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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
7785893c-e43f-47e1-80eb-91d1fc3e6a0f | stefann-scene-text-editor-using-font-adaptive | 1903.01192 | null | https://arxiv.org/abs/1903.01192v3 | https://arxiv.org/pdf/1903.01192v3.pdf | STEFANN: Scene Text Editor using Font Adaptive Neural Network | Textual information in a captured scene plays an important role in scene interpretation and decision making. Though there exist methods that can successfully detect and interpret complex text regions present in a scene, to the best of our knowledge, there is no significant prior work that aims to modify the textual information in an image. The ability to edit text directly on images has several advantages including error correction, text restoration and image reusability. In this paper, we propose a method to modify text in an image at character-level. We approach the problem in two stages. At first, the unobserved character (target) is generated from an observed character (source) being modified. We propose two different neural network architectures - (a) FANnet to achieve structural consistency with source font and (b) Colornet to preserve source color. Next, we replace the source character with the generated character maintaining both geometric and visual consistency with neighboring characters. Our method works as a unified platform for modifying text in images. We present the effectiveness of our method on COCO-Text and ICDAR datasets both qualitatively and quantitatively. | ['Prasun Roy', 'Umapada Pal', 'Subhankar Ghosh', 'Saumik Bhattacharya'] | 2019-03-04 | stefann-scene-text-editor-using-font-adaptive-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Roy_STEFANN_Scene_Text_Editor_Using_Font_Adaptive_Neural_Network_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Roy_STEFANN_Scene_Text_Editor_Using_Font_Adaptive_Neural_Network_CVPR_2020_paper.pdf | cvpr-2020-6 | ['scene-text-editing'] | ['computer-vision'] | [ 1.00891924e+00 -1.62184000e-01 2.85656661e-01 -4.32715148e-01
-1.92591518e-01 -5.62003255e-01 6.03142083e-01 2.65753448e-01
-4.18788254e-01 4.81422305e-01 1.76881284e-01 -2.08935365e-01
3.26248199e-01 -7.34349012e-01 -9.97825146e-01 -4.95600134e-01
7.78041422e-01 1.69984967e-01 6.20160460e-01 -1.19660936e-01
5.41911960e-01 7.51753628e-01 -1.31180537e+00 5.65007567e-01
9.77743804e-01 6.14198327e-01 5.68321705e-01 8.51218164e-01
-2.37943277e-01 9.97638047e-01 -6.69084847e-01 -5.09498239e-01
2.74726510e-01 -5.58533013e-01 -7.51899242e-01 6.88855410e-01
6.16098166e-01 -5.01116693e-01 -3.92147720e-01 1.29580557e+00
2.56436616e-01 1.23779764e-02 7.44902909e-01 -8.64456236e-01
-1.13550985e+00 8.66766989e-01 -9.49269772e-01 -2.48130225e-03
2.35295035e-02 -8.72725770e-02 6.03626192e-01 -8.79297376e-01
6.38172090e-01 1.13866532e+00 3.68097574e-01 3.68837714e-01
-1.06930339e+00 -3.27064782e-01 3.19375843e-01 1.30182326e-01
-1.05493116e+00 -4.00890231e-01 9.37696338e-01 -5.34335673e-01
4.65461493e-01 2.75281191e-01 5.89964688e-01 8.28969300e-01
3.38190436e-01 9.02437270e-01 8.55819583e-01 -8.28509748e-01
2.48243976e-02 2.00593188e-01 -1.38742462e-01 7.20297098e-01
2.58598864e-01 -2.80391574e-01 -4.71534550e-01 3.32506448e-01
6.66975379e-01 2.03829542e-01 -3.68396312e-01 -2.58828968e-01
-1.22226429e+00 6.04333997e-01 3.14692348e-01 1.97048768e-01
-2.29079381e-01 3.24557930e-01 1.81359738e-01 -3.66829671e-02
3.90597254e-01 3.73536557e-01 -1.07388653e-01 2.11212829e-01
-1.07321894e+00 -2.13399544e-01 3.26311946e-01 8.66913676e-01
6.73994780e-01 2.91929185e-01 -2.75705367e-01 9.98629928e-01
1.64278373e-01 7.39018679e-01 3.50938499e-01 -6.26800299e-01
5.26300728e-01 5.98593712e-01 -8.81395265e-02 -1.37073231e+00
-1.72162995e-01 -9.67759416e-02 -1.03034413e+00 3.95196974e-01
1.51945755e-01 -4.66278195e-02 -1.18464959e+00 1.27334833e+00
1.61292538e-01 -3.01094323e-01 -1.49213970e-01 7.04922378e-01
4.88810033e-01 9.09050703e-01 -3.42781514e-01 1.34266078e-01
1.20944130e+00 -1.24999237e+00 -8.30737293e-01 -4.53888208e-01
1.73414364e-01 -1.24073505e+00 9.58698869e-01 3.90516430e-01
-1.13846123e+00 -5.13483107e-01 -1.09108889e+00 -3.81366074e-01
-4.84912187e-01 5.49540877e-01 8.33461806e-02 4.62799102e-01
-1.00779772e+00 2.09206805e-01 -5.93852103e-01 -5.34612834e-01
2.53109187e-01 1.45337507e-01 -1.81917816e-01 -6.17824234e-02
-8.46727729e-01 7.00129151e-01 5.53891838e-01 2.79378027e-01
-6.87935889e-01 -3.96166325e-01 -6.21061683e-01 4.46524918e-02
5.12614071e-01 -5.66828251e-01 1.12045145e+00 -1.60289657e+00
-1.56085670e+00 6.21014774e-01 -2.39881188e-01 -2.50150293e-01
7.73356497e-01 -4.69671227e-02 -3.66972059e-01 3.52856338e-01
1.09969489e-01 8.48229587e-01 1.15995896e+00 -1.66445673e+00
-9.36440766e-01 -1.19918123e-01 -1.07090101e-01 3.59358519e-01
-3.58792812e-01 -1.40332401e-01 -1.01080036e+00 -1.21846175e+00
2.06920743e-01 -7.80691624e-01 1.80586502e-02 5.48440695e-01
-7.22250521e-01 3.32668036e-01 1.04394352e+00 -9.46879387e-01
1.18608749e+00 -2.04663968e+00 2.58195102e-02 2.65382379e-01
2.09234565e-01 3.72969985e-01 -1.68788016e-01 4.70943153e-01
1.39937866e-02 1.76952943e-01 -5.34410477e-01 -4.03809041e-01
-2.48754621e-01 1.00171849e-01 -4.46250647e-01 3.62518966e-01
1.02907576e-01 8.21382105e-01 -5.52440643e-01 -5.56165576e-01
5.59812069e-01 4.81343925e-01 -4.64108050e-01 -2.04957902e-01
-3.61142009e-01 2.22049832e-01 -2.75968462e-01 5.78513145e-01
8.80208969e-01 -2.11013362e-01 4.19870988e-02 -4.51221168e-01
-2.46412590e-01 -3.16503793e-01 -1.15825605e+00 1.37117505e+00
-1.31537572e-01 1.12883973e+00 -1.93288270e-02 -7.58238494e-01
7.90026844e-01 -1.94049150e-01 2.64892220e-01 -7.62674093e-01
9.79979560e-02 -3.67942266e-02 -1.03350088e-01 -4.31224167e-01
1.17099214e+00 3.59595925e-01 1.94720998e-01 4.83870029e-01
-5.82841516e-01 -6.68587089e-02 3.82388026e-01 2.20091715e-01
6.26183271e-01 -7.68586323e-02 1.72812283e-01 -8.07345882e-02
5.55663168e-01 1.63327545e-01 2.03663751e-01 9.75447536e-01
1.10429287e-01 9.60688412e-01 4.11829025e-01 -4.03629273e-01
-1.34134626e+00 -9.09190655e-01 2.88842116e-02 1.09240019e+00
3.32684070e-01 -4.43997197e-02 -8.63728464e-01 -3.59250575e-01
-1.81645527e-01 6.63888693e-01 -6.86014712e-01 4.51252200e-02
-6.18516028e-01 -4.18432504e-01 6.66273177e-01 5.83775640e-01
9.79062140e-01 -1.06874597e+00 -4.67212826e-01 2.63203699e-02
-2.99850136e-01 -1.08039248e+00 -1.01831448e+00 1.68243378e-01
-6.34678483e-01 -7.82418251e-01 -5.64093888e-01 -1.01633418e+00
1.13022137e+00 5.10396719e-01 6.81535840e-01 8.02400857e-02
-2.86879271e-01 1.77758664e-01 -4.54052836e-01 -2.80541062e-01
-6.95133209e-01 -1.11231424e-01 -4.33314860e-01 3.22376132e-01
-1.88892037e-01 6.19228482e-02 -4.28945333e-01 1.73638821e-01
-1.55430162e+00 6.16241992e-01 7.83594906e-01 6.85480118e-01
6.00301504e-01 3.71648341e-01 -1.82215497e-01 -1.11996269e+00
5.85130453e-01 9.90523025e-02 -6.57649100e-01 5.01358330e-01
-4.78470951e-01 1.36979252e-01 7.39277899e-01 -3.68662655e-01
-1.34732652e+00 5.04704416e-01 7.47014731e-02 -2.19574437e-01
-1.89151853e-01 4.21871901e-01 -1.60321832e-01 -2.22932752e-02
4.74387944e-01 5.61885774e-01 -2.26719171e-01 -4.60967839e-01
3.53154838e-01 7.85901308e-01 7.22714901e-01 -3.78385335e-01
8.46145093e-01 8.44769597e-01 -1.62964374e-01 -1.11596906e+00
-7.44553030e-01 -2.04078153e-01 -9.60976541e-01 -1.96775481e-01
1.13277149e+00 -6.37097538e-01 -1.57829359e-01 7.97495067e-01
-1.26525116e+00 -3.26043427e-01 -1.05031673e-02 7.53584653e-02
-1.99264720e-01 8.19781601e-01 -4.60758328e-01 -4.78388548e-01
-2.82710135e-01 -1.11058617e+00 1.07074273e+00 2.64156967e-01
2.24994600e-01 -9.56413209e-01 -2.58730620e-01 3.66507888e-01
2.44235814e-01 2.26274222e-01 1.04502034e+00 -1.02313221e-01
-7.26371229e-01 3.08014397e-02 -5.54284453e-01 3.77902240e-01
3.86723250e-01 5.06574273e-01 -8.06171775e-01 -1.25478163e-01
-3.76535237e-01 2.29507864e-01 9.69402552e-01 3.54651242e-01
1.35242474e+00 -3.15131485e-01 -1.97013721e-01 5.53144455e-01
1.41156852e+00 5.14121592e-01 8.57457638e-01 4.67302024e-01
1.05209029e+00 4.83435392e-01 3.26121658e-01 4.66767907e-01
1.85089573e-01 4.53757793e-01 3.03171247e-01 -5.12575865e-01
-4.20500547e-01 -3.65876287e-01 3.70585918e-01 5.45895696e-01
3.77457798e-01 -9.17263806e-01 -9.53742504e-01 4.20556664e-01
-1.96841359e+00 -8.83942127e-01 -5.01111031e-01 2.01063561e+00
6.25098526e-01 1.01178259e-01 -4.19418544e-01 9.13790055e-03
1.20944059e+00 1.06580846e-01 -5.52824795e-01 -3.66204798e-01
-4.96583849e-01 -2.35754877e-01 8.24668765e-01 5.39715469e-01
-1.08606529e+00 1.09303403e+00 5.97575140e+00 6.91293597e-01
-1.29007280e+00 -2.60044843e-01 7.22690344e-01 9.00667459e-02
-2.06897736e-01 4.16604578e-02 -5.74032426e-01 4.97394025e-01
6.78073466e-02 1.00221470e-01 5.09615898e-01 5.14232576e-01
4.22911227e-01 -4.02745754e-01 -9.75613236e-01 9.73499775e-01
6.11677170e-01 -1.47066534e+00 6.92282379e-01 -6.51357025e-02
1.11080611e+00 -3.27797800e-01 1.25894070e-01 -3.45746785e-01
1.57278463e-01 -8.34586442e-01 1.30173278e+00 7.11335361e-01
8.04072320e-01 -6.78759217e-01 4.49549675e-01 1.20781451e-01
-1.11697412e+00 1.39964193e-01 -4.41735536e-01 3.12199950e-01
1.26088247e-01 6.33230329e-01 -7.51641035e-01 3.50376636e-01
6.18926883e-01 7.16533482e-01 -1.02729344e+00 8.44928920e-01
-2.58831561e-01 2.44552821e-01 -7.32593015e-02 7.69512355e-02
2.03095719e-01 -3.44830096e-01 3.52933049e-01 1.15424228e+00
1.87530503e-01 -2.09742919e-01 1.43766314e-01 9.25811350e-01
-3.16741914e-01 9.43950936e-02 -3.72134715e-01 -2.84948111e-01
2.57972002e-01 1.04906619e+00 -1.14257634e+00 -4.35421228e-01
-3.63404185e-01 1.67397892e+00 1.07634226e-02 4.12175417e-01
-8.75890493e-01 -5.27017593e-01 1.53709846e-02 7.25950152e-02
4.98026252e-01 -2.25412101e-01 -5.56274474e-01 -1.12138867e+00
2.13972449e-01 -8.72954547e-01 1.18001200e-01 -1.27956700e+00
-1.02146983e+00 5.33615768e-01 -2.79589355e-01 -1.10862708e+00
2.31057078e-01 -7.15554595e-01 -4.64670360e-01 7.12309897e-01
-1.28196156e+00 -1.28529012e+00 -5.90399981e-01 5.94038427e-01
8.61145914e-01 5.51021695e-02 1.49964824e-01 2.90693492e-01
-5.86406350e-01 3.40946645e-01 6.81785226e-01 4.46678400e-01
9.09757674e-01 -1.16942489e+00 3.81972343e-01 1.36216128e+00
7.89793674e-03 3.05501342e-01 6.08564794e-01 -1.00495720e+00
-1.13760316e+00 -1.32272077e+00 5.92091203e-01 -1.73983037e-01
3.92052293e-01 -4.07350183e-01 -7.75415778e-01 6.56924546e-01
4.35012996e-01 -3.90347183e-01 1.07719965e-01 -5.86141527e-01
-9.59582180e-02 -1.32626861e-01 -8.07720423e-01 9.73445952e-01
5.56346595e-01 -3.77667397e-01 -4.12675053e-01 2.24970073e-01
4.32908207e-01 -5.14558017e-01 -2.69754410e-01 8.13761875e-02
5.95228434e-01 -9.74534810e-01 7.08814859e-01 -3.65794078e-02
7.12782204e-01 -7.45916009e-01 -1.02462433e-01 -1.10833907e+00
-2.34962359e-01 -3.02262843e-01 4.48621631e-01 1.34463060e+00
4.40568268e-01 -3.82461488e-01 6.00436389e-01 4.29446220e-01
-1.84875101e-01 -1.87063545e-01 -4.35298800e-01 -2.96847641e-01
-5.42023517e-02 -2.53442645e-01 4.67431426e-01 7.72188783e-01
-6.23055160e-01 2.54044384e-01 -8.23941588e-01 3.04936737e-01
4.04711783e-01 9.26092789e-02 7.83425629e-01 -9.59227502e-01
-1.11213028e-01 -5.39347589e-01 -3.50027084e-02 -1.21634054e+00
-1.96340814e-01 -6.12842023e-01 3.44114780e-01 -1.86738443e+00
3.88544172e-01 -9.93887931e-02 2.10453525e-01 3.86111885e-01
-9.27620977e-02 3.84429604e-01 3.12948316e-01 2.84233749e-01
-3.15730125e-01 4.35816824e-01 1.41922152e+00 -4.65535700e-01
-1.39375210e-01 -5.18735707e-01 -6.43253267e-01 8.25466156e-01
8.44047189e-01 -4.28402394e-01 -3.18101227e-01 -9.81984437e-01
2.74124652e-01 -2.35219091e-01 3.09470475e-01 -9.17409182e-01
3.75640094e-01 -3.10388446e-01 8.37526798e-01 -1.02029204e+00
9.51951370e-02 -8.77515674e-01 1.32668108e-01 4.58022296e-01
-4.43091601e-01 2.80070901e-01 9.66206715e-02 7.07447290e-01
-8.06349367e-02 -5.35803854e-01 1.15686595e+00 -2.96967495e-02
-8.80219042e-01 1.94231290e-02 -8.29229355e-01 -2.70737618e-01
9.15622950e-01 -4.33162153e-01 -5.84433377e-01 -4.70273614e-01
-1.98637679e-01 4.69146930e-02 8.41500163e-01 5.37726343e-01
8.16366494e-01 -1.04713809e+00 -5.70344388e-01 3.00216854e-01
1.92613155e-01 -9.35514569e-02 3.07504624e-01 5.34857750e-01
-1.21598542e+00 2.20677137e-01 -1.43903375e-01 -6.35410130e-01
-1.41522133e+00 5.50877094e-01 3.72034490e-01 9.00269821e-02
-6.63512111e-01 4.23691243e-01 5.11869311e-01 -1.65696949e-01
1.79679409e-01 -4.96172726e-01 -1.43691190e-02 -3.26598398e-02
5.13901949e-01 2.37433970e-01 1.24548441e-02 -6.94531679e-01
-2.59384383e-02 8.25785279e-01 -3.24419886e-01 -2.43782416e-01
1.09442544e+00 -5.61719179e-01 -3.77345949e-01 4.33244109e-01
8.73509884e-01 3.40281039e-01 -1.19357014e+00 -1.74572811e-01
-2.52991170e-01 -5.99968195e-01 2.40235358e-01 -9.95007098e-01
-1.15690446e+00 8.15654814e-01 6.86030447e-01 1.18818998e-01
1.24909329e+00 -2.77328938e-01 6.05262041e-01 6.47029042e-01
-3.03546965e-01 -1.47998798e+00 4.54928279e-01 5.76952100e-01
8.88449967e-01 -1.19231927e+00 8.98682922e-02 -4.64781940e-01
-8.61149549e-01 1.53724396e+00 6.59737170e-01 9.18357372e-02
2.10150793e-01 1.47562757e-01 1.50844499e-01 -3.03412005e-02
-2.91178644e-01 -1.97868407e-01 4.22772020e-01 3.92511785e-01
2.22008809e-01 -1.41159192e-01 7.79602379e-02 -1.20598316e-01
7.05458596e-03 -3.33986491e-01 1.13020766e+00 1.05198574e+00
-6.29025638e-01 -9.14820790e-01 -7.58310199e-01 4.83615160e-01
-3.41369897e-01 -3.37084234e-01 -7.87947536e-01 6.93214655e-01
9.17230323e-02 1.02124095e+00 2.70460248e-01 -1.54909462e-01
1.29360944e-01 -2.66852409e-01 2.90803283e-01 -5.75636089e-01
-3.29834521e-01 3.19501758e-01 -2.35896975e-01 -1.11766919e-01
-4.38644230e-01 -5.44758499e-01 -1.18956065e+00 -4.26698804e-01
-3.00921887e-01 -3.01866323e-01 7.70546913e-01 7.43162096e-01
1.75635636e-01 7.38976061e-01 5.37976861e-01 -5.72862566e-01
9.55211818e-02 -7.44181335e-01 -6.67023540e-01 4.86171633e-01
3.12168777e-01 -1.97584525e-01 -2.10721314e-01 7.69330263e-01] | [11.785948753356934, 1.8597841262817383] |
729ce678-3bd6-4539-9dc9-e4d034c77a66 | tome-a-two-stage-approach-for-model-based | 2305.11161 | null | https://arxiv.org/abs/2305.11161v1 | https://arxiv.org/pdf/2305.11161v1.pdf | TOME: A Two-stage Approach for Model-based Retrieval | Recently, model-based retrieval has emerged as a new paradigm in text retrieval that discards the index in the traditional retrieval model and instead memorizes the candidate corpora using model parameters. This design employs a sequence-to-sequence paradigm to generate document identifiers, which enables the complete capture of the relevance between queries and documents and simplifies the classic indexretrieval-rerank pipeline. Despite its attractive qualities, there remain several major challenges in model-based retrieval, including the discrepancy between pre-training and fine-tuning, and the discrepancy between training and inference. To deal with the above challenges, we propose a novel two-stage model-based retrieval approach called TOME, which makes two major technical contributions, including the utilization of tokenized URLs as identifiers and the design of a two-stage generation architecture. We also propose a number of training strategies to deal with the training difficulty as the corpus size increases. Extensive experiments and analysis on MS MARCO and Natural Questions demonstrate the effectiveness of our proposed approach, and we investigate the scaling laws of TOME by examining various influencing factors. | ['Haifeng Wang', 'Ji-Rong Wen', 'Hua Wu', 'Jing Liu', 'Wayne Xin Zhao', 'Ruiyang Ren'] | 2023-05-18 | null | null | null | null | ['natural-questions'] | ['miscellaneous'] | [ 1.22956291e-01 -4.25043851e-01 -3.43879312e-01 -3.27123404e-02
-1.30516410e+00 -6.85071886e-01 9.93199944e-01 1.11588247e-01
-5.62747061e-01 5.29405236e-01 1.19330630e-01 -4.36210871e-01
-4.04746622e-01 -5.11881232e-01 -3.62039596e-01 -2.53895819e-01
2.15569064e-01 7.47675598e-01 5.55979788e-01 -4.35298949e-01
8.58779907e-01 1.84823051e-01 -1.61974704e+00 3.10012400e-01
1.16118586e+00 9.73704517e-01 3.60979050e-01 8.05949688e-01
-5.69932699e-01 5.52793980e-01 -8.92773092e-01 -4.27824438e-01
1.53300732e-01 -2.23282516e-01 -9.88468945e-01 -4.56636429e-01
3.58788550e-01 -5.26576638e-01 -4.61180806e-01 8.65794420e-01
7.22481012e-01 5.95993325e-02 8.17940712e-01 -9.32678759e-01
-7.77022421e-01 8.03268969e-01 -4.99151736e-01 2.77197480e-01
5.49190104e-01 -1.14868648e-01 1.24084663e+00 -8.81985247e-01
4.58366632e-01 1.25126636e+00 3.71142745e-01 4.91320163e-01
-6.85697556e-01 -4.86292809e-01 4.75522205e-02 3.38593930e-01
-1.54780531e+00 -4.02120799e-01 5.03136396e-01 -2.87169516e-01
9.24811423e-01 4.70979720e-01 2.93961614e-01 8.17707717e-01
-8.00949335e-02 1.00435269e+00 7.04323173e-01 -9.62815285e-01
5.07418551e-02 1.84307292e-01 5.44370174e-01 6.10494792e-01
3.48673165e-01 -1.10466890e-01 -3.75681698e-01 -8.34552109e-01
5.04706323e-01 -1.05582967e-01 -1.22002833e-01 3.51664722e-02
-1.02072084e+00 7.34858632e-01 -1.62480861e-01 2.55456090e-01
-3.43477637e-01 1.82642385e-01 5.13057530e-01 1.98461071e-01
2.19673499e-01 6.85464799e-01 -3.92577916e-01 -1.59273908e-01
-1.16947818e+00 3.26379567e-01 9.02996302e-01 1.26673198e+00
5.86325169e-01 -3.53179663e-01 -4.71326262e-01 9.84221518e-01
4.33011413e-01 5.58129489e-01 9.32806313e-01 -6.03617191e-01
6.60495162e-01 4.71907467e-01 3.77462983e-01 -9.27657723e-01
3.13839205e-02 -2.46785656e-01 -4.22899157e-01 -6.15423322e-01
1.42616972e-01 1.56523332e-01 -7.66900718e-01 1.39212430e+00
2.33015224e-01 -1.86671287e-01 -1.29839465e-01 6.67123735e-01
5.38701594e-01 6.05621338e-01 9.27419215e-02 -1.91145033e-01
1.52008414e+00 -1.17149580e+00 -7.69142210e-01 -8.16545412e-02
7.67284811e-01 -1.11533642e+00 1.17868650e+00 6.13169819e-02
-1.29112399e+00 -5.04875481e-01 -9.23092544e-01 -1.32605031e-01
-4.47864413e-01 1.36751190e-01 6.31061018e-01 5.18653929e-01
-9.99836445e-01 3.97835523e-01 -5.01215816e-01 -3.33579659e-01
-3.08046252e-01 2.20713824e-01 2.27391452e-01 7.62277171e-02
-1.68276346e+00 7.99235165e-01 5.77009380e-01 9.03991982e-02
-5.79483867e-01 -4.04612213e-01 -3.34612876e-01 4.38973486e-01
3.92281741e-01 -7.31248856e-01 1.54409456e+00 -5.28611839e-01
-1.46117198e+00 3.61252367e-01 -4.07721937e-01 -5.01925908e-02
2.78871715e-01 -4.05656904e-01 -4.37140971e-01 3.45998615e-01
-6.95403218e-02 1.26594380e-01 9.08035338e-01 -1.09556627e+00
-6.01514399e-01 8.25310796e-02 6.43512681e-02 2.66403705e-01
-6.44186378e-01 3.74922454e-01 -9.51685250e-01 -6.23319566e-01
1.19018424e-02 -8.04142237e-01 -1.27368674e-01 -4.37230080e-01
-2.87841588e-01 -6.56436443e-01 3.99115831e-01 -4.70609695e-01
2.02337956e+00 -1.80688524e+00 -2.87083864e-01 4.02689695e-01
1.31867617e-01 6.74084544e-01 -4.38108683e-01 1.09439313e+00
4.02845711e-01 4.46789920e-01 1.10430606e-01 -1.61676973e-01
2.57328063e-01 -1.22087948e-01 -6.68299019e-01 -2.19554469e-01
-1.17819710e-02 9.91593122e-01 -1.05977750e+00 -1.11483872e+00
-3.17548007e-01 2.04970971e-01 -4.23357129e-01 4.46758330e-01
-4.74465460e-01 -2.55557179e-01 -9.23918843e-01 6.06063604e-01
6.26989782e-01 -4.34553146e-01 9.71373469e-02 8.12665746e-02
-6.64173206e-03 7.86044061e-01 -1.11083138e+00 1.49537885e+00
-3.08749050e-01 1.23889439e-01 -5.37685724e-03 -5.10780275e-01
8.25599730e-01 4.66168672e-01 2.04206347e-01 -7.59958744e-01
-6.19778335e-02 4.11071301e-01 -3.30814123e-01 -7.77105808e-01
1.30146527e+00 2.45619550e-01 3.30260433e-02 1.06959188e+00
-2.30595335e-01 5.47895990e-02 5.76361239e-01 7.00280309e-01
1.11687732e+00 1.22345246e-01 2.93051060e-02 -7.71160647e-02
6.62434757e-01 2.56186854e-02 1.99881494e-01 1.19017911e+00
2.15847075e-01 4.75509971e-01 2.47598097e-01 -1.03123695e-01
-8.01863134e-01 -5.96693218e-01 1.07544594e-01 1.34978950e+00
1.56460568e-01 -7.25438237e-01 -7.57270575e-01 -6.86881959e-01
-1.70648471e-02 5.38330853e-01 -2.28135824e-01 -3.46271366e-01
-8.21819901e-01 -7.08112538e-01 6.81424618e-01 2.89594889e-01
5.33147193e-02 -1.01400280e+00 -1.94582462e-01 3.67683440e-01
-5.21199107e-01 -7.68688142e-01 -8.94601285e-01 -2.83062339e-01
-8.61112177e-01 -9.12090540e-01 -5.90021551e-01 -8.10646236e-01
6.59293890e-01 6.94351077e-01 1.23577964e+00 7.98888743e-01
-2.00322837e-01 5.15113235e-01 -6.42197728e-01 -2.58197308e-01
-5.45789182e-01 7.88272619e-01 -8.86567608e-02 -3.72789085e-01
4.51731920e-01 -1.43769413e-01 -7.93777347e-01 3.02109927e-01
-1.42640829e+00 -3.19860458e-01 8.30624938e-01 1.04950917e+00
2.88971841e-01 -2.09023222e-01 8.83784890e-01 -9.04912531e-01
1.38001883e+00 -5.55349231e-01 -6.89761758e-01 9.81830060e-01
-1.25458562e+00 4.35149997e-01 6.11682415e-01 -6.35101080e-01
-1.03705406e+00 -4.18469071e-01 -5.88097945e-02 -5.14966361e-02
4.68538702e-01 9.01052594e-01 2.44448096e-01 1.13288477e-01
4.31005448e-01 4.66647118e-01 3.24448501e-03 -8.08274209e-01
5.31032503e-01 1.07921636e+00 1.44501716e-01 -9.65553343e-01
9.59406376e-01 2.44477913e-02 -3.81893903e-01 -4.41114455e-01
-7.81540751e-01 -8.78319800e-01 -5.22202015e-01 3.67431305e-02
1.10305481e-01 -6.80405855e-01 -7.00007319e-01 2.41138339e-01
-1.36972260e+00 2.14700416e-01 -1.55938461e-01 3.40963870e-01
-1.81432188e-01 8.31199527e-01 -8.19508433e-01 -8.12991142e-01
-1.04798746e+00 -9.92074370e-01 1.01632464e+00 1.87854961e-01
-1.68223619e-01 -9.33767796e-01 3.60750526e-01 2.48853669e-01
7.26847351e-01 -6.49933755e-01 1.35011721e+00 -9.22322154e-01
-6.77085578e-01 -5.10991693e-01 -3.70122284e-01 1.35666654e-01
-1.18814021e-01 6.80866838e-02 -6.68873191e-01 -3.46589029e-01
-1.98343396e-01 -3.48301023e-01 6.60161197e-01 -2.32577324e-01
9.65618134e-01 -4.09131676e-01 -3.70377779e-01 1.59050539e-01
1.39126015e+00 1.19358540e-01 5.75012326e-01 5.45348823e-01
3.70385945e-01 4.90248948e-01 7.03897953e-01 4.49121296e-01
4.09466147e-01 5.99947691e-01 1.94923896e-02 5.75107932e-02
6.72672992e-04 -5.82970679e-01 1.28182486e-01 1.62320983e+00
2.55457759e-01 -3.50619972e-01 -5.78221262e-01 5.04966080e-01
-1.79614663e+00 -9.54000890e-01 3.42020169e-02 2.35044551e+00
1.22415257e+00 -3.03419232e-02 -1.19748391e-01 -1.23988964e-01
6.66725338e-01 7.24388286e-02 -3.53855193e-01 -1.83996812e-01
9.80706587e-02 2.86296159e-01 3.28050852e-01 6.61078215e-01
-6.52124941e-01 1.11829782e+00 7.55657482e+00 1.13370836e+00
-9.67104673e-01 -8.10507759e-02 4.50046584e-02 2.82107256e-02
-4.43335623e-01 3.46638948e-01 -1.25082016e+00 5.64382613e-01
1.02724922e+00 -5.92476249e-01 4.32652384e-01 7.90669203e-01
-1.64773166e-02 1.06037922e-01 -9.75827992e-01 6.18709385e-01
2.08125249e-01 -1.12745070e+00 5.91952801e-01 -2.42395312e-01
5.02956927e-01 -1.06442489e-01 -9.19638500e-02 5.81639051e-01
1.75586492e-01 -5.41437626e-01 7.56431818e-01 7.19245732e-01
5.72762668e-01 -4.47746128e-01 5.78799546e-01 4.15126264e-01
-1.13158464e+00 1.10036321e-02 -6.48042679e-01 2.21828803e-01
1.72436461e-01 2.93276966e-01 -9.66273844e-01 6.33041143e-01
2.05013439e-01 9.59415138e-02 -8.02308381e-01 1.15363312e+00
-9.74111930e-02 6.30378962e-01 -2.15930313e-01 -4.16136861e-01
1.05495937e-01 -6.51251674e-02 3.29723001e-01 1.36527205e+00
2.48811573e-01 -1.32933125e-01 1.54967710e-01 5.96204638e-01
-1.39578104e-01 2.51060218e-01 -1.86605960e-01 -8.28709826e-02
1.05432904e+00 1.30537522e+00 -2.49657854e-01 -7.11392522e-01
-2.85192758e-01 6.67407095e-01 4.54362690e-01 5.17419279e-01
-5.49394250e-01 -8.78704429e-01 1.09985292e-01 1.34944186e-01
3.19045484e-01 -2.95407355e-01 2.14963511e-01 -1.19828165e+00
3.80988628e-01 -1.15199685e+00 3.86348993e-01 -5.32817125e-01
-1.26504600e+00 6.79432571e-01 3.25367779e-01 -1.27505505e+00
-7.29402304e-01 -1.26718432e-01 -3.85577768e-01 1.04380536e+00
-1.93472981e+00 -9.27420795e-01 3.84597704e-02 2.47374579e-01
4.69045013e-01 -2.72943806e-02 8.76657844e-01 5.37017047e-01
-4.13979560e-01 9.38311040e-01 6.85472071e-01 4.57173958e-02
9.31971729e-01 -9.21583891e-01 6.12659752e-01 6.88640296e-01
-1.99955069e-02 1.40045500e+00 3.24302733e-01 -4.66528416e-01
-1.71987987e+00 -7.56251931e-01 1.31202984e+00 -4.34420854e-01
8.34355414e-01 -2.95215994e-01 -1.07961714e+00 2.92655110e-01
1.84956491e-01 -4.93377894e-01 6.44110441e-01 6.99281320e-02
-5.96620202e-01 -1.64824709e-01 -7.37901151e-01 6.35198057e-01
6.98281765e-01 -7.72860527e-01 -7.72626400e-01 4.30845708e-01
8.97678196e-01 -9.31145102e-02 -8.41023684e-01 3.69366147e-02
8.10048580e-01 -4.97291237e-01 9.46409583e-01 -6.01510167e-01
6.30741268e-02 -1.59201279e-01 1.23842858e-01 -9.00992274e-01
-3.62165630e-01 -1.07773089e+00 -4.73833829e-01 1.43764818e+00
4.42087173e-01 -7.14704692e-01 5.28161585e-01 7.82128274e-01
5.16788550e-02 -7.67211914e-01 -6.85106099e-01 -8.38864684e-01
1.40493229e-01 1.09483227e-02 9.88802433e-01 4.96066570e-01
1.84694037e-01 5.37226975e-01 -3.12612385e-01 -2.38687560e-01
3.20328355e-01 1.15697123e-01 6.89124703e-01 -1.24785221e+00
-4.27061051e-01 -4.52573001e-01 2.94191122e-01 -1.87288058e+00
-8.16826615e-03 -8.04714262e-01 1.33302435e-01 -1.46882308e+00
5.04387558e-01 -7.14063823e-01 -4.48186755e-01 1.30210117e-01
-6.21445239e-01 -3.32709730e-01 6.07319735e-02 9.08373296e-01
-1.02094591e+00 5.41503668e-01 1.17556083e+00 -1.15000062e-01
-6.32509738e-02 -4.29918990e-02 -8.43608081e-01 1.48147061e-01
5.06163061e-01 -7.11683571e-01 -5.96270680e-01 -8.23305607e-01
4.19095188e-01 2.07819268e-01 -1.23803720e-01 -5.66915035e-01
6.10716403e-01 -6.18751161e-02 -1.40805557e-01 -8.56828928e-01
1.14417277e-01 -6.06242478e-01 -3.36903781e-01 3.44621360e-01
-8.34982693e-01 6.65910721e-01 -1.58505902e-01 5.89897394e-01
-2.30658039e-01 -8.57309818e-01 3.14994812e-01 -2.00990796e-01
-3.37295920e-01 2.98377395e-01 -4.48475331e-01 3.27651650e-01
2.98677593e-01 9.98331383e-02 -5.73172212e-01 -3.25002879e-01
1.52822375e-01 3.44070226e-01 3.15780580e-01 5.23518205e-01
5.12476742e-01 -1.18917692e+00 -6.51039302e-01 2.08228856e-01
2.52667934e-01 -2.33887166e-01 -2.45715175e-02 3.81477326e-01
-4.65987116e-01 9.71177757e-01 4.21436757e-01 -2.05638438e-01
-1.10066116e+00 5.80621719e-01 -1.70135677e-01 -9.65029657e-01
-1.58381388e-01 4.43993032e-01 -1.93445131e-01 -2.88536519e-01
3.60986382e-01 1.58078317e-02 -3.42170060e-01 3.43473926e-02
8.11180055e-01 2.86852688e-01 2.57245213e-01 -2.83999711e-01
5.21394350e-02 3.18241864e-01 -8.58735800e-01 -3.91605884e-01
8.74313772e-01 -3.54639947e-01 -4.42924947e-01 1.12522505e-01
1.19792640e+00 4.23631072e-03 -5.06007910e-01 -6.34224117e-01
5.59782445e-01 -4.34194803e-01 -8.99097174e-02 -8.23218286e-01
-4.98523384e-01 7.87174284e-01 3.37959886e-01 4.42873746e-01
8.96546781e-01 -3.64108592e-01 1.37582016e+00 7.96131551e-01
4.03945386e-01 -1.29987991e+00 1.49342760e-01 9.72649217e-01
6.50004327e-01 -8.62275839e-01 1.66425645e-01 -2.01960001e-02
-2.00226873e-01 9.83687997e-01 4.13056731e-01 2.73283988e-01
5.32162070e-01 -1.71872810e-01 3.40049773e-01 -1.29846841e-01
-1.05087543e+00 -1.16793290e-02 4.09762144e-01 2.84895569e-01
6.20370865e-01 -3.92655879e-01 -8.96622479e-01 4.19532806e-01
-1.90971456e-02 2.72350349e-02 1.66433945e-01 1.32083809e+00
-4.85987872e-01 -1.70657134e+00 -2.65552849e-01 5.27537704e-01
-6.52923048e-01 -6.85647190e-01 -3.84901971e-01 6.54105961e-01
-6.00576043e-01 1.03742719e+00 -2.40252316e-01 -3.98030102e-01
2.49447688e-01 3.45574379e-01 1.84021190e-01 -5.94359934e-01
-1.01432812e+00 1.13833010e-01 2.62493198e-03 -1.78293630e-01
-9.17258486e-02 -3.25353593e-01 -8.93154263e-01 -1.22689061e-01
-1.01382673e+00 9.50914502e-01 8.52641761e-01 7.27716148e-01
8.66865754e-01 1.13544278e-01 9.44537997e-01 -4.59357202e-01
-1.52343440e+00 -1.32926404e+00 -3.93193871e-01 4.02183950e-01
1.65939748e-01 -4.14174706e-01 -4.96765494e-01 -1.57236367e-01] | [11.461852073669434, 7.636738300323486] |
622d6fc9-e91b-4be5-b25d-1f854bdc93de | implicit-autoencoder-for-point-cloud-self | 2201.00785 | null | https://arxiv.org/abs/2201.00785v4 | https://arxiv.org/pdf/2201.00785v4.pdf | Implicit Autoencoder for Point Cloud Self-supervised Representation Learning | This paper advocates the use of implicit surface representation in autoencoder-based self-supervised 3D representation learning. The most popular and accessible 3D representation, i.e., point clouds, involves discrete samples of the underlying continuous 3D surface. This discretization process introduces sampling variations on the 3D shape, making it challenging to develop transferable knowledge of the true 3D geometry. In the standard autoencoding paradigm, the encoder is compelled to encode not only the 3D geometry but also information on the specific discrete sampling of the 3D shape into the latent code. This is because the point cloud reconstructed by the decoder is considered unacceptable unless there is a perfect mapping between the original and the reconstructed point clouds. This paper introduces the Implicit AutoEncoder (IAE), a simple yet effective method that addresses the sampling variation issue by replacing the commonly-used point-cloud decoder with an implicit decoder. The implicit decoder reconstructs a continuous representation of the 3D shape, independent of the imperfections in the discrete samples. Extensive experiments demonstrate that the proposed IAE achieves state-of-the-art performance across various self-supervised learning benchmarks. | ['Gang Hua', 'Hao Kang', 'Chen Song', 'QiXing Huang', 'Li Guan', 'Haoxiang Li', 'Zhenpei Yang', 'Siming Yan'] | 2022-01-03 | null | null | null | null | ['3d-point-cloud-linear-classification'] | ['computer-vision'] | [-3.08941193e-02 3.07092994e-01 8.35236385e-02 -3.30941468e-01
-6.52828097e-01 -4.09144431e-01 5.89247286e-01 -2.18635835e-02
-2.01843549e-02 3.22835773e-01 3.61689366e-02 4.50704619e-02
2.90142626e-01 -1.10965419e+00 -1.37405968e+00 -8.08161318e-01
1.39223918e-01 9.74456966e-01 -1.68880165e-01 -3.89932096e-02
3.45324501e-02 7.13686705e-01 -1.86903679e+00 1.33360684e-01
7.02171683e-01 1.12023306e+00 3.99765313e-01 3.10893118e-01
-6.13728940e-01 2.57676572e-01 -6.25265718e-01 -1.76523671e-01
4.11816806e-01 -1.49104923e-01 -3.52939218e-01 4.23861235e-01
3.94954920e-01 -4.31644499e-01 -3.01259696e-01 1.18609202e+00
-2.86582368e-03 -3.29619259e-01 1.13448048e+00 -9.51756001e-01
-7.89461851e-01 6.42483681e-02 -1.62362114e-01 -4.65754837e-01
2.19029427e-01 -2.86271960e-01 7.56172597e-01 -1.01922882e+00
4.70845193e-01 1.07130456e+00 8.18506598e-01 2.73114264e-01
-1.31632364e+00 -3.37736428e-01 -2.59398818e-01 -1.58190340e-01
-1.70921028e+00 -1.88214079e-01 1.33715427e+00 -7.78636634e-01
8.91453385e-01 4.66321893e-02 9.03627455e-01 1.08447933e+00
3.04392248e-01 5.09981990e-01 1.09154022e+00 -5.74388504e-01
7.26963460e-01 5.26363313e-01 -2.40575477e-01 5.97169161e-01
2.31422395e-01 2.14461908e-01 -2.81862140e-01 -3.20108801e-01
1.35691488e+00 2.07844660e-01 -1.09339297e-01 -9.01130974e-01
-8.41740489e-01 8.56339335e-01 5.03007710e-01 1.40234560e-01
-6.42330885e-01 2.68825274e-02 6.62209094e-03 3.57743621e-01
8.12995434e-01 1.33917794e-01 -3.72311264e-01 -5.17577380e-02
-9.37075973e-01 -6.83985511e-03 7.93368280e-01 1.21174157e+00
1.26012695e+00 4.29634094e-01 5.08008659e-01 5.02954900e-01
5.59801459e-01 5.96324086e-01 5.00514030e-01 -9.23657835e-01
1.37250394e-01 6.72607422e-01 1.02286071e-01 -1.05185580e+00
3.14075232e-01 -4.72009003e-01 -9.51582253e-01 5.74500442e-01
-7.64612854e-02 1.66000783e-01 -8.51936936e-01 1.36476529e+00
3.61594379e-01 4.27736998e-01 2.08392173e-01 7.67212212e-01
6.97053254e-01 7.99620569e-01 -5.19558787e-01 7.55686685e-02
8.04082870e-01 -3.32486689e-01 -5.34409940e-01 -1.84300706e-01
2.16837525e-01 -3.53741914e-01 6.81775093e-01 1.15251906e-01
-1.04037964e+00 -8.13360453e-01 -1.41016972e+00 -1.25504509e-01
-2.63698399e-01 3.33717614e-02 4.31405038e-01 3.02833945e-01
-1.04305387e+00 8.60272765e-01 -1.11063147e+00 -1.01753898e-01
3.59197706e-01 2.92148501e-01 -6.37340665e-01 9.09929499e-02
-8.08488846e-01 9.13023472e-01 1.56907916e-01 -1.67647302e-01
-7.94130385e-01 -6.49863780e-01 -9.56691802e-01 -4.30049822e-02
-1.55705914e-01 -7.06854880e-01 9.36769426e-01 -1.08514965e+00
-1.57666719e+00 9.73674357e-01 -2.73641825e-01 -3.66003931e-01
2.42351130e-01 -1.78490579e-01 -9.23005193e-02 6.61152825e-02
-1.05792228e-02 5.43528855e-01 1.36852431e+00 -1.71305668e+00
7.07379133e-02 -5.81673026e-01 -2.10372090e-01 2.63117313e-01
-5.01889177e-02 -9.41345334e-01 -2.88091838e-01 -4.92167085e-01
5.64912617e-01 -8.00607264e-01 4.95171435e-02 2.99973696e-01
-2.35989645e-01 -3.01051233e-02 8.62387657e-01 -4.86129075e-01
3.69830549e-01 -2.21098161e+00 2.03603119e-01 1.11142859e-01
8.11726823e-02 -5.21083362e-02 -3.78842391e-02 4.80278999e-01
-1.20183237e-01 -1.31819397e-01 -3.85347992e-01 -6.89882219e-01
-7.76488632e-02 6.05125606e-01 -5.17400146e-01 7.22254038e-01
4.33091640e-01 6.85154319e-01 -7.78978825e-01 -1.98280469e-01
4.58132863e-01 9.58819926e-01 -6.01065099e-01 5.35746992e-01
-3.43863308e-01 5.11382222e-01 -5.55217624e-01 4.17178512e-01
7.97045290e-01 -4.17802989e-01 -3.73090714e-01 -1.84404984e-01
-9.85944867e-02 3.72629642e-01 -1.07977569e+00 2.14318037e+00
-5.66475809e-01 4.56765473e-01 -1.52181298e-01 -1.11845803e+00
1.47494555e+00 5.70999444e-01 5.44597626e-01 -2.37237260e-01
1.10107854e-01 3.20524812e-01 -4.55042601e-01 -7.15533867e-02
3.62529427e-01 -4.58288729e-01 1.55535012e-01 3.23633134e-01
2.82707304e-01 -8.10077548e-01 -9.57399487e-01 -2.40614280e-01
7.75202990e-01 3.71824950e-01 3.04120451e-01 -9.59913433e-02
2.93324828e-01 -1.09847644e-02 3.92011225e-01 4.58222926e-01
3.62878591e-01 9.95815337e-01 1.13615386e-01 -5.68011522e-01
-1.50087333e+00 -1.29263341e+00 -3.61565769e-01 -1.67266712e-01
-7.98581261e-03 -1.13781013e-01 -7.24775016e-01 -4.09464538e-01
3.20815653e-01 8.87566984e-01 -8.26968312e-01 -2.79106349e-01
-1.35651931e-01 5.31732030e-02 1.32399082e-01 3.30318838e-01
4.45122689e-01 -5.74799359e-01 -6.88058794e-01 2.06776381e-01
1.25720561e-01 -1.11850667e+00 3.66528295e-02 2.96232104e-01
-1.37397790e+00 -6.85334682e-01 -4.80204791e-01 -7.33432114e-01
9.47592854e-01 1.51045784e-01 1.11978936e+00 -1.46607295e-01
2.02210024e-01 5.20321310e-01 -3.27495873e-01 -4.62211341e-01
-7.35795557e-01 -3.10111851e-01 1.17856093e-01 2.55120426e-01
5.81664503e-01 -1.01375842e+00 -6.68528825e-02 -3.03046145e-02
-7.20418751e-01 1.97386950e-01 4.68200177e-01 9.17188942e-01
1.24984121e+00 1.21206798e-01 1.98006719e-01 -6.85380638e-01
-2.10654978e-02 -5.30636787e-01 -6.30415022e-01 -1.98889256e-01
-3.27085942e-01 4.41881806e-01 6.96556926e-01 -3.33829224e-01
-7.30195940e-01 3.01994443e-01 -3.88921142e-01 -1.25890100e+00
-4.75382239e-01 3.52556914e-01 -2.64982224e-01 -1.29202828e-01
6.44024074e-01 7.29359865e-01 5.08256912e-01 -8.86976242e-01
1.68951645e-01 9.06580687e-01 4.75335717e-01 -5.19198239e-01
9.50613976e-01 7.51495361e-01 4.02009599e-02 -1.21684718e+00
-6.81968331e-01 -2.18380675e-01 -1.06260979e+00 4.88256626e-02
8.23243320e-01 -1.16116679e+00 -2.44190887e-01 2.87059188e-01
-1.48229170e+00 -2.38623098e-01 -8.28390658e-01 4.33140069e-01
-9.22173202e-01 2.89734155e-01 -2.31873959e-01 -7.99489617e-01
-2.58194596e-01 -1.14699614e+00 1.57023168e+00 -3.26321661e-01
-6.65880665e-02 -9.89584386e-01 1.65054381e-01 -4.85724546e-02
1.48652941e-01 4.74390805e-01 1.00523484e+00 -3.60640883e-01
-6.94249034e-01 -5.39745569e-01 2.42158547e-01 8.60954404e-01
3.34710568e-01 -2.40734771e-01 -1.19619215e+00 -2.08677769e-01
7.01485157e-01 -1.24890849e-01 4.07329679e-01 3.27450782e-01
1.23152113e+00 -3.73192251e-01 -9.46392491e-02 8.60896587e-01
1.63533962e+00 -3.32078747e-02 5.50648034e-01 6.77725300e-02
6.03868604e-01 3.43459249e-01 7.43495449e-02 5.36580741e-01
4.35905010e-01 5.39882720e-01 7.42947638e-01 1.93239659e-01
-8.38612989e-02 -7.77818799e-01 1.78379133e-01 1.50233781e+00
-3.15794982e-02 1.40727192e-01 -8.14396381e-01 4.06776845e-01
-1.31844330e+00 -6.78079426e-01 2.73328364e-01 2.45090270e+00
7.27524459e-01 1.84369870e-02 -4.99351054e-01 4.34449106e-01
5.15259326e-01 1.15642078e-01 -7.95272529e-01 -3.33487213e-01
-4.28611301e-02 3.71021301e-01 2.66992956e-01 4.79996204e-01
-8.97383630e-01 6.13062143e-01 5.87160301e+00 3.90584946e-01
-1.32767987e+00 5.78082949e-02 1.50987610e-01 2.91188657e-01
-5.03370285e-01 -5.15496209e-02 -5.89962065e-01 5.35532534e-01
9.88358676e-01 -9.35316533e-02 4.36386943e-01 1.01014006e+00
-1.27507314e-01 3.57603908e-01 -1.41433942e+00 1.08753395e+00
1.03459477e-01 -1.46376407e+00 3.31968784e-01 2.05669358e-01
6.16643667e-01 1.21921591e-01 3.39284018e-02 1.57776400e-01
-5.74059896e-02 -8.89981389e-01 1.01445246e+00 8.51073265e-01
1.21731091e+00 -6.26384377e-01 6.29118025e-01 7.57429659e-01
-8.22799385e-01 3.33597004e-01 -5.88461339e-01 -2.23698720e-01
-1.07484937e-01 8.72600555e-01 -9.31958556e-01 5.09755611e-01
6.55483425e-01 9.25778806e-01 -1.33174837e-01 6.48058951e-01
-8.72058496e-02 5.49554527e-01 -5.29979289e-01 1.96369171e-01
5.04203551e-02 -4.16410327e-01 9.11006331e-01 6.46273255e-01
6.50778949e-01 1.30032018e-01 5.81622720e-02 1.36133945e+00
-3.33232917e-02 -2.22992748e-01 -1.10753918e+00 -1.32514566e-01
5.82137823e-01 5.16968191e-01 -1.61994040e-01 -1.63119569e-01
-4.28157389e-01 9.42698419e-01 4.02573645e-01 3.31671089e-01
-2.35837370e-01 -7.23649710e-02 8.00890923e-01 2.51849115e-01
6.86058700e-01 -5.16688287e-01 -5.05375326e-01 -9.79317725e-01
1.15120128e-01 -5.91083527e-01 -3.05612445e-01 -9.97349143e-01
-1.41141021e+00 6.38291717e-01 -9.25254524e-02 -1.70148432e+00
-5.15901148e-01 -5.58859885e-01 -3.32903892e-01 1.05562806e+00
-1.40192294e+00 -8.88243437e-01 -3.08494925e-01 4.51418817e-01
3.73039186e-01 -2.57108063e-01 1.34761620e+00 -4.61044237e-02
1.91852793e-01 3.06724191e-01 3.40865642e-01 9.14698318e-02
8.96190554e-02 -1.25094628e+00 6.02142155e-01 1.82964921e-01
4.42299038e-01 4.07875925e-01 6.91090167e-01 -6.71437740e-01
-1.89516163e+00 -1.19760418e+00 5.15375316e-01 -5.26074469e-01
1.93384051e-01 -4.91365403e-01 -1.31184208e+00 8.26608062e-01
-2.37334624e-01 2.91974992e-01 4.97367173e-01 -2.81458437e-01
-3.01931620e-01 -1.28000276e-02 -1.24170375e+00 1.51680410e-01
7.20411301e-01 -9.41958666e-01 -9.45243001e-01 1.44201785e-01
8.64495873e-01 -6.85180604e-01 -1.00931954e+00 2.96490520e-01
3.07314992e-01 -9.43410397e-01 1.02403855e+00 -2.36941949e-01
6.96290493e-01 -2.71223783e-01 -6.89344764e-01 -1.50158143e+00
-1.81035325e-01 -1.26696959e-01 -6.30817711e-01 8.22967649e-01
1.10769179e-02 -6.03118241e-01 1.02749193e+00 3.37941736e-01
-3.31548631e-01 -9.28720951e-01 -1.25723433e+00 -8.33878696e-01
4.04286176e-01 -5.99304914e-01 1.01474321e+00 1.05077362e+00
-4.86431926e-01 1.56765133e-01 -8.67830291e-02 5.54824531e-01
8.90361905e-01 2.91472465e-01 7.29000151e-01 -1.51739371e+00
-3.67806643e-01 3.59153636e-02 -9.29265857e-01 -1.46742702e+00
3.99866700e-01 -1.07369518e+00 -3.62782516e-02 -1.33208823e+00
-3.35428536e-01 -5.88597059e-01 1.04490936e-01 8.37810412e-02
5.70902109e-01 -1.15803413e-01 -8.87494460e-02 3.47739220e-01
1.56242698e-01 1.13349986e+00 1.23259449e+00 -1.44431502e-01
-1.26618609e-01 -1.53221861e-02 -2.90906847e-01 8.10614288e-01
4.83938783e-01 -5.29727221e-01 -3.22291523e-01 -8.86296690e-01
-2.04847157e-02 2.34050751e-01 5.06555617e-01 -1.15028489e+00
3.04281026e-01 1.06429093e-01 5.96138656e-01 -1.03412819e+00
8.20542336e-01 -1.29035687e+00 4.62948322e-01 -7.38994102e-04
-7.44615495e-02 -1.35482565e-01 2.16881186e-01 8.31601799e-01
-3.05790603e-01 -2.05177188e-01 5.53535104e-01 -2.71878034e-01
-2.27366179e-01 5.31378865e-01 1.81644950e-02 -1.93987012e-01
5.66264570e-01 -6.72803164e-01 4.43761766e-01 -3.61705959e-01
-4.92315084e-01 -4.47886616e-01 9.61742759e-01 3.29551548e-01
9.31098819e-01 -1.59676552e+00 -5.96433103e-01 9.65317428e-01
1.43778697e-01 7.12993622e-01 1.97312422e-02 2.12399457e-02
-4.19387490e-01 3.41964960e-01 -1.25959367e-01 -1.00792909e+00
-4.89292115e-01 4.26548272e-01 4.47498143e-01 3.98020387e-01
-1.08554912e+00 5.28734207e-01 2.74148043e-02 -5.71934879e-01
1.29843906e-01 -3.34196538e-01 2.36653939e-01 -3.84912610e-01
1.37671709e-01 -6.18972704e-02 2.17403337e-01 -9.12181318e-01
1.89952599e-03 7.33609736e-01 2.53529191e-01 -2.75849900e-03
1.54942501e+00 1.36820629e-01 -1.10152565e-01 1.21907854e+00
1.58538103e+00 -1.97340786e-01 -1.61503100e+00 -4.77726400e-01
-3.93543720e-01 -5.30166924e-01 5.41746140e-01 -2.98941374e-01
-9.01065171e-01 1.09103942e+00 4.77046430e-01 2.21577093e-01
7.16373503e-01 -2.10009888e-02 6.65616930e-01 3.56689006e-01
7.97229886e-01 -6.95673108e-01 -2.63026446e-01 6.06466711e-01
1.08612120e+00 -1.10719550e+00 8.60231146e-02 -1.39781386e-01
-3.72640848e-01 1.14066732e+00 2.11256474e-01 -6.50696814e-01
1.05064440e+00 1.47602156e-01 -1.11952260e-01 -3.30584079e-01
-6.83117032e-01 2.91977137e-01 3.94864768e-01 7.51834154e-01
5.41608259e-02 1.84561625e-01 6.62514806e-01 5.61536610e-01
-5.81579566e-01 -9.09084529e-02 3.66898626e-01 7.65490770e-01
-2.76099473e-01 -8.27221692e-01 -5.25819242e-01 4.54174101e-01
2.14681551e-01 2.13363156e-01 -2.81073421e-01 7.40238249e-01
1.10401101e-01 2.66595334e-01 7.18905985e-01 -4.50241327e-01
2.66094595e-01 2.47240692e-01 5.72454035e-01 -7.56006718e-01
9.79336072e-03 -1.77820146e-01 -5.60968757e-01 -3.61085773e-01
-2.18222171e-01 -6.97732508e-01 -1.36076152e+00 -9.93289649e-02
-1.78362340e-01 2.81189740e-01 1.13579917e+00 8.30510020e-01
5.57160974e-01 1.63754001e-01 9.30990398e-01 -1.21015775e+00
-7.78509855e-01 -7.98758626e-01 -8.01188767e-01 3.74362499e-01
9.05413866e-01 -9.47604537e-01 -5.77710629e-01 3.52631621e-02] | [8.750897407531738, -3.6044998168945312] |
d5b19f41-eefd-453d-8ce4-bd2a0dbe27cc | a-new-lda-formulation-with-covariates | 2202.11527 | null | https://arxiv.org/abs/2202.11527v1 | https://arxiv.org/pdf/2202.11527v1.pdf | A new LDA formulation with covariates | The Latent Dirichlet Allocation (LDA) model is a popular method for creating mixed-membership clusters. Despite having been originally developed for text analysis, LDA has been used for a wide range of other applications. We propose a new formulation for the LDA model which incorporates covariates. In this model, a negative binomial regression is embedded within LDA, enabling straight-forward interpretation of the regression coefficients and the analysis of the quantity of cluster-specific elements in each sampling units (instead of the analysis being focused on modeling the proportion of each cluster, as in Structural Topic Models). We use slice sampling within a Gibbs sampling algorithm to estimate model parameters. We rely on simulations to show how our algorithm is able to successfully retrieve the true parameter values and the ability to make predictions for the abundance matrix using the information given by the covariates. The model is illustrated using real data sets from three different areas: text-mining of Coronavirus articles, analysis of grocery shopping baskets, and ecology of tree species on Barro Colorado Island (Panama). This model allows the identification of mixed-membership clusters in discrete data and provides inference on the relationship between covariates and the abundance of these clusters. | ['Denis Valle', 'Rafael Izbicki', 'Gilson Shimizu'] | 2022-02-18 | null | null | null | null | ['topic-models'] | ['natural-language-processing'] | [ 4.32641059e-02 -8.03508535e-02 -1.91572860e-01 -3.30676109e-01
-3.55903894e-01 -3.52039784e-01 7.63579547e-01 5.10522902e-01
-2.45091230e-01 4.49026078e-01 2.95863867e-01 -3.64581078e-01
-1.71754852e-01 -1.02240229e+00 -4.15606350e-01 -9.87537622e-01
-5.64014196e-01 9.15420890e-01 8.59654844e-02 3.42591763e-01
2.21788123e-01 2.79061228e-01 -1.49520946e+00 -4.10595946e-02
6.44072473e-01 3.43486637e-01 4.48366821e-01 4.30662334e-01
-2.51877457e-01 2.25033984e-01 -4.65719819e-01 -9.29866210e-02
2.12309062e-02 -3.61837536e-01 -2.54178166e-01 3.23691458e-01
-1.55234426e-01 -1.79649502e-01 2.15973571e-01 8.37867022e-01
5.59747107e-02 -4.75286245e-02 1.22749448e+00 -1.38109195e+00
7.54456520e-02 6.98555112e-01 -1.03383183e+00 4.60810140e-02
1.64808258e-01 -1.61234692e-01 1.10712004e+00 -5.50118089e-01
5.74572921e-01 1.42261481e+00 6.67679012e-01 -2.76736796e-01
-1.64436078e+00 -6.41808271e-01 -5.11360653e-02 6.34650961e-02
-1.56014085e+00 -1.24371119e-01 5.14980853e-01 -8.49629521e-01
4.97151583e-01 5.86182289e-02 8.32778871e-01 7.27589905e-01
1.95415571e-01 7.35679090e-01 8.84149075e-01 -4.69052315e-01
6.34419739e-01 1.41425759e-01 1.44706905e-01 2.89637625e-01
4.08813030e-01 -2.74146318e-01 -3.28660190e-01 -9.12337303e-01
5.61471879e-01 2.67139494e-01 1.24796428e-01 -6.97726250e-01
-8.97852063e-01 1.44868958e+00 9.65717956e-02 1.48366198e-01
-5.56045651e-01 2.13313773e-02 2.88985759e-01 -2.11535648e-01
9.76331770e-01 -1.24991760e-01 -1.57517344e-01 1.76169313e-02
-1.20740664e+00 3.76711041e-01 9.38726127e-01 5.83042562e-01
7.42576540e-01 -4.04982954e-01 1.47012100e-01 8.66189480e-01
7.62891233e-01 4.82066274e-01 3.17266732e-01 -6.87105477e-01
8.64371061e-02 5.57490766e-01 3.11275750e-01 -1.08037472e+00
-3.18011761e-01 8.22721794e-02 -6.14502847e-01 -6.95261434e-02
5.38416207e-01 -1.87951311e-01 -7.43350565e-01 1.68020308e+00
6.20713234e-01 1.64913669e-01 -5.07704258e-01 4.57433611e-01
8.64506587e-02 9.46215391e-01 4.17460769e-01 -5.08871078e-01
1.40549958e+00 -2.17658743e-01 -6.15306854e-01 -9.15882885e-02
5.37632823e-01 -4.68293279e-01 8.38752925e-01 3.52942675e-01
-6.68785751e-01 -1.10142320e-01 -6.00744605e-01 3.46895695e-01
-3.21840376e-01 -1.69104710e-01 6.19766414e-01 6.59116805e-01
-7.47788548e-01 2.03889161e-01 -1.06021428e+00 -9.01466429e-01
2.81225801e-01 1.90266386e-01 3.91679108e-02 -1.18831731e-01
-8.93440962e-01 5.34265637e-01 -2.79416726e-03 -1.86218947e-01
-6.01464748e-01 -5.81244409e-01 -5.13223708e-01 3.28921974e-01
1.04434051e-01 -5.21775424e-01 7.28734255e-01 -6.19711459e-01
-9.55887794e-01 7.51709878e-01 -2.92516321e-01 -3.70970339e-01
2.77855724e-01 2.70592093e-01 9.32711661e-02 8.70603044e-03
4.01793718e-01 6.45689130e-01 6.99462354e-01 -1.11485863e+00
-6.31906390e-01 -6.72691882e-01 -4.58564341e-01 8.21498781e-02
-2.39946648e-01 2.40413193e-02 -2.44050071e-01 -5.87976098e-01
2.70202637e-01 -1.05571663e+00 -3.87414992e-01 -1.94129750e-01
-1.69086024e-01 -3.67365956e-01 5.09890318e-01 -6.05987549e-01
1.31961071e+00 -2.38615751e+00 1.07830033e-01 4.33965802e-01
2.30217054e-01 -4.25587982e-01 1.19171254e-01 8.12493563e-01
1.71888068e-01 2.11541429e-02 -6.56693339e-01 -3.44814777e-01
-1.99053194e-02 2.43052587e-01 -2.18368724e-01 8.37285697e-01
-7.78519064e-02 1.82346001e-01 -7.35135376e-01 -4.87788945e-01
1.90601423e-01 5.14286518e-01 -6.06751442e-01 1.65587291e-02
-1.79193452e-01 1.36841759e-01 -1.72213048e-01 2.44897828e-01
6.82326317e-01 -2.00588495e-01 7.52416730e-01 3.43532056e-01
-1.76874474e-01 1.09164223e-01 -1.26296723e+00 9.32695746e-01
-1.90697193e-01 5.70287645e-01 3.16245884e-01 -9.25805211e-01
9.11688209e-01 9.20045301e-02 8.55591476e-01 7.81948492e-03
5.81411794e-02 -1.01100080e-01 8.25507864e-02 -2.10495383e-01
3.73131365e-01 -3.35434079e-01 -1.75114498e-01 9.87652302e-01
-2.83095568e-01 -1.59573838e-01 3.72841567e-01 2.73892552e-01
9.18766081e-01 -2.12749973e-01 5.63370287e-01 -4.97562826e-01
-2.82771252e-02 2.68990040e-01 3.81851345e-01 5.33850133e-01
3.38662267e-02 3.23209226e-01 7.32380450e-01 -3.30752879e-01
-1.34213412e+00 -1.09233642e+00 -5.62013924e-01 1.16294646e+00
-2.14116439e-01 -4.09154266e-01 -6.31641626e-01 -6.00000098e-02
1.84935227e-01 7.92809248e-01 -8.90118718e-01 5.97939603e-02
-4.58106399e-02 -1.42957652e+00 5.35353161e-02 2.44985595e-02
-4.78754863e-02 -9.55840707e-01 -5.82876921e-01 2.17190161e-01
-2.07859248e-01 -6.09637678e-01 -1.52759314e-01 3.57845783e-01
-9.38845754e-01 -1.14509773e+00 -4.69181359e-01 -5.48507035e-01
7.69066274e-01 1.59226358e-01 7.87447274e-01 -2.46029064e-01
-4.09532309e-01 3.66190910e-01 -3.24044079e-01 -4.10369307e-01
-5.32633722e-01 5.44855334e-02 1.68691263e-01 8.63346756e-02
9.61468637e-01 -6.60369635e-01 -3.30977142e-01 4.50278014e-01
-1.06725323e+00 -1.75364807e-01 2.71511018e-01 6.72263741e-01
3.31964493e-01 6.44947886e-02 5.42457640e-01 -8.82821918e-01
7.03127861e-01 -1.26363230e+00 -6.04700029e-01 2.36794278e-02
-4.97611374e-01 -2.59191573e-01 -1.57825425e-02 -5.46980858e-01
-7.36729503e-01 1.71450049e-01 3.19980949e-01 -1.74289912e-01
-2.67195821e-01 7.25054801e-01 -3.53388414e-02 6.20548964e-01
4.81569707e-01 -3.28161865e-02 3.35248917e-01 -6.11318469e-01
3.68173212e-01 9.70432580e-01 3.20832729e-02 -4.23011631e-01
2.90563703e-01 7.72359908e-01 -9.59921554e-02 -1.27124012e+00
-3.68179977e-01 -8.85550320e-01 -8.49017680e-01 -2.26775795e-01
7.26380527e-01 -9.05986011e-01 -6.95650935e-01 2.86207259e-01
-7.12190807e-01 -4.43817258e-01 -1.96259066e-01 5.52559733e-01
-5.67309678e-01 4.49072152e-01 -4.21091199e-01 -1.01223063e+00
4.42067869e-02 -9.41766858e-01 9.44373369e-01 -2.92311668e-01
-5.49706399e-01 -1.25122738e+00 5.84534585e-01 9.26921815e-02
9.17563587e-02 1.70452848e-01 1.23369861e+00 -7.57462502e-01
-1.36539862e-01 -2.89609283e-01 -1.17379747e-01 -1.70247987e-01
2.57420421e-01 3.88660282e-01 -6.19093716e-01 -3.41592312e-01
1.76854834e-01 2.41567075e-01 6.85995102e-01 1.04340971e+00
7.02575028e-01 -2.34026402e-01 -7.47370362e-01 1.36381015e-01
1.23057497e+00 2.43513510e-02 5.53434491e-01 8.95310417e-02
3.51339996e-01 1.00506425e+00 4.80012596e-01 9.06616986e-01
3.80722612e-01 7.16488123e-01 2.46038407e-01 8.53281766e-02
3.92177999e-01 -6.53415397e-02 2.59229869e-01 6.57096386e-01
4.43646640e-01 -2.18193889e-01 -1.27084172e+00 8.64745736e-01
-1.87791646e+00 -8.91323447e-01 -5.61081886e-01 2.42990637e+00
7.18840420e-01 -2.06289306e-01 6.54085815e-01 -5.66830598e-02
1.06031585e+00 -1.93464737e-02 -4.69278455e-01 -2.89138079e-01
2.16070563e-01 -1.43027008e-01 4.25979674e-01 6.11735344e-01
-9.01799560e-01 6.19641185e-01 7.44307899e+00 8.19040656e-01
-6.30683839e-01 -6.48666844e-02 7.19559431e-01 -1.23257510e-01
-1.98126912e-01 2.07180515e-01 -8.13137531e-01 7.27579892e-01
9.59171593e-01 -2.70134032e-01 2.67451167e-01 6.08358324e-01
5.74921668e-01 -7.27075160e-01 -8.72671247e-01 5.02501070e-01
-1.04695037e-02 -7.00585961e-01 -1.81357950e-01 5.70322275e-01
6.08252048e-01 5.35924509e-02 -1.25976503e-01 -9.33358446e-03
7.34860301e-01 -6.80926383e-01 3.34853649e-01 2.88870782e-01
5.67443907e-01 -7.64709830e-01 4.86570001e-01 5.57631969e-01
-7.46859610e-01 -1.17862009e-01 -5.04321694e-01 -3.03188920e-01
2.15095490e-01 1.02989507e+00 -1.35013092e+00 -2.12047338e-01
6.92027092e-01 5.66707850e-01 -3.33579779e-01 1.14102209e+00
2.07238607e-02 9.71554935e-01 -9.43414688e-01 -1.89441983e-02
8.82711485e-02 -6.53334618e-01 5.99772334e-01 1.03312814e+00
3.95160586e-01 -1.43191695e-01 1.79928422e-01 8.80893886e-01
3.41549903e-01 3.52801681e-01 -6.05822027e-01 -4.41290736e-02
6.58179402e-01 1.13354266e+00 -1.20423269e+00 -2.68609017e-01
-3.26491535e-01 2.82513082e-01 1.57691985e-01 2.64868557e-01
-4.69517410e-01 4.09729518e-02 5.97001195e-01 6.16301060e-01
6.91694021e-01 -3.96627963e-01 -3.51678312e-01 -8.47649038e-01
-4.92869765e-01 -5.11210978e-01 4.01044577e-01 -4.98284757e-01
-1.55995071e+00 -9.62092727e-02 4.93401647e-01 -9.54367101e-01
-4.65240300e-01 -1.73813835e-01 -6.41856670e-01 8.64662826e-01
-7.50290215e-01 -8.20979059e-01 -1.09971194e-02 2.15211660e-01
3.25659335e-01 8.12862888e-02 6.85979247e-01 4.22282424e-03
-5.73699057e-01 -1.76031187e-01 8.94673884e-01 -2.38516152e-01
6.21025503e-01 -1.06661212e+00 3.24852139e-01 4.30426627e-01
4.78268079e-02 7.19996870e-01 1.01345229e+00 -1.06961894e+00
-6.94486320e-01 -6.95970595e-01 7.76868880e-01 -2.03173667e-01
1.02036226e+00 -6.35553300e-01 -7.85131514e-01 7.12919474e-01
-9.34072658e-02 -6.48264229e-01 1.17357457e+00 3.97332430e-01
1.65967783e-03 1.74264640e-01 -1.09329987e+00 4.28807884e-01
2.79456854e-01 -7.20136762e-02 -1.80262744e-01 4.24243212e-01
2.29445174e-01 4.13410693e-01 -8.02892148e-01 -7.71270618e-02
5.77081680e-01 -7.96872437e-01 6.85888529e-01 -4.01077002e-01
2.86067843e-01 -2.36740336e-01 -2.03116536e-01 -1.20071113e+00
-5.04168987e-01 -2.99394391e-02 1.43504113e-01 1.25102460e+00
1.46348372e-01 -2.99879849e-01 7.36022174e-01 5.62365770e-01
5.45796812e-01 -4.67868090e-01 -9.68946218e-01 -4.02933627e-01
1.04836941e-01 -2.30999991e-01 4.49418247e-01 9.93552089e-01
1.42684191e-01 4.28538501e-01 -2.87050992e-01 1.90288261e-01
9.13681328e-01 9.46541503e-02 8.75831723e-01 -1.68333340e+00
-1.82652399e-01 -2.12143779e-01 -2.81577945e-01 -7.15516031e-01
4.42052931e-02 -5.53051889e-01 1.31429002e-01 -1.39232445e+00
6.41955853e-01 -7.58649230e-01 4.94816452e-01 2.64852166e-01
-1.93392292e-01 1.47749990e-01 -5.43636866e-02 6.75736368e-01
-2.51111656e-01 3.45531285e-01 5.08279443e-01 1.02063075e-01
-4.35323894e-01 1.56659335e-01 -3.81715089e-01 6.92668021e-01
6.38054073e-01 -8.17249775e-01 -2.62871921e-01 1.14197917e-01
4.19858009e-01 -1.08569175e-01 2.28677258e-01 -5.40813029e-01
1.12340361e-01 -2.12769836e-01 1.24246381e-01 -7.99346983e-01
1.55168921e-01 -8.82293582e-01 6.41001940e-01 4.15085942e-01
-2.61347443e-01 -2.42791578e-01 6.78126812e-02 8.72530699e-01
7.07521476e-03 -4.59816396e-01 5.72324276e-01 -1.13661904e-02
-5.36893494e-02 3.40262800e-02 -1.03071952e+00 -3.04739028e-01
1.11046493e+00 -5.43031991e-02 6.91635534e-02 -6.37823582e-01
-8.99589717e-01 2.61549443e-01 6.41939461e-01 1.30525121e-04
1.11856330e-02 -1.04928863e+00 -8.59441876e-01 1.40527591e-01
-1.06143327e-02 -2.48207841e-02 1.17618978e-01 7.98052609e-01
-4.83414203e-01 2.37665012e-01 -2.70713288e-02 -9.12335932e-01
-1.26203644e+00 7.27633178e-01 -1.78120658e-01 -3.83377582e-01
-3.69358897e-01 1.70791343e-01 5.03306448e-01 -3.78402442e-01
2.06993856e-02 6.48158267e-02 -3.22670698e-01 6.71605349e-01
3.01147878e-01 5.44159830e-01 -3.26044410e-01 -5.23100913e-01
-2.15330422e-01 2.05220163e-01 4.74251285e-02 -2.43705660e-01
1.68177176e+00 -4.63856637e-01 -5.02781808e-01 1.09494126e+00
9.76701319e-01 -6.61753342e-02 -1.19025135e+00 -1.54115990e-01
4.35463339e-02 -2.76860267e-01 2.13599298e-03 -3.81689906e-01
-5.44536591e-01 8.58514607e-01 4.17696834e-01 7.40842402e-01
6.39442801e-01 2.10357860e-01 1.94074646e-01 -1.33821696e-01
2.44569138e-01 -7.16097414e-01 -3.00371766e-01 6.92162961e-02
4.55415010e-01 -9.26117063e-01 4.71013576e-01 -3.61346573e-01
-3.47792685e-01 8.01834047e-01 -3.64956595e-02 -8.81970748e-02
9.76131260e-01 1.67787582e-01 -3.05549175e-01 -4.18728232e-01
-6.60954714e-01 4.12544385e-02 -1.75727233e-01 4.40610707e-01
4.36172307e-01 3.58527511e-01 -4.94891644e-01 2.64561623e-01
-1.43464684e-01 -2.76854634e-01 6.78997636e-01 6.81528211e-01
-6.87019646e-01 -1.07536972e+00 -6.62842333e-01 7.91808546e-01
-4.26485479e-01 -8.58940110e-02 -3.71257931e-01 7.44729042e-01
-2.15875134e-01 7.59913981e-01 6.33099854e-01 1.23740323e-01
-2.52218634e-01 3.21549445e-01 8.64920691e-02 -7.57323205e-01
-3.24891299e-01 5.34725249e-01 -1.28545463e-01 1.18532993e-01
-5.64626813e-01 -1.28084898e+00 -7.32811153e-01 -5.96722066e-01
-5.05707443e-01 4.90010649e-01 1.03445494e+00 7.08078265e-01
2.05074131e-01 6.72837570e-02 7.79425502e-01 -8.52868557e-01
-1.45932570e-01 -1.29134643e+00 -1.16856420e+00 2.99925864e-01
-2.87441853e-02 -6.95884764e-01 -7.28272498e-01 2.48340532e-01] | [10.311253547668457, 6.890551567077637] |
acf6e12d-1728-41bc-b2e3-ff5e2a5f1e35 | multi-script-handwritten-digit-recognition | 2106.08267 | null | https://arxiv.org/abs/2106.08267v1 | https://arxiv.org/pdf/2106.08267v1.pdf | Multi-script Handwritten Digit Recognition Using Multi-task Learning | Handwritten digit recognition is one of the extensively studied area in machine learning. Apart from the wider research on handwritten digit recognition on MNIST dataset, there are many other research works on various script recognition. However, it is not very common for multi-script digit recognition which encourage the development of robust and multipurpose systems. Additionally working on multi-script digit recognition enables multi-task learning, considering the script classification as a related task for instance. It is evident that multi-task learning improves model performance through inductive transfer using the information contained in related tasks. Therefore, in this study multi-script handwritten digit recognition using multi-task learning will be investigated. As a specific case of demonstrating the solution to the problem, Amharic handwritten character recognition will also be experimented. The handwritten digits of three scripts including Latin, Arabic and Kannada are studied to show that multi-task models with reformulation of the individual tasks have shown promising results. In this study a novel way of using the individual tasks predictions was proposed to help classification performance and regularize the different loss for the purpose of the main task. This finding has outperformed the baseline and the conventional multi-task learning models. More importantly, it avoided the need for weighting the different losses of the tasks, which is one of the challenges in multi-task learning. | ['Randolf Scholz', 'Durga Prasad Sharma', 'Lars Schmidt-Thieme', 'Mesay Samuel Gondere'] | 2021-06-15 | null | null | null | null | ['handwritten-digit-recognition'] | ['computer-vision'] | [ 1.59386545e-01 -4.90253747e-01 -3.63258086e-02 -6.70056045e-01
-6.76334321e-01 -5.64586759e-01 7.54387915e-01 -2.89771497e-01
-6.05901599e-01 7.81825542e-01 -1.42494649e-01 -1.91758603e-01
-3.68176728e-01 -4.51060712e-01 -5.39154291e-01 -8.91931057e-01
3.68868679e-01 5.37756324e-01 1.94842547e-01 -3.80291305e-02
4.66626227e-01 7.77516067e-01 -1.29983711e+00 8.06995094e-01
6.07251167e-01 7.63032377e-01 3.66456568e-01 6.68819487e-01
2.60391161e-02 1.14879262e+00 -7.70929277e-01 -7.52996802e-01
2.54011184e-01 -3.18767041e-01 -7.91908979e-01 2.69399256e-01
5.87951303e-01 -3.60844582e-01 -4.69344147e-02 6.56913638e-01
6.51729763e-01 7.24678487e-02 8.52075696e-01 -9.78569210e-01
-7.01784849e-01 6.11324728e-01 -6.65142119e-01 1.32082015e-01
-1.64490510e-02 -1.81142420e-01 8.09691906e-01 -8.81963968e-01
2.70658612e-01 1.33812213e+00 7.13534713e-01 2.42458493e-01
-1.04609764e+00 -4.78816748e-01 4.86923680e-02 5.74524999e-01
-1.13144028e+00 -2.18080580e-02 6.90814376e-01 -4.94024694e-01
9.57683921e-01 1.82468936e-01 -6.63408116e-02 1.29234159e+00
3.27393353e-01 9.65486526e-01 1.47812331e+00 -7.88262606e-01
-2.72968292e-01 5.60181022e-01 4.36602294e-01 4.46522623e-01
1.23774391e-02 -1.70834363e-01 -2.15733826e-01 1.69733956e-01
6.50525153e-01 -4.19547111e-02 1.90254748e-01 -1.50616765e-01
-1.15648079e+00 8.76771688e-01 -2.67186575e-02 9.35884714e-01
-3.48352909e-01 -2.97077578e-02 7.63856530e-01 6.03216946e-01
2.57025957e-01 1.72246814e-01 -5.86246192e-01 -1.26428768e-01
-1.00943255e+00 2.06176750e-02 8.94186020e-01 4.93430495e-01
5.95989466e-01 3.84743840e-01 -3.42778832e-01 1.23892796e+00
3.29970926e-01 7.79831931e-02 6.90442145e-01 -2.97178656e-01
6.98325276e-01 6.41762078e-01 -3.21874887e-01 -7.49264300e-01
-2.03243896e-01 -3.17901999e-01 -1.02625215e+00 5.87773144e-01
8.69944453e-01 -2.46141672e-01 -9.63101685e-01 1.10202014e+00
-2.33672962e-01 -2.34945029e-01 8.59754533e-02 7.57048190e-01
2.32017010e-01 5.56290507e-01 1.21961318e-01 1.99709058e-01
1.36220193e+00 -1.09719253e+00 -7.05773473e-01 -2.62963325e-01
3.92141372e-01 -1.16407907e+00 7.48115897e-01 8.74143422e-01
-5.11343896e-01 -9.72210288e-01 -9.88774240e-01 1.79541156e-01
-6.28559589e-01 8.90623689e-01 6.88464165e-01 1.05248106e+00
-5.74248850e-01 5.73092997e-01 -5.75481176e-01 -4.22568113e-01
4.32754695e-01 3.09624583e-01 -3.65218461e-01 -3.64351690e-01
-8.79486740e-01 1.57108343e+00 6.89990342e-01 2.64612921e-02
-8.27440917e-01 -2.99647570e-01 -3.89146149e-01 -2.27228671e-01
-3.51607753e-03 2.69729421e-02 9.45344031e-01 -1.44309878e+00
-1.45005643e+00 9.83354747e-01 1.83216736e-01 -3.93091023e-01
9.47065175e-01 -1.99788660e-01 -4.77137268e-01 -1.55964583e-01
-8.76435488e-02 2.35884532e-01 1.08359516e+00 -9.40075219e-01
-4.89652276e-01 -4.06388611e-01 -3.50615442e-01 -3.62605490e-02
-5.06032765e-01 4.51728404e-01 -1.65577605e-01 -1.03615439e+00
-2.57352114e-01 -7.26411104e-01 7.12247044e-02 -2.05528527e-01
-1.83157131e-01 -3.49086493e-01 1.19285452e+00 -9.43907380e-01
8.93308997e-01 -2.24549413e+00 2.48267934e-01 1.86088637e-01
-4.23661321e-01 4.92790878e-01 -2.98104614e-01 4.33795840e-01
-1.62555546e-01 -1.75012335e-01 -1.90576360e-01 -3.00205082e-01
-9.08517987e-02 4.03608382e-01 -2.37986267e-01 2.26362318e-01
3.83785248e-01 5.14847338e-01 -2.42792219e-01 -2.91728228e-01
3.77567530e-01 4.69720066e-01 1.32757962e-01 -1.11155279e-01
2.41579972e-02 3.13252121e-01 -2.93542475e-01 8.14950645e-01
6.55376315e-01 -1.36134145e-03 8.73567611e-02 -4.13888991e-01
-1.61442891e-01 -2.80182570e-01 -1.39058185e+00 1.33671987e+00
-5.42721450e-01 7.14641035e-01 -3.28437805e-01 -1.49832952e+00
1.37361693e+00 5.34632385e-01 3.00703049e-01 -5.99161923e-01
-1.66583419e-01 5.64302444e-01 4.65661645e-01 -5.14469504e-01
3.21669281e-01 -2.49770671e-01 1.03344388e-01 5.43419540e-01
3.15673709e-01 -5.88671677e-02 2.44832247e-01 -4.25243974e-01
5.81315756e-01 3.29912573e-01 2.46070161e-01 -6.05955012e-02
8.36015761e-01 -1.37244970e-01 2.99157739e-01 6.55710459e-01
-1.54040903e-01 4.39953268e-01 2.68847466e-01 -5.74950933e-01
-1.20912135e+00 -4.77325618e-01 -3.14000309e-01 1.30172873e+00
-7.28917897e-01 3.06634575e-01 -1.90446660e-01 -6.67357206e-01
2.58365333e-01 6.59953058e-01 -5.57040632e-01 3.05187047e-01
-7.29630411e-01 -1.27591181e+00 9.86992538e-01 5.23402214e-01
9.31227446e-01 -1.07426941e+00 -5.43074369e-01 3.39007080e-01
2.16870725e-01 -1.03964269e+00 -4.57446799e-02 6.86901987e-01
-9.52856541e-01 -1.03528261e+00 -1.31593227e+00 -9.61069167e-01
1.88348889e-01 -1.81288570e-01 5.45519948e-01 -4.42627877e-01
-5.41845441e-01 3.00936103e-01 -4.12650138e-01 -4.14285392e-01
-6.21479988e-01 2.07047522e-01 -2.08034351e-01 4.18997169e-01
4.77385908e-01 -3.24004173e-01 2.56363928e-01 4.86018151e-01
-7.40541220e-01 -1.69994816e-01 1.09610903e+00 8.99829984e-01
1.75501525e-01 -7.90165514e-02 9.74082708e-01 -6.12049758e-01
5.20204842e-01 -2.24986553e-01 -4.46665466e-01 7.62751162e-01
-4.45662916e-01 1.11746952e-01 6.93679631e-01 -6.50552452e-01
-1.32646310e+00 2.10112885e-01 -4.67615686e-02 -1.92034453e-01
-3.92838657e-01 4.82576728e-01 -1.58215351e-02 -3.41600090e-01
6.46226645e-01 5.81921101e-01 1.45430788e-02 -5.49221456e-01
-9.88752022e-02 9.25728858e-01 -9.05124936e-03 -4.97868389e-01
4.21521455e-01 -6.50408268e-02 6.83462396e-02 -9.85034466e-01
-6.36913657e-01 -2.22493246e-01 -9.90061939e-01 -3.68193418e-01
8.76964211e-01 -7.26514161e-01 -6.20116770e-01 1.20892739e+00
-1.30637538e+00 -3.69887173e-01 1.58469945e-01 6.62235022e-01
-4.34445649e-01 5.99746823e-01 -5.98102272e-01 -9.49293494e-01
1.34869620e-01 -1.32667863e+00 6.01417184e-01 -1.66339986e-02
5.12485839e-02 -1.27169192e+00 -1.09525092e-01 5.24922192e-01
5.39241135e-01 1.48745522e-01 1.08639133e+00 -1.07664704e+00
-5.17002583e-01 -2.74849862e-01 -3.76566410e-01 9.31631923e-01
5.05487204e-01 6.76508546e-02 -1.19530535e+00 -1.83675408e-01
8.32929015e-02 -8.32694530e-01 1.05763066e+00 2.07173392e-01
9.46494699e-01 -2.38087010e-02 7.63942078e-02 1.56296134e-01
1.60200560e+00 4.93260056e-01 6.94685459e-01 7.38861740e-01
6.46863282e-01 6.34186268e-01 4.86761838e-01 4.91497487e-01
6.76610172e-02 8.03843141e-01 2.37017795e-02 1.79245800e-01
-3.20100933e-01 4.71851915e-01 5.94581425e-01 5.86579502e-01
-3.20553988e-01 8.67263451e-02 -1.21477461e+00 2.63642102e-01
-1.71331263e+00 -9.95029449e-01 -3.31829399e-01 1.89399838e+00
9.03699219e-01 7.46658351e-03 1.86663359e-01 4.98927265e-01
7.00334847e-01 1.31321713e-01 -4.53316659e-01 -7.23291814e-01
-5.46286941e-01 1.87080637e-01 4.79477316e-01 2.25454390e-01
-1.64179909e+00 7.34628499e-01 5.74104691e+00 8.19788754e-01
-1.29041743e+00 5.74742220e-02 6.50744319e-01 4.41406429e-01
4.47501153e-01 -4.19973910e-01 -1.14187741e+00 3.76588404e-01
7.75421500e-01 4.57166545e-02 2.92861998e-01 8.53634715e-01
4.75601889e-02 -2.76716739e-01 -1.14623821e+00 8.73821735e-01
4.83919501e-01 -8.14733267e-01 3.33786130e-01 -1.65287793e-01
6.26519263e-01 -1.10739373e-01 1.13423027e-01 3.67129534e-01
1.18967600e-01 -1.03017259e+00 6.96327925e-01 6.00118637e-01
4.32601959e-01 -5.26242971e-01 9.72528994e-01 4.03771192e-01
-7.47980416e-01 -4.60096836e-01 -2.85168201e-01 -1.12933218e-02
-1.04124457e-01 4.55857664e-01 -6.08470738e-01 8.12987983e-01
4.38435733e-01 9.19456005e-01 -9.10026193e-01 1.15166807e+00
4.95276861e-02 5.71999073e-01 -7.39467284e-03 -1.59184888e-01
5.08674145e-01 -2.16784224e-01 1.49833322e-01 1.60659564e+00
4.05587018e-01 -5.06261110e-01 1.21858746e-01 5.45770049e-01
2.20603377e-01 1.98501840e-01 -6.94066048e-01 -1.41307041e-01
-1.17899515e-01 1.09588921e+00 -6.80619240e-01 -3.05474073e-01
-7.50977993e-01 1.34389532e+00 5.95408585e-03 1.79293066e-01
-6.86653078e-01 -5.12578011e-01 1.42179564e-01 -4.87515271e-01
6.05224967e-01 -3.73191059e-01 -6.99978471e-01 -9.67148900e-01
1.85082704e-01 -1.02922463e+00 3.56962025e-01 -3.06054473e-01
-1.56608915e+00 3.48702163e-01 -2.10661411e-01 -1.11995459e+00
-2.27742895e-01 -1.11870372e+00 -6.02749765e-01 1.17767966e+00
-1.55984104e+00 -1.60929036e+00 -1.35609135e-01 6.68674529e-01
8.84591758e-01 -8.60055387e-01 1.01700830e+00 7.32637346e-01
-6.96127951e-01 7.83401728e-01 4.15488154e-01 3.94236326e-01
1.07894754e+00 -1.17192626e+00 -1.18396208e-01 7.20800340e-01
3.84722501e-02 1.37483180e-02 2.58908093e-01 -4.93815124e-01
-1.06739318e+00 -1.02355564e+00 7.95504689e-01 -1.71706542e-01
7.11690307e-01 -1.74852703e-02 -1.01485515e+00 8.15405965e-01
2.31228814e-01 -4.25255179e-01 8.24178100e-01 -1.81037739e-01
-4.66101736e-01 -2.76147753e-01 -1.08918512e+00 7.37671554e-02
1.70792356e-01 -4.60468471e-01 -7.46718109e-01 3.94494474e-01
-2.50844806e-01 1.24684744e-01 -9.48939562e-01 8.10579136e-02
5.73001266e-01 -8.56523931e-01 9.91280973e-01 -5.06937146e-01
6.31901681e-01 -1.64345391e-02 -4.36120808e-01 -1.24619484e+00
-2.00449109e-01 4.69846316e-02 1.31164581e-01 1.56088948e+00
4.37030107e-01 -5.92916608e-01 4.89026129e-01 3.84090632e-01
-5.98278046e-02 -2.71320879e-01 -9.87104237e-01 -1.21026886e+00
4.40428406e-01 -2.53436446e-01 -5.61252609e-02 1.14987063e+00
-5.41697443e-01 1.50511742e-01 -6.59439147e-01 -1.76979482e-01
6.88897431e-01 1.76891387e-02 3.97096574e-01 -1.27711058e+00
-4.90795791e-01 -5.18885553e-01 -4.44174916e-01 -4.81461257e-01
2.09117115e-01 -1.02415550e+00 -3.13656271e-01 -1.41409767e+00
8.72185901e-02 -4.90264326e-01 -1.53310016e-01 8.03783000e-01
3.88052850e-03 2.01541752e-01 5.34238994e-01 2.26803318e-01
-1.96241483e-01 1.58123121e-01 1.17829049e+00 -2.81590760e-01
7.91986380e-03 4.44094211e-01 -2.48113573e-01 5.26869893e-01
1.00346792e+00 -2.67505497e-01 4.15360443e-02 -5.68381488e-01
-2.44837016e-01 -1.13497749e-01 4.57678288e-01 -1.10836244e+00
2.91812450e-01 2.91413777e-02 6.74352646e-01 -5.88100672e-01
4.81184900e-01 -8.84951234e-01 -7.43492544e-02 6.59790576e-01
-2.37806037e-01 -7.92439505e-02 4.76436049e-01 1.76124334e-01
-5.57184994e-01 -6.11329675e-01 1.01451552e+00 -2.97724307e-01
-1.11520803e+00 -2.47240067e-01 -7.86251307e-01 -4.63758111e-01
1.34195173e+00 -5.14679432e-01 -3.70067582e-02 1.94601819e-01
-8.59640419e-01 -1.60481319e-01 -2.58444101e-01 5.30092597e-01
5.15080512e-01 -1.20095253e+00 -1.09179270e+00 2.99600977e-02
-5.56377992e-02 -6.30917311e-01 2.75983781e-01 7.71578908e-01
-3.33110482e-01 7.06184030e-01 -9.48973358e-01 -3.76959473e-01
-1.61669719e+00 1.52817726e-01 4.57472175e-01 -5.01841843e-01
-1.31938547e-01 5.70576489e-01 -5.10459661e-01 -5.35022497e-01
5.75933576e-01 -1.66613370e-01 -3.84781122e-01 6.94168985e-01
2.72919834e-01 6.37344956e-01 3.10407519e-01 -3.37414205e-01
-2.75194615e-01 6.98845088e-01 -5.24254799e-01 4.80323285e-03
1.58317852e+00 3.35294366e-01 -2.47984961e-01 8.14569890e-01
9.99443829e-01 -5.27617395e-01 -1.15293586e+00 -1.26153558e-01
7.18303323e-01 -3.84996980e-01 -1.83957651e-01 -1.36994433e+00
-8.01580012e-01 1.07331634e+00 7.53799856e-01 -7.95277208e-02
1.13771820e+00 -4.47432965e-01 1.50081083e-01 8.68856966e-01
4.35223311e-01 -1.25401974e+00 4.33580011e-01 7.98188627e-01
1.09539652e+00 -1.48326313e+00 -8.39226916e-02 5.31950817e-02
-9.51321781e-01 1.86416113e+00 5.92880368e-01 -1.01134427e-01
4.64792460e-01 2.89243072e-01 4.37812880e-02 2.42195219e-01
-2.94307023e-01 7.20321313e-02 2.19403759e-01 6.17599308e-01
7.96035171e-01 -8.39973688e-02 -2.40996018e-01 5.74389994e-01
4.71775770e-01 2.64356852e-01 6.13527179e-01 8.95262659e-01
-2.40863815e-01 -1.68195236e+00 -6.13618493e-01 6.77625299e-01
-7.38726616e-01 1.71465613e-02 -4.24587786e-01 8.17806780e-01
1.15493841e-01 5.88916063e-01 -2.57250786e-01 -2.14249358e-01
3.50466222e-01 5.48121154e-01 6.90600276e-01 -5.34475386e-01
-1.18014967e+00 -4.03961122e-01 5.76557443e-02 5.09438403e-02
-6.10003471e-01 -8.77030492e-01 -5.44765592e-01 -1.09876975e-01
-1.82615384e-01 -2.65597761e-01 9.74757671e-01 9.53327715e-01
-1.98334396e-01 5.80790043e-01 5.57936013e-01 -7.80729413e-01
-1.09428918e+00 -1.21562231e+00 -6.81751788e-01 1.64281428e-01
1.49184674e-01 -4.37642217e-01 -4.14844230e-02 3.02021265e-01] | [11.822907447814941, 2.606790781021118] |
4f3f4435-c15a-4c15-98d3-e4a438b42036 | a-deep-cascade-model-for-multi-document | 1811.11374 | null | http://arxiv.org/abs/1811.11374v1 | http://arxiv.org/pdf/1811.11374v1.pdf | A Deep Cascade Model for Multi-Document Reading Comprehension | A fundamental trade-off between effectiveness and efficiency needs to be
balanced when designing an online question answering system. Effectiveness
comes from sophisticated functions such as extractive machine reading
comprehension (MRC), while efficiency is obtained from improvements in
preliminary retrieval components such as candidate document selection and
paragraph ranking. Given the complexity of the real-world multi-document MRC
scenario, it is difficult to jointly optimize both in an end-to-end system. To
address this problem, we develop a novel deep cascade learning model, which
progressively evolves from the document-level and paragraph-level ranking of
candidate texts to more precise answer extraction with machine reading
comprehension. Specifically, irrelevant documents and paragraphs are first
filtered out with simple functions for efficiency consideration. Then we
jointly train three modules on the remaining texts for better tracking the
answer: the document extraction, the paragraph extraction and the answer
extraction. Experiment results show that the proposed method outperforms the
previous state-of-the-art methods on two large-scale multi-document benchmark
datasets, i.e., TriviaQA and DuReader. In addition, our online system can
stably serve typical scenarios with millions of daily requests in less than
50ms. | ['Rui Wang', 'Haiqing Chen', 'Ming Yan', 'Jiangnan Xia', 'Bin Bi', 'Zhongzhou Zhao', 'Luo Si', 'Chen Wu', 'Wei Wang', 'Ji Zhang'] | 2018-11-28 | null | null | null | null | ['triviaqa'] | ['miscellaneous'] | [-1.80808250e-02 -2.20878869e-01 8.22809618e-03 -3.67024362e-01
-1.25803339e+00 -6.43534005e-01 5.41606665e-01 5.65618932e-01
-7.79958963e-01 5.36783159e-01 3.00367445e-01 -5.63681245e-01
-3.65389556e-01 -8.09407830e-01 -5.78152716e-01 -2.87222207e-01
4.37815189e-01 7.02253461e-01 5.27729750e-01 -4.05290693e-01
6.40282452e-01 2.43325029e-02 -1.47635376e+00 5.21795571e-01
1.38646245e+00 1.00264657e+00 3.39552432e-01 1.06929922e+00
-4.49639022e-01 9.69237387e-01 -7.47064888e-01 -6.03885651e-01
-1.05369754e-01 -2.93981701e-01 -1.03050411e+00 -1.67238757e-01
4.80700910e-01 -9.48028028e-01 -2.40047753e-01 7.42552698e-01
8.35150659e-01 9.31068361e-02 5.73268712e-01 -6.12388015e-01
-7.41496444e-01 7.68795609e-01 -5.66986859e-01 4.57254171e-01
4.65065628e-01 1.87053427e-01 1.36365080e+00 -9.96884346e-01
1.49595648e-01 1.35455120e+00 -8.78848508e-02 2.06606790e-01
-4.87333149e-01 -3.03568959e-01 3.42063516e-01 4.43002403e-01
-9.22913790e-01 -5.51092863e-01 4.71967936e-01 1.83085147e-02
9.63799179e-01 4.58856225e-01 8.22585672e-02 7.03878701e-01
1.53355330e-01 1.38903296e+00 6.00299180e-01 -4.18915510e-01
5.81962802e-02 5.41296154e-02 5.38638115e-01 5.97229540e-01
1.54122204e-01 -5.32522380e-01 -3.94196332e-01 8.43719244e-02
2.14349747e-01 -6.29202053e-02 -3.55753750e-01 8.46221074e-02
-1.07470453e+00 8.23801160e-01 2.46325254e-01 1.76415622e-01
-3.41628373e-01 -1.08243637e-01 5.42445481e-01 5.99225283e-01
4.03451204e-01 4.44826782e-01 -6.83616519e-01 -3.70550565e-02
-9.83015716e-01 3.98349851e-01 9.84722972e-01 1.00879991e+00
4.36479688e-01 -4.93842930e-01 -8.42651367e-01 9.82806563e-01
3.81377131e-01 6.06031358e-01 5.04415274e-01 -6.40390813e-01
1.08383358e+00 7.00169444e-01 2.99746633e-01 -9.44095671e-01
-3.95569682e-01 -7.54824221e-01 -7.44327486e-01 -6.38325512e-01
4.91629660e-01 -2.67181158e-01 -7.04750717e-01 1.29708624e+00
3.85417432e-01 -4.78809416e-01 -6.86682016e-02 1.00675452e+00
1.06432354e+00 8.81543815e-01 1.87686067e-02 -2.35556588e-01
1.63990080e+00 -1.51872253e+00 -7.64015973e-01 -2.75835276e-01
5.64243853e-01 -9.47714984e-01 1.32755959e+00 1.89730182e-01
-1.43307161e+00 -5.65782070e-01 -9.23894584e-01 -6.89740658e-01
-2.76916713e-01 4.76685613e-01 3.37756127e-01 1.65138006e-01
-7.53557086e-01 1.93608776e-01 -5.40048361e-01 -3.30385081e-02
2.03358442e-01 2.02480823e-01 2.85561383e-01 -1.53085738e-01
-1.37657166e+00 8.20453703e-01 1.34271562e-01 2.87101269e-01
-6.73340142e-01 -4.89517570e-01 -2.87124306e-01 6.67338014e-01
5.93702972e-01 -1.12847793e+00 1.78541148e+00 -5.29192507e-01
-1.64270663e+00 5.38477421e-01 -2.18898907e-01 -1.33427650e-01
5.63282609e-01 -7.77841985e-01 -1.85753748e-01 4.02079254e-01
6.50702342e-02 2.64572293e-01 9.19058442e-01 -8.14371943e-01
-1.04665470e+00 -3.92060697e-01 5.06592393e-01 5.64098239e-01
-5.21621346e-01 7.02482462e-02 -9.94900227e-01 -2.62617469e-01
-1.71291098e-01 -3.09268951e-01 1.40556559e-01 -3.56839269e-01
-3.27325821e-01 -7.51920164e-01 4.39519107e-01 -9.70959187e-01
1.61553049e+00 -1.70493114e+00 2.72151321e-01 -1.24223024e-01
3.40606034e-01 3.43374163e-01 -4.81579870e-01 5.77665269e-01
4.71128494e-01 -1.06408435e-04 1.58737674e-01 -2.97554284e-01
1.41212985e-01 -4.00962323e-01 -4.45666581e-01 7.65055493e-02
2.39133656e-01 1.10723984e+00 -1.01653767e+00 -7.02529013e-01
-2.20519036e-01 -2.24687420e-02 -4.72245038e-01 5.26244879e-01
-5.32166541e-01 2.44477596e-02 -9.10383582e-01 7.19138563e-01
4.77743447e-01 -6.32513106e-01 -2.50429899e-01 3.62647213e-02
9.57003310e-02 6.76621020e-01 -8.25617790e-01 1.59185970e+00
-6.17720366e-01 3.31153989e-01 1.43860936e-01 -6.99653089e-01
5.04898727e-01 1.24626510e-01 6.34918064e-02 -1.28320503e+00
1.90188870e-01 4.24241960e-01 1.67787932e-02 -7.79629648e-01
9.03609991e-01 5.99589586e-01 3.47439498e-02 7.40493119e-01
-7.15210363e-02 -6.84930682e-02 5.63719809e-01 6.19210482e-01
1.08412910e+00 -2.16066286e-01 -1.24263884e-02 -2.14329988e-01
9.03885663e-01 -7.77792856e-02 -1.06040016e-01 9.19370532e-01
5.39926514e-02 4.04345244e-01 4.73845452e-01 -9.66650918e-02
-7.57268727e-01 -8.73297095e-01 1.49471611e-01 1.64827716e+00
3.11179757e-01 -2.71746814e-01 -9.46380496e-01 -8.91708612e-01
-1.17084488e-01 5.29168427e-01 -1.98748395e-01 -1.83901474e-01
-9.09423053e-01 -5.63745677e-01 1.60483524e-01 3.71138752e-01
7.09968507e-01 -1.10743666e+00 -6.57396853e-01 4.12727624e-01
-5.20914614e-01 -9.05156910e-01 -7.77668953e-01 -1.50423110e-01
-7.56494403e-01 -1.06476390e+00 -9.33146179e-01 -8.26278210e-01
4.71179903e-01 5.95750928e-01 1.55081367e+00 6.26082957e-01
7.74444044e-02 2.95949101e-01 -6.82558477e-01 -5.27603805e-01
-4.76474501e-02 8.62202287e-01 -4.46432739e-01 -1.61504015e-01
3.54746610e-01 -7.63315037e-02 -1.14112842e+00 2.92210251e-01
-1.03722382e+00 -8.87851864e-02 9.78087366e-01 9.10359204e-01
3.64929318e-01 -8.52497891e-02 9.66219246e-01 -7.84609020e-01
1.43650317e+00 -5.16690016e-01 -5.39663136e-01 9.21905041e-01
-4.49175298e-01 1.56016931e-01 8.40543866e-01 -3.86614710e-01
-1.11157525e+00 -4.39346433e-01 -4.31398660e-01 1.54099479e-01
2.25478902e-01 7.43117452e-01 1.24735022e-02 5.40345728e-01
4.45730299e-01 3.68947715e-01 -3.93875420e-01 -4.85400677e-01
4.99712467e-01 1.20850527e+00 4.07243490e-01 -5.10586083e-01
6.26043975e-01 -1.06735779e-02 -6.10327005e-01 -5.15732467e-01
-1.27918518e+00 -7.31357396e-01 -4.62739855e-01 -8.25546831e-02
5.25508046e-01 -8.48055959e-01 -8.73206973e-01 4.90310580e-01
-1.41891515e+00 3.94849665e-03 7.90729672e-02 1.28607884e-01
-2.26480320e-01 3.47456634e-01 -8.03510249e-01 -6.89947844e-01
-1.17228961e+00 -9.96361017e-01 1.43407297e+00 5.52322865e-01
-2.60466263e-02 -6.63365781e-01 -9.97614637e-02 8.28342915e-01
6.35441780e-01 -6.46890283e-01 1.20661378e+00 -7.07018375e-01
-8.74797225e-01 -3.67114991e-01 -4.77132261e-01 1.83184177e-01
-1.11931242e-01 -8.01719725e-02 -8.38721812e-01 -4.73087817e-01
2.00150728e-01 -5.99789619e-01 1.10606360e+00 5.56274578e-02
1.37054873e+00 -4.34683204e-01 4.78744833e-03 8.31077024e-02
1.03015816e+00 2.90733706e-02 4.51128244e-01 3.14927161e-01
4.74809617e-01 7.60397196e-01 8.62268567e-01 3.08877945e-01
8.17058086e-01 3.22572291e-01 3.45524698e-01 4.98958789e-02
-3.70571874e-02 -2.40256131e-01 2.39735171e-01 1.49470258e+00
4.32468057e-01 -7.86016047e-01 -8.42192948e-01 5.55874348e-01
-1.94571543e+00 -6.60559297e-01 -5.62271103e-02 1.95593655e+00
9.87148702e-01 3.28934163e-01 -7.71272928e-02 4.18628044e-02
3.68329674e-01 2.07733765e-01 -8.10133457e-01 -2.30536029e-01
7.15525001e-02 1.58230513e-01 7.56541416e-02 4.61449265e-01
-1.02505386e+00 9.32919025e-01 5.39902067e+00 1.03671002e+00
-8.48166108e-01 2.07670648e-02 7.72580862e-01 -2.01281816e-01
-4.09358054e-01 -2.00284988e-01 -1.03891909e+00 4.62833494e-01
9.38418567e-01 -1.54774517e-01 3.95017326e-01 5.57735384e-01
6.67359233e-02 -1.77262545e-01 -1.03840292e+00 7.42319643e-01
1.51168242e-01 -1.06878054e+00 3.12794209e-01 -4.31658328e-01
4.86403197e-01 -9.41468030e-03 3.15609910e-02 7.66374886e-01
-2.21838783e-02 -8.41700256e-01 6.99999988e-01 6.61975920e-01
3.66298884e-01 -8.22934806e-01 8.25389028e-01 8.08678448e-01
-9.96740162e-01 -2.97454327e-01 -5.28218210e-01 1.19427182e-01
1.39228120e-01 6.49602711e-01 -3.93759489e-01 5.10652959e-01
5.89131653e-01 2.08308712e-01 -7.64511645e-01 9.67000425e-01
-3.91250610e-01 5.09608805e-01 -1.94989860e-01 -6.48561656e-01
3.58753234e-01 -4.82832119e-02 1.61255181e-01 1.06859624e+00
9.97312590e-02 3.08431476e-01 -2.06462100e-01 5.87360620e-01
-5.44808686e-01 4.56880540e-01 1.36513278e-01 -7.10000023e-02
5.76069474e-01 1.50432062e+00 -3.51923466e-01 -4.41138119e-01
-5.95927715e-01 9.25947309e-01 6.52236700e-01 4.07421052e-01
-7.75647700e-01 -8.29143643e-01 5.24593517e-02 -9.73190069e-02
3.49301279e-01 -1.53027609e-01 -1.57066315e-01 -1.43291914e+00
5.01220047e-01 -1.39049590e+00 4.81274754e-01 -5.68049610e-01
-1.20306706e+00 5.47438025e-01 -3.93199712e-01 -9.23025787e-01
-2.23358408e-01 -3.68632168e-01 -6.66249692e-01 8.60574603e-01
-2.04554558e+00 -9.14872050e-01 -2.94235170e-01 4.01661754e-01
9.96905208e-01 1.00273505e-01 3.91883135e-01 6.27903700e-01
-7.67552733e-01 8.08365941e-01 2.48073682e-01 -1.25192730e-02
6.76293373e-01 -1.22646177e+00 5.02044976e-01 8.31778467e-01
1.54217146e-02 7.60118008e-01 2.81901240e-01 -3.67675751e-01
-1.80512869e+00 -8.13098073e-01 1.10246420e+00 -5.54090679e-01
5.58171928e-01 -2.26191118e-01 -1.01715147e+00 1.47772253e-01
4.66328502e-01 -5.60924709e-01 3.24799955e-01 1.09888822e-01
-4.90703583e-02 -2.49080434e-01 -6.20788813e-01 7.45538592e-01
6.89099193e-01 -6.31552398e-01 -7.16208220e-01 6.51175737e-01
9.87363994e-01 -5.94503224e-01 -6.40172303e-01 4.25728649e-01
5.62586308e-01 -6.28736734e-01 9.09758568e-01 -5.79097033e-01
7.72678435e-01 -8.07142183e-02 1.83466762e-01 -1.15053797e+00
-1.15915641e-01 -4.66444403e-01 -6.52409613e-01 1.15598881e+00
6.82203531e-01 -1.73829883e-01 4.53385144e-01 5.73559403e-01
4.37295437e-03 -1.23974502e+00 -6.08342886e-01 -2.31732696e-01
2.70255119e-01 7.59633929e-02 8.24826658e-01 3.12735856e-01
-1.46610498e-01 1.07719672e+00 -1.55993685e-01 -3.88389602e-02
2.47415349e-01 5.46262324e-01 8.22571754e-01 -9.92897570e-01
-4.18583453e-01 -6.79966927e-01 5.83198547e-01 -2.08509851e+00
-2.05055967e-01 -5.17705739e-01 7.23947734e-02 -1.81900811e+00
3.81401122e-01 1.21276192e-02 -2.64633685e-01 -6.13534488e-02
-8.33535790e-01 -6.94940925e-01 6.76134229e-02 3.32628608e-01
-1.34998286e+00 8.27763975e-01 1.64588201e+00 -3.26715738e-01
-1.10126659e-01 1.54090077e-01 -9.41475451e-01 4.81147528e-01
7.13330090e-01 -2.37402186e-01 -5.54886520e-01 -1.24676526e+00
5.59796870e-01 4.67167169e-01 -2.70285606e-01 -5.00024438e-01
6.90460742e-01 8.24305341e-02 2.19940260e-01 -1.06665921e+00
-8.65481049e-02 -4.94389862e-01 -9.76346433e-01 2.70611703e-01
-6.68453455e-01 3.23136270e-01 -7.19914287e-02 5.16810596e-01
-3.55195969e-01 -3.83936793e-01 4.63960975e-01 -5.19058891e-02
-3.31131816e-01 2.65480489e-01 -2.10099354e-01 3.11448425e-01
5.01182437e-01 3.29681993e-01 -9.12605464e-01 -6.29938364e-01
-4.20613214e-03 1.03325152e+00 -2.62444049e-01 7.22912371e-01
6.96934819e-01 -8.91327739e-01 -1.05348527e+00 -5.34945019e-02
4.62313294e-02 3.42661142e-01 3.25201273e-01 7.40934074e-01
-6.01532996e-01 8.91720891e-01 3.15498203e-01 -4.13382709e-01
-1.03327417e+00 3.62041980e-01 1.87631279e-01 -8.89463365e-01
-1.62407488e-01 8.88913274e-01 -3.39386519e-03 -4.24200892e-01
6.67134345e-01 -4.94571894e-01 -5.99239945e-01 3.17250073e-01
8.77043664e-01 4.35591936e-01 3.29448760e-01 4.19393331e-02
-3.05155553e-02 3.84211898e-01 -6.34091675e-01 1.61393031e-01
1.10037601e+00 -5.49795866e-01 -2.26100728e-01 -2.88870819e-02
1.09858465e+00 -5.16868792e-02 -8.36196125e-01 -3.89252722e-01
2.05496401e-01 -1.35101780e-01 8.25539008e-02 -1.00279498e+00
-9.18771327e-01 1.09730470e+00 2.10901201e-01 2.14002058e-01
1.35172236e+00 -1.90281332e-01 1.38472188e+00 9.73908961e-01
9.11771655e-02 -1.29812717e+00 3.53890508e-01 8.53135347e-01
1.12420523e+00 -1.47091794e+00 7.44326189e-02 -7.63807148e-02
-3.10036123e-01 1.10840285e+00 8.59006941e-01 1.44465014e-01
3.38190079e-01 -1.85155511e-01 -1.97083578e-02 -3.69116545e-01
-1.12540114e+00 -1.30886763e-01 6.86375558e-01 -2.61227280e-01
5.41580737e-01 -1.52064517e-01 -6.10390604e-01 7.86417425e-01
-1.29963219e-01 -2.18199387e-01 2.14091450e-01 8.92501295e-01
-7.97787964e-01 -9.05369759e-01 -9.82068777e-02 7.96432018e-01
-7.40332901e-01 -3.37681323e-01 -4.49254960e-01 4.49230403e-01
-4.56004202e-01 1.33790338e+00 -3.74481529e-02 -1.30989090e-01
5.26094198e-01 -5.32038249e-02 1.56815946e-01 -5.35217524e-01
-9.73810613e-01 4.05474938e-02 1.95251465e-01 -4.23144758e-01
-1.89308584e-01 -3.13528836e-01 -1.12407899e+00 -3.49217862e-01
-8.72628033e-01 3.12978476e-01 6.00324929e-01 1.04886198e+00
5.55206180e-01 6.79565370e-01 7.27655590e-01 -1.54487774e-01
-1.19997859e+00 -1.25809312e+00 1.68294385e-02 1.75798953e-01
5.33099234e-01 -1.65988375e-02 -2.46839136e-01 -3.31830263e-01] | [11.162005424499512, 7.856837749481201] |
85785d55-1324-4a27-9a73-11f6d8730f93 | auxiliary-task-based-deep-reinforcement-1 | 2302.14312 | null | https://arxiv.org/abs/2302.14312v1 | https://arxiv.org/pdf/2302.14312v1.pdf | Auxiliary Task-based Deep Reinforcement Learning for Quantum Control | Due to its property of not requiring prior knowledge of the environment, reinforcement learning has significant potential for quantum control problems. In this work, we investigate the effectiveness of continuous control policies based on deep deterministic policy gradient. To solve the sparse reward signal in quantum learning control problems, we propose an auxiliary task-based deep reinforcement learning (AT-DRL) for quantum control. In particular, we first design a guided reward function based on the fidelity of quantum states that enables incremental fidelity improvement. Then, we introduce the concept of an auxiliary task whose network shares parameters with the main network to predict the reward provided by the environment (called the main task). The auxiliary task learns synchronously with the main task, allowing one to select the most relevant features of the environment, thus aiding the agent in comprehending how to achieve the desired state. The numerical simulations demonstrate that the proposed AT-DRL can provide a solution to the sparse reward in quantum systems, and has great potential in designing control pulses that achieve efficient quantum state preparation. | ['Daoyi Dong', 'Sen Kuang', 'Hailan Ma', 'Shumin Zhou'] | 2023-02-28 | null | null | null | null | ['continuous-control'] | ['playing-games'] | [ 7.97831491e-02 -1.04812821e-02 -8.74207541e-02 -5.09163481e-04
-5.70082784e-01 -9.95425358e-02 4.92107093e-01 -7.28311092e-02
-7.49502420e-01 1.05340946e+00 3.04563344e-02 -2.69691616e-01
-1.74028650e-01 -1.00680363e+00 -7.22685575e-01 -1.36195290e+00
8.79905969e-02 8.09698254e-02 -9.65191647e-02 -6.48994982e-01
5.84778845e-01 2.79161602e-01 -1.02226639e+00 -4.51097429e-01
1.11833000e+00 8.26685786e-01 7.41236091e-01 2.36369833e-01
3.48376781e-01 8.61160934e-01 -3.81755322e-01 4.06429231e-01
1.72780722e-01 -8.04608881e-01 -5.31738222e-01 -3.12507480e-01
-4.37617242e-01 -5.46361029e-01 -1.05742145e+00 1.24593294e+00
9.41420674e-01 6.04787767e-01 4.77051258e-01 -6.92462802e-01
-5.84805727e-01 7.79494822e-01 4.02201153e-02 3.16475272e-01
-7.29025900e-02 6.71873510e-01 8.82507622e-01 -3.28397751e-01
6.13570929e-01 9.61701334e-01 -1.66782781e-01 9.21094298e-01
-1.24377394e+00 -7.02185273e-01 -2.25755766e-01 5.29171765e-01
-1.17333806e+00 -4.20217216e-01 8.14415097e-01 -6.79710433e-02
6.96164966e-01 -4.35346663e-01 6.18921638e-01 9.29494739e-01
4.29503500e-01 6.12347364e-01 1.35756063e+00 -5.04208505e-01
7.30078459e-01 1.29317641e-02 -1.57057732e-01 7.77998686e-01
-2.10112318e-01 9.52548981e-01 -5.34854114e-01 -3.30682918e-02
9.08078492e-01 -2.29221374e-01 -3.61498088e-01 -2.09480524e-01
-1.33337891e+00 9.01994169e-01 8.68844032e-01 2.45174661e-01
-8.40310216e-01 6.38790846e-01 1.51892737e-01 3.45538914e-01
-2.80335546e-01 9.46086824e-01 4.34080958e-02 -1.52977094e-01
-3.85704041e-01 3.42816710e-01 2.57581174e-01 6.86311603e-01
1.02687478e+00 3.41018528e-01 -7.49081969e-01 2.85571843e-01
1.95932806e-01 8.88066232e-01 2.43305191e-01 -1.31009555e+00
2.41998047e-01 -1.81984052e-01 4.69293296e-01 -2.98218668e-01
-3.34503949e-01 -4.09874320e-01 -8.34731162e-01 1.92919478e-01
2.40589902e-02 -5.78973889e-01 -7.17925370e-01 2.06166625e+00
2.79315323e-01 6.21702336e-02 3.84812802e-01 1.06239998e+00
3.19006115e-01 9.27766919e-01 -1.92228407e-01 -4.55942094e-01
8.47754776e-01 -7.29539871e-01 -8.95663798e-01 2.08771005e-02
5.63568830e-01 -1.93528682e-01 8.89621139e-01 1.15696996e-01
-8.19309950e-01 -5.09083331e-01 -1.09658873e+00 3.56487632e-01
1.52006969e-01 1.44950384e-02 6.59919739e-01 2.10966587e-01
-8.22801888e-01 1.22267091e+00 -6.75861359e-01 1.01516172e-01
1.12334408e-01 5.14979243e-01 1.03568479e-01 1.12070423e-02
-1.73895407e+00 1.09574378e+00 7.69137740e-01 1.22942373e-01
-1.44891214e+00 -1.62748650e-01 -5.63760877e-01 1.73237324e-01
6.26524091e-01 -6.85899854e-01 1.35839128e+00 -1.99863181e-01
-2.27178335e+00 5.11816666e-02 -1.87703893e-02 -4.18165058e-01
-5.47325648e-02 9.46743116e-02 2.21628919e-02 3.07881624e-01
1.75915539e-01 5.78586936e-01 1.08020234e+00 -6.28745377e-01
-5.09044707e-01 -1.75411299e-01 2.54357159e-01 3.09834629e-01
-1.45388156e-01 -4.55907851e-01 6.66531399e-02 -1.18293114e-01
-6.32398278e-02 -1.07589996e+00 -6.93309903e-01 -5.90794563e-01
-2.51246572e-01 -5.08279264e-01 3.51568818e-01 -7.69133195e-02
9.92176890e-01 -2.13188815e+00 4.88323361e-01 1.14191130e-01
1.47212729e-01 2.83818513e-01 -2.23426521e-01 6.28730893e-01
3.94455254e-01 -2.62555897e-01 -4.30722758e-02 2.75540709e-01
7.30557591e-02 1.65893868e-01 -3.46517712e-01 3.27829361e-01
2.12259069e-01 9.81776595e-01 -1.27696693e+00 -1.69481277e-01
1.05042182e-01 1.83667559e-02 -5.76967657e-01 5.05343437e-01
-3.96711141e-01 1.07659853e+00 -1.19779730e+00 6.40950277e-02
3.08365017e-01 -2.91856289e-01 7.02655166e-02 -3.62333767e-02
-4.71581399e-01 6.01585686e-01 -8.30754936e-01 1.54997361e+00
-2.93173790e-01 1.74883142e-01 2.28074551e-01 -1.20479608e+00
1.07304931e+00 2.58912981e-01 4.24482405e-01 -1.44786727e+00
4.35830176e-01 3.36344659e-01 4.26403642e-01 -4.67470974e-01
4.80121970e-01 -3.49732697e-01 -1.70794412e-01 5.67995429e-01
2.31744066e-01 -5.71078002e-01 1.72277004e-01 3.35526973e-01
9.92437124e-01 1.23024479e-01 2.12459385e-01 -1.77691668e-01
5.75124979e-01 -1.76268741e-01 6.37698591e-01 1.15143478e+00
-4.71431434e-01 -2.66026706e-01 4.17095184e-01 -1.79581389e-01
-1.26651430e+00 -7.43993044e-01 -1.03147412e-02 7.28548050e-01
5.48848867e-01 -1.98076114e-01 -4.76981163e-01 -3.46105784e-01
-1.20126732e-01 7.51506388e-01 -5.18544793e-01 -6.75394893e-01
-6.16596341e-01 -4.79716957e-01 1.36370480e-01 2.37454791e-02
8.02798212e-01 -1.45868385e+00 -5.22761226e-01 6.14483893e-01
5.46299247e-03 -9.41794991e-01 -5.13461173e-01 5.70654571e-01
-7.79628754e-01 -7.37488866e-01 -5.38270295e-01 -3.68171632e-01
4.34932739e-01 2.35682175e-01 4.20261979e-01 -1.31190956e-01
1.92919374e-03 1.67845353e-01 -2.44120568e-01 8.06331169e-03
-5.09910643e-01 9.33636054e-02 4.06293064e-01 -2.17344448e-01
5.19692339e-02 -5.24447143e-01 -6.01876140e-01 -7.50176609e-02
-6.48777127e-01 7.38612935e-02 8.12784493e-01 1.27902675e+00
6.03141248e-01 -5.24675027e-02 7.66582131e-01 -3.49070102e-01
8.97792697e-01 -3.06514114e-01 -8.73877943e-01 6.73674196e-02
-7.19715118e-01 8.93522203e-01 9.27989423e-01 -2.27216721e-01
-8.93417478e-01 -1.46648273e-01 -1.62975058e-01 -2.18283728e-01
2.38851860e-01 4.62171048e-01 3.14792365e-01 -3.27893883e-01
6.36060834e-01 7.08263338e-01 1.18655078e-01 -2.20086146e-03
4.33138728e-01 4.88340169e-01 1.31098285e-01 -9.65574384e-01
8.07018995e-01 7.37589598e-02 5.26242197e-01 -6.72291398e-01
-9.33396101e-01 2.83339265e-04 -4.17367876e-01 -2.16523200e-01
8.95175517e-01 -6.74839854e-01 -1.32755995e+00 3.32304478e-01
-1.01677775e+00 -5.62503874e-01 -3.82596374e-01 9.60018992e-01
-1.07818043e+00 2.90556103e-01 -7.41680205e-01 -9.62067366e-01
-4.06728685e-01 -1.39042854e+00 5.60893416e-01 6.85641646e-01
7.39703119e-01 -5.54922700e-01 1.41709432e-01 -9.87099111e-02
5.02254307e-01 -1.29962221e-01 9.04912353e-01 -7.71351606e-02
-9.95182931e-01 3.49269420e-01 -7.40425512e-02 2.26170391e-01
-1.21159822e-01 -5.24289012e-01 -6.55764937e-01 -5.93203425e-01
1.73409387e-01 -9.03622925e-01 9.50481117e-01 3.85245025e-01
8.47421348e-01 6.83418214e-02 -5.73410019e-02 4.70440418e-01
1.16373122e+00 4.85944241e-01 6.30736709e-01 4.15362000e-01
4.35343444e-01 2.13020384e-01 8.70723307e-01 6.01837814e-01
2.06604853e-01 6.52882814e-01 5.49697757e-01 3.48534316e-01
3.71543139e-01 -3.64565849e-01 5.38232923e-01 8.89577806e-01
-2.14322731e-01 4.03869957e-01 -2.27848426e-01 2.21008109e-03
-2.00437379e+00 -1.13497758e+00 3.63668740e-01 2.27471948e+00
8.50614846e-01 5.65521568e-02 -1.15090959e-01 -3.46914798e-01
7.50677645e-01 2.07152531e-01 -9.95475769e-01 -3.78897458e-01
2.43091285e-01 3.77933711e-01 4.43398207e-01 4.12646443e-01
-6.84229314e-01 1.27945900e+00 5.66958666e+00 1.05273736e+00
-1.32367325e+00 -6.93230927e-02 1.11222066e-01 2.50398040e-01
-2.22723410e-01 3.02978784e-01 -9.72615480e-01 5.43103516e-01
9.01497245e-01 -4.31258470e-01 1.18117905e+00 6.36579931e-01
6.00588083e-01 -1.81796476e-01 -8.54540467e-01 7.09612012e-01
-7.15795934e-01 -1.41589308e+00 -1.62864178e-01 2.89965451e-01
7.61775732e-01 1.31025851e-01 1.35663435e-01 7.45640099e-01
2.01919302e-01 -6.43303871e-01 5.60504019e-01 6.92185342e-01
7.85083175e-01 -8.43471825e-01 4.45449889e-01 8.40939522e-01
-7.94051826e-01 -2.99003124e-01 -6.80918157e-01 -2.21602410e-01
1.27517521e-01 1.49282783e-01 -6.77099645e-01 4.45110470e-01
1.26301587e-01 6.48230553e-01 4.90075350e-02 7.89740264e-01
-7.00466096e-01 4.52245981e-01 -1.58248827e-01 -7.04503298e-01
5.97203910e-01 -6.05715692e-01 4.81374443e-01 4.55533445e-01
4.50436950e-01 4.74799961e-01 3.84307653e-01 1.31878924e+00
-1.93946838e-01 -1.41371578e-01 -6.30629539e-01 -4.96126920e-01
6.28448904e-01 1.07687581e+00 -3.01887900e-01 -1.59714803e-01
5.92701249e-02 7.85304487e-01 6.93014860e-01 3.99507254e-01
-7.39297748e-01 -6.29304469e-01 3.18484604e-01 -4.95637298e-01
4.87762392e-01 -4.67491508e-01 3.56731355e-01 -1.17738533e+00
-4.38609183e-01 -6.35487616e-01 -2.91405261e-01 -4.36292350e-01
-9.23005760e-01 2.92967439e-01 -4.26906228e-01 -9.49190140e-01
-3.23754162e-01 -3.18828136e-01 -7.05251276e-01 1.12342620e+00
-1.79829848e+00 -4.05578256e-01 1.10793814e-01 7.50432014e-01
-5.63231669e-02 -2.46362686e-01 8.35664034e-01 8.98771570e-04
-6.76509500e-01 2.91994631e-01 5.92223585e-01 -1.02164015e-01
5.92704475e-01 -1.31002164e+00 5.83945699e-02 5.45348465e-01
-1.69997200e-01 7.14393616e-01 7.09028900e-01 -4.55455631e-01
-1.76337290e+00 -7.16355622e-01 3.04885656e-01 3.60312968e-01
6.18582070e-01 -8.81027430e-02 -5.67998827e-01 5.75992092e-02
3.51132900e-02 -2.98528206e-02 8.07569399e-02 -1.60030916e-01
6.59222677e-02 -2.20304489e-01 -9.85094368e-01 4.70729083e-01
5.69024682e-01 -8.76596749e-01 -3.30831319e-01 3.01144719e-01
9.96252358e-01 -3.61608356e-01 -7.16099441e-01 2.17667446e-01
1.84566170e-01 -6.28643632e-01 6.28761292e-01 -7.37264037e-01
3.32699984e-01 -3.81815612e-01 -3.04850936e-03 -1.71160364e+00
-7.32019424e-01 -1.01611030e+00 -1.07476674e-01 4.56731677e-01
1.93712488e-01 -6.04360163e-01 7.22078323e-01 3.18387330e-01
-2.81134278e-01 -6.43803060e-01 -9.90335047e-01 -6.53732181e-01
1.73228279e-01 2.25443561e-02 3.70169431e-01 6.69103384e-01
2.85335064e-01 5.98456919e-01 -6.98359787e-01 1.94875985e-01
6.89148307e-01 1.49603531e-01 5.49501061e-01 -6.41712189e-01
-6.36062741e-01 -2.83021033e-01 7.52567425e-02 -1.34143233e+00
1.94442838e-01 -8.69815886e-01 4.06548351e-01 -1.31887639e+00
2.44397461e-01 -6.49673164e-01 -6.00702167e-01 6.73668534e-02
-4.88631427e-01 -4.38349664e-01 3.74909401e-01 3.22420120e-01
-7.38811493e-01 1.37042868e+00 2.02466416e+00 -5.97075447e-02
-3.67264450e-01 9.30935591e-02 -4.84118730e-01 1.17595037e-02
8.32790375e-01 -6.10379457e-01 -2.54999340e-01 -1.37330294e-01
1.61812797e-01 7.50336707e-01 7.97428712e-02 -9.89169180e-01
3.99542809e-01 -4.20109421e-01 -4.44790460e-02 -2.31078401e-01
4.21977222e-01 -2.34906197e-01 -6.09783053e-01 1.00883174e+00
-5.55098772e-01 -4.52000111e-01 -2.74666369e-01 7.50066757e-01
-7.43254973e-03 -6.01626515e-01 9.91691053e-01 -2.61809289e-01
-6.07572258e-01 5.58671415e-01 -4.39260304e-01 5.73067255e-02
8.04051399e-01 6.83264673e-01 -3.98579806e-01 -3.94054621e-01
-8.43379080e-01 5.09784102e-01 6.93546236e-02 -4.30675521e-02
6.21374846e-01 -1.39354563e+00 -4.05430853e-01 5.75898355e-03
-7.22266287e-02 -3.66519600e-01 2.69153655e-01 6.17788553e-01
-1.25096887e-01 7.27419615e-01 -3.88652861e-01 -3.53184909e-01
-3.60908955e-01 6.68779969e-01 4.65643257e-01 -3.52923810e-01
-4.42745566e-01 4.81191278e-01 -7.46403709e-02 -3.67209375e-01
1.11152925e-01 1.55012030e-02 -1.55126631e-01 -4.73387986e-01
4.86413956e-01 1.13699906e-01 -3.06486666e-01 -1.24656871e-01
1.62391916e-01 1.44962043e-01 -1.84114397e-01 -3.02767962e-01
1.29417920e+00 7.99641311e-02 -5.62156066e-02 1.56906381e-01
8.33793521e-01 -4.72084194e-01 -1.56214845e+00 -6.34730041e-01
-1.57131985e-01 -1.19347624e-01 5.07401466e-01 -4.52520221e-01
-5.46628535e-01 1.07046008e+00 5.85953712e-01 1.38396353e-01
8.06466162e-01 -1.49126738e-01 8.21008742e-01 1.14900160e+00
9.38973665e-01 -1.37734795e+00 4.44142342e-01 1.06216896e+00
4.56121087e-01 -1.23450649e+00 -2.08214372e-01 4.18104619e-01
-6.43020928e-01 1.16484213e+00 6.41507268e-01 -3.36074889e-01
2.56665200e-01 -3.86742890e-01 -4.11776334e-01 -2.54838645e-01
-6.63606048e-01 -5.78468680e-01 -1.23374119e-01 4.75948006e-01
5.92738017e-02 1.92892045e-01 -4.28939849e-01 3.26236673e-02
1.05593249e-01 5.87547990e-03 6.24482751e-01 8.51390243e-01
-9.38309133e-01 -1.42238533e+00 -1.50516689e-01 3.07335436e-01
-1.52159274e-01 -5.14686741e-02 1.17489591e-01 3.53981376e-01
-3.48607637e-02 6.70983672e-01 -3.68319571e-01 -4.32270467e-01
2.90429562e-01 -2.93335557e-01 8.54913116e-01 -7.22847521e-01
-2.68401474e-01 3.03577706e-02 -1.55067772e-01 -6.43056571e-01
-1.81635872e-01 -4.24148977e-01 -1.61466229e+00 -3.02749753e-01
-5.40774703e-01 6.77551568e-01 6.46047950e-01 1.18881810e+00
3.95670623e-01 6.22045934e-01 1.26114345e+00 -9.36331451e-01
-1.36815405e+00 -9.71965671e-01 -8.18072259e-01 -3.18357232e-03
4.99684364e-01 -8.03614080e-01 -2.01461866e-01 -7.93440580e-01] | [4.143901824951172, 1.987600326538086] |
c930978c-945c-4a70-b0ef-00ef4d8556b5 | hallucinating-pose-compatible-scenes | 2112.06909 | null | https://arxiv.org/abs/2112.06909v2 | https://arxiv.org/pdf/2112.06909v2.pdf | Hallucinating Pose-Compatible Scenes | What does human pose tell us about a scene? We propose a task to answer this question: given human pose as input, hallucinate a compatible scene. Subtle cues captured by human pose -- action semantics, environment affordances, object interactions -- provide surprising insight into which scenes are compatible. We present a large-scale generative adversarial network for pose-conditioned scene generation. We significantly scale the size and complexity of training data, curating a massive meta-dataset containing over 19 million frames of humans in everyday environments. We double the capacity of our model with respect to StyleGAN2 to handle such complex data, and design a pose conditioning mechanism that drives our model to learn the nuanced relationship between pose and scene. We leverage our trained model for various applications: hallucinating pose-compatible scene(s) with or without humans, visualizing incompatible scenes and poses, placing a person from one generated image into another scene, and animating pose. Our model produces diverse samples and outperforms pose-conditioned StyleGAN2 and Pix2Pix/Pix2PixHD baselines in terms of accurate human placement (percent of correct keypoints) and quality (Frechet inception distance). | ['Alexei A. Efros', 'Tim Brooks'] | 2021-12-13 | null | null | null | null | ['scene-generation'] | ['computer-vision'] | [ 4.15632844e-01 1.16066858e-01 5.17866254e-01 -3.52094293e-01
-6.42610669e-01 -8.05312991e-01 8.68050933e-01 -4.40836608e-01
-2.97131509e-01 4.86751646e-01 6.80696130e-01 2.37813760e-02
4.27814931e-01 -6.73208237e-01 -1.14808381e+00 -2.80273974e-01
1.61076486e-01 5.79213858e-01 2.34148018e-02 -3.21294695e-01
-1.35380745e-01 3.95028800e-01 -1.50283432e+00 5.05070806e-01
4.34938937e-01 6.74325168e-01 3.25281650e-01 1.00498164e+00
4.67786133e-01 8.88909936e-01 -9.20441687e-01 -4.75432098e-01
5.51711798e-01 -6.04426086e-01 -5.57903051e-01 4.14306551e-01
9.95449185e-01 -6.70960367e-01 -5.84550083e-01 7.70474494e-01
6.95060492e-01 1.70739591e-01 5.64987063e-01 -1.33450568e+00
-7.41996765e-01 3.43258262e-01 -5.54675817e-01 -2.60953978e-02
9.81294394e-01 8.48988354e-01 6.99134469e-01 -8.95994544e-01
1.00791144e+00 1.76815736e+00 6.37591124e-01 6.97687268e-01
-1.54626417e+00 -4.50476229e-01 1.72941938e-01 -1.52325884e-01
-1.18026304e+00 -4.24378783e-01 5.83538651e-01 -6.12482607e-01
6.50594056e-01 5.74833274e-01 9.74407434e-01 2.12304521e+00
-5.40337786e-02 8.81442428e-01 1.11015630e+00 -1.15420423e-01
2.33527794e-01 -1.15922689e-01 -4.70523804e-01 4.88826215e-01
9.24215987e-02 1.71808019e-01 -5.36441565e-01 -1.02993418e-02
1.37508059e+00 1.95676044e-01 -3.28176826e-01 -4.35446292e-01
-1.80063558e+00 4.95798558e-01 6.96531057e-01 -3.34557623e-01
-3.62143308e-01 5.10600686e-01 2.97960378e-02 1.86902322e-02
1.32446870e-01 1.12389636e+00 -1.71609223e-01 -6.66202884e-03
-5.89608431e-01 9.28440213e-01 5.87704659e-01 1.29855907e+00
3.94017339e-01 8.67748484e-02 -5.02265930e-01 3.75107259e-01
-7.70571530e-02 8.95964921e-01 -8.06705561e-03 -1.28734267e+00
3.67869049e-01 3.20884854e-01 4.08068776e-01 -1.21005654e+00
-2.85245746e-01 -2.76077896e-01 -6.82625353e-01 2.80244648e-01
5.40831923e-01 -3.92597556e-01 -1.13213480e+00 1.98807156e+00
3.58569980e-01 1.39280766e-01 -2.33840376e-01 1.21412933e+00
8.80813003e-01 7.12332726e-01 2.87253559e-01 4.58384722e-01
1.51329994e+00 -1.01337910e+00 -4.23811734e-01 -6.44230127e-01
5.59010096e-02 -5.69538772e-01 1.56111956e+00 2.72293210e-01
-1.31474888e+00 -9.11066413e-01 -7.59684622e-01 -5.14428437e-01
-1.93409488e-01 1.05830310e-02 6.51757002e-01 1.68811992e-01
-9.44247067e-01 5.29509246e-01 -7.14563191e-01 -5.10112047e-01
3.17582875e-01 2.17602868e-02 -6.62165344e-01 -1.20798737e-01
-9.38700974e-01 8.76325369e-01 1.22816369e-01 -6.89538987e-03
-1.25109613e+00 -7.93654263e-01 -1.15876031e+00 -1.43957451e-01
3.09101373e-01 -1.35503161e+00 1.14044118e+00 -8.13234687e-01
-1.04682863e+00 9.49002981e-01 1.51231006e-01 -3.13335687e-01
1.04218554e+00 -5.85528910e-01 -2.11387828e-01 2.20021829e-01
3.14913750e-01 1.30806911e+00 8.38893533e-01 -1.59249914e+00
-2.25690141e-01 -3.49989831e-01 2.23223016e-01 4.22837526e-01
5.49193956e-02 -1.08797483e-01 -6.38855517e-01 -9.27351952e-01
1.97162107e-02 -1.03152966e+00 -4.19773430e-01 3.35214168e-01
-8.56661737e-01 3.91894996e-01 7.79436231e-01 -7.37246573e-01
5.85345507e-01 -2.11902189e+00 5.25295019e-01 -1.84993342e-01
4.24581259e-01 -1.41468778e-01 -3.41577023e-01 5.26425481e-01
-1.24904796e-01 3.26774940e-02 -1.36342691e-02 -8.15176427e-01
2.32719287e-01 1.23213202e-01 -5.17377794e-01 3.14786553e-01
2.85278022e-01 1.45474839e+00 -1.03011823e+00 -3.24574351e-01
4.70376521e-01 6.99605584e-01 -8.95425498e-01 6.08672678e-01
-5.07786393e-01 8.91495705e-01 -1.32224262e-01 5.92266083e-01
5.72176933e-01 -3.03777039e-01 -1.31094143e-01 -4.57539558e-01
4.26024944e-01 9.41287503e-02 -1.01262796e+00 2.10425353e+00
-1.79101646e-01 5.87139070e-01 -2.03612715e-01 -2.03143898e-02
5.80470026e-01 5.18096127e-02 1.10743329e-01 -4.35205787e-01
3.21101733e-02 -3.57571930e-01 -3.28826368e-01 -7.00682700e-01
6.73989713e-01 -1.45899668e-01 -5.05450666e-01 2.63015896e-01
6.10686317e-02 -5.70145726e-01 -1.52471706e-01 3.11579704e-01
1.07784855e+00 4.73480701e-01 1.08618468e-01 -4.44726162e-02
-3.89505804e-01 -1.76838592e-01 1.37731105e-01 7.83047915e-01
-1.11460082e-01 1.30558205e+00 6.43044233e-01 -6.56490445e-01
-1.52044499e+00 -1.51527631e+00 3.41166317e-01 1.16276526e+00
3.97352546e-01 -3.07867110e-01 -7.78918922e-01 -4.62420940e-01
5.51993623e-02 7.72470772e-01 -1.03773856e+00 -1.52619094e-01
-5.60481727e-01 -2.32951805e-01 4.97254699e-01 7.16488242e-01
5.12291312e-01 -1.37890291e+00 -1.03784513e+00 -2.32163101e-01
-3.88360292e-01 -1.33387351e+00 -7.54769623e-01 -1.86639339e-01
-3.25838417e-01 -9.34902012e-01 -8.79659414e-01 -5.54318309e-01
8.05226088e-01 1.09672941e-01 1.60331416e+00 -3.88660342e-01
-4.99096841e-01 3.89600217e-01 -1.51471958e-01 -2.70595849e-01
-1.72023490e-01 -2.95401424e-01 5.97833879e-02 -5.77136911e-02
-2.68330555e-02 -7.66273618e-01 -8.62060785e-01 1.83196977e-01
-8.31217229e-01 5.83691716e-01 5.53588271e-01 6.06976926e-01
3.18034440e-01 -4.29336697e-01 2.13514958e-02 -1.04796028e+00
5.11113882e-01 -3.69464487e-01 -7.91090429e-02 6.87999055e-02
2.81121373e-01 -1.87808260e-01 3.15111607e-01 -6.22949719e-01
-9.07475173e-01 1.71299174e-01 9.58390385e-02 -9.38645303e-01
-3.77204508e-01 -1.75737932e-01 -4.70566571e-01 4.86155570e-01
1.16746747e+00 8.46349075e-02 -2.16795862e-01 -2.88026363e-01
8.89291048e-01 -7.86803290e-02 1.33033371e+00 -8.71397913e-01
1.03506899e+00 5.80381811e-01 -2.00217336e-01 -5.41480362e-01
-6.77592874e-01 -1.86581705e-02 -7.20482767e-01 -5.32620028e-02
1.39640379e+00 -1.22574723e+00 -8.00135672e-01 3.96038562e-01
-1.34817731e+00 -5.57725370e-01 -4.65909809e-01 3.53633650e-02
-7.52286315e-01 -1.79188311e-01 -6.18307471e-01 -4.70027089e-01
4.28021327e-02 -1.09905457e+00 1.69650471e+00 -6.55010119e-02
-9.04229224e-01 -6.86948657e-01 -1.22187160e-01 3.05005759e-01
2.41434500e-01 1.11630321e+00 4.50914741e-01 -1.52202472e-01
-6.89200699e-01 -9.09902751e-02 -1.71446338e-01 4.08645235e-02
-8.31948593e-02 -2.46877685e-01 -9.97145712e-01 -3.32928330e-01
-2.26375952e-01 -6.26375914e-01 5.38884997e-01 8.64604339e-02
1.19262087e+00 -5.22193789e-01 -2.05141544e-01 9.10761654e-01
1.08106971e+00 -1.84121460e-01 7.94827163e-01 4.95979823e-02
1.21702182e+00 5.43805063e-01 2.42043629e-01 4.96293336e-01
2.93058366e-01 6.71452284e-01 4.62082863e-01 -4.81487602e-01
-2.34277472e-01 -1.05625105e+00 2.67281920e-01 -7.06339180e-02
-1.14665031e-02 -4.53860313e-01 -8.61029685e-01 3.58635932e-01
-1.55829871e+00 -1.20601988e+00 2.49703407e-01 1.80067813e+00
7.29329944e-01 1.81096688e-01 3.31001014e-01 -2.05864653e-01
4.91024286e-01 3.87400627e-01 -6.84057057e-01 9.32681486e-02
-2.64386386e-01 -2.55135931e-02 1.85522556e-01 3.62426639e-01
-1.10352981e+00 9.55509126e-01 6.81922913e+00 2.96853065e-01
-1.11591673e+00 -2.22882345e-01 8.00283432e-01 -3.71551394e-01
-5.78645587e-01 -2.76431829e-01 -3.10997069e-01 3.98822725e-01
3.36532325e-01 2.40414709e-01 4.98827428e-01 9.98081625e-01
1.20840125e-01 7.74988160e-02 -1.64714587e+00 1.29792356e+00
2.59301245e-01 -1.12589180e+00 3.42770010e-01 1.67093854e-02
8.00373197e-01 -3.43887806e-01 6.30529702e-01 2.46452466e-01
6.72942162e-01 -1.51858127e+00 1.21806180e+00 5.87504089e-01
1.16358566e+00 -3.89350206e-01 1.57398462e-01 2.31590748e-01
-7.57749021e-01 4.37547192e-02 -4.08375412e-02 -3.43128324e-01
4.31165069e-01 1.39877722e-01 -9.50559735e-01 6.83308989e-02
6.63521409e-01 5.65560162e-01 -8.40617478e-01 4.93338078e-01
-3.61148030e-01 2.97949719e-03 -1.96643487e-01 2.22435027e-01
1.40521945e-02 1.05308801e-01 5.58814764e-01 1.22166514e+00
2.30750933e-01 2.64451593e-01 2.87111789e-01 1.40502369e+00
-4.47620712e-02 -5.50207496e-01 -8.35548759e-01 1.64289936e-01
6.64545596e-01 1.01528394e+00 -5.03100634e-01 -4.54243392e-01
9.90951881e-02 1.63381779e+00 1.70488536e-01 6.40644968e-01
-9.72886145e-01 1.03856865e-02 7.88482845e-01 4.37607497e-01
4.58464585e-02 -3.03962022e-01 -4.85241771e-01 -1.25585735e+00
5.29013090e-02 -1.25048935e+00 2.02285852e-02 -1.51467824e+00
-1.17593646e+00 7.26119101e-01 1.99589923e-01 -1.19568622e+00
-4.39257145e-01 -5.14526248e-01 -5.51346183e-01 8.63170922e-01
-4.18959439e-01 -1.56948447e+00 -7.13671386e-01 6.20577514e-01
6.49569571e-01 7.55547732e-02 8.09836566e-01 1.91816259e-02
-2.66396493e-01 5.69559216e-01 -6.83296323e-01 2.83126146e-01
7.73287535e-01 -1.46044767e+00 1.08070743e+00 8.22326541e-01
2.05688342e-01 8.54966760e-01 1.03315365e+00 -6.05022371e-01
-1.52983093e+00 -1.19341624e+00 4.54425842e-01 -1.38137877e+00
1.73554927e-01 -9.09276903e-01 -3.23709756e-01 1.06270564e+00
1.45096749e-01 2.68504322e-01 2.96792626e-01 -9.76930931e-02
-7.22338557e-01 3.13989222e-01 -1.08071029e+00 1.22377694e+00
1.52090883e+00 -7.72608697e-01 -5.89553475e-01 2.70946205e-01
1.15650797e+00 -7.68162191e-01 -4.48505849e-01 3.18433821e-01
6.16921604e-01 -1.11453116e+00 1.38394797e+00 -8.19723487e-01
1.09189677e+00 -3.68627220e-01 -3.35736871e-01 -1.42209828e+00
-4.88532722e-01 -7.43174613e-01 9.23325599e-04 6.75694883e-01
1.39976874e-01 -1.32715359e-01 6.95214927e-01 8.41225624e-01
-3.65929082e-02 -6.11117423e-01 -2.43461668e-01 -4.30181623e-01
-8.70674327e-02 -3.21407288e-01 7.73899734e-01 8.61075282e-01
-2.77259946e-01 4.16722983e-01 -7.67895460e-01 1.32361427e-01
5.69641471e-01 -7.68938242e-03 1.58046341e+00 -6.36742413e-01
-7.61563897e-01 -1.68796450e-01 -4.72428024e-01 -1.26625204e+00
-2.11165488e-01 -3.93733472e-01 -1.85732059e-02 -1.37168455e+00
3.29452872e-01 -1.94572583e-01 3.29131067e-01 4.35911119e-01
-4.54895467e-01 5.66979229e-01 6.03470147e-01 1.63795277e-01
-6.81479275e-01 4.50000435e-01 1.72730899e+00 1.09288972e-02
3.43445688e-02 -5.03753424e-01 -8.46928775e-01 7.20834732e-01
1.86278537e-01 1.01198576e-01 -7.45343208e-01 -8.54896724e-01
1.24891162e-01 2.89218247e-01 1.10963869e+00 -1.28859508e+00
-5.70248216e-02 -2.39392668e-01 1.16259038e+00 -3.37108493e-01
8.70438278e-01 -5.77785075e-01 6.51370347e-01 3.05477351e-01
-6.49807215e-01 4.03774619e-01 1.39818087e-01 6.81829870e-01
2.13518649e-01 6.85449898e-01 5.74794114e-01 -4.98581588e-01
-7.55945086e-01 4.26834881e-01 2.63075531e-01 3.05264860e-01
9.99400079e-01 -1.84775576e-01 -3.05843025e-01 -8.09393585e-01
-9.14511502e-01 4.87198420e-02 9.56871748e-01 7.57680535e-01
6.57006919e-01 -1.49933088e+00 -7.70722926e-01 3.84863198e-01
2.09317341e-01 3.29020500e-01 3.26704621e-01 1.37221247e-01
-7.52548099e-01 -8.14100280e-02 -4.89376187e-01 -7.76574194e-01
-8.79247069e-01 8.24159861e-01 2.38499001e-01 2.70750731e-01
-7.49073386e-01 1.23131955e+00 9.33713913e-01 -4.66809005e-01
1.37079000e-01 -2.39493832e-01 4.50032681e-01 -3.69152963e-01
5.30377269e-01 1.41406223e-01 -5.73394895e-01 -6.16010964e-01
-9.96603668e-02 2.84857780e-01 2.36381754e-01 -5.95498085e-01
1.11135077e+00 6.54277354e-02 2.20672831e-01 3.59958500e-01
1.22705221e+00 -1.78941209e-02 -1.95038819e+00 -9.02216230e-03
-7.50799060e-01 -7.13707745e-01 -6.63499653e-01 -1.00458086e+00
-6.50161505e-01 8.01965356e-01 2.68054396e-01 -3.75506014e-01
7.80454397e-01 1.04417816e-01 6.96643770e-01 2.73411244e-01
6.09002411e-01 -5.60249209e-01 8.08787644e-01 2.18194962e-01
1.38077903e+00 -1.24494779e+00 -2.29026124e-01 -2.88695723e-01
-9.52118278e-01 5.05417645e-01 1.05000162e+00 -4.19475496e-01
1.62501872e-01 2.91238219e-01 1.60778359e-01 -2.68359452e-01
-6.29130960e-01 9.27197412e-02 3.37168157e-01 7.72461593e-01
1.38078004e-01 3.27747226e-01 5.37705839e-01 5.11873841e-01
-9.37423944e-01 -4.13860261e-01 1.99970275e-01 6.92762017e-01
-1.04076385e-01 -4.90885526e-01 -5.91916502e-01 1.19528873e-03
1.33071497e-01 -3.05208098e-02 -6.22771680e-01 7.68275440e-01
4.12070543e-01 6.42927825e-01 1.40448853e-01 -6.62127435e-01
5.23079693e-01 -2.85702884e-01 7.17912436e-01 -7.33022392e-01
-3.26063901e-01 -2.76974943e-02 -3.70755047e-02 -8.92596066e-01
3.47601660e-02 -5.91313422e-01 -6.87956810e-01 -4.70156610e-01
3.97006989e-01 -4.57849890e-01 2.93412209e-01 6.60340369e-01
4.44158196e-01 6.51475072e-01 2.91331828e-01 -1.53759992e+00
-4.16984618e-01 -9.71758246e-01 -2.72238195e-01 1.14280319e+00
3.03861946e-01 -5.07904112e-01 -1.58258155e-01 4.32819247e-01] | [11.83080768585205, -0.6829266548156738] |
91770b92-ac0b-46c4-ae08-faf61048867b | leveraging-native-language-speech-for-accent | 1712.08992 | null | http://arxiv.org/abs/1712.08992v2 | http://arxiv.org/pdf/1712.08992v2.pdf | Leveraging Native Language Speech for Accent Identification using Deep Siamese Networks | The problem of automatic accent identification is important for several
applications like speaker profiling and recognition as well as for improving
speech recognition systems. The accented nature of speech can be primarily
attributed to the influence of the speaker's native language on the given
speech recording. In this paper, we propose a novel accent identification
system whose training exploits speech in native languages along with the
accented speech. Specifically, we develop a deep Siamese network-based model
which learns the association between accented speech recordings and the native
language speech recordings. The Siamese networks are trained with i-vector
features extracted from the speech recordings using either an unsupervised
Gaussian mixture model (GMM) or a supervised deep neural network (DNN) model.
