080693e+00 -1.09059715e+00 1.73307732e-01 7.80290902e-01
-6.27183259e-01 -7.66746342e-01 -5.04206419e-01 2.15937510e-01
-4.58553940e-01 2.01852307e-01 -2.16214448e-01 -6.72356904e-01
-7.78508261e-02 -9.03135777e-01 1.16108298e+00 6.20811641e-01
5.31740785e-02 -8.86417389e-01 -1.70037046e-01 5.66025913e-01
4.67314087e-02 2.59471625e-01 8.32004547e-01 -7.73381412e-01
-5.27342796e-01 1.44973611e-02 -6.28465116e-01 5.98835528e-01
2.79166520e-01 -3.58003527e-01 -1.04422903e+00 -1.32324293e-01
9.25896466e-02 -3.66235018e-01 9.57659781e-01 7.06698179e-01
1.32493544e+00 -2.69603524e-02 -5.30770838e-01 5.05232036e-01
1.38320017e+00 1.52684957e-01 2.21593648e-01 1.64346531e-01
9.25864279e-01 9.31861639e-01 1.11035788e+00 3.18798125e-01
3.95680219e-01 5.65886557e-01 4.89404261e-01 -2.82692045e-01
-4.71259922e-01 -4.79953915e-01 1.36844050e-02 2.99524158e-01
7.62840748e-01 -3.16584677e-01 -9.91151452e-01 7.63909519e-01
-1.77784657e+00 -4.03334826e-01 -2.22651765e-01 2.22380733e+00
1.13850355e+00 4.87077832e-01 1.55581102e-01 -9.57821608e-02
1.21240294e+00 4.88980114e-03 -1.27865362e+00 3.07954103e-01
-1.95926622e-01 -6.10730909e-02 8.46983552e-01 4.11367029e-01
-1.24964368e+00 1.21065652e+00 5.24042749e+00 1.22129369e+00
-1.17502880e+00 3.25665206e-01 1.33785665e+00 1.15616150e-01
-3.28940988e-01 4.91713034e-03 -8.89455080e-01 8.28538358e-01
5.32397628e-01 7.22459033e-02 1.43742740e-01 1.16449344e+00
9.33900252e-02 -2.60750085e-01 -8.01130950e-01 8.27195168e-01
-3.32756042e-02 -1.10381484e+00 -2.00032786e-01 5.87277561e-02
1.22677279e+00 3.65786217e-02 1.66890129e-01 2.10537687e-01
5.23175895e-01 -6.84678614e-01 9.64783370e-01 7.05353916e-02
9.82797027e-01 -5.82046211e-01 4.58187014e-01 6.00764394e-01
-9.80739713e-01 -1.62400559e-01 -3.61229181e-01 2.77869701e-01
1.36361405e-01 9.32798862e-01 -8.18666935e-01 -2.22083461e-02
5.06318271e-01 8.69680345e-01 -5.77957749e-01 1.00253415e+00
-2.45713875e-01 8.87385249e-01 -2.88399577e-01 3.70656461e-01
3.03959489e-01 -2.47324213e-01 2.14624181e-01 8.72210503e-01
1.67860061e-01 -1.18178517e-01 4.30220008e-01 9.56386507e-01
-7.36615136e-02 6.69611543e-02 -1.02087356e-01 1.05258606e-01
4.85727459e-01 1.05512750e+00 -1.14610887e+00 -3.44233811e-01
-1.51400581e-01 1.06428802e+00 2.17723042e-01 4.08356428e-01
-8.94725680e-01 3.70155908e-02 4.45664048e-01 2.70770997e-01
3.07259619e-01 1.70834571e-01 -9.58538175e-01 -9.34346199e-01
4.54983525e-02 -5.82014382e-01 4.91919130e-01 -6.51837826e-01
-1.36945355e+00 3.88617873e-01 -2.22183645e-01 -1.49351919e+00
8.58521312e-02 -2.26841807e-01 -5.48133969e-01 7.11645484e-01
-1.76098752e+00 -1.09492600e+00 -5.37722886e-01 5.40330946e-01
8.06089342e-01 2.88493223e-02 2.49287456e-01 4.07615811e-01
-6.40826106e-01 5.69982290e-01 4.48840052e-01 -1.22115202e-01
9.69460726e-01 -1.40277016e+00 1.63426414e-01 7.33769894e-01
1.61785688e-02 -8.69991630e-02 6.93557560e-01 -7.36921191e-01
-4.76222128e-01 -1.55922449e+00 4.20059294e-01 -1.40321672e-01
3.66136819e-01 -3.13511938e-01 -1.14062238e+00 4.70784932e-01
-2.20957071e-01 2.35758618e-01 4.76073772e-01 -2.82962680e-01
-2.02221230e-01 -2.52486408e-01 -1.33429360e+00 4.47121441e-01
9.60933328e-01 -7.69793987e-02 -3.64188015e-01 5.50928056e-01
7.27840006e-01 -5.13978720e-01 -2.62225717e-01 4.24935788e-01
4.94530201e-02 -7.99892426e-01 7.76942730e-01 -1.25970155e-01
5.35531521e-01 -5.58080316e-01 -7.91611671e-02 -1.35911441e+00
-2.09899828e-01 -4.10540700e-02 1.54815808e-01 1.57580590e+00
5.40342271e-01 -4.89446312e-01 1.53169930e+00 6.46460235e-01
-2.84667313e-01 -5.83143711e-01 -8.90291154e-01 -5.61185837e-01
1.61748424e-01 -1.51188105e-01 2.81561017e-01 9.66488063e-01
-5.79598546e-01 3.08928668e-01 -1.39383063e-01 2.81729311e-01
8.97198856e-01 3.64286840e-01 7.01752186e-01 -1.26358140e+00
-2.83692747e-01 -3.64946663e-01 -1.17747366e-01 -1.10585570e+00
4.24681932e-01 -8.87364030e-01 5.96728623e-01 -1.41890979e+00
3.90780002e-01 -1.09411490e+00 -2.67603785e-01 5.23198426e-01
-4.82766330e-01 6.47383034e-01 -2.07903683e-01 5.42349458e-01
-6.46715522e-01 6.42364740e-01 1.44012189e+00 -1.97911203e-01
-2.70665735e-01 1.83754146e-01 -9.59384859e-01 7.56545067e-01
7.15060294e-01 -7.53830850e-01 -7.32963264e-01 -3.22406232e-01
-3.55328232e-01 -1.66724220e-01 1.14707932e-01 -1.08396518e+00
1.13251820e-01 -4.52682585e-01 5.13735294e-01 -6.16480470e-01
9.29649919e-02 -8.68940234e-01 8.26954320e-02 2.45087355e-01
-5.59750319e-01 -7.94366598e-01 -4.32396941e-02 8.41947138e-01
-2.91138530e-01 -2.28494748e-01 1.40385747e+00 -1.28569957e-02
-8.30934465e-01 3.80656153e-01 -9.86947212e-03 4.36244041e-01
1.34331930e+00 -4.77944851e-01 -1.95522055e-01 -1.23282664e-01
-6.24931276e-01 5.02304316e-01 8.36641490e-01 1.95174456e-01
1.05822638e-01 -1.12351215e+00 -5.95106423e-01 1.65531293e-01
3.13206404e-01 7.98374116e-01 4.56300050e-01 5.89272678e-01
-2.50854164e-01 6.16596872e-03 -1.00735649e-01 -9.23124909e-01
-9.87788498e-01 4.51407164e-01 3.45459044e-01 -1.71257243e-01
-5.62030375e-01 1.15170884e+00 7.69765615e-01 -5.97329795e-01
2.44854540e-01 3.55237653e-03 -1.84148416e-01 1.61684856e-01
2.82015234e-01 -2.97593866e-02 -1.68844923e-01 -6.73066378e-01
-2.37227902e-01 7.33616114e-01 -1.67798623e-01 2.74367854e-02
1.20085549e+00 -3.97121817e-01 2.06086546e-01 4.15977001e-01
1.14137542e+00 -2.80195206e-01 -2.03100848e+00 -5.79712629e-01
9.26252361e-03 -5.12546241e-01 1.41298056e-01 -9.54917490e-01
-1.47197676e+00 6.97341621e-01 7.81256974e-01 -1.75805017e-01
1.24063826e+00 2.64690548e-01 9.87392843e-01 -2.71292537e-01
9.82317403e-02 -1.39721298e+00 2.15165347e-01 1.40730441e-01
1.49365619e-01 -1.51030827e+00 -3.35612178e-01 -7.77803242e-01
-1.12280250e+00 7.25480258e-01 9.64636505e-01 -1.17873058e-01
5.11489332e-01 -6.73774481e-02 2.97612667e-01 3.19462232e-02
-3.46670091e-01 -2.39644870e-01 2.39961952e-01 7.21592069e-01
-2.91366428e-02 2.11513534e-01 -5.88737475e-03 5.12762725e-01
1.18407927e-01 -1.53082013e-02 2.84421533e-01 6.36555493e-01
-7.24960625e-01 -1.09760892e+00 -5.41231096e-01 6.73009634e-01
-2.50202239e-01 3.76152918e-02 -2.52625912e-01 2.78253466e-01
3.80005270e-01 8.05467129e-01 1.98822796e-01 -1.62854403e-01
1.20760977e-01 1.46886840e-01 1.82644203e-02 -8.40097904e-01
-2.81876710e-04 3.04826796e-01 -3.70126963e-01 -2.41529882e-01
-3.31198722e-01 -7.98814893e-01 -1.48238409e+00 1.17187195e-01
-3.50360423e-01 2.14103553e-02 5.39267123e-01 9.22880769e-01
2.84796685e-01 4.90456045e-01 7.89200783e-01 -6.54679477e-01
-5.45667052e-01 -8.14009130e-01 -8.27468634e-01 5.66615760e-01
2.36239508e-01 -7.09646642e-01 -4.25889194e-01 2.92149156e-01] | [9.65475845336914, 1.3061491250991821] |
b166a890-6bfe-4445-9a3a-b7f06da80a7b | skeleton-dml-deep-metric-learning-for | 2012.13823 | null | https://arxiv.org/abs/2012.13823v2 | https://arxiv.org/pdf/2012.13823v2.pdf | Skeleton-DML: Deep Metric Learning for Skeleton-Based One-Shot Action Recognition | One-shot action recognition allows the recognition of human-performed actions with only a single training example. This can influence human-robot-interaction positively by enabling the robot to react to previously unseen behaviour. We formulate the one-shot action recognition problem as a deep metric learning problem and propose a novel image-based skeleton representation that performs well in a metric learning setting. Therefore, we train a model that projects the image representations into an embedding space. In embedding space the similar actions have a low euclidean distance while dissimilar actions have a higher distance. The one-shot action recognition problem becomes a nearest-neighbor search in a set of activity reference samples. We evaluate the performance of our proposed representation against a variety of other skeleton-based image representations. In addition, we present an ablation study that shows the influence of different embedding vector sizes, losses and augmentation. Our approach lifts the state-of-the-art by 3.3% for the one-shot action recognition protocol on the NTU RGB+D 120 dataset under a comparable training setup. With additional augmentation our result improved over 7.7%. | ['Dietrich Paulus', 'Nick Theisen', 'Simon Häring', 'Raphael Memmesheimer'] | 2020-12-26 | null | null | null | null | ['one-shot-3d-action-recognition'] | ['computer-vision'] | [ 6.70068622e-01 3.16800028e-01 -2.61095554e-01 -4.76449519e-01
-8.93066704e-01 8.14057440e-02 8.84550095e-01 -2.44629562e-01
-9.19141233e-01 4.99695569e-01 4.01441574e-01 3.50274861e-01
-1.27300367e-01 -4.12624508e-01 -6.99272335e-01 -8.41295123e-01
-1.02174222e-01 5.57264447e-01 4.11150575e-01 -2.53162861e-01
2.34026223e-01 6.64333701e-01 -1.82719541e+00 3.86413544e-01
-4.35121730e-03 9.69634473e-01 1.02264859e-01 7.55115271e-01
2.93840468e-01 1.05988967e+00 -6.46643758e-01 1.84292831e-02
4.64241982e-01 -4.83138829e-01 -9.20472562e-01 2.69655675e-01
2.38812834e-01 -4.54858273e-01 -7.40350842e-01 6.02401614e-01
5.49703360e-01 4.07723665e-01 8.62452984e-01 -1.52567327e+00
-5.85130394e-01 3.68134156e-02 -3.00458997e-01 5.53789251e-02
7.37925231e-01 3.88312370e-01 6.71169996e-01 -7.40658224e-01
8.29075873e-01 1.19209337e+00 4.17872995e-01 8.58460248e-01
-1.15929699e+00 -9.67343450e-02 -1.19065858e-01 6.15534663e-01
-1.13511741e+00 -4.90629554e-01 6.07703567e-01 -3.38647306e-01
1.47691488e+00 -8.01683590e-02 5.44331908e-01 1.49810791e+00
-4.35918123e-02 1.07683635e+00 6.56912148e-01 -4.46339548e-01
3.93274963e-01 -3.79463643e-01 -6.06123805e-02 5.30553818e-01
1.32388607e-01 1.90876111e-01 -6.97199166e-01 2.07431808e-01
5.36303639e-01 3.38432342e-01 -1.69264153e-01 -1.01157641e+00
-1.41832435e+00 7.63009548e-01 4.87579048e-01 4.73644137e-01
-3.93972933e-01 5.27610362e-01 5.91951370e-01 3.70157838e-01
4.62574922e-02 5.18763125e-01 -4.07555073e-01 -7.30298400e-01
-4.54803646e-01 -4.10636216e-02 7.69420147e-01 5.81457913e-01
7.17202127e-01 -9.40874368e-02 -1.96173906e-01 7.46413052e-01
3.94501723e-02 4.85147446e-01 8.44844222e-01 -1.19735885e+00
3.84552687e-01 6.87613666e-01 1.26772195e-01 -8.25590014e-01
-4.69431639e-01 3.01538229e-01 -3.48093748e-01 6.34502769e-01
5.12950003e-01 2.00205520e-01 -1.03351152e+00 1.61994612e+00
1.65483236e-01 7.89880306e-02 3.89461100e-01 1.01939011e+00
3.12738895e-01 3.11808616e-01 -2.00306401e-01 2.25454777e-01
1.02443063e+00 -1.09369874e+00 -6.05022907e-01 -4.85042721e-01
1.30414915e+00 -2.60770231e-01 1.24600410e+00 2.59447604e-01
-5.30314922e-01 -3.77278388e-01 -1.36066759e+00 -1.29848018e-01
-5.29429853e-01 1.86076522e-01 4.85023469e-01 4.95240837e-01
-8.30471575e-01 8.82853031e-01 -1.19494033e+00 -8.85912001e-01
4.59645629e-01 4.36310410e-01 -1.09727108e+00 -2.41843641e-01
-8.47166598e-01 1.26013267e+00 3.19108546e-01 -2.87311584e-01
-1.04671621e+00 -1.39292777e-01 -1.11349404e+00 -2.44616345e-01
4.99733627e-01 -2.48486131e-01 1.21838737e+00 -7.46948600e-01
-1.74601769e+00 1.10653746e+00 8.74486119e-02 -7.64554739e-01
5.90343416e-01 -4.15775150e-01 -2.26931632e-01 2.89871752e-01
8.51713568e-02 7.99838960e-01 7.00125873e-01 -7.43986964e-01
-3.93920690e-01 -5.93213499e-01 2.02298269e-01 2.52807170e-01
-2.52561361e-01 -1.77043393e-01 -2.20018193e-01 -3.20660204e-01
-1.86583288e-02 -1.16856337e+00 -1.10274792e-01 5.03971696e-01
7.37856999e-02 -6.41169250e-02 8.23775768e-01 -1.65502608e-01
5.70450246e-01 -2.38545680e+00 5.31129658e-01 -2.14179471e-01
-2.68554747e-01 1.34055063e-01 -4.06514227e-01 5.03225982e-01
-1.32913888e-01 -3.99860561e-01 -4.65286851e-01 -3.92617762e-01
2.41023660e-01 6.79956317e-01 2.22569034e-01 8.57747376e-01
2.45539650e-01 8.66424620e-01 -9.39288676e-01 -3.64304870e-01
5.99689782e-01 5.34287810e-01 -3.91750067e-01 2.89496243e-01
1.81266367e-01 2.12421671e-01 -6.67229518e-02 6.27700865e-01
5.00193574e-02 -1.83485670e-03 1.23038515e-01 -1.01607271e-01
3.47913623e-01 -1.38406893e-02 -1.04004836e+00 2.42464948e+00
-4.52477425e-01 7.83790410e-01 -3.97305310e-01 -1.17558813e+00
9.74753916e-01 1.46895885e-01 7.46857285e-01 -8.75658453e-01
1.40095174e-01 1.41472623e-01 1.39997378e-01 -6.06517196e-01
3.01944047e-01 -5.37534654e-02 -3.29679966e-01 7.36414135e-01
3.25712889e-01 1.79591775e-01 2.11367100e-01 1.08995095e-01
1.78445268e+00 4.74677861e-01 4.47222292e-01 1.86335832e-01
3.66744995e-01 -1.40772745e-01 2.38153726e-01 4.58303958e-01
-6.87257767e-01 6.23337388e-01 4.34458584e-01 -6.45130575e-01
-8.75468612e-01 -9.78453994e-01 1.92955583e-01 1.25353789e+00
2.22956821e-01 -2.83599555e-01 -6.00094020e-01 -9.01318908e-01
8.89310986e-03 7.97467053e-01 -1.12534690e+00 -7.72803664e-01
-6.75775051e-01 -2.82592952e-01 6.47720814e-01 7.94901550e-01
4.40742522e-01 -1.38446641e+00 -1.35657227e+00 -5.54921627e-02
9.35180336e-02 -1.20235026e+00 -9.76520628e-02 5.18941760e-01
-7.52077460e-01 -1.36522436e+00 -9.10942495e-01 -6.71971023e-01
6.86220109e-01 2.14920729e-01 8.17002892e-01 -3.67559195e-01
-4.88525629e-01 9.85140502e-01 -7.69759536e-01 -5.01964130e-02
-4.05717999e-01 -1.87125936e-01 2.67926395e-01 1.46411434e-01
7.19390750e-01 -4.28629279e-01 -5.71367443e-01 4.97244537e-01
-6.59095705e-01 -2.92876512e-01 8.30554247e-01 8.58766615e-01
4.41612393e-01 -5.26754439e-01 2.83823684e-02 -2.16688573e-01
2.69151568e-01 -1.85309574e-01 7.77095035e-02 1.88928589e-01
-2.98463732e-01 6.32859409e-01 2.07088798e-01 -7.31327295e-01
-7.49057770e-01 6.11348927e-01 3.35417479e-01 -6.26081288e-01
-4.00929898e-01 6.72792122e-02 -2.73838401e-01 -1.33207455e-01
8.99918079e-01 2.49859005e-01 3.25543284e-01 -4.65013474e-01
6.30324841e-01 6.75201714e-01 6.44476175e-01 -3.05742115e-01
3.32856178e-01 6.89819515e-01 1.12344421e-01 -8.10375035e-01
-4.69254822e-01 -6.23394907e-01 -1.10937870e+00 -3.03910375e-01
9.18663144e-01 -6.21075928e-01 -6.17145300e-01 5.05656838e-01
-1.05892825e+00 -6.11848295e-01 -7.42894113e-01 6.69826388e-01
-1.31035626e+00 2.80747443e-01 -2.05218285e-01 -7.31803536e-01
8.01787972e-02 -1.13330162e+00 1.39058745e+00 -3.03139865e-01
-5.51239371e-01 -4.24277216e-01 4.92233962e-01 2.68305838e-01
4.59282135e-04 4.47284877e-01 3.18029255e-01 -9.68028605e-01
-8.99233148e-02 -6.21119797e-01 1.11658223e-01 3.90077382e-01
3.27661753e-01 -4.30793464e-01 -8.64805937e-01 -2.10719571e-01
7.86262285e-03 -7.45563328e-01 9.27075446e-01 -1.45647511e-01
7.00765669e-01 3.85975949e-02 -2.60781169e-01 3.56109291e-01
9.65809584e-01 2.06867635e-01 1.05450332e+00 6.49887741e-01
5.18622756e-01 4.79080021e-01 8.71879339e-01 4.77244854e-01
-1.23556182e-01 8.84598374e-01 6.36683285e-01 2.07394376e-01
-9.20811519e-02 -1.01408996e-01 6.78693175e-01 2.40615308e-01
-7.26385564e-02 3.22098918e-02 -9.00827527e-01 4.93943840e-01
-2.05850530e+00 -1.11080110e+00 4.54943746e-01 2.07344818e+00
4.39797610e-01 1.99431628e-01 2.52361834e-01 5.38857698e-01
4.08545047e-01 3.30867499e-01 -6.94514930e-01 -4.84473258e-01
5.44091202e-02 1.53751284e-01 5.95013618e-01 1.54986635e-01
-1.25380445e+00 8.51370454e-01 5.94124985e+00 4.52539921e-01
-9.49702919e-01 1.28963113e-01 3.60584781e-02 -4.94288355e-01
6.24035776e-01 -3.53657216e-01 -2.77987152e-01 2.75886536e-01
1.05543137e+00 -5.08027039e-02 2.42960989e-01 1.18060541e+00
-3.84561322e-03 -4.05304223e-01 -1.65172946e+00 1.34009254e+00
6.07228935e-01 -9.09939766e-01 -1.21973626e-01 1.60521582e-01
3.27757627e-01 1.52471691e-01 -1.78941384e-01 4.97254461e-01
1.61962777e-01 -1.05658054e+00 5.16334236e-01 5.81427813e-01
8.28863502e-01 -4.58246529e-01 6.34857178e-01 2.22964719e-01
-9.55564559e-01 -2.89307415e-01 -3.10159445e-01 -2.43019715e-01
1.71543583e-01 -2.35973209e-01 -8.40160429e-01 3.02259356e-01
4.87290740e-01 1.08887219e+00 -5.23671210e-01 7.74452329e-01
-2.71143556e-01 -8.58586729e-02 -2.39697814e-01 -5.67958467e-02
3.66426259e-01 9.90636647e-02 2.93179512e-01 9.61588383e-01
2.51138359e-01 7.46790618e-02 -8.92458260e-02 3.98984641e-01
4.14419696e-02 -2.03863382e-01 -1.05530190e+00 -1.63943887e-01
-1.29522197e-02 6.95990562e-01 -5.97502530e-01 -2.20918313e-01
-4.18463230e-01 1.69686878e+00 4.94747102e-01 1.17926188e-01
-8.03818882e-01 -5.03545523e-01 1.01839960e+00 -1.38568655e-01
4.73475516e-01 -4.06493604e-01 1.47276863e-01 -8.71056378e-01
1.48729712e-01 -7.56307602e-01 2.82504141e-01 -7.57434845e-01
-7.16918647e-01 3.47779244e-01 2.33366080e-02 -1.61667597e+00
-5.66290796e-01 -9.36082423e-01 -2.67981350e-01 1.49337739e-01
-9.43969607e-01 -8.60904813e-01 -3.23535353e-01 6.04345143e-01
7.19904363e-01 -2.48573512e-01 1.19438505e+00 4.34639566e-02
-3.82005274e-01 4.79857653e-01 1.19844049e-01 2.23405123e-01
6.83137655e-01 -1.07324851e+00 4.05175894e-01 5.88715971e-01
4.78546292e-01 2.11218879e-01 5.67406654e-01 -3.28856528e-01
-1.45367587e+00 -7.78052330e-01 6.17547452e-01 -7.33693242e-01
5.96843243e-01 -2.97901750e-01 -6.69888616e-01 6.14844859e-01
-2.11189777e-01 3.40478927e-01 7.07238317e-01 -2.35861167e-01
-5.16386509e-01 -7.62454718e-02 -1.27626836e+00 5.49057901e-01
1.45438647e+00 -6.13650024e-01 -9.34333980e-01 2.58872777e-01
4.68703747e-01 5.29486537e-02 -7.20629454e-01 2.80842721e-01
7.97725201e-01 -1.08076668e+00 9.51256633e-01 -9.42197263e-01
2.79346794e-01 -1.71608731e-01 -5.99956334e-01 -1.15037155e+00
-1.22760333e-01 -4.25280482e-01 -1.76697552e-01 3.94788146e-01
1.31340474e-01 -3.47510606e-01 9.03184831e-01 7.04349637e-01
-3.24005969e-02 -6.80991054e-01 -1.26426053e+00 -1.28070796e+00
-3.26662183e-01 -5.92669547e-01 2.54243463e-01 4.86376435e-01
4.06080991e-01 6.93521425e-02 -4.65193689e-01 -4.08056885e-01
4.00973916e-01 -2.22770095e-01 1.13270926e+00 -1.00221002e+00
-1.33774281e-01 -2.43676379e-01 -1.34020698e+00 -9.95237112e-01
3.85785103e-01 -8.20445359e-01 3.45159918e-01 -1.51275218e+00
3.64750773e-02 3.12471420e-01 -5.11465371e-01 7.26309180e-01
4.03696507e-01 2.46410280e-01 3.53487134e-01 1.91217080e-01
-1.04485524e+00 9.79652226e-01 8.71962368e-01 -3.68518889e-01
1.13771416e-01 -4.01627004e-01 -4.54452820e-02 6.09141827e-01
7.51108825e-01 -6.21095181e-01 -3.62939984e-01 -3.91151339e-01
-2.07354382e-01 -2.87615657e-01 4.24032807e-01 -1.41582060e+00
1.79896742e-01 -1.04151089e-02 1.79056242e-01 -1.69867501e-01
9.20451403e-01 -1.04833364e+00 -9.44154784e-02 7.46523440e-01
-5.13092220e-01 -2.74093866e-01 -1.63843352e-02 8.83424461e-01
-2.51187966e-03 -1.22005150e-01 6.75198257e-01 -1.06003903e-01
-1.22187579e+00 -5.64609701e-03 -4.27148491e-01 2.64028218e-02
1.42271984e+00 -8.54114473e-01 -8.64464343e-02 -2.96413779e-01
-8.46243739e-01 -1.16489232e-01 6.81203365e-01 6.45682693e-01
7.50383973e-01 -1.50025105e+00 -2.93738127e-01 3.16367120e-01
5.65796077e-01 -5.04424214e-01 -9.63412002e-02 8.61423314e-01
-3.93941432e-01 4.85401690e-01 -7.35182405e-01 -5.75349271e-01
-1.22287548e+00 4.95177150e-01 3.92519742e-01 -5.09064905e-02
-8.41847718e-01 7.09589779e-01 -1.44317061e-01 -4.33348596e-01
6.10621333e-01 -1.85161635e-01 -9.17984024e-02 -6.53351992e-02
5.64640641e-01 7.46137261e-01 -5.02557531e-02 -9.90338445e-01
-7.37239718e-01 4.85840231e-01 9.18798819e-02 -3.04657489e-01
1.30778456e+00 2.58699149e-01 4.50408369e-01 8.05956781e-01
1.64812386e+00 -9.91306722e-01 -1.47027886e+00 -1.26462147e-01
3.14402431e-01 -7.23377287e-01 -2.03578755e-01 -5.97923338e-01
-8.16052794e-01 8.29634011e-01 9.30307865e-01 -3.07594061e-01
7.34960318e-01 2.28100598e-01 6.67569876e-01 1.00580251e+00
6.32270694e-01 -1.22054279e+00 8.13433826e-01 6.06765568e-01
1.12726891e+00 -1.49306202e+00 -1.68329682e-02 2.85706997e-01
-8.86859655e-01 1.02699733e+00 5.09219706e-01 -2.70261317e-01
3.08853477e-01 9.34552997e-02 7.19259456e-02 -3.35533798e-01
-6.03515446e-01 -4.46056038e-01 -5.65022863e-02 8.86925101e-01
9.91884694e-02 3.07134110e-02 -1.14840388e-01 2.40141585e-01
1.25879928e-01 1.60131767e-01 4.11615580e-01 1.33726346e+00
-3.90230507e-01 -9.98587310e-01 -2.43627541e-02 2.22206622e-01
9.71719697e-02 4.84379381e-01 -7.84077168e-01 8.77023816e-01
-2.61720240e-01 8.57605517e-01 2.44973794e-01 -6.65799141e-01
6.34368479e-01 3.43452513e-01 7.54696846e-01 -6.83490396e-01
-4.15305346e-02 -5.79240441e-01 1.14848360e-01 -1.33083844e+00
-7.12554276e-01 -8.65011275e-01 -1.38340998e+00 1.51867703e-01
7.28691965e-02 -3.32072586e-01 7.37874568e-01 9.55263317e-01
4.05120999e-01 3.48856419e-01 3.94985288e-01 -1.22859836e+00
-7.96350420e-01 -1.11766934e+00 -5.97148716e-01 7.80038953e-01
1.15998469e-01 -1.15508866e+00 -5.51647067e-01 1.14767707e-03] | [8.053242683410645, 0.6409252285957336] |
53139a63-4ae5-49c5-abfc-78cf8b0d84e4 | klpt-kurdish-language-processing-toolkit | null | null | https://aclanthology.org/2020.nlposs-1.11 | https://aclanthology.org/2020.nlposs-1.11.pdf | KLPT – Kurdish Language Processing Toolkit | Despite the recent advances in applying language-independent approaches to various natural language processing tasks thanks to artificial intelligence, some language-specific tools are still essential to process a language in a viable manner. Kurdish language is a less-resourced language with a remarkable diversity in dialects and scripts and lacks basic language processing tools. To address this issue, we introduce a language processing toolkit to handle such a diversity in an efficient way. Our toolkit is composed of fundamental components such as text preprocessing, stemming, tokenization, lemmatization and transliteration and is able to get further extended by future developers. The project is publicly available. | ['Sina Ahmadi'] | null | null | null | null | emnlp-nlposs-2020-11 | ['transliteration'] | ['natural-language-processing'] | [-9.10389274e-02 -3.23488414e-01 -1.56654865e-01 -4.93944585e-01
-6.41922593e-01 -1.01611948e+00 8.37250948e-01 3.53997678e-01
-7.51382887e-01 6.10493064e-01 3.56774807e-01 -8.26122642e-01
1.37882411e-01 -7.75113523e-01 1.72036752e-01 -1.76438868e-01
1.06968597e-01 5.44242561e-01 -5.32107651e-02 -6.48668289e-01
4.07863438e-01 6.70206010e-01 -1.16315532e+00 4.07539815e-01
9.86257553e-01 9.31530818e-02 2.85978377e-01 6.40630364e-01
-1.05636644e+00 6.20499432e-01 -4.34217036e-01 -3.99914861e-01
2.05226183e-01 -2.63244331e-01 -1.30658674e+00 -1.16924636e-01
-1.38558954e-01 -1.00476108e-02 1.65611282e-01 9.82971668e-01
4.50679779e-01 -1.27260566e-01 3.87161165e-01 -7.05456734e-01
-4.75109369e-01 1.06438494e+00 -1.88781023e-01 1.69138357e-01
7.57420182e-01 1.72394007e-01 7.05707550e-01 -1.05857205e+00
8.35745990e-01 1.39295697e+00 3.65433246e-01 4.87829030e-01
-9.04376745e-01 -4.32558894e-01 6.53647557e-02 -2.83535421e-01
-1.62102211e+00 -3.83511961e-01 4.72825736e-01 -4.05492365e-01
1.41091907e+00 2.59381741e-01 3.70647371e-01 7.10443616e-01
1.83597673e-03 6.39851570e-01 1.32061517e+00 -1.05766463e+00
-6.73358887e-02 3.50783706e-01 3.93464297e-01 5.99094987e-01
7.11295903e-02 -7.20260918e-01 -2.58736134e-01 -1.15631282e-01
3.95462394e-01 -2.26738855e-01 1.20423093e-01 3.46937537e-01
-1.39008224e+00 8.13210964e-01 -2.59475112e-01 9.56858575e-01
-2.35896051e-01 -4.53701913e-01 8.41418743e-01 5.40156960e-01
4.79103982e-01 6.62143707e-01 -8.14024091e-01 -3.95022333e-01
-1.15551019e+00 2.02412978e-01 1.15850902e+00 1.12940049e+00
7.50281751e-01 8.09586048e-02 -5.62921688e-02 9.33093607e-01
2.75470793e-01 4.33876663e-01 7.08389759e-01 -6.63759470e-01
7.23573446e-01 8.63794088e-01 -1.15531035e-01 -6.03974462e-01
-4.37236547e-01 -7.78078958e-02 -6.75222456e-01 -5.66116795e-02
5.19066691e-01 -2.68535703e-01 -8.71805906e-01 1.28467321e+00
2.57812321e-01 -6.93328917e-01 1.72088563e-01 3.67648244e-01
8.24502587e-01 7.42523313e-01 3.63040447e-01 -2.13001788e-01
1.77175021e+00 -7.11163163e-01 -8.66186082e-01 -4.00584042e-01
8.86268616e-01 -1.50796080e+00 1.44924760e+00 4.24634755e-01
-1.21205902e+00 -1.14023551e-01 -7.40235925e-01 -4.82362002e-01
-1.07490587e+00 8.11131150e-02 7.47386754e-01 1.02758563e+00
-1.16285038e+00 1.80366635e-01 -4.67512727e-01 -8.66851866e-01
4.76746634e-02 1.45838052e-01 -5.00306666e-01 -6.97705001e-02
-1.30515432e+00 9.30166721e-01 6.86963975e-01 7.59821534e-02
2.09573619e-02 -4.39051837e-01 -1.04270017e+00 -3.08296382e-01
4.52957571e-01 -5.72992384e-01 1.36107945e+00 -7.32363820e-01
-1.79217744e+00 1.37157202e+00 -3.45162034e-01 -2.36714870e-01
5.12309909e-01 -2.87345946e-01 -5.80239892e-01 -1.01903968e-01
2.58146316e-01 1.22155175e-01 5.27755439e-01 -5.95862508e-01
-7.70535946e-01 -3.64813805e-01 -2.81778574e-01 2.33732760e-01
-3.86639893e-01 1.13844681e+00 -7.29355693e-01 -9.78872359e-01
-6.67929798e-02 -6.81057334e-01 -4.15125817e-01 -4.77689534e-01
-3.18937600e-01 -3.94758880e-01 4.70315874e-01 -9.60840344e-01
1.60088241e+00 -1.96877658e+00 -2.08613634e-01 7.65733719e-02
-9.94532555e-03 5.75473368e-01 -1.24327227e-01 1.12364447e+00
8.56683254e-02 5.87395430e-01 -3.76575172e-01 -4.10671324e-01
1.83869064e-01 4.15752351e-01 -3.49674344e-01 2.04908401e-01
3.38316649e-01 7.01624990e-01 -1.05035186e+00 -5.86609006e-01
4.74431992e-01 5.02323151e-01 -1.07310057e-01 -7.78813288e-02
-4.77851704e-02 1.15904950e-01 -3.91088068e-01 9.94134784e-01
6.97141230e-01 4.97071177e-01 3.01427215e-01 5.27696788e-01
-8.45756292e-01 7.34234571e-01 -1.40650249e+00 1.67927778e+00
-6.68172121e-01 5.79538584e-01 5.09602785e-01 -8.29700470e-01
8.42931151e-01 4.17712510e-01 3.69136371e-02 -2.51413405e-01
1.97187945e-01 4.72121924e-01 -2.14531019e-01 -6.51774824e-01
9.13908124e-01 -8.47293884e-02 -4.22399223e-01 6.23396873e-01
6.65439516e-02 -4.75605041e-01 8.06021690e-01 3.04576993e-01
9.14545894e-01 -4.05327827e-02 9.24745202e-01 -5.62000394e-01
1.05395329e+00 2.80597925e-01 5.05734622e-01 5.85390866e-01
-1.87843844e-01 3.86266649e-01 2.21949905e-01 -3.39959264e-01
-1.08621645e+00 -8.00925016e-01 2.19724774e-02 1.41956568e+00
-6.90288305e-01 -8.39977860e-01 -8.82838368e-01 -4.79163736e-01
-2.76739776e-01 8.50794435e-01 -1.73853096e-02 3.90229881e-01
-8.30217302e-01 -5.40354967e-01 1.03917706e+00 1.29575998e-01
4.55112606e-01 -1.34646332e+00 -1.38945878e-01 4.11393076e-01
-6.34059608e-02 -1.31633878e+00 -1.90033138e-01 1.24626726e-01
-7.04219937e-01 -6.00944996e-01 -6.23921514e-01 -1.17273307e+00
4.73156989e-01 2.07475498e-01 1.20188832e+00 1.43405333e-01
-1.79173410e-01 2.75799572e-01 -4.74447548e-01 -7.04092741e-01
-7.71326184e-01 4.31240261e-01 -5.62500656e-02 -3.92257869e-01
6.57041371e-01 -1.74047470e-01 -8.01423043e-02 -2.76161820e-01
-1.22566259e+00 -1.16990328e-01 3.68866801e-01 2.48713717e-01
2.02547014e-01 2.22619176e-01 2.07745507e-01 -1.16865540e+00
1.12575638e+00 -2.56950021e-01 -3.03947687e-01 2.54280686e-01
-3.63599658e-01 1.01491228e-01 1.09458661e+00 -1.10271886e-01
-1.17702508e+00 1.56028777e-01 -5.53311944e-01 7.60239959e-01
-6.87076390e-01 7.63614416e-01 -3.29451680e-01 1.03904739e-01
3.99939984e-01 4.13236022e-01 -3.14806819e-01 -7.83153176e-01
6.30058587e-01 1.02496147e+00 5.14574766e-01 -5.38811207e-01
1.11354029e+00 3.45106393e-01 -5.43183744e-01 -1.32984447e+00
-3.18316132e-01 -8.26586008e-01 -9.99825716e-01 5.80655225e-03
5.59388638e-01 -8.66901577e-01 -1.12558648e-01 6.24735236e-01
-1.42168391e+00 -1.71609730e-01 -6.53721467e-02 5.15694618e-01
4.81393933e-02 6.59986317e-01 -7.70672798e-01 -7.53198147e-01
-7.57730424e-01 -7.08351254e-01 8.10428798e-01 1.28951073e-01
-5.77108383e-01 -1.18137336e+00 1.88157827e-01 2.30925307e-01
5.16547680e-01 -2.28144713e-02 1.12157869e+00 -8.13716471e-01
3.56376451e-03 -1.92441493e-01 -3.89117301e-02 3.94096583e-01
2.85455048e-01 5.82511723e-01 -5.71267605e-01 -2.45780274e-02
6.21694550e-02 4.71402332e-02 5.23883462e-01 -5.22139966e-02
6.00867867e-01 -4.58168447e-01 -6.37661368e-02 3.75314534e-01
1.43110323e+00 2.32396815e-02 4.95046735e-01 7.84693241e-01
4.21523511e-01 9.04427111e-01 4.16007251e-01 3.63015324e-01
5.82051754e-01 1.47470176e-01 -2.98262775e-01 -2.29079396e-01
3.27257775e-02 1.52678834e-02 6.51859641e-01 1.29533458e+00
6.90650642e-02 -2.10952237e-02 -1.56943607e+00 6.83973491e-01
-1.40966976e+00 -7.96785116e-01 -4.84608233e-01 1.90175176e+00
1.20815444e+00 1.28949851e-01 9.73098055e-02 2.89131492e-01
4.99693692e-01 5.93522042e-02 1.93324268e-01 -1.24277282e+00
-3.22158515e-01 3.24996114e-01 3.64056945e-01 6.74351573e-01
-1.04309869e+00 1.65100694e+00 7.27646017e+00 6.89182758e-01
-1.26538873e+00 -1.83731437e-01 8.81602243e-02 4.21326995e-01
-2.75230974e-01 1.70546755e-01 -9.13141966e-01 1.80429071e-01
9.23355758e-01 -5.57983637e-01 6.29971266e-01 7.03716040e-01
6.40254617e-01 -2.76395213e-02 -6.45733893e-01 7.82862008e-01
6.85325637e-02 -1.12941074e+00 2.15526804e-01 -3.35156471e-01
2.41205290e-01 3.17709625e-01 -2.29339615e-01 5.28712630e-01
3.51304114e-01 -9.74118590e-01 7.31730580e-01 1.08898789e-01
5.21297097e-01 -7.37849474e-01 6.48876548e-01 5.48810244e-01
-1.22766399e+00 2.64200866e-01 -2.89083898e-01 -6.71882570e-01
2.05604494e-01 7.56007433e-01 -9.39434290e-01 7.74588048e-01
5.66189408e-01 3.80628049e-01 -8.98092687e-01 8.50614965e-01
-2.98931211e-01 7.58577824e-01 -2.73607492e-01 -1.46998227e-01
3.97364199e-01 -5.54007769e-01 5.79533637e-01 2.04005551e+00
1.18397325e-01 -8.59534293e-02 3.62192094e-01 3.01168293e-01
2.48962820e-01 1.03593457e+00 -6.52956128e-01 -3.86909902e-01
3.26533318e-01 1.52563155e+00 -1.34275186e+00 -5.99814117e-01
-6.18278384e-01 1.02551401e+00 2.65551358e-01 3.50328237e-01
-2.82046527e-01 -9.22945917e-01 5.62111855e-01 1.68734044e-01
3.64856683e-02 -9.58026350e-01 -5.08595347e-01 -1.28956616e+00
1.25728045e-02 -1.26583540e+00 5.93554020e-01 -3.85528862e-01
-1.20025086e+00 8.74549925e-01 -4.55066636e-02 -8.29572141e-01
-3.46061558e-01 -8.51379693e-01 -5.62694073e-01 1.11819255e+00
-1.52602160e+00 -1.15907502e+00 1.44143760e-01 6.43170416e-01
6.02847576e-01 -2.79457867e-01 1.17945385e+00 4.78340447e-01
-7.03535676e-01 2.63148576e-01 1.13209769e-01 4.62074876e-01
8.99462223e-01 -1.40330398e+00 6.31058991e-01 1.10878026e+00
1.26108050e-01 1.23159814e+00 8.22500288e-01 -4.20871675e-01
-1.45612240e+00 -9.78507757e-01 1.99371541e+00 -3.64434749e-01
1.08378363e+00 -7.75076985e-01 -9.88090098e-01 5.40448785e-01
6.49890900e-01 -4.30635631e-01 1.01299167e+00 1.34509057e-01
-2.94970900e-01 1.22614481e-01 -1.00900388e+00 9.94030476e-01
7.20751643e-01 -7.54358411e-01 -1.12698698e+00 4.75184679e-01
5.08186996e-01 -9.96872783e-02 -6.35238230e-01 -5.29513210e-02
2.38626003e-01 -5.21822333e-01 5.33551216e-01 -5.38764894e-01
-4.81501259e-02 -4.48177099e-01 2.25231081e-01 -9.13421631e-01
-1.50728837e-01 -1.28889751e+00 5.50222158e-01 1.84641123e+00
5.99112272e-01 -8.99228990e-01 2.82586396e-01 5.69968820e-01
3.68081555e-02 -1.26866519e-01 -6.28408551e-01 -8.95904660e-01
4.51179802e-01 -8.74314547e-01 5.20352364e-01 1.13996792e+00
4.07499969e-01 6.32981598e-01 -1.35412127e-01 -2.12401330e-01
1.74762830e-01 4.13334742e-02 8.06184173e-01 -1.15786052e+00
2.42373347e-01 -7.07902730e-01 -7.35424906e-02 -5.36030293e-01
2.97391474e-01 -1.06930566e+00 2.97523104e-03 -1.82723248e+00
-1.61462560e-01 -1.49933249e-01 6.33556694e-02 5.97215056e-01
-3.38020176e-02 1.34310484e-01 1.51041627e-01 5.68375662e-02
-4.90970999e-01 1.66659951e-02 7.04797387e-01 7.33826458e-02
-6.00058556e-01 -6.96402416e-02 -8.94249439e-01 1.09150624e+00
1.08167267e+00 -6.22580945e-01 1.21433541e-01 -5.31424642e-01
4.49234933e-01 -6.14491343e-01 -3.90287220e-01 -6.79515362e-01
3.36971730e-01 -2.92726517e-01 4.63895425e-02 -5.89017868e-01
-4.02710080e-01 -4.62059021e-01 -3.06565464e-01 4.56411332e-01
-7.17184544e-02 5.45842290e-01 4.34648752e-01 -3.02169561e-01
-3.62668961e-01 -4.57650453e-01 4.80212271e-01 -4.99496877e-01
-9.20049965e-01 1.68987811e-01 -1.44114661e+00 1.68002576e-01
8.38827431e-01 -2.55170405e-01 7.53047317e-02 -5.83621822e-02
-1.40518099e-01 1.14197187e-01 6.67852581e-01 5.57566881e-01
1.05106160e-01 -8.20581198e-01 -9.92167115e-01 3.59009445e-01
4.76046540e-02 -2.59545833e-01 -2.78523445e-01 4.40405667e-01
-1.11507773e+00 6.72040105e-01 -2.18160540e-01 6.62534609e-02
-1.40322495e+00 5.55424213e-01 -1.14253864e-01 -5.15479982e-01
-5.67851901e-01 3.33885729e-01 -6.01295948e-01 -9.60441232e-01
4.72482890e-02 -6.34833694e-01 -5.03330112e-01 1.48342460e-01
7.90269554e-01 2.06455529e-01 1.79248020e-01 -9.79619741e-01
-6.84569538e-01 6.23155236e-01 -1.54880896e-01 -4.62079704e-01
1.11987472e+00 -4.13152307e-01 -8.41707349e-01 6.12806439e-01
9.00169730e-01 6.89391494e-01 -1.27510607e-01 -2.76546836e-01
7.69802094e-01 -2.26291612e-01 -1.08639367e-01 -6.25013530e-01
-3.41891348e-01 6.81440353e-01 -9.76064056e-02 4.51751918e-01
1.09078348e+00 -2.17405111e-01 7.34109402e-01 6.01192594e-01
3.71947028e-02 -1.62251723e+00 -6.53034151e-01 1.45977306e+00
7.16162920e-01 -1.03107250e+00 7.54014105e-02 -6.80357397e-01
-4.93342072e-01 1.36317551e+00 5.06904833e-02 1.91221684e-01
8.24936986e-01 6.09689474e-01 7.07291842e-01 4.26269360e-02
-4.41484213e-01 -4.47248191e-01 1.67448685e-01 5.87658882e-01
1.08298194e+00 2.80108564e-02 -1.12119389e+00 6.40850067e-01
-5.96013427e-01 -8.80646259e-02 5.04006267e-01 1.18799007e+00
-3.31492037e-01 -1.74249887e+00 -6.58094406e-01 -7.70058036e-02
-9.39899027e-01 -5.74518442e-01 -7.10901618e-01 9.32644129e-01
-4.71090712e-02 1.07675958e+00 -3.38876188e-01 8.10101926e-02
2.72971630e-01 2.65910327e-01 1.38667166e-01 -1.06394315e+00
-1.11209822e+00 9.39511321e-03 2.03159213e-01 -2.27655292e-01
-2.78189480e-01 -7.05039144e-01 -1.45097196e+00 -4.79329288e-01
2.24479467e-01 1.90762207e-01 7.88316548e-01 1.11513019e+00
1.29210770e-01 1.23561211e-02 1.89587161e-01 -5.38374782e-01
-2.24228173e-01 -9.10294354e-01 -5.44453621e-01 1.61708087e-01
-4.79253009e-04 2.07262635e-01 -3.63504022e-01 1.14292517e-01] | [10.36864185333252, 10.074274063110352] |
d1866b80-cc88-44a4-84f3-bcb389ce7009 | automatic-differentiation-for-adjoint-stencil | 1907.02818 | null | https://arxiv.org/abs/1907.02818v1 | https://arxiv.org/pdf/1907.02818v1.pdf | Automatic Differentiation for Adjoint Stencil Loops | Stencil loops are a common motif in computations including convolutional neural networks, structured-mesh solvers for partial differential equations, and image processing. Stencil loops are easy to parallelise, and their fast execution is aided by compilers, libraries, and domain-specific languages. Reverse-mode automatic differentiation, also known as algorithmic differentiation, autodiff, adjoint differentiation, or back-propagation, is sometimes used to obtain gradients of programs that contain stencil loops. Unfortunately, conventional automatic differentiation results in a memory access pattern that is not stencil-like and not easily parallelisable. In this paper we present a novel combination of automatic differentiation and loop transformations that preserves the structure and memory access pattern of stencil loops, while computing fully consistent derivatives. The generated loops can be parallelised and optimised for performance in the same way and using the same tools as the original computation. We have implemented this new technique in the Python tool PerforAD, which we release with this paper along with test cases derived from seismic imaging and computational fluid dynamics applications. | ['Gerard Gorman', 'Fabio Luporini', 'Jan Hückelheim', 'Navjot Kukreja', 'Sri Hari Krishna Narayanan', 'Paul Hovland'] | 2019-07-05 | null | null | null | null | ['seismic-imaging'] | ['miscellaneous'] | [-1.55047506e-01 -2.96964616e-01 6.48372293e-01 -9.19373631e-02
2.17758063e-02 -6.35721743e-01 5.45615971e-01 1.31593034e-01
-6.01524174e-01 8.20072174e-01 -8.97249058e-02 -7.32089579e-01
1.38593717e-02 -1.08352673e+00 -4.48978931e-01 -6.54390275e-01
-3.91016752e-01 1.95596650e-01 4.86101985e-01 -4.07410830e-01
3.62078100e-01 9.66497242e-01 -1.58653653e+00 1.17953420e-01
8.28406513e-01 9.79384899e-01 -2.13033080e-01 1.04704714e+00
-5.87257981e-01 1.09031856e+00 -3.77321765e-02 -1.11205943e-01
2.50951827e-01 -4.76509899e-01 -8.17923129e-01 -9.71380174e-02
6.71700314e-02 -2.10781306e-01 2.04746187e-01 7.68536389e-01
4.39819008e-01 1.22719519e-01 6.54071748e-01 -8.86327505e-01
1.44187257e-01 1.58521459e-01 -4.40080434e-01 5.81277981e-02
1.14529327e-01 1.70308560e-01 4.67257023e-01 -1.06616855e+00
5.37056983e-01 1.09537911e+00 1.41904032e+00 2.65828073e-01
-1.48409128e+00 -2.29421079e-01 -5.84896326e-01 -5.34671247e-01
-1.26168025e+00 -1.05103359e-01 5.94914615e-01 -7.86728561e-01
1.06831682e+00 6.81140721e-01 8.75766039e-01 2.70520300e-01
5.86577117e-01 2.31750861e-01 1.10631430e+00 -3.32046211e-01
4.94624764e-01 -1.98193714e-02 -4.43861149e-02 8.00459087e-01
-7.91748762e-02 2.55942047e-01 -1.20551445e-01 -4.22064066e-01
1.17969918e+00 -2.90810764e-01 -2.51985520e-01 -4.57821578e-01
-1.27526879e+00 1.04010713e+00 2.66786933e-01 5.00207782e-01
-4.41893041e-01 3.29337686e-01 8.06543708e-01 2.55772233e-01
5.89890540e-01 5.23430586e-01 -4.24713105e-01 -3.38568151e-01
-1.30033398e+00 6.16844237e-01 1.40746057e+00 4.94004577e-01
9.79622066e-01 3.92650306e-01 3.68944108e-02 6.18723094e-01
-4.79339696e-02 2.24218681e-01 5.45478702e-01 -1.28060544e+00
-2.10037410e-01 5.75037539e-01 -8.32821131e-02 -1.20347250e+00
-8.19318414e-01 -2.76407719e-01 -1.28322101e+00 8.70388627e-01
3.91391367e-01 -5.00532985e-01 -5.51340461e-01 1.00640690e+00
5.74669838e-01 4.37087305e-02 5.11659123e-02 7.23001361e-01
6.43544912e-01 7.92908192e-01 -8.46093222e-02 1.69215742e-02
8.92141283e-01 -7.23165274e-01 -2.50337511e-01 9.89070237e-02
1.21516144e+00 -8.79453599e-01 8.67778599e-01 2.57336736e-01
-1.12104917e+00 -2.35524878e-01 -8.18106830e-01 -1.57795891e-01
-3.36131036e-01 -8.72065946e-02 7.31291592e-01 4.07427281e-01
-1.39735472e+00 1.15593278e+00 -9.44013596e-01 2.81460285e-01
1.96380522e-02 3.15861881e-01 -1.53721973e-01 7.85266519e-01
-9.62375522e-01 6.34458840e-01 3.09287399e-01 9.06340331e-02
-2.19710663e-01 -1.05161142e+00 -8.57180834e-01 5.12041934e-02
-1.48108542e-01 -5.51110506e-01 1.24305940e+00 -1.38518369e+00
-1.64644420e+00 9.18321788e-01 6.56409338e-02 -5.97569466e-01
9.41158831e-01 1.66817889e-01 5.23025766e-02 -2.39312515e-01
-5.77045567e-02 3.97109777e-01 8.23681116e-01 -7.90372550e-01
-4.32549030e-01 1.14160724e-01 -3.34004998e-01 -4.17394228e-02
3.31587903e-02 1.47078894e-02 1.86772477e-02 -6.98076546e-01
6.11699820e-02 -7.24208653e-01 -6.09303772e-01 3.37893933e-01
-5.97741380e-02 2.71785617e-01 6.42442882e-01 -8.66284370e-01
1.04556298e+00 -2.21828485e+00 1.46039978e-01 5.47816813e-01
2.75690973e-01 3.72246504e-01 3.67774516e-01 4.12810415e-01
-1.80899203e-01 -2.85188705e-01 -8.78061652e-01 -1.47318514e-02
-2.53870696e-01 2.91525245e-01 -2.55802393e-01 4.89467502e-01
2.19690859e-01 5.34082115e-01 -7.84432650e-01 -4.21115100e-01
1.35470912e-01 7.31733024e-01 -9.31208134e-01 7.67491534e-02
-1.84509948e-01 6.81093156e-01 -3.26907218e-01 1.92360476e-01
7.72800446e-01 2.59030215e-03 -1.07008768e-02 -6.30032942e-02
-1.07988346e+00 2.23673508e-01 -1.57007504e+00 1.31892037e+00
-7.99178302e-01 6.57141805e-01 5.84641278e-01 -1.15208113e+00
1.21768165e+00 2.55439162e-01 4.08227682e-01 -4.38680887e-01
9.47825089e-02 7.26923883e-01 -2.49519348e-01 -2.79234976e-01
6.69046998e-01 -1.62814166e-02 2.19766825e-01 3.64523321e-01
-2.36269355e-01 -5.97998559e-01 5.33650875e-01 1.80855822e-02
9.55526352e-01 4.02688593e-01 3.43101323e-01 -8.64750803e-01
8.87721121e-01 4.54071671e-01 3.61411095e-01 4.66184825e-01
4.73508507e-01 6.36918545e-01 9.88194168e-01 -8.23056221e-01
-1.41669095e+00 -6.11381471e-01 -3.52768719e-01 7.97164261e-01
-4.52940971e-01 -3.56324881e-01 -9.36121404e-01 2.28020102e-01
-6.67855218e-02 3.00432563e-01 -4.16295826e-01 2.75964350e-01
-9.97934461e-01 -6.21734858e-01 7.34398842e-01 5.37909031e-01
8.11738193e-01 -1.17233109e+00 -1.14084315e+00 6.34892762e-01
6.35651350e-01 -5.99410415e-01 -1.11599609e-01 3.68059814e-01
-1.12434077e+00 -8.63673985e-01 -9.40793633e-01 -7.59709120e-01
5.63674510e-01 -4.56418693e-01 1.34685135e+00 2.73321807e-01
-3.27513665e-01 1.69125535e-02 1.40149578e-01 4.54609022e-02
-7.70361900e-01 1.02251060e-01 -3.50422293e-01 -7.25866631e-02
-3.27713341e-01 -8.59918654e-01 -4.12273109e-01 -4.56195921e-02
-1.06622362e+00 3.87805164e-01 7.43908957e-02 9.77225959e-01
5.64994097e-01 -6.16536252e-02 8.08891654e-02 -8.45515907e-01
8.67782712e-01 -2.38082156e-01 -1.13616502e+00 -2.84246475e-01
-3.88571501e-01 4.84163284e-01 9.74867404e-01 -3.14016789e-01
-8.95515144e-01 3.93145323e-01 -6.94258451e-01 -1.84705600e-01
1.00935977e-02 6.80800200e-01 3.56374860e-01 -6.59196734e-01
9.67156827e-01 1.90800726e-01 3.38584185e-01 -4.75266129e-01
9.19718444e-02 3.29980761e-01 7.28780627e-01 -6.09532714e-01
4.44365829e-01 5.05801976e-01 4.55772728e-01 -1.01486802e+00
4.12289724e-02 -9.19911489e-02 -4.88864571e-01 -1.86322689e-01
7.71840394e-01 -4.32972997e-01 -5.80293775e-01 1.03276777e+00
-1.44093204e+00 -6.16290987e-01 -5.53443968e-01 1.63744703e-01
-4.17709768e-01 2.04710320e-01 -6.96121454e-01 -4.53753233e-01
-4.29277927e-01 -1.24825132e+00 7.61139214e-01 1.88481927e-01
-3.49575073e-01 -1.32894540e+00 3.13436538e-01 -7.04494894e-01
9.43412483e-01 6.07523918e-01 7.00308979e-01 -2.08275467e-01
-2.31188998e-01 2.48983335e-02 -2.48701423e-01 3.79278779e-01
-2.24737376e-01 3.26198846e-01 -7.36874044e-01 5.77432588e-02
1.84347034e-01 2.82125533e-01 6.58198893e-01 3.00787628e-01
1.01750875e+00 -3.81352812e-01 -5.71875274e-02 1.12332463e+00
1.50231767e+00 -1.11993656e-01 5.49430966e-01 3.29071671e-01
5.59614837e-01 5.82567513e-01 -9.25208181e-02 5.01745224e-01
-4.53846566e-02 3.27233523e-01 5.65349683e-02 -4.38873708e-01
2.00184390e-01 1.92267194e-01 9.62960124e-02 6.01046264e-01
-3.08993876e-01 6.59852207e-01 -1.33392107e+00 4.42877561e-01
-1.66375434e+00 -6.07333422e-01 -9.90623713e-01 1.85714626e+00
1.12261653e+00 -6.87214136e-02 7.47250393e-02 4.69353229e-01
3.10691535e-01 -4.72192131e-02 -1.00801043e-01 -1.11441457e+00
9.53892246e-03 5.63875556e-01 8.78477812e-01 7.83564925e-01
-9.49526131e-01 6.34930849e-01 6.34759188e+00 5.25631070e-01
-1.67127192e+00 4.45449725e-02 4.53698158e-01 3.86008233e-01
-3.04501265e-01 7.24671483e-02 -4.81829524e-01 2.16241986e-01
8.29650521e-01 -3.92070599e-03 4.42660809e-01 9.99544024e-01
1.86973304e-01 -3.56748044e-01 -7.72159576e-01 7.31971622e-01
-4.40037489e-01 -1.81368589e+00 -5.81339896e-01 -2.11858496e-01
6.65953219e-01 -1.05292976e-01 -3.52955401e-01 -3.35277282e-02
1.65715486e-01 -8.96313667e-01 7.25129604e-01 3.44708890e-01
5.83014488e-01 -9.15319085e-01 4.39296246e-01 2.59545326e-01
-1.08100414e+00 2.78569132e-01 -9.28398296e-02 -5.28534770e-01
2.34554306e-01 1.08856773e+00 -4.89027321e-01 1.20926619e-01
7.89738774e-01 5.03144920e-01 -3.68169248e-01 9.81251299e-01
1.66830331e-01 4.72477466e-01 -4.44212437e-01 -1.67541713e-01
5.23930371e-01 -6.29815817e-01 6.70341432e-01 1.48839116e+00
3.72987002e-01 -9.00793895e-02 -2.46682271e-01 1.10266662e+00
6.11370027e-01 3.38190913e-01 -5.23881137e-01 1.39495492e-01
-1.64981619e-01 1.17697632e+00 -9.66195762e-01 -3.18914592e-01
-5.14221787e-01 7.95058310e-01 7.23708048e-03 3.61905485e-01
-5.79749107e-01 -7.25418329e-01 6.91489756e-01 4.19869840e-01
2.16732681e-01 -7.23485351e-01 -6.39383554e-01 -8.93828332e-01
-1.04681589e-01 -5.99527538e-01 -3.87070216e-02 -7.07159996e-01
-9.54383969e-01 8.53051484e-01 5.68461306e-02 -7.36729205e-01
-6.69755757e-01 -9.36699986e-01 -8.24316025e-01 1.16408801e+00
-1.21009743e+00 -5.35018504e-01 -2.25221694e-01 5.42080164e-01
-5.18538356e-02 5.74789904e-02 8.37318003e-01 3.51979107e-01
-9.18142125e-02 4.06398922e-02 3.59076858e-01 2.30084658e-01
3.55369933e-02 -1.12418997e+00 5.53175688e-01 8.50422084e-01
-5.77665806e-01 3.15486431e-01 8.24341536e-01 -6.26483142e-01
-1.30998564e+00 -9.50838566e-01 1.02217674e+00 4.30994928e-01
9.83040869e-01 -1.88505277e-01 -1.32408321e+00 4.99769121e-01
1.61150202e-01 2.38470972e-01 1.85700208e-02 -3.33042264e-01
8.37420821e-02 2.12479569e-02 -1.09057653e+00 4.61063772e-01
4.35509086e-01 -2.12266907e-01 -1.28523946e-01 2.05172271e-01
3.10978085e-01 -1.00103629e+00 -7.55967200e-01 2.27402031e-01
4.20589775e-01 -1.36595654e+00 9.04555142e-01 -5.23332097e-02
4.25504684e-01 -3.40029478e-01 2.21670002e-01 -1.09297931e+00
4.05942556e-03 -9.01185870e-01 1.13981292e-01 1.05334651e+00
1.59922227e-01 -1.15327501e+00 3.79039496e-01 5.33592880e-01
-1.27168775e-01 -6.64464116e-01 -9.23217654e-01 -6.45420492e-01
4.81638998e-01 -3.97904396e-01 6.51780546e-01 8.63526404e-01
-2.24676087e-01 -2.47015357e-01 6.60702288e-02 -1.67936996e-01
4.14456248e-01 -7.60772452e-02 6.36805892e-01 -1.21620095e+00
-3.11118543e-01 -7.25829363e-01 -4.66864735e-01 -6.52008355e-01
2.13901073e-01 -8.73741746e-01 7.87790716e-02 -1.21763217e+00
-6.63605332e-01 -7.55816638e-01 7.20100999e-01 3.99373114e-01
4.07566100e-01 3.13348591e-01 -2.68902034e-01 -7.31367394e-02
4.52746898e-01 8.21800604e-02 1.01092732e+00 2.63771236e-01
-5.72438300e-01 -2.15291724e-01 3.05505586e-03 9.87290978e-01
8.40274096e-01 -3.19121718e-01 -2.98961680e-02 -3.12729627e-01
6.34811699e-01 4.05084938e-02 8.40613484e-01 -1.19292176e+00
1.96913645e-01 2.19292715e-02 4.53104079e-02 -3.63043487e-01
-6.11095242e-02 -5.36665022e-01 6.46773577e-01 6.71961606e-01
-2.06176370e-01 4.05033708e-01 3.46378833e-01 -4.33080554e-01
-2.85776556e-01 -4.78675961e-01 9.64119852e-01 -3.32130104e-01
-6.85032308e-01 3.94009613e-02 -7.93684304e-01 1.58335164e-01
6.81089520e-01 -1.52098939e-01 4.45832938e-01 -1.67705089e-01
-4.96973157e-01 -1.42860442e-01 6.28440559e-01 -3.20284665e-01
3.23259234e-01 -1.22496986e+00 -7.76493967e-01 6.51202738e-01
-6.88422918e-01 4.06739146e-01 2.29132976e-02 9.82315958e-01
-1.60454631e+00 1.71620604e-02 -2.45812953e-01 -7.72104919e-01
-9.54997659e-01 1.86008796e-01 9.69798386e-01 -3.17925304e-01
-8.91841173e-01 5.48571825e-01 2.08993498e-02 -6.40121400e-01
-2.13195905e-01 -5.66249132e-01 1.52966872e-01 -1.09328583e-01
5.33601224e-01 5.18818676e-01 2.93056160e-01 -4.94428128e-01
-2.20600426e-01 7.63384283e-01 7.23410368e-01 -3.10531467e-01
1.32781780e+00 3.92374724e-01 -7.66397655e-01 2.79427081e-01
1.34582639e+00 -3.10993902e-02 -1.33153021e+00 1.68225765e-01
-4.76252139e-02 -1.05484352e-01 2.71666944e-01 -1.06633082e-01
-1.27756250e+00 1.00091088e+00 2.66410679e-01 4.70315069e-01
1.06155229e+00 -5.05373240e-01 7.39529788e-01 1.26465648e-01
6.84096813e-02 -9.31857169e-01 -6.22556627e-01 9.50202286e-01
9.63228047e-01 -5.73535085e-01 -4.74685878e-02 -3.91576380e-01
-1.62841693e-01 1.48124206e+00 2.77305335e-01 -6.06115341e-01
9.67342377e-01 1.20009708e+00 -8.02865326e-02 -1.13820523e-01
-3.33499521e-01 8.74686316e-02 4.19273600e-03 3.29535872e-01
8.88202190e-01 -1.88190982e-01 -7.85189748e-01 1.19283065e-01
-3.11607808e-01 2.92805612e-01 5.09627521e-01 1.24206829e+00
-2.33935371e-01 -1.03655064e+00 -7.20268905e-01 2.48729229e-01
-2.78138548e-01 -3.28519523e-01 8.30405299e-03 6.88505292e-01
-7.79661015e-02 1.73219204e-01 4.15715426e-01 8.54259059e-02
1.10847294e-01 1.24586046e-01 4.31784123e-01 -5.60938381e-02
-9.64905500e-01 -2.02320471e-01 1.48718745e-01 -5.84018469e-01
-3.26019317e-01 -5.32327771e-01 -1.84009182e+00 -6.03490770e-01
1.73076287e-01 3.11208218e-01 7.86441565e-01 9.02796388e-01
3.22634190e-01 5.45292497e-01 3.56118411e-01 -1.35201192e+00
-2.70283967e-01 -4.59787458e-01 -5.27643323e-01 1.26309963e-02
4.12197500e-01 -4.37564075e-01 -4.41833645e-01 7.94029608e-02] | [6.516347408294678, 3.3516931533813477] |
13511506-0da0-4aa5-94a5-597d4d583725 | density-propagation-and-improved-bounds-on | null | null | http://papers.nips.cc/paper/4723-density-propagation-and-improved-bounds-on-the-partition-function | http://papers.nips.cc/paper/4723-density-propagation-and-improved-bounds-on-the-partition-function.pdf | Density Propagation and Improved Bounds on the Partition Function | Given a probabilistic graphical model, its density of states is a function that, for any likelihood value, gives the number of configurations with that probability. We introduce a novel message-passing algorithm called Density Propagation (DP) for estimating this function. We show that DP is exact for tree-structured graphical models and is, in general, a strict generalization of both sum-product and max-product algorithms. Further, we use density of states and tree decomposition to introduce a new family of upper and lower bounds on the partition function. For any tree decompostion, the new upper bound based on finer-grained density of state information is provably at least as tight as previously known bounds based on convexity of the log-partition function, and strictly stronger if a general condition holds. We conclude with empirical evidence of improvement over convex relaxations and mean-field based bounds. | ['Ashish Sabharwal', 'Carla P. Gomes', 'Stefano Ermon', 'Bart Selman'] | 2012-12-01 | null | null | null | neurips-2012-12 | ['tree-decomposition'] | ['graphs'] | [ 1.72883883e-01 3.15031499e-01 -4.58311260e-01 -2.54705876e-01
-8.48580182e-01 -8.78762305e-01 5.66280425e-01 4.02920991e-01
-2.31510531e-02 1.04819357e+00 5.46329767e-02 -7.40460157e-01
-5.22855997e-01 -9.68019664e-01 -8.02707851e-01 -7.80833483e-01
-7.45228708e-01 9.79472399e-01 4.08590525e-01 2.76816368e-01
1.89673200e-01 7.11151481e-01 -1.11117840e+00 -1.28951505e-01
6.38035953e-01 6.79012895e-01 9.65962466e-03 1.16630661e+00
-7.38849640e-02 6.42415285e-01 -4.09005672e-01 -3.18081349e-01
-2.69639224e-01 -3.90024275e-01 -1.14512324e+00 4.63055931e-02
-2.07816139e-02 -2.62726426e-01 -5.03159106e-01 1.22649419e+00
9.23429430e-03 -4.84483577e-02 8.37813318e-01 -1.52899098e+00
-2.33274892e-01 1.02752364e+00 -7.78037488e-01 3.64240646e-01
5.12101710e-01 -4.96262908e-01 1.13833892e+00 -2.07138151e-01
5.61686397e-01 1.24140823e+00 6.04271829e-01 7.02981278e-02
-2.01254082e+00 -3.56533498e-01 3.15052718e-01 -6.15153685e-02
-1.65996099e+00 -8.07053670e-02 1.65559724e-01 -3.67829442e-01
8.82665575e-01 5.79003274e-01 4.54134494e-01 5.87216258e-01
3.08730274e-01 6.52764082e-01 1.06293118e+00 -5.84489405e-01
4.63979155e-01 -3.60053405e-02 4.49241281e-01 8.98224652e-01
8.47605646e-01 2.90816859e-03 -4.16291118e-01 -9.02195334e-01
8.00982058e-01 -1.96639538e-01 -3.60306948e-01 -7.54910707e-01
-8.76809299e-01 6.69924796e-01 -1.12544104e-01 5.74038178e-02
-6.96161985e-02 5.63771665e-01 3.08782212e-03 -2.07915053e-01
3.54625225e-01 -2.53666490e-01 -4.07259166e-01 -3.48645031e-01
-1.05902708e+00 3.89186412e-01 1.40374255e+00 1.05324602e+00
8.61296296e-01 -3.91014695e-01 -1.92525476e-01 2.87438661e-01
5.62613249e-01 8.37764919e-01 -7.00772285e-01 -9.60389912e-01
3.63086015e-01 1.04912281e-01 5.51663637e-01 -6.00961447e-01
-2.98915178e-01 -3.73886198e-01 -7.44214594e-01 -1.82154402e-01
6.03678048e-01 -1.37836829e-01 -9.07800317e-01 2.24777603e+00
3.06482971e-01 8.94951597e-02 -4.44918990e-01 3.84193063e-01
-2.11856142e-01 1.14072454e+00 1.03521277e-03 -7.03785717e-01
1.05971956e+00 -9.94085968e-02 -5.77429533e-01 1.01905717e-02
5.59920847e-01 -3.55359524e-01 3.64332885e-01 2.79067904e-01
-1.01625395e+00 2.66145349e-01 -8.85987341e-01 3.69050622e-01
1.11031771e-01 -4.48545933e-01 9.68535483e-01 1.23049629e+00
-1.24827564e+00 6.11695349e-01 -1.38766932e+00 -3.52049381e-01
-3.54093350e-02 3.25483114e-01 -1.62337720e-01 -9.17127579e-02
-7.29279101e-01 7.20716298e-01 4.87670690e-01 4.13555698e-03
-1.13948298e+00 -5.73125541e-01 -7.60518014e-01 3.58085454e-01
2.09040120e-01 -6.37381315e-01 1.26398194e+00 -2.85023055e-03
-1.22228372e+00 5.55167437e-01 -5.18011987e-01 -5.50929248e-01
8.19116756e-02 2.84667343e-01 -1.53978879e-03 1.99456781e-01
-7.12757756e-04 5.82386507e-03 3.08645844e-01 -1.43398631e+00
-5.42688608e-01 -4.19669449e-01 2.97843367e-01 -9.45134554e-03
2.00906575e-01 8.83547962e-03 -3.41321498e-01 2.01483309e-01
2.94473708e-01 -8.93950045e-01 -5.08612871e-01 -2.25596339e-01
-9.73764420e-01 -1.83635667e-01 2.95315653e-01 -2.70103484e-01
1.59836459e+00 -1.72295868e+00 1.65524811e-01 6.80298030e-01
3.11811596e-01 -3.56798232e-01 2.31146768e-01 8.26416373e-01
-1.72975212e-02 4.85773921e-01 -5.15520453e-01 -3.21629912e-01
3.20099205e-01 4.53229725e-01 -3.43211710e-01 8.86460900e-01
-1.16157122e-01 3.07663262e-01 -8.28052521e-01 -6.45209312e-01
1.66312307e-01 1.69050828e-01 -7.13968158e-01 -1.50864139e-01
-2.68429071e-01 -4.21876460e-02 -5.67058802e-01 1.73916996e-01
1.14844310e+00 -4.08495039e-01 6.97867393e-01 2.94805884e-01
-1.48323610e-01 3.23985457e-01 -1.38753915e+00 1.26676333e+00
-2.76955396e-01 2.29116008e-01 4.12133962e-01 -8.11560094e-01
2.69847959e-01 1.99136972e-01 4.18632001e-01 2.62653232e-01
1.85924202e-01 -1.43033117e-01 -2.39852130e-01 7.81926513e-02
3.89446378e-01 -4.43169475e-01 -4.51093763e-01 7.28859663e-01
1.87094733e-01 -2.33214814e-02 3.05475205e-01 8.74098837e-01
1.33049309e+00 -6.93041608e-02 2.71293432e-01 -5.85687220e-01
5.89537956e-02 -3.14167440e-01 3.64746571e-01 1.35032499e+00
-2.10613981e-01 2.34148249e-01 1.16933870e+00 3.44058722e-01
-9.56391275e-01 -1.66984737e+00 -5.43615282e-01 1.06567872e+00
2.79441953e-01 -1.00438011e+00 -8.42934608e-01 -2.96460003e-01
-6.00797161e-02 8.36135268e-01 -6.72588587e-01 -3.85378599e-02
1.68998365e-03 -1.09538460e+00 5.52518129e-01 3.91843408e-01
1.13450021e-01 -1.87345088e-01 -1.85019717e-01 1.72794968e-01
-1.54439270e-01 -1.04815304e+00 -3.30269694e-01 5.33190846e-01
-9.72058535e-01 -9.17246878e-01 -3.36622119e-01 -2.92052269e-01
5.62817752e-01 1.43009815e-02 8.96229148e-01 -1.54703543e-01
1.69377610e-01 5.51314354e-01 6.60974532e-02 5.77344634e-02
-5.30442476e-01 8.73314142e-02 6.09607296e-03 -2.43825659e-01
-1.02384895e-01 -8.12112391e-01 -1.48733094e-01 1.11844111e-02
-8.25640917e-01 3.29785608e-02 3.34819019e-01 4.57549959e-01
5.61429143e-01 4.96588111e-01 -1.37141243e-01 -7.99652636e-01
4.70095724e-01 -5.64828634e-01 -1.01393354e+00 3.27771068e-01
-3.50421160e-01 6.48799837e-01 2.78561473e-01 1.10052511e-01
-1.19226229e+00 1.90205574e-01 2.47666582e-01 1.71467587e-02
-1.03166156e-01 7.07917035e-01 -1.66684493e-01 1.16828367e-01
4.96584892e-01 2.88151711e-01 -3.19591254e-01 -2.76051432e-01
5.79111278e-01 2.69154459e-01 5.06992340e-01 -1.15237665e+00
7.26936758e-01 6.73806429e-01 6.45994782e-01 -7.73108244e-01
-9.20144618e-01 -3.27370852e-01 -4.94240284e-01 -6.74110325e-03
5.66236496e-01 -4.01834995e-01 -1.09366035e+00 3.12357128e-01
-1.21171641e+00 -3.68528396e-01 1.83631070e-02 3.89439553e-01
-7.67093480e-01 6.71170950e-01 -7.00641811e-01 -1.67710412e+00
2.58139193e-01 -6.63845837e-01 9.08853114e-01 6.19666465e-02
1.40363067e-01 -1.14010167e+00 5.09407938e-01 -1.29623502e-01
1.19490974e-01 2.94044852e-01 1.04527700e+00 -3.79428506e-01
-7.21606433e-01 -3.65265250e-01 -3.95933837e-01 -6.42753094e-02
-4.52679604e-01 3.87794405e-01 -5.56483865e-01 -2.74750620e-01
-2.56350011e-01 1.81053415e-01 8.70872140e-01 7.49775648e-01
9.67743278e-01 -4.36036289e-01 -9.19646859e-01 2.13517323e-01
1.62933397e+00 -2.28461280e-01 5.16035736e-01 -2.00969234e-01
4.41505879e-01 8.35990682e-02 6.58003166e-02 7.36171484e-01
5.46630621e-01 4.54651326e-01 1.99583709e-01 4.45048213e-01
3.41777056e-01 -5.49655676e-01 4.06356037e-01 4.46703821e-01
-1.66710913e-01 -3.98927599e-01 -9.71492589e-01 6.13856733e-01
-1.88877976e+00 -1.18731737e+00 -3.51154745e-01 2.63209367e+00
1.02557635e+00 4.11479592e-01 3.58763814e-01 4.47913520e-02
1.13446891e+00 -2.51171570e-02 -2.58535683e-01 -4.53026950e-01
1.87307581e-01 6.57634854e-01 1.13568735e+00 1.03352785e+00
-7.50234365e-01 5.71066320e-01 8.39918423e+00 9.33064282e-01
-2.72904664e-01 3.25061530e-01 2.35337436e-01 4.23290692e-02
-5.69275320e-01 6.28150403e-01 -8.46256614e-01 3.28990161e-01
1.29033065e+00 -5.43603480e-01 5.16063035e-01 7.96434820e-01
-3.87403667e-02 -6.66723013e-01 -1.21845067e+00 4.65519935e-01
-4.85464066e-01 -1.09897089e+00 -3.19275409e-01 5.94000041e-01
7.74324059e-01 -2.68428195e-02 -2.30196625e-01 4.03866591e-03
1.17725790e+00 -8.97296309e-01 7.59073794e-01 2.21796408e-01
6.02377594e-01 -1.07696009e+00 5.36294878e-01 6.23642802e-01
-1.43977404e+00 1.44210666e-01 -1.89986944e-01 -2.28493556e-01
5.86585939e-01 1.00646746e+00 -5.90961814e-01 5.01681566e-01
2.88827062e-01 3.13564241e-02 -2.68123914e-02 1.23564065e+00
-2.95325577e-01 1.03202856e+00 -1.10885358e+00 -8.39845762e-02
6.16977215e-02 -1.03706293e-01 6.74256742e-01 1.44652057e+00
6.92858696e-02 6.25812709e-02 2.86600739e-01 1.02381968e+00
1.17429860e-01 -5.10183334e-01 -3.63588363e-01 -7.50537440e-02
8.37851763e-01 1.11899126e+00 -8.99154842e-01 -3.11260670e-01
-1.49240941e-02 6.77417338e-01 3.90487194e-01 3.38109404e-01
-9.46315587e-01 -3.41960937e-01 5.81972599e-01 8.58381093e-02
6.04786336e-01 -5.94907045e-01 -2.23045349e-01 -1.04352140e+00
-1.46084249e-01 -3.10215130e-02 5.38993776e-01 -4.14427310e-01
-1.01451826e+00 5.18834963e-02 6.35552049e-01 -3.73029023e-01
-2.54072458e-01 -3.94918621e-01 -5.49920976e-01 8.77241313e-01
-9.19484675e-01 -8.11158597e-01 4.19414043e-01 3.79675090e-01
-4.30315375e-01 7.78678179e-01 9.73734081e-01 -9.26450789e-02
-5.34492075e-01 3.06611776e-01 2.68109202e-01 -1.79957762e-01
-1.22949786e-01 -1.60755348e+00 1.25455499e-01 1.19346285e+00
1.29632056e-01 6.78447247e-01 1.21615887e+00 -6.26642942e-01
-1.52042079e+00 -7.95694530e-01 7.37153769e-01 -5.10716200e-01
9.24301744e-01 -5.82716525e-01 -5.60022533e-01 1.15896213e+00
1.13523819e-01 -1.77997515e-01 6.13696039e-01 7.48127878e-01
-4.21111137e-01 3.64782244e-01 -1.22360599e+00 1.74208775e-01
1.11382914e+00 -5.81906736e-01 -2.56000072e-01 4.98354435e-01
4.39200878e-01 -3.19732100e-01 -9.94011939e-01 1.97746038e-01
4.46341515e-01 -1.00337029e+00 5.95883191e-01 -5.56339860e-01
-1.07716322e-01 -5.39679766e-01 -4.84028727e-01 -9.17725563e-01
-6.57103598e-01 -9.68770087e-01 -4.37590331e-01 1.10618091e+00
5.68031311e-01 -7.23801434e-01 7.82665670e-01 5.96663892e-01
1.50733277e-01 -7.71921217e-01 -1.15377843e+00 -1.02294242e+00
1.24549344e-01 -7.93423891e-01 2.98545480e-01 5.08571386e-01
5.92560053e-01 3.23109567e-01 -1.86156899e-01 7.06457555e-01
1.13656056e+00 1.25648558e-01 4.39093977e-01 -1.16695583e+00
-6.78581178e-01 -3.86311144e-01 -6.56827867e-01 -1.39131927e+00
7.53665641e-02 -9.14641738e-01 3.57185975e-02 -1.70723271e+00
9.14254189e-01 -5.43743670e-01 -1.81834757e-01 2.74118930e-01
2.03103647e-01 -2.06001341e-01 -7.33589903e-02 2.22330559e-02
-9.18980837e-01 3.48547429e-01 9.89277422e-01 3.36924523e-01
6.52454570e-02 3.11903358e-01 -6.63674414e-01 6.04611218e-01
5.16395152e-01 -6.72986865e-01 -2.06218466e-01 5.38397431e-02
5.54444551e-01 5.13989091e-01 3.56431246e-01 -7.37841785e-01
2.27589935e-01 -3.82032692e-01 -3.63691598e-02 -1.03264630e+00
3.02501738e-01 -3.98433119e-01 3.55329156e-01 4.51648533e-01
-2.64447570e-01 -3.96234870e-01 9.11517367e-02 1.02748907e+00
3.47193837e-01 -3.33795667e-01 7.44490564e-01 1.73434779e-01
-3.68802138e-02 3.66566062e-01 -8.66081417e-01 7.87558109e-02
9.31283593e-01 -7.56091252e-02 -2.48854309e-01 -7.31363118e-01
-9.90053892e-01 2.43457615e-01 3.54538530e-01 -5.58728635e-01
1.03998140e-01 -1.18711793e+00 -5.35797656e-01 -2.84931004e-01
-1.22169994e-01 -7.33688995e-02 2.29317337e-01 9.66284335e-01
-4.25735742e-01 4.98901457e-01 3.54142904e-01 -6.08516455e-01
-9.93177950e-01 5.13490736e-01 3.19387764e-01 -5.34125507e-01
-1.42187074e-01 7.94121742e-01 2.00959027e-01 -1.31447554e-01
7.44327605e-02 -4.94383991e-01 6.92375660e-01 -6.15589321e-01
4.57746238e-01 4.73302275e-01 -4.05564427e-01 -4.74562943e-01
-5.99537432e-01 1.87984988e-01 -9.75271165e-02 -6.70883715e-01
1.04914010e+00 -2.66723633e-01 -4.20431644e-01 4.48905766e-01
9.08120573e-01 7.98689798e-02 -1.13344049e+00 -1.11622766e-01
-1.76552162e-01 -2.98907667e-01 7.53158778e-02 -6.81108236e-01
-6.06328070e-01 6.06992781e-01 1.25475019e-01 9.73052025e-01
7.00197756e-01 7.33553827e-01 2.81224042e-01 4.05031204e-01
7.57762194e-01 -6.44005775e-01 -7.63002574e-01 5.34633994e-01
3.79716516e-01 -3.88461381e-01 1.31617293e-01 -7.56362975e-01
1.28478790e-02 7.86090136e-01 -7.14779869e-02 -4.42459323e-02
9.80504751e-01 8.03841650e-01 -1.02881134e+00 -9.28837582e-02
-9.99922991e-01 -3.16758752e-01 -2.35416993e-01 7.66876221e-01
3.56665909e-01 6.61170721e-01 -2.27886811e-01 5.70490122e-01
-3.73437881e-01 -8.25598761e-02 5.97610116e-01 9.64441478e-01
-8.62334013e-01 -1.02880049e+00 -4.50518608e-01 4.63068992e-01
-4.05029297e-01 -9.99263301e-03 -2.55137146e-01 5.48023820e-01
-3.76497656e-01 1.11305499e+00 1.43009812e-01 -1.42144747e-02
-8.02782252e-02 -5.00695929e-02 1.22422504e+00 -6.54378176e-01
3.29448462e-01 -2.61982650e-01 3.53296489e-01 -3.93197328e-01
-1.20865040e-01 -7.03828752e-01 -1.13621128e+00 -1.09414911e+00
-9.80494440e-01 3.22110564e-01 6.32159293e-01 1.08719516e+00
2.13341191e-01 1.25831902e-01 5.29123843e-01 -6.61309779e-01
-7.31757343e-01 -9.01763499e-01 -1.06552005e+00 -2.94516295e-01
1.21599190e-01 -7.04307795e-01 -5.85122526e-01 -1.99550197e-01] | [7.160953998565674, 4.8393778800964355] |
c9641e18-b4ac-46ae-a2b0-2ff15225303b | a-survey-on-video-action-recognition-in | 2206.01038 | null | https://arxiv.org/abs/2206.01038v1 | https://arxiv.org/pdf/2206.01038v1.pdf | A Survey on Video Action Recognition in Sports: Datasets, Methods and Applications | To understand human behaviors, action recognition based on videos is a common approach. Compared with image-based action recognition, videos provide much more information. Reducing the ambiguity of actions and in the last decade, many works focused on datasets, novel models and learning approaches have improved video action recognition to a higher level. However, there are challenges and unsolved problems, in particular in sports analytics where data collection and labeling are more sophisticated, requiring sport professionals to annotate data. In addition, the actions could be extremely fast and it becomes difficult to recognize them. Moreover, in team sports like football and basketball, one action could involve multiple players, and to correctly recognize them, we need to analyse all players, which is relatively complicated. In this paper, we present a survey on video action recognition for sports analytics. We introduce more than ten types of sports, including team sports, such as football, basketball, volleyball, hockey and individual sports, such as figure skating, gymnastics, table tennis, tennis, diving and badminton. Then we compare numerous existing frameworks for sports analysis to present status quo of video action recognition in both team sports and individual sports. Finally, we discuss the challenges and unsolved problems in this area and to facilitate sports analytics, we develop a toolbox using PaddlePaddle, which supports football, basketball, table tennis and figure skating action recognition. | ['Dejing Dou', 'Jun Cheng', 'Feixiang Lu', 'Ning Ding', 'Haoyi Xiong', 'Jian Bian', 'Qingzhong Wang', 'Fei Wu'] | 2022-06-02 | null | null | null | null | ['sports-analytics'] | ['computer-vision'] | [ 1.47300795e-01 -3.22847545e-01 -5.68101168e-01 9.82833207e-02
-1.27247810e-01 -6.10400200e-01 2.58006424e-01 4.85307351e-02
-4.48532879e-01 5.25363743e-01 2.75020391e-01 3.11837643e-02
-3.04352641e-01 -8.51474166e-01 -6.49225056e-01 -5.93020797e-01
-2.05415696e-01 1.50701493e-01 5.39973557e-01 -3.94156665e-01
3.72540027e-01 4.61091340e-01 -2.13779378e+00 7.04718351e-01
4.09798801e-01 9.26359117e-01 -1.19083755e-01 8.86820197e-01
-7.31978789e-02 1.38875663e+00 -6.79034531e-01 -5.66361785e-01
3.85874927e-01 -6.03783786e-01 -8.60869229e-01 3.49742889e-01
5.29516757e-01 -2.93490708e-01 -6.75264895e-01 9.24393713e-01
2.56086022e-01 3.28459442e-01 1.81359872e-01 -1.62614834e+00
-1.55184001e-01 5.30294657e-01 -4.08272266e-01 5.60882390e-01
8.38639736e-01 2.43712857e-01 6.43114805e-01 -5.50561726e-01
6.25917196e-01 1.09696090e+00 5.43256223e-01 3.86230737e-01
-4.04543102e-01 -8.24370682e-01 3.37477058e-01 8.60361278e-01
-1.05279708e+00 -1.66092873e-01 3.29353154e-01 -5.97605824e-01
6.21451914e-01 4.64405358e-01 1.47618186e+00 9.71324503e-01
2.65115947e-02 1.36404765e+00 8.67670059e-01 -2.76359886e-01
1.48136460e-04 -4.79959905e-01 3.40493798e-01 3.24095309e-01
1.27253875e-01 -2.06919417e-01 -8.57486129e-01 4.79416668e-01
8.51142049e-01 3.76984209e-01 2.43413132e-02 -1.76930085e-01
-1.45979178e+00 4.41604346e-01 -1.45822778e-01 1.23741567e-01
-2.98201799e-01 7.96674043e-02 7.48729944e-01 2.77239919e-01
3.99167836e-02 -3.58512178e-02 1.08115580e-02 -1.00135398e+00
-7.76692152e-01 6.34158432e-01 6.34240687e-01 7.33626306e-01
6.12023652e-01 1.73864663e-01 -2.39452556e-01 6.64668620e-01
-1.03619933e-01 1.90735117e-01 4.10885334e-01 -1.11291683e+00
6.23676002e-01 8.51350129e-01 -4.84192297e-02 -1.15466166e+00
-4.04539526e-01 1.82455987e-01 -7.31001198e-01 2.52876937e-01
5.91414213e-01 1.32157043e-01 -8.41560662e-01 1.01849174e+00
4.40203667e-01 5.56873202e-01 -7.30966311e-03 1.20389581e+00
1.08067095e+00 4.36351627e-01 1.96128666e-01 -2.87667125e-01
1.50886726e+00 -1.22832310e+00 -8.90997052e-01 -1.60627306e-01
6.05456471e-01 -7.82528102e-01 7.98443377e-01 9.10042703e-01
-1.01882708e+00 -8.80827904e-01 -6.59416378e-01 7.77416751e-02
-3.59717041e-01 2.76675552e-01 8.20769131e-01 6.37834787e-01
-3.44041467e-01 6.08187497e-01 -1.05383909e+00 -3.62848490e-01
3.71042043e-01 2.97482193e-01 -6.97809577e-01 -4.57804501e-02
-1.11893654e+00 7.00390577e-01 8.22220922e-01 1.03884190e-01
-8.20514679e-01 -4.60066020e-01 -9.33603108e-01 -2.10038513e-01
1.02398908e+00 -4.27014194e-03 1.11102748e+00 -1.02978802e+00
-1.38234115e+00 9.83228028e-01 1.46037966e-01 -5.38371861e-01
5.69332242e-01 -4.78256077e-01 -5.89769065e-01 2.15896413e-01
2.41861448e-01 1.32056877e-01 3.57469827e-01 -4.46726412e-01
-1.09819865e+00 -2.79412955e-01 6.42175972e-01 3.36340874e-01
-2.12454051e-01 3.68628085e-01 -6.82568848e-01 -7.37460077e-01
1.88629702e-01 -9.10173953e-01 -4.94660996e-02 -1.67847931e-01
-1.80530459e-01 -4.71788466e-01 8.50987375e-01 -5.65205157e-01
1.52073336e+00 -2.36581683e+00 2.64805883e-01 -5.62278442e-02
1.18711270e-01 4.98953551e-01 1.03536263e-01 5.84983647e-01
-1.46018013e-01 -1.01885550e-01 1.78251803e-01 3.89611781e-01
-1.72877386e-01 6.12730265e-01 9.92362797e-02 2.19278798e-01
-2.64024846e-02 6.89370394e-01 -1.01002538e+00 -6.78259194e-01
5.95598578e-01 -5.21733910e-02 -4.78639096e-01 -1.17107760e-02
1.71336457e-01 4.53297704e-01 -5.02330542e-01 1.03554595e+00
3.64929885e-01 2.00120509e-01 2.82305688e-01 -1.78739637e-01
-4.30043906e-01 1.02457963e-01 -1.91955149e+00 1.59251344e+00
1.98677797e-02 5.43479860e-01 2.97258317e-04 -1.47327352e+00
6.78216219e-01 2.19813034e-01 1.01704526e+00 -6.63016081e-01
8.87045451e-03 1.85565595e-02 9.64799449e-02 -1.11990559e+00
5.83505869e-01 1.70778297e-02 -8.09364468e-02 4.04036790e-01
-1.69309825e-01 3.11011344e-01 1.13618755e+00 1.17863029e-01
1.05624819e+00 4.26219106e-01 5.66736460e-01 3.81009936e-01
7.15136766e-01 2.50523329e-01 9.14724290e-01 5.56054354e-01
-3.29604596e-01 5.40859222e-01 5.39574027e-01 -9.47318017e-01
-6.61677539e-01 -9.39612269e-01 3.13701153e-01 1.23397744e+00
4.17080879e-01 -8.10726643e-01 -5.80073297e-01 -5.90430200e-01
-4.68755588e-02 -1.57597438e-01 -4.05166060e-01 -1.09131657e-01
-1.05009592e+00 -5.72814107e-01 5.64835906e-01 8.04624379e-01
9.28165078e-01 -1.24732637e+00 -6.23471916e-01 2.80911744e-01
-5.27394354e-01 -1.32141733e+00 -2.41326496e-01 -3.93296808e-01
-7.74185121e-01 -1.71594799e+00 -6.03803098e-01 -6.54557526e-01
1.87090948e-01 4.37168986e-01 9.97354984e-01 2.98342913e-01
-4.88614708e-01 6.10207081e-01 -8.34865093e-01 -5.12024760e-01
-2.38909543e-01 -1.95428655e-01 2.89072901e-01 2.09302381e-01
6.59946859e-01 -2.39200950e-01 -5.86030424e-01 6.30039454e-01
-9.29663002e-01 -5.01409755e-04 5.97832263e-01 3.76499087e-01
6.50905788e-01 1.76065966e-01 -7.12015107e-02 -4.26241279e-01
1.12111516e-01 -3.16460162e-01 -1.56966656e-01 3.06728780e-01
-1.54656228e-02 -6.29659474e-01 3.01638961e-01 -8.01040769e-01
-6.97004676e-01 2.64343560e-01 -1.56405121e-01 -4.64558691e-01
-3.77644062e-01 3.45421106e-01 -9.89240706e-02 -1.27009362e-01
6.46376669e-01 4.89620388e-01 1.88457534e-01 -5.86497426e-01
-8.89523998e-02 7.45600104e-01 4.83667672e-01 -3.94463509e-01
4.30182040e-01 5.07427692e-01 1.70031376e-02 -1.00873458e+00
-7.30665624e-01 -6.27257586e-01 -8.93786132e-01 -1.13966680e+00
1.28898585e+00 -9.10846293e-01 -1.26352084e+00 8.53666246e-01
-6.99702203e-01 -7.79530257e-02 -3.55382085e-01 9.90470409e-01
-5.27379870e-01 8.06919336e-01 -7.27890670e-01 -8.39649737e-01
1.62318751e-01 -1.02701902e+00 8.30713034e-01 4.57594842e-01
-1.40444398e-01 -5.34895122e-01 7.03787804e-02 9.56479251e-01
-2.06301764e-01 4.67545182e-01 1.56293616e-01 -4.06813860e-01
-6.47003889e-01 -3.09030682e-01 2.07735911e-01 4.38737690e-01
1.28332585e-01 2.56570309e-01 -1.82841673e-01 1.88343916e-02
-5.12345433e-01 -3.67684841e-01 5.53335309e-01 3.31087798e-01
1.30857348e+00 2.68312730e-02 -2.42804468e-01 4.32392329e-01
7.14361310e-01 4.76882577e-01 1.01334584e+00 6.55907214e-01
7.51042783e-01 6.17558777e-01 1.23921943e+00 3.82775813e-01
3.84713173e-01 8.54628980e-01 4.35683697e-01 1.04238562e-01
-1.10608734e-01 -1.31420538e-01 5.38515031e-01 7.89990842e-01
-1.22303474e+00 -1.62247252e-02 -8.04469645e-01 2.47214884e-01
-2.00071836e+00 -1.46889639e+00 -5.69476128e-01 2.01348519e+00
3.85523409e-01 6.62246123e-02 7.63160288e-01 8.00765395e-01
7.18899667e-01 1.49366334e-01 -2.40626335e-01 -1.54236197e-01
-3.55676003e-02 1.71892583e-01 4.35685545e-01 -8.16846192e-02
-1.50723779e+00 9.39940155e-01 6.06076050e+00 9.01173472e-01
-9.83488142e-01 -1.22684814e-01 9.50804725e-02 -2.88091630e-01
4.92438525e-01 -3.12903747e-02 -7.41670370e-01 5.56730568e-01
4.33635056e-01 -8.04591030e-02 7.84196183e-02 1.05761468e+00
2.78241426e-01 -2.84278572e-01 -8.49360228e-01 1.34926009e+00
2.32221067e-01 -1.37612283e+00 -6.68780282e-02 -3.05046495e-02
3.55012059e-01 -4.43897963e-01 -5.22162318e-01 5.99080503e-01
5.24606556e-02 -7.63602197e-01 7.67919004e-01 5.05397499e-01
3.08758199e-01 -5.86120844e-01 6.35759950e-01 5.03539443e-01
-1.46385157e+00 -3.75170201e-01 -2.56815434e-01 -7.50918150e-01
3.18058461e-01 1.38952836e-01 2.20352381e-01 7.80985832e-01
1.16711843e+00 1.29889822e+00 -4.14137453e-01 1.35022998e+00
-2.17593744e-01 4.63279545e-01 -5.80868460e-02 -3.58502418e-02
1.84366435e-01 -4.49096590e-01 3.79307419e-01 8.84666443e-01
3.19520593e-01 5.99486887e-01 7.77764559e-01 -1.00901358e-01
3.43039244e-01 2.48597935e-01 -3.55710506e-01 -2.49488071e-01
-1.47727624e-01 1.16859281e+00 -9.51501727e-01 -5.40444434e-01
-6.31634295e-01 7.17805088e-01 -1.40901476e-01 2.25991830e-02
-9.82434034e-01 -1.28171980e-01 9.81704473e-01 4.83152539e-01
-4.15731259e-02 -3.71166080e-01 3.13861251e-01 -1.44455445e+00
1.45520076e-01 -1.39505255e+00 8.19548607e-01 -5.37965298e-01
-7.41188049e-01 3.11729740e-02 3.04350674e-01 -1.91001546e+00
-2.38976181e-01 -8.16773355e-01 -4.38475102e-01 -5.44934385e-02
-7.19121456e-01 -8.78843129e-01 -5.61028481e-01 7.59294868e-01
9.56056476e-01 -2.10609332e-01 1.76841393e-01 8.32530022e-01
-7.53267288e-01 1.31092221e-01 -4.42186594e-01 6.22138798e-01
4.64845508e-01 -8.18748295e-01 -5.53581119e-02 8.01350057e-01
4.35024649e-01 1.27180234e-01 4.61224437e-01 -6.90810025e-01
-1.55930102e+00 -5.49338818e-01 4.67006743e-01 -2.86468029e-01
5.99049568e-01 -1.02099776e-01 -7.64175355e-01 7.37522185e-01
-2.28996247e-01 5.08377776e-02 8.19374084e-01 3.21877003e-02
1.05413266e-01 -2.95802236e-01 -5.09918988e-01 5.46709180e-01
1.49086010e+00 -1.36303738e-01 -5.79968035e-01 5.52699387e-01
-7.47983903e-02 -7.88799584e-01 -8.23458910e-01 3.09267104e-01
8.64242196e-01 -1.15682268e+00 1.15903151e+00 -1.09319758e+00
5.74994385e-01 -5.40424287e-01 2.43828416e-01 -6.98376954e-01
-2.24609330e-01 -2.69810885e-01 -1.18349172e-01 9.95957196e-01
-2.93355197e-01 -1.03207484e-01 1.24912155e+00 5.12853026e-01
-4.12454665e-01 -3.71936202e-01 -1.05058289e+00 -1.04609346e+00
-3.51903051e-01 -6.57623649e-01 3.71446818e-01 8.32901478e-01
2.64035076e-01 -2.35178411e-01 -9.46183443e-01 -2.66721129e-01
4.27563936e-01 1.32325485e-01 1.22898638e+00 -1.09284759e+00
-3.36406410e-01 -3.24355781e-01 -1.42062414e+00 -1.06585920e+00
-1.82260200e-01 -5.12406707e-01 -3.30423802e-01 -1.63651466e+00
6.80235028e-02 1.35433137e-01 -1.26077145e-01 5.72343528e-01
-6.97470158e-02 5.63094437e-01 4.82987434e-01 2.44751856e-01
-8.35885525e-01 -1.09778956e-01 1.49296141e+00 -2.15277195e-01
-4.91028465e-02 3.52670372e-01 -9.59237218e-02 1.00723588e+00
5.67063451e-01 -3.16671938e-01 -1.84357107e-01 -2.80661613e-01
3.47274482e-01 1.80540666e-01 3.91135722e-01 -1.35964239e+00
2.90798455e-01 -6.06720984e-01 1.82369694e-01 -4.98589367e-01
3.61989349e-01 -4.16187704e-01 3.19081753e-01 7.64341891e-01
3.23699340e-02 -1.21430196e-02 -1.40219674e-01 4.60747778e-01
-8.20214510e-01 -1.30159602e-01 5.24724960e-01 -4.63359535e-01
-1.38348639e+00 3.60730380e-01 -9.54856157e-01 7.02207163e-02
1.76893878e+00 -8.26431036e-01 -1.06875710e-01 -4.19542342e-01
-1.28777540e+00 3.80169660e-01 4.68039736e-02 7.49148309e-01
5.34078836e-01 -1.49711967e+00 -6.96364582e-01 4.36302461e-02
1.95155486e-01 -3.20415497e-01 8.47069502e-01 1.25545335e+00
-9.16879296e-01 1.60717145e-02 -7.36066997e-01 -5.98774076e-01
-1.88476813e+00 4.84853387e-01 4.99614999e-02 -3.08346990e-02
-7.46140778e-01 5.17777264e-01 -1.55027956e-01 -3.89748849e-02
3.56550664e-01 -3.88688177e-01 -8.82398486e-01 4.01793420e-01
7.72322536e-01 9.44074690e-01 -1.65866345e-01 -8.13934147e-01
-4.72004861e-01 7.60538280e-01 2.34613672e-01 5.34955442e-01
1.05891252e+00 1.49663508e-01 -3.87631766e-02 2.98599750e-01
5.04715025e-01 -1.59016222e-01 -1.02253270e+00 1.02806509e-01
-1.11276619e-01 -7.62829959e-01 -4.58056659e-01 -1.73487216e-01
-1.23974633e+00 8.55288088e-01 5.56937099e-01 4.87761945e-01
1.13127935e+00 -1.80677623e-01 7.65561402e-01 3.13952804e-01
5.96740901e-01 -1.66058540e+00 3.84851813e-01 4.27283973e-01
6.10729039e-01 -1.29415429e+00 2.08010063e-01 -7.10786283e-01
-8.93121064e-01 1.50590396e+00 8.11691880e-01 -1.64868459e-01
5.29973805e-01 2.49785230e-01 4.88333888e-02 -1.42610073e-01
-2.96182275e-01 -5.17469883e-01 2.68989652e-01 4.51166987e-01
3.57558161e-01 -2.39324383e-02 -8.31656218e-01 5.64432442e-01
-1.25696078e-01 4.06059176e-01 6.83009088e-01 1.27714813e+00
-5.02643585e-01 -1.37089491e+00 -6.75752461e-01 5.51072538e-01
-6.92966580e-01 4.31673169e-01 -3.22431564e-01 1.14668190e+00
6.34647012e-01 8.83525312e-01 1.92015409e-01 -6.44473672e-01
8.15956175e-01 -3.52641121e-02 4.33493495e-01 -5.46664119e-01
-7.70029724e-01 -2.19923154e-01 2.95336157e-01 -8.95483851e-01
-9.25666809e-01 -7.61718035e-01 -1.13193989e+00 -5.01970947e-01
1.43874779e-01 1.14950784e-01 2.67907143e-01 1.05097914e+00
-4.57856357e-02 2.81629503e-01 1.96772709e-01 -7.78821290e-01
1.83185861e-01 -7.61175632e-01 -8.55919182e-01 6.17158711e-01
-3.07664782e-01 -1.03038871e+00 -1.37256503e-01 2.82362431e-01] | [7.869900226593018, 0.32400843501091003] |
13a0010d-68b0-495b-855e-38ac6dfa28a2 | re-examining-linear-embeddings-for-high-1 | 2001.11659 | null | https://arxiv.org/abs/2001.11659v2 | https://arxiv.org/pdf/2001.11659v2.pdf | Re-Examining Linear Embeddings for High-Dimensional Bayesian Optimization | Bayesian optimization (BO) is a popular approach to optimize expensive-to-evaluate black-box functions. A significant challenge in BO is to scale to high-dimensional parameter spaces while retaining sample efficiency. A solution considered in existing literature is to embed the high-dimensional space in a lower-dimensional manifold, often via a random linear embedding. In this paper, we identify several crucial issues and misconceptions about the use of linear embeddings for BO. We study the properties of linear embeddings from the literature and show that some of the design choices in current approaches adversely impact their performance. We show empirically that properly addressing these issues significantly improves the efficacy of linear embeddings for BO on a range of problems, including learning a gait policy for robot locomotion. | ['Roberto Calandra', 'Benjamin Letham', 'Akshara Rai', 'Eytan Bakshy'] | 2020-01-31 | null | http://proceedings.neurips.cc/paper/2020/hash/10fb6cfa4c990d2bad5ddef4f70e8ba2-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/10fb6cfa4c990d2bad5ddef4f70e8ba2-Paper.pdf | neurips-2020-12 | ['misconceptions'] | ['miscellaneous'] | [-1.90402627e-01 -4.43542637e-02 -2.84735054e-01 -1.35600805e-01
-6.36924207e-01 -3.38492841e-01 3.23918462e-01 -5.11425324e-02
-5.68838954e-01 6.86442375e-01 1.87494904e-01 -2.40746081e-01
-3.34697872e-01 -5.91572344e-01 -7.35778928e-01 -7.91946173e-01
-4.31516945e-01 3.47116292e-01 2.14361832e-01 -1.43433154e-01
5.15106738e-01 5.55839300e-01 -1.41774035e+00 -4.90917623e-01
6.55947268e-01 7.82097399e-01 -1.42785842e-02 1.02417409e+00
2.50141412e-01 3.09234232e-01 -6.17236733e-01 -3.51322263e-01
1.78949147e-01 -1.33128464e-01 -6.28187060e-01 -1.04441203e-01
3.01765680e-01 -4.35268134e-01 -1.90076366e-01 9.41438794e-01
5.91028988e-01 2.94045210e-01 1.14864743e+00 -1.53188801e+00
-5.95908880e-01 2.31035396e-01 -1.82520658e-01 2.12506726e-01
2.02783886e-02 2.13570267e-01 1.20638740e+00 -9.64819312e-01
2.65703708e-01 1.50116456e+00 1.04992700e+00 4.52293545e-01
-1.47667444e+00 -2.31316313e-01 -1.36031121e-01 8.67489204e-02
-1.41642952e+00 -5.35307109e-01 5.99645853e-01 -6.08340144e-01
9.50035274e-01 -6.01611286e-02 7.60726094e-01 9.16492999e-01
5.77565670e-01 7.11331427e-01 7.82806516e-01 -3.31582040e-01
5.81456482e-01 4.03369218e-02 1.03524417e-01 6.55521214e-01
8.20876896e-01 3.10882479e-01 -5.96989274e-01 -5.42485952e-01
7.49661803e-01 -4.29863125e-01 -8.96762535e-02 -1.02164590e+00
-9.62869406e-01 1.25690973e+00 4.36664104e-01 -1.33957472e-02
-1.29701331e-01 6.77052498e-01 2.74013609e-01 -4.76808473e-02
2.69232959e-01 9.38785791e-01 -4.41460729e-01 -5.51704228e-01
-5.43533146e-01 5.98192692e-01 9.75869179e-01 8.37765455e-01
7.41547108e-01 4.87885438e-02 3.11797202e-01 7.02202916e-01
7.39306271e-01 3.10260803e-01 5.82136989e-01 -1.41462779e+00
2.86805600e-01 4.66326624e-02 4.96765554e-01 -1.25374413e+00
-4.45302188e-01 4.16415073e-02 -2.71127701e-01 3.10351789e-01
6.87398970e-01 -3.03855151e-01 -5.12527943e-01 1.69528627e+00
4.72477227e-01 -8.44592527e-02 3.69587615e-02 7.40814388e-01
8.41986164e-02 6.34769738e-01 -1.00232765e-01 2.53173918e-01
9.71286118e-01 -9.45261776e-01 -7.65342295e-01 -5.25349975e-01
7.61079550e-01 -5.46542108e-01 1.22622585e+00 3.56532156e-01
-9.02178109e-01 -1.61049768e-01 -1.45494592e+00 -1.66062817e-01
-4.13186282e-01 1.94745108e-01 6.00195050e-01 9.73013580e-01
-9.21616733e-01 8.99950027e-01 -1.09474564e+00 -4.95453089e-01
2.65576094e-01 5.31578183e-01 -1.63487673e-01 -4.35421877e-02
-8.01706016e-01 1.36479092e+00 3.37398201e-01 2.70080883e-02
-6.48759842e-01 -5.88557601e-01 -9.93305922e-01 -1.31068960e-01
1.84171394e-01 -5.87371349e-01 1.23927319e+00 -6.93434328e-02
-1.84142745e+00 2.91478604e-01 -1.19770635e-02 -5.35663605e-01
4.23362315e-01 -7.73255587e-01 8.79810601e-02 1.24895275e-01
-1.93592440e-02 8.04316938e-01 1.07920980e+00 -9.59654570e-01
-3.61202985e-01 -3.32524061e-01 -3.65016460e-02 1.54978842e-01
-6.54782832e-01 -3.96022439e-01 -2.03652740e-01 -6.32490575e-01
2.32639667e-02 -1.21674836e+00 -4.16746467e-01 2.45872498e-01
-9.24571976e-02 -2.35382989e-01 7.55377293e-01 -4.73363161e-01
1.25505459e+00 -2.11411023e+00 4.48740065e-01 -2.18342524e-02
2.90974565e-02 -1.30018830e-01 6.55701756e-02 4.87618864e-01
2.14906409e-01 3.08953434e-01 -3.77214253e-01 -3.69822770e-01
5.05631506e-01 5.19270241e-01 -2.00597048e-01 1.00291693e+00
4.84507173e-01 7.34499991e-01 -9.26696897e-01 -5.86607277e-01
3.32784474e-01 5.44779003e-01 -1.05081260e+00 2.60181665e-01
-4.64804582e-02 -4.66352887e-03 -3.94818872e-01 4.21756506e-01
3.86233658e-01 -2.17607487e-02 4.37926874e-02 -1.83301359e-01
7.85972401e-02 3.64521116e-01 -1.30397737e+00 1.46135378e+00
-3.29570025e-01 7.51968443e-01 1.52789550e-02 -1.02138066e+00
6.96253955e-01 -1.29031435e-01 7.06137538e-01 4.04921435e-02
7.37930387e-02 2.33960584e-01 -2.52548195e-02 -5.12937903e-01
6.56083524e-01 -1.60376072e-01 -3.77719253e-02 3.06407362e-01
1.02171294e-01 -5.30429542e-01 8.92459676e-02 -2.30938286e-01
1.09198487e+00 2.85969198e-01 3.98073405e-01 -6.13383591e-01
1.71397310e-02 1.17797069e-01 5.98592877e-01 6.51383877e-01
-5.44747531e-01 5.65562367e-01 4.42541957e-01 -9.31127220e-02
-1.22067988e+00 -1.17560363e+00 -4.41906095e-01 1.02693748e+00
9.80833098e-02 -6.22425735e-01 -6.52786374e-01 -4.50364023e-01
3.95828962e-01 6.77593708e-01 -6.51899099e-01 -3.69410068e-01
-6.14054441e-01 -1.02653277e+00 5.13203621e-01 8.19285035e-01
5.06788641e-02 -4.70609456e-01 -1.06649351e+00 3.66940886e-01
1.24499239e-01 -8.44339550e-01 -4.52745199e-01 4.75069016e-01
-1.17793143e+00 -9.46358562e-01 -5.32267809e-01 -4.64009732e-01
6.42503262e-01 2.74702418e-03 8.72768581e-01 -3.24153513e-01
-4.18992013e-01 5.13718545e-01 -1.80378288e-01 -1.97044328e-01
-2.59728432e-01 1.08473592e-01 3.55352044e-01 -5.93081117e-01
3.42843741e-01 -3.57449055e-01 -4.44395959e-01 4.53518331e-01
-7.22833276e-01 -4.44794774e-01 5.01347721e-01 1.08858359e+00
2.32539624e-01 1.03226475e-01 4.16979343e-01 -3.67768049e-01
9.01012301e-01 -4.37187344e-01 -6.91855907e-01 -1.44394651e-01
-6.90892398e-01 6.60491943e-01 3.61310214e-01 -6.21446848e-01
-3.71961474e-01 -1.03178717e-01 -1.63640738e-01 -3.30432802e-01
1.72248736e-01 3.66207510e-01 -7.56825134e-02 -2.74779499e-01
6.21801436e-01 -3.56590182e-01 2.55306989e-01 -4.23257113e-01
6.48006797e-01 5.17418146e-01 2.00814024e-01 -1.04328334e+00
6.39749289e-01 2.99779415e-01 1.85817212e-01 -1.07274032e+00
-5.67716122e-01 -3.29500675e-01 -6.34383440e-01 9.97934118e-03
7.93662667e-01 -5.69400430e-01 -6.58845723e-01 4.88101989e-02
-8.52659225e-01 -4.86917257e-01 -4.21462357e-01 6.76422298e-01
-1.07312381e+00 3.07208598e-01 -5.02922297e-01 -8.50318551e-01
1.99005455e-01 -1.42541063e+00 9.92020786e-01 9.01043341e-02
-5.64604878e-01 -1.19232523e+00 4.72331673e-01 -1.00427652e-02
5.34397364e-01 4.37126532e-02 1.02313173e+00 -3.17602098e-01
-1.92766502e-01 -6.81564286e-02 1.04838900e-01 4.53629911e-01
9.88342464e-02 2.69754112e-01 -8.90301645e-01 -3.62816095e-01
7.11305961e-02 -4.26156670e-01 6.57598019e-01 6.53583646e-01
9.78053808e-01 -3.09949338e-01 -3.31989408e-01 5.35555840e-01
1.28081381e+00 -6.07959181e-02 2.95954734e-01 5.97018003e-01
5.02316415e-01 6.49453580e-01 8.05136800e-01 4.93176937e-01
2.80157596e-01 6.41579449e-01 3.25095415e-01 4.54870224e-01
3.09081614e-01 -3.08411151e-01 7.32736468e-01 8.62685978e-01
5.40789962e-01 -4.23795432e-02 -8.68421137e-01 7.31089354e-01
-1.90951157e+00 -5.33694446e-01 3.14987630e-01 2.28118753e+00
7.95580745e-01 1.67713240e-01 3.50311786e-01 1.73964754e-01
6.18693650e-01 9.60562900e-02 -7.01765954e-01 -5.54049373e-01
1.30436897e-01 4.30958811e-03 8.66011083e-01 6.02785587e-01
-1.25254774e+00 7.09081888e-01 8.23336506e+00 5.30488908e-01
-7.66434908e-01 -1.98668689e-02 2.51524031e-01 -2.03062855e-02
1.02361448e-01 -3.81511897e-02 -1.04713213e+00 3.36411446e-01
1.13771474e+00 -2.44891420e-01 4.18729603e-01 1.12129104e+00
1.28289476e-01 -1.79027647e-01 -1.50219095e+00 9.01538193e-01
7.11108148e-02 -1.06674731e+00 -3.23472291e-01 2.10210443e-01
5.63865125e-01 -9.81654078e-02 3.23132634e-01 3.38455349e-01
5.05683601e-01 -1.30639243e+00 5.08888543e-01 1.76825672e-01
2.86419958e-01 -7.39762723e-01 4.83630061e-01 9.01458412e-02
-1.06145334e+00 -3.21936280e-01 -6.52860641e-01 4.74757478e-02
2.27297038e-01 4.38056082e-01 -8.15476477e-01 -3.95515449e-02
8.40144396e-01 7.08668411e-01 -5.61177731e-01 1.25704372e+00
-1.15689836e-01 7.10280418e-01 -7.90695786e-01 -3.82619530e-01
2.97413260e-01 -3.96506697e-01 6.25406325e-01 1.05463660e+00
3.35668802e-01 -4.26103711e-01 -1.17251784e-01 8.60534489e-01
3.06663156e-01 -1.62585050e-01 -1.01383281e+00 -4.05826330e-01
5.52142143e-01 6.98981404e-01 -3.34453493e-01 6.75104633e-02
-1.90391436e-01 5.45902252e-01 3.68074834e-01 2.91214287e-01
-7.77024150e-01 -5.53411961e-01 1.10689867e+00 -9.08423141e-02
5.99304497e-01 -8.87543619e-01 -3.94080251e-01 -8.62397254e-01
2.02761181e-02 -5.84202230e-01 3.22567999e-01 -4.98905599e-01
-1.44504750e+00 -1.54046878e-01 4.36576128e-01 -1.02815807e+00
-3.57357174e-01 -1.10530078e+00 -3.21529210e-01 4.72220063e-01
-1.19318092e+00 -4.90112096e-01 1.02541588e-01 2.68180016e-02
3.84251177e-01 8.46875012e-02 7.06006587e-01 1.63236752e-01
-6.34693146e-01 5.09647310e-01 6.63136363e-01 -1.53616905e-01
6.98903918e-01 -1.57139635e+00 4.53275174e-01 6.10007763e-01
-5.11417631e-03 9.20223296e-01 1.14633739e+00 -3.74818057e-01
-1.85880065e+00 -8.13663185e-01 3.36272180e-01 -6.51505411e-01
9.08900976e-01 -3.47854882e-01 -6.26244605e-01 7.44737327e-01
-7.09259585e-02 -1.61777601e-01 7.54650295e-01 2.57095486e-01
6.44543953e-03 1.26047790e-01 -1.22850430e+00 9.57567930e-01
7.08110452e-01 -4.46699470e-01 -8.44061375e-01 1.37009650e-01
5.96614182e-01 -2.24580243e-01 -1.16486704e+00 3.77999574e-01
8.28101873e-01 -6.26674712e-01 1.32917166e+00 -7.75159717e-01
1.85326248e-01 -2.70252943e-01 -4.62605149e-01 -1.51130247e+00
-1.53098598e-01 -7.75833070e-01 -4.70985830e-01 8.30683351e-01
8.69671069e-03 -8.61193597e-01 8.79509509e-01 9.17615294e-01
4.71418723e-02 -8.82636786e-01 -1.13943565e+00 -1.03139639e+00
7.01005936e-01 -3.57304692e-01 2.56704658e-01 4.29652750e-01
1.84739798e-01 1.78832889e-01 -3.14578086e-01 3.02959353e-01
6.99302614e-01 -3.29554826e-01 1.06464386e+00 -9.72109973e-01
-2.19370842e-01 -4.78407085e-01 -6.64264202e-01 -1.50027013e+00
9.78333727e-02 -5.14986813e-01 3.63462448e-01 -1.20384300e+00
-2.73727506e-01 -5.35492063e-01 -1.49999350e-01 1.68770887e-02
-9.09022912e-02 -1.31412685e-01 -4.13259752e-02 -6.43382221e-02
-2.04352066e-01 1.04357147e+00 7.94143558e-01 -1.16520689e-03
-1.25223756e-01 -3.17706257e-01 -4.89055097e-01 7.98227608e-01
9.15004492e-01 -7.16756761e-01 -3.25393617e-01 -4.71562177e-01
8.90802145e-02 -4.59841549e-01 2.45313480e-01 -1.01475692e+00
1.46234050e-01 -2.96242148e-01 2.28921399e-01 -5.68959534e-01
6.95220232e-01 -6.71456158e-01 -2.40512311e-01 4.26542580e-01
-2.97301590e-01 5.14927566e-01 3.48433733e-01 8.00368667e-01
1.16330430e-01 -7.34010458e-01 8.72207761e-01 2.46426493e-01
-4.96071011e-01 -6.40942995e-03 -6.04328275e-01 3.65214586e-01
9.65366602e-01 -1.96191326e-01 -2.46849760e-01 -1.55437186e-01
-4.99268532e-01 2.32649893e-01 6.55280650e-01 3.35517138e-01
5.65919042e-01 -1.45551467e+00 -1.76757261e-01 3.58649716e-02
1.95852909e-02 2.35038102e-02 -4.22709495e-01 7.49533474e-01
-8.84800375e-01 4.72652704e-01 -2.64384210e-01 -6.19181573e-01
-8.69344056e-01 3.39316875e-01 3.43522519e-01 9.09916759e-02
-6.03256047e-01 7.36811697e-01 -2.71273941e-01 -6.12086654e-01
3.88858646e-01 -6.13980949e-01 1.90509588e-01 -1.54524490e-01
8.77330229e-02 8.02248836e-01 -2.89509505e-01 -4.00368065e-01
-3.73432338e-01 6.31329358e-01 1.80846587e-01 -6.96432829e-01
1.42707455e+00 3.86442919e-03 1.58524752e-01 9.32367802e-01
1.47246146e+00 -2.84729064e-01 -1.44125366e+00 1.20553955e-01
2.89752066e-01 -6.70717537e-01 2.17454717e-01 -5.23391254e-02
-4.84316647e-01 1.11523688e+00 6.83634400e-01 4.38465141e-02
3.97562504e-01 -3.49065572e-01 6.50281489e-01 8.43436062e-01
3.84149283e-01 -1.46037877e+00 3.27835292e-01 4.36980397e-01
7.76545227e-01 -1.19759369e+00 2.50074834e-01 -1.07631534e-01
-5.51008284e-01 1.30602920e+00 4.84067589e-01 -4.63292748e-01
9.00418937e-01 3.06834966e-01 -2.98870534e-01 -1.30324349e-01
-6.64583683e-01 9.96968821e-02 2.13607382e-02 7.07255900e-01
1.82686731e-01 -1.13633923e-01 -2.83663273e-01 2.57795364e-01
-3.87650609e-01 -2.66874999e-01 4.75406259e-01 1.33201778e+00
-5.93280077e-01 -1.10150898e+00 -5.53603828e-01 4.27290559e-01
-2.61222690e-01 2.46132910e-01 -1.86979026e-01 7.96414435e-01
-3.65257025e-01 1.01413369e+00 -5.83167411e-02 -3.03116888e-01
3.45151089e-02 1.54373527e-01 6.84358299e-01 -4.80314046e-01
-5.15444428e-02 -1.78486094e-01 1.09789245e-01 -6.40107214e-01
-2.02965796e-01 -9.79575634e-01 -1.01130319e+00 -2.12529436e-01
-4.80520934e-01 1.93071321e-01 9.53046262e-01 8.41732800e-01
2.82079697e-01 2.03896463e-01 3.31199408e-01 -1.11377645e+00
-1.19834673e+00 -7.88826168e-01 -4.90693599e-01 4.01256830e-02
5.87196708e-01 -1.31092155e+00 -6.74399257e-01 -1.04394943e-01] | [6.750175952911377, 4.0211639404296875] |
efc34d0f-30d9-41de-8128-ded50cdf7636 | coarse-to-fine-knowledge-selection-for | 2302.11849 | null | https://arxiv.org/abs/2302.11849v1 | https://arxiv.org/pdf/2302.11849v1.pdf | Coarse-to-Fine Knowledge Selection for Document Grounded Dialogs | Multi-document grounded dialogue systems (DGDS) belong to a class of conversational agents that answer users' requests by finding supporting knowledge from a collection of documents. Most previous studies aim to improve the knowledge retrieval model or propose more effective ways to incorporate external knowledge into a parametric generation model. These methods, however, focus on retrieving knowledge from mono-granularity language units (e.g. passages, sentences, or spans in documents), which is not enough to effectively and efficiently capture precise knowledge in long documents. This paper proposes Re3G, which aims to optimize both coarse-grained knowledge retrieval and fine-grained knowledge extraction in a unified framework. Specifically, the former efficiently finds relevant passages in a retrieval-and-reranking process, whereas the latter effectively extracts finer-grain spans within those passages to incorporate into a parametric answer generation model (BART, T5). Experiments on DialDoc Shared Task demonstrate the effectiveness of our method. | ['Cam-Tu Nguyen', 'Yongbin Li', 'Haiyang Yu', 'Cheng Fu', 'Haomin Fu', 'Yeqin Zhang'] | 2023-02-23 | null | null | null | null | ['answer-generation'] | ['natural-language-processing'] | [-8.43926445e-02 1.84184343e-01 -2.62837470e-01 -7.62303919e-02
-1.34984004e+00 -7.68613040e-01 9.32045460e-01 3.36718768e-01
-4.36173737e-01 1.24060452e+00 7.89589405e-01 -8.98956060e-02
-4.63030875e-01 -9.28642511e-01 -2.20395982e-01 -3.92753065e-01
1.97965994e-01 9.47763026e-01 5.12002051e-01 -6.34258091e-01
7.04691708e-01 1.87361062e-01 -1.30159044e+00 6.40258253e-01
1.19229245e+00 5.09515107e-01 3.97999406e-01 8.18792582e-01
-8.09632361e-01 9.96401131e-01 -1.00195599e+00 -2.67839819e-01
-3.26201260e-01 -4.38840181e-01 -1.63442290e+00 -9.37772244e-02
1.28981322e-01 -7.17370093e-01 -1.11405075e-01 7.44332075e-01
6.06280863e-01 4.74078238e-01 7.10759997e-01 -8.21229935e-01
-6.70829237e-01 9.04224336e-01 -9.38973352e-02 1.00481607e-01
8.22716296e-01 -2.53269851e-01 1.21085727e+00 -8.68411183e-01
6.79176271e-01 1.68028057e+00 3.42742592e-01 5.38806915e-01
-9.01586235e-01 -9.97072682e-02 1.04630955e-01 1.81493834e-01
-1.29048645e+00 -5.06644964e-01 6.01910949e-01 -2.40042537e-01
1.27188265e+00 4.51542258e-01 3.36705595e-01 7.98512936e-01
-2.16324955e-01 1.15048039e+00 7.01175153e-01 -7.16113448e-01
1.35334209e-01 3.25217009e-01 4.74406362e-01 6.91680908e-01
4.76840362e-02 -4.59405124e-01 -8.25243533e-01 -7.11740017e-01
7.02298045e-01 -3.80574167e-01 -4.55223978e-01 8.09664875e-02
-1.13500893e+00 1.10886371e+00 -1.09559521e-01 5.01754999e-01
-4.50384706e-01 -4.02207762e-01 4.84904587e-01 3.27893496e-01
4.99343753e-01 8.71501148e-01 -5.68382621e-01 -5.23218155e-01
-7.97830641e-01 6.58301592e-01 1.38572168e+00 1.22391152e+00
7.93101490e-01 -6.14369392e-01 -8.08229208e-01 1.22555876e+00
2.44829252e-01 4.59083080e-01 6.11011624e-01 -1.07763493e+00
6.90307796e-01 8.52845490e-01 4.88604009e-01 -8.06999385e-01
-2.75797725e-01 -2.33183112e-02 -4.48130041e-01 -7.17170179e-01
3.63716841e-01 -2.39025667e-01 -3.31612825e-01 1.51086462e+00
4.11455542e-01 -3.95096451e-01 5.74472368e-01 7.84848571e-01
1.07233274e+00 8.51006269e-01 -9.04592499e-02 -4.06883240e-01
1.56118691e+00 -1.19312918e+00 -8.29312921e-01 -9.43759233e-02
9.14883196e-01 -8.91280174e-01 1.14583504e+00 9.19183493e-02
-1.29660618e+00 -3.41034651e-01 -6.41515255e-01 -3.79825801e-01
-4.08273667e-01 4.09808531e-02 4.45529282e-01 4.63881940e-01
-1.14540529e+00 2.60610551e-01 -7.38248155e-02 -2.85053521e-01
-1.98191911e-01 -6.72825566e-03 1.92267194e-01 -5.99143431e-02
-1.76467454e+00 9.93277490e-01 5.97255766e-01 -5.82969561e-02
-6.72358572e-01 -5.43940306e-01 -6.18841827e-01 1.14626974e-01
6.91377699e-01 -1.00620627e+00 1.68501723e+00 -3.47862005e-01
-1.84325993e+00 4.38789636e-01 -2.77965933e-01 -2.56892830e-01
1.24207444e-01 -4.16845709e-01 -5.18317372e-02 4.64580208e-01
2.74389863e-01 5.24868667e-01 5.58881462e-01 -1.26823473e+00
-9.15457487e-01 -1.82287723e-01 6.52667940e-01 8.47695351e-01
-4.68810409e-01 6.13913983e-02 -7.73364842e-01 -4.33798552e-01
-2.34614775e-01 -8.03894699e-01 -5.34325140e-03 -8.34084332e-01
-5.90604603e-01 -8.25744569e-01 6.05275214e-01 -6.74691439e-01
1.67416811e+00 -1.39223468e+00 3.32004339e-01 1.90991923e-01
2.79739320e-01 4.51423168e-01 -1.85281798e-01 1.00423539e+00
8.28084230e-01 1.55474365e-01 1.10852368e-01 -6.80102333e-02
2.79172152e-01 2.61718363e-01 -5.26291549e-01 -4.54664946e-01
-1.84793487e-01 9.42112565e-01 -1.20773244e+00 -8.43856275e-01
-2.08837032e-01 1.21502653e-01 -3.94392490e-01 3.21156591e-01
-6.37413800e-01 1.81629494e-01 -1.11047578e+00 4.93899047e-01
1.94746956e-01 -2.58644998e-01 1.47639722e-01 4.82726516e-03
-1.61347687e-01 5.46063960e-01 -8.89806747e-01 1.70385265e+00
-7.56457746e-01 2.40163580e-01 -1.18447039e-02 -4.09935772e-01
7.65736699e-01 4.51359570e-01 1.94219097e-01 -6.66013479e-01
-1.57571286e-01 2.16531619e-01 -3.39894861e-01 -6.76254988e-01
1.27311301e+00 1.38591696e-02 -3.25952232e-01 7.89749146e-01
-7.86709115e-02 -2.67240375e-01 5.10980964e-01 7.26806223e-01
9.94224608e-01 -7.52065182e-02 3.50429296e-01 -2.03225523e-01
7.56095827e-01 1.90460071e-01 1.10278182e-01 1.19171989e+00
1.62268639e-01 2.03568250e-01 3.22852343e-01 1.12604499e-01
-6.29004538e-01 -6.84595168e-01 1.50764167e-01 1.50989604e+00
1.35554045e-01 -7.50622213e-01 -9.35005248e-01 -8.59554350e-01
2.53657121e-02 8.49287212e-01 -1.93495721e-01 -1.92044452e-01
-6.12999141e-01 -4.48848397e-01 8.72829616e-01 2.64045268e-01
7.23351598e-01 -1.20287931e+00 -2.71290451e-01 5.60592353e-01
-8.79282355e-01 -9.09083366e-01 -7.95624435e-01 -2.45806277e-01
-7.79331028e-01 -9.42545652e-01 -9.37702060e-01 -6.89003289e-01
2.96880037e-01 5.00322878e-01 1.32679021e+00 2.23745897e-01
1.26241222e-01 8.04659367e-01 -8.48303854e-01 -2.09125102e-01
-5.06325543e-01 5.30574501e-01 -8.44795555e-02 -2.41381258e-01
4.39611465e-01 -1.25619590e-01 -4.22395617e-01 3.48944515e-01
-8.87311161e-01 -1.33875096e-02 5.34605086e-01 1.09578443e+00
2.52836674e-01 -3.09801027e-02 1.11539268e+00 -8.12504232e-01
1.72705579e+00 -4.77161676e-01 -2.17828691e-01 8.42504501e-01
-6.04175508e-01 4.58663225e-01 4.71479505e-01 -1.37308240e-01
-1.53118098e+00 -6.61119282e-01 -8.19728300e-02 1.99747309e-01
5.14284298e-02 9.19870973e-01 -1.12595320e-01 1.70626000e-01
6.92756653e-01 5.74141800e-01 -1.63875878e-01 -5.26033461e-01
7.25707531e-01 1.06642258e+00 1.40879244e-01 -1.29580140e+00
2.66439289e-01 2.86242552e-02 -5.33304453e-01 -1.02390206e+00
-9.74691093e-01 -9.75264072e-01 -6.53297484e-01 -1.86827630e-01
4.54523474e-01 -6.37106419e-01 -5.89912117e-01 1.86577007e-01
-1.39674866e+00 -2.34340474e-01 -2.54578263e-01 2.70940185e-01
-6.45282507e-01 6.73487842e-01 -7.07649171e-01 -8.06826472e-01
-8.40171874e-01 -7.14385688e-01 1.28338587e+00 3.71638656e-01
-4.77429509e-01 -1.16353500e+00 3.93733978e-01 7.22202122e-01
3.50421309e-01 -3.13900769e-01 1.13730073e+00 -9.20323551e-01
-5.35784781e-01 2.13142727e-02 -1.71485588e-01 6.51238710e-02
2.90616959e-01 -3.10799837e-01 -7.44663060e-01 -1.40378699e-01
-1.39026776e-01 -6.57802999e-01 8.29968452e-01 1.49565851e-02
7.61525154e-01 -8.21999311e-01 -3.57512295e-01 -2.85351634e-01
8.00458133e-01 3.73623878e-01 5.40999889e-01 3.76880616e-01
2.97010154e-01 8.49167168e-01 8.32506239e-01 5.60676038e-01
7.30960250e-01 8.08505058e-01 -3.85307550e-01 3.86371642e-01
-7.90654048e-02 -7.79469684e-02 1.20724104e-01 1.12268817e+00
-1.65941253e-01 -3.96551400e-01 -8.05983841e-01 7.76075244e-01
-2.20113134e+00 -1.11430001e+00 2.80951262e-01 1.89128101e+00
1.48072994e+00 -1.53117687e-01 3.91875133e-02 -2.95815170e-01
6.70613945e-01 9.31203067e-02 -4.29151237e-01 -3.87936950e-01
-6.65761530e-02 -4.73597124e-02 -1.68053657e-01 8.17928135e-01
-5.35753489e-01 1.12829494e+00 5.98461151e+00 1.18227124e+00
-2.92604387e-01 -6.12485781e-02 5.48935682e-02 4.99866195e-02
-4.85881537e-01 -4.14596349e-02 -1.31594312e+00 2.36728400e-01
9.24456418e-01 -6.51751816e-01 3.00527215e-01 6.24118209e-01
1.07157655e-01 -3.03448081e-01 -9.94355261e-01 5.55840254e-01
1.49670064e-01 -1.36274958e+00 5.73763371e-01 -1.68564007e-01
8.21646631e-01 -5.33685505e-01 -3.64184111e-01 7.27967381e-01
5.12994170e-01 -6.65146351e-01 4.01207924e-01 9.02595937e-01
2.28912786e-01 -8.09391320e-01 7.04083800e-01 6.93635702e-01
-1.01480532e+00 2.05915630e-01 -3.81664783e-01 2.45614469e-01
3.17532182e-01 3.24328363e-01 -1.34393835e+00 7.66281664e-01
3.56949955e-01 1.94575280e-01 -3.11324090e-01 7.70578027e-01
-1.41930804e-01 3.20159674e-01 -1.98144279e-02 -6.25479043e-01
6.03129148e-01 -5.34418318e-03 5.75721085e-01 1.53171623e+00
1.69175610e-01 5.35622537e-01 2.39691943e-01 5.64878464e-01
-1.39789969e-01 3.65662515e-01 -3.75583678e-01 -1.10811502e-01
7.51273572e-01 1.18199372e+00 -3.53093654e-01 -6.96056008e-01
-2.25263923e-01 8.38858545e-01 4.51284677e-01 3.75083536e-01
-2.59307534e-01 -7.74070203e-01 4.30025309e-01 -2.76545763e-01
8.26656818e-02 -2.32384607e-01 1.95439443e-01 -1.12922084e+00
1.38023794e-02 -1.09464598e+00 5.82007527e-01 -5.78474760e-01
-1.36710072e+00 5.10507286e-01 3.10024500e-01 -1.03317499e+00
-8.35817814e-01 -8.88992399e-02 -2.25765392e-01 9.71626937e-01
-1.69308507e+00 -9.57545817e-01 -8.72402191e-02 7.59901047e-01
8.70845914e-01 -1.49685457e-01 1.15894032e+00 6.43636286e-02
-2.20447302e-01 3.89401138e-01 4.14365947e-01 9.46236122e-03
7.85492659e-01 -1.37085056e+00 -6.18177392e-02 2.58294761e-01
8.35950207e-03 1.13787913e+00 4.14267719e-01 -7.81068385e-01
-1.28104639e+00 -8.43334138e-01 1.25334775e+00 -4.89149660e-01
4.70541418e-01 -7.23574962e-03 -1.07708263e+00 3.15500408e-01
4.06960964e-01 -8.82236600e-01 8.53602231e-01 4.22947764e-01
-1.88653260e-01 1.61994949e-01 -9.63919938e-01 6.79218709e-01
8.73846054e-01 -7.91863799e-01 -1.20411623e+00 5.11259556e-01
8.09886575e-01 -4.84113872e-01 -9.72094536e-01 2.86103729e-02
4.74491715e-01 -6.26827598e-01 1.06692111e+00 -6.61923289e-01
2.68175691e-01 -2.03403398e-01 -7.84085542e-02 -1.56444764e+00
-1.02002971e-01 -7.01143026e-01 -5.12486815e-01 1.45278692e+00
5.58046103e-01 -5.09436905e-01 3.92010003e-01 7.03191459e-01
-2.40499601e-01 -6.49651349e-01 -6.83233380e-01 -6.59107029e-01
8.45966265e-02 -2.23409366e-02 8.71099710e-01 8.49794686e-01
5.91219842e-01 7.47319341e-01 -3.76968235e-01 -7.73621127e-02
3.97784084e-01 4.73031104e-01 7.33052671e-01 -1.12935770e+00
-1.56509459e-01 -5.31409025e-01 3.97480518e-01 -1.63532138e+00
2.64537901e-01 -6.98196471e-01 2.50728011e-01 -2.05096316e+00
2.19195098e-01 -4.70701069e-01 3.85404192e-02 3.85927856e-01
-4.08313721e-01 -3.65261704e-01 -5.70846349e-02 4.03832257e-01
-1.19212496e+00 8.00357282e-01 1.50167847e+00 -2.32793748e-01
-5.93798280e-01 1.02524973e-01 -9.70246553e-01 5.65116584e-01
6.35268807e-01 -2.58210063e-01 -7.38987386e-01 -3.62246007e-01
2.63590962e-01 4.62950200e-01 5.09222969e-02 -3.21481526e-01
6.87878013e-01 -3.12779993e-01 -2.64323764e-02 -6.88208342e-01
4.28468525e-01 -2.29633763e-01 -4.73455697e-01 2.54941676e-02
-6.90451026e-01 -1.55818030e-01 2.35713765e-01 5.25356948e-01
-5.33278465e-01 -6.67706847e-01 7.58812800e-02 -4.29271162e-01
-8.32421064e-01 -4.00579199e-02 -6.26506925e-01 5.43657780e-01
4.28455710e-01 4.92180064e-02 -6.80278122e-01 -5.59803307e-01
-4.30467874e-01 6.76470816e-01 -1.00839056e-01 6.35922074e-01
7.40447998e-01 -1.35544860e+00 -8.25545967e-01 -4.08861578e-01
2.39846095e-01 -1.82131771e-02 2.83720344e-01 3.83992702e-01
-1.57886103e-01 1.23319530e+00 1.80467352e-01 -1.85500547e-01
-1.27802956e+00 -1.40996100e-02 2.35336810e-01 -7.71626890e-01
-3.51985663e-01 7.83055007e-01 -8.07205886e-02 -6.05051577e-01
2.95829505e-01 -1.54412627e-01 -7.53854394e-01 4.82329845e-01
8.33678186e-01 5.20363331e-01 3.77442501e-02 -3.10133010e-01
-6.18163049e-02 4.32334632e-01 -5.32959521e-01 -4.26291436e-01
8.88156950e-01 -5.55315793e-01 -3.17942828e-01 3.36661398e-01
9.50436175e-01 1.09884106e-01 -7.84377337e-01 -8.74626935e-01
3.75810593e-01 -2.67395586e-01 4.72538732e-02 -9.84193623e-01
-3.29340100e-01 4.41241026e-01 -1.78897738e-01 5.10090590e-01
8.57509017e-01 1.26299232e-01 9.21904445e-01 1.12935913e+00
5.62044203e-01 -1.42907465e+00 3.93879831e-01 1.05875111e+00
1.21190178e+00 -9.03047204e-01 -9.57165584e-02 -1.80214867e-01
-7.43963003e-01 1.30336249e+00 6.48779631e-01 7.86091924e-01
1.90242663e-01 -1.08581349e-01 -3.34655121e-02 -3.30122799e-01
-8.97335410e-01 -4.32647318e-01 4.45976883e-01 2.95247018e-01
3.19511801e-01 -1.31229281e-01 -5.19303083e-01 7.39699364e-01
-2.35161334e-01 -7.32166544e-02 3.72426689e-01 1.19181538e+00
-9.19446349e-01 -1.20244110e+00 -3.86351883e-01 4.87775475e-01
-1.18221462e-01 -3.15735579e-01 -6.59624338e-01 5.21830320e-01
-4.06280190e-01 1.43153179e+00 -3.82532090e-01 -2.68842041e-01
4.85614657e-01 3.16100806e-01 3.14247280e-01 -8.09535205e-01
-8.07908118e-01 1.13065885e-02 7.76211441e-01 -7.73804486e-02
-5.25118828e-01 -4.43773985e-01 -1.26898062e+00 -2.15786949e-01
-6.61794841e-01 1.04584169e+00 2.76552498e-01 1.03010976e+00
4.55771536e-01 3.63377005e-01 5.49571574e-01 -4.64153022e-01
-7.26822376e-01 -1.27508140e+00 -5.77473223e-01 -2.69758096e-03
2.30185091e-01 -4.50007021e-01 -2.15500861e-01 -1.99696958e-01] | [12.249053001403809, 8.037956237792969] |
684a7882-5d7f-4020-89dc-cb075cff90d4 | fireball-characteristics-derivable-from | 2102.06574 | null | https://arxiv.org/abs/2102.06574v2 | https://arxiv.org/pdf/2102.06574v2.pdf | Fireball characteristics derivable from acoustic data | Near field acoustical signals from fireballs (ranges<200 km), when detected by dense ground networks, may be used to estimate the orientation of the trajectory of a fireball (Pujol et al., 2005) as well as fragmentation locations (Kalenda et al., 2014; Edwards and Hildebrand, 2004). Distinguishing ballistic arrivals (from the cylindrical shock of the fireball)from fragmentation generated signals (quasi-spherical sources) remains a challenge, but are obtainable through analysis of the acoustic path and the timing observed at ground instruments. Here we describe an integrated computer code, termed the Bolide Acoustic Modelling program or BAM, to estimate fireball trajectories and energetics. We develop a new methodology for measuring energy release from bolide fragmentation episodes solely from acoustic measurements and incorporate this into BAM. We also explore the sensitivity of seismo-acoustic fireball solutions and energy estimates to uncertainty in the underlying atmospheric model. Applying BAM to the Stubenberg meteorite producing fireball, we find the total fireball energy from ballistic arrivals to be approximately $5 \times 10^{10}$J which compares favorably to the optical estimate of $4.36 \times 10^{10}$J. The combined fragmentation energy of the Stubenberg event from acoustic data was found to be $1.47^{+0.28}_{-0.12} \times 10^{10}$J, roughly one third of the ballistic or optical total energy. We also show that measuring fireball velocities from acoustic data alone is very challenging but may be possible for slow, deeply penetrating fireballs with shallow entry angles occurring over dense seismic/infrasound networks. | ['Pavel Spurný', 'Denis Vida', 'Peter Brown', 'Luke McFadden'] | 2021-02-12 | null | null | null | null | ['acoustic-modelling'] | ['speech'] | [-3.04846130e-02 -1.89498991e-01 5.55722117e-01 3.47550541e-01
-1.19481599e+00 -7.94656575e-01 1.44887149e-01 6.81123361e-02
-5.74244916e-01 6.90392852e-01 -2.27617677e-02 -3.14451575e-01
-3.77994120e-01 -9.57171500e-01 -6.61156178e-01 -1.00733256e+00
-4.11876112e-01 5.58185935e-01 4.55897689e-01 -1.81740358e-01
1.38342157e-01 7.23261416e-01 -1.14515209e+00 -6.24581240e-02
2.86283523e-01 9.83275115e-01 1.79636344e-01 1.01041400e+00
3.24306548e-01 4.68915015e-01 -6.71482921e-01 7.85111487e-02
2.40940042e-02 -4.93139744e-01 -6.22295022e-01 -7.41221786e-01
4.65059914e-02 -4.46036339e-01 -1.63292348e-01 5.06528139e-01
5.89437783e-01 4.52519894e-01 9.81908739e-01 -7.29141653e-01
3.72495294e-01 4.77944791e-01 -5.38262606e-01 7.23193347e-01
2.81091213e-01 -3.42203714e-02 6.97276354e-01 -7.74490833e-01
1.88805833e-01 4.66630727e-01 8.06195199e-01 -7.99158867e-03
-8.66088808e-01 -9.88861859e-01 -7.10165679e-01 1.30492091e-01
-1.55440676e+00 -7.23130465e-01 5.65768540e-01 -7.82975972e-01
1.37356102e+00 3.59934241e-01 6.34352386e-01 7.27824330e-01
-2.09018700e-02 -1.02559499e-01 7.23378003e-01 -7.29738891e-01
2.73926318e-01 -1.58217132e-01 -1.38820171e-01 4.74100947e-01
3.88790488e-01 4.16641921e-01 -6.10499680e-01 -4.29084480e-01
5.63059807e-01 -6.04435623e-01 -4.97916073e-01 7.46630847e-01
-8.98433328e-01 7.17615545e-01 1.07666455e-01 -2.72748666e-03
-5.34573853e-01 4.94036466e-01 6.67004362e-02 -9.44893509e-02
5.40825009e-01 4.05111760e-01 -2.03764603e-01 -7.32223809e-01
-1.26762950e+00 7.66763508e-01 7.75255442e-01 4.46016669e-01
6.19774699e-01 4.09598798e-01 6.37453079e-01 5.56866646e-01
5.54478645e-01 1.17427981e+00 -1.48382470e-01 -8.45608771e-01
4.84018058e-01 -4.15478647e-01 1.79706752e-01 -9.28307891e-01
-5.66790998e-01 -2.22744912e-01 -1.23762466e-01 1.55745894e-01
7.73577929e-01 -5.27088404e-01 -4.33500916e-01 1.47819948e+00
3.38460326e-01 2.47507304e-01 5.03027774e-02 7.79255688e-01
6.82447970e-01 1.11644936e+00 -1.48837656e-01 -4.55837011e-01
1.44421124e+00 3.40763777e-02 2.96213105e-02 -5.79432249e-01
2.40344465e-01 -9.82533991e-01 3.20587128e-01 2.26855874e-01
-1.23830700e+00 9.18567628e-02 -8.13936710e-01 7.79848099e-01
3.22119325e-01 -4.13658559e-01 1.14349484e-01 6.09548807e-01
-6.71125174e-01 3.22684586e-01 -1.28556967e+00 -1.53097883e-01
-2.69155741e-01 -1.79800428e-02 1.67941481e-01 3.59511763e-01
-1.07621527e+00 4.28802103e-01 9.66961030e-03 2.10392952e-01
-1.04172397e+00 -9.00666058e-01 -5.85371494e-01 -6.99325697e-03
2.81701803e-01 -2.31926814e-01 1.40741897e+00 3.54411379e-02
-1.02148819e+00 2.07218677e-01 1.07383966e-01 -3.33836019e-01
-1.03904568e-01 6.02148660e-02 -4.09793794e-01 8.06417704e-01
3.23920488e-01 -1.01022407e-01 2.92233407e-01 -1.17229176e+00
-5.16756713e-01 -1.57701269e-01 -1.69483364e-01 -4.73068431e-02
9.40893143e-02 5.50998807e-01 4.36945707e-01 -6.13132954e-01
4.27442282e-01 -9.51470137e-01 6.29201857e-03 -5.37242830e-01
-3.00383747e-01 1.69439510e-01 4.44737226e-01 -7.96855390e-01
8.99529636e-01 -1.78980291e+00 -2.53415912e-01 5.29125512e-01
-1.72969531e-02 -2.38085613e-01 4.75394845e-01 9.87442493e-01
2.08553419e-01 2.48007879e-01 -6.50323212e-01 1.58960685e-01
-2.77918547e-01 -2.04450637e-01 -3.77781957e-01 7.15036154e-01
-2.03855366e-01 3.90743762e-02 -6.82677746e-01 -1.16747342e-01
-1.00121170e-01 3.08556050e-01 -4.11338001e-01 8.54767188e-02
2.67847687e-01 1.58650160e-01 -3.70489955e-01 8.21511865e-01
1.10445988e+00 6.84611738e-01 -3.43609631e-01 -1.27539292e-01
-8.42179000e-01 4.30163056e-01 -1.24466038e+00 9.71502542e-01
-3.94381315e-01 7.33277023e-01 8.86308670e-01 -8.43759120e-01
9.31719303e-01 6.36746705e-01 4.58317190e-01 -4.34900671e-01
-8.17977637e-02 5.47617674e-01 2.86986269e-02 -8.94895196e-01
5.36341429e-01 -1.16522515e+00 -1.37603879e-01 7.96793938e-01
-1.83055289e-02 -4.26223963e-01 -4.16864529e-02 2.01485842e-01
1.27736342e+00 -7.90736377e-02 -3.01795602e-01 -2.97487974e-01
-1.59215942e-01 1.38417482e-01 5.19704282e-01 6.85111225e-01
2.71294359e-02 7.31457293e-01 3.19361985e-01 -2.53237505e-02
-9.03117120e-01 -1.31532907e+00 -4.48115557e-01 7.32394218e-01
7.74198920e-02 -3.77560228e-01 -5.30790865e-01 1.58272251e-01
-1.45180732e-01 6.64848864e-01 -1.00379899e-01 -3.54853272e-01
-6.90540671e-01 -1.32499015e+00 1.20597112e+00 7.15998590e-01
1.75076313e-02 -1.09380198e+00 -9.29386616e-01 4.50838447e-01
-7.20234573e-01 -1.04680610e+00 1.13885626e-01 4.08742040e-01
-5.50946414e-01 -6.19073868e-01 -3.54808569e-01 -2.19972625e-01
1.33298621e-01 8.07416216e-02 9.03367519e-01 -4.76595387e-02
-1.98875383e-01 5.34715414e-01 -6.76894426e-01 -7.79064775e-01
-3.81143868e-01 -4.60578144e-01 5.29774129e-01 -3.45511705e-01
1.92000911e-01 -8.02202523e-01 -7.14446306e-01 5.33356607e-01
-6.70652986e-01 -5.38718522e-01 -3.23308222e-02 2.31135845e-01
4.08820599e-01 2.86441982e-01 5.63016593e-01 1.38652802e-01
1.88801497e-01 -7.85872757e-01 -5.00177801e-01 -5.00247836e-01
1.87603035e-03 -4.72503215e-01 4.62706745e-01 -5.18129356e-02
-1.12035728e+00 -6.58002913e-01 -5.64062715e-01 4.28431891e-02
-4.49857533e-01 7.91394949e-01 2.98684418e-01 -2.15030834e-01
8.47384632e-01 1.70885131e-01 -2.91556209e-01 -5.30296147e-01
-3.50264430e-01 7.84222722e-01 7.60426939e-01 -9.42955196e-01
8.59786153e-01 8.09464693e-01 -8.62188339e-02 -1.51541615e+00
-3.01696330e-01 -6.00370049e-01 -1.14802152e-01 -5.68604112e-01
7.65285552e-01 -7.75443852e-01 -8.03922534e-01 3.59696776e-01
-1.02332819e+00 -3.44390988e-01 -9.34360325e-02 1.25548458e+00
-3.76270235e-01 2.57676274e-01 -6.03583813e-01 -1.36090624e+00
-3.66764098e-01 -8.68602693e-01 8.23765278e-01 9.92079675e-02
-4.13561374e-01 -7.39433229e-01 6.24290824e-01 4.62529123e-01
2.76361495e-01 7.17736244e-01 5.46156943e-01 -1.15289070e-01
-3.21751118e-01 -4.46616530e-01 1.43860057e-01 7.79982507e-02
-2.03427315e-01 4.09935057e-01 -1.04118943e+00 -1.84216797e-01
4.37618703e-01 -1.81830406e-01 1.01621354e+00 7.86247611e-01
7.13331252e-02 -1.14406804e-02 -4.36868131e-01 7.29355156e-01
1.52521849e+00 5.86181343e-01 2.66273677e-01 4.53960299e-01
3.07284772e-01 5.35863876e-01 4.30555940e-01 7.57769942e-01
2.50901934e-02 4.95661616e-01 4.36845571e-01 3.77113938e-01
1.64980784e-01 2.84541935e-01 4.80656236e-01 8.76827717e-01
-9.63875294e-01 -3.52178931e-01 -1.26556182e+00 7.23695576e-01
-9.78595912e-01 -1.42781150e+00 -6.36761487e-01 2.42866778e+00
4.86908913e-01 1.72754347e-01 8.02497640e-02 2.40796804e-01
4.66506183e-01 -8.75578225e-02 2.06879780e-01 -7.08283424e-01
2.32329026e-01 6.55795097e-01 7.53626764e-01 6.91765010e-01
-3.86696815e-01 2.30082989e-01 5.72353649e+00 3.86436403e-01
-1.00157166e+00 1.92309737e-01 -1.80245608e-01 -4.03701037e-01
-5.65191865e-01 2.26168662e-01 -7.07045138e-01 4.24798727e-01
1.33428752e+00 -7.05710575e-02 2.83502996e-01 5.50487339e-02
5.13120294e-01 -8.24097037e-01 -4.70196396e-01 4.91385430e-01
-1.96257755e-01 -1.12437165e+00 -6.67314887e-01 1.90098852e-01
1.97124377e-01 2.73529589e-01 -5.46868682e-01 -4.96463552e-02
-7.41930231e-02 -5.98077893e-01 1.34333503e+00 4.90357667e-01
8.62738848e-01 -1.15137780e+00 5.13969243e-01 6.08648777e-01