We perform several accent identification experiments using the CSLU Foreign
Accented English (FAE) corpus. In these experiments, our proposed approach
using deep Siamese networks yield significant relative performance improvements
of 15.4 percent on a 10-class accent identification task, over a baseline
DNN-based classification system that uses GMM i-vectors. Furthermore, we
present a detailed error analysis of the proposed accent identification system. | ['Sriram Ganapathy', 'Aditya Siddhant', 'Preethi Jyothi'] | 2017-12-25 | null | null | null | null | ['speaker-profiling'] | ['speech'] | [ 2.58592423e-02 -6.31005615e-02 -5.34374118e-02 -7.88466036e-01
-7.66388535e-01 -6.04087353e-01 4.05833632e-01 -2.12649047e-01
-6.89056218e-01 5.20702660e-01 4.94950563e-01 -5.06560445e-01
2.79397547e-01 -2.86533684e-01 -5.53522646e-01 -7.10256517e-01
7.05639226e-03 6.94725871e-01 -4.73903745e-01 -2.82249749e-01
1.38016799e-02 6.68046653e-01 -1.00481093e+00 -2.75940299e-01
7.49519289e-01 8.92133176e-01 1.86782658e-01 8.63198459e-01
-3.11109662e-01 8.58014107e-01 -1.07464683e+00 -2.92112172e-01
-7.50554949e-02 -3.26013803e-01 -9.49849069e-01 1.29119337e-01
5.16954482e-01 -3.06878723e-02 -2.49315441e-01 1.37897861e+00
4.50885117e-01 3.89420778e-01 5.61458945e-01 -8.16077471e-01
-4.26527560e-01 1.04531455e+00 -1.92630932e-01 5.30321360e-01
-1.70188636e-01 -3.52818996e-01 7.90844917e-01 -9.79306400e-01
2.38635421e-01 1.32180905e+00 4.52042848e-01 5.70898294e-01
-1.11652195e+00 -7.91552126e-01 1.29996255e-01 1.22865431e-01
-1.59691060e+00 -8.94586325e-01 1.08803546e+00 -2.68349171e-01
7.94742703e-01 2.98920095e-01 7.15562403e-02 1.15976584e+00
-3.41304988e-02 9.25257564e-01 9.73030567e-01 -6.68387473e-01
4.10161555e-01 1.92551449e-01 3.94901305e-01 2.89939821e-01
-6.53064430e-01 3.86858582e-02 -5.27854562e-01 7.84043223e-03
4.57204670e-01 -4.68997657e-01 -9.06563029e-02 7.81364292e-02
-1.08379424e+00 1.00024450e+00 1.78873688e-01 4.66467202e-01
-5.74854076e-01 -3.47061396e-01 4.63443726e-01 1.28582433e-01
6.28761828e-01 2.35322580e-01 -6.31087780e-01 -2.26537719e-01
-9.03854668e-01 -8.24785009e-02 9.17982578e-01 6.72401428e-01
5.29501915e-01 6.55043900e-01 3.87928963e-01 1.54219985e+00
3.26719701e-01 7.11311817e-01 9.75081205e-01 -7.28077829e-01
3.24573547e-01 8.86682887e-04 -3.07441920e-01 -5.42011738e-01
-3.49457562e-01 -6.83798432e-01 -1.01386511e+00 1.25606045e-01
4.04202759e-01 -5.75661898e-01 -9.78737473e-01 2.00989580e+00
1.77042976e-01 1.23163179e-01 5.82256794e-01 7.05072045e-01
4.35373098e-01 7.46083200e-01 1.56851783e-01 -1.79563269e-01
1.28328824e+00 -1.05654776e+00 -1.04733169e+00 -4.32022095e-01
4.76529598e-01 -1.14179289e+00 8.20817590e-01 1.57664135e-01
-8.74050438e-01 -8.49528491e-01 -1.02246988e+00 3.62878859e-01
-5.81618428e-01 3.37557465e-01 7.54509354e-03 1.00427580e+00
-1.11171782e+00 -1.12475134e-01 -8.22696090e-01 -2.72422224e-01
1.94351301e-02 6.31988823e-01 -4.42102343e-01 4.75354701e-01
-1.32589805e+00 7.46143818e-01 5.00641465e-01 8.29741359e-02
-5.54867566e-01 -5.21209002e-01 -1.04425621e+00 1.52675554e-01
-1.09321900e-01 1.26609981e-01 1.57345474e+00 -1.25615621e+00
-1.73844326e+00 6.08925581e-01 -4.78275478e-01 -8.05186749e-01
-1.43927122e-02 -3.26496363e-01 -9.13249314e-01 -2.57665396e-01
-7.42105320e-02 4.69960332e-01 7.95187235e-01 -8.98387969e-01
-8.79502654e-01 -5.77538610e-01 -4.38350320e-01 2.78961837e-01
-3.20030749e-01 3.90051454e-01 6.35357276e-02 -1.03885281e+00
3.06755042e-04 -1.12051368e+00 -6.93250522e-02 -1.07782042e+00
-7.09829152e-01 -3.64639878e-01 1.12878132e+00 -9.25848484e-01
1.06916010e+00 -2.62272811e+00 -4.88115586e-02 3.93147379e-01
-1.50862008e-01 4.99124855e-01 -2.55440045e-02 1.54424328e-02
-3.23929131e-01 -9.03892294e-02 -1.08704895e-01 -6.23410881e-01
1.47035956e-01 2.17604399e-01 -4.04068917e-01 3.06382418e-01
1.45581871e-01 2.85304993e-01 -3.86218756e-01 -2.29088470e-01
2.50624388e-01 6.21883035e-01 -1.43682256e-01 5.18142879e-01
1.63804471e-01 3.19093674e-01 9.83453915e-02 3.96783203e-01
8.20290804e-01 5.86816728e-01 4.08636957e-01 -1.09793216e-01
-2.34234840e-01 6.16795242e-01 -1.18413043e+00 1.34562647e+00
-7.26514518e-01 7.73063600e-01 5.18161118e-01 -9.85879183e-01
1.23737824e+00 5.23655355e-01 9.06842574e-03 -2.63331741e-01
-1.34360287e-02 3.02954823e-01 3.72342527e-01 1.71256587e-01
6.23626292e-01 -2.57197350e-01 -2.04162106e-01 2.71187633e-01
4.49489594e-01 3.33487600e-01 -7.72348046e-02 -2.70572975e-02
5.94497979e-01 -4.90638286e-01 1.97401673e-01 -5.12705386e-01
7.82652736e-01 -3.00096422e-01 6.68252587e-01 5.73812068e-01
-4.54062134e-01 3.65330458e-01 9.97251272e-02 -3.54348958e-01
-7.94081986e-01 -1.21629679e+00 -3.83867383e-01 1.44953501e+00
-3.91789824e-01 -4.71284688e-02 -1.19561958e+00 -8.07164967e-01
-3.66549700e-01 1.00269639e+00 -3.74954760e-01 9.30064395e-02
-8.86321366e-01 -6.98326230e-01 9.05858278e-01 8.38453472e-01
6.02109611e-01 -1.19016206e+00 3.08492601e-01 3.08789879e-01
-1.44083813e-01 -1.23768818e+00 -9.89635766e-01 6.65184736e-01
-3.02470863e-01 -4.24403727e-01 -5.23421109e-01 -1.40942562e+00
2.96725243e-01 -4.52401489e-01 7.80863106e-01 -5.76861680e-01
3.31339687e-01 8.45245570e-02 -2.22193405e-01 -4.14212286e-01
-9.75312948e-01 6.89767063e-01 7.19764531e-01 4.65633035e-01
9.18872654e-01 -3.54753882e-01 5.14644943e-02 1.82688445e-01
-6.09805465e-01 -5.78582942e-01 5.76242745e-01 1.13384116e+00
4.03159231e-01 2.25603834e-01 9.01698291e-01 -8.46020162e-01
4.40087348e-01 -3.17958117e-01 -6.55066431e-01 -1.10782333e-01
-2.39347383e-01 9.48558748e-02 7.19515264e-01 -3.61331761e-01
-1.34354937e+00 4.38663483e-01 -9.87080872e-01 -3.05874556e-01
-8.58950078e-01 5.67732692e-01 -6.08224690e-01 1.38044208e-01
2.86648154e-01 3.57557595e-01 7.64765888e-02 -7.17171073e-01
1.74544349e-01 1.29123735e+00 1.06873822e+00 -2.84506142e-01
4.39745903e-01 -2.95602828e-01 -6.42932117e-01 -1.40002763e+00
-6.66865110e-01 -5.13842881e-01 -8.25052202e-01 3.02900579e-02
1.05199504e+00 -8.21262777e-01 -6.00498676e-01 8.84896338e-01
-1.24755669e+00 -2.44048283e-01 4.32952726e-03 7.39650548e-01
-3.57124686e-01 1.07143708e-01 -8.28904569e-01 -9.28225338e-01
-2.33598292e-01 -1.43648374e+00 7.73662210e-01 3.00265998e-01
-3.82809967e-01 -1.34938443e+00 2.36179724e-01 4.24243361e-01
5.65425932e-01 -5.76293826e-01 8.21214378e-01 -1.52092421e+00
-6.16981946e-02 -7.17225596e-02 2.66343683e-01 9.87698197e-01
5.32018900e-01 -1.59726486e-01 -1.42462647e+00 -1.60596639e-01
1.85867786e-01 -1.27419844e-01 3.66609275e-01 5.89956105e-01
8.70623291e-01 -2.92287648e-01 1.06282704e-01 6.14395976e-01
8.71203303e-01 7.49154866e-01 3.83372009e-01 2.93327391e-01
7.76551902e-01 4.81319100e-01 2.95748502e-01 -1.18375994e-01
2.55355090e-01 6.57511592e-01 -3.21213380e-02 -3.52692246e-01
-5.72036617e-02 -1.24093637e-01 5.00204206e-01 1.53082943e+00
3.45662594e-01 -1.89103603e-01 -1.05559099e+00 7.09185719e-01
-1.21826792e+00 -8.22512984e-01 2.91830003e-01 2.20001864e+00
8.01284611e-01 1.69593439e-01 2.29084641e-01 1.81948349e-01
1.05747354e+00 2.77674526e-01 -3.58235985e-01 -6.87070191e-01
-4.34402674e-01 5.70340097e-01 4.22060043e-01 9.99727368e-01
-1.73479581e+00 1.06814110e+00 6.14880514e+00 8.51349354e-01
-1.38483047e+00 1.31986842e-01 8.97377551e-01 4.19905335e-01
3.68478060e-01 -5.24280488e-01 -9.88695264e-01 4.20714468e-01
1.55125511e+00 -8.77575055e-02 2.48279244e-01 1.04532754e+00
2.78230384e-02 3.08532298e-01 -8.74654353e-01 8.75621676e-01
9.89602134e-02 -1.10926020e+00 -1.64638385e-01 2.37095103e-01
4.99789685e-01 3.38570625e-01 3.55508655e-01 3.63292217e-01
3.92462462e-01 -9.82799947e-01 6.24675632e-01 -3.46060768e-02
4.90026444e-01 -1.26662803e+00 1.19130754e+00 1.67668164e-01
-1.00398290e+00 2.97289133e-01 -1.95989441e-02 1.37019575e-01
9.04263183e-03 2.17584357e-01 -1.27486658e+00 1.95418939e-01
5.12931168e-01 1.68274403e-01 -2.20565766e-01 3.70958507e-01
1.21952668e-01 1.36736524e+00 -2.71996886e-01 2.11705100e-02
3.12181056e-01 -2.53582686e-01 6.60651147e-01 1.68952191e+00
5.43151572e-02 -3.18467051e-01 2.07337067e-01 5.27063906e-01
-3.48186255e-01 2.86891699e-01 -3.41073841e-01 -2.72098213e-01
6.97246850e-01 1.03565574e+00 -3.91496271e-01 -3.64167601e-01
-4.19802755e-01 9.86708105e-01 4.21391904e-01 5.26716828e-01
-4.45346653e-01 -7.60642827e-01 1.20763457e+00 -4.18527514e-01
3.95959735e-01 -3.33490223e-01 -1.70894146e-01 -7.45508134e-01
-2.42457166e-01 -1.03420413e+00 1.70599017e-02 -2.62603998e-01
-1.32109332e+00 1.19061995e+00 -4.00192201e-01 -5.79741478e-01
-8.26773405e-01 -7.31662571e-01 -6.57476246e-01 1.41940773e+00
-1.13094246e+00 -8.10049593e-01 2.94887304e-01 3.88801068e-01
9.44840431e-01 -8.89239013e-01 1.26274395e+00 3.83635998e-01
-7.86820412e-01 9.35766160e-01 3.48356277e-01 8.87555063e-01
7.68799245e-01 -1.67112267e+00 8.34236801e-01 9.92028177e-01
4.54808950e-01 6.24437332e-01 5.64688206e-01 -2.24742293e-01
-8.96387279e-01 -1.24523616e+00 1.26132476e+00 -1.80436403e-01
7.12610006e-01 -4.94545758e-01 -9.86434281e-01 9.59535360e-01
6.40111327e-01 -2.44858682e-01 1.08031285e+00 3.94725025e-01
-1.43449157e-01 -2.52355933e-01 -8.77661765e-01 2.94961005e-01
3.69308680e-01 -9.43166256e-01 -7.64389038e-01 1.15463533e-01
6.91688359e-01 -3.29383790e-01 -8.61050963e-01 1.44668072e-01
2.37075225e-01 -6.92702711e-01 7.64229178e-01 -6.53652251e-01
-1.38508409e-01 -2.67942846e-01 -6.06453836e-01 -1.90567017e+00
-2.53789365e-01 -4.21905249e-01 3.40391189e-01 1.64559352e+00
6.49497271e-01 -6.00636601e-01 7.07721591e-01 2.79776275e-01
-2.90796369e-01 -5.74235655e-02 -1.10946643e+00 -7.37362206e-01
2.53787130e-01 -4.46867675e-01 7.44024634e-01 1.18143046e+00
-1.91746429e-01 7.00466454e-01 -1.74567208e-01 6.07290864e-01
5.29207587e-01 -3.96644443e-01 5.14811337e-01 -9.32651997e-01
-9.86153185e-02 -4.26958114e-01 -7.52851486e-01 -9.07602072e-01
8.40232790e-01 -6.29271507e-01 4.79093254e-01 -5.82818747e-01
-3.26713145e-01 -3.95174265e-01 -5.59993625e-01 2.11724773e-01
-2.42702693e-01 4.39424850e-02 1.59490053e-02 -1.72567919e-01
-1.09622553e-01 4.95841682e-01 1.98583946e-01 -3.74096423e-01
-2.96868861e-01 3.76230687e-01 -2.20895335e-01 8.64278615e-01
1.16743350e+00 -2.16278568e-01 -1.71076044e-01 -3.29173684e-01
-7.93133557e-01 -3.47109698e-02 -1.70533657e-01 -1.04987907e+00
4.10681456e-01 1.54646173e-01 2.18545228e-01 -5.71326911e-01
4.89486009e-01 -5.41634083e-01 -1.30393475e-01 5.41241616e-02
-4.70844448e-01 2.07473397e-01 3.66834730e-01 1.01877920e-01
-9.22682405e-01 -1.67094037e-01 1.01809049e+00 2.53627598e-01
-7.27977335e-01 -9.93225947e-02 -9.02429879e-01 -3.36711891e-02
3.90158147e-01 3.48426133e-01 -1.85763136e-01 -4.93936121e-01
-1.02456057e+00 -2.41051689e-01 7.00391456e-02 5.06803274e-01
3.10127288e-01 -1.44659483e+00 -7.44796097e-01 8.00229907e-01
-2.11644202e-01 -3.76265228e-01 2.21378461e-01 4.28184479e-01
-3.40705395e-01 5.92412114e-01 -1.07841186e-01 -6.05455399e-01
-1.33300400e+00 3.94959152e-01 6.04936779e-01 8.30868408e-02
6.75121695e-02 1.06544983e+00 5.46058953e-01 -9.98797953e-01
5.95044017e-01 -1.74498990e-01 -2.09866762e-01 -2.74305165e-01
5.47572911e-01 1.72462732e-01 2.17991173e-01 -1.33126342e+00
-5.68877816e-01 6.24941289e-02 -5.28216124e-01 -4.46018517e-01
8.22144985e-01 -1.93429723e-01 -1.21525154e-02 6.57875538e-01
1.31221116e+00 4.41978306e-01 -6.84593737e-01 -2.56389648e-01
2.54598200e-01 1.49036884e-01 5.02783597e-01 -7.98505783e-01
-1.07906115e+00 7.82037377e-01 9.19025242e-01 3.18274856e-01
1.07954121e+00 -1.67891718e-02 7.59524167e-01 3.66877317e-01
-2.16601089e-01 -1.37610066e+00 -6.35817111e-01 8.41879249e-01
5.29055893e-01 -1.31287086e+00 -1.02511573e+00 -1.84969932e-01
-7.28923619e-01 9.91601884e-01 5.00998020e-01 1.35276884e-01
9.75103080e-01 5.00437498e-01 1.02113116e+00 2.09129542e-01
-2.96737820e-01 -1.07735269e-01 1.80608973e-01 6.07575417e-01
8.52148831e-01 5.39183617e-01 2.96635002e-01 6.17286026e-01
-7.58716285e-01 -6.81993723e-01 3.02349687e-01 5.67644298e-01
-3.04247826e-01 -1.23218322e+00 -7.27942526e-01 1.49700746e-01
-7.34823942e-01 -2.82833189e-01 -4.64816332e-01 4.80858624e-01
-9.75070223e-02 1.16304529e+00 5.26371479e-01 -5.80840409e-01
2.52057642e-01 6.32347047e-01 -1.99681744e-01 -3.43388975e-01
-4.24024284e-01 3.77327472e-01 2.29522660e-01 -5.27747683e-02
-4.55218554e-01 -7.93261230e-01 -1.03324485e+00 -1.92112073e-01
-1.30229235e-01 6.57018483e-01 1.06155801e+00 1.15560412e+00
1.35916695e-01 4.97354835e-01 8.81800115e-01 -6.00006640e-01
-5.46599925e-01 -1.41491675e+00 -1.03926110e+00 1.27209172e-01
7.06778467e-01 -3.96827012e-01 -4.19595331e-01 3.06459218e-01] | [14.3590669631958, 6.700863361358643] |
f2a8c440-eff9-4078-b673-1e29b9a0afa5 | multi-task-item-attribute-graph-pre-training | 2306.14462 | null | https://arxiv.org/abs/2306.14462v1 | https://arxiv.org/pdf/2306.14462v1.pdf | Multi-task Item-attribute Graph Pre-training for Strict Cold-start Item Recommendation | Recommendation systems suffer in the strict cold-start (SCS) scenario, where the user-item interactions are entirely unavailable. The ID-based approaches completely fail to work. Cold-start recommenders, on the other hand, leverage item contents to map the new items to the existing ones. However, the existing SCS recommenders explore item contents in coarse-grained manners that introduce noise or information loss. Moreover, informative data sources other than item contents, such as users' purchase sequences and review texts, are ignored. We explore the role of the fine-grained item attributes in bridging the gaps between the existing and the SCS items and pre-train a knowledgeable item-attribute graph for SCS item recommendation. Our proposed framework, ColdGPT, models item-attribute correlations into an item-attribute graph by extracting fine-grained attributes from item contents. ColdGPT then transfers knowledge into the item-attribute graph from various available data sources, i.e., item contents, historical purchase sequences, and review texts of the existing items, via multi-task learning. To facilitate the positive transfer, ColdGPT designs submodules according to the natural forms of the data sources and coordinates the multiple pre-training tasks via unified alignment-and-uniformity losses. Our pre-trained item-attribute graph acts as an implicit, extendable item embedding matrix, which enables the SCS item embeddings to be easily acquired by inserting these items and propagating their attributes' embeddings. We carefully process three public datasets, i.e., Yelp, Amazon-home, and Amazon-sports, to guarantee the SCS setting for evaluation. Extensive experiments show that ColdGPT consistently outperforms the existing SCS recommenders by large margins and even surpasses models that are pre-trained on 75-224 times more, cross-domain data on two out of four datasets. | ['Philip S. Yu', 'Chenyu You', 'Hao Peng', 'Zhiwei Liu', 'Chen Wang', 'Liangwei Yang', 'Yuwei Cao'] | 2023-06-26 | null | null | null | null | ['multi-task-learning'] | ['methodology'] | [-3.51168513e-01 -5.27022302e-01 -7.34159410e-01 -4.89905596e-01
-7.24970222e-01 -7.71866024e-01 3.60539764e-01 -1.34755373e-01
-4.67623264e-01 5.64366400e-01 5.24736702e-01 -2.32650135e-02
-3.00181836e-01 -6.96426988e-01 -7.58956611e-01 -7.05345929e-01
1.33669019e-01 5.95372200e-01 -1.04511298e-01 -3.23054880e-01
1.31309390e-01 -1.13569424e-01 -1.26795495e+00 3.95526439e-01
1.03333414e+00 1.03567195e+00 3.00207049e-01 1.52051464e-01
-3.78927141e-01 5.36556095e-02 -2.01405212e-01 -7.28141308e-01
4.25925106e-01 -1.98836233e-02 -3.24304342e-01 -2.43279621e-01
1.70417398e-01 -2.83238351e-01 -2.95029491e-01 8.15769255e-01
2.63799459e-01 3.60662639e-01 7.64025152e-01 -1.31657839e+00
-1.63895905e+00 8.75260174e-01 -5.93636572e-01 5.56018651e-02
3.31105590e-01 6.05801344e-02 1.49758315e+00 -1.38636625e+00
4.31508750e-01 1.00379634e+00 7.57904887e-01 4.24281448e-01
-1.02508891e+00 -8.26317072e-01 6.80350602e-01 1.09894753e-01
-1.33781171e+00 1.56875700e-02 5.82759559e-01 -3.84981513e-01
8.61763477e-01 4.16963100e-02 4.68427211e-01 1.61375415e+00
8.95710364e-02 9.11423802e-01 7.46519864e-01 -1.02674246e-01
6.83217719e-02 4.64307398e-01 5.78369737e-01 1.33281663e-01
2.69290060e-01 5.22904992e-02 -5.78290522e-01 -2.26322651e-01
5.17545044e-01 7.00751603e-01 -2.47122437e-01 -4.14716482e-01
-1.23541415e+00 8.88189673e-01 2.60139972e-01 2.25104570e-01
-4.85445827e-01 -3.92768592e-01 4.25657243e-01 5.87469995e-01
5.25247693e-01 4.62373376e-01 -9.64487731e-01 1.11203574e-01
-4.14246023e-01 4.81147058e-02 8.18873882e-01 1.45220208e+00
8.66479993e-01 -5.01453541e-02 -1.00733489e-01 9.24780726e-01
5.11778235e-01 6.84233785e-01 7.05097556e-01 -1.36297002e-01
7.73970842e-01 4.80378091e-01 1.58853590e-01 -9.20119822e-01
-1.05401866e-01 -7.74977565e-01 -7.04582393e-01 -6.25166655e-01
1.26051232e-01 -2.60077477e-01 -6.76421165e-01 1.93016613e+00
3.32908899e-01 3.59899789e-01 1.92155037e-02 1.06057525e+00
7.19871640e-01 8.09392571e-01 1.89167902e-01 -7.40460902e-02
1.33507812e+00 -1.22548294e+00 -5.99482179e-01 -1.65107325e-01
7.82364964e-01 -5.97614706e-01 1.63251424e+00 4.43900287e-01
-5.54800451e-01 -7.81726718e-01 -9.74366069e-01 4.60955054e-02
-5.99900723e-01 3.58541071e-01 6.64130807e-01 4.61077869e-01
-3.79792422e-01 4.29950505e-01 -4.04859215e-01 -6.29596785e-02
2.75252163e-01 3.64115924e-01 -4.69436944e-01 -2.79507756e-01
-1.53179467e+00 5.41455269e-01 1.76030979e-01 -9.23295468e-02
-7.26349115e-01 -1.17219067e+00 -6.71674073e-01 2.16829479e-01
5.52919447e-01 -8.06938946e-01 7.74603486e-01 -9.96093571e-01
-1.34489346e+00 3.67647499e-01 1.20426893e-01 7.48064043e-03
-5.09980991e-02 -7.30535448e-01 -9.44473565e-01 -4.33943093e-01
1.20619029e-01 4.86332476e-02 8.99407208e-01 -1.14479446e+00
-7.41582274e-01 -4.24323827e-01 3.15659866e-02 3.61898661e-01
-8.77924263e-01 -1.86776459e-01 -7.98904240e-01 -9.19748187e-01
-3.50756705e-01 -1.16467786e+00 -7.28709064e-03 -3.82404804e-01
-3.15101981e-01 -3.39003205e-01 5.11476576e-01 -3.57011437e-01
1.41264689e+00 -2.44669461e+00 1.16708376e-01 3.09482187e-01
1.07298188e-01 2.80215591e-01 -6.26008630e-01 5.34653008e-01
2.97487527e-02 7.05112368e-02 2.77581483e-01 -5.77674866e-01
2.48083383e-01 4.46360469e-01 -4.46850836e-01 2.37714350e-01
-1.84516668e-01 1.01669288e+00 -8.63214374e-01 -8.32422450e-02
-2.05240659e-02 4.18325394e-01 -6.62291110e-01 3.35401297e-01
-1.64032981e-01 2.34919980e-01 -7.78389275e-01 4.98269975e-01
6.57002747e-01 -3.75533849e-01 1.12668067e-01 -3.44772696e-01
2.70668477e-01 5.21306038e-01 -1.19397771e+00 1.98799181e+00
-7.35752642e-01 -1.63896337e-01 -2.79427290e-01 -8.03139687e-01
9.50900555e-01 1.72863498e-01 5.35321414e-01 -7.85071433e-01
-1.68341503e-01 1.70658872e-01 -9.75128114e-02 -3.68982464e-01
6.49622440e-01 -7.96392038e-02 -3.79906356e-01 7.65579522e-01
3.33426446e-01 7.83612490e-01 -5.88131621e-02 3.67529601e-01
8.27076077e-01 1.75067931e-01 1.31713063e-01 -1.77859396e-01
4.05429184e-01 -4.10700500e-01 7.43360579e-01 6.91786528e-01
2.40121245e-01 6.81319356e-01 1.52076799e-02 -4.05752003e-01
-8.74746740e-01 -1.22238517e+00 -4.27834578e-02 1.74613428e+00
3.67479235e-01 -6.72479391e-01 -1.55225113e-01 -1.25665951e+00
4.83388841e-01 6.93480551e-01 -8.85788918e-01 -5.16653717e-01
-3.74853939e-01 -6.96733057e-01 1.05496578e-01 7.83465564e-01
-1.76156685e-01 -8.88088882e-01 4.70682770e-01 4.27066475e-01
-7.29923099e-02 -7.99344420e-01 -1.15835321e+00 2.41695598e-01
-5.52807629e-01 -7.82346666e-01 -4.45508003e-01 -4.80106115e-01
5.21696031e-01 5.99702656e-01 1.23890734e+00 2.79752798e-02
3.01027954e-01 2.65834242e-01 -8.01568747e-01 -3.94371860e-02
5.87460548e-02 4.32350069e-01 5.03037989e-01 2.87303329e-01
8.50425184e-01 -6.84073746e-01 -7.41490424e-01 7.53206849e-01
-8.82347703e-01 -3.28716904e-01 6.89913869e-01 1.09908938e+00
7.33394206e-01 -2.66171604e-01 1.11060762e+00 -1.38948691e+00
6.03552282e-01 -1.14934242e+00 -1.68801278e-01 4.29449916e-01
-1.13484299e+00 5.59503213e-02 1.07835233e+00 -8.76460791e-01
-1.12454617e+00 -3.61473024e-01 -2.07642719e-01 -5.40380716e-01
-1.70721635e-02 6.93620861e-01 -5.25813639e-01 4.65627789e-01
4.50468630e-01 1.17382102e-01 -2.86915451e-01 -9.87460256e-01
7.13753998e-01 6.98580444e-01 3.08856636e-01 -6.67212844e-01
8.87210906e-01 6.02640063e-02 -6.55931652e-01 -1.41732544e-01
-1.18960452e+00 -8.71928811e-01 -6.13659382e-01 4.10867065e-01
4.36476648e-01 -1.05380166e+00 -4.02216494e-01 8.00064877e-02
-7.23173201e-01 1.05242759e-01 -5.38475633e-01 7.75332987e-01
-1.74114123e-01 1.76846936e-01 -7.48160064e-01 -2.24162623e-01
-5.77421784e-01 -8.41488779e-01 8.80251586e-01 1.53666809e-01
-1.31581664e-01 -1.14448297e+00 2.75473565e-01 1.98776737e-01
4.81918037e-01 -6.04038477e-01 1.10625422e+00 -1.35386658e+00
-4.17893887e-01 -2.75311351e-01 -1.51930317e-01 5.21722555e-01
3.70239675e-01 -2.45851591e-01 -6.66881502e-01 -4.35816199e-01
-2.06246823e-01 -2.02052012e-01 7.94793189e-01 1.67454049e-01
9.91572797e-01 -4.51552093e-01 -4.75556552e-01 8.19089711e-01
1.48441672e+00 1.37345567e-01 4.04932886e-01 2.26930857e-01
1.03288615e+00 2.58078426e-01 8.67787004e-01 5.39022863e-01
6.37932062e-01 7.49315679e-01 1.28267705e-01 8.56954455e-02
1.24548197e-01 -7.26664007e-01 3.70858550e-01 1.36438680e+00
1.29740521e-01 -5.81350803e-01 -2.80685782e-01 4.01241869e-01
-1.78690922e+00 -7.59277642e-01 -3.75982821e-02 2.27241278e+00
9.16434526e-01 4.44253460e-02 6.60502240e-02 -4.63523030e-01
3.81914347e-01 -8.41185302e-02 -7.09096074e-01 -1.38539106e-01
-9.62690488e-02 -8.10235962e-02 3.64918500e-01 9.22230855e-02
-1.08604741e+00 8.72041106e-01 4.97106695e+00 9.61957932e-01
-7.83353031e-01 3.45806450e-01 1.59164950e-01 -3.65603298e-01
-8.57583940e-01 -1.13368571e-01 -1.26139045e+00 9.82954502e-01
8.70194733e-01 -1.83787018e-01 3.94524693e-01 1.08716500e+00
-1.83121219e-01 5.36317527e-01 -1.25476086e+00 7.41836369e-01
2.09512964e-01 -1.09423351e+00 1.63995549e-01 9.37447250e-02
8.05571318e-01 2.08392292e-01 4.30605769e-01 8.97366166e-01
5.15507042e-01 -6.52070880e-01 4.51313823e-01 5.70471764e-01
8.20698559e-01 -7.07665145e-01 8.32137465e-01 3.02284926e-01
-1.30068445e+00 -2.25275695e-01 -6.07271314e-01 2.42904544e-01
2.44451225e-01 6.54934764e-01 -3.04918110e-01 9.25301731e-01
7.48758733e-01 1.24642408e+00 -4.39477414e-01 8.69130075e-01
-7.90237561e-02 7.92528868e-01 -1.44812763e-01 6.43822253e-02
2.23551616e-01 -6.06939256e-01 2.38144249e-01 1.07742178e+00
5.26593029e-01 8.75120610e-02 3.09406996e-01 6.17066383e-01
-2.74058580e-01 4.02618974e-01 -5.10067225e-01 2.60854699e-02
7.27349877e-01 1.37820959e+00 -2.93184835e-02 -2.74946749e-01
-9.44369793e-01 8.65828156e-01 4.25366640e-01 4.95379508e-01
-8.67677569e-01 -2.69392401e-01 1.03159618e+00 -2.63507795e-02
6.08051598e-01 5.44397049e-02 -1.47280782e-01 -1.48912656e+00
-1.41098937e-02 -1.04966104e+00 5.69360495e-01 -4.53975201e-01
-2.20155263e+00 6.47463799e-01 -2.12545604e-01 -1.53588736e+00
-1.57433733e-01 -4.32307124e-01 -3.97536367e-01 1.03195620e+00
-1.39422655e+00 -1.30961657e+00 -1.56937866e-03 8.08775544e-01
5.09928882e-01 -4.38992381e-01 7.94303000e-01 6.65332317e-01
-6.54859126e-01 1.21904516e+00 5.94392955e-01 -1.91749446e-02
1.09309864e+00 -1.17301285e+00 4.20230299e-01 5.27525902e-01
4.27266777e-01 1.17297661e+00 1.90424368e-01 -7.05969512e-01
-1.73897791e+00 -1.16005301e+00 6.43332779e-01 -7.69048154e-01
8.18562090e-01 -5.88180482e-01 -1.30902469e+00 9.12993014e-01
4.32178192e-02 7.89168403e-02 1.23235464e+00 8.06783795e-01
-9.07600045e-01 -3.66702199e-01 -6.44379795e-01 4.28327650e-01
1.20007789e+00 -6.15305781e-01 -7.96922743e-01 2.71534890e-01
1.00495887e+00 2.32920870e-02 -1.07921422e+00 2.02073157e-01
6.83751285e-01 -4.09518123e-01 1.12442124e+00 -1.07459283e+00
3.61742973e-01 -1.52235150e-01 -2.41852939e-01 -1.75744760e+00
-7.64962494e-01 -4.45722342e-01 -1.86702415e-01 1.44676244e+00
7.69451737e-01 -7.27620482e-01 6.32409930e-01 5.52525997e-01
-4.67055023e-01 -7.99073160e-01 -5.23117900e-01 -9.29709136e-01
1.06637567e-01 -2.90560335e-01 1.13261199e+00 1.12382901e+00
-5.03199780e-03 7.46929765e-01 -7.56070018e-01 -3.33632678e-02
3.00920695e-01 5.28898299e-01 7.78732121e-01 -1.24516726e+00
-5.73052943e-01 -7.63383210e-02 2.02470914e-01 -1.38287258e+00
2.58983344e-01 -1.09868360e+00 -1.76897779e-01 -1.18101692e+00
3.41305286e-01 -1.13453603e+00 -1.03753948e+00 5.32669485e-01
-4.89917219e-01 -3.26142386e-02 4.63835569e-03 4.87752318e-01
-8.06509793e-01 6.62948191e-01 1.32148337e+00 5.35675474e-02
-2.57127434e-01 1.78741187e-01 -1.09659433e+00 4.56288546e-01
5.98928571e-01 -6.64041698e-01 -7.59267807e-01 -5.50000370e-01
5.13544202e-01 -2.51844347e-01 -2.41826281e-01 -4.16624755e-01
2.00305045e-01 -2.14414284e-01 3.30203891e-01 -5.38889408e-01
1.65330157e-01 -1.03972507e+00 2.73146480e-01 -2.65985519e-01
-5.59811711e-01 2.26697132e-01 -1.82268023e-01 1.06376505e+00
-1.10160798e-01 -1.93337843e-01 2.52531946e-01 1.20658204e-01
-5.72659433e-01 8.31190169e-01 4.99921553e-02 2.18333319e-01
7.96123445e-01 -1.71867758e-02 -5.04557908e-01 -6.78726658e-02
-6.03601933e-01 3.02700430e-01 3.64145875e-01 9.30474699e-01
5.50307631e-01 -1.63638973e+00 -5.85597992e-01 4.74024147e-01
5.15145242e-01 -3.85337889e-01 5.72913945e-01 7.60042608e-01
5.03827393e-01 2.99775153e-01 -1.63407475e-01 -2.55335182e-01
-8.73798132e-01 1.12426543e+00 -2.69283295e-01 -4.53653097e-01
-6.21792078e-01 8.52770925e-01 5.40580928e-01 -5.82207739e-01
1.12468325e-01 -1.25288010e-01 -2.81644374e-01 2.49085099e-01
5.18941998e-01 3.22629213e-02 5.84776662e-02 -4.65863526e-01
-2.19528362e-01 5.88017523e-01 -7.15712547e-01 2.85112470e-01
1.39868069e+00 -5.47557950e-01 3.12804312e-01 3.89653951e-01
1.46626806e+00 2.33446762e-01 -1.14549160e+00 -9.25907910e-01
-1.17263712e-01 -6.35742664e-01 -2.27210790e-01 -8.92666698e-01
-1.31900764e+00 7.00279653e-01 2.28753567e-01 4.59816828e-02
1.04859519e+00 7.93743953e-02 1.15706515e+00 2.63414711e-01
5.32806396e-01 -1.11095977e+00 1.46253645e-01 4.38861907e-01
6.12347782e-01 -1.17117560e+00 -1.21693499e-01 -2.67100960e-01
-1.00823843e+00 6.47863984e-01 7.94249356e-01 -7.60956034e-02
9.02816832e-01 1.29096091e-01 -7.41159692e-02 9.26040635e-02
-1.03202319e+00 -2.38613356e-02 6.50000453e-01 6.41113102e-01
2.44818807e-01 2.29551151e-01 -1.79429904e-01 1.62334180e+00
1.64404958e-01 -3.86084765e-02 1.63160697e-01 4.12377149e-01
-8.46069008e-02 -1.37184823e+00 2.26436421e-01 8.03925753e-01
-3.18778634e-01 -4.48276013e-01 3.47971879e-02 6.25960052e-01
1.70356259e-01 6.35785222e-01 -6.62941709e-02 -7.87919521e-01
4.67767477e-01 -3.70410900e-03 -1.07312746e-01 -8.07970107e-01
-7.44393885e-01 5.39732911e-02 4.28254120e-02 -4.05564696e-01
5.69553673e-03 -6.20873868e-01 -9.82914925e-01 -2.96571285e-01
-5.84655225e-01 4.71060991e-01 3.74930263e-01 9.31555569e-01
7.97403336e-01 2.93131202e-01 1.09405708e+00 -3.93616766e-01
-9.72192109e-01 -9.09546316e-01 -8.46155286e-01 7.77019799e-01
1.29891401e-02 -8.45976889e-01 -3.46664399e-01 -1.23841867e-01] | [10.161707878112793, 5.566991806030273] |
10b0c246-cdb0-4032-b5f3-65e10df3350b | self-supervised-3d-mesh-reconstruction-from | null | null | http://openaccess.thecvf.com//content/CVPR2021/html/Hu_Self-Supervised_3D_Mesh_Reconstruction_From_Single_Images_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Hu_Self-Supervised_3D_Mesh_Reconstruction_From_Single_Images_CVPR_2021_paper.pdf | Self-Supervised 3D Mesh Reconstruction From Single Images | Recent single-view 3D reconstruction methods reconstruct object's shape and texture from a single image with only 2D image-level annotation. However, without explicit 3D attribute-level supervision, it is still difficult to achieve satisfying reconstruction accuracy. In this paper, we propose a Self-supervised Mesh Reconstruction (SMR) approach to enhance 3D mesh attribute learning process. Our approach is motivated by observations that (1) 3D attributes from interpolation and prediction should be consistent, and (2) feature representation of landmarks from all images should be consistent. By only requiring silhouette mask annotation, our SMR can be trained in an end-to-end manner and generalizes to reconstruct natural objects of birds, cows, motorbikes, etc. Experiments demonstrate that our approach improves both 2D supervised and unsupervised 3D mesh reconstruction on multiple datasets. We also show that our model can be adapted to other image synthesis tasks, e.g., novel view generation, shape transfer, and texture transfer, with promising results. Our code is publicly available at https://github.com/Jia-Research-Lab. | ['Jiaya Jia', 'Shu Liu', 'Xiaogang Xu', 'LiWei Wang', 'Tao Hu'] | 2021-06-19 | null | null | null | cvpr-2021-1 | ['single-view-3d-reconstruction'] | ['computer-vision'] | [ 1.99213564e-01 2.47259423e-01 -1.93190277e-01 -6.13233507e-01
-8.86574268e-01 -4.25737530e-01 3.98401320e-01 -3.52524109e-02
3.11422080e-01 5.67758203e-01 1.09124467e-01 1.28684729e-01
2.10476011e-01 -8.06834698e-01 -1.22603118e+00 -3.85396063e-01
4.52160418e-01 8.66932273e-01 2.44431198e-01 -4.26994264e-02
1.17371298e-01 6.96956992e-01 -1.53868043e+00 3.27759922e-01
5.33196628e-01 1.08069193e+00 3.79272372e-01 4.25316215e-01
-1.28014922e-01 4.63038921e-01 2.42920339e-01 -1.97994277e-01
3.83576781e-01 -3.84719789e-01 -9.11866605e-01 5.93675435e-01
7.06968307e-01 -4.34591115e-01 -9.37566161e-03 8.72852385e-01
4.25823927e-01 -2.41037816e-01 8.10640156e-01 -9.84248221e-01
-7.61927962e-01 -2.97119413e-02 -8.01349103e-01 -7.01339126e-01
4.88923103e-01 -4.76000980e-02 7.83720493e-01 -1.22356284e+00
1.02023208e+00 1.24788618e+00 6.75353348e-01 7.33945429e-01
-1.54401326e+00 -4.85757262e-01 7.15189353e-02 -2.61348307e-01
-1.27706718e+00 -4.65185910e-01 1.03860712e+00 -5.48744977e-01
4.35937852e-01 1.25275850e-01 7.47751534e-01 9.23930109e-01
-2.01363787e-02 8.69798243e-01 1.28565514e+00 -3.86317134e-01
7.42976069e-02 1.81523070e-01 -5.17229736e-01 1.01092446e+00
-1.45627543e-01 3.04099955e-02 -5.56400061e-01 -6.64834529e-02
1.33719397e+00 5.61612584e-02 -1.19322009e-01 -1.07796860e+00
-1.43339550e+00 5.35895944e-01 1.78987771e-01 -1.92503005e-01
-2.59290725e-01 2.73476332e-01 1.78872809e-01 2.97635883e-01
9.92225826e-01 -1.67071596e-02 -6.76053464e-01 9.72923711e-02
-5.32365680e-01 2.26065248e-01 5.39921761e-01 1.37346768e+00
1.08752143e+00 -2.72078533e-02 3.02393854e-01 1.02752125e+00
3.32511514e-01 7.55267382e-01 -1.22220986e-01 -1.24899459e+00
9.66633856e-02 4.47031051e-01 1.04740068e-01 -8.12567055e-01
-1.52044311e-01 6.36080327e-03 -8.43595922e-01 4.38421369e-01
9.11029205e-02 2.64894962e-01 -9.50744510e-01 1.54307818e+00
7.39831626e-01 2.17519164e-01 -2.53323853e-01 9.07410800e-01
9.35741603e-01 2.89634407e-01 -3.56934696e-01 1.15654673e-02
1.21915913e+00 -8.58963609e-01 -3.53165656e-01 1.55126546e-02
3.45153213e-01 -9.68460619e-01 1.25162888e+00 2.78998464e-01
-1.36213422e+00 -5.33815563e-01 -6.82756484e-01 -3.64003509e-01
6.40938655e-02 9.33583546e-03 5.45279682e-01 1.69841051e-01
-8.76817167e-01 6.12804711e-01 -1.00776267e+00 -4.79257554e-01
7.27535963e-01 2.58890063e-01 -7.04039693e-01 -1.75638840e-01
-5.09819925e-01 6.91635311e-01 8.40707272e-02 -2.20253155e-01
-9.63520288e-01 -8.40820849e-01 -9.28535998e-01 -4.98174697e-01
2.61521369e-01 -1.22150493e+00 1.00991523e+00 -8.43260348e-01
-1.64089549e+00 1.48868072e+00 -3.29121530e-01 6.59053326e-02
4.68907714e-01 -1.32571012e-01 3.05530369e-01 1.34021163e-01
2.48244464e-01 8.42825234e-01 7.55452156e-01 -1.90624392e+00
-4.42231625e-01 -5.89428961e-01 -6.41943738e-02 3.17543715e-01
6.79682344e-02 -3.04112732e-01 -5.32937169e-01 -6.87156618e-01
5.06386459e-01 -9.69693601e-01 -2.91278213e-01 7.15346932e-01
-3.28379959e-01 -5.45168780e-02 7.12373912e-01 -6.02005482e-01
4.07823622e-01 -2.11550641e+00 3.77328664e-01 1.73471034e-01
1.00455388e-01 -3.01106840e-01 -3.45462710e-02 3.48462224e-01
1.15004949e-01 -2.27322225e-02 -5.14123201e-01 -6.48570597e-01
-6.22920878e-02 3.20174813e-01 -1.31133556e-01 5.93694091e-01
3.03580374e-01 9.60860014e-01 -8.43925834e-01 -6.52063370e-01
5.02667129e-01 5.45445085e-01 -8.24832261e-01 3.83257985e-01
-4.47258621e-01 1.16109288e+00 -5.71153164e-01 8.76378715e-01
7.80227125e-01 -4.75136191e-01 -1.05416916e-01 -4.00715113e-01
6.31005643e-03 7.35212639e-02 -1.12406576e+00 2.31625390e+00
-8.28908443e-01 2.58170575e-01 1.74930215e-01 -9.83687103e-01
9.98483539e-01 3.26110959e-01 6.95973933e-01 -3.66677910e-01
-1.37728319e-01 3.22806597e-01 -6.52515233e-01 -3.59269053e-01
7.29422569e-02 -2.93373913e-01 1.00034818e-01 5.29167652e-01
4.20415401e-02 -7.55471885e-01 -4.36435133e-01 2.28569675e-02
4.95307177e-01 7.78970718e-01 1.77150175e-01 -2.32708082e-01
3.39981973e-01 1.28744068e-02 4.72304165e-01 1.53799728e-01
3.56157243e-01 1.18777704e+00 1.98885307e-01 -5.32749414e-01
-1.47671020e+00 -1.22608876e+00 -4.09256846e-01 6.39706612e-01
3.57209712e-01 -1.79962441e-01 -5.55014729e-01 -7.06402779e-01
1.86428264e-01 3.01263630e-01 -6.54430926e-01 1.84226662e-01
-5.84772289e-01 -1.65262356e-01 8.44028220e-02 5.13513327e-01
3.05951566e-01 -8.38284731e-01 -3.37501794e-01 5.62455915e-02
-2.53690273e-01 -1.24669719e+00 -6.92218184e-01 -2.02294648e-01
-1.13111746e+00 -8.92072678e-01 -7.64814615e-01 -1.03008974e+00
1.15338969e+00 4.12332833e-01 1.18819225e+00 3.36595595e-01
-1.99210927e-01 6.39118493e-01 -3.02018344e-01 -2.50862181e-01
-4.82191354e-01 -2.02036113e-01 1.01135327e-02 1.05066359e-01
-2.22231701e-01 -1.04332733e+00 -7.81446159e-01 5.96036971e-01
-6.82176709e-01 5.44755876e-01 4.02838111e-01 9.67319369e-01
1.37652099e+00 -4.60446984e-01 4.36488241e-01 -1.18042922e+00
-1.33147299e-01 -1.60333410e-01 -4.75665599e-01 1.49630651e-01
-5.65085113e-01 -7.78245777e-02 4.35777068e-01 -3.06185246e-01
-1.09865391e+00 5.56685686e-01 -3.66657525e-01 -8.50706875e-01
-4.77703959e-01 8.47131833e-02 -3.21829855e-01 -1.08880676e-01
4.35679227e-01 3.40670556e-01 3.59976441e-01 -7.20947087e-01
5.17798543e-01 4.98644888e-01 3.81580710e-01 -8.06261122e-01
9.05688882e-01 9.49862778e-01 1.34858519e-01 -6.96759760e-01
-8.99915695e-01 -3.50901932e-01 -9.41250741e-01 -2.30422974e-01
8.61069441e-01 -1.18727994e+00 -5.77084720e-01 3.62439781e-01
-1.13780951e+00 -4.26829338e-01 -3.69444996e-01 4.45635498e-01
-1.18448138e+00 3.53277951e-01 -5.42652071e-01 -4.67742443e-01
-3.04615885e-01 -1.11164796e+00 1.68131840e+00 -1.86636627e-01
-1.56110689e-01 -9.30156171e-01 -1.51352689e-01 5.98193467e-01
9.35281888e-02 5.75701416e-01 9.41292644e-01 4.80493493e-02
-7.83972025e-01 1.48799375e-01 -1.27153486e-01 3.97809684e-01
3.33375841e-01 -1.36566550e-01 -8.31307709e-01 -1.69811428e-01
-1.75406843e-01 -6.45181239e-01 5.39208531e-01 3.36990237e-01
1.40318823e+00 -2.01587200e-01 -2.28794232e-01 8.21116507e-01
1.33395588e+00 -2.49214202e-01 5.11951566e-01 -8.11061040e-02
1.05878747e+00 8.73982430e-01 6.96829855e-01 3.80362153e-01
6.51833653e-01 8.84100378e-01 5.78775465e-01 -3.14483553e-01
-4.73989487e-01 -6.23386741e-01 1.09037176e-01 9.48599935e-01
-2.74740309e-01 1.74961254e-01 -7.98841476e-01 5.93113601e-01
-1.73431444e+00 -5.98176837e-01 -2.99550563e-01 2.26496935e+00
9.48248684e-01 -1.62164271e-01 1.38663752e-02 -2.21595064e-01
5.68834901e-01 -1.90117806e-01 -6.66474819e-01 -9.54204276e-02
4.81570773e-02 2.63145715e-01 2.79005438e-01 6.72546625e-01
-9.12746310e-01 1.08192301e+00 5.11774683e+00 8.78919423e-01
-9.28266525e-01 2.59418547e-01 7.41432250e-01 1.05282418e-01
-7.19695091e-01 1.81493089e-01 -4.72474992e-01 2.17772633e-01
1.38266236e-01 1.34457663e-01 3.48690122e-01 8.04868996e-01
3.74702029e-02 7.23292679e-02 -1.36340904e+00 1.11410749e+00
1.39486045e-01 -1.52635038e+00 1.55887261e-01 1.35722876e-01
1.00695145e+00 -2.12204158e-01 -7.71884993e-02 -2.25831628e-01
2.21520722e-01 -9.30907667e-01 8.31976235e-01 5.55218875e-01
1.21697950e+00 -6.69472814e-01 3.28522861e-01 4.25139844e-01
-1.12169409e+00 6.27028108e-01 -3.15241516e-01 3.25479209e-01
4.47849363e-01 6.18252456e-01 -5.97552121e-01 8.56128454e-01
7.05525875e-01 1.03103828e+00 -2.35770687e-01 5.37116885e-01
-1.66553169e-01 2.18171731e-01 -2.20998213e-01 3.95424992e-01
-2.36969709e-01 -2.92751491e-01 3.94958079e-01 7.64984369e-01
3.00235629e-01 2.72082239e-01 2.41054431e-01 9.20566320e-01
-6.94882795e-02 1.77708179e-01 -8.41578305e-01 3.58969897e-01
3.23426276e-01 1.11047769e+00 -7.95495391e-01 -1.72626078e-01
-3.68442208e-01 1.22206640e+00 4.52863425e-01 2.93995328e-02
-5.97728491e-01 1.67227969e-01 4.84516889e-01 5.44140875e-01
3.21151823e-01 -2.91559577e-01 -4.99552488e-01 -1.19610035e+00
1.15489893e-01 -5.34675479e-01 -1.70546025e-01 -9.55607712e-01
-1.53941417e+00 3.23685288e-01 -2.98530664e-02 -1.67492771e+00
2.57226615e-03 -3.50638986e-01 -1.66103780e-01 6.54038072e-01
-1.36467850e+00 -1.64771748e+00 -3.08326423e-01 5.71815252e-01
6.69727385e-01 6.41039666e-03 1.03908253e+00 2.96445310e-01
4.70991991e-02 3.40805948e-01 -1.33982554e-01 -1.95563883e-02
8.29984188e-01 -1.17495370e+00 4.64965820e-01 2.26182655e-01
1.57124445e-01 2.55431116e-01 3.61981094e-01 -6.68693304e-01
-1.65616179e+00 -1.29011786e+00 5.82045376e-01 -6.76811457e-01
1.22523241e-01 -4.91923273e-01 -8.36293101e-01 7.42928505e-01
-1.44802444e-02 3.66338760e-01 4.85025555e-01 -5.74313141e-02
-2.77703553e-01 -3.82774137e-02 -1.15746665e+00 5.34482598e-01
1.52618766e+00 -4.11792397e-01 -1.10180050e-01 5.54307044e-01
5.93182564e-01 -9.14323688e-01 -1.29720569e+00 6.52031362e-01
7.77826369e-01 -8.64422798e-01 1.15954912e+00 -5.25646806e-01
8.72233331e-01 -4.14586574e-01 -4.37866807e-01 -1.14567423e+00
-1.38238668e-01 -3.52429390e-01 -2.72584893e-02 1.13252842e+00
3.28990251e-01 -3.35825413e-01 7.72682369e-01 3.81833524e-01
-2.91520745e-01 -1.08600938e+00 -8.61043334e-01 -7.16629446e-01
9.91838127e-02 -1.89473331e-01 5.76682627e-01 9.97724950e-01
-5.65228760e-01 3.81310701e-01 -6.42461777e-01 1.99013710e-01
8.47127140e-01 7.07092166e-01 1.11208987e+00 -1.26948857e+00
-3.48049343e-01 -1.37736991e-01 -3.10307950e-01 -1.44149733e+00
2.52815455e-01 -1.18212497e+00 8.14061314e-02 -1.65007854e+00
3.14499825e-01 -8.41633141e-01 2.56553024e-01 3.69562864e-01
1.30778059e-01 6.06379330e-01 -5.53349927e-02 2.66336143e-01
-4.15153146e-01 8.56006265e-01 1.82166398e+00 8.28060433e-02
9.39187855e-02 2.05094069e-02 -4.78861213e-01 9.08182621e-01
7.66940892e-01 -4.57833171e-01 -4.07938629e-01 -6.50811672e-01
2.25525349e-02 3.40084195e-01 5.86075187e-01 -3.84057790e-01
-1.79164872e-01 -3.47629309e-01 5.13538301e-01 -6.51121974e-01
5.36605835e-01 -9.59439874e-01 4.79609191e-01 1.16394557e-01
-1.86567277e-01 -1.04251698e-01 9.58092883e-03 6.45060062e-01
7.71591440e-02 8.49094316e-02 8.78059030e-01 -4.36672300e-01
-2.73898005e-01 8.08918715e-01 2.16869697e-01 7.17239231e-02
1.01091003e+00 -3.41897547e-01 1.73403680e-01 -3.43583405e-01
-8.34964573e-01 1.83290303e-01 1.19236135e+00 2.95138657e-01
9.44377661e-01 -1.71416020e+00 -1.01723921e+00 2.81015933e-01
2.92155743e-01 6.90952301e-01 2.66660333e-01 8.55034888e-01
-5.68538368e-01 -7.93746039e-02 -1.94543347e-01 -1.11570227e+00
-1.20607471e+00 3.54619324e-01 1.42623752e-01 9.26467404e-02
-9.13508713e-01 6.76068366e-01 6.10964298e-01 -9.58118141e-01
7.73016959e-02 -6.47298172e-02 3.40084761e-01 -4.19242233e-01
9.06876847e-02 -3.02687436e-02 3.50613706e-02 -7.08840191e-01
-1.90846100e-01 1.26152611e+00 4.54462096e-02 -6.66963235e-02
1.52905679e+00 -1.87978640e-01 -1.96963489e-01 5.95726311e-01
1.29949605e+00 1.45463452e-01 -1.45638597e+00 -2.95083582e-01
-4.36036855e-01 -7.38308549e-01 -1.94913894e-01 -4.34365809e-01
-1.20966625e+00 8.94987345e-01 3.98419797e-01 -2.89757788e-01
1.03051519e+00 4.15947556e-01 8.63049924e-01 -4.04396132e-02
8.75248909e-01 -7.08727837e-01 1.95502132e-01 1.95501015e-01
1.24399185e+00 -1.45978022e+00 1.19360067e-01 -7.89404154e-01
-6.64582312e-01 9.32705820e-01 6.34704590e-01 -2.70090520e-01
8.66770983e-01 1.52568907e-01 -1.47258639e-01 -3.12903106e-01
-7.79342175e-01 -4.20366600e-02 3.29566866e-01 6.68003500e-01
5.13378799e-01 8.28082785e-02 -1.34234026e-01 2.00831771e-01
-6.59580529e-02 -4.74640653e-02 2.63220102e-01 8.29617858e-01
-1.27055377e-01 -1.28542078e+00 -1.21675812e-01 4.05831933e-01
-2.15777725e-01 6.06929548e-02 -2.18037665e-01 5.96425593e-01
9.13326740e-02 4.75123703e-01 1.71474945e-02 -3.08732271e-01
5.02968132e-01 -1.58062562e-01 9.05917346e-01 -9.52421308e-01
1.45627512e-02 3.05621713e-01 4.12608460e-02 -5.99754274e-01
-7.33447492e-01 -7.62776017e-01 -1.25032890e+00 -2.05432400e-01
-2.31709436e-01 -2.51954913e-01 4.77736443e-01 5.62994003e-01
5.92885196e-01 5.50065711e-02 9.64229345e-01 -1.03710377e+00
-1.57941446e-01 -6.65433824e-01 -5.09596825e-01 6.55386627e-01
2.45265648e-01 -8.57220113e-01 -3.20587829e-02 6.59209251e-01] | [8.546640396118164, -3.053508758544922] |
7346992c-1305-4009-82a9-15827bb731c8 | team-dadefrni-at-case-2021-task-1-document | null | null | https://aclanthology.org/2021.case-1.22 | https://aclanthology.org/2021.case-1.22.pdf | Team “DaDeFrNi” at CASE 2021 Task 1: Document and Sentence Classification for Protest Event Detection | This paper accompanies our top-performing submission to the CASE 2021 shared task, which is hosted at the workshop on Challenges and Applications of Automated Extraction of Socio-political Events from Text. Subtasks 1 and 2 of Task 1 concern the classification of newspaper articles and sentences into “conflict” versus “not conflict”-related in four different languages. Our model performs competitively in both subtasks (up to 0.8662 macro F1), obtaining the highest score of all contributions for subtask 1 on Hindi articles (0.7877 macro F1). We describe all experiments conducted with the XLM-RoBERTa (XLM-R) model and report results obtained in each binary classification task. We propose supplementing the original training data with additional data on political conflict events. In addition, we provide an analysis of unigram probability estimates and geospatial references contained within the original training corpus. | ['Niklas Stoehr', 'Dennis Atzenhofer', 'Daniel Vegh', 'Francesco Re'] | null | null | null | null | acl-case-2021-8 | ['sentence-classification'] | ['natural-language-processing'] | [ 2.14633822e-01 1.80027723e-01 -3.07488978e-01 -5.80906510e-01
-1.52303004e+00 -8.01822603e-01 1.52655888e+00 5.36480427e-01
-8.47500741e-01 1.11154032e+00 9.11501527e-01 -7.97835827e-01
-1.29052490e-01 -7.45705545e-01 -7.12777972e-01 -3.05889547e-01
-1.56070188e-01 5.54380000e-01 6.80416971e-02 -3.85256290e-01
5.53242266e-01 1.17835738e-01 -1.03896391e+00 7.53117681e-01
7.51926184e-01 6.04778647e-01 -9.19899866e-02 8.43418717e-01
-2.45148674e-01 1.09221959e+00 -8.38876605e-01 -7.71774471e-01
-1.07328124e-01 -2.83546329e-01 -1.21808767e+00 -7.08853722e-01
7.10345149e-01 2.17464715e-01 -3.22776318e-01 7.74823725e-01
3.59016091e-01 -2.19799981e-01 1.14502740e+00 -8.46398234e-01
-2.61617750e-01 1.36575365e+00 -8.78160417e-01 9.39348578e-01
4.88735676e-01 -4.26903337e-01 1.26560020e+00 -9.14466023e-01
9.73717391e-01 1.34884596e+00 7.73482800e-01 -1.89552590e-01
-1.22817218e+00 -7.14175701e-01 1.52734607e-01 2.90557772e-01
-1.16475546e+00 -5.33165812e-01 2.89597571e-01 -7.13302195e-01
1.20216143e+00 6.87058568e-01 3.55343223e-01 1.34279025e+00
3.64298284e-01 7.87797868e-01 1.62168467e+00 -4.99333650e-01
-1.99165300e-01 5.03918203e-03 4.60050702e-01 2.81257093e-01
2.05618814e-01 -2.49530867e-01 -5.95379293e-01 -5.39687812e-01
9.49798748e-02 -7.80710578e-01 1.33123156e-02 6.64687932e-01
-1.50632298e+00 9.07346904e-01 1.56237423e-01 5.75636029e-01
-3.80816877e-01 -7.88134933e-02 2.40688384e-01 3.34850192e-01
1.21728694e+00 3.96525770e-01 -4.73462075e-01 -2.64719129e-01
-1.03742635e+00 5.37039638e-01 1.06471634e+00 8.04398537e-01
2.90104330e-01 -5.95067799e-01 -4.04107362e-01 9.93796170e-01
-8.72270837e-02 8.37271810e-01 4.25531603e-02 -6.95591152e-01
1.20264339e+00 1.19502045e-01 2.00109273e-01 -1.33959377e+00
-7.65720129e-01 -4.55552697e-01 -6.18871987e-01 -4.51214850e-01
5.42755067e-01 -4.55341816e-01 -7.20939934e-01 1.65965044e+00
2.28003070e-01 -2.40206197e-01 4.38669957e-02 3.21403861e-01
1.25936091e+00 9.15657222e-01 4.40971285e-01 -4.44759130e-01
1.73251164e+00 -7.18632042e-01 -7.24429011e-01 -3.21527392e-01
5.75837135e-01 -1.14892113e+00 5.46932161e-01 4.45603170e-02
-9.15893853e-01 -2.67485023e-01 -8.21120560e-01 -6.92828819e-02
-5.75733185e-01 7.20541626e-02 3.76190156e-01 2.00777948e-01
-7.50443578e-01 2.80470103e-01 -3.21521431e-01 -4.77463186e-01
6.22939765e-02 -1.71229020e-01 -3.15439969e-01 3.01829159e-01
-1.58936429e+00 1.31419349e+00 1.83588341e-02 -1.86520144e-01
-3.27809602e-01 -5.90382993e-01 -6.79965913e-01 -2.60217160e-01
2.65088499e-01 -9.53827947e-02 1.28653574e+00 -4.22826797e-01
-7.08019197e-01 1.44978094e+00 -6.68652207e-02 -5.73178232e-01
7.23638475e-01 -3.31824780e-01 -7.50924647e-01 -1.32187176e-02
7.95215666e-01 1.53646410e-01 2.38737226e-01 -9.69265819e-01
-1.31570578e+00 -2.78305352e-01 -8.36209431e-02 1.14985995e-01
-9.31371301e-02 7.84916759e-01 -8.83450434e-02 -1.04149270e+00
-1.80644542e-01 -8.05837452e-01 -1.36577234e-01 -1.04926264e+00
-5.75597227e-01 -6.21083319e-01 1.32779196e-01 -1.27603340e+00
1.46929741e+00 -1.76041555e+00 7.87492022e-02 1.67465836e-01
1.37409270e-01 -2.40111649e-01 -6.40717447e-02 5.99952936e-01
2.68467446e-03 5.19750297e-01 -1.97277576e-01 -9.35147032e-02
7.54143149e-02 -2.16452301e-01 -6.66530550e-01 4.46067154e-01
6.73215613e-02 7.59346485e-01 -9.94279265e-01 -7.78639793e-01
-2.16889158e-01 -1.05624542e-01 8.31775367e-02 -4.60081965e-01
1.29815936e-01 2.59758085e-01 -3.16992164e-01 3.30333441e-01
5.59100389e-01 -1.24035306e-01 4.13339645e-01 -2.14656040e-01
-7.54108489e-01 8.34085643e-01 -6.59246981e-01 1.02292418e+00
-5.44226348e-01 1.04385030e+00 2.50736594e-01 -8.29453826e-01
6.95000529e-01 1.15509532e-01 4.73462284e-01 -8.23390663e-01
-7.50444010e-02 2.78569579e-01 -6.09094724e-02 -3.99325907e-01
7.21676290e-01 8.58966038e-02 -9.43355620e-01 5.42426467e-01
-1.60219088e-01 -2.56163746e-01 6.44843340e-01 4.05239433e-01
1.18776596e+00 -1.44899100e-01 6.56148851e-01 -7.74727285e-01
3.64230305e-01 1.82503298e-01 4.38188910e-01 1.11742091e+00
-1.02345668e-01 4.37894255e-01 6.46788001e-01 -6.29447997e-01
-9.79781210e-01 -9.61714745e-01 -6.25960648e-01 1.21315932e+00
-2.02454463e-01 -6.90423667e-01 -4.36780065e-01 -9.01707292e-01
-1.68500781e-01 1.09654129e+00 -7.86253273e-01 6.85727060e-01
-9.92996514e-01 -1.39615715e+00 8.30425262e-01 1.59090891e-01
3.80683422e-01 -9.17664528e-01 -4.50091094e-01 2.35067576e-01
-9.32730675e-01 -1.29837120e+00 -8.87246877e-02 2.59912491e-01
-2.93402910e-01 -1.21105587e+00 -6.99776649e-01 -7.78248072e-01
-6.81906417e-02 -1.23811461e-01 1.37272966e+00 -4.79930073e-01
-1.92965686e-01 -6.58343211e-02 -4.71744537e-01 -6.50564969e-01
-5.15994430e-01 5.73480964e-01 -3.16669047e-01 -3.88408333e-01
4.81469423e-01 -1.70918301e-01 5.94622120e-02 -1.27106994e-01
-2.70064503e-01 4.40026224e-01 2.33408943e-01 6.90466702e-01
1.07087746e-01 -2.25642934e-01 4.89754945e-01 -9.70956206e-01
6.01466477e-01 -8.37425411e-01 -5.87709546e-01 3.63506913e-01
-7.84242302e-02 -1.78970456e-01 2.68547475e-01 6.61841184e-02
-1.22563875e+00 -3.97156417e-01 -1.42224997e-01 9.45341885e-01
-7.71434084e-02 9.63267207e-01 4.75286186e-01 5.68977118e-01
7.25365639e-01 -2.39405558e-01 -5.18996894e-01 -5.42770863e-01
2.25822523e-01 1.04108572e+00 6.85851574e-01 -6.72722816e-01
6.49953902e-01 3.48045021e-01 -2.78076530e-01 -1.09889269e+00
-1.34526110e+00 -3.97008330e-01 -4.47851062e-01 -2.07679763e-01
1.05169272e+00 -1.01622617e+00 -1.75116241e-01 4.79884773e-01
-1.31529295e+00 -3.61771673e-01 1.21888913e-01 6.67886555e-01
-2.72319734e-01 9.22677666e-02 -6.62556410e-01 -8.43249202e-01
-1.65204749e-01 -6.50481224e-01 1.01338208e+00 -2.01592341e-01
-6.29959226e-01 -8.59550714e-01 3.87872905e-01 5.07905126e-01
2.31635615e-01 7.05875158e-01 1.10275483e+00 -8.20021629e-01
2.73272365e-01 -9.11603589e-03 -2.40825370e-01 -4.06505257e-01
2.85297055e-02 2.47520670e-01 -7.44903743e-01 7.83581957e-02
-1.45363778e-01 -3.30826312e-01 1.38637769e+00 5.40982068e-01
6.88852847e-01 -5.01876414e-01 -5.95891416e-01 1.73813209e-01
9.66398597e-01 -1.83445439e-02 3.75308096e-01 8.91377389e-01
2.60301054e-01 1.05935884e+00 5.65745652e-01 4.27342683e-01
8.27954054e-01 7.69035816e-01 -1.54884800e-01 -1.48782119e-01
-1.91877022e-01 -1.13699548e-01 3.05621564e-01 7.94746816e-01
-2.09813625e-01 -1.96176067e-01 -1.50007665e+00 8.54045272e-01
-1.84122193e+00 -1.34498191e+00 -8.63182664e-01 1.87480986e+00
1.17096543e+00 3.64461541e-01 1.54402018e-01 -2.58573517e-02
9.58381355e-01 5.32090366e-01 2.52144277e-01 -3.96892905e-01
-5.10237396e-01 2.48185173e-01 6.53973997e-01 9.60235834e-01
-1.81620729e+00 9.81698692e-01 7.13690901e+00 9.88532186e-01
-6.56790495e-01 2.30150014e-01 8.35660875e-01 -1.21004350e-01
-2.80838430e-01 -4.21782285e-02 -9.45161045e-01 5.69046080e-01
1.32060301e+00 -4.23345208e-01 -6.45300150e-02 3.12582821e-01
6.75965846e-02 -5.06969571e-01 -6.62445843e-01 4.50584173e-01
-2.13080142e-02 -1.55733621e+00 -1.88171044e-01 9.30509493e-02
7.87835002e-01 6.40664637e-01 -2.59935915e-01 4.87988651e-01
7.02044547e-01 -9.25852537e-01 1.07106030e+00 3.02558511e-01
7.44036138e-01 -5.69475710e-01 8.71502876e-01 4.32353318e-01
-1.03964603e+00 9.24532562e-02 5.96537739e-02 -2.44847953e-01
4.82575923e-01 7.53525853e-01 -5.09829521e-01 7.15289831e-01
9.40578520e-01 6.63618386e-01 -6.71329856e-01 6.48917735e-01
-4.40761030e-01 1.00192118e+00 -4.64528590e-01 -2.44410470e-01
3.66146296e-01 2.39859194e-01 8.46218705e-01 1.88718069e+00
6.33898154e-02 3.21568638e-01 9.99915153e-02 1.59578249e-01
-1.12900786e-01 4.09553379e-01 -5.69060445e-01 -4.90710093e-03
6.45972908e-01 1.11149609e+00 -7.95713127e-01 -5.98236322e-01
-4.34051961e-01 4.58584726e-01 4.36434001e-01 1.67728946e-01
-9.01043832e-01 -5.18849015e-01 9.82104167e-02 8.94236378e-03
2.41322443e-02 -1.61022127e-01 -4.78050321e-01 -1.09543073e+00
-2.07699671e-01 -8.14211607e-01 7.27398992e-01 -3.80746484e-01
-1.58347940e+00 7.80844510e-01 5.70494413e-01 -5.95337749e-01
-3.51178139e-01 -3.39098424e-01 -3.64948690e-01 9.80657935e-01
-1.20858848e+00 -1.20470512e+00 5.97953275e-02 2.34865658e-02
5.18081605e-01 -3.83855075e-01 8.15872371e-01 2.45585009e-01
-3.01782668e-01 2.89006084e-01 2.46587515e-01 3.35570425e-01
1.04064989e+00 -1.22643864e+00 8.52598310e-01 5.89356422e-01
1.89380035e-01 3.24337065e-01 8.31228852e-01 -9.08934057e-01
-4.86975700e-01 -8.64658654e-01 2.03067327e+00 -7.15125740e-01
1.25618231e+00 -4.57107127e-01 -2.62219131e-01 5.36980212e-01
4.85090792e-01 -8.87465775e-01 7.40635753e-01 6.57154977e-01
-4.32011396e-01 2.71622866e-01 -8.61762166e-01 5.93106508e-01
8.96080852e-01 -5.07835805e-01 -1.00312805e+00 8.32907319e-01
4.98706073e-01 -3.23062092e-01 -9.06929433e-01 3.91033709e-01
5.44153154e-01 -4.50562388e-01 7.93056488e-01 -7.96870291e-01
1.00216746e+00 8.70243609e-02 -3.74207675e-01 -1.10621500e+00
-5.43321013e-01 -4.22731280e-01 3.64245474e-01 1.24115634e+00
9.60616350e-01 -4.51621234e-01 1.75637871e-01 1.00860998e-01
1.70907572e-01 -3.22813481e-01 -1.15424871e+00 -5.90240240e-01
4.91070330e-01 -4.65734571e-01 1.63246810e-01 1.40269327e+00
5.44042177e-02 7.84799278e-01 -3.34615052e-01 -2.57032476e-02
6.61819518e-01 2.90082723e-01 6.47991002e-01 -1.25756705e+00
5.23165055e-03 -6.49452269e-01 8.56730714e-02 -4.43008721e-01
3.44294697e-01 -1.01922596e+00 -4.65921797e-02 -1.87074697e+00
5.07141054e-01 -3.22930932e-01 4.83044125e-02 4.64122862e-01
-3.63794208e-01 4.33921844e-01 2.01049879e-01 5.69775820e-01
-3.86094302e-01 -1.51120927e-02 4.71446514e-01 -3.78686577e-01
1.35134086e-01 -1.11765854e-01 -7.09559679e-01 8.06104958e-01
7.38137603e-01 -6.65791869e-01 3.60768974e-01 -6.07696593e-01
4.24691290e-01 -3.59726958e-02 2.50898659e-01 -7.54756987e-01
4.42051701e-02 -4.61747795e-01 2.99780697e-01 -9.69262958e-01
1.83040872e-02 -7.83279315e-02 -3.50417532e-02 4.70474154e-01
-7.33473122e-01 2.76881546e-01 1.56336322e-01 2.97979951e-01
-2.48970404e-01 -8.91062990e-02 4.08672780e-01 -8.41073319e-02
-3.50856930e-01 -3.05527538e-01 -8.92094851e-01 5.39766967e-01
8.28226447e-01 3.93872440e-01 -8.38232160e-01 -3.99824172e-01
-7.82438695e-01 3.25249702e-01 -1.03699185e-01 4.93020415e-01
-1.71859950e-01 -1.08230066e+00 -1.68029261e+00 -4.56551075e-01
-6.82891905e-02 -6.35457277e-01 -2.30109575e-03 9.28297639e-01
-5.83724320e-01 6.28956914e-01 4.47220989e-02 -9.08757299e-02
-1.37103093e+00 2.58949064e-02 -2.22308040e-01 -7.17526555e-01
-3.14033598e-01 7.59677470e-01 1.86703764e-02 -5.17907262e-01
-2.71068402e-02 9.56592932e-02 -5.76167762e-01 6.70204580e-01
7.14051843e-01 5.62532246e-01 -4.37994115e-02 -9.61254954e-01
-6.61040664e-01 3.35267216e-01 -2.33961597e-01 -7.65283167e-01
1.69508052e+00 3.26292366e-02 -3.88095438e-01 6.43428802e-01
1.09019125e+00 5.84891737e-01 -3.03587735e-01 -2.54240036e-01
6.24553382e-01 9.35815051e-02 8.41684714e-02 -1.19085538e+00
-2.67715424e-01 3.94162536e-01 -9.37516466e-02 6.11908674e-01
3.97394419e-01 2.59995192e-01 4.26907569e-01 3.31327170e-01
3.29554945e-01 -1.48169935e+00 -4.98979688e-01 1.08123958e+00
1.25979984e+00 -1.15158856e+00 4.71540809e-01 -5.04512787e-01
-4.98029590e-01 9.48154271e-01 1.54203013e-01 1.30309090e-01
7.22497344e-01 3.97896945e-01 7.22417608e-02 -3.68021429e-01
-7.33339906e-01 -1.61402877e-02 4.12317693e-01 2.45675087e-01
9.60953414e-01 3.35625321e-01 -1.20231497e+00 6.11583650e-01
-6.47244990e-01 -5.78448296e-01 3.52085143e-01 9.37535048e-01
-3.80396813e-01 -8.83174181e-01 -4.62846845e-01 6.32932067e-01
-9.04121399e-01 -5.18839955e-01 -1.00800312e+00 9.88200426e-01
-1.17250055e-01 1.18153322e+00 3.22022796e-01 -3.60930473e-01
2.12338492e-01 -1.26049414e-01 3.24643642e-01 -4.25020754e-01
-9.82753754e-01 9.16928202e-02 1.16819465e+00 -1.19596139e-01
-5.99475384e-01 -9.85814273e-01 -7.30472624e-01 -5.23539901e-01
3.79773267e-02 4.69497770e-01 7.43961573e-01 9.12939489e-01
9.71801206e-03 3.54375809e-01 4.90338475e-01 -6.30658329e-01
-4.83250767e-01 -1.35743701e+00 -2.18858108e-01 2.09700093e-01
1.49829179e-01 -3.66182446e-01 -4.44586575e-01 -6.16227016e-02] | [9.011157035827637, 9.809412002563477] |
4aa4d49d-7fb7-4c45-8131-15b13cea894a | sketch-based-medical-image-retrieval | 2303.03633 | null | https://arxiv.org/abs/2303.03633v1 | https://arxiv.org/pdf/2303.03633v1.pdf | Sketch-based Medical Image Retrieval | The amount of medical images stored in hospitals is increasing faster than ever; however, utilizing the accumulated medical images has been limited. This is because existing content-based medical image retrieval (CBMIR) systems usually require example images to construct query vectors; nevertheless, example images cannot always be prepared. Besides, there can be images with rare characteristics that make it difficult to find similar example images, which we call isolated samples. Here, we introduce a novel sketch-based medical image retrieval (SBMIR) system that enables users to find images of interest without example images. The key idea lies in feature decomposition of medical images, whereby the entire feature of a medical image can be decomposed into and reconstructed from normal and abnormal features. By extending this idea, our SBMIR system provides an easy-to-use two-step graphical user interface: users first select a template image to specify a normal feature and then draw a semantic sketch of the disease on the template image to represent an abnormal feature. Subsequently, it integrates the two kinds of input to construct a query vector and retrieves reference images with the closest reference vectors. Using two datasets, ten healthcare professionals with various clinical backgrounds participated in the user test for evaluation. As a result, our SBMIR system enabled users to overcome previous challenges, including image retrieval based on fine-grained image characteristics, image retrieval without example images, and image retrieval for isolated samples. Our SBMIR system achieves flexible medical image retrieval on demand, thereby expanding the utility of medical image databases. | ['Ryuji Hamamoto', 'Tatsuya Harada', 'Yusuke Kurose', 'Amina Bolatkan', 'Nobuji Kouno', 'Satoshi Nakamura', 'Yukihiro Yoshida', 'Yasuyuki Takamizawa', 'Masamichi Takahashi', 'Hirokazu Watanabe', 'Mototaka Miyake', 'Takaaki Mizuno', 'Ryuichiro Hataya', 'Lin Gu', 'Kazuma Kobayashi'] | 2023-03-07 | null | null | null | null | ['medical-image-retrieval', 'medical-image-retrieval'] | ['computer-vision', 'medical'] | [ 2.54673541e-01 -4.04708624e-01 -2.64620423e-01 -2.08412826e-01
-7.93987989e-01 -4.73520607e-01 3.38267386e-01 3.78522128e-01
-3.37252975e-01 8.14577863e-02 -4.12212349e-02 -2.54147440e-01
-4.42347556e-01 -9.09568250e-01 1.58647809e-03 -7.64608920e-01
3.61076087e-01 4.13310349e-01 3.47782195e-01 -1.43180668e-01
2.92622298e-01 7.70644307e-01 -1.64142692e+00 4.73171383e-01
7.63359368e-01 9.98838544e-01 7.12417662e-01 4.41473931e-01
-5.44954479e-01 3.16052914e-01 -9.78322864e-01 -1.04244962e-01
1.43954381e-01 -4.41304624e-01 -5.70167065e-01 3.41370642e-01
2.77949553e-02 -7.03181982e-01 -4.78829980e-01 9.30764854e-01
7.75162280e-01 -1.31067842e-01 6.78477764e-01 -1.24187553e+00
-8.92790616e-01 -2.11989321e-02 -4.40508902e-01 4.90066744e-02
4.80320990e-01 1.34134572e-02 4.69435900e-01 -9.47242379e-01
8.65603209e-01 1.17426813e+00 1.49956256e-01 6.70089364e-01
-7.34366894e-01 -7.12426841e-01 -1.48463011e-01 2.79771566e-01
-1.75679564e+00 -1.24014221e-01 7.39618897e-01 -3.55459124e-01
3.04610252e-01 7.40443826e-01 7.79049397e-01 3.76097262e-01
1.84466898e-01 1.08632779e+00 5.44793129e-01 -2.38177538e-01
-2.34529283e-02 2.96447664e-01 2.33838037e-02 8.02877784e-01
1.63876683e-01 -3.17843974e-01 -7.53799826e-02 -4.17410374e-01
1.11337149e+00 1.05601263e+00 -5.20123482e-01 -3.27298224e-01
-1.49375105e+00 5.38827479e-01 5.25294423e-01 6.82186902e-01
-3.46517622e-01 -3.15773487e-01 4.47515100e-01 1.99487567e-01
-1.46629155e-01 1.60891190e-01 -3.42056490e-02 3.29528064e-01
-8.38318706e-01 -7.93321803e-02 4.48317915e-01 9.97774303e-01
6.20495439e-01 -5.51959991e-01 -4.11913723e-01 1.11664188e+00
1.41727567e-01 8.00880253e-01 9.58666801e-01 -6.75646365e-01
8.92023742e-02 9.68411803e-01 -2.66576204e-02 -1.49789357e+00
-1.33341908e-01 8.85820296e-03 -1.15164280e+00 -1.50989369e-01
9.29889604e-02 5.86656749e-01 -9.71903086e-01 1.12504768e+00
1.51667222e-01 -1.92347869e-01 9.43479910e-02 9.83529866e-01
1.28756845e+00 5.84828377e-01 -3.60224210e-02 -2.71137148e-01
1.75329101e+00 -8.06374252e-01 -8.51853251e-01 2.20445797e-01
5.71417093e-01 -8.56563985e-01 1.31337178e+00 3.25067639e-01
-8.40815365e-01 -4.73583758e-01 -1.00110412e+00 3.70447226e-02
-5.63769341e-01 5.21722555e-01 4.12680268e-01 3.23071659e-01
-8.17295074e-01 3.33124459e-01 -6.17411017e-01 -3.23415697e-01
3.54168385e-01 2.21883208e-01 -5.85791051e-01 -4.69407529e-01
-9.76414859e-01 4.98509824e-01 1.91659868e-01 -8.59757364e-02
-4.21099663e-01 -7.69402862e-01 -7.47384071e-01 -4.74638976e-02
2.07563475e-01 -7.26857185e-01 8.28523159e-01 -7.07598209e-01
-9.20520246e-01 1.09000814e+00 -1.34440914e-01 3.26418608e-01
5.11861145e-01 1.51626632e-01 -7.53363073e-01 8.06028843e-01
2.38666475e-01 6.27947688e-01 7.50478506e-01 -1.24568534e+00
-4.61047322e-01 -2.79820830e-01 -1.39849320e-01 1.42513260e-01
-4.32183683e-01 -1.33586928e-01 -1.20393658e+00 -7.81047046e-01
3.38268310e-01 -5.81972897e-01 -2.18201041e-01 6.08465493e-01
-3.31859440e-01 -1.28321856e-01 1.00534713e+00 -4.80369955e-01
1.62097704e+00 -2.48794222e+00 -3.58868599e-01 5.46391726e-01
3.78969967e-01 4.04935718e-01 -2.74702758e-01 5.33149004e-01
7.22163022e-02 2.54988611e-01 -3.95390671e-03 1.39867201e-01
-4.35382396e-01 1.96837068e-01 -3.51742841e-02 1.17576286e-01
-8.04444402e-02 9.32219326e-01 -9.84638691e-01 -1.20224833e+00
3.85300636e-01 5.59335053e-01 -2.16900364e-01 2.65956968e-01
3.94586205e-01 2.53843337e-01 -8.37378383e-01 9.53306377e-01
7.53968060e-01 -5.56444347e-01 5.39066456e-02 -5.09677410e-01
2.61669040e-01 -5.23207366e-01 -1.08715010e+00 1.64931500e+00
-2.89253622e-01 2.41122380e-01 -1.31112710e-01 -7.03853846e-01
9.18668628e-01 4.88605857e-01 7.70772576e-01 -7.56819844e-01
-4.14857566e-02 2.92071164e-01 -4.20795202e-01 -8.88976455e-01
2.07306191e-01 1.63866803e-02 1.12421632e-01 5.41212380e-01
-4.27720487e-01 -1.01810679e-01 1.97769791e-01 4.25392568e-01
8.97989929e-01 -4.27165836e-01 3.90928745e-01 9.81745869e-03
8.18934798e-01 6.53721616e-02 2.80969024e-01 7.78106809e-01
-6.04217015e-02 9.33791637e-01 5.65248765e-02 -6.68906033e-01
-8.52449834e-01 -1.19534254e+00 -2.96904713e-01 6.15797162e-01
6.59935415e-01 -5.89540184e-01 -4.16978329e-01 -5.83811581e-01
-3.73280421e-02 -6.82573253e-03 -4.27021295e-01 -6.29361868e-02
-3.82177949e-01 -3.85124922e-01 2.77006328e-01 2.07839876e-01
4.35359150e-01 -1.30836427e+00 -7.88067818e-01 1.82480574e-01
-2.54587024e-01 -4.78459179e-01 -1.02019656e+00 -7.87424445e-01
-8.01638544e-01 -1.36871600e+00 -1.33574140e+00 -1.02660573e+00
1.31840253e+00 7.92129934e-01 9.37261283e-01 9.29317892e-01
-1.08651900e+00 5.85840106e-01 -3.42275828e-01 -2.99142629e-01
-3.74379754e-01 -2.80144215e-01 -1.58952653e-01 -3.21644805e-02
2.68076748e-01 -6.34960756e-02 -1.28054583e+00 5.27544260e-01
-1.65477967e+00 2.44788349e-01 8.34564030e-01 1.02099311e+00
9.06890571e-01 -1.42167564e-02 5.00804245e-01 -7.41461754e-01
6.99050248e-01 -3.49400043e-01 -1.93652898e-01 8.81465018e-01
-5.54221153e-01 -1.04889624e-01 4.80792850e-01 -6.27485096e-01
-5.37694037e-01 6.49322709e-03 -7.96511918e-02 -6.22873843e-01
1.11464914e-02 5.45503795e-01 -5.62412031e-02 1.13866650e-01
3.94201905e-01 6.28477097e-01 5.25583982e-01 -4.15022582e-01
8.46206471e-02 1.03296089e+00 5.58785737e-01 -1.83341935e-01
5.18319309e-01 4.47507948e-01 -2.32905731e-01 -7.58134425e-01
-3.47277224e-01 -8.20055068e-01 -4.73035514e-01 -2.33381927e-01
6.67145252e-01 -6.74729049e-01 -7.10044682e-01 2.45516330e-01
-8.94737124e-01 2.74778187e-01 -4.89886180e-02 3.33482713e-01
-1.20561659e-01 6.81735218e-01 -5.41010082e-01 -5.00289261e-01
-6.77121580e-01 -1.37238109e+00 1.24524307e+00 3.06990683e-01
3.03760800e-03 -7.01531112e-01 -4.25797254e-01 9.21955034e-02
5.17885566e-01 6.28480911e-02 1.11140120e+00 -5.40237606e-01
-6.05354071e-01 -6.38675749e-01 -3.22665006e-01 -9.14052576e-02
7.14062691e-01 7.39010200e-02 -4.23350930e-01 -3.98675054e-01
-1.42648607e-01 5.75461313e-02 5.72580457e-01 1.96995154e-01
1.53026748e+00 -3.43246609e-01 -8.66229653e-01 3.75746936e-01
1.39684236e+00 6.31858826e-01 7.97678351e-01 1.67329133e-01
4.65353787e-01 4.87439811e-01 7.67142475e-01 3.58560383e-01
1.32988974e-01 5.08033097e-01 -2.19494812e-02 -5.18921852e-01
7.03466162e-02 -1.91902950e-01 -3.29094350e-01 1.01584017e+00
1.59922212e-01 5.57147302e-02 -7.53340900e-01 5.02893806e-01
-1.62996173e+00 -8.22722495e-01 3.55830610e-01 2.25514197e+00
8.77731681e-01 -4.45180863e-01 -2.05850884e-01 1.43569723e-01
8.55555296e-01 -1.80853710e-01 -5.05369186e-01 2.99682766e-01
1.67626128e-01 -1.70488674e-02 3.66377831e-02 -1.37771117e-02
-8.95609617e-01 3.79389286e-01 6.06379128e+00 1.26700902e+00
-1.43000114e+00 -1.37880564e-01 6.93204880e-01 9.44714099e-02
-5.31918883e-01 -5.17406285e-01 -4.56145495e-01 5.08259237e-01
9.51738432e-02 -4.41143274e-01 4.07164991e-02 9.29422438e-01
-1.16174305e-02 -3.08153536e-02 -9.41513836e-01 1.55563247e+00
1.75300807e-01 -1.54288638e+00 7.82364845e-01 -7.17436820e-02
3.44384134e-01 -5.65997064e-01 4.22789939e-02 2.18199845e-02
-3.53952199e-01 -9.98657107e-01 1.51512921e-01 7.49469519e-01
1.36873782e+00 -5.67349136e-01 7.12276280e-01 3.87240767e-01
-1.26766217e+00 1.23137444e-01 -4.97583121e-01 6.79170966e-01
-1.34707615e-01 3.42563123e-01 -8.82932782e-01 7.04926908e-01
5.45947313e-01 4.51873988e-01 -6.92710221e-01 1.25234127e+00
3.01030159e-01 -1.31901354e-01 -2.03149959e-01 -9.67054628e-03
2.97758561e-02 -1.44536838e-01 1.52459294e-01 1.21873212e+00
6.80806696e-01 5.46151459e-01 3.83835793e-01 6.04831457e-01
9.78498086e-02 6.67250991e-01 -7.48074710e-01 2.33935602e-02
6.23262405e-01 1.41638839e+00 -1.09356225e+00 -6.75642073e-01
-3.48428428e-01 1.13642442e+00 -4.34387505e-01 3.61058325e-01
-3.65842760e-01 -8.29233229e-01 2.40021855e-01 1.92680225e-01
4.54318570e-03 2.63524532e-01 2.25483313e-01 -1.11372578e+00
-1.04949638e-01 -9.87380743e-01 6.64990485e-01 -9.06984210e-01
-1.19795358e+00 7.79162109e-01 -5.42805158e-02 -1.85517299e+00
-2.43701994e-01 -3.16908270e-01 -4.14734215e-01 8.58271956e-01
-9.90385175e-01 -1.04508221e+00 -7.12444782e-01 9.27646458e-01
5.04963219e-01 -2.03571528e-01 1.12809610e+00 5.01182079e-01
-3.02045316e-01 6.42847180e-01 1.56011507e-01 3.66465062e-01
8.04289758e-01 -7.88666189e-01 -1.73759639e-01 1.95942923e-01
-3.41250598e-02 1.01402223e+00 6.31181300e-02 -5.72034001e-01
-1.43849814e+00 -8.13253582e-01 5.61887145e-01 -1.76812455e-01
1.27719611e-01 1.86324462e-01 -9.41234827e-01 1.66587174e-01
-1.34858817e-01 2.95850575e-01 8.75235260e-01 -8.05398345e-01
-1.87622949e-01 -2.32294261e-01 -1.36340880e+00 9.14255857e-01
6.85702324e-01 -5.43498099e-01 -4.48820323e-01 3.20869565e-01
5.55207789e-01 -3.11993688e-01 -9.93843913e-01 4.56638753e-01
8.13879967e-01 -5.73034048e-01 1.30492783e+00 -2.68400222e-01
2.18987763e-01 -5.85699916e-01 6.12331666e-02 -8.47752929e-01
-1.50582537e-01 -1.74400613e-01 4.94804949e-01 9.00155365e-01
1.69265270e-01 -6.36169851e-01 4.17705327e-01 6.65442884e-01
2.04675332e-01 -1.03912055e+00 -5.41121364e-01 -5.09042680e-01
-4.23212528e-01 -2.66279001e-02 8.83227885e-01 8.97672534e-01
5.12125865e-02 -2.32742205e-01 -2.03166101e-02 -1.24238625e-01
3.30925345e-01 4.35733140e-01 6.16607845e-01 -1.17218959e+00
-5.58707751e-02 -5.99906504e-01 -5.45963705e-01 -9.96886015e-01
-5.04644513e-01 -9.20532525e-01 -1.85923845e-01 -1.85088170e+00
5.32633066e-01 -7.71128535e-01 -4.99838203e-01 5.09587884e-01
-2.63545215e-01 6.46441698e-01 1.93366274e-01 7.84407079e-01
-5.65987885e-01 -5.80745675e-02 1.69541287e+00 -4.19380754e-01
-1.36287391e-01 -2.03109775e-02 -6.77915990e-01 4.58656937e-01
6.00722075e-01 -2.74007976e-01 -5.57425320e-01 -1.91746414e-01
-4.95564416e-02 4.74793792e-01 2.40501329e-01 -7.63171315e-01
3.04930925e-01 -2.04772100e-01 6.44940495e-01 -9.94418204e-01
1.85425237e-01 -9.25927401e-01 4.74118471e-01 8.15059960e-01
-8.65465179e-02 2.38872394e-01 5.72492406e-02 3.11694652e-01
-5.63464940e-01 -3.94954920e-01 3.81846666e-01 -5.08271217e-01
-6.90356493e-01 5.94754517e-01 -2.75424808e-01 -4.10838157e-01
1.02714396e+00 -4.32395697e-01 -2.79495895e-01 -4.10594106e-01
-6.57302082e-01 2.25845441e-01 5.92191100e-01 3.95914048e-01
1.25153959e+00 -1.64121902e+00 -5.37003040e-01 4.54746455e-01
6.07735217e-01 2.78962636e-03 6.50795817e-01 7.81493247e-01
-1.01847696e+00 3.49195719e-01 -9.91694108e-02 -7.81620204e-01
-1.71871650e+00 1.06799173e+00 5.91066964e-02 -1.23145163e-01
-1.03261161e+00 2.98503846e-01 6.71207607e-01 -2.62940060e-02
1.95424989e-01 -1.00625284e-01 -2.92282522e-01 1.68902338e-01
1.08007848e+00 -1.01017199e-01 9.79300439e-02 -5.04189372e-01
-4.27838624e-01 8.75154078e-01 -2.50200570e-01 5.65618388e-02
9.88658726e-01 -1.18146323e-01 -1.56357095e-01 2.23717168e-01
1.56584883e+00 5.05441092e-02 -6.13127649e-01 -3.90176147e-01
-3.17045987e-01 -7.49783278e-01 -4.11013126e-01 -7.08060801e-01
-1.30503953e+00 7.52660334e-01 7.71861672e-01 1.37997165e-01
1.41788268e+00 1.62136316e-01 9.43269134e-01 4.42260653e-01
5.14172256e-01 -7.14996517e-01 2.41067797e-01 -2.06834346e-01
1.02705467e+00 -1.17211330e+00 4.45672758e-02 -4.47562188e-01
-6.97117448e-01 1.25732827e+00 1.91121742e-01 1.34684265e-01
7.43512630e-01 4.06699851e-02 5.49875319e-01 -4.64968830e-01
-3.94179076e-01 7.43417144e-02 5.39992034e-01 4.80905980e-01
1.91798553e-01 -5.92831755e-04 -5.42407513e-01 5.92972219e-01
2.07350776e-01 1.32832363e-01 1.05871469e-01 1.03414309e+00
-3.68246257e-01 -1.20569777e+00 -6.03302598e-01 6.55752540e-01
-6.06196404e-01 -1.96456954e-01 -4.96920124e-02 7.36084819e-01
-6.95459470e-02 6.89765573e-01 1.52950346e-01 -3.01143169e-01
4.06096995e-01 -1.65984213e-01 2.38147676e-01 -5.05904794e-01
-3.49397480e-01 1.55323803e-01 -6.70270205e-01 -4.65898156e-01
-1.68146104e-01 -6.18235283e-02 -1.38171446e+00 1.41533893e-02
-2.57670850e-01 3.05388331e-01 6.49682283e-01 5.28441668e-01
4.67659682e-01 2.98605889e-01 5.82809210e-01 -6.24363065e-01
-1.17341243e-01 -6.84739351e-01 -5.46502531e-01 8.43277156e-01
3.17000896e-01 -4.51902688e-01 -2.32829954e-02 2.61905044e-01] | [14.374837875366211, -1.5511887073516846] |
8be67383-72d6-49b4-b051-1e1c18867891 | fp-age-leveraging-face-parsing-attention-for | 2106.11145 | null | https://arxiv.org/abs/2106.11145v2 | https://arxiv.org/pdf/2106.11145v2.pdf | FP-Age: Leveraging Face Parsing Attention for Facial Age Estimation in the Wild | Image-based age estimation aims to predict a person's age from facial images. It is used in a variety of real-world applications. Although end-to-end deep models have achieved impressive results for age estimation on benchmark datasets, their performance in-the-wild still leaves much room for improvement due to the challenges caused by large variations in head pose, facial expressions, and occlusions. To address this issue, we propose a simple yet effective method to explicitly incorporate facial semantics into age estimation, so that the model would learn to correctly focus on the most informative facial components from unaligned facial images regardless of head pose and non-rigid deformation. To this end, we design a face parsing-based network to learn semantic information at different scales and a novel face parsing attention module to leverage these semantic features for age estimation. To evaluate our method on in-the-wild data, we also introduce a new challenging large-scale benchmark called IMDB-Clean. This dataset is created by semi-automatically cleaning the noisy IMDB-WIKI dataset using a constrained clustering method. Through comprehensive experiment on IMDB-Clean and other benchmark datasets, under both intra-dataset and cross-dataset evaluation protocols, we show that our method consistently outperforms all existing age estimation methods and achieves a new state-of-the-art performance. To the best of our knowledge, our work presents the first attempt of leveraging face parsing attention to achieve semantic-aware age estimation, which may be inspiring to other high level facial analysis tasks. Code and data are available on \url{https://github.com/ibug-group/fpage}. | ['Maja Pantic', 'Yujiang Wang', 'Jie Shen', 'Yiming Lin'] | 2021-06-21 | null | null | null | null | ['age-estimation', 'face-parsing', 'age-estimation'] | ['computer-vision', 'computer-vision', 'miscellaneous'] | [-1.84243336e-01 3.41185361e-01 -7.99914822e-02 -1.14058614e+00
-7.65235484e-01 -1.61407813e-01 4.00802970e-01 -3.06413829e-01
-3.18233162e-01 2.46217921e-01 3.56042594e-01 5.00420749e-01
3.02195936e-01 -4.85825032e-01 -5.75494349e-01 -6.04925394e-01
-1.02071822e-01 5.80086768e-01 -3.59383821e-01 2.37809405e-01
-1.18455991e-01 2.67244637e-01 -1.82437706e+00 -6.12122065e-04
6.41913176e-01 1.29836011e+00 -2.76675135e-01 3.69666517e-01
5.51538393e-02 4.65087265e-01 -3.31562728e-01 -1.01887250e+00
1.89709976e-01 -2.94247828e-02 -8.28509569e-01 1.25874743e-01
1.01144540e+00 -7.37799227e-01 -2.66538143e-01 9.63796735e-01
1.00408399e+00 -1.70454279e-01 4.52880085e-01 -1.65152466e+00
-5.99954426e-01 3.58166814e-01 -8.32483768e-01 -2.43988618e-01
3.38693291e-01 -1.36815496e-02 6.48452878e-01 -8.46527457e-01
4.61832970e-01 1.68830359e+00 7.31145084e-01 1.23262227e+00
-7.29176462e-01 -1.08772409e+00 4.04185504e-01 2.12176085e-01
-1.24031937e+00 -9.52983797e-01 8.07494104e-01 -2.56428689e-01
4.83952731e-01 -4.01518382e-02 2.81101167e-01 1.38771975e+00
-3.63468409e-01 9.36672211e-01 1.15209985e+00 -1.67055994e-01
6.69510439e-02 -2.88084984e-01 2.33315349e-01 1.06098819e+00
-4.72886860e-02 -9.88549441e-02 -8.55619013e-01 -2.29399458e-01
2.76040167e-01 -2.28880811e-02 -9.98178311e-03 -1.89035505e-01
-6.58875108e-01 6.33818209e-01 3.20443749e-01 -1.79735765e-01
-1.86046779e-01 3.09154660e-01 3.95771056e-01 -7.15173036e-02
7.81246305e-01 -1.38844758e-01 -8.02383244e-01 -2.01929942e-01
-7.97373772e-01 3.78346771e-01 5.99046171e-01 7.18080401e-01
5.97736478e-01 -1.19493857e-01 3.80010009e-02 1.15205288e+00
4.95197445e-01 5.67737937e-01 3.80600870e-01 -1.24755454e+00
2.53757715e-01 4.65783924e-01 -4.36101556e-01 -5.21140456e-01
-6.25669599e-01 2.58589699e-03 -4.90379155e-01 1.82953045e-01
6.76763475e-01 -1.12811357e-01 -1.36252224e+00 2.50045109e+00
7.39702821e-01 1.69682756e-01 -2.62171268e-01 8.30897391e-01
9.98686850e-01 1.15136899e-01 5.40117860e-01 -1.29314691e-01
1.76563835e+00 -9.07468975e-01 -5.67887008e-01 -4.00014699e-01
5.34219027e-01 -4.99964476e-01 1.02040911e+00 3.09034824e-01
-1.20577157e+00 -3.79082948e-01 -5.51303148e-01 -2.57897973e-01
-3.14232141e-01 2.27433696e-01 8.04791152e-01 8.50168109e-01
-1.13299036e+00 4.39165652e-01 -1.13959789e+00 -6.96506560e-01
9.69466507e-01 5.79042673e-01 -8.18123698e-01 -2.76039660e-01
-7.79793441e-01 5.19186139e-01 -1.37135208e-01 1.64581820e-01
-7.54541814e-01 -8.77990425e-01 -1.05461991e+00 -2.05194280e-01
3.31157565e-01 -1.02014363e+00 1.35380924e+00 -1.13935447e+00
-1.28228939e+00 1.33751357e+00 -4.79433954e-01 -4.11402471e-02
3.15711498e-01 -5.47740221e-01 -2.33501166e-01 1.30971447e-01
1.19409382e-01 1.10725141e+00 9.91399884e-01 -1.01596665e+00
-2.07169279e-01 -1.36181808e+00 -2.37275437e-01 7.23518059e-02
-7.06797421e-01 5.77731192e-01 -9.08612490e-01 -5.97462773e-01
-1.30708143e-01 -1.04470181e+00 2.25807682e-01 4.16896850e-01
-1.76187024e-01 -4.76533800e-01 6.65646076e-01 -1.11826909e+00
1.07832813e+00 -1.94021845e+00 1.85956344e-01 -1.33252382e-01
3.18271607e-01 1.51311025e-01 -2.52393723e-01 -1.84139416e-01
-1.59714788e-01 -4.58843168e-03 -1.55006960e-01 -1.16515684e+00
7.69331306e-02 3.37451436e-02 2.23471805e-01 4.61316079e-01
1.72491238e-01 7.36599743e-01 -4.76048678e-01 -6.43054664e-01
-1.76876843e-01 7.69534111e-01 -6.67754054e-01 4.74946976e-01
-4.26775068e-02 4.73000169e-01 -3.76804084e-01 1.19981778e+00
9.09149349e-01 1.56493768e-01 -2.99427267e-02 -3.57468277e-01
5.58240831e-01 -1.56171635e-01 -8.04708242e-01 1.97245359e+00
-3.51430774e-01 -5.98378703e-02 4.39781517e-01 -6.80461109e-01
7.20903039e-01 7.63761401e-02 5.51592827e-01 -6.52161419e-01
3.16466749e-01 -6.83317631e-02 -3.59090626e-01 -4.49393600e-01
5.47153652e-02 2.16758214e-02 -8.82102326e-02 2.80682772e-01
3.80394787e-01 3.10037553e-01 8.03290009e-02 1.84477955e-01
9.39684868e-01 2.93621153e-01 -8.14547464e-02 -1.62108727e-02
4.30731684e-01 -8.57414544e-01 9.70068038e-01 6.11989237e-02
-6.78239346e-01 8.03726852e-01 5.33907473e-01 -5.52026212e-01
-6.72963977e-01 -1.14761662e+00 -9.68217254e-02 1.65823424e+00
-4.11583573e-01 -6.95141315e-01 -1.33153903e+00 -9.88309085e-01
1.34917915e-01 1.68167368e-01 -1.01496387e+00 -8.87845457e-02
-3.69102329e-01 -9.40321863e-01 5.94007194e-01 6.51651502e-01
4.79717374e-01 -8.57513070e-01 6.29279986e-02 -2.98593730e-01
-1.44535542e-01 -1.39160979e+00 -6.70158267e-01 -4.63255495e-01
-6.71045065e-01 -1.11008680e+00 -8.07229280e-01 -6.34759963e-01
8.99622083e-01 -2.81629205e-01 1.16967714e+00 2.39757374e-01
-3.17551434e-01 7.41057098e-01 -2.04728663e-01 -7.36049533e-01
-1.80299184e-03 1.79794878e-01 2.53616750e-01 1.35379910e-01
7.41710126e-01 -7.89350331e-01 -8.88958454e-01 2.10374892e-01
-5.26368260e-01 -1.06184691e-01 2.74146438e-01 4.81668770e-01
4.66322094e-01 -4.58255082e-01 6.41201735e-01 -7.64153302e-01
9.72917899e-02 -4.04230058e-01 -3.07790250e-01 1.81432322e-01
-6.24299645e-01 -1.29454676e-02 1.30414546e-01 -1.56125292e-01
-1.11817968e+00 2.66785771e-01 -5.00301361e-01 -4.31892902e-01
-3.62469107e-01 9.53888893e-03 -7.99420774e-01 1.50790945e-01
2.07440257e-01 -2.45133683e-01 4.53322947e-01 -7.35639870e-01
2.22871095e-01 7.49893546e-01 6.69770539e-01 -8.58274102e-01
7.20091045e-01 6.29571199e-01 -7.40857646e-02 -4.86318737e-01
-1.12520242e+00 -1.34988859e-01 -5.84972739e-01 -2.11144760e-01
9.57177043e-01 -1.39736295e+00 -8.30456614e-01 1.24836934e+00
-8.95708442e-01 -4.25556153e-01 3.54181439e-01 1.02568902e-01
-4.97045398e-01 4.37494665e-01 -7.59734571e-01 -7.39163756e-01
-9.12416875e-01 -1.05217814e+00 1.60615766e+00 3.16702366e-01
-3.34759325e-01 -7.22055316e-01 -9.28394496e-02 9.87992406e-01
1.03834547e-01 3.39323699e-01 4.90411878e-01 -4.85450596e-01
-1.16967194e-01 -1.31753281e-01 -1.84287801e-01 4.25105542e-01
6.40605614e-02 6.77497163e-02 -1.35438704e+00 -3.16303819e-01
-4.08499449e-01 -7.31108248e-01 9.31049883e-01 3.76657218e-01
1.73363900e+00 -2.46176898e-01 -1.79738507e-01 9.53712881e-01
1.00002694e+00 -4.36317682e-01 7.46979952e-01 4.48861904e-03
1.02597880e+00 9.16978836e-01 5.30406594e-01 6.22100830e-01
9.67066824e-01 6.97187245e-01 4.73290533e-01 -3.36227119e-02
-3.00850004e-01 -2.68836409e-01 4.01617825e-01 4.72144723e-01
-1.35350361e-01 5.00056408e-02 -8.39147627e-01 4.71112043e-01
-1.57235646e+00 -6.84713125e-01 3.99728894e-01 2.14473057e+00
9.52002883e-01 -1.64234191e-01 6.48745358e-01 -3.07080477e-01
6.45162702e-01 2.12230772e-01 -7.11231232e-01 -3.72619569e-01
8.75269547e-02 3.88883650e-01 1.33121282e-01 2.03616172e-01
-1.15740490e+00 1.00882363e+00 5.03491926e+00 5.39204180e-01
-1.10224521e+00 2.74478257e-01 1.19025350e+00 -4.59029078e-01
2.33216546e-02 -4.66860533e-01 -9.99366879e-01 6.97619140e-01
1.03056014e+00 2.22616300e-01 4.63360757e-01 1.07886136e+00
2.38519739e-02 4.18212824e-02 -1.16548705e+00 1.19963145e+00
4.38951790e-01 -6.28381312e-01 -1.87003821e-01 1.13589205e-01
3.54679704e-01 -1.30008608e-02 3.27972263e-01 3.08679402e-01
6.68721199e-02 -1.11259091e+00 7.25898981e-01 4.02773082e-01
9.99831140e-01 -8.54902029e-01 5.24127722e-01 -1.47608042e-01
-1.11698294e+00 -1.50885820e-01 -1.66947275e-01 3.06735355e-02
7.17063621e-02 6.12200141e-01 -3.60129893e-01 1.12217732e-01
1.15793931e+00 5.64238429e-01 -8.79236877e-01 5.37661076e-01
-1.46174192e-01 4.34737027e-01 -3.06249619e-01 5.27783453e-01
-4.49428409e-01 1.06216721e-01 -1.41760977e-02 8.15428615e-01
4.48197842e-01 1.12307072e-01 -1.91168427e-01 4.57838625e-01
-5.01983166e-01 1.45917177e-01 -2.43931085e-01 4.00486998e-02
4.32110906e-01 1.54175293e+00 -3.39496106e-01 -6.51909336e-02
-5.37769377e-01 1.13795626e+00 7.56937921e-01 7.18906596e-02
-8.76631498e-01 1.23795912e-01 1.36860216e+00 2.80676395e-01
1.25347570e-01 -7.64351934e-02 -8.36593956e-02 -1.09107876e+00
1.78758100e-01 -9.21604693e-01 5.74729800e-01 -7.96164691e-01
-1.46664369e+00 4.43659186e-01 -7.99456146e-03 -4.73911673e-01
-3.13731879e-01 -6.94201529e-01 -6.33504629e-01 5.83397567e-01
-1.37830377e+00 -1.84896791e+00 -5.88323355e-01 6.86484993e-01
2.96623975e-01 -1.49797693e-01 8.60857606e-01 4.10693467e-01
-9.61901784e-01 1.31720006e+00 -4.11313832e-01 4.40290004e-01
1.31910431e+00 -1.18812716e+00 4.56993788e-01 7.44117081e-01
-1.48013666e-01 5.49247503e-01 5.65959394e-01 -5.52959621e-01
-1.39030850e+00 -1.14399838e+00 8.41943324e-01 -7.82625020e-01
2.63254404e-01 -7.46056080e-01 -8.18039060e-01 8.00283730e-01
-6.40258342e-02 3.14790815e-01 8.07983041e-01 5.29140830e-01
-9.09180224e-01 -4.31665450e-01 -1.35215068e+00 4.66277272e-01
1.56854880e+00 -3.71675581e-01 -1.43542737e-01 2.18812525e-01
7.44744301e-01 -4.38439399e-01 -1.18531382e+00 6.91485822e-01
8.72745156e-01 -1.01174235e+00 1.07935929e+00 -5.64581692e-01
4.89088118e-01 1.13013521e-01 -1.26853706e-02 -1.08009017e+00
3.23742740e-02 -6.37470245e-01 -4.11321193e-01 1.88994932e+00
1.52762122e-02 -5.38994670e-01 1.16450632e+00 1.20836794e+00
2.84244888e-03 -1.00513184e+00 -7.80034065e-01 -4.56889778e-01
2.59486645e-01 -4.24821764e-01 9.21110094e-01 6.90733790e-01
-4.06569481e-01 1.42880976e-01 -3.99976552e-01 2.46528625e-01
1.00892615e+00 -2.21700177e-01 1.00233662e+00 -1.24682367e+00
4.34572175e-02 -2.34923571e-01 -5.58625579e-01 -3.58992845e-01
8.82627964e-01 -6.65726125e-01 -2.43108585e-01 -1.17627573e+00
4.93179232e-01 -3.05378169e-01 -1.54149309e-01 7.95189142e-01
-5.49216449e-01 4.20979351e-01 1.36443987e-01 -2.52335399e-01
-6.26540303e-01 6.87295198e-01 9.05218542e-01 6.16094470e-02
2.37667426e-01 -1.61365625e-02 -1.05758727e+00 7.93377936e-01
9.18739974e-01 -3.09348464e-01 -3.76609504e-01 -5.44478953e-01
-7.65910149e-02 -2.98240423e-01 2.26792008e-01 -8.48174572e-01
-5.23077063e-02 1.55529724e-02 6.18296921e-01 -2.62740105e-01
6.45452201e-01 -4.54082370e-01 -8.01418945e-02 3.74590121e-02
-5.41475639e-02 1.60724565e-01 1.07307896e-01 3.71465087e-01
-1.30295893e-03 2.39315867e-01 8.75881374e-01 1.03805761e-03
-7.95301735e-01 1.03268766e+00 3.18110168e-01 1.44779727e-01
9.12363112e-01 -3.35880816e-02 -4.37743962e-01 -4.73434359e-01
-7.19759047e-01 3.81480157e-01 7.87482440e-01 6.31266296e-01
3.96156639e-01 -1.39500380e+00 -8.08111966e-01 2.63616264e-01
4.34974134e-01 -6.97018653e-02 4.83316272e-01 7.21749961e-01
-8.64869431e-02 -2.33772844e-01 -2.31979772e-01 -4.27659065e-01
-1.80528986e+00 5.61586082e-01 3.47345173e-01 1.99383777e-03
-8.13258663e-02 1.20019341e+00 4.00424719e-01 -7.09434032e-01
5.61455786e-01 1.00761823e-01 1.05440123e-02 1.19945258e-01
8.85758460e-01 2.16797039e-01 8.58805478e-02 -9.15259182e-01
-5.25382340e-01 8.27069521e-01 -2.11571977e-01 2.07769603e-01
1.46075726e+00 -2.76067287e-01 -2.53965735e-01 -1.43533066e-01
1.48420715e+00 -7.27241635e-02 -1.35148704e+00 -1.85504645e-01
-2.27034032e-01 -4.76691484e-01 -9.81727093e-02 -8.25688124e-01
-1.71182775e+00 7.71871507e-01 8.94915581e-01 -7.51751781e-01
1.45194387e+00 3.58516425e-01 1.04663396e+00 -3.88474911e-02
3.21892917e-01 -1.26897764e+00 2.54095137e-01 1.30671754e-01
8.20585907e-01 -1.57630694e+00 2.65747197e-02 -5.14081538e-01
-6.42808378e-01 8.19434166e-01 1.07948756e+00 3.32506120e-01
7.71075726e-01 2.43738472e-01 2.32796341e-01 -1.00530401e-01
-6.83147609e-01 -3.05559337e-01 1.73597574e-01 8.70218873e-01
5.77686012e-01 1.89293977e-02 -6.30601943e-02 1.00467706e+00
-4.04276341e-01 1.55989274e-01 4.97971810e-02 6.57859504e-01
-1.40551120e-01 -1.43003786e+00 -4.02359635e-01 4.51694548e-01
-8.44745994e-01 1.54441580e-01 -4.40933228e-01 5.67871571e-01
1.71871096e-01 7.41736531e-01 1.57788515e-01 -2.96220988e-01
1.18610367e-01 3.98618579e-01 8.03741992e-01 -4.52274173e-01
-3.01260620e-01 -2.09862620e-01 2.88895071e-01 -9.93938744e-01
-4.07047927e-01 -9.55506980e-01 -1.20622182e+00 -5.76461256e-01
2.51384556e-01 -3.50907147e-01 8.72920811e-01 6.65377557e-01
6.09805465e-01 1.83081850e-02 5.09524763e-01 -9.13813114e-01
-3.93243641e-01 -9.80602026e-01 -5.20073712e-01 8.46564949e-01
-3.01449336e-02 -8.28035057e-01 -2.92893112e-01 9.03020948e-02] | [13.48222541809082, 0.8208431005477905] |
4350acc5-ddd9-4445-a6f6-f307f8e4b2c1 | sophia-a-scalable-stochastic-second-order | 2305.14342 | null | https://arxiv.org/abs/2305.14342v1 | https://arxiv.org/pdf/2305.14342v1.pdf | Sophia: A Scalable Stochastic Second-order Optimizer for Language Model Pre-training | Given the massive cost of language model pre-training, a non-trivial improvement of the optimization algorithm would lead to a material reduction on the time and cost of training. Adam and its variants have been state-of-the-art for years, and more sophisticated second-order (Hessian-based) optimizers often incur too much per-step overhead. In this paper, we propose Sophia, Second-order Clipped Stochastic Optimization, a simple scalable second-order optimizer that uses a light-weight estimate of the diagonal Hessian as the pre-conditioner. The update is the moving average of the gradients divided by the moving average of the estimated Hessian, followed by element-wise clipping. The clipping controls the worst-case update size and tames the negative impact of non-convexity and rapid change of Hessian along the trajectory. Sophia only estimates the diagonal Hessian every handful of iterations, which has negligible average per-step time and memory overhead. On language modeling with GPT-2 models of sizes ranging from 125M to 770M, Sophia achieves a 2x speed-up compared with Adam in the number of steps, total compute, and wall-clock time. Theoretically, we show that Sophia adapts to the curvature in different components of the parameters, which can be highly heterogeneous for language modeling tasks. Our run-time bound does not depend on the condition number of the loss. | ['Tengyu Ma', 'Percy Liang', 'David Hall', 'Zhiyuan Li', 'Hong Liu'] | 2023-05-23 | null | null | null | null | ['stochastic-optimization'] | ['methodology'] | [-2.93434292e-01 -8.39237198e-02 -1.20146021e-01 -2.99291790e-01
-1.03796160e+00 -5.00682354e-01 3.08268577e-01 2.80965120e-01
-8.91837716e-01 3.22606951e-01 -2.20637918e-01 -7.30207324e-01
1.63883969e-01 -4.19995755e-01 -9.76194918e-01 -5.40527880e-01
-3.27629298e-01 7.06011415e-01 2.33004361e-01 -1.27103493e-01
2.31022492e-01 2.73570150e-01 -8.41584384e-01 -5.29757142e-02
8.14472198e-01 1.12124145e+00 1.49348363e-01 9.75188792e-01
-7.44394660e-02 5.78423142e-01 -2.59094059e-01 -4.36526179e-01
3.60464752e-01 -1.49335817e-01 -5.66513717e-01 4.49185483e-02
4.72622991e-01 -5.03478527e-01 -3.66616935e-01 1.00469398e+00
5.59829056e-01 2.58164823e-01 3.65555406e-01 -1.17970157e+00
-7.60925412e-02 6.22004926e-01 -7.47681737e-01 1.51548028e-01
-2.50186592e-01 1.80970758e-01 9.67481732e-01 -1.07162666e+00
3.57296914e-01 1.34304476e+00 8.63194942e-01 3.60896945e-01
-1.07806528e+00 -4.83281195e-01 4.64815557e-01 -9.84426066e-02
-1.42129171e+00 -3.43036085e-01 3.60799372e-01 -3.49969476e-01
1.20809174e+00 2.01377496e-01 3.94973934e-01 2.77814835e-01
1.83610216e-01 8.46820772e-01 6.06842101e-01 -2.44609222e-01
2.27378055e-01 2.97108680e-01 5.83255291e-02 1.07646549e+00
5.53476587e-02 -1.80792794e-01 -4.41072762e-01 -5.05180359e-01
6.31584942e-01 -1.13983735e-01 -7.99086988e-02 -1.63043991e-01
-9.12464857e-01 7.79130459e-01 2.40081370e-01 5.98757379e-02
-1.85026586e-01 4.82568771e-01 4.41737413e-01 3.77542228e-01
6.72484219e-01 3.82377863e-01 -8.37466836e-01 -7.02419400e-01
-1.28391623e+00 2.92612225e-01 1.02298081e+00 8.56427133e-01
6.47880495e-01 1.01281807e-01 1.16764881e-01 7.20369339e-01
2.27057934e-01 5.40908277e-01 5.00004947e-01 -7.72747397e-01
8.50417972e-01 4.69382763e-01 1.35228589e-01 -9.67398822e-01
-5.69605291e-01 -7.13635802e-01 -6.25444293e-01 -1.01418227e-01
6.69321477e-01 -5.39944887e-01 -8.04700375e-01 1.66129911e+00
5.58740497e-01 3.88911329e-02 -2.62091637e-01 7.92002916e-01
1.90832987e-02 7.81498671e-01 -2.83665091e-01 -5.71530424e-02
1.18658674e+00 -1.38295197e+00 -7.04734325e-02 -4.56199020e-01
1.08436573e+00 -7.91090071e-01 1.37565017e+00 2.26182818e-01
-1.46250880e+00 -3.48053500e-02 -1.07154083e+00 -3.39147836e-01
-5.37766740e-02 3.20974320e-01 4.65300977e-01 5.17272532e-01
-9.26912725e-01 9.74621892e-01 -1.36555529e+00 2.49089465e-01
6.71213716e-02 5.85882127e-01 -1.08810619e-01 2.02174798e-01
-8.01048100e-01 6.82220697e-01 7.17521310e-02 1.63973078e-01
-5.34720242e-01 -1.45331800e+00 -6.12325013e-01 3.59181851e-01
4.08012927e-01 -4.54526186e-01 1.25250912e+00 -6.92012131e-01
-1.76554251e+00 6.78613663e-01 -5.22208571e-01 -8.27811897e-01
1.08220947e+00 -4.60703462e-01 1.86057210e-01 -2.04703152e-01
-2.06958771e-01 4.08208638e-01 7.69931495e-01 -6.61892474e-01
-4.38797683e-01 -4.26507622e-01 -1.33516148e-01 3.53270918e-01
-5.32074749e-01 -6.33802637e-02 -9.65458512e-01 -5.18510103e-01
1.11188851e-01 -1.23148870e+00 -4.83639538e-01 8.90220329e-02
-1.85485125e-01 1.78865552e-01 4.94949460e-01 -7.68270433e-01
1.49594808e+00 -2.42073488e+00 -1.52684048e-01 1.23838283e-01
1.09740473e-01 2.98843026e-01 3.18921655e-02 2.91788191e-01
3.98698658e-01 1.98133752e-01 -5.37932992e-01 -9.05992389e-01
1.66180715e-01 9.50456187e-02 -3.83768827e-01 6.11949444e-01
-4.96635772e-03 5.64829707e-01 -8.01828980e-01 -2.82313734e-01
-1.86456576e-01 6.18429005e-01 -8.60581100e-01 -3.85123491e-02
-2.23197863e-01 6.58465326e-02 -2.60865301e-01 2.81268507e-02
7.68993020e-01 -4.66359049e-01 1.78104296e-01 1.29785314e-01
-4.44079973e-02 7.23773479e-01 -1.36354351e+00 1.30364931e+00
-9.48201358e-01 6.77955925e-01 3.29352409e-01 -6.12060010e-01
5.15929639e-01 -2.11496148e-02 2.32890114e-01 -3.49179178e-01
-1.97886787e-02 5.34024179e-01 -1.24327242e-01 5.85423037e-02
4.43706751e-01 7.76241794e-02 2.63859868e-01 4.64292020e-01
-3.88307452e-01 -2.45614886e-01 4.21076685e-01 1.94584876e-01
8.83182526e-01 -2.33336985e-01 -5.31660113e-03 -3.73320282e-01
4.24987853e-01 -1.51841164e-01 4.73742872e-01 6.07898891e-01
1.02737702e-01 4.23338711e-01 7.56422818e-01 -4.32652950e-01
-1.15179396e+00 -6.89619422e-01 3.58360335e-02 1.31460583e+00
-2.40809843e-01 -6.10711038e-01 -8.57901394e-01 -5.12494981e-01
1.70191899e-01 6.32021725e-01 -3.12620521e-01 -8.08253959e-02
-8.91032696e-01 -1.04487109e+00 5.29805362e-01 5.35730720e-01
4.42309409e-01 -3.68697375e-01 -5.65032363e-01 3.38352740e-01
3.34081411e-01 -1.15472722e+00 -1.19054294e+00 1.62809223e-01
-1.09200168e+00 -6.11338317e-01 -7.52686262e-01 -6.70147061e-01
9.72123563e-01 1.31507188e-01 1.18455648e+00 2.29739174e-01
1.84165630e-02 -1.84021927e-02 3.06996465e-01 -2.73080409e-01
-2.15544581e-01 3.49635750e-01 -5.88713400e-02 -6.12268262e-02
-2.23871291e-01 -5.79792321e-01 -5.87519825e-01 1.88444391e-01
-5.76206505e-01 9.58292335e-02 2.65505016e-01 9.94956613e-01
7.13672161e-01 -7.93240294e-02 -1.80605426e-01 -8.01216364e-01
5.40794194e-01 -2.31016859e-01 -1.16699266e+00 8.41563419e-02
-7.70241380e-01 4.41735208e-01 9.21123862e-01 -7.40856171e-01
-6.34169877e-01 9.82750878e-02 -3.28160636e-02 -4.35899824e-01
8.06776583e-01 5.73027432e-01 2.68528074e-01 -1.66910738e-01
2.55295575e-01 2.75421917e-01 -1.02460437e-01 -6.33726656e-01
3.79648417e-01 1.90896213e-01 3.86653125e-01 -5.56637883e-01
6.44541323e-01 4.87014443e-01 1.99612051e-01 -8.86245191e-01
-7.95811415e-01 -3.32561702e-01 -2.12321803e-01 3.77460301e-01
4.26333934e-01 -9.09652412e-01 -8.96320105e-01 5.02673268e-01
-9.60689366e-01 -7.39236295e-01 -1.74340189e-01 5.18435240e-01
-3.53309304e-01 2.95360684e-01 -8.97485852e-01 -7.99684703e-01
-8.30090046e-01 -1.24135661e+00 9.72656488e-01 -8.34793597e-02
7.56924748e-02 -1.11674261e+00 -1.45297170e-01 1.17085259e-02
4.97766316e-01 -3.13741267e-01 9.09206629e-01 -5.03852308e-01
-3.80492449e-01 -3.61629725e-01 -2.46211082e-01 4.55332160e-01
-4.10151988e-01 2.50214133e-02 -3.81405205e-01 -4.84964937e-01
2.73908675e-01 6.31820187e-02 8.60158980e-01 3.65764499e-01
9.32013571e-01 -7.03696430e-01 -1.20291375e-01 9.38828468e-01
1.27977288e+00 -1.55176714e-01 8.38869810e-02 1.84559494e-01
9.01462972e-01 -6.57768548e-02 4.17570919e-01 6.29997373e-01
4.92760450e-01 6.66329682e-01 1.50993377e-01 -2.15891495e-01
1.78365141e-01 -2.25761861e-01 6.31464660e-01 1.20691001e+00
3.43599677e-01 3.65097937e-03 -1.09566092e+00 4.30860072e-01
-1.71142769e+00 -4.60319579e-01 3.11906319e-02 2.50348425e+00
1.07448816e+00 4.36099291e-01 1.23256028e-01 -1.58763275e-01
2.79873908e-01 6.76724240e-02 -7.65093923e-01 -5.73271275e-01
7.03993961e-02 4.14039716e-02 1.10101473e+00 1.26578057e+00
-8.63797545e-01 1.09361613e+00 6.23218393e+00 1.03279543e+00
-1.53482997e+00 1.68604463e-01 7.50403881e-01 -6.34729624e-01
-3.75215597e-02 7.62961134e-02 -1.05955720e+00 5.06692052e-01
1.10994768e+00 -2.18854517e-01 7.21140862e-01 1.24948382e+00
2.17423633e-01 1.17137888e-02 -9.91374493e-01 9.11065280e-01
-2.18655184e-01 -1.17844701e+00 -2.58467972e-01 1.56106547e-01
7.97267318e-01 5.28398275e-01 2.08805025e-01 2.52718478e-01
2.55307317e-01 -7.37009108e-01 7.28364587e-01 3.49997357e-02
4.17932063e-01 -9.05964494e-01 5.76271355e-01 5.30229926e-01
-1.12941742e+00 -2.16870345e-02 -4.39910471e-01 -1.13067679e-01
4.20289665e-01 9.45588410e-01 -9.04041290e-01 -1.36751339e-01
4.20232296e-01 1.43306658e-01 -2.91927040e-01 7.80643463e-01
-1.05705597e-02 8.78534436e-01 -1.16064334e+00 -2.10649714e-01
5.80622196e-01 -4.65302140e-01 5.51989317e-01 1.43920135e+00
2.14888498e-01 -1.91943496e-01 2.92388737e-01 4.18122977e-01
-2.61027604e-01 2.74203807e-01 1.96737573e-01 -2.15150446e-01
4.57146704e-01 9.84736979e-01 -5.49618840e-01 -6.67495310e-01
-3.85472506e-01 9.24996614e-01 4.95191902e-01 2.30342284e-01
-1.06388032e+00 -1.94021508e-01 7.88090467e-01 3.96506041e-01
5.75739980e-01 -7.48761892e-01 -6.03490949e-01 -1.12913465e+00
6.28992617e-01 -7.56991863e-01 1.58874631e-01 -7.88774416e-02
-7.61390209e-01 4.99960482e-01 -2.32637018e-01 -5.39737046e-01
-3.05216044e-01 -4.87407774e-01 -4.14817363e-01 7.85003424e-01
-1.21188521e+00 -6.30305648e-01 2.34505668e-01 8.59062970e-02
5.21477342e-01 1.94130599e-01 4.39758182e-01 6.07047796e-01
-7.71976471e-01 1.08994651e+00 4.80327278e-01 -2.01559141e-01
4.77160335e-01 -1.12260497e+00 8.34596515e-01 7.41993666e-01
-3.56382318e-02 7.29534090e-01 9.13886786e-01 -3.61687809e-01
-1.48149610e+00 -8.84728789e-01 1.23892927e+00 -8.49738270e-02
9.77494240e-01 -5.37820458e-01 -8.42510045e-01 6.69537783e-01
-3.03960264e-01 9.88042727e-02 8.54754448e-02 9.18165371e-02
-2.99963892e-01 -2.64780551e-01 -8.27335954e-01 9.39721048e-01
7.23451197e-01 -4.61064607e-01 2.60685742e-01 6.67897463e-01
7.17242241e-01 -8.02396894e-01 -7.46194661e-01 9.38339308e-02
6.14983439e-01 -5.67128003e-01 8.17492247e-01 -6.30952954e-01
9.37672704e-02 -8.00247341e-02 -4.40362990e-02 -9.18471336e-01
4.96921614e-02 -1.17265809e+00 -4.03022200e-01 8.00561547e-01
9.00893629e-01 -7.35090137e-01 9.04915512e-01 9.13689971e-01
1.16540939e-02 -1.52560019e+00 -7.96611130e-01 -8.77440572e-01
2.52362221e-01 -5.13746679e-01 3.96889269e-01 5.64348102e-01
-1.58120617e-01 2.47154713e-01 -4.70898330e-01 1.04091920e-01
2.27297083e-01 -1.92175075e-01 8.78417730e-01 -6.17384076e-01
-8.30512643e-01 -5.83947778e-01 1.23828920e-02 -1.65769756e+00
-7.91331157e-02 -6.74507320e-01 -1.59754992e-01 -1.00886691e+00
4.67844456e-02 -4.26569730e-01 2.21259240e-02 3.66737574e-01
-4.38919127e-01 -1.66426793e-01 4.74118620e-01 4.06628922e-02
-3.64512593e-01 4.16233838e-01 8.82656097e-01 -5.84777556e-02
-6.51004612e-01 1.43432111e-01 -2.42825732e-01 9.98638928e-01
7.35447526e-01 -6.72648609e-01 -4.48837489e-01 -8.03095937e-01
4.75182891e-01 -5.36250956e-02 -4.91395481e-02 -8.54972601e-01
3.60449493e-01 -4.52562645e-02 -1.74216405e-01 -5.12428463e-01
5.75378418e-01 -4.12393659e-01 -1.78792119e-01 5.33537805e-01
-3.87266397e-01 6.04979277e-01 3.79008323e-01 2.07983479e-01
-1.44950803e-02 -2.97015369e-01 1.01127088e+00 -9.94098634e-02
2.73003709e-02 4.48805720e-01 -1.95200160e-01 4.09651816e-01
5.76946437e-01 1.05029427e-01 1.96355268e-01 -4.87624705e-01
-4.69082445e-01 3.50425184e-01 4.06240016e-01 9.03858840e-02
1.73835725e-01 -8.82977426e-01 -6.70025826e-01 1.74201485e-02
-7.04042852e-01 7.50271082e-02 3.51454467e-01 1.13270402e+00
-6.92951620e-01 4.19111788e-01 6.61409020e-01 -4.80297863e-01
-1.24492037e+00 4.55797195e-01 4.25072074e-01 -9.72339630e-01
-6.17105007e-01 1.23196697e+00 1.80975839e-01 -3.56783748e-01
3.85970592e-01 -4.69913393e-01 6.20904982e-01 -1.83315963e-01
5.74750721e-01 6.25058413e-01 2.43111864e-01 -2.64487177e-01
-4.80378866e-01 5.30156672e-01 -3.46081406e-01 -4.39035535e-01
1.15936351e+00 3.32005024e-02 -5.69398701e-02 4.42352861e-01
1.54446077e+00 6.28074482e-02 -1.42458999e+00 -1.07462451e-01
-9.94306877e-02 -1.38087302e-01 2.16503039e-01 -4.89225715e-01
-1.25651467e+00 1.04241610e+00 4.22092557e-01 -2.07394674e-01
8.92736197e-01 -4.59456027e-01 1.14532435e+00 4.76629853e-01
1.96781129e-01 -1.22334957e+00 -2.76388168e-01 8.32949698e-01
7.04399645e-01 -1.02850831e+00 2.60463387e-01 -3.66985500e-01
-5.49285650e-01 8.71172369e-01 3.01623285e-01 -2.40053475e-01
8.44554484e-01 4.99236286e-01 -2.84471512e-02 2.88375705e-01
-1.06121445e+00 4.23782676e-01 2.60752469e-01 -1.48366496e-01
4.62152213e-01 1.94240376e-01 -5.13314605e-01 3.49565327e-01
-5.75129688e-01 -3.97759020e-01 9.88053530e-02 7.39353538e-01
-3.38544041e-01 -8.48901153e-01 -1.07006490e-01 2.30458930e-01
-6.31760716e-01 -5.97579241e-01 8.20302777e-03 6.62073612e-01
-3.29349369e-01 7.13450491e-01 2.19000176e-01 -3.70343849e-02
2.34833181e-01 3.65361571e-02 3.45650554e-01 -2.06685036e-01
-7.24943936e-01 1.27228841e-01 8.66066217e-02 -6.50035381e-01
5.68806648e-01 -4.14342940e-01 -1.59899497e+00 -7.77073205e-01
-2.87166089e-01 9.37592834e-02 9.39379275e-01 9.37312067e-01
7.51283169e-01 6.56589940e-02 3.95768970e-01 -7.37507105e-01
-1.24161756e+00 -8.03540587e-01 -1.81786433e-01 1.64957136e-01
4.13771242e-01 -2.54717857e-01 -5.14179826e-01 -1.94405451e-01] | [8.27196979522705, 3.5317296981811523] |
4e66a2c6-a464-47b6-bcbe-d3b1dd2eb6e0 | cluda-contrastive-learning-in-unsupervised | 2208.14227 | null | https://arxiv.org/abs/2208.14227v2 | https://arxiv.org/pdf/2208.14227v2.pdf | CLUDA : Contrastive Learning in Unsupervised Domain Adaptation for Semantic Segmentation | In this work, we propose CLUDA, a simple, yet novel method for performing unsupervised domain adaptation (UDA) for semantic segmentation by incorporating contrastive losses into a student-teacher learning paradigm, that makes use of pseudo-labels generated from the target domain by the teacher network. More specifically, we extract a multi-level fused-feature map from the encoder, and apply contrastive loss across different classes and different domains, via source-target mixing of images. We consistently improve performance on various feature encoder architectures and for different domain adaptation datasets in semantic segmentation. Furthermore, we introduce a learned-weighted contrastive loss to improve upon on a state-of-the-art multi-resolution training approach in UDA. We produce state-of-the-art results on GTA $\rightarrow$ Cityscapes (74.4 mIOU, +0.6) and Synthia $\rightarrow$ Cityscapes (67.2 mIOU, +1.4) datasets. CLUDA effectively demonstrates contrastive learning in UDA as a generic method, which can be easily integrated into any existing UDA for semantic segmentation tasks. Please refer to the supplementary material for the details on implementation. | ['Rahul Tallamraju', 'Shuaib Ahmed', 'Anuraag Bhattacharya', 'Jaswin Kasi', 'Midhun Vayyat'] | 2022-08-27 | null | null | null | null | ['synthetic-to-real-translation'] | ['computer-vision'] | [ 2.57225424e-01 1.86804563e-01 -1.06509201e-01 -5.86038172e-01
-1.28668106e+00 -6.56661034e-01 5.49803019e-01 -1.90772012e-01
-7.25941241e-01 5.75526059e-01 -1.07947931e-01 6.42635226e-02
-3.93859576e-03 -9.55991924e-01 -1.03660488e+00 -5.39070785e-01
2.24950820e-01 8.08624506e-01 4.83330369e-01 -1.94627732e-01
-1.91923901e-01 2.33795479e-01 -1.38129878e+00 3.96344513e-01
1.06036329e+00 9.92187381e-01 4.06091183e-01 5.74332595e-01
-1.57498807e-01 5.10977566e-01 -6.88878059e-01 -5.32370210e-01
1.70307145e-01 -5.61470926e-01 -1.13127983e+00 6.16973154e-02
7.04422534e-01 -1.48900062e-01 -8.00339803e-02 9.73038137e-01
4.33227628e-01 3.61807883e-01 9.55807030e-01 -9.32901621e-01
-7.59184897e-01 5.31809211e-01 -6.51069164e-01 2.48924971e-01
-7.30251968e-02 4.37266380e-02 1.04826438e+00 -7.75839150e-01
7.77275741e-01 1.30717623e+00 5.95307052e-01 7.03251660e-01
-1.53258133e+00 -8.21554601e-01 3.05458069e-01 4.19841334e-02
-1.15260184e+00 -1.51216045e-01 7.00286508e-01 -4.83374774e-01
9.94880497e-01 -1.30545020e-01 4.12830234e-01 1.13325715e+00
-3.39596778e-01 1.20447946e+00 1.17898273e+00 -4.72741902e-01
1.53477520e-01 6.29624575e-02 -5.95664978e-03 7.26788163e-01
-2.19128922e-01 3.66653688e-02 -4.42577422e-01 3.04441392e-01
8.12190354e-01 -3.55843246e-01 1.49562731e-01 -5.34524798e-01
-7.74609804e-01 1.08132184e+00 6.26496077e-01 7.15803504e-02
4.21889173e-03 -5.30433655e-03 4.36696947e-01 2.53854990e-01
8.63003492e-01 3.58570367e-01 -7.96512008e-01 -3.42337489e-02
-9.56683934e-01 4.21285033e-01 3.13284308e-01 9.78558004e-01
9.61369872e-01 7.23562017e-02 -1.02500588e-01 1.28134286e+00
1.83669955e-01 5.77716053e-01 4.65983540e-01 -1.37371075e+00
2.55317181e-01 3.63798440e-01 -1.60118386e-01 -2.78982580e-01
-3.09883982e-01 -4.51230139e-01 -5.36786914e-01 4.03627276e-01
6.13938332e-01 -9.03416052e-02 -1.30837572e+00 1.97997546e+00
3.40503454e-01 3.37196738e-01 2.81765461e-01 7.87809193e-01
1.01587296e+00 5.24577618e-01 3.21861058e-01 3.52162421e-01
1.40530765e+00 -1.39217508e+00 -2.18099624e-01 -4.59837317e-01
5.13508320e-01 -6.35930479e-01 1.40296006e+00 4.24731851e-01
-1.34619069e+00 -8.27794015e-01 -8.65652144e-01 -3.90888363e-01
-5.53143680e-01 1.93710849e-01 3.46359164e-01 4.12263095e-01
-1.04413140e+00 5.40147066e-01 -9.82572258e-01 -4.61979240e-01
8.59219015e-01 1.55251294e-01 -7.00554326e-02 -1.73946217e-01
-1.15524828e+00 7.73066103e-01 4.99201804e-01 -6.80834293e-01
-1.15082753e+00 -9.87561464e-01 -1.11242974e+00 -5.15917875e-02
3.80655020e-01 -6.58506095e-01 1.57897162e+00 -1.22700918e+00
-1.67577648e+00 1.51537824e+00 6.94805235e-02 -7.22890854e-01
5.00325680e-01 -3.28777820e-01 -4.18948889e-01 4.26056683e-01
5.98035336e-01 1.29655278e+00 7.86298990e-01 -1.32930481e+00
-8.58571053e-01 -3.85309815e-01 2.61360556e-02 3.09916079e-01
5.86930960e-02 -1.06363013e-01 -4.61438686e-01 -8.55315089e-01
-1.48709327e-01 -7.13479459e-01 -1.97277993e-01 -8.06161538e-02
-8.26396495e-02 -3.74064505e-01 6.74244761e-01 -6.93337262e-01
7.21772671e-01 -2.27728963e+00 1.85376167e-01 -6.28341064e-02
4.68834341e-02 3.84788841e-01 -3.75309795e-01 -1.40909344e-01
-5.72286732e-02 -1.21259838e-01 -8.39631438e-01 -4.64996040e-01
9.16938931e-02 3.09642494e-01 -2.01797009e-01 1.22396819e-01
5.50371766e-01 8.13047171e-01 -9.82370973e-01 -4.37564522e-01
2.17226312e-01 4.31331813e-01 -6.54448867e-01 2.85222352e-01
-5.24741709e-01 6.36033416e-01 -2.99317896e-01 3.54438514e-01
8.31874728e-01 -2.79015481e-01 -1.29909158e-01 1.50538892e-01
5.92971444e-02 3.47538650e-01 -8.59414279e-01 2.39750099e+00
-6.90489054e-01 4.97852415e-01 1.74958318e-01 -1.36380208e+00
9.17336166e-01 -2.08671279e-02 3.71569008e-01 -1.06261563e+00
3.23920213e-02 3.98834795e-01 -2.74135888e-01 -2.03215539e-01
3.38469625e-01 -1.97540969e-01 -3.23560148e-01 2.89676964e-01
7.42258072e-01 -5.17181277e-01 4.93794441e-01 9.70482603e-02
8.83628547e-01 5.93247831e-01 8.31231177e-02 -3.47280353e-01
4.98767138e-01 2.51417547e-01 5.52817166e-01 6.64279163e-01
-2.60875374e-01 1.01451218e+00 3.52294624e-01 -2.17220232e-01
-1.00621331e+00 -1.54095304e+00 -3.72211069e-01 1.47614706e+00
7.25391805e-02 2.48810742e-02 -1.04324150e+00 -1.07888103e+00
4.21466865e-03 7.77187824e-01 -6.10669613e-01 -9.96372029e-02
-6.32106900e-01 -6.58716559e-01 6.27311826e-01 8.52643013e-01
8.60413134e-01 -1.28401530e+00 -4.91277844e-01 1.50451690e-01
-8.57880190e-02 -1.21213996e+00 -3.01433057e-01 3.95221412e-01
-7.86052406e-01 -7.95415223e-01 -1.09132135e+00 -1.15561664e+00
4.31243211e-01 -3.23541276e-02 1.60545576e+00 -5.00501812e-01
-2.75322229e-01 3.48533183e-01 -2.68783391e-01 -2.60071456e-01
-4.43114966e-01 3.30562085e-01 -4.09119457e-01 -3.34907830e-01
4.46223617e-01 -5.01849174e-01 -5.45815706e-01 2.60344267e-01
-8.65169466e-01 5.95882051e-02 3.36660534e-01 9.08796549e-01
9.44336057e-01 -1.45473182e-01 6.65897191e-01 -1.30084491e+00
2.90107369e-01 -4.14349586e-01 -7.65644014e-01 4.79678847e-02
-4.43139583e-01 9.74529535e-02 7.77056754e-01 -1.94625944e-01
-1.34610224e+00 1.04278490e-01 -5.31670690e-01 -4.23727244e-01
-6.66804731e-01 2.06420809e-01 -2.20079631e-01 3.30266207e-01
8.14077497e-01 7.62793198e-02 -2.02787369e-01 -6.68961227e-01
8.42092037e-01 5.81522226e-01 7.77837873e-01 -8.54023039e-01
6.04900241e-01 5.07349014e-01 -4.70936507e-01 -6.43549979e-01
-1.01818657e+00 -5.21025538e-01 -7.74227977e-01 2.44226024e-01
1.13341737e+00 -1.32422805e+00 -1.07618704e-01 6.95867598e-01
-7.84369767e-01 -1.02062356e+00 -6.64073110e-01 2.95729041e-01
-8.98000836e-01 1.11813866e-01 -6.75883949e-01 -1.78077638e-01
-1.96666047e-01 -1.11963642e+00 1.16683471e+00 4.69400644e-01
-1.66092291e-01 -1.11179757e+00 1.79625779e-01 5.79276502e-01
2.76926219e-01 7.10492134e-02 8.54534924e-01 -6.53271377e-01
-3.48173738e-01 4.36945975e-01 -5.11005342e-01 6.63964927e-01
1.65745802e-03 -2.37246305e-01 -1.22429848e+00 -2.08096027e-01
-4.01037604e-01 -6.93682849e-01 1.30156493e+00 5.77903569e-01
1.24640501e+00 1.13373406e-01 -2.21421033e-01 8.24485183e-01
1.30458581e+00 1.60690457e-01 4.06496018e-01 5.51024497e-01
6.15783930e-01 5.85541189e-01 6.74851656e-01 2.24014357e-01
5.66857994e-01 5.76424658e-01 1.71239018e-01 -4.48518068e-01
-6.32420540e-01 -2.83452451e-01 2.12496579e-01 5.93353629e-01
2.48878002e-01 -2.96749156e-02 -8.67946506e-01 1.06525600e+00
-1.64712346e+00 -5.72176158e-01 1.70582980e-01 1.83022940e+00
1.11086047e+00 2.44369134e-01 4.70994592e-01 -2.44740531e-01
6.55947089e-01 1.70417592e-01 -7.40023077e-01 -5.56192935e-01
-1.52691573e-01 9.04233396e-01 4.99530733e-01 4.74361181e-01
-1.42936802e+00 1.40395582e+00 5.86541700e+00 1.19566786e+00
-1.03164530e+00 3.45743716e-01 7.91127443e-01 6.06986359e-02
-3.27471823e-01 -3.22590679e-01 -6.05851293e-01 4.95646417e-01
9.56279755e-01 4.01213691e-02 4.18511093e-01 1.01778936e+00
-2.21029177e-01 -9.09094512e-02 -8.78224969e-01 7.96558619e-01
-3.94834541e-02 -1.19753993e+00 -1.60748899e-01 -3.48491877e-01
1.01198888e+00 3.66715550e-01 3.89233857e-01 6.37068629e-01
8.86241019e-01 -8.45211923e-01 6.67349577e-01 -7.79851079e-02
1.01618016e+00 -8.89279008e-01 3.52563530e-01 2.14725267e-02
-1.14328992e+00 1.84342653e-01 -2.99816906e-01 2.46035799e-01
2.35038083e-02 4.77095574e-01 -5.97857356e-01 5.75761676e-01
1.20021856e+00 9.47837651e-01 -4.93562937e-01 6.35161757e-01
-5.68094313e-01 7.11522937e-01 -3.01696599e-01 5.27086556e-01
4.75569725e-01 -2.49962673e-01 2.27899596e-01 1.41175675e+00
1.47279605e-01 -1.59602091e-01 1.64336145e-01 1.14416015e+00
-4.39619303e-01 2.60617714e-02 -4.07298833e-01 1.98135749e-01
1.93091184e-01 1.07956862e+00 -7.74289787e-01 -5.77712238e-01
-5.64865351e-01 1.16756034e+00 5.41992068e-01 4.73483354e-01
-8.72187436e-01 -4.43404973e-01 6.65368617e-01 -3.07736956e-02
6.39232993e-01 6.01032972e-02 -3.95542353e-01 -1.07986796e+00
-2.85469949e-01 -7.97844768e-01 7.13660121e-01 -7.05167472e-01
-1.39380765e+00 5.62750220e-01 1.64673433e-01 -1.01825142e+00
-2.31614307e-01 -6.10695541e-01 -4.59804267e-01 8.36283207e-01
-1.85658491e+00 -1.21445858e+00 -1.18723974e-01 6.16714656e-01
8.97485137e-01 -2.81583965e-01 8.20907593e-01 3.09756368e-01
-5.09892523e-01 6.40620530e-01 2.93530047e-01 3.18098813e-01
9.08767343e-01 -1.61286879e+00 6.50620461e-01 7.86859095e-01
2.94342428e-01 -1.18978664e-01 2.62655675e-01 -2.49695316e-01
-2.70184040e-01 -1.26794040e+00 6.11839771e-01 -3.57506454e-01
5.30358672e-01 -4.21563834e-01 -1.08740532e+00 9.07111168e-01
4.57963824e-01 9.26002413e-02 6.09678924e-01 2.03596696e-01
-5.41089535e-01 -1.39214575e-01 -1.50203168e+00 4.21430588e-01
1.05265260e+00 -4.48709875e-01 -6.87045932e-01 1.09891206e-01
8.91122162e-01 -5.92516720e-01 -1.02188063e+00 4.40193504e-01
2.32257679e-01 -9.52459931e-01 1.11441088e+00 -4.50648606e-01
6.54189050e-01 -5.97510822e-02 -5.44626778e-03 -1.51644075e+00
-3.61685127e-01 -1.72773451e-01 2.29361564e-01 1.26401615e+00
5.10783315e-01 -6.03012145e-01 9.78311896e-01 1.73123226e-01
-4.45486575e-01 -4.73372847e-01 -1.00270772e+00 -7.82219410e-01
8.31444621e-01 -3.15057486e-01 4.59923446e-01 1.00202680e+00
-4.30221170e-01 4.17223603e-01 1.70474872e-01 4.21932079e-02
7.00841010e-01 2.37081707e-01 5.84297955e-01 -1.13380587e+00
-4.31684256e-01 -6.55476511e-01 -1.12266047e-02 -1.22055960e+00
4.30006325e-01 -1.15661919e+00 2.61824839e-02 -1.56311011e+00
1.14828676e-01 -5.66077948e-01 -4.81502891e-01 5.11705458e-01
-1.67345643e-01 5.21116853e-01 9.66871977e-02 -4.59073670e-02
-6.88861132e-01 6.00924909e-01 1.38467515e+00 -3.35321903e-01
-1.08361885e-01 -1.35169491e-01 -6.18171334e-01 8.20189357e-01
8.84345293e-01 -5.31660199e-01 -5.73290646e-01 -7.30665088e-01
-3.67026806e-01 -2.23838896e-01 2.01772720e-01 -9.63331223e-01
-2.74200559e-01 1.30296052e-02 3.62267882e-01 -4.01230752e-01
3.76034886e-01 -4.90832359e-01 -4.56425071e-01 2.62300074e-01
-3.52879494e-01 -3.14359039e-01 5.18812299e-01 2.42630497e-01
-4.72041637e-01 -2.37733170e-01 1.22867620e+00 -3.58618677e-01
-1.05641019e+00 1.66174665e-01 -2.20727101e-01 7.69483507e-01
8.22970390e-01 -1.06467254e-01 -3.05083871e-01 7.67528405e-03
-1.00183403e+00 3.64463001e-01 5.10336220e-01 2.58133709e-01
3.62847269e-01 -1.15362275e+00 -7.72635043e-01 3.91856194e-01
1.85552895e-01 6.84514701e-01 3.66584241e-01 4.27970052e-01
-5.54915130e-01 1.25708923e-01 -4.58235383e-01 -7.58379340e-01
-8.55588853e-01 2.89865017e-01 4.13356155e-01 -4.64660823e-01
-5.45125127e-01 1.19236803e+00 6.08087838e-01 -9.21267569e-01
5.74697256e-02 -3.05719525e-01 -4.83972766e-02 1.38729531e-02
2.36476883e-01 2.94311196e-01 7.51293376e-02 -5.24059594e-01
-3.28030169e-01 7.10775197e-01 -1.89695224e-01 -2.69073993e-01
1.32175016e+00 -1.44149780e-01 3.35162431e-01 2.55174816e-01
1.21148932e+00 -2.82302707e-01 -1.83842981e+00 -3.85336250e-01
-1.25806272e-01 -2.13121027e-01 -1.16157703e-01 -1.31900442e+00
-1.28831601e+00 1.07761335e+00 7.36550927e-01 -1.85067393e-02
1.34175825e+00 3.66732270e-01 9.41807270e-01 -9.08633024e-02
1.34360462e-01 -1.36132824e+00 2.14211658e-01 6.12845302e-01
4.90390211e-01 -1.44197702e+00 -3.15904140e-01 -4.00837392e-01
-8.63183618e-01 6.67364836e-01 7.35076547e-01 -5.00364363e-01
5.35408914e-01 1.51678815e-01 3.48804832e-01 5.19510582e-02
-3.91462147e-01 -6.03036165e-01 1.45344630e-01 1.02560663e+00
4.44411010e-01 1.39924034e-01 -1.19004607e-01 8.10197949e-01
-2.14247718e-01 -1.24042764e-01 2.17953756e-01 7.15541482e-01
-2.99357116e-01 -1.27016211e+00 -1.19409963e-01 1.66642010e-01
-5.58362186e-01 -4.73853983e-02 -6.78719804e-02 8.29971433e-01
3.69933724e-01 7.17822134e-01 3.93607289e-01 7.99330324e-02
4.50677007e-01 1.78152964e-01 6.79422319e-01 -7.44295299e-01
-5.28669715e-01 3.31665993e-01 5.04952557e-02 -4.24863756e-01
-4.75799263e-01 -7.25487947e-01 -1.59543860e+00 1.21798486e-01
1.73144609e-01 3.02294996e-02 4.42127883e-01 8.17342639e-01
3.31762522e-01 7.27348924e-01 3.89808565e-01 -6.66217446e-01
-1.12670384e-01 -9.43707168e-01 -6.89356029e-01 7.18669415e-01
1.43876880e-01 -8.33743274e-01 -1.56170577e-01 2.65957981e-01] | [9.733423233032227, 1.386281967163086] |
26546ad5-bd88-4ddd-80c7-aa6f53b8d9bf | harmonious-semantic-line-detection-via | 2104.06903 | null | https://arxiv.org/abs/2104.06903v1 | https://arxiv.org/pdf/2104.06903v1.pdf | Harmonious Semantic Line Detection via Maximal Weight Clique Selection | A novel algorithm to detect an optimal set of semantic lines is proposed in this work. We develop two networks: selection network (S-Net) and harmonization network (H-Net). First, S-Net computes the probabilities and offsets of line candidates. Second, we filter out irrelevant lines through a selection-and-removal process. Third, we construct a complete graph, whose edge weights are computed by H-Net. Finally, we determine a maximal weight clique representing an optimal set of semantic lines. Moreover, to assess the overall harmony of detected lines, we propose a novel metric, called HIoU. Experimental results demonstrate that the proposed algorithm can detect harmonious semantic lines effectively and efficiently. Our codes are available at https://github.com/dongkwonjin/Semantic-Line-MWCS. | ['Chang-Su Kim', 'Seong-Gyun Jeong', 'Wonhui Park', 'Dongkwon Jin'] | 2021-04-14 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Jin_Harmonious_Semantic_Line_Detection_via_Maximal_Weight_Clique_Selection_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Jin_Harmonious_Semantic_Line_Detection_via_Maximal_Weight_Clique_Selection_CVPR_2021_paper.pdf | cvpr-2021-1 | ['line-detection'] | ['computer-vision'] | [ 2.64180191e-02 -1.81162767e-02 -2.55808830e-01 -1.96500033e-01
-4.50976819e-01 -4.94555771e-01 1.37818307e-01 1.98378846e-01
-6.12168713e-03 5.42170107e-01 5.39970361e-02 1.21367641e-01
-3.62517208e-01 -1.29883754e+00 -4.12819654e-01 -3.86761904e-01
5.30815795e-02 2.23203853e-01 5.39165914e-01 -2.96636969e-01
4.49618995e-01 2.08458751e-01 -1.20984995e+00 -1.37262359e-01
1.23778713e+00 6.39242530e-01 9.43317115e-02 2.15734243e-01
-1.45402983e-01 3.22555691e-01 -6.42836928e-01 -1.55616626e-01
1.99177250e-01 -8.76761079e-01 -6.99252069e-01 2.26137578e-01
-1.52044576e-02 1.28215462e-01 -3.59313309e-01 1.40486336e+00
4.37300086e-01 2.49488037e-02 2.80966818e-01 -1.51113713e+00
-4.53194022e-01 1.08450365e+00 -9.35700893e-01 -2.42118418e-01
4.98709649e-01 -8.60699639e-02 1.27964592e+00 -9.97177124e-01
7.53456593e-01 1.07395673e+00 6.45976126e-01 3.21766175e-02
-8.98019910e-01 -7.84041941e-01 -9.09232125e-02 3.36376488e-01
-1.87085605e+00 -1.53967723e-01 1.17060804e+00 -8.34504142e-02
2.17161238e-01 6.07002914e-01 1.02391934e+00 2.11643890e-01
-3.90027821e-01 8.91092479e-01 8.00833881e-01 -5.79571068e-01
1.40845925e-01 -1.03489220e-01 4.68877815e-02 8.85331869e-01
4.80347037e-01 -3.82524461e-01 -6.74097180e-01 -1.17129073e-01
6.37638450e-01 -4.24125314e-01 -4.18429106e-01 -3.48349720e-01
-9.73041713e-01 7.69258022e-01 4.70334381e-01 5.30250490e-01
2.60169711e-02 -5.27889617e-02 2.50489026e-01 5.68391904e-02
1.98355138e-01 2.90633410e-01 7.37053454e-02 1.69264629e-01
-1.03653669e+00 7.68780038e-02 5.70874155e-01 1.15877521e+00
1.12216902e+00 -2.68672198e-01 -2.22035810e-01 1.10335302e+00
1.35175809e-01 3.73297364e-01 -6.81310669e-02 -8.63716364e-01
3.46568555e-01 9.46904778e-01 -1.19172506e-01 -1.52028847e+00
-6.22602403e-01 -5.29444575e-01 -8.71199250e-01 -3.31284851e-01
1.01589650e-01 -1.55184388e-01 -3.42153728e-01 1.55854559e+00
4.55118001e-01 3.77763718e-01 -3.00621420e-01 8.81076276e-01
7.43952751e-01 5.74642003e-01 -3.39483351e-01 -3.52991551e-01
1.26202095e+00 -1.01927757e+00 -7.60952592e-01 4.91500534e-02
4.49802339e-01 -1.01623487e+00 1.00128794e+00 2.40203410e-01
-1.21510839e+00 -4.21048939e-01 -1.17638433e+00 1.52910650e-01
-2.12662771e-01 4.13858861e-01 5.39248586e-01 4.97576803e-01
-1.09991074e+00 4.56167400e-01 -3.99919331e-01 -4.41827953e-01
3.23422432e-01 2.85474092e-01 1.34146251e-02 9.85803530e-02
-1.16675138e+00 3.40240985e-01 7.52784431e-01 2.31495142e-01
-2.24306747e-01 -3.75578940e-01 -9.18123662e-01 2.84902751e-01
6.30387962e-01 -7.25861013e-01 9.29819405e-01 -7.60935843e-01
-1.22816670e+00 7.66960263e-01 -1.86110497e-01 -1.78771690e-02
4.30441171e-01 8.67761225e-02 -6.95506692e-01 3.84495437e-01
2.78363198e-01 5.28343260e-01 2.53689378e-01 -1.62089062e+00
-9.12594914e-01 -2.35584319e-01 -2.49981880e-01 2.69889623e-01
-5.82952380e-01 -5.80627024e-02 -1.22211111e+00 -6.52090430e-01
4.61525321e-01 -8.30242574e-01 -1.31056204e-01 -2.36829251e-01
-1.28563428e+00 -3.76915991e-01 6.24565363e-01 -5.91038167e-01
1.92235935e+00 -2.10698724e+00 -1.68104470e-01 1.08803272e+00
2.84759521e-01 -5.64905182e-02 -2.44827121e-01 6.90816045e-01
9.65492278e-02 2.40733907e-01 -4.02284712e-01 1.12165242e-01
-3.69114056e-03 -1.74011737e-01 1.38859689e-01 4.32955056e-01
-1.18850261e-01 6.54403865e-01 -8.08384299e-01 -6.38806343e-01
1.19400389e-01 5.07603288e-02 -3.76495987e-01 -1.41986936e-01
6.72845468e-02 2.19908673e-02 -6.06551588e-01 6.65998161e-01
9.62342978e-01 -2.93975174e-01 5.80788136e-01 -3.64674956e-01
-3.49289715e-01 -9.47467983e-02 -1.60390544e+00 1.62452924e+00
8.18245336e-02 4.17268664e-01 -3.29997152e-01 -8.21304917e-01
1.41373789e+00 -1.71697512e-01 7.61521161e-01 -7.33094394e-01
1.17796190e-01 4.49381292e-01 -1.54071003e-01 -2.26689473e-01
7.98998713e-01 5.77164412e-01 -3.60761940e-01 2.28255317e-01
-2.28720143e-01 -6.16919436e-02 8.17330897e-01 4.04954433e-01
8.92548680e-01 -1.39633901e-02 2.94756711e-01 -3.67144763e-01
7.53186584e-01 3.25149506e-01 1.08236670e+00 6.66657269e-01
-1.40429184e-01 5.15428901e-01 6.60728157e-01 -7.32187033e-02
-1.03816390e+00 -9.68448639e-01 -9.67019983e-03 6.12781525e-01
1.02964056e+00 -7.99991369e-01 -9.02376413e-01 -2.60731071e-01
-7.92655647e-02 5.89966893e-01 -3.53510231e-01 -3.19082409e-01
-6.29801869e-01 -5.71759522e-01 3.53890628e-01 1.46225855e-01
7.00306594e-01 -9.26293373e-01 -2.81567037e-01 1.78259730e-01
-3.29922050e-01 -7.91843951e-01 -8.24204087e-01 -4.75029171e-01
-4.79895622e-01 -1.32871926e+00 -4.57822204e-01 -1.18766248e+00
9.68541384e-01 5.60841620e-01 9.55812156e-01 4.38923359e-01
-3.53364080e-01 -4.50528748e-02 -3.88429701e-01 -5.71153164e-02
-5.61090931e-02 1.41539171e-01 -2.93322623e-01 -1.00967929e-01
3.28858078e-01 -4.74519342e-01 -8.74304891e-01 5.58589160e-01
-5.11598706e-01 2.54649460e-01 3.12002450e-01 5.60487628e-01
9.24934447e-01 4.73466307e-01 4.61831063e-01 -9.06710684e-01
9.03490722e-01 -2.25323528e-01 -8.72012317e-01 5.69187701e-01
-4.72333133e-01 -3.96020442e-01 5.70623279e-01 1.88197583e-01
-9.81919587e-01 6.08277954e-02 -3.28227542e-02 -1.14643008e-01
1.93354890e-01 5.67853510e-01 -1.02378421e-01 1.80732436e-03
1.48847997e-01 1.19367346e-01 -5.03653169e-01 -2.14522958e-01
5.55090547e-01 5.08561313e-01 8.13792646e-01 -3.63175899e-01
1.16482496e+00 6.23182476e-01 -1.16604365e-01 -6.98224723e-01
-6.05499685e-01 -7.10732341e-01 -5.42816162e-01 -5.30786693e-01
6.20185912e-01 -6.77875638e-01 -7.38844991e-01 4.70260441e-01
-9.66160297e-01 2.41372343e-02 -2.11383685e-01 2.79237658e-01
-3.12383384e-01 4.72505897e-01 -5.69465160e-01 -6.88424289e-01
-4.82037961e-01 -6.20763600e-01 6.81424439e-01 7.69988656e-01
-4.12317812e-01 -7.98522592e-01 6.65794984e-02 1.97697714e-01
-1.37160748e-01 3.87209266e-01 7.20285118e-01 -4.14688230e-01
-5.29465675e-01 -3.35346651e-03 -3.32416505e-01 -2.57976949e-01
2.19269127e-01 4.91875708e-01 -4.09935176e-01 -1.66826054e-01
-6.46112561e-01 8.40413272e-02 8.10459435e-01 5.71919143e-01
1.18684518e+00 -9.50689614e-02 -5.71794093e-01 7.09090531e-01
1.62020767e+00 4.39577997e-01 7.25488484e-01 5.22870183e-01
6.23600423e-01 5.30017793e-01 8.07278574e-01 8.06149125e-01
5.31323135e-01 6.03104293e-01 3.33328582e-02 -5.75582683e-01
-4.50076535e-02 -5.76756895e-01 -1.17137365e-01 1.27676284e+00
1.00342646e-01 -4.76215571e-01 -8.55242252e-01 5.53111732e-01
-2.18698430e+00 -8.31804574e-01 -5.65479040e-01 2.05141950e+00
5.25500119e-01 3.09635788e-01 3.49442422e-01 4.37133700e-01
1.19815195e+00 -2.44761910e-02 -4.00284082e-01 -1.22137949e-01
-5.11392534e-01 -1.67816877e-02 3.35231185e-01 3.93122971e-01
-9.40234721e-01 1.15378022e+00 5.36772299e+00 1.13129103e+00
-6.28712893e-01 -1.64846361e-01 5.83093941e-01 1.26598105e-01
-6.30504966e-01 1.94221720e-01 -5.66069782e-01 5.06617785e-01
-3.60574499e-02 -5.28878868e-01 1.20499350e-01 5.42496264e-01
5.08681178e-01 -2.68533409e-01 -3.66582096e-01 8.37194264e-01
-1.17132012e-02 -1.40813124e+00 -3.15039717e-02 -4.02075708e-01
8.62435758e-01 -4.88042742e-01 -1.98270917e-01 -1.47226304e-01
4.25574720e-01 -3.72821540e-01 5.70273817e-01 3.90289128e-01
7.31127024e-01 -1.19736028e+00 5.01707852e-01 -9.24410149e-02
-1.69217467e+00 1.13127381e-01 -3.51521194e-01 4.79599297e-01
2.54090875e-01 8.16010773e-01 -3.76759231e-01 9.91932452e-01
6.97883546e-01 6.89828753e-01 -5.35496771e-01 1.54946971e+00
-6.12725198e-01 4.42051232e-01 -2.50266820e-01 -1.06696606e-01
1.04139693e-01 -7.20830739e-01 6.04671717e-01 1.01635635e+00
5.86050808e-01 4.81253378e-02 3.19619119e-01 9.11177337e-01
-2.09165990e-01 5.39983749e-01 -1.08386770e-01 1.74391553e-01
1.14142489e+00 1.54224455e+00 -1.22643352e+00 -1.44433424e-01
-2.36461535e-01 9.10594761e-01 2.37392277e-01 3.53914142e-01
-7.88509786e-01 -9.63998795e-01 2.19310433e-01 -1.10437170e-01
1.02641605e-01 -1.28393680e-01 -4.21641350e-01 -8.95151496e-01
4.08535963e-03 -5.03656507e-01 7.18758523e-01 -8.28889191e-01
-9.28179801e-01 3.41154903e-01 -3.13137054e-01 -1.33210957e+00
4.37659055e-01 8.78076032e-02 -1.14420986e+00 4.35982227e-01
-1.28713655e+00 -8.46527994e-01 -6.65362000e-01 5.87533176e-01
3.93542200e-01 1.02485791e-01 2.97497183e-01 3.10655981e-01
-9.08166707e-01 6.21354222e-01 1.48927137e-01 2.11227775e-01
6.03260040e-01 -1.03322184e+00 1.53071359e-01 8.86008978e-01
1.50949895e-01 4.86634314e-01 6.17150545e-01 -7.04294562e-01
-1.03367031e+00 -8.70110571e-01 8.81432056e-01 5.10706127e-01
5.92789292e-01 -3.29153955e-01 -8.11556876e-01 2.73686171e-01
3.36199135e-01 -4.72894043e-01 6.63421929e-01 -1.41740963e-02
4.70005944e-02 -1.76917627e-01 -9.00448263e-01 7.56407619e-01
1.29772127e+00 -1.27582461e-01 -1.65314317e-01 4.52076793e-01
7.20287502e-01 -2.78172195e-01 -9.14085090e-01 3.82909685e-01
4.18392569e-01 -1.11639357e+00 8.48834634e-01 2.68842876e-01
3.32334071e-01 -7.21820831e-01 4.30868506e-01 -1.36877596e+00
-5.53747952e-01 -8.30972314e-01 3.88313562e-01 1.46841347e+00
5.90919554e-01 -4.99116600e-01 7.83768654e-01 -2.80268729e-01
-2.19960034e-01 -7.54942000e-01 -4.85891700e-01 -7.51474321e-01
-4.77929175e-01 -4.11341749e-02 1.02536547e+00 1.25241709e+00
3.13758105e-01 2.41845086e-01 -3.78772885e-01 1.53537109e-01
8.13024342e-01 5.86245000e-01 6.49696648e-01 -1.09980977e+00
-9.58812982e-02 -7.74226069e-01 -1.21109352e-01 -6.91024721e-01
-7.92295858e-02 -8.55035543e-01 -1.08253621e-01 -1.93974507e+00
3.55135173e-01 -5.33566415e-01 -4.27263767e-01 3.81767571e-01
-3.23392928e-01 3.62079233e-01 1.87248826e-01 3.58400524e-01
-8.51070583e-01 6.67032480e-01 1.35390115e+00 1.47975143e-02
-5.48880517e-01 -2.70564646e-01 -6.75048649e-01 8.95299792e-01
1.35901093e+00 -4.27728713e-01 -2.70908564e-01 -1.97052941e-01
2.20688045e-01 -6.82476675e-03 -3.37088034e-02 -1.04825914e+00
4.95770335e-01 -2.42905423e-01 3.31027508e-01 -8.78784716e-01
-7.08567277e-02 -3.78392816e-01 2.18910679e-01 7.27065444e-01
-2.58584917e-01 6.98059623e-04 -2.42891703e-02 2.51583129e-01
-4.37272936e-01 -1.38650686e-01 6.15501821e-01 2.40752324e-01
-1.01257622e+00 2.11964458e-01 -2.60594357e-02 -2.27123667e-02
1.19802022e+00 -4.26379919e-01 -5.79504490e-01 -3.08706760e-01
-5.90156794e-01 8.27914834e-01 4.54491585e-01 4.36649472e-01
8.03405941e-01 -1.51878440e+00 -7.42782950e-01 4.67863977e-02
3.34316969e-01 -2.38868386e-01 2.91760951e-01 6.66964889e-01
-7.72721827e-01 -6.40665879e-03 -1.02180257e-01 -2.78922021e-01
-1.42813921e+00 1.19672984e-01 5.54387607e-02 5.74978404e-02
-5.50882638e-01 8.38796794e-01 9.67725366e-03 -3.53614092e-01
-6.96291476e-02 2.34984040e-01 -3.55253309e-01 -4.62142192e-02
-1.86052248e-02 5.97355366e-01 -3.91984463e-01 -6.55146658e-01
-4.29725945e-01 8.05782020e-01 3.02083492e-01 1.25212982e-01
1.20468950e+00 -2.15442792e-01 -5.16069472e-01 -3.44956270e-03
8.40773582e-01 3.36181939e-01 -8.02484512e-01 -1.29579559e-01
3.41916412e-01 -5.12152493e-01 -2.94266582e-01 -3.57417285e-01
-1.15518069e+00 2.39300817e-01 2.58329242e-01 2.97342390e-01
1.49291754e+00 -3.13340686e-02 9.76407528e-01 -8.19140952e-03
1.63301185e-01 -1.53584838e+00 -1.51148401e-02 1.69968560e-01
6.44913137e-01 -7.25640893e-01 7.76396245e-02 -1.15571785e+00
-3.60080779e-01 1.11964667e+00 8.25514972e-01 -2.98415180e-02
3.94837677e-01 2.12395430e-01 -5.61104976e-02 -4.12991971e-01
-3.59706551e-01 -4.22294259e-01 1.51610643e-01 2.20285714e-01
2.27047563e-01 1.47829682e-01 -8.47302973e-01 3.86173725e-01
-5.15818894e-01 -1.10877320e-01 6.08642042e-01 7.10452557e-01
-7.66823113e-01 -1.08789861e+00 -4.26438838e-01 2.33955070e-01
1.60773590e-01 -1.25761339e-02 -6.94924235e-01 7.42456913e-01
3.27985436e-01 1.21269250e+00 2.43009672e-01 -6.36769891e-01
5.15659690e-01 -5.00918448e-01 2.07495958e-01 -2.59973079e-01
-4.12023991e-01 5.25321066e-01 3.24011296e-01 -3.77389073e-01
-4.03591931e-01 -2.33562753e-01 -1.59073985e+00 -3.32178473e-01
-5.64013481e-01 4.88532454e-01 2.62069404e-01 3.69621098e-01
3.45992565e-01 6.12927198e-01 9.87659633e-01 -9.19799507e-02
1.09442279e-01 -4.56041515e-01 -8.12169135e-01 3.84488821e-01
-4.19201046e-01 -3.12660456e-01 -1.63400218e-01 -1.55228719e-01] | [7.582265377044678, 4.667274475097656] |
5dd2facb-9fe4-437b-a4f8-db71bb020478 | fourier-domain-optimization-for-image | 1809.04187 | null | http://arxiv.org/abs/1809.04187v1 | http://arxiv.org/pdf/1809.04187v1.pdf | Fourier-Domain Optimization for Image Processing | Image optimization problems encompass many applications such as spectral
fusion, deblurring, deconvolution, dehazing, matting, reflection removal and
image interpolation, among others. With current image sizes in the order of
megabytes, it is extremely expensive to run conventional algorithms such as
gradient descent, making them unfavorable especially when closed-form solutions
can be derived and computed efficiently. This paper explains in detail the
framework for solving convex image optimization and deconvolution in the
Fourier domain. We begin by explaining the mathematical background and
motivating why the presented setups can be transformed and solved very
efficiently in the Fourier domain. We also show how to practically use these
solutions, by providing the corresponding implementations. The explanations are
aimed at a broad audience with minimal knowledge of convolution and image
optimization. The eager reader can jump to Section 3 for a footprint of how to
solve and implement a sample optimization function, and Section 5 for the more
complex cases. | ['Sabine Süsstrunk', 'Frederike Dümbgen', 'Majed El Helou', 'Radhakrishna Achanta'] | 2018-09-11 | null | null | null | null | ['reflection-removal'] | ['computer-vision'] | [ 5.79100847e-01 -1.73637554e-01 3.77183408e-01 -1.61695749e-01
-7.58037329e-01 -4.89965349e-01 1.49848640e-01 -3.42236489e-01
-6.25168562e-01 7.14128971e-01 -5.28907217e-02 -3.93855751e-01
-2.24019825e-01 -1.54621065e-01 -4.28256035e-01 -8.80515277e-01
-1.74792185e-01 -1.95521593e-01 -4.01139081e-01 -1.62347276e-02
3.22273284e-01 5.06388128e-01 -1.55347884e+00 -7.84302223e-03
8.80492687e-01 9.73345697e-01 3.66251439e-01 1.01058447e+00
-1.14532135e-01 5.70654213e-01 -4.91731495e-01 -3.73712808e-01
3.61855447e-01 -4.83576596e-01 -6.40664518e-01 5.43575168e-01
3.98537546e-01 -5.68768561e-01 -1.77180856e-01 1.37310040e+00
6.62510157e-01 2.47990116e-01 5.07371724e-01 -9.60840225e-01
-6.65850401e-01 1.02224410e-01 -6.95211947e-01 3.66748460e-02
3.08584899e-01 2.72464547e-02 5.17786205e-01 -1.09408617e+00
2.71589279e-01 9.46462035e-01 9.97622669e-01 3.16753298e-01
-1.06989789e+00 -1.72002912e-01 1.08551339e-03 2.41509587e-01
-1.29562438e+00 -7.19640911e-01 4.55589414e-01 -2.53083348e-01
8.74984741e-01 4.58798587e-01 6.16233468e-01 3.71770591e-01
-1.40072167e-01 6.63294077e-01 1.26025915e+00 -7.77704656e-01
-1.38129085e-01 1.91795498e-01 1.63242072e-01 7.25094259e-01
-9.37343761e-02 1.35050625e-01 -4.61417139e-01 -1.66949615e-01
9.16996241e-01 -9.41876173e-02 -7.44262516e-01 7.05312192e-02
-1.30377495e+00 9.33502495e-01 2.93071657e-01 1.70626834e-01
-4.09978777e-01 2.91488290e-01 1.71655625e-01 4.64235038e-01
6.13339841e-01 2.09158584e-01 -1.80017278e-01 3.76730636e-02
-1.23807442e+00 3.41511309e-01 8.62780094e-01 7.96729445e-01
7.16409504e-01 4.02126044e-01 1.07184067e-01 9.75473762e-01
2.70724505e-01 6.82735085e-01 2.33567283e-01 -1.37536931e+00
1.77624542e-02 -4.49331850e-01 3.34713399e-01 -7.97445059e-01
-3.47636998e-01 -2.63634771e-01 -9.68431532e-01 5.09685993e-01
4.69206303e-01 -4.71560687e-01 -6.50134206e-01 1.12262702e+00
9.55525264e-02 3.32699627e-01 -1.45558402e-01 1.16977692e+00
8.94751072e-01 8.23994040e-01 -5.37264168e-01 -6.52390003e-01
1.69420326e+00 -1.02655768e+00 -8.09716702e-01 -1.98310196e-01
2.06169188e-01 -1.37006199e+00 7.30664730e-01 5.82432449e-01
-1.33982146e+00 -1.97394475e-01 -1.03016508e+00 -2.69380480e-01
-1.30627468e-01 2.43140608e-01 6.87864721e-01 9.00398612e-01
-1.25556183e+00 7.54853427e-01 -6.88127100e-01 -4.11018342e-01
1.90488800e-01 3.91172528e-01 7.75429141e-03 -3.10591254e-02
-7.16319382e-01 1.04513812e+00 1.60169899e-02 3.66619289e-01
-5.14210343e-01 -8.75176728e-01 -7.02457190e-01 -1.54525191e-01
1.39718741e-01 -8.13109696e-01 1.19362462e+00 -9.24412847e-01
-1.69781280e+00 1.10911906e+00 -5.43171465e-01 -4.97792661e-01
6.41616642e-01 -2.69114822e-01 -2.05358773e-01 1.70150548e-01
-3.17265809e-01 2.45934516e-01 1.41484284e+00 -9.31327462e-01
-3.73906434e-01 -2.15905085e-02 5.97372502e-02 3.86078030e-01
-1.55602992e-01 3.06050152e-01 -4.29710299e-01 -7.13209689e-01
1.81264415e-01 -7.82025039e-01 -2.81181186e-01 2.48674572e-01
-5.32610536e-01 5.14795303e-01 7.81593204e-01 -9.65303779e-01
1.00833941e+00 -2.10911131e+00 -3.08108870e-02 1.26770392e-01
2.56201446e-01 1.66931391e-01 -3.44823264e-02 4.97268915e-01
-4.78523463e-01 -3.83890152e-01 -5.70889652e-01 -4.74180490e-01
-7.63305835e-03 -1.79235920e-01 -4.07918841e-01 1.09591532e+00
-1.62616611e-01 7.64159858e-01 -6.99104667e-01 -6.12168424e-02
4.41962510e-01 7.88094044e-01 -5.36264241e-01 2.92404462e-03
2.74168670e-01 5.32703936e-01 -1.34430408e-01 5.56265116e-01
9.74910378e-01 -3.60544980e-01 -1.60511464e-01 -5.81483543e-01
-5.58077216e-01 -4.92435656e-02 -1.59817183e+00 1.30824769e+00
-6.40534282e-01 1.14810395e+00 8.71945620e-01 -1.30736113e+00
4.60028052e-01 3.14047694e-01 5.85729241e-01 -3.00529033e-01
-1.59657523e-01 3.95941108e-01 -3.44079494e-01 -5.24551928e-01
6.52880192e-01 -4.31291074e-01 4.47111398e-01 7.29817092e-01
-2.86215842e-01 -4.80162442e-01 1.61649778e-01 1.45030189e-02
7.88382232e-01 -1.94492072e-01 4.17857647e-01 -3.25542241e-01
7.84178972e-01 -1.51212281e-02 -1.69766918e-01 7.90857136e-01
9.38105434e-02 8.47037733e-01 7.43176639e-02 -4.17790294e-01
-1.14191508e+00 -8.48042369e-01 -3.87043297e-01 8.92734706e-01
2.23016992e-01 -2.81246156e-01 -9.15775537e-01 1.28395781e-01
-2.13275746e-01 3.57755899e-01 -1.33505881e-01 3.73669624e-01
-5.65535665e-01 -1.15648961e+00 2.34617189e-01 -3.00805899e-03
7.42761970e-01 -6.76671207e-01 -7.15577304e-01 -8.28484446e-03
-4.57681328e-01 -1.03134346e+00 -5.94302535e-01 2.15486884e-01
-1.09527814e+00 -7.84207106e-01 -1.29477715e+00 -6.39598012e-01
8.21504831e-01 7.46435463e-01 9.97314155e-01 4.89452183e-01
-5.83497167e-01 7.17572808e-01 6.50301352e-02 -2.30585009e-01
-1.47316784e-01 -4.54163581e-01 -3.80212702e-02 1.40539110e-02
-3.10634412e-02 -7.07432568e-01 -7.18733609e-01 1.47570059e-01
-9.56493020e-01 9.62358043e-02 2.63045043e-01 9.92780328e-01
4.92196351e-01 8.65103081e-02 -5.32769300e-02 -8.58543038e-01
8.26909661e-01 5.40084019e-02 -8.66405606e-01 8.00291598e-02
-5.98514736e-01 5.60761578e-02 3.81397754e-01 -2.32762605e-01
-9.83267069e-01 1.02782339e-01 -2.21523777e-01 -2.44139731e-01
1.14070646e-01 4.33101565e-01 3.78055841e-01 -7.31630206e-01
7.47353792e-01 3.09411198e-01 2.33566344e-01 -4.94893730e-01
6.04556918e-01 5.53353608e-01 8.22717369e-01 -3.79828274e-01
1.04975235e+00 8.32642078e-01 4.29366380e-02 -1.41432345e+00
-7.04961419e-01 -5.94392121e-01 -3.17264140e-01 -3.16110015e-01
6.94949329e-01 -7.65561044e-01 -8.93021047e-01 5.69141805e-01
-1.34057021e+00 -3.43153894e-01 -2.69314170e-01 5.59823453e-01
-8.89304996e-01 7.32262433e-01 -6.55513704e-01 -8.08429301e-01
-3.34498733e-01 -1.26353467e+00 9.40080941e-01 3.12632263e-01
-1.33499065e-02 -1.28145623e+00 -2.20024228e-01 3.17674249e-01
7.67096996e-01 -3.69908176e-02 5.56754589e-01 1.59707651e-01
-7.42287278e-01 -1.25373587e-01 -4.77995843e-01 3.74216348e-01
3.08111720e-02 -1.43261790e-01 -1.28907311e+00 -4.62993920e-01
4.55426216e-01 -6.25427440e-02 8.65805209e-01 8.29836905e-01
1.26186395e+00 -2.94345319e-01 -1.97012633e-01 1.26242077e+00
1.55771375e+00 -3.52417409e-01 7.91596770e-01 2.57509232e-01
4.20331568e-01 6.85098767e-01 2.48292968e-01 5.37660420e-01
-7.82987848e-02 6.37656569e-01 4.27956879e-01 -4.45237845e-01
-2.24968672e-01 6.53252363e-01 3.82425636e-01 5.91744423e-01
-4.90332097e-01 -8.65564421e-02 -3.75605583e-01 2.64800847e-01
-1.62182522e+00 -1.17048192e+00 -4.49737012e-01 2.27833915e+00
5.93999922e-01 -5.87796867e-01 -2.32082978e-02 1.88939586e-01
8.22673440e-01 2.25111127e-01 -3.74818593e-01 -4.08156276e-01
-2.84200281e-01 3.54348302e-01 1.05967247e+00 9.84149814e-01
-1.29486120e+00 4.91558701e-01 7.64741516e+00 7.72146761e-01
-1.21142411e+00 2.05805898e-01 4.34699565e-01 -1.91940755e-01
-2.30251208e-01 1.38701592e-02 -3.79412860e-01 2.74702519e-01
6.82398736e-01 -1.34366527e-01 9.97722030e-01 3.61121356e-01
5.10372937e-01 -3.51874083e-01 -7.30576277e-01 1.72484016e+00
4.13654149e-02 -1.36332822e+00 -6.75096333e-01 8.34651813e-02
7.45104432e-01 1.33289352e-01 1.28366798e-01 -4.00885910e-01
-4.95063215e-02 -1.16433024e+00 5.18601060e-01 4.02689725e-01
6.99397624e-01 -4.28482056e-01 2.38978252e-01 3.08846325e-01
-7.85980701e-01 1.13332666e-01 -4.41868097e-01 -4.63498533e-02
5.22524238e-01 1.06072319e+00 -1.85849503e-01 4.43470269e-01
5.57905078e-01 5.70141256e-01 1.68008804e-01 1.32222331e+00
-7.55229443e-02 3.38933676e-01 -6.89332664e-01 2.62451649e-01
2.10941419e-01 -8.96537066e-01 7.48202264e-01 1.36868143e+00
8.11863303e-01 3.10004234e-01 -2.88957089e-01 7.74142206e-01
2.16401696e-01 7.75588118e-03 -2.50577003e-01 1.26967393e-02
-3.26113813e-02 1.39559007e+00 -7.68414497e-01 -2.16825366e-01
-8.62636328e-01 1.44035983e+00 -2.65299350e-01 7.93624222e-01
-7.19824970e-01 -6.72902584e-01 7.84962773e-01 7.36787973e-04
3.68350834e-01 -3.69193912e-01 -3.77695173e-01 -1.28605974e+00
-3.95453982e-02 -9.17185903e-01 1.31798342e-01 -9.80459452e-01
-1.11135674e+00 3.80915582e-01 4.33142809e-03 -1.03016293e+00
-1.85254365e-01 -9.27885234e-01 -7.01500058e-01 1.26392829e+00
-1.75763249e+00 -5.56271434e-01 -4.52545375e-01 7.94736624e-01
3.46935570e-01 4.74920422e-02 6.90217316e-01 5.86633921e-01
-3.45607609e-01 2.48002738e-01 5.39199233e-01 -1.98195815e-01
6.15838349e-01 -1.00563884e+00 1.15850911e-01 1.04656863e+00
2.90323626e-02 6.96240604e-01 1.16519666e+00 -8.44955519e-02
-1.56222260e+00 -4.65677857e-01 6.83723390e-01 2.31103748e-01
5.99479735e-01 1.48234954e-02 -8.48352432e-01 4.24188167e-01
4.81979609e-01 8.99878517e-02 4.60075408e-01 -2.82462329e-01
4.57953252e-02 7.37734511e-02 -1.10725558e+00 4.95353013e-01
8.40445995e-01 -6.10241354e-01 -1.28356274e-02 9.21781242e-01
1.09784808e-02 -8.01941633e-01 -5.18008292e-01 -1.78230003e-01
5.57086170e-01 -1.33179760e+00 1.39362633e+00 -1.26440361e-01
2.11674049e-01 -4.40668195e-01 3.61892628e-03 -1.16678488e+00
-1.63638338e-01 -1.16514540e+00 2.10416298e-02 6.62866533e-01
1.73003986e-01 -9.76360619e-01 6.36793315e-01 4.75814700e-01
-7.06634298e-02 -3.80863547e-01 -8.26333284e-01 -6.07840002e-01
-3.06241959e-01 -4.84428406e-01 2.34291613e-01 8.20386350e-01
-1.81359977e-01 -4.14558053e-02 -6.70870602e-01 3.25811744e-01
1.00358295e+00 2.94103414e-01 6.41879380e-01 -7.14654505e-01
-5.76147556e-01 -5.39199412e-01 -1.97988883e-01 -1.64530051e+00
-1.13033965e-01 -8.78671348e-01 1.96502641e-01 -1.42124927e+00
3.17623951e-02 -2.84446716e-01 3.41959655e-01 -6.49663061e-02
-1.29821241e-01 6.38281822e-01 2.94220507e-01 2.92658508e-01
1.07107600e-02 1.30285591e-01 1.24426699e+00 -3.65287364e-02
2.35056374e-02 2.00847238e-01 -7.99605370e-01 8.85573864e-01
5.78723609e-01 -2.28810415e-01 -3.03287983e-01 -6.87281907e-01
2.97482938e-01 1.43196210e-02 7.18178988e-01 -7.17768192e-01
3.89323026e-01 5.85251972e-02 3.01571131e-01 -4.26480144e-01
7.66797245e-01 -8.06959212e-01 2.89688647e-01 2.63272852e-01
-7.42331594e-02 -1.02330245e-01 2.48923108e-01 2.07979396e-01
-3.33562434e-01 -9.09440875e-01 1.02600694e+00 -3.06115180e-01
-6.53448761e-01 2.03041151e-01 -4.59961683e-01 -1.42527580e-01
7.83820271e-01 -4.88769412e-01 -1.78417265e-01 -1.04096508e+00
-8.62340689e-01 -7.50950798e-02 5.05030036e-01 -4.56925899e-01
6.71083033e-01 -1.10162544e+00 -8.13188374e-01 1.79938316e-01
-5.32188535e-01 -3.00774515e-01 4.34851289e-01 1.44659460e+00
-1.00934494e+00 2.10074499e-01 2.25242168e-01 -5.36484480e-01
-1.47395921e+00 2.74748236e-01 6.57930672e-01 2.40990948e-02
-8.11720908e-01 1.15103030e+00 3.75329942e-01 1.90234128e-02
8.78838897e-02 -2.87685804e-02 2.51843840e-01 -8.90481174e-02
9.21669304e-01 6.78767800e-01 2.61392325e-01 -5.09836614e-01
-2.04996705e-01 1.03153288e+00 3.21886897e-01 -2.64637500e-01
1.33082545e+00 -5.32329023e-01 -4.72511381e-01 -1.28233969e-01
1.33662891e+00 1.59438774e-01 -1.32593310e+00 -1.13596827e-01
-4.36860442e-01 -5.58827221e-01 4.33586568e-01 -2.53318429e-01
-1.23525906e+00 1.09884834e+00 6.64418519e-01 4.45215762e-01
1.56243801e+00 -1.70931429e-01 6.78891182e-01 4.29154187e-01
-2.90245265e-02 -8.89136791e-01 -2.36064926e-01 5.65279186e-01
8.99021924e-01 -9.87717927e-01 5.24468958e-01 -5.82650065e-01
-1.20144941e-01 1.54049075e+00 -3.78889769e-01 -1.66097075e-01
8.15893829e-01 6.19830191e-01 1.69821799e-01 -2.39565391e-02
-1.59748018e-01 -2.24428937e-01 2.64814317e-01 7.80770421e-01
6.95520222e-01 -2.30187729e-01 -4.47313748e-02 -1.90612346e-01
-2.30784535e-01 -5.93325198e-02 6.77716732e-01 7.60583580e-01
-5.27430236e-01 -1.21006143e+00 -1.04736555e+00 4.35114741e-01
-5.98371387e-01 -5.58612347e-01 1.76787838e-01 3.62946898e-01
-2.30170593e-01 8.02407384e-01 -6.14238121e-02 3.55480254e-01
8.73812512e-02 -9.68930423e-02 9.08420801e-01 -3.56582493e-01
-4.00307357e-01 3.39810073e-01 -5.93727268e-03 -5.46069860e-01
-7.29700267e-01 -6.94714606e-01 -1.03045213e+00 -7.92196453e-01
-3.20703864e-01 -1.04152232e-01 8.51841807e-01 8.42831314e-01
-3.10100354e-02 1.97953612e-01 3.12158167e-01 -1.21879923e+00
-5.03986359e-01 -5.72016180e-01 -8.63323271e-01 4.47645336e-02
7.72777200e-01 -2.81272531e-01 -4.81740296e-01 5.17738044e-01] | [11.651957511901855, -2.6571545600891113] |
f7b3a4b3-31c1-4ac3-b8c7-6d869e84d59f | feta-a-benchmark-for-few-sample-task-transfer | 2205.06262 | null | https://arxiv.org/abs/2205.06262v2 | https://arxiv.org/pdf/2205.06262v2.pdf | FETA: A Benchmark for Few-Sample Task Transfer in Open-Domain Dialogue | Task transfer, transferring knowledge contained in related tasks, holds the promise of reducing the quantity of labeled data required to fine-tune language models. Dialogue understanding encompasses many diverse tasks, yet task transfer has not been thoroughly studied in conversational AI. This work explores conversational task transfer by introducing FETA: a benchmark for few-sample task transfer in open-domain dialogue. FETA contains two underlying sets of conversations upon which there are 10 and 7 tasks annotated, enabling the study of intra-dataset task transfer; task transfer without domain adaptation. We utilize three popular language models and three learning algorithms to analyze the transferability between 132 source-target task pairs and create a baseline for future work. We run experiments in the single- and multi-source settings and report valuable findings, e.g., most performance trends are model-specific, and span extraction and multiple-choice tasks benefit the most from task transfer. In addition to task transfer, FETA can be a valuable resource for future research into the efficiency and generalizability of pre-training datasets and model architectures, as well as for learning settings such as continual and multitask learning. | ['William Yang Wang', 'Jay Pujara', 'Lise Getoor', 'Deepak Ramachandran', 'Luke Yoffe', 'Connor Pryor', 'Pegah Jandaghi', 'Yi-Lin Tuan', 'Alon Albalak'] | 2022-05-12 | null | null | null | null | ['dialogue-understanding'] | ['natural-language-processing'] | [ 4.31786686e-01 2.69134253e-01 -1.70285016e-01 -6.99210405e-01
-1.10274923e+00 -9.91086364e-01 9.36218143e-01 -1.75431103e-01
-5.79199076e-01 1.19639158e+00 6.16112888e-01 -3.65204751e-01
1.73557654e-01 -3.37868363e-01 -3.74545068e-01 -3.98348361e-01
-5.45827709e-02 1.05934989e+00 6.09068833e-02 -6.41579151e-01
-8.69991444e-03 -1.84878245e-01 -7.35714853e-01 8.60188425e-01
8.65818441e-01 6.72113478e-01 1.07305788e-01 7.42655039e-01
-1.95275843e-01 8.08823824e-01 -8.89357448e-01 -6.63600028e-01
6.04359806e-02 -6.22948825e-01 -1.56785488e+00 -3.27172056e-02
4.17906225e-01 -3.54288399e-01 -1.64656535e-01 2.25470066e-01
8.40101898e-01 3.92026335e-01 1.02716827e+00 -1.41250169e+00
-7.08763242e-01 7.99458504e-01 -2.28389785e-01 2.20951453e-01
4.41083282e-01 4.20481503e-01 1.06471407e+00 -8.64622176e-01
5.80117762e-01 1.61129117e+00 6.57404184e-01 1.03990912e+00
-1.39673436e+00 -7.18607664e-01 -1.51082188e-01 -1.07021175e-01
-5.82583070e-01 -8.84840429e-01 4.68534648e-01 -5.66126466e-01
1.12489402e+00 4.34201434e-02 1.83263585e-01 1.75098550e+00
-6.54538423e-02 1.12165964e+00 1.33954048e+00 -5.42934060e-01
-2.12229580e-01 3.80466670e-01 2.83396482e-01 3.82229120e-01
-1.25089601e-01 -1.82599485e-01 -9.28491950e-01 -3.17179948e-01
4.57461864e-01 -6.17040813e-01 -3.09651673e-01 -1.02705106e-01
-1.54867363e+00 1.12417841e+00 1.13829166e-01 2.96130300e-01
-5.99177517e-02 -1.49413243e-01 1.03692055e+00 9.22741830e-01
9.29376841e-01 8.87259483e-01 -8.88603449e-01 -6.66310370e-01
-1.51097521e-01 3.62001926e-01 1.44486988e+00 1.09550774e+00
7.65289307e-01 -2.44837150e-01 -5.28810859e-01 1.37738121e+00
-2.07150012e-01 4.25314248e-01 5.30554235e-01 -1.01661050e+00
1.06357241e+00 3.80909383e-01 1.08321160e-02 -3.31838727e-01
-4.72350359e-01 1.98148027e-01 -5.46118617e-01 -4.20344293e-01
9.68799651e-01 -8.75154018e-01 -1.91029191e-01 1.90358603e+00
6.98352605e-02 -2.87690729e-01 4.86282408e-01 4.36802894e-01
9.73274112e-01 6.36697352e-01 2.75711596e-01 -1.14633970e-01
1.53239608e+00 -1.25755882e+00 -5.92365623e-01 -5.68467200e-01
1.24704111e+00 -8.73242974e-01 1.41127300e+00 -2.47281063e-02
-1.03681827e+00 -3.12435478e-01 -4.69128489e-01 -3.71509403e-01
-2.48127535e-01 -2.45832235e-01 6.89348519e-01 4.29716051e-01
-1.09565473e+00 1.35653406e-01 -3.53987336e-01 -6.61751866e-01
2.35429600e-01 6.80268556e-02 -3.10986012e-01 -1.19779564e-01
-1.58217657e+00 1.45840466e+00 4.14341450e-01 -3.68012398e-01
-9.72857535e-01 -9.85062242e-01 -8.43653798e-01 4.09168713e-02
3.29310358e-01 -8.05721402e-01 1.98666120e+00 -1.00827134e+00
-1.80893111e+00 1.12550628e+00 -2.30222374e-01 -5.36046267e-01
4.57755417e-01 -3.00820112e-01 3.37910168e-02 -1.92363024e-01
1.45865604e-01 8.73597145e-01 4.25429732e-01 -1.02027202e+00
-6.95133567e-01 -6.87682331e-02 1.74385503e-01 7.43429184e-01
-6.51567340e-01 1.52452812e-01 1.74937099e-01 -4.28547114e-01
-8.09722126e-01 -1.16535592e+00 2.74403602e-01 -4.16577876e-01
-2.38237754e-01 -7.74344027e-01 5.69464743e-01 -6.89735949e-01
9.48366106e-01 -1.87784421e+00 2.22626150e-01 -4.39372241e-01
9.56854671e-02 1.95519596e-01 -4.10385311e-01 8.56373727e-01
2.26623029e-01 1.09958299e-01 -3.49836677e-01 -3.80015016e-01
-1.87598653e-02 9.55241267e-03 -1.92533419e-01 -1.18089207e-01
2.55560577e-01 1.12537301e+00 -8.75171244e-01 -4.07197863e-01
-1.05495788e-01 -6.71615079e-02 -4.20364112e-01 6.05059087e-01
-3.26158673e-01 7.39535511e-01 -4.60245788e-01 -1.69037487e-02
1.04687653e-01 -3.00600022e-01 7.03926831e-02 7.39678442e-02
3.03671539e-01 9.80803847e-01 -2.63189822e-01 1.99988484e+00
-8.63986671e-01 9.68353868e-01 3.55875850e-01 -7.50099838e-01
9.11079586e-01 6.30416512e-01 2.93063670e-01 -5.73933959e-01
8.07924662e-03 1.89190060e-02 4.75292712e-01 -5.03082275e-01
6.51984334e-01 -3.23799789e-01 -4.62133586e-01 1.10012770e+00
2.77007252e-01 -3.63269418e-01 1.44522563e-01 3.52106690e-01
9.43936586e-01 -2.21656099e-01 2.30185598e-01 -4.49214160e-01
2.65675575e-01 3.01051468e-01 8.31110105e-02 5.97883999e-01
-3.54401767e-01 2.76260171e-02 5.22523999e-01 -1.74584091e-01
-8.73758733e-01 -7.74604023e-01 -8.46023858e-02 2.16501307e+00
-2.87441581e-01 -6.71145692e-02 -8.19683790e-01 -8.21728468e-01
2.29215667e-01 9.11408007e-01 -6.11831605e-01 -3.10088784e-01
-5.72365820e-01 -6.40328944e-01 9.01761413e-01 4.50363904e-01
5.22016466e-01 -1.41699874e+00 -4.01837490e-02 2.13384524e-01
-7.71804869e-01 -1.12377453e+00 -8.27825665e-01 1.18164763e-01
-7.16219306e-01 -9.44824755e-01 -8.48005831e-01 -9.79451180e-01
3.97981815e-02 4.45605725e-01 1.71859121e+00 -2.27158234e-01
9.38189775e-02 6.10473096e-01 -3.66806239e-01 -6.45938337e-01
-9.37955797e-01 7.49088585e-01 -3.47690657e-02 -1.57279372e-01
6.02537453e-01 -3.37860763e-01 -1.84906706e-01 5.98035157e-01
-2.38566339e-01 3.04546684e-01 5.32402277e-01 1.19696796e+00
-4.70104337e-01 -8.92325819e-01 1.32424080e+00 -1.22341311e+00
1.54720080e+00 -7.82223403e-01 7.97066614e-02 3.90027374e-01
-2.59660512e-01 -2.42490526e-02 3.54613006e-01 -4.72278088e-01
-1.65117967e+00 -4.79080856e-01 2.65507847e-01 2.87535042e-01
-2.47383013e-01 4.84310985e-01 4.63535674e-02 2.84360349e-01
1.04487324e+00 7.33864605e-02 3.93940479e-01 -4.18351859e-01
6.33677304e-01 1.09167969e+00 1.34052977e-01 -1.08027053e+00
3.28147352e-01 -5.89645505e-02 -7.70659447e-01 -1.06213379e+00
-8.70627403e-01 -4.74522114e-01 -6.76866531e-01 -5.69133796e-02
9.46912766e-01 -1.08748794e+00 -7.44650006e-01 4.57664251e-01
-1.48891747e+00 -1.29751229e+00 -9.79990438e-02 4.02094811e-01
-7.11104929e-01 -7.60314390e-02 -1.04751682e+00 -4.48554367e-01
-4.90043849e-01 -1.02377510e+00 9.19220448e-01 -8.24481472e-02
-8.21014941e-01 -1.50163162e+00 3.55024338e-01 7.15610623e-01
5.82668304e-01 -3.00928950e-01 1.05295193e+00 -1.42004001e+00
-2.66938820e-03 2.92962670e-01 -1.70564458e-01 3.28068465e-01
3.81903142e-01 -5.75289667e-01 -1.25200844e+00 -5.53065717e-01
-1.52885765e-01 -1.38881302e+00 6.38315797e-01 1.19237430e-01
5.45567334e-01 -2.49006256e-01 -3.13038290e-01 -3.41702141e-02
4.17769194e-01 2.13977501e-01 9.78035480e-02 1.19896367e-01
5.93289554e-01 1.17228925e+00 5.16359866e-01 1.58915699e-01
7.20139384e-01 7.26488888e-01 -3.42902333e-01 4.87971827e-02
-2.36563966e-01 -5.72349019e-02 6.06547058e-01 1.19595742e+00
1.74076408e-01 -3.50411147e-01 -1.21229029e+00 6.48784280e-01
-1.65123403e+00 -8.03036928e-01 2.76490748e-01 2.02229714e+00
1.55718243e+00 -6.75835460e-02 4.64648485e-01 -5.26452720e-01
6.51702464e-01 1.21623926e-01 -5.41344643e-01 -6.17088795e-01
9.88884494e-02 -3.41264606e-02 5.86777776e-02 6.61047399e-01
-9.76350963e-01 1.10424399e+00 6.27233362e+00 7.70224988e-01
-7.98576176e-01 3.87773216e-01 7.51795113e-01 9.26492214e-02
-1.70656905e-01 -2.38828227e-01 -8.09732199e-01 2.86303461e-01
1.21948624e+00 -6.21612787e-01 4.05486673e-01 7.60424078e-01
-1.11037567e-01 6.88710809e-02 -1.82061291e+00 6.13163114e-01
4.93266247e-02 -9.65694726e-01 1.66050140e-02 -5.66614745e-03
8.10938954e-01 1.94756210e-01 -1.64480194e-01 9.58393037e-01
7.32450902e-01 -1.06516755e+00 9.21504721e-02 6.64268732e-02
9.22280431e-01 -4.49417144e-01 6.64561510e-01 5.64509869e-01
-6.12613559e-01 1.70427993e-01 -1.52391151e-01 -2.79272646e-01
1.57026619e-01 -1.35990873e-01 -1.52510953e+00 2.46002257e-01
3.86227638e-01 7.15447247e-01 -2.93678254e-01 4.21658933e-01
6.98708883e-03 8.95623207e-01 -8.15057382e-02 -2.58223921e-01
1.79836422e-01 -9.87297148e-02 3.65710765e-01 1.63759160e+00
-2.01673165e-01 1.57479830e-02 4.06940192e-01 4.89137918e-01
-5.44955909e-01 2.34251440e-01 -9.56283629e-01 -1.08476281e-01
8.82218003e-01 1.06852555e+00 -1.30850673e-01 -4.04815048e-01
-6.29764438e-01 8.76366913e-01 6.82086110e-01 3.45903754e-01
-3.36614966e-01 -3.11449587e-01 7.26631463e-01 -1.72697037e-01
-3.64372283e-01 -1.77821919e-01 -4.14835721e-01 -1.08985305e+00
-2.49051228e-01 -1.30824804e+00 4.06497568e-01 -4.43024397e-01
-1.93695819e+00 5.61298788e-01 3.08471620e-01 -7.30379820e-01
-8.55703354e-01 -5.25948822e-01 -7.14337707e-01 1.10413611e+00
-1.26781440e+00 -1.25672472e+00 -2.22030267e-01 7.37499177e-01
1.21583652e+00 -5.27692497e-01 1.21639788e+00 7.64461011e-02
-3.38196754e-01 8.94934833e-01 5.87224737e-02 4.41758156e-01
1.59176159e+00 -1.26269722e+00 6.23213470e-01 -1.23855136e-02
-1.30279049e-01 4.55366790e-01 3.85492742e-01 -4.03266072e-01
-1.14507926e+00 -9.25224423e-01 8.15078795e-01 -1.09005904e+00
8.39178979e-01 -6.96311653e-01 -1.14494634e+00 1.20037699e+00
9.18292701e-01 -5.71255624e-01 8.98405671e-01 7.49782681e-01
-2.73480177e-01 9.76420492e-02 -8.22515786e-01 5.66766262e-01
9.62165058e-01 -7.95316994e-01 -8.87566447e-01 6.00938201e-01
9.96806383e-01 -4.25912738e-01 -1.14821160e+00 4.86666560e-02
3.96124542e-01 -6.35442436e-01 7.32377887e-01 -1.01730359e+00
5.51081955e-01 8.49863708e-01 2.05646306e-01 -1.87123215e+00
-1.48847550e-01 -8.32309961e-01 2.02335492e-01 1.58429587e+00
9.89058614e-01 -8.08056712e-01 4.04093355e-01 8.96912813e-01
-3.71013373e-01 -4.36745971e-01 -7.29442000e-01 -6.90801382e-01
8.57767582e-01 1.91686880e-02 2.31009707e-01 1.37150323e+00
3.15097243e-01 1.28969479e+00 -3.98549110e-01 -7.34537244e-01
1.62523806e-01 -2.12112889e-01 1.29921806e+00 -1.24377322e+00
-2.68467128e-01 -5.12893677e-01 4.90454793e-01 -1.21788228e+00
6.73247099e-01 -1.17873704e+00 1.80759251e-01 -1.22335887e+00
3.00143212e-01 -7.67115057e-01 2.09101006e-01 6.49720192e-01
-3.83129597e-01 -1.66740060e-01 2.77388543e-01 4.13816839e-01
-5.95465183e-01 7.64482796e-01 1.48204124e+00 -4.64283042e-02
-4.51545089e-01 1.60109058e-01 -6.84472024e-01 5.34088492e-01
9.32289362e-01 -3.94544929e-01 -7.51890957e-01 -7.12895274e-01
-3.02414596e-01 3.31719548e-01 -1.08222343e-01 -3.82697523e-01
-2.94500291e-02 -9.83492061e-02 -1.72158867e-01 1.25728697e-01
5.75416982e-01 -4.22532916e-01 -7.47775674e-01 1.30987301e-01
-1.11819291e+00 4.50493023e-02 5.31095088e-01 3.83976400e-01
-1.64182529e-01 -1.67809695e-01 5.76285005e-01 -2.08380595e-01
-6.18994355e-01 -1.23185227e-02 -5.24339259e-01 1.05247331e+00
8.51338327e-01 1.04443900e-01 -7.43314385e-01 -8.10462475e-01
-5.57395399e-01 6.06377125e-01 5.31672016e-02 7.05216527e-01
6.11267053e-02 -1.18591881e+00 -1.33744943e+00 -2.69840389e-01
3.96564424e-01 2.23839544e-02 1.28349409e-01 6.83451056e-01
3.61614004e-02 6.12588286e-01 -3.77796203e-01 -6.07764721e-01
-1.20101786e+00 1.54198661e-01 2.18104810e-01 -6.05874717e-01
-2.05346987e-01 9.21036601e-01 5.35803378e-01 -9.47680056e-01
2.13684484e-01 -2.16212124e-02 -1.07620470e-01 3.90384167e-01
4.23852265e-01 4.69062209e-01 -1.95654646e-01 -3.06177676e-01
-6.54373989e-02 -9.09616724e-02 -4.60309654e-01 -3.87219578e-01
9.40256357e-01 -2.87198871e-01 -9.20641050e-02 9.71063733e-01
1.23302746e+00 -3.49911273e-01 -1.16486311e+00 -6.73048496e-01
2.10765943e-01 6.03573136e-02 -6.05142832e-01 -1.21572495e+00
-3.67871284e-01 1.21227920e+00 1.42282732e-02 3.26147646e-01
5.81738412e-01 1.20796382e-01 7.90826619e-01 7.60267138e-01
2.87323922e-01 -9.04040158e-01 4.97747391e-01 1.16456795e+00
1.25638998e+00 -1.57383168e+00 -4.73919243e-01 -3.03934932e-01
-1.20670807e+00 7.88828731e-01 1.11844862e+00 1.90927878e-01
2.81213880e-01 9.57383290e-02 2.50068277e-01 -5.48825860e-02
-1.14150739e+00 1.58477217e-01 2.14067727e-01 6.34484768e-01
1.07778931e+00 9.89070088e-02 -7.73551315e-02 5.82310557e-01
-3.61525089e-01 -1.71561152e-01 4.01068240e-01 7.56887734e-01
-2.14942694e-01 -1.12336683e+00 -5.03542349e-02 4.96805787e-01
-1.93739057e-01 -4.45852458e-01 -8.32197249e-01 9.64338720e-01
-6.25180185e-01 1.15690207e+00 1.42298743e-01 -2.91080952e-01
3.47717345e-01 7.78564334e-01 3.41331661e-01 -9.99999344e-01
-1.00242710e+00 -3.70687276e-01 9.87698615e-01 -1.05908528e-01
-4.00550604e-01 -5.38546383e-01 -7.84399927e-01 -3.83003026e-01
-4.06523138e-01 4.96527344e-01 2.78753132e-01 9.50518191e-01
5.04415929e-01 3.94351304e-01 4.71441120e-01 -8.32542896e-01
-8.94459903e-01 -1.84711182e+00 -1.60802960e-01 5.74633658e-01
2.61611998e-01 -6.31663024e-01 -1.73923582e-01 2.04032704e-01] | [12.592072486877441, 8.165547370910645] |
62c2e5a9-fcb8-4e1a-8aa8-50a8da1aa9c5 | exploring-smoothness-and-class-separation-for | 2203.01324 | null | https://arxiv.org/abs/2203.01324v3 | https://arxiv.org/pdf/2203.01324v3.pdf | Exploring Smoothness and Class-Separation for Semi-supervised Medical Image Segmentation | Semi-supervised segmentation remains challenging in medical imaging since the amount of annotated medical data is often scarce and there are many blurred pixels near the adhesive edges or in the low-contrast regions. To address the issues, we advocate to firstly constrain the consistency of pixels with and without strong perturbations to apply a sufficient smoothness constraint and further encourage the class-level separation to exploit the low-entropy regularization for the model training. Particularly, in this paper, we propose the SS-Net for semi-supervised medical image segmentation tasks, via exploring the pixel-level smoothness and inter-class separation at the same time. The pixel-level smoothness forces the model to generate invariant results under adversarial perturbations. Meanwhile, the inter-class separation encourages individual class features should approach their corresponding high-quality prototypes, in order to make each class distribution compact and separate different classes. We evaluated our SS-Net against five recent methods on the public LA and ACDC datasets. Extensive experimental results under two semi-supervised settings demonstrate the superiority of our proposed SS-Net model, achieving new state-of-the-art (SOTA) performance on both datasets. The code is available at https://github.com/ycwu1997/SS-Net. | ['Jianfei Cai', 'ZongYuan Ge', 'Qianyi Wu', 'Zhonghua Wu', 'Yicheng Wu'] | 2022-03-02 | null | null | null | null | ['semi-supervised-medical-image-segmentation'] | ['computer-vision'] | [ 5.02422273e-01 4.45462435e-01 -4.73561525e-01 -5.26532769e-01
-7.73828030e-01 -1.85782522e-01 1.45357028e-01 -1.59235820e-02
-4.60016429e-01 6.38505399e-01 3.40742208e-02 -5.83734587e-02
4.39376980e-02 -4.56920147e-01 -5.29526114e-01 -1.16458392e+00
1.54502690e-01 9.08974037e-02 5.10873616e-01 1.83437049e-01
-1.54525951e-01 3.90212148e-01 -9.15604413e-01 2.58481711e-01
1.38258970e+00 1.10482311e+00 4.07947779e-01 2.94228494e-01
6.75777197e-02 5.63329935e-01 -1.07754678e-01 -1.26924142e-01
4.30282831e-01 -3.41245204e-01 -8.47882211e-01 4.37687546e-01
4.60060716e-01 -2.88764447e-01 -4.20474708e-01 1.58573318e+00
5.06932437e-01 1.17117517e-01 4.99966055e-01 -9.57193673e-01
-6.47461414e-01 2.57410437e-01 -8.97842824e-01 2.70493895e-01
-2.53753483e-01 4.65024322e-01 7.02453673e-01 -4.26233977e-01
7.21109509e-01 7.05804050e-01 5.66802561e-01 6.65254474e-01
-1.43108284e+00 -5.16321242e-01 1.22750133e-01 3.16617638e-02
-1.27052212e+00 -3.65375310e-01 1.10211694e+00 -3.66791725e-01
1.99393317e-01 3.33663702e-01 4.56155866e-01 8.33572149e-01
1.87189385e-01 1.18782485e+00 1.25541520e+00 -2.24008068e-01
1.15151800e-01 7.98616484e-02 2.37027809e-01 8.56506288e-01
1.35130808e-01 1.61181781e-02 -2.38924138e-02 -1.42207682e-01
8.53693545e-01 1.85000807e-01 -7.85918534e-01 -5.12692511e-01
-1.26398015e+00 5.77225626e-01 7.40827262e-01 4.63554800e-01
-4.17115271e-01 -2.90499657e-01 5.17708659e-01 -2.17071623e-01
6.79814517e-01 1.35418147e-01 -3.02796006e-01 2.85306215e-01
-1.11870205e+00 -1.01709820e-01 3.21167618e-01 8.72506678e-01
4.36275095e-01 -8.05417076e-02 -4.87485290e-01 9.39842582e-01
3.15177619e-01 1.85788274e-01 6.11751080e-01 -8.60322952e-01
4.05434012e-01 5.66010714e-01 -1.90218329e-01 -9.70185459e-01
-3.24096918e-01 -6.80593312e-01 -1.33608270e+00 6.32751808e-02
5.92763364e-01 -5.80543578e-02 -1.34718311e+00 1.62610102e+00
6.23334765e-01 5.50308466e-01 -1.19352706e-01 1.04960501e+00
9.34791148e-01 5.69483876e-01 4.81440406e-03 -4.97460157e-01
1.27539599e+00 -1.12351120e+00 -9.65361118e-01 -3.92545223e-01
5.56470573e-01 -5.15753806e-01 1.13559389e+00 6.16592467e-02
-1.01468444e+00 -4.97161984e-01 -9.63668227e-01 1.17257975e-01
1.47507802e-01 2.35365406e-01 3.77174377e-01 4.44940120e-01
-6.58297062e-01 5.92192769e-01 -1.28536940e+00 1.96758032e-01
1.14052069e+00 2.42748082e-01 -2.26214722e-01 -1.84919223e-01
-1.08184040e+00 3.52929592e-01 4.72850978e-01 3.75897199e-01
-5.50108075e-01 -8.38428557e-01 -9.50623691e-01 -1.90583572e-01
3.85481179e-01 -4.95733380e-01 8.93533409e-01 -1.06367850e+00
-1.24355900e+00 1.15648806e+00 -2.82295141e-02 -3.19750935e-01
9.21355486e-01 2.11726025e-01 -3.27233851e-01 4.75401729e-01
2.23927885e-01 7.87274837e-01 6.65197372e-01 -1.43795347e+00
-3.92235398e-01 -4.38548714e-01 -3.86386305e-01 1.68373585e-01
-2.99111545e-01 -2.50497073e-01 -6.22741759e-01 -1.02814639e+00
2.93888539e-01 -1.05014420e+00 -5.36613286e-01 3.87596667e-01
-8.14960122e-01 9.56256464e-02 7.41632104e-01 -7.35577703e-01
1.16255414e+00 -2.33772540e+00 -1.22517139e-01 1.63418755e-01
3.08347821e-01 4.24559444e-01 -1.37981586e-02 -6.23718560e-01
-1.13445885e-01 1.18241850e-02 -9.76232588e-01 -2.60884076e-01
-3.99025768e-01 3.05054039e-01 9.87338871e-02 8.29549670e-01
1.20142810e-01 9.88224506e-01 -1.02832997e+00 -8.96098316e-01
2.84345955e-01 4.27221924e-01 -3.98839444e-01 2.36537114e-01
-1.65958162e-02 7.06003904e-01 -7.67338574e-01 6.74881160e-01
1.01805365e+00 -4.34110105e-01 1.33214025e-02 -2.76649296e-01
3.07434887e-01 -1.03638962e-01 -1.02651858e+00 1.74604690e+00
-6.32314160e-02 4.69248027e-01 3.53727758e-01 -1.07364762e+00
5.26763618e-01 2.90975899e-01 7.99990177e-01 -4.92762029e-01
2.32867301e-01 3.13533932e-01 -7.33410288e-03 -6.17702961e-01
-8.26872438e-02 -2.10952476e-01 3.47356290e-01 7.68014491e-02
-1.29481971e-01 -1.50090843e-01 1.24256551e-01 1.31426737e-01
5.59033215e-01 1.33624440e-02 1.13119565e-01 -4.92535919e-01
5.81499338e-01 -1.43668279e-01 1.08566713e+00 5.25670409e-01
-7.28654027e-01 1.00402987e+00 3.92228812e-01 -1.71268582e-01
-5.72534263e-01 -9.77203488e-01 -6.17131352e-01 4.01304781e-01
4.87567276e-01 1.20687798e-01 -9.65973973e-01 -1.20137370e+00
-3.08018178e-01 5.10926187e-01 -7.71021664e-01 -2.30372012e-01
-5.61329067e-01 -9.94577110e-01 3.85662109e-01 4.13398206e-01
6.51059747e-01 -1.12830222e+00 -4.58687007e-01 1.36087492e-01
-4.54543710e-01 -1.23085272e+00 -8.99987400e-01 1.30011946e-01
-1.02186513e+00 -9.84967828e-01 -1.22516966e+00 -1.15576017e+00
1.21304178e+00 1.35432094e-01 8.66288602e-01 1.45412624e-01
-4.44442272e-01 -1.93197146e-01 -1.88003242e-01 -2.17952549e-01
-4.26938027e-01 -2.88718322e-04 -3.58252406e-01 1.46349773e-01
-5.95211722e-02 -4.10648733e-01 -8.47824752e-01 3.80995303e-01
-1.16154087e+00 2.40038291e-01 4.85533118e-01 1.17604053e+00
1.06159866e+00 1.22555174e-01 3.70941281e-01 -1.08946800e+00
1.37021407e-01 -3.00896019e-01 -3.83699030e-01 3.62026066e-01
-5.50802350e-01 -2.52193362e-01 5.60787857e-01 -5.19352436e-01
-1.04866755e+00 4.29177433e-01 -2.36348823e-01 -4.12363261e-01
-2.93598920e-01 3.11981082e-01 -2.50550389e-01 -6.91179857e-02
4.35611755e-01 5.44012129e-01 2.55160660e-01 -1.51874125e-01
1.75830930e-01 5.72330117e-01 5.74871778e-01 -3.16346705e-01
7.02924728e-01 8.93151224e-01 -1.75131127e-01 -6.94873452e-01
-1.16726208e+00 -5.02446115e-01 -6.45784378e-01 -1.79904521e-01
1.11788583e+00 -7.71089852e-01 -1.92067310e-01 7.53950894e-01
-8.51473212e-01 -6.22146845e-01 -4.20552999e-01 4.56349671e-01
-5.22595167e-01 6.80977523e-01 -8.47063899e-01 -3.20199132e-01
-5.59119523e-01 -1.53080916e+00 8.70294333e-01 5.56659877e-01
2.02224061e-01 -1.14613402e+00 -9.02110040e-02 5.66245973e-01
1.15200810e-01 4.76294845e-01 6.13154769e-01 -5.27617216e-01
-3.66134405e-01 -1.01792969e-01 -2.87333041e-01 7.11156905e-01
5.28207362e-01 -1.21038154e-01 -8.04409385e-01 -3.41758907e-01
2.14133486e-01 -3.99837434e-01 1.02864730e+00 9.51385021e-01
1.73883295e+00 -1.35109276e-01 -3.56947750e-01 1.13213992e+00
1.13691747e+00 -5.11614680e-02 6.23773158e-01 2.33294845e-01
8.90692651e-01 5.41992068e-01 7.31930196e-01 1.30727097e-01
8.47062245e-02 5.13157547e-01 3.09713960e-01 -6.59745932e-01
-2.85264790e-01 -1.11720981e-02 -8.76188353e-02 6.31950557e-01
2.52059817e-01 -9.15284902e-02 -8.38459074e-01 6.80803955e-01
-1.78881657e+00 -6.49716020e-01 -1.55242488e-01 2.01911950e+00
1.27929652e+00 1.62126735e-01 -3.74423638e-02 -1.85048655e-01
8.49798083e-01 3.63350570e-01 -7.46334791e-01 2.66936660e-01
-6.64431304e-02 -1.06046528e-01 7.37288654e-01 4.19831902e-01
-1.48529685e+00 6.26794696e-01 5.19973373e+00 1.25476778e+00
-1.12613320e+00 1.20721668e-01 1.40653956e+00 -7.05386326e-02
-9.09079015e-02 -3.37362766e-01 -4.37593818e-01 7.70714104e-01
2.33351305e-01 1.05814509e-01 1.53367659e-02 6.95172012e-01
2.90232033e-01 -4.28304821e-02 -6.16707146e-01 8.17062199e-01
-1.56643882e-01 -1.39414537e+00 -1.78830817e-01 -3.47273946e-02
1.06945455e+00 7.74763674e-02 7.15894625e-02 -1.92343235e-01
1.19632678e-02 -7.97730148e-01 4.73822355e-01 2.43717045e-01
8.33191752e-01 -6.44815981e-01 6.77253187e-01 3.86696368e-01
-1.02723753e+00 2.30607525e-01 -3.17041337e-01 6.67168260e-01
3.30948263e-01 8.65085483e-01 -4.35248584e-01 6.63687885e-01
6.69442832e-01 1.02177906e+00 -4.91929770e-01 1.01214576e+00
-2.07077503e-01 7.50408649e-01 -2.31998473e-01 4.86121684e-01
3.73651743e-01 -4.83284682e-01 5.94176769e-01 1.30428755e+00
-1.61328297e-02 2.98700422e-01 3.86345893e-01 8.87216330e-01
-7.50382543e-02 7.33505487e-02 6.26130588e-03 2.49474883e-01
1.67089477e-01 1.32093644e+00 -1.14836895e+00 -3.69255453e-01
-2.09568650e-01 1.06001723e+00 8.37761560e-04 4.53046858e-01
-8.70546937e-01 -2.76571393e-01 3.39256704e-01 2.17464998e-01
1.16795354e-01 8.96979943e-02 -5.95633805e-01 -1.25619292e+00
2.53179163e-01 -7.59918213e-01 6.03076339e-01 -4.42197084e-01
-1.32499874e+00 6.58317566e-01 -1.63705572e-01 -1.41939342e+00
3.59197527e-01 -3.36436480e-01 -7.54035234e-01 6.25582933e-01
-1.85938394e+00 -1.03438520e+00 -3.44985574e-01 5.44743121e-01
5.19914746e-01 3.12693194e-02 3.50740701e-01 4.17319924e-01
-8.55538487e-01 7.16866195e-01 3.43716353e-01 4.25041676e-01
8.20942402e-01 -1.21791613e+00 -1.65127851e-02 9.35651362e-01
-1.25102982e-01 3.36803168e-01 4.70491976e-01 -6.83569908e-01
-7.40457833e-01 -1.19719589e+00 3.03743660e-01 1.80580970e-02
5.52803934e-01 -1.94516659e-01 -1.29940796e+00 3.82874221e-01
-1.84703209e-02 8.55946660e-01 5.42217076e-01 -5.71064174e-01
1.35907397e-01 -9.60697532e-02 -1.46906781e+00 5.21884203e-01
8.52354765e-01 -2.79824853e-01 -3.48495901e-01 6.07911587e-01
7.46041119e-01 -7.20253170e-01 -8.64525378e-01 5.81737459e-01
1.32189170e-01 -8.68319035e-01 1.00775218e+00 -3.11382204e-01
6.06829643e-01 -3.69860202e-01 2.52118975e-01 -1.30695975e+00
-3.05842847e-01 -5.69216967e-01 2.39915140e-02 1.13768494e+00
3.44826519e-01 -7.38325596e-01 9.76166010e-01 6.55431688e-01
-3.84002119e-01 -1.20186985e+00 -1.09024620e+00 -6.66256726e-01
1.13714218e-01 -1.69321582e-01 2.21962512e-01 1.14282632e+00
-3.33719879e-01 -2.70427585e-01 -2.19603673e-01 2.47909486e-01
8.59058440e-01 8.95558000e-02 1.95460409e-01 -7.88799226e-01
-3.81914437e-01 -5.61079502e-01 -2.97953248e-01 -1.11984336e+00
2.90655077e-01 -1.14237046e+00 2.92345256e-01 -1.41833711e+00
4.43740189e-01 -6.81183279e-01 -5.91300130e-01 5.37236154e-01
-6.69019818e-01 3.44050825e-01 -5.06720766e-02 3.35551262e-01
-4.62602824e-01 6.99110210e-01 1.89998949e+00 -3.29135597e-01
-2.25991800e-01 2.36263633e-01 -6.97032392e-01 8.20004880e-01
7.38365412e-01 -5.31707406e-01 -4.31269735e-01 -2.94632375e-01
-4.99566078e-01 -1.05241247e-01 4.05425400e-01 -7.04822063e-01
-3.80048365e-03 -2.83405989e-01 3.55373114e-01 -3.59289020e-01
-1.40456378e-01 -8.75699043e-01 -2.73330599e-01 7.81823754e-01
-4.86096829e-01 -7.23100841e-01 1.49339885e-01 6.27583206e-01
-3.10860395e-01 -2.26409554e-01 1.37354553e+00 1.72445811e-02
-3.89261156e-01 6.97581470e-01 1.33228630e-01 2.73322284e-01
1.20110977e+00 -3.25055271e-01 -1.04466878e-01 2.63482276e-02
-8.05788398e-01 5.54834962e-01 4.85052019e-01 1.01552799e-01
6.50045812e-01 -1.16445053e+00 -6.92764878e-01 1.69666767e-01
3.93140391e-02 5.92326701e-01 6.49696767e-01 1.27427709e+00
-5.56083858e-01 -3.40457186e-02 -1.34532481e-01 -7.85048306e-01
-1.22563708e+00 4.95802701e-01 6.54192805e-01 -4.11869019e-01
-9.15842772e-01 9.79250610e-01 6.03847921e-01 -3.73226494e-01
2.37420797e-01 -3.98410469e-01 -5.23699373e-02 -3.23283821e-01
4.57603157e-01 1.18699774e-01 -1.52887896e-01 -6.41386271e-01
-4.57936078e-01 4.25225258e-01 -2.67189831e-01 4.08995748e-01
1.30218315e+00 -9.11312699e-02 -7.27630258e-02 5.04787043e-02
1.35091758e+00 -2.00485274e-01 -1.70993495e+00 -4.38449591e-01
-3.09366047e-01 -5.84613383e-01 4.34110224e-01 -5.70097208e-01
-1.64858508e+00 7.52972841e-01 8.52134764e-01 -1.02173770e-02
1.27039719e+00 2.82701179e-02 1.14199877e+00 -1.63838312e-01
-1.60930455e-01 -1.16437650e+00 -6.29920810e-02 -6.82169273e-02
5.36338031e-01 -1.49294424e+00 1.95413813e-01 -8.18312705e-01
-9.97188807e-01 8.16036701e-01 4.94973928e-01 -7.40081891e-02
6.76087737e-01 2.36324504e-01 2.11711183e-01 -1.79260999e-01
-7.79807046e-02 6.67892843e-02 6.40485406e-01 5.21250784e-01
3.57027471e-01 1.56902164e-01 -1.05120428e-01 5.24951637e-01
4.03054595e-01 5.65629229e-02 3.88089687e-01 8.89793575e-01
-2.59138018e-01 -7.75773704e-01 -2.40097836e-01 6.30216956e-01
-6.81699395e-01 -6.21470883e-02 7.07075074e-02 4.67082232e-01
1.14591964e-01 7.02537298e-01 -1.13357849e-01 1.65797591e-01
1.25661105e-01 -4.25327510e-01 2.90966481e-01 -5.30556381e-01
-3.36106509e-01 4.41380233e-01 -3.27523828e-01 -4.88532007e-01
-4.41314280e-01 -8.41925204e-01 -1.59072673e+00 2.43163124e-01
-3.75638098e-01 -9.42466110e-02 2.10433424e-01 7.88008094e-01
2.29297742e-01 6.24280632e-01 8.44192982e-01 -7.13832378e-01
-5.70992291e-01 -6.66474164e-01 -6.17484212e-01 6.13800287e-01
3.86201352e-01 -3.95871907e-01 -3.48495156e-01 2.93812901e-01] | [14.613163948059082, -2.05411434173584] |
c0d2a9a8-b23a-4ee8-b0c1-2090899ea265 | rethinking-spatial-invariance-of-1 | 2206.05253 | null | https://arxiv.org/abs/2206.05253v2 | https://arxiv.org/pdf/2206.05253v2.pdf | Rethinking Spatial Invariance of Convolutional Networks for Object Counting | Previous work generally believes that improving the spatial invariance of convolutional networks is the key to object counting. However, after verifying several mainstream counting networks, we surprisingly found too strict pixel-level spatial invariance would cause overfit noise in the density map generation. In this paper, we try to use locally connected Gaussian kernels to replace the original convolution filter to estimate the spatial position in the density map. The purpose of this is to allow the feature extraction process to potentially stimulate the density map generation process to overcome the annotation noise. Inspired by previous work, we propose a low-rank approximation accompanied with translation invariance to favorably implement the approximation of massive Gaussian convolution. Our work points a new direction for follow-up research, which should investigate how to properly relax the overly strict pixel-level spatial invariance for object counting. We evaluate our methods on 4 mainstream object counting networks (i.e., MCNN, CSRNet, SANet, and ResNet-50). Extensive experiments were conducted on 7 popular benchmarks for 3 applications (i.e., crowd, vehicle, and plant counting). Experimental results show that our methods significantly outperform other state-of-the-art methods and achieve promising learning of the spatial position of objects. | ['Alexander G. Hauptmann', 'Xiao Wu', 'Jingkuan Song', 'Hong Li', 'Qi Dai', 'Zhi-Qi Cheng'] | 2022-06-10 | rethinking-spatial-invariance-of | http://openaccess.thecvf.com//content/CVPR2022/html/Cheng_Rethinking_Spatial_Invariance_of_Convolutional_Networks_for_Object_Counting_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Cheng_Rethinking_Spatial_Invariance_of_Convolutional_Networks_for_Object_Counting_CVPR_2022_paper.pdf | cvpr-2022-1 | ['object-counting'] | ['computer-vision'] | [ 5.33137843e-02 -3.04121524e-01 1.68094020e-02 -4.79832351e-01
-1.63922787e-01 -4.22181278e-01 7.19173253e-01 -6.47468194e-02
-8.26858222e-01 7.27309525e-01 2.65055932e-02 -4.48882401e-01
1.53848127e-01 -1.26552212e+00 -7.74021626e-01 -5.34197569e-01
2.34041303e-01 3.31889689e-01 6.86033845e-01 1.41255930e-01
5.55505633e-01 6.62601233e-01 -1.51068950e+00 1.91782966e-01
9.49208438e-01 8.37270975e-01 2.80239910e-01 7.03769326e-01
-4.23963606e-01 1.38411260e+00 -7.27109790e-01 -5.17995059e-01
7.07061663e-02 -1.29299015e-01 -7.31458724e-01 -3.12158823e-01
7.86917329e-01 -7.37575591e-01 -5.11351049e-01 1.56467915e+00
3.65475416e-01 2.43450493e-01 8.64513159e-01 -1.18386221e+00
-1.28450477e+00 7.08799779e-01 -9.90153968e-01 6.63547337e-01
-3.23974133e-01 2.19268724e-01 6.09996378e-01 -1.01880121e+00
4.22458351e-02 1.41610634e+00 8.35871577e-01 3.66200268e-01
-9.43449080e-01 -8.38818967e-01 1.74192280e-01 1.88348562e-01
-1.84351766e+00 -7.05386847e-02 3.01410049e-01 -3.83152574e-01
7.86166847e-01 2.51542181e-01 4.36372012e-01 7.56111085e-01
-5.86209968e-02 7.57993639e-01 1.09594059e+00 -2.86457360e-01
3.72581273e-01 -9.11110714e-02 2.57650375e-01 8.93872440e-01
7.57506430e-01 -1.83237106e-01 -1.46981820e-01 -1.90797523e-01
1.40554249e+00 -1.75662860e-02 1.78594559e-01 9.21553059e-04
-1.23792243e+00 9.16841686e-01 9.23278809e-01 4.71051067e-01
-2.86805183e-01 4.63834524e-01 3.46257776e-01 -4.23178375e-01
6.20008469e-01 2.70176888e-01 -1.83418766e-01 -8.12671557e-02
-1.03672898e+00 3.17438453e-01 6.52898729e-01 9.49473441e-01
5.86104214e-01 8.00126940e-02 -9.06408548e-01 6.78420663e-01
3.09599172e-02 6.72154605e-01 1.31193295e-01 -1.12283313e+00
4.35594320e-01 6.68953538e-01 1.46910727e-01 -1.21343327e+00
-4.28037822e-01 -2.26799905e-01 -1.25201273e+00 -2.35435143e-02
8.79122436e-01 1.45767644e-01 -1.05661154e+00 1.54410458e+00
1.24870464e-01 6.11309230e-01 -4.57825810e-01 9.29419935e-01
8.28710198e-01 4.73760575e-01 3.74513000e-01 3.50361347e-01
1.58117568e+00 -7.35987306e-01 -4.84652698e-01 -7.29427338e-02
5.19674301e-01 -5.38589656e-01 1.24860358e+00 7.14368671e-02
-9.81296360e-01 -7.07669675e-01 -1.01539266e+00 -1.90573335e-01
-6.19601130e-01 6.02849007e-01 1.12662745e+00 7.68421888e-01
-7.65616894e-01 6.74886584e-01 -9.33877110e-01 -2.99564660e-01
9.82110083e-01 1.75534993e-01 4.44453508e-02 -1.17952108e-01
-9.80589271e-01 8.14571738e-01 5.42434514e-01 -2.42503546e-02
-4.29995120e-01 -7.47428596e-01 -6.22849286e-01 2.10794196e-01
-2.61335488e-04 -7.18563795e-01 1.27051270e+00 -4.01213944e-01
-1.16348994e+00 8.96038055e-01 -8.43453109e-02 -4.99288797e-01
4.10926700e-01 7.84779564e-02 5.88316880e-02 2.96579488e-03
4.62674439e-01 1.08248127e+00 3.63603532e-01 -1.02277744e+00
-9.77121472e-01 -4.10100639e-01 1.91658542e-01 -1.50323421e-01
-7.41447747e-01 -3.90920416e-02 -4.71967876e-01 -5.75410366e-01
1.96455464e-01 -6.99523628e-01 -1.80308849e-01 4.60244641e-02
-3.93082589e-01 -4.16037470e-01 8.75832677e-01 -4.87032562e-01
1.28855169e+00 -1.98398066e+00 -4.85571861e-01 9.40265041e-03
3.77849221e-01 3.90671760e-01 -6.11284003e-02 -3.66430998e-01
2.30177999e-01 1.60995811e-01 -1.65925786e-01 -2.98985392e-01
-3.33451889e-02 1.89629152e-01 -3.57023299e-01 7.05519974e-01
3.60046804e-01 1.21732795e+00 -1.08393884e+00 -6.94955468e-01
4.38443661e-01 7.36705005e-01 -5.80735564e-01 -2.45184571e-01
-5.66129908e-02 3.21928650e-01 -4.90718931e-01 8.14128160e-01
1.24573016e+00 -5.04125059e-01 -2.76014119e-01 -3.73625249e-01
-2.74468899e-01 5.03568314e-02 -1.11427963e+00 1.23852789e+00
-1.72512263e-01 4.84311610e-01 -3.70159358e-01 -7.19794095e-01
8.74172270e-01 -3.57133627e-01 2.77178586e-01 -7.03929007e-01
3.70630950e-01 2.57601678e-01 7.14903548e-02 5.35310060e-02
8.43643606e-01 2.97666758e-01 -4.53526415e-02 5.09942621e-02
6.91177621e-02 -1.11974925e-01 4.32797730e-01 1.55336201e-01
1.05321360e+00 -2.27242187e-01 2.81211585e-01 -6.08677030e-01
4.50565189e-01 -4.52528149e-02 3.58286083e-01 1.14403427e+00
-6.09723806e-01 7.27930784e-01 5.93102098e-01 -7.28888392e-01
-1.11618376e+00 -1.02549255e+00 -3.62477690e-01 1.30083609e+00
1.04295045e-01 -5.72188646e-02 -9.58085239e-01 -6.77660048e-01
1.63471866e-02 6.49927735e-01 -8.16921532e-01 3.12994234e-02
-7.11702824e-01 -1.14509130e+00 1.15625739e+00 1.15810776e+00
1.26538432e+00 -9.19928253e-01 -4.77582723e-01 9.54345465e-02
-2.62717694e-01 -1.40096331e+00 -5.60604572e-01 2.23507285e-01
-6.26110971e-01 -8.93738806e-01 -9.09073651e-01 -6.48877561e-01
7.78848886e-01 5.54014623e-01 1.19373906e+00 1.82604745e-01
-2.55043536e-01 -5.12272976e-02 -5.63080832e-02 -6.39072835e-01
-1.88719127e-02 3.31819355e-01 -7.72118941e-03 -5.32954335e-01
9.69761968e-01 -4.64668602e-01 -7.81048536e-01 4.56769228e-01
-9.77467835e-01 -2.60042012e-01 4.25435007e-01 6.88467264e-01
5.67042112e-01 2.17443869e-01 3.75603050e-01 -8.24324429e-01
6.75005972e-01 -3.29207689e-01 -8.52670074e-01 1.28143013e-01
-4.24899608e-02 -1.02568731e-01 5.83995700e-01 -7.17298627e-01
-8.91859055e-01 5.69663867e-02 -2.17656828e-02 -4.28377271e-01
3.27702835e-02 -8.68243948e-02 9.35616419e-02 -2.54193962e-01
9.39257264e-01 2.44801491e-01 -5.18886209e-01 4.29065339e-03
4.16042536e-01 4.53419805e-01 7.59267926e-01 -6.60010338e-01
7.54388511e-01 8.43546569e-01 7.36186802e-02 -7.14145362e-01
-9.89917159e-01 -5.41478276e-01 -6.72378600e-01 5.37555963e-02
1.05008638e+00 -8.49790692e-01 -1.06849515e+00 6.27799153e-01
-1.51700902e+00 -1.29294083e-01 -3.54418486e-01 3.66023928e-01
-3.60119611e-01 2.88515389e-01 -8.34229767e-01 -8.19749653e-01
-3.12176466e-01 -9.82866764e-01 1.24067855e+00 4.99369025e-01
4.06058915e-02 -7.83543944e-01 -7.92722777e-02 -2.02124547e-02
8.13901246e-01 -1.14976771e-01 6.85419142e-01 -4.65010822e-01
-5.94647646e-01 -7.84602985e-02 -1.19391453e+00 3.49538028e-01
-1.38563097e-01 1.82054996e-01 -1.29492974e+00 7.06832856e-02
-2.64816344e-01 -1.87953293e-01 1.23427951e+00 7.79877305e-01
1.79680908e+00 -1.46681488e-01 -2.69203156e-01 7.61451364e-01
1.20498919e+00 -2.04402745e-01 1.07205606e+00 1.77662507e-01
8.75932395e-01 2.05488697e-01 2.77455777e-01 4.67024982e-01
5.94898164e-01 3.61891717e-01 4.67637330e-01 -2.35694334e-01
-2.16938809e-01 -4.67780232e-01 -1.74856126e-01 5.81170857e-01
-5.00652730e-01 -1.53103262e-01 -9.33789790e-01 3.48017931e-01
-1.88214076e+00 -1.15116370e+00 -3.39696616e-01 1.91837537e+00
5.46693802e-01 7.07542524e-02 1.78207397e-01 -3.50230820e-02
8.16984951e-01 3.39000337e-02 -3.71064425e-01 -5.52501604e-02
-2.27844760e-01 2.67138153e-01 1.19820833e+00 1.73186123e-01
-1.48998642e+00 1.17867088e+00 6.24392176e+00 1.29542625e+00
-8.18165839e-01 1.45599008e-01 1.08098066e+00 2.36824244e-01
2.10539028e-01 -3.59314084e-01 -1.16914117e+00 5.61960638e-01
5.29205203e-01 1.50241509e-01 4.25806612e-01 1.07795978e+00
-1.16091177e-01 -2.67392457e-01 -7.34480560e-01 9.93498504e-01
-1.37073651e-01 -1.42460346e+00 -5.32568060e-02 5.50172739e-02
8.94756734e-01 1.15429923e-01 2.36122176e-01 8.12594175e-01
6.65101290e-01 -1.31141317e+00 6.34558797e-01 5.48981845e-01
8.04056764e-01 -7.48890579e-01 9.07044172e-01 5.05305290e-01
-1.53285837e+00 5.31592108e-02 -1.16004598e+00 -9.14157405e-02
-1.45174384e-01 8.28316927e-01 -5.48340201e-01 -1.96475275e-02
8.18253756e-01 2.47810274e-01 -7.76819050e-01 1.07139862e+00
-4.45553847e-02 4.73625958e-01 -4.64450330e-01 -2.43528098e-01
4.10223603e-01 1.27696740e-02 4.07447387e-03 1.47314143e+00
3.28199238e-01 8.18643197e-02 1.34765759e-01 1.38446498e+00
-3.10295075e-01 -5.71606010e-02 -6.24128103e-01 1.21056944e-01
5.89776695e-01 1.45162511e+00 -1.35084009e+00 -4.71499354e-01
-3.01516175e-01 9.35187995e-01 5.78466713e-01 5.99037781e-02
-1.34904063e+00 -3.42020094e-01 3.75783950e-01 2.75333583e-01
3.90777171e-01 -2.32175708e-01 -5.78609884e-01 -9.45765376e-01
-1.33104458e-01 -4.62999940e-01 2.01763824e-01 -4.93973076e-01
-1.58137465e+00 2.61126667e-01 1.05228126e-01 -7.02778935e-01
2.88619161e-01 -7.55543768e-01 -7.95763493e-01 9.50221300e-01
-1.54047859e+00 -1.19838762e+00 -6.13063633e-01 4.86328006e-01
2.90388286e-01 4.59327456e-03 5.75924516e-01 5.97540736e-01
-1.95626527e-01 5.45251667e-01 -2.51181632e-01 5.38099051e-01
4.12958354e-01 -1.24533796e+00 7.23645449e-01 6.55549884e-01
1.68291762e-01 7.15603888e-01 1.45133793e-01 -6.81467175e-01
-8.71985078e-01 -1.45279384e+00 7.22748935e-01 -7.81613767e-01
6.32147074e-01 -4.17525798e-01 -8.83199632e-01 4.65471953e-01
-2.12955832e-01 5.73525667e-01 1.45313710e-01 -1.06205113e-01
-3.92759562e-01 3.09222758e-01 -1.39968956e+00 4.99197572e-01
1.19917488e+00 -4.16182786e-01 2.04967000e-02 2.55827367e-01
6.51010692e-01 -4.43745226e-01 -4.24364328e-01 6.06822789e-01
4.51937795e-01 -1.02887797e+00 1.21974468e+00 -3.19883019e-01
5.24948359e-01 -4.89305466e-01 -4.48496103e-01 -7.99718916e-01
-8.51905227e-01 1.72864854e-01 -3.00681114e-01 1.22308004e+00
2.33414397e-02 -4.38074797e-01 9.97919142e-01 3.44784051e-01
1.86749682e-01 -6.53343379e-01 -9.88451242e-01 -7.95138657e-01
4.57979321e-01 -3.18436146e-01 8.18387628e-01 8.77532780e-01
-4.11639184e-01 6.68664500e-02 -4.67759408e-02 2.86094457e-01
5.06874084e-01 -5.36726296e-01 8.02947879e-01 -1.19083655e+00
1.08507134e-01 -7.74157524e-01 -6.41340077e-01 -1.43373322e+00
6.40495569e-02 -7.44928598e-01 -4.99677099e-02 -1.42581964e+00
6.63819849e-01 -4.64897543e-01 -1.46342605e-01 3.18647385e-01
-4.97099549e-01 6.25010788e-01 1.35276616e-01 9.71502811e-02
-7.86683500e-01 3.60553235e-01 1.57984006e+00 -3.21739465e-01
3.09862918e-03 1.07169263e-01 -7.97276974e-01 1.02782869e+00
9.39068913e-01 -3.71029109e-01 -2.79058188e-01 -5.41843951e-01
1.31329492e-01 -7.47199297e-01 7.51206338e-01 -1.20159423e+00
4.28641379e-01 -1.99366719e-01 8.68826985e-01 -8.65051031e-01
8.54272246e-02 -6.21549785e-01 -5.17334223e-01 3.08047861e-01
3.57232033e-03 -8.12476575e-02 3.56738985e-01 3.94204617e-01
8.65882486e-02 -2.74926841e-01 9.81234431e-01 -3.69197547e-01
-6.62782252e-01 4.61837053e-01 1.92121714e-02 1.67180523e-01
8.73884082e-01 -3.05777490e-01 -8.23032260e-01 -2.70245746e-02
-1.67435080e-01 -6.73717037e-02 1.44461155e-01 1.39216155e-01
3.82273853e-01 -1.50529575e+00 -7.50039995e-01 1.06537730e-01
-4.02142331e-02 1.43640190e-01 8.51182565e-02 8.28734517e-01
-7.03949153e-01 5.90176105e-01 -4.16449159e-02 -8.13445807e-01
-7.08613813e-01 4.19881880e-01 4.11040276e-01 -5.19895017e-01
-4.72045988e-01 1.10278690e+00 3.14171702e-01 -6.31325722e-01
3.56580794e-01 -8.72120559e-01 -1.60959333e-01 -2.21885175e-01
8.73997331e-01 5.66893578e-01 2.80599296e-02 -4.68924791e-01
-4.68319505e-01 2.75281072e-01 -4.23134342e-02 2.83952385e-01
1.13113642e+00 2.30045766e-01 -5.01743816e-02 1.89122513e-01
9.61699188e-01 -1.74509630e-01 -1.28762960e+00 -2.46141106e-01
-1.21779285e-01 -8.25888991e-01 1.04155250e-01 -4.97540653e-01
-1.13697886e+00 8.51521194e-01 5.55835664e-01 3.21995795e-01
8.35193455e-01 -1.04438014e-01 5.05980790e-01 4.74670351e-01
5.19212365e-01 -1.14899373e+00 1.17320970e-01 9.43242550e-01
4.97503251e-01 -1.44145226e+00 8.89496431e-02 -3.76469165e-01
-3.87389809e-01 8.98451209e-01 1.11370707e+00 -3.87410909e-01
5.88191688e-01 7.22046614e-01 -4.25132692e-01 -2.92065501e-01
-9.27908123e-02 -3.25115681e-01 1.93414643e-01 7.75581539e-01
6.73022270e-01 2.30990842e-01 1.20249450e-01 5.23750842e-01
-2.26295263e-01 1.01329498e-01 2.17247665e-01 7.87318230e-01
-7.21929193e-01 -4.04886931e-01 -6.97870433e-01 7.61614919e-01
-4.61916268e-01 -3.46421510e-01 -2.15304941e-01 8.56165290e-01
4.23794806e-01 5.24684131e-01 4.36448485e-01 -7.75015727e-02
5.17985106e-01 -2.89892018e-01 7.92415977e-01 -4.07516837e-01
-3.79093945e-01 -4.25087839e-01 -4.53862727e-01 -3.96767765e-01
-3.06951374e-01 -2.47138038e-01 -1.23643696e+00 -1.08755577e+00
-6.12321854e-01 -5.44751585e-01 5.13253272e-01 5.40834367e-01
-1.05978005e-01 8.31335604e-01 4.14194241e-02 -9.16890919e-01
-7.15391576e-01 -1.23651373e+00 -6.54286802e-01 2.14553714e-01
-1.76600665e-01 -8.09201062e-01 -2.28367597e-01 -1.74394757e-01] | [8.72249698638916, -0.03189729526638985] |
47545e04-652c-41d8-84ca-709294026dba | adaptive-direction-guided-structure-tensor | 2001.05717 | null | https://arxiv.org/abs/2001.05717v1 | https://arxiv.org/pdf/2001.05717v1.pdf | Adaptive Direction-Guided Structure Tensor Total Variation | Direction-guided structure tensor total variation (DSTV) is a recently proposed regularization term that aims at increasing the sensitivity of the structure tensor total variation (STV) to the changes towards a predetermined direction. Despite of the plausible results obtained on the uni-directional images, the DSTV model is not applicable to the multi-directional images of real-world. In this study, we build a two-stage framework that brings adaptivity to DSTV. We design an alternative to STV, which encodes the first-order information within a local neighborhood under the guidance of spatially varying directional descriptors (i.e., orientation and the dose of anisotropy). In order to estimate those descriptors, we propose an efficient preprocessor that captures the local geometry based on the structure tensor. Through the extensive experiments, we demonstrate how beneficial the involvement of the directional information in STV is, by comparing the proposed method with the state-of-the-art analysis-based denoising models, both in terms of restoration quality and computational efficiency. | ['Mustafa E. Kamasak', 'Ezgi Demircan-Tureyen'] | 2020-01-16 | null | null | null | null | ['art-analysis'] | ['computer-vision'] | [ 1.08062215e-01 -4.36071217e-01 3.38514954e-01 -2.75671899e-01
-4.46060181e-01 -3.21808815e-01 4.68424886e-01 -7.07754046e-02
-2.01789036e-01 4.09943938e-01 4.93286639e-01 1.12997182e-01
-7.31803834e-01 -8.61550391e-01 -4.83951181e-01 -1.27341986e+00
-1.37440145e-01 -1.41470045e-01 3.16433787e-01 -3.76585603e-01
4.93413031e-01 6.35780811e-01 -1.21008229e+00 1.25304580e-01
9.61592436e-01 9.28120434e-01 1.80697367e-01 2.79943049e-01
7.00856522e-02 4.16773587e-01 -8.08207467e-02 -1.65786415e-01
3.16166133e-01 -3.12139302e-01 -4.88264829e-01 3.20148021e-01
4.17564273e-01 -2.78242409e-01 -3.05128902e-01 1.12407458e+00
5.40788114e-01 1.65676445e-01 6.77036703e-01 -5.95374644e-01
-7.21604586e-01 5.73434681e-02 -7.15811968e-01 3.72732282e-01
2.12881282e-01 1.57345280e-01 9.29247320e-01 -1.09022355e+00
9.37996745e-01 1.21463156e+00 6.28803015e-01 9.28529650e-02
-1.28116310e+00 -1.25829831e-01 1.64284959e-01 3.11859727e-01
-1.16522205e+00 -2.85768420e-01 1.62783992e+00 -6.42387271e-01
3.11759889e-01 3.13991100e-01 4.23630714e-01 7.80953288e-01
5.35477698e-01 4.61988598e-01 1.28859961e+00 -4.86416399e-01
2.82911360e-01 -3.92030925e-01 4.40262377e-01 5.53979993e-01
1.53302282e-01 3.21638197e-01 -5.22454679e-01 -2.86825746e-01
8.47974539e-01 -2.40340650e-01 -5.58018565e-01 -7.45569348e-01
-1.27713025e+00 7.26274550e-01 5.22405386e-01 5.17734826e-01
-7.22391665e-01 -4.29916382e-02 4.78439718e-01 5.99417016e-02
7.56694019e-01 2.98636854e-01 -1.92888111e-01 1.56974778e-01
-7.22054958e-01 1.20868735e-01 1.62765160e-01 3.56625110e-01
8.93898845e-01 2.08663315e-01 -3.16389173e-01 1.00234842e+00
5.48443973e-01 4.71308082e-01 1.93461105e-01 -9.97672498e-01
2.47159645e-01 5.30213773e-01 -5.21030091e-02 -1.65525448e+00
-4.51724499e-01 -7.86997795e-01 -1.21390545e+00 1.74037918e-01
4.10009325e-01 2.22939014e-01 -6.90018892e-01 1.62511218e+00
5.85502148e-01 1.00017950e-01 -1.93940595e-01 1.11579740e+00
6.32153094e-01 5.93080878e-01 -3.03145498e-01 -4.96100873e-01
1.05759239e+00 -7.15144098e-01 -9.44512188e-01 1.50075808e-01
4.20221806e-01 -1.03176153e+00 8.90234649e-01 4.53376561e-01
-8.89229357e-01 -5.01975000e-01 -9.06847239e-01 -1.12647824e-02
1.27555326e-01 1.52597472e-01 3.67401987e-01 4.60389435e-01
-1.09522605e+00 6.37946725e-01 -9.52307582e-01 -1.90721780e-01
-1.01762176e-01 -8.13415721e-02 -3.48677635e-01 -2.16382936e-01
-7.91631699e-01 5.32763898e-01 -2.19805300e-01 6.80869520e-01
-6.61807716e-01 -7.26818979e-01 -4.95542169e-01 -2.37738475e-01
1.65614009e-01 -4.86516207e-01 2.82386154e-01 -6.04357839e-01
-1.51292312e+00 4.66677755e-01 -2.70359069e-01 1.79042026e-01
5.44461131e-01 -6.64084852e-02 -8.21602121e-02 4.65334356e-01
1.31396487e-01 -8.16941932e-02 9.62452292e-01 -1.65862060e+00
1.23642065e-01 -6.53366685e-01 -4.97594588e-02 5.54660195e-03
-3.90385270e-01 -1.71933770e-01 -3.15253526e-01 -1.04741156e+00
7.98670888e-01 -9.49111879e-01 -3.72310519e-01 2.75313944e-01
-4.75199103e-01 1.55482277e-01 7.22741544e-01 -1.01244414e+00
1.20248353e+00 -2.28580666e+00 5.84886611e-01 5.03721893e-01
2.14305967e-01 7.75504932e-02 -8.74672383e-02 5.12485802e-01
4.67556939e-02 -9.82521251e-02 -6.59995019e-01 -4.10522744e-02
-3.15283000e-01 3.23110878e-01 4.12864145e-03 7.42591977e-01
-1.25468895e-01 2.82041818e-01 -7.77347684e-01 -3.80787969e-01
3.50804448e-01 7.72441506e-01 -6.66799247e-01 1.63473681e-01
2.35075906e-01 8.16563964e-01 -7.95516074e-01 5.16172886e-01
1.20891309e+00 1.95893243e-01 7.89982155e-02 -8.43576550e-01
-4.41534787e-01 -2.06797570e-01 -1.25466323e+00 1.44000554e+00
-3.22286695e-01 2.51288712e-01 4.60598260e-01 -1.12178576e+00
1.03909743e+00 2.38309413e-01 6.94170415e-01 -6.98196411e-01
-6.27299119e-03 3.77914965e-01 -2.57411804e-02 -5.59704006e-01
1.79060206e-01 -6.99784383e-02 3.72649759e-01 1.34849995e-01
-1.37654364e-01 1.45873781e-02 2.71229446e-01 3.42246778e-02
9.09985721e-01 4.97711636e-02 1.83270313e-02 -7.37644970e-01
9.37828004e-01 -2.59498656e-01 6.89135551e-01 3.96126866e-01
-2.61248231e-01 6.86686635e-01 3.32598895e-01 -4.54210609e-01
-1.07714319e+00 -9.59320188e-01 -5.10831714e-01 4.78040040e-01
2.62870103e-01 -2.55931318e-01 -6.81176901e-01 -2.28641748e-01
-1.75753877e-01 3.29069287e-01 -5.54258406e-01 -1.14215307e-01
-7.86924124e-01 -1.16466546e+00 -4.45768088e-02 1.45351395e-01
8.41570258e-01 -5.70064187e-01 -1.67707533e-01 2.20470160e-01
-2.34831616e-01 -9.63039577e-01 -4.25500751e-01 -2.28187144e-01
-1.25860667e+00 -1.00887203e+00 -7.82280326e-01 -6.07614636e-01
8.05698812e-01 3.84228796e-01 7.75363624e-01 1.72924116e-01
-9.54664126e-02 5.85533082e-01 -4.78302211e-01 3.85396212e-01
-1.72690749e-01 -2.98471272e-01 -2.18372837e-01 7.17981637e-01
-2.35622659e-01 -8.47352505e-01 -8.94724727e-01 5.29056191e-01
-1.06742144e+00 -8.93754885e-02 5.21375358e-01 8.61010611e-01
8.94890368e-01 9.20527950e-02 1.80222884e-01 -5.76753557e-01
7.44267285e-01 -2.05387279e-01 -3.70150834e-01 1.98135316e-01
-6.24276340e-01 1.45418048e-01 5.16851723e-01 -2.60169268e-01
-1.22047436e+00 1.90502424e-02 -1.69068590e-01 -1.05641857e-01
1.98177204e-01 6.65294766e-01 -1.03369243e-01 -6.18373454e-01
4.09350246e-01 4.94622856e-01 3.27908285e-02 -8.80297482e-01
2.17054516e-01 1.67106405e-01 2.76054829e-01 -7.07770705e-01
9.55406964e-01 8.27942312e-01 4.53016192e-01 -1.14468884e+00
-4.91869181e-01 -5.54383099e-01 -8.20677340e-01 -4.13780421e-01
8.41543555e-01 -5.45182288e-01 -4.66750294e-01 9.42832768e-01
-1.18651545e+00 3.62454504e-02 1.70193072e-02 5.41469097e-01
-3.55234474e-01 9.25123811e-01 -6.55380189e-01 -5.53728402e-01
-3.15055043e-01 -1.30011547e+00 9.46866393e-01 -2.43617997e-01
3.88064653e-01 -1.14664233e+00 1.74322516e-01 3.49797219e-01
6.07146323e-01 3.99122179e-01 1.11137366e+00 1.22101448e-01
-7.31377721e-01 -1.08718025e-02 -1.04698099e-01 7.02970982e-01
1.31789237e-01 1.37390196e-01 -5.15269578e-01 -2.52624869e-01
5.01917839e-01 2.98706353e-01 7.46486843e-01 7.32008874e-01
1.04276145e+00 -1.66904181e-01 1.00045308e-01 7.91339517e-01
1.66204751e+00 1.50294844e-02 7.56337285e-01 6.28501773e-01
7.47718573e-01 6.38889313e-01 5.14147460e-01 4.97618765e-01
1.91061810e-01 8.51485848e-01 5.93785584e-01 -1.76874474e-01
-3.38216245e-01 1.46276832e-01 1.00493945e-01 1.30731380e+00
-6.81139648e-01 4.22216132e-02 -7.04101145e-01 5.78907430e-01
-1.72527969e+00 -7.95657218e-01 -6.55574381e-01 2.10408854e+00
6.77249730e-01 -1.47541746e-01 -4.65905964e-01 2.48077080e-01
6.62797451e-01 6.26094103e-01 -2.78432757e-01 -2.93737203e-01
-3.53058875e-01 -2.10665435e-01 3.00848782e-01 7.76225746e-01
-8.81804228e-01 3.87064159e-01 6.47863817e+00 8.21492851e-01
-1.21485496e+00 -6.28666133e-02 4.41731036e-01 4.02312964e-01
-5.12925386e-01 -1.49969254e-02 -2.51942247e-01 4.75696683e-01
4.03097659e-01 1.92717448e-01 5.43765187e-01 4.79622602e-01
6.83046162e-01 7.05414638e-02 -4.68553960e-01 7.91926682e-01
5.69857359e-02 -1.07797945e+00 2.80711830e-01 9.01508406e-02
6.63543701e-01 -2.69042104e-01 -2.32959930e-02 -3.52800578e-01
-3.92631918e-01 -4.26544547e-01 6.77970171e-01 8.49702537e-01
2.37143517e-01 -5.04915178e-01 8.61911535e-01 8.26841444e-02
-1.15230954e+00 4.60131057e-02 -4.23537254e-01 1.80956170e-01
4.30408537e-01 1.23526728e+00 -1.78276867e-01 7.89371789e-01
6.94915831e-01 8.30351830e-01 -5.30357897e-01 9.11992073e-01
-3.42832506e-01 7.39433885e-01 -2.25882649e-01 4.25079167e-01
2.70979106e-01 -9.99940574e-01 1.03597260e+00 7.68949449e-01
3.86445522e-01 2.21517757e-01 3.28476951e-02 6.50709391e-01
3.61154139e-01 2.66254842e-01 -2.79010475e-01 4.07644689e-01
8.92744809e-02 1.27710605e+00 -5.19568145e-01 1.78160071e-02
-2.53262967e-01 8.04254651e-01 -5.02737872e-02 7.74192214e-01
-3.58554751e-01 -2.96522081e-02 4.50665206e-01 3.69092882e-01
3.56972873e-01 -6.06439412e-01 -5.38697302e-01 -1.16358864e+00
6.12601697e-01 -7.23517895e-01 6.36488572e-02 -8.39794636e-01
-1.45176351e+00 6.97406650e-01 1.09333294e-02 -1.27924037e+00
9.07566249e-02 -6.82841241e-01 -5.44280648e-01 6.84586346e-01
-1.59479451e+00 -1.22931707e+00 -4.39276367e-01 6.21710062e-01
2.98903137e-01 1.99120566e-01 4.09778953e-01 4.59608734e-01
-5.40192902e-01 1.03685677e-01 5.85425258e-01 -1.00910537e-01
6.06836617e-01 -1.00353134e+00 -2.20332354e-01 1.05907857e+00
-4.10426408e-01 7.78692245e-01 1.00518012e+00 -6.39932275e-01
-1.60592926e+00 -6.91147804e-01 6.22109711e-01 2.88614947e-02
6.92113400e-01 -8.27008579e-03 -9.73815620e-01 4.26889956e-01
-7.41930157e-02 3.11786354e-01 2.76066303e-01 -2.20365971e-02
-2.65783131e-01 -4.05251682e-01 -1.24431062e+00 5.04702866e-01
1.04989362e+00 -4.66842860e-01 -3.73530716e-01 2.33123064e-01
3.91579926e-01 -1.02077775e-01 -1.13286316e+00 5.99200845e-01
3.70794475e-01 -1.06699145e+00 1.26462615e+00 -5.81685752e-02
5.02397299e-01 -6.21921957e-01 -4.61286366e-01 -1.33874333e+00
-7.07919776e-01 -5.32847226e-01 1.29962280e-01 1.22929847e+00
5.14982231e-02 -7.45678246e-01 3.50951344e-01 4.10722345e-01
-4.62575287e-01 -7.93813050e-01 -1.28296435e+00 -6.94306016e-01
-8.46940000e-03 -2.33548805e-01 2.16782257e-01 8.59648764e-01
-5.23033559e-01 -4.01905105e-02 -5.09818733e-01 4.63788807e-01
9.68444705e-01 6.76833466e-02 4.34818208e-01 -1.03274345e+00
-4.86549810e-02 -8.75387788e-02 -6.53523147e-01 -1.07595026e+00
-2.80982435e-01 -6.03166521e-01 -7.42550427e-03 -1.58441424e+00
1.24248236e-01 -4.13483620e-01 -3.43838334e-01 1.20064840e-02
-8.05944055e-02 2.73519635e-01 -2.96287220e-02 4.97661144e-01
-3.07541844e-02 9.29729998e-01 1.75623643e+00 -9.13432166e-02
-1.11234821e-01 -2.91701674e-01 -3.92498702e-01 7.00547636e-01
4.24929231e-01 -2.52485096e-01 -3.32552195e-01 -7.30794072e-01
1.02836482e-01 -9.40315574e-02 4.86923218e-01 -8.21652889e-01
5.90688959e-02 -1.77369803e-01 4.21230085e-02 -5.46296120e-01
2.74839133e-01 -7.77611554e-01 8.35527331e-02 1.51651040e-01
-1.13311760e-01 -8.40132535e-02 -1.39378324e-01 6.87868893e-01
-4.97290939e-01 -2.11111620e-01 9.05497432e-01 8.61795396e-02
-5.65545857e-01 1.86728999e-01 -3.30700427e-01 -2.56883651e-01
6.17895126e-01 -3.09523433e-01 -3.40601146e-01 -1.45103753e-01
-6.47877395e-01 -2.41199166e-01 4.75317061e-01 -4.18527462e-02
6.71194732e-01 -1.33554995e+00 -7.63761818e-01 2.86793649e-01
-7.18722641e-02 -1.53545439e-01 7.09859967e-01 1.21385765e+00
-7.02161789e-01 5.52893505e-02 -2.40614682e-01 -8.00550401e-01
-1.14456546e+00 4.47812498e-01 3.79369497e-01 -3.20499718e-01
-8.13140869e-01 5.50719142e-01 3.57215106e-01 -5.12927413e-01
-1.67896181e-01 -3.45466167e-01 -2.83716589e-01 -1.10892080e-01
3.37530613e-01 6.80404007e-01 2.33565003e-01 -9.32372272e-01
-3.30836982e-01 1.33829772e+00 2.45888457e-01 -3.27859595e-02
1.62522221e+00 -4.66256112e-01 -6.44144714e-01 1.18855670e-01
1.28635240e+00 2.45156854e-01 -1.36533487e+00 -1.08955204e-01
-1.48379669e-01 -7.99315989e-01 7.06647456e-01 -5.72132111e-01
-1.49896002e+00 6.43950522e-01 8.54026079e-01 1.51662961e-01
1.34834886e+00 -3.02525103e-01 8.11466634e-01 6.62441179e-02
3.98004293e-01 -8.47417176e-01 2.33734757e-01 2.24589080e-01
1.33978856e+00 -9.29734290e-01 3.53269398e-01 -9.58870053e-01
-2.89593518e-01 1.25890088e+00 1.71534628e-01 -4.10700887e-01
9.10180032e-01 -1.17965117e-01 1.80175453e-01 -4.91724014e-01
-1.85406953e-01 2.06488535e-01 5.55956125e-01 5.02062917e-01
4.91310507e-01 -1.89719528e-01 -9.65488136e-01 2.00471450e-02
1.90188736e-01 -2.50778437e-01 4.91020322e-01 8.30142975e-01
-3.08794379e-01 -1.15857697e+00 -7.02389956e-01 1.13054648e-01
-3.86233926e-01 5.98476417e-02 -6.15784861e-02 5.40041685e-01
1.11760318e-01 1.04069865e+00 -2.58615613e-01 -2.27589965e-01
5.06319582e-01 -4.91689295e-01 3.12746793e-01 1.58351138e-02
-2.65167177e-01 3.40733647e-01 1.41605586e-01 -6.64328277e-01
-8.63482475e-01 -6.96896195e-01 -8.68681669e-01 -1.42657384e-01
-3.24655741e-01 -2.84043550e-02 8.49898934e-01 8.81978750e-01
2.81936616e-01 3.95322442e-01 9.49429214e-01 -9.42380607e-01
-4.01718855e-01 -8.51643682e-01 -7.96767414e-01 7.10687876e-01
3.52418989e-01 -8.53411973e-01 -8.61690998e-01 -1.08748944e-02] | [11.41195297241211, -2.438284158706665] |
1975f26b-e6f0-4dd8-ab68-28cb910996ce | how-should-agents-ask-questions-for-situated | 2106.06504 | null | https://arxiv.org/abs/2106.06504v1 | https://arxiv.org/pdf/2106.06504v1.pdf | How Should Agents Ask Questions For Situated Learning? An Annotated Dialogue Corpus | Intelligent agents that are confronted with novel concepts in situated environments will need to ask their human teammates questions to learn about the physical world. To better understand this problem, we need data about asking questions in situated task-based interactions. To this end, we present the Human-Robot Dialogue Learning (HuRDL) Corpus - a novel dialogue corpus collected in an online interactive virtual environment in which human participants play the role of a robot performing a collaborative tool-organization task. We describe the corpus data and a corresponding annotation scheme to offer insight into the form and content of questions that humans ask to facilitate learning in a situated environment. We provide the corpus as an empirically-grounded resource for improving question generation in situated intelligent agents. | ['Matthew Marge', 'Matthias Scheutz', 'Gordon Briggs', 'Antonio Roque', 'Felix Gervits'] | 2021-06-11 | null | https://aclanthology.org/2021.sigdial-1.37 | https://aclanthology.org/2021.sigdial-1.37.pdf | sigdial-acl-2021-7 | ['novel-concepts'] | ['reasoning'] | [ 3.30982506e-01 1.01325643e+00 4.86710638e-01 -4.66480166e-01
-6.36827826e-01 -9.36613977e-01 9.97799695e-01 1.92144111e-01
-4.27968770e-01 9.91391659e-01 5.85654318e-01 -5.14393210e-01
-1.15535609e-01 -5.07898629e-01 -1.83571696e-01 -1.84479684e-01
-1.59210384e-01 1.02721167e+00 1.91763252e-01 -8.38035524e-01
4.16708380e-01 2.32547864e-01 -1.31909680e+00 5.00590503e-01
8.57356787e-01 8.25937912e-02 1.02423382e+00 8.52847815e-01
-3.27433199e-01 1.54344869e+00 -7.29184330e-01 -1.45927355e-01
-3.02628819e-02 -8.43018532e-01 -1.92964804e+00 2.29055807e-01
-1.92283005e-01 -6.38803899e-01 -2.27803856e-01 5.15936315e-01
2.53486693e-01 8.40383172e-01 6.01681650e-01 -1.34713638e+00
-2.57908940e-01 8.37087095e-01 4.89937246e-01 -7.23036155e-02
1.31281781e+00 2.78229117e-01 1.06753004e+00 -5.92740715e-01
1.27494419e+00 1.51288366e+00 4.40389723e-01 8.89401793e-01
-6.93580568e-01 2.61534393e-01 -5.40884361e-02 7.01289535e-01
-7.46820211e-01 -4.76249129e-01 7.58242011e-01 -4.45952952e-01
1.30285907e+00 7.92528987e-02 4.31699753e-01 1.11215580e+00
-7.54842535e-02 8.81775320e-01 5.62621117e-01 -6.96374238e-01
2.25969434e-01 5.40237725e-01 -2.54916400e-02 6.54979289e-01
-3.64327043e-01 -1.63300184e-03 -6.51268899e-01 -2.85648167e-01
1.03642392e+00 -6.78965211e-01 -6.56573623e-02 -6.91548288e-01
-1.37892604e+00 1.00264335e+00 1.56533688e-01 6.90258443e-01
-3.37479353e-01 -1.58004791e-01 7.20177770e-01 7.69843459e-01
-1.15431353e-01 1.30523193e+00 -7.76083887e-01 -8.07595909e-01
6.14595234e-01 5.33670008e-01 1.51163733e+00 1.07940733e+00
6.02304339e-01 -3.58129263e-01 2.30728716e-01 9.51648533e-01
3.87558192e-01 -4.01128605e-02 3.78071964e-01 -1.71497071e+00
4.87242103e-01 6.24233544e-01 6.03419483e-01 -7.20483363e-01
-6.47756517e-01 5.04864931e-01 6.21754006e-02 2.79266089e-01
5.92882514e-01 -4.77976859e-01 1.66732684e-01 1.34382832e+00
8.62819672e-01 -5.91839552e-01 8.18146765e-01 6.56225979e-01
1.22833860e+00 6.48944080e-01 -1.55554876e-01 -2.92951852e-01
1.55235672e+00 -1.32248104e+00 -8.97966266e-01 -2.63666242e-01
1.31356776e+00 -4.22440350e-01 1.27929652e+00 2.29544967e-01
-7.85020709e-01 -5.86416423e-01 -6.77284837e-01 -5.17919242e-01
-3.46810430e-01 -3.40121210e-01 7.50965059e-01 2.80769765e-01
-8.24915707e-01 2.75562942e-01 -3.00637513e-01 -1.10299563e+00
-3.74695718e-01 -4.88120504e-02 -4.72647905e-01 1.86799139e-01
-1.14567447e+00 1.44084179e+00 5.69590747e-01 -1.37854204e-01
-8.75057101e-01 5.93417101e-02 -1.39351940e+00 -3.73082578e-01
8.96432459e-01 -4.08242404e-01 2.13307667e+00 -5.28583944e-01
-1.91543579e+00 7.67017603e-01 2.20428050e-01 -2.24259198e-01
3.19403708e-01 -1.52161896e-01 2.00275794e-01 3.75047684e-01
4.84732598e-01 5.35949111e-01 -2.16914490e-02 -1.51229513e+00
-1.02735126e+00 -3.21926653e-01 7.61520684e-01 8.57642591e-01
3.32377881e-01 1.41764581e-01 1.62802085e-01 -5.85192069e-02
7.78035246e-05 -9.37289298e-01 -3.63636494e-01 -4.11407530e-01
-2.65001971e-02 -7.68456399e-01 6.76066995e-01 -5.58776736e-01
9.23615247e-02 -1.92866695e+00 2.93541998e-01 -2.66617715e-01
3.23891401e-01 2.73943562e-02 -2.46610269e-01 1.09887862e+00
3.65467966e-01 -2.12068260e-01 7.78921321e-02 6.12471849e-02
3.08840334e-01 4.50001717e-01 6.57686666e-02 -1.31421283e-01
-1.49183616e-01 7.39160836e-01 -1.42009938e+00 -6.49583638e-01
4.82974917e-01 -9.20850039e-02 -3.85301858e-01 8.09269011e-01
-4.86791998e-01 8.40405405e-01 -7.23169982e-01 5.57926968e-02
-2.37482443e-01 4.97274920e-02 4.18044060e-01 6.35373592e-01
-1.66012153e-01 6.14910901e-01 -7.29170680e-01 1.85756207e+00
-6.60202622e-01 9.88536060e-01 4.00168121e-01 -4.95131105e-01
1.01256704e+00 6.00600481e-01 1.02916189e-01 -5.56603789e-01
2.45857090e-01 -8.56297687e-02 1.41201675e-01 -1.25199807e+00
8.22641075e-01 -1.40258163e-01 -4.26613837e-01 8.30156684e-01
8.49225521e-02 -7.43724406e-01 1.19037263e-01 3.50137711e-01
1.36449838e+00 8.49146619e-02 6.61569357e-01 4.74300832e-02
5.03477573e-01 5.91621995e-01 1.42198727e-01 1.00296056e+00
-6.13406003e-01 -2.84141809e-01 5.26135325e-01 -5.77434421e-01
-9.23590720e-01 -7.43884265e-01 2.23479077e-01 1.52684712e+00
1.71820238e-01 -3.71667415e-01 -6.69868886e-01 -6.06734633e-01
-7.18976915e-01 1.02416468e+00 -4.89540815e-01 2.20782772e-01
-5.25602341e-01 4.01281029e-01 4.44885939e-01 2.19420373e-01
7.35589564e-01 -1.86744869e+00 -1.39314580e+00 4.49206442e-01
-6.51221097e-01 -1.27571082e+00 2.06930935e-03 2.84197450e-01
-5.00252128e-01 -1.32942700e+00 -2.46496752e-01 -1.32369983e+00
3.89854074e-01 2.05279782e-01 9.94750798e-01 3.59173715e-02
5.92002235e-02 1.16522443e+00 -1.19591975e+00 -4.88957942e-01
-9.79646206e-01 1.48358822e-01 -1.37068853e-01 -6.86374485e-01
1.01041719e-01 -3.90059531e-01 1.76789239e-01 5.40122390e-01
-5.53082526e-01 2.96559960e-01 5.98711036e-02 9.58153546e-01
-5.65860689e-01 -9.22901556e-03 5.26931882e-01 -7.35555649e-01
1.37138653e+00 -4.34473306e-01 -1.84447721e-01 4.33114111e-01
1.52205080e-01 6.67770719e-03 3.00764084e-01 -2.97692329e-01
-1.72172034e+00 1.79515287e-01 1.59074254e-02 7.80002952e-01
-7.15987980e-01 5.60923100e-01 -3.35705161e-01 1.55685246e-01
1.04512584e+00 8.97254888e-03 -4.96634319e-02 -2.55456418e-01
6.28415048e-01 7.83521712e-01 7.58740306e-01 -1.06788623e+00
4.79178071e-01 -1.88611746e-01 -4.31949168e-01 -1.21606970e+00
-2.75631040e-01 -5.76995850e-01 -9.51052725e-01 -5.14659166e-01
8.66224945e-01 -4.75646496e-01 -1.13902748e+00 2.52842486e-01
-1.54525375e+00 -1.29605222e+00 -4.43726629e-01 2.52767384e-01
-1.18415260e+00 3.12036604e-01 -7.41369426e-01 -1.22027385e+00
8.49090517e-02 -9.88737345e-01 7.19482064e-01 2.17279077e-01
-7.80313909e-01 -1.16546667e+00 3.56128275e-01 7.39310145e-01
6.05291240e-02 2.05616012e-01 9.84441638e-01 -1.24763203e+00
-7.73249492e-02 3.79612446e-01 -6.00883365e-03 -3.09047997e-01
2.00133786e-01 -6.73941731e-01 -5.40726304e-01 2.34869152e-01
3.89954448e-01 -9.99424934e-01 -3.45102429e-01 -4.82003719e-01
1.11360095e-01 -4.10858572e-01 -8.95959884e-02 -5.17687857e-01
5.27677298e-01 7.69169390e-01 4.93313938e-01 5.35349727e-01
2.60880560e-01 1.52600622e+00 1.07641459e+00 5.70144355e-01
1.06404829e+00 6.65043473e-01 3.58842462e-02 2.99901187e-01
4.42943543e-01 -2.71920383e-01 3.08622420e-01 5.15461624e-01
6.83132112e-02 -1.22874483e-01 -1.32058108e+00 6.85852706e-01
-2.15125179e+00 -8.90367031e-01 -3.43034357e-01 1.21039367e+00
9.98307109e-01 -4.32713866e-01 1.81653395e-01 -6.12281561e-02
5.70163846e-01 -6.52353931e-03 -3.67901951e-01 -5.58310032e-01
3.01483065e-01 -4.27356511e-01 -5.11899173e-01 8.86413693e-01
-6.23671114e-01 1.39567780e+00 5.78647804e+00 3.62524837e-02
-3.71677242e-02 1.76770672e-01 5.81722520e-02 5.48456609e-01
1.43049300e-01 -9.54875350e-03 -1.53993979e-01 -4.27064896e-01
7.28248179e-01 -1.93274364e-01 6.00385725e-01 9.93481815e-01
3.29518735e-01 -7.94837713e-01 -1.58120584e+00 9.00772095e-01
1.04046769e-01 -8.56125474e-01 -3.88026029e-01 3.12072206e-02
2.78154343e-01 -3.09358388e-01 -3.76410455e-01 6.00135624e-01
8.70744228e-01 -8.52390528e-01 4.20391530e-01 3.46757233e-01
1.93619862e-01 -5.17441750e-01 8.77775669e-01 9.02967572e-01
-9.49456871e-01 -2.25761563e-01 -8.68226215e-02 -7.39247322e-01
4.37668681e-01 -6.32551014e-01 -1.84509754e+00 1.90843046e-01
5.38750768e-01 9.16546285e-02 -3.43994409e-01 5.47636569e-01
-3.60163927e-01 -9.88707766e-02 -3.01727988e-02 -5.96914709e-01
2.42639244e-01 -2.04510152e-01 6.03005528e-01 6.24789059e-01
-1.41158178e-01 9.45239842e-01 3.56190443e-01 6.48446918e-01
2.86358565e-01 1.17316619e-01 -1.23364127e+00 5.37681319e-02
5.25417924e-01 1.13745368e+00 -6.16409957e-01 -2.54087061e-01
-1.83664739e-01 1.13753688e+00 2.95769304e-01 1.94364801e-01
-1.86462149e-01 -7.01966286e-01 3.27706277e-01 -3.25690180e-01
-3.99097860e-01 -5.76076150e-01 1.31025268e-02 -6.28529787e-01
-2.38716885e-01 -1.30763865e+00 5.11212312e-02 -1.35884726e+00
-1.14035165e+00 6.64454222e-01 1.86912581e-01 -7.12775588e-01
-1.08187473e+00 -5.58122993e-01 -5.34015656e-01 3.12937617e-01
-5.83236158e-01 -9.35367584e-01 -6.19635224e-01 5.36100388e-01
1.09161067e+00 -4.56864744e-01 1.03904212e+00 -5.67195237e-01
2.34745368e-02 -1.03273816e-01 -2.92711496e-01 2.17607856e-01
6.32993937e-01 -1.09952843e+00 5.78071594e-01 -1.26057684e-01
4.39883098e-02 4.94683146e-01 9.47610974e-01 -7.86635339e-01
-1.35612643e+00 -4.51821923e-01 9.30673718e-01 -1.09786010e+00
6.76348209e-01 -2.47931242e-01 -9.15605426e-01 9.42827284e-01
6.40038371e-01 -7.98599064e-01 8.38083565e-01 2.16813996e-01
2.13208705e-01 7.59442925e-01 -1.43748784e+00 8.33601058e-01
1.29878473e+00 -9.01093781e-01 -1.66221690e+00 5.38832664e-01
9.74706769e-01 -4.56382602e-01 -6.68035507e-01 5.25767058e-02
2.82363385e-01 -8.08212817e-01 5.36057234e-01 -7.13375866e-01
3.09279382e-01 -2.05941826e-01 -2.13123038e-01 -1.49751842e+00
7.46060312e-02 -8.84184957e-01 5.60832143e-01 1.05266941e+00
4.95645314e-01 -6.51202261e-01 4.44922417e-01 1.20020151e+00
-1.83746904e-01 -1.65673837e-01 -7.57085979e-01 -3.98352891e-01
-6.48844813e-04 -4.90154654e-01 4.91094977e-01 1.06598985e+00
1.36397052e+00 1.04263115e+00 5.78328073e-02 -1.42557785e-01
1.53278679e-01 -7.24918067e-01 1.33348525e+00 -1.47840416e+00
-4.95637916e-02 -2.62526553e-02 7.29385912e-02 -1.21203578e+00
5.25182784e-01 -3.48077327e-01 8.01673651e-01 -1.86332774e+00
-9.56650972e-02 -2.29221210e-01 9.36800778e-01 2.65579551e-01
5.57922423e-02 -5.47593057e-01 3.96320432e-01 1.16582938e-01
-1.14190447e+00 8.02646160e-01 1.28613889e+00 2.05696821e-01
-7.46784627e-01 -2.28202507e-01 -3.98020804e-01 1.06136477e+00
1.00236952e+00 -6.80631101e-02 -6.48624241e-01 -3.13225180e-01
3.32124144e-01 6.18384898e-01 2.71812171e-01 -7.60451555e-01
6.71876252e-01 -4.34228510e-01 -1.23736791e-01 -4.43675756e-01
3.78372043e-01 -8.80383134e-01 -2.09143192e-01 3.60736310e-01
-8.06024194e-01 -1.25977889e-01 8.06443319e-02 2.22195789e-01
-4.14877981e-02 -7.73669302e-01 3.44984531e-02 -5.04589498e-01
-1.09075570e+00 -6.36826575e-01 -1.38255692e+00 1.74200922e-01
1.13175833e+00 -1.38091758e-01 -3.37717026e-01 -1.18199301e+00
-7.76344538e-01 6.94922864e-01 5.34118891e-01 4.54259545e-01
7.09496737e-01 -9.68695819e-01 -5.89918911e-01 -2.48940140e-01
4.07927155e-01 1.80770531e-01 2.22960748e-02 -1.34031298e-02
-7.41076112e-01 5.52594721e-01 -4.34928745e-01 -9.48151425e-02
-1.17745435e+00 1.03749670e-01 4.34584260e-01 -1.56272814e-01
-5.00420451e-01 8.42993498e-01 3.13952923e-01 -1.30271733e+00
4.13947403e-01 -4.72267903e-02 -7.62155533e-01 1.06462568e-01
5.58111846e-01 4.47298110e-01 -4.58267123e-01 -5.61835587e-01
7.56503940e-02 -1.32747650e-01 1.20879300e-01 -8.60153973e-01
1.09882653e+00 -6.43240035e-01 -1.17895111e-01 9.32520628e-01
7.35593736e-01 -4.53555465e-01 -9.72403049e-01 -5.05910158e-01
4.63263363e-01 -2.39696681e-01 -7.01764882e-01 -8.75828981e-01
2.25433007e-01 4.90284413e-01 1.52459413e-01 7.81624019e-01
2.04479992e-01 7.01078773e-01 5.11708796e-01 1.68768096e+00
1.01302099e+00 -1.50103402e+00 8.03685844e-01 1.48383069e+00
1.54169309e+00 -1.42137265e+00 -3.03412139e-01 -4.85902965e-01
-1.17298591e+00 1.12289739e+00 1.00086260e+00 5.98781765e-01
5.90504184e-02 -4.58073765e-02 4.34569001e-01 -4.87628847e-01
-8.57858598e-01 -3.35976958e-01 -3.67669284e-01 1.33175826e+00
2.34312505e-01 -1.04755141e-01 1.18625537e-01 5.28507233e-01
-7.82723784e-01 -3.89800400e-01 9.01083112e-01 1.39714193e+00
-6.58415318e-01 -1.03477275e+00 -6.43452883e-01 5.08738086e-02
4.02173907e-01 3.83754134e-01 -1.42153656e+00 7.36092627e-01
-1.35643259e-01 1.77753520e+00 -9.63956192e-02 -2.04992041e-01
7.15100467e-01 2.84519166e-01 4.79555607e-01 -1.12250972e+00
-7.70415246e-01 -5.63802063e-01 8.96937072e-01 -3.70196134e-01
-3.60648394e-01 -1.00510502e+00 -1.71975875e+00 -7.30756447e-02
-1.90931708e-01 6.54420257e-01 5.64419508e-01 1.25863445e+00
-9.42398608e-02 3.82225096e-01 4.80012745e-01 -1.16894960e+00
-2.42223829e-01 -1.33376241e+00 -3.23081046e-01 1.80170998e-01
2.00967342e-01 -4.21797872e-01 -2.42900848e-01 4.19495031e-02] | [4.426083564758301, 0.7429066896438599] |
e7ee855c-0b8a-4c6d-97ea-b68689c40741 | learning-sequence-encoders-for-temporal | 1809.03202 | null | http://arxiv.org/abs/1809.03202v1 | http://arxiv.org/pdf/1809.03202v1.pdf | Learning Sequence Encoders for Temporal Knowledge Graph Completion | Research on link prediction in knowledge graphs has mainly focused on static
multi-relational data. In this work we consider temporal knowledge graphs where
relations between entities may only hold for a time interval or a specific
point in time. In line with previous work on static knowledge graphs, we
propose to address this problem by learning latent entity and relation type
representations. To incorporate temporal information, we utilize recurrent
neural networks to learn time-aware representations of relation types which can
be used in conjunction with existing latent factorization methods. The proposed
approach is shown to be robust to common challenges in real-world KGs: the
sparsity and heterogeneity of temporal expressions. Experiments show the
benefits of our approach on four temporal KGs. The data sets are available
under a permissive BSD-3 license 1. | ['Sebastijan Dumančić', 'Alberto García-Durán', 'Mathias Niepert'] | 2018-09-10 | learning-sequence-encoders-for-temporal-1 | https://aclanthology.org/D18-1516 | https://aclanthology.org/D18-1516.pdf | emnlp-2018-10 | ['temporal-knowledge-graph-completion'] | ['knowledge-base'] | [-9.95042250e-02 3.06056082e-01 -1.01110256e+00 -2.19580501e-01
-1.85750321e-01 -5.44063210e-01 5.56069493e-01 4.88718927e-01
2.57142987e-02 8.04629564e-01 1.89902991e-01 -5.10194004e-01
-6.46849871e-01 -1.20287466e+00 -6.25257194e-01 -2.39399105e-01
-8.05431783e-01 5.15033603e-01 2.65838712e-01 -2.07790673e-01
-2.71553457e-01 5.62709391e-01 -1.13517618e+00 3.83083403e-01
5.52428782e-01 8.03277433e-01 -4.47247833e-01 5.44424295e-01
-4.76713739e-02 1.21739161e+00 -3.48569363e-01 -5.38272262e-01
2.31221423e-01 2.03775689e-01 -1.24829102e+00 -6.10273182e-02
1.09282859e-01 -2.09233284e-01 -8.27322960e-01 6.51457369e-01
1.10221200e-01 3.91418368e-01 5.39591551e-01 -1.46206725e+00
-7.34730303e-01 9.76489544e-01 -4.93007720e-01 6.73388124e-01
4.79288846e-01 -5.01501977e-01 1.32633090e+00 -4.76633370e-01
9.87703264e-01 1.13243055e+00 6.52983963e-01 3.65652777e-02
-1.19421113e+00 -4.82028872e-01 6.11146212e-01 4.48197424e-01
-1.58426881e+00 -4.26419973e-01 6.34698033e-01 -4.84903961e-01
1.47585988e+00 6.23218529e-03 6.15804136e-01 9.36893046e-01
3.30528140e-01 7.29244232e-01 7.38534868e-01 -4.16705102e-01
-1.35368839e-01 -3.00974905e-01 4.26647663e-01 7.61873126e-01
3.63965482e-01 -6.74942732e-02 -8.00380886e-01 -3.18499535e-01
7.93269455e-01 5.75887747e-02 -3.11077069e-02 -6.07686400e-01
-1.05151582e+00 7.89734960e-01 4.42385465e-01 5.91095686e-01
-4.62088794e-01 3.45149904e-01 4.89550322e-01 5.88690937e-01
7.76419938e-01 -8.02377090e-02 -7.64754355e-01 -8.07216838e-02
-6.76944911e-01 1.43989399e-01 9.95475352e-01 1.09908199e+00
6.83915019e-01 8.31393301e-02 -1.34991050e-01 6.01829529e-01
2.02268779e-01 -1.54988587e-01 2.18838274e-01 -6.50864661e-01
5.79234183e-01 8.68267834e-01 1.08836934e-01 -1.14399838e+00
-5.77121913e-01 -5.09850919e-01 -5.40190458e-01 -5.04839480e-01
2.07899630e-01 -3.06886006e-02 -1.01838803e+00 1.74479008e+00
4.42594141e-01 4.81565714e-01 2.81735927e-01 3.66537690e-01
8.15324545e-01 4.69300479e-01 1.12625413e-01 -5.79731584e-01
1.06328678e+00 -7.72016287e-01 -1.04364407e+00 1.68022424e-01
1.02578545e+00 -3.62435132e-01 2.56005108e-01 -7.96948671e-02
-1.03635430e+00 -3.12774517e-02 -7.95079112e-01 3.27734314e-02
-9.30073261e-01 -1.11200675e-01 1.23568451e+00 5.14569283e-01
-1.13409436e+00 7.20250130e-01 -1.19591296e+00 -7.14195430e-01
2.27362677e-01 5.03963292e-01 -3.17321479e-01 3.45000140e-02
-1.71318948e+00 7.77462304e-01 8.06175649e-01 3.84981483e-01
-5.99995136e-01 -5.59974730e-01 -9.26952541e-01 8.65842253e-02
9.15141881e-01 -8.25396299e-01 1.14041483e+00 -5.45938611e-01
-1.15807796e+00 4.47890162e-01 -3.59303117e-01 -7.91444659e-01
1.44839540e-01 -2.09107131e-01 -1.08422399e+00 7.07232803e-02
1.34094074e-01 -1.46665759e-02 4.42504108e-01 -9.39715922e-01
-4.32043970e-01 -2.68608034e-01 5.41634560e-01 1.07101247e-01
-3.75512540e-01 3.62840444e-02 -7.13817537e-01 -6.27924383e-01
1.46541014e-01 -8.93736422e-01 -1.64454758e-01 -7.24120140e-01
-5.84524095e-01 -3.24919760e-01 8.29674959e-01 -6.06696129e-01
1.74683022e+00 -1.59254491e+00 2.38133743e-01 4.90240872e-01
8.90073702e-02 5.70843145e-02 2.17410937e-01 8.43632221e-01
-3.57044667e-01 2.14577362e-01 3.46540064e-01 -6.90419525e-02
-1.63909435e-01 5.12679100e-01 -3.33635926e-01 3.37461561e-01
5.23766428e-02 1.18855631e+00 -9.52467918e-01 -4.04719293e-01
-2.08944783e-01 3.61924410e-01 1.59541648e-02 -1.83192611e-01
-3.39480788e-01 -6.71466673e-03 -6.16373122e-01 8.96124005e-01
1.86745793e-01 -6.64314449e-01 8.42037499e-01 -2.53610820e-01
6.70360550e-02 3.54682565e-01 -1.20852530e+00 1.63233435e+00
-3.00567389e-01 1.22090563e-01 -3.82801443e-01 -1.08483613e+00
4.95710373e-01 7.19719708e-01 8.12810838e-01 -4.66954648e-01
-7.72015154e-02 -6.21804036e-02 1.95849612e-02 -2.62658298e-01
7.26409376e-01 1.43937275e-01 -6.31309533e-03 4.79211032e-01
3.11397225e-01 5.91044962e-01 7.19687998e-01 6.71275675e-01
1.33806908e+00 2.26921767e-01 4.68560696e-01 1.13654137e-02
1.92413107e-01 -1.13253079e-01 7.51471817e-01 4.49999094e-01
6.17848225e-02 -3.03522587e-01 7.14791417e-01 -4.96018142e-01
-5.32394290e-01 -9.79768395e-01 -4.42203181e-03 9.42634583e-01
-1.36731863e-01 -9.69263911e-01 1.82637051e-01 -9.13764477e-01
3.07033151e-01 4.27803248e-01 -8.74324918e-01 -1.65656328e-01
-4.06216204e-01 -5.92520237e-01 3.06987762e-01 7.56061375e-01
-2.31833495e-02 -5.85653245e-01 -1.05666004e-01 4.12101358e-01
-4.18101549e-02 -1.27076948e+00 -1.26786962e-01 3.35983306e-01
-9.17568326e-01 -1.19034731e+00 -2.01872975e-01 -4.42374796e-01
4.88475591e-01 1.70722052e-01 1.17207420e+00 -1.13387525e-01
-4.94172014e-02 8.65107894e-01 -5.90063214e-01 -8.23387951e-02
-4.64891382e-02 4.82668221e-01 3.01034689e-01 -3.99907725e-03
2.65897661e-01 -8.08739245e-01 -2.30649307e-01 1.53965220e-01
-9.31812584e-01 -8.36141631e-02 3.53542000e-01 7.64070272e-01
4.38390523e-01 5.75065732e-01 6.36787117e-01 -1.29670858e+00
6.39808118e-01 -1.01598132e+00 -4.45704043e-01 6.78348958e-01
-9.40927088e-01 5.11777438e-02 1.44596577e-01 -4.52112347e-01
-1.07575440e+00 -1.09688319e-01 5.82601011e-01 -5.76727331e-01
3.48145157e-01 1.31141198e+00 3.61951478e-02 -4.07012403e-02
2.57447124e-01 -1.90412283e-01 -5.39193153e-01 -2.27042988e-01
5.99113345e-01 2.67838724e-02 2.31761813e-01 -7.73818016e-01
9.11050081e-01 4.76772904e-01 1.58527032e-01 -4.25767809e-01
-7.74083018e-01 -5.93703687e-01 -9.49582994e-01 -1.70917690e-01
4.28892136e-01 -1.19074619e+00 -7.06150174e-01 6.32262453e-02
-7.20667720e-01 -2.76367396e-01 -1.47747040e-01 3.94526422e-01
-3.79856974e-01 3.17932963e-01 -7.96222627e-01 -8.84197056e-01
-8.30984786e-02 -4.04091120e-01 7.55932927e-01 1.87239461e-02
-2.59275913e-01 -1.42583644e+00 4.48865145e-02 2.17365220e-01
1.62055373e-01 3.91097158e-01 9.82126296e-01 -6.99883461e-01
-8.67394447e-01 -2.03714266e-01 1.93728600e-02 -3.71382833e-01
5.53267360e-01 8.49091560e-02 -4.55513209e-01 -3.33033055e-01
-7.29315817e-01 -3.07805955e-01 1.05635452e+00 3.35898042e-01
6.84206069e-01 -4.79631960e-01 -8.27634692e-01 4.24278170e-01
1.42048681e+00 2.77151346e-01 1.45490125e-01 2.72735327e-01
8.31443787e-01 5.07063508e-01 6.94802582e-01 5.79545975e-01
7.98245966e-01 8.88705730e-01 1.15250438e-01 1.65306956e-01
1.25188038e-01 -4.25259411e-01 1.49535984e-01 8.83010268e-01
-4.13042396e-01 -3.49780411e-01 -1.05692518e+00 9.86898780e-01
-2.40219355e+00 -1.12384892e+00 -5.32089025e-02 1.80788994e+00
7.09562421e-01 2.02962652e-01 2.49401018e-01 8.78509805e-02
5.02049387e-01 3.77466708e-01 -4.52405781e-01 -3.44582677e-01
-1.46926522e-01 -7.03823194e-02 7.26457417e-01 4.03689504e-01
-1.17698097e+00 1.07597649e+00 6.70403337e+00 4.88544017e-01
-8.06347489e-01 2.15862229e-01 2.32600242e-01 -2.65101790e-01
-5.45685232e-01 3.44956338e-01 -7.13965297e-01 5.19114882e-02
1.22591043e+00 -5.10046124e-01 3.33595693e-01 4.39574808e-01
8.04111511e-02 -4.20158962e-03 -1.08603036e+00 5.96714616e-01
-2.68102109e-01 -1.46267033e+00 -1.04510210e-01 1.21667571e-01
8.30984056e-01 -1.07219502e-01 9.67651606e-03 4.55328047e-01
8.75351548e-01 -8.79641116e-01 2.76392877e-01 6.82348430e-01
6.18189514e-01 -7.36044943e-01 5.39284945e-01 -2.99098119e-02
-1.74802220e+00 -7.19193965e-02 9.23631713e-02 -1.49243593e-01
3.06491792e-01 6.39954567e-01 -1.15726507e+00 1.45596969e+00
7.13624716e-01 1.04472399e+00 -4.87540513e-01 6.76902652e-01
-1.86715186e-01 5.23078322e-01 -3.33441764e-01 3.66553247e-01
-7.07048997e-02 1.72398817e-02 3.06011945e-01 9.04055178e-01
2.08410367e-01 4.47427621e-03 3.43068063e-01 5.06089509e-01
-8.15152153e-02 8.30637142e-02 -9.53667104e-01 -5.83334506e-01
5.14324844e-01 1.12039101e+00 -8.60281110e-01 -2.38536835e-01
-7.10046828e-01 7.40118980e-01 6.89141691e-01 7.86208034e-01
-5.53574204e-01 3.87341566e-02 3.08839917e-01 6.21359050e-02
5.50711513e-01 -5.44724762e-01 2.82375425e-01 -1.45973229e+00
3.44713666e-02 -3.22415620e-01 1.08139396e+00 -5.26442409e-01
-1.29162073e+00 3.20361555e-01 4.43542570e-01 -8.45112503e-01
-4.71455097e-01 -1.38523549e-01 -2.61469036e-01 6.53058112e-01
-1.46861935e+00 -1.77311242e+00 3.10705394e-01 8.39001358e-01
-4.99352291e-02 -1.11653276e-01 8.97364914e-01 3.71825486e-01
-7.37352490e-01 4.45540905e-01 6.86988384e-02 2.76077300e-01
5.35053670e-01 -1.27990723e+00 4.14698243e-01 9.91013169e-01
5.08799255e-01 1.09770012e+00 5.52071035e-01 -1.16509855e+00
-1.45692599e+00 -1.11609316e+00 1.04203153e+00 -3.29650134e-01
1.25494289e+00 -2.51453727e-01 -9.31150556e-01 1.51138115e+00
5.57830092e-03 3.49571288e-01 9.63595152e-01 1.00344384e+00
-5.36034882e-01 -1.15645215e-01 -6.83086157e-01 3.74440372e-01
1.31570137e+00 -7.70507991e-01 -4.56391215e-01 4.47417736e-01
8.75362277e-01 -5.79682350e-01 -1.55114877e+00 7.64718831e-01
5.22386193e-01 -3.10831398e-01 9.08999741e-01 -1.02982032e+00
-2.41535413e-03 -3.94507080e-01 -1.45714479e-02 -1.10566711e+00
-4.43262726e-01 -7.44974554e-01 -1.12541962e+00 1.42191124e+00
4.95852530e-01 -5.49671054e-01 9.45108652e-01 5.95695078e-01
2.39169225e-01 -8.72425556e-01 -1.00463009e+00 -9.64390099e-01
-2.80406177e-01 -3.29984874e-01 5.24524271e-01 1.42151225e+00
3.45324934e-01 4.32870120e-01 -6.28869116e-01 5.20280063e-01
2.63464183e-01 5.95138729e-01 4.90329891e-01 -1.30853152e+00
-3.74277949e-01 -9.61345583e-02 -4.95953888e-01 -6.01137459e-01
5.04071534e-01 -1.02849603e+00 -7.39078045e-01 -1.76764774e+00
-2.21500918e-02 -3.56755972e-01 -7.27664769e-01 9.42945421e-01
-1.63433850e-02 -2.80170828e-01 -2.05751315e-01 1.88764259e-01
-7.75132000e-01 4.53612894e-01 7.25605428e-01 -2.78694898e-01
-2.79250801e-01 1.22110039e-01 -4.77096468e-01 2.38847196e-01
6.38124526e-01 -2.77974337e-01 -9.12177086e-01 -8.01135600e-02
7.16510773e-01 4.21138376e-01 5.73948659e-02 -4.34426427e-01
4.44104135e-01 -2.37839863e-01 1.05212420e-01 -7.02707767e-01
4.41350937e-01 -8.96240413e-01 6.35464370e-01 2.30027705e-01
-1.79466397e-01 8.43254253e-02 2.71410763e-01 1.11093819e+00
-3.97759557e-01 2.36822814e-01 -4.30767462e-02 -4.14920785e-03
-9.82836008e-01 5.90250492e-01 -1.60980299e-01 -3.77037048e-01
1.19948769e+00 -1.47883311e-01 -4.84940112e-01 -4.86061245e-01
-1.18485069e+00 5.56640327e-01 1.90997347e-01 6.01052105e-01
3.94370019e-01 -1.57606947e+00 -2.94356942e-01 -2.73618460e-01
3.10911775e-01 -2.26381302e-01 4.20164734e-01 8.75198781e-01
1.80491939e-01 5.59376895e-01 7.85286278e-02 -1.90249950e-01
-1.18879068e+00 8.83400559e-01 1.86870039e-01 -8.04457307e-01
-4.69271719e-01 6.15339637e-01 -3.55381101e-01 -1.18587419e-01
4.81903106e-02 -3.99476260e-01 -4.75867748e-01 3.99920285e-01
-5.34783527e-02 2.12922826e-01 5.38246222e-02 -7.36720383e-01
-4.89405692e-01 2.47371912e-01 -4.13345337e-01 2.06728086e-01
1.47930157e+00 -1.25166968e-01 -2.03425020e-01 1.04555464e+00
9.24260259e-01 7.40719363e-02 -7.12417126e-01 -6.82874739e-01
5.66447258e-01 -3.06170255e-01 -1.43299073e-01 -6.47023141e-01
-1.18668699e+00 1.77261278e-01 2.13454247e-01 5.44642627e-01
9.45612490e-01 1.71667174e-01 5.50403833e-01 3.14150482e-01
5.33662677e-01 -9.32232022e-01 -2.38366976e-01 5.34666777e-01
4.31342840e-01 -1.00142968e+00 4.44685876e-01 -7.42005289e-01
-4.33620542e-01 1.03261411e+00 5.54453731e-01 2.82764435e-01
8.67429674e-01 1.28350675e-01 -4.11202461e-01 -5.13283908e-01
-1.26461840e+00 -4.24362093e-01 3.29909027e-01 4.55093980e-01
6.08134270e-01 2.17616841e-01 -2.55535126e-01 4.37079012e-01
1.13891274e-01 2.45384663e-01 5.58524609e-01 1.31469941e+00
1.36526331e-01 -1.54469836e+00 1.05203874e-01 5.99179566e-01
-6.38428926e-01 -5.10254875e-02 -3.08065087e-01 7.84478128e-01
7.54300365e-03 8.94434631e-01 -1.26134679e-01 -4.67655867e-01
2.65637308e-01 3.11365515e-01 4.79334205e-01 -7.53934145e-01
-3.24377984e-01 -8.07454735e-02 5.80172002e-01 -6.06886029e-01
-9.21095014e-01 -8.19299996e-01 -1.16062737e+00 -2.41716087e-01
-4.06735867e-01 1.59521267e-01 3.10841277e-02 8.48670363e-01
5.38277090e-01 8.83728147e-01 2.18337163e-01 -1.51577041e-01
7.98925087e-02 -8.72042120e-01 -9.08106685e-01 1.92826152e-01
1.86228439e-01 -1.07363153e+00 -4.98820804e-02 7.32333511e-02] | [8.55846118927002, 7.89893102645874] |
1b7784c7-61f8-4dd4-8027-34600277a8a9 | persuasive-dialogue-understanding-the | 2011.09954 | null | https://arxiv.org/abs/2011.09954v2 | https://arxiv.org/pdf/2011.09954v2.pdf | Persuasive Dialogue Understanding: the Baselines and Negative Results | Persuasion aims at forming one's opinion and action via a series of persuasive messages containing persuader's strategies. Due to its potential application in persuasive dialogue systems, the task of persuasive strategy recognition has gained much attention lately. Previous methods on user intent recognition in dialogue systems adopt recurrent neural network (RNN) or convolutional neural network (CNN) to model context in conversational history, neglecting the tactic history and intra-speaker relation. In this paper, we demonstrate the limitations of a Transformer-based approach coupled with Conditional Random Field (CRF) for the task of persuasive strategy recognition. In this model, we leverage inter- and intra-speaker contextual semantic features, as well as label dependencies to improve the recognition. Despite extensive hyper-parameter optimizations, this architecture fails to outperform the baseline methods. We observe two negative results. Firstly, CRF cannot capture persuasive label dependencies, possibly as strategies in persuasive dialogues do not follow any strict grammar or rules as the cases in Named Entity Recognition (NER) or part-of-speech (POS) tagging. Secondly, the Transformer encoder trained from scratch is less capable of capturing sequential information in persuasive dialogues than Long Short-Term Memory (LSTM). We attribute this to the reason that the vanilla Transformer encoder does not efficiently consider relative position information of sequence elements. | ['Soujanya Poria', 'Amir Hussain', 'Navonil Majumder', 'Deepanway Ghosal', 'Hui Chen'] | 2020-11-19 | null | null | null | null | ['intent-recognition', 'dialogue-understanding'] | ['natural-language-processing', 'natural-language-processing'] | [ 8.85776699e-01 6.34585321e-01 -2.45207846e-01 -7.08136857e-01
-6.87571704e-01 -4.83693957e-01 1.24500310e+00 1.69021841e-02
-7.58946836e-01 9.72435832e-01 8.76704335e-01 -8.90079558e-01
4.88859303e-02 -6.06936812e-01 -3.22094113e-01 -3.46967399e-01
3.31412524e-01 3.01615328e-01 -1.43660277e-01 -5.72468877e-01
4.32221085e-01 1.22561842e-01 -1.34547961e+00 4.58699465e-01
7.93004274e-01 7.51372159e-01 2.50180095e-01 1.08622539e+00
-5.77320993e-01 1.71713638e+00 -8.90041828e-01 -3.39707851e-01
-4.25094843e-01 -8.17302883e-01 -1.39428949e+00 -6.96090087e-02
-4.55858232e-03 -2.26600587e-01 -2.30169043e-01 7.19868064e-01
4.48268116e-01 2.78123379e-01 4.63375747e-01 -7.27032006e-01
-5.24216413e-01 1.00751281e+00 1.46880597e-01 1.44584656e-01
6.29450083e-01 -4.03996073e-02 1.11123538e+00 -4.46899980e-01
4.72651809e-01 1.50947106e+00 7.12049246e-01 9.19056952e-01
-9.89607692e-01 -3.13396342e-02 2.38829628e-01 3.35471720e-01
-5.91032088e-01 -5.52508235e-01 9.65689898e-01 -7.32133761e-02
1.14974892e+00 4.38901097e-01 5.42181075e-01 1.90261877e+00
2.31093630e-01 1.16899395e+00 1.28501463e+00 -6.23949885e-01
1.58398077e-01 -2.07815748e-02 5.41892409e-01 6.82969868e-01
-6.20203435e-01 1.02072306e-01 -4.05588120e-01 -3.54492038e-01
3.12965602e-01 -9.90074053e-02 -2.01716170e-01 3.49427342e-01
-1.14783537e+00 1.28417218e+00 1.08141385e-01 5.09770036e-01
-4.46061373e-01 -1.49070341e-02 6.97583914e-01 3.49636793e-01
3.46090972e-01 5.77389240e-01 -7.06429422e-01 -7.89073586e-01
-3.61978352e-01 -1.82857230e-01 1.40391695e+00 3.95661741e-01
5.91730237e-01 1.81549624e-01 -4.37621772e-01 1.08634126e+00
1.53582841e-01 1.94990069e-01 6.99137509e-01 -9.12034154e-01
2.93317080e-01 5.02942979e-01 -3.60431857e-02 -8.93741786e-01
-4.92392033e-01 -1.66603267e-01 -8.60796690e-01 -2.53015906e-01
4.71044689e-01 -3.98199201e-01 -6.30396008e-01 1.83237839e+00
4.68960144e-02 -5.90231456e-02 4.07945991e-01 7.52685606e-01
1.08027661e+00 6.41986132e-01 4.54664379e-01 -3.97021592e-01
1.60403609e+00 -1.05097198e+00 -9.97126102e-01 -6.24972820e-01
7.08620071e-01 -7.12393463e-01 1.10516286e+00 -6.87977746e-02
-7.00546563e-01 -5.38327336e-01 -9.37597990e-01 -1.38944119e-01
-4.59510326e-01 -3.35740745e-02 8.92672658e-01 6.85507119e-01
-7.24708021e-01 3.88852030e-01 -4.31983024e-01 -2.77727872e-01
1.96779281e-01 1.86581984e-02 -4.04304378e-02 3.44257742e-01
-1.81673801e+00 1.21914041e+00 1.59321591e-01 2.97248960e-01
-6.51775658e-01 -3.02580744e-01 -1.13427114e+00 1.62532777e-01
6.90186679e-01 -6.59475029e-01 1.59786308e+00 -1.16411602e+00
-2.32470441e+00 7.19043434e-01 -2.02787474e-01 -6.85079336e-01
3.63728970e-01 -1.62144497e-01 -3.67840350e-01 -1.51404232e-01
-5.42758442e-02 5.24927855e-01 9.08965409e-01 -7.13417470e-01
-5.95734954e-01 -1.68564335e-01 2.81695127e-01 4.21224475e-01
-5.80088049e-02 3.71170700e-01 2.90052265e-01 -2.77621746e-01
-4.79383647e-01 -7.81404138e-01 -4.39569324e-01 -9.24048603e-01
-5.41354299e-01 -9.02619064e-01 6.18043005e-01 -5.94299018e-01
1.08541930e+00 -1.79380286e+00 -1.02353320e-01 -2.81931221e-01
1.78637117e-01 4.84123975e-01 -1.15474552e-01 5.39207697e-01
1.03221841e-01 1.86579868e-01 3.24280560e-02 -2.74040192e-01
3.31247538e-01 4.81192529e-01 -4.08093750e-01 1.29888549e-01
1.76947489e-01 1.18525589e+00 -8.15626562e-01 -3.62671673e-01
1.91327289e-01 5.01092315e-01 -8.57690647e-02 4.23580468e-01
-4.56508398e-01 4.59917575e-01 -6.90139174e-01 -4.41404805e-03
1.86031684e-01 -4.54103529e-01 5.51383734e-01 2.31187433e-01
-1.51977524e-01 1.36912382e+00 -4.92309898e-01 1.40980923e+00
-7.78270245e-01 6.43850982e-01 9.65552703e-02 -1.15597665e+00
8.77211511e-01 5.35488486e-01 -1.61041263e-02 -8.96910965e-01
3.01590204e-01 5.32878675e-02 1.70830771e-01 -2.48632312e-01
6.98594511e-01 -1.76344112e-01 -4.29681748e-01 8.00230026e-01
-1.96615323e-01 2.51304328e-01 -3.12830210e-02 2.19485536e-01
1.19556785e+00 -1.90776125e-01 6.63263381e-01 -1.00349851e-01
9.03190374e-01 -1.41644657e-01 6.44951165e-01 1.18863440e+00
-2.30585169e-02 1.74496710e-01 6.72897041e-01 -4.81930256e-01
-6.63318455e-01 -5.28505445e-01 1.28711864e-01 1.48499405e+00
-2.48366341e-01 -4.59339470e-01 -6.57355785e-01 -1.17172611e+00
-4.28772479e-01 1.10793972e+00 -6.90876424e-01 -1.00102127e-01
-9.71069992e-01 -4.56230074e-01 7.72923708e-01 3.34735006e-01
6.85379803e-01 -1.53229761e+00 -7.14563012e-01 5.58077455e-01
-2.54775345e-01 -9.72993731e-01 -4.12227929e-01 3.55405480e-01
-4.24435347e-01 -1.17010891e+00 -2.16994807e-01 -6.39935613e-01
1.60833374e-02 9.01516993e-03 1.22150898e+00 7.17937946e-02
2.23006666e-01 4.89903569e-01 -3.49502921e-01 -3.22732717e-01
-1.02113903e+00 2.97962964e-01 -2.94705212e-01 -1.42953202e-01
5.79163730e-01 -3.37076545e-01 -3.13458651e-01 2.67509669e-01
-5.24160683e-01 2.65125066e-01 4.28447813e-01 1.32713902e+00
-1.04526058e-01 -2.99574137e-01 8.50504935e-01 -1.22184229e+00
1.24767804e+00 -2.14029789e-01 -9.94888693e-02 2.99921185e-01
-5.79415739e-01 1.69238776e-01 8.42746139e-01 -4.25535291e-01
-1.49289930e+00 -4.67997938e-01 -7.41672993e-01 3.35944295e-01
-5.41358650e-01 6.22713566e-01 -1.50595307e-01 5.73629260e-01
5.51910043e-01 4.61429685e-01 2.91309476e-01 -5.32752216e-01
2.58549094e-01 8.59670103e-01 2.90077180e-01 -3.69784117e-01
-1.54268891e-02 2.09075920e-02 -4.42313880e-01 -9.22766566e-01
-1.32487166e+00 -2.16989547e-01 -2.92260706e-01 -1.00459315e-01
1.14802289e+00 -4.16651189e-01 -1.20202529e+00 3.70636344e-01
-1.41834891e+00 -5.35063744e-01 -2.14557022e-01 4.01040524e-01
-6.71083093e-01 5.75057328e-01 -9.04177129e-01 -1.08735752e+00
-5.42334139e-01 -9.85341549e-01 8.19380462e-01 2.84311056e-01
-5.46589553e-01 -1.27888942e+00 -2.35038459e-01 6.95836961e-01
6.66510463e-01 -2.27755815e-01 9.27699447e-01 -1.28282404e+00
-9.70903598e-03 -9.70238727e-03 7.94918314e-02 5.53115904e-01
1.80162966e-01 -5.79828143e-01 -7.79217839e-01 2.86210388e-01
4.91130769e-01 -5.17412782e-01 7.79089689e-01 1.83885872e-01
7.90832162e-01 -8.41033816e-01 1.49289425e-02 -1.81842238e-01
4.99437034e-01 2.89959848e-01 8.57564628e-01 2.76332676e-01
5.69137692e-01 7.64805555e-01 4.93885547e-01 1.65003687e-01
7.06988037e-01 5.31364977e-01 -3.29475850e-02 -2.84042154e-02
-1.30461812e-01 -4.66300428e-01 6.37256801e-01 6.44279957e-01
2.61250317e-01 -3.47811401e-01 -8.15949380e-01 2.15454385e-01
-1.84477079e+00 -1.21361685e+00 -2.86415726e-01 1.84269309e+00
1.25820208e+00 1.96264401e-01 -9.34026092e-02 3.66992429e-02
7.23138511e-01 6.05513692e-01 -1.42334312e-01 -1.00508595e+00
-2.81172931e-01 1.49688758e-02 5.76773696e-02 8.97649705e-01
-1.07164466e+00 1.05301893e+00 5.80508327e+00 8.73062193e-01
-1.02643383e+00 2.96525866e-01 7.73376465e-01 6.89181447e-01
-3.70909512e-01 7.75747523e-02 -9.09346223e-01 3.85216624e-01
1.29286981e+00 2.16042891e-01 3.09212238e-01 6.54987693e-01
2.23230392e-01 -1.84385836e-01 -1.06207383e+00 5.49261570e-01
1.08295381e-01 -1.50558758e+00 -3.83545190e-01 9.36412346e-03
1.36959180e-01 -4.62577231e-02 -2.21541777e-01 6.99313402e-01
6.46199524e-01 -1.17028081e+00 4.69556212e-01 5.11485398e-01
4.07665521e-01 -5.16560555e-01 8.96235704e-01 7.99289584e-01
-4.87455994e-01 8.96704420e-02 -1.81411043e-01 -4.43109751e-01
4.70627427e-01 3.11109632e-01 -1.44649863e+00 2.51344502e-01
1.39459103e-01 5.51422238e-01 -3.28977942e-01 -6.21986873e-02
-6.73991263e-01 1.11193287e+00 -2.32816607e-01 -6.89843833e-01
5.91121495e-01 -1.38352275e-01 7.41871476e-01 1.42057455e+00
-4.01612878e-01 7.64943883e-02 2.62643397e-01 5.65930307e-01
-1.36755789e-02 -1.67112313e-02 -5.19127250e-01 -3.98934007e-01
2.21226469e-01 1.12025130e+00 -3.23770285e-01 -5.35876393e-01
-3.75966251e-01 7.93399155e-01 4.19646263e-01 1.39056206e-01
-5.97003579e-01 6.18852228e-02 5.17767012e-01 -2.98817426e-01
2.53940433e-01 -1.18408784e-01 -4.15091872e-01 -9.68218684e-01
-3.57818812e-01 -1.18689120e+00 5.28482616e-01 -3.79854977e-01
-1.43403649e+00 7.97727704e-01 -3.40094268e-01 -4.15514201e-01
-9.25042093e-01 -4.16583538e-01 -7.70825446e-01 9.29819524e-01
-1.52688324e+00 -1.10054088e+00 1.43915802e-01 2.95230597e-01
9.71807063e-01 -1.80359527e-01 1.24496448e+00 -1.49119526e-01
-3.76496315e-01 3.65660578e-01 -4.38144714e-01 1.82509467e-01
5.44430733e-01 -1.40136480e+00 2.93113887e-01 4.09142077e-01
-6.61395714e-02 7.28684843e-01 7.55266309e-01 -5.52759111e-01
-1.29271424e+00 -6.25907958e-01 1.61719668e+00 -5.24293184e-01
7.68463790e-01 -4.39287901e-01 -9.46998596e-01 5.27876794e-01
7.36301959e-01 -6.01435006e-01 9.28787589e-01 6.78139508e-01
-3.26593727e-01 4.76320356e-01 -7.86873937e-01 7.89587617e-01
1.13132262e+00 -7.39290118e-01 -1.14219666e+00 3.28629434e-01
6.39269531e-01 -2.05345213e-01 -6.89972699e-01 2.49615252e-01
3.09825450e-01 -9.04753983e-01 7.04705060e-01 -8.52403700e-01
5.00059903e-01 2.02116594e-01 1.03008755e-01 -1.11455858e+00
-1.07384384e-01 -9.09222066e-01 -2.92332172e-02 1.23072088e+00
5.57309806e-01 -7.71753728e-01 5.41915476e-01 5.91189504e-01
-2.40622371e-01 -6.09625638e-01 -1.00278783e+00 -2.21496210e-01
-3.30680348e-02 -4.41645771e-01 4.67063367e-01 9.18507755e-01
2.01206878e-01 1.61238766e+00 -7.34881103e-01 -5.51768780e-01
4.06323560e-03 2.08581924e-01 7.33407557e-01 -1.12620211e+00
-2.83555329e-01 -5.43876290e-01 7.20886737e-02 -1.67220092e+00
9.15771902e-01 -6.73728704e-01 3.49924117e-01 -1.56630635e+00
-3.43166851e-02 -4.98510420e-01 1.37960240e-01 7.09812403e-01
-3.23399127e-01 -2.66093791e-01 -7.95547888e-02 -7.28494376e-02
-8.81310344e-01 9.29090381e-01 1.05064940e+00 -2.51716077e-01
-1.98568121e-01 3.19883913e-01 -8.60428393e-01 7.76997209e-01
9.15678740e-01 -3.34572434e-01 -2.48601601e-01 -1.16360582e-01
1.47749305e-01 4.40284103e-01 2.86292940e-01 -2.03478336e-01
2.82939315e-01 -2.02531487e-01 1.77059998e-03 -4.29813236e-01
4.22713578e-01 -3.57449353e-01 -3.36127758e-01 3.07462513e-01
-8.61780941e-01 -8.72707590e-02 -2.16867894e-01 7.50179887e-01
-2.30391100e-01 -4.51294273e-01 4.09553051e-01 -2.76032686e-01
-6.18557453e-01 -3.40377539e-01 -1.28106308e+00 -4.61966358e-02
2.84423172e-01 -1.46542326e-01 -5.04143178e-01 -7.78373241e-01
-5.05723596e-01 1.99629143e-01 8.54238719e-02 6.92517519e-01
4.69706088e-01 -8.35226357e-01 -8.44761193e-01 -2.75777299e-02
-2.70610631e-01 -3.76282096e-01 3.68824840e-01 9.93017137e-01
1.51128277e-01 7.20925629e-01 1.99535474e-01 -4.56683964e-01
-1.34694731e+00 2.72419125e-01 3.98165315e-01 -6.95272624e-01
-8.10344040e-01 7.23788798e-01 1.58822581e-01 -7.88933873e-01
2.26802990e-01 2.77296931e-01 -7.65585124e-01 -2.22790111e-02
5.54314852e-01 1.76478788e-01 9.13498830e-03 -6.45734906e-01
-3.71063232e-01 -3.18754077e-01 -3.99158478e-01 -1.38238788e-01
1.19829273e+00 -1.92788869e-01 -1.93645746e-01 4.29225564e-01
1.08304822e+00 2.12252796e-01 -9.99171019e-01 -2.66438872e-01
3.80209386e-01 -7.07142726e-02 -4.64013368e-02 -1.05845296e+00
-2.65194029e-01 7.21511364e-01 -2.39173751e-02 6.64910376e-01
6.01888478e-01 5.28919585e-02 8.15471947e-01 6.96195900e-01
5.21686152e-02 -1.29819310e+00 -1.82390586e-02 1.28583920e+00
9.09677505e-01 -1.10769093e+00 -4.50921655e-01 -3.62469494e-01
-1.02511632e+00 1.02383971e+00 3.78381789e-01 2.68780559e-01
4.06298429e-01 1.67662591e-01 3.37003827e-01 -2.89856166e-01
-1.14833546e+00 -1.52133793e-01 1.89489812e-01 3.99768263e-01
7.20874369e-01 1.08395174e-01 -7.69153297e-01 6.81527078e-01
-3.86611193e-01 -2.94154733e-01 5.01044095e-01 9.13743258e-01
-3.32537055e-01 -1.29068553e+00 4.70586494e-02 5.53778291e-01
-6.53202295e-01 -3.59586418e-01 -8.39591980e-01 3.42158109e-01
-3.52447659e-01 1.53991878e+00 -4.92033884e-02 -2.68553048e-01
1.62564024e-01 3.49230021e-01 2.02672452e-01 -7.44231224e-01
-1.07136524e+00 -4.60227951e-02 1.04918170e+00 -3.40994209e-01
-6.33307219e-01 -6.76503420e-01 -1.16612065e+00 -3.03598255e-01
-2.47198954e-01 6.72672987e-01 6.32563055e-01 1.38605177e+00
3.15384358e-01 6.11262679e-01 6.33012295e-01 -3.87922138e-01
-1.03607333e+00 -1.44931924e+00 -5.81355095e-02 2.62074947e-01
1.21727295e-01 -4.18120682e-01 -4.20116723e-01 -3.83753240e-01] | [12.759254455566406, 7.78432559967041] |
594a203c-6a9b-4753-a057-7253c328df61 | unsupervised-domain-adaptation-for-clinical | null | null | https://aclanthology.org/W17-2320 | https://aclanthology.org/W17-2320.pdf | Unsupervised Domain Adaptation for Clinical Negation Detection | Detecting negated concepts in clinical texts is an important part of NLP information extraction systems. However, generalizability of negation systems is lacking, as cross-domain experiments suffer dramatic performance losses. We examine the performance of multiple unsupervised domain adaptation algorithms on clinical negation detection, finding only modest gains that fall well short of in-domain performance. | ['Hadi Amiri', 'Guergana Savova', 'Timothy Miller', 'Steven Bethard'] | 2017-08-01 | null | null | null | ws-2017-8 | ['negation-detection'] | ['natural-language-processing'] | [ 3.38607162e-01 4.26318944e-01 -6.46702945e-01 -6.18327498e-01
-9.66190696e-01 -6.42635465e-01 1.26636431e-01 1.03558278e+00
-7.15852022e-01 1.23154795e+00 4.73333538e-01 -5.51794946e-01
-3.80341351e-01 -4.24897015e-01 -2.76377738e-01 -2.46056840e-01
-2.74453551e-01 8.64392936e-01 2.16471404e-01 -4.64720309e-01
-2.17871461e-02 4.45530206e-01 -6.34992480e-01 7.37304389e-01
8.15463364e-01 2.45112106e-01 -7.74732709e-01 6.69196308e-01
-2.18033671e-01 9.95016992e-01 -8.84339511e-01 -4.56784278e-01
-2.31379032e-01 -5.68216503e-01 -8.08239937e-01 -2.61538893e-01
6.38144091e-02 -1.72448799e-01 -1.71946794e-01 1.20919478e+00
5.63011527e-01 -3.79024804e-01 6.70362949e-01 -9.00007427e-01
-5.13158917e-01 7.36837387e-01 -3.87794040e-02 7.53535390e-01
7.15236306e-01 1.05410494e-01 1.10241210e+00 -5.13184786e-01
1.20762360e+00 8.09328735e-01 8.48907351e-01 6.73308849e-01
-1.12636733e+00 -7.26638317e-01 -1.21446349e-01 -3.11678678e-01
-1.23894429e+00 -5.90037048e-01 3.29986304e-01 -2.44176030e-01
1.60397267e+00 6.07401468e-02 4.83561188e-01 9.72119510e-01
5.32532752e-01 6.47295535e-01 8.48548651e-01 -4.03540313e-01
2.80556887e-01 2.38661170e-01 4.23484623e-01 5.50856113e-01
7.42384493e-01 -1.13525823e-01 -6.39496207e-01 -8.78534198e-01
2.66219437e-01 -4.12727416e-01 -2.76324868e-01 1.34768113e-01
-9.33875024e-01 8.95896137e-01 -7.21799582e-02 6.27976120e-01
-3.95165473e-01 -4.46749985e-01 9.10312831e-01 7.26017177e-01
5.77012241e-01 6.63469732e-01 -9.33751881e-01 -4.02047306e-01
-9.41545784e-01 3.63828570e-01 1.08623278e+00 8.52322936e-01
-9.32103619e-02 -3.16990107e-01 8.56040865e-02 6.36337399e-01
-2.45767176e-01 4.02019888e-01 6.63197637e-01 -5.36949039e-01
4.25037771e-01 7.10412264e-01 9.93561931e-03 -8.96292686e-01
-8.44338357e-01 -2.83145040e-01 -5.18584847e-01 -4.48025346e-01
4.43595618e-01 -5.63744783e-01 -7.31933832e-01 1.66190696e+00
1.75425887e-01 -4.35692012e-01 5.81359506e-01 5.04457474e-01
8.67767632e-01 -9.61574242e-02 7.36090779e-01 -5.27957857e-01
1.49419820e+00 -2.57470816e-01 -1.13898706e+00 -6.46007240e-01
1.03992999e+00 -7.42151320e-01 3.96980882e-01 6.48571193e-01
-1.23368621e+00 3.61517102e-01 -9.72842455e-01 -5.97626157e-02
-1.72039449e-01 -3.94035071e-01 7.34199286e-01 6.78705633e-01
-7.02153563e-01 4.69670504e-01 -8.86218071e-01 -5.85769296e-01
6.02111161e-01 6.24114096e-01 -7.41333902e-01 -1.80292889e-01
-1.49766839e+00 1.33889306e+00 5.60971022e-01 -3.72460693e-01
-1.29606098e-01 -8.20956826e-01 -8.94925416e-01 -2.09205732e-01
2.54170507e-01 -7.01150417e-01 1.49520385e+00 -7.86445856e-01
-8.28677952e-01 1.21741426e+00 -2.47291237e-01 -6.55521750e-01
3.47658575e-01 -1.44720659e-01 -1.03731346e+00 4.20254707e-01
1.75141901e-01 3.93323183e-01 1.93324164e-01 -3.71166825e-01
-5.48583090e-01 -3.26662421e-01 -3.37532192e-01 2.87131220e-01
-4.12731051e-01 3.58656436e-01 1.59693673e-01 -4.86396998e-01
-9.93577093e-02 -4.99502748e-01 -4.76389557e-01 -9.35484245e-02
-2.25656137e-01 -3.74710679e-01 4.59963918e-01 -5.44485569e-01
1.23225641e+00 -2.23252082e+00 -5.12101650e-01 1.03899084e-01
4.15791541e-01 3.50605100e-01 1.89538851e-01 3.90978038e-01
-4.99018401e-01 2.35955775e-01 -2.33218864e-01 4.30386364e-01
-2.95870662e-01 4.77653176e-01 -9.44607928e-02 4.56618875e-01
7.43223906e-01 8.69067669e-01 -1.20225978e+00 -9.18749213e-01
-2.85163939e-01 4.56565470e-02 -7.36214101e-01 -3.28096092e-01
-2.44584173e-01 -1.46592781e-01 -6.23643637e-01 8.06044817e-01
4.33491915e-01 -4.07609373e-01 9.00470853e-01 4.40709889e-02
2.32058898e-01 8.07230473e-01 -7.20197022e-01 1.36142814e+00
5.69131374e-01 2.91859269e-01 1.50056958e-01 -1.04063594e+00
5.82933962e-01 6.85356796e-01 7.10590720e-01 -6.67577744e-01
1.45116031e-01 3.92553687e-01 4.71190959e-01 -7.41206884e-01
2.77468503e-01 -1.13674331e+00 -1.80187970e-01 1.82710618e-01
4.34349924e-02 -1.12688161e-01 2.62600511e-01 4.47903097e-01
1.59469628e+00 -4.43893403e-01 1.18782377e+00 -3.93573344e-01
1.82813704e-01 9.06569660e-01 8.23548257e-01 6.41153276e-01
-6.92645192e-01 3.70012909e-01 1.14469302e+00 -2.30407834e-01
-7.83779621e-01 -7.58282065e-01 -7.56308436e-01 6.76204503e-01
-4.43975538e-01 -6.49901688e-01 -4.31394964e-01 -1.13970637e+00
2.67577350e-01 5.06870627e-01 -5.94998121e-01 -2.72733718e-01
-4.83987123e-01 -1.26734388e+00 1.12067640e+00 6.35911584e-01
-3.89381677e-01 -1.02855337e+00 -5.57868063e-01 5.24050653e-01
-1.65309846e-01 -1.56409466e+00 -2.10645318e-01 8.59118938e-01
-1.22091103e+00 -1.58692622e+00 -3.17888349e-01 -1.03158152e+00
5.65827012e-01 -3.66915673e-01 1.40785444e+00 6.42964095e-02
-3.21724534e-01 2.28543669e-01 -3.30472022e-01 -8.06774080e-01
-7.88111866e-01 -8.97460580e-02 2.32545167e-01 -1.23430383e+00
1.40622640e+00 -2.07315460e-01 -4.55605090e-01 -1.01489238e-01
-1.01102984e+00 -8.73987257e-01 8.76187623e-01 1.03695035e+00
4.86917853e-01 -7.51153827e-02 9.60740328e-01 -1.70609128e+00
1.18013835e+00 -6.39713824e-01 2.57380642e-02 -6.01447076e-02
-9.08485591e-01 -7.75148794e-02 4.71740723e-01 -2.85010904e-01
-7.70261407e-01 9.51183401e-03 -2.89988041e-01 2.56624579e-01
-4.23895180e-01 7.31402755e-01 2.24755690e-01 3.35499376e-01
1.26336300e+00 -2.89636850e-01 2.20933169e-01 -9.97833461e-02
5.52619696e-02 6.15089536e-01 3.48614424e-01 -2.63072819e-01
3.41424435e-01 2.67727137e-01 -3.27914298e-01 -8.88270676e-01
-9.70824063e-01 -8.00587118e-01 -2.56709844e-01 6.94218814e-01
7.38106668e-01 -1.03225660e+00 -3.83877099e-01 -2.56089270e-01
-8.36828709e-01 -1.96839925e-02 -6.47537351e-01 6.35336995e-01
-5.22132516e-02 4.93530840e-01 -9.99766946e-01 -2.71084189e-01
-4.31062222e-01 -6.29795849e-01 6.32787824e-01 -2.07440048e-01
-9.60400105e-01 -9.63844657e-01 5.23063540e-01 1.85017496e-01
-1.71962962e-01 1.51340142e-01 1.17756939e+00 -1.67834556e+00
4.24741030e-01 -6.11001849e-01 -4.95415665e-02 2.93494254e-01
2.59520888e-01 -2.77188957e-01 -7.11165965e-01 -1.15251884e-01
1.31341949e-01 -4.89402205e-01 6.18981540e-01 4.01623040e-01
4.75411028e-01 -2.36639440e-01 -8.33846390e-01 1.28060132e-02
1.23855984e+00 1.70515344e-01 4.56911772e-01 2.50078648e-01
-9.44895968e-02 3.94615203e-01 5.91316879e-01 4.26035255e-01
-1.56501420e-02 -7.43908063e-02 -3.87470096e-01 -1.15199640e-01
1.44860744e-01 9.91982892e-02 1.80132806e-01 4.85723525e-01
3.59649390e-01 -2.03176022e-01 -1.28886330e+00 7.97372341e-01
-1.39889300e+00 -7.96636522e-01 -1.43224686e-01 1.50138223e+00
1.41707015e+00 8.24212790e-01 9.09074694e-02 1.04107291e-01
2.18348861e-01 -3.06512684e-01 -5.05034208e-01 -6.68002486e-01
-3.11304659e-01 7.65952110e-01 4.37444150e-01 3.03553104e-01
-9.53154862e-01 8.06264877e-01 8.45637703e+00 1.28252476e-01
-7.35236108e-01 4.43756394e-02 3.65864187e-01 -1.60543606e-01
-6.89785108e-02 -2.40072414e-01 -6.70658767e-01 -1.10817868e-02
1.18531191e+00 -3.85267764e-01 -4.39921021e-01 9.68206525e-01
1.66029841e-01 -1.04306027e-01 -1.37975800e+00 5.78312278e-01
1.76405191e-01 -1.19559157e+00 -1.26173809e-01 7.56197497e-02
7.60684013e-01 3.09343159e-01 -2.90262401e-01 4.28391099e-01
3.56076658e-01 -9.70798850e-01 -7.78210163e-02 1.54530644e-01
6.91560566e-01 -6.28777623e-01 1.09034312e+00 2.01367557e-01
-3.65471125e-01 7.12542608e-02 -3.50681424e-01 -9.48223323e-02
9.91124064e-02 9.60605025e-01 -1.49773526e+00 2.50001669e-01
5.29458284e-01 5.62915325e-01 -2.60222197e-01 8.08395982e-01
-2.59727508e-01 7.32280135e-01 -3.78483474e-01 -1.19585551e-01
1.94417074e-01 3.38452816e-01 5.47368944e-01 1.77973866e+00
-3.86878848e-01 6.95384920e-01 1.15774460e-01 4.05523986e-01
-3.40065956e-01 4.30994213e-01 -9.82578456e-01 -5.60434997e-01
1.91498071e-01 9.01341081e-01 -6.43998921e-01 -6.54594302e-01
-6.47586465e-01 6.21114075e-01 1.36453182e-01 5.34880124e-02
-5.41217864e-01 -4.63706404e-01 8.36723387e-01 1.11728825e-01
1.38218984e-01 2.17503279e-01 -4.19689178e-01 -1.16370296e+00
8.43937099e-02 -1.42625844e+00 1.12054181e+00 -3.10543299e-01
-1.56401420e+00 3.57891440e-01 -1.43655434e-01 -9.60597157e-01
-3.83814394e-01 -8.04168940e-01 1.85149368e-02 5.87440491e-01
-1.49369085e+00 -5.69727421e-01 2.47571617e-01 5.83442330e-01
4.98605929e-02 3.12806778e-02 1.33052003e+00 4.35676128e-01
-1.99155286e-01 8.31643045e-01 -3.74453105e-02 5.27562141e-01
1.21294749e+00 -1.18455815e+00 1.32989794e-01 5.14743805e-01
-2.03795016e-01 9.74556744e-01 7.62038946e-01 -1.02456999e+00
-8.80833387e-01 -7.36731350e-01 1.51402569e+00 -6.92548573e-01
8.22240174e-01 1.06488280e-02 -1.20507967e+00 9.75836575e-01
1.93679422e-01 -1.73675820e-01 1.20520294e+00 5.80601215e-01
-3.26525599e-01 3.44369411e-01 -1.49835074e+00 3.18778276e-01
7.96728909e-01 -6.21472776e-01 -1.30774653e+00 7.09202170e-01
5.38794935e-01 -5.42934895e-01 -1.18770289e+00 5.65161347e-01
2.93100893e-01 -4.03327495e-01 8.78187001e-01 -1.33907390e+00
4.90383744e-01 1.60377100e-01 1.97895408e-01 -8.60635042e-01
-2.15941325e-01 -2.89503187e-01 1.43769220e-01 5.77048123e-01
9.32253838e-01 -6.18312180e-01 1.18557990e+00 8.73139560e-01
-2.41161920e-02 -4.80617642e-01 -6.33978963e-01 -4.25211638e-01
4.97032166e-01 -3.05651903e-01 -3.73146497e-02 1.48762989e+00
1.20180607e+00 6.42545938e-01 3.43287200e-01 1.61022082e-01
3.29518437e-01 -2.85476476e-01 5.44603951e-02 -1.20018673e+00
-2.70639181e-01 -4.04661328e-01 -6.04648411e-01 -4.50327873e-01
1.30296901e-01 -9.08523440e-01 -2.57472731e-02 -1.40289593e+00
5.81748962e-01 -7.68898893e-03 -5.17384171e-01 9.85556364e-01
-2.56082028e-01 1.70182019e-01 -4.78929877e-01 -3.52908857e-02
-7.19823062e-01 -1.68687738e-02 8.50581825e-01 -2.43162155e-01
-3.60997349e-01 -3.83198559e-01 -1.06405497e+00 9.18127418e-01
7.81497598e-01 -1.06648481e+00 -3.88363153e-01 -2.09062353e-01
4.25014943e-01 1.88037083e-01 -9.04469565e-02 -7.68883109e-01
3.15169543e-01 -2.13677391e-01 4.66380119e-01 -4.10240591e-01
-2.65479892e-01 -4.98780280e-01 -4.56605464e-01 8.58307123e-01
-4.15604413e-01 2.00762957e-01 6.98890328e-01 5.84198773e-01
-4.80594248e-01 -2.45955676e-01 6.46581233e-01 -3.71252805e-01
-3.24216515e-01 -2.89932042e-01 -6.38797343e-01 8.56605172e-01
6.55945182e-01 -1.75005808e-01 -3.33944350e-01 1.76767826e-01
-1.04171360e+00 1.62027285e-01 3.29552382e-01 6.44969195e-02
6.03209913e-01 -6.16112947e-01 -6.82341635e-01 -1.70456678e-01
3.71994734e-01 -1.70573071e-02 3.12352534e-02 9.79852974e-01
-6.40004218e-01 8.04530799e-01 3.52185108e-02 -3.96689624e-01
-1.43268001e+00 6.90980136e-01 4.10565704e-01 -5.73994517e-01
-7.90835500e-01 8.16574454e-01 -5.11720292e-02 -3.34365994e-01
1.72225200e-02 -5.09023488e-01 -5.21564409e-02 -2.68252520e-03
5.49940765e-01 -3.14472377e-01 5.62150061e-01 9.95105430e-02
-9.70744014e-01 -2.20787808e-01 -5.04687071e-01 -1.05730399e-01
1.38156438e+00 5.81379592e-01 -6.59884661e-02 9.34977084e-02
8.55131209e-01 8.04441944e-02 -7.84265846e-02 -9.35434923e-02
4.46518540e-01 1.73123270e-01 -7.10946545e-02 -1.20058072e+00
-3.99557620e-01 2.60971606e-01 3.00638586e-01 -9.64355767e-02
1.09807396e+00 1.45320101e-02 6.80561006e-01 7.31577754e-01
-2.40826607e-02 -1.17454600e+00 -2.15980679e-01 4.56621855e-01
2.95144945e-01 -1.32790554e+00 4.51346010e-01 -3.65755707e-01
-7.98681974e-01 8.94707859e-01 4.32944089e-01 -2.24547032e-02
8.02508056e-01 6.71557546e-01 3.45111787e-01 -8.18441689e-01
-9.99794126e-01 1.38055444e-01 -1.35639906e-01 5.96714139e-01
9.03825700e-01 9.34766978e-02 -9.31182623e-01 7.96433091e-01
6.36473522e-02 5.02769053e-01 5.14918804e-01 1.42317224e+00
-1.85635999e-01 -1.30771494e+00 -1.45760730e-01 7.11400807e-01
-1.14737725e+00 -5.43284357e-01 -7.02639103e-01 1.28947830e+00
-4.41627167e-02 8.25162649e-01 -2.90321290e-01 8.62233415e-02
8.63350809e-01 6.03252709e-01 3.70604753e-01 -9.03218806e-01
-9.30467606e-01 4.48837355e-02 6.61207795e-01 -4.41292584e-01
-5.16379595e-01 -9.05259192e-01 -1.71107554e+00 -1.69951886e-01
-3.70060086e-01 6.52963638e-01 4.01585698e-02 9.73585606e-01
3.82480413e-01 4.82159615e-01 -2.43737563e-01 7.29025781e-01
-7.47633338e-01 -9.23462391e-01 -5.20217717e-01 6.65614605e-01
4.66808498e-01 -1.96614444e-01 -2.43260860e-01 1.12040810e-01] | [8.464381217956543, 8.794323921203613] |
638ed437-025c-4ce6-834f-e83e7c83b5f5 | neighborhood-aware-scalable-temporal-network | 2209.01084 | null | https://arxiv.org/abs/2209.01084v3 | https://arxiv.org/pdf/2209.01084v3.pdf | Neighborhood-aware Scalable Temporal Network Representation Learning | Temporal networks have been widely used to model real-world complex systems such as financial systems and e-commerce systems. In a temporal network, the joint neighborhood of a set of nodes often provides crucial structural information useful for predicting whether they may interact at a certain time. However, recent representation learning methods for temporal networks often fail to extract such information or depend on online construction of structural features, which is time-consuming. To address the issue, this work proposes Neighborhood-Aware Temporal network model (NAT). For each node in the network, NAT abandons the commonly-used one-single-vector-based representation while adopting a novel dictionary-type neighborhood representation. Such a dictionary representation records a downsampled set of the neighboring nodes as keys, and allows fast construction of structural features for a joint neighborhood of multiple nodes. We also design a dedicated data structure termed N-cache to support parallel access and update of those dictionary representations on GPUs. NAT gets evaluated over seven real-world large-scale temporal networks. NAT not only outperforms all cutting-edge baselines by averaged 1.2% and 4.2% in transductive and inductive link prediction accuracy, respectively, but also keeps scalable by achieving a speed-up of 4.1-76.7x against the baselines that adopt joint structural features and achieves a speed-up of 1.6-4.0x against the baselines that cannot adopt those features. The link to the code: https: //github.com/Graph-COM/Neighborhood-Aware-Temporal-Network. | ['Pan Li', 'Yuhong Luo'] | 2022-09-02 | null | null | null | null | ['inductive-link-prediction'] | ['graphs'] | [-1.86345652e-01 -2.77565539e-01 -6.92137361e-01 -2.54539937e-01
-3.89352739e-02 -5.19791782e-01 6.97507977e-01 6.22990251e-01
-2.92974383e-01 5.86420596e-01 6.07548356e-02 -6.01236403e-01
-3.27597111e-01 -1.29612935e+00 -5.02764285e-01 -7.18316436e-01
-6.41275525e-01 5.65773726e-01 6.38651013e-01 -3.61498445e-01
3.66725847e-02 3.83333772e-01 -1.08615911e+00 9.67256278e-02
3.60737592e-01 1.02707338e+00 4.28026915e-02 3.67940217e-01
-2.42871329e-01 5.68769157e-01 -1.48159042e-01 -1.88445657e-01
3.74323756e-01 -8.40703174e-02 -7.00235307e-01 -4.96008307e-01
-8.99641216e-03 -2.68170893e-01 -9.87850189e-01 7.81118333e-01
2.64390200e-01 2.36993298e-01 3.01573813e-01 -1.42697859e+00
-4.91766572e-01 9.72249448e-01 -7.27596641e-01 5.43885469e-01
1.31171092e-01 1.05982588e-03 1.25354064e+00 -6.25860989e-01
7.68056393e-01 1.03708148e+00 8.08620155e-01 2.51277328e-01
-1.19889438e+00 -8.88386190e-01 3.47172141e-01 2.55900562e-01
-1.44380796e+00 -3.26597720e-01 7.45686233e-01 -1.54728130e-01
1.09904742e+00 1.64664403e-01 8.14555228e-01 1.00001037e+00
2.42595702e-01 5.09354234e-01 7.65040040e-01 1.73134226e-02
-9.39800777e-03 -2.98008204e-01 3.97675484e-01 7.95072675e-01
2.06996933e-01 9.51144993e-02 -6.35301471e-01 -5.67352474e-01
1.01451397e+00 4.95681822e-01 -2.28744950e-02 -1.07747428e-01
-1.44487989e+00 8.97662699e-01 8.23878169e-01 6.48616374e-01
-3.80733997e-01 3.55001003e-01 6.76128626e-01 6.03920400e-01
4.58452284e-01 -6.36920109e-02 -5.75512230e-01 -2.08611608e-01
-6.15478635e-01 -1.10105649e-01 8.78380656e-01 9.09462750e-01
7.43616879e-01 -2.43493114e-02 6.06705174e-02 7.08535314e-01
1.95522606e-01 1.72240332e-01 3.59330744e-01 -5.65866530e-01
4.99890834e-01 9.10947025e-01 -3.55823398e-01 -1.52042258e+00
-5.22453189e-01 -6.48593843e-01 -1.26912677e+00 -5.26544273e-01
2.99826294e-01 1.59208302e-03 -5.95768869e-01 1.69139993e+00
7.48989761e-01 6.91070557e-01 -1.22184433e-01 6.63046896e-01
8.27419341e-01 9.66471434e-01 -1.76422819e-01 -3.87668222e-01
1.18423951e+00 -9.90963042e-01 -4.44215059e-01 1.48718327e-01
9.59204197e-01 -5.58564663e-01 6.46014452e-01 1.43246740e-01
-7.71364391e-01 -2.60098010e-01 -9.05424774e-01 3.58584411e-02
-4.32448089e-01 -4.35951829e-01 9.51117754e-01 1.79307297e-01
-1.18293905e+00 9.30684507e-01 -9.88061428e-01 -6.59161627e-01
1.42634407e-01 6.97280169e-01 -5.27662277e-01 -5.26180454e-02
-1.31951857e+00 4.01262522e-01 2.55529433e-01 3.56069803e-02
-7.29044139e-01 -7.54544795e-01 -6.24870539e-01 -9.07325521e-02
4.59200472e-01 -4.42294359e-01 7.68019259e-01 -7.53512502e-01
-1.09688771e+00 2.95554966e-01 -3.48330975e-01 -5.82526207e-01
2.01891184e-01 1.49466023e-02 -7.27195263e-01 7.96777308e-02
1.86480749e-02 2.34746426e-01 5.53177595e-01 -7.67509520e-01
-3.79483491e-01 -2.47638762e-01 1.61031023e-01 4.92744446e-02
-9.59110081e-01 -1.56436577e-01 -7.95985997e-01 -8.05451751e-01
3.13607574e-01 -1.04494250e+00 -4.65362787e-01 3.56164068e-01
-4.43579465e-01 -3.82236242e-01 1.11070478e+00 -2.21130952e-01
1.57624650e+00 -1.91551745e+00 -5.01770945e-03 6.45780921e-01
6.52045429e-01 3.01097155e-01 -3.57888967e-01 7.12949157e-01
-7.25297630e-02 1.20898411e-01 1.30709827e-01 -2.63212800e-01
-4.49615568e-01 5.14284074e-01 -1.81980893e-01 4.88224477e-01
-2.88977265e-01 8.26807022e-01 -1.06283677e+00 -4.65217412e-01
1.12841666e-01 6.35150135e-01 -3.39562356e-01 -9.35592279e-02
-1.48401584e-03 2.10694849e-01 -7.74559259e-01 5.44113517e-01
3.89899045e-01 -6.57802165e-01 6.69500589e-01 -3.51303786e-01
1.70167685e-02 4.71125633e-01 -1.08824933e+00 1.58865249e+00
-2.30257139e-01 4.19907510e-01 -8.58561993e-02 -1.18802214e+00
9.68337655e-01 2.70438492e-01 9.37895477e-01 -7.22398818e-01
4.97548804e-02 2.18251437e-01 1.48159415e-01 6.34687692e-02
2.82792777e-01 3.44144166e-01 6.61302591e-03 7.36159861e-01
-2.39880219e-01 6.81778371e-01 1.20171323e-01 5.53410172e-01
1.63602233e+00 -4.16223377e-01 1.48735955e-01 -3.26217592e-01
4.80453432e-01 -1.31565481e-01 8.59122634e-01 5.40852547e-01
-1.15118735e-01 9.55315456e-02 6.98124588e-01 -8.84659052e-01
-7.87393391e-01 -7.96804786e-01 1.24374054e-01 1.17179406e+00
2.76457965e-01 -9.26939368e-01 -1.34845763e-01 -7.33817101e-01