-1.52012444e+00 2.08231509e-01 -4.08628494e-01 -4.04261410e-01
9.05811489e-01 8.10925841e-01 -1.02180171e+00 7.53058553e-01
1.22607207e+00 -6.04811248e-05 1.65007129e-01 9.47204947e-01
8.74125883e-02 1.40641689e+00 -1.20041370e+00 8.99428800e-02
4.10359174e-01 -3.95035982e-01 1.00575125e+00 1.27502859e+00
9.03071821e-01 5.16330242e-01 -3.50531757e-01 8.42122436e-01
2.24487126e-01 -1.95725456e-01 -4.74719644e-01 1.90101415e-02
5.37123442e-01 1.15722048e+00 -1.03108907e+00 1.55807301e-01
-1.42348483e-01 1.41680911e-01 -4.38039750e-01 4.19648886e-01
-1.03646398e+00 -7.71413386e-01 5.57144761e-01 5.34831822e-01
1.89466983e-01 -4.79768068e-01 -1.19566685e-02 -5.31212032e-01
2.07043275e-01 -7.79380500e-02 1.45902470e-01 -7.34610736e-01
-6.71283185e-01 5.42847514e-01 5.44946253e-01 -1.09206760e+00
-3.31184536e-01 -3.02679211e-01 -1.25740838e+00 8.55277240e-01
-6.69453681e-01 -4.71281618e-01 -3.07351053e-01 1.42847925e-01
-9.15863588e-02 1.89940929e-01 7.26046324e-01 2.26247087e-01
-3.57991308e-01 1.35697678e-01 5.45437932e-01 -2.10837577e-03
3.53922576e-01 -9.22742844e-01 2.87276268e-01 7.70235598e-01
8.93491432e-02 2.28762522e-01 1.26036787e+00 -7.89703786e-01
-1.35896325e+00 -6.08595908e-01 6.14863575e-01 -3.61309111e-01
8.64302576e-01 -3.40701640e-01 -8.71063232e-01 4.68097508e-01
1.83984265e-01 -1.38636500e-01 7.72675693e-01 -2.53501773e-01
2.12881267e-01 -1.72908261e-01 -1.17556834e+00 7.72805214e-02
6.45629823e-01 -4.08502311e-01 -6.22995377e-01 2.43977591e-01
3.48443091e-01 -1.48809224e-01 -8.33974779e-01 6.22088432e-01
8.30299735e-01 -1.02325022e+00 9.66069877e-01 -1.39202550e-01
8.84141475e-02 -1.39264762e-01 -3.34729612e-01 -1.06789923e+00
-6.02021106e-02 -5.44822335e-01 4.02577996e-01 1.48632216e+00
5.25493622e-01 -3.78535330e-01 6.58549428e-01 1.26442373e-01
-4.08560932e-01 -3.45294535e-01 -1.44934690e+00 -7.94217885e-01
3.00510615e-01 -1.13214922e+00 -5.58560230e-02 4.74400401e-01
1.95762828e-01 -2.41665766e-01 1.44849628e-01 7.52140582e-01
8.61767173e-01 -2.20746234e-01 6.86534941e-02 -9.42478478e-01
-6.02433503e-01 -9.01791267e-03 -2.09491998e-01 -3.03548634e-01
-4.71604496e-01 -7.02094913e-01 4.65220362e-01 -1.36110616e+00
-1.38764724e-01 -6.93589628e-01 -5.96038029e-02 1.49444893e-01
5.00537276e-01 4.08107549e-01 -9.97837335e-02 3.10111716e-02
4.88348633e-01 2.54076362e-01 4.06621724e-01 1.79751426e-01
-1.89197093e-01 1.92247495e-01 -6.26502838e-03 1.03899324e+00
6.98589444e-01 -1.03068507e+00 -8.60375315e-02 -4.71585244e-01
7.96113729e-01 5.13075471e-01 6.34059906e-01 -1.25820076e+00
1.90328076e-01 -2.94889867e-01 -1.31347805e-01 -6.01996958e-01
3.70321810e-01 -2.74820805e-01 5.41851521e-01 3.31290871e-01
3.35011750e-01 -1.26533121e-01 4.34111059e-01 4.28245038e-01
-2.16959715e-01 -9.67265904e-01 5.85878134e-01 -2.22327292e-01
8.00564736e-02 -1.32201865e-01 -1.05140984e+00 2.33905479e-01
7.96330392e-01 -1.09279208e-01 -1.96821019e-01 -4.19126242e-01
-6.45087957e-01 -2.97055304e-01 1.78362668e-01 -3.80529583e-01
5.01821160e-01 -8.69743347e-01 -9.42536950e-01 3.19006443e-02
-3.63262385e-01 1.59020647e-01 4.71176893e-01 1.09075475e+00
-1.29430366e+00 7.33317137e-02 2.04392001e-01 -5.42215943e-01
-9.07531977e-01 -1.22998551e-01 5.52667499e-01 3.99003118e-01
-4.66543943e-01 1.35088599e+00 -5.53039163e-02 1.01617046e-01
-3.47691238e-01 -1.94361493e-01 2.00087234e-01 4.30989087e-01
4.81865436e-01 1.05071020e+00 4.01247472e-01 -6.67571902e-01
-6.61921799e-01 8.12855601e-01 3.05442393e-01 -5.65963566e-01
1.30538213e+00 -4.73725051e-02 -1.76527023e-01 4.36833322e-01
1.08032179e+00 3.85366261e-01 -9.55383062e-01 2.36800686e-01
-4.36991662e-01 -1.68634936e-01 3.14852059e-01 -4.23258513e-01
-8.10297728e-01 8.21795881e-01 1.23427451e-01 5.57488978e-01
8.71203065e-01 4.76253182e-01 8.72416854e-01 1.28953740e-01
4.05618608e-01 -1.04754794e+00 -3.09840053e-01 3.02586824e-01
8.02212358e-01 -5.06401241e-01 -1.86843872e-02 -1.64936230e-01
4.13729809e-03 1.43761802e+00 2.22393394e-01 -1.67051032e-01
9.26734746e-01 7.94645667e-01 -3.45723219e-02 -1.63632601e-01
-4.51326549e-01 2.16668874e-01 -3.11962724e-01 3.14735919e-01
1.41170427e-01 4.93614599e-02 -2.70017087e-01 7.77678192e-01
-4.96604234e-01 -3.05021703e-01 1.06023657e+00 1.26574719e+00
-8.42439651e-01 -5.21776617e-01 -1.23975062e+00 3.12657475e-01
-5.90760589e-01 -1.13047056e-01 7.66313449e-02 5.82341552e-01
1.36679247e-01 1.16845858e+00 4.05092269e-01 -2.50940889e-01
3.95055294e-01 8.39179195e-03 3.88910592e-01 -4.48260844e-01
-3.40976268e-01 5.65498710e-01 4.94751066e-01 -7.58198742e-03
-5.05465090e-01 -9.78944004e-01 -1.74422348e+00 -3.68705064e-01
-6.16362453e-01 6.18080974e-01 6.61807775e-01 1.00561333e+00
-4.69247162e-01 2.75732011e-01 6.30303442e-01 -1.24443448e+00
-1.09212488e-01 -9.87797856e-01 -1.16796613e+00 -4.77550179e-01
3.73621851e-01 -7.52182305e-01 -1.22293031e+00 7.85028040e-02] | [6.809706211090088, 2.8264431953430176] |
c40abd5e-bebb-4e9e-b0ba-85926bd64f4c | query-understanding-in-the-age-of-large | 2306.16004 | null | https://arxiv.org/abs/2306.16004v1 | https://arxiv.org/pdf/2306.16004v1.pdf | Query Understanding in the Age of Large Language Models | Querying, conversing, and controlling search and information-seeking interfaces using natural language are fast becoming ubiquitous with the rise and adoption of large-language models (LLM). In this position paper, we describe a generic framework for interactive query-rewriting using LLMs. Our proposal aims to unfold new opportunities for improved and transparent intent understanding while building high-performance retrieval systems using LLMs. A key aspect of our framework is the ability of the rewriter to fully specify the machine intent by the search engine in natural language that can be further refined, controlled, and edited before the final retrieval phase. The ability to present, interact, and reason over the underlying machine intent in natural language has profound implications on transparency, ranking performance, and a departure from the traditional way in which supervised signals were collected for understanding intents. We detail the concept, backed by initial experiments, along with open questions for this interactive query understanding framework. | ['Vinay Setty', 'Abhijit Anand', 'Venktesh V', 'Avishek Anand'] | 2023-06-28 | null | null | null | null | ['retrieval'] | ['methodology'] | [ 3.48595411e-01 6.08577430e-01 -4.07698154e-01 -6.26170278e-01
-7.84586906e-01 -9.47133482e-01 1.19501936e+00 2.95794189e-01
-5.27448535e-01 1.32781044e-01 7.01275766e-01 -6.15507901e-01
-1.53023735e-01 -4.76259768e-01 -1.45711243e-01 3.49110812e-01
1.12482823e-01 5.43431699e-01 1.01890169e-01 -6.36234105e-01
4.50367212e-01 3.49627972e-01 -1.46503818e+00 7.03085899e-01
7.13896632e-01 6.69502199e-01 3.41882259e-01 8.95452917e-01
-5.31030536e-01 1.04314125e+00 -4.89338726e-01 -2.91336924e-01
-1.18172400e-01 3.96392681e-03 -1.31466603e+00 -4.96743590e-01
3.83223593e-01 -4.26902860e-01 -1.15184737e-02 6.98287785e-01
8.72785076e-02 1.80346798e-02 5.11768401e-01 -1.11627746e+00
-8.75180781e-01 4.36033458e-01 2.59107351e-01 -2.65097201e-01
1.11130583e+00 1.20724231e-01 1.40627110e+00 -7.01963365e-01
9.63474691e-01 1.52730465e+00 1.46470424e-02 6.06638968e-01
-1.29940057e+00 -1.93114042e-01 3.77165079e-01 -2.73714185e-01
-1.15389693e+00 -9.05195355e-01 3.22148800e-01 -4.19587612e-01
1.17112410e+00 9.63376820e-01 3.77206922e-01 7.31768489e-01
1.31988540e-01 8.68260920e-01 7.47665107e-01 -8.61890733e-01
-2.92362310e-02 7.84031153e-01 4.43085462e-01 8.40858459e-01
-4.95518278e-03 1.09438285e-01 -7.35938311e-01 -4.24980611e-01
5.99641860e-01 -1.27945274e-01 -1.57252580e-01 -4.78880614e-01
-1.25284338e+00 7.18251884e-01 2.92695522e-01 2.77902901e-01
-1.85346097e-01 8.29812139e-02 2.82120615e-01 7.74495542e-01
2.44964808e-01 1.22249365e+00 -3.59488189e-01 -1.85917914e-01
-7.81812787e-01 3.83778423e-01 1.36812866e+00 1.15563381e+00
6.72035277e-01 -5.10467112e-01 -5.60813010e-01 7.62037098e-01
6.78360403e-01 5.39446235e-01 2.56680429e-01 -1.30321455e+00
1.96060359e-01 8.44876230e-01 4.78975952e-01 -8.72053087e-01
-2.25847960e-01 -9.21830535e-02 -9.08727050e-02 -9.86369625e-02
6.76420107e-02 2.03627437e-01 -5.29324234e-01 1.66953683e+00
-1.34667277e-01 -1.03325260e+00 2.64008909e-01 6.44624770e-01
5.64678669e-01 5.51336706e-01 3.91477615e-01 -1.77530959e-01
1.46527541e+00 -5.23898125e-01 -7.04268038e-01 -3.30517262e-01
1.07712722e+00 -7.49812007e-01 1.54758537e+00 2.02280447e-01
-1.00018513e+00 -3.56250733e-01 -7.86675155e-01 -7.37482131e-01
-4.89354789e-01 -1.14495018e-02 9.62840617e-01 3.41029346e-01
-1.35433984e+00 -4.41312604e-02 -6.20766997e-01 -6.49312556e-01
-1.87056661e-01 2.51927525e-01 -4.29209694e-02 1.68140575e-01
-1.27964365e+00 9.73103940e-01 3.99960503e-02 -2.67811507e-01
-3.36788327e-01 -5.90150654e-01 -7.45219469e-01 2.86826827e-02
3.67628187e-01 -8.92332256e-01 1.88796985e+00 -7.36566901e-01
-1.36737704e+00 1.07602656e+00 -5.44722915e-01 -2.80305564e-01
2.82217979e-01 -7.00043857e-01 -4.27073985e-01 -1.57909542e-01
2.96460450e-01 8.92617464e-01 3.66807729e-01 -1.23434770e+00
-5.68886042e-01 -4.84995246e-01 5.32967687e-01 4.76259828e-01
-2.03446940e-01 2.41671383e-01 -7.93780088e-01 -1.87532634e-01
-4.90668379e-02 -8.19503188e-01 -1.72724262e-01 2.32151330e-01
-3.24788153e-01 -4.09264028e-01 6.42984509e-01 -3.24978977e-01
1.77179503e+00 -2.10116458e+00 -5.74268848e-02 2.30339855e-01
3.47107828e-01 1.70899779e-01 -9.01522562e-02 1.15492296e+00
1.68469831e-01 5.75723290e-01 3.37583691e-01 -1.92954853e-01
4.35099989e-01 -1.41799627e-02 -9.03097868e-01 -2.03055337e-01
-1.05609469e-01 9.92397964e-01 -7.51546144e-01 -6.05047524e-01
1.52029112e-01 1.23986945e-01 -8.41052413e-01 5.17427742e-01
-7.05855012e-01 4.34488952e-02 -8.27266276e-01 5.38817108e-01
-1.96245521e-01 -3.81462514e-01 1.67880788e-01 -1.19907990e-01
-5.60392141e-01 7.58011162e-01 -8.68167639e-01 1.79006100e+00
-8.73458922e-01 8.62756193e-01 3.20085853e-01 -1.35794654e-01
5.92902601e-01 3.23338419e-01 -6.90319669e-03 -1.00451076e+00
-4.82840478e-01 1.55275196e-01 -3.48912507e-01 -7.93418407e-01
8.42581093e-01 2.43513823e-01 -4.13865566e-01 9.13139641e-01
-4.19521779e-01 -5.05810022e-01 1.78760514e-01 6.56328976e-01
7.38151491e-01 -5.56445867e-02 3.27368319e-01 -3.30878556e-01
4.02257472e-01 2.71567762e-01 -3.47874731e-01 1.38605893e+00
1.94705844e-01 -1.00852540e-02 2.95073032e-01 -5.51181018e-01
-6.15527153e-01 -7.77314723e-01 4.13850546e-02 1.65807068e+00
1.80760652e-01 -8.09387684e-01 -4.85687226e-01 -1.07742064e-01
-1.59251630e-01 1.26968038e+00 -6.11815639e-02 -1.90341786e-01
-3.67690504e-01 1.68712586e-01 5.25382578e-01 -6.75899833e-02
3.01422328e-01 -1.07296658e+00 -1.03694236e+00 5.17257340e-02
-2.80410230e-01 -9.04467344e-01 -6.17023051e-01 -6.79632872e-02
-8.99455845e-01 -7.48907685e-01 -8.66877288e-02 -5.50274849e-01
3.93188685e-01 2.14364663e-01 1.41933835e+00 3.90106201e-01
-1.62058577e-01 9.43546593e-01 -1.83863014e-01 -5.34137785e-01
-6.31758749e-01 3.26751113e-01 -2.05057099e-01 -3.84846300e-01
4.13337141e-01 -2.10887536e-01 -6.42063379e-01 9.28951651e-02
-1.16575527e+00 5.01829982e-01 5.98087013e-01 2.93303341e-01
6.73424527e-02 -6.63711131e-01 1.55552983e-01 -1.34240890e+00
1.30431402e+00 -1.78839788e-01 -5.22653997e-01 7.77664185e-01
-1.15631187e+00 7.32956111e-01 1.27898395e-01 -5.13334461e-02
-1.31246710e+00 -1.72266334e-01 1.39654294e-01 3.29508394e-01
-1.79599524e-01 6.98670924e-01 3.39368880e-02 8.98963138e-02
9.25446391e-01 9.80653707e-03 -3.83809535e-03 -5.12883306e-01
1.10025060e+00 8.41498673e-01 3.49385917e-01 -6.82724714e-01
5.76622605e-01 3.12610716e-01 -5.41518211e-01 -9.40383673e-01
-5.07206619e-01 -7.18842208e-01 -2.47285053e-01 -1.18345693e-02
5.71145833e-01 -7.72561252e-01 -8.45475852e-01 -1.24713302e-01
-1.30740237e+00 -2.94449151e-01 -3.12780976e-01 5.15908673e-02
-5.20501316e-01 2.37496391e-01 -3.88064772e-01 -1.07430971e+00
-4.98104572e-01 -1.02618110e+00 1.42857456e+00 7.03622997e-02
-1.16427982e+00 -1.01229215e+00 1.95203543e-01 4.75835145e-01
6.96950138e-01 -5.91848433e-01 1.42329729e+00 -8.91045272e-01
-8.69541228e-01 -3.99220437e-01 -2.75198907e-01 -3.76085162e-01
6.04521073e-02 -3.35607887e-03 -9.90257025e-01 8.26080367e-02
4.90014590e-02 -4.90323454e-01 2.93426484e-01 -1.86203718e-01
7.73661494e-01 -6.43811464e-01 -6.18821919e-01 3.26696843e-01
1.24255359e+00 3.59673858e-01 3.54768425e-01 2.00665742e-01
1.77417472e-01 8.20672214e-01 5.04069388e-01 1.02878988e-01
5.07609367e-01 7.77762890e-01 -4.02481528e-03 -6.17068782e-02
1.10264391e-01 -6.57667339e-01 1.93072140e-01 4.52330172e-01
4.43741828e-01 -3.25361401e-01 -9.39836621e-01 1.73411608e-01
-1.70887756e+00 -7.50073195e-01 3.62681597e-01 2.12994313e+00
9.13466454e-01 1.66762564e-02 -4.71820712e-01 -5.81731200e-01
4.36145589e-02 3.86095375e-01 -5.39192438e-01 -8.69558394e-01
2.95496017e-01 4.44028415e-02 1.24739006e-01 1.27631211e+00
-5.35344064e-01 1.12089479e+00 7.41173410e+00 2.21734390e-01
-9.36224580e-01 -3.33135992e-01 2.96310782e-01 9.80483592e-02
-9.68046963e-01 3.17184478e-01 -5.97436428e-01 -3.36742282e-01
8.33120048e-01 -3.31125617e-01 6.87692821e-01 7.47802734e-01
3.60094726e-01 -1.93819940e-01 -1.71092772e+00 8.76741230e-01
-4.45845090e-02 -1.38085723e+00 5.09542763e-01 -2.07218006e-01
-3.34330066e-03 -7.06687868e-02 -1.53815104e-02 4.93891537e-01
4.94751304e-01 -8.89638126e-01 5.24331987e-01 8.82763326e-01
9.31967974e-01 4.71336320e-02 -2.86698639e-02 7.31649339e-01
-8.35368633e-01 -1.40424207e-01 2.23791331e-01 -2.81437188e-01
-4.61925194e-02 2.51829922e-02 -8.01897168e-01 1.17903456e-01
2.92742699e-01 1.08274229e-01 -6.24117553e-01 4.18711364e-01
1.06783137e-01 8.47570896e-02 -3.20028216e-01 -4.83256817e-01
-3.41653004e-02 -6.44015148e-02 7.44977415e-01 1.31861830e+00
-1.87087610e-01 1.47712037e-01 2.39523426e-01 1.01119065e+00
7.77660534e-02 2.18418807e-01 -1.05039096e+00 -5.24942577e-01
4.84513581e-01 9.66059744e-01 -1.13774464e-01 -5.02869844e-01
-4.15678561e-01 9.54737127e-01 1.61280304e-01 7.72863686e-01
-1.25768110e-01 -3.45806420e-01 5.77293992e-01 2.57705718e-01
-6.88639402e-01 -2.06574365e-01 -1.89824998e-01 -1.31979942e+00
9.60495248e-02 -1.38086259e+00 4.60229784e-01 -1.16595018e+00
-7.43731320e-01 4.89320785e-01 5.26700079e-01 -5.54753661e-01
-8.43588233e-01 -3.38254184e-01 -4.70315963e-01 1.06696010e+00
-1.07368779e+00 -9.17724669e-01 3.87809193e-03 2.10753828e-01
8.48613739e-01 1.68512866e-01 1.28600490e+00 -4.15309444e-02
9.35732126e-02 1.75150946e-01 -2.94967204e-01 -1.93216950e-01
8.20965886e-01 -1.13375485e+00 5.37878335e-01 4.10952926e-01
5.39467037e-01 1.54427254e+00 7.64057457e-01 -3.81137699e-01
-1.78466344e+00 -4.11906213e-01 1.37832117e+00 -1.02828920e+00
4.97512013e-01 -5.85036755e-01 -7.33060896e-01 8.15122545e-01
4.69312996e-01 -9.20316339e-01 7.34004259e-01 5.73360205e-01
-4.66963679e-01 -1.01740390e-01 -6.58536613e-01 1.11207402e+00
9.63854849e-01 -1.30199218e+00 -8.27484548e-01 4.59210545e-01
1.17067683e+00 -2.56540626e-01 -3.74264985e-01 2.69478709e-01
9.08194363e-01 -8.21222007e-01 8.59189034e-01 -7.07888842e-01
-1.58003375e-01 -4.76767607e-02 -1.97995022e-01 -7.11244941e-01
-3.05275060e-02 -1.02278769e+00 1.37865275e-01 8.51019084e-01
7.89599419e-01 -6.55691326e-01 4.60258782e-01 1.65018070e+00
8.85549113e-02 -4.51746136e-01 -3.87363851e-01 2.80547654e-03
-2.63069659e-01 -6.89317584e-01 3.71462405e-01 3.69599134e-01
6.30063117e-01 6.92947567e-01 8.91414806e-02 5.91960689e-03
9.43674296e-02 3.14135075e-01 8.69524360e-01 -1.28086412e+00
-2.90588915e-01 -5.08835733e-01 2.62965798e-01 -1.76662588e+00
-2.08631113e-01 -8.52219462e-01 -6.54037073e-02 -1.62789559e+00
9.04516280e-02 -3.71110171e-01 2.15891391e-01 2.53147662e-01
1.27505228e-01 -5.38846910e-01 2.56865352e-01 7.66989946e-01
-9.55621123e-01 1.93657860e-01 1.00052381e+00 -9.00253505e-02
-5.75863540e-01 1.74730346e-01 -1.05596685e+00 6.56464934e-01
4.15038317e-01 -1.30705684e-01 -9.68946099e-01 -7.44981587e-01
6.84759021e-01 4.09747094e-01 3.03251624e-01 -5.86513102e-01
7.18891740e-01 -2.59478003e-01 -8.40395540e-02 -3.56663525e-01
3.87591332e-01 -8.92746925e-01 4.35745157e-02 3.80394191e-01
-1.37230086e+00 2.65832156e-01 2.80698277e-02 3.62482190e-01
-1.98110819e-01 -7.96887800e-02 1.02681234e-01 -3.36185157e-01
-7.69203365e-01 -1.21817812e-01 -6.81580544e-01 2.13495880e-01
3.44210476e-01 1.73711237e-02 -1.98694170e-01 -9.76560652e-01
-6.80300355e-01 4.79401767e-01 5.61422467e-01 7.55345106e-01
4.03649032e-01 -7.19401598e-01 2.63047088e-02 3.81850630e-01
5.55698276e-01 -2.54527062e-01 -4.72109318e-01 9.02317911e-02
-3.39740306e-01 1.30412269e+00 4.63622302e-01 -4.46575910e-01
-1.01851654e+00 3.76279861e-01 2.91018248e-01 -3.45682442e-01
-1.63294822e-01 5.89433312e-01 4.51288581e-01 -5.12184083e-01
5.43817341e-01 -5.47559798e-01 -2.36341715e-01 -2.51439631e-01
4.68927979e-01 -2.14483991e-01 -2.70848691e-01 -9.98882949e-02
-2.27313921e-01 2.89844304e-01 -3.43511462e-01 -7.74212480e-01
6.90381169e-01 -4.68883514e-01 -3.85580659e-01 7.92592943e-01
1.10357094e+00 1.73018157e-01 -3.32673460e-01 -3.42198104e-01
4.72730786e-01 -1.32558495e-01 -1.81555480e-01 -1.03210747e+00
-4.00006175e-02 6.55680478e-01 4.57885295e-01 5.64499259e-01
8.13012898e-01 2.51421273e-01 3.22290450e-01 1.32631159e+00
4.00009155e-01 -9.33900893e-01 -6.52599707e-02 4.99024868e-01
1.18068123e+00 -1.19183791e+00 -7.16412961e-02 -2.31311813e-01
-5.78031957e-01 1.04079199e+00 5.43395579e-01 6.38630152e-01
5.54038584e-01 -1.43717732e-02 5.13217747e-01 -7.85959959e-01
-1.26927602e+00 6.50438219e-02 4.11908656e-01 3.56470764e-01
9.74330842e-01 -7.70368055e-02 -1.15767874e-01 3.70067619e-02
-2.25970820e-01 1.27662733e-01 -3.44769992e-02 1.02629316e+00
-5.88907540e-01 -1.35298991e+00 -2.03019474e-02 4.00592655e-01
-3.22358727e-01 -4.56777126e-01 -9.01227236e-01 8.01498830e-01
-5.01694560e-01 1.29238474e+00 -9.17031989e-02 -1.13444798e-01
4.83548254e-01 4.95251685e-01 -1.38977990e-01 -1.06489265e+00
-6.04684114e-01 -3.10729861e-01 5.72913408e-01 -9.97124851e-01
-1.26847371e-01 -2.91903079e-01 -1.09012830e+00 1.89639971e-01
-2.94394463e-01 7.25894690e-01 6.35222375e-01 7.26017237e-01
7.35396862e-01 -2.05963820e-01 3.45136970e-01 -5.21347225e-02
-6.05599940e-01 -7.26944923e-01 1.19502470e-01 3.19909573e-01
6.14893138e-01 3.38986740e-02 -2.90980995e-01 2.46763110e-01] | [11.891278266906738, 7.862659931182861] |
a4fe36b2-08d3-409c-8968-b9e7c1d3fae1 | pretrained-language-models-are-all-you-need | null | null | https://openreview.net/forum?id=OA-sluxOzcY | https://openreview.net/pdf?id=OA-sluxOzcY | Pretrained Language Models Are All You Need For Text-to-SQL Schema Linking | The use of Exact Match based Schema Linking (EMSL) has become standard in text-to-SQL: many state-of-the-art text-to-SQL models employ EMSL, and their performance drops significantly when the EMSL component is removed. In this work, however, we demonstrate that EMSL reduces robustness, rendering models vulnerable to synonym substitution and typos. Instead of relying on EMSL to make up for deficiencies in question-schema encoding, we show that by utilizing the pre-trained language model as the encoder, we can improve the performance without using EMSL, and thus the model is more robust. Our experiments suggest that EMSL is not the icing on the cake, but it is the one that introduces the vulnerability, and it can be replaced by better input encoding. | ['Anonymous'] | 2021-12-17 | null | null | null | acl-arr-december-2022-12 | ['text-to-sql'] | ['computer-code'] | [ 5.59559427e-02 1.83272541e-01 -1.66712910e-01 -2.45617792e-01
-8.52281034e-01 -7.85218000e-01 4.24967051e-01 6.09563410e-01
-4.82104599e-01 3.65229815e-01 5.06411970e-01 -6.07998788e-01
-1.88847049e-03 -9.36742783e-01 -1.12166297e+00 2.50991076e-01
5.48821867e-01 2.43097022e-01 6.62784874e-01 -6.32988751e-01
3.11266124e-01 9.23714489e-02 -1.34682441e+00 6.48325145e-01
9.39500928e-01 4.87007707e-01 -2.81882912e-01 2.66429842e-01
-6.88757300e-01 9.70631897e-01 -8.10251832e-01 -9.06613886e-01
2.76050836e-01 -3.38251233e-01 -9.34046566e-01 -5.13343990e-01
7.40650654e-01 -2.58261532e-01 -3.44445735e-01 8.70507836e-01
4.53743666e-01 -2.76957512e-01 1.90509766e-01 -1.15948629e+00
-6.14478886e-01 9.86905038e-01 -2.86243290e-01 -2.33463287e-01
6.42881513e-01 -1.31754383e-01 1.14176869e+00 -9.82832134e-01
9.24251616e-01 1.15393806e+00 9.85341430e-01 3.07300627e-01
-1.27028489e+00 -6.00072086e-01 -3.06215975e-02 -1.45478593e-02
-1.39227772e+00 -5.46913147e-01 4.38328385e-01 -2.70375073e-01
1.37586808e+00 4.48862702e-01 3.16371739e-01 1.09420240e+00
1.43526629e-01 6.15467489e-01 8.75854194e-01 -6.46678805e-01
-3.08440849e-02 4.29458797e-01 1.44584686e-01 7.22956896e-01
6.15745068e-01 -1.96893811e-02 -5.57569742e-01 -4.68663603e-01
4.59642768e-01 -3.35826278e-01 -1.02491796e-01 -3.19459647e-01
-7.63485670e-01 7.14695036e-01 2.15416551e-01 2.23164529e-01
8.73073936e-03 2.03742146e-01 6.80656135e-01 5.41350126e-01
9.15214866e-02 8.92663538e-01 -3.97907257e-01 -3.01190913e-01
-1.25530601e+00 3.87127846e-01 9.92702544e-01 1.39221776e+00
7.55193949e-01 -1.92210898e-01 1.07678376e-01 8.10065925e-01
8.84686932e-02 2.93002069e-01 4.43357497e-01 -1.02666593e+00
6.77039862e-01 1.10841000e+00 1.49529902e-02 -7.98586190e-01
-1.97907507e-01 -4.47658777e-01 4.94872294e-02 -3.18383090e-02
3.43004525e-01 1.48596436e-01 -6.81896329e-01 1.72657681e+00
-1.30556867e-01 -3.30444992e-01 9.94064063e-02 3.04153681e-01
6.67363226e-01 3.83234203e-01 1.28116742e-01 2.46017784e-01
1.13728714e+00 -6.65055454e-01 -7.34900773e-01 -3.89226258e-01
1.25275886e+00 -9.47315156e-01 1.29493570e+00 8.13801512e-02
-1.17345953e+00 -2.98495293e-01 -1.28068602e+00 -2.77955115e-01
-7.44207561e-01 -2.54861593e-01 4.36642289e-01 6.67130053e-01
-8.92776191e-01 5.56005657e-01 -5.75764000e-01 -9.71570671e-01
1.16104791e-02 -1.03808932e-01 -5.54180026e-01 -1.06526360e-01
-1.46170151e+00 1.15278399e+00 5.83809197e-01 -4.51254606e-01
-1.12855442e-01 -8.07428360e-01 -9.07159984e-01 1.09786496e-01
8.58191550e-01 -7.26027489e-01 1.05951536e+00 -6.43207073e-01
-1.02317834e+00 4.19701815e-01 -2.21764669e-01 -2.43209884e-01
4.28617537e-01 -3.20454270e-01 -6.55863106e-01 -1.07285991e-01
1.42550796e-01 3.96437854e-01 3.67446721e-01 -1.35775292e+00
-3.66194636e-01 -3.42400707e-02 3.12350452e-01 -1.27575636e-01
-4.14653093e-01 3.73836428e-01 -6.52630329e-01 -7.09557295e-01
-1.15006864e-02 -8.29901993e-01 3.13748568e-01 3.49793509e-02
-3.23384315e-01 2.66340910e-03 6.12483680e-01 -7.85745442e-01
1.93795645e+00 -2.04984665e+00 -3.48072261e-01 4.67290312e-01
-1.18115231e-01 5.12244344e-01 -2.80383557e-01 1.24098134e+00
-1.36792079e-01 8.70931506e-01 -3.38107109e-01 -1.66693211e-01
-3.21631767e-02 3.21168005e-01 -4.49790388e-01 -6.57396093e-02
4.06524092e-01 9.49870586e-01 -6.26133859e-01 -6.35648727e-01
-2.01210856e-01 9.72929820e-02 -8.63017261e-01 9.97441188e-02
-4.57670927e-01 -3.77556294e-01 -1.85839370e-01 5.04981935e-01
5.48581481e-01 -1.14495106e-01 2.89176196e-01 -2.75400221e-01
-1.32817432e-01 8.71259272e-01 -1.43943691e+00 1.74470937e+00
-5.26544929e-01 2.94439346e-01 -2.23771572e-01 -4.31164563e-01
9.73582447e-01 2.52742201e-01 2.07071230e-01 -1.09594846e+00
-5.04639685e-01 5.60966134e-01 -2.77273268e-01 -7.21128702e-01
7.11289048e-01 1.48217931e-01 -2.47054711e-01 5.32493234e-01
-1.40030369e-01 -9.68479260e-04 3.68342310e-01 5.75887263e-01
1.25217509e+00 4.37551558e-01 1.61491871e-01 1.64689049e-02
2.90783167e-01 3.25998574e-01 6.26037478e-01 9.58170116e-01
4.27543283e-01 4.00431216e-01 4.55634981e-01 -4.77417633e-02
-1.26871502e+00 -9.76545393e-01 -1.82776213e-01 7.57461846e-01
-1.63428053e-01 -1.08016908e+00 -7.59377658e-01 -8.70392978e-01
2.82969385e-01 1.22397459e+00 -2.64135361e-01 -4.16262031e-01
-9.50655818e-01 -3.23262602e-01 1.20264053e+00 7.42525756e-01
2.79806763e-01 -6.76907063e-01 -4.68502730e-01 5.90300083e-01
-2.90434599e-01 -1.00110984e+00 -3.85036409e-01 1.77648202e-01
-7.43651748e-01 -9.44367170e-01 8.44897553e-02 -1.94670156e-01
3.39015841e-01 5.36292195e-02 1.40804350e+00 5.19245088e-01
1.77079509e-03 2.35510513e-01 -4.71103728e-01 -4.12357479e-01
-7.64142036e-01 3.04413319e-01 -4.29520667e-01 -5.01098752e-01
5.32316208e-01 -3.57950091e-01 -2.15873688e-01 2.76998162e-01
-1.32836020e+00 -3.86680141e-02 1.86912507e-01 7.71589816e-01
2.06662923e-01 -9.10315439e-02 6.71780348e-01 -1.46321857e+00
4.67468530e-01 -3.83844346e-01 -4.76555198e-01 5.66700399e-01
-1.11214960e+00 3.43324810e-01 6.22193277e-01 7.52460510e-02
-8.91697347e-01 -3.43492061e-01 -4.48976815e-01 -9.83916745e-02
-3.74845080e-02 8.79590511e-01 -2.50026345e-01 -4.00786251e-02
7.30953932e-01 -1.62928000e-01 -1.85923293e-01 -8.11500371e-01
4.78041440e-01 6.72966599e-01 3.44422758e-01 -6.50464833e-01
8.35471272e-01 2.56699264e-01 -3.57105136e-01 -3.73750538e-01
-5.31307936e-01 -1.65806115e-01 -2.87961155e-01 6.04852028e-02
5.70266545e-01 -9.10097241e-01 -2.48686969e-01 -4.05363254e-02
-1.21383154e+00 -8.77899155e-02 -3.96118253e-01 8.55004564e-02
-2.61359215e-01 4.67073411e-01 -6.15290701e-01 -4.93813068e-01
-3.45663205e-02 -1.12837291e+00 7.73873746e-01 -1.37318641e-01
-7.21655250e-01 -7.96751559e-01 -3.93462889e-02 3.58261436e-01
6.83689237e-01 -3.15491222e-02 1.68662179e+00 -9.59641397e-01
-5.79251647e-01 -3.38020891e-01 -1.84699193e-01 1.89549819e-01
1.88463032e-01 3.14323246e-01 -8.35140347e-01 -3.45597684e-01
-1.49064958e-01 -1.28470287e-01 5.16002774e-01 -4.90702242e-01
1.00389063e+00 -4.66396838e-01 -6.81830570e-02 6.65644109e-01
1.89890826e+00 2.69186199e-01 8.23207617e-01 7.32214689e-01
6.87634230e-01 6.26511455e-01 3.86494011e-01 1.46488830e-01
6.95560455e-01 8.48510742e-01 4.00451839e-01 -2.38913208e-01
-2.98704863e-01 -8.60745609e-01 3.02419096e-01 9.90594089e-01
6.60214961e-01 -4.88347352e-01 -1.00305331e+00 4.87742037e-01
-1.91232598e+00 -8.19437206e-01 -3.02374452e-01 2.35308313e+00
9.95540559e-01 2.90854305e-01 -5.90146072e-02 6.86745569e-02
2.96635360e-01 -1.07737631e-01 -2.70767361e-01 -7.66576529e-01
-3.85771185e-01 3.07892114e-01 8.13477695e-01 5.69907129e-01
-5.05798042e-01 9.65339661e-01 6.93643951e+00 5.68229318e-01
-1.12366986e+00 4.03149836e-02 -1.72338709e-01 -1.36943027e-01
-9.40486193e-01 5.29397547e-01 -8.46213639e-01 6.20305181e-01
9.69448805e-01 -9.46896300e-02 3.87812048e-01 5.63967586e-01
-4.44670916e-02 -3.55384089e-02 -1.32943928e+00 4.30886090e-01
1.34952724e-01 -1.32737744e+00 3.20418060e-01 -1.86458513e-01
2.94168174e-01 -2.50123709e-01 -2.77069420e-01 2.93158025e-01
3.28966886e-01 -8.56508732e-01 1.03337967e+00 6.12142384e-01
5.36210954e-01 -6.31616771e-01 8.06650758e-01 2.17541426e-01
-8.54403317e-01 -4.10763435e-02 -2.47217387e-01 3.52844238e-01
7.67608285e-02 4.05923784e-01 -6.79407477e-01 8.53500187e-01
5.04279554e-01 3.03216696e-01 -1.17584407e+00 9.08647478e-01
-1.25633761e-01 6.09153092e-01 -2.54464716e-01 2.63272583e-01
-7.84953088e-02 2.05747023e-01 5.00925481e-01 1.49029112e+00
4.72176254e-01 -5.35755455e-01 1.26421541e-01 9.33898568e-01
-1.79956675e-01 3.12907457e-01 -7.28445947e-01 -2.58161187e-01
9.30715263e-01 5.54322422e-01 -1.01502933e-01 -4.40852344e-01
-9.07846451e-01 9.44859862e-01 4.22445446e-01 3.63769352e-01
-7.09428072e-01 -6.56580269e-01 5.22521198e-01 6.67221487e-01
4.05350655e-01 -5.88888042e-02 -5.15945315e-01 -1.06995559e+00
3.77391458e-01 -1.45566916e+00 5.85453510e-01 -9.48432624e-01
-1.13268852e+00 3.55453700e-01 1.10432915e-01 -8.95692766e-01
-3.69744718e-01 -1.65016979e-01 -3.06881249e-01 1.01630378e+00
-1.56045461e+00 -1.08050764e+00 1.34546727e-01 3.53363633e-01
3.18254419e-02 1.13768406e-01 8.21014047e-01 7.27400899e-01
-5.44036746e-01 1.16000295e+00 -8.09705853e-02 1.36720836e-01
1.20685756e+00 -1.14460886e+00 7.80850351e-01 1.10429442e+00
1.15884393e-01 1.39892745e+00 7.43790627e-01 -9.11265433e-01
-1.53821540e+00 -1.09552991e+00 1.45638537e+00 -7.04015255e-01
7.72973239e-01 -4.86582935e-01 -1.52852368e+00 6.44807637e-01
2.92386889e-01 -3.36512595e-01 7.05612302e-01 -7.65883178e-02
-1.04043067e+00 -1.66143939e-01 -9.84581292e-01 7.55662680e-01
9.89214420e-01 -1.05360889e+00 -7.14684904e-01 -7.92817865e-03
9.64220881e-01 -2.90909261e-01 -1.15644038e+00 2.72364289e-01
4.63653386e-01 -7.83042610e-01 7.70655870e-01 -8.82654309e-01
4.48428154e-01 -4.93461460e-01 -3.75824183e-01 -1.01972795e+00
-1.57023937e-01 -4.63735193e-01 -3.62500921e-02 1.69238281e+00
6.81697607e-01 -6.53390348e-01 5.32276452e-01 9.00490940e-01
-2.61158139e-01 -5.10022759e-01 -8.06586564e-01 -1.12371969e+00
3.08855474e-01 -4.18986112e-01 1.12780058e+00 9.42456007e-01
1.56700328e-01 1.40620008e-01 -2.39886522e-01 7.48839229e-02
1.88132048e-01 -1.21160738e-01 8.14262569e-01 -1.00425339e+00
-3.74100626e-01 -2.29124919e-01 4.87979800e-02 -7.18358397e-01
-1.19630031e-01 -1.02429473e+00 -1.10330246e-01 -1.73106790e+00
-4.28556912e-02 -6.98624015e-01 -1.61505342e-01 7.14388728e-01
-2.08769873e-01 -1.38378263e-01 5.44115067e-01 1.79239422e-01
-2.20731214e-01 -2.25384980e-02 6.06632292e-01 2.26462148e-02
-1.65565480e-02 -4.34211463e-01 -1.01817942e+00 3.82739514e-01
6.51951849e-01 -9.86082137e-01 -4.79061365e-01 -8.38696837e-01
8.52224588e-01 -8.76074284e-02 3.21063370e-01 -8.40189159e-01
3.47861797e-01 -5.86253637e-03 -8.23688060e-02 -3.05703968e-01
2.04764396e-01 -1.00646746e+00 4.17893142e-01 3.60974193e-01
-6.25904083e-01 5.90333760e-01 4.67386633e-01 5.96317947e-02
-1.60927847e-01 -7.36192346e-01 5.05860090e-01 -1.92242444e-01
-4.33550328e-01 -1.54709667e-01 -5.05834997e-01 4.37865555e-01
3.83361965e-01 -1.74806505e-01 -8.75141025e-01 -3.49041551e-01
-8.60278010e-02 2.99114976e-02 1.15003896e+00 5.86556137e-01
2.86617935e-01 -1.32632339e+00 -4.77705449e-01 3.22674960e-01
6.02951884e-01 -4.96438086e-01 -1.86378360e-01 3.41691405e-01
-3.70737582e-01 3.88056666e-01 8.33805799e-02 -6.23871051e-02
-1.02387834e+00 6.54121399e-01 1.95206538e-01 -2.20158592e-01
-5.79866409e-01 4.90495533e-01 -3.58110994e-01 -5.09687304e-01
2.92304903e-01 -1.95965543e-01 1.94628105e-01 -9.37508047e-02
1.63866684e-01 1.60374612e-01 6.23644471e-01 -2.81686425e-01
-4.90143269e-01 3.73974711e-01 -2.56370127e-01 -1.99832797e-01
1.09199429e+00 -1.14376433e-01 -8.46608579e-02 4.18883413e-01
1.21276653e+00 4.94389921e-01 -4.76318628e-01 -4.97883111e-01
3.99656087e-01 -5.04611492e-01 -1.28767267e-01 -1.15540051e+00
-7.06728160e-01 8.26922238e-01 3.33144069e-01 2.03700185e-01
8.33726108e-01 -3.41394335e-01 1.05405271e+00 4.06359762e-01
4.34766531e-01 -1.30185354e+00 -2.65801311e-01 6.15567505e-01
7.92922258e-01 -9.17657495e-01 7.28800818e-02 -5.12665212e-01
-5.16232491e-01 1.03783369e+00 6.36555433e-01 2.69651294e-01
2.92983741e-01 7.10570276e-01 1.82725310e-01 -1.93105355e-01
-7.86709309e-01 -6.71796352e-02 -1.28936291e-01 3.62937003e-01
7.06354439e-01 -3.77526343e-01 -3.91973287e-01 4.82556403e-01
-2.25265294e-01 1.56011581e-01 7.08193243e-01 1.37756085e+00
-1.82172298e-01 -1.79047012e+00 -2.21428007e-01 6.47871554e-01
-5.70367455e-01 -4.92661268e-01 -5.43460965e-01 9.72543359e-01
2.84230709e-01 8.68716896e-01 -1.04725792e-03 -5.64353466e-01
8.49807560e-01 6.62979424e-01 2.74445891e-01 -6.96409702e-01
-1.24572289e+00 -2.71789759e-01 6.12723112e-01 -7.95194685e-01
2.17452332e-01 -4.54551607e-01 -1.11711287e+00 -5.49802840e-01
-3.37268919e-01 9.49838832e-02 7.95038640e-01 6.61943078e-01
7.24110305e-01 4.36455339e-01 2.92082787e-01 4.90572780e-01
-6.10628605e-01 -5.51360250e-01 -8.28040484e-03 6.78688049e-01
8.69070068e-02 -4.63807523e-01 -1.17228463e-01 -1.51592687e-01] | [9.690458297729492, 7.890426158905029] |
3a430fc0-0b2c-4d1f-83cc-080a86ad1988 | multiple-object-tracking-with-context | 1411.7935 | null | http://arxiv.org/abs/1411.7935v2 | http://arxiv.org/pdf/1411.7935v2.pdf | Multiple object tracking with context awareness | Multiple people tracking is a key problem for many applications such as
surveillance, animation or car navigation, and a key input for tasks such as
activity recognition. In crowded environments occlusions and false detections
are common, and although there have been substantial advances in recent years,
tracking is still a challenging task. Tracking is typically divided into two
steps: detection, i.e., locating the pedestrians in the image, and data
association, i.e., linking detections across frames to form complete
trajectories.
For the data association task, approaches typically aim at developing new,
more complex formulations, which in turn put the focus on the optimization
techniques required to solve them. However, they still utilize very basic
information such as distance between detections. In this thesis, I focus on the
data association task and argue that there is contextual information that has
not been fully exploited yet in the tracking community, mainly social context
and spatial context coming from different views. | ['Laura Leal-Taixé'] | 2014-11-24 | null | null | null | null | ['multiple-people-tracking'] | ['computer-vision'] | [-5.49898902e-03 -4.96971369e-01 -2.12255418e-01 9.65994038e-03
-1.78505138e-01 -5.50045788e-01 6.16431773e-01 4.13295418e-01
-5.92499554e-01 7.13120878e-01 1.15107827e-01 -5.76460399e-02
4.50732969e-02 -5.01293242e-01 -3.49538952e-01 -8.18290949e-01
3.60007621e-02 2.45220661e-01 7.33492315e-01 1.88202769e-01
8.51743370e-02 7.39037693e-01 -1.75725770e+00 -2.95513004e-01
5.19757569e-01 6.30078256e-01 1.57172814e-01 7.18441606e-01
-2.29652584e-01 6.09923780e-01 -5.40735424e-01 -4.43512917e-01
2.38851514e-02 -2.71162063e-01 -3.32404822e-01 3.75547349e-01
3.20044160e-01 -4.98542525e-02 -1.89485192e-01 1.11865866e+00
3.12426597e-01 5.91165423e-01 2.20989630e-01 -1.62500799e+00
9.02202502e-02 -1.02325827e-02 -6.96717501e-01 3.28294814e-01
3.40164155e-01 5.28181009e-02 7.71509290e-01 -7.59625137e-01
6.83392644e-01 1.20500648e+00 5.78235209e-01 4.46046293e-01
-1.17252469e+00 -5.53397596e-01 5.65978110e-01 3.53678197e-01
-1.38093948e+00 -3.93215865e-01 6.40124083e-01 -9.38212514e-01
3.08359206e-01 3.94219905e-01 8.19243610e-01 9.55472589e-01
-2.36295968e-01 9.65539575e-01 8.37541759e-01 -2.56943583e-01
1.99196458e-01 3.65182340e-01 3.60533357e-01 3.13800663e-01
5.15860975e-01 9.45726596e-03 -4.34437394e-01 -1.24532983e-01
5.72359085e-01 2.32962385e-01 -1.71670333e-01 -5.20916045e-01
-1.04544806e+00 8.59916687e-01 2.87157357e-01 5.41516006e-01
-4.49946523e-01 -1.29185945e-01 1.88439012e-01 -8.15508142e-02
3.57496232e-01 8.88801366e-02 7.55737126e-02 -2.24896878e-01
-9.60382640e-01 6.50519848e-01 7.23311245e-01 9.00186658e-01
7.95646608e-01 -2.37663612e-01 -2.37081394e-01 4.21988308e-01
4.03140664e-01 5.20580888e-01 -8.38802010e-02 -7.64735281e-01
3.48435611e-01 9.00233924e-01 5.82498312e-01 -1.30746222e+00
-4.13788736e-01 -3.77902150e-01 -7.27166414e-01 2.44197801e-01
8.61390948e-01 -3.76725376e-01 -4.21727628e-01 1.81948388e+00
8.69636536e-01 3.40660542e-01 -2.79481858e-01 9.57411289e-01
5.80588162e-01 3.34450960e-01 2.30269358e-01 -4.26401407e-01
1.45684731e+00 -8.15365911e-01 -8.41248333e-01 -4.71071333e-01
4.07166660e-01 -7.67560780e-01 2.57668942e-01 1.44026861e-01
-9.75916147e-01 -5.94420731e-01 -7.41191804e-01 1.38576180e-01
-4.87557858e-01 1.88667178e-01 3.63591641e-01 7.12425530e-01
-7.47899115e-01 7.56244585e-02 -7.84339726e-01 -6.76469684e-01
3.37352514e-01 3.11060280e-01 -2.76416421e-01 -1.25667870e-01
-8.37341726e-01 9.29533958e-01 3.21061790e-01 1.58589929e-01
-4.48411345e-01 -2.60509193e-01 -6.66935921e-01 -1.79307923e-01
9.13944840e-01 -6.27234340e-01 9.16908681e-01 -7.76595294e-01
-9.18667495e-01 7.59955406e-01 -6.14828408e-01 -4.09474432e-01
9.07630265e-01 -3.56379658e-01 -3.57093930e-01 -3.15747559e-01
2.46966749e-01 3.36598098e-01 8.28504086e-01 -1.26002979e+00
-1.16340351e+00 -4.86567408e-01 2.08178371e-01 1.63504571e-01
-1.31596610e-01 5.23118854e-01 -8.30289841e-01 -5.31187117e-01
-1.00120410e-01 -1.20900524e+00 -3.19644451e-01 1.11132734e-01
-1.89802572e-01 -4.05136347e-01 1.09240913e+00 -5.84881902e-01
1.31773388e+00 -2.04319406e+00 3.78153026e-01 1.48605853e-01
2.62586206e-01 5.32406867e-01 3.52566302e-01 3.18538129e-01
2.59033650e-01 -1.92825675e-01 8.84601399e-02 -6.02909386e-01
-7.88766295e-02 1.20263942e-01 -3.47831547e-02 5.61376393e-01
1.27164662e-01 5.59106052e-01 -1.10469139e+00 -5.58423877e-01
4.68800187e-01 5.27459562e-01 -1.05320260e-01 -4.54269908e-03
-9.16284546e-02 9.20493424e-01 -5.77223957e-01 4.47867066e-01
5.85808873e-01 -1.61091223e-01 -5.59335016e-03 1.01556577e-01
-6.84755325e-01 -2.93139696e-01 -1.63423371e+00 1.26792395e+00
-2.58529596e-02 8.25497448e-01 4.09658760e-01 -1.01799881e+00
5.45401692e-01 3.23215842e-01 7.71790862e-01 -4.44367498e-01
1.96325690e-01 -4.29897904e-02 4.50520329e-02 -5.15808284e-01
7.39541173e-01 1.44307524e-01 1.07912868e-01 3.44133675e-01
-5.79280019e-01 5.84005415e-01 6.50775969e-01 8.33893567e-03
9.68000531e-01 1.05113685e-01 3.88567716e-01 1.25147805e-01
8.73141587e-01 3.87010664e-01 7.37288237e-01 6.84989154e-01
-3.99028331e-01 2.22242922e-01 2.01234803e-01 -4.31149691e-01
-8.35151613e-01 -6.14686370e-01 9.79906097e-02 1.01730108e+00
5.06824136e-01 -3.67495239e-01 -7.21281648e-01 -4.79385912e-01
-8.49092081e-02 1.21616080e-01 -5.42583823e-01 3.09703290e-01
-8.79056454e-01 -6.73635185e-01 1.57000676e-01 5.24286091e-01
3.95828128e-01 -7.86120534e-01 -9.96711791e-01 3.89503270e-01
-2.69430459e-01 -1.45239341e+00 -4.13452595e-01 -3.24727774e-01
-6.47241116e-01 -1.28137672e+00 -9.62492526e-01 -3.44254702e-01
5.99140584e-01 7.50391662e-01 9.26446855e-01 2.74683535e-01
-4.48224038e-01 5.32050669e-01 -3.51668745e-01 -6.77662253e-01
-5.47597669e-02 -7.60427909e-03 5.65333664e-02 4.41513598e-01
4.25257117e-01 -5.05431928e-02 -5.22622347e-01 7.25923240e-01
-6.24084771e-01 -3.92936058e-02 5.31599760e-01 3.50422055e-01
4.24798787e-01 1.75815701e-01 1.34751216e-01 -6.65617764e-01
3.52899104e-01 -3.50520819e-01 -8.52730095e-01 2.74068385e-01
1.19668148e-01 -3.80906433e-01 1.71834096e-01 -5.22148430e-01
-9.37266529e-01 3.21548909e-01 1.48367407e-02 -3.56659740e-01
-5.27773201e-01 2.30384022e-01 -4.06408608e-01 -7.42408410e-02
3.15986842e-01 1.00808507e-02 -5.12734316e-02 -2.19955519e-01
8.14769566e-02 3.19749743e-01 1.47581577e-01 -3.05565357e-01
8.94658625e-01 7.24581420e-01 1.93943262e-01 -1.45571935e+00
-8.74933600e-01 -1.04600203e+00 -8.32934260e-01 -6.72797501e-01
1.02827561e+00 -8.54874432e-01 -8.45709443e-01 3.20333153e-01
-1.22149730e+00 -1.57878397e-03 -3.99841443e-02 5.77978611e-01
-1.73823953e-01 3.85539174e-01 2.34124810e-02 -1.29538500e+00
1.67904496e-01 -1.21186543e+00 9.78013217e-01 5.35960793e-01
-2.88724840e-01 -1.14610803e+00 8.57363641e-02 3.80346358e-01
3.42140436e-01 4.77009356e-01 2.95302510e-01 -5.08113682e-01
-9.96741056e-01 -5.35402298e-01 -1.83572263e-01 -1.28964305e-01
1.24804489e-01 9.01686996e-02 -9.09500420e-01 -2.82839656e-01
-1.26458138e-01 3.03107738e-01 5.53828537e-01 5.89931071e-01
5.65332294e-01 7.60435089e-02 -9.00184691e-01 8.51714388e-02
9.91510987e-01 3.89796376e-01 3.32762599e-01 2.65692085e-01
7.17414320e-01 9.43470359e-01 9.75120604e-01 2.48255134e-01
4.53651041e-01 1.07831883e+00 3.58741760e-01 -6.52407780e-02
-8.77603441e-02 1.98707417e-01 2.12822765e-01 2.65950888e-01
-4.23687577e-01 -2.30767682e-01 -8.35514605e-01 4.65907454e-01
-2.48689747e+00 -1.27583981e+00 -4.61660922e-01 2.38780522e+00
7.31903240e-02 7.05615012e-03 6.19153857e-01 2.44820327e-01
1.13399446e+00 7.05151632e-02 -5.40688992e-01 3.80791426e-01
3.03682219e-02 -4.08959150e-01 3.62516195e-01 5.19524336e-01
-1.45308971e+00 7.04996049e-01 5.30349493e+00 5.45422912e-01
-9.39347923e-01 3.07761636e-02 7.85413235e-02 3.18115167e-02
4.43950206e-01 4.79455851e-02 -1.26990354e+00 6.83457732e-01
3.47618908e-01 3.96256819e-02 8.15271586e-02 6.77556872e-01
4.82576191e-01 -5.38183689e-01 -9.87134576e-01 1.12275064e+00
-6.52661845e-02 -9.17393923e-01 -4.73168880e-01 4.10262674e-01
5.97606778e-01 -4.03844416e-01 -1.62150830e-01 1.52820051e-01
7.10779727e-02 -5.81323326e-01 6.82753861e-01 6.45253241e-01
-4.04926054e-02 -5.71006119e-01 7.23878860e-01 6.76100314e-01
-1.60474277e+00 -5.11913486e-02 -1.72445923e-01 -1.30526587e-01
4.72190022e-01 6.53099000e-01 -4.24257457e-01 7.28521705e-01
5.64768255e-01 8.26587558e-01 -4.02367592e-01 1.62550938e+00
-2.46894106e-01 9.88426581e-02 -3.31139505e-01 -1.88480824e-01
1.93317816e-01 -4.64184403e-01 8.87895346e-01 1.08590961e+00
2.91080564e-01 -8.61473083e-02 7.44189203e-01 5.95906198e-01
3.08834732e-01 -5.09675369e-02 -6.15526378e-01 2.00467542e-01
3.52526933e-01 1.27118552e+00 -1.04772472e+00 -3.18450153e-01
-6.78748369e-01 5.48409045e-01 7.61945024e-02 3.19462717e-01
-9.61823046e-01 1.03010058e-01 9.91168201e-01 3.14556092e-01
2.39438042e-01 -4.85443354e-01 1.14104189e-01 -1.05465508e+00
1.76242232e-01 -5.48331141e-01 3.21779281e-01 -2.55124390e-01
-8.78176332e-01 1.55671418e-01 1.85703725e-01 -1.28885067e+00
-1.90907061e-01 -5.11593938e-01 -6.18060887e-01 8.23490024e-01
-1.48081446e+00 -9.94484007e-01 -6.55010283e-01 6.72891259e-01
5.47150493e-01 2.17308909e-01 3.11087251e-01 6.70953989e-01
-9.73562300e-01 -3.57592069e-02 4.50915471e-03 3.48577052e-01
4.98369306e-01 -1.00215197e+00 4.83426191e-02 1.26799977e+00
1.11711442e-01 5.08430779e-01 9.07647848e-01 -6.93229735e-01
-1.28730953e+00 -1.12295783e+00 8.48380804e-01 -4.85334486e-01
6.36347532e-01 -2.78401613e-01 -7.41369605e-01 5.64073265e-01
-1.45954400e-01 1.02560699e-01 5.43026030e-01 4.14075181e-02
3.62502962e-01 1.84791505e-01 -7.81247199e-01 6.98579192e-01
1.04562652e+00 -1.87623650e-01 -2.27315441e-01 1.51400149e-01
7.83696175e-02 -4.45651174e-01 -4.69533324e-01 6.99901804e-02
5.85929930e-01 -9.12388504e-01 1.16167760e+00 -2.98175305e-01
-4.27615583e-01 -6.98967040e-01 1.25933141e-01 -9.21408832e-01
-1.82812899e-01 -4.77138609e-01 -2.68211365e-01 1.37720692e+00
-1.86916068e-02 -4.64528203e-01 1.00442350e+00 8.92879426e-01
2.02338517e-01 -2.47888744e-01 -8.74146819e-01 -8.55804622e-01
-4.99548912e-01 -3.89962524e-01 2.84798741e-01 8.21934402e-01
-4.62444425e-01 3.69355619e-01 -5.67552805e-01 3.67506325e-01
8.76560271e-01 2.24007860e-01 1.22700715e+00 -1.74103713e+00
1.72152326e-01 -5.66837430e-01 -5.44172287e-01 -1.10334027e+00
-7.56829157e-02 -2.91284382e-01 1.17944248e-01 -1.48162043e+00
1.42626002e-01 -5.78050315e-01 2.16761813e-01 8.47812518e-02
-4.44310069e-01 2.48087086e-02 3.62905800e-01 2.45689586e-01
-8.60724628e-01 2.62861013e-01 9.95102465e-01 -1.65038332e-01
-2.00028673e-01 5.57914495e-01 -3.58531415e-01 9.36516225e-01
7.15009511e-01 -4.73357916e-01 -1.69545308e-01 -3.67488533e-01
1.61431357e-01 2.06972938e-02 7.54275203e-01 -1.35073173e+00
7.35641479e-01 -3.65873635e-01 2.43639529e-01 -7.73647130e-01
7.28859007e-01 -1.13078153e+00 3.77889484e-01 6.03231728e-01
8.81948546e-02 1.56609848e-01 1.68711275e-01 8.01646292e-01
-3.00911069e-01 -4.28190380e-01 5.13622642e-01 -2.41863489e-01
-9.68538165e-01 3.79153937e-01 -3.53849262e-01 2.22800933e-02
1.47532928e+00 -4.94277805e-01 -9.65054929e-02 -2.20683366e-01
-1.13265717e+00 4.60806429e-01 5.63748717e-01 4.73540366e-01
2.64352888e-01 -1.05842328e+00 -6.14186108e-01 -1.19312190e-01
3.73291485e-02 5.40822074e-02 1.87325343e-01 1.21598661e+00
-2.57423669e-02 4.60011363e-01 9.17548165e-02 -9.44263160e-01
-1.83405662e+00 5.13248324e-01 4.09100167e-02 -3.13257635e-01
-5.94516635e-01 5.28108478e-01 2.20113888e-01 2.30231240e-01
4.80788112e-01 -1.40882611e-01 -5.89089990e-01 6.06579721e-01
7.95801759e-01 7.52544820e-01 -1.62650406e-01 -1.14968383e+00
-4.63901311e-01 9.18372869e-01 1.99457526e-01 9.74429250e-02
1.01024914e+00 -3.52926463e-01 1.61152467e-01 4.63910520e-01
5.34364462e-01 -2.51535233e-02 -1.28440571e+00 -4.45908785e-01
3.74101013e-01 -8.16693187e-01 -2.49459043e-01 -1.04063123e-01
-9.76451993e-01 8.85994911e-01 5.49213231e-01 6.14954233e-01
7.15717494e-01 4.88691814e-02 4.46566939e-01 2.66277671e-01
4.85879928e-01 -1.01137388e+00 2.23802291e-02 5.33150256e-01
4.86251920e-01 -1.44421768e+00 1.05290808e-01 -5.42349637e-01
-5.36034167e-01 7.72545516e-01 4.61276740e-01 1.70606807e-01
5.71844220e-01 1.76632777e-01 -1.85369506e-01 -9.98425707e-02
-2.91907072e-01 -9.87670243e-01 2.21708864e-01 5.51661611e-01
2.15712234e-01 -1.11116141e-01 -1.02903664e-01 1.84692606e-01
1.73585027e-01 1.35542145e-02 1.88961223e-01 1.11178505e+00
-5.77105284e-01 -1.03859079e+00 -8.01303089e-01 1.98185489e-01
-4.37599897e-01 4.88141000e-01 -2.90142566e-01 9.87625599e-01
4.79116887e-01 1.22544134e+00 2.17968263e-02 2.39951331e-02
5.68943441e-01 -8.04237723e-02 4.18356329e-01 -5.14399886e-01
-5.19816935e-01 -2.73957811e-02 -1.93487313e-02 -3.24387312e-01
-1.05210042e+00 -1.13600481e+00 -7.48048663e-01 -3.48422438e-01
-5.41430831e-01 8.71000066e-02 7.45196283e-01 1.09852588e+00
-4.94948821e-03 5.66590726e-01 2.05329016e-01 -1.08740127e+00
1.43602565e-01 -5.81784189e-01 -2.39034548e-01 4.93555486e-01
4.17658865e-01 -1.17514312e+00 -2.97607817e-02 -3.21107060e-02] | [6.530906677246094, -1.9067034721374512] |
a63a091f-806d-46ec-95d3-28da1c8a1191 | a-novel-transformer-network-with-shifted | 2208.01252 | null | https://arxiv.org/abs/2208.01252v1 | https://arxiv.org/pdf/2208.01252v1.pdf | A Novel Transformer Network with Shifted Window Cross-Attention for Spatiotemporal Weather Forecasting | Earth Observatory is a growing research area that can capitalize on the powers of AI for short time forecasting, a Now-casting scenario. In this work, we tackle the challenge of weather forecasting using a video transformer network. Vision transformer architectures have been explored in various applications, with major constraints being the computational complexity of Attention and the data hungry training. To address these issues, we propose the use of Video Swin-Transformer, coupled with a dedicated augmentation scheme. Moreover, we employ gradual spatial reduction on the encoder side and cross-attention on the decoder. The proposed approach is tested on the Weather4Cast2021 weather forecasting challenge data, which requires the prediction of 8 hours ahead future frames (4 per hour) from an hourly weather product sequence. The dataset was normalized to 0-1 to facilitate using the evaluation metrics across different datasets. The model results in an MSE score of 0.4750 when provided with training data, and 0.4420 during transfer learning without using training data, respectively. | ['Panos Liatsis', 'Hasan Al Marzouqi', 'Alabi Bojesomo'] | 2022-08-02 | null | null | null | null | ['weather-forecasting'] | ['miscellaneous'] | [ 1.67181507e-01 2.95209289e-02 2.97539979e-01 -5.16561627e-01
-4.39453930e-01 -4.78704065e-01 8.69730711e-01 -3.75711352e-01
-4.65811521e-01 6.96253419e-01 3.21922123e-01 -3.48535836e-01
1.00123324e-01 -6.31926298e-01 -8.70624006e-01 -7.52916455e-01
-2.33526990e-01 4.57475940e-03 1.08519286e-01 -8.32750276e-02
-1.45929689e-02 1.33830765e-02 -1.49752569e+00 2.88770139e-01
7.17819631e-01 1.35615206e+00 5.65366924e-01 9.40859556e-01
2.96506315e-01 1.00595105e+00 -4.43766564e-01 -4.13134396e-01
5.00793576e-01 -1.70182258e-01 -3.48975271e-01 -2.60376241e-02
4.90492314e-01 -6.54825509e-01 -4.72174287e-01 8.60119641e-01
5.85656703e-01 1.99388772e-01 3.54413807e-01 -1.20664346e+00
-4.63353664e-01 2.79860109e-01 -3.44501972e-01 6.31522179e-01
-2.18580112e-01 1.33577839e-01 9.39736485e-01 -9.98220921e-01
2.44258299e-01 8.91428709e-01 6.60776436e-01 3.32894146e-01
-6.82376325e-01 -5.63647807e-01 1.93138823e-01 8.42560709e-01
-1.09377980e+00 -4.34387624e-01 4.88131344e-01 -5.58412373e-01
1.20324659e+00 2.47784391e-01 5.61331570e-01 9.33977902e-01
3.31227422e-01 5.73168218e-01 1.08027577e+00 -1.91156149e-01
1.99886903e-01 9.29167271e-02 -6.29682392e-02 1.74224406e-01
-2.44846657e-01 4.08258736e-01 -4.60339367e-01 3.19314390e-01
3.53073061e-01 -1.13357333e-02 -6.42565012e-01 1.83106214e-01
-1.08122909e+00 7.54985332e-01 5.80428362e-01 8.26429725e-02
-5.53822339e-01 1.50107175e-01 4.92140025e-01 5.72079003e-01
8.87375712e-01 1.34357974e-01 -7.13303208e-01 -1.32879719e-01
-1.22082293e+00 2.75151968e-01 5.88944972e-01 8.39273751e-01
4.27196532e-01 4.03980464e-01 -3.41555625e-02 4.63704884e-01
3.50779921e-01 7.28655100e-01 5.97168922e-01 -5.76830089e-01
7.10655272e-01 1.10471845e-01 2.08614528e-01 -9.01677251e-01
-4.75536972e-01 -6.69305205e-01 -1.02362478e+00 2.00909987e-01
1.84418149e-02 -6.15114033e-01 -1.10878849e+00 1.54455590e+00
2.20766857e-01 8.00519526e-01 3.29033226e-01 1.18370807e+00
7.04654217e-01 1.42014849e+00 -4.73167785e-02 -2.47782230e-01
1.28579223e+00 -1.35909629e+00 -7.47160077e-01 -1.47649512e-01
4.39668000e-01 -7.96548963e-01 9.12288189e-01 4.66515213e-01
-7.99476266e-01 -8.18076968e-01 -1.06817532e+00 3.42562310e-02
-3.81616145e-01 3.69668454e-01 1.84817523e-01 3.65268826e-01
-1.08967805e+00 6.02825046e-01 -7.30360806e-01 -2.25893080e-01
-6.87141269e-02 9.25531704e-03 -1.96625367e-01 -5.84971532e-02
-1.46157134e+00 1.00382364e+00 5.32466412e-01 4.53756154e-01
-1.11249745e+00 -1.06154883e+00 -6.64549053e-01 4.33110118e-01
7.71026462e-02 -4.87304181e-01 1.15863228e+00 -1.13334966e+00
-1.67329812e+00 1.47019595e-01 2.51193494e-01 -1.05405903e+00
5.98064601e-01 -6.95915759e-01 -8.03573012e-01 7.84440935e-02
-2.06706166e-01 7.85689771e-01 1.04638898e+00 -5.70636988e-01
-9.53820527e-01 -2.91182101e-02 1.65360034e-01 4.27885354e-01
-5.91244757e-01 -6.04935698e-02 -1.70596421e-01 -9.21394885e-01
-4.83953565e-01 -1.03072882e+00 -9.47472379e-02 -1.71533406e-01
9.77375880e-02 1.54797494e-01 1.12047637e+00 -1.18266058e+00
1.09115100e+00 -2.28083253e+00 2.72260517e-01 -1.50581613e-01
-1.76477939e-01 5.49910545e-01 -1.88678741e-01 3.65068644e-01
-1.78519592e-01 -2.85142630e-01 -2.54719734e-01 -3.66215408e-01
-2.35642735e-02 2.73404807e-01 -8.35345149e-01 4.90706474e-01
2.37781689e-01 5.53888381e-01 -4.41192329e-01 1.47885576e-01
3.09675932e-01 7.50526965e-01 -4.72561657e-01 5.36323965e-01
-3.17738712e-01 5.88896692e-01 -3.90854962e-02 1.61077842e-01
8.90868723e-01 -1.38977647e-01 -1.00730546e-01 -2.30565548e-01
-5.05824983e-01 2.21522614e-01 -8.86493206e-01 1.48641562e+00
-5.52323699e-01 9.26052034e-01 -4.86847498e-02 -9.64271665e-01
7.56360412e-01 7.59173274e-01 1.62240267e-01 -1.04051876e+00
2.77419500e-02 1.38665875e-02 6.10323735e-02 -6.13787532e-01
4.25856292e-01 -1.00578181e-01 4.17899460e-01 1.45575672e-01
2.04063933e-02 2.77386215e-02 1.05263628e-01 -2.67321199e-01
8.31808567e-01 2.96326548e-01 -3.26403379e-02 -2.64074713e-01
4.52835202e-01 -8.77752379e-02 4.22668308e-01 1.96082309e-01
8.94243196e-02 6.24518871e-01 9.54875648e-02 -9.03816402e-01
-1.25688207e+00 -4.06499654e-01 -1.77129492e-01 1.02058339e+00
-2.27034405e-01 -2.67748475e-01 -6.67134225e-01 -3.91248673e-01
-2.58390039e-01 8.06449771e-01 -6.29177213e-01 7.57241100e-02
-4.78331208e-01 -9.17594969e-01 4.05071646e-01 4.60189313e-01
8.27382147e-01 -8.03594649e-01 -1.14778435e+00 2.80618012e-01
-2.66036361e-01 -1.42663097e+00 -2.53909498e-01 8.86657089e-02
-6.93750739e-01 -6.68288827e-01 -1.22472763e+00 -3.84881228e-01
1.90773785e-01 2.93970466e-01 9.06473577e-01 -3.01882774e-01
1.65817160e-02 2.08037406e-01 -5.62002838e-01 -3.85399133e-01
1.49611801e-01 1.56156734e-01 -1.01288065e-01 3.00359786e-01
-3.15528698e-02 -5.64102769e-01 -7.35158622e-01 2.44863555e-01
-8.54513645e-01 2.94569165e-01 4.77295369e-01 9.87095535e-01
1.81352198e-01 -1.33409649e-01 4.76795465e-01 -4.51216459e-01
1.67336032e-01 -7.89464712e-01 -1.05591440e+00 2.90652484e-01
-6.10382736e-01 -1.17327005e-01 8.41318369e-01 -2.15817079e-01
-1.17142129e+00 -3.16420309e-02 6.25202432e-03 -6.82525396e-01
2.40608025e-02 9.13099885e-01 7.37860501e-02 6.77840561e-02
3.00573647e-01 3.83148015e-01 -3.86890322e-01 -5.80755234e-01
2.92128950e-01 6.64547145e-01 5.16665697e-01 5.54112606e-02
5.57019889e-01 2.07752094e-01 -1.31579220e-01 -8.28786433e-01
-8.52002382e-01 -2.07674652e-01 -2.80805498e-01 -3.28633755e-01
9.90356028e-01 -1.36128998e+00 -4.02872115e-01 4.48395193e-01
-1.00939047e+00 -4.18876648e-01 1.05450943e-01 8.66019547e-01
-2.03238457e-01 3.01666707e-01 -4.86275434e-01 -6.64472103e-01
-7.11010098e-01 -9.10363674e-01 7.60851443e-01 2.11740687e-01
2.66810864e-01 -7.91656852e-01 2.70036429e-01 2.27879554e-01
8.17656636e-01 1.67163163e-01 5.70668697e-01 -3.51699620e-01
-6.92667305e-01 -2.38976832e-02 -3.96104068e-01 5.78052104e-01
-1.77919447e-01 -2.30071053e-01 -1.35233986e+00 -5.86862862e-01
2.24207133e-01 -3.35530430e-01 9.16625857e-01 2.99729079e-01
1.08259058e+00 -4.77868289e-01 1.68107569e-01 8.77539158e-01
1.31156051e+00 3.22202593e-01 6.76814556e-01 3.38092119e-01
6.27595782e-01 5.10749757e-01 6.29704177e-01 7.07716644e-01
5.52187324e-01 6.47413969e-01 6.51525438e-01 -1.52828813e-01
4.02219445e-02 9.97165740e-02 5.25504708e-01 9.05244827e-01
-2.73412466e-01 -6.90561950e-01 -9.61409330e-01 5.43777466e-01
-1.77820623e+00 -9.68703389e-01 4.77199741e-02 2.08447027e+00
3.86654407e-01 1.81273371e-02 -3.44708174e-01 2.20674425e-01
3.13554198e-01 4.35880482e-01 -4.36940700e-01 -1.95726722e-01
-3.29807326e-02 1.68263286e-01 6.45334899e-01 5.30803919e-01
-1.40835249e+00 6.86835587e-01 5.58385181e+00 4.46642697e-01
-1.75129509e+00 1.30852804e-01 7.31776714e-01 -1.48016691e-01
-6.33375114e-03 -1.96145579e-01 -8.22873056e-01 7.60521650e-01
1.57180262e+00 -1.94363594e-01 3.46124828e-01 6.97813213e-01
3.29135418e-01 8.82054791e-02 -7.45288968e-01 8.06501567e-01
1.29202589e-01 -1.24294329e+00 -1.12398593e-02 -2.70697147e-01
6.44999981e-01 4.74996954e-01 1.88267633e-01 3.85840774e-01
-2.33659998e-01 -9.81463492e-01 8.22307467e-01 5.58556080e-01
7.09866405e-01 -6.17609441e-01 1.03926420e+00 4.29185927e-01
-1.26035666e+00 -1.42273724e-01 -4.86455113e-01 -4.10648108e-01
2.29319051e-01 6.30236804e-01 -9.42877650e-01 7.06175864e-01
1.10309446e+00 9.50733483e-01 -2.44857460e-01 1.11388993e+00
1.52780667e-01 9.76007223e-01 -4.55284983e-01 3.18386316e-01
5.06620169e-01 -1.52102798e-01 3.11264127e-01 1.15549088e+00
8.13314319e-01 1.89636365e-01 7.03334808e-02 1.70334160e-01
1.24805026e-01 3.89804319e-02 -4.96846944e-01 1.11318968e-01
-7.99078941e-02 1.20449233e+00 -2.48044655e-01 -5.24768531e-01
-7.70752251e-01 1.05925906e+00 -2.41619479e-02 4.12267983e-01
-1.29104590e+00 -1.88766956e-01 5.31617045e-01 -1.48572952e-01
8.65531564e-01 -2.38337636e-01 5.29640764e-02 -1.19574547e+00
2.45919541e-01 -6.78026438e-01 1.89509764e-01 -1.05629849e+00
-8.75418484e-01 1.02435100e+00 -1.32651459e-02 -1.56589448e+00
-2.77377546e-01 -5.46187699e-01 -5.67276657e-01 9.64435279e-01
-2.15490460e+00 -1.18831336e+00 -5.83409011e-01 3.84907454e-01
7.26060092e-01 -2.48505533e-01 7.56860077e-01 9.17121351e-01
-5.20487309e-01 4.09085542e-01 2.53681153e-01 -2.27500856e-01
5.88789880e-01 -1.03807449e+00 6.98234081e-01 1.10931468e+00
1.02183580e-01 -7.61327296e-02 8.16236258e-01 -2.38517016e-01
-1.17505372e+00 -1.49455810e+00 1.07842219e+00 6.71100318e-02
4.99189228e-01 -2.61774689e-01 -1.04065704e+00 6.57675683e-01
7.66538441e-01 2.27496997e-01 3.72839063e-01 -5.18307030e-01
-2.59444624e-01 -2.38450497e-01 -8.17231536e-01 2.89422750e-01
4.48016793e-01 -3.35448921e-01 -4.39956814e-01 3.80260348e-01
7.38947809e-01 -5.42472184e-01 -7.96191692e-01 4.53991890e-01
3.93199503e-01 -9.41701233e-01 6.98904514e-01 -3.76683891e-01
4.92490709e-01 -5.54214358e-01 -3.35305810e-01 -1.51338339e+00
-4.30162609e-01 -4.45676029e-01 -4.61188741e-02 9.74425852e-01
5.59781611e-01 -4.02742237e-01 5.24869084e-01 3.79096866e-01
-4.05573100e-01 -4.58266139e-01 -1.15309095e+00 -5.89183092e-01
-1.19709201e-01 -3.52278411e-01 5.52615225e-01 9.27283466e-01
-3.74100953e-01 3.94221395e-01 -1.15887618e+00 5.51770806e-01
2.47327596e-01 7.71051943e-02 5.94173431e-01 -9.16926026e-01
-5.49573421e-01 1.29159078e-01 -2.84978032e-01 -1.26916122e+00
-1.16927877e-01 -4.75488812e-01 -2.20543649e-02 -1.10752952e+00
-2.67622828e-01 5.76536879e-02 -3.70775402e-01 4.65270340e-01
3.68759334e-02 2.13300198e-01 3.56773108e-01 8.21157843e-02
-2.35295743e-01 9.45101261e-01 7.84978330e-01 -1.34091616e-01
2.37337351e-01 2.06332095e-03 4.97301817e-02 6.11936152e-01
8.90013039e-01 -2.81192809e-01 -6.74240470e-01 -9.95794773e-01
2.04314664e-01 2.24220648e-01 3.98484677e-01 -1.25007236e+00
2.48050898e-01 1.16579756e-01 3.47883195e-01 -6.53841913e-01
5.82810044e-01 -1.10603261e+00 4.02122170e-01 4.86944646e-01
-1.40676603e-01 6.11692190e-01 4.30124491e-01 5.36137104e-01
-6.49367452e-01 3.80276889e-02 6.77445173e-01 1.40516773e-01
-1.13230014e+00 2.42659077e-01 -3.51439506e-01 -1.15909837e-01
1.10875070e+00 1.89430133e-01 -2.93748021e-01 -4.24260288e-01
-6.34892166e-01 3.02048296e-01 -8.85475725e-02 5.25004864e-01
5.23830950e-01 -1.05853832e+00 -9.85739589e-01 4.00764644e-01
-1.70909930e-02 -3.17973673e-01 5.43188691e-01 8.04534376e-01
-7.19999909e-01 5.70884049e-01 -3.70879680e-01 -5.68424881e-01
-1.12664485e+00 4.76818562e-01 4.50007737e-01 -9.83948484e-02
-8.21970642e-01 7.12924838e-01 3.61179322e-01 -1.54109553e-01
2.48249367e-01 -8.35930407e-01 -4.07008171e-01 2.04531893e-01
7.99860060e-01 3.54094714e-01 3.48328590e-01 -3.25122178e-01
-1.95443273e-01 5.35517633e-01 9.97682917e-04 -1.71810612e-01
1.56008959e+00 -2.31901258e-01 3.19355875e-01 3.62381876e-01
9.72868145e-01 -5.20844638e-01 -1.66878867e+00 -2.18251035e-01
-8.18245113e-02 -4.33421731e-01 4.41223502e-01 -8.24877918e-01
-1.34810162e+00 1.14215767e+00 9.93977547e-01 2.33611912e-01
1.31501436e+00 -6.74922943e-01 8.95153642e-01 3.66366684e-01
-3.71790230e-02 -7.73168325e-01 -4.83996779e-01 8.00178647e-01
1.08130896e+00 -1.41434443e+00 -2.92426586e-01 2.29226016e-02
-8.18036497e-01 1.04410756e+00 3.72508198e-01 -1.30615816e-01
8.44007373e-01 2.89312810e-01 2.83715010e-01 1.61809906e-01
-1.28938031e+00 7.42111132e-02 3.67760420e-01 1.43409878e-01
4.93981421e-01 -1.22393593e-02 -9.06483755e-02 3.34258705e-01
-1.29863203e-01 3.36756885e-01 3.49756896e-01 6.64316416e-01
-3.07599515e-01 -6.67656004e-01 -2.66437888e-01 9.38553959e-02
-5.56476414e-01 -5.39199233e-01 3.26772690e-01 4.11815673e-01
8.16239119e-02 7.81864464e-01 2.57606566e-01 -3.59090954e-01
3.25982451e-01 4.54210080e-02 -1.59805082e-02 -3.38676199e-02
-5.96403480e-01 8.71049091e-02 3.20114829e-02 -3.96147937e-01
-6.25651419e-01 -5.72821438e-01 -7.14008629e-01 -2.71407723e-01
-5.28585576e-02 1.96236625e-01 6.63090110e-01 9.44572330e-01
6.39479101e-01 5.92237055e-01 7.28112340e-01 -9.90336597e-01
-4.96184796e-01 -1.19918346e+00 -3.59112769e-01 4.16075699e-02
7.33551323e-01 -4.30837035e-01 -3.57835472e-01 4.77233231e-01] | [6.749782562255859, 2.824658155441284] |
62b237f9-6810-4cfb-bd41-ebc5c95c9b7a | online-mutual-foreground-segmentation-for | 1809.02851 | null | http://arxiv.org/abs/1809.02851v2 | http://arxiv.org/pdf/1809.02851v2.pdf | Online Mutual Foreground Segmentation for Multispectral Stereo Videos | The segmentation of video sequences into foreground and background regions is
a low-level process commonly used in video content analysis and smart
surveillance applications. Using a multispectral camera setup can improve this
process by providing more diverse data to help identify objects despite adverse
imaging conditions. The registration of several data sources is however not
trivial if the appearance of objects produced by each sensor differs
substantially. This problem is further complicated when parallax effects cannot
be ignored when using close-range stereo pairs. In this work, we present a new
method to simultaneously tackle multispectral segmentation and stereo
registration. Using an iterative procedure, we estimate the labeling result for
one problem using the provisional result of the other. Our approach is based on
the alternating minimization of two energy functions that are linked through
the use of dynamic priors. We rely on the integration of shape and appearance
cues to find proper multispectral correspondences, and to properly segment
objects in low contrast regions. We also formulate our model as a frame
processing pipeline using higher order terms to improve the temporal coherence
of our results. Our method is evaluated under different configurations on
multiple multispectral datasets, and our implementation is available online. | ['Guillaume-Alexandre Bilodeau', 'Pierre-Luc St-Charles', 'Robert Bergevin'] | 2018-09-08 | null | null | null | null | ['foreground-segmentation'] | ['computer-vision'] | [ 7.28432536e-01 -6.90822423e-01 2.62346745e-01 -2.95846522e-01
-6.56855226e-01 -8.79163861e-01 6.08130217e-01 2.84255505e-01
-5.61659515e-01 5.11744320e-01 -2.58267969e-01 1.06467418e-01
-1.59334779e-01 -8.16295624e-01 -4.78395581e-01 -9.80052769e-01
3.20799768e-01 2.35004246e-01 8.67877543e-01 4.78713168e-03
1.83194950e-01 6.85661733e-01 -1.74111485e+00 2.32444733e-01
9.08065856e-01 8.38648081e-01 5.55958509e-01 6.53205037e-01
8.40316527e-03 4.76854295e-01 -2.16861621e-01 -1.98491827e-01
5.71217239e-01 -2.93091357e-01 -7.57109046e-01 7.30811119e-01
5.76295853e-01 -3.68431568e-01 3.04808319e-01 1.33521998e+00
2.43889868e-01 3.52683276e-01 4.30333734e-01 -8.95631611e-01
3.33488375e-01 -4.75934260e-02 -8.26749504e-01 3.66687089e-01
3.70311856e-01 1.65220946e-01 6.30517304e-01 -4.86580193e-01
5.59695125e-01 9.25832689e-01 6.27404213e-01 -8.20878074e-02
-1.46231377e+00 -1.04478948e-01 7.82887042e-02 2.07942456e-01
-1.29850924e+00 -5.46565115e-01 8.99332643e-01 -7.17018902e-01
3.28234136e-01 3.04515123e-01 8.26387227e-01 5.69817960e-01
-1.17199026e-01 2.77053595e-01 1.31833768e+00 -4.97857720e-01
1.56021982e-01 -3.05558965e-02 1.09368823e-01 3.51644516e-01
3.01214963e-01 9.41519737e-02 -2.15986088e-01 -2.15308636e-01
6.99005663e-01 1.27846032e-01 -5.72366893e-01 -5.20199478e-01
-1.07785404e+00 4.72708851e-01 1.09553412e-01 6.23003006e-01
-5.73697925e-01 -1.00858517e-01 -2.29622629e-02 -9.67195258e-03
6.92127883e-01 7.78849889e-03 -2.10372448e-01 2.38545015e-01
-1.26328301e+00 2.89648864e-02 4.63489920e-01 3.91408950e-01
1.04062009e+00 -2.72017688e-01 8.33968222e-02 8.31728578e-01
3.35123211e-01 5.42822182e-01 2.95028333e-02 -1.09908509e+00
9.29112136e-02 2.79738456e-01 3.80707651e-01 -1.10351634e+00
-4.32558626e-01 -2.44978026e-01 -5.02517641e-01 2.84733772e-01
7.55672455e-01 -4.83369865e-02 -6.28263652e-01 1.30543017e+00
7.35885084e-01 3.39542985e-01 -1.41757786e-01 1.04689348e+00
4.46831822e-01 7.56131947e-01 7.27402493e-02 -5.19526899e-01
1.33817577e+00 -4.66984034e-01 -6.24406040e-01 -2.07995981e-01
1.05998076e-01 -1.09634888e+00 4.60117251e-01 3.64542693e-01
-1.07454586e+00 -5.24938524e-01 -7.49846101e-01 1.88265741e-01
-2.06209570e-01 2.07257032e-01 2.48856157e-01 7.11725354e-01
-1.05125976e+00 6.53100133e-01 -7.70443201e-01 -4.98851925e-01
1.94993615e-02 2.16944709e-01 -2.17888057e-01 -1.14791751e-01
-8.07373762e-01 9.03798342e-01 5.58865368e-01 3.09226602e-01
-4.19358760e-01 -3.88704717e-01 -8.80971372e-01 -2.89215744e-01
6.13184810e-01 -4.86743748e-01 8.10223222e-01 -1.28164101e+00
-1.38214219e+00 1.07184756e+00 -1.50613606e-01 -1.44336239e-01
5.84483385e-01 1.79489776e-02 -2.26014763e-01 4.10396665e-01
1.55133963e-01 3.21869135e-01 9.04471874e-01 -1.41486251e+00
-7.97715485e-01 -3.40309530e-01 1.11037262e-01 3.43903035e-01
7.16193067e-03 2.19207108e-01 -7.91907430e-01 -5.45739532e-01
2.72757292e-01 -8.76248360e-01 -1.99249506e-01 5.85387684e-02
-4.10890356e-02 4.24126595e-01 7.84286916e-01 -9.77742195e-01
7.66089976e-01 -2.09088826e+00 3.00286949e-01 3.17474425e-01
-1.02538466e-01 3.08096409e-01 -3.02246044e-04 3.67121063e-02
1.34337991e-01 -4.73345518e-01 -6.71732128e-01 -1.33876530e-02
-5.94641805e-01 5.82603514e-02 1.35944605e-01 8.84068906e-01
5.98734021e-02 3.13240081e-01 -9.09976661e-01 -5.82300842e-01
6.86160564e-01 5.08199990e-01 -2.59990305e-01 2.50016034e-01
-1.90110490e-01 8.10036600e-01 -2.77038038e-01 5.20160735e-01
9.71267223e-01 -1.12880776e-02 2.80251503e-01 -6.20856941e-01
-5.40110350e-01 -2.82567918e-01 -1.62533200e+00 1.63483226e+00
-2.35101655e-01 5.47565281e-01 4.75110978e-01 -9.15553689e-01
6.84736431e-01 1.93051934e-01 8.86551499e-01 -4.29655850e-01
2.51158237e-01 3.11741292e-01 -2.43925571e-01 -6.39224946e-01
6.34291649e-01 -1.79335214e-02 3.74010891e-01 2.43889511e-01
-5.50570190e-02 -3.57371569e-01 4.64384109e-01 -3.37342978e-01
5.72467804e-01 4.29668188e-01 2.90022552e-01 -2.42237031e-01
8.22270691e-01 1.04590498e-01 6.31997049e-01 5.59104800e-01
-1.11437172e-01 7.17615545e-01 1.05277084e-01 -3.77467006e-01
-8.91675353e-01 -8.34384680e-01 -2.42149785e-01 6.62726462e-01
4.30590689e-01 7.46991932e-02 -7.75383830e-01 -3.71299297e-01
-3.73667568e-01 4.22464669e-01 -2.02329248e-01 3.51969630e-01
-5.71867824e-01 -1.18635547e+00 -8.16386007e-03 -1.11380234e-01
6.78094327e-01 -5.90152502e-01 -9.59864378e-01 2.80767471e-01
-5.48615873e-01 -1.43245697e+00 -3.56760204e-01 3.31491753e-02
-8.29390228e-01 -1.41780496e+00 -8.83249879e-01 -4.29932863e-01
5.29501677e-01 7.36455917e-01 9.40822124e-01 2.66269092e-02
-2.20992148e-01 6.80923522e-01 -4.07620817e-01 6.93729371e-02
-4.38731998e-01 -2.88503349e-01 -2.33725041e-01 6.32633030e-01
-5.44124097e-03 -4.16971236e-01 -4.62287158e-01 4.34347302e-01
-1.40056252e+00 2.59996712e-01 3.04960966e-01 5.38920760e-01
6.70989573e-01 2.10158914e-01 -2.75869787e-01 -5.16400933e-01
1.07941478e-02 -2.35389978e-01 -1.22886145e+00 4.30639952e-01
-3.61498296e-02 -1.66494742e-01 2.13588804e-01 -3.30420226e-01
-1.23013127e+00 6.09752715e-01 5.28605469e-02 -2.86128491e-01
-4.68366921e-01 2.60802478e-01 -2.39234760e-01 -4.12513047e-01
2.87913561e-01 2.53038019e-01 -1.80799067e-02 -4.02541935e-01
2.41385505e-01 5.34727275e-01 5.40213525e-01 -4.48449224e-01
7.27799356e-01 9.01589990e-01 1.32601202e-01 -1.27948987e+00
-7.20266998e-01 -8.19944978e-01 -8.87594581e-01 -7.63231575e-01
1.14536417e+00 -7.38242328e-01 -4.05830622e-01 6.64631963e-01
-1.37739778e+00 -2.50844032e-01 -1.25465721e-01 5.24022996e-01
-4.31791544e-01 7.98331320e-01 -1.63856596e-01 -8.67398918e-01
4.17909324e-02 -1.43303919e+00 1.12104893e+00 2.98392355e-01
3.47322762e-01 -1.06773305e+00 1.07493788e-01 4.27730620e-01
2.55992144e-01 3.97327065e-01 2.64318317e-01 -8.50091130e-02
-7.76027441e-01 2.11886182e-01 -2.38218427e-01 3.09366316e-01
1.46378130e-01 1.02373958e-01 -1.06870496e+00 -1.64601773e-01
2.84570187e-01 3.26014042e-01 7.39877343e-01 8.04373741e-01
8.06474090e-01 1.45539224e-01 -1.80604964e-01 7.03656673e-01
1.67433918e+00 1.77116901e-01 4.97104049e-01 3.54799241e-01
7.09266067e-01 1.16339159e+00 8.15563619e-01 3.26670498e-01
2.73807466e-01 1.11668372e+00 4.45218384e-01 -2.36921787e-01
-9.73014981e-02 5.76159775e-01 3.29746872e-01 2.12483838e-01
-4.46180582e-01 -1.53893784e-01 -9.86502826e-01 5.08531213e-01
-1.81568050e+00 -1.08638656e+00 -6.94414556e-01 2.35656023e+00
5.20627439e-01 -3.80855948e-01 1.39402747e-01 1.11217732e-02
1.04004395e+00 1.13864891e-01 -1.81322590e-01 2.71046579e-01
-3.69561374e-01 -1.44686177e-01 6.87255979e-01 6.06463790e-01
-1.43434668e+00 5.76023698e-01 5.53414345e+00 6.21380508e-01
-1.11321652e+00 1.85985893e-01 6.00104451e-01 1.10211298e-01
-5.93225732e-02 1.68694496e-01 -5.90714216e-01 5.88303089e-01
5.00354528e-01 2.60523230e-01 4.60967749e-01 2.09137768e-01
4.40525562e-01 -8.00826669e-01 -7.83095419e-01 1.11210227e+00
1.48747370e-01 -1.14742947e+00 -3.17119539e-01 6.99817063e-03
6.45815670e-01 -1.74926054e-02 -2.40567297e-01 -7.70592570e-01
-1.27264321e-01 -3.68700415e-01 8.86483967e-01 6.96283519e-01
3.44764143e-01 -4.33941394e-01 5.51330924e-01 1.81681603e-01
-1.46282852e+00 4.39150557e-02 -2.03732727e-03 2.51041800e-01
4.92480844e-01 7.74648190e-01 -3.18702877e-01 9.71016705e-01
7.84344614e-01 8.34936678e-01 -6.29235685e-01 1.36698484e+00
8.63639042e-02 -1.01467094e-03 -3.53163242e-01 4.73889947e-01
-5.88372089e-02 -8.46404374e-01 7.98977554e-01 1.09356332e+00
4.42606837e-01 1.74587220e-01 3.82262826e-01 8.07286680e-01
6.04198873e-01 6.61387071e-02 -4.75544482e-01 4.00489300e-01
-6.14310894e-03 1.49956822e+00 -1.09729481e+00 -1.14562050e-01
-5.71792960e-01 1.06473613e+00 -2.05473915e-01 3.74522716e-01
-7.59978235e-01 1.39195248e-01 3.85458082e-01 2.13134900e-01
2.48659149e-01 -3.79217982e-01 8.04614574e-02 -1.39376128e+00
1.44075379e-01 -7.06273437e-01 4.04891908e-01 -9.39774334e-01
-9.66335833e-01 5.37113369e-01 3.75129163e-01 -1.47653961e+00
-1.67116120e-01 -4.57266599e-01 -3.62748861e-01 8.59388888e-01
-1.74276137e+00 -1.24673355e+00 -4.82647836e-01 6.95766270e-01
4.05305088e-01 3.31385612e-01 4.23992097e-01 5.18761098e-01
-4.71555829e-01 -3.61706078e-01 2.52066612e-01 -1.92656070e-01
5.94224036e-01 -9.38094914e-01 -3.46685350e-01 1.40411639e+00
-2.74758469e-02 7.50127658e-02 8.54228854e-01 -6.40576780e-01
-1.04915905e+00 -9.92894530e-01 5.85918963e-01 -1.03943869e-01
4.61921722e-01 9.41160917e-02 -8.64596903e-01 3.36840779e-01
1.81539342e-01 3.91640095e-03 5.74951649e-01 -5.25049031e-01
2.44194865e-01 -2.54271120e-01 -9.96889710e-01 1.47404119e-01
4.87704009e-01 -2.18914017e-01 -2.67074138e-01 2.29099900e-01
2.09936231e-01 -3.61455172e-01 -7.36549020e-01 3.96992326e-01
3.59682649e-01 -1.16114283e+00 1.08384013e+00 3.39599490e-01
-7.40557164e-02 -8.25602055e-01 -2.04707190e-01 -1.08415484e+00
2.07475685e-02 -3.88806880e-01 4.06053990e-01 1.27264559e+00
-1.18104376e-01 -5.27404547e-01 3.87001216e-01 4.67098445e-01
1.21488445e-01 9.14873257e-02 -8.49198937e-01 -6.55846834e-01
-5.92426419e-01 -3.13461512e-01 2.03503236e-01 8.16797674e-01
-7.12890446e-01 -1.12365313e-01 -3.63054633e-01 5.96442103e-01
1.08892941e+00 3.64695728e-01 6.18646502e-01 -1.51725852e+00
-4.19074386e-01 -3.21897089e-01 -3.19276839e-01 -7.27842629e-01
6.58582151e-02 -4.78983194e-01 1.68081462e-01 -1.29114377e+00
3.12040448e-01 -3.63810241e-01 2.08379090e-01 5.44091649e-02
-8.51842239e-02 5.55654109e-01 2.63416916e-01 3.17360699e-01
-4.51346815e-01 1.57291859e-01 9.53077018e-01 -1.35434270e-01
-2.43480101e-01 -7.87603203e-03 -1.20644867e-02 9.49999750e-01
5.83275616e-01 -4.04313952e-01 -1.30613759e-01 -5.45373917e-01
-7.03227706e-03 1.54490814e-01 7.96913922e-01 -1.08237171e+00
2.94732451e-01 -3.42465252e-01 2.83730924e-01 -4.97402817e-01
5.14186203e-01 -1.29737389e+00 6.99981332e-01 3.57223988e-01
9.80885625e-02 -2.26969510e-01 2.02039033e-01 5.45182526e-01
-1.94651484e-01 -6.30946517e-01 1.13095760e+00 -1.66383490e-01
-8.48558068e-01 1.57720163e-01 -5.73052704e-01 -3.78123611e-01
1.13025784e+00 -2.84154594e-01 1.03129186e-01 -2.31177822e-01
-7.29079306e-01 1.34755764e-02 7.81347394e-01 5.05542494e-02
4.27255958e-01 -9.51956451e-01 -6.14088297e-01 1.65671423e-01
-7.33192414e-02 -1.63976982e-01 4.96777654e-01 1.16811216e+00
-7.83611059e-01 7.41611347e-02 -3.09064686e-01 -9.49899256e-01
-1.65493047e+00 2.24589840e-01 7.75625348e-01 -2.08783284e-01
-2.92796493e-01 3.05957824e-01 4.95027192e-02 -9.44999605e-02
-3.81321311e-02 -2.90945590e-01 -5.24177194e-01 3.17106098e-01
5.89136422e-01 6.05975688e-01 1.51704386e-01 -1.11870086e+00
-4.34219420e-01 1.04298794e+00 3.94249737e-01 -3.36270332e-01
1.44489288e+00 -5.11553884e-01 -4.52356160e-01 3.38099718e-01
9.30543482e-01 -2.45941728e-02 -1.45039630e+00 -2.57105619e-01
-1.32037103e-01 -7.53678322e-01 3.55903804e-01 -3.51205915e-01
-1.13968050e+00 5.74847341e-01 8.78738582e-01 4.56328481e-01
1.54386556e+00 -1.41789481e-01 3.51806581e-01 -1.57540381e-01
3.04197907e-01 -1.08658147e+00 -2.09796622e-01 3.06800872e-01
5.31025410e-01 -1.41151607e+00 2.02958450e-01 -7.61246741e-01
-3.38034749e-01 1.45556104e+00 1.70097098e-01 2.09255844e-01
4.79292631e-01 4.95257899e-02 4.53705117e-02 -9.07329246e-02
-1.37565462e-02 -8.03311229e-01 4.48590010e-01 5.91649055e-01
3.01027536e-01 -2.65569627e-01 -4.52161700e-01 -9.98755693e-02
5.19297659e-01 -1.87653273e-01 5.67377508e-01 9.06321585e-01
-4.38765109e-01 -1.29523659e+00 -9.71363008e-01 -1.14367856e-02
-5.05915761e-01 1.18771419e-01 -6.43986389e-02 4.42118943e-01
3.80563140e-01 1.26201379e+00 1.29363507e-01 -3.46916467e-02
1.75608799e-01 -2.27844432e-01 6.16820812e-01 -3.50345701e-01
-4.87917900e-01 5.78967154e-01 -1.03201680e-02 -6.06733561e-01
-1.42433298e+00 -1.07677686e+00 -6.27822101e-01 -1.02914805e-02
-4.21806365e-01 -1.46123126e-01 7.98111022e-01 8.63089859e-01
-8.58575106e-02 3.49852175e-01 6.38963699e-01 -1.35643470e+00
7.66173154e-02 -4.20306981e-01 -5.59869289e-01 4.47573572e-01
3.08196247e-01 -6.63592637e-01 -2.07273498e-01 6.13482594e-01] | [9.861888885498047, -2.371594190597534] |
51fcc53e-a7cc-4acb-a7aa-275c8483e357 | satisfiability-aided-language-models-using | 2305.09656 | null | https://arxiv.org/abs/2305.09656v2 | https://arxiv.org/pdf/2305.09656v2.pdf | Satisfiability-Aided Language Models Using Declarative Prompting | Prior work has combined chain-of-thought prompting in large language models (LLMs) with programmatic representations to perform effective and transparent reasoning. While such an approach works very well for tasks that only require forward reasoning (e.g., straightforward arithmetic), it is less effective for constraint solving problems that require more sophisticated planning and search. In this paper, we propose a new satisfiability-aided language modeling (SATLM) approach for improving the reasoning capabilities of LLMs. We use an LLM to generate a declarative task specification rather than an imperative program and leverage an off-the-shelf automated theorem prover to derive the final answer. This approach has two key advantages. The declarative specification is closer to the problem description than the reasoning steps are, so the LLM can parse it out of the description more accurately. Furthermore, by offloading the actual reasoning task to an automated theorem prover, our approach can guarantee the correctness of the answer with respect to the parsed specification and avoid planning errors in the solving process. We evaluate SATLM on 6 different datasets and show that it consistently outperforms program-aided LMs in an imperative paradigm. In particular, SATLM outperforms program-aided LMs by 23% on a challenging subset of the GSM arithmetic reasoning dataset; SATLM also achieves a new SoTA on LSAT, surpassing previous models that are trained on the full training set. | ['Greg Durrett', 'Isil Dillig', 'Qiaochu Chen', 'Xi Ye'] | 2023-05-16 | null | null | null | null | ['arithmetic-reasoning'] | ['reasoning'] | [ 1.72674671e-01 7.17778087e-01 -1.20900363e-01 -3.33199829e-01
-1.06699932e+00 -7.97308087e-01 5.54068208e-01 3.07696313e-01
-2.75968798e-02 4.93401974e-01 2.01928541e-01 -1.17434001e+00
-1.17703145e-02 -1.10065269e+00 -9.82332826e-01 2.17041433e-01
1.52346805e-01 7.87213862e-01 4.18240815e-01 -2.65411764e-01
5.64935096e-02 3.95311899e-02 -1.24643111e+00 7.85332024e-01
1.04584551e+00 6.40443802e-01 -5.09687848e-02 8.54679406e-01
-4.15575296e-01 1.55147445e+00 -3.85142535e-01 -3.93973708e-01
2.49302104e-01 -2.95932025e-01 -1.20926619e+00 -2.83967048e-01
3.95767152e-01 -4.21628058e-01 4.86097559e-02 8.40440571e-01
-3.17682087e-01 -1.50347933e-01 3.14369351e-01 -1.38165402e+00
-9.55404714e-02 1.20593047e+00 -1.64686263e-01 -3.94902408e-01
8.37930143e-01 3.05712372e-01 1.20969939e+00 -4.34081942e-01
6.91785634e-01 1.43694031e+00 6.14059329e-01 5.66519380e-01
-1.60883546e+00 -3.66544008e-01 3.08195889e-01 1.40931994e-01
-1.41329098e+00 -4.41125691e-01 4.39675301e-01 -3.48864555e-01
1.64420223e+00 4.80375469e-01 4.64389235e-01 3.44684809e-01
4.84803855e-01 8.42575490e-01 9.39804316e-01 -7.22344816e-01
4.38137949e-01 2.11598605e-01 4.06749338e-01 9.01326239e-01
1.95540726e-01 -6.07964993e-02 -2.77123094e-01 -2.01200277e-01
3.83491367e-01 -4.63819742e-01 -5.95277548e-02 -4.52864051e-01
-1.13918006e+00 7.82972753e-01 5.34468703e-02 1.96914703e-01
-9.98689532e-02 4.34156239e-01 4.70316470e-01 5.29025555e-01
-6.72003925e-02 8.55498850e-01 -7.19052136e-01 -2.02150524e-01
-1.23696637e+00 7.49510229e-01 1.49598622e+00 1.08205354e+00
3.93948019e-01 -2.08052509e-02 -1.06862009e-01 9.87957343e-02
3.26470137e-01 4.19872701e-01 4.80923131e-02 -1.30906749e+00
7.14027107e-01 1.00105917e+00 1.14062563e-01 -7.54726350e-01
-3.82455498e-01 -2.52940625e-01 -1.08752333e-01 3.97595137e-01
6.43380165e-01 -1.54481813e-01 -5.66698015e-01 1.69646573e+00
1.73717439e-01 -1.83145493e-01 4.77211922e-01 5.99659920e-01
5.45074403e-01 9.91085827e-01 3.04836724e-02 -1.14388876e-01
1.29264843e+00 -1.15409946e+00 -3.49433422e-01 -5.84243476e-01
1.18497849e+00 -2.45718047e-01 9.60504830e-01 6.91116631e-01
-1.46785498e+00 -1.47384509e-01 -1.02994609e+00 -1.17275901e-01
2.56301314e-02 -1.90433636e-01 1.17964625e+00 1.60043761e-01
-1.22733545e+00 1.45064011e-01 -9.41353679e-01 4.16417234e-02
2.14921042e-01 3.84650409e-01 -1.90594181e-01 -4.76697028e-01
-8.95524621e-01 9.98361170e-01 7.42828786e-01 -1.23901039e-01
-8.19942176e-01 -1.02979290e+00 -1.25400960e+00 2.59513199e-01
9.92548287e-01 -7.12064981e-01 1.98949587e+00 -7.23083198e-01
-1.45599318e+00 5.19042671e-01 -3.37217718e-01 -7.44127870e-01
5.12998521e-01 -1.38256952e-01 -9.55554992e-02 1.37913348e-02
7.45085478e-02 5.33422887e-01 2.41615459e-01 -9.02561009e-01
-4.73499477e-01 -6.97552711e-02 9.50790942e-01 -5.19146286e-02
3.37325633e-01 1.26449525e-01 -3.89526695e-01 2.81617083e-02
2.09817171e-01 -7.81033933e-01 -2.69828677e-01 -6.43992871e-02
-3.33820760e-01 -3.53901982e-01 3.53060663e-01 -6.33058965e-01
1.35671413e+00 -1.98504293e+00 3.29891622e-01 1.90073222e-01
3.98057222e-01 2.54297704e-01 -6.16829516e-03 5.89909017e-01
-1.49274603e-01 1.46866143e-01 -1.63722873e-01 -6.92634983e-03
5.55370748e-01 2.55273163e-01 -6.88453794e-01 1.31968744e-02
4.49729174e-01 1.23875833e+00 -7.86780953e-01 -8.04143608e-01
1.73302099e-01 -2.81952560e-01 -1.30932844e+00 2.58793175e-01
-1.22570062e+00 -8.32249820e-02 -3.79701167e-01 5.96231341e-01
5.39021671e-01 -2.59059161e-01 5.38294554e-01 3.70831400e-01
-1.27413183e-01 7.99591780e-01 -1.35481358e+00 1.76832485e+00
-8.84278715e-01 4.41837817e-01 1.47982016e-01 -6.44625008e-01
4.84786242e-01 2.95925856e-01 -1.68998435e-01 -5.71086586e-01
-1.92536995e-01 2.70392507e-01 2.34373271e-01 -4.01656836e-01
6.56571463e-02 -3.59909445e-01 -4.13355768e-01 7.55534947e-01
-2.63575941e-01 -7.66698122e-01 4.64231551e-01 4.90329921e-01
1.29083884e+00 5.41185319e-01 5.57767272e-01 -1.84711829e-01
7.90603518e-01 6.56183064e-01 5.07094026e-01 8.13793123e-01
4.02953565e-01 -4.94822208e-03 1.03843701e+00 -5.67626953e-01
-6.98881567e-01 -7.57906318e-01 1.02536775e-01 9.28190708e-01
-2.61605263e-01 -1.08602428e+00 -7.96968639e-01 -6.76779568e-01
-3.93978097e-02 1.46268702e+00 3.47009338e-02 -1.72964945e-01
-8.07722270e-01 1.02047198e-01 7.93710589e-01 6.26221359e-01
4.42183137e-01 -9.67549205e-01 -8.56139362e-01 2.73782611e-01
-3.04248363e-01 -1.20689857e+00 6.19079918e-02 2.07431719e-01
-7.00338721e-01 -1.16423523e+00 4.45404410e-01 -5.15230417e-01
7.67846048e-01 -1.72419116e-01 1.22392750e+00 3.22835147e-01
-3.50885168e-02 2.00160921e-01 -1.85320586e-01 -3.82100135e-01
-8.24957788e-01 -6.03737943e-02 -3.71653497e-01 -7.10836232e-01
2.29242474e-01 -3.55477989e-01 3.66227776e-01 -1.88153952e-01
-7.37780631e-01 7.24182606e-01 3.91505808e-01 5.77325404e-01
3.11231077e-01 5.20554900e-01 1.17343618e-03 -1.09321892e+00
4.97084051e-01 -1.84047371e-01 -1.05073845e+00 3.81896526e-01
-5.67944348e-01 5.54631770e-01 1.09551096e+00 -1.49337143e-01
-1.00037062e+00 -2.96845508e-04 6.68318570e-02 2.51906030e-02
-5.64592704e-02 9.08667386e-01 -1.09120414e-01 1.68566763e-01
5.99165142e-01 2.46593773e-01 -1.40896797e-01 4.02719267e-02
2.60319322e-01 1.98184043e-01 5.87116003e-01 -1.38677633e+00
1.14076996e+00 -4.66312580e-02 1.98580906e-01 -1.19930863e-01
-8.87794137e-01 9.54286680e-02 -2.03629643e-01 2.84306586e-01
4.68941599e-01 -7.32538044e-01 -1.07264614e+00 3.27499397e-02
-1.36529124e+00 -9.50187087e-01 -3.45286906e-01 1.23864897e-01
-8.69058907e-01 6.76388815e-02 -5.86964548e-01 -1.13319123e+00
-3.21934193e-01 -1.28257751e+00 1.02160418e+00 -2.58135974e-01
-8.70039999e-01 -9.73201215e-01 -7.01574981e-02 3.94053847e-01
3.41302544e-01 1.64682209e-01 1.63099384e+00 -7.98368037e-01
-7.69853711e-01 -4.03094202e-01 -2.41809711e-01 2.05445275e-01
-3.68968219e-01 4.41998504e-02 -6.84560716e-01 1.35016233e-01
9.33538005e-02 -5.73067427e-01 2.41009429e-01 6.32727668e-02
9.87099528e-01 -6.88966870e-01 -1.95318952e-01 4.71033305e-01
1.27966070e+00 -6.60340190e-02 6.83473825e-01 3.93331498e-01
2.24612758e-01 4.64941144e-01 7.41416931e-01 8.69395509e-02
8.28813791e-01 5.11939824e-01 2.52007067e-01 3.43346655e-01
1.45432223e-02 -5.65872073e-01 5.48526764e-01 2.65607417e-01
1.38110623e-01 1.11557849e-01 -1.54007626e+00 2.33581945e-01
-2.08362079e+00 -8.85098159e-01 -2.95180857e-01 1.87140715e+00
1.33208191e+00 4.39624012e-01 -1.81469262e-01 3.54497552e-01
-1.88682839e-01 -1.22700468e-01 -2.33848497e-01 -8.77373874e-01
4.07458365e-01 6.28693223e-01 6.09368738e-03 1.09424555e+00
-7.28293955e-01 1.19408059e+00 6.22730017e+00 4.66917634e-01
-9.20101285e-01 -1.90944597e-01 -9.16361734e-02 -5.24737835e-02
-5.06307602e-01 5.34334362e-01 -6.43981993e-01 -2.00966135e-01
1.27171266e+00 -3.44140291e-01 8.35385025e-01 1.05824208e+00
6.41159266e-02 -4.26206321e-01 -1.83668053e+00 5.05805612e-01
-1.45623878e-01 -1.46635509e+00 -1.01410178e-02 -2.62220770e-01
3.13955873e-01 -4.68265295e-01 -2.77735770e-01 9.45987701e-01
4.72440243e-01 -1.26673031e+00 1.18837976e+00 4.00134861e-01
4.84346330e-01 -6.77207470e-01 8.11891675e-01 1.00057101e+00
-1.06125963e+00 -2.88924962e-01 1.37695968e-01 -8.60701025e-01
-5.54150604e-02 2.51956552e-01 -1.18167341e+00 5.95587850e-01
8.62199366e-02 2.81427681e-01 -6.17441833e-01 6.25366569e-01
-7.32055962e-01 4.58644420e-01 -4.04481679e-01 -1.61936563e-02
3.04872841e-01 2.19418064e-01 3.62592101e-01 1.22370398e+00
-2.09373161e-01 3.24002057e-01 3.72918427e-01 1.44965720e+00
4.01161730e-01 -2.86256284e-01 -4.03223604e-01 -4.91089523e-01
3.11216980e-01 9.92089808e-01 -2.80808479e-01 -5.54302037e-01
-3.44306678e-01 4.01323736e-01 4.59146172e-01 3.21425915e-01
-9.45587575e-01 -2.08372414e-01 2.79110581e-01 1.88485295e-01
5.34749813e-02 -5.77162266e-01 -4.24779475e-01 -1.18605840e+00
2.90179759e-01 -1.45366108e+00 3.17065954e-01 -1.03234637e+00
-4.95458186e-01 2.61759967e-01 4.56668615e-01 -4.77673203e-01
-7.19416976e-01 -6.56685889e-01 -6.21169567e-01 9.99373198e-01
-1.39692593e+00 -1.25380278e+00 -3.27867270e-02 2.96544611e-01
5.26885271e-01 2.84086287e-01 1.23037505e+00 -1.50965586e-01
-3.13286692e-01 3.37372959e-01 -9.61348295e-01 -5.97029403e-02
2.38980964e-01 -1.31002760e+00 2.95901269e-01 1.05136907e+00
-1.22528538e-01 1.15244651e+00 7.45861113e-01 -5.57137668e-01
-2.21968102e+00 -9.90649521e-01 1.14206672e+00 -6.70421183e-01
1.02527785e+00 -3.82763475e-01 -6.75328255e-01 1.24330080e+00
-1.26794308e-01 -3.18536103e-01 2.73111492e-01 1.92049965e-01
-6.46870792e-01 4.04965989e-02 -9.61798251e-01 7.45429158e-01
8.01166356e-01 -5.70470333e-01 -1.13442409e+00 4.39604521e-01
9.61577177e-01 -1.03202808e+00 -6.89447284e-01 8.84003118e-02
4.65314776e-01 -7.21787930e-01 7.15990841e-01 -8.18131983e-01
9.02945459e-01 -6.63759708e-01 -4.65555459e-01 -7.64287114e-01
-8.72368887e-02 -6.56311333e-01 -6.22808754e-01 9.87844288e-01
6.55908048e-01 -7.28483617e-01 5.12905240e-01 1.34275639e+00
-2.60899395e-01 -1.01574647e+00 -5.36804855e-01 -5.46829700e-01
1.71695635e-01 -1.20982277e+00 7.02633739e-01 4.28936094e-01
6.73872411e-01 3.57737333e-01 3.00813496e-01 2.53176540e-01
3.74135107e-01 5.46468556e-01 9.20758843e-01 -1.03659046e+00
-8.15373540e-01 -4.26689357e-01 -4.13525626e-02 -8.08551311e-01
7.77666152e-01 -1.41345620e+00 2.02582657e-01 -1.85300660e+00
2.58169085e-01 -4.59699363e-01 3.58861387e-01 1.10021734e+00
2.99518973e-01 -5.02299786e-01 3.37996989e-01 -1.45933896e-01
-7.93193102e-01 -9.70840082e-02 1.12588239e+00 -3.16477567e-01
-6.62511364e-02 6.22864701e-02 -9.33177352e-01 9.10898507e-01
4.58381325e-01 -2.33224049e-01 -4.92102504e-01 -5.17788529e-01
7.05976248e-01 5.44301808e-01 5.40005505e-01 -1.06075740e+00
5.54595411e-01 -6.56501830e-01 -2.78591692e-01 -1.19310930e-01
9.98016521e-02 -9.31494117e-01 2.48097360e-01 6.78337634e-01
-6.79051638e-01 -1.03749827e-01 6.49666190e-01 2.62553338e-02
-5.05374596e-02 -4.56460625e-01 3.76931727e-01 -1.75689265e-01
-6.06920838e-01 -3.11196029e-01 -4.48875874e-01 1.85240045e-01
1.02252674e+00 3.29184234e-02 -3.94235700e-01 -3.03949982e-01
-4.77562338e-01 5.39824009e-01 6.20765626e-01 -5.13008498e-02
5.41930318e-01 -7.75294244e-01 -4.27652895e-01 2.27264032e-01
1.05219401e-01 5.16310215e-01 -3.45460117e-01 1.08792615e+00
-6.99657142e-01 9.58105087e-01 1.86163425e-01 -3.19779038e-01
-1.05117583e+00 5.80159783e-01 3.51946950e-01 -6.42702818e-01
-6.91920221e-01 6.69682384e-01 1.61878332e-01 -7.37479687e-01
1.34288654e-01 -1.06655669e+00 4.07226503e-01 -6.30965054e-01
6.84114695e-01 -7.07217678e-02 8.49463716e-02 1.42677799e-01
-7.59990036e-01 1.87467098e-01 -2.55837645e-02 -1.60713360e-01
1.39936769e+00 5.72191477e-01 -6.55127883e-01 3.65590900e-01
4.35225427e-01 1.31955191e-01 -6.74276948e-01 -2.84338295e-01
1.64350465e-01 -1.06647149e-01 8.45469385e-02 -1.10264838e+00
-3.79959852e-01 6.24657571e-01 -7.33523309e-01 1.55572578e-01
8.77204537e-01 6.99523911e-02 5.41683257e-01 9.30684388e-01
7.95681179e-01 -6.90497458e-01 -1.29788965e-01 8.94019961e-01
8.54405940e-01 -8.86560082e-01 9.72472951e-02 -6.11042678e-01
-4.92653877e-01 1.32929921e+00 7.54917026e-01 -5.82048409e-02
-8.54228362e-02 9.07486737e-01 -4.18690085e-01 -1.54868081e-01
-1.35141623e+00 2.65895844e-01 -2.51036510e-02 3.38798642e-01
4.66937006e-01 9.50841457e-02 2.81665251e-02 1.00925744e+00
-6.30846620e-01 4.18395936e-01 5.83428264e-01 1.24440575e+00
-2.67601937e-01 -1.10332716e+00 -4.96807277e-01 2.34488547e-01
-1.88604876e-01 -3.02272588e-01 -4.21739578e-01 1.09335041e+00
-3.21792066e-02 1.03926599e+00 -2.27634326e-01 -2.01284766e-01
2.35908881e-01 4.29824620e-01 9.34037030e-01 -1.16496825e+00
-5.35499811e-01 -4.37192917e-01 7.95028985e-01 -9.62195635e-01
1.31282479e-01 -5.04146576e-01 -1.87972391e+00 -6.26152396e-01
-1.14703670e-01 2.22674668e-01 3.72409821e-01 1.24634874e+00
4.09589000e-02 7.81075537e-01 -8.56294408e-02 -4.74325627e-01
-9.31093812e-01 -3.58753562e-01 -3.64700332e-02 -1.59001499e-01
2.76043713e-01 -1.57052428e-01 -1.62133873e-01 7.87574053e-02] | [9.164511680603027, 7.227541923522949] |
ec50dd5a-d27a-49aa-b9e7-c8b899e9242f | unsupervised-content-based-image-retrieval-at | null | null | https://iopscience.iop.org/article/10.1088/1742-6596/1950/1/012059/meta | https://iopscience.iop.org/article/10.1088/1742-6596/1950/1/012059/pdf | Unsupervised Content based Image Retrieval at Different Precision Level by Combining Multiple Features | Image retrieval is a procedure of finding appropriate images in the image database. There are two types of image retrieval systems in common practice. These are the text-based image retrieval (TBIR) system and content-based image retrieval (CBIR) system. The content based system is proven to be more effective in which the visual contents of the images are extracted and described by multi-dimensional feature vectors. In this work, several models are developed by combining different image features in a combination of two and three. To begin with, three different models based on the combination of two features, viz., color with shape, shape with texture, and color with texture are designed. A three features based model is considered with color, shape, and texture in the next step. The retrieval rate of the mentioned models is assessed in terms of precisions. The results are obtained using COREL standard database. This study shows that the images can be better retrieved using three features based model in contrast to models using two features. | ['Mohd Atif Jamil', 'S. M. Zakariya'] | 2021-01-20 | null | null | null | icmai-2021-1 | ['content-based-image-retrieval'] | ['computer-vision'] | [ 8.58914107e-02 -8.55123758e-01 -1.24125093e-01 -8.13878849e-02
-6.39618993e-01 -4.28939134e-01 6.99027061e-01 4.30921972e-01
-6.36570752e-01 3.72601539e-01 3.07269454e-01 1.37539329e-02
-7.94949830e-01 -7.23545015e-01 8.18582550e-02 -6.89581275e-01
2.80570716e-01 1.85431197e-01 3.82623523e-01 -3.79417002e-01
9.03973818e-01 8.83344352e-01 -2.16406035e+00 2.42241278e-01
3.27999890e-01 8.57937038e-01 3.45338941e-01 6.61398828e-01
-4.93972093e-01 8.30307961e-01 -4.87325579e-01 1.60769075e-01
2.13637933e-01 -2.27401718e-01 -7.72637665e-01 3.14566821e-01
-1.33138701e-01 -3.82140368e-01 -1.24777704e-01 8.86756241e-01
5.77449262e-01 5.69877820e-03 9.91022468e-01 -1.04742861e+00
-8.82917821e-01 -2.37359166e-01 -6.95386767e-01 1.27201125e-01
6.91492140e-01 -4.01197016e-01 4.65425849e-01 -1.11019242e+00
6.72578156e-01 1.39404237e+00 2.06581075e-02 -2.29055006e-02
-6.04767799e-01 -4.11748648e-01 -4.48193491e-01 5.98357379e-01
-1.73600101e+00 -3.69529389e-02 8.11177671e-01 -5.97871661e-01
7.32698202e-01 6.38305664e-01 5.06854296e-01 -1.68647349e-01
4.81759429e-01 3.86653453e-01 1.45884669e+00 -1.11551702e+00
-1.28926858e-01 5.09994030e-01 4.33769792e-01 5.09554923e-01
3.71553481e-01 5.69681264e-02 -3.09337437e-01 -3.15621585e-01
5.95685005e-01 5.24969041e-01 1.55043509e-02 -2.57554710e-01
-8.80109131e-01 9.54191744e-01 3.46476108e-01 9.51197445e-01
-5.24541318e-01 -1.93665341e-01 4.55201387e-01 1.86569706e-01
2.08846554e-01 1.68934673e-01 -4.78747152e-02 3.16976428e-01
-9.12197769e-01 -5.37649868e-03 4.14480358e-01 7.55571246e-01
7.57523715e-01 -1.84881613e-01 -1.59599692e-01 1.11564851e+00
8.55641246e-01 6.82117343e-01 9.12509203e-01 -5.61062694e-01
-3.50954056e-01 8.50255668e-01 -3.02809943e-02 -1.39341426e+00
-7.23901838e-02 3.44188005e-01 -4.60943311e-01 3.48594338e-01
-1.68163240e-01 4.67963994e-01 -1.31001437e+00 7.82810509e-01
1.07743055e-01 -5.44915736e-01 2.03333631e-01 7.63352692e-01
1.36872077e+00 8.94262433e-01 1.41898483e-01 -7.78569356e-02
1.73928833e+00 -7.19437301e-01 -9.23369169e-01 3.61406237e-01
3.44307333e-01 -1.43330479e+00 6.16443515e-01 3.27886581e-01
-9.05207455e-01 -5.24274766e-01 -1.00580490e+00 5.02375290e-02
-1.10063028e+00 4.20409203e-01 2.77915746e-01 6.66980207e-01
-1.33555114e+00 5.88784106e-02 -7.37929568e-02 -7.17733204e-01
-3.67424965e-01 2.63856828e-01 -5.81606328e-01 -2.66990542e-01
-8.98400068e-01 1.10128021e+00 4.08072799e-01 -1.67701006e-01
-1.71354920e-01 1.01219893e-01 -6.36311948e-01 -1.99345887e-01
-5.19151390e-02 -1.97488680e-01 5.51125050e-01 -9.92278397e-01
-1.19877839e+00 1.07741070e+00 -1.92246869e-01 1.99671924e-01
1.86327666e-01 2.14722842e-01 -4.06883985e-01 8.65519464e-01
9.49417576e-02 5.12626231e-01 6.13917828e-01 -1.61880362e+00
-5.91593027e-01 -3.85862708e-01 -1.02753043e-01 3.87382388e-01
-1.69750124e-01 6.94217920e-01 -7.42461205e-01 -3.44564438e-01
3.61686945e-01 -8.32310140e-01 5.42351119e-02 7.26833940e-02
-5.36238262e-03 -2.74036109e-01 9.56760824e-01 -7.51381695e-01
1.10634983e+00 -2.10148835e+00 -3.81375998e-01 7.43061304e-01
-3.48690711e-02 6.04037523e-01 -7.98011273e-02 8.56358707e-01
3.79640386e-02 4.92504746e-01 4.17324841e-01 3.13470840e-01
-1.85873479e-01 1.24891676e-01 3.62055421e-01 2.29740053e-01
-1.12549715e-01 4.11447167e-01 -4.73085910e-01 -1.08455753e+00
6.96276307e-01 7.69991577e-01 1.20061450e-01 -7.57603422e-02
5.28419197e-01 -1.01092041e-01 -7.90484667e-01 9.59739625e-01
7.76334882e-01 -7.13060349e-02 2.32710671e-02 -2.11130664e-01
-2.56999880e-01 -4.55725580e-01 -1.20740509e+00 8.37464571e-01
-2.73778260e-01 4.71251339e-01 -2.92673171e-01 -6.40917242e-01
1.18951344e+00 6.83106363e-01 6.35814667e-01 -1.11866105e+00
3.11882406e-01 3.44361693e-01 -6.84035718e-02 -9.12903666e-01
8.14089775e-01 1.62902042e-01 4.39017683e-01 3.72105986e-01
-1.45238683e-01 -1.20322309e-01 4.86972690e-01 1.21734917e-01
5.23260772e-01 -1.36792794e-01 4.30020839e-01 -3.35332572e-01
8.57910156e-01 1.61301181e-01 -1.66613758e-01 5.42317986e-01
-1.01180896e-01 5.36095440e-01 -6.39885962e-02 -3.38099390e-01
-1.12217951e+00 -5.66378117e-01 -1.96612358e-01 6.04572415e-01
4.28172022e-01 -1.95440918e-01 -2.71154910e-01 3.84659357e-02
-1.44086316e-01 5.49846627e-02 -4.23655093e-01 9.12651643e-02
-7.99860954e-02 -5.31161547e-01 2.35904735e-02 -5.59861735e-02
7.74407446e-01 -1.21998680e+00 -7.50822067e-01 -2.12044030e-01
8.67752135e-02 -4.33099240e-01 -1.48658246e-01 -3.83002937e-01
-8.32336664e-01 -1.09438264e+00 -1.06668496e+00 -9.39186871e-01
8.75904202e-01 8.89970064e-01 8.08284223e-01 9.63969767e-01
-8.05839658e-01 8.85828078e-01 -1.06606579e+00 -4.55547690e-01
-3.56672555e-01 -2.67544627e-01 -4.70135838e-01 -1.37864828e-01
5.49369931e-01 1.63614884e-01 -7.91837215e-01 1.66902438e-01
-1.56651449e+00 -1.89866960e-01 9.68089879e-01 6.11913204e-01
7.14861572e-01 2.92465925e-01 3.06549162e-01 -6.69691324e-01
8.02578568e-01 -3.20164204e-01 -3.54390532e-01 7.37140059e-01
-8.96567523e-01 -1.61336705e-01 -3.86528932e-02 -2.34257966e-01
-8.21739137e-01 -1.06917722e-02 6.13578409e-02 -1.81876034e-01
-2.00220227e-01 8.25732529e-01 2.94642478e-01 -5.53288341e-01
4.18288410e-01 6.21227264e-01 3.57614785e-01 -3.58193427e-01
-1.42331854e-01 1.06994545e+00 -2.54688948e-01 -3.30145806e-01
5.76625764e-01 1.41688570e-01 1.54623657e-01 -1.06787694e+00
-2.03900725e-01 -1.14692163e+00 -4.69605833e-01 -6.39304340e-01
9.31538105e-01 -6.49866402e-01 -3.85281712e-01 5.10831654e-01
-7.26876557e-01 5.88092446e-01 2.74625808e-01 7.21328795e-01
-1.27561435e-01 5.83737433e-01 -4.19864684e-01 -1.03780186e+00
-7.12623358e-01 -1.33743036e+00 9.11176682e-01 4.72677678e-01
2.88452268e-01 -7.40294337e-01 -1.55872747e-01 2.86849529e-01
7.67668188e-01 1.97865382e-01 1.03494358e+00 -5.77615201e-01
-3.80577117e-01 -7.47565806e-01 -5.08037746e-01 2.08591625e-01
3.75650495e-01 4.66435134e-01 -4.65512007e-01 -1.66657120e-01
-1.53570309e-01 -4.12753299e-02 4.82125282e-01 3.47968280e-01
7.83014774e-01 -4.69379053e-02 -2.64391303e-01 -1.69124052e-01
2.32807922e+00 1.02475202e+00 1.12005150e+00 7.16185749e-01
1.13944612e-01 5.60981214e-01 9.40391183e-01 2.74827898e-01
-4.92938645e-02 4.43713605e-01 6.50009289e-02 -2.46459693e-01
-1.83337137e-01 2.64209658e-01 -2.03340992e-01 7.64519751e-01
-1.46123976e-01 -1.24862261e-01 -9.49846148e-01 4.92085487e-01
-1.53546500e+00 -9.48665023e-01 -1.56141981e-01 2.22907400e+00
4.29891318e-01 -3.00636441e-01 6.57037273e-02 4.27348614e-01
6.52776122e-01 -1.98144302e-01 1.05796948e-01 -6.40008211e-01
5.44436797e-02 2.31703579e-01 5.56131661e-01 2.91583270e-01
-8.73035908e-01 5.20874560e-01 6.72466707e+00 8.55488002e-01
-1.31821084e+00 -2.29576841e-01 4.89200532e-01 4.73844618e-01
-1.71888158e-01 1.21673778e-01 -5.42967141e-01 4.69726026e-01
5.60175419e-01 -3.48368913e-01 2.34056517e-01 5.20010114e-01
2.11200267e-01 -7.97345817e-01 -1.81515917e-01 1.12185109e+00
4.28196132e-01 -1.04278100e+00 6.08551502e-01 1.06179290e-01
3.18201482e-01 -3.72263044e-01 2.90515274e-03 -1.66289717e-01
-2.31469497e-01 -8.66542101e-01 5.20574808e-01 9.49113131e-01
5.95844090e-01 -7.55153358e-01 1.08274400e+00 -8.99809673e-02
-1.14899147e+00 1.60395235e-01 -4.27638978e-01 3.55276674e-01
-3.25122714e-01 -1.72780398e-02 -3.66240025e-01 8.47573638e-01
8.20024967e-01 2.71933913e-01 -9.24103558e-01 1.65436387e+00
3.54328334e-01 1.62756890e-01 -1.74885601e-01 -3.53945643e-01
1.93449914e-01 -4.66151506e-01 1.88353434e-01 1.00511479e+00
5.36832154e-01 2.60346085e-01 -1.96913704e-02 4.80354130e-01
4.45897222e-01 8.44388187e-01 -8.10615957e-01 3.52057442e-02
5.76128304e-01 1.38011408e+00 -1.05299914e+00 -5.47819078e-01
-4.37991589e-01 6.35555804e-01 -5.99481821e-01 1.53259128e-01
-2.35767514e-01 -7.62001097e-01 -3.19857776e-01 1.21372163e-01
7.90018514e-02 -1.18272662e-01 2.81280845e-01 -5.74361920e-01
-2.22193092e-01 -7.95014918e-01 2.99401462e-01 -1.22797942e+00
-1.03613222e+00 7.31129766e-01 6.18322968e-01 -1.42694366e+00
-3.60846147e-02 -6.14664435e-01 -2.75405705e-01 1.17610765e+00
-1.47933483e+00 -1.31197679e+00 -2.79993594e-01 7.17472970e-01
4.43928808e-01 -3.29412520e-01 9.19682443e-01 4.37060565e-01
-2.70643115e-01 1.18321441e-01 4.49490190e-01 1.31046131e-01
5.62156200e-01 -9.39817309e-01 -8.57412577e-01 5.35006702e-01
-2.01598063e-01 6.37276411e-01 5.15723944e-01 -5.00186563e-01
-1.44907200e+00 -3.87455434e-01 9.38233674e-01 2.75809556e-01
1.70439959e-01 4.93477434e-01 -5.97864330e-01 9.74211022e-02
3.65350693e-01 -2.78964549e-01 7.72755682e-01 -4.89212036e-01
-1.31502137e-01 -3.18894237e-02 -1.49221170e+00 3.78359526e-01
-1.70917928e-01 -3.71702343e-01 -3.85157317e-01 2.69336820e-01
9.10009593e-02 9.94901434e-02 -1.09869671e+00 1.47969782e-01
8.08739603e-01 -7.55271256e-01 1.19289911e+00 1.23800434e-01
1.82310149e-01 -6.22568429e-01 -2.83387065e-01 -6.48337960e-01
-2.73000300e-01 2.35733524e-01 9.78711367e-01 1.24787748e+00
3.06759775e-01 -6.65931880e-01 2.05897629e-01 6.58257306e-01
4.54908401e-01 -3.50723594e-01 -4.21522588e-01 -6.94402993e-01
-9.10791382e-02 4.09537792e-01 3.88324380e-01 5.57157874e-01
-1.07122451e-01 1.30519629e-01 -1.44290805e-01 -3.24391484e-01
3.23005438e-01 1.30518049e-01 3.87280911e-01 -1.38440597e+00
3.44805419e-01 -3.99906635e-01 -7.55100787e-01 -1.00454807e-01
-4.57147956e-01 -5.75724721e-01 -2.04907462e-01 -1.99823844e+00
7.26868927e-01 -4.42811638e-01 -2.69759089e-01 3.19611192e-01
2.27414705e-02 4.82203454e-01 5.17653883e-01 7.16680586e-01
-3.51270705e-01 -1.14780329e-01 1.18318069e+00 -8.80546346e-02
-8.08944106e-02 -2.82190859e-01 -4.98161793e-01 3.97105813e-01
8.95051122e-01 -3.95812988e-01 -4.41404909e-01 -1.43346652e-01
1.53948799e-01 8.42797309e-02 1.59642905e-01 -8.03759038e-01
1.81416437e-01 -1.27503753e-01 5.23193836e-01 -9.28006947e-01
4.49750215e-01 -1.13262129e+00 5.35027802e-01 5.47333300e-01
-2.66460896e-01 4.73915875e-01 9.29906368e-02 3.21569651e-01
-6.30553961e-01 -7.78369009e-01 7.30699301e-01 -4.14491087e-01
-1.11425531e+00 -2.48596147e-02 -6.56941354e-01 -8.55374396e-01
1.08404136e+00 -7.84458995e-01 -2.54070550e-01 -5.38147628e-01
-6.88489914e-01 -1.75890148e-01 4.05195236e-01 3.20855469e-01
1.02593827e+00 -1.45650077e+00 -5.56043863e-01 -1.00918829e-01
4.33963507e-01 -7.61400342e-01 3.29417855e-01 7.60665834e-01
-1.40951693e+00 6.38147950e-01 -6.06932402e-01 -3.71764541e-01
-1.99251819e+00 6.97115600e-01 1.74215868e-01 -1.83470726e-01
-2.69275099e-01 -1.20857079e-02 -2.58280486e-01 8.97301733e-03
-2.77230237e-02 2.79966623e-01 -1.08759069e+00 1.70785144e-01
7.38238931e-01 2.49093309e-01 1.27737850e-01 -1.30859649e+00
-3.95196617e-01 1.27264106e+00 -1.50195807e-01 -2.79262751e-01
1.06443131e+00 -3.46480370e-01 -5.80877662e-01 2.76240706e-01
1.33602405e+00 -8.46654400e-02 1.72715280e-02 -3.00045133e-01
1.35511279e-01 -8.12243998e-01 2.33742967e-01 -9.51827168e-01
-1.09819984e+00 4.74447548e-01 1.24126840e+00 5.09026945e-01
1.42212260e+00 -3.66420656e-01 2.76186675e-01 1.45485938e-01
2.91933566e-01 -1.17929995e+00 5.65869920e-03 1.90500423e-01
9.61744308e-01 -1.26533031e+00 3.14879954e-01 -3.13755095e-01
-4.60870534e-01 1.34944427e+00 6.87459484e-02 -2.22139031e-01
1.21024632e+00 -1.31695390e-01 1.96304873e-01 -4.51527327e-01
-5.41500270e-01 -6.24006450e-01 5.48477113e-01 4.82847393e-01
7.09100425e-01 -6.47793263e-02 -1.26536775e+00 -1.63331062e-01
3.80366236e-01 1.23402260e-01 4.39277470e-01 1.49441755e+00
-7.97028542e-01 -1.33325922e+00 -8.65182817e-01 5.94753265e-01
-7.80930817e-01 -8.91750976e-02 -4.75501150e-01 1.01915777e+00
-1.34322196e-01 1.42387938e+00 -2.25411281e-01 -2.53992647e-01
1.33954212e-01 -7.78123140e-02 4.24413383e-01 -3.25517476e-01
-5.56766987e-01 2.56038636e-01 -2.76353806e-01 -9.44270119e-02
-9.52916026e-01 -3.51506114e-01 -9.40715492e-01 -3.34951311e-01
-6.11848295e-01 2.82592535e-01 1.26286018e+00 7.64703035e-01
2.25543886e-01 1.17131710e-01 5.43845236e-01 -6.56053722e-01
9.62365139e-03 -9.14802432e-01 -7.98619688e-01 3.83267701e-01
-1.19686816e-02 -5.85420191e-01 -1.86522737e-01 1.17068946e-01] | [10.763493537902832, 0.03575511276721954] |
f384b243-35cc-4fe4-80d9-5ae7ee1bc9b0 | dfcanet-dense-feature-calibration-attention | 2111.00919 | null | https://arxiv.org/abs/2111.00919v1 | https://arxiv.org/pdf/2111.00919v1.pdf | DFCANet: Dense Feature Calibration-Attention Guided Network for Cross Domain Iris Presentation Attack Detection | An iris presentation attack detection (IPAD) is essential for securing personal identity is widely used iris recognition systems. However, the existing IPAD algorithms do not generalize well to unseen and cross-domain scenarios because of capture in unconstrained environments and high visual correlation amongst bonafide and attack samples. These similarities in intricate textural and morphological patterns of iris ocular images contribute further to performance degradation. To alleviate these shortcomings, this paper proposes DFCANet: Dense Feature Calibration and Attention Guided Network which calibrates the locally spread iris patterns with the globally located ones. Uplifting advantages from feature calibration convolution and residual learning, DFCANet generates domain-specific iris feature representations. Since some channels in the calibrated feature maps contain more prominent information, we capitalize discriminative feature learning across the channels through the channel attention mechanism. In order to intensify the challenge for our proposed model, we make DFCANet operate over nonsegmented and non-normalized ocular iris images. Extensive experimentation conducted over challenging cross-domain and intra-domain scenarios highlights consistent outperforming results. Compared to state-of-the-art methods, DFCANet achieves significant gains in performance for the benchmark IIITD CLI, IIIT CSD and NDCLD13 databases respectively. Further, a novel incremental learning-based methodology has been introduced so as to overcome disentangled iris-data characteristics and data scarcity. This paper also pursues the challenging scenario that considers soft-lens under the attack category with evaluation performed under various cross-domain protocols. The code will be made publicly available. | ['Raghavendra Ramachandra', 'Sumantra Dutta Roy', 'Aman Verma', 'Gaurav Jaswal'] | 2021-11-01 | null | null | null | null | ['cross-domain-iris-presentation-attack'] | ['computer-vision'] | [ 3.92272174e-01 -3.04871172e-01 -3.63027304e-01 -2.17123806e-01
-4.43162650e-01 -6.31805301e-01 3.43141854e-01 -2.38512650e-01
-1.55299917e-01 6.55833840e-01 2.79236078e-01 -3.60742927e-01
-6.18130982e-01 -2.52186209e-01 -5.60758889e-01 -7.75653064e-01
-1.40637487e-01 3.89309600e-02 -5.31049728e-01 1.25195011e-01
4.05858010e-01 6.25992835e-01 -1.52631557e+00 2.28037342e-01
9.51676011e-01 7.59298325e-01 -6.89363301e-01 7.92226136e-01
2.42764100e-01 1.95993572e-01 -5.57459295e-01 -5.52146018e-01
8.46873045e-01 -4.85985518e-01 -6.15542710e-01 4.45248127e-01
1.13730168e+00 -4.56439644e-01 -2.44998887e-01 1.04906547e+00
9.86468136e-01 -3.07004929e-01 5.10933936e-01 -1.12688303e+00
-7.63850808e-01 5.53286113e-02 -1.08871710e+00 2.84391195e-01
3.67365628e-01 5.76670289e-01 6.08696759e-01 -4.62783366e-01
3.89890254e-01 9.23011303e-01 5.72496533e-01 5.60172677e-01
-1.19431317e+00 -9.62943077e-01 -1.14880756e-01 -1.36012346e-01
-1.36178243e+00 -2.07697228e-01 5.62440813e-01 -3.52841258e-01
6.30234659e-01 2.45936111e-01 4.39332068e-01 1.08539581e+00
-4.73068990e-02 5.38189232e-01 1.84624279e+00 -5.23270488e-01
-5.62625468e-01 3.75098944e-01 1.60807192e-01 5.34082055e-01
4.59189683e-01 7.86359966e-01 -4.82694596e-01 6.94547407e-03
9.28436100e-01 -1.27588496e-01 -4.80187416e-01 -2.72970885e-01
-1.12052166e+00 4.08877879e-01 3.02739650e-01 2.50010252e-01
-2.60353118e-01 -5.41583657e-01 2.81387031e-01 4.70992565e-01
-9.35502201e-02 5.24514675e-01 -4.24020737e-01 -1.47164449e-01
-6.80933535e-01 1.36380559e-02 4.57240343e-01 8.09803665e-01
5.04190981e-01 -2.92838681e-02 -3.65361959e-01 5.83506644e-01
2.13139623e-01 5.60802400e-01 2.72532731e-01 -2.75426537e-01
3.87295604e-01 7.87330270e-01 7.94141889e-02 -9.08827484e-01
-4.02413756e-01 -1.00849175e+00 -8.34613144e-01 4.66076136e-01
5.50337970e-01 -4.35459316e-01 -1.17852354e+00 1.34920585e+00
4.42210287e-01 7.55172610e-01 3.28667521e-01 1.14307940e+00
7.73375511e-01 -7.24420324e-02 -7.40268677e-02 -3.95357870e-02
1.35659623e+00 -6.01529717e-01 -4.41813648e-01 3.68352026e-01
1.78961039e-01 -1.25701320e+00 7.39562869e-01 5.95781326e-01
-7.18892455e-01 -7.33585238e-01 -1.02668726e+00 9.56405476e-02
-2.69925654e-01 3.88819933e-01 6.27094328e-01 1.16883886e+00
-9.40102100e-01 4.40712459e-02 -3.30410689e-01 -5.06188095e-01
7.22541153e-01 1.02229440e+00 -6.06971860e-01 -4.45369035e-02
-8.21238458e-01 4.69516695e-01 2.46051833e-01 4.71739992e-02
-5.33034742e-01 -6.53988302e-01 -5.18395066e-01 -3.84818077e-01
7.45946094e-02 -6.92987680e-01 5.86876690e-01 -1.11078823e+00
-1.61920917e+00 1.13693726e+00 -8.42169002e-02 -5.86178482e-01
2.12035552e-01 -5.94013045e-03 -7.46016860e-01 -2.12236606e-02
-3.48691732e-01 2.98781961e-01 1.22985601e+00 -1.11309838e+00
-6.99824870e-01 -5.73469520e-01 7.83406943e-03 2.91728646e-01
-2.38153234e-01 1.77335963e-01 -5.56015134e-01 -6.45846903e-01
-7.74856433e-02 -9.70291257e-01 1.97362170e-01 -1.97826967e-01
-7.48490334e-01 6.52680099e-02 7.68539131e-01 -4.89722729e-01
1.09718764e+00 -2.33823800e+00 -1.06598020e-01 4.94082689e-01
2.71909207e-01 8.96739125e-01 -2.07844362e-01 5.48221963e-03
-3.44300002e-01 -9.89232212e-02 5.56974970e-02 -8.66075456e-02
-1.31323189e-01 4.07812782e-02 -1.48284912e-01 7.57118225e-01
3.11755985e-01 5.76257348e-01 -4.36897725e-01 -3.72979671e-01
4.14999664e-01 7.10885227e-01 -5.66805482e-01 1.26838654e-01
2.75404692e-01 8.35362613e-01 -4.02594686e-01 1.09317577e+00
1.26127195e+00 -2.25204080e-01 -9.54919383e-02 -3.75051230e-01
-1.19931497e-01 -3.25300813e-01 -1.31051791e+00 1.57822061e+00
-1.29368186e-01 4.08086658e-01 -1.73642542e-02 -7.35776424e-01
8.56078506e-01 4.20031816e-01 2.82924831e-01 -6.27297699e-01
3.85840029e-01 2.40397632e-01 2.94428289e-01 -6.99127316e-01
2.50078648e-01 -2.77130026e-02 3.89800489e-01 3.26253265e-01
1.33456588e-01 3.83162439e-01 -3.29756677e-01 -1.94420621e-01
4.48085725e-01 2.14754995e-02 3.27408612e-01 1.93180963e-02
8.77003253e-01 -2.52904087e-01 3.77450705e-01 6.05040193e-01
-5.38628697e-01 6.34278357e-01 4.85321671e-01 -4.44598049e-01
-7.03481555e-01 -9.12247062e-01 -8.93700123e-01 4.03432548e-01
2.51146495e-01 -2.40527187e-02 -6.33788168e-01 -8.57322693e-01
2.98737586e-01 -1.22602768e-01 -8.23599041e-01 3.42303127e-01
-6.72723576e-02 -9.94870365e-01 7.21172333e-01 8.66960883e-02
7.37455845e-01 -5.13177693e-01 -6.47478551e-02 -2.24331215e-01
4.24087763e-01 -9.27538991e-01 -4.79111016e-01 -4.02241975e-01
-4.66628253e-01 -1.61138928e+00 -7.08409011e-01 -7.83686459e-01
9.07069623e-01 1.41218096e-01 7.21462250e-01 7.92739391e-02
-7.48581648e-01 4.99091715e-01 -2.73939997e-01 -6.50654554e-01
1.18444517e-01 -4.33863848e-02 1.51344508e-01 9.12984312e-01
1.12911928e+00 -3.14408302e-01 -8.64303410e-01 3.17698956e-01
-5.00964761e-01 -3.84941041e-01 8.27451468e-01 1.13435423e+00
5.69930375e-01 1.76193401e-01 2.61307627e-01 -8.57705891e-01
6.80282176e-01 -4.15294409e-01 -9.29193258e-01 1.32415861e-01
-9.18516934e-01 -2.09034041e-01 2.14432374e-01 -4.89134431e-01
-1.03812408e+00 4.59981412e-02 3.56245339e-01 -5.19595921e-01
-4.59901243e-01 3.91088612e-02 8.60857219e-02 -8.05995822e-01
8.16219091e-01 2.62228429e-01 3.33758771e-01 -4.63595867e-01
-8.11450705e-02 1.16724670e+00 6.54246867e-01 -6.10903323e-01
1.12144697e+00 3.67494971e-01 1.60314336e-01 -6.20475888e-01
-5.06552815e-01 -5.94017804e-01 -4.89179432e-01 6.33544177e-02
7.80014217e-01 -1.04971480e+00 -1.17868674e+00 9.02961612e-01
-7.08872557e-01 3.03852409e-01 1.15742117e-01 8.76954496e-01
-8.40828717e-02 4.08663392e-01 -2.84242839e-01 -6.69328451e-01
-2.62147397e-01 -1.32087278e+00 9.08738315e-01 8.93454850e-01
1.91479668e-01 -8.46895754e-01 9.72188786e-02 6.90372646e-01
4.89485681e-01 4.04657245e-01 5.81382632e-01 -7.28092372e-01
-8.23579431e-01 -3.31172317e-01 -6.25569880e-01 4.89606172e-01
3.46310347e-01 8.05795640e-02 -1.29494584e+00 -6.21740043e-01
-3.01730186e-01 -2.50595659e-01 3.57876897e-01 5.13154209e-01
1.08198333e+00 -3.89050283e-02 -1.82663977e-01 1.27160704e+00
1.65947032e+00 1.30957678e-01 7.43326187e-01 3.64482433e-01
4.64377671e-01 5.17987013e-01 4.08076227e-01 4.76786852e-01
1.52295930e-02 5.92162192e-01 3.17096680e-01 -7.18235850e-01
-4.22368258e-01 -1.03904650e-01 -2.48023897e-01 7.05103055e-02
-3.29858959e-01 -1.52068943e-01 -7.35973537e-01 5.74217379e-01
-1.24292374e+00 -6.90511882e-01 -8.06190297e-02 2.41414189e+00
9.07629788e-01 -2.37699479e-01 3.68304253e-02 2.40509608e-03
6.69134140e-01 -1.12596437e-01 -7.26942718e-01 -4.06473756e-01
-5.69419265e-01 7.70208776e-01 6.27735496e-01 3.69528472e-01
-1.32198715e+00 7.62266695e-01 5.57428503e+00 5.76815069e-01
-1.51641035e+00 -2.01935440e-01 5.64681172e-01 -1.24238797e-01
2.39015892e-01 -2.28834882e-01 -9.37689483e-01 3.65598619e-01
6.84893012e-01 1.99017495e-01 4.69701588e-01 1.94692910e-01
-1.25204518e-01 3.18952799e-02 -7.16951251e-01 1.37523961e+00
2.29389548e-01 -1.19383550e+00 -1.44882873e-01 4.84059781e-01
9.93654847e-01 -2.39907410e-02 8.74532402e-01 -2.35151961e-01
4.02918793e-02 -1.40823615e+00 -4.25964177e-01 6.79576457e-01
1.35505176e+00 -7.12617695e-01 9.46540952e-01 -4.14162308e-01
-7.13246465e-01 -1.83052212e-01 -1.79957122e-01 1.83605656e-01
-6.16555452e-01 2.45157052e-02 -1.19109070e+00 7.39624798e-01
5.54562211e-01 9.65387881e-01 -7.01778293e-01 1.36587334e+00
9.11369175e-03 5.63732803e-01 -2.49766886e-01 4.08329636e-01
6.24680519e-02 -1.51261687e-01 7.69072175e-01 8.86430383e-01
1.52040318e-01 5.70847699e-03 -1.84597433e-01 6.47093356e-01
3.41722108e-02 2.78774232e-01 -5.14578104e-01 1.99406184e-02
2.66948074e-01 9.94666457e-01 5.13524525e-02 1.31944060e-01
-7.34618247e-01 7.60849118e-01 -2.45703086e-01 6.29869640e-01
-5.41933954e-01 -5.06035268e-01 1.07332253e+00 1.43255249e-01
2.32929140e-01 2.01055914e-01 -2.69811064e-01 -1.12699509e+00
-7.91535005e-02 -1.45015311e+00 4.31976408e-01 -4.05733615e-01
-1.41342318e+00 7.51800418e-01 -3.96373808e-01 -1.79738796e+00
1.40074953e-01 -8.30489933e-01 -3.99915099e-01 1.45964921e+00
-1.96634293e+00 -1.71137881e+00 -3.56011152e-01 1.29959476e+00
-7.44491667e-02 -8.28277767e-01 9.24276412e-01 4.71925467e-01
-8.15210879e-01 1.39980769e+00 4.36655469e-02 4.02388662e-01
1.27516055e+00 -1.16075563e+00 -1.40525149e-02 1.04754925e+00
-3.71673815e-02 9.65075314e-01 3.97284240e-01 -4.20826912e-01
-1.49423683e+00 -8.44674110e-01 4.85225230e-01 -4.92004246e-01
3.52045894e-01 7.17168003e-02 -6.58867896e-01 5.82126737e-01
4.40037161e-01 3.81969005e-01 9.33830142e-01 2.84286797e-01
-5.39396286e-01 -3.24778259e-01 -1.26383543e+00 2.68135309e-01
5.99238157e-01 -7.02950180e-01 -2.31706262e-01 2.41558924e-01
1.50368154e-01 -7.95948088e-01 -1.06720459e+00 3.40834320e-01
5.05463719e-01 -1.20735371e+00 9.25619900e-01 -8.29931021e-01
-3.87091748e-02 -6.40448749e-01 1.61674749e-02 -7.31997550e-01
1.79125562e-01 -1.22510493e+00 1.21065371e-01 1.34106052e+00
3.68080527e-01 -1.01669490e+00 9.23861921e-01 4.70212251e-01
2.81555831e-01 -5.67107141e-01 -8.70791495e-01 -3.64417911e-01
-5.42134382e-02 3.98866385e-02 9.24751282e-01 1.08915639e+00
-2.90406048e-01 -2.41894975e-01 -6.52130067e-01 8.65224838e-01
9.95608747e-01 2.23087743e-01 1.20561445e+00 -1.18378270e+00
-3.39850843e-01 -1.09792441e-01 -9.24657106e-01 -7.51745164e-01
-2.08049297e-01 -5.58519423e-01 -8.01702499e-01 -4.54455316e-01
1.39663890e-01 -5.94876170e-01 -5.30112863e-01 5.54879189e-01
-1.30317494e-01 6.00926816e-01 -7.55571797e-02 1.01974763e-01
-1.37803614e-01 -1.68408960e-01 1.55899775e+00 -1.25711024e-01
-3.07147801e-01 2.98438162e-01 -9.38692808e-01 7.66962543e-02
7.17352927e-01 -4.72831316e-02 -5.54981351e-01 -1.93804204e-01
-2.47722864e-01 -4.60635405e-03 4.92351592e-01 -1.01655650e+00
4.16612118e-01 5.28811328e-02 5.66625357e-01 -3.21434885e-01
1.23919755e-01 -9.64652061e-01 4.97729927e-02 1.92591548e-01
-1.82525942e-03 -4.81223196e-01 4.33226705e-01 5.97755432e-01
-4.98216808e-01 3.84550512e-01 9.83352542e-01 2.40925506e-01
-4.26283926e-01 5.97357035e-01 2.78274328e-01 -1.40243163e-02
1.06227851e+00 -7.60995090e-01 -6.05083704e-01 3.52554693e-04
-6.78234279e-01 9.68670473e-02 5.65462828e-01 4.80103701e-01
5.02214849e-01 -8.59979570e-01 -8.60212684e-01 1.23675728e+00
2.67258167e-01 -2.63156623e-01 5.40541053e-01 1.17775810e+00
-3.86178225e-01 6.77070439e-01 -5.28684855e-01 -7.19767153e-01
-1.66032124e+00 6.02349579e-01 8.02109241e-01 1.21777162e-01
-4.06901479e-01 9.81018662e-01 8.57902020e-02 -2.72899687e-01
4.07610923e-01 -8.69124755e-02 -2.24314779e-01 -1.73563495e-01
5.58151424e-01 1.58744529e-02 -6.58018375e-03 -6.51724398e-01
-1.97883323e-02 1.19716215e+00 -5.79583228e-01 5.62596083e-01
9.70436275e-01 -2.18008935e-01 -9.60152894e-02 -5.94863296e-01
1.16761160e+00 3.48957121e-01 -1.33315337e+00 -4.77610707e-01
-4.80944037e-01 -9.90907490e-01 1.65493742e-01 -1.27614284e+00
-1.15807021e+00 8.21146250e-01 1.42767239e+00 -3.32883477e-01
1.60132575e+00 -5.16210854e-01 5.30590713e-01 -3.30880612e-01
5.26701659e-02 -6.33956790e-01 -3.09398800e-01 -1.40272835e-02
4.98523861e-01 -1.59248734e+00 4.90969419e-02 -2.59165257e-01
-5.92304885e-01 1.07020855e+00 6.68408632e-01 7.55509213e-02
7.20890164e-01 5.24942800e-02 5.38793325e-01 -2.88124323e-01
-4.28007506e-02 -2.62356579e-01 6.74471438e-01 1.09435034e+00
3.01931947e-01 3.49821784e-02 -1.29258364e-01 3.75389129e-01
-9.19870064e-02 2.01314941e-01 4.58117425e-01 5.76537728e-01
3.23380351e-01 -1.54579794e+00 -5.98283529e-01 4.35862094e-01
-7.47905970e-01 -1.31186709e-01 -1.91002145e-01 9.48014736e-01
5.52496493e-01 8.03971231e-01 -1.33298710e-01 -5.17129660e-01
1.86134964e-01 -1.57966703e-01 3.98619235e-01 -2.33519956e-01
-9.26613808e-01 2.46164396e-01 -2.00817451e-01 -4.46521133e-01
-7.14472175e-01 -6.42374754e-01 -6.79746628e-01 -3.43862534e-01
-3.24420780e-01 -1.34606585e-01 5.31026840e-01 5.00456333e-01
7.48582840e-01 1.69114679e-01 7.66926825e-01 -1.68777883e-01
-4.39303815e-01 -8.53841066e-01 -8.13483119e-01 3.62665266e-01
9.87275898e-01 -6.79437637e-01 -4.51780885e-01 1.21765805e-03] | [3.743922472000122, -3.63281512260437] |
b8991af1-34a4-422b-9402-e607b1bc04b7 | perspectives-on-ai-architectures-and-co | 2304.03748 | null | https://arxiv.org/abs/2304.03748v1 | https://arxiv.org/pdf/2304.03748v1.pdf | Perspectives on AI Architectures and Co-design for Earth System Predictability | Recently, the U.S. Department of Energy (DOE), Office of Science, Biological and Environmental Research (BER), and Advanced Scientific Computing Research (ASCR) programs organized and held the Artificial Intelligence for Earth System Predictability (AI4ESP) workshop series. From this workshop, a critical conclusion that the DOE BER and ASCR community came to is the requirement to develop a new paradigm for Earth system predictability focused on enabling artificial intelligence (AI) across the field, lab, modeling, and analysis activities, called ModEx. The BER's `Model-Experimentation', ModEx, is an iterative approach that enables process models to generate hypotheses. The developed hypotheses inform field and laboratory efforts to collect measurement and observation data, which are subsequently used to parameterize, drive, and test model (e.g., process-based) predictions. A total of 17 technical sessions were held in this AI4ESP workshop series. This paper discusses the topic of the `AI Architectures and Co-design' session and associated outcomes. The AI Architectures and Co-design session included two invited talks, two plenary discussion panels, and three breakout rooms that covered specific topics, including: (1) DOE HPC Systems, (2) Cloud HPC Systems, and (3) Edge computing and Internet of Things (IoT). We also provide forward-looking ideas and perspectives on potential research in this co-design area that can be achieved by synergies with the other 16 session topics. These ideas include topics such as: (1) reimagining co-design, (2) data acquisition to distribution, (3) heterogeneous HPC solutions for integration of AI/ML and other data analytics like uncertainty quantification with earth system modeling and simulation, and (4) AI-enabled sensor integration into earth system measurements and observations. Such perspectives are a distinguishing aspect of this paper. | ['Philip W. Jones', 'Ivy B. Peng', 'Matthew R. Norman', 'Sarat S. Sreepathi', 'Tushar Krishna', 'James C. Hoe', 'Maya B. Gokhale', 'Simon D. Hammond', 'Mahantesh Halappanavar', 'James A. Ang', 'Maruti K. Mudunuru'] | 2023-04-07 | null | null | null | null | ['edge-computing'] | ['time-series'] | [-3.04676116e-01 -3.88735622e-01 2.28620619e-01 -1.37956932e-01
-3.21656048e-01 -5.80884576e-01 7.20792353e-01 4.00081873e-01
3.50733787e-01 6.50425673e-01 2.21944660e-01 -6.78907037e-01
-3.52759868e-01 -9.65776086e-01 -5.46425045e-01 -8.13983798e-01
-5.81989348e-01 4.86983478e-01 -2.63318289e-02 -7.86305666e-02
9.71778780e-02 9.96844769e-01 -1.81032383e+00 -1.36406526e-01
8.23475122e-01 1.28276646e+00 7.01476261e-02 8.04590583e-01
2.38629162e-01 6.54068053e-01 -7.77475685e-02 8.37549865e-01
2.87235856e-01 -2.58133203e-01 -3.71752203e-01 -7.65236020e-01
-4.64149266e-01 3.66024151e-02 1.79444075e-01 8.07004750e-01
5.45532465e-01 1.69282526e-01 3.98733795e-01 -1.45858967e+00
-1.64252252e-03 1.69137046e-01 -2.90425181e-01 3.53823245e-01
5.15761897e-02 6.41664922e-01 5.48934281e-01 -7.68762052e-01
2.97965229e-01 7.86140084e-01 8.34274173e-01 -1.32211015e-01
-1.13536763e+00 -8.16668510e-01 -1.23869248e-01 4.65475954e-02
-1.47605431e+00 -7.46203244e-01 3.51077110e-01 -9.35344875e-01
1.58594394e+00 8.74957681e-01 1.20631361e+00 3.18510234e-01
7.16871142e-01 1.68059319e-01 1.38454843e+00 -2.98819393e-01
9.01036203e-01 1.08133651e-01 2.67947227e-01 -5.38884066e-02
6.17351651e-01 8.69176507e-01 -2.39427537e-01 -3.99681717e-01
2.73170531e-01 -5.43127418e-01 -1.96305081e-01 1.56670511e-01
-8.41292739e-01 5.08404434e-01 2.88422883e-01 3.93801779e-01
-7.72039711e-01 2.50762433e-01 2.81140655e-01 1.47738159e-01
7.66194522e-01 8.93485665e-01 -6.91812098e-01 -1.85098246e-01
-9.84276175e-01 4.06077474e-01 9.55034316e-01 5.95620513e-01
4.70752180e-01 5.66579759e-01 1.45886019e-01 1.76562935e-01
8.50227416e-01 1.09600890e+00 8.95463862e-03 -1.02066934e+00
-1.92151964e-01 1.62204146e-01 6.07250571e-01 -7.29444742e-01
-8.98542404e-01 -1.87192261e-01 -6.41438961e-01 3.58694881e-01
-4.50517535e-01 -7.30139971e-01 -8.98755968e-01 1.22372234e+00
4.20743197e-01 1.46105900e-01 4.60400045e-01 8.41010332e-01
8.86933386e-01 1.23508167e+00 2.19762146e-01 -1.37248918e-01
1.14446712e+00 -4.23454076e-01 -5.55882156e-01 -8.89601856e-02
4.60872293e-01 -5.44925213e-01 3.94665211e-01 -9.66062211e-03
-1.01718020e+00 -3.05224180e-01 -1.08431089e+00 5.11709511e-01
-7.58199990e-01 -4.99182075e-01 9.37450409e-01 5.57780206e-01
-1.20124316e+00 6.11642659e-01 -1.41396356e+00 -7.30964780e-01
-2.55334854e-01 1.56105027e-01 2.56558150e-01 3.20490807e-01
-1.34385848e+00 1.10912442e+00 3.23642910e-01 6.41317889e-02
-8.51658165e-01 -1.31311095e+00 -5.81001163e-01 2.83575267e-01
-4.20455456e-01 -9.06004548e-01 9.70919847e-01 -8.80809426e-02
-1.59359336e+00 3.45785737e-01 -2.97390148e-02 -5.21770298e-01
8.42728093e-02 6.59725666e-02 -9.17133331e-01 -1.61817521e-01
1.97047908e-02 4.60373193e-01 -1.33552030e-01 -1.04539013e+00
-6.54776692e-01 -5.03480136e-01 -5.83419263e-01 3.01735196e-02
3.37683201e-01 4.33509618e-01 3.02162230e-01 -2.40524225e-02
2.82971591e-01 -1.14582884e+00 -3.02530438e-01 -3.25352937e-01
-9.74300951e-02 -7.09235445e-02 9.13115740e-01 -6.42693937e-01
9.95152771e-01 -1.94467521e+00 -2.38263384e-01 4.47032541e-01
-1.27273515e-01 4.40360047e-02 3.31487566e-01 9.32432771e-01
-1.90339699e-01 2.65333354e-01 -1.13509328e-03 4.98576015e-02
8.99354964e-02 -2.34848827e-01 -6.00786030e-01 3.14987957e-01
-4.56059061e-04 4.69141960e-01 -7.97592461e-01 4.04545188e-01
6.89153016e-01 1.79713696e-01 -7.27050975e-02 1.40415207e-01
-5.63579559e-01 7.87692785e-01 -4.00691569e-01 8.68781924e-01
8.63508582e-01 3.27427238e-02 -9.70193446e-02 -1.05136283e-01
-1.11567426e+00 2.02226654e-01 -1.25641465e+00 8.59947383e-01
-5.36557555e-01 7.18083560e-01 3.71591479e-01 -4.75163966e-01
1.02498245e+00 5.07881761e-01 7.97096074e-01 -5.65546155e-01
-8.34832266e-02 6.34330750e-01 2.80671120e-01 -6.08899474e-01
5.71469128e-01 -2.95085520e-01 1.69632405e-01 3.55395615e-01
-4.09618706e-01 -7.64730871e-01 -2.01555088e-01 -2.31558055e-01
9.58720505e-01 2.38016725e-01 8.45494345e-02 -1.17452693e+00
6.53412864e-02 4.80099529e-01 5.47318339e-01 2.45533779e-01
-2.95901775e-01 -7.32372850e-02 3.06462869e-02 -6.30830765e-01
-1.32605541e+00 -6.51798069e-01 -8.00577164e-01 5.13624072e-01
2.60444302e-02 -4.55493063e-01 -1.60924122e-01 5.41518748e-01
4.17549074e-01 1.14399409e+00 -3.42449754e-01 1.38276786e-01
3.58519070e-02 -1.28889966e+00 4.76330698e-01 5.43646038e-01
6.01222157e-01 -5.34885585e-01 -8.32451463e-01 4.17919576e-01
1.61060348e-01 -8.50824475e-01 7.00906456e-01 4.31862503e-01
-7.97560155e-01 -4.93263543e-01 2.75728200e-02 1.56354561e-01
3.35945003e-02 1.49051115e-01 9.33291197e-01 -1.57890096e-01
-1.68724120e-01 3.69047433e-01 -1.47918031e-01 -1.06336296e+00
-2.16675609e-01 -5.00422120e-01 5.68048537e-01 -6.89612985e-01
4.05902267e-01 -8.71127367e-01 -7.94598937e-01 4.66610134e-01
-6.31436169e-01 1.65875867e-01 1.18951373e-01 3.58112097e-01
6.56581461e-01 1.92962453e-01 6.09785557e-01 -2.76586920e-01
9.83358398e-02 -9.61240590e-01 -1.33423865e+00 2.89875239e-01
-1.03709602e+00 -1.71120912e-01 4.88572717e-01 3.63257289e-01
-1.05236351e+00 -1.29453883e-01 -2.01396719e-01 -1.11525044e-01
-2.07024723e-01 8.43439400e-01 -2.17721447e-01 -3.80810827e-01
7.48238981e-01 -7.68019482e-02 -1.02712169e-01 -2.28031546e-01
-2.60836035e-01 9.57186162e-01 2.39450812e-01 -9.07425225e-01
6.77766502e-01 5.56593060e-01 2.62248456e-01 -1.28868330e+00
4.59631421e-02 -3.83435518e-01 1.44688144e-01 -5.56378961e-01
9.51253414e-01 -1.29299128e+00 -3.85753393e-01 4.19338346e-01
-8.51571977e-01 -6.45740449e-01 -2.83319682e-01 9.62822139e-01
-8.79378617e-02 -3.91625792e-01 -2.66854763e-01 -1.21278811e+00
-7.05935836e-01 -1.15410435e+00 1.00315487e+00 6.79215729e-01
-4.66366351e-01 -9.88472700e-01 5.99184334e-01 -3.61697073e-03
1.21491051e+00 7.08572924e-01 6.20203078e-01 -3.77077371e-01
-8.03899646e-01 -1.89902917e-01 -1.68772981e-01 -1.92276642e-01
-2.49959871e-01 6.21933639e-01 -1.35041296e+00 -3.01441550e-01
5.61279319e-02 -7.58485347e-02 3.88831347e-01 5.85336208e-01
9.89056468e-01 5.60577475e-02 -7.10444748e-01 8.92092466e-01
1.68039238e+00 6.47438407e-01 7.58758128e-01 4.17111546e-01
1.48524210e-01 5.67302823e-01 4.10395473e-01 6.93181217e-01
1.74810454e-01 3.51141095e-01 2.38694444e-01 8.59822929e-02
6.08118065e-02 1.49774700e-01 2.00644702e-01 8.97098005e-01
-5.63860655e-01 -1.38324603e-01 -1.46012664e+00 4.11757022e-01
-1.45580494e+00 -9.14162755e-01 -5.66648960e-01 2.26611328e+00
1.37472525e-01 -1.89143479e-01 -3.72325301e-01 -2.76943922e-01
3.36662710e-01 1.32028088e-01 -8.71778905e-01 -5.72004259e-01
1.12966925e-01 2.96008736e-01 7.25914121e-01 4.91964638e-01
-7.11410105e-01 6.04537845e-01 6.31354523e+00 2.64527857e-01
-1.51678705e+00 1.41362995e-01 5.06789327e-01 1.88523024e-01
-8.44803154e-01 3.88543755e-01 -9.57633853e-01 4.76072162e-01
1.81551015e+00 -9.27317321e-01 5.61818182e-01 7.32832134e-01
6.24044240e-01 -4.73499358e-01 -7.03318775e-01 2.55967259e-01
-6.19043052e-01 -1.52493978e+00 -7.27854729e-01 1.70948461e-01
9.21095014e-01 9.92951870e-01 -4.74760294e-01 3.50002885e-01
7.44691074e-01 -8.69114578e-01 4.37611908e-01 8.09310079e-01
8.48291457e-01 -4.38138574e-01 7.28677034e-01 -5.63051403e-02
-1.40654969e+00 -1.46346271e-01 -4.99187827e-01 -5.36453068e-01
9.98197123e-02 1.03763151e+00 -3.80818576e-01 9.52019930e-01
1.16658747e+00 1.49907529e-01 1.07837088e-01 1.19561362e+00
3.04432094e-01 6.53148770e-01 -9.62498307e-01 -1.35128945e-01
-6.38576690e-03 -2.31957704e-01 8.70525181e-01 9.46372926e-01
6.92326128e-01 7.44772613e-01 -2.31138855e-01 1.02348626e+00
5.92107952e-01 -4.13723350e-01 -7.85371959e-01 -2.16055036e-01
9.07580018e-01 1.28312159e+00 -6.94361567e-01 -3.22310120e-01
-1.46429926e-01 6.90105185e-02 -6.13944411e-01 3.52345139e-01
-7.65968680e-01 -3.28200519e-01 1.09704280e+00 1.30982444e-01
-3.05882543e-01 -2.57958174e-01 -8.04831684e-01 -8.43406677e-01
-4.56809253e-01 -4.47959006e-01 1.59787744e-01 -1.32597983e+00
-1.07315850e+00 4.91122901e-01 5.25723279e-01 -1.00247264e+00
-3.72175455e-01 -4.22474772e-01 -1.16774774e+00 1.22939289e+00
-1.34862721e+00 -7.69065857e-01 -7.56293356e-01 -2.79527992e-01
-1.82764959e-02 6.61164448e-02 1.22022986e+00 1.50851905e-01
-6.21297300e-01 -1.95236564e-01 1.02696884e+00 -7.59311438e-01
2.59946018e-01 -9.71534610e-01 5.85030198e-01 7.76059151e-01
-7.34063923e-01 2.37531126e-01 1.07096946e+00 -1.03098178e+00
-2.00680447e+00 -1.27849007e+00 6.43101633e-01 -2.23087981e-01
1.14618731e+00 3.77803594e-02 -7.15352714e-01 7.75930345e-01
3.39958608e-01 -1.28658041e-02 7.57454515e-01 3.36645581e-02
4.63435382e-01 -3.09847206e-01 -1.27310991e+00 3.80089462e-01
5.54600239e-01 -3.39541346e-01 -1.69941902e-01 6.85890436e-01
7.57318199e-01 -5.70899844e-01 -1.41716301e+00 9.36473548e-01
7.01403201e-01 -6.16778970e-01 5.31024218e-01 -4.38179076e-01
3.42002809e-01 -7.36078680e-01 -5.18528342e-01 -1.38679469e+00
-4.14621949e-01 -5.29686809e-01 3.49969305e-02 1.01335073e+00
3.62846702e-01 -9.43491518e-01 2.43257701e-01 1.34400761e+00
-7.78784573e-01 -5.08359134e-01 -9.34382617e-01 -9.45464849e-01
2.25769803e-01 -6.57660007e-01 1.00161207e+00 1.00088668e+00
2.42828488e-01 -6.41103014e-02 2.52431422e-01 9.26693439e-01
5.99004865e-01 3.39320600e-01 6.03794515e-01 -1.42932415e+00
-8.78691599e-02 -2.77751297e-01 -5.25283515e-01 -4.35319394e-02
-6.53578281e-01 -6.18201017e-01 1.09481081e-01 -1.58406603e+00
-1.14006981e-01 -8.26555729e-01 -1.10109203e-01 1.83293402e-01
2.25300387e-01 -3.46326262e-01 1.10162549e-01 5.40217042e-01
2.05341488e-01 5.80946565e-01 5.96078813e-01 1.19604386e-01
-3.21883380e-01 -1.02972634e-01 -3.70249659e-01 2.85745829e-01
1.07329500e+00 -2.66840339e-01 -9.77704301e-02 -3.50593030e-01
4.75013196e-01 3.22047174e-01 4.45532382e-01 -1.66471350e+00
2.59074628e-01 -4.99548078e-01 5.65287113e-01 -6.12269759e-01
1.82174519e-02 -1.03521597e+00 1.29943943e+00 6.16260707e-01
4.16406423e-01 7.06557259e-02 9.79907393e-01 5.61078787e-02
5.06585091e-03 1.03664018e-01 7.22616255e-01 -4.97059561e-02
-4.70502228e-01 1.56373426e-01 -6.50207281e-01 -4.63374585e-01
1.37911320e+00 1.59110576e-01 -8.29849780e-01 -1.14919931e-01
-7.09995389e-01 8.34766924e-01 6.66277349e-01 -1.77328691e-01
-9.58173573e-02 -1.02749085e+00 -8.84246409e-01 1.31884381e-01
-1.10543519e-01 1.58708975e-01 5.30133605e-01 8.56622756e-01
-8.26840878e-01 7.38100886e-01 -2.14857161e-01 -7.54904330e-01
-5.81730306e-01 4.18852806e-01 7.16259122e-01 -1.13264993e-01
-5.09463727e-01 3.70068818e-01 -8.79548267e-02 -5.20150304e-01
-3.06166321e-01 -4.12356406e-01 2.49718592e-01 -1.65373534e-01
6.52754605e-01 5.16062021e-01 4.13185567e-01 -1.57073379e-01
-4.06840116e-01 3.40948910e-01 7.15460122e-01 -1.86106771e-01
1.64812195e+00 1.13828788e-02 -2.96674758e-01 8.13593686e-01
6.77015483e-01 -2.50136048e-01 -9.95710552e-01 4.14989769e-01
-1.35225258e-04 -1.29782185e-01 5.27142942e-01 -1.14380598e+00
-7.94282675e-01 5.94764471e-01 7.90012956e-01 4.69979793e-01
1.16323125e+00 -1.70557141e-01 4.16515887e-01 2.88771421e-01
2.99251974e-01 -1.16386795e+00 -7.91344345e-01 5.15050828e-01
1.33929813e+00 -7.95300961e-01 3.10809046e-01 1.09737985e-01
-1.35402933e-01 1.09610510e+00 7.85279214e-01 2.22751081e-01
1.15549862e+00 8.56696665e-01 -2.11640403e-01 -5.39166868e-01
-9.64262247e-01 1.88765058e-03 -2.82908827e-01 2.83919215e-01
4.16617900e-01 6.34359181e-01 -1.89010143e-01 6.73010290e-01
-1.70819059e-01 2.83696324e-01 2.74895698e-01 1.02539563e+00
-6.73580289e-01 -6.28585696e-01 -9.29665685e-01 6.80658162e-01
5.72315557e-03 -1.96866855e-01 -7.67235234e-02 6.09064758e-01
-1.24300778e-01 7.89960086e-01 3.87049317e-01 -6.46051586e-01
1.27005965e-01 1.47580519e-01 -1.82341084e-01 -1.60630986e-01
-4.94725674e-01 -1.42399505e-01 5.54050624e-01 -5.11223376e-01
-5.20041995e-02 -9.14941192e-01 -1.30596662e+00 -8.24619234e-01
-3.85118216e-01 7.09901333e-01 1.42962444e+00 6.93794966e-01
8.80081773e-01 8.23237240e-01 7.87456214e-01 -1.15247238e+00
-6.30005971e-02 -1.24121749e+00 -6.84357643e-01 -5.98741114e-01
-1.65463388e-01 -5.96343577e-01 -6.41070366e-01 -5.52949309e-01] | [6.482177734375, 3.2069108486175537] |
8c276c18-ae5e-4410-b751-7758a9cbeab5 | monte-carlo-tree-search-based-variable | 2210.01628 | null | https://arxiv.org/abs/2210.01628v2 | https://arxiv.org/pdf/2210.01628v2.pdf | Monte Carlo Tree Search based Variable Selection for High Dimensional Bayesian Optimization | Bayesian optimization (BO) is a class of popular methods for expensive black-box optimization, and has been widely applied to many scenarios. However, BO suffers from the curse of dimensionality, and scaling it to high-dimensional problems is still a challenge. In this paper, we propose a variable selection method MCTS-VS based on Monte Carlo tree search (MCTS), to iteratively select and optimize a subset of variables. That is, MCTS-VS constructs a low-dimensional subspace via MCTS and optimizes in the subspace with any BO algorithm. We give a theoretical analysis of the general variable selection method to reveal how it can work. Experiments on high-dimensional synthetic functions and real-world problems (i.e., NAS-bench problems and MuJoCo locomotion tasks) show that MCTS-VS equipped with a proper BO optimizer can achieve state-of-the-art performance. | ['Chao Qian', 'Xiaobin Huang', 'Ke Xue', 'Lei Song'] | 2022-10-04 | null | null | null | null | ['variable-selection'] | ['methodology'] | [-1.46332040e-01 -6.85663700e-01 -3.93346816e-01 -1.65707842e-01
-7.96561539e-01 -2.89126128e-01 9.83302519e-02 -3.64847749e-01
-3.43219995e-01 1.18973005e+00 -3.16650718e-01 -3.09331149e-01
-3.08467120e-01 -6.87113702e-01 -5.16444683e-01 -1.12636232e+00
-2.81032681e-01 9.12891507e-01 2.11837724e-01 -1.58462599e-01
2.68958926e-01 2.63159424e-01 -1.44775319e+00 -2.74267346e-01
1.01926601e+00 9.52200949e-01 2.34641328e-01 6.64040446e-01
2.97924113e-02 -1.13204710e-01 -5.92152774e-01 -2.58171111e-01
3.89944434e-01 -5.78422725e-01 -6.37325466e-01 -1.49588034e-01
5.52964862e-03 -6.05650917e-02 -1.45540744e-01 8.84119987e-01
7.09118068e-01 4.45366472e-01 5.47943294e-01 -1.37641382e+00
-1.61124002e-02 4.94012654e-01 -5.65430760e-01 1.14067711e-01
1.69659033e-01 5.12354434e-01 1.02761054e+00 -7.93522000e-01
6.69506192e-01 1.49289584e+00 5.22334337e-01 5.57600558e-01
-1.67679298e+00 -6.08784974e-01 1.33833453e-01 4.25780147e-01
-1.41620004e+00 -1.88183591e-01 6.16042137e-01 -4.71797615e-01
6.87131405e-01 4.66496170e-01 9.39129651e-01 9.96531665e-01
4.35222089e-01 9.66878355e-01 1.07929718e+00 -1.33420855e-01
9.71754968e-01 -2.45786488e-01 5.58609143e-02 8.01544905e-01
2.71069109e-01 3.60138804e-01 -1.10839927e+00 -5.62493086e-01
4.99232709e-01 -2.77045637e-01 -2.88305432e-01 -9.55351651e-01
-1.23670256e+00 1.26092780e+00 1.54835507e-01 -2.62968630e-01
-2.77448475e-01 2.86622912e-01 2.96822131e-01 8.35638344e-02
3.57942820e-01 6.88708305e-01 -7.35444903e-01 -3.85516822e-01
-1.11592066e+00 8.79868209e-01 8.44484985e-01 8.06963384e-01
6.30344272e-01 9.74697992e-02 -2.86771953e-01 7.64922261e-01
4.51915205e-01 5.88657737e-01 3.32830936e-01 -9.48699594e-01
5.35167098e-01 3.43288809e-01 1.24027461e-01 -7.43204117e-01
-6.20007277e-01 -5.62556088e-01 -8.87952924e-01 3.82224053e-01
3.92870575e-01 -4.09655333e-01 -8.25147390e-01 1.70207787e+00
7.99146891e-01 2.68592715e-01 -2.31397256e-01 1.09488404e+00
3.46596330e-01 5.33071816e-01 -3.60751450e-01 -5.46002686e-01
9.38639581e-01 -1.10486221e+00 -6.93862438e-01 -4.01188254e-01
4.77020234e-01 -1.15411408e-01 1.00656164e+00 6.75385773e-01
-9.18380976e-01 -1.68515861e-01 -1.20061374e+00 4.23237413e-01
-3.70837189e-02 -2.41221189e-01 1.12194359e+00 9.13829982e-01
-5.29939532e-01 7.70489514e-01 -1.15706635e+00 3.65377814e-02
5.54671884e-01 4.53607082e-01 -9.98506024e-02 -4.55049798e-02
-1.02022433e+00 6.11799002e-01 4.25230533e-01 1.66564494e-01
-1.24089980e+00 -4.37169135e-01 -7.86759734e-01 2.76411958e-02
1.03382695e+00 -1.10238922e+00 9.86237705e-01 -1.92035630e-01
-1.94516647e+00 2.03457296e-01 -6.52056396e-01 -4.05071169e-01
5.39190054e-01 -3.88622403e-01 3.39457914e-02 -1.78998113e-01
-8.10111240e-02 1.81061208e-01 1.09367752e+00 -9.46476817e-01
-3.40194464e-01 -8.27328920e-01 -3.37746680e-01 1.80509910e-01
-2.20196381e-01 9.37094465e-02 -6.47603333e-01 -5.79310238e-01
5.32988787e-01 -1.17993307e+00 -7.69358337e-01 -1.16991863e-01
-6.51875079e-01 -5.19909710e-02 5.10007620e-01 -3.42143923e-01
1.80492079e+00 -1.84788108e+00 1.04277265e+00 3.89105767e-01
1.69138893e-01 4.06773575e-02 1.14825711e-01 1.67021796e-01
2.56716281e-01 2.00112745e-01 -7.67571151e-01 -5.18297374e-01
1.73514545e-01 3.20272148e-01 -8.08000639e-02 7.17604756e-01
-1.06166370e-01 5.75824022e-01 -9.63041067e-01 -5.79872489e-01
1.37946159e-01 -6.09183032e-03 -1.05388951e+00 2.54973352e-01
-5.46491206e-01 5.45394063e-01 -8.19584370e-01 9.00223434e-01
6.56396806e-01 -1.75504208e-01 2.13088334e-01 2.84630835e-01
4.34387736e-02 -9.83387604e-02 -1.68911886e+00 1.87354791e+00
-1.11326672e-01 4.81297970e-01 6.80349022e-02 -9.26757991e-01
7.35636413e-01 -2.24480048e-01 3.02068383e-01 -3.36138368e-01
1.87666491e-01 2.65407562e-01 -5.28990999e-02 -4.67944086e-01
2.67866105e-01 7.77567830e-03 -1.80140749e-01 1.79438069e-01
4.81375074e-03 -3.84366065e-01 5.05312204e-01 -9.41961780e-02
1.11276484e+00 3.11382145e-01 4.85321820e-01 -2.67711222e-01
5.52398324e-01 1.06788971e-01 1.01977336e+00 9.53581452e-01
-1.99469879e-01 5.10396838e-01 5.52601218e-01 -3.75595063e-01
-5.98524272e-01 -1.00578797e+00 -3.71919334e-01 1.06443667e+00
3.05040717e-01 -7.88663864e-01 -6.69610322e-01 -6.35763824e-01
2.75683612e-01 9.08784688e-01 -4.86940652e-01 -9.81255472e-02
-6.25698030e-01 -1.35199356e+00 9.15565714e-02 1.15476795e-01
4.05072331e-01 -6.93124592e-01 -8.91708553e-01 4.70189571e-01
-6.97527304e-02 -5.76643348e-01 -8.71691331e-02 4.19114888e-01
-1.07301414e+00 -8.73849511e-01 -5.43771327e-01 -1.29753634e-01
2.08201781e-01 7.63581842e-02 1.12186635e+00 -1.11628279e-01
-3.62566799e-01 -4.14458334e-01 -1.75783664e-01 -1.98426649e-01
-4.67804000e-02 1.68529153e-01 2.38007948e-01 -2.17493325e-01
1.20628953e-01 -4.38720465e-01 -4.43192273e-01 7.06482291e-01
-4.92361248e-01 -2.63020862e-02 2.92644471e-01 1.44771981e+00
9.42644656e-01 3.54257494e-01 1.56943575e-01 -6.90510631e-01
4.59618509e-01 -4.30704951e-01 -1.16193926e+00 8.79440382e-02
-6.57201946e-01 4.96781707e-01 4.16060895e-01 -3.18485618e-01
-7.56433606e-01 1.23741254e-01 -2.01422777e-02 -4.14287210e-01
1.47693917e-01 7.23972142e-01 -3.14841866e-01 -1.38156950e-01
5.49191117e-01 3.05428822e-02 -2.53966987e-01 -7.37850070e-01
1.15921788e-01 4.53543544e-01 2.04313129e-01 -7.40505755e-01
9.44683254e-01 4.58732039e-01 5.42113185e-01 -5.89785576e-01
-7.37799704e-01 -2.28955016e-01 -4.18336719e-01 -1.34540558e-01
6.76737607e-01 -4.87717122e-01 -9.12443101e-01 4.81516838e-01
-8.06725264e-01 -4.61822420e-01 1.64161697e-02 5.72073102e-01
-8.02280486e-01 3.96412760e-02 -1.33233532e-01 -9.58408475e-01
-5.99839315e-02 -1.59894884e+00 1.01513970e+00 3.14171940e-01
-1.19187295e-01 -6.73110604e-01 3.60647619e-01 4.78603363e-01
3.22266847e-01 3.06034356e-01 8.07564616e-01 -2.20880240e-01
-7.77672529e-01 1.25615999e-01 3.68988067e-01 -1.43468156e-01
-4.53188032e-01 1.84265554e-01 -5.75667918e-01 -4.31482553e-01
1.88730821e-01 -3.09205800e-01 9.03673947e-01 6.59424782e-01
1.45251465e+00 -1.10916056e-01 -5.69110811e-01 1.25778496e+00
1.39737201e+00 2.35392660e-01 3.17011118e-01 7.15746820e-01
4.40177262e-01 1.85752153e-01 1.08809590e+00 8.58638942e-01
3.61499786e-02 1.13230288e+00 5.93426943e-01 1.95910096e-01
4.48137701e-01 8.74428153e-02 2.85909146e-01 5.58529735e-01
2.01368496e-01 -2.99004465e-01 -1.01217818e+00 2.16224194e-01
-2.09734178e+00 -6.95983946e-01 -2.50207484e-01 2.29071760e+00
9.63361263e-01 2.00468600e-01 2.37253696e-01 1.51180312e-01
7.23972261e-01 2.15948120e-01 -1.10646963e+00 -2.21848071e-01
-1.60221517e-01 2.76169002e-01 5.65439284e-01 3.24624866e-01
-1.10142088e+00 9.85014319e-01 6.83611727e+00 1.23661852e+00
-8.75397146e-01 -8.45902413e-03 4.49590415e-01 -4.07237172e-01
9.06969681e-02 1.12910300e-01 -1.09312475e+00 7.16787100e-01
7.02310860e-01 -2.14675307e-01 9.30384874e-01 1.08296859e+00
2.67811716e-01 -3.76266718e-01 -9.95586932e-01 1.05326438e+00
-4.47496139e-02 -1.27164924e+00 -3.64585847e-01 -4.57573608e-02
9.57626700e-01 -7.08599985e-02 -3.77505347e-02 3.85752082e-01
3.82448465e-01 -1.31980860e+00 5.48922718e-01 1.46202207e-01
3.91409427e-01 -8.29369187e-01 7.95793176e-01 4.03195053e-01
-9.77258027e-01 -2.41104007e-01 -2.87828207e-01 1.45277888e-01
5.15418470e-01 6.37216270e-01 -4.72050369e-01 6.43721163e-01
9.14619625e-01 4.79298830e-01 -4.33501542e-01 1.51340020e+00
-4.52564210e-01 8.60712171e-01 -5.79474688e-01 -3.97860855e-01
5.10873683e-02 -5.10275543e-01 1.03202760e+00 8.01919699e-01
2.09800050e-01 -2.64714897e-01 1.39401451e-01 9.27481055e-01
2.21349984e-01 -1.70195531e-02 -1.16344050e-01 -1.55731328e-02
6.83221340e-01 8.36345077e-01 -4.83484358e-01 -2.24411055e-01
1.23347208e-01 6.68250203e-01 2.78542221e-01 4.17376399e-01
-9.38741207e-01 -4.24169570e-01 9.06160474e-01 -5.33592999e-01
6.46268308e-01 -4.85744029e-01 -7.56207764e-01 -1.28973603e+00
-4.77521755e-02 -1.25137579e+00 4.23984617e-01 -4.84690011e-01
-9.17475939e-01 2.90094465e-01 3.65344621e-02 -9.54555392e-01
-2.11659625e-01 -5.64368367e-01 -3.55376691e-01 8.20338607e-01
-1.10175490e+00 -2.34859675e-01 -2.25357309e-01 4.71034497e-01
6.77954495e-01 -2.54643440e-01 6.82450175e-01 9.46569741e-02
-1.20938516e+00 6.03678584e-01 6.61264300e-01 -5.60621142e-01
4.02609229e-01 -1.20273101e+00 3.06925684e-01 7.01584220e-01
-6.58555627e-02 7.85121083e-01 1.19839549e+00 -6.31757319e-01
-2.03627515e+00 -6.45303071e-01 2.54957944e-01 -2.01430574e-01
6.62985742e-01 -4.32090193e-01 -7.15326309e-01 2.49409884e-01
-3.52507889e-01 -2.99329519e-01 4.83568013e-01 4.68079180e-01
1.78963318e-01 -5.03639095e-02 -1.14759290e+00 9.24478829e-01
1.21368277e+00 3.32175121e-02 -4.36904728e-01 6.20884776e-01
6.23037934e-01 -6.97591007e-01 -6.87843800e-01 5.00911772e-01
6.20898068e-01 -1.05490720e+00 1.14605713e+00 -8.05320323e-01
2.36475304e-01 -5.23786068e-01 -2.63426214e-01 -1.58055389e+00
-9.78048816e-02 -1.01943350e+00 -5.95418274e-01 8.58809590e-01
1.71178401e-01 -6.76165104e-01 9.68288541e-01 3.82940263e-01
6.54219836e-02 -1.11982894e+00 -1.55851340e+00 -1.07386529e+00
-5.97887188e-02 -5.26181579e-01 8.27032268e-01 5.36313891e-01
-2.68243521e-01 3.36434513e-01 -4.42521900e-01 1.52528673e-01
1.05843019e+00 2.72668153e-01 1.23208368e+00 -1.03121281e+00
-8.88449967e-01 -4.94425416e-01 -3.20053935e-01 -1.24228442e+00
1.77409708e-01 -5.43839633e-01 2.58325189e-01 -1.15976191e+00
2.27330059e-01 -4.95181918e-01 -4.29823697e-02 9.49025154e-03
-6.85670972e-01 -3.16292703e-01 -2.79977918e-02 5.84187806e-02
-5.72052538e-01 1.15194190e+00 1.12812984e+00 -6.00366183e-02
-3.71478409e-01 1.06308803e-01 -2.09408581e-01 6.04113758e-01
6.75932825e-01 -9.61598039e-01 -4.80950773e-02 -2.17642173e-01
2.92379349e-01 3.18464965e-01 6.29962757e-02 -9.13024902e-01
7.86324814e-02 -7.57667124e-01 1.44703716e-01 -8.04483712e-01
6.35676742e-01 -4.70736444e-01 2.26282880e-01 5.96486330e-01
-6.51604161e-02 3.44888978e-02 6.57524215e-03 6.37192249e-01
-7.13502988e-02 -5.09683371e-01 7.86451936e-01 9.34836268e-02
-4.46577132e-01 2.89713323e-01 -3.85275573e-01 1.93880334e-01
9.86159861e-01 -6.06040321e-02 -1.96342587e-01 2.19464302e-04
-5.92777848e-01 8.40373516e-01 5.62159002e-01 -1.54684588e-01
5.37200570e-01 -1.10847092e+00 -6.24236047e-01 2.71597058e-01
-1.13351598e-01 2.78893530e-01 1.00324363e-01 7.65181422e-01
-4.89506304e-01 3.11061502e-01 1.47157386e-01 -9.34458554e-01
-1.30235660e+00 5.00401497e-01 2.16115966e-01 -5.26018262e-01
-4.98971134e-01 1.21770871e+00 -9.12540182e-02 -4.91445720e-01
1.62340939e-01 1.00851513e-03 1.27856597e-01 -1.55899912e-01
4.14959162e-01 7.26708353e-01 -2.06472740e-01 -8.49071220e-02
-4.36865866e-01 5.10867476e-01 3.61936927e-01 -4.10398245e-01
1.47404134e+00 1.86873283e-02 -5.32306433e-02 5.43108404e-01
9.58379447e-01 -2.05043912e-01 -1.19039822e+00 -6.24895729e-02
5.85263520e-02 -1.04371142e+00 3.87412667e-01 -5.30193925e-01
-1.10371649e+00 8.15292716e-01 4.80290204e-01 -7.03012496e-02
9.95954692e-01 -1.88011423e-01 6.53233290e-01 7.00810194e-01
7.60651529e-01 -1.23502553e+00 -9.20344666e-02 6.85569704e-01
7.76021481e-01 -1.16970229e+00 3.63506019e-01 -3.53211403e-01
-5.31291664e-01 9.74477291e-01 5.55152476e-01 -5.10932207e-02
5.32616675e-01 -9.00345482e-03 -4.97376800e-01 -1.95053548e-01
-9.69724596e-01 -1.86723366e-01 -7.25877434e-02 2.66567498e-01
-1.55032739e-01 1.44880742e-01 -4.90166217e-01 6.19695842e-01
-3.25037688e-01 -1.37095563e-02 6.65415078e-02 1.02979326e+00
-2.85198867e-01 -1.04280269e+00 -7.60735929e-01 5.66021740e-01
-1.74936011e-01 -1.25458390e-01 -4.77683991e-02 5.94672322e-01
-1.14229612e-01 9.87722397e-01 -4.12919611e-01 -4.33333844e-01
4.72889468e-02 5.48653305e-03 4.45656002e-01 -3.82224053e-01
-5.37959695e-01 1.40750512e-01 1.92853749e-01 -9.15442944e-01
2.30250880e-02 -1.07688379e+00 -9.19204116e-01 -2.83569247e-01
-6.90459788e-01 2.94621617e-01 9.85345960e-01 9.64084089e-01
2.46525675e-01 6.62142038e-01 7.14198470e-01 -8.85249257e-01
-1.00677609e+00 -6.38682544e-01 -6.65792823e-01 3.40954326e-02
1.88019335e-01 -1.41007078e+00 -7.68434763e-01 -5.21683872e-01] | [6.5266804695129395, 4.003716468811035] |
95a10a8f-c525-41f7-8a1d-4ec738444437 | using-deep-learning-with-large-aggregated | 2201.01669 | null | https://arxiv.org/abs/2201.01669v3 | https://arxiv.org/pdf/2201.01669v3.pdf | Using Deep Learning with Large Aggregated Datasets for COVID-19 Classification from Cough | The Covid-19 pandemic has been one of the most devastating events in recent history, claiming the lives of more than 5 million people worldwide. Even with the worldwide distribution of vaccines, there is an apparent need for affordable, reliable, and accessible screening techniques to serve parts of the World that do not have access to Western medicine. Artificial Intelligence can provide a solution utilizing cough sounds as a primary screening mode for COVID-19 diagnosis. This paper presents multiple models that have achieved relatively respectable performance on the largest evaluation dataset currently presented in academic literature. Through investigation of a self-supervised learning model (Area under the ROC curve, AUC = 0.807) and a convolutional nerual network (CNN) model (AUC = 0.802), we observe the possibility of model bias with limited datasets. Moreover, we observe that performance increases with training data size, showing the need for the worldwide collection of data to help combat the Covid-19 pandemic with non-traditional means. | ['Aaron Broukhim', 'Wei Chen', 'Christian Canham', 'Praveen Govindan', 'Akanksha Rajput', 'Gunvant Chaudhari', 'Jaclyn Xiao', 'Jennifer Ranjani J.', 'Daniel C. H. Tan', 'Esin Darici Haritaoglu', 'Mert Pilanci', 'Amil Khanzada', 'Laura Gomezjurado', 'Minami Yamaura', 'Nicholas Rasmussen'] | 2022-01-05 | null | null | null | null | ['covid-19-detection'] | ['medical'] | [ 3.22115943e-02 -4.87260334e-02 -1.67779759e-01 4.35445160e-02
-3.91648561e-01 -6.78127110e-01 4.24435616e-01 7.03500569e-01
-9.34441268e-01 6.60294056e-01 -2.64216885e-02 -6.36230767e-01
-1.09023631e-01 -6.89041734e-01 -2.97064155e-01 -4.39670742e-01
-3.03737193e-01 5.99954963e-01 -3.19439769e-02 -3.31433445e-01
-2.52325181e-02 5.80326319e-01 -1.24243963e+00 -2.01948378e-02
1.15006649e+00 8.75344396e-01 2.46129289e-01 7.57001460e-01
1.67246357e-01 2.23142222e-01 -8.15885961e-01 -2.15458930e-01
9.18394774e-02 -3.24629754e-01 -5.32307625e-01 -7.35282242e-01
-6.24822676e-02 -6.51474297e-01 2.60671020e-01 5.43826938e-01
5.43558419e-01 -2.40872189e-01 6.42503560e-01 -7.15793550e-01
-5.05034029e-01 3.22016180e-02 -2.36121759e-01 4.10047561e-01
2.62368590e-01 2.52739727e-01 5.72937250e-01 -4.29698378e-01
6.51835144e-01 8.21527660e-01 8.88481379e-01 5.48118412e-01
-9.27801609e-01 -6.03017986e-01 -4.66596782e-01 1.09508540e-02
-1.09358478e+00 2.66120255e-01 2.90084153e-01 -5.30437350e-01
1.05596387e+00 4.07689989e-01 8.83721888e-01 8.99626911e-01
2.86541313e-01 -3.73161435e-02 1.15764844e+00 -1.19881809e-01
-1.61124393e-02 4.01907444e-01 -1.63384542e-01 5.85327983e-01
8.31177235e-01 1.73806369e-01 -1.57644209e-02 -5.77206254e-01
4.37321126e-01 2.86738753e-01 -1.28696382e-01 8.98046121e-02
-1.03754938e+00 9.80049431e-01 3.25889856e-01 5.41606188e-01
-4.88085330e-01 -2.42547512e-01 4.73856091e-01 1.46955639e-01
4.85364616e-01 8.17624748e-01 -5.98030746e-01 -3.43117118e-02
-8.99948418e-01 2.53766656e-01 5.84368587e-01 -1.55503184e-01
2.94602185e-01 2.87851249e-03 3.33954915e-02 5.27626097e-01
1.01977855e-01 8.86092365e-01 4.00778621e-01 -4.75578159e-01
1.69378549e-01 8.36556077e-01 2.99786627e-01 -1.13189280e+00
-8.04812491e-01 -6.06036901e-01 -8.43785942e-01 -3.04475188e-01
3.88209194e-01 -5.30593455e-01 -6.97780728e-01 1.69243419e+00
2.34497786e-01 9.00899619e-02 -4.19992395e-02 6.97551012e-01
5.85978031e-01 5.57859004e-01 2.46586502e-01 -1.91443622e-01
1.38698220e+00 -3.42692226e-01 -3.89978617e-01 5.46766408e-02
5.42756617e-01 -4.04981792e-01 7.26238728e-01 3.82075459e-01
-6.56076849e-01 -2.45658010e-01 -1.00912368e+00 2.67312825e-01
-6.46611452e-01 -1.28689259e-01 5.21143377e-01 9.59394157e-01
-8.65444362e-01 4.04527694e-01 -8.37811828e-01 -7.86598444e-01
2.50114650e-01 3.95734727e-01 -2.75160164e-01 -1.20242506e-01
-1.24850190e+00 1.21552777e+00 3.68654013e-01 1.32349983e-01
-9.13148999e-01 -5.42405069e-01 -3.84170920e-01 -2.19810188e-01
-8.61162320e-02 -6.10714912e-01 5.57611704e-01 -4.67058837e-01
-6.30303800e-01 8.26866627e-01 1.89370860e-03 -6.06173754e-01
3.29070151e-01 -3.93283188e-01 -5.15515685e-01 4.15757388e-01
-9.75746363e-02 6.85134411e-01 5.09263694e-01 -9.71249878e-01
-5.82446277e-01 -4.89289016e-01 -1.72095418e-01 -9.93334800e-02
-5.41521192e-01 2.49649882e-01 2.76707232e-01 -3.94624680e-01
-5.25338173e-01 -9.95803833e-01 -3.27795774e-01 -2.65341043e-01
-1.68415159e-01 -2.28575870e-01 8.01922023e-01 -1.11392832e+00
1.13700032e+00 -1.77617049e+00 -2.52049208e-01 1.43002331e-01
3.48320335e-01 8.24100614e-01 -2.25792360e-02 6.11681342e-01
2.90741950e-01 3.87820542e-01 -3.35553885e-01 1.81922436e-01
-4.20559108e-01 1.09059833e-01 -8.08905885e-02 5.18827140e-01
4.41447556e-01 8.01787496e-01 -1.02307677e+00 -1.70822918e-01
2.57820755e-01 8.56385171e-01 -4.16635960e-01 3.90713900e-01
-2.32603878e-01 5.84776998e-01 -5.23000717e-01 5.55967331e-01
4.57197130e-01 -5.12992918e-01 2.01422587e-01 2.90167153e-01
-1.97199002e-01 -2.49530058e-02 -4.56350207e-01 1.06856656e+00
-1.81148753e-01 4.89899367e-01 -1.93701759e-01 -7.19599962e-01
6.85028732e-01 4.84493434e-01 5.89222848e-01 -6.10433877e-01
3.86993319e-01 4.18288708e-01 2.35814258e-01 -6.86271310e-01
1.90763816e-01 -9.29561108e-02 1.97563127e-01 2.69586951e-01
-2.15637207e-01 1.82316408e-01 7.87932798e-02 -1.16671704e-01
9.76489246e-01 -2.33833879e-01 1.89626142e-01 -2.84700304e-01
3.43908489e-01 4.48151171e-01 2.13635296e-01 6.44120038e-01
-2.30492264e-01 4.13980991e-01 2.38691553e-01 -5.52165627e-01
-8.95738423e-01 -8.74108195e-01 -4.07945305e-01 6.53400838e-01
-2.20640734e-01 9.38128456e-02 -7.15034425e-01 -2.24199414e-01
-1.01963066e-01 4.72006649e-01 -6.51087224e-01 2.69553680e-02
-3.48714083e-01 -1.00418985e+00 9.13678110e-01 2.46434584e-01
5.31638503e-01 -9.76157248e-01 -1.32967842e+00 1.78829193e-01
-3.33849907e-01 -8.97845507e-01 1.77499682e-01 1.68099627e-01
-7.87176728e-01 -1.13615859e+00 -1.00636828e+00 -5.02840340e-01
4.17448223e-01 6.06694743e-02 6.94794178e-01 4.75893945e-01
-5.70400417e-01 3.06987375e-01 -2.39827529e-01 -7.72965968e-01
-4.26827699e-01 1.12675503e-01 9.04396176e-02 -4.65379626e-01
4.78617400e-01 -7.83480108e-02 -1.15242851e+00 -3.62154394e-02
-8.39563668e-01 -2.91006595e-01 5.26703358e-01 5.45003057e-01
3.05123657e-01 -3.37727755e-01 8.88823152e-01 -6.13384962e-01
6.42420828e-01 -9.48425114e-01 -4.29715812e-01 1.35943452e-02
-9.03014541e-01 -4.38430101e-01 5.84440231e-01 -3.76342326e-01
-7.25815058e-01 -4.30144012e-01 -3.07306170e-01 -5.34053445e-02
-2.69528419e-01 6.64603829e-01 7.13088334e-01 2.16341004e-01
6.95979357e-01 -5.41866384e-02 1.75348446e-01 -5.15934348e-01
-5.78518212e-02 7.51190424e-01 4.44488138e-01 -6.21681698e-02
7.47831345e-01 6.49654448e-01 3.87619920e-02 -1.22734284e+00
-6.98383510e-01 -6.27888918e-01 -1.12684861e-01 -1.68863371e-01
1.35035205e+00 -8.14793646e-01 -7.24947095e-01 3.78213942e-01
-1.07801414e+00 -6.14201277e-02 3.19429517e-01 7.04191327e-01
1.19919322e-01 1.35934576e-01 -4.34493423e-01 -9.28994834e-01
-7.94886291e-01 -6.65711403e-01 6.07738435e-01 2.99233854e-01
-5.37699163e-01 -1.02351213e+00 7.19990253e-01 5.80318093e-01
9.21296656e-01 7.64503241e-01 9.87569749e-01 -1.13494921e+00
-2.27292538e-01 -5.32331169e-01 -2.52445817e-01 3.96386921e-01
3.60087544e-01 -2.01076552e-01 -1.18105674e+00 -4.38440412e-01
7.81638827e-03 -3.90432417e-01 7.80174613e-01 3.94702077e-01
7.53975987e-01 -2.76429802e-01 -2.95369357e-01 1.34786189e-01
1.48533642e+00 5.88098228e-01 4.14842576e-01 2.38193229e-01
3.74158889e-01 6.36140406e-01 2.36006111e-01 2.49148354e-01
4.45379347e-01 1.67517126e-01 5.60143471e-01 -2.87430376e-01
2.55566120e-01 -2.05216065e-01 -1.10218517e-01 5.38286328e-01
-4.06836927e-01 -2.79048473e-01 -1.16098988e+00 7.02792346e-01
-1.20305562e+00 -9.12407815e-01 -9.56827104e-02 2.50623107e+00
4.53763992e-01 2.70368401e-02 2.97206700e-01 -7.33033195e-02
3.70230049e-01 -2.62778908e-01 -4.76046681e-01 -6.18285656e-01
-7.79651478e-02 2.81126767e-01 2.31807575e-01 2.42188238e-02
-9.44531262e-01 2.78305084e-01 6.97763777e+00 2.20366225e-01
-1.44635427e+00 6.09549917e-02 7.37089753e-01 2.02547878e-01
-9.92859378e-02 -3.13623041e-01 -3.40340883e-01 4.99945015e-01
1.20501983e+00 4.29207087e-02 2.68285125e-01 3.97555202e-01
4.15024698e-01 -1.35912552e-01 -4.31008130e-01 3.54306877e-01
1.22314848e-01 -1.13574934e+00 -2.33864188e-01 1.10126950e-01
7.02660263e-01 4.33027744e-01 9.43250209e-02 -7.41128158e-03
-1.41989693e-01 -1.22789037e+00 5.75168617e-02 3.78499687e-01
8.70429158e-01 -6.56226575e-01 1.09026945e+00 3.54508042e-01
-8.57956827e-01 -5.63590825e-02 -2.50586402e-02 -1.06835917e-01
4.87241633e-02 4.25222933e-01 -1.18815339e+00 3.14422697e-01
7.80782402e-01 1.89653132e-02 -5.08107483e-01 1.11902654e+00
1.10339664e-01 8.98709357e-01 -4.59807783e-01 -3.39787245e-01
1.95977047e-01 -6.34002462e-02 3.44916582e-01 1.20416462e+00
3.06573957e-01 -1.34120896e-01 -9.00673792e-02 5.26240349e-01
2.33594745e-01 2.39942789e-01 -7.57931709e-01 -5.11463761e-01
3.59008223e-01 1.01678813e+00 -7.72321999e-01 -5.82428798e-02
-5.68659127e-01 6.11116588e-01 -5.20909578e-02 1.67503580e-01
-7.04536438e-01 -3.29326540e-01 4.24000233e-01 2.63612062e-01
4.82184067e-02 2.86560208e-02 -1.72513068e-01 -6.05044663e-01
-4.18783665e-01 -8.97771537e-01 5.18887281e-01 -3.81505787e-01
-1.03191268e+00 4.53196466e-01 2.72792820e-02 -8.14152241e-01
-3.34546566e-01 -5.89189649e-01 -5.86418748e-01 9.13411915e-01
-1.50239062e+00 -1.07321298e+00 -1.55745139e-02 1.60326764e-01
5.15570259e-03 4.47002463e-02 1.18996215e+00 1.54471830e-01
-3.03653747e-01 2.20839813e-01 1.58450365e-01 -9.91086196e-03
4.47114289e-01 -8.09590876e-01 1.92079484e-01 4.35655504e-01
-4.48844165e-01 1.04770935e+00 7.53955603e-01 -8.72020900e-01
-1.22748685e+00 -9.12561715e-01 1.15686035e+00 -4.31679636e-01
5.07247627e-01 4.47185487e-02 -8.17718685e-01 3.32781732e-01
2.64193743e-01 -6.96753919e-01 8.86170566e-01 -1.08432874e-01
-2.61239350e-01 8.18586648e-02 -1.53174639e+00 2.91245192e-01
3.92666131e-01 -3.89762700e-01 -6.58146501e-01 1.86393395e-01
7.34590232e-01 -2.13277750e-02 -7.80748546e-01 6.05753660e-01
8.73774767e-01 -8.51768792e-01 8.85575831e-01 -6.51023626e-01
2.94858038e-01 -5.51857986e-02 1.35881796e-01 -1.01944292e+00
1.25208482e-01 -3.40513438e-01 8.22852105e-02 7.21642733e-01
5.91926873e-01 -9.31128144e-01 4.84535635e-01 3.77650410e-01
2.50907987e-01 -9.80408311e-01 -7.88501680e-01 -5.76559305e-01
1.64179355e-01 -2.23202854e-01 4.47043926e-01 9.25780356e-01
-6.12505227e-02 8.14129785e-02 -4.35059726e-01 1.48046557e-02
3.77756298e-01 -5.93284629e-02 3.52849901e-01 -1.44737709e+00
-1.73352912e-01 -1.37588382e-01 -1.84110239e-01 -1.64908722e-01
-5.24720073e-01 -6.69326961e-01 -3.72910112e-01 -1.79853153e+00
4.76638943e-01 -3.13913018e-01 -5.09417713e-01 4.70246643e-01
-3.45292576e-02 3.90492588e-01 1.35741010e-02 1.61260571e-02
-9.41434130e-02 -4.28330116e-02 1.19644201e+00 6.81470195e-03
-2.27048442e-01 -1.36268079e-01 -7.13171780e-01 5.79150081e-01
1.26873052e+00 -5.31550109e-01 -4.42957044e-01 -2.72168159e-01
5.62129736e-01 -2.48963125e-02 3.97511095e-01 -8.73292744e-01
-2.47018531e-01 -2.07835570e-01 4.91565794e-01 -6.66586697e-01
2.30505794e-01 -7.53189683e-01 3.50462586e-01 1.29447889e+00
1.02577163e-02 2.71574408e-01 2.83307701e-01 3.40869874e-01
2.39812568e-01 1.78452693e-02 6.77861571e-01 4.00760658e-02
-1.59072265e-01 1.27506023e-02 -5.92571735e-01 2.71146685e-01
8.68668497e-01 2.87684007e-03 -6.98194623e-01 -1.82229564e-01
-2.63489962e-01 1.96251839e-01 3.57906759e-01 5.01604259e-01
4.49203074e-01 -5.34526646e-01 -6.89867556e-01 8.95137787e-02
1.22043431e-01 -3.09710443e-01 2.48356655e-01 9.46326613e-01
-9.27133381e-01 7.70483792e-01 -3.79580587e-01 -4.36682612e-01
-1.05418217e+00 5.91021895e-01 2.27646500e-01 -2.54957229e-01
-4.18209881e-01 4.98758018e-01 -2.56244570e-01 -3.00305665e-01
1.08679138e-01 -1.59256548e-01 -4.15686816e-01 5.86687736e-02
6.32351577e-01 7.81062484e-01 -1.11515215e-02 -4.73496318e-01
-5.47909796e-01 2.44286940e-01 1.43708974e-01 1.58010378e-01
1.31120455e+00 2.42516965e-01 -1.70624956e-01 2.49945611e-01
9.02834833e-01 -5.59521802e-02 -5.32884359e-01 5.39833546e-01
-5.54319024e-02 1.34166880e-02 -2.43881911e-01 -1.14326656e+00
-5.64694822e-01 8.47976029e-01 1.18134153e+00 3.25836003e-01
9.29688573e-01 -1.31413996e-01 8.13152134e-01 4.22881395e-01
3.52719516e-01 -7.03792632e-01 -1.59401506e-01 2.73089796e-01
4.80687857e-01 -1.12240791e+00 -1.78585127e-01 2.99250513e-01
-4.45242167e-01 5.37972987e-01 1.46086752e-01 7.41870701e-03
5.97575366e-01 6.65093213e-02 2.88908482e-01 -3.54547292e-01
-5.70683777e-01 -1.92088217e-01 1.35748670e-01 8.06696475e-01
5.73371410e-01 3.64784896e-01 -9.24465120e-01 4.19227481e-01
7.53498301e-02 1.25931785e-01 1.57787755e-01 7.09587932e-01
-6.61399901e-01 -9.84756172e-01 -3.11509997e-01 9.61669624e-01
-8.15070271e-01 -2.63079554e-02 -5.72516620e-01 9.29525316e-01
3.94400656e-01 1.21692836e+00 2.28908788e-02 -1.92198008e-01
1.95259545e-02 1.76281527e-01 2.40707874e-01 -3.87280077e-01
-7.39929974e-01 -9.97226611e-02 1.41663328e-01 -1.63808316e-01
-5.38608968e-01 -3.91207635e-01 -1.32487321e+00 -2.84971625e-01
-2.16679141e-01 1.90952092e-01 8.08118939e-01 9.38902199e-01
3.00728977e-01 6.42211884e-02 4.02299851e-01 -3.90824348e-01
-4.13297832e-01 -8.89719129e-01 -3.33012372e-01 2.43394211e-01
4.92396593e-01 -4.06425804e-01 -1.90004021e-01 -2.62560546e-01] | [15.548295974731445, -1.6508674621582031] |
6ff5664c-6660-496c-9787-e1e3e813494b | gu-yi-yu-qian-tao-ming-ming-shi-ti-shi-bie | null | null | https://aclanthology.org/2022.ccl-1.37 | https://aclanthology.org/2022.ccl-1.37.pdf | 古汉语嵌套命名实体识别数据集的构建和应用研究(Construction and application of classical Chinese nested named entity recognition data set) | “本文聚焦研究较少的古汉语嵌套命名实体识别任务,以《史记》作为原始语料,针对古文意义丰富而导致的实体分类模糊问题,分别构建了基于字词本义和语境义2个标注标准的古汉语嵌套命名实体数据集,探讨了数据集的实体分类原则和标注格式,并用RoBERTa-classical-chinese+GlobalPointer模型进行对比试验,标准一数据集F1值为80.42%,标准二F1值为77.43%,以此确定了数据集的标注标准。之后对比了六种预训练模型配合GlobalPointer在古汉语嵌套命名实体识别任务上的表现。最终试验结果:RoBERTa-classical-chinese模型F1值为84.71%,表现最好。” | ['Genhui Liu', 'Jinzhu Liu', 'Zhiqiang Xie'] | null | null | null | null | ccl-2022-10 | ['nested-named-entity-recognition'] | ['natural-language-processing'] | [-1.05992138e+00 -8.14990282e-01 2.08533540e-01 2.48288482e-01
-6.13713920e-01 -8.59294355e-01 3.60294133e-01 7.90150642e-01
-2.95110345e-01 3.93532664e-01 7.95863569e-01 -1.03492007e-01
-1.93099096e-01 -8.90956163e-01 -5.29613435e-01 -9.20150757e-01
-4.87034738e-01 1.38527286e+00 3.31108809e-01 -9.41700161e-01
6.87802494e-01 5.52671313e-01 -6.58839583e-01 -4.82489966e-04
1.01154649e+00 4.69045162e-01 8.15085709e-01 5.60465336e-01
-4.42329906e-02 7.28211403e-01 -1.31595492e+00 3.14988881e-01
-4.80656832e-01 -1.34791806e-01 -4.68052864e-01 4.84445304e-01
-4.71198380e-01 -9.59460437e-01 -1.23876595e+00 1.12786388e+00
1.64650217e-01 -9.32452083e-02 9.85819101e-01 -3.50766391e-01
-1.06154716e+00 7.43020415e-01 -1.79317653e-01 5.21384954e-01
-2.38834217e-01 1.05184577e-01 5.48805594e-01 -6.51822090e-01
5.18596292e-01 1.49476838e+00 1.54670680e+00 3.00089151e-01
-1.25910294e+00 -1.04026341e+00 7.34306693e-01 -4.47766334e-01
-1.79508829e+00 4.97714669e-01 3.34389985e-01 -1.76269472e-01
8.80367994e-01 5.20196736e-01 1.11799562e+00 1.74244022e+00
4.52400237e-01 1.28552246e+00 1.40812838e+00 5.53824067e-01
1.31175548e-01 1.08523262e+00 -3.13151747e-01 -3.03913713e-01
6.44903898e-01 1.15505733e-01 1.44450083e-01 6.60456002e-01
5.37029743e-01 7.77413473e-02 1.17609855e-02 1.45632362e+00
-1.49044275e+00 1.44268787e+00 3.31490159e-01 1.60854757e+00
3.75668742e-02 -1.02279341e+00 -5.49926996e-01 5.04686952e-01
2.29768053e-01 1.39331490e-01 8.17875713e-02 1.78690404e-01
-1.75891697e-01 4.07407761e-01 8.56515288e-01 1.34987819e+00
1.04776382e+00 4.79124159e-01 -3.84880900e-01 6.26190186e-01
1.34639037e+00 9.29103136e-01 1.38604689e+00 -6.90274060e-01
-7.17419684e-01 5.36872089e-01 -4.47817594e-01 -1.53296399e+00
-9.44925725e-01 -1.54853892e+00 -7.10186720e-01 -9.79312479e-01
-2.91722044e-02 -4.50252950e-01 -6.33490562e-01 1.10378635e+00
1.99230686e-01 -9.50822830e-01 2.83037812e-01 7.78596640e-01
5.40438652e-01 1.59187663e+00 1.01172412e-02 -1.88319504e-01
1.35836339e+00 -4.32094038e-01 -4.13238853e-01 -2.13823523e-02
-3.32959265e-01 -1.77572680e+00 1.73342836e+00 2.30732977e-01
-7.43917763e-01 -2.49822751e-01 -1.39656317e+00 3.66396666e-01
3.56310122e-02 -3.57003361e-02 2.72813320e-01 1.37973237e+00
-1.30031049e+00 1.08973406e-01 -8.29598010e-01 -1.25062275e+00
-2.16092631e-01 3.55324209e-01 3.28802407e-01 -7.00874776e-02
-1.11625051e+00 8.43727708e-01 4.90122959e-02 4.40136999e-01
-1.56717408e+00 -5.29121518e-01 -4.68083173e-01 7.99395815e-02
9.76003051e-01 2.03981802e-01 1.88852191e+00 -1.05970418e+00
-9.26048815e-01 3.99988174e-01 -5.24148405e-01 7.12552667e-02
3.06951702e-01 2.70855129e-01 -2.11929515e-01 1.29759923e-01
2.06555471e-01 3.95724595e-01 3.54561687e-01 -1.94027913e+00
-6.52994633e-01 -1.05468774e+00 2.23557636e-01 -3.79705690e-02
2.64874995e-01 1.22992611e+00 -4.31865066e-01 -1.05088782e+00
-1.08397426e-02 -2.04639387e+00 -1.35471523e-01 -1.59931827e+00
-6.04768217e-01 -5.87595487e-03 7.92205811e-01 -1.78747803e-01
1.23417747e+00 -2.12721252e+00 -8.89823794e-01 8.99560094e-01
8.96009654e-02 7.35365629e-01 1.98306546e-01 2.03073096e+00
6.21318936e-01 1.37608016e+00 -8.69631052e-01 7.69646227e-01
-1.35489509e-01 -7.00940192e-02 8.99202824e-01 -2.17222914e-01
-1.95028991e-01 -3.87999192e-02 -4.69503582e-01 2.81001329e-02
1.24011822e-01 4.45224851e-01 6.77971244e-01 1.83924288e-01
4.13614929e-01 7.69824237e-02 -1.38289347e-01 1.15277028e+00
1.34964025e+00 -3.11884791e-01 7.84827411e-01 3.24868202e-01
-1.01692736e+00 -2.36705318e-01 -2.00503588e+00 1.28762758e+00
1.37856752e-01 5.59323370e-01 3.47644866e-01 -7.25013316e-01
1.29158700e+00 1.17360497e+00 8.05568814e-01 -3.25748235e-01
3.31477255e-01 8.54012728e-01 5.28764606e-01 -1.24944612e-01
1.28876281e+00 -9.78242874e-01 4.22827750e-01 3.95719379e-01
-4.68398005e-01 8.09725225e-02 3.87050688e-01 -6.62887633e-01
1.04899883e+00 -7.64479876e-01 -1.81940675e-01 -9.52963293e-01
5.36072791e-01 -8.60511046e-03 3.10472727e-01 9.86919925e-02
-5.30062735e-01 6.77758873e-01 -4.40436274e-01 5.02570748e-01
-1.54314905e-01 -1.40786731e+00 -1.77378371e-01 1.00286162e+00
3.84159870e-02 -1.21213877e+00 -7.00823545e-01 -6.57627821e-01
-6.57386899e-01 8.17718685e-01 2.55362809e-01 -1.48207590e-01
-5.98873019e-01 -5.71146786e-01 1.64607525e-01 -1.39801785e-01
1.59109688e+00 -1.10709298e+00 7.87552446e-03 -1.39931709e-01
-1.23618729e-01 -4.92087513e-01 -9.60014224e-01 -6.25363410e-01
-1.66572726e+00 -7.42589295e-01 5.05349301e-02 -1.38598907e+00
9.40738678e-01 2.71378160e-01 6.51576161e-01 6.02894247e-01
2.50271469e-01 2.62184232e-01 -3.28144491e-01 -8.15731108e-01
-1.84182018e-01 7.29842961e-01 -1.37556449e-01 8.70609999e-01
7.06365585e-01 9.44400206e-02 -1.11271274e+00 7.05071807e-01
-1.25517845e+00 -6.96865395e-02 3.87059003e-01 -1.00969873e-01
5.63054025e-01 2.50982076e-01 1.62650466e-01 -1.01582575e+00
1.09416544e+00 2.08283260e-01 -1.06583941e+00 2.67673403e-01
-1.16567445e+00 -9.83425438e-01 1.36141956e+00 -5.04774988e-01
-1.03284943e+00 -1.11271954e+00 -5.01793861e-01 -2.71435589e-01
1.39893264e-01 4.77552861e-01 -2.34487057e-01 6.76399708e-01
-2.54721791e-01 3.72675240e-01 2.92447865e-01 -9.96135712e-01
-1.43274739e-01 7.11670935e-01 3.38852167e-01 -8.87044370e-01
7.63295829e-01 9.34145525e-02 -3.10569525e-01 -7.31194198e-01
3.05733532e-01 -3.09226334e-01 -5.23947120e-01 6.85329735e-01
9.42232430e-01 -1.30588746e+00 -2.52909511e-01 7.15746760e-01
-5.09985089e-01 4.64676797e-01 -3.35885763e-01 1.86002457e+00
-9.83651057e-02 -3.90514642e-01 -7.05746830e-01 -6.07580543e-01
3.20866778e-02 -2.40245056e+00 1.17270410e-01 2.03197384e+00
-1.02147199e-01 -8.51778269e-01 -2.90806051e-02 8.89581069e-02
9.64206696e-01 -1.88128710e-01 9.02004361e-01 -9.20507550e-01
-9.86322224e-01 -3.09009671e-01 1.45648465e-01 9.97400582e-01
8.67144167e-01 6.28843546e-01 3.66276473e-01 -7.24147931e-02
-3.06599408e-01 6.22343063e-01 3.55019480e-01 5.88397026e-01
3.04630511e-02 -5.81214547e-01 1.30643293e-01 6.12204969e-01
2.05564690e+00 1.21307838e+00 1.01668513e+00 1.79802203e+00
-3.97550464e-01 -4.15474512e-02 2.91582614e-01 4.08653468e-01
4.85516757e-01 8.92599881e-01 -4.36194003e-01 1.13100350e+00
-9.24283624e-01 -1.74474657e-01 9.73039687e-01 1.33738935e+00
-1.23496372e-02 2.68667608e-01 -1.92678839e-01 9.10183489e-01
-2.13640666e+00 -1.78342497e+00 5.21863252e-02 2.63618946e+00
5.24585366e-01 -4.12450880e-01 9.06506926e-02 -1.98452100e-01
1.37126112e+00 -1.71105191e-01 3.33529830e-01 -4.83492911e-01
4.13211167e-01 1.07000148e+00 5.35010874e-01 3.30817699e-01
-7.29182601e-01 3.17392319e-01 1.25760269e+00 7.73732781e-01
-2.24897003e+00 -7.46259540e-02 5.57569146e-01 -1.03095591e+00
-7.52675653e-01 8.54219675e-01 -6.48609161e-01 1.96559036e+00
1.16010177e+00 -1.24105442e+00 7.24120885e-02 1.82823882e-01
1.02966332e+00 3.51738125e-01 -3.00976634e-01 1.42359912e+00
-9.59222671e-03 -1.71432757e+00 -4.51154560e-02 2.89583683e-01
7.55787849e-01 -4.72586095e-01 -1.77377433e-01 4.31212634e-01
1.01067223e-01 -9.99074876e-01 1.41482413e+00 9.70781446e-02
2.34726831e-01 -1.10311747e+00 1.38157201e+00 1.71802398e-02
-1.23057270e+00 -4.46450919e-01 7.71411210e-02 -1.47504479e-01
4.68048453e-01 -9.38934460e-02 1.12289645e-01 6.33646905e-01
-6.19864836e-02 -5.73658235e-02 -6.10987306e-01 2.19900417e+00
-9.81423974e-01 3.04502875e-01 -5.16076945e-02 -1.44300258e+00
8.86260092e-01 -6.06413007e-01 8.51298690e-01 1.48662710e+00
4.71174419e-01 9.71167743e-01 2.58111298e-01 9.75102067e-01
-5.52610397e-01 3.69183928e-01 2.06797466e-01 3.75039876e-01
8.41088235e-01 8.71202290e-01 -1.34401619e+00 -8.10270846e-01
-3.23630631e-01 -7.77774513e-01 -3.25399786e-01 -4.87671852e-01
-1.80484086e-01 -9.27867770e-01 -3.71761620e-02 4.14005578e-01
-4.61986512e-01 -1.03337550e+00 3.42548400e-01 -4.36414480e-01
-3.53845179e-01 -1.18737674e+00 7.92660788e-02 -5.30966163e-01
-9.42005455e-01 1.99007854e-01 5.87054849e-01 -1.02020562e+00
4.21127379e-01 -1.36661184e+00 -7.99687356e-02 7.90096402e-01
-2.16663942e-01 -8.42369378e-01 -1.51572764e-01 -7.76110739e-02
1.34466755e+00 1.56636104e-01 8.03862214e-01 8.71159613e-01
-7.91501641e-01 6.06533051e-01 1.47490752e+00 5.15103042e-01
4.50631320e-01 -2.00812840e+00 -9.39297855e-01 4.55177665e-01
6.15211964e-01 6.04690135e-01 -1.21153481e-02 -5.61252594e-01
-1.35660529e+00 -6.75202608e-01 1.00323737e+00 6.14509225e-01
3.72095883e-01 -4.16749626e-01 -2.45436579e-01 1.27138913e+00
7.77501702e-01 -4.01933789e-01 5.93300581e-01 -2.98456252e-01
5.39419591e-01 -1.08292067e+00 -1.46007800e+00 1.16190851e+00
4.20498580e-01 1.17792986e-01 -5.72402000e-01 2.97815204e-02
4.06013936e-01 -9.49949086e-01 -1.44364333e+00 -1.56707689e-02
1.47223997e+00 -7.14618981e-01 -3.50305915e-01 -8.84313807e-02
2.63862554e-02 -3.72415841e-01 -2.63184756e-01 -1.26417315e+00
3.24425399e-01 -1.16009402e+00 1.43736386e+00 1.21665287e+00
5.17671525e-01 -1.22449768e+00 8.07513893e-01 1.39299500e+00
-9.95176136e-01 -4.05348390e-01 -1.79622686e+00 -9.94105816e-01
5.61612546e-01 -6.93333894e-02 4.05068249e-01 -4.62289974e-02
-4.35944796e-01 1.76640257e-01 3.02883089e-01 -2.63521969e-01
-3.18838626e-01 -8.19385350e-01 4.04795676e-01 -9.20435488e-01
7.91994095e-01 -6.51211679e-01 -2.13590696e-01 -4.18416113e-01
-1.11852777e+00 -1.84168503e-01 -1.75941050e+00 -1.79057121e+00
3.04174185e-01 -3.71907502e-01 -1.10838592e-01 -3.96393150e-01
1.07337236e+00 1.42449811e-01 1.08819711e+00 9.28040326e-01
2.04193980e-01 -2.41970390e-01 1.31183910e+00 -3.63885015e-01
-6.39875591e-01 2.74604648e-01 -9.44086730e-01 6.80398345e-01
5.92775643e-01 -2.24098954e-02 -1.68386638e-01 -7.45273948e-01
1.95493862e-01 -1.37376726e+00 -1.10417676e+00 -9.16682363e-01
8.23282748e-02 -1.00391142e-01 2.00401872e-01 4.96579885e-01
2.47911736e-01 -5.86006463e-01 5.90486288e-01 4.11444038e-01
1.37887582e-01 1.15814734e+00 1.69771478e-01 -7.70666420e-01
-2.13271379e-01 -1.05928349e+00 5.96664906e-01 -5.03800213e-01
-2.97659785e-01 -4.01349723e-01 -8.08832943e-01 -5.41483819e-01
1.30723464e+00 4.19289954e-02 -9.66667712e-01 8.64536017e-02
-6.56234324e-01 -3.81069392e-01 5.32905400e-01 7.56408930e-01
1.19378901e+00 -1.57577193e+00 -2.11176991e+00 4.61797006e-02
-5.58834612e-01 1.10009693e-01 2.57008642e-01 1.29183388e+00
-9.44510460e-01 6.63053155e-01 -5.04475236e-01 -4.38049436e-01
-2.18213797e+00 5.07790267e-01 -3.63193333e-01 -6.59133643e-02
-3.72351170e-01 4.76724744e-01 -2.74421543e-01 2.14084104e-01
-7.45214894e-02 -5.41538119e-01 -3.82214457e-01 1.86147749e-01
3.22185338e-01 4.28516626e-01 -2.03474164e-02 -1.40463376e+00
-6.87002480e-01 1.48103023e+00 -7.85330772e-01 -4.25775796e-01
2.03985885e-01 -1.72421467e-02 -4.65717375e-01 2.71890610e-02
1.15472066e+00 1.06611884e+00 6.99561555e-03 8.52127671e-01
-6.69835031e-01 -5.29021144e-01 -8.70598614e-01 -1.70599639e+00
-5.05695045e-01 5.00761449e-01 3.03729266e-01 -5.81720531e-01
9.03543413e-01 -4.01764750e-01 1.13274157e+00 1.75595835e-01
1.41986787e+00 -2.07337618e+00 -8.97150457e-01 1.19303846e+00
9.07478273e-01 -1.44591630e+00 2.39255920e-01 -1.01648070e-01
9.60114673e-02 1.68427825e+00 5.85348547e-01 -5.47810853e-01
9.84617054e-01 -2.75747031e-01 -2.17261061e-01 -4.41907495e-02
-9.60877910e-02 -6.03104532e-01 -8.01639408e-02 4.00870293e-01
1.04357934e+00 2.13039130e-01 -1.59635377e+00 5.41789889e-01
-1.12843168e+00 1.46871477e-01 1.16831875e+00 1.05019820e+00
-6.38596296e-01 -6.03110015e-01 -1.60086274e+00 3.57704043e-01
-9.79048252e-01 7.70541608e-01 1.22459017e-01 1.43569565e+00
4.70169187e-01 2.50903511e+00 -2.97574908e-01 -1.08131254e+00
1.72862458e+00 -3.30458015e-01 5.40539443e-01 -8.59612450e-02
-1.66057241e+00 -8.47868845e-02 1.60215795e-01 5.07118165e-01
-2.63780862e-01 -1.88426852e-01 -1.09476662e+00 -1.32826829e+00
-2.53266543e-01 2.19511032e+00 1.27391565e+00 5.15313685e-01
-2.37978384e-01 7.60878801e-01 9.75800574e-01 -6.21785641e-01
-9.75690961e-01 -1.18480289e+00 -1.02354181e+00 -5.32528937e-01
-7.01994970e-02 -4.12339658e-01 -1.03262019e+00 -7.02671528e-01] | [-3.316066265106201, 6.907612323760986] |
958ab2a4-a697-4c94-9776-a3cd130f6ca1 | gaussian-processes-with-context-supported | 1703.08653 | null | http://arxiv.org/abs/1703.08653v3 | http://arxiv.org/pdf/1703.08653v3.pdf | Gaussian Processes with Context-Supported Priors for Active Object Localization | We devise an algorithm using a Bayesian optimization framework in conjunction
with contextual visual data for the efficient localization of objects in still
images. Recent research has demonstrated substantial progress in object
localization and related tasks for computer vision. However, many current
state-of-the-art object localization procedures still suffer from inaccuracy
and inefficiency, in addition to failing to provide a principled and
interpretable system amenable to high-level vision tasks. We address these
issues with the current research.
Our method encompasses an active search procedure that uses contextual data
to generate initial bounding-box proposals for a target object. We train a
convolutional neural network to approximate an offset distance from the target
object. Next, we use a Gaussian Process to model this offset response signal
over the search space of the target. We then employ a Bayesian active search
for accurate localization of the target.
In experiments, we compare our approach to a state-of-theart bounding-box
regression method for a challenging pedestrian localization task. Our method
exhibits a substantial improvement over this baseline regression method. | ['Melanie Mitchell', 'Jordan Witte', 'Anthony D. Rhodes', 'Bruno Jedynak'] | 2017-03-25 | null | null | null | null | ['active-object-localization'] | ['computer-vision'] | [ 1.40057430e-01 -3.05168301e-01 -1.93275083e-02 -7.05250621e-01
-1.35174322e+00 -2.82237679e-01 6.04342282e-01 1.83724642e-01
-8.10976028e-01 6.71935916e-01 -2.74067149e-02 1.82392243e-02
1.72209755e-01 -3.22164476e-01 -8.28306675e-01 -8.23818922e-01
6.37216866e-02 5.67320585e-01 7.51346707e-01 3.26282561e-01
4.56009924e-01 6.81174040e-01 -1.31343997e+00 -3.02769821e-02
6.00541413e-01 9.64190841e-01 4.48679656e-01 8.55327547e-01
1.85766578e-01 5.39497256e-01 -7.85040855e-01 2.33976226e-02
2.55120486e-01 -2.15727855e-02 -3.67082715e-01 6.66391179e-02
8.10456216e-01 -3.28131825e-01 -1.89735085e-01 1.01375520e+00
6.51890337e-01 5.01487195e-01 6.19815707e-01 -1.07960820e+00
-4.26850468e-01 2.49545500e-02 -7.01214373e-01 4.22658175e-01
3.34462076e-01 2.90301561e-01 8.12277973e-01 -1.24528372e+00
3.50238413e-01 1.37123704e+00 8.89858305e-01 4.10494775e-01
-1.54607427e+00 -2.95656323e-01 3.03002447e-01 1.44706801e-01
-1.58564484e+00 -8.12183499e-01 6.48879647e-01 -5.33165812e-01
8.58155310e-01 -5.30229397e-02 5.00533462e-01 9.70086157e-01
1.64385274e-01 8.49423826e-01 1.03984129e+00 -7.14205146e-01
4.49634820e-01 8.60697404e-02 8.65317509e-02 8.20467949e-01
2.31607735e-01 3.89623731e-01 -5.65366030e-01 -3.87504220e-01
7.41426766e-01 -1.84374869e-01 -2.13548709e-02 -7.92661011e-01
-1.10515046e+00 9.18125689e-01 8.24125886e-01 -2.34191060e-01
-4.66971695e-01 6.65424109e-01 -9.10378769e-02 -6.33986771e-01
5.43923974e-01 2.08975643e-01 -2.19470650e-01 7.54432827e-02
-1.20186150e+00 4.56800044e-01 6.54308498e-01 9.01059985e-01
6.26000047e-01 -1.68993011e-01 -2.81311572e-01 6.52320623e-01
1.00677049e+00 4.48871970e-01 -1.27674028e-01 -9.33882952e-01
2.91092396e-01 2.73185670e-01 5.63278615e-01 -8.21793497e-01
-1.79232791e-01 -2.91964710e-01 7.41345529e-03 6.94796026e-01
6.85089648e-01 2.14446839e-02 -1.01218796e+00 1.32368016e+00
4.89209175e-01 2.43892252e-01 -3.73416424e-01 1.13081539e+00
6.45646691e-01 6.78385258e-01 2.75270581e-01 1.58401132e-01
1.23998713e+00 -1.23564768e+00 -3.52894962e-01 -6.50247753e-01
1.63902834e-01 -9.02542531e-01 6.46001518e-01 2.39888251e-01
-9.62411225e-01 -6.76239789e-01 -1.11959147e+00 -2.21372787e-02
-2.26503968e-01 4.99975890e-01 5.88766396e-01 6.12809658e-01
-9.37684536e-01 2.48149693e-01 -1.22136748e+00 -4.76173729e-01
6.23384893e-01 3.93304080e-01 -2.50999898e-01 1.53035119e-01
-4.02876168e-01 1.22798932e+00 3.56577963e-01 1.97810724e-01
-1.06476068e+00 -5.69016695e-01 -9.80010986e-01 -3.18892211e-01
4.40929294e-01 -5.79646945e-01 1.45386934e+00 -8.36857677e-01
-1.40082991e+00 7.85704315e-01 -4.94552493e-01 -6.05318367e-01
3.58347148e-01 -5.79618812e-01 6.85961470e-02 4.73032221e-02
4.33012605e-01 9.53234553e-01 8.83493900e-01 -1.44633663e+00
-8.82360399e-01 -3.33390743e-01 -2.06988707e-01 1.69391274e-01
3.95199865e-01 4.93965507e-01 -7.19453633e-01 -4.69964892e-01
3.36189777e-01 -8.03029239e-01 -6.90673828e-01 6.17706954e-01
-3.15800518e-01 -2.22593755e-01 8.24137390e-01 -4.68202889e-01
7.70604610e-01 -1.98425961e+00 -1.15215831e-01 1.17670432e-01
1.07470237e-01 -1.87497698e-02 7.24585401e-03 -6.81934431e-02
3.05914223e-01 -2.35229060e-01 3.18089873e-02 -8.80836904e-01
-4.03661728e-02 -9.87053439e-02 -4.10229772e-01 9.37136114e-01
3.73662889e-01 9.51031923e-01 -9.43896770e-01 -5.72003961e-01
3.37626696e-01 6.72218263e-01 -4.11139786e-01 3.35303664e-01
-2.13873222e-01 4.62208480e-01 -4.90633041e-01 8.87687802e-01
4.21340466e-01 -1.57591268e-01 -5.21654308e-01 -1.87960669e-01
-4.73966956e-01 1.68787971e-01 -1.19289505e+00 1.68285811e+00
-1.30415782e-01 9.35021222e-01 7.80872181e-02 -9.12377417e-01
1.07723475e+00 -1.29190475e-01 4.08313304e-01 -2.93490976e-01
8.81272629e-02 2.80771911e-01 -6.28200173e-02 -2.37303749e-01
6.06808543e-01 3.25650424e-01 1.82085231e-01 2.04908960e-02
1.58320889e-01 -5.96746728e-02 -1.21649832e-01 -1.09577633e-01
9.65707839e-01 7.37490416e-01 4.95808333e-01 -2.98410714e-01
3.12234640e-01 2.57005572e-01 3.15726548e-01 1.24349165e+00
-6.93741381e-01 7.96590626e-01 -2.04617813e-01 -5.04293621e-01
-1.08022583e+00 -1.18317199e+00 -2.16704488e-01 1.07883656e+00
3.02433550e-01 -1.36352554e-01 -6.10644400e-01 -6.24076843e-01
1.99392461e-03 5.51486015e-01 -4.52392042e-01 2.29383901e-01
-8.14679801e-01 -8.77787590e-01 3.22415709e-01 7.97689140e-01
4.79502439e-01 -8.14370871e-01 -1.21239567e+00 3.74540031e-01
-3.28658521e-02 -1.11158597e+00 -4.33342189e-01 3.34007055e-01
-6.49966717e-01 -8.18502069e-01 -6.65280640e-01 -7.39298224e-01
8.63062978e-01 3.07502389e-01 9.98104274e-01 -1.55403644e-01
-7.56049156e-01 6.09376073e-01 -3.54973190e-02 -7.15430260e-01
-6.96265623e-02 -3.77449036e-01 4.03965376e-02 -1.74159184e-01
6.41927958e-01 8.30917209e-02 -8.46792459e-01 4.74694103e-01
-1.19241439e-01 -2.91887343e-01 5.05020499e-01 7.55938768e-01
8.79591703e-01 -2.12631643e-01 4.45397496e-02 -1.92186460e-02
2.84102499e-01 -2.06936061e-01 -1.29238617e+00 1.36888340e-01
-1.24953769e-01 7.39506409e-02 -5.70952296e-02 -6.77197397e-01
-9.78616238e-01 7.48468816e-01 -8.10333900e-03 -1.82539374e-01
-3.96517366e-01 9.98248979e-02 -4.81669605e-02 -6.21379912e-01
9.41885650e-01 -6.74468204e-02 -3.05322200e-01 -3.31211269e-01
3.69760871e-01 5.63660443e-01 7.46950269e-01 -6.92872703e-01
7.75983274e-01 7.42924511e-01 1.04646757e-01 -7.62204647e-01
-1.03175104e+00 -8.96497011e-01 -9.46944475e-01 -3.13758165e-01
9.67716932e-01 -8.59522283e-01 -6.78276420e-01 1.97454721e-01
-1.45894480e+00 -3.60613793e-01 -1.15522601e-01 7.74604082e-01
-7.56930113e-01 2.62202621e-01 -3.20779383e-02 -1.23458815e+00
-2.63499375e-02 -1.38529146e+00 1.44298875e+00 4.09290999e-01
-1.36629626e-01 -7.87801564e-01 2.67924547e-01 2.81695515e-01
4.78343338e-01 1.62329614e-01 1.19800113e-01 -6.70386612e-01
-1.19544053e+00 -4.47209090e-01 -3.15912664e-01 -8.28785822e-02
-2.18866035e-01 1.70176420e-02 -9.79261994e-01 -2.69620597e-01
2.07656454e-02 -1.95734039e-01 8.65535796e-01 9.61759746e-01
8.43843579e-01 6.16164133e-02 -7.85764992e-01 6.58532858e-01
1.38898921e+00 3.09265584e-01 3.62173826e-01 6.30517304e-01
4.01148528e-01 4.48651075e-01 9.05336142e-01 1.95282176e-01
1.90646037e-01 1.12159002e+00 3.85533869e-01 -1.54805690e-01
-1.48352817e-01 -1.56049341e-01 1.56923398e-01 -2.41876110e-01
-2.28342451e-02 -1.18321449e-01 -1.15486431e+00 5.61284006e-01
-2.08949518e+00 -8.01574409e-01 -1.19908964e-02 2.42434645e+00
5.19662142e-01 3.36518198e-01 1.35514066e-01 -3.29909772e-01
7.26921380e-01 -1.74311295e-01 -4.55511630e-01 1.84306458e-01
2.28971928e-01 2.36258726e-03 7.77718127e-01 6.60687327e-01
-1.68105698e+00 1.05764198e+00 6.83476448e+00 5.40033221e-01
-8.64327729e-01 1.48229048e-01 3.59529227e-01 2.25948513e-01
6.29096985e-01 2.83194661e-01 -1.48791826e+00 1.96045190e-01
6.24955654e-01 3.46786439e-01 7.88718462e-02 1.26436758e+00
3.77756536e-01 -5.55824518e-01 -1.31860340e+00 1.04011011e+00
3.48791510e-01 -1.08250833e+00 -5.90873003e-01 -3.62029150e-02
4.70760018e-01 1.35422498e-01 -3.39219831e-02 5.72432689e-02
3.10188740e-01 -9.54673231e-01 1.13323486e+00 6.04435563e-01
1.93914905e-01 -2.99152344e-01 5.12163877e-01 3.96766663e-01
-1.16926515e+00 6.44748798e-03 -4.82199967e-01 9.04803574e-02
3.87291610e-01 2.31633663e-01 -1.10185981e+00 -1.77671880e-01
7.79697239e-01 1.72617808e-01 -7.62647808e-01 1.91908801e+00
-4.18139368e-01 4.26562995e-01 -5.79828560e-01 -3.27573419e-01
1.96223989e-01 1.33622184e-01 9.25489664e-01 1.38205636e+00
1.06586076e-01 -1.02793619e-01 5.75487614e-01 1.20162284e+00
3.30921948e-01 -2.91930512e-02 -3.61823529e-01 4.16747183e-01
5.36800742e-01 1.08337653e+00 -1.04619682e+00 -3.01134914e-01
-3.10329467e-01 8.53086770e-01 4.24355149e-01 4.93262440e-01
-9.25347030e-01 -3.13026607e-01 2.15334252e-01 1.26706120e-02
5.95336199e-01 -6.37698650e-01 -1.98470190e-01 -1.01610625e+00
1.09549034e-02 -4.20476556e-01 -4.53136526e-02 -9.21073616e-01
-1.21072996e+00 3.88041466e-01 3.89861852e-01 -8.62199783e-01
-2.80487835e-01 -7.87631214e-01 -5.77710032e-01 1.10205722e+00
-1.25821602e+00 -1.23799324e+00 -3.61554623e-01 2.86252737e-01
7.85409689e-01 -1.98912919e-01 6.70833766e-01 7.73697579e-03
-2.05026641e-01 2.96686262e-01 -8.98792893e-02 3.33675593e-01
7.46534944e-01 -1.33751106e+00 6.55099213e-01 1.24492800e+00
2.63056308e-01 8.48604977e-01 9.10151184e-01 -7.68469691e-01
-1.13363934e+00 -8.94577742e-01 5.79898000e-01 -9.84305263e-01
6.91696584e-01 -6.93974435e-01 -3.76899898e-01 7.65707672e-01
-1.27702519e-01 5.14491796e-01 3.45754325e-01 1.22285880e-01
-2.66780883e-01 -3.07116415e-02 -1.03881979e+00 6.15996718e-01
6.52485013e-01 -3.27460527e-01 -8.41273844e-01 2.38557100e-01
3.67780447e-01 -5.20732641e-01 -2.46618211e-01 3.85058075e-01
6.31471038e-01 -7.32166171e-01 1.34724784e+00 -3.78339857e-01
-2.58119732e-01 -7.12907314e-01 -2.43924335e-01 -9.89042878e-01
-3.98353398e-01 -4.66555327e-01 -3.47744435e-01 8.71996760e-01
2.00340956e-01 -3.08967769e-01 9.48780000e-01 6.51211917e-01
-9.61619914e-02 -5.81126273e-01 -1.11560535e+00 -6.27764761e-01
-3.51381719e-01 -4.63251263e-01 -7.54645243e-02 -5.11258915e-02
-4.64622498e-01 2.57158220e-01 -1.46145821e-01 4.09041852e-01
1.28630805e+00 -1.08639166e-01 9.93644834e-01 -1.07492852e+00
-2.07370013e-01 -3.87207389e-01 -7.85784721e-01 -1.53561711e+00
1.87947638e-02 -2.04197735e-01 9.19917107e-01 -1.58995724e+00
2.34564081e-01 -5.08131325e-01 -4.24643829e-02 2.44666815e-01
-2.29443774e-01 4.28141207e-01 5.47487885e-02 2.29277730e-01
-8.17347288e-01 2.33472005e-01 6.09963834e-01 -2.31703043e-01
-3.03818882e-01 4.06727672e-01 -2.89700329e-01 9.59671140e-01
5.71775138e-01 -6.83099210e-01 4.17290401e-04 -2.85163254e-01
-1.76966235e-01 -2.64595568e-01 9.90209401e-01 -1.12618971e+00
7.21850753e-01 -3.08123697e-02 8.58938217e-01 -1.03076899e+00
7.90146232e-01 -8.88707161e-01 -2.76460111e-01 3.37736577e-01
-2.74266034e-01 -1.62892848e-01 1.84084177e-01 6.72299564e-01
1.29073828e-01 -4.75702435e-01 9.87521887e-01 -1.25312090e-01
-8.91541302e-01 2.41334662e-01 -4.80007231e-01 -2.75629401e-01
1.03872633e+00 -3.76723707e-01 -2.02288166e-01 -8.86518955e-02
-7.73598015e-01 -7.74469674e-02 2.82439977e-01 4.00554866e-01
8.08583260e-01 -1.15158570e+00 -6.73625708e-01 1.34722114e-01
1.13578521e-01 2.40939371e-02 -4.13253486e-01 8.45685959e-01
-6.06012106e-01 5.25141120e-01 1.44677475e-01 -1.16013849e+00
-1.39565086e+00 5.18287182e-01 5.21472991e-01 1.25348553e-01
-4.97170150e-01 1.03908062e+00 7.19902515e-02 -1.47113815e-01
6.65271163e-01 -3.16120625e-01 7.14457631e-02 -3.41396004e-01
6.94159448e-01 3.98016572e-01 -2.26982549e-01 -8.42296660e-01
-5.65230131e-01 6.18417382e-01 6.62636459e-02 -3.93480003e-01
1.07823789e+00 -1.66429967e-01 9.37896371e-02 3.87170345e-01
8.75213206e-01 4.39565219e-02 -1.69402528e+00 -2.24037886e-01
3.17886531e-01 -7.71634638e-01 2.71459818e-01 -6.41227186e-01
-3.26242924e-01 7.00823545e-01 9.09496486e-01 -2.53448188e-01
5.66762865e-01 1.82868376e-01 5.28841913e-02 5.76644003e-01
5.24663150e-01 -9.16783154e-01 2.58110374e-01 3.15710515e-01
7.46708870e-01 -1.65970087e+00 3.83613646e-01 -3.51102710e-01
-2.14603499e-01 1.21329629e+00 7.11471558e-01 -3.76775295e-01
6.38739288e-01 4.72894639e-01 9.23228413e-02 -1.06330387e-01
-3.24389428e-01 -2.69828379e-01 6.22313142e-01 8.04435790e-01
2.04374790e-01 -1.97880194e-01 1.98071167e-01 1.58562317e-01
1.97176054e-01 -1.62112936e-01 3.47846858e-02 1.08100331e+00
-8.95620883e-01 -7.65075564e-01 -9.70917463e-01 -1.43999919e-01
-4.38273102e-01 -1.46894470e-01 -1.88419417e-01 7.27534831e-01
1.30850345e-01 9.56145644e-01 1.32471263e-01 3.11097801e-01
2.09674180e-01 -1.98604688e-01 7.42960632e-01 -7.60151505e-01
-1.33960724e-01 5.42526543e-01 -1.94416732e-01 -6.44855738e-01
-5.62942803e-01 -8.28475535e-01 -1.01205385e+00 4.49148864e-01
-7.87314475e-01 -1.30820751e-01 1.06665432e+00 8.68774831e-01
-1.90097056e-02 1.38007164e-01 1.31722420e-01 -1.47260189e+00
-6.18871987e-01 -8.02250743e-01 -4.19325233e-02 -1.81560189e-01
5.12351751e-01 -9.43373024e-01 -2.37676322e-01 9.05032232e-02] | [7.639462471008301, -2.428853988647461] |
f1fbcefe-699d-42a9-8e27-6a87970a3ab8 | pseudo-label-guided-cross-video-pixel | 2207.09664 | null | https://arxiv.org/abs/2207.09664v1 | https://arxiv.org/pdf/2207.09664v1.pdf | Pseudo-label Guided Cross-video Pixel Contrast for Robotic Surgical Scene Segmentation with Limited Annotations | Surgical scene segmentation is fundamentally crucial for prompting cognitive assistance in robotic surgery. However, pixel-wise annotating surgical video in a frame-by-frame manner is expensive and time consuming. To greatly reduce the labeling burden, in this work, we study semi-supervised scene segmentation from robotic surgical video, which is practically essential yet rarely explored before. We consider a clinically suitable annotation situation under the equidistant sampling. We then propose PGV-CL, a novel pseudo-label guided cross-video contrast learning method to boost scene segmentation. It effectively leverages unlabeled data for a trusty and global model regularization that produces more discriminative feature representation. Concretely, for trusty representation learning, we propose to incorporate pseudo labels to instruct the pair selection, obtaining more reliable representation pairs for pixel contrast. Moreover, we expand the representation learning space from previous image-level to cross-video, which can capture the global semantics to benefit the learning process. We extensively evaluate our method on a public robotic surgery dataset EndoVis18 and a public cataract dataset CaDIS. Experimental results demonstrate the effectiveness of our method, consistently outperforming the state-of-the-art semi-supervised methods under different labeling ratios, and even surpassing fully supervised training on EndoVis18 with 10.1% labeling. | ['Pheng-Ann Heng', 'Qi Dou', 'Guangyong Chen', 'Yueming Jin', 'Zixu Zhao', 'Yang Yu'] | 2022-07-20 | null | null | null | null | ['scene-segmentation'] | ['computer-vision'] | [ 4.48001653e-01 4.21832949e-01 -7.05119133e-01 -5.42042017e-01
-9.73703921e-01 -4.44811642e-01 1.37665376e-01 9.52883216e-04
-4.27403390e-01 6.92040443e-01 2.28929207e-01 -3.32801074e-01
-1.08140379e-01 -3.07584643e-01 -8.56904566e-01 -8.63925576e-01
4.09209877e-01 2.79255211e-01 8.86507481e-02 9.69896093e-02
1.50610553e-02 1.87376454e-01 -1.26369679e+00 5.15508533e-01
9.34510529e-01 1.03712070e+00 4.70563412e-01 1.89783826e-01
2.96228658e-02 1.24360287e+00 -5.30766845e-02 -1.18951626e-01
3.71279120e-01 -3.77014875e-01 -1.02887475e+00 5.97963154e-01
5.32677174e-01 -3.84836197e-01 -4.13795978e-01 1.16091907e+00
3.07025909e-01 -1.11560464e-01 4.77357566e-01 -8.19201589e-01
-3.19851607e-01 6.78882122e-01 -6.17822707e-01 2.70267166e-02
1.14670567e-01 2.27483466e-01 6.80109739e-01 -8.30930710e-01
1.07492459e+00 1.02940619e+00 3.22611630e-01 5.40114045e-01
-1.06712008e+00 -6.43144906e-01 4.16697115e-01 3.17471884e-02
-1.28445339e+00 -1.97887972e-01 1.01630235e+00 -5.73486924e-01
1.27784073e-01 5.06310686e-02 7.74400175e-01 1.07755291e+00
2.40450248e-01 1.16661417e+00 1.43730068e+00 -2.66304195e-01
1.09951258e-01 1.84886307e-02 2.85429973e-02 1.26553178e+00
1.66347772e-01 2.02650353e-01 -3.45570296e-01 1.30893528e-01
1.13357246e+00 4.59558785e-01 -6.80941761e-01 -8.46842051e-01
-1.61297166e+00 5.69789290e-01 7.37547815e-01 1.39346719e-01
-3.62748414e-01 2.84588099e-01 4.40708607e-01 5.73472939e-02
2.92999923e-01 3.94223422e-01 -3.93179744e-01 7.68666912e-04
-7.84009099e-01 -3.76824617e-01 6.34677410e-01 1.34500170e+00
6.82417691e-01 -1.23103604e-01 -3.76025736e-01 6.85792983e-01
2.77812958e-01 1.57158509e-01 5.18809021e-01 -1.09444571e+00
1.09101892e-01 8.10264409e-01 -3.18409763e-02 -5.33278465e-01
-6.01124048e-01 -6.02977395e-01 -8.36876392e-01 1.22202784e-01
4.06712443e-01 4.53712605e-02 -1.21816480e+00 1.34753764e+00
3.10280710e-01 5.94980478e-01 1.04368083e-01 1.24664676e+00
9.08628047e-01 9.74673033e-02 2.17597187e-01 -6.04735851e-01
1.31019247e+00 -1.43819296e+00 -8.23943436e-01 -2.05550194e-01
1.17085636e+00 -6.51362419e-01 1.13858199e+00 4.43084657e-01
-8.78172338e-01 -3.75906020e-01 -9.98721838e-01 -2.38642663e-01
2.07471326e-01 5.51512003e-01 9.54081535e-01 2.95221031e-01
-8.34120870e-01 3.20242405e-01 -1.01242208e+00 -2.52982024e-02
7.72175491e-01 4.41977650e-01 -4.62829769e-01 -7.29920030e-01
-6.88907027e-01 6.69971049e-01 5.27584255e-01 2.06635937e-01
-1.27890968e+00 -8.14345837e-01 -1.04204047e+00 -4.78952050e-01
8.25856507e-01 -5.22052824e-01 1.17569613e+00 -1.16592145e+00
-1.44560313e+00 1.25301909e+00 -4.97328006e-02 -3.23909700e-01
6.79884911e-01 -8.83572325e-02 -3.05080377e-02 4.08673763e-01
-2.69129127e-02 9.06821251e-01 8.21851552e-01 -1.66020274e+00
-4.78675187e-01 -2.30649859e-01 2.78155923e-01 4.97404724e-01
-1.62974939e-01 -4.14390385e-01 -8.38616133e-01 -6.95020616e-01
4.25977379e-01 -1.35222876e+00 -6.42362773e-01 4.28574622e-01
-4.14421767e-01 2.53769159e-02 5.23620129e-01 -5.04503191e-01
8.78383934e-01 -2.19703817e+00 2.09200904e-01 3.85571532e-02
2.93513149e-01 4.15582731e-02 1.44787773e-01 -4.27771211e-01
1.02791572e-02 -4.36607182e-01 -5.60685039e-01 -4.40616429e-01
-5.56020319e-01 5.39529026e-01 6.05905131e-02 8.18064392e-01
-7.32428506e-02 1.02752292e+00 -1.36010242e+00 -9.44092035e-01
4.76373374e-01 2.93957770e-01 -8.16822886e-01 2.91202068e-01
-1.05362624e-01 1.00156331e+00 -6.32684052e-01 1.00160992e+00
5.39761961e-01 -4.26710695e-01 2.63475895e-01 -4.62592870e-01
1.31576166e-01 -1.81953251e-01 -6.56413794e-01 2.56685162e+00
-5.93280077e-01 3.47831428e-01 6.67548105e-02 -1.10601532e+00
7.70082176e-01 3.74425389e-02 8.56269419e-01 -4.31380481e-01
4.68039423e-01 4.33412731e-01 -1.99314460e-01 -6.36970878e-01
2.13499516e-01 8.53272676e-02 -1.77389234e-01 -8.17100704e-02
7.67114237e-02 -3.59925747e-01 -5.04751429e-02 2.17160270e-01
7.58729160e-01 3.90972376e-01 3.28619570e-01 -4.06779885e-01
5.69923818e-01 3.24580103e-01 7.57616162e-01 5.60401559e-01
-5.44965327e-01 8.22504699e-01 4.94650513e-01 -2.76089340e-01
-4.44090188e-01 -8.87632430e-01 -2.59776264e-01 6.91782296e-01
7.71243632e-01 -1.27727598e-01 -7.23546386e-01 -1.24252939e+00
-1.46775484e-01 3.65692735e-01 -7.41810441e-01 -3.80928278e-01
-6.28657043e-01 -3.51942033e-01 -1.27837092e-01 5.32269120e-01
3.00097674e-01 -9.99296486e-01 -4.65703517e-01 4.69344258e-02
-3.95970464e-01 -1.50033116e+00 -5.69166481e-01 3.52771878e-01
-9.81204987e-01 -1.35641491e+00 -7.55039334e-01 -1.23229778e+00
1.25129569e+00 6.40666008e-01 9.19490993e-01 9.89448652e-02
-5.01415610e-01 6.07518017e-01 -4.05272007e-01 -1.12844542e-01
-2.82612830e-01 -1.61673024e-01 -1.51158169e-01 3.02483197e-02
1.80775020e-02 2.69724918e-03 -9.92460668e-01 3.94921035e-01
-8.18482876e-01 4.39982802e-01 9.22732234e-01 1.13210022e+00
1.15743244e+00 -4.77043927e-01 2.08459422e-01 -1.39328349e+00
-1.13587357e-01 -4.49199319e-01 -4.07409489e-01 3.73805642e-01
-5.02716660e-01 -1.27566949e-01 3.80505711e-01 -3.07104975e-01
-1.09397447e+00 6.61047697e-01 3.05518985e-01 -9.32837963e-01
-2.56676860e-02 4.87769842e-01 1.65104195e-01 -5.03843367e-01
4.52701747e-01 1.43007427e-01 2.82787800e-01 1.05031036e-01
5.56531608e-01 4.99670088e-01 7.43254781e-01 -5.91716886e-01
4.00120884e-01 8.61186087e-01 -5.85311651e-02 -4.31877881e-01
-1.24590027e+00 -8.00868690e-01 -7.06430554e-01 -4.62978125e-01
1.00630450e+00 -1.30676663e+00 -5.48180103e-01 2.12091520e-01
-8.78157616e-01 -5.33387303e-01 -2.93474942e-01 7.83215761e-01
-8.56449604e-01 5.05108356e-01 -5.54177284e-01 -3.23502392e-01
-2.01349467e-01 -1.70645463e+00 1.28204179e+00 1.01038314e-01
1.94404125e-01 -9.24006581e-01 -3.55429143e-01 7.45769680e-01
-4.42825910e-03 3.40744823e-01 4.52694058e-01 -4.39287186e-01
-9.15634513e-01 -3.91549282e-02 -4.81700033e-01 3.44665378e-01
2.53157109e-01 -4.69945431e-01 -8.08184326e-01 -4.06749845e-01
3.36098634e-02 -6.24387085e-01 1.01086986e+00 5.96660614e-01
1.75833797e+00 1.09071761e-01 -4.59054083e-01 1.00796318e+00
1.28421450e+00 -1.09605514e-01 4.75603938e-01 2.71780610e-01
1.07578981e+00 6.70904875e-01 1.35125804e+00 3.44377667e-01
5.59976935e-01 4.83581454e-01 6.97470665e-01 -4.34337139e-01
-3.95496368e-01 -2.27082372e-01 1.06169425e-01 8.32979023e-01
-2.20986139e-02 1.51105642e-01 -7.15509117e-01 4.77946132e-01
-2.05896902e+00 -3.04719418e-01 9.94153619e-02 1.93092549e+00
1.02534950e+00 -6.35598078e-02 -4.91240591e-01 -2.38857850e-01
4.36521024e-01 7.45252371e-02 -6.34080410e-01 3.88033718e-01
6.44214526e-02 -4.12111022e-02 9.18261528e-01 3.80004376e-01
-1.35598314e+00 1.25294340e+00 5.24388599e+00 8.25018287e-01
-1.16257203e+00 3.21656287e-01 9.05825019e-01 -1.02460571e-01
-1.95007473e-01 -3.85690778e-02 -4.44521636e-01 2.44664773e-01
2.43055522e-01 2.62200385e-01 2.57048130e-01 1.03978920e+00
2.39855453e-01 -2.06726521e-01 -1.11714756e+00 1.39715064e+00
3.76810074e-01 -1.39663863e+00 7.42594525e-02 -1.32882699e-01
1.06752157e+00 -1.81786716e-01 -7.90354237e-02 3.08690220e-01
1.11345768e-01 -8.88307989e-01 4.98065323e-01 5.39977491e-01
9.23805952e-01 -4.28980350e-01 7.98241794e-01 1.44520709e-02
-9.58960414e-01 -2.09751934e-01 -3.36886972e-01 4.37700570e-01
9.09141228e-02 4.96957600e-01 -7.03085840e-01 6.27789497e-01
4.93821919e-01 1.19256020e+00 -4.35533047e-01 1.03799307e+00
-2.40729868e-01 3.33013058e-01 7.96550214e-02 3.23289186e-01
4.41929132e-01 -1.64094076e-01 1.68918222e-01 1.05520701e+00
1.81146547e-01 4.01234299e-01 7.69687653e-01 3.97262454e-01
-2.11096644e-01 1.90739945e-01 -4.30415273e-01 1.99439123e-01
2.25410178e-01 1.43646955e+00 -8.73299122e-01 -3.62927973e-01
-4.44513738e-01 1.16581154e+00 3.12815577e-01 4.79406834e-01
-8.80088925e-01 1.85076505e-01 2.48233646e-01 -2.64564832e-03
-1.59678627e-02 -2.28223592e-01 -3.07298362e-01 -1.25727093e+00
-4.08870801e-02 -6.27340734e-01 5.13413727e-01 -6.76764011e-01
-8.73904705e-01 5.86130917e-01 3.98397446e-04 -1.81537485e+00
-5.60243800e-03 -7.78221130e-01 -1.39697075e-01 1.87111676e-01
-2.05892730e+00 -1.54645026e+00 -9.95739996e-01 9.37109172e-01
7.60601163e-01 3.67042981e-02 5.62456250e-01 2.18251467e-01
-3.72154236e-01 4.74245459e-01 -1.43353224e-01 2.68751867e-02
1.05629802e+00 -1.08490956e+00 -4.96809185e-01 5.31304121e-01
-6.73059300e-02 4.14513946e-01 1.99624583e-01 -4.85752225e-01
-1.54038084e+00 -1.29988623e+00 1.36452904e-02 -2.54518181e-01
4.91162449e-01 -1.12134330e-01 -5.63348830e-01 7.56911218e-01
-2.13663373e-02 6.82900846e-01 6.60811245e-01 -3.69301081e-01
-5.75538864e-03 -3.08793560e-02 -1.18510234e+00 7.14545846e-01
1.47866535e+00 -4.23408419e-01 -2.99438655e-01 7.76720822e-01
1.00208962e+00 -9.65350807e-01 -1.06461203e+00 7.43705571e-01
1.84428364e-01 -7.87960470e-01 9.67839241e-01 -1.96564615e-01
5.28670967e-01 -3.64622831e-01 -5.75343184e-02 -9.66219127e-01
1.38444612e-02 -5.66385746e-01 -6.89938068e-02 7.05316305e-01
2.04563767e-01 -4.33635980e-01 9.53190029e-01 5.80037117e-01
-7.93454707e-01 -1.09482276e+00 -9.02695954e-01 -4.13973600e-01
-2.54781485e-01 -4.24071133e-01 -1.19150698e-01 1.13102853e+00
8.05135071e-02 -3.43563437e-01 -4.19875413e-01 1.61043450e-01
7.83951938e-01 2.12911695e-01 5.92093468e-01 -9.46586311e-01
-1.59018919e-01 -2.26397127e-01 -5.74167490e-01 -1.19913375e+00
3.08431685e-01 -1.19158912e+00 3.61158192e-01 -1.60368967e+00
3.52244586e-01 -6.78888142e-01 -5.85203767e-01 5.23970366e-01
-2.23570049e-01 4.08315897e-01 -1.89465165e-01 4.49306071e-01
-8.46030593e-01 5.43350935e-01 1.94762743e+00 -4.11903739e-01
-1.22503176e-01 -1.29805446e-01 -6.36269629e-01 8.19726825e-01
4.16176766e-01 -3.58517736e-01 -7.21521318e-01 -2.98484713e-01
-3.21346164e-01 4.30467315e-02 3.69514495e-01 -8.65693867e-01
3.56583923e-01 -1.09614760e-01 1.72114164e-01 -3.05445194e-01
3.02655399e-01 -9.65062320e-01 -3.67204398e-01 6.14214838e-01
-4.13015962e-01 -6.39697611e-01 -3.64368968e-03 8.48947406e-01
-3.71050000e-01 -1.75536737e-01 8.38056624e-01 -1.62792325e-01
-1.11503196e+00 7.18130291e-01 1.70955390e-01 4.46089581e-02
1.37750518e+00 -9.39247534e-02 -1.63788676e-01 6.43153042e-02
-8.79641533e-01 4.67140526e-01 5.37086129e-01 5.08746147e-01
9.09539640e-01 -1.17913949e+00 -4.50838357e-01 1.57432795e-01
4.95742321e-01 5.40514946e-01 5.71346283e-01 1.39939523e+00
-7.28905857e-01 2.49989659e-01 -9.09530893e-02 -1.13349354e+00
-1.08729994e+00 6.39563680e-01 2.63836473e-01 -1.45073637e-01
-8.97946000e-01 9.17893410e-01 6.19148612e-01 -3.96323711e-01
4.55003411e-01 -6.74648225e-01 -1.45801172e-01 -2.66850978e-01
1.77606493e-01 -2.38951534e-01 -1.51988372e-01 -5.25590003e-01
-3.38885128e-01 7.71163404e-01 -3.68611306e-01 3.16201121e-01
8.62577796e-01 -1.81590930e-01 2.20921263e-02 2.83525884e-01
1.27157617e+00 -3.25809062e-01 -1.79720211e+00 -3.70605409e-01
-2.32492208e-01 -6.21054113e-01 4.15433586e-01 -6.23111010e-01
-1.40506434e+00 5.33097684e-01 7.76277959e-01 -6.09311104e-01
1.09945118e+00 2.09161445e-01 7.48664618e-01 3.12009782e-01
8.11391294e-01 -9.09028888e-01 2.68789500e-01 1.18873410e-01
7.77833045e-01 -1.78907406e+00 1.27484933e-01 -9.60165977e-01
-1.18789828e+00 9.77129400e-01 7.51666903e-01 -1.32621676e-01
6.09458387e-01 2.30034105e-02 2.98620135e-01 -2.79339582e-01
-2.65097380e-01 -2.53671229e-01 4.10941809e-01 4.36616153e-01
3.90837163e-01 1.44637167e-01 -2.58357018e-01 4.65156615e-01
2.29803666e-01 2.69406915e-01 5.54181755e-01 1.09616506e+00
-2.26664260e-01 -7.03964949e-01 1.25763223e-01 3.80015045e-01
-1.97733030e-01 -4.07276489e-02 5.83755262e-02 6.90407395e-01
3.67530808e-02 7.16565788e-01 -1.47104979e-01 -8.65906477e-02
1.25431135e-01 -3.96684289e-01 5.87078691e-01 -6.82083011e-01
-3.14363062e-01 2.24515676e-01 7.65010938e-02 -9.63337600e-01
-7.80413568e-01 -4.88640249e-01 -1.65819514e+00 4.70039874e-01
-2.37179264e-01 -1.13266513e-01 5.79382360e-01 9.40071344e-01
1.26026869e-01 9.00015950e-01 7.19572186e-01 -1.05642474e+00
-2.40142152e-01 -6.55549943e-01 -4.32012379e-01 4.55870479e-01
2.69883305e-01 -9.71840262e-01 -2.31013149e-01 1.68079466e-01] | [14.273147583007812, -2.989182710647583] |
8c2f4378-d656-4375-9da6-1cc90b17ac9b | towards-effective-multi-label-recognition | 2207.05137 | null | https://arxiv.org/abs/2207.05137v1 | https://arxiv.org/pdf/2207.05137v1.pdf | Towards Effective Multi-Label Recognition Attacks via Knowledge Graph Consistency | Many real-world applications of image recognition require multi-label learning, whose goal is to find all labels in an image. Thus, robustness of such systems to adversarial image perturbations is extremely important. However, despite a large body of recent research on adversarial attacks, the scope of the existing works is mainly limited to the multi-class setting, where each image contains a single label. We show that the naive extensions of multi-class attacks to the multi-label setting lead to violating label relationships, modeled by a knowledge graph, and can be detected using a consistency verification scheme. Therefore, we propose a graph-consistent multi-label attack framework, which searches for small image perturbations that lead to misclassifying a desired target set while respecting label hierarchies. By extensive experiments on two datasets and using several multi-label recognition models, we show that our method generates extremely successful attacks that, unlike naive multi-label perturbations, can produce model predictions consistent with the knowledge graph. | ['Ehsan Elhamifar', 'Hassan Mahmood'] | 2022-07-11 | null | null | null | null | ['multi-label-learning'] | ['methodology'] | [ 9.23065186e-01 1.44995570e-01 -2.97181100e-01 -3.05022538e-01
-8.60281885e-01 -1.17296827e+00 4.38862860e-01 2.46015266e-01
4.10278654e-03 5.13786614e-01 -6.44774020e-01 -3.22606027e-01
3.59399989e-03 -7.57074475e-01 -1.10088789e+00 -7.41528571e-01
1.22292608e-01 4.29159403e-01 4.01461899e-01 -2.44909879e-02
1.41270131e-01 6.24632537e-01 -1.12765038e+00 3.37756366e-01
2.26489648e-01 8.28317225e-01 -8.45874310e-01 7.46451199e-01
4.40300941e-01 1.02806652e+00 -7.56406605e-01 -8.83855700e-01
6.97141886e-01 -5.17354548e-01 -1.06112003e+00 4.41200644e-01
1.16149831e+00 7.30220750e-02 -3.13413382e-01 1.84735012e+00
2.80975938e-01 -1.75665900e-01 6.05319619e-01 -2.03150678e+00
-7.43536115e-01 4.59747314e-01 -4.89401698e-01 -1.25778437e-01
3.90083402e-01 -2.86474009e-03 9.12459373e-01 -3.01821738e-01
5.83282709e-01 1.36364305e+00 7.24885285e-01 7.86326647e-01
-1.44553506e+00 -1.06328273e+00 1.77690834e-01 2.36254141e-01
-1.43487656e+00 -1.21138267e-01 6.55960858e-01 -3.52583379e-01
3.46865922e-01 5.59061348e-01 -4.95564565e-02 1.18508041e+00
3.63956630e-01 3.48356068e-01 1.64301050e+00 -5.82815766e-01
1.13859639e-01 2.43084460e-01 1.04104534e-01 9.34631944e-01
2.31384322e-01 3.31677288e-01 -3.60975415e-01 -5.83039582e-01
3.04097801e-01 -1.83684051e-01 -3.09691548e-01 -7.26034522e-01
-9.15976405e-01 8.60043466e-01 2.66447753e-01 5.14029488e-02
2.05397129e-01 3.06221634e-01 4.42238539e-01 6.90003932e-01
2.18371704e-01 5.08784413e-01 -3.21781904e-01 7.31368840e-01
-4.54509050e-01 9.35038775e-02 7.87012219e-01 9.77407873e-01
9.97116446e-01 -2.05126449e-01 6.39021471e-02 2.72495568e-01
2.66857356e-01 7.22218573e-01 2.38323197e-01 -8.12180281e-01
1.00177497e-01 5.44192314e-01 -1.49341911e-01 -1.46742392e+00
-4.80025291e-01 -1.64256811e-01 -9.57847059e-01 4.62379843e-01
4.75302339e-01 2.31670849e-02 -7.91374147e-01 2.12982059e+00
4.98956352e-01 7.73518443e-01 2.05430463e-01 5.80031753e-01
4.59591448e-01 1.27192810e-01 2.35466942e-01 -4.27926004e-01
1.24235129e+00 -9.35899973e-01 -4.60010022e-01 -2.28083774e-01
7.89476633e-01 -7.23457038e-01 7.19556153e-01 2.63886273e-01
-6.28911555e-01 -4.44820404e-01 -1.05042613e+00 5.16019583e-01
-4.55204248e-01 -6.54306054e-01 3.75056267e-01 1.17973495e+00
-7.72415340e-01 2.99788773e-01 -3.11685622e-01 -2.48984486e-01
3.33957046e-01 5.20150065e-01 -7.68437922e-01 -3.01984340e-01
-1.33387077e+00 9.05999005e-01 5.15570104e-01 -3.10883254e-01
-1.16301858e+00 -4.29350466e-01 -9.75015938e-01 -3.45268011e-01
6.74800992e-01 -3.72895360e-01 1.06110740e+00 -1.09861088e+00
-9.28032637e-01 1.29062521e+00 1.56686574e-01 -4.84587163e-01
4.61923391e-01 3.87554824e-01 -6.86470747e-01 3.44185561e-01
4.74893115e-02 4.91404533e-01 1.18553090e+00 -1.71933246e+00
-6.01049840e-01 -3.57728213e-01 4.41425085e-01 -1.39918476e-01
-3.22136968e-01 3.68517339e-01 -7.45593831e-02 -6.34929061e-01
9.56774652e-02 -1.58777237e+00 -3.46028805e-01 -1.61027953e-01
-8.66672397e-01 1.52462855e-01 1.12965071e+00 -5.04473709e-02
8.14851522e-01 -2.06214976e+00 4.64743935e-02 5.32811642e-01
2.28612319e-01 3.26274842e-01 -3.18373740e-01 2.16889352e-01
-4.61380184e-01 5.65344274e-01 -3.05940449e-01 -1.52629271e-01
3.98999564e-02 2.81665295e-01 -8.33423316e-01 9.37718332e-01
-6.16819672e-02 9.41754818e-01 -8.39400649e-01 -5.21020055e-01
1.06106438e-01 -1.65603142e-02 -2.72062540e-01 7.14418218e-02
-2.18892202e-01 4.74155724e-01 -2.44094729e-01 9.21352267e-01
8.79896402e-01 -5.14983416e-01 4.27254260e-01 -1.72211662e-01
8.03656757e-01 -7.70804644e-01 -1.32758033e+00 1.04651070e+00
-1.02432981e-01 1.55007094e-01 -1.30066201e-01 -9.33858216e-01
6.42587602e-01 3.51884633e-01 4.71053004e-01 -3.52019638e-01
1.89064771e-01 4.99055646e-02 -2.29877517e-01 -1.36601701e-01
8.90810341e-02 -3.25992614e-01 -5.29342949e-01 7.58154035e-01
-1.40450895e-01 -1.49170250e-01 -1.46027550e-01 3.95114511e-01
1.27213168e+00 -4.35407043e-01 3.47018808e-01 6.30896166e-02
7.92704403e-01 -8.64781067e-02 7.50277579e-01 1.12905490e+00
-5.80949783e-01 3.52480650e-01 5.48536897e-01 -5.02530038e-01
-7.33148992e-01 -8.60132396e-01 -1.23706326e-01 8.07357311e-01
7.36308634e-01 -3.40671122e-01 -7.50451624e-01 -1.38509274e+00
2.30936959e-01 2.36058131e-01 -6.73059583e-01 -6.11018121e-01
-3.96779984e-01 -6.50072813e-01 1.05884361e+00 -6.14177948e-03
4.72293735e-01 -1.05595911e+00 -1.18870303e-01 -7.26530105e-02
-6.88938648e-02 -1.50335503e+00 -5.75729012e-01 5.69537980e-03
-2.89048851e-01 -1.79554307e+00 2.92983931e-02 -8.80272985e-01
9.88776922e-01 1.70530379e-01 9.19840038e-01 4.26864505e-01
-2.63102621e-01 7.08755016e-01 -2.45166376e-01 -2.53700554e-01
-1.10519671e+00 -2.25692242e-01 3.34849566e-01 4.49393690e-01
1.86821401e-01 -3.34592402e-01 4.75008376e-02 7.37412274e-01
-1.32874739e+00 -2.97364891e-01 2.41701216e-01 8.22851479e-01
8.18583608e-01 4.69064534e-01 6.62978530e-01 -1.47413850e+00
3.20667505e-01 -3.10976088e-01 -7.99918592e-01 7.96382010e-01
-7.37582445e-01 -1.85522854e-01 8.31959963e-01 -8.14501822e-01
-4.40735161e-01 4.31063116e-01 2.44069934e-01 -6.35918736e-01
-3.69411111e-01 8.64972621e-02 -3.89492154e-01 -1.07762420e+00
9.00596857e-01 5.47393747e-02 -9.88471508e-02 1.86830804e-01
4.45297867e-01 4.14042473e-01 6.52324498e-01 -5.11654615e-01
1.34535432e+00 4.84405518e-01 7.96171606e-01 -2.54352331e-01
-1.21442533e+00 -2.66773611e-01 -5.80710113e-01 -3.23796362e-01
6.51162684e-01 -6.91461802e-01 -1.02189279e+00 8.92334819e-01
-9.50843573e-01 -1.48843765e-01 1.25648398e-02 2.47329567e-02
-7.60743141e-01 8.37736666e-01 -6.29522026e-01 -4.14421231e-01
-9.13636386e-02 -1.39448130e+00 8.45436871e-01 -5.50174415e-02
3.04384381e-02 -1.06682026e+00 9.09936801e-02 5.22142053e-01
-9.33020189e-03 6.32996023e-01 1.16287220e+00 -1.08289599e+00
-7.18241930e-01 -4.98195618e-01 -6.17297478e-02 3.91277283e-01
1.49871379e-01 -2.29303643e-01 -9.78199422e-01 -6.60448730e-01
-6.71046451e-02 -1.00353014e+00 5.60030758e-01 -1.88105479e-01
1.18417966e+00 -4.02136415e-01 -4.68177050e-01 5.35743952e-01
1.59064770e+00 5.57109527e-02 4.58854169e-01 3.98762584e-01
8.64972055e-01 2.44716913e-01 7.04199493e-01 1.30260497e-01
1.17168270e-01 7.87888169e-01 8.88536990e-01 -6.86348602e-02
5.79673126e-02 -3.05964611e-02 1.43145502e-01 1.90663919e-01
6.41319036e-01 -4.26389754e-01 -7.79976726e-01 1.09258093e-01
-1.90884936e+00 -1.00352538e+00 1.80771351e-02 2.08813024e+00
1.02013457e+00 1.10160500e-01 -1.00380868e-01 1.79450139e-01
1.12332976e+00 2.13524267e-01 -7.94655204e-01 -3.30509216e-01
-2.70172477e-01 2.91653499e-02 7.59648979e-01 6.10572994e-01
-1.63796687e+00 1.12825584e+00 7.29098606e+00 8.40939462e-01
-9.88917887e-01 4.59246337e-02 5.54276824e-01 3.54371578e-01
-1.31655380e-01 1.37875870e-01 -8.87870252e-01 1.27294034e-01
8.74894440e-01 -2.55928576e-01 4.23483878e-01 8.75612795e-01
-6.33955240e-01 3.11870039e-01 -1.16279829e+00 8.15499961e-01
4.53222960e-01 -1.10423577e+00 1.59413874e-01 3.71041596e-02
1.05453789e+00 -4.96540695e-01 2.46751308e-01 1.01924375e-01
8.14047158e-01 -9.78537381e-01 5.24684787e-01 1.33905858e-01
1.01596951e+00 -9.32180822e-01 4.41382617e-01 4.33605164e-01
-9.69584584e-01 -1.45415515e-01 -4.11410719e-01 3.57476443e-01
-9.68332961e-02 4.48609710e-01 -6.24577045e-01 5.66391766e-01
2.35900089e-01 4.19317484e-01 -8.68705988e-01 6.70735598e-01
-5.03376186e-01 4.43909883e-01 -1.02915671e-02 5.78986049e-01
-5.67049161e-02 2.27904171e-01 3.46955508e-01 8.66535068e-01
-1.95150360e-01 -1.12060681e-01 8.18531513e-01 2.30852455e-01
-5.57325125e-01 9.93302017e-02 -9.33391690e-01 2.38964349e-01
5.70223093e-01 1.29109168e+00 -7.97029972e-01 -2.11899534e-01
-3.92473191e-01 1.11088049e+00 3.78865391e-01 1.46241397e-01
-1.13201797e+00 -1.08511552e-01 6.00157022e-01 -4.81121808e-01
-2.86376644e-02 1.24527492e-01 -1.19322382e-01 -1.10476673e+00
-1.17108047e-01 -1.60706162e+00 6.91371322e-01 -4.00239110e-01
-1.31928873e+00 4.89100605e-01 -2.57595301e-01 -1.13190293e+00
-1.65435150e-01 -4.52085346e-01 -1.54072687e-01 6.02642894e-01
-1.42954314e+00 -1.46857584e+00 -1.94923058e-01 9.50862825e-01
5.12595363e-02 -2.20641971e-01 1.32731462e+00 1.52528882e-01
-5.50218344e-01 1.15891218e+00 -1.97381556e-01 1.29668877e-01
1.07899857e+00 -1.13180768e+00 3.37275237e-01 1.12891912e+00
5.53216577e-01 2.25423872e-01 4.85373765e-01 -8.05794477e-01
-1.15764046e+00 -1.50550056e+00 7.50643790e-01 -6.75396800e-01
7.89373398e-01 -2.52637237e-01 -9.03171122e-01 1.14099634e+00
-1.21306311e-02 6.13164902e-01 9.51231360e-01 -2.62769938e-01
-1.13548851e+00 1.17082901e-01 -1.59334528e+00 5.46105802e-01
1.00834060e+00 -9.15309250e-01 -1.32298291e-01 7.59223402e-01
6.63745701e-01 -5.30175090e-01 -1.02785146e+00 6.96235895e-01
2.21256718e-01 -6.20217860e-01 1.03120458e+00 -9.22500908e-01
4.82344441e-02 -5.41274071e-01 -2.24597469e-01 -1.22081208e+00
-4.07197326e-01 -5.64210773e-01 -2.82408781e-02 1.02675200e+00
2.46979535e-01 -6.35828972e-01 7.91342556e-01 6.57352269e-01
2.50552505e-01 -2.95137644e-01 -9.44338620e-01 -1.01506519e+00
-2.79013123e-02 -1.94308057e-01 5.18741190e-01 1.36168277e+00
-2.76745528e-01 4.84835841e-02 -8.25903475e-01 9.95468318e-01
1.05094719e+00 1.79881439e-01 8.78523469e-01 -1.13975799e+00
-3.35636050e-01 -2.90148258e-02 -8.82831752e-01 -4.03329790e-01
9.18137610e-01 -1.02296484e+00 1.99406873e-02 -8.30916643e-01
3.12801093e-01 -5.37764013e-01 -4.18927193e-01 9.69033599e-01
-1.97304711e-01 9.85340595e-01 2.79976636e-01 3.37486267e-01
-9.01953697e-01 -7.60384947e-02 1.01501095e+00 -6.48518443e-01
4.84476328e-01 1.37172461e-01 -7.91740596e-01 7.91729271e-01
7.24722564e-01 -9.48782623e-01 -4.83551830e-01 1.41441301e-01
3.39393348e-01 1.10352284e-03 4.25644964e-01 -1.05960608e+00
3.88084888e-01 -4.50857759e-01 -4.73285932e-03 -1.65634789e-02
1.40993103e-01 -1.14201880e+00 4.34949785e-01 6.96474850e-01
-5.95405161e-01 -9.02900919e-02 1.51358590e-01 8.77514899e-01
-1.77530438e-01 -4.38630313e-01 1.23932505e+00 -6.08650260e-02
-8.51143599e-01 6.60962045e-01 -1.20839454e-01 1.36491582e-01
1.71461689e+00 4.42703515e-02 -6.96294308e-01 -2.72203147e-01
-8.48124206e-01 9.07278359e-02 7.52201736e-01 4.12193835e-01
4.15046275e-01 -1.48890793e+00 -5.85478842e-01 2.95762420e-01
4.23111528e-01 -3.90245527e-01 1.48388162e-01 2.27319703e-01
-2.02501535e-01 2.84804143e-02 -1.79242775e-01 -4.38356489e-01
-1.79317677e+00 1.25434422e+00 5.42928934e-01 -4.56655711e-01
-2.53133953e-01 8.00173342e-01 2.07332700e-01 -4.56092209e-01
3.69275987e-01 4.92059022e-01 1.30563781e-01 -2.36379385e-01
5.88738620e-01 5.28133102e-02 2.44632196e-02 -1.09009194e+00
-5.29886186e-01 8.97890627e-01 -2.69286245e-01 2.66200840e-01
6.43966258e-01 -2.02653520e-02 -3.61545056e-01 -1.07841820e-01
1.26504397e+00 1.04493923e-01 -8.67783129e-01 -5.06446660e-01
1.49620546e-03 -6.50363505e-01 -4.60000634e-01 -6.96697533e-01
-1.14316118e+00 2.72778600e-01 6.61752999e-01 4.58104849e-01
1.25804949e+00 -1.23818830e-01 4.91585702e-01 6.33500040e-01
6.98948562e-01 -6.04569077e-01 4.09478247e-01 3.79186720e-01
5.52310228e-01 -1.46773398e+00 -7.09695891e-02 -8.99438977e-01
-4.73528653e-01 7.64532149e-01 8.45078647e-01 3.97136714e-03
7.22500443e-01 2.49587938e-01 5.27744353e-01 -1.13866650e-01
-5.15305817e-01 1.09553710e-01 5.69186062e-02 8.49111080e-01
-4.32809681e-01 -4.91389260e-03 1.36596933e-01 -3.51162138e-03
1.36906758e-01 -3.62863630e-01 7.58307099e-01 8.17301989e-01
-2.56776363e-01 -1.61705744e+00 -6.81977272e-01 4.39648591e-02
-7.25705802e-01 1.08397454e-01 -5.89579105e-01 5.31685948e-01
2.46954203e-01 1.20457780e+00 -4.80433226e-01 -7.72379696e-01
3.59198302e-01 2.13080421e-01 5.35742402e-01 -6.52559400e-01
-7.58848011e-01 -5.35673797e-01 -1.36225954e-01 -5.77124178e-01
-5.56908607e-01 -4.13322449e-01 -1.05468583e+00 -4.28751141e-01
-5.59831381e-01 -8.70661139e-02 2.58087337e-01 9.13203299e-01
1.79868177e-01 1.79305196e-01 1.15606022e+00 -1.74972326e-01
-9.05512750e-01 -5.02064168e-01 -7.84480095e-01 1.02900624e+00
2.53455132e-01 -5.75657606e-01 -6.04589164e-01 2.15527028e-01] | [5.681197643280029, 7.74666166305542] |
183a9419-3dbf-489e-8f73-f1ec89368844 | unsupervised-learning-of-style-aware-facial | 2306.10006 | null | https://arxiv.org/abs/2306.10006v2 | https://arxiv.org/pdf/2306.10006v2.pdf | Unsupervised Learning of Style-Aware Facial Animation from Real Acting Performances | This paper presents a novel approach for text/speech-driven animation of a photo-realistic head model based on blend-shape geometry, dynamic textures, and neural rendering. Training a VAE for geometry and texture yields a parametric model for accurate capturing and realistic synthesis of facial expressions from a latent feature vector. Our animation method is based on a conditional CNN that transforms text or speech into a sequence of animation parameters. In contrast to previous approaches, our animation model learns disentangling/synthesizing different acting-styles in an unsupervised manner, requiring only phonetic labels that describe the content of training sequences. For realistic real-time rendering, we train a U-Net that refines rasterization-based renderings by computing improved pixel colors and a foreground matte. We compare our framework qualitatively/quantitatively against recent methods for head modeling as well as facial animation and evaluate the perceived rendering/animation quality in a user-study, which indicates large improvements compared to state-of-the-art approaches | ['Peter Eisert', 'Anna Hilsmann', 'Wolfgang Paier'] | 2023-06-16 | null | null | null | null | ['neural-rendering'] | ['computer-vision'] | [ 3.74007642e-01 1.44998699e-01 8.73529837e-02 -6.85835958e-01
-7.93091595e-01 -3.67758900e-01 8.67679417e-01 -5.31049311e-01
5.69715686e-02 3.08715373e-01 5.02315760e-01 6.59139007e-02
6.17063701e-01 -7.84009695e-01 -8.69453490e-01 -6.72398746e-01
-3.65554057e-02 7.34004259e-01 -2.56001323e-01 -2.65401900e-01
-1.56613365e-01 8.75647545e-01 -2.05030203e+00 3.96364123e-01
5.02905190e-01 8.09146404e-01 -2.26036698e-01 1.08868039e+00
-1.98403582e-01 7.91063309e-01 -5.34407556e-01 -4.89713520e-01
4.35328968e-02 -5.92872918e-01 -4.80352610e-01 4.87944424e-01
9.32464004e-01 -7.50211477e-01 -1.41297430e-01 8.72278333e-01
5.65874517e-01 1.75725460e-01 9.10952568e-01 -1.27071714e+00
-6.23836815e-01 2.45198682e-01 -4.02629107e-01 -5.06582439e-01
7.03272581e-01 4.39495325e-01 6.09410167e-01 -7.59829998e-01
8.36173654e-01 1.83270264e+00 4.00677264e-01 9.72489238e-01
-1.54252195e+00 -8.64493728e-01 4.06811275e-02 -2.72758573e-01
-1.38151693e+00 -6.42443120e-01 9.84353364e-01 -5.60018539e-01
6.90578461e-01 4.95619982e-01 1.16918266e+00 1.43550014e+00
-7.47713298e-02 8.81262958e-01 1.15621543e+00 -4.20478106e-01
2.67917395e-01 -1.19890489e-01 -5.28167486e-01 9.98908639e-01
-6.63682044e-01 3.21176112e-01 -5.04982591e-01 -1.71788290e-01
1.49505103e+00 -5.01836240e-01 -1.24174699e-01 -3.07397783e-01
-9.89714980e-01 7.43307590e-01 2.04578731e-02 -1.98532730e-01
-3.47618431e-01 5.25505245e-01 3.10102254e-01 -1.24799289e-01
7.92889476e-01 9.69463810e-02 2.17539612e-02 -3.47206146e-01
-1.45730758e+00 6.49462640e-01 7.01141655e-01 8.78906190e-01
4.83346015e-01 7.51106381e-01 -1.93474993e-01 7.65545309e-01
4.37625706e-01 8.25339377e-01 1.16796136e-01 -1.59538555e+00
-1.94307581e-01 4.20777351e-02 6.92485049e-02 -8.68126988e-01
-1.78692073e-01 1.47809520e-01 -7.68031657e-01 8.23242366e-01
2.73882359e-01 -6.40818402e-02 -1.15014255e+00 1.84557867e+00
3.94598931e-01 6.08287513e-01 -4.65756238e-01 8.53125453e-01
7.85136104e-01 6.79778814e-01 3.51675421e-01 -1.63401231e-01
1.12641287e+00 -9.06693578e-01 -9.60738599e-01 2.92137116e-01
2.81459928e-01 -7.10562587e-01 1.32117915e+00 5.37002385e-01
-1.68097305e+00 -6.42021656e-01 -7.64652908e-01 -3.48687708e-01
2.29092576e-02 1.42808989e-01 5.62083781e-01 8.48055422e-01
-1.67408562e+00 6.51244462e-01 -8.94615889e-01 -1.26053363e-01
3.54223907e-01 2.20101804e-01 -2.07041800e-01 4.78710979e-01
-7.16515481e-01 7.21760929e-01 -3.12459916e-01 -1.90221310e-01
-9.90616858e-01 -8.06060970e-01 -1.08261240e+00 2.12966148e-02
-4.17790890e-01 -9.19247329e-01 1.61689878e+00 -1.37265337e+00
-2.56619310e+00 1.10026395e+00 -3.50037485e-01 -1.04525745e-01
7.55878448e-01 -7.47542754e-02 -2.46350870e-01 1.76675782e-01
-3.33437264e-01 1.27096760e+00 1.24974036e+00 -1.57755911e+00
-1.66369125e-01 -1.22168958e-02 -4.51922454e-02 3.09547991e-01
1.41414152e-02 3.01585317e-01 -6.64897859e-01 -7.86498547e-01
-4.37044978e-01 -9.31860983e-01 -1.01521790e-01 5.25565088e-01
-2.98657686e-01 8.92973542e-02 9.05411601e-01 -7.75528967e-01
9.39780533e-01 -1.98278928e+00 3.98648024e-01 1.01358898e-01
2.21603662e-01 4.27309498e-02 -1.92814782e-01 -3.56209092e-02
-3.40853512e-01 9.29655209e-02 -2.09255546e-01 -1.16815400e+00
1.84333801e-01 2.92575330e-01 -5.02808154e-01 4.84053969e-01
1.55201375e-01 9.06394839e-01 -7.35303521e-01 -5.86419106e-01
6.48978949e-01 1.17506897e+00 -7.83776402e-01 5.12626290e-01
-6.61139011e-01 8.83735776e-01 4.41301689e-02 5.25141597e-01
7.64734685e-01 7.22498074e-02 1.59128025e-01 -1.15648888e-01
-1.94734931e-01 7.17091709e-02 -9.57218885e-01 1.77553976e+00
-7.06647933e-01 8.08508873e-01 1.84494555e-01 -3.57217282e-01
6.55687153e-01 5.68093359e-01 2.77408719e-01 -6.18117988e-01
4.74482000e-01 -1.20729938e-01 -5.71481466e-01 -4.02612060e-01
5.25023699e-01 -1.13227405e-01 1.46560520e-01 4.96842563e-01
-9.68998969e-02 -7.05325544e-01 -3.71061474e-01 -3.97705249e-02
3.83905083e-01 7.99276233e-01 -1.96842134e-01 -1.35887027e-01
1.84785485e-01 -5.54542065e-01 -1.98034346e-02 1.97800398e-01
1.03882343e-01 9.40050900e-01 5.33615112e-01 -3.96802694e-01
-1.36155570e+00 -1.04746270e+00 1.22383535e-01 1.29519987e+00
-4.26362902e-01 -2.72687018e-01 -1.42955506e+00 -1.16572110e-02
-2.97450542e-01 1.00306880e+00 -9.27773714e-01 2.45291263e-01
-8.87054205e-01 -3.01652908e-01 6.43081367e-01 5.09968638e-01
5.76314963e-02 -1.29732811e+00 -5.55250585e-01 9.15070646e-04
2.64371000e-02 -1.22525358e+00 -6.36859953e-01 -4.59680498e-01
-6.44815028e-01 -4.71544743e-01 -9.03793693e-01 -7.20207095e-01
6.57287121e-01 -3.29604417e-01 1.24629283e+00 2.15872884e-01
-2.52857685e-01 3.97321910e-01 1.55088350e-01 -1.53914019e-01
-8.82013142e-01 -4.04634237e-01 7.74810538e-02 1.74824238e-01
-4.56390232e-01 -9.35039699e-01 -7.86787331e-01 1.95194781e-02
-9.33857203e-01 7.35025764e-01 -8.47818926e-02 3.98834229e-01
6.17778480e-01 -6.70742214e-01 -1.89106241e-01 -8.32681656e-01
5.33359051e-01 -7.56561384e-02 -6.75498009e-01 -2.00585579e-03
-2.29550987e-01 5.78375682e-02 5.03097057e-01 -8.20613146e-01
-1.25797808e+00 1.97365075e-01 -5.48726499e-01 -8.38537931e-01
-4.89033341e-01 -1.89154953e-01 -2.59610504e-01 -1.80158354e-02
5.25652468e-01 1.34649396e-01 1.77469105e-01 -2.25953668e-01
9.72988307e-01 3.01239133e-01 9.04562473e-01 -8.76336753e-01
7.94199169e-01 6.56539381e-01 2.42233854e-02 -9.65971231e-01
-1.47619933e-01 2.64428616e-01 -7.18582809e-01 -5.56327581e-01
1.11453676e+00 -8.03396702e-01 -1.04541850e+00 8.80483449e-01
-1.38794732e+00 -9.95085895e-01 -3.57179075e-01 3.72106060e-02
-9.97226179e-01 2.78750598e-01 -8.24057937e-01 -9.48721826e-01
-2.04547822e-01 -1.50490522e+00 1.72521758e+00 -3.16868126e-02
-5.48313856e-01 -9.76369679e-01 8.68622512e-02 5.05021475e-02
5.31476438e-01 8.05054784e-01 9.15965080e-01 2.07802698e-01
-5.84781110e-01 1.49411410e-01 -8.19409937e-02 1.94402281e-02
-2.48425677e-02 8.49786282e-01 -1.41548073e+00 -3.19324285e-02
-3.27945620e-01 -1.99807718e-01 2.17588127e-01 7.39946187e-01
1.40771484e+00 -5.62199533e-01 -1.79636497e-02 1.20011091e+00
8.84026468e-01 1.44879833e-01 8.57363880e-01 -1.56035796e-01
1.02831626e+00 5.39379537e-01 -1.15566133e-02 6.51291907e-01
5.14782071e-01 9.06008840e-01 3.69296849e-01 -3.78872782e-01
-4.94218737e-01 -4.51315761e-01 4.44501042e-01 7.53393531e-01
-2.90784508e-01 -1.21003717e-01 -4.77205485e-01 1.81646019e-01
-1.31555676e+00 -9.24719989e-01 1.54697582e-01 2.03604317e+00
1.03160334e+00 -1.47309706e-01 4.77051079e-01 -9.48295370e-02
4.80779588e-01 3.31774503e-01 -3.63185823e-01 -8.12589943e-01
-7.63719678e-02 7.12480068e-01 4.30401862e-02 8.69591236e-01
-7.83686340e-01 1.30841565e+00 7.08522940e+00 7.77724087e-01
-1.67032754e+00 -1.58898588e-02 9.04206812e-01 -1.36925057e-01
-7.65833139e-01 -4.21512574e-01 -4.68146861e-01 1.37691185e-01
1.13348174e+00 1.64645329e-01 6.92017198e-01 8.16912293e-01
6.40743554e-01 2.85975248e-01 -1.19776213e+00 1.00260484e+00
2.19023526e-01 -1.62495065e+00 4.46733057e-01 7.31535405e-02
7.66117513e-01 -3.60783696e-01 6.29493773e-01 1.02039553e-01
5.57517290e-01 -1.32300925e+00 1.32969236e+00 6.47043765e-01
1.57775426e+00 -5.47199011e-01 -2.44559810e-01 5.42685948e-02
-1.15441668e+00 6.11681283e-01 2.70956635e-01 1.26951993e-01
4.01167303e-01 -1.73872456e-01 -6.25046790e-01 -9.97054949e-02
5.12174666e-01 3.84124845e-01 -2.23251671e-01 4.41417038e-01
-3.20861071e-01 6.41458094e-01 -2.65180409e-01 2.29547232e-01
-1.10337496e-01 -1.73613816e-01 3.51200491e-01 1.52128756e+00
1.51929885e-01 1.87522545e-01 -2.20921263e-02 1.13593125e+00
-1.57598779e-02 2.53961563e-01 -3.75509530e-01 -5.68366349e-02
2.43871078e-01 1.08536088e+00 -5.94014585e-01 -5.52708447e-01
-2.29961071e-02 1.24173677e+00 1.40390709e-01 4.62906629e-01
-1.04938471e+00 2.99125761e-01 8.46362889e-01 2.81860441e-01
8.70857462e-02 -5.39405346e-01 -4.08612877e-01 -1.03885579e+00
-5.49230576e-01 -9.56174195e-01 -3.67827773e-01 -1.18294799e+00
-8.27522993e-01 1.06068945e+00 2.40614504e-01 -9.44288135e-01
-7.97400355e-01 -3.24666977e-01 -6.73801541e-01 9.55975592e-01
-1.32447672e+00 -1.64752865e+00 -3.81204098e-01 5.72630107e-01
7.21783459e-01 1.56007960e-01 1.17151427e+00 1.20270804e-01
-3.20557088e-01 8.83461356e-01 -3.58848661e-01 5.45096695e-02
5.11789739e-01 -1.14458656e+00 1.14376032e+00 4.92096245e-01
6.21739496e-03 4.60741043e-01 1.03526545e+00 -4.82824564e-01
-1.26014531e+00 -8.84176373e-01 4.61804867e-01 -3.80875438e-01
3.15142274e-01 -6.06577575e-01 -8.69513512e-01 7.76525855e-01
3.31115872e-01 5.54729588e-02 4.52906370e-01 -3.98105651e-01
-6.00399256e-01 3.71683389e-01 -1.14198744e+00 1.02777600e+00
1.01621222e+00 -7.12243319e-01 2.46651303e-02 1.44699782e-01
6.99486256e-01 -9.71294701e-01 -6.95170224e-01 -6.21735258e-03
1.00525749e+00 -1.04521585e+00 9.76262271e-01 -4.12097007e-01
5.95207274e-01 -1.07514605e-01 -1.65942367e-02 -1.15480638e+00
-1.31208256e-01 -1.10981071e+00 -6.03861250e-02 1.06537378e+00
-2.65347138e-02 -5.52968681e-02 8.18288028e-01 9.09432352e-01
-3.45173813e-02 -5.25602818e-01 -6.83581054e-01 -2.02065960e-01
2.06166893e-01 -7.51310766e-01 9.23521817e-01 9.16892290e-01
-4.52762514e-01 -1.24424800e-01 -5.49862802e-01 7.36401752e-02
7.08883047e-01 -9.48358402e-02 1.03905404e+00 -9.89622653e-01
-1.82399958e-01 -7.49853313e-01 -2.23623499e-01 -1.09237707e+00
6.97093964e-01 -6.65137172e-01 4.10316065e-02 -1.16495097e+00
-1.28230184e-01 -1.62183583e-01 6.10927045e-01 4.26436782e-01
1.95309430e-01 3.44669253e-01 2.14506552e-01 -8.92673712e-03
-7.78059661e-02 6.47408545e-01 1.47465587e+00 1.34090886e-01
-2.11338744e-01 -1.61093995e-01 -1.76816195e-01 1.01238918e+00
3.67389023e-01 -5.11757396e-02 -3.61575902e-01 -4.78719234e-01
-4.99482900e-02 5.24240434e-01 5.52071631e-01 -8.46118510e-01
-8.94163921e-02 -3.56465459e-01 3.39296877e-01 -3.87326241e-01
9.43175197e-01 -4.54118907e-01 5.00141263e-01 2.56286394e-02
-6.22741997e-01 1.27096117e-01 2.83079535e-01 5.50376065e-02
1.94821194e-01 4.47743893e-01 1.05427825e+00 -3.63548249e-02
-3.17921340e-01 4.81046915e-01 -5.23823321e-01 -1.24028772e-01
3.92781258e-01 -3.27277541e-01 6.12905882e-02 -1.08464813e+00
-9.94887710e-01 -4.32914764e-01 8.58606279e-01 3.62071812e-01
7.96555102e-01 -1.44366026e+00 -7.74216354e-01 5.88731766e-01
-3.09362382e-01 -1.70583978e-01 3.60982180e-01 2.86772668e-01
-9.48651671e-01 -1.02986872e-01 -4.04716879e-01 -6.73640668e-01