2.89026856e-01 2.27623299e-01 -7.42142797e-01 -2.65225172e-01
-7.01256692e-01 -7.03658044e-01 3.97258133e-01 5.82892656e-01
5.00316083e-01 -6.09571278e-01 1.12959214e-01 4.17556912e-01
-4.13528495e-02 -8.75754714e-01 -6.31750822e-01 1.48075759e-01
-1.28470504e+00 -1.16577208e+00 -8.95129815e-02 -6.92069709e-01
6.68046117e-01 5.73042274e-01 9.41765189e-01 5.57986081e-01
-1.61707886e-02 4.19154502e-02 -4.60428834e-01 3.51763636e-01
-7.05459639e-02 3.43075216e-01 3.77703160e-01 1.66270837e-01
2.56037354e-01 -1.13671660e+00 -8.50474179e-01 5.85715055e-01
-6.62308097e-01 -4.06651087e-02 4.04506147e-01 8.62911046e-01
6.53119683e-01 1.69889331e-01 5.19962013e-01 -9.28475380e-01
4.09770608e-01 -1.02127504e+00 -5.12466669e-01 1.29642874e-01
-8.48065674e-01 -4.33224812e-02 8.72416317e-01 -8.32647145e-01
-3.91299754e-01 -5.96721135e-02 2.64891297e-01 -6.23970687e-01
4.40019041e-01 7.93389082e-01 3.33206683e-01 -1.12342998e-01
4.21612263e-01 2.34631911e-01 6.65923581e-02 -3.28389555e-01
1.85936853e-01 2.97829807e-01 9.61918980e-02 -7.02805459e-01
1.03516877e+00 5.29308259e-01 2.59267181e-01 -5.83892703e-01
-6.44412696e-01 -5.87999344e-01 -5.31469822e-01 -7.41099715e-02
1.90106988e-01 -8.61250043e-01 -8.24733913e-01 2.75762230e-01
-9.26723480e-01 -4.06626791e-01 -4.01690342e-02 3.38040113e-01
1.25330076e-01 2.80726522e-01 -9.52575684e-01 -3.64711523e-01
-5.26699901e-01 -7.35121906e-01 4.87567008e-01 -1.07794091e-01
-1.71131387e-01 -1.29426086e+00 -8.36152658e-02 6.91493973e-02
5.93236387e-01 2.75950313e-01 8.85171235e-01 -8.02933097e-01
-6.44480050e-01 -2.40844578e-01 -3.00145060e-01 -9.86543819e-02
2.25140005e-01 1.19862713e-01 -3.68855536e-01 -6.39707088e-01
-5.05042672e-01 2.53485958e-03 8.15219998e-01 7.15481415e-02
1.06099212e+00 -3.64094079e-01 -8.33180785e-01 6.45499587e-01
1.30594492e+00 1.86422512e-01 3.29813987e-01 6.69590756e-02
8.65974665e-01 3.44576985e-01 6.38365269e-01 5.42372108e-01
5.77414274e-01 6.25552893e-01 5.54969132e-01 9.92657542e-02
1.67550772e-01 -3.30728620e-01 3.68605554e-01 1.35593867e+00
-1.14096448e-01 -3.59724462e-01 -9.80751932e-01 8.70122373e-01
-2.20050216e+00 -7.85157681e-01 -3.20264369e-01 2.05439830e+00
8.53286862e-01 3.96520764e-01 2.52474248e-01 1.20158374e-01
8.35511029e-01 4.49682266e-01 -8.06755424e-01 -9.17746946e-02
1.81471370e-02 -1.48032587e-02 6.44553363e-01 3.85568231e-01
-8.92513156e-01 1.07866395e+00 5.12000704e+00 1.05258322e+00
-1.16836190e+00 2.29057476e-01 6.24576867e-01 -1.67288765e-01
-2.88032919e-01 2.53339857e-01 -7.78495491e-01 4.16271687e-01
1.18272185e+00 -3.67595434e-01 5.21951437e-01 6.20522559e-01
2.73028493e-01 1.67120054e-01 -1.09426892e+00 8.79601717e-01
-4.60377097e-01 -1.64346170e+00 2.02059343e-01 2.78713465e-01
6.16777301e-01 4.45565969e-01 1.52441105e-02 1.61660597e-01
4.84858394e-01 -8.45181942e-01 2.45298520e-01 3.61957282e-01
7.97533751e-01 -7.02214658e-01 6.33052170e-01 2.39021897e-01
-1.94761419e+00 -9.02618766e-02 -2.92018592e-01 -2.19888315e-01
9.05913592e-04 1.01519656e+00 -6.80720091e-01 5.35943270e-01
8.20964575e-01 1.32985044e+00 -2.02676803e-01 7.46478915e-01
1.36883587e-01 7.50539720e-01 -7.56315053e-01 -1.96194127e-01
4.20535445e-01 -2.30234966e-01 4.87590402e-01 8.67583096e-01
3.57532084e-01 3.23909104e-01 3.29551876e-01 3.03126574e-01
-4.07784969e-01 9.06525254e-02 -7.47579277e-01 -1.75830543e-01
1.04599595e+00 1.40739107e+00 -8.26782525e-01 -3.04499567e-01
-5.47210515e-01 8.34745526e-01 4.75734353e-01 2.17997745e-01
-8.90985370e-01 -3.12496841e-01 8.91513646e-01 3.66806388e-01
3.38357627e-01 -5.14361620e-01 2.15890561e-03 -9.44404364e-01
3.58670130e-02 -6.14908516e-01 5.45915186e-01 -2.50624359e-01
-1.42551696e+00 7.50747740e-01 -2.24225879e-01 -1.27511716e+00
6.42939582e-02 -3.00027698e-01 -6.16624415e-01 5.43909729e-01
-1.24662328e+00 -1.10087681e+00 -1.81379646e-01 9.99309838e-01
1.32540777e-01 -2.04612732e-01 8.03452075e-01 5.38324654e-01
-7.40912735e-01 6.58185720e-01 2.89141029e-01 4.17156011e-01
5.41795492e-01 -8.61538291e-01 6.21771336e-01 6.08756959e-01
2.48279512e-01 9.56587374e-01 3.34611267e-01 -8.52191150e-01
-1.72853625e+00 -1.39110410e+00 9.49167073e-01 -8.93383101e-02
1.19427204e+00 -3.76815736e-01 -8.98521483e-01 7.26145089e-01
-1.16670161e-01 7.37416208e-01 5.71088493e-01 3.63591343e-01
-7.37800717e-01 -5.85525155e-01 -1.01459873e+00 7.08656967e-01
1.59088004e+00 -5.62588215e-01 -4.49564978e-02 5.12480140e-01
9.52689230e-01 -1.40079364e-01 -1.36935008e+00 2.69780040e-01
5.17600894e-01 -6.79098845e-01 1.02061403e+00 -5.63423932e-01
8.77651498e-02 -3.05031359e-01 -9.09785926e-02 -9.71854389e-01
-5.30004144e-01 -9.03855443e-01 -5.94639063e-01 1.15308869e+00
3.54221821e-01 -1.09680140e+00 9.56275284e-01 2.46411547e-01
2.47484699e-01 -1.08667541e+00 -1.03422236e+00 -8.00297558e-01
-2.42336690e-01 -4.48589355e-01 6.90393150e-01 1.46315813e+00
-1.08393282e-02 4.71827388e-01 -3.92924041e-01 1.44277409e-01
5.24282932e-01 2.58263797e-01 6.58922613e-01 -1.45370793e+00
-2.33526245e-01 -4.42133456e-01 -4.84611332e-01 -1.25798249e+00
4.83800359e-02 -1.03838980e+00 -7.24525034e-01 -1.39790499e+00
1.35818899e-01 -1.04850209e+00 -6.82499886e-01 9.10560846e-01
3.01991791e-01 1.27016217e-01 -1.96257517e-01 4.75725323e-01
-6.92984939e-01 5.55764914e-01 1.14932775e+00 -1.44977361e-01
-2.41026342e-01 -1.12355009e-01 -5.30875266e-01 4.54969317e-01
9.26399648e-01 -6.21578753e-01 -6.73777461e-01 -4.62361723e-01
3.38183850e-01 3.08893651e-01 2.29144663e-01 -6.75014853e-01
7.02839196e-01 -1.41928792e-01 7.61532709e-02 -5.76871455e-01
5.49571693e-01 -9.58301306e-01 3.43147576e-01 5.99603057e-01
-1.24069259e-01 3.81444603e-01 4.85518686e-02 1.03682017e+00
-6.82663023e-02 3.86132002e-01 3.29828471e-01 1.28969505e-01
-6.73189461e-01 9.07715321e-01 4.65623243e-03 -1.09268941e-01
1.17481029e+00 -2.05123663e-01 -4.47047979e-01 -4.14924592e-01
-6.19364560e-01 4.94818032e-01 2.41702586e-01 4.75317299e-01
6.49694264e-01 -1.48177564e+00 -5.05073845e-01 4.04521115e-02
-4.03871574e-02 -4.08801250e-02 1.61480114e-01 1.05924869e+00
-4.47345376e-01 2.47427776e-01 -1.95082743e-02 -6.10179245e-01
-1.33066595e+00 5.17876208e-01 2.91971445e-01 -5.45399487e-01
-7.46722579e-01 8.39964569e-01 -2.37130210e-01 -3.59492779e-01
6.83430880e-02 -2.37238839e-01 -1.31357342e-01 5.58493054e-03
2.35326782e-01 4.01060492e-01 -4.75534014e-02 -5.93068779e-01
-4.84532386e-01 5.45486569e-01 -3.54074150e-01 2.68471539e-01
1.60344255e+00 -1.84572116e-02 -5.84256768e-01 5.22324443e-01
1.42876697e+00 -2.48357117e-01 -8.17918360e-01 -9.24920261e-01
5.18445335e-02 -4.11890119e-01 2.52657861e-01 -2.78675318e-01
-1.74231076e+00 4.17734712e-01 1.99940607e-01 3.53011489e-01
1.09086204e+00 -6.12278953e-02 1.12171745e+00 6.34786069e-01
7.12900281e-01 -7.12590814e-01 1.78266745e-02 6.76782906e-01
5.78277886e-01 -1.01701581e+00 2.37809852e-01 -6.17679358e-01
-1.73431218e-01 1.07015407e+00 6.48191452e-01 -1.67533204e-01
1.26035583e+00 1.44220665e-01 -3.25414985e-01 -3.84320885e-01
-1.30674958e+00 1.17263794e-01 1.04215555e-01 3.23107004e-01
3.60938311e-01 1.61527023e-01 -2.32179940e-01 3.36487979e-01
1.16969258e-01 -4.60765153e-01 8.42824727e-02 9.56779063e-01
-6.61006868e-02 -1.50151348e+00 1.20921321e-01 8.07295918e-01
-1.04898192e-01 -2.87264496e-01 -2.08316550e-01 7.30914295e-01
-1.47746429e-01 9.07479882e-01 2.43862733e-01 -8.13137591e-01
6.63217083e-02 -3.93945307e-01 -3.64636704e-02 -5.21962702e-01
-8.69329572e-01 5.92780411e-02 2.00818673e-01 -8.16704273e-01
-4.65360165e-01 -6.84554458e-01 -1.34994268e+00 -1.05997145e+00
-2.86353081e-01 2.72785783e-01 2.24741340e-01 5.57233393e-01
6.42932177e-01 3.54891062e-01 7.65379429e-01 -4.89931405e-01
-1.52326867e-01 -7.72793114e-01 -5.59798002e-01 1.34444639e-01
2.15510949e-01 -7.30309665e-01 -2.53127187e-01 -4.12226170e-01] | [7.22967529296875, 6.021739482879639] |
e911c154-8ab8-45d7-86dd-abc95731bdc8 | diverse-multi-answer-retrieval-with-1 | 2211.16029 | null | https://arxiv.org/abs/2211.16029v1 | https://arxiv.org/pdf/2211.16029v1.pdf | Diverse Multi-Answer Retrieval with Determinantal Point Processes | Often questions provided to open-domain question answering systems are ambiguous. Traditional QA systems that provide a single answer are incapable of answering ambiguous questions since the question may be interpreted in several ways and may have multiple distinct answers. In this paper, we address multi-answer retrieval which entails retrieving passages that can capture majority of the diverse answers to the question. We propose a re-ranking based approach using Determinantal point processes utilizing BERT as kernels. Our method jointly considers query-passage relevance and passage-passage correlation to retrieve passages that are both query-relevant and diverse. Results demonstrate that our re-ranking technique outperforms state-of-the-art method on the AmbigQA dataset. | ['Manish Shrivastava', 'Nikhil Rayaprolu', 'Poojitha Nandigam'] | 2022-11-29 | diverse-multi-answer-retrieval-with | https://aclanthology.org/2022.coling-1.194 | https://aclanthology.org/2022.coling-1.194.pdf | coling-2022-10 | ['point-processes', 'open-domain-question-answering'] | ['methodology', 'natural-language-processing'] | [-2.83719242e-01 -1.12223633e-01 1.92234024e-01 -2.95096189e-01
-2.24864054e+00 -1.27912307e+00 5.86035430e-01 4.21134442e-01
-3.78261507e-01 1.07537901e+00 3.70546579e-01 -5.44916630e-01
-7.21845269e-01 -8.11533988e-01 -4.81995225e-01 -1.59970984e-01
3.51558715e-01 1.23186934e+00 9.56304550e-01 -8.76786888e-01
5.75433850e-01 5.47683835e-02 -1.26065290e+00 6.24185026e-01
1.49860132e+00 6.56011343e-01 1.17602095e-01 1.22526562e+00
-8.26472104e-01 7.87887394e-01 -8.84413302e-01 -5.60968935e-01
8.35744757e-03 -6.00453675e-01 -1.60992348e+00 -6.36667728e-01
8.93061817e-01 -1.25748590e-01 -4.71819416e-02 8.17733109e-01
4.78524357e-01 5.34663975e-01 8.20028067e-01 -7.05940843e-01
-1.12620199e+00 -1.23145849e-01 7.12660840e-03 8.66853118e-01
1.21784174e+00 -3.02981436e-01 1.52000535e+00 -1.09876525e+00
4.83497024e-01 1.39926445e+00 1.37694135e-01 4.21323240e-01
-1.15500474e+00 -5.81886172e-02 -1.77175686e-01 5.31532168e-01
-1.02300322e+00 -1.27586901e-01 4.00868088e-01 -6.56258315e-02
8.40906382e-01 6.49509311e-01 -3.04374937e-02 6.39698029e-01
3.11435282e-01 6.87253237e-01 1.01201797e+00 -5.05367935e-01
2.34201029e-01 -2.20026076e-01 7.28837848e-01 3.67669851e-01
-4.17817421e-02 -4.56797957e-01 -3.84356111e-01 -7.79798865e-01
1.99381039e-01 -3.57278854e-01 -3.25388253e-01 1.24988459e-01
-9.18668807e-01 1.06774402e+00 5.20173311e-02 5.88463843e-02
-5.58229983e-01 -2.53668159e-01 1.80338800e-01 9.45469260e-01
1.97277397e-01 9.19765294e-01 -6.43770397e-01 -2.14012519e-01
-6.31001949e-01 7.78292537e-01 1.43011248e+00 6.65329874e-01
7.75435925e-01 -8.94421339e-01 -8.68285477e-01 1.12102842e+00
2.78923124e-01 5.96876144e-01 4.20137197e-01 -1.13949692e+00
6.86093569e-01 6.52356744e-01 6.13807201e-01 -6.76808536e-01
3.03065479e-02 -6.66799396e-02 -4.20011133e-02 -5.34258068e-01
4.59307373e-01 -2.10206471e-02 -7.72456706e-01 1.29030216e+00
4.99657273e-01 -3.94277483e-01 4.72554535e-01 9.37103033e-01
1.18823099e+00 9.80987132e-01 1.12294592e-01 2.33207271e-01
1.74800766e+00 -1.23281884e+00 -7.55391598e-01 -1.43404946e-01
2.39364848e-01 -1.35072196e+00 1.32925951e+00 3.32880840e-02
-1.24901772e+00 -4.73305404e-01 -4.98371273e-01 -5.68000317e-01
-2.54573524e-01 -1.21281713e-01 8.47333297e-02 4.22298610e-01
-1.20302701e+00 -1.01693824e-01 -7.74394535e-03 -2.91862160e-01
-2.05832452e-01 4.66200225e-02 -4.59564552e-02 -4.72632527e-01
-1.63692474e+00 1.09090853e+00 7.00399429e-02 -3.10573936e-01
-7.36038148e-01 -6.97784066e-01 -4.93350416e-01 3.03212196e-01
5.05776703e-01 -1.20216501e+00 1.73693514e+00 -3.12353015e-01
-1.23197317e+00 5.43901265e-01 -7.39768267e-01 -3.70640084e-02
1.46222822e-02 -4.61166769e-01 -6.15683198e-01 9.49822485e-01
6.40303135e-01 3.65850925e-01 7.11390674e-01 -1.20939910e+00
-7.76047111e-01 -3.14357996e-01 6.42927647e-01 5.51178217e-01
3.00352961e-01 6.91525564e-02 -4.52687442e-01 -6.32728636e-02
4.37631607e-01 -4.81318235e-01 -3.95403430e-03 -5.25258482e-01
-1.56069934e-01 -7.31174290e-01 4.86990392e-01 -9.77204263e-01
1.37270963e+00 -1.46554744e+00 -1.28591448e-01 2.34495495e-02
1.69250354e-01 2.31398061e-01 -4.85954016e-01 1.04712510e+00
4.27144766e-01 1.63856596e-01 -2.75976886e-03 2.34469518e-01
2.50130624e-01 2.64777631e-01 -8.07806849e-01 -1.88488901e-01
3.61677736e-01 1.18102741e+00 -1.09215558e+00 -7.64796615e-01
-4.13696170e-01 3.77679989e-02 -3.03708017e-01 4.66899782e-01
-6.54388547e-01 3.27564031e-01 -1.01233435e+00 9.09980059e-01
5.22108793e-01 -4.31280106e-01 -4.34691936e-01 2.15917498e-01
3.20373029e-01 7.38876939e-01 -7.64650345e-01 1.45174468e+00
-4.93487805e-01 3.63668531e-01 -7.86919296e-02 -3.65766585e-01
7.34504640e-01 3.49144489e-01 -9.81197655e-02 -1.00382257e+00
-4.25811768e-01 5.10866046e-01 -2.03479990e-01 -8.23556781e-01
1.05077314e+00 -1.52815834e-01 -9.41703692e-02 4.57473069e-01
1.23754404e-01 -4.59257931e-01 6.11686170e-01 5.39308071e-01
1.35805416e+00 -1.55758604e-01 1.04678050e-01 -2.98137099e-01
1.02034593e+00 3.21548849e-01 -1.63916066e-01 1.22580349e+00
-3.37097436e-01 6.24579370e-01 4.94802296e-01 1.09718703e-01
-6.90510094e-01 -1.57634842e+00 -8.16894546e-02 1.44053543e+00
2.72517741e-01 -1.51416227e-01 -4.23002452e-01 -6.74289286e-01
-1.21214613e-02 8.24148476e-01 -3.93777519e-01 1.45169184e-01
-4.40493584e-01 -2.45130032e-01 5.71608484e-01 1.22085355e-01
3.18794698e-01 -9.59017992e-01 -7.97655135e-02 2.63361603e-01
-9.45777714e-01 -7.67241836e-01 -6.15606904e-01 -1.39776841e-01
-9.09668028e-01 -1.33293450e+00 -8.84332657e-01 -9.50114191e-01
3.10595185e-01 5.43129683e-01 1.78418255e+00 1.33435175e-01
1.89652875e-01 1.10311520e+00 -5.78756213e-01 -4.85381112e-02
-3.45825911e-01 2.53120303e-01 -6.70944750e-01 -3.72989058e-01
8.15955102e-01 -1.66031122e-01 -8.88069093e-01 3.59744161e-01
-1.25657475e+00 -9.52053785e-01 4.04813141e-01 1.02682829e+00
6.85676455e-01 -5.66410065e-01 1.28643692e+00 -6.37197793e-01
1.71763313e+00 -9.23474193e-01 -4.19003904e-01 1.01242507e+00
-4.60248053e-01 3.17347288e-01 5.48629761e-01 -2.17662692e-01
-1.38385355e+00 -7.97281086e-01 -2.88375825e-01 1.82803318e-01
-1.71908036e-01 6.67866528e-01 2.56337076e-01 -8.43114629e-02
8.89362574e-01 2.52649695e-01 -3.83318305e-01 -3.99260461e-01
6.16807997e-01 5.16754687e-01 1.40850693e-01 -8.62212718e-01
7.01690853e-01 2.16828525e-01 -2.65625596e-01 -7.89867401e-01
-1.16221964e+00 -1.41828525e+00 -2.75610059e-01 -1.29975259e-01
7.62294233e-01 -6.25228286e-01 -6.02332056e-01 -1.91213921e-01
-1.33607411e+00 4.40808624e-01 -2.96107471e-01 2.99592644e-01
-4.08181995e-01 7.28771627e-01 -7.68285096e-01 -7.85667479e-01
-5.26184857e-01 -8.44369769e-01 1.20958221e+00 6.42673016e-01
-2.89316595e-01 -1.14212465e+00 7.21259058e-01 9.48728383e-01
6.14678919e-01 -4.69730943e-01 1.16109979e+00 -1.25547302e+00
-9.31516826e-01 -3.06994259e-01 -1.52210265e-01 2.26996541e-02
-1.82416551e-02 -4.55502063e-01 -7.51168787e-01 -3.09798289e-02
2.32634559e-01 -7.20453441e-01 9.45511401e-01 1.29817531e-01
3.86178434e-01 -2.99465537e-01 2.60867029e-01 -3.32382500e-01
1.41298926e+00 1.96821745e-02 7.34727561e-01 4.48455304e-01
3.66251543e-02 8.17725658e-01 7.51913965e-01 -9.50980186e-02
7.28525698e-01 1.24814354e-01 1.77303683e-02 3.83685887e-01
1.13510378e-01 -2.42345661e-01 9.87596512e-02 1.00886405e+00
4.33198601e-01 -4.43782449e-01 -9.16554391e-01 1.01725495e+00
-1.69477522e+00 -1.19047713e+00 -5.13477325e-01 2.06371331e+00
9.68988419e-01 -3.44773889e-01 -1.55967578e-01 -4.75681335e-01
4.70902264e-01 7.90385753e-02 -3.38494956e-01 -5.15037060e-01
-3.14387411e-01 7.06049323e-01 -1.68195829e-01 9.63164091e-01
-7.06420600e-01 8.17709029e-01 6.98097372e+00 8.73602331e-01
-8.33948255e-02 8.29189047e-02 4.09698188e-01 3.59240472e-01
-9.59407985e-01 3.67247522e-01 -7.89192021e-01 -2.51987632e-02
1.01458299e+00 -5.00481009e-01 2.48969868e-01 4.68262225e-01
-2.72380263e-01 -7.50711620e-01 -8.97170544e-01 4.72341180e-01
2.92537928e-01 -1.02213955e+00 5.74994445e-01 -4.41271603e-01
9.12004292e-01 -9.54867154e-02 -7.49348570e-03 6.85312927e-01
4.40601170e-01 -9.94980216e-01 1.46255746e-01 9.76356983e-01
1.00861505e-01 -5.64526677e-01 8.50466371e-01 4.29871768e-01
-9.52604651e-01 9.01969299e-02 -5.54780245e-01 1.39235824e-01
2.70295739e-01 2.36852825e-01 -8.06170464e-01 9.33745444e-01
7.13666260e-01 -3.01829427e-01 -7.93604553e-01 1.20663130e+00
-2.25014627e-01 8.74796808e-01 -2.40757540e-01 -5.05523622e-01
5.40938556e-01 -4.09356982e-01 6.50317490e-01 8.50289643e-01
2.40718707e-01 5.33223331e-01 5.53824194e-02 7.67125607e-01
-6.49623126e-02 4.95621592e-01 -3.24702919e-01 5.85572086e-02
5.28320849e-01 8.24929774e-01 -1.37965024e-01 -3.36845338e-01
-5.87246656e-01 1.18828511e+00 3.58530253e-01 8.07820380e-01
-2.89117604e-01 -5.92150211e-01 2.84423947e-01 -3.88480544e-01
2.90424854e-01 -2.04683319e-02 1.86503798e-01 -1.16268528e+00
4.36254442e-01 -1.12878489e+00 9.61967826e-01 -9.33522165e-01
-1.98815656e+00 6.16599262e-01 -4.10105549e-02 -9.30577338e-01
-4.51895177e-01 -3.30190688e-01 -4.70204741e-01 1.51355398e+00
-2.01652598e+00 -7.15745687e-01 9.58646536e-02 6.72941089e-01
6.94606066e-01 4.19344120e-02 1.00604808e+00 2.03161389e-01
3.00269872e-01 1.86239049e-01 4.96480942e-01 -1.94997966e-01
1.02540922e+00 -1.52900481e+00 4.36661690e-02 5.91482699e-01
2.18301430e-01 1.05416298e+00 5.22948861e-01 -6.01727307e-01
-1.43764496e+00 -5.02633452e-01 1.51878643e+00 -1.04691100e+00
8.08995187e-01 4.48622584e-01 -1.34573436e+00 3.43636394e-01
4.91979092e-01 -4.86952186e-01 1.05717301e+00 2.96144158e-01
-3.69617999e-01 -2.84829717e-02 -1.02205956e+00 6.34784639e-01
1.07638054e-01 -1.05799389e+00 -1.60469520e+00 3.71458054e-01
9.03306663e-01 -3.22491288e-01 -6.63548350e-01 1.94148213e-01
1.34717688e-01 -7.60272801e-01 1.17723930e+00 -9.62947845e-01
3.46916139e-01 -3.42239767e-01 -3.65179360e-01 -1.12235820e+00
-1.61585689e-01 -4.39235777e-01 -2.49879852e-01 9.14197981e-01
7.83050656e-01 -7.12390065e-01 3.63288999e-01 7.00777769e-01
-1.92137007e-02 -6.86722875e-01 -1.25366235e+00 -6.25823200e-01
5.57782829e-01 2.27478769e-04 5.62551558e-01 5.67227185e-01
-8.84028226e-02 8.89706731e-01 2.20256642e-01 3.28403920e-01
2.51297057e-01 4.64251965e-01 4.04853284e-01 -1.08613265e+00
-1.13255233e-01 -3.34178627e-01 1.75503880e-01 -1.56400156e+00
-2.75631920e-02 -6.14812315e-01 1.42390326e-01 -2.00198531e+00
1.91499278e-01 -7.93882087e-02 -2.47832760e-01 -3.53184640e-01
-9.33359683e-01 -1.51362821e-01 -1.73383966e-01 3.50674689e-01
-1.21160007e+00 6.13319874e-01 1.57414281e+00 -1.51879594e-01
-1.23464139e-02 6.58824295e-02 -8.03872347e-01 1.46593302e-01
8.91901791e-01 -3.85520011e-01 -6.53964996e-01 -5.19471884e-01
6.50139451e-01 6.30080342e-01 3.01384240e-01 -6.33956075e-01
5.43621838e-01 -1.27420589e-01 9.57241356e-02 -9.24174130e-01
3.50089639e-01 -4.13890153e-01 -5.20343900e-01 -5.55609632e-03
-6.89664841e-01 5.77509046e-01 1.45801634e-01 9.61930871e-01
-7.14078188e-01 -8.10557425e-01 8.12123418e-02 -4.43443209e-01
-5.37383199e-01 6.05529249e-02 -5.12559891e-01 9.67002153e-01
4.40116823e-01 1.62902743e-01 -7.28236318e-01 -7.98215151e-01
-3.71072084e-01 8.36746991e-01 8.44562054e-02 4.71470177e-01
8.25818062e-01 -1.19180858e+00 -1.04663444e+00 -6.33148849e-01
4.40288514e-01 -2.48944789e-01 4.42083150e-01 4.44506437e-01
-6.54353201e-01 1.04118133e+00 3.59809875e-01 -3.69867831e-01
-1.03571475e+00 2.02435613e-01 3.45411837e-01 -7.48419344e-01
-3.63105461e-02 8.74964118e-01 -1.08253837e-01 -9.09464598e-01
-9.12683830e-02 -1.01179637e-01 -5.93058944e-01 2.33599156e-01
6.37542367e-01 4.16743040e-01 1.92869797e-01 -3.64076734e-01
-8.12478364e-02 4.76720840e-01 -1.56702220e-01 -6.60971880e-01
5.78280330e-01 -5.18855274e-01 -4.99270767e-01 5.43907166e-01
1.15027893e+00 2.69578993e-01 -4.35968518e-01 -4.57171917e-01
2.95341939e-01 -5.19526005e-01 -3.92890960e-01 -1.25042987e+00
-4.05286402e-02 7.74325252e-01 3.80709797e-01 5.77530801e-01
1.03501832e+00 4.58720386e-01 1.05468512e+00 1.05026245e+00
2.62691319e-01 -1.08467841e+00 4.72727597e-01 1.15734017e+00
1.05399883e+00 -1.36224163e+00 -4.36376840e-01 -1.72275364e-01
-5.49325764e-01 1.11205351e+00 6.15658522e-01 1.21609911e-01
2.54554570e-01 -7.42881775e-01 5.51496983e-01 -6.43358350e-01
-1.01639557e+00 -5.27352870e-01 8.82891893e-01 2.50732243e-01
3.82479519e-01 -1.72015026e-01 -8.25830579e-01 4.95644212e-01
-1.47738010e-01 -4.37249273e-01 3.01162213e-01 1.03543544e+00
-8.55615735e-01 -1.13635278e+00 -7.34854639e-01 6.18540585e-01
-6.30766869e-01 -5.52893460e-01 -5.20112634e-01 3.79557222e-01
-7.29524672e-01 1.59149373e+00 -3.45103204e-01 1.49446815e-01
3.79801512e-01 6.46900117e-01 4.37568575e-01 -6.77241445e-01
-7.15260029e-01 -3.54827642e-01 2.78907806e-01 -2.00102746e-01
-2.28068501e-01 -3.58772665e-01 -1.09832835e+00 -5.54485321e-02
-5.45829952e-01 9.61422920e-01 2.36688569e-01 9.62320030e-01
4.03710872e-01 2.95712948e-01 3.50090176e-01 3.66859317e-01
-1.15873158e+00 -9.65627551e-01 -2.75717974e-01 3.91073108e-01
7.10829496e-01 -2.09697112e-01 -3.35758805e-01 -2.68007338e-01] | [11.364776611328125, 7.8683905601501465] |
f737e1d5-6d72-4c3f-8d0d-2f46e073cdec | effective-modeling-of-encoder-decoder | 1911.09886 | null | https://arxiv.org/abs/1911.09886v1 | https://arxiv.org/pdf/1911.09886v1.pdf | Effective Modeling of Encoder-Decoder Architecture for Joint Entity and Relation Extraction | A relation tuple consists of two entities and the relation between them, and often such tuples are found in unstructured text. There may be multiple relation tuples present in a text and they may share one or both entities among them. Extracting such relation tuples from a sentence is a difficult task and sharing of entities or overlapping entities among the tuples makes it more challenging. Most prior work adopted a pipeline approach where entities were identified first followed by finding the relations among them, thus missing the interaction among the relation tuples in a sentence. In this paper, we propose two approaches to use encoder-decoder architecture for jointly extracting entities and relations. In the first approach, we propose a representation scheme for relation tuples which enables the decoder to generate one word at a time like machine translation models and still finds all the tuples present in a sentence with full entity names of different length and with overlapping entities. Next, we propose a pointer network-based decoding approach where an entire tuple is generated at every time step. Experiments on the publicly available New York Times corpus show that our proposed approaches outperform previous work and achieve significantly higher F1 scores. | ['Hwee Tou Ng', 'Tapas Nayak'] | 2019-11-22 | null | null | null | null | ['joint-entity-and-relation-extraction'] | ['natural-language-processing'] | [ 0.2119807 0.6159588 -0.19698708 -0.45471698 -0.76540697 -0.65523124
0.32015967 0.8602835 -0.38539782 1.0919371 0.48316276 -0.34265032
-0.02811869 -1.1613059 -0.893169 -0.09423227 -0.14599106 0.9645997
0.4483424 -0.24546425 -0.09353817 0.16337223 -1.2865055 0.66716677
0.9191947 0.6916563 0.249408 0.56590694 -0.7950328 0.8817389
-0.66513383 -1.2285706 0.18819834 -0.48075104 -1.1160526 -0.0922793
-0.09676333 -0.12160069 -0.32355246 1.011471 0.2558863 -0.0682069
0.41335276 -1.003052 -0.5152016 1.2235998 -0.5164585 0.14168222
0.5610215 -0.6072378 1.3509557 -0.7124257 0.8315434 1.2118506
0.38801837 0.2386313 -0.84790987 -0.4751191 -0.03922564 0.1645029
-1.4694331 -0.50405496 0.10443547 -0.17320983 1.7320564 0.4046423
0.32964805 0.5006616 0.18349728 0.51608056 0.13485351 -0.5842481
-0.11893824 0.1256702 0.3132692 0.3952957 0.47955722 -0.4431356
-0.6797807 -0.19816656 0.22925043 -0.21591459 -0.04583479 0.07410844
-1.262023 0.5444788 0.19743507 0.2555836 -0.64008224 -0.21767336
0.32878697 0.2346199 0.5719913 0.26769996 -0.7596202 -0.23165648
-0.51114196 0.43962035 1.3298687 1.4083495 0.818696 -0.9453586
-0.14929639 0.7604274 0.23261327 0.09043219 0.36294124 -0.45124033
1.2417843 1.0433036 0.22563946 -1.1158845 -0.3096132 -0.08371612
-0.57143927 -0.7393642 0.2370923 -0.44353718 -0.65908194 1.6056445
0.60241616 0.20513202 0.62485665 0.52300084 1.0643463 0.7771639
0.0743768 -0.391859 1.755051 -0.9169698 -1.0096024 -0.38060895
0.8521995 -0.88008255 0.20620234 -0.19408685 -1.125051 -0.18074173
-1.0126219 -0.2743926 -0.54011613 0.29077592 0.6927923 0.35687822
-0.51471883 0.5925197 -0.84360784 -0.40883443 0.08443679 0.54744375
-0.5398713 0.1706352 -1.5270836 0.978354 0.88253605 0.25309253
0.08196238 -0.19742072 -1.0816791 0.29209706 0.6212728 -0.77915967
1.3718499 -0.5428892 -1.0295221 0.56186545 -0.97997737 -0.7067232
0.03183423 -0.33359796 -0.6907652 -0.19517548 0.36367807 0.4364855
-0.03864313 -0.8470649 -0.99966127 -0.35938504 0.3107253 0.52565485
0.02341017 0.4284447 -0.70197964 -0.24496752 0.5164091 -0.85085654
-0.07348863 -0.7058519 -0.81897056 -0.48112708 0.26495603 -0.6786778
1.6646343 -1.8511789 0.02245712 0.04859463 0.26853105 0.32114384
0.18127048 0.95134765 -0.14555377 0.31526005 -0.32068312 -0.2961467
-0.08408574 0.42015734 -0.38538098 -0.07161094 0.61954135 1.1165297
-0.7289357 -0.59350634 -0.5144823 0.43942213 -0.10058915 0.31608617
-0.38983145 0.11233459 -0.45149532 0.37514144 0.75149924 -0.44818988
0.6125509 -0.1997966 -0.0887036 0.8716813 -1.3067366 1.342741
-0.09758535 0.37955028 -0.3927787 -0.71919644 0.97969085 0.74866766
0.2763024 -0.5947662 -0.02774538 0.39531705 0.2283085 -0.7667456
0.9321685 -0.04234008 0.10041048 0.40984944 0.0203289 0.4617818
0.7082374 0.35642117 1.28707 -0.0847882 0.5235237 0.4283791
0.53384864 0.10779475 0.9650975 0.40419123 0.32832268 0.46762025
0.9356906 -0.3069295 -1.0859044 -0.60033715 0.07230468 0.5406804
0.3276675 -0.87065524 -0.42731765 -0.70735806 -0.05695068 0.6132057
-0.5008159 0.16799112 -0.8870643 -0.7113299 0.45744792 0.37989846
0.33759192 -0.8518658 -0.4346923 0.5561825 -0.6576112 -1.5339072
-0.1127851 0.29685166 -0.7281996 -0.91749805 -0.3437504 -0.8464882
0.6798986 -0.07698481 1.2803202 0.03792385 0.24281286 -0.5985893
-0.58864814 -0.403475 -0.3685225 0.37729332 -0.24800682 0.03312554
0.84780353 -0.28821197 -0.09968466 -0.01994181 -0.8101716 0.30558646
0.5499429 0.5907509 0.4576829 0.32134867 0.55289924 -1.5977556
0.67108995 -0.89242816 -0.37788057 0.6735589 -0.49439225 0.32450685
0.42098737 -0.13097852 -1.0695349 0.2794015 -0.125883 0.28575426
-0.10503023 1.0057471 -0.33173653 0.6812515 0.36759 -0.06357009
-0.40005872 -0.5066201 0.2919442 0.7344694 0.28203267 -0.3384275
0.5110454 0.05222606 -0.21599923 -0.33190084 -0.95129585 -0.4765731
-0.87209755 0.31852752 0.8044541 -0.98499554 -0.67440027 0.04462361
-1.7915108 0.4054334 -0.12797335 0.55417174 0.046308 0.2807916
-0.6355363 -0.7120617 -0.37384266 -1.0748543 0.84191567 0.3737852
-0.49541232 -0.6778489 0.21719693 0.25445494 0.00742512 -0.09844055
1.1319957 -1.2537557 -0.6210506 -0.2710445 -0.3893138 -0.33959824
0.23079154 -0.12962763 -0.46102068 0.21729256 -0.44144645 0.06334962
0.6154771 -0.17860849 0.35171467 -0.5271206 -0.6882702 0.2001283
1.2703003 0.56953907 0.66990274 0.22015828 0.6513129 0.9767354
0.5441978 0.34986863 0.85545814 0.63306075 0.09328115 0.22126544
0.16770333 -0.39991373 0.15319292 0.9338829 0.15606008 -0.84026456
-0.8667761 0.6845962 -2.1182692 -0.78713065 -0.47960836 2.0942993
1.080963 0.07712258 -0.23589431 -0.08913506 0.8792367 -0.23294964
-0.4233682 -0.36476132 -0.258232 0.31135538 0.75613314 0.61942905
-1.1000545 1.1585838 5.853016 0.28713694 -0.66847134 -0.06879065
0.42883992 0.06066865 -0.4858195 0.4407282 -1.2802713 0.30106786
1.2840561 -0.50477695 0.01671145 0.17928259 -0.08115128 -0.37338352
-1.1950552 0.82953984 -0.05485135 -1.3293 -0.15380679 -0.14407705
0.4479439 -0.06276774 -0.62532717 0.12575606 0.42638493 -0.94754785
0.46073872 0.47758445 0.44767296 -0.8394138 1.0363005 0.4698343
-1.3837659 0.226396 -0.5495389 -0.12339744 0.46425736 0.76378155
-1.0048058 0.9853393 0.46857426 0.65752375 -0.35082668 0.9930739
-0.37161326 0.37905863 -0.35232067 -0.322797 -0.04814541 -0.22134514
0.4456246 1.274606 0.4271856 0.42946854 -0.00626904 0.6698428
-0.38117245 0.38829818 -0.5799242 -0.25791213 0.7409528 1.0603948
-0.8318796 -0.57919997 -0.7074483 1.1685756 0.7292482 0.34364238
-0.6387303 -0.6748807 0.58311176 -0.08563109 0.5130125 -0.13222475
-0.13061534 -1.3965986 0.5583962 -0.7435183 0.35580397 -0.48553452
-1.0250126 0.9793916 -0.16354553 -1.0699278 -0.5021346 -0.13588907
-0.02374631 1.174467 -1.3229358 -0.80638164 0.20742914 0.1277999
0.2577708 -0.09955677 1.0339848 0.615442 -0.77936417 0.52855766
-0.1199724 0.5798851 0.52830416 -1.0969994 0.9573845 1.0572008
0.5749067 0.98806435 0.58034956 -1.0184224 -1.3013074 -0.78490674
2.2281845 -0.19057733 0.56438047 -0.50095737 -1.0908086 0.974421
0.4982508 -0.33443692 0.7869424 0.28835404 -0.17463072 0.08761882
-0.83642185 0.45691893 0.86110955 -0.40727496 -0.71961474 0.4160364
1.1679115 -0.78773224 -0.759785 0.51036483 0.46048284 -0.6664102
0.5957827 -0.7856987 0.62875473 -0.46417215 -0.12030332 -1.1575803
0.06030535 -0.5372003 -0.6407073 1.6175724 1.2820276 -0.59387153
0.8632158 1.1346166 0.18359327 -0.84937155 -0.7728403 -0.49618748
-0.33093023 -0.3277328 1.0215491 0.7547841 0.30431297 0.9502466
-0.18148963 0.44722632 0.0241063 0.2779163 0.4841334 -1.1336595
-0.32480302 0.176514 -0.32257402 -1.4066621 0.1667088 -0.9187395
0.05399479 -1.9155717 0.2637212 -0.65166414 0.1432833 0.5853218
-0.31104904 -0.40906826 -0.09928901 0.23548803 -0.48994353 0.27608824
0.7839904 -0.05088957 -0.30017287 0.02872027 -0.8328243 0.40867877
0.47213796 -0.81646013 -0.44853953 -0.7898102 0.7629488 0.5242546
-0.28717366 -0.6333025 0.6592397 0.1657285 0.15246797 -0.73464507
0.22760135 -0.96023613 0.40490723 0.06696514 -0.32062715 0.31652302
0.03265743 0.4882264 -0.67076826 -0.43808967 0.03567974 -0.07750297
-0.5353952 0.19948557 -0.04098861 0.05571219 1.030262 0.13324343
-0.44585764 -0.36411428 -0.77313495 0.29625434 0.02001073 0.49140847
0.57515776 -1.1037223 -0.8625524 0.03478226 0.18422422 0.24690142
-0.10372321 0.5729086 -0.65797204 0.5895309 0.27747843 -0.01646417
-1.5679861 0.37549382 0.2041492 -0.6231175 -0.486923 1.1818426
-0.12404227 -0.5834009 0.01274576 -0.36604652 -0.58286905 0.24288087
0.60269356 -0.13765232 0.39447978 -1.0054871 -0.41653854 0.20211616
-0.41730478 -0.22104903 1.2378685 -0.2753607 -0.68936497 0.5072129
1.1980008 0.16351794 -0.51657516 -0.6042649 0.675267 -0.3822201
-0.6202051 -0.58452886 -0.87924653 0.47452122 -0.03099611 0.35749203
1.0157958 0.3694214 1.245902 0.37042972 0.2807559 -0.6918003
-0.6812619 0.9347761 0.3579257 -1.2918324 -0.03379803 -0.9974517
-0.81620073 1.1089145 0.4166695 0.28796405 0.54200816 0.5019603
0.08440001 -0.15978041 -1.1946032 -0.4344712 0.32300672 0.17205422
0.91584116 0.11835734 -0.85275763 0.756546 -0.44294444 -0.17199796
0.36422095 0.98967654 -0.25273943 -1.8496548 0.14477797 0.63491535
-0.89994675 -0.51449937 -0.47423357 0.21032768 0.3378372 1.2091081
0.25798053 -0.5491621 0.6004037 0.33180326 0.2247289 -0.9621013
-0.76521844 -0.09766208 0.75605565 -0.27980825 -0.44207996 -0.791858
-1.7083459 -0.23328431 -0.525309 0.4917641 0.46661356 1.1484454
0.6925421 0.39591953 0.3262309 0.1739784 0.12340182 -0.8979169
-0.5506197 0.21192743 0.21244855 -0.3539139 0.30090767 0.20879443] | [9.24360179901123, 8.704204559326172] |
6a3bbc22-73b1-446b-85b9-84439c0cd140 | deep-graph-reprogramming | 2304.14593 | null | https://arxiv.org/abs/2304.14593v1 | https://arxiv.org/pdf/2304.14593v1.pdf | Deep Graph Reprogramming | In this paper, we explore a novel model reusing task tailored for graph neural networks (GNNs), termed as "deep graph reprogramming". We strive to reprogram a pre-trained GNN, without amending raw node features nor model parameters, to handle a bunch of cross-level downstream tasks in various domains. To this end, we propose an innovative Data Reprogramming paradigm alongside a Model Reprogramming paradigm. The former one aims to address the challenge of diversified graph feature dimensions for various tasks on the input side, while the latter alleviates the dilemma of fixed per-task-per-model behavior on the model side. For data reprogramming, we specifically devise an elaborated Meta-FeatPadding method to deal with heterogeneous input dimensions, and also develop a transductive Edge-Slimming as well as an inductive Meta-GraPadding approach for diverse homogenous samples. Meanwhile, for model reprogramming, we propose a novel task-adaptive Reprogrammable-Aggregator, to endow the frozen model with larger expressive capacities in handling cross-domain tasks. Experiments on fourteen datasets across node/graph classification/regression, 3D object recognition, and distributed action recognition, demonstrate that the proposed methods yield gratifying results, on par with those by re-training from scratch. | ['DaCheng Tao', 'Xinchao Wang', 'Yiding Yang', 'Li Ju', 'Chongbin Yuan', 'Yongcheng Jing'] | 2023-04-28 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Jing_Deep_Graph_Reprogramming_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Jing_Deep_Graph_Reprogramming_CVPR_2023_paper.pdf | cvpr-2023-1 | ['3d-object-recognition', 'object-recognition', 'action-recognition-in-videos', 'graph-classification'] | ['computer-vision', 'computer-vision', 'computer-vision', 'graphs'] | [ 5.55663168e-01 3.53829265e-01 -7.86394104e-02 -1.35736927e-01
-3.59364659e-01 -5.71895301e-01 6.03381276e-01 -2.71569759e-01
-2.45333731e-01 7.08283663e-01 -1.39686704e-01 -1.74125955e-01
-3.98863047e-01 -7.38055110e-01 -8.03317368e-01 -8.52557778e-01
2.80875444e-01 5.67783356e-01 3.47012207e-02 -4.08767760e-01
-4.88254502e-02 5.51141322e-01 -1.19194078e+00 -6.89922124e-02
9.75158751e-01 8.76131296e-01 7.84906521e-02 4.44560736e-01
-9.07094404e-02 4.39576715e-01 -2.93437839e-01 -7.34321117e-01
3.37908089e-01 -2.51378775e-01 -4.17877316e-01 1.68857709e-01
3.20192933e-01 1.09374247e-01 -3.94995093e-01 9.57527161e-01
6.74141526e-01 2.47925371e-02 7.05916584e-01 -1.54165518e+00
-8.78085196e-01 7.10460246e-01 -4.13605362e-01 -2.87452582e-02
-5.25518246e-02 4.20285165e-01 7.32765973e-01 -6.76978528e-01
8.58797073e-01 9.92630303e-01 9.26981270e-01 9.89050627e-01
-1.36865258e+00 -4.45509404e-01 4.15666491e-01 -2.00375959e-01
-1.17233813e+00 -4.53584343e-01 1.11634612e+00 -4.43957627e-01
8.96507025e-01 7.83580914e-02 7.70227909e-01 1.85690176e+00
3.15469682e-01 4.28385198e-01 7.02144146e-01 -9.11802128e-02
3.02833557e-01 -6.73085675e-02 9.57904235e-02 8.00375879e-01
4.58248138e-01 6.35694116e-02 -5.51625431e-01 -4.09900025e-02
6.89360738e-01 8.52092803e-02 -2.90698290e-01 -7.43125916e-01
-1.11794269e+00 4.99031514e-01 4.00255412e-01 3.89512807e-01
-4.34902042e-01 3.70402008e-01 5.95948756e-01 5.22662640e-01
8.40057194e-01 4.19407457e-01 -6.97249532e-01 3.24683301e-02
-4.44340616e-01 1.27544329e-01 8.50518227e-01 1.26232326e+00
7.88108170e-01 3.59701037e-01 -3.02770168e-01 7.44428992e-01
7.31629878e-02 1.51530772e-01 5.34707248e-01 -3.51541579e-01
6.23109281e-01 1.04602301e+00 -4.47166353e-01 -1.05476165e+00
-7.92383194e-01 -9.97193336e-01 -1.42011690e+00 -1.83888465e-01
1.59581125e-01 -2.05474913e-01 -1.26074529e+00 2.14796352e+00
6.01314962e-01 2.94371158e-01 6.00090697e-02 5.62720120e-01
8.78680706e-01 2.73072660e-01 1.20017968e-01 2.10370477e-02
1.15316379e+00 -1.19154072e+00 -5.12228906e-01 -3.92442733e-01
1.05205441e+00 1.25368431e-01 1.14845288e+00 4.28315192e-01
-8.37551713e-01 -5.58769822e-01 -7.96442688e-01 -1.11908242e-01
-6.53379738e-01 7.58081451e-02 7.90828764e-01 6.57038927e-01
-1.31133866e+00 6.24569714e-01 -6.49576247e-01 -5.12258232e-01
5.57901144e-01 4.41936463e-01 -7.48751342e-01 -1.24238200e-01
-9.83580828e-01 7.14728117e-01 6.44994617e-01 3.42357904e-01
-1.05498469e+00 -9.48523760e-01 -6.96652174e-01 8.26620311e-02
6.92703485e-01 -1.39945090e+00 5.21447837e-01 -8.09327662e-01
-1.59937060e+00 7.57724643e-01 3.86030108e-01 -1.88085601e-01
6.19370878e-01 1.74859494e-01 -4.49053943e-01 -3.08470339e-01
-2.27570221e-01 5.87420344e-01 1.18883681e+00 -1.23056483e+00
2.87644695e-02 -6.79060936e-01 3.38169485e-02 2.18118191e-01
-8.42837155e-01 -5.69157898e-01 -5.74817717e-01 -8.14620078e-01
-7.06878081e-02 -1.06034577e+00 -3.63827169e-01 -1.29173130e-01
-5.95125318e-01 -1.01865642e-01 7.22784817e-01 -3.71659547e-01
1.16221309e+00 -2.17543459e+00 7.96193480e-01 2.39708856e-01
6.60837770e-01 3.56665999e-01 -6.26541376e-01 5.13368547e-01
-1.82295874e-01 1.02439374e-01 -4.99435365e-01 -6.54112041e-01
4.38813530e-02 5.23906350e-01 1.08711652e-01 2.80326933e-01
3.41997415e-01 1.29747927e+00 -8.79836738e-01 -3.98510337e-01
-1.57247052e-01 3.32092941e-01 -7.23389089e-01 -3.99975479e-02
-3.13196182e-01 6.16016388e-01 -6.41027868e-01 6.43247902e-01
8.01479459e-01 -3.88769925e-01 2.76749343e-01 -3.63465637e-01
3.76895845e-01 -2.73480535e-01 -9.88492787e-01 2.24954128e+00
-5.39624691e-01 -1.47067621e-01 1.87724367e-01 -1.37585306e+00
1.20627892e+00 -1.23697864e-02 3.60499829e-01 -4.78110045e-01
2.96094567e-01 5.56080528e-02 -1.04559034e-01 -4.47947741e-01
3.42820972e-01 -7.49409124e-02 -2.99424469e-01 2.00443521e-01
5.46155214e-01 1.16112316e-02 1.09778158e-01 1.29162893e-01
1.60335541e+00 3.83798093e-01 1.34179950e-01 -3.39412987e-01
4.15897161e-01 -6.25721067e-02 5.13570726e-01 5.61463952e-01
-1.30937127e-02 3.56476486e-01 9.10098553e-01 -4.50056642e-01
-6.62553310e-01 -7.57450759e-01 3.60915154e-01 1.30391777e+00
-5.56675121e-02 -3.96586269e-01 -8.43971312e-01 -1.18667567e+00
1.98466018e-01 5.24710655e-01 -9.70182776e-01 -7.77123332e-01
-3.29026371e-01 -9.13969338e-01 7.90830433e-01 4.86169577e-01
3.79097313e-01 -8.23617697e-01 -2.99934857e-03 1.35985970e-01
2.15051174e-01 -1.14910102e+00 -4.14721221e-01 4.08452809e-01
-9.60059643e-01 -7.43446231e-01 -6.56049371e-01 -8.26367140e-01
8.13328981e-01 1.12425849e-01 1.03310668e+00 1.52592257e-01
-9.50893015e-02 3.12792629e-01 -5.00247180e-01 -8.26243460e-02
-5.58904290e-01 7.05590963e-01 1.18670464e-01 3.49763542e-01
-1.38707548e-01 -1.05675972e+00 -1.79545149e-01 1.85169086e-01
-1.14051294e+00 3.51927787e-01 7.08466649e-01 9.83216166e-01
5.71210682e-01 -3.15127254e-01 1.10006797e+00 -1.20684791e+00
8.30085933e-01 -7.37987995e-01 -2.85934359e-01 6.11765444e-01
-7.88325310e-01 1.47675142e-01 1.10190094e+00 -6.09742224e-01
-9.19462144e-01 1.01368539e-01 -1.30459126e-02 -7.66931534e-01
5.65156229e-02 5.72582483e-01 -5.26331306e-01 -3.68746042e-01
9.02075887e-01 3.33263129e-01 2.13174090e-01 -6.50209904e-01
6.54367864e-01 3.08937967e-01 4.79658812e-01 -5.44310272e-01
1.07327163e+00 1.64739355e-01 3.74890745e-01 -3.74553680e-01
-4.47494686e-01 2.90870480e-02 -7.52451360e-01 -2.47760400e-01
8.92103672e-01 -7.60986567e-01 -5.09868383e-01 8.21808040e-01
-1.00142956e+00 -6.55510783e-01 -4.26699191e-01 9.52852368e-02
-4.71794546e-01 2.67036766e-01 -2.01321781e-01 -1.34037390e-01
-5.01735687e-01 -9.51829553e-01 1.06198502e+00 1.10578276e-01
2.95658082e-01 -1.31463039e+00 1.71601713e-01 2.76930422e-01
4.85782295e-01 5.73103249e-01 1.05049300e+00 -1.05517018e+00
-2.40167618e-01 -1.65211916e-01 -2.83178926e-01 4.22662735e-01
3.03992063e-01 -4.21245068e-01 -8.08531344e-01 -4.37114626e-01
-3.13216269e-01 -4.33042318e-01 8.94287467e-01 -5.71723096e-02
1.26283348e+00 -1.61180809e-01 -5.52601337e-01 9.69877601e-01
1.20758951e+00 -1.37298092e-01 3.66696417e-01 -3.56385075e-02
1.23478627e+00 3.58181953e-01 2.52283007e-01 3.45845312e-01
4.12305921e-01 5.96866369e-01 6.69158578e-01 -1.51026845e-01
-2.09537029e-01 -5.03233433e-01 1.08628750e-01 8.15370977e-01
-1.35158792e-01 -7.62766778e-01 -9.85096872e-01 3.88026863e-01
-1.93888903e+00 -4.49761868e-01 -5.68373613e-02 1.61591244e+00
4.74897027e-01 1.93311274e-02 -1.18614413e-01 -1.58536270e-01
6.66280925e-01 1.80434301e-01 -9.90441561e-01 -2.78372288e-01
-1.91849582e-02 2.24299446e-01 3.54379952e-01 8.53915438e-02
-7.52651751e-01 1.26739848e+00 5.29219055e+00 1.07437789e+00
-1.03170621e+00 2.54649222e-01 4.21820760e-01 1.04918294e-01
-6.17274225e-01 5.23729138e-02 -6.92096412e-01 3.02174658e-01
8.38442504e-01 -1.47839308e-01 6.98740065e-01 9.02545810e-01
-1.36485070e-01 3.97127628e-01 -1.20119309e+00 8.25685203e-01
1.79526478e-01 -1.28206074e+00 4.88842368e-01 1.62572846e-01
5.88045180e-01 -6.50874525e-02 -1.30610242e-01 6.03959262e-01
2.28240207e-01 -8.25893104e-01 3.63645852e-01 7.17559397e-01
8.91340137e-01 -4.27912205e-01 4.13660049e-01 6.06288850e-01
-1.11412203e+00 -1.21989682e-01 -2.50096321e-01 8.74274820e-02
3.71183977e-02 6.15487456e-01 -7.87930548e-01 1.21109080e+00
4.12427008e-01 9.97126698e-01 -9.45312083e-01 4.86924201e-01
1.15803488e-01 1.02720834e-01 -1.98335171e-01 9.24588069e-02
1.47819325e-01 -3.62597048e-01 6.27998710e-01 9.32074249e-01
4.36653882e-01 -1.88465714e-01 -9.58781466e-02 9.06323612e-01
-4.11276072e-01 -3.33221965e-02 -1.08258080e+00 -2.02535629e-01
2.19347998e-01 1.61484194e+00 -7.61837184e-01 -1.37617618e-01
-7.26470649e-02 1.07620060e+00 7.88300276e-01 3.91926914e-01
-1.01746094e+00 -1.91883981e-01 3.51110578e-01 -1.92871630e-01
2.03966513e-01 -1.67556405e-01 -1.51290730e-01 -1.34932423e+00
5.82737774e-02 -8.89810264e-01 3.79137903e-01 -6.98312938e-01
-1.44984198e+00 6.78692698e-01 3.18660289e-02 -1.15420914e+00
-1.08629778e-01 -6.03548348e-01 -5.72247446e-01 5.12865067e-01
-1.32375562e+00 -1.64034653e+00 -5.84139228e-01 8.61927629e-01
2.07561359e-01 -3.19997042e-01 6.06016099e-01 5.37604332e-01
-9.68729794e-01 8.39824915e-01 -1.48272574e-01 -2.01979816e-01
4.99345094e-01 -9.19189334e-01 7.07610250e-01 7.77603686e-01
-8.16144124e-02 4.60197121e-01 2.65043467e-01 -7.58528411e-01
-1.83628404e+00 -1.57972479e+00 4.39741820e-01 -3.24531585e-01
5.85409164e-01 -8.35829020e-01 -9.24304366e-01 8.67827415e-01
-2.36878783e-01 3.12975645e-01 4.52854156e-01 1.31365299e-01
-2.69179702e-01 -4.01530027e-01 -1.20486653e+00 5.65412402e-01
2.09829593e+00 -2.15967715e-01 -1.61564350e-01 3.99172485e-01
1.04511166e+00 -3.90474290e-01 -1.06017673e+00 5.74397743e-01
2.29144856e-01 -6.39477015e-01 9.26633894e-01 -9.94647145e-01
1.15451626e-01 -4.46172468e-02 3.74745131e-02 -1.67786932e+00
-4.81960803e-01 -8.73437762e-01 -1.46110266e-01 1.51791346e+00
5.24888873e-01 -1.00165653e+00 8.29369068e-01 4.20588583e-01
-8.46392870e-01 -8.00666869e-01 -1.22826040e+00 -9.08003092e-01
-1.68513518e-03 -2.95963645e-01 7.83623457e-01 1.01021719e+00
-4.03527737e-01 5.28531253e-01 -4.75087106e-01 -4.08490859e-02
4.85407621e-01 -2.86101460e-01 1.13186073e+00 -1.25856662e+00
-4.27180409e-01 -3.72354388e-01 -3.97915423e-01 -8.64041269e-01
4.82428372e-01 -1.42321980e+00 -3.66867900e-01 -1.45116758e+00
-5.00317588e-02 -5.10813653e-01 -2.35727534e-01 9.46458697e-01
4.60103340e-02 1.34290665e-01 1.28019467e-01 -1.17901646e-01
-4.38585877e-01 9.43235695e-01 1.59322202e+00 -6.73771575e-02
-3.19532335e-01 -1.04084074e-01 -1.05501819e+00 4.47637498e-01
6.44891858e-01 -6.62002385e-01 -8.88839900e-01 -6.42722666e-01
4.71317232e-01 1.20606106e-02 6.59377635e-01 -9.83078897e-01
2.25766927e-01 -1.06387347e-01 3.60328220e-02 -7.52809420e-02
2.56064922e-01 -9.28144872e-01 6.78870320e-01 3.43019992e-01
-5.88162728e-02 1.06333710e-01 2.55104750e-01 1.00600243e+00
2.09052250e-01 -7.18141496e-02 5.11976182e-01 3.65584381e-02
-6.12462103e-01 6.02341831e-01 1.47304982e-01 1.32646874e-01
1.30636299e+00 -3.42024565e-01 -5.95727801e-01 1.42706886e-01
-9.01565373e-01 2.70783007e-01 4.81886446e-01 7.01533854e-01
4.00473148e-01 -1.31104791e+00 -5.54821968e-01 5.53464234e-01
3.43847394e-01 2.16905004e-03 6.78781331e-01 1.07131016e+00
-1.35514796e-01 -1.03874587e-01 -2.51658201e-01 -5.34562409e-01
-7.76340008e-01 7.46815503e-01 3.26102108e-01 -7.86842763e-01
-6.03833914e-01 8.01712573e-01 2.70606518e-01 -8.47940564e-01
-7.43055576e-03 -5.11445642e-01 -5.16203940e-02 2.51043797e-01
-3.45939845e-01 3.57024729e-01 3.81712466e-01 -1.80167511e-01
-2.96789318e-01 5.95811486e-01 -3.27979103e-02 4.61867929e-01
1.51693797e+00 -2.81743500e-02 2.00354885e-02 2.01560885e-01
1.14573598e+00 -4.57489282e-01 -1.31129062e+00 -1.42882943e-01
-2.79378444e-01 -8.56336132e-02 7.68465572e-04 -7.94918716e-01
-1.35063231e+00 5.54761589e-01 2.53450990e-01 3.46544266e-01
1.31626689e+00 -7.38299936e-02 5.70497870e-01 4.98184144e-01
4.23885226e-01 -8.58662248e-01 1.29631549e-01 4.14631307e-01
1.06391966e+00 -9.14850295e-01 -2.28708848e-01 -3.51915061e-01
-6.76578045e-01 8.83496404e-01 7.69016922e-01 -5.19128479e-02
7.02541530e-01 6.73218668e-02 -5.45517504e-01 -3.49907070e-01
-9.10655856e-01 -6.55631945e-02 9.75075960e-02 9.00555730e-01
-3.17732722e-01 -1.25315607e-01 -1.30330637e-01 8.14839005e-01
4.91198897e-02 3.57326627e-01 2.21710563e-01 7.06408858e-01
9.19288993e-02 -1.20338368e+00 1.90946609e-01 5.28937340e-01
1.31976366e-01 -6.96186200e-02 -5.72126031e-01 1.07015991e+00
2.04358771e-01 5.18355012e-01 -1.97986975e-01 -8.85175765e-01
5.52088380e-01 1.57457873e-01 4.48053390e-01 -6.90105677e-01
-8.66084099e-01 -3.90745968e-01 3.18671525e-01 -4.66638863e-01
-2.04202145e-01 -2.16075301e-01 -8.60354960e-01 -1.82396263e-01
-2.78239310e-01 -1.75701618e-01 6.05457544e-01 8.75347853e-01
9.78416443e-01 8.76592278e-01 5.46598911e-01 -1.05192149e+00
-8.43696535e-01 -1.05905557e+00 -6.93941295e-01 3.58342350e-01
1.19389817e-01 -8.36437404e-01 -4.68948841e-01 -1.06919095e-01] | [7.302810192108154, 6.272518157958984] |
94f68bbb-f6f6-49e2-a804-14983bac95e2 | near-infrared-depth-independent-image | 2203.14085 | null | https://arxiv.org/abs/2203.14085v1 | https://arxiv.org/pdf/2203.14085v1.pdf | Near-Infrared Depth-Independent Image Dehazing using Haar Wavelets | We propose a fusion algorithm for haze removal that combines color information from an RGB image and edge information extracted from its corresponding NIR image using Haar wavelets. The proposed algorithm is based on the key observation that NIR edge features are more prominent in the hazy regions of the image than the RGB edge features in those same regions. To combine the color and edge information, we introduce a haze-weight map which proportionately distributes the color and edge information during the fusion process. Because NIR images are, intrinsically, nearly haze-free, our work makes no assumptions like existing works that rely on a scattering model and essentially designing a depth-independent method. This helps in minimizing artifacts and gives a more realistic sense to the restored haze-free image. Extensive experiments show that the proposed algorithm is both qualitatively and quantitatively better on several key metrics when compared to existing state-of-the-art methods. | ['Hassan Foroosh', 'Shengnan Hu', 'Ankit Sharma', 'Sumit Laha'] | 2022-03-26 | null | null | null | null | ['image-dehazing'] | ['computer-vision'] | [ 4.63118702e-02 -4.64360982e-01 4.53214407e-01 -7.19282478e-02
-3.74371350e-01 -1.31872073e-01 1.60646185e-01 -1.74930423e-01
-2.63650358e-01 4.32259947e-01 1.06668867e-01 -5.97226899e-04
-1.34088844e-01 -1.05244982e+00 -3.02018702e-01 -1.29597092e+00
2.74662584e-01 -2.68440813e-01 4.30966467e-01 -5.09898245e-01
1.57277197e-01 5.38917065e-01 -1.73219204e+00 7.10153505e-02
9.73546743e-01 1.16689658e+00 3.00648361e-01 5.22074103e-01
-3.98205630e-02 9.72455740e-01 -2.14456752e-01 -7.56292120e-02
6.68724835e-01 -6.03309274e-01 -1.72481149e-01 4.00360882e-01
3.21636587e-01 -4.03192133e-01 -5.71738482e-01 1.27358413e+00
3.34957898e-01 3.20700347e-01 4.69430417e-01 -9.87697423e-01
-8.05839300e-01 -1.78528085e-01 -8.60415936e-01 3.69173624e-02
1.79666251e-01 -1.05224632e-01 5.74641168e-01 -1.08962476e+00
3.56579989e-01 6.35551393e-01 4.94328886e-01 3.93996798e-02
-7.62705624e-01 -4.47817832e-01 -1.63799658e-01 4.14025187e-01
-1.61937118e+00 -3.52032959e-01 1.26297927e+00 -1.22796327e-01
4.07424271e-01 4.32569832e-01 8.95702779e-01 1.29101232e-01
5.18737793e-01 6.41728222e-01 1.52593565e+00 -7.89648116e-01
3.34287643e-01 8.09367150e-02 -4.46478697e-03 6.69777274e-01
5.92319965e-01 1.18704960e-01 -5.53697228e-01 1.25523984e-01
4.74070668e-01 6.24422371e-01 -6.59199893e-01 -3.05227190e-01
-8.52701426e-01 5.06522238e-01 5.94360828e-01 3.04838508e-01
-7.69903541e-01 1.28711304e-02 -6.33532643e-01 3.55884314e-01
5.83138108e-01 -1.26492172e-01 8.92591756e-03 4.39690471e-01
-1.04188418e+00 -1.13245964e-01 3.53177041e-01 7.13992178e-01
1.31407642e+00 1.80752978e-01 3.40508252e-01 6.27521515e-01
6.36617899e-01 9.89072382e-01 -3.65561619e-02 -1.16529417e+00
-1.72384247e-01 4.93649781e-01 4.15497959e-01 -9.73831534e-01
-1.47520259e-01 -2.54768819e-01 -8.08189332e-01 9.48827684e-01
1.30178601e-01 1.04777500e-01 -1.21501327e+00 1.10504520e+00
4.25537199e-01 8.76225829e-02 3.05975825e-01 1.01674175e+00
8.36384952e-01 7.34535813e-01 -5.48864722e-01 -2.89777189e-01
1.22904742e+00 -8.17488372e-01 -1.17546713e+00 -5.34369946e-02
-1.75556511e-01 -9.49620426e-01 6.17769182e-01 6.22924030e-01
-1.05485702e+00 -2.76345640e-01 -1.17036200e+00 -7.35030770e-02
-4.99226332e-01 -1.59050763e-01 5.47833443e-01 7.90976882e-01
-1.39716625e+00 2.55676240e-01 -6.00295782e-01 -2.92739838e-01
-1.52356848e-01 7.52506719e-04 -3.17469180e-01 -6.57048583e-01
-8.93243432e-01 1.00813198e+00 2.85994206e-02 4.32987303e-01
-3.71964842e-01 -3.85392398e-01 -8.52214336e-01 -1.68144003e-01
2.01325640e-01 -6.25367641e-01 4.49038476e-01 -9.04194236e-01
-1.46009195e+00 5.07224143e-01 -3.02970260e-01 1.96621194e-01
1.73665762e-01 -2.71342605e-01 -5.86527288e-01 5.37224293e-01
-3.87593269e-01 1.36110500e-01 1.04434323e+00 -2.02967000e+00
-5.70498109e-01 -3.81894439e-01 -2.22567804e-02 2.99905449e-01
-1.77661031e-01 -2.92565048e-01 -6.20614648e-01 -5.51688433e-01
6.87624812e-01 -6.98153317e-01 -1.84037253e-01 1.97831362e-01
-2.83138841e-01 4.70430940e-01 1.03658688e+00 -7.02123225e-01
1.04805410e+00 -2.41116595e+00 -2.31815632e-02 2.84227192e-01
6.12703741e-01 -3.53276432e-02 -8.89542419e-03 6.10257506e-01
6.48377240e-02 -2.81181008e-01 -3.65915209e-01 -3.26670259e-01
-3.93903852e-01 1.02748677e-01 1.26150876e-01 1.04213369e+00
-1.96272492e-01 5.35336852e-01 -8.53264272e-01 -3.77251089e-01
6.97622299e-01 9.12668288e-01 -9.57906172e-02 2.24148974e-01
4.98762220e-01 3.70889246e-01 -2.40144357e-01 1.01172435e+00
1.23780727e+00 1.87910542e-01 -3.62549782e-01 -3.24969888e-01
-4.74649608e-01 -1.94316283e-01 -1.02967191e+00 1.46231306e+00
-4.49207813e-01 6.90133929e-01 4.93975341e-01 -5.97949207e-01
8.90277922e-01 2.36875698e-01 6.58943832e-01 -8.71750653e-01
1.97735161e-01 1.61880404e-01 -4.70904648e-01 -4.94551778e-01
6.13436878e-01 -4.87004995e-01 5.09739518e-01 2.91795671e-01
-2.74716139e-01 -3.68288964e-01 -2.48056695e-01 1.03431590e-01
9.69278276e-01 4.71351994e-03 1.00794733e-01 -2.97783911e-01
4.68969166e-01 -1.16762079e-01 4.72632974e-01 7.55051792e-01
-3.11251462e-01 9.00581002e-01 -4.31191206e-01 -3.76744658e-01
-7.75560975e-01 -1.25580072e+00 -1.48758456e-01 4.88168329e-01
8.24866414e-01 -5.63628785e-02 -5.41574597e-01 -7.56018758e-02
-1.58749968e-01 6.70291245e-01 -7.99366415e-01 -2.01418519e-01
-2.10344478e-01 -7.96076477e-01 -1.65667072e-01 4.73111542e-03
7.34848499e-01 -6.60156667e-01 -5.89446843e-01 -1.26657067e-02
-2.66067475e-01 -1.05132604e+00 -1.01033761e-03 1.16753161e-01
-9.26155269e-01 -1.00103092e+00 -7.90837824e-01 -4.87399668e-01
7.92800426e-01 1.27330959e+00 1.07878840e+00 3.99181932e-01
-2.18732581e-01 6.55608296e-01 -9.04109478e-01 -5.44119060e-01
8.96385014e-02 -8.18807364e-01 -2.74012983e-01 5.30715287e-01
3.84744525e-01 -5.62892973e-01 -9.67475951e-01 1.24024777e-02
-1.19220233e+00 2.77959913e-01 7.78266728e-01 5.57849109e-01
4.85209674e-01 6.60568714e-01 -1.69982105e-01 -5.34906209e-01
-1.08721316e-01 -5.45010209e-01 -6.08483255e-01 2.36183584e-01
-7.77953446e-01 -1.44269109e-01 2.67631531e-01 1.51792407e-01
-1.21904075e+00 1.88142687e-01 1.88635364e-01 -4.07764256e-01
-6.86640143e-02 2.66623646e-01 -1.70624420e-01 -6.67855978e-01
2.61370420e-01 5.15181363e-01 -6.67935386e-02 -4.18197006e-01
4.14559275e-01 6.56648457e-01 6.53756618e-01 -8.29531923e-02
1.37505019e+00 1.16747022e+00 1.82251677e-01 -1.30598760e+00
-4.67710078e-01 -9.60074902e-01 -4.86101031e-01 -5.84349513e-01
8.91774118e-01 -1.03106916e+00 -6.01772070e-01 6.87003374e-01
-9.78345811e-01 -1.46995753e-01 -1.08064920e-01 5.87954462e-01
-1.07006386e-01 5.21050274e-01 -5.76855302e-01 -1.37931192e+00
-1.82078511e-01 -9.09035861e-01 1.06724119e+00 3.41151059e-01
7.07185805e-01 -1.04728425e+00 1.49382483e-02 3.97322029e-01
8.45588326e-01 4.76087004e-01 5.11921287e-01 4.32037741e-01
-9.68437552e-01 -1.70020774e-01 -3.74112308e-01 4.46503192e-01
5.29203176e-01 9.21783522e-02 -1.25437367e+00 -5.92220649e-02
4.69227165e-01 2.53554702e-01 1.27120471e+00 4.57425505e-01
4.98903602e-01 -1.03651285e-02 1.08614422e-01 7.78501868e-01
2.09234285e+00 1.42006613e-02 1.19424534e+00 5.06227314e-01
7.68326223e-01 6.53756738e-01 5.90835512e-01 3.86020094e-01
4.39373881e-01 5.43817163e-01 8.67917120e-01 -7.96832860e-01
-4.02663201e-01 4.57406908e-01 3.87248576e-01 6.98057950e-01
-4.49694842e-01 -4.61061269e-01 -4.43233371e-01 5.98231852e-01
-1.62947929e+00 -7.01871455e-01 -5.62387705e-01 2.23103261e+00
4.26875293e-01 -4.58941758e-01 -3.28391582e-01 4.30954158e-01
5.13460457e-01 1.59212038e-01 1.45741031e-02 5.66722453e-02
-5.53312778e-01 3.21092963e-01 8.32777798e-01 7.36569941e-01
-8.34964812e-01 6.38458490e-01 6.57969856e+00 3.33893597e-01
-1.07836628e+00 1.36747301e-01 1.70310244e-01 7.41234794e-02
-6.57298863e-01 2.57067401e-02 2.02861000e-02 3.03964347e-01
5.91519654e-01 1.86496899e-01 7.06277251e-01 3.13539833e-01
3.83201510e-01 -7.00353801e-01 -3.59959513e-01 1.10441256e+00
1.81301743e-01 -9.31063533e-01 -1.74929231e-01 1.06839098e-01
8.96391809e-01 1.13178007e-01 1.83054134e-01 -5.01654446e-01
3.29635322e-01 -7.63240635e-01 9.97407436e-01 9.78386402e-01
4.50019270e-01 -6.21592343e-01 1.08333373e+00 -6.58833385e-02
-1.19929826e+00 6.45339787e-02 -4.26774889e-01 -9.37051997e-02
1.93850085e-01 1.23597193e+00 -2.11153105e-02 9.27181661e-01
1.08214045e+00 6.17284715e-01 -5.49422383e-01 1.11508596e+00
-3.32379103e-01 3.47077489e-01 -3.59637976e-01 5.24853528e-01
1.36339113e-01 -7.76222587e-01 3.26781392e-01 1.02152431e+00
5.62095225e-01 5.71664214e-01 -1.10607170e-01 7.50847638e-01
2.34904543e-01 -1.27654120e-01 -6.94542050e-01 2.98894286e-01
1.74124345e-01 1.43157160e+00 -8.74686837e-01 -1.75815165e-01
-8.55885565e-01 1.26504970e+00 -4.39200014e-01 7.99609005e-01
-4.29804415e-01 -4.76373494e-01 7.00590968e-01 5.74864857e-02
4.81006593e-01 -3.80983680e-01 -3.79261106e-01 -1.14031124e+00
-2.25326214e-02 -3.48946959e-01 -1.61879845e-02 -1.27305686e+00
-1.12135601e+00 5.11193931e-01 -1.81959733e-01 -1.29445744e+00
4.31119055e-01 -6.94685996e-01 -6.44360065e-01 8.42101812e-01
-2.24990177e+00 -1.26053858e+00 -9.84429955e-01 9.41383183e-01
8.50722715e-02 5.45127332e-01 5.53092539e-01 2.92891741e-01
-3.55306447e-01 -1.19920447e-02 4.33443069e-01 -1.99385360e-01
5.58390677e-01 -1.25155628e+00 -2.62839943e-01 1.38742757e+00
-8.68257582e-02 6.80680096e-01 9.30077732e-01 -4.50978577e-01
-1.84959674e+00 -8.02515805e-01 3.67377281e-01 -2.96612918e-01
2.41707325e-01 -1.34108692e-01 -8.72810960e-01 2.11737797e-01
5.17392337e-01 4.91630971e-01 6.70381308e-01 -5.06177604e-01
-3.59532893e-01 -3.18242788e-01 -1.32847226e+00 1.78159460e-01
5.19837677e-01 -4.72338676e-01 -3.70292485e-01 4.24089134e-02
3.65744203e-01 -1.64907575e-01 -6.65445924e-01 4.24094588e-01
7.44324207e-01 -1.56374454e+00 1.03722286e+00 2.95923144e-01
8.71454328e-02 -1.00005591e+00 -6.65228546e-01 -1.19909799e+00
-5.52996039e-01 -3.46915901e-01 -1.21431472e-02 8.40995729e-01
1.65005192e-01 -6.30447865e-01 5.35872400e-01 2.54781961e-01
-3.21611017e-01 -2.66773045e-01 -6.20846868e-01 -6.58296525e-01
-3.45231414e-01 -2.12420508e-01 3.72054309e-01 8.97247493e-01
-4.68396485e-01 -1.87729821e-01 -5.27525365e-01 7.51968205e-01
1.16516197e+00 2.72225171e-01 4.99841183e-01 -1.32506824e+00
5.29928058e-02 -2.22313017e-01 -4.60307896e-01 -3.86353463e-01
-2.68744767e-01 -3.41014802e-01 4.53380078e-01 -1.90350604e+00
3.03752452e-01 -3.74691516e-01 -6.37511790e-01 4.94461596e-01
-3.61844093e-01 9.55503225e-01 1.24352224e-01 3.16395223e-01
-4.06591445e-01 7.47908473e-01 1.13334811e+00 -1.47827581e-01
-1.39500663e-01 -3.94369543e-01 -6.84328735e-01 5.30288756e-01
5.11167049e-01 -3.61758202e-01 -2.14879453e-01 -4.34273005e-01
2.92212278e-01 -2.37679020e-01 3.10628980e-01 -1.15511453e+00
3.41145545e-01 -4.17997926e-01 5.65694392e-01 -4.84456360e-01
5.56647420e-01 -1.33853209e+00 4.19582188e-01 2.56274521e-01
4.96901721e-01 -1.55828059e-01 -7.19558671e-02 6.42189085e-01
-3.02354902e-01 1.84917748e-02 9.36314940e-01 -1.44156545e-01
-8.82470906e-01 8.99043083e-02 -4.83748317e-01 -6.90068364e-01
8.10696483e-01 -6.10867381e-01 -4.18883622e-01 -8.61928046e-01
-2.78372198e-01 -3.48445386e-01 1.18685400e+00 -8.34941342e-02
9.95772839e-01 -1.06063616e+00 -7.35218644e-01 2.62379825e-01
3.95761728e-01 -1.95416510e-01 2.78703183e-01 1.11224186e+00
-9.44075644e-01 -2.39393041e-01 -6.56195134e-02 -3.49143237e-01
-1.05470812e+00 5.75166941e-01 1.96922347e-01 3.25332552e-01
-1.03938401e+00 5.30528426e-01 3.12863141e-01 3.98383997e-02
-1.79877564e-01 -1.10003002e-01 4.52574007e-02 -4.29706007e-01
7.54356563e-01 5.16850114e-01 3.40257078e-01 -9.15404916e-01
-3.44628632e-01 9.45308864e-01 4.85311002e-01 -2.77976543e-01
1.64294887e+00 -6.24351978e-01 -6.56389594e-01 4.23364818e-01
8.99194300e-01 6.37434006e-01 -1.14472234e+00 -2.46829525e-01
-4.58850890e-01 -9.65046823e-01 8.64440560e-01 -7.68908441e-01
-1.64984131e+00 6.57531321e-01 9.58154917e-01 2.62054920e-01
1.85747445e+00 -1.95044845e-01 7.48018920e-01 1.58981800e-01
5.39279997e-01 -6.82512879e-01 -3.87951843e-02 1.47909164e-01
4.69765693e-01 -1.11842823e+00 4.89803016e-01 -6.55799687e-01
-2.74346381e-01 1.11650968e+00 6.15581647e-02 -5.39362133e-02
7.59981394e-01 1.89230248e-01 4.55677480e-01 -3.80866170e-01
-1.00280315e-01 -6.22831285e-01 2.87390053e-01 7.27191567e-01
2.56878555e-01 -1.40265256e-01 -7.87907019e-02 -3.49135883e-02
2.19624192e-01 -1.65831387e-01 8.79583359e-01 1.18813527e+00
-9.40894604e-01 -9.02818739e-01 -9.48001266e-01 5.68534136e-02
-3.13730687e-01 -2.74297595e-01 -2.32939631e-01 7.28114009e-01
1.66465059e-01 1.44398594e+00 -1.97486997e-01 -6.18356049e-01
-6.15052357e-02 -2.05823734e-01 5.53796947e-01 -1.86910585e-01
6.25791773e-02 3.53517413e-01 -4.39973652e-01 -6.55322790e-01
-8.30260634e-01 -3.42818052e-01 -1.18754470e+00 -4.53645378e-01
-6.18864000e-01 1.65484756e-01 9.79767263e-01 8.38685155e-01
1.07244417e-01 4.93141592e-01 9.40923750e-01 -1.10427678e+00
4.70347315e-01 -7.51231492e-01 -1.24067843e+00 2.50336468e-01
8.82820368e-01 -7.96691239e-01 -8.20609689e-01 -1.30083458e-02] | [10.849465370178223, -3.1669416427612305] |
9a3e476c-739b-4eb2-aa92-b33d6bcd1385 | a-novel-policy-for-pre-trained-deep | 2101.00738 | null | https://arxiv.org/abs/2101.00738v2 | https://arxiv.org/pdf/2101.00738v2.pdf | A novel policy for pre-trained Deep Reinforcement Learning for Speech Emotion Recognition | Reinforcement Learning (RL) is a semi-supervised learning paradigm which an agent learns by interacting with an environment. Deep learning in combination with RL provides an efficient method to learn how to interact with the environment is called Deep Reinforcement Learning (deep RL). Deep RL has gained tremendous success in gaming - such as AlphaGo, but its potential have rarely being explored for challenging tasks like Speech Emotion Recognition (SER). The deep RL being used for SER can potentially improve the performance of an automated call centre agent by dynamically learning emotional-aware response to customer queries. While the policy employed by the RL agent plays a major role in action selection, there is no current RL policy tailored for SER. In addition, extended learning period is a general challenge for deep RL which can impact the speed of learning for SER. Therefore, in this paper, we introduce a novel policy - "Zeta policy" which is tailored for SER and apply Pre-training in deep RL to achieve faster learning rate. Pre-training with cross dataset was also studied to discover the feasibility of pre-training the RL Agent with a similar dataset in a scenario of where no real environmental data is not available. IEMOCAP and SAVEE datasets were used for the evaluation with the problem being to recognize four emotions happy, sad, angry and neutral in the utterances provided. Experimental results show that the proposed "Zeta policy" performs better than existing policies. The results also support that pre-training can reduce the training time upon reducing the warm-up period and is robust to cross-corpus scenario. | ['Jiajun Liu', 'Björn W. Schuller', 'Sara Khalifa', 'Rajib Rana', 'Thejan Rajapakshe'] | 2021-01-04 | null | null | null | null | ['cross-corpus'] | ['computer-vision'] | [-2.28519782e-01 1.68143496e-01 2.63426095e-01 -4.86993313e-01
-4.65841055e-01 -5.42701840e-01 4.64011371e-01 -1.32265836e-01
-6.75744116e-01 9.07630920e-01 1.02917232e-01 -9.37908515e-02
1.07217714e-01 -5.35157621e-01 -4.76426303e-01 -7.42032886e-01
-2.10897744e-01 5.98057747e-01 -1.49938762e-01 -7.90972233e-01
1.78066909e-01 5.55903435e-01 -1.70927548e+00 5.63566506e-01
3.52376342e-01 8.74900937e-01 4.19788450e-01 9.17462885e-01
-5.61793149e-02 1.37901175e+00 -9.25470412e-01 1.80831939e-01
2.84435660e-01 -4.37966764e-01 -1.13297355e+00 2.10963693e-02
-6.04404509e-01 -1.91442966e-01 1.56489953e-01 5.61736763e-01
1.18831635e+00 6.96260333e-01 2.50969946e-01 -1.32071590e+00
6.84989989e-02 4.20349658e-01 -1.88513771e-01 1.40340194e-01
5.10159969e-01 2.87815303e-01 7.61258543e-01 -3.66110921e-01
4.53281462e-01 1.31955016e+00 3.64936262e-01 8.72323871e-01
-5.65272391e-01 -7.46151030e-01 1.10341385e-02 2.26141125e-01
-9.07164633e-01 -3.12398136e-01 9.46022809e-01 3.95107754e-02
1.39306188e+00 3.04877847e-01 4.49130416e-01 1.31360924e+00
-5.61849475e-02 9.33298647e-01 1.53576648e+00 -4.66256082e-01
5.20228088e-01 6.10055804e-01 -3.57495040e-01 4.72377300e-01
-8.47559392e-01 4.99413997e-01 -4.23175812e-01 -9.24642831e-02
3.44496042e-01 -5.81407309e-01 4.90675382e-02 2.16187146e-02
-7.37558722e-01 1.07195187e+00 1.03049643e-01 4.11978096e-01
-6.50027156e-01 -4.63382527e-02 1.07183087e+00 8.55558753e-01
3.01692724e-01 9.19694781e-01 -7.72814929e-01 -7.03177691e-01
-3.39180321e-01 2.33708009e-01 9.31485474e-01 4.57704306e-01
6.91717803e-01 6.45404100e-01 1.75750051e-02 1.15379083e+00
1.23755969e-01 1.51117995e-01 8.47787797e-01 -9.96175349e-01
5.97499944e-02 5.15427947e-01 1.56486064e-01 -5.20603895e-01
-7.40033090e-01 -4.05067354e-01 -3.89482677e-01 5.55299103e-01
-6.45356625e-02 -8.54591489e-01 -5.92041910e-01 1.62796605e+00
4.38228011e-01 3.83433998e-01 8.38663518e-01 8.86709452e-01
9.35809612e-01 9.17976201e-01 1.43687099e-01 -4.44103479e-01
1.09335971e+00 -1.00845289e+00 -7.62150764e-01 -2.74495989e-01
7.24841475e-01 -6.96044445e-01 1.17411673e+00 7.39766359e-01
-8.69640052e-01 -4.98357713e-01 -9.03017700e-01 6.67175472e-01
-5.94677269e-01 -1.00715786e-01 7.91017771e-01 6.18576765e-01
-1.06326425e+00 2.25342035e-01 -4.40997303e-01 -4.31232601e-01
-1.06534153e-01 8.32271695e-01 -2.23347396e-01 3.58348697e-01
-1.55275667e+00 1.01928151e+00 5.30182958e-01 -4.55232635e-02
-1.08749366e+00 -1.42778367e-01 -6.67657435e-01 -2.00867504e-02
3.94453883e-01 -5.68162240e-02 1.58973908e+00 -1.77959073e+00
-2.49366283e+00 6.01373374e-01 3.48675519e-01 -5.46547294e-01
3.75527322e-01 -3.42588536e-02 -7.84948766e-01 4.54538651e-02
-4.35213238e-01 8.36403191e-01 7.73949564e-01 -1.50345480e+00
-7.16301441e-01 -2.43028365e-02 3.86791468e-01 7.01385081e-01
5.52181937e-02 2.92661279e-01 6.34402037e-02 -2.33161733e-01
-6.92184687e-01 -1.11552334e+00 -2.89209366e-01 -1.11421359e+00
1.51753724e-01 -4.31716532e-01 1.03517127e+00 -3.41928095e-01
8.60725999e-01 -2.19133067e+00 -3.04041028e-01 3.22925508e-01
-5.30615091e-01 6.86273992e-01 -2.89841980e-01 7.02069402e-01
-3.32673550e-01 -1.10358261e-01 1.12444453e-01 -8.06751137e-04
1.87915079e-02 5.97975194e-01 -7.71722943e-02 6.81060180e-02
3.49227011e-01 5.95052779e-01 -8.89905870e-01 -2.32404485e-01
3.36997747e-01 4.85199451e-01 -5.98164082e-01 9.70169485e-01
-3.03540587e-01 7.38598943e-01 -4.30186391e-01 3.07054490e-01
3.30276579e-01 3.39249074e-01 1.12609081e-01 1.89684674e-01
-1.40810654e-01 1.05799176e-01 -1.29545546e+00 1.34538519e+00
-8.60645175e-01 4.81491297e-01 1.65733963e-01 -1.38420379e+00
1.32907438e+00 6.98407412e-01 6.06387794e-01 -1.17304170e+00
3.44251335e-01 3.49666514e-02 3.24949950e-01 -1.06446052e+00
4.60783094e-01 -4.16799545e-01 -1.69190198e-01 4.90017414e-01
1.34850405e-02 -1.69852637e-02 -1.47583693e-01 -1.56452790e-01
1.19631433e+00 3.36975724e-01 3.00890058e-01 7.40758926e-02
7.81965911e-01 -3.62430364e-02 6.92721844e-01 5.43084085e-01
-4.92868364e-01 -7.87366852e-02 4.37299103e-01 -4.47873831e-01
-5.05838752e-01 -3.69154304e-01 4.42316532e-01 1.63203442e+00
-1.94605947e-01 1.15538843e-01 -6.23851180e-01 -7.31562495e-01
-5.77198982e-01 9.19329226e-01 -5.01732290e-01 -2.61763602e-01
-5.12818992e-01 -6.95987284e-01 6.76436245e-01 2.53982663e-01
6.73293889e-01 -2.10514331e+00 -1.22651517e+00 4.73036587e-01
-1.52026877e-01 -1.14085579e+00 2.00030595e-01 7.65735149e-01
-4.93582100e-01 -6.78572953e-01 -2.31515586e-01 -8.53093088e-01
3.39697808e-01 -2.68938541e-01 1.13825929e+00 -6.01893067e-02
-6.26117662e-02 6.55550957e-01 -8.27083886e-01 -8.32820475e-01
-7.52749026e-01 -1.07793979e-01 1.05641633e-01 2.92348191e-02
3.64558786e-01 -3.98067564e-01 -4.62916017e-01 2.68960088e-01
-7.66817927e-01 -2.65979618e-01 7.01053560e-01 9.31465149e-01
1.32784158e-01 3.42415005e-01 1.23786676e+00 -9.68965769e-01
1.13828409e+00 -5.73992670e-01 -1.90746263e-01 1.41327053e-01
-3.38427871e-01 2.25022629e-01 6.90571249e-01 -5.34623444e-01
-1.36944151e+00 -1.23213530e-02 -4.99375463e-01 -2.27023512e-01
-5.60441136e-01 3.99848729e-01 -5.25837354e-02 -7.37255951e-03
4.77589548e-01 -8.50210562e-02 2.65502501e-02 -5.76739609e-02
-8.11602846e-02 1.02104557e+00 2.02513248e-01 -5.74884355e-01
1.53999031e-01 -9.86011885e-03 -2.90405750e-01 -7.66394377e-01
-5.49122155e-01 -6.42525196e-01 -1.12117995e-02 -5.71703017e-01
7.89113045e-01 -8.71764362e-01 -1.26425958e+00 5.86439148e-02
-8.28518510e-01 -8.13499928e-01 -1.66854277e-01 5.20882308e-01
-8.42371404e-01 -2.10181087e-01 -3.52455884e-01 -1.27548552e+00
-5.02064586e-01 -1.18781543e+00 5.80697000e-01 3.98181915e-01
-3.53437096e-01 -1.02227390e+00 4.61439282e-01 4.62064505e-01
4.86106426e-01 2.33815849e-01 6.73790097e-01 -1.02335238e+00
3.57030749e-01 1.43636495e-01 2.32223511e-01 6.25100732e-01
-3.47553194e-02 -1.99901402e-01 -1.26930201e+00 -1.53321654e-01
1.00550555e-01 -1.14282095e+00 1.84581265e-01 -2.84559149e-02
1.02796543e+00 -2.64698654e-01 4.30921018e-01 1.72582313e-01
1.31305337e+00 8.88369024e-01 6.78811014e-01 7.53249943e-01
2.31402934e-01 8.14721644e-01 1.15623331e+00 7.65321612e-01
2.28252098e-01 4.31205660e-01 5.90118706e-01 -2.53130645e-01
3.61652583e-01 1.28229305e-01 8.27689946e-01 7.35807061e-01
-3.95157076e-02 -2.52722979e-01 -7.46847391e-01 2.40561709e-01
-1.93985319e+00 -1.16627038e+00 5.86199045e-01 1.83464730e+00
7.32724190e-01 8.64036009e-02 2.76256949e-01 2.00224489e-01
3.01708281e-01 2.18897983e-02 -5.80002964e-01 -1.47866416e+00
-7.40174577e-02 6.29908442e-01 1.66740134e-01 6.03472710e-01
-1.00481439e+00 1.19208336e+00 5.36205435e+00 6.04848742e-01
-1.65084231e+00 1.04228936e-01 6.68722630e-01 1.67916901e-02
1.94484457e-01 -3.79691482e-01 -4.31988925e-01 9.45971087e-02
1.35887659e+00 3.65671486e-01 7.43208587e-01 1.03283060e+00
7.36624002e-01 -2.04462841e-01 -7.86709070e-01 9.92943227e-01
-5.87895028e-02 -7.91724920e-01 -4.61830527e-01 -3.07125092e-01
5.09690404e-01 2.64544815e-01 -5.37085645e-02 1.23473608e+00
5.48659325e-01 -1.07745969e+00 2.71397769e-01 1.98324427e-01
3.48540336e-01 -1.43039548e+00 1.10686409e+00 5.51491976e-01
-7.14405596e-01 -3.78461093e-01 -2.63865888e-01 -2.00749382e-01
-1.74274907e-01 -3.55275124e-01 -1.52887821e+00 3.43856543e-01
7.49219239e-01 1.58253521e-01 -1.52397841e-01 4.86303777e-01
-6.54916242e-02 7.98325300e-01 -2.63076220e-02 -3.18527699e-01
7.58857727e-01 -9.89051536e-02 3.70204389e-01 1.45355392e+00
1.36969104e-01 2.40541220e-01 3.42565954e-01 1.32868379e-01
1.33569330e-01 4.02276009e-01 -7.30364561e-01 -9.70475823e-02
1.42109707e-01 1.40215886e+00 -4.98396546e-01 -9.08330306e-02
-2.18976438e-01 9.72300231e-01 3.97261292e-01 4.19400275e-01
-8.64565551e-01 -1.72624484e-01 5.42119443e-01 -1.94660112e-01
2.36026525e-01 2.36264452e-01 2.43782535e-01 -4.28963989e-01
-5.66695988e-01 -1.58255076e+00 3.71943831e-01 -9.06012297e-01
-9.33180809e-01 1.03791845e+00 -2.15197176e-01 -1.17343998e+00
-7.68786967e-01 -4.89696175e-01 -7.16938198e-01 5.53257763e-01
-1.58194757e+00 -9.21336651e-01 -1.24652520e-01 9.03626740e-01
8.42405379e-01 -6.85194373e-01 1.21199954e+00 1.43452108e-01
-6.33237541e-01 3.53526443e-01 -2.11728245e-01 -2.79437248e-02
7.47480273e-01 -1.34846187e+00 -5.57354033e-01 2.25833446e-01
-6.32078201e-02 5.48904352e-02 9.65851009e-01 -2.28232309e-01
-1.37575841e+00 -1.00973415e+00 2.50452489e-01 2.01455221e-01
2.98860162e-01 -2.23516017e-01 -7.48355567e-01 4.17279452e-01
8.16974759e-01 -1.50155902e-01 9.42076683e-01 5.67731299e-02
1.82765245e-01 -1.85754269e-01 -1.39644706e+00 4.65541452e-01
1.88351184e-01 -3.69653255e-01 -2.33114555e-01 1.59779653e-01
5.10983884e-01 -3.61264735e-01 -7.73882866e-01 4.37770247e-01
3.03674400e-01 -1.12364280e+00 6.16396964e-01 -8.44110727e-01
3.99632845e-03 5.47352340e-03 -8.80014747e-02 -1.79082572e+00
5.18117696e-02 -9.17096436e-01 5.94946407e-02 1.20818484e+00
3.83260280e-01 -5.95022738e-01 7.68962681e-01 4.83057261e-01
-1.52952284e-01 -9.17934597e-01 -5.80921471e-01 -4.79869157e-01
-2.09778070e-01 -5.90430796e-01 4.87104863e-01 1.14974558e+00
7.30309859e-02 5.43801427e-01 -5.87864161e-01 1.24387927e-01
-2.23671690e-01 -3.15443426e-01 9.10677612e-01 -8.36924851e-01
-6.08583212e-01 -2.31032103e-01 -1.65802479e-01 -3.10396343e-01
6.17397010e-01 -5.53882718e-01 2.50543296e-01 -1.05671251e+00
-3.54275763e-01 -5.81322491e-01 -6.48496568e-01 5.29194534e-01
1.42421335e-01 -2.06602886e-01 1.23536624e-01 -3.55053484e-01
-7.21467495e-01 6.34072661e-01 1.07727790e+00 1.03220366e-01
-6.22057498e-01 1.32477492e-01 -4.26433951e-01 6.64218485e-01
1.38499296e+00 -4.39586550e-01 -8.34894061e-01 1.32580340e-01
3.34213495e-01 4.27015007e-01 -1.46058589e-01 -9.70949054e-01
-9.33501497e-02 -3.25514674e-01 6.71154484e-02 -2.38841966e-01
4.95361924e-01 -9.33279753e-01 -1.07180409e-01 2.34945685e-01
-5.45105517e-01 4.24646169e-01 5.01766682e-01 2.46831656e-01
-4.33100581e-01 -3.69285315e-01 8.92832637e-01 -3.70246798e-01
-1.11113691e+00 -1.85192019e-01 -7.50640750e-01 1.04292892e-01
1.31425357e+00 1.33782171e-03 1.50854900e-01 -7.82339394e-01
-7.48998404e-01 4.66347575e-01 -1.57576755e-01 5.96038997e-01
6.82263017e-01 -1.01713717e+00 -6.02081597e-01 4.68158424e-02
9.00789257e-03 -2.29229018e-01 1.66809186e-01 6.24056518e-01
-4.19754505e-01 1.78705364e-01 -6.41001761e-01 -2.44473532e-01
-1.52882934e+00 3.86385590e-01 6.64268911e-01 -6.71069741e-01
-2.50697881e-01 7.92780340e-01 -1.27932355e-01 -8.80820453e-01
5.84074557e-01 1.27885371e-01 -7.03577459e-01 9.71920323e-03
3.25699210e-01 1.18877850e-01 1.47548512e-01 -5.41806936e-01
-2.86225647e-01 -5.12128025e-02 -9.91880894e-02 -4.65081573e-01
1.55308163e+00 7.17540458e-02 1.74853489e-01 7.86235213e-01
1.16377211e+00 -2.02364013e-01 -1.09122789e+00 2.41420362e-02
1.04994245e-01 9.18518901e-02 2.83224016e-01 -1.28468871e+00
-1.12706149e+00 7.85350204e-01 1.07308221e+00 8.40056017e-02
1.32141411e+00 -4.55016792e-01 5.39607882e-01 7.22526968e-01
2.13615522e-01 -1.90903509e+00 5.29734492e-01 7.38616288e-01
9.55522656e-01 -1.74799430e+00 -4.36517179e-01 2.89792269e-01
-1.55043197e+00 1.01199818e+00 1.00690317e+00 -9.94353443e-02
4.47749496e-01 4.64786619e-01 6.10438943e-01 -2.77548999e-01
-1.22812009e+00 -3.47736180e-01 -1.77666903e-01 6.34911835e-01
6.16714358e-01 2.82034248e-01 -6.11721864e-03 3.60786557e-01
-1.11930661e-01 -9.93864089e-02 5.34820378e-01 1.12702954e+00
-2.67800808e-01 -1.28058076e+00 -2.45402440e-01 8.34089816e-02
-5.98331094e-01 2.30722398e-01 -5.48890948e-01 8.42400193e-01
3.38448673e-01 1.19510782e+00 -1.65460423e-01 -5.32080412e-01
3.85615379e-01 2.22369120e-01 1.26157030e-01 -5.82400084e-01
-1.36179745e+00 6.36047721e-02 6.50265694e-01 -5.84688008e-01
-6.95471525e-01 -6.13897860e-01 -1.71206570e+00 1.89974736e-02
-5.78437597e-02 4.87673730e-01 7.71372795e-01 1.09253776e+00
2.15444073e-01 7.46125460e-01 1.07919550e+00 -5.92228353e-01
-5.13993561e-01 -9.58079517e-01 -3.64775091e-01 3.70160460e-01
1.76222712e-01 -5.98122358e-01 -1.97199449e-01 -3.19874734e-01] | [13.32573127746582, 6.04916524887085] |
e420cd3f-2427-45c3-b8ad-30b4608d8cef | v2ifi-in-vehicle-vital-sign-monitoring-via | 2110.14848 | null | https://arxiv.org/abs/2110.14848v1 | https://arxiv.org/pdf/2110.14848v1.pdf | V2iFi: in-Vehicle Vital Sign Monitoring via Compact RF Sensing | Given the significant amount of time people spend in vehicles, health issues under driving condition have become a major concern. Such issues may vary from fatigue, asthma, stroke, to even heart attack, yet they can be adequately indicated by vital signs and abnormal activities. Therefore, in-vehicle vital sign monitoring can help us predict and hence prevent these issues. Whereas existing sensor-based (including camera) methods could be used to detect these indicators, privacy concern and system complexity both call for a convenient yet effective and robust alternative. This paper aims to develop V2iFi, an intelligent system performing monitoring tasks using a COTS impulse radio mounted on the windshield. V2iFi is capable of reliably detecting driver's vital signs under driving condition and with the presence of passengers, thus allowing for potentially inferring corresponding health issues. Compared with prior work based on Wi-Fi CSI, V2iFi is able to distinguish reflected signals from multiple users, and hence provide finer-grained measurements under more realistic settings. We evaluate V2iFi both in lab environments and during real-life road tests; the results demonstrate that respiratory rate, heart rate, and heart rate variability can all be estimated accurately. Based on these estimation results, we further discuss how machine learning models can be applied on top of V2iFi so as to improve both physiological and psychological wellbeing in driving environments. | ['Xu Zhang', 'Jun Luo', 'Chao Cai', 'Zhe Chen', 'Tianyue Zheng'] | 2021-10-28 | null | null | null | null | ['heart-rate-variability'] | ['medical'] | [ 1.63467273e-01 -4.17338870e-02 -2.43735716e-01 -2.54486531e-01
-4.05639052e-01 -2.18674049e-01 1.04011282e-01 -6.59660175e-02
-3.66657704e-01 7.97498763e-01 -1.28401443e-01 -4.27468300e-01
-1.01968810e-01 -5.76146483e-01 -9.03141424e-02 -7.07101941e-01
3.49878967e-02 -2.60056138e-01 1.36166811e-01 1.36397004e-01
-9.82312560e-02 5.88157654e-01 -1.97180617e+00 -1.79651067e-01
5.57715535e-01 1.36602104e+00 -1.32942423e-01 8.99741113e-01
4.01418507e-01 7.66053677e-01 -1.04414308e+00 1.40693784e-01
-6.21191040e-03 -1.53331548e-01 6.87115043e-02 -3.66979957e-01
1.78148091e-01 -5.91314018e-01 -4.83961016e-01 2.67762125e-01
9.00303423e-01 -2.45879553e-02 2.87921399e-01 -1.78438282e+00
3.59913468e-01 -1.90373778e-01 -7.34816417e-02 4.14051145e-01
5.08754671e-01 3.64511102e-01 2.44260341e-01 -5.40986538e-01
-1.40363961e-01 6.81571960e-01 8.19560885e-01 5.09093106e-01
-8.00790548e-01 -1.06536138e+00 -4.83139217e-01 6.65247440e-01
-1.51541018e+00 -7.61601746e-01 6.40021503e-01 -2.43348449e-01
7.71184862e-01 6.90491676e-01 7.15389073e-01 1.12456250e+00
2.98952818e-01 4.33835894e-01 1.05644929e+00 9.13288072e-02
2.08021432e-01 6.15071595e-01 2.14762136e-01 2.51507103e-01
5.09216309e-01 2.87303835e-01 -6.16053700e-01 -1.95171311e-01
2.01125458e-01 2.37254053e-03 -6.34831786e-01 2.70843446e-01
-1.18084478e+00 2.07011834e-01 -2.51538873e-01 8.56637880e-02
-5.66843390e-01 3.28612119e-01 2.24424988e-01 2.32598618e-01
1.58375576e-01 1.71643928e-01 -2.12608978e-01 -5.62754989e-01
-8.26541305e-01 -1.34077370e-01 8.15844953e-01 6.80139959e-01
5.09034395e-01 1.33998662e-01 -3.67110163e-01 5.69779456e-01
4.13492501e-01 1.25224388e+00 1.92595050e-01 -9.85362411e-01
1.40576109e-01 -4.87859920e-02 3.52581590e-01 -1.10315073e+00
-8.04336131e-01 -5.65273464e-01 -5.27057111e-01 1.32316723e-01
7.26160780e-02 -5.94269216e-01 -3.61920446e-01 1.48783684e+00
2.75055975e-01 8.82229686e-01 5.27916998e-02 8.70613873e-01
8.99886727e-01 3.54628712e-01 -6.09172322e-02 -5.59005558e-01
1.47864234e+00 -3.11899006e-01 -1.17496777e+00 -3.84265214e-01
4.55875754e-01 -3.65979284e-01 5.39143741e-01 6.66759670e-01
-7.34006643e-01 -7.20046580e-01 -1.22189915e+00 5.64369321e-01
-1.31471470e-01 -3.07079628e-02 1.92040741e-01 1.35790956e+00
-9.69317973e-01 1.03622362e-01 -6.48055255e-01 -3.61801296e-01
3.25340480e-02 8.20665881e-02 -8.93340632e-02 -1.20573424e-01
-1.50949085e+00 1.10089755e+00 -3.04166913e-01 5.17700791e-01
-5.54355919e-01 -7.28488147e-01 -5.63164711e-01 -1.52647108e-01
3.21892470e-01 -5.05303085e-01 1.10931778e+00 2.51376778e-02
-1.50332391e+00 3.35109740e-01 -3.82477582e-01 -5.50056815e-01
4.53938723e-01 -2.21669063e-01 -1.21563208e+00 3.10723335e-01
-2.42655247e-01 2.17607722e-01 7.75653303e-01 -7.07296789e-01
-5.11600018e-01 -9.52135921e-02 -2.04501405e-01 -1.04454443e-01
-2.16746300e-01 -7.10390732e-02 8.45802799e-02 3.45943481e-01
-2.28808850e-01 -9.96938348e-01 1.96275830e-01 2.59513427e-02
-1.54233590e-01 2.78040376e-02 1.00630128e+00 -5.10421455e-01
1.35172033e+00 -2.03733206e+00 -1.05149162e+00 1.83632016e-01
2.92147547e-01 7.40060508e-01 2.17084318e-01 2.04524532e-01
3.15308511e-01 -4.64742891e-02 2.72696465e-01 1.49977757e-02
-1.28464133e-01 3.40117007e-01 4.58823815e-02 8.11139703e-01
1.19713262e-01 7.17630863e-01 -6.40464425e-01 -1.82327166e-01
8.05086672e-01 8.27951670e-01 -3.12146768e-02 1.77087083e-01
7.56320715e-01 5.97536683e-01 -2.88932592e-01 4.66730833e-01
7.21901417e-01 3.79697263e-01 -3.55284989e-01 -2.31278390e-01
-2.29712933e-01 2.07338333e-01 -1.10551870e+00 6.69814110e-01
-8.79464686e-01 1.09728920e+00 2.08297566e-01 -1.07250702e+00
1.05403459e+00 8.28442574e-01 4.81205791e-01 -1.14535248e+00
2.75281429e-01 -7.14280754e-02 -7.33522698e-02 -1.24855232e+00
1.34266438e-02 -9.72627774e-02 -6.99199662e-02 2.12186337e-01
-6.86981976e-01 2.24557389e-02 -3.21896702e-01 -7.75869563e-02
1.22933102e+00 -5.72718143e-01 1.64346665e-01 -3.98227125e-02
7.01335311e-01 -5.67922890e-01 4.20712322e-01 6.74571216e-01
-6.59260154e-01 2.36223474e-01 -6.17767312e-03 1.07932901e-02
-6.39725244e-03 -9.91643488e-01 -3.64207327e-01 4.27515537e-01
4.06330168e-01 8.07133541e-02 -4.89864349e-01 6.15969226e-02
1.59941688e-01 1.24266434e+00 -2.69409359e-01 -5.30835032e-01
-1.46868020e-01 -5.15559733e-01 8.38959098e-01 4.64405358e-01
6.28148377e-01 -6.65030301e-01 -1.14746165e+00 3.60482186e-01
-6.12336755e-01 -1.46060169e+00 3.24475430e-02 -3.41893137e-02
-2.41896480e-01 -9.06425476e-01 -2.54444659e-01 1.59096763e-01
8.44338238e-02 9.39242959e-01 7.78785408e-01 2.19863668e-01
-6.18980587e-01 9.53922331e-01 1.22072451e-01 -1.02370358e+00
-2.30606019e-01 -5.98953664e-01 3.28358740e-01 4.90620166e-01
5.09882152e-01 -3.30021828e-01 -1.00316715e+00 7.96958923e-01
-3.47430527e-01 -4.43738282e-01 3.86821479e-01 6.50641471e-02
1.33208930e-01 1.35355949e-01 9.99969125e-01 -3.40459347e-01
5.42686045e-01 -7.04906404e-01 -3.45267475e-01 -1.04588859e-01
-7.98736691e-01 -4.74486351e-01 1.60078436e-01 -1.80385277e-01
-8.91954720e-01 -3.84658992e-01 -2.75845200e-01 -1.32248208e-01
-6.56631291e-01 3.32957059e-01 -2.53065765e-01 -7.48973852e-03
5.42686462e-01 -8.21557492e-02 2.28029087e-01 6.11088015e-02
-1.00969322e-01 1.22019422e+00 7.29822636e-01 2.60329962e-01
6.36739910e-01 4.53309417e-01 1.91913515e-01 -1.49191463e+00
-5.23737907e-01 -1.05909324e+00 -1.59289807e-01 -9.69472229e-01
7.79610455e-01 -1.16153741e+00 -1.22444975e+00 3.71180892e-01
-7.14988053e-01 -6.98524565e-02 1.39925048e-01 1.00650382e+00
-2.24931866e-01 2.14916527e-01 -1.38936311e-01 -1.49282753e+00
-9.66277346e-02 -6.32261097e-01 7.17315495e-01 2.37312227e-01
-4.98797953e-01 -9.65616703e-01 4.95771598e-03 1.01176476e+00
9.47977185e-01 4.87129003e-01 1.52728604e-02 -1.19003765e-01
-2.12863624e-01 -8.13376904e-01 1.05017908e-01 3.36464167e-01
7.36263469e-02 -1.90338686e-01 -1.66093516e+00 -3.14720459e-02
3.03838491e-01 7.61928037e-02 3.95261705e-01 3.86651665e-01
9.68402863e-01 -1.09595768e-01 -5.56704104e-01 3.66775453e-01
9.76493835e-01 2.00711235e-01 1.07305801e+00 -5.94229326e-02
3.37707967e-01 5.61771154e-01 8.83886397e-01 6.12867057e-01
4.35895056e-01 6.62462950e-01 6.67221546e-01 -3.56867790e-01
-7.81498998e-02 2.41667867e-01 4.50179100e-01 4.10989583e-01
-2.20777601e-01 -5.59903324e-01 -6.07246816e-01 2.76690632e-01
-1.31902075e+00 -1.22133052e+00 -8.97199214e-01 2.57491398e+00
1.97230130e-01 6.87174052e-02 2.47109905e-01 6.26742780e-01
7.45050907e-01 -2.00534135e-01 -5.11841714e-01 -6.89511299e-01
2.68330127e-01 2.37268209e-02 9.33249712e-01 3.00704330e-01
-5.34890890e-01 -1.76157326e-01 6.46526718e+00 1.76249892e-01
-1.53587425e+00 1.71654731e-01 3.47636938e-01 -4.06610698e-01
-1.86121970e-01 -5.12249470e-01 -7.49443233e-01 4.57686365e-01
1.82191038e+00 -2.50460431e-02 1.37148798e-01 5.77774286e-01
9.32350278e-01 -4.31932837e-01 -7.68719494e-01 1.19299185e+00
6.38213158e-02 -6.64876640e-01 -1.09328341e+00 1.99240416e-01
-7.87565857e-02 1.52850807e-01 -4.58674366e-03 3.75721067e-01
-7.53476322e-01 -9.60632086e-01 4.34093624e-01 6.44526720e-01
8.87691617e-01 -7.98500717e-01 8.33949089e-01 5.82870424e-01
-1.17257762e+00 -2.94719841e-02 -1.07728742e-01 -2.82944322e-01
2.89265811e-01 9.84235644e-01 -9.41040933e-01 1.80551931e-01
4.32838678e-01 2.68707097e-01 -2.66596168e-01 1.18220079e+00
-1.97930381e-01 1.17295432e+00 -4.43749219e-01 -1.20211929e-01
-3.24747384e-01 2.09938988e-01 3.51555198e-01 1.11725664e+00
6.45927072e-01 2.62156397e-01 -1.43633783e-01 5.32542586e-01
6.17620230e-01 -2.75759250e-01 -7.92660356e-01 3.73504460e-01
4.62299556e-01 1.47144425e+00 -3.12618107e-01 -1.66820899e-01
-7.52047539e-01 4.87035036e-01 -7.35651910e-01 5.67236006e-01
-1.38343990e+00 -5.64229727e-01 1.12783134e+00 5.87219894e-01
3.27347293e-02 -2.14755327e-01 -1.81915537e-01 -6.21018291e-01
-9.91212483e-03 -1.18469097e-01 -5.13726957e-02 -7.68128991e-01
-4.07155216e-01 3.04283351e-01 1.36182383e-02 -1.53347003e+00
-9.71432477e-02 -2.71021605e-01 -7.72635937e-01 7.23045409e-01
-2.03275037e+00 -3.44983131e-01 -8.31430018e-01 6.53294384e-01
4.34339754e-02 3.50815028e-01 6.27003312e-01 6.68840408e-01
-7.91380525e-01 8.09181035e-01 -1.13465622e-01 -3.95133734e-01
6.91635072e-01 -6.06199384e-01 8.25605765e-02 7.51372397e-01
-3.62212539e-01 1.51754841e-01 7.61762500e-01 -2.50942498e-01
-1.64231896e+00 -1.21294451e+00 8.98825049e-01 -4.51364696e-01
2.74445981e-01 -3.02050740e-01 -7.43450701e-01 2.28734072e-02
4.77572903e-02 4.71855968e-01 8.16866100e-01 -2.21938968e-01
8.58216435e-02 -5.68805575e-01 -1.21871066e+00 2.53165573e-01
7.16174304e-01 -5.40728331e-01 -1.69760719e-01 2.15571612e-01
3.58495228e-02 -4.47397456e-02 -8.74556303e-01 3.20330858e-01
7.51142144e-01 -1.03960276e+00 1.04521334e+00 1.58391088e-01
-6.18445933e-01 -2.78124481e-01 3.30988407e-01 -1.15443623e+00
-7.99321830e-02 -8.66545320e-01 -9.18897465e-02 1.12981927e+00
3.21541518e-01 -1.32545567e+00 3.46486837e-01 1.04526043e+00
-4.45301346e-02 -3.02373856e-01 -9.26943779e-01 -9.14103031e-01
-6.75383508e-01 -1.23552966e+00 4.18871909e-01 3.66241634e-01
3.17803413e-01 1.34572104e-01 -6.72108054e-01 5.13921380e-01
5.42107642e-01 -4.82897341e-01 8.76319110e-01 -1.41654968e+00
-5.90077974e-02 4.28882651e-02 -7.01200902e-01 -7.89085448e-01
-1.61172271e-01 -2.66992122e-01 4.44083363e-01 -1.45193624e+00
-4.87591952e-01 -4.14214045e-01 -5.86167395e-01 2.13778973e-01
-1.84418827e-01 6.70931339e-01 -3.97056937e-01 -3.03181201e-01
-2.33337462e-01 8.80022496e-02 8.97136033e-01 8.06183554e-03
-9.28083584e-02 5.88911116e-01 -4.97198999e-01 4.65830535e-01
9.35345769e-01 -3.24676931e-01 -8.03211570e-01 7.09202960e-02
2.73513794e-01 6.90143764e-01 6.68918788e-01 -1.62246692e+00
2.50927001e-01 -2.13201374e-01 1.64610758e-01 -4.79742557e-01
5.68821251e-01 -1.00073206e+00 3.67053658e-01 7.47820258e-01
-2.79409755e-02 -4.00716454e-01 4.56367582e-01 6.75262392e-01
2.51821298e-02 1.38160706e-01 6.61670268e-01 3.80688906e-01
-5.79655111e-01 1.32811725e-01 -1.20030653e+00 -1.30264387e-01
1.16325510e+00 -5.32737255e-01 -5.52356720e-01 -8.71939838e-01
-2.78518647e-01 2.50759929e-01 -7.34001249e-02 4.80396003e-01
7.37709403e-01 -1.02819359e+00 -6.22082174e-01 4.82814789e-01
3.76251876e-01 -6.27999544e-01 3.57299149e-01 1.50221169e+00
-1.75685987e-01 7.58305132e-01 5.02449237e-02 -9.36856270e-01
-1.46481824e+00 4.39031780e-01 3.58501703e-01 5.83953440e-01
-6.01061344e-01 5.42880476e-01 -5.03476411e-02 2.92639941e-01
2.59259433e-01 -4.84188110e-01 -3.11873436e-01 4.07283008e-02
9.99997318e-01 7.73941576e-01 3.68531823e-01 -6.20805442e-01
-6.83888853e-01 5.59903562e-01 6.51157856e-01 -6.15057871e-02
5.68938375e-01 -8.33027959e-01 5.53025007e-01 4.00000334e-01
1.04796803e+00 1.36019856e-01 -1.07034385e+00 9.24350843e-02
-4.37793434e-01 -3.79402071e-01 5.23108125e-01 -7.12907493e-01
-1.14107132e+00 1.14035404e+00 1.06576288e+00 4.98736471e-01
1.22493339e+00 -2.47893095e-01 1.07323885e+00 2.25053638e-01
4.06924009e-01 -9.44909573e-01 -2.65619278e-01 -2.45834023e-01
3.52499515e-01 -1.09980953e+00 -2.17592195e-01 -4.45455372e-01
-6.41436100e-01 1.02174520e+00 1.58606052e-01 5.14348567e-01
7.70667255e-01 5.34008563e-01 5.88793695e-01 -3.80768329e-02
-8.73476267e-01 -4.52139318e-01 1.14552319e-01 9.79400098e-01
1.89348742e-01 1.59383029e-01 1.55372899e-02 2.66765445e-01
-1.43880844e-02 1.03904709e-01 7.00655103e-01 6.92706585e-01
-6.62601411e-01 -4.29359645e-01 -6.15111530e-01 5.14767647e-01
-4.14411277e-01 4.59541023e-01 2.44923756e-02 5.14974892e-01
2.41056100e-01 1.96411669e+00 1.91884954e-02 -6.02702200e-01
5.77627361e-01 1.28856763e-01 5.42484820e-02 -4.84529957e-02
-2.60420144e-01 -2.67894536e-01 5.21291554e-01 -9.51987088e-01
-5.48347831e-01 -6.88515425e-01 -1.08925688e+00 -2.94639438e-01
-5.42353630e-01 -7.04822456e-03 1.07589424e+00 1.15468681e+00
3.84140372e-01 8.11938047e-01 9.07028675e-01 -3.73281956e-01
-1.39927924e-01 -7.22696304e-01 -7.05686152e-01 4.62619700e-02
6.03026211e-01 -8.06392133e-01 -7.81758010e-01 -2.33796194e-01] | [13.384800910949707, 2.8136484622955322] |
cb9818b3-7724-4386-93fc-9e7f2153fa7b | low-dimensional-denoising-embedding | 2103.17099 | null | https://arxiv.org/abs/2103.17099v1 | https://arxiv.org/pdf/2103.17099v1.pdf | Low-dimensional Denoising Embedding Transformer for ECG Classification | The transformer based model (e.g., FusingTF) has been employed recently for Electrocardiogram (ECG) signal classification. However, the high-dimensional embedding obtained via 1-D convolution and positional encoding can lead to the loss of the signal's own temporal information and a large amount of training parameters. In this paper, we propose a new method for ECG classification, called low-dimensional denoising embedding transformer (LDTF), which contains two components, i.e., low-dimensional denoising embedding (LDE) and transformer learning. In the LDE component, a low-dimensional representation of the signal is obtained in the time-frequency domain while preserving its own temporal information. And with the low dimensional embedding, the transformer learning is then used to obtain a deeper and narrower structure with fewer training parameters than that of the FusingTF. Experiments conducted on the MIT-BIH dataset demonstrates the effectiveness and the superior performance of our proposed method, as compared with state-of-the-art methods. | ['Wenwu Wang', 'Xinxin Wang', 'Pengming Feng', 'Wenbo Wang', 'Jian Guan'] | 2021-03-31 | null | null | null | null | ['ecg-classification'] | ['medical'] | [ 2.29690656e-01 -3.01274657e-01 2.36081824e-01 -2.68491417e-01
-5.01830578e-01 -1.84335560e-01 -3.75358872e-02 -5.57509474e-02
-3.08677614e-01 5.99944472e-01 2.35483631e-01 -4.45183329e-02
-3.23744178e-01 -6.08883917e-01 -1.23159565e-01 -1.09961760e+00
-2.63883442e-01 -2.40171254e-01 1.01642638e-01 9.74582694e-03
-1.66992366e-01 2.09978521e-01 -1.04483688e+00 2.16108635e-01
9.67121720e-01 1.27568161e+00 -1.92047656e-02 3.71671945e-01
9.59299281e-02 5.55494785e-01 -4.61936593e-01 -9.36669186e-02
2.09945943e-02 -6.59863710e-01 -4.06898707e-01 -1.63578212e-01
-2.05741078e-01 -2.41740078e-01 -6.90358818e-01 8.33276153e-01
8.77303064e-01 -3.48077789e-02 5.58044970e-01 -9.22455788e-01
-2.61750698e-01 1.86295673e-01 -5.11744738e-01 3.55078042e-01
6.97748270e-03 -1.07254155e-01 4.90320086e-01 -9.00839746e-01
4.02855098e-01 1.03956091e+00 1.12773407e+00 3.38452131e-01
-1.38980675e+00 -9.23649549e-01 -9.92847607e-02 4.19985622e-01
-1.49493265e+00 -2.13444963e-01 1.26462293e+00 -4.05238658e-01
4.54612941e-01 1.83846325e-01 8.43959689e-01 9.76462662e-01
5.77895045e-01 6.36806488e-01 1.20918036e+00 -1.10591285e-01
-2.74132646e-04 -9.83221978e-02 1.41516626e-01 6.14883423e-01
-2.71019600e-02 2.52247989e-01 -4.01104927e-01 -3.25001895e-01
8.04307878e-01 2.99293578e-01 -7.45729387e-01 -3.39703709e-01
-1.24904299e+00 6.44582987e-01 3.32132310e-01 4.75519389e-01
-5.15673757e-01 7.93863367e-03 7.94365346e-01 3.98818970e-01
6.51868165e-01 -2.91460566e-02 -3.12875092e-01 -2.44775295e-01
-9.12876546e-01 -2.80254483e-02 5.63234091e-01 3.82633716e-01
4.83438790e-01 1.41843423e-01 -2.22530141e-01 8.66347551e-01
2.48986095e-01 2.42202148e-01 7.77024567e-01 -6.15107596e-01
4.36352909e-01 4.70133990e-01 -2.30122313e-01 -1.18162870e+00
-4.47103649e-01 -7.70904422e-01 -1.66075516e+00 -4.90078144e-02
6.56792000e-02 -1.96603805e-01 -7.13904798e-01 1.81252265e+00
2.32144833e-01 8.41499448e-01 2.58767068e-01 8.31596792e-01
7.42978036e-01 7.33651876e-01 -9.64649990e-02 -5.28926313e-01
1.16526043e+00 -5.78603864e-01 -1.16409421e+00 1.42475128e-01
4.72653449e-01 -5.62230170e-01 6.34341896e-01 4.60227549e-01
-8.21236193e-01 -7.46501029e-01 -1.17818785e+00 2.14536767e-02
1.92182250e-02 1.56105459e-01 4.22045410e-01 4.34166133e-01
-7.39187837e-01 8.52092028e-01 -9.29572761e-01 -2.49822866e-02
4.68590796e-01 2.51256317e-01 -4.05438066e-01 -1.47336021e-01
-1.58005381e+00 5.80956519e-01 1.58763736e-01 3.98081481e-01
-7.55677104e-01 -6.47143900e-01 -8.11722040e-01 1.26957446e-01
7.10308999e-02 -3.96848023e-01 6.90375090e-01 -4.25876170e-01
-1.35753548e+00 3.73645484e-01 -2.02969313e-01 -3.58550102e-01
4.05886859e-01 -1.19232908e-01 -5.60027122e-01 4.14876580e-01
-5.27814962e-02 4.72706519e-02 1.19176662e+00 -9.05520797e-01
-2.28290007e-01 -6.08176768e-01 -3.26195359e-01 1.00314379e-01
-7.26617515e-01 -5.14742017e-01 -1.95327550e-01 -1.13537288e+00
6.24720693e-01 -7.29492903e-01 -7.11683333e-02 7.02141449e-02
-6.73493370e-02 -3.23036052e-02 1.06919014e+00 -1.06133854e+00
1.74998450e+00 -2.65310335e+00 5.32237291e-01 1.73347443e-01
5.69558978e-01 3.95733178e-01 7.01986849e-02 4.60443377e-01
-3.45160931e-01 6.37638196e-02 -3.93448770e-01 -3.02313626e-01
-3.41846108e-01 2.69385278e-01 -2.04899721e-02 6.41404092e-01
1.48315290e-02 6.10912263e-01 -8.25636446e-01 -5.08092463e-01
2.96062112e-01 8.11410904e-01 -4.15194660e-01 2.31256425e-01
7.17121065e-01 6.71595156e-01 -6.05801702e-01 4.32023108e-01
7.88315177e-01 -2.03791596e-02 1.85896456e-01 -8.25435877e-01
-2.51159109e-02 1.05113156e-01 -1.14769006e+00 1.90389681e+00
-2.82578826e-01 3.22838962e-01 1.54433204e-02 -1.24606562e+00
9.97543037e-01 6.96648479e-01 7.88099289e-01 -6.20738864e-01
2.26901859e-01 2.41817400e-01 7.86451101e-02 -6.33255899e-01
-1.39997452e-01 -3.78736436e-01 5.02166115e-02 5.03993630e-01
1.34535417e-01 4.24595237e-01 -9.86356884e-02 -7.00253323e-02
1.13201284e+00 1.11924866e-02 1.28365561e-01 -4.04388368e-01
7.42381752e-01 -6.97574019e-01 9.16855395e-01 3.29699963e-01
-2.82786578e-01 4.03337151e-01 4.76057589e-01 -4.71519053e-01
-7.18283713e-01 -8.81858826e-01 -4.20233846e-01 2.60123163e-01
2.12205350e-01 -4.95351255e-01 -6.53394401e-01 -5.79614639e-01
8.52886066e-02 1.45002007e-01 -6.23251200e-01 -7.51442671e-01
-7.97407448e-01 -7.97026873e-01 6.77917480e-01 5.87088406e-01
7.22114325e-01 -7.66893268e-01 -6.07973397e-01 4.80454028e-01
-3.52226704e-01 -8.80213141e-01 -6.82403922e-01 2.95396715e-01
-1.38993454e+00 -8.75169158e-01 -9.06293929e-01 -7.89670825e-01
4.61915612e-01 -9.38173942e-03 6.60011709e-01 5.57817556e-02
-2.27746710e-01 2.51942277e-02 -4.16918904e-01 -1.79167256e-01
-1.55878579e-02 -1.14120811e-01 9.80614424e-02 3.79759490e-01
1.90943062e-01 -9.97944415e-01 -8.60279024e-01 2.94599056e-01
-9.06474471e-01 -6.92581981e-02 7.56134272e-01 1.29153800e+00
5.97517610e-01 4.01572257e-01 1.01456177e+00 -8.15490186e-01
8.36831927e-01 -2.64784932e-01 -1.02158472e-01 -6.14741407e-02
-7.99645960e-01 1.22831231e-02 7.26447701e-01 -5.72113335e-01
-8.48388374e-01 -1.85775623e-01 -3.29142779e-01 -7.90448487e-01
2.82051116e-01 6.58716798e-01 -7.41438195e-02 -9.43611562e-02
2.65919626e-01 7.07708776e-01 3.35388839e-01 -8.98160040e-01
-1.36115253e-01 7.36295283e-01 4.16246891e-01 -3.19340646e-01
7.51620233e-01 2.98697442e-01 7.05691725e-02 -6.48228288e-01
-5.90234697e-01 -3.96292090e-01 -6.21847451e-01 1.83987897e-02
8.70233119e-01 -8.29060555e-01 -5.43851733e-01 7.22263753e-01
-9.51546550e-01 1.69189766e-01 -3.93587321e-01 8.53492200e-01
-3.92861307e-01 7.01807916e-01 -8.38771939e-01 -7.49104261e-01
-6.05121374e-01 -1.01179707e+00 7.39560843e-01 -8.96910504e-02
3.49866860e-02 -9.36233103e-01 4.19330001e-02 -1.04581386e-01
2.91134983e-01 4.74794239e-01 1.25057781e+00 -2.50293255e-01
-1.48068532e-01 -3.00845087e-01 -1.90592818e-02 8.95007551e-01
2.59893745e-01 -6.47241592e-01 -9.35862303e-01 -5.66433847e-01
7.95810580e-01 -7.47504272e-03 8.20315242e-01 3.98344696e-01
1.30437410e+00 -1.25262380e-01 -3.11274886e-01 9.54675317e-01
1.38928926e+00 4.84999508e-01 7.86903977e-01 5.96858859e-02
5.89113653e-01 1.04267649e-01 5.65327287e-01 4.96946812e-01
1.35895893e-01 4.31155056e-01 -1.92868020e-02 -4.16241705e-01
5.47598265e-02 -1.70946091e-01 6.76024407e-02 1.29950011e+00
-2.03331873e-01 7.36004114e-02 -5.52917361e-01 3.96987855e-01
-1.91664076e+00 -8.13978434e-01 9.24443826e-02 2.25008869e+00
9.26710129e-01 1.21939011e-01 -7.25955293e-02 8.63335490e-01
5.67177653e-01 2.63214797e-01 -6.95391297e-01 -9.95312929e-02
9.45975780e-02 3.64652336e-01 1.10140830e-01 6.59467652e-02
-1.12831450e+00 1.44288525e-01 5.70677185e+00 8.17763925e-01
-1.12925518e+00 2.51123071e-01 4.22251821e-01 3.64920646e-01
2.10436974e-02 -2.21076578e-01 -2.35812366e-01 7.12126076e-01
7.83097625e-01 -2.29636520e-01 3.08209121e-01 3.76450330e-01
2.77469963e-01 3.17364693e-01 -9.80309486e-01 1.47101021e+00
6.67313263e-02 -1.10521364e+00 -2.04777420e-01 1.21852458e-02
1.38662830e-01 -5.76453745e-01 -1.55722395e-01 3.63938630e-01
-8.36144328e-01 -9.14669335e-01 4.11193639e-01 7.16706216e-01
1.04571164e+00 -9.76324379e-01 1.09784687e+00 4.01324302e-01
-1.44963658e+00 -1.84811085e-01 -2.52614826e-01 1.20393164e-01
1.64643660e-01 8.16458702e-01 -7.54143298e-02 9.13951337e-01
8.28942835e-01 1.40293908e+00 -2.13271201e-01 8.62168550e-01
-2.04560924e-02 7.44282067e-01 -9.33510959e-02 3.79630238e-01
-5.92680052e-02 -5.54811060e-01 7.13257432e-01 1.06222701e+00
2.56458431e-01 4.80034322e-01 2.23651379e-01 4.91959065e-01
-3.64697054e-02 -5.19464202e-02 -3.97896111e-01 -4.82528545e-02
3.12373638e-01 1.16436541e+00 -2.43601501e-01 -3.29455167e-01
-3.19084167e-01 1.09851444e+00 -9.78148431e-02 3.93489540e-01
-7.38491774e-01 -8.36902320e-01 5.18253684e-01 9.00552124e-02
2.80741543e-01 -7.37139136e-02 -2.14419603e-01 -1.21896279e+00
2.39480600e-01 -9.36984420e-01 3.74209911e-01 -3.34157944e-01
-1.31574631e+00 8.49542379e-01 -1.95849270e-01 -1.64608181e+00
-4.52689566e-02 -3.34160268e-01 -4.85197604e-01 1.19958377e+00
-1.52526176e+00 -9.35213804e-01 -3.58892143e-01 8.25921297e-01
2.61463702e-01 -6.17647097e-02 9.96637642e-01 8.38581800e-01
-4.44741994e-01 6.20677590e-01 2.48203427e-01 2.50350982e-01
6.60287440e-01 -1.08118224e+00 -2.78890282e-01 5.67243695e-01
-4.15104687e-01 5.60467482e-01 3.25612515e-01 -3.73933285e-01
-1.38023245e+00 -1.05575478e+00 6.92742705e-01 1.69426143e-01
3.00459802e-01 -3.80534649e-01 -1.09135580e+00 4.14850265e-01
-5.85521013e-02 1.13409095e-01 8.48962069e-01 -5.05029783e-02
-4.78903092e-02 -6.24410033e-01 -1.24938381e+00 2.55468786e-01
7.92661786e-01 -6.90826178e-01 -7.37975359e-01 -7.88289383e-02
4.97465074e-01 -3.27202260e-01 -1.42579734e+00 6.62881196e-01
8.20734322e-01 -7.80253649e-01 1.00094151e+00 -2.12013319e-01
1.70998335e-01 -4.05655861e-01 9.22916457e-03 -1.35224593e+00
-6.32613242e-01 -5.45247555e-01 -2.52357155e-01 1.08428574e+00
-1.67517632e-01 -8.37318659e-01 3.95676762e-01 -6.75345585e-02
-2.77023017e-01 -1.17210340e+00 -1.15095162e+00 -6.89176679e-01
-2.16384754e-01 -1.65831015e-01 3.69392455e-01 9.19119596e-01
1.70139074e-02 5.68239152e-01 -7.17509687e-01 -7.53518939e-02
8.47984552e-01 6.57832175e-02 1.52342930e-01 -1.37974787e+00
-1.78916544e-01 8.92804563e-02 -7.48348534e-01 -1.03592324e+00
-7.37721249e-02 -8.98118675e-01 -1.57633886e-01 -1.31785011e+00
1.68859109e-01 -4.29804921e-01 -8.47981989e-01 3.32642555e-01
-1.04846492e-01 3.08276147e-01 6.94587268e-03 3.05115402e-01
-7.66820163e-02 8.12730789e-01 1.49329698e+00 -1.09154686e-01
-2.77112156e-01 -1.51399216e-02 -5.10119319e-01 5.93707740e-01
6.55378878e-01 -6.03722274e-01 -5.44073582e-01 -2.08598465e-01
-4.26595688e-01 3.91738057e-01 2.21884623e-01 -1.09588373e+00
2.34372616e-02 4.78787243e-01 7.10373819e-01 -7.12528288e-01
5.63809335e-01 -8.92202437e-01 2.94232816e-01 7.83560634e-01
-1.03153542e-01 -8.11263621e-02 1.68968052e-01 7.40162909e-01
-6.37660861e-01 7.99606293e-02 9.38031614e-01 3.56495269e-02
-2.32247248e-01 5.15198469e-01 -1.61440521e-01 -9.54318568e-02
8.52411330e-01 -2.99431026e-01 1.64320603e-01 -1.17689058e-01
-1.09172177e+00 9.47377831e-02 -1.22382550e-03 1.69483870e-01
9.92474139e-01 -1.69177127e+00 -6.40830755e-01 5.80383062e-01
-4.13239934e-02 -1.47192210e-01 6.65331006e-01 1.33536386e+00
-2.06695527e-01 1.61149606e-01 -2.18326002e-01 -7.71864772e-01
-1.20444298e+00 4.45318073e-01 4.08593237e-01 -5.76637626e-01
-1.28406858e+00 4.44630682e-01 2.95407534e-01 -2.94542895e-03
2.48804346e-01 -3.62764210e-01 -3.36473256e-01 2.14322090e-01
5.08600652e-01 3.72125953e-01 7.74606541e-02 -4.67788339e-01
-5.10794282e-01 8.82581294e-01 5.41002750e-02 1.72748387e-01
1.44730818e+00 -6.99508637e-02 -3.17392409e-01 5.59694588e-01
1.51051664e+00 -2.93312997e-01 -1.00496173e+00 -3.33204567e-01
-2.88473427e-01 -5.14590144e-01 4.53247875e-01 -5.49749792e-01
-1.20219839e+00 1.28067374e+00 1.24201727e+00 7.98753202e-02
1.65559959e+00 -5.48639953e-01 1.24161637e+00 7.45197162e-02
2.57233202e-01 -8.69619668e-01 1.96693745e-02 1.22789674e-01
8.63525271e-01 -7.96818376e-01 -9.89806280e-02 -2.97993928e-01
-4.34095263e-01 1.22373796e+00 8.69480297e-02 -2.01172575e-01
1.09853089e+00 2.38998741e-01 2.84780301e-02 -1.13828734e-01
-4.59539771e-01 2.31639534e-01 2.71802723e-01 5.01519203e-01
3.55247706e-01 -8.95101279e-02 -6.35191917e-01 9.87700224e-01
3.35778445e-01 2.99797446e-01 1.39774576e-01 9.21395004e-01
-1.55178204e-01 -1.17019141e+00 -1.20795012e-01 6.87501013e-01
-6.54700279e-01 -5.82881048e-02 1.38791338e-01 5.04083455e-01
2.99115598e-01 7.29398370e-01 -2.61830270e-01 -7.06018209e-01
4.51989532e-01 2.45367914e-01 3.65394562e-01 -4.05867279e-01
-4.99731421e-01 5.69160879e-01 -2.38467827e-01 -6.22704923e-01
-4.04973924e-01 -5.03223479e-01 -1.07235885e+00 -2.16992106e-02
-2.04686150e-01 3.68343145e-01 3.12962234e-01 6.86807632e-01
3.39729041e-01 9.00405586e-01 8.73256683e-01 -6.28474534e-01
-7.05496192e-01 -1.03108168e+00 -1.10360897e+00 5.07229388e-01
6.41912460e-01 -6.66996539e-01 -5.15433311e-01 3.45531479e-02] | [14.150919914245605, 3.258960247039795] |
152e5508-c048-496a-972d-73d398f3ee27 | 190513313 | 1905.13313 | null | https://arxiv.org/abs/1905.13313v5 | https://arxiv.org/pdf/1905.13313v5.pdf | Technical Report of the Video Event Reconstruction and Analysis (VERA) System -- Shooter Localization, Models, Interface, and Beyond | Every minute, hundreds of hours of video are uploaded to social media sites and the Internet from around the world. This material creates a visual record of the experiences of a significant percentage of humanity and can help illuminate how we live in the present moment. When properly analyzed, this video can also help analysts to reconstruct events of interest, including war crimes, human rights violations, and terrorist acts. Machine learning and computer vision can play a crucial role in this process. In this technical report, we describe the Video Event Reconstruction and Analysis (VERA) system. This new tool brings together a variety of capabilities we have developed over the past few years (including video synchronization and geolocation to order unstructured videos lacking metadata over time and space, and sound recognition algorithms) to enable the reconstruction and analysis of events captured on video. Among other uses, VERA enables the localization of a shooter from just a few videos that include the sound of gunshots. To demonstrate the efficacy of this suite of tools, we present the results of estimating the shooter's location of the Las Vegas Shooting in 2017 and show that VERA accurately predicts the shooter's location using only the first few gunshots. We then point out future directions that can help improve the system and further reduce unnecessary human labor in the process. All of the components of VERA run through a web interface that enables human-in-the-loop verification to ensure accurate estimations. All relevant source code, including the web interface and machine learning models, is freely available on Github. We hope that researchers and software developers will be inspired to improve and expand this system moving forward to better meet the needs of human rights and public safety. | ['Alexander Hauptmann', 'Junwei Liang', 'Jay D. Aronson'] | 2019-05-26 | null | null | null | null | ['shooter-localization', 'gunshot-detection', 'video-synchronization'] | ['audio', 'audio', 'computer-vision'] | [-5.23081347e-02 -3.71125817e-01 -1.29346903e-02 -3.18095274e-02
-8.73615742e-01 -8.15477729e-01 4.97506797e-01 3.74329239e-01
-3.83347660e-01 2.31936723e-01 4.50986624e-01 -3.57986271e-01
1.57605298e-02 -7.09413469e-01 -3.80909711e-01 -3.11285138e-01
-2.92164296e-01 6.58254921e-02 4.77435231e-01 -1.83177106e-02
3.95571887e-01 6.98830485e-01 -1.36318219e+00 4.66128200e-01
-2.78531499e-02 8.92931223e-01 1.10669471e-01 7.71366894e-01
2.96894163e-01 1.24867165e+00 -7.38204896e-01 -2.04414129e-01
1.70349181e-01 -1.78750157e-01 -6.99137986e-01 -3.54092449e-01
2.54683733e-01 -8.11482966e-01 -6.33792222e-01 5.16720533e-01
4.46777225e-01 1.26659393e-01 8.20904970e-02 -1.37941897e+00
2.09633425e-01 3.24539781e-01 -4.57196414e-01 7.06070721e-01
9.93067861e-01 2.95042396e-01 3.87769014e-01 -6.29028082e-01
7.43831336e-01 8.36325884e-01 1.03798437e+00 -2.99951900e-02
-5.73724210e-01 -9.38025594e-01 -4.07185405e-01 5.81403494e-01
-1.56927645e+00 -5.39815307e-01 6.18198216e-01 -6.83422327e-01
1.12664747e+00 3.21821809e-01 6.81306541e-01 1.19439924e+00
4.93839383e-02 4.73130852e-01 3.80927175e-01 -4.07853901e-01
7.69735575e-02 -2.72563726e-01 -2.45573506e-01 6.84298575e-01
-9.70716327e-02 -1.06294021e-01 -7.77115941e-01 -5.20685971e-01
5.61896384e-01 3.13365191e-01 -3.07788432e-01 1.05361819e-01
-1.37788117e+00 7.50346839e-01 5.88280186e-02 2.83872098e-01
-5.18504798e-01 6.68602511e-02 6.46630168e-01 -1.66300401e-01
3.61255825e-01 2.39130810e-01 9.94696245e-02 -9.41353381e-01
-1.17163301e+00 2.32551754e-01 6.46960139e-01 6.29372537e-01
5.22824585e-01 -8.12493712e-02 1.72257110e-01 4.88697410e-01
-7.76618421e-02 3.67774516e-01 2.79411245e-02 -1.25438154e+00
5.88837087e-01 2.37818629e-01 2.20046386e-01 -1.59099555e+00
-5.94351709e-01 1.43679753e-01 -1.51210263e-01 4.21066657e-02
4.80028927e-01 -3.95375639e-01 -2.01654464e-01 1.09954906e+00
5.23218691e-01 7.05751479e-01 -6.23094261e-01 8.48199606e-01
6.27049983e-01 9.04754043e-01 -1.35624200e-01 -1.20095596e-01
1.33075547e+00 -2.77452528e-01 -7.50757515e-01 -9.48130712e-02
6.31279171e-01 -9.22242105e-01 6.05732679e-01 4.90792751e-01
-9.06097174e-01 -1.90608129e-01 -8.51916492e-01 2.35337660e-01
-2.03987211e-01 -2.10234195e-01 4.87733513e-01 4.55267638e-01
-6.63781583e-01 6.39092624e-01 -1.17182219e+00 -7.37969398e-01
5.10770679e-01 -1.37330636e-01 -5.14670014e-01 -9.72644463e-02
-1.03857636e+00 9.44136500e-01 1.50946125e-01 -4.83806804e-02
-8.62978518e-01 -6.95431113e-01 -8.41280162e-01 -2.14014560e-01
4.92890805e-01 -4.66877185e-02 1.31671286e+00 -3.93229246e-01
-7.91522264e-01 6.61459863e-01 2.65601054e-02 -5.14095008e-01
4.41142619e-01 -2.72417367e-01 -5.97141862e-01 6.85129881e-01
4.26466882e-01 9.21664089e-02 6.35450065e-01 -6.93359256e-01
-9.95697975e-01 -2.07330540e-01 7.58738816e-02 -1.43481806e-01
-2.91996270e-01 8.46875429e-01 -4.77596760e-01 -5.11062384e-01
-2.87347943e-01 -7.24210083e-01 3.18467408e-01 -1.48683369e-01
3.55712883e-02 3.62809032e-01 1.19585073e+00 -1.23559570e+00
1.53250909e+00 -2.26263499e+00 -3.92877370e-01 9.59077850e-02
-1.71651304e-01 2.48788580e-01 2.29133546e-01 1.04258990e+00
-5.12004308e-02 -5.21977022e-02 1.11025348e-01 -1.26080692e-01
-2.65609682e-01 -2.38288552e-01 -5.98183513e-01 8.67382109e-01
-3.38522226e-01 2.78240800e-01 -1.13208663e+00 -2.91492641e-01
4.94779289e-01 5.59849143e-01 -3.53983581e-01 3.02396435e-02
3.74927193e-01 5.48708975e-01 -2.54324615e-01 7.34944880e-01
3.48240346e-01 1.93626031e-01 3.56934480e-02 -1.61870420e-01
-4.49453205e-01 2.58746266e-01 -1.11687863e+00 1.56464577e+00
-2.52631515e-01 1.12923396e+00 3.09370279e-01 -7.51327276e-01
4.96215612e-01 6.10182524e-01 7.01482475e-01 -4.88347530e-01
2.10364223e-01 -1.79899141e-01 -6.44512057e-01 -9.56045687e-01
6.62715971e-01 1.85126737e-01 -1.81949809e-01 8.37396264e-01
-2.24638790e-01 1.72229484e-02 2.36432254e-01 4.48494166e-01
1.58186817e+00 3.63227352e-02 3.16234887e-01 5.56361735e-01
-7.05043077e-02 2.62919486e-01 3.52568507e-01 6.41092181e-01
-3.52787346e-01 6.60420775e-01 3.55372012e-01 -5.21976411e-01
-1.05737293e+00 -8.08578193e-01 7.56589770e-02 8.86115611e-01
-1.91372275e-01 -6.99915946e-01 -7.70017624e-01 -2.72366017e-01
-3.26696545e-01 9.56493199e-01 -3.19857568e-01 9.17000771e-02
-6.07861757e-01 -2.91098833e-01 8.95174325e-01 5.50938368e-01
3.90910119e-01 -1.00723195e+00 -1.05450892e+00 -1.45664613e-03
-8.12799275e-01 -1.35585392e+00 -3.15297365e-01 -5.90356946e-01
-3.70550603e-01 -1.31487381e+00 -2.27752373e-01 -2.31504306e-01
2.23548830e-01 5.56839883e-01 5.95429003e-01 1.94213554e-01
-6.25306308e-01 9.29381430e-01 -8.05636287e-01 -4.79633570e-01
-2.89142370e-01 -4.05845791e-01 8.91917199e-02 -5.51587753e-02
2.78734982e-01 -7.10890651e-01 -3.60510468e-01 2.44604155e-01
-9.67656732e-01 7.78212994e-02 -3.98168772e-01 4.29328009e-02
3.77058238e-02 2.49300137e-01 1.93721667e-01 -4.73651469e-01
4.85971600e-01 -1.06385314e+00 -5.10394514e-01 -1.23489387e-01
5.66518381e-02 -8.67481947e-01 3.49366277e-01 -1.54294133e-01
-6.30798995e-01 4.54714615e-03 1.66814439e-02 -5.95171988e-01
-3.87297660e-01 5.74214160e-01 2.74079263e-01 2.45028794e-01
6.36066616e-01 -8.05057660e-02 -1.83244571e-01 -4.47672963e-01
-8.49722475e-02 1.04623461e+00 9.46869493e-01 -6.11986071e-02
6.41036570e-01 1.00786948e+00 -2.57883459e-01 -8.95080030e-01
-6.39612973e-01 -7.70419955e-01 -4.73846465e-01 -8.81478965e-01
7.43742764e-01 -8.84060562e-01 -8.85088146e-01 5.96134126e-01
-1.20825958e+00 -2.43979469e-01 1.35447890e-01 7.05562711e-01
-4.90738183e-01 4.03831512e-01 -4.78634447e-01 -8.98553789e-01
-6.44481853e-02 -6.38880551e-01 7.71988750e-01 2.42212251e-01
-6.38655007e-01 -7.36687005e-01 2.06185162e-01 7.40763962e-01
2.62161523e-01 7.11534381e-01 2.56450057e-01 -4.76966172e-01
-4.09725368e-01 -8.09526682e-01 -5.20807095e-02 -4.00003977e-02
1.30585164e-01 1.54215783e-01 -9.13574219e-01 -2.59373218e-01
-9.38587934e-02 8.73301923e-02 4.64735359e-01 3.84963959e-01
8.59588385e-01 -3.61240834e-01 -3.75788093e-01 5.59568286e-01
1.06625390e+00 5.09380758e-01 7.85494149e-01 5.89382231e-01
6.85341418e-01 4.89472270e-01 8.75981510e-01 9.88082290e-01
4.67487395e-01 8.12495172e-01 6.77527368e-01 2.44278640e-01
9.28335488e-02 -3.86359394e-01 4.86018062e-01 3.24602813e-01
-4.08445477e-01 -2.01279551e-01 -1.12462509e+00 6.73271477e-01
-1.68169212e+00 -1.84295356e+00 -1.18960112e-01 2.36908436e+00
2.24187717e-01 -1.39593795e-01 4.34272230e-01 2.81411171e-01
8.27521086e-01 3.29402089e-01 -8.89435131e-03 -3.86426896e-01
3.58466804e-01 3.96050848e-02 6.47305369e-01 4.69269514e-01
-1.19190979e+00 6.45884812e-01 6.42345953e+00 7.48653710e-01
-1.29203892e+00 2.76298285e-01 -5.47882728e-03 -5.31912088e-01
1.82451203e-01 2.87009031e-01 -4.13580537e-01 6.39624536e-01
1.15170741e+00 -1.08733006e-01 8.01000953e-01 7.32110858e-01
9.17364001e-01 -5.14693618e-01 -7.02831805e-01 1.05557740e+00
4.03042614e-01 -1.85700464e+00 -6.19584203e-01 -3.56383957e-02
1.04681537e-01 7.08861351e-02 -3.50884289e-01 -4.74925153e-02
-9.51968580e-02 -7.33790696e-01 1.04320097e+00 6.26940906e-01
6.92734003e-01 -9.48929429e-01 4.70901459e-01 5.65361619e-01
-1.09051597e+00 -2.31247783e-01 9.50807109e-02 -5.37262559e-01
8.25855732e-01 4.10775393e-01 -1.09359980e+00 3.32924664e-01
1.04181790e+00 7.82241046e-01 -4.48176593e-01 1.29490662e+00
-1.25873044e-01 8.59503031e-01 -3.97279143e-01 3.51715893e-01
2.02154531e-03 1.95357203e-01 9.05845284e-01 1.31180799e+00
6.15883172e-01 4.00643378e-01 -2.13889163e-02 3.40109438e-01
3.46944153e-01 -3.21317375e-01 -8.34670007e-01 -6.45735040e-02
7.04249859e-01 1.24831808e+00 -8.57468128e-01 -4.09581363e-02
-3.39677513e-01 7.87338555e-01 -3.11502963e-02 1.11439861e-01
-1.18909323e+00 -5.69822609e-01 4.85514015e-01 7.96576619e-01
1.80226207e-01 -5.48689365e-01 1.55600488e-01 -9.33108807e-01
-2.62924936e-02 -8.46310139e-01 5.78849018e-01 -1.24391842e+00
-6.57453179e-01 4.19304937e-01 3.68194103e-01 -1.32259047e+00
-5.06352544e-01 -6.32857829e-02 -8.83953154e-01 2.92737573e-01
-6.38898849e-01 -9.59788561e-01 -3.92187983e-01 5.21679878e-01
3.25169712e-01 -3.01271677e-02 7.39033520e-01 6.25939727e-01
-5.14947891e-01 7.08106905e-02 -1.51647329e-01 4.59294379e-01
7.39809692e-01 -4.27918017e-01 3.73312771e-01 1.11966157e+00
3.54842544e-01 3.23147655e-01 8.34846973e-01 -9.02435720e-01
-1.23849940e+00 -7.79129863e-01 8.44247758e-01 -6.91634297e-01
1.02622187e+00 -2.54017800e-01 -4.97839361e-01 1.08279335e+00
3.12329270e-02 -2.99965203e-01 7.80986607e-01 -1.05080269e-01
-1.76766559e-01 1.97239351e-02 -9.78811681e-01 3.16353947e-01
5.52886903e-01 -8.24700594e-01 -4.78947371e-01 5.06525636e-01
1.14645243e-01 -5.83543003e-01 -6.00850523e-01 1.73518687e-01
5.64637899e-01 -1.29457366e+00 9.16951001e-01 -2.56683379e-01
2.31838837e-01 -4.27557856e-01 -4.44593392e-02 -8.72141480e-01
-9.33916494e-03 -7.07472503e-01 7.01884702e-02 1.22323990e+00
-6.90826401e-02 -4.60074037e-01 3.75805080e-01 7.59206474e-01
-5.31050600e-02 -1.72192708e-01 -1.03998971e+00 -5.91028571e-01
-6.80248678e-01 -1.32007289e+00 4.59351212e-01 1.06970143e+00
5.16815841e-01 -3.10776055e-01 -6.78279936e-01 4.27975267e-01
3.90461594e-01 -2.85806537e-01 8.42631638e-01 -7.10367322e-01
-2.00988963e-01 7.09160492e-02 -7.36231804e-01 -3.11065227e-01
-2.20871538e-01 -5.60640216e-01 -3.39521348e-01 -1.38912880e+00
1.80032700e-01 -2.12834910e-01 2.39755839e-01 6.37049198e-01
4.43412989e-01 5.15663683e-01 3.93261373e-01 3.48896831e-01
-6.31381094e-01 -2.01557383e-01 4.88648504e-01 1.41515270e-01
-1.40192732e-02 6.61071315e-02 -3.36620420e-01 1.13514173e+00
7.92578697e-01 -7.65507221e-01 -2.94733103e-02 -5.50682724e-01
2.85159081e-01 3.53111893e-01 8.54057193e-01 -1.31463981e+00
4.55498248e-01 -2.10904628e-01 2.93676138e-01 -6.91037655e-01
5.76179743e-01 -8.05306971e-01 5.21650434e-01 1.92317978e-01
6.68813065e-02 1.50466725e-01 4.28636491e-01 4.15919900e-01
-1.85084060e-01 -2.08571643e-01 3.52466196e-01 -3.14160511e-02
-5.88939011e-01 -4.83153164e-02 -8.22815657e-01 -6.74288943e-02
1.44269323e+00 -3.05127323e-01 -4.59703177e-01 -1.02056420e+00
-6.20008588e-01 -1.32998414e-02 5.37375331e-01 4.72426176e-01
4.48945314e-01 -1.09060431e+00 -5.81845164e-01 -4.60978076e-02
-1.72049999e-01 -5.62762141e-01 6.11239791e-01 9.61661220e-01
-1.07696819e+00 2.06240430e-01 -3.15640748e-01 -3.55968237e-01
-1.46554840e+00 4.59798425e-01 2.56945416e-02 2.08752170e-01
-7.12617576e-01 5.07463336e-01 -5.22106767e-01 1.02192245e-01
2.68287241e-01 1.35823160e-01 -2.17739120e-01 2.82825798e-01
1.26622069e+00 9.04164612e-01 2.88608093e-02 -1.13739467e+00
-6.52227581e-01 1.61378905e-01 2.61888236e-01 -2.58363187e-01
1.54673529e+00 -2.87411898e-01 4.50386889e-02 3.43595356e-01
1.02317929e+00 4.09111261e-01 -1.21579218e+00 2.29434162e-01
-3.47734243e-01 -8.20077419e-01 1.31081477e-01 -4.85873222e-01
-8.92620862e-01 8.95788014e-01 3.96827638e-01 3.70199740e-01
1.09984112e+00 8.26285109e-02 1.11009479e+00 -1.00778751e-02
6.13164306e-01 -1.14859855e+00 -2.36234680e-01 2.72442788e-01
7.28595853e-01 -8.22685122e-01 2.48036504e-01 -5.83127029e-02
-7.26993501e-01 1.19402754e+00 4.44306573e-03 3.82212289e-02
5.69499433e-01 5.25163352e-01 4.91357781e-02 -2.68303961e-01
-4.89651114e-01 5.60679473e-02 -2.85264969e-01 7.62557507e-01
3.72647047e-02 -2.28629094e-02 8.60580876e-02 4.93650317e-01
-2.19078660e-01 1.08558796e-01 7.74637103e-01 1.13313174e+00
-4.50462669e-01 -7.50268042e-01 -9.54709828e-01 1.69229969e-01
-7.31571436e-01 1.14133544e-01 -2.05493867e-01 5.21551847e-01
1.89859003e-01 1.34468949e+00 2.27572352e-01 -5.41963935e-01
2.07113639e-01 -2.16616273e-01 2.64679670e-01 -3.53293628e-01
-6.80827200e-01 -2.44348466e-01 3.88896495e-01 -9.93625402e-01
-2.30765045e-01 -9.93213475e-01 -1.23043942e+00 -7.56897748e-01
-2.21938770e-02 -2.98483800e-02 8.96900952e-01 8.75791669e-01
4.18604523e-01 5.12281917e-02 4.11350131e-01 -1.32148612e+00
1.95234537e-01 -6.88951910e-01 -4.16237950e-01 2.18742475e-01
1.65533587e-01 -5.13564944e-01 -4.03248668e-01 3.65566999e-01] | [8.305780410766602, 0.3097838759422302] |
31a97517-8093-4c2b-af1b-58c1105ae6b0 | redirtrans-latent-to-latent-translation-for-1 | 2305.11452 | null | https://arxiv.org/abs/2305.11452v1 | https://arxiv.org/pdf/2305.11452v1.pdf | ReDirTrans: Latent-to-Latent Translation for Gaze and Head Redirection | Learning-based gaze estimation methods require large amounts of training data with accurate gaze annotations. Facing such demanding requirements of gaze data collection and annotation, several image synthesis methods were proposed, which successfully redirected gaze directions precisely given the assigned conditions. However, these methods focused on changing gaze directions of the images that only include eyes or restricted ranges of faces with low resolution (less than $128\times128$) to largely reduce interference from other attributes such as hairs, which limits application scenarios. To cope with this limitation, we proposed a portable network, called ReDirTrans, achieving latent-to-latent translation for redirecting gaze directions and head orientations in an interpretable manner. ReDirTrans projects input latent vectors into aimed-attribute embeddings only and redirects these embeddings with assigned pitch and yaw values. Then both the initial and edited embeddings are projected back (deprojected) to the initial latent space as residuals to modify the input latent vectors by subtraction and addition, representing old status removal and new status addition. The projection of aimed attributes only and subtraction-addition operations for status replacement essentially mitigate impacts on other attributes and the distribution of latent vectors. Thus, by combining ReDirTrans with a pretrained fixed e4e-StyleGAN pair, we created ReDirTrans-GAN, which enables accurately redirecting gaze in full-face images with $1024\times1024$ resolution while preserving other attributes such as identity, expression, and hairstyle. Furthermore, we presented improvements for the downstream learning-based gaze estimation task, using redirected samples as dataset augmentation. | ['Truong Nguyen', 'Ning Bi', 'Lei Wang', 'Zhen Wang', 'Shiwei Jin'] | 2023-05-19 | redirtrans-latent-to-latent-translation-for | http://openaccess.thecvf.com//content/CVPR2023/html/Jin_ReDirTrans_Latent-to-Latent_Translation_for_Gaze_and_Head_Redirection_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Jin_ReDirTrans_Latent-to-Latent_Translation_for_Gaze_and_Head_Redirection_CVPR_2023_paper.pdf | cvpr-2023-1 | ['gaze-estimation'] | ['computer-vision'] | [ 4.80757535e-01 3.34565729e-01 -1.78296074e-01 -7.40205109e-01
-1.67971924e-01 -5.39220572e-01 2.92365819e-01 -7.73753226e-01
-3.09315026e-01 5.33983231e-01 2.52148211e-01 -1.37665616e-02
1.37915716e-01 -4.37702239e-01 -7.46186137e-01 -7.99502254e-01
3.26547891e-01 -2.30699569e-01 -3.08992684e-01 -8.63782912e-02
2.47898817e-01 2.86186635e-01 -2.14835501e+00 3.84086892e-02
8.05787086e-01 1.04354930e+00 -6.39767498e-02 4.86842245e-01
-2.83394270e-02 5.65666199e-01 -5.75434566e-01 -4.06324089e-01
3.80226970e-01 -4.86941725e-01 -2.79898584e-01 -5.22626415e-02
9.00117576e-01 -6.22362733e-01 -3.99385504e-02 9.43857312e-01
5.15709698e-01 7.55642205e-02 5.61672390e-01 -1.57251072e+00
-1.11954272e+00 1.41116261e-01 -1.10580826e+00 -2.34628364e-01
5.35034359e-01 3.59282464e-01 8.00945640e-01 -8.84275436e-01
4.70542282e-01 1.14753890e+00 4.53017324e-01 1.16664279e+00
-1.18655396e+00 -1.44200993e+00 2.17675969e-01 1.75063580e-01
-1.53863716e+00 -7.81192839e-01 7.60981560e-01 -3.02141339e-01
6.16404831e-01 5.75213015e-01 4.46857929e-01 1.36502981e+00
-1.52249292e-01 3.60970795e-01 1.12597525e+00 -4.91347075e-01
-2.25473881e-01 2.62008488e-01 -1.70678139e-01 6.18768930e-01
9.41809919e-03 -1.43384933e-01 -7.88039684e-01 1.60600469e-01
7.10332870e-01 1.32982641e-01 -6.63473964e-01 -4.20795441e-01
-1.16914797e+00 6.82795405e-01 5.55894077e-01 -2.23469108e-01
-1.84447140e-01 8.50683898e-02 2.84309275e-02 -1.28664691e-02
4.79717493e-01 4.80774492e-01 -4.89056289e-01 -5.65442741e-02
-7.52191722e-01 -1.78749457e-01 1.48871496e-01 1.15929425e+00
1.15487683e+00 1.11766234e-01 -3.23362172e-01 8.10140610e-01
5.29619932e-01 9.52067256e-01 5.37027597e-01 -9.25932765e-01
4.01019484e-01 7.73761392e-01 -4.96817306e-02 -9.92313862e-01
-2.45507449e-01 3.27828303e-02 -6.85919583e-01 3.38303328e-01
2.88763583e-01 -2.32019782e-01 -1.20589066e+00 2.32132554e+00
4.20283318e-01 1.27188444e-01 -2.24261850e-01 1.00126517e+00
8.31544578e-01 5.65874219e-01 6.36010990e-02 -2.55558252e-01
1.51801515e+00 -7.30219245e-01 -9.79349494e-01 -2.42003962e-01
5.27796388e-01 -7.94814944e-01 1.69820058e+00 7.80009627e-02
-7.65445828e-01 -4.79142666e-01 -1.02296185e+00 -4.33401525e-01
-1.41864032e-01 3.90279770e-01 4.92396921e-01 1.04925156e+00
-1.37755156e+00 6.22679219e-02 -6.02305174e-01 -5.89032993e-02
5.10194898e-01 8.63075614e-01 -5.87486446e-01 5.75818047e-02
-9.89829063e-01 5.46534896e-01 -2.30911881e-01 3.58901829e-01
-3.43540728e-01 -8.19175899e-01 -1.20621777e+00 1.53937310e-01
3.52010727e-01 -4.45112318e-01 8.98746490e-01 -1.19607770e+00
-1.91834903e+00 8.66944313e-01 -5.52202344e-01 3.04015338e-01
6.77740574e-02 -1.23596787e-01 -4.67978865e-01 -2.28939578e-01
-7.44111463e-02 1.11183250e+00 1.54130828e+00 -1.02680457e+00
-3.76627982e-01 -6.35060370e-01 1.47687107e-01 3.03089082e-01
-8.85002494e-01 3.95850129e-02 -6.23972058e-01 -4.78857130e-01
-5.92154860e-02 -1.10453641e+00 4.72979724e-01 1.44180298e-01
-4.02394861e-01 -1.34040508e-02 1.07815552e+00 -5.00183642e-01
1.30812371e+00 -2.35103679e+00 2.21917003e-01 1.13526210e-01
5.79350650e-01 8.69868994e-02 -3.93679947e-01 -2.78428257e-01
-5.14972210e-01 1.29446685e-01 -2.14976966e-02 -7.30480909e-01
-3.64384316e-02 1.00381359e-01 -3.20430487e-01 5.19778073e-01
1.75737709e-01 7.86125958e-01 -5.39907157e-01 -3.71633053e-01
1.53582960e-01 8.24997604e-01 -8.44823956e-01 2.98382699e-01
4.80412059e-02 4.08917964e-01 -7.91760087e-02 5.16894042e-01
9.14187551e-01 -3.63797583e-02 4.51264456e-02 -4.11170304e-01
-3.95089872e-02 2.57887006e-01 -8.22909951e-01 1.34820843e+00
-5.19124389e-01 9.62264419e-01 -2.31633224e-02 -1.37674183e-01
1.02169573e+00 6.06344640e-03 3.56537640e-01 -7.69894481e-01
3.66788000e-01 -2.98572242e-01 -1.91375855e-02 -4.14175808e-01
6.37633801e-01 1.43173486e-01 -1.90824606e-02 6.79817677e-01
1.26521736e-02 2.44255811e-01 -2.61657923e-01 -7.90670067e-02
5.17211199e-01 4.00251687e-01 -1.05396464e-01 1.01856872e-01
4.96311277e-01 -5.77341437e-01 6.14452660e-01 -7.64478222e-02
-1.13869682e-01 8.67659748e-01 7.38471389e-01 1.14318426e-03
-8.83992553e-01 -7.46788681e-01 -2.55481787e-02 1.49690413e+00
7.59229586e-02 -5.15869975e-01 -1.09010112e+00 -8.89237046e-01
-2.84248292e-01 8.04224372e-01 -1.11055601e+00 -3.91545951e-01
-5.42942286e-01 -6.24406040e-01 5.07526517e-01 3.79002869e-01
3.18417937e-01 -1.04709029e+00 -4.26336944e-01 -5.60636520e-01
-5.30810542e-02 -6.49677098e-01 -1.07352948e+00 -2.55944908e-01
-3.14499676e-01 -1.14406407e+00 -7.83470809e-01 -5.71512818e-01
1.21177518e+00 2.22277686e-01 4.60652918e-01 -1.19943582e-01
1.17303897e-02 7.46776462e-02 -2.48612449e-01 -3.32892597e-01
-5.59824426e-03 2.02482082e-02 3.27613294e-01 5.23036361e-01
5.06737471e-01 -3.58988523e-01 -7.36019135e-01 5.07815301e-01
-7.25641489e-01 1.93148777e-01 3.08092892e-01 8.17372918e-01
2.79880702e-01 -6.06674492e-01 2.76769966e-01 -8.30349147e-01
4.96901900e-01 -9.05135199e-02 -5.47905684e-01 1.34713531e-01
-8.29670608e-01 4.07899842e-02 4.11300659e-01 -6.96238399e-01
-1.23556232e+00 -1.99720591e-01 1.25102103e-02 -9.42918181e-01
-6.72815815e-02 -1.65546745e-01 -4.78401095e-01 -4.50037234e-02
6.74312949e-01 -5.57527971e-03 3.30306828e-01 -1.55252963e-01
4.68978554e-01 9.85243440e-01 3.49723756e-01 -6.12367764e-02
7.47670531e-01 2.57934004e-01 -2.21865490e-01 -5.55555701e-01
-6.14861846e-01 3.08650080e-02 -6.59168243e-01 -1.26250267e-01
8.47130656e-01 -7.95140564e-01 -1.04667306e+00 7.18740106e-01
-8.70549023e-01 -3.02400291e-01 -2.69507110e-01 3.36429775e-01
-1.99539334e-01 4.49445322e-02 -1.27364054e-01 -5.95224082e-01
-4.16855335e-01 -1.27081239e+00 1.10320985e+00 4.98690695e-01
-4.03657585e-01 -5.60233414e-01 -1.76535234e-01 3.53973985e-01
4.32690054e-01 -4.11503110e-03 7.81010151e-01 -1.03253447e-01
-3.47258061e-01 -6.63596243e-02 -3.94302040e-01 3.57062191e-01
5.48647344e-01 1.97134838e-01 -1.33489919e+00 -3.35757315e-01
-8.02092701e-02 -2.15457946e-01 5.22919536e-01 2.68014073e-01
1.23369753e+00 -4.13534105e-01 -2.96209484e-01 1.17092478e+00
7.81029880e-01 9.67845097e-02 7.63125360e-01 1.84522197e-01
1.23466027e+00 7.75319815e-01 5.01512527e-01 2.96285510e-01
5.63316703e-01 6.97420955e-01 4.95623112e-01 -2.28052810e-01
-2.67670035e-01 -3.25200975e-01 5.84364712e-01 4.34657097e-01
-6.53779507e-02 -3.30787420e-01 -4.97765064e-01 3.25224847e-01
-1.23108327e+00 -7.58369148e-01 9.96350497e-02 2.31065607e+00
1.14675236e+00 -2.13989615e-01 -2.33263999e-01 -1.12288848e-01
8.61267686e-01 2.90557325e-01 -9.15476978e-01 -3.41732919e-01
3.33955176e-02 3.67233992e-01 5.25664151e-01 2.38807291e-01
-7.14075863e-01 9.40459907e-01 5.55531216e+00 5.89565396e-01
-1.77969003e+00 1.62339136e-01 4.63862479e-01 -6.24169230e-01
-6.05140746e-01 -2.70911783e-01 -1.00839639e+00 5.83487988e-01
6.62286639e-01 3.24634075e-01 6.56927943e-01 6.55915618e-01
6.39397055e-02 2.23069742e-01 -1.02411294e+00 1.24808514e+00
4.32029963e-01 -8.46674144e-01 -1.33386865e-01 3.04108441e-01
5.79498649e-01 -3.34191620e-01 7.88699150e-01 3.32079411e-01
-2.21865505e-01 -1.16504228e+00 6.85444832e-01 4.13793951e-01
1.89083636e+00 -5.63538194e-01 4.64152813e-01 -1.38638809e-01
-6.87500775e-01 -1.58272222e-01 -8.01323503e-02 2.50359662e-02
-8.13813582e-02 -4.09484357e-02 -9.49252129e-01 -1.39598669e-02
8.17889571e-01 6.42979741e-01 -8.59845459e-01 1.60127729e-01
-4.87538010e-01 4.41472709e-01 -4.10390377e-01 1.27213970e-01
-3.18001419e-01 -1.73154563e-01 3.48466486e-01 5.12784958e-01
4.88417327e-01 -1.08342931e-01 -7.85657406e-01 9.46748435e-01
-3.33717883e-01 -1.38466835e-01 -3.69238198e-01 1.88450783e-01
5.79631507e-01 1.33105016e+00 -1.86831132e-01 1.45322382e-02
-2.65837342e-01 9.65204418e-01 1.77234977e-01 4.39832687e-01
-1.02789056e+00 -4.77723062e-01 1.16115236e+00 4.16874319e-01
2.99916536e-01 1.86683774e-01 -3.28712881e-01 -1.28833652e+00
-6.79703206e-02 -8.41997325e-01 -1.66862812e-02 -1.15781689e+00
-8.66385102e-01 7.58679867e-01 4.60733287e-02 -1.14489174e+00
-3.91742468e-01 -4.85779852e-01 -4.23038960e-01 1.19353127e+00
-1.47288442e+00 -1.53642869e+00 -6.65351570e-01 8.88669789e-01
2.04266548e-01 -2.12413669e-01 9.54585075e-01 2.46453404e-01
-9.70870018e-01 1.44272327e+00 -2.61723936e-01 -6.93285931e-03
1.36106801e+00 -9.91721332e-01 1.86699063e-01 7.03360021e-01
-2.47738034e-01 9.76518810e-01 4.59984124e-01 -3.38468164e-01
-1.21146989e+00 -9.63278472e-01 6.68502927e-01 -4.80237156e-01
1.59778908e-01 -7.72616506e-01 -8.81943345e-01 9.13657963e-01
2.38449335e-01 1.32446764e-02 7.52314746e-01 1.45838618e-01
-4.03206885e-01 -3.56475472e-01 -1.12305701e+00 9.08121288e-01
9.81273293e-01 -7.00512290e-01 -1.54884920e-01 -1.79511011e-01
6.29979968e-01 -6.40230834e-01 -4.83513981e-01 4.66776431e-01
7.31962323e-01 -9.48871493e-01 6.97112203e-01 -1.40706033e-01
2.82902658e-01 -4.20390695e-01 1.45229220e-01 -1.09659076e+00
-1.50169626e-01 -7.37498641e-01 -8.17063451e-02 1.52045465e+00
2.49961615e-01 -8.60401750e-01 8.55186462e-01 8.66854668e-01
3.68646160e-02 -6.40431046e-01 -7.86084056e-01 8.66094381e-02
-4.01707649e-01 -2.30893329e-01 1.12223494e+00 1.03518891e+00
-2.51423538e-01 3.69894087e-01 -5.87832868e-01 1.08802989e-01
3.95255029e-01 -2.00481802e-01 1.10415018e+00 -8.29574764e-01
1.93747133e-01 -2.60304004e-01 -3.57575506e-01 -1.17567980e+00
3.53997797e-01 -4.66065586e-01 -8.39822218e-02 -6.55037344e-01
-1.05615929e-01 -6.35976136e-01 -1.80244774e-01 8.96749675e-01
-2.75617510e-01 8.03734124e-01 1.51940927e-01 2.67510116e-01
7.51322061e-02 6.64467633e-01 1.41720438e+00 -3.53989489e-02
-3.50994736e-01 -6.34692535e-02 -9.99508262e-01 6.05466485e-01
5.36200702e-01 -3.24751854e-01 -7.27218032e-01 -6.32541776e-01
4.24545854e-01 -3.33977073e-01 2.56794363e-01 -6.17238224e-01
1.31574422e-01 -1.05639838e-01 5.56373894e-01 -3.12575102e-01
6.67844117e-01 -6.67681813e-01 4.55809645e-02 -2.59876966e-01
-3.34728479e-01 5.59849385e-03 3.00071180e-01 3.08310539e-01
-5.11619449e-02 1.53631121e-02 6.63947701e-01 5.62890887e-01
-5.00626385e-01 3.72530520e-01 1.43287733e-01 -3.28251749e-01
1.22768378e+00 -7.48878717e-01 -3.87702882e-01 -5.33649683e-01
-4.94372308e-01 4.00391668e-02 8.74896824e-01 7.41285682e-01
6.38701200e-01 -1.25100553e+00 -3.24685872e-01 1.06446326e+00
1.17184103e-01 1.26307100e-01 4.64960277e-01 8.89333427e-01
-3.29047024e-01 1.92584485e-01 -5.22170484e-01 -6.72026455e-01
-1.48728633e+00 5.20181596e-01 1.89476430e-01 4.73331839e-01
-1.22635089e-01 1.09280753e+00 7.16335356e-01 -5.31664550e-01
1.35942232e-02 -1.77930310e-01 -5.86669743e-01 2.80403286e-01
5.37164450e-01 5.71837910e-02 -2.86982860e-02 -9.63285565e-01
-4.09160316e-01 8.91271412e-01 -1.86551064e-01 9.80156846e-03
1.13571572e+00 -5.04881978e-01 -1.57786712e-01 6.63051009e-02
1.34406817e+00 5.17817259e-01 -1.56221223e+00 9.90038961e-02
-7.62881160e-01 -7.59197116e-01 8.00769106e-02 -6.00340486e-01
-1.28834987e+00 1.00741899e+00 8.50317359e-01 -2.86647022e-01
1.48034978e+00 -2.25391269e-01 5.84387720e-01 -1.90610081e-01
2.75998600e-02 -6.43501103e-01 -3.60259637e-02 2.50837386e-01
7.63137698e-01 -1.19793320e+00 -9.80202109e-02 -3.05675894e-01
-8.69275868e-01 8.62756729e-01 1.00812578e+00 3.70321214e-01
5.14225781e-01 6.63758442e-02 3.51085275e-01 -9.12359729e-02
-5.59494376e-01 7.12546706e-02 6.34791374e-01 7.76844621e-01
3.70747566e-01 -1.43042758e-01 2.38951728e-01 4.48571205e-01
-5.73366582e-01 -1.26646683e-01 4.91085440e-01 5.15622377e-01
1.64800718e-01 -1.00417209e+00 -5.41939020e-01 4.87761885e-01
-3.35628688e-01 -2.00661808e-01 -1.41701311e-01 7.29287446e-01
4.52457249e-01 5.86345077e-01 5.24243355e-01 -6.73194408e-01
2.30278656e-01 3.62123661e-02 4.26802307e-01 -7.40034938e-01
-2.78426677e-01 1.31556332e-01 -2.17935756e-01 -5.29940069e-01
-3.14545333e-01 -6.01314247e-01 -9.61126149e-01 -3.93254876e-01
-7.02284396e-01 -2.70364970e-01 6.95449471e-01 6.58905864e-01
7.36480117e-01 5.71828604e-01 7.15668321e-01 -1.01305234e+00
-3.54327977e-01 -1.31842768e+00 -5.07133007e-01 3.84844065e-01
5.74522018e-01 -8.59912395e-01 -4.54509884e-01 3.27149481e-01] | [14.06370735168457, 0.0138698760420084] |
97e7ad1f-3adc-4c4b-b5b8-72e186995676 | spacecraft-depth-completion-based-on-the-gray | 2208.14030 | null | https://arxiv.org/abs/2208.14030v1 | https://arxiv.org/pdf/2208.14030v1.pdf | Spacecraft depth completion based on the gray image and the sparse depth map | Perceiving the three-dimensional (3D) structure of the spacecraft is a prerequisite for successfully executing many on-orbit space missions, and it can provide critical input for many downstream vision algorithms. In this paper, we propose to sense the 3D structure of spacecraft using light detection and ranging sensor (LIDAR) and a monocular camera. To this end, Spacecraft Depth Completion Network (SDCNet) is proposed to recover the dense depth map based on gray image and sparse depth map. Specifically, SDCNet decomposes the object-level spacecraft depth completion task into foreground segmentation subtask and foreground depth completion subtask, which segments the spacecraft region first and then performs depth completion on the segmented foreground area. In this way, the background interference to foreground spacecraft depth completion is effectively avoided. Moreover, an attention-based feature fusion module is also proposed to aggregate the complementary information between different inputs, which deduces the correlation between different features along the channel and the spatial dimension sequentially. Besides, four metrics are also proposed to evaluate object-level depth completion performance, which can more intuitively reflect the quality of spacecraft depth completion results. Finally, a large-scale satellite depth completion dataset is constructed for training and testing spacecraft depth completion algorithms. Empirical experiments on the dataset demonstrate the effectiveness of the proposed SDCNet, which achieves 0.25m mean absolute error of interest and 0.759m mean absolute truncation error, surpassing state-of-the-art methods by a large margin. The spacecraft pose estimation experiment is also conducted based on the depth completion results, and the experimental results indicate that the predicted dense depth map could meet the needs of downstream vision tasks. | ['WeiChun Chen', 'Xinlong Chen', 'Yu Chen', 'Zhiqiang Yan', 'Hongyuan Wang', 'Xiang Liu'] | 2022-08-30 | null | null | null | null | ['depth-completion', 'foreground-segmentation'] | ['computer-vision', 'computer-vision'] | [ 4.54699574e-03 -2.80504227e-01 1.04781061e-01 -3.75915468e-01
-3.74905348e-01 -5.52735209e-01 5.91839254e-01 -4.58403915e-01
-3.71238112e-01 4.11818177e-01 -2.78792698e-02 -5.65087870e-02
-2.35096425e-01 -6.05088532e-01 -3.70826274e-01 -8.56744468e-01
1.76030044e-02 5.30387342e-01 2.39132613e-01 9.64085683e-02
2.70386338e-01 6.47874773e-01 -1.45184934e+00 -2.69327879e-01
9.52451408e-01 1.39260697e+00 7.07706094e-01 2.62899816e-01
-1.86760515e-01 1.70633867e-01 -5.22164643e-01 3.13650846e-01
5.06760716e-01 3.12526412e-02 -2.11644441e-01 6.93737864e-01
2.42139220e-01 -6.26581848e-01 -4.59030837e-01 1.26750755e+00
2.16549058e-02 1.16493985e-01 3.91597152e-01 -1.03260589e+00
-1.25820562e-01 -2.37079173e-01 -6.16920948e-01 4.08673346e-01
6.98229447e-02 6.27948701e-01 6.37407959e-01 -1.06744480e+00
2.41461620e-01 1.29786849e+00 2.81637341e-01 1.76215217e-01
-7.08093226e-01 -6.61697745e-01 3.79000574e-01 -1.36497036e-01
-1.30563104e+00 -6.70269579e-02 6.89318478e-01 -5.19016623e-01
6.25500858e-01 1.76764458e-01 7.65469313e-01 4.85459328e-01
2.02184901e-01 6.19519830e-01 9.03804123e-01 9.84408706e-02
2.06866235e-01 -2.53141314e-01 4.44092117e-02 8.37544382e-01
6.81578517e-01 4.24364239e-01 -3.48074585e-01 3.00450742e-01
1.07920623e+00 3.03005099e-01 -6.56788826e-01 -1.77980602e-01
-1.63566685e+00 6.27107918e-01 6.81701481e-01 2.09048819e-02
-2.13388398e-01 -1.68141097e-01 4.67753783e-02 -7.37924799e-02
2.39618003e-01 1.77770510e-01 -4.74698581e-02 1.91004798e-01
-7.63707638e-01 -4.79721278e-02 4.51224267e-01 1.10287631e+00
9.43429112e-01 3.67897391e-01 -9.07057449e-02 1.50278494e-01
7.67819643e-01 9.71208096e-01 2.57859468e-01 -9.85092044e-01
6.22363389e-01 9.32433784e-01 3.10245901e-01 -8.06108594e-01
-4.85107332e-01 -6.50051773e-01 -9.06840742e-01 3.58928472e-01
1.97444871e-01 -1.49815440e-01 -9.28086877e-01 1.16872895e+00
5.09976983e-01 1.30479500e-01 2.59396732e-01 1.71912253e+00
7.94025958e-01 8.66052926e-01 -3.64216864e-01 -2.45731860e-01
1.30522358e+00 -7.13659823e-01 -4.12443936e-01 -9.07695055e-01
2.11984187e-01 -3.96975607e-01 7.54007220e-01 4.91491050e-01
-6.54414475e-01 -6.39128506e-01 -1.29811263e+00 6.02810048e-02
2.35557690e-01 4.96884495e-01 8.52060020e-01 3.73487502e-01
-7.66653061e-01 5.50790280e-02 -8.39490235e-01 -1.79311410e-01
2.50296146e-01 2.89789140e-01 -2.07876414e-01 -4.27409500e-01
-7.86190033e-01 4.53857005e-01 5.15690804e-01 5.58364511e-01
-1.31693304e+00 -4.91194189e-01 -7.90203273e-01 -4.65153642e-02
2.93219209e-01 -9.29847717e-01 9.74907279e-01 -6.98730290e-01
-1.03968334e+00 7.46148825e-01 -6.46460727e-02 -2.69608378e-01
3.46730590e-01 -7.33664376e-04 -3.50702226e-01 3.21128517e-01
2.22788855e-01 7.52454996e-01 8.72174203e-01 -1.36055589e+00
-1.04910982e+00 -9.24502432e-01 -7.36579224e-02 7.05770195e-01
-7.46268183e-02 -5.12569010e-01 -8.80169630e-01 -3.09031963e-01
7.99517334e-01 -6.42006457e-01 -1.62678152e-01 3.23187709e-01
-1.54715568e-01 7.38630593e-02 9.21907306e-01 -6.34927869e-01
6.41689777e-01 -2.30845523e+00 4.24658924e-01 8.65910426e-02
2.75821447e-01 -1.63724899e-01 5.61136082e-02 -1.29718438e-01
3.42384011e-01 -5.27986526e-01 -4.44212765e-01 -4.11665112e-01
-9.23025757e-02 2.33768910e-01 -3.38119179e-01 8.05978477e-01
-5.56336008e-02 6.06108427e-01 -6.59865618e-01 -1.75740689e-01
4.00028169e-01 1.76357791e-01 -1.87248185e-01 1.86376020e-01
-3.88921797e-01 6.97906196e-01 -7.36232758e-01 1.09534514e+00
9.95669484e-01 -1.24353759e-01 -1.68579426e-02 -3.80403638e-01
-5.52082062e-01 -1.41229942e-01 -9.29295421e-01 2.07334685e+00
-2.14575678e-01 6.45400524e-01 5.23977816e-01 -6.86644077e-01
1.37361014e+00 -3.69668722e-01 2.75838792e-01 -8.22590947e-01
3.58758986e-01 1.41476274e-01 -2.49034926e-01 -3.08885902e-01
5.72784722e-01 -4.83126134e-01 -2.50820249e-01 -7.95235038e-02
-3.25359285e-01 -3.64497751e-01 -1.93682015e-01 1.50996163e-01
7.75292158e-01 -2.01178230e-02 -1.85042262e-01 -3.12069207e-01
6.18546605e-01 3.80689949e-01 1.01464379e+00 1.54401660e-01
-2.13979125e-01 3.62513870e-01 5.54142185e-02 -2.74463475e-01
-7.58459687e-01 -1.02844095e+00 9.55029130e-02 1.66260928e-01
1.11482525e+00 2.24232078e-01 -4.41090643e-01 -3.64121586e-01
1.73768118e-01 5.24340272e-01 -2.07380846e-01 -2.47818634e-01
-1.33243352e-01 -6.21500194e-01 2.19668910e-01 5.26148140e-01
9.28610742e-01 -4.85533774e-01 -7.48726189e-01 -2.29093414e-02
-3.67151976e-01 -1.19216037e+00 -2.85835028e-01 7.73610845e-02
-1.20859003e+00 -1.02813280e+00 -5.37980974e-01 -7.86176205e-01
7.54021823e-01 1.04185736e+00 6.97590053e-01 -2.40998361e-02
-3.16819727e-01 4.12600607e-01 -9.42891911e-02 -1.16571732e-01
4.22849059e-01 -5.18567383e-01 3.29445571e-01 6.87695816e-02
5.09603381e-01 -2.73430169e-01 -7.69243121e-01 4.04220849e-01
-8.51039410e-01 1.40412554e-01 9.72763419e-01 6.92765355e-01
7.48050511e-01 2.95345157e-01 -1.81666110e-02 1.43263657e-02
9.52324714e-04 -3.16989660e-01 -1.25108123e+00 -1.92717329e-01
-3.16548675e-01 -1.66717649e-01 2.92274147e-01 -6.45752996e-02
-1.13324177e+00 1.51469111e-01 2.61956275e-01 -9.33336139e-01
9.55702551e-03 4.10716921e-01 -6.40881121e-01 -7.94120505e-02
1.92598030e-01 5.34820318e-01 1.81072786e-01 -4.41939265e-01
-1.24641977e-01 5.98857701e-01 7.36793399e-01 -2.96860278e-01
1.08778775e+00 1.12016988e+00 1.17040291e-01 -8.53691578e-01
-6.99720085e-01 -7.00646996e-01 -4.39834416e-01 -3.88979375e-01
9.48379159e-01 -1.45962763e+00 -8.87566268e-01 4.13284540e-01
-1.00206101e+00 -2.37711996e-01 4.00382951e-02 8.74322891e-01
-2.13463172e-01 7.09853768e-01 -3.83570075e-01 -8.04665685e-01
-2.30605304e-01 -1.06260467e+00 1.35253930e+00 4.87524837e-01
6.24412179e-01 -7.81260908e-01 -4.38047856e-01 5.20504177e-01
-3.82918976e-02 1.22930981e-01 5.79528213e-01 5.29269390e-02
-1.44725943e+00 1.43146198e-02 -6.40371799e-01 1.02759227e-01
8.29356313e-02 -4.81644869e-01 -9.07727540e-01 -5.78822136e-01
4.60670382e-01 8.69346708e-02 1.07670915e+00 5.39636493e-01
8.32399189e-01 1.73340157e-01 -3.90931368e-01 9.74950314e-01
1.52402425e+00 2.67284036e-01 5.85550845e-01 5.85370302e-01
7.20350862e-01 6.15587234e-01 1.40129673e+00 8.33534062e-01
2.66705573e-01 3.26639205e-01 1.15896833e+00 -1.05612189e-01
3.46267670e-02 1.20683769e-02 5.13979256e-01 4.59567875e-01
2.44739756e-01 -8.13949257e-02 -7.90372729e-01 4.09028202e-01
-1.45791209e+00 -7.07245708e-01 -2.96877176e-01 2.26886535e+00
5.00817932e-02 2.32750118e-01 -3.11460823e-01 -2.70928472e-01
7.41550922e-01 2.87430078e-01 -9.64208484e-01 5.39162874e-01
-2.42539272e-01 -6.37533069e-01 6.31066799e-01 4.79640335e-01
-1.05581987e+00 8.24947715e-01 4.73064232e+00 4.40437585e-01
-1.06714034e+00 -2.79845864e-01 2.32183665e-01 -1.06070086e-01
-5.39135993e-01 1.23334453e-01 -1.16086829e+00 4.39371914e-01
1.73928872e-01 -5.15787080e-02 5.10606766e-01 8.90328825e-01
2.98896104e-01 -3.97760659e-01 -8.65951240e-01 1.42312276e+00
2.25536644e-01 -1.12091208e+00 1.59253344e-01 3.14100653e-01
5.73566914e-01 1.69028491e-01 7.94858709e-02 1.04170233e-01
-5.08371182e-02 -6.45448267e-01 7.98084557e-01 6.80604219e-01
9.83328760e-01 -7.44423568e-01 7.75081277e-01 6.76707745e-01
-1.50164580e+00 -4.80054080e-01 -5.92924178e-01 -1.15373574e-01
2.68217444e-01 1.00779486e+00 -7.20205128e-01 8.47572446e-01
7.94171989e-01 8.74021173e-01 -4.10289049e-01 1.17465818e+00
-1.68283463e-01 1.95340463e-03 -4.22722042e-01 1.47867590e-01
5.15900195e-01 -6.29345059e-01 8.67790580e-01 7.61004150e-01
7.12948918e-01 2.26940408e-01 4.51538354e-01 1.02296543e+00
1.36525407e-01 -6.14465237e-01 -5.26917100e-01 -6.96967030e-03
4.57590103e-01 1.46425140e+00 -6.01785183e-01 -3.25347543e-01
-4.85848248e-01 9.91935074e-01 -2.61852562e-01 3.74641955e-01
-8.37437570e-01 -2.47034162e-01 9.44020689e-01 -6.93608671e-02
4.50263619e-01 -6.29575312e-01 -3.55686337e-01 -1.21912062e+00
1.09371834e-01 -3.45760524e-01 1.68217584e-01 -1.06501365e+00
-1.12643921e+00 3.88303578e-01 -3.44589829e-01 -1.83210611e+00
3.01992714e-01 -8.04210126e-01 -5.71658731e-01 1.01228738e+00
-1.74030662e+00 -1.02064764e+00 -1.16104162e+00 5.16746163e-01
7.68901825e-01 -1.82913676e-01 1.21628933e-01 6.22428954e-02
-5.13000965e-01 -2.39622802e-01 1.56227365e-01 -1.07028484e-01
4.77033705e-01 -7.98603594e-01 2.98896521e-01 1.13209248e+00
-4.19008493e-01 3.47684681e-01 4.51229900e-01 -9.97808576e-01
-2.06013417e+00 -1.36271179e+00 2.43985683e-01 -1.54548526e-01
5.30741572e-01 -7.15042427e-02 -7.32703507e-01 6.54890597e-01
-3.21660459e-01 -1.55962273e-01 1.29404843e-01 -3.68062854e-01
3.12995128e-02 -3.51057708e-01 -1.03768897e+00 3.64842504e-01
1.08038306e+00 -2.99726576e-01 -7.29450881e-01 2.20106572e-01
8.62112224e-01 -5.72283387e-01 -5.59061944e-01 6.16250336e-01
1.32287174e-01 -1.09876776e+00 9.92106199e-01 1.76111534e-01
5.76598197e-02 -1.03043640e+00 -3.96971464e-01 -9.68575418e-01
-4.15851325e-01 -1.37111291e-01 5.68098240e-02 9.57562029e-01
-1.18293352e-01 -5.03029764e-01 8.44007134e-01 3.33091944e-01
-8.99766386e-01 -2.39178792e-01 -1.04939795e+00 -8.94231498e-01
-5.14045060e-01 -2.73825705e-01 4.32533205e-01 7.49713719e-01
-5.28891683e-01 3.23644698e-01 -4.45057685e-03 9.62071002e-01
1.05033767e+00 6.73494995e-01 7.44124591e-01 -1.55169082e+00
-1.19826518e-01 -1.74379900e-01 -3.94400746e-01 -1.78093314e+00
5.58916964e-02 -8.32219243e-01 3.10117483e-01 -1.79863989e+00
5.05231619e-02 -3.87183636e-01 7.72903040e-02 -2.55927388e-02
2.38124561e-02 -6.08782955e-02 6.85927924e-03 4.96974260e-01
-6.11944020e-01 1.00217569e+00 1.36493027e+00 -4.16497558e-01
-1.08306728e-01 -8.02928507e-02 -4.41450119e-01 6.79512560e-01
5.45753121e-01 -2.16827206e-02 -3.97108436e-01 -7.59193897e-01
-2.73082674e-01 4.24808949e-01 6.54022336e-01 -1.34882712e+00
4.43888843e-01 -2.02024773e-01 7.94992983e-01 -1.10881877e+00
8.13182950e-01 -9.71804917e-01 -6.45078486e-03 6.50906980e-01
5.54378390e-01 -2.38381073e-01 1.82590827e-01 1.00465214e+00
-4.07034814e-01 -5.77201843e-02 6.36286616e-01 -1.36885926e-01
-1.43650472e+00 6.17559195e-01 -1.73119947e-01 -2.38506317e-01
1.14958668e+00 -5.02182484e-01 -3.26315939e-01 3.68139669e-02
-3.08589578e-01 6.14033461e-01 7.21071064e-01 2.53085375e-01
1.08738434e+00 -1.33138180e+00 -4.55498278e-01 6.70809746e-01
2.65684277e-01 5.79456329e-01 4.98542368e-01 9.13862228e-01
-6.81689203e-01 6.24327421e-01 -1.53218582e-01 -1.08691156e+00
-9.56088841e-01 4.80236769e-01 2.71472305e-01 2.87276089e-01
-8.28289390e-01 8.54328871e-01 7.58662701e-01 -1.90975964e-01
2.13670000e-01 -6.23437762e-01 5.06344624e-02 -6.10040501e-02
7.24190652e-01 1.49652004e-01 -1.23767830e-01 -4.62063849e-01
-4.02348459e-01 7.44271934e-01 1.81966916e-01 -9.61024985e-02
1.01632166e+00 -5.69729924e-01 -1.52363311e-02 2.18205929e-01
7.49193609e-01 -3.04319918e-01 -2.01909971e+00 -2.36453459e-01
-3.57983202e-01 -8.38528931e-01 4.72550333e-01 -4.94618773e-01
-1.14210761e+00 1.08669496e+00 5.59426248e-01 -3.15411717e-01
1.25656962e+00 5.37610194e-03 7.16417432e-01 4.57938194e-01
6.96494222e-01 -7.71088660e-01 2.43943393e-01 6.12268925e-01
6.17458701e-01 -1.34437537e+00 1.10273041e-01 -3.90922636e-01
-7.96504796e-01 1.05583405e+00 8.44888806e-01 -1.26606664e-02
3.91202309e-04 1.03266127e-01 -1.10762581e-01 -3.22811216e-01
-3.78204584e-01 -3.21422607e-01 1.99139848e-01 5.57138741e-01
-3.60391170e-01 -5.68310767e-02 1.65204614e-01 6.44623458e-01
2.26235121e-01 -3.28027278e-01 2.53667116e-01 7.43339837e-01
-1.13343644e+00 -3.70849520e-01 -8.90381038e-01 1.93968743e-01
4.41427946e-01 3.50388885e-03 -3.69880915e-01 7.46699214e-01
7.42596900e-03 8.60243142e-01 3.62197101e-01 -5.31310141e-01
1.96089238e-01 -4.63171363e-01 3.36675376e-01 -4.91261631e-01
-4.46785763e-02 3.40754211e-01 -4.36240844e-02 -5.49995244e-01
-1.88556179e-01 -5.85307062e-01 -1.68073237e+00 -1.64026484e-01
-2.04606101e-01 6.31984994e-02 7.17639625e-01 8.56454372e-01
1.85516685e-01 5.79450369e-01 7.65367866e-01 -1.14957917e+00
-3.16928029e-01 -7.78105438e-01 -9.68577445e-01 -3.77019234e-02
4.12159652e-01 -9.07529533e-01 -7.45938003e-01 -1.32987514e-01] | [7.990692138671875, -2.4145913124084473] |
63cd7359-e145-4848-b647-004d3ac90711 | composing-structure-aware-batches-for-1 | null | null | https://aclanthology.org/2022.findings-acl.239 | https://aclanthology.org/2022.findings-acl.239.pdf | Composing Structure-Aware Batches for Pairwise Sentence Classification | Identifying the relation between two sentences requires datasets with pairwise annotations. In many cases, these datasets contain instances that are annotated multiple times as part of different pairs. They constitute a structure that contains additional helpful information about the inter-relatedness of the text instances based on the annotations. This paper investigates how this kind of structural dataset information can be exploited during training.We propose three batch composition strategies to incorporate such information and measure their performance over 14 heterogeneous pairwise sentence classification tasks. Our results show statistically significant improvements (up to 3.9%) - independent of the pre-trained language model - for most tasks compared to baselines that follow a standard training procedure. Further, we see that even this baseline procedure can profit from having such structural information in a low-resource setting. | ['Iryna Gurevych', 'Tilman Beck', 'Andreas Waldis'] | null | null | null | null | findings-acl-2022-5 | ['sentence-classification'] | ['natural-language-processing'] | [ 3.26423734e-01 3.49932700e-01 -6.83325380e-02 -8.94271076e-01
-1.18205142e+00 -9.40177381e-01 7.27786839e-01 8.85462821e-01
-6.54971361e-01 8.28508019e-01 7.72288978e-01 -1.75825655e-01
5.99609576e-02 -3.02681893e-01 -7.14008272e-01 -6.83912396e-01
-1.32011101e-01 5.99632382e-01 1.93536937e-01 -3.87894034e-01
2.25324526e-01 -4.95360680e-02 -1.21751177e+00 8.39460015e-01
5.04355848e-01 7.70747900e-01 1.23274170e-01 6.13947749e-01
-1.75618544e-01 8.53316784e-01 -7.83572316e-01 -7.50268817e-01
2.55747229e-01 -1.54938698e-01 -1.25335550e+00 -9.35096219e-02
7.67791271e-01 7.93800950e-02 -8.25423077e-02 6.87022328e-01
5.04706979e-01 3.89842726e-02 5.47664225e-01 -9.35678124e-01
-1.19218521e-01 1.24997425e+00 -4.64796066e-01 3.94737601e-01
5.65988064e-01 -2.28005126e-01 1.63000596e+00 -8.70789051e-01
8.63694489e-01 9.64780629e-01 9.30124700e-01 1.65468797e-01
-1.20575285e+00 -3.19273144e-01 -1.56835802e-02 1.17553391e-01
-1.21212089e+00 -6.81858361e-01 7.00680614e-01 -3.78030956e-01
1.42303407e+00 5.20808518e-01 7.64408186e-02 1.19021595e+00
-1.35188296e-01 7.19118774e-01 8.21267366e-01 -6.92045450e-01
-2.14926124e-01 2.10552335e-01 6.62967682e-01 3.71694565e-01
1.53046638e-01 -6.12835288e-01 -5.38474321e-01 -1.93283260e-01
-2.00823858e-01 -4.61031854e-01 -3.20199370e-01 -2.54509896e-01
-1.25995624e+00 5.04765153e-01 2.98655421e-01 7.89311230e-01
2.54185051e-02 -2.31944963e-01 9.66791630e-01 4.88439441e-01
8.72440338e-01 8.56451213e-01 -8.93620968e-01 -2.62412101e-01
-7.98621893e-01 4.10548858e-02 1.11543941e+00 1.06004977e+00
6.80195153e-01 -6.88912749e-01 -4.17087078e-01 1.01296306e+00
-1.32143572e-01 -6.29767701e-02 5.87336063e-01 -7.10610271e-01
1.26896679e+00 6.44720614e-01 -6.17606863e-02 -8.89249742e-01
-6.23449087e-01 -3.61889601e-01 -7.50757515e-01 -6.90116644e-01
6.64649785e-01 -3.11637640e-01 -1.94765806e-01 1.80489516e+00
3.72545600e-01 -6.63396642e-02 1.83271199e-01 4.92768317e-01
9.71194386e-01 2.92408228e-01 -1.33178830e-01 -1.43118978e-01
1.33017838e+00 -1.08629370e+00 -5.68717599e-01 -2.53851265e-01
1.40204167e+00 -9.03761029e-01 9.60706353e-01 -9.75028127e-02
-9.78203356e-01 -4.36052382e-01 -8.92778933e-01 -3.28776240e-01
-2.84749925e-01 1.39077261e-01 3.79413396e-01 5.26811481e-01
-1.10958135e+00 6.54105723e-01 -4.11405921e-01 -4.12730008e-01
1.83218330e-01 3.28010350e-01 -7.35905111e-01 -7.03849569e-02
-1.24561310e+00 1.07570028e+00 4.26784456e-01 -7.11794049e-02
-1.01733573e-01 -9.06059623e-01 -9.22818899e-01 -1.59206763e-02
3.54923427e-01 -5.37572265e-01 1.21802044e+00 -9.25561428e-01
-8.85272622e-01 1.27007771e+00 -5.26000023e-01 -5.32614648e-01
5.36659300e-01 -3.59899580e-01 2.99549811e-02 3.85115109e-02
1.43490151e-01 5.01596093e-01 3.02363634e-01 -1.10709226e+00
-3.17916512e-01 -3.05101186e-01 3.45086932e-01 3.63807142e-01
-5.42271018e-01 4.78016853e-01 -3.75258207e-01 -7.59898245e-01
-2.55141973e-01 -9.49103773e-01 -8.30456614e-02 -4.91101474e-01
-7.42956936e-01 -4.20944512e-01 5.87595522e-01 -7.31560647e-01
1.11785758e+00 -2.00739431e+00 2.55367935e-01 5.03668450e-02
1.64733812e-01 2.81141788e-01 -3.70079428e-01 8.82529557e-01
-2.34478816e-01 3.91945541e-01 -2.95758635e-01 -9.02373075e-01
3.04086618e-02 1.58787683e-01 -1.31126970e-01 2.67343789e-01
2.55894333e-01 8.48690331e-01 -7.77500391e-01 -5.09571075e-01
-2.18990445e-01 2.65935242e-01 -3.81342769e-01 3.90913069e-01
5.89325503e-02 6.08918786e-01 -1.44040331e-01 1.76187485e-01
3.80199492e-01 -2.24667281e-01 6.22532725e-01 -3.03795069e-01
2.54992932e-01 1.01967525e+00 -7.85090268e-01 1.66391325e+00
-4.63104844e-01 7.16391921e-01 -4.86847721e-02 -1.22613966e+00
7.33536184e-01 2.64465779e-01 4.19305712e-01 -4.33216721e-01
9.00948718e-02 1.09155953e-01 3.41554642e-01 -6.27549767e-01
6.99724436e-01 -5.34922257e-02 -1.88387230e-01 5.58601618e-01
8.49619359e-02 6.49062768e-02 4.94470119e-01 4.54797506e-01
1.42755222e+00 -1.98029861e-01 4.20793802e-01 -4.31575984e-01
7.81875908e-01 -3.25386673e-01 4.66747850e-01 8.54181528e-01
4.17603739e-02 6.48446679e-01 7.90503979e-01 -2.70225376e-01
-1.22165215e+00 -4.16722417e-01 -3.46312195e-01 1.04431653e+00
-2.11426571e-01 -9.36635852e-01 -6.19777083e-01 -1.00563526e+00
-1.46786258e-01 5.26084900e-01 -6.81740522e-01 -1.35461479e-01
-7.90223241e-01 -7.61578679e-01 7.71579623e-01 7.12476313e-01
2.46852770e-01 -7.16332436e-01 -1.49804473e-01 -5.71514182e-02
-5.57045698e-01 -1.62523448e+00 -4.00974840e-01 3.83347839e-01
-7.56836534e-01 -1.03701723e+00 -2.79947340e-01 -7.81761825e-01
6.67070031e-01 1.29758820e-01 1.68055594e+00 3.80943894e-01
-2.14776509e-02 2.65777856e-01 -6.79599583e-01 -1.29844710e-01
-4.75491524e-01 4.45984632e-01 -1.13799572e-01 -1.72534898e-01
4.27496344e-01 -5.52049756e-01 -5.73709868e-02 1.47872165e-01
-4.57910508e-01 -9.42620561e-02 3.91043901e-01 9.45160389e-01
1.60206810e-01 -6.92029670e-02 3.98509026e-01 -1.32980394e+00
4.76830363e-01 -5.07949650e-01 -7.69308433e-02 5.60751021e-01
-7.37213641e-02 2.41090611e-01 6.09953344e-01 -1.97839901e-01
-9.51143205e-01 -1.05701536e-01 -2.31221542e-02 1.39775932e-01
-2.41744384e-01 6.18186712e-01 -3.21225047e-01 2.28508517e-01
3.69592845e-01 -3.22185636e-01 -1.74003333e-01 -5.71540475e-01
2.17948660e-01 7.44201541e-01 9.69816744e-02 -8.51277411e-01
6.13457620e-01 5.77207059e-02 -3.23177967e-03 -6.32029772e-01
-1.29570603e+00 -7.54895091e-01 -1.17731869e+00 2.40048066e-01
7.11760342e-01 -9.29917395e-01 -4.24577534e-01 2.09703699e-01
-1.36577117e+00 -3.81418467e-01 -9.35627967e-02 4.00296211e-01
-1.64812431e-01 5.61819315e-01 -8.31926882e-01 -6.07061684e-01
-1.35683596e-01 -1.04302168e+00 1.12888515e+00 -3.78088742e-01
-6.42728925e-01 -1.24409044e+00 1.79076679e-02 7.65617013e-01
1.57927319e-01 -1.89749300e-01 8.60044122e-01 -1.39117861e+00
-7.27778748e-02 -3.26975286e-01 -8.28059018e-02 3.24423045e-01
1.77219778e-01 -5.74877784e-02 -1.09713840e+00 -2.97651857e-01
-3.55862230e-02 -4.35312778e-01 8.46355498e-01 -7.13199154e-02
1.17654634e+00 -4.60380197e-01 -2.46830821e-01 1.59652621e-01
1.04660094e+00 -3.60708624e-01 3.02569658e-01 2.58121818e-01
9.31126177e-01 1.14254677e+00 5.60978472e-01 2.77225763e-01
6.25423789e-01 1.06713045e+00 1.72680169e-01 -2.65542362e-02
-4.75162826e-02 1.52797159e-02 4.02190059e-01 1.23003948e+00
-4.30142228e-03 -3.97909969e-01 -9.67889786e-01 5.07109702e-01
-1.81519854e+00 -1.02738678e+00 -4.53377128e-01 2.07213402e+00
1.28106034e+00 1.43860295e-01 1.01628743e-01 2.07607239e-01
6.25475764e-01 2.27678999e-01 -6.71257302e-02 -4.96701837e-01
-4.58254129e-01 1.64957315e-01 4.51777488e-01 6.23904109e-01
-1.28829861e+00 6.84793591e-01 6.67703438e+00 5.50707519e-01
-5.45037806e-01 3.26275192e-02 7.22810686e-01 -1.03144810e-01
-3.30243230e-01 1.33660622e-02 -9.79139268e-01 4.56830084e-01
1.15434635e+00 6.54336289e-02 -1.86099976e-01 3.74035031e-01
-4.22626697e-02 -1.46831900e-01 -1.48991323e+00 5.46076834e-01
3.06978464e-01 -1.06536782e+00 -1.68350339e-01 -1.18719898e-01
6.13261342e-01 1.93673506e-01 -3.27510953e-01 2.79888779e-01
3.08263004e-01 -7.88373172e-01 5.71815133e-01 2.20736817e-01
2.76420534e-01 -5.23618400e-01 1.36561561e+00 5.48890889e-01
-1.01506186e+00 5.17858304e-02 -3.35784733e-01 -2.30439052e-01
2.17072368e-01 1.00812161e+00 -9.20124292e-01 9.21126723e-01
6.58194184e-01 9.75224674e-01 -1.11133122e+00 8.20883274e-01
-2.92721599e-01 7.16848612e-01 -2.62859911e-01 -6.65691122e-02
8.75039101e-02 -1.27492165e-02 5.25300860e-01 1.50541461e+00
-5.55123668e-03 -1.44513845e-01 1.11174554e-01 2.23099738e-01
-4.79560107e-01 3.53606075e-01 -7.28681862e-01 7.62295276e-02
4.00537729e-01 1.35394824e+00 -5.40974736e-01 -4.21907395e-01
-6.79135263e-01 8.90725315e-01 8.87899876e-01 -1.44034550e-01
-6.15814328e-01 -2.17846587e-01 4.78611410e-01 -1.57034636e-01
3.17609280e-01 -2.09941298e-01 -2.52856702e-01 -1.30430019e+00
5.11707127e-01 -8.77238810e-01 4.79761422e-01 -6.35679424e-01
-1.74517584e+00 6.62323177e-01 6.45417571e-02 -9.17328954e-01
-2.86073387e-01 -5.90873837e-01 -5.16181529e-01 7.41617203e-01
-1.24801517e+00 -1.26747215e+00 -3.36583657e-03 1.72542423e-01
3.59419078e-01 -3.38192768e-02 9.62037146e-01 3.79681915e-01
-7.47722328e-01 9.81283069e-01 2.47258097e-02 5.36666155e-01
1.12460363e+00 -1.45248342e+00 3.02908480e-01 8.58484745e-01
6.14191949e-01 8.64873469e-01 6.86221004e-01 -4.35429484e-01
-1.00132978e+00 -7.63383210e-01 1.70988631e+00 -1.03581774e+00
9.08328533e-01 -7.57577837e-01 -1.04901016e+00 8.76860499e-01
7.14345455e-01 -1.51432380e-02 1.21773052e+00 9.21973646e-01
-8.26073408e-01 -2.75317654e-02 -8.36181521e-01 3.45807135e-01
1.26367974e+00 -7.87338197e-01 -8.91724825e-01 5.87659359e-01
8.55636120e-01 -4.80463445e-01 -1.08987463e+00 5.36886215e-01
2.50479043e-01 -9.57587540e-01 8.61984670e-01 -1.08511949e+00
7.86855638e-01 1.41261742e-01 -3.18217218e-01 -1.19055772e+00
6.17170036e-02 -4.11891639e-01 2.23010674e-01 1.80197716e+00
8.67845476e-01 -4.90442097e-01 4.27033097e-01 8.77598345e-01
-2.89717704e-01 -5.94091713e-01 -8.80546808e-01 -7.73427188e-01
3.06661159e-01 -3.00164759e-01 4.47621584e-01 1.35375571e+00
5.28768718e-01 8.07685316e-01 -8.23516548e-02 -1.78698972e-01
1.84953585e-01 -2.07817480e-02 7.55332708e-01 -1.10147285e+00
-3.62854004e-01 -4.23008621e-01 -5.64761937e-01 -7.45087743e-01
9.04247463e-01 -1.26514912e+00 1.96924582e-02 -1.27563119e+00
5.36681414e-01 -5.49044788e-01 -2.00266510e-01 7.08044827e-01
-6.39786482e-01 1.38409659e-01 1.13924749e-01 2.47684434e-01
-8.56977880e-01 2.44408786e-01 6.64681911e-01 -1.20606713e-01
1.66787028e-01 -2.05471009e-01 -6.92514539e-01 5.77400386e-01
9.83296037e-01 -4.92303461e-01 -2.92065740e-01 -7.39669442e-01
4.48806226e-01 -2.86972404e-01 1.13632105e-01 -7.11008132e-01
2.27391109e-01 3.48425388e-01 4.10807040e-03 -3.98736089e-01
2.19416171e-01 -8.50300074e-01 -2.12987959e-01 6.79662824e-02
-9.75917399e-01 2.33259901e-01 1.93577766e-01 4.53340113e-01
-4.76980001e-01 -5.56946516e-01 2.92149246e-01 -7.04983473e-02
-3.39062124e-01 -2.69874603e-01 -1.83645800e-01 4.84717876e-01
8.81020606e-01 3.22437853e-01 -5.78056395e-01 -3.14209491e-01
-5.49833536e-01 1.39477089e-01 4.00608122e-01 4.53346610e-01
-9.35913175e-02 -1.05996883e+00 -8.43426108e-01 -1.70415416e-01
2.99859941e-01 -9.11351964e-02 7.40323216e-02 1.23485720e+00
3.65479179e-02 4.34439689e-01 1.34584829e-01 -4.45411831e-01
-1.75571132e+00 3.73551786e-01 1.10916547e-01 -7.51465023e-01
-5.49526751e-01 1.04492736e+00 -1.39470458e-01 -7.75039911e-01
9.42482874e-02 -1.34377167e-01 -2.81503588e-01 4.27691251e-01
3.73349786e-01 -8.45206156e-03 4.33516532e-01 -9.20246065e-01
-5.44406056e-01 3.77216578e-01 -2.80308068e-01 -4.46759053e-02
1.38914442e+00 -2.33784035e-01 -6.28550947e-01 7.93089628e-01
1.55677617e+00 4.45640206e-01 -6.29893243e-01 -4.66065079e-01
5.15080035e-01 -2.96820164e-01 -4.71935064e-01 -6.94705427e-01
-8.61140490e-01 9.17482257e-01 -3.47268820e-01 2.66985714e-01
7.58742571e-01 2.32887149e-01 5.32978356e-01 5.78002810e-01
1.89178273e-01 -8.70778739e-01 -6.77999705e-02 7.52517879e-01
8.09831440e-01 -1.55399573e+00 1.52340248e-01 -8.11098933e-01
-7.76959896e-01 9.73043323e-01 5.26049972e-01 1.68677002e-01
5.50716937e-01 4.42588747e-01 -3.21832895e-02 -2.63424907e-02
-1.01545525e+00 -2.23006785e-01 5.32458842e-01 4.28890020e-01
1.03322637e+00 -1.72823429e-01 -5.19503832e-01 6.03621721e-01
-3.28722805e-01 -7.29483306e-01 3.04986447e-01 8.33596706e-01
-3.60975191e-02 -1.61948729e+00 2.23020941e-01 4.63251919e-01
-6.36314869e-01 -3.93979877e-01 -8.49469781e-01 5.86374342e-01
-1.82077870e-01 1.10952032e+00 1.06547363e-01 -3.76818985e-01
2.20422715e-01 3.36758643e-01 4.87745434e-01 -8.25210690e-01
-1.12570953e+00 -3.76908541e-01 8.97038043e-01 -4.14931655e-01
-8.65381539e-01 -8.62076879e-01 -9.87600982e-01 -7.78086483e-02
-3.88422430e-01 2.84413189e-01 2.51999199e-01 1.25423658e+00
3.71566713e-01 4.37094271e-01 5.62862396e-01 -5.88841200e-01
-4.62915629e-01 -1.27815270e+00 -1.71702251e-01 8.39855731e-01
7.93320164e-02 -3.86367023e-01 -4.79738206e-01 2.62268409e-02] | [9.594117164611816, 8.897502899169922] |
ec22d3bc-bbd4-41d1-b2d0-2ac6707be6db | incorporating-knowledge-into-document | 2301.11719 | null | https://arxiv.org/abs/2301.11719v4 | https://arxiv.org/pdf/2301.11719v4.pdf | The Exploration of Knowledge-Preserving Prompts for Document Summarisation | Despite the great development of document summarisation techniques nowadays, factual inconsistencies between the generated summaries and the original texts still occur from time to time. This study explores the possibility of adopting prompts to incorporate factual knowledge into generated summaries. We specifically study prefix-tuning that uses a set of trainable continuous prefix prompts together with discrete natural language prompts to aid summary generation. Experimental results demonstrate that the trainable prefixes can help the summarisation model extract information from discrete prompts precisely, thus generating knowledge-preserving summaries that are factually consistent with the discrete prompts. The ROUGE improvements of the generated summaries indicate that explicitly adding factual knowledge into the summarisation process could boost the overall performance, showing great potential for applying it to other natural language processing tasks. | ['Makhmoor Fiza', 'Alireza Seyed Shakeri', 'Wei Emma Zhang', 'Chen Chen'] | 2023-01-27 | null | null | null | null | ['document-summarization'] | ['natural-language-processing'] | [ 5.70094287e-01 6.85225546e-01 -3.25641662e-01 -3.97499233e-01
-1.34430516e+00 -8.81156027e-01 1.09913027e+00 7.22738087e-01
-2.64124751e-01 1.15048301e+00 1.15449786e+00 -2.90807337e-01
-2.15870127e-01 -6.76638663e-01 -6.03626907e-01 -3.81835997e-02
6.25237748e-02 6.17965877e-01 1.02063827e-01 -3.76269192e-01
7.94020593e-01 5.82462586e-02 -1.39436817e+00 8.86121631e-01
1.34741998e+00 3.32112849e-01 2.25383058e-01 1.06119657e+00
-5.28322041e-01 9.05226409e-01 -1.40953946e+00 -4.17583048e-01
-7.90229291e-02 -6.06918454e-01 -1.07792282e+00 1.40919704e-02
7.83187211e-01 -4.48407799e-01 -3.04515343e-02 7.13779688e-01
4.10957783e-01 5.16769648e-01 7.72015870e-01 -7.78767169e-01
-5.88459611e-01 1.21674812e+00 -9.35970098e-02 3.40835273e-01
1.08762217e+00 1.03058942e-01 1.15574682e+00 -4.12557602e-01
7.18911648e-01 1.26886547e+00 5.18000066e-01 4.05743599e-01
-9.43467140e-01 -1.74094811e-01 1.70090586e-01 1.14699490e-01
-7.44262218e-01 -6.58235252e-01 7.98651516e-01 -2.12095857e-01
1.46951318e+00 6.17339611e-01 4.81381029e-01 8.94903481e-01
3.85747015e-01 1.02230752e+00 5.62816978e-01 -6.56199872e-01
6.45140782e-02 4.89399694e-02 4.33762997e-01 4.21779841e-01
5.44219077e-01 -2.56990224e-01 -6.87035322e-01 -4.99624789e-01
2.33172908e-01 -5.21828353e-01 -4.26962823e-01 3.41178060e-01
-1.20290399e+00 8.49632144e-01 9.95518640e-02 4.11900073e-01
-8.07501853e-01 -2.52600443e-02 8.01398993e-01 2.79106975e-01
3.94866228e-01 1.26650631e+00 -2.10589752e-01 -3.75394106e-01
-1.36395919e+00 8.57872069e-01 1.15922248e+00 1.01025581e+00
5.03882766e-01 1.47890419e-01 -9.47754383e-01 6.64490223e-01
-2.20283613e-01 3.65069181e-01 6.64699256e-01 -1.05833888e+00
1.03313684e+00 7.54927993e-01 4.84384030e-01 -1.11204875e+00
-3.06127131e-01 -3.14003915e-01 -4.92805541e-01 -4.94889319e-01
1.26581520e-01 -3.21443319e-01 -6.33959353e-01 1.27842176e+00
3.99760716e-02 -2.48333752e-01 3.93563539e-01 3.60162675e-01
9.90633607e-01 1.06741917e+00 2.38323398e-02 -5.44216096e-01
1.23126566e+00 -9.25746143e-01 -1.05982518e+00 -4.73706186e-01
8.03697824e-01 -8.19414258e-01 1.21190417e+00 3.12349737e-01
-1.29178882e+00 -5.52152634e-01 -1.01767206e+00 -2.36688390e-01
2.74455808e-02 2.24906042e-01 5.73352814e-01 3.11014950e-01
-1.03808224e+00 8.68143678e-01 -5.48997045e-01 -3.19001287e-01
3.43313336e-01 5.91764823e-02 -1.70436233e-01 -1.37095317e-01
-1.27563131e+00 8.38700891e-01 8.79371762e-01 -2.78259724e-01
-2.99140275e-01 -8.74976277e-01 -9.46961820e-01 3.21869135e-01
5.83573222e-01 -1.08870268e+00 1.85694814e+00 -6.60301745e-01
-1.36895108e+00 2.50478208e-01 -4.06295270e-01 -7.74841905e-01
4.62748826e-01 -4.23592150e-01 -1.60603568e-01 4.30218309e-01
3.02669287e-01 7.08748996e-01 5.08547783e-01 -1.18227315e+00
-9.12581921e-01 8.44379365e-02 3.40370871e-02 4.90066767e-01
-2.63128519e-01 -9.75851640e-02 5.28488271e-02 -6.59377396e-01
-2.78505653e-01 -5.21412075e-01 -2.43538722e-01 -7.88028777e-01
-8.74648452e-01 -6.77757978e-01 5.65966249e-01 -9.82574642e-01
1.59626079e+00 -1.45723200e+00 -1.80029914e-01 -1.07311867e-01
-1.39471814e-01 4.40729797e-01 -3.52169961e-01 1.07514167e+00
2.96632141e-01 1.80931315e-01 -3.63170654e-01 -9.42682028e-02
9.21354145e-02 9.18880999e-02 -8.43279421e-01 -3.48575383e-01
2.77436614e-01 1.16241848e+00 -1.29962718e+00 -4.63755310e-01
-3.80065814e-02 4.48102169e-02 -4.38359737e-01 4.47206676e-01
-7.89460480e-01 1.27239734e-01 -4.72135812e-01 -4.68842126e-02
2.19031751e-01 2.40220991e-03 -8.27341750e-02 -2.99744513e-02
-4.01313007e-02 1.12202168e+00 -8.22174668e-01 1.65387523e+00
-3.73698264e-01 7.02144265e-01 -5.86878955e-01 -6.95911109e-01
9.61367488e-01 4.66844261e-01 -3.55815329e-02 -6.26524150e-01
-1.68238416e-01 7.45498240e-02 -8.93444344e-02 -6.80776894e-01
1.38837552e+00 -1.93948805e-01 -3.11485499e-01 7.09206104e-01
-8.98657665e-02 -6.36463404e-01 8.44555914e-01 7.19881594e-01
1.09213090e+00 -2.05241114e-01 6.71758294e-01 -7.16597959e-02
4.60095555e-01 5.85519433e-01 1.26951441e-01 1.17398751e+00
3.02914262e-01 5.45260489e-01 6.23250484e-01 -1.74013361e-01
-1.01836503e+00 -7.72762835e-01 4.50251013e-01 7.75313199e-01
-3.70548517e-01 -8.43364835e-01 -7.10485637e-01 -7.63462186e-01
-1.28375307e-01 1.65935469e+00 -3.83563936e-01 -3.22737157e-01
-7.86318243e-01 -2.70810753e-01 6.69362843e-01 6.59728289e-01
3.72988462e-01 -1.27006960e+00 -8.92014802e-01 5.93739331e-01
-7.16619253e-01 -9.88387883e-01 -6.05209768e-01 -3.97240147e-02
-1.06825328e+00 -9.26032782e-01 -5.66037774e-01 -4.68513548e-01
4.29955602e-01 1.95525020e-01 1.31537783e+00 -4.92694415e-02
3.57612520e-01 5.27828038e-01 -4.49353069e-01 -6.77338481e-01
-1.14249623e+00 5.41790247e-01 -3.01455319e-01 -6.61153972e-01
2.07031637e-01 -3.19152206e-01 -2.59092331e-01 -3.48613858e-01
-1.23620486e+00 3.34520072e-01 4.84848499e-01 6.28302038e-01
1.70260966e-01 1.14120170e-01 1.15450096e+00 -1.17150581e+00
1.53770614e+00 -2.34493598e-01 -2.26519667e-02 3.09007376e-01
-4.09606159e-01 3.75556648e-01 1.03573847e+00 -1.97400615e-01
-1.34544027e+00 -4.18471307e-01 -9.38679948e-02 2.55121171e-01
-2.65725315e-01 8.70239854e-01 1.82581320e-02 8.00053656e-01
9.46756244e-01 5.27441204e-01 -2.44367957e-01 -1.46037832e-01
6.17464960e-01 6.82911456e-01 8.70064914e-01 -7.32052088e-01
5.35003185e-01 -8.94755404e-03 -4.65502888e-01 -9.06772316e-01
-1.11624646e+00 -4.84498709e-01 -3.76411051e-01 2.76378430e-02
2.84978569e-01 -6.29486978e-01 2.33491254e-03 -7.00621903e-02
-1.52109122e+00 -2.69512981e-01 -6.74334466e-01 5.82317188e-02
-7.09789574e-01 6.27836883e-01 -3.69934797e-01 -6.65849984e-01
-8.90643537e-01 -5.51364958e-01 1.09758914e+00 4.30787504e-01
-1.05389428e+00 -1.01490498e+00 1.97560042e-01 3.19267124e-01
2.61829257e-01 2.61304826e-01 1.17663813e+00 -1.22973001e+00
-3.71010184e-01 -4.85512257e-01 6.64978847e-02 3.09725732e-01
4.44571853e-01 2.07380459e-01 -7.01468825e-01 -3.26829702e-02
-6.22855313e-02 -1.52962625e-01 8.38200688e-01 4.42056119e-01
5.96837997e-01 -1.31466162e+00 -3.02823544e-01 -1.13364041e-01
9.21436548e-01 1.44562781e-01 5.77338755e-01 3.50251764e-01
5.70778966e-01 8.28191102e-01 7.04282701e-01 4.59126651e-01
5.34189284e-01 3.08569372e-01 -1.97736546e-01 3.91124845e-01
-2.96032935e-01 -6.23909593e-01 2.26274401e-01 1.00379002e+00
3.17179441e-01 -5.94166934e-01 -6.92814112e-01 9.70619321e-01
-1.80000567e+00 -1.31152344e+00 -3.68935019e-02 2.01171923e+00
1.21254206e+00 3.55719060e-01 1.27047986e-01 2.49985859e-01
5.29808939e-01 5.12958348e-01 -2.22129196e-01 -9.56395626e-01
-1.13979846e-01 -3.36599760e-02 7.24394759e-03 6.53390110e-01
-7.37251520e-01 8.49889159e-01 6.49426556e+00 7.75435746e-01
-9.11666214e-01 -5.69106281e-01 1.73031658e-01 -9.11116153e-02
-7.39239037e-01 7.58511648e-02 -7.53658533e-01 3.53077322e-01
1.22566807e+00 -1.07550561e+00 -7.70697324e-03 5.60963750e-01
6.50492728e-01 -3.41115952e-01 -1.35838711e+00 2.40095973e-01
1.32242858e-01 -1.69028723e+00 7.86337793e-01 -3.78773510e-01
9.67547774e-01 -4.68524724e-01 -4.21730638e-01 3.53544384e-01
4.40790176e-01 -8.41437459e-01 5.78373730e-01 5.61509550e-01
4.05888498e-01 -9.13353026e-01 8.11106086e-01 6.73228741e-01
-6.80061460e-01 8.81651863e-02 -3.98258716e-01 -2.48016477e-01
4.86996204e-01 6.45229936e-01 -1.72834766e+00 9.91847336e-01
-2.26457775e-01 7.55544186e-01 -7.74809897e-01 1.11167920e+00
-5.94417274e-01 7.70925045e-01 -2.24220619e-01 -3.63339037e-01
3.66224110e-01 7.64158592e-02 8.97241116e-01 1.73950958e+00
2.62340516e-01 1.41985476e-01 1.38168827e-01 6.76259041e-01
-1.29843444e-01 1.61850572e-01 -7.60171533e-01 -2.90664464e-01
5.92807770e-01 9.06712115e-01 -5.23674428e-01 -9.01961982e-01
1.21927761e-01 8.29826474e-01 1.87443241e-01 3.05370569e-01
-2.98433274e-01 -6.51907563e-01 9.28183720e-02 2.51987111e-02
2.70746768e-01 -9.01790336e-02 -6.26953542e-01 -9.07262027e-01
2.46091545e-01 -1.09897840e+00 4.90357131e-01 -8.03754747e-01
-1.09211886e+00 5.21874070e-01 3.31179678e-01 -8.75500143e-01
-7.52076745e-01 1.71824142e-01 -1.09315598e+00 6.07717335e-01
-1.44904029e+00 -6.01028621e-01 -7.29299635e-02 -1.91302434e-01
1.03722775e+00 1.37430787e-01 8.15939367e-01 -3.27017963e-01
-1.85992181e-01 4.69415694e-01 -6.64360747e-02 -1.07484832e-01
7.47866035e-01 -1.48477197e+00 7.78583586e-01 1.06526887e+00
3.03338796e-01 8.52338076e-01 1.35278451e+00 -9.14762974e-01
-9.57198203e-01 -1.12629950e+00 1.43023217e+00 -4.97245312e-01
4.94335502e-01 1.51768163e-01 -1.31861734e+00 6.34702146e-01
7.67479360e-01 -1.03669035e+00 6.14426136e-01 -4.38369103e-02
-2.00812817e-01 2.33618647e-01 -8.68543506e-01 6.55429423e-01
7.31077909e-01 -3.33333164e-01 -1.54542148e+00 3.28026354e-01
9.22494173e-01 -6.21473074e-01 -5.22863925e-01 2.81545222e-01
2.12106690e-01 -6.04215622e-01 7.08325207e-01 -7.75027990e-01
9.91270781e-01 -9.19407159e-02 3.15281123e-01 -1.81364214e+00
-8.28351155e-02 -8.12290907e-01 -4.65390354e-01 1.59868217e+00
3.92151207e-01 -4.44013983e-01 4.74015206e-01 6.81477726e-01
-4.39609110e-01 -4.35208082e-01 -6.02837145e-01 -7.85350919e-01
2.02479690e-01 -3.07450145e-02 6.41578972e-01 6.49355114e-01
5.04164457e-01 8.66555154e-01 -6.58084974e-02 -1.68903053e-01
2.68442780e-01 2.17105940e-01 9.34131324e-01 -1.14115179e+00
-1.42761663e-01 -5.91907859e-01 1.96186990e-01 -1.28191853e+00
8.68716687e-02 -9.18402433e-01 3.95002842e-01 -2.34444284e+00
7.42630586e-02 2.16461602e-03 2.93259650e-01 4.48929042e-01
-8.22908640e-01 -5.82242370e-01 2.63021052e-01 2.67277621e-02
-6.72573566e-01 6.11283541e-01 1.27715087e+00 -3.09509665e-01
-6.20145738e-01 7.77920559e-02 -1.23980737e+00 5.06996870e-01
7.37102509e-01 -5.82279027e-01 -6.37801230e-01 -3.77934664e-01
2.66307294e-02 3.22505385e-01 -1.52473422e-02 -9.26465929e-01
5.27430654e-01 -2.03027815e-01 -3.06247908e-04 -8.57141078e-01
-1.42822132e-01 -1.92186713e-01 -1.22401901e-01 2.01041013e-01
-9.84387219e-01 1.51601940e-01 5.86063564e-01 5.15987992e-01
-3.57196212e-01 -4.48415488e-01 1.88703239e-01 -2.12632835e-01
-4.38944906e-01 -3.96724671e-01 -6.72243536e-01 5.29038250e-01
4.94841278e-01 -4.05048817e-01 -6.72509134e-01 -5.17632544e-01
-4.65908721e-02 2.78258353e-01 3.19719672e-01 3.38207603e-01
4.26400930e-01 -9.11324084e-01 -1.08348954e+00 -2.58128911e-01
8.19619894e-02 2.97701836e-01 6.39920235e-02 2.82502115e-01
-2.66139388e-01 9.91560996e-01 -5.60734756e-02 -3.94465625e-01
-1.10750294e+00 2.48685554e-01 -1.49576157e-01 -7.58573472e-01
-9.07428741e-01 4.82344061e-01 -2.26021767e-01 -2.63643920e-01
1.55739322e-01 -6.73009336e-01 -4.62531060e-01 2.70336211e-01
9.80346084e-01 4.84570801e-01 3.23620021e-01 -4.01690491e-02
-3.60685587e-02 -4.72773947e-02 -4.70238268e-01 -3.04412603e-01
1.32780111e+00 -1.59702554e-01 8.96092430e-02 5.60639203e-01
8.15555036e-01 2.84674615e-01 -1.12675583e+00 -2.43957013e-01
7.54703104e-01 -3.41289848e-01 -3.14736933e-01 -1.20670581e+00
-2.03135192e-01 5.37673473e-01 -6.13943040e-01 5.08370101e-01
8.88389528e-01 -1.00436755e-01 1.04140210e+00 8.81067336e-01
3.98743823e-02 -9.44648683e-01 9.92794558e-02 7.62951016e-01
1.18432891e+00 -8.16223025e-01 2.12993503e-01 -2.88961023e-01
-9.09018934e-01 1.24223876e+00 4.05478746e-01 -1.91696659e-01
-4.69603240e-01 -3.19481501e-03 1.48701280e-01 -1.17550507e-01
-1.03565478e+00 2.39231184e-01 4.40464765e-01 7.18554676e-01
5.75739145e-01 -6.64663836e-02 -6.24461710e-01 7.40482807e-01
-8.31370533e-01 9.70359817e-02 1.22408652e+00 9.17073727e-01
-8.21712792e-01 -1.00129986e+00 -3.01766127e-01 8.99890661e-01
-2.21636638e-01 -2.27576137e-01 -8.09565246e-01 5.59535623e-01
-6.73936665e-01 1.20837617e+00 -5.46102896e-02 1.87263787e-01
7.49990880e-01 4.47413445e-01 3.87852043e-01 -1.06668270e+00
-9.97545004e-01 -3.12772274e-01 7.70926535e-01 -9.30580050e-02
-1.60240859e-01 -7.35817611e-01 -1.48123860e+00 -2.33413398e-01
-2.62380600e-01 7.31023550e-01 3.41301113e-01 1.13704729e+00
5.16409993e-01 8.15171838e-01 3.20558101e-01 -6.83625162e-01
-8.06679964e-01 -1.22151768e+00 8.40061456e-02 3.24026108e-01
5.29464841e-01 -4.08641621e-02 -1.59280196e-01 2.54793197e-01] | [12.400884628295898, 9.393078804016113] |
1daf6f89-6da1-41b0-93dc-77dadc0a39d1 | all-in-1-at-ijcnlp-2017-task-4-short-text | null | null | https://aclanthology.org/I17-4024 | https://aclanthology.org/I17-4024.pdf | All-In-1 at IJCNLP-2017 Task 4: Short Text Classification with One Model for All Languages | We present All-In-1, a simple model for multilingual text classification that does not require any parallel data. It is based on a traditional Support Vector Machine classifier exploiting multilingual word embeddings and character n-grams. Our model is simple, easily extendable yet very effective, overall ranking 1st (out of 12 teams) in the IJCNLP 2017 shared task on customer feedback analysis in four languages: English, French, Japanese and Spanish. | ['Barbara Plank'] | 2017-12-01 | all-in-1-at-ijcnlp-2017-task-4-short-text-1 | https://aclanthology.org/I17-4024 | https://aclanthology.org/I17-4024.pdf | ijcnlp-2017-12 | ['multilingual-word-embeddings', 'multilingual-text-classification'] | ['methodology', 'miscellaneous'] | [-6.46601260e-01 -3.33499402e-01 -7.01109052e-01 -4.22266245e-01
-9.76532161e-01 -9.08014357e-01 7.56743252e-01 9.43792522e-01
-1.04385781e+00 8.53858888e-01 5.71303606e-01 -1.05537117e+00
1.65602267e-01 -1.39712244e-01 -3.00354809e-01 -1.36485711e-01
1.84118718e-01 9.39291239e-01 -7.09674880e-02 -8.56588364e-01
2.26755694e-01 8.75124186e-02 -8.14409316e-01 5.46939671e-01
8.18870187e-01 5.79648077e-01 -1.23397291e-01 1.06838977e+00
-2.75093526e-01 6.95748389e-01 -1.59408614e-01 -8.67186010e-01
-7.75237978e-02 1.51819095e-01 -1.07621217e+00 -3.31866592e-01
5.63098669e-01 5.09694964e-02 -1.22067235e-01 6.24049664e-01
6.32060528e-01 -1.44038513e-01 8.96696627e-01 -8.64226222e-01
-1.19726455e+00 1.07894015e+00 -4.94846255e-01 6.25412226e-01
8.82330775e-01 -3.54148954e-01 1.79396176e+00 -1.43296432e+00
8.70000124e-01 1.42926490e+00 9.26230311e-01 9.98758301e-02
-1.36647856e+00 -5.17234623e-01 3.78661364e-01 2.32581109e-01
-8.57307196e-01 -1.53092474e-01 2.62273103e-01 -7.26321042e-01
1.67660081e+00 1.91840246e-01 4.51581657e-01 1.35000265e+00
4.74360764e-01 1.17645264e+00 1.24694097e+00 -1.04641712e+00
-3.56722504e-01 7.79769003e-01 7.79970706e-01 7.45799720e-01
5.80109050e-03 -2.95489520e-01 -6.61724210e-01 -4.98860836e-01
-9.48310494e-02 -2.37379223e-01 -3.15425217e-01 -2.50880867e-01
-1.24291527e+00 1.39771509e+00 5.03122993e-02 5.70615053e-01
-2.99461409e-02 -4.75598663e-01 7.40828216e-01 1.02627826e+00
9.89408493e-01 4.92383689e-01 -1.26382351e+00 -1.00662969e-01
-6.45280361e-01 1.96358114e-01 9.35698211e-01 5.82813323e-01
5.76416492e-01 -1.75445452e-01 2.16982812e-01 1.25018716e+00
3.67098242e-01 4.04535860e-01 1.04479420e+00 -7.53027573e-02
6.92529678e-01 5.94675124e-01 -8.62552822e-02 -7.15011358e-01
-7.77768612e-01 -3.61547530e-01 -2.66157746e-01 -1.83691889e-01
6.30504116e-02 -2.34678537e-01 -1.93499282e-01 9.34804916e-01
-5.52888438e-02 -6.46011591e-01 2.41402611e-01 6.52923763e-01
9.06530738e-01 5.41114450e-01 -2.36327842e-01 1.16147690e-01
1.29586387e+00 -1.58813131e+00 -5.03192067e-01 -2.39602730e-01
1.29407084e+00 -1.40495634e+00 1.16447043e+00 7.57685363e-01
-5.53189874e-01 -3.59905154e-01 -1.09223223e+00 -4.53693181e-01
-8.59992564e-01 2.42915526e-01 7.68481672e-01 7.95696735e-01
-1.16324258e+00 2.07174972e-01 -2.46661112e-01 -7.26132810e-01
-1.34948924e-01 1.63232714e-01 -9.13529515e-01 -3.64887685e-01
-1.24277198e+00 1.25727141e+00 3.46310288e-01 -4.40771341e-01
-2.50411838e-01 -6.74159348e-01 -1.05075634e+00 -5.39859414e-01
-3.51463675e-01 -1.08584382e-01 1.52448654e+00 -9.58931625e-01
-1.33020842e+00 1.14189374e+00 -2.64630497e-01 -3.10371578e-01
4.12570983e-01 -6.08733892e-01 -1.02793992e+00 -4.19840246e-01
3.16238433e-01 4.20702130e-01 4.62945133e-01 -6.93988323e-01
-8.19555283e-01 -2.87993193e-01 -3.29948843e-01 2.13815510e-01
-7.69206226e-01 7.11764574e-01 5.11354953e-02 -6.74893379e-01
-2.40970448e-01 -8.30273509e-01 -1.82462677e-01 -6.48801804e-01
-2.10850134e-01 -8.14173222e-01 4.28421676e-01 -1.00443745e+00
1.20422447e+00 -1.78500748e+00 7.42617026e-02 -1.77580595e-01
3.29857655e-02 8.46952423e-02 -4.35235381e-01 8.93803239e-01
-3.58753711e-01 3.07147801e-01 4.75178182e-01 -4.43312317e-01
1.84126049e-01 8.38497356e-02 -3.58567178e-01 6.83976054e-01
3.39204490e-01 9.00445759e-01 -1.06484377e+00 -3.42172533e-01
7.65514094e-03 2.47527987e-01 -5.56820393e-01 -1.17860585e-01
1.31866440e-01 8.85205492e-02 -1.99849214e-02 6.91549659e-01
4.60263640e-01 3.08755487e-01 5.54308891e-01 5.67200072e-02
-4.74198461e-01 7.78115213e-01 -8.11093986e-01 1.69469857e+00
-7.31391490e-01 1.01048613e+00 -4.70303707e-02 -9.01935041e-01
9.09443736e-01 6.07126772e-01 1.54351816e-01 -8.33860159e-01
1.63001433e-01 8.33953261e-01 -4.60974798e-02 -3.71912599e-01
5.79314828e-01 4.87573780e-02 -6.07069790e-01 8.57477307e-01
5.94439089e-01 6.88531920e-02 4.98564124e-01 3.93392742e-01
7.40993381e-01 -1.13088310e-01 6.42072856e-01 -6.88505411e-01
4.94536161e-01 1.01834111e-01 1.71431124e-01 5.15470147e-01
-2.96240687e-01 2.62362331e-01 3.50237250e-01 -8.89522851e-01
-1.08335841e+00 -7.95990646e-01 -4.91294622e-01 2.08255243e+00
-7.04558671e-01 -7.30935454e-01 -4.67332304e-02 -1.14125216e+00
4.24322873e-01 6.08270109e-01 -4.72384900e-01 4.57847059e-01
-4.19567138e-01 -8.03086996e-01 5.15663445e-01 3.13148469e-01
-5.42449534e-01 -8.94272029e-01 1.92957580e-01 4.47877705e-01
-1.88333526e-01 -1.00788629e+00 -7.10726917e-01 7.74784625e-01
-7.22074270e-01 -1.06605589e+00 -5.51234782e-01 -1.35785472e+00
2.42719993e-01 6.13832697e-02 1.67449093e+00 -3.62676471e-01
-2.96360672e-01 2.83506125e-01 -7.65777349e-01 -5.50585687e-01
-3.79752904e-01 5.66230297e-01 5.06131351e-01 -2.74755269e-01
1.00062096e+00 2.51002871e-02 1.37237608e-02 3.46098170e-02
-2.91930646e-01 -5.52780628e-01 5.17021656e-01 1.10042489e+00
7.65900835e-02 -4.40476000e-01 6.87666953e-01 -1.10116303e+00
1.24117970e+00 -6.10932767e-01 -2.69326959e-02 2.96407968e-01
-9.02613580e-01 -4.69809212e-02 7.60169566e-01 -3.65246117e-01
-3.91008437e-01 -2.59271324e-01 -4.65126991e-01 5.01604617e-01
2.44008861e-02 7.42626131e-01 4.68664736e-01 1.59709737e-01
9.29238617e-01 -1.08196018e-02 -3.96302938e-01 -8.83953869e-01
7.63617992e-01 1.17807209e+00 3.17814648e-02 -4.10470933e-01
3.76180619e-01 -1.21425271e-01 -9.58018661e-01 -7.25606799e-01
-8.59071612e-01 -1.09069681e+00 -1.04098666e+00 1.64389625e-01
6.58325911e-01 -1.17093849e+00 -2.80470192e-01 1.86004579e-01
-1.08042657e+00 5.23353256e-02 2.49466524e-01 7.68169880e-01
1.03815898e-01 1.81797072e-01 -1.27482510e+00 -6.09636724e-01
-5.52567422e-01 -8.74850214e-01 7.63714373e-01 -4.50522870e-01
-7.28493452e-01 -1.61275756e+00 6.39073372e-01 3.75267684e-01
4.27489370e-01 -5.03088951e-01 1.15631354e+00 -1.26734614e+00
6.08558238e-01 -5.11031628e-01 -3.71234417e-02 7.22353697e-01
-5.75532392e-02 1.53767678e-03 -8.18380892e-01 -6.65723383e-01
-3.59718978e-01 -1.03867447e+00 1.14087534e+00 -7.44845122e-02
4.36068863e-01 -3.28152359e-01 -3.73854116e-02 -7.41123334e-02
1.36916649e+00 -3.67478073e-01 -1.64353877e-01 6.90723956e-01
6.74100041e-01 6.19335473e-01 3.08717012e-01 2.29628205e-01
8.69012237e-01 4.43362743e-01 7.68442452e-02 -8.43253210e-02
3.17094117e-01 -2.30823368e-01 8.51767838e-01 1.86701620e+00
4.25426453e-01 -1.58378810e-01 -1.06683588e+00 7.43493497e-01
-1.74153614e+00 -4.22179997e-01 -6.58356905e-01 1.81841910e+00
1.03814089e+00 1.03107855e-01 2.36855030e-01 2.50725985e-01
1.03908315e-01 1.11018375e-01 6.73975721e-02 -1.42350233e+00
-4.41045374e-01 6.42013490e-01 6.01259112e-01 7.67719567e-01
-1.18822932e+00 9.51438665e-01 7.23585176e+00 5.18246531e-01
-8.15234840e-01 6.61675394e-01 3.11289012e-01 -1.34873912e-02
-4.76604432e-01 -8.68609920e-02 -1.11177790e+00 1.24243766e-01
1.19647312e+00 -8.13724026e-02 6.30432740e-02 7.71381974e-01
-3.72914314e-01 1.10112324e-01 -1.25104880e+00 7.16354728e-01
5.16888499e-01 -1.16355383e+00 2.05969781e-01 -2.55201489e-01
8.75398517e-01 7.05347061e-01 7.29824901e-02 8.82439256e-01
7.19301939e-01 -1.14051831e+00 7.07038283e-01 7.46240169e-02
4.85392511e-01 -9.49317098e-01 1.04185283e+00 4.55264360e-01
-8.20639551e-01 -1.70097828e-01 -3.39384079e-01 -1.55405462e-01
1.16586350e-01 5.02877176e-01 -8.52659345e-01 4.93983358e-01
8.31559479e-01 1.06608558e+00 -7.23195553e-01 4.47764605e-01
-2.15861633e-01 8.05463910e-01 9.30903628e-02 -1.76042289e-01
4.82405335e-01 1.97233051e-01 2.97045588e-01 1.98493493e+00
2.75053382e-02 -8.86881649e-01 7.11563945e-01 -3.10440272e-01
-1.91048652e-01 1.10290408e+00 -6.22664571e-01 -1.08209730e-03
1.78210884e-02 1.30596352e+00 -2.11541086e-01 -1.50299907e-01
-8.46293807e-01 1.01255369e+00 8.19373071e-01 -5.29214274e-04
-2.09972635e-01 -5.29059529e-01 2.72723615e-01 -8.33803117e-02
2.33518228e-01 -1.71268210e-01 -9.90328658e-03 -1.34836030e+00
-2.07721237e-02 -1.27583086e+00 6.01204395e-01 -3.06215078e-01
-1.76855898e+00 8.63717198e-01 -7.95696437e-01 -8.86674047e-01
-2.69788355e-01 -1.38057649e+00 -1.01884797e-01 1.08342516e+00
-1.90712667e+00 -1.27214193e+00 5.54480374e-01 5.63147962e-01
6.04445994e-01 -8.00337851e-01 1.44070899e+00 6.33058548e-01
-5.92621267e-01 9.49105680e-01 5.58607697e-01 1.86933890e-01
1.45134985e+00 -1.67147899e+00 3.46609473e-01 1.77025825e-01
4.86673176e-01 5.45058429e-01 3.86476666e-01 -1.78847998e-01
-1.22181320e+00 -6.82091951e-01 2.16403818e+00 -9.87922251e-01
1.51543486e+00 -6.99425638e-01 -4.14403528e-01 9.32127774e-01
8.53931606e-01 -2.15330884e-01 1.35927701e+00 9.23192561e-01
-9.67904449e-01 4.17216122e-03 -6.75023496e-01 3.19101423e-01
1.91451862e-01 -8.80537629e-01 -7.91089475e-01 7.90170014e-01
7.46353626e-01 1.17039084e-01 -1.00009978e+00 -2.62618251e-02
6.70181096e-01 -6.44234717e-01 6.62924111e-01 -1.08634698e+00
4.65857446e-01 3.30258191e-01 -4.75775778e-01 -1.82161772e+00
-4.70443726e-01 -3.42013568e-01 3.10230702e-01 1.02372313e+00
1.14426017e+00 -9.85896587e-01 1.52661696e-01 -4.15637285e-01
-1.19202636e-01 -6.94772124e-01 -9.85647202e-01 -7.37722576e-01
7.73055792e-01 -7.01052368e-01 7.57829323e-02 1.47323942e+00
6.31746292e-01 9.74493384e-01 -2.55173326e-01 -5.05638123e-01
1.71200484e-01 -2.02814266e-01 5.69239080e-01 -1.59454858e+00
-3.90324980e-01 -8.11348975e-01 -3.35588038e-01 -1.00126135e+00
4.14777160e-01 -1.64508700e+00 -3.81334573e-01 -1.56132495e+00
2.50785023e-01 -2.57771224e-01 -6.16751313e-01 4.44873810e-01
-1.01161525e-01 3.22345436e-01 -8.35605338e-02 1.15739286e-01
-8.40225220e-01 8.40783715e-02 6.11085057e-01 -3.99672061e-01
3.82779360e-01 -2.32560605e-01 -8.89998138e-01 4.26889986e-01
5.76972246e-01 -6.51622355e-01 1.08662471e-01 -7.80800521e-01
7.62467563e-01 -4.25053000e-01 -2.20056757e-01 -2.43347347e-01
-1.15963146e-01 2.74784476e-01 2.89321482e-01 -4.32660013e-01
-2.10216254e-01 -5.76245606e-01 -7.07738400e-01 3.26854974e-01
-4.88990963e-01 8.62462044e-01 2.29730099e-01 4.35390294e-01
-4.33528692e-01 -4.55609709e-01 3.19596559e-01 2.31868271e-02
-6.25939786e-01 -4.40363847e-02 -8.01334202e-01 8.17066282e-02
6.24599099e-01 5.16336262e-01 -4.25816029e-01 -3.29480506e-02
-6.71798289e-01 5.17139435e-01 1.07301772e-01 9.82361019e-01
1.97075829e-01 -1.47261298e+00 -1.33238602e+00 2.78317332e-01
7.99065709e-01 -1.07072425e+00 -2.85246402e-01 9.37723696e-01
-5.51537931e-01 1.05224407e+00 8.71700570e-02 -3.26952010e-01
-1.18051267e+00 3.81338745e-01 -3.61604725e-05 -8.96168292e-01
-1.17300741e-01 1.15953660e+00 -6.14267170e-01 -1.54828143e+00
8.53677988e-02 -2.74273753e-01 -5.77528119e-01 7.07357109e-01
6.11858308e-01 2.13694111e-01 5.92730522e-01 -6.99263453e-01
-6.50798261e-01 3.34451467e-01 -5.71877778e-01 -4.00081605e-01
1.32563925e+00 9.20174737e-03 -5.82705915e-01 1.27681887e+00
1.61718357e+00 3.74472260e-01 -6.47869557e-02 -4.16681260e-01
5.30184388e-01 -2.69563925e-02 1.30663544e-01 -8.91555786e-01
-4.84350353e-01 9.77743149e-01 5.58420420e-01 3.18936616e-01
2.38862216e-01 1.16851386e-02 7.96174765e-01 5.82922280e-01
3.97184342e-01 -1.44221318e+00 -2.14326270e-02 1.06102991e+00
8.76503885e-01 -1.54553151e+00 -8.70338529e-02 3.56945008e-01
-6.84149861e-01 1.43490934e+00 2.39961952e-01 -1.08764805e-01
1.01856041e+00 1.30184054e-01 6.39623106e-01 5.05203754e-02
-1.14162135e+00 7.13732913e-02 4.96695846e-01 4.76938725e-01
1.36187077e+00 4.37826574e-01 -6.82371736e-01 6.13641798e-01
-4.83728021e-01 -4.74870652e-01 2.56117374e-01 8.96149933e-01
-3.46922666e-01 -1.55929077e+00 1.42160291e-02 5.72597563e-01
-8.79117727e-01 -6.84800327e-01 -4.65670258e-01 6.12292469e-01
-1.01674370e-01 1.19345212e+00 -2.45520957e-02 -4.52058375e-01
4.54220802e-01 3.90919805e-01 2.15437442e-01 -8.04252267e-01
-9.14601624e-01 -1.12088673e-01 5.53530276e-01 -1.19657457e-01
-2.15787292e-01 -9.24649239e-01 -6.42110169e-01 -4.08065379e-01
-3.74861002e-01 2.36015528e-01 9.32725966e-01 8.82780015e-01
4.38307747e-02 1.69386521e-01 7.50190794e-01 -3.89997423e-01
-5.66110492e-01 -1.46846414e+00 -5.72432995e-01 1.84045285e-01
3.54609460e-01 -2.99025059e-01 -5.31020761e-01 -2.75165796e-01] | [10.852616310119629, 9.780913352966309] |
5d92187d-eb78-43a3-9ff7-341ea206329a | multi-modal-unsupervised-pre-training-for | 2207.07894 | null | https://arxiv.org/abs/2207.07894v1 | https://arxiv.org/pdf/2207.07894v1.pdf | Multi-Modal Unsupervised Pre-Training for Surgical Operating Room Workflow Analysis | Data-driven approaches to assist operating room (OR) workflow analysis depend on large curated datasets that are time consuming and expensive to collect. On the other hand, we see a recent paradigm shift from supervised learning to self-supervised and/or unsupervised learning approaches that can learn representations from unlabeled datasets. In this paper, we leverage the unlabeled data captured in robotic surgery ORs and propose a novel way to fuse the multi-modal data for a single video frame or image. Instead of producing different augmentations (or 'views') of the same image or video frame which is a common practice in self-supervised learning, we treat the multi-modal data as different views to train the model in an unsupervised manner via clustering. We compared our method with other state of the art methods and results show the superior performance of our approach on surgical video activity recognition and semantic segmentation. | ['Omid Mohareri', 'Muhammad Abdullah Jamal'] | 2022-07-16 | null | null | null | null | ['unsupervised-pre-training'] | ['methodology'] | [ 5.62941611e-01 3.96729022e-01 -5.08798122e-01 -5.83260894e-01
-7.56920099e-01 -6.21541381e-01 2.79003888e-01 3.03927600e-01
-5.77221811e-01 3.96379381e-01 4.43870038e-01 -1.71280086e-01
-2.08496317e-01 -3.93498629e-01 -6.71897352e-01 -8.89623106e-01
3.34350854e-01 4.78182793e-01 7.91320875e-02 1.98764279e-01
1.51602412e-02 3.75513464e-01 -1.64451742e+00 6.41862631e-01
6.36548996e-01 7.83771455e-01 3.24443758e-01 7.70662487e-01
-2.66201794e-01 1.10172737e+00 -1.54246345e-01 1.51558623e-01
3.36850077e-01 -5.85397780e-01 -1.01327837e+00 6.24586999e-01
2.79297233e-01 -2.25838721e-01 -4.16358352e-01 9.50462103e-01
3.76173943e-01 -1.60633326e-02 5.32278538e-01 -1.03196979e+00
1.30089045e-01 4.36132789e-01 -3.24366063e-01 3.16594154e-01
1.56690866e-01 -5.57824709e-02 2.40671024e-01 -5.61942875e-01
1.02024686e+00 6.86475873e-01 2.47736692e-01 7.74811983e-01
-9.76120710e-01 -3.21607113e-01 1.98448181e-01 6.79363310e-02
-9.29299295e-01 -5.74341059e-01 8.00992727e-01 -6.76651239e-01
3.43070984e-01 1.28882706e-01 7.80022264e-01 1.21059608e+00
5.34536727e-02 9.94138360e-01 1.12859213e+00 -5.73635161e-01
3.71097594e-01 1.38825625e-01 1.23836780e-02 9.11102653e-01
1.60959765e-01 6.35161847e-02 -5.47990203e-01 2.14694217e-01
8.47920179e-01 7.43822455e-01 -2.56106585e-01 -1.03541481e+00
-1.60701752e+00 6.35938168e-01 1.26010761e-01 4.45402592e-01
-4.05717164e-01 -4.41151150e-02 4.53670710e-01 1.84530735e-01
1.52837113e-01 3.11810672e-01 -4.31644887e-01 -3.18419933e-01
-9.86841202e-01 -4.39994603e-01 8.51484478e-01 8.99361372e-01
7.21202493e-01 -4.02193040e-01 1.96416508e-02 6.30084217e-01
3.13882470e-01 -2.02569813e-02 8.59372556e-01 -1.06976473e+00
2.32454553e-01 8.97727370e-01 -5.96368574e-02 -4.76092607e-01
-5.76961517e-01 -1.21035464e-01 -6.83986902e-01 2.81469584e-01
4.92793381e-01 -2.16584504e-01 -1.30293036e+00 1.28608096e+00
1.64449021e-01 1.14787482e-01 2.54836470e-01 8.31846178e-01
1.01634336e+00 7.83696249e-02 2.94343904e-02 -6.17673635e-01
9.83642042e-01 -1.31254244e+00 -9.43285227e-01 -1.88624293e-01
9.98068035e-01 -5.76604247e-01 7.76661098e-01 5.48097074e-01
-9.37068939e-01 -3.91458035e-01 -9.45735037e-01 2.52842307e-01
-4.05848026e-01 3.08338523e-01 7.55410135e-01 6.22746885e-01
-7.81512022e-01 4.11446840e-01 -1.35042679e+00 -4.93526101e-01
7.25254655e-01 5.54578364e-01 -8.21180880e-01 -3.81471634e-01
-2.12053552e-01 5.80130100e-01 5.31938851e-01 1.32504636e-02
-1.23785150e+00 -5.22445500e-01 -8.65516126e-01 -2.87449867e-01
7.84276247e-01 -4.06923980e-01 1.17119753e+00 -1.40090752e+00
-1.51044774e+00 1.18083060e+00 -1.36491552e-01 -3.21453691e-01
3.26838940e-01 -2.42088541e-01 -1.50397047e-01 4.98407096e-01
-2.58844256e-01 5.62828779e-01 8.03444624e-01 -1.42963493e+00
-5.50251722e-01 -5.83546698e-01 1.50265902e-01 3.12076360e-01
-4.81034368e-01 -2.03216076e-01 -5.62786460e-01 -5.13952553e-01
4.06970888e-01 -1.31190503e+00 -4.55637664e-01 -8.10752511e-02
-2.40985736e-01 2.71208256e-01 9.05667484e-01 -3.91755372e-01
1.05544758e+00 -2.18237019e+00 4.34941649e-01 1.02640860e-01
3.40955347e-01 1.76905975e-01 1.86196849e-01 2.10073248e-01
-2.32403323e-01 -2.29749680e-01 -3.74642968e-01 -2.54898787e-01
-6.17883265e-01 5.98235190e-01 3.55300397e-01 5.56024194e-01
-9.00461003e-02 6.95277572e-01 -1.26476753e+00 -8.24177682e-01
4.52482134e-01 1.20315671e-01 -6.64985240e-01 3.83533925e-01
2.31326721e-03 1.28699958e+00 -4.15979207e-01 7.11740732e-01
1.48633927e-01 -3.42671424e-01 5.08570671e-01 -3.88071656e-01
1.73395485e-01 -1.22894754e-03 -1.08644855e+00 2.79566216e+00
-4.09596354e-01 4.09003347e-01 -4.21991944e-02 -1.53895879e+00
5.08342505e-01 5.09688675e-01 1.21584272e+00 -2.15899989e-01
3.43402654e-01 2.59678941e-02 7.37237409e-02 -1.03012764e+00
5.19296788e-02 -1.13974355e-01 -2.34767934e-03 5.27554333e-01
6.88673258e-01 1.06840201e-01 2.22307861e-01 2.07063869e-01
1.31972694e+00 4.10206527e-01 3.55181307e-01 -1.05024688e-01
5.53438425e-01 3.17716181e-01 5.51254094e-01 6.46075308e-01
-2.92886436e-01 8.82469356e-01 3.56396288e-01 -6.00179911e-01
-7.91565955e-01 -1.01452196e+00 1.38331085e-01 8.63839209e-01
2.03944623e-01 -5.56562662e-01 -8.42415869e-01 -1.25554442e+00
-4.19726104e-01 6.04216419e-02 -8.10816348e-01 -1.50251672e-01
-5.17732203e-01 -6.20267749e-01 -4.08371066e-04 7.63317347e-01
-3.53091732e-02 -9.08542037e-01 -8.16827834e-01 1.41529948e-01
-2.89101303e-01 -1.31321859e+00 -2.15993270e-01 5.83036184e-01
-1.33710432e+00 -1.30305207e+00 -5.98462045e-01 -9.20516789e-01
1.33584690e+00 3.26929599e-01 9.54805195e-01 -1.47582233e-01
-5.39692640e-01 9.90373075e-01 -6.05270624e-01 -5.79432487e-01
-4.64752704e-01 2.63051987e-02 1.62635475e-01 2.44635805e-01
2.51168162e-01 -4.10953432e-01 -7.66755342e-01 2.78179646e-01
-1.23823690e+00 1.73936024e-01 8.10544789e-01 8.12900960e-01
8.56107354e-01 -2.51909852e-01 2.03181416e-01 -1.33276904e+00
4.03754860e-02 -5.44268548e-01 -3.33580524e-01 2.34629244e-01
-4.88607317e-01 1.32261500e-01 3.80343914e-01 -3.82084668e-01
-1.08507597e+00 8.36157441e-01 3.54189336e-01 -9.61842597e-01
-6.39671922e-01 5.20211756e-01 -1.02081798e-01 -7.94582441e-02
6.09809458e-01 1.37581001e-03 4.92515564e-01 -4.03176397e-01
3.41756374e-01 7.11812019e-01 6.73525751e-01 -2.70111650e-01
6.28303289e-01 8.85429382e-01 -1.71521798e-01 -4.66205865e-01
-9.37188327e-01 -9.98414993e-01 -1.18639410e+00 -5.13762653e-01
1.12199068e+00 -8.89228582e-01 -3.72998327e-01 9.17313993e-02
-6.34578526e-01 -1.39455646e-01 -5.22652030e-01 9.89948809e-01
-8.65073979e-01 3.18876326e-01 -2.88945526e-01 -6.01460338e-01
-1.05459101e-01 -1.35123551e+00 9.98672843e-01 3.80901337e-01
-4.54768166e-02 -9.47710395e-01 2.80807108e-01 7.71759927e-01
4.26614545e-02 3.94681185e-01 4.31067824e-01 -9.09224570e-01
-5.99863112e-01 -4.13160831e-01 1.95073381e-01 4.99464393e-01
5.73070168e-01 -2.86606640e-01 -8.95281017e-01 -2.68708795e-01
2.26002455e-01 -2.93409526e-01 5.26888132e-01 4.03910607e-01
1.48410642e+00 -1.09274238e-01 -5.34999967e-01 6.63358450e-01
1.39400494e+00 1.90016702e-01 6.75148904e-01 2.68684894e-01
7.28256464e-01 7.74089992e-01 9.94006932e-01 2.03586921e-01
1.12920888e-01 2.83986568e-01 5.07466555e-01 -2.91954041e-01
1.56899214e-01 -7.16588199e-02 1.31851181e-01 9.96397555e-01
-4.94084001e-01 1.77929252e-01 -1.02397418e+00 5.97205222e-01
-2.16115189e+00 -8.68859112e-01 2.35666528e-01 2.30737519e+00
6.10537112e-01 -9.24374089e-02 1.25607669e-01 1.04524009e-01
4.47275639e-01 -1.23729490e-01 -4.14801031e-01 2.05042418e-02
3.17920983e-01 1.67065293e-01 6.03372395e-01 -6.73143193e-02
-1.32944679e+00 6.52671218e-01 6.08056068e+00 2.86182135e-01
-1.07414532e+00 2.70455271e-01 4.41893876e-01 -4.57619309e-01
1.22481994e-01 -5.50866239e-02 -2.35238120e-01 2.52195686e-01
1.01083231e+00 1.57938793e-01 2.59626180e-01 9.71946895e-01
2.64196277e-01 -3.86795402e-01 -1.37128067e+00 1.32805777e+00
4.01097298e-01 -1.55395365e+00 -2.35326424e-01 2.14851320e-01
8.71148527e-01 6.10930584e-02 -3.90466601e-01 -5.84629886e-02
1.86873659e-01 -9.71023500e-01 1.20228522e-01 7.51185060e-01
5.22848725e-01 -3.17214608e-01 8.96970749e-01 4.93009478e-01
-8.38345706e-01 -2.32554853e-01 1.26457110e-01 4.31132108e-01
-1.25869364e-01 7.56181851e-02 -7.01990306e-01 6.93578005e-01
9.17103827e-01 1.07104516e+00 -5.20893991e-01 1.07853091e+00
9.98794585e-02 3.98030460e-01 -1.03113562e-01 6.25523210e-01
2.49041811e-01 -2.64175922e-01 2.98336834e-01 9.19981539e-01
6.93648085e-02 9.04970542e-02 3.74777704e-01 2.37608016e-01
-4.37932462e-02 4.34495918e-02 -8.25373292e-01 -3.24928939e-01
-2.01031342e-01 1.56843042e+00 -1.02080417e+00 -4.78018075e-01
-8.13634515e-01 9.74041879e-01 1.13936039e-02 1.46685302e-01
-5.41844487e-01 1.28609911e-01 1.65167391e-01 1.57909229e-01
1.19646341e-01 -2.70752460e-01 5.31194219e-03 -1.40457559e+00
4.47378121e-03 -8.56564164e-01 6.67933643e-01 -5.62836409e-01
-9.65928674e-01 4.53796029e-01 4.91271764e-02 -2.05217099e+00
-3.20713967e-01 -5.80217361e-01 -3.65271777e-01 8.25658813e-02
-1.50607347e+00 -1.07191694e+00 -9.08839464e-01 9.47027981e-01
7.20242739e-01 -3.72464955e-01 1.06256056e+00 2.09123045e-01
-4.50693876e-01 1.10003345e-01 3.33980948e-01 3.64297181e-01
9.13407385e-01 -1.18661797e+00 -7.03713477e-01 6.38440192e-01
3.18737298e-01 5.35445929e-01 3.07847977e-01 -3.25912982e-01
-1.82812297e+00 -8.22707891e-01 -3.91294993e-03 -5.23417532e-01
4.67676908e-01 -1.70077473e-01 -5.82401812e-01 9.13928747e-01
2.06143335e-01 5.49431384e-01 1.30576956e+00 -2.17726097e-01
1.44169912e-01 -1.20112509e-01 -1.04441464e+00 2.64057755e-01
1.11989927e+00 -3.33607346e-01 -6.53431952e-01 5.25618732e-01
3.21299940e-01 -6.02910221e-01 -1.19499195e+00 6.83316231e-01
5.41643083e-01 -9.20931041e-01 7.32799470e-01 -8.33195329e-01
5.16024768e-01 -4.03011769e-01 9.52890366e-02 -1.05965638e+00
3.59012306e-01 -5.67587614e-01 -1.39066830e-01 7.53183722e-01
2.14663923e-01 -1.93109140e-01 1.29091835e+00 4.72236276e-01
-4.11696345e-01 -5.86012483e-01 -7.29925096e-01 -4.43457603e-01
-3.10732991e-01 -2.67288923e-01 -1.28295079e-01 1.06447458e+00
3.63563925e-01 -1.31591707e-02 -8.60193297e-02 5.40666357e-02
5.75416565e-01 2.04454198e-01 7.93969572e-01 -1.25675535e+00
-1.75682917e-01 1.90591335e-01 -6.56276524e-01 -5.28140306e-01
5.57579957e-02 -8.58582497e-01 3.52524847e-01 -1.67244291e+00
3.47421318e-01 -3.24448287e-01 -5.60108006e-01 4.67835933e-01
1.15426846e-01 1.87666506e-01 -6.39186054e-02 5.51876664e-01
-9.10039008e-01 1.27649024e-01 1.11280560e+00 -1.64850459e-01
-4.59748477e-01 -3.61803062e-02 -4.54191804e-01 1.04146528e+00
5.81073761e-01 -6.05715215e-01 -6.17559373e-01 -3.63643646e-01
-3.26931059e-01 2.93614835e-01 7.80921131e-02 -1.10090113e+00
4.71246362e-01 -5.38256615e-02 3.42528135e-01 -4.43562955e-01
1.05123736e-01 -1.26187396e+00 -2.18485482e-02 4.70915526e-01
-3.04198563e-01 -1.93059206e-01 -3.27254906e-02 8.95233154e-01
-6.03816271e-01 -2.32651532e-01 6.92548573e-01 -6.77863181e-01
-7.46113062e-01 3.85402173e-01 -5.50668836e-01 -1.21787898e-01
1.29943097e+00 -3.67222190e-01 -5.68296798e-02 -1.41985223e-01
-1.40956807e+00 2.89957598e-02 5.75816453e-01 7.34173000e-01
7.08521426e-01 -1.05179644e+00 -2.58146882e-01 1.90010071e-01
5.30641019e-01 3.95507902e-01 3.75446290e-01 1.37978256e+00
-5.12304187e-01 4.25232500e-01 -4.50972170e-01 -9.71449792e-01
-1.25899899e+00 7.13358343e-01 1.80145428e-01 -3.51796806e-01
-7.58284152e-01 2.59346783e-01 2.24148124e-01 -4.87682790e-01
3.64054561e-01 -1.16338201e-01 -4.39914435e-01 1.18849508e-01
3.27641755e-01 -2.04077940e-02 2.26424605e-01 -5.03594160e-01
-2.80379176e-01 3.38007838e-01 -1.87582791e-01 1.33675590e-01
1.43994319e+00 -7.60465786e-02 7.43331015e-02 6.72113359e-01
1.19476736e+00 -2.68931091e-01 -1.19115508e+00 -3.30756873e-01
1.74762055e-01 -3.66504699e-01 2.51882672e-01 -5.24312317e-01
-1.42405391e+00 6.36151314e-01 9.65767622e-01 -2.21153349e-01
1.16777110e+00 1.91182360e-01 4.81023014e-01 4.14319545e-01
4.27056789e-01 -1.27027345e+00 3.23834151e-01 -6.02039061e-02
3.91450018e-01 -1.67539263e+00 -5.90431467e-02 -4.89083111e-01
-8.76197457e-01 1.29272795e+00 5.58683157e-01 -4.14606221e-02
6.71539068e-01 4.19073671e-01 3.81019950e-01 -2.54800826e-01
-4.48226660e-01 -4.06924844e-01 1.95278406e-01 5.73973596e-01
4.45769697e-01 -2.75504887e-01 -1.77012965e-01 5.21925330e-01
3.18340480e-01 5.36612451e-01 6.96763217e-01 1.66557932e+00
-7.04631507e-02 -1.29063725e+00 -2.43623927e-01 6.76072896e-01
-5.19770801e-01 1.50245368e-01 -1.31970361e-01 6.25347733e-01
-6.31470531e-02 9.19607699e-01 -4.78866138e-02 -2.55948991e-01
2.40411341e-01 2.75024593e-01 6.49143159e-01 -1.01014221e+00
-5.23291171e-01 3.56417567e-01 -1.09507062e-01 -9.30494130e-01
-1.23630071e+00 -7.44773149e-01 -1.32575452e+00 5.32503843e-01
-6.95901811e-02 4.11814302e-02 8.70107234e-01 1.09878802e+00
2.40216240e-01 8.28047574e-01 6.29886627e-01 -9.91912246e-01
1.18213087e-01 -7.31647730e-01 -4.09155399e-01 7.00909853e-01
3.34118009e-01 -7.32030511e-01 -2.35464096e-01 7.56424308e-01] | [14.18350887298584, -3.0535244941711426] |
67c71575-08fa-454c-b9a4-9f6c6e7a1426 | towards-summarizing-multiple-documents-with | 2305.01498 | null | https://arxiv.org/abs/2305.01498v1 | https://arxiv.org/pdf/2305.01498v1.pdf | Towards Summarizing Multiple Documents with Hierarchical Relationships | Most existing multi-document summarization (MDS) datasets lack human-generated and genuine (i.e., not synthetic) summaries or source documents with explicit inter-document relationships that a summary must capture. To enhance the capabilities of MDS systems we present PeerSum, a novel dataset for generating meta-reviews of scientific papers, where the meta-reviews are highly abstractive and genuine summaries of reviews and corresponding discussions. These source documents have rich inter-document relationships of an explicit hierarchical structure with cross-references and often feature conflicts. As there is a scarcity of research that incorporates hierarchical relationships into MDS systems through attention manipulation on pre-trained language models, we additionally present Rammer (Relationship-aware Multi-task Meta-review Generator), a meta-review generation model that uses sparse attention based on the hierarchical relationships and a multi-task objective that predicts several metadata features in addition to the standard text generation objective. Our experimental results show that PeerSum is a challenging dataset, and Rammer outperforms other strong baseline MDS models under various evaluation metrics. | ['Jey Han Lau', 'Eduard Hovy', 'Miao Li'] | 2023-05-02 | null | null | null | null | ['review-generation', 'multi-document-summarization', 'document-summarization'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 1.36572748e-01 5.91177821e-01 -4.68619436e-01 -2.67988145e-01
-1.52041280e+00 -4.83385265e-01 1.30375195e+00 4.04818684e-01
1.59330338e-01 1.20312715e+00 1.12244499e+00 -2.84840554e-01
6.59966618e-02 -5.92476130e-01 -8.64830673e-01 -2.57915795e-01
3.27495992e-01 6.13989532e-01 -1.72041640e-01 -8.09102654e-02
7.75650620e-01 -1.22194082e-01 -1.53351092e+00 8.61825049e-01
1.41009462e+00 3.75395566e-01 3.07567179e-01 1.06795895e+00
-4.77589518e-01 8.91580284e-01 -1.38089049e+00 -6.95313811e-01
-2.07713068e-01 -5.50684810e-01 -9.11130309e-01 2.70362571e-02
8.87727141e-01 -1.69829682e-01 -1.86012372e-01 5.77557027e-01
8.66836250e-01 7.23556951e-02 1.25944865e+00 -1.40089095e+00
-1.46699727e+00 1.62314558e+00 -6.94806576e-01 1.47137195e-01
5.19609809e-01 -1.03174962e-01 1.31781328e+00 -1.13007152e+00
8.68157923e-01 1.59813559e+00 3.15817505e-01 3.95548016e-01
-8.64642024e-01 -6.91281974e-01 2.08362207e-01 -4.05538362e-03
-8.54302764e-01 -5.68566859e-01 6.29145086e-01 -2.28384361e-01
1.24422264e+00 3.53075266e-01 2.29522169e-01 1.59712422e+00
6.66767716e-01 8.77067864e-01 5.19305706e-01 -3.39903295e-01
2.68337369e-01 1.63586423e-01 3.04356188e-01 1.94290847e-01
7.46929407e-01 -6.89936519e-01 -9.92700636e-01 -1.50790662e-01
1.75516203e-01 -2.94924915e-01 3.43930684e-02 2.26645440e-01
-1.68646646e+00 6.21780396e-01 9.26589891e-02 8.84539932e-02
-3.36230934e-01 1.17149770e-01 6.60805345e-01 1.79671627e-02
7.28987336e-01 8.33060205e-01 -2.49823302e-01 8.03706795e-02
-1.33338737e+00 6.41212046e-01 8.39033186e-01 1.39553261e+00
2.91699439e-01 1.34327903e-01 -8.86800230e-01 8.07038486e-01
5.06739877e-02 6.14797533e-01 7.20730066e-01 -8.50800335e-01
1.04082882e+00 7.87261367e-01 1.44561589e-01 -1.18191230e+00
-3.04241002e-01 -6.32244110e-01 -1.10692000e+00 -5.74110508e-01
-3.89795959e-01 -1.11487739e-01 -5.41590691e-01 1.36507928e+00
-9.45762992e-02 -1.81985080e-01 5.19279540e-01 2.46617898e-01
2.02461505e+00 1.06350243e+00 -6.05837815e-02 -3.56327921e-01
1.20887005e+00 -1.48570573e+00 -1.04113173e+00 -3.88047010e-01
6.53934896e-01 -7.80305445e-01 1.24103796e+00 2.96945035e-01
-1.49322605e+00 -6.52329326e-01 -1.32026672e+00 -4.35028762e-01
-5.90143979e-01 5.76813519e-01 4.66460347e-01 -4.90611978e-02
-1.20529890e+00 5.62665403e-01 -2.25411683e-01 -1.88386500e-01
4.18783695e-01 -1.15970194e-01 -2.81947285e-01 1.12291694e-01
-1.12882102e+00 7.66971946e-01 1.73098877e-01 -3.09299767e-01
-7.71022916e-01 -1.10865247e+00 -1.13036597e+00 1.40299305e-01
1.85731888e-01 -1.27252173e+00 1.21750391e+00 -1.93686128e-01
-1.04266381e+00 7.44424641e-01 -3.12770844e-01 -3.63371581e-01
2.59373903e-01 -3.93771917e-01 -4.33530748e-01 1.53549626e-01
7.99414456e-01 8.43175471e-01 6.02980494e-01 -1.48562622e+00
-3.94323498e-01 -3.76803316e-02 -9.35797393e-02 3.09540719e-01
-4.37735438e-01 3.55977342e-02 -2.45550901e-01 -1.07599831e+00
-5.94128132e-01 -6.13578498e-01 -1.44485399e-01 -7.98084199e-01
-1.42645407e+00 -5.06216586e-01 5.67251265e-01 -4.90777135e-01
1.61613095e+00 -1.32966149e+00 2.58691728e-01 -5.09188533e-01
2.83206642e-01 1.12903424e-01 -6.43045127e-01 9.48514760e-01
-2.70813424e-03 6.40963197e-01 -2.37271190e-01 -8.22533965e-01
1.87470421e-01 -3.14887643e-01 -9.00836170e-01 -2.66072214e-01
3.52123618e-01 1.19685960e+00 -1.13244355e+00 -6.71336889e-01
-2.35052168e-01 2.35045165e-01 -1.24351807e-01 3.27290088e-01
-4.34823513e-01 7.34168366e-02 -4.07701105e-01 5.28883457e-01
5.02583921e-01 -6.98430181e-01 -3.67985606e-01 -3.02877456e-01
-5.98818772e-02 7.84305215e-01 -8.27392459e-01 1.86245871e+00
-5.49729347e-01 6.86292827e-01 -5.74506700e-01 -7.13357329e-01
1.00942361e+00 2.55709738e-01 3.75113517e-01 -3.54907572e-01
-1.43266797e-01 1.55802399e-01 -5.05543590e-01 -2.70052642e-01
1.38745809e+00 3.83689076e-01 -4.71448898e-01 1.03546071e+00
4.94136028e-02 -6.72958732e-01 9.10774708e-01 1.30760050e+00
1.34318328e+00 -5.49023151e-02 2.87828535e-01 -1.58933669e-01
4.44530696e-01 3.48522551e-02 1.54805034e-01 1.06237960e+00
6.71346009e-01 1.04925823e+00 7.74355769e-01 8.38879272e-02
-1.09089124e+00 -9.81774509e-01 2.92503387e-02 6.73695326e-01
-4.95283008e-02 -1.14857113e+00 -4.44848537e-01 -7.82254398e-01
1.34222537e-01 1.23564577e+00 -7.16878474e-01 -3.06265384e-01
-2.00924382e-01 -8.87042940e-01 5.51621974e-01 6.00893021e-01
8.90724435e-02 -1.26708090e+00 -1.32577211e-01 2.78385788e-01
-2.65223712e-01 -1.13779914e+00 -7.62977302e-01 -1.47614591e-02
-6.82975888e-01 -9.32517290e-01 -8.21579456e-01 -5.53022981e-01
6.81295931e-01 4.17456359e-01 1.80742073e+00 -1.65129572e-01
-9.74106491e-02 3.84404242e-01 -3.94462109e-01 -7.46376038e-01
-7.56392419e-01 5.38711905e-01 -9.36222225e-02 -6.40800774e-01
6.35169521e-02 -4.30356354e-01 -2.52787709e-01 -1.25436768e-01
-1.01970994e+00 4.34326828e-01 8.06685627e-01 8.28380942e-01
4.16241050e-01 -2.94696838e-01 1.54063213e+00 -1.04909098e+00
1.35983336e+00 -6.87303483e-01 7.46077597e-02 5.14272809e-01
-6.22694254e-01 1.77973911e-01 6.95691884e-01 -3.54872614e-01
-1.15775919e+00 -6.14878118e-01 4.20302421e-01 -3.36094238e-02
3.94064993e-01 6.88359976e-01 -3.00501496e-01 8.14668179e-01
7.54358530e-01 4.07008052e-01 -1.81364462e-01 -2.97533274e-01
7.56493449e-01 7.84024119e-01 6.89460456e-01 -5.39609134e-01
5.54786384e-01 1.64804250e-01 -1.44311070e-01 -6.46413147e-01
-1.24249959e+00 -1.64774716e-01 -5.21404207e-01 1.19147770e-01
5.23521066e-01 -1.16593742e+00 -9.46804285e-02 1.45471931e-01
-1.70379925e+00 2.88519505e-02 -4.35383946e-01 1.36304781e-01
-4.00866538e-01 5.23164988e-01 -6.83439314e-01 -5.02129853e-01
-1.19249165e+00 -1.02358866e+00 1.59428000e+00 2.19953552e-01
-7.31863201e-01 -9.93067026e-01 1.21877596e-01 3.86335015e-01
3.25709641e-01 1.84168831e-01 1.05929232e+00 -8.26568186e-01
-4.39347804e-01 -1.13818392e-01 -2.88930058e-01 1.08913653e-01
2.95468092e-01 5.87658644e-01 -6.66222572e-01 2.62708869e-02
-4.59801257e-01 -5.35167396e-01 1.19238329e+00 6.97507799e-01
1.09409666e+00 -8.54960322e-01 -5.11942565e-01 1.53888732e-01
6.87350154e-01 -8.59916285e-02 4.60823715e-01 2.78100133e-01
8.84867668e-01 6.79468453e-01 5.61555326e-01 7.14226484e-01
1.25063527e+00 1.56789377e-01 1.28286824e-01 -6.25302792e-02
-3.80823374e-01 -3.87535244e-01 5.86710453e-01 1.58723533e+00
6.54267609e-01 -8.64232838e-01 -5.35325825e-01 8.15138638e-01
-1.87874198e+00 -1.08104491e+00 -4.13485199e-01 1.70251489e+00
1.29646671e+00 1.22351743e-01 -4.66471352e-02 -2.28185296e-01
7.90564239e-01 5.41509628e-01 -6.77219033e-01 -6.59916639e-01
-6.90736890e-01 -1.07305057e-01 -3.12415343e-02 1.71685383e-01
-8.96560550e-01 6.94364727e-01 6.32455158e+00 7.60461748e-01
-6.53298497e-01 -3.64921778e-01 6.03506505e-01 -4.30420578e-01
-9.91101086e-01 -1.46683261e-01 -1.20868313e+00 5.85019827e-01
1.00830960e+00 -9.62359607e-01 -4.15753573e-01 7.79255867e-01
1.73771188e-01 5.35844825e-02 -1.36972857e+00 7.16327190e-01
5.64424694e-01 -1.93011129e+00 9.45197642e-01 -7.81712234e-02
1.32062066e+00 -3.76439035e-01 1.74020767e-01 5.15304446e-01
6.64184451e-01 -1.08039963e+00 6.13102019e-01 6.14128351e-01
7.57553101e-01 -7.50414729e-01 6.53739512e-01 2.88589686e-01
-8.32381248e-01 2.08856940e-01 -5.94269156e-01 3.57528269e-01
3.91693413e-01 1.05728590e+00 -7.57916749e-01 8.54304552e-01
3.09313983e-01 1.47503221e+00 -9.43726301e-01 6.86234951e-01
-2.68038124e-01 2.74862856e-01 3.00857216e-01 -3.42393994e-01
1.76943749e-01 4.92289737e-02 7.69560337e-01 1.50689650e+00
6.66613758e-01 -3.98167372e-01 -9.68617201e-02 1.06891382e+00
-5.32591522e-01 1.71028465e-01 -7.93610930e-01 -4.62765962e-01
7.45446563e-01 1.49900055e+00 -3.18844527e-01 -7.82044768e-01
-1.42461255e-01 5.17990232e-01 1.09148575e-02 1.78028956e-01
-5.55281639e-01 -6.81083918e-01 2.43929282e-01 -1.30884543e-01
6.40833825e-02 2.05838203e-01 -7.68623352e-01 -1.43414390e+00
1.54275373e-01 -1.06290984e+00 3.03557277e-01 -1.30629599e+00
-1.43777943e+00 8.00234079e-01 6.00874573e-02 -1.24397433e+00
-6.00972593e-01 1.49589747e-01 -1.13632500e+00 7.17130125e-01
-1.37161195e+00 -1.51391268e+00 -8.11287165e-02 -2.50049144e-01
1.14034700e+00 -5.16364396e-01 6.01660550e-01 -1.03474185e-01
-6.57670677e-01 5.73214054e-01 -4.88065444e-02 -3.83592755e-01
1.19946408e+00 -1.61580777e+00 9.53592956e-01 8.48927557e-01
-1.75087765e-01 9.08125520e-01 7.39885867e-01 -1.26902544e+00
-1.26770246e+00 -1.31833804e+00 1.36784625e+00 -7.92484164e-01
7.59664297e-01 -3.18981379e-01 -9.45748389e-01 4.10821736e-01
9.40714478e-01 -7.34819472e-01 9.59078670e-01 -4.89404686e-02
-1.05056815e-01 7.25864545e-02 -5.80943584e-01 8.63192439e-01
9.75960612e-01 -2.60416389e-01 -8.51894140e-01 6.12688482e-01
1.25128376e+00 -3.37377310e-01 -7.75357783e-01 4.75787044e-01
1.90621972e-01 -4.79786068e-01 1.00470352e+00 -7.92693615e-01
1.70313132e+00 -2.33481437e-01 3.24862778e-01 -1.78135037e+00
-2.45359316e-01 -7.02188730e-01 -6.71927989e-01 1.91258121e+00
6.56736851e-01 7.19395429e-02 3.68214518e-01 4.47390109e-01
-5.72937489e-01 -7.85596013e-01 -2.61662245e-01 -4.44580436e-01
2.93874949e-01 -2.06710175e-02 9.24611926e-01 7.82900453e-01
2.88861513e-01 1.04188061e+00 -4.24579263e-01 -3.29557657e-01
5.21020293e-01 3.47040534e-01 1.08561969e+00 -1.09909189e+00
1.78897887e-01 -7.78897405e-01 3.31859648e-01 -9.07906651e-01
7.08750606e-01 -1.16971672e+00 -1.27282858e-01 -2.44594646e+00
5.00823259e-01 -4.32330407e-02 1.52267262e-01 1.43012315e-01
-6.76933885e-01 -3.25564772e-01 -1.29985213e-01 1.67147741e-01
-1.14329636e+00 1.05204380e+00 1.51281393e+00 -4.94941682e-01
-1.13779731e-01 -1.08279280e-01 -1.57131219e+00 2.19359159e-01
3.60955775e-01 -5.18056750e-01 -6.84005320e-01 -4.42707002e-01
5.48035920e-01 1.64427951e-01 -8.49284083e-02 -6.40512884e-01
5.20713508e-01 -6.48701489e-02 4.49482083e-01 -1.45125711e+00
5.56605682e-02 1.87879533e-01 -1.47599623e-01 -9.73485187e-02
-1.16982424e+00 5.59034407e-01 1.21581435e-01 5.01474500e-01
-1.57610089e-01 -1.95573270e-01 2.70583387e-02 -2.86079437e-01
1.51023582e-01 1.07476883e-01 -3.91805232e-01 4.12289470e-01
4.99037474e-01 1.26940846e-01 -1.15035093e+00 -5.45430958e-01
5.81975691e-02 6.73556209e-01 3.94236684e-01 8.68708313e-01
6.85381472e-01 -1.34410954e+00 -1.35322487e+00 -3.52717489e-01
2.70126134e-01 4.37257558e-01 1.96380824e-01 2.76920617e-01
-1.80103965e-02 6.72988892e-01 2.40649730e-02 -2.25250065e-01
-9.39799428e-01 5.56533575e-01 -4.77648288e-01 -7.97701180e-01
-6.20769441e-01 6.03301108e-01 2.17568517e-01 -5.15863538e-01
1.65975049e-01 -4.97948736e-01 -5.70214570e-01 4.46318001e-01
9.76994634e-01 5.31689644e-01 1.66063368e-01 -3.59899849e-01
-1.42487735e-01 1.07543066e-01 -4.86263245e-01 -5.20265698e-02
1.54389858e+00 -2.29236320e-01 -4.08023983e-01 5.65883517e-01
1.07404244e+00 2.54640013e-01 -7.87272036e-01 -1.96746826e-01
1.33765161e-01 -5.51390909e-02 -7.25771859e-02 -8.72442544e-01
-7.21953034e-01 7.02343643e-01 -6.27817690e-01 2.24894986e-01
5.12190819e-01 6.98845163e-02 9.46903288e-01 5.26997685e-01
-3.24454337e-01 -1.12277901e+00 6.70229077e-01 5.37693620e-01
1.64350080e+00 -1.19794512e+00 6.32654250e-01 -1.47233710e-01
-1.16076303e+00 9.98349369e-01 9.50008035e-01 1.56956866e-01
2.95775324e-01 4.11537886e-01 -1.84085011e-01 -2.31136352e-01
-1.60204816e+00 3.57172430e-01 7.02399075e-01 6.96992815e-01
7.38810420e-01 -2.98693955e-01 -2.84169257e-01 1.14548862e+00
-5.96482098e-01 -2.78078467e-01 1.09652615e+00 6.21196330e-01
-2.22015217e-01 -9.27295327e-01 -1.49784610e-01 1.01786590e+00
-5.35897970e-01 -4.22263294e-01 -8.08607280e-01 3.48503858e-01
-6.53095305e-01 1.08925509e+00 5.29539995e-02 -2.07315475e-01
1.34108171e-01 -2.58920103e-01 -5.67134060e-02 -1.05103719e+00
-8.94580781e-01 -1.99599639e-01 3.26422364e-01 6.34476170e-02
-3.38706166e-01 -4.68400717e-01 -1.21253800e+00 -3.22670519e-01
-8.75946432e-02 3.89886111e-01 6.70525014e-01 7.72024930e-01
9.61959958e-01 1.07915592e+00 6.32522166e-01 -9.35978591e-01
-5.70297778e-01 -1.26445329e+00 -3.14735621e-01 2.48914838e-01
2.44285583e-01 -3.11977595e-01 -3.11002612e-01 2.01022848e-01] | [12.461299896240234, 9.537919044494629] |
c7429cac-e95a-424e-a16e-c34f287a8cc3 | rethinking-dimensionality-reduction-in-grid | 2209.09464 | null | https://arxiv.org/abs/2209.09464v4 | https://arxiv.org/pdf/2209.09464v4.pdf | Rethinking Dimensionality Reduction in Grid-based 3D Object Detection | Bird's eye view (BEV) is widely adopted by most of the current point cloud detectors due to the applicability of well-explored 2D detection techniques. However, existing methods obtain BEV features by simply collapsing voxel or point features along the height dimension, which causes the heavy loss of 3D spatial information. To alleviate the information loss, we propose a novel point cloud detection network based on a Multi-level feature dimensionality reduction strategy, called MDRNet. In MDRNet, the Spatial-aware Dimensionality Reduction (SDR) is designed to dynamically focus on the valuable parts of the object during voxel-to-BEV feature transformation. Furthermore, the Multi-level Spatial Residuals (MSR) is proposed to fuse the multi-level spatial information in the BEV feature maps. Extensive experiments on nuScenes show that the proposed method outperforms the state-of-the-art methods. The code will be available upon publication. | ['Zhiheng Li', 'Chengjie Wang', 'Yong liu', 'Qiang Nie', 'Kai Wu', 'Jianlin Liu', 'Jinli Liao', 'Yikang Ding', 'Ying Chen', 'Dihe Huang'] | 2022-09-20 | null | null | null | null | ['cloud-detection'] | ['computer-vision'] | [-3.01092751e-02 -6.08698368e-01 -3.92012149e-02 -1.84791982e-01
-2.26349562e-01 -4.47072059e-01 5.05179226e-01 1.45957604e-01
-2.38668591e-01 6.31247833e-02 -1.61308706e-01 -1.21439636e-01
-3.25245917e-01 -1.11014366e+00 -2.53445983e-01 -5.86348057e-01
2.09296778e-01 1.55346185e-01 9.03769195e-01 -2.75694937e-01
5.38232684e-01 1.21613503e+00 -1.79694879e+00 -8.09871033e-02
8.27966511e-01 1.06281340e+00 4.26662773e-01 1.80625603e-01
-3.74309927e-01 -1.26590967e-01 -7.71903917e-02 1.12240396e-01
5.49740613e-01 8.75862762e-02 -8.85057673e-02 7.53005873e-03
2.75478095e-01 -3.68242443e-01 -2.66125292e-01 1.16966045e+00
3.69172573e-01 2.31247872e-01 6.41391695e-01 -1.26092577e+00
-2.38648623e-01 -3.31589550e-01 -1.14613473e+00 2.23107949e-01
1.71935603e-01 -2.82624122e-02 6.53131366e-01 -1.42045105e+00
7.19087601e-01 1.25179493e+00 4.31675136e-01 1.22949980e-01
-9.51225400e-01 -8.02565813e-01 1.56905323e-01 1.18138380e-01
-1.82000446e+00 -1.03591576e-01 1.26799989e+00 -5.25394619e-01
8.24582160e-01 5.74040651e-01 7.22903430e-01 1.84586287e-01
6.81278631e-02 3.92850786e-01 1.00631714e+00 -3.76836896e-01
4.59476635e-02 1.08434655e-01 -3.10776141e-02 5.62538803e-01
4.17869925e-01 2.93332219e-01 -4.12269354e-01 -2.42965326e-01
9.76105213e-01 4.56595927e-01 -1.21873297e-01 -9.05392051e-01
-9.69053566e-01 9.04915869e-01 7.22993553e-01 1.64592594e-01
-3.28940958e-01 -2.04488382e-01 1.29512161e-01 -1.28109902e-01
6.67195082e-01 5.74781299e-02 -1.88736722e-01 9.48632956e-02
-8.82277429e-01 4.10481066e-01 2.05603018e-01 9.88906145e-01
9.12782669e-01 -2.17153862e-01 -3.36103626e-02 6.45178556e-01
5.87173998e-01 5.34402549e-01 1.81093961e-02 -7.66001105e-01
5.47454119e-01 1.27235329e+00 -2.87598744e-02 -1.41094840e+00
-4.66063499e-01 -2.68931717e-01 -8.20484579e-01 5.71630418e-01
-2.47592747e-01 3.42620045e-01 -8.96779656e-01 1.05751741e+00
8.30165386e-01 -1.87759716e-02 -2.75937110e-01 1.12005389e+00
1.02393341e+00 6.44135773e-01 -4.13892895e-01 -5.09176888e-02
1.19358242e+00 -5.05047083e-01 -2.29806736e-01 4.58216742e-02
3.10317636e-01 -7.73520291e-01 7.96957910e-01 1.22331232e-01
-9.21860278e-01 -4.16433692e-01 -1.10607445e+00 -1.13370843e-01
-4.62635070e-01 -2.85165850e-02 5.41484118e-01 4.57664698e-01
-6.43204927e-01 4.51176077e-01 -7.62921095e-01 -3.69491667e-01
4.74984944e-01 3.19894999e-01 -2.52952933e-01 -1.48073882e-01
-7.13798165e-01 4.13442314e-01 2.45790690e-01 1.06379740e-01
-4.53371614e-01 -7.20988035e-01 -6.51084721e-01 9.18357596e-02
4.53457385e-01 -6.02188826e-01 7.03562856e-01 -9.32376459e-02
-1.16999853e+00 7.15448022e-01 -3.40594321e-01 1.06546775e-01
4.98375863e-01 2.51522344e-02 -2.76237339e-01 2.50695497e-01
2.16455668e-01 5.89179933e-01 7.41103888e-01 -1.43249619e+00
-9.19126689e-01 -7.09947050e-01 -1.60450310e-01 4.50180769e-01
-3.17520231e-01 7.07810074e-02 -5.48291683e-01 -4.60551828e-01
8.63284469e-01 -6.78259790e-01 -1.64634183e-01 2.51854748e-01
-4.48570609e-01 -3.76599908e-01 1.24717748e+00 -2.65221179e-01
1.30131102e+00 -2.51424432e+00 4.54130918e-02 4.29632813e-01
5.44486940e-01 1.04908787e-01 -2.64899898e-02 4.49639380e-01
2.16199458e-03 4.35812436e-02 -1.15743689e-01 -1.22461855e-01
-2.84713179e-01 -1.17476657e-01 -6.78604022e-02 6.88776612e-01
3.81334350e-02 4.24904257e-01 -8.22232842e-01 -5.30786455e-01
7.02772200e-01 4.86976475e-01 -3.62267792e-01 -2.88321190e-02
8.39270353e-02 3.71051490e-01 -6.74923897e-01 7.81058908e-01
1.33449769e+00 1.44198081e-02 -5.90682089e-01 -2.98563004e-01
-7.72520542e-01 -3.19159031e-02 -1.35261285e+00 1.51321137e+00
-5.06646708e-02 2.53769547e-01 1.07180141e-01 -4.12958324e-01
1.21516311e+00 -1.06603883e-01 7.26162076e-01 -5.71306705e-01
1.34106696e-01 1.52184397e-01 -3.97825278e-02 -1.14004605e-01
6.03862166e-01 3.48438807e-02 2.38379955e-01 4.11583632e-02
-4.99637663e-01 -3.54605138e-01 -1.43538326e-01 1.04325287e-01
8.45050693e-01 -6.05250262e-02 3.97872567e-01 -1.72221333e-01
7.37679362e-01 2.38157347e-01 7.38862514e-01 3.58212769e-01
-1.01298064e-01 6.68321371e-01 1.91414073e-01 -3.24802786e-01
-9.74057198e-01 -1.03646660e+00 -3.77729237e-01 6.19314730e-01
5.85296452e-01 -4.38996643e-01 -4.93045270e-01 -4.65713739e-01
2.88674206e-01 5.89853525e-01 -2.54427731e-01 -3.74619216e-02
-4.99358088e-01 -4.51251090e-01 -7.66536444e-02 4.53691006e-01
6.04085088e-01 -6.01781785e-01 -7.06268549e-01 -1.24760596e-02
2.57360995e-01 -8.15006256e-01 -2.77126819e-01 -1.56381264e-01
-1.12427926e+00 -9.25300956e-01 -5.20963967e-01 -4.42210346e-01
6.57157481e-01 1.10136259e+00 6.57261431e-01 2.61276841e-01
-2.55466253e-01 3.57012637e-02 -4.74436104e-01 -3.82744342e-01
5.17842211e-02 9.41252410e-02 1.53109282e-01 -1.87998831e-01
7.36093879e-01 -5.63581407e-01 -8.45668137e-01 4.13484186e-01
-7.67384648e-01 2.03735352e-01 4.26131696e-01 5.00174820e-01
1.02503014e+00 3.50688159e-01 -4.98655923e-02 -3.53671491e-01
2.73208767e-01 -3.05329651e-01 -9.97265577e-01 -9.14519578e-02
-6.18887842e-01 -2.44078830e-01 2.27638677e-01 -6.43096045e-02
-8.60359073e-01 2.26717368e-01 5.53272814e-02 -7.05217719e-01
-2.30704501e-01 1.22277707e-01 -3.72228146e-01 -4.89257842e-01
2.57502973e-01 3.27741623e-01 -2.60309696e-01 -7.51394868e-01
3.64590228e-01 8.22230816e-01 2.72339553e-01 -1.49695247e-01
1.15269506e+00 9.24000621e-01 3.99727464e-01 -9.36735630e-01
-3.77504289e-01 -9.49612737e-01 -9.14127946e-01 -2.67332256e-01
8.96256387e-01 -8.22555780e-01 -6.92035437e-01 2.97884464e-01
-1.29754162e+00 4.28921431e-01 -2.02958465e-01 5.12390673e-01
-1.70078412e-01 3.97677302e-01 -1.99341878e-01 -8.04112315e-01
-3.40045244e-01 -1.17093098e+00 1.24268818e+00 3.63354027e-01
2.17528537e-01 -4.67928827e-01 7.36340806e-02 2.93807480e-02
7.55621642e-02 2.66093016e-01 9.80491996e-01 -3.21821243e-01
-9.88518834e-01 -1.58924088e-01 -6.58262849e-01 1.53016791e-01
1.54529527e-01 1.55255422e-01 -7.72697866e-01 -2.55209327e-01
1.57494336e-01 5.00934422e-01 7.59733677e-01 3.10662895e-01
1.16601181e+00 2.77011573e-01 -6.87345326e-01 8.48742664e-01
1.56952894e+00 1.72428653e-01 4.11413372e-01 4.40020710e-01
1.01579142e+00 4.86207664e-01 8.49063575e-01 5.91022193e-01
3.13578069e-01 7.54070342e-01 7.74723291e-01 -2.13913500e-01
-5.31861670e-02 -2.40910381e-01 -1.65121004e-01 8.55843127e-01
-2.56998509e-01 1.68837924e-02 -9.55938101e-01 3.59197199e-01
-1.63339233e+00 -6.11152232e-01 -5.31174839e-01 2.30321503e+00
6.00478873e-02 1.89741880e-01 -6.06473312e-02 5.70118837e-02
8.11786294e-01 2.22559884e-01 -6.13148928e-01 -1.06299132e-01
1.14180315e-02 -8.94793049e-02 8.91645730e-01 1.36539713e-01
-1.00905442e+00 8.64356458e-01 5.55703163e+00 9.45524454e-01
-9.78470862e-01 1.68185815e-01 -4.61729057e-02 -7.42705986e-02
-2.72284836e-01 1.04845554e-01 -9.95694637e-01 3.64627570e-01
7.76463076e-02 -1.03278346e-01 3.48651141e-01 1.06120050e+00
4.66985226e-01 -2.52815038e-01 -7.77068377e-01 1.27468169e+00
1.10199237e-02 -1.10046840e+00 4.05876301e-02 3.12763304e-01
4.69229907e-01 1.82620063e-01 -1.03236996e-01 -1.35121211e-01
-5.87576032e-02 -5.58209479e-01 5.46735346e-01 4.55297261e-01
9.22357261e-01 -9.42231596e-01 4.47837681e-01 5.14776647e-01
-1.74755824e+00 -6.71129152e-02 -7.41697788e-01 3.75230089e-02
1.98252410e-01 6.79138839e-01 -5.57721198e-01 7.60326922e-01
8.41010928e-01 7.14955270e-01 -8.28673840e-01 1.22381973e+00
-4.33613285e-02 7.57239312e-02 -5.82011402e-01 3.16611491e-02
2.98925340e-01 -4.72979724e-01 9.85183060e-01 7.48271525e-01
6.15430415e-01 1.16540439e-01 1.60382792e-01 1.01905525e+00
1.95612997e-01 2.70468801e-01 -7.38024175e-01 4.10893977e-01
6.99210942e-01 1.36203086e+00 -9.63578641e-01 -9.11506116e-02
-6.42402053e-01 7.73033500e-01 9.65133384e-02 -5.06187184e-03
-4.14376676e-01 -5.58803201e-01 6.86279237e-01 5.24599552e-01
5.31692863e-01 -3.22113335e-01 -4.15973842e-01 -8.98181796e-01
1.71031073e-01 -3.20803642e-01 9.15718451e-02 -8.94422412e-01
-1.19366491e+00 3.70016694e-01 3.24399054e-01 -1.69109881e+00
1.60060674e-01 -5.44688106e-01 -6.07994854e-01 9.61258590e-01
-1.57324064e+00 -1.01611149e+00 -6.03196740e-01 4.50188518e-01
5.06173790e-01 -4.29598838e-02 4.47833151e-01 2.29697704e-01
-3.90480638e-01 1.13575958e-01 2.86239207e-01 -1.98984519e-01
3.90290111e-01 -8.55399072e-01 4.43289548e-01 8.21092427e-01
-1.34951606e-01 4.82688397e-01 3.71277243e-01 -9.90469873e-01
-1.36049247e+00 -1.04645765e+00 5.57713389e-01 -2.50774950e-01
2.88804382e-01 -4.42744464e-01 -1.07878089e+00 2.00263217e-01
-2.59370446e-01 1.50060192e-01 2.78703511e-01 -2.73166627e-01
-1.87632173e-01 -2.18007594e-01 -1.22696102e+00 4.76819634e-01
1.15787566e+00 -2.25831792e-01 -5.29395163e-01 1.77421615e-01
7.90655792e-01 -5.09736955e-01 -7.45580375e-01 7.39868581e-01
4.76098835e-01 -1.13636351e+00 1.10310340e+00 -6.66058511e-02
8.47014189e-02 -7.92521238e-01 -2.87895232e-01 -1.11919653e+00
-7.49258935e-01 -5.60632832e-02 -2.20523626e-02 1.24996543e+00
-1.81179568e-01 -5.66870987e-01 9.47329998e-01 2.88274765e-01
-1.27176180e-01 -6.93669319e-01 -1.23929632e+00 -8.63865376e-01
-1.92938313e-01 -3.49295288e-01 1.00685954e+00 7.18754888e-01
-4.79851276e-01 2.66950607e-01 1.26765981e-01 4.45883632e-01
6.95412219e-01 3.27563584e-01 8.39365721e-01 -1.71843827e+00
2.53661633e-01 -5.31856298e-01 -7.15861380e-01 -9.39849198e-01
-3.22705567e-01 -8.39401066e-01 -1.11476682e-01 -1.58863139e+00
2.79874444e-01 -5.82546234e-01 -1.91460729e-01 -3.19997966e-02
-7.96532333e-02 -2.53405068e-02 3.39201748e-01 4.26462829e-01
-3.13413233e-01 5.89810669e-01 1.32316267e+00 2.00002626e-01
-5.03844976e-01 -9.17174667e-02 -3.60682547e-01 9.05367017e-01
6.72109485e-01 -6.07182741e-01 -2.71290064e-01 -3.44689101e-01
3.66549902e-02 -9.73688513e-02 4.92663801e-01 -1.13408148e+00
2.71750152e-01 -3.54330570e-01 5.68377852e-01 -1.67742324e+00
5.51398456e-01 -1.14667654e+00 1.55107334e-01 2.88544565e-01
3.53958488e-01 2.18365461e-01 3.94521318e-02 5.96187830e-01
-1.10329181e-01 -1.57638267e-01 7.84259737e-01 -2.76286662e-01
-7.30547547e-01 5.50963819e-01 6.53315783e-02 -4.46948707e-01
1.25178051e+00 -5.34116030e-01 -1.35807678e-01 1.12427376e-01
-1.89779952e-01 2.58208573e-01 1.00234294e+00 3.38866711e-01
1.09196305e+00 -1.54284370e+00 -5.18283963e-01 4.49110776e-01
4.53081578e-01 4.41603184e-01 4.07601207e-01 7.31228352e-01
-6.03978395e-01 4.27903414e-01 3.59610119e-03 -8.71557415e-01
-1.41392267e+00 7.60565698e-01 8.31862465e-02 3.01986467e-02
-9.91192698e-01 5.79562664e-01 3.32667410e-01 -3.37812573e-01
-5.63053899e-02 -1.89334512e-01 -2.78728515e-01 -7.58014172e-02
3.93596083e-01 7.39518106e-01 1.28644899e-01 -9.24371839e-01
-6.22542322e-01 1.02238345e+00 -7.02457726e-02 7.65612423e-02
1.29016340e+00 -1.83163151e-01 -2.73904413e-01 2.78200209e-01
1.09847522e+00 1.17607020e-01 -1.07286859e+00 -2.85340786e-01
-2.42253214e-01 -9.38203335e-01 4.23943996e-01 -1.35910213e-01
-9.77062225e-01 9.90227401e-01 1.04759789e+00 5.73122650e-02
1.07531393e+00 1.93760805e-02 7.46811450e-01 2.08822682e-01
4.54228789e-01 -1.00042045e+00 -3.25486749e-01 3.34367484e-01
7.99169481e-01 -1.09892380e+00 4.67664897e-01 -1.01960266e+00
-2.51709580e-01 1.10017228e+00 8.13371241e-01 -2.55498230e-01
9.07167912e-01 6.28668070e-02 -3.24739814e-01 -5.87186992e-01
-1.96019381e-01 -1.37975618e-01 3.17121327e-01 5.55335343e-01
-1.06942184e-01 -1.34666944e-02 -3.97749126e-01 2.84161419e-01
-3.84247527e-02 -1.34915233e-01 1.98392302e-01 1.02788854e+00
-6.64873540e-01 -9.75641072e-01 -8.02559912e-01 6.36769712e-01
1.81686327e-01 -6.15012571e-02 -2.76550442e-01 9.29202735e-01
2.93160766e-01 7.18837321e-01 3.43896002e-01 -5.88189900e-01
6.12478733e-01 -3.22228521e-01 2.77721941e-01 -7.36793935e-01
-1.49231613e-01 2.69024700e-01 -5.10200143e-01 -6.55084014e-01
-3.77829194e-01 -7.82506645e-01 -1.27672112e+00 -2.96867609e-01
-5.05766034e-01 -1.81777358e-01 8.42035413e-01 5.73812187e-01
4.74990755e-01 2.59569585e-01 9.25577641e-01 -1.07453859e+00
-2.30985031e-01 -7.48752773e-01 -7.70752847e-01 2.08929688e-01
2.00144827e-01 -1.14535069e+00 -4.49096590e-01 -5.17325580e-01] | [7.8930983543396, -2.9453837871551514] |
32bb49ed-9526-4e37-b0f1-6773bd105131 | thematic-cohesion-measuring-terms | null | null | https://aclanthology.org/L14-1742 | https://aclanthology.org/L14-1742.pdf | Thematic Cohesion: measuring terms discriminatory power toward themes | We present a new measure of thematic cohesion. This measure associates each term with a weight representing its discriminatory power toward a theme, this theme being itself expressed by a list of terms (a thematic lexicon). This thematic cohesion criterion can be used in many applications, such as query expansion, computer-assisted translation, or iterative construction of domain-specific lexicons and corpora. The measure is computed in two steps. First, a set of documents related to the terms is gathered from the Web by querying a Web search engine. Then, we produce an oriented co-occurrence graph, where vertices are the terms and edges represent the fact that two terms co-occur in a document. This graph can be interpreted as a recommendation graph, where two terms occurring in a same document means that they recommend each other. This leads to using a random walk algorithm that assigns a global importance value to each vertex of the graph. After observing the impact of various parameters on those importance values, we evaluate their correlation with retrieval effectiveness. | ['Claude de Loupy', "Cl{\\'e}ment de Groc", 'Xavier Tannier'] | 2014-05-01 | null | null | null | lrec-2014-5 | ['text-clustering'] | ['natural-language-processing'] | [ 2.79866815e-01 1.61529317e-01 -4.16424990e-01 -1.46533445e-01
-4.53612536e-01 -6.62784338e-01 9.27454472e-01 9.47770834e-01
-4.09084052e-01 3.61243188e-01 5.74048281e-01 -2.03847051e-01
-7.88059771e-01 -1.06691408e+00 -2.26736054e-01 -7.18809545e-01
-2.43663222e-01 7.09921718e-01 3.28437328e-01 -3.70212108e-01
5.60721755e-01 2.48959467e-01 -1.62551248e+00 2.34928846e-01
7.22829103e-01 7.57779717e-01 2.97682047e-01 4.66863550e-02
-4.92201298e-01 2.92464703e-01 -6.24971926e-01 -3.65271628e-01
-1.50108054e-01 -5.33386230e-01 -9.62744772e-01 7.63509199e-02
-5.11595421e-02 4.83564109e-01 3.25454414e-01 1.14814246e+00
-1.18218139e-01 2.20292777e-01 1.05868757e+00 -9.00797307e-01
-1.97615042e-01 8.85175586e-01 -4.63268131e-01 1.21688060e-01
6.64759457e-01 -7.20296323e-01 1.53815913e+00 -7.03909695e-01
9.21098650e-01 1.17255163e+00 1.32280275e-01 -1.73386514e-01
-1.17419088e+00 -4.73332033e-02 6.86445534e-02 9.83829424e-02
-1.28071773e+00 2.34771237e-01 7.48100877e-01 -5.96738458e-01
6.64785326e-01 5.78329146e-01 8.15770745e-01 4.99799043e-01
1.45304397e-01 9.29824412e-02 7.10613608e-01 -7.54478514e-01
3.51702809e-01 3.65909904e-01 4.30893272e-01 3.79331857e-01
4.45887774e-01 -5.54893017e-01 -5.76866448e-01 -5.83209395e-01
1.08790696e-01 -2.35160530e-01 -3.08264524e-01 -6.76764846e-01
-9.11215544e-01 8.62140298e-01 4.10668701e-01 9.84743595e-01
-5.98772705e-01 -5.44241443e-02 4.06554341e-01 5.15589535e-01
4.26394880e-01 4.42275017e-01 1.13114491e-02 2.36943409e-01
-5.35037756e-01 2.56350994e-01 8.93192828e-01 5.17897427e-01
8.34330738e-01 -8.74995351e-01 1.80034712e-02 8.10472548e-01
6.34009719e-01 1.07465044e-01 6.62317336e-01 -4.21003968e-01
2.51594007e-01 1.12155676e+00 3.50789838e-02 -1.58152473e+00
-3.74225020e-01 -3.34087968e-01 -3.87300372e-01 -1.09336473e-01
1.54097423e-01 3.03298026e-01 -2.49635845e-01 1.72777832e+00
4.77785677e-01 -3.38157386e-01 -2.17848434e-03 6.89694703e-01
5.73956907e-01 6.47797525e-01 1.99926332e-01 -6.18858159e-01
1.68308604e+00 -4.07262176e-01 -6.53936028e-01 1.17333189e-01
5.39661527e-01 -9.22080696e-01 8.49847674e-01 3.74540150e-01
-7.72621989e-01 -1.68838337e-01 -8.80100012e-01 1.90317973e-01
-7.10586369e-01 -1.76257715e-01 2.54599363e-01 3.95247877e-01
-1.04188180e+00 5.28196514e-01 -3.19740921e-01 -8.02436829e-01
-4.33026820e-01 1.10116087e-01 -2.55951971e-01 1.76202133e-01
-1.23176801e+00 7.86962569e-01 6.12297952e-01 -3.96222949e-01
-4.86510843e-01 -7.63323754e-02 -5.22260666e-01 3.35177034e-01
4.38291550e-01 -6.41334832e-01 6.64428532e-01 -8.51513863e-01
-7.87796319e-01 1.13135576e+00 -2.05790535e-01 -2.80794621e-01
-3.40426676e-02 1.32011324e-01 -7.30036914e-01 1.19067490e-01
2.37560034e-01 -4.87250984e-02 7.05600560e-01 -1.40862978e+00
-8.42299938e-01 -5.39263606e-01 2.53061682e-01 5.05153954e-01
-7.65372455e-01 2.74276346e-01 -8.81509423e-01 -5.15222490e-01
5.10189295e-01 -7.83398569e-01 7.34458044e-02 -5.14166951e-01
-3.48040789e-01 -6.78097546e-01 3.86347294e-01 -2.33280852e-01
1.74042094e+00 -1.87801886e+00 5.86066246e-01 9.83058393e-01
3.41553360e-01 -2.25262582e-01 1.09319463e-01 9.13461745e-01
6.59651160e-02 2.53282070e-01 -7.44463177e-03 1.80147886e-01
1.73296416e-04 -2.43991539e-02 -1.15604877e-01 1.57083675e-01
-3.52307826e-01 2.80097723e-01 -1.12816930e+00 -6.44813061e-01
-1.02992512e-01 3.86334300e-01 -3.18374664e-01 -1.64252520e-01
-4.31749165e-01 -1.26931146e-01 -9.06168938e-01 1.30591765e-01
9.63986516e-02 -3.77355844e-01 8.77286971e-01 -2.86897123e-01
-1.72735929e-01 5.25339365e-01 -1.13526392e+00 1.24814177e+00
-3.67242485e-01 3.35047483e-01 -4.17972296e-01 -9.68371868e-01
1.13663030e+00 2.62480915e-01 6.04621410e-01 -7.00309992e-01
1.97081625e-01 3.75672728e-01 -2.32999653e-01 -4.65485007e-01
8.57873976e-01 1.73106983e-01 -2.12571416e-02 6.88677490e-01
-1.18170112e-01 1.77238613e-01 7.19251573e-01 4.41718221e-01
9.26975787e-01 -3.29693288e-01 3.99377495e-01 -6.95019901e-01
8.62082064e-01 -9.40168947e-02 1.21240050e-03 4.83936489e-01
5.65420270e-01 3.96767333e-02 7.99195290e-01 -2.86905468e-01
-9.08880115e-01 -7.84185648e-01 -2.03126788e-01 1.18381953e+00
3.26344281e-01 -9.73481119e-01 -5.78519881e-01 -3.09237391e-01
-1.55239133e-03 5.57930887e-01 -8.66204798e-01 -2.21105039e-01
-2.38881767e-01 -4.63348210e-01 -1.34497210e-01 -1.78969398e-01
-6.10919185e-02 -9.45462108e-01 -5.62776387e-01 1.23276047e-01
-3.20599139e-01 -4.25958663e-01 -2.17389017e-01 1.03007928e-01
-8.10442030e-01 -9.77086961e-01 -4.78559464e-01 -9.32122409e-01
5.76812744e-01 3.60085845e-01 1.46360373e+00 3.83586198e-01
1.46142453e-01 2.11548805e-01 -7.61713088e-01 2.07379751e-04
-4.60287631e-01 2.71428645e-01 1.17338903e-01 1.12939648e-01
4.68896836e-01 -5.43948770e-01 -4.64617372e-01 4.74392235e-01
-1.14418960e+00 -3.70858669e-01 5.64501137e-02 4.36798006e-01
7.34985054e-01 2.94001251e-01 3.35343748e-01 -9.23792839e-01
1.22483635e+00 -7.23031044e-01 -6.12524450e-01 7.01594532e-01
-8.06341112e-01 2.66825944e-01 1.44954383e-01 -2.36596242e-01
-6.35615051e-01 -3.55418265e-01 1.67305812e-01 2.89520264e-01
3.10091853e-01 1.27125072e+00 2.25997791e-02 2.71927118e-01
7.81895578e-01 9.14808363e-02 -3.73770833e-01 -4.18129593e-01
5.35599172e-01 6.73869431e-01 1.19149230e-01 -5.00168502e-01
4.41938788e-01 2.78046042e-01 1.04922801e-01 -6.74647033e-01
-6.41459882e-01 -7.57064342e-01 -2.49849096e-01 -4.16443288e-01
7.47582138e-01 -3.35344493e-01 -6.38338029e-01 -2.14466617e-01
-8.00861716e-01 5.08997917e-01 -1.89409479e-01 5.22629738e-01
-1.01275183e-01 2.89183676e-01 -7.02862814e-02 -6.90074921e-01
-3.18985313e-01 -7.81338692e-01 6.20646179e-01 1.95848092e-01
-5.22199452e-01 -1.07792008e+00 5.19142091e-01 1.46189049e-01
1.40812665e-01 -6.93314243e-04 1.43849456e+00 -9.36730504e-01
-5.50244041e-02 -3.40191811e-01 6.02305457e-02 -1.43424541e-01
2.11267203e-01 4.14832905e-02 -4.20631111e-01 -2.30789751e-01
-6.52091354e-02 1.54843135e-02 6.23982966e-01 1.98994055e-01
5.92410862e-01 -2.98160136e-01 -7.11547792e-01 -3.50849032e-02
1.46157968e+00 5.12189806e-01 4.51261699e-01 6.92148149e-01
4.14078027e-01 8.97879660e-01 6.10924125e-01 3.36120158e-01
1.14806361e-01 1.09980989e+00 2.60228693e-01 2.54850149e-01
3.51240456e-01 -2.08270952e-01 -4.38596979e-02 1.04153168e+00
-1.14603885e-01 -6.41959369e-01 -9.35117066e-01 5.85271895e-01
-1.77690721e+00 -8.41050267e-01 -3.41528356e-01 2.47719193e+00
7.44813502e-01 3.00542265e-01 2.77266383e-01 1.53640166e-01
8.20306957e-01 8.62167776e-02 -5.40371425e-02 -3.76907259e-01
9.78693590e-02 -2.08212346e-01 9.36815068e-02 7.45982468e-01
-6.19234502e-01 7.83473432e-01 6.01162434e+00 8.72278988e-01
-8.51326346e-01 -2.28102475e-01 3.89786690e-01 1.52173191e-01
-8.88111353e-01 1.40376300e-01 -5.69543660e-01 3.24570894e-01
6.30201399e-01 -7.89851248e-01 1.22094199e-01 6.36862218e-01
2.20727902e-02 -3.04908037e-01 -7.15538085e-01 4.52688098e-01
2.82085299e-01 -9.45051432e-01 4.28620577e-01 2.27922231e-01
5.33627808e-01 -3.38088155e-01 -4.47531343e-02 -2.08308756e-01
2.00704351e-01 -4.98667836e-01 3.49622160e-01 5.18073559e-01
4.30623889e-01 -8.48510265e-01 5.36706984e-01 1.59029827e-01
-1.34272134e+00 2.57420301e-01 -3.56500715e-01 3.34638327e-01
1.34983376e-01 8.10978889e-01 -6.31244123e-01 8.36129963e-01
6.45279586e-01 2.43183360e-01 -2.74856776e-01 9.71791387e-01
-1.89775988e-01 3.81067604e-01 -3.12956125e-01 -5.07830858e-01
2.84909785e-01 -7.40084231e-01 7.77665257e-01 9.49593186e-01
4.36995894e-01 -1.51575983e-01 -1.91732589e-02 5.84251225e-01
-1.45153850e-01 8.42354894e-01 -6.77458167e-01 -1.29227698e-01
5.83546638e-01 1.40316844e+00 -1.18598473e+00 -3.91854912e-01
-3.54247004e-01 6.52732074e-01 2.97200143e-01 1.16603345e-01
-4.44637835e-01 -4.08718795e-01 3.85783404e-01 2.28199154e-01
9.08980668e-02 6.02002256e-02 1.49003476e-01 -7.04236627e-01
-3.28236236e-03 -7.27684200e-01 5.10457635e-01 -6.50139749e-01
-9.12274063e-01 1.02443385e+00 1.39633015e-01 -1.29193294e+00
-4.61810172e-01 -2.19235510e-01 -2.41476402e-01 8.90368879e-01
-1.01832771e+00 -4.59052503e-01 -2.20861509e-01 7.12606251e-01
6.28052726e-02 -2.47169942e-01 1.01310158e+00 2.26707980e-01
-4.25916836e-02 1.32013634e-01 3.25255454e-01 -2.91275352e-01
4.88715529e-01 -1.11784256e+00 1.39147922e-01 4.95582968e-01
5.62847972e-01 9.13116097e-01 8.78210306e-01 -7.74924695e-01
-8.47111404e-01 -4.89137322e-01 1.66525364e+00 -2.37564132e-01
9.11197126e-01 2.91562779e-03 -8.74753833e-01 3.41316849e-01
2.84210980e-01 -6.13417447e-01 8.31168771e-01 6.15063846e-01
-3.28763247e-01 -1.12902395e-01 -8.01646829e-01 6.17895067e-01
9.12426651e-01 -4.35330600e-01 -7.44845212e-01 5.55665314e-01
7.23048866e-01 9.19209644e-02 -8.81136119e-01 -1.10165309e-02
4.95730668e-01 -8.60093355e-01 8.48082900e-01 -5.45636296e-01
3.99380803e-01 -3.33150893e-01 2.33422965e-02 -1.44783795e+00
-5.85997343e-01 -3.35564107e-01 1.92667618e-01 1.17205775e+00
6.78177178e-01 -3.52355242e-01 3.39503050e-01 1.31781384e-01
2.32286453e-01 -5.04572451e-01 -7.97357440e-01 -6.27973795e-01
-2.61437148e-01 -3.57739553e-02 6.30203187e-01 1.08479857e+00
6.76901996e-01 7.00364709e-01 -1.41621903e-01 -2.03663439e-01
3.45722914e-01 3.96775812e-01 2.13400304e-01 -1.70771396e+00
-1.25440165e-01 -5.63976467e-01 -4.10182804e-01 -5.45201838e-01
-2.35682562e-01 -1.13667691e+00 -2.44225636e-01 -1.65845561e+00
2.54872859e-01 -4.91047591e-01 -5.19412637e-01 1.70960426e-01
2.22859439e-02 -2.06262022e-02 -1.27720967e-01 5.93849242e-01
-7.09345222e-01 1.27952591e-01 8.17068219e-01 -3.49620283e-02
-3.22350144e-01 -1.25441715e-01 -7.38387167e-01 7.44194031e-01
5.88558018e-01 -7.55105555e-01 -5.37220597e-01 -1.63657278e-01
1.00750363e+00 1.30297288e-01 -1.98293954e-01 -7.65325904e-01
2.36322030e-01 -1.34347090e-02 -3.28811646e-01 -1.86385497e-01
-7.66658336e-02 -1.12337601e+00 5.68088055e-01 5.46322763e-01
-5.25074780e-01 2.58132935e-01 -1.86868086e-01 5.07448673e-01
-4.74972010e-01 -7.90122092e-01 2.49710530e-01 6.30915239e-02
-4.63242054e-01 -1.78452998e-01 -4.29482996e-01 -1.56560898e-01
8.25058222e-01 -4.55886610e-02 -6.59447303e-03 -4.01938647e-01
-7.91240394e-01 1.36028707e-01 3.55164707e-01 5.82553387e-01
3.14276457e-01 -1.42556822e+00 -5.74012995e-01 -2.76966184e-01
6.21518910e-01 -5.88887513e-01 -1.48431987e-01 5.44151783e-01
-3.82940024e-01 3.50144565e-01 -6.73205266e-03 -4.46146399e-01
-1.57731521e+00 6.27970576e-01 1.06515025e-03 -5.40544152e-01
-4.46438521e-01 5.49141586e-01 1.10281387e-03 1.11728236e-01
1.76877111e-01 -6.46142289e-02 -1.07545948e+00 6.21372581e-01
5.69315612e-01 2.04776347e-01 2.04477981e-01 -7.28605688e-01
-4.70142215e-01 8.82479429e-01 -1.81564778e-01 -3.56667906e-01
1.18079007e+00 -1.58515617e-01 -5.93175352e-01 5.98404586e-01
1.07805431e+00 1.88322634e-01 -1.82573810e-01 -4.26146418e-01
4.86999989e-01 -3.44423711e-01 3.08885355e-03 -6.23877823e-01
-9.21110272e-01 2.77597550e-02 3.70373160e-01 1.15660036e+00
1.09016824e+00 3.35963041e-01 2.99035944e-02 5.27137220e-01
3.72250736e-01 -1.09061778e+00 1.74527057e-02 5.44438660e-01
8.51807594e-01 -6.19472325e-01 -3.08707934e-02 -5.51171422e-01
-5.20043671e-01 1.04855692e+00 -1.40897095e-01 9.49059203e-02
6.99793577e-01 -5.71260639e-02 -8.74582827e-02 -6.86743021e-01
-5.53749323e-01 -3.12094301e-01 7.17930615e-01 1.48689002e-01
7.44289339e-01 -1.74738877e-02 -1.46218061e+00 1.31298929e-01
-2.48973206e-01 -1.31626949e-01 2.48691235e-02 6.83040440e-01
-8.15235078e-01 -1.40736270e+00 -2.17105061e-01 4.17230904e-01
-2.06999868e-01 -1.66464567e-01 -7.15085506e-01 3.74272913e-01
-8.14509299e-03 1.00629151e+00 2.76786685e-01 -4.95358795e-01
3.17633688e-01 8.94728079e-02 1.38566047e-01 -5.27385116e-01
-6.78592920e-01 2.98636854e-01 2.44038045e-01 -2.42747575e-01
-5.97993672e-01 -5.35167098e-01 -1.18080890e+00 -2.05655515e-01
-4.80060130e-01 7.63319433e-01 7.52525508e-01 7.70238519e-01
1.32125422e-01 4.98172283e-01 6.08798742e-01 -2.07046703e-01
5.36510535e-02 -8.48702371e-01 -7.83450484e-01 7.31975615e-01
-2.48913258e-01 -7.07931697e-01 -4.87555236e-01 -5.45383543e-02] | [10.160544395446777, 8.384496688842773] |
56efaebd-5eed-4706-97c4-8480800581a0 | distilling-style-from-image-pairs-for-global | 2209.15165 | null | https://arxiv.org/abs/2209.15165v2 | https://arxiv.org/pdf/2209.15165v2.pdf | Distilling Style from Image Pairs for Global Forward and Inverse Tone Mapping | Many image enhancement or editing operations, such as forward and inverse tone mapping or color grading, do not have a unique solution, but instead a range of solutions, each representing a different style. Despite this, existing learning-based methods attempt to learn a unique mapping, disregarding this style. In this work, we show that information about the style can be distilled from collections of image pairs and encoded into a 2- or 3-dimensional vector. This gives us not only an efficient representation but also an interpretable latent space for editing the image style. We represent the global color mapping between a pair of images as a custom normalizing flow, conditioned on a polynomial basis of the pixel color. We show that such a network is more effective than PCA or VAE at encoding image style in low-dimensional space and lets us obtain an accuracy close to 40 dB, which is about 7-10 dB improvement over the state-of-the-art methods. | ['Rafal K. Mantiuk', 'Param Hanji', 'Aamir Mustafa'] | 2022-09-30 | null | null | null | null | ['tone-mapping', 'inverse-tone-mapping'] | ['computer-vision', 'computer-vision'] | [ 5.74396610e-01 -1.62484705e-01 8.67254974e-04 -5.38714468e-01
-4.97754604e-01 -7.16207504e-01 5.74131191e-01 -4.06485319e-01
-2.79077470e-01 6.58908784e-01 1.89079598e-01 -5.05770259e-02
4.51306626e-02 -7.67348945e-01 -6.46649957e-01 -7.82348156e-01
2.66055793e-01 2.24320009e-01 -1.39117718e-01 -2.57174134e-01
2.57645488e-01 5.61272204e-01 -1.35999918e+00 2.86374986e-01
6.72786117e-01 1.03806221e+00 9.37329307e-02 9.42200899e-01
-3.98390293e-01 6.89762294e-01 -5.70040941e-01 -5.84921241e-01
4.91617799e-01 -7.86431670e-01 -7.98100591e-01 1.49523094e-01
9.05481815e-01 -3.82116377e-01 -4.20947373e-01 1.36896598e+00
3.02711040e-01 -1.63223088e-01 8.10311675e-01 -1.09119391e+00
-1.13318765e+00 3.67992908e-01 -5.39179683e-01 -2.77016044e-01
2.49918789e-01 -9.09475759e-02 1.03208637e+00 -6.33910477e-01
7.65854657e-01 1.30307603e+00 3.92067492e-01 6.23136938e-01
-1.70838344e+00 -3.87112558e-01 -7.89286196e-02 2.66138226e-01
-1.14965522e+00 -2.76242733e-01 1.19138455e+00 -4.48389739e-01
2.33215898e-01 5.54768920e-01 6.79095149e-01 1.09710872e+00
1.97546661e-01 5.33671975e-01 1.69887877e+00 -7.69751847e-01
8.33725706e-02 4.77264613e-01 -7.83117712e-02 8.00601125e-01
-9.63564515e-02 8.42765272e-02 -5.18589497e-01 2.70923618e-02
1.06159735e+00 -2.70981550e-01 -3.50186259e-01 -7.42129624e-01
-1.30664849e+00 8.94888103e-01 4.70517844e-01 1.27513573e-01
-1.77447796e-01 2.72053301e-01 3.53326350e-02 5.22596657e-01
2.93913275e-01 7.25236714e-01 -2.06449702e-01 -1.26672879e-01
-8.26978564e-01 -3.92039493e-02 7.85659134e-01 7.17635453e-01
1.11053848e+00 2.78898269e-01 -1.76543966e-01 8.83784831e-01
1.33439219e-02 5.97913742e-01 2.88090825e-01 -1.35558105e+00
1.52445897e-01 2.27051571e-01 8.00056383e-02 -1.24427581e+00
-6.18444569e-02 -2.02376217e-01 -1.01779211e+00 7.69083858e-01
5.87565601e-01 -2.27133289e-01 -8.91474307e-01 2.19287157e+00
-1.57603532e-01 9.71348137e-02 -9.09328237e-02 7.04128861e-01
1.14549026e-01 8.40434909e-01 -2.95210719e-01 -1.29244700e-01
1.26992142e+00 -8.52972746e-01 -9.00698483e-01 -2.47956976e-01
4.73498031e-02 -1.04282212e+00 1.33703268e+00 5.85335791e-01
-1.22242057e+00 -6.31096184e-01 -1.22468829e+00 -3.66017610e-01
-3.93811703e-01 3.83326054e-01 6.33081019e-01 9.68915462e-01
-1.40732062e+00 6.99948788e-01 -3.66594970e-01 -2.43761241e-01
6.68007582e-02 1.68735668e-01 -4.62229162e-01 1.08244345e-02
-9.08429563e-01 9.31143165e-01 -1.69446096e-01 1.14691220e-01
-4.73793477e-01 -7.75298595e-01 -8.07145834e-01 4.03914489e-02
4.96201850e-02 -8.47396851e-01 7.83041120e-01 -1.23357213e+00
-1.98676860e+00 9.47804809e-01 -1.53006643e-01 -4.80411090e-02
6.00397229e-01 -1.20761745e-01 -5.42902827e-01 1.22956164e-01
-1.45745546e-01 8.34696054e-01 1.34892631e+00 -1.55981255e+00
-3.50134224e-01 -2.13579118e-01 5.06972671e-02 1.57533124e-01
-8.08726907e-01 1.15295939e-01 -6.07679546e-01 -9.44095254e-01
1.04828477e-01 -1.07852018e+00 -1.36474222e-01 6.73663795e-01
-3.65615129e-01 6.71131551e-01 8.75141919e-01 -7.80032158e-01
1.20351887e+00 -2.14047074e+00 6.83764637e-01 2.83551812e-01
1.88509986e-01 4.72539514e-02 -4.75486666e-01 4.70005609e-02
-2.93604493e-01 1.04168117e-01 -3.87925208e-01 -5.10065675e-01
1.19393170e-01 3.47290814e-01 -3.37664962e-01 4.02181715e-01
3.80687833e-01 5.71238577e-01 -7.44665086e-01 -1.61246628e-01
2.80546963e-01 7.43754268e-01 -6.35391891e-01 2.60266066e-01
1.36575267e-01 4.46169674e-01 -2.89788153e-02 1.28045559e-01
9.22806323e-01 1.25019597e-02 2.99960762e-01 -5.85157454e-01
-4.58970256e-02 -1.18205719e-01 -1.41027474e+00 1.89834535e+00
-6.85344636e-01 9.25033331e-01 1.32457670e-02 -8.68279517e-01
1.09194207e+00 4.97169755e-02 4.89631981e-01 -6.41056478e-01
6.74160272e-02 -1.80791281e-02 -2.44515955e-01 -9.92966965e-02
6.13498330e-01 -1.71950698e-01 -1.15871884e-01 6.03769541e-01
2.87715316e-01 -2.13746279e-01 1.09260425e-01 5.12327179e-02
8.41915727e-01 7.72546511e-03 -6.57462403e-02 -2.79600829e-01
7.52321064e-01 -6.30023062e-01 4.31173414e-01 6.65333867e-01
6.17559850e-02 8.78184497e-01 8.74620080e-01 -4.06568617e-01
-1.31045401e+00 -1.28246248e+00 -1.58867359e-01 6.41991138e-01
4.38360572e-02 -3.81505907e-01 -8.95850420e-01 -3.41803491e-01
-1.46578684e-01 5.62418103e-01 -7.61731088e-01 -4.25837398e-01
-6.02711320e-01 -6.60587013e-01 4.01798904e-01 3.72539371e-01
6.75727129e-01 -7.34847069e-01 -2.00048268e-01 -4.24118973e-02
-1.03613667e-01 -1.05361593e+00 -6.75533652e-01 1.81390807e-01
-7.77913988e-01 -6.83401227e-01 -9.68605220e-01 -6.83402240e-01
7.83913314e-01 -5.68617694e-02 1.08216655e+00 -3.59673619e-01
-2.73490757e-01 4.29418653e-01 2.84420140e-02 -1.64406970e-02
-6.62383139e-01 -2.37889975e-01 8.71232748e-02 5.59103847e-01
1.15697715e-03 -7.25177884e-01 -4.85752195e-01 1.98855504e-01
-9.04049456e-01 1.53860375e-01 5.04447222e-01 9.78569031e-01
5.79024374e-01 -2.09402293e-02 3.66135091e-02 -1.09386563e+00
6.72897041e-01 1.50224254e-01 -6.40672922e-01 4.00346577e-01
-6.57529771e-01 5.49775124e-01 7.08351612e-01 -4.68847215e-01
-1.40079308e+00 3.50222111e-01 -9.36043933e-02 -3.97146255e-01
-1.55843318e-01 9.68456343e-02 -3.12988132e-01 -5.78463018e-01
8.45763624e-01 7.61613175e-02 1.67315498e-01 -7.23969221e-01
9.49617386e-01 3.24059308e-01 8.40233505e-01 -6.12558246e-01
1.13963139e+00 4.37089890e-01 1.95811599e-01 -6.16044164e-01
-5.11973798e-01 -1.03779368e-01 -7.88540244e-01 -2.29344159e-01
9.54561353e-01 -6.68224037e-01 -3.02119881e-01 6.31559134e-01
-1.04536641e+00 -3.30561638e-01 -4.32867467e-01 2.32017130e-01
-8.17979634e-01 4.49055940e-01 -7.82769263e-01 -3.93650323e-01
5.84435575e-02 -1.16429830e+00 8.33605409e-01 1.14758886e-01
6.99421484e-03 -9.73851204e-01 -4.31557931e-02 2.68971431e-03
6.57368004e-01 1.32450521e-01 1.20386422e+00 3.63298804e-01
-5.33616662e-01 -5.14993556e-02 -3.77501428e-01 6.54741049e-01
4.50182021e-01 9.75306928e-02 -1.05846667e+00 -2.69041091e-01
1.48656681e-01 -6.59754593e-03 8.57432127e-01 3.10576141e-01
1.32127512e+00 -2.55212814e-01 1.86064750e-01 1.09468973e+00
1.63140523e+00 3.75951901e-02 9.66936588e-01 2.13833526e-01
8.20626974e-01 4.93533224e-01 -2.41432600e-02 1.99242577e-01
-8.64990950e-02 9.22814071e-01 1.15054950e-01 -5.05041540e-01
-4.59431052e-01 -1.38905689e-01 3.53662819e-01 6.65173233e-01
-8.35626051e-02 -2.45608892e-02 -4.73381817e-01 6.84591755e-02
-1.36714435e+00 -8.97748649e-01 8.34900588e-02 2.29661345e+00
1.06307876e+00 6.12332784e-02 -7.75609389e-02 1.86836958e-01
6.45539522e-01 3.06617856e-01 -4.22506750e-01 -7.40117908e-01
-3.76686573e-01 3.34458709e-01 5.97667634e-01 7.67529964e-01
-9.89877462e-01 8.21103930e-01 7.47503996e+00 6.40119255e-01
-1.47572124e+00 -1.33798957e-01 7.68290877e-01 2.64064312e-01
-4.81645316e-01 -6.14319779e-02 -2.89801359e-01 3.58315378e-01
7.10177600e-01 -7.12157190e-02 9.50914681e-01 5.91911614e-01
-1.81847408e-01 1.46708056e-01 -1.06407738e+00 1.20816135e+00
3.80289257e-01 -1.31808019e+00 2.86914796e-01 5.99144399e-02
8.83547962e-01 -5.73807418e-01 6.02397323e-01 -7.97599852e-02
7.95441642e-02 -7.25851297e-01 8.29095721e-01 7.06396341e-01
1.23148417e+00 -6.03711963e-01 1.52926564e-01 -1.77437186e-01
-7.59517014e-01 4.03335690e-02 -5.15604496e-01 1.22000411e-01
3.64468843e-02 4.06206101e-01 -1.27590436e-03 2.68238962e-01
5.91510236e-01 7.12621927e-01 -8.65541756e-01 7.93833017e-01
-3.52614641e-01 4.03595299e-01 -9.34767947e-02 3.58386278e-01
4.57116868e-03 -5.59904277e-01 5.18085301e-01 1.19273174e+00
5.37983716e-01 -9.79279578e-02 -3.40940088e-01 1.19241881e+00
-1.74812749e-01 -9.61794332e-02 -4.56432581e-01 -6.08201846e-02
6.64719800e-03 1.47742045e+00 -5.18464386e-01 -2.94496328e-01
-4.60256666e-01 1.56621885e+00 1.30728334e-01 5.79796493e-01
-6.92818105e-01 -6.21298909e-01 9.64687407e-01 -3.89826536e-01
3.40649277e-01 -3.09363008e-01 -4.88630682e-01 -1.33183241e+00
-3.52856913e-03 -9.49502587e-01 -8.75192434e-02 -1.03234875e+00
-1.30172539e+00 7.70195246e-01 -2.21536741e-01 -1.28089893e+00
-1.86834663e-01 -1.13574028e+00 -4.68328327e-01 1.12320387e+00
-1.53840244e+00 -9.18130159e-01 -4.16153163e-01 7.14417219e-01
-4.88799103e-02 -3.36562485e-01 1.11104548e+00 4.36100334e-01
-3.62106174e-01 8.38902116e-01 3.24371248e-01 1.19991980e-01
1.09299147e+00 -1.60711539e+00 3.18510652e-01 9.30008113e-01
4.07533318e-01 4.70149636e-01 8.19793224e-01 -7.48305842e-02
-1.52363849e+00 -7.75950015e-01 5.76149940e-01 -4.34575021e-01
5.71660161e-01 -4.50411320e-01 -9.13674295e-01 5.38360178e-01
3.91986102e-01 -1.33975234e-03 6.11522734e-01 1.21017769e-01
-6.75986230e-01 -3.74118179e-01 -1.02754474e+00 8.25520098e-01
8.88086319e-01 -8.97080421e-01 -1.89475223e-01 4.53581996e-02
5.07417142e-01 -3.94866318e-01 -8.03761482e-01 -1.34658679e-01
5.05391777e-01 -1.11959243e+00 1.14018750e+00 -6.33538365e-01
5.90360105e-01 -3.27198148e-01 -2.53081918e-01 -1.46787882e+00
-6.22257411e-01 -5.70776105e-01 1.37534350e-01 1.23005688e+00
4.15132672e-01 -3.89556229e-01 7.61626959e-01 6.53757334e-01
1.42790154e-01 -3.91757488e-01 -5.58719158e-01 -6.13303483e-01
1.42047659e-01 -2.42003024e-01 6.54258609e-01 9.49122787e-01
-4.22027141e-01 2.89064109e-01 -8.82375181e-01 9.98664126e-02
6.89930499e-01 1.03260800e-01 6.44536257e-01 -9.95186269e-01
-4.67321426e-01 -6.75438106e-01 -6.42407596e-01 -1.02292395e+00
2.77412444e-01 -6.68390512e-01 -1.81032449e-01 -1.08820510e+00
1.06349103e-01 -4.98241931e-01 -4.97314543e-01 3.40851665e-01
3.17513337e-03 5.87358773e-01 3.40025008e-01 -1.78782061e-01
-1.08203486e-01 4.42230463e-01 1.39966249e+00 -3.10010880e-01
-8.40526596e-02 -2.96405762e-01 -8.61404836e-01 5.88182986e-01
6.06996894e-01 -1.87997282e-01 -3.92272681e-01 -7.72399366e-01
2.51996517e-01 6.35688826e-02 3.73972714e-01 -1.07242692e+00
-1.83570106e-02 -1.58943936e-01 6.83325768e-01 7.51956925e-02
7.32303321e-01 -8.15204561e-01 3.18325639e-01 1.03546843e-01
-5.76132774e-01 -3.05640716e-02 8.54499191e-02 4.98716056e-01
-5.24188399e-01 -2.60083020e-01 9.09046233e-01 7.69345537e-02
-8.86787415e-01 1.58718348e-01 -2.33614415e-01 -3.39927524e-01
4.88534838e-01 -1.80756927e-01 -2.28638694e-01 -7.08010316e-01
-7.54599869e-01 -4.79579806e-01 7.00910270e-01 4.65965062e-01
4.59781855e-01 -1.77460992e+00 -5.00153303e-01 7.01220870e-01
-5.04108816e-02 -8.36218834e-01 2.97094434e-01 3.81734699e-01
-6.20123386e-01 7.49323368e-02 -9.25830901e-01 -4.77131635e-01
-1.16054201e+00 5.67693949e-01 4.50689584e-01 8.68307501e-02
-3.87444973e-01 8.30026984e-01 1.83443725e-01 -2.29611382e-01
6.19783252e-03 -1.62283197e-01 -5.69400601e-02 -1.16750039e-02
4.64475036e-01 5.56070283e-02 -1.09697185e-01 -8.02053928e-01
3.29885818e-02 1.13499033e+00 1.06513076e-01 -3.62909138e-01
1.01997375e+00 -2.20252872e-01 -4.12759691e-01 4.07683462e-01
1.57999945e+00 2.26245895e-01 -1.47398388e+00 -2.59165823e-01
-3.57915044e-01 -1.06050539e+00 3.32896233e-01 -7.05493093e-01
-1.42542922e+00 1.04071510e+00 9.27420914e-01 2.34706402e-01
1.53323328e+00 -3.64091069e-01 4.09322530e-01 2.70770192e-01
2.46506974e-01 -9.92871761e-01 2.59720594e-01 3.39065343e-01
1.08502507e+00 -1.14017975e+00 -2.67849062e-02 -4.18880016e-01
-5.98848939e-01 1.30569696e+00 2.75833517e-01 -1.99586913e-01
6.12074077e-01 2.05281556e-01 3.97250772e-01 2.85915613e-01
-3.72895777e-01 1.46935329e-01 6.10860467e-01 6.62314832e-01
3.75457853e-01 1.08369924e-01 -4.31787446e-02 8.08131695e-02
-3.52962375e-01 -2.63604790e-01 6.24333084e-01 4.98178065e-01
-1.17422998e-01 -1.68761325e+00 -4.08698797e-01 2.83826351e-01
-1.63375244e-01 6.70773685e-02 -3.53281796e-01 4.23156857e-01
-9.56957508e-03 5.83551705e-01 1.49836034e-01 -4.76305842e-01
2.88918436e-01 3.76078524e-02 8.37856948e-01 -2.52390444e-01
-5.20401485e-02 3.13252471e-02 -2.13473663e-01 -7.97419727e-01
-2.46391803e-01 -5.79801142e-01 -5.37514031e-01 -3.85358751e-01
9.66306105e-02 -2.29440272e-01 7.36988127e-01 4.23601061e-01
7.88224488e-02 4.96013880e-01 7.31829166e-01 -9.54035342e-01
-2.93624967e-01 -3.43686968e-01 -8.51836503e-01 7.42263615e-01
4.12876338e-01 -5.83178759e-01 -3.21632057e-01 4.33425993e-01] | [11.645468711853027, -0.49930983781814575] |
5a67d5c5-751f-455b-b28a-04ef082d38ab | abstractive-and-extractive-text-summarization | 1807.08000 | null | http://arxiv.org/abs/1807.08000v2 | http://arxiv.org/pdf/1807.08000v2.pdf | Abstractive and Extractive Text Summarization using Document Context Vector and Recurrent Neural Networks | Sequence to sequence (Seq2Seq) learning has recently been used for
abstractive and extractive summarization. In current study, Seq2Seq models have
been used for eBay product description summarization. We propose a novel
Document-Context based Seq2Seq models using RNNs for abstractive and extractive
summarizations. Intuitively, this is similar to humans reading the title,
abstract or any other contextual information before reading the document. This
gives humans a high-level idea of what the document is about. We use this idea
and propose that Seq2Seq models should be started with contextual information
at the first time-step of the input to obtain better summaries. In this manner,
the output summaries are more document centric, than being generic, overcoming
one of the major hurdles of using generative models. We generate
document-context from user-behavior and seller provided information. We train
and evaluate our models on human-extracted-golden-summaries. The
document-contextual Seq2Seq models outperform standard Seq2Seq models.