-1.45862794e+00 2.55090326e-01 4.84144628e-01 1.32373273e-01
-7.05158591e-01 8.88535202e-01 5.01162767e-01 -2.77211338e-01
3.25876594e-01 -5.56278586e-01 -6.70777783e-02 -2.46367872e-01
4.43851829e-01 4.93144281e-02 -2.72560090e-01 -1.20424068e+00
1.27815992e-01 7.49667585e-01 3.59527558e-01 -7.60870337e-01
1.13888979e+00 1.30224181e-02 -4.42261659e-02 4.01153266e-01
1.39796138e+00 1.19371921e-01 -1.62397182e+00 1.02530241e-01
-6.27620995e-01 -4.28640872e-01 -1.06679145e-02 -5.57086170e-01
-1.22331333e+00 1.07649446e+00 6.50860786e-01 -3.41108024e-01
9.99528289e-01 -2.72576213e-01 9.59526122e-01 -3.77842247e-01
1.51301205e-01 -7.57192910e-01 1.53258905e-01 3.85849535e-01
1.09933031e+00 -7.90734291e-01 -2.45836005e-01 -5.65159500e-01
-7.97874629e-01 1.15582657e+00 4.51655596e-01 -2.88372964e-01
7.23004878e-01 7.67614841e-01 2.35936403e-01 -1.39183268e-01
-5.38372099e-01 6.03147373e-02 4.86028522e-01 6.90792561e-01
6.41332626e-01 3.08125436e-01 3.50785434e-01 7.22922534e-02
-8.27682137e-01 1.56182557e-01 3.65716815e-01 5.05703270e-01
-4.60859686e-02 -1.02368295e+00 -2.76264489e-01 -1.49258211e-01
-2.62947530e-01 -1.44597918e-01 -1.09633647e-01 6.89890981e-01
3.51387588e-03 4.49430615e-01 4.49251622e-01 -3.97724330e-01
1.40754595e-01 2.19857782e-01 9.39910233e-01 -4.28297520e-01
-5.67146659e-01 5.38309097e-01 1.35789346e-02 -7.49963760e-01
-4.27411497e-01 -6.64789736e-01 -1.33129430e+00 -4.88570094e-01
2.14877382e-01 -3.50565583e-01 8.15791965e-01 7.49624372e-01
1.88579157e-01 5.51123559e-01 6.34444833e-01 -1.62702751e+00
-2.33882689e-03 -5.88062823e-01 -4.90131199e-01 6.02357388e-01
3.79208416e-01 -6.28547251e-01 -2.91735172e-01 6.08128726e-01] | [12.792254447937012, -0.45074620842933655] |
76eaead6-af95-49e8-9566-c2680560c60f | online-segment-to-segment-neural-transduction | 1609.08194 | null | http://arxiv.org/abs/1609.08194v1 | http://arxiv.org/pdf/1609.08194v1.pdf | Online Segment to Segment Neural Transduction | We introduce an online neural sequence to sequence model that learns to
alternate between encoding and decoding segments of the input as it is read. By
independently tracking the encoding and decoding representations our algorithm
permits exact polynomial marginalization of the latent segmentation during
training, and during decoding beam search is employed to find the best
alignment path together with the predicted output sequence. Our model tackles
the bottleneck of vanilla encoder-decoders that have to read and memorize the
entire input sequence in their fixed-length hidden states before producing any
output. It is different from previous attentive models in that, instead of
treating the attention weights as output of a deterministic function, our model
assigns attention weights to a sequential latent variable which can be
marginalized out and permits online generation. Experiments on abstractive
sentence summarization and morphological inflection show significant
performance gains over the baseline encoder-decoders. | ['Lei Yu', 'Jan Buys', 'Phil Blunsom'] | 2016-09-26 | online-segment-to-segment-neural-transduction-1 | https://aclanthology.org/D16-1138 | https://aclanthology.org/D16-1138.pdf | emnlp-2016-11 | ['abstractive-sentence-summarization'] | ['natural-language-processing'] | [ 9.33064282e-01 8.77809346e-01 -3.55141640e-01 -3.74396622e-01
-1.23183644e+00 -9.01189506e-01 5.08644342e-01 1.24682248e-01
-5.79111695e-01 7.79073417e-01 5.39491832e-01 -6.77431941e-01
6.42884791e-01 -7.02130198e-01 -1.22632313e+00 -5.43459415e-01
1.37414977e-01 8.44116569e-01 -1.91327155e-01 1.01666056e-01
3.78513694e-01 -1.39758646e-01 -9.25824106e-01 5.19499183e-01
9.03728902e-01 3.66673052e-01 7.06356347e-01 1.36358988e+00
-3.34324300e-01 7.51801431e-01 -7.29301512e-01 -5.55774391e-01
5.41552790e-02 -9.26115811e-01 -1.27039051e+00 2.00732440e-01
3.68744850e-01 -5.23782730e-01 -2.76346087e-01 9.96270180e-01
3.45614046e-01 -1.21031655e-02 7.86090374e-01 -6.67167306e-01
-1.07396185e+00 1.17146182e+00 -3.69358927e-01 2.47095436e-01
3.80486459e-01 4.08334017e-01 1.38272214e+00 -5.50975144e-01
7.15167105e-01 1.05478585e+00 2.68739849e-01 7.15895593e-01
-1.54785824e+00 1.72088332e-02 3.01710039e-01 7.95331821e-02
-9.29151535e-01 -6.60187006e-01 2.14524060e-01 -4.32353795e-01
1.79167724e+00 2.17816085e-01 6.04536712e-01 9.46575999e-01
4.37830180e-01 1.12049103e+00 5.00820160e-01 -6.26864791e-01
2.31196702e-01 -4.09428746e-01 2.75204331e-01 8.25328529e-01
-2.68737435e-01 -3.81845355e-01 -5.07655025e-01 8.12691525e-02
6.14657521e-01 -4.56632078e-01 -4.23194528e-01 1.67378440e-01
-1.10410786e+00 8.13553274e-01 1.24109715e-01 -8.43462199e-02
-4.13621575e-01 4.46717024e-01 4.30383146e-01 3.75761002e-01
4.62347537e-01 4.03895766e-01 -4.73024428e-01 -5.32724321e-01
-1.41095626e+00 3.58848237e-02 9.01245832e-01 1.09126997e+00
6.12879157e-01 -3.83498035e-02 -5.28528512e-01 6.17331862e-01
7.15844706e-02 3.03900123e-01 5.66203952e-01 -9.97191668e-01
9.07510698e-01 2.42849737e-01 8.87138546e-02 -1.77481636e-01
8.19480270e-02 -3.47951740e-01 -6.66859806e-01 -1.57462820e-01
3.76537502e-01 -2.52910078e-01 -1.24323499e+00 1.86984253e+00
-1.98284715e-01 5.26712090e-02 1.29643187e-01 6.34233832e-01
1.74836606e-01 1.19837260e+00 -2.51179151e-02 -4.74427849e-01
1.14135098e+00 -1.37024379e+00 -7.00924397e-01 -6.90921903e-01
6.65509224e-01 -4.73187774e-01 9.78061557e-01 2.26308733e-01
-2.03265405e+00 -3.91858757e-01 -1.08878601e+00 -6.66727662e-01
2.24318847e-01 -3.61708775e-02 1.80539951e-01 1.42148957e-01
-1.56113791e+00 8.69504809e-01 -1.26970589e+00 4.10446413e-02
3.00566107e-01 4.00213659e-01 -1.24462306e-01 2.08891153e-01
-9.41805422e-01 9.70766902e-01 6.54723704e-01 2.55752206e-01
-9.53401744e-01 -4.49187517e-01 -8.35903764e-01 2.97273248e-01
3.67547981e-02 -1.22446156e+00 1.88788950e+00 -1.35569012e+00
-1.87187743e+00 8.07491899e-01 -8.10481787e-01 -8.75084519e-01
3.81393969e-01 -3.89145017e-01 3.83712977e-01 1.35456488e-01
-4.02225628e-02 8.44519615e-01 8.46264005e-01 -6.82542205e-01
-4.75910753e-01 6.70841113e-02 -2.09885295e-02 5.79368174e-01
2.60155439e-01 5.04472665e-02 -6.29405737e-01 -4.76532102e-01
1.23591255e-02 -8.51086438e-01 -3.08755279e-01 -3.94385725e-01
-7.97259390e-01 -1.45551311e-02 8.78821872e-03 -1.17919290e+00
1.39677775e+00 -1.75680232e+00 9.82367218e-01 -1.85533285e-01
1.08625146e-03 8.12410265e-02 -2.48171180e-01 4.95218486e-01
4.33053588e-03 3.04595470e-01 -7.08785355e-01 -7.21272767e-01
9.69903693e-02 5.08549988e-01 -5.35797954e-01 2.94496328e-01
3.75934333e-01 1.37083101e+00 -9.58319485e-01 -3.94492477e-01
-3.24663877e-01 5.55797433e-03 -7.70212293e-01 5.48063099e-01
-7.41721153e-01 2.60937721e-01 -6.48189932e-02 8.75355378e-02
3.21941495e-01 -3.99928182e-01 3.90283436e-01 4.93539661e-01
-1.58204153e-01 9.95735228e-01 -5.39705276e-01 2.07429743e+00
-4.47915912e-01 7.49203563e-01 9.49154347e-02 -9.01761472e-01
3.71455431e-01 4.93170112e-01 -3.72697592e-01 -4.32097673e-01
4.33857851e-02 8.90083015e-02 -2.68650502e-02 -3.28567713e-01
8.86036694e-01 -4.79786545e-02 -2.44048417e-01 8.96119058e-01
2.25202009e-01 -1.68832261e-02 1.33575425e-01 4.95098591e-01
9.64808643e-01 5.33482075e-01 2.96657115e-01 -2.41728164e-02
2.91602612e-01 6.36494979e-02 4.13522422e-01 9.05437768e-01
3.44232857e-01 7.63102889e-01 9.94310558e-01 -5.45107648e-02
-1.55520773e+00 -1.19767511e+00 4.08052385e-01 1.23855174e+00
-3.27477515e-01 -2.50195116e-01 -1.30684054e+00 -4.03633565e-01
-4.81690943e-01 1.28842723e+00 -4.89794195e-01 -9.63593423e-02
-1.13984120e+00 -5.95826685e-01 6.75601840e-01 6.49025321e-01
-2.22609621e-02 -1.27777731e+00 -6.32167220e-01 5.12278140e-01
-5.22178471e-01 -6.02184713e-01 -9.46606755e-01 6.16310477e-01
-1.05105293e+00 -4.25642401e-01 -6.20896518e-01 -1.03090930e+00
9.79527950e-01 -3.90424222e-01 1.15996265e+00 6.86631352e-02
-6.96519343e-03 -1.35976613e-01 -5.13365231e-02 1.71930809e-02
-9.24392939e-01 6.61344409e-01 -5.78620851e-01 -2.51533031e-01
-5.02470993e-02 -5.81807733e-01 -4.41244632e-01 -4.03998226e-01
-7.37753093e-01 8.74193668e-01 6.62488937e-01 8.58060658e-01
5.78937232e-01 -7.53134012e-01 3.00400436e-01 -1.14579022e+00
6.63148105e-01 -6.61109626e-01 -4.98233825e-01 4.55974609e-01
-2.93605685e-01 6.86480105e-01 9.01742280e-01 -2.83398390e-01
-1.03765857e+00 4.69740517e-02 -2.86093265e-01 5.92501573e-02
-1.35267407e-01 5.74037552e-01 -1.91050321e-01 8.91549647e-01
2.72937357e-01 5.65890908e-01 1.60557315e-01 -5.38194656e-01
7.60191560e-01 5.89442134e-01 9.35739279e-01 -4.64485943e-01
5.53682208e-01 -1.22067854e-01 -4.22012776e-01 -4.52023089e-01
-7.94291019e-01 -2.10945830e-02 -9.90223527e-01 2.43845627e-01
1.15546489e+00 -6.74742758e-01 -4.28869545e-01 5.04275739e-01
-1.78358340e+00 -8.70380521e-01 -5.43666422e-01 6.08325824e-02
-9.84813094e-01 4.76034969e-01 -1.21937966e+00 -7.77617335e-01
-5.93770325e-01 -1.10232937e+00 1.09280396e+00 8.75509977e-02
-6.42347515e-01 -9.63898063e-01 4.21666354e-02 1.55176327e-01
3.01737458e-01 -1.22910179e-01 1.35864770e+00 -5.31769753e-01
-1.01507151e+00 2.99436171e-02 -2.04594471e-02 2.76929915e-01
-2.70419627e-01 -6.78381100e-02 -6.88686132e-01 -2.27017924e-01
-5.45670614e-02 -3.36081177e-01 1.18606055e+00 3.16308677e-01
9.50450659e-01 -9.44651723e-01 -1.75005347e-01 7.17741489e-01
1.18551695e+00 8.01979005e-02 7.35865891e-01 9.39317346e-02
5.67920923e-01 3.81201565e-01 2.09420081e-02 1.15802959e-01
4.33088720e-01 2.99042702e-01 2.86895335e-01 2.34255344e-01
-3.17257121e-02 -7.00932443e-01 8.28697979e-01 1.29348111e+00
3.69337983e-02 -9.13145125e-01 -6.97705567e-01 6.60955429e-01
-1.91489875e+00 -1.06678391e+00 -8.53846839e-04 2.12463450e+00
1.30862677e+00 1.66842565e-01 -2.06017748e-01 -1.49162650e-01
6.03413880e-01 3.06534559e-01 -5.99804699e-01 -1.12242377e+00
5.30392751e-02 5.90065956e-01 7.80094385e-01 1.18010366e+00
-7.61805236e-01 1.27091932e+00 7.07956123e+00 5.61874986e-01
-8.50001991e-01 2.62704074e-01 7.48003900e-01 -4.90476638e-01
-6.23186946e-01 2.38470495e-01 -7.41921246e-01 4.65117544e-01
1.53999221e+00 -1.15573164e-02 9.04916584e-01 3.62943202e-01
3.02136570e-01 -1.36486426e-01 -1.42657137e+00 3.25681388e-01
4.63276841e-02 -1.33794379e+00 2.05108285e-01 -4.45413329e-02
5.60677707e-01 3.14327255e-02 -2.42215465e-04 6.26348332e-02
6.11228764e-01 -1.23628485e+00 1.16253316e+00 6.69273734e-01
9.55024362e-01 -6.06563687e-01 2.67023623e-01 8.06460679e-01
-5.72922111e-01 4.39913049e-02 -3.80079478e-01 -3.82519513e-01
7.79826045e-01 -7.78618976e-02 -1.10187972e+00 2.97900707e-01
-2.25690037e-01 4.63002950e-01 -1.78489968e-01 6.20175838e-01
-8.23970735e-01 1.07173967e+00 4.78182882e-02 9.21980664e-02
3.74109387e-01 -3.90578777e-01 5.68865895e-01 1.54541755e+00
2.17087299e-01 4.23849113e-02 9.86191779e-02 9.05215740e-01
-2.97865778e-01 -2.61447616e-02 -9.32944268e-02 -2.78003931e-01
1.82233036e-01 7.10017920e-01 -6.13237321e-01 -6.72845721e-01
-1.01634212e-01 1.82818115e+00 7.29543567e-01 5.35620153e-01
-8.08183849e-01 -3.26242566e-01 3.99985313e-01 -5.69504453e-03
6.91603065e-01 -4.99282748e-01 -5.21939397e-01 -1.17473662e+00
-1.27614453e-01 -6.69054508e-01 1.41114905e-01 -8.17124844e-01
-5.47779918e-01 4.68826950e-01 -1.50720134e-01 -3.74444187e-01
-6.98935628e-01 -1.88066259e-01 -8.77684295e-01 1.46776295e+00
-1.42049527e+00 -9.01904762e-01 5.13112307e-01 -2.36671805e-01
1.04439366e+00 2.88609743e-01 8.81753504e-01 -1.55788898e-01
-6.57479525e-01 6.39010012e-01 2.30995566e-01 1.05716005e-01
2.27014601e-01 -1.62542546e+00 1.32464266e+00 1.12948859e+00
8.70485082e-02 8.04842889e-01 7.21754014e-01 -7.59038091e-01
-1.30677080e+00 -1.02595222e+00 1.52487516e+00 -3.49338919e-01
5.35510421e-01 -5.66573381e-01 -9.93156374e-01 1.28769970e+00
9.21151638e-01 -7.69023001e-01 5.20566344e-01 -3.50922704e-01
-9.71378461e-02 6.42302752e-01 -5.35054624e-01 6.56644464e-01
1.15616333e+00 -5.73185384e-01 -8.76460791e-01 2.99201697e-01
1.07637620e+00 -6.82219207e-01 -3.49160731e-01 -3.83459866e-01
4.21903938e-01 -4.94087100e-01 5.47903895e-01 -8.13628078e-01
9.50002730e-01 -1.39698431e-01 1.40974194e-01 -1.34452856e+00
-4.82220888e-01 -1.02326024e+00 -3.95781845e-01 9.84746277e-01
9.10345852e-01 -3.71535391e-01 8.17277849e-01 4.42522854e-01
-5.68521976e-01 -7.22794235e-01 -8.52309287e-01 -1.98049873e-01
2.96940804e-01 -2.07959443e-01 4.69052672e-01 2.13037163e-01
3.79792303e-02 6.69331312e-01 -3.59054953e-01 7.25452080e-02
4.09355521e-01 3.02922904e-01 4.37245637e-01 -5.48787236e-01
-9.66309845e-01 -5.50068676e-01 2.39709526e-01 -1.97104800e+00
2.78071046e-01 -1.36664140e+00 6.03747249e-01 -1.89096022e+00
3.32316428e-01 1.80803195e-01 2.50830166e-02 5.62396109e-01
-3.79484862e-01 -7.24561363e-02 1.59834072e-01 1.89320356e-01
-5.61358273e-01 3.75455648e-01 9.53257740e-01 -1.08714245e-01
-8.00703019e-02 -2.07986701e-02 -6.97638214e-01 3.54551017e-01
6.79362595e-01 -5.62081099e-01 -4.08102661e-01 -1.05087340e+00
4.07566726e-01 7.23910928e-01 5.53618036e-02 -2.44208500e-01
4.21513796e-01 -1.52175769e-01 1.64935231e-01 -6.42317832e-01
1.82677045e-01 2.74451412e-02 4.67132963e-02 4.88584489e-01
-1.01837015e+00 3.83221895e-01 -1.52589589e-01 5.54675817e-01
1.35452822e-01 -6.86210811e-01 5.71184993e-01 -4.74551052e-01
-1.99309379e-01 -1.35497283e-02 -9.73889947e-01 2.97264129e-01
5.14764309e-01 -3.67441148e-01 -1.17518939e-01 -4.81248170e-01
-9.08990264e-01 2.99009949e-01 7.42290616e-01 1.30895644e-01
3.87109786e-01 -6.27270639e-01 -7.40042329e-01 1.98754340e-01
-5.93318820e-01 4.69125986e-01 5.04868403e-02 5.44086039e-01
-8.47363889e-01 5.22998393e-01 -4.23016064e-02 -2.21157506e-01
-1.09729779e+00 4.32050079e-01 1.76788062e-01 -5.75404286e-01
-5.30426919e-01 1.09379053e+00 1.14504606e-01 -1.33414671e-01
2.71006078e-01 -4.88949835e-01 1.92847177e-01 -1.29014000e-01
5.35118043e-01 2.85298765e-01 -1.80037379e-01 -3.99071306e-01
1.33082747e-01 9.14407745e-02 -3.37266028e-01 -5.12885988e-01
1.28003824e+00 -3.18277150e-01 -5.20568013e-01 6.97759807e-01
1.20028436e+00 -1.42108679e-01 -1.51853335e+00 -3.34881544e-02
2.22623587e-01 1.21626273e-01 -2.75161296e-01 -8.31006348e-01
-4.71361369e-01 1.14050972e+00 -3.05006325e-01 -1.28000244e-01
8.06254268e-01 -1.66725032e-02 1.26854551e+00 1.44605547e-01
-9.20037106e-02 -1.01287508e+00 -1.59921721e-01 9.43535089e-01
5.02548277e-01 -5.30286789e-01 -3.67447674e-01 -7.92618766e-02
-6.21302545e-01 1.14730835e+00 3.12293142e-01 -1.87006563e-01
-1.18758366e-01 5.54868937e-01 -2.47380480e-01 2.95170546e-01
-1.27422941e+00 2.85123974e-01 6.00854382e-02 1.59351349e-01
7.04787970e-01 1.33759484e-01 -4.07368243e-01 4.54888165e-01
-5.75358808e-01 -1.39477387e-01 7.40674973e-01 9.98803914e-01
-7.29780793e-01 -1.33279192e+00 -2.47487798e-02 4.12117362e-01
-6.87212884e-01 -6.03865981e-01 -2.98504651e-01 4.10728641e-02
-2.27937803e-01 5.24722219e-01 5.63573420e-01 9.04157311e-02
-1.13888256e-01 8.38359714e-01 5.68826318e-01 -1.07237947e+00
-9.02372837e-01 7.85092339e-02 1.17017858e-01 -2.94438004e-01
3.31935674e-01 -7.70713866e-01 -1.51639402e+00 -9.31571573e-02
-2.10695639e-01 2.95651168e-01 5.51326096e-01 1.03307092e+00
3.28492999e-01 6.65388525e-01 3.50113779e-01 -8.54901910e-01
-9.03263390e-01 -9.20174241e-01 9.97257158e-02 6.56819269e-02
6.10520482e-01 3.36087644e-01 -8.27178881e-02 5.29162765e-01] | [11.515830039978027, 9.059093475341797] |
d50893fe-5787-4373-8b38-bc56b4c1559d | deep-neural-network-techniques-for-monaural | 2212.00369 | null | https://arxiv.org/abs/2212.00369v2 | https://arxiv.org/pdf/2212.00369v2.pdf | Deep neural network techniques for monaural speech enhancement: state of the art analysis | Deep neural networks (DNN) techniques have become pervasive in domains such as natural language processing and computer vision. They have achieved great success in these domains in task such as machine translation and image generation. Due to their success, these data driven techniques have been applied in audio domain. More specifically, DNN models have been applied in speech enhancement domain to achieve denosing, dereverberation and multi-speaker separation in monaural speech enhancement. In this paper, we review some dominant DNN techniques being employed to achieve speech separation. The review looks at the whole pipeline of speech enhancement from feature extraction, how DNN based tools are modelling both global and local features of speech and model training (supervised and unsupervised). We also review the use of speech-enhancement pre-trained models to boost speech enhancement process. The review is geared towards covering the dominant trends with regards to DNN application in speech enhancement in speech obtained via a single speaker. | ['Peter Ochieng'] | 2022-12-01 | null | null | null | null | ['art-analysis', 'speech-separation', 'speaker-separation'] | ['computer-vision', 'speech', 'speech'] | [ 4.03088987e-01 1.69845410e-02 3.29150796e-01 -3.19200158e-01
-7.07944572e-01 -1.91427514e-01 7.79623568e-01 -1.69581160e-01
-4.31761742e-01 3.99831593e-01 8.99746776e-01 -1.46623492e-01
-1.92963690e-01 -4.36227024e-01 -2.15932310e-01 -7.93682992e-01
2.20420584e-01 -1.33851439e-01 -1.63122952e-01 -6.19983912e-01
-2.30499264e-03 5.40877879e-01 -2.00724745e+00 4.61897194e-01
4.00739759e-01 7.71853268e-01 4.96645361e-01 1.35202360e+00
-1.20062009e-02 8.52613986e-01 -8.18317175e-01 -1.49261668e-01
1.80003315e-01 -6.42201781e-01 -5.35204768e-01 2.89182007e-01
1.40148297e-01 -2.95606196e-01 -5.79555213e-01 1.07614124e+00
1.41904378e+00 3.85279924e-01 4.67626065e-01 -7.44424403e-01
-5.12362480e-01 5.48986852e-01 -2.87402302e-01 5.85628033e-01
7.80466869e-02 -1.71951413e-01 5.10904908e-01 -1.04112983e+00
2.32129246e-01 1.20099461e+00 5.53650677e-01 6.88758671e-01
-6.53935194e-01 -3.42011988e-01 -5.54736853e-01 4.51343268e-01
-7.55514324e-01 -9.66419697e-01 8.36165369e-01 -3.46078314e-02
1.39254618e+00 3.12506288e-01 3.89925480e-01 1.00139415e+00
-7.03923553e-02 8.25034678e-01 1.29499173e+00 -8.78042042e-01
5.92839122e-02 6.85629994e-02 -2.25360706e-01 -2.77445197e-01
-4.50737208e-01 6.61829948e-01 -9.45724726e-01 4.42924649e-01
6.25194371e-01 -4.83815372e-01 1.89936161e-03 4.25645500e-01
-8.08413148e-01 6.89602256e-01 -1.63989812e-02 8.05373430e-01
-8.13902438e-01 1.34372516e-02 7.42279947e-01 6.06664777e-01
6.52549148e-01 2.11111754e-01 -4.74100411e-01 -4.62055981e-01
-1.27537203e+00 2.40372777e-01 3.09605598e-01 6.88612700e-01
1.75485581e-01 9.21276331e-01 -2.10028380e-01 1.52123570e+00
2.85636812e-01 5.57181954e-01 8.07119727e-01 -9.11214411e-01
1.21883847e-01 -2.03341827e-01 -2.94380307e-01 -5.14374018e-01
-1.92504510e-01 -6.48245871e-01 -8.98603201e-01 6.05988026e-01
-2.97476768e-01 -3.87616336e-01 -1.36308742e+00 1.39490342e+00
3.73258054e-01 1.39084205e-01 4.71848249e-01 8.87451828e-01
1.16247797e+00 8.93791854e-01 7.71191046e-02 -2.30022371e-01
1.48653972e+00 -9.29833829e-01 -1.27458274e+00 -1.83154285e-01
-2.24609543e-02 -1.59828651e+00 3.64327312e-01 7.41344452e-01
-1.45216012e+00 -9.83510315e-01 -9.57528532e-01 -2.31207415e-01
-4.67067897e-01 -1.20465316e-01 1.34063721e-01 9.00312185e-01
-1.53687155e+00 5.46454012e-01 -6.80174053e-01 -4.18247074e-01
2.98508674e-01 4.81788188e-01 -3.33441138e-01 3.00845087e-01
-1.37181568e+00 1.11475587e+00 2.80359268e-01 1.72783092e-01
-1.32162476e+00 -3.17223489e-01 -7.91595459e-01 -6.26619114e-03
-9.65851992e-02 -5.01430392e-01 1.77246809e+00 -1.06403112e+00
-1.83307278e+00 8.30305815e-01 -1.60647631e-01 -8.65041316e-01
6.45290241e-02 -2.13046774e-01 -9.35654879e-01 8.47680271e-02
-4.23590302e-01 7.35994339e-01 1.23822248e+00 -9.28440690e-01
-8.02033246e-01 -1.66700527e-01 -3.11384946e-01 7.63743401e-01
-2.41139188e-01 7.55620062e-01 1.85959965e-01 -1.01482463e+00
-9.22538340e-03 -2.14966506e-01 6.78143278e-02 -4.24265087e-01
-5.53888120e-02 -7.26547316e-02 1.37736595e+00 -1.29195440e+00
9.66171205e-01 -2.20903349e+00 -3.11740004e-02 -2.32435420e-01
1.35644510e-01 9.20158982e-01 -1.59806743e-01 5.80309927e-01
-4.33827102e-01 -3.54676157e-01 -1.41040847e-01 -6.81109190e-01
6.77766576e-02 3.42707455e-01 -2.15945765e-01 1.99087903e-01
1.20024659e-01 6.23819232e-01 -5.06953955e-01 -3.68839562e-01
8.22935641e-01 1.20629275e+00 -2.78233826e-01 2.78541952e-01
2.92886585e-01 4.15222317e-01 3.43004137e-01 5.29763281e-01
8.64515483e-01 8.51135671e-01 -3.03850740e-01 -1.62447155e-01
-3.73383373e-01 4.91220653e-01 -1.22209620e+00 1.54766309e+00
-6.20595038e-01 1.18915093e+00 6.87962174e-01 -1.24004948e+00
1.03417528e+00 1.10957158e+00 2.55422741e-01 -7.43618667e-01
4.22401071e-01 3.73884439e-01 2.85636067e-01 -6.54976189e-01
5.87455332e-01 -7.05561817e-01 6.93083763e-01 1.01421289e-01
7.77529955e-01 -3.16669732e-01 -5.20377327e-03 -2.34369189e-01
7.19412506e-01 -2.42868453e-01 4.11865234e-01 2.55379155e-02
7.96756864e-01 -4.91904616e-01 1.72395572e-01 5.02138317e-01
-6.54478729e-01 7.98893273e-01 -1.44771472e-01 -7.09158927e-02
-1.44774187e+00 -8.77844214e-01 -1.18473664e-01 1.37282145e+00
-5.34232914e-01 4.28404100e-02 -1.00535393e+00 4.70168963e-02
-8.26210558e-01 6.13175333e-01 -3.11984748e-01 -1.17893741e-01
-4.45007324e-01 -5.81680775e-01 8.10645938e-01 4.43266809e-01
6.04937792e-01 -1.59909511e+00 -2.24767223e-01 4.87721890e-01
-1.20944068e-01 -9.70661402e-01 -8.92176032e-02 4.71937269e-01
-9.01450038e-01 -2.39937112e-01 -1.25703061e+00 -1.26815534e+00
4.15238552e-02 1.53888717e-01 7.65296400e-01 -4.27983403e-01
-3.44750941e-01 4.09503043e-01 -4.89429623e-01 -1.02871215e+00
-8.53567183e-01 -2.06358582e-01 2.46062160e-01 -1.19497627e-01
6.18105650e-01 -1.04154885e+00 -5.96676469e-01 -2.44530700e-02
-1.12422919e+00 -3.01711410e-01 7.43011475e-01 8.13915670e-01
2.48800531e-01 4.13474202e-01 7.97862053e-01 -5.39927594e-02
1.15003991e+00 -6.18576892e-02 -2.55483210e-01 -2.96453029e-01
-2.59067506e-01 -3.95794660e-01 3.59686881e-01 -3.94777834e-01
-1.40547490e+00 -2.67249137e-01 -9.25460577e-01 -2.74951249e-01
-8.04291546e-01 4.19342250e-01 -1.22390740e-01 6.96826801e-02
8.26862216e-01 4.67604250e-01 2.67321110e-01 -7.69363940e-01
2.30556190e-01 1.32016563e+00 6.73505425e-01 3.42100300e-02
4.09800410e-01 2.51058102e-01 -2.07523331e-01 -1.25068903e+00
-3.47723454e-01 -7.51329720e-01 -3.42917293e-01 -4.04998451e-01
9.33583140e-01 -8.99894118e-01 -1.69315264e-01 1.00637329e+00
-1.26550424e+00 -1.75367013e-01 -5.60082853e-01 7.01774657e-01
-4.63356525e-01 1.20720915e-01 -6.84390843e-01 -1.09833109e+00
-7.64119208e-01 -1.19838834e+00 8.88999641e-01 5.96022546e-01
8.00430682e-03 -1.10574889e+00 3.51242155e-01 3.02441806e-01
9.34259295e-01 -2.65433460e-01 3.60307336e-01 -7.89373696e-01
2.26681173e-01 -4.01504710e-02 1.85272649e-01 1.29943836e+00
3.01240474e-01 -2.99322575e-01 -1.76766872e+00 5.33811226e-02
4.82616454e-01 -3.01918164e-02 5.03110170e-01 1.02454448e+00
6.33181691e-01 -2.92128175e-02 2.78859675e-01 4.12602961e-01
1.09426463e+00 4.94800240e-01 9.08387661e-01 2.02480569e-01
1.54603630e-01 7.74578273e-01 3.84232312e-01 2.77473599e-01
-3.85207325e-01 3.60502094e-01 3.24034035e-01 -3.95093560e-01
-9.18583453e-01 1.28491759e-01 3.95617545e-01 1.21497202e+00
-2.47067183e-01 -2.49644443e-01 -6.63919687e-01 8.95072401e-01
-1.26008594e+00 -1.08923769e+00 -6.58604205e-02 1.79820037e+00
7.93125510e-01 -5.10494560e-02 8.71433169e-02 7.96254396e-01
9.79022503e-01 2.56095141e-01 -8.19858089e-02 -1.13151169e+00
-4.71563190e-01 9.30147409e-01 1.22154415e-01 7.96242893e-01
-1.14891624e+00 7.92582214e-01 6.38672638e+00 1.07227743e+00
-1.23982716e+00 4.87332821e-01 4.06339645e-01 -3.60340364e-02
2.24281445e-01 -5.21731496e-01 -4.36027646e-01 5.52146100e-02
1.35358548e+00 5.46162874e-02 4.78314638e-01 6.64693236e-01
7.20860779e-01 3.51264179e-02 -4.78143990e-01 1.24843645e+00
-3.76752857e-03 -1.11642301e+00 -1.66906282e-01 -2.55056620e-02
8.47806692e-01 2.71474183e-01 3.63801390e-01 2.84138381e-01
-3.15089040e-02 -1.08725631e+00 5.04563749e-01 1.27598315e-01
6.71184480e-01 -1.11865020e+00 9.48550522e-01 2.39769936e-01
-9.35852468e-01 1.75666134e-03 -3.22859526e-01 -1.42876148e-01
4.66192067e-01 6.78406179e-01 -1.07825565e+00 4.71957415e-01
8.70271027e-01 3.81458342e-01 1.62307605e-01 1.05730438e+00
-2.48056948e-01 7.57669687e-01 -2.27884129e-01 2.94209689e-01
2.29439527e-01 1.00411795e-01 9.39059496e-01 1.44333136e+00
4.12916362e-01 6.16447628e-02 -7.29194283e-01 4.00271565e-01
8.86912048e-02 1.00506775e-01 -4.76197869e-01 -1.57243013e-01
1.86245188e-01 1.39845693e+00 -3.39810699e-01 -3.62506777e-01
-2.92650640e-01 8.11900616e-01 -4.96578246e-01 3.38881373e-01
-6.32158279e-01 -8.51249635e-01 8.30045700e-01 1.34837016e-01
3.75371575e-01 -4.81810197e-02 -2.81712443e-01 -3.59172881e-01
-2.63703197e-01 -1.21025729e+00 -7.95753449e-02 -1.15732789e+00
-8.39824200e-01 9.93167877e-01 -1.33748308e-01 -1.10853529e+00
-3.97840023e-01 -9.47811782e-01 -6.79387152e-01 1.30829835e+00
-1.41313136e+00 -1.04343176e+00 -2.50505395e-02 5.98751307e-01
9.83318448e-01 -4.65536565e-01 6.82335198e-01 8.72136712e-01
-5.46281859e-02 2.57111341e-01 1.23576216e-01 3.50152552e-02
8.34966481e-01 -1.18016839e+00 3.18097562e-01 1.15135324e+00
1.32472560e-01 4.08909768e-01 1.10793698e+00 -2.58424550e-01
-6.80048466e-01 -8.19127381e-01 1.23077631e+00 2.25532174e-01
2.38523260e-01 -1.66312322e-01 -5.90677679e-01 3.18274170e-01
1.05989039e+00 -2.79780179e-01 5.75788796e-01 -3.44876468e-01
3.38218898e-01 -3.73788476e-01 -1.20789528e+00 4.01949465e-01
4.97552186e-01 -6.95881128e-01 -7.03030467e-01 9.28441286e-02
5.28454244e-01 -5.79225600e-01 -6.42688394e-01 2.76465535e-01
1.90219983e-01 -1.11322391e+00 1.07420242e+00 -5.19033432e-01
5.48779488e-01 -3.03912163e-01 -1.98530212e-01 -1.57560432e+00
-1.11452090e-02 -1.06612003e+00 -9.61215645e-02 1.46328104e+00
1.20273806e-01 -4.42914486e-01 4.06915933e-01 -8.41546953e-02
-6.68970764e-01 -2.68655568e-01 -1.04012609e+00 -4.54427689e-01
-1.74004689e-01 -7.81831741e-01 4.41999137e-01 6.65014863e-01
-2.94073075e-01 4.03006166e-01 -6.25358224e-01 1.10581025e-01
4.02698189e-01 -9.31710362e-01 4.46141362e-01 -9.14238393e-01
-2.18418285e-01 -3.98574620e-01 -6.66178405e-01 -8.47651660e-01
-2.64042407e-01 -5.04838943e-01 2.43797496e-01 -1.78853405e+00
-2.68476963e-01 4.40404654e-01 -3.57053459e-01 1.06870867e-01
1.22224800e-01 2.88209051e-01 -7.82290027e-02 -3.81696880e-01
8.52137357e-02 4.47618306e-01 1.25316405e+00 2.12109205e-03
-1.85913682e-01 1.76431268e-01 -8.15705717e-01 6.58968329e-01
1.11016881e+00 -3.78204197e-01 -5.22783518e-01 -4.78252023e-01
-2.28406973e-02 -7.61251301e-02 3.72391611e-01 -1.21430695e+00
2.98463166e-01 3.67332429e-01 2.18102366e-01 -7.36900747e-01
6.08230650e-01 -8.33975673e-01 -6.27939682e-03 1.24702767e-01
-3.08654398e-01 -7.22153410e-02 4.50711817e-01 1.29055828e-01
-9.60211694e-01 -3.98634017e-01 1.23826814e+00 -5.12460098e-02
-7.86201775e-01 -8.92767236e-02 -9.58198488e-01 -3.23272079e-01
6.09463692e-01 -3.71807605e-01 1.58612907e-01 -9.03309405e-01
-1.21197510e+00 -4.94967908e-01 -4.48571861e-01 3.29446077e-01
7.14849353e-01 -1.22388554e+00 -9.14966464e-01 1.82094038e-01
-4.85485822e-01 -3.94548588e-02 7.13574827e-01 1.06866419e+00
-5.61669767e-01 4.82023120e-01 -3.06604803e-01 -5.31421065e-01
-1.77743983e+00 3.35972577e-01 7.20853686e-01 -1.50048379e-02
-3.10693175e-01 1.17619145e+00 1.26381516e-02 -4.20209259e-01
3.34124625e-01 -1.98439538e-01 -4.72234428e-01 -1.39417246e-01
7.61821866e-01 7.57802784e-01 6.42865121e-01 -7.24113822e-01
-6.63108230e-02 2.17843086e-01 -9.56684873e-02 -8.03448319e-01
1.63471413e+00 -4.34437066e-01 -1.50829241e-01 7.18726441e-02
1.14923251e+00 1.83819011e-02 -9.61015224e-01 -1.17322914e-01
-3.31293404e-01 4.49641943e-02 7.60756731e-01 -1.01456761e+00
-1.01597071e+00 1.46191370e+00 1.09231365e+00 3.44922394e-01
1.81549621e+00 -1.92880288e-01 7.19131708e-01 -1.13957182e-01
-3.87883723e-01 -1.51709652e+00 2.08471753e-02 7.39381731e-01
1.00321114e+00 -1.08123040e+00 -3.37976962e-01 6.59535974e-02
-5.73115706e-01 1.23949552e+00 1.71296552e-01 4.40220870e-02
8.14037263e-01 6.69249117e-01 4.82117504e-01 -6.13813847e-02
-4.14433539e-01 -4.32718903e-01 2.91275769e-01 1.15177369e+00
7.98972428e-01 -1.59717962e-01 -3.63508537e-02 4.08622473e-01
-5.03487766e-01 -2.70866584e-02 2.87400246e-01 9.38763559e-01
-7.06008852e-01 -1.29722095e+00 -8.81999910e-01 3.58576551e-02
-9.06053901e-01 -5.25095940e-01 -3.45466256e-01 5.00868201e-01
3.11684698e-01 1.29921794e+00 -1.44625813e-01 -2.97200173e-01
4.85881627e-01 1.05744019e-01 3.67600858e-01 -3.94733161e-01
-8.55441630e-01 7.39606261e-01 2.12240383e-01 1.01542160e-01
-7.12811470e-01 -5.84237039e-01 -8.40996802e-01 -2.93501675e-01
-3.10561001e-01 7.04540834e-02 1.11137044e+00 9.08295333e-01
1.87514707e-01 1.07938516e+00 4.10939038e-01 -1.03762996e+00
-3.45689297e-01 -1.55445707e+00 -8.30537558e-01 -1.81267392e-02
7.48388886e-01 -2.12226465e-01 -4.77990776e-01 4.01965290e-01] | [15.118825912475586, 5.901556491851807] |
ba09d6de-3d41-4d17-ba70-caa89cd5d12a | robust-neural-architecture-search | 2304.02845 | null | https://arxiv.org/abs/2304.02845v2 | https://arxiv.org/pdf/2304.02845v2.pdf | Robust Neural Architecture Search | Neural Architectures Search (NAS) becomes more and more popular over these years. However, NAS-generated models tends to suffer greater vulnerability to various malicious attacks. Lots of robust NAS methods leverage adversarial training to enhance the robustness of NAS-generated models, however, they neglected the nature accuracy of NAS-generated models. In our paper, we propose a novel NAS method, Robust Neural Architecture Search (RNAS). To design a regularization term to balance accuracy and robustness, RNAS generates architectures with both high accuracy and good robustness. To reduce search cost, we further propose to use noise examples instead adversarial examples as input to search architectures. Extensive experiments show that RNAS achieves state-of-the-art (SOTA) performance on both image classification and adversarial attacks, which illustrates the proposed RNAS achieves a good tradeoff between robustness and accuracy. | ['Weiping Wang', 'Yong liu', 'Jian Li', 'Xunyu Zhu'] | 2023-04-06 | null | null | null | null | ['architecture-search'] | ['methodology'] | [-1.98305771e-01 -4.16222125e-01 -3.76852453e-02 -2.57845223e-01
-9.04249489e-01 -8.61023128e-01 5.82743645e-01 -7.07202375e-01
-2.02098459e-01 5.38512349e-01 7.17008784e-02 -4.48243499e-01
4.80623432e-02 -7.78117359e-01 -7.20142424e-01 -7.22584188e-01
3.22930783e-01 1.16480842e-01 3.66327494e-01 -4.02364165e-01
1.43116862e-01 6.85672462e-01 -9.48253393e-01 1.93958029e-01
9.27801549e-01 1.14298809e+00 -2.75712311e-01 5.98422110e-01
-8.26411881e-03 7.47623682e-01 -7.76307583e-01 -7.76683450e-01
7.32068837e-01 -1.58355460e-01 -4.67133164e-01 -6.85228407e-01
4.73582625e-01 -4.13165390e-01 -9.38952208e-01 1.33779085e+00
9.44061816e-01 3.50599959e-02 4.76069927e-01 -1.35079753e+00
-1.22898459e+00 7.81992555e-01 -3.15412015e-01 2.88943738e-01
2.24404763e-02 7.00589836e-01 8.38027775e-01 -8.55391264e-01
2.50424027e-01 1.28023124e+00 8.18446875e-01 9.26391065e-01
-8.59081089e-01 -1.33298469e+00 7.56665319e-02 1.58721000e-01
-1.26672757e+00 -5.73531508e-01 9.65353429e-01 -6.00043871e-02
8.19762349e-01 6.54898226e-01 1.54938534e-01 1.72181058e+00
1.20740674e-01 8.10719311e-01 9.66774583e-01 -2.27492433e-02
2.35054240e-01 -7.41112754e-02 -2.87199649e-03 6.56210721e-01
2.46786177e-01 5.54784656e-01 -1.29880011e-01 -3.99927586e-01
6.83229983e-01 7.10961819e-02 -1.57840535e-01 -7.09868874e-03
-8.26109827e-01 8.66319537e-01 8.39678884e-01 1.07438445e-01
-1.54143989e-01 2.72409558e-01 6.71451032e-01 4.39797848e-01
7.67745599e-02 8.59001994e-01 -4.68728155e-01 8.89613852e-02
-5.28831303e-01 1.75073102e-01 4.62172300e-01 7.46827185e-01
1.27763718e-01 7.91971087e-01 -2.13998869e-01 9.50062037e-01
2.90103316e-01 7.27210581e-01 8.99745584e-01 -6.59465551e-01
4.08058345e-01 7.44797945e-01 -3.43547344e-01 -1.13103807e+00
-1.13163650e-01 -6.70962453e-01 -1.19210422e+00 3.28930497e-01
-4.81903329e-02 -1.22442536e-01 -1.20873284e+00 1.87679636e+00
1.32960483e-01 3.34477872e-01 3.41211677e-01 9.28979456e-01
9.65689600e-01 7.69528270e-01 -3.02100219e-02 7.89459124e-02
9.67142522e-01 -1.19742477e+00 -3.40541273e-01 -2.29498848e-01
2.42416002e-02 -7.30024874e-01 1.30887234e+00 1.83526233e-01
-1.02420247e+00 -4.99892652e-01 -1.11123550e+00 4.21478808e-01
-3.92357290e-01 -9.82132182e-02 3.38805616e-01 8.95320356e-01
-7.68507659e-01 4.34391081e-01 -6.73655927e-01 6.11558892e-02
5.37503541e-01 4.52834666e-01 -2.53932625e-01 3.37252170e-01
-1.36428523e+00 7.35561907e-01 1.95999771e-01 -8.77904743e-02
-1.25911379e+00 -4.23587114e-01 -5.84360421e-01 1.82794109e-01
3.19562376e-01 -5.26046276e-01 1.13729692e+00 -8.76103878e-01
-1.55628729e+00 5.74425042e-01 5.54512858e-01 -5.91554701e-01
4.82162476e-01 -2.78804809e-01 -7.01726556e-01 -1.35838524e-01
-2.83009171e-01 3.86426121e-01 9.98636663e-01 -1.23176110e+00
4.98278253e-02 -1.78309679e-01 5.65644838e-02 -2.76761383e-01
-7.18164563e-01 2.58618087e-01 -5.50433278e-01 -1.29437423e+00
-1.36775225e-01 -1.18682981e+00 -3.81135941e-01 -6.55823275e-02
-6.17579639e-01 2.07911944e-03 9.86819685e-01 -4.55275327e-01
1.49640739e+00 -2.03816032e+00 -3.95247201e-03 3.48622203e-01
1.02659352e-01 1.03679752e+00 -6.36145234e-01 2.48352885e-01
-2.31001005e-01 5.89300275e-01 -1.97560370e-01 1.15985647e-01
7.28812590e-02 1.26772627e-01 -7.78908551e-01 1.24132685e-01
1.81255311e-01 1.11362326e+00 -6.56411231e-01 -1.60842940e-01
-1.92412943e-01 2.96834677e-01 -6.78997695e-01 5.19527793e-01
-2.09384128e-01 1.01426393e-01 -8.11207592e-01 1.04936242e+00
7.05187738e-01 -2.84721524e-01 -2.40278706e-01 -1.25773147e-01
3.91616642e-01 -1.08468369e-01 -8.17965448e-01 1.19730079e+00
-3.58141154e-01 3.58740926e-01 1.23130091e-01 -5.05672097e-01
9.43688750e-01 1.00553535e-01 9.12121087e-02 -7.31149614e-01
3.00895333e-01 2.26823166e-01 8.87456611e-02 -3.35435539e-01
8.54994804e-02 3.93014550e-01 -6.78490326e-02 3.10655534e-01
-2.85711557e-01 1.41263092e-02 -4.13869202e-01 1.07723869e-01
1.36276209e+00 -2.26570129e-01 7.57138506e-02 2.96669845e-02
6.23164773e-01 -1.71866059e-01 6.90565705e-01 8.12318861e-01
-4.26410437e-01 5.97828686e-01 1.32542074e-01 -6.52563870e-01
-1.00032985e+00 -1.14484692e+00 1.00565046e-01 1.01950812e+00
1.41281649e-01 -2.79171199e-01 -8.90994310e-01 -1.04136312e+00
-5.20190559e-02 4.37888861e-01 -6.37976408e-01 -7.54843771e-01
-8.29063475e-01 -6.51727557e-01 1.44248021e+00 5.43114185e-01
9.10697103e-01 -1.36280310e+00 -1.55893534e-01 -7.76420534e-02
1.59400359e-01 -7.69211650e-01 -8.54131937e-01 -2.11506292e-01
-5.85874915e-01 -1.06638837e+00 -6.31425142e-01 -7.98545420e-01
5.50475180e-01 7.81674162e-02 9.12356377e-01 4.74145710e-01
-2.44325936e-01 -5.36550023e-02 -3.50745529e-01 -3.37668240e-01
-6.37433112e-01 1.73742503e-01 5.02691209e-01 -2.33775169e-01
-1.15851283e-01 -6.70547307e-01 -5.65785766e-01 5.82766771e-01
-1.22115099e+00 -5.63901663e-01 9.26079571e-01 1.10186160e+00
3.08945835e-01 -2.37330034e-01 7.88962722e-01 -5.38628221e-01
1.11892128e+00 -5.37003577e-01 -7.46497095e-01 4.48760033e-01
-1.02221847e+00 2.18824357e-01 1.13724124e+00 -8.79079163e-01
-6.67252302e-01 -8.16908106e-02 -5.02934337e-01 -1.20291924e+00
6.79218844e-02 3.36492896e-01 -2.12040782e-01 -5.84888697e-01
1.13280380e+00 4.49296117e-01 1.52257755e-01 -3.91969174e-01
2.97210872e-01 6.38878703e-01 5.89396536e-01 -5.89151442e-01
1.29691422e+00 -1.07372351e-01 -1.92799285e-01 -2.24042341e-01
-5.47274113e-01 1.16421911e-03 1.11491844e-01 1.42401651e-01
4.59939629e-01 -6.69751346e-01 -8.43617797e-01 6.25201643e-01
-9.68125761e-01 -2.53362536e-01 1.13053948e-01 4.06647064e-02
-2.10252300e-01 3.86765152e-01 -7.91984081e-01 -6.36306643e-01
-1.11428165e+00 -1.47915852e+00 5.85352004e-01 3.13358158e-01
1.08163789e-01 -5.50865233e-01 1.33469954e-01 2.62541771e-01
9.68613863e-01 3.57343942e-01 8.19771826e-01 -1.06786144e+00
-7.11860180e-01 -4.19422120e-01 -2.81173795e-01 4.78942543e-01
9.90652516e-02 9.79932174e-02 -9.68051910e-01 -3.03057104e-01
8.50408077e-02 -5.61418355e-01 7.70268261e-01 -1.01777352e-01
1.47195041e+00 -1.02844596e+00 -1.91424623e-01 9.12550867e-01
1.35645163e+00 4.30917174e-01 7.83251405e-01 5.51722646e-01
6.87263548e-01 -9.41442773e-02 3.85447443e-01 6.74332306e-02
-2.50495493e-01 7.41485596e-01 9.30817902e-01 1.39356717e-01
-1.56827532e-02 -2.76342720e-01 5.59076905e-01 7.67417133e-01
2.17571422e-01 -2.39657164e-01 -1.00562465e+00 2.37644628e-01
-1.69855201e+00 -1.15652812e+00 4.83063698e-01 1.81610525e+00
1.07257116e+00 3.36288214e-01 1.06414212e-02 -1.75243229e-01
6.21304870e-01 3.11464041e-01 -1.09035945e+00 -2.95743644e-01
-1.98935583e-01 1.35902181e-01 4.54905391e-01 1.30263627e-01
-1.10752213e+00 1.20432925e+00 7.30605984e+00 1.23980439e+00
-1.31580317e+00 -7.81053081e-02 7.71819711e-01 -1.85929120e-01
-3.47654104e-01 -4.00009841e-01 -6.10441923e-01 7.55883873e-01
8.43340635e-01 -1.18223891e-01 5.68354249e-01 1.48831880e+00
-2.28584111e-01 1.13066936e+00 -5.45179546e-01 1.03396702e+00
5.09070233e-02 -1.60707259e+00 4.98106480e-01 -1.65374413e-01
7.92619705e-01 2.05083311e-01 5.16828835e-01 6.39840484e-01
7.42829084e-01 -1.20975530e+00 5.01661003e-01 2.81787992e-01
8.17351639e-01 -7.53017545e-01 6.76299572e-01 2.87162840e-01
-9.99925077e-01 -4.51932997e-01 -3.17307502e-01 3.64003301e-01
-1.66234776e-01 -6.82422239e-03 -5.62743068e-01 2.80536354e-01
6.61410093e-01 1.31652012e-01 -8.10924470e-01 9.72969472e-01
-6.28665984e-02 7.40241051e-01 -1.45976007e-01 -3.21881086e-01
5.05849361e-01 1.93767995e-03 6.40369475e-01 9.03238773e-01
2.22323179e-01 5.34364358e-02 2.87874430e-01 8.65489721e-01
-5.39782763e-01 -5.64394295e-02 -8.49310637e-01 -1.97935298e-01
9.08663988e-01 1.06897223e+00 -2.07974926e-01 -1.34013727e-01
3.68374661e-02 7.74513185e-01 3.14351052e-01 2.33130708e-01
-1.18387926e+00 -4.29348916e-01 1.02330291e+00 -1.88275039e-01
-8.67043287e-02 -3.36897373e-02 -3.74826521e-01 -1.02923882e+00
-5.69797540e-03 -1.67630172e+00 4.25095648e-01 -7.09200263e-01
-1.69425571e+00 1.04055023e+00 -5.29762387e-01 -1.37430942e+00
-7.55998641e-02 -5.33319950e-01 -9.11364019e-01 6.41806066e-01
-1.19419885e+00 -1.30042481e+00 -1.58799827e-01 6.97360277e-01
8.04558933e-01 -9.50706005e-01 9.05431509e-01 2.19274178e-01
-8.83680880e-01 1.33471024e+00 1.77459598e-01 5.70451796e-01
8.01816881e-01 -7.79302120e-01 9.50708151e-01 1.05320764e+00
3.53814662e-02 1.09569252e+00 4.26199466e-01 -6.12841964e-01
-1.57982481e+00 -1.20511925e+00 -8.03229958e-02 -5.77166378e-01
7.94305205e-01 6.53903745e-03 -9.61117744e-01 5.35125017e-01
1.65426880e-01 1.15978502e-01 5.35593629e-01 -3.20860505e-01
-1.11030173e+00 -2.94416904e-01 -1.36513186e+00 9.95740354e-01
1.04322720e+00 -6.95478082e-01 -4.81993198e-01 2.73532659e-01
1.27326238e+00 -4.79444474e-01 -5.44879913e-01 6.83979213e-01
5.11102617e-01 -8.75090897e-01 1.38277435e+00 -8.40295672e-01
2.58072227e-01 -2.86533147e-01 -2.49535099e-01 -1.30981135e+00
-4.83779550e-01 -8.58472943e-01 -2.85399824e-01 1.01237023e+00
4.28882390e-01 -8.30012739e-01 1.01423180e+00 6.49141490e-01
-2.13540271e-01 -1.17720938e+00 -7.82435596e-01 -1.05816031e+00
2.17237264e-01 -1.10216647e-01 1.01504457e+00 1.22027123e+00
-4.62720245e-01 9.70211625e-02 -6.16717339e-01 2.75150746e-01
3.92750889e-01 -1.90363050e-01 7.93004870e-01 -8.96757305e-01
-2.87264466e-01 -7.97338068e-01 -2.88805097e-01 -7.17929363e-01
3.23075831e-01 -6.63708031e-01 -1.38668805e-01 -8.94564867e-01
-1.21103659e-01 -7.38015592e-01 -5.99495053e-01 8.15871179e-01
-4.17472064e-01 3.27589750e-01 2.01777279e-01 4.84620363e-01
-4.24527586e-01 6.52339220e-01 9.58549976e-01 -4.55149919e-01
-1.79756582e-01 4.94366921e-02 -8.43382359e-01 6.79943144e-01
1.03626657e+00 -3.98974001e-01 -4.58807766e-01 -6.30879283e-01
6.10597469e-02 -2.21556947e-01 1.23005703e-01 -1.11749935e+00
3.65451902e-01 -3.75218123e-01 4.18638974e-01 -1.80238098e-01
3.79495978e-01 -8.36139381e-01 2.30877176e-01 8.01015735e-01
-4.56051856e-01 5.43224335e-01 3.15259963e-01 5.37391961e-01
-1.72803238e-01 -1.18274756e-01 1.02944911e+00 -1.59278035e-01
-5.18091500e-01 7.12311506e-01 1.81302443e-01 -9.39777121e-02
1.00733817e+00 -4.22988944e-02 -7.30857968e-01 -2.76459038e-01
-4.15928125e-01 2.82291979e-01 4.81344998e-01 6.66660309e-01
9.99267042e-01 -1.64087129e+00 -6.25922143e-01 2.91074008e-01
-5.53945377e-02 -2.76351839e-01 3.65429409e-02 2.50425994e-01
-6.62877619e-01 2.34964266e-01 -2.44084120e-01 -1.51543111e-01
-1.43651497e+00 8.65957677e-01 4.03218210e-01 -2.72370517e-01
-2.42357284e-01 1.10181618e+00 -1.34026289e-01 -7.22288787e-01
5.42439580e-01 3.15132380e-01 -8.55546296e-02 -5.22830606e-01
5.97906530e-01 3.08471680e-01 -1.87148601e-01 -3.90397757e-01
-4.06287640e-01 4.03121591e-01 -3.17236513e-01 2.85451531e-01
1.22050893e+00 4.72806185e-01 2.74294969e-02 -3.46644759e-01
1.06813693e+00 2.58628339e-01 -1.09059417e+00 -1.17062427e-01
-2.02813908e-01 -5.45889974e-01 -1.49363324e-01 -9.80555356e-01
-1.40720105e+00 6.25695169e-01 7.44235933e-01 2.95391649e-01
1.15915132e+00 -4.95438218e-01 1.28294015e+00 7.31707990e-01
2.00359687e-01 -6.93173468e-01 5.46506643e-01 7.03608334e-01
1.08480203e+00 -1.21292913e+00 -2.53513902e-01 -1.29339576e-01
-6.84001327e-01 9.60110188e-01 1.09450674e+00 -3.81835580e-01
5.00266612e-01 2.72487134e-01 3.74107391e-01 4.98498045e-02
-7.72767544e-01 2.39090875e-01 2.85660416e-01 6.28081560e-01
-1.28141344e-01 -2.97895163e-01 1.49212509e-01 7.27332175e-01
-1.66480109e-01 -4.47584927e-01 9.61712524e-02 7.04847038e-01
-4.06258941e-01 -1.20404923e+00 -4.85507220e-01 3.71969074e-01
-8.07264686e-01 -3.99903148e-01 -7.81231940e-01 5.35573363e-01
-3.26787382e-01 7.43489444e-01 -5.88540137e-01 -1.11093712e+00
4.17470098e-01 -2.71234568e-02 -1.55217499e-01 -1.82656914e-01
-1.04851067e+00 -3.17463964e-01 -1.63982794e-01 -6.89263642e-01
2.51255631e-01 1.90974146e-01 -1.06880987e+00 -4.47951913e-01
-4.59882081e-01 2.74748862e-01 5.88326156e-01 4.92162228e-01
7.35291243e-01 4.76287305e-01 9.34731483e-01 -5.51622808e-01
-1.39389575e+00 -7.25453556e-01 5.74376658e-02 4.63912159e-01
2.60931373e-01 -3.10275674e-01 -5.52139103e-01 -3.80507171e-01] | [5.669137477874756, 7.954410076141357] |
5a5f89ea-d5b5-48e0-919b-6c3681b5180b | aligning-multilingual-word-embeddings-for | 1910.03291 | null | https://arxiv.org/abs/1910.03291v1 | https://arxiv.org/pdf/1910.03291v1.pdf | Aligning Multilingual Word Embeddings for Cross-Modal Retrieval Task | In this paper, we propose a new approach to learn multimodal multilingual embeddings for matching images and their relevant captions in two languages. We combine two existing objective functions to make images and captions close in a joint embedding space while adapting the alignment of word embeddings between existing languages in our model. We show that our approach enables better generalization, achieving state-of-the-art performance in text-to-image and image-to-text retrieval task, and caption-caption similarity task. Two multimodal multilingual datasets are used for evaluation: Multi30k with German and English captions and Microsoft-COCO with English and Japanese captions. | ['Alireza Mohammadshahi', 'Karl Aberer', 'Remi Lebret'] | 2019-10-08 | aligning-multilingual-word-embeddings-for-2 | https://aclanthology.org/D19-6402 | https://aclanthology.org/D19-6402.pdf | emnlp-ws-2019-11 | ['multilingual-word-embeddings'] | ['methodology'] | [-2.56591700e-02 -2.79211819e-01 -2.97035664e-01 -4.63303328e-01
-1.66042709e+00 -7.64748871e-01 9.40836191e-01 2.40865558e-01
-8.98367405e-01 5.77230334e-01 4.51337606e-01 2.14198232e-02
3.26337308e-01 -1.03429332e-01 -7.98872530e-01 -2.73158073e-01
2.53100783e-01 6.43787444e-01 6.65260628e-02 -1.58333108e-01
1.11041613e-01 2.66721398e-01 -1.20730758e+00 6.26547396e-01
6.28751516e-01 8.00259113e-01 3.64211231e-01 9.86309230e-01
-1.40670344e-01 4.68617350e-01 -1.44894242e-01 -1.00240123e+00
7.85681903e-02 -9.77056697e-02 -8.20054471e-01 5.26896678e-02
1.22642946e+00 -6.12094589e-02 -6.67533755e-01 1.04463315e+00
8.88232768e-01 -2.51963645e-01 9.14204359e-01 -1.26568377e+00
-1.55683994e+00 8.99814367e-02 -6.76787674e-01 1.23064898e-01
7.40219355e-01 -3.50308642e-02 1.13780046e+00 -1.33732104e+00
9.92302418e-01 1.61934114e+00 3.13762844e-01 6.49355412e-01
-1.15798318e+00 -3.13776612e-01 -1.12975523e-01 2.73691505e-01
-1.58052373e+00 -4.57841963e-01 2.03879401e-01 -4.79958087e-01
9.15388644e-01 1.86305940e-01 1.26358956e-01 1.28510451e+00
2.36852869e-01 9.83577549e-01 8.48413706e-01 -6.70901179e-01
-4.33857739e-01 7.13342130e-01 -2.07167551e-01 6.75663710e-01
-4.43738662e-02 -2.13944167e-01 -5.04589140e-01 -3.30339134e-01
4.83606368e-01 -3.99423063e-01 -2.44209751e-01 -5.55672348e-01
-1.63141179e+00 1.02711880e+00 3.71555090e-01 4.25978124e-01
-1.22865550e-01 2.57152945e-01 5.26207924e-01 3.46499085e-01
3.69736075e-01 4.42515373e-01 -1.04736410e-01 1.50006548e-01
-5.26881337e-01 5.10091446e-02 4.26008016e-01 1.18404496e+00
8.05863202e-01 -3.51509452e-01 -3.11621636e-01 1.21526396e+00
5.29238164e-01 1.29207659e+00 7.83754826e-01 -7.46258020e-01
7.17886508e-01 1.99751690e-01 1.31292254e-01 -9.66837525e-01
1.93685398e-01 8.61248448e-02 -4.97235596e-01 -3.90587807e-01
1.95956882e-02 1.43848538e-01 -8.69546711e-01 1.62027252e+00
-3.96318845e-02 -9.54480246e-02 5.95933914e-01 9.30523157e-01
1.03185713e+00 8.33470702e-01 2.66945213e-01 2.48193771e-01
1.57998705e+00 -1.39778757e+00 -7.66282618e-01 -5.87699234e-01
5.07038176e-01 -1.46574998e+00 1.02625942e+00 -4.20997053e-01
-1.06289923e+00 -5.94255149e-01 -7.89746463e-01 -4.38980550e-01
-6.57808483e-01 3.08960646e-01 1.86453611e-01 3.28078330e-01
-1.21282458e+00 -2.09812447e-01 -1.59480825e-01 -8.45343769e-01
2.04245653e-02 3.68746854e-02 -9.08194721e-01 -6.25281155e-01
-1.26929986e+00 1.38696611e+00 2.85012364e-01 -2.98337698e-01
-8.51385415e-01 -3.86886269e-01 -1.22637045e+00 -3.99032921e-01
-3.40857029e-01 -7.03368008e-01 1.20573437e+00 -1.12801242e+00
-8.32371473e-01 1.61523318e+00 -3.21729392e-01 -2.29191333e-01
2.72651732e-01 -1.47891968e-01 -7.37455785e-01 5.60333788e-01
3.78635824e-01 1.59178090e+00 8.89255881e-01 -1.44578540e+00
-4.19129521e-01 -2.06281632e-01 -1.36854708e-01 5.56161106e-01
-6.57395959e-01 3.83007348e-01 -1.10993874e+00 -4.40232784e-01
-2.90278286e-01 -1.06281698e+00 -1.45939644e-02 2.15688586e-01
-2.07527727e-03 -2.57462710e-01 6.18172944e-01 -8.35792661e-01
8.92642438e-01 -2.06835628e+00 3.96949261e-01 -2.53004253e-01
-1.75522983e-01 1.64510131e-01 -1.04173505e+00 6.87632382e-01
-8.34609941e-02 8.41793790e-02 6.06088340e-02 -6.88299954e-01
3.01471263e-01 4.02092934e-01 -3.28807622e-01 4.12427336e-01
3.13476175e-01 1.36056149e+00 -9.11125362e-01 -1.20943236e+00
1.87727183e-01 6.60648882e-01 -5.90649918e-02 4.99401271e-01
1.43481255e-01 5.12396097e-02 -2.96197236e-01 7.64651775e-01
6.35635734e-01 -1.00167997e-01 -3.42252813e-02 -5.48624098e-01
2.61473596e-01 -4.41913486e-01 -6.82055116e-01 1.91536796e+00
-7.46978998e-01 1.17789388e+00 3.48362364e-02 -6.18883073e-01
6.92547917e-01 6.10859990e-01 2.21866682e-01 -8.75782788e-01
3.68766151e-02 4.27920669e-01 -5.98275959e-01 -8.63829911e-01
8.75910521e-01 5.41553162e-02 -4.20772552e-01 2.19534650e-01
3.05787057e-01 -4.02850360e-01 3.96896362e-01 4.17526394e-01
3.76243770e-01 -1.10272408e-01 5.82390577e-02 -6.70430064e-02
6.86617553e-01 -1.02072179e-01 -1.90144151e-01 6.42990589e-01
-2.38795787e-01 8.40130508e-01 1.80673659e-01 -3.08503062e-01
-1.49074507e+00 -1.40805924e+00 -2.46052787e-01 1.27500594e+00
1.26616552e-01 -2.16915563e-01 -5.43698013e-01 -6.19062960e-01
4.03878391e-02 5.00848055e-01 -6.31058872e-01 9.60444435e-02
-3.50622207e-01 -3.96023601e-01 7.34877884e-01 2.56979883e-01
8.71121958e-02 -9.71425533e-01 6.19200468e-02 -1.63347796e-01
-5.35348117e-01 -1.77095199e+00 -1.28098428e+00 -4.44762558e-01
-4.44420576e-01 -1.02267218e+00 -1.45186973e+00 -1.57071054e+00
7.24120438e-01 2.03093499e-01 1.45057523e+00 -3.84271145e-01
-4.97615308e-01 1.34162414e+00 -2.45147616e-01 -8.55381489e-02
-6.35542810e-01 -4.68895845e-02 1.00176699e-01 -3.36498278e-03
4.18474734e-01 4.17689458e-02 -3.50045592e-01 2.68624723e-01
-1.32921445e+00 -1.03070714e-01 7.25155175e-01 8.19562912e-01
5.52801788e-01 -1.16008997e+00 2.93654025e-01 1.86425429e-02
8.85268152e-01 -3.47837776e-01 -4.01154131e-01 9.42156494e-01
-3.44154924e-01 2.07990021e-01 1.17527090e-01 -7.43188858e-01
-5.35208702e-01 2.14427024e-01 8.61510932e-02 -6.80853724e-01
1.80723220e-02 4.21720177e-01 8.05161968e-02 -3.70824933e-01
4.66724485e-01 3.96836698e-01 -1.36487365e-01 -2.20187768e-01
9.52188075e-01 1.07654655e+00 8.55200708e-01 -7.30539441e-01
7.17656076e-01 2.17565909e-01 -4.08620328e-01 -7.72549748e-01
-6.10028386e-01 -7.20689416e-01 -7.35697746e-01 -2.70164728e-01
1.30367661e+00 -1.33836555e+00 -2.20241204e-01 2.10613623e-01
-1.59191692e+00 3.27847838e-01 4.14788499e-02 6.67020857e-01
-5.08731782e-01 6.32991135e-01 -5.07158279e-01 -3.23052734e-01
-5.31515837e-01 -1.45903587e+00 1.65170157e+00 1.19821690e-01
1.72130123e-01 -1.33138359e+00 4.55135912e-01 4.98155862e-01
2.85382807e-01 -5.52687235e-02 6.55466557e-01 -6.01749837e-01
-4.23484057e-01 -5.36119401e-01 -6.94849849e-01 4.71908182e-01
-1.50184510e-02 -1.90107182e-01 -8.20161760e-01 -6.09599948e-01
-6.94336772e-01 -9.24945712e-01 7.81537056e-01 6.72432259e-02
4.72031444e-01 -6.97646141e-02 -3.16510111e-01 3.22358191e-01
1.66988099e+00 -3.11378807e-01 6.56742632e-01 4.31331962e-01
6.52878582e-01 6.17546856e-01 5.02031207e-01 -6.15837760e-02
6.08366907e-01 9.20206189e-01 2.31160924e-01 -2.85654545e-01
-2.54788071e-01 -3.31035286e-01 5.05024076e-01 1.13319612e+00
6.54568911e-01 -4.35694128e-01 -9.31666374e-01 1.21097231e+00
-1.82540023e+00 -7.09452927e-01 1.30965903e-01 2.02567315e+00
9.93032217e-01 -5.18768072e-01 -3.13111663e-01 -1.01836586e+00
9.35961425e-01 1.16441920e-01 -1.37037680e-01 -5.55420160e-01
-6.30759478e-01 -9.60445479e-02 6.11616552e-01 7.89295852e-01
-1.22882128e+00 1.09642196e+00 6.91349411e+00 8.89804721e-01
-9.16217685e-01 4.26113397e-01 5.13702154e-01 2.18969256e-01
-4.92315024e-01 -2.84625649e-01 -6.55595481e-01 2.52502859e-02
8.42709661e-01 -2.37139001e-01 2.55079627e-01 5.60730517e-01
-2.50610441e-01 1.67940393e-01 -1.26926565e+00 1.51623249e+00
9.17324901e-01 -1.31383669e+00 6.03859305e-01 -2.16886431e-01
1.05710483e+00 3.39167058e-01 3.95506054e-01 1.24174803e-01
1.98110752e-02 -1.15152574e+00 5.66504896e-01 5.59040606e-01
1.05297232e+00 -4.82619226e-01 8.79566610e-01 -3.84161890e-01
-1.17616343e+00 3.78124088e-01 -5.15385509e-01 8.12785208e-01
4.04738486e-01 -4.05024365e-02 -7.11689234e-01 5.86305976e-01
6.86267316e-01 6.48846745e-01 -1.06594348e+00 9.60682750e-01
6.68614432e-02 -2.07531229e-01 3.50172408e-02 -5.96310869e-02
6.24299347e-01 -6.45764619e-02 5.10902643e-01 1.64575362e+00
4.26601231e-01 -6.88050807e-01 2.48678297e-01 4.82340395e-01
-3.12350273e-01 4.95482713e-01 -8.74044478e-01 -2.93707371e-01
2.79465109e-01 1.31403339e+00 -8.90547335e-02 -5.57330012e-01
-5.97661495e-01 1.42965770e+00 2.84471184e-01 5.95488012e-01
-8.15023363e-01 -4.65287268e-01 5.97273529e-01 -2.97346413e-01
2.46398300e-01 -1.97858155e-01 4.72334325e-01 -1.30212855e+00
7.14600980e-02 -9.58353162e-01 3.54963571e-01 -1.38328421e+00
-1.86386240e+00 1.01117182e+00 5.64098693e-02 -1.17038965e+00
-2.13998139e-01 -8.59938085e-01 -1.25805020e-01 1.04508471e+00
-1.76872432e+00 -1.73736691e+00 -1.14311464e-01 6.93922460e-01
4.62499291e-01 -4.42837834e-01 9.74600494e-01 6.83606625e-01
-1.02188572e-01 8.70721757e-01 5.50523996e-01 3.06877464e-01
1.57243168e+00 -1.01564300e+00 2.19943210e-01 5.48790991e-01
3.44247669e-01 3.64395589e-01 6.76081777e-01 -3.45861763e-01
-1.39199531e+00 -1.03167474e+00 1.40066576e+00 -6.25940979e-01
1.03876030e+00 -3.35614324e-01 -5.33081293e-01 6.65406823e-01
1.29239571e+00 -3.80164497e-02 6.12464666e-01 -3.44385296e-01
-7.25524306e-01 -1.44934123e-02 -8.48776042e-01 8.01958084e-01
3.20902526e-01 -1.10052061e+00 -7.74929225e-01 7.90266514e-01
8.24018657e-01 -2.76018649e-01 -1.05808127e+00 1.07294478e-01
6.89235628e-01 -3.19802463e-01 1.33450329e+00 -7.86272168e-01
4.78217065e-01 -1.90210968e-01 -7.75100946e-01 -1.26368892e+00
1.76591519e-02 -4.43689853e-01 4.27399784e-01 1.18793297e+00
5.39466262e-01 -4.60255742e-01 1.23401359e-01 1.88243300e-01
2.96936631e-02 -3.74289721e-01 -1.22762346e+00 -7.85328090e-01
4.31745529e-01 -1.67703867e-01 1.53607205e-01 9.28325117e-01
-1.04761183e-01 6.12154126e-01 -5.97777724e-01 2.15629205e-01
6.15273058e-01 -1.11847721e-01 7.42880762e-01 -5.56777537e-01
1.32902205e-01 -4.25644547e-01 -4.84569848e-01 -1.07140458e+00
6.66426003e-01 -1.09706724e+00 3.67827192e-02 -1.66410375e+00
7.55455256e-01 1.92977533e-01 -9.76960734e-02 3.05219442e-01
-1.37664020e-01 7.51293421e-01 3.61363411e-01 2.10515037e-01
-1.11422741e+00 6.29328370e-01 1.24916148e+00 -7.42482483e-01
4.44680125e-01 -7.91979015e-01 -2.85477817e-01 1.53371215e-01
3.40699673e-01 -4.73750263e-01 -5.96561991e-02 -1.07729268e+00
1.11700073e-01 -9.78173222e-03 4.06613946e-01 -5.93349516e-01
1.45760149e-01 9.84167680e-02 1.68930829e-01 -4.85321850e-01
6.63300216e-01 -7.37235725e-01 -2.58885175e-01 2.14731053e-01
-6.68493271e-01 8.81480575e-01 5.34154773e-01 5.48462868e-01
-6.61262989e-01 -5.26052237e-01 7.12068439e-01 1.35433882e-01
-7.87109792e-01 2.99234271e-01 -3.97742480e-01 2.93556124e-01
8.78653765e-01 1.31655604e-01 -5.77369392e-01 -8.01648974e-01
-6.27624989e-01 6.34390056e-01 6.39325023e-01 1.13026524e+00
8.34814072e-01 -2.04157257e+00 -1.21168339e+00 -2.11163729e-01
9.94158626e-01 -1.06115103e+00 1.32571086e-01 6.62304461e-01
-6.52204812e-01 8.79175425e-01 -2.08498269e-01 -7.85042405e-01
-1.57003021e+00 6.35425508e-01 3.90207767e-01 -2.13474572e-01
2.30535090e-01 7.03702927e-01 1.74543768e-01 -8.82231474e-01
2.58437276e-01 1.50141135e-01 -8.83186236e-02 1.49087131e-01
4.37074721e-01 -1.86071411e-01 -3.80931854e-01 -1.36519837e+00
-4.99004126e-01 1.08531451e+00 -9.17296186e-02 -4.85573769e-01
8.03789437e-01 -5.01360834e-01 -4.46631074e-01 4.10525590e-01
1.95961702e+00 -8.11923444e-02 -5.66317677e-01 -4.14210320e-01
-1.32826671e-01 -4.40667450e-01 -1.82427108e-01 -7.03927636e-01
-6.48242712e-01 1.11513710e+00 1.24631119e+00 -3.03930253e-01
8.34963679e-01 5.15352309e-01 8.92046154e-01 5.61357200e-01
-1.06369890e-01 -1.09059167e+00 5.62142432e-01 5.11506021e-01
1.11512649e+00 -1.74343574e+00 -1.42156854e-01 1.58110172e-01
-1.02666390e+00 1.13670421e+00 4.76433665e-01 2.44004145e-01
1.76430956e-01 -3.54895055e-01 5.51781714e-01 -3.50025343e-03
-6.71699345e-01 -3.00304472e-01 9.28639770e-01 7.78012276e-01
5.37197053e-01 -1.14440218e-01 -5.35112284e-02 -9.79730114e-02
3.52143407e-01 -4.64038938e-01 2.61654198e-01 7.56213069e-01
-2.70281553e-01 -1.30228364e+00 -5.89477122e-01 -1.27291605e-01
-3.93973351e-01 -5.65926731e-01 -4.78337914e-01 7.51950383e-01
-2.37606421e-01 8.43706131e-01 2.13066980e-01 -7.07384571e-03
2.54712015e-01 1.13573976e-01 6.86393380e-01 -3.93294036e-01
-3.00705820e-01 1.22968800e-01 3.89958099e-02 -3.31496626e-01
-5.69952071e-01 -5.13820827e-01 -6.37274683e-01 1.37678266e-01
-2.02735141e-01 2.95927942e-01 7.88782597e-01 6.49966598e-01
4.45014715e-01 5.28197363e-02 4.50525999e-01 -7.37051904e-01
-3.28183532e-01 -1.05131042e+00 -1.57485142e-01 7.51228571e-01
2.55169243e-01 -1.47820652e-01 -1.97312698e-01 2.79504299e-01] | [11.213693618774414, 1.5315526723861694] |
8ced091f-857b-4253-98ea-03ccbade08e2 | a-dataset-and-reranking-method-for-multimodal | null | null | https://aclanthology.org/W18-1814 | https://aclanthology.org/W18-1814.pdf | A Dataset and Reranking Method for Multimodal MT of User-Generated Image Captions | null | ['Julian Hitschler', 'Shigehiko Schamoni', 'Stefan Riezler'] | 2018-03-01 | null | null | null | ws-2018-3 | ['multimodal-machine-translation'] | ['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] | [-7.2270050048828125, 3.682457208633423] |
312edb1f-a1f9-489a-9a77-e00f7ad1a854 | adversarial-distortion-learning-for-medical | 2204.14100 | null | https://arxiv.org/abs/2204.14100v1 | https://arxiv.org/pdf/2204.14100v1.pdf | Adversarial Distortion Learning for Medical Image Denoising | We present a novel adversarial distortion learning (ADL) for denoising two- and three-dimensional (2D/3D) biomedical image data. The proposed ADL consists of two auto-encoders: a denoiser and a discriminator. The denoiser removes noise from input data and the discriminator compares the denoised result to its noise-free counterpart. This process is repeated until the discriminator cannot differentiate the denoised data from the reference. Both the denoiser and the discriminator are built upon a proposed auto-encoder called Efficient-Unet. Efficient-Unet has a light architecture that uses the residual blocks and a novel pyramidal approach in the backbone to efficiently extract and re-use feature maps. During training, the textural information and contrast are controlled by two novel loss functions. The architecture of Efficient-Unet allows generalizing the proposed method to any sort of biomedical data. The 2D version of our network was trained on ImageNet and tested on biomedical datasets whose distribution is completely different from ImageNet; so, there is no need for re-training. Experimental results carried out on magnetic resonance imaging (MRI), dermatoscopy, electron microscopy and X-ray datasets show that the proposed method achieved the best on each benchmark. Our implementation and pre-trained models are available at https://github.com/mogvision/ADL. | ['Jussi Tohka', 'Alejandra Sierra', 'Mohammad Khateri', 'Morteza Ghahremani'] | 2022-04-29 | null | null | null | null | ['color-image-denoising', 'medical-image-denoising', 'grayscale-image-denoising'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 4.36478466e-01 1.49553522e-01 4.32122141e-01 -2.74281502e-01
-5.80925882e-01 -2.58534223e-01 3.95597339e-01 -2.15390041e-01
-6.42769158e-01 6.08512044e-01 8.71021450e-02 4.37147617e-02
3.71938348e-02 -8.20387900e-01 -6.95657015e-01 -1.13245952e+00
2.33970750e-02 2.92647570e-01 3.64226073e-01 -7.20897242e-02
-6.54125512e-02 5.32883108e-01 -1.03878021e+00 3.84855926e-01
6.84003413e-01 1.07869911e+00 3.95735264e-01 7.85236001e-01
1.62478372e-01 7.00081587e-01 -5.04103303e-01 -3.58268142e-01
5.34249604e-01 -4.95850146e-01 -5.62942505e-01 -1.41438125e-02
1.66860893e-01 -5.05596161e-01 -4.83054638e-01 1.08379149e+00
1.02618408e+00 -1.73652396e-02 7.80598223e-01 -8.71351421e-01
-8.85488629e-01 2.16980681e-01 -5.02395749e-01 4.34011191e-01
-8.27984810e-02 1.71814188e-01 2.26071060e-01 -7.24808276e-01
7.43647397e-01 1.13758075e+00 7.24391758e-01 7.59759128e-01
-1.22675109e+00 -5.49792349e-01 -2.90185511e-01 2.19793782e-01
-1.21853054e+00 -2.48431697e-01 8.13548028e-01 -4.04988647e-01
6.16331398e-01 1.95373759e-01 5.15413105e-01 1.23601174e+00
8.98612499e-01 3.95744562e-01 1.39145720e+00 -2.38912567e-01
1.83791757e-01 -5.66354441e-03 -2.62643516e-01 5.86783350e-01
2.85913628e-02 1.17703132e-01 6.91045728e-03 1.46276392e-02
1.09213781e+00 1.75009221e-01 -4.83445674e-01 -3.49563658e-01
-8.16849172e-01 5.77706575e-01 6.66572094e-01 5.62929273e-01
-7.24847615e-01 -8.50572214e-02 4.31222349e-01 3.67014378e-01
4.03095633e-01 -5.53298835e-03 3.20644826e-02 2.23982409e-01
-9.16987002e-01 2.14361846e-02 4.33690131e-01 5.30633986e-01
5.18697977e-01 1.83500536e-02 -7.76425973e-02 8.41322839e-01
1.47603646e-01 3.20995390e-01 1.03191662e+00 -8.09636950e-01
1.24580272e-01 4.10449654e-01 -2.60822058e-01 -1.06035602e+00
-3.21175098e-01 -6.87277913e-01 -1.42662001e+00 5.95307589e-01
5.26740216e-02 -5.30790724e-02 -1.32267237e+00 1.58676827e+00
4.55262482e-01 2.68445969e-01 -1.50095046e-01 1.14613271e+00
8.98805201e-01 5.10394871e-01 -1.11982107e-01 -1.14531390e-01
1.20992506e+00 -8.41756642e-01 -9.37397420e-01 -1.34309858e-01
8.61742646e-02 -8.67916882e-01 5.87928474e-01 5.45485735e-01
-1.50146079e+00 -7.25945413e-01 -1.23619604e+00 -2.27240026e-01
-3.01388532e-01 -4.72482778e-02 1.36771500e-01 3.89721602e-01
-1.04231989e+00 7.78578997e-01 -1.07256973e+00 1.83158796e-02
5.28735101e-01 3.93497348e-01 -6.75067306e-01 -6.50780722e-02
-1.28712797e+00 9.16773558e-01 2.00262859e-01 4.39001918e-01
-1.21976590e+00 -4.31664675e-01 -7.14708269e-01 -1.16664588e-01
-2.00135529e-01 -8.91214252e-01 9.07407463e-01 -1.17660558e+00
-1.54281092e+00 1.11501920e+00 1.59205571e-01 -4.86071914e-01
8.09725940e-01 7.20459223e-02 -5.55808842e-01 3.12401056e-01
1.81505933e-01 3.31376910e-01 1.21455503e+00 -1.26988840e+00
-2.26525456e-01 -4.39390153e-01 -2.42521077e-01 6.31768256e-02
-5.20338975e-02 -3.68425101e-01 -3.82675767e-01 -1.01283562e+00
1.62512079e-01 -6.30391479e-01 -1.46446034e-01 4.32721674e-02
-3.83575678e-01 3.54418159e-01 7.31477261e-01 -9.56130445e-01
1.09643745e+00 -2.24602699e+00 3.38400900e-01 2.75685638e-01
4.27170873e-01 6.13346934e-01 -2.49052793e-01 2.95750767e-01
-5.56228697e-01 -1.33785143e-01 -4.60592180e-01 -3.47842276e-01
-2.82155305e-01 2.20020667e-01 1.63380325e-01 7.61740088e-01
1.21466778e-01 7.48025060e-01 -7.06615686e-01 -4.30213541e-01
3.50159436e-01 9.38082218e-01 -4.49716091e-01 3.42422277e-01
3.57457638e-01 7.30266213e-01 -4.03471380e-01 3.60826701e-01
1.08951604e+00 -1.05635025e-01 6.36944771e-02 -4.57246989e-01
1.37044519e-01 2.51470283e-02 -1.25951028e+00 1.77689934e+00
-4.34946060e-01 2.21263260e-01 4.11063135e-01 -1.31226242e+00
8.36533308e-01 4.60597366e-01 4.04037893e-01 -9.59863782e-01
4.09421444e-01 3.71584624e-01 1.01007827e-01 -7.35706568e-01
-1.85850129e-01 -5.20194829e-01 3.07973891e-01 1.83878288e-01
2.23072812e-01 -9.42948554e-03 -1.09311469e-01 -3.63641456e-02
1.21201849e+00 3.71495225e-02 3.13826710e-01 -1.78615212e-01
8.41056526e-01 -4.05216902e-01 5.60535729e-01 5.09981751e-01
-2.41270721e-01 7.61802495e-01 3.42076957e-01 -3.47600251e-01
-1.12650883e+00 -1.35388362e+00 -3.01267415e-01 4.31783795e-01
3.13662994e-03 9.24514681e-02 -1.05383277e+00 -6.12889409e-01
-2.01574340e-01 2.27328300e-01 -8.42802346e-01 -2.79146492e-01
-5.46078384e-01 -5.66353321e-01 4.12614226e-01 3.06999356e-01
6.80593610e-01 -1.02075267e+00 -4.35185224e-01 2.51099169e-01
4.42609750e-02 -8.22754622e-01 -5.74860275e-01 2.84424961e-01
-9.16266561e-01 -9.67917144e-01 -8.87275517e-01 -1.11880732e+00
9.37464058e-01 -1.90707922e-01 7.98949659e-01 -2.95606703e-02
-4.23065573e-01 1.87216289e-02 -1.28888637e-01 -3.23154360e-01
-5.84670901e-01 -1.82642266e-01 -1.44152701e-01 2.25324675e-01
4.17441390e-02 -1.03497958e+00 -1.12858212e+00 2.94557154e-01
-1.30202925e+00 -6.70630038e-02 8.55798960e-01 1.05946946e+00
9.00702238e-01 1.35328144e-01 5.26416481e-01 -9.13151324e-01
6.64292574e-01 -3.90539825e-01 -2.77223438e-01 -1.31675690e-01
-4.75068331e-01 2.95176879e-02 7.86476731e-01 -3.26739371e-01
-9.47938442e-01 -5.44309951e-02 -5.22930205e-01 -5.72140098e-01
-1.06638670e-01 1.43685281e-01 -1.30604014e-01 4.66491003e-03
6.51569903e-01 3.00420880e-01 1.85374618e-01 -6.76725686e-01
8.76254588e-03 7.92548656e-01 9.28909421e-01 -9.81293544e-02
6.93038642e-01 7.41973579e-01 1.35223329e-01 -5.12311697e-01
-3.06844354e-01 -1.36864588e-01 -4.68459278e-01 -4.72976007e-02
9.92493331e-01 -8.50048423e-01 -4.61859196e-01 8.31956565e-01
-1.10252631e+00 -1.49415150e-01 -3.45227420e-01 5.15126228e-01
-4.74478334e-01 3.37428242e-01 -7.11991310e-01 -3.07061434e-01
-7.26612270e-01 -1.36924922e+00 8.55334222e-01 4.02514599e-02
9.85038131e-02 -1.17295420e+00 2.63377577e-01 3.03025812e-01
5.70090353e-01 6.34939969e-01 8.89129639e-01 -5.91175616e-01
-3.40663284e-01 -3.32762599e-01 1.13281399e-01 1.04068124e+00
3.71585608e-01 -4.98748332e-01 -1.02020633e+00 -5.04558146e-01
7.64385879e-01 -2.77171254e-01 9.14224803e-01 6.00498259e-01
1.27808499e+00 -2.18724757e-01 -9.03435126e-02 9.33306873e-01
1.48745775e+00 2.61997968e-01 1.01640546e+00 3.21119905e-01
5.60288668e-01 2.21117646e-01 1.43072009e-02 6.69266656e-02
1.01308607e-01 4.99809474e-01 5.98516345e-01 -4.71067816e-01
-4.75985736e-01 -2.26742905e-02 1.89779311e-01 1.02228189e+00
-2.43279725e-01 -1.37478575e-01 -6.86501384e-01 5.56618690e-01
-1.46314859e+00 -9.44938958e-01 -3.78563330e-02 1.99011052e+00
1.10833776e+00 1.80575401e-01 -2.27396131e-01 2.73965061e-01
7.66036808e-01 -2.19681375e-02 -7.15433419e-01 -5.35377443e-01
-4.77802120e-02 8.48613977e-01 4.01918262e-01 6.15977228e-01
-1.05706882e+00 2.48293772e-01 5.55196238e+00 9.69095647e-01
-1.42784131e+00 3.22324395e-01 6.76587343e-01 -5.42645939e-02
-1.06508456e-01 -4.66639102e-01 -2.19367892e-01 6.65581882e-01
7.42912650e-01 1.08647756e-02 3.56790841e-01 5.69781125e-01
2.66444564e-01 3.39809470e-02 -9.64197278e-01 9.61200356e-01
1.33738453e-02 -1.13213885e+00 8.07360411e-02 -2.07275555e-01
5.76450050e-01 -4.45096493e-02 1.89554125e-01 -4.62410003e-02
-1.10513784e-01 -1.20286381e+00 4.43637878e-01 7.43356824e-01
9.16659713e-01 -7.33327866e-01 9.99898791e-01 3.17999184e-01
-7.50750482e-01 2.12681934e-01 -2.85054117e-01 1.50803581e-01
-1.27406150e-01 7.53392994e-01 -6.17002368e-01 7.23166585e-01
6.92237258e-01 6.95616961e-01 -2.80643076e-01 7.52509356e-01
-2.67774582e-01 2.19845414e-01 1.74167976e-02 6.55570030e-01
1.89106718e-01 -1.82773978e-01 6.78642154e-01 1.15110874e+00
1.88993737e-01 4.59733121e-02 -2.39557907e-01 7.64419317e-01
-1.68547615e-01 -1.31175369e-01 -3.30664277e-01 4.25023496e-01
1.36454448e-01 1.33782542e+00 -5.53036869e-01 -1.96964189e-01
-1.66410401e-01 1.23544490e+00 1.36399614e-02 2.83475339e-01
-8.02008331e-01 -6.03589296e-01 3.99468124e-01 3.89321536e-01
5.32978833e-01 2.57050812e-01 -1.94843158e-01 -9.89648700e-01
8.09163004e-02 -9.94910717e-01 1.98611587e-01 -8.36025357e-01
-1.57835960e+00 9.76290882e-01 -1.66569322e-01 -1.12031662e+00
-1.83346167e-01 -6.98032379e-01 -5.96738875e-01 1.25950193e+00
-1.49774623e+00 -9.30791974e-01 -4.79923874e-01 7.89794147e-01
2.53045410e-01 -1.14366114e-01 7.78022289e-01 6.79275811e-01
-3.99215370e-01 6.13350093e-01 2.91377246e-01 2.24399939e-01
7.34609723e-01 -1.21742976e+00 1.86939746e-01 8.26781750e-01
-5.08430064e-01 5.82789958e-01 6.05978489e-01 -5.60216069e-01
-9.97447371e-01 -1.13403726e+00 4.76666719e-01 -5.45364693e-02
4.09972131e-01 -3.33029956e-01 -9.31010127e-01 6.11581445e-01
5.48383355e-01 2.30213836e-01 5.12686968e-01 -8.26929629e-01
-2.52387617e-02 -2.95545727e-01 -1.65104485e+00 2.71576673e-01
7.67789364e-01 -3.28485578e-01 -7.16259360e-01 1.72494665e-01
1.77941635e-01 -6.94187462e-01 -1.19342017e+00 3.54044318e-01
4.99852240e-01 -1.22476864e+00 1.03179121e+00 -4.68306206e-02
6.47504747e-01 -3.47264349e-01 4.18549143e-02 -1.41962886e+00
-4.63154942e-01 -4.80524004e-01 7.00861886e-02 8.83503795e-01
-1.49200976e-01 -8.03444564e-01 4.35411751e-01 5.82769737e-02
-3.53172004e-01 -1.13936496e+00 -1.08018935e+00 -6.29716933e-01
2.92519748e-01 -3.44235934e-02 4.50477213e-01 7.82917261e-01
-5.80451787e-01 2.21141383e-01 -2.22773612e-01 1.29770383e-01
8.33002210e-01 -3.40112627e-01 3.39314997e-01 -8.00285101e-01
-3.91943097e-01 -1.72929645e-01 -7.09365249e-01 -8.26205671e-01
-1.68591946e-01 -1.07978010e+00 -2.30053440e-01 -1.56029046e+00
1.29890546e-01 -1.76274270e-01 -3.95911366e-01 4.44735706e-01
-6.46110103e-02 5.31636119e-01 -7.97774643e-02 2.33630776e-01
2.75427252e-02 4.22996968e-01 1.68602395e+00 -2.44424775e-01
8.99921283e-02 -1.09645121e-01 -4.81536329e-01 6.93687081e-01
7.78723419e-01 -6.12778425e-01 -4.54408139e-01 -5.29920816e-01
-2.95764655e-01 -1.21387579e-01 6.13568485e-01 -1.08886874e+00
2.06924662e-01 3.62353384e-01 7.62112856e-01 -2.66062498e-01
1.82502776e-01 -9.33688521e-01 4.86185282e-01 6.83681786e-01
-1.31850570e-01 6.64389739e-03 9.11412984e-02 3.25423211e-01
-3.62917364e-01 -2.32002199e-01 1.23877621e+00 -2.33842403e-01
-2.62366831e-01 3.19760859e-01 -1.25140890e-01 -6.36195811e-03
1.03831244e+00 -3.54475409e-01 -1.50013208e-01 -1.63259730e-01
-1.09047794e+00 -1.03060611e-01 4.35859412e-01 4.06876281e-02
9.00778353e-01 -1.39320242e+00 -7.44279444e-01 5.64445138e-01
-4.37493354e-01 1.90069780e-01 5.38450718e-01 9.68695760e-01
-7.67194152e-01 1.28116235e-01 -6.40208125e-01 -4.36393261e-01
-1.04849195e+00 5.89426816e-01 7.77097583e-01 -6.09865129e-01
-6.90320730e-01 7.60378599e-01 3.30503255e-01 -2.67612696e-01
1.96927056e-01 -3.88371527e-01 -8.94307438e-03 -2.93692410e-01
5.01059592e-01 3.39696527e-01 3.98118556e-01 -5.56092203e-01
-3.75731349e-01 5.71254551e-01 -2.21250400e-01 3.88273895e-02
1.77086329e+00 -6.31515160e-02 -3.57279360e-01 2.50642061e-01
1.49208128e+00 -2.10530132e-01 -1.14511263e+00 -7.80776665e-02
-5.59144080e-01 -2.61361778e-01 4.00977343e-01 -8.35115552e-01
-1.49392664e+00 8.50559115e-01 1.11758268e+00 1.21863939e-01
1.70958090e+00 -3.85450810e-01 9.62321162e-01 -2.67943025e-01
1.38781145e-02 -8.07774961e-01 7.10488707e-02 1.44373685e-01
1.09112120e+00 -8.83720160e-01 5.79246171e-02 -1.41330838e-01
-4.23114479e-01 9.89192009e-01 3.55505645e-01 -6.39550209e-01
9.03004289e-01 4.48167741e-01 1.29900128e-01 -2.96568036e-01
-4.50973958e-01 2.14357525e-01 2.66235769e-01 7.05004215e-01
2.62402743e-01 -3.30441803e-01 -5.49119353e-01 7.43119657e-01
-2.54179966e-02 2.87905097e-01 3.75176400e-01 9.05813873e-01
-9.62256342e-02 -1.09792566e+00 -4.08645183e-01 3.33050460e-01
-7.75728524e-01 1.79986916e-02 1.12408493e-02 6.74945772e-01
5.88092387e-01 6.31141961e-01 1.53330803e-01 -3.20075244e-01
5.54310799e-01 -1.03015028e-01 5.21722555e-01 -3.91914576e-01
-8.33689570e-01 2.37457499e-01 -3.28675628e-01 -6.38532519e-01
-4.85989779e-01 -3.21118563e-01 -1.23082387e+00 -2.88762420e-01
-8.43418855e-03 -8.86747241e-03 6.73047960e-01 5.56245208e-01
3.22758347e-01 9.49714363e-01 7.87863314e-01 -7.05726147e-01
-4.98401493e-01 -9.73882139e-01 -6.69091284e-01 7.63859451e-01
6.39273107e-01 -4.83469427e-01 -4.09989774e-01 2.93586850e-01] | [13.223081588745117, -2.5087804794311523] |
51c8786f-0d92-4985-a8f4-931b81236415 | language-learning-using-speech-to-image | 1909.03795 | null | https://arxiv.org/abs/1909.03795v1 | https://arxiv.org/pdf/1909.03795v1.pdf | Language learning using Speech to Image retrieval | Humans learn language by interaction with their environment and listening to other humans. It should also be possible for computational models to learn language directly from speech but so far most approaches require text. We improve on existing neural network approaches to create visually grounded embeddings for spoken utterances. Using a combination of a multi-layer GRU, importance sampling, cyclic learning rates, ensembling and vectorial self-attention our results show a remarkable increase in image-caption retrieval performance over previous work. Furthermore, we investigate which layers in the model learn to recognise words in the input. We find that deeper network layers are better at encoding word presence, although the final layer has slightly lower performance. This shows that our visually grounded sentence encoder learns to recognise words from the input even though it is not explicitly trained for word recognition. | ['Stefan L. Frank', 'Mirjam Ernestus', 'Danny Merkx'] | 2019-09-09 | null | null | null | null | ['grounded-language-learning'] | ['natural-language-processing'] | [ 4.72736299e-01 3.72989357e-01 3.06282938e-01 -4.14356858e-01
-7.37862110e-01 -4.94856507e-01 1.00990808e+00 7.09452108e-02
-6.68233871e-01 4.64388907e-01 7.76266873e-01 -3.15047145e-01
3.95636469e-01 -7.74196208e-01 -8.54967594e-01 -5.59734583e-01
-3.71050611e-02 4.39459294e-01 4.51149344e-02 -3.17385972e-01
3.73755366e-01 4.66166735e-01 -1.50346184e+00 7.29372859e-01
8.35523680e-02 6.00059807e-01 3.77853453e-01 1.28438675e+00
-2.92194188e-01 1.08707988e+00 -8.26380670e-01 -1.40049651e-01
-7.68696144e-02 -6.76549673e-01 -9.77962375e-01 1.02552392e-01
6.33393466e-01 -5.05486250e-01 -3.53875488e-01 6.38114631e-01
6.18134499e-01 1.87854975e-01 8.56745958e-01 -6.09435201e-01
-1.42977762e+00 4.68961537e-01 -1.38522997e-01 2.79578179e-01
5.64766586e-01 2.45271280e-01 1.11927867e+00 -1.32223415e+00
6.71679378e-01 1.72061944e+00 3.12745184e-01 8.01110446e-01
-1.20396185e+00 -1.52766690e-01 2.28053391e-01 1.79458871e-01
-1.23715639e+00 -6.85564041e-01 4.29800421e-01 -3.07046205e-01
1.56823456e+00 2.62045205e-01 6.79961145e-01 1.24225724e+00
-8.37150291e-02 9.77739215e-01 8.95865858e-01 -9.56797004e-01
1.35763496e-01 3.56589168e-01 -1.49153247e-01 7.61411667e-01
-1.24436639e-01 3.69443819e-02 -6.81592464e-01 1.27966091e-01
9.56779957e-01 -2.03206703e-01 -3.34093750e-01 -5.84363103e-01
-1.34653401e+00 8.96258712e-01 8.40455592e-01 3.90275866e-01
-4.94014353e-01 6.08759820e-01 3.30380112e-01 2.28927732e-01
3.91318947e-01 4.60902601e-01 -2.44754702e-01 -2.06856765e-02
-9.18616235e-01 -1.13417588e-01 5.30103147e-01 5.71065068e-01
4.82643962e-01 2.37259045e-01 -2.96764612e-01 7.90907145e-01
4.61413145e-01 4.89832550e-01 6.99409306e-01 -7.12099731e-01
8.34612697e-02 2.30220318e-01 -9.30480361e-02 -8.61514628e-01
-8.87662172e-02 -3.11027139e-01 -3.21745515e-01 3.95825148e-01
2.75729179e-01 4.67247553e-02 -1.21541607e+00 1.61755824e+00
-2.72725701e-01 -5.62641695e-02 4.83488590e-01 1.10749435e+00
8.85292828e-01 9.24791992e-01 1.97343379e-01 2.53513128e-01
1.35325241e+00 -1.05918574e+00 -5.89637995e-01 -5.90158522e-01
5.88560343e-01 -4.47740436e-01 1.30943096e+00 5.02549827e-01
-1.19430315e+00 -6.65283263e-01 -1.12492216e+00 -4.89291251e-01
-5.60629487e-01 1.25674158e-01 5.99959373e-01 3.84835988e-01
-1.56218016e+00 9.37366635e-02 -4.50912654e-01 -6.60804093e-01
3.90764713e-01 5.53092845e-02 -4.80637074e-01 -1.62349388e-01
-1.11849546e+00 1.26128757e+00 3.45921785e-01 -2.64273752e-02
-1.12159669e+00 -7.54919648e-02 -1.22181809e+00 6.72083721e-02
3.09321210e-02 -6.39601290e-01 1.41747904e+00 -1.43628204e+00
-1.30420995e+00 1.06329846e+00 -1.34505495e-01 -6.64529204e-01
1.21594332e-01 -2.22605675e-01 -1.00088052e-01 6.54960632e-01
-3.90141308e-01 1.43608904e+00 8.41444135e-01 -1.54556382e+00
-9.46153849e-02 -1.61988169e-01 2.05413535e-01 3.18968982e-01
-2.13777289e-01 4.43313606e-02 -2.05513760e-01 -5.74431241e-01
-1.88793749e-01 -5.09756923e-01 5.36443330e-02 4.89466816e-01
3.70171815e-02 -3.75522733e-01 6.82299972e-01 -6.95508718e-01
6.15719378e-01 -2.16443324e+00 3.06208551e-01 -1.36315092e-01
4.22938131e-02 3.91022682e-01 -5.35985947e-01 7.48629272e-01
-1.90397903e-01 1.32939145e-01 -7.86782876e-02 -4.03269529e-01
9.72548500e-02 4.97597694e-01 -3.72312158e-01 3.35307419e-01
6.66635692e-01 1.30081987e+00 -8.73207390e-01 -3.97864640e-01
4.07348901e-01 1.01456833e+00 -4.92649168e-01 1.22164220e-01
-2.83867478e-01 -8.53853598e-02 3.00273299e-02 1.97211862e-01
1.37494341e-01 -2.23030463e-01 7.96044692e-02 1.15866542e-01
1.35586560e-01 5.50307810e-01 -7.20238864e-01 1.92973983e+00
-6.98663592e-01 1.28808105e+00 8.89902338e-02 -1.24529922e+00
8.25394034e-01 5.87273180e-01 -3.95680785e-01 -1.02537203e+00
2.62247548e-02 -1.66159153e-01 3.31071727e-02 -7.70370126e-01
3.54147971e-01 -3.97891372e-01 1.26381129e-01 3.33246112e-01
3.46420944e-01 -1.33109689e-01 -1.27391905e-01 5.19927502e-01
1.00124717e+00 1.17491178e-01 2.80292898e-01 -7.43503720e-02
4.18399066e-01 -6.10872731e-02 -3.94715011e-01 7.68942833e-01
-8.61078203e-02 9.56416965e-01 3.50525916e-01 -4.76315618e-01
-1.04742873e+00 -1.25808287e+00 9.36812609e-02 1.45184040e+00
-2.98553228e-01 -1.82429343e-01 -7.24119067e-01 -4.74952519e-01
-3.57813388e-01 1.20583451e+00 -7.72785723e-01 -3.68134439e-01
-2.72872657e-01 -9.09772664e-02 5.14934123e-01 6.74733579e-01
1.76622272e-01 -1.75869167e+00 -8.10244322e-01 1.31151110e-01
1.26014739e-01 -9.48454142e-01 -1.54943272e-01 2.68428743e-01
-4.23654377e-01 -5.46046495e-01 -1.08966649e+00 -1.16956401e+00
8.90802145e-01 2.19133824e-01 1.42680693e+00 3.72910231e-01
-6.33975148e-01 8.74607623e-01 -5.21765709e-01 -5.09302378e-01
-3.68966341e-01 -3.65198076e-01 -2.63987958e-01 -2.24176198e-01
6.35175884e-01 -7.78960735e-02 -6.15924597e-01 -2.55252570e-01
-1.00242293e+00 1.71058644e-02 7.84604371e-01 8.72084260e-01
4.31910604e-02 -5.37166893e-01 3.92179489e-01 -3.67424488e-01
8.66384447e-01 -2.60146141e-01 -5.58728212e-03 1.12788469e-01
1.37750907e-02 3.34483862e-01 1.35504857e-01 -4.95349735e-01
-7.67130375e-01 2.20978484e-01 -3.42598379e-01 -1.07267708e-01
-6.16559505e-01 2.40532055e-01 1.48141369e-01 2.09025204e-01
9.01594937e-01 5.21623433e-01 9.30354297e-02 -1.76428318e-01
8.18827808e-01 8.11918080e-01 4.95846361e-01 -2.82771975e-01
4.80040997e-01 3.90790790e-01 -6.74213469e-01 -1.37112749e+00
-6.79964721e-01 -3.15748751e-01 -5.78078210e-01 -1.11704983e-01
1.29404747e+00 -9.43329036e-01 -4.93036330e-01 -7.85997137e-02
-1.50505626e+00 -4.39887583e-01 -3.51304978e-01 5.01750171e-01
-5.03796637e-01 2.51661807e-01 -4.24049854e-01 -1.21097219e+00
-1.75595120e-01 -9.07984972e-01 1.32421398e+00 -4.88418192e-02
-5.06202340e-01 -9.56916451e-01 -2.91523188e-02 2.29689211e-01
5.25350690e-01 -9.88271609e-02 7.31905758e-01 -6.90899670e-01
-5.00468671e-01 7.70873502e-02 -4.45555329e-01 4.06661898e-01
-8.75074267e-02 -1.98006615e-01 -1.21194553e+00 -3.11258823e-01
-2.80173004e-01 -8.14225495e-01 1.25993943e+00 2.36853898e-01
8.29154670e-01 -5.23136497e-01 -4.95732054e-02 1.02374457e-01
1.31639397e+00 6.93360269e-02 8.95083964e-01 3.18448186e-01
4.38962638e-01 8.48175824e-01 -6.79383725e-02 -8.72973702e-04
2.73729503e-01 3.97362471e-01 4.34252292e-01 -3.50833684e-01
-4.44815964e-01 -2.59400815e-01 6.23054087e-01 4.43335503e-01
2.89209187e-01 -6.48661137e-01 -1.07898438e+00 9.92267013e-01
-1.39870203e+00 -1.03802538e+00 2.64430165e-01 1.85634077e+00
8.68481576e-01 2.45144978e-01 -1.35413911e-02 1.78515419e-01
2.09952936e-01 1.06581725e-01 -3.19928527e-01 -1.00741148e+00
-1.50282994e-01 4.53707933e-01 4.24541607e-02 7.18072951e-01
-7.48671830e-01 1.07629681e+00 7.27764654e+00 2.51771301e-01
-1.15692914e+00 -3.60532925e-02 4.17852759e-01 -2.25120068e-01
-3.44341904e-01 -3.42882782e-01 -3.85230631e-01 -1.72186885e-02
1.11729038e+00 3.51955980e-01 3.44787657e-01 6.14962280e-01
9.55986383e-04 -1.06732830e-01 -1.16668797e+00 8.79890323e-01
6.75248623e-01 -1.12612045e+00 3.84015322e-01 -1.65291503e-01
3.39907646e-01 8.24494138e-02 8.34436268e-02 3.30619693e-01
3.43962908e-01 -1.66353297e+00 6.87634051e-01 7.83389628e-01
6.08092248e-01 -5.49973488e-01 4.39009219e-01 4.31749284e-01
-7.42807090e-01 1.70726016e-01 -3.65736663e-01 -3.99528772e-01
7.04751983e-02 2.95411539e-03 -1.14012301e+00 -8.24466869e-02
5.29613137e-01 3.56727332e-01 -6.79570913e-01 8.39334190e-01
-4.28111076e-01 6.43117964e-01 -1.19869798e-01 -5.01309812e-01
5.27727962e-01 4.36277330e-01 2.49491379e-01 1.71161497e+00
1.60674065e-01 6.36551678e-02 -1.74245700e-01 7.66152918e-01
1.12647630e-01 1.42141983e-01 -1.01184499e+00 -3.03889960e-01
-5.35211563e-02 9.50467587e-01 -5.23412049e-01 -5.22890627e-01
-5.80401123e-01 1.38690770e+00 4.04133350e-01 6.17753506e-01
-2.62806833e-01 -5.12937129e-01 5.79918206e-01 6.63839281e-02
6.86021984e-01 -7.10530043e-01 1.16170086e-01 -8.79874706e-01
-1.46306679e-01 -8.14729691e-01 -1.08611591e-01 -1.55146551e+00
-1.04163349e+00 6.29345000e-01 -1.70200661e-01 -5.11196077e-01
-6.28768802e-01 -1.03274739e+00 -4.09268945e-01 9.05837357e-01
-1.29703009e+00 -1.08005500e+00 -6.46117926e-02 2.81612009e-01
9.08002377e-01 -6.18850105e-02 1.20209324e+00 -2.43728772e-01
6.18810579e-02 3.65331829e-01 -2.00889066e-01 3.61352950e-01
5.02077639e-01 -1.42004526e+00 6.22636378e-01 6.10136032e-01
9.58688796e-01 8.09078574e-01 8.02828074e-01 -3.05856824e-01
-1.25604475e+00 -5.05885005e-01 1.03467357e+00 -5.66052496e-01
5.11580765e-01 -8.82361531e-01 -1.03484881e+00 6.17702603e-01
1.01372385e+00 -9.74959359e-02 6.13040328e-01 -1.31670639e-01
-5.79962313e-01 3.69383484e-01 -9.13468897e-01 5.66703439e-01
8.00369442e-01 -8.91552210e-01 -1.32077718e+00 4.93569940e-01
8.67424190e-01 -2.42914055e-02 -2.56745309e-01 -2.59634465e-01
4.67532486e-01 -8.25419307e-01 1.38033748e+00 -8.57386172e-01
5.93904257e-01 -1.50885075e-01 -2.38979056e-01 -1.31372011e+00
-2.12263137e-01 -6.12595975e-02 1.52235061e-01 1.09770656e+00
5.55929959e-01 -3.74366671e-01 5.97527444e-01 2.54625827e-01
1.66369796e-01 -4.66417938e-01 -7.22768128e-01 -4.68031615e-01
3.17005455e-01 -3.77755731e-01 6.07684329e-02 6.06215656e-01
2.13044696e-02 8.27931881e-01 -2.30006605e-01 -5.28401555e-03
3.63942116e-01 -3.83513987e-01 4.56264079e-01 -9.85117614e-01
-3.55630577e-01 -4.27695572e-01 -5.43517530e-01 -1.29936647e+00
9.91725251e-02 -8.71839762e-01 4.05567884e-01 -2.17707968e+00
1.26004741e-01 3.79066259e-01 -2.61932105e-01 5.79281390e-01
1.29120693e-01 5.47048509e-01 1.78472146e-01 -2.05157191e-01
-7.76400149e-01 4.88579184e-01 1.27372813e+00 -2.29172438e-01
2.86536306e-01 -6.74754977e-01 -7.51929939e-01 6.21293604e-01
6.66314304e-01 -1.14797309e-01 -4.30374056e-01 -9.06115115e-01
2.71622181e-01 -2.06871793e-01 7.54174769e-01 -8.99850547e-01
1.81852743e-01 3.25293660e-01 8.78663242e-01 -4.15167034e-01
5.98572433e-01 -6.13862038e-01 -5.67381442e-01 4.38506812e-01
-8.04727554e-01 7.54542276e-02 4.08604652e-01 4.78535652e-01
-2.81392604e-01 -4.48857635e-01 5.44314325e-01 -6.25608146e-01
-9.95911956e-01 -4.61539090e-01 -9.27425802e-01 -2.23158728e-02
5.53465009e-01 -3.10893208e-01 -7.82182515e-02 -8.53708506e-01
-9.21204865e-01 4.19894466e-03 3.07448059e-01 6.12519026e-01
1.04623926e+00 -1.25124073e+00 -8.13616693e-01 2.61600316e-01
3.13950986e-01 -1.89489454e-01 -2.20949739e-01 2.31906608e-01
-5.82983792e-01 6.58365905e-01 -1.23010337e-01 -6.20787144e-01
-1.26977408e+00 5.92436850e-01 3.59699011e-01 4.77903098e-01
-5.89258909e-01 1.34357738e+00 3.38590920e-01 -4.81216043e-01
5.38195133e-01 -3.62308562e-01 -2.64719307e-01 2.01928727e-02
7.74956286e-01 -4.05856907e-01 -1.76936507e-01 -8.91180336e-01
-4.35512960e-01 6.16793334e-01 -2.47566798e-03 -5.87883353e-01
1.32065606e+00 -9.71528515e-02 1.71930999e-01 6.57768309e-01
1.42731786e+00 -2.93020636e-01 -1.16008162e+00 -2.38859817e-01
-1.05965003e-01 -1.05712824e-01 1.92575976e-01 -9.99577165e-01
-6.83147728e-01 1.56385064e+00 6.17680728e-01 2.18303844e-01
9.17630792e-01 4.20200497e-01 3.18587482e-01 6.91432178e-01
-3.51169892e-02 -9.24779892e-01 6.24011695e-01 5.79851985e-01
1.33311081e+00 -1.28291750e+00 -1.49797603e-01 3.09581190e-01
-7.68284917e-01 1.09800673e+00 3.74723822e-01 -3.69047672e-01
2.37111568e-01 1.60386726e-01 3.35519969e-01 -2.32446313e-01
-9.24940348e-01 -4.93127137e-01 4.29272324e-01 7.87292600e-01
7.15804756e-01 -2.25368455e-01 9.04176459e-02 -2.59226952e-02
-2.53658324e-01 -4.05232757e-01 2.90993899e-01 8.11255157e-01
-7.95327485e-01 -8.84724021e-01 -3.60052437e-01 1.44016773e-01
-2.14782625e-01 -4.93650168e-01 -9.58771884e-01 7.00897574e-01
-1.45644337e-01 8.23436320e-01 4.24790800e-01 5.71950674e-02
2.70280302e-01 4.65511739e-01 6.22772217e-01 -9.59332287e-01
-4.41397160e-01 -2.16015540e-02 1.15457840e-01 -3.51168066e-01
-3.97073537e-01 -5.41912079e-01 -1.45474851e+00 2.72036284e-01
-1.17125943e-01 8.43413547e-02 9.10107255e-01 9.21655536e-01
1.46581784e-01 6.76407874e-01 -2.24037431e-02 -9.50187385e-01
-2.74972856e-01 -1.07489181e+00 -1.36355355e-01 3.15960318e-01
7.42990255e-01 -3.93784463e-01 -4.68307346e-01 1.58441126e-01] | [10.82374095916748, 1.7444604635238647] |
bf7ce675-f56d-4e39-b57d-d2a74ed21d42 | voxlingua107-a-dataset-for-spoken-language-1 | 2011.12998 | null | https://arxiv.org/abs/2011.12998v1 | https://arxiv.org/pdf/2011.12998v1.pdf | VoxLingua107: a Dataset for Spoken Language Recognition | This paper investigates the use of automatically collected web audio data for the task of spoken language recognition. We generate semi-random search phrases from language-specific Wikipedia data that are then used to retrieve videos from YouTube for 107 languages. Speech activity detection and speaker diarization are used to extract segments from the videos that contain speech. Post-filtering is used to remove segments from the database that are likely not in the given language, increasing the proportion of correctly labeled segments to 98%, based on crowd-sourced verification. The size of the resulting training set (VoxLingua107) is 6628 hours (62 hours per language on the average) and it is accompanied by an evaluation set of 1609 verified utterances. We use the data to build language recognition models for several spoken language identification tasks. Experiments show that using the automatically retrieved training data gives competitive results to using hand-labeled proprietary datasets. The dataset is publicly available. | ['Tanel Alumäe', 'Jörgen Valk'] | 2020-11-25 | null | null | null | null | ['spoken-language-identification'] | ['speech'] | [ 8.68403614e-02 5.54545373e-02 -3.23976696e-01 -5.46728373e-01
-1.62621844e+00 -7.79457510e-01 5.73874652e-01 2.04378031e-02
-7.57535279e-01 6.86852694e-01 7.05216825e-01 4.53516804e-02
4.80706513e-01 -1.35808393e-01 -7.46062815e-01 -6.02076769e-01
5.30953370e-02 5.11841714e-01 3.10978800e-01 2.53583014e-01
4.12691295e-01 1.83299914e-01 -1.88039827e+00 4.50758040e-01
6.69670045e-01 9.17575240e-01 1.56629607e-01 1.06541455e+00
-2.56505489e-01 8.69819343e-01 -9.03815866e-01 4.84012738e-02
-4.39493880e-02 -3.98761809e-01 -8.49521279e-01 4.80795532e-01
6.76833391e-01 -2.91969538e-01 -2.44905293e-01 9.44192648e-01
6.91637099e-01 8.61601457e-02 4.49674904e-01 -1.17877376e+00
-5.52486144e-02 7.45123267e-01 1.72976963e-02 5.07732630e-01
1.01824188e+00 -7.22819269e-02 9.32496786e-01 -1.25027025e+00
9.78082001e-01 1.20467901e+00 2.26423994e-01 5.54937184e-01
-8.26725364e-01 -7.02010036e-01 -2.85096496e-01 9.83066633e-02
-1.91280949e+00 -1.30072165e+00 4.90378439e-01 -4.58284199e-01
1.17423308e+00 3.13396960e-01 4.18270409e-01 1.17801988e+00
-5.43731034e-01 9.92380619e-01 6.83278382e-01 -7.75777638e-01
2.85330325e-01 5.40094793e-01 8.14820677e-02 5.62423348e-01
-3.26806337e-01 -3.14563930e-01 -1.06428814e+00 -3.15863103e-01
1.16208889e-01 -7.92074800e-01 -7.05220997e-01 -2.65874770e-02
-1.47210503e+00 7.52908528e-01 -3.20511341e-01 4.06394988e-01
-3.16034466e-01 -5.13347387e-01 6.56076312e-01 4.32395786e-01
6.43677056e-01 1.58465415e-01 -2.72054076e-01 -5.33820450e-01
-1.22992265e+00 2.12530002e-01 1.24367607e+00 1.16396296e+00
6.40353620e-01 1.79532290e-01 -2.92497724e-02 1.15676165e+00
2.88501203e-01 8.71341228e-01 7.22693980e-01 -1.04577398e+00
7.05292702e-01 4.93460834e-01 1.20227069e-01 -6.56911492e-01
-2.39604786e-02 2.77621746e-01 -2.14562248e-02 -4.36702669e-01
6.02237105e-01 -2.84305423e-01 -7.21138418e-01 1.36712968e+00
2.75733858e-01 -6.75046742e-02 1.80218652e-01 8.09872985e-01
1.04199719e+00 1.10098839e+00 -8.74008164e-02 -4.63499486e-01
1.20709777e+00 -8.80319834e-01 -7.21449912e-01 3.59103456e-02
5.97007692e-01 -7.81818628e-01 1.05743444e+00 5.77571154e-01
-7.95210421e-01 -3.41514945e-01 -7.75196075e-01 1.91699937e-01
-2.22305700e-01 3.50801259e-01 -2.73042414e-02 7.35458314e-01
-1.10393524e+00 -3.62209678e-01 -3.63907129e-01 -5.63156426e-01
1.29455119e-01 1.66489646e-01 -5.10029137e-01 -1.06082693e-01
-1.16338825e+00 3.23925376e-01 2.89140403e-01 -2.53339022e-01
-1.35455322e+00 -1.84063315e-01 -1.05039561e+00 -1.59135580e-01
5.35974085e-01 3.04813862e-01 1.38038206e+00 -1.27663088e+00
-1.35543263e+00 1.12058949e+00 -5.61617494e-01 -6.02519631e-01
4.88968253e-01 5.54411039e-02 -7.09038854e-01 7.34521449e-01
2.99879909e-01 8.52892816e-01 1.07951057e+00 -8.44757378e-01
-9.38244760e-01 -5.35878018e-02 -4.33370888e-01 2.46448517e-01
-6.06371760e-01 5.64200997e-01 -9.16126966e-01 -5.30759037e-01
-1.89431489e-01 -9.80692387e-01 5.18692493e-01 -4.07221884e-01
-5.06650984e-01 -5.31973541e-01 7.84104586e-01 -1.14041352e+00
1.30060649e+00 -2.32112002e+00 -5.07775769e-02 2.22188637e-01
-9.47542414e-02 1.87112629e-01 -4.32037830e-01 3.92616004e-01
2.37787634e-01 1.48625001e-01 9.22144428e-02 -4.25722182e-01
-1.28621325e-01 -1.16095841e-01 -3.75230491e-01 4.26210761e-01
-1.12161748e-01 3.39125067e-01 -8.54931235e-01 -7.89787829e-01
-1.36335762e-02 4.71127957e-01 -3.39736164e-01 3.42598677e-01
-1.68469220e-01 2.55442202e-01 -2.26433039e-01 8.51577520e-01
1.63577706e-01 5.19756079e-01 -2.49616936e-01 1.93365768e-01
-2.28087172e-01 7.31528163e-01 -1.13651586e+00 1.60925519e+00
-3.41519743e-01 1.17233491e+00 2.67716944e-01 -6.69503808e-01
9.21875060e-01 8.66569459e-01 3.96733761e-01 -3.91213626e-01
1.26528284e-02 3.26112807e-01 -2.87742347e-01 -7.87463248e-01
3.67960483e-01 5.44373930e-01 -2.93929964e-01 6.06105208e-01
4.76713806e-01 -2.37129495e-01 6.73130929e-01 3.69740129e-01
9.37447548e-01 -3.45179737e-01 2.94958502e-02 -1.11025140e-01
9.24444318e-01 2.41970524e-01 3.91461313e-01 6.38276756e-01
-5.23015380e-01 4.95723248e-01 3.55521560e-01 -1.35732517e-01
-1.05033171e+00 -7.32095361e-01 7.33084306e-02 1.22603679e+00
-5.73761165e-01 -5.03135026e-01 -1.17900753e+00 -6.81966782e-01
-3.53066504e-01 6.99223578e-01 -1.37811244e-01 7.60539696e-02
-2.42052630e-01 9.33616310e-02 8.68241489e-01 4.78902794e-02
4.18524325e-01 -1.41436028e+00 -2.64735013e-01 1.16937041e-01
-5.62307477e-01 -1.28522992e+00 -9.21504378e-01 -1.78460807e-01
-1.49692893e-01 -9.64213610e-01 -8.34693372e-01 -1.21005797e+00
5.96271932e-01 2.12353677e-01 9.37405884e-01 -1.84623361e-01
-2.71380216e-01 8.26239228e-01 -6.05223715e-01 -4.25333828e-01
-8.77079010e-01 1.13711648e-01 5.20097256e-01 2.31128901e-01
8.91503870e-01 1.83646768e-01 4.15415578e-02 3.49990845e-01
-5.32602012e-01 -4.42378849e-01 3.72169875e-02 6.32516384e-01
4.46337819e-01 1.49556594e-02 5.43375194e-01 -2.47620821e-01
6.60543680e-01 -2.44507059e-01 -6.45651519e-01 1.72651485e-01
-1.03343047e-01 -2.35298544e-01 2.34863207e-01 -7.26245522e-01
-8.70290577e-01 5.62653124e-01 4.22426574e-02 -4.48177844e-01
-4.64101315e-01 4.74946052e-01 -1.22150416e-02 1.03601798e-01
5.07557571e-01 6.95027649e-01 9.29019973e-02 -3.39624792e-01
-4.21064608e-02 1.45230925e+00 5.84591806e-01 -8.92603621e-02
2.20280379e-01 4.86837327e-02 -1.11539793e+00 -1.71534920e+00
-4.12788421e-01 -8.55928719e-01 -4.61235076e-01 -4.25661445e-01
8.37432921e-01 -1.31849051e+00 -4.33042020e-01 5.39790869e-01
-9.50269759e-01 -2.64763921e-01 -1.47235468e-02 7.08785117e-01
-4.92672056e-01 2.17463136e-01 -4.07626212e-01 -1.29607713e+00
-4.06826437e-01 -1.10369658e+00 1.25548995e+00 9.41957813e-03
-4.80170041e-01 -4.92565423e-01 1.21656619e-01 7.05720961e-01
-2.02977490e-02 -5.31177580e-01 1.29707605e-01 -1.27941561e+00
-3.84289533e-01 -5.58974683e-01 3.37096840e-01 5.06539226e-01
3.88822742e-02 2.74579883e-01 -1.21126568e+00 -2.51427799e-01
-2.24319130e-01 -7.57236719e-01 6.92402184e-01 2.89690882e-01
7.01926112e-01 -4.91386056e-01 -1.41177997e-01 -1.35538384e-01
6.72991455e-01 3.45627964e-01 3.84003103e-01 8.63216743e-02
4.13815767e-01 8.73528481e-01 7.65564680e-01 4.55961198e-01
3.85177463e-01 6.41673028e-01 -2.23861486e-01 4.07060504e-01
4.83373068e-02 -3.98807347e-01 8.89848471e-01 1.02360380e+00
2.92196304e-01 -4.59998399e-01 -1.17716718e+00 1.00565076e+00
-1.25957525e+00 -1.18645799e+00 2.50454009e-01 2.33166695e+00
9.50646281e-01 8.23231693e-03 6.62038028e-01 3.97690505e-01
8.59030604e-01 -6.13692142e-02 -1.18946597e-01 -1.51930466e-01
-6.69110492e-02 -2.94654429e-01 3.12496483e-01 7.71327615e-01
-1.28687072e+00 9.18673217e-01 6.48774767e+00 7.90083945e-01
-1.15295398e+00 5.97185344e-02 5.05427122e-01 -3.14132601e-01
1.52677864e-01 -4.44048673e-01 -1.15473139e+00 6.08902931e-01
1.42434609e+00 -2.77831227e-01 4.33775127e-01 9.02448237e-01
5.82047999e-01 -4.18859184e-01 -9.21420574e-01 1.27343118e+00
6.28929913e-01 -9.91300404e-01 -5.58122620e-02 -1.07637495e-01
4.96742725e-01 3.88045281e-01 -3.21430147e-01 3.53928506e-01
-3.95448431e-02 -6.67173088e-01 1.00931180e+00 2.13509068e-01
7.51136899e-01 -7.96764374e-01 5.96259534e-01 6.40436232e-01
-1.06705177e+00 7.05991909e-02 -7.27454945e-02 3.49423110e-01
9.87299811e-03 2.31224343e-01 -1.51393795e+00 -2.36316323e-01
9.01916504e-01 5.23497999e-01 -5.75607598e-01 1.04074955e+00
-1.44491225e-01 9.21908557e-01 -7.13221490e-01 -5.25480092e-01
2.56168228e-02 1.47543475e-01 8.44353735e-01 1.40682852e+00
3.11607838e-01 -3.25443119e-01 1.73785701e-01 3.53927970e-01
-2.41604641e-01 4.55798447e-01 -8.51025462e-01 -4.47531432e-01
9.47453260e-01 9.71780300e-01 -5.31960189e-01 -6.83682561e-01
-4.29064542e-01 8.40530634e-01 -1.36859313e-01 4.27663624e-01
-4.74846661e-01 -6.06370151e-01 3.55847806e-01 -6.92834035e-02
3.64127964e-01 -1.34846717e-01 5.72776973e-01 -1.23113012e+00
1.91595942e-01 -1.30001295e+00 3.15238953e-01 -6.53822839e-01
-7.68942952e-01 8.00775945e-01 -7.53544793e-02 -1.24484849e+00
-7.71622121e-01 -2.01616600e-01 -2.05149993e-01 6.73990190e-01
-1.05356061e+00 -6.50299072e-01 -2.74383456e-01 5.34779906e-01
1.30408990e+00 -7.36815095e-01 8.75220537e-01 4.42383736e-01
-4.13351059e-01 5.11273563e-01 -1.06627360e-01 5.08633196e-01
9.59071219e-01 -8.74418855e-01 1.49887487e-01 7.39616215e-01
5.63836157e-01 3.14639002e-01 6.31477296e-01 -5.45450091e-01
-1.29598188e+00 -9.14970100e-01 1.63592589e+00 -5.77191055e-01
6.75135553e-01 -7.22209334e-01 -7.98920393e-01 4.10212308e-01
2.51317561e-01 -8.82533044e-02 7.59297371e-01 -5.48032403e-01
-3.03951293e-01 3.05816997e-03 -1.02289796e+00 1.37833282e-01
6.47760212e-01 -1.09962666e+00 -6.30838454e-01 5.04696310e-01
4.50566888e-01 -1.13740496e-01 -6.25204682e-01 -3.19923431e-01
4.40474629e-01 -5.69765508e-01 6.64276361e-01 -2.44348288e-01
-4.37907763e-02 -2.09008917e-01 -2.01405928e-01 -1.10098505e+00
5.99524975e-01 -7.19436944e-01 1.72689825e-01 1.67402303e+00
7.69238532e-01 -2.58384943e-01 5.45877576e-01 5.21839499e-01
2.41232514e-01 5.32206660e-03 -1.27559209e+00 -7.16003478e-01
-4.76791531e-01 -4.51538086e-01 1.84873134e-01 7.74685383e-01
2.88854748e-01 5.09530187e-01 -2.71657467e-01 4.68902066e-02
4.28759307e-01 -2.45930105e-01 8.83315325e-01 -9.29073095e-01
1.93950474e-01 -5.10687642e-02 -3.22598457e-01 -9.68685329e-01
5.28504491e-01 -6.54319406e-01 4.86345351e-01 -1.04300654e+00
-4.12110537e-02 -2.57774480e-02 2.38826811e-01 5.50354421e-01
3.05170536e-01 2.72991598e-01 -9.53303576e-02 2.58871049e-01
-8.21679175e-01 2.98908204e-01 4.45872873e-01 -5.27102470e-01
-4.76594239e-01 2.34703515e-02 -7.55668655e-02 5.39122760e-01
6.13290012e-01 -4.17349875e-01 -2.27968648e-01 -1.98952079e-01
-3.03932041e-01 3.60086598e-02 -3.33492830e-02 -9.71613705e-01
1.92326427e-01 3.00107718e-01 -1.00573219e-01 -6.80016279e-01
3.94668102e-01 -7.02195883e-01 -3.07650238e-01 2.95813918e-01
-6.55944526e-01 -1.86544180e-01 5.09851687e-02 4.21633720e-01
-6.73869967e-01 -3.40500861e-01 4.08471406e-01 4.73434068e-02
-8.44173074e-01 -1.00549094e-01 -9.82623935e-01 1.27893671e-01
9.77663696e-01 -3.79275829e-02 7.31592327e-02 -1.06210780e+00
-7.02149272e-01 1.84666872e-01 2.72698104e-01 7.19313204e-01
5.71069479e-01 -1.06419480e+00 -8.30983341e-01 2.55763143e-01
3.48124504e-01 -3.59688491e-01 -1.72740474e-01 6.34844780e-01
-4.44862634e-01 8.30928445e-01 3.04374695e-01 -7.85411596e-01
-1.81116712e+00 3.37358832e-01 3.16763967e-02 5.61387479e-01
-3.08339566e-01 9.07241762e-01 -4.98420388e-01 -3.13961416e-01
9.36591685e-01 -3.40230554e-01 -3.73146832e-01 4.10840809e-01
9.99360025e-01 1.80668741e-01 5.25586419e-02 -1.10859871e+00
-6.36871576e-01 1.10833809e-01 9.50507075e-02 -6.50541663e-01
1.13434851e+00 -3.23231041e-01 1.28183171e-01 7.54774153e-01
1.35518169e+00 3.83326888e-01 -1.00519836e+00 -2.23483011e-01
1.65503845e-01 -2.49178797e-01 1.28886491e-01 -5.01821280e-01
-7.50577629e-01 6.44523919e-01 5.70444703e-01 3.44274491e-01
7.91226983e-01 2.73726642e-01 4.90753889e-01 6.48894787e-01
4.92201388e-01 -1.40429401e+00 -5.13487905e-02 5.67654133e-01
1.02674520e+00 -1.49680531e+00 -3.59866589e-01 -3.82126987e-01
-7.99336016e-01 9.42728519e-01 3.30143243e-01 1.15437701e-01
4.21029598e-01 7.46437386e-02 3.92282039e-01 1.46099925e-01
-7.07015693e-01 -6.70467988e-02 2.21186414e-01 5.50120533e-01
3.63053530e-01 -1.25889227e-01 -1.76220700e-01 2.84183234e-01
-2.58174419e-01 -1.31980136e-01 5.13330877e-01 9.10954535e-01
-6.26022637e-01 -8.17687988e-01 -5.57510674e-01 2.40385503e-01
-3.89619708e-01 -8.78312737e-02 -8.43997061e-01 4.48657811e-01
-3.37945014e-01 1.33525777e+00 2.30811208e-01 -1.48576945e-01
1.30164444e-01 5.46368718e-01 7.45118558e-02 -7.23309219e-01
-3.47198725e-01 5.23472071e-01 7.09282756e-01 -3.42282951e-01
-5.35357177e-01 -9.64581311e-01 -9.89599168e-01 1.74885586e-01
-3.13209444e-01 5.07365227e-01 8.45130384e-01 7.46433616e-01
3.26656401e-01 -5.21040373e-02 7.83094704e-01 -1.05586362e+00
-3.76333654e-01 -1.13373578e+00 -2.99228370e-01 2.75046051e-01
5.42374372e-01 -4.42554176e-01 -8.06817114e-01 6.02973580e-01] | [14.228422164916992, 6.250761985778809] |
0a561cb5-d065-4e15-9b75-08685286e464 | r-2vos-robust-referring-video-object | 2207.01203 | null | https://arxiv.org/abs/2207.01203v1 | https://arxiv.org/pdf/2207.01203v1.pdf | R^2VOS: Robust Referring Video Object Segmentation via Relational Multimodal Cycle Consistency | Referring video object segmentation (R-VOS) aims to segment the object masks in a video given a referring linguistic expression to the object. It is a recently introduced task attracting growing research attention. However, all existing works make a strong assumption: The object depicted by the expression must exist in the video, namely, the expression and video must have an object-level semantic consensus. This is often violated in real-world applications where an expression can be queried to false videos, and existing methods always fail in such false queries due to abusing the assumption. In this work, we emphasize that studying semantic consensus is necessary to improve the robustness of R-VOS. Accordingly, we pose an extended task from R-VOS without the semantic consensus assumption, named Robust R-VOS ($\mathrm{R}^2$-VOS). The $\mathrm{R}^2$-VOS task is essentially related to the joint modeling of the primary R-VOS task and its dual problem (text reconstruction). We embrace the observation that the embedding spaces have relational consistency through the cycle of text-video-text transformation, which connects the primary and dual problems. We leverage the cycle consistency to discriminate the semantic consensus, thus advancing the primary task. Parallel optimization of the primary and dual problems are enabled by introducing an early grounding medium. A new evaluation dataset, $\mathrm{R}^2$-Youtube-VOS, is collected to measure the robustness of R-VOS models against unpaired videos and expressions. Extensive experiments demonstrate that our method not only identifies negative pairs of unrelated expressions and videos, but also improves the segmentation accuracy for positive pairs with a superior disambiguating ability. Our model achieves the state-of-the-art performance on Ref-DAVIS17, Ref-Youtube-VOS, and the novel $\mathrm{R}^2$-Youtube-VOS dataset. | ['Bhiksha Raj', 'Yan Lu', 'Xiao Li', 'Xiaohao Xu', 'Jinglu Wang', 'Xiang Li'] | 2022-07-04 | null | null | null | null | ['referring-expression-segmentation', 'referring-video-object-segmentation'] | ['computer-vision', 'computer-vision'] | [ 1.46493569e-01 -7.37763718e-02 -2.59380728e-01 -4.17578936e-01
-7.84204483e-01 -5.20712614e-01 1.68090940e-01 -4.29865211e-01
-2.17829257e-01 3.49934727e-01 -1.57124430e-01 -1.25030279e-01
-6.40265271e-02 -4.53224272e-01 -8.79606128e-01 -5.39235950e-01
2.05536380e-01 5.39638065e-02 4.16010588e-01 -1.73434228e-01
9.90451779e-03 1.42979726e-01 -1.50292659e+00 2.98706979e-01
6.84242785e-01 1.52543807e+00 1.46851808e-01 2.63957411e-01
-2.11224675e-01 9.44866359e-01 -6.32746756e-01 -7.99531341e-01
3.64428580e-01 -6.91332102e-01 -8.41560245e-01 2.75153488e-01
5.47589242e-01 -1.24453478e-01 -4.16151822e-01 1.50358748e+00
1.56772211e-01 1.50608867e-01 4.80605960e-01 -1.88537264e+00
-7.65244186e-01 2.97197193e-01 -7.43094027e-01 2.35290393e-01
6.02707028e-01 1.96278915e-01 1.32669306e+00 -9.19597805e-01
7.82549083e-01 1.14524698e+00 5.42442679e-01 4.93139178e-01
-9.08962667e-01 -9.51613128e-01 5.01380682e-01 3.17878187e-01
-1.80812287e+00 -2.20418364e-01 7.87294745e-01 -5.66367865e-01
5.96077025e-01 4.63470697e-01 6.69627368e-01 1.15659130e+00
-2.12838411e-01 1.21831095e+00 9.23949122e-01 9.81753990e-02
8.32657665e-02 4.98853177e-02 9.11680758e-02 7.78662324e-01
-5.84317110e-02 -3.52022469e-01 -6.32295012e-01 2.92745829e-01
7.65474677e-01 -1.35876104e-01 -6.85189009e-01 -5.49871065e-02
-1.13522542e+00 6.16150916e-01 2.31454849e-01 3.64055991e-01
-9.47092324e-02 6.31835088e-02 5.46086669e-01 2.55127460e-01
2.56166726e-01 2.25300163e-01 -3.02878439e-01 -2.32088044e-01
-1.02789688e+00 9.66775566e-02 6.14249110e-01 1.57972193e+00
5.94788134e-01 8.57871175e-02 -1.39554828e-01 7.32991874e-01
3.74098271e-01 3.90787393e-01 3.37745696e-01 -1.20792663e+00
4.92143959e-01 6.46655321e-01 -1.07725680e-01 -1.38269007e+00
-7.63992313e-03 -1.23280913e-01 -7.78953493e-01 -2.40800589e-01
2.81467080e-01 5.96208833e-02 -6.13078475e-01 1.86465847e+00
3.86199385e-01 4.74733949e-01 7.68398819e-03 1.17760587e+00
1.21505034e+00 5.47265589e-01 1.37073286e-02 -4.20220286e-01
1.38893819e+00 -9.35643852e-01 -9.38041985e-01 -1.46799060e-02
5.35939157e-01 -6.28006577e-01 1.17018366e+00 3.08099300e-01
-9.39391911e-01 -6.15492642e-01 -9.63564932e-01 -1.31293520e-01
-1.45465121e-01 -1.08549502e-02 3.30347061e-01 3.74817163e-01
-8.16570878e-01 4.59905684e-01 -4.57200617e-01 -2.19990835e-01
5.97904563e-01 4.64426428e-01 -5.31686008e-01 -3.18513326e-02
-1.32342303e+00 4.19080853e-01 2.64055043e-01 3.47593009e-01
-5.75419724e-01 -4.67815369e-01 -9.05341625e-01 -2.49962285e-01
9.56260562e-01 -4.60093051e-01 8.54558527e-01 -1.38748872e+00
-1.25110459e+00 1.30555904e+00 -3.20433646e-01 -1.24439627e-01
8.67070735e-01 -2.17643112e-01 -7.33756900e-01 3.95547837e-01
4.93567705e-01 5.73748350e-01 1.16148269e+00 -1.31336200e+00
-7.11740553e-01 -4.09307808e-01 2.49860659e-01 6.07988378e-03
-2.13726163e-01 1.94556519e-01 -1.18286848e+00 -9.74052429e-01
4.70175087e-01 -9.16142881e-01 3.40930253e-01 3.43404785e-02
-6.30067170e-01 -5.07427096e-01 1.07165813e+00 -6.63668573e-01
1.43432999e+00 -2.39372659e+00 3.92207623e-01 1.53154165e-01
2.72094607e-01 2.48496994e-01 -1.13261618e-01 -8.63397270e-02
-2.64109105e-01 5.07368803e-01 -2.20119074e-01 -2.51741916e-01
5.01547493e-02 4.22803909e-01 -4.05623555e-01 5.78285873e-01
4.00084436e-01 9.87075627e-01 -8.62795234e-01 -9.72892225e-01
-4.54580933e-02 1.48831636e-01 -4.10596162e-01 3.48318696e-01
-2.91388124e-01 4.35678869e-01 -5.80256343e-01 8.92454684e-01
4.99852121e-01 -4.05061573e-01 8.49685669e-02 -7.07448900e-01
1.91385910e-01 -3.14274073e-01 -1.32197583e+00 1.51317585e+00
1.05130367e-01 6.30102754e-01 4.03093025e-02 -1.19432557e+00
8.26016307e-01 1.85790449e-01 8.27163339e-01 -6.30680084e-01
2.51951098e-01 2.37577066e-01 -3.48501116e-01 -1.07110584e+00
4.99831170e-01 -1.65313289e-01 -1.67434022e-01 3.43510024e-02
7.14494437e-02 -3.01715255e-01 2.19516248e-01 3.43294710e-01
8.66564810e-01 4.98455524e-01 -2.97811106e-02 -8.64309743e-02
7.42744148e-01 -5.95024228e-02 8.98616016e-01 5.30012310e-01
-4.77993637e-01 6.43656194e-01 8.58445764e-01 -7.16578811e-02
-6.19476616e-01 -9.89808321e-01 -9.03087407e-02 9.02400374e-01
8.41760516e-01 -6.45458639e-01 -7.74893224e-01 -1.09612107e+00
-2.25686818e-01 6.53557301e-01 -6.51031733e-01 -3.07208039e-02
-4.95376617e-01 -3.79140645e-01 7.74017811e-01 6.02254272e-01
7.19466567e-01 -8.69701982e-01 -3.08940589e-01 -1.22349545e-01
-5.94204962e-01 -1.69372010e+00 -6.39624476e-01 -2.11391196e-01
-4.78632152e-01 -1.23963165e+00 -5.86616814e-01 -8.16139877e-01
7.24191129e-01 1.84498325e-01 9.31689799e-01 2.99014717e-01
1.57037675e-02 7.17346966e-01 -5.37521541e-01 -3.81653989e-03
-2.62865216e-01 -3.32466871e-01 -3.77218947e-02 4.28490072e-01
4.22422111e-01 -2.44230434e-01 -3.37922513e-01 9.00034368e-01
-1.22228301e+00 1.63154930e-01 1.77023962e-01 7.97990084e-01
9.98034537e-01 6.69320896e-02 4.22350109e-01 -5.07704854e-01
1.54133424e-01 -4.70857620e-01 -4.92342412e-01 3.94629657e-01
-1.73969999e-01 -9.34451967e-02 5.86341918e-01 -5.70948064e-01
-6.58742011e-01 3.44514623e-02 -1.39566034e-01 -1.14379478e+00
4.16177996e-02 4.94342953e-01 -5.83949924e-01 1.63587496e-01
9.76091996e-02 2.17199013e-01 -2.61088371e-01 -9.62867439e-02
2.74937242e-01 6.61361635e-01 8.01222742e-01 -6.58576012e-01
7.25918114e-01 6.79548860e-01 -4.61086705e-02 -9.29117918e-01
-8.62938046e-01 -6.56343937e-01 -5.17551005e-01 -4.18819517e-01
1.15191650e+00 -9.00771618e-01 -8.74080837e-01 3.57609510e-01
-1.24680018e+00 -9.24554542e-02 -2.20968574e-01 2.67733663e-01
-7.24842072e-01 6.99061811e-01 -2.79317915e-01 -7.22629786e-01
9.03861597e-02 -1.36427927e+00 1.20258653e+00 5.42401895e-02
-3.59929115e-01 -6.79049551e-01 -6.26802802e-01 7.00788796e-01
-2.27160454e-01 2.77099967e-01 7.06024468e-01 -6.06447816e-01
-6.08578086e-01 -3.32921073e-02 -5.56407452e-01 5.88828623e-01
8.93802047e-02 2.68585354e-01 -7.26875067e-01 2.35761553e-02
2.12783322e-01 -2.95053303e-01 7.59235024e-01 5.72526120e-02
1.36703098e+00 -2.66108721e-01 -4.10976037e-02 7.26643324e-01
1.16820395e+00 2.84619987e-01 7.53019333e-01 1.54658645e-01
9.03616786e-01 6.81798816e-01 1.00308430e+00 2.05202594e-01
4.26521897e-01 8.45759988e-01 5.29657543e-01 1.20932236e-01
3.27143818e-03 -1.71604529e-01 5.39888203e-01 8.46163392e-01
-1.53233454e-01 -3.30470055e-01 -7.25377858e-01 5.17360270e-01
-2.05008674e+00 -1.00615954e+00 -3.24060440e-01 1.88154078e+00
7.52080202e-01 -3.49425748e-02 -2.70433631e-02 1.36847243e-01
9.66143727e-01 3.49481970e-01 -6.63500071e-01 1.25630125e-01
-4.11401004e-01 -6.36453107e-02 1.50398254e-01 1.18795134e-01
-1.14806807e+00 1.09684360e+00 4.58321238e+00 1.20927489e+00
-1.24008942e+00 2.33189672e-01 6.56834602e-01 -4.14550491e-02
-1.62310421e-01 -7.38966241e-02 -7.37436950e-01 6.78144872e-01
3.18108708e-01 -7.07802027e-02 2.75775313e-01 7.53929853e-01
1.70303911e-01 -7.62297735e-02 -1.27636397e+00 1.41392004e+00
3.62597913e-01 -1.09102225e+00 4.06035781e-03 -3.97791296e-01
6.43837214e-01 -2.05967098e-01 5.87036088e-02 2.26241797e-01
-1.53765827e-01 -9.36244845e-01 1.22797036e+00 2.15226412e-01
1.15820181e+00 -3.08012933e-01 6.92254603e-01 -3.04134488e-02
-1.50822198e+00 7.09708259e-02 1.43149033e-01 4.30833995e-01
3.07628363e-01 2.38555908e-01 -1.22371167e-01 8.95915270e-01
1.12740767e+00 1.12907052e+00 -4.37020540e-01 5.19424736e-01
-3.83662999e-01 4.37326431e-01 -2.69476861e-01 1.00350797e-01
1.81111097e-01 -3.26728493e-01 8.01025152e-01 1.17133546e+00
1.77200809e-01 3.40003490e-01 1.59028605e-01 1.16790903e+00
-3.31175655e-01 2.03438506e-01 -4.03100818e-01 -1.67700231e-01
2.28788003e-01 9.75225806e-01 -8.37678730e-01 -3.96629155e-01
-4.85317469e-01 1.26845086e+00 -4.98525845e-03 5.26795864e-01
-1.44575131e+00 -1.71577588e-01 6.27883911e-01 6.54948354e-02
6.00683749e-01 -3.77607474e-04 -2.59790346e-02 -1.43078220e+00
3.80837083e-01 -9.51000214e-01 4.51064795e-01 -1.00932634e+00
-1.32440352e+00 5.75371563e-01 5.73150069e-02 -1.52276742e+00
4.31612022e-02 -5.40206194e-01 -1.48098052e-01 4.67223883e-01
-1.38764203e+00 -1.12248230e+00 -5.84278584e-01 9.87407863e-01
5.86551130e-01 6.94975406e-02 2.96326399e-01 4.33286279e-01
-8.50263417e-01 6.90171242e-01 -3.08466643e-01 5.01206458e-01
5.97140789e-01 -7.49180079e-01 -1.31639898e-01 8.71587276e-01
3.55175138e-01 5.16393781e-01 6.66002989e-01 -7.40678012e-01
-1.46485639e+00 -1.37608635e+00 4.57674861e-01 -2.88158566e-01
8.35996091e-01 -1.92827731e-01 -1.00925648e+00 8.60945165e-01
-1.71370819e-01 4.01799798e-01 5.20399868e-01 -4.64488864e-01
-4.76617336e-01 -1.72395967e-02 -1.06142271e+00 7.94709980e-01
1.40172195e+00 -7.88590074e-01 -5.81920981e-01 3.05550903e-01
8.40910435e-01 -5.56569695e-01 -9.97114301e-01 5.18267035e-01
4.71896797e-01 -1.01799047e+00 8.02245796e-01 -7.34284818e-01
6.41603112e-01 -5.53283334e-01 -5.92250705e-01 -4.99926716e-01
2.46492758e-01 -8.83438706e-01 -7.58068562e-02 1.47670889e+00
5.62206283e-02 -3.64430696e-01 7.02382505e-01 7.36602545e-01
-1.24069102e-01 -8.68612409e-01 -1.25075340e+00 -1.08240950e+00
-2.63539582e-01 -8.74904871e-01 3.03593159e-01 1.27081537e+00
-1.50692075e-01 1.14273183e-01 -3.45557839e-01 1.80967912e-01
2.77756393e-01 1.04820080e-01 7.82597363e-01 -7.39528656e-01
-2.89115816e-01 -4.37722236e-01 -7.71519244e-01 -1.42356658e+00
4.49511379e-01 -7.60827601e-01 -4.98754438e-03 -1.16629589e+00
-5.82462139e-02 -3.63224477e-01 -1.53247520e-01 4.95295346e-01
-2.34337628e-01 4.00556833e-01 4.15309280e-01 4.17792559e-01
-1.12462878e+00 6.70009375e-01 1.17825961e+00 -2.81115532e-01
-1.10237405e-01 -2.84957647e-01 -5.70176780e-01 9.61368680e-01
2.56489128e-01 -3.90337259e-01 -2.63333201e-01 -2.95760810e-01
3.77121121e-01 6.51456416e-02 6.33440673e-01 -6.08491302e-01
1.49249285e-01 -1.92414090e-01 -1.29586324e-01 -4.94722009e-01
3.48726392e-01 -1.02326512e+00 3.51111561e-01 5.05843014e-02
-1.24591790e-01 7.52397552e-02 3.73081793e-03 5.99097908e-01
-4.95073110e-01 -2.13217184e-01 6.25738740e-01 -3.36210020e-02
-1.09939718e+00 3.41819853e-01 9.29798186e-02 4.68472451e-01
1.41149139e+00 -5.79183936e-01 -1.45236790e-01 -4.97177452e-01
-6.76334143e-01 4.25745249e-01 3.96292150e-01 6.38784647e-01
7.81255960e-01 -1.28312564e+00 -3.93715799e-01 3.03370003e-02
2.80232489e-01 2.88202405e-01 2.13421226e-01 1.36470008e+00
-2.58100867e-01 -6.47026822e-02 9.80858728e-02 -8.67157817e-01
-1.45446706e+00 6.43229723e-01 2.96894640e-01 1.60774723e-01
-4.27099228e-01 9.70424294e-01 3.36163223e-01 1.17393494e-01
2.40178704e-01 -3.54250550e-01 -1.16491914e-01 1.78111851e-01
2.15621039e-01 3.27918530e-01 -2.62296021e-01 -1.31562340e+00
-3.54442328e-01 8.30489874e-01 2.14633629e-01 8.67040530e-02
9.45727587e-01 -3.60117614e-01 -2.81070888e-01 4.82042164e-01
1.51944089e+00 -5.38777597e-02 -1.07631719e+00 -2.10629240e-01
-1.70840129e-01 -6.37256444e-01 -2.19480932e-01 -2.95974404e-01
-1.43873334e+00 6.79693758e-01 2.65976518e-01 1.67775497e-01
1.20328915e+00 2.76568502e-01 8.52342725e-01 2.25472208e-02
3.08703423e-01 -1.16834831e+00 3.81824911e-01 2.14085624e-01
8.97839069e-01 -1.40948081e+00 -1.52839236e-02 -8.10707748e-01
-8.64372730e-01 1.00270331e+00 7.15886056e-01 1.75631046e-01
4.20289755e-01 1.94283053e-02 -7.39996582e-02 -4.01175648e-01
-1.94085628e-01 -2.01275870e-01 4.02205884e-01 2.78688520e-01
9.46801379e-02 -2.24066183e-01 -2.24537462e-01 1.01459110e+00
-7.39439949e-03 4.08001132e-02 1.47654161e-01 7.34462798e-01
2.13133693e-01 -7.07152188e-01 -2.34397218e-01 3.23757499e-01
-5.69853127e-01 -1.96043635e-03 -4.71389890e-01 9.40083921e-01
3.83062661e-01 9.66801584e-01 5.95133156e-02 -6.80199921e-01
3.88757080e-01 -1.25095934e-01 2.87692457e-01 -2.06367925e-01
-5.22116423e-01 1.72792092e-01 -7.39629343e-02 -8.46979618e-01
-8.46608579e-01 -7.15881646e-01 -1.59548914e+00 -1.44644603e-01