Moreover, generating human extracted summaries is prohibitively expensive to
scale, we therefore propose a semi-supervised technique for extracting
approximate summaries and using it for training Seq2Seq models at scale.
Semi-supervised models are evaluated against human extracted summaries and are
found to be of similar efficacy. We provide side by side comparison for
abstractive and extractive summarizers (contextual and non-contextual) on same
evaluation dataset. Overall, we provide methodologies to use and evaluate the
proposed techniques for large document summarization. Furthermore, we found
these techniques to be highly effective, which is not the case with existing
techniques. | ['Nish Parikh', 'Gyanit Singh', 'Chandra Khatri'] | 2018-07-20 | null | null | null | null | ['extractive-document-summarization'] | ['natural-language-processing'] | [ 5.95926642e-01 2.90670693e-01 -3.38609397e-01 -3.68322164e-01
-1.03157043e+00 -7.71267533e-01 7.89740920e-01 4.63259697e-01
-4.11941200e-01 9.98236477e-01 1.05154502e+00 5.72835915e-02
8.03943798e-02 -5.63728213e-01 -6.57738745e-01 -3.30479562e-01
1.85967952e-01 4.90370721e-01 -6.67147413e-02 -2.81034499e-01
7.38482952e-01 1.76235750e-01 -1.47335660e+00 7.62257218e-01
1.13879621e+00 5.04447520e-01 3.17464978e-01 1.38209283e+00
-2.60253847e-01 8.33313644e-01 -1.32847404e+00 -4.73283678e-01
-1.28404573e-01 -9.72474277e-01 -8.03887010e-01 1.13459788e-02
6.53918684e-01 -5.70155382e-01 -5.92146581e-03 8.17506015e-01
7.80004263e-01 2.95721024e-01 9.71841931e-01 -8.14331055e-01
-5.71285009e-01 1.09600270e+00 -4.87108886e-01 1.44790545e-01
8.71820867e-01 5.07781208e-02 1.25244415e+00 -4.49388593e-01
7.20355451e-01 1.26357150e+00 5.35269201e-01 6.36454046e-01
-9.28091109e-01 -1.65672868e-01 -1.06253386e-01 -2.30454877e-01
-6.68177247e-01 -4.92227107e-01 5.30798674e-01 3.97130176e-02
1.30450273e+00 6.51223183e-01 5.67567229e-01 1.48026359e+00
5.39526224e-01 1.30469108e+00 6.14623904e-01 -3.16577762e-01
3.27416003e-01 2.10689485e-01 3.65620017e-01 1.12892814e-01
4.62417424e-01 -2.32831776e-01 -5.86592615e-01 -1.46366403e-01
2.44220346e-01 4.14546803e-02 -1.65861100e-01 2.87036121e-01
-1.11901021e+00 9.46548343e-01 9.85332355e-02 2.35547334e-01
-7.35472322e-01 1.37809783e-01 1.03633320e+00 1.75899714e-01
5.39258778e-01 8.79326463e-01 -2.21273735e-01 -5.12712121e-01
-1.54473817e+00 6.51951671e-01 1.17639279e+00 1.24979055e+00
1.80452570e-01 3.42656642e-01 -8.44430685e-01 8.31677854e-01
-2.13120684e-01 6.42031610e-01 8.24130833e-01 -9.62934196e-01
6.25085831e-01 3.89503598e-01 1.64264381e-01 -8.43438327e-01
-1.87235817e-01 -5.69792032e-01 -1.00269973e+00 -4.81763005e-01
-2.64867067e-01 -2.48898521e-01 -8.11145186e-01 1.20653033e+00
-3.44785452e-01 -2.64588833e-01 5.33728898e-01 5.56037724e-01
1.25347781e+00 1.22144818e+00 -4.97411005e-02 -4.98153210e-01
1.23668253e+00 -1.18996572e+00 -1.04325271e+00 -1.06083415e-01
6.42721534e-01 -8.50827038e-01 9.56084192e-01 3.93784821e-01
-1.27677310e+00 -6.55804455e-01 -1.16330028e+00 -9.09927934e-02
-2.73104727e-01 3.48161787e-01 3.51473868e-01 5.61973810e-01
-9.17887866e-01 9.41453576e-01 -5.84854484e-01 -5.64643621e-01
2.61615694e-01 2.48592556e-01 -7.60189965e-02 1.45552710e-01
-1.02258801e+00 7.11086631e-01 8.85723650e-01 -1.49082676e-01
-7.20490694e-01 -5.66401362e-01 -1.02796710e+00 3.68395180e-01
3.06032866e-01 -9.90866959e-01 1.74132836e+00 -7.97358274e-01
-1.54889750e+00 1.01245813e-01 -3.22101504e-01 -7.88744509e-01
3.80659312e-01 -6.11170352e-01 -3.31756651e-01 2.26338282e-01
2.77380794e-01 6.15039945e-01 6.43552303e-01 -1.07278955e+00
-6.21728539e-01 -1.22442380e-01 -1.12820283e-01 3.46038908e-01
-2.69704640e-01 3.37849488e-03 -7.05931559e-02 -7.23852813e-01
-6.19387627e-01 -7.15248227e-01 -1.66234851e-01 -1.01679373e+00
-9.13063467e-01 -4.31265324e-01 5.73676765e-01 -9.46678638e-01
1.68507719e+00 -1.73037076e+00 -2.35005355e-04 -1.90935716e-01
-2.37562098e-02 6.82116210e-01 -3.88745815e-01 1.01634264e+00
3.04853022e-01 2.42548078e-01 -2.48875633e-01 -6.00020885e-01
5.61891049e-02 2.74368152e-02 -5.00392437e-01 -2.86288619e-01
2.09414214e-01 1.18245363e+00 -1.03887999e+00 -5.90636373e-01
-9.64855626e-02 1.58622995e-01 -5.09106874e-01 2.89472461e-01
-4.45238709e-01 7.37178922e-02 -5.12114465e-01 3.36962402e-01
4.61448938e-01 1.58667192e-01 -1.34860262e-01 -1.65717378e-01
-3.11819203e-02 4.92650032e-01 -8.09587955e-01 1.78061223e+00
-5.97238481e-01 6.47952437e-01 -5.29451430e-01 -6.72513247e-01
8.51258457e-01 2.55856037e-01 -8.80815648e-03 -7.42000043e-01
8.53606090e-02 2.34405145e-01 -3.22934568e-01 -5.84372222e-01
1.42398012e+00 3.52493394e-03 -4.17654157e-01 5.97068727e-01
2.92084426e-01 -4.89740431e-01 6.86562955e-01 6.23050511e-01
9.89965558e-01 2.06114516e-01 7.72734225e-01 -3.60099156e-03
3.47648889e-01 9.64382812e-02 1.96114674e-01 1.19417357e+00
3.38877171e-01 8.35780859e-01 6.63487315e-01 -3.79334167e-02
-1.07928240e+00 -7.30140746e-01 4.73121673e-01 9.91294861e-01
-2.12779358e-01 -8.92735898e-01 -1.03385425e+00 -8.91166151e-01
-2.78779656e-01 1.43242109e+00 -4.00252968e-01 -2.16961667e-01
-6.59102440e-01 -2.04642385e-01 6.63564324e-01 7.23269165e-01
3.80709171e-01 -1.36919129e+00 -8.51761580e-01 3.79553348e-01
-2.28649527e-01 -9.17443097e-01 -9.51199353e-01 3.92674878e-02
-1.01260364e+00 -4.61357236e-01 -1.04511869e+00 -4.89603072e-01
1.94537178e-01 5.50422668e-02 1.19808936e+00 -3.09984148e-01
1.37367740e-01 2.21589431e-01 -6.24735773e-01 -8.19000125e-01
-9.90599453e-01 5.72222769e-01 -1.06824718e-01 -4.23369676e-01
1.65622011e-01 -3.27321023e-01 -5.49169362e-01 -3.12823743e-01
-1.19011188e+00 5.15910313e-02 1.05347574e+00 7.55534172e-01
2.20687836e-01 -5.99832274e-02 1.07694376e+00 -1.23482430e+00
1.55559373e+00 -1.62713826e-01 6.69453964e-02 3.12968463e-01
-3.93460900e-01 3.42166066e-01 1.04200709e+00 -4.07341301e-01
-1.13786578e+00 -2.30644315e-01 -2.37480462e-01 -2.88423933e-02
-1.94336265e-01 7.32317448e-01 1.57141145e-02 9.14325774e-01
7.96578169e-01 5.10946810e-01 -7.89108872e-02 -2.91283846e-01
4.12690789e-01 1.10190904e+00 4.82015878e-01 -2.52011180e-01
3.33964020e-01 -1.02389492e-01 -2.93690056e-01 -1.04843855e+00
-1.04224741e+00 -6.59278274e-01 -3.26245397e-01 6.48194328e-02
7.09764540e-01 -5.36496401e-01 -5.54958820e-01 4.85706516e-02
-1.51951206e+00 5.60637303e-02 -6.34212315e-01 3.12764972e-01
-7.32070744e-01 6.19097769e-01 -5.33608735e-01 -8.64460230e-01
-1.16611326e+00 -8.05608988e-01 1.40733039e+00 4.91822362e-01
-9.18015122e-01 -9.34933066e-01 2.74187475e-01 1.89689040e-01
5.10088146e-01 3.07762086e-01 7.32821405e-01 -1.40208137e+00
-1.69745758e-02 -4.43041265e-01 -2.86155492e-02 6.78050756e-01
2.59034932e-01 -8.95477608e-02 -6.64139330e-01 -1.37780577e-01
-1.11586936e-02 -2.81204641e-01 1.07773352e+00 5.82056463e-01
7.77642965e-01 -9.76217330e-01 -1.33695588e-01 -1.26353770e-01
1.11250246e+00 5.65170571e-02 6.16602838e-01 -1.68864196e-03
5.64591825e-01 6.15854084e-01 8.92536759e-01 5.73111475e-01
1.88851207e-01 3.55397433e-01 -8.74059349e-02 1.74846724e-01
-2.06896007e-01 -6.43387794e-01 7.25191712e-01 9.74438906e-01
-1.36104077e-02 -7.95737982e-01 -3.66722137e-01 6.50254250e-01
-1.95753396e+00 -1.30540514e+00 6.03107677e-04 1.90916777e+00
8.82599354e-01 3.86433393e-01 3.55414033e-01 -2.12249666e-01
4.10383552e-01 3.07611704e-01 -4.44150835e-01 -1.28477907e+00
3.66557040e-03 1.47996262e-01 3.11061740e-01 3.14594209e-01
-7.24661827e-01 8.59256446e-01 5.87954807e+00 1.02004147e+00
-8.00705910e-01 -3.18009526e-01 4.23280239e-01 -2.91274130e-01
-4.57649291e-01 -2.26913691e-01 -9.30002868e-01 5.02113402e-01
1.38637459e+00 -5.07825196e-01 -5.51600121e-02 8.52650940e-01
4.19380069e-01 -2.54030347e-01 -1.49767196e+00 7.64496386e-01
4.95578349e-01 -1.41913533e+00 7.56484389e-01 -8.18602741e-02
9.48334217e-01 -3.01619232e-01 -1.67506739e-01 6.51600480e-01
1.10507168e-01 -1.02702665e+00 4.98874187e-01 5.60088575e-01
5.12289882e-01 -1.03008473e+00 1.25339186e+00 6.38126612e-01
-7.38096356e-01 8.38984400e-02 -4.44508642e-01 2.10966185e-01
4.37184125e-01 4.64491546e-01 -1.16329765e+00 9.59742486e-01
1.72802418e-01 8.85947347e-01 -5.21442831e-01 8.07481945e-01
-8.44702348e-02 4.76117402e-01 3.74073051e-02 -6.70637488e-01
4.31872249e-01 -5.46813793e-02 7.75524676e-01 1.82113957e+00
5.02067685e-01 -8.86927545e-02 3.10940389e-02 7.23397136e-01
-2.02318609e-01 2.53632128e-01 -7.46239066e-01 -4.75232124e-01
1.97660819e-01 1.11013901e+00 -5.34657240e-01 -7.26372600e-01
-7.68367276e-02 1.21122074e+00 -1.73075795e-01 2.11383298e-01
-5.85329831e-01 -8.35823715e-01 -1.02193259e-01 -4.01366055e-02
4.37812120e-01 1.71076551e-01 -1.42169431e-01 -9.84305620e-01
6.08103909e-02 -1.22855604e+00 3.19879413e-01 -7.68791735e-01
-9.58098173e-01 7.47887969e-01 3.67030025e-01 -1.19971228e+00
-8.42602968e-01 -3.33165377e-02 -8.95578504e-01 6.00908637e-01
-1.35309911e+00 -1.05983138e+00 -6.52190000e-02 -1.32296532e-01
1.38875139e+00 -2.19498888e-01 6.74134016e-01 -3.54780167e-01
-2.00010672e-01 5.33451080e-01 1.57497540e-01 -2.27995858e-01
7.77846992e-01 -1.42984760e+00 6.44334435e-01 7.56546140e-01
1.53719947e-01 9.33345258e-01 1.17609751e+00 -9.78099704e-01
-1.19599724e+00 -1.15796494e+00 1.08678341e+00 -4.15392756e-01
2.94762373e-01 -2.69800961e-01 -7.73516595e-01 4.70571280e-01
9.47239161e-01 -9.91706610e-01 7.10086644e-01 -4.00146618e-02
1.41265616e-01 -6.51984228e-05 -9.64450479e-01 6.84947193e-01
8.26933026e-01 -2.15875894e-01 -1.04342794e+00 3.05539012e-01
1.07438111e+00 -3.11644316e-01 -6.28605604e-01 2.74049938e-01
5.97340703e-01 -9.19326007e-01 5.25652647e-01 -5.51671386e-01
9.94350731e-01 2.98744980e-02 2.92842686e-02 -1.76844323e+00
1.60977527e-01 -8.80604267e-01 -4.17814136e-01 1.58528113e+00
5.75010717e-01 -3.26570541e-01 6.17635429e-01 2.54728328e-02
-3.72976601e-01 -6.66305959e-01 -1.66248560e-01 -8.60469937e-01
-7.12008178e-02 7.44934976e-02 5.61569631e-01 2.92476058e-01
1.17812425e-01 9.59940135e-01 -4.58568454e-01 -3.96339029e-01
3.71444017e-01 2.70392478e-01 9.32849288e-01 -7.99309909e-01
-2.69714028e-01 -4.97566104e-01 -8.19176808e-02 -1.24589241e+00
1.64499879e-01 -6.99501336e-01 9.09748748e-02 -1.95811784e+00
5.66583812e-01 4.37784106e-01 1.88449681e-01 1.33669749e-02
-2.95376241e-01 -2.49068975e-01 1.94134772e-01 -2.80576199e-02
-9.15910900e-01 6.89546585e-01 1.23473763e+00 -2.26875052e-01
-5.48977017e-01 3.44551533e-01 -9.65675890e-01 2.73614645e-01
8.52278590e-01 -4.52151597e-01 -6.20914161e-01 4.13840227e-02
2.05606192e-01 2.57049292e-01 -5.86154312e-02 -7.84230351e-01
3.41693670e-01 3.99073213e-02 2.69334108e-01 -1.17591727e+00
8.59651864e-02 -3.29156160e-01 -2.53795445e-01 4.42719102e-01
-9.85494196e-01 1.85831249e-01 2.40330979e-01 6.50344133e-01
-3.89089376e-01 -7.20047295e-01 1.76887706e-01 -3.42551500e-01
-3.15866143e-01 -1.71209455e-01 -3.67414802e-01 9.75719765e-02
6.09513879e-01 -4.30595011e-01 -1.81081682e-01 -9.22643244e-01
-3.33967149e-01 1.99060202e-01 1.97749898e-01 4.33836907e-01
7.16986656e-01 -9.48340774e-01 -1.13245308e+00 -2.50109047e-01
1.26443595e-01 -1.47396997e-01 2.55424112e-01 4.86212462e-01
-4.63056296e-01 9.43213284e-01 -1.45023748e-01 -5.43250382e-01
-1.44803715e+00 5.04359066e-01 -2.61117667e-01 -6.92506015e-01
-4.35765415e-01 2.56053418e-01 3.14495862e-02 -3.27255994e-01
9.77709591e-02 -6.36536121e-01 -5.73430002e-01 3.70281190e-01
7.84975529e-01 5.86872995e-01 1.51920766e-01 -3.06985408e-01
1.14612497e-01 2.32508048e-01 -6.37152791e-01 -1.58448994e-01
1.37924349e+00 -2.98418701e-02 1.90907270e-02 5.17307937e-01
1.33770299e+00 1.59441635e-01 -7.66535163e-01 6.82144985e-02
2.02691868e-01 4.46068421e-02 -2.52061248e-01 -8.61972988e-01
-3.69089782e-01 6.73574269e-01 -3.24474983e-02 4.86943811e-01
9.89839613e-01 -3.92713733e-02 1.07908428e+00 6.10758424e-01
-1.21470675e-01 -1.18310773e+00 2.77026892e-01 4.64872688e-01
1.08618534e+00 -1.12287033e+00 3.39386642e-01 -2.13315692e-02
-1.07507825e+00 1.11413097e+00 3.55839640e-01 -1.29697964e-01
-2.54583031e-01 -4.60691284e-03 -4.27844286e-01 -1.58745483e-01
-8.18714261e-01 1.21698871e-01 4.89966393e-01 4.33591515e-01
6.49003446e-01 4.73402143e-02 -4.69749659e-01 7.70208180e-01
-6.33452654e-01 -4.23153602e-02 1.00657654e+00 1.00970232e+00
-4.74826336e-01 -9.18708324e-01 -6.55646473e-02 7.36649275e-01
-6.08911276e-01 -1.80247024e-01 -7.34204233e-01 6.70953214e-01
-6.54520631e-01 1.11923444e+00 -1.79161489e-01 -3.34188826e-02
5.43308318e-01 4.18047868e-02 3.65254581e-01 -9.88236547e-01
-9.51736033e-01 1.16259024e-01 5.39312124e-01 -3.00921619e-01
-3.74967754e-01 -6.82752132e-01 -1.08656073e+00 -2.76587725e-01
-3.82511824e-01 4.42777365e-01 7.12081492e-01 8.72823298e-01
5.60606182e-01 8.10389817e-01 6.37116194e-01 -8.86366844e-01
-9.51188207e-01 -1.36571431e+00 -3.60791206e-01 3.44152033e-01
4.95104343e-01 2.16949433e-01 -2.34318167e-01 1.02111958e-01] | [12.44977855682373, 9.415557861328125] |
15b4a4d5-efff-4e4c-a33f-08ddb3018235 | large-scale-pre-trained-models-are | 2303.15975 | null | https://arxiv.org/abs/2303.15975v2 | https://arxiv.org/pdf/2303.15975v2.pdf | Large-scale Pre-trained Models are Surprisingly Strong in Incremental Novel Class Discovery | Discovering novel concepts from unlabelled data and in a continuous manner is an important desideratum of lifelong learners. In the literature such problems have been partially addressed under very restricted settings, where either access to labelled data is provided for discovering novel concepts (e.g., NCD) or learning occurs for a limited number of incremental steps (e.g., class-iNCD). In this work we challenge the status quo and propose a more challenging and practical learning paradigm called MSc-iNCD, where learning occurs continuously and unsupervisedly, while exploiting the rich priors from large-scale pre-trained models. To this end, we propose simple baselines that are not only resilient under longer learning scenarios, but are surprisingly strong when compared with sophisticated state-of-the-art methods. We conduct extensive empirical evaluation on a multitude of benchmarks and show the effectiveness of our proposed baselines, which significantly raises the bar. | ['Elisa Ricci', 'Nicu Sebe', 'Zhun Zhong', 'Subhankar Roy', 'Mingxuan Liu'] | 2023-03-28 | null | null | null | null | ['novel-class-discovery', 'novel-class-discovery', 'novel-concepts'] | ['computer-vision', 'methodology', 'reasoning'] | [ 5.07031024e-01 2.58775800e-01 -4.85850424e-01 -3.79412830e-01
-7.50881791e-01 -6.72969162e-01 9.57470298e-01 4.71241862e-01
-8.12873304e-01 1.02224910e+00 4.17387411e-02 -2.92623043e-01
-1.98684379e-01 -4.37111646e-01 -9.59268153e-01 -5.55091321e-01
-1.44261628e-01 5.52555799e-01 3.46299976e-01 -5.54963062e-03
7.94810355e-02 2.89704978e-01 -2.01208138e+00 -7.64307976e-02
9.74427223e-01 7.85635650e-01 -8.48705322e-03 3.87764245e-01
-6.73841238e-02 5.84553063e-01 -2.22815603e-01 -4.62392241e-01
1.80078477e-01 -2.88296312e-01 -8.13768864e-01 2.18097284e-01
4.66035187e-01 -6.96300939e-02 -2.82971151e-02 8.80364478e-01
4.70492065e-01 3.38677943e-01 7.50953019e-01 -1.09581757e+00
-2.78807044e-01 6.13768637e-01 -6.76017880e-01 2.12106913e-01
2.34573543e-01 3.00095286e-02 1.07525408e+00 -1.17933583e+00
3.34664553e-01 1.08914101e+00 6.02534771e-01 7.24639833e-01
-1.21023881e+00 -6.85064375e-01 7.10614562e-01 2.39854380e-01
-1.25216806e+00 -6.10597014e-01 5.56436718e-01 -1.99593961e-01
5.74130476e-01 1.31320551e-01 3.36844653e-01 1.39965129e+00
-2.28538647e-01 1.21063900e+00 1.23544693e+00 -6.42808616e-01
4.30089414e-01 1.47921070e-01 3.57960135e-01 4.57289100e-01
4.13465410e-01 1.71965912e-01 -6.87059641e-01 -5.97253777e-02
2.25722313e-01 1.80990607e-01 -1.42324656e-01 -4.92341757e-01
-1.21063530e+00 7.09379196e-01 7.47049053e-04 2.22704858e-01
-2.40059942e-01 -2.20313191e-01 3.51763338e-01 4.25427914e-01
7.23410726e-01 3.49440962e-01 -7.67675281e-01 -3.17486674e-01
-9.50053811e-01 2.97508150e-01 7.06251144e-01 1.03950465e+00
8.45840156e-01 -3.07640642e-01 -1.75568581e-01 8.46648991e-01
-5.75772822e-02 9.38366875e-02 6.30332589e-01 -4.09879267e-01
3.84515733e-01 2.51296520e-01 9.01575759e-02 -4.33760643e-01
-4.84706610e-01 -9.19886410e-01 -8.33114922e-01 -1.56650975e-01
3.76024544e-01 -3.45025718e-01 -9.32655692e-01 1.86187124e+00
5.76280117e-01 8.56979311e-01 2.50313640e-01 4.03240532e-01
5.10798931e-01 3.32639903e-01 1.15226924e-01 -5.44517398e-01
9.03864980e-01 -1.07959318e+00 -5.53531528e-01 -3.52167577e-01
6.36950970e-01 -3.26148480e-01 1.21670818e+00 8.17870915e-01
-1.02010810e+00 -6.54177010e-01 -9.35486495e-01 1.18455626e-01
-2.58492202e-01 -2.86714017e-01 6.92423105e-01 5.74721038e-01
-8.53410721e-01 5.07821560e-01 -8.46017241e-01 -4.27070379e-01
5.72162628e-01 1.02762818e-01 -1.13602251e-01 -4.86679792e-01
-9.24128175e-01 5.21533847e-01 7.66620576e-01 -1.31170735e-01
-1.26138532e+00 -8.17152441e-01 -6.08496964e-01 -6.47730306e-02
9.20904756e-01 -3.79522234e-01 1.51288390e+00 -8.02045107e-01
-1.44218791e+00 8.29943895e-01 -5.64270318e-02 -8.07889938e-01
8.52953315e-01 -7.21310377e-01 -3.47670704e-01 1.02684163e-01
2.71932431e-03 6.86840296e-01 9.48767364e-01 -1.43891275e+00
-1.21756673e+00 -3.08674127e-01 2.58825779e-01 2.97060490e-01
-8.61381888e-01 -2.55308449e-01 -5.08255899e-01 -6.88499331e-01
-1.99863072e-02 -9.16996181e-01 -2.68963426e-01 -3.48251253e-01
-3.02682132e-01 -6.73166811e-01 8.65104795e-01 -9.24637169e-02
1.26778018e+00 -2.06589532e+00 -4.26116912e-03 -1.39230704e-02
1.71647057e-01 3.54532540e-01 -7.75345936e-02 3.08198631e-01
2.09290236e-02 8.90996829e-02 -4.80294347e-01 -5.27501702e-01
5.53444363e-02 3.01401794e-01 -3.66621256e-01 2.55761385e-01
2.30740398e-01 9.56323862e-01 -1.25026524e+00 -3.43165845e-01
-1.21395350e-01 1.30890638e-01 -4.65287745e-01 3.19703579e-01
-4.89325076e-01 5.87246716e-01 -3.61088246e-01 5.97515881e-01
4.44979846e-01 -3.61290336e-01 1.96326107e-01 3.15148979e-01
2.14712904e-03 1.34123683e-01 -1.15277052e+00 1.96962714e+00
-5.31103134e-01 1.79252550e-01 -2.84353018e-01 -1.34285986e+00
7.04112411e-01 3.56887668e-01 2.54126430e-01 -5.64568639e-01
-2.96837002e-01 2.28238747e-01 -2.13826686e-01 -3.91154677e-01
3.70531797e-01 -3.99077743e-01 6.03740551e-02 5.42695522e-01
2.60710478e-01 2.12125883e-01 2.48100802e-01 2.30556473e-01
1.18938255e+00 1.90442607e-01 4.54462349e-01 -2.51835406e-01
4.07564610e-01 -2.87819654e-01 6.24217331e-01 1.29264951e+00
-1.74198285e-01 4.00428206e-01 2.20699012e-01 -6.90884590e-02
-5.04078746e-01 -1.13418174e+00 -2.12236896e-01 1.48795557e+00
5.49376830e-02 -3.69033307e-01 -3.14764798e-01 -1.41549885e+00
-4.08638120e-02 6.95654750e-01 -5.11216402e-01 -3.40019196e-01
-4.65740651e-01 -8.62899065e-01 2.59946167e-01 6.46658719e-01
3.30522537e-01 -7.92475462e-01 -4.77534056e-01 3.29295754e-01
1.10764578e-01 -1.11993742e+00 -1.81731954e-01 5.27084589e-01
-1.08638012e+00 -9.62338269e-01 -7.68715739e-01 -8.83542538e-01
5.99182665e-01 3.28712732e-01 1.44883406e+00 -8.83244053e-02
2.41251811e-02 5.18919528e-01 -6.12895846e-01 -6.29295051e-01
8.28431845e-02 3.97295564e-01 2.83329189e-01 2.56449044e-01
4.03970450e-01 -9.26788151e-01 -4.57599729e-01 2.40500569e-01
-1.11675048e+00 3.72488098e-03 1.01663661e+00 9.60589588e-01
7.42057621e-01 4.16997552e-01 1.31784225e+00 -1.58178067e+00
3.70724976e-01 -9.07476783e-01 -3.83765429e-01 3.10912371e-01
-9.34740782e-01 8.38089585e-02 7.42476404e-01 -6.23639226e-01
-1.20391703e+00 -4.91556637e-02 -4.92833331e-02 -3.41872156e-01
-4.84281570e-01 6.64217234e-01 -2.89669901e-01 2.16765359e-01
5.83206236e-01 3.05168152e-01 -4.62391973e-01 -7.98366666e-01
4.88791227e-01 5.70639133e-01 7.54522860e-01 -9.70842838e-01
1.17114198e+00 5.64768791e-01 -1.91799104e-01 -7.51556873e-01
-1.52114820e+00 -7.05541134e-01 -7.80309081e-01 6.29240423e-02
2.11451665e-01 -1.16623807e+00 -2.25308180e-01 3.84998620e-01
-5.83274722e-01 -4.62481260e-01 -5.94369411e-01 5.30204952e-01
-3.42832029e-01 3.40053290e-01 -2.39257067e-01 -8.40487123e-01
-3.90405059e-02 -5.20707786e-01 7.07164586e-01 3.23532015e-01
6.12501837e-02 -1.10008347e+00 8.13859850e-02 2.20135674e-01
1.28449619e-01 2.30534703e-01 7.52728522e-01 -1.08131540e+00
-3.74980003e-01 -1.27942665e-02 5.32108475e-04 4.68993217e-01
2.56116092e-01 -5.66110075e-01 -1.16178298e+00 -6.17037535e-01
-2.02843487e-01 -9.49881494e-01 1.32008886e+00 -7.92231858e-02
1.62828028e+00 -3.39929432e-01 -3.99569809e-01 3.53931159e-01
1.14063954e+00 1.09206503e-02 1.15938313e-01 1.50479078e-01
5.55597603e-01 6.56120956e-01 9.50161755e-01 4.79752302e-01
5.14049292e-01 4.41203296e-01 2.36483708e-01 2.87070200e-02
6.38810024e-02 -3.49784881e-01 1.86342016e-01 9.47051942e-01
1.41800009e-02 -3.39494765e-01 -1.01311648e+00 9.25848663e-01
-1.93461430e+00 -5.17575085e-01 3.22725236e-01 2.20885134e+00
1.29040766e+00 6.05010092e-01 1.76345840e-01 3.52003098e-01
5.74149251e-01 1.49716824e-01 -8.64109516e-01 1.17264502e-01
-7.81613663e-02 5.64675033e-01 1.24350414e-01 3.19875062e-01
-1.30367208e+00 8.50136697e-01 5.83878422e+00 1.01541948e+00
-7.35553443e-01 2.26685166e-01 5.29281974e-01 -1.70040876e-01
-3.10151696e-01 1.00113474e-01 -9.73469436e-01 2.92376906e-01
1.04470849e+00 -9.19699147e-02 1.08794384e-01 7.36363351e-01
-1.54058874e-01 -6.55285716e-02 -1.46944475e+00 8.72371435e-01
2.58043408e-01 -9.37900066e-01 -5.67718372e-02 -2.23865714e-02
1.11981869e+00 -1.67192146e-02 2.10927501e-01 7.65979469e-01
5.24641454e-01 -9.11001801e-01 4.90413010e-01 4.97987121e-01
7.35311151e-01 -1.00687587e+00 4.77681816e-01 6.90498531e-01
-1.06255758e+00 -2.42653370e-01 -2.20329344e-01 -1.17788732e-01
-1.28513366e-01 9.33952272e-01 -7.28649497e-01 7.95464396e-01
5.91218591e-01 1.15000951e+00 -8.83486927e-01 1.05793357e+00
-4.66593087e-01 1.35378265e+00 -2.77585626e-01 1.72307402e-01
3.48567277e-01 1.95221663e-01 3.47910464e-01 1.29710519e+00
9.04307216e-02 3.83964973e-03 3.59416693e-01 4.05926675e-01
-5.96783757e-01 1.09353110e-01 -3.99737000e-01 -1.37722027e-02
4.86916721e-01 1.15015852e+00 -5.69682360e-01 -3.28877151e-01
-7.59265363e-01 7.43195117e-01 5.81954658e-01 4.91668969e-01
-4.15471792e-01 -1.93215579e-01 2.16812178e-01 1.47787660e-01
5.19866288e-01 -3.61819863e-02 1.83251679e-01 -1.34366262e+00
2.33838752e-01 -8.73866558e-01 8.52946043e-01 -4.17171977e-04
-1.71607697e+00 1.93454638e-01 1.29483595e-01 -1.17338133e+00
-1.74208134e-01 -2.45739415e-01 -5.78031540e-01 2.46500865e-01
-2.15286112e+00 -9.16498959e-01 -1.20592266e-01 6.31089151e-01
7.45728314e-01 -1.38100594e-01 6.49098635e-01 2.49462187e-01
-5.28345942e-01 8.59793127e-01 4.77406442e-01 -2.25932017e-01
7.58715570e-01 -1.52303874e+00 3.47467303e-01 9.36910927e-01
4.95668173e-01 4.14384812e-01 4.80513036e-01 -4.56459552e-01
-1.17494833e+00 -1.46194875e+00 7.02107072e-01 -4.32426721e-01
5.37410975e-01 -6.74922228e-01 -1.14854610e+00 7.12038398e-01
-2.16326848e-01 3.31006348e-01 7.84308553e-01 6.06577516e-01
-3.87613505e-01 -2.74897248e-01 -8.06304991e-01 3.62284005e-01
1.26246119e+00 -3.31570059e-01 -7.61843085e-01 4.13923621e-01
9.02650833e-01 -1.46372437e-01 -6.12016261e-01 6.45678222e-01
3.82303119e-01 -8.25112760e-01 9.42660570e-01 -9.27788675e-01
2.38503143e-01 -8.27887431e-02 3.05529740e-02 -1.17561460e+00
8.72281641e-02 -8.27847302e-01 -7.87428558e-01 1.42698348e+00
3.69640559e-01 -6.25247240e-01 8.98880720e-01 1.33785442e-01
-1.60829142e-01 -1.02165067e+00 -8.90009403e-01 -1.12958741e+00
2.84204125e-01 -5.79629064e-01 4.24270779e-01 1.11991262e+00
-1.91053465e-01 5.00267744e-01 -4.61688846e-01 5.44662066e-02
7.66482830e-01 9.00824070e-02 7.03794777e-01 -1.59693789e+00
-3.68481129e-01 -9.03498232e-02 -2.10808396e-01 -1.40896893e+00
4.34485227e-01 -8.71854782e-01 1.89832360e-01 -1.18389595e+00
4.08252418e-01 -6.58122897e-01 -8.75867426e-01 5.06170154e-01
-7.16451764e-01 4.71858233e-02 -7.42546394e-02 1.77409187e-01
-1.23895776e+00 7.97646224e-01 9.38136399e-01 9.14627388e-02
-2.29396880e-01 3.12022775e-01 -1.00698245e+00 7.35251129e-01
7.73532510e-01 -5.70114434e-01 -7.69768655e-01 -2.26525947e-01
2.37136096e-01 -3.16997200e-01 2.75712579e-01 -1.19808292e+00
3.80554765e-01 -1.18799962e-01 1.30500108e-01 -6.51419580e-01
1.41306296e-01 -6.96977735e-01 -5.69233060e-01 2.07285821e-01
-6.62737966e-01 -3.18946242e-01 1.36539377e-02 1.31003058e+00
-1.98207811e-01 -1.77542642e-01 7.11043417e-01 -6.49169907e-02
-6.90585136e-01 6.08630776e-01 4.36600335e-02 7.40049660e-01
1.04956090e+00 2.52712611e-03 -9.99887958e-02 -2.22946107e-01
-6.94893420e-01 4.62140739e-01 2.11597487e-01 5.11736274e-01
5.58875024e-01 -1.14142370e+00 -8.50188792e-01 -6.48486465e-02
5.87841868e-01 5.37846088e-01 6.01590984e-02 6.97055817e-01
2.31204987e-01 4.43630666e-01 2.81688184e-01 -6.57131970e-01
-7.88601696e-01 7.42256224e-01 -9.05972198e-02 -4.79321301e-01
-8.14255178e-01 1.01255500e+00 4.71328586e-01 -5.88560402e-01
7.20788121e-01 -1.66373223e-01 -3.37349445e-01 1.24826729e-01
5.78574359e-01 2.35905617e-01 1.62623778e-01 -2.29941204e-01
-1.88732013e-01 1.94085762e-01 -5.65363824e-01 -2.15236153e-02
1.46684325e+00 -3.02852213e-01 3.96942586e-01 8.14629197e-01
9.20537829e-01 -1.29132301e-01 -1.53766108e+00 -8.99307191e-01
4.53973889e-01 -3.26806933e-01 -1.01407878e-01 -9.55028296e-01
-7.56031275e-01 7.76700497e-01 5.37526309e-01 -7.67873675e-02
1.18038726e+00 1.83513865e-01 8.03192079e-01 7.59577990e-01
3.49232942e-01 -1.20799577e+00 3.27821553e-01 5.22135973e-01
3.92014116e-01 -1.48266053e+00 -8.98872912e-02 -1.40585095e-01
-4.66194272e-01 7.29663014e-01 7.55480886e-01 1.90801486e-01
7.52007067e-01 -1.05310522e-01 -1.47701845e-01 -6.35153204e-02
-1.06563783e+00 -3.95706385e-01 3.04381847e-01 5.08814096e-01
2.63784468e-01 -1.37369022e-01 -3.65645409e-01 7.96415389e-01
6.61392789e-03 -7.41579309e-02 2.93464333e-01 1.39135325e+00
-5.56903481e-01 -1.15134466e+00 -9.15760249e-02 5.98048270e-01
-5.68675518e-01 -1.39320627e-01 -2.75320441e-01 7.14027703e-01
2.81579286e-01 1.11516643e+00 -1.92246124e-01 -8.78418908e-02
2.77890056e-01 2.87827611e-01 3.15888792e-01 -1.00480008e+00
-3.08524162e-01 -1.62350506e-01 -3.25481296e-02 -3.49018037e-01
-6.70945883e-01 -8.49246383e-01 -1.00548422e+00 3.13956678e-01
-2.99147725e-01 4.73425984e-01 2.62490809e-01 1.24530625e+00
2.06120670e-01 4.15238708e-01 7.02886283e-01 -4.10353661e-01
-7.04238355e-01 -8.99168491e-01 -5.81136584e-01 3.03072214e-01
4.91729170e-01 -7.70249188e-01 -4.82611984e-01 1.75195664e-01] | [9.729674339294434, 3.2416365146636963] |
7b02d3ee-61d0-455f-839e-50bfc430c576 | copula-based-sensitivity-analysis-for-multi | 2102.09412 | null | https://arxiv.org/abs/2102.09412v3 | https://arxiv.org/pdf/2102.09412v3.pdf | Copula-based Sensitivity Analysis for Multi-Treatment Causal Inference with Unobserved Confounding | Recent work has focused on the potential and pitfalls of causal identification in observational studies with multiple simultaneous treatments. Building on previous work, we show that even if the conditional distribution of unmeasured confounders given treatments were known exactly, the causal effects would not in general be identifiable, although they may be partially identified. Given these results, we propose a sensitivity analysis method for characterizing the effects of potential unmeasured confounding, tailored to the multiple treatment setting, that can be used to characterize a range of causal effects that are compatible with the observed data. Our method is based on a copula factorization of the joint distribution of outcomes, treatments, and confounders, and can be layered on top of arbitrary observed data models. We propose a practical implementation of this approach making use of the Gaussian copula, and establish conditions under which causal effects can be bounded. We also describe approaches for reasoning about effects, including calibrating sensitivity parameters, quantifying robustness of effect estimates, and selecting models that are most consistent with prior hypotheses. | ['Alexander Franks', "Alexander D'Amour", 'Jiajing Zheng'] | 2021-02-18 | null | null | null | null | ['causal-identification'] | ['reasoning'] | [ 3.97775769e-01 1.61474332e-01 -7.70700872e-01 -3.30809474e-01
-7.72181749e-01 -7.75625408e-01 4.40185517e-01 4.58464921e-01
-1.88990235e-01 1.01336932e+00 6.68085754e-01 -6.01438403e-01
-8.08308005e-01 -7.99741745e-01 -8.79016995e-01 -5.69040835e-01
-2.84012258e-01 5.80540240e-01 -1.79627031e-01 4.85904455e-01
6.30590916e-02 3.15345585e-01 -1.24137974e+00 2.28159800e-02
1.01081109e+00 -5.18927202e-02 -2.19287023e-01 5.73439240e-01
3.92944187e-01 4.38908160e-01 -2.98168451e-01 -4.10052180e-01
-4.09220979e-02 -3.50855201e-01 -6.06684387e-01 -5.70290140e-04
2.51815557e-01 -4.56677020e-01 4.43500988e-02 8.25884700e-01
4.31652158e-01 -1.46511480e-01 1.05798924e+00 -1.55503798e+00
-7.10035563e-01 9.71000612e-01 -5.37721217e-01 4.21148501e-02
3.96844506e-01 -3.66846435e-02 9.23677742e-01 -4.89177227e-01
4.94952768e-01 1.66668034e+00 6.65634096e-01 3.09098929e-01
-1.74964857e+00 -7.43643820e-01 3.04375798e-01 -3.69514287e-01
-1.11199605e+00 -4.47444260e-01 9.28114057e-02 -6.06002688e-01
4.33206886e-01 3.09363931e-01 3.08405131e-01 1.13575053e+00
4.02757257e-01 3.52433324e-01 1.09532905e+00 -5.89125514e-01
3.36289793e-01 -2.52605863e-02 3.98345262e-01 2.21525550e-01
8.73347819e-01 5.88543713e-01 -1.89155519e-01 -1.04433191e+00
8.02060425e-01 5.28755374e-02 -3.57817173e-01 -5.57012022e-01
-1.17461622e+00 1.07041204e+00 -4.54317145e-02 -1.09175779e-01
-6.58670425e-01 3.44129294e-01 1.74678504e-01 -1.59477085e-01
3.99858832e-01 3.25348556e-01 -6.31169260e-01 2.23405987e-01
-5.15095890e-01 5.94596446e-01 6.54801905e-01 8.70528817e-01
4.28346008e-01 -5.90858698e-01 -4.41239506e-01 3.65503877e-01
1.75311491e-01 9.01762009e-01 -1.92367852e-01 -1.13139665e+00
2.38728762e-01 3.14472437e-01 6.92979932e-01 -5.20246804e-01
-5.78085542e-01 8.90722349e-02 -5.61839402e-01 -1.59394786e-01
7.35026062e-01 -5.91709971e-01 -8.42330933e-01 2.24100423e+00
3.79549563e-01 3.90688986e-01 -8.60368088e-02 4.27587479e-01
1.82669535e-01 1.66577280e-01 6.44609511e-01 -6.60232008e-01
1.24717200e+00 -1.34651959e-01 -8.01269174e-01 6.38395473e-02
6.80717111e-01 -6.24802232e-01 6.80119395e-01 -2.53729038e-02
-1.35006094e+00 1.48421386e-02 -4.34112579e-01 1.43154263e-01
7.95859247e-02 -1.59841970e-01 8.53507638e-01 9.80761766e-01
-9.24307525e-01 6.38254941e-01 -8.69742155e-01 -3.97634238e-01
1.35845691e-01 5.84329247e-01 -1.66138336e-01 -2.22681332e-02
-1.05923581e+00 8.03735077e-01 6.54459223e-02 -1.07101172e-01
-1.07294488e+00 -1.35168850e+00 -6.43096745e-01 4.30736572e-01
5.72676241e-01 -1.23173141e+00 1.16418242e+00 -7.48688817e-01
-7.19075084e-01 3.90044004e-01 -4.70172822e-01 -1.01548724e-01
3.01717818e-01 6.99746683e-02 -1.97298586e-01 -2.76212357e-02
4.33677256e-01 6.94190562e-02 3.88192654e-01 -1.25435936e+00
-6.15694702e-01 -5.97432494e-01 8.74466971e-02 1.74397409e-01
4.96273078e-02 5.25141835e-01 1.74069759e-02 -2.35117480e-01
-1.34166241e-01 -9.59538758e-01 -6.00036025e-01 -4.27432775e-01
-5.53825021e-01 3.36247794e-02 3.35942991e-02 -2.17430040e-01
1.22427571e+00 -1.92401135e+00 -2.82513630e-02 3.55272472e-01
1.43962309e-01 -4.94455904e-01 -6.61328882e-02 6.41383290e-01
-4.40197885e-01 5.15668690e-01 -2.33747303e-01 -1.62966907e-01
7.63612688e-02 1.88898608e-01 -4.51662511e-01 8.58489335e-01
1.20542705e-01 7.64238775e-01 -7.96461105e-01 -2.59814829e-01
1.01022966e-01 2.34560654e-01 -7.09328949e-01 4.62363437e-02
1.06794819e-01 3.97971004e-01 -5.78426123e-01 2.42954999e-01
8.30755889e-01 -3.39421183e-01 7.22049475e-01 3.42211664e-01
-2.60227472e-01 2.06064165e-01 -1.45304179e+00 6.84827566e-01
-2.28604496e-01 -1.31970391e-01 1.72037974e-01 -1.00202048e+00
2.42325500e-01 6.97599471e-01 6.52332067e-01 3.72566171e-02
6.45840690e-02 3.50294448e-02 2.73768846e-02 -5.21091759e-01
4.05665189e-02 -6.32389963e-01 -1.59856051e-01 5.66650271e-01
-1.29247993e-01 3.07605147e-01 -1.95120815e-02 1.54546067e-01
1.08622730e+00 -2.70095497e-01 6.32850766e-01 -6.29179955e-01
-3.39305662e-02 -8.42663273e-02 6.72426164e-01 1.12655962e+00
-6.13560202e-03 4.87182498e-01 9.73832190e-01 1.32869989e-01
-9.91745353e-01 -1.28837907e+00 -5.67800701e-01 7.09064066e-01
-1.25832036e-01 -9.53793228e-02 -5.45044661e-01 -3.31770927e-01
4.69861239e-01 7.90970385e-01 -1.11840701e+00 1.65082663e-02
-1.41897976e-01 -1.42242384e+00 4.70148087e-01 6.33448184e-01
-4.30339754e-01 -5.46227396e-01 -2.95920193e-01 1.27970353e-01
1.77254491e-02 -6.38496816e-01 -3.41541886e-01 5.30523323e-02
-1.00969720e+00 -1.64109564e+00 -5.72200537e-01 -1.45210966e-01
6.98213339e-01 2.48931482e-01 1.02747619e+00 -5.12334006e-03
1.81030124e-01 6.93528116e-01 9.31173265e-02 -7.28338003e-01
-4.98237938e-01 -5.19338071e-01 2.48859733e-01 -1.74119428e-01
3.83576810e-01 -5.78889310e-01 -6.74525917e-01 1.23574875e-01
-8.37006927e-01 -3.48323077e-01 3.08040559e-01 8.48644376e-01
3.50074112e-01 -5.83439730e-02 7.43974388e-01 -1.26118267e+00
6.49612069e-01 -9.31788445e-01 -7.64108658e-01 5.70391178e-01
-8.01916301e-01 -2.72019636e-02 2.76866466e-01 -6.33084357e-01
-1.29208875e+00 -2.25041255e-01 5.17119348e-01 -8.31048861e-02
-1.68242007e-01 7.79218972e-01 -3.96238267e-01 2.59756625e-01
6.35111868e-01 -7.19709992e-01 -2.29430795e-02 -4.23030615e-01
4.90287662e-01 4.43613678e-01 1.36171862e-01 -8.82300556e-01
2.28732020e-01 7.06048250e-01 2.44551718e-01 -2.16383293e-01
-5.95747530e-01 -3.04295480e-01 -3.34525555e-01 3.70842785e-01
8.14307511e-01 -8.67334783e-01 -1.13933539e+00 -2.09403019e-02
-9.14913952e-01 -3.04533243e-01 -1.96925685e-01 1.10243797e+00
-5.91607213e-01 1.98817313e-01 -3.64003688e-01 -1.14677489e+00
2.46247321e-01 -1.17001104e+00 9.59298670e-01 -1.67316534e-02
-4.30148840e-01 -1.46331918e+00 3.44805658e-01 1.70692205e-02
-2.75321100e-02 2.78025150e-01 1.41443360e+00 -6.54598117e-01
-3.93446207e-01 -1.29009828e-01 -2.78166920e-01 -3.40239525e-01
2.34206513e-01 3.20236892e-01 -7.60446131e-01 -2.37004653e-01
-1.58226199e-03 1.67892128e-01 4.66687709e-01 1.58469093e+00
8.49985123e-01 -3.84423703e-01 -8.59083056e-01 1.81608453e-01
1.42803442e+00 1.84110835e-01 3.83145869e-01 1.29326116e-02
3.85503054e-01 8.01976442e-01 3.01358640e-01 5.45951366e-01
5.80208421e-01 5.41399360e-01 1.81837305e-01 -2.22682729e-01
4.94240820e-01 -4.56341654e-01 3.33575830e-02 -2.11637035e-01
-1.19235390e-03 -2.66177535e-01 -7.38054514e-01 8.77205491e-01
-1.89554453e+00 -1.00822377e+00 -6.74204230e-01 2.65109062e+00
7.72169173e-01 -1.86774805e-01 5.34792185e-01 -2.72886425e-01
1.00327682e+00 -5.93534291e-01 -4.21613485e-01 -4.98491347e-01
-7.29155093e-02 2.69599110e-01 1.17733860e+00 7.62011647e-01
-8.48538399e-01 6.55910522e-02 8.79020500e+00 2.53440917e-01
-4.81775492e-01 1.35580942e-01 5.64504743e-01 9.91267338e-02
-8.20139050e-01 6.93511486e-01 -6.69258535e-01 3.51348430e-01
1.19165850e+00 -8.29277396e-01 3.82720120e-02 3.46427768e-01
9.19655561e-01 -3.59688342e-01 -1.42960799e+00 -3.26313414e-02
-5.55392385e-01 -9.61835086e-01 -1.17371783e-01 4.15422112e-01
9.43298340e-01 -4.19814676e-01 7.49666467e-02 -9.41431001e-02
1.05439746e+00 -1.17918360e+00 3.06177139e-01 3.31454009e-01
7.18041539e-01 -6.19337499e-01 6.01527810e-01 3.27956408e-01
-5.86192966e-01 -3.29131812e-01 -3.86207402e-01 -3.53481889e-01
1.24201171e-01 8.50224614e-01 -7.98354030e-01 7.35846460e-01
2.92417526e-01 4.42260057e-01 8.16809237e-02 1.15486968e+00
-3.03005129e-01 9.85168099e-01 -3.05141538e-01 4.40273076e-01
-2.61160016e-01 -5.76811656e-02 3.46833497e-01 8.54880691e-01
4.53890681e-01 4.18647915e-01 -3.08965240e-02 9.57931757e-01
1.30852208e-01 2.16020122e-01 -6.99529231e-01 1.53609455e-01
6.14545882e-01 6.37495339e-01 -5.08267939e-01 -2.74933010e-01
-5.95749497e-01 1.04093842e-01 -1.78717840e-02 6.50611162e-01
-7.13712633e-01 4.61169273e-01 6.03996933e-01 4.07789797e-02
-1.12152576e-01 2.80972481e-01 -5.83117425e-01 -1.13737774e+00
-1.59924462e-01 -8.44973862e-01 8.21522474e-01 -6.38182700e-01
-1.42379880e+00 -5.17328382e-01 7.79976070e-01 -7.19217777e-01
-1.30586103e-01 -3.27457458e-01 -6.84594095e-01 1.25169337e+00
-9.18422043e-01 -9.67802584e-01 4.19914007e-01 5.34667671e-01
-1.81486651e-01 6.01016581e-01 1.04492474e+00 -1.55747207e-02
-5.78904629e-01 3.47356945e-01 3.38036984e-01 -4.68112856e-01
8.61682892e-01 -1.34365106e+00 -4.81182858e-02 7.14262247e-01
-5.09368062e-01 1.04660583e+00 9.71409738e-01 -1.16167319e+00
-1.15125465e+00 -7.40946054e-01 9.12593722e-01 -6.50176406e-01
1.04476976e+00 -1.91126734e-01 -7.72938907e-01 1.27713752e+00
-1.22202688e-03 -5.88631034e-01 8.57384682e-01 9.73599195e-01
-2.38746911e-01 1.44399971e-01 -1.10136902e+00 7.58967698e-01
9.17531013e-01 -9.12408233e-02 -4.65621561e-01 4.55758214e-01
6.11714065e-01 -1.49904102e-01 -9.32293952e-01 5.83751440e-01
6.45709157e-01 -6.95257366e-01 9.32855189e-01 -1.14693069e+00
3.97494376e-01 4.92515378e-02 2.75455229e-02 -1.31285107e+00
-4.97370094e-01 -5.32820344e-01 5.22831202e-01 1.11174893e+00
6.26703441e-01 -6.91968143e-01 3.78865421e-01 1.44193113e+00
3.41777325e-01 -1.56059071e-01 -8.38801801e-01 -6.49489343e-01
6.18544161e-01 -3.19899291e-01 7.26583600e-01 1.07985032e+00
2.95320094e-01 2.16639638e-01 -4.64136302e-01 7.06808746e-01
7.98955977e-01 1.66086778e-01 5.54146230e-01 -1.29044592e+00
-4.24538523e-01 -6.35626316e-02 -1.24278128e-01 -3.57116938e-01
1.61328748e-01 -4.87215728e-01 -2.52023697e-01 -1.35375810e+00
1.04756010e+00 -5.62196612e-01 -1.15984902e-01 4.45833832e-01
-6.73100591e-01 -4.39382523e-01 -1.63819268e-01 -2.51890998e-03
6.89824522e-02 3.87347229e-02 1.08819318e+00 1.74063087e-01
-3.23731750e-01 2.42850065e-01 -1.14827108e+00 6.21891141e-01
5.41951597e-01 -7.38947153e-01 -6.27851188e-01 -1.77534536e-01
2.21303418e-01 6.82924807e-01 6.74480081e-01 -3.92734669e-02
2.20642164e-02 -8.25497210e-01 2.03433484e-01 -2.13395417e-01
-2.07780674e-01 -8.00763845e-01 8.05034816e-01 3.65447998e-01
-7.03185022e-01 3.10448073e-02 3.77354711e-01 6.36271715e-01
2.21391857e-01 -4.04407293e-01 3.31746697e-01 8.20099935e-02
2.26565391e-01 1.04155973e-01 -3.94767195e-01 -6.11567125e-02
9.35178518e-01 2.61803895e-01 -5.83377063e-01 -5.20435452e-01
-9.64111507e-01 4.57524061e-01 6.34082496e-01 -3.73184271e-02
1.32468358e-01 -1.16412854e+00 -8.66046190e-01 -2.27849111e-01
-6.59057125e-02 -5.60478151e-01 5.32647848e-01 1.06255758e+00
2.48135731e-01 4.87182170e-01 1.98138580e-01 -2.22227290e-01
-1.31357718e+00 1.24609423e+00 3.13820690e-01 -2.74731785e-01
-2.76006073e-01 1.01435095e-01 9.53184724e-01 -1.47745728e-01
-1.38172716e-01 -2.60108829e-01 4.50441651e-02 -1.65447250e-01
5.62828541e-01 5.10572016e-01 -3.60823572e-01 -2.27575198e-01
-4.99833435e-01 3.31770837e-01 2.77801901e-01 -3.73097360e-01
1.19892597e+00 -4.61928099e-01 -3.86851311e-01 3.79626513e-01
8.74776185e-01 5.23659587e-01 -1.03935540e+00 1.18096687e-01
-3.26502286e-02 -4.13272768e-01 -3.01856041e-01 -7.80769467e-01
-5.52294016e-01 4.81722832e-01 2.42749929e-01 4.36654657e-01
1.15945029e+00 -4.28577997e-02 -3.54097426e-01 -3.76021415e-01
1.10502183e-01 -4.73000675e-01 -7.12620497e-01 -2.40664870e-01
5.87395430e-01 -1.05405915e+00 2.81828582e-01 -8.17517817e-01
-3.31271768e-01 7.84874082e-01 -2.14335267e-02 -7.82822296e-02
7.99668729e-01 3.79636228e-01 -1.43435627e-01 -2.58247465e-01
-8.96940947e-01 -1.22214094e-01 2.30807558e-01 7.68970728e-01
4.18849021e-01 5.18735111e-01 -9.18185413e-01 6.93708718e-01
1.68471277e-01 2.99796402e-01 9.93035674e-01 6.21973515e-01
7.70913810e-02 -1.30320251e+00 -8.76345515e-01 5.63138366e-01
-9.14060652e-01 -1.67135850e-01 -1.73386455e-01 9.87920344e-01
-4.32827696e-02 1.30895007e+00 5.85250370e-02 3.04020584e-01
5.35827696e-01 -4.54665236e-02 4.28044736e-01 -6.53529704e-01
-2.38507196e-01 3.64022672e-01 2.00408369e-01 -3.11538786e-01
-8.63608837e-01 -1.10196984e+00 -8.51963758e-01 -6.26480877e-01
-5.43729246e-01 3.08575988e-01 1.31310478e-01 1.00177622e+00
2.30235979e-01 2.66405851e-01 6.72902703e-01 -2.31744409e-01
-7.09508359e-01 -5.91071904e-01 -7.63611019e-01 3.70791078e-01
3.41905326e-01 -8.97826612e-01 -4.96336818e-01 8.75607785e-03] | [7.93298864364624, 5.266634941101074] |
a61c6cb1-5115-4b13-9fd6-d130ea1f009d | inferring-methodological-meta-knowledge-from | null | null | https://aclanthology.org/Y16-1002 | https://aclanthology.org/Y16-1002.pdf | Inferring Methodological Meta-knowledge from Large Biomedical Corpora | null | ['Goran Nenadic'] | 2016-10-01 | inferring-methodological-meta-knowledge-from-1 | https://aclanthology.org/Y16-1002 | https://aclanthology.org/Y16-1002.pdf | paclic-2016-10 | ['temporal-information-extraction'] | ['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] | [-7.3386149406433105, 3.8228201866149902] |
c2f57446-191d-49cc-80cc-bb38c636249f | low-compute-and-fully-parallel-computer | null | null | http://openaccess.thecvf.com/content_iccv_2017/html/Fanello_Low_Compute_and_ICCV_2017_paper.html | http://openaccess.thecvf.com/content_ICCV_2017/papers/Fanello_Low_Compute_and_ICCV_2017_paper.pdf | Low Compute and Fully Parallel Computer Vision With HashMatch | Numerous computer vision problems such as stereo depth estimation, object-class segmentation and foreground/background segmentation can be formulated as per-pixel image labeling tasks. Given one or many images as input, the desired output of these methods is usually a spatially smooth assignment of labels. The large amount of such computer vision problems has lead to significant research efforts, with the state of art moving from CRF-based approaches to deep CNNs and more recently, hybrids of the two. Although these approaches have significantly advanced the state of the art, the vast majority has solely focused on improving quantitative results and are not designed for low-compute scenarios. In this paper, we present a new general framework for a variety of computer vision labeling tasks, called HashMatch. Our approach is designed to be both fully parallel, i.e. each pixel is independently processed, and low-compute, with a model complexity an order of magnitude less than existing CNN and CRF-based approaches. We evaluate HashMatch extensively on several problems such as disparity estimation, image retrieval, feature approximation and background subtraction, for which HashMatch achieves high computational efficiency while producing high quality results.
| ['Sean Ryan Fanello', 'Julien Valentin', 'Carlo Ciliberto', 'Philip Davidson', 'Vladimir Tankovich', 'Shahram Izadi', 'Christoph Rhemann', 'Adarsh Kowdle'] | 2017-10-01 | null | null | null | iccv-2017-10 | ['stereo-depth-estimation'] | ['computer-vision'] | [ 5.49823999e-01 -5.62820286e-02 -1.09700322e-01 -4.75401849e-01
-9.94581461e-01 -5.56086600e-01 5.91801167e-01 2.84624875e-01
-5.60262501e-01 5.32481194e-01 -3.82133126e-01 -4.10179675e-01
4.18708891e-01 -7.72228181e-01 -5.71747184e-01 -6.67653143e-01
4.71683294e-01 6.24604762e-01 7.63150036e-01 1.94339320e-01
4.84136879e-01 7.38179862e-01 -1.81868410e+00 -7.71039650e-02
6.22266948e-01 1.26583672e+00 1.75477341e-02 8.57501447e-01
-4.75839108e-01 7.14698374e-01 -2.57492870e-01 -4.64434922e-01
3.58023524e-01 -3.16014200e-01 -9.40093219e-01 5.34578025e-01
9.66244757e-01 -4.36530143e-01 -1.68799683e-01 1.24344206e+00
4.14119124e-01 1.39382124e-01 4.56451207e-01 -1.19495511e+00
-1.08661100e-01 8.79946724e-02 -7.84557998e-01 2.00245708e-01
-3.45412306e-02 2.12667305e-02 6.70501471e-01 -8.20073962e-01
5.58092594e-01 1.25160146e+00 6.77045465e-01 4.93608713e-01
-1.36702716e+00 -2.97309101e-01 2.10031405e-01 2.61933524e-02
-1.45105994e+00 -3.54800969e-01 5.60198545e-01 -6.62985086e-01
8.41522276e-01 2.14733444e-02 6.06101930e-01 1.53126225e-01
-1.38657570e-01 8.95274341e-01 1.38342822e+00 -6.47375464e-01
3.68581980e-01 -4.45112810e-02 2.14791417e-01 8.30957055e-01
7.35411420e-02 -6.26761466e-02 -3.37176830e-01 5.80113232e-02
1.05670214e+00 4.72648703e-02 -6.74685463e-02 -5.16372323e-01
-1.14586461e+00 8.85879397e-01 2.98183233e-01 -1.38374735e-02
-1.94470152e-01 3.12210560e-01 3.22573841e-01 -7.34205320e-02
6.72899365e-01 -1.16419300e-01 -3.08205694e-01 6.15354739e-02
-1.37663805e+00 6.40145838e-01 8.56867850e-01 9.41012204e-01
1.28872383e+00 -2.27399066e-01 7.14124739e-02 6.85354173e-01
1.60709009e-01 6.00726068e-01 5.35808280e-02 -1.41506290e+00
1.95314124e-01 5.14791846e-01 1.13151073e-01 -1.00726581e+00
-4.41299438e-01 8.25837255e-02 -7.45323777e-01 4.75870371e-01
7.08059728e-01 -4.82667312e-02 -1.35317528e+00 1.56602597e+00
6.07771218e-01 1.12131245e-01 -2.69963533e-01 9.09913301e-01
7.11180091e-01 7.39594996e-01 -8.71646330e-02 -7.53527433e-02
1.15193498e+00 -1.07489729e+00 -3.98608327e-01 -4.68976140e-01
3.66442263e-01 -1.06134653e+00 5.91706216e-01 6.09633088e-01
-1.39069211e+00 -4.75642890e-01 -8.14448595e-01 -4.77205455e-01
-3.67705286e-01 -1.82605982e-01 8.10927629e-01 7.90758908e-01
-1.38897681e+00 5.37655234e-01 -9.08500433e-01 -4.07068461e-01
7.43404567e-01 5.98672628e-01 -8.13530087e-02 -4.46286052e-01
-5.48501074e-01 8.07312787e-01 3.74476373e-01 -1.62783749e-02
-7.51565039e-01 -4.09513235e-01 -9.27327752e-01 -1.91465974e-01
4.26317722e-01 -7.46666133e-01 1.55470133e+00 -1.13299632e+00
-1.43349373e+00 1.41151619e+00 -3.33480656e-01 -4.80837107e-01
6.00256264e-01 -5.19566983e-02 2.41208494e-01 1.66715607e-01
1.24625623e-01 1.27660728e+00 6.79752171e-01 -1.19317365e+00
-1.14112699e+00 -5.80097139e-01 8.40721000e-03 3.20749253e-01
2.35452615e-02 2.49181613e-01 -9.59548473e-01 -3.45341742e-01
2.19327912e-01 -9.19632375e-01 -6.54133320e-01 4.67897207e-01
-1.58034697e-01 -1.64850056e-01 7.40459561e-01 -3.59544098e-01
7.09056079e-01 -1.92335308e+00 4.45924886e-02 -9.24473777e-02
2.98560917e-01 3.28025430e-01 2.08330289e-01 -1.79779734e-02
3.80410850e-01 -1.81101695e-01 -6.17747545e-01 -6.06560171e-01
-1.31707475e-01 4.25945818e-01 -1.63086802e-01 8.62465084e-01
2.73542374e-01 8.58395875e-01 -7.54685223e-01 -8.85088861e-01
6.75379992e-01 5.66604495e-01 -4.75263149e-01 1.47682428e-01
-4.54278857e-01 3.56510222e-01 -2.68877745e-01 8.28794241e-01
9.91985977e-01 -3.52817476e-01 -7.36904666e-02 -1.31695559e-02
-5.39546371e-01 1.08867742e-01 -1.22499979e+00 1.74312925e+00
-1.54896826e-01 8.54303777e-01 4.12233919e-01 -1.20282567e+00
7.53943741e-01 -5.07006794e-02 6.74582362e-01 -6.01679504e-01
2.74115473e-01 3.32858980e-01 -3.25720400e-01 -1.03467844e-01
7.01338053e-01 -8.50305259e-02 2.11457208e-01 2.02731311e-01
-1.79309070e-01 -4.81453836e-01 2.85219193e-01 3.18859443e-02
9.57858443e-01 2.57636398e-01 4.29800957e-01 -3.51774871e-01
5.01946390e-01 4.93044704e-01 6.47718191e-01 7.54871428e-01
-4.05144751e-01 1.08425999e+00 2.71013141e-01 -4.34639215e-01
-1.01185405e+00 -9.67215538e-01 -3.13574016e-01 8.78341794e-01
5.62395990e-01 1.59832407e-02 -1.14489698e+00 -3.41310859e-01
-9.11460221e-02 5.94560988e-02 -3.47684890e-01 5.23185670e-01
-8.67620230e-01 -6.40358686e-01 3.38026583e-01 6.65413976e-01
7.80740261e-01 -1.08471680e+00 -1.03464925e+00 3.34541142e-01
-2.72028204e-02 -1.36488473e+00 -3.57339352e-01 3.63318831e-01
-1.05989552e+00 -9.86034930e-01 -9.39112425e-01 -1.01600635e+00
5.98622918e-01 4.46832180e-01 1.39867651e+00 6.42131269e-02
-5.19867539e-01 3.45730722e-01 -3.56352367e-02 -3.51089478e-01
5.94330840e-02 -2.07199529e-02 -4.88799155e-01 3.99165638e-02
5.44224322e-01 -3.73294175e-01 -6.98253810e-01 1.84762925e-01
-1.20830190e+00 1.91802129e-01 5.73227406e-01 7.04324067e-01
1.03381741e+00 -1.57862708e-01 4.23194766e-02 -1.13478136e+00
6.39626458e-02 -2.05047522e-02 -1.03765130e+00 8.63033980e-02
-4.33549851e-01 -2.54778177e-01 3.45092952e-01 -1.73898146e-01
-9.28020656e-01 5.57126164e-01 -3.45878392e-01 -4.02002931e-01
-5.42244852e-01 2.04291537e-01 -1.78722754e-01 -3.51595581e-01
2.83865333e-01 2.32808694e-01 -9.47110355e-02 -3.56394440e-01
5.15296400e-01 5.26641905e-01 9.06064749e-01 -5.12825012e-01
5.16964197e-01 8.84289801e-01 2.15440571e-01 -9.26863968e-01
-8.73925507e-01 -8.82726312e-01 -8.50638926e-01 -2.92976260e-01
1.25344193e+00 -8.60943854e-01 -4.26468879e-01 8.68647635e-01
-1.39542413e+00 -5.98019898e-01 -1.86235562e-01 6.29245490e-02
-7.50612676e-01 4.02357996e-01 -5.97182453e-01 -7.71544874e-01
-1.16271101e-01 -1.40080822e+00 1.50906658e+00 5.47185838e-01
2.10066751e-01 -1.10504711e+00 -5.51351681e-02 3.45256448e-01
4.07837927e-01 3.76055121e-01 7.42428005e-01 -2.10877106e-01
-1.15712857e+00 -1.44943416e-01 -7.55051255e-01 3.18706632e-01
-2.02657551e-01 -8.52209479e-02 -1.18663239e+00 -3.52057628e-02
-2.76045445e-02 -5.09527087e-01 9.50133502e-01 8.95446599e-01
1.19398153e+00 2.33196288e-01 -4.32945907e-01 7.53604710e-01
1.68501890e+00 2.01382369e-01 7.88454175e-01 2.53286213e-01
8.10146391e-01 7.09432483e-01 7.47632325e-01 1.55986831e-01
5.74382901e-01 7.80261517e-01 5.33024251e-01 -6.23737812e-01
-2.92387396e-01 2.08317265e-01 -9.91851240e-02 5.66160381e-01
1.22731537e-01 -2.28295639e-01 -1.08941424e+00 6.70771658e-01
-2.00104380e+00 -7.43081868e-01 -2.53282309e-01 2.14566541e+00
5.28766215e-01 9.48220491e-02 1.54908597e-01 1.44127935e-01
8.55370879e-01 9.27633345e-02 -6.29335701e-01 -2.86218822e-01
-1.87337935e-01 3.90669048e-01 7.24858046e-01 5.49537599e-01
-1.41792452e+00 1.13345253e+00 6.29915571e+00 7.72388816e-01
-1.15214205e+00 3.76206413e-02 1.08847916e+00 2.77146399e-01
9.11738425e-02 9.02209878e-02 -9.79400396e-01 1.71936035e-01
5.09762526e-01 1.96517035e-01 2.36821726e-01 9.10925746e-01
-1.43067077e-01 -6.77015185e-01 -1.11629891e+00 1.32834268e+00
1.68366343e-01 -1.46921778e+00 -1.45542964e-01 1.43481210e-01
1.04077077e+00 3.13078225e-01 -1.77480020e-02 -6.60099909e-02
2.99747407e-01 -9.95787621e-01 7.16843963e-01 2.66189158e-01
6.64775789e-01 -6.45056486e-01 6.28185034e-01 2.84559697e-01
-1.16657186e+00 3.12349319e-01 -3.53868723e-01 -1.46962807e-01
3.82988781e-01 7.50253618e-01 -3.36503655e-01 2.48541176e-01
7.44428158e-01 6.09251380e-01 -4.53174293e-01 1.24238741e+00
1.99936822e-01 3.33352178e-01 -4.90248293e-01 7.61187449e-02
4.24375266e-01 -2.93714464e-01 1.49635896e-02 1.43083036e+00
8.28082580e-03 2.52683133e-01 7.33761370e-01 7.62808800e-01
-1.25542358e-01 4.02540714e-03 -5.13300896e-01 3.18290323e-01
2.51716495e-01 1.36863041e+00 -1.43604147e+00 -6.15673423e-01
-7.88671672e-01 8.85164976e-01 2.75922149e-01 1.83658183e-01
-6.80039048e-01 -2.61204332e-01 6.63089156e-01 2.55171806e-01
5.41954458e-01 -3.78623426e-01 -6.50212884e-01 -1.00499225e+00
-4.18466181e-02 -7.02109635e-01 1.62426755e-01 -5.93892813e-01
-1.18656003e+00 4.80773896e-01 -4.61529307e-02 -9.03926373e-01
-9.92803872e-02 -8.81487370e-01 -3.42346936e-01 8.11381102e-01
-1.87968040e+00 -1.02074182e+00 -5.43064654e-01 6.62957847e-01
7.47684956e-01 4.14688736e-01 5.44794500e-01 6.50267959e-01
-4.18289483e-01 -6.72715483e-03 -3.19821155e-03 1.54267862e-01
4.53451425e-01 -1.41142845e+00 4.85693753e-01 1.15724051e+00
5.51598631e-02 1.89923346e-01 3.99281293e-01 -3.48615676e-01
-1.15618742e+00 -1.17865705e+00 8.87794495e-01 -3.82738620e-01
3.18551511e-01 -3.16965133e-01 -7.82164931e-01 5.59012711e-01
1.16136134e-01 4.50901240e-01 9.50036794e-02 -3.23819876e-01
2.08092406e-02 2.80070473e-02 -1.10183525e+00 4.78706241e-01
9.29327846e-01 -3.92543495e-01 -1.38009861e-01 3.59466016e-01
3.26037318e-01 -7.50364125e-01 -4.73693669e-01 3.71838629e-01
2.78726518e-01 -1.20897233e+00 9.97491956e-01 -1.30486503e-01
4.82413948e-01 -5.64021170e-01 -3.34355593e-01 -6.64366484e-01
2.54386738e-02 -2.88853437e-01 1.75225377e-01 1.21148252e+00
2.77779810e-02 -3.49420518e-01 1.15230596e+00 7.88309753e-01
-2.23651215e-01 -7.11154580e-01 -8.90302896e-01 -5.44665754e-01
6.04585186e-02 -3.96572500e-01 1.82869732e-01 6.41355038e-01
-7.38042593e-01 2.54378110e-01 -1.56062499e-01 9.62327980e-03
8.79403770e-01 5.07256269e-01 9.38707471e-01 -1.28163958e+00
-1.21470787e-01 -7.29102850e-01 -6.29277706e-01 -1.67965925e+00
-5.81184519e-04 -6.85074389e-01 5.37631631e-01 -1.64073682e+00
2.33078644e-01 -4.86565173e-01 2.01604411e-01 3.70137900e-01
-8.34552273e-02 6.42000616e-01 -3.17806117e-02 1.05100490e-01
-8.07921171e-01 1.23724572e-01 9.91519332e-01 -1.81462660e-01
-2.85030734e-02 7.76064768e-02 -2.79975533e-01 9.71761048e-01
6.63535416e-01 -4.22267258e-01 -1.72412574e-01 -6.15153551e-01
-1.28142923e-01 1.21347234e-01 5.24694026e-01 -1.07316065e+00
4.13761169e-01 -3.21695119e-01 2.80453891e-01 -7.73760855e-01
3.34994406e-01 -5.99837542e-01 -1.23326421e-01 2.15097293e-01
-6.23234436e-02 1.76320896e-01 1.67419463e-01 3.91730100e-01
-4.94334996e-01 -2.36086875e-01 1.26438904e+00 -3.08447778e-01
-1.08729768e+00 4.92130160e-01 -3.84424448e-01 2.85903662e-01
9.41264510e-01 -4.66267556e-01 -1.06927738e-01 -1.66912109e-01
-5.23174047e-01 -2.36585420e-02 7.33918428e-01 -9.92405862e-02
5.66310525e-01 -9.29811597e-01 -5.62614381e-01 -6.52553812e-02
-5.73422685e-02 6.33993804e-01 1.38292238e-02 7.76486516e-01
-1.04607749e+00 5.84694564e-01 -1.42611861e-01 -1.05832744e+00
-1.28171968e+00 4.17566478e-01 3.32691967e-01 -1.90384433e-01
-6.88250899e-01 1.01507556e+00 4.85192955e-01 -1.74930379e-01
4.46539134e-01 -2.85825789e-01 1.24162465e-01 -1.79028869e-01
4.63985890e-01 3.35157156e-01 1.64348975e-01 -5.52461803e-01
-4.06513304e-01 9.39865887e-01 -5.44463135e-02 -1.90095797e-01
1.04724741e+00 -2.74630755e-01 -2.52850443e-01 2.97654182e-01
1.08317232e+00 -3.63942325e-01 -1.58428741e+00 -4.09857303e-01
1.55992910e-01 -6.78809941e-01 4.25632924e-01 -4.35757965e-01
-1.40482259e+00 1.20166123e+00 7.09009647e-01 1.50290802e-01
1.36609089e+00 5.68902008e-02 8.23832333e-01 2.68576533e-01
5.74183702e-01 -9.99977231e-01 -9.65779424e-02 4.70092028e-01
1.24652959e-01 -1.59419787e+00 8.90436918e-02 -6.47179723e-01
-3.02505016e-01 1.04539347e+00 6.23347521e-01 -2.86230713e-01
6.62588835e-01 5.69608748e-01 2.32342452e-01 -1.88645303e-01
-2.47939914e-01 -6.61384463e-01 1.32191658e-01 6.64335608e-01
4.72259492e-01 -2.03236714e-01 -1.67019889e-01 -2.45559976e-01
2.51712352e-01 6.66905120e-02 3.75428379e-01 1.06284618e+00
-6.85431659e-01 -1.15826452e+00 -3.86982560e-01 3.36337745e-01
-5.15001714e-01 -2.15617940e-01 -4.08044845e-01 8.15457880e-01
2.23936245e-01 8.22073162e-01 4.42475498e-01 2.20256478e-01
1.60458177e-01 -1.44469216e-01 6.88566089e-01 -6.55414820e-01
-4.75210220e-01 1.92613780e-01 -1.88038915e-01 -6.00364149e-01
-9.40353394e-01 -8.54717493e-01 -1.25825822e+00 -3.68388832e-01
-4.26811486e-01 -2.80465990e-01 7.93346882e-01 9.67664897e-01
-5.08787483e-02 2.00318441e-01 2.78375864e-01 -1.29622674e+00
-2.27971911e-01 -4.76924270e-01 -5.59135258e-01 2.53563970e-01
3.19491953e-01 -5.93060136e-01 -9.78367403e-03 4.15333360e-01] | [9.484844207763672, 0.2207546979188919] |
3d150ad0-64e5-4690-8754-296a440e8c24 | 3d-regression-neural-network-for-the | 1802.05914 | null | http://arxiv.org/abs/1802.05914v2 | http://arxiv.org/pdf/1802.05914v2.pdf | 3D Regression Neural Network for the Quantification of Enlarged Perivascular Spaces in Brain MRI | Enlarged perivascular spaces (EPVS) in the brain are an emerging imaging
marker for cerebral small vessel disease, and have been shown to be related to
increased risk of various neurological diseases, including stroke and dementia.