-5.01759648e-01 3.47609594e-02 2.52736807e-01 1.03916633e+00
2.68210590e-01 4.21197295e-01 5.34513950e-01 -4.71255422e-01
-2.25282013e-01 -2.56261617e-01 -5.67166209e-01 9.94579077e-01
8.22398663e-02 -7.33413637e-01 -4.95137900e-01 4.33700711e-01] | [9.828984260559082, 0.6351532936096191] |
5b35bb31-4973-44dc-9853-f99e7ea0cfd1 | a-comprehensive-benchmark-for-covid-19 | 2209.07805 | null | https://arxiv.org/abs/2209.07805v3 | https://arxiv.org/pdf/2209.07805v3.pdf | A Comprehensive Benchmark for COVID-19 Predictive Modeling Using Electronic Health Records in Intensive Care | The COVID-19 pandemic has posed a heavy burden to the healthcare system worldwide and caused huge social disruption and economic loss. Many deep learning models have been proposed to conduct clinical predictive tasks such as mortality prediction for COVID-19 patients in intensive care units using Electronic Health Record (EHR) data. Despite their initial success in certain clinical applications, there is currently a lack of benchmarking results to achieve a fair comparison so that we can select the optimal model for clinical use. Furthermore, there is a discrepancy between the formulation of traditional prediction tasks and real-world clinical practice in intensive care. To fill these gaps, we propose two clinical prediction tasks, Outcome-specific length-of-stay prediction and Early mortality prediction for COVID-19 patients in intensive care units. The two tasks are adapted from the naive length-of-stay and mortality prediction tasks to accommodate the clinical practice for COVID-19 patients. We propose fair, detailed, open-source data-preprocessing pipelines and evaluate 17 state-of-the-art predictive models on two tasks, including 5 machine learning models, 6 basic deep learning models and 6 deep learning predictive models specifically designed for EHR data. We provide benchmarking results using data from two real-world COVID-19 EHR datasets. One dataset is publicly available without needing any inquiry and another dataset can be accessed on request. We provide fair, reproducible benchmarking results for two tasks. We deploy all experiment results and models on an online platform. We also allow clinicians and researchers to upload their data to the platform and get quick prediction results using our trained models. We hope our efforts can further facilitate deep learning and machine learning research for COVID-19 predictive modeling. | ['Ewen M. Harrison', 'Liantao Ma', 'Wen Tang', 'Yasha Wang', 'Wenqing Wang', 'Yinghao Zhu', 'Junyi Gao'] | 2022-09-16 | null | null | null | null | ['mortality-prediction', 'length-of-stay-prediction'] | ['medical', 'medical'] | [-2.97008958e-02 -3.72287720e-01 -1.65631741e-01 -3.75660986e-01
-6.83944702e-01 -9.98881608e-02 -1.34029344e-01 7.11843312e-01
-6.04698360e-01 7.77991056e-01 2.70550609e-01 -7.40893364e-01
-3.88828158e-01 -6.12734437e-01 -3.02187681e-01 -4.80717808e-01
-4.98867184e-01 1.13892615e+00 -2.94958383e-01 1.57987267e-01
-2.11725160e-01 3.84950548e-01 -7.58227050e-01 7.38991261e-01
6.28583610e-01 7.78547525e-01 6.21258877e-02 8.11994672e-01
4.24658984e-01 8.90183926e-01 -3.95473719e-01 -2.31790915e-01
2.56639928e-01 -3.43002170e-01 -7.92645037e-01 -6.42566025e-01
-3.08027089e-01 -5.05509377e-01 -2.28675902e-01 2.70435035e-01
1.04452252e+00 -3.69651586e-01 7.76202798e-01 -1.09541595e+00
-3.09556961e-01 3.05704832e-01 -9.65311602e-02 4.02964324e-01
9.95181650e-02 4.05246943e-01 5.60181797e-01 -6.16909385e-01
4.53009039e-01 7.47762561e-01 1.32822692e+00 5.55729091e-01
-8.87997985e-01 -7.17986822e-01 -3.91581476e-01 1.64786652e-01
-1.23018253e+00 -1.15608079e-02 9.51742288e-03 -9.05571759e-01
1.27837884e+00 1.18814126e-01 8.20327342e-01 1.41331780e+00
7.05789924e-01 4.82991308e-01 5.89419007e-01 1.43586159e-01
4.28892672e-02 -1.94085732e-01 1.40968561e-01 5.67438900e-01
3.92471403e-01 1.31622404e-01 -8.65237713e-02 -7.58995533e-01
5.43130517e-01 1.04569387e+00 -3.66389155e-01 -4.33386303e-02
-1.45130050e+00 8.30873251e-01 1.90886855e-01 -2.95681553e-03
-7.90854275e-01 -1.80301905e-01 8.61057758e-01 2.81279445e-01
3.92660677e-01 4.28697646e-01 -1.34000790e+00 -4.05332774e-01
-6.39057815e-01 3.69165272e-01 8.49117815e-01 7.81474173e-01
1.33717582e-01 -3.47750098e-01 -3.97917330e-01 8.09278369e-01
2.35288113e-01 4.95482475e-01 5.24888337e-01 -6.59437835e-01
2.83671588e-01 4.64987069e-01 2.04820156e-01 -6.97180152e-01
-9.28915203e-01 -4.53167737e-01 -1.23854792e+00 -4.51071680e-01
1.33779734e-01 -6.30820215e-01 -7.36535609e-01 1.35781598e+00
-4.06651795e-02 5.68210185e-01 2.33319685e-01 6.03084028e-01
1.03560543e+00 4.45581168e-01 3.58754814e-01 -4.51355100e-01
1.63814759e+00 -7.28169382e-01 -6.24601245e-01 2.07229614e-01
1.34899366e+00 -5.13040721e-01 6.31984591e-01 3.52504849e-01
-8.23804259e-01 -2.54583489e-02 -5.29638648e-01 2.84071445e-01
-2.85371631e-01 -5.59310205e-02 3.25509548e-01 1.97219089e-01
-7.88158953e-01 5.18944442e-01 -1.21209455e+00 -6.78495169e-01
5.97521663e-01 3.20317686e-01 -3.10489446e-01 -7.65309557e-02
-1.30328357e+00 9.76253927e-01 2.24431574e-01 -1.24461725e-01
-1.08345020e+00 -1.20869827e+00 -6.04749620e-01 -3.29509042e-02
-1.85189903e-01 -1.50477719e+00 1.31414461e+00 -1.96821585e-01
-7.22443402e-01 1.06813157e+00 -9.11660269e-02 -5.75321972e-01
5.62368453e-01 -4.21639949e-01 -4.39783454e-01 -1.07142195e-01
-9.33338776e-02 6.43320382e-02 1.58921644e-01 -8.46188545e-01
-6.25409663e-01 -4.47506547e-01 -5.69747508e-01 -1.27681673e-01
-1.76949352e-01 4.22152191e-01 -9.65150297e-02 -6.03492200e-01
-6.02278948e-01 -9.50277388e-01 -6.06454611e-01 -3.96898955e-01
-2.08761409e-01 -1.54959038e-01 6.53633773e-01 -8.22154462e-01
1.49982917e+00 -1.87501514e+00 -4.13632840e-01 -2.79047787e-01
5.78136265e-01 7.55117297e-01 -9.01073739e-02 6.36656225e-01
-1.33340240e-01 2.11186007e-01 -2.44323730e-01 -3.91541332e-01
-5.53091109e-01 2.00973824e-01 -1.80652380e-01 4.22730029e-01
2.10178480e-01 1.07407165e+00 -1.05213618e+00 -4.17692900e-01
1.80784419e-01 6.52194977e-01 -7.89151192e-01 8.90308738e-01
6.57210872e-02 8.53949666e-01 -5.90774119e-01 5.14570832e-01
5.57566226e-01 -7.65005112e-01 2.05832183e-01 1.38276726e-01
3.25308859e-01 2.52949774e-01 -2.88733929e-01 1.27995181e+00
-3.24180454e-01 1.11971445e-01 -2.07181647e-01 -1.17678750e+00
7.58690715e-01 7.00427294e-01 1.05285943e+00 -1.79095730e-01
2.03644067e-01 2.90949419e-02 3.89501676e-02 -9.43430066e-01
-9.60035473e-02 -3.24673355e-01 1.14331692e-02 4.41325396e-01
-3.48955512e-01 2.86916643e-01 -2.69348592e-01 -1.01787232e-01
1.54540515e+00 -3.24390888e-01 5.79842925e-01 -1.55044615e-01
5.28010279e-02 1.99082181e-01 9.75685537e-01 7.73538053e-01
-4.80352312e-01 7.10060179e-01 3.39483798e-01 -1.18053007e+00
-8.57349694e-01 -1.07249081e+00 -5.45714617e-01 8.79363298e-01
-4.52518046e-01 -5.67646265e-01 -3.52156222e-01 -5.98591983e-01
4.68657874e-02 4.65568453e-01 -6.81518614e-01 -7.00489730e-02
-6.78343654e-01 -1.24418223e+00 8.65469038e-01 7.19878554e-01
3.97982867e-03 -1.25450420e+00 -1.00062442e+00 5.23583829e-01
-3.43200386e-01 -8.85443091e-01 -1.74258977e-01 2.19212919e-01
-9.82438982e-01 -1.42720902e+00 -5.43690860e-01 -7.32918024e-01
2.25679904e-01 -1.55880868e-01 1.33303475e+00 3.47985059e-01
-5.64534545e-01 1.96816474e-01 -2.31630564e-01 -1.05078101e+00
-4.72114861e-01 9.58987251e-02 3.67755413e-01 -5.52910626e-01
8.31458986e-01 -2.16511086e-01 -1.02597702e+00 2.28532270e-01
-6.73790216e-01 1.16231911e-01 3.02190691e-01 1.04149890e+00
6.93034589e-01 -4.82483804e-01 9.21069562e-01 -1.00724769e+00
6.17484391e-01 -1.04160261e+00 -1.46346897e-01 6.76707774e-02
-9.85508263e-01 -3.87474358e-01 7.87777841e-01 2.06054188e-02
-5.64576268e-01 -1.24413431e-01 -4.54089075e-01 -5.29069304e-01
-4.08785313e-01 7.91875839e-01 5.40305972e-01 8.58215153e-01
5.46141624e-01 2.03409493e-02 1.91174179e-01 -6.76333964e-01
-3.24639916e-01 1.13342547e+00 1.89460397e-01 -2.47369781e-01
1.05966538e-01 1.90151811e-01 -7.97949955e-02 -3.25784683e-01
-9.69883084e-01 -7.23671615e-01 -4.08932060e-01 3.61251920e-01
1.34245074e+00 -1.36910641e+00 -1.00854719e+00 5.46064436e-01
-1.12805474e+00 -4.38450962e-01 6.01549000e-02 6.53512359e-01
-5.44449031e-01 2.10754070e-02 -1.11867380e+00 -4.40861881e-01
-9.71938312e-01 -1.19531572e+00 1.00240743e+00 -3.30110639e-01
-4.76612031e-01 -1.37883425e+00 6.49931192e-01 2.88353354e-01
6.48860157e-01 6.00748181e-01 1.19152856e+00 -1.25961924e+00
-8.53346810e-02 -2.84027338e-01 -1.95031777e-01 2.34615371e-01
2.77892202e-01 -1.57154053e-02 -7.88822114e-01 -6.53434515e-01
-5.09131029e-02 -4.30373937e-01 8.16772461e-01 6.46769106e-01
1.64650416e+00 -2.85821527e-01 -8.00955057e-01 1.08857155e+00
1.21684754e+00 5.16456246e-01 4.55619544e-01 2.84908175e-01
7.17611134e-01 1.91651911e-01 4.91624206e-01 9.39086199e-01
7.45448887e-01 3.59691232e-01 2.69609094e-01 -4.31985766e-01
5.92052281e-01 3.81942838e-02 -1.97781235e-01 1.00853848e+00
-4.72707391e-01 -2.82827228e-01 -1.60890973e+00 5.20957351e-01
-2.02645111e+00 -8.81334901e-01 -4.09158707e-01 2.08044791e+00
8.53620589e-01 -4.44501549e-01 9.48642716e-02 -3.53990346e-01
3.54357660e-01 -4.11840081e-01 -5.59228361e-01 -5.43617368e-01
3.42480481e-01 1.44480988e-01 1.05324484e-01 4.28541414e-02
-1.23681307e+00 4.56460238e-01 7.21924114e+00 2.40828305e-01
-1.19576526e+00 2.60200113e-01 1.13178682e+00 -2.69961298e-01
2.25752383e-01 -3.99886459e-01 -5.57426929e-01 5.63355982e-01
1.29394805e+00 -1.18313618e-01 -3.55740152e-02 8.50093186e-01
7.09622741e-01 6.45692050e-01 -1.39625883e+00 1.08103526e+00
-3.90807569e-01 -1.56735992e+00 -1.61585048e-01 4.65828087e-03
5.55598497e-01 6.52462661e-01 -1.91469252e-01 4.68267798e-01
4.04666394e-01 -1.44323277e+00 -1.89113677e-01 7.05912948e-01
1.20839465e+00 -5.14525652e-01 1.30568409e+00 3.64944577e-01
-7.77822733e-01 -1.72335938e-01 -2.52610594e-01 -1.17194854e-01
3.25931519e-01 5.94802201e-01 -1.26521957e+00 5.16506433e-01
1.01104546e+00 1.00468481e+00 -1.38273180e-01 1.08238494e+00
3.80708665e-01 9.24058497e-01 1.08942375e-01 3.57200950e-01
-1.62543342e-01 1.07063115e-01 1.51415899e-01 1.60748911e+00
5.14848948e-01 3.66354227e-01 2.56548911e-01 5.54528356e-01
-5.68423793e-02 2.29065388e-01 -7.76831388e-01 -2.38188189e-02
2.89878786e-01 1.00811219e+00 -6.50677830e-02 -4.76091713e-01
-4.80251104e-01 5.61019957e-01 1.81775272e-01 2.47148320e-01
-9.21452940e-01 -1.22951016e-01 1.16005492e+00 4.77250576e-01
-2.00931542e-02 2.05354705e-01 -5.61870217e-01 -1.29222357e+00
-5.50522923e-01 -1.01533496e+00 9.77836728e-01 -4.60541785e-01
-1.80294645e+00 6.82948053e-01 -3.11783571e-02 -1.28367138e+00
-5.06526947e-01 -5.85731089e-01 -5.90758502e-01 8.75954628e-01
-1.41509175e+00 -8.18411291e-01 -3.91199738e-01 6.19522870e-01
1.81693584e-01 -4.36692357e-01 1.51660168e+00 4.01571214e-01
-6.63280904e-01 5.23889840e-01 3.53882730e-01 3.42975825e-01
1.01375365e+00 -8.18205178e-01 3.30168605e-01 2.62266099e-01
-8.60459089e-01 9.04035807e-01 3.31710488e-01 -6.87243700e-01
-1.07695448e+00 -1.63586819e+00 1.13638115e+00 -9.73349392e-01
5.08606076e-01 -7.40842940e-03 -1.06094623e+00 1.09588432e+00
1.90250084e-01 2.82037724e-02 1.35635853e+00 5.34360781e-02
1.21833831e-02 -1.51435321e-03 -1.12200725e+00 1.70990214e-01
9.72091436e-01 -1.61508307e-01 -6.53861344e-01 5.61250210e-01
7.95093536e-01 -2.41671339e-01 -1.45672870e+00 1.02316833e+00
5.54879963e-01 -7.56474257e-01 9.77774799e-01 -1.28147900e+00
7.97442973e-01 -3.21168751e-02 2.19654769e-01 -1.32445312e+00
-5.30504107e-01 -3.41285586e-01 -2.01220199e-01 3.76942337e-01
4.40544248e-01 -9.72977757e-01 4.56376374e-01 5.94530880e-01
-2.38407552e-01 -1.46715128e+00 -6.47245467e-01 -3.43562871e-01
3.47881645e-01 -3.79403710e-01 7.32650399e-01 1.38361144e+00
1.36440828e-01 2.82110810e-01 -5.94546556e-01 -6.47358876e-03
2.37510577e-01 1.42962232e-01 7.87757218e-01 -1.31395626e+00
-4.29256320e-01 -1.85255602e-01 -2.85747707e-01 -5.15013039e-01
-2.77933925e-01 -8.41205716e-01 -2.70518929e-01 -1.95027173e+00
7.46372581e-01 -7.57680237e-01 -9.29132044e-01 8.93608212e-01
-3.92786294e-01 2.76340246e-02 -6.68234229e-02 5.19180477e-01
-5.38128793e-01 2.98816890e-01 1.04862273e+00 3.81535441e-02
-3.27753514e-01 1.58839330e-01 -5.86876154e-01 5.41517317e-01
9.09603715e-01 -6.97926283e-01 -2.57323027e-01 -4.78243381e-01
9.94097516e-02 4.69930828e-01 1.58034593e-01 -6.94499850e-01
-2.17998132e-01 -3.97067696e-01 5.77152073e-01 -5.26680648e-01
-7.18267933e-02 -6.74896240e-01 3.93904090e-01 1.04714930e+00
-1.83966875e-01 4.92143780e-01 2.42896900e-01 3.41812730e-01
-5.11505790e-02 3.64469230e-01 7.56341219e-01 -6.92965463e-02
-1.84471890e-01 8.61496568e-01 -7.24823475e-01 2.53723532e-01
1.18361926e+00 1.70418322e-01 -5.31399071e-01 -1.87511504e-01
-8.15428257e-01 4.34591025e-01 3.59269738e-01 3.70473921e-01
7.01853633e-01 -9.98735905e-01 -1.27147400e+00 3.13670576e-01
5.11277080e-01 1.71128497e-01 3.05606663e-01 1.38836503e+00
-9.45713878e-01 6.20283186e-01 -1.27660602e-01 -6.59293294e-01
-1.24149489e+00 9.66141164e-01 5.08404851e-01 -5.97154856e-01
-8.79975140e-01 4.22394544e-01 3.88527542e-01 -5.63011467e-01
1.79497629e-01 -4.58735764e-01 -1.30802080e-01 -3.16864133e-01
7.53898561e-01 2.73349106e-01 1.73418030e-01 -3.42277080e-01
-7.78161466e-01 1.26371577e-01 -9.39921588e-02 8.20673347e-01
1.67580450e+00 2.70853996e-01 -2.59017497e-01 3.11252832e-01
1.32355499e+00 -5.13026476e-01 -7.13769376e-01 -4.87130471e-02
-2.48027295e-01 -2.41347998e-01 -3.70309204e-01 -8.59830260e-01
-7.99049258e-01 9.37638283e-01 8.67413461e-01 -2.81559508e-02
1.22547138e+00 -4.56775688e-02 1.10558879e+00 4.53061551e-01
2.23807082e-01 -5.59768260e-01 -1.75767839e-01 7.79542506e-01
5.74886143e-01 -1.39642823e+00 -2.29213476e-01 2.36508138e-02
-7.89089918e-01 7.98421323e-01 4.80968982e-01 1.40939131e-01
1.16540039e+00 4.72105861e-01 5.76523066e-01 -3.68275464e-01
-1.32476556e+00 3.85248274e-01 -1.20394185e-01 6.53206646e-01
1.02796316e+00 5.79741895e-01 -3.52518171e-01 1.08292496e+00
2.75208820e-02 4.75540072e-01 2.94180959e-01 8.23444903e-01
-1.69182345e-02 -1.10173833e+00 -5.82344197e-02 1.05142725e+00
-9.55677032e-01 -4.10084456e-01 6.94393646e-03 3.25863570e-01
2.13789821e-01 7.82141089e-01 -3.35619459e-03 -5.09096444e-01
2.43541136e-01 2.38121569e-01 -1.13701582e-01 -6.59042299e-01
-7.22394109e-01 -2.44575322e-01 5.63401803e-02 -4.73855764e-01
-8.06043074e-02 -5.43745518e-01 -1.38596642e+00 -4.54837501e-01
2.84409642e-01 -1.27412817e-02 2.92528600e-01 7.80982316e-01
1.05623257e+00 6.60530686e-01 2.22247899e-01 -3.67502570e-01
-6.26326919e-01 -1.02897131e+00 -3.38335872e-01 5.50519884e-01
5.10551572e-01 -3.25024515e-01 -1.08987287e-01 2.52174497e-01] | [7.965155601501465, 6.194892406463623] |
b71abe62-d732-4976-aeb2-4688d1633b0c | emotion-dynamics-modeling-via-bert | 2104.07252 | null | https://arxiv.org/abs/2104.07252v2 | https://arxiv.org/pdf/2104.07252v2.pdf | Emotion Dynamics Modeling via BERT | Emotion dynamics modeling is a significant task in emotion recognition in conversation. It aims to predict conversational emotions when building empathetic dialogue systems. Existing studies mainly develop models based on Recurrent Neural Networks (RNNs). They cannot benefit from the power of the recently-developed pre-training strategies for better token representation learning in conversations. More seriously, it is hard to distinguish the dependency of interlocutors and the emotional influence among interlocutors by simply assembling the features on top of RNNs. In this paper, we develop a series of BERT-based models to specifically capture the inter-interlocutor and intra-interlocutor dependencies of the conversational emotion dynamics. Concretely, we first substitute BERT for RNNs to enrich the token representations. Then, a Flat-structured BERT (F-BERT) is applied to link up utterances in a conversation directly, and a Hierarchically-structured BERT (H-BERT) is employed to distinguish the interlocutors when linking up utterances. More importantly, a Spatial-Temporal-structured BERT, namely ST-BERT, is proposed to further determine the emotional influence among interlocutors. Finally, we conduct extensive experiments on two popular emotion recognition in conversation benchmark datasets and demonstrate that our proposed models can attain around 5\% and 10\% improvement over the state-of-the-art baselines, respectively. | ['Jianping Shen', 'Haiqin Yang'] | 2021-04-15 | null | null | null | null | ['emotion-recognition-in-conversation'] | ['natural-language-processing'] | [-3.48062277e-01 -9.62826163e-02 -5.38477581e-03 -6.87524021e-01
-5.09929061e-01 -4.30610776e-01 5.26695490e-01 -1.39868215e-01
-2.20683232e-01 6.01948321e-01 7.13165522e-01 -1.96790155e-02
2.20122516e-01 -3.26374978e-01 -1.17793716e-01 -8.99054825e-01
-3.48348394e-02 2.54018873e-01 -4.72182959e-01 -8.24837863e-01
2.20227595e-02 2.90609330e-01 -1.28689349e+00 6.08347476e-01
8.48146319e-01 1.06174970e+00 -7.48272091e-02 5.14579415e-01
-5.69076955e-01 1.43031073e+00 -7.79692948e-01 -3.44911814e-01
-3.69374692e-01 -7.52921939e-01 -9.54654336e-01 -5.25050424e-02
-7.85526633e-01 -1.07083984e-01 -2.50059307e-01 6.63434684e-01
7.07488894e-01 4.22093391e-01 5.90683222e-01 -1.35022211e+00
-1.64738372e-01 1.01719069e+00 -5.33544362e-01 1.21897556e-01
4.40602243e-01 -2.73484290e-01 1.14953220e+00 -7.56012619e-01
3.34995478e-01 1.55377841e+00 5.26135266e-01 6.91942155e-01
-6.69343650e-01 -7.75167465e-01 4.94120955e-01 4.84642595e-01
-7.98259974e-01 -6.19436920e-01 1.32562947e+00 -2.46400312e-01
9.97381330e-01 3.43149751e-01 5.25945663e-01 1.45706630e+00
-7.10946098e-02 1.25607014e+00 9.36927915e-01 -2.17719182e-01
-1.44472033e-01 1.89365044e-01 4.34063047e-01 1.63725927e-01
-9.20931756e-01 -1.42157063e-01 -6.47333980e-01 -2.53191948e-01
3.58254969e-01 -1.15564242e-01 -3.66194338e-01 2.02399656e-01
-8.11308503e-01 1.01224613e+00 4.65648085e-01 5.08357942e-01
-5.69730639e-01 -1.69479012e-01 1.12487555e+00 3.73584718e-01
6.20727181e-01 2.16892511e-01 -4.70616758e-01 -9.07748103e-01
-3.78741473e-01 -1.75305083e-01 9.95197892e-01 7.75286257e-01
5.84092379e-01 -3.84638608e-02 -3.46635222e-01 1.52593744e+00
1.16038941e-01 -2.45497331e-01 6.71976507e-01 -6.86225355e-01
4.41997737e-01 6.10365868e-01 5.59418909e-02 -9.09453452e-01
-5.25940895e-01 -2.42765725e-01 -1.01250851e+00 -4.57087249e-01
7.27323582e-03 -6.38813913e-01 -2.11896479e-01 1.79339826e+00
4.08791929e-01 2.41806582e-01 4.09909338e-01 8.52993488e-01
1.11081243e+00 8.87533963e-01 -9.97636188e-03 -5.92516661e-01
1.33314824e+00 -1.42655325e+00 -1.02077603e+00 6.89346790e-02
8.80494356e-01 -7.26429999e-01 8.44420910e-01 4.26307023e-01
-9.57688272e-01 -4.99276042e-01 -6.51598513e-01 -1.10202199e-02
-3.09261739e-01 1.83269624e-02 7.33531237e-01 3.63988072e-01
-6.23558044e-01 2.12240309e-01 -4.61027712e-01 6.13285154e-02
-3.32798958e-02 2.17299059e-01 -2.00378433e-01 1.88854188e-01
-1.74623752e+00 8.84805858e-01 6.46624938e-02 6.96934879e-01
-5.09675026e-01 -3.43088627e-01 -9.17284429e-01 6.23864308e-02
9.29078609e-02 -6.83296993e-02 1.42291081e+00 -1.24474871e+00
-2.08720779e+00 5.78980803e-01 -4.77430165e-01 -1.27819732e-01
3.42311680e-01 -2.46129662e-01 -5.98865390e-01 2.87657138e-02
-2.88868278e-01 4.87046450e-01 4.54514205e-01 -1.32542801e+00
-6.56418502e-01 -4.72124703e-02 1.61624193e-01 5.51785350e-01
-4.52878386e-01 4.59886461e-01 -4.06740546e-01 -5.60971737e-01
-2.01898515e-01 -9.00459409e-01 3.58444825e-02 -7.77583420e-01
-4.88597929e-01 -7.30741203e-01 9.48154151e-01 -5.14676809e-01
1.29392612e+00 -2.23954415e+00 2.44003907e-01 -2.53883470e-03
6.39213398e-02 2.71712065e-01 -2.65988171e-01 6.70836091e-01
-1.87508792e-01 -1.56254873e-01 1.79523647e-01 -6.62983775e-01
2.57411808e-01 4.19435322e-01 -4.73410994e-01 2.09248364e-01
2.54813492e-01 8.08262229e-01 -8.85340035e-01 -3.49452645e-01
1.35240078e-01 6.50348663e-01 -3.80629539e-01 5.46937883e-01
7.47273117e-02 5.91762245e-01 -3.93019319e-01 4.25129354e-01
4.47705358e-01 1.04808308e-01 2.76496381e-01 -1.33373812e-01
-8.48565996e-02 8.12477529e-01 -6.81779087e-01 1.38982737e+00
-9.01541233e-01 6.02914691e-01 4.20444310e-01 -1.08813190e+00
1.31804681e+00 7.15390563e-01 6.22941494e-01 -5.05334377e-01
5.42482436e-01 -4.21577543e-02 2.85538286e-01 -5.91985285e-01
6.90995812e-01 -3.95558000e-01 -5.50462544e-01 6.98507249e-01
-5.93873784e-02 2.34964177e-01 -6.47193613e-03 5.99327572e-02
8.46251190e-01 -1.86577484e-01 3.69049348e-02 1.66774854e-01
8.60288322e-01 -5.72921991e-01 9.92931426e-01 2.67432511e-01
-6.13109231e-01 2.17459947e-01 9.41201031e-01 -2.77874261e-01
-3.50494057e-01 -6.08498812e-01 3.88079137e-02 1.82671225e+00
1.26115829e-01 -2.82351613e-01 -5.15761197e-01 -7.63411939e-01
-4.19508278e-01 6.46978855e-01 -7.72349954e-01 -2.62414277e-01
-5.08330643e-01 -8.01503420e-01 7.20667660e-01 5.75663388e-01
5.01311243e-01 -1.62476289e+00 -1.61801204e-01 5.03278673e-01
-7.97726989e-01 -9.45211411e-01 -4.17733133e-01 5.44749737e-01
-1.71658725e-01 -7.69164503e-01 -4.39149439e-01 -8.60351205e-01
2.15908751e-01 1.79571182e-01 1.01777208e+00 -2.42745876e-02
1.73132658e-01 6.97577223e-02 -8.76158357e-01 -4.18381184e-01
-2.23786011e-01 1.85217947e-01 3.58416117e-04 2.46444121e-01
5.32846391e-01 -6.11049294e-01 -3.69718343e-01 6.20939255e-01
-4.71726418e-01 -3.95558961e-02 2.31112912e-01 1.19880378e+00
-5.29645346e-02 1.82046033e-02 1.11527133e+00 -8.08893085e-01
1.03451955e+00 -7.46788859e-01 3.62120152e-01 1.61941260e-01
1.39611259e-01 -1.53285012e-01 7.68642068e-01 -7.28118896e-01
-1.33897758e+00 -2.71438867e-01 -4.30133075e-01 -3.29431295e-01
-9.53441113e-02 6.87564373e-01 -4.23751652e-01 4.62043613e-01
7.16141015e-02 6.17671274e-02 -2.45173365e-01 -2.30450153e-01
4.19486642e-01 1.16726017e+00 3.53711784e-01 -9.29099500e-01
-6.67954013e-02 1.64513126e-01 -8.34418058e-01 -7.74619937e-01
-9.62206423e-01 -7.35565066e-01 -4.48613614e-01 -4.33320701e-01
7.28915691e-01 -8.27611864e-01 -1.04964519e+00 6.72512352e-01
-1.35419738e+00 -3.94797117e-01 1.52529955e-01 4.51837122e-01
-4.74490941e-01 1.17620744e-01 -1.15126431e+00 -1.11375964e+00
-4.65705216e-01 -1.07080030e+00 7.57260323e-01 3.99638444e-01
-5.17971635e-01 -1.14612412e+00 1.48554295e-01 5.48782229e-01
2.71619737e-01 1.37412146e-01 8.60017478e-01 -9.29357648e-01
4.80492234e-01 6.80314079e-02 -1.69452056e-01 4.48972851e-01
3.46524268e-01 1.83096770e-02 -1.19889987e+00 4.41575870e-02
2.53136188e-01 -7.76178777e-01 6.25160813e-01 9.44225490e-02
9.44638789e-01 -2.87114412e-01 -1.14234954e-01 3.11629295e-01
4.20923352e-01 5.32765806e-01 6.27675891e-01 1.03345826e-01
6.70256734e-01 1.13367593e+00 7.77873516e-01 7.33574569e-01
7.68297374e-01 5.18849790e-01 2.74143964e-01 -5.49003892e-02
4.86661822e-01 4.16490324e-02 6.63923442e-01 1.44842339e+00
-8.31749365e-02 -3.24992597e-01 -6.05197251e-01 4.36844558e-01
-2.17389226e+00 -1.02911842e+00 -1.49577141e-01 1.48725653e+00
1.04972601e+00 -4.44617979e-02 2.59335130e-01 2.15741441e-01
9.46345925e-01 5.20762682e-01 -6.67997360e-01 -9.79477286e-01
-1.49590805e-01 -1.01911500e-01 -4.88934457e-01 4.45118636e-01
-8.84240508e-01 1.05772591e+00 5.04951668e+00 8.91808748e-01
-1.28346145e+00 1.64117198e-02 8.22188973e-01 -2.43584514e-01
-9.16439518e-02 -3.89747292e-01 -6.76962793e-01 5.72379053e-01
8.63549054e-01 -2.00771168e-02 4.17310655e-01 9.17439163e-01
5.34854412e-01 2.71121711e-01 -1.10815990e+00 1.14425135e+00
1.54868007e-01 -7.82938600e-01 -3.80571932e-01 -2.09715143e-01
4.24811035e-01 -1.45394012e-01 -2.28779316e-01 1.01156044e+00
4.72293586e-01 -1.08752418e+00 3.98139775e-01 3.45425814e-01
2.09417373e-01 -1.30881095e+00 1.11569238e+00 4.77672964e-01
-1.39435291e+00 -5.65061159e-02 -2.81307787e-01 -3.18907589e-01
3.46464753e-01 3.15529197e-01 -6.78181052e-01 6.82999432e-01
5.97747743e-01 9.68556702e-01 1.91848174e-01 4.75907713e-01
-3.79849941e-01 6.80093586e-01 1.22440944e-03 -2.98369467e-01
5.65155685e-01 -3.11034918e-01 2.39704132e-01 1.47486854e+00
-7.55947381e-02 2.82343626e-01 5.58360070e-02 4.58788991e-01
-3.26910108e-01 2.22756147e-01 -1.29111633e-01 -5.10002002e-02
6.15915835e-01 1.61796033e+00 -2.15275958e-01 -2.84842521e-01
-1.35024503e-01 1.09266198e+00 6.24599814e-01 3.50989610e-01
-9.79443312e-01 -6.30919397e-01 1.01606834e+00 -6.89425290e-01
2.98589230e-01 6.59961477e-02 1.80067152e-01 -1.02230680e+00
-9.98068601e-02 -9.70340431e-01 3.54196638e-01 -8.00086558e-01
-1.72335458e+00 9.88614082e-01 -3.86722058e-01 -1.06440473e+00
-5.58482826e-01 -2.96473026e-01 -1.27244020e+00 8.89294624e-01
-1.39984488e+00 -1.10590625e+00 -2.68738419e-01 7.19741464e-01
5.93653798e-01 4.78103459e-02 9.20418322e-01 2.84998089e-01
-1.06411839e+00 7.33775496e-01 3.56033519e-02 6.16983294e-01
7.88019240e-01 -1.03022766e+00 -2.82959819e-01 2.07524449e-01
-2.61685401e-01 6.35234594e-01 4.98734802e-01 -6.77007884e-02
-9.84064937e-01 -8.82126212e-01 1.04362416e+00 -1.43820262e-02
8.50274026e-01 -6.48658216e-01 -1.09906781e+00 5.10743439e-01
4.55775321e-01 -2.27564633e-01 1.10626292e+00 7.76074469e-01
-4.15705562e-01 -4.51653227e-02 -8.68151784e-01 5.47526538e-01
6.64435565e-01 -7.44007468e-01 -7.45956063e-01 6.60952106e-02
6.36141479e-01 -3.75543565e-01 -8.79509568e-01 3.76305759e-01
6.77689731e-01 -1.28834450e+00 6.82448924e-01 -6.82344854e-01
6.47884011e-01 2.19409525e-01 1.00903265e-01 -1.62045586e+00
1.23026650e-02 -9.60246444e-01 9.24226418e-02 1.83255291e+00
1.50225535e-01 -6.64310932e-01 5.88937759e-01 4.62870479e-01
-3.93987358e-01 -1.12578881e+00 -8.59718382e-01 -2.99776137e-01
1.89051643e-01 -4.81282234e-01 6.13388419e-01 1.40658820e+00
8.78754020e-01 6.81434870e-01 -6.73755407e-01 -2.41834611e-01
-2.14463249e-01 2.34558046e-01 6.91410124e-01 -9.67275620e-01
-1.83663458e-01 -7.70066559e-01 4.44318950e-02 -1.32478774e+00
8.31321716e-01 -5.66960037e-01 3.92688751e-01 -1.36658895e+00
-2.54586916e-02 -6.93457305e-01 -5.88013470e-01 4.35649812e-01
-3.19795370e-01 -1.43938169e-01 7.12854788e-03 4.44050394e-02
-6.90116227e-01 1.14846182e+00 1.06358981e+00 3.19463480e-03
-5.74554086e-01 4.55831997e-02 -7.32383251e-01 9.74664986e-01
9.79072452e-01 -2.77547598e-01 -4.29938108e-01 -2.59263953e-03
1.88207421e-02 4.63696480e-01 -1.08068205e-01 -3.51593226e-01
2.27057070e-01 -1.29471377e-01 -2.25273203e-02 -7.67585337e-01
7.53137469e-01 -3.79225075e-01 -4.38630790e-01 -6.38867021e-02
-5.63903093e-01 -7.56511763e-02 1.39716968e-01 3.04068953e-01
-6.70340121e-01 -5.59279695e-02 5.38598120e-01 3.51108685e-02
-2.79920399e-01 9.48233828e-02 -6.72197759e-01 1.07782364e-01
8.12352061e-01 1.25874281e-01 -4.34831113e-01 -8.08966398e-01
-6.37031794e-01 6.54051304e-01 -1.39635503e-01 6.22830451e-01
4.77009773e-01 -1.45824504e+00 -5.61284065e-01 -2.25455910e-01
9.61733907e-02 1.70587432e-02 6.65789485e-01 1.01736522e+00
2.04461768e-01 2.01439261e-01 -2.21361697e-01 -1.97113752e-01
-1.52231443e+00 1.57678261e-01 4.96817201e-01 -6.55688763e-01
-2.96645671e-01 1.13547289e+00 4.48538363e-01 -8.02067399e-01
4.58170921e-01 -2.62681842e-01 -5.06125212e-01 4.98625040e-01
4.58819270e-01 3.04065496e-01 -2.69740224e-01 -9.25578713e-01
-3.68041813e-01 5.57896011e-02 -3.11923921e-01 -9.64641869e-02
1.38369310e+00 -3.36761773e-01 -3.50322157e-01 8.24645340e-01
1.31972492e+00 7.14171454e-02 -1.08542979e+00 -2.62853235e-01
-1.51531219e-01 1.70366049e-01 -1.17858581e-01 -7.23000228e-01
-9.54935491e-01 1.19445920e+00 -2.15762123e-01 3.67521107e-01
1.09679401e+00 -5.94313666e-02 1.10257089e+00 4.43595797e-01
-3.33819874e-02 -1.32091308e+00 2.89328545e-01 1.09339809e+00
1.03233135e+00 -1.14317977e+00 -5.71194530e-01 -2.67283827e-01
-1.25627756e+00 1.08041966e+00 7.16124058e-01 3.89784649e-02
5.38881600e-01 2.80336410e-01 4.79559094e-01 -1.77324593e-01
-1.08407545e+00 -1.03196941e-01 5.35809956e-02 2.80467600e-01
7.90292382e-01 6.32442981e-02 -9.46440771e-02 1.11705530e+00
-5.00705600e-01 -4.13707078e-01 3.38651955e-01 6.79555595e-01
-1.97487965e-01 -1.26337600e+00 -2.21613601e-01 1.76775470e-01
-2.93677241e-01 -9.23309028e-02 -7.09558249e-01 6.34161472e-01
1.52612969e-01 1.51688421e+00 1.58694416e-01 -7.51759052e-01
2.34585851e-01 2.58994520e-01 -4.54679951e-02 -4.71689135e-01
-1.22662616e+00 2.81480968e-01 5.23787439e-01 -3.14531416e-01
-6.80573046e-01 -5.66871762e-01 -1.49984539e+00 -2.88559735e-01
-3.85310531e-01 7.49554634e-01 3.93708766e-01 1.14309740e+00
3.00313324e-01 7.09273636e-01 1.22913647e+00 -1.00627482e+00
-1.57404035e-01 -1.48159170e+00 -6.22671843e-01 3.99792612e-01
2.26467296e-01 -7.57305861e-01 -6.62955880e-01 -2.55165696e-01] | [13.046117782592773, 6.06048059463501] |
d8dc0f86-7f6f-4dc9-a614-08306bd8e1f0 | acoustic-absement-in-detail-quantifying | 2304.06183 | null | https://arxiv.org/abs/2304.06183v2 | https://arxiv.org/pdf/2304.06183v2.pdf | Acoustic absement in detail: Quantifying acoustic differences across time-series representations of speech data | The speech signal is a consummate example of time-series data. The acoustics of the signal change over time, sometimes dramatically. Yet, the most common type of comparison we perform in phonetics is between instantaneous acoustic measurements, such as formant values. In the present paper, I discuss the concept of absement as a quantification of differences between two time-series. I then provide an experimental example of absement applied to phonetic analysis for human and/or computer speech recognition. The experiment is a template-based speech recognition task, using dynamic time warping to compare the acoustics between recordings of isolated words. A recognition accuracy of 57.9% was achieved. The results of the experiment are discussed in terms of using absement as a tool, as well as the implications of using acoustics-only models of spoken word recognition with the word as the smallest discrete linguistic unit. | ['Matthew C. Kelley'] | 2023-04-12 | null | null | null | null | ['dynamic-time-warping'] | ['time-series'] | [ 2.70917535e-01 -4.28799301e-01 2.75007755e-01 -5.90962470e-01
-7.51076400e-01 -7.44274974e-01 6.17342174e-01 -8.28970000e-02
-6.24353290e-01 1.18463397e-01 3.40863466e-01 -5.39620101e-01
-1.95397839e-01 -3.85412514e-01 -1.51459515e-01 -7.95465112e-01
-1.73107252e-01 9.96722355e-02 6.85120746e-02 -2.50207365e-01
3.44124526e-01 6.24830961e-01 -2.02248430e+00 2.53412813e-01
3.01149905e-01 9.45693493e-01 2.06249759e-01 9.50023234e-01
-5.00544012e-01 -7.42878169e-02 -1.20230377e+00 -7.32416734e-02
1.90353811e-01 -4.96844053e-01 -6.69636428e-01 2.12219283e-01
1.81315154e-01 1.86471611e-01 -2.35595450e-01 9.25094128e-01
5.52993715e-01 5.40811062e-01 5.47813773e-01 -1.08156729e+00
-3.27327162e-01 6.32696748e-01 3.61449897e-01 5.97108006e-01
6.64213121e-01 -9.54846367e-02 7.53356159e-01 -7.99551427e-01
3.14357698e-01 1.38754570e+00 6.05646193e-01 1.94349319e-01
-1.24941194e+00 -4.23369259e-01 -1.37153029e-01 4.13695902e-01
-1.39918685e+00 -8.03710282e-01 7.79229701e-01 -5.00000238e-01
1.27374744e+00 7.85533845e-01 7.70507216e-01 8.09870660e-01
2.83380717e-01 4.53690410e-01 1.40389478e+00 -8.61191452e-01
1.56829983e-01 4.23609205e-02 2.58055151e-01 1.88468583e-02
-5.46312809e-01 2.89681882e-01 -4.90405411e-01 -1.18513852e-01
4.32795525e-01 -3.41308504e-01 -2.88178116e-01 3.73381734e-01
-1.16706014e+00 5.17692149e-01 -2.52023399e-01 1.11416423e+00
-3.72097164e-01 -1.88199833e-01 6.40328526e-01 8.00580084e-01
4.07886028e-01 3.67668480e-01 -3.39355767e-01 -8.73609960e-01
-9.40211177e-01 1.66860893e-01 8.12930882e-01 3.61034334e-01
1.80432778e-02 4.81143564e-01 3.45599085e-01 1.29826844e+00
2.53619313e-01 5.81409454e-01 1.05785227e+00 -6.54716730e-01
3.52256820e-02 -2.58713275e-01 -1.05459474e-01 -8.63544524e-01
-1.00674309e-01 1.42851621e-01 -4.17540856e-02 1.53185949e-01
7.89308667e-01 2.62389243e-01 -9.36882734e-01 1.51762831e+00
4.08305600e-02 3.24629545e-02 8.82732943e-02 6.18059218e-01
4.20017272e-01 1.09404588e+00 -2.06710115e-01 -8.53486300e-01
1.23692334e+00 -4.05504018e-01 -1.29319632e+00 7.65398378e-03
2.54629552e-01 -1.35706890e+00 1.36883783e+00 7.25030243e-01
-9.62289870e-01 -6.88138664e-01 -1.01125395e+00 1.62517667e-01
-6.43300951e-01 -3.83085132e-01 1.92846179e-01 9.20225561e-01
-9.43468988e-01 6.56702340e-01 -8.54779303e-01 -4.46500838e-01
-6.84738159e-01 1.18971027e-01 -1.75264940e-01 5.81744730e-01
-1.21040201e+00 9.72165346e-01 3.72972302e-02 1.80837363e-01
-2.46671230e-01 -4.69631881e-01 -7.97966480e-01 -1.68219104e-01
-1.93420410e-01 6.21595621e-01 1.84559453e+00 -8.86458933e-01
-1.78691971e+00 8.15654457e-01 -4.13938820e-01 -2.52820164e-01
1.87327638e-01 1.12951182e-01 -1.38672400e+00 -6.38747215e-02
-1.15715742e-01 -7.40912929e-02 9.66023505e-01 -6.49583697e-01
-4.83109266e-01 -2.15262651e-01 -7.32374310e-01 -1.67542532e-01
1.59264237e-01 5.03466427e-01 1.88858882e-01 -7.53407359e-01
4.40170705e-01 -7.97118485e-01 2.40831658e-01 -3.25261563e-01
-6.01052940e-02 -4.82609957e-01 9.30049002e-01 -9.91481781e-01
1.40707922e+00 -2.64125896e+00 -3.14570099e-01 3.80068690e-01
-5.38873196e-01 1.90217525e-01 1.22893400e-01 6.46613300e-01
-4.75269556e-01 8.27481598e-02 -2.43196368e-01 -1.88631251e-01
4.35350128e-02 3.89457911e-01 -6.85149968e-01 6.12532496e-01
-2.10933954e-01 4.27692980e-01 -7.81106174e-01 -6.64628521e-02
4.02898133e-01 6.61447108e-01 2.19377905e-01 6.36230782e-02
4.15606558e-01 2.10714608e-01 2.71075666e-01 3.08549494e-01
2.51100391e-01 9.90145862e-01 -1.06878765e-02 -1.60227969e-01
-8.74355197e-01 8.98073256e-01 -1.14807975e+00 1.38303053e+00
-5.13212919e-01 1.25096118e+00 4.69918735e-02 -9.95085835e-01
1.22965431e+00 8.12200129e-01 3.24727297e-01 -7.83545077e-01
2.00421810e-01 6.93101168e-01 6.41276240e-01 -6.22040629e-01
5.43228805e-01 -3.79126698e-01 5.61573319e-02 2.87375569e-01
-1.48873091e-01 -8.41031075e-01 1.70606107e-01 -4.32303995e-01
5.96399069e-01 -3.62004161e-01 2.67938107e-01 -5.35068154e-01
4.55237925e-01 -2.38780513e-01 1.72854230e-01 2.45531306e-01
-3.40013862e-01 4.16485935e-01 1.66794121e-01 -1.41985729e-01
-1.08971381e+00 -1.05273509e+00 -6.42491698e-01 9.03746367e-01
-3.70439261e-01 -2.37394243e-01 -7.99159586e-01 2.69897342e-01
-6.88113570e-02 1.00588632e+00 -3.26815397e-01 -1.45917252e-01
-6.60357296e-01 -4.15573835e-01 7.86768794e-01 2.77331948e-01
-2.81317621e-01 -1.26949108e+00 -5.48697531e-01 5.69520116e-01
-4.77715433e-02 -9.59326804e-01 -7.24368274e-01 4.44172293e-01
-8.26963663e-01 -5.23569286e-01 -4.66601044e-01 -9.16195631e-01
-9.01760440e-03 1.83899909e-01 5.59782147e-01 -3.04649711e-01
-3.49729687e-01 6.31050229e-01 -3.83221954e-01 -4.55249906e-01
-9.05073881e-01 -5.16967535e-01 5.89413881e-01 1.40377209e-01
4.97287065e-01 -8.49500775e-01 -1.64354011e-01 6.29503846e-01
-7.20161080e-01 -6.23931646e-01 -1.35278916e-02 5.80882132e-01
5.17087579e-01 3.19626957e-01 6.77341878e-01 -2.43759453e-01
9.53721941e-01 -1.06942594e-01 -3.84395123e-01 -8.24191794e-02
-4.80024308e-01 -2.88855612e-01 5.64883769e-01 -8.04803014e-01
-9.17013943e-01 -2.99545914e-01 -4.54700619e-01 -1.93883553e-01
-3.97839427e-01 5.56415975e-01 -1.41995534e-01 9.15979594e-02
6.72222793e-01 5.22292435e-01 4.01130348e-01 -7.33950794e-01
1.39341652e-01 1.24169827e+00 9.93684709e-01 -4.26816881e-01
4.07298505e-01 1.16956487e-01 -5.96563995e-01 -1.62466037e+00
-7.93877840e-02 -6.89731896e-01 -7.50028491e-01 -5.68201363e-01
5.13163388e-01 -2.68130749e-01 -5.56315958e-01 7.21082389e-01
-1.08691740e+00 -3.07667434e-01 -5.04630327e-01 8.65139842e-01
-7.30620325e-01 3.84173155e-01 -5.08949399e-01 -1.28984249e+00
-2.66610924e-02 -1.11024916e+00 8.14363003e-01 -4.47491482e-02
-5.06921172e-01 -9.63548005e-01 1.55448198e-01 1.63648084e-01
3.53287667e-01 -6.72387704e-02 6.01377726e-01 -8.87897134e-01
4.17086869e-01 -4.29418921e-01 6.61952972e-01 7.17250764e-01
5.74039340e-01 5.21898985e-01 -1.40737927e+00 -1.69186354e-01
6.03573203e-01 2.16200545e-01 1.21403880e-01 3.70576560e-01
8.80373180e-01 -1.75137267e-01 9.54624191e-02 1.22732203e-02
9.20630097e-01 1.19777632e+00 6.73493266e-01 -8.29398818e-03
1.09197609e-01 9.85115528e-01 8.16080332e-01 1.22491844e-01
-2.27113441e-01 8.88184607e-01 -2.30016127e-01 4.76420671e-01
-9.39027146e-02 -1.46731855e-02 7.52865672e-01 1.95216155e+00
1.45781845e-01 -1.14072062e-01 -1.01253879e+00 8.76630187e-01
-1.14400351e+00 -1.05764210e+00 -3.16498518e-01 2.53481340e+00
8.28155339e-01 2.85589248e-01 1.38597220e-01 1.05585420e+00
9.29126382e-01 2.82623231e-01 -9.21369344e-02 -1.22444773e+00
1.73791610e-02 3.84230465e-01 6.55363053e-02 8.65155578e-01
-6.58507049e-01 5.84943473e-01 7.46883249e+00 8.71676266e-01
-1.57081389e+00 -5.00240549e-02 6.61271662e-02 5.03937826e-02
-5.60946427e-02 -2.88320392e-01 -1.97883978e-01 6.05767667e-01
1.72032392e+00 -5.90445399e-01 5.08972406e-01 4.68554705e-01
6.56323791e-01 -9.21220705e-02 -1.11056685e+00 1.10658574e+00
-1.05965085e-01 -5.49064994e-01 -3.58812243e-01 9.52783413e-03
1.11892320e-01 -2.25758925e-01 1.49451032e-01 1.17005132e-01
-6.28152668e-01 -9.89661217e-01 1.13406944e+00 3.48435938e-01
6.82029605e-01 -5.69725096e-01 3.42329830e-01 1.70797616e-01
-1.32910240e+00 2.19127551e-01 -1.27253786e-01 -3.35906297e-01
5.92448592e-01 4.20115083e-01 -9.31507766e-01 1.62518367e-01
5.00204206e-01 3.49818200e-01 1.03953935e-01 1.11721635e+00
2.50341356e-01 9.41619515e-01 -5.22806108e-01 -1.89900145e-01
1.86740935e-01 -3.37995172e-01 7.54148185e-01 1.64485097e+00
5.68071246e-01 1.38249651e-01 -1.59277886e-01 3.82694304e-01
6.90120816e-01 3.29836398e-01 -5.63248754e-01 -5.74126720e-01
8.43231380e-01 7.22588837e-01 -7.33555436e-01 -3.08308899e-01
-4.32719022e-01 6.97778046e-01 -6.27636731e-01 2.81106442e-01
-5.26287913e-01 -8.12027633e-01 9.89851475e-01 1.72103895e-03
-7.47862831e-02 -5.45436978e-01 5.44102006e-02 -5.64531028e-01
1.29771620e-01 -8.21856260e-01 -1.73353299e-01 -7.33023465e-01
-1.21211040e+00 6.71105921e-01 2.88948894e-01 -1.24770284e+00
-6.40484810e-01 -8.58452380e-01 -7.36184835e-01 1.14954376e+00
-6.64885461e-01 -1.30900785e-01 1.40359461e-01 3.72309238e-01
9.41465080e-01 -6.30844459e-02 9.73400891e-01 4.23105985e-01
-9.68805477e-02 3.90506059e-01 2.41908282e-01 -7.92590305e-02
5.08537531e-01 -1.29301548e+00 8.48640203e-01 6.72778785e-01
5.26893556e-01 6.07756197e-01 1.11718452e+00 -3.03679913e-01
-1.08500171e+00 -3.52937579e-01 1.16336298e+00 -1.48112446e-01
9.72719967e-01 -4.10337210e-01 -1.29957068e+00 2.87078291e-01
7.92374760e-02 -3.74097317e-01 7.87265420e-01 -1.80117935e-01
-2.52725601e-01 -2.10443661e-01 -9.70395923e-01 4.79901254e-01
7.97519743e-01 -1.31602144e+00 -1.27355146e+00 2.53891498e-01
8.25572133e-01 -2.64127791e-01 -1.19942796e+00 -1.08733833e-01
9.11700368e-01 -8.97843659e-01 8.60536337e-01 -1.98813796e-01
-3.89592260e-01 -2.04874411e-01 -3.83890271e-01 -1.50121474e+00
5.15369810e-02 -8.86796653e-01 5.64135790e-01 1.25599980e+00
2.47639552e-01 -9.24434781e-01 8.04392993e-02 3.61351699e-01
-3.99958938e-01 -1.23727769e-01 -1.31781888e+00 -1.38519931e+00
-5.66031747e-02 -1.21912766e+00 4.71894205e-01 8.77461076e-01
4.60420102e-01 -9.00376216e-02 9.26601663e-02 -1.47118308e-02
2.12946773e-01 -2.64254332e-01 2.47119188e-01 -1.22013199e+00
-8.44020247e-02 -6.28567100e-01 -7.41916358e-01 -7.33504653e-01
6.52103266e-03 -3.99065375e-01 5.55683553e-01 -8.42651784e-01
-8.99197936e-01 -1.74414486e-01 -2.38396853e-01 1.24901295e-01
3.56284201e-01 -1.07598655e-01 3.64154458e-01 8.56600627e-02
6.60705626e-01 3.06553841e-01 8.96240413e-01 7.20662763e-03
-3.29315931e-01 3.13993543e-01 2.22430915e-01 5.37371993e-01
7.72180974e-01 -3.82723212e-01 -5.38016438e-01 -2.81169116e-02
-3.89908880e-01 2.97424078e-01 7.80401826e-02 -8.60629082e-01
2.31074676e-01 -5.03284670e-02 -3.66038591e-01 -4.24168080e-01
6.68692470e-01 -9.51871693e-01 6.35982633e-01 5.67264557e-01
-3.36301297e-01 4.29875761e-01 4.39677715e-01 3.27974647e-01
-6.35338068e-01 -4.24680293e-01 7.49828935e-01 1.81775510e-01
-7.63558745e-01 -3.71222705e-01 -9.54517245e-01 -3.14581960e-01
5.67575753e-01 -5.38897693e-01 -6.98201358e-02 -4.46847916e-01
-1.07773316e+00 -6.18330359e-01 -1.24622323e-02 6.85370207e-01
3.06270480e-01 -1.28997040e+00 -2.98630387e-01 4.08080459e-01
-1.15798719e-01 -6.84098780e-01 1.07027672e-01 8.46686065e-01
-3.60712558e-01 6.29312992e-01 -2.49852046e-01 -6.31955087e-01
-1.63146818e+00 3.45507205e-01 5.24305403e-01 6.47901893e-01
-4.90858793e-01 7.08346725e-01 -2.25784495e-01 -2.74180681e-01
3.11424166e-01 -7.96406686e-01 -1.19701572e-01 4.43463832e-01
6.30932271e-01 4.89988685e-01 3.85044336e-01 -9.67803776e-01
-4.84712213e-01 5.17749190e-01 1.04218952e-01 -9.99464512e-01
9.80381489e-01 -3.48284930e-01 -2.29149058e-01 1.55691373e+00
1.46485746e+00 2.27867827e-01 -6.71129405e-01 -5.80219142e-02
2.96849370e-01 -6.06080472e-01 -3.97162549e-02 -7.01733649e-01
-4.27005589e-01 8.40876460e-01 8.71540964e-01 1.00663996e+00
1.11855686e+00 -3.03706020e-01 6.22651577e-01 2.81599134e-01
2.96566844e-01 -1.22337615e+00 -4.31815267e-01 5.68313658e-01
1.12370741e+00 -7.10424423e-01 -4.86072987e-01 -3.24287057e-01
-3.74231994e-01 1.32944524e+00 -1.13811933e-01 3.94695625e-02
8.03757966e-01 3.08627397e-01 5.71970940e-01 1.69643298e-01
-5.52383780e-01 3.82211320e-02 3.15532714e-01 5.92338800e-01
7.89124727e-01 2.91769743e-01 -7.09883392e-01 2.55809903e-01
-9.81517673e-01 -4.80770648e-01 6.43548906e-01 9.39214706e-01
-5.50803006e-01 -1.24431813e+00 -8.57737005e-01 2.32239217e-01
-5.41639805e-01 -1.32781053e-02 -3.55267197e-01 6.36191428e-01
-2.96571434e-01 1.38435948e+00 4.57431853e-01 -5.07055461e-01
6.63599610e-01 7.37653673e-01 2.39224464e-01 -4.41907167e-01
-6.92602694e-01 3.68688762e-01 1.67409986e-01 -4.45849508e-01
-4.05481160e-01 -1.14468729e+00 -1.20861983e+00 -3.47438186e-01
-2.12011054e-01 5.80334723e-01 1.32844734e+00 9.78144526e-01
-3.13242555e-01 5.72802365e-01 7.26093113e-01 -6.92186177e-01
-5.02362609e-01 -1.08609581e+00 -1.05648255e+00 1.99314728e-01
5.66393256e-01 -3.91712546e-01 -8.12959850e-01 2.15076789e-01] | [14.678549766540527, 6.078554630279541] |
c1919c5b-1a95-42a7-93c5-aa026db279b1 | unsupervised-complementary-aware-multi | 2112.04701 | null | https://arxiv.org/abs/2112.04701v1 | https://arxiv.org/pdf/2112.04701v1.pdf | Unsupervised Complementary-aware Multi-process Fusion for Visual Place Recognition | A recent approach to the Visual Place Recognition (VPR) problem has been to fuse the place recognition estimates of multiple complementary VPR techniques simultaneously. However, selecting the optimal set of techniques to use in a specific deployment environment a-priori is a difficult and unresolved challenge. Further, to the best of our knowledge, no method exists which can select a set of techniques on a frame-by-frame basis in response to image-to-image variations. In this work, we propose an unsupervised algorithm that finds the most robust set of VPR techniques to use in the current deployment environment, on a frame-by-frame basis. The selection of techniques is determined by an analysis of the similarity scores between the current query image and the collection of database images and does not require ground-truth information. We demonstrate our approach on a wide variety of datasets and VPR techniques and show that the proposed dynamic multi-process fusion (Dyn-MPF) has superior VPR performance compared to a variety of challenging competitive methods, some of which are given an unfair advantage through access to the ground-truth information. | ['Michael Milford', 'Tobias Fischer', 'Stephen Hausler'] | 2021-12-09 | null | null | null | null | ['visual-place-recognition'] | ['computer-vision'] | [ 2.64666080e-01 -6.86995625e-01 2.41534617e-02 -3.65246236e-01
-1.10928237e+00 -8.60198438e-01 8.06281388e-01 3.83860648e-01
-4.62419242e-01 4.78559971e-01 -1.73066258e-01 -7.43478686e-02
-4.33778256e-01 -4.34419721e-01 -5.44017851e-01 -6.71567500e-01
2.41284631e-02 4.94938374e-01 6.20053291e-01 -3.86917591e-01
7.28880882e-01 8.63807380e-01 -1.92619419e+00 6.72973245e-02
5.74422717e-01 1.02397847e+00 5.75901926e-01 6.40966475e-01
1.68963999e-01 5.79024196e-01 -6.19290650e-01 -9.68117174e-03
6.40149891e-01 -3.13940525e-01 -5.71036518e-01 1.80584535e-01
5.40497184e-01 5.13865389e-02 -9.86964107e-02 9.82454419e-01
6.97644532e-01 6.71187997e-01 5.83296895e-01 -1.23593795e+00
-2.75735319e-01 -5.39700352e-02 -7.37871826e-01 6.07137859e-01
7.55403101e-01 -3.42338756e-02 9.35555339e-01 -1.04559350e+00
9.13144588e-01 9.79733944e-01 2.58709282e-01 1.98754799e-02
-1.36646593e+00 -3.89364481e-01 6.39405400e-02 3.58603567e-01
-2.03547525e+00 -6.81457520e-01 8.79132152e-01 -3.92145365e-01
8.84471834e-01 3.20286334e-01 3.90939474e-01 7.64218032e-01
2.09469292e-02 6.54044271e-01 1.23398495e+00 -5.44272363e-01
5.43935359e-01 1.25926090e-02 -2.79687226e-01 4.67969954e-01
1.16715975e-01 1.32165909e-01 -1.01577997e+00 -4.26373899e-01
5.45333505e-01 -3.04621726e-01 -4.26191002e-01 -8.14628720e-01
-1.23440385e+00 5.18924594e-01 3.72059762e-01 4.30413812e-01
-4.95338708e-01 1.23452000e-01 1.83450449e-02 1.63824290e-01
1.25927910e-01 3.74754369e-01 -2.19733879e-01 -1.06569357e-01
-1.32326901e+00 3.93399656e-01 4.73482311e-01 6.78893566e-01
1.01535666e+00 -1.24451645e-01 1.12222629e-02 7.58091807e-01
5.43794990e-01 6.97089851e-01 1.38109848e-01 -1.01486623e+00
4.30501133e-01 9.90296528e-02 2.72095531e-01 -1.51806235e+00
-2.80540111e-03 -1.75574757e-02 -3.00858557e-01 3.22061211e-01
1.64101049e-01 2.12751627e-01 -8.13008308e-01 1.44731641e+00
4.31391567e-01 -2.17503737e-04 5.16416430e-02 1.03561521e+00
5.11748731e-01 6.81421697e-01 -1.17697477e-01 -3.95844072e-01
9.88105357e-01 -5.79471886e-01 -6.56043112e-01 -4.55633491e-01
2.56425917e-01 -9.36864734e-01 4.19626772e-01 1.84717044e-01
-6.60145164e-01 -3.48370880e-01 -1.24575353e+00 2.26885900e-01
-3.57621610e-01 6.62847683e-02 1.31443590e-01 5.03817677e-01
-1.34282494e+00 1.88835174e-01 -5.03809810e-01 -8.19337428e-01
1.41132429e-01 5.88727176e-01 -6.97711945e-01 -1.35961294e-01
-7.34795094e-01 1.13714254e+00 4.57660437e-01 2.29931790e-02
-1.08594775e+00 -1.97939947e-01 -7.72752821e-01 -2.47185037e-01
7.02908874e-01 -4.24961120e-01 1.00108147e+00 -9.84778762e-01
-1.30193055e+00 7.47440696e-01 -4.02167886e-01 -5.73384285e-01
4.73431438e-01 2.38722965e-01 -2.00176507e-01 3.06703359e-01
2.46742889e-01 7.59801447e-01 9.45419133e-01 -1.62549508e+00
-8.64826858e-01 -4.19604510e-01 1.15353480e-01 7.01722205e-01
2.55479515e-01 2.09966525e-01 -6.15478814e-01 -4.00386184e-01
4.11595345e-01 -1.02995086e+00 -3.20392370e-01 4.28460352e-02
5.29168658e-02 7.01646134e-02 8.82184029e-01 -4.27913815e-01
9.13355410e-01 -2.23577642e+00 2.84868002e-01 5.15995860e-01
-1.04528191e-02 -3.15072909e-02 1.15777478e-02 5.19556403e-01
1.56443223e-01 1.18095456e-02 -2.65661567e-01 -3.00385803e-01
-2.48027161e-01 3.50346804e-01 -2.36384720e-01 9.27288413e-01
-2.37945318e-01 3.32259417e-01 -1.01463449e+00 -8.52911413e-01
7.27980316e-01 2.64780611e-01 -2.03621045e-01 3.19146216e-01
1.86055392e-01 4.57315981e-01 -4.77149576e-01 7.91441500e-01
6.41823232e-01 1.40085772e-01 5.62121868e-02 -1.11505307e-01
-2.82216340e-01 -2.67006546e-01 -1.49076593e+00 1.86101639e+00
-1.63226709e-01 7.34599113e-01 -1.67472214e-02 -6.71708047e-01
1.20544159e+00 1.79764003e-01 7.26135314e-01 -7.06201077e-01
1.49270758e-01 4.17477310e-01 -1.13346148e-02 -1.39734223e-01
9.11745429e-01 2.53923442e-02 -2.06112534e-01 3.01540226e-01
2.56483763e-01 -8.54000822e-02 2.62597382e-01 9.04457495e-02
1.24124658e+00 2.95080394e-02 7.51493990e-01 -2.20633134e-01
6.74042284e-01 2.68347729e-02 5.11737823e-01 1.01103413e+00
-7.23246932e-01 9.09751892e-01 -4.85553145e-02 -4.95860547e-01
-7.84320891e-01 -9.68878150e-01 -8.60890150e-02 8.79631162e-01
6.71563983e-01 -2.89974213e-01 -3.02544594e-01 -5.14084876e-01
-2.59913474e-01 5.84028542e-01 -5.97985923e-01 2.58996427e-01
-3.22953641e-01 -3.79902840e-01 3.80428165e-01 1.14415690e-01
5.84301829e-01 -9.57153857e-01 -1.16390419e+00 5.90845868e-02
-3.82214218e-01 -1.18765068e+00 -2.66360164e-01 1.42654419e-01
-4.49308217e-01 -8.99761081e-01 -6.35556340e-01 -3.44995350e-01
6.43639147e-01 8.35106671e-01 7.99732745e-01 5.38466647e-02
1.13299035e-01 6.76694930e-01 -6.14970207e-01 -1.02626041e-01
-1.95518285e-01 -2.45719980e-02 1.55084729e-01 4.07572418e-01
1.23155285e-02 -4.04793590e-01 -3.20665091e-01 6.16464615e-01
-7.95042038e-01 -1.71380445e-01 4.14957941e-01 5.38121879e-01
1.08181918e+00 1.86887100e-01 -1.16260648e-01 -2.65585184e-01
4.80655909e-01 -3.61891061e-01 -7.33011842e-01 6.05910599e-01
-3.47010791e-01 7.15675876e-02 6.63032085e-02 -2.92805612e-01
-9.66981053e-01 4.14011210e-01 2.10802004e-01 -9.24232006e-01
-1.81148618e-01 3.71879160e-01 -1.27260163e-01 -6.25691593e-01
8.58441055e-01 3.15423697e-01 -4.35331017e-01 -1.34087875e-01
4.83146131e-01 5.09037614e-01 6.92019403e-01 -6.07411802e-01
8.55166316e-01 5.87268412e-01 5.00381514e-02 -8.81909907e-01
-5.47365248e-01 -1.04694819e+00 -7.40895927e-01 -6.70305729e-01
8.61933589e-01 -7.14391053e-01 -2.55896538e-01 1.24902710e-01
-1.20534682e+00 2.02055443e-02 -5.41393496e-02 1.50264978e-01
-7.72003591e-01 3.69718015e-01 3.42817158e-01 -1.08006084e+00
-1.41779095e-01 -1.46094131e+00 1.19629157e+00 2.72471488e-01
-7.21888468e-02 -5.83995521e-01 1.67174235e-01 3.49472284e-01
2.67683715e-01 3.73861104e-01 1.03085518e-01 -5.73646963e-01
-8.80378246e-01 -1.10364594e-01 1.59852840e-02 -1.55990198e-01
1.15641735e-01 1.20143004e-01 -8.56344104e-01 -3.27437639e-01
-1.85610354e-02 1.38981506e-01 4.45567787e-01 4.05414313e-01
7.24282801e-01 1.40609786e-01 -4.16369438e-01 5.72821736e-01
1.69480312e+00 5.18416107e-01 6.85042560e-01 7.25511730e-01
6.71374679e-01 3.67312759e-01 1.07802331e+00 3.36060375e-01
4.27076191e-01 1.17729223e+00 4.07591969e-01 2.19515190e-01
1.82424888e-01 -2.25204185e-01 8.39906111e-02 2.21196339e-01
-2.47243643e-01 -4.52546328e-01 -1.23875511e+00 7.61144221e-01
-2.11435723e+00 -1.12013888e+00 3.16774040e-01 2.45531130e+00
-5.92542104e-02 -1.39792085e-01 2.28759963e-02 1.37878805e-01
8.43623459e-01 5.73180139e-01 -2.42865488e-01 -1.38489217e-01
-2.29046106e-01 -2.38933489e-01 7.95355737e-01 3.17934722e-01
-1.17093122e+00 8.23111415e-01 6.52573490e+00 9.37885225e-01
-1.10189700e+00 1.94853067e-01 1.98583931e-01 2.24667385e-01
-3.55306081e-02 3.87530267e-01 -4.61162657e-01 3.20310831e-01
7.01391518e-01 -3.62293541e-01 6.20733917e-01 8.25849652e-01
4.52727079e-02 -6.57681823e-01 -1.00079787e+00 1.62986207e+00
4.93154317e-01 -1.39197254e+00 -7.07800537e-02 1.45387635e-01
7.13315547e-01 1.47800341e-01 1.23237066e-01 -1.51513740e-01
1.25845388e-01 -7.14165986e-01 1.19465482e+00 3.87165695e-01
3.22309256e-01 -7.92131424e-01 7.46573031e-01 3.01533490e-01
-1.47352505e+00 -1.70181587e-01 -2.19225809e-01 2.84004956e-01
3.22274178e-01 -2.16870476e-02 -8.74668121e-01 9.53281045e-01
9.06155050e-01 4.20825928e-01 -8.21908534e-01 1.30820942e+00
-5.67418076e-02 2.74455976e-02 -5.00423908e-01 1.59292251e-01
2.26573914e-01 6.30232766e-02 1.01599526e+00 9.74569738e-01
4.89613831e-01 1.58691928e-01 3.70127022e-01 4.76705819e-01
2.40480170e-01 3.69090021e-01 -1.03839612e+00 3.09029788e-01
7.00307667e-01 1.20542276e+00 -1.07433641e+00 2.30676141e-02
-2.99495496e-02 7.68427253e-01 2.93762773e-01 3.30945522e-01
-6.05320632e-01 2.25414351e-01 4.33624059e-01 1.04085498e-01
4.74683970e-01 -4.90004927e-01 -1.10999770e-01 -9.56803441e-01
7.98129216e-02 -8.27040315e-01 5.87163270e-01 -1.16108680e+00
-9.19052124e-01 8.29656065e-01 3.93285483e-01 -1.53007805e+00
-2.72324830e-01 -1.28875971e-01 -2.44701430e-01 7.00680077e-01
-1.31035936e+00 -1.00809836e+00 -2.86971509e-01 5.57957828e-01
5.79837978e-01 -2.69800127e-01 6.06198788e-01 1.22408643e-01
-1.38772205e-01 2.34487683e-01 -6.95376322e-02 -1.25320226e-01
6.23509526e-01 -9.35520828e-01 -1.66219741e-03 1.34988570e+00
6.29591107e-01 4.36405599e-01 1.08019650e+00 -4.74422365e-01
-1.44868374e+00 -7.16681600e-01 5.88708460e-01 -5.75963140e-01
2.92275578e-01 -1.34677336e-01 -7.91644037e-01 4.49134976e-01
8.89145955e-02 3.25935215e-01 6.40982628e-01 -4.76154163e-02
-1.53467432e-01 -2.15193599e-01 -1.36086595e+00 4.97129977e-01
8.01602483e-01 -6.05523646e-01 -7.92029381e-01 4.42033745e-02
1.82390198e-01 -8.00845742e-01 -4.97635841e-01 3.79533142e-01
2.53505528e-01 -9.43523943e-01 1.01648486e+00 3.01771700e-01
-1.03266105e-01 -1.22246420e+00 -9.38335240e-01 -1.12700737e+00
-2.03652933e-01 -5.16277075e-01 1.55563802e-01 1.12872767e+00
1.81332201e-01 -3.41540605e-01 4.13541406e-01 6.10011101e-01
3.24566215e-01 -3.87425750e-01 -1.38726902e+00 -6.41583145e-01
-8.71686339e-01 -5.55830300e-01 4.18140799e-01 6.81501448e-01
-4.71413404e-01 3.05448938e-02 -5.80552220e-01 5.55001438e-01
8.34965229e-01 1.76814720e-01 9.10312653e-01 -9.46781576e-01
-1.25270188e-01 -6.01564907e-02 -1.04077148e+00 -6.44625783e-01
-1.23298038e-02 -6.32414639e-01 4.17572439e-01 -1.53597140e+00
-5.24735893e-04 -4.45813030e-01 -3.44338566e-01 2.92402089e-01
1.94793597e-01 2.24492282e-01 6.93695486e-01 5.79647064e-01
-1.01856065e+00 4.73981589e-01 6.35506809e-01 -2.49479398e-01
-4.14373964e-01 -3.72130066e-01 -4.28214759e-01 5.19762158e-01
4.55520958e-01 -7.39876747e-01 -4.80526179e-01 -4.72117513e-01
3.17635238e-02 3.05813938e-01 1.72214240e-01 -1.31497514e+00
5.75615466e-01 -4.81473714e-01 1.72963515e-01 -8.64490867e-01
4.96961534e-01 -9.84730065e-01 7.14907229e-01 -1.00575618e-01
-6.43074885e-02 4.12664682e-01 1.11544169e-01 8.89855444e-01
-2.42645010e-01 -2.52059013e-01 7.66731381e-01 -1.20240606e-01
-1.45320714e+00 2.13278040e-01 -3.49746495e-01 -3.31085682e-01
1.35276663e+00 -6.10638618e-01 1.03427784e-03 -3.33410561e-01
-3.41724128e-01 1.43263415e-01 8.07267249e-01 5.65787137e-01
7.78655171e-01 -1.34191763e+00 -4.21028465e-01 3.05066314e-02
4.59399462e-01 -1.19226508e-01 4.43834215e-01 5.37958562e-01
-6.34992063e-01 1.29380912e-01 -4.45198536e-01 -9.67779100e-01
-1.40745735e+00 7.06392765e-01 4.41356122e-01 -1.07216373e-01
-3.77076417e-01 6.40842140e-01 -3.54895294e-02 -9.80117992e-02
-4.96655703e-02 2.65518963e-01 -2.42159680e-01 1.68563053e-01
5.07857025e-01 1.35434434e-01 2.14077920e-01 -1.49497604e+00
-7.89323270e-01 6.95001185e-01 2.09332541e-01 -6.45456493e-01
1.14838982e+00 -4.56985027e-01 9.05084461e-02 5.08047283e-01
8.95151675e-01 -1.71319157e-01 -1.09521961e+00 -1.55621663e-01
-3.72407772e-02 -1.16921365e+00 1.46958381e-01 -5.03377020e-01
-9.42542732e-01 3.85575145e-01 9.54395831e-01 -8.04718956e-02
1.19312406e+00 -3.16565260e-02 -8.95597320e-03 5.17685473e-01
1.17238593e+00 -1.19804704e+00 -7.27720261e-02 2.01276064e-01
9.44017470e-01 -1.21993124e+00 2.38611892e-01 -1.91515282e-01
-6.61071181e-01 8.61349165e-01 3.79723340e-01 -1.74875811e-01
4.88642424e-01 6.66056797e-02 1.68240622e-01 -1.98347732e-01
-5.62230647e-01 -4.57786053e-01 2.23466024e-01 9.38692927e-01
-2.66402453e-01 1.04033776e-01 -2.11224094e-01 -2.77597100e-01
8.58144313e-02 -1.30671933e-01 5.00146985e-01 1.22815967e+00
-4.20614749e-01 -9.76255417e-01 -8.15966308e-01 1.95114538e-01
-1.07008204e-01 3.32011253e-01 -1.97018713e-01 6.16490364e-01
2.14864358e-01 1.08319736e+00 -2.43330210e-01 -4.33098316e-01
4.01487559e-01 -2.14480504e-01 5.16762674e-01 -3.82604808e-01
-4.67414826e-01 -6.30381033e-02 4.88869846e-05 -5.26987195e-01
-8.85499299e-01 -9.57549691e-01 -8.63485336e-01 -2.59799451e-01
-3.88054699e-01 -9.68784913e-02 6.36097372e-01 1.13764465e+00
3.23042154e-01 1.88782737e-01 5.77269256e-01 -1.42067170e+00
-1.05198063e-01 -3.95354807e-01 -4.94966537e-01 3.09573382e-01
6.60672486e-02 -1.18096542e+00 -2.24342212e-01 -4.93344069e-02] | [7.506068229675293, -1.8909430503845215] |
6148db23-f731-4e86-b71e-ba3284b6310b | unsupervised-improvement-of-audio-text-cross | 2305.01864 | null | https://arxiv.org/abs/2305.01864v2 | https://arxiv.org/pdf/2305.01864v2.pdf | Unsupervised Improvement of Audio-Text Cross-Modal Representations | Recent advances in using language models to obtain cross-modal audio-text representations have overcome the limitations of conventional training approaches that use predefined labels. This has allowed the community to make progress in tasks like zero-shot classification, which would otherwise not be possible. However, learning such representations requires a large amount of human-annotated audio-text pairs. In this paper, we study unsupervised approaches to improve the learning framework of such representations with unpaired text and audio. We explore domain-unspecific and domain-specific curation methods to create audio-text pairs that we use to further improve the model. We also show that when domain-specific curation is used in conjunction with a soft-labeled contrastive loss, we are able to obtain significant improvement in terms of zero-shot classification performance on downstream sound event classification or acoustic scene classification tasks. | ['Paris Smaragdis', 'Fabio Ayres', 'Tiago Tavares', 'Junkai Wu', 'Krishna Subramani', 'Cem Subakan', 'Zhepei Wang'] | 2023-05-03 | null | null | null | null | ['acoustic-scene-classification', 'scene-classification'] | ['audio', 'computer-vision'] | [ 6.41396642e-01 -6.20075464e-02 2.60629743e-01 -6.42092764e-01
-1.59384787e+00 -5.27695894e-01 7.15443492e-01 4.38076973e-01
-5.20987868e-01 4.08934206e-01 4.04492408e-01 -1.33147985e-01
5.54360040e-02 -3.75222415e-01 -4.39148396e-01 -4.73516047e-01
-2.16900166e-02 3.13520849e-01 3.58315468e-01 -2.04162113e-02
1.85659360e-02 1.75358996e-01 -1.86904848e+00 6.47665322e-01
2.90570736e-01 8.86379540e-01 7.31156319e-02 8.73157442e-01
-1.93706200e-01 7.73989320e-01 -4.89772469e-01 -8.96571055e-02
1.44790515e-01 -7.15975523e-01 -8.75807106e-01 -1.26709864e-01
6.21780694e-01 9.04748309e-03 3.46609242e-02 6.67841494e-01
7.69764781e-01 5.59388995e-01 7.97841132e-01 -1.13071060e+00
-9.75267217e-02 7.61248350e-01 -4.34702516e-01 1.55593067e-01
3.64075482e-01 -2.92951703e-01 1.31436121e+00 -8.06330264e-01
4.50577497e-01 1.26225853e+00 8.26008499e-01 4.84400362e-01
-1.56722498e+00 -8.48953784e-01 -3.16791534e-02 1.99865043e-01
-1.43362534e+00 -9.05174971e-01 8.70198667e-01 -4.71022129e-01
9.92971420e-01 3.17096144e-01 1.89929783e-01 1.07243288e+00
-2.71061540e-01 6.61481559e-01 8.36774468e-01 -1.01072395e+00
1.29281104e-01 1.16415240e-01 9.80992019e-02 4.64165419e-01
-3.99988025e-01 8.19368195e-03 -7.82955647e-01 -2.26954520e-01
3.01483393e-01 -4.37257439e-01 -6.40606582e-02 -3.31315666e-01
-1.02507186e+00 8.72010887e-01 2.12569699e-01 6.80843949e-01
6.90703094e-02 2.35572517e-01 8.19129109e-01 3.92810136e-01
7.27329731e-01 6.14553511e-01 -2.61796117e-01 -2.35809863e-01
-1.38644505e+00 1.20029628e-01 6.69757545e-01 7.68567801e-01
6.71046376e-01 1.90090761e-01 -2.10902631e-01 1.26947665e+00
8.48429203e-02 1.55524407e-02 5.40057659e-01 -1.00676823e+00
3.18970174e-01 1.58805400e-02 -9.44224223e-02 -5.73277533e-01
-3.83410811e-01 -2.13512465e-01 -2.54822284e-01 9.57474858e-02
3.58889520e-01 -1.03530213e-01 -1.04785430e+00 1.90775573e+00
4.48001549e-02 5.79618812e-01 5.11540584e-02 6.11872792e-01
6.13473415e-01 7.23919749e-01 2.48624057e-01 -2.03611284e-01
1.26056945e+00 -8.13731372e-01 -6.85323536e-01 -3.49398069e-02
9.46482241e-01 -1.01102650e+00 1.33612442e+00 3.83375734e-01
-1.01633477e+00 -5.08179605e-01 -9.83573437e-01 -1.00064948e-01
-2.66104400e-01 -8.67230743e-02 5.08388519e-01 7.08320975e-01
-6.87944353e-01 4.69824135e-01 -8.62383783e-01 -5.00121593e-01
2.76278645e-01 1.32157892e-01 -2.91209847e-01 -1.43877072e-02
-1.16423774e+00 5.95465422e-01 2.44704798e-01 -4.04622465e-01
-1.10822630e+00 -8.92182827e-01 -9.79751110e-01 1.74304694e-01
3.02039862e-01 -2.78882205e-01 1.57356238e+00 -1.00853539e+00
-1.23932242e+00 8.18258464e-01 -1.64569110e-01 -3.86910915e-01
2.56934404e-01 -8.26372132e-02 -3.75573844e-01 2.98703909e-01
1.05194308e-01 7.84660339e-01 9.00914788e-01 -1.16724956e+00
-7.35696495e-01 -1.11786671e-01 -4.05481532e-02 1.36388138e-01
-7.01197207e-01 4.70270395e-01 -2.99979210e-01 -8.73410583e-01
1.19118858e-02 -9.39421892e-01 -8.20355117e-02 -4.14897352e-02
-2.68301755e-01 -8.94022360e-02 7.15412974e-01 -4.85441774e-01
8.78971398e-01 -2.48723054e+00 -1.11681998e-01 1.92627236e-02
-2.32338980e-01 1.51065916e-01 -2.83077925e-01 5.13088882e-01
-1.75029695e-01 9.76273045e-02 -3.23055327e-01 -8.80356908e-01
4.10607122e-02 1.86511233e-01 -4.73664463e-01 1.33912027e-01
2.33382732e-01 3.22020859e-01 -9.00516570e-01 -6.15439773e-01
1.59697697e-01 6.50935113e-01 -8.44934821e-01 2.00498790e-01
-1.45256907e-01 4.73160863e-01 -1.02898382e-01 2.49170169e-01
8.12884867e-02 1.52045354e-01 2.55318582e-02 -5.92081174e-02
-2.35752854e-02 6.96843207e-01 -1.12813973e+00 1.90189743e+00
-8.66350591e-01 9.36801493e-01 8.95020142e-02 -1.24728143e+00
6.24181986e-01 7.03093171e-01 4.87885326e-01 -2.52101541e-01
9.84879583e-02 1.74378052e-01 -5.52919284e-02 -3.78638804e-01
4.49263096e-01 -8.08961511e-01 -1.39926672e-01 4.18983549e-01
4.77801591e-01 -3.00645381e-01 1.05840534e-01 7.94880688e-02
1.12925994e+00 4.98422608e-02 -1.69100706e-02 -3.84847745e-02
3.41829151e-01 -1.63024604e-01 3.37828070e-01 7.54271388e-01
-3.83366235e-02 9.98929918e-01 2.15647310e-01 1.83917165e-01
-9.07199085e-01 -1.04245412e+00 -3.73652279e-01 1.68720341e+00
-3.08034390e-01 -6.32409751e-01 -5.78539252e-01 -4.56127763e-01
-2.75686651e-01 9.67475235e-01 -2.84520954e-01 -1.87015101e-01
-3.77498209e-01 -5.76209664e-01 9.12793756e-01 7.22752094e-01
-2.77927015e-02 -8.54237974e-01 -3.72813195e-01 2.83125788e-01
-2.90763706e-01 -1.22654378e+00 -2.49745935e-01 7.37426102e-01
-5.85526049e-01 -6.35584235e-01 -7.90302753e-01 -8.76707196e-01
2.83939153e-01 1.48544773e-01 7.51704037e-01 -3.15671951e-01
-5.06395221e-01 5.54874480e-01 -6.27431273e-01 -4.76271778e-01
-3.84556413e-01 7.93412924e-02 -1.15227886e-02 5.64967319e-02
1.80293202e-01 -6.59139037e-01 -4.08214692e-05 1.27968952e-01
-1.01869094e+00 -1.43264830e-01 6.69457391e-02 9.05678272e-01
2.61200905e-01 6.27146587e-02 7.84371614e-01 -7.82491684e-01
5.96165299e-01 -2.51843661e-01 -1.89386368e-01 -2.62735486e-02
-2.36020967e-01 1.48864001e-01 5.13040602e-01 -6.19009674e-01
-1.09309709e+00 5.75934239e-02 -3.13563704e-01 -6.47277117e-01
-3.54044527e-01 5.79846263e-01 8.82921889e-02 -6.51122187e-04
8.63418281e-01 -1.11935198e-01 -3.74296248e-01 -6.26610458e-01
3.62585425e-01 9.73234892e-01 3.87180686e-01 -6.42945707e-01
5.40397108e-01 2.54886568e-01 -1.33203998e-01 -1.10123432e+00
-1.00066972e+00 -7.05282211e-01 -4.81624275e-01 -9.12257284e-02
9.26272213e-01 -9.82092619e-01 -3.79328206e-02 1.67612076e-01
-8.56806755e-01 -3.91687363e-01 -4.62335616e-01 7.48777568e-01
-7.99054027e-01 2.89640009e-01 -5.54889917e-01 -1.09285212e+00
1.85671613e-01 -8.88203502e-01 1.24749076e+00 -2.74138123e-01
-5.76122522e-01 -9.11082387e-01 1.67050242e-01 3.43549818e-01
3.21020722e-01 -2.04747338e-02 9.70604241e-01 -9.94275272e-01
-1.32336631e-01 -2.98804969e-01 3.81925739e-02 4.90658492e-01
1.02590092e-01 -9.13860425e-02 -1.55614555e+00 -1.84930041e-01
-2.10476413e-01 -7.92190194e-01 1.05122924e+00 1.93031698e-01
1.24960136e+00 1.80705845e-01 -2.55126595e-01 3.65843147e-01
1.00653398e+00 1.56040311e-01 3.39248508e-01 -1.18143283e-01
4.52996731e-01 7.95164585e-01 6.10650599e-01 4.37707782e-01
9.83289182e-02 9.72825706e-01 -1.15822576e-01 -9.18528736e-02
-3.59207243e-01 -3.08453947e-01 3.24073285e-01 7.73134410e-01
2.22629637e-01 -3.17577511e-01 -9.88776505e-01 9.40725148e-01
-1.59397364e+00 -1.04588604e+00 1.00539014e-01 2.14833236e+00
1.03400338e+00 2.39844605e-01 7.34483749e-02 6.51356697e-01
7.34914541e-01 1.76134571e-01 -6.35403991e-02 -3.09983313e-01
3.23425889e-01 4.71455574e-01 1.44249182e-02 5.90824962e-01
-1.21883035e+00 9.23624396e-01 6.50384760e+00 8.50935280e-01
-1.13430572e+00 3.47521335e-01 2.55802661e-01 -3.43606204e-01
-2.54804462e-01 4.72877212e-02 -6.75711513e-01 1.82684198e-01
1.35986745e+00 8.87834951e-02 2.57283777e-01 4.69588816e-01
1.58980444e-01 -6.31130785e-02 -1.37943053e+00 1.03081536e+00
2.35685959e-01 -9.78901029e-01 -1.51676536e-01 -3.12848091e-01
4.85780865e-01 -6.51656762e-02 -7.30558485e-02 5.26195109e-01
2.24318251e-01 -8.16708386e-01 6.46491110e-01 2.55282462e-01
8.60360742e-01 -7.01214790e-01 3.64512801e-01 2.15736344e-01
-1.22882998e+00 -7.07923025e-02 -1.90798879e-01 -4.16953340e-02
3.34028304e-01 4.12267715e-01 -1.07055604e+00 5.03654480e-01
5.96224606e-01 5.46592236e-01 -3.11832607e-01 1.12680519e+00
-6.32876679e-02 1.10199749e+00 -4.61958051e-01 2.10257903e-01
2.04070479e-01 2.64545441e-01 5.88312149e-01 1.48766983e+00
3.82396460e-01 -1.96670622e-01 3.88656676e-01 5.19640565e-01
1.70762576e-02 2.88031667e-01 -7.02026665e-01 -2.79559553e-01
4.57817525e-01 1.06244218e+00 -6.48096621e-01 -3.96955550e-01
-3.37199986e-01 8.98355842e-01 1.90812856e-01 2.63193339e-01
-6.42680526e-01 -7.66071081e-01 5.72416365e-01 1.74182847e-01
2.95169681e-01 -1.82241470e-01 -1.87820457e-02 -1.01831472e+00
-2.51238167e-01 -5.69124043e-01 4.86908853e-01 -8.49631071e-01
-1.34522164e+00 4.76677030e-01 2.57905900e-01 -1.24106908e+00
-4.78232801e-01 -3.02292913e-01 -6.82457089e-01 7.26998627e-01
-1.31525540e+00 -1.21917558e+00 3.54377069e-02 6.13602459e-01
6.40788913e-01 -1.78545535e-01 1.12537909e+00 6.12218022e-01
-2.05157712e-01 7.17221141e-01 -4.62119654e-03 1.53285176e-01
1.21186185e+00 -1.10521865e+00 -4.67780344e-02 7.17669785e-01
5.88990867e-01 3.76089275e-01 9.11480725e-01 -2.98735708e-01
-7.55998433e-01 -1.06251955e+00 9.46679056e-01 -3.51063013e-01
7.95805156e-01 -7.21264422e-01 -8.42076719e-01 8.20954084e-01
-1.28463283e-03 1.13392569e-01 1.14664340e+00 5.80385745e-01
-5.31257212e-01 -1.23410203e-01 -8.40256810e-01 3.95177513e-01
9.25357580e-01 -1.10272145e+00 -8.48665774e-01 3.31012309e-01
7.77088284e-01 7.63081992e-03 -7.31419384e-01 3.60115051e-01
4.30310726e-01 -5.72227597e-01 8.87908101e-01 -8.43147397e-01
1.86312288e-01 -2.03825653e-01 -4.82838899e-01 -1.21449447e+00
5.76692596e-02 -4.94282186e-01 4.12543833e-01 1.67518079e+00
5.07992566e-01 -2.93363631e-01 4.36372876e-01 3.17796469e-01
-4.51932341e-01 -8.20430964e-02 -1.16045141e+00 -9.94847536e-01
1.77781671e-01 -9.33362067e-01 7.89555758e-02 1.04917097e+00
4.06524897e-01 5.75116038e-01 -4.33826298e-01 4.52694930e-02
2.90717095e-01 5.18085435e-02 5.86064458e-01 -1.29981816e+00
-4.63411927e-01 -2.37259492e-01 -6.03234410e-01 -7.40419924e-01
4.29645181e-01 -1.19746840e+00 4.40008640e-01 -1.28906608e+00
3.20683653e-03 -5.84651768e-01 -6.41872346e-01 7.42486477e-01
1.33863743e-02 5.55366278e-01 2.33886749e-01 -7.82899782e-02
-6.20891690e-01 5.51705599e-01 5.80064893e-01 -1.37686387e-01
-2.16335431e-01 -1.05618455e-01 -4.98189360e-01 5.78106046e-01
6.70518398e-01 -7.65430570e-01 -5.24151087e-01 -3.01193595e-01
-1.14860244e-01 5.55635132e-02 3.05128455e-01 -1.34077024e+00
6.57627657e-02 1.91454947e-01 1.27431020e-01 -2.07343832e-01
8.22665393e-01 -6.44474268e-01 -5.38455956e-02 3.78478430e-02
-9.97779548e-01 -5.73571205e-01 3.49268347e-01 5.69653273e-01
-5.51910520e-01 -6.71182632e-01 9.12673414e-01 6.77992180e-02
-4.40788597e-01 -2.13379756e-01 -7.39281416e-01 1.97118878e-01
8.53105068e-01 1.35518700e-01 1.36514008e-01 -6.46597385e-01
-9.79859948e-01 -8.68665129e-02 2.53298670e-01 4.27736938e-01
3.13288331e-01 -1.30433404e+00 -6.40050054e-01 1.47099286e-01
4.33835238e-01 -4.01155174e-01 9.84925181e-02 8.11982274e-01
-5.76837212e-02 3.19726408e-01 9.68986377e-02 -5.86067677e-01
-1.57415140e+00 2.87540078e-01 2.28510469e-01 -7.35232309e-02
-5.88984311e-01 1.02336407e+00 3.06876719e-01 -4.82579857e-01
5.30999482e-01 -2.25635648e-01 -1.05237231e-01 3.65521461e-01
5.41383624e-01 1.05582476e-01 2.35497400e-01 -5.93793154e-01
-2.86685407e-01 4.15956259e-01 -6.78601814e-03 -6.99500084e-01
1.32881951e+00 -7.77665079e-02 4.55501139e-01 1.03576875e+00
1.24717653e+00 1.65656686e-01 -1.00089717e+00 -1.51489273e-01
1.64584428e-01 -4.02723283e-01 3.01509798e-01 -6.43831432e-01
-6.39625788e-01 1.38528323e+00 7.05854595e-01 2.91957319e-01
1.07465553e+00 1.15865529e-01 6.51032507e-01 5.28547049e-01
2.09854737e-01 -1.04430449e+00 2.14676619e-01 6.15653217e-01
5.78860581e-01 -1.15848672e+00 -3.44156832e-01 -3.19654644e-01
-4.98105526e-01 9.60236609e-01 2.86626846e-01 9.53942239e-02
6.72120333e-01 3.53016049e-01 1.18712120e-01 2.86335386e-02
-8.69759798e-01 -5.56338310e-01 2.29455858e-01 6.62676752e-01
7.48161614e-01 -1.80217877e-01 -2.87941769e-02 4.98817444e-01
-1.51650175e-01 -1.73881769e-01 3.48210990e-01 1.12090337e+00
-6.20226622e-01 -1.30623007e+00 -3.33496362e-01 4.21416342e-01
-6.93008423e-01 -3.25419039e-01 -3.07768017e-01 5.19055665e-01
-7.67474324e-02 1.19946849e+00 3.26399982e-01 -2.62594640e-01
1.97214872e-01 7.41935432e-01 5.00033915e-01 -1.07138681e+00
-5.37497818e-01 4.62804854e-01 5.28247237e-01 -1.57223046e-01
-5.62593102e-01 -7.18990445e-01 -1.36290061e+00 2.33818933e-01
-5.07990241e-01 2.03051269e-01 6.21762514e-01 9.32456493e-01
1.67726457e-01 7.42672443e-01 5.05884707e-01 -1.01091528e+00
-3.91724169e-01 -1.07632577e+00 -5.49374044e-01 4.49297458e-01
4.54332650e-01 -8.37096274e-01 -4.68361080e-01 3.81311268e-01] | [15.219592094421387, 5.092373371124268] |
5937937e-2806-4722-8ad6-36652538a274 | an-empirical-evaluation-of-similarity | 1401.3973 | null | http://arxiv.org/abs/1401.3973v1 | http://arxiv.org/pdf/1401.3973v1.pdf | An Empirical Evaluation of Similarity Measures for Time Series Classification | Time series are ubiquitous, and a measure to assess their similarity is a
core part of many computational systems. In particular, the similarity measure
is the most essential ingredient of time series clustering and classification
systems. Because of this importance, countless approaches to estimate time
series similarity have been proposed. However, there is a lack of comparative
studies using empirical, rigorous, quantitative, and large-scale assessment
strategies. In this article, we provide an extensive evaluation of similarity
measures for time series classification following the aforementioned
principles. We consider 7 different measures coming from alternative measure
`families', and 45 publicly-available time series data sets coming from a wide
variety of scientific domains. We focus on out-of-sample classification
accuracy, but in-sample accuracies and parameter choices are also discussed.
Our work is based on rigorous evaluation methodologies and includes the use of
powerful statistical significance tests to derive meaningful conclusions. The
obtained results show the equivalence, in terms of accuracy, of a number of
measures, but with one single candidate outperforming the rest. Such findings,
together with the followed methodology, invite researchers on the field to
adopt a more consistent evaluation criteria and a more informed decision
regarding the baseline measures to which new developments should be compared. | ['Joan Serrà', 'Josep Lluis Arcos'] | 2014-01-16 | null | null | null | null | ['time-series-clustering'] | ['time-series'] | [ 1.87829718e-01 -7.43132651e-01 -1.62596479e-01 -3.62993568e-01
-5.93169570e-01 -7.19714522e-01 7.38422215e-01 5.64243376e-01
-3.70335877e-01 5.26658118e-01 -1.07834399e-01 -2.13228330e-01
-7.99388409e-01 -5.29954433e-01 9.16527882e-02 -8.21934998e-01
-6.60504580e-01 1.47700265e-01 1.54825628e-01 -2.72965729e-01
6.14148378e-01 5.96405268e-01 -1.92970240e+00 -2.30518937e-01
8.87900352e-01 1.29781282e+00 -2.19748244e-01 2.84341931e-01
-3.54202464e-02 3.76934230e-01 -9.16780651e-01 -9.96248946e-02
7.92031586e-02 -5.80690384e-01 -5.21470010e-01 -5.50321937e-02
-1.28807887e-01 4.32485104e-01 1.48585498e-01 9.28136289e-01
5.57404697e-01 3.84516388e-01 7.42688417e-01 -1.61926591e+00
-4.79277521e-01 4.43817765e-01 -2.49208316e-01 4.76646721e-01
5.26557982e-01 -7.50156771e-03 1.06362951e+00 -4.21511948e-01
3.84543806e-01 6.53195322e-01 9.00898218e-01 -1.04678966e-01
-1.16368675e+00 -6.19086742e-01 -6.81166351e-03 5.43974996e-01
-1.42908859e+00 -1.80386722e-01 1.09850299e+00 -5.19350231e-01
6.99737251e-01 7.10596323e-01 6.15252495e-01 8.22893083e-01
2.53953218e-01 2.06748039e-01 1.57571435e+00 -4.73134041e-01
5.99779069e-01 1.15928575e-02 4.70079243e-01 3.11265178e-02
3.39259386e-01 -5.12697399e-02 -7.80178607e-02 -4.79143113e-01
3.63912731e-01 1.24867558e-01 -2.42032245e-01 -4.21258062e-01
-1.22950280e+00 6.96612716e-01 -6.09525554e-02 1.17072594e+00
-3.26315314e-01 -4.13306832e-01 7.91987121e-01 7.33696938e-01
6.43413901e-01 4.71122861e-01 -3.28650326e-01 -4.90782082e-01
-1.01576102e+00 1.95728078e-01 7.92515695e-01 4.87915635e-01
3.18190873e-01 4.11384441e-02 -2.14822683e-02 7.83816576e-01
-9.75519940e-02 3.36250305e-01 9.18244779e-01 -6.92720175e-01
1.64333865e-01 5.23242712e-01 -1.06175222e-01 -1.44144690e+00
-4.83261198e-01 -5.42823315e-01 -1.03711116e+00 -4.48481273e-03
3.13240677e-01 9.65031162e-02 -2.68603742e-01 1.58174515e+00
1.57438412e-01 3.12057763e-01 -1.90244317e-01 7.06389546e-01
3.56694281e-01 4.99119520e-01 -1.32946894e-01 -8.10246170e-01
1.32431269e+00 -3.84044111e-01 -7.36820996e-01 4.91468400e-01
4.51396972e-01 -9.66030359e-01 8.86677682e-01 5.33414900e-01
-7.49825239e-01 -6.31592155e-01 -1.00546336e+00 5.49639583e-01
-5.63120008e-01 -1.00585796e-01 6.72181070e-01 5.96744597e-01
-1.03755856e+00 9.45448935e-01 -8.17770302e-01 -7.05480158e-01
-2.35146731e-01 1.95302427e-01 -2.37771079e-01 4.07662600e-01
-1.10693848e+00 8.49485219e-01 2.15167865e-01 3.10285147e-02
-8.94397646e-02 -4.69610929e-01 -4.17320549e-01 -1.71661049e-01
-9.10181832e-03 -2.74097681e-01 1.05224788e+00 -8.45679879e-01
-1.16723931e+00 7.67307937e-01 -3.54680582e-03 -5.22784948e-01
4.24365729e-01 2.32461482e-01 -1.04898274e+00 9.60893556e-02
1.03273064e-01 -2.75981188e-01 7.06517160e-01 -8.10364246e-01
-5.99993825e-01 -4.38248128e-01 -2.14011937e-01 -1.40945375e-01
-5.06037235e-01 4.50576663e-01 7.14656860e-02 -1.02712548e+00
2.78177679e-01 -8.84314537e-01 -2.17197493e-01 -3.44225198e-01
-5.36014363e-02 -5.55993974e-01 6.44542634e-01 -3.99825782e-01
1.80385780e+00 -2.26958203e+00 3.34890676e-03 4.78744507e-01
6.86362758e-02 3.06754764e-02 1.36709273e-01 9.09098804e-01
-4.58944649e-01 3.14155333e-02 -4.63997006e-01 -1.15366824e-01
2.17275202e-01 -1.33631364e-01 -5.52473962e-01 7.95728445e-01
7.68291252e-03 2.78278559e-01 -8.64634931e-01 -3.24938089e-01
4.58420038e-01 2.76435226e-01 -5.56795783e-02 8.88119936e-02
2.75832027e-01 4.12096888e-01 -5.44691026e-01 4.79241401e-01
3.20703804e-01 -2.20019698e-01 -1.39545072e-02 -1.74730271e-01
-4.49809313e-01 1.08391382e-01 -1.27247608e+00 1.27957129e+00
-3.58004980e-02 6.52458727e-01 -4.15455133e-01 -1.52799308e+00
1.09687793e+00 4.72770065e-01 1.07534671e+00 -6.28775716e-01
4.02492493e-01 6.01611495e-01 3.23784173e-01 -3.87146235e-01
4.05263543e-01 4.73599173e-02 -2.05598325e-01 4.62380379e-01
-1.87990084e-01 -1.00046106e-01 6.01552427e-01 -4.10486311e-01
1.04204297e+00 -1.50511637e-01 7.87358463e-01 -5.05016685e-01
7.36105204e-01 -3.16215493e-02 2.62903154e-01 4.28638607e-01
-5.43836296e-01 3.67237866e-01 3.13939005e-01 -4.18437868e-01
-9.47620153e-01 -7.52458215e-01 -4.85211700e-01 7.49345839e-01
-8.64051506e-02 -4.73637551e-01 -4.13374066e-01 -3.40942033e-02
-5.35320416e-02 5.18684804e-01 -6.44856989e-01 -1.49855185e-02
-4.12374228e-01 -9.55019116e-01 4.45190668e-01 1.84288606e-01
1.90916985e-01 -1.00555789e+00 -8.20408881e-01 3.06677938e-01
-1.34278938e-01 -8.85388017e-01 -1.28880247e-01 2.53462315e-01
-1.27573776e+00 -1.18163633e+00 -8.58064771e-01 -5.33253610e-01
2.83720672e-01 4.69428658e-01 8.71409595e-01 -1.46729993e-02
-9.13544074e-02 4.10324425e-01 -8.30074430e-01 -3.86580437e-01
-2.97238171e-01 -1.22804139e-02 2.69477546e-01 5.81427850e-02
6.08926415e-01 -9.50003922e-01 -4.67604429e-01 6.49777412e-01
-8.59085679e-01 -6.48222864e-01 3.12576026e-01 4.25447345e-01
3.94300818e-01 4.26044732e-01 6.67591989e-01 -2.46900976e-01
9.72612083e-01 -6.52459979e-01 -4.79761392e-01 2.17745334e-01
-9.36447024e-01 -1.91111490e-01 9.12132144e-01 -4.83944237e-01
-3.46902937e-01 -6.26331508e-01 2.66111344e-02 -3.37642819e-01
-2.41595060e-01 6.88124001e-01 4.43666309e-01 -1.16773717e-01
6.57674909e-01 3.76759410e-01 -2.33857539e-02 -4.58447307e-01
2.05620974e-02 6.05807424e-01 4.81753200e-01 -6.17685854e-01
7.87389874e-01 4.05687660e-01 -5.89173473e-02 -1.04444456e+00
-3.73572648e-01 -7.70872891e-01 -5.36473632e-01 -2.72422731e-01
5.19216895e-01 -4.33726966e-01 -5.23277402e-01 5.30249119e-01
-8.26017737e-01 1.85116678e-01 -2.93107390e-01 8.07324350e-01
-6.17563069e-01 6.54819667e-01 -3.63231927e-01 -9.59163964e-01
-2.26679206e-01 -9.12248075e-01 7.55464017e-01 3.57799679e-02
-6.57772124e-01 -1.23875833e+00 3.26974213e-01 -6.40434928e-06
5.58791459e-01 5.97922444e-01 7.78558791e-01 -1.09623170e+00
9.21729878e-02 -4.53792155e-01 2.22903401e-01 1.79494366e-01
4.47689772e-01 4.18263137e-01 -9.40941751e-01 -4.11554009e-01
3.92782748e-01 1.97685063e-01 2.45922774e-01 4.58833277e-01
1.21439886e+00 1.07730448e-01 -1.56536698e-01 2.66383588e-01
1.34131229e+00 6.61722422e-01 4.52099770e-01 5.88559628e-01
1.20295554e-01 7.43599892e-01 8.08885157e-01 7.98856616e-01
1.56202152e-01 7.51433194e-01 7.85025023e-03 1.71334222e-01
3.45982850e-01 4.30527925e-01 2.62034893e-01 1.50563955e+00
-4.52502191e-01 -1.47344097e-01 -8.74680400e-01 3.84706825e-01
-1.67824578e+00 -1.33454406e+00 -3.07025641e-01 2.52538228e+00
5.49574792e-01 1.64459899e-01 5.97892880e-01 1.03784668e+00
8.67962599e-01 2.12649286e-01 -1.84564829e-01 -2.46090487e-01
-1.97880968e-01 1.48642585e-01 6.28419220e-02 -3.39193568e-02
-1.14669085e+00 1.97834358e-01 6.76340437e+00 8.28089297e-01
-1.43534815e+00 -2.66086161e-01 4.79513884e-01 3.26377928e-01
-1.04776889e-01 -6.30770102e-02 -2.07165778e-01 7.16803610e-01
1.33357453e+00 -8.41079295e-01 1.91878483e-01 5.52944362e-01
5.22889078e-01 1.19467959e-01 -9.26538229e-01 1.25178218e+00
9.83708054e-02 -7.90326118e-01 -3.72098595e-01 -1.11908838e-03
5.22897482e-01 -2.21111640e-01 -5.55656031e-02 1.64407358e-01
-3.19400668e-01 -6.33789122e-01 6.20005667e-01 5.41483402e-01
3.53850245e-01 -7.08411932e-01 8.02541494e-01 1.00330107e-01
-1.56216550e+00 -1.08855233e-01 -5.42821959e-02 -4.88522857e-01
1.63213432e-01 8.13380539e-01 -1.65389463e-01 1.09992552e+00
7.24424362e-01 8.36052835e-01 -4.88651007e-01 1.33536303e+00
2.71473020e-01 6.88745737e-01 -3.26410383e-01 -3.39986563e-01
9.84869897e-02 -5.10941744e-01 4.37210858e-01 9.11842942e-01
6.97176695e-01 -1.49629898e-02 3.61753404e-02 4.63826865e-01
6.75955594e-01 6.06536388e-01 -7.50070691e-01 -3.64500523e-01
6.96493268e-01 1.13690364e+00 -1.20817924e+00 -3.75130445e-01
-5.77219963e-01 5.10423541e-01 -1.44826785e-01 1.01930104e-01
-9.27391112e-01 -5.75653076e-01 6.54482365e-01 -2.08192170e-01
-2.86596995e-02 -3.66951913e-01 -3.59415323e-01 -1.00583351e+00
3.12245458e-01 -1.00057399e+00 6.76366746e-01 -4.20224249e-01
-1.62140417e+00 8.03550422e-01 4.54293609e-01 -2.00347805e+00
-5.36785722e-01 -4.58734959e-01 -7.01619327e-01 5.08614063e-01
-1.09314764e+00 -3.65449607e-01 -3.72352779e-01 5.51657319e-01
3.34251285e-01 -1.37203664e-01 9.41045582e-01 6.45002902e-01
-4.93911058e-01 3.36762846e-01 3.54645967e-01 -1.27532199e-01
6.21238232e-01 -9.62002337e-01 1.99475333e-01 7.25138426e-01
-4.59254533e-02 7.11467087e-01 1.12904298e+00 -3.54052186e-01
-1.22506809e+00 -6.03262544e-01 9.70623851e-01 -3.04538697e-01
1.16308904e+00 2.23980155e-02 -9.96897519e-01 1.63967431e-01
1.32752270e-01 -2.79585123e-01 1.09418738e+00 1.83566898e-01
-2.95155048e-02 -1.96214035e-01 -9.34093058e-01 5.62379777e-01
9.11993504e-01 -5.51782370e-01 -9.39384997e-01 1.48539290e-01
3.37129414e-01 2.92011559e-01 -1.42993975e+00 5.36705673e-01
5.25276542e-01 -1.22158670e+00 8.73791337e-01 -3.24059039e-01
-3.34477844e-03 -3.08796257e-01 -3.04811746e-01 -1.21454298e+00
-3.35376561e-01 -6.42477930e-01 7.38188177e-02 1.33369887e+00
6.33566752e-02 -1.06855953e+00 3.92497391e-01 3.02609026e-01
-1.57564744e-01 -5.85455239e-01 -8.14697623e-01 -1.21661699e+00
-6.31993487e-02 -8.14361572e-01 6.55839860e-01 1.35299385e+00
3.77806604e-01 2.55948864e-02 -1.54997781e-01 -1.43068433e-01
6.67001486e-01 5.14483452e-01 6.16763115e-01 -1.69112575e+00
4.09009159e-02 -9.80406821e-01 -8.49702477e-01 -3.03995520e-01
-4.60517779e-02 -6.62817717e-01 -2.80200928e-01 -1.10554254e+00
-9.00753886e-02 -3.39604139e-01 -6.80393755e-01 2.43097879e-02
3.78738195e-02 1.81451768e-01 3.36192362e-02 6.28145874e-01
-3.31857413e-01 5.43154001e-01 9.06602681e-01 6.69704527e-02
-1.75487518e-01 2.98682898e-01 -4.61614579e-01 7.09011793e-01
9.42157984e-01 -2.75330156e-01 -6.61345065e-01 1.08450316e-01
-1.31252035e-01 3.21065485e-02 1.64538562e-01 -1.35024202e+00
2.18361706e-01 -2.62971908e-01 -8.17117468e-02 -5.38460970e-01
8.56947526e-02 -1.03994274e+00 5.39667904e-01 6.33958578e-01
-2.16273457e-01 7.15413392e-01 3.85501911e-03 3.08007240e-01
-5.89183211e-01 -1.49943456e-01 5.90043008e-01 2.05972731e-01
-8.27881455e-01 2.26978213e-01 -1.87869936e-01 5.06985597e-02
1.18795204e+00 -5.84183872e-01 -2.00844891e-02 -3.19596946e-01
-5.43391407e-01 -7.56968185e-02 5.50076008e-01 6.62435293e-01
2.93847591e-01 -1.55359828e+00 -6.74484611e-01 -2.32306477e-02
4.30490941e-01 -7.33026922e-01 2.05568582e-01 1.20533252e+00
-3.99155140e-01 6.54739261e-01 -4.08750236e-01 -7.46496141e-01
-1.10053384e+00 7.14528501e-01 3.58181708e-02 -8.13486651e-02
-5.10743856e-01 3.21304873e-02 -2.61403710e-01 -1.56865343e-01
1.89707562e-01 -6.28677726e-01 -3.92844111e-01 4.36257213e-01
5.01164615e-01 6.96035266e-01 2.29657993e-01 -6.86903179e-01
-5.58425546e-01 8.31637740e-01 5.00862539e-01 -1.10005431e-01
1.32581854e+00 -3.76989357e-02 -2.34651014e-01 1.06457412e+00
1.16732252e+00 -1.40053287e-01 -4.52531338e-01 -1.64236519e-02
4.92773414e-01 -4.97677803e-01 -4.26679820e-01 -1.59367487e-01
-6.25718296e-01 5.20543277e-01 6.88051045e-01 1.12571526e+00
1.48115170e+00 -2.99038261e-01 3.09800565e-01 1.63392887e-01
6.72904372e-01 -9.99167144e-01 -1.92057177e-01 3.22871357e-01
8.43603611e-01 -1.05575287e+00 -5.23207821e-02 -3.86689276e-01
-2.47879520e-01 1.07090783e+00 7.22483397e-02 -2.36161187e-01
7.43890882e-01 1.01855606e-01 -1.50619466e-02 -8.34432170e-02
-5.84852695e-01 -1.70652717e-01 5.17952621e-01 3.52438450e-01
9.14442360e-01 1.54876197e-02 -1.09783161e+00 4.98596072e-01
-3.66709769e-01 -1.11604676e-01 1.31062731e-01 8.83811533e-01
-3.57332647e-01 -1.14590347e+00 -5.56021512e-01 5.46724916e-01
-4.20258552e-01 3.23831886e-01 -2.53892362e-01 1.03970289e+00
-1.65556163e-01 1.22989607e+00 -1.96218565e-02 -4.70218062e-01
4.56871182e-01 5.01045352e-03 2.26769924e-01 -1.17632359e-01
-6.80604041e-01 -2.58786455e-02 -1.15134455e-01 -4.23225641e-01
-1.00743520e+00 -9.83747244e-01 -8.73542964e-01 -5.25896668e-01
-2.41864249e-01 6.51923954e-01 5.97117603e-01 9.17923808e-01
2.12970972e-01 4.27755833e-01 1.05656421e+00 -7.42962182e-01
-6.55025840e-01 -9.77012813e-01 -7.95881689e-01 8.08419228e-01
4.54258621e-02 -7.29885399e-01 -7.06530273e-01 -1.96223557e-01] | [7.265561580657959, 3.3236923217773438] |
ff3b5bdb-0c70-46c3-a381-378e1074b253 | language-guided-audio-visual-source | 2303.16342 | null | https://arxiv.org/abs/2303.16342v1 | https://arxiv.org/pdf/2303.16342v1.pdf | Language-Guided Audio-Visual Source Separation via Trimodal Consistency | We propose a self-supervised approach for learning to perform audio source separation in videos based on natural language queries, using only unlabeled video and audio pairs as training data. A key challenge in this task is learning to associate the linguistic description of a sound-emitting object to its visual features and the corresponding components of the audio waveform, all without access to annotations during training. To overcome this challenge, we adapt off-the-shelf vision-language foundation models to provide pseudo-target supervision via two novel loss functions and encourage a stronger alignment between the audio, visual and natural language modalities. During inference, our approach can separate sounds given text, video and audio input, or given text and audio input alone. We demonstrate the effectiveness of our self-supervised approach on three audio-visual separation datasets, including MUSIC, SOLOS and AudioSet, where we outperform state-of-the-art strongly supervised approaches despite not using object detectors or text labels during training. | ['Kate Saenko', 'Bryan Russell', 'Oriol Nieto', 'Justin Salamon', 'Bryan A. Plummer', 'Andrea Burns', 'Arijit Ray', 'Reuben Tan'] | 2023-03-28 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Tan_Language-Guided_Audio-Visual_Source_Separation_via_Trimodal_Consistency_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Tan_Language-Guided_Audio-Visual_Source_Separation_via_Trimodal_Consistency_CVPR_2023_paper.pdf | cvpr-2023-1 | ['audio-source-separation'] | ['audio'] | [ 3.48811090e-01 -2.12995782e-01 -1.91730514e-01 -3.39527220e-01
-1.30281222e+00 -1.05851877e+00 5.59892356e-01 1.14684820e-01
-3.77906531e-01 3.43366861e-01 2.63870835e-01 2.97807902e-01
1.72933303e-02 -7.39136040e-02 -9.65819299e-01 -4.15681243e-01
-1.49308890e-01 2.74300456e-01 4.28425282e-01 3.97947758e-01
6.27220096e-03 1.40520364e-01 -2.03718710e+00 6.84821963e-01
2.56883293e-01 1.30954909e+00 1.89961746e-01 1.28148699e+00
1.16723927e-03 1.08069301e+00 -2.61022985e-01 -9.53198746e-02
1.77071124e-01 -4.62099940e-01 -6.89995587e-01 3.43237102e-01
1.04637086e+00 -1.86793298e-01 -3.20877343e-01 8.99322510e-01
5.92955351e-01 1.21041529e-01 6.91667914e-01 -1.54746938e+00
-4.41740096e-01 7.36570239e-01 -4.38933611e-01 1.32326737e-01
6.66290104e-01 1.46257773e-01 1.39109254e+00 -1.26315880e+00
3.91340554e-01 1.11384737e+00 6.76782310e-01 4.45805758e-01
-1.37092471e+00 -7.23835766e-01 1.35349050e-01 2.35118121e-01
-1.39542544e+00 -1.13496697e+00 9.05483246e-01 -6.33320212e-01
7.06005096e-01 2.20041737e-01 4.69000638e-01 1.21688807e+00
-6.39584303e-01 1.11330676e+00 7.38841176e-01 -6.65513575e-01
2.66539097e-01 3.98313105e-01 -5.88131286e-02 7.01684117e-01
-4.35307592e-01 4.78132144e-02 -1.19355476e+00 -2.16725245e-01
4.70506579e-01 -3.16216677e-01 -3.73245835e-01 -5.99499226e-01
-1.08927131e+00 4.16705132e-01 -9.72896814e-03 1.44325241e-01
5.55031151e-02 3.64393562e-01 4.43216115e-01 3.03993613e-01
4.82442588e-01 2.43030712e-01 -4.54234064e-01 -1.77565932e-01
-1.27251434e+00 -2.67274138e-02 7.16603041e-01 1.02034569e+00
6.06754780e-01 1.59164026e-01 -1.21311046e-01 7.62879670e-01
4.50291067e-01 6.75963640e-01 3.01742226e-01 -1.08034968e+00
4.00321335e-01 4.91601415e-02 -2.83388160e-02 -4.99437004e-01
-7.67781436e-02 -2.53062755e-01 -2.18537524e-01 1.32753327e-01
3.21166873e-01 -9.36982334e-02 -8.49631429e-01 1.73831308e+00
2.21355721e-01 6.77140415e-01 2.52463464e-02 9.90177274e-01
1.15069354e+00 5.42523861e-01 -1.53122172e-01 -2.20783845e-01
1.18679690e+00 -1.22480309e+00 -5.38280964e-01 -2.76270330e-01
1.12324543e-01 -8.10480058e-01 1.07420969e+00 4.81682211e-01
-1.14254761e+00 -9.04842317e-01 -8.55694711e-01 -3.85644995e-02
-7.38063678e-02 5.89125752e-01 1.72239155e-01 2.82216042e-01
-7.22726822e-01 4.20885921e-01 -8.18972945e-01 -2.67044872e-01
5.12795627e-01 1.01893373e-01 -4.38814640e-01 1.76155597e-01
-7.83427179e-01 2.38676056e-01 1.61933854e-01 -2.58606702e-01
-1.60171616e+00 -8.51941288e-01 -9.85937536e-01 7.31867999e-02
6.85499728e-01 -3.94616902e-01 1.43939435e+00 -1.44899583e+00
-1.41999257e+00 9.76909041e-01 -1.47401035e-01 -5.09434760e-01
4.01760072e-01 -5.55749357e-01 -4.00269121e-01 6.96188033e-01
8.22639987e-02 8.78552198e-01 1.69104159e+00 -1.36000621e+00
-7.60142505e-01 -1.56416520e-01 9.39688981e-02 2.08477378e-01
-4.85612303e-01 1.92477390e-01 -7.22279906e-01 -7.10682988e-01
-2.94344217e-01 -7.87688792e-01 4.55666065e-01 4.81673837e-01
-3.75156581e-01 -2.32095495e-01 9.17006791e-01 -4.56491500e-01
7.49122679e-01 -2.69875336e+00 2.74466544e-01 -1.59133166e-01
8.69081542e-02 1.63506895e-01 -4.26873922e-01 1.32686362e-01
-1.16614580e-01 -3.59952629e-01 -1.36595950e-01 -7.47781456e-01
8.38693306e-02 -8.49151313e-02 -7.58679509e-01 4.68909293e-01
4.10307407e-01 5.07429540e-01 -1.14429998e+00 -8.30270231e-01
7.05164298e-02 5.40203094e-01 -6.04886591e-01 6.74527824e-01
-5.32535791e-01 3.88655394e-01 3.90832350e-02 7.44902372e-01
2.43195653e-01 -1.67484507e-01 -3.32711674e-02 -3.27231646e-01
1.46175921e-01 2.84930795e-01 -1.45036614e+00 2.17196941e+00
-4.38595563e-01 1.00396776e+00 5.54442823e-01 -9.55116212e-01
5.97444296e-01 6.72244310e-01 4.74154860e-01 -2.14116469e-01
-9.29105878e-02 1.41146451e-01 -4.85317081e-01 -8.17199230e-01
-2.47221272e-02 -1.93665043e-01 9.33189020e-02 5.66314936e-01
8.53180885e-01 -2.64839441e-01 1.54742390e-01 2.87806034e-01
9.44307387e-01 4.96393472e-01 -1.35995775e-01 2.52720743e-01
5.21910012e-01 -2.99227506e-01 2.51595080e-01 7.29902685e-01
-8.05721432e-02 9.20099914e-01 2.20853612e-01 1.58103451e-01
-6.00341260e-01 -1.32064128e+00 5.04360534e-03 1.82074523e+00
-1.84008777e-01 -6.69812441e-01 -5.61717391e-01 -8.33017528e-01
2.41778437e-02 3.23552221e-01 -3.40337604e-01 1.28053557e-02
-2.10361972e-01 1.89415235e-02 9.35423374e-01 6.32993996e-01
-7.57199600e-02 -9.00459170e-01 -4.43112940e-01 -1.30554348e-01
-2.56886393e-01 -1.53061390e+00 -6.70769095e-01 4.28121656e-01
-4.19712842e-01 -1.07423997e+00 -7.11903095e-01 -9.08559799e-01
4.13240820e-01 3.56868595e-01 1.19460726e+00 -2.57301211e-01
-5.15216947e-01 1.14980972e+00 -2.80233860e-01 -4.10744727e-01
-4.04112041e-01 -2.77965844e-01 2.76263267e-01 4.54182297e-01
2.96095274e-02 -7.27140903e-01 -3.26695979e-01 1.10666268e-01
-7.77628779e-01 -2.21293747e-01 1.99507952e-01 5.93563199e-01
5.97661257e-01 -1.72167607e-02 4.30738926e-01 -3.51465732e-01
1.94957331e-01 -3.35856467e-01 -4.43835527e-01 2.31789261e-01
-1.11266062e-01 -5.09342924e-02 6.98539436e-01 -8.23529184e-01
-7.53645897e-01 6.32640362e-01 3.78422678e-01 -1.13180900e+00
-5.17630875e-01 1.80549666e-01 -1.74078077e-01 7.48744980e-02
7.43954897e-01 1.44741461e-01 -1.69252902e-01 -7.23300576e-01
5.84314346e-01 9.14961874e-01 1.03590202e+00 -4.71628457e-01
7.62565434e-01 7.31712043e-01 -2.10593879e-01 -1.03527999e+00
-1.26427686e+00 -9.47047114e-01 -8.02177846e-01 -2.90167898e-01
8.86772990e-01 -1.44509840e+00 -6.10982060e-01 3.70916247e-01
-1.16727138e+00 -1.34740427e-01 -5.05000353e-01 7.70522773e-01
-7.69519091e-01 4.15902495e-01 -3.04971308e-01 -1.13249910e+00
-8.97513479e-02 -8.02457333e-01 1.43519342e+00 -9.32659674e-03
-1.53937444e-01 -7.77000606e-01 2.34329402e-01 5.89226127e-01
-6.73073381e-02 -2.01855913e-01 5.17719269e-01 -8.89534891e-01
-4.77321088e-01 -1.14364706e-01 -1.71310425e-01 6.71377301e-01
2.57436574e-01 9.12062600e-02 -1.68416488e+00 -3.09129387e-01
-1.22010581e-01 -1.03415287e+00 1.16629374e+00 1.35365397e-01
1.00052691e+00 -2.28615761e-01 -5.00919670e-02 7.19424367e-01
1.07345903e+00 -3.87959063e-01 -1.63094491e-01 -2.30424985e-01
8.91690493e-01 6.78974926e-01 5.40786266e-01 5.11422813e-01
1.26984581e-01 7.16896832e-01 5.34648836e-01 -1.55313835e-01
-4.73492861e-01 -4.60170478e-01 8.58313024e-01 4.86671805e-01
3.32474709e-01 -1.06742859e-01 -6.75541282e-01 6.95636749e-01
-1.78876519e+00 -1.02451575e+00 1.53669089e-01 2.18002319e+00
1.14718676e+00 5.25585338e-02 3.75127733e-01 4.08182979e-01
5.73668063e-01 1.69337317e-01 -5.52097559e-01 2.66098917e-01
-1.11656547e-01 2.43454814e-01 7.26092011e-02 4.51051474e-01
-1.58081341e+00 7.30760336e-01 6.39556360e+00 7.31133342e-01
-1.25055480e+00 6.03910349e-02 -4.17431369e-02 -6.92140937e-01
-2.15527881e-02 6.68790787e-02 -5.71953952e-01 2.16930404e-01
9.90780771e-01 1.70284629e-01 6.66098773e-01 7.18956172e-01
5.90636991e-02 3.64339389e-02 -1.84848547e+00 1.33876252e+00
6.51378155e-01 -1.23255372e+00 -9.83724073e-02 -5.23092985e-01
3.84350866e-01 3.82239670e-02 -2.06144787e-02 1.28757954e-01
-2.10495293e-02 -9.21542764e-01 1.15129364e+00 4.65194464e-01
8.61108005e-01 -3.11029702e-01 3.32288146e-02 3.21473837e-01
-1.39271700e+00 -8.69947448e-02 7.10667148e-02 1.55537203e-01
6.70728609e-02 2.55095154e-01 -8.84579659e-01 4.02448952e-01
1.01323473e+00 1.01684952e+00 -5.96341848e-01 1.00156021e+00
-2.56776750e-01 1.00552917e+00 -4.22042876e-01 4.88981247e-01
-1.16329536e-01 4.42867011e-01 9.89129126e-01 1.47326291e+00
4.01009284e-02 -4.30280566e-01 5.65950871e-01 8.02997947e-01
-1.35977715e-01 4.25339192e-02 -4.55137491e-01 -3.54357034e-01
4.31830853e-01 1.19973648e+00 -4.81060535e-01 -3.11260134e-01
-4.07185137e-01 9.45748985e-01 5.81006147e-02 3.53688687e-01
-8.67128432e-01 -3.39750171e-01 5.08041382e-01 1.31739378e-01
6.43447399e-01 -3.84638011e-02 2.09630221e-01 -1.18330467e+00
1.75035030e-01 -9.39964175e-01 4.81948107e-01 -1.11805201e+00
-1.35787010e+00 3.57640803e-01 -6.17374964e-02 -1.62206900e+00
-3.64077896e-01 -5.46695650e-01 -4.70696449e-01 4.16521668e-01
-1.51313937e+00 -1.28218257e+00 -1.51285604e-01 8.68249893e-01
6.99010670e-01 -4.43190306e-01 8.39419186e-01 3.53775054e-01
-1.50556579e-01 5.02938807e-01 -8.25993270e-02 3.95946056e-01
1.23856175e+00 -1.26550329e+00 -5.51062115e-02 5.83198786e-01
1.21200275e+00 9.60690379e-02 8.17826569e-01 -2.30722085e-01
-1.52262568e+00 -1.04380929e+00 4.77467716e-01 -5.16337633e-01
8.96153986e-01 -7.94381917e-01 -7.94422448e-01 5.44079483e-01
2.65524089e-01 4.87152189e-01 9.90543485e-01 4.41425815e-02
-8.22507918e-01 -3.46055984e-01 -5.69395363e-01 1.79875106e-01
9.22127306e-01 -1.29495776e+00 -7.92830944e-01 4.75417376e-01
9.06805158e-01 -3.68307203e-01 -3.55989337e-01 2.88663149e-01
5.95289648e-01 -8.41142178e-01 1.10757589e+00 -8.11338723e-01
2.85985351e-01 -5.59588313e-01 -2.91780651e-01 -1.01529264e+00
1.08270915e-02 -9.06986117e-01 -4.00941521e-01 1.60860276e+00
3.07365537e-01 2.25637004e-01 4.89656508e-01 -8.06422830e-02
1.20867141e-01 -1.11554638e-01 -9.53662276e-01 -8.19706738e-01
-5.91769695e-01 -8.68001044e-01 -2.52617802e-02 8.47061217e-01
6.58771023e-02 7.25408673e-01 -6.37135267e-01 3.49690497e-01
8.10666740e-01 2.89469749e-01 9.13702309e-01 -1.33404636e+00
-8.21476102e-01 -6.45336956e-02 -5.21928608e-01 -9.17630851e-01
5.97649872e-01 -8.36101055e-01 3.28619212e-01 -1.10006750e+00
1.86518177e-01 6.35574982e-02 -3.60010922e-01 6.61573052e-01
1.89041927e-01 6.59456611e-01 3.32025707e-01 2.22232267e-01
-1.17142892e+00 6.33103848e-01 6.22238517e-01 -5.25926769e-01
-1.68133497e-01 3.00798062e-02 -4.97944266e-01 1.04057646e+00
2.14594632e-01 -5.99926889e-01 -5.17449081e-01 -6.21521890e-01
1.57534227e-01 1.82293072e-01 7.69967914e-01 -1.11135101e+00
4.14108634e-01 7.99102336e-02 2.25818753e-01 -4.80127037e-01
5.89909375e-01 -9.35212135e-01 -3.26178223e-01 -1.55008465e-01
-8.35161746e-01 -5.27643085e-01 2.93947339e-01 9.45158422e-01
-4.43595916e-01 -1.84219539e-01 6.69245601e-01 1.31576940e-01
-3.97978157e-01 8.07168707e-02 -3.05222243e-01 3.69161814e-01
8.49491000e-01 1.37267694e-01 -1.34871658e-02 -6.31661057e-01
-1.11211824e+00 1.86810166e-01 2.22867742e-01 6.80815041e-01
7.05432951e-01 -1.30182850e+00 -7.45982766e-01 2.10896313e-01
3.80631715e-01 -1.03124335e-01 1.44278735e-01 6.74006641e-01
9.58476067e-02 2.08686233e-01 6.25974312e-02 -1.02625477e+00
-1.70223987e+00 6.18032753e-01 2.18060285e-01 3.84585261e-01
-4.72401202e-01 1.03249919e+00 2.71642089e-01 -1.55729249e-01
1.10004532e+00 -3.94871712e-01 -1.83780566e-01 2.79141873e-01
6.40136242e-01 2.42161408e-01 -8.37167203e-02 -6.52768731e-01
-5.06742537e-01 6.27493501e-01 3.37580353e-01 -5.11093199e-01
1.24217689e+00 -1.76570907e-01 1.09438151e-01 1.08855140e+00
1.17991710e+00 4.16062117e-01 -1.50777221e+00 -6.51632071e-01
-1.91359192e-01 -3.59214157e-01 1.33289769e-01 -6.58806264e-01
-9.23372030e-01 1.16230738e+00 7.22584128e-01 1.59689918e-01
1.08499908e+00 3.93886834e-01 4.76100951e-01 5.20930767e-01
-9.45487246e-02 -1.07959139e+00 7.01685786e-01 3.86616945e-01
7.66027033e-01 -1.35961294e+00 -1.21440515e-01 -2.60407537e-01
-6.61090374e-01 1.06930614e+00 3.62926900e-01 -1.24551982e-01
7.34784245e-01 5.25586128e-01 2.84106702e-01 5.11911605e-03
-1.05396760e+00 -5.34119368e-01 7.83783019e-01 6.85667574e-01
4.31146860e-01 -3.55710685e-01 7.84001708e-01 6.92339063e-01
1.23399822e-02 -4.19473052e-02 1.19746648e-01 9.08170581e-01
-3.56522083e-01 -7.94061542e-01 -4.24549639e-01 -2.10189391e-02
-6.83916032e-01 -2.55292118e-01 -7.51225710e-01 1.31603986e-01
1.36464208e-01 1.12410247e+00 1.74759701e-01 -3.06356430e-01
1.05868593e-01 3.80140781e-01 6.33154392e-01 -8.54266644e-01
-4.31124687e-01 6.05223477e-01 -5.64263798e-02 -4.52591568e-01
-8.94989014e-01 -5.76107144e-01 -1.32074833e+00 6.83840811e-01
-3.99197310e-01 3.46149594e-01 4.11113650e-01 1.04958570e+00
3.18603188e-01 4.67702121e-01 7.50384212e-01 -1.16064775e+00
-5.42023838e-01 -8.09271514e-01 -5.81673682e-01 4.78722841e-01
1.00048852e+00 -5.82121789e-01 -6.94789886e-01 8.16761732e-01] | [14.764171600341797, 4.928518772125244] |
17f9a28f-3745-468c-a7e9-ee8a441a9256 | information-theoretic-evaluation-of-privacy | 2106.06046 | null | https://arxiv.org/abs/2106.06046v5 | https://arxiv.org/pdf/2106.06046v5.pdf | Information Theoretic Evaluation of Privacy-Leakage, Interpretability, and Transferability for Trustworthy AI | In order to develop machine learning and deep learning models that take into account the guidelines and principles of trustworthy AI, a novel information theoretic trustworthy AI framework is introduced. A unified approach to "privacy-preserving interpretable and transferable learning" is considered for studying and optimizing the tradeoffs between privacy, interpretability, and transferability aspects. A variational membership-mapping Bayesian model is used for the analytical approximations of the defined information theoretic measures for privacy-leakage, interpretability, and transferability. The approach consists of approximating the information theoretic measures via maximizing a lower-bound using variational optimization. The study presents a unified information theoretic approach to study different aspects of trustworthy AI in a rigorous analytical manner. The approach is demonstrated through numerous experiments on benchmark datasets and a real-world biomedical application concerned with the detection of mental stress on individuals using heart rate variability analysis. | ['Bernhard Freudenthaler', 'Lukas Fischer', 'Bernhard A. Moser', 'Mohit Kumar'] | 2021-06-06 | null | null | null | null | ['heart-rate-variability'] | ['medical'] | [ 7.67653733e-02 5.89754283e-01 -2.12171480e-01 -8.23654413e-01
-6.17263973e-01 -3.03940743e-01 3.70078743e-01 4.11768287e-01
-5.02705753e-01 8.78564298e-01 -2.26490013e-02 3.94780375e-03
-5.53852081e-01 -4.97055322e-01 -5.45235693e-01 -7.51517951e-01
-1.47007987e-01 3.23876023e-01 -6.41721189e-01 4.04124588e-01
8.16132054e-02 4.02709097e-01 -9.77342904e-01 2.45731026e-01
9.84551668e-01 1.16717410e+00 -9.30400372e-01 1.89450219e-01
4.77999538e-01 4.41571146e-01 -4.15490299e-01 -9.71185207e-01
5.78965783e-01 -5.70565820e-01 -8.87739062e-01 -3.37685704e-01
1.58387184e-01 -5.92304587e-01 -4.16498780e-02 1.54355013e+00
3.11280310e-01 -2.46700235e-02 8.71135890e-01 -1.69477296e+00
-1.00739110e+00 8.18799615e-01 -3.75099853e-02 -5.31270169e-03
1.94346923e-02 6.62312508e-02 1.19273376e+00 -4.28366929e-01
2.77963728e-01 1.23673201e+00 5.17468870e-01 7.02502370e-01
-1.26619947e+00 -4.80096668e-01 -2.73338199e-01 2.49111935e-01
-1.51510775e+00 -4.66799885e-01 5.53064167e-01 -5.46337426e-01
1.62585095e-01 5.40951908e-01 6.91554487e-01 9.83316541e-01
7.88784862e-01 6.45752013e-01 1.22579837e+00 -4.03886557e-01
8.02746415e-01 9.26146150e-01 5.56759953e-01 8.55000019e-01
7.90076554e-01 4.42896634e-01 -8.34511399e-01 -6.20694697e-01
3.06636274e-01 -9.45081413e-02 -7.94533044e-02 -6.24821365e-01
-8.75333965e-01 1.03216541e+00 2.63196856e-01 2.53463060e-01
-4.36928093e-01 1.36109591e-01 4.95007753e-01 4.01899070e-01
9.65921462e-01 3.75686467e-01 -1.21423863e-01 3.53328973e-01
-7.94973731e-01 1.35445222e-01 1.09672570e+00 7.88311779e-01
4.06124234e-01 3.96780297e-02 -3.93982798e-01 8.14672210e-04
8.56829047e-01 5.05794823e-01 3.92252021e-02 -1.29775774e+00
9.96193150e-04 4.03897703e-01 1.99976757e-01 -1.12223089e+00
9.57992021e-03 -4.36342418e-01 -9.59638476e-01 3.77206802e-01
4.37658697e-01 -1.22613996e-01 -2.79227167e-01 1.77469397e+00
3.13010782e-01 -1.32092789e-01 2.86612332e-01 9.89902675e-01
5.31812429e-01 4.15104419e-01 2.89588720e-01 -4.46371406e-01
1.33650029e+00 -5.68092875e-02 -1.13835680e+00 4.14171100e-01
2.99839556e-01 1.44432694e-01 8.57106924e-01 4.90213007e-01
-1.10620201e+00 -6.11870252e-02 -1.34158039e+00 -2.70634562e-01
-2.87787795e-01 -1.36479735e-01 5.13561845e-01 1.34619284e+00
-9.35406685e-01 6.56796992e-01 -7.70020425e-01 2.46665683e-02
7.93227255e-01 3.01827490e-01 -3.52565408e-01 3.82708251e-01
-1.36637604e+00 8.88377964e-01 2.72631019e-01 3.65725398e-01
-1.04122591e+00 -8.09763193e-01 -6.45274818e-01 3.48694533e-01
-1.67586431e-01 -1.07460678e+00 6.68496251e-01 -1.26852846e+00
-1.49570692e+00 1.02327430e+00 2.28706703e-01 -8.17312360e-01
9.63013470e-01 -2.06341073e-01 -3.14500369e-02 1.40972987e-01
-2.12950259e-01 3.50356847e-01 9.62792158e-01 -9.51759815e-01
5.29643744e-02 -9.08354342e-01 -2.74842441e-01 1.66512042e-01
-6.30278170e-01 -1.88605085e-01 5.67523360e-01 -3.46583545e-01
-2.30570629e-01 -7.37677276e-01 5.37813641e-02 7.11998105e-01
-3.92241567e-01 1.14747241e-01 4.39544499e-01 -6.20632589e-01
7.88740933e-01 -2.15149355e+00 2.17018351e-01 5.77242315e-01
6.05798602e-01 9.45525691e-02 1.83221549e-01 -5.14027514e-02
3.81766737e-01 3.78291488e-01 -3.60993624e-01 -4.69150037e-01
4.84184951e-01 -1.26714274e-01 -4.93548438e-02 1.02921510e+00
-2.34387919e-01 7.99078643e-01 -5.84031820e-01 -4.19492930e-01
1.91094398e-01 7.30635047e-01 -4.50076044e-01 1.89429343e-01
2.90120915e-02 3.33034009e-01 -6.82967007e-01 4.23924595e-01
7.40573406e-01 4.55671474e-02 2.56510526e-01 -8.65435824e-02
2.62515247e-01 -3.52929890e-01 -7.88771451e-01 1.42001569e+00
-1.13153486e-02 7.23782361e-01 2.02426672e-01 -8.66471112e-01
1.02520943e+00 4.95324433e-01 3.33447576e-01 -2.29047835e-01
5.99514782e-01 2.64210235e-02 -2.57327378e-01 -3.55718851e-01
1.05223127e-01 -4.43571478e-01 -1.94601744e-01 5.95253885e-01
2.50699837e-02 1.33568227e-01 -7.56200671e-01 1.13361143e-01
5.92766881e-01 -2.31074914e-01 6.24848545e-01 -9.99938428e-01
4.76484120e-01 -4.62560505e-01 5.40447295e-01 7.09011257e-01
-8.01302135e-01 2.14026272e-01 5.53163946e-01 -7.45633781e-01
-8.88068855e-01 -1.10126364e+00 -4.26048368e-01 5.11975646e-01
-9.75116156e-03 -6.59296140e-02 -1.26714849e+00 -4.72739607e-01
1.82654113e-01 1.00803638e+00 -1.04099643e+00 -6.74321890e-01
4.45959747e-01 -7.42179394e-01 7.51306653e-01 -2.33645067e-02
6.92858398e-01 -6.72900975e-01 -8.90241981e-01 -3.67273480e-01
-7.46577978e-02 -7.63965130e-01 -2.11609468e-01 -1.91651270e-01
-7.68441856e-01 -6.21957302e-01 -2.80982435e-01 8.62815455e-02
4.89552349e-01 -4.83037293e-01 8.35381746e-01 -6.92482293e-02
-2.68147111e-01 6.32671058e-01 5.62160760e-02 -9.68417823e-01
-7.58464456e-01 -2.33395815e-01 2.14037091e-01 5.61296821e-01
7.96272337e-01 -3.83712500e-01 -5.79731822e-01 -9.32362825e-02
-8.82570267e-01 -3.80879790e-01 2.90818304e-01 6.64193869e-01
4.70314294e-01 -1.49206027e-01 5.95153153e-01 -8.14078033e-01
9.67408717e-01 -5.56099892e-01 -8.66346538e-01 6.16038144e-01
-1.38698483e+00 2.38605082e-01 3.11131567e-01 5.25453091e-02
-1.18766963e+00 1.24440361e-02 3.76673162e-01 -4.26783800e-01
-3.52753140e-02 2.95217663e-01 -6.05295599e-01 -2.64450610e-01
6.89141273e-01 4.95221429e-02 4.73640919e-01 -1.05933905e-01
4.87054884e-01 8.59336793e-01 2.84602344e-01 -5.45418501e-01
3.61439526e-01 6.09055758e-01 2.82366455e-01 -7.24795640e-01
-5.74034095e-01 1.68013811e-01 -6.85130179e-01 -2.81556070e-01
8.93570900e-01 -5.65893829e-01 -1.25592482e+00 3.65693033e-01
-1.09944761e+00 3.67176175e-01 -7.50227690e-01 6.44655943e-01
-7.40707040e-01 4.61115628e-01 -4.31870103e-01 -1.27787971e+00
-8.87680292e-01 -1.07963681e+00 7.07591891e-01 -1.13982372e-01
-2.96536177e-01 -1.29792058e+00 7.54645541e-02 5.87326825e-01
2.69499570e-01 5.38227797e-01 8.33133459e-01 -1.08813202e+00
-6.83949053e-01 -3.77101541e-01 5.62590808e-02 7.61630833e-01
-2.42471099e-01 -2.07023054e-01 -1.26534414e+00 -1.50063232e-01
5.95998645e-01 -1.40750349e-01 4.34210211e-01 5.29202700e-01
7.74594963e-01 -8.85883272e-01 6.96531087e-02 6.48991108e-01
1.39646232e+00 8.80211070e-02 5.51904380e-01 -1.81033731e-01
3.08950365e-01 8.67147207e-01 3.33459228e-01 8.68784010e-01
1.26159251e-01 2.39022300e-01 3.92208606e-01 3.51905048e-01
6.68074489e-01 -1.18921258e-01 3.49733233e-01 1.91537455e-01
1.05841935e-01 -3.36672366e-01 -5.49968779e-01 2.15970024e-01
-1.79275239e+00 -8.88156414e-01 8.81108344e-02 2.32936239e+00
7.25242198e-01 -3.54392260e-01 7.95491934e-02 -6.93873242e-02
6.26943171e-01 -2.47400686e-01 -7.13174880e-01 -8.21021736e-01
3.91334295e-02 -2.38395080e-01 4.19525832e-01 6.29008412e-01
-9.72110271e-01 4.26361293e-01 6.99491501e+00 4.45837259e-01
-7.73307741e-01 3.09657812e-01 1.15739679e+00 -2.45978329e-02
-8.35443437e-01 -9.36588049e-02 -3.78589571e-01 4.62713897e-01
1.20694578e+00 -5.25056064e-01 4.59734589e-01 6.87736273e-01
2.13979051e-01 1.13891222e-01 -1.48307681e+00 7.53617823e-01
-7.19585642e-02 -1.32234728e+00 6.89303130e-02 2.96572924e-01
3.09421629e-01 -3.54378223e-01 5.01555860e-01 -5.09285390e-01
9.30325538e-02 -1.18878686e+00 8.45491409e-01 9.60325599e-01
5.05248487e-01 -7.58722007e-01 7.68696904e-01 3.25782925e-01
-1.79972231e-01 1.32321496e-03 -4.26249593e-01 7.77215660e-02
-4.62159872e-01 7.87851393e-01 -5.34616232e-01 5.83048522e-01
6.37658894e-01 5.62700272e-01 -8.92413929e-02 6.02881670e-01
1.10580027e-01 4.44599301e-01 -3.47186655e-01 -3.14705104e-01
-2.46642172e-01 -5.34894526e-01 8.63865554e-01 8.64326894e-01
8.39741379e-02 1.18576430e-01 -5.44614017e-01 1.76869631e+00
-8.58906358e-02 2.36751229e-01 -8.78969729e-01 2.04776134e-02
5.80410004e-01 1.02074707e+00 -3.54817986e-01 5.83568774e-02
8.39523897e-02 8.03358495e-01 -6.36910498e-02 7.09894150e-02
-5.80605209e-01 -6.79404214e-02 5.95141768e-01 -6.86704963e-02
-3.16517800e-01 2.43860647e-01 -8.16181481e-01 -1.16536546e+00
7.17092529e-02 -8.42310369e-01 6.09385014e-01 -2.43055463e-01
-1.43869066e+00 4.97678876e-01 5.26808977e-01 -9.13483799e-01
-1.23348042e-01 -3.63628924e-01 -3.54521453e-01 9.47248042e-01
-1.02318871e+00 -1.10138369e+00 1.28465548e-01 6.39522612e-01
-3.32866967e-01 -5.15253544e-01 1.08016992e+00 -8.00694823e-02
-5.47579885e-01 1.13759863e+00 3.77356648e-01 -1.75657168e-01
5.13140820e-02 -1.01018643e+00 -1.77421510e-01 9.02838826e-01
9.70345661e-02 7.10082889e-01 8.30420792e-01 -3.60792965e-01
-1.41328084e+00 -8.78230155e-01 7.74047315e-01 -5.74987352e-01
2.65837729e-01 -4.14289951e-01 -7.72653282e-01 7.69437611e-01
2.86905646e-01 1.46830887e-01 1.14966297e+00 -8.06302726e-02
-3.94640207e-01 -4.68921304e-01 -2.04168320e+00 1.40260115e-01
4.25185978e-01 -6.70371294e-01 -4.52677190e-01 3.03795993e-01
6.20095491e-01 2.00253665e-01 -1.27227497e+00 -5.16652949e-02
6.99831307e-01 -1.00829899e+00 7.03992069e-01 -6.10600412e-01
-6.98785186e-02 7.94691369e-02 -1.89548627e-01 -1.03059351e+00
-9.46889147e-02 -1.18830705e+00 -5.09757325e-02 1.06529152e+00
2.34681308e-01 -9.43823576e-01 7.67439127e-01 1.51795495e+00
5.11229098e-01 -3.75484496e-01 -1.39629090e+00 -5.88292837e-01
3.89603496e-01 -5.00806868e-02 6.58381402e-01 9.75785673e-01
3.55569690e-01 -7.01041222e-02 -5.86283147e-01 1.64809078e-01
1.38157117e+00 -2.87894279e-01 1.47085086e-01 -1.48480785e+00
5.58223501e-02 4.98632714e-02 -7.49175251e-01 -6.16529509e-02
4.94026065e-01 -9.49475646e-01 -1.02243662e-01 -5.70577323e-01
3.99643809e-01 -1.27464429e-01 -5.94610155e-01 2.18837306e-01
2.34483331e-01 -2.10721374e-01 1.93158701e-01 8.21522772e-02
-3.88024122e-01 7.25170612e-01 6.69925213e-01 -1.45748094e-01
1.25072584e-01 2.22420126e-01 -8.95816505e-01 5.39422035e-01
7.32495189e-01 -7.40782022e-01 -9.18712318e-01 -2.38138780e-01
1.49078414e-01 2.52041221e-01 9.36980784e-01 -7.48115659e-01
1.69262126e-01 -9.22043100e-02 1.85136452e-01 5.26144952e-02
3.59299332e-01 -1.23097432e+00 4.67788517e-01 6.38015151e-01
-1.14359927e+00 -4.42656070e-01 -9.30454135e-02 8.58472228e-01
9.03291553e-02 -2.01796606e-01 9.62111831e-01 -2.03010384e-02
-1.17605301e-02 4.31240618e-01 -3.52318585e-01 -5.10959141e-02
1.19820857e+00 -5.92471808e-02 -1.14702083e-01 -5.17010510e-01
-8.28106344e-01 -1.96793407e-01 8.19753185e-02 -4.68332954e-02
9.46286380e-01 -1.05086458e+00 -1.10461569e+00 3.72190028e-01
-1.54351369e-02 -5.05753875e-01 6.09752499e-02 5.76335907e-01
-4.05438662e-01 4.76014048e-01 -5.39896607e-01 -3.42563778e-01
-1.09358847e+00 4.09984529e-01 6.65564418e-01 8.51868540e-02
-3.53974998e-01 5.25735438e-01 6.25875890e-02 -1.71957940e-01
4.35144395e-01 -2.06541285e-01 1.75876319e-01 -2.66166717e-01
4.50617373e-01 5.48779011e-01 -9.27983448e-02 -4.36618686e-01
-4.23117757e-01 -3.55046727e-02 9.10373703e-02 -3.61340195e-01
1.01604021e+00 -3.52632076e-01 -4.76413518e-01 4.35235381e-01
1.33399475e+00 -5.65615654e-01 -1.04346776e+00 -1.70044750e-01
1.05798021e-01 -5.02510488e-01 3.22564095e-01 -8.00894797e-01
-1.04162240e+00 1.12513912e+00 1.02121019e+00 2.61168659e-01
9.85996425e-01 -3.87840152e-01 3.96625847e-01 5.78315556e-01
3.09347987e-01 -9.47101176e-01 -4.86383021e-01 -3.14785630e-01
8.64513397e-01 -1.07651925e+00 2.68088430e-01 -1.55447841e-01
-7.25754321e-01 1.00709653e+00 1.19948186e-01 3.14882919e-02
1.10422206e+00 1.81559592e-01 -1.65228069e-01 -1.74060822e-01
-5.43804944e-01 7.82633901e-01 4.86688286e-01 7.16009796e-01
1.45767117e-02 4.89165485e-01 -5.71470737e-01 9.92650509e-01
1.03692681e-01 2.61901081e-01 3.85774106e-01 4.32470322e-01
-7.74773657e-02 -6.85832500e-01 -3.75757553e-02 7.39226714e-02
-8.17652583e-01 -9.41941813e-02 -6.88460231e-01 5.83102763e-01
-1.93174109e-01 8.33417535e-01 -4.93610166e-02 -4.30333436e-01
-1.62726596e-01 1.04539350e-01 2.24429056e-01 1.67444702e-02
-7.69666910e-01 -5.19147098e-01 -1.05009265e-01 -5.15658379e-01
-5.15785277e-01 -7.41799057e-01 -8.15237224e-01 -7.08074033e-01
-7.78985918e-02 3.62785161e-01 6.45015001e-01 9.16800559e-01
6.30055785e-01 -2.15633288e-01 5.90963721e-01 -5.47218658e-02
-1.17253709e+00 -3.74199927e-01 -8.53255868e-01 2.60194331e-01
4.26814407e-01 -1.46380141e-01 -5.58691800e-01 -1.11959197e-01] | [6.303143501281738, 6.625558376312256] |
463c9219-1a83-4f1f-9915-1c522976bd8a | coarse-to-fine-classification-via-parametric | 1405.4308 | null | http://arxiv.org/abs/1405.4308v1 | http://arxiv.org/pdf/1405.4308v1.pdf | Coarse-to-Fine Classification via Parametric and Nonparametric Models for Computer-Aided Diagnosis | Classification is one of the core problems in Computer-Aided Diagnosis (CAD),
targeting for early cancer detection using 3D medical imaging interpretation.
High detection sensitivity with desirably low false positive (FP) rate is
critical for a CAD system to be accepted as a valuable or even indispensable
tool in radiologists' workflow. Given various spurious imagery noises which
cause observation uncertainties, this remains a very challenging task. In this
paper, we propose a novel, two-tiered coarse-to-fine (CTF) classification
cascade framework to tackle this problem. We first obtain
classification-critical data samples (e.g., samples on the decision boundary)
extracted from the holistic data distributions using a robust parametric model
(e.g., \cite{Raykar08}); then we build a graph-embedding based nonparametric
classifier on sampled data, which can more accurately preserve or formulate the
complex classification boundary. These two steps can also be considered as
effective "sample pruning" and "feature pursuing + $k$NN/template matching",
respectively. Our approach is validated comprehensively in colorectal polyp
detection and lung nodule detection CAD systems, as the top two deadly cancers,
using hospital scale, multi-site clinical datasets. The results show that our
method achieves overall better classification/detection performance than
existing state-of-the-art algorithms using single-layer classifiers, such as
the support vector machine variants \cite{Wang08}, boosting \cite{Slabaugh10},
logistic regression \cite{Ravesteijn10}, relevance vector machine
\cite{Raykar08}, $k$-nearest neighbor \cite{Murphy09} or spectral projections
on graph \cite{Cai08}. | ['Xiaojing Ye', 'Shipeng Yu', 'Meizhu Liu', 'Le Lu'] | 2014-05-16 | null | null | null | null | ['lung-nodule-detection'] | ['medical'] | [ 4.28581864e-01 2.53971100e-01 -4.20818508e-01 -1.15729935e-01
-1.04691970e+00 -2.69225240e-01 3.07479233e-01 4.59265590e-01
-1.01469215e-02 4.84606534e-01 -6.71049729e-02 -7.53542304e-01
-7.96721637e-01 -7.51924694e-01 -3.94921035e-01 -9.46008444e-01
-1.13097243e-01 5.20112157e-01 3.99147063e-01 1.48631819e-02
1.95224911e-01 6.14767075e-01 -1.36776006e+00 3.62043649e-01
7.14494228e-01 1.20566523e+00 1.22954957e-01 7.28558481e-01
3.11781373e-02 5.82551837e-01 -2.13347580e-02 -2.57241756e-01
8.36131871e-02 -2.35583782e-01 -6.36601627e-01 1.09862061e-02
1.01405106e-01 -1.28699645e-01 -1.96023107e-01 1.19193590e+00
6.60107255e-01 -3.22089553e-01 1.30757868e+00 -6.81282520e-01
-5.37643969e-01 2.05292985e-01 -7.58049846e-01 3.13761175e-01
1.97195485e-01 -9.18526873e-02 6.58341348e-01 -1.11498141e+00
3.79632592e-01 8.32583547e-01 1.05316448e+00 4.31620538e-01
-1.11714864e+00 -5.21293342e-01 -2.83979207e-01 -1.83809604e-02
-1.33851290e+00 -2.43812688e-02 7.05189049e-01 -7.74571061e-01
4.31531936e-01 7.00419068e-01 7.19000399e-01 9.33372319e-01
6.72766745e-01 4.89126384e-01 1.12989938e+00 -6.03259444e-01
7.38638863e-02 4.43096817e-01 3.10264975e-01 1.13331270e+00
7.14224637e-01 3.17809671e-01 -1.08968884e-01 -5.88855863e-01
9.16181684e-01 1.90442532e-01 -2.96507120e-01 -4.98451024e-01
-1.07238352e+00 9.76409853e-01 5.74469030e-01 1.40039951e-01
-3.39513630e-01 -1.78668499e-01 3.67761225e-01 3.33877318e-02
4.34471935e-01 1.21280923e-01 -2.33935282e-01 4.28067774e-01
-7.95746624e-01 -4.27373461e-02 7.57934988e-01 8.41100693e-01
4.24920946e-01 -1.75185308e-01 -4.33942199e-01 8.09152544e-01
5.15426099e-01 2.83524781e-01 6.39586091e-01 -3.15394580e-01
1.16536999e-02 7.04842865e-01 -1.65135846e-01 -1.05720997e+00
-7.29897797e-01 -8.38392913e-01 -1.42438376e+00 8.11824501e-02
2.92475402e-01 2.58301079e-01 -8.30940783e-01 9.71439660e-01
5.50871491e-01 2.18495920e-01 -1.05283566e-01 6.22963488e-01
1.03402269e+00 7.66371489e-02 1.24982707e-01 -3.00880462e-01
1.68271446e+00 -6.51416600e-01 -2.11603805e-01 1.52236715e-01
7.50734448e-01 -7.87341356e-01 8.67392421e-01 4.65227365e-01
-5.73977709e-01 -5.85255325e-01 -1.05556834e+00 1.55927449e-01
-2.12961823e-01 7.82710433e-01 7.16274142e-01 9.94307399e-01
-8.33136201e-01 3.34129363e-01 -9.02035773e-01 -3.62024367e-01
6.87010586e-01 3.91277313e-01 -2.66458631e-01 -5.03091574e-01
-6.66960776e-01 7.57746518e-01 8.79814103e-02 1.85153931e-01
-8.44861627e-01 -9.26852822e-01 -7.13018239e-01 -3.45530480e-01
3.62046927e-01 -8.01155746e-01 7.60415018e-01 -4.61763591e-01
-1.08459532e+00 1.17714322e+00 7.64955021e-03 -3.53071988e-01
7.69362926e-01 1.65114909e-01 -3.61357838e-01 1.88086689e-01
1.09822981e-01 3.02526653e-01 9.04586375e-01 -1.18168163e+00
-4.74005103e-01 -6.11304760e-01 -3.94351304e-01 -5.84503412e-02
-2.13873073e-01 -2.39355475e-01 -2.38925770e-01 -7.95527756e-01
4.59507793e-01 -9.84860182e-01 -5.28800189e-01 2.96928436e-01
-5.48203051e-01 -2.65881926e-01 7.18261778e-01 -8.34904313e-01
1.29631472e+00 -2.10359383e+00 -7.63384849e-02 6.48979723e-01
5.38458288e-01 1.99497879e-01 4.55079973e-01 8.83590896e-03
-1.63634509e-01 1.47208095e-01 -2.85716921e-01 1.10876806e-01
-3.18759948e-01 -6.02304609e-03 3.81029248e-01 8.10796797e-01
2.77551264e-01 9.62731421e-01 -8.34344089e-01 -7.99365282e-01
4.51085508e-01 3.73458207e-01 -3.57080191e-01 8.19462016e-02
7.59612843e-02 3.66290212e-01 -6.66913748e-01 1.04846525e+00
6.43459022e-01 -5.24172187e-01 1.05223946e-01 -6.48373425e-01
1.98579729e-01 -4.71286595e-01 -1.23028982e+00 1.30581450e+00
-3.17781329e-01 1.44153491e-01 1.50684696e-02 -1.29067075e+00
1.05239844e+00 3.41004461e-01 4.61419851e-01 -1.23345204e-01
2.80981421e-01 6.32744193e-01 -2.34937798e-02 -6.02536261e-01
-1.76639110e-01 -1.65923387e-01 -1.83326646e-03 -2.24926993e-01
-1.89540848e-01 -2.75590986e-01 -2.84348398e-01 -1.93479970e-01
1.11103511e+00 -1.92574009e-01 9.72810507e-01 -7.15899885e-01
8.88896286e-01 3.29239279e-01 1.71426117e-01 8.14787686e-01
-3.25150818e-01 7.08730578e-01 4.78502333e-01 -2.75890350e-01
-6.38381720e-01 -1.09176981e+00 -6.18690968e-01 4.65364873e-01
-8.87467861e-02 2.96584785e-01 -4.47621375e-01 -8.92798424e-01
2.20634520e-01 3.26353431e-01 -7.93678105e-01 -3.69686157e-01
-4.14445370e-01 -1.03973711e+00 4.71157670e-01 4.85165745e-01
4.12112802e-01 -5.59526324e-01 -5.48243701e-01 1.79075807e-01
2.60137945e-01 -6.72491848e-01 -1.32741764e-01 4.67609763e-01
-1.04992878e+00 -1.53837180e+00 -8.84702742e-01 -9.55282927e-01
8.53703737e-01 1.67851895e-01 9.72783744e-01 1.80285215e-01
-1.08132303e+00 4.16327894e-01 -3.97760570e-01 -5.07154167e-01
-5.04555464e-01 -6.84227720e-02 -2.31056899e-01 -4.72883023e-02
4.10809249e-01 -2.24016413e-01 -7.74198651e-01 2.29577452e-01
-6.82946622e-01 -2.45446607e-01 1.20018685e+00 1.40497279e+00
1.14845610e+00 1.41823098e-01 6.24045849e-01 -1.22382271e+00
4.70565915e-01 -5.94876766e-01 -4.36233640e-01 2.54790276e-01
-5.59872150e-01 -1.37738824e-01 6.27500236e-01 -3.19008589e-01
-7.25765467e-01 1.41702026e-01 -1.53190926e-01 -5.25142372e-01
-2.27524683e-01 6.07425928e-01 1.40508823e-02 -4.60387737e-01
1.12052441e+00 3.35480332e-01 1.48213357e-01 -2.57392913e-01
-1.10151283e-01 7.04634845e-01 3.14319611e-01 -2.16822639e-01
7.75311649e-01 4.53925192e-01 6.36913180e-01 -9.13903117e-01
-7.28441238e-01 -9.53046262e-01 -6.06352270e-01 1.35385999e-02
8.09067428e-01 -8.43585014e-01 -3.36987555e-01 1.48046151e-01
-7.01074362e-01 1.95570841e-01 -2.99533606e-01 7.61315584e-01
-3.02051127e-01 4.58783329e-01 -4.18181330e-01 -8.42026770e-01
-4.23685700e-01 -1.19527304e+00 1.17233026e+00 8.69736820e-02
-1.24864802e-01 -1.10070944e+00 -1.52741760e-01 2.83018351e-01
2.89752662e-01 6.58867836e-01 1.18832958e+00 -7.91364789e-01
-4.17954803e-01 -5.99817872e-01 -5.20169735e-01 4.28156286e-01
2.64476866e-01 -4.26232209e-03 -8.71647000e-01 -3.70696574e-01
9.86314416e-02 -5.40720634e-02 7.74744928e-01 7.69623578e-01
1.49718535e+00 -2.15147492e-02 -8.81072760e-01 5.52753448e-01
1.61779714e+00 8.97568464e-02 4.31862950e-01 2.90059485e-03
5.84502876e-01 3.72974902e-01 6.85195267e-01 4.20391858e-01
3.20886299e-02 1.98279187e-01 4.04701620e-01 -2.04689175e-01
-3.36089253e-01 -7.17405677e-02 -2.51181871e-01 6.96826160e-01
-9.32125673e-02 8.56688544e-02 -8.91052604e-01 4.85276371e-01
-1.43608534e+00 -5.76187432e-01 -5.89868128e-01 2.20333457e+00
5.09070933e-01 1.64524600e-01 -1.63354874e-01 3.80328447e-01
6.19484484e-01 -2.21440077e-01 -4.10531074e-01 -3.89508940e-02
1.49035469e-01 5.73127389e-01 6.10405624e-01 2.18704537e-01
-1.45610631e+00 2.17027470e-01 4.52797508e+00 1.17352653e+00
-1.01310933e+00 2.15669304e-01 9.73678529e-01 4.91116434e-01
-8.47306922e-02 -3.07139844e-01 -7.86912024e-01 2.77215362e-01
6.68765187e-01 2.97180057e-01 -2.14562863e-01 9.66992021e-01
1.52541548e-01 -2.31542289e-01 -1.10998464e+00 1.02631593e+00
1.35920107e-01 -1.48238635e+00 2.51094922e-02 1.22086935e-01
4.67315555e-01 -3.08678091e-01 1.51932120e-01 2.41970271e-01
-8.31836089e-02 -1.28701663e+00 9.84326527e-02 5.27427197e-01
1.30090630e+00 -4.39918727e-01 9.96478975e-01 5.41339636e-01
-1.30597973e+00 4.27907668e-02 -3.34597766e-01 5.29222012e-01
-2.38121316e-01 1.02709699e+00 -1.06070769e+00 1.12965727e+00
7.30491817e-01 6.19804263e-01 -7.82776415e-01 1.08383524e+00
1.28085136e-01 6.64763570e-01 -1.74000353e-01 -3.57234359e-01
-1.10439047e-01 5.92209101e-02 3.85298043e-01 1.25876057e+00
5.69203496e-01 3.69084418e-01 2.74964929e-01 4.62536037e-01
2.28774741e-01 3.09171855e-01 -6.62325144e-01 2.84866273e-01
2.87793636e-01 1.36101234e+00 -9.37610447e-01 -2.10459545e-01
-3.41088653e-01 6.50156736e-01 -5.35444655e-02 -7.62537122e-02
-7.68331051e-01 -2.19486997e-01 -9.00428817e-02 5.04587471e-01
2.38870069e-01 1.13975987e-01 -4.07352775e-01 -1.01351655e+00
6.19100779e-02 -7.68121302e-01 6.29375339e-01 -2.02981517e-01
-1.51727211e+00 4.48729813e-01 2.58822404e-02 -1.54544926e+00
1.01715460e-01 -1.14407849e+00 -4.51535016e-01 7.84841418e-01
-1.61500275e+00 -1.29658842e+00 -6.24350011e-01 6.87383652e-01
4.02277142e-01 -1.37494549e-01 1.06233549e+00 3.58157694e-01
-2.95228243e-01 7.57719636e-01 2.14200005e-01 7.60786533e-02
4.73659694e-01 -1.23475325e+00 -5.35770178e-01 3.74887705e-01
-1.67707965e-01 8.75694901e-02 2.80590713e-01 -6.89350247e-01
-1.51180339e+00 -1.45054674e+00 3.89873534e-01 -3.28838080e-01
3.86169225e-01 4.40988727e-02 -7.75738120e-01 4.51527148e-01
-5.10956228e-01 5.30747473e-01 8.03175151e-01 -4.21743780e-01
5.84542006e-02 -1.55974865e-01 -1.51311469e+00 2.10585624e-01
9.32315886e-01 -1.54014856e-01 -3.65751952e-01 5.74346185e-01
3.14917147e-01 -5.36566734e-01 -1.04363966e+00 8.37089598e-01
4.41594809e-01 -9.93310750e-01 1.25831258e+00 -4.19310391e-01
1.42014936e-01 -1.38859749e-01 -2.50783175e-01 -8.58261764e-01
-4.81247962e-01 -1.02336124e-01 -2.96933968e-02 7.81863511e-01
5.62160790e-01 -6.52050674e-01 1.03330696e+00 3.24968882e-02
-3.46605420e-01 -1.10336733e+00 -1.16826606e+00 -5.52252710e-01
1.58266366e-01 -2.62482613e-01 -8.58326927e-02 8.64079535e-01
-3.98809075e-01 -1.80296108e-01 2.99469680e-02 5.25644600e-01
8.17199647e-01 -2.71625400e-01 4.80492622e-01 -1.48718250e+00
-5.51172256e-01 -4.70688075e-01 -6.58312023e-01 -6.39501929e-01
-4.69969988e-01 -1.07496119e+00 -2.27103367e-01 -1.56237018e+00
2.02336729e-01 -7.15780497e-01 -2.86504537e-01 4.89741052e-03
-1.76304892e-01 1.48154557e-01 -5.94804227e-01 2.72879779e-01
-2.20795386e-02 6.48062900e-02 1.20755875e+00 -2.31597811e-01
-8.37995950e-03 4.35970962e-01 -7.15716898e-01 7.11484075e-01
4.41889346e-01 -3.67488474e-01 -2.70632535e-01 2.04998836e-01
-1.04877941e-01 3.41916054e-01 8.46130610e-01 -1.02561879e+00
2.50472307e-01 -1.97492894e-02 7.77141392e-01 -5.99218547e-01
2.48357415e-01 -8.30809474e-01 2.04665914e-01 1.01230216e+00
4.29275781e-02 -2.06500396e-01 1.17653891e-01 9.60344374e-01
-2.24314004e-01 -3.48287135e-01 1.05528915e+00 -3.84504735e-01
-3.97830576e-01 4.58749563e-01 -7.21791014e-02 -1.49274096e-01
1.40329671e+00 -3.57978672e-01 -1.06105357e-01 2.23953992e-01
-9.16590810e-01 5.83559424e-02 -1.66048512e-01 -4.00267728e-02
7.05577254e-01 -1.13881278e+00 -9.26634669e-01 2.96284735e-01
3.40516806e-01 3.99198472e-01 4.00972605e-01 1.45661592e+00
-6.59833491e-01 2.72582322e-01 3.45014393e-01 -1.01976502e+00
-1.24860120e+00 4.86219019e-01 4.86008167e-01 -7.01138437e-01
-3.58670115e-01 1.16497624e+00 -8.14591199e-02 -3.64499152e-01
-5.32357348e-03 -7.04112470e-01 -2.55172193e-01 9.49471220e-02
1.21173672e-01 5.35519898e-01 4.19419825e-01 -2.90746748e-01
-5.03139973e-01 8.00239205e-01 -1.62383452e-01 5.44491172e-01
1.07767117e+00 1.93130478e-01 1.71629190e-02 1.75979450e-01
1.26449001e+00 -7.63562098e-02 -6.57843053e-01 -1.51558489e-01
-4.98660468e-02 -3.13284457e-01 4.81488109e-02 -7.22952187e-01
-6.75238371e-01 1.00341225e+00 1.01210427e+00 2.30281949e-01
1.06922042e+00 1.89730734e-01 2.24933192e-01 1.06454335e-01
2.66740948e-01 -7.20760107e-01 4.62075472e-02 -1.75980344e-01
8.84819627e-01 -1.54707921e+00 3.40127349e-01 -8.49536896e-01
-3.00957322e-01 1.42589247e+00 3.20296079e-01 -2.94435561e-01
1.23055708e+00 -6.63373526e-03 -1.46448925e-01 -2.60380238e-01
-3.39029461e-01 3.09018921e-02 6.08823419e-01 6.83898032e-01
2.61526287e-01 3.69515896e-01 -4.13235873e-01 9.16981876e-01
7.46413320e-02 9.68641639e-02 1.60057858e-01 8.97471070e-01
-5.17758191e-01 -7.95997739e-01 -5.18143475e-01 1.24873090e+00
-4.23411876e-01 -9.38541070e-02 2.01703291e-02 1.16876221e+00
3.42586815e-01 5.60656786e-01 -4.57165003e-01 -2.47788444e-01
5.15207291e-01 -8.79736394e-02 3.81438076e-01 -7.83402383e-01
-5.63992321e-01 2.56302595e-01 -2.10612919e-03 -1.93618909e-01
-3.43928844e-01 -7.85250008e-01 -1.00278342e+00 1.72842681e-01
-6.08555496e-01 -1.75993338e-01 3.99847984e-01 7.15999246e-01
6.90968409e-02 7.22989738e-01 6.14424884e-01 -5.38456023e-01
-1.14049888e+00 -8.06573808e-01 -5.15458047e-01 3.06848660e-02
2.10513502e-01 -7.19411075e-01 -5.77325940e-01 -2.58019138e-02] | [15.133688926696777, -2.3720474243164062] |
cf668266-7b88-4cad-9e26-43a277739074 | disclda-discriminative-learning-for | null | null | http://papers.nips.cc/paper/3599-disclda-discriminative-learning-for-dimensionality-reduction-and-classification | http://papers.nips.cc/paper/3599-disclda-discriminative-learning-for-dimensionality-reduction-and-classification.pdf | DiscLDA: Discriminative Learning for Dimensionality Reduction and Classification | Probabilistic topic models (and their extensions) have become popular as models of latent structures in collections of text documents or images. These models are usually treated as generative models and trained using maximum likelihood estimation, an approach which may be suboptimal in the context of an overall classification problem. In this paper, we describe DiscLDA, a discriminative learning framework for such models as Latent Dirichlet Allocation (LDA) in the setting of dimensionality reduction with supervised side information. In DiscLDA, a class-dependent linear transformation is introduced on the topic mixture proportions. This parameter is estimated by maximizing the conditional likelihood using Monte Carlo EM. By using the transformed topic mixture proportions as a new representation of documents, we obtain a supervised dimensionality reduction algorithm that uncovers the latent structure in a document collection while preserving predictive power for the task of classification. We compare the predictive power of the latent structure of DiscLDA with unsupervised LDA on the 20 Newsgroup ocument classification task. | ['Michael. I. Jordan', 'Simon Lacoste-Julien', 'Fei Sha'] | 2008-12-01 | null | null | null | neurips-2008-12 | ['supervised-dimensionality-reduction'] | ['computer-vision'] | [ 1.43525422e-01 1.95920482e-01 -4.03071225e-01 -4.25651789e-01
-8.40034902e-01 -5.34832299e-01 1.21539187e+00 -6.93380758e-02
-1.03033185e-01 2.80537963e-01 6.74348474e-01 1.13952674e-01
-2.46974543e-01 -5.55920601e-01 -1.80746302e-01 -1.21958673e+00
2.03705892e-01 8.47694218e-01 -1.15715392e-01 4.74779427e-01
3.56008500e-01 2.71098554e-01 -1.44937849e+00 2.87884295e-01
5.86047173e-01 4.92512077e-01 1.56134993e-01 4.40609843e-01
-1.71829224e-01 4.65225577e-01 -4.48340714e-01 -1.18720777e-01
-2.78342422e-02 -2.30915740e-01 -6.90743148e-01 8.51490319e-01
2.95315892e-01 -1.70222655e-01 -2.48328224e-01 8.40772271e-01
1.41374812e-01 8.20467398e-02 1.55390072e+00 -1.31376350e+00
-1.59878135e-01 5.09621084e-01 -7.17611730e-01 -4.49083857e-02
5.69868796e-02 -5.44238567e-01 1.13189018e+00 -1.02262175e+00
7.59937048e-01 1.65238762e+00 3.91286403e-01 7.53600225e-02
-1.85152173e+00 -3.40577036e-01 5.21555590e-03 -2.92841461e-03
-1.42117321e+00 -4.02594715e-01 9.85404789e-01 -9.91844475e-01
5.00188410e-01 -4.59768102e-02 3.39927614e-01 1.07217038e+00
3.81013662e-01 1.21596003e+00 9.35142696e-01 -6.94686413e-01
5.90994596e-01 4.10736740e-01 3.74715567e-01 2.97908962e-01
1.61315426e-01 -4.62020189e-01 -6.37781084e-01 -9.46740806e-01
4.06256825e-01 3.47541049e-02 4.97735627e-02 -1.22739780e+00
-8.24109018e-01 1.45864284e+00 -2.27324560e-01 2.66035408e-01
-4.34249789e-01 7.93063492e-02 2.58015841e-01 -8.99627879e-02
1.13357103e+00 1.27474174e-01 7.00421538e-03 8.27958509e-02
-1.26450956e+00 3.86296093e-01 9.16377187e-01 6.31623328e-01
7.64868736e-01 -2.93186486e-01 -1.76956758e-01 9.43339705e-01
7.78273225e-01 3.50605130e-01 6.52625740e-01 -8.01312864e-01
7.36759901e-02 4.63745892e-01 5.06486557e-03 -8.08021605e-01
-1.40411109e-01 -3.11079025e-01 -6.07800066e-01 -1.15983337e-01
1.15006752e-01 1.69695914e-01 -8.26630592e-01 1.61727202e+00
4.11212504e-01 -1.53055012e-01 -1.48398597e-02 3.21050704e-01
1.67577535e-01 9.15669858e-01 1.84070140e-01 -4.80150908e-01
1.17980075e+00 -6.99981749e-01 -1.00519300e+00 4.79768850e-02
7.39113510e-01 -8.16504359e-01 6.66467309e-01 6.27096593e-01
-7.72264063e-01 -1.73039272e-01 -7.36060262e-01 -4.26482223e-02
-1.92562386e-01 3.66079301e-01 5.67733705e-01 8.17543685e-01
-8.32744837e-01 2.24491671e-01 -1.10419762e+00 -3.78087640e-01
3.87781501e-01 1.60050809e-01 -1.23608299e-01 -2.91239798e-01
-6.05251789e-01 4.95565772e-01 4.08023149e-01 -5.04732251e-01
-1.04774785e+00 -4.49368566e-01 -8.04826796e-01 1.08546846e-01
1.49075761e-01 -3.42627257e-01 1.07398891e+00 -3.27447772e-01
-1.24661338e+00 9.52407002e-01 -5.86356878e-01 -4.09355581e-01
3.28450620e-01 -2.99121082e-01 5.45170866e-02 2.01180533e-01
2.90549308e-01 5.86760581e-01 1.50112820e+00 -1.28731978e+00
-4.02119428e-01 -6.54635370e-01 -6.03162825e-01 1.93294287e-01
-5.16219020e-01 -4.79421169e-02 -3.88137937e-01 -7.60624170e-01
6.24487221e-01 -1.18219316e+00 -5.51125258e-02 -2.60751814e-01
-3.91162753e-01 -7.10669637e-01 1.27838552e+00 -6.31417155e-01
1.08395779e+00 -2.31405044e+00 3.21413845e-01 3.02349418e-01
3.91739547e-01 -1.67650744e-01 1.73340052e-01 4.38940734e-01
-1.09605961e-01 -1.28186136e-01 -3.78287941e-01 -8.91772628e-01
2.01489344e-01 1.65035918e-01 -7.28867233e-01 7.99057424e-01
-1.80862352e-01 5.38957655e-01 -5.24395287e-01 -5.46786547e-01
1.02419712e-01 5.88319719e-01 -6.82820082e-01 3.02897915e-02
-3.20847362e-01 1.21310152e-01 -5.19068837e-01 2.11234108e-01
7.02718794e-01 -3.11939001e-01 4.43030089e-01 7.63725489e-02
6.79277033e-02 3.14250648e-01 -1.06190193e+00 1.71574306e+00
-2.59722881e-02 9.48091567e-01 -1.99383512e-01 -9.09425676e-01
9.96631384e-01 4.94371742e-01 7.69211709e-01 2.12484986e-01
-6.77306503e-02 -2.41023496e-01 -4.33715373e-01 -2.31162131e-01
4.90839869e-01 -1.41726330e-01 -2.22505167e-01 9.88602817e-01
2.64120638e-01 -1.41240805e-01 1.94668949e-01 5.12351334e-01
6.59971595e-01 -1.23582175e-03 3.40384513e-01 -4.05969739e-01
-1.46639384e-02 -2.52535241e-03 3.17055844e-02 7.88278401e-01
3.37456077e-01 4.81999904e-01 7.04757750e-01 -1.59325123e-01
-1.29924560e+00 -1.10697401e+00 -6.52155340e-01 1.07949972e+00
-5.40143847e-01 -6.55447423e-01 -7.36364007e-01 -5.67086756e-01
-1.37711197e-01 8.50382030e-01 -5.09586632e-01 -2.56758749e-01
-2.02122867e-01 -1.42579591e+00 1.59376532e-01 2.36628484e-02
1.73547685e-01 -4.43399638e-01 -1.98739707e-01 2.40821883e-01
-1.80651903e-01 -9.51956570e-01 -6.18178844e-02 2.27350846e-01
-1.04999447e+00 -8.17284107e-01 -8.99570405e-01 -4.93815482e-01
7.57082939e-01 1.79360718e-01 7.22595632e-01 -8.88272941e-01
-2.07757428e-01 5.88347077e-01 -1.68339044e-01 -1.97808161e-01
-5.89104652e-01 1.64849043e-01 2.76901484e-01 3.01363915e-01
7.32883215e-01 -5.70962310e-01 -9.34726968e-02 2.29768232e-01
-1.09261501e+00 1.66700743e-02 4.02076244e-01 7.12094009e-01
3.69998038e-01 3.63178343e-01 2.11480290e-01 -9.53347445e-01
8.11956584e-01 -5.97746909e-01 -5.54473639e-01 1.47740141e-01
-7.06840932e-01 2.31329292e-01 -2.65399605e-01 -5.05628765e-01
-1.29181623e+00 9.14446563e-02 4.96730477e-01 -5.07456422e-01
-4.13735241e-01 5.05964816e-01 -2.83827841e-01 3.73245895e-01
5.79770923e-01 4.02836859e-01 2.98930220e-02 -9.62010026e-01
4.55973893e-01 9.30114925e-01 1.77086115e-01 -5.55839956e-01
4.63942349e-01 8.57819796e-01 3.22561376e-02 -1.00044644e+00
-1.07224488e+00 -9.11623180e-01 -1.00680757e+00 1.29435271e-01
9.01281953e-01 -1.00848424e+00 -1.10897003e-02 4.07025874e-01
-1.00159037e+00 1.17300153e-01 -2.89729774e-01 7.59869516e-01
-7.82432556e-01 5.12427211e-01 -2.96299666e-01 -8.24675858e-01
1.32424921e-01 -9.00460780e-01 1.34920609e+00 -2.67130911e-01
-4.74272668e-01 -1.32504845e+00 4.85975146e-01 3.11180800e-01
-8.86175260e-02 -2.49283090e-02 1.21400809e+00 -9.02732611e-01
-3.00625205e-01 -6.22642815e-01 1.06937708e-02 3.01029831e-01
3.19540232e-01 -1.40851572e-01 -1.18884003e+00 -4.33931321e-01
4.13467854e-01 -4.72447230e-03 1.06761789e+00 7.14528024e-01
9.77046251e-01 -4.41773832e-01 -7.62091279e-01 2.60011971e-01
1.04840255e+00 1.47459097e-02 3.70167226e-01 6.78986907e-02
5.63685656e-01 6.89539015e-01 2.93291211e-01 7.31652260e-01
1.75415322e-01 7.28040516e-01 -1.79807708e-01 2.51269639e-01
1.23865709e-01 -2.51004696e-01 2.86442250e-01 6.82712138e-01
3.52819204e-01 -3.65493447e-01 -1.15956378e+00 5.98557472e-01
-1.76155841e+00 -6.07237935e-01 -4.34635431e-02 1.98783517e+00
7.75460601e-01 -9.96409357e-02 6.07219338e-02 1.56474516e-01
8.72448206e-01 2.12147251e-01 -2.93710947e-01 3.56966071e-02
-2.78824549e-02 -3.16785246e-01 7.95023069e-02 3.90347391e-01
-1.31101263e+00 7.69066691e-01 6.97354603e+00 1.07843018e+00
-6.10398173e-01 2.54268616e-01 3.77719015e-01 -7.75410384e-02
-1.70091391e-01 3.27284843e-01 -1.19929993e+00 5.15596449e-01
7.87777543e-01 -1.72464803e-01 -1.82441056e-01 1.04497612e+00
3.17914158e-01 -3.90068471e-01 -1.26704526e+00 8.85795593e-01
4.42668766e-01 -1.04609919e+00 3.80447626e-01 8.04861009e-01
1.08486223e+00 -2.11053282e-01 3.88650447e-01 1.86556131e-02
4.80153501e-01 -6.88932359e-01 2.55935967e-01 4.92937297e-01
3.23516876e-01 -5.99279940e-01 4.68275964e-01 6.01821244e-01
-3.50138843e-01 4.07596193e-02 -4.54118371e-01 2.90147632e-01
1.87320530e-01 9.87290084e-01 -1.25141931e+00 -1.71821579e-01
4.02743399e-01 8.38754594e-01 -4.11941439e-01 9.77832615e-01
-9.33068022e-02 8.45155001e-01 -4.30084378e-01 3.96571547e-01
1.79668218e-01 -3.23207200e-01 1.00227928e+00 1.10322404e+00
1.69834211e-01 -3.02061200e-01 2.02327445e-01 9.50684905e-01
6.04226254e-04 -5.63751645e-02 -5.99378705e-01 -1.46715105e-01
3.17793041e-01 9.89578426e-01 -8.49210203e-01 -5.16853333e-01
6.17379732e-02 6.68122292e-01 1.04209948e-02 3.66923600e-01
-2.01404735e-01 3.24470878e-01 3.24027658e-01 2.14979574e-01
5.21621406e-01 -5.94145000e-01 -8.53617415e-02 -1.24011552e+00
-2.52878398e-01 -6.52080357e-01 3.48992139e-01 -6.67993605e-01
-1.21541893e+00 2.51746416e-01 7.81583428e-01 -1.18429530e+00
-6.28770411e-01 -3.58974606e-01 -4.02303725e-01 5.02252996e-01
-1.05347657e+00 -1.09961534e+00 9.39726532e-02 4.23021793e-01
7.13970125e-01 -5.00478804e-01 8.98264945e-01 -2.88499862e-01
-2.64533728e-01 -3.01374914e-03 9.68183100e-01 -1.86308905e-01
7.37648487e-01 -1.27091026e+00 -8.44042562e-03 5.43217838e-01
2.49510989e-01 6.49807215e-01 8.66056859e-01 -7.50234902e-01
-7.75796950e-01 -8.81534815e-01 9.94764268e-01 -4.76415724e-01
5.99833667e-01 -7.09497809e-01 -8.73529017e-01 7.86200643e-01
-4.60867882e-02 -5.46055555e-01 1.21492648e+00 3.61361116e-01
-3.85974079e-01 4.99364167e-01 -9.69082177e-01 2.67990768e-01
2.05236658e-01 -6.25786304e-01 -6.13420606e-01 7.92736948e-01
3.67290378e-01 1.91700995e-01 -7.60210574e-01 -9.83918086e-02
4.79236573e-01 -4.99647170e-01 8.20127130e-01 -5.47183931e-01
2.46583715e-01 -6.48908224e-03 -3.74897480e-01 -1.21264327e+00
-2.85084516e-01 -4.82201278e-01 -3.42915624e-01 1.22494149e+00
-7.43553415e-02 -2.75165915e-01 9.08398330e-01 5.66079915e-01
4.17398989e-01 -2.52892315e-01 -8.97483528e-01 -6.87951446e-01
1.33961290e-01 -4.87172276e-01 1.61946520e-01 9.89471257e-01
-4.61650826e-02 3.82987291e-01 -4.57073063e-01 -4.10458222e-02
1.17338729e+00 1.14157841e-01 8.50334346e-01 -1.62456393e+00
-1.43144146e-01 -3.76340300e-02 -4.73796070e-01 -1.02367938e+00
4.88363117e-01 -7.17542708e-01 -1.00566074e-01 -1.25574934e+00
6.43294871e-01 -2.22793788e-01 2.84953475e-01 2.33168498e-01
3.13208610e-01 -4.53733690e-02 -1.34053692e-01 9.81493533e-01
-5.45901716e-01 7.53010094e-01 6.90800905e-01 -1.13962285e-01
-3.19775283e-01 2.96781331e-01 -3.40493709e-01 1.08761740e+00
6.11723065e-01 -1.08261073e+00 -3.07771474e-01 -1.15471803e-01
2.38072172e-01 -2.87838191e-01 3.30860049e-01 -5.26985288e-01
2.90115029e-01 2.37073898e-01 3.35830659e-01 -9.72109377e-01
6.68544710e-01 -5.28771698e-01 -5.77306421e-03 1.06316052e-01
-6.15088761e-01 -6.51975513e-01 -2.56215155e-01 8.39648724e-01
-3.54862869e-01 -4.71252322e-01 6.16185486e-01 -4.37662750e-02
-2.69554317e-01 7.09057897e-02 -9.82777953e-01 -4.35916603e-01
8.12410057e-01 -8.41385573e-02 5.77709787e-02 -5.22183537e-01
-1.32503426e+00 -1.39064491e-01 3.40059668e-01 2.98343748e-01
3.67476761e-01 -1.31150830e+00 -7.21413314e-01 2.22480536e-01
7.90442005e-02 2.32446752e-02 5.90931140e-02 7.20888495e-01
1.63527191e-01 7.82616317e-01 1.70063704e-01 -8.44604492e-01
-1.37835824e+00 3.81188303e-01 2.71627051e-03 -4.97712195e-01
-5.45673370e-01 4.18870389e-01 7.68582582e-01 -2.41784200e-01
1.46537423e-01 5.49531691e-02 -3.98103118e-01 5.32735765e-01
3.47965121e-01 4.43798155e-01 -1.05606809e-01 -6.12321258e-01
3.21491510e-02 3.64336312e-01 -4.87594157e-01 -4.37046617e-01
1.53270507e+00 -5.04530847e-01 -2.78708607e-01 6.89500690e-01
1.45875084e+00 -2.01160312e-01 -1.27427471e+00 -7.17480183e-01
3.47984731e-01 -4.68196392e-01 5.15874922e-01 -1.26465470e-01
-3.76309425e-01 9.00223255e-01 5.55862844e-01 4.07552719e-01
6.90024018e-01 5.30118763e-01 1.36320159e-01 4.52886790e-01
-7.97309875e-02 -1.09271348e+00 1.88911393e-01 3.98532629e-01
7.64655113e-01 -1.10677075e+00 4.26935345e-01 -3.84636968e-01
-4.22915250e-01 1.01124251e+00 -2.48443305e-01 -2.51164734e-01
1.07483149e+00 1.46340579e-01 -5.14436126e-01 -5.38742602e-01
-8.41175795e-01 2.31220737e-01 6.08618021e-01 5.68691432e-01
2.76900113e-01 8.83480161e-02 -9.24386382e-02 3.91128123e-01
-3.01068366e-01 -2.49330565e-01 3.96719873e-01 9.04787779e-01
-6.11798942e-01 -1.03663194e+00 -6.33226335e-01 4.33848530e-01
-5.48440635e-01 -3.61312218e-02 -3.88218939e-01 4.09071296e-01
-2.24891901e-01 7.55676866e-01 3.43595207e-01 2.44000897e-01
-5.35939872e-01 5.79365015e-01 2.30714858e-01 -8.36236358e-01
4.25350338e-01 8.62284243e-01 -2.01147452e-01 -5.23281805e-02
-5.53639352e-01 -1.32844770e+00 -5.71804643e-01 1.60748273e-01
-5.38474739e-01 3.25933903e-01 1.16800821e+00 1.24748862e+00
9.38697159e-02 5.27173914e-02 6.99786544e-01 -8.57685983e-01
-5.63748777e-01 -1.10839319e+00 -1.00499022e+00 1.80222571e-01
1.17575414e-01 -6.81239486e-01 -6.20769739e-01 5.82961440e-01] | [10.343573570251465, 6.907923698425293] |
a2b636a6-fbc9-4fab-8c4f-f0b643790cce | multi-stage-object-detection-with-group | 1608.05159 | null | http://arxiv.org/abs/1608.05159v1 | http://arxiv.org/pdf/1608.05159v1.pdf | Multi-stage Object Detection with Group Recursive Learning | Most of existing detection pipelines treat object proposals independently and
predict bounding box locations and classification scores over them separately.
However, the important semantic and spatial layout correlations among proposals
are often ignored, which are actually useful for more accurate object
detection. In this work, we propose a new EM-like group recursive learning
approach to iteratively refine object proposals by incorporating such context
of surrounding proposals and provide an optimal spatial configuration of object
detections. In addition, we propose to incorporate the weakly-supervised object
segmentation cues and region-based object detection into a multi-stage
architecture in order to fully exploit the learned segmentation features for
better object detection in an end-to-end way. The proposed architecture
consists of three cascaded networks which respectively learn to perform
weakly-supervised object segmentation, object proposal generation and recursive
detection refinement. Combining the group recursive learning and the
multi-stage architecture provides competitive mAPs of 78.6% and 74.9% on the
PASCAL VOC2007 and VOC2012 datasets respectively, which outperforms many
well-established baselines [10] [20] significantly. | ['Xiaodan Liang', 'Jianshu Li', 'Jiashi Feng', 'Tingfa Xu', 'Shuicheng Yan', 'Jianan Li'] | 2016-08-18 | null | null | null | null | ['object-proposal-generation'] | ['computer-vision'] | [ 8.53636190e-02 3.76722887e-02 -1.24354273e-01 -6.36807919e-01
-7.78937876e-01 -5.07649183e-01 4.30780560e-01 3.18554133e-01
-7.93137133e-01 2.06672952e-01 -2.06749812e-01 -5.43945469e-02
3.80896926e-01 -6.10805154e-01 -7.99159467e-01 -5.69198132e-01
3.80172320e-02 4.69721019e-01 1.08709049e+00 1.98983356e-01
3.85772943e-01 5.92108190e-01 -1.35679519e+00 3.04238737e-01
8.31109226e-01 1.05940354e+00 7.14701116e-01 6.73529565e-01
-9.11004618e-02 4.43727106e-01 -2.66019076e-01 -1.94767028e-01
2.51512229e-01 3.80881280e-02 -5.99493802e-01 4.03856277e-01
5.81929266e-01 -4.95560646e-01 1.25772074e-01 9.98877048e-01
3.43992382e-01 1.29180029e-01 7.15423524e-01 -7.81476140e-01
-1.81963071e-01 5.06627321e-01 -1.15029442e+00 3.37427884e-01
-2.14570180e-01 3.52300227e-01 1.15260613e+00 -1.11809385e+00
3.15209359e-01 1.28358626e+00 4.31030899e-01 3.77195567e-01
-1.16943049e+00 -8.41554880e-01 6.80232465e-01 2.25041479e-01
-1.50565851e+00 -2.03895181e-01 8.11460435e-01 -4.31816608e-01
7.11458921e-01 3.29499207e-02 5.57952821e-01 5.49830973e-01
-1.85754701e-01 1.28715372e+00 6.95928037e-01 -3.09636861e-01
4.99826483e-02 2.92773455e-01 1.95260048e-01 6.77364707e-01
3.03177446e-01 -1.02252021e-01 -1.07797854e-01 1.23775870e-01
7.78638721e-01 2.58066982e-01 1.77099720e-01 -5.63087940e-01
-1.15780056e+00 7.09562659e-01 1.15948880e+00 1.15991749e-01
-4.10436988e-01 1.25137776e-01 3.18840861e-01 -6.26973629e-01
4.25291985e-01 3.94993007e-01 -4.38043743e-01 5.01263797e-01
-1.14537060e+00 2.98410565e-01 1.26082063e-01 9.31507587e-01
9.39770877e-01 -2.64695674e-01 -5.85530341e-01 8.81387353e-01
7.82903910e-01 3.79159003e-01 1.07222110e-01 -6.75996006e-01
6.00464582e-01 9.23046827e-01 3.77042711e-01 -8.60388577e-01
-5.65533102e-01 -9.35030222e-01 -4.72030610e-01 -8.08661804e-02
3.01401287e-01 -3.31299379e-02 -1.32882822e+00 1.34497607e+00
6.47586524e-01 4.45214301e-01 -2.59193391e-01 1.14910829e+00
1.01061165e+00 6.85462892e-01 3.56825382e-01 1.33732200e-01
1.38779151e+00 -1.44073260e+00 -2.86660522e-01 -5.79568148e-01
5.51686883e-01 -9.89905596e-01 7.43020236e-01 8.70820880e-02
-1.09861386e+00 -8.91638100e-01 -9.45128620e-01 -2.57216066e-01
-1.27312671e-02 8.41688395e-01 6.64834976e-01 3.07523549e-01
-7.26487696e-01 6.43357635e-02 -1.02586353e+00 -9.42476373e-03
1.05100536e+00 3.68062168e-01 1.46997914e-01 -1.50798514e-01
-4.05921012e-01 6.05169237e-01 8.51876140e-01 5.02629817e-01
-1.01094997e+00 -5.87083578e-01 -7.65852213e-01 -3.24009545e-02
5.68655789e-01 -4.75309044e-01 1.24951625e+00 -6.06247246e-01
-1.10042500e+00 8.26251686e-01 -3.80697966e-01 -5.05153716e-01
6.76894724e-01 -4.50150102e-01 1.72241420e-01 -3.31002735e-02
2.80496418e-01 1.23535311e+00 6.18984103e-01 -1.35913289e+00
-1.24003601e+00 -4.15899605e-01 -1.15601011e-01 1.42733887e-01
-8.54206979e-02 -2.93800719e-02 -1.00737071e+00 -4.59924757e-01
5.42367935e-01 -8.28690112e-01 -6.86739206e-01 5.72487675e-02
-6.27217472e-01 -6.06542945e-01 1.05853379e+00 -4.12923872e-01
9.56880808e-01 -1.91683197e+00 5.54800071e-02 5.66889234e-02
1.21527947e-01 3.56624633e-01 -2.25377366e-01 -2.23968789e-01
3.34527969e-01 -9.27814320e-02 -1.77407056e-01 -6.63590014e-01
-1.67366803e-01 -1.39224350e-01 -2.47502536e-01 4.78912801e-01
6.57880068e-01 1.17406702e+00 -8.85482430e-01 -6.70135617e-01
5.07875025e-01 4.14159834e-01 -7.31088281e-01 3.65604460e-01
-4.83527362e-01 5.57716310e-01 -7.00226903e-01 6.95938408e-01
9.08559561e-01 -2.21942887e-01 -9.32761356e-02 -3.95978004e-01
-2.46769130e-01 1.44723743e-01 -1.27308583e+00 1.61374271e+00
-1.67531043e-01 4.88020718e-01 -2.44464893e-02 -8.58986199e-01
9.27481592e-01 -1.75127909e-01 1.02984682e-01 -5.87727845e-01
1.01833887e-01 2.79504329e-01 1.82293326e-01 -2.89025426e-01
4.80134040e-01 2.51464814e-01 1.05391249e-01 1.54642165e-01
-4.80572283e-02 -1.54956421e-02 2.11390078e-01 5.75384162e-02
5.42995751e-01 2.40070239e-01 1.86420336e-01 -1.27921090e-01
7.51538754e-01 -5.30935712e-02 7.12589562e-01 7.70351112e-01
-2.66914964e-01 7.91627645e-01 1.65926248e-01 -3.87470961e-01
-9.26165819e-01 -9.98640418e-01 -2.39518940e-01 1.37392962e+00
5.79739273e-01 -7.40606934e-02 -6.67483866e-01 -8.57815087e-01
-1.65884882e-01 5.79421461e-01 -4.55509752e-01 1.22360438e-01
-8.83806586e-01 -8.75136912e-01 1.69870302e-01 8.91542554e-01
6.94093227e-01 -1.16905570e+00 -6.29875064e-01 3.52526873e-01
-1.61138531e-02 -1.22462022e+00 -5.30342162e-01 1.08832680e-01
-9.34837639e-01 -9.84303951e-01 -7.91541755e-01 -1.01240027e+00
8.84716928e-01 4.19123828e-01 7.77367115e-01 1.68387726e-01
-5.68207800e-01 -2.09190652e-01 -2.60498941e-01 -4.86847281e-01
1.04265176e-01 3.14546645e-01 -3.08946401e-01 1.32778585e-01
2.85856277e-01 -1.13764271e-01 -1.01195681e+00 5.67509055e-01
-6.43240750e-01 3.15018475e-01 1.10465872e+00 6.04528427e-01
8.43740165e-01 -2.41318449e-01 5.16955197e-01 -8.18862259e-01
-8.20086077e-02 -1.83741540e-01 -8.80436599e-01 1.74364343e-01
-2.10907951e-01 -1.05701022e-01 2.06277952e-01 -2.00746641e-01
-1.18265712e+00 5.76231360e-01 -2.32781902e-01 -3.29944611e-01
-2.88882941e-01 -3.67917977e-02 -3.23887140e-01 -8.59638825e-02
3.93076509e-01 3.28272671e-01 -6.50215268e-01 -6.90287530e-01
5.24759829e-01 5.10954320e-01 6.00012600e-01 -3.28794748e-01
8.61795366e-01 5.72334349e-01 -3.65394592e-01 -4.38551188e-01
-1.17670298e+00 -9.99758422e-01 -1.01337790e+00 -8.79502296e-02
1.05515325e+00 -1.07524061e+00 -4.81580555e-01 4.22150224e-01
-1.43590653e+00 -2.31603831e-01 1.21984612e-02 3.69591177e-01
-1.76775932e-01 2.38570854e-01 -4.32320207e-01 -9.35371637e-01
-3.61240655e-01 -1.25220919e+00 1.51413834e+00 6.10784769e-01
2.15930298e-01 -4.56398576e-01 -3.44305456e-01 4.68248099e-01
1.22872777e-01 3.03101800e-02 3.84658635e-01 -6.62217081e-01
-1.10745180e+00 -3.48494411e-01 -9.07783747e-01 1.72527730e-01
-1.27599746e-01 -6.34122565e-02 -9.87562537e-01 -3.56168151e-01
-4.25567776e-01 -9.08229426e-02 1.35901797e+00 5.86081624e-01
1.37099683e+00 -1.49935307e-02 -7.41008341e-01 5.29382288e-01
1.28924274e+00 4.87772636e-02 3.00479949e-01 1.57104000e-01
9.42847848e-01 6.10643327e-01 9.57713246e-01 3.78888249e-01
3.52303177e-01 6.83857441e-01 5.95083773e-01 -2.83090353e-01
-5.74835390e-02 -2.68359333e-01 -3.00665386e-02 2.50082165e-01
9.74702463e-03 -5.01104211e-03 -8.87851000e-01 7.20842123e-01
-1.94092560e+00 -5.15992999e-01 -2.79211760e-01 1.98941910e+00
5.76311588e-01 5.68648279e-01 2.90922493e-01 -2.78858155e-01
9.28217053e-01 1.13620490e-01 -7.92281985e-01 2.37226605e-01
1.22319795e-01 -4.03646491e-02 6.55505776e-01 3.49811196e-01
-1.53732085e+00 1.30509472e+00 5.15687227e+00 8.88307512e-01
-8.78816664e-01 1.34647295e-01 8.51888120e-01 -9.17167589e-02
1.32714540e-01 -9.52120200e-02 -1.34207213e+00 2.21570969e-01
1.75407484e-01 5.11880457e-01 -2.58267283e-01 9.61620688e-01
3.02686095e-01 -1.81202814e-01 -1.00280333e+00 8.78315270e-01
-1.89610854e-01 -1.30707824e+00 -9.65540707e-02 -1.66708663e-01
1.03319299e+00 2.42688730e-01 3.10295336e-02 3.57718110e-01
8.86561871e-02 -8.50015640e-01 9.32733655e-01 2.28454873e-01
1.11822605e-01 -7.57006347e-01 8.71642053e-01 5.28668344e-01
-1.52408457e+00 -2.68240899e-01 -4.96762395e-01 1.57533109e-01
2.44171232e-01 5.02461851e-01 -1.06606233e+00 1.96524590e-01
7.91563570e-01 7.41944849e-01 -9.61479962e-01 1.65215969e+00
-3.96148413e-01 6.87352657e-01 -4.08308834e-01 -1.79071710e-01
5.94830453e-01 9.71070305e-02 4.92429614e-01 1.45552683e+00
-1.54273644e-01 1.59908071e-01 6.39427722e-01 1.22509289e+00
-6.05155081e-02 1.28081769e-01 2.13501617e-01 2.58080512e-01
4.05387312e-01 1.68067217e+00 -1.28429735e+00 -3.20262641e-01
-2.85006583e-01 8.18608463e-01 4.33289766e-01 3.14451933e-01
-9.45466936e-01 -3.17748666e-01 2.43261650e-01 1.38918310e-01
8.10531974e-01 -2.66930312e-01 -3.99714649e-01 -7.89020061e-01
1.62695095e-01 -2.68288106e-01 2.23527148e-01 -4.49119896e-01
-1.05997097e+00 3.07480276e-01 -2.05739111e-01 -1.02701592e+00
1.16868623e-01 -5.31772852e-01 -8.82223427e-01 7.88693011e-01
-1.76558125e+00 -1.49763465e+00 -3.88184279e-01 3.24630849e-02
8.65906000e-01 1.71106204e-01 1.11216128e-01 2.77328342e-01
-8.43818128e-01 5.18773913e-01 -1.60071597e-01 4.65472758e-01
5.04561543e-01 -1.24798524e+00 4.61419523e-01 9.44294274e-01
2.19342247e-01 4.55602795e-01 4.07294154e-01 -6.54371738e-01
-8.57774317e-01 -1.64104033e+00 4.53151971e-01 -4.22312409e-01
2.42503420e-01 -6.05513096e-01 -9.84821320e-01 4.66363966e-01
-8.19407180e-02 3.42238039e-01 9.05243233e-02 7.69056603e-02
-2.76301980e-01 -1.29166871e-01 -7.56328106e-01 5.37959993e-01
1.02968490e+00 -9.11284313e-02 -4.21233982e-01 5.15243173e-01
7.59363592e-01 -4.73767698e-01 -2.31048584e-01 6.15132987e-01
2.98518509e-01 -8.77551734e-01 1.10278571e+00 -4.25550014e-01
1.33415774e-01 -7.73311734e-01 1.34935647e-01 -6.29598081e-01
-3.72879863e-01 -2.82773614e-01 4.98009920e-02 1.16149056e+00
5.73461950e-01 -2.39553645e-01 1.00732517e+00 3.71594548e-01
-1.20633885e-01 -8.99466276e-01 -6.58514023e-01 -2.98831612e-01
-2.08152100e-01 -5.77189922e-01 2.93857127e-01 2.83181131e-01
-6.55316532e-01 3.64218891e-01 -2.55054105e-02 4.24351156e-01
7.51088977e-01 4.29777503e-01 8.65856469e-01 -1.05039954e+00
-1.72673255e-01 -7.36238003e-01 -5.34943759e-01 -1.64914930e+00
-9.67875570e-02 -7.22150743e-01 5.32199025e-01 -1.55443525e+00
6.31652534e-01 -6.66027784e-01 -4.45739806e-01 3.98499876e-01
-6.86900020e-01 5.97834527e-01 1.07827365e-01 3.59573305e-01
-1.20724595e+00 5.14864206e-01 1.22897005e+00 -1.39313191e-01
-4.59785461e-01 3.81972939e-01 -5.12497425e-01 9.10355031e-01
5.52634895e-01 -4.20854419e-01 -1.40727326e-01 -3.91152829e-01
-3.76363754e-01 -4.12262678e-01 7.12128341e-01 -9.82429206e-01
4.52532917e-01 3.18183079e-02 7.32926071e-01 -1.33861005e+00
2.76295632e-01 -5.79180062e-01 -6.71702504e-01 4.69341397e-01
-3.89985621e-01 -5.31447828e-01 2.87449598e-01 9.31994677e-01
-6.76466301e-02 -2.07703531e-01 9.72781122e-01 -6.98578730e-02
-1.00383317e+00 4.23538387e-01 1.93032343e-02 -5.81926890e-02
1.24775743e+00 -2.13107780e-01 6.57709017e-02 2.26559609e-01
-6.43948615e-01 6.84226990e-01 5.24651818e-02 5.98029673e-01
5.78623116e-01 -9.80179250e-01 -8.32371473e-01 1.02757327e-01
1.66192532e-01 8.05503011e-01 2.70003289e-01 8.01359951e-01
-4.43081051e-01 5.05847871e-01 1.10823043e-01 -1.17682850e+00
-1.16566646e+00 3.03078681e-01 3.46122533e-01 -6.41616061e-02
-5.85040927e-01 1.42882288e+00 6.91914618e-01 -4.34810638e-01
3.93225551e-01 -4.81828690e-01 -2.81266749e-01 -9.15790275e-02
4.90504533e-01 1.15075193e-01 -1.79614097e-01 -6.75027072e-01
-4.83380497e-01 7.52213061e-01 -5.70834875e-01 1.61364630e-01
1.16751385e+00 -1.62079066e-01 1.00501209e-01 2.11783484e-01
1.07349503e+00 -2.93541282e-01 -1.64824474e+00 -5.75023413e-01
3.15734893e-01 -3.84893775e-01 1.72838181e-01 -7.02835977e-01
-1.21752822e+00 1.05046558e+00 6.67684317e-01 -2.60362238e-01
9.01089966e-01 2.79531658e-01 7.51695931e-01 3.30369234e-01
8.06104168e-02 -9.91387963e-01 2.71291852e-01 2.39948496e-01
6.64836109e-01 -1.57177126e+00 1.27194285e-01 -7.69906759e-01
-3.51027310e-01 9.40139472e-01 1.06911552e+00 -2.62468189e-01
3.57262403e-01 6.05781861e-02 -5.95735610e-02 3.44226696e-02
-3.95221710e-01 -4.84361708e-01 8.39699626e-01 3.19723248e-01
3.36894602e-01 5.74799627e-02 -2.11119846e-01 7.15826690e-01
2.60968179e-01 -3.42105865e-01 1.84805319e-02 7.21240461e-01
-9.17302191e-01 -7.84538507e-01 -4.62801993e-01 5.04929066e-01
-2.56117970e-01 9.39199030e-02 -3.11159730e-01 6.63034022e-01
4.65982467e-01 7.77125478e-01 2.65320718e-01 -2.88167708e-02
2.57949948e-01 -2.46652573e-01 2.79089570e-01 -9.68174279e-01
-4.93180871e-01 4.82224315e-01 -3.09375346e-01 -4.55822736e-01
-3.69924545e-01 -7.68060803e-01 -1.55161893e+00 3.80864531e-01
-9.05682623e-01 -1.16701022e-01 5.91331601e-01 1.10380852e+00
1.90283954e-01 6.28417969e-01 3.51311296e-01 -1.27333593e+00
-1.63944021e-01 -1.10747755e+00 -1.49357259e-01 1.14125714e-01
1.91509694e-01 -6.20499313e-01 6.15998991e-02 -7.57217705e-02] | [9.32315444946289, 0.6379354000091553] |
bb6171c6-33c8-4498-8fd1-220d5a3b9f22 | pdfnet-pointwise-dense-flow-network-for-urban | 2109.10083 | null | https://arxiv.org/abs/2109.10083v1 | https://arxiv.org/pdf/2109.10083v1.pdf | PDFNet: Pointwise Dense Flow Network for Urban-Scene Segmentation | In recent years, using a deep convolutional neural network (CNN) as a feature encoder (or backbone) is the most commonly observed architectural pattern in several computer vision methods, and semantic segmentation is no exception. The two major drawbacks of this architectural pattern are: (i) the networks often fail to capture small classes such as wall, fence, pole, traffic light, traffic sign, and bicycle, which are crucial for autonomous vehicles to make accurate decisions. (ii) due to the arbitrarily increasing depth, the networks require massive labeled data and additional regularization techniques to converge and to prevent the risk of over-fitting, respectively. While regularization techniques come at minimal cost, the collection of labeled data is an expensive and laborious process. In this work, we address these two drawbacks by proposing a novel lightweight architecture named point-wise dense flow network (PDFNet). In PDFNet, we employ dense, residual, and multiple shortcut connections to allow a smooth gradient flow to all parts of the network. The extensive experiments on Cityscapes and CamVid benchmarks demonstrate that our method significantly outperforms baselines in capturing small classes and in few-data regimes. Moreover, our method achieves considerable performance in classifying out-of-the training distribution samples, evaluated on Cityscapes to KITTI dataset. | ['Venkata Satya Sai Ajay Daliparthi'] | 2021-09-21 | null | null | null | null | ['scene-segmentation'] | ['computer-vision'] | [-6.57724738e-02 -1.47623241e-01 -3.80410969e-01 -5.42046964e-01
-1.13125801e-01 -4.48114127e-01 6.08245850e-01 -2.39886642e-01
-4.70420599e-01 6.36451304e-01 -3.29752624e-01 -3.60068589e-01
-4.83978763e-02 -9.47533190e-01 -7.75838792e-01 -5.65586567e-01
1.07089065e-01 3.28767151e-01 6.42166018e-01 -2.60067016e-01
3.24371941e-02 7.10836291e-01 -1.49313295e+00 -1.83941051e-01
1.11477172e+00 1.19155645e+00 9.60631669e-02 1.45922244e-01
-4.80959833e-01 7.49709547e-01 -3.28910500e-01 -5.84937751e-01
6.24016702e-01 -5.67970686e-02 -7.41453350e-01 3.43213499e-01
6.56401634e-01 -3.90812337e-01 -4.74473476e-01 1.12155986e+00
1.06050901e-01 2.61099339e-01 6.81856573e-01 -1.30909181e+00
-4.09435004e-01 2.65805185e-01 -9.00906861e-01 1.66035607e-01
-2.51865000e-01 3.52691114e-01 8.70935261e-01 -7.82985568e-01
5.22685826e-01 1.11642051e+00 7.68864632e-01 4.08694834e-01
-9.81394231e-01 -7.02525914e-01 4.23838675e-01 1.73973933e-01
-1.35414016e+00 -2.61583030e-01 8.51868510e-01 -4.49520141e-01
6.10098481e-01 -1.41322520e-02 6.59818470e-01 1.08067060e+00
-1.63587064e-01 8.60757649e-01 6.16303742e-01 3.15088853e-02
9.62378606e-02 1.21103153e-01 3.14845383e-01 8.73014987e-01
3.17365319e-01 2.02434789e-03 8.57130289e-02 2.17854977e-01
7.27259278e-01 2.94627070e-01 -1.28997326e-01 -3.82989168e-01
-5.15997589e-01 9.37840402e-01 8.28027368e-01 7.55588785e-02
-1.33836553e-01 1.35137841e-01 4.03137326e-01 4.85904552e-02
4.37255919e-01 -8.52456167e-02 -3.35720450e-01 -6.19633682e-02
-8.88520539e-01 1.62048906e-01 5.49245059e-01 9.72920835e-01
1.09787345e+00 2.01656356e-01 1.61520526e-01 9.94842291e-01
2.14459464e-01 4.47742730e-01 2.79188722e-01 -8.66449654e-01
7.65600622e-01 1.02303994e+00 -1.16304807e-01 -1.12308073e+00
-4.47061628e-01 -6.29042149e-01 -1.14270878e+00 2.59498209e-01
6.59794807e-01 -2.11612672e-01 -1.27651095e+00 1.33682168e+00
3.52744430e-01 3.60709816e-01 -2.33126000e-01 1.00867105e+00
8.51251483e-01 5.46938896e-01 2.07539752e-01 2.85038054e-01
1.06974924e+00 -1.24209070e+00 -3.07679206e-01 -4.82961923e-01
5.47285557e-01 -5.26656091e-01 1.05436707e+00 2.46153176e-02
-7.44675756e-01 -6.16971672e-01 -9.41892684e-01 -2.62612134e-01
-4.86346304e-01 1.72953844e-01 7.47566223e-01 5.42725861e-01
-7.50867903e-01 5.38114786e-01 -7.93976068e-01 -1.54037625e-01
9.02523160e-01 1.95174083e-01 -3.60099584e-01 -2.06235051e-01
-9.56161439e-01 5.69357753e-01 2.04531506e-01 4.03645813e-01
-7.14624226e-01 -8.03626597e-01 -1.07062411e+00 1.65085346e-01
4.24295992e-01 -3.26247185e-01 9.51870978e-01 -1.15655911e+00
-1.47995842e+00 6.55901968e-01 3.43250521e-02 -5.65932810e-01
9.09017920e-01 -2.93521285e-01 -2.84233660e-01 9.80700999e-02
7.58970827e-02 8.10245752e-01 8.24730754e-01 -1.10275507e+00
-8.40679765e-01 -1.71359345e-01 1.74859405e-01 -8.12150091e-02
-2.53372848e-01 -2.77439386e-01 -6.07585192e-01 -4.09152687e-01
1.59002900e-01 -7.96555340e-01 -4.39199418e-01 1.62408233e-01
-5.76811314e-01 -3.21196347e-01 1.21079433e+00 -1.71885610e-01
9.80433226e-01 -2.28807616e+00 -2.81161606e-01 2.73729235e-01
2.66240984e-01 6.96753085e-01 -7.15411529e-02 -2.62635518e-02
1.64363563e-01 1.83857292e-01 -4.21961546e-01 -3.92233670e-01
-1.57474712e-01 4.17962492e-01 -2.37435102e-01 5.14140904e-01
4.74070996e-01 9.64872718e-01 -7.98416734e-01 -4.37218428e-01
5.65886974e-01 6.52973354e-01 -4.11549419e-01 -1.24297850e-01
-1.40058532e-01 3.86902452e-01 -6.03841782e-01 6.87315404e-01
8.77472579e-01 -3.00050706e-01 -2.38782451e-01 -1.06363736e-01
-2.25597307e-01 7.08201602e-02 -1.19110394e+00 1.33626378e+00
-4.10359710e-01 7.20633626e-01 1.20608628e-01 -1.25846779e+00
9.97374356e-01 -9.57271308e-02 3.86951447e-01 -8.74184370e-01
2.91618496e-01 1.98699921e-01 -4.37108614e-02 -5.27924836e-01
3.49564016e-01 2.60890961e-01 3.05930208e-02 8.20462406e-02
-1.20724976e-01 1.73882678e-01 4.09917146e-01 -3.21644098e-02
6.96775734e-01 -6.87651187e-02 -5.09831682e-02 -2.97229141e-01
6.73042893e-01 4.47285920e-02 1.02972853e+00 5.42754650e-01
-4.50222164e-01 7.74682462e-01 6.68666303e-01 -9.96116161e-01
-9.29916620e-01 -7.99851000e-01 -2.21403688e-01 8.37627769e-01
3.18948507e-01 2.26216000e-02 -7.19079316e-01 -9.47203338e-01
-1.27210654e-02 2.88086563e-01 -5.75131297e-01 6.28637383e-03
-7.57674992e-01 -6.40703738e-01 4.95909363e-01 8.09639513e-01
1.05468678e+00 -8.94602120e-01 -6.28820479e-01 1.53946877e-01
-6.94680735e-02 -1.41440475e+00 -4.20329124e-01 4.58901972e-02
-7.89324164e-01 -1.33850896e+00 -7.26634443e-01 -9.78050292e-01
7.73374617e-01 4.54575509e-01 9.61915851e-01 1.91951886e-01
-3.09972256e-01 -2.05582038e-01 -2.14954302e-01 -1.38487965e-01
1.38011262e-01 3.11263204e-01 -3.80458295e-01 3.25305074e-01
3.94726366e-01 -5.01318157e-01 -8.48011434e-01 5.72210312e-01
-7.90262222e-01 2.54512206e-02 4.13846314e-01 7.49940336e-01
5.49575448e-01 4.67807502e-02 3.85847002e-01 -1.07413447e+00
1.47157595e-01 -5.95067561e-01 -8.35427999e-01 -4.50520776e-02
-3.00018191e-01 -2.40160063e-01 8.97312284e-01 -2.53530234e-01
-1.00323629e+00 2.42888242e-01 -3.15073937e-01 -5.66398680e-01
-4.01940018e-01 1.68900430e-01 -1.52395517e-01 -7.56926462e-02
4.77236301e-01 6.41518012e-02 -3.44339684e-02 -4.60436404e-01
2.23408133e-01 5.42823613e-01 4.12088722e-01 -3.45996112e-01
8.80026758e-01 8.08048368e-01 1.30654499e-01 -1.00362802e+00
-8.71440053e-01 -5.25820076e-01 -6.80209577e-01 -1.12136021e-01
7.97562838e-01 -8.33996892e-01 -6.17185533e-01 5.76853693e-01
-8.66182387e-01 -4.94437903e-01 -2.64661670e-01 2.04705119e-01
-1.83417752e-01 2.81047910e-01 -5.22481501e-01 -5.33254981e-01
-2.35934481e-01 -1.45163298e+00 7.97866404e-01 7.59859622e-01
1.57030597e-01 -1.01620460e+00 -4.32051212e-01 3.44785988e-01
6.18495524e-01 4.20186102e-01 6.64000869e-01 -3.28230888e-01
-6.75399899e-01 -2.54071444e-01 -6.01400614e-01 5.58614612e-01
2.40064025e-01 3.44010860e-01 -8.77700388e-01 8.62983230e-04
-3.99754852e-01 -3.74566317e-01 1.16905129e+00 4.22364682e-01
1.41803753e+00 -1.48035556e-01 -3.10945153e-01 9.88924563e-01
1.29839993e+00 6.21748529e-02 6.23401582e-01 2.88577765e-01
9.37815726e-01 6.42565548e-01 3.49748760e-01 1.99795395e-01
5.11067152e-01 4.64077681e-01 6.10475659e-01 -2.82847434e-01
-3.16785991e-01 -1.35249868e-01 4.90649007e-02 4.97160345e-01
-1.12733036e-01 -1.00768790e-01 -8.59400511e-01 7.07524955e-01
-1.77817035e+00 -7.72986472e-01 -3.76830071e-01 1.90181255e+00
3.69622529e-01 3.91230911e-01 1.42296210e-01 1.60659000e-01
5.63820422e-01 2.17356488e-01 -6.53713226e-01 -1.25472084e-01
-1.61211804e-01 1.38357013e-01 6.71729982e-01 3.86805087e-01
-1.40614903e+00 1.02694941e+00 5.40720129e+00 9.07471180e-01
-1.45463479e+00 -3.15182991e-02 9.40278471e-01 7.73454383e-02
-1.17665734e-02 -2.85905570e-01 -9.00897086e-01 6.45721793e-01
4.91976142e-01 4.17165935e-01 1.32932767e-01 1.01015544e+00
2.24012777e-01 -5.69593050e-02 -6.22217953e-01 9.68833864e-01
-2.71265268e-01 -1.38026750e+00 1.03128818e-03 -6.31063897e-03
8.75926077e-01 4.95646656e-01 7.43475035e-02 2.97773629e-01
4.33248878e-01 -1.03606808e+00 6.51260793e-01 1.19684063e-01
7.79228747e-01 -9.78134573e-01 7.22147107e-01 3.37560356e-01
-1.42254150e+00 -1.72937065e-01 -5.45747697e-01 7.24977553e-02
1.91468045e-01 7.18562245e-01 -2.72632986e-01 4.05731678e-01
8.72223616e-01 8.79042268e-01 -4.39708710e-01 1.22875667e+00
-3.77632529e-01 7.12683022e-01 -5.96221030e-01 1.41173914e-01
8.49797368e-01 -3.57116550e-01 2.64733970e-01 1.09350634e+00
3.65997516e-02 -1.98287740e-01 4.31679726e-01 8.44835639e-01
-2.48452827e-01 -1.10287830e-01 -5.12015581e-01 2.80874699e-01
1.83441639e-01 1.51996863e+00 -1.10704255e+00 -2.27589473e-01
-6.35274172e-01 6.45760775e-01 4.34680611e-01 5.60969234e-01
-8.67887080e-01 -4.28054005e-01 8.64527285e-01 2.18472674e-01
4.76688057e-01 -2.23812476e-01 -4.55575436e-01 -9.79831934e-01
1.49193123e-01 -4.14689273e-01 1.46001488e-01 -8.68350044e-02
-1.19571292e+00 6.40519202e-01 -2.91389942e-01 -1.30157125e+00
8.25153813e-02 -7.31939018e-01 -8.30404341e-01 6.12605095e-01
-1.96641719e+00 -1.09723043e+00 -5.86410105e-01 6.23733222e-01
7.40667403e-01 -9.22147408e-02 2.02328071e-01 8.12693894e-01
-8.85420203e-01 5.95589280e-01 2.25764122e-02 5.92900634e-01
1.78476363e-01 -1.05756664e+00 4.44054663e-01 7.36003280e-01
7.86772445e-02 2.85449117e-01 2.21018761e-01 -3.51486355e-01
-8.54898155e-01 -1.41210818e+00 5.89730263e-01 -1.33817298e-02
5.02564430e-01 -3.79410535e-01 -9.69112039e-01 5.77044606e-01
-6.97964132e-02 6.28944337e-01 2.57237911e-01 -2.90926043e-02
-3.73863161e-01 -3.96560550e-01 -1.07848167e+00 5.39189935e-01
1.03079724e+00 -7.48220980e-02 -1.90249592e-01 2.16897741e-01
5.31507254e-01 -4.28086281e-01 -3.62672031e-01 4.58841562e-01
4.13466752e-01 -1.04729927e+00 9.12496507e-01 -4.31477606e-01
4.81677175e-01 -3.62334490e-01 1.36899561e-01 -1.19873095e+00
-1.86576828e-01 -4.29292291e-01 3.94230969e-02 1.16787219e+00
3.83983135e-01 -7.86901116e-01 1.16128695e+00 5.06452382e-01
-3.30570191e-01 -1.02257669e+00 -9.77297604e-01 -7.65437841e-01
1.25106975e-01 -5.17325878e-01 6.87399268e-01 8.39930892e-01
-7.31559038e-01 1.92524254e-01 -2.34113783e-01 8.35095905e-03
6.56023204e-01 -9.10771787e-02 8.17741454e-01 -1.36457193e+00
1.71113253e-01 -8.00781369e-01 -5.30878067e-01 -1.28036046e+00
1.63355455e-01 -7.43720472e-01 -1.61490604e-01 -1.52043879e+00
-1.38624802e-01 -8.77852976e-01 -6.66626170e-02 4.68479067e-01
-7.11276010e-02 3.97501945e-01 7.62651786e-02 1.09121069e-01
-4.94989902e-01 6.21450901e-01 1.37955999e+00 -3.39593798e-01
-1.75580367e-01 3.29079419e-01 -5.12925446e-01 1.06253874e+00
9.13096964e-01 -3.65676552e-01 -5.78915238e-01 -6.51656032e-01
6.53146058e-02 -4.36194867e-01 5.40839255e-01 -9.82865572e-01
2.63121754e-01 -2.26497054e-01 3.32029730e-01 -7.06003487e-01
1.42019823e-01 -1.08890450e+00 -2.28053078e-01 3.57326001e-01
7.71062747e-02 -1.01906866e-01 1.71104163e-01 4.91234839e-01
-4.31923538e-01 -2.36054987e-01 1.00262868e+00 -5.24335355e-02
-9.61753845e-01 6.55025959e-01 -1.44493088e-01 3.35845619e-01
1.15060461e+00 -4.45284069e-01 -3.53778332e-01 -5.46449572e-02
-2.82156259e-01 6.68659806e-01 1.56525791e-01 5.61066389e-01
4.34020698e-01 -1.16270947e+00 -4.73599792e-01 4.92406636e-01
-8.13564211e-02 8.27572942e-01 3.37204814e-01 7.91114271e-01
-9.22212541e-01 3.19731027e-01 -8.97562951e-02 -7.43093193e-01
-7.33958781e-01 1.39856726e-01 5.81331074e-01 -1.57403424e-01
-9.53589022e-01 8.02407861e-01 3.29661518e-01 -5.26611507e-01
4.32399541e-01 -4.74715561e-01 -3.51041973e-01 5.81541210e-02
4.12661016e-01 5.04534781e-01 1.13793418e-01 -7.65488505e-01
-3.17777634e-01 6.64481997e-01 -2.52491534e-01 5.19651353e-01
1.38688421e+00 -1.59291059e-01 1.52188957e-01 7.09973555e-03
1.18707275e+00 -3.31696302e-01 -1.86454463e+00 -2.51048863e-01
-2.05368754e-02 -4.03546780e-01 -2.59587937e-03 -4.09958988e-01
-1.84491789e+00 1.10090435e+00 4.01697904e-01 2.39230812e-01
8.27639878e-01 -1.86434075e-01 1.17137647e+00 3.24652374e-01
1.79667100e-01 -1.26036859e+00 -1.79754451e-01 7.50104249e-01
4.91207123e-01 -1.40046763e+00 -2.75514454e-01 -5.51821470e-01
-5.18629372e-01 1.08521950e+00 9.31548655e-01 -3.81374687e-01
9.40334797e-01 1.88075483e-01 2.25516781e-01 -1.25408471e-01
-2.95763463e-01 -3.44861329e-01 2.50070274e-01 4.98029232e-01
1.35803074e-01 -6.54021576e-02 -8.61654505e-02 3.73188853e-01
-1.01126380e-01 -1.93308786e-01 3.85984600e-01 7.49242067e-01
-3.98502290e-01 -7.33559132e-01 -1.43262763e-02 5.90535641e-01
-5.60020864e-01 1.95116550e-01 -1.58103362e-01 1.00142777e+00
4.01395202e-01 8.80486608e-01 3.84806186e-01 -1.34204999e-01
4.91155118e-01 -1.65868461e-01 -4.22199555e-02 -3.08286309e-01
-5.45417488e-01 -1.36344701e-01 -2.10008010e-01 -7.40872622e-01
-3.61220896e-01 -3.02451432e-01 -1.35908115e+00 -3.52190703e-01
-2.66129017e-01 -1.31089181e-01 4.99667913e-01 1.10793447e+00
3.48619401e-01 5.34428656e-01 7.05398142e-01 -7.58604407e-01
-3.84164393e-01 -7.90698051e-01 -4.29588318e-01 5.99082947e-01
4.01316434e-01 -9.08287108e-01 -3.38841945e-01 -8.19713920e-02] | [9.476289749145508, 0.2662179470062256] |
e5628cac-0af6-41cb-91f6-d95bea3a4ddd | automatic-construction-of-an-annotated-corpus | null | null | https://aclanthology.org/2022.lrec-1.755 | https://aclanthology.org/2022.lrec-1.755.pdf | Automatic Construction of an Annotated Corpus with Implicit Aspects | Aspect-based sentiment analysis (ABSA) is a task that involves classifying the polarity of aspects of the products or services described in users’ reviews. Most previous work on ABSA has focused on explicit aspects, which appear as explicit words or phrases in the sentences of the review. However, users often express their opinions toward the aspects indirectly or implicitly, in which case the specific name of an aspect does not appear in the review. The current datasets used for ABSA are mainly annotated with explicit aspects. This paper proposes a novel method for constructing a corpus that is automatically annotated with implicit aspects. The main idea is that sentences containing explicit and implicit aspects share a similar context. First, labeled sentences with explicit aspects and unlabeled sentences that include implicit aspects are collected. Next, clustering is performed on these sentences so that similar sentences are merged into the same cluster. Finally, the explicit aspects are propagated to the unlabeled sentences in the same cluster, in order to construct a labeled dataset containing implicit aspects. The results of our experiments on mobile phone reviews show that our method of identifying the labels of implicit aspects achieves a maximum accuracy of 82%. | ['Kiyoaki Shirai', 'Aye Aye Mar'] | null | null | null | null | lrec-2022-6 | ['aspect-based-sentiment-analysis'] | ['natural-language-processing'] | [ 1.11539438e-01 2.25388318e-01 -5.04847705e-01 -9.29541588e-01
-5.29713154e-01 -7.33013511e-01 4.63973999e-01 5.67644238e-01
-2.47929841e-01 5.71910858e-01 3.41266274e-01 -2.12125629e-01
4.49685901e-01 -7.82182395e-01 -2.79300570e-01 -6.90964937e-01
5.72469294e-01 3.72228146e-01 1.34258226e-01 -2.87445426e-01
4.71663445e-01 -5.71092106e-02 -1.24843919e+00 5.35415530e-01
7.39270091e-01 7.18163550e-01 2.78947894e-02 2.41591692e-01
-9.58316684e-01 6.77822232e-01 -9.86454487e-01 -7.70536065e-01
-1.51424184e-01 -6.58530056e-01 -7.90326297e-01 7.09417582e-01
5.07323071e-02 2.21788511e-01 6.32253110e-01 1.17193890e+00
4.86911200e-02 -1.72969863e-01 6.95072234e-01 -1.08258772e+00
-3.28398705e-01 6.60141885e-01 -8.22290421e-01 -7.66657665e-02
2.28892073e-01 -4.58997518e-01 1.43582594e+00 -1.01739979e+00
4.34595048e-01 1.05132389e+00 5.33632398e-01 3.20255935e-01
-8.13212812e-01 -5.37509680e-01 9.16759312e-01 -7.69161209e-02
-1.01859498e+00 -7.15413392e-02 9.23052669e-01 -3.42720270e-01
8.75842452e-01 3.21235687e-01 1.01706731e+00 3.33742321e-01
1.82344109e-01 8.96211267e-01 1.06783962e+00 -4.57176566e-01
5.26721239e-01 9.53702927e-01 9.84353602e-01 1.46198288e-01
5.54730237e-01 -9.74705696e-01 -3.02349210e-01 -3.41544777e-01
-3.53669107e-01 -1.12697653e-01 9.50775295e-02 -2.86432266e-01
-6.59616709e-01 9.94260192e-01 4.71867174e-02 5.59688747e-01
-5.46870649e-01 -4.60500807e-01 5.32754958e-01 3.63452658e-02
8.18227887e-01 5.67252934e-01 -8.37899745e-01 1.55261960e-02
-4.66468722e-01 7.07188323e-02 1.10151696e+00 1.12650526e+00
1.32980788e+00 -3.32230985e-01 1.37456909e-01 9.56763625e-01
5.51872432e-01 5.89953542e-01 4.40363199e-01 -4.89204913e-01
4.53770041e-01 1.41669858e+00 7.25166574e-02 -1.17820096e+00
-1.61838353e-01 -3.98295373e-01 -3.30523878e-01 -1.24032505e-01
-2.14947581e-01 -4.02561486e-01 -8.26454580e-01 1.34296191e+00
4.35853392e-01 -3.34220678e-01 3.93976629e-01 6.19852364e-01
9.76548135e-01 7.79287755e-01 3.00738096e-01 -4.15637821e-01
1.74800467e+00 -1.07727766e+00 -1.04192328e+00 -4.51522976e-01
7.21202850e-01 -1.04474699e+00 1.07336795e+00 2.68568516e-01
-5.58503091e-01 -2.98606545e-01 -9.88181233e-01 2.92309314e-01
-6.19307101e-01 2.65540272e-01 6.97602034e-01 7.07215726e-01
-8.20942760e-01 -1.35324329e-01 -3.67988408e-01 -3.41313899e-01
1.65030330e-01 2.12056339e-01 -3.00679088e-01 3.20565790e-01
-9.42183971e-01 4.78933603e-01 7.79059753e-02 4.16286588e-02
-1.35502487e-01 -2.18583986e-01 -1.16828680e+00 4.55056243e-02
4.42099929e-01 -2.84567356e-01 1.40047240e+00 -1.71701872e+00
-1.05004168e+00 9.74057376e-01 -8.77756119e-01 -2.85855439e-02
-4.10910249e-01 -2.95182168e-01 -7.06095695e-01 4.49344963e-02
5.68551362e-01 2.31361270e-01 7.19735324e-01 -1.56323647e+00
-9.22361553e-01 -6.66327953e-01 4.82427657e-01 4.54182088e-01
-4.30917978e-01 1.74425527e-01 -8.31157804e-01 -5.01665831e-01
5.37764020e-02 -1.02791893e+00 -4.30275321e-01 -6.86857104e-01
-5.97856820e-01 -3.86507660e-01 1.01939356e+00 -3.48078787e-01
1.33841765e+00 -2.04707408e+00 -3.32835257e-01 4.04623419e-01
1.12156019e-01 2.58506745e-01 -3.66824307e-02 2.65160918e-01
-1.09184176e-01 2.59374499e-01 -4.72923636e-01 -2.27305502e-01
-2.54819244e-01 2.23372862e-01 -3.37747991e-01 -1.81792036e-01
2.07636625e-01 5.80257773e-01 -8.68544102e-01 -6.08261049e-01
-3.10010374e-01 9.78858992e-02 -4.02239949e-01 7.47836605e-02
-3.07830960e-01 2.88967103e-01 -8.94569993e-01 5.32315850e-01
6.85343146e-01 -1.65625021e-01 3.86879325e-01 -3.35697263e-01
-1.98957026e-01 7.65243709e-01 -8.66313696e-01 9.11098659e-01
-5.82068205e-01 6.66292906e-01 -1.33588789e-02 -9.33915436e-01
1.04753494e+00 4.32175159e-01 5.11134803e-01 -3.83891821e-01
1.85345516e-01 2.34653026e-01 -1.60242151e-02 -4.14245814e-01
8.74515176e-01 -3.41203183e-01 -3.00344437e-01 6.79954231e-01
-1.51876882e-01 -6.05888546e-01 6.65082574e-01 4.38962042e-01
5.64788282e-01 -2.03985974e-01 6.93587005e-01 -2.37303004e-01
1.12326849e+00 3.03630978e-01 6.70106232e-01 2.67576069e-01
-1.40765011e-01 4.25134659e-01 8.19426835e-01 -3.06125581e-01
-9.25515354e-01 -4.57113057e-01 -2.95271650e-02 8.03815067e-01
2.98328072e-01 -1.02216649e+00 -7.41021156e-01 -1.14401150e+00
-4.41795677e-01 8.00712526e-01 -8.27548981e-01 2.58625060e-01
-2.54584968e-01 -7.67143726e-01 -3.14834356e-01 2.90552825e-01
3.88539910e-01 -1.20840812e+00 -8.33886787e-02 1.92733422e-01
-1.94672838e-01 -1.08186638e+00 -4.85528678e-01 9.64289978e-02
-6.84976697e-01 -1.28472984e+00 -4.89701927e-01 -9.40198600e-01
1.09800673e+00 4.54690099e-01 1.12491572e+00 -1.28252394e-02
4.71078277e-01 2.48934150e-01 -7.18237400e-01 -7.19550967e-01
-2.80262262e-01 6.99834302e-02 -2.32517585e-01 2.53157049e-01
1.11773062e+00 -2.48168677e-01 -3.88789445e-01 3.64325106e-01
-7.21843958e-01 -2.86959529e-01 7.00275421e-01 3.22965860e-01
7.85576046e-01 1.90752000e-01 6.65097594e-01 -1.75799584e+00
8.46611381e-01 -4.25271660e-01 -1.66944057e-01 1.11455001e-01
-7.34125972e-01 -2.38128915e-01 3.93443316e-01 -2.13698640e-01
-1.28986609e+00 -1.02650799e-01 -3.74971807e-01 4.12078500e-01
-2.01641127e-01 8.84984970e-01 -4.04574960e-01 5.51018894e-01
2.45169520e-01 6.94767386e-02 -1.71269387e-01 -1.61352754e-01
1.16689704e-01 9.83220041e-01 -2.55247504e-01 -1.13112785e-01
4.06506181e-01 3.31724048e-01 -4.56489474e-01 -9.62009966e-01
-1.56794691e+00 -1.02768517e+00 -5.06696284e-01 -1.71993271e-01
7.81220138e-01 -8.71290863e-01 -3.01526114e-02 4.25183982e-01
-1.02152646e+00 2.58944303e-01 -4.07347769e-01 5.31875789e-01
-1.18240237e-01 4.97565061e-01 -3.68018121e-01 -9.31054771e-01
-5.22371113e-01 -1.26332581e+00 9.33522880e-01 5.43017626e-01
-8.83123279e-01 -8.57280850e-01 2.55500585e-01 4.16461229e-01
1.33392572e-01 -4.29747015e-01 1.01261139e+00 -1.03532016e+00
1.00739850e-02 -5.01583040e-01 1.61607206e-01 6.02378547e-01
8.58978629e-01 1.62804261e-01 -9.48353231e-01 1.60692587e-01
4.51047629e-01 1.18029185e-01 4.08341467e-01 3.67090255e-01
6.39466763e-01 -4.28703487e-01 -3.89461815e-01 -2.44491264e-01
1.13046014e+00 7.25887120e-01 4.59533006e-01 3.89218241e-01
5.51662683e-01 9.71504629e-01 1.18152726e+00 1.05677329e-01
4.00963813e-01 4.20544863e-01 7.11686313e-02 -7.48138353e-02
4.17075306e-01 1.10649042e-01 5.13834059e-01 1.47031617e+00
3.82109344e-01 -1.51178107e-01 -4.86155659e-01 9.64537621e-01
-1.54949653e+00 -7.27954328e-01 -5.76350212e-01 1.90195525e+00
1.04167640e+00 3.49126041e-01 -1.59718040e-02 2.84468830e-01
9.11319196e-01 3.38682026e-01 -3.71425301e-01 -9.72389400e-01
-9.28418636e-02 -5.77406250e-02 -1.67046487e-01 5.39217770e-01
-1.22091770e+00 8.03785264e-01 5.41755342e+00 5.23966134e-01
-8.18105578e-01 1.46749513e-02 8.03677261e-01 3.25896889e-01
-8.15589547e-01 2.79459268e-01 -1.03099668e+00 3.71405721e-01
6.26656473e-01 -3.26489210e-01 -4.13940847e-01 1.24590254e+00
2.82792509e-01 -5.14158368e-01 -7.01862216e-01 3.89574349e-01
4.39318240e-01 -7.78378665e-01 3.61809492e-01 -4.38339636e-02
1.09832036e+00 -3.30502480e-01 -5.28717451e-02 2.93463886e-01
-1.16091572e-01 -3.03798586e-01 4.39023405e-01 4.12456803e-02
3.05945814e-01 -1.02237594e+00 1.36271656e+00 2.26225436e-01
-1.36864495e+00 4.46037203e-01 -3.82534146e-01 -5.38263656e-03
3.14662308e-01 1.13052773e+00 -5.71602285e-01 4.12465841e-01
6.55848861e-01 1.02300429e+00 -5.08646667e-01 7.15521157e-01
-7.19808519e-01 7.20996141e-01 9.09190103e-02 -5.10631025e-01
3.53941292e-01 -7.64915168e-01 5.05109966e-01 1.12424314e+00
1.36001691e-01 8.05467367e-02 4.99883480e-03 3.08145344e-01
-1.01699278e-01 7.92339683e-01 -8.31017971e-01 -2.77675807e-01
-2.29688045e-02 1.59967077e+00 -9.26053226e-01 -7.82154858e-01
-8.74164641e-01 5.98351717e-01 -9.70883220e-02 2.86911637e-01
-3.88874143e-01 -6.15695238e-01 6.46796048e-01 3.39866988e-03
5.54063916e-01 2.45305195e-01 -5.13958097e-01 -1.10635996e+00
3.65031362e-01 -8.26196194e-01 2.26121679e-01 -6.73487604e-01
-1.29403055e+00 1.00113225e+00 -4.04031575e-01 -1.46268165e+00
-1.83161750e-01 -3.94407362e-01 -8.96773040e-01 8.33962977e-01
-1.39262438e+00 -7.14833617e-01 -2.42656440e-01 2.08980560e-01
7.43976235e-01 -7.31818080e-02 8.03231597e-01 1.94100231e-01
-2.67538667e-01 1.50584623e-01 -4.11642134e-01 9.91062149e-02
6.97085738e-01 -1.27137792e+00 2.30891928e-01 6.99079216e-01
5.19500859e-02 1.04275072e+00 7.30503500e-01 -9.20721948e-01
-6.40676141e-01 -9.27533984e-01 1.57026470e+00 -3.60349357e-01
7.80296564e-01 -1.91590980e-01 -8.80663872e-01 7.47817159e-01
5.54548085e-01 -5.73861361e-01 1.30143297e+00 4.78696436e-01
-1.56926334e-01 -1.78906158e-01 -8.84650230e-01 6.53977454e-01
2.19342694e-01 -6.60546780e-01 -9.33268368e-01 1.04256786e-01
7.87669718e-01 2.53692806e-01 -3.66271704e-01 4.64030474e-01
3.21482331e-01 -7.19280899e-01 1.94693238e-01 -4.57966298e-01
5.17025054e-01 -5.78481555e-01 1.16033144e-01 -1.39056253e+00
1.72336832e-01 -3.09039801e-02 2.47464791e-01 1.73601675e+00
9.56538618e-01 -5.89897096e-01 8.00254941e-01 6.71288550e-01
-6.30788058e-02 -8.31346929e-01 -3.04206073e-01 -2.71478385e-01
-3.54182571e-01 -4.80648875e-01 6.34737790e-01 9.14207220e-01
4.43825752e-01 1.02607036e+00 1.41332805e-01 -1.24552108e-01
6.78880364e-02 6.68266237e-01 7.55432308e-01 -1.05650258e+00
1.09964900e-01 -2.13926166e-01 -2.08310395e-01 -8.87178779e-01
2.02042162e-01 -4.30396497e-01 2.58214116e-01 -1.70829690e+00
5.85634887e-01 -5.26816070e-01 -3.68448831e-02 2.87728369e-01
-6.55471444e-01 2.86756814e-01 -1.85318962e-01 1.99818373e-01
-7.27481008e-01 6.24190211e-01 1.10379601e+00 -3.70449871e-01
-4.73729700e-01 6.08126104e-01 -1.21910357e+00 1.17323565e+00
9.85424519e-01 -4.89816636e-01 -5.88128567e-01 1.42422676e-01
4.52985853e-01 -5.50689459e-01 -5.87612331e-01 -5.02729595e-01
-1.42465621e-01 -9.62702259e-02 -1.14910759e-01 -1.00338054e+00
2.55564988e-01 -1.05630326e+00 -2.75574356e-01 2.66250968e-01
-3.21600229e-01 1.61514536e-01 2.43240092e-02 3.97871792e-01
-7.78183579e-01 -6.54859185e-01 3.47824693e-01 -1.25386566e-01
-5.35283446e-01 1.05751492e-01 -8.97602022e-01 1.11426905e-01
1.09655428e+00 -5.32383285e-02 -1.04539752e-01 -6.55642688e-01
-7.61088729e-01 1.89334244e-01 4.26845610e-01 3.76513034e-01
4.44227874e-01 -1.07300270e+00 -3.37872118e-01 -1.36450261e-01
5.01683414e-01 8.94522015e-03 6.53414056e-03 7.51379013e-01
-1.79857105e-01 4.88848567e-01 3.20254415e-01 -4.32563007e-01
-1.54351068e+00 6.61014616e-01 -5.34752570e-02 -3.81368876e-01
-2.31874034e-01 4.40764397e-01 6.34231269e-01 -7.29325354e-01
-1.68406919e-01 -1.35217682e-01 -1.06938601e+00 5.79966009e-01
6.10451877e-01 -2.22179964e-01 8.34638476e-02 -1.13741076e+00
-4.25555050e-01 8.20551813e-01 -3.80443990e-01 -2.31863230e-01
1.26970947e+00 -3.27697277e-01 -6.17814660e-01 9.41347837e-01
1.19747460e+00 7.28704214e-01 -4.18517232e-01 -2.46584043e-01
8.58687311e-02 -2.13149726e-01 -4.11294311e-01 -5.58900416e-01
-1.25522757e+00 6.22621834e-01 -4.49881107e-02 4.93953824e-01
1.07241499e+00 3.64438206e-01 5.37679255e-01 4.48515981e-01
1.93462491e-01 -1.24556935e+00 -3.70006859e-02 6.69349790e-01
6.44057631e-01 -1.14515662e+00 1.76989019e-01 -8.64829540e-01
-1.24924648e+00 7.04246104e-01 7.97014117e-01 1.32298142e-01
8.14659178e-01 1.97875977e-01 5.89509785e-01 -4.96495366e-01
-6.45552695e-01 -1.93380043e-01 2.72470713e-01 4.03241277e-01
7.06678033e-01 6.12356607e-03 -1.04266489e+00 8.80139470e-01
-2.29478657e-01 -4.70889062e-01 7.54819751e-01 1.05840707e+00
-4.02640313e-01 -1.35951519e+00 6.95421221e-03 5.92542052e-01
-8.26884806e-01 -2.56205350e-01 -8.19222391e-01 5.00997007e-01
-2.15789247e-02 1.30614102e+00 8.59004557e-02 -1.95641905e-01
3.80066991e-01 1.24268167e-01 -3.52425247e-01 -1.17466331e+00
-8.47765386e-01 2.34306410e-01 4.96939898e-01 -2.59323835e-01
-9.38406408e-01 -6.74403310e-01 -1.38719344e+00 2.15607628e-01
-5.18528521e-01 1.05072987e+00 8.50267112e-01 1.24183071e+00
1.03467524e-01 3.12465757e-01 9.81646478e-01 -2.38521829e-01
2.95663953e-01 -9.73832190e-01 -6.50525451e-01 4.15019035e-01
2.56807745e-01 -2.20433578e-01 -5.72242916e-01 7.49118626e-02] | [11.356721878051758, 6.705973148345947] |
9b698e9a-12f2-4a92-bf93-1cf9e3e16813 | laplacian-convolutional-representation-for | 2212.01529 | null | https://arxiv.org/abs/2212.01529v2 | https://arxiv.org/pdf/2212.01529v2.pdf | Laplacian Convolutional Representation for Traffic Time Series Imputation | Spatiotemporal traffic data imputation is of great significance in intelligent transportation systems and data-driven decision-making processes. To make an accurate reconstruction from partially observed traffic data, we assert the importance of characterizing both global and local trends in traffic time series. In the literature, substantial prior works have demonstrated the effectiveness of utilizing low-rankness property of traffic data by matrix/tensor completion models. In this study, we first introduce a Laplacian kernel to temporal regularization for characterizing local trends in traffic time series, which can be formulated in the form of circular convolution. Then, we develop a low-rank Laplacian convolutional representation (LCR) model by putting the nuclear norm of a circulant matrix and the Laplacian temporal regularization together, which is proved to meet a unified framework that takes a fast Fourier transform (FFT) solution in a relatively low time complexity. Through extensive experiments on some traffic datasets, we demonstrate the superiority of LCR for imputing traffic time series of various time series behaviors (e.g., data noises and strong/weak periodicity). The proposed LCR model is an efficient and effective solution to large-scale traffic data imputation over the existing baseline models. Despite the LCR's application to time series data, the key modeling idea lies in bridging the low-rank models and the Laplacian regularization through FFT, which is also applicable to image inpainting. The adapted datasets and Python implementation are publicly available at https://github.com/xinychen/transdim. | ['Lijun Sun', 'Nicolas Saunier', 'Zhanhong Cheng', 'Xinyu Chen'] | 2022-12-03 | null | null | null | null | ['image-inpainting', 'traffic-data-imputation'] | ['computer-vision', 'time-series'] | [ 1.41185939e-01 -7.23020732e-01 -2.23699376e-01 -2.32312486e-01
-6.84180260e-01 -2.88688868e-01 4.54576552e-01 -5.15638232e-01
-2.25852340e-01 6.45552278e-01 4.01734710e-01 -5.99231839e-01
-4.59782034e-01 -6.76318347e-01 -7.96357214e-01 -8.65496933e-01
1.14401698e-01 -5.38816163e-03 -8.77906978e-02 -2.65504748e-01
8.47800523e-02 7.02416301e-01 -1.63342142e+00 1.88438743e-01
1.16416001e+00 1.01106882e+00 2.58033741e-02 2.27146357e-01
-2.36209661e-01 9.28931057e-01 -1.38639301e-01 -4.02822167e-01
3.81772488e-01 3.64441872e-02 -3.01918179e-01 2.18830541e-01
3.33824575e-01 -1.84761658e-01 -1.00684488e+00 9.42044079e-01
2.25753129e-01 3.47744316e-01 4.77411985e-01 -1.73701441e+00
-7.60329127e-01 1.40525430e-01 -8.17666650e-01 2.98944175e-01
-6.35452643e-02 2.84673482e-01 5.96638024e-01 -9.50879753e-01
2.48302683e-01 1.08039796e+00 8.40475738e-01 1.21158056e-01
-1.33775353e+00 -7.85598338e-01 5.28947003e-02 5.20657897e-01
-1.59328258e+00 -5.21922112e-01 1.16755867e+00 -6.07086480e-01
5.23406506e-01 4.76178706e-01 1.99457482e-01 1.10698736e+00
1.10396989e-01 7.12976038e-01 1.15715849e+00 2.25155309e-01
-1.59739897e-01 -3.08749914e-01 2.26716071e-01 3.24660599e-01
2.20032945e-01 8.29712749e-02 -2.59169996e-01 -1.24661148e-01
6.23922706e-01 5.13208330e-01 -2.36936621e-02 8.83303285e-02
-1.28703642e+00 6.20914817e-01 1.84800893e-01 9.91308615e-02
-5.30995905e-01 3.07501614e-01 6.02184296e-01 2.46009618e-01
6.92724586e-01 -4.82054055e-01 -2.09339708e-01 -1.63822845e-01
-8.65503669e-01 1.04652315e-01 2.84588814e-01 8.80634129e-01
9.32858706e-01 3.61702383e-01 -3.99387747e-01 9.26105320e-01
7.57306814e-02 9.75514710e-01 2.06556112e-01 -1.08290136e+00
8.58806908e-01 3.53190601e-01 1.30465487e-04 -1.42082953e+00
-2.97406018e-01 -3.91940743e-01 -1.40440166e+00 -1.71083078e-01
4.32733506e-01 -1.91850916e-01 -4.88168776e-01 1.82861209e+00
2.00256750e-01 9.57117558e-01 -5.20063788e-02 9.46782887e-01
5.86229563e-01 8.14080775e-01 -5.07065207e-02 -3.15176785e-01
1.26377106e+00 -5.25501192e-01 -1.06190610e+00 1.48472041e-01
4.78979647e-01 -9.25262272e-01 1.00053179e+00 -3.03171412e-03
-6.78258300e-01 -6.63329899e-01 -5.35930216e-01 -2.94231087e-01
-3.93142790e-01 5.26894450e-01 6.38881683e-01 4.17263776e-01
-7.11352944e-01 2.51971304e-01 -7.59438932e-01 -2.10501596e-01
3.35377306e-01 5.49975075e-02 -4.27145839e-01 -2.44541392e-01
-1.19651210e+00 5.32725275e-01 -2.81467199e-01 6.88830972e-01
-5.59301436e-01 -9.80933309e-01 -6.72225237e-01 -1.80548787e-01
4.28594887e-01 -4.77188617e-01 6.14037454e-01 -5.27999163e-01
-1.09238958e+00 3.73916686e-01 -5.79411983e-01 -3.64986598e-01
7.00999975e-01 -1.57098278e-01 -8.78994167e-01 -1.01606719e-01
3.16055059e-01 2.26119589e-02 9.70229268e-01 -9.21017587e-01
-2.95506001e-01 -2.97801286e-01 -2.27901742e-01 -2.76370734e-01
-3.80630910e-01 -7.96486661e-02 -5.36632180e-01 -1.03254664e+00
1.52987644e-01 -9.52522457e-01 -3.09626341e-01 3.13456878e-02
-3.26210529e-01 -1.63190082e-01 1.08100641e+00 -9.44109797e-01
1.40874696e+00 -2.19114161e+00 -2.71005750e-01 2.91075587e-01
1.51672915e-01 1.74918190e-01 -2.99447626e-01 6.38617635e-01
-3.98844630e-01 -1.20695509e-01 -4.09248173e-01 -2.89053857e-01
1.11983985e-01 2.48257875e-01 -7.63897121e-01 7.00717568e-01
3.34423721e-01 1.00950980e+00 -6.57908916e-01 -3.06374997e-01
4.53726977e-01 7.84520864e-01 -3.74443769e-01 -6.11745939e-02
1.90262720e-01 7.18235433e-01 -5.34308016e-01 4.46977973e-01
1.13777208e+00 -2.28963438e-02 -3.48862827e-01 -7.97957480e-01
-5.43593228e-01 9.30231437e-02 -1.25379002e+00 1.51213121e+00
-5.25779128e-01 7.38534153e-01 1.00118235e-01 -1.19905853e+00
8.36847901e-01 2.26222843e-01 9.49954391e-01 -9.41341460e-01
-3.02624330e-02 1.42412424e-01 -3.22868526e-01 -8.66274476e-01
4.69060987e-01 1.36860937e-01 1.06309660e-01 4.09902364e-01
-5.49852192e-01 5.22657692e-01 4.16540742e-01 1.14841007e-01
1.06014144e+00 7.06147626e-02 -4.10783738e-01 -7.76259527e-02
8.56440485e-01 -7.76716098e-02 7.45183051e-01 4.47965711e-01
3.17774713e-03 6.04579747e-01 4.12645817e-01 -4.96313751e-01
-1.04022050e+00 -9.38864410e-01 -2.65450448e-01 8.60493839e-01
-1.33228227e-01 -2.32907534e-01 -3.40780348e-01 -2.33535066e-01
1.63502293e-03 5.32421887e-01 -4.53490496e-01 -7.75897205e-02
-8.56359124e-01 -9.52908158e-01 6.50991976e-01 3.13114882e-01
6.94916785e-01 -6.17355406e-01 1.51862815e-01 1.38245031e-01
-7.37652838e-01 -1.41001427e+00 -8.84324729e-01 -4.02888477e-01
-8.10438991e-01 -8.58274162e-01 -5.66163421e-01 -4.04541999e-01
7.30522096e-01 9.19787705e-01 7.13116348e-01 -7.53162354e-02
-2.87271291e-01 3.84587586e-01 -2.09797576e-01 1.66187719e-01
3.85877676e-02 -1.15468703e-01 1.50764197e-01 7.96955347e-01
3.99417311e-01 -9.84531820e-01 -5.68248808e-01 5.48602402e-01
-1.11977649e+00 2.21905231e-01 5.71894586e-01 5.88744044e-01
5.46053767e-01 2.60765642e-01 5.47275722e-01 -5.68334818e-01
7.06262171e-01 -8.51056814e-01 -5.92770934e-01 1.15484476e-01
-2.55608410e-01 -1.24356136e-01 7.40967810e-01 -5.18862724e-01
-1.12346053e+00 -5.48539199e-02 -1.86012477e-01 -8.90198231e-01
-1.03049427e-01 6.72713757e-01 -1.40146792e-01 4.75811251e-02
2.36566752e-01 6.06459498e-01 1.83971688e-01 -5.62747598e-01
4.37065989e-01 4.83298391e-01 4.86203015e-01 -9.22368526e-01
1.20584381e+00 8.40184212e-01 2.59692401e-01 -1.08724618e+00
-5.83214939e-01 -5.04275680e-01 -5.98914504e-01 -4.17573750e-01
6.80822611e-01 -1.02910650e+00 -9.86550868e-01 4.31959093e-01
-1.01292682e+00 6.50657117e-02 -2.63734519e-01 6.62714779e-01
-3.96647483e-01 7.40066230e-01 -4.12391096e-01 -8.20875943e-01
8.72990191e-02 -1.01358855e+00 8.51340294e-01 -2.47834876e-01
3.74787718e-01 -9.44200933e-01 -9.90136191e-02 6.57947540e-01
5.87259650e-01 2.80768126e-01 6.79170072e-01 -1.22871630e-01
-7.79558420e-01 -1.82948351e-01 -4.78716105e-01 4.46043104e-01
1.45894736e-01 4.94094857e-04 -7.78161466e-01 -5.43116257e-02
1.09845279e-02 1.56073406e-01 1.01219380e+00 4.43869799e-01
1.45838332e+00 -4.19072300e-01 -7.11321039e-03 7.69499183e-01
1.20597684e+00 -1.49041206e-01 9.86566305e-01 -1.20099679e-01
1.01803362e+00 7.71957874e-01 5.30690730e-01 5.19085705e-01
5.80436468e-01 7.23371387e-01 8.33469927e-02 -1.69537529e-01
-1.96970403e-01 -2.77911603e-01 5.62449694e-01 1.30507469e+00
-2.67580688e-01 9.78177413e-02 -8.01829934e-01 5.56249797e-01
-2.21942616e+00 -1.37395871e+00 -7.71392226e-01 2.22243643e+00
4.07970995e-01 -2.00296089e-01 1.54209182e-01 7.63973221e-02
8.30678761e-01 3.15541804e-01 -5.78355074e-01 -3.25078815e-02
-3.56705815e-01 -1.06050797e-01 7.74719536e-01 3.60548079e-01
-1.07596874e+00 6.81635141e-01 5.51298761e+00 1.26592851e+00
-1.03341532e+00 3.00730228e-01 4.68795568e-01 1.91887110e-01
-3.82158846e-01 -3.75846736e-02 -4.65952039e-01 7.54811883e-01
1.08201885e+00 -2.39238992e-01 7.01309025e-01 2.96201855e-01
1.07203639e+00 3.55258316e-01 -5.30217886e-01 1.13197422e+00
-1.58040717e-01 -1.35200334e+00 1.49688721e-01 2.50303745e-01
6.60021007e-01 1.00340165e-01 3.56944948e-01 2.06945822e-01
2.53879633e-02 -8.45012963e-01 3.68639171e-01 8.93421888e-01
8.05352509e-01 -7.20644712e-01 4.60635066e-01 3.18855315e-01
-1.65233457e+00 -1.03920765e-01 -4.76932943e-01 -1.46557286e-01
4.26552773e-01 9.14614439e-01 -8.61610696e-02 8.47152472e-01
5.49875021e-01 1.16414142e+00 -4.55523521e-01 8.91874790e-01
1.74453244e-01 7.79300570e-01 -3.22783619e-01 6.20911837e-01
1.01115674e-01 -1.01268888e+00 5.30165195e-01 1.12280500e+00
4.96975332e-01 1.54619843e-01 2.47900531e-01 8.94098639e-01
1.18599730e-02 9.18995067e-02 -7.24012554e-01 -8.89900327e-03
3.67978871e-01 1.32581520e+00 -3.40286821e-01 -2.10858993e-02
-8.29763114e-01 6.32746041e-01 4.23192009e-02 8.01056564e-01
-1.25753868e+00 -2.35367373e-01 9.81406093e-01 2.39701703e-01
1.66989177e-01 -6.78927183e-01 -3.02207530e-01 -1.41928291e+00
4.18829948e-01 -6.83424532e-01 1.24571212e-01 -6.82335556e-01
-1.67180765e+00 3.07551950e-01 3.32599692e-02 -1.85498691e+00
2.20693067e-01 -4.76641774e-01 -7.42203712e-01 7.70640850e-01
-1.63942182e+00 -1.50310636e+00 -3.93545806e-01 1.00339484e+00
3.72658998e-01 -1.33488834e-01 2.38461420e-01 1.03007138e+00
-1.02885020e+00 3.57017726e-01 3.79983008e-01 1.68363199e-01
5.00881553e-01 -5.64494669e-01 3.19322854e-01 1.06439066e+00
-1.99527562e-01 7.66863585e-01 5.75551808e-01 -6.96600020e-01
-1.81580889e+00 -1.62124860e+00 8.22137773e-01 -2.25092322e-01
1.08115780e+00 -2.00829998e-01 -9.69232023e-01 6.97528839e-01
-1.29220277e-01 2.47295931e-01 6.63885057e-01 -1.53539658e-01
-5.26688457e-01 -6.91072941e-01 -8.70776057e-01 6.86160862e-01
1.17198098e+00 -7.52305388e-01 -2.62589097e-01 4.57535237e-01
6.27241671e-01 4.68975455e-02 -8.85624170e-01 4.08109754e-01
4.61559117e-01 -6.04729831e-01 1.21743834e+00 -5.46825767e-01
2.51442403e-01 -8.97829592e-01 -3.39785337e-01 -6.81550562e-01
-3.67002815e-01 -8.39963198e-01 -1.16451241e-01 1.49890697e+00
1.13383010e-01 -7.29573846e-01 5.12479663e-01 6.07381821e-01
-1.84011430e-01 -3.70087177e-01 -1.18211663e+00 -7.94235528e-01
-9.51164737e-02 -1.03689694e+00 5.54755688e-01 9.27955687e-01
-5.68576455e-01 1.19291037e-01 -1.03244925e+00 4.18558419e-01
9.38367248e-01 4.51450944e-02 9.78765428e-01 -1.08052492e+00
2.80550420e-01 -2.41218239e-01 -3.34623337e-01 -1.19735885e+00
4.90144491e-01 -1.00549340e+00 -2.75527090e-01 -1.28494906e+00
9.19717848e-02 -4.35638011e-01 -4.11851585e-01 2.96227336e-01
1.55823618e-01 3.65231454e-01 -2.29730159e-02 4.75684702e-01
-4.38873500e-01 9.00189102e-01 1.12860215e+00 -2.42091045e-01
3.94312516e-02 1.80311456e-01 -5.19160688e-01 4.34303284e-01
6.88273787e-01 -3.41108143e-01 -4.92056578e-01 -5.63546538e-01
2.01239839e-01 4.98554111e-02 6.44215465e-01 -9.53601956e-01
5.41048884e-01 -3.51665974e-01 2.74999756e-02 -8.81579876e-01
3.27634633e-01 -1.22213006e+00 6.24943733e-01 1.27051726e-01
-1.71338052e-01 2.50381380e-01 1.70495674e-01 8.08558226e-01
-2.43573040e-01 3.10114831e-01 4.00411189e-01 2.95730531e-01
-7.00835645e-01 7.24422812e-01 -5.32052934e-01 8.21827445e-03
8.09009552e-01 -2.19599724e-01 -3.91279399e-01 -3.96987498e-01
-5.19467056e-01 2.90537238e-01 2.64550336e-02 6.11197352e-01
6.86238170e-01 -1.81356323e+00 -9.39403892e-01 4.28942859e-01
2.08756458e-02 -5.66673815e-01 8.13395619e-01 1.46592093e+00
-2.56587446e-01 3.33792120e-01 -9.46395546e-02 -6.10206008e-01
-7.95708537e-01 6.17292285e-01 4.23984453e-02 2.52194554e-02
-6.86211884e-01 -8.12822580e-03 2.31025696e-01 -5.09381950e-01
4.57423888e-02 -2.61305243e-01 -9.09612179e-02 8.09599534e-02
4.98691678e-01 8.37435663e-01 7.07699135e-02 -1.17130876e+00
-2.87541419e-01 9.94384646e-01 3.12514514e-01 1.28912613e-01
1.40364242e+00 -4.24410045e-01 -4.04995322e-01 4.29831445e-01
1.48752296e+00 2.27631345e-01 -1.36036491e+00 -4.03125584e-01
-1.09665193e-01 -5.81713378e-01 -2.51521599e-02 -1.41336858e-01
-1.39566123e+00 9.02860940e-01 4.94021297e-01 7.73300678e-02
1.22361422e+00 -4.56328750e-01 1.04715240e+00 3.86801153e-01
1.50312036e-01 -8.57293308e-01 -2.08758354e-01 6.22193396e-01
9.60014880e-01 -1.11201847e+00 -7.52758160e-02 -5.38151979e-01
-5.44215322e-01 1.03129399e+00 2.95013785e-01 -1.87181354e-01
9.04890776e-01 1.47038773e-01 -8.06121603e-02 -1.88307418e-03
-6.91173553e-01 -5.95010877e-01 3.88168484e-01 5.02696753e-01
2.95662493e-01 5.38374931e-02 -3.97110820e-01 4.76239324e-01
4.69318591e-02 4.01628800e-02 2.96489477e-01 4.22994614e-01
-2.23883186e-02 -9.89214778e-01 -4.92328078e-01 3.76927346e-01
-3.01034123e-01 2.52387077e-02 2.21097499e-01 4.24517602e-01
2.01110289e-01 1.26431978e+00 1.12954043e-01 -5.23906648e-01
4.80791062e-01 -1.14424184e-01 -1.47925362e-01 1.14945844e-01
-2.13432014e-01 1.30854130e-01 -2.59558689e-02 -7.18862355e-01
-5.82653642e-01 -7.44237959e-01 -8.33977699e-01 -8.45112860e-01
7.64057934e-02 7.29949176e-02 4.61952776e-01 8.32977891e-01
6.43216848e-01 5.16353488e-01 8.72566700e-01 -6.97146654e-01
-2.06948057e-01 -6.69473648e-01 -4.58548278e-01 7.81913936e-01
4.72298324e-01 -6.65207922e-01 -4.47971165e-01 1.80354521e-01] | [6.570418834686279, 2.1183528900146484] |
c37cac19-c806-47a8-b41d-649c74f8b00a | museformer-transformer-with-fine-and-coarse | 2210.10349 | null | https://arxiv.org/abs/2210.10349v2 | https://arxiv.org/pdf/2210.10349v2.pdf | Museformer: Transformer with Fine- and Coarse-Grained Attention for Music Generation | Symbolic music generation aims to generate music scores automatically. A recent trend is to use Transformer or its variants in music generation, which is, however, suboptimal, because the full attention cannot efficiently model the typically long music sequences (e.g., over 10,000 tokens), and the existing models have shortcomings in generating musical repetition structures. In this paper, we propose Museformer, a Transformer with a novel fine- and coarse-grained attention for music generation. Specifically, with the fine-grained attention, a token of a specific bar directly attends to all the tokens of the bars that are most relevant to music structures (e.g., the previous 1st, 2nd, 4th and 8th bars, selected via similarity statistics); with the coarse-grained attention, a token only attends to the summarization of the other bars rather than each token of them so as to reduce the computational cost. The advantages are two-fold. First, it can capture both music structure-related correlations via the fine-grained attention, and other contextual information via the coarse-grained attention. Second, it is efficient and can model over 3X longer music sequences compared to its full-attention counterpart. Both objective and subjective experimental results demonstrate its ability to generate long music sequences with high quality and better structures. | ['Tie-Yan Liu', 'Tao Qin', 'Shikun Zhang', 'Wei Ye', 'Xu Tan', 'Wei Hu', 'Rui Wang', 'Peiling Lu', 'Botao Yu'] | 2022-10-19 | null | null | null | null | ['music-generation', 'music-generation'] | ['audio', 'music'] | [ 2.16667801e-01 -2.20344335e-01 9.51913521e-02 3.42037588e-01
-7.86776066e-01 -4.78368074e-01 2.06070080e-01 9.21257511e-02
1.71782188e-02 7.39280939e-01 6.40647829e-01 1.56856388e-01
-3.91238749e-01 -7.93400824e-01 -5.34279883e-01 -6.88200295e-01
2.05458961e-02 4.33405429e-01 1.16777979e-01 -4.66624498e-01
5.26614666e-01 5.57215605e-03 -1.63291931e+00 3.69825870e-01
1.00654173e+00 8.21601987e-01 5.30118346e-01 5.88601410e-01
-2.30232701e-01 5.95378578e-01 -8.60328317e-01 -2.97613114e-01
-4.55024242e-02 -8.53497624e-01 -7.43777037e-01 -3.71462822e-01
3.20302397e-01 -2.90629297e-01 -3.17852162e-02 9.43500996e-01
7.14193821e-01 2.88396150e-01 3.83301407e-01 -1.01219225e+00
-6.54752135e-01 1.28181386e+00 -6.42698765e-01 1.59708440e-01
4.86228377e-01 -8.86958614e-02 1.49949360e+00 -8.06887448e-01
2.49223068e-01 1.14794707e+00 6.45292401e-01 4.29337174e-01
-1.03554630e+00 -8.02730918e-01 1.42348424e-01 2.63534784e-01
-1.51802349e+00 -2.14629754e-01 8.77951026e-01 -3.38152379e-01
7.89014459e-01 3.69933277e-01 8.42585266e-01 9.01736677e-01
4.59522456e-02 7.02351332e-01 5.73679209e-01 -4.56210732e-01
-3.71556869e-03 -5.71183026e-01 -2.06650853e-01 3.32939446e-01
1.60909757e-01 -1.63806584e-02 -7.69645572e-01 -2.05313176e-01
9.77959335e-01 -2.06505865e-01 -1.33741722e-01 2.79118389e-01
-1.55168140e+00 7.24539578e-01 3.68389100e-01 6.37449026e-01
-6.98321283e-01 3.56948197e-01 3.43519151e-01 -1.26280442e-01
3.44151743e-02 8.60059083e-01 -2.21645042e-01 -5.34748197e-01
-1.23343837e+00 5.64129472e-01 3.48416388e-01 9.59403515e-01
6.61805093e-01 3.18204105e-01 -8.14437747e-01 9.42730069e-01
7.68488348e-02 1.17098697e-01 7.70669699e-01 -8.84121954e-01
5.43216705e-01 4.03132200e-01 2.04323847e-02 -9.68950629e-01
-5.68429567e-02 -6.69255733e-01 -1.12309670e+00 -1.19391549e-02
1.44895747e-01 -1.77334864e-02 -6.80689216e-01 1.80457211e+00
-7.60622397e-02 4.86864328e-01 -1.38014436e-01 9.12801087e-01
8.56923580e-01 7.43996561e-01 -1.11107603e-01 -2.86686748e-01
1.43548989e+00 -9.90583897e-01 -7.03141510e-01 1.40937893e-02
-4.79868948e-02 -1.12455487e+00 1.38883483e+00 3.31893981e-01
-1.38446200e+00 -1.04261184e+00 -1.08145797e+00 -4.54201810e-02
1.94260970e-01 1.37668058e-01 5.28658032e-01 8.43008533e-02
-8.24789941e-01 8.32917452e-01 -3.16090822e-01 8.58358219e-02
1.73937604e-01 2.70052046e-01 2.80909061e-01 3.21777821e-01
-1.36004150e+00 2.58055985e-01 5.61230361e-01 -2.73295715e-02
-6.30284309e-01 -7.03866899e-01 -6.08418465e-01 5.69851577e-01
2.28830248e-01 -8.33980441e-01 1.31836581e+00 -7.96797931e-01
-1.53889561e+00 2.92048573e-01 -2.58145720e-01 -3.62244099e-01
6.88421950e-02 -3.61128777e-01 -6.98152259e-02 -2.13413890e-02
2.46714503e-01 7.14685917e-01 6.63468421e-01 -9.77704465e-01
-6.49316430e-01 4.26307172e-02 1.46680415e-01 3.68940741e-01
-3.72917384e-01 -1.66361723e-02 -6.57719672e-01 -1.37620854e+00
1.81255117e-01 -9.52647269e-01 -1.55209288e-01 -7.18000114e-01
-7.63721347e-01 -2.14162558e-01 5.52962899e-01 -4.87542093e-01
1.76511705e+00 -2.27750731e+00 3.66177112e-01 6.08980805e-02
-7.78608322e-02 3.24366063e-01 -3.18767220e-01 6.57892764e-01
-9.02464613e-02 2.17100427e-01 -9.83675867e-02 -4.39261682e-02
5.46849631e-02 -7.54456520e-02 -4.47413832e-01 -5.08364856e-01
1.97065994e-02 1.01245642e+00 -1.05245852e+00 -4.10887301e-01
-6.09611757e-02 2.88984299e-01 -8.34045887e-01 1.06419757e-01
-2.38525838e-01 5.22259653e-01 -5.44209182e-01 3.34742546e-01
1.07937038e-01 -1.70887351e-01 -1.55814484e-01 -3.17503095e-01
-1.83169529e-01 6.31094456e-01 -1.30992401e+00 1.77032697e+00
-3.58158767e-01 2.67840892e-01 -3.85393500e-01 -3.67546320e-01
1.03877914e+00 5.48254013e-01 4.78797019e-01 -6.07299805e-01
-1.65378869e-01 2.24989057e-01 2.44064271e-01 -1.79408297e-01
8.69396031e-01 -2.95585901e-01 -3.66038680e-01 6.45513535e-01
-2.34361589e-01 -9.06252861e-02 5.23553491e-01 2.35331416e-01
9.04280245e-01 9.69823077e-02 3.79527897e-01 6.28890693e-02
4.56847191e-01 -4.12060291e-01 8.41357231e-01 7.37731695e-01
2.40250811e-01 8.26169372e-01 4.58505064e-01 -1.65304780e-01
-9.91870880e-01 -9.82146561e-01 4.34340864e-01 1.29385626e+00
1.15148246e-01 -8.93607080e-01 -8.12748194e-01 -1.86975613e-01
-1.18442841e-01 6.75431728e-01 -4.72380817e-01 -3.36274385e-01
-6.60830081e-01 -4.17688876e-01 7.34869540e-01 7.15058804e-01
4.40699130e-01 -1.73251879e+00 -5.15964270e-01 6.74215853e-01
-6.41463757e-01 -4.45467740e-01 -1.00517023e+00 -8.42510238e-02
-8.20171118e-01 -8.57392430e-01 -9.28058922e-01 -7.37506330e-01
2.16548100e-01 7.14960098e-02 1.23971808e+00 1.73499808e-01
-6.57404587e-02 -2.43902624e-01 -6.59704089e-01 -3.57623547e-01
-7.35033453e-02 3.61086398e-01 -1.21262439e-01 6.84113428e-02
1.86585113e-02 -7.14180887e-01 -4.26905900e-01 1.09322481e-01
-7.69343019e-01 2.10734636e-01 7.85018802e-01 1.00827348e+00
7.85722733e-01 1.97606966e-01 8.83890092e-01 -6.51364207e-01
9.67958927e-01 -7.41799101e-02 -1.31359622e-01 -5.98270334e-02
-1.98952064e-01 5.76150138e-03 7.67347217e-01 -4.62018043e-01
-7.74742723e-01 -1.77284852e-01 -2.37664506e-01 -4.14065778e-01
3.91511023e-02 7.65996933e-01 -1.61332399e-01 5.55770636e-01
4.25375164e-01 4.00054783e-01 -5.00449121e-01 -5.66937864e-01
1.51238412e-01 3.57003987e-01 6.51790142e-01 -8.05894613e-01
8.30223560e-01 -1.07099183e-01 -2.24936917e-01 -5.45813501e-01
-9.19622898e-01 -1.47665486e-01 -5.09905696e-01 -4.52003404e-02
8.28245699e-01 -7.25079238e-01 -6.53322577e-01 3.82307291e-01
-9.90048587e-01 -1.83965802e-01 -6.48513019e-01 4.93948609e-01
-6.83982432e-01 3.39654535e-01 -8.28094959e-01 -6.78834915e-01
-7.00853407e-01 -1.00005817e+00 1.19430256e+00 3.68289173e-01
-6.93964899e-01 -4.26680356e-01 9.95339709e-04 7.74781704e-02
2.43910566e-01 2.60512352e-01 1.26720250e+00 -2.19716132e-01
-6.09535456e-01 2.79530566e-02 -2.50697881e-02 1.72897950e-01
3.08682501e-01 1.55285448e-01 -5.74030042e-01 -1.69744432e-01
-5.01770377e-01 -1.49224371e-01 7.10916162e-01 4.48709458e-01
1.25559175e+00 -3.28472197e-01 -4.06679511e-02 4.75793600e-01
1.04866970e+00 4.78630513e-01 7.78827667e-01 2.62675166e-01
7.86383390e-01 2.63919383e-01 7.40522623e-01 6.59614682e-01
1.14698201e-01 1.01381874e+00 2.30313435e-01 -6.65232725e-03
-2.85523266e-01 -6.25589788e-01 2.94499069e-01 1.20203352e+00
-4.71139133e-01 -1.15834713e-01 -4.85590309e-01 6.14968777e-01
-1.83191907e+00 -1.34771132e+00 -7.96198323e-02 2.37946963e+00
8.77890468e-01 1.11163184e-01 2.44313970e-01 5.79173386e-01
7.69673645e-01 2.22292066e-01 -4.77634877e-01 -3.49301577e-01
-6.46075010e-02 5.43084204e-01 -2.70734280e-01 1.42640591e-01
-7.36281037e-01 1.04822719e+00 5.97568178e+00 1.05701828e+00
-7.87535369e-01 -2.41137624e-01 2.78286576e-01 -3.75332952e-01
-4.95519906e-01 -3.07616275e-02 -6.96275830e-01 6.44089818e-01
5.92936814e-01 -4.26674604e-01 6.53187454e-01 5.23224354e-01
3.46195102e-01 2.16972902e-01 -1.00854015e+00 8.60889971e-01
-5.62299192e-02 -1.34469295e+00 5.25076210e-01 8.23990479e-02
9.98181283e-01 -4.71881628e-01 7.96428844e-02 3.86347055e-01
6.90217167e-02 -1.02012110e+00 1.18341696e+00 5.32220602e-01
8.91703010e-01 -1.23578739e+00 7.69624949e-01 2.42028013e-01
-1.67999506e+00 -6.76688403e-02 -3.44102949e-01 -1.29260570e-01
3.26635867e-01 5.46350718e-01 -6.30032599e-01 6.90118730e-01
6.29173636e-01 7.00000703e-01 -1.94461986e-01 1.12025583e+00
-3.88481528e-01 7.14459538e-01 3.90983447e-02 -1.64836203e-03
3.95166337e-01 -1.84646249e-01 6.58884585e-01 1.07787752e+00
9.66652155e-01 1.55489102e-01 2.73436219e-01 8.83626282e-01
1.07521657e-02 2.80476809e-01 -1.69574738e-01 -7.80781880e-02
6.54474378e-01 9.66764092e-01 -4.35172677e-01 -4.88759756e-01
-1.23611316e-02 9.22697067e-01 2.40096480e-01 2.84784257e-01
-8.22955847e-01 -8.35190713e-01 6.45243526e-01 1.88509464e-01
5.91552198e-01 -5.18706515e-02 -4.01323140e-01 -7.63040960e-01
-1.25189140e-01 -1.10536909e+00 3.82504076e-01 -1.01669466e+00
-1.18772459e+00 7.75266767e-01 -3.25308561e-01 -1.36909616e+00
-4.40824270e-01 8.07198733e-02 -9.98383105e-01 1.22297382e+00
-9.52934504e-01 -8.93266737e-01 -2.23139882e-01 4.95996147e-01
6.69491470e-01 -1.12097286e-01 9.64270413e-01 2.56472528e-01
-3.00584614e-01 7.34962523e-01 -1.24318302e-01 1.52903527e-01
5.84386945e-01 -1.28628063e+00 3.84531826e-01 5.42510331e-01
5.60727596e-01 7.95308530e-01 4.89571184e-01 -6.02374911e-01
-8.17294955e-01 -1.05125308e+00 1.24842703e+00 8.79693702e-02
3.29672962e-01 -2.89504230e-02 -8.82328033e-01 3.42080146e-01
1.55772477e-01 -6.26734018e-01 7.98280954e-01 4.10636008e-01
-2.22518414e-01 -1.85065314e-01 -6.07523620e-01 8.12536478e-01
1.08126497e+00 -3.78411800e-01 -8.07468593e-01 -1.24042451e-01
6.82543576e-01 -3.70694935e-01 -8.18112314e-01 4.82338190e-01
6.68568015e-01 -1.03429210e+00 9.37207758e-01 -2.85905749e-01
7.05946743e-01 -6.85573041e-01 -3.52691859e-02 -1.44807708e+00
-9.27391529e-01 -8.21638763e-01 -1.25009656e-01 1.53076684e+00
4.72015709e-01 -1.91874504e-01 6.98843777e-01 -3.10080200e-01
-4.45374668e-01 -6.73474371e-01 -7.24394619e-01 -6.94798470e-01
-1.94443256e-01 -2.93850064e-01 9.67593670e-01 7.81129599e-01
7.11534731e-03 6.32209778e-01 -7.34521329e-01 -1.31152511e-01
3.67074579e-01 6.39943182e-01 6.33434236e-01 -1.26041484e+00
-7.34029412e-01 -7.57371843e-01 -1.20869398e-01 -1.08610487e+00
-2.20120445e-01 -8.90719891e-01 2.11777553e-01 -1.59364378e+00
2.78704375e-01 -4.73983765e-01 -6.81536496e-01 4.70750362e-01
-6.16694510e-01 4.79409844e-01 5.44217288e-01 2.45805800e-01
-3.46119016e-01 5.93813181e-01 1.53157544e+00 -1.65602431e-01
-4.93865401e-01 1.51429892e-01 -9.39521909e-01 6.65205479e-01
7.71249712e-01 -4.33456600e-01 -5.04191518e-01 -3.24799359e-01
1.62835717e-01 1.57560423e-01 1.66347146e-01 -1.24260330e+00
1.16544500e-01 -7.25416467e-02 4.79902953e-01 -9.62239087e-01
2.50385523e-01 -1.85994685e-01 4.76623714e-01 4.18157816e-01
-5.45033574e-01 1.19710557e-01 2.11995259e-01 3.11380386e-01
-5.30637264e-01 -4.04789031e-01 5.71388721e-01 -2.75611222e-01
-3.81062150e-01 3.71667087e-01 -2.83431977e-01 1.37845755e-01
6.04261518e-01 -2.08484262e-01 1.28392085e-01 -4.22675610e-01
-7.86092699e-01 -1.00714758e-01 2.12271020e-01 6.08626783e-01
5.66628277e-01 -1.76151562e+00 -9.47305560e-01 2.40831032e-01
1.31704971e-01 1.77687377e-01 4.46090758e-01 4.93266433e-01
-4.00520787e-02 3.95905256e-01 -2.54407376e-01 -3.66508514e-01
-1.38429356e+00 4.30514663e-01 -1.65017039e-01 -6.31076157e-01
-6.45059109e-01 9.48908091e-01 4.04648155e-01 1.22395404e-01
2.09185213e-01 -4.57227468e-01 -3.55855048e-01 1.26323789e-01
7.04295635e-01 3.25210452e-01 -1.80995420e-01 -6.03039443e-01
-1.68852583e-01 9.73944724e-01 -2.01710314e-02 -2.25629866e-01
1.05554664e+00 1.42177805e-01 -1.70067593e-01 5.82833827e-01
5.08718371e-01 3.92394722e-01 -1.17329860e+00 -3.25861126e-02
-3.59867699e-02 -4.65190798e-01 -1.54233083e-01 -7.96422541e-01
-1.00950372e+00 7.21091747e-01 -1.31169230e-01 2.48601526e-01
1.32108128e+00 -6.65176064e-02 1.11290669e+00 7.80682778e-03
4.81955469e-01 -1.05957973e+00 3.59813303e-01 7.98742175e-01
1.18812490e+00 -2.60502458e-01 -2.38143921e-01 -2.02399731e-01
-6.76445901e-01 8.42016578e-01 6.81301355e-01 -7.04740882e-02
2.29255706e-02 1.24241732e-01 -1.20971985e-01 2.40352657e-02
-8.37723613e-01 -3.42511028e-01 5.96144974e-01 3.35341871e-01
8.26150000e-01 1.23129673e-01 -2.38514334e-01 9.37249005e-01
-6.97560251e-01 -2.02195235e-02 2.60171682e-01 4.15220708e-01
-5.85735559e-01 -1.28650522e+00 -4.66676652e-01 4.93055522e-01
-4.79395986e-01 -4.84669745e-01 -4.67386037e-01 4.89201367e-01
5.78417122e-01 9.92031574e-01 9.07947570e-02 -5.57772934e-01
6.39272153e-01 -9.29064583e-04 3.58756602e-01 -7.64017463e-01
-9.27135408e-01 5.44487655e-01 -1.13497190e-01 -3.65916520e-01
-1.71644211e-01 -7.35095501e-01 -1.34920681e+00 -2.51732498e-01
-2.00509861e-01 4.39553320e-01 1.75165311e-01 6.26808941e-01
4.35346216e-01 1.13734412e+00 6.28569901e-01 -1.08809388e+00
-3.72013479e-01 -1.27926660e+00 -7.56025076e-01 4.32244539e-01
2.21665297e-03 -4.83702332e-01 -6.55245483e-02 6.25726804e-02] | [15.974321365356445, 5.536772727966309] |
7056888f-03b5-408f-aaab-eaa13e1d08d1 | learning-from-miscellaneous-other-class-words | 2106.15167 | null | https://arxiv.org/abs/2106.15167v1 | https://arxiv.org/pdf/2106.15167v1.pdf | Learning from Miscellaneous Other-Class Words for Few-shot Named Entity Recognition | Few-shot Named Entity Recognition (NER) exploits only a handful of annotations to identify and classify named entity mentions. Prototypical network shows superior performance on few-shot NER. However, existing prototypical methods fail to differentiate rich semantics in other-class words, which will aggravate overfitting under few shot scenario. To address the issue, we propose a novel model, Mining Undefined Classes from Other-class (MUCO), that can automatically induce different undefined classes from the other class to improve few-shot NER. With these extra-labeled undefined classes, our method will improve the discriminative ability of NER classifier and enhance the understanding of predefined classes with stand-by semantic knowledge. Experimental results demonstrate that our model outperforms five state-of-the-art models in both 1-shot and 5-shots settings on four NER benchmarks. We will release the code upon acceptance. The source code is released on https: //github.com/shuaiwa16/OtherClassNER.git. | ['Juanzi Li', 'Lei Hou', 'Minghui Liu', 'Yixin Cao', 'Bin Xu', 'Shuai Wang', 'Meihan Tong'] | 2021-06-29 | null | https://aclanthology.org/2021.acl-long.487 | https://aclanthology.org/2021.acl-long.487.pdf | acl-2021-5 | ['miscellaneous', 'few-shot-ner'] | ['miscellaneous', 'natural-language-processing'] | [-2.74271160e-01 2.01031044e-01 -3.00920248e-01 -6.16509676e-01
-7.36577451e-01 -6.66949093e-01 5.18067122e-01 1.25127152e-01
-6.07970417e-01 8.15706253e-01 4.57068741e-01 1.33863047e-01
-9.21739116e-02 -9.39351559e-01 -1.80377126e-01 -3.93316984e-01
2.82648563e-01 3.50365996e-01 5.68741858e-01 -3.17710012e-01
2.17003301e-01 1.53820261e-01 -1.61617231e+00 3.36006314e-01
8.64253938e-01 4.95819479e-01 7.62411058e-02 3.10304046e-01
-7.89318681e-01 8.79475832e-01 -5.01565218e-01 -6.50933683e-01
1.04485594e-01 -4.79895532e-01 -1.00604105e+00 -4.81224120e-01
1.36094123e-01 1.85417235e-01 -1.74411923e-01 1.17329264e+00
8.07660639e-01 7.37619102e-01 6.65139019e-01 -9.92930174e-01
-9.32988346e-01 1.08858836e+00 -1.69518456e-01 5.13882756e-01
1.47491366e-01 2.22601164e-02 1.21780825e+00 -1.16198468e+00
8.52868199e-01 9.67111349e-01 8.68348598e-01 1.01335359e+00
-7.40698993e-01 -7.75645375e-01 1.57295708e-02 2.41708443e-01
-1.53672457e+00 -5.74968040e-01 3.68927896e-01 -8.92853662e-02
1.17971444e+00 1.95580989e-01 -5.12597198e-03 1.18460011e+00
-4.13145006e-01 6.56877935e-01 9.58252728e-01 -2.84012318e-01
4.74558502e-01 5.05509973e-02 1.00923431e+00 4.98987913e-01
3.95667136e-01 -3.02668381e-02 -1.91599578e-01 -1.80770323e-01
9.63198170e-02 2.65814871e-01 -2.15663716e-01 6.62055314e-02
-8.35581481e-01 7.90005982e-01 3.21154207e-01 8.08164060e-01
-3.28066945e-01 -3.08826059e-01 3.90314221e-01 8.37329924e-02
4.00237828e-01 5.06374061e-01 -7.49801219e-01 -3.13820988e-01
-7.95792222e-01 -2.63644785e-01 1.13443255e+00 1.22862160e+00
9.17668223e-01 -1.10683933e-01 -5.15452564e-01 1.12161493e+00
-5.79106323e-02 1.92137152e-01 8.03585470e-01 -5.29954731e-01
2.83290058e-01 6.81587160e-01 6.66691661e-02 -4.49137360e-01
-5.37170291e-01 -2.59908944e-01 -5.01452446e-01 -3.71348143e-01
1.67423397e-01 -4.54251558e-01 -1.26242673e+00 1.35250747e+00
5.59037566e-01 7.73025632e-01 2.40015879e-01 7.68636346e-01
1.39995074e+00 7.53551245e-01 5.56325734e-01 -4.14192080e-02
1.64757097e+00 -1.08154023e+00 -7.87067831e-01 -2.92856336e-01
9.91330683e-01 -5.45993268e-01 1.08269560e+00 -3.75373065e-01
-3.98150265e-01 -3.77415359e-01 -8.33189487e-01 -6.59378693e-02
-1.01719010e+00 1.19727040e-02 5.20278633e-01 9.17195857e-01
-3.54402602e-01 7.53391504e-01 -6.57272339e-01 -8.24073672e-01
4.08120811e-01 -2.57337280e-02 -2.78279275e-01 -2.05225348e-02
-1.75470126e+00 9.26187396e-01 8.37084115e-01 -3.62082630e-01
-6.01045251e-01 -1.06475759e+00 -9.65732396e-01 3.79373163e-01
6.07119977e-01 -3.77092391e-01 1.32703042e+00 -3.41786206e-01
-1.18975413e+00 8.75482023e-01 -6.00600690e-02 -3.58321279e-01
-4.02370095e-02 -1.24235153e-01 -8.21379721e-01 -1.20978236e-01
4.90014404e-01 5.05127847e-01 2.00992841e-02 -1.00952387e+00
-7.03781724e-01 -1.57628670e-01 1.60744473e-01 -6.55565709e-02
-4.95847821e-01 2.49007151e-01 -7.38218278e-02 -4.53238338e-01
-3.16920549e-01 -5.39889693e-01 -2.28581220e-01 -7.18426049e-01
-4.68454659e-01 -6.57299101e-01 6.81037247e-01 -1.01865847e-02
1.44091797e+00 -2.21756649e+00 -6.89586103e-01 -3.38448554e-01
-3.68024632e-02 6.26367033e-01 -1.86192811e-01 5.30378282e-01
-2.13654250e-01 4.06080365e-01 -1.20667815e-01 -4.73568328e-02
1.96863309e-01 2.15562239e-01 -1.52044490e-01 1.82305396e-01
2.21219957e-01 8.91332984e-01 -1.31299996e+00 -5.36352038e-01
1.15708105e-01 4.57579195e-01 -9.42878872e-02 1.74636453e-01
1.62273958e-01 7.80414492e-02 -5.41607976e-01 6.70049131e-01
6.17656291e-01 -2.38123938e-01 -8.14974084e-02 -2.18442649e-01
-2.74700075e-01 4.43114519e-01 -1.36281085e+00 1.52483058e+00
-2.71796584e-01 1.05542429e-01 -6.87980711e-01 -7.56515384e-01
9.68947768e-01 5.48333168e-01 9.09427628e-02 -3.06259304e-01
4.19825345e-01 2.20306054e-01 -1.12405598e-01 -5.32032132e-01
5.00663638e-01 -4.39763606e-01 -2.98785210e-01 2.37964451e-01
6.47167385e-01 5.12601674e-01 5.94753802e-01 7.64249191e-02
1.21784246e+00 3.27699259e-02 8.67822528e-01 -1.93437591e-01
1.54204652e-01 1.21990167e-01 9.88955379e-01 9.94980812e-01
-6.91188693e-01 5.52976966e-01 1.06881559e-01 -3.60461503e-01
-8.27585280e-01 -9.51390564e-01 -1.29888222e-01 1.54149103e+00
8.39344412e-02 -3.81128222e-01 -7.69746125e-01 -1.04031861e+00
-4.32056963e-01 1.35927880e+00 -6.33622766e-01 -1.21418580e-01
-2.97850251e-01 -9.06458199e-01 1.06945074e+00 5.66377282e-01
4.01705891e-01 -1.17752516e+00 -4.66751605e-01 1.95138291e-01
-2.21297905e-01 -1.03776562e+00 -6.50539458e-01 4.39842671e-01
-5.59216142e-01 -1.10813594e+00 -8.78633022e-01 -9.36526060e-01
3.18983525e-01 3.30119431e-01 1.07557225e+00 -7.65507668e-02
-3.31413597e-01 2.81162858e-01 -8.89332533e-01 -4.31762666e-01
-6.29912838e-02 3.42581332e-01 -5.07809445e-02 -1.09939009e-01
1.23021185e+00 -4.72078741e-01 -3.74782830e-01 2.23142222e-01
-7.18709588e-01 -4.89762664e-01 3.81382942e-01 7.21718788e-01
4.00682569e-01 -1.18904309e-02 8.50830376e-01 -1.38025653e+00
2.87795335e-01 -1.07335210e+00 7.94365257e-03 4.43941891e-01
-6.39357924e-01 1.23528920e-01 7.19077110e-01 -5.42229176e-01
-1.45864868e+00 6.61244020e-02 -9.31175798e-02 -3.67603838e-01
-6.55038297e-01 4.12655503e-01 -2.58824319e-01 4.15877312e-01
9.75647688e-01 3.64440493e-02 -8.51682663e-01 -7.49729574e-01
5.95625997e-01 8.79289627e-01 3.69532645e-01 -4.20532733e-01
5.16211092e-01 3.24258238e-01 -5.43203890e-01 -8.51574421e-01
-1.38494802e+00 -1.26964009e+00 -7.10916042e-01 -6.16891906e-02
9.50566053e-01 -9.26193118e-01 4.89641307e-03 9.03126150e-02
-1.15157616e+00 1.00501634e-01 -4.38368320e-01 3.80941004e-01
-9.68711823e-02 2.16307208e-01 -6.25599146e-01 -8.91909182e-01
-5.45605421e-01 -4.54636633e-01 8.94217372e-01 8.65620792e-01
-2.12001204e-01 -8.90542150e-01 3.80428851e-01 2.22984016e-01
2.03691155e-01 -4.18953486e-02 5.36569357e-01 -1.72053075e+00
-3.16835716e-02 -3.35484684e-01 -3.30675654e-02 -9.60629210e-02
3.35765854e-02 -2.95969278e-01 -1.16413963e+00 1.86823726e-01
-2.04698071e-01 -2.06485450e-01 9.38777208e-01 -2.66072825e-02
5.80392718e-01 -1.00543357e-01 -5.42821348e-01 2.72282004e-01
1.52212989e+00 2.11444557e-01 7.13144481e-01 3.34696084e-01
7.05829084e-01 5.48479497e-01 7.70353317e-01 4.31575656e-01
3.55825543e-01 2.38806918e-01 -1.53680276e-02 2.46619031e-01
-2.87387162e-01 -3.21509361e-01 1.66436151e-01 8.48089814e-01
-1.61022916e-01 -3.25623512e-01 -1.17174876e+00 9.17660594e-01
-1.68877029e+00 -1.36321843e+00 -2.01655254e-01 1.81214917e+00
9.58847344e-01 -2.09681969e-02 -7.24469572e-02 -3.40535045e-01
1.41718936e+00 2.45879143e-01 -4.25020695e-01 -1.10634178e-01
-1.02255121e-01 4.10503656e-01 5.54170132e-01 1.37611955e-01
-1.24826324e+00 1.22750831e+00 5.09389639e+00 1.13757205e+00
-5.50935566e-01 6.76552117e-01 4.08419609e-01 2.35764205e-01
-2.30808668e-02 2.10686818e-01 -1.43592942e+00 4.70855922e-01
1.43015075e+00 -4.64837074e-01 -1.46960944e-01 1.08909369e+00
-1.13759466e-01 3.63059342e-01 -8.37140024e-01 6.74560547e-01
1.97608396e-01 -1.16857100e+00 9.99773573e-03 -3.27978820e-01
8.18640351e-01 3.45528603e-01 -6.98029995e-01 9.63067412e-01
4.60590541e-01 -5.96003294e-01 3.49693149e-01 6.13059103e-01
4.91011292e-01 -7.61170030e-01 1.05714881e+00 4.67290252e-01
-1.46084642e+00 -5.04285097e-02 -6.58454657e-01 5.96245192e-02
2.48573452e-01 5.01582742e-01 -4.86103833e-01 6.46550357e-01
7.05321848e-01 6.20065749e-01 -3.89673084e-01 1.25308836e+00
-2.97892898e-01 6.66209757e-01 -8.29891115e-02 -4.35722828e-01
3.51832300e-01 2.23364100e-01 5.75803339e-01 1.69658077e+00
2.63118058e-01 7.29281306e-01 2.61454254e-01 6.30691528e-01
-2.43210390e-01 4.60140944e-01 -5.55196643e-01 -3.38023454e-01
8.58045042e-01 1.43669236e+00 -1.09047949e+00 -6.43494070e-01
-6.40300155e-01 9.22263741e-01 4.70341861e-01 1.60632655e-01
-9.99192297e-01 -1.21403563e+00 5.18213272e-01 -1.66762620e-01
5.52023888e-01 3.13008338e-01 -1.85575500e-01 -1.43927491e+00
-4.45477724e-01 -3.34364742e-01 9.75884438e-01 -5.87906003e-01
-1.87640631e+00 6.04883552e-01 -1.40009433e-01 -1.13987076e+00
1.92326099e-01 -4.74696338e-01 -8.51247370e-01 3.14800948e-01
-1.33247304e+00 -1.02984345e+00 -2.21281275e-01 3.21420312e-01
8.19716215e-01 -2.07332727e-02 1.07643998e+00 5.87084174e-01
-6.62984431e-01 6.30001366e-01 6.21077530e-02 6.55887067e-01
8.76001775e-01 -1.23888278e+00 3.60876888e-01 9.27800477e-01
2.92454988e-01 6.66929424e-01 6.06907904e-01 -7.53545880e-01
-7.10797131e-01 -1.37827027e+00 1.41523933e+00 -5.90755284e-01
7.71043241e-01 -2.73214802e-02 -8.87706459e-01 7.28170753e-01
2.00213656e-01 1.29578829e-01 1.23554087e+00 2.68298388e-01
-6.38055563e-01 2.75452852e-01 -1.20739734e+00 5.23924232e-01
1.16385698e+00 -4.14563209e-01 -1.29513490e+00 1.67460486e-01
8.94987404e-01 -2.56421249e-02 -9.79763806e-01 3.23396534e-01
2.00503364e-01 -7.04904020e-01 8.52902234e-01 -9.34267163e-01
2.42000356e-01 -3.32426369e-01 -1.18224807e-01 -1.18803275e+00
-4.00891811e-01 -2.97358543e-01 -1.13172449e-01 1.77202904e+00
6.82903409e-01 -5.81255615e-01 4.88397062e-01 5.85837543e-01
-3.08813483e-01 -3.71325761e-01 -7.85773933e-01 -1.25981092e+00
5.73619828e-02 -1.65205821e-01 6.72213972e-01 1.59522390e+00
2.15563044e-01 7.10203350e-01 -2.17590854e-01 3.21888804e-01
4.86345768e-01 -1.16379492e-01 2.44641185e-01 -1.29137981e+00
5.23433238e-02 -1.78323090e-01 -5.10768473e-01 -5.31519294e-01
2.88296908e-01 -1.16312468e+00 3.66249293e-01 -1.49230409e+00
5.99669397e-01 -2.66639054e-01 -7.98542857e-01 1.03418374e+00
-5.38347065e-01 3.00887227e-01 2.20917687e-01 1.10163175e-01
-1.09623063e+00 4.41307336e-01 6.12555861e-01 3.38031724e-02
-1.60281941e-01 -3.69124264e-01 -7.65098572e-01 6.95921898e-01
8.94349694e-01 -9.66326952e-01 -2.64682889e-01 -1.19841948e-01
-7.42542595e-02 -4.18241113e-01 -4.41872366e-02 -1.15677762e+00
4.90813404e-01 -2.22297937e-01 4.78681736e-02 -3.26150149e-01
9.45342984e-03 -4.08844650e-01 8.07495043e-02 2.50902832e-01
-4.11637723e-01 -4.03724372e-01 -1.02889940e-01 6.84887826e-01
-1.07454509e-01 -9.31401372e-01 8.41703296e-01 -3.66642684e-01
-1.36519778e+00 3.71806979e-01 -1.88284412e-01 7.30086684e-01
1.02139890e+00 -1.12880111e-01 -5.85360348e-01 7.53342882e-02
-8.84298444e-01 1.11460224e-01 2.05482498e-01 5.87895334e-01
2.58250177e-01 -1.33678389e+00 -6.52624667e-01 -4.28588212e-01
5.10373354e-01 -5.59530556e-01 4.68944728e-01 6.50221765e-01
-1.67123884e-01 2.19807029e-01 6.36349395e-02 -1.31737106e-02
-1.06863666e+00 6.60988271e-01 3.23368579e-01 -1.93173096e-01
-7.37462878e-01 7.98780441e-01 2.35036314e-02 -8.28755796e-01
7.76830176e-03 8.06761459e-02 -6.50368750e-01 2.53355592e-01
9.49209332e-01 6.01185322e-01 -1.26933470e-01 -6.88329816e-01
-5.30250430e-01 1.80189312e-01 -1.27523854e-01 2.62935042e-01
1.42673838e+00 -1.22448042e-01 8.64320174e-02 8.33791435e-01
1.08193374e+00 -6.92007393e-02 -6.91971242e-01 -1.41746670e-01