Automatic quantification of EPVS would greatly help to advance research into
its etiology and its potential as a risk indicator of disease. We propose a
convolutional network regression method to quantify the extent of EPVS in the
basal ganglia from 3D brain MRI. We first segment the basal ganglia and
subsequently apply a 3D convolutional regression network designed for small
object detection within this region of interest. The network takes an image as
input, and outputs a quantification score of EPVS. The network has
significantly more convolution operations than pooling ones and no final
activation, allowing it to span the space of real numbers. We validated our
approach using a dataset of 2000 brain MRI scans scored visually. Experiments
with varying sizes of training and test sets showed that a good performance can
be achieved with a training set of only 200 scans. With a training set of 1000
scans, the intraclass correlation coefficient (ICC) between our scoring method
and the expert's visual score was 0.74. Our method outperforms by a large
margin - more than 0.10 - four more conventional automated approaches based on
intensities, scale-invariant feature transform, and random forest. We show that
the network learns the structures of interest and investigate the influence of
hyper-parameters on the performance. We also evaluate the reproducibility of
our network using a set of 60 subjects scanned twice (scan-rescan
reproducibility). On this set our network achieves an ICC of 0.93, while the
intrarater agreement reaches 0.80. Furthermore, the automatic EPVS scoring
correlates similarly to age as visual scoring. | ['Wiro Niessen', 'Gerda Bortsova', 'Florian Dubost', 'Meike Vernooij', 'Marleen de Bruijne', 'M. Arfan Ikram', 'Hieab Adams'] | 2018-02-16 | null | null | null | null | ['small-object-detection'] | ['computer-vision'] | [-1.13287330e-01 1.05001613e-01 -5.09443134e-02 -6.68432117e-01
-7.52536535e-01 -4.02193218e-01 4.79721069e-01 1.49408638e-01
-1.09835410e+00 6.27862453e-01 1.86390817e-01 -2.86198944e-01
1.44502625e-01 -8.38759065e-01 -3.60648781e-01 -5.48997819e-01
-7.05700278e-01 4.49578673e-01 6.80216789e-01 2.72255749e-01
1.16136156e-01 8.49924922e-01 -8.93431127e-01 1.75123230e-01
6.14552796e-01 1.06018543e+00 2.52515614e-01 5.46174288e-01
-1.01697231e-02 4.57359105e-01 -7.22193360e-01 -2.07551047e-01
4.36416358e-01 -1.72861412e-01 -7.12339520e-01 -1.85434267e-01
4.88205701e-01 -5.57491243e-01 -3.14562321e-01 9.26600337e-01
8.09405327e-01 -6.05704486e-02 9.31901872e-01 -8.03330243e-01
-4.36977327e-01 3.91024292e-01 -7.59695172e-01 9.58227694e-01
-2.66307473e-01 2.58631766e-01 6.10131919e-01 -6.53348327e-01
6.80660844e-01 8.83548439e-01 6.76731050e-01 4.05719459e-01
-1.39629853e+00 -8.02502453e-01 -3.90117526e-01 3.98817271e-01
-1.36928689e+00 -1.92140356e-01 1.55032218e-01 -8.61929655e-01
8.52954805e-01 -3.87439579e-02 9.42850709e-01 7.10426092e-01
4.38102454e-01 2.43537575e-01 1.49219620e+00 -1.60789385e-01
3.67487758e-01 -1.19237021e-01 3.19434762e-01 7.17273712e-01
2.06681624e-01 1.42294988e-01 8.94081518e-02 -1.54397473e-01
1.25178075e+00 -1.19537443e-01 -2.09160328e-01 -3.20673794e-01
-1.20902061e+00 1.13677692e+00 8.69593441e-01 4.19279337e-01
-3.13247859e-01 5.03291423e-03 5.22087038e-01 1.61102727e-01
2.53209949e-01 1.96935505e-01 -2.05260232e-01 2.68618643e-01
-1.14728534e+00 -5.01228347e-02 3.76377732e-01 5.06213121e-02
6.11243658e-02 -2.13178799e-01 -2.78379917e-01 1.11294580e+00
-1.15176797e-01 4.12825614e-01 8.19974542e-01 -1.09047735e+00
2.16558948e-01 4.87104416e-01 -1.93185344e-01 -5.30464232e-01
-1.04070628e+00 -5.17725646e-01 -9.67360079e-01 9.66632247e-01
9.27646995e-01 -2.28060275e-01 -9.40404117e-01 1.62378454e+00
4.72299755e-02 -3.52882117e-01 -5.23217082e-01 1.18639004e+00
7.18183756e-01 6.63889870e-02 2.57578760e-01 -1.51119620e-01
1.54632986e+00 -5.67529738e-01 -6.93823919e-02 -1.91551223e-01
5.80998480e-01 -2.43501812e-01 1.12327456e+00 1.85188383e-01
-1.08337903e+00 -2.43237823e-01 -8.84737790e-01 1.60972506e-01
-1.89299136e-01 3.67254585e-01 4.37689990e-01 8.76362205e-01
-1.35010684e+00 4.47583020e-01 -9.78100240e-01 -3.61322790e-01
9.63634908e-01 4.48123783e-01 -7.43960857e-01 2.12205112e-01
-8.17286372e-01 1.37761855e+00 1.30007222e-01 -1.65622562e-01
-5.86930215e-01 -8.30375791e-01 -3.16439033e-01 -1.75842289e-02
-1.37154788e-01 -6.24704242e-01 8.81735861e-01 -6.03020191e-01
-1.12762988e+00 1.18133771e+00 -9.42575037e-02 -5.66801250e-01
8.69102240e-01 2.46637225e-01 -2.11608887e-01 6.62367225e-01
3.61837178e-01 7.16813684e-01 4.19055879e-01 -7.58552969e-01
-3.52856666e-01 -6.36353433e-01 -2.39429578e-01 -2.22080216e-01
-6.60571381e-02 3.76750141e-01 -6.73037916e-02 -5.76135576e-01
2.07937732e-01 -6.73285961e-01 -4.91711825e-01 5.05305409e-01
2.98098428e-03 3.97988595e-03 4.59256731e-02 -9.76562679e-01
6.57438159e-01 -1.89265800e+00 -3.27405959e-01 6.50595427e-01
8.31630349e-01 3.60877886e-02 -8.17193687e-02 -4.97524798e-01
-3.71295333e-01 2.21292481e-01 -2.88608909e-01 1.93334654e-01
-4.65065300e-01 -2.00624481e-01 2.67974049e-01 7.61669755e-01
8.50426629e-02 9.08492625e-01 -5.86435795e-01 -4.99349028e-01
1.05060183e-01 5.98800719e-01 -4.01720285e-01 -4.38401029e-02
6.11922979e-01 3.39681745e-01 -2.46968538e-01 2.28881076e-01
4.91974890e-01 -5.25993049e-01 -2.66147945e-02 -1.94720432e-01
-5.17698005e-02 -2.23554820e-01 -6.97295606e-01 1.23931122e+00
-1.88668698e-01 9.05279815e-01 -6.38812780e-02 -1.06227541e+00
8.84208381e-01 1.84387922e-01 4.90420520e-01 -8.74368191e-01
4.38799948e-01 1.88644886e-01 5.27211130e-01 -4.49468255e-01
-5.43319941e-01 -3.09569061e-01 3.53677899e-01 5.27848125e-01
-1.22288018e-02 2.28805274e-01 2.65072912e-01 1.89259619e-01
1.41080201e+00 -4.77700233e-01 5.51320553e-01 -2.69759029e-01
4.09698516e-01 -7.93896616e-02 2.20801696e-01 7.85968542e-01
-5.87126374e-01 6.88550353e-01 9.51096654e-01 -5.81604779e-01
-1.20541465e+00 -1.47247767e+00 -6.32502913e-01 6.39733136e-01
-3.12515765e-01 6.24858215e-02 -8.57443333e-01 -5.46579301e-01
-5.82835823e-02 2.12576210e-01 -9.39996362e-01 2.07590595e-01
-5.82671702e-01 -1.21271968e+00 5.91424584e-01 9.14432526e-01
5.51508069e-01 -8.83038938e-01 -8.96699846e-01 1.28315404e-01
-7.79627636e-02 -1.08677876e+00 -1.49958193e-01 2.14690133e-03
-1.00530791e+00 -1.28583169e+00 -1.27842295e+00 -7.86406100e-01
8.03159714e-01 -6.80344701e-02 8.75761271e-01 9.86617897e-03
-8.43758702e-01 7.43294060e-02 -1.06351376e-02 -2.96574179e-02
-2.44835123e-01 -1.20351464e-01 -4.86396141e-02 -5.11053920e-01
2.98184276e-01 -8.15230370e-01 -1.01846874e+00 3.76015246e-01
-3.83519650e-01 -2.10531667e-01 8.59731376e-01 6.87443435e-01
4.86747354e-01 -5.22139013e-01 5.63037813e-01 -4.64201212e-01
7.08683670e-01 -3.10717285e-01 -6.33568406e-01 1.23990260e-01
-4.06244248e-01 -9.67830792e-02 4.52434689e-01 -5.87660193e-01
-5.30849993e-01 -7.00027272e-02 1.05892032e-01 -6.47616237e-02
-2.34356388e-01 2.04224706e-01 4.34246838e-01 -3.33145350e-01
1.03167796e+00 -1.57632187e-01 4.16266382e-01 -2.07179472e-01
2.88337886e-01 5.24083674e-01 6.84096932e-01 -2.28792205e-01
4.32521462e-01 6.48599744e-01 2.16780871e-01 -7.61912823e-01
-4.08740073e-01 -4.31020588e-01 -8.72220814e-01 -3.96277726e-01
1.12216985e+00 -6.80076420e-01 -6.51811600e-01 3.42122689e-02
-9.66432512e-01 -5.65969884e-01 -5.76479249e-02 9.13738728e-01
-3.65146041e-01 3.31500590e-01 -6.88024938e-01 -3.16066951e-01
-6.11790240e-01 -1.36722291e+00 4.58922327e-01 3.14893387e-02
-3.95030260e-01 -8.24357569e-01 7.04313740e-02 2.14484289e-01
6.21980608e-01 4.33837533e-01 1.15929842e+00 -7.26281822e-01
-5.72548769e-02 -3.80722046e-01 -7.53548563e-01 2.42472038e-01
-1.86086059e-01 -3.49048585e-01 -7.45036602e-01 -1.02357283e-01
-2.45016083e-01 -1.55580074e-01 9.34261978e-01 9.45281267e-01
1.20873797e+00 1.47728771e-01 -1.45321071e-01 4.19527233e-01
1.25686491e+00 1.23627886e-01 6.35819077e-01 6.91664279e-01
2.14164108e-01 5.63899517e-01 -2.79134661e-01 1.27318293e-01
1.51235148e-01 5.44060409e-01 1.73235893e-01 -4.14302588e-01
-3.96594018e-01 7.00902462e-01 1.26860335e-01 3.69482815e-01
-6.21021748e-01 5.07165551e-01 -1.00058520e+00 3.48706931e-01
-1.20461845e+00 -9.70102191e-01 -3.67388397e-01 2.27695465e+00
6.91709578e-01 2.57168174e-01 5.86421847e-01 -6.33352324e-02
7.92011201e-01 -2.79828548e-01 -5.55921376e-01 -2.15174258e-01
-4.69991863e-02 5.89415610e-01 8.19817960e-01 3.50388348e-01
-9.96957123e-01 4.67649698e-01 7.03797150e+00 2.81435907e-01
-1.27806604e+00 4.13116992e-01 7.87573159e-01 -4.20106232e-01
3.46698731e-01 -5.38282394e-01 -2.39004865e-01 3.31273615e-01
7.95752108e-01 -7.58651644e-02 4.65113789e-01 6.58739269e-01
5.29630542e-01 -2.04514220e-01 -6.87657297e-01 8.44113171e-01
2.53426284e-02 -1.18594575e+00 -2.86990613e-01 1.74926817e-01
3.13077688e-01 5.42769492e-01 -1.64698705e-01 2.20915508e-02
2.52379507e-01 -1.25485444e+00 6.33084476e-01 5.27142346e-01
1.17996252e+00 -3.96938473e-01 8.90234053e-01 -5.20658158e-02
-9.36915636e-01 -1.81643665e-02 -2.82683104e-01 1.86475590e-01
1.20427832e-01 6.06630445e-01 -9.65998173e-01 -3.93928319e-01
9.33949649e-01 1.05473146e-01 -7.90791392e-01 1.63193238e+00
-2.70571798e-01 4.96462584e-01 -3.34459275e-01 7.31220767e-02
7.38919750e-02 -9.16321427e-02 3.36414129e-01 1.12609100e+00
2.06070125e-01 1.96839690e-01 -1.70302406e-01 1.00173688e+00
9.95608419e-02 4.70659882e-01 -4.17246185e-02 4.84812260e-01
2.28127703e-01 1.28405035e+00 -1.22497153e+00 -3.48191619e-01
-5.84125638e-01 6.04735613e-01 4.84379262e-01 3.57468128e-01
-5.74731469e-01 -3.51355970e-01 1.78562790e-01 4.59521472e-01
2.13386878e-01 -1.97774783e-01 -6.17005706e-01 -9.10556793e-01
5.55340461e-02 -3.86677682e-01 3.43796343e-01 -7.48972356e-01
-1.35708618e+00 6.37524664e-01 -5.14087565e-02 -8.78535330e-01
-4.72495481e-02 -8.07838321e-01 -7.98238158e-01 1.11698735e+00
-1.31678271e+00 -5.38969874e-01 -4.98839468e-01 5.59815168e-01
4.03038710e-02 -2.11291984e-01 8.83187771e-01 1.36172414e-01
-3.97509158e-01 4.28626925e-01 -1.27763242e-01 5.93487084e-01
6.61069930e-01 -1.23109615e+00 2.12361977e-01 6.49677455e-01
-4.51868504e-01 6.50251210e-01 2.68286973e-01 -5.45126677e-01
-3.88538063e-01 -9.87098753e-01 5.13504028e-01 -2.07119271e-01
8.58588517e-01 1.20363981e-02 -8.47144663e-01 5.37449360e-01
-2.69204855e-01 4.01967555e-01 8.47358763e-01 -1.55154303e-01
-5.30577719e-01 3.77166905e-02 -1.34073579e+00 3.91422719e-01
8.29969227e-01 -2.36293659e-01 -7.26055622e-01 3.02921742e-01
8.39454401e-03 -1.22510180e-01 -1.22676218e+00 1.83689460e-01
9.93026793e-01 -1.02548027e+00 1.20106983e+00 -5.29815018e-01
4.21699405e-01 -7.37040490e-02 1.80255219e-01 -1.16913486e+00
-5.54465353e-01 2.40774319e-01 3.42097729e-01 4.31620002e-01
5.04028499e-01 -7.64854491e-01 7.82836378e-01 6.38701081e-01
4.97377105e-02 -7.99660444e-01 -1.22793078e+00 -7.99054146e-01
5.65553784e-01 -2.83326954e-01 2.02609405e-01 5.60422421e-01
3.13338973e-02 -3.79619114e-02 3.86947244e-01 5.77081367e-02
8.51094007e-01 -2.19962925e-01 3.47741097e-01 -1.46432710e+00
-3.65331359e-02 -1.00048983e+00 -8.39691579e-01 -3.64788294e-01
-2.96133254e-02 -1.31869173e+00 -4.20048565e-01 -1.71202707e+00
6.16049469e-01 -3.50410581e-01 -3.25655252e-01 7.02996910e-01
5.82772493e-02 6.02314711e-01 -4.84276973e-02 3.28954697e-01
-2.61760484e-02 -5.91979995e-02 1.24308896e+00 -5.68362102e-02
-2.54609913e-01 -2.83929974e-01 -4.76478398e-01 7.22990096e-01
1.21954465e+00 -3.44207793e-01 -7.36070648e-02 -1.36068389e-01
-3.09604496e-01 -4.46508043e-02 8.33097696e-01 -9.91692245e-01
-1.34214938e-01 1.96528554e-01 1.06485105e+00 -1.05175212e-01
4.00988646e-02 -4.93570298e-01 -2.43140042e-01 7.96401978e-01
-4.18761760e-01 -5.03515303e-02 2.00843066e-02 6.64056465e-02
2.93938786e-01 -2.05080897e-01 1.10774934e+00 -3.95180225e-01
-4.29472268e-01 2.57164747e-01 -6.81511521e-01 1.77544251e-01
9.19920146e-01 -2.83120960e-01 -4.60928082e-01 -1.28957123e-01
-1.15936697e+00 -4.95925248e-02 2.79698431e-01 3.39641012e-02
5.45339584e-01 -1.09369361e+00 -8.83224070e-01 2.60131508e-02
-8.26050937e-02 -5.08663177e-01 7.25394636e-02 1.38492811e+00
-8.04948270e-01 1.92159101e-01 -7.08986461e-01 -6.72291279e-01
-1.30913067e+00 2.01005954e-02 7.41250694e-01 -1.96654469e-01
-1.09782100e+00 6.21275067e-01 1.09868079e-01 -2.02492941e-02
1.55500144e-01 -4.55496222e-01 -5.13154984e-01 -2.86350138e-02
9.72345293e-01 5.70259809e-01 1.01959199e-01 -5.88378012e-01
-3.33853006e-01 6.03733540e-01 -1.23782739e-01 -2.38520011e-01
1.51668477e+00 1.52493790e-01 -2.18000874e-01 1.13885015e-01
1.26025212e+00 -2.54922897e-01 -9.87550020e-01 -3.54564711e-02
-2.37673074e-01 -1.61745518e-01 4.43893254e-01 -9.96801436e-01
-1.34706414e+00 8.05467725e-01 1.37276495e+00 8.84990469e-02
7.25730836e-01 3.09963197e-01 4.88298208e-01 5.34898378e-02
3.27102125e-01 -6.74383700e-01 -2.25052878e-01 4.81991917e-02
8.75385404e-01 -1.16897762e+00 7.14834481e-02 -2.29148760e-01
-4.50930774e-01 1.35769057e+00 3.36394042e-01 -3.42549860e-01
6.11465812e-01 2.82663494e-01 2.26469755e-01 -3.06034535e-01
-9.79587659e-02 -6.29395694e-02 2.79241592e-01 8.31253290e-01
4.47991997e-01 3.26426685e-01 -6.32423520e-01 6.78459823e-01
-5.06720304e-01 2.50230551e-01 5.12108088e-01 5.72828114e-01
-7.17440724e-01 -7.28242457e-01 -3.70271295e-01 1.10383034e+00
-5.73838115e-01 -1.96580105e-02 -1.31620005e-01 8.81843984e-01
4.51400988e-02 5.38497806e-01 2.72286892e-01 1.10825792e-01
3.65392536e-01 -9.00329207e-04 6.35208488e-01 -3.14677626e-01
-4.40502524e-01 -2.65261084e-02 -1.00418672e-01 -6.91561520e-01
-3.85882735e-01 -7.16669261e-01 -1.58108151e+00 -5.78788370e-02
4.55852086e-03 -2.21861809e-01 5.51550090e-01 8.33577037e-01
-9.70395356e-02 4.66341943e-01 2.47486129e-01 -7.19624996e-01
-3.54004174e-01 -9.13476050e-01 -6.91052496e-01 3.04858536e-01
1.19271368e-01 -8.78845751e-01 -4.73794132e-01 4.10600267e-02] | [14.152261734008789, -2.0642173290252686] |
3137ecd5-4eab-4ed6-a54c-272580ef33fd | gtnet-generative-transfer-network-for-zero | 2001.06812 | null | https://arxiv.org/abs/2001.06812v2 | https://arxiv.org/pdf/2001.06812v2.pdf | GTNet: Generative Transfer Network for Zero-Shot Object Detection | We propose a Generative Transfer Network (GTNet) for zero shot object detection (ZSD). GTNet consists of an Object Detection Module and a Knowledge Transfer Module. The Object Detection Module can learn large-scale seen domain knowledge. The Knowledge Transfer Module leverages a feature synthesizer to generate unseen class features, which are applied to train a new classification layer for the Object Detection Module. In order to synthesize features for each unseen class with both the intra-class variance and the IoU variance, we design an IoU-Aware Generative Adversarial Network (IoUGAN) as the feature synthesizer, which can be easily integrated into GTNet. Specifically, IoUGAN consists of three unit models: Class Feature Generating Unit (CFU), Foreground Feature Generating Unit (FFU), and Background Feature Generating Unit (BFU). CFU generates unseen features with the intra-class variance conditioned on the class semantic embeddings. FFU and BFU add the IoU variance to the results of CFU, yielding class-specific foreground and background features, respectively. We evaluate our method on three public datasets and the results demonstrate that our method performs favorably against the state-of-the-art ZSD approaches. | ['Shizhen Zhao', 'Lerenhan Li', 'Changqian Yu', 'Zhong Ji', 'Nong Sang', 'Yuanjie Shao', 'Changxin Gao'] | 2020-01-19 | null | null | null | null | ['zero-shot-object-detection'] | ['computer-vision'] | [ 2.83377975e-01 2.75593102e-01 1.98110074e-01 -2.86655128e-01
-6.58935368e-01 -3.60166848e-01 8.40451062e-01 -3.66763592e-01
-7.76933208e-02 4.64108944e-01 -5.70244938e-02 -6.12907857e-02
4.37669992e-01 -1.40077007e+00 -1.10829628e+00 -7.19022512e-01
1.69161722e-01 3.01069438e-01 6.36739612e-01 7.49199688e-02
-2.37090915e-01 3.13118160e-01 -1.52916455e+00 3.87520730e-01
6.73394859e-01 1.35037673e+00 4.64983396e-02 6.46380663e-01
3.32761742e-02 8.00460637e-01 -9.29050446e-01 -3.28725129e-01
5.07657826e-01 -4.56334472e-01 -3.95190209e-01 9.98661816e-02
4.40237522e-01 -7.50469804e-01 -7.41547287e-01 9.58944201e-01
6.26772702e-01 2.68901855e-01 1.07553887e+00 -1.70426333e+00
-1.33509457e+00 2.75343627e-01 -3.50311041e-01 1.32002264e-01
3.64779718e-02 5.42454958e-01 7.20633805e-01 -1.06276464e+00
6.92542911e-01 1.59694982e+00 3.35906506e-01 7.75120020e-01
-1.14468753e+00 -8.17520797e-01 1.44807994e-01 5.36198057e-02
-1.31026089e+00 -1.69311658e-01 6.42744303e-01 -5.53825259e-01
1.01667070e+00 -5.48934862e-02 6.84816778e-01 1.45821500e+00
3.77371728e-01 1.00959206e+00 5.72838008e-01 -2.51273662e-01
5.36843300e-01 1.20418154e-01 3.57279787e-03 7.23803580e-01
3.64402920e-01 3.24880332e-01 -3.07270527e-01 8.08813721e-02
1.04043067e+00 3.48050356e-01 -8.56619999e-02 -5.20489633e-01
-8.38691056e-01 1.20469797e+00 8.68483961e-01 -2.05416620e-01
-1.43522307e-01 4.44812238e-01 5.71171120e-02 2.82761663e-01
5.55954337e-01 4.20225561e-02 -2.49157891e-01 4.71790582e-01
-3.90800744e-01 2.26141050e-01 7.37061024e-01 1.33612335e+00
1.13933110e+00 2.21736804e-01 -8.08917582e-01 7.13146210e-01
3.55960935e-01 8.57546449e-01 3.75361800e-01 -5.83385170e-01
2.12069720e-01 8.29854548e-01 -1.21996999e-01 -7.50379026e-01
7.44418725e-02 -4.92317975e-01 -4.74242389e-01 3.72726351e-01
-4.67854515e-02 -2.88457394e-01 -1.54190099e+00 1.96908772e+00
6.34004235e-01 6.91797316e-01 1.73451126e-01 7.69952953e-01
1.20682073e+00 8.02711904e-01 1.27167672e-01 4.89632457e-01
1.32073653e+00 -1.07213104e+00 -2.02104867e-01 -4.67546046e-01
3.76288980e-01 -5.24421751e-01 8.65947187e-01 -3.56823117e-01
-6.52245462e-01 -6.17593288e-01 -1.15456140e+00 -1.68982714e-01
-5.89085162e-01 7.09374994e-02 4.27444398e-01 4.84129548e-01
-8.36073816e-01 2.62134016e-01 -7.73897350e-01 -3.42842847e-01
8.68090689e-01 1.22497253e-01 -1.93553850e-01 -2.89112628e-01
-1.21776927e+00 7.31433213e-01 6.10884428e-01 -2.92417139e-01
-1.77828062e+00 -8.80126059e-01 -1.20474994e+00 2.06421599e-01
4.72648591e-01 -1.08151519e+00 1.23544395e+00 -7.45220482e-01
-1.39122355e+00 8.07890713e-01 1.78412959e-01 -2.74632275e-01
4.86799419e-01 -9.85720381e-02 -2.78699458e-01 1.81461722e-01
3.53990734e-01 8.65051687e-01 1.36425066e+00 -1.26210904e+00
-7.44099796e-01 -1.40905026e-02 8.82510543e-02 -1.59204826e-02
-5.45478575e-02 -4.08787757e-01 -4.18793827e-01 -9.05651450e-01
-1.68349192e-01 -6.47703052e-01 2.03415558e-01 2.85718620e-01
-5.85676432e-01 -1.25573754e-01 1.21791017e+00 -3.88100803e-01
7.17372417e-01 -2.44704723e+00 -1.21575959e-01 -3.40166725e-02
4.20286864e-01 1.86865315e-01 -3.57171863e-01 -8.56833681e-02
8.16208944e-02 -1.95424572e-01 -1.57764211e-01 -2.24508420e-01
3.22089553e-01 2.23665372e-01 -5.44896245e-01 1.19491473e-01
9.44026828e-01 1.39537060e+00 -1.28034580e+00 -7.14186653e-02
2.95311868e-01 6.02008760e-01 -8.18520427e-01 4.74333107e-01
-4.82397020e-01 -1.71922315e-02 -7.19737113e-01 8.25861812e-01
6.42055154e-01 -2.94099629e-01 -3.71193141e-01 -1.43830106e-01
3.99313688e-01 1.38022795e-01 -1.01030648e+00 1.09183848e+00
-1.63706496e-01 2.81781465e-01 -4.02993709e-01 -4.91465032e-01
1.03600550e+00 -1.90136116e-02 -1.50107276e-02 -4.66098160e-01
3.54225427e-01 -2.05875710e-02 -1.39147013e-01 -1.04125045e-01
1.81965381e-01 -2.53162235e-01 -2.10031629e-01 4.95690763e-01
8.07211161e-01 -1.89839453e-01 -1.85488760e-01 5.09736001e-01
1.48081779e+00 4.27902713e-02 3.41435261e-02 -6.95143715e-02
1.50581077e-01 -3.93606663e-01 5.66849768e-01 9.17503774e-01
-3.44284475e-01 8.07529807e-01 6.08481526e-01 -3.65858108e-01
-9.45279300e-01 -1.74664009e+00 1.20209798e-01 1.04554272e+00
2.81098217e-01 4.47083684e-03 -9.18391109e-01 -1.04804468e+00
4.49021250e-01 7.97647715e-01 -1.02961755e+00 -9.04454291e-01
-6.32233396e-02 -6.02318704e-01 4.05144542e-01 9.20533597e-01
6.07525468e-01 -1.39836466e+00 -4.94081825e-01 1.94055468e-01
3.72361273e-01 -1.12086475e+00 -6.53290153e-01 2.10333228e-01
-3.24760526e-01 -1.01291919e+00 -6.80441022e-01 -8.16528261e-01
7.80524194e-01 2.42430255e-01 1.02068973e+00 -1.80557087e-01
-7.74938226e-01 5.41859150e-01 -4.50935185e-01 -7.83249557e-01
-4.54270720e-01 -4.88756001e-02 -1.10381424e-01 2.72165328e-01
6.17606044e-01 -3.43295902e-01 -6.49353147e-01 2.15449587e-01
-1.12865245e+00 1.07436152e-02 5.23267150e-01 8.85098219e-01
5.19455016e-01 -4.10395384e-01 7.51565576e-01 -7.66520083e-01
1.57663941e-01 -7.02771902e-01 -5.80624104e-01 1.49312615e-01
-1.23429984e-01 -1.74689829e-01 3.29431623e-01 -6.94403708e-01
-9.36619461e-01 -1.61644921e-01 1.98324069e-01 -9.61180806e-01
2.74239965e-02 -1.82662353e-01 -6.39482796e-01 1.22334719e-01
6.92891598e-01 2.37252355e-01 -2.67667532e-01 -1.83310032e-01
6.24792099e-01 5.44428051e-01 6.59913421e-01 -4.19944346e-01
1.16057885e+00 4.04195935e-01 -3.60074401e-01 -4.36401129e-01
-1.07164848e+00 -7.80376494e-02 -4.52404112e-01 -4.06307653e-02
1.14755893e+00 -1.11001980e+00 -1.04094833e-01 8.55628252e-01
-9.84762907e-01 -5.29379368e-01 -8.25959623e-01 1.46762103e-01
-5.96843898e-01 -3.36203337e-01 -5.24662495e-01 -4.96658981e-01
-3.67228270e-01 -1.04960454e+00 1.40709662e+00 5.41855812e-01
1.45750925e-01 -9.13540542e-01 -2.57241249e-01 7.07015069e-03
3.98059368e-01 6.38053417e-01 9.79612052e-01 -8.76257122e-01
-9.25760031e-01 -2.20113352e-01 -5.50211668e-01 6.99732184e-01
2.66631603e-01 -2.03427821e-01 -1.27771246e+00 -4.10546958e-01
-1.51074499e-01 -4.72265273e-01 1.04231310e+00 1.85085073e-01
8.89239311e-01 -1.87713116e-01 -4.60771978e-01 7.11380780e-01
1.22258174e+00 1.68445975e-01 6.29250467e-01 9.38148126e-02
8.39109302e-01 -8.12768266e-02 3.03670168e-01 4.43275064e-01
2.51342773e-01 2.69396245e-01 5.22052705e-01 -5.20125702e-02
-7.21137881e-01 -5.56631327e-01 4.79660213e-01 1.82977766e-01
4.44439709e-01 -4.78628874e-01 -6.93474829e-01 5.28885543e-01
-1.62003160e+00 -8.88291836e-01 5.49266756e-01 1.91627908e+00
6.21151865e-01 2.64815986e-01 -1.10818297e-01 -2.33829960e-01
7.98485160e-01 1.57963976e-01 -8.57457042e-01 -1.01084121e-01
2.59982795e-02 3.90010476e-01 7.57596567e-02 2.96776533e-01
-1.14431405e+00 1.18251026e+00 5.51278925e+00 7.39673436e-01
-1.01174617e+00 2.04446390e-01 5.32104611e-01 -1.68268457e-01
-3.31042558e-01 -2.06631094e-01 -1.05715072e+00 6.14547312e-01
5.26724339e-01 -1.73388198e-01 1.97378471e-01 1.23143041e+00
-5.55968106e-01 1.05303898e-01 -1.21158171e+00 5.56326985e-01
1.91524252e-01 -1.13686180e+00 5.32994330e-01 -7.49905631e-02
9.03118789e-01 1.24190792e-01 1.46987364e-01 8.79085302e-01
5.96319616e-01 -8.62504601e-01 6.69038951e-01 4.32056218e-01
9.47960794e-01 -8.88944685e-01 5.70785999e-01 1.20416157e-01
-1.14061832e+00 3.97108532e-02 -6.00584984e-01 1.93411618e-01
1.76907354e-03 6.34796321e-01 -1.01251388e+00 4.31306899e-01
5.87208807e-01 6.42368436e-01 -5.41951776e-01 8.17244053e-01
-5.71446180e-01 5.05467832e-01 -3.10249388e-01 1.82105809e-01
2.71097749e-01 4.19660062e-02 6.45306706e-01 1.01766920e+00
2.59185731e-01 -1.24338828e-02 2.44576886e-01 1.37257135e+00
-5.50307930e-01 -3.76711875e-01 -4.42521930e-01 -5.09755090e-02
6.02626204e-01 1.19037724e+00 -7.17191279e-01 -6.54642522e-01
-4.82960016e-01 1.22631204e+00 3.60568076e-01 4.19655830e-01
-1.02979815e+00 -6.99950576e-01 8.97835076e-01 4.17425558e-02
8.09925973e-01 4.12454963e-01 1.02428244e-02 -1.40674305e+00
-1.29558727e-01 -5.51045835e-01 4.48493361e-01 -7.86117435e-01
-1.69642484e+00 5.11627018e-01 -1.22902907e-01 -1.05720711e+00
-9.04371738e-02 -7.80490398e-01 -8.55885684e-01 9.93301809e-01
-1.48833644e+00 -1.45870006e+00 -6.81168616e-01 6.05964124e-01
3.97711039e-01 -3.93318117e-01 7.21269965e-01 2.29848191e-01
-7.09541857e-01 8.32284212e-01 -3.09881270e-02 5.65814376e-01
6.82958901e-01 -1.18746281e+00 7.46806085e-01 8.32371235e-01
-1.92354590e-01 4.05798107e-01 1.88767821e-01 -8.78140748e-01
-1.17300260e+00 -1.95293581e+00 3.50970089e-01 -7.70221055e-01
6.54871941e-01 -8.11634839e-01 -8.49538922e-01 1.03064632e+00
-4.16369826e-01 6.56360567e-01 4.79706079e-01 -4.45684254e-01
-7.26736605e-01 3.79476063e-02 -1.34224367e+00 4.05726731e-01
1.16898954e+00 -5.66878855e-01 -8.36625159e-01 1.30303308e-01
1.23801470e+00 -5.16672790e-01 -5.13769388e-01 4.86004531e-01
3.78414243e-01 -6.58583045e-01 9.85476434e-01 -6.40372515e-01
5.80615938e-01 -4.69106227e-01 -3.63477796e-01 -1.23459768e+00
-3.87265593e-01 -2.06837684e-01 -5.10246992e-01 1.33709431e+00
2.26160526e-01 -7.71807909e-01 5.85450351e-01 4.70805466e-01
-3.08184654e-01 -7.32308567e-01 -9.27321970e-01 -1.05122936e+00
-3.98528241e-02 -2.18512088e-01 7.25986779e-01 7.11146295e-01
-8.00132692e-01 5.39074302e-01 -8.20238441e-02 2.72893369e-01
6.74492538e-01 1.93909705e-01 9.14767325e-01 -1.06164742e+00
-5.16772151e-01 -1.38029873e-01 -7.45688558e-01 -7.28457987e-01
1.67082902e-02 -1.07646406e+00 6.15207739e-02 -1.46358752e+00
4.48137164e-01 -1.66461140e-01 -4.13624972e-01 7.71749377e-01
-5.45680106e-01 3.91043663e-01 3.80089313e-01 -1.88276544e-01
-4.99108791e-01 9.81624246e-01 1.35371757e+00 -2.21858084e-01
-1.26344681e-01 -1.36330038e-01 -6.26315951e-01 6.68566465e-01
2.77747005e-01 -6.63436353e-01 -4.78421122e-01 -4.66489434e-01
-4.32196856e-01 -6.26346588e-01 8.49486947e-01 -1.09565091e+00
-1.24836564e-01 6.60382509e-02 9.16661561e-01 -4.38464135e-01
3.46791804e-01 -5.51468432e-01 -1.36550188e-01 5.21804154e-01
1.05089493e-01 -6.79755688e-01 1.97939590e-01 7.49488056e-01
1.02155410e-01 1.92796346e-02 9.21933591e-01 6.99568167e-02
-7.77609587e-01 6.89768493e-01 -3.53040658e-02 1.46478951e-01
1.33222270e+00 -9.13825929e-02 -6.04077995e-01 -1.95941076e-01
-5.42768955e-01 3.34477097e-01 4.16288793e-01 8.55621576e-01
8.15042377e-01 -1.58510733e+00 -6.45148993e-01 6.57139301e-01
3.97445858e-01 3.93845975e-01 2.04761043e-01 1.92583278e-01
-1.80292964e-01 -1.40177742e-01 -8.78603086e-02 -5.40150464e-01
-5.63545167e-01 8.00511718e-01 3.22125822e-01 6.61425525e-03
-6.08135402e-01 1.25274289e+00 1.05733967e+00 -5.68209469e-01
9.51606855e-02 -3.25044692e-01 3.48514944e-01 -1.38047561e-02
6.91933155e-01 3.04727167e-01 -2.24862292e-01 -3.47717226e-01
-3.00691217e-01 2.54968405e-01 -1.41223952e-01 1.52126610e-01
1.22909677e+00 2.42627978e-01 2.70107776e-01 2.75019258e-01
1.33524823e+00 -5.13540506e-01 -1.67834771e+00 -4.34194654e-01
-4.43186671e-01 -3.90340865e-01 -8.11326131e-03 -9.44351137e-01
-1.05825520e+00 8.20320606e-01 6.49504900e-01 -1.50307164e-01
1.03767669e+00 3.14973980e-01 9.26725864e-01 1.96381494e-01
4.76613551e-01 -7.13011444e-01 5.29433906e-01 7.67465472e-01
8.03728342e-01 -1.13332677e+00 -4.27799910e-01 -5.59895694e-01
-5.63577116e-01 6.92167699e-01 9.89586771e-01 -6.20938241e-01
6.31712258e-01 3.46797705e-01 7.81535264e-03 -7.19594359e-02
-7.27159619e-01 -3.76916915e-01 4.35179204e-01 8.19288611e-01
-2.96849817e-01 3.21714096e-02 5.41622519e-01 9.78025258e-01
-3.09265554e-02 2.49058846e-02 4.13088165e-02 1.13533950e+00
-6.13103271e-01 -7.29827285e-01 -1.70890287e-01 6.48580730e-01
2.79697087e-02 -1.71672836e-01 -3.43217164e-01 8.17214429e-01
4.51259434e-01 5.67731500e-01 3.28658372e-01 -5.66400886e-01
3.70174527e-01 2.10846141e-01 4.56169158e-01 -1.10839438e+00
-3.40648830e-01 -1.92140341e-01 -2.86130488e-01 -5.35859048e-01
3.27659279e-01 -2.46360257e-01 -1.38193524e+00 2.68504098e-02
-3.85860384e-01 -3.42453361e-01 2.92164087e-01 8.84086013e-01
5.71118653e-01 8.92767787e-01 7.50133812e-01 -9.20164526e-01
-6.19264364e-01 -9.57839966e-01 -6.10499799e-01 5.31995296e-01
4.12145585e-01 -1.02487040e+00 -4.75244999e-01 1.38687760e-01] | [9.789335250854492, 2.162303924560547] |
abf02605-a97a-4972-8ee6-c5ad1cf024d5 | feddebug-systematic-debugging-for-federated | 2301.03553 | null | https://arxiv.org/abs/2301.03553v1 | https://arxiv.org/pdf/2301.03553v1.pdf | FedDebug: Systematic Debugging for Federated Learning Applications | In Federated Learning (FL), clients train a model locally and share it with a central aggregator to build a global model. Impermissibility to access client's data and collaborative training makes FL appealing for applications with data-privacy concerns such as medical imaging. However, these FL characteristics pose unprecedented challenges for debugging. When a global model's performance deteriorates, finding the round and the clients responsible is a major pain point. Developers resort to trial-and-error debugging with subsets of clients, hoping to increase the accuracy or let future FL rounds retune the model, which are time-consuming and costly. We design a systematic fault localization framework, FedDebug, that advances the FL debugging on two novel fronts. First, FedDebug enables interactive debugging of realtime collaborative training in FL by leveraging record and replay techniques to construct a simulation that mirrors live FL. FedDebug's {\em breakpoint} can help inspect an FL state (round, client, and global model) and seamlessly move between rounds and clients' models, enabling a fine-grained step-by-step inspection. Second, FedDebug automatically identifies the client responsible for lowering global model's performance without any testing data and labels--both are essential for existing debugging techniques. FedDebug's strengths come from adapting differential testing in conjunction with neurons activations to determine the precise client deviating from normal behavior. FedDebug achieves 100\% to find a single client and 90.3\% accuracy to find multiple faulty clients. FedDebug's interactive debugging incurs 1.2\% overhead during training, while it localizes a faulty client in only 2.1\% of a round's training time. With FedDebug, we bring effective debugging practices to federated learning, improving the quality and productivity of FL application developers. | ['Muhammad Ali Gulzar', 'Ali Anwar', 'Waris Gill'] | 2023-01-09 | null | null | null | null | ['fault-localization'] | ['computer-code'] | [-3.57881367e-01 6.89318404e-02 -3.56681138e-01 -5.68566740e-01
-1.17213082e+00 -7.44832933e-01 -1.98283553e-01 9.14266482e-02
-7.76490197e-02 5.51850677e-01 -4.58675623e-01 -6.91224873e-01
-8.73944387e-02 -5.81360936e-01 -8.04256380e-01 -6.40004992e-01
-4.36470568e-01 2.74629921e-01 2.06221089e-01 3.99926841e-01
-1.80396855e-01 3.79460484e-01 -1.39897704e+00 7.42774427e-01
7.31705368e-01 9.30720568e-01 -2.11805731e-01 9.12994683e-01
1.33789375e-01 1.22679317e+00 -9.41379726e-01 -2.36667082e-01
1.51807725e-01 -3.23416620e-01 -8.80227745e-01 -3.13711822e-01
2.20374256e-01 -7.60184586e-01 1.75672844e-01 8.90429616e-01
4.13247645e-01 -3.17769110e-01 -1.38196170e-01 -1.78525543e+00
-1.13216303e-01 7.90146172e-01 -4.41880196e-01 9.03898850e-02
2.32586682e-01 5.27100682e-01 8.96537900e-01 -4.20002043e-01
4.92458761e-01 6.06527746e-01 1.00519657e+00 7.14758337e-01
-1.20604694e+00 -9.81016338e-01 1.28182292e-01 -1.15845546e-01
-1.31744647e+00 -6.39935911e-01 2.78055966e-01 -5.22309542e-01
1.01372457e+00 6.68557703e-01 5.81337735e-02 8.83645713e-01
2.47941077e-01 6.07551336e-01 8.37751567e-01 -1.44484654e-01
6.28406644e-01 4.31717858e-02 2.79467970e-01 9.42161322e-01
1.02937497e-01 2.97109410e-02 -7.00176716e-01 -8.44308317e-01
5.90467274e-01 1.41153619e-01 -4.76542681e-01 -9.59726423e-02
-8.13711882e-01 3.73023063e-01 1.20582864e-01 1.02564476e-01
4.71592583e-02 3.75008732e-01 5.29600799e-01 6.18868768e-01
1.99229762e-01 5.23505569e-01 -8.59043539e-01 -4.32442367e-01
-9.14425135e-01 7.13030174e-02 1.02081871e+00 8.67546320e-01
8.84679615e-01 -1.40689135e-01 -1.27955467e-01 4.18296725e-01
1.49457995e-02 5.32545298e-02 4.89509940e-01 -1.29597366e+00
1.51088580e-01 9.13291037e-01 -8.70318785e-02 -6.65080965e-01
-2.57736415e-01 -4.31883097e-01 -4.39801425e-01 6.73636615e-01
3.20654362e-01 -4.67897683e-01 -3.64189357e-01 1.79166877e+00
5.50716519e-01 3.36549759e-01 -2.85063088e-01 5.75957835e-01
2.74028271e-01 2.18166187e-01 -4.25556660e-01 -8.88179094e-02
9.60305452e-01 -8.87027204e-01 -1.24625161e-01 -9.48060583e-03
1.21329331e+00 -2.83474028e-01 1.06690764e+00 6.57435477e-01
-1.00293076e+00 -1.44957140e-01 -9.86028552e-01 3.66519690e-01
1.07093610e-01 -1.13398895e-01 6.00772142e-01 7.78117120e-01
-1.28048587e+00 7.70643175e-01 -1.39009178e+00 1.65905610e-01
6.08926892e-01 6.41099155e-01 -4.82144952e-01 -8.78926069e-02
-4.34690505e-01 3.44456524e-01 -2.41687164e-01 -2.30608657e-01
-1.14589512e+00 -9.97195959e-01 -6.14471018e-01 2.26416841e-01
2.81622052e-01 -5.75056493e-01 1.79361200e+00 -8.12877893e-01
-1.04657006e+00 4.52933490e-01 -8.84158239e-02 -2.45343283e-01
7.05304623e-01 1.06324598e-01 -4.85208273e-01 -1.47108480e-01
1.77313834e-01 -7.91418999e-02 6.67361498e-01 -1.07722306e+00
-7.23020434e-01 -3.12421203e-01 -1.93671048e-01 -4.42512780e-01
-1.47706538e-01 3.39989290e-02 -2.58438230e-01 -2.10145712e-02
-1.11349091e-01 -3.94762874e-01 -4.29713586e-03 2.53675967e-01
-3.57346475e-01 2.75743343e-02 1.10728765e+00 -3.35092634e-01
1.30666411e+00 -2.37823343e+00 -7.67101228e-01 4.63098079e-01
9.08887029e-01 1.73082471e-01 -2.60586962e-02 3.15473884e-01
-8.75033885e-02 2.08544791e-01 4.86870483e-02 -5.87953031e-01
-9.28215086e-02 2.44827852e-01 -1.28166974e-01 5.58101475e-01
-5.14537841e-02 6.97358429e-01 -8.68942976e-01 -3.77140909e-01
-3.35629970e-01 1.62839860e-01 -9.20653105e-01 4.92517620e-01
-3.20788592e-01 3.58530223e-01 -5.00414371e-01 9.23980176e-01
4.52446759e-01 -7.93371499e-01 4.64129448e-01 1.87197015e-01
-1.18607819e-01 4.39333498e-01 -1.13373113e+00 1.24129367e+00
-4.82879162e-01 3.83512676e-01 5.10981560e-01 -6.37630045e-01
8.13987970e-01 4.30783123e-01 4.79477465e-01 -5.41289568e-01
1.03576750e-01 3.12280208e-01 -1.93475515e-01 -4.64540094e-01
-1.05488203e-01 3.70170683e-01 -2.26756018e-02 1.32455122e+00
-3.64862829e-02 5.62368870e-01 -3.57559860e-01 7.18683079e-02
2.13277078e+00 -2.26905067e-02 1.90675948e-02 1.21631594e-02
-4.85386215e-02 -4.15402092e-02 9.48419690e-01 9.82936800e-01
-3.22370946e-01 5.20524800e-01 9.39452648e-01 -4.04686272e-01
-5.63249648e-01 -8.87770534e-01 1.60943046e-01 1.27628732e+00
-3.81320417e-01 -4.61555421e-01 -8.82234514e-01 -1.17746389e+00
1.67599306e-01 5.72916865e-01 -6.85836434e-01 -3.15096557e-01
-4.21025962e-01 -2.68023431e-01 8.58782232e-01 4.08418864e-01
3.95690769e-01 -8.17219079e-01 -1.08281159e+00 1.64322644e-01
6.15181886e-02 -3.69846523e-01 -7.40869224e-01 6.56695187e-01
-7.17576742e-01 -1.41469967e+00 -3.32127698e-02 -4.26761627e-01
8.33099186e-01 -2.13625468e-02 1.04446995e+00 5.18600166e-01
-6.11095786e-01 2.36389592e-01 -2.20142454e-01 4.37418371e-02
-6.44422412e-01 -1.91336945e-01 -1.93022117e-01 -1.13047138e-01
1.46615699e-01 -8.70953262e-01 -5.68395615e-01 7.44314969e-01
-6.23202443e-01 -2.88906038e-01 2.01133430e-01 1.00116932e+00
4.63448495e-01 -2.76096668e-02 5.80465853e-01 -1.55854952e+00
4.76264596e-01 -7.12008297e-01 -9.31646109e-01 3.10165346e-01
-1.20074606e+00 -1.94640700e-02 8.95670295e-01 -5.65543890e-01
-7.44444788e-01 -1.45222601e-02 1.03826565e-03 -8.44374537e-01
-1.16109066e-01 3.55962217e-01 -9.04663056e-02 -1.14004411e-01
1.14344347e+00 -5.26656322e-02 4.14622188e-01 -6.09840870e-01
-1.24063157e-01 7.01113105e-01 7.08682954e-01 -7.28067160e-01
2.48362526e-01 1.09034747e-01 -6.86505497e-01 9.49595049e-02
-3.60146433e-01 -2.77280957e-01 6.14116788e-02 -4.45085578e-02
1.82890758e-01 -6.77412629e-01 -1.44564903e+00 4.47187036e-01
-9.15980220e-01 -8.24608505e-01 -3.10117871e-01 2.83778068e-02
-1.31415635e-01 2.70082485e-02 -6.37857854e-01 -6.49785042e-01
-5.25864720e-01 -1.27400565e+00 7.80262649e-01 -3.79142864e-03
-5.53848982e-01 -7.09698975e-01 1.52422488e-01 2.71593034e-01
5.71637273e-01 2.72801071e-01 8.62182617e-01 -9.75543797e-01
-6.06845260e-01 -4.79270726e-01 5.28013408e-02 4.17501062e-01
1.58870652e-01 8.68711770e-02 -1.44057405e+00 -4.64458883e-01
2.44619146e-01 -2.94978917e-01 9.46899969e-03 -5.19250073e-02
1.45414209e+00 -7.85301745e-01 -4.47070181e-01 9.26007569e-01
1.10497797e+00 3.74843031e-01 2.28524521e-01 8.38637128e-02
4.47773606e-01 -8.93639401e-03 2.93929249e-01 7.64351130e-01
1.61647573e-01 1.28421187e-01 6.19004190e-01 -2.11301427e-02
1.23892330e-01 -8.38587806e-02 4.30948287e-01 3.60377520e-01
6.51225209e-01 -2.27319449e-02 -1.04685307e+00 2.97294110e-01
-1.83419073e+00 -7.47941375e-01 1.14035435e-01 2.41486645e+00
1.13739777e+00 1.65861212e-02 4.98876534e-02 7.92925339e-03
3.75986785e-01 -3.58245224e-01 -1.03006673e+00 -4.36688423e-01
5.25712967e-01 2.91236460e-01 5.15635133e-01 2.19829544e-01
-5.94441891e-01 3.26957613e-01 5.76427126e+00 5.18616021e-01
-1.48342824e+00 3.83421451e-01 9.20945823e-01 -2.49962464e-01
-4.19118732e-01 2.85713803e-02 -6.85102820e-01 5.05452394e-01
1.12677944e+00 -1.80097699e-01 5.60118437e-01 1.43231690e+00
2.07660124e-02 -2.68896013e-01 -1.69563496e+00 9.29420590e-01
-1.59937650e-01 -1.53493845e+00 -8.21406603e-01 9.21783224e-02
5.14270008e-01 3.46467197e-01 -1.27471790e-01 2.25529268e-01
8.27026069e-01 -1.15162528e+00 6.08202398e-01 2.86872268e-01
1.09586287e+00 -8.09505224e-01 6.65368676e-01 7.06530035e-01
-7.94415355e-01 -1.46564633e-01 1.81952536e-01 2.98619717e-02
-3.37895542e-01 6.55093908e-01 -1.29909706e+00 1.93485081e-01
1.04464006e+00 9.91500095e-02 -5.25722027e-01 8.75276625e-01
1.74560100e-01 8.89016211e-01 -3.34195763e-01 1.59897700e-01
-3.36560577e-01 3.60404193e-01 1.05272397e-01 8.87744009e-01
1.61016122e-01 -1.18581824e-01 1.19024098e-01 1.04905164e+00
-2.23435581e-01 -3.18878382e-01 -2.80070662e-01 2.60587363e-03
9.13557410e-01 1.18962216e+00 -2.90107191e-01 -1.50795296e-01
-2.48281494e-01 8.59724402e-01 4.73362923e-01 1.52056530e-01
-8.32788467e-01 -4.97456551e-01 1.10394490e+00 1.24152355e-01
1.02657609e-01 1.49758786e-01 -4.28260267e-01 -7.56605148e-01
2.49905840e-01 -1.27090168e+00 6.46486700e-01 -4.42181677e-01
-1.25753522e+00 8.28859925e-01 -4.52895075e-01 -9.99159217e-01
-5.02398372e-01 -5.11874147e-02 -8.49449396e-01 8.81795585e-01
-7.49586284e-01 -8.45496535e-01 -4.09969479e-01 8.84852886e-01
-9.07097533e-02 1.17559172e-01 1.09683108e+00 4.21629459e-01
-7.73779690e-01 1.39152443e+00 -5.86304963e-02 1.14052899e-01
8.11027348e-01 -9.86833990e-01 4.98300701e-01 8.34845245e-01
-1.91000313e-01 9.22904074e-01 3.09448510e-01 -6.28989041e-01
-1.56639934e+00 -1.30609727e+00 7.52077758e-01 -4.85184848e-01
6.10323846e-01 -5.20743072e-01 -1.14723635e+00 1.05866325e+00
-4.06684160e-01 6.71265721e-01 7.03341484e-01 3.40041906e-01
-8.10824394e-01 -4.56575662e-01 -1.63132703e+00 2.94373304e-01
6.81535423e-01 -7.71292388e-01 2.65717834e-01 4.02752757e-01
7.81286538e-01 -5.84358931e-01 -9.55397666e-01 -1.21802934e-01
4.68231201e-01 -1.38525105e+00 3.60427052e-01 -4.04746979e-01
1.16666757e-01 -3.29403371e-01 3.51214148e-02 -9.33722913e-01
7.08528012e-02 -1.09717500e+00 -1.87443048e-01 1.28817618e+00
4.92852241e-01 -1.11626112e+00 1.00053382e+00 1.21517193e+00
-3.42741013e-01 -9.31183577e-01 -8.66474152e-01 -6.18672252e-01
-3.38066578e-01 -7.60437787e-01 1.05642211e+00 1.16926706e+00
2.83941805e-01 -3.63535106e-01 -6.51596263e-02 4.52318817e-01
4.54022288e-01 -7.40219355e-02 8.64877045e-01 -7.96319544e-01
-1.08720779e+00 -1.36988595e-01 -7.76245221e-02 -6.30714417e-01
-1.01486295e-01 -8.04599941e-01 2.67482489e-01 -7.41960347e-01
6.38244599e-02 -9.81062710e-01 -2.09376380e-01 1.45644176e+00
1.52012616e-01 -1.44330099e-01 -1.74654946e-01 1.88516304e-01
-6.13855302e-01 -3.54270965e-01 6.65363967e-01 2.11875349e-01
-3.51901948e-01 3.33097219e-01 -6.83364987e-01 4.19315308e-01
6.10686839e-01 -6.64176464e-01 -4.31503326e-01 -5.21539927e-01
3.08522075e-01 4.60086793e-01 5.79169691e-01 -1.03891301e+00
7.53788531e-01 -2.56376322e-02 2.87860394e-01 -1.47959605e-01
-1.55026034e-01 -8.13848555e-01 6.70241892e-01 4.40902740e-01
-2.52343535e-01 3.50862980e-01 6.92001879e-02 3.67893547e-01
2.31933352e-02 -1.34022787e-01 7.36732483e-01 -1.05724305e-01
-1.84250906e-01 3.57543796e-01 -2.54342616e-01 1.07080750e-01
1.07332778e+00 -1.58623859e-01 -7.00668037e-01 -2.09888056e-01
-6.04796469e-01 4.53945428e-01 8.21250379e-01 -1.67228401e-01
3.49985719e-01 -5.40307403e-01 -8.14037696e-02 6.89154208e-01
5.57054877e-02 2.11492747e-01 3.50904852e-01 9.88397062e-01
-3.85606289e-01 -2.20527083e-01 1.48830250e-01 -6.12255692e-01
-1.39588559e+00 3.78750980e-01 7.20634580e-01 -7.94555768e-02
-7.01768041e-01 1.26490331e+00 -6.18439987e-02 -5.89372694e-01
5.80984533e-01 -4.40130532e-01 7.88417101e-01 -3.86376500e-01
8.10436964e-01 4.66468513e-01 6.15709007e-01 3.07691991e-01
-5.92414558e-01 -4.05909687e-01 -2.10491836e-01 1.25761449e-01
1.38959062e+00 1.77148297e-01 -3.13402504e-01 4.64701861e-01
1.28037095e+00 1.11094892e-01 -1.50997865e+00 5.03752008e-02
-1.35976091e-01 -4.92133170e-01 -2.67004664e-03 -1.25565994e+00
-1.29194629e+00 4.37469572e-01 5.56682467e-01 -2.79039405e-02
1.18526506e+00 6.69146404e-02 6.73018277e-01 2.23919094e-01
7.66628325e-01 -5.56808412e-01 -2.12733001e-01 1.75182968e-02
3.00926507e-01 -1.01719344e+00 -2.28447855e-01 -9.48269479e-03
-3.81514847e-01 9.68889058e-01 9.28693295e-01 2.17969894e-01
5.37621439e-01 1.01535118e+00 2.46968374e-01 -2.54182965e-01
-1.37960148e+00 8.11214745e-01 -3.46581966e-01 5.03473401e-01
9.19256359e-02 8.90017673e-02 4.85446483e-01 1.01717150e+00
-6.96876496e-02 3.52661252e-01 4.06397283e-01 1.40108621e+00
-7.72972777e-02 -1.19005811e+00 -4.28739518e-01 8.30662549e-01
-4.56919402e-01 1.22463025e-01 -5.63153811e-03 5.54912865e-01
2.18083903e-01 8.40633214e-01 9.18751284e-02 -6.29543960e-01
8.17400776e-03 2.04365179e-01 1.32035576e-02 -7.46616006e-01
-1.15294993e+00 -4.04190309e-02 9.88090113e-02 -1.25904298e+00
5.50174356e-01 -5.50561249e-01 -1.51907086e+00 -6.50172174e-01
-3.66227418e-01 3.14422071e-01 6.99578881e-01 7.95567334e-01
8.81557584e-01 4.12669152e-01 9.95789587e-01 -2.89660394e-01
-8.60429466e-01 -3.19181859e-01 -6.42337203e-01 -1.16803795e-01
6.40836775e-01 -1.46640912e-01 -5.04910588e-01 1.03339562e-02] | [6.057352542877197, 6.4258880615234375] |
177a7b69-0bb9-4ee0-8436-475aaa0e1ac0 | neural-probabilistic-logic-programming-in | 2303.04660 | null | https://arxiv.org/abs/2303.04660v2 | https://arxiv.org/pdf/2303.04660v2.pdf | Neural Probabilistic Logic Programming in Discrete-Continuous Domains | Neural-symbolic AI (NeSy) allows neural networks to exploit symbolic background knowledge in the form of logic. It has been shown to aid learning in the limited data regime and to facilitate inference on out-of-distribution data. Probabilistic NeSy focuses on integrating neural networks with both logic and probability theory, which additionally allows learning under uncertainty. A major limitation of current probabilistic NeSy systems, such as DeepProbLog, is their restriction to finite probability distributions, i.e., discrete random variables. In contrast, deep probabilistic programming (DPP) excels in modelling and optimising continuous probability distributions. Hence, we introduce DeepSeaProbLog, a neural probabilistic logic programming language that incorporates DPP techniques into NeSy. Doing so results in the support of inference and learning of both discrete and continuous probability distributions under logical constraints. Our main contributions are 1) the semantics of DeepSeaProbLog and its corresponding inference algorithm, 2) a proven asymptotically unbiased learning algorithm, and 3) a series of experiments that illustrate the versatility of our approach. | ['Luc De Raedt', 'Angelika Kimmig', 'Giuseppe Marra', 'Robin Manhaeve', 'Pedro Zuidberg Dos Martires', 'Lennert De Smet'] | 2023-03-08 | null | null | null | null | ['probabilistic-programming'] | ['methodology'] | [-4.21934351e-02 5.53592086e-01 -5.30560315e-01 -5.75097024e-01
-7.56129324e-01 -5.82666457e-01 8.36347401e-01 -3.65559310e-02
-2.32860535e-01 1.05772972e+00 -9.31856111e-02 -6.27745628e-01
-4.58672613e-01 -1.35142577e+00 -1.08992147e+00 -4.33824807e-01
-3.52753758e-01 9.44317877e-01 2.32253924e-01 2.63417393e-01
-4.51127365e-02 5.69306791e-01 -1.64628744e+00 2.00367928e-01
6.47078395e-01 1.16283214e+00 -2.77966321e-01 5.66976726e-01
-3.59645247e-01 1.35784113e+00 -3.50946277e-01 -4.71722871e-01
-2.68197149e-01 -1.08743710e-02 -6.17755175e-01 -9.89763975e-01
3.85471880e-01 -5.39101958e-01 -3.98617566e-01 1.27039981e+00
8.66450444e-02 1.73068002e-01 9.84550178e-01 -1.71911395e+00
-4.92880970e-01 1.54936647e+00 4.25329208e-02 -1.89510267e-02
9.91333723e-02 -1.26342382e-02 1.19251299e+00 -4.42931056e-01
2.97624141e-01 1.73138082e+00 8.85539651e-01 4.66128588e-01
-1.44513750e+00 -4.82397079e-01 -4.17340808e-02 1.60641715e-01
-1.55212951e+00 -3.00784737e-01 7.41135120e-01 -2.01407194e-01
1.15066540e+00 3.31339054e-02 5.34569979e-01 1.05076969e+00
2.94481158e-01 1.24934709e+00 1.01160049e+00 -7.91648269e-01
7.76624322e-01 1.00176625e-01 2.28797466e-01 7.11630583e-01
6.82429373e-02 6.06447577e-01 -6.12965465e-01 -2.65207410e-01
6.02215767e-01 -3.42352152e-01 3.47680539e-01 -3.38627338e-01
-8.43117356e-01 9.73538339e-01 2.49935780e-02 -1.45188579e-02
-1.56990960e-01 1.13106096e+00 3.90178651e-01 -4.55885082e-02
-4.18703891e-02 2.41474569e-01 -5.16725123e-01 -4.11286116e-01
-1.06546140e+00 7.36857057e-01 1.31159127e+00 9.55302477e-01
5.40672839e-01 2.05601990e-01 -6.04199283e-02 4.94884521e-01
6.94133878e-01 9.26317751e-01 -8.50879997e-02 -1.45110667e+00
4.92897630e-02 6.82313219e-02 -7.97797143e-02 -8.17587554e-01
-5.86876385e-02 1.46792337e-01 -4.26932603e-01 5.51475227e-01
5.55179060e-01 -3.33266497e-01 -8.12887609e-01 2.18455124e+00
-2.21308157e-01 1.95582911e-01 3.30575079e-01 2.27826074e-01
5.05570352e-01 6.39529109e-01 2.28126094e-01 1.66716695e-01
1.00942361e+00 -5.18497638e-02 -6.07593656e-01 9.18289647e-03
3.23119909e-01 -3.63361277e-02 8.43165100e-01 5.93956769e-01
-1.14112937e+00 1.23226628e-01 -9.92584109e-01 -5.11997901e-02
-5.76673090e-01 -4.13478404e-01 1.21735251e+00 9.11146939e-01
-1.02105558e+00 7.58957982e-01 -1.28889990e+00 2.85552919e-01
8.16195130e-01 4.88641173e-01 7.23126680e-02 -1.22566424e-01
-1.67793369e+00 1.08794522e+00 1.09351182e+00 -8.48799795e-02
-9.71643627e-01 -6.98171914e-01 -1.19509685e+00 4.36462611e-01
6.42790020e-01 -4.10757601e-01 1.55712342e+00 -6.62361205e-01
-1.84035552e+00 3.55228364e-01 7.57289901e-02 -9.22290683e-01
4.30309445e-01 -2.18916759e-01 -1.53045282e-01 1.08194038e-01
-3.07110757e-01 9.28316534e-01 2.26876825e-01 -9.03759122e-01
-5.88617682e-01 5.08355908e-02 3.90842974e-01 -5.72433323e-02
1.37489110e-01 -7.61286020e-02 -1.89833894e-01 -1.69332266e-01
-2.82274261e-02 -5.41940510e-01 2.45805085e-02 6.41320273e-02
-7.12238848e-01 -4.77469146e-01 3.26726645e-01 -1.31710589e-01
9.07558799e-01 -1.87872577e+00 2.40620896e-02 7.77074993e-01
4.71977033e-02 -3.06818366e-01 4.37976420e-01 2.00584158e-01
1.63795114e-01 2.40520239e-01 -2.70920008e-01 -6.06354997e-02
9.07233477e-01 7.75287271e-01 -7.70885289e-01 1.89605281e-01
3.38488311e-01 9.76527512e-01 -9.60427940e-01 -9.45064902e-01
2.94661611e-01 2.92331696e-01 -7.46226966e-01 -1.77282825e-01
-1.01171112e+00 -4.64717776e-01 -3.08450639e-01 8.11841071e-01
4.16946173e-01 6.80438355e-02 3.86389047e-01 2.16733649e-01
4.68519293e-02 5.04412055e-01 -1.43397903e+00 1.35165870e+00
-4.65290517e-01 6.09084308e-01 -3.43164802e-01 -7.28105545e-01
5.57445347e-01 2.90716499e-01 -1.80842495e-03 -1.15129307e-01
3.37982327e-01 1.82211459e-01 -1.63982585e-02 -1.26420319e-01
4.16794270e-01 -4.97670352e-01 -4.81378645e-01 5.96630394e-01
4.66702789e-01 -7.14368641e-01 2.48330683e-01 2.64341116e-01
8.76125097e-01 6.63268030e-01 2.25945428e-01 -3.67844492e-01
1.39018282e-01 -1.17022626e-01 5.02021849e-01 1.29784286e+00
-5.32927141e-02 8.13725740e-02 1.17494738e+00 -2.40445256e-01
-7.13821769e-01 -1.60683882e+00 -4.58047539e-01 1.06110001e+00
-2.25429624e-01 -2.56131440e-01 -4.46883142e-01 -5.19077539e-01
2.99248368e-01 1.52913725e+00 -5.15553415e-01 -1.73950806e-01
-1.81261763e-01 -7.02733457e-01 1.26546204e+00 8.87017071e-01
2.11817846e-01 -1.37064075e+00 -4.65008169e-01 2.30767131e-01
2.66297817e-01 -7.22694576e-01 3.45834315e-01 7.39081562e-01
-7.45222509e-01 -8.50507140e-01 1.71042085e-02 -3.59713018e-01
2.04646979e-02 -1.03276896e+00 1.16434336e+00 -5.32119513e-01
2.06467301e-01 3.66306037e-01 3.84403110e-01 -7.21158087e-01
-7.90465355e-01 -1.16144173e-01 2.09749743e-01 -8.10401082e-01
6.71843529e-01 -1.04894900e+00 3.86543959e-01 -2.20503211e-01
-8.70428681e-01 5.12287579e-03 4.49755073e-01 8.41571808e-01
4.93644178e-01 6.62326157e-01 4.20357287e-01 -9.69948411e-01
6.09599948e-01 -5.04411757e-01 -1.10813200e+00 5.08197606e-01
-5.27926981e-01 5.65636635e-01 4.84024405e-01 -3.23050261e-01
-1.01306570e+00 -1.50104448e-01 -9.87254307e-02 -3.08268040e-01
-2.95517087e-01 9.93888080e-01 -4.41978842e-01 4.00652766e-01
6.82784796e-01 1.48081690e-01 1.25361010e-01 4.71986197e-02
5.86409271e-01 2.55099505e-01 8.96985769e-01 -1.52316034e+00
3.94232064e-01 2.45689318e-01 3.43579680e-01 -2.18535528e-01
-8.74524891e-01 4.65294868e-01 -4.03544664e-01 2.17650621e-03
4.00428027e-01 -5.57819784e-01 -1.31401455e+00 6.14464879e-01
-8.60356510e-01 -7.36690998e-01 -5.25006831e-01 4.36781526e-01
-9.83253062e-01 1.18771188e-01 -7.10687995e-01 -1.39654243e+00
1.78984776e-02 -9.76988614e-01 7.27882564e-01 2.50641942e-01
-4.51690018e-01 -1.36227751e+00 -1.66105136e-01 -5.15027583e-01
1.59786716e-01 2.98497379e-01 1.32670641e+00 -8.76973212e-01
-5.69742262e-01 -2.89900392e-01 -1.60195559e-01 5.16516566e-01
-5.60613811e-01 3.44753891e-01 -1.11169994e+00 5.78704059e-01
-3.44574481e-01 -7.49385238e-01 7.03335941e-01 5.24129272e-01
1.27035534e+00 -3.21071386e-01 -2.54054487e-01 4.19128209e-01
1.35813379e+00 -3.03413328e-02 6.11518621e-01 1.86145112e-01
3.37029636e-01 4.27990943e-01 1.11558616e-01 2.85946757e-01
5.15717983e-01 2.98587769e-01 4.13354307e-01 7.43719161e-01
2.04595521e-01 -4.79191989e-01 3.83073866e-01 1.56292304e-01
2.79490501e-02 -2.07445040e-01 -1.30793822e+00 4.03719217e-01
-1.85018063e+00 -1.31628621e+00 3.55327487e-01 1.89581704e+00
1.68348110e+00 7.15221405e-01 -1.28483191e-01 3.31436157e-01
5.69181383e-01 -1.27117932e-01 -5.86095989e-01 -5.82519591e-01
-1.54697686e-01 5.29607475e-01 4.08117056e-01 8.05204988e-01
-1.13898420e+00 9.42130268e-01 6.94255400e+00 1.02050483e+00
-6.35002613e-01 -2.61887491e-01 3.39268386e-01 -4.12259661e-02
-6.67138696e-01 -1.14414683e-02 -8.65040720e-01 5.31123638e-01
1.31675601e+00 -2.21722052e-01 6.89232707e-01 1.15575302e+00
-4.76743281e-01 -3.81180942e-01 -1.58165193e+00 6.68294251e-01
-1.81535870e-01 -1.54786634e+00 -2.97318902e-02 -1.42079353e-01
4.35455561e-01 6.86818361e-02 7.30908588e-02 7.54819810e-01
1.38571870e+00 -1.45326090e+00 1.24658203e+00 8.89467239e-01
9.81926501e-01 -1.28782272e+00 1.05689371e+00 4.43035334e-01
-3.49592894e-01 -3.57678719e-02 -2.15638056e-01 -1.83515936e-01
7.72445798e-02 8.13293934e-01 -8.72248590e-01 2.53074318e-01
3.92621249e-01 5.06057799e-01 -1.77832041e-02 5.82418978e-01
-8.08785081e-01 8.79427016e-01 -1.02286756e+00 -2.86614358e-01
1.69740453e-01 2.99727768e-01 3.11101943e-01 1.33843994e+00
-1.07750408e-02 -1.49820879e-01 1.08108886e-01 1.55703318e+00
8.37767348e-02 -8.98563981e-01 -4.68215555e-01 -3.70839626e-01
9.09723938e-01 5.44654965e-01 -4.29029286e-01 -4.71210897e-01
-2.18207706e-02 1.19991668e-01 3.03054869e-01 3.53290945e-01
-7.74624050e-01 -6.11763775e-01 3.34149987e-01 -5.01447260e-01
2.19181910e-01 -2.40124822e-01 -7.42451370e-01 -8.25013816e-01
-3.44437271e-01 -6.71452820e-01 4.18497175e-01 -7.23357022e-01
-1.35801268e+00 3.38178128e-02 7.47819960e-01 -1.96888953e-01
-9.03056085e-01 -9.45227087e-01 -5.00577688e-01 1.20776343e+00
-1.39507246e+00 -1.19781816e+00 4.70498830e-01 1.78589851e-01
-2.73033887e-01 -2.12721542e-01 1.12866187e+00 -2.27961540e-01
-2.71414608e-01 6.57194078e-01 1.32019401e-01 1.16847061e-01
1.74505413e-01 -1.82413888e+00 7.43869394e-02 6.89567029e-01
-3.92541662e-02 8.64700675e-01 9.13292050e-01 -6.00495458e-01
-1.48141623e+00 -7.43670404e-01 7.75515318e-01 -6.39524460e-01
1.07540059e+00 -3.40389222e-01 -6.44927680e-01 9.92499471e-01
-3.09044570e-01 6.14405870e-02 6.22010946e-01 6.21919751e-01
-7.08606422e-01 1.25980064e-01 -1.34421074e+00 6.89876020e-01
5.77205181e-01 -7.64520526e-01 -1.07129037e+00 3.89317535e-02
5.79378903e-01 -7.58115172e-01 -8.20047081e-01 2.16357440e-01
1.05166352e+00 -8.57306480e-01 8.68296921e-01 -5.67500234e-01
7.48663723e-01 -2.51242757e-01 -7.37402618e-01 -8.02787483e-01
8.41240808e-02 -4.67737019e-01 -8.93860281e-01 1.17984438e+00
4.26793069e-01 -6.74885392e-01 7.21035779e-01 9.36497033e-01
-1.36958912e-01 -7.28340447e-01 -1.08749449e+00 -6.95826769e-01
4.86876488e-01 -1.48643255e+00 7.06820786e-01 4.73524272e-01
1.70527071e-01 -3.74362588e-01 1.06858060e-01 3.06868136e-01
7.08084345e-01 5.75470813e-02 3.02159607e-01 -1.17802083e+00
-6.56589627e-01 -6.35600388e-01 -6.39565468e-01 -6.31412625e-01
7.87189364e-01 -1.06208003e+00 5.17639399e-01 -1.19143617e+00
2.12825537e-02 -7.23489642e-01 -4.39332783e-01 9.57693875e-01
2.31161937e-01 6.15894720e-02 -1.10696360e-01 -2.12027594e-01
-6.67281508e-01 4.37698066e-01 5.27192771e-01 -5.38869500e-02
6.26888871e-02 3.22844530e-03 -7.35345781e-01 1.19317281e+00
6.85377300e-01 -5.04869640e-01 -4.53911871e-01 -1.88436940e-01
8.49094212e-01 -2.89422143e-02 7.76246250e-01 -1.02841806e+00
2.35530153e-01 -4.70973939e-01 4.21355516e-01 -7.38953769e-01
1.86781317e-01 -5.66174626e-01 -2.57643670e-01 1.75286338e-01
-6.55144691e-01 -4.45922285e-01 5.53574681e-01 4.27916467e-01
-5.15701920e-02 -5.65021217e-01 5.15846729e-01 -2.49729790e-02
-7.82224596e-01 -1.89478397e-01 -5.34963965e-01 2.41364762e-01
5.68866909e-01 3.83676812e-02 -1.49715662e-01 -3.46274495e-01
-7.62297928e-01 3.36870879e-01 2.58304209e-01 -2.73090214e-01
4.38225627e-01 -1.17453897e+00 -3.44256222e-01 -9.90419313e-02
-1.86878100e-01 4.78574336e-01 -9.06029493e-02 3.26785326e-01
-5.29709756e-01 4.62988675e-01 -6.02108352e-02 -5.12526155e-01
-4.69691068e-01 2.05052286e-01 4.61833566e-01 -8.43684152e-02
-4.78704154e-01 8.95651042e-01 -3.35631520e-01 -7.21317649e-01
6.49217188e-01 -7.07721174e-01 6.80376366e-02 -3.12666386e-01
6.91450238e-01 1.80635273e-01 -3.72351557e-01 2.11718485e-01
-4.23606038e-01 -2.85420924e-01 5.16970791e-02 -6.60107076e-01
1.37283790e+00 2.99095452e-01 -3.93893272e-01 1.07550645e+00
4.17284429e-01 -8.76525864e-02 -1.38922501e+00 -2.93882370e-01
2.97715962e-01 4.72892635e-02 1.47762030e-01 -1.35413206e+00
-3.10872734e-01 9.71614659e-01 9.46036875e-02 7.62039125e-02
6.08301520e-01 1.42043754e-01 3.93946558e-01 9.37730789e-01
4.56133723e-01 -8.54398012e-01 -5.31030416e-01 1.04324937e+00
5.64345181e-01 -8.28447223e-01 6.12731948e-02 9.56927910e-02
-3.87055516e-01 1.47232533e+00 3.73550773e-01 -5.30618876e-02
5.91785967e-01 9.05011594e-01 -3.81074727e-01 -8.24357718e-02
-8.36439371e-01 1.66081592e-01 1.21374540e-01 8.72168660e-01
2.99005002e-01 2.65156657e-01 5.94805956e-01 1.04862547e+00
-6.11581326e-01 4.74974900e-01 4.74986225e-01 1.05480278e+00
-3.46683592e-01 -9.15843666e-01 -3.78768772e-01 4.60574895e-01
-5.97269595e-01 -4.37834710e-01 1.02985270e-01 8.91970694e-01
2.18202531e-01 5.21540761e-01 1.03823043e-01 7.36111328e-02
-2.82874107e-01 4.98447299e-01 7.46302187e-01 -7.39959598e-01
1.03588246e-01 -5.26208460e-01 3.18649918e-01 -3.57734650e-01
-1.73414633e-01 -7.93156981e-01 -1.53663611e+00 -5.38030207e-01
-1.84166610e-01 4.37140651e-02 6.92075253e-01 1.19767904e+00
-4.54588644e-02 6.48683965e-01 -1.92071378e-01 -6.73594534e-01
-1.00860679e+00 -7.48515368e-01 -9.80746031e-01 -4.98950422e-01
1.49196014e-01 -7.87749350e-01 -2.84893990e-01 -1.15048565e-01] | [8.674386978149414, 6.923358917236328] |
460870af-1390-44d1-92af-46a21a4d6bb7 | towards-argument-aware-abstractive | 2306.00672 | null | https://arxiv.org/abs/2306.00672v1 | https://arxiv.org/pdf/2306.00672v1.pdf | Towards Argument-Aware Abstractive Summarization of Long Legal Opinions with Summary Reranking | We propose a simple approach for the abstractive summarization of long legal opinions that considers the argument structure of the document. Legal opinions often contain complex and nuanced argumentation, making it challenging to generate a concise summary that accurately captures the main points of the legal opinion. Our approach involves using argument role information to generate multiple candidate summaries, then reranking these candidates based on alignment with the document's argument structure. We demonstrate the effectiveness of our approach on a dataset of long legal opinions and show that it outperforms several strong baselines. | ['Diane Litman', 'Yang Zhong', 'Mohamed Elaraby'] | 2023-06-01 | null | null | null | null | ['abstractive-text-summarization'] | ['natural-language-processing'] | [ 3.20969403e-01 8.18790197e-01 -8.33052933e-01 -4.54430580e-01
-1.46457040e+00 -1.14104402e+00 8.84754479e-01 1.00489628e+00
-2.48274133e-01 1.18145919e+00 1.44531417e+00 -6.95276916e-01
-1.59514278e-01 -4.44875807e-01 -4.64861095e-01 -1.24565512e-01
4.66598392e-01 6.17369473e-01 1.59320906e-01 -6.28513992e-01
1.07658386e+00 2.52854347e-01 -1.22492898e+00 8.20902109e-01
1.59712934e+00 4.16879922e-01 -5.75728834e-01 7.38964379e-01
-3.80082011e-01 1.31716669e+00 -1.57997525e+00 -1.30311084e+00
-1.64520089e-03 -4.57619637e-01 -1.07669723e+00 -2.70074368e-01
1.04903281e+00 -5.37813246e-01 1.75625160e-01 9.30905819e-01
5.02349973e-01 -2.25329578e-01 1.08553851e+00 -8.55173707e-01
-4.04142022e-01 1.13414454e+00 -5.76050282e-01 4.97513950e-01
7.88423777e-01 -3.79124731e-01 1.54494691e+00 -3.18795085e-01
1.03723323e+00 1.55051076e+00 5.48244357e-01 2.87126958e-01
-8.75068903e-01 -1.26634032e-01 6.14473403e-01 -1.03219420e-01
-1.65801272e-01 -6.06071115e-01 8.71281564e-01 -3.38061959e-01
8.88493061e-01 2.10348397e-01 8.65655541e-01 9.63684499e-01
4.08759654e-01 1.03867221e+00 7.33398736e-01 -3.00425798e-01
3.21252704e-01 -3.60577226e-01 9.92696643e-01 2.30387941e-01
1.06104875e+00 -7.00223386e-01 -2.54733443e-01 -1.21961355e+00
-2.78195918e-01 -3.36395681e-01 -1.33902219e-03 2.01700687e-01
-8.56556952e-01 9.72047031e-01 -1.25235077e-02 2.25496024e-01
-6.90291405e-01 1.15358971e-01 6.39475882e-01 3.14515412e-01
5.86336255e-01 7.98465908e-01 -2.22211510e-01 -3.57176602e-01
-1.12854683e+00 1.05016947e+00 1.16020894e+00 5.29250681e-01
1.91304192e-01 -3.46915156e-01 -7.75480211e-01 7.81170130e-01
4.14599814e-02 6.81246877e-01 3.61641273e-02 -1.41083777e+00
1.01087308e+00 9.97385025e-01 4.05779719e-01 -1.02053618e+00
1.16974838e-01 -3.14691216e-01 -2.51754761e-01 5.26172295e-02
1.06921077e-01 -5.30526042e-01 -7.29739785e-01 1.18590832e+00
1.80629224e-01 -8.77820194e-01 6.07987702e-01 1.56528294e-01
1.12633288e+00 7.47124374e-01 6.16012178e-02 -4.63466525e-01
1.00589085e+00 -7.82887876e-01 -8.66536975e-01 -1.54087707e-01
6.22789562e-01 -7.52124965e-01 6.83267534e-01 2.38195583e-01
-1.53004754e+00 1.93141595e-01 -1.25796092e+00 -4.49985355e-01
1.35376379e-01 1.78677410e-01 5.13760984e-01 2.93910593e-01
-7.35681713e-01 5.79495251e-01 -1.65960476e-01 5.27902804e-02
7.28855073e-01 -9.76147801e-02 -8.45817104e-02 -1.44728022e-02
-9.78204429e-01 1.01819038e+00 2.30759576e-01 -1.78451955e-01
-4.98597231e-03 -7.30589867e-01 -8.63344371e-01 1.81040496e-01
5.76440573e-01 -8.16197693e-01 1.69413984e+00 -3.35638314e-01
-1.02316940e+00 5.49066007e-01 -5.93784571e-01 -5.15657425e-01
6.45847440e-01 -4.85766679e-01 -1.29438385e-01 4.54121858e-01
7.21796691e-01 2.61402130e-01 4.78556842e-01 -1.27885449e+00
-8.01739037e-01 -2.28245169e-01 7.42468834e-01 3.56258154e-01
2.71975016e-03 2.88093597e-01 1.19615868e-01 -7.23448634e-01
-2.24476665e-01 -5.66745758e-01 -4.99257356e-01 -5.91970444e-01
-6.20456398e-01 -6.51435018e-01 6.31889284e-01 -6.20390296e-01
1.64765048e+00 -1.44364953e+00 -8.21556151e-02 2.59562820e-01
3.57282102e-01 3.48358929e-01 -1.01944104e-01 9.62698221e-01
2.28121102e-01 5.13486326e-01 -3.29967737e-01 2.02481896e-01
2.47696623e-01 2.73284554e-01 -1.10043311e+00 4.32524038e-03
-3.03759929e-02 1.15476847e+00 -1.11849201e+00 -8.52588713e-01
-4.00522709e-01 -1.54872864e-01 -6.10194802e-01 -2.52213717e-01
-5.51070750e-01 -5.09796329e-02 -9.27147269e-01 5.86433887e-01
3.51155698e-01 -4.49698903e-02 1.93035930e-01 -3.50554466e-01
-7.38230869e-02 8.25913012e-01 -4.94503438e-01 1.24421942e+00
1.87697373e-02 6.16829038e-01 -4.16473858e-02 -7.35784292e-01
6.33535147e-01 2.10722864e-01 2.86381334e-01 -5.46224952e-01
7.74173904e-03 5.10482788e-01 -4.28696387e-02 -5.19388795e-01
7.64554977e-01 1.62811443e-01 -6.63599610e-01 1.05658054e+00
-6.03100002e-01 -6.83038414e-01 1.20225453e+00 1.00473082e+00
1.10811961e+00 -3.15340936e-01 6.92487895e-01 -2.36119732e-01
5.06095409e-01 6.04376554e-01 3.96335930e-01 1.14942920e+00
2.35963911e-01 3.63724530e-01 1.27426541e+00 -7.10878313e-01
-8.14207494e-01 -7.78989434e-01 1.02662683e-01 4.68977660e-01
-4.45829518e-02 -9.62642372e-01 -7.74878979e-01 -1.37708676e+00
2.92921811e-01 1.09063315e+00 -7.39019573e-01 1.76444367e-01
-9.77615893e-01 -3.81537616e-01 4.67884064e-01 3.81751180e-01
2.33131334e-01 -8.17956090e-01 -9.08099532e-01 2.64389068e-01
-4.69664931e-01 -7.36831903e-01 -4.76372689e-01 -5.04863560e-01
-8.07586312e-01 -1.53521597e+00 -6.32547200e-01 -2.04971597e-01
7.90685236e-01 8.77277926e-02 1.22922421e+00 4.73223984e-01
3.12186509e-01 1.61161080e-01 -2.22211674e-01 -8.29242587e-01
-9.12473857e-01 2.94386566e-01 -5.80400050e-01 -6.33070767e-01
-1.06241494e-01 -4.32453334e-01 -3.20303202e-01 -4.25847441e-01
-1.00635421e+00 -2.92199582e-01 5.91993868e-01 5.88111162e-01
1.71790510e-01 -4.99104768e-01 8.50763381e-01 -1.53537691e+00
1.96536970e+00 -2.85341829e-01 -2.38814145e-01 9.20309365e-01
-5.30838072e-01 4.19084698e-01 4.07615900e-01 -7.48826517e-03
-1.26448274e+00 -8.02716672e-01 -3.26264858e-01 5.76508462e-01
4.25530881e-01 9.28345025e-01 2.62186080e-01 5.84986329e-01
8.57424557e-01 -3.40928197e-01 -2.46822342e-01 -8.30288082e-02
6.66148603e-01 6.59534156e-01 6.08586729e-01 -9.79061604e-01
6.03523016e-01 4.79556084e-01 -2.18402773e-01 -5.76141000e-01
-1.62388766e+00 -4.23101038e-02 -4.27117378e-01 -1.43277317e-01
4.07094389e-01 -5.43802500e-01 -2.99869567e-01 -2.47296274e-01
-1.77244914e+00 3.53930622e-01 -5.27564943e-01 4.78920750e-02
-1.81938663e-01 7.64544308e-01 -5.92402756e-01 -7.15873241e-01
-9.71366942e-01 -8.62582564e-01 1.20323503e+00 2.03745529e-01
-1.01813853e+00 -8.16803515e-01 4.65040445e-01 7.29516149e-01
3.49361748e-01 6.71653986e-01 1.39252257e+00 -9.19865310e-01
7.13662580e-02 -5.86674392e-01 -5.99135272e-03 1.17094710e-01
1.33862540e-01 5.86226225e-01 -2.63603240e-01 1.37175590e-01
-1.80337012e-01 -4.51468080e-01 1.07408690e+00 6.19002342e-01
3.75717640e-01 -1.07501733e+00 -3.55346084e-01 -3.51993442e-01
6.18433237e-01 3.68883908e-02 5.21108627e-01 3.56112003e-01
3.27822626e-01 8.76335144e-01 8.33122909e-01 2.21038103e-01
7.00109482e-01 1.75636992e-01 -9.81123187e-03 1.41449973e-01
2.01310590e-03 -2.34896243e-01 2.11489618e-01 7.10795462e-01
-1.25928491e-01 -4.68625605e-01 -6.77447557e-01 7.70687282e-01
-2.23062587e+00 -1.54159391e+00 -3.02021682e-01 1.49710596e+00
6.22229815e-01 4.38607931e-01 5.85683472e-02 6.59892708e-02
5.85030496e-01 7.78825939e-01 -5.50850451e-01 -1.03585637e+00
-4.57340628e-01 7.32484460e-02 7.10118636e-02 6.61686897e-01
-7.14571238e-01 7.33421564e-01 7.69197130e+00 5.71676970e-01
-5.78094840e-01 -3.44352245e-01 6.66613579e-01 -2.16255203e-01
-1.23647058e+00 3.95550370e-01 -7.73367465e-01 2.53060728e-01
5.53243339e-01 -1.02157354e+00 -5.62528193e-01 7.57597208e-01
2.39393353e-01 -3.77699822e-01 -8.15121830e-01 3.60009164e-01
3.97020221e-01 -2.04446149e+00 8.06337118e-01 1.54339775e-01
1.24086654e+00 -3.05507034e-01 -3.85685235e-01 5.45491204e-02
8.90304506e-01 -7.41387963e-01 9.42449629e-01 2.06152469e-01
2.45279580e-01 -7.05766499e-01 9.43645358e-01 2.98765898e-01
-6.31209493e-01 -3.64052862e-01 -2.12995231e-01 -1.64692193e-01
7.40970433e-01 7.64134109e-01 -5.48777878e-01 7.50634849e-01
1.50262833e-01 1.01475108e+00 -5.23334920e-01 7.86605477e-01
-8.26736152e-01 4.88638669e-01 -1.33088278e-02 -3.59317899e-01
5.38913429e-01 -2.16300622e-01 8.35009873e-01 1.08146286e+00
2.82261908e-01 3.97165716e-01 1.76398560e-01 3.26072842e-01
-4.56499994e-01 2.50868738e-01 -7.14639902e-01 -1.92375273e-01
5.89828372e-01 9.36625540e-01 -4.53000844e-01 -1.00331533e+00
3.70595567e-02 2.96113461e-01 2.52748340e-01 2.03342959e-01
-4.37197447e-01 -3.21184486e-01 3.45885843e-01 -1.34194419e-01
2.31284052e-01 2.48701781e-01 -2.09668875e-01 -1.28764677e+00
5.98236203e-01 -1.28930807e+00 9.31196570e-01 -7.81286657e-01
-1.05269372e+00 7.39279032e-01 3.41616511e-01 -1.03441894e+00
-6.19316995e-01 -2.53053624e-02 -1.25406528e+00 3.97763461e-01
-1.41055596e+00 -9.52139437e-01 3.49145025e-01 -1.40446678e-01
6.71844482e-01 1.51777640e-01 4.15891379e-01 -2.66784877e-01
-2.71184832e-01 1.59551740e-01 -2.86471695e-01 -1.32496417e-01
6.16497874e-01 -1.37241876e+00 6.02174938e-01 8.80696654e-01
3.97498067e-03 1.05389428e+00 9.69693124e-01 -1.05770338e+00
-9.24479425e-01 -4.50305134e-01 1.46931195e+00 -5.71790159e-01
7.42326677e-01 2.17922211e-01 -7.18654811e-01 4.57469910e-01
9.09535825e-01 -8.83867621e-01 7.63912320e-01 1.00702725e-01
-5.36009848e-01 -2.58432310e-02 -1.01799178e+00 9.33022201e-01
8.68542671e-01 -1.95263699e-01 -1.65382314e+00 4.01636571e-01
4.78860945e-01 -3.42415422e-01 -4.73101497e-01 4.67306703e-01
7.49813557e-01 -7.89576530e-01 6.08788371e-01 -8.23976874e-01
8.64576995e-01 -1.83549091e-01 2.62432098e-01 -1.37156630e+00
1.38438568e-01 -7.92318761e-01 -1.61096305e-01 1.06024218e+00
1.00521696e+00 -6.17365658e-01 6.60497844e-01 8.18756998e-01
-2.64722705e-01 -8.51015508e-01 -6.70602143e-01 -1.55059189e-01
2.75692612e-01 6.70943856e-02 6.90075636e-01 6.23664558e-01
5.78305185e-01 8.52668524e-01 -3.11500654e-02 -3.13235432e-01
5.15378535e-01 8.11100721e-01 8.45710814e-01 -1.65455246e+00
3.14864486e-01 -7.48831630e-01 7.62935355e-02 -1.00402939e+00
4.42039102e-01 -5.26293516e-01 -3.96341860e-01 -2.59613371e+00
3.07766736e-01 2.54566312e-01 2.80882359e-01 2.88195759e-01
-4.44768071e-01 -4.58847106e-01 -5.72205661e-03 4.10848707e-01
-1.02737510e+00 2.97050893e-01 1.17081618e+00 -6.51739657e-01
-1.35377809e-01 1.97141171e-02 -1.67112839e+00 8.80649447e-01
5.52365422e-01 -4.96289164e-01 -5.31447351e-01 -5.19622147e-01
6.90507054e-01 1.47977903e-01 -4.04994607e-01 -3.73755395e-01
5.61397731e-01 -2.50476390e-01 1.48775969e-02 -1.25771141e+00
3.79109383e-02 -1.73394550e-02 -3.20259243e-01 5.61666131e-01
-9.62949574e-01 6.80886090e-01 -1.74543504e-02 3.88317525e-01
-5.95610678e-01 -4.05717283e-01 2.19586007e-02 -1.81293458e-01
6.68180659e-02 -2.19265655e-01 -5.11838198e-01 5.45324147e-01
6.02409720e-01 -2.20339224e-01 -1.24107635e+00 -7.13090360e-01
-1.77743614e-01 6.80201232e-01 6.12207651e-01 1.43764704e-01
5.84911883e-01 -9.52624917e-01 -1.38208222e+00 -6.82002187e-01
1.11068226e-01 -7.32010081e-02 -4.07447256e-02 5.08906126e-01
-5.78931332e-01 5.36034882e-01 1.66496038e-01 -7.53447041e-02
-1.56340241e+00 1.16804857e-02 -1.17216960e-01 -8.21214855e-01
-6.84779048e-01 2.17522249e-01 -3.76939893e-01 -1.20113470e-01
-5.20678796e-02 -6.09601378e-01 -9.43968773e-01 6.25486135e-01
8.15113246e-01 4.21817899e-01 -1.82042830e-02 -6.10274851e-01
-2.86813468e-01 7.12908030e-01 -5.72248697e-01 -4.01885718e-01
1.61675096e+00 1.89552039e-01 -6.68642581e-01 6.34634346e-02
7.68526733e-01 1.00970352e+00 -7.21483588e-01 1.02475837e-01
5.27830482e-01 -2.00276151e-01 -4.61803198e-01 -8.68543446e-01
-4.43523169e-01 2.32674211e-01 -8.08735430e-01 3.26899141e-01
5.76592326e-01 2.69336760e-01 7.72368312e-01 9.74408209e-01
-5.49837984e-02 -1.17755103e+00 -2.14462616e-02 5.90325713e-01
1.35267568e+00 -7.71221042e-01 6.83628678e-01 -4.21318382e-01
-8.35537970e-01 1.25437999e+00 9.48778018e-02 -1.50263831e-01
1.16695948e-01 1.23063870e-01 5.26185751e-01 -8.48858178e-01
-1.09962726e+00 2.35529363e-01 4.34160948e-01 1.91714093e-02
4.09441173e-01 -3.76151085e-01 -1.30073893e+00 3.52603644e-01
-5.68486750e-01 -2.85169959e-01 1.12856662e+00 1.27596009e+00
-5.77821136e-01 -1.23614502e+00 -2.80856341e-01 8.46730649e-01
-1.00014818e+00 -1.73307240e-01 -1.28956556e+00 5.37286758e-01
-5.32332718e-01 1.43136322e+00 -4.45266664e-01 3.27956676e-01
7.95814514e-01 -1.97526202e-01 2.93485522e-01 -7.20485032e-01
-9.69971359e-01 -4.03784029e-02 1.10753334e+00 -3.25471550e-01
-6.86849654e-01 -7.87689507e-01 -1.16680717e+00 -1.29161060e-01
-5.70630245e-02 1.10696495e+00 4.71699625e-01 1.23493469e+00
5.28300881e-01 1.91358000e-01 1.65043294e-01 -4.02722567e-01
-7.86570668e-01 -7.72182584e-01 1.26346856e-01 5.88365912e-01
5.83074212e-01 -3.49270523e-01 -2.96117276e-01 -2.70794034e-01] | [12.16167163848877, 9.594408988952637] |
6ca88a44-bac6-4eec-bf10-dffc5d36f152 | temporal-wasserstein-non-negative-matrix | 1912.03463 | null | https://arxiv.org/abs/1912.03463v1 | https://arxiv.org/pdf/1912.03463v1.pdf | Temporal Wasserstein non-negative matrix factorization for non-rigid motion segmentation and spatiotemporal deconvolution | Motion segmentation for natural images commonly relies on dense optic flow to yield point trajectories which can be grouped into clusters through various means including spectral clustering or minimum cost multicuts. However, in biological imaging scenarios, such as fluorescence microscopy or calcium imaging, where the signal to noise ratio is compromised and intensity fluctuations occur, optical flow may be difficult to approximate. To this end, we propose an alternative paradigm for motion segmentation based on optimal transport which models the video frames as time-varying mass represented as histograms. Thus, we cast motion segmentation as a temporal non-linear matrix factorization problem with Wasserstein metric loss. The dictionary elements of this factorization yield segmentation of motion into coherent objects while the loading coefficients allow for time-varying intensity signal of the moving objects to be captured. We demonstrate the use of the proposed paradigm on a simulated multielectrode drift scenario, as well as calcium indicating fluorescence microscopy videos of the nematode Caenorhabditis elegans (C. elegans). The latter application has the added utility of extracting neural activity of the animal in freely conducted behavior. | ['Erdem Varol', 'Amin Nejatbakhsh', 'Conor McGrory'] | 2019-12-07 | null | null | null | null | ['motion-segmentation'] | ['computer-vision'] | [ 2.45836511e-01 -4.08101320e-01 2.28039801e-01 -1.55992597e-01
-1.94043249e-01 -9.03794706e-01 1.16912372e-01 -8.70879889e-02
-1.00411177e+00 9.39071655e-01 -5.99601269e-01 1.50171652e-01
-2.16323540e-01 -8.29325467e-02 -8.30822349e-01 -1.24488318e+00
-2.50045151e-01 1.79093778e-01 1.34244129e-01 5.02028465e-01
5.42829573e-01 6.01377606e-01 -1.45379686e+00 -1.57351688e-01
6.11522198e-01 6.19770288e-01 4.91376638e-01 1.03321075e+00
7.49066919e-02 3.68102670e-01 -3.81691694e-01 1.27271935e-01
2.63225913e-01 -3.68433386e-01 -5.91432512e-01 2.17022344e-01
2.49708876e-01 -4.35059756e-01 9.92534757e-02 1.32289445e+00
2.07325414e-01 4.35884237e-01 7.15407729e-01 -1.31638753e+00
4.64897417e-02 2.29204699e-01 -8.89282286e-01 4.24711585e-01
3.39353047e-02 1.05372809e-01 5.83631098e-01 -6.00366890e-01
9.95989501e-01 1.20405912e+00 3.43931735e-01 8.53018463e-01
-1.71867228e+00 -2.34725773e-01 2.27667481e-01 4.36732359e-02
-1.00564659e+00 -3.45692515e-01 5.25768578e-01 -8.00530970e-01
5.16660094e-01 4.65924153e-03 7.50968158e-01 8.74809742e-01
5.14059842e-01 8.38380873e-01 7.80409455e-01 -2.18199380e-02
5.26858509e-01 -3.05564940e-01 2.50358790e-01 5.38324833e-01
3.32397670e-01 -3.29762071e-01 -5.04183471e-01 -2.71398157e-01
8.36649239e-01 1.35925725e-01 -4.96156693e-01 -5.47112584e-01
-1.38681638e+00 5.49489439e-01 2.75845621e-02 2.44591519e-01
-1.45448700e-01 3.56988013e-01 2.06774607e-01 -9.65675786e-02
3.87461856e-02 1.85530424e-01 -2.15110570e-01 -5.97761683e-02
-1.18144345e+00 4.98722136e-01 4.66369212e-01 6.40198171e-01
8.67040575e-01 2.40337532e-02 2.20014855e-01 2.67857224e-01
7.32620120e-01 3.90569806e-01 7.07064569e-01 -1.69605052e+00
1.77855305e-02 4.65712219e-01 4.26428318e-01 -8.20821941e-01
-4.97352600e-01 3.29592943e-01 -6.74261272e-01 5.23527741e-01
9.45769429e-01 -4.16443229e-01 -8.11000466e-01 1.98465657e+00
5.62117934e-01 4.30601925e-01 4.98177409e-02 1.08713555e+00
7.77267051e-05 5.73762238e-01 -4.20952998e-02 -7.51312852e-01
9.79283333e-01 -3.88450950e-01 -9.45052505e-01 7.19950795e-02
6.65725052e-01 -4.24118608e-01 5.63884616e-01 5.01890838e-01
-1.08684075e+00 -2.25454181e-01 -1.04922676e+00 -4.45843004e-02
-1.74796596e-01 -1.13835596e-01 4.88641411e-01 3.15700024e-01
-9.95959401e-01 9.86021280e-01 -1.73224735e+00 -3.82697344e-01
4.09858763e-01 4.83910352e-01 -3.38079959e-01 2.31681064e-01
-4.38239485e-01 3.67470652e-01 2.97354311e-01 3.40760857e-01
-1.01431108e+00 -6.12091899e-01 -5.39473712e-01 -3.22002828e-01
-2.89669901e-01 -7.58671880e-01 8.53634715e-01 -8.96126390e-01
-1.33537161e+00 7.35137999e-01 -3.55984300e-01 -5.21224201e-01
5.96501470e-01 -1.56430304e-01 3.56195986e-01 5.67925513e-01
2.14646220e-01 9.76585746e-01 9.03254330e-01 -1.15137041e+00
-6.01783395e-01 -7.34921515e-01 -1.39743641e-01 2.70900905e-01
-1.52803570e-01 -1.79571271e-01 -4.81504090e-02 -3.44780773e-01
1.94883928e-01 -1.16181946e+00 -4.96606290e-01 4.09932733e-01
-2.02442095e-01 2.24510118e-01 1.18103111e+00 -2.98967898e-01
9.70832765e-01 -2.20151401e+00 6.93378508e-01 -1.04782678e-01
4.05525774e-01 -1.43947480e-02 1.28008053e-01 5.55641577e-03
2.84585804e-01 -1.31442234e-01 -6.25791609e-01 -2.19197288e-01
-3.84873152e-01 2.56601751e-01 7.83672631e-02 9.38587070e-01
1.81710005e-01 4.34035063e-01 -1.01002443e+00 -5.44258475e-01
1.04849130e-01 5.51772654e-01 -6.23203695e-01 -8.95408243e-02
-2.61457592e-01 9.61301386e-01 -5.71678042e-01 3.90395135e-01
6.82596803e-01 -1.41586185e-01 9.81038660e-02 -2.09598422e-01
-3.51793081e-01 -6.53589189e-01 -1.12436926e+00 2.03370476e+00
2.26243690e-01 7.09286094e-01 6.63240135e-01 -1.09049201e+00
3.74307156e-01 4.72834706e-02 1.05360091e+00 1.53748065e-01
2.49499738e-01 2.53517449e-01 2.19457611e-01 -5.43020368e-01
4.89392042e-01 -3.12640518e-02 3.41990918e-01 3.59972477e-01
1.22695848e-01 9.70656350e-02 5.26630461e-01 2.80626178e-01
1.11554921e+00 5.14001906e-01 -4.68084961e-02 -5.46981037e-01
2.40857914e-01 4.15123887e-02 6.69897854e-01 3.06638211e-01
-5.65995753e-01 4.34208572e-01 3.49120170e-01 -1.09350495e-01
-1.05192530e+00 -9.44020808e-01 -3.95369560e-01 3.35385650e-01
3.41235638e-01 1.31100625e-01 -1.12222695e+00 -9.31933299e-02
-1.37631238e-01 8.23713690e-02 -4.58084464e-01 1.86630577e-01
-5.49954712e-01 -1.17617309e+00 4.85680908e-01 1.29069462e-01
1.82659507e-01 -8.10226738e-01 -1.31033051e+00 5.22070408e-01
-2.08841085e-01 -1.18434513e+00 -1.80870563e-01 3.94864559e-01
-1.28510153e+00 -1.04855216e+00 -9.51836646e-01 -6.85315967e-01
7.81915188e-01 3.51595312e-01 3.63361061e-01 -1.81657344e-01
-6.90107524e-01 6.42306328e-01 -6.08847775e-02 2.24267645e-03
1.45312607e-01 -4.36205089e-01 3.26331973e-01 2.72994548e-01
3.39305729e-01 -7.16933787e-01 -9.38889205e-01 2.12814942e-01
-1.23956931e+00 -2.99366117e-01 1.69221405e-03 8.01179469e-01
9.14560020e-01 -1.08741365e-01 2.80431181e-01 -5.34367204e-01
2.87902802e-01 -4.71143335e-01 -7.91370451e-01 -1.08289056e-01
-5.89954183e-02 -8.38470981e-02 4.05891418e-01 -8.15721035e-01
-8.45684826e-01 6.64504707e-01 4.55554396e-01 -6.52423918e-01
-1.88664094e-01 1.89561307e-01 1.56137869e-01 -3.00616920e-01
5.54186344e-01 1.21420249e-01 1.50509521e-01 -4.46537770e-02
2.60484844e-01 4.10521805e-01 6.52404666e-01 -5.42524457e-01
4.89129722e-01 1.08109057e+00 3.28153074e-01 -1.26677024e+00
-9.77069214e-02 -6.54634953e-01 -8.43805909e-01 -5.97012877e-01
1.28300381e+00 -5.80187500e-01 -1.12108111e+00 6.72147930e-01
-1.28792274e+00 -4.22391355e-01 -6.28620461e-02 8.20161223e-01
-9.16201591e-01 7.21700490e-01 -8.37374270e-01 -1.10891688e+00
-2.44891681e-02 -1.34681559e+00 9.98323321e-01 4.79934752e-01
-1.08236514e-01 -1.02864587e+00 1.38769522e-01 1.13709711e-01
-8.78370553e-02 5.44456959e-01 7.26471782e-01 1.26926294e-02
-8.02244365e-01 6.79949969e-02 1.16221674e-01 9.63533893e-02
-8.23970139e-03 5.17230868e-01 -9.64824319e-01 -2.63558596e-01
2.53045619e-01 -1.46611899e-01 8.48937333e-01 1.17228067e+00
6.74233556e-01 1.92574754e-01 -5.47700584e-01 6.66156113e-01
1.40784848e+00 4.89615649e-01 3.45767915e-01 2.13338330e-01
9.08497095e-01 1.01855087e+00 5.66785753e-01 5.10318756e-01
-2.85776425e-02 5.08165121e-01 4.89765912e-01 4.50057417e-01
3.81682843e-01 3.86322737e-01 4.71909314e-01 6.70249164e-01
-1.54105186e-01 -3.05658579e-01 -6.27312958e-01 6.94565654e-01
-2.06763387e+00 -9.21112359e-01 -4.11878705e-01 2.21974730e+00
5.32563686e-01 -1.40261009e-01 2.66785681e-01 2.46548772e-01
8.31414759e-01 -3.00860614e-01 -9.01538789e-01 3.61261629e-02
-8.71027112e-02 -4.26028818e-01 5.14815331e-01 5.26901782e-01
-1.01968110e+00 6.49675071e-01 5.41801596e+00 3.39297503e-01
-1.07750762e+00 -1.87923968e-01 3.83779049e-01 -2.28577077e-01
-5.32164052e-02 -2.31487639e-02 -8.77512932e-01 6.95991039e-01
8.20381880e-01 -1.92657724e-01 3.94648165e-01 3.00698608e-01
6.16331577e-01 -5.59154987e-01 -1.20292115e+00 1.18742800e+00
-3.63364607e-01 -9.59690750e-01 -7.12690428e-02 2.74373442e-01
4.98035043e-01 -1.02359429e-01 1.55431867e-01 -5.42354107e-01
1.74410753e-02 -5.22755206e-01 5.09647489e-01 6.40993357e-01
2.84073800e-01 -3.70156556e-01 2.48222411e-01 5.04671276e-01
-9.61058855e-01 -9.09190029e-02 -5.24175406e-01 4.16439474e-02
5.42914927e-01 5.21956921e-01 -3.44948798e-01 4.22798693e-02
5.80305398e-01 1.04041517e+00 -4.97657713e-03 1.25606441e+00
6.11334741e-01 3.73966098e-01 -6.54646456e-01 1.06332786e-01
3.47383380e-01 -8.34354162e-01 7.82295287e-01 9.60161626e-01
5.18483222e-01 1.23289771e-01 -8.46555606e-02 8.98875356e-01
1.58631146e-01 -8.93203467e-02 -6.34739339e-01 -2.23532185e-01
7.84599409e-02 1.32361698e+00 -1.41808271e+00 2.26798579e-02
-3.38394672e-01 9.88089263e-01 3.93018275e-02 6.34635568e-01
-4.65963989e-01 -2.50121653e-01 8.08121562e-01 2.14879550e-02
1.41456246e-01 -4.65355426e-01 1.04254685e-01 -1.30968463e+00
-1.51266292e-01 -3.04945618e-01 5.24645522e-02 -5.42971432e-01
-9.29076552e-01 7.81921148e-02 -4.17457446e-02 -1.39902306e+00
-5.86041510e-02 -6.62860930e-01 -3.08061987e-01 3.89647007e-01
-1.11238813e+00 -5.87229908e-01 -7.97461718e-02 6.32045448e-01
5.77215135e-01 1.83258459e-01 4.61840153e-01 4.10733551e-01
-7.07399487e-01 -1.21095434e-01 4.64881033e-01 -9.48897004e-02
5.58314741e-01 -1.30406690e+00 -2.24453002e-01 1.03696573e+00
-9.09525156e-03 6.54260933e-01 8.63597393e-01 -6.06353164e-01
-1.22741783e+00 -9.27719116e-01 8.11367780e-02 -2.28103176e-01
9.43272114e-01 -1.97009608e-01 -1.03288329e+00 6.72047138e-01
-6.62315711e-02 2.53665537e-01 6.59210801e-01 -9.88190413e-01
3.53019357e-01 2.11170807e-01 -1.21807253e+00 6.41906798e-01
6.84681416e-01 -2.24822536e-01 -5.67946285e-02 2.26563275e-01
1.22361325e-01 -2.06597582e-01 -8.47890437e-01 6.04336746e-02
8.02562773e-01 -7.00025737e-01 6.84605896e-01 -5.49113095e-01
2.61063039e-01 -7.59544730e-01 -1.51431754e-01 -1.09478962e+00
6.81099342e-03 -8.10746074e-01 -5.23736067e-02 1.08145905e+00
1.71220630e-01 -1.69245809e-01 9.74026978e-01 5.56837559e-01
5.57491630e-02 -2.23277077e-01 -9.67961729e-01 -4.94337469e-01
-1.56557746e-02 -1.38035432e-01 -3.71842623e-01 8.54013264e-01
5.64842671e-02 8.06516185e-02 -2.53991336e-01 2.16731489e-01
1.10483301e+00 -1.04209252e-01 4.15845543e-01 -1.30436587e+00
-4.24955934e-02 -2.41347477e-01 -9.16522205e-01 -1.21936953e+00
3.02997828e-01 -6.40098989e-01 3.94407094e-01 -9.78028357e-01
2.86598235e-01 -3.31328362e-02 -9.22791660e-02 -2.06380844e-01
-1.65830292e-02 1.18204631e-01 -1.17143683e-01 4.26630199e-01
-6.04316413e-01 2.64731318e-01 1.34573293e+00 -1.47150382e-01
-3.72380108e-01 4.40440215e-02 4.94379289e-02 8.92588437e-01
5.19484699e-01 -5.60106277e-01 -5.42937100e-01 -4.33710277e-01
6.84263632e-02 3.16996425e-01 2.85947978e-01 -1.02591491e+00
4.82675701e-01 1.51694221e-02 3.38860422e-01 -4.07557577e-01
4.16962177e-01 -9.06041145e-01 2.73823649e-01 5.17003119e-01
-2.99474329e-01 2.36351397e-02 2.76986547e-02 1.09788847e+00
-2.64699966e-01 -3.86466652e-01 1.05183947e+00 -2.91111469e-01
-5.04436851e-01 2.85868853e-01 -9.89163697e-01 2.25037318e-02
1.24923790e+00 -5.01239181e-01 -1.80601269e-01 -8.19916129e-02
-1.06500793e+00 1.99890405e-01 5.79717219e-01 -1.69009313e-01
7.16022253e-01 -9.98219848e-01 -1.82972208e-01 -1.03121586e-01
-3.32026541e-01 2.19788969e-01 4.44033235e-01 1.28616083e+00
-7.61159956e-01 3.53456065e-02 -5.59217691e-01 -1.05508506e+00
-1.20600796e+00 4.36213434e-01 2.92928576e-01 2.65086353e-01
-3.93177181e-01 8.07080150e-01 4.19569671e-01 1.31354600e-01
5.57963662e-02 -6.16331339e-01 -4.09591287e-01 3.03512961e-01
5.46945095e-01 6.63968503e-01 -2.90680587e-01 -8.55611324e-01
-2.71191239e-01 9.19956863e-01 1.07599879e-02 -2.62588352e-01
1.12861025e+00 -6.06831312e-01 -1.72178552e-01 8.90109658e-01
1.15140617e+00 -4.80328023e-01 -1.78205383e+00 4.20621127e-01
1.84720308e-01 -2.42419869e-01 2.86245509e-03 1.55118048e-01
-1.06593311e+00 1.04184055e+00 7.52116501e-01 2.94614255e-01
8.81760895e-01 -4.86301035e-01 6.83258772e-01 4.71225619e-01
4.01144862e-01 -1.00098586e+00 4.73458283e-02 8.17063078e-02
1.14661641e-01 -1.09687316e+00 -2.30180919e-01 -1.05118930e-01
-2.85613805e-01 1.23433077e+00 2.31505886e-01 -3.88660014e-01
6.09226584e-01 5.30937910e-01 1.32491231e-01 1.50009552e-02
-6.05736017e-01 4.17953208e-02 -3.70358765e-01 7.35100925e-01
2.32396245e-01 -1.74143776e-01 -3.48103136e-01 7.11320788e-02
3.30851346e-01 5.29149882e-02 9.16098237e-01 1.02831054e+00
-4.57722127e-01 -6.64866745e-01 -2.94113487e-01 3.13445032e-01
-9.24351990e-01 3.18486542e-01 -1.14676774e-01 5.02835989e-01
-1.31357402e-01 7.33639359e-01 4.75129038e-02 5.28307483e-02
-3.55870500e-02 3.91264213e-03 6.38823867e-01 -3.74467164e-01
-1.39987677e-01 6.77922249e-01 -6.64477468e-01 -6.27713084e-01
-1.18350971e+00 -1.02996290e+00 -1.66729105e+00 4.23275791e-02
-2.41963908e-01 1.09483548e-01 8.42364490e-01 8.23798120e-01
-4.49676774e-02 2.82883763e-01 2.78805315e-01 -1.38758361e+00
-2.60012820e-02 -5.62962949e-01 -1.14173532e+00 4.04640615e-01
6.65158987e-01 -7.30716765e-01 -8.43157113e-01 8.21650624e-01] | [8.7245512008667, -1.1259485483169556] |
c231061b-1a29-4d50-8e20-c454cca7c86c | attention-based-natural-language-person | 1705.08923 | null | http://arxiv.org/abs/1705.08923v1 | http://arxiv.org/pdf/1705.08923v1.pdf | Attention-based Natural Language Person Retrieval | Following the recent progress in image classification and captioning using
deep learning, we develop a novel natural language person retrieval system
based on an attention mechanism. More specifically, given the description of a
person, the goal is to localize the person in an image. To this end, we first
construct a benchmark dataset for natural language person retrieval. To do so,
we generate bounding boxes for persons in a public image dataset from the
segmentation masks, which are then annotated with descriptions and attributes
using the Amazon Mechanical Turk. We then adopt a region proposal network in
Faster R-CNN as a candidate region generator. The cropped images based on the
region proposals as well as the whole images with attention weights are fed
into Convolutional Neural Networks for visual feature extraction, while the
natural language expression and attributes are input to Bidirectional Long
Short- Term Memory (BLSTM) models for text feature extraction. The visual and
text features are integrated to score region proposals, and the one with the
highest score is retrieved as the output of our system. The experimental
results show significant improvement over the state-of-the-art method for
generic object retrieval and this line of research promises to benefit search
in surveillance video footage. | ['Jie Yu', 'Demetri Terzopoulos', 'Tao Zhou', 'Muhao Chen'] | 2017-05-24 | null | null | null | null | ['person-retrieval'] | ['computer-vision'] | [ 7.65164942e-02 -3.60952020e-01 -2.02551320e-01 -5.31105578e-01
-8.99903774e-01 -5.90889931e-01 6.63947940e-01 -5.48895560e-02
-8.54700744e-01 5.23177981e-01 2.92519927e-01 3.28363389e-01
2.78805822e-01 -8.35964084e-01 -6.77753508e-01 -5.52937090e-01
1.87678784e-01 6.59115970e-01 1.54606164e-01 -5.88531829e-02
1.41558513e-01 4.38310325e-01 -1.25087261e+00 5.56602418e-01
4.14789617e-01 1.41308129e+00 1.00463554e-01 6.06306672e-01
-5.86715527e-02 5.75449646e-01 -7.90359080e-01 -5.66270411e-01
2.84130275e-01 -2.91048914e-01 -8.78856957e-01 5.48535921e-02
5.19636810e-01 -7.89830744e-01 -7.85869539e-01 8.76755595e-01
6.33454502e-01 2.95660943e-01 7.78515816e-01 -1.09702229e+00
-1.11635840e+00 1.51618600e-01 -5.79663038e-01 2.23280579e-01
7.32793868e-01 3.43880206e-01 9.47618544e-01 -1.10988855e+00
8.32436919e-01 1.44809973e+00 6.80143908e-02 8.95748138e-01
-7.09635854e-01 -6.25795901e-01 3.49845022e-01 1.86450467e-01
-1.79777098e+00 -2.13557526e-01 5.42465389e-01 -4.71037596e-01
8.66385162e-01 5.28282374e-02 8.41376185e-01 1.26660705e+00
1.36901274e-01 1.18091130e+00 3.78557265e-01 -1.28026545e-01
-2.00133413e-01 3.77677441e-01 2.56296158e-01 9.39790249e-01
6.83882833e-02 -2.96772514e-02 -3.29231560e-01 -1.79381266e-01
6.64194345e-01 4.54772443e-01 -1.42249331e-01 -2.56502390e-01
-1.35480881e+00 8.89394462e-01 1.08158672e+00 -2.32790057e-02
-5.26180208e-01 1.31314933e-01 3.60794842e-01 2.29938142e-02
4.85728651e-01 2.21714199e-01 -5.41025177e-02 6.05092466e-01
-1.15779960e+00 7.40623295e-01 3.38055879e-01 1.00586903e+00
6.18806481e-01 -5.65989375e-01 -1.10624492e+00 8.27189207e-01
5.18238306e-01 8.21585834e-01 4.15341258e-01 -4.63135600e-01
6.56621575e-01 9.32598591e-01 3.24378192e-01 -9.25127804e-01
-6.40014857e-02 -2.59108603e-01 -7.96320200e-01 -3.50342661e-01
2.14048177e-01 4.72459458e-02 -1.09265971e+00 1.54384255e+00
1.17052294e-01 -3.39777857e-01 1.18913837e-01 1.48098600e+00
1.12767506e+00 9.54546869e-01 2.20226943e-01 4.07921553e-01
1.80761170e+00 -1.32811630e+00 -4.62588519e-01 -4.06771183e-01
6.88037723e-02 -4.45759356e-01 7.43584752e-01 -7.21039101e-02
-9.76922333e-01 -7.59025037e-01 -7.20168889e-01 -3.24037850e-01
-7.21363366e-01 5.32229841e-01 2.21657008e-01 1.08616568e-01
-1.07726836e+00 -8.13333839e-02 -2.71244317e-01 -6.26974344e-01
7.70812690e-01 2.48771966e-01 -3.68665278e-01 -2.12278947e-01
-1.36589432e+00 6.61567330e-01 4.58942920e-01 2.54541487e-01
-1.20632207e+00 -7.97966495e-02 -1.04005134e+00 1.60225675e-01
1.81736723e-01 -1.04632080e+00 1.01476741e+00 -1.21043646e+00
-8.22681963e-01 1.35069680e+00 -1.92769170e-01 -5.96448302e-01
5.74134111e-01 -2.87857175e-01 -1.22892052e-01 5.96463025e-01
4.91511047e-01 1.53670716e+00 9.80985582e-01 -1.03498268e+00
-5.86623013e-01 -3.88349771e-01 4.84565087e-02 1.58735678e-01
-3.69842738e-01 6.11704946e-01 -1.14125586e+00 -6.11102223e-01
-3.45512778e-01 -8.08264613e-01 -1.71996549e-01 3.63586605e-01
-5.59608519e-01 -7.44159162e-01 6.15110874e-01 -9.61642921e-01
8.86571884e-01 -1.99150646e+00 1.89749852e-01 3.28940332e-01
6.13601990e-02 2.46257380e-01 -3.85436743e-01 3.40417176e-02
1.77676469e-01 4.47549112e-02 3.19946706e-02 -4.05455917e-01
1.45819096e-03 -3.43048692e-01 -5.70357382e-01 4.49231863e-01
6.23891830e-01 1.49446058e+00 -8.61016333e-01 -1.01497793e+00
1.83117464e-01 4.46928173e-01 -2.61237830e-01 5.87665141e-01
-4.21100438e-01 2.68317759e-01 -8.36803854e-01 8.37782323e-01
2.62431026e-01 -2.63255864e-01 -6.39836431e-01 2.69985981e-02
1.76544502e-01 -1.24253854e-01 -6.20950997e-01 1.65787721e+00
1.32104173e-01 6.45068824e-01 -1.54966652e-01 -8.02117944e-01
9.77668822e-01 2.97146499e-01 3.18309247e-01 -7.35603750e-01
1.40552565e-01 -5.66720478e-02 -5.56028485e-01 -6.82606578e-01
6.76408470e-01 4.28138435e-01 -4.92803395e-01 3.92856479e-01
-7.22880755e-03 2.90332556e-01 2.22758546e-01 4.51329321e-01
7.00637758e-01 1.97267234e-01 2.46354216e-03 1.02947503e-01
7.62755334e-01 1.58593372e-01 2.22888932e-01 8.65230143e-01
-4.00922805e-01 8.19299459e-01 2.07042292e-01 -8.91430497e-01
-1.34981108e+00 -1.00360453e+00 1.10410146e-01 1.21066582e+00
1.44671366e-01 -3.37801985e-02 -9.94961500e-01 -7.44493127e-01
-5.59500679e-02 1.15948804e-01 -7.61736691e-01 -1.61409184e-01
-5.78089118e-01 -1.93680122e-01 5.99890411e-01 6.13748193e-01
9.15636063e-01 -1.79206717e+00 -6.62068784e-01 -1.34628490e-01
-3.47628027e-01 -1.13603914e+00 -9.84011471e-01 -5.73871672e-01
-1.76718280e-01 -9.16988552e-01 -1.57088494e+00 -1.30423892e+00
9.21625733e-01 1.37823582e-01 1.04658258e+00 1.20848902e-01
-6.30157053e-01 5.46078026e-01 -2.39965141e-01 -2.50405759e-01
1.61468029e-01 1.56270176e-01 -1.79278702e-01 2.44666323e-01
7.51893342e-01 6.10247970e-01 -1.04575872e+00 1.17189221e-01
-9.39580321e-01 -7.06252828e-02 6.58214629e-01 8.09630632e-01
7.28444397e-01 -3.85718703e-01 4.84456152e-01 -2.48336196e-01
6.93315327e-01 -4.68356073e-01 -6.60047412e-01 5.36937058e-01
8.20648372e-02 -7.56194815e-02 4.09479052e-01 -4.13096100e-01
-8.37163091e-01 2.69354701e-01 1.94465339e-01 -5.74887514e-01
-2.39484847e-01 1.16794720e-01 -8.97657946e-02 3.69186461e-01
4.26457733e-01 5.80089867e-01 -2.23893076e-01 -5.35130613e-02
5.14833450e-01 1.01133788e+00 5.91784239e-01 -3.93042862e-01
7.05908358e-01 4.94310766e-01 -6.42126501e-01 -5.97626865e-01
-1.03767753e+00 -7.26375461e-01 -6.46630347e-01 -3.35371405e-01
1.50998318e+00 -1.10296965e+00 -6.72520459e-01 2.48846337e-01
-1.57970166e+00 -1.83150433e-02 5.12434840e-02 2.11043388e-01
-3.40750337e-01 9.56522226e-02 -5.66820979e-01 -9.16787624e-01
-9.89168167e-01 -1.10848200e+00 1.74525046e+00 5.34408033e-01
-1.37294576e-01 -5.02295196e-01 -1.85807109e-01 6.15758896e-01
3.15217793e-01 4.81613874e-02 5.95204890e-01 -8.41063678e-01
-9.00576115e-01 -7.22603559e-01 -7.09232032e-01 1.99958146e-01
-4.13478702e-01 -2.25888237e-01 -8.64096463e-01 -5.18935025e-01
-4.95578259e-01 -5.97627044e-01 1.30785513e+00 4.11045283e-01
1.41609895e+00 -3.26488912e-01 -7.78858781e-01 4.72930759e-01
1.14098537e+00 -1.08227106e-02 5.13319016e-01 2.76450247e-01
5.85939705e-01 8.10510993e-01 7.70705402e-01 2.30318859e-01
3.11239511e-01 7.13730931e-01 1.40120849e-01 -3.90750200e-01
-3.56003009e-02 -3.83885235e-01 3.73940110e-01 -1.30713671e-01
1.85260370e-01 -4.93497461e-01 -8.46691191e-01 7.96672463e-01
-1.91856980e+00 -1.10186279e+00 4.92044836e-01 2.04736280e+00
6.35804772e-01 -1.48784593e-01 2.53364712e-01 -6.40297234e-01
9.12018359e-01 2.58748174e-01 -5.04981697e-01 2.35898644e-02
-1.01987712e-01 -9.01872292e-02 2.76131421e-01 9.99330506e-02
-1.54986560e+00 1.32699108e+00 5.38999367e+00 6.41571105e-01
-8.62627089e-01 -6.51009157e-02 9.08156633e-01 -2.52488256e-01
1.04289882e-01 -5.28983712e-01 -1.09816825e+00 3.25672507e-01
5.88989675e-01 7.01743588e-02 3.28523189e-01 8.24399710e-01
1.79860324e-01 1.70555770e-01 -1.15810323e+00 1.07050669e+00
6.57455385e-01 -9.76476669e-01 5.39026916e-01 -1.07200190e-01
7.81998098e-01 4.42906059e-02 2.01719299e-01 3.92538697e-01
-6.62292494e-03 -1.24058294e+00 8.00511301e-01 1.02617133e+00
7.36879349e-01 -7.20879972e-01 1.00903428e+00 2.48839408e-01
-1.28971028e+00 -2.66794711e-01 -7.28561997e-01 4.78840470e-01
9.38141122e-02 9.55005456e-03 -7.63842821e-01 1.26087353e-01
8.08054626e-01 5.91818810e-01 -8.02441657e-01 1.17355645e+00
-3.02412778e-01 1.69054270e-01 -9.20616612e-02 -4.01544064e-01
4.59472060e-01 -7.27929771e-02 4.27774012e-01 1.54059923e+00
7.02259541e-02 -2.79685669e-02 4.86611485e-01 1.35982060e+00
-5.37622213e-01 2.53328621e-01 -8.00980210e-01 -2.89318319e-02
2.37955168e-01 1.40838540e+00 -6.68145359e-01 -8.32917988e-01
-3.75617802e-01 1.33533990e+00 3.53184938e-01 7.37204731e-01
-9.20730412e-01 -4.38820183e-01 2.77872741e-01 3.07407584e-02
3.09569955e-01 2.12441117e-01 3.03487480e-01 -1.06806123e+00
2.72966921e-01 -7.00966597e-01 5.31769812e-01 -1.10116470e+00
-1.51181960e+00 8.42288733e-01 2.91574951e-02 -9.73372102e-01
-2.73249269e-01 -5.35982251e-01 -4.31950331e-01 1.07903850e+00
-1.39459741e+00 -1.44339454e+00 -5.03344357e-01 7.05980182e-01
8.40561628e-01 -5.22160292e-01 7.01967239e-01 2.11363226e-01
-5.03967285e-01 5.31753957e-01 -4.09145921e-01 1.01161265e+00
7.46668041e-01 -9.12708640e-01 6.28284276e-01 6.87070072e-01
2.00647622e-01 8.35872769e-01 7.85311125e-03 -9.00222480e-01
-1.01161814e+00 -1.60018170e+00 1.01179004e+00 -6.29018307e-01
2.74118900e-01 -6.37092888e-01 -6.63371503e-01 5.93738735e-01
3.25611413e-01 2.63025075e-01 3.22497547e-01 -5.97850978e-01
-3.15111428e-01 5.35345525e-02 -1.12636399e+00 6.25781953e-01
7.68631339e-01 -6.96255326e-01 -7.43936062e-01 6.77899659e-01
8.06787968e-01 -6.63783327e-02 -3.87667567e-01 5.21783419e-02
6.37218118e-01 -2.78942078e-01 1.28522992e+00 -9.08314466e-01
4.28686053e-01 -2.59794980e-01 3.63557078e-02 -6.93873703e-01
-2.05780193e-01 -1.44793853e-01 3.81308466e-01 1.18661559e+00
4.44731593e-01 -2.07976043e-01 7.17646301e-01 5.89270949e-01
2.60954142e-01 -6.95347488e-01 -6.60776556e-01 -3.49093586e-01
-5.35655655e-02 7.10917264e-02 4.50429022e-01 2.61659026e-01
-3.55333805e-01 3.81443173e-01 -3.56031597e-01 3.46771702e-02
6.72768831e-01 2.46933967e-01 6.71753466e-01 -9.82203245e-01
1.36996076e-01 -3.58684957e-01 -2.89050639e-01 -1.29177439e+00
4.45209771e-01 -1.02943468e+00 4.02986676e-01 -1.81254315e+00
9.25050676e-01 1.54116219e-02 -4.18500394e-01 4.46851671e-01
-3.96973699e-01 5.76938272e-01 2.14233950e-01 2.96431839e-01
-1.35553873e+00 5.73605359e-01 1.09138775e+00 -6.56813920e-01
-1.42991826e-01 6.50148764e-02 -3.87376845e-01 5.43569446e-01
5.68619728e-01 -2.72957116e-01 5.86156584e-02 -4.73412424e-01
-5.83276860e-02 6.29246235e-02 9.71280396e-01 -7.69465804e-01
3.73853743e-01 1.75063282e-01 1.16407204e+00 -9.83393967e-01
4.83404130e-01 -7.22798884e-01 -5.46480060e-01 4.21637982e-01
-8.59368742e-01 2.36035779e-01 -9.37855840e-02 5.28836787e-01
-1.98424309e-01 -1.47123754e-01 5.20497262e-01 -3.89255911e-01
-6.76300406e-01 7.26761699e-01 -2.27700770e-01 -9.09396410e-02
1.00714004e+00 4.29922864e-02 -3.50401223e-01 -4.31042016e-01
-5.89975119e-01 6.42055511e-01 2.97656298e-01 7.97389448e-01
1.04126239e+00 -1.40291679e+00 -9.29177940e-01 3.43762577e-01
4.80437011e-01 -1.10260285e-02 -1.23061668e-02 2.13187799e-01
-4.31863964e-01 9.32742834e-01 -3.16226971e-03 -6.79778695e-01
-1.26875162e+00 8.17512393e-01 4.33971763e-01 -1.62744015e-01
-4.04648393e-01 9.91771221e-01 4.28020149e-01 -1.48744628e-01
5.17888606e-01 -1.65273428e-01 -4.46615189e-01 -3.67816025e-03
7.65424192e-01 -1.96502462e-01 -4.68711913e-01 -9.69050348e-01
-5.10008991e-01 7.11264133e-01 -2.24921808e-01 -2.20510125e-01
1.08230817e+00 -7.22051365e-03 -1.79254159e-01 -5.21006994e-02
1.31765676e+00 -4.86324161e-01 -1.19527996e+00 -4.28018034e-01
-1.07182175e-01 -2.14295015e-01 -2.86898106e-01 -7.15317011e-01
-1.08937013e+00 9.79422331e-01 7.87157655e-01 -1.00019976e-01
7.97228158e-01 5.36995351e-01 9.96361494e-01 6.20091736e-01
1.76655442e-01 -1.05155432e+00 3.86636317e-01 2.78722912e-01
1.05413032e+00 -1.37818170e+00 -8.00404251e-02 1.31090403e-01
-7.06346095e-01 8.85383010e-01 7.04761088e-01 -4.32172328e-01
1.76166162e-01 -3.38416159e-01 -1.51186790e-02 -3.13186198e-01
-5.83054781e-01 -3.99045616e-01 7.06574261e-01 5.67335546e-01
2.17242151e-01 -4.69633304e-02 1.41827121e-01 6.24972701e-01
-5.33507653e-02 -1.40761718e-01 -2.82955229e-01 3.64136010e-01
-6.11894011e-01 -6.63127959e-01 -5.75545371e-01 4.04992849e-01
-5.07045925e-01 -1.22093283e-01 -8.33332121e-01 4.08171088e-01
1.09473370e-01 8.33119810e-01 2.45097518e-01 -1.20123464e-03
2.55945146e-01 1.48418918e-01 1.34908885e-01 -6.05199695e-01
-7.42145419e-01 -1.99136022e-03 -2.75754482e-01 -4.54906970e-01
-1.45818919e-01 -3.97397727e-01 -1.18984568e+00 3.70613843e-01
-1.66301146e-01 1.02399491e-01 4.55962777e-01 9.16797996e-01
1.28957152e-01 4.27268833e-01 3.82190466e-01 -7.55077004e-01
-3.75106871e-01 -1.12797201e+00 7.78554827e-02 6.88689828e-01
3.57268393e-01 -3.83562148e-01 1.15478158e-01 1.88229680e-01] | [14.643486976623535, 0.8472605347633362] |
d1bdc0cf-a824-4b3d-91e4-c1c050473f08 | one-shot-domain-adaptation-in-video-based | 2301.00812 | null | https://arxiv.org/abs/2301.00812v3 | https://arxiv.org/pdf/2301.00812v3.pdf | One-shot domain adaptation in video-based assessment of surgical skills | Deep Learning (DL) has achieved automatic and objective assessment of surgical skills. However, DL models are data-hungry and restricted to their training domain. This prevents them from transitioning to new tasks where data is limited. Hence, domain adaptation is crucial to implement DL in real life. Here, we propose a meta-learning model, A-VBANet, that can deliver domain-agnostic surgical skill classification via one-shot learning. We develop the A-VBANet on five laparoscopic and robotic surgical simulators. Additionally, we test it on operating room (OR) videos of laparoscopic cholecystectomy. Our model successfully adapts with accuracies up to 99.5% in one-shot and 99.9% in few-shot settings for simulated tasks and 89.7% for laparoscopic cholecystectomy. For the first time, we provide a domain-agnostic procedure for video-based assessment of surgical skills. A significant implication of this approach is that it allows the use of data from surgical simulators to assess performance in the operating room. | ['Suvranu De', 'Xavier Intes', 'Gene Yang', 'Steven Schwaitzberg', 'Erim Yanik'] | 2022-12-16 | null | null | null | null | ['one-shot-learning'] | ['methodology'] | [-4.13172394e-02 1.48089081e-01 -3.63186240e-01 -3.89496386e-01
-7.87664711e-01 -6.44069016e-01 1.52433336e-01 2.80428410e-01
-8.47440541e-01 4.14237857e-01 5.04165739e-02 -4.71628606e-01
-4.51356560e-01 -4.90858614e-01 -6.80844128e-01 -2.50581235e-01
-2.39969790e-01 5.55589736e-01 8.50846469e-02 -3.33676040e-01
2.31933203e-02 5.22183061e-01 -1.36735654e+00 5.81644177e-01
9.34816062e-01 7.43957758e-01 4.53732252e-01 9.25550342e-01
1.41989633e-01 8.73475254e-01 -4.03352320e-01 -2.11869583e-01
4.59028125e-01 -2.51024693e-01 -7.83216059e-01 -2.08714202e-01
3.68192911e-01 -5.30778229e-01 -5.58436096e-01 8.75948370e-01
6.83532119e-01 2.25670218e-01 4.38666880e-01 -9.25183356e-01
-1.96297660e-01 4.57759231e-01 4.73830178e-02 3.39486688e-01
2.79778600e-01 4.07548755e-01 2.99509376e-01 -5.77827394e-01
8.61472487e-01 6.94676936e-01 8.12353194e-01 8.40402424e-01
-9.46403801e-01 -6.21957481e-01 -1.53616562e-01 -3.57169434e-02
-9.37133908e-01 -2.42235079e-01 3.96623939e-01 -6.51483834e-01
7.72736549e-01 -1.61488622e-01 1.04600894e+00 1.42119443e+00
6.76712036e-01 8.82505596e-01 1.15823901e+00 -1.40061870e-01
2.47383550e-01 2.21833542e-01 -2.49430776e-01 8.55107844e-01
4.61642534e-01 4.53523010e-01 -4.60426331e-01 2.21523181e-01
9.43629086e-01 3.49754989e-01 -1.48886174e-01 -8.04183662e-01
-1.40937316e+00 6.96677685e-01 4.49200898e-01 2.61516988e-01
-4.14530039e-01 7.55442679e-02 8.07021260e-01 5.63016236e-01
-2.19885372e-02 1.11550713e+00 -5.11411011e-01 -6.26126111e-01
-8.25583458e-01 -1.10918358e-02 1.10769725e+00 1.15931571e+00
2.80127913e-01 -8.87131691e-03 -2.96647668e-01 7.33339369e-01
-3.06857765e-01 1.06155843e-01 9.57885802e-01 -1.11458039e+00
1.37739286e-01 5.08756101e-01 8.74143243e-02 -5.79692185e-01
-8.68402719e-01 -6.13776803e-01 -5.64656377e-01 3.97978634e-01
1.10825330e-01 -3.16311121e-01 -1.05966258e+00 1.49310076e+00
5.98431528e-02 3.76505524e-01 2.00915083e-01 9.11237061e-01
1.25665438e+00 6.43103123e-02 2.58519113e-01 -6.84410930e-02
1.15084291e+00 -1.14459658e+00 -5.05701602e-01 -6.12258911e-01
9.72044826e-01 -5.52659273e-01 1.17711234e+00 7.28635490e-01
-1.15618908e+00 -7.40645349e-01 -1.05789900e+00 1.73849463e-01
-5.81353307e-02 1.48543075e-01 8.57153893e-01 4.90354300e-01
-8.31959903e-01 7.20413148e-01 -1.24480343e+00 -3.68523866e-01
3.13712150e-01 4.92410839e-01 -6.02035582e-01 -1.30931467e-01
-9.16377187e-01 1.04691505e+00 4.44401860e-01 -2.97932267e-01
-1.38761151e+00 -1.01054609e+00 -1.08729112e+00 1.27750844e-01
5.10827184e-01 -8.02978396e-01 1.88233805e+00 -1.04765058e+00
-1.77791822e+00 1.16136920e+00 5.73940337e-01 -5.46627462e-01
5.66631734e-01 -3.89367193e-01 -4.87222552e-01 2.32407585e-01
-9.84381586e-02 4.28193867e-01 5.11016786e-01 -1.03524756e+00
-6.09980583e-01 -8.50049257e-02 3.86295736e-01 3.76119494e-01
-3.27085137e-01 -5.68711162e-01 -5.62173724e-01 -4.71090078e-01
-1.81580216e-01 -1.24652469e+00 -5.75418711e-01 8.68876055e-02
2.93707907e-01 3.48377436e-01 1.72675833e-01 -4.24591243e-01
1.18141806e+00 -2.36339569e+00 1.39199331e-01 -1.90170541e-01
2.50404835e-01 6.25913501e-01 -3.00270140e-01 4.74505961e-01
4.32220660e-02 -3.29268187e-01 2.89971288e-02 -8.25857371e-02
-3.27646941e-01 2.75181979e-01 2.21414313e-01 4.25922096e-01
-5.38672358e-02 8.98003042e-01 -1.47067893e+00 -6.32924795e-01
4.19191360e-01 1.77255906e-02 -1.07296836e+00 5.81344664e-01
1.09998994e-01 7.29751885e-01 -3.73893201e-01 6.13371074e-01
3.28160226e-01 -2.77800977e-01 3.70647937e-01 -1.23418391e-01
1.19474053e-01 5.74806854e-02 -7.59721458e-01 2.62266278e+00
-9.66685593e-01 5.95380366e-01 1.80286512e-01 -8.86002362e-01
7.17613578e-01 3.20739537e-01 7.99295366e-01 -7.18136132e-01
3.52410436e-01 2.57466733e-01 2.74228752e-01 -9.95577514e-01
3.86940151e-01 -2.92061210e-01 -4.88634646e-01 1.27445713e-01
7.47448206e-01 -5.19730270e-01 1.91242918e-01 1.02806628e-01
1.42529356e+00 2.14961559e-01 7.28122354e-01 -3.18744957e-01
2.75177985e-01 4.85455126e-01 5.06109834e-01 1.06330013e+00
-7.38307595e-01 5.65855026e-01 1.18815809e-01 -7.31780887e-01
-7.17894912e-01 -1.29701662e+00 3.51390429e-02 1.22953308e+00
3.30671340e-01 -2.89357573e-01 -4.75892842e-01 -8.39908123e-01
1.84674457e-01 6.04572356e-01 -7.98239887e-01 -8.57175767e-01
-5.08237004e-01 -1.32960841e-01 1.67198941e-01 6.18234813e-01
1.34246841e-01 -1.11894214e+00 -9.55928028e-01 4.93719727e-01
2.94511095e-02 -1.13845193e+00 -2.80352801e-01 3.16994548e-01
-1.11786258e+00 -1.22726309e+00 -6.15697324e-01 -1.04807949e+00
5.40887952e-01 1.45110235e-01 1.31252491e+00 -2.27341995e-01
-6.08810723e-01 8.61230612e-01 -2.51373678e-01 -5.94383359e-01
-7.28363395e-01 3.03010553e-01 2.09543049e-01 -5.34659207e-01
2.63763547e-01 -3.02420139e-01 -1.17079222e+00 2.49511257e-01
-5.92316151e-01 7.79132247e-02 1.26129174e+00 1.04064822e+00
4.30607766e-01 -6.14185452e-01 4.19249564e-01 -1.29776168e+00
6.67127371e-01 -6.34899199e-01 -4.68119532e-01 9.14827809e-02
-6.93540990e-01 -8.09363574e-02 5.91460764e-01 -8.16035867e-01
-1.11783457e+00 7.13130683e-02 -4.61813435e-03 -8.08477700e-01
-1.35493264e-01 7.85842896e-01 5.70983291e-01 -1.67771608e-01
1.16207898e+00 -1.20069766e-02 2.39155233e-01 -2.42651910e-01
-3.68041173e-02 4.98556703e-01 9.03110087e-01 -5.21092772e-01
3.73988360e-01 2.48838276e-01 -9.85721126e-02 -2.72177964e-01
-9.55177426e-01 -7.47141540e-01 -5.42749703e-01 -3.05593610e-01
5.24405360e-01 -1.20293093e+00 -8.22368622e-01 6.52979836e-02
-3.77853245e-01 -8.54944766e-01 -4.13378745e-01 1.00498819e+00
-1.00992370e+00 6.62767291e-02 -7.24403739e-01 -2.86990702e-01
-3.46564412e-01 -1.25394773e+00 7.82116294e-01 2.49986008e-01
-2.95581400e-01 -1.08600271e+00 3.98982495e-01 1.82313159e-01
5.03172398e-01 3.00175607e-01 3.70540828e-01 -7.54417837e-01
-2.88235068e-01 -5.35132289e-01 1.60497040e-01 3.20774883e-01
1.50637165e-01 -4.25027221e-01 -7.57856429e-01 -8.12465250e-01
-6.66536242e-02 -5.34216225e-01 4.89642262e-01 4.67545122e-01
1.55887699e+00 8.59287009e-02 -4.39035475e-01 1.07337320e+00
1.36724162e+00 7.06481934e-02 2.56764263e-01 6.28965080e-01
1.83888167e-01 2.13271812e-01 1.12087238e+00 4.81958896e-01
2.01600850e-01 4.25070852e-01 5.20377457e-01 -2.09656641e-01
-2.15726525e-01 -2.45684177e-01 1.21694036e-01 9.32668507e-01
-5.87575845e-02 4.69410196e-02 -1.23861742e+00 8.12662184e-01
-1.62618554e+00 -7.60813177e-01 4.44389194e-01 2.28038001e+00
7.87373960e-01 1.94150940e-01 -1.74897820e-01 -5.49740136e-01
1.53262511e-01 -3.69323522e-01 -7.32353926e-01 -6.15331113e-01
6.77347422e-01 4.33205068e-01 5.38980961e-01 -1.64228398e-03
-1.14936268e+00 8.06510508e-01 6.47302055e+00 4.78842616e-01
-1.26078248e+00 1.17119268e-01 1.33844927e-01 -4.16952968e-01
3.52499187e-01 -4.34317619e-01 -2.90387273e-01 3.16522360e-01
1.07762349e+00 -3.96131545e-01 2.05404356e-01 1.37321174e+00
-1.01050518e-01 -1.26535967e-01 -1.35850716e+00 1.05945337e+00
-1.99271850e-02 -1.23719311e+00 -2.67857403e-01 -3.60150784e-01
8.85132432e-01 1.72751367e-01 1.08184308e-01 1.03425598e+00
4.17047471e-01 -1.05586576e+00 2.39269406e-01 4.46831614e-01
1.22862446e+00 -6.31147206e-01 1.00627041e+00 5.23293674e-01
-7.27829576e-01 -4.37126964e-01 -1.99962795e-01 9.46308896e-02
-1.58514693e-01 -5.31412773e-02 -1.43361533e+00 4.26262707e-01
7.98590779e-01 6.90906048e-01 -1.85837671e-01 1.31008625e+00
1.87324863e-02 3.19517612e-01 1.32548317e-01 1.41948789e-01
5.10582447e-01 2.32537150e-01 2.87432164e-01 1.52890587e+00
5.38870037e-01 1.46719396e-01 3.81046355e-01 1.54540524e-01
-2.11057022e-01 1.27561092e-02 -7.72762001e-01 -5.59881032e-02
2.97989875e-01 1.40353560e+00 -3.90949279e-01 -1.97542161e-01
-4.27578658e-01 1.10065031e+00 4.50439125e-01 1.86560661e-01
-6.25616848e-01 -3.88908148e-01 7.48546660e-01 2.13923696e-02
1.80130843e-02 -1.68949291e-01 -1.86753511e-01 -1.03631353e+00
-4.89856213e-01 -9.85941768e-01 5.41136563e-01 -6.63119495e-01
-1.04067326e+00 6.12114787e-01 -7.09823295e-02 -1.91564035e+00
-4.40395206e-01 -9.50428307e-01 -3.02945971e-01 4.63642359e-01
-1.55049980e+00 -8.36031914e-01 -1.00388384e+00 5.82183182e-01
7.35604823e-01 -3.85114640e-01 9.87471700e-01 1.40553325e-01
-1.32050976e-01 6.44450784e-01 1.94838256e-01 1.50329664e-01
1.24379098e+00 -1.25314128e+00 3.89752574e-02 3.67712706e-01
-2.99749434e-01 5.73083639e-01 6.88838959e-01 -4.15101826e-01
-1.56122494e+00 -1.07697034e+00 1.94072515e-01 -6.32799983e-01
5.76679766e-01 -1.40959680e-01 -8.13754261e-01 8.40073705e-01
-2.50070635e-02 1.90894634e-01 1.01705039e+00 1.18881278e-01
4.20394316e-02 -1.70586675e-01 -1.34640992e+00 6.92593932e-01
1.28357363e+00 -2.64568180e-01 -8.28473508e-01 4.08897817e-01
5.06223261e-01 -1.29145980e+00 -1.30556834e+00 7.46462762e-01
7.69021690e-01 -9.57439661e-01 1.09336877e+00 -7.78872669e-01
7.27447987e-01 3.74677449e-01 3.89177352e-01 -1.68133748e+00
-3.62010121e-01 -4.30644661e-01 -1.26840264e-01 1.49150386e-01
2.54547328e-01 -3.90877455e-01 7.75565803e-01 2.76515424e-01
-7.26976633e-01 -6.88356519e-01 -6.71309412e-01 -9.06951249e-01
3.50241996e-02 -3.13163549e-01 2.29971185e-01 1.04478312e+00
2.64198005e-01 -2.33785018e-01 -2.02223629e-01 3.11184321e-02
3.13536078e-01 1.19221047e-01 9.40298140e-01 -1.25822818e+00
-4.77753431e-01 -3.99926335e-01 -5.32438874e-01 -7.85682797e-01
8.05880874e-03 -9.70240474e-01 9.84249860e-02 -1.60231411e+00
2.81005353e-01 -5.19720674e-01 -6.86934829e-01 2.84386992e-01
-7.43452385e-02 -5.24326190e-02 7.14382306e-02 9.37250704e-02
-7.19117880e-01 1.61863759e-01 1.57722414e+00 1.07167289e-01
-3.42361659e-01 2.61198014e-01 -4.31254834e-01 7.33239651e-01
9.23048556e-01 -4.82804626e-01 -5.21109939e-01 -4.46591347e-01
4.04388011e-02 3.66965681e-01 3.92161869e-02 -1.37220371e+00
4.76841629e-01 -2.50102103e-01 2.63974547e-01 -4.09847088e-02
3.04861069e-01 -9.20772314e-01 7.24442825e-02 9.71321523e-01
-2.92696565e-01 -8.60470086e-02 7.48500347e-01 5.56989431e-01
-2.34449625e-01 -2.06018016e-01 8.44541848e-01 -5.83503425e-01
-1.28142881e+00 5.06381392e-01 -4.60912734e-01 2.21558928e-01
1.22602999e+00 -3.09009612e-01 -8.15923959e-02 -2.49409586e-01
-9.88464653e-01 3.44683081e-01 4.63560164e-01 5.79994977e-01
7.01695442e-01 -1.05434227e+00 -7.17817605e-01 3.68412495e-01
7.28621304e-01 6.48997873e-02 7.29999959e-01 8.82272840e-01
-9.48911250e-01 4.25453305e-01 -7.35570669e-01 -7.70765841e-01
-1.07441759e+00 7.52531350e-01 5.38178921e-01 -4.83010232e-01
-7.82664180e-01 8.71176660e-01 3.15877080e-01 -7.53576756e-01
3.31392407e-01 -4.26908135e-01 -7.46040270e-02 -3.69215190e-01
2.98896164e-01 -2.80295052e-02 1.85186207e-01 2.64043063e-01
-2.10912168e-01 1.79579288e-01 -2.10495263e-01 2.59206206e-01
1.28414142e+00 1.62141293e-01 6.17309093e-01 5.10052979e-01
8.52737546e-01 -3.29588473e-01 -1.42638111e+00 -3.39727819e-01
-5.14395721e-02 -5.18042088e-01 2.02775404e-01 -1.19086134e+00
-7.57211506e-01 7.61861563e-01 8.10575604e-01 -4.99171287e-01
1.17455649e+00 -1.15753792e-01 5.57863414e-01 6.35196924e-01
7.53665864e-01 -1.17944419e+00 4.19503450e-01 3.34937841e-01
7.67979622e-01 -1.59488261e+00 -1.77592114e-01 1.24449003e-02
-8.64478230e-01 1.07385004e+00 1.03382337e+00 -1.22583248e-01
4.95723307e-01 3.75015050e-01 3.33875328e-01 -2.56428748e-01
-7.10000098e-01 3.67579646e-02 4.39449698e-01 6.31971061e-01
5.53405523e-01 1.96930870e-01 -1.94232434e-01 6.78198993e-01
-1.17131829e-01 6.02888703e-01 7.23914802e-01 1.26888835e+00
-3.12548578e-01 -6.36162817e-01 2.13388637e-01 6.34794533e-01
-4.14217502e-01 9.70699564e-02 1.99602708e-01 1.15347838e+00
-2.72380590e-01 6.09658003e-01 6.52572364e-02 -2.08352953e-01
7.09805310e-01 -3.81815210e-02 7.65456021e-01 -1.09721017e+00
-9.76947367e-01 -3.12011302e-01 3.61676551e-02 -8.36478472e-01
-1.23261683e-01 -3.28800410e-01 -1.01321578e+00 -2.85767436e-01
1.32777482e-01 1.64177954e-01 7.06099093e-01 4.28910524e-01
4.87984776e-01 1.09754550e+00 3.19939733e-01 -7.39252925e-01
-7.41421998e-01 -9.43999648e-01 -5.43319404e-01 6.40322983e-01
4.17904615e-01 -7.73672581e-01 -9.12464187e-02 -1.75533295e-01] | [14.071737289428711, -3.3710811138153076] |
bbf73999-f8e9-4b0b-95cf-00758fd1e2ab | fairness-in-visual-clustering-a-novel | 2304.07408 | null | https://arxiv.org/abs/2304.07408v1 | https://arxiv.org/pdf/2304.07408v1.pdf | Fairness in Visual Clustering: A Novel Transformer Clustering Approach | Promoting fairness for deep clustering models in unsupervised clustering settings to reduce demographic bias is a challenging goal. This is because of the limitation of large-scale balanced data with well-annotated labels for sensitive or protected attributes. In this paper, we first evaluate demographic bias in deep clustering models from the perspective of cluster purity, which is measured by the ratio of positive samples within a cluster to their correlation degree. This measurement is adopted as an indication of demographic bias. Then, a novel loss function is introduced to encourage a purity consistency for all clusters to maintain the fairness aspect of the learned clustering model. Moreover, we present a novel attention mechanism, Cross-attention, to measure correlations between multiple clusters, strengthening faraway positive samples and improving the purity of clusters during the learning process. Experimental results on a large-scale dataset with numerous attribute settings have demonstrated the effectiveness of the proposed approach on both clustering accuracy and fairness enhancement on several sensitive attributes. | ['Khoa Luu', 'Kaushik Roy', 'Marios Savvides', 'Chi Nhan Duong', 'Xuan-Bac Nguyen'] | 2023-04-14 | null | null | null | null | ['deep-clustering', 'deep-clustering'] | ['miscellaneous', 'natural-language-processing'] | [-3.93636316e-01 -9.14302468e-02 -1.54249594e-01 -8.44635844e-01
-4.09242928e-01 -1.96026891e-01 3.06419075e-01 5.58298528e-01
-5.98321199e-01 5.40028632e-01 3.15660477e-01 3.36019993e-01
-2.54675031e-01 -8.82429957e-01 -2.18857512e-01 -1.08815527e+00
1.04333743e-01 4.42191720e-01 -4.34064746e-01 1.81435734e-01
2.79983670e-01 2.83052832e-01 -1.23539102e+00 -7.03838840e-02
1.34201241e+00 9.54592645e-01 -3.59717011e-01 -2.04287544e-01
5.80173545e-02 5.53679943e-01 -6.09683692e-01 -7.28064954e-01
2.99259603e-01 -3.51223469e-01 -5.38154125e-01 5.30766882e-02
3.01314771e-01 -2.30061427e-01 -3.08388900e-02 1.26547873e+00
8.56844485e-01 2.58739978e-01 7.22547889e-01 -1.60339403e+00
-7.44296968e-01 9.39065278e-01 -1.05399084e+00 5.69013739e-03
-4.91463363e-01 1.05530471e-01 1.21500146e+00 -4.65318799e-01
6.91998899e-02 1.45564246e+00 3.63947421e-01 6.12352073e-01
-1.37240160e+00 -1.10547781e+00 1.45871088e-01 1.57823026e-01
-1.58532226e+00 -4.51485246e-01 8.20234358e-01 -4.62858707e-01
5.54886051e-02 7.11894333e-02 2.98531383e-01 6.62765026e-01
-2.61354774e-01 4.25146580e-01 9.88010764e-01 -4.51851152e-02
5.06627858e-01 4.71302062e-01 3.89960051e-01 2.37743661e-01
2.89281577e-01 -5.62169738e-02 -2.86249697e-01 6.93751723e-02
4.15833652e-01 8.12228993e-02 -1.80883072e-02 -5.54414868e-01
-8.82654667e-01 1.05422127e+00 8.64904881e-01 8.80496800e-02
-3.24923843e-01 9.55144390e-02 5.59501946e-01 -1.36137426e-01
6.82029903e-01 3.61910701e-01 -1.07710898e-01 1.60756126e-01
-9.58730519e-01 5.94043694e-02 2.15632409e-01 8.45261455e-01
7.42455304e-01 -7.33006895e-02 -7.18943715e-01 9.63525951e-01
2.33773336e-01 4.06994253e-01 3.05285215e-01 -1.20036674e+00
3.50663930e-01 9.99871135e-01 1.54856564e-02 -1.24631715e+00
-3.46594214e-01 -7.53086209e-01 -1.38526440e+00 1.36588976e-01
4.01258051e-01 -6.73504472e-02 -3.61212909e-01 2.19903183e+00
5.72235823e-01 -2.87762403e-01 -1.99800059e-01 1.17110085e+00
5.92347443e-01 4.41366762e-01 4.22987312e-01 -3.19139659e-01
1.23410594e+00 -6.83620691e-01 -7.88062513e-01 2.14954391e-01
4.07146871e-01 -2.47966588e-01 1.21298063e+00 1.24690130e-01
-9.83893573e-01 -5.57388127e-01 -8.09088171e-01 -7.47118220e-02
-2.20238343e-01 2.02098504e-01 5.56908965e-01 7.54177094e-01
-7.08673656e-01 4.82837975e-01 -5.50764740e-01 -1.83407202e-01
1.03559279e+00 4.94936943e-01 -1.29358992e-01 3.64103578e-02
-1.23987317e+00 3.70250225e-01 3.77546817e-01 -1.23925796e-02
-6.13206446e-01 -8.96598041e-01 -6.09019995e-01 6.05398595e-01
1.52550682e-01 -5.58667004e-01 5.95876098e-01 -1.23142207e+00
-1.06182408e+00 1.07934713e+00 5.22851497e-02 -3.10324073e-01
7.51654506e-01 -1.37344152e-01 -1.49933651e-01 -6.00314513e-02
2.67916262e-01 8.64708066e-01 4.59453166e-01 -1.43877280e+00
-6.17793798e-01 -7.20173955e-01 -8.60256478e-02 3.57383698e-01
-9.14803267e-01 -1.03259377e-01 -2.60504276e-01 -5.80647290e-01
-3.05824488e-01 -6.18770361e-01 -3.17622453e-01 -1.13220721e-01
-5.77515483e-01 -4.05888557e-01 3.77638370e-01 -4.40584302e-01
1.37656438e+00 -2.35298467e+00 6.59225062e-02 5.61573446e-01
5.56463897e-01 1.26520753e-01 2.01627407e-02 -2.05405205e-01
-7.44597763e-02 2.78043568e-01 -3.95155847e-01 -4.09983128e-01
2.82690346e-01 -2.25807041e-01 1.98283046e-02 5.16203344e-01
3.75334293e-01 5.76932013e-01 -8.51437628e-01 -6.63074374e-01
1.14573501e-01 5.20841897e-01 -6.71117127e-01 3.10679227e-01
2.17383131e-01 4.77741599e-01 -3.04087967e-01 3.70211840e-01
1.11104405e+00 -1.39045984e-01 1.82805061e-01 -2.24897638e-01
2.09693700e-01 -1.44234657e-01 -1.08154964e+00 1.24114299e+00
-1.82726756e-01 3.23290639e-02 1.21416144e-01 -1.04184306e+00
1.21607161e+00 -5.28051704e-02 6.18673980e-01 -7.57743955e-01
4.54819292e-01 -1.37497604e-01 7.25657493e-02 -2.84545515e-02
4.51328099e-01 -3.06571543e-01 -1.42543003e-01 5.59237659e-01
-3.27771753e-01 4.21546638e-01 2.74950683e-01 3.63621503e-01
4.42829728e-01 -5.34865260e-01 -7.81568699e-03 -6.27021074e-01
4.68664229e-01 -3.33989590e-01 1.02766526e+00 4.95688528e-01
-6.28325582e-01 5.03916442e-01 7.65745282e-01 -2.38013849e-01
-1.07607400e+00 -9.75161076e-01 -1.65282875e-01 1.29057384e+00
2.04355001e-01 -2.22378090e-01 -1.04366517e+00 -7.19969571e-01
1.17181949e-01 6.18043721e-01 -9.14350390e-01 -5.82605541e-01
-1.88051358e-01 -1.23071766e+00 4.61278588e-01 6.39459789e-01
6.37313247e-01 -9.97022688e-01 -3.50833029e-01 -2.19999000e-01
-3.28508556e-01 -6.97888136e-01 -6.78190112e-01 1.22305498e-01
-6.95239961e-01 -1.12848485e+00 -5.95998824e-01 -5.44409752e-01
8.87631118e-01 6.77184984e-02 1.10462356e+00 2.40804985e-01
-1.43527389e-01 -2.57492423e-01 -1.48524523e-01 -2.80215263e-01
-7.70116299e-02 1.65785775e-01 1.82655051e-01 3.50956053e-01
6.69229984e-01 -4.47186559e-01 -6.89152181e-01 3.15255135e-01
-8.28013480e-01 -2.96345264e-01 3.57150853e-01 6.99699879e-01
4.36953247e-01 2.59338945e-01 8.48812997e-01 -9.05261993e-01
6.73184991e-01 -5.07276595e-01 -5.12645960e-01 1.94735497e-01
-9.00956154e-01 -1.61234081e-01 6.31794810e-01 -3.51832867e-01
-1.05580389e+00 -2.12276176e-01 2.05958441e-01 -3.19859564e-01
-5.42423427e-02 1.46382749e-01 -8.20636630e-01 5.47206700e-01
4.56277132e-01 -1.58360153e-01 1.11062609e-01 -1.91024676e-01
4.81396109e-01 6.62849545e-01 4.08697963e-01 -6.70563817e-01
5.74009359e-01 6.72622025e-01 -2.54601669e-02 -2.50329256e-01
-9.03956711e-01 -3.71537447e-01 -7.85464823e-01 -3.77736054e-02
8.01737249e-01 -1.00358677e+00 -1.12241232e+00 4.79437262e-01
-6.86188459e-01 -1.89493909e-01 -2.52781332e-01 3.53083640e-01
-1.62713870e-01 3.37890923e-01 -5.57653427e-01 -8.75567079e-01
-6.78662539e-01 -9.82231200e-01 6.59563541e-01 4.10741806e-01
-9.79828313e-02 -8.46194685e-01 -8.52512121e-02 4.10395652e-01
3.21888536e-01 3.02453548e-01 1.07550645e+00 -7.45538652e-01
-1.22282021e-01 6.92829117e-02 -6.55369699e-01 5.17024219e-01
1.85886055e-01 1.49098441e-01 -9.91203904e-01 -4.19436127e-01
-2.78215289e-01 -4.39799994e-01 1.01523793e+00 5.70160329e-01
1.52112210e+00 -2.42526397e-01 -2.32242420e-01 6.94116116e-01
1.31479645e+00 5.38025275e-02 6.48496389e-01 2.36324906e-01
8.97225142e-01 9.91388083e-01 6.93469286e-01 7.90146172e-01
4.90156204e-01 3.56779546e-01 3.90892476e-01 -2.98222750e-01
2.27795094e-01 5.55645116e-02 -8.08617473e-02 5.57196856e-01
1.03933334e-01 -2.70806029e-02 -8.47214580e-01 5.68539977e-01
-1.82734048e+00 -9.56993222e-01 -7.66531825e-02 2.32523680e+00
8.97287548e-01 -6.83371630e-03 4.22153950e-01 1.39599904e-01
1.43490744e+00 -2.18417168e-01 -7.65202165e-01 -1.16271220e-01
-1.91312596e-01 -2.10032746e-01 2.00420156e-01 3.16538364e-01
-1.19569373e+00 8.00026298e-01 5.30376101e+00 9.33044732e-01
-9.64081109e-01 -8.09881743e-03 1.42890382e+00 -3.61822337e-01
-2.79910922e-01 -3.26821178e-01 -5.04617512e-01 8.75217021e-01
4.34472591e-01 -2.16088429e-01 3.44349176e-01 6.88182175e-01
3.84069800e-01 2.31751308e-01 -1.02890182e+00 9.46882427e-01
-1.79138258e-01 -7.71303236e-01 -5.24128750e-02 1.61489606e-01
6.67239368e-01 -5.38208723e-01 4.31281835e-01 2.36920565e-01
5.99093616e-01 -1.04325020e+00 5.65758348e-01 3.15264821e-01
7.23143280e-01 -1.45832896e+00 7.89266765e-01 -1.04690520e-02
-8.29628706e-01 -2.53098726e-01 -5.90923965e-01 1.89690903e-01
-2.05105603e-01 1.05005336e+00 -2.99792200e-01 3.38979065e-01
9.77152944e-01 6.29997253e-01 -7.37802625e-01 8.76773596e-01
6.55041784e-02 4.14602935e-01 -1.17271177e-01 1.93988204e-01
3.20239440e-02 -5.22933662e-01 -1.73718277e-02 1.06144655e+00
1.10226944e-01 1.04063660e-01 1.88978352e-02 1.23601115e+00
-4.96898890e-01 2.73384035e-01 -1.02169126e-01 2.52636909e-01
8.94499123e-01 1.68783212e+00 -6.29534125e-01 -2.36374393e-01
1.40629634e-01 5.58958471e-01 5.55261374e-01 1.16379090e-01
-1.00538599e+00 -4.65602517e-01 7.41986990e-01 -6.06670342e-02
2.14692950e-02 3.67870927e-01 -5.84086597e-01 -9.67371583e-01
-1.73814520e-01 -7.68939018e-01 7.00211525e-01 -2.94012219e-01
-1.78133309e+00 3.71192634e-01 -3.57183367e-01 -1.07678592e+00
5.32260656e-01 7.63061345e-02 -6.45469248e-01 8.35318029e-01
-1.46740818e+00 -1.07414126e+00 -6.27908945e-01 7.82067835e-01
-2.18024984e-01 -1.97283164e-01 4.06267554e-01 6.88249707e-01
-1.00670135e+00 1.08518136e+00 2.69308418e-01 3.48540515e-01
1.17847133e+00 -1.37724102e+00 -1.93956837e-01 6.70116842e-01
-5.40792704e-01 8.09181392e-01 3.69924873e-01 -4.01671886e-01
-5.34349442e-01 -1.39644778e+00 5.51027775e-01 -2.66799331e-01
1.16408497e-01 -4.81103331e-01 -1.00189757e+00 2.30859444e-01
5.33939712e-02 -1.17483601e-01 1.10873687e+00 2.97401637e-01
-5.29911876e-01 -6.87217355e-01 -1.44207859e+00 3.30374122e-01
7.68196821e-01 -1.54217228e-01 9.71425921e-02 1.10960834e-01
6.32488549e-01 1.73247680e-01 -1.09350777e+00 3.61511558e-01
3.26697946e-01 -1.04512000e+00 8.51983070e-01 -5.10641992e-01
6.97248340e-01 -2.96995044e-01 -4.50568795e-02 -1.42071044e+00
-7.43739426e-01 -1.06645979e-01 2.35121578e-01 2.01829815e+00
4.09068093e-02 -2.69195586e-01 8.24469924e-01 8.12141776e-01
1.57536343e-01 -4.43416983e-01 -6.20940924e-01 -3.70020658e-01
4.85645443e-01 2.44051635e-01 9.87163365e-01 1.42506683e+00
-1.07250251e-01 5.38110673e-01 -3.71374846e-01 6.37792125e-02
9.71147776e-01 1.21216983e-01 6.78084254e-01 -1.42056096e+00
2.84044445e-01 -7.60551393e-01 -1.55605584e-01 -2.10786313e-01
3.55319142e-01 -9.10842597e-01 -5.49817830e-02 -1.16592813e+00
8.10525656e-01 -7.70714641e-01 -5.90491593e-01 3.29511315e-01
-6.61406279e-01 1.82053551e-01 1.37072489e-01 2.68177897e-01
-8.23471069e-01 9.41611469e-01 9.22785282e-01 -2.52801895e-01
-5.11791110e-02 -2.72320986e-01 -1.33394134e+00 3.52550030e-01
8.61944497e-01 -5.22699714e-01 -3.78903896e-01 -2.92551488e-01
1.20057566e-02 -4.17531759e-01 1.70889720e-01 -8.99469078e-01
-2.07771901e-02 -2.48105317e-01 6.44421279e-01 -3.71536136e-01
-8.28980356e-02 -8.80431652e-01 -1.70680836e-01 4.21761453e-01
-7.59728551e-01 -4.98781651e-02 -1.16799362e-01 4.75748062e-01
-1.07648402e-01 3.73024911e-01 1.34837568e+00 7.67406598e-02
-2.67346948e-01 5.60894668e-01 2.93397307e-01 3.10107231e-01
1.20804155e+00 1.57808870e-01 -3.91637385e-01 -2.17215672e-01
-4.53875631e-01 8.04817975e-01 6.61411047e-01 2.44546354e-01
3.45158041e-01 -1.69290245e+00 -9.45640087e-01 -5.13178781e-02
3.41787785e-01 9.56015140e-02 3.05646509e-01 7.74284422e-01
-2.03784183e-01 -2.17581820e-02 -5.19057631e-01 -5.40320814e-01
-1.25201499e+00 9.50977683e-01 3.82419735e-01 -2.36599118e-01
1.85880959e-02 7.12976396e-01 5.49482107e-01 -6.74097061e-01
5.31036377e-01 1.59749478e-01 -3.44679236e-01 3.09213072e-01
5.37171364e-01 6.24794602e-01 -3.99516225e-02 -5.50832987e-01
-5.62561333e-01 3.09824675e-01 -4.31603678e-02 2.44993076e-01
1.30778825e+00 -3.69005919e-01 -5.26522279e-01 1.18823148e-01
1.08701587e+00 -1.81401238e-01 -1.40884542e+00 -1.71169952e-01
-1.45541325e-01 -6.12376630e-01 -8.16152692e-02 -7.76212811e-01
-1.60583901e+00 1.20441437e+00 8.56971145e-01 -3.47444527e-02
1.17802680e+00 -2.47658387e-01 6.41915381e-01 6.56829157e-04
-1.88296765e-01 -1.37960017e+00 2.55654991e-01 2.65218150e-02
3.50631267e-01 -1.56279492e+00 -4.11530472e-02 -1.73907086e-01
-1.01052523e+00 6.35366142e-01 1.06258535e+00 6.07261695e-02
4.93070871e-01 -7.28958100e-02 1.58919707e-01 -1.20808996e-01
-3.50795060e-01 -3.93739995e-03 8.39274675e-02 6.58783913e-01
6.63728416e-01 2.38969818e-01 -3.56290638e-01 8.73358190e-01
-1.39149085e-01 -3.43457878e-01 3.19465697e-01 2.22758785e-01
-2.53044099e-01 -6.95033133e-01 -3.41831267e-01 3.73530656e-01
-6.51637912e-01 -7.65485615e-02 -5.20355523e-01 4.63338912e-01
3.13941032e-01 1.12008929e+00 4.44226593e-01 -2.89674371e-01
1.01037823e-01 -2.46927306e-01 1.69320479e-01 -3.33394170e-01
-8.20334017e-01 6.12156391e-02 -4.30157006e-01 -3.69526207e-01
-4.83600318e-01 -3.41053784e-01 -1.20300329e+00 -8.08349609e-01
-2.50275910e-01 3.88174385e-01 3.36985230e-01 4.51276720e-01
4.48645771e-01 4.24018502e-01 1.04784358e+00 -2.82623172e-01
-5.65812051e-01 -8.80097091e-01 -9.92593944e-01 1.05628955e+00
4.73450907e-02 -6.13640547e-01 -3.09885085e-01 -7.54102170e-02] | [9.2808198928833, 3.84321928024292] |
46bf5510-4f16-44a1-b085-0578f2e1d7ad | instance-segmentation-of-fallen-trees-in | 2105.01998 | null | https://arxiv.org/abs/2105.01998v1 | https://arxiv.org/pdf/2105.01998v1.pdf | Instance segmentation of fallen trees in aerial color infrared imagery using active multi-contour evolution with fully convolutional network-based intensity priors | In this paper, we introduce a framework for segmenting instances of a common object class by multiple active contour evolution over semantic segmentation maps of images obtained through fully convolutional networks. The contour evolution is cast as an energy minimization problem, where the aggregate energy functional incorporates a data fit term, an explicit shape model, and accounts for object overlap. Efficient solution neighborhood operators are proposed, enabling optimization through metaheuristics such as simulated annealing. We instantiate the proposed framework in the context of segmenting individual fallen stems from high-resolution aerial multispectral imagery. We validated our approach on 3 real-world scenes of varying complexity. The test plots were situated in regions of the Bavarian Forest National Park, Germany, which sustained a heavy bark beetle infestation. Evaluations were performed on both the polygon and line segment level, showing that the multi-contour segmentation can achieve up to 0.93 precision and 0.82 recall. An improvement of up to 7 percentage points (pp) in recall and 6 in precision compared to an iterative sample consensus line segment detection was achieved. Despite the simplicity of the applied shape parametrization, an explicit shape model incorporated into the energy function improved the results by up to 4 pp of recall. Finally, we show the importance of using a deep learning based semantic segmentation method as the basis for individual stem detection. Our method is a step towards increased accessibility of automatic fallen tree mapping, due to higher cost efficiency of aerial imagery acquisition compared to laser scanning. The precise fallen tree maps could be further used as a basis for plant and animal habitat modeling, studies on carbon sequestration as well as soil quality in forest ecosystems. | ['Marco Heurich', 'Wei Yao', 'Jacquelyn Shelton', 'Przemyslaw Polewski'] | 2021-05-05 | null | null | null | null | ['line-segment-detection'] | ['computer-vision'] | [ 5.89333236e-01 2.51047432e-01 3.28005366e-02 -1.90212473e-01
-5.42834044e-01 -6.53298140e-01 4.88571107e-01 6.38952434e-01
-7.93353856e-01 7.10908771e-01 -5.71985185e-01 -3.20976883e-01
-5.50994992e-01 -1.20439517e+00 -5.68118572e-01 -7.35866249e-01
-3.34549636e-01 7.30783761e-01 4.38850224e-01 -1.39472470e-01
1.19309239e-01 1.11523116e+00 -1.76154006e+00 -3.56510431e-02
1.12462604e+00 9.55263734e-01 6.38998210e-01 4.92318034e-01
-2.34985724e-01 -2.03997210e-01 -4.51063007e-01 -1.42813921e-01
3.28461766e-01 4.54633869e-02 -7.12280452e-01 4.51847017e-01
3.09858143e-01 -1.19770542e-01 6.66707873e-01 1.18585277e+00
1.01458758e-01 -9.82651338e-02 8.00814092e-01 -9.59294736e-01
2.78625190e-02 4.76442248e-01 -5.91199577e-01 -1.91070542e-01
-1.97560191e-01 2.10096687e-01 1.02676475e+00 -6.58843994e-01
7.04869747e-01 7.43984401e-01 8.25378120e-01 -1.33847997e-01
-1.74937582e+00 -2.04926774e-01 9.98165086e-02 1.42769501e-01
-1.57141817e+00 -7.00041056e-02 4.87792492e-01 -5.55710554e-01
8.37596357e-01 4.72354621e-01 1.10778058e+00 2.36429185e-01
4.15121642e-04 6.02837265e-01 1.01581395e+00 -3.96537602e-01
5.26769876e-01 8.78930688e-02 6.23793788e-02 5.36308467e-01
5.85285783e-01 1.22976489e-01 1.82362020e-01 -8.35994035e-02
9.19915259e-01 -3.06857497e-01 -2.26707056e-01 -7.89368033e-01
-7.55930483e-01 9.60002065e-01 9.97575700e-01 5.69940031e-01
-7.12048888e-01 3.68130207e-03 1.87048614e-01 -1.89972371e-01
6.04656398e-01 4.18291509e-01 -3.97786409e-01 4.60493654e-01
-1.52010357e+00 3.45506310e-01 7.80731678e-01 4.25242811e-01
8.98509502e-01 -7.18968213e-02 2.08579779e-01 7.64493883e-01
3.53130162e-01 7.91662812e-01 -6.98231831e-02 -9.40522373e-01
-4.02284600e-02 9.51209188e-01 1.65234149e-01 -1.06442952e+00
-5.83921731e-01 -5.73859453e-01 -7.41984725e-01 7.39969015e-01
4.00189012e-01 5.76154143e-02 -1.05959940e+00 1.31536043e+00
3.02202404e-01 -4.38987911e-01 -1.59715906e-01 8.87502611e-01
2.14999869e-01 4.82593536e-01 2.58235961e-01 -4.53622818e-01
1.28741169e+00 -3.63560975e-01 -4.24062163e-01 -7.53254667e-02
4.83538270e-01 -5.33521652e-01 6.55651510e-01 4.21702117e-01
-9.08776283e-01 -4.69471812e-01 -1.22708154e+00 6.05213761e-01
-8.18202436e-01 4.32875901e-01 5.87726712e-01 7.20656157e-01
-1.04080641e+00 8.94484162e-01 -8.53610158e-01 -5.35430551e-01
6.80015564e-01 5.08158922e-01 -2.00210288e-01 4.01907712e-01
-8.27655077e-01 9.17865038e-01 7.61348724e-01 5.75511813e-01
-3.68606031e-01 -4.24409539e-01 -5.65336287e-01 2.22074330e-01
3.67529213e-01 -3.60017955e-01 7.70965576e-01 -1.20894456e+00
-1.15280378e+00 1.29601181e+00 3.15907270e-01 -9.19428587e-01
7.37832785e-01 -1.74705535e-01 -9.31419581e-02 3.08495730e-01
1.90539390e-01 8.16176414e-01 5.20036399e-01 -1.48513734e+00
-6.25562906e-01 -6.29051805e-01 -1.40195400e-01 2.14464869e-02
4.65005152e-02 -3.30644548e-01 1.63336203e-01 -4.20949012e-01
4.66806322e-01 -9.12339747e-01 -4.76108700e-01 1.91841587e-01
-7.07388371e-02 4.04716223e-01 7.42095292e-01 -7.08625913e-01
1.00998878e+00 -1.73764825e+00 2.68562168e-01 5.84422708e-01
-5.08321729e-03 4.30236071e-01 1.41966194e-01 2.02965245e-01
-7.24919513e-02 3.09820175e-01 -1.19089544e+00 1.82515919e-01
-1.47434726e-01 3.40253770e-01 1.81994773e-02 3.43960613e-01
3.91979039e-01 7.38537073e-01 -6.78914428e-01 -4.64360297e-01
4.92736220e-01 4.65475589e-01 -5.69042340e-02 -1.04499489e-01
-4.15925384e-01 -8.59429464e-02 -1.65613309e-01 6.08293355e-01
9.89759982e-01 1.81722552e-01 2.05247596e-01 -7.39262179e-02
-7.11512804e-01 -4.21019256e-01 -1.43355179e+00 1.42406487e+00
-2.82849073e-01 3.99845541e-01 4.35310096e-01 -1.03967142e+00
1.29478180e+00 6.95796907e-02 7.62987196e-01 -4.53800708e-01
4.57108207e-02 5.79429328e-01 -1.97582375e-02 -4.15889993e-02
6.62061512e-01 -4.79012765e-02 4.33911949e-01 1.13027478e-02
-2.29921136e-02 -5.88304102e-01 3.69463444e-01 -3.20253640e-01
4.66694891e-01 3.94358665e-01 3.62561673e-01 -7.89299011e-01
6.62470996e-01 5.88178754e-01 9.82006118e-02 3.55337977e-01
-1.97555587e-01 4.35425073e-01 4.07048166e-01 -3.79229516e-01
-9.65459406e-01 -1.06661713e+00 -7.13598669e-01 6.68754220e-01
-2.07332429e-03 1.82083156e-02 -1.00389898e+00 -4.34890836e-01
1.19520545e-01 6.97142363e-01 -4.63609457e-01 1.85681462e-01
-4.35135901e-01 -1.27369058e+00 4.07510281e-01 2.22549841e-01
9.61987019e-01 -1.22168791e+00 -1.24543619e+00 5.11129498e-01
-3.12767960e-02 -7.56368816e-01 4.92463738e-01 6.94244504e-01
-1.28729415e+00 -1.04206336e+00 -8.99422348e-01 -3.38927746e-01
3.58259469e-01 -2.73008607e-02 1.03895092e+00 9.14747193e-02
-6.56915963e-01 2.95921624e-01 -2.92495638e-01 -2.49615327e-01
-4.68973011e-01 4.22764748e-01 -4.77892935e-01 -1.82494536e-01
1.68890372e-01 -5.35216093e-01 -5.20584762e-01 2.75582463e-01
-9.18616891e-01 -1.05221562e-01 5.56663811e-01 6.55501127e-01
8.72193038e-01 1.02622099e-01 2.89022863e-01 -7.32918739e-01
2.58285284e-01 -1.81962132e-01 -1.03349864e+00 2.56874293e-01
-7.39494324e-01 -2.09515318e-01 1.96098864e-01 -2.17149928e-02
-8.36829841e-01 6.52882576e-01 -6.10379949e-02 1.55079544e-01
-6.00801349e-01 5.13979733e-01 -1.67797640e-01 -3.16447288e-01
8.72887313e-01 3.24992789e-03 -1.55773712e-02 -2.20656350e-01
3.47210348e-01 4.38392311e-01 5.11812508e-01 -3.41100365e-01
7.50111699e-01 4.81900424e-01 2.41750851e-01 -1.20994341e+00
-1.89889759e-01 -5.41907370e-01 -1.09413242e+00 -5.32961667e-01
9.54232574e-01 -4.45541561e-01 -4.64760065e-01 5.00423729e-01
-1.06033003e+00 -3.94633025e-01 -5.50044417e-01 2.09573835e-01