6.51105583e-01 -1.37283310e-01 -1.20832905e-01 -8.53122115e-01
-7.93934405e-01 5.89694917e-01 5.67595184e-01 1.78076521e-01
6.99070811e-01 2.32962012e-01 1.13094103e+00 5.88237464e-01
4.97162372e-01 -1.21820307e+00 -3.90633315e-01 9.46034610e-01
1.65115166e-02 -1.27538919e+00 -3.60045105e-01 -5.83754420e-01
-7.70404220e-01 1.00865245e+00 7.24841833e-01 -1.15886807e-01
8.20926189e-01 7.77224153e-02 2.05260068e-01 -3.06785434e-01
-6.88613653e-01 -8.79111469e-01 1.80351943e-01 5.48123896e-01
6.66428387e-01 1.02493048e-01 -7.15546191e-01 1.22655427e+00
8.78412053e-02 -6.42711744e-02 5.94523489e-01 9.90931928e-01
-9.13646817e-01 -9.12274897e-01 6.92869164e-03 3.86357576e-01
-7.00776219e-01 -5.89919150e-01 -3.38871688e-01 6.64652944e-01
4.18410450e-01 1.16690934e+00 -1.06337287e-01 -1.20842613e-01
5.19397199e-01 7.51474857e-01 -1.06492631e-01 -1.13973272e+00
-6.86565220e-01 -3.52567792e-01 4.68279898e-01 -2.79247910e-01
-5.01998186e-01 -3.57884347e-01 -1.67415202e+00 -1.05820335e-01
-6.70535743e-01 6.11716151e-01 2.22046882e-01 1.09449708e+00
5.22681117e-01 3.13775271e-01 3.33416790e-01 -3.21339101e-01
-5.85023344e-01 -1.15240955e+00 -6.85080588e-01 6.30540133e-01
-2.74957210e-01 -8.59977722e-01 -5.59141338e-01 5.53285517e-03] | [9.643937110900879, 9.346454620361328] |
401f125b-d05a-408f-be57-77c246f9837c | transferable-knowledge-based-multi | 2107.12618 | null | https://arxiv.org/abs/2107.12618v1 | https://arxiv.org/pdf/2107.12618v1.pdf | Transferable Knowledge-Based Multi-Granularity Aggregation Network for Temporal Action Localization: Submission to ActivityNet Challenge 2021 | This technical report presents an overview of our solution used in the submission to 2021 HACS Temporal Action Localization Challenge on both Supervised Learning Track and Weakly-Supervised Learning Track. Temporal Action Localization (TAL) requires to not only precisely locate the temporal boundaries of action instances, but also accurately classify the untrimmed videos into specific categories. However, Weakly-Supervised TAL indicates locating the action instances using only video-level class labels. In this paper, to train a supervised temporal action localizer, we adopt Temporal Context Aggregation Network (TCANet) to generate high-quality action proposals through ``local and global" temporal context aggregation and complementary as well as progressive boundary refinement. As for the WSTAL, a novel framework is proposed to handle the poor quality of CAS generated by simple classification network, which can only focus on local discriminative parts, rather than locate the entire interval of target actions. Further inspired by the transfer learning method, we also adopt an additional module to transfer the knowledge from trimmed videos (HACS Clips dataset) to untrimmed videos (HACS Segments dataset), aiming at promoting the classification performance on untrimmed videos. Finally, we employ a boundary regression module embedded with Outer-Inner-Contrastive (OIC) loss to automatically predict the boundaries based on the enhanced CAS. Our proposed scheme achieves 39.91 and 29.78 average mAP on the challenge testing set of supervised and weakly-supervised temporal action localization track respectively. | ['Yu Qiao', 'Wei Wu', 'Weihao Gan', 'Dongliang Wang', 'Yukun Li', 'Peiqin Zhuang', 'Haisheng Su'] | 2021-07-27 | null | null | null | null | ['weakly-supervised-temporal-action'] | ['computer-vision'] | [ 4.83662784e-01 -1.94005966e-02 -5.18405735e-01 -2.82543868e-01
-1.12220883e+00 -4.87618357e-01 5.05553603e-01 -3.78491402e-01
-4.13133889e-01 6.70195282e-01 2.79638737e-01 2.47594580e-01
-7.93745443e-02 -2.35441402e-01 -7.72794425e-01 -8.60219717e-01
-3.86470079e-01 6.52434826e-02 7.64828503e-01 2.05701753e-01
1.18251517e-01 3.10270637e-01 -1.50566101e+00 8.40530336e-01
7.80572236e-01 1.51709008e+00 1.40727356e-01 5.38097501e-01
1.20324485e-01 1.31595111e+00 -4.66570586e-01 1.01270981e-01
4.28495914e-01 -6.58483326e-01 -7.96021163e-01 1.33344263e-01
7.07006037e-01 -2.83944905e-01 -3.51733208e-01 7.30102241e-01
2.89874196e-01 3.08750659e-01 4.79555190e-01 -1.45712817e+00
-3.22834462e-01 4.23123777e-01 -5.65102994e-01 5.06696522e-01
5.04768431e-01 3.36245418e-01 8.21247339e-01 -8.77619505e-01
8.52132201e-01 8.13624024e-01 7.10697412e-01 5.49270928e-01
-8.06326807e-01 -6.89343214e-01 4.34829801e-01 8.53099406e-01
-1.37054074e+00 -4.86622632e-01 8.20906162e-01 -4.94336635e-01
7.29124546e-01 1.57286748e-01 6.72405005e-01 1.21073627e+00
1.75458699e-01 1.14025187e+00 1.04268229e+00 -1.34307757e-01
2.19243690e-01 -1.47705778e-01 -3.97477955e-01 6.12968862e-01
-6.36371195e-01 1.40823036e-01 -7.01794505e-01 3.09045315e-01
6.95593536e-01 -3.65666710e-02 -3.15734953e-01 -3.20671082e-01
-1.45303929e+00 2.94990718e-01 5.21287441e-01 5.71945250e-01
-4.49908674e-01 1.80686280e-01 6.43283367e-01 4.32687849e-02
4.27404732e-01 1.22415721e-01 -5.99485040e-01 -4.57436889e-01
-1.20736217e+00 -1.69333994e-01 1.52705550e-01 1.00902343e+00
5.93406618e-01 2.64210328e-02 -7.32734740e-01 5.27682245e-01
-1.25667676e-01 4.12103862e-01 5.61029971e-01 -1.05427146e+00
7.14079380e-01 5.21262050e-01 1.60732374e-01 -9.89071369e-01
-2.75260061e-01 -4.56789911e-01 -6.76794469e-01 -3.10692061e-02
4.12803918e-01 7.47587979e-02 -9.46028471e-01 1.62752366e+00
3.69366378e-01 7.73621440e-01 -1.06941797e-01 1.03629947e+00
4.33973908e-01 6.85399294e-01 4.09818113e-01 -3.90578508e-01
9.74889040e-01 -1.30186796e+00 -8.24016273e-01 -3.06700952e-02
9.08249140e-01 -4.85339701e-01 1.05122328e+00 3.98836672e-01
-9.42522466e-01 -8.65452290e-01 -9.17380691e-01 1.66806370e-01
-2.44565994e-01 7.88536966e-01 3.57080728e-01 2.85538360e-02
-9.35427904e-01 6.37365103e-01 -8.92667294e-01 -1.81654394e-01
5.78809679e-01 1.86488345e-01 -6.19328797e-01 -1.76423773e-01
-1.20846272e+00 6.67100012e-01 6.99564874e-01 3.11680734e-01
-1.18286991e+00 -6.39213502e-01 -7.94514954e-01 -1.63190648e-01
7.09289968e-01 -3.83931794e-03 9.33432817e-01 -1.57499719e+00
-1.31073272e+00 5.32982707e-01 1.68735581e-03 -7.58206129e-01
6.41231418e-01 -2.57174909e-01 -5.97906768e-01 7.38491833e-01
3.24252814e-01 9.66203213e-01 8.66540372e-01 -8.87326181e-01
-1.15164089e+00 -6.76064640e-02 -3.84093001e-02 2.89707601e-01
-2.98841357e-01 -5.09084500e-02 -6.64215803e-01 -8.94738436e-01
1.00525893e-01 -9.12039399e-01 6.04855530e-02 7.50631019e-02
3.32473335e-03 -3.93934757e-01 1.23250687e+00 -1.16284239e+00
1.36113966e+00 -2.31117535e+00 2.62901187e-01 -6.54154047e-02
-1.33074835e-01 2.20869511e-01 -3.12900513e-01 6.13578744e-02
-1.24789864e-01 -2.14789361e-01 -1.80485874e-01 -2.64146149e-01
-1.51898205e-01 1.12275496e-01 -1.98959216e-01 4.34715390e-01
2.97734082e-01 8.94222558e-01 -1.11446536e+00 -8.12654555e-01
4.42714900e-01 1.88057035e-01 -4.16378021e-01 1.21215798e-01
-2.35816672e-01 7.60127246e-01 -4.15709227e-01 8.93812776e-01
4.28110480e-01 1.62819605e-02 -8.11557844e-03 -6.20000780e-01
-1.33026093e-01 -7.26239309e-02 -1.05817556e+00 2.01496482e+00
-3.25958192e-01 4.53587562e-01 -1.42339319e-01 -1.17760956e+00
6.79133117e-01 5.18396735e-01 1.08817267e+00 -9.34641361e-01
-2.10601658e-01 1.10862538e-01 -2.78453678e-01 -7.38237441e-01
6.81775287e-02 1.68286458e-01 -1.21836483e-01 1.28233269e-01
3.09378028e-01 4.81492966e-01 2.71087170e-01 5.64527139e-02
1.23840189e+00 9.03766513e-01 -9.72917024e-03 1.24869026e-01
8.49741280e-01 7.65823647e-02 9.52782094e-01 3.61501634e-01
-7.50733852e-01 5.97366989e-01 3.71324509e-01 -4.12714303e-01
-7.06615746e-01 -1.03803325e+00 1.37040466e-01 1.08796418e+00
3.03504497e-01 -3.59733284e-01 -7.11342573e-01 -1.48248947e+00
-2.82810628e-01 3.75775754e-01 -8.48423243e-01 -2.32905135e-01
-7.67304063e-01 -1.24072701e-01 5.23646355e-01 8.56604278e-01
1.12539470e+00 -1.33083725e+00 -4.40599680e-01 1.36701822e-01
-5.43107510e-01 -1.39103615e+00 -8.09317946e-01 -7.83554744e-03
-6.80078924e-01 -1.27797794e+00 -7.32300043e-01 -8.61564159e-01
4.87566590e-01 4.19296250e-02 5.14762998e-01 -2.97923535e-01
-1.08779922e-01 4.76913810e-01 -7.68876433e-01 3.86283159e-01
-7.83153158e-03 -1.72621071e-01 6.86108619e-02 5.68686128e-01
2.82430708e-01 -4.40009832e-01 -8.22842121e-01 7.58708000e-01
-6.30448997e-01 9.77158174e-02 7.95322478e-01 6.47454262e-01
7.96772361e-01 5.27964123e-02 6.88113093e-01 -1.73363954e-01
-1.92817926e-01 -3.77007693e-01 -3.11629206e-01 4.69716460e-01
-2.93049365e-01 -3.55900735e-01 7.15676665e-01 -6.59094989e-01
-1.14171994e+00 5.17795742e-01 1.06125504e-01 -9.29445863e-01
-2.93272436e-01 3.10105503e-01 -1.23679437e-01 -1.55402809e-01
5.50557137e-01 5.97266138e-01 -2.61513442e-01 -4.01956439e-01
2.52232373e-01 4.55463499e-01 6.99703097e-01 -3.63405704e-01
6.40077114e-01 5.00751078e-01 -1.56689093e-01 -4.12965000e-01
-9.57434356e-01 -7.50891566e-01 -1.00648344e+00 -8.19256365e-01
1.35858011e+00 -9.59529757e-01 -4.51778263e-01 3.93038064e-01
-9.09227490e-01 -6.20048344e-01 -3.38321686e-01 7.72272110e-01
-8.60337615e-01 3.81328076e-01 -4.62053776e-01 -5.75053036e-01
-1.28076911e-01 -9.09199893e-01 1.20300150e+00 1.10404760e-01
1.72317941e-02 -7.52771974e-01 -8.81150141e-02 6.05616808e-01
1.48408756e-01 4.43946958e-01 3.42012137e-01 -5.15550435e-01
-7.24153697e-01 -3.95133913e-01 -1.43652171e-01 7.31259346e-01
1.90273419e-01 -2.19319940e-01 -7.78827429e-01 -2.24488705e-01
-2.89369196e-01 -4.78841215e-01 8.10942173e-01 5.03341377e-01
1.29291117e+00 -1.56306058e-01 -3.47038925e-01 4.97569203e-01
1.03620517e+00 4.50784534e-01 8.37856829e-01 1.41263604e-01
7.09342122e-01 5.01662850e-01 1.46029985e+00 3.50464791e-01
3.45866266e-03 1.05068517e+00 2.54749238e-01 1.84904272e-03
-3.75481874e-01 -4.19909507e-01 8.53693366e-01 5.52884698e-01
-3.88497621e-01 4.95198928e-02 -5.53310931e-01 4.86682057e-01
-2.07642841e+00 -1.23612988e+00 1.24838009e-01 2.09764385e+00
6.54120147e-01 2.23373443e-01 2.91742384e-01 1.15615986e-01
8.31319451e-01 2.50330746e-01 -4.58072275e-01 2.37544820e-01
-9.90222618e-02 4.44011502e-02 4.47747827e-01 1.70999154e-01
-1.68772411e+00 1.08168280e+00 5.15489864e+00 1.39434433e+00
-1.04495788e+00 3.34985346e-01 5.03801286e-01 -1.50028750e-01
3.84124398e-01 -9.89481583e-02 -5.58561146e-01 6.93539679e-01
7.24015594e-01 2.44387269e-01 1.45262778e-01 9.28111434e-01
7.01487541e-01 -2.48688623e-01 -1.07159173e+00 9.36540306e-01
2.90987402e-01 -1.26004064e+00 -1.61080718e-01 -3.23487103e-01
8.94836605e-01 -2.85477579e-01 -3.42253000e-02 6.90057397e-01
-3.85121346e-01 -6.90542936e-01 8.30424666e-01 6.85506105e-01
9.46174502e-01 -5.31151772e-01 8.16080034e-01 2.96559364e-01
-1.71060002e+00 -5.00453353e-01 9.67926253e-03 2.37345934e-01
1.17381752e-01 3.36038768e-02 -6.29273713e-01 6.73372030e-01
8.53526652e-01 1.33416164e+00 -6.79901540e-01 1.08575392e+00
-3.11684638e-01 6.61452770e-01 -5.87345175e-02 3.16931963e-01
4.38821942e-01 -2.61304807e-02 3.32258284e-01 1.11342108e+00
2.95907080e-01 1.43627346e-01 5.77297390e-01 4.48710769e-01
3.16604942e-01 1.18232325e-01 -3.51897508e-01 3.78024206e-02
1.56241715e-01 1.24295771e+00 -7.56818235e-01 -4.40146595e-01
-2.65211254e-01 1.21335971e+00 1.20188678e-02 4.90055948e-01
-1.38060236e+00 -1.14970185e-01 1.59329429e-01 1.11207083e-01
3.94366264e-01 -2.05518082e-01 2.14412343e-02 -1.04132760e+00
2.11427107e-01 -7.28450298e-01 6.27551079e-01 -9.47359800e-01
-8.44485223e-01 3.74398619e-01 7.92131498e-02 -1.96205318e+00
-1.02097780e-01 -3.01589340e-01 -5.64990222e-01 3.37590098e-01
-1.17004120e+00 -1.53377259e+00 -4.65288222e-01 8.36223185e-01
9.32577193e-01 -1.14144117e-01 3.44794929e-01 6.70463145e-01
-5.27352631e-01 6.73361480e-01 -3.32063109e-01 1.88623920e-01
9.17104483e-01 -1.03577328e+00 -4.13242966e-01 9.33006108e-01
8.13276321e-02 -3.03620603e-02 3.53878647e-01 -7.96212852e-01
-9.47478712e-01 -1.57803893e+00 6.48161411e-01 -3.58735472e-01
6.43325984e-01 -1.42470583e-01 -6.40727341e-01 6.92817986e-01
-9.38320160e-03 3.40311110e-01 3.03028226e-01 -3.65108281e-01
-3.90066504e-02 -3.31902504e-01 -1.03874147e+00 3.95832270e-01
1.26318681e+00 -5.71407974e-01 -4.36012059e-01 5.34904063e-01
6.33191705e-01 -1.73135296e-01 -8.73558819e-01 7.41989553e-01
5.05823135e-01 -8.22849095e-01 7.59211123e-01 -5.11781752e-01
4.95711923e-01 -6.67825818e-01 -1.19005300e-01 -7.06135094e-01
-2.63438046e-01 -3.86648178e-01 -3.23956609e-01 1.13995862e+00
2.06245810e-01 -6.36579916e-02 1.04111087e+00 4.52082902e-02
-6.21971130e-01 -8.14397931e-01 -1.29507232e+00 -9.84979630e-01
-4.31019455e-01 -5.24944305e-01 -4.43880074e-03 9.33432519e-01
-5.65353269e-03 -1.47358134e-01 -6.40118062e-01 1.71807304e-01
3.72928351e-01 -3.34016606e-03 5.62078893e-01 -5.47348261e-01
-2.98978657e-01 -1.71507791e-01 -7.12411821e-01 -1.21543324e+00
2.35140905e-01 -8.09201419e-01 2.60495692e-01 -1.25421965e+00
3.97049915e-03 -3.98354501e-01 -6.76198125e-01 7.51783729e-01
1.11382194e-01 6.42736971e-01 1.91902921e-01 1.60293862e-01
-1.28283858e+00 6.74313784e-01 1.35591555e+00 -2.72904545e-01
-1.71367690e-01 2.10676417e-01 2.52151400e-01 5.88156939e-01
5.87736249e-01 -3.08392584e-01 -4.93336141e-01 2.96678413e-02
-3.44272077e-01 2.82962263e-01 6.59632266e-01 -1.49796486e+00
2.59259373e-01 -3.51527095e-01 4.62394118e-01 -7.76906788e-01
3.68487418e-01 -9.30543542e-01 -1.03950920e-02 3.78833652e-01
-4.90101516e-01 -3.14381093e-01 8.42166841e-02 7.60072529e-01
-4.55992103e-01 7.15382621e-02 7.29548573e-01 5.40036671e-02
-1.31885421e+00 4.76251066e-01 -2.78106749e-01 -9.85092893e-02
1.54422474e+00 -3.93467903e-01 -1.76893443e-01 -2.25400954e-01
-1.31669819e+00 3.37669343e-01 7.81294033e-02 5.23738146e-01
6.82221055e-01 -1.66378891e+00 -3.48390758e-01 2.92197485e-02
2.30521530e-01 -3.87133330e-01 8.22314322e-01 1.48567140e+00
-3.09071362e-01 4.54920232e-01 -2.61202544e-01 -7.36730039e-01
-1.21543705e+00 7.62961745e-01 3.85017425e-01 -2.52093166e-01
-6.96836948e-01 7.75186360e-01 2.08052337e-01 3.70159076e-04
5.05482078e-01 -2.88684428e-01 -3.56801331e-01 1.58865392e-01
5.32122314e-01 4.18111742e-01 -1.78375155e-01 -7.73420513e-01
-4.85507995e-01 7.15947568e-01 1.98694229e-01 7.72040412e-02
8.80695879e-01 -2.03359835e-02 3.77811313e-01 3.15848500e-01
1.30139768e+00 -2.21496269e-01 -1.82291412e+00 -1.28861174e-01
1.42656997e-01 -3.63399893e-01 -1.02456920e-01 -1.04440296e+00
-1.09980476e+00 7.84218729e-01 9.13289428e-01 -2.61548281e-01
1.47813356e+00 -3.07803750e-02 8.23579371e-01 -2.43265666e-02
4.16056663e-01 -1.51697111e+00 5.51445603e-01 2.73467839e-01
1.02475441e+00 -1.12557411e+00 -2.32023016e-01 -3.64592582e-01
-9.56379473e-01 9.89519656e-01 9.09272909e-01 -9.47325230e-02
4.47772145e-01 -5.93213215e-02 -1.28677413e-01 -4.91322577e-03
-6.48308754e-01 -2.57703394e-01 5.17847359e-01 6.82375669e-01
-8.56707171e-02 -2.00174645e-01 -3.45319301e-01 5.53653419e-01
5.76151788e-01 2.71737814e-01 4.45332788e-02 8.29711497e-01
-4.42769796e-01 -7.00468719e-01 -1.12813979e-01 2.56087750e-01
-1.78876415e-01 1.20835520e-01 -2.80559927e-01 7.76494443e-01
7.08566844e-01 8.31795990e-01 1.05964541e-01 -8.00041914e-01
2.59611070e-01 1.33522049e-01 3.18195194e-01 -4.15774494e-01
-4.00084496e-01 3.12623292e-01 1.86555147e-01 -1.02039325e+00
-7.35062659e-01 -7.92618334e-01 -1.35971117e+00 2.92347163e-01
-1.20078079e-01 1.76580533e-01 1.57695264e-01 1.08540380e+00
3.53732228e-01 5.76253474e-01 7.11419702e-01 -1.00073576e+00
-2.41114229e-01 -1.09899950e+00 -5.15554070e-01 4.63993937e-01
1.66813880e-01 -8.69231224e-01 -4.67417747e-01 4.65646416e-01] | [8.432629585266113, 0.6539883613586426] |
bae0a0ac-3a2a-44af-87b1-1b9c71dadfd0 | lip-sync-matters-a-novel-multimodal-forgery | null | null | https://ieeexplore.ieee.org/document/9980296 | http://www.apsipa.org/proceedings/2022/APSIPA%202022/ThPM2-1/1570840237.pdf | Lip Sync Matters: A Novel Multimodal Forgery Detector | Deepfake technology has advanced a lot, but it is a double-sided sword for the community. One can use it for beneficial purposes, such as restoring vintage content in old movies, or for nefarious purposes, such as creating fake footage to manipulate the public and distribute non-consensual pornography. A lot of work has been done to combat its improper use by detecting fake footage with good performance thanks to the availability of numerous public datasets and unimodal deep learning-based models. However, these methods are insufficient to detect multimodal manipulations, such as both visual and acoustic. This work proposes a novel lip-reading-based multi-modal Deepfake detection method called “Lip Sync Matters.” It targets high-level semantic features to exploit the mismatch between the lip sequence extracted from the video and the synthetic lip sequence generated from the audio by the Wav2lip model to detect forged videos. Experimental results show that the proposed method outperforms several existing unimodal, ensemble, and multimodal methods on the publicly available multimodal FakeAVCeleb dataset. | ['Hsin-Min Wang', 'Yu Tsao', 'Yan-Tsung Peng', 'Sarwar Khan', 'Ammarah Hashmi', 'Sahibzada Adil Shahzad'] | 2022-11-07 | null | null | null | apsipa-asc-2022-2022-11 | ['face-swapping'] | ['computer-vision'] | [-6.57773949e-03 -3.18205774e-01 -3.05033505e-01 1.07341483e-02
-1.03265214e+00 -4.07040417e-01 4.87595350e-01 -2.73911953e-01
-1.20721199e-01 5.74071050e-01 4.94700849e-01 1.15651794e-01
5.25440156e-01 -3.63003612e-01 -6.68433428e-01 -9.35640216e-01
5.68261921e-01 -1.43530279e-01 2.35692620e-01 -4.05310661e-01
6.15419388e-01 2.14878306e-01 -1.90303874e+00 7.01664209e-01
5.23256004e-01 1.20046318e+00 -1.37220034e-02 5.57437658e-01
-5.84737435e-02 5.06405056e-01 -6.27158344e-01 -8.91404569e-01
-1.85449213e-01 -5.65065861e-01 -3.29530627e-01 -1.10586122e-01
6.26780748e-01 -7.78375864e-01 -5.57569861e-01 1.42320192e+00
9.88151610e-01 -2.96342969e-01 4.34894413e-01 -1.76230383e+00
-4.83571976e-01 2.76698142e-01 -7.52880573e-01 -1.30210221e-01
5.51774502e-01 2.87309289e-01 5.67799747e-01 -1.02229118e+00
4.53673065e-01 1.49261117e+00 7.27806211e-01 6.53450906e-01
-7.67963886e-01 -1.14497685e+00 -6.54195189e-01 6.99407816e-01
-1.26021242e+00 -1.04836023e+00 1.02794814e+00 -2.90106803e-01
4.36787993e-01 2.69977182e-01 5.58249831e-01 1.69998181e+00
-1.54685071e-02 1.15272808e+00 9.75242734e-01 -3.44696969e-01
-2.68792629e-01 2.70116508e-01 -3.37403864e-01 7.30834663e-01
-7.50113651e-02 7.02519342e-02 -1.07396019e+00 -2.61516750e-01
1.88188732e-01 -3.05382341e-01 -6.26857340e-01 -1.47367626e-01
-1.14299130e+00 8.74073803e-01 -1.03495479e-01 4.42639351e-01
-2.18986068e-02 5.87252825e-02 8.92833114e-01 2.96359628e-01
3.27246368e-01 9.21795145e-02 -3.90546918e-02 -4.17596906e-01
-9.85398829e-01 2.41512746e-01 4.13237244e-01 4.28480297e-01
3.61467868e-01 2.10623309e-01 9.67214629e-02 1.07282865e+00
4.09615010e-01 8.04373264e-01 8.54520082e-01 -7.64918387e-01
6.05646312e-01 1.37990519e-01 3.32140595e-01 -1.77350879e+00
-1.20349556e-01 3.31782699e-01 -7.29672134e-01 7.11259767e-02
5.00191808e-01 1.47035629e-01 -4.85462159e-01 1.48410916e+00
2.57495642e-01 1.63163200e-01 -1.78462967e-01 1.21246505e+00
1.17185736e+00 5.87998450e-01 -5.33349775e-02 4.00479883e-02
1.24455774e+00 -9.25050020e-01 -1.21245587e+00 4.37565707e-02
4.75187272e-01 -1.21447790e+00 1.13731897e+00 7.05237746e-01
-9.82433677e-01 -4.67814982e-01 -1.00767958e+00 -8.68320912e-02
-4.46209580e-01 3.22721928e-01 2.62926131e-01 1.12958729e+00
-7.28079379e-01 3.30701113e-01 -1.45318136e-01 -3.47931564e-01
6.65229142e-01 -1.08023956e-01 -5.82495511e-01 -1.92274243e-01
-1.44762075e+00 7.76361167e-01 2.37808809e-01 2.91904598e-01
-1.04090023e+00 -1.17455445e-01 -9.28162456e-01 -1.44046023e-01
3.54642272e-01 -1.54588610e-01 8.16887856e-01 -1.28576303e+00
-1.43069685e+00 1.15668118e+00 -1.13395028e-01 -1.64165825e-01
7.79455006e-01 -2.71579415e-01 -7.40983367e-01 2.33366698e-01
-1.19547360e-01 6.89062655e-01 1.61628938e+00 -1.20535696e+00
-3.34638208e-01 -3.94321233e-01 -3.63275766e-01 4.90008034e-02
-4.50278997e-01 2.91761935e-01 -1.76303878e-01 -8.38953018e-01
-1.07435375e-01 -7.73147464e-01 8.52776408e-01 2.75225677e-02
-7.40302980e-01 -5.81784686e-03 1.58282268e+00 -1.34894717e+00
1.05573535e+00 -2.49343443e+00 -5.01232035e-02 -2.75244236e-01
1.69355515e-02 5.61271548e-01 -2.19241366e-01 5.72890162e-01
1.14560120e-01 1.79915413e-01 -1.70754530e-02 -3.70210111e-01
7.84549862e-02 -2.46668398e-01 -4.38996047e-01 8.72821033e-01
-2.25206167e-01 7.97764361e-01 -9.17732894e-01 -4.69582587e-01
2.24873155e-01 6.66613817e-01 -2.27628499e-01 1.19244657e-01
4.80966382e-02 2.88344353e-01 -1.53412698e-02 1.03676128e+00
1.09414613e+00 5.27550995e-01 -1.01959586e-01 -5.91535866e-01
6.08781446e-03 8.32588412e-04 -7.68455863e-01 1.61558759e+00
-1.73802629e-01 1.21640801e+00 4.25933689e-01 -1.01460350e+00
6.44630134e-01 4.00785804e-01 2.39016533e-01 -5.31397998e-01
4.93005723e-01 3.51104081e-01 -2.81545371e-01 -1.08753121e+00
7.16979980e-01 -1.97477311e-01 -1.16382122e-01 2.39030495e-01
-1.22014042e-02 -1.52251422e-01 -5.01576364e-01 1.62315950e-01
6.03525281e-01 1.32347584e-01 -1.25309691e-01 3.01714778e-01
7.58465469e-01 -4.90425676e-01 1.45818576e-01 3.67315948e-01
-7.02729762e-01 4.85221863e-01 4.34256583e-01 -9.08993110e-02
-9.71535385e-01 -9.30869699e-01 8.32479447e-02 1.02754283e+00
4.31834847e-01 -1.33474752e-01 -8.10518146e-01 -6.18084311e-01
1.23438761e-01 5.21266162e-01 -2.76690245e-01 -3.93340737e-01
-2.40663484e-01 -1.94547489e-01 1.23087990e+00 -1.03036255e-01
1.01988292e+00 -1.13394010e+00 -2.34415710e-01 -8.79991800e-02
-1.00646985e+00 -1.30886614e+00 -6.58331037e-01 -7.27645814e-01
-1.18499495e-01 -1.06726134e+00 -1.01928496e+00 -8.11629772e-01
8.22673924e-03 5.65751791e-01 4.58655894e-01 9.37401503e-02
-3.39477330e-01 2.78638393e-01 -3.40416402e-01 -2.35970408e-01
-7.88839579e-01 -4.71057206e-01 8.24600905e-02 6.05997860e-01
4.32218671e-01 -5.40094301e-02 -4.39197332e-01 4.78064984e-01
-8.43654156e-01 -6.86479807e-02 2.50753701e-01 9.42109346e-01
-1.53630510e-01 -5.23020141e-02 8.48276019e-01 -1.10975817e-01
7.17304409e-01 -3.98726672e-01 -1.54548839e-01 -3.22818160e-02
-6.66001588e-02 -4.50351179e-01 1.85924813e-01 -6.99433804e-01
-1.04400063e+00 -4.76682872e-01 -3.36902946e-01 -6.55397356e-01
-2.79866189e-01 -7.16793686e-02 -5.51347315e-01 -2.04414502e-01
8.03009346e-02 3.25770795e-01 2.55209953e-01 -3.93311203e-01
3.02728742e-01 1.25246775e+00 6.89249814e-01 -8.37851986e-02
5.00780523e-01 7.45499015e-01 -1.95511922e-01 -1.26179981e+00
-3.83469313e-01 -4.84035671e-01 -5.48100919e-02 -7.83819318e-01
8.32697093e-01 -7.05450356e-01 -1.12858641e+00 1.50031877e+00
-1.53344154e+00 1.62086323e-01 4.83739108e-01 3.12816292e-01
-4.23095047e-01 1.13164592e+00 -6.80421829e-01 -8.85436118e-01
-6.99056610e-02 -1.27466393e+00 1.36897886e+00 -1.52690768e-01
-9.03540552e-02 -6.24129891e-01 -1.13732740e-01 1.15926623e+00
3.45035076e-01 2.58376360e-01 6.62825108e-01 -4.47128892e-01
-3.79794031e-01 -3.03063273e-01 -4.08448905e-01 7.10100591e-01
1.04104392e-01 -6.81030601e-02 -1.38474905e+00 -8.42641443e-02
-5.79294115e-02 -8.37283671e-01 1.03223968e+00 3.71777192e-02
1.19145548e+00 -5.24946213e-01 -1.65744185e-01 3.65021497e-01
1.03436673e+00 -9.26959217e-02 1.07557118e+00 -3.02527170e-03
6.73929989e-01 7.51065373e-01 5.75531662e-01 5.69555461e-01
1.39069930e-01 8.01450491e-01 9.08610284e-01 1.87565371e-01
-2.68462598e-01 -5.23737311e-01 7.32802331e-01 6.68206811e-01
1.81908667e-01 -4.85540181e-01 -5.83688855e-01 6.26985431e-01
-1.59550297e+00 -1.37357140e+00 -3.18677425e-01 2.08993173e+00
5.63410521e-01 -1.68799013e-01 1.15785301e-01 4.60501045e-01
1.11226368e+00 4.06032503e-01 -3.52306306e-01 -5.70529997e-01
-4.92804348e-01 -2.94606447e-01 2.79866606e-01 2.05795676e-01
-1.19966412e+00 1.10972309e+00 4.70806408e+00 1.43926787e+00
-1.33895421e+00 5.27824044e-01 3.39093775e-01 1.81548968e-01
-1.91364195e-02 -5.95557272e-01 -5.04398644e-01 8.65839720e-01
5.82013905e-01 4.24169332e-01 3.24164420e-01 6.48507714e-01
5.83041668e-01 -3.83317322e-01 -5.36787570e-01 1.53376055e+00
7.95611382e-01 -1.25180948e+00 -1.97738722e-01 2.46656518e-02
2.98049450e-01 -2.87232131e-01 3.75488877e-01 1.13086358e-01
-5.00778496e-01 -8.97540331e-01 7.91226685e-01 2.81442076e-01
9.27339315e-01 -8.37242961e-01 8.33079159e-01 3.30966324e-01
-7.66996503e-01 4.48816456e-02 -6.01626337e-02 3.99862647e-01
1.87905371e-01 2.07286209e-01 -7.07572877e-01 8.90423432e-02
7.63729930e-01 6.31544411e-01 -9.11872163e-02 9.63054597e-01
-1.54750168e-01 5.48777759e-01 -2.39316914e-02 6.76715225e-02
1.97154749e-02 1.34659350e-01 1.01477754e+00 1.21365333e+00
3.68746638e-01 -5.29037416e-01 -4.27109569e-01 6.27266049e-01
-2.83268660e-01 1.83253065e-01 -1.00981700e+00 -4.31049198e-01
2.66547382e-01 1.17744720e+00 -2.06959069e-01 -2.28553079e-02
-2.39519686e-01 1.38539112e+00 -4.13361013e-01 1.59102544e-01
-1.27564573e+00 -4.74305689e-01 6.95246458e-01 1.92082062e-01
1.23206250e-01 -3.28401364e-02 1.46333858e-01 -1.21346188e+00
-1.81179807e-01 -1.09884810e+00 3.01550981e-03 -1.25150406e+00
-1.17186630e+00 1.74190015e-01 -4.75762337e-01 -1.23713326e+00
5.11041097e-02 -4.86973196e-01 -3.69167507e-01 5.72673738e-01
-1.42066896e+00 -1.42726123e+00 -4.58532870e-01 7.83479154e-01
5.88129342e-01 -3.65475863e-01 4.86322641e-01 5.48998058e-01
-2.74041563e-01 9.08497155e-01 6.36989847e-02 3.58133107e-01
1.23709881e+00 -3.36239219e-01 -2.80487210e-01 5.08547485e-01
-2.26977468e-01 -1.09087182e-02 8.08525980e-01 -4.71198261e-01
-1.46141899e+00 -6.90790176e-01 8.55987728e-01 -1.13613419e-01
6.64014637e-01 -2.57181704e-01 -7.43089616e-01 6.31930456e-02
4.32472080e-01 -4.27210301e-01 4.61000890e-01 -5.92134178e-01
-5.49185097e-01 -7.83438906e-02 -1.49224257e+00 3.84512067e-01
5.95107615e-01 -7.74007022e-01 -3.94492179e-01 2.35899389e-01
4.67228472e-01 -5.53142987e-02 -2.64313459e-01 3.35018128e-01
1.01872838e+00 -1.37200737e+00 9.37524259e-01 -4.29409325e-01
4.89724785e-01 3.65973487e-02 -3.56944323e-01 -1.28636801e+00
4.61081296e-01 -8.69949818e-01 2.11835857e-02 1.38199306e+00
-1.12035684e-01 -5.28691411e-01 5.67618251e-01 -3.88174020e-02
-1.06218606e-01 -1.02365367e-01 -1.19599688e+00 -6.51826978e-01
-3.14326704e-01 -6.19027257e-01 3.38887960e-01 1.25520015e+00
4.12198044e-02 1.01946019e-01 -1.06663835e+00 4.70486060e-02
7.15966225e-01 -2.38362625e-01 9.97932017e-01 -7.81646073e-01
1.12375915e-01 -4.21588570e-01 -6.28046930e-01 -9.13047373e-01
4.06383216e-01 -4.65490192e-01 -4.12423164e-02 -7.69603908e-01
5.04245497e-02 8.86212513e-02 3.76563780e-02 2.57340401e-01
3.20578128e-01 6.53029382e-01 3.10114056e-01 1.59894675e-01
-2.46270716e-01 6.15041792e-01 1.31777966e+00 -4.14867073e-01
1.80942506e-01 -1.78655788e-01 -3.30345005e-01 9.55290854e-01
6.13361299e-01 -3.58512431e-01 7.27491230e-02 -7.83195049e-02
3.15408587e-01 4.22999859e-01 8.31208110e-01 -8.37316871e-01
-1.41894519e-01 1.28777782e-02 1.00475006e-01 -7.03170180e-01
8.51276815e-01 -7.40835249e-01 -3.19494724e-01 2.52711982e-01
-2.27198720e-01 -2.94893652e-01 2.01209918e-01 6.95177078e-01
-4.35760826e-01 -1.63932011e-01 1.15179944e+00 2.52619255e-02
-6.77114546e-01 -1.96557552e-01 -4.26059425e-01 2.97198500e-02
8.70663106e-01 -2.40907967e-01 -6.16676450e-01 -9.89322424e-01
-5.02574503e-01 -2.78504550e-01 3.76455277e-01 6.26180470e-01
1.06787527e+00 -1.29176486e+00 -7.31201470e-01 2.18349546e-01
1.84001774e-01 -9.33236718e-01 5.89959145e-01 1.01790094e+00
-3.04977983e-01 2.31762797e-01 -3.66944283e-01 -4.70332652e-01
-1.61167538e+00 5.52009523e-01 3.15967619e-01 3.69427890e-01
-3.47942650e-01 7.82132030e-01 1.36403553e-02 -3.48654352e-02
3.39600265e-01 8.46307054e-02 -6.33360147e-02 4.49555695e-01
6.05443001e-01 6.69453263e-01 5.62243769e-03 -1.20183039e+00
-3.95518631e-01 2.19918266e-01 3.81615400e-01 -5.99761605e-02
8.31141651e-01 -5.34568846e-01 -7.31447488e-02 3.21131617e-01
1.49023223e+00 4.04509097e-01 -7.17014313e-01 1.81366429e-01
-5.54740787e-01 -9.01733816e-01 1.98470533e-01 -8.66827190e-01
-1.03293121e+00 1.32329714e+00 7.00650036e-01 2.19638720e-01
9.37743187e-01 -5.97744733e-02 1.43562663e+00 3.89800183e-02
3.19850564e-01 -1.22503829e+00 6.38708234e-01 1.52549809e-02
1.15174031e+00 -1.47668350e+00 -4.18305874e-01 -1.94417194e-01
-9.36094224e-01 1.16968644e+00 2.66319662e-01 3.10435593e-01
4.83742148e-01 -4.17910330e-02 7.25625753e-02 1.98259175e-01
-1.74356133e-01 -3.51867974e-02 7.49736130e-02 8.24278235e-01
5.27865775e-02 1.78395212e-02 -1.19233958e-01 3.91295016e-01
2.16964856e-02 2.64081322e-02 6.71588123e-01 3.92697424e-01
-4.79407489e-01 -6.84681952e-01 -6.96199656e-01 5.64001128e-02
-6.57636702e-01 -1.25430718e-01 -6.00784361e-01 8.21204722e-01
3.17649662e-01 1.30459857e+00 -2.74580032e-01 -6.22850895e-01
-1.35219529e-01 3.84957671e-01 4.84984398e-01 2.67771751e-01
-3.17125648e-01 1.80778310e-01 2.84561068e-01 -6.18885994e-01
-5.37101030e-01 -5.13386726e-01 -6.16048396e-01 -7.21852243e-01
-5.19029617e-01 -3.45009625e-01 1.12134075e+00 9.37966108e-01
1.59183457e-01 -2.60318547e-01 6.32360816e-01 -1.07264256e+00
-4.00642931e-01 -9.80171978e-01 -4.74962801e-01 6.52623296e-01
6.51497662e-01 -8.91051412e-01 -7.39131987e-01 1.09143788e-02] | [12.956488609313965, 1.2933027744293213] |
498307bb-1013-44ff-9a28-2c5719b99cd4 | recovery-of-the-fetal-electrocardiogram-for | 1904.09525 | null | https://arxiv.org/abs/1904.09525v2 | https://arxiv.org/pdf/1904.09525v2.pdf | Recovery of the fetal electrocardiogram for morphological analysis from two trans-abdominal channels via optimal shrinkage | We propose a novel algorithm to recover fetal electrocardiogram (ECG) for both the fetal heart rate analysis and morphological analysis of its waveform from two or three trans-abdominal maternal ECG channels. We design an algorithm based on the optimal-shrinkage and the nonlocal Euclidean median under the wave-shape manifold model. For the fetal heart rate analysis, the algorithm is evaluated on publicly available database, 2013 PhyioNet/Computing in Cardiology Challenge, set A. For the morphological analysis, we propose to simulate semi-real databases by mixing the MIT-BIH Normal Sinus Rhythm Database and MITDB Arrhythmia Database. For the fetal R peak detection, the proposed algorithm outperforms all algorithms under comparison. For the morphological analysis, the algorithm provides an encouraging result in recovery of the fetal ECG waveform, including PR, QT and ST intervals, even when the fetus has arrhythmia. To the best of our knowledge, this is the first work focusing on recovering the fetal ECG for morphological analysis from two or three channels with an algorithm potentially applicable for continuous fetal electrocardiographic monitoring, which creates the potential for long term monitoring purpose. | ['Hau-Tieng Wu', 'Piers Barker', 'Salim Idriss', 'Stephen Miller', 'Pei-Chun Su'] | 2019-04-21 | null | null | null | null | ['morphological-analysis'] | ['natural-language-processing'] | [ 4.76100743e-01 1.74677655e-01 3.62116784e-01 -2.34527230e-01
-7.59098828e-01 -7.67761350e-01 -2.51005471e-01 2.18722299e-01
7.96494633e-02 5.62830746e-01 -2.60960579e-01 -4.05533493e-01
-5.43734133e-01 -4.91779327e-01 -4.07591999e-01 -9.38372135e-01
-7.02746570e-01 3.63179177e-01 -5.25923312e-01 3.38933975e-01
7.30919242e-02 9.08440828e-01 -8.69111896e-01 7.75531754e-02
8.25847328e-01 1.03008723e+00 -3.58682662e-01 1.12341702e+00
5.87245524e-01 1.90622672e-01 -4.81838495e-01 -2.82093346e-01
5.17367959e-01 -1.00375581e+00 -4.17558968e-01 -2.47986168e-01
2.11803928e-01 -2.53210515e-01 -1.09336466e-01 6.06632411e-01
1.36759174e+00 -2.06875652e-01 7.81824827e-01 -6.85435891e-01
2.11173490e-01 5.20328701e-01 -9.20201898e-01 6.39360309e-01
3.88823837e-01 -4.35803711e-01 3.23035359e-01 -1.16766965e+00
5.84839880e-01 4.59959745e-01 1.25651360e+00 2.20938936e-01
-1.03439319e+00 -6.45623088e-01 -8.77293169e-01 -1.84989814e-02
-1.85011196e+00 -4.79656875e-01 7.94526935e-01 -4.12652791e-01
4.72719282e-01 5.93671322e-01 9.08166111e-01 4.38722998e-01
1.45515293e-01 3.88071053e-02 9.93554533e-01 -5.53508043e-01
-3.68838795e-02 -2.35352114e-01 -6.92380667e-02 8.26599181e-01
5.05668700e-01 -3.96000482e-02 -4.21956033e-01 -6.19647801e-01
8.70707273e-01 -3.86202723e-01 -5.11273205e-01 -1.49013966e-01
-1.18780315e+00 3.58436853e-01 -4.40035492e-01 5.43866158e-01
-7.21483171e-01 3.14298235e-02 4.34337914e-01 6.57392859e-01
4.73173797e-01 3.38937730e-01 -2.43246153e-01 -2.27039620e-01
-1.48582876e+00 6.40908554e-02 8.52033734e-01 8.92934978e-01
3.56164902e-01 4.69039120e-02 -6.62717298e-02 7.53646135e-01
4.75866228e-01 6.71020806e-01 3.26528788e-01 -1.17495668e+00
3.37911963e-01 3.71842980e-02 -3.03601995e-02 -1.11189437e+00
-8.22750866e-01 -4.80347574e-01 -1.12069190e+00 -2.08671704e-01
4.47808117e-01 -7.39292264e-01 -6.26548588e-01 1.35582852e+00
5.40782988e-01 8.50757957e-01 2.81106859e-01 6.78418934e-01
1.21182442e+00 2.88342178e-01 -6.29617751e-01 -1.18149757e+00
1.33644605e+00 -3.91338132e-02 -1.03797483e+00 5.99829495e-01
6.18888497e-01 -7.99399495e-01 -1.02065466e-02 8.41298044e-01
-1.47364450e+00 -3.12241524e-01 -1.00545359e+00 5.12041867e-01
3.42998773e-01 2.92020947e-01 4.25222144e-02 1.08879602e+00
-1.12295210e+00 8.05588782e-01 -1.15237415e+00 -3.14477891e-01
3.27561468e-01 2.00593635e-01 -5.38513958e-01 8.44669268e-02
-1.00578415e+00 9.60583508e-01 -5.13073755e-03 3.67612332e-01
-4.42581475e-01 -8.77925873e-01 -8.19668353e-01 -2.04816222e-01
-3.12682331e-01 -3.88327628e-01 6.36361599e-01 -4.27991599e-01
-1.29947293e+00 9.91356254e-01 -2.39287406e-01 -3.62367392e-01
6.98602140e-01 1.48037091e-01 -5.15454829e-01 4.78497654e-01
-1.27943039e-01 -1.71830401e-01 1.01714420e+00 -9.32341158e-01
2.76215635e-02 -6.93158925e-01 -9.67873454e-01 1.07623570e-01
7.92427063e-02 1.20579734e-01 -1.86117306e-01 -8.18005443e-01
8.37634563e-01 -9.44164634e-01 -5.32135256e-02 -4.84735727e-01
-3.58818546e-02 5.33015728e-01 2.35485777e-01 -1.14977694e+00
1.56701052e+00 -2.34176469e+00 -1.45815872e-02 8.59575987e-01
2.77107477e-01 1.58746690e-01 2.59738445e-01 5.31610548e-01
-4.75648791e-01 9.17811468e-02 -6.37091756e-01 -2.06228673e-01
-6.82147503e-01 -1.93968080e-02 7.02745318e-02 1.40259397e+00
-1.99746117e-01 6.80228591e-01 -6.09696805e-01 -6.22892261e-01
-1.32642284e-01 4.52702433e-01 -1.93903521e-01 2.09074050e-01
1.03064740e+00 7.83839703e-01 -1.19882286e-01 4.41979110e-01
1.05629921e+00 2.70790279e-01 4.09116447e-01 -1.13338619e-01
-1.19873725e-01 -2.32714906e-01 -1.31567156e+00 2.04777575e+00
6.52513281e-02 6.36157632e-01 4.31771785e-01 -1.00758708e+00
1.09850192e+00 1.35247767e+00 8.41620862e-01 -1.11518279e-01
1.67097658e-01 4.02500153e-01 4.06986028e-01 -9.75794017e-01
-3.31329584e-01 -6.23870134e-01 4.75236952e-01 6.96886659e-01
7.07822889e-02 -2.78210551e-01 2.04744786e-01 -1.29909560e-01
1.11136556e+00 -2.03332398e-02 6.14944756e-01 -6.84125841e-01
5.34276426e-01 -8.00041616e-01 6.16062164e-01 5.86775422e-01
-7.34539479e-02 1.37908530e+00 4.95493114e-01 -5.42326570e-01
-5.08722365e-01 -7.92609394e-01 -7.28929579e-01 2.38338336e-01
-1.81623593e-01 -1.65115565e-01 -7.41098106e-01 -3.97781938e-01
-3.53693590e-02 2.97575712e-01 -7.11415112e-01 1.56895831e-01
-9.11196053e-01 -1.09662044e+00 1.18065572e+00 3.10612023e-01
-1.61110312e-01 -8.28446150e-01 -9.44411397e-01 4.18041021e-01
-2.64473915e-01 -7.16510952e-01 -2.31349036e-01 -1.57268345e-01
-1.28797579e+00 -1.06033719e+00 -1.14339924e+00 -6.70889676e-01
5.77921808e-01 -4.34180737e-01 8.58938158e-01 2.18353793e-01
-6.09245002e-01 2.79291958e-01 -2.67740548e-01 -7.10076213e-01
-4.86136556e-01 -2.96611339e-01 1.80943340e-01 5.23052931e-01
-4.58600551e-01 -1.06008422e+00 -8.96095872e-01 2.51395822e-01
-4.96415168e-01 -2.87325472e-01 3.17757219e-01 4.07224953e-01
5.23194075e-01 -3.06499571e-01 1.05712235e+00 -9.70023215e-01
4.30369437e-01 -7.54827321e-01 -2.23164305e-01 1.10822894e-01
-1.05822051e+00 -5.55213928e-01 3.05344820e-01 9.78376437e-03
-8.97644639e-01 -9.59846675e-02 -3.40689898e-01 -2.63586909e-01
-1.19277298e-01 5.96298158e-01 2.10337073e-01 -2.04769269e-01
8.61589670e-01 1.61821544e-01 7.21155852e-02 -4.07523215e-01
3.98509344e-03 4.99354243e-01 6.65208042e-01 -4.65995848e-01
5.52241683e-01 5.36520481e-01 5.37785709e-01 -9.44137871e-01
-2.37170048e-02 -7.23540723e-01 -8.04357052e-01 -1.16959259e-01
5.84863603e-01 -5.28274000e-01 -2.10131496e-01 3.28014940e-01
-1.33966267e+00 1.50689155e-01 -3.81596923e-01 8.77880454e-01
-6.95165455e-01 6.93089306e-01 -5.67103326e-01 -9.16027546e-01
-8.94263923e-01 -7.39461541e-01 4.34202701e-01 -1.89763814e-01
-2.99379200e-01 -9.91090059e-01 5.60786784e-01 -7.36263916e-02
2.95019567e-01 1.11568558e+00 4.96359378e-01 -5.78508615e-01
2.38706484e-01 -5.44140816e-01 2.25921974e-01 2.90148586e-01
3.23198974e-01 1.92284867e-01 -9.60480213e-01 -5.32313824e-01
4.50539410e-01 4.27433223e-01 3.42010289e-01 6.00269496e-01
7.71651804e-01 2.20964346e-02 -8.17292407e-02 1.12058008e+00
1.22420931e+00 4.18497890e-01 7.46550739e-01 -4.77054954e-01
4.66894001e-01 4.60777014e-01 5.86223125e-01 9.46898282e-01
1.14524096e-01 1.78490523e-02 1.30398661e-01 -3.67060512e-01
8.24938156e-03 4.28176194e-01 -2.68475085e-01 1.01732159e+00
-8.73864472e-01 2.42465679e-02 -1.16575623e+00 6.98032260e-01
-1.60387409e+00 -8.39236200e-01 -5.74201465e-01 2.35056734e+00
7.90078342e-01 -5.09746373e-01 -3.66800725e-02 5.28949678e-01
7.41635799e-01 -1.73550278e-01 -1.69445127e-01 -5.50790727e-01
-1.24955866e-02 7.42550731e-01 2.31336400e-01 3.04131210e-01
-6.79287016e-01 -1.93009526e-01 6.29382038e+00 2.44513080e-01
-1.19585657e+00 4.46364015e-01 5.16379952e-01 1.39840767e-01
2.52172034e-02 -1.82909459e-01 -4.33331430e-02 1.60928279e-01
1.13829327e+00 -2.67483681e-01 1.20615736e-01 1.40138820e-01
3.22240382e-01 6.96967542e-02 -1.10826254e+00 1.34349489e+00
1.96588174e-01 -1.24237311e+00 -6.20349288e-01 -2.31033951e-01
7.21506596e-01 -1.55059561e-01 -2.81498641e-01 -3.93450975e-01
-1.25432909e+00 -1.19926035e+00 4.44033951e-01 8.89446199e-01
1.58162761e+00 -7.70249546e-01 7.78625309e-01 4.33644265e-01
-1.07580626e+00 3.46087903e-01 -2.08487026e-02 2.22057134e-01
-7.54065663e-02 6.41217470e-01 -9.92518365e-01 8.52910936e-01
4.23256546e-01 9.28688228e-01 -2.31086299e-01 1.34764898e+00
1.02215089e-01 1.12065530e+00 -3.61839920e-01 7.91504920e-01
-5.56272149e-01 -3.81850690e-01 9.20797884e-01 1.45056248e+00
7.64864087e-01 7.41229415e-01 -4.28878576e-01 4.86580521e-01
1.27803519e-01 5.80252230e-01 -9.35203254e-01 4.66198474e-01
3.92097771e-01 1.39595640e+00 -8.76578987e-01 -2.55204022e-01
-3.22596788e-01 4.40238804e-01 -3.63130659e-01 4.77944911e-01
-5.93942285e-01 -8.27563107e-01 -2.07833827e-01 4.09118831e-01
-2.31319115e-01 2.12924570e-01 -8.27489257e-01 -8.25594962e-01
6.07073195e-02 -9.12228346e-01 6.37250006e-01 -2.71170735e-01
-5.70188344e-01 5.03820598e-01 3.56340595e-02 -1.26026797e+00
-4.28587347e-01 2.63728704e-02 -9.86576676e-01 1.08741522e+00
-1.05789828e+00 -5.53288043e-01 -4.85199392e-02 4.50797588e-01
-1.23116702e-01 -1.31385192e-01 1.09390342e+00 6.86312199e-01
-1.56050295e-01 7.95988262e-01 -2.12853521e-01 5.33870943e-02
5.82705855e-01 -1.32237804e+00 6.13478906e-02 8.55648935e-01
-6.93563148e-02 8.77252340e-01 8.10939610e-01 -4.81168687e-01
-1.44428849e+00 -7.60714710e-01 7.32865632e-01 -4.35908139e-01
-1.55104980e-01 1.01460874e-01 -9.24818575e-01 5.50198376e-01
1.60709202e-01 5.58360279e-01 1.03761148e+00 -4.17858481e-01
4.02609587e-01 -3.63315165e-01 -1.47293818e+00 -2.32630536e-01
5.61017454e-01 4.34147604e-02 -5.77262282e-01 1.12817280e-01
-1.56276286e-01 -7.18306243e-01 -1.65765452e+00 1.01239073e+00
1.05269265e+00 -8.56989980e-01 6.18944764e-01 -1.90460205e-01
-7.43744569e-03 -3.17234725e-01 2.85939068e-01 -9.92465615e-01
-1.36118978e-01 -1.67114484e+00 -2.72440404e-01 1.18966329e+00
6.11817777e-01 -7.72027314e-01 4.82436210e-01 1.50814354e-01
-1.61148071e-01 -8.05332661e-01 -1.24379551e+00 -2.11136699e-01
-1.77815840e-01 -4.84273940e-01 2.93675244e-01 9.64325786e-01
1.79984733e-01 -1.19282000e-01 -3.41855913e-01 4.62197244e-01
7.41730154e-01 2.42625445e-01 3.40850621e-01 -1.16325462e+00
-1.87610283e-01 -1.06283091e-02 -4.29156959e-01 -8.60134289e-02
-3.32165956e-01 -8.70045662e-01 -8.30207616e-02 -1.29272556e+00
-6.68777972e-02 -5.87314487e-01 -6.12156510e-01 -3.69622484e-02
-2.18201846e-01 5.75060606e-01 -2.53928840e-01 -1.41182452e-01
1.24886110e-02 -3.20004970e-01 1.13203251e+00 6.61806405e-01
-2.87883282e-01 5.53828537e-01 -4.06505197e-01 8.83012176e-01
6.63641751e-01 -8.12585950e-01 -4.02648956e-01 4.33718152e-02
3.80429506e-01 1.14022517e+00 -1.00657463e-01 -8.66546571e-01
-5.60841486e-02 2.68015295e-01 4.76630598e-01 -7.04124272e-01
2.49923021e-02 -5.82653582e-01 5.96277833e-01 5.47423601e-01
-8.53728205e-02 4.62516725e-01 8.81701782e-02 2.09226906e-01
-1.82060763e-01 -2.49939963e-01 1.08856857e+00 -3.16826478e-02
4.52061921e-01 5.39416671e-01 -6.49519265e-01 3.16080511e-01
9.21301961e-01 -2.88833559e-01 3.39881301e-01 -3.42202961e-01
-1.32980716e+00 -1.21513261e-02 6.01743683e-02 -2.44129822e-01
9.93460655e-01 -1.05948400e+00 -1.46371293e+00 5.81249535e-01
-2.12479949e-01 1.10201901e-02 3.29358131e-01 1.85444129e+00
-1.20214987e+00 -1.72447726e-01 -3.04930843e-02 -6.94124043e-01
-1.45550704e+00 1.80495486e-01 5.67736745e-01 9.37456936e-02
-7.44568348e-01 8.71333182e-01 -1.40701443e-01 -2.72952467e-02
1.96643304e-02 -2.33064994e-01 -3.87505054e-01 3.61376792e-01
5.91604710e-01 9.68357384e-01 5.83566844e-01 -6.60555363e-01
-4.59120721e-01 8.76621604e-01 6.53826892e-01 -1.97889194e-01
1.16771686e+00 -4.48258072e-01 -6.26648247e-01 5.00940502e-01
1.16970229e+00 4.68180776e-01 -5.40143251e-01 2.69526839e-01
-9.53777134e-02 -3.95931184e-01 -3.24774057e-01 -5.38489580e-01
-1.22181535e+00 1.03312886e+00 9.25367475e-01 3.85445207e-01
1.34422970e+00 -4.72501159e-01 3.96526754e-01 1.87388330e-03
2.64908314e-01 -6.07380688e-01 -4.82751459e-01 -5.27794920e-02
1.09451377e+00 -5.45102358e-01 3.39509130e-01 -1.80010676e-01
-4.08188760e-01 1.38421464e+00 -2.24219456e-01 -2.38815203e-01
9.36224997e-01 5.62964082e-01 5.60496688e-01 -2.34039128e-01
-3.98078650e-01 3.79932761e-01 4.99531001e-01 5.25791883e-01
7.65371144e-01 1.35225996e-01 -9.49318826e-01 4.93545055e-01
-8.76112841e-03 4.32521142e-02 7.91976213e-01 8.47491622e-01
-1.36058450e-01 -7.10991919e-01 -4.58494633e-01 5.27229249e-01
-1.44040155e+00 -2.05212638e-01 7.34828264e-02 3.66815388e-01
2.95507610e-01 9.01853502e-01 -5.74829578e-01 3.89217943e-01
3.66132647e-01 3.46131265e-01 8.20369720e-01 -7.03766704e-01
-6.45634294e-01 4.50227767e-01 1.00333996e-01 -3.35968435e-01
-2.33743325e-01 -1.11523151e+00 -1.47180486e+00 3.58474165e-01
-5.67884803e-01 3.70256901e-01 8.71041059e-01 5.85890532e-01
2.55076766e-01 3.48536104e-01 8.57727230e-01 -4.87678707e-01
-2.45870098e-01 -1.13556302e+00 -1.31298637e+00 1.83920324e-01
7.41099417e-01 -1.44609645e-01 -1.60333291e-01 2.78190881e-01] | [14.185968399047852, 3.1508004665374756] |
d6a84008-44d0-4cf7-b138-9a0344e3613d | uod-universal-one-shot-detection-of | 2306.07615 | null | https://arxiv.org/abs/2306.07615v3 | https://arxiv.org/pdf/2306.07615v3.pdf | UOD: Universal One-shot Detection of Anatomical Landmarks | One-shot medical landmark detection gains much attention and achieves great success for its label-efficient training process. However, existing one-shot learning methods are highly specialized in a single domain and suffer domain preference heavily in the situation of multi-domain unlabeled data. Moreover, one-shot learning is not robust that it faces performance drop when annotating a sub-optimal image. To tackle these issues, we resort to developing a domain-adaptive one-shot landmark detection framework for handling multi-domain medical images, named Universal One-shot Detection (UOD). UOD consists of two stages and two corresponding universal models which are designed as combinations of domain-specific modules and domain-shared modules. In the first stage, a domain-adaptive convolution model is self-supervised learned to generate pseudo landmark labels. In the second stage, we design a domain-adaptive transformer to eliminate domain preference and build the global context for multi-domain data. Even though only one annotated sample from each domain is available for training, the domain-shared modules help UOD aggregate all one-shot samples to detect more robust and accurate landmarks. We investigated both qualitatively and quantitatively the proposed UOD on three widely-used public X-ray datasets in different anatomical domains (i.e., head, hand, chest) and obtained state-of-the-art performances in each domain. | ['S. Kevin Zhou', 'Zaiyi Liu', 'Qingsong Yao', 'Quan Quan', 'Heqin Zhu'] | 2023-06-13 | null | null | null | null | ['one-shot-learning'] | ['methodology'] | [ 2.19089642e-01 1.51893616e-01 -4.54468250e-01 -4.18375194e-01
-1.39511442e+00 -1.30190194e-01 4.19154704e-01 2.40834996e-01
-5.99787295e-01 5.13172448e-01 7.60350972e-02 1.26134977e-01
-1.75711408e-01 -6.45846725e-01 -1.95851400e-01 -8.01785290e-01
2.23445341e-01 5.59936285e-01 6.77075148e-01 -5.53149693e-02
-9.69207361e-02 3.49292457e-01 -1.25797606e+00 1.13121957e-01
8.93004298e-01 9.36177373e-01 5.94297349e-01 3.55286539e-01
-2.14491174e-01 6.22801542e-01 -2.17305452e-01 -1.44137889e-01
3.85741927e-02 -4.37920451e-01 -8.99412632e-01 2.99222648e-01
-9.80950966e-02 -2.18432084e-01 -1.69914797e-01 1.14967549e+00
9.93996859e-01 3.20863426e-01 8.26436222e-01 -7.94698894e-01
-4.29899752e-01 3.30341697e-01 -6.75719619e-01 3.35038543e-01
3.42935957e-02 1.17689751e-01 9.48951185e-01 -1.13746560e+00
9.72836018e-01 6.68242514e-01 5.49508691e-01 7.87932098e-01
-9.54155207e-01 -4.07530397e-01 5.50596928e-03 2.32492939e-01
-1.55493033e+00 -3.73930752e-01 9.95331228e-01 -4.33230639e-01
4.49178576e-01 -4.12574895e-02 8.50532800e-02 8.99915040e-01
2.14087516e-01 8.51440072e-01 8.30465138e-01 -3.61772239e-01
5.33141434e-01 1.53431207e-01 2.28024021e-01 7.44448245e-01
-1.27177149e-01 -5.62788509e-02 -1.84147686e-01 -2.01775968e-01
6.64785862e-01 3.97837967e-01 -7.91507587e-02 -6.37551486e-01
-1.14744246e+00 7.96867609e-01 7.93835938e-01 7.29488552e-01
-5.37062049e-01 -4.54543173e-01 5.83676398e-01 4.35794406e-02
4.93018448e-01 3.28837961e-01 -1.90555900e-01 2.10661799e-01
-9.99308050e-01 -3.18805516e-01 5.10537922e-01 8.72471035e-01
8.05218101e-01 -2.93863893e-01 -4.70736742e-01 1.30692589e+00
2.05546897e-02 -9.92199928e-02 8.79438877e-01 -3.45192909e-01
1.68349385e-01 7.16157556e-01 -9.54485238e-02 -5.05589843e-01
-8.63768339e-01 -5.48223734e-01 -8.63903344e-01 -1.93499569e-02
3.87018085e-01 -2.04083592e-01 -1.42632461e+00 1.65400982e+00
6.05653644e-01 2.05659389e-01 2.24322677e-02 1.12064219e+00
1.20376813e+00 2.69460052e-01 3.66128564e-01 -2.66967565e-01
1.63595200e+00 -1.19667840e+00 -5.54015100e-01 -4.01167601e-01
8.43809247e-01 -6.02252603e-01 1.01554644e+00 -1.36235192e-01
-7.39167809e-01 -5.19604981e-01 -1.14916730e+00 -7.76505023e-02
-3.57617348e-01 7.44780004e-02 3.06783110e-01 3.37089092e-01
-5.81663072e-01 3.96822035e-01 -8.67901206e-01 -4.50716972e-01
7.29126394e-01 1.94469377e-01 -5.20575464e-01 -4.63263184e-01
-1.06138122e+00 8.99441719e-01 3.52642953e-01 -3.72943640e-01
-1.21599519e+00 -6.43535256e-01 -1.00073004e+00 -4.53995019e-02
7.28701293e-01 -5.56454897e-01 1.24958599e+00 -6.92837358e-01
-1.24529707e+00 1.18694472e+00 -1.32763460e-02 -1.73189774e-01
4.95553166e-01 3.07495624e-01 -5.50223768e-01 3.42652708e-01
4.65065211e-01 4.96751308e-01 8.41560245e-01 -1.16129279e+00
-6.74446166e-01 -5.29257178e-01 -6.87728375e-02 2.57444680e-01
-4.21675861e-01 -1.73174050e-02 -6.27894580e-01 -7.42643297e-01
2.99365789e-01 -7.12628126e-01 -5.74116230e-01 3.22414964e-01
-3.27100039e-01 -1.10303327e-01 6.11303449e-01 -3.96187425e-01
1.16500902e+00 -2.40500784e+00 1.17719427e-01 -8.87021422e-03
3.89878482e-01 2.36165166e-01 -2.86130346e-02 5.08647598e-02
-1.14333361e-01 -4.41177934e-01 -5.66319883e-01 -4.74235266e-01
-3.12574536e-01 4.25464869e-01 4.04285818e-01 5.93556345e-01
2.52242267e-01 6.87452316e-01 -1.18722796e+00 -1.08805621e+00
3.16449255e-01 1.75417751e-01 -3.98117572e-01 2.14168712e-01
6.86137239e-03 6.67357981e-01 -5.99706590e-01 9.57941055e-01
5.72030902e-01 -5.80800235e-01 7.08614290e-02 -1.04599826e-01
2.34134272e-02 -2.39493877e-01 -1.08791161e+00 2.28205085e+00
-7.19370127e-01 4.18159738e-02 -1.06625780e-02 -1.01490915e+00
7.77089894e-01 4.92983043e-01 9.02976632e-01 -7.83054709e-01
3.52287024e-01 4.99509066e-01 -1.20566944e-02 -5.28506935e-01
5.62273413e-02 -6.53497577e-01 -1.89868525e-01 4.49996889e-01
6.06423259e-01 1.46480039e-01 3.80376838e-02 1.04989424e-01
1.23686397e+00 -2.05180094e-01 7.69575000e-01 -2.01988578e-01
7.01291084e-01 8.13841671e-02 7.59630620e-01 5.99550664e-01
-6.83928788e-01 1.01036978e+00 1.88305005e-01 -4.41133887e-01
-7.63656497e-01 -1.02746010e+00 -3.97979140e-01 1.44642246e+00
4.67490703e-01 -1.44816533e-01 -5.55091619e-01 -1.01091027e+00
-2.57625431e-01 5.48042834e-01 -8.67259145e-01 -3.96864504e-01
-2.52300709e-01 -8.72060120e-01 2.72189856e-01 5.58400154e-01
3.60378176e-01 -1.01136363e+00 -8.48441780e-01 3.67125571e-01
-8.46572146e-02 -8.54508996e-01 -5.76764226e-01 5.49147129e-01
-7.27071404e-01 -1.23890460e+00 -1.44576478e+00 -1.05561757e+00
9.47790921e-01 2.34954447e-01 8.87759030e-01 -1.93783805e-01
-6.04551911e-01 2.74683028e-01 -4.80957687e-01 -3.68294656e-01
-1.48834288e-01 1.44576311e-01 -1.61528438e-01 2.10638523e-01
4.37765867e-01 -3.14908117e-01 -7.00267315e-01 5.68872631e-01
-7.75121689e-01 -2.45207131e-01 8.42075288e-01 1.18846619e+00
8.44249547e-01 -1.58437386e-01 6.48028970e-01 -1.01312184e+00
3.96477610e-01 -6.58699691e-01 -7.81783909e-02 4.12967563e-01
-3.87067497e-01 -6.20880835e-02 6.41562283e-01 -4.55707341e-01
-1.06908059e+00 3.38727653e-01 -2.75806427e-01 -4.63659286e-01
-1.92863032e-01 3.91116112e-01 1.49076022e-02 -1.98811889e-01
9.31451976e-01 1.94917515e-01 9.70374793e-02 -5.36270797e-01
2.91465938e-01 6.57102942e-01 6.42154157e-01 -3.62107903e-01
5.26229918e-01 6.65935636e-01 -2.72026658e-01 -7.48287916e-01
-1.00655723e+00 -1.00777411e+00 -8.45531881e-01 -1.57484144e-01
9.88804579e-01 -8.29247057e-01 1.97285727e-01 3.98441702e-01
-7.87792742e-01 -6.46251738e-02 -6.22794092e-01 4.82127368e-01
-5.99669695e-01 2.92308986e-01 -4.72069770e-01 -4.27442193e-01
-5.18255234e-01 -1.35495996e+00 1.20034242e+00 3.61543506e-01
-8.37133303e-02 -8.54620755e-01 2.77792633e-01 -8.61734003e-02
2.33624220e-01 7.06352443e-02 8.16068590e-01 -9.90766644e-01
-1.18235424e-02 -4.70402151e-01 -3.30979764e-01 1.82662025e-01
2.89740235e-01 -5.46384275e-01 -1.05781078e+00 -3.05468738e-01
1.34644493e-01 -4.04768169e-01 7.77996123e-01 5.01373529e-01
1.01192904e+00 1.62183106e-01 -5.33881903e-01 5.37397087e-01
1.51552093e+00 1.59822017e-01 3.34594220e-01 1.45804673e-01
4.54526484e-01 4.78216469e-01 9.41349208e-01 5.65145969e-01
2.52460867e-01 6.25199616e-01 2.52279043e-01 -3.90115380e-01
-1.60689935e-01 -1.07516438e-01 -2.91892439e-01 5.18604457e-01
1.32295966e-01 1.64540187e-01 -1.12564099e+00 7.74006367e-01
-1.75464153e+00 -6.06822431e-01 4.52649117e-01 2.05175781e+00
7.83863783e-01 2.23536968e-01 2.56591856e-01 -9.43690073e-03
7.62551248e-01 2.78648704e-01 -6.44426346e-01 1.78628832e-01
2.43476421e-01 2.24517673e-01 3.48392248e-01 1.09385908e-01
-1.37747467e+00 8.29112530e-01 5.38967037e+00 1.07862449e+00
-1.13081837e+00 7.90323198e-01 5.59100747e-01 -1.05003014e-01
7.93819875e-02 -2.09751606e-01 -5.06209075e-01 4.96102482e-01
5.00441968e-01 -1.76669866e-01 -3.45094204e-01 1.34770405e+00
-9.18301120e-02 -2.02169865e-01 -9.24566269e-01 1.13744843e+00
2.06044689e-02 -1.34252441e+00 -3.58551562e-01 -2.10309625e-01
6.24186039e-01 1.09987371e-01 -2.49931008e-01 5.01891732e-01
-9.15631559e-03 -6.19482398e-01 2.84592241e-01 3.00537497e-01
1.12613559e+00 -6.58764124e-01 9.00875211e-01 5.12592256e-01
-1.19860351e+00 -1.74502999e-01 -5.78308940e-01 3.00259590e-01
2.77778000e-01 4.10101444e-01 -7.94032216e-01 7.21259832e-01
4.98752654e-01 6.63801014e-01 -3.43326658e-01 1.27801096e+00
3.71082909e-02 1.81641072e-01 -1.71073332e-01 1.92387894e-01
5.35582900e-01 2.40563273e-01 3.94967675e-01 9.31834877e-01
7.41685405e-02 3.57137114e-01 4.04126048e-01 4.93003517e-01
-3.48790735e-02 3.04368049e-01 -4.58358586e-01 3.84609729e-01
3.77292365e-01 1.51795602e+00 -9.87195194e-01 -5.42220116e-01
-6.36890590e-01 9.87607658e-01 2.31080681e-01 7.13546276e-02
-7.44521320e-01 -4.51175511e-01 4.25682664e-01 1.93464980e-01
9.41503197e-02 1.15070008e-01 -3.66456091e-01 -1.06722653e+00
-2.89673954e-01 -3.85767847e-01 8.73827279e-01 -3.20429385e-01
-1.47737694e+00 6.36000395e-01 -1.29955009e-01 -1.68549776e+00
-8.03670958e-02 -3.95121247e-01 -8.27437937e-01 6.25940263e-01
-1.53599370e+00 -1.04395759e+00 -5.72194517e-01 7.49926746e-01
8.08665693e-01 -1.68343425e-01 1.00983202e+00 7.27540553e-01
-7.03325748e-01 7.27820992e-01 7.99212679e-02 2.85338253e-01
1.03162718e+00 -1.14719200e+00 1.94027424e-02 7.95200408e-01
-1.17878154e-01 3.95470828e-01 4.01816517e-01 -5.66711783e-01
-1.02548730e+00 -1.00575745e+00 4.54487860e-01 -4.42057140e-02
5.23676872e-01 -1.13000154e-01 -9.98846531e-01 3.84655267e-01
-3.81682932e-01 6.66300654e-01 8.92531335e-01 -9.56525803e-02
7.75417918e-03 4.77903187e-02 -1.51549017e+00 3.44371438e-01
9.88717139e-01 -5.33818126e-01 -7.42451906e-01 4.45085168e-01
4.91512030e-01 -5.11431158e-01 -1.09378314e+00 4.31415886e-01
2.57761955e-01 -6.77504182e-01 9.30749297e-01 -4.98010397e-01
3.08651149e-01 -2.00602144e-01 -9.94864479e-02 -1.19627428e+00
-3.89465064e-01 -1.84434503e-01 1.08657077e-01 9.46109056e-01
2.95531720e-01 -4.66623157e-01 8.23817253e-01 5.16524136e-01
-4.68439519e-01 -8.79440606e-01 -1.20560884e+00 -7.18349099e-01
2.13708077e-02 -2.85151094e-01 3.13497365e-01 1.00163221e+00
9.83474329e-02 4.19018775e-01 -3.17702591e-01 2.09261454e-03
7.46370792e-01 1.38722897e-01 4.60961133e-01 -1.18950272e+00
-1.73599884e-01 -2.34152421e-01 -5.26726246e-01 -7.33289301e-01
-2.29122475e-01 -1.01238728e+00 3.03321540e-01 -1.64261556e+00
3.31479490e-01 -5.96645713e-01 -7.14109182e-01 5.67729473e-01
-3.29184592e-01 2.61529982e-01 -7.86565989e-02 3.59366000e-01
-8.03072512e-01 4.97484744e-01 1.39162743e+00 -1.52351409e-01
-1.89194173e-01 1.53199792e-01 -6.34285688e-01 7.28073537e-01
4.39989567e-01 -5.93445361e-01 -4.96380180e-01 -1.20459348e-01
-5.69062233e-01 1.76183388e-01 1.87917411e-01 -1.26178670e+00
4.92627621e-01 1.31132767e-01 3.09334666e-01 -5.23189247e-01
2.29532763e-01 -8.34545851e-01 -3.61698508e-01 5.44311047e-01
-1.54975027e-01 -5.39734244e-01 -9.39524397e-02 7.10262656e-01
-4.00651395e-01 -4.04535979e-01 1.40900302e+00 -4.83245403e-01
-1.32346869e+00 6.59301221e-01 5.47753610e-02 2.96118319e-01
1.41367352e+00 -2.26092800e-01 -2.65081823e-02 1.69070005e-01
-1.12395573e+00 3.78131270e-01 4.19608682e-01 3.84928703e-01
6.69322312e-01 -1.26027381e+00 -5.48435569e-01 2.63957530e-01
6.93671405e-01 2.17673823e-01 7.88156271e-01 1.08568859e+00
-2.43369728e-01 1.03475571e-01 -3.16050202e-01 -7.46988177e-01
-9.10013497e-01 8.33036423e-01 4.48892504e-01 -3.88138771e-01
-8.47418070e-01 1.00841439e+00 4.00948077e-01 -3.87707472e-01
2.32139677e-01 -5.15067615e-02 -3.18871707e-01 4.25308496e-01
7.37120867e-01 1.72964677e-01 2.24961787e-01 -6.49174392e-01
-6.89135432e-01 6.53813183e-01 -3.29881102e-01 2.11357251e-01
1.28155994e+00 1.21995201e-02 3.36857706e-01 4.43423897e-01
1.29898822e+00 -5.32905221e-01 -1.21131194e+00 -7.64276564e-01
2.05186263e-01 -2.75863081e-01 5.66243678e-02 -5.79847813e-01
-1.03092504e+00 8.87395024e-01 8.10623944e-01 -3.96977693e-01
1.19019699e+00 2.90510803e-01 8.76516163e-01 -9.91721228e-02
5.72642922e-01 -1.24605000e+00 2.89544493e-01 2.09667236e-01
4.11481142e-01 -1.63729823e+00 -5.79050928e-02 -2.99945086e-01
-9.67545569e-01 8.79215002e-01 6.99797869e-01 -1.09462112e-01
8.17618966e-01 8.96503776e-02 5.74267060e-02 -4.12888914e-01
-3.97297859e-01 -4.82949138e-01 4.02314961e-01 5.39416611e-01
2.04888329e-01 1.00249209e-01 -4.57247704e-01 9.06250238e-01
5.33341229e-01 2.74810523e-01 2.60224892e-03 1.21563900e+00
-6.92731023e-01 -9.05092776e-01 -1.97460175e-01 5.91881037e-01
-3.47406834e-01 1.57589793e-01 7.00697303e-02 6.81418002e-01
1.61700726e-01 5.96279621e-01 -1.67118445e-01 -3.82307738e-01
3.60296100e-01 1.09225772e-01 2.15594172e-01 -1.04802275e+00
-4.29516792e-01 1.35725215e-01 -2.51607686e-01 -5.45288861e-01
-2.23055989e-01 -4.32107002e-01 -1.45578241e+00 3.21038276e-01
-3.15873206e-01 -2.29412336e-02 3.63722265e-01 9.91320431e-01
1.43823355e-01 7.41380394e-01 6.10293508e-01 -7.60088980e-01
-5.89091778e-01 -8.69608462e-01 -7.39112020e-01 4.52889144e-01
1.63677901e-01 -1.11710119e+00 -6.19990155e-02 -4.33320075e-01] | [14.605606079101562, -2.0855934619903564] |
1e80334e-2934-4098-b8de-684d3cd4ad0f | dense-sparse-retrieval-using-sparse-language | 2304.00114 | null | https://arxiv.org/abs/2304.00114v1 | https://arxiv.org/pdf/2304.00114v1.pdf | Dense Sparse Retrieval: Using Sparse Language Models for Inference Efficient Dense Retrieval | Vector-based retrieval systems have become a common staple for academic and industrial search applications because they provide a simple and scalable way of extending the search to leverage contextual representations for documents and queries. As these vector-based systems rely on contextual language models, their usage commonly requires GPUs, which can be expensive and difficult to manage. Given recent advances in introducing sparsity into language models for improved inference efficiency, in this paper, we study how sparse language models can be used for dense retrieval to improve inference efficiency. Using the popular retrieval library Tevatron and the MSMARCO, NQ, and TriviaQA datasets, we find that sparse language models can be used as direct replacements with little to no drop in accuracy and up to 4.3x improved inference speeds | ['ChengXiang Zhai', 'Daniel Campos'] | 2023-03-31 | null | null | null | null | ['triviaqa'] | ['miscellaneous'] | [-2.44712636e-01 -6.77542925e-01 -7.28637278e-01 -2.73501366e-01
-1.23041153e+00 -6.51631176e-01 9.73489165e-01 3.74219030e-01
-5.07192552e-01 5.72387099e-01 3.40293497e-01 -7.58866370e-01
-2.40673140e-01 -7.80951619e-01 -4.01790440e-01 -4.47571307e-01
1.64696485e-01 7.13814020e-01 2.49760389e-01 -3.91317546e-01
4.86710876e-01 5.47182500e-01 -1.39384174e+00 2.51842707e-01
4.13495958e-01 8.39034140e-01 4.55201790e-02 6.00690901e-01
-4.61306036e-01 6.72719657e-01 -5.22113800e-01 -2.97008961e-01
2.38426238e-01 1.36015326e-01 -6.63142502e-01 -7.06419051e-01
4.94605273e-01 -4.53610033e-01 -1.00943029e+00 6.63361251e-01
6.19569123e-01 4.15626287e-01 4.61259514e-01 -7.82630920e-01
-4.94395107e-01 4.91643071e-01 -5.29535353e-01 4.41397935e-01
4.84669924e-01 -2.81477589e-02 1.45455992e+00 -9.29354370e-01
6.21100008e-01 1.35304987e+00 4.59569216e-01 1.09616779e-01
-1.23527551e+00 -7.26355195e-01 1.97189376e-02 3.37276191e-01
-1.64615452e+00 -4.60373223e-01 3.85519773e-01 6.27716035e-02
1.46865022e+00 5.24501562e-01 5.32589674e-01 9.65640962e-01
4.13936406e-01 8.91490042e-01 7.33632267e-01 -5.70993900e-01
8.23647305e-02 5.79942390e-02 4.94610488e-01 6.21518672e-01
3.61160934e-01 -2.06585359e-02 -7.52263784e-01 -7.95098424e-01
5.77766478e-01 2.09794417e-01 1.88353404e-01 -1.64200291e-02
-9.35245931e-01 1.23874998e+00 3.16385001e-01 2.31434658e-01
-3.11273396e-01 6.17486358e-01 5.95149100e-01 2.55148649e-01
5.89022040e-01 5.30484676e-01 -2.53346145e-01 -4.28434700e-01
-1.29957950e+00 7.99145341e-01 8.89159083e-01 9.36816990e-01
6.88525856e-01 3.79409157e-02 -3.97473305e-01 1.00658607e+00
3.38697582e-01 8.51723433e-01 3.92794788e-01 -7.80385911e-01
4.07979399e-01 3.97920221e-01 -1.41544584e-02 -1.04386818e+00
-2.43378848e-01 -5.18711090e-01 -4.85900283e-01 -7.08101332e-01
-4.67771851e-02 2.18285099e-01 -1.04666269e+00 1.23554528e+00
2.27111816e-01 9.36020613e-02 -2.21290559e-01 6.80875599e-01
7.12470591e-01 1.01022959e+00 1.06322661e-01 1.18662015e-01
1.40931165e+00 -9.09247816e-01 -5.79564750e-01 -3.87896001e-01
1.05376983e+00 -1.13749301e+00 8.91984522e-01 3.60788077e-01
-1.12792885e+00 -1.00701135e-02 -8.11330974e-01 -5.34816742e-01
-4.47015762e-01 -1.85373247e-01 1.37246943e+00 4.82139379e-01
-9.48596179e-01 3.02576780e-01 -1.02851403e+00 -2.86655098e-01
9.28025767e-02 3.24961126e-01 -4.52143066e-02 -5.41875541e-01
-1.25248885e+00 8.63296211e-01 1.77948669e-01 4.36008982e-02
-5.41674435e-01 -9.21002388e-01 -6.78456068e-01 3.28733295e-01
4.36721116e-01 -5.51786661e-01 1.24142826e+00 7.04246238e-02
-1.26333416e+00 3.95283312e-01 -5.32087147e-01 -5.95905900e-01
-1.87387928e-01 -4.29634869e-01 -3.08627009e-01 5.36198355e-02
-1.42574534e-01 3.12255532e-01 5.16317129e-01 -4.84047771e-01
-2.55082250e-01 -2.21734077e-01 1.31219730e-01 1.28869057e-01
-3.46701920e-01 3.24856520e-01 -1.24949360e+00 -6.90134823e-01
2.42779657e-01 -1.05661821e+00 -3.86995345e-01 -3.01099181e-01
-2.57086426e-01 -5.76213241e-01 6.26020372e-01 -5.54564595e-01
1.45016801e+00 -1.89241624e+00 3.23709436e-02 6.06608927e-01
2.33537495e-01 3.36388141e-01 -4.46862392e-02 7.42299557e-01
5.58180690e-01 -1.90225840e-02 2.10025892e-01 -1.92494944e-01
2.49044240e-01 2.94740498e-01 -6.72893465e-01 3.15260410e-01
-1.15580112e-01 1.15954709e+00 -9.09066677e-01 -4.86942708e-01
2.80018240e-01 7.18972385e-01 -8.50050211e-01 -1.29319787e-01
-3.46987814e-01 -1.76093951e-01 -7.47807622e-01 7.37134874e-01
3.61886829e-01 -5.09160697e-01 4.38114673e-01 -1.00772362e-02
1.64209619e-01 7.92668223e-01 -9.63534653e-01 1.60826254e+00
-8.58429551e-01 7.34447360e-01 9.91942883e-02 -8.62638712e-01
6.43237591e-01 1.74097456e-02 3.08273166e-01 -1.06893158e+00
-6.13653846e-02 3.22169930e-01 -2.70069659e-01 -2.01317817e-02
1.11285448e+00 -5.53377531e-03 -2.91491717e-01 6.59955502e-01
1.37189375e-02 -3.77370924e-01 3.88019800e-01 5.64697981e-01
1.16549110e+00 -2.87255555e-01 2.49518268e-02 -2.94808865e-01
2.06753641e-01 1.12156145e-01 3.32902730e-01 1.08034492e+00
5.97568452e-01 2.90584236e-01 2.08633885e-01 -4.44836855e-01
-8.33859563e-01 -7.65160859e-01 -3.33022565e-01 1.47567821e+00
-1.40794396e-01 -9.75851536e-01 -5.21753728e-02 -1.62800312e-01
2.01703757e-01 6.98178351e-01 8.28388482e-02 -2.00628102e-01
-6.94178581e-01 -9.13889408e-01 7.09989309e-01 6.35398269e-01
1.37521224e-02 -5.49551368e-01 -1.30777985e-01 1.58823147e-01
1.17924586e-01 -1.17197382e+00 -3.56672913e-01 2.08655372e-01
-8.78771663e-01 -5.71682215e-01 -7.14555919e-01 -4.47432637e-01
2.92100817e-01 3.80915105e-01 1.45316434e+00 3.38095188e-01
-6.13233030e-01 4.57326591e-01 -1.76786035e-01 -2.10778907e-01
-2.49247206e-03 4.38452899e-01 1.70367822e-01 -7.35756278e-01
5.49454629e-01 -3.32592994e-01 -6.74158037e-01 -2.13737711e-02
-9.34975684e-01 -3.71270955e-01 4.96239543e-01 1.00144613e+00
4.77205276e-01 -2.91461825e-01 1.39789164e-01 -9.16171253e-01
8.36343408e-01 -6.83768570e-01 -1.00606930e+00 2.27955967e-01
-1.01140356e+00 4.44658786e-01 2.27853000e-01 -2.45685115e-01
-7.41717935e-01 -4.93264586e-01 -4.00600970e-01 -5.51196277e-01
5.42147219e-01 9.85788643e-01 6.03184581e-01 -2.60979593e-01
4.83542770e-01 2.47636780e-01 -3.05291682e-01 -6.65346622e-01
5.27285218e-01 6.59614205e-01 8.34855810e-02 -8.26446652e-01
6.01130843e-01 2.70561785e-01 1.30092055e-01 -9.20813262e-01
-6.85468733e-01 -8.72891307e-01 -5.33225387e-03 4.62184787e-01
4.57154244e-01 -1.11583304e+00 -5.55890501e-01 -1.54424131e-01
-9.42775607e-01 -1.70553058e-01 3.58732678e-02 4.91917282e-01
-5.26197348e-03 4.82584000e-01 -8.69711995e-01 -7.73615599e-01
-6.06807888e-01 -1.27672017e+00 1.38775837e+00 1.32390866e-02
-2.48757288e-01 -1.13619840e+00 7.65593275e-02 3.80652457e-01
7.81252146e-01 -5.98881662e-01 1.19158113e+00 -7.56026328e-01
-8.53303790e-01 -6.25004947e-01 -4.49544072e-01 -3.89265940e-02
-3.43563855e-01 -2.79433072e-01 -6.42387271e-01 -5.29085159e-01
-4.39400941e-01 -3.96470815e-01 9.13107097e-01 1.04634501e-02
1.32361138e+00 -1.55491248e-01 -5.09985507e-01 5.96426129e-01
1.34948182e+00 -1.10154502e-01 5.18410623e-01 -9.54699237e-03
5.55519640e-01 -1.47964925e-01 5.08140266e-01 3.89239848e-01
1.50990605e-01 1.08216822e+00 -2.79418707e-01 1.82478458e-01
-1.59454316e-01 -1.99263781e-01 1.79733828e-01 1.22575164e+00
2.03031719e-01 -3.27427417e-01 -1.03919935e+00 5.94840288e-01
-1.78417695e+00 -7.65472770e-01 1.81290135e-01 2.31544662e+00
8.61359060e-01 6.73031341e-03 -3.76834214e-01 -3.37626219e-01
8.85734409e-02 3.90009344e-01 -5.00764251e-01 -6.38987720e-01
1.87905189e-02 9.22207534e-01 6.91357791e-01 7.05845892e-01
-5.59432149e-01 1.15341544e+00 7.75656176e+00 1.27500927e+00
-1.17782569e+00 1.84510365e-01 3.50575119e-01 -7.15862811e-01
-5.69106042e-01 7.28504211e-02 -1.15236115e+00 3.46242994e-01
1.37623394e+00 -2.55221218e-01 5.53657770e-01 8.08254778e-01
-1.83591589e-01 -1.65721580e-01 -1.03357017e+00 1.25642192e+00
3.49654467e-03 -1.82379174e+00 2.75175542e-01 1.90321356e-01
6.94192052e-01 5.20091593e-01 9.66688693e-02 7.84687757e-01
4.19526219e-01 -1.09435439e+00 9.21100602e-02 3.98051888e-01
6.31335557e-01 -6.82164431e-01 7.28567064e-01 1.61872730e-01
-1.03089511e+00 6.05794862e-02 -5.14119327e-01 1.58212315e-02
1.58328757e-01 6.29082382e-01 -7.60645390e-01 4.94310737e-01
5.99627674e-01 2.83570260e-01 -3.60471964e-01 7.93991089e-01
1.73681736e-01 8.56806278e-01 -7.80575156e-01 -3.89113873e-01
5.72586775e-01 -8.76659676e-02 4.10098672e-01 1.29884028e+00
3.78381193e-01 -8.26049671e-02 3.82053405e-01 5.26449859e-01
-3.87830138e-01 1.39543861e-01 -7.13963687e-01 -4.36075956e-01
6.85666800e-01 1.01729774e+00 -4.37560648e-01 -5.09435296e-01
-5.48438907e-01 7.50417709e-01 3.26232970e-01 5.12433052e-01
-6.68924809e-01 -5.65834403e-01 7.64517605e-01 8.70339945e-02
3.92399460e-01 -6.36348784e-01 -3.43180336e-02 -1.21343529e+00
-1.19416930e-01 -9.83150542e-01 3.51845860e-01 -5.53081214e-01
-1.17656994e+00 1.42475352e-01 2.29552761e-01 -5.30063152e-01
-7.24916279e-01 -5.77630758e-01 -8.31744149e-02 1.16493237e+00
-1.31994939e+00 -8.50772202e-01 1.98216841e-01 4.20226216e-01
4.56352085e-01 -3.26960623e-01 9.08845365e-01 5.61770558e-01
-2.44684458e-01 7.57419825e-01 5.79730213e-01 -9.02049616e-02
5.51900685e-01 -6.63578153e-01 5.61562121e-01 4.67295736e-01
6.86966777e-01 1.47177696e+00 6.18506849e-01 -5.67786634e-01
-2.39430594e+00 -7.04228640e-01 1.03603578e+00 -5.86128294e-01
9.82317328e-01 -5.86172283e-01 -6.97929084e-01 6.57234609e-01
1.41785830e-01 1.00019775e-01 6.39087141e-01 8.31839263e-01
-5.91811240e-01 2.17015352e-02 -7.29911685e-01 7.49356806e-01
8.43458891e-01 -1.10613048e+00 -3.76277298e-01 8.29847753e-01
8.83817911e-01 -5.06407678e-01 -9.49818552e-01 2.39771843e-01
6.32949412e-01 -3.77103388e-01 1.43402958e+00 -6.69693887e-01
1.22000296e-02 1.35741800e-01 -4.91528213e-01 -8.10335040e-01
-1.86121449e-01 -6.23955905e-01 -5.01106322e-01 7.88722873e-01
3.37097615e-01 -6.73374891e-01 8.23852658e-01 7.43295312e-01
3.18224341e-01 -9.77658987e-01 -9.96097088e-01 -6.36246204e-01
3.02815944e-01 -8.00896466e-01 5.90320408e-01 5.63722849e-01
-1.31225899e-01 4.43632752e-01 -2.99354702e-01 -1.61476597e-01
3.51986855e-01 3.35855156e-01 6.06380999e-01 -8.72848213e-01
-4.04798031e-01 -4.95498061e-01 -3.82714838e-01 -1.60258341e+00
3.54726255e-01 -1.12052143e+00 -2.63020813e-01 -1.44688451e+00
3.01285714e-01 -7.73335636e-01 -4.67527241e-01 3.55358154e-01
-4.97202910e-02 4.74245042e-01 6.00969605e-02 1.71005860e-01
-8.21733534e-01 4.77845460e-01 6.14431441e-01 -2.83948958e-01
6.98723793e-02 -1.58987567e-01 -6.70836151e-01 3.46788824e-01
3.56160402e-01 -2.79155672e-01 -5.03390968e-01 -9.08418715e-01
8.16581845e-01 1.50115892e-01 1.31745905e-01 -6.01395905e-01
4.86003846e-01 9.29530561e-02 1.64272919e-01 -8.70318711e-01
8.18549812e-01 -4.46731359e-01 -2.77389169e-01 2.01985717e-01
-4.17565376e-01 3.93190503e-01 4.78470445e-01 7.33063161e-01
-2.70087898e-01 -3.19788724e-01 -1.17970747e-03 -9.90851372e-02
-3.97209734e-01 3.38740975e-01 -5.01827538e-01 2.08189338e-01
3.07047844e-01 4.42273527e-01 -1.87900826e-01 -5.31816721e-01
-1.58090368e-01 2.94237226e-01 8.36902782e-02 3.87279332e-01
4.97954935e-01 -1.28707314e+00 -4.28744018e-01 3.79050933e-02
2.60060094e-02 -1.79699212e-01 6.37018606e-02 6.00317359e-01
-5.30438364e-01 1.34204757e+00 5.05268097e-01 -6.88796699e-01
-1.21194220e+00 3.17109525e-01 -2.75804400e-01 -6.14643633e-01
-5.52563250e-01 9.74178016e-01 -2.18329698e-01 -4.62410599e-01
2.16407210e-01 -2.60678411e-01 3.07296902e-01 -1.40029013e-01
5.62984943e-01 2.80144721e-01 4.56367046e-01 -2.14713365e-01
-4.04870570e-01 3.45739931e-01 -5.28354883e-01 -2.16998443e-01
1.22966230e+00 1.81869924e-01 -2.37985328e-01 3.87221217e-01
1.43426847e+00 2.18819454e-01 -2.16133565e-01 -5.65943897e-01
2.54632324e-01 -5.10300457e-01 6.57885730e-01 -4.82984692e-01
-8.28728855e-01 8.37017238e-01 3.05455029e-01 1.08178340e-01
6.42261446e-01 1.35417596e-01 1.17947698e+00 1.12344539e+00
7.19105482e-01 -8.56271267e-01 -4.51243937e-01 8.34958255e-01
5.44697344e-01 -1.18671465e+00 6.57361627e-01 -8.38575885e-02
-1.52194470e-01 8.52383137e-01 5.71639389e-02 -1.56206667e-01
7.42726684e-01 2.47471571e-01 -3.42367411e-01 -4.75586146e-01
-1.19588053e+00 5.11014648e-02 5.83180606e-01 -5.67266764e-03
6.44042313e-01 -1.15017600e-01 -3.61636996e-01 1.51606083e-01
-1.67391002e-01 -2.61260629e-01 -1.26215458e-01 1.09016621e+00
-2.04334810e-01 -1.59717894e+00 -1.98705047e-01 9.48749065e-01
-6.24545455e-01 -7.31608033e-01 -8.61798320e-03 6.14762604e-01
-7.80133724e-01 7.25896716e-01 2.06439301e-01 -2.12510034e-01
-1.27557576e-01 2.96511203e-01 5.46710610e-01 -7.71588027e-01
-6.66138470e-01 -1.58557333e-02 3.60923082e-01 -9.85990465e-01
4.82624620e-02 -5.51502168e-01 -1.09457648e+00 -5.32473862e-01
-3.71926576e-01 4.67241585e-01 8.61595273e-01 1.00749123e+00
7.86847711e-01 3.35523695e-01 1.29315063e-01 -6.17351115e-01
-9.12182152e-01 -8.61582518e-01 -2.43119881e-01 -4.81231548e-02
7.36029968e-02 -5.94141245e-01 -2.40603536e-01 -6.62255764e-01] | [11.470147132873535, 7.578574180603027] |
8c452a2a-3410-4935-8be0-8e226e02edd8 | dream3d-zero-shot-text-to-3d-synthesis-using | 2212.14704 | null | https://arxiv.org/abs/2212.14704v2 | https://arxiv.org/pdf/2212.14704v2.pdf | Dream3D: Zero-Shot Text-to-3D Synthesis Using 3D Shape Prior and Text-to-Image Diffusion Models | Recent CLIP-guided 3D optimization methods, such as DreamFields and PureCLIPNeRF, have achieved impressive results in zero-shot text-to-3D synthesis. However, due to scratch training and random initialization without prior knowledge, these methods often fail to generate accurate and faithful 3D structures that conform to the input text. In this paper, we make the first attempt to introduce explicit 3D shape priors into the CLIP-guided 3D optimization process. Specifically, we first generate a high-quality 3D shape from the input text in the text-to-shape stage as a 3D shape prior. We then use it as the initialization of a neural radiance field and optimize it with the full prompt. To address the challenging text-to-shape generation task, we present a simple yet effective approach that directly bridges the text and image modalities with a powerful text-to-image diffusion model. To narrow the style domain gap between the images synthesized by the text-to-image diffusion model and shape renderings used to train the image-to-shape generator, we further propose to jointly optimize a learnable text prompt and fine-tune the text-to-image diffusion model for rendering-style image generation. Our method, Dream3D, is capable of generating imaginative 3D content with superior visual quality and shape accuracy compared to state-of-the-art methods. | ['Shenghua Gao', 'XiaoHu Qie', 'Ying Shan', 'Yan-Pei Cao', 'Weihao Cheng', 'Xintao Wang', 'Jiale Xu'] | 2022-12-28 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Xu_Dream3D_Zero-Shot_Text-to-3D_Synthesis_Using_3D_Shape_Prior_and_Text-to-Image_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Xu_Dream3D_Zero-Shot_Text-to-3D_Synthesis_Using_3D_Shape_Prior_and_Text-to-Image_CVPR_2023_paper.pdf | cvpr-2023-1 | ['text-to-shape-generation', 'text-to-3d'] | ['computer-vision', 'computer-vision'] | [ 5.43132544e-01 1.58796072e-01 4.39486533e-01 -1.52139932e-01
-7.55856574e-01 -7.00701237e-01 9.13163066e-01 -4.60630447e-01
1.75424799e-01 4.66524303e-01 1.37658924e-01 -2.88546383e-01
1.22565448e-01 -9.86004591e-01 -8.44562352e-01 -5.97318530e-01
5.49478114e-01 7.05037594e-01 -2.23375279e-02 -3.91225070e-01
2.68125236e-01 6.78529739e-01 -1.50460982e+00 2.48680428e-01
1.19557357e+00 1.04880607e+00 3.84587437e-01 8.76073599e-01
-4.09835011e-01 6.35629535e-01 -6.29642069e-01 -3.33032906e-01
4.26632315e-01 -8.40526044e-01 -3.54651719e-01 4.58675414e-01
4.53844309e-01 -4.26256150e-01 -2.14312047e-01 7.61248767e-01
7.81986475e-01 1.71488225e-01 9.70300198e-01 -8.96366060e-01
-9.74272609e-01 -1.40397903e-02 -7.59742796e-01 -4.67666298e-01
4.66344237e-01 3.83631080e-01 5.62259257e-01 -1.00914788e+00
1.00554705e+00 1.33021092e+00 4.03274924e-01 6.64143920e-01
-1.55749929e+00 -3.53318900e-01 -2.10847825e-01 -3.55206549e-01
-1.31128824e+00 -4.17076439e-01 1.09719348e+00 -5.48581719e-01
7.70234048e-01 4.11215186e-01 7.71431267e-01 1.11044323e+00
1.69954687e-01 6.61196530e-01 1.16418564e+00 -6.70392632e-01
2.12495744e-01 1.05774529e-01 -8.38659167e-01 7.26589680e-01
-3.51900488e-01 3.33271146e-01 -5.12667060e-01 1.66467819e-02
1.32941258e+00 -4.22763467e-01 -3.01839113e-01 -3.08868915e-01
-1.16169012e+00 6.67186737e-01 4.26142275e-01 7.00259283e-02
-4.99197572e-01 2.95494169e-01 -2.08835974e-01 -8.49793702e-02
9.06387866e-01 6.77608371e-01 5.19079119e-02 4.20176350e-02
-1.11049581e+00 5.78453958e-01 6.88599050e-01 1.02527058e+00
6.20343089e-01 4.79039609e-01 -6.30110085e-01 9.84525442e-01
2.96005934e-01 9.34572577e-01 7.93906152e-02 -1.21467340e+00
2.41450727e-01 4.10586178e-01 1.84927106e-01 -7.46788740e-01
3.51454131e-02 -2.63396472e-01 -8.30164909e-01 6.78746581e-01
1.56834230e-01 -1.43401504e-01 -1.25874662e+00 1.40239942e+00
5.37606001e-01 -7.31604407e-03 -6.92873672e-02 1.00639951e+00
9.96803880e-01 9.71034348e-01 -3.66737485e-01 2.43018307e-02
1.08267510e+00 -9.53036666e-01 -6.47435784e-01 -5.32796420e-02
2.70673305e-01 -1.09974277e+00 1.35049736e+00 1.17518656e-01
-1.68827486e+00 -4.90948498e-01 -9.51655269e-01 -2.79542297e-01
4.51939330e-02 4.68299575e-02 2.47543573e-01 5.03016353e-01
-1.14745390e+00 4.46035266e-01 -3.60074043e-01 2.59298994e-03
5.67405760e-01 -1.16367184e-01 -3.46227176e-02 -6.66569695e-02
-7.94330120e-01 7.79959857e-01 4.16918918e-02 -3.56184691e-01
-8.54400933e-01 -1.04101074e+00 -6.92594826e-01 -8.66272524e-02
2.98351467e-01 -1.23375809e+00 1.19954479e+00 -9.51980352e-01
-2.14314842e+00 1.17436004e+00 -3.61742005e-02 -8.23919252e-02
5.02858102e-01 1.57420844e-01 1.76090375e-01 1.75540686e-01
3.89562338e-03 1.03041327e+00 1.23804200e+00 -1.71107101e+00
-1.03942469e-01 1.16820261e-01 6.21789647e-03 5.32358944e-01
7.71309584e-02 -2.61814296e-01 -7.47100413e-01 -1.06562579e+00
9.97120701e-03 -6.62544072e-01 -2.30698332e-01 5.04674196e-01
-5.66619098e-01 2.35197395e-01 8.89214814e-01 -3.48176956e-01
6.79309905e-01 -1.94271910e+00 3.45997483e-01 9.28716958e-02
1.75803050e-01 1.96148068e-01 -4.09901053e-01 4.09496039e-01
1.78992469e-02 2.03625366e-01 -4.04847682e-01 -7.49601543e-01
9.93301421e-02 5.55891916e-02 -4.87499326e-01 1.05635568e-01
3.43450904e-01 1.22745895e+00 -7.35797167e-01 -3.82999778e-01
4.88247097e-01 9.49618697e-01 -8.76752317e-01 6.05203569e-01
-7.52888680e-01 8.75470817e-01 -5.23383737e-01 5.42074621e-01
7.83026516e-01 -2.19944194e-01 -2.72070765e-01 -3.99725467e-01
-2.29835987e-01 -8.90802220e-02 -7.44025290e-01 1.96521056e+00
-7.10905492e-01 5.58692396e-01 1.97204679e-01 -6.19107246e-01
1.08426809e+00 1.35239854e-01 5.64345419e-01 -9.94543970e-01
1.95700139e-01 2.38333270e-01 -4.48035717e-01 -3.45298767e-01
6.83469534e-01 -5.06821275e-01 4.07692641e-02 5.64568698e-01
2.70369779e-02 -1.34680569e+00 -9.63557884e-02 2.02760890e-01
6.25427723e-01 6.11021221e-01 -2.46517986e-01 -1.33406967e-01
4.11202043e-01 -7.25927651e-02 -5.58455959e-02 5.78301251e-01
6.28269434e-01 1.44738722e+00 3.83577585e-01 -2.55716473e-01
-1.63429892e+00 -1.16161239e+00 4.08988781e-02 5.43022215e-01
9.57913101e-02 -2.93741345e-01 -1.06050372e+00 -4.92055774e-01
-2.66579688e-01 1.18203223e+00 -5.33004344e-01 8.30429420e-03
-4.44673121e-01 -5.04147172e-01 4.80219603e-01 8.97981301e-02
5.62783778e-01 -1.01731563e+00 -3.46126169e-01 2.85482138e-01
-1.76856458e-01 -1.12183046e+00 -9.27015841e-01 -1.46739542e-01
-7.30502963e-01 -5.57634711e-01 -1.24904692e+00 -5.24866998e-01
9.84742284e-01 1.36698976e-01 1.32600772e+00 1.70258909e-01
-5.37065268e-01 4.61895376e-01 -2.29152918e-01 -4.10958290e-01
-8.56693983e-01 -4.08997178e-01 -4.10836756e-01 9.52699631e-02
-6.68877900e-01 -6.90458596e-01 -7.88364649e-01 3.30709308e-01
-1.20896721e+00 7.57743537e-01 3.53686661e-01 6.92596555e-01
9.12269533e-01 -1.62198558e-01 1.27035215e-01 -4.71268594e-01
6.74472392e-01 4.17377017e-02 -6.73649788e-01 3.13100740e-02
-2.87311852e-01 2.75741160e-01 6.57033324e-01 -5.70251584e-01
-1.31714630e+00 6.09834194e-02 -5.92783689e-01 -8.10582936e-01
-2.71022245e-02 2.66212989e-02 -1.67812586e-01 -3.27511460e-01
8.85081768e-01 4.73079771e-01 1.81658398e-02 -4.00521517e-01
7.94003665e-01 3.33587050e-01 5.14042914e-01 -8.54389846e-01
1.23451030e+00 5.61763465e-01 1.56840861e-01 -9.64611650e-01
-8.96924615e-01 2.38657460e-01 -4.69031423e-01 -3.41137856e-01
1.06940460e+00 -6.93297565e-01 -4.00986582e-01 6.95512116e-01
-1.40026772e+00 -9.98284698e-01 -8.04664791e-01 -4.66772206e-02
-1.10147667e+00 2.15861812e-01 -3.63307446e-01 -6.98852658e-01
-5.59229970e-01 -1.23138678e+00 1.65150011e+00 1.16966166e-01
4.03860398e-03 -1.06431460e+00 4.70859818e-02 2.49719203e-01
7.31835306e-01 5.82242966e-01 1.06974471e+00 4.09341335e-01
-8.86686325e-01 1.03256613e-01 -3.66244346e-01 2.63647437e-01
6.77970126e-02 -3.11521590e-02 -9.13208842e-01 5.41260056e-02
2.79197991e-02 -4.06933099e-01 4.93504196e-01 6.63835526e-01
1.14877510e+00 -2.85161167e-01 -7.58685693e-02 1.18652236e+00
1.42115736e+00 1.99041492e-03 8.65207136e-01 -1.19544022e-01
8.60622823e-01 4.18034971e-01 2.26404712e-01 5.28762400e-01
2.50681192e-01 1.03069103e+00 3.24798733e-01 -4.98785228e-01
-8.79714310e-01 -6.59791529e-01 1.21141836e-01 5.65626621e-01
-1.25842810e-01 -6.37082696e-01 -3.80659938e-01 1.32876441e-01
-1.36497021e+00 -8.89579892e-01 -1.49061888e-01 2.13033938e+00
9.85450804e-01 -7.59418979e-02 -2.22873434e-01 -1.05434038e-01
4.03727084e-01 2.74250656e-01 -5.62331915e-01 -4.91030663e-01
-3.42319936e-01 3.70545328e-01 1.79464892e-01 7.70529568e-01
-5.03656924e-01 1.23654556e+00 6.11178589e+00 1.26065493e+00
-1.32662523e+00 -3.87859344e-03 9.13061380e-01 -1.58221111e-01
-1.06059039e+00 -8.54521170e-02 -5.24916768e-01 1.98079705e-01
5.40699244e-01 -1.68388113e-01 7.52448857e-01 3.70501012e-01
5.83589673e-01 -7.99811780e-02 -9.11561966e-01 1.20535123e+00
1.94515288e-01 -1.73769522e+00 4.34124470e-01 4.70229052e-02
1.21188581e+00 -4.30556118e-01 2.84032017e-01 -1.42791480e-01
1.89666450e-01 -1.25423551e+00 1.18758261e+00 8.50805640e-01
1.49200344e+00 -5.46685219e-01 -2.00938135e-01 4.19246554e-01
-9.69481230e-01 4.52407449e-01 -1.97156653e-01 4.88014221e-01
6.17288530e-01 8.37350309e-01 -6.19656384e-01 4.48908657e-01
2.64296651e-01 5.08340657e-01 -8.83516297e-02 8.72160316e-01
-2.99342185e-01 2.00427517e-01 -2.70138353e-01 1.39557302e-01
2.03211889e-01 -3.94875616e-01 8.45458031e-01 8.84034276e-01
7.95447469e-01 4.05760854e-01 3.43996808e-02 1.84021556e+00
-3.15322340e-01 1.42140791e-01 -7.98503280e-01 -8.31553489e-02
1.14025064e-01 1.17610800e+00 -6.31207347e-01 -3.12214643e-01
-8.75841230e-02 1.52155697e+00 9.86003354e-02 5.77445269e-01
-8.24769139e-01 -3.27762306e-01 3.24253917e-01 3.94710749e-01
2.62350887e-01 -4.22874779e-01 -5.43137133e-01 -1.00731993e+00
-1.37690455e-01 -7.41361737e-01 -4.43864822e-01 -1.51535928e+00
-1.23813105e+00 7.52590895e-01 -8.74931514e-02 -1.12601709e+00
-2.61101156e-01 -3.73231411e-01 -7.81641066e-01 1.29373145e+00
-1.50564098e+00 -1.41238248e+00 -3.37177932e-01 6.32368207e-01
7.25244462e-01 -2.20040157e-02 6.83380425e-01 -4.67326045e-02
9.20718536e-02 3.49268347e-01 -1.89720839e-03 -3.85266721e-01
6.03965521e-01 -1.04599154e+00 8.25498283e-01 5.38171470e-01
-7.52731934e-02 7.02443570e-02 6.16571903e-01 -7.89030135e-01
-1.71777964e+00 -1.02290869e+00 4.55720603e-01 -6.18170857e-01
1.47475049e-01 -4.90017772e-01 -6.72012746e-01 3.84585410e-02
4.23760951e-01 -1.87168568e-01 -3.15783313e-04 -6.49853885e-01
3.05005728e-04 2.01019660e-01 -1.31800759e+00 9.20013428e-01
1.18335259e+00 -4.81387734e-01 -2.74852425e-01 3.33193004e-01
8.13406587e-01 -8.84252667e-01 -7.25211203e-01 1.37347803e-01
2.41154596e-01 -1.04170954e+00 1.22282588e+00 4.92482334e-02
7.68886566e-01 -3.24299157e-01 -7.91460574e-02 -1.56059813e+00
-1.71725273e-01 -1.19956410e+00 1.40236411e-02 1.07803321e+00
3.95576626e-01 -2.02441052e-01 7.95088887e-01 4.38863665e-01
-4.00323093e-01 -6.56205833e-01 -6.77621424e-01 -5.01725435e-01
1.80598199e-01 -5.23754060e-01 6.91252053e-01 6.78352296e-01
-5.39596796e-01 3.28698993e-01 -4.90024179e-01 -1.90944463e-01
8.47096920e-01 3.14274758e-01 9.42304730e-01 -7.80387282e-01
-3.77286315e-01 -5.60027957e-01 2.10067704e-01 -1.62266874e+00
-1.87428147e-01 -9.65741694e-01 2.57752746e-01 -1.74667728e+00
-6.56507835e-02 -6.38493717e-01 7.03577340e-01 1.03221782e-01
-1.31472334e-01 4.86235291e-01 3.97883981e-01 5.01568131e-02
-1.66901853e-02 1.04604542e+00 2.17432189e+00 -7.03857541e-02
-5.56207001e-01 -2.64616698e-01 -6.24831319e-01 4.01212305e-01
4.39346462e-01 -2.29453191e-01 -6.21337533e-01 -6.57446563e-01
2.30221465e-01 2.94026047e-01 4.49226648e-01 -7.98315167e-01
-4.47650552e-02 -1.85910016e-01 5.16380370e-01 -7.12536633e-01
7.33274877e-01 -5.64369798e-01 1.69208765e-01 -2.42988933e-02
-2.70145655e-01 -3.50752503e-01 2.88071960e-01 2.46582806e-01
1.75772071e-01 -1.44308418e-01 1.00315559e+00 -2.49919444e-01
-1.00116268e-01 5.92050314e-01 -2.24890858e-01 3.79767835e-01
7.22329438e-01 -3.19441885e-01 -1.38069525e-01 -7.10906029e-01
-5.21821439e-01 -1.76953033e-01 8.36225152e-01 3.77903968e-01
9.13932860e-01 -1.65251577e+00 -8.90024722e-01 5.58651388e-01
-1.28306389e-01 2.00856924e-01 3.47756892e-01 4.98048961e-01
-6.39998317e-01 1.75569758e-01 -7.70207727e-03 -8.09461594e-01
-7.74925590e-01 3.98791343e-01 6.18245244e-01 2.20075175e-02
-9.17816579e-01 7.51588523e-01 6.68523610e-01 -6.06785715e-01
-7.36911735e-03 -2.02272728e-01 3.42982143e-01 -4.16807085e-01
4.39005107e-01 -1.32735729e-01 -1.08019032e-01 -5.54005444e-01
2.22972289e-01 1.16723168e+00 4.24568146e-01 -5.92552364e-01
1.34699976e+00 -1.81899667e-02 2.12067351e-01 -6.75802538e-03
1.10051441e+00 1.49305835e-01 -1.76972950e+00 -4.07825001e-02
-8.14852118e-01 -7.32089460e-01 3.89673173e-01 -9.21181858e-01
-1.14765775e+00 9.64446902e-01 3.47889662e-01 1.04418416e-02
1.03507948e+00 -4.67445068e-02 1.00412524e+00 -1.21588528e-01
2.18240678e-01 -9.01805341e-01 5.02477229e-01 4.85240161e-01
1.40376890e+00 -8.42545509e-01 -9.44415852e-02 -4.54961866e-01
-6.19380116e-01 1.11199975e+00 3.72274339e-01 -5.23117743e-02
4.40628648e-01 5.12760818e-01 4.21362519e-02 -2.33762324e-01
-6.19185209e-01 -7.03507438e-02 7.70889819e-01 8.16449821e-01
3.08930248e-01 -2.47944117e-01 1.11889794e-01 8.43570009e-02
-2.26947814e-01 1.23557687e-01 4.45802212e-01 5.45741022e-01
-2.50415176e-01 -1.20021307e+00 -5.67889035e-01 8.80131051e-02
-5.50710708e-02 -3.36873204e-01 -3.37225586e-01 3.26618075e-01
-1.96033828e-02 6.27584398e-01 5.63156158e-02 -2.20728278e-01
5.82985997e-01 -1.58550501e-01 9.14946258e-01 -4.73375767e-01
-3.84513378e-01 4.83124316e-01 3.71432416e-02 -5.72273493e-01
-1.59927741e-01 -2.11082458e-01 -1.15309215e+00 -3.52227271e-01
-4.60816734e-02 -2.64869303e-01 8.52245092e-01 6.52198732e-01
5.24280369e-01 6.11595094e-01 7.93122172e-01 -1.70602381e+00
-5.72956391e-02 -5.21784186e-01 -4.57576364e-01 4.37958449e-01
9.42883193e-02 -4.61908251e-01 -2.60830969e-01 1.55774623e-01] | [9.283080101013184, -3.3041858673095703] |
2d9146be-b3f7-4f2e-97aa-a442048c404a | learning-low-resource-end-to-end-goal | null | null | https://aclanthology.org/2020.acl-main.57 | https://aclanthology.org/2020.acl-main.57.pdf | Learning Low-Resource End-To-End Goal-Oriented Dialog for Fast and Reliable System Deployment | Existing end-to-end dialog systems perform less effectively when data is scarce. To obtain an acceptable success in real-life online services with only a handful of training examples, both fast adaptability and reliable performance are highly desirable for dialog systems. In this paper, we propose the Meta-Dialog System (MDS), which combines the advantages of both meta-learning approaches and human-machine collaboration. We evaluate our methods on a new extended-bAbI dataset and a transformed MultiWOZ dataset for low-resource goal-oriented dialog learning. Experimental results show that MDS significantly outperforms non-meta-learning baselines and can achieve more than 90{\%} per-turn accuracies with only 10 dialogs on the extended-bAbI dataset. | ['Xiaodan Zhu', 'Hangyu Li', 'Yinpei Dai', 'Chengguang Tang', 'Jian Sun', 'Yongbin Li'] | 2020-07-01 | null | null | null | acl-2020-6 | ['goal-oriented-dialog', 'dialog-learning'] | ['natural-language-processing', 'natural-language-processing'] | [-3.62114131e-01 2.77480811e-01 -2.86612183e-01 -9.18441892e-01
-1.13470995e+00 -5.26305377e-01 8.92120838e-01 -2.68190771e-01
-5.43998897e-01 1.08399153e+00 3.46181422e-01 -4.11291927e-01
1.59569502e-01 -2.94934809e-01 8.88645090e-03 -2.54261315e-01
2.83918202e-01 1.10441244e+00 4.46759164e-01 -9.78083849e-01
-3.54950614e-02 -1.41473070e-01 -8.78101170e-01 6.20386541e-01
9.15539682e-01 7.23195255e-01 2.77989835e-01 6.31913245e-01
-4.39442784e-01 9.09031212e-01 -8.48549366e-01 -8.12505424e-01
-8.17049108e-03 -2.52897143e-01 -1.16593313e+00 -5.32775931e-02
2.88771719e-01 -8.40721011e-01 -3.71036083e-01 5.39797544e-01
6.50358081e-01 6.00608647e-01 4.36363429e-01 -1.28515697e+00
-3.33540976e-01 8.92578125e-01 8.44292492e-02 -1.60312504e-01
5.37878036e-01 3.94193798e-01 1.22063160e+00 -1.04987109e+00
5.67360103e-01 1.61984539e+00 3.14099818e-01 1.09000361e+00
-1.09741747e+00 -5.41940987e-01 2.38830060e-01 6.41345903e-02
-6.87319279e-01 -8.89819086e-01 5.81351876e-01 -6.72516078e-02
1.13838291e+00 7.13705420e-02 -1.20658785e-01 1.42954373e+00
-1.45765871e-01 1.04404545e+00 1.08881485e+00 -2.90125400e-01
8.53648335e-02 4.93580610e-01 3.57587576e-01 4.63042915e-01
-7.34139085e-01 -6.98034912e-02 -7.50617802e-01 -1.43793672e-01
2.83443630e-01 -1.81147531e-01 1.25607640e-01 9.09163952e-02
-1.37330663e+00 9.69304740e-01 1.39950305e-01 2.04372451e-01
5.40733822e-02 -4.97628659e-01 4.26380396e-01 6.05970740e-01
3.83766741e-01 3.00453335e-01 -6.76721513e-01 -6.91452742e-01
-1.62246406e-01 4.97452348e-01 1.34769893e+00 1.47228324e+00
3.34303647e-01 -5.29632866e-02 -1.44703791e-01 1.48007119e+00
1.53063729e-01 4.48631793e-01 6.10282004e-01 -1.19881392e+00
1.03386927e+00 7.99006760e-01 4.33381587e-01 -2.12588742e-01
-5.23715973e-01 4.21028972e-01 -6.90722167e-01 1.62326507e-02
9.04357135e-01 -3.40396225e-01 -2.24361420e-01 1.82733774e+00
3.50400984e-01 -4.41402435e-01 4.34809893e-01 6.38212383e-01
1.27581787e+00 6.67514741e-01 2.75691897e-01 -2.04510406e-01
1.25331903e+00 -1.51288843e+00 -8.45456362e-01 -3.79780203e-01
7.63101459e-01 -7.82423139e-01 1.59610963e+00 2.06759498e-01
-1.23569715e+00 -6.64511919e-01 -7.46976733e-01 -1.60720468e-01
-4.39105481e-01 -7.03546405e-02 5.69504142e-01 6.39145911e-01
-9.13895667e-01 2.51659781e-01 -3.70389074e-01 -3.76183271e-01
-3.01021814e-01 5.29445171e-01 -3.53653580e-01 2.07204387e-01
-1.24509227e+00 1.11156809e+00 3.49981457e-01 -2.73489267e-01
-6.49378240e-01 -6.31645262e-01 -5.56891739e-01 -1.96398199e-02
6.88400745e-01 -2.27735490e-01 2.10626674e+00 -2.35841453e-01
-2.22014260e+00 7.44849682e-01 -1.20538786e-01 -3.56944799e-01
7.27254748e-01 -2.79445678e-01 -4.10293788e-01 -2.56814331e-01
-3.74874383e-01 9.04749334e-01 3.16198111e-01 -1.17481530e+00
-1.02382970e+00 -1.97752848e-01 5.08362591e-01 3.77839059e-01
-5.36429763e-01 -9.04618576e-02 -3.15590799e-01 -1.95367783e-01
-1.42357230e-01 -1.11680949e+00 -7.01819956e-02 -4.71845537e-01
-3.01569492e-01 -7.63908863e-01 1.03411603e+00 -5.47177136e-01
9.30361688e-01 -1.87130737e+00 1.58638313e-01 -5.84292293e-01
1.71341613e-01 4.68195170e-01 -2.77045369e-01 7.48434365e-01
4.49314415e-01 -1.62042975e-01 2.91453749e-01 -7.42392600e-01
2.00770304e-01 1.56503424e-01 -3.31558734e-01 -1.97102949e-01
-2.49725342e-01 8.21987569e-01 -9.76549447e-01 -6.11563861e-01
6.14263356e-01 -1.53820738e-01 -5.44008076e-01 1.02430117e+00
-5.25938749e-01 7.76869953e-01 -1.72230259e-01 6.38621271e-01
1.52546152e-01 -1.29493117e-01 4.74868685e-01 -4.72971089e-02
5.72445765e-02 5.37046909e-01 -6.70764565e-01 1.91560507e+00
-8.53628874e-01 3.44457060e-01 3.66831183e-01 -5.09400666e-01
1.22723138e+00 5.24219751e-01 3.51348966e-01 -6.88729465e-01
1.96215287e-01 1.58363596e-01 1.75545692e-01 -2.68227488e-01
7.11897612e-01 -4.47069183e-02 -4.52678472e-01 5.78469515e-01
5.11101902e-01 -1.39370322e-01 1.92460701e-01 3.80724519e-01
7.51700580e-01 -2.70009905e-01 2.29914770e-01 -1.05660588e-01
6.84079051e-01 7.81400595e-03 4.06836569e-01 7.73519337e-01
-6.84691668e-01 1.54334068e-01 2.01434389e-01 -3.71059567e-01
-7.33495533e-01 -1.04370022e+00 2.14869022e-01 1.90570545e+00
8.42018351e-02 -3.54360580e-01 -7.13445187e-01 -9.75687206e-01
-1.95613086e-01 1.00047851e+00 -3.51371579e-02 -1.05454661e-01
-6.57454014e-01 -5.20724833e-01 4.00778353e-01 5.28933585e-01
1.03386915e+00 -1.13341045e+00 -8.68301913e-02 4.57355797e-01
-5.60982406e-01 -1.44941449e+00 -4.83614057e-01 -1.11462854e-01
-8.04648161e-01 -6.92278206e-01 -4.68782216e-01 -7.67457187e-01
-9.74511206e-02 5.00491202e-01 1.28737831e+00 -7.67256245e-02
2.56476581e-01 2.63654292e-01 -5.78897476e-01 -2.38125205e-01
-9.74602938e-01 5.55829823e-01 1.99083760e-01 -5.01079857e-01
5.87362587e-01 -5.43503642e-01 -2.68215388e-01 8.51891994e-01
-1.37774184e-01 3.01950216e-01 3.36399078e-01 1.27741420e+00
-3.28958511e-01 -5.78214169e-01 1.25333750e+00 -8.36943805e-01
9.56447184e-01 -2.70635217e-01 -4.73109901e-01 5.80983937e-01
-4.44840223e-01 -2.20527053e-01 5.20844519e-01 -6.22033894e-01
-1.73309994e+00 -1.05108604e-01 -3.43037009e-01 7.00438991e-02
-3.03590536e-01 1.90592125e-01 -5.81695735e-01 2.54150331e-01
5.00830710e-01 4.46769260e-02 1.40855417e-01 -6.75199151e-01
8.47068369e-01 1.38097942e+00 7.00216770e-01 -8.77566159e-01
2.93811888e-01 -2.72524394e-02 -6.26722395e-01 -8.91138852e-01
-8.06604505e-01 -5.21826386e-01 -8.56360793e-01 -4.79412079e-01
7.02561080e-01 -8.67057085e-01 -1.29456806e+00 4.44674402e-01
-1.21447432e+00 -9.47586477e-01 1.70413688e-01 3.34684521e-01
-7.53368855e-01 2.56461084e-01 -7.83210635e-01 -1.15044820e+00
-5.96806467e-01 -1.31574821e+00 9.17162418e-01 2.21599534e-01
-4.63044763e-01 -1.14721882e+00 5.98475933e-02 9.05671954e-01
4.40933228e-01 -5.41203976e-01 9.84792709e-01 -1.32078969e+00
-3.05499732e-01 -8.60641599e-02 -7.38386214e-02 2.72724360e-01
3.34973633e-01 -4.05219615e-01 -9.95882452e-01 -4.38019812e-01
-1.47562608e-01 -1.11640751e+00 9.94482264e-02 -2.76369244e-01
6.04704201e-01 -4.04821634e-01 -6.74830973e-02 -2.86778584e-02
7.73468614e-01 6.61714852e-01 6.89195171e-02 4.05652970e-02
3.48391980e-01 1.02759123e+00 1.03453052e+00 5.83739936e-01
8.34316969e-01 1.04378331e+00 2.24790320e-01 3.16753536e-01
-8.80870968e-02 -2.87737697e-01 2.96347857e-01 1.06131864e+00
2.09854558e-01 -4.08203423e-01 -9.43195820e-01 4.26968217e-01
-2.17467690e+00 -8.15662861e-01 2.95062691e-01 2.06309676e+00
1.14440775e+00 2.79229015e-01 6.24160647e-01 -4.50929940e-01
5.15368283e-01 2.35486940e-01 -5.70007145e-01 -2.68396825e-01
1.08021200e-01 -3.68148893e-01 -1.33062333e-01 8.71374905e-01
-1.03200352e+00 1.61512780e+00 6.47949648e+00 6.93935871e-01
-5.96810162e-01 4.73632902e-01 6.36634171e-01 -1.06088623e-01
3.05405587e-01 -1.70190424e-01 -1.28559196e+00 4.41014230e-01
1.18273950e+00 -9.83817503e-02 5.41188180e-01 1.23212695e+00
4.31882441e-02 6.16901368e-02 -1.27618039e+00 9.06004667e-01
-1.96010321e-01 -1.09546459e+00 -1.52065381e-01 -5.29396646e-02
3.71917337e-01 -1.25054628e-01 -1.51418634e-02 1.10608137e+00
1.01301515e+00 -7.38041341e-01 3.31375241e-01 1.06471941e-01
5.41407287e-01 -5.78209817e-01 6.97662950e-01 8.48001778e-01
-8.48634720e-01 -1.63343906e-01 -3.82697970e-01 1.38131455e-01
5.38170755e-01 -2.79501081e-01 -1.31050205e+00 3.07007343e-01
5.08581102e-01 3.13522398e-01 -2.20332339e-01 3.12866032e-01
1.40932873e-01 3.71952623e-01 -1.26810834e-01 -4.41277504e-01
2.81921595e-01 -2.91313231e-01 4.32155311e-01 1.15452683e+00
-2.05228716e-01 4.01353389e-01 5.01091599e-01 4.62650895e-01
-2.02984720e-01 1.49792716e-01 -5.66384673e-01 9.14784446e-02
8.38352025e-01 1.38306046e+00 -3.56713124e-02 -4.52711493e-01
-7.59290397e-01 8.50288272e-01 6.28663599e-01 1.61359712e-01
-5.46823382e-01 -1.38378277e-01 7.44713962e-01 -5.02715230e-01
-3.61552984e-01 -4.16547865e-01 -2.76430845e-01 -1.16903841e+00
-3.25914532e-01 -1.25729167e+00 5.00143170e-01 -5.74127674e-01
-1.53722799e+00 8.93735588e-01 2.55240381e-01 -1.07030809e+00
-1.06866658e+00 -7.82446682e-01 -6.58477724e-01 7.48204291e-01
-1.02526510e+00 -1.64030933e+00 -4.19164896e-01 6.82589054e-01
1.34739208e+00 -8.12605083e-01 1.13924849e+00 8.55441615e-02
-5.07156551e-01 9.18029308e-01 1.78052172e-01 2.43010774e-01
1.29995477e+00 -1.48351610e+00 5.31003535e-01 2.01891307e-02
-1.00935414e-01 6.47196293e-01 6.83950126e-01 -5.63750744e-01
-1.37112880e+00 -6.68089509e-01 8.80759478e-01 -7.86579728e-01
7.57851779e-01 -6.06438160e-01 -8.51981699e-01 8.58842731e-01
5.57327569e-01 -6.98640347e-01 6.61827207e-01 6.73092902e-01
-2.99501479e-01 -1.01376951e-01 -1.19084394e+00 1.05770147e+00
1.13603294e+00 -4.08788472e-01 -9.85831618e-01 6.26260102e-01
7.87864268e-01 -6.35836244e-01 -1.04156411e+00 3.64357144e-01
5.18643975e-01 -9.66858923e-01 1.09002483e+00 -1.06574821e+00
1.14118323e-01 5.76248527e-01 -4.09717590e-01 -1.43995297e+00
1.89755633e-01 -8.49638879e-01 -2.15707928e-01 1.62664115e+00
5.34220517e-01 -4.46698129e-01 8.45149875e-01 1.15904975e+00
-1.54321447e-01 -5.60007870e-01 -1.05653703e+00 -8.88144970e-01
3.29506308e-01 -2.03954563e-01 6.85710490e-01 7.02583790e-01
8.51498067e-01 8.48156154e-01 -8.15253496e-01 -3.92707378e-01
3.74926776e-01 1.73962519e-01 1.36695516e+00 -1.11333132e+00
-2.87049323e-01 -2.74607569e-01 2.84986079e-01 -1.49735689e+00
6.78571641e-01 -5.13666332e-01 1.29357293e-01 -1.05049849e+00
3.30275856e-02 -4.91693944e-01 1.40333697e-01 5.12186050e-01
-2.24462211e-01 -3.18336487e-01 2.94935137e-01 3.49333555e-01
-8.77112031e-01 8.86535645e-01 9.84558821e-01 -2.46641740e-01
-5.98462760e-01 3.93543601e-01 -3.36394161e-01 7.18132615e-01
9.58822370e-01 1.03692368e-01 -6.29983723e-01 -3.07654440e-01
-5.35878479e-01 6.55408382e-01 -2.09693864e-01 -7.62275875e-01
4.34198827e-01 -2.57642508e-01 -3.76404524e-01 -7.38132477e-01
1.05071390e+00 -6.19325995e-01 -5.61902523e-01 4.83873971e-02
-7.43016720e-01 3.63588631e-02 3.47343497e-02 3.25920165e-01
-9.89618525e-02 -2.60439008e-01 8.50311756e-01 -2.72128344e-01
-7.74423242e-01 3.99243496e-02 -3.75947148e-01 1.59172252e-01
7.35411167e-01 3.43550712e-01 -7.69442141e-01 -1.13304901e+00
-7.31050491e-01 5.68140030e-01 1.02333084e-01 8.48478913e-01
3.50926191e-01 -1.07591629e+00 -5.92738211e-01 -2.00429559e-01
2.95630425e-01 -3.68566602e-01 3.85374784e-01 3.71651083e-01
-2.64858571e-03 6.76488161e-01 -2.85946339e-01 -4.75160003e-01
-1.50127149e+00 2.99835771e-01 3.36151361e-01 -4.32141453e-01
-4.84757483e-01 6.01867855e-01 -7.93004930e-02 -1.10657525e+00
7.04236865e-01 8.21883306e-02 -1.10526435e-01 -3.67627516e-02
7.19262064e-01 4.41058159e-01 -1.27851591e-01 -5.20966887e-01
-3.26295905e-02 -3.30360681e-01 -6.21093273e-01 -7.85259068e-01
1.00290883e+00 -3.46580118e-01 3.18485945e-01 6.90030813e-01
8.82764995e-01 -3.65669638e-01 -1.34762192e+00 -6.65450871e-01
1.98804751e-01 -4.09333616e-01 -4.84465450e-01 -1.20949209e+00
-4.84976023e-01 1.11140287e+00 4.82944131e-01 1.80506274e-01
5.23488104e-01 1.01169176e-01 1.06658137e+00 1.20922625e+00
7.61041105e-01 -1.23344326e+00 6.02761924e-01 8.81700039e-01
1.00232291e+00 -1.85824406e+00 -3.96377742e-01 -2.84111112e-01
-1.24640310e+00 9.19769287e-01 1.33479810e+00 4.77059275e-01
5.15380681e-01 2.45907661e-02 6.01922989e-01 1.87852964e-01
-1.14724433e+00 7.09686941e-03 -7.86632597e-02 5.91466486e-01
4.89051521e-01 -1.13494888e-01 -1.34870755e-02 1.04882324e+00
-2.39627659e-01 -4.36499447e-01 1.31574631e-01 8.92205000e-01
-5.87097466e-01 -1.46769857e+00 -7.83644337e-03 1.45588517e-01
-1.78971142e-02 9.95671526e-02 -7.87152767e-01 1.15008104e+00
-7.65513062e-01 1.59009373e+00 -2.38003626e-01 -7.52426326e-01
5.72645783e-01 7.05770314e-01 4.60139886e-02 -4.92527902e-01
-8.58205378e-01 1.79220811e-02 5.89441478e-01 -5.10458350e-01
-2.81610727e-01 -4.07813072e-01 -1.03694999e+00 -5.79810977e-01
-3.26036751e-01 2.21301526e-01 4.56469893e-01 9.78240192e-01
2.02823758e-01 8.78350958e-02 6.81894064e-01 -7.92281508e-01
-1.23972690e+00 -1.59744108e+00 -4.07872260e-01 4.08154577e-01
-2.13195249e-01 -7.02934921e-01 -2.40362868e-01 -1.26635402e-01] | [12.860274314880371, 8.041276931762695] |
77942b41-28ce-47fe-8db4-6c77bb3092f5 | definition-extraction-feature-analysis-from | null | null | https://aclanthology.org/2020.cogalex-1.10 | https://aclanthology.org/2020.cogalex-1.10.pdf | Definition Extraction Feature Analysis: From Canonical to Naturally-Occurring Definitions | Textual definitions constitute a fundamental source of knowledge when seeking the meaning of words, and they are the cornerstone of lexical resources like glossaries, dictionaries, encyclopedia or thesauri. In this paper, we present an in-depth analytical study on the main features relevant to the task of definition extraction. Our main goal is to study whether linguistic structures from canonical (the Aristotelian or genus et differentia model) can be leveraged to retrieve definitions from corpora in different domains of knowledge and textual genres alike. To this end, we develop a simple linear classifier and analyze the contribution of several (sets of) linguistic features. Finally, as a result of our experiments, we also shed light on the particularities of existing benchmarks as well as the most challenging aspects of the task. | ['Jose Camacho-Collados', 'Luis Espinosa Anke', 'Mireia Roig Mirapeix'] | null | null | null | null | coling-cogalex-2020-12 | ['definition-extraction'] | ['natural-language-processing'] | [ 6.75749257e-02 -2.83644140e-01 -6.45508170e-01 -4.32941131e-02
-3.94975692e-01 -1.23084891e+00 1.27549005e+00 7.42066681e-01
-7.25666940e-01 8.11564863e-01 4.88425374e-01 -4.89778727e-01
-5.16679585e-01 -7.28714764e-01 -1.45219952e-01 -4.37232196e-01
1.62637785e-01 1.64933264e-01 -1.22034997e-01 -5.95482826e-01
5.27326286e-01 3.81328613e-01 -1.68158829e+00 9.59953442e-02
4.69158560e-01 1.18784845e+00 1.54086258e-02 3.17341872e-02
-5.62095523e-01 4.64804202e-01 -6.20262086e-01 -6.37915432e-01
-4.77100387e-02 -1.63490370e-01 -1.08468091e+00 -2.39275455e-01
4.37825441e-01 4.37348336e-01 -1.31119743e-01 1.01559937e+00
1.07741706e-01 1.16317838e-01 9.46134150e-01 -7.61425257e-01
-3.20805132e-01 1.07279646e+00 -1.81238949e-01 5.24120092e-01
4.95476276e-01 -1.45838708e-01 1.47810137e+00 -8.35626006e-01
9.42230225e-01 1.01442409e+00 5.58550954e-01 1.92946389e-01
-9.68762398e-01 -2.67745078e-01 8.59825909e-02 2.92233586e-01
-1.41289270e+00 -5.64145803e-01 8.06565702e-01 -4.32757944e-01
8.37538302e-01 4.68784690e-01 5.23224592e-01 1.09548938e+00
-5.44758663e-02 4.31773335e-01 1.24300349e+00 -8.93794119e-01
-3.82029563e-02 2.70358503e-01 5.78149140e-01 5.55965424e-01
5.24142802e-01 -1.85477033e-01 -6.23172343e-01 -1.43672347e-01
4.90073174e-01 -3.48186970e-01 -4.64531809e-01 -8.32541510e-02
-1.14320242e+00 8.89259458e-01 2.58143507e-02 9.49531496e-01
-1.47590265e-01 -1.23120341e-02 6.70424879e-01 4.77968514e-01
3.92123222e-01 8.64047170e-01 -6.28686965e-01 -3.14832747e-01
-6.94008708e-01 1.73909858e-01 1.08098364e+00 9.06246841e-01
5.96180558e-01 -4.94125724e-01 8.68835822e-02 9.32462752e-01
1.02874562e-01 3.08595300e-01 6.51465058e-01 -4.43597585e-01
2.49481022e-01 6.47840023e-01 -1.25372941e-02 -9.65635657e-01
-5.84754884e-01 -6.00053251e-01 -2.99926162e-01 -6.38748169e-01
4.50949937e-01 -1.08443759e-01 -3.76190841e-01 1.48647237e+00
2.71126151e-01 -3.79146218e-01 -1.78229764e-01 5.41564226e-01
1.20850122e+00 2.53234535e-01 4.20978107e-02 -3.64034563e-01
1.56782055e+00 -4.33469921e-01 -5.88685036e-01 -8.92768279e-02
6.74988508e-01 -8.67519021e-01 1.07081532e+00 2.12801158e-01
-7.64612854e-01 -2.75390297e-01 -8.35009694e-01 -2.27774143e-01
-9.37641025e-01 1.62843853e-01 6.91070735e-01 4.49148327e-01
-3.76074225e-01 4.20343399e-01 -4.30699468e-01 -4.86866385e-01
7.24947527e-02 -2.72218406e-01 -2.14230508e-01 8.95508528e-02
-1.25286281e+00 1.30672014e+00 7.19702721e-01 -2.75717169e-01
-3.16194445e-01 -7.64772475e-01 -8.16485047e-01 -1.53582022e-01
8.25677276e-01 -4.74386871e-01 9.95740294e-01 -8.95652175e-01
-8.74435306e-01 1.39112461e+00 1.83307733e-02 -3.70416582e-01
5.26003875e-02 -1.68344736e-01 -4.58780557e-01 -4.72212546e-02
3.88921469e-01 4.47345823e-02 4.94758457e-01 -1.11415017e+00
-8.27405691e-01 -2.48349190e-01 4.61663693e-01 -9.39883515e-02
-5.76631606e-01 1.75423101e-01 -3.92316908e-01 -7.66602218e-01
-2.18008444e-01 -6.83602452e-01 3.23966473e-01 -3.46668154e-01
-4.73037541e-01 -7.81576633e-01 1.37760136e-02 -3.84985119e-01
2.00600648e+00 -2.07319236e+00 5.67581296e-01 2.58358777e-01
3.94620299e-01 -2.31336951e-02 3.38530719e-01 9.06460941e-01
-2.65302602e-02 4.81692612e-01 -1.51681632e-01 -9.01368260e-02
4.71013665e-01 4.25108880e-01 -4.37410474e-01 2.02061445e-01
-1.61264777e-01 1.03522170e+00 -9.66996074e-01 -5.68388164e-01
2.37984627e-01 1.49199605e-01 4.16912138e-02 -2.17405796e-01
-3.79092038e-01 -9.37767886e-03 -6.18239164e-01 4.44573134e-01
-2.19055340e-01 -2.23095015e-01 4.59707916e-01 -3.47918570e-01
-4.12264377e-01 1.08811378e+00 -8.24648201e-01 1.58057380e+00
-7.90277123e-01 7.02733815e-01 -2.30817989e-01 -1.01870608e+00
4.32373405e-01 -8.23820010e-04 3.08225065e-01 -5.83136082e-01
5.92671931e-01 3.64733696e-01 4.49458659e-02 -3.95606488e-01
4.77006197e-01 -2.97354877e-01 -4.61061239e-01 2.87225097e-01
1.32938772e-01 -2.82118201e-01 8.03989410e-01 1.03194311e-01
8.70593607e-01 -2.40646988e-01 1.06694794e+00 -6.41453862e-01
7.39262640e-01 2.90030688e-01 1.88566595e-01 5.63364685e-01
3.08026254e-01 2.39610568e-01 6.10573411e-01 -6.05612457e-01
-8.50598335e-01 -8.57836366e-01 -7.27778494e-01 9.12308931e-01
-4.02182303e-02 -1.11944103e+00 -3.46730918e-01 -6.72673523e-01
8.88093337e-02 8.76217484e-01 -7.56351292e-01 3.75183895e-02
-5.23724854e-01 -5.92037380e-01 6.12132967e-01 1.84513941e-01
2.73663282e-01 -9.61270213e-01 -8.44739139e-01 -3.80504914e-02
-1.06208012e-01 -1.22312534e+00 -2.28172809e-01 3.69370997e-01
-1.01083502e-01 -1.15793860e+00 -2.60938019e-01 -6.56881988e-01
-1.29861543e-02 -6.72067255e-02 1.69455099e+00 2.22502142e-01
-2.21149981e-01 5.25745392e-01 -7.80878782e-01 -6.60153568e-01
-2.74077117e-01 6.23655856e-01 3.91950198e-02 -2.59129226e-01
5.98231733e-01 -4.75972116e-01 -2.37516966e-02 -1.63749456e-01
-1.08595538e+00 -3.56159747e-01 3.39260966e-01 5.67779422e-01
4.09421802e-01 -3.05373445e-02 5.08862078e-01 -1.18449271e+00
9.55580473e-01 -5.98187745e-01 -4.97457713e-01 2.89670676e-01
-6.14522219e-01 3.68989497e-01 7.41895676e-01 -3.87285352e-01
-5.07276058e-01 -3.38001013e-01 -1.57352790e-01 -4.18537855e-02
-9.61235911e-03 1.24629998e+00 -8.51311758e-02 3.49142812e-02
7.85291076e-01 2.86355466e-01 -3.61609012e-01 -7.37752736e-01
5.81947327e-01 6.08627081e-01 3.28047574e-01 -8.29496801e-01
7.34296679e-01 3.12194973e-01 7.86739141e-02 -1.18994653e+00
-1.45834911e+00 -8.18142414e-01 -7.94136345e-01 -6.70283437e-02
5.07778585e-01 -3.82257372e-01 -2.80425936e-01 -1.41941875e-01
-9.73820627e-01 8.09198692e-02 -3.34312439e-01 1.55144170e-01
-2.44363859e-01 4.78835881e-01 -2.91845948e-01 -4.95291322e-01
-2.59628147e-01 -5.43342888e-01 8.88295114e-01 8.95628482e-02
-5.80951512e-01 -1.48972714e+00 1.87646225e-01 1.01614393e-01
2.65824348e-01 5.53418040e-01 1.22432101e+00 -8.50569665e-01
1.00294747e-04 1.31136075e-01 2.20152497e-01 1.86475828e-01
4.12602127e-01 -6.06514513e-02 -4.48570430e-01 1.28375843e-01
2.67755315e-02 -3.05944175e-01 1.18681431e+00 1.69684167e-03
7.76456594e-01 -4.60544288e-01 -1.97358474e-01 3.97145599e-01
1.59929943e+00 -4.43610817e-01 1.70995533e-01 6.23556137e-01
3.76291156e-01 6.95496082e-01 5.24525583e-01 4.50613230e-01
3.55265796e-01 7.61269152e-01 1.41504053e-02 3.61570716e-01
-2.09108755e-01 1.66435558e-02 5.47470078e-02 1.05783796e+00
3.48957703e-02 -1.05247080e-01 -1.28495884e+00 8.40995252e-01
-1.48072469e+00 -8.45237195e-01 -9.32281464e-02 2.11485791e+00
1.34838903e+00 2.10632190e-01 2.54753202e-01 1.85436800e-01
3.23202610e-01 2.44749427e-01 5.57881370e-02 -2.85152078e-01
-2.36609608e-01 8.13040614e-01 4.59978223e-01 3.14469784e-01
-1.31482029e+00 1.08739555e+00 5.90867424e+00 1.18872631e+00
-9.27554488e-01 -1.02435179e-01 -1.27491325e-01 -9.52522829e-02
-4.35455739e-01 -5.64516373e-02 -8.82594049e-01 5.24136722e-01
7.90705085e-01 -4.95143861e-01 2.86233604e-01 5.58837652e-01
-2.87521034e-01 -4.91542593e-02 -1.33628047e+00 1.08511353e+00
4.63935673e-01 -1.37258506e+00 2.63879269e-01 -7.81006068e-02
5.18349230e-01 3.00102402e-02 -7.00059310e-02 1.09927624e-01
1.11243837e-02 -1.16552258e+00 8.20256293e-01 3.18110853e-01
8.16142201e-01 -6.36546016e-01 7.64575720e-01 2.96930432e-01
-1.06131780e+00 1.02960341e-01 -2.53575236e-01 -3.19211245e-01
1.02759011e-01 5.39584517e-01 -4.49373096e-01 7.76830196e-01
2.12257713e-01 8.64131749e-01 -6.85296297e-01 7.78226554e-01
-5.83068490e-01 7.46622324e-01 -4.38209593e-01 -4.89217848e-01
3.76479000e-01 -3.60420346e-02 5.92918932e-01 1.60778809e+00
-1.82342678e-01 1.72840968e-01 1.53238446e-01 7.97663569e-01
-2.27391660e-01 5.47791302e-01 -5.03475189e-01 -4.03819233e-01
5.89958251e-01 1.31692278e+00 -6.96854949e-01 -1.51657537e-01
-5.63223004e-01 2.57939726e-01 4.19777244e-01 6.64604455e-02
-6.26195729e-01 -4.64750081e-01 5.75853169e-01 1.83249395e-02
2.28503630e-01 -5.58097601e-01 -3.16457838e-01 -1.47583401e+00
-4.44246363e-03 -9.10665810e-01 5.26494324e-01 -2.50677764e-01
-1.62484169e+00 3.16991419e-01 2.77533501e-01 -9.31883633e-01
-3.52285445e-01 -1.01116192e+00 -1.88575342e-01 5.68347633e-01
-1.53589380e+00 -1.21473217e+00 8.94332957e-03 3.31445575e-01
4.79440063e-01 -2.37684119e-02 9.31728482e-01 2.68590838e-01
-4.42381531e-01 3.90357733e-01 1.10720180e-01 4.97887164e-01
4.98976141e-01 -1.19017076e+00 1.26509324e-01 6.10116720e-01
9.55764472e-01 1.03173494e+00 7.77192116e-01 -3.04130107e-01
-1.28140569e+00 -6.29117310e-01 1.17695403e+00 -8.28589797e-01
1.15384018e+00 -1.92227617e-01 -7.46590376e-01 5.86946070e-01
2.14781612e-01 -5.88710845e-01 8.88195455e-01 7.18410850e-01
-5.54171860e-01 1.31839618e-01 -5.55756748e-01 6.06668830e-01
1.09737325e+00 -5.73732018e-01 -1.07619512e+00 2.35863373e-01
4.38094318e-01 -2.28767186e-01 -1.14846039e+00 2.64958829e-01
8.44353497e-01 -5.87472677e-01 8.47053766e-01 -8.30481231e-01
5.87202191e-01 -1.37174148e-02 -1.84594885e-01 -1.15148067e+00
-1.61930639e-02 -4.35661256e-01 7.17584491e-02 1.34219217e+00
7.88791835e-01 -4.28752184e-01 1.79032922e-01 2.52799481e-01
-1.73275976e-03 -8.00768197e-01 -1.05053782e+00 -7.55725026e-01
6.42806709e-01 -6.19115829e-01 2.32129186e-01 1.23161054e+00
3.63893628e-01 1.02266550e+00 1.64279088e-01 -4.95090604e-01
3.10726315e-01 3.48622233e-01 4.12366182e-01 -1.29427660e+00
-7.33445734e-02 -8.22469056e-01 -5.21632195e-01 -1.03999507e+00
4.86765087e-01 -1.28093946e+00 -2.97395289e-01 -1.35129881e+00
1.00485370e-01 -3.59540105e-01 -2.72320449e-01 4.49444473e-01
-8.87627453e-02 1.46116316e-01 3.13341081e-01 2.63680756e-01
-6.65848792e-01 3.05554777e-01 7.09788501e-01 -1.32889077e-01
-6.54919446e-02 -3.03799868e-01 -7.74618626e-01 8.63839507e-01
8.62340569e-01 -4.75152552e-01 -1.47101924e-01 -6.09184325e-01
7.78603137e-01 -5.93167961e-01 9.54510793e-02 -3.94779861e-01
9.91497263e-02 -2.78252572e-01 -1.05941564e-01 -7.36901462e-02
1.69070914e-01 -6.97779238e-01 -3.07891518e-01 1.21192120e-01
-3.63701522e-01 1.13585941e-01 1.06129065e-01 1.03839055e-01
-2.97036290e-01 -5.31349123e-01 5.76586902e-01 -2.72311896e-01
-6.54410422e-01 -1.82641849e-01 -2.86194414e-01 7.81479239e-01
6.21292293e-01 2.19057098e-01 -1.23871543e-01 -1.55256629e-01
-3.78762871e-01 2.04640955e-01 5.30185699e-01 5.35650432e-01
2.80404121e-01 -1.07372880e+00 -7.75571764e-01 -3.32923651e-01
3.94547582e-01 -3.38564008e-01 -6.52305782e-01 7.56767511e-01
-3.27907920e-01 7.72995532e-01 -2.14146706e-03 -2.23437831e-01
-1.19165075e+00 6.92327976e-01 1.69598088e-01 -2.20491126e-01
-3.94111782e-01 6.39686882e-01 9.11855400e-02 -2.66506165e-01
1.44419685e-01 -5.24559140e-01 -5.92756212e-01 4.09818798e-01
4.44553167e-01 2.70968914e-01 2.85582338e-02 -8.26372385e-01
-5.20736635e-01 5.72169900e-01 1.75801322e-01 -1.73605144e-01
1.32540679e+00 -2.76785135e-01 -4.47675079e-01 1.10685706e+00
1.11262500e+00 4.15633500e-01 -1.28179193e-01 -3.33364189e-01
6.58899963e-01 -8.24890807e-02 -1.59365147e-01 -9.39609587e-01
-4.45742697e-01 5.21504998e-01 -1.98516205e-01 6.20052516e-01
9.84605908e-01 5.41751027e-01 5.08976877e-01 5.05140066e-01
3.85788590e-01 -1.00347030e+00 -2.83687472e-01 8.85615885e-01
9.69530702e-01 -8.59539628e-01 1.58537254e-01 -5.43689072e-01
-3.57713282e-01 1.31303751e+00 -8.71190429e-02 -4.44344878e-02
8.27693403e-01 9.27343741e-02 -8.93468186e-02 -6.91979408e-01
-7.25672781e-01 -8.38000298e-01 6.24509573e-01 1.87493786e-01
7.15729535e-01 -7.12044686e-02 -1.24987221e+00 9.29198384e-01
-7.30460823e-01 -4.90508854e-01 3.07567358e-01 6.34236515e-01
-3.61879826e-01 -1.27720094e+00 3.24225463e-02 5.53077221e-01
-9.19035017e-01 -5.34367800e-01 -1.10174763e+00 1.09522128e+00
3.60859990e-01 8.77964854e-01 -2.36574814e-01 1.21594162e-03
1.09046511e-01 1.91581339e-01 6.54725432e-01 -8.91995609e-01
-7.38397002e-01 -1.93990439e-01 5.32779276e-01 -8.87201503e-02
-6.69250429e-01 -6.91426396e-01 -1.04382896e+00 -2.62577355e-01
-3.77639383e-01 2.83344418e-01 4.61719394e-01 1.35733032e+00
-1.04182512e-01 2.07971379e-01 2.60773271e-01 -2.69720823e-01
-4.72878844e-01 -1.09279430e+00 -4.91298556e-01 6.83769703e-01
8.54360685e-03 -7.99036264e-01 -5.42172492e-01 1.10663556e-01] | [10.238306045532227, 9.182602882385254] |
c9d3d3cc-bfbe-435a-90d0-652b6f33361c | lisens-a-scalable-architecture-for-video | 1503.04267 | null | http://arxiv.org/abs/1503.04267v1 | http://arxiv.org/pdf/1503.04267v1.pdf | LiSens --- A Scalable Architecture for Video Compressive Sensing | The measurement rate of cameras that take spatially multiplexed measurements
by using spatial light modulators (SLM) is often limited by the switching speed
of the SLMs. This is especially true for single-pixel cameras where the
photodetector operates at a rate that is many orders-of-magnitude greater than
the SLM. We study the factors that determine the measurement rate for such
spatial multiplexing cameras (SMC) and show that increasing the number of
pixels in the device improves the measurement rate, but there is an optimum
number of pixels (typically, few thousands) beyond which the measurement rate
does not increase. This motivates the design of LiSens, a novel imaging
architecture, that replaces the photodetector in the single-pixel camera with a
1D linear array or a line-sensor. We illustrate the optical architecture
underlying LiSens, build a prototype, and demonstrate results of a range of
indoor and outdoor scenes. LiSens delivers on the promise of SMCs: imaging at a
megapixel resolution, at video rate, using an inexpensive low-resolution
sensor. | ['Jian Wang', 'Mohit Gupta', 'Aswin C. Sankaranarayanan'] | 2015-03-14 | null | null | null | null | ['video-compressive-sensing'] | ['computer-vision'] | [ 8.18311572e-01 -1.58971116e-01 3.14862058e-02 -1.19378507e-01
-1.92005605e-01 -5.96393585e-01 4.50431466e-01 -7.58876085e-01
-7.88386881e-01 8.75108063e-01 -2.14741170e-01 -3.93925697e-01
1.94499582e-01 -6.55389130e-01 -7.43357599e-01 -7.00118780e-01
3.36761057e-01 -1.64898202e-01 8.58540118e-01 2.85592765e-01
1.91942781e-01 3.89479518e-01 -1.64263892e+00 1.95136771e-01
3.45522702e-01 1.25166214e+00 5.46105504e-01 1.02020657e+00
1.64994970e-01 7.65169561e-01 -8.66287529e-01 2.55396307e-01
2.98549861e-01 -6.60750210e-01 -9.91247073e-02 2.56991476e-01
1.02771997e+00 -9.28837478e-01 -4.03535604e-01 7.39723325e-01
2.59367581e-02 -3.88195217e-01 3.11149746e-01 -1.01928568e+00
-1.31570563e-01 -6.94579259e-02 -6.02608263e-01 4.39239353e-01
4.32194829e-01 3.90001088e-01 4.49252576e-01 -3.96972597e-01
6.43170238e-01 1.08613694e+00 6.25549555e-01 3.75391930e-01
-1.46399736e+00 -7.12473691e-01 -4.06649709e-01 -3.53134006e-01
-1.15701091e+00 -8.13425004e-01 1.73966661e-01 -3.46103311e-01
1.19276321e+00 -4.02642973e-02 5.69753945e-01 6.05204701e-01
7.24071383e-01 -3.39237660e-01 1.68477237e+00 -3.63931715e-01
1.97922438e-01 3.25396895e-01 -1.30464137e-01 4.38771576e-01
6.53036594e-01 1.73385039e-01 -9.09922123e-01 2.11175025e-01
1.53609896e+00 3.92686836e-02 -3.94271433e-01 -5.80625348e-02
-1.07946074e+00 2.62031794e-01 1.14454091e-01 3.60048234e-01
5.55749461e-02 5.57585478e-01 -2.54827261e-01 5.23239315e-01
4.29039216e-03 8.51620436e-01 3.00503343e-01 -2.53415853e-01
-9.96604443e-01 -4.45112377e-01 6.77986562e-01 1.12958360e+00
5.55297554e-01 -2.46577322e-01 8.94816741e-02 3.46253037e-01
2.04262719e-01 5.80473840e-01 1.61590680e-01 -1.56247032e+00
4.05757844e-01 5.58836460e-01 5.08178771e-01 -2.15594351e-01
-1.56453043e-01 -3.21983844e-02 -2.87983596e-01 1.02705491e+00
6.28681183e-01 -3.23368579e-01 -5.61783731e-01 1.07597744e+00
-1.35582551e-01 -8.68589878e-02 -1.42791241e-01 1.10703838e+00
3.89999539e-01 6.74740136e-01 -4.93638724e-01 -5.41801214e-01
1.20580459e+00 -5.18480182e-01 -6.75589263e-01 -2.32566103e-01
2.62579322e-01 -9.48667467e-01 9.88876224e-01 4.93147850e-01
-1.19282210e+00 -4.24106568e-01 -1.47420895e+00 -3.89037654e-02
1.60927996e-01 2.79668570e-01 3.25924493e-02 7.94792175e-01
-9.65022027e-01 5.38990676e-01 -9.26406920e-01 -6.89740360e-01
1.19184799e-01 6.65021360e-01 -1.60243452e-01 -1.36916846e-01
-2.65498042e-01 7.47768342e-01 4.05617245e-02 -5.59961379e-01
-1.22096010e-01 -7.56787241e-01 -1.05354756e-01 -9.07950774e-02
2.05946326e-01 -7.74667263e-01 1.02288914e+00 -4.84112203e-01
-2.02584887e+00 9.10732746e-01 -2.30433166e-01 -6.88283205e-01
3.21989119e-01 -1.20162502e-01 -4.31740135e-01 5.68069220e-01
-1.29257336e-01 5.22115946e-01 7.83788204e-01 -8.17259192e-01
-8.97354782e-01 -9.42483470e-02 2.00602248e-01 -1.45628855e-01
-2.83251911e-01 2.66479284e-01 -4.80692312e-02 4.04458314e-01
1.77733347e-01 -7.38071322e-01 4.03589532e-02 3.34349364e-01
-1.22464612e-01 3.72550189e-01 1.39714158e+00 2.04839334e-01
9.72221076e-01 -2.40963125e+00 -5.01807094e-01 -4.38772351e-01
3.42680305e-01 3.18287164e-01 2.52025612e-02 6.64246857e-01
7.24957228e-01 -2.85157233e-01 -4.79440689e-02 -1.25491008e-01
-6.91921711e-01 2.18991384e-01 -7.59508014e-02 4.81754929e-01
-2.50127204e-02 2.57942557e-01 -5.58898628e-01 -2.18770385e-01
4.17969912e-01 1.88497201e-01 -3.69230509e-01 1.22028857e-01
-1.08940073e-03 7.56592095e-01 -1.02993451e-01 4.89708841e-01
8.17699313e-01 -5.31400144e-01 2.92061806e-01 -3.20931256e-01
-1.22672904e+00 5.20863831e-01 -1.28837442e+00 1.08832836e+00
-2.29321003e-01 1.20253444e+00 4.23025638e-01 -1.00265257e-01
1.05467522e+00 1.69900522e-01 4.75041330e-01 -7.75700271e-01
-1.83038250e-01 3.89376432e-01 -3.16591084e-01 -5.84430456e-01
5.95258355e-01 -6.03267372e-01 2.85943538e-01 2.46652991e-01
-7.77427256e-02 -3.66829187e-01 4.82146591e-01 -1.45910963e-01
1.42077267e+00 -1.55887216e-01 1.94728449e-01 -3.31923366e-01
1.44362584e-01 2.09112130e-02 4.36628610e-01 8.00009966e-01
3.54118198e-02 4.76100862e-01 2.52463043e-01 -1.17311418e-01
-1.44587934e+00 -1.17605734e+00 -3.66370231e-01 3.30456406e-01
5.90324879e-01 -4.27396923e-01 -4.38387275e-01 5.10640562e-01
1.79720595e-01 1.35856420e-01 2.90070981e-01 3.92619938e-01
-4.25468206e-01 -4.42833006e-01 2.18879342e-01 4.29911852e-01
9.76145148e-01 -2.49349922e-02 -1.35951996e+00 4.11163002e-01
4.78831321e-01 -1.87896526e+00 -9.70244482e-02 -1.08623065e-01
-8.96497309e-01 -1.09818888e+00 -7.93872476e-02 -3.59496713e-01
4.86616760e-01 6.55900180e-01 8.29632640e-01 -6.05794787e-01
-4.31423694e-01 7.45857120e-01 -1.75362658e-02 -1.57568067e-01
-1.20754592e-01 -4.34269190e-01 2.80086488e-01 -1.40580967e-01
3.88739347e-01 -8.13572049e-01 -1.02628386e+00 5.49145043e-01
-6.38475060e-01 2.62151271e-01 7.14042544e-01 3.98079723e-01
3.28755766e-01 -4.83137406e-02 -8.49530101e-02 -6.56763732e-01
2.94097438e-02 1.28085271e-01 -1.51483440e+00 -3.62460375e-01
-6.02618635e-01 -2.39682943e-01 4.63214159e-01 -4.25392210e-01
-8.81366670e-01 2.00606987e-01 4.58867490e-01 -1.66227475e-01
-2.46569738e-02 -1.70410156e-01 2.68580854e-01 -8.14349592e-01
7.10296512e-01 1.32942870e-02 2.09674224e-01 -2.64206737e-01
2.28518210e-02 7.50399411e-01 6.38376474e-01 -8.19485821e-03
5.29946446e-01 1.10743177e+00 5.54943085e-01 -1.36256433e+00
-1.81561604e-01 -6.04277134e-01 -5.15506446e-01 -3.58226508e-01
8.62029552e-01 -1.22438872e+00 -9.88186240e-01 4.91460294e-01
-1.18368959e+00 -2.11542726e-01 -3.33415806e-01 9.98882711e-01
-3.14960033e-01 -1.52758092e-01 -8.84392142e-01 -8.66224647e-01
3.01845282e-01 -1.06493831e+00 1.01066351e+00 5.29089987e-01
-2.75632620e-01 -5.37906110e-01 -2.58266717e-01 4.84003216e-01
7.02478647e-01 1.93194479e-01 5.45117319e-01 6.48916304e-01
-1.62352753e+00 -8.79524350e-02 -4.95546401e-01 3.29171956e-01
1.06411427e-01 9.69057679e-02 -9.66474235e-01 -3.89171958e-01
2.66043216e-01 -1.24943100e-01 7.24021077e-01 7.22212315e-01
4.39133257e-01 7.52682798e-03 -6.83884859e-01 7.13203490e-01
1.99914658e+00 4.23658013e-01 8.56527984e-01 1.26522616e-01
3.09031665e-01 3.09408680e-02 2.12910742e-01 3.22863132e-01
-1.26361817e-01 8.46645832e-01 3.12929899e-01 1.26269042e-01
-2.81415313e-01 9.02238954e-03 6.09768569e-01 6.77978635e-01
-2.37371281e-01 -2.77111143e-01 -3.77073169e-01 -4.60385438e-03
-1.09628820e+00 -9.64022756e-01 -4.85303253e-01 2.45924020e+00
5.89033425e-01 -6.13650866e-02 4.31205705e-02 -2.15458691e-01
5.38567662e-01 1.88417405e-01 -3.65291953e-01 -1.40464067e-01
-2.19046235e-01 1.43962977e-02 1.44565356e+00 6.70529425e-01
-7.64642775e-01 5.32870531e-01 7.96768618e+00 3.22973490e-01
-1.41723025e+00 1.35741562e-01 1.89997852e-01 -7.52809167e-01
5.84212132e-02 7.84870982e-02 -1.38869703e+00 9.20262218e-01
8.59605134e-01 1.12417214e-01 3.50815445e-01 3.06349307e-01
5.20434558e-01 -9.99707699e-01 -1.25388229e+00 1.30374825e+00
-2.58601815e-01 -1.40747619e+00 -2.58523971e-01 5.33176720e-01
7.29269743e-01 -2.18011495e-02 1.68694183e-01 -7.25263119e-01
-1.33574322e-01 -6.94224775e-01 2.20928729e-01 5.22571921e-01
1.25345695e+00 -3.36353965e-02 1.66221738e-01 2.93231070e-01
-8.70550632e-01 -2.63699114e-01 -6.72392190e-01 -5.75656950e-01
2.80432105e-01 1.03951705e+00 -7.40059912e-01 -2.98619390e-01
5.96581519e-01 4.08444345e-01 -2.46651135e-02 1.04896402e+00
1.05981171e-01 5.33448517e-01 -6.41860664e-01 -1.33558720e-01
5.46863563e-02 -7.54401386e-01 8.75610888e-01 8.87184203e-01
8.64743173e-01 4.55729157e-01 -2.23265260e-01 1.08425939e+00
-8.34101960e-02 -8.41767609e-01 -6.86292708e-01 6.81983232e-02
7.69417644e-01 1.20879090e+00 -4.61215913e-01 -1.98830783e-01
-1.02552211e+00 9.65410173e-01 -3.20924550e-01 2.96649933e-01
-2.41446629e-01 -3.77817005e-01 8.53694916e-01 7.82370508e-01
2.45984212e-01 -9.40388799e-01 -2.40586460e-01 -8.02428186e-01
2.70946234e-01 -3.43461007e-01 -2.95076221e-01 -9.00583982e-01
-9.32285428e-01 -1.35386661e-01 -3.39507312e-01 -1.44669628e+00
2.81746268e-01 -1.01165533e+00 -3.54971528e-01 8.58182132e-01
-1.37040007e+00 -7.01547146e-01 -4.19252366e-01 2.77259529e-01
1.31807849e-01 1.41131237e-01 3.59162152e-01 3.46404970e-01
-9.74592492e-02 -3.85323986e-02 5.61653078e-01 -1.33030578e-01
7.93484747e-01 -8.01488101e-01 9.98738408e-02 9.33829427e-01
-3.29340339e-01 5.73375881e-01 5.61228275e-01 -3.93251479e-01
-1.79033101e+00 -7.17863977e-01 6.67981625e-01 -5.16136229e-01
4.12173957e-01 -3.65523964e-01 -3.21391344e-01 4.85264957e-01
3.33048373e-01 1.37852980e-02 6.08999014e-01 -4.47270632e-01
-2.54748221e-02 -5.93747437e-01 -1.29414797e+00 4.22674358e-01
1.33767223e+00 -8.24982524e-01 -3.79161388e-02 2.62847096e-01
5.11109114e-01 -4.09705997e-01 -7.99324870e-01 8.62147883e-02
9.78603840e-01 -1.31261611e+00 6.33084059e-01 3.62685591e-01
4.88914698e-01 -6.33863091e-01 -1.99799925e-01 -8.43461931e-01
-4.02581900e-01 -8.42561066e-01 2.49107152e-01 1.00174046e+00
1.53305650e-01 -1.19466114e+00 9.36375737e-01 6.42518938e-01
3.14022601e-02 -2.70669937e-01 -1.25356388e+00 -1.24354076e+00
-5.88669300e-01 2.37849802e-01 2.28172556e-01 3.73839825e-01
1.27568111e-01 4.88152742e-01 -2.63898522e-01 3.07202548e-01
9.53919470e-01 -1.61678806e-01 5.90568423e-01 -9.28722084e-01
-4.62581068e-01 -2.06551641e-01 -1.03807604e+00 -1.85910010e+00
-8.78448069e-01 -3.75256352e-02 -4.02545214e-01 -1.15996861e+00
9.21365395e-02 -1.53817847e-01 2.68875927e-01 -4.35791761e-01
3.63153249e-01 2.70468384e-01 3.65738273e-01 4.88418788e-01
-4.97271508e-01 -2.37561733e-01 1.21988058e+00 2.27417022e-01
-3.70653421e-01 -1.37785614e-01 -1.79357544e-01 3.53138685e-01
3.49340647e-01 -8.14592019e-02 -2.51380295e-01 -5.82090795e-01
8.37525427e-02 -1.29532302e-02 6.78556025e-01 -1.73091972e+00
7.67176092e-01 -2.81523634e-02 5.13346791e-01 -2.56533384e-01
8.39219272e-01 -9.28361118e-01 5.29034913e-01 5.42305768e-01
-4.55237851e-02 -3.24707806e-01 -1.52929723e-01 3.62376183e-01
-1.64633110e-01 -1.69276372e-02 1.17359340e+00 -3.40185046e-01
-8.12945187e-01 -1.77092090e-01 -6.71916723e-01 -2.75801361e-01
1.02124965e+00 -8.10205221e-01 -1.02247775e+00 -1.36989191e-01
-1.27012894e-01 -2.13436827e-01 9.96904671e-01 -2.57242799e-01
4.72396284e-01 -1.13284373e+00 1.09462179e-01 4.24676567e-01
-5.24753481e-02 -2.27995485e-01 -9.08942372e-02 9.72566366e-01
-9.43449557e-01 7.07699955e-01 -2.74237871e-01 -9.68013644e-01
-1.47049057e+00 6.57822415e-02 5.09608030e-01 3.23273540e-01
-6.51547372e-01 8.39552701e-01 -1.54756278e-01 5.01103818e-01
-2.26824090e-01 -4.78379846e-01 3.63928974e-01 -4.92598861e-01
7.52976537e-01 6.93215549e-01 -3.19691211e-01 -3.95116545e-02
-6.76363856e-02 1.03972793e+00 2.63847589e-01 -2.83854425e-01
1.04530013e+00 -5.34313738e-01 -3.21855217e-01 7.34018385e-01
9.62807953e-01 5.71727492e-02 -1.70280766e+00 -2.29686394e-01
-5.75264335e-01 -9.75773215e-01 1.16335146e-01 -3.50710541e-01
-5.06216049e-01 6.90522492e-01 7.10131526e-01 2.08099902e-01
1.03269935e+00 7.34245181e-02 7.07403481e-01 2.54946649e-01
6.56903744e-01 -1.40414572e+00 -3.06876283e-03 1.43561840e-01
1.53208315e-01 -1.07425439e+00 3.00879151e-01 -8.04711759e-01
-5.13189025e-02 1.23180556e+00 5.40584326e-01 -3.29362303e-02
3.15704674e-01 8.49577606e-01 -6.05845302e-02 -1.69027552e-01
-6.86866462e-01 7.41442591e-02 -4.25728768e-01 4.48680311e-01
5.28100252e-01 2.54017562e-01 -3.62857223e-01 -3.54091078e-01
2.44478919e-02 4.39196944e-01 8.44745755e-01 1.14998305e+00
-1.00701571e+00 -1.14851022e+00 -6.60600603e-01 7.24551618e-01
3.41608673e-02 1.22756161e-01 -1.05278105e-01 4.57141548e-01
2.94288874e-01 1.23901057e+00 6.07546628e-01 -2.60550499e-01
4.01286840e-01 -1.35777399e-01 9.54747915e-01 -5.54050088e-01
8.04113001e-02 -9.42156613e-02 1.58471465e-01 -7.29561269e-01
-6.87854707e-01 -7.25255728e-01 -7.94394732e-01 -5.89198709e-01
-5.73287085e-02 -5.04012406e-01 6.46100640e-01 6.87371433e-01
5.96402645e-01 1.67984515e-01 6.65263593e-01 -4.82403159e-01
-1.84489742e-01 -6.01941288e-01 -1.13131893e+00 -1.90725788e-01
6.29464865e-01 -3.62952113e-01 -2.70418525e-01 2.17275620e-01] | [9.605571746826172, -2.5944254398345947] |
9b937ee4-9cc3-477f-a5f6-7c31477dff43 | amorphous-fortress-observing-emergent | 2306.13169 | null | https://arxiv.org/abs/2306.13169v1 | https://arxiv.org/pdf/2306.13169v1.pdf | Amorphous Fortress: Observing Emergent Behavior in Multi-Agent FSMs | We introduce a system called Amorphous Fortress -- an abstract, yet spatial, open-ended artificial life simulation. In this environment, the agents are represented as finite-state machines (FSMs) which allow for multi-agent interaction within a constrained space. These agents are created by randomly generating and evolving the FSMs; sampling from pre-defined states and transitions. This environment was designed to explore the emergent AI behaviors found implicitly in simulation games such as Dwarf Fortress or The Sims. We apply the hill-climber evolutionary search algorithm to this environment to explore the various levels of depth and interaction from the generated FSMs. | ['Julian Togelius', 'Sam Earle', 'Dipika Rajesh', 'M Charity'] | 2023-06-22 | null | null | null | null | ['artificial-life'] | ['miscellaneous'] | [-7.65639246e-02 1.69572517e-01 2.61601597e-01 5.70538044e-02
2.86700100e-01 -7.09800899e-01 1.10190129e+00 -1.99685887e-01
-3.56935501e-01 1.13650358e+00 -1.57424331e-01 -2.12612003e-01
-1.88184872e-01 -1.03804290e+00 -2.43724257e-01 -5.72923899e-01
-9.42526162e-01 1.03924870e+00 6.12200916e-01 -7.00662315e-01
1.28095090e-01 5.57975054e-01 -2.01833653e+00 8.24437477e-03
1.05362558e+00 1.24436870e-01 2.45204926e-01 1.11460221e+00
8.56464505e-02 1.11784267e+00 -9.44203496e-01 2.42536455e-01
2.85257757e-01 -9.27906454e-01 -6.79586470e-01 1.35695904e-01
-8.67511511e-01 1.22579090e-01 -5.53472519e-01 8.85083318e-01
3.59996296e-02 6.07350647e-01 4.12627637e-01 -1.49539280e+00
-1.54809758e-01 6.74492121e-01 -7.68902153e-02 2.04357147e-01
6.69928730e-01 8.14735591e-01 1.86130136e-01 -2.15783060e-01
8.75567377e-01 1.38778925e+00 2.67833859e-01 8.63874018e-01
-1.21522081e+00 -3.08382630e-01 -2.62829095e-01 -9.65168923e-02
-1.77863383e+00 -4.37408864e-01 2.65150964e-01 -2.02640682e-01
1.53015435e+00 7.52182961e-01 1.50404894e+00 8.36076021e-01
8.39952707e-01 5.29494107e-01 1.16930056e+00 -5.14578998e-01
1.09429622e+00 -1.60818428e-01 -2.96269447e-01 7.68894553e-01
2.15344906e-01 7.14638650e-01 -4.13796574e-01 -7.93708324e-01
1.06568491e+00 -3.64388704e-01 1.81334138e-01 -4.72732842e-01
-1.09293342e+00 5.37134171e-01 -5.27891563e-04 2.69197494e-01
-7.10436225e-01 4.09139186e-01 1.15158679e-02 3.33515644e-01
-1.65013596e-01 1.17012513e+00 -1.52572393e-01 -5.34731388e-01
-7.38648295e-01 7.23054111e-01 1.17598331e+00 6.82565987e-01
4.63220954e-01 3.21045667e-01 1.29817426e-01 -2.15075687e-02
4.03570712e-01 1.43316641e-01 3.69447917e-01 -1.38224256e+00
-4.67220217e-01 7.61132002e-01 5.37559748e-01 -2.34429941e-01
-3.97885293e-01 -1.98004991e-01 -3.64058882e-01 8.61131132e-01
6.43514283e-03 -6.01870656e-01 -1.03724957e+00 1.49124706e+00
5.79815507e-01 2.91304708e-01 5.16590595e-01 6.50613070e-01
2.02055648e-01 1.01916862e+00 1.11302339e-01 -4.69091654e-01
1.02585828e+00 -8.84841919e-01 -6.71728790e-01 -1.86879143e-01
3.67542684e-01 2.05088276e-02 6.89542055e-01 2.15860948e-01
-1.50446582e+00 4.85470891e-02 -1.32992816e+00 9.44166243e-01
-4.33340728e-01 -8.37438881e-01 7.68978596e-01 6.41838253e-01
-1.37965930e+00 6.40984297e-01 -1.47770107e+00 -3.80138725e-01
-1.12117216e-01 5.39314806e-01 2.31156290e-01 5.93463778e-01
-1.19264066e+00 1.16945124e+00 7.08644807e-01 -3.46200347e-01
-1.41990364e+00 7.93362260e-02 -6.36775672e-01 4.87664901e-02
5.14091671e-01 -9.56820965e-01 1.35288489e+00 -1.07667387e+00
-2.11859941e+00 6.21919930e-01 2.18889005e-02 -5.31465948e-01
2.77393460e-01 6.11317098e-01 -5.85418463e-01 1.27316028e-01
-2.67689198e-01 4.74778473e-01 4.37183350e-01 -1.50033092e+00
-2.18079284e-01 -1.08123891e-01 9.54148322e-02 6.75434530e-01
4.11407292e-01 4.13720012e-01 1.41826823e-01 -2.06105962e-01
-2.41264969e-01 -1.36561036e+00 -1.16125607e+00 -8.47740173e-01
-2.97059238e-01 5.19995801e-02 5.18999398e-01 1.17442504e-01
1.19889200e+00 -1.71584439e+00 5.82328618e-01 4.59792107e-01
1.85738355e-01 2.54379828e-02 -6.19595461e-02 1.17872632e+00
3.04858416e-01 5.54071255e-02 -2.08836019e-01 1.31600335e-01
2.12681368e-01 4.05880153e-01 2.19068602e-01 2.46653482e-01
-2.46306404e-01 8.83031547e-01 -1.08232999e+00 -6.23969495e-01
3.08502484e-02 -3.22668478e-02 -4.69520122e-01 2.84505695e-01
-7.67977118e-01 5.60038865e-01 -8.36040616e-01 4.33009446e-01
3.26923311e-01 -1.99259549e-01 5.21898687e-01 1.12050998e+00
-7.05697417e-01 3.64514403e-02 -1.16260850e+00 1.58066738e+00
-6.12583943e-02 2.49585286e-01 1.97856456e-01 -1.97538003e-01
7.45002329e-01 2.80260295e-01 4.14824694e-01 -3.68872166e-01
3.26947123e-01 2.14512214e-01 5.59248924e-01 -2.66641527e-01
7.47003675e-01 -9.20806229e-02 -4.26253349e-01 7.84522057e-01
-2.55238622e-01 -1.00909078e+00 5.97426891e-01 3.36415112e-01
1.60458076e+00 9.10500064e-02 5.03222466e-01 -6.60539269e-01
1.29179671e-01 5.74893951e-01 4.40840065e-01 1.14551353e+00
-3.35066289e-01 -4.79251705e-02 9.71195027e-02 -6.28840864e-01
-1.04011548e+00 -1.30689752e+00 2.16816694e-01 8.32138896e-01
5.49392998e-01 -7.89528191e-01 -1.06670904e+00 2.05217339e-02
-2.94439375e-01 1.17683649e+00 -5.29008508e-01 -3.98496628e-01
-5.32942772e-01 -7.05606103e-01 3.92677873e-01 -1.12746172e-01
4.99701619e-01 -1.80740440e+00 -1.71200061e+00 5.04541337e-01
4.71505821e-01 -3.01160097e-01 -7.23903775e-02 3.84396315e-01
-5.91825426e-01 -4.96358722e-01 -6.49738088e-02 -6.01848722e-01
4.20358956e-01 -5.24848759e-01 1.30662715e+00 3.27559650e-01
-4.96798605e-01 3.00820500e-01 -1.72230631e-01 -2.88055509e-01
-1.10988617e+00 -9.99322459e-02 2.36649677e-01 -5.07811666e-01
4.96453494e-02 -7.47802019e-01 -4.53066826e-01 2.73343474e-01
-9.83833194e-01 4.26262319e-01 1.63558468e-01 7.38621831e-01
1.64375275e-01 5.41342258e-01 4.19350982e-01 -1.60809070e-01
1.03938770e+00 -5.13094783e-01 -7.67728746e-01 3.79576445e-01
-2.26292953e-01 1.64964557e-01 7.35614240e-01 -5.68618774e-01
-1.04555321e+00 -1.38576686e-01 3.04767996e-01 -1.00415044e-01
-3.93021226e-01 4.55432981e-01 -1.52335707e-02 -4.24519181e-02
8.11575234e-01 6.52544260e-01 1.54269397e-01 2.28956774e-01
3.41586918e-01 6.48779213e-01 4.36174035e-01 -9.42415595e-01
5.38987756e-01 1.75392434e-01 -5.19551560e-02 -7.98898518e-01
3.04693073e-01 4.81331587e-01 -9.87118408e-02 -4.97002423e-01
7.12295055e-01 -2.72894174e-01 -9.75281060e-01 7.59234488e-01
-7.87469089e-01 -1.03683877e+00 -8.84938061e-01 2.04840228e-02
-9.00711060e-01 -1.82523176e-01 -7.39300370e-01 -1.36033964e+00
1.34773587e-03 -1.20789099e+00 6.22052550e-01 8.10912728e-01
-8.34294558e-01 -7.01603830e-01 7.67893672e-01 -5.07845759e-01
5.87806761e-01 5.26909947e-01 6.66734099e-01 -5.34660041e-01
-6.72810972e-01 1.97328478e-01 8.31440389e-01 -6.65428042e-01
1.23185562e-02 4.91294980e-01 -1.60787553e-01 -5.68330407e-01
-1.54257476e-01 -1.61619797e-01 -1.42972872e-01 3.49454671e-01
5.85506439e-01 -3.80592942e-01 -7.52470315e-01 3.82600009e-01
1.05976617e+00 1.19107234e+00 6.70093596e-01 7.16481149e-01
-2.79535890e-01 4.77713346e-01 3.94633561e-01 1.07664430e+00
3.70030075e-01 4.68481749e-01 3.79923224e-01 2.21862808e-01
4.14027929e-01 -1.67267814e-01 3.45666558e-01 3.80082011e-01
-1.19223624e-01 -4.78022218e-01 -1.25280857e+00 5.53035319e-01
-1.99486458e+00 -1.29104042e+00 3.38946462e-01 1.81441724e+00
1.01134872e+00 8.89737234e-02 3.33596528e-01 -4.05172348e-01
8.69025767e-01 6.46065474e-02 -8.02199543e-01 -8.11337829e-01
-2.04319417e-01 3.58689964e-01 9.15258825e-02 5.06403089e-01
-5.15583217e-01 1.13968158e+00 7.72802544e+00 5.99171281e-01
-6.26313388e-01 -2.94480294e-01 3.89842749e-01 -6.14820242e-01
-2.50307024e-01 3.55033010e-01 -7.50702694e-02 6.20246112e-01
1.45563900e+00 -8.93809676e-01 1.22483754e+00 2.88907319e-01
4.50358421e-01 -5.33109963e-01 -7.59764791e-01 4.47250575e-01
-5.84458649e-01 -1.54346681e+00 -2.83195406e-01 2.14146376e-01
1.16353488e+00 -1.00641839e-01 -6.92099938e-03 1.48148015e-01
1.65243649e+00 -1.32171023e+00 1.09148622e+00 3.88189226e-01
5.78073740e-01 -1.24187803e+00 3.50681297e-03 8.56874943e-01
-1.10497928e+00 -3.77568156e-01 7.39310756e-02 -4.40701514e-01
5.50675809e-01 -1.87473983e-01 -4.77992654e-01 4.45942469e-02
6.51060164e-01 -2.13643655e-01 -1.34976223e-01 9.85276282e-01
1.05438679e-01 1.64484352e-01 -5.04851818e-01 -7.59623349e-01
5.42314172e-01 -7.48851240e-01 1.04323494e+00 7.09122658e-01
3.98684829e-01 7.05351770e-01 2.67455935e-01 1.13113260e+00
6.24857724e-01 -5.92737079e-01 -8.17240894e-01 -4.20208216e-01
9.39577222e-01 8.81869435e-01 -1.12487721e+00 -5.46835303e-01
3.10392380e-01 8.35169494e-01 -6.75502419e-02 3.63118589e-01
-9.04476106e-01 -5.30826092e-01 9.13375378e-01 1.28798991e-01
-1.14419594e-01 -5.48464656e-01 -1.00474823e-02 -1.09285557e+00
-9.10493612e-01 -1.44030106e+00 1.54394194e-01 -8.35673869e-01
-6.55238330e-01 1.02963305e+00 2.97710389e-01 -5.81207752e-01
-9.54828143e-01 1.19548872e-01 -8.91365707e-01 7.30889142e-01
-1.94405600e-01 -6.31528020e-01 1.78750381e-01 5.82175255e-01
4.58694160e-01 -5.14021397e-01 8.93923044e-01 -8.25425744e-01
-6.86905086e-01 -4.10995223e-02 3.61467093e-01 -5.74752867e-01
-4.85501975e-01 -1.15333915e+00 8.81092548e-01 8.48339796e-01
-2.31181741e-01 8.45033407e-01 1.34551835e+00 -1.05800569e+00
-1.67455935e+00 -4.42390174e-01 1.80555925e-01 -3.51335496e-01
6.11234009e-01 -4.69342321e-01 -6.11496568e-01 7.25227892e-01
8.56146634e-01 -2.87111044e-01 2.96422750e-01 -4.32464361e-01
6.08553052e-01 7.23747313e-01 -1.42492306e+00 1.44509101e+00
1.42735970e+00 -7.35142231e-02 -5.06932437e-01 -9.60925072e-02
5.82504094e-01 -5.17256320e-01 -6.33321226e-01 5.75364493e-02
3.96050811e-01 -1.04086053e+00 7.19001710e-01 -7.52770483e-01
8.96390378e-02 -4.85976160e-01 -5.37600601e-04 -1.86844516e+00
-4.67861325e-01 -1.53105533e+00 -2.87008822e-01 6.77085519e-01
2.30110884e-01 -1.00087535e+00 8.39142799e-01 8.44412148e-01
-1.16004169e-01 -5.67425013e-01 -8.50204051e-01 -9.16753471e-01
5.53492568e-02 2.76050478e-01 1.28773844e+00 7.42645741e-01
6.60899818e-01 1.44700827e-02 2.10534096e-01 5.91456406e-02
6.10707343e-01 2.68043667e-01 4.03034091e-01 -9.68019485e-01
-7.34049082e-01 -5.06771564e-01 -3.95190939e-02 -5.04670024e-01
2.26254061e-01 -3.76653105e-01 3.48250955e-01 -1.21692455e+00
1.38598084e-01 -5.36535144e-01 2.08218306e-01 1.94899485e-01
2.71992505e-01 -4.95005280e-01 3.20576131e-01 1.72224864e-01
-9.10999596e-01 8.35645795e-01 1.21703207e+00 3.73997897e-01
-5.20176947e-01 -2.64551401e-01 -3.16850305e-01 7.26697326e-01
7.93790817e-01 -3.99452239e-01 -5.86758792e-01 2.34294280e-01
1.88851744e-01 8.18782806e-01 7.29958117e-02 -1.14467013e+00
2.47084290e-01 -1.02811122e+00 -1.04577698e-01 -3.02542627e-01
5.25633216e-01 -4.15590078e-01 1.21503460e+00 1.01422620e+00
-2.83666670e-01 6.57198071e-01 1.78921059e-01 1.67113170e-01
1.78593457e-01 -2.10057423e-01 7.86798596e-01 -6.64881647e-01
-7.22289920e-01 -1.01583049e-01 -1.45800555e+00 -4.60232887e-03
1.74505293e+00 -5.28666079e-01 -3.39096308e-01 -4.40732330e-01
-9.00684178e-01 4.41355109e-01 1.38688529e+00 -2.74227858e-02
4.90006983e-01 -1.05608058e+00 -5.86622953e-01 2.91664660e-01
-3.05720538e-01 -1.00283064e-01 -6.87177256e-02 -2.35883463e-02
-9.61419404e-01 -1.15144543e-01 -7.82795727e-01 -2.10473672e-01
-9.04233277e-01 5.08069456e-01 8.92274022e-01 -3.47806334e-01
-4.87437159e-01 5.64126015e-01 2.17317902e-02 -5.25222421e-01
-5.57525158e-01 1.65625110e-01 5.53609729e-02 -6.45627975e-01
3.72451603e-01 5.99512815e-01 -5.39592445e-01 -3.87010396e-01
-5.78725398e-01 -2.43927807e-01 5.21169782e-01 -9.73433793e-01
1.39614558e+00 -2.05387138e-02 -3.71302277e-01 3.75168145e-01
1.61403939e-01 -2.19610080e-01 -1.39575052e+00 4.67642784e-01
4.97592725e-02 -2.90787876e-01 -4.08867300e-01 -7.92281747e-01
-3.00639182e-01 -1.36215091e-01 1.50729984e-01 8.76652062e-01
9.08651769e-01 -1.05564937e-01 3.84270519e-01 2.96764225e-01
1.08237648e+00 -1.09485972e+00 1.01214059e-01 7.79510140e-01
6.07862115e-01 -2.23333105e-01 -3.41618389e-01 1.31990865e-01
-6.74736679e-01 7.53664851e-01 8.98715973e-01 -4.20605063e-01
1.57801375e-01 1.26294184e+00 -2.05467224e-01 -4.73625422e-01
-1.51964366e+00 -4.93951514e-02 -5.57079554e-01 8.10922921e-01
-1.43350795e-01 3.37293863e-01 -1.54568449e-01 2.43907094e-01
-5.12658775e-01 -7.03106001e-02 1.08198416e+00 1.50786054e+00
-6.66459143e-01 -1.03000271e+00 -6.98991060e-01 1.59479186e-01
8.86215195e-02 2.55021572e-01 -7.55926847e-01 7.98202813e-01
-7.31351525e-02 1.07067418e+00 3.03446561e-01 -2.92605728e-01
-4.93452139e-02 -2.02555612e-01 7.00296581e-01 -5.77379405e-01
-9.68179584e-01 -4.21021193e-01 4.05716836e-01 -5.09846807e-01
2.81105816e-01 -1.07050097e+00 -1.66816080e+00 -7.12216139e-01
-1.33709431e-01 7.96112597e-01 4.16768521e-01 4.88538563e-01
3.86735082e-01 5.72773039e-01 5.41486263e-01 -1.13691127e+00
-4.87966537e-01 -5.37638605e-01 -8.51727128e-01 -6.94250837e-02
-5.91519736e-02 -5.44444799e-01 -1.68211207e-01 -3.48217696e-01] | [3.710787773132324, 1.5218619108200073] |
695d62e5-495b-4a54-a78e-e2e070ba9394 | humanmac-masked-motion-completion-for-human | 2302.03665 | null | https://arxiv.org/abs/2302.03665v2 | https://arxiv.org/pdf/2302.03665v2.pdf | HumanMAC: Masked Motion Completion for Human Motion Prediction | Human motion prediction is a classical problem in computer vision and computer graphics, which has a wide range of practical applications. Previous effects achieve great empirical performance based on an encoding-decoding style. The methods of this style work by first encoding previous motions to latent representations and then decoding the latent representations into predicted motions. However, in practice, they are still unsatisfactory due to several issues, including complicated loss constraints, cumbersome training processes, and scarce switch of different categories of motions in prediction. In this paper, to address the above issues, we jump out of the foregoing style and propose a novel framework from a new perspective. Specifically, our framework works in a masked completion fashion. In the training stage, we learn a motion diffusion model that generates motions from random noise. In the inference stage, with a denoising procedure, we make motion prediction conditioning on observed motions to output more continuous and controllable predictions. The proposed framework enjoys promising algorithmic properties, which only needs one loss in optimization and is trained in an end-to-end manner. Additionally, it accomplishes the switch of different categories of motions effectively, which is significant in realistic tasks, e.g., the animation task. Comprehensive experiments on benchmarks confirm the superiority of the proposed framework. The project page is available at https://lhchen.top/Human-MAC. | ['Tongliang Liu', 'Xiaobo Xia', 'Yiren Pang', 'Yewen Li', 'Jiawei Zhang', 'Ling-Hao Chen'] | 2023-02-07 | null | null | null | null | ['motion-prediction'] | ['computer-vision'] | [ 4.23230737e-01 -1.82705685e-01 -2.99966037e-01 -1.96938545e-01
-4.15181518e-01 -1.17198907e-01 5.61884820e-01 -4.38807249e-01
-2.34745473e-01 4.78798687e-01 3.56018066e-01 -1.58766210e-01
1.37683108e-01 -6.89469755e-01 -6.23926163e-01 -1.06772840e+00
3.09961796e-01 1.05689235e-01 3.25606555e-01 -8.46638680e-02
1.06057018e-01 2.13535786e-01 -1.26135302e+00 9.38568711e-02
7.96947896e-01 7.85945833e-01 5.00901043e-01 4.61418062e-01
1.29170373e-01 9.63065624e-01 -4.65876013e-02 -2.84561723e-01
1.48838237e-01 -6.07248127e-01 -4.75887746e-01 4.75803941e-01
9.93412510e-02 -4.33114320e-01 -6.18527472e-01 1.12834072e+00
3.95708442e-01 2.93617010e-01 4.72520053e-01 -9.88334537e-01
-6.02340221e-01 3.08746010e-01 -7.34913409e-01 -1.54142499e-01
3.00632894e-01 4.00793612e-01 9.28808689e-01 -9.05189455e-01
4.60134894e-01 1.38050187e+00 4.52783048e-01 7.72969127e-01
-1.19619977e+00 -5.81423163e-01 3.27443868e-01 2.00149506e-01
-1.20120454e+00 -6.01775467e-01 9.12932694e-01 -6.96484089e-01
5.38399279e-01 2.15382725e-01 4.53677803e-01 1.23131919e+00
1.18575715e-01 1.09332466e+00 6.70728981e-01 -1.45824224e-01
2.36024573e-01 -2.31987715e-01 -1.47087604e-01 6.63021564e-01
5.34763420e-03 -8.86943284e-03 -2.99712747e-01 -1.11170830e-02
1.01027572e+00 3.58189881e-01 -6.76450312e-01 -3.67069572e-01
-1.41059721e+00 7.40244269e-01 3.33259374e-01 1.02998227e-01
-4.31574255e-01 2.68225759e-01 2.39921406e-01 -6.17413670e-02
3.81502420e-01 -1.79938912e-01 -6.65933937e-02 -9.58446711e-02
-9.89148855e-01 2.88625002e-01 4.19219404e-01 8.42403352e-01
6.48083210e-01 2.00313449e-01 -2.37669945e-01 7.46432066e-01
4.05724227e-01 3.94747794e-01 6.87793791e-01 -1.09145427e+00
7.03758717e-01 1.06976226e-01 3.11710000e-01 -1.57276511e+00
-1.49035752e-01 -3.14671099e-01 -1.39363801e+00 1.20618820e-01
3.62382501e-01 -1.20025270e-01 -8.19030762e-01 1.85448563e+00
3.86408985e-01 6.30593121e-01 -6.51780963e-02 1.16686630e+00
3.40838909e-01 1.05941212e+00 1.65028661e-01 -4.50821936e-01
1.11536217e+00 -1.36501801e+00 -8.31493974e-01 -2.73629755e-01
4.74497408e-01 -7.84348071e-01 1.08078849e+00 4.62421864e-01
-1.10164630e+00 -9.41568136e-01 -7.57605314e-01 -3.91207933e-02
4.04580683e-01 3.93042594e-01 6.66347504e-01 3.02200973e-01
-8.72397900e-01 7.35194266e-01 -1.20819032e+00 -1.83405906e-01
2.40612164e-01 2.39859462e-01 -2.30318516e-01 -8.50860998e-02
-9.39735889e-01 2.12802410e-01 2.65839666e-01 5.33308029e-01
-8.14493835e-01 -1.70166850e-01 -9.47625637e-01 6.96042553e-02
5.43380678e-01 -9.70037103e-01 1.19112647e+00 -1.04309106e+00
-1.80572724e+00 2.72786260e-01 -3.83950680e-01 -2.02103376e-01
8.30112994e-01 -4.40058827e-01 -2.21210644e-01 9.87008139e-02
8.32994804e-02 5.62995553e-01 1.17314053e+00 -9.76709127e-01
-6.56533301e-01 -1.93598002e-01 -2.04590052e-01 1.52423039e-01
-3.65288854e-01 -3.06899637e-01 -8.95684421e-01 -1.05057144e+00
2.18654990e-01 -1.02597189e+00 -6.73369288e-01 5.16704395e-02
-2.90982485e-01 -1.15809090e-01 6.05689645e-01 -5.85017204e-01
1.33607864e+00 -2.38771987e+00 6.26513362e-01 -1.60128534e-01
2.62650937e-01 2.43074983e-01 5.21246530e-02 3.03729326e-01
1.18847825e-02 -7.76420236e-02 -4.62841690e-01 -6.82547390e-01
-9.98433009e-02 1.74001783e-01 -6.26248896e-01 4.83384311e-01
1.08857691e-01 7.27057755e-01 -9.64203358e-01 -3.97078902e-01
2.78466731e-01 4.63714659e-01 -7.23492384e-01 4.38143373e-01
-2.46659368e-01 9.39800680e-01 -7.56763339e-01 3.84286225e-01
6.00104928e-01 -4.20702815e-01 1.19626455e-01 -6.94535077e-02
-8.18245262e-02 6.72990531e-02 -1.27021611e+00 1.94913912e+00
-3.57441962e-01 3.55356842e-01 -3.79721820e-02 -1.15724313e+00
7.85275757e-01 3.19562674e-01 3.60501856e-01 -1.43635601e-01
-2.43488490e-03 1.25343546e-01 -1.78052470e-01 -7.71674097e-01
4.12457198e-01 -1.41465902e-01 -1.42617552e-02 1.04194894e-01
-3.24042678e-01 2.78611690e-01 -6.41596736e-03 -2.25168858e-02
8.91027093e-01 5.30621350e-01 2.68574029e-01 1.70660749e-01
7.33743548e-01 -4.94891480e-02 1.01512372e+00 3.86659533e-01
-1.43948719e-01 7.21983194e-01 4.15935248e-01 -3.97928059e-01
-9.05309558e-01 -8.44175220e-01 2.88282841e-01 7.21169293e-01
4.32853878e-01 -3.14353317e-01 -4.55833048e-01 -4.27075624e-01
-3.79649788e-01 4.19734448e-01 -4.17025805e-01 -2.08958879e-01
-9.21614110e-01 -7.02123284e-01 1.72783017e-01 6.12292767e-01
4.71976519e-01 -1.08412755e+00 -5.61872363e-01 4.24051076e-01
-3.47467333e-01 -1.16305768e+00 -6.32638395e-01 -4.43419158e-01
-1.16805398e+00 -5.75022638e-01 -1.13215995e+00 -9.93340909e-01
6.32243574e-01 5.59399426e-01 7.44321167e-01 6.96266666e-02
-7.94405341e-02 3.01499176e-03 -3.34786534e-01 2.95502722e-01
-2.01618418e-01 -9.48489234e-02 2.38918941e-02 4.68062699e-01
1.61960632e-01 -6.54564679e-01 -1.00424695e+00 3.79585922e-01
-9.72588122e-01 3.55410039e-01 8.53068411e-01 1.09103429e+00
6.38722301e-01 1.16532773e-01 2.29407534e-01 -7.83756137e-01
2.66665131e-01 -6.81059480e-01 -5.99758625e-01 3.81147079e-02
-2.04244196e-01 -1.63221471e-02 8.87888551e-01 -4.67601836e-01
-1.25569081e+00 5.32612741e-01 -3.43000352e-01 -7.51737773e-01
-3.04063916e-01 3.74594450e-01 -4.15889233e-01 2.72679687e-01
1.65781677e-01 4.72597241e-01 2.18840986e-02 -7.18290567e-01
3.03632021e-01 3.47499698e-01 7.11227953e-01 -4.75923300e-01
9.47763324e-01 6.27288818e-01 -6.53386712e-02 -8.36873770e-01
-4.73533958e-01 -4.32091713e-01 -4.62779850e-01 -1.13906160e-01
9.88548279e-01 -1.00208616e+00 -6.45399749e-01 6.12770081e-01
-1.21771193e+00 -2.93796360e-01 4.86975908e-02 8.01571250e-01
-6.05896354e-01 9.57863092e-01 -6.93926156e-01 -7.22644985e-01
-1.18096046e-01 -1.36800539e+00 1.01101005e+00 2.61560529e-01
-1.24148399e-01 -8.80488575e-01 9.14679393e-02 1.95221305e-01
9.85489488e-02 2.59625077e-01 7.13180304e-01 9.76289138e-02
-8.40305328e-01 -8.23426247e-02 8.82403646e-03 1.93418965e-01
1.30227432e-01 -6.14519529e-02 -7.41619170e-01 -3.71701956e-01
1.52732655e-01 -1.62858099e-01 1.11945164e+00 3.97726506e-01
1.42019272e+00 -3.16189170e-01 -3.89477104e-01 8.72118652e-01
1.36352766e+00 2.25507259e-01 7.41991758e-01 1.87095761e-01
9.30723727e-01 7.00217426e-01 5.91570437e-01 6.40148163e-01
2.68491685e-01 8.21003199e-01 4.47189271e-01 3.63033861e-02
5.06082848e-02 -3.98919344e-01 5.46537101e-01 1.13414109e+00
-3.41243535e-01 -3.27169240e-01 -6.57138586e-01 4.89114076e-01
-2.16954446e+00 -1.10798943e+00 -1.49424344e-01 2.18647408e+00
6.37797832e-01 1.17369585e-01 7.41656572e-02 1.02527112e-01
7.92334974e-01 4.99692142e-01 -5.08523464e-01 7.61827081e-02
2.40649760e-01 -1.09288149e-01 1.71580255e-01 4.52995777e-01
-1.11214244e+00 1.01112843e+00 4.78108454e+00 9.14322674e-01
-1.13903809e+00 7.15962276e-02 6.31506622e-01 1.52613446e-01
-2.59756774e-01 1.19433209e-01 -7.63123274e-01 6.48918927e-01
3.25539351e-01 8.21661502e-02 2.63680249e-01 8.26397479e-01
5.15387893e-01 1.81155011e-01 -9.15069699e-01 1.16269255e+00
-1.80156007e-01 -1.22939014e+00 2.28146672e-01 -3.66157666e-02
5.44355989e-01 -4.12172884e-01 1.01267919e-01 2.82186121e-01
-3.13207395e-02 -9.18140531e-01 7.85419703e-01 6.53939128e-01
6.10328674e-01 -5.58382452e-01 5.11627674e-01 6.33911610e-01
-1.36535156e+00 -4.73014452e-02 -5.89403629e-01 -3.74358058e-01
5.86030483e-01 5.78762412e-01 -1.64084077e-01 6.09207511e-01
4.54365224e-01 1.08729374e+00 -1.25378445e-01 9.84182715e-01
-4.13360596e-01 6.86411381e-01 3.23438756e-02 1.35961443e-01
2.96008825e-01 -5.22815347e-01 5.12894452e-01 9.98461783e-01
5.18619895e-01 1.94846541e-01 4.36714411e-01 7.38754690e-01
6.11009821e-02 1.83133036e-02 -4.74211305e-01 9.37074497e-02
2.12987453e-01 1.12710321e+00 -6.36704743e-01 -3.18579912e-01
-5.43143451e-01 1.31171775e+00 2.56741554e-01 5.13537705e-01
-9.94606018e-01 -2.18516365e-01 5.36816776e-01 -1.05743203e-02
3.90218765e-01 -4.46061373e-01 -1.53290868e-01 -1.54028392e+00
2.57683486e-01 -7.13128269e-01 8.83715004e-02 -3.37413728e-01
-1.17443120e+00 5.88292718e-01 -2.02608049e-01 -1.67924011e+00
-3.31565946e-01 -4.34270859e-01 -7.48921990e-01 7.61875391e-01
-1.35286272e+00 -8.69278848e-01 -5.32006443e-01 6.23084009e-01
9.01888251e-01 4.49575484e-02 4.26867604e-01 4.78560627e-01
-8.42808843e-01 4.83159781e-01 2.22805232e-01 1.49025559e-01
6.11354470e-01 -9.11018670e-01 5.08457422e-01 1.07453144e+00
-4.91233021e-02 5.83712041e-01 6.78582966e-01 -6.21336997e-01
-1.33427048e+00 -1.21472263e+00 7.16759562e-01 3.10330037e-02
6.09340787e-01 -1.96856946e-01 -9.16897595e-01 6.64028168e-01
6.89645782e-02 -9.58907306e-02 4.15102392e-01 -3.52623373e-01
1.38405174e-01 8.50235745e-02 -3.63587320e-01 9.17802989e-01
1.17399502e+00 -2.07869232e-01 -5.11819303e-01 2.88172096e-01
6.86096966e-01 -4.13089871e-01 -5.11587262e-01 4.27132130e-01
4.96890575e-01 -9.65160489e-01 9.71828938e-01 -4.61176485e-01
8.60936642e-01 -5.19119263e-01 -3.10490411e-02 -1.01916599e+00
-5.59244752e-01 -9.64627206e-01 -2.22113520e-01 1.12668836e+00
6.68187886e-02 -5.44384062e-01 8.94467831e-01 3.19285959e-01
-1.71948597e-01 -1.01017666e+00 -6.95913374e-01 -6.59679890e-01
-1.41271397e-01 -3.20526212e-01 3.28534663e-01 7.83006549e-01
-3.83358449e-01 3.41109693e-01 -1.06214213e+00 2.77429581e-01
6.66977406e-01 4.30192709e-01 1.04622710e+00 -8.29845965e-01
-9.13964868e-01 -4.12126154e-01 -3.40433031e-01 -2.01827979e+00
3.46755460e-02 -5.32749295e-01 2.79126763e-01 -1.37931943e+00
1.17576577e-01 -3.28953117e-01 -5.39146811e-02 2.08146691e-01
-5.34096241e-01 2.74546482e-02 3.83408189e-01 7.76729345e-01
-4.70509350e-01 9.89089072e-01 1.39007974e+00 -1.21126831e-01
-1.56588942e-01 2.95731634e-01 -3.53551894e-01 1.07497811e+00
7.97833383e-01 -3.65924388e-01 -5.84710956e-01 -7.58481503e-01
-1.18455522e-01 4.24325168e-01 5.51472068e-01 -9.97051239e-01
2.29516223e-01 -3.00083160e-01 3.31842244e-01 -4.24845129e-01
5.38990915e-01 -6.83757424e-01 1.82183161e-01 6.86892629e-01
-2.37348914e-01 5.20421639e-02 -2.81405061e-01 1.06185377e+00
-3.86984766e-01 -1.10266022e-01 7.01068759e-01 -1.42109707e-01
-9.48699474e-01 6.15284145e-01 -2.83617914e-01 -1.10643893e-01
9.35071349e-01 -1.96308464e-01 -6.94437942e-04 -4.72417593e-01
-8.63168895e-01 1.92361817e-01 5.50607622e-01 3.48147362e-01
7.00255215e-01 -1.37451458e+00 -5.90968192e-01 1.72201976e-01
-2.56295145e-01 2.37014890e-01 5.56811750e-01 1.08157980e+00
-6.08994961e-01 1.42546296e-01 -8.69255513e-02 -7.84768403e-01
-8.98200333e-01 6.70246661e-01 -6.13553524e-02 -2.60649025e-01
-9.34956908e-01 7.38354206e-01 7.81561136e-01 7.87586048e-02
4.00526673e-01 -2.30566546e-01 -3.24153006e-01 -2.55710095e-01
5.59804022e-01 2.68208981e-01 -5.28840005e-01 -7.66239703e-01
-6.77049458e-02 8.07029366e-01 4.18013260e-02 6.60555158e-03
1.19383681e+00 -3.60743940e-01 7.79439807e-02 3.74423921e-01
1.15910327e+00 1.11819565e-01 -1.60991311e+00 -2.62370259e-01
1.03442110e-02 -7.62878478e-01 -2.82301009e-01 7.99418613e-02
-1.13303792e+00 1.15531230e+00 3.22889328e-01 -1.16543770e-02
1.30254209e+00 -3.33317190e-01 1.27253008e+00 1.53770357e-01
2.71323174e-01 -7.46217966e-01 1.10901639e-01 2.95804292e-01
6.59023523e-01 -1.17439747e+00 -1.31532341e-01 -5.04041314e-01
-7.57930934e-01 1.18008709e+00 6.00184083e-01 -2.45122075e-01
4.84673738e-01 -1.05837412e-01 -1.06278084e-01 1.97535589e-01
-7.17147827e-01 -7.59647638e-02 1.26164317e-01 1.66411266e-01
4.28575277e-01 -6.32548705e-03 -4.62421477e-01 6.54005051e-01
1.31609425e-01 2.44640633e-01 3.03324133e-01 7.28214622e-01
-4.31143254e-01 -1.04433727e+00 -4.36306506e-01 -2.53327359e-02
-3.54947925e-01 7.94754028e-02 1.82849541e-01 4.93284792e-01
8.12712163e-02 8.37716520e-01 -1.85655653e-01 -4.81292963e-01
1.72798946e-01 -3.02104235e-01 8.70966613e-02 -4.88838255e-01
2.87313778e-02 5.77576935e-01 -2.27499470e-01 -7.10249722e-01
-6.65610492e-01 -6.91650152e-01 -1.10330689e+00 -2.64171571e-01
-7.84683377e-02 8.68600681e-02 6.39115795e-02 8.35209250e-01
3.25870663e-01 4.55070883e-01 7.56707788e-01 -1.16301501e+00
-7.33920574e-01 -7.44900823e-01 -3.90128076e-01 5.68511605e-01
2.68698931e-01 -7.23594308e-01 -2.59238541e-01 4.16501373e-01] | [10.739822387695312, -1.0810681581497192] |
87bdaee8-81d1-4f01-9429-9d46f5eb29fe | remote-sensing-image-classification-with-the | 2104.00704 | null | https://arxiv.org/abs/2104.00704v1 | https://arxiv.org/pdf/2104.00704v1.pdf | Remote Sensing Image Classification with the SEN12MS Dataset | Image classification is one of the main drivers of the rapid developments in deep learning with convolutional neural networks for computer vision. So is the analogous task of scene classification in remote sensing. However, in contrast to the computer vision community that has long been using well-established, large-scale standard datasets to train and benchmark high-capacity models, the remote sensing community still largely relies on relatively small and often application-dependend datasets, thus lacking comparability. With this letter, we present a classification-oriented conversion of the SEN12MS dataset. Using that, we provide results for several baseline models based on two standard CNN architectures and different input data configurations. Our results support the benchmarking of remote sensing image classification and provide insights to the benefit of multi-spectral data and multi-sensor data fusion over conventional RGB imagery. | ['Yu-Lun Wu', 'Michael Schmitt'] | 2021-04-01 | null | null | null | null | ['remote-sensing-image-classification'] | ['miscellaneous'] | [ 5.76970816e-01 -4.70226556e-01 8.89782906e-02 -4.92248356e-01
-6.74097776e-01 -5.51756203e-01 7.73324013e-01 7.07701147e-02
-8.67049456e-01 6.39549255e-01 6.56238869e-02 -7.99801528e-01
-5.32799959e-01 -1.16099894e+00 -5.11835694e-01 -7.99574792e-01
6.03687624e-03 1.22710288e-01 -2.47656003e-01 -4.60008115e-01
-1.56928599e-01 9.86035585e-01 -1.73040283e+00 3.87257218e-01
4.88722175e-01 1.36862504e+00 1.92690104e-01 8.57042432e-01
-1.15353307e-02 6.16403461e-01 -1.88571274e-01 -4.74476591e-02
5.39405704e-01 -1.38338014e-01 -9.04571235e-01 1.87041208e-01
6.97327793e-01 -3.41066003e-01 -2.95462400e-01 8.57597172e-01
7.07776070e-01 8.36215541e-02 2.53754020e-01 -8.46832991e-01
-5.27883649e-01 3.17242056e-01 -2.33599037e-01 4.50689316e-01
-3.80655646e-01 3.20751667e-01 9.76968467e-01 -6.53404355e-01
3.67111653e-01 1.01118219e+00 1.03193784e+00 6.84973896e-02
-1.30099726e+00 -4.48642254e-01 -1.29475996e-01 7.22188354e-02
-1.26936698e+00 -4.90618229e-01 4.18010473e-01 -4.65325743e-01
1.17064905e+00 5.18380046e-01 7.57651448e-01 7.03157127e-01
-8.42167959e-02 1.68036714e-01 1.57856989e+00 -3.83602738e-01
3.70486900e-02 -5.64452335e-02 1.86336517e-01 1.33992553e-01
2.90880203e-01 3.85939449e-01 -2.15952903e-01 1.97064564e-01
6.01565123e-01 2.94849455e-01 -3.13978523e-01 9.26868692e-02
-1.31505907e+00 9.60258842e-01 1.05265641e+00 4.03191835e-01
-6.52937829e-01 2.63483882e-01 3.41938615e-01 4.26754087e-01
6.76509500e-01 2.75468498e-01 -7.09984422e-01 1.73332945e-01
-1.12238526e+00 2.28941962e-01 4.62635607e-01 4.34535235e-01
1.07224381e+00 1.93430126e-01 2.70395011e-01 7.57481039e-01
2.44811922e-01 7.58901060e-01 1.67626917e-01 -7.12076306e-01
3.80077958e-01 7.40884483e-01 -1.15856342e-01 -7.98090279e-01
-6.25056446e-01 -6.50941133e-01 -1.24688852e+00 5.25518298e-01
2.39556804e-01 -9.70602408e-02 -8.73977661e-01 1.13623154e+00
-1.12276167e-01 -1.95623055e-01 2.84606129e-01 9.17976499e-01
9.62465048e-01 5.67972720e-01 1.33661777e-01 4.01406676e-01
1.04587770e+00 -6.48538470e-01 -5.97620197e-02 -4.24803674e-01
4.32289600e-01 -5.63904285e-01 8.53083193e-01 2.70847917e-01
-2.17574000e-01 -6.90285802e-01 -1.02680123e+00 -8.79292786e-02
-9.31822777e-01 1.37558118e-01 1.04207945e+00 6.48348510e-01
-1.17831981e+00 6.74472153e-01 -6.24778032e-01 -8.92403781e-01
6.54247224e-01 1.25497147e-01 -6.18888319e-01 -1.66639447e-01
-1.01409352e+00 1.14416397e+00 7.49932289e-01 6.08939826e-01
-7.99895227e-01 -7.12387145e-01 -6.18942261e-01 -4.51600179e-02
-1.48775682e-01 -3.82822871e-01 1.04381609e+00 -1.39068663e+00
-1.00538349e+00 1.00918746e+00 3.47016424e-01 -6.11279130e-01
4.81999427e-01 -6.84697703e-02 -4.36962038e-01 -1.42716989e-01
-1.21860415e-01 8.47624540e-01 4.09208059e-01 -1.17637038e+00
-7.00874329e-01 -3.33564103e-01 3.18539172e-01 2.13088337e-02
-1.57480732e-01 4.82150912e-02 2.50784665e-01 -2.54874647e-01
1.57455117e-01 -7.78755486e-01 -4.06929255e-01 2.96675235e-01
-4.32166420e-02 2.89667219e-01 9.51747715e-01 -5.53402781e-01
5.00025928e-01 -2.22686553e+00 -1.07155390e-01 1.37244374e-01
5.48263639e-02 3.68999690e-01 -3.23208243e-01 3.78142178e-01
-4.10681099e-01 1.88458264e-01 -5.11381745e-01 1.36370270e-03
-2.54775405e-01 4.96764034e-01 -4.80838418e-01 7.25345075e-01
4.06532019e-01 9.83898342e-01 -6.80615723e-01 -1.30778579e-02
7.24030495e-01 6.96038485e-01 -1.48035139e-01 5.94267026e-02
-1.05028212e-01 5.86316407e-01 -1.50208846e-01 7.07905829e-01
8.34847033e-01 -2.63993710e-01 -1.26159051e-03 -3.86826962e-01
-4.61310118e-01 1.53864324e-01 -9.56787467e-01 1.56953871e+00
-5.52315414e-01 1.02212262e+00 1.77491367e-01 -1.18916297e+00
9.15950894e-01 1.01767853e-01 4.14081395e-01 -8.96092474e-01
1.83655366e-01 4.10088688e-01 1.53179511e-01 -3.46904278e-01
5.52625954e-01 -3.49537104e-01 3.44769835e-01 2.31552050e-01
5.96466139e-02 -2.21367955e-01 -6.57363385e-02 -3.07435751e-01
5.99276960e-01 3.66267562e-01 1.82553858e-01 -3.90844852e-01
3.69835258e-01 4.41502839e-01 3.25760618e-03 7.41093457e-01
-1.22060962e-01 6.74347043e-01 -1.06718935e-01 -1.04561412e+00
-1.22094750e+00 -5.83443880e-01 -6.78675234e-01 1.06840432e+00
-3.05176526e-01 8.83529857e-02 -2.61634350e-01 -1.59844890e-01
1.18653759e-01 2.15303659e-01 -5.81402957e-01 3.25438023e-01
-2.33649224e-01 -1.45653677e+00 7.69169211e-01 3.99164677e-01
1.03715587e+00 -1.16801786e+00 -1.03861356e+00 2.82674700e-01
5.58946319e-02 -1.01444113e+00 8.69764030e-01 6.45835996e-01
-1.03680646e+00 -1.27431774e+00 -4.85461265e-01 -2.49526650e-01
5.73188253e-02 6.43745542e-01 1.41942298e+00 1.33930713e-01
-3.40021998e-01 2.55940884e-01 -4.95318860e-01 -6.62488461e-01
-3.81241649e-01 3.69172961e-01 -3.45214963e-01 -1.25963449e-01
4.55263942e-01 -5.46856523e-01 -6.43248856e-01 -9.43599790e-02
-1.30232179e+00 1.51447356e-01 8.67900133e-01 7.12389112e-01
3.53390366e-01 1.70359835e-02 2.24734887e-01 -6.10638678e-01
1.36681022e-02 -5.20079911e-01 -7.43048728e-01 1.84760913e-01
-7.47098386e-01 -2.38296643e-01 3.26668352e-01 2.41601527e-01
-8.86820734e-01 2.17616484e-01 -3.04960191e-01 5.53842708e-02
-7.66509354e-01 1.09939051e+00 1.24617726e-01 -3.93442661e-01
1.04782486e+00 3.22319359e-01 -4.90580015e-02 -4.45695639e-01
4.17372644e-01 8.89095843e-01 4.17067736e-01 -3.86860788e-01
8.25417697e-01 7.73001790e-01 1.36147425e-01 -1.18863726e+00
-7.64638960e-01 -5.74010432e-01 -8.64199638e-01 -1.47237405e-01
9.42648709e-01 -1.26923561e+00 -2.23401248e-01 8.32519650e-01
-9.86060560e-01 -6.55703127e-01 -2.77853608e-01 3.55428338e-01
-1.70686632e-01 8.93847719e-02 -1.62798464e-01 -8.11458111e-01
-3.78746390e-01 -8.72054100e-01 1.05073631e+00 1.23881385e-01
3.85246485e-01 -8.69712114e-01 1.32700741e-01 2.12024152e-01
1.00769603e+00 5.36813736e-01 6.03580952e-01 -2.65108436e-01
-7.21157730e-01 -2.64079630e-01 -7.77803361e-01 7.16328979e-01
2.17928767e-01 1.35851309e-01 -1.52822781e+00 -2.64280558e-01
-2.96432942e-01 -7.09351957e-01 1.38315403e+00 2.24774286e-01
1.09049165e+00 1.80872642e-02 -2.18111891e-02 1.08190370e+00
2.02684712e+00 -1.85263917e-01 8.28579426e-01 8.73007417e-01
6.91041827e-01 5.76882064e-01 2.54152745e-01 1.95024520e-01
3.01811725e-01 2.25715086e-01 8.42210829e-01 -7.49803424e-01
-1.65770983e-03 3.87184232e-01 -9.77451876e-02 2.26505235e-01
-6.09271288e-01 -6.71359301e-02 -1.27119279e+00 5.22323132e-01
-1.63083541e+00 -1.03313053e+00 -3.88441205e-01 1.85878360e+00
5.61502576e-01 -3.07977498e-01 -2.75401473e-01 3.75606775e-01
2.93192416e-01 6.05839789e-01 -3.77336442e-01 6.27071336e-02
-6.86824799e-01 5.51936984e-01 1.10972738e+00 1.69957563e-01
-1.55115414e+00 8.46235156e-01 6.62013674e+00 3.58887762e-01
-1.71243548e+00 1.52738497e-01 7.93508410e-01 2.08915532e-01
9.52741038e-03 -5.41009679e-02 -5.91416359e-01 -1.55852467e-01
1.26588309e+00 2.86949486e-01 4.97704566e-01 7.91130781e-01
2.68545657e-01 -3.19710463e-01 -8.72762382e-01 1.00573289e+00
-2.26857483e-01 -1.68513191e+00 6.58231555e-04 1.46852836e-01
7.15257943e-01 1.07772195e+00 -1.25079036e-01 3.17334421e-02
3.09738219e-01 -1.17185223e+00 8.04422140e-01 3.84580135e-01
9.80389118e-01 -3.43678743e-01 9.60756838e-01 1.01862371e-01
-1.20062947e+00 -2.77612805e-01 -6.20210469e-01 -4.42309588e-01
-3.04604769e-01 6.09936297e-01 -3.62174720e-01 7.53375649e-01
1.18815267e+00 9.52440023e-01 -8.73602808e-01 9.77877378e-01
-6.27285661e-03 7.26692915e-01 -4.26265866e-01 4.09551561e-01
5.96337199e-01 -2.94410974e-01 -2.32782643e-02 1.39603889e+00
3.41139376e-01 1.18344640e-02 -3.15164849e-02 7.16161549e-01
8.72175768e-02 -1.84911177e-01 -8.92562747e-01 -1.13356732e-01
1.43723339e-01 1.71624231e+00 -6.29403114e-01 -1.22640468e-01
-5.88110924e-01 5.25829494e-01 -3.41940261e-02 3.71172011e-01
-6.04966760e-01 -2.65258588e-02 7.03477740e-01 -3.22533585e-02
1.86375082e-01 -5.63562870e-01 -4.19224262e-01 -1.01151752e+00
-2.14655876e-01 -7.91800141e-01 2.66774863e-01 -8.84539187e-01
-1.21774614e+00 7.09608138e-01 -8.79967809e-02 -1.12898195e+00
1.95208102e-01 -1.02398956e+00 -5.70759535e-01 1.11081409e+00
-2.54377556e+00 -1.67627847e+00 -9.54974651e-01 6.42720819e-01
7.92489424e-02 -1.33598492e-01 1.16568983e+00 2.76970267e-01
-3.87883693e-01 -1.58388928e-01 4.59024847e-01 4.78160799e-01
3.52471530e-01 -1.06004906e+00 3.96588266e-01 9.70656216e-01
2.51826972e-01 2.55977482e-01 3.14594537e-01 -5.86421266e-02
-1.36825526e+00 -1.51443493e+00 5.71569145e-01 -2.16415152e-01
8.06929708e-01 -7.79488236e-02 -8.07600737e-01 7.26536214e-01
2.80996799e-01 3.02961588e-01 8.06639135e-01 -9.32995081e-02
-5.60558379e-01 -5.77982008e-01 -9.65542138e-01 7.31381327e-02
7.44128227e-01 -7.33030260e-01 -2.61135638e-01 3.48975450e-01
2.58061290e-01 -1.19428173e-01 -9.12137091e-01 4.84918118e-01
5.37554860e-01 -1.26323891e+00 1.06555831e+00 -5.98030746e-01
6.29585207e-01 -3.94539386e-01 -7.76845276e-01 -1.23885345e+00
-4.90581989e-01 2.84371823e-01 7.78301537e-01 8.19329262e-01
3.67177665e-01 -6.42622292e-01 6.09448671e-01 1.52633443e-01
-2.48939976e-01 -4.00608294e-02 -1.07143509e+00 -6.35861874e-01
3.59643579e-01 -6.18643403e-01 7.20980465e-01 1.09434116e+00
-9.81498897e-01 1.54765561e-01 -1.21905640e-01 3.54034990e-01
4.77721244e-01 3.20214629e-01 8.74969304e-01 -1.65591621e+00
-1.31104231e-01 -6.01888180e-01 -4.67037767e-01 -2.96859920e-01
-2.48977505e-02 -9.15686369e-01 1.30353361e-01 -1.71990526e+00
1.85217306e-01 -7.38958657e-01 -4.77933377e-01 9.31129575e-01
1.25240579e-01 7.47435927e-01 6.38384148e-02 3.46512228e-01
-2.39588693e-01 3.02938968e-01 7.39051938e-01 -3.82479966e-01
2.27598846e-01 -3.55814546e-01 -5.64541221e-01 5.16718030e-01
1.05123007e+00 -3.13097596e-01 7.06213415e-02 -7.49706089e-01
5.72775841e-01 -5.54033577e-01 9.77882564e-01 -1.43024850e+00
-9.20338333e-02 -3.54694009e-01 5.65268993e-01 -5.55792153e-01
4.41048294e-02 -1.12670541e+00 4.19938862e-01 5.28369486e-01
-5.50778676e-03 -6.01888150e-02 4.34577793e-01 1.68479890e-01
-3.76465023e-01 2.84243673e-01 9.18253660e-01 -5.13096809e-01
-1.31108785e+00 4.54734296e-01 -3.33305359e-01 -5.38424134e-01
5.81815839e-01 -3.11918169e-01 -6.49492562e-01 2.86718248e-03
-4.36872333e-01 -2.56893963e-01 3.50536764e-01 3.68617743e-01
3.44418526e-01 -1.06778502e+00 -1.08201754e+00 3.13470989e-01
5.04654884e-01 2.33123958e-01 1.17257103e-01 8.11898232e-01
-9.65532422e-01 4.81993824e-01 -6.14450395e-01 -8.47176254e-01
-9.55903411e-01 7.32524693e-02 8.87309372e-01 7.93040693e-02
-4.14139777e-01 7.05558777e-01 -3.28792423e-01 -8.05569530e-01
-2.63280034e-01 -3.48283738e-01 -5.19873090e-02 1.87992454e-01
4.39620286e-01 1.88914418e-01 5.53109586e-01 -8.42112303e-01
-3.92587662e-01 3.29144567e-01 5.86969137e-01 1.14675999e-01
1.93902183e+00 1.27984419e-01 -5.14637470e-01 4.78265494e-01
9.33912337e-01 -6.53032899e-01 -1.23762405e+00 -2.50115126e-01
2.60572899e-02 -3.40400577e-01 5.48436046e-01 -9.47560370e-01
-1.35686982e+00 1.15734887e+00 1.12272763e+00 2.93534905e-01
1.20906973e+00 -1.45377934e-01 1.44632161e-01 9.17268515e-01
3.42353374e-01 -7.78328478e-01 -4.93124157e-01 7.60258853e-01
9.72371280e-01 -1.62730944e+00 2.26024866e-01 5.02789430e-02
-1.32648498e-01 1.31184363e+00 3.34698021e-01 3.93393151e-02
7.23088324e-01 1.04444936e-01 5.68304002e-01 -3.81617814e-01
-2.04197332e-01 -6.11881614e-01 1.25358894e-01 7.21589148e-01
7.09336638e-01 1.92865282e-01 1.64257601e-01 -5.79077862e-02
-6.08226620e-02 3.45978796e-01 2.82641679e-01 1.02230310e+00
-6.41756952e-01 -1.03886747e+00 -5.49941838e-01 5.05978167e-01
-3.53997856e-01 -5.20643651e-01 -3.11284155e-01 8.73482287e-01
4.49392557e-01 8.62149656e-01 1.93298146e-01 -3.77703637e-01
2.35118002e-01 -1.19256098e-02 1.60002246e-01 -4.51700360e-01
-8.77739906e-01 -2.17473984e-01 7.67342150e-02 -5.91602266e-01
-1.17084622e+00 -5.66956580e-01 -5.71778953e-01 -5.79531014e-01
-2.51829386e-01 -2.68716246e-01 1.05525291e+00 9.83592153e-01
2.31084079e-01 3.94944638e-01 3.93028498e-01 -1.33461869e+00
-4.31660622e-01 -1.37477732e+00 -6.29821718e-01 2.79700696e-01
4.95062560e-01 -3.36193025e-01 -2.02595428e-01 1.30441999e-02] | [9.620850563049316, -1.5550668239593506] |
8e29a6a9-bcf6-4ec3-8734-7026aaa29eb7 | defending-against-insertion-based-textual | 2305.02394 | null | https://arxiv.org/abs/2305.02394v1 | https://arxiv.org/pdf/2305.02394v1.pdf | Defending against Insertion-based Textual Backdoor Attacks via Attribution | Textual backdoor attack, as a novel attack model, has been shown to be effective in adding a backdoor to the model during training. Defending against such backdoor attacks has become urgent and important. In this paper, we propose AttDef, an efficient attribution-based pipeline to defend against two insertion-based poisoning attacks, BadNL and InSent. Specifically, we regard the tokens with larger attribution scores as potential triggers since larger attribution words contribute more to the false prediction results and therefore are more likely to be poison triggers. Additionally, we further utilize an external pre-trained language model to distinguish whether input is poisoned or not. We show that our proposed method can generalize sufficiently well in two common attack scenarios (poisoning training data and testing data), which consistently improves previous methods. For instance, AttDef can successfully mitigate both attacks with an average accuracy of 79.97% (56.59% up) and 48.34% (3.99% up) under pre-training and post-training attack defense respectively, achieving the new state-of-the-art performance on prediction recovery over four benchmark datasets. | ['V. G. Vinod Vydiswaran', 'Chaowei Xiao', 'Wei Ping', 'Zhuofeng Wu', 'Jiazhao Li'] | 2023-05-03 | null | null | null | null | ['backdoor-attack'] | ['adversarial'] | [-1.64675545e-02 -3.66434932e-01 -3.07083488e-01 9.83945876e-02
-8.85159433e-01 -1.21346033e+00 6.18269622e-01 5.23420572e-01
-3.45100641e-01 5.89642346e-01 -2.89123170e-02 -7.37082958e-01
3.91667873e-01 -1.04846430e+00 -9.62503254e-01 -5.62724710e-01
-9.23091248e-02 3.38732064e-01 4.54839587e-01 -2.90476084e-01
4.11383569e-01 5.65348387e-01 -7.30774403e-01 7.02510238e-01
4.48060244e-01 5.12328386e-01 -5.81586421e-01 5.16038835e-01
2.13917971e-01 9.07263637e-01 -1.21283948e+00 -1.09524620e+00
4.52185363e-01 -1.39060272e-02 -7.37397969e-01 -7.90419936e-01
3.28034282e-01 -5.57954967e-01 -6.43289387e-01 1.02529383e+00
6.92635238e-01 -2.85961509e-01 4.78225797e-01 -1.40459883e+00
-4.60331202e-01 1.11234903e+00 -7.28079915e-01 4.37168062e-01
3.69023234e-01 5.94964981e-01 8.86110604e-01 -8.07817996e-01
1.62356898e-01 1.15123606e+00 6.12046659e-01 7.74751365e-01
-9.99409497e-01 -1.38838887e+00 1.50450738e-02 2.75735676e-01
-1.33064020e+00 -1.56832233e-01 4.99399096e-01 -2.30717719e-01
1.13891757e+00 5.75928986e-01 -5.05114719e-02 1.80829787e+00
3.15094829e-01 8.44184935e-01 8.95903051e-01 -2.46073261e-01
3.00379507e-02 2.58657277e-01 3.21559936e-01 2.75209635e-01
5.43752372e-01 3.99472833e-01 -5.12439311e-01 -8.67948949e-01
9.33697298e-02 2.41257310e-01 -1.67739704e-01 5.61653376e-01
-6.36433840e-01 1.14170861e+00 4.63047713e-01 -3.93035375e-02
-3.78223270e-01 1.76822409e-01 5.89636505e-01 -2.39095856e-02
2.50855535e-01 6.88771486e-01 -5.49488544e-01 2.12721959e-01
-6.66994452e-01 3.51949364e-01 7.63964117e-01 5.52700102e-01
2.72310436e-01 2.74282455e-01 -6.40157402e-01 4.28158104e-01
1.84894145e-01 6.10320270e-01 1.78864881e-01 -1.52323529e-01
8.96451473e-01 3.18500012e-01 1.17108971e-02 -9.28495467e-01
-9.23791230e-02 -9.70874190e-01 -5.02894998e-01 -3.42170477e-01
3.63956243e-01 -1.85344920e-01 -7.02972531e-01 1.89563310e+00
3.56067449e-01 5.75931609e-01 1.11549750e-01 4.94463563e-01
2.49136552e-01 6.02122545e-01 5.19732773e-01 6.52471185e-02
1.48468828e+00 -7.21520185e-01 -4.05878514e-01 -2.50619024e-01
8.51136267e-01 -6.15097523e-01 1.13411498e+00 7.02141464e-01
-7.34145224e-01 5.06885052e-02 -1.11837447e+00 5.49425662e-01
-6.55494988e-01 -4.88261402e-01 5.39708853e-01 1.11105001e+00
-1.86016902e-01 5.62945664e-01 -5.44705808e-01 1.52782410e-01
6.34690225e-01 1.59457251e-01 -2.82323211e-01 -1.37032881e-01
-1.87970054e+00 9.28217173e-01 3.75380784e-01 -4.46121752e-01
-1.51245379e+00 -9.86510038e-01 -2.99061835e-01 4.35046293e-02
4.53009456e-01 -2.22848013e-01 1.11155391e+00 1.81362495e-01
-9.39462304e-01 5.55742741e-01 2.33149067e-01 -9.04085338e-01
6.45177543e-01 -6.62635624e-01 -5.43537617e-01 2.24782024e-02
-3.18123889e-03 1.25400633e-01 9.31735992e-01 -1.04343712e+00
-3.82929385e-01 -2.97883093e-01 1.52774245e-01 -2.95685738e-01
-1.12960720e+00 6.42935693e-01 -8.28559473e-02 -9.82454658e-01
-5.53797483e-01 -7.20798314e-01 -1.90769643e-01 -5.84329128e-01
-1.10856354e+00 -1.59041464e-01 1.03575599e+00 -6.65480971e-01
1.68962801e+00 -1.76410282e+00 -3.15869927e-01 1.95007131e-01
1.74425423e-01 8.52083623e-01 3.22341500e-03 7.71954954e-01
-3.06697398e-01 5.03156960e-01 -2.85113603e-01 -2.71806985e-01
4.11853120e-02 5.40896552e-03 -1.36732841e+00 4.68776643e-01
3.06255639e-01 7.54709840e-01 -8.38572621e-01 1.58945646e-03
2.12648548e-02 3.27688932e-01 -6.89474046e-01 3.77863914e-01
-1.90212235e-01 2.39594534e-01 -3.70749205e-01 8.58264625e-01
8.33671331e-01 8.42869356e-02 -1.10511236e-01 -3.75553370e-02
2.31003255e-01 5.60012341e-01 -8.34528387e-01 8.74425173e-01
-2.70244300e-01 9.33611020e-02 -4.27533418e-01 -4.00610119e-01
7.17606306e-01 3.47181886e-01 -6.45990437e-03 -4.06679809e-01
2.74165303e-01 7.99003094e-02 2.11110860e-02 -8.81590098e-02
2.81019866e-01 -1.45676494e-01 -3.84077579e-01 7.74148405e-01
-1.93234041e-01 4.93384093e-01 3.79355787e-03 6.72849417e-01
1.53743219e+00 -5.53619504e-01 1.73650473e-01 7.31767938e-02
5.33595085e-01 -9.84567553e-02 5.38351893e-01 1.05998302e+00
-1.01388559e-01 1.37800932e-01 5.61348677e-01 -3.85269910e-01
-7.18849242e-01 -1.02164721e+00 1.25576584e-02 1.04377079e+00
-1.07335851e-01 -7.65715837e-01 -8.79849255e-01 -1.22732568e+00
2.21490413e-01 1.18821430e+00 -4.91831571e-01 -8.09596956e-01
-6.71380341e-01 -1.04686677e+00 1.73840129e+00 6.43109500e-01
4.55973715e-01 -8.56865346e-01 -4.69264060e-01 2.48213127e-01
-1.82570726e-01 -1.14972532e+00 -4.59908843e-01 4.00188357e-01
-4.37844008e-01 -1.20988441e+00 -1.64824635e-01 4.99204844e-02
3.69914889e-01 1.98085546e-01 8.32340539e-01 4.00156409e-01
-4.58141267e-01 -2.84401238e-01 -4.18077618e-01 -5.22294343e-01
-5.57999671e-01 5.44390455e-03 5.08767247e-01 -1.20834611e-01
4.34466422e-01 -3.74354541e-01 -2.78999984e-01 5.47177434e-01
-1.12368298e+00 -4.45118397e-01 4.37170863e-01 7.10484803e-01
2.56565452e-01 6.87446669e-02 6.19100034e-01 -1.23735774e+00
7.04004943e-01 -9.78542805e-01 -3.75176966e-01 2.33179644e-01
-3.93639296e-01 -1.09685548e-01 1.20733082e+00 -7.54448771e-01
-4.97996390e-01 -2.75425643e-01 -5.31922698e-01 -8.21604371e-01
-1.63576022e-01 3.18384767e-01 -4.40500975e-01 -4.89902347e-02
1.08126891e+00 2.33747646e-01 -7.02651858e-01 -5.40329933e-01
2.96191007e-01 5.74537575e-01 5.46201706e-01 -7.69166768e-01
1.45667970e+00 2.02510566e-01 -7.79714957e-02 -3.71342093e-01
-9.14531052e-01 -3.79326880e-01 -1.74127236e-01 1.46206066e-01
4.28686947e-01 -6.74932063e-01 -7.81964540e-01 7.08627999e-01
-1.21970475e+00 -2.04267621e-01 3.14224720e-01 -4.83772764e-03
3.09996545e-01 7.75614023e-01 -8.95244539e-01 -9.28468466e-01
-8.40920389e-01 -1.16109931e+00 5.09935498e-01 -2.09962606e-01
-1.11054614e-01 -8.69805038e-01 -4.28956561e-03 2.98902214e-01
3.81632119e-01 4.40848738e-01 9.46982265e-01 -1.41603398e+00
-1.01015382e-01 -7.00858057e-01 -1.15497056e-02 2.86952049e-01
-1.90946147e-01 1.55672282e-02 -1.28832567e+00 -5.07206559e-01
5.75280339e-02 -5.13031244e-01 8.51257443e-01 -5.21291852e-01
1.40108943e+00 -6.47263169e-01 -4.86523956e-01 5.57555199e-01
1.19026303e+00 -1.62437148e-02 8.66487205e-01 4.04169500e-01
7.33824372e-01 3.59532267e-01 5.96883953e-01 7.02242315e-01
-1.77148774e-01 9.47121680e-01 8.11689973e-01 1.74351022e-01
1.95762768e-01 -8.53287339e-01 8.06274652e-01 8.27603415e-02
6.06698394e-01 -7.15885937e-01 -1.11973870e+00 8.79520103e-02
-1.37722576e+00 -1.14294171e+00 -4.75046068e-01 2.34526587e+00
9.58451211e-01 4.40075636e-01 1.49462432e-01 5.61532795e-01
6.56575859e-01 6.92417398e-02 -5.73547125e-01 -4.05174613e-01
-1.81562975e-01 5.04407406e-01 9.38642085e-01 3.66947562e-01
-1.43918502e+00 1.33069646e+00 6.16748905e+00 1.25049722e+00
-9.50590134e-01 1.99374273e-01 5.15032709e-01 -1.31943271e-01
-3.07249017e-02 -3.13061401e-02 -1.27160347e+00 6.93111360e-01
1.28947699e+00 1.01710334e-01 2.77417213e-01 7.57318676e-01
-2.41165891e-01 5.55622339e-01 -9.13159072e-01 6.03795469e-01
1.60748318e-01 -1.17417359e+00 5.18433154e-01 1.46574691e-01
2.84126759e-01 -2.07224444e-01 5.67565382e-01 5.48022330e-01
6.92337394e-01 -1.17713726e+00 8.75700891e-01 -1.82880282e-01
6.66218996e-01 -1.01192832e+00 6.68561101e-01 7.58825064e-01
-9.03773606e-01 -3.62149954e-01 -4.25506979e-01 1.22730993e-01
2.97876894e-02 5.07759213e-01 -1.03177774e+00 5.61826766e-01
6.27852023e-01 4.11247224e-01 -7.05801725e-01 8.09186220e-01
-7.76332021e-01 1.34435320e+00 -4.58954394e-01 1.38367742e-01
2.04106554e-01 6.53476715e-01 7.38646150e-01 1.33212399e+00
-1.36965379e-01 1.01519696e-01 3.47320765e-01 7.64568746e-01
-3.12922448e-01 -1.85687631e-01 -4.76022154e-01 -1.17533624e-01
8.81293535e-01 1.09137762e+00 -2.73059160e-01 -2.57067174e-01
7.24825114e-02 8.38235378e-01 3.41681480e-01 -5.78117371e-02
-1.45433652e+00 -3.50031137e-01 8.13024521e-01 2.12740507e-02
-5.65071069e-02 9.29488838e-02 -2.07997620e-01 -1.20141792e+00
-3.18970859e-01 -1.17279696e+00 8.02293181e-01 -2.77938753e-01
-1.55573761e+00 5.18250704e-01 4.53880168e-02 -1.07187152e+00
6.83223456e-02 -6.67219222e-01 -1.09226167e+00 8.94746900e-01
-1.37827134e+00 -1.17472684e+00 2.02690467e-01 6.57952726e-01
3.33267272e-01 -3.94637585e-01 9.00513291e-01 5.06065965e-01
-1.19817114e+00 1.47261882e+00 -3.73613060e-01 4.50384021e-01
9.00431812e-01 -9.15558815e-01 8.08122754e-01 1.38580322e+00
2.65917987e-01 1.02296698e+00 7.80787468e-01 -1.02774906e+00
-1.49869919e+00 -1.42128289e+00 6.31202757e-01 -1.04266119e+00
8.93738449e-01 -6.27610505e-01 -1.12916589e+00 5.14124036e-01
-1.55941769e-01 -1.92018434e-01 1.08683538e+00 -2.02089190e-01
-1.31244302e+00 3.87096554e-01 -1.39991927e+00 6.19090557e-01
7.99057484e-01 -4.44986522e-01 -3.92565131e-01 5.43379426e-01
9.80309129e-01 -3.44727427e-01 -5.88931680e-01 2.82466650e-01
2.30076686e-01 -6.89217925e-01 1.32208753e+00 -1.08896959e+00
3.71567577e-01 -2.17992976e-01 -3.71427923e-01 -8.89318228e-01
-3.81230474e-01 -6.70137346e-01 -6.25499666e-01 1.28556156e+00
5.33248365e-01 -9.49779272e-01 7.23441958e-01 1.98584110e-01
-2.64302865e-02 -8.06824923e-01 -6.87801838e-01 -9.19052482e-01
4.61397588e-01 -6.71406865e-01 8.60453665e-01 1.21383572e+00
-3.07014193e-02 4.54950221e-02 -7.45297790e-01 6.84741795e-01
9.00417268e-01 -4.28280741e-01 9.08441961e-01 -9.30601656e-01
-5.47681689e-01 -1.30301669e-01 -1.14669107e-01 -6.34078383e-01
1.44972056e-01 -1.03641999e+00 -5.36468089e-01 -7.64985442e-01
2.77833939e-01 -3.93067330e-01 -7.09041655e-01 1.33991659e+00
-5.27870893e-01 5.74852049e-01 2.13628843e-01 3.87475222e-01
-2.74020910e-01 2.57863879e-01 3.89870167e-01 -2.18538404e-01
2.50864863e-01 8.68906751e-02 -9.24082518e-01 4.65182096e-01
1.02618206e+00 -1.05078292e+00 -8.88311416e-02 -3.01784724e-01
1.74090177e-01 -3.48484159e-01 4.06843692e-01 -7.96142578e-01
1.70044348e-01 -3.90119217e-02 3.21263075e-01 -6.76147223e-01
1.78293750e-01 -5.11862218e-01 1.28017263e-02 1.01481569e+00
-4.10832942e-01 2.38448158e-01 5.09448767e-01 6.83040857e-01
2.34604403e-01 -1.73515260e-01 7.72856712e-01 1.89044252e-01
-1.69507250e-01 4.99393702e-01 -2.60836869e-01 1.38045549e-01
1.22700906e+00 1.73484057e-01 -9.28037226e-01 -4.80870157e-02
-1.79154038e-01 1.48971051e-01 2.73014158e-01 4.59970444e-01
8.31986487e-01 -1.02202058e+00 -8.44681144e-01 2.09718525e-01
2.29729071e-01 -7.72585332e-01 8.06814432e-02 6.08435333e-01
-2.16850743e-01 3.99922937e-01 3.93289030e-02 -9.90107358e-02
-1.41794491e+00 9.19338703e-01 1.27261460e-01 -6.66846156e-01
-4.29268181e-01 1.08128846e+00 -7.15933368e-02 -3.14507842e-01
2.58255094e-01 3.08587819e-01 2.17178967e-02 -3.45536590e-01
8.52348208e-01 5.51533580e-01 8.78723860e-02 -5.35668612e-01
-6.44792318e-01 -4.64300029e-02 -6.12584949e-01 1.87913820e-01
9.68933880e-01 9.61754799e-01 -1.43837988e-01 -2.49770522e-01
9.74269390e-01 4.13605332e-01 -7.17635393e-01 -1.27104029e-01
-5.30895442e-02 -7.28044152e-01 -1.39125913e-01 -1.35216308e+00
-8.88935387e-01 1.18877804e+00 2.32365072e-01 2.15283558e-01
6.88043773e-01 -2.75961667e-01 1.18144953e+00 2.72389740e-01
4.98008460e-01 -3.10527265e-01 2.83018976e-01 5.61111569e-01
5.87349176e-01 -6.84916377e-01 2.26708502e-02 -5.81858754e-01
-5.62070072e-01 7.48463809e-01 7.64476240e-01 -7.81116784e-02
1.81367204e-01 3.92635226e-01 -1.95148587e-01 1.19180605e-01
-1.01660430e+00 5.61317623e-01 -2.33808253e-02 4.01186049e-01
4.07327525e-02 1.59881130e-01 2.99842283e-02 1.04458070e+00
-1.69746622e-01 -8.01632285e-01 3.49604577e-01 8.13452303e-01
-3.43950242e-01 -1.46905363e+00 -7.63926685e-01 3.06071758e-01
-9.87505734e-01 -3.92254025e-01 -8.14476848e-01 3.83838028e-01
-7.07082301e-02 1.23117244e+00 -5.86490035e-01 -1.02478790e+00
3.44183981e-01 1.15644664e-01 2.21686568e-02 -6.69347346e-01
-1.41202033e+00 -3.93581063e-01 -7.99409971e-02 -4.75477695e-01
7.30238855e-01 -1.94714755e-01 -1.17970514e+00 -6.47571921e-01
-7.16733336e-01 3.14722598e-01 4.54879582e-01 6.03772938e-01
4.41224188e-01 4.06449378e-01 8.60484838e-01 -2.44177327e-01
-1.24786508e+00 -8.20913792e-01 -1.31423160e-01 4.29787129e-01
-1.91864252e-01 -8.69306266e-01 -6.60266101e-01 -6.11733675e-01] | [5.986425876617432, 7.832669258117676] |
3049fd6d-2d48-452b-a3df-0533cdf9d5a5 | in-game-toxic-language-detection-shared-task | 2211.05995 | null | https://arxiv.org/abs/2211.05995v3 | https://arxiv.org/pdf/2211.05995v3.pdf | In-game Toxic Language Detection: Shared Task and Attention Residuals | In-game toxic language becomes the hot potato in the gaming industry and community. There have been several online game toxicity analysis frameworks and models proposed. However, it is still challenging to detect toxicity due to the nature of in-game chat, which has extremely short length. In this paper, we describe how the in-game toxic language shared task has been established using the real-world in-game chat data. In addition, we propose and introduce the model/framework for toxic language token tagging (slot filling) from the in-game chat. The data and code will be released. | ['Soyeon Caren Han', 'Feiqi Cao', 'Weixuan Wu', 'Yuanzhe Jia'] | 2022-11-11 | null | null | null | null | ['slot-filling'] | ['natural-language-processing'] | [-2.69648671e-01 -3.16340566e-01 3.64617705e-02 2.95226544e-01
-1.10568249e+00 -9.43785429e-01 -1.34952934e-02 3.69761176e-02
-4.63239551e-01 8.83484900e-01 3.99540335e-01 -3.24050456e-01
-1.13497593e-01 -7.60474920e-01 -5.97415725e-03 -3.67998958e-01
-1.98332980e-01 1.79528490e-01 6.95377886e-01 -4.38150913e-01
1.99779972e-01 -5.17099490e-03 -1.09768498e+00 5.61446011e-01
4.64792311e-01 2.76183397e-01 4.23643440e-01 8.96175385e-01
-3.12767893e-01 1.57418406e+00 -1.04983222e+00 -7.94233859e-01
2.10876048e-01 -7.37086117e-01 -1.17221010e+00 -3.82697046e-01
-5.81403077e-01 -1.80558577e-01 -3.37266654e-01 1.14429843e+00
8.92346680e-01 1.63156420e-01 2.09102958e-01 -1.44420981e+00
-2.93366730e-01 9.69616592e-01 -2.61982352e-01 3.62601787e-01
9.34074342e-01 1.67554691e-01 1.07036340e+00 1.65647883e-02
8.50067317e-01 9.26234841e-01 8.43248010e-01 7.79848754e-01
-3.96627426e-01 -9.33198154e-01 -2.71539629e-01 2.10464567e-01
-1.42411935e+00 9.74129215e-02 5.17535627e-01 -7.57865191e-01
1.40867400e+00 3.60388100e-01 7.48516619e-01 1.37870634e+00
4.21340019e-01 6.75732553e-01 1.06067765e+00 -1.47691041e-01
3.76307636e-01 -1.11104801e-01 -2.91496009e-01 4.45641667e-01
-3.66771191e-01 -5.75678907e-02 -8.80202174e-01 -4.52618450e-01
3.73362213e-01 -3.52528632e-01 3.70539546e-01 6.36009336e-01
1.31966516e-01 1.18836379e+00 -3.10370147e-01 4.69734520e-01
-1.67365074e-01 2.25617915e-01 1.06849670e+00 3.28950197e-01
1.08564508e+00 3.37667793e-01 -5.30374408e-01 -1.30205560e+00
-5.31957507e-01 4.03257459e-01 1.10807717e+00 6.65766299e-01
1.48729965e-01 -1.45350263e-01 -4.45056707e-02 1.04336882e+00
4.82095838e-01 -2.77855992e-01 4.10853803e-01 -6.61179483e-01
2.90552765e-01 3.24312955e-01 -2.77953207e-01 -6.83263481e-01
-5.71429312e-01 -3.33062746e-02 -2.70326138e-02 1.39363229e-01
3.72364879e-01 -4.18976903e-01 -2.84619510e-01 1.61423433e+00
-2.54329909e-02 2.92060673e-01 -1.37715548e-01 5.32304764e-01
1.15803111e+00 3.60999197e-01 9.30159509e-01 -6.30738363e-02
1.52372050e+00 -5.69292665e-01 -7.86198556e-01 -2.13100702e-01
1.14672112e+00 -7.31113791e-01 1.19026721e+00 6.29546344e-01
-7.87639737e-01 1.65615693e-01 -7.57958949e-01 -1.20611787e-02
-6.85515404e-01 -8.09708178e-01 8.62294495e-01 1.48245895e+00
-8.62833560e-01 4.36568141e-01 -3.92477363e-01 -7.58928955e-01
3.81926477e-01 3.24833810e-01 -2.11609408e-01 9.81466323e-02
-1.72837675e+00 9.78562057e-01 4.18417484e-01 -6.60822749e-01
-9.52944160e-01 -6.13142371e-01 -7.66163528e-01 -3.48684996e-01
4.48056668e-01 -2.51690783e-02 1.53635991e+00 -2.95934111e-01
-1.84645295e+00 8.33746493e-01 6.54388428e-01 3.83649990e-02
3.86884093e-01 -1.28950819e-01 -2.68397838e-01 -4.70798165e-01
4.43008363e-01 2.29890868e-02 -3.31493109e-01 -5.93000948e-01
-6.19674563e-01 -1.62557632e-01 7.23029435e-01 3.25189918e-01
-2.17680931e-01 1.28377926e+00 -3.79417837e-01 -3.90835673e-01
-7.62315035e-01 -5.37703454e-01 -5.89850605e-01 -6.45288587e-01
-2.00760216e-01 -3.59083295e-01 3.74199808e-01 -7.91580081e-01
1.72331071e+00 -2.06885982e+00 -4.65689093e-01 -2.27356941e-01
3.78859192e-01 2.32165139e-02 -1.62751004e-01 1.05315614e+00
2.42361240e-02 5.60108364e-01 2.21884340e-01 -1.91752583e-01
2.99137652e-01 1.12309888e-01 9.28558856e-02 3.16707909e-01
-3.82399380e-01 6.64887369e-01 -1.28718305e+00 -4.93796676e-01
1.59397438e-01 1.98259279e-02 -5.09976387e-01 2.16765627e-01
-3.18564862e-01 1.96798041e-01 -6.49708807e-01 6.20208681e-01
5.64751625e-01 5.15653610e-01 1.33464307e-01 6.77286029e-01
-5.57273507e-01 7.31933594e-01 -6.87361538e-01 1.89437068e+00
-1.89099878e-01 2.44447887e-01 -9.17552933e-02 -4.28157717e-01
4.45320308e-01 6.10218167e-01 8.38484883e-01 -8.50684464e-01
4.72467005e-01 2.26531234e-02 3.39893848e-01 -8.17721546e-01
7.46266663e-01 -4.64052826e-01 -8.64819825e-01 7.50978231e-01
1.76927865e-01 -2.29712501e-01 6.12522781e-01 2.76215494e-01
1.85362756e+00 1.36989161e-01 5.08756340e-01 5.89466877e-02
1.38129473e-01 1.03358500e-01 7.39816904e-01 6.86762989e-01
-5.57668507e-01 3.65624756e-01 9.96820867e-01 7.48748332e-02
-6.13258719e-01 -6.23528540e-01 3.63623142e-01 1.64098084e+00
-3.77153337e-01 -1.36200404e+00 -1.18007910e+00 -6.03926063e-01
-6.29455566e-01 3.85571867e-01 -3.76410902e-01 -1.08779654e-01
1.59170672e-01 -9.37582254e-01 1.56316650e+00 2.15688765e-01
2.67140448e-01 -1.48254800e+00 -2.98961520e-01 8.34365606e-01
-3.28435600e-01 -9.14478481e-01 -3.32484722e-01 4.55412209e-01
-4.34475057e-02 -1.29795003e+00 5.73490076e-02 -5.08130670e-01
-6.05049789e-01 -3.35049666e-02 9.99794066e-01 -7.58927129e-03
-3.66777807e-01 -4.85590957e-02 -9.40097809e-01 -6.59777045e-01
-3.94989938e-01 1.28955632e-01 -1.29432291e-01 -7.94120073e-01
8.11231613e-01 -5.89070499e-01 1.02135293e-01 3.32788974e-02
-9.92095768e-01 -3.16276401e-01 -1.71291262e-01 2.18845621e-01
4.33411673e-02 3.38403136e-01 5.95079780e-01 -1.29739809e+00
1.39019728e+00 -1.00455117e+00 -2.33690515e-01 -1.21519081e-01
1.81934997e-01 -8.04811716e-01 4.45076764e-01 -1.41179860e-01
-6.53275907e-01 -2.95873469e-04 -1.02968514e+00 3.28873008e-01
-1.34229481e-01 1.20619559e+00 -5.41924417e-01 -1.00759156e-01
6.79552674e-01 -2.90787578e-01 -3.98964584e-01 -5.51133394e-01
1.99093908e-01 1.04083717e+00 -5.92830032e-02 -6.06159508e-01
2.33028859e-01 -1.82557032e-01 -6.05581105e-01 -8.27636778e-01
-4.76447374e-01 -7.52531946e-01 -2.76160359e-01 -7.29905903e-01
9.99605060e-01 -8.93517911e-01 -1.05304158e+00 8.38593185e-01
-1.15677142e+00 -7.11647451e-01 -1.85094088e-01 2.04781756e-01
-4.86088902e-01 5.09432912e-01 -1.29050016e+00 -1.19038379e+00
-7.93441609e-02 -1.13804007e+00 4.14653510e-01 1.31436035e-01
-6.50536478e-01 -9.91201282e-01 8.05827677e-01 3.71642381e-01
1.91625372e-01 2.54191041e-01 6.66009724e-01 -8.34350467e-01
7.10397214e-02 -1.89540386e-01 1.47385597e-01 -1.24126799e-01
2.48506680e-01 3.72333126e-03 -1.01261604e+00 2.04405993e-01
1.70083269e-01 -6.81424320e-01 1.19250342e-01 2.53754288e-01
8.46291125e-01 6.80513009e-02 1.53814450e-01 3.38645577e-02
1.22037613e+00 7.70859659e-01 1.06656265e+00 5.09233594e-01
3.84521991e-01 6.51162088e-01 7.17759728e-01 9.52872455e-01
4.18054909e-01 6.42505407e-01 2.62384772e-01 3.10080349e-01
2.07230732e-01 -5.02528906e-01 8.25166345e-01 1.10809004e+00
-5.37221245e-02 -6.73239112e-01 -9.07518744e-01 5.17055511e-01
-1.81902933e+00 -1.14076769e+00 -5.26005387e-01 1.57118845e+00
9.00304019e-01 1.44227520e-01 6.14978731e-01 5.11192679e-02
3.31965387e-01 1.18577972e-01 -4.88945059e-02 -1.10583448e+00
-7.32669383e-02 7.74987519e-01 6.27857983e-01 6.92959845e-01
-1.10472870e+00 1.46870911e+00 6.90992641e+00 1.59460807e+00
-5.83923459e-01 9.04269278e-01 3.57109129e-01 -6.11077622e-02
-3.63734178e-02 7.23230913e-02 -1.67923972e-01 7.04503238e-01
1.16941571e+00 -6.70734763e-01 4.85068828e-01 6.87185109e-01
7.08883643e-01 -4.12010312e-01 -5.88400483e-01 7.36681104e-01
-7.64482319e-02 -9.18093979e-01 -4.70228881e-01 3.26508462e-01
5.72214685e-02 4.03976887e-01 -1.71321839e-01 4.46014345e-01
1.10302591e+00 -1.16971564e+00 7.18109488e-01 -7.62672424e-02
1.01267755e+00 -8.05628359e-01 8.29515576e-01 5.86167276e-01
-1.24717128e+00 4.61364212e-03 -4.56357688e-01 -9.77782845e-01
9.96352509e-02 1.46306992e-01 -5.72532117e-01 6.07014954e-01
1.00682819e+00 7.18349755e-01 -4.97745246e-01 1.37961411e+00
-2.38641262e-01 7.71634638e-01 -1.20383359e-01 -4.73742992e-01
3.64517987e-01 -3.87040824e-01 1.42850682e-01 1.26891255e+00
2.65432894e-01 4.52317864e-01 4.07414496e-01 6.76859438e-01
-1.97358299e-02 3.90747845e-01 -9.77886200e-01 -5.31994760e-01
2.77548403e-01 1.11782193e+00 -8.18419874e-01 1.61107093e-01
-5.66339135e-01 1.11709774e+00 9.97265503e-02 -2.53343999e-01
-1.00460601e+00 -4.28926945e-01 1.12291062e+00 -1.23184115e-01
-4.37291473e-01 5.53886779e-02 -2.99480557e-01 -5.68654358e-01
-4.95203793e-01 -8.92327130e-01 5.59505343e-01 -8.18448961e-01
-1.61745334e+00 6.30946219e-01 -2.19239414e-01 -1.33904505e+00
-2.25415632e-01 -6.19505703e-01 -7.96870947e-01 8.54670584e-01
-3.15234184e-01 -9.34518695e-01 2.54753411e-01 4.57839638e-01
6.29615128e-01 -7.03984797e-02 1.15891314e+00 7.88371384e-01
-5.71246564e-01 6.39195979e-01 -6.57533556e-02 2.23521233e-01
5.71173847e-01 -9.64686632e-01 4.82359707e-01 4.85755891e-01
-1.47909135e-01 2.56257653e-01 6.21817827e-01 -9.99018610e-01
-8.09027791e-01 -9.22861159e-01 1.12480795e+00 -6.65111125e-01
1.24797249e+00 -9.06910121e-01 -4.38184768e-01 2.91939735e-01
4.40767109e-01 -7.22728908e-01 1.56147921e+00 1.63546309e-01
-3.46834511e-01 6.56441927e-01 -1.40874505e+00 7.21184552e-01
1.20558977e+00 -1.10694253e+00 -6.24708347e-02 6.39544189e-01
5.86167157e-01 -4.26114142e-01 -7.14871407e-01 -5.90975106e-01
2.91063100e-01 -6.48848116e-01 3.57876509e-01 -8.72975349e-01
5.86057663e-01 -1.85894325e-01 -7.29321316e-02 -1.23806047e+00
-4.39720333e-01 -8.04630578e-01 9.69069779e-01 1.54155505e+00
4.46522444e-01 -2.47172490e-01 1.04925573e+00 8.98616433e-01
-1.78575754e-01 -1.38819829e-01 -1.21458614e+00 -8.65832925e-01
3.73844922e-01 -1.38792574e+00 2.99262881e-01 1.02044177e+00
1.12954438e+00 4.37116057e-01 -6.78311408e-01 -2.49047786e-01
1.21801376e-01 -9.43131864e-01 3.68361920e-01 -9.81555045e-01
-4.48035300e-01 -1.46220535e-01 -7.62347281e-01 -4.67893541e-01
2.55799562e-01 -7.70083487e-01 2.15464026e-01 -1.44474840e+00
7.42232502e-01 -2.40073234e-01 -1.98250830e-01 7.64954209e-01
1.28097877e-01 5.29690921e-01 3.02677333e-01 -3.71390015e-01
-1.13097787e+00 2.24654198e-01 4.84345615e-01 -2.11729798e-02
-2.27769822e-01 -1.27256051e-01 -1.07071614e+00 7.47061312e-01
1.11013901e+00 -1.12934935e+00 -4.52787340e-01 -8.35022256e-02
6.87400579e-01 2.01570779e-01 -1.45175233e-01 -9.12966073e-01
3.75141114e-01 -3.96607727e-01 -6.30799234e-01 -1.20851673e-01
1.53468296e-01 -2.76382267e-01 5.72400630e-01 2.96242446e-01
-1.01732507e-01 1.92552000e-01 4.43570137e-01 2.35092193e-01
-8.64937752e-02 -6.03509665e-01 2.90594935e-01 -3.92070591e-01
-6.61088169e-01 2.74600595e-01 -1.58879852e+00 3.45183462e-01
9.09641623e-01 -4.03024465e-01 -2.44710341e-01 -7.51325846e-01
-6.03899956e-01 -2.81196623e-03 3.73923361e-01 5.18652976e-01
4.45419878e-01 -1.02206874e+00 -6.25657916e-01 -5.80313027e-01
1.84226006e-01 -8.08558345e-01 6.81040406e-01 4.25737798e-01
-8.20691884e-01 1.22885495e-01 -4.38095212e-01 4.40580487e-01
-1.52097833e+00 2.56194919e-01 3.54228884e-01 -6.35058284e-01
-5.99866249e-02 8.86493742e-01 1.25250325e-01 -6.11917734e-01
-3.52853276e-02 1.71996936e-01 -5.48219144e-01 -1.42562523e-01
6.32140279e-01 2.72163510e-01 2.34383479e-01 -6.66743219e-01
-5.90442359e-01 -1.49520025e-01 1.02074064e-01 -3.86474133e-01
1.35316741e+00 1.85610071e-01 -5.22445321e-01 5.23017883e-01
1.04627979e+00 3.45389634e-01 -7.20379055e-01 2.34372959e-01
4.51816171e-01 -3.77224922e-01 -2.79854506e-01 -1.03680468e+00
-8.85082066e-01 4.53915983e-01 4.17698503e-01 3.14600945e-01
8.17974031e-01 1.47084575e-02 7.32967377e-01 -1.21532932e-01
8.81877244e-01 -1.39397657e+00 -2.60227174e-01 1.13566589e+00
4.82108891e-01 -4.07224923e-01 -3.17571998e-01 -8.60064209e-01
-8.02004218e-01 4.11811441e-01 6.61369324e-01 1.08563468e-01
8.98978412e-01 9.01147068e-01 3.79244626e-01 -5.32890677e-01
-8.88260126e-01 -5.28601527e-01 -8.18447411e-01 1.05905044e+00
8.84945035e-01 2.02546731e-01 -9.67822015e-01 1.67937422e+00
-4.07167405e-01 2.23797351e-01 1.13536274e+00 9.88777399e-01
-1.92502007e-01 -1.66568768e+00 -1.10431790e-01 3.40533286e-01
-1.25896704e+00 -5.75536132e-01 -1.25010252e+00 6.13260210e-01
5.82106590e-01 1.68252838e+00 -5.26776493e-01 -1.16680527e+00
4.15295660e-01 -4.02420908e-02 1.95575982e-01 -1.21435905e+00
-1.48172486e+00 -2.09427569e-02 6.14245355e-01 -6.56134009e-01
-1.16493508e-01 -5.05909979e-01 -1.23264825e+00 -7.39785910e-01
-2.36696750e-01 6.38655841e-01 2.12788552e-01 9.32271600e-01
-1.96621846e-02 5.34123421e-01 1.05768830e-01 -2.33712301e-01
2.37773135e-01 -1.10329497e+00 -1.35376608e+00 2.57562362e-02
-8.64271700e-01 -4.84348387e-01 -2.08077013e-01 -3.20876747e-01] | [8.966323852539062, 10.449366569519043] |
5e3036db-6513-4b62-bb8a-9b01809d6f63 | semantic-prediction-which-one-should-come | 2110.02829 | null | https://arxiv.org/abs/2110.02829v1 | https://arxiv.org/pdf/2110.02829v1.pdf | Semantic Prediction: Which One Should Come First, Recognition or Prediction? | The ultimate goal of video prediction is not forecasting future pixel-values given some previous frames. Rather, the end goal of video prediction is to discover valuable internal representations from the vast amount of available unlabeled video data in a self-supervised fashion for downstream tasks. One of the primary downstream tasks is interpreting the scene's semantic composition and using it for decision-making. For example, by predicting human movements, an observer can anticipate human activities and collaborate in a shared workspace. There are two main ways to achieve the same outcome, given a pre-trained video prediction and pre-trained semantic extraction model; one can first apply predictions and then extract semantics or first extract semantics and then predict. We investigate these configurations using the Local Frequency Domain Transformer Network (LFDTN) as the video prediction model and U-Net as the semantic extraction model on synthetic and real datasets. | ['and Sven Behnke', 'Jan Nogga', 'Hafez Farazi'] | 2021-10-06 | null | null | null | null | ['video-prediction'] | ['computer-vision'] | [ 5.51672816e-01 3.04796666e-01 -3.35343182e-01 -5.82262695e-01
-2.08100885e-01 -3.15134168e-01 7.37937331e-01 -1.10866927e-01
-1.70667931e-01 7.53306627e-01 7.28076935e-01 4.70446944e-02
2.34721288e-01 -6.90179348e-01 -8.54268253e-01 -4.24034685e-01
-2.63394825e-02 2.11532321e-03 5.28651416e-01 1.18267983e-01
2.67414749e-01 2.26842389e-01 -1.80845070e+00 8.94868016e-01
1.74204215e-01 1.34243429e+00 6.65967822e-01 7.22277641e-01
-5.51717691e-02 1.86775112e+00 -1.80172950e-01 -2.98087914e-02
1.46880120e-01 -6.72196448e-01 -1.10037220e+00 3.67778361e-01
-1.88813999e-01 -3.56400669e-01 -6.20873094e-01 9.18713391e-01
-9.10953656e-02 3.03733706e-01 5.84712327e-01 -1.20854032e+00
-2.57747859e-01 6.93232536e-01 -1.52326614e-01 2.63251662e-01
5.99161565e-01 1.99074224e-01 8.21509480e-01 -5.47358990e-01
9.85112906e-01 9.88028824e-01 3.18313509e-01 5.38377404e-01
-8.82236183e-01 -2.21665695e-01 4.04435873e-01 6.29686236e-01
-1.05540669e+00 -5.97130477e-01 7.61481404e-01 -5.33314466e-01
8.94809484e-01 1.27367541e-01 7.44371831e-01 1.39933085e+00
3.45757194e-02 8.39125812e-01 6.77794099e-01 -1.26077563e-01
3.42681229e-01 1.05304713e-03 -1.35206044e-01 5.97067416e-01
-3.72626871e-01 -5.82189560e-02 -8.71774971e-01 1.67323038e-01
6.35156214e-01 3.76344264e-01 -4.51511741e-01 -1.34220898e-01
-1.71039724e+00 2.59619057e-01 2.42040709e-01 4.01687354e-01
-6.96542084e-01 3.54708374e-01 4.42001730e-01 4.55693632e-01
4.58601087e-01 3.23916376e-01 -6.02349520e-01 -3.83889943e-01
-9.49980795e-01 1.04984101e-02 6.43033803e-01 9.07024443e-01
8.69635940e-01 -4.18659300e-02 -2.66664356e-01 3.37931126e-01
3.32539916e-01 1.34403378e-01 8.62538815e-01 -1.08161962e+00
4.48663056e-01 4.59223628e-01 2.10383251e-01 -1.03904998e+00
-1.21273018e-01 -2.48047058e-02 -5.35807431e-01 -1.25652626e-01
4.70484525e-01 -3.12899947e-01 -9.19931889e-01 1.67491317e+00
7.59457797e-02 7.76608706e-01 2.99428880e-01 1.18851316e+00
6.89775884e-01 1.05123985e+00 4.58213419e-01 -3.38685244e-01
1.13135946e+00 -1.05263948e+00 -5.63254058e-01 -4.60323125e-01
7.09512413e-01 -4.36074555e-01 4.56670612e-01 2.82603592e-01
-8.20247650e-01 -8.83753896e-01 -9.15546536e-01 -1.41811455e-02
-1.49375081e-01 2.61514723e-01 4.41966206e-01 -2.08259672e-01
-9.35926080e-01 8.63743901e-01 -9.61344540e-01 -5.85940003e-01
4.24768150e-01 1.69335365e-01 -5.05012274e-01 -1.46500263e-02
-1.06567979e+00 6.43347919e-01 5.93624532e-01 -9.00778640e-03
-1.04927313e+00 -4.14177239e-01 -7.91164398e-01 2.84191817e-01
4.26120222e-01 -6.32189870e-01 1.16180468e+00 -1.86672831e+00
-1.30220079e+00 6.15391195e-01 -4.39812273e-01 -7.17385590e-01
2.38845780e-01 -1.87516257e-01 -4.09978300e-01 2.23229110e-01
8.46889243e-02 7.54920065e-01 7.62023032e-01 -1.04133558e+00
-9.71291482e-01 -4.38686699e-01 -1.62092887e-03 3.71493280e-01
-2.22311422e-01 -1.71143055e-01 -4.54422593e-01 -5.54866493e-01
1.69316992e-01 -7.21844435e-01 -3.04484129e-01 -1.50824055e-01
-1.81255624e-01 -5.45041375e-02 1.02687573e+00 -9.86414433e-01
1.15796006e+00 -2.14183354e+00 2.65367925e-01 1.70779571e-01
1.37828037e-01 8.55876133e-02 -8.52002352e-02 3.25436413e-01
-1.61344066e-01 -3.33092272e-01 2.27871612e-01 -2.16038357e-02
-3.92454386e-01 -2.19166558e-02 -5.04221022e-01 1.29923895e-01
1.97264716e-01 8.73527944e-01 -1.17491817e+00 -4.70991433e-01
3.39762539e-01 2.06615210e-01 -4.97737676e-01 5.07366836e-01
-6.91148698e-01 7.14958131e-01 -8.03826392e-01 4.22546864e-01
-9.16410387e-02 -5.54043472e-01 7.94643044e-01 -5.02021134e-01
-8.91747177e-02 2.48364910e-01 -9.65982676e-01 1.85717440e+00
-2.75905967e-01 1.02808082e+00 -4.92500305e-01 -1.53419137e+00
7.82522142e-01 5.44615328e-01 8.17239285e-01 -6.81633472e-01
1.04998723e-01 -3.56934667e-02 -3.59589279e-01 -9.63796675e-01
3.73309493e-01 -6.24567084e-02 6.45842627e-02 4.82455432e-01
4.11399812e-01 7.64115691e-01 -9.42210853e-02 1.93247616e-01
1.45337903e+00 7.56703556e-01 3.94969583e-01 3.07702199e-02
5.93080521e-01 2.49057338e-01 6.30284965e-01 3.52675140e-01
-6.46529123e-02 6.25164628e-01 6.06060505e-01 -8.72528315e-01
-9.81813014e-01 -7.94077754e-01 6.16980135e-01 1.17126167e+00
4.03892398e-01 -6.22629881e-01 -5.78236163e-01 -6.63142323e-01
-3.92493725e-01 6.52858198e-01 -4.29736495e-01 -1.80746004e-01
-4.19961840e-01 -5.53233996e-02 1.22418569e-03 7.67659903e-01
4.74673122e-01 -1.27877986e+00 -1.10488975e+00 2.88018882e-01
-6.65937603e-01 -1.28047311e+00 -4.98826168e-02 5.58744483e-02
-6.77639604e-01 -1.14094961e+00 -3.33115399e-01 -7.94991434e-01
7.22985208e-01 3.93720567e-01 9.92646217e-01 9.76806283e-02
1.06373899e-01 4.37180787e-01 -4.62560385e-01 1.23134807e-01
-2.20327839e-01 -2.98160613e-01 -3.92420590e-02 4.06202912e-01
6.66127443e-01 -5.19293249e-01 -7.17737019e-01 1.73411086e-01
-5.59968889e-01 6.70428216e-01 3.88528705e-01 5.85306764e-01
6.29111052e-01 1.05810575e-02 3.07663947e-01 -8.01752925e-01
1.50025794e-02 -6.91789567e-01 -2.95490414e-01 4.66004282e-01
-1.11411057e-01 6.98927939e-02 8.71815801e-01 -4.16334182e-01
-1.37048960e+00 5.45102894e-01 1.88755184e-01 -8.24022830e-01
-3.55182618e-01 4.50799823e-01 -3.20560515e-01 5.78825533e-01
4.40895289e-01 4.46068853e-01 -1.62277862e-01 -1.61628097e-01
2.21854985e-01 6.38211608e-01 5.72258472e-01 -4.09373611e-01
3.49805117e-01 5.36714971e-01 -1.77535132e-01 -8.67458522e-01
-1.04583752e+00 -4.46435899e-01 -7.57883012e-01 -6.70636833e-01
1.24525142e+00 -1.17564631e+00 -7.52690852e-01 1.79957911e-01
-1.21306026e+00 -5.05317211e-01 -3.86650831e-01 6.20907307e-01
-9.70413804e-01 1.85854226e-01 -2.84874052e-01 -6.54496193e-01
6.17823415e-02 -1.03472853e+00 1.02450454e+00 2.24058777e-01
-4.65386271e-01 -8.79902065e-01 -1.71322137e-01 4.71373111e-01
2.53193341e-02 2.83774078e-01 6.04370117e-01 -6.70508683e-01
-9.60717142e-01 7.23819435e-02 -1.75421193e-01 1.37073010e-01
1.42720297e-01 -1.79559812e-01 -9.99697328e-01 1.65155724e-01
7.92219192e-02 -2.50409096e-01 6.84653819e-01 4.61555570e-01
1.51391292e+00 -5.04435658e-01 -5.88449299e-01 4.84257072e-01
1.42691219e+00 2.63217121e-01 6.69256389e-01 2.26075262e-01
6.74799562e-01 8.08340371e-01 7.82564700e-01 5.35415888e-01
5.49176633e-01 5.35350740e-01 2.70450562e-01 2.86350995e-01
-9.92314070e-02 -7.28342593e-01 5.19107640e-01 5.79138577e-01
-3.67906064e-01 -2.61004955e-01 -9.34937119e-01 5.17773688e-01
-2.60234427e+00 -1.54059362e+00 1.57943547e-01 2.01191831e+00
3.73224139e-01 1.84597429e-02 -1.12339929e-02 -2.61412077e-02
6.20744884e-01 1.63748085e-01 -5.53438604e-01 9.60012823e-02
2.49073014e-01 -3.09112549e-01 3.88024539e-01 1.12967297e-01
-1.17926216e+00 1.09786189e+00 5.46093464e+00 5.05642891e-01
-1.19176900e+00 2.15796590e-01 1.06285238e+00 -1.48471206e-01
3.45369913e-02 3.12960505e-01 -5.07670701e-01 5.88675022e-01
1.14346552e+00 -1.33173734e-01 4.26389247e-01 7.58127332e-01
5.09171903e-01 -2.27511033e-01 -1.37596667e+00 1.02754116e+00
1.38384765e-02 -1.74361348e+00 8.47049281e-02 -1.65652722e-01
5.32378197e-01 -1.99695155e-02 -4.86570477e-01 1.30998120e-01
2.20836088e-01 -8.39280128e-01 9.51386213e-01 1.06285572e+00
6.32282197e-01 -3.83638740e-01 4.57162708e-01 6.29327059e-01
-1.38100064e+00 -3.74510854e-01 -2.98892379e-01 -4.20187294e-01
3.06325167e-01 2.52243876e-01 -6.45450711e-01 1.99096203e-01
6.78359628e-01 1.30365872e+00 -1.46396905e-01 6.71575606e-01
-2.70996720e-01 6.25876009e-01 -5.52546047e-02 9.00003910e-02
2.18338147e-01 -1.56732813e-01 3.08794916e-01 8.78854990e-01
4.34102476e-01 4.52387512e-01 3.08901697e-01 6.41734004e-01
-3.34840491e-02 -1.43971801e-01 -7.24739730e-01 -1.32309303e-01
2.29867503e-01 1.06234896e+00 -9.32988286e-01 -6.08808577e-01
-6.56978488e-01 1.15925944e+00 3.78052980e-01 7.01933086e-01
-9.85279202e-01 2.10728541e-01 7.34337449e-01 2.92456150e-01
3.04024905e-01 -7.98662752e-02 -2.11190376e-02 -1.44905400e+00
7.92890117e-02 -6.03932738e-01 2.87960678e-01 -1.23296535e+00
-8.36872101e-01 5.09951591e-01 -3.10777664e-01 -1.55734968e+00
-5.60674667e-01 -4.32083309e-01 -5.61593652e-01 4.04161870e-01
-1.36324477e+00 -1.25753009e+00 -4.71329868e-01 7.19260216e-01
9.71377373e-01 -3.11614782e-01 6.42675459e-01 -1.78930797e-02
-1.12981826e-01 -2.82253981e-01 -1.73282444e-01 2.88310111e-01
2.98164248e-01 -8.69741917e-01 1.85643643e-01 1.01901603e+00
4.63131666e-01 9.94132087e-02 6.87546611e-01 -7.26497769e-01
-1.48588586e+00 -1.38961971e+00 9.02307212e-01 -4.00911093e-01
6.62809670e-01 -2.54442573e-01 -6.62512839e-01 1.06284070e+00
-5.85565120e-02 1.47677526e-01 5.56835473e-01 -2.59520471e-01
-5.79203181e-02 6.44093379e-02 -7.36934364e-01 6.68204427e-01
1.50938666e+00 -4.93519723e-01 -6.35529041e-01 3.58501792e-01
6.67282820e-01 -2.31171310e-01 -8.77416015e-01 1.70982957e-01
7.18612611e-01 -1.09369218e+00 8.98165882e-01 -1.00375092e+00
9.41999674e-01 -3.10997754e-01 -4.13015664e-01 -1.15069389e+00
-3.78867537e-01 -4.88145083e-01 -3.09490800e-01 1.00047660e+00
3.68551910e-01 -6.58464357e-02 1.06751192e+00 8.38411868e-01
1.37911722e-01 -5.38357019e-01 -4.80555862e-01 -2.88235247e-01
-8.67992282e-01 -6.95531726e-01 6.24667823e-01 7.77271926e-01
3.23490202e-01 4.70863819e-01 -6.74449742e-01 4.55203801e-02
3.49687815e-01 2.95393318e-01 8.05587351e-01 -9.86550272e-01
-3.07981819e-01 -6.28494285e-03 -7.62226403e-01 -1.32000816e+00
3.99561584e-01 -7.08393633e-01 9.73739699e-02 -1.62473071e+00
2.15202793e-01 -2.29110777e-01 -3.93472731e-01 5.21344364e-01
1.08041123e-01 -2.14257594e-02 3.77708435e-01 3.66405457e-01
-1.03466165e+00 4.63457018e-01 1.04992783e+00 1.34744300e-02
-1.23283237e-01 1.28862858e-01 -5.12366176e-01 8.02190304e-01
5.21956801e-01 -1.81623429e-01 -7.89276123e-01 -4.68468487e-01
2.53967077e-01 6.76639259e-01 4.55243021e-01 -1.23048091e+00
3.94493788e-01 -4.00976926e-01 7.48179495e-01 -4.01130706e-01
3.19729120e-01 -1.06714892e+00 3.63034636e-01 2.61401445e-01
-5.83381593e-01 -6.48390353e-02 -3.22158515e-01 8.22637618e-01
-5.34626484e-01 8.13386515e-02 3.26453418e-01 -3.59311312e-01
-1.55378664e+00 2.91831434e-01 -6.35671675e-01 -2.17755020e-01
1.42638814e+00 -6.00380301e-01 -1.58914998e-01 -5.63978434e-01
-1.13890696e+00 1.71496674e-01 4.43343580e-01 7.57663727e-01
7.16827571e-01 -1.29619074e+00 -3.38954180e-01 3.21333408e-01
7.36929998e-02 -2.17094332e-01 3.14509153e-01 6.07860863e-01
-3.30917984e-01 3.92242342e-01 -4.53650504e-01 -6.44631326e-01
-1.07800329e+00 6.61544085e-01 2.23930657e-01 -8.44449848e-02
-7.55319297e-01 6.29034758e-01 4.06625897e-01 8.78085569e-02
1.59653410e-01 -1.53948367e-01 -4.97681975e-01 -1.39019594e-01
6.93234146e-01 1.14764728e-01 -3.14084858e-01 -8.86729360e-01
-1.78681031e-01 9.14296806e-02 4.10583049e-01 3.35020870e-02
1.67931175e+00 -3.57816368e-01 1.37380704e-01 5.32455444e-01
1.14703822e+00 -7.31613338e-01 -1.61652696e+00 -2.79141963e-01
2.35461771e-01 -5.02302527e-01 -7.27801817e-03 -4.19361234e-01
-1.28647482e+00 6.77577853e-01 2.88235337e-01 3.99568491e-02
1.20633900e+00 1.42642021e-01 7.58761704e-01 4.15295929e-01
6.57132149e-01 -1.28706956e+00 2.91367650e-01 2.56556690e-01
4.16673750e-01 -1.15155828e+00 -1.75821334e-01 -3.03423017e-01
-1.01214051e+00 1.32134414e+00 7.49971747e-01 1.00165829e-02
7.64206886e-01 -2.17502587e-03 -1.46829262e-01 -3.01027268e-01
-1.19814765e+00 -1.99353635e-01 2.78448999e-01 5.15724778e-01
5.22588909e-01 -2.33054869e-02 1.68889076e-01 7.53704846e-01
1.00088410e-01 5.74889898e-01 4.03163224e-01 7.75212765e-01
-4.86690849e-01 -7.39758372e-01 2.78309193e-02 7.80969560e-01
-2.99832970e-01 2.88059801e-01 -1.46723151e-01 1.08234122e-01
1.70289084e-01 8.14498961e-01 2.10784972e-01 -7.53854454e-01
-2.14428380e-02 2.23904029e-01 2.94786632e-01 -6.79791868e-01
-2.85193145e-01 -9.43024307e-02 2.14352638e-01 -1.12353444e+00
-9.54126418e-01 -7.85980225e-01 -1.25023007e+00 -2.23103515e-03
1.72834620e-01 -1.37956208e-03 4.56647992e-01 1.28906369e+00
4.97374892e-01 4.91958439e-01 5.73439002e-01 -8.65580201e-01
-8.07547420e-02 -6.59158349e-01 -3.42862040e-01 8.29805374e-01
1.23607546e-01 -4.97933000e-01 2.22991854e-02 9.00237799e-01] | [8.36875057220459, 0.42859703302383423] |
bca36115-97d9-4b1a-9985-dec6aebaf435 | swapped-goal-conditioned-offline | 2302.08865 | null | https://arxiv.org/abs/2302.08865v1 | https://arxiv.org/pdf/2302.08865v1.pdf | Swapped goal-conditioned offline reinforcement learning | Offline goal-conditioned reinforcement learning (GCRL) can be challenging due to overfitting to the given dataset. To generalize agents' skills outside the given dataset, we propose a goal-swapping procedure that generates additional trajectories. To alleviate the problem of noise and extrapolation errors, we present a general offline reinforcement learning method called deterministic Q-advantage policy gradient (DQAPG). In the experiments, DQAPG outperforms state-of-the-art goal-conditioned offline RL methods in a wide range of benchmark tasks, and goal-swapping further improves the test results. It is noteworthy, that the proposed method obtains good performance on the challenging dexterous in-hand manipulation tasks for which the prior methods failed. | ['Joni-Kristen Kämäräinen', 'Joni Pajarinen', 'Dingding Cai', 'Huiling Wang', 'Wenyan Yang'] | 2023-02-17 | null | null | null | null | ['offline-rl'] | ['playing-games'] | [-6.61941245e-02 -1.46386936e-01 -2.95191497e-01 -1.24239521e-02
-9.33434546e-01 -4.24551159e-01 4.13263977e-01 -2.00660095e-01
-6.65051401e-01 1.48660326e+00 -5.33852093e-02 -3.91275913e-01
-3.92018259e-01 -6.49012446e-01 -9.20530975e-01 -7.19408095e-01
-3.71195465e-01 4.20315504e-01 1.60446614e-01 -4.77198541e-01
4.51438129e-01 3.30570132e-01 -1.37831891e+00 -1.37874350e-01
1.42100084e+00 7.61689961e-01 5.04155397e-01 3.62121582e-01
2.76475608e-01 1.02579784e+00 -5.82259953e-01 -1.05498388e-01
5.29508471e-01 -7.00672805e-01 -5.57297587e-01 -2.12140188e-01
1.76147409e-02 -7.71273792e-01 -6.90746248e-01 1.05984139e+00
5.70596337e-01 8.53617489e-01 4.35279578e-01 -1.30478191e+00
-7.46039629e-01 5.83644211e-01 -2.81005174e-01 4.58013564e-02
4.98313546e-01 6.59006655e-01 7.22923040e-01 -6.53828919e-01
6.72674179e-01 1.27964962e+00 4.77982193e-01 7.42156565e-01
-1.02162516e+00 -7.59220898e-01 5.83534896e-01 1.82785884e-01
-9.40358877e-01 2.07295522e-01 7.06874609e-01 -3.24020326e-01
8.58702183e-01 -2.13942364e-01 8.08222353e-01 1.53031957e+00
2.58487940e-01 1.07160807e+00 1.57411742e+00 -1.24289960e-01
6.32309616e-01 -5.19825101e-01 -3.46742511e-01 7.75374353e-01
1.33299455e-01 1.09352160e+00 -5.22467375e-01 -3.69608514e-02
9.46875513e-01 2.60193236e-02 -5.39482385e-02 -5.10688007e-01
-1.19856465e+00 9.40722406e-01 4.96300310e-01 -1.25804141e-01
-6.73740208e-01 2.61605978e-01 4.69459683e-01 5.82776964e-01
2.93650299e-01 9.07186210e-01 -5.43809950e-01 -5.95852613e-01
-5.26974499e-01 9.51840997e-01 6.37163877e-01 1.13835800e+00
4.98731136e-01 3.32491308e-01 -5.66681087e-01 6.57525599e-01
-9.27742422e-02 5.29471815e-01 6.51052296e-01 -1.16559148e+00
8.09595108e-01 3.21639091e-01 5.36177039e-01 -5.59432864e-01
-3.27390671e-01 -4.32284564e-01 -3.71244520e-01 5.27461112e-01
6.72932625e-01 -6.29947782e-01 -1.05511510e+00 1.80921280e+00
3.70940059e-01 1.00393362e-01 2.58199006e-01 1.06321561e+00
4.98060584e-01 5.18687785e-01 2.74778187e-01 -3.80229205e-01
3.22910219e-01 -1.27925515e+00 -7.78184652e-01 -2.71745861e-01
5.27848363e-01 -3.28490324e-02 1.46046090e+00 6.94935620e-01
-8.89960051e-01 -5.66247165e-01 -1.00342000e+00 3.13493639e-01
-2.45151594e-01 9.66084283e-03 8.70621085e-01 2.15943500e-01
-5.61956525e-01 1.18534076e+00 -9.58839059e-01 4.46673073e-02
4.73988801e-01 2.87058890e-01 -1.61359355e-01 -1.23132408e-01
-1.19652557e+00 1.09046531e+00 8.67921293e-01 -7.89097622e-02
-1.77424741e+00 -5.41006923e-01 -8.05276155e-01 -2.29396433e-01
1.04797900e+00 -3.02318454e-01 1.45142555e+00 -7.01287210e-01
-2.29002237e+00 2.42332920e-01 3.69816899e-01 -5.55231512e-01
9.24950242e-01 -6.17654622e-01 -1.45540282e-01 3.32313403e-02
-6.69848099e-02 6.11423373e-01 9.76854384e-01 -1.21598685e+00
-7.32731760e-01 -2.79367685e-01 2.16115281e-01 5.08323491e-01
-6.66229352e-02 -5.00025094e-01 -5.04612327e-02 -8.73283744e-01
-3.19630384e-01 -9.58964050e-01 -6.57694757e-01 -5.76537132e-01
-1.90996364e-01 -4.31744754e-01 5.40853798e-01 -7.30101883e-01
1.20301950e+00 -1.82232630e+00 4.37980503e-01 -3.88665423e-02
-2.52678573e-01 2.80004263e-01 -3.89505982e-01 5.77701032e-01
2.20446914e-01 -4.13901776e-01 -2.30557635e-01 -4.08319682e-02
2.06605360e-01 4.07036752e-01 -5.54775715e-01 2.57565409e-01
-1.52286172e-01 1.07617271e+00 -1.58404446e+00 -2.69464880e-01
3.95969748e-02 -2.10777536e-01 -6.37718022e-01 5.37737250e-01
-7.27840543e-01 9.54782307e-01 -6.71788752e-01 7.10820198e-01
2.62172103e-01 1.48418173e-01 1.44015923e-01 5.76169193e-01
-2.32621562e-02 1.39241755e-01 -8.48071635e-01 2.10830784e+00
-2.65417993e-01 1.57904983e-01 -4.02106822e-01 -9.63015914e-01
1.05508590e+00 5.34421653e-02 5.09340525e-01 -8.63441825e-01
1.10370986e-01 3.24874312e-01 1.72333047e-01 -6.80118799e-01
6.05259955e-01 -8.45161527e-02 -9.80690122e-02 2.51497060e-01
2.96050102e-01 -3.16323429e-01 4.25614119e-01 -6.88455850e-02
9.89470065e-01 9.33189869e-01 2.81722605e-01 -1.65463924e-01
3.11356157e-01 1.39931530e-01 7.34092653e-01 1.09485447e+00
-5.52416563e-01 1.54611528e-01 5.00688016e-01 -2.69420385e-01
-8.55548680e-01 -1.09751761e+00 3.99155766e-01 1.37404966e+00
1.62255913e-01 -1.68503970e-01 -6.21264935e-01 -1.18202615e+00
4.40589190e-01 9.63991344e-01 -6.83746219e-01 -3.86138618e-01
-7.06771672e-01 -4.38861549e-01 3.34596217e-01 8.31983209e-01
5.79477370e-01 -1.47984338e+00 -6.86229229e-01 4.49160784e-01
2.09003408e-02 -8.50066960e-01 -3.83425951e-01 1.48996100e-01
-1.14783335e+00 -1.09337473e+00 -9.49416578e-01 -6.45220578e-01
5.10093868e-01 -1.16161242e-01 7.83042431e-01 -3.11379611e-01
1.56474113e-02 3.55498105e-01 -5.73266089e-01 -4.33092326e-01
-3.73193055e-01 5.71169704e-02 2.97229081e-01 -6.00013494e-01
1.44477352e-01 -4.13697749e-01 -6.46804810e-01 2.17333764e-01
-4.03937787e-01 -2.71364987e-01 4.54917848e-01 1.37477875e+00
6.66985631e-01 -3.23084854e-02 1.09059393e+00 -6.19213104e-01
1.20296514e+00 -4.68445331e-01 -9.69674587e-01 1.78090960e-01
-8.14682305e-01 1.93640709e-01 1.10390115e+00 -9.23573792e-01
-1.12036490e+00 -1.18385576e-01 -2.45812107e-02 -6.42424226e-01
4.53664996e-02 5.00908077e-01 2.35765517e-01 -8.62956271e-02
8.55322540e-01 3.61364335e-01 1.70130879e-01 -4.15330142e-01
3.87176722e-01 3.03022772e-01 5.10024011e-01 -1.03274345e+00
4.10272360e-01 2.08852105e-02 6.21089488e-02 -2.25088030e-01
-9.24151778e-01 4.99928482e-02 -2.47740045e-01 -3.33170921e-01
4.82767373e-01 -7.22301066e-01 -1.01803565e+00 5.19700110e-01
-5.92429996e-01 -1.27779651e+00 -5.17545342e-01 7.81751812e-01
-1.28194380e+00 2.40747154e-01 -6.97588682e-01 -1.00655675e+00
-1.72772408e-01 -1.27856481e+00 4.92414206e-01 4.71336305e-01
2.04450250e-01 -8.17046702e-01 1.68270439e-01 -1.11907907e-01
2.01664418e-01 3.53496373e-01 5.66600561e-01 -5.79826534e-01
-3.43013316e-01 6.46809861e-02 1.54641330e-01 4.64476347e-01
1.44402739e-02 -4.74110812e-01 -4.28404063e-01 -8.80261064e-01
-1.78068802e-01 -1.10286558e+00 7.43063927e-01 2.50154287e-01
1.38825214e+00 -1.73116177e-01 -1.20399162e-01 5.35705268e-01
1.28890765e+00 7.08613753e-01 4.39096212e-01 5.07724702e-01
4.07693088e-01 2.73765087e-01 1.50212419e+00 7.21237898e-01
1.90982640e-01 2.48382524e-01 6.44934952e-01 3.37265611e-01
1.78570598e-01 -9.64247942e-01 5.41941524e-01 3.68828297e-01
-3.62570733e-01 -4.92181666e-02 -5.16992390e-01 5.09745121e-01
-2.20344996e+00 -7.93181360e-01 3.85715187e-01 2.19062185e+00
9.52739716e-01 1.68744251e-01 5.47317624e-01 -3.33553344e-01
3.92977685e-01 -7.76883662e-02 -1.25126362e+00 -2.47768685e-01
2.24318892e-01 3.49608809e-01 5.80207884e-01 4.73157555e-01
-9.06005025e-01 1.34567237e+00 6.98489666e+00 9.90951002e-01
-7.93986976e-01 -3.17461267e-02 2.36757934e-01 -1.44926682e-01
1.23388641e-01 -1.41852289e-01 -7.44781315e-01 4.86565411e-01
5.40289044e-01 -1.97171196e-01 1.20341063e+00 1.15165031e+00
2.15194106e-01 -2.92754531e-01 -1.14757264e+00 7.77753055e-01
-2.56355643e-01 -8.29124272e-01 -2.73393452e-01 -1.27377927e-01
1.05973136e+00 3.18206730e-03 1.21258847e-01 1.11054957e+00
9.26287830e-01 -1.04075587e+00 5.94086945e-01 5.64774156e-01
9.03309941e-01 -1.04326713e+00 4.31497633e-01 6.57905221e-01
-6.21816754e-01 -6.93434656e-01 -4.93747830e-01 -2.93007374e-01
8.91276728e-03 -7.20484927e-02 -5.88427722e-01 5.71117580e-01
4.18455362e-01 7.10844398e-01 -1.84824258e-01 9.61475611e-01
-8.32519054e-01 6.16946995e-01 -1.12827718e-01 -3.89943391e-01
7.35238731e-01 -4.87538248e-01 5.92413127e-01 6.43670142e-01
3.42132539e-01 2.12733701e-01 4.77281421e-01 9.93759453e-01
1.08445309e-01 2.82596406e-02 -8.07791531e-01 -3.52960795e-01
4.32329029e-01 7.97480702e-01 -3.30125600e-01 -2.78091490e-01
6.95304340e-03 9.70063210e-01 7.87727118e-01 7.08800316e-01
-8.06784093e-01 -5.48095822e-01 4.03963447e-01 -4.64157879e-01
3.98083031e-01 -4.88485008e-01 3.42885442e-02 -1.00181222e+00
-3.34250741e-02 -1.15315080e+00 4.91635889e-01 -5.18815935e-01
-1.27098763e+00 2.46999413e-01 1.41124681e-01 -1.40527332e+00
-5.60631752e-01 -7.15638638e-01 -3.10353398e-01 6.34824812e-01
-1.41653061e+00 -5.94918609e-01 -1.69324338e-01 7.67829955e-01
6.13089859e-01 -5.22608757e-01 6.75309360e-01 -1.51317149e-01
-4.34899837e-01 6.93808734e-01 4.18207377e-01 -1.63168326e-01
5.58481574e-01 -1.29807782e+00 7.08113238e-02 5.59294343e-01
-3.21751863e-01 5.46851695e-01 7.23162234e-01 -8.92428041e-01
-1.55139852e+00 -9.10633743e-01 -7.86869824e-02 -5.59529811e-02
6.91653311e-01 -2.02046096e-01 -6.34155512e-01 7.86597431e-01
9.66033861e-02 -5.77065870e-02 1.07373774e-01 -2.53641196e-02
1.18094206e-01 3.37113962e-02 -1.15362298e+00 9.51553941e-01
1.39609766e+00 -1.54360712e-01 -8.07437420e-01 4.68775481e-01
8.05331469e-01 -8.76117408e-01 -9.19824719e-01 5.49245298e-01
4.66211110e-01 -6.47715509e-01 7.23916173e-01 -1.19230294e+00
6.63674891e-01 -9.19260010e-02 1.11112453e-01 -2.06233382e+00
-1.99711472e-01 -9.98352647e-01 -5.54308414e-01 5.94558716e-01
7.26103112e-02 -8.20872724e-01 7.54190564e-01 2.71570265e-01
-1.86955184e-01 -1.01162279e+00 -6.73046708e-01 -1.43681550e+00
5.45637250e-01 -1.86524808e-01 6.43625081e-01 6.99010015e-01
5.13214529e-01 -6.78273514e-02 -6.28755748e-01 -2.38545582e-01
7.52450109e-01 2.88742334e-01 9.88387108e-01 -6.16239250e-01
-5.69071472e-01 -2.61417270e-01 1.91701710e-01 -1.44039261e+00
3.46530825e-01 -6.83095098e-01 3.54554534e-01 -1.35308981e+00
-1.82215586e-01 -5.07211566e-01 -3.89734209e-01 3.77473265e-01
-5.51343918e-01 -4.19874012e-01 4.17255789e-01 -9.08575654e-02
-7.71334231e-01 1.25975633e+00 2.13903880e+00 4.02042866e-02
-5.28469145e-01 1.37677148e-01 -4.01367396e-01 4.75340754e-01
1.01717329e+00 -4.65508223e-01 -7.11032867e-01 -8.67941305e-02
3.22351814e-03 5.82906306e-01 8.84561315e-02 -9.22083437e-01
-2.94059478e-02 -7.00744808e-01 2.03596473e-01 -5.40009916e-01
1.56296864e-01 -5.16368866e-01 -3.81284237e-01 7.80448675e-01
-5.58715522e-01 4.71149720e-02 2.29177490e-01 7.90605068e-01
-3.69378030e-02 -2.92638958e-01 5.63840389e-01 -4.00070220e-01
-7.41423130e-01 4.42096889e-01 -3.27079535e-01 4.27574366e-01
1.19964325e+00 7.56750628e-03 -2.37259403e-01 -3.98637235e-01
-7.80384600e-01 7.59096801e-01 1.30803302e-01 3.95001829e-01
8.40627611e-01 -1.38520908e+00 -5.25048494e-01 7.87943527e-02
-5.60943186e-02 -3.84794250e-02 1.65101558e-01 6.36273682e-01
-3.17496955e-01 3.20556730e-01 -4.36234921e-01 -1.69516116e-01
-4.51749623e-01 9.15186465e-01 3.75835180e-01 -6.39354289e-01
-7.48310208e-01 7.08407879e-01 -1.89958394e-01 -7.17533231e-01
5.20109832e-01 -4.21094954e-01 1.91923119e-02 -4.32808280e-01
2.74342448e-01 7.60380566e-01 -2.78730124e-01 1.62633032e-01
1.12007089e-01 8.38276837e-03 -1.36690855e-01 -2.77520418e-01
1.29999876e+00 2.55637765e-01 5.06503642e-01 4.17410553e-01
5.66727400e-01 -5.48954368e-01 -2.18599772e+00 -7.30107352e-02
-2.11412758e-02 -6.77847028e-01 -2.91319132e-01 -1.18839049e+00
-7.41022408e-01 5.67376137e-01 5.33632696e-01 -1.27956480e-01
8.96899700e-01 -6.20319724e-01 7.48555303e-01 8.07424247e-01
9.89500701e-01 -1.72052288e+00 4.34338242e-01 9.09620345e-01
1.17157853e+00 -1.33270228e+00 6.21367283e-02 1.74611300e-01
-1.03787434e+00 9.71759200e-01 1.19207871e+00 -6.43672526e-01
2.40950242e-01 -4.64534461e-02 -2.06071064e-01 1.07256427e-01
-7.03225732e-01 -3.43963981e-01 -8.66494849e-02 7.34417975e-01
-1.01126939e-01 2.14986533e-01 -8.29515278e-01 5.75703204e-01
-2.50527322e-01 5.10000706e-01 1.28670931e-01 1.24435472e+00
-2.58239895e-01 -9.76839423e-01 -1.93732724e-01 3.71742100e-01
-1.88358754e-01 2.73498833e-01 -1.04966693e-01 1.01803017e+00
-2.27280751e-01 8.08287144e-01 -3.99945617e-01 -4.18251187e-01
4.46805477e-01 -4.32457291e-02 1.01683784e+00 -3.06889534e-01
-6.24066830e-01 -5.61261922e-02 7.40578771e-02 -1.00697529e+00
-1.43150210e-01 -6.05974793e-01 -1.50844359e+00 5.46246730e-02
-2.06485346e-01 1.94372963e-02 2.67202377e-01 9.84747052e-01
1.74407467e-01 6.44384086e-01 6.32091105e-01 -8.77890646e-01
-1.51802862e+00 -1.22338676e+00 -6.87710762e-01 6.23162091e-01
1.85392052e-01 -1.30358005e+00 -9.92692187e-02 -3.93780977e-01] | [4.139947414398193, 2.058342695236206] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.