-7.46192634e-01 2.91376233e-01 1.34597439e-02 -1.13902259e+00
-5.16123593e-01 -1.00771296e+00 1.03705537e+00 2.36671165e-01
1.92540474e-02 -8.36730242e-01 2.56794738e-03 1.42912358e-01
3.87174755e-01 5.37459433e-01 7.92208016e-01 -2.88437128e-01
-6.28127217e-01 -1.09897546e-01 -4.84927207e-01 3.79049540e-01
-1.64921451e-02 4.60020423e-01 -9.77602720e-01 -2.45700292e-02
-9.70342606e-02 4.83595906e-03 1.06265700e+00 6.47963047e-01
8.27528596e-01 1.98056147e-01 -4.35538024e-01 3.89417052e-01
2.00558639e+00 3.91555548e-01 6.66147709e-01 5.18749774e-01
3.04214120e-01 8.67725849e-01 7.77662516e-01 2.73733228e-01
-3.33738923e-01 7.45585561e-01 9.24373329e-01 -3.95557463e-01
3.02497242e-02 4.37913150e-01 -4.95129339e-02 -1.82914123e-01
-5.10497749e-01 -7.97841996e-02 -1.25379670e+00 6.38884127e-01
-1.77093363e+00 -9.53777969e-01 -7.33180761e-01 2.49984121e+00
3.02542418e-01 3.75959307e-01 1.56806335e-01 4.78643477e-01
8.49848688e-01 2.46380009e-02 -3.57759535e-01 -3.60209703e-01
-3.39092851e-01 4.89197940e-01 9.92176592e-01 6.53524458e-01
-1.27069235e+00 1.07933533e+00 5.39749908e+00 7.12837338e-01
-1.11642611e+00 -1.63096085e-01 3.38242352e-01 3.07321399e-01
1.63815901e-01 9.66486707e-02 -6.13385439e-01 7.27463216e-02
6.91277742e-01 1.58321500e-01 3.26533020e-01 7.02790916e-01
2.87926883e-01 -6.93776786e-01 -3.15885484e-01 4.92036402e-01
-3.93591166e-01 -1.05496252e+00 -1.75035268e-01 3.59872311e-01
4.14452702e-01 3.37525941e-02 -3.29015881e-01 -1.91884309e-01
3.50096375e-02 -9.34237301e-01 7.44594693e-01 4.82270032e-01
5.77173471e-01 -8.00433993e-01 7.08520055e-01 3.27692330e-01
-1.49489200e+00 -2.00397577e-02 -1.94155231e-01 7.07251728e-02
1.83638647e-01 5.44551373e-01 -7.63165891e-01 7.92748213e-01
5.03491104e-01 2.03315914e-01 -7.63102353e-01 1.22086275e+00
2.69676670e-02 3.95161092e-01 -8.27465713e-01 -4.51454800e-03
5.31852841e-01 -6.36132300e-01 5.90307295e-01 1.27906489e+00
2.58799762e-01 -6.82961419e-02 6.06049560e-02 1.26686990e+00
5.20696402e-01 4.14993882e-01 -4.03481185e-01 3.36692967e-02
2.02950567e-01 1.45637023e+00 -1.54146874e+00 -3.09434254e-02
2.78014541e-01 1.01679289e+00 -6.37751957e-03 1.01273142e-01
-7.20340729e-01 -2.29556993e-01 5.17860726e-02 3.77680421e-01
3.54513675e-01 -3.48489106e-01 -5.61954379e-01 -5.52913368e-01
-5.90502052e-03 -2.94342935e-01 1.57626376e-01 -7.67614067e-01
-6.72255278e-01 5.59252203e-01 2.82732248e-01 -7.96666384e-01
-1.77308619e-01 -7.84032285e-01 -3.62056464e-01 9.89515007e-01
-1.22398901e+00 -1.25024390e+00 -5.24776638e-01 -1.76661778e-02
3.73034865e-01 1.43349886e-01 1.15667713e+00 -1.26984105e-01
-2.94191718e-01 -2.47728556e-01 2.27927461e-01 -3.74056369e-01
-1.12652183e-01 -1.39592016e+00 1.47450238e-01 7.11394966e-01
1.05141163e-01 5.91799915e-02 6.33091748e-01 -7.15371370e-01
-4.82257664e-01 -1.00564039e+00 6.43770874e-01 2.27247193e-01
5.33808053e-01 -6.54882938e-02 -1.09192431e+00 2.19419226e-01
1.85713470e-02 -2.83735573e-01 3.10380936e-01 -2.03521118e-01
3.36102754e-01 -6.82159066e-02 -1.45049214e+00 2.46695220e-01
9.44748878e-01 -9.77781340e-02 -2.78630644e-01 2.94247985e-01
9.89289656e-02 2.65135653e-02 -8.63211274e-01 6.96657598e-01
6.62244499e-01 -1.23962927e+00 1.03852332e+00 -2.89940946e-02
3.36027205e-01 -1.62842661e-01 -3.29689503e-01 -1.09183633e+00
-4.52367097e-01 8.26017873e-04 4.41823274e-01 1.17245674e+00
3.54457021e-01 -4.23497826e-01 8.78349125e-01 1.16777465e-01
1.29604228e-02 -5.40463865e-01 -9.89923418e-01 -7.52134800e-01
1.51353225e-01 -3.75643671e-01 2.35144272e-01 4.55441743e-01
-7.13873386e-01 -5.22093363e-02 4.61010903e-01 3.80026966e-01
7.60695398e-01 1.51113793e-01 4.29840088e-01 -1.93684411e+00
9.36935917e-02 -9.23477888e-01 -5.59968233e-01 -3.49252552e-01
1.83470011e-01 -8.18892300e-01 -1.28903287e-02 -1.61085534e+00
-1.96331739e-01 -6.26816094e-01 1.18307799e-01 3.86884421e-01
1.55614644e-01 2.79293656e-01 1.84672147e-01 2.13043004e-01
3.22102398e-01 3.68022680e-01 7.28590071e-01 -1.41740024e-01
-4.21751201e-01 2.34210521e-01 1.53598143e-02 8.88534367e-01
9.50940967e-01 -4.21537340e-01 -1.66327983e-01 -9.59419534e-02
-4.99388948e-02 -3.93986441e-02 8.80203426e-01 -1.24018455e+00
7.70102115e-03 -1.17127642e-01 3.76574367e-01 -9.30633545e-01
3.70738834e-01 -1.40492189e+00 5.85621417e-01 9.38779414e-01
-4.46673594e-02 -4.84133452e-01 5.25759280e-01 2.50118017e-01
6.20241053e-02 -7.74651527e-01 1.09996474e+00 -3.46672863e-01
-7.89862454e-01 -5.57082333e-03 -4.71380055e-01 -5.34181774e-01
1.11188483e+00 -6.69793963e-01 5.35935834e-02 1.42002717e-01
-1.08900476e+00 -8.80256016e-03 6.09064877e-01 3.46575603e-02
3.34575057e-01 -9.38827634e-01 -6.69364035e-01 1.94311410e-01
6.83367923e-02 1.75546005e-03 1.32561311e-01 7.27197766e-01
-1.08170378e+00 3.67257059e-01 -6.18826628e-01 -9.10591483e-01
-1.38271856e+00 2.64841229e-01 7.29106665e-01 -3.36819530e-01
-2.59271771e-01 4.04053956e-01 -2.16174796e-01 -3.46239597e-01
1.24881351e-02 -4.35851961e-01 -3.32787454e-01 5.74564338e-01
-1.04445040e-01 5.38629651e-01 2.72563130e-01 -8.77941310e-01
-4.15563345e-01 6.95900619e-01 4.19748366e-01 -2.43477613e-01
1.38809836e+00 5.01101539e-02 -8.38225856e-02 3.87490362e-01
6.22376204e-01 -3.80387843e-01 -1.18781173e+00 2.09790021e-01
4.95404303e-01 -1.68619275e-01 4.84590858e-01 -1.07854426e+00
-1.13064885e+00 6.84882462e-01 1.19444907e+00 6.17144525e-01
1.18611324e+00 -3.51195067e-01 1.17501512e-01 2.62153298e-01
4.61942405e-01 -1.09025896e+00 -6.93250775e-01 2.17267275e-01
1.06875730e+00 -1.17117178e+00 2.54426718e-01 -6.31819010e-01
-1.56992286e-01 1.40259087e+00 2.21120119e-01 -1.44817799e-01
4.51748252e-01 3.40445161e-01 -1.24661095e-01 -3.29326481e-01
3.04557346e-02 -9.05632734e-01 3.84525210e-02 7.51256049e-01
2.30004653e-01 4.40913796e-01 -7.09372163e-01 2.05655426e-01
-8.31754506e-03 2.06213892e-02 1.97825089e-01 8.65416467e-01
-8.60710144e-01 -9.34648395e-01 -6.95327759e-01 3.57701510e-01
3.79593205e-03 3.12399659e-02 -7.20115304e-01 1.25889766e+00
3.89541209e-01 5.48534274e-01 1.55729115e-01 3.50613669e-02
4.55219358e-01 1.72722071e-01 5.06386280e-01 -3.98330688e-01
-8.21775138e-01 1.84319615e-01 -1.07192248e-01 -2.97268063e-01
-6.89217091e-01 -8.19697797e-01 -1.29806066e+00 -4.14101891e-02
-5.99885643e-01 6.41748533e-02 1.11501706e+00 6.87160194e-01
-1.38290748e-01 3.86030555e-01 3.00021827e-01 -1.04274917e+00
-2.65772939e-01 -9.54050124e-01 -9.14898753e-01 -2.25198045e-01
-1.54201925e-01 -8.00460279e-01 -3.43531013e-01 1.24225616e-01] | [9.312124252319336, -1.6124305725097656] |
22bfc8af-835c-4555-a3ca-e8d70a8e03b3 | query-expansion-with-locally-trained-word | 1605.07891 | null | http://arxiv.org/abs/1605.07891v2 | http://arxiv.org/pdf/1605.07891v2.pdf | Query Expansion with Locally-Trained Word Embeddings | Continuous space word embeddings have received a great deal of attention in
the natural language processing and machine learning communities for their
ability to model term similarity and other relationships. We study the use of
term relatedness in the context of query expansion for ad hoc information
retrieval. We demonstrate that word embeddings such as word2vec and GloVe, when
trained globally, underperform corpus and query specific embeddings for
retrieval tasks. These results suggest that other tasks benefiting from global
embeddings may also benefit from local embeddings. | ['Nick Craswell', 'Bhaskar Mitra', 'Fernando Diaz'] | 2016-05-25 | query-expansion-with-locally-trained-word-1 | https://aclanthology.org/P16-1035 | https://aclanthology.org/P16-1035.pdf | acl-2016-8 | ['ad-hoc-information-retrieval'] | ['natural-language-processing'] | [-9.50230211e-02 -2.75106907e-01 -4.92881685e-01 -2.64914125e-01
-7.59721935e-01 -5.68384290e-01 1.12883878e+00 8.54079485e-01
-9.98658061e-01 2.09378138e-01 8.30424428e-01 -4.92424935e-01
-5.27907491e-01 -6.79432631e-01 1.47772238e-01 -1.56938672e-01
-3.18776757e-01 4.42296654e-01 1.67688355e-01 -6.19604230e-01
5.85589409e-01 4.73890930e-01 -1.29285049e+00 -8.96546990e-02
4.28884566e-01 6.19363785e-01 4.01827786e-03 5.44027686e-01
-3.43462020e-01 1.84691146e-01 -4.96206135e-01 -1.22040316e-01
-3.42660882e-02 1.27489731e-01 -9.24494565e-01 -5.85197628e-01
3.82446647e-01 -1.83624864e-01 -8.65516424e-01 6.50343180e-01
7.78114259e-01 8.47752094e-01 8.60472858e-01 -9.00184572e-01
-1.31384075e+00 2.22508982e-01 -2.76293755e-01 6.86142325e-01
4.43473130e-01 -2.28288502e-01 1.72987008e+00 -1.09075260e+00
6.63551271e-01 1.28785300e+00 2.79750735e-01 2.41225019e-01
-1.27294040e+00 -2.52015829e-01 7.84645453e-02 3.23202550e-01
-1.70034170e+00 -1.08206935e-01 5.44623852e-01 -3.41144532e-01
1.61206758e+00 2.97431380e-01 3.39841902e-01 8.31055760e-01
2.14432836e-01 5.77261627e-01 3.31705272e-01 -6.85044050e-01
-1.87011570e-01 1.68668866e-01 5.40279210e-01 1.73923731e-01
3.22384775e-01 8.35079625e-02 -2.71506578e-01 -6.10956609e-01
4.65843379e-01 4.06870991e-01 -2.27016836e-01 -2.33192831e-01
-1.02405190e+00 1.31222415e+00 8.63846421e-01 8.09133649e-01
-5.60259163e-01 6.52631581e-01 6.50194883e-01 3.92388165e-01
6.99295640e-01 1.21865094e+00 -3.94028395e-01 -2.39809349e-01
-4.69285369e-01 3.55585128e-01 5.32681346e-01 8.83756876e-01
8.21533680e-01 -3.91835839e-01 -3.49297076e-01 1.24050808e+00
4.15371239e-01 4.18583810e-01 1.08934081e+00 -3.93782794e-01
-7.58559406e-02 4.29077953e-01 7.62724504e-03 -1.27897608e+00
-2.09750399e-01 -1.66010693e-01 -4.77044545e-02 -5.04273355e-01
-2.69133568e-01 2.41637915e-01 -9.06914473e-01 1.61927176e+00
3.04858629e-02 -2.82262564e-01 5.09994924e-02 7.71654248e-01
6.35751188e-01 5.09556293e-01 2.96475291e-01 -7.52304075e-03
1.46640682e+00 -8.01287591e-01 -7.31351137e-01 -5.00816941e-01
1.23458648e+00 -9.09265995e-01 1.00301850e+00 -2.86918849e-01
-8.01169038e-01 -3.36233646e-01 -7.86969900e-01 -5.31710088e-01
-1.03644300e+00 -6.47102714e-01 9.04992759e-01 2.86642104e-01
-1.23558283e+00 2.01125398e-01 -6.70380116e-01 -9.49105144e-01
1.42226312e-02 5.08750863e-02 -3.20557445e-01 -5.63319802e-01
-1.49832380e+00 1.45762384e+00 3.92202169e-01 -3.09761405e-01
-3.85739595e-01 -5.79627931e-01 -1.11443162e+00 3.15022737e-01
-6.61856160e-02 -4.88008052e-01 1.10699785e+00 -1.45351261e-01
-4.81418967e-01 7.54813135e-01 -2.39418805e-01 -2.97051847e-01
-6.52224839e-01 -4.77938592e-01 -5.66886723e-01 1.58160940e-01
-6.63269088e-02 7.45789945e-01 4.31478083e-01 -6.25232637e-01
-2.15673849e-01 -4.24664289e-01 -1.88628882e-02 3.82518262e-01
-1.08629441e+00 3.60077262e-01 -5.17996073e-01 -9.50990140e-01
-2.23150805e-01 -8.22559237e-01 -2.56159097e-01 2.11081922e-01
3.52934450e-01 -1.11059856e+00 7.65511394e-01 -4.82355297e-01
1.65929139e+00 -2.18263221e+00 5.44122308e-02 1.62428483e-01
2.16363460e-01 4.80941266e-01 -6.68311834e-01 1.20019329e+00
5.97821437e-02 5.64817250e-01 3.11704516e-01 5.76288179e-02
2.06725881e-01 3.60414565e-01 -4.31333601e-01 2.91060627e-01
1.47968680e-01 1.35704517e+00 -1.13162458e+00 -5.29387593e-01
-1.08370008e-02 6.92224622e-01 -6.16700351e-01 2.41576806e-01
-8.73524416e-03 -6.75062299e-01 -7.25950539e-01 4.16413814e-01
6.28683046e-02 -1.58456981e-01 1.15508497e-01 -1.51188420e-02
1.91534683e-01 4.87541705e-01 -3.14802736e-01 1.98678279e+00
-7.30987608e-01 1.07391191e+00 -2.68323004e-01 -7.49596417e-01
7.86646366e-01 4.21622694e-01 4.65895295e-01 -9.96969819e-01
1.38367582e-02 2.12108772e-02 8.23653117e-02 -6.37902260e-01
1.15085614e+00 -2.63835460e-01 -2.31620923e-01 8.83392096e-01
1.05618894e-01 -2.19714314e-01 5.16947592e-03 6.45378053e-01
1.33228528e+00 -5.24074376e-01 4.32669133e-01 -3.29498380e-01
1.45903185e-01 5.84705919e-02 -9.35926139e-02 6.16643131e-01
-1.51713356e-01 4.18085128e-01 -2.03182071e-01 -6.69342801e-02
-8.85816395e-01 -9.67742324e-01 -4.06616360e-01 1.66776073e+00
1.77900687e-01 -6.85141563e-01 2.01155409e-01 -4.44425076e-01
6.09497547e-01 8.31901193e-01 -8.03367078e-01 -6.97986722e-01
-4.34672743e-01 -2.16696501e-01 5.65980136e-01 8.43651950e-01
-5.36699831e-01 -9.96006191e-01 -1.69384882e-01 2.65578717e-01
2.77627945e-01 -7.83135295e-01 -9.54590142e-01 1.88977018e-01
-8.54089737e-01 -8.73175085e-01 -8.08850825e-01 -9.85333323e-01
3.36950660e-01 7.43222952e-01 1.20273864e+00 4.39872175e-01
-6.44084394e-01 1.33227575e+00 -8.74950469e-01 -3.60660136e-01
9.11830962e-02 1.02739491e-01 2.87500858e-01 -6.53588176e-01
1.18603683e+00 -4.53661382e-01 -7.68772006e-01 1.74184218e-01
-1.53305244e+00 -1.11696184e+00 3.48912179e-01 1.04548693e+00
1.82588056e-01 -5.69619060e-01 7.78327644e-01 -4.12517071e-01
1.66346598e+00 -7.93102741e-01 3.09161399e-03 6.23577297e-01
-1.27746260e+00 3.61136496e-01 -1.37360930e-01 -6.14065289e-01
-5.36304832e-01 -8.76089990e-01 9.21678320e-02 -4.27364707e-01
1.66166112e-01 1.04950464e+00 5.22653878e-01 -9.50106904e-02
9.91593003e-01 6.97936639e-02 1.45579493e-02 -5.81087649e-01
9.47544932e-01 8.56370091e-01 -1.70338139e-01 -5.85428476e-01
7.80105829e-01 4.74764891e-02 -5.00939190e-01 -9.50879574e-01
-3.33989531e-01 -1.21871746e+00 -2.22031757e-01 3.16089392e-01
9.82286334e-01 -7.38042712e-01 3.80570954e-03 -5.43183863e-01
-1.24242020e+00 3.74284416e-01 -4.72291470e-01 7.24752247e-01
-7.56637380e-03 4.02489632e-01 -4.56253022e-01 -5.48290014e-01
-3.51185501e-01 -6.36923611e-01 1.05789030e+00 -2.20105220e-02
-6.18984640e-01 -1.59603846e+00 6.68098509e-01 -4.65711057e-02
9.90210176e-01 -2.44385287e-01 1.22588515e+00 -1.43793619e+00
-9.41047370e-02 -8.90029073e-01 -1.75813884e-01 2.78026998e-01
4.33925599e-01 -4.09355551e-01 -7.11440980e-01 -4.33758885e-01
-4.04979974e-01 -5.53129315e-01 1.07130253e+00 1.26645928e-02
7.30174541e-01 -1.67213678e-01 -5.09910524e-01 1.73080102e-01
1.53796029e+00 1.67775080e-01 3.11899483e-01 2.89795905e-01
5.04034936e-01 5.86694181e-01 4.46830630e-01 2.42650852e-01
1.36152301e-02 6.71226740e-01 1.64220676e-01 1.41346484e-01
4.88759838e-02 -3.81864458e-01 -1.20467626e-01 8.82091343e-01
4.46008265e-01 -5.07190645e-01 -1.05066025e+00 1.06469667e+00
-1.64500535e+00 -7.51647651e-01 3.67356896e-01 2.01298189e+00
8.65122616e-01 -3.59106690e-01 -2.07834587e-01 -2.66271889e-01
3.64679396e-01 6.87699437e-01 -4.24499542e-01 -9.61355984e-01
1.03925481e-01 6.22349501e-01 2.58363396e-01 5.65182567e-01
-6.01647437e-01 1.10781181e+00 7.13343573e+00 7.65712082e-01
-6.86014891e-01 3.53665985e-02 -8.77557844e-02 -1.80374943e-02
-7.77396381e-01 -1.57056004e-02 -3.87147218e-01 -5.97328953e-02
9.38190639e-01 -8.05239737e-01 2.29031965e-01 8.35394323e-01
-2.35433400e-01 2.47051895e-01 -1.43216085e+00 8.90693724e-01
4.31929886e-01 -1.03329015e+00 2.84021139e-01 1.79319009e-01
4.90572870e-01 2.41698980e-01 1.76401541e-01 6.21528268e-01
3.25189084e-01 -1.41805458e+00 -3.56132269e-01 2.16753513e-01
7.04340875e-01 -3.79184306e-01 6.50499284e-01 1.24212177e-02
-1.02922428e+00 -3.95216942e-02 -5.86766779e-01 -7.32934773e-02
2.93706805e-01 2.23477513e-01 -1.02675962e+00 2.11206377e-01
9.13469046e-02 6.78116858e-01 -7.64895499e-01 9.21381414e-01
-4.60430309e-02 4.26005304e-01 -1.84657514e-01 -4.03992593e-01
5.46286941e-01 5.40719181e-02 3.41733694e-01 1.52184308e+00
1.81513831e-01 3.04606464e-02 -6.32621869e-02 4.46301043e-01
-2.00081021e-01 4.77230966e-01 -1.12855482e+00 -8.77376258e-01
7.24786758e-01 9.51931357e-01 -5.73689006e-02 -2.87752867e-01
-6.07146740e-01 7.38700390e-01 3.38169664e-01 6.37907565e-01
-2.79542834e-01 -7.92816460e-01 1.34912372e+00 -1.88126534e-01
3.57257992e-01 -4.94757682e-01 1.09783359e-01 -9.26913202e-01
-1.01094119e-01 -3.07331622e-01 5.93605340e-01 -7.81985939e-01
-1.74374247e+00 4.44280595e-01 2.11813763e-01 -9.58594143e-01
-7.01172054e-01 -5.72350144e-01 -4.39959973e-01 1.08419394e+00
-1.46834242e+00 -9.67335582e-01 1.95737451e-01 4.26870108e-01
4.87025201e-01 -1.18660726e-01 1.19670248e+00 1.43444821e-01
2.89518256e-02 7.88336694e-01 4.42705184e-01 1.16561167e-01
8.89674008e-01 -1.10002208e+00 2.63068587e-01 2.66835421e-01
7.04855382e-01 1.23688960e+00 5.36220968e-01 -4.10477877e-01
-1.59191799e+00 -7.14926660e-01 1.48283017e+00 -7.62747407e-01
9.47767138e-01 -6.52536452e-02 -1.03195894e+00 7.08025634e-01
6.31407976e-01 3.35355327e-02 1.04771805e+00 7.03149498e-01
-9.15012836e-01 1.35500193e-01 -8.98433030e-01 6.08745694e-01
8.49080861e-01 -1.17264688e+00 -1.17669666e+00 6.94771051e-01
1.41794288e+00 4.41558003e-01 -1.19740522e+00 1.47894353e-01
5.31185508e-01 -7.79489577e-02 1.42480719e+00 -1.13442075e+00
1.59330189e-01 2.76338071e-01 -5.30077577e-01 -1.49974227e+00
-5.51630437e-01 -2.21473277e-01 -1.87627614e-01 9.01720464e-01
3.95338118e-01 -7.26481676e-01 1.21582516e-01 8.94394398e-01
7.74713829e-02 -6.29481554e-01 -7.53663480e-01 -9.42542791e-01
6.80831313e-01 -3.63439977e-01 3.80312383e-01 1.02206504e+00
4.37606245e-01 6.72834933e-01 2.01892793e-01 -2.68537313e-01
2.41047740e-02 -3.45089287e-02 3.59318703e-01 -1.18706167e+00
-3.69258225e-02 -6.74323499e-01 -9.20217216e-01 -1.27536631e+00
2.64684141e-01 -1.26023233e+00 -3.25207487e-02 -1.74034905e+00
3.97949517e-02 -2.18280882e-01 -8.84261787e-01 1.44958928e-01
-4.76464093e-01 1.73601598e-01 -5.47235757e-02 2.08648100e-01
-6.34631932e-01 8.78470302e-01 8.53171468e-01 -2.27077931e-01
9.28425640e-02 -7.67060399e-01 -9.21227217e-01 6.37516528e-02
5.06271422e-01 -3.99196208e-01 -6.20642781e-01 -1.00180805e+00
3.88894677e-01 -3.02182019e-01 -1.07883096e-01 -3.32543224e-01
3.67152691e-01 -1.60392180e-01 -1.16650186e-01 -2.08998681e-03
5.73660493e-01 -9.18975890e-01 -6.28847361e-01 1.86477810e-01
-9.28674996e-01 6.01039708e-01 2.10834816e-01 9.61211979e-01
-6.50002658e-01 -3.32966954e-01 -3.41748074e-02 1.89999808e-02
-7.60229409e-01 3.70186239e-01 -4.44682330e-01 4.66784477e-01
5.74300885e-01 -6.51539043e-02 -2.85541385e-01 -5.49946308e-01
-4.50908005e-01 4.17323023e-01 3.43243837e-01 9.23090100e-01
9.12413478e-01 -1.60645306e+00 -6.45135224e-01 2.31203414e-03
8.93914104e-01 -4.58014369e-01 -3.22396129e-01 4.50954348e-01
-9.81433094e-02 1.00418794e+00 3.47678602e-01 -2.03496739e-01
-1.04801571e+00 9.26824570e-01 -2.01589957e-01 -1.97871462e-01
-3.41319948e-01 1.09815979e+00 5.07601351e-02 -2.10838169e-01
2.36954793e-01 -1.50484413e-01 -2.92657793e-01 2.16546103e-01
6.30442083e-01 9.36201140e-02 1.31988935e-02 -5.36236286e-01
-5.86095035e-01 4.56255019e-01 -5.10240197e-01 -4.37196285e-01
1.38159871e+00 -4.04054262e-02 -1.76840335e-01 3.58834475e-01
2.08454609e+00 -2.70699114e-01 7.00940415e-02 -7.58883059e-01
5.47318995e-01 -7.66350508e-01 3.81790787e-01 -5.18373728e-01
-4.89395827e-01 1.01195860e+00 5.40472746e-01 4.71711516e-01
8.25435519e-01 3.75494719e-01 1.00109172e+00 8.59206975e-01
2.01994613e-01 -1.00023174e+00 1.68420658e-01 6.80648327e-01
9.10504401e-01 -1.10087430e+00 1.94566727e-01 4.62782115e-01
-2.86568254e-01 1.01772499e+00 4.10520196e-01 -2.89903998e-01
1.00396371e+00 -3.86877656e-01 -2.14209259e-02 -5.09567440e-01
-9.89731610e-01 -5.29310942e-01 6.99702561e-01 6.80561900e-01
9.10259843e-01 -1.66969106e-01 -8.48658383e-01 3.64920646e-02
2.31614083e-01 -3.19264680e-01 -1.15565374e-01 1.19449127e+00
-6.33114815e-01 -1.32880163e+00 8.12771246e-02 6.28106833e-01
-1.86313584e-01 -6.60388231e-01 -5.96365631e-01 5.46753168e-01
-5.52366078e-01 9.59762573e-01 3.63073587e-01 -4.55498576e-01
1.18880346e-01 4.35944259e-01 2.14789882e-01 -1.09304047e+00
-6.96515620e-01 -1.76855102e-01 2.47112453e-01 -4.70504344e-01
-1.68097943e-01 -1.95449293e-01 -8.60689282e-01 1.97177380e-02
-6.69525206e-01 5.63934863e-01 7.06640899e-01 7.34952152e-01
6.05159163e-01 3.52610439e-01 4.82983977e-01 -3.65380257e-01
-7.51232624e-01 -1.48711848e+00 -6.24129772e-01 7.76916146e-01
1.83625504e-01 -5.02208531e-01 -6.85470045e-01 -5.43844223e-01] | [10.78756332397461, 8.414078712463379] |
d91a530a-14d4-4176-b331-bfdb0096814a | the-greedy-dirichlet-process-filter-an-online | 1811.05911 | null | http://arxiv.org/abs/1811.05911v2 | http://arxiv.org/pdf/1811.05911v2.pdf | The Greedy Dirichlet Process Filter - An Online Clustering Multi-Target Tracker | Reliable collision avoidance is one of the main requirements for autonomous
driving. Hence, it is important to correctly estimate the states of an unknown
number of static and dynamic objects in real-time. Here, data association is a
major challenge for every multi-target tracker. We propose a novel multi-target
tracker called Greedy Dirichlet Process Filter (GDPF) based on the
non-parametric Bayesian model called Dirichlet Processes and the fast posterior
computation algorithm Sequential Updating and Greedy Search (SUGS). By adding a
temporal dependence we get a real-time capable tracking framework without the
need of a previous clustering or data association step. Real-world tests show
that GDPF outperforms other multi-target tracker in terms of accuracy and
stability. | ['Hans-Joachim Wuensche', 'Benjamin Naujoks', 'Patrick Burger'] | 2018-11-14 | null | null | null | null | ['online-clustering'] | ['computer-vision'] | [-4.89752293e-01 -4.22011703e-01 -1.47358431e-02 -2.14529052e-01
-5.21200716e-01 -4.30557877e-01 8.06813180e-01 1.61077008e-01
-5.78636348e-01 6.85810685e-01 -4.71759140e-01 -1.31829649e-01
-3.47600907e-01 -6.93927944e-01 -5.34169257e-01 -9.02027190e-01
-1.05322890e-01 1.09976685e+00 1.15474272e+00 1.54230362e-02
1.27383526e-02 5.92788815e-01 -1.56036735e+00 -8.04166615e-01
7.50057042e-01 8.30255151e-01 5.34628093e-01 6.59747362e-01
1.79411098e-01 -2.78385669e-01 -4.65987831e-01 5.67933172e-02
2.04981878e-01 1.81862220e-01 3.58543068e-01 -1.87020555e-01
-1.70413293e-02 -2.00982377e-01 2.96595600e-02 1.02033257e+00
4.76636052e-01 4.09065485e-01 6.40365362e-01 -1.45010734e+00
2.94555247e-01 1.15727343e-01 -8.12190652e-01 3.62507582e-01
-2.70914048e-01 2.59269029e-01 2.61647075e-01 -6.95224345e-01
9.10188034e-02 1.42086208e+00 7.08489895e-01 3.56427222e-01
-9.84591067e-01 -9.99026775e-01 3.28123987e-01 4.35492635e-01
-1.41434491e+00 -3.05287957e-01 3.20053041e-01 -5.79417527e-01
3.57980490e-01 -2.57193092e-02 6.28312111e-01 8.43712211e-01
8.13965082e-01 2.99566329e-01 7.55425096e-01 8.75865519e-02
5.47615528e-01 -6.63722530e-02 2.41350740e-01 3.76560539e-01
1.04996908e+00 4.33117300e-01 -4.58213478e-01 -4.11330134e-01
4.79852408e-01 -1.35357666e-03 2.42287576e-01 -5.08111417e-01
-1.13130975e+00 7.94933915e-01 -1.07128724e-01 -2.41819909e-03
-4.60248321e-01 4.97665405e-01 1.57365769e-01 -1.87573418e-01
1.71106473e-01 -2.47270197e-01 -7.91661516e-02 -2.42513776e-01
-7.64210641e-01 4.05207038e-01 6.07850075e-01 8.71952295e-01
6.86898530e-01 2.25446582e-01 -4.64688279e-02 1.76537514e-01
8.61566663e-01 1.27996027e+00 2.43722111e-01 -4.73750204e-01
-9.98962224e-02 1.46066368e-01 7.22295821e-01 -8.94404948e-01
-4.59854603e-01 -5.53006589e-01 -8.33114207e-01 5.44693947e-01
1.75995320e-01 -2.80353487e-01 -6.56514049e-01 1.49805629e+00
1.00864244e+00 6.17549598e-01 1.52834347e-02 8.19575012e-01
3.24759811e-01 6.12258971e-01 3.14879715e-01 -4.25268859e-01
1.39549327e+00 -4.18669134e-01 -9.96092677e-01 -4.32179481e-01
1.02042429e-01 -6.71884358e-01 1.83499560e-01 3.88802826e-01
-4.04642105e-01 -7.51216471e-01 -1.19851601e+00 6.41881764e-01
-1.12642393e-01 1.38621256e-01 4.42670196e-01 6.75475121e-01
-6.77333117e-01 6.12936430e-02 -1.25766122e+00 -3.79277647e-01
-7.73415565e-02 5.24436355e-01 -6.97147921e-02 7.61090741e-02
-1.00128007e+00 1.12201536e+00 6.26813054e-01 7.89812282e-02
-1.20420635e+00 -4.32542205e-01 -5.42506695e-01 -2.64819950e-01
4.27176595e-01 -5.00962973e-01 1.08020151e+00 6.78249672e-02
-1.41083252e+00 9.36659351e-02 -4.20268923e-01 -7.35488355e-01
5.46757042e-01 -4.76071119e-01 -5.79739630e-01 -2.32806131e-01
1.62445381e-01 4.20158178e-01 1.06216657e+00 -1.20374620e+00
-1.07040930e+00 -2.42400393e-01 -6.46664858e-01 3.16006362e-01
1.14717618e-01 -1.17854707e-01 -3.95429105e-01 -1.90328911e-01
1.86384663e-01 -1.18757117e+00 -3.83777618e-01 -7.25325793e-02
-4.72961105e-02 -6.24831855e-01 1.43688023e+00 -3.01394820e-01
8.47006917e-01 -2.16002798e+00 -1.68578386e-01 2.63715297e-01
1.17843524e-01 2.29195029e-01 2.50619620e-01 1.45464376e-01
3.81737322e-01 -5.00408530e-01 1.28411397e-01 -3.75492543e-01
1.44533738e-01 3.56741816e-01 -3.31669509e-01 9.04344440e-01
-1.35270745e-01 3.89680624e-01 -8.38494420e-01 -5.50303817e-01
4.89687800e-01 3.41749668e-01 -1.19096003e-01 -1.31137684e-01
-4.13131088e-01 7.27534831e-01 -6.91034257e-01 1.82249472e-01
9.67105627e-01 1.23767719e-01 -3.62130180e-02 -4.93459553e-02
-6.58152640e-01 -3.01174939e-01 -1.74264705e+00 1.11672890e+00
3.16706151e-02 3.99454892e-01 2.29029641e-01 -6.01907849e-01
1.11926317e+00 2.44890675e-01 5.78338623e-01 -3.99895817e-01
2.32049882e-01 8.74108151e-02 -4.93985191e-02 3.16705555e-02
7.98191249e-01 3.59285693e-03 -3.61895919e-01 2.62440830e-01
-4.01777744e-01 -6.13627993e-02 2.07008407e-01 1.93580344e-01
1.03838015e+00 -1.42222062e-01 1.07332319e-01 -3.83775532e-01
3.50737363e-01 1.53763473e-01 1.18834877e+00 7.45407522e-01
-4.11713213e-01 -1.62666425e-01 -2.59287864e-01 -2.28005290e-01
-6.69091582e-01 -1.11143911e+00 -1.58594623e-01 7.45054901e-01
7.07274318e-01 -1.73484206e-01 -9.00646895e-02 -1.79156378e-01
3.53910893e-01 5.55537760e-01 -2.86365896e-01 -1.48576453e-01
-5.93847871e-01 -8.95159483e-01 2.12400153e-01 2.41718009e-01
3.77918124e-01 -2.66443253e-01 -1.07697749e+00 4.97658879e-01
-4.08217646e-02 -1.09242773e+00 -1.62955105e-01 6.31106198e-02
-6.20251298e-01 -9.57705975e-01 -1.67331353e-01 -1.89033329e-01
3.18207800e-01 5.20249486e-01 4.96193707e-01 -3.18622589e-01
-1.00851491e-01 2.94325560e-01 -4.37278338e-02 -8.11705410e-01
-2.73381650e-01 -2.08857596e-01 6.46765649e-01 1.36968791e-01
2.43814141e-01 -3.53723764e-01 -3.47711116e-01 7.10442483e-01
-4.03019935e-01 -4.10125367e-02 7.46890962e-01 2.93120354e-01
8.77530873e-01 4.56481516e-01 4.32571530e-01 1.02365261e-03
2.44614989e-01 -4.92973268e-01 -1.44474018e+00 -5.53110056e-02
-2.79902905e-01 -1.93058420e-02 1.00649022e-01 -6.60623372e-01
-1.09001517e+00 3.22060853e-01 2.53581882e-01 -6.40787601e-01
-6.87443987e-02 2.01844394e-01 -2.21541598e-01 -2.18469590e-01
2.85700291e-01 1.55823335e-01 -5.65170348e-02 -4.54553217e-01
1.86493337e-01 3.09115559e-01 6.19635105e-01 -4.04502809e-01
1.17480183e+00 8.17527413e-01 5.64012408e-01 -9.67229605e-01
-3.11857551e-01 -7.52313554e-01 -5.15198827e-01 -6.20581925e-01
9.77713943e-01 -1.08073545e+00 -1.04660976e+00 4.19064701e-01
-1.11027741e+00 -1.03507906e-01 -3.05309091e-02 9.41259146e-01
-3.84892195e-01 4.54508722e-01 -2.60856915e-02 -1.08228922e+00
-1.14473633e-01 -9.27502871e-01 8.25126410e-01 4.79934633e-01
1.71839461e-01 -7.37041593e-01 4.35893446e-01 -7.05493242e-02
4.48191792e-01 2.95485884e-01 7.48070702e-02 -4.58677828e-01
-8.65268409e-01 -2.94591337e-01 -7.18624191e-03 -3.75871330e-01
-1.16488650e-01 1.73704341e-01 -4.05611187e-01 -6.18254364e-01
1.22493111e-01 3.08720827e-01 7.30572224e-01 5.72655976e-01
1.89904526e-01 1.82809860e-01 -1.04727292e+00 6.75934628e-02
1.31927550e+00 2.94772297e-01 8.17893595e-02 1.68240517e-01
4.47610676e-01 1.90068454e-01 1.41750336e+00 7.09386885e-01
6.56129062e-01 8.15473378e-01 6.54445052e-01 3.56114000e-01
8.75149071e-02 1.50875254e-02 4.92121309e-01 6.18029416e-01
3.06985050e-01 -1.39468238e-01 -1.09073114e+00 7.17060685e-01
-2.21939182e+00 -9.17303026e-01 -8.02636921e-01 2.42818356e+00
4.25120592e-01 4.53423709e-01 8.56680274e-02 -1.50115132e-01
1.03778481e+00 -3.53027344e-01 -6.89137816e-01 3.21303964e-01
2.20494330e-01 -2.90286571e-01 1.05028903e+00 5.60883403e-01
-1.29358077e+00 7.11632550e-01 5.98825788e+00 9.89928067e-01
-8.57344806e-01 3.93555075e-01 -2.91685998e-01 -2.35439334e-02
2.96441346e-01 1.46127254e-01 -1.82660878e+00 6.71613574e-01
1.25562096e+00 -4.22477573e-01 -2.93684870e-01 7.78276443e-01
5.15694141e-01 -8.14356387e-01 -6.22064173e-01 8.69504750e-01
-2.62943178e-01 -7.50877738e-01 -4.56184745e-01 1.65579483e-01
4.61119026e-01 2.76977509e-01 -1.77367657e-01 2.18415365e-01
8.95309150e-01 -3.12506318e-01 8.84309530e-01 7.59133577e-01
1.82861045e-01 -7.27451384e-01 6.28094792e-01 6.95695460e-01
-1.64997447e+00 -1.73799917e-02 -4.51570392e-01 1.37702242e-01
7.13793933e-01 9.63242114e-01 -6.87120497e-01 6.22285843e-01
7.33303130e-01 4.72804308e-01 -3.39018553e-01 1.64811754e+00
-3.27461585e-02 5.35522223e-01 -1.06933677e+00 -3.34132724e-02
1.49558067e-01 -4.05415818e-02 1.04989755e+00 7.33503938e-01
4.69877690e-01 -9.00594145e-03 7.42118061e-01 3.69183302e-01
7.67082274e-01 -4.67902869e-01 -2.99305052e-01 3.85569185e-01
8.22944582e-01 1.27027881e+00 -9.24902558e-01 -4.15422052e-01
-1.22046217e-01 1.26632854e-01 -3.24776053e-01 -8.82092305e-03
-1.18041909e+00 -3.18013169e-02 7.97547460e-01 -8.81386176e-02
5.08491814e-01 -8.25296879e-01 -1.33342400e-01 -6.92503214e-01
-2.29784995e-01 -1.31645769e-01 2.82137603e-01 -3.63135487e-01
-1.03644598e+00 1.21706657e-01 3.99772346e-01 -1.52743745e+00
-1.69898599e-01 -2.44479924e-02 -4.32397157e-01 7.66444623e-01
-1.53428674e+00 -9.99615371e-01 -4.41049337e-01 5.69313109e-01
4.28480744e-01 -5.38014807e-02 3.71151298e-01 3.41917723e-01
-5.57089746e-01 -5.13737798e-02 1.59849554e-01 -5.29567719e-01
8.18919778e-01 -8.19112837e-01 3.53945941e-01 1.01278377e+00
-3.09279829e-01 1.02428347e-01 1.29275584e+00 -1.30337095e+00
-1.55196893e+00 -1.16080344e+00 4.48872954e-01 -2.94362038e-01
6.27910197e-01 -2.21206561e-01 -7.84001052e-01 5.78153491e-01
-1.81666873e-02 -1.73528835e-01 2.33256638e-01 -1.10518672e-01
3.12374681e-01 -2.83885568e-01 -8.94094348e-01 1.77080959e-01
3.99023652e-01 2.65360653e-01 -5.24672091e-01 2.63151169e-01
4.22857970e-01 -3.99799526e-01 -5.63147724e-01 5.95442593e-01
3.86992693e-01 -5.49429238e-01 7.43921459e-01 2.52542913e-01
-1.03070247e+00 -1.04865265e+00 5.25173321e-02 -1.04739428e+00
-5.20883858e-01 -5.36311209e-01 -2.11616248e-01 1.00965452e+00
-1.95762306e-01 -8.41351092e-01 6.85018897e-01 3.73976022e-01
-1.84549704e-01 1.30716726e-01 -1.35570168e+00 -1.21974063e+00
-5.87440431e-01 -5.07800817e-01 2.80176431e-01 5.58935523e-01
-4.62192714e-01 2.19474867e-01 -4.40329909e-01 9.52716291e-01
1.41807759e+00 -5.44933118e-02 1.22090983e+00 -1.84661651e+00
-1.31808847e-01 -8.58204346e-03 -5.16839445e-01 -1.01366365e+00
-2.65927743e-02 -8.57504085e-02 6.03290975e-01 -1.50178647e+00
9.14311931e-02 -7.69223571e-01 4.85315174e-02 1.55401066e-01
-3.15769464e-02 -7.75541514e-02 1.45896701e-02 4.90152657e-01
-1.02763212e+00 9.32708085e-01 7.00345695e-01 7.92578682e-02
-1.93511117e-02 3.79009128e-01 3.54929678e-02 6.13788128e-01
7.62388110e-01 -9.70619857e-01 -1.90315589e-01 -3.36612135e-01
1.19516831e-02 7.95223638e-02 3.48934084e-01 -1.59695482e+00
9.09474671e-01 -3.82383257e-01 1.18022434e-01 -1.53182578e+00
7.78062940e-01 -9.90581870e-01 6.32352114e-01 8.74904871e-01
5.81417561e-01 1.81984782e-01 5.33549309e-01 1.26246214e+00
-2.37505436e-02 1.88788325e-02 9.20735121e-01 3.57465416e-01
-1.04255331e+00 3.70809108e-01 -7.22197354e-01 -4.66329634e-01
1.70227683e+00 -2.24733576e-01 -1.79616690e-01 -1.89157709e-01
-6.63105726e-01 7.72965550e-01 3.01932812e-01 5.48178971e-01
4.55917239e-01 -1.20515621e+00 -5.52934945e-01 1.30931720e-01
-1.13762192e-01 4.76718135e-03 3.98148179e-01 9.80805397e-01
-5.87929636e-02 5.48334301e-01 -1.25534430e-01 -1.05643511e+00
-1.34469712e+00 3.14336091e-01 6.64055496e-02 -1.46667451e-01
-4.31399435e-01 3.73994231e-01 8.72344449e-02 9.13312435e-02
1.44298315e-01 -1.25735357e-01 -1.32433221e-01 -5.24627492e-02
5.70912600e-01 4.01061594e-01 -1.75445721e-01 -6.73571825e-01
-4.53562140e-01 6.81577086e-01 2.13374626e-02 -3.11942101e-01
8.92830551e-01 -3.95819455e-01 6.39429912e-02 5.55324793e-01
4.99154441e-02 -1.70052126e-01 -1.69734859e+00 -1.56377062e-01
2.73112893e-01 -4.51823175e-01 3.10682565e-01 -2.44698837e-01
-5.43244720e-01 4.77231055e-01 1.01513803e+00 -8.15161094e-02
5.12009203e-01 -4.88360375e-02 5.54387093e-01 3.68428499e-01
6.96874797e-01 -1.00101805e+00 -1.45053668e-02 7.23787844e-01
3.80213410e-01 -1.08041298e+00 1.00143157e-01 -3.62104923e-01
-2.90456384e-01 6.54213011e-01 6.93157852e-01 -1.80011630e-01
9.78414893e-01 4.18064624e-01 -1.94356903e-01 -5.75216115e-02
-9.43030834e-01 -3.87957573e-01 2.42377669e-02 6.28146589e-01
-5.73171377e-01 1.56548262e-01 -3.61225039e-01 4.27646846e-01
1.05928913e-01 -1.77210718e-01 2.64317691e-01 1.11926544e+00
-1.05702937e+00 -9.09933984e-01 -8.84590745e-01 1.82962641e-01
1.48381349e-02 5.30610204e-01 3.51131350e-01 5.97552299e-01
1.10805877e-01 1.05033350e+00 2.64714569e-01 -2.10609049e-01
1.97452068e-01 -1.92218438e-01 2.24533871e-01 -4.05129313e-01
-1.69343069e-01 3.15939099e-01 -1.25419796e-01 -4.73885596e-01
-3.71385753e-01 -9.38580513e-01 -1.31945956e+00 -3.67758960e-01
-6.04519546e-01 4.06198442e-01 1.15138388e+00 9.95155811e-01
5.50232351e-01 3.75772029e-01 2.91210473e-01 -9.37814832e-01
-5.03958106e-01 -9.73470151e-01 -5.19765079e-01 -1.84742630e-01
1.73467547e-01 -1.32966208e+00 -1.47279456e-01 -1.90193996e-01] | [6.558304786682129, -2.0658016204833984] |
fdbf352f-34b1-4aab-96b4-240e7512db31 | a-denoised-mean-teacher-for-domain-adaptive | 2306.14749 | null | https://arxiv.org/abs/2306.14749v2 | https://arxiv.org/pdf/2306.14749v2.pdf | A denoised Mean Teacher for domain adaptive point cloud registration | Point cloud-based medical registration promises increased computational efficiency, robustness to intensity shifts, and anonymity preservation but is limited by the inefficacy of unsupervised learning with similarity metrics. Supervised training on synthetic deformations is an alternative but, in turn, suffers from the domain gap to the real domain. In this work, we aim to tackle this gap through domain adaptation. Self-training with the Mean Teacher is an established approach to this problem but is impaired by the inherent noise of the pseudo labels from the teacher. As a remedy, we present a denoised teacher-student paradigm for point cloud registration, comprising two complementary denoising strategies. First, we propose to filter pseudo labels based on the Chamfer distances of teacher and student registrations, thus preventing detrimental supervision by the teacher. Second, we make the teacher dynamically synthesize novel training pairs with noise-free labels by warping its moving inputs with the predicted deformations. Evaluation is performed for inhale-to-exhale registration of lung vessel trees on the public PVT dataset under two domain shifts. Our method surpasses the baseline Mean Teacher by 13.5/62.8%, consistently outperforms diverse competitors, and sets a new state-of-the-art accuracy (TRE=2.31mm). Code is available at https://github.com/multimodallearning/denoised_mt_pcd_reg. | ['Mattias P. Heinrich', 'Alexander Bigalke'] | 2023-06-26 | null | null | null | null | ['point-cloud-registration'] | ['computer-vision'] | [ 2.64194101e-01 4.03300464e-01 -4.96370718e-02 -3.28884363e-01
-1.54604626e+00 -6.56503439e-01 5.42012691e-01 2.29048342e-01
-4.74779934e-01 6.42962515e-01 6.33359775e-02 8.63225982e-02
-3.00853521e-01 -6.12141609e-01 -6.45948470e-01 -1.27454615e+00
2.87038833e-01 9.28027332e-01 3.32729399e-01 -1.72004372e-01
-1.90835103e-01 4.48357671e-01 -1.13713491e+00 3.34904432e-01
1.14348578e+00 7.85234749e-01 6.95929956e-03 3.27935308e-01
6.63294271e-02 3.44975948e-01 -6.22881293e-01 -6.61209762e-01
3.88523340e-01 -2.89923310e-01 -7.56790400e-01 8.52324516e-02
7.02070892e-01 -8.07885006e-02 -1.62388876e-01 9.57045853e-01
9.70604837e-01 2.29542945e-02 7.02958286e-01 -9.33709502e-01
-3.69916707e-01 2.69498140e-01 -5.80858290e-01 -4.32290137e-02
7.40105063e-02 1.63036004e-01 4.29085284e-01 -8.31111670e-01
8.00805449e-01 8.07724893e-01 9.32722151e-01 9.33355927e-01
-1.61861730e+00 -6.85448766e-01 -3.84890050e-01 -1.13038421e-01
-1.42124856e+00 -3.76499891e-01 6.37487710e-01 -5.66236973e-01
4.16867375e-01 4.08737034e-01 3.15629452e-01 1.27149498e+00
9.40648764e-02 2.48457387e-01 1.40366697e+00 -1.63605824e-01
1.66307554e-01 1.71263978e-01 -3.94267231e-01 3.83673400e-01
-8.43360499e-02 2.66725779e-01 -4.84417170e-01 -3.64566535e-01
4.34737682e-01 -1.93056136e-01 -4.77144629e-01 -7.27236509e-01
-1.29062307e+00 6.39478385e-01 5.65606356e-01 4.16084647e-01
-2.80257702e-01 -9.18463990e-02 2.82902151e-01 3.43179196e-01
8.16331685e-01 2.43572176e-01 -3.83488744e-01 6.35819063e-02
-1.07938755e+00 1.46452412e-01 5.80881894e-01 7.24325657e-01
5.11728644e-01 -2.34386936e-01 -1.38197169e-01 8.40756536e-01
3.15838695e-01 4.77689266e-01 5.84506691e-01 -9.20774817e-01
2.53087997e-01 3.03139359e-01 -3.67899001e-01 -6.33291721e-01
-4.31103557e-01 -6.90576732e-01 -1.10817468e+00 3.47290814e-01
7.43823588e-01 3.62292826e-02 -1.03368104e+00 1.62635672e+00
9.61835563e-01 4.77108330e-01 -1.45220403e-02 8.27066004e-01
1.06656969e+00 6.48271143e-02 2.02717945e-01 -3.17608714e-01
1.33999598e+00 -6.13807738e-01 -5.82443118e-01 1.24413766e-01
7.70391822e-01 -9.17953491e-01 7.03743398e-01 5.12937486e-01
-1.04788756e+00 -4.67951208e-01 -5.09355545e-01 1.59503102e-01
-3.05682868e-02 -7.23522902e-02 2.81258021e-02 7.93292463e-01
-1.15943778e+00 8.42532814e-01 -1.10500073e+00 -2.91325539e-01
7.53487766e-01 6.99173391e-01 -5.83557904e-01 -3.40572782e-02
-9.15957868e-01 8.87225330e-01 1.06126972e-01 -1.73213556e-01
-6.95243061e-01 -1.17918491e+00 -4.72164035e-01 -4.88577485e-01
1.49898678e-01 -8.49829555e-01 1.06315541e+00 -8.72259974e-01
-1.59663320e+00 1.24776614e+00 1.35038048e-01 -4.16933209e-01
1.05638862e+00 1.26579016e-01 -2.50116020e-01 2.57685840e-01
1.55646771e-01 7.80087650e-01 9.91376519e-01 -1.38619494e+00
-2.39321902e-01 -5.75221300e-01 -6.96031928e-01 1.59515277e-01
-2.26333484e-01 -1.70164704e-01 -1.25310600e-01 -8.73244584e-01
3.52977484e-01 -1.16889715e+00 -2.65537500e-01 1.44863293e-01
-1.02946378e-01 -1.83468446e-01 6.16056263e-01 -7.18635261e-01
6.53496265e-01 -2.31174278e+00 -2.93684900e-02 4.15181875e-01
4.10045147e-01 2.41247237e-01 -7.86284730e-02 5.84587380e-02
-4.41740245e-01 -1.56384662e-01 -6.58289313e-01 -5.64492106e-01
-3.46230358e-01 3.25858146e-01 -1.32785931e-01 8.51300895e-01
7.06418091e-03 7.63437450e-01 -9.98806953e-01 -7.12652326e-01
1.10654809e-01 7.35818982e-01 -4.71472502e-01 1.73645973e-01
7.04823956e-02 1.34642160e+00 -2.65792519e-01 4.98305827e-01
1.04236937e+00 -7.38002639e-03 1.42885044e-01 -1.71171591e-01
1.88960463e-01 1.38847873e-01 -9.93709445e-01 1.99592412e+00
-2.29975760e-01 8.94270465e-02 1.83949694e-01 -1.01350796e+00
1.02151501e+00 7.55677164e-01 1.06925023e+00 -5.19663215e-01
1.08756267e-01 5.79402506e-01 -7.00612739e-02 -4.54243124e-01
-1.20781817e-01 -5.50009787e-01 1.59432575e-01 5.98511510e-02
2.28442624e-01 -4.83269930e-01 -3.28676581e-01 -7.10804090e-02
1.15999818e+00 2.19976380e-01 -3.18648806e-03 -2.25465313e-01
5.43040156e-01 5.24219573e-02 5.72686851e-01 3.40680271e-01
-3.52589309e-01 1.09648275e+00 2.18329266e-01 -2.91031659e-01
-7.15216339e-01 -1.36246514e+00 -5.75736165e-01 7.59308577e-01
-7.55667388e-02 -1.92132488e-01 -9.13664699e-01 -1.02497947e+00
7.52171036e-03 4.98967856e-01 -5.40837348e-01 -1.77619457e-01
-7.66098499e-01 -7.14084208e-01 7.07545936e-01 2.26694435e-01
1.81032076e-01 -6.73463881e-01 -2.78714746e-01 1.69254020e-01
-4.25848663e-01 -1.03096473e+00 -3.93151373e-01 8.37411433e-02
-1.32831514e+00 -1.01022220e+00 -9.90161359e-01 -7.14539349e-01
8.44418526e-01 -1.48243025e-01 1.16266346e+00 -9.22891218e-03
-2.21132636e-02 4.01235849e-01 -1.75710276e-01 -3.18166405e-01
-7.92458892e-01 2.64031261e-01 3.69716845e-02 7.63559639e-02
1.28022850e-01 -9.27962840e-01 -7.64001250e-01 5.66942155e-01
-9.95447993e-01 -2.72886783e-01 5.05632758e-01 1.03532696e+00
1.01493394e+00 -2.60340601e-01 4.66005802e-01 -9.73099291e-01
2.65366286e-01 -3.24763447e-01 -4.67144787e-01 9.14029777e-02
-7.52971172e-01 -1.09003656e-01 3.53365660e-01 -5.43364108e-01
-1.09916759e+00 4.46955442e-01 -3.74623656e-01 -6.33027375e-01
-5.08137882e-01 -9.55218915e-04 -8.15983489e-02 -4.26189721e-01
1.19594562e+00 1.29189432e-01 3.28042388e-01 -5.71073353e-01
3.74180228e-01 5.45055628e-01 7.81274319e-01 -7.14316010e-01
1.11032057e+00 7.54951000e-01 1.43987164e-01 -5.97194970e-01
-5.91758311e-01 -4.96470541e-01 -9.17191505e-01 -1.96453229e-01
8.61895561e-01 -9.57342505e-01 -3.69381279e-01 5.00476956e-01
-1.08053553e+00 -2.28250727e-01 -7.39716172e-01 5.30869305e-01
-5.83653390e-01 5.24733007e-01 -3.33379239e-01 -1.74511224e-01
-6.01560652e-01 -1.11744797e+00 1.06359553e+00 -7.69124031e-02
-2.18162015e-01 -9.48529482e-01 3.55370402e-01 6.55121088e-01
4.02815402e-01 6.27402961e-01 3.86444867e-01 -9.89282906e-01
-3.55709523e-01 -5.15526645e-02 1.66851461e-01 5.22448838e-01
1.93497181e-01 -3.72278512e-01 -1.15087700e+00 -4.36248183e-01
5.55532873e-01 -1.39963731e-01 8.10563684e-01 3.21672648e-01
1.08414829e+00 -1.00960404e-01 -2.99644142e-01 7.62616158e-01
1.07728267e+00 -2.93834686e-01 6.32319987e-01 2.24824980e-01
7.59103656e-01 8.83939028e-01 5.29258847e-01 1.82923421e-01
2.54241616e-01 6.88236535e-01 5.58005035e-01 -1.42152652e-01
-4.07490075e-01 2.98097469e-02 1.16978534e-01 8.37546527e-01
-2.10478321e-01 1.97351933e-01 -1.16830528e+00 5.86046040e-01
-1.47813272e+00 -6.21754527e-01 -4.44072723e-01 2.40441608e+00
1.10271931e+00 2.61392444e-02 9.96916220e-02 -9.51254740e-03
5.83305180e-01 -7.42357746e-02 -2.11499572e-01 1.06435277e-01
1.11742236e-01 5.77663422e-01 5.85586369e-01 3.52498531e-01
-1.22832990e+00 6.40592694e-01 4.48223448e+00 1.09805119e+00
-1.16502857e+00 6.60449982e-01 4.96721864e-01 -5.46204299e-02
-1.73456132e-01 -1.57526538e-01 -3.33156466e-01 4.53042895e-01
9.16343689e-01 2.61552215e-01 1.06160820e-01 6.09976292e-01
-2.31349631e-03 1.12923026e-01 -1.09702671e+00 8.45701814e-01
7.05986191e-03 -1.06137037e+00 -3.13228101e-01 2.00069591e-01
8.00693214e-01 2.63727665e-01 1.48738042e-01 -8.26955140e-02
-2.81656650e-03 -9.96754885e-01 3.04126471e-01 4.61960375e-01
1.07064462e+00 -5.47991633e-01 7.11786747e-01 3.93745095e-01
-8.38274956e-01 4.16771710e-01 -1.11643516e-01 5.84317327e-01
-5.73188700e-02 7.80411184e-01 -1.09240317e+00 8.17835152e-01
6.94578111e-01 6.05978906e-01 -4.67754513e-01 1.15461278e+00
-2.81635135e-01 6.06320858e-01 -4.28831100e-01 6.88191295e-01
-2.33037055e-01 -1.93326727e-01 8.20926428e-01 1.00300312e+00
3.49974990e-01 8.80327076e-02 -2.81574521e-02 6.04580820e-01
-1.63443759e-01 1.10598765e-01 -5.05554855e-01 5.99557877e-01
3.47067237e-01 1.36677599e+00 -6.10628843e-01 -8.03244859e-02
-9.26410705e-02 8.05211723e-01 -4.53271456e-02 4.12315503e-02
-7.81971395e-01 3.58203426e-02 3.48259687e-01 4.95657563e-01
1.09052442e-01 1.54597372e-01 -2.89395481e-01 -9.11699772e-01
5.65907620e-02 -1.02717912e+00 6.02689266e-01 -3.64234775e-01
-1.52739930e+00 6.82327926e-01 3.01787034e-02 -1.69523036e+00
-1.19147502e-01 -2.48599350e-01 -4.90140319e-01 8.41754496e-01
-1.51663661e+00 -1.27403605e+00 -4.57078338e-01 6.96680665e-01
1.75402254e-01 6.91488981e-02 9.44263577e-01 6.14330173e-01
5.00303693e-02 8.38225901e-01 3.60998422e-01 1.09508283e-01
1.32568932e+00 -1.45568478e+00 2.63912737e-01 3.05385441e-01
1.29103065e-01 2.06835747e-01 6.30238533e-01 -5.73647082e-01
-8.93572211e-01 -1.15718222e+00 8.69774818e-01 -8.33557189e-01
4.60680336e-01 -1.65851370e-01 -1.21796679e+00 4.37644154e-01
9.55841616e-02 4.23447818e-01 7.27900386e-01 -3.25166047e-01
-2.85836756e-01 -4.04016167e-01 -1.62565923e+00 2.49055177e-01
8.18261743e-01 -2.93753862e-01 -6.51088059e-01 6.57522082e-01
4.07768726e-01 -1.00801301e+00 -1.48362267e+00 6.02441907e-01
2.98379153e-01 -8.46869826e-01 1.19344985e+00 -9.58469063e-02
3.64854112e-02 -2.84108073e-01 2.37817794e-01 -1.35109270e+00
1.37295295e-03 -6.92633629e-01 2.13969424e-01 1.31468141e+00
1.58210531e-01 -8.85797858e-01 1.08148301e+00 4.10891593e-01
-2.03026012e-01 -8.35386992e-01 -1.57485998e+00 -7.51212656e-01
6.15235448e-01 -2.40780696e-01 3.91921133e-01 1.24558079e+00
-3.77243519e-01 6.09476492e-02 5.26698604e-02 3.17337960e-01
9.66047883e-01 -2.66822815e-01 7.60069370e-01 -1.39515221e+00
-2.60515481e-01 -2.46854097e-01 -2.68476963e-01 -6.45649612e-01
1.08899474e-01 -1.27575624e+00 -7.33815180e-03 -1.21169639e+00
-1.29951671e-01 -5.51657677e-01 -1.16594657e-01 4.76013660e-01
1.60310119e-01 4.93647665e-01 1.66065004e-02 5.36922157e-01
-2.58530855e-01 4.96544540e-01 1.47533214e+00 -1.75792977e-01
-1.79510683e-01 3.61167043e-01 -3.28089088e-01 6.70879662e-01
9.36645389e-01 -1.06218195e+00 -1.69350624e-01 -2.58872837e-01
-1.90647095e-01 6.46980256e-02 6.01479232e-01 -1.04900050e+00
3.82436901e-01 1.91210359e-01 2.71493763e-01 -4.91403252e-01
2.12199241e-01 -1.08406067e+00 4.50491160e-01 5.98343909e-01
-1.03018083e-01 -1.80707932e-01 8.55022818e-02 5.40432811e-01
-2.05906019e-01 -1.24930546e-01 1.17279470e+00 -3.05166692e-02
7.13160485e-02 2.82316923e-01 3.39000411e-02 3.35283577e-01
8.89330745e-01 -2.64810771e-01 -1.72601312e-01 -1.37089845e-02
-9.87360716e-01 1.63038522e-01 4.32444870e-01 1.05654478e-01
4.84535873e-01 -1.14749289e+00 -1.05998015e+00 1.90562516e-01
5.69789186e-02 6.98319435e-01 4.26837325e-01 1.54491806e+00
-3.64214212e-01 -1.36171579e-01 1.10560812e-01 -1.12519658e+00
-1.46936131e+00 2.91212201e-01 7.29867220e-01 -4.74585950e-01
-5.72244942e-01 9.86702740e-01 2.36105427e-01 -8.86674583e-01
1.61832094e-01 -2.17610851e-01 7.20890611e-02 1.18071690e-01
-5.73735721e-02 3.88175577e-01 6.52357459e-01 -8.12674820e-01
-4.84404325e-01 8.60143661e-01 -6.48117065e-02 1.03355378e-01
1.33127201e+00 1.70987666e-01 -1.77173480e-01 -3.81581783e-02
1.13280702e+00 1.96471602e-01 -1.13271070e+00 -3.99703920e-01
2.20681429e-02 -4.45606917e-01 -1.06231064e-01 -9.00497019e-01
-1.24286473e+00 8.05418372e-01 1.11716700e+00 -1.35373443e-01
1.27743506e+00 1.93660766e-01 9.14874732e-01 -1.11106552e-01
-5.85736532e-04 -8.71413291e-01 4.83202711e-02 1.42459154e-01
7.46555388e-01 -1.32418907e+00 1.39217392e-01 -5.31632125e-01
-5.83197653e-01 8.07549417e-01 3.48856121e-01 -9.18520540e-02
7.04295218e-01 1.80995673e-01 4.94418949e-01 -2.86579251e-01
-3.42893541e-01 7.83963054e-02 5.78200102e-01 8.13233852e-01
3.58614683e-01 2.15945318e-02 -3.61660093e-01 3.16908568e-01
-2.95109481e-01 -9.94773135e-02 2.00436078e-02 7.90279508e-01
-3.25788893e-02 -1.52281833e+00 -7.17690945e-01 2.55300790e-01
-7.07380593e-01 9.97624621e-02 -2.89502949e-01 7.57459700e-01
4.80838567e-01 6.62955821e-01 -1.18912421e-01 -1.60884649e-01
6.82994545e-01 3.31738181e-02 4.71936733e-01 -5.61868846e-01
-1.19181561e+00 4.50866580e-01 -2.89943665e-01 -4.62940156e-01
-5.19655526e-01 -8.77944648e-01 -1.19954550e+00 -7.10017458e-02
-2.51478195e-01 1.35063887e-01 6.15494490e-01 6.37287378e-01
3.46916467e-01 4.17281657e-01 7.49751091e-01 -7.41986811e-01
-1.00866210e+00 -7.22270370e-01 -2.99220264e-01 7.42370486e-01
3.58643502e-01 -5.90545595e-01 -5.56578815e-01 1.07527599e-01] | [14.59825611114502, -2.139113426208496] |
35d5bf95-e386-4130-a5c6-6479aa101e76 | a-hybrid-approach-to-vietnamese-word | null | null | https://ieeexplore.ieee.org/document/7800279 | https://ieeexplore.ieee.org/document/7800279 | A hybrid approach to Vietnamese word segmentation | Word segmentation is the very first task for Vietnamese language processing. Word-segmented text is the input of almost other NLP tasks. This task faces some challenges due to specific characteristics of the language. As in many other Asian languages such as Japanese, Korean and Chinese, white spaces in Vietnamese are not always used as word separators and a word may contain one or more syllables. In this paper, we propose an efficient hybrid approach to detect word boundary for Vietnamese texts using logistic regression as a binary classifier combining with longest matching algorithm. First, longest matching algorithm is used to catch words that contain more than two syllables in input sentence. Next, the system utilizes the classifier to determine the boundary of 2-syllable words and proper names. Then, the predictions having low confidence conducted by the classifier are verified by a dictionary to get the final result. Our system can achieve an F-measure of 98.82% which is the most accurate result for Vietnamese word segmentation to the best of our knowledge. Moreover, the system also has a high speed. It can run word segmentation for nearly 34k tokens per second. | ['Anh-Cuong Le', 'Tuan-Phong Nguyen'] | 2016-12-29 | null | null | null | null | ['vietnamese-word-segmentation'] | ['natural-language-processing'] | [-3.82222161e-02 -3.12272102e-01 -3.29664439e-01 -2.15261579e-01
-5.76835573e-01 -5.52318215e-01 -1.28515795e-01 3.51969212e-01
-9.94725168e-01 7.36179948e-01 -1.87282190e-01 -7.21932471e-01
5.78372359e-01 -8.44688594e-01 -1.01812534e-01 -6.63027287e-01
4.25541490e-01 6.08551860e-01 5.87628543e-01 -2.08811909e-01
4.53750044e-01 4.86548424e-01 -9.56299424e-01 4.14293958e-03
1.20349038e+00 4.19298500e-01 4.56146270e-01 4.62112874e-01
-5.43548763e-01 1.03233410e-02 -6.77087307e-01 -3.44126850e-01
1.16810396e-01 -3.98099184e-01 -6.74701095e-01 3.73735316e-02
-2.87861049e-01 -2.88978159e-01 1.71471253e-01 1.19900846e+00
2.87172943e-01 5.95969893e-02 7.26937771e-01 -6.42601609e-01
-1.93298712e-01 1.04829490e+00 -7.86432207e-01 3.68563861e-01
1.91956043e-01 -1.46307021e-01 1.03144717e+00 -1.03864360e+00
4.44523633e-01 9.56306398e-01 1.04246289e-01 4.49340731e-01
-7.13689387e-01 -7.43118703e-01 1.75364137e-01 6.21310957e-02
-1.71790552e+00 -6.08317032e-02 4.53141779e-01 -3.86141539e-01
1.06551707e+00 1.53280988e-01 5.18920243e-01 4.46074098e-01
2.16479182e-01 7.53257573e-01 9.06329691e-01 -6.76771343e-01
2.53329556e-02 2.53909796e-01 4.90178466e-01 6.23239636e-01
3.26790690e-01 -4.32080865e-01 -4.88384953e-03 2.26618752e-01
3.32639843e-01 -1.98065981e-01 -9.09368098e-02 8.23962748e-01
-1.18777657e+00 9.28302586e-01 -7.04626888e-02 6.52742207e-01
-2.01841742e-01 -5.07202804e-01 4.47111875e-01 -2.20282868e-01
4.76398058e-02 1.23694025e-01 -4.19851154e-01 -1.40795425e-01
-1.03212035e+00 4.47747596e-02 9.49149013e-01 1.00209224e+00
7.04108298e-01 1.61019087e-01 1.78645879e-01 8.28523099e-01
2.41506934e-01 5.85931957e-01 7.29698122e-01 -2.28545770e-01
6.21910930e-01 5.06197929e-01 -7.54324049e-02 -7.44627953e-01
-3.04208547e-01 -1.74629509e-01 -6.28803313e-01 -1.94791183e-01
6.81240916e-01 -5.16767442e-01 -1.06365919e+00 1.13595450e+00
4.05951113e-01 -2.81769961e-01 2.54247487e-01 8.88577461e-01
7.22060323e-01 1.20149136e+00 3.32913041e-01 -4.73398000e-01
1.72719586e+00 -1.00683188e+00 -8.95011544e-01 -2.64290690e-01
5.33132136e-01 -1.17124736e+00 9.12407339e-01 4.53208715e-01
-7.00416982e-01 -4.71077770e-01 -1.01273930e+00 -4.07286584e-02
-5.01093090e-01 2.83350021e-01 3.27668786e-01 8.62265408e-01
-3.06761682e-01 7.97130391e-02 -7.38983512e-01 -2.81456500e-01
-1.23988062e-01 4.39484507e-01 -3.07100732e-02 -3.50750536e-02
-1.26537108e+00 6.79705620e-01 9.65253830e-01 2.58137614e-01
-1.48648992e-01 -4.43942808e-02 -9.19371367e-01 4.99044806e-02
5.42880058e-01 1.49997473e-01 1.14068878e+00 -8.54675889e-01
-1.24309051e+00 7.06918955e-01 -5.24503648e-01 -9.04146358e-02
3.39246720e-01 2.00928077e-02 -6.44999623e-01 6.59731328e-02
2.49111816e-01 4.97278929e-01 3.87133181e-01 -7.96954453e-01
-1.12601709e+00 -2.51585960e-01 -4.81592834e-01 2.41324872e-01
-1.43885180e-01 4.11650389e-01 -7.10629463e-01 -6.23838842e-01
4.00181621e-01 -7.20262170e-01 -2.91274607e-01 -8.42448711e-01
-4.16483283e-01 -5.98601818e-01 6.52893782e-01 -1.16299295e+00
1.77479649e+00 -2.12646556e+00 -3.98608983e-01 4.87497389e-01
-1.47949383e-01 4.29675221e-01 1.99078515e-01 2.76714623e-01
3.18088174e-01 3.91642451e-01 -2.94138551e-01 2.01091021e-01
-9.09680128e-02 2.94273704e-01 -7.94478059e-02 2.19593942e-01
3.08310390e-01 6.21013463e-01 -5.93300045e-01 -1.09661388e+00
-8.37532505e-02 1.12785555e-01 -7.68297613e-02 -1.66881904e-02
-3.79520431e-02 -1.68170687e-02 -4.06505942e-01 8.40922952e-01
6.59789145e-01 4.37724710e-01 4.19590712e-01 2.29125589e-01
-5.96071661e-01 3.19340974e-01 -1.34191787e+00 1.06105542e+00
-9.05871317e-02 2.87778258e-01 -1.66265443e-01 -8.60921323e-01
1.15742397e+00 3.79715264e-01 6.74279854e-02 -5.62058330e-01
3.95030975e-01 6.91925943e-01 4.50444937e-01 -5.49565315e-01
8.23661327e-01 -1.34868681e-01 -3.89895618e-01 2.48679563e-01
-2.64271140e-01 -3.91163714e-02 6.89689994e-01 -1.28515795e-01
5.30724525e-01 -1.03859641e-01 5.87602913e-01 -1.95969671e-01
7.95138419e-01 4.16149318e-01 1.21750939e+00 3.68635386e-01
-3.16090822e-01 4.86748904e-01 4.88401085e-01 -1.53106079e-02
-1.09247375e+00 -6.67586565e-01 -8.04104879e-02 9.52057362e-01
1.09778427e-01 -1.69558734e-01 -1.05752099e+00 -6.67892754e-01
-3.21438044e-01 6.79738164e-01 6.93188012e-02 3.11456054e-01
-1.07675266e+00 -5.69699943e-01 5.50336778e-01 4.90430504e-01
5.66267252e-01 -1.26414251e+00 -4.11120057e-01 5.10466039e-01
-1.87848583e-01 -1.34205258e+00 -7.66689658e-01 2.27759555e-01
-5.94619095e-01 -8.70453835e-01 -7.14077711e-01 -1.59315801e+00
6.77078247e-01 7.19868317e-02 3.81426305e-01 2.86450028e-01
-1.00210041e-01 -6.30939305e-01 -6.27073169e-01 -3.52394849e-01
-3.02632153e-01 4.41465855e-01 4.33809729e-03 -2.32040450e-01
8.15574467e-01 2.64648676e-01 -3.04226905e-01 2.26522997e-01
-7.84457743e-01 -1.10737994e-01 6.82528615e-01 5.00666559e-01
7.30160058e-01 3.94573987e-01 3.83609474e-01 -1.02576733e+00
5.64446688e-01 -4.95623499e-01 -7.33787596e-01 3.52550954e-01
-4.54015255e-01 -1.65157929e-01 9.96528983e-01 -6.54738724e-01
-1.07429004e+00 3.18426788e-01 -6.64110899e-01 4.42802489e-01
-3.08593839e-01 7.78399587e-01 -4.52569276e-01 3.26556176e-01
5.22552058e-02 3.78379077e-01 -3.42142701e-01 -4.46577370e-01
-1.03576228e-01 1.20488989e+00 4.09027040e-01 -2.40028724e-01
6.52231932e-01 -1.86549664e-01 -3.71041447e-01 -1.12063289e+00
-3.40945154e-01 -6.91490114e-01 -8.77926707e-01 8.58456939e-02
1.26877820e+00 -5.16996682e-01 -7.80703664e-01 5.04692733e-01
-1.26948142e+00 -5.79762496e-02 4.18877482e-01 7.79587865e-01
2.68287122e-01 7.12008595e-01 -7.73800015e-01 -1.04561484e+00
-5.11676490e-01 -1.19516969e+00 5.08361578e-01 7.94015169e-01
-2.27316886e-01 -7.14825153e-01 -3.50113153e-01 2.67710090e-01
-1.25149697e-01 -2.31442422e-01 1.10364616e+00 -1.14338696e+00
-1.12797998e-01 -2.45527923e-01 -5.58933243e-02 2.03331798e-01
-3.98063846e-02 3.64846885e-01 -4.31192458e-01 -2.21077353e-02
-6.48126975e-02 1.43217459e-01 8.12206149e-01 1.29281849e-01
6.44874513e-01 -1.75003186e-01 -2.10341513e-01 2.21136823e-01
1.42740417e+00 1.08275414e+00 5.44324100e-01 2.83886135e-01
5.91162145e-01 7.08306968e-01 1.21638834e+00 2.93906897e-01
4.26072627e-01 2.39185378e-01 -1.23129062e-01 -1.93195775e-01
3.66984963e-01 -2.10907757e-01 4.48333681e-01 1.29330623e+00
3.17184716e-01 -4.77074623e-01 -1.33392227e+00 5.49313843e-01
-1.44063413e+00 -5.26651800e-01 -5.66039622e-01 2.23841929e+00
9.80591476e-01 3.38251770e-01 1.30860537e-01 5.01721919e-01
1.05335319e+00 1.61747131e-02 -2.11971417e-01 -7.89819419e-01
-4.82749976e-02 4.58840162e-01 6.35859013e-01 7.93425798e-01
-1.07051408e+00 1.57103336e+00 5.37029791e+00 1.05178118e+00
-1.18813539e+00 -8.11291784e-02 5.53394020e-01 5.08077919e-01
-7.59226233e-02 1.31588385e-01 -1.44095874e+00 6.56489134e-01
6.69187069e-01 -9.42623839e-02 1.17654860e-01 5.68801582e-01
2.56277025e-01 -5.04373074e-01 -4.55451399e-01 7.91958153e-01
1.21433362e-01 -7.03237414e-01 -7.76093155e-02 1.82784908e-02
5.07746875e-01 -3.52956176e-01 -2.03625530e-01 2.69590408e-01
-1.26601920e-01 -8.45827520e-01 7.12423503e-01 -2.81456131e-02
7.77980566e-01 -1.09044409e+00 9.42291081e-01 6.88437700e-01
-1.22657335e+00 2.39727959e-01 -5.57536483e-01 2.12850664e-02
3.20222914e-01 4.11831558e-01 -1.05430532e+00 2.25838333e-01
2.98021287e-01 5.64056411e-02 -3.08825016e-01 1.04735303e+00
-5.64632118e-01 9.43395555e-01 -4.56157446e-01 -5.98972023e-01
5.54533780e-01 -5.17895877e-01 1.98926941e-01 1.57843292e+00
4.13747728e-01 3.86306584e-01 5.50606132e-01 4.14549559e-01
1.88127562e-01 9.11790609e-01 -1.61917716e-01 -3.30735594e-01
5.74655294e-01 1.12447011e+00 -1.36102939e+00 -4.57368314e-01
-4.57730055e-01 1.01233006e+00 1.22652456e-01 3.86605918e-01
-7.71161199e-01 -1.07267892e+00 2.03648522e-01 -1.43664435e-01
3.64612490e-01 -6.60704076e-01 -4.29930270e-01 -1.00147331e+00
-9.67869908e-02 -8.26357186e-01 4.21433300e-01 -1.14339449e-01
-7.43415654e-01 6.86207175e-01 -3.52965266e-01 -7.66545773e-01
-1.06255017e-01 -8.55255961e-01 -6.99904978e-01 1.06669700e+00
-1.49579954e+00 -6.83757007e-01 2.44897828e-02 2.96001256e-01
9.31975663e-01 -1.55249164e-01 7.11832225e-01 2.49881640e-01
-1.19197106e+00 6.00326121e-01 -5.67540992e-03 7.18402386e-01
6.58224642e-01 -1.15522540e+00 3.57874453e-01 1.25481188e+00
3.64164375e-02 5.35846770e-01 4.12445754e-01 -1.14616656e+00
-9.57278609e-01 -6.11275733e-01 1.68169594e+00 4.66497302e-01
4.56394613e-01 -3.98772597e-01 -9.55044389e-01 4.21347320e-01
1.76765740e-01 -5.12762845e-01 8.56553137e-01 -1.57665148e-01
3.12769972e-02 6.75090998e-02 -9.18121040e-01 6.65864527e-01
3.52490902e-01 -1.17166378e-01 -6.10773921e-01 2.16334149e-01
6.94986939e-01 -4.88306791e-01 -6.12499416e-01 6.61057979e-02
4.99439210e-01 -5.43317020e-01 2.22657368e-01 -2.83592254e-01
7.24133998e-02 -4.50922400e-01 -1.99744683e-02 -7.73713231e-01
-1.36317208e-01 -5.16430318e-01 6.33707881e-01 1.56337750e+00
9.38073039e-01 -6.59481347e-01 7.27784812e-01 4.60525304e-01
-3.66809405e-02 -5.20452082e-01 -6.51102126e-01 -6.48241282e-01
1.27088264e-01 -3.19140732e-01 6.58259928e-01 7.48467207e-01
9.20768306e-02 4.62439030e-01 -1.03432506e-01 1.55189052e-01
3.50656688e-01 4.28384840e-02 2.04248205e-01 -8.86460841e-01
1.83172692e-02 -4.55186903e-01 -2.28610020e-02 -1.08460963e+00
3.41319382e-01 -7.62697756e-01 3.56270999e-01 -1.30451417e+00
1.01347260e-01 -5.90275466e-01 -2.48128101e-02 4.82163072e-01
-4.16357428e-01 1.27676964e-01 9.56534147e-02 6.77712113e-02
-2.31471613e-01 7.05390349e-02 1.01440859e+00 1.09496582e-02
-5.39679945e-01 9.57822204e-02 -4.41048533e-01 5.98807752e-01
1.16778588e+00 -5.20217538e-01 1.81744650e-01 -4.84153539e-01
1.76141355e-02 5.48596447e-03 -6.03766084e-01 -5.75254798e-01
3.85272771e-01 -5.53867817e-01 3.11226636e-01 -9.15940821e-01
-1.73111573e-01 -7.52226710e-01 -2.30284333e-01 7.36867845e-01
7.09906816e-02 2.88871914e-01 2.35563572e-02 -4.27423045e-02
-2.38955647e-01 -9.09206629e-01 6.62140727e-01 -1.61225483e-01
-9.29842949e-01 1.00194894e-01 -8.77543688e-01 1.74762607e-02
1.25631475e+00 -5.53037703e-01 -1.08796291e-01 6.20755628e-02
-2.75817513e-01 4.83279735e-01 1.75573766e-01 1.98536694e-01
6.89549625e-01 -7.59633839e-01 -7.92243481e-01 3.22104722e-01
-1.73135653e-01 -8.73153284e-02 -6.92557544e-02 6.61015332e-01
-1.19043291e+00 4.37417120e-01 6.09258562e-02 -2.32893646e-01
-1.48531353e+00 5.55640042e-01 -1.31309584e-01 -1.01330385e-01
-3.20589870e-01 5.21683395e-01 -2.80415118e-01 -8.38331357e-02
1.83266550e-01 -2.27385670e-01 -6.02751315e-01 2.18424395e-01
3.64914149e-01 3.44621360e-01 -2.18439892e-01 -9.88798320e-01
-4.64738488e-01 6.94330156e-01 -2.60246366e-01 -4.58994478e-01
7.05291212e-01 -1.58329867e-02 -3.12945575e-01 3.77786607e-01
9.63362694e-01 4.51871276e-01 -5.07285953e-01 -1.21036112e-01
3.48042339e-01 -2.59647608e-01 -1.82906345e-01 -5.89352906e-01
-6.97110057e-01 9.33515847e-01 1.61714196e-01 -1.54780727e-02
8.90223563e-01 -1.94231004e-01 1.31987226e+00 2.43819386e-01
1.27586320e-01 -1.59531522e+00 -4.55246985e-01 8.98445606e-01
2.90838063e-01 -1.27607298e+00 -1.88398018e-01 -6.73019290e-01
-8.93839061e-01 1.28142774e+00 8.01647127e-01 2.45459620e-02
5.12063444e-01 2.61518627e-01 2.76535511e-01 4.77765650e-01
-1.65886164e-01 -6.42474294e-01 1.53280586e-01 3.00447822e-01
5.74954569e-01 1.96028590e-01 -1.45902646e+00 6.98291540e-01
-4.42025930e-01 -6.35663927e-01 5.00305235e-01 7.67615855e-01
-9.18039382e-01 -1.39030635e+00 -4.70178485e-01 2.50757575e-01
-8.75091136e-01 -3.14001918e-01 -2.39753470e-01 8.61119211e-01
1.82686925e-01 1.30409980e+00 1.40280247e-01 -3.25654119e-01
1.32776991e-01 2.22227141e-01 1.06491402e-01 -8.88569772e-01
-6.93247557e-01 6.94366455e-01 -5.52820340e-02 8.47317874e-02
7.85590336e-02 -6.45556986e-01 -2.06671739e+00 -2.61293948e-01
-5.21213591e-01 5.35138309e-01 7.73369193e-01 9.94055033e-01
-4.74040002e-01 5.78151979e-02 5.61897933e-01 -1.80074826e-01
-2.60614067e-01 -8.43534052e-01 -9.10880864e-01 -3.35753895e-02
-1.06885605e-01 -1.23484805e-01 -1.91539273e-01 -8.60024765e-02] | [10.158830642700195, 10.147780418395996] |
8600deb0-f216-40a4-8a5e-77ccded78b08 | optical-flow-and-scene-flow-estimation-a | null | null | https://www.sciencedirect.com/science/article/abs/pii/S0031320321000480 | https://www.sciencedirect.com/science/article/abs/pii/S0031320321000480 | Optical flow and scene flow estimation: A survey | Motion analysis is one of the most fundamental and challenging problems in the field of computer vision, which can be widely applied in many areas, such as autonomous driving, action recognition, scene understanding, and robotics. In general, the displacement field between subsequent frames can be divided into two types: optical flow and scene flow. The optical flow represents the pixel motion of adjacent frames. In contrast, the scene flow is a 3D motion field of the dynamic scene between two frames. Traditional approaches for the estimation of optical flow and scene flow usually leverage the variational technique, which can be solved as an energy minimization process. In recent years, deep learning has emerged as a powerful technique for learning feature representations directly from data. It has led to remarkable progress in the field of optical flow and scene flow estimation. In this paper, we provide a comprehensive survey of optical flow and scene flow estimation. First, we briefly review the pioneering approaches that use variational technique and then we delve in detail into the deep learning-based approaches. Furthermore, we present insightful observations on evaluation issues, specifically benchmark datasets, evaluation metrics, and state-of-the-art performance. Finally, we give the promising directions for future research. To the best of our knowledge, we are the first to review both optical flow and scene flow estimation, and the first to cover both traditional and deep learning-based approaches. | ['XiangdongKong', 'NingLv', 'XuezhiXiang', 'MingliangZhai'] | 2021-02-01 | null | null | null | pattern-recognition-2021-2 | ['scene-flow-estimation'] | ['computer-vision'] | [ 3.30421887e-02 -6.37894809e-01 -4.44030315e-01 -1.32381946e-01
-2.16584489e-01 -3.81619155e-01 4.91548479e-01 -2.14311451e-01
-4.45381552e-01 6.80562198e-01 9.87655371e-02 -1.52027637e-01
-6.31512627e-02 -5.80358803e-01 -4.59740192e-01 -7.92292118e-01
-1.43107295e-01 -2.53827065e-01 3.56210530e-01 -2.19889414e-02
5.31464636e-01 5.46829402e-01 -1.40130877e+00 -1.87857345e-01
5.36551476e-01 9.40760672e-01 2.27443278e-01 8.75859022e-01
-2.09662676e-01 1.33592367e+00 -3.25403363e-01 -1.20581709e-01
9.00050327e-02 -5.71560919e-01 -1.10341167e+00 2.98671067e-01
5.55356920e-01 -7.37127066e-01 -9.62186217e-01 9.14366126e-01
3.04696113e-01 6.24615192e-01 4.58240300e-01 -1.51873457e+00
-5.36615491e-01 -1.38670534e-01 -6.80540979e-01 8.36951017e-01
4.49918360e-01 2.97302425e-01 8.42803299e-01 -7.44010448e-01
7.33465791e-01 1.27529109e+00 1.80370048e-01 5.16968727e-01
-7.48212695e-01 -2.14182734e-01 3.80997956e-01 8.50000858e-01
-8.79621089e-01 -3.88427943e-01 9.80389655e-01 -8.54149342e-01
8.02870631e-01 -6.38919249e-02 7.83384442e-01 6.74239218e-01
4.65896934e-01 1.31809318e+00 6.30928159e-01 -2.80945480e-01
6.79388195e-02 -3.71245146e-01 8.68986920e-02 8.54308128e-01
-4.45107967e-02 1.49049357e-01 -5.47926307e-01 1.68978423e-01
9.55464780e-01 5.30413017e-02 -5.31293333e-01 -5.01746595e-01
-1.38011980e+00 9.22279298e-01 4.27899063e-01 1.74200952e-01
-2.37409532e-01 4.68589693e-01 4.16817933e-01 -1.28351659e-01
3.51261705e-01 7.56113008e-02 -1.11824892e-01 -3.59489471e-01
-9.06428814e-01 4.41604346e-01 6.61842585e-01 6.36811435e-01
9.60937500e-01 3.93288881e-01 -1.14608943e-01 6.56474411e-01
4.96971309e-01 4.17189628e-01 3.92192185e-01 -1.65188849e+00
1.84760734e-01 1.66059256e-01 8.15200508e-02 -1.26884472e+00
-1.60706490e-01 2.49311086e-02 -7.18210757e-01 2.13085711e-01
4.75762904e-01 -2.53542185e-01 -5.18649876e-01 1.42471826e+00
3.91960293e-01 8.57106924e-01 -5.89897819e-02 1.09653819e+00
9.99228835e-01 7.86245346e-01 -2.61162549e-01 -3.56821209e-01
9.36843276e-01 -1.30645537e+00 -1.03475201e+00 -1.94691643e-01
5.21397293e-01 -8.84286940e-01 4.60951269e-01 5.58352359e-02
-1.25996816e+00 -7.39124119e-01 -9.26715374e-01 -3.97958964e-01
-1.04967862e-01 -1.90370008e-01 8.56233239e-01 2.43180305e-01
-1.03556800e+00 5.78160524e-01 -1.24672270e+00 -4.63615716e-01
5.58883071e-01 1.09698012e-01 -2.00400829e-01 -2.35270351e-01
-1.05404663e+00 7.58145452e-01 1.88924462e-01 3.11010152e-01
-9.94704247e-01 -6.27281725e-01 -1.15412116e+00 -2.83719182e-01
2.67302483e-01 -8.21196258e-01 1.36670125e+00 -5.37582278e-01
-1.61733007e+00 7.72280335e-01 -7.15124130e-01 -3.70050460e-01
5.45222759e-01 -3.84245336e-01 -1.75294012e-01 2.06671417e-01
8.03741962e-02 5.97587168e-01 6.44330561e-01 -8.94129932e-01
-9.30821598e-01 -6.74806759e-02 4.15004820e-01 1.78551644e-01
2.02527903e-02 1.52832568e-01 -4.36714530e-01 -3.37075710e-01
-1.17884047e-01 -9.24714327e-01 -3.21784794e-01 3.67934078e-01
-3.43505412e-01 -2.85252124e-01 1.06750572e+00 -2.30860770e-01
1.25256038e+00 -2.01282668e+00 1.98537990e-01 -5.31578064e-01
3.67925823e-01 4.28526103e-01 8.19881409e-02 1.94318518e-01
1.19080208e-01 -2.52095938e-01 -3.78048301e-01 -3.57076555e-01
-2.74955750e-01 3.19342226e-01 -4.12419289e-01 9.00691032e-01
1.27605051e-01 1.08374345e+00 -1.34038436e+00 -5.09156823e-01
7.74730623e-01 6.06887519e-01 -5.12196064e-01 9.15828943e-02
6.35413453e-02 8.23339403e-01 -6.24418557e-01 2.95555800e-01
6.18715703e-01 -2.57501513e-01 -2.90119022e-01 -1.22035459e-01
-3.74083906e-01 1.79556221e-01 -1.17690015e+00 1.86803257e+00
-2.78804213e-01 1.33869720e+00 -2.37645656e-02 -1.27164543e+00
5.83859980e-01 9.51388404e-02 1.03492200e+00 -4.25984472e-01
1.71434030e-01 1.11321658e-01 -9.37118083e-02 -8.33913743e-01
7.11483359e-01 6.10737875e-02 3.06774676e-01 3.67121637e-01
-1.01137221e-01 -1.48593724e-01 5.22498906e-01 7.11429939e-02
7.48775184e-01 3.57921362e-01 3.66777033e-01 1.25940023e-02
1.08370948e+00 1.49415145e-02 6.22785807e-01 4.34668332e-01
-9.32345808e-01 4.24350023e-01 2.15625465e-01 -8.38475823e-01
-6.53900385e-01 -9.30074275e-01 -1.17847674e-01 6.11671805e-01
6.05584145e-01 -2.56330103e-01 -4.69699979e-01 -3.75891387e-01
-8.65617320e-02 1.39432505e-01 -4.21058983e-01 -5.51255867e-02
-9.62311864e-01 -5.68553627e-01 2.61835575e-01 5.81399798e-01
8.26771677e-01 -1.09582055e+00 -7.65275240e-01 1.76262289e-01
-5.72949588e-01 -1.55954230e+00 -5.89990020e-01 -6.30565584e-01
-9.33923662e-01 -1.25864434e+00 -8.08889389e-01 -8.90529513e-01
1.93783075e-01 7.92217612e-01 1.00978935e+00 1.35692388e-01
-3.99744570e-01 6.48070574e-01 -1.47581100e-01 -5.03443219e-02
-1.39034525e-01 -1.79448888e-01 -6.42596418e-03 2.33868390e-01
3.87489498e-01 -4.03947085e-01 -9.04173017e-01 1.68384954e-01
-8.22242677e-01 -2.03964815e-01 3.97323593e-02 5.69341838e-01
6.12271011e-01 -1.80594206e-01 -6.39935303e-03 -5.35681725e-01
2.84329712e-01 -4.05313849e-01 -5.60753524e-01 -5.14645204e-02
4.50516352e-03 -4.70863953e-02 4.40381944e-01 -8.17184150e-02
-1.05762351e+00 5.81885837e-02 -8.81918594e-02 -6.40067041e-01
-2.00281188e-01 4.22336847e-01 2.17746332e-01 -2.41316766e-01
2.92490244e-01 2.65980482e-01 -1.08705923e-01 -2.03264520e-01
4.50317889e-01 3.58894825e-01 6.41659796e-01 -3.38742644e-01
5.42739034e-01 1.13682687e+00 2.98580468e-01 -1.17115974e+00
-9.65284467e-01 -8.40546966e-01 -8.90031636e-01 -6.05351388e-01
1.14667666e+00 -6.00874066e-01 -8.99786830e-01 8.67541671e-01
-1.51879656e+00 -4.30420995e-01 -1.35406032e-01 8.38002861e-01
-8.70876491e-01 7.84166932e-01 -7.74835467e-01 -7.11875141e-01
7.31041804e-02 -1.49128985e+00 1.03506732e+00 4.85075086e-01
9.35147777e-02 -1.69732654e+00 3.95655602e-01 3.82079273e-01
1.03037119e-01 5.16203940e-01 2.01497674e-01 1.88859373e-01
-9.30227280e-01 6.62368312e-02 -2.18645304e-01 3.42231810e-01
2.83869803e-01 2.72372127e-01 -9.79955852e-01 -2.10522681e-01
1.10666610e-01 -7.80453011e-02 9.76317585e-01 9.32195723e-01
9.56387997e-01 2.02089146e-01 -3.79427284e-01 1.02259886e+00
1.38600469e+00 3.35955739e-01 6.01276934e-01 3.12999904e-01
1.02788568e+00 7.06671655e-01 5.44727862e-01 3.49639177e-01
5.43292522e-01 6.92464650e-01 3.24919015e-01 5.40313870e-02
-2.97047764e-01 -1.86496116e-02 3.75404179e-01 7.53895819e-01
-2.90414959e-01 -2.05150485e-01 -6.93585157e-01 5.36160946e-01
-2.00338793e+00 -1.21491122e+00 -4.22054678e-01 1.88774741e+00
2.49531493e-01 -1.52426451e-01 -2.20053848e-02 1.40253574e-01
7.71654546e-01 7.32797682e-01 -6.43471897e-01 -2.58252263e-01
2.73696501e-02 -8.55133310e-02 4.01618451e-01 9.62097943e-01
-1.38840568e+00 1.17061496e+00 6.34557867e+00 3.78599435e-01
-1.31462872e+00 -4.77760583e-02 3.48103225e-01 1.52629903e-02
1.70774564e-01 6.43402040e-02 -7.61825979e-01 4.70970571e-01
5.28046191e-01 -2.34110072e-01 3.24961007e-01 5.83718657e-01
7.22209454e-01 -3.38911474e-01 -1.01482165e+00 1.28898740e+00
8.27292576e-02 -1.57357228e+00 -4.02831212e-02 1.74362704e-01
9.44672406e-01 1.66580558e-01 -2.47653406e-02 -4.57899794e-02
8.17672163e-02 -7.41801500e-01 3.97151858e-01 5.75615764e-01
5.02414942e-01 -5.04341125e-01 5.85765243e-01 2.35326052e-01
-1.46745837e+00 7.63162971e-02 -3.33915889e-01 -4.53762054e-01
7.89643645e-01 6.29793882e-01 6.57741278e-02 5.68100691e-01
7.27767825e-01 1.64051318e+00 2.63586994e-02 1.27470112e+00
-1.89275041e-01 3.63318503e-01 3.08596082e-02 1.43552363e-01
5.21795571e-01 -4.64215666e-01 6.87133551e-01 1.12754250e+00
-3.03116664e-02 6.07797643e-03 5.31252742e-01 7.82599866e-01
9.02022719e-02 -2.32816145e-01 -6.73777997e-01 1.86663911e-01
9.44208503e-02 1.18068993e+00 -5.59129953e-01 -3.79830092e-01
-7.90743947e-01 9.22980845e-01 8.01575109e-02 6.50584042e-01
-7.73922145e-01 -3.40255082e-01 1.13521183e+00 -2.08921149e-01
5.82295209e-02 -7.73089647e-01 7.81606436e-02 -1.45727754e+00
-1.08529106e-01 -7.27980807e-02 2.11562723e-01 -6.91168606e-01
-9.44667816e-01 3.30179363e-01 1.13097958e-01 -1.36642170e+00
-3.93317372e-01 -6.90293014e-01 -7.89449632e-01 7.57100582e-01
-2.17350531e+00 -5.35249591e-01 -6.95090652e-01 6.70219302e-01
9.02823091e-01 1.95272341e-01 1.89846292e-01 3.09905618e-01
-7.39388287e-01 1.00302510e-02 2.41235659e-01 5.42697847e-01
4.88100648e-01 -1.08416986e+00 4.48906779e-01 1.18413889e+00
2.46961102e-01 2.97980607e-01 4.56195354e-01 -2.18772814e-01
-1.30847847e+00 -1.09382617e+00 8.43516946e-01 -3.45637918e-01
7.87808955e-01 3.55069563e-02 -8.11630130e-01 6.41802669e-01
2.76577860e-01 5.20763457e-01 4.39016253e-01 -6.56858146e-01
3.07752699e-01 3.75557281e-02 -6.86942458e-01 4.96724546e-01
9.44257855e-01 -4.98693645e-01 -2.42360055e-01 2.07428530e-01
4.16712224e-01 -6.24593437e-01 -6.50487781e-01 1.64794862e-01
6.10455096e-01 -1.21452069e+00 1.01741838e+00 -5.81056952e-01
4.90912646e-01 -3.83598626e-01 6.88033784e-03 -1.15395796e+00
-4.02675182e-01 -8.60000014e-01 -4.64548141e-01 7.07113743e-01
-3.29948843e-01 -4.79283422e-01 9.35014606e-01 3.78868431e-01
-2.80202746e-01 -7.33749628e-01 -5.98506093e-01 -6.20287061e-01
1.06268801e-01 -6.31466448e-01 3.56501580e-04 8.88013124e-01
-3.15789878e-01 3.13972980e-01 -3.31179917e-01 -9.89335179e-02
7.09304273e-01 1.36857301e-01 8.52437794e-01 -1.02678275e+00
-9.98314917e-02 -6.98976696e-01 -7.17772543e-01 -1.83103931e+00
5.20288289e-01 -6.81769133e-01 2.11155117e-01 -1.77407062e+00
1.91415399e-02 7.98109695e-02 -1.36135772e-01 -1.33885860e-01
-3.76894772e-01 2.08264887e-01 3.29838485e-01 4.11101282e-01
-5.05351543e-01 5.79174340e-01 1.80740261e+00 -1.28724754e-01
-2.76341319e-01 1.04122825e-01 -1.62658229e-01 8.80881667e-01
7.90315866e-01 -1.65556148e-02 -5.98194718e-01 -7.26571798e-01
-2.03377843e-01 5.78472670e-03 4.12763834e-01 -7.77669609e-01
3.90642881e-01 -4.67533350e-01 4.50281873e-02 -5.19032657e-01
3.11535567e-01 -5.10407269e-01 -4.86519873e-01 4.86751765e-01
-1.16336592e-01 3.77646200e-02 5.24404980e-02 6.21006668e-01
-6.11412585e-01 -1.09305404e-01 9.35174406e-01 -2.26485282e-01
-1.40690672e+00 7.87748933e-01 -5.51176846e-01 2.99069405e-01
1.09103918e+00 -3.69330615e-01 -1.59393445e-01 -5.22978365e-01
-4.37243342e-01 2.94456750e-01 1.68910697e-01 5.57843268e-01
7.95244634e-01 -1.19986868e+00 -6.71419561e-01 5.18149957e-02
-2.20235974e-01 1.75365493e-01 4.16370183e-01 1.11313164e+00
-9.07050848e-01 6.65371776e-01 -1.64710149e-01 -9.00020838e-01
-9.94608760e-01 5.71343541e-01 5.83287656e-01 1.29816597e-02
-5.67721426e-01 7.22968996e-01 6.18146539e-01 1.06503665e-01
2.06034482e-01 -3.79256099e-01 -3.92644018e-01 -2.03945756e-01
7.73191690e-01 9.39459920e-01 -3.31704259e-01 -1.06304133e+00
-5.51947594e-01 9.93754923e-01 1.66665539e-01 4.37118858e-02
8.49383593e-01 -4.28360164e-01 -2.05512315e-01 5.68594635e-01
1.65213203e+00 -3.68451387e-01 -1.54817402e+00 -1.75306991e-01
-1.78288817e-01 -6.26224875e-01 3.22721094e-01 1.01511955e-01
-1.55151021e+00 1.38403749e+00 4.06836569e-01 2.78283328e-01
1.06174481e+00 -1.01341374e-01 1.14227974e+00 1.59237832e-01
2.89682925e-01 -8.09364259e-01 2.52791345e-01 7.61162102e-01
3.98326814e-01 -1.34967387e+00 5.05126175e-03 -4.44039732e-01
-2.96395272e-01 1.21769464e+00 4.47414875e-01 -4.07457709e-01
1.00134444e+00 3.22372913e-02 8.34384933e-02 -1.83128137e-02
-5.17381370e-01 -3.93000871e-01 3.13405752e-01 6.40766263e-01
6.85753286e-01 -3.28579485e-01 -6.54259548e-02 -3.68454039e-01
1.87110305e-01 2.61518031e-01 6.71861768e-01 9.94229853e-01
-4.98862803e-01 -9.64818895e-01 -1.62639081e-01 1.15728833e-01
-3.29408258e-01 1.76819265e-01 -7.80430436e-02 5.97952545e-01
2.04374436e-02 1.12662494e+00 4.00846869e-01 -6.81853220e-02
2.91831166e-01 -4.05803263e-01 7.27914870e-01 -3.23177606e-01
1.77757055e-01 -1.10157624e-01 -4.06169891e-01 -9.20022309e-01
-1.19682622e+00 -8.02504659e-01 -1.23637009e+00 -5.44217706e-01
-5.76665662e-02 -7.86919221e-02 3.68954003e-01 1.09800828e+00
6.21116124e-02 5.22101939e-01 6.97412491e-01 -1.04163837e+00
1.50214314e-01 -6.15524173e-01 -5.38196325e-01 3.02249968e-01
8.84537458e-01 -9.62892711e-01 -3.68513912e-01 5.10735095e-01] | [8.766536712646484, -1.7167549133300781] |
96825d15-554b-4673-97df-e312096fde04 | gaining-insights-into-unrecognized-user | 2204.05158 | null | https://arxiv.org/abs/2204.05158v2 | https://arxiv.org/pdf/2204.05158v2.pdf | Gaining Insights into Unrecognized User Utterances in Task-Oriented Dialog Systems | The rapidly growing market demand for automatic dialogue agents capable of goal-oriented behavior has caused many tech-industry leaders to invest considerable efforts into task-oriented dialog systems. The success of these systems is highly dependent on the accuracy of their intent identification -- the process of deducing the goal or meaning of the user's request and mapping it to one of the known intents for further processing. Gaining insights into unrecognized utterances -- user requests the systems fail to attribute to a known intent -- is therefore a key process in continuous improvement of goal-oriented dialog systems. We present an end-to-end pipeline for processing unrecognized user utterances, deployed in a real-world, commercial task-oriented dialog system, including a specifically-tailored clustering algorithm, a novel approach to cluster representative extraction, and cluster naming. We evaluated the proposed components, demonstrating their benefits in the analysis of unrecognized user requests. | ['Ateret Anaby-Tavor', 'Gaurav Pandey', 'Vineet Kumar', 'David Boaz', 'Matan Vetzler', 'Ella Rabinovich'] | 2022-04-11 | null | null | null | null | ['goal-oriented-dialog'] | ['natural-language-processing'] | [ 1.16257720e-01 2.09665313e-01 1.63395613e-01 -8.53899717e-01
-8.18640292e-01 -8.42051744e-01 7.40919292e-01 3.16430807e-01
-2.57255793e-01 4.26935017e-01 6.84990525e-01 -4.57348794e-01
-3.07239387e-02 -2.20252037e-01 4.64381605e-01 -3.13646406e-01
-1.96715832e-01 1.34523046e+00 1.23697728e-01 -7.03588903e-01
5.59710026e-01 3.25591445e-01 -1.26778996e+00 5.04764557e-01
6.44935191e-01 6.65808797e-01 2.47900516e-01 1.01843226e+00
-4.55791801e-01 1.13494420e+00 -8.81583333e-01 -3.94238420e-02
-2.29635462e-02 -7.42746413e-01 -1.38788772e+00 4.98294860e-01
-1.12346508e-01 -3.25301170e-01 7.72727728e-02 8.30419898e-01
2.01146469e-01 4.87144113e-01 6.63562417e-01 -1.44476092e+00
-1.48956984e-01 7.10024238e-01 1.17088035e-01 -1.68430313e-01
7.27628350e-01 1.32755443e-01 1.37536478e+00 -3.79426807e-01
5.51690519e-01 1.36410797e+00 2.57435590e-01 9.18957770e-01
-1.28375995e+00 -2.62116641e-01 -5.51961623e-02 -2.85403989e-02
-1.24862921e+00 -8.93535376e-01 6.52910471e-01 -6.61005020e-01
1.21005702e+00 5.58546543e-01 5.94106428e-02 8.59388769e-01
-1.65949255e-01 4.53974903e-01 1.02062333e+00 -5.31061292e-01
5.36981702e-01 5.00546098e-01 5.50155461e-01 4.74977702e-01
-4.97208893e-01 -5.02469897e-01 -4.76452976e-01 -4.42801595e-01
1.77334577e-01 -4.08958226e-01 -1.43441081e-01 -1.86710956e-03
-7.44758964e-01 1.01825106e+00 -6.12819940e-02 6.50419652e-01
-4.55805838e-01 -3.67742330e-01 5.15959203e-01 4.98671442e-01
5.06883860e-01 7.88291872e-01 -4.96251941e-01 -6.30853713e-01
-5.56868255e-01 3.43101412e-01 1.41645825e+00 1.23718596e+00
7.27484524e-01 -2.74319530e-01 -2.70014942e-01 9.95196581e-01
4.07465339e-01 4.60370071e-03 7.12962508e-01 -9.66991663e-01
3.11379641e-01 1.21444464e+00 3.54306370e-01 -7.54821599e-01
-8.24773729e-01 2.57991225e-01 -2.23256171e-01 4.80200686e-02
6.39089823e-01 -3.89113426e-01 -4.21580404e-01 1.55972445e+00
3.05871546e-01 -6.42121851e-01 3.81044805e-01 9.38172758e-01
6.39131486e-01 4.48606372e-01 1.86817273e-01 -3.68301064e-01
1.65364671e+00 -5.17309546e-01 -7.75388360e-01 -2.22356170e-01
7.68033326e-01 -9.50682342e-01 1.19204891e+00 2.15315893e-01
-6.24088466e-01 -3.13775539e-01 -5.76648295e-01 1.04095422e-01
-2.07323477e-01 -4.18686196e-02 4.80613977e-01 7.66185999e-01
-1.04682887e+00 3.03629100e-01 -4.03123349e-01 -8.20217609e-01
-3.80609334e-01 6.46260381e-01 -2.97349870e-01 2.77567029e-01
-8.79484057e-01 8.32503557e-01 5.13001680e-01 -3.30861837e-01
-4.66631502e-01 -1.95397928e-01 -7.51054049e-01 2.28329837e-01
4.65781331e-01 -1.73873544e-01 1.80346179e+00 -5.46830595e-01
-1.68825841e+00 8.79103959e-01 -3.04667413e-01 -3.08887899e-01
1.68503940e-01 6.14645481e-02 -4.36789066e-01 -1.96579751e-02
1.68597862e-01 4.59214926e-01 7.87143826e-01 -1.08082151e+00
-1.09097767e+00 -5.44370294e-01 1.13775926e-02 4.74234283e-01
-5.65448701e-01 4.33890462e-01 -2.13017210e-01 5.83781768e-03
2.15539798e-01 -1.05390716e+00 -1.34888202e-01 -8.42217743e-01
-5.90877295e-01 -8.91068459e-01 9.81322944e-01 -6.80264950e-01
1.12864912e+00 -1.98222971e+00 -1.71433613e-02 4.73565124e-02
4.51983482e-01 1.51960984e-01 2.28307918e-01 8.42522085e-01
8.57696459e-02 7.03613162e-02 -6.21002242e-02 -5.21800935e-01
3.63551676e-01 1.55597717e-01 -3.24652314e-01 1.57259583e-01
1.68528389e-02 4.37141061e-01 -9.85291958e-01 -6.18263483e-01
5.66093683e-01 -2.27839679e-01 -4.23621863e-01 1.04332805e+00
-7.22123206e-01 6.11285329e-01 -6.08139277e-01 1.31683171e-01
2.16045424e-01 -2.12356612e-01 5.53806245e-01 1.86817590e-02
-2.43160471e-01 7.59324491e-01 -7.61975288e-01 1.44191813e+00
-4.49544191e-01 5.19324243e-01 5.05244136e-01 -5.95716774e-01
1.01534688e+00 6.08872473e-01 5.61915517e-01 -2.63984621e-01
3.67336363e-01 1.18111791e-02 2.72382051e-01 -7.75468588e-01
8.74961972e-01 -2.27284789e-01 -4.87776518e-01 8.85759115e-01
2.29559287e-01 -1.57306105e-01 3.36975306e-01 5.04697621e-01
1.23043442e+00 -5.10515928e-01 4.94813919e-01 -3.01689893e-01
6.75346494e-01 6.61824822e-01 2.16241091e-01 5.47424138e-01
-3.66985112e-01 3.91999036e-01 5.52474439e-01 -9.87860113e-02
-7.78397024e-01 -4.64372456e-01 3.21110994e-01 1.84749210e+00
-3.35774928e-01 -4.88069147e-01 -7.65821695e-01 -8.88167143e-01
-3.03162307e-01 1.19335878e+00 -2.72251070e-01 2.14167289e-04
-5.32329381e-01 -3.74086857e-01 5.43227255e-01 -3.37765142e-02
4.30193365e-01 -1.16081011e+00 -5.28608143e-01 4.79153693e-01
-4.94017690e-01 -1.14736152e+00 -4.26719338e-01 2.99288929e-01
-3.05922955e-01 -1.21660125e+00 -1.87877297e-01 -6.55552983e-01
4.06497180e-01 2.20993742e-01 1.18858051e+00 1.43782407e-01
-1.12011423e-03 5.85314333e-01 -6.46578074e-01 -2.30800033e-01
-1.31322813e+00 3.16128939e-01 1.20451242e-01 7.48580992e-02
8.47290218e-01 -2.15923890e-01 -1.62034824e-01 5.62268376e-01
-5.05720973e-01 -1.54758871e-01 1.34505751e-02 5.81755340e-01
-4.85306978e-01 4.88798767e-02 6.84111774e-01 -9.54797149e-01
1.46396971e+00 -3.47043276e-01 -4.21831846e-01 1.91099793e-01
-4.52753037e-01 1.26351684e-01 8.78224611e-01 -1.12248287e-01
-1.25035489e+00 2.00374246e-01 -5.28377116e-01 2.39999499e-02
-7.98048556e-01 2.60982424e-01 -1.20793998e-01 3.41238499e-01
7.89077103e-01 2.04559281e-01 2.27513224e-01 -5.60148418e-01
4.70068276e-01 1.52556539e+00 5.03430605e-01 -5.47437072e-01
3.02715451e-01 5.35180159e-02 -7.01286912e-01 -1.24713528e+00
-4.51449692e-01 -1.25903773e+00 -5.65097809e-01 -3.40052485e-01
9.06776786e-01 -5.85114658e-01 -1.13077235e+00 1.96343422e-01
-1.32615888e+00 -2.40606278e-01 -2.55796481e-02 1.93727896e-01
-6.04013801e-01 3.58830750e-01 -5.07216036e-01 -1.30087221e+00
-4.10095215e-01 -1.22976899e+00 8.22697163e-01 2.59154797e-01
-1.17985284e+00 -9.73814666e-01 1.27708599e-01 6.94894135e-01
4.70155478e-01 -5.94266534e-01 1.13953733e+00 -1.96409535e+00
1.11355484e-01 -2.54285991e-01 -1.02137392e-02 7.54556134e-02
4.04479831e-01 -2.94367135e-01 -1.08897257e+00 -1.16551906e-01
4.44557756e-01 -4.98024821e-01 3.10317092e-02 4.24840581e-03
1.34393066e-01 -5.36104441e-01 -2.10323870e-01 -2.41287485e-01
7.59233177e-01 6.79710925e-01 1.06204465e-01 -1.34648934e-01
2.95031488e-01 1.25660741e+00 6.55681610e-01 7.07955718e-01
5.11138380e-01 9.66035664e-01 2.53405601e-01 2.97871143e-01
1.03916951e-01 2.28115432e-02 1.94773808e-01 6.83251739e-01
4.90024745e-01 -3.65767062e-01 -8.81485522e-01 5.05490839e-01
-1.99063718e+00 -8.06441009e-01 -1.23090960e-01 1.88683248e+00
8.66570055e-01 -1.23177273e-02 5.63919842e-01 -4.57010828e-02
7.95247197e-01 -2.72466578e-02 -3.56775105e-01 -5.49435019e-01
6.01181924e-01 -3.04572433e-01 -9.02940258e-02 6.65986121e-01
-9.58493769e-01 1.31897902e+00 5.91925144e+00 4.35413182e-01
-1.02504039e+00 -1.54402666e-02 5.15085816e-01 4.83700246e-01
1.32938549e-01 -6.42304569e-02 -8.88085663e-01 2.08981872e-01
1.09060168e+00 -4.29210216e-01 5.46301007e-01 1.16435409e+00
4.88442034e-01 -9.38106328e-02 -1.39423859e+00 4.60160106e-01
7.33903944e-02 -1.01484692e+00 -2.72591442e-01 -3.99146155e-02
-7.98702687e-02 -2.37770200e-01 -4.13285911e-01 4.74504709e-01
9.74935949e-01 -7.30592668e-01 3.60862702e-01 -1.40993908e-01
3.69627386e-01 -4.19362903e-01 5.40538013e-01 6.55304372e-01
-8.03328454e-01 -8.59134048e-02 -1.87702358e-01 -1.47916272e-01
2.45842680e-01 -9.95725691e-02 -1.90786099e+00 1.10262074e-01
2.22190052e-01 8.03861097e-02 -4.00193065e-01 4.61071640e-01
5.06202057e-02 6.41483724e-01 -2.81372339e-01 -4.27180022e-01
1.82659045e-01 -1.45239756e-01 6.63552165e-01 1.39739227e+00
-2.12771431e-01 3.94544899e-01 5.58047950e-01 7.69260585e-01
1.45007685e-01 2.57300824e-01 -6.00313365e-01 -4.13628012e-01
5.35348535e-01 1.48340189e+00 -8.72995615e-01 -3.81595284e-01
-2.05641702e-01 8.60185742e-01 3.02129507e-01 1.30894454e-02
-3.62508923e-01 -4.80718255e-01 7.80281842e-01 -2.10553393e-01
-2.24821940e-01 -2.63213426e-01 -2.68978328e-01 -7.37429261e-01
-2.94212222e-01 -1.25515044e+00 5.73172212e-01 -3.64292890e-01
-1.31598687e+00 9.60214674e-01 -2.01454163e-02 -8.58702004e-01
-1.05440795e+00 -3.81623358e-01 -6.82284057e-01 1.00101399e+00
-6.56769216e-01 -1.00772810e+00 -1.78825840e-01 7.31723785e-01
1.08564627e+00 -4.67766613e-01 1.17119420e+00 -1.13270007e-01
-4.41364855e-01 1.78706065e-01 -3.02160650e-01 2.86925465e-01
8.40391338e-01 -1.34082448e+00 3.03093046e-01 4.89691079e-01
3.12566012e-03 9.88383353e-01 1.24166393e+00 -6.85124397e-01
-1.40347302e+00 -8.64649773e-01 1.18275905e+00 -3.45244318e-01
7.40189016e-01 -6.51976228e-01 -1.00533724e+00 7.05801785e-01
4.94746894e-01 -1.01327252e+00 8.55030060e-01 3.92933697e-01
-3.45087834e-02 2.88211137e-01 -1.16247451e+00 7.14092612e-01
5.66629171e-01 -4.52869862e-01 -9.98830199e-01 6.48853421e-01
7.56737769e-01 -1.19575202e-01 -7.11259425e-01 -2.29716271e-01
1.54446959e-01 -1.07213044e+00 6.10415578e-01 -7.82365978e-01
-1.01104416e-01 -1.78005341e-02 -1.37814283e-01 -1.33114970e+00
-2.21995205e-01 -9.98117387e-01 5.55616856e-01 1.44547260e+00
3.38106811e-01 -4.53122616e-01 8.70207489e-01 1.45660508e+00
-2.99165726e-01 7.52054676e-02 -8.04416358e-01 -3.24529767e-01
-2.96998709e-01 -5.57286084e-01 4.09696639e-01 9.81686831e-01
1.05411553e+00 1.18034756e+00 -4.77092415e-01 3.26033831e-02
2.08236441e-01 1.64716229e-01 1.17287350e+00 -1.29608881e+00
-1.54257178e-01 -4.71603066e-01 -1.54427171e-01 -1.32794940e+00
2.77109325e-01 -5.96560001e-01 5.93946338e-01 -1.23128092e+00
-2.13087276e-01 -5.18388987e-01 5.27215838e-01 4.92004871e-01
-1.59950718e-01 -3.00975591e-01 1.37649357e-01 5.42735159e-01
-8.03964972e-01 3.48941028e-01 4.35556501e-01 2.10906728e-03
-7.46052861e-01 3.21798563e-01 -7.24987090e-01 7.54279971e-01
8.98142159e-01 -4.99438345e-01 -4.35456693e-01 2.18970627e-01
-2.42466614e-01 3.47065240e-01 -2.89374918e-01 -9.79467809e-01
5.94375253e-01 -3.07601511e-01 -3.01259488e-01 -4.77900326e-01
4.51109886e-01 -1.01772463e+00 -1.63173392e-01 3.14054281e-01
-8.98498833e-01 8.81690457e-02 -5.79520985e-02 2.58010179e-01
-1.98440943e-02 -6.85850799e-01 6.68012261e-01 -3.27490658e-01
-9.32280242e-01 -2.78835714e-01 -1.11080360e+00 1.49044111e-01
1.06455302e+00 1.13616176e-01 -4.01377559e-01 -9.51927364e-01
-6.90062523e-01 3.60429227e-01 3.97120118e-01 5.71288645e-01
4.01251793e-01 -5.31662226e-01 -6.87461495e-01 1.14520848e-01
3.37838769e-01 -5.56088448e-01 -1.30689636e-01 4.30500358e-01
-3.40120375e-01 7.23135114e-01 -4.43349108e-02 -6.23863578e-01
-1.72631264e+00 3.71819288e-01 1.54011548e-01 -4.21994358e-01
-4.78967607e-01 5.50563097e-01 -2.19615847e-02 -6.24533951e-01
4.77370083e-01 3.32779624e-02 -5.19195020e-01 9.01016891e-02
5.75397372e-01 2.32491255e-01 1.68726757e-01 -8.92335415e-01
-4.11372632e-01 -3.85062903e-01 -4.12773967e-01 -4.78791803e-01
9.53841269e-01 -4.23185349e-01 -2.15314075e-01 5.42097986e-01
1.09481919e+00 -2.63197124e-01 -6.58768952e-01 -1.26776904e-01
5.32288611e-01 -1.59146458e-01 -3.21861356e-01 -7.86525071e-01
-2.45028824e-01 6.49999142e-01 2.21807629e-01 1.16300488e+00
6.66293859e-01 1.31166294e-01 5.71964741e-01 7.52203107e-01
5.36195576e-01 -1.12181413e+00 1.72997117e-01 1.02267182e+00
8.43351781e-01 -1.46795881e+00 -3.07185799e-01 -4.43725675e-01
-1.11781144e+00 1.09729350e+00 7.06459939e-01 4.22068894e-01
4.06506777e-01 7.45697692e-02 5.81011772e-01 -5.21314919e-01
-9.33230877e-01 -3.12486470e-01 6.62739575e-02 6.40178800e-01
7.05893338e-01 -8.92333873e-03 -2.26035431e-01 3.09954494e-01
-3.68761629e-01 -4.91604775e-01 4.43681657e-01 9.84489858e-01
-7.44210064e-01 -1.05828321e+00 -3.54926497e-01 4.31425005e-01
-2.46843785e-01 -7.75761753e-02 -1.10558343e+00 5.64490855e-01
-5.89297175e-01 1.93325460e+00 2.44594049e-02 -7.54906297e-01
4.13298130e-01 5.66049457e-01 -2.37810731e-01 -1.10556138e+00
-1.20199358e+00 1.30480155e-01 7.19612479e-01 -2.87789196e-01
-1.64129272e-01 -5.80134153e-01 -1.47343147e+00 -3.37308466e-01
-4.62807536e-01 7.55105078e-01 8.11286867e-01 1.07889163e+00
4.50169742e-01 1.15309790e-01 8.40875030e-01 -5.73158801e-01
-7.62698174e-01 -1.41612077e+00 -3.41042727e-01 8.57219934e-01
-5.20580895e-02 -2.08999559e-01 -2.32518256e-01 2.18429491e-01] | [12.70824909210205, 7.814077854156494] |
d9fae795-fc52-4486-a374-33633e9d77aa | optimized-directed-roadmap-graph-for-multi | 2003.12924 | null | https://arxiv.org/abs/2003.12924v1 | https://arxiv.org/pdf/2003.12924v1.pdf | Optimized Directed Roadmap Graph for Multi-Agent Path Finding Using Stochastic Gradient Descent | We present a novel approach called Optimized Directed Roadmap Graph (ODRM). It is a method to build a directed roadmap graph that allows for collision avoidance in multi-robot navigation. This is a highly relevant problem, for example for industrial autonomous guided vehicles. The core idea of ODRM is, that a directed roadmap can encode inherent properties of the environment which are useful when agents have to avoid each other in that same environment. Like Probabilistic Roadmaps (PRMs), ODRM's first step is generating samples from C-space. In a second step, ODRM optimizes vertex positions and edge directions by Stochastic Gradient Descent (SGD). This leads to emergent properties like edges parallel to walls and patterns similar to two-lane streets or roundabouts. Agents can then navigate on this graph by searching their path independently and solving occurring agent-agent collisions at run-time. Using the graphs generated by ODRM compared to a non-optimized graph significantly fewer agent-agent collisions happen. We evaluate our roadmap with both, centralized and decentralized planners. Our experiments show that with ODRM even a simple centralized planner can solve problems with high numbers of agents that other multi-agent planners can not solve. Additionally, we use simulated robots with decentralized planners and online collision avoidance to show how agents are a lot faster on our roadmap than on standard grid maps. | ['Marc Toussaint', 'Christian Henkel'] | 2020-03-29 | null | null | null | null | ['multi-agent-path-finding'] | ['playing-games'] | [-1.57584339e-01 7.49028623e-01 3.29443961e-01 1.29284948e-01
-3.70674551e-01 -5.52544057e-01 8.13903391e-01 3.42500865e-01
-5.18539488e-01 1.16662061e+00 -1.14150919e-01 -2.85661519e-01
-5.03482640e-01 -1.35792160e+00 -7.64772415e-01 -7.60152161e-01
-8.87562096e-01 1.48138523e+00 6.73885942e-01 -7.77041733e-01
3.72938246e-01 6.33798301e-01 -1.35905302e+00 -4.19558823e-01
7.62654722e-01 5.63843362e-03 8.56023014e-01 9.28412855e-01
-1.50692374e-01 5.74712813e-01 -3.36401552e-01 2.45773748e-01
3.75961870e-01 -1.07399812e-02 -8.95691752e-01 -1.72973335e-01
-3.21370393e-01 4.34283130e-02 -2.16752067e-01 1.01706910e+00
2.46467918e-01 3.77980143e-01 9.27771032e-01 -1.65241015e+00
-1.14998214e-01 7.32910693e-01 -4.35853809e-01 -4.21758831e-01
4.71948534e-01 3.42066914e-01 5.69604099e-01 -4.80012387e-01
9.25317883e-01 1.73037398e+00 4.03843880e-01 2.85395384e-01
-1.24549949e+00 -2.35330582e-01 2.21787527e-01 -3.62622226e-03
-1.14836395e+00 1.06000118e-01 3.88246506e-01 -3.24181437e-01
9.79735494e-01 -1.01071566e-01 5.04999697e-01 7.14247286e-01
9.75784540e-01 2.89572120e-01 8.58879268e-01 -2.11150363e-01
6.07164323e-01 -1.11200266e-01 -1.93052709e-01 1.01264334e+00
5.40669143e-01 3.59238684e-01 -1.63363710e-01 -1.43752888e-01
8.88727605e-01 -3.48918766e-01 9.79182273e-02 -1.19975543e+00
-1.49873710e+00 1.02800190e+00 6.68241620e-01 -7.24730864e-02
-5.64410686e-01 4.95512515e-01 2.50712782e-01 3.13992769e-01
-1.99725330e-01 6.18264437e-01 -3.09689581e-01 -6.40641078e-02
8.97287428e-02 7.91231096e-01 1.08514559e+00 1.12551439e+00
1.41274321e+00 -1.10548161e-01 4.13049072e-01 4.59922254e-01
5.51291287e-01 8.78602803e-01 7.94265345e-02 -1.22061491e+00
5.31678855e-01 4.63598669e-01 4.53054607e-01 -1.26030612e+00
-1.06288612e+00 6.21502697e-02 -8.48381042e-01 1.12808859e+00
4.33930755e-01 -5.40510356e-01 -8.24826300e-01 1.58829403e+00
5.53748667e-01 -2.21595556e-01 5.83713233e-01 6.82543993e-01
3.17641705e-01 7.53332138e-01 -3.01925004e-01 1.75024703e-01
1.15399361e+00 -1.13256907e+00 -3.38044018e-01 -4.76295799e-01
8.69859874e-01 -2.01224685e-01 6.61100805e-01 2.85175383e-01
-8.16633523e-01 -7.70996064e-02 -1.03797662e+00 4.00972873e-01
-5.69726765e-01 -5.18413186e-01 6.04580104e-01 3.30271542e-01
-1.42239118e+00 6.42639935e-01 -8.98744047e-01 -5.17576814e-01
8.13769326e-02 3.95884722e-01 -4.49971139e-01 4.60562063e-03
-9.51534986e-01 1.24784279e+00 4.22589093e-01 -2.49074668e-01
-1.31113219e+00 -1.00894138e-01 -1.10838747e+00 -3.00813377e-01
4.81154203e-01 -7.29826391e-01 1.11223054e+00 -7.01084584e-02
-1.81556702e+00 3.05143535e-01 -1.42335683e-01 -6.64216459e-01
5.05126119e-01 4.66429256e-02 1.36417419e-01 -1.42377298e-02
5.91827273e-01 9.09459054e-01 4.20297384e-01 -1.66893923e+00
-8.81162465e-01 -2.03794152e-01 1.88155249e-01 6.12684250e-01
6.62159324e-01 -5.50954282e-01 2.97393464e-02 2.42123887e-01
1.62779093e-01 -1.37667608e+00 -1.16232955e+00 -3.41820687e-01
-7.65760064e-01 -3.94236922e-01 6.69281960e-01 1.28762126e-01
4.14600223e-01 -1.61779022e+00 4.75076437e-01 6.40179694e-01
1.76747561e-01 -2.57800847e-01 -4.00937200e-01 9.96338725e-01
4.36282396e-01 1.06774710e-01 -3.51162463e-01 -1.05252333e-01
2.45638773e-01 5.06356597e-01 -7.78150791e-03 5.22732794e-01
-1.90946594e-01 7.46666312e-01 -1.40368450e+00 -3.26146632e-01
3.64702255e-01 7.39828944e-02 -4.20132905e-01 -3.13109875e-01
-4.43221629e-01 4.07775462e-01 -6.33620799e-01 6.14318773e-02
6.69203401e-01 1.77618340e-02 3.57820600e-01 5.31009197e-01
-6.52283669e-01 1.84066862e-01 -1.43842888e+00 1.52895951e+00
-4.82978135e-01 6.69506133e-01 2.43845016e-01 -7.27308393e-01
1.10623944e+00 -1.37551084e-01 2.74895996e-01 -6.27070844e-01
-1.15853108e-01 3.10639352e-01 -9.35184509e-02 -1.12891987e-01
7.20287144e-01 3.77516627e-01 -3.90332490e-01 5.53667784e-01
-5.36796212e-01 -7.26481259e-01 3.68021637e-01 4.21696603e-01
1.45264900e+00 6.56327372e-03 1.45216852e-01 -6.60524368e-01
4.01843488e-01 7.04955935e-01 4.38371181e-01 1.07158351e+00
2.29500607e-02 2.22828284e-01 4.61789191e-01 -7.69584358e-01
-9.53425109e-01 -1.24934363e+00 3.87326568e-01 5.94154775e-01
9.43351507e-01 -4.20493722e-01 -5.91534019e-01 -4.19028133e-01
8.80394578e-02 7.57541597e-01 -5.16606092e-01 1.45056069e-01
-8.56070936e-01 -6.31043017e-01 3.53653245e-02 -2.39538804e-01
5.80645442e-01 -1.32560027e+00 -1.21206450e+00 5.53579926e-01
7.04307780e-02 -7.20907748e-01 -3.78503017e-02 3.67057383e-01
-5.02161682e-01 -1.21701956e+00 -2.73603022e-01 -1.03531289e+00
9.02389288e-01 5.85485697e-01 1.09678018e+00 -2.31588975e-01
-1.18166119e-01 3.80424351e-01 -3.19033563e-01 -4.49914038e-01
-7.23042071e-01 2.38382950e-01 3.09098780e-01 -6.69659376e-01
-2.36974210e-01 -7.55913615e-01 -3.75502050e-01 6.27242863e-01
-2.68946201e-01 2.77992338e-01 6.59818649e-01 4.42732751e-01
5.63217342e-01 6.67657316e-01 3.78262788e-01 -5.35916150e-01
8.87835979e-01 -6.27817810e-01 -1.09863389e+00 -1.67986080e-01
-3.73359770e-01 5.12634695e-01 5.64119339e-01 -9.51273069e-02
-5.09661913e-01 3.24943751e-01 2.81990319e-01 2.80033529e-01
-9.95279700e-02 4.76019621e-01 1.00418746e-01 -1.66814461e-01
7.15350747e-01 6.26900867e-02 3.33999515e-01 5.31337224e-02
5.26129186e-01 6.68004304e-02 2.68974900e-01 -6.49309158e-01
1.03912663e+00 6.63759053e-01 6.12520576e-01 -9.20119405e-01
2.10749373e-01 -1.49260005e-02 -4.20866579e-01 -2.69997746e-01
8.42347443e-01 -3.65104944e-01 -1.22970307e+00 3.49679708e-01
-1.33332717e+00 -1.05551159e+00 -2.88882136e-01 3.10643435e-01
-9.80400383e-01 1.12538621e-01 -2.92458534e-01 -9.68733251e-01
1.83737606e-01 -1.35113847e+00 8.32508802e-01 2.51670152e-01
-1.73561990e-01 -1.04160273e+00 7.25284159e-01 -5.62386751e-01
4.36915547e-01 5.37563026e-01 7.31371999e-01 -4.45123136e-01
-1.05023849e+00 1.99698791e-01 -3.58985760e-03 -7.95487583e-01
2.58340597e-01 -1.59530938e-01 -8.18023980e-02 -2.22564846e-01
-5.17426968e-01 -3.98812443e-02 3.73223573e-01 5.49968660e-01
9.01204646e-02 -4.25556123e-01 -1.01926553e+00 9.01942030e-02
1.40795326e+00 3.65609288e-01 7.02399373e-01 1.12720740e+00
5.38441300e-01 9.05272603e-01 8.75356853e-01 3.11323076e-01
8.92037392e-01 7.80924022e-01 1.07140911e+00 8.00782591e-02
2.31674135e-01 -2.18878359e-01 5.57279587e-01 2.73949951e-01
5.42272255e-02 -2.70131856e-01 -1.20383549e+00 6.89456761e-01
-2.36000824e+00 -7.26872265e-01 -6.30075395e-01 2.14919019e+00
2.58668482e-01 1.31691903e-01 1.62534744e-01 -2.82073021e-01
7.97148883e-01 -7.51483142e-02 -5.09487867e-01 -5.21857738e-01
-1.24915712e-01 -3.37180406e-01 1.17614472e+00 1.25491333e+00
-9.33832645e-01 1.13445556e+00 6.19058609e+00 5.01935124e-01
-4.48371321e-01 -2.58910894e-01 5.23807481e-02 4.73490804e-01
-2.45135292e-01 1.62963256e-01 -8.44743311e-01 2.38784552e-01
8.24056923e-01 -2.17714265e-01 8.98571849e-01 9.15017664e-01
4.34972197e-01 -5.28841794e-01 -8.10376227e-01 7.30313301e-01
-5.95558465e-01 -1.43788099e+00 -1.15187928e-01 5.05781174e-01
7.86208212e-01 5.14495194e-01 -2.37251654e-01 1.86782002e-01
1.77743375e+00 -1.05468631e+00 7.41796672e-01 3.32319498e-01
4.57362570e-02 -8.94246221e-01 6.14530027e-01 5.96105754e-01
-1.53053606e+00 -2.50363015e-02 -4.53982502e-01 -1.47221610e-01
7.52465487e-01 4.27668989e-01 -1.43359256e+00 7.42994010e-01
5.88524520e-01 3.29061270e-01 -4.46642488e-02 1.12424850e+00
-4.09051389e-01 -3.68735880e-01 -6.95509255e-01 -5.03717959e-01
6.49654865e-01 -6.78309917e-01 1.06572449e+00 8.00609529e-01
3.94255072e-01 -3.42490822e-01 4.33913708e-01 7.42442727e-01
6.47835314e-01 -3.09960544e-01 -1.29069471e+00 4.69986886e-01
7.76808560e-01 1.20877016e+00 -1.16384947e+00 -1.50063112e-01
1.35903433e-01 7.14214444e-01 4.03505594e-01 3.72734189e-01
-7.31604576e-01 -6.44844174e-01 8.92220199e-01 -9.68693271e-02
1.22974716e-01 -8.31425190e-01 -3.30913477e-02 -3.73096049e-01
-5.71325243e-01 -5.89145005e-01 -1.40782937e-01 -7.18773425e-01
-8.31261873e-01 6.69718266e-01 5.00953794e-02 -1.00798082e+00
-7.01790392e-01 -5.81096947e-01 -5.96671760e-01 6.46122158e-01
-1.43170607e+00 -7.90160954e-01 -2.46569678e-01 4.74598527e-01
5.29548168e-01 -1.74657017e-01 9.08553004e-01 -4.92468596e-01
-2.03967392e-01 -3.47883999e-01 2.25059450e-01 -2.31169954e-01
1.26151919e-01 -1.56162727e+00 9.31120336e-01 5.91719985e-01
-1.25960782e-01 2.98900962e-01 1.03607607e+00 -1.04012227e+00
-1.60949230e+00 -1.08210492e+00 4.23543036e-01 -3.83885920e-01
6.83659077e-01 -3.34562331e-01 -4.28283393e-01 6.60407245e-01
3.75585914e-01 -5.50982058e-01 -3.20322931e-01 -1.32031485e-01
3.08163911e-01 1.55397713e-01 -1.17029977e+00 1.08341634e+00
1.19203532e+00 2.21930429e-01 -2.69310534e-01 6.23056412e-01
7.26994455e-01 -3.03105414e-01 -2.28983387e-01 1.17069431e-01
2.53349304e-01 -9.79383469e-01 9.69500482e-01 -1.00576445e-01
-1.19943522e-01 -7.68894136e-01 -5.31045347e-02 -2.09083080e+00
-4.04373586e-01 -1.12715197e+00 5.30914307e-01 6.25049949e-01
6.38681829e-01 -1.31587946e+00 8.46242964e-01 2.52258897e-01
-3.74357373e-01 -3.95129234e-01 -9.20877576e-01 -1.07212448e+00
7.57614598e-02 -1.20161593e-01 6.42886937e-01 4.25849497e-01
1.89590454e-01 3.03828597e-01 3.13297212e-02 6.48384690e-01
9.64111328e-01 -1.56313375e-01 1.26732731e+00 -1.36669731e+00
9.63105261e-02 -5.33267379e-01 -3.62764299e-01 -9.77582812e-01
1.93078235e-01 -7.08951116e-01 5.44542849e-01 -2.20453000e+00
-3.56835783e-01 -1.11697233e+00 5.46342432e-01 2.89805353e-01
4.64285195e-01 -3.09688777e-01 9.03394353e-03 1.71497956e-01
-6.90625548e-01 5.34725666e-01 1.27382386e+00 -1.01805933e-01
-6.98389411e-01 -2.03431740e-01 -1.29661396e-01 7.83389747e-01
1.16109860e+00 -5.35945237e-01 -6.25355244e-01 -3.67659032e-01
6.26747727e-01 1.46583408e-01 1.99852079e-01 -1.15740371e+00
5.84946692e-01 -4.07476068e-01 -4.77647066e-01 -6.05917752e-01
3.90021950e-01 -6.66945457e-01 5.29377639e-01 9.48600233e-01
1.48817614e-01 6.09746456e-01 1.96910053e-01 8.20799828e-01
2.48094618e-01 -2.80247688e-01 5.20684004e-01 -5.02009988e-01
-7.86469817e-01 5.93438819e-02 -1.11607945e+00 -1.12018421e-01
1.37594283e+00 -2.53733903e-01 -6.21752739e-01 -5.79309046e-01
-6.47966981e-01 8.82074654e-01 8.68690252e-01 1.13868125e-01
7.33962059e-01 -1.13448083e+00 -8.33013952e-01 -7.06812814e-02
-2.82648385e-01 5.46238005e-01 -1.54551163e-01 5.22380948e-01
-1.12236845e+00 3.94493103e-01 -4.46233124e-01 -6.25888705e-01
-9.45921361e-01 4.80984211e-01 3.26140493e-01 -4.66162682e-01
-9.64295506e-01 5.72943449e-01 4.32802439e-01 -8.98657918e-01
-1.49876073e-01 -3.11953366e-01 -2.74542361e-01 -3.53557795e-01
3.12506080e-01 5.74724138e-01 -4.73754555e-01 -5.05818963e-01
-5.36237895e-01 7.36801267e-01 1.43939972e-01 -6.74827278e-01
1.39194155e+00 -3.99514198e-01 -2.52450168e-01 2.43545771e-01
5.08886456e-01 1.17488541e-01 -1.33019376e+00 4.66552854e-01
1.65321335e-01 -1.54024422e-01 -3.77968103e-01 -2.93262213e-01
-5.97570240e-01 2.66702682e-01 2.36781269e-01 7.56889462e-01
2.01433688e-01 1.14146419e-01 3.52552056e-01 7.73479640e-01
1.32830226e+00 -9.19647634e-01 1.41040057e-01 1.02595425e+00
9.56465185e-01 -9.44103837e-01 -2.09560916e-01 -5.49540639e-01
-6.65562332e-01 1.06881094e+00 3.33473742e-01 -6.24493122e-01
3.81097823e-01 4.72914189e-01 -8.31432194e-02 -4.11989510e-01
-6.17935479e-01 -4.34461981e-01 -6.44156873e-01 1.26108384e+00
-6.43993437e-01 2.39768282e-01 1.80881292e-01 -2.84550846e-01
-5.86826801e-01 -4.83499676e-01 1.05956948e+00 9.56387281e-01
-1.04082608e+00 -1.11202848e+00 -5.06692469e-01 8.23638290e-02
4.83312458e-01 2.98682570e-01 -2.02398077e-01 1.11657369e+00
-3.10523033e-01 1.17073178e+00 1.43980637e-01 -3.80464584e-01
2.68792391e-01 -5.26874065e-01 3.15356523e-01 -6.08768165e-01
-4.78063244e-03 -5.90166628e-01 5.36592185e-01 -7.76535988e-01
-1.60441056e-01 -7.24532127e-01 -1.97485304e+00 -3.53595346e-01
1.84555706e-02 5.52929282e-01 9.04981136e-01 6.96566939e-01
5.35439312e-01 4.28397477e-01 5.38549304e-01 -1.49280727e+00
2.84996629e-02 -4.25771087e-01 -4.71405983e-01 -2.69621015e-01
4.15647686e-01 -1.04741073e+00 -3.44188690e-01 -5.26852906e-01] | [4.932121753692627, 1.5520479679107666] |
9c57e292-1f8a-4ca2-861c-0fe42f7bfed0 | synthetic-human-model-dataset-for-skeleton | 1903.02679 | null | http://arxiv.org/abs/1903.02679v1 | http://arxiv.org/pdf/1903.02679v1.pdf | Synthetic Human Model Dataset for Skeleton Driven Non-rigid Motion Tracking and 3D Reconstruction | We introduce a synthetic dataset for evaluating non-rigid 3D human
reconstruction based on conventional RGB-D cameras. The dataset consist of
seven motion sequences of a single human model. For each motion sequence
per-frame ground truth geometry and ground truth skeleton are given. The
dataset also contains skinning weights of the human model. More information
about the dataset can be found at:
https://research.csiro.au/robotics/our-work/databases/synthetic-human-model-dataset/ | ['Peyman Moghadam', 'Shafeeq Elanattil'] | 2019-03-07 | null | null | null | null | ['3d-human-reconstruction'] | ['computer-vision'] | [-6.18558116e-02 1.12575695e-01 -2.49314949e-01 -2.74492025e-01
-4.70371574e-01 -3.08822691e-01 3.19359928e-01 -4.42625105e-01
-4.31823879e-01 2.75263876e-01 3.26643497e-01 2.24115476e-01
4.51717973e-01 -3.91671151e-01 -6.89204991e-01 -4.30635482e-01
1.99341834e-01 6.83847666e-01 3.98275644e-01 -3.63791846e-02
-3.41754667e-02 6.68828785e-01 -1.15238976e+00 -1.05851904e-01
-1.32476300e-01 7.09763885e-01 2.46017680e-01 1.07559109e+00
8.60830069e-01 5.40389597e-01 -1.20099992e-01 -5.37538350e-01
7.79479980e-01 -4.96999204e-01 -8.46216619e-01 5.21389544e-01
3.39466840e-01 -6.45808399e-01 -1.02141035e+00 9.08283293e-01
8.73666644e-01 1.92415476e-01 3.98508340e-01 -1.22373676e+00
-1.18339874e-01 -9.52794179e-02 -5.56126654e-01 7.28253275e-02
1.23035288e+00 3.41102898e-01 4.49710369e-01 -8.97791386e-01
1.06005192e+00 1.27870345e+00 6.41766906e-01 9.91947353e-01
-9.61496592e-01 -4.59532619e-01 -4.98226792e-01 -7.92529956e-02
-1.39166737e+00 -6.27669990e-01 8.28809202e-01 -4.85614896e-01
7.38932788e-01 2.30010062e-01 1.33066392e+00 1.41550910e+00
9.00915042e-02 9.35409427e-01 6.58133328e-01 -4.33763146e-01
-3.85273397e-02 -5.35974741e-01 -9.16646346e-02 1.10836565e+00
3.67248386e-01 2.92115808e-01 -5.84828377e-01 1.22385986e-01
1.63300920e+00 1.44705728e-01 -2.40681916e-01 -9.42301095e-01
-1.66980290e+00 2.94710338e-01 2.51571685e-01 -4.27990332e-02
-4.40814644e-01 4.01113570e-01 4.06079203e-01 -8.24868307e-02
-1.85615998e-02 -2.43547350e-01 -4.75822896e-01 -3.75269204e-01
-6.31899774e-01 5.93942463e-01 4.91176426e-01 1.33646047e+00
5.70048809e-01 2.49054357e-02 6.98566914e-01 3.90401483e-01
2.98498541e-01 6.73770964e-01 5.12539268e-01 -1.89594710e+00
3.15867215e-01 4.95552570e-01 4.79609251e-01 -9.78555620e-01
-4.82991517e-01 3.92232507e-01 -7.66355097e-01 9.59142670e-02
5.57939231e-01 1.73380636e-02 -9.82787490e-01 1.12562573e+00
6.86197221e-01 -2.47845110e-02 -1.61757126e-01 1.42249334e+00
1.04142594e+00 -7.12201968e-02 -1.85703978e-01 2.25378588e-01
1.16729200e+00 -9.96140420e-01 -5.46388328e-01 -2.46818796e-01
2.79776961e-01 -6.30932689e-01 9.54328537e-01 3.78581971e-01
-1.44469261e+00 -5.76578021e-01 -9.78440821e-01 -4.03910309e-01
1.03574671e-01 1.12621419e-01 3.49875838e-01 4.41822022e-01
-8.06642532e-01 4.50031847e-01 -1.42683184e+00 -8.58334661e-01
1.70618653e-01 1.79328591e-01 -1.05418503e+00 -3.00709695e-01
-7.96475768e-01 1.05582952e+00 3.47933292e-01 7.65394345e-02
-1.18819070e+00 -1.35228828e-01 -1.15765631e+00 -1.07980275e+00
2.15163097e-01 -1.11905730e+00 1.54604387e+00 -6.02310598e-01
-1.61672139e+00 1.64317024e+00 1.77731570e-02 -2.01965183e-01
9.24616516e-01 -5.12103975e-01 -2.08006904e-01 6.31296992e-01
-1.25985473e-01 6.95469439e-01 7.75072217e-01 -1.32651794e+00
1.50158750e-02 -6.58253729e-01 -2.20119327e-01 1.46090403e-01
5.91640413e-01 2.41297096e-01 -8.73735130e-01 -8.36386442e-01
3.22079599e-01 -1.43781543e+00 -4.89836007e-01 2.52257019e-01
-6.71802402e-01 5.29800951e-01 3.21214259e-01 -9.93299782e-01
7.86362588e-01 -1.78280520e+00 5.33782125e-01 -4.63684499e-02
5.11430912e-02 -1.37247995e-01 1.05360597e-01 1.45787358e-01
-2.02821940e-01 -2.67869145e-01 -3.30774188e-01 -5.49174964e-01
-7.25955665e-02 4.33926910e-01 4.14376885e-01 1.00204980e+00
-3.60745072e-01 8.29909027e-01 -8.33720684e-01 -7.61774600e-01
5.77857554e-01 7.01689243e-01 -1.55158892e-01 3.82122159e-01
2.31124029e-01 9.52036858e-01 -4.03166294e-01 1.01549339e+00
2.88797557e-01 -5.84580600e-02 -4.28717025e-02 -9.65021029e-02
3.25669885e-01 -2.23203748e-01 -1.46806657e+00 2.40030980e+00
2.38348588e-01 7.78106675e-02 1.35415256e-01 -5.89307129e-01
6.00933075e-01 5.28168857e-01 8.19188774e-01 -3.78216207e-01
5.79715371e-01 5.27892895e-02 -4.82056469e-01 -6.54141486e-01
4.79869038e-01 -3.26207913e-02 -5.51728569e-02 3.60788137e-01
-8.63216370e-02 -3.72569740e-01 9.17501524e-02 2.30525527e-03
1.04911757e+00 9.68761981e-01 4.15695131e-01 5.10047078e-02
2.80295312e-01 3.63300502e-01 8.13397408e-01 2.36856714e-02
-6.77771091e-01 1.09272349e+00 -1.83974758e-01 -8.28468442e-01
-1.42632592e+00 -1.38153791e+00 -5.46678714e-02 5.10918796e-01
4.82430160e-01 -2.77262688e-01 -9.83238757e-01 -1.97996553e-02
8.31583142e-02 3.39379638e-01 -6.00640476e-01 8.53553861e-02
-7.28286147e-01 -7.55801797e-02 6.47147894e-01 7.60416687e-01
5.71424305e-01 -1.03131533e+00 -1.40253925e+00 -2.43860468e-01
-4.80399877e-01 -1.37847066e+00 -4.68652308e-01 -2.24117801e-01
-1.30701184e+00 -1.47026050e+00 -1.04589307e+00 -4.52237815e-01
1.00235915e+00 2.82919139e-01 1.14787197e+00 2.07148731e-01
-6.67289197e-01 1.05053735e+00 -4.26511526e-01 -1.50271297e-01
-3.52758467e-01 -4.02298570e-01 4.39503729e-01 -5.24811685e-01
2.83504158e-01 -2.69409925e-01 -9.13840890e-01 7.52391934e-01
-6.24892294e-01 3.25405389e-01 1.45530820e-01 2.71783352e-01
1.12626827e+00 -4.32172894e-01 -6.44880176e-01 -3.21363717e-01
-4.35556769e-02 -1.67138070e-01 -4.11691099e-01 -4.89790365e-02
1.98789895e-01 -4.62596416e-01 -3.50052476e-01 -3.90409082e-01
-8.31310987e-01 8.50741684e-01 1.27186194e-01 -8.51889431e-01
-6.79423749e-01 -1.36999100e-01 -3.10633838e-01 1.02090672e-01
7.05463648e-01 -4.17731926e-02 1.00886926e-01 -5.16041338e-01
4.48132217e-01 2.44283900e-01 1.25120008e+00 -8.72641683e-01
6.86387300e-01 6.75087214e-01 3.60618941e-02 -8.44959378e-01
-3.69704455e-01 -7.37762690e-01 -1.54832470e+00 -6.90920174e-01
1.15930617e+00 -1.06094742e+00 -5.39568186e-01 8.30754340e-01
-1.04088032e+00 -8.21107388e-01 -4.07746434e-01 7.75722027e-01
-1.33013344e+00 6.19136631e-01 -9.01742280e-01 -5.47464669e-01
-1.34088010e-01 -1.16605353e+00 1.31769812e+00 9.88087058e-02
-7.09707618e-01 -9.81046498e-01 3.51016372e-01 7.39788771e-01
-2.50451088e-01 9.39567506e-01 -4.07622345e-02 -9.36327130e-02
-1.94354311e-01 -4.28228050e-01 4.73113358e-01 1.58749074e-01
-2.64869202e-02 6.91819116e-02 -6.44455910e-01 -3.08565706e-01
-5.90060465e-02 -4.33684379e-01 3.22631985e-01 3.93018097e-01
5.97463489e-01 -4.51667458e-02 -1.40921712e-01 5.57561576e-01
1.20217216e+00 -3.73780400e-01 7.88190663e-01 6.20243788e-01
9.01793420e-01 6.02579653e-01 7.64816165e-01 6.92181647e-01
4.48043555e-01 8.28666747e-01 4.22720462e-01 8.35815147e-02
-1.73629627e-01 -3.11586410e-01 5.20494759e-01 7.94268548e-01
-8.62809181e-01 1.15332328e-01 -1.33062744e+00 5.37163138e-01
-1.67233300e+00 -1.02554393e+00 -4.51037139e-01 2.18298697e+00
6.90983236e-01 -2.14125514e-01 8.23929787e-01 4.57504869e-01
6.77563071e-01 -4.38709296e-02 -6.12647176e-01 4.52600792e-02
-9.80966091e-02 -1.20416440e-01 5.97665906e-01 5.87853014e-01
-9.94210720e-01 8.70957553e-01 5.71899652e+00 1.03895836e-01
-6.56492352e-01 1.18880704e-01 -1.12501614e-01 -3.05031300e-01
1.70873567e-01 -7.70820752e-02 -5.54306984e-01 1.80688381e-01
1.00485897e+00 -5.58151677e-02 2.24759385e-01 7.86455989e-01
3.57919633e-01 -4.31650490e-01 -9.40732002e-01 1.10284698e+00
7.79020339e-02 -8.50206614e-01 -8.55481774e-02 2.01038271e-01
6.02587759e-01 4.48319875e-02 -2.48147279e-01 -5.00669062e-01
4.00475293e-01 -6.72035336e-01 1.17535663e+00 9.12098706e-01
1.00740647e+00 -3.97857457e-01 3.08597863e-01 3.20532054e-01
-1.10377288e+00 3.52657914e-01 -4.78853762e-01 1.26975387e-01
3.75556141e-01 3.68854940e-01 -3.93552601e-01 6.83493912e-01
1.19693041e+00 9.81465518e-01 -6.30899847e-01 9.78083730e-01
-5.45590281e-01 2.03274012e-01 -2.59137899e-01 6.14669204e-01
-4.65272784e-01 -4.06416655e-01 6.40231669e-01 1.17496836e+00
2.58614868e-01 5.25054693e-01 2.68852830e-01 2.92234242e-01
2.62722015e-01 -5.17942831e-02 -8.93715262e-01 1.49251103e-01
1.91711605e-01 1.04813778e+00 -8.53247941e-01 -1.79855555e-01
-2.47482330e-01 1.49611068e+00 -1.44154862e-01 2.91956693e-01
-7.29582012e-01 2.15733722e-02 5.30200839e-01 4.08578664e-01
-7.60693243e-03 -8.53863895e-01 -2.01437980e-01 -1.26771188e+00
-3.79697084e-02 -7.31040835e-01 4.71321106e-01 -1.26943922e+00
-7.39324033e-01 2.21771240e-01 3.78964484e-01 -1.46310365e+00
-4.22308981e-01 -4.92552072e-01 -6.61031976e-02 5.00337839e-01
-6.28457963e-01 -1.30533957e+00 -7.28682518e-01 1.17380083e+00
6.39354587e-01 6.20962344e-02 9.18027103e-01 4.07297090e-02
-6.14258230e-01 2.61012614e-01 -5.18724561e-01 5.30856371e-01
7.52527177e-01 -1.07066321e+00 7.20812142e-01 8.89731109e-01
-1.21886857e-01 6.19353235e-01 8.66504252e-01 -9.44943190e-01
-1.75653338e+00 -7.46805847e-01 4.32717919e-01 -1.09561348e+00
4.57003474e-01 -9.02734399e-02 -5.05518138e-01 1.25978386e+00
-2.50084512e-02 4.44376379e-01 6.31927073e-01 -7.56569147e-01
4.03386354e-02 5.34802496e-01 -1.37147629e+00 5.20086288e-01
1.47791886e+00 -3.06619316e-01 -8.20318401e-01 2.14756235e-01
3.98106843e-01 -1.24673212e+00 -1.42134607e+00 2.86282927e-01
1.00038826e+00 -9.41238105e-01 1.36934590e+00 -4.28838938e-01
4.82794672e-01 -6.23460472e-01 -6.69911563e-01 -7.67665505e-01
-6.10555010e-03 -5.64103663e-01 -5.12503564e-01 5.82338452e-01
-2.12662145e-01 4.50447723e-02 1.01526380e+00 7.68798530e-01
1.95411995e-01 -3.85412484e-01 -9.45335031e-01 -8.82782578e-01
-2.97075808e-01 -6.45377576e-01 1.62432179e-01 7.94415236e-01
-1.62795588e-01 -1.39986426e-01 -3.98742557e-01 2.90619582e-01
1.04729390e+00 -2.69995719e-01 1.32211959e+00 -1.00133228e+00
-1.68071225e-01 -9.07887146e-02 -7.83458531e-01 -7.89420724e-01
6.45770133e-02 -5.03125250e-01 -5.45402877e-02 -1.51110399e+00
-4.96201366e-02 1.92257300e-01 4.31305826e-01 4.83888716e-01
2.83850700e-01 6.22365117e-01 1.87107190e-01 4.67254221e-01
-7.20365405e-01 3.38711351e-01 1.39039958e+00 2.46628493e-01
9.25349519e-02 8.76525566e-02 1.92303404e-01 1.13324761e+00
7.67967284e-01 -5.09910345e-01 -8.04794859e-03 -4.02891427e-01
-1.20310858e-01 5.28501272e-01 7.52149403e-01 -1.15201986e+00
2.85284698e-01 -2.82135576e-01 8.21400642e-01 -4.58396405e-01
6.00835204e-01 -9.30910766e-01 9.81946647e-01 8.50089967e-01
-7.99076706e-02 5.82565248e-01 1.61090009e-02 4.69136655e-01
1.64427862e-01 -2.35596821e-02 9.67562914e-01 -7.76924372e-01
-9.18490350e-01 3.86192679e-01 -8.85534659e-02 8.61033872e-02
1.03118598e+00 -8.34354758e-01 2.36109242e-01 -3.77873182e-01
-1.10459769e+00 -2.50534508e-02 1.37200928e+00 3.34127933e-01
8.89474213e-01 -1.57773638e+00 -5.98022163e-01 1.25025108e-01
1.08211316e-01 2.52313435e-01 -9.39634722e-03 8.62497807e-01
-1.20581007e+00 2.12748721e-01 -8.76004279e-01 -5.71168005e-01
-1.33599317e+00 4.63326126e-01 6.22846425e-01 3.69194180e-01
-9.11415100e-01 4.05579686e-01 -3.08683723e-01 -6.28607392e-01
2.32266232e-01 -1.68037042e-01 4.96375591e-01 -5.60384631e-01
4.05853689e-01 1.07800210e+00 -2.73742110e-01 -1.39443719e+00
-5.69505095e-01 1.03384197e+00 7.24776089e-01 -5.37486374e-01
1.27136827e+00 -3.27078432e-01 7.63972774e-02 6.15358055e-01
9.74795163e-01 -2.32067004e-01 -1.40621448e+00 -1.78216413e-01
-4.42224890e-02 -6.87401652e-01 -4.70779300e-01 -3.37584585e-01
-1.06786549e+00 7.64848709e-01 5.44298172e-01 -5.51297069e-01
9.24883962e-01 9.87929925e-02 7.89067447e-01 3.14025551e-01
8.18716407e-01 -1.02033806e+00 2.93203712e-01 3.17622870e-01
1.19860733e+00 -1.18813956e+00 5.39594650e-01 -2.70019412e-01
-8.52173805e-01 1.00048208e+00 4.40213025e-01 -3.79447281e-01
5.54764450e-01 2.11899698e-01 3.85873139e-01 -3.51745307e-01
-8.60646144e-02 -1.93529561e-01 3.41717079e-02 1.00902510e+00
3.02288920e-01 2.04846725e-01 1.78963020e-01 3.88478339e-01
-6.58955157e-01 2.00475782e-01 7.57575095e-01 1.13114214e+00
-7.74256513e-02 -9.12894964e-01 -7.40682185e-01 -1.87865749e-01
-4.15114939e-01 6.99444175e-01 -5.48764348e-01 9.95555401e-01
-2.12213457e-01 7.17865109e-01 -1.80148289e-01 -5.23864508e-01
8.59019995e-01 -5.44673949e-03 1.14551568e+00 -3.81780654e-01
-2.59943515e-01 4.47291918e-02 1.48053139e-01 -1.15143239e+00
-8.91752005e-01 -1.13927376e+00 -1.62182021e+00 -5.20201743e-01
3.70096922e-01 -4.22905535e-01 6.51071429e-01 2.43139699e-01
2.43590605e-02 -1.79811090e-01 1.28504962e-01 -1.67944098e+00
-1.73368782e-01 -7.06704378e-01 -6.86131895e-01 6.89416468e-01
1.78278744e-01 -8.11788440e-01 -2.22444013e-01 8.14657569e-01] | [7.114962100982666, -0.9234035611152649] |
76e461e8-d04f-4d07-8874-69eeff161afc | two-steps-to-risk-sensitivity | 2111.06803 | null | https://arxiv.org/abs/2111.06803v1 | https://arxiv.org/pdf/2111.06803v1.pdf | Two steps to risk sensitivity | Distributional reinforcement learning (RL) -- in which agents learn about all the possible long-term consequences of their actions, and not just the expected value -- is of great recent interest. One of the most important affordances of a distributional view is facilitating a modern, measured, approach to risk when outcomes are not completely certain. By contrast, psychological and neuroscientific investigations into decision making under risk have utilized a variety of more venerable theoretical models such as prospect theory that lack axiomatically desirable properties such as coherence. Here, we consider a particularly relevant risk measure for modeling human and animal planning, called conditional value-at-risk (CVaR), which quantifies worst-case outcomes (e.g., vehicle accidents or predation). We first adopt a conventional distributional approach to CVaR in a sequential setting and reanalyze the choices of human decision-makers in the well-known two-step task, revealing substantial risk aversion that had been lurking under stickiness and perseveration. We then consider a further critical property of risk sensitivity, namely time consistency, showing alternatives to this form of CVaR that enjoy this desirable characteristic. We use simulations to examine settings in which the various forms differ in ways that have implications for human and animal planning and behavior. | ['Peter Dayan', 'Chris Gagne'] | 2021-11-12 | null | http://proceedings.neurips.cc/paper/2021/hash/ba530cdf0a884348613f2aaa3a5ba5e8-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/ba530cdf0a884348613f2aaa3a5ba5e8-Paper.pdf | neurips-2021-12 | ['distributional-reinforcement-learning'] | ['methodology'] | [ 2.38409676e-02 3.19751889e-01 -2.51742452e-01 -3.80283326e-01
-4.33592707e-01 -5.63482881e-01 7.50175178e-01 5.65798759e-01
-1.21855521e+00 7.67830789e-01 2.43146703e-01 -6.29582226e-01
-6.57012105e-01 -7.16158807e-01 -4.63046461e-01 -7.66708672e-01
-5.94652772e-01 2.93303668e-01 -1.22670144e-01 -1.73155859e-01
4.03658509e-01 4.60519105e-01 -1.23595953e+00 -6.20279372e-01
7.55021214e-01 7.82600641e-01 2.08990395e-01 2.91140497e-01
3.76749277e-01 5.71475267e-01 -2.51807660e-01 -6.59591377e-01
3.23833078e-01 -2.00160682e-01 -5.06797254e-01 -3.17799270e-01
-2.75613606e-01 -5.29201925e-01 -9.62180942e-02 1.31189013e+00
3.46856385e-01 4.14912611e-01 1.04256320e+00 -1.27753150e+00
-6.28848791e-01 7.57801652e-01 -3.93626750e-01 2.99475849e-01
2.67546505e-01 5.09612679e-01 1.17356408e+00 -1.71655595e-01
2.35824704e-01 1.58565438e+00 4.11512136e-01 5.43829203e-01
-1.65678084e+00 -3.25717479e-01 2.57623225e-01 -1.27907638e-02
-8.83180141e-01 -1.63130209e-01 5.43543518e-01 -7.15458632e-01
6.85689807e-01 8.99524800e-03 5.48025608e-01 1.10630703e+00
8.98548782e-01 3.92506689e-01 1.65553021e+00 -2.09603027e-01
8.39343309e-01 -8.59477147e-02 6.49455041e-02 1.18528470e-01
6.46661282e-01 9.69959497e-01 -1.48138165e-01 -4.07080770e-01
6.07887983e-01 1.32585421e-01 -1.57012716e-02 -4.44389582e-01
-1.10881460e+00 9.33971047e-01 1.28258750e-01 6.90559372e-02
-7.73469985e-01 1.32407203e-01 2.18774557e-01 4.50196952e-01
1.21729359e-01 5.26076734e-01 -2.29946643e-01 -9.43387076e-02
-4.12374377e-01 7.46189296e-01 6.15117431e-01 4.77502882e-01
4.02299970e-01 4.93762866e-02 -2.50865877e-01 3.05615932e-01
5.03286779e-01 4.36784655e-01 2.27159247e-01 -1.09536493e+00
4.13325161e-01 -1.92028940e-01 5.49493313e-01 -8.56279373e-01
-7.51458824e-01 -1.96205109e-01 -4.74286079e-01 8.39259565e-01
6.78599119e-01 -4.65044737e-01 -5.31006813e-01 2.54097795e+00
1.22106902e-01 -3.18766207e-01 2.44097397e-01 7.34670043e-01
-3.33132207e-01 3.47448766e-01 8.26346219e-01 -6.30031526e-01
1.18832374e+00 1.24879302e-02 -4.94523853e-01 -4.47639436e-01
4.38162982e-01 -1.23360820e-01 1.22065151e+00 4.99772996e-01
-1.08839810e+00 2.43747458e-01 -9.37666059e-01 2.96279788e-01
-2.08918732e-02 -9.55792129e-01 7.35767305e-01 6.83696389e-01
-9.21677768e-01 1.12057424e+00 -6.99262679e-01 -3.36271614e-01
4.64192897e-01 -5.99617846e-02 -1.59899354e-01 2.21979454e-01
-1.29284704e+00 1.17834914e+00 2.01379687e-01 -6.50317296e-02
-1.29392493e+00 -4.67483222e-01 -6.60631657e-01 3.28661621e-01
6.64620340e-01 -4.56995755e-01 1.60084617e+00 -8.82508636e-01
-1.39293325e+00 6.48830831e-01 3.38104010e-01 -6.74532413e-01
5.87421358e-01 -6.93614632e-02 -2.32184842e-01 -1.72573894e-01
2.66789079e-01 3.51409733e-01 5.99836886e-01 -8.13654304e-01
-2.33313605e-01 -5.50660133e-01 1.69479594e-01 4.78894502e-01
1.45243451e-01 2.47639567e-01 8.83753777e-01 -7.26288915e-01
-2.57288903e-01 -7.96526849e-01 -7.20701039e-01 -7.96598420e-02
-1.46103293e-01 -4.64822590e-01 -3.69013458e-01 -1.13745436e-01
1.06495547e+00 -2.12910509e+00 -8.82388577e-02 1.07985444e-01
-4.94493842e-02 -3.39279234e-01 -1.32058010e-01 6.54867232e-01
-2.55745500e-01 2.53924072e-01 -4.53126431e-01 -5.31699173e-02
7.11448908e-01 1.28922060e-01 -4.12635595e-01 7.57766187e-01
3.01849544e-02 6.46789193e-01 -9.73698854e-01 -1.41585320e-01
-7.19432905e-02 -5.14396131e-02 -6.31136835e-01 5.20015731e-02
-2.61830896e-01 8.82179141e-02 -5.13208151e-01 2.31728747e-01
3.61616641e-01 3.12659711e-01 2.81639010e-01 5.97610652e-01
-3.78603995e-01 3.43369544e-01 -1.04597461e+00 1.13173842e+00
-7.44185448e-02 2.51977503e-01 -3.95806693e-02 -8.09516191e-01
2.83565193e-01 1.76252097e-01 1.57181755e-01 -7.27259636e-01
3.88852149e-01 7.17227384e-02 3.75488639e-01 -3.39222819e-01
3.59890640e-01 -1.07601142e+00 -4.86482680e-01 7.82760799e-01
-1.39369398e-01 -1.27254859e-01 -1.10297181e-01 1.32192552e-01
8.92193437e-01 1.02436893e-01 8.79976630e-01 -8.53964806e-01
-9.51976553e-02 -2.20488727e-01 9.17205095e-01 8.03674340e-01
-7.45277107e-01 -3.66880111e-02 1.06447220e+00 -2.82482360e-03
-8.31891656e-01 -1.45439923e+00 -2.75899619e-01 8.27569366e-01
-1.63124558e-02 6.62104636e-02 -7.25993812e-01 -1.93492025e-01
2.88036913e-01 1.56758535e+00 -8.04407358e-01 -4.23311889e-01
-5.59066422e-02 -8.68340194e-01 2.61369020e-01 5.19615889e-01
-5.42625710e-02 -1.15124416e+00 -1.31007588e+00 7.24635348e-02
3.31015676e-01 -2.26625040e-01 -4.22969192e-01 4.28430796e-01
-8.13249946e-01 -9.12944257e-01 -6.94437861e-01 -1.20231621e-01
3.20708185e-01 -6.61403090e-02 1.00980818e+00 -2.81461388e-01
9.04705599e-02 7.65378177e-01 8.05155933e-02 -7.20584154e-01
-1.98176458e-01 -9.12488818e-01 7.13777781e-01 -2.40941405e-01
5.03151000e-01 -6.15923464e-01 -6.80053830e-01 2.26405561e-02
-9.17073369e-01 -4.79265779e-01 3.83776277e-01 5.86399972e-01
4.67882067e-01 1.75517261e-01 8.48380506e-01 -7.01436460e-01
8.93976450e-01 -8.47445309e-01 -9.27970111e-01 1.56117335e-01
-9.11448181e-01 2.20981315e-01 3.47274601e-01 -5.38236320e-01
-1.33202803e+00 -4.57083642e-01 8.17231089e-02 1.28216967e-01
-1.93289176e-01 6.42376840e-01 -3.19423199e-01 4.18051749e-01
5.83594382e-01 -6.34041205e-02 6.95764497e-02 -2.76863307e-01
3.66026431e-01 2.23899007e-01 2.71387786e-01 -9.30648088e-01
6.03047192e-01 8.01257864e-02 3.45858544e-01 -6.40510201e-01
-6.48406506e-01 3.73190880e-01 1.61513295e-02 -1.99803695e-01
1.10686743e+00 -6.84395909e-01 -1.32527876e+00 4.66910779e-01
-6.72801495e-01 -6.74944580e-01 -7.23327279e-01 8.08312535e-01
-1.40171111e+00 3.17054600e-01 -4.33576614e-01 -1.36092567e+00
3.39196384e-01 -1.00029755e+00 3.97350967e-01 2.73712635e-01
-1.84560657e-01 -1.02799523e+00 2.40549207e-01 -4.03549373e-01
4.38076645e-01 2.73664534e-01 1.30542231e+00 -8.32976103e-01
-2.70044208e-01 2.14910790e-01 2.85238922e-01 -1.28803188e-02
-1.33941144e-01 -4.34703559e-01 -6.40482605e-01 -1.44553512e-01
6.22037590e-01 -4.98022437e-01 6.40794754e-01 7.39773452e-01
6.70302331e-01 -5.94137371e-01 3.58989537e-02 1.61501735e-01
1.46781135e+00 6.82874203e-01 5.31866074e-01 4.67534810e-01
-9.32249725e-02 1.10866463e+00 8.53239357e-01 7.57793784e-01
5.41588962e-01 5.83776951e-01 6.19494319e-01 7.00821579e-01
7.49642313e-01 -5.48377395e-01 5.63065469e-01 -5.17964475e-02
1.35926604e-01 -1.85221612e-01 -7.53082514e-01 4.51600879e-01
-1.62125278e+00 -1.41342950e+00 4.65095073e-01 2.70287251e+00
8.63483846e-01 3.31791103e-01 5.22886634e-01 -1.86564520e-01
6.00899220e-01 1.58140749e-01 -7.68778443e-01 -6.51107848e-01
7.18198791e-02 -2.66460598e-01 6.86907768e-01 6.54894114e-01
-7.82605112e-01 5.65736413e-01 7.03915596e+00 5.90395093e-01
-6.31728172e-01 -1.60270005e-01 9.70178068e-01 -1.87161565e-01
-7.16022134e-01 1.11244339e-02 -3.49746346e-01 3.80638897e-01
1.12444675e+00 -8.70479107e-01 4.66231585e-01 9.63485181e-01
3.56681913e-01 -6.00659549e-01 -1.56809890e+00 5.36526501e-01
-4.30141896e-01 -5.92200637e-01 -4.47776854e-01 1.76213145e-01
3.00107867e-01 -2.93165743e-01 3.96870375e-01 2.88071245e-01
9.28025663e-01 -1.14161706e+00 1.32560086e+00 6.73361123e-01
5.37342489e-01 -9.24097478e-01 3.56884122e-01 7.25242794e-01
-4.91000891e-01 -4.46081936e-01 -4.43942130e-01 -4.75806296e-01
3.07281524e-01 3.56052160e-01 -2.72065520e-01 -1.95372086e-02
4.26620692e-01 2.73677975e-01 -1.57912508e-01 9.16815519e-01
-1.61254749e-01 3.21410149e-01 -3.89731109e-01 -2.13259816e-01
3.31600487e-01 -3.13257784e-01 5.78619838e-01 6.44232869e-01
3.17241579e-01 4.50352222e-01 -1.90852851e-01 1.05525744e+00
4.61263597e-01 -1.22281998e-01 -9.51624870e-01 -1.58167541e-01
5.51148653e-01 7.44545519e-01 -6.24563277e-01 2.84651071e-02
-2.85937011e-01 3.40383470e-01 3.11602890e-01 3.41361493e-01
-6.37335777e-01 -1.20933447e-02 9.57244813e-01 -6.64602891e-02
-1.39356583e-01 -2.24058509e-01 -2.98371613e-01 -8.91243815e-01
-1.40248686e-01 -6.35385096e-01 4.61750776e-01 -3.44435543e-01
-1.52831662e+00 7.39202425e-02 5.37763000e-01 -9.53016579e-01
-3.60216111e-01 -6.25996411e-01 -6.38569772e-01 1.02576220e+00
-1.34363866e+00 -3.28236997e-01 8.18252265e-01 4.90743458e-01
2.54957229e-01 2.10101679e-01 5.89873850e-01 -2.63765216e-01
-4.46771801e-01 3.72610718e-01 1.00341827e-01 -5.17997444e-01
4.75750089e-01 -1.45035303e+00 2.81015366e-01 6.66035652e-01
-2.67310411e-01 6.60734653e-01 9.88564432e-01 -8.38401973e-01
-1.23815489e+00 -7.05133080e-01 8.14590037e-01 -4.76850539e-01
8.65698278e-01 -9.67796072e-02 -5.17426670e-01 8.85890126e-01
-2.83961277e-03 -5.41532218e-01 5.96739411e-01 6.99483156e-02
-3.45150530e-01 1.62900880e-01 -1.46689022e+00 1.23770618e+00
9.88489747e-01 -3.59986544e-01 -1.01964486e+00 -1.75744090e-02
7.41792560e-01 3.05301428e-01 -5.02688169e-01 3.57142836e-02
5.89689076e-01 -1.22077394e+00 8.06549728e-01 -8.54143202e-01
2.32995331e-01 4.21044827e-02 -5.43569744e-01 -1.51303840e+00
-6.39188170e-01 -7.79223919e-01 7.07405448e-01 9.83747125e-01
1.90078706e-01 -1.07546985e+00 3.08406949e-01 1.15550137e+00
3.58675681e-02 -5.89349687e-01 -1.26688993e+00 -9.99695539e-01
8.33493114e-01 -5.95556974e-01 4.40165907e-01 7.19718754e-01
4.60737020e-01 -7.54842013e-02 -3.13025445e-01 -6.31091595e-02
1.11807275e+00 -2.86872774e-01 9.64472517e-02 -1.38291192e+00
-3.64535362e-01 -6.73162699e-01 -2.82544255e-01 -6.76680982e-01
2.50541866e-01 -6.74295843e-01 2.92603046e-01 -9.74495053e-01
2.94134766e-01 -4.71990436e-01 -4.20286715e-01 2.41795301e-01
-7.12746456e-02 -6.16356730e-01 4.15495217e-01 -8.85610357e-02
-2.70816445e-01 8.09845328e-01 8.94449234e-01 2.46565133e-01
-1.63457766e-01 1.14468202e-01 -1.16775203e+00 1.11985958e+00
8.26440573e-01 -5.98362029e-01 -7.04020977e-01 -9.01584551e-02
5.80842257e-01 7.83486485e-01 4.92125064e-01 -5.95088005e-01
-2.18876190e-02 -1.22551322e+00 3.74288782e-02 -1.01935491e-01
2.71304306e-02 -8.53343606e-01 2.14284062e-01 6.16467595e-01
-8.56660485e-01 3.79719943e-01 2.06009615e-02 9.96595800e-01
3.66838902e-01 -3.25962067e-01 8.86959255e-01 -5.24068624e-02
-4.49285150e-01 2.62792706e-01 -9.07428324e-01 2.34683871e-01
1.34810293e+00 1.07147843e-01 -4.12556738e-01 -4.86928642e-01
-6.27863526e-01 2.55071282e-01 5.69361091e-01 1.08658358e-01
5.52655160e-01 -1.21568239e+00 -6.04266107e-01 -1.83545485e-01
3.51704806e-02 -5.23645401e-01 4.53736871e-01 8.75973463e-01
-1.33461580e-02 2.89962173e-01 -4.55996126e-01 1.98079139e-01
-5.20870030e-01 1.04170239e+00 3.57694685e-01 -7.76447579e-02
-5.36472857e-01 6.70124233e-01 8.08540404e-01 1.56527068e-02
9.34891999e-02 -2.49257669e-01 -1.59319207e-01 1.90585583e-01
6.18283451e-01 4.12189335e-01 -5.67105949e-01 -6.24807537e-01
-4.82466429e-01 1.36194289e-01 6.45300895e-02 -7.18664408e-01
1.26336086e+00 -3.44401687e-01 7.26336762e-02 7.60631919e-01
3.71960014e-01 -2.68745214e-01 -1.74772418e+00 -1.29947275e-01
3.75407457e-01 -4.63422298e-01 -1.37759984e-01 -8.47977638e-01
-4.15269375e-01 9.15864348e-01 4.40224886e-01 3.59276265e-01
9.69956696e-01 -9.87220109e-02 8.99560750e-02 4.27534908e-01
7.64752746e-01 -1.15911651e+00 -2.11234137e-01 2.51063913e-01
9.58810985e-01 -9.34781969e-01 -5.91706745e-02 3.11507374e-01
-8.85148466e-01 6.75480306e-01 2.03476250e-01 -2.47515455e-01
8.08918238e-01 1.07309610e-01 -1.73812672e-01 3.94937880e-02
-1.13717222e+00 -1.89370364e-01 -2.05066085e-01 8.68850172e-01
1.65913150e-01 6.57893479e-01 -5.28122962e-01 1.07099342e+00
-2.26438522e-01 -1.40708476e-01 9.03237760e-01 9.19309556e-01
-6.46376014e-01 -8.25731218e-01 -2.37593144e-01 4.79412556e-01
-7.10495830e-01 -1.18185811e-01 8.66049994e-03 6.11639261e-01
-3.86329144e-01 9.19710338e-01 1.30697802e-01 -1.15381658e-01
2.71176696e-01 1.46980137e-01 4.52011764e-01 -6.40332818e-01
-2.97405094e-01 -1.05300918e-02 -7.01189041e-02 -7.70816743e-01
-2.96705306e-01 -1.16671014e+00 -1.01121056e+00 -4.19519216e-01
6.02242798e-02 -1.30566526e-02 3.08363795e-01 9.29221690e-01
-1.63051978e-01 5.14747947e-02 5.61544240e-01 -8.65330219e-01
-1.53245127e+00 -5.27501225e-01 -1.30071449e+00 2.30907366e-01
5.44984460e-01 -9.41298306e-01 -6.52141929e-01 -5.54066479e-01] | [4.250586986541748, 2.6534652709960938] |
3a020e01-8aaa-476a-bf06-232c1bf345e8 | cenet-toward-concise-and-efficient-lidar | 2207.12691 | null | https://arxiv.org/abs/2207.12691v1 | https://arxiv.org/pdf/2207.12691v1.pdf | CENet: Toward Concise and Efficient LiDAR Semantic Segmentation for Autonomous Driving | Accurate and fast scene understanding is one of the challenging task for autonomous driving, which requires to take full advantage of LiDAR point clouds for semantic segmentation. In this paper, we present a \textbf{concise} and \textbf{efficient} image-based semantic segmentation network, named \textbf{CENet}. In order to improve the descriptive power of learned features and reduce the computational as well as time complexity, our CENet integrates the convolution with larger kernel size instead of MLP, carefully-selected activation functions, and multiple auxiliary segmentation heads with corresponding loss functions into architecture. Quantitative and qualitative experiments conducted on publicly available benchmarks, SemanticKITTI and SemanticPOSS, demonstrate that our pipeline achieves much better mIoU and inference performance compared with state-of-the-art models. The code will be available at https://github.com/huixiancheng/CENet. | ['Guo-Qiang Xiao', 'Xian-Feng Han', 'Hui-Xian Cheng'] | 2022-07-26 | null | null | null | null | ['robust-3d-semantic-segmentation', 'lidar-semantic-segmentation'] | ['computer-vision', 'computer-vision'] | [ 2.89866608e-02 7.45330080e-02 -1.82254672e-01 -9.24754024e-01
-6.39693320e-01 -2.98788697e-01 2.90426642e-01 -1.80009022e-01
-7.79172838e-01 6.05199933e-01 -3.72019053e-01 -4.42934036e-01
-8.66284892e-02 -9.61956620e-01 -1.02426851e+00 -5.21541476e-01
3.67675006e-01 5.67297101e-01 5.85092068e-01 1.52485529e-02
2.15662941e-01 2.88191766e-01 -1.60089707e+00 -3.78709435e-02
1.20988548e+00 1.40858555e+00 5.83768189e-01 4.04078543e-01
-2.82124937e-01 4.67789769e-01 -2.72517025e-01 -2.99846023e-01
2.78807729e-01 1.36813283e-01 -8.77432227e-01 -2.60778725e-01
4.37137157e-01 -4.51386690e-01 -3.56899947e-01 1.21553099e+00
2.76758581e-01 3.95726681e-01 4.40761179e-01 -1.36076617e+00
-3.86899620e-01 5.90455413e-01 -5.31177282e-01 2.30488837e-01
-2.91543365e-01 3.79491061e-01 7.97061265e-01 -8.35379660e-01
2.39835352e-01 1.23087490e+00 6.33059263e-01 3.27894837e-01
-9.17915463e-01 -1.18337297e+00 4.23943341e-01 4.25732136e-01
-1.44955623e+00 -3.47320646e-01 4.72926289e-01 -3.25372636e-01
9.12276983e-01 1.28290489e-01 4.97154266e-01 8.76918316e-01
-3.06221209e-02 9.95649338e-01 8.96400928e-01 1.30270958e-01
1.07184500e-01 4.06760350e-02 5.81945062e-01 9.06058133e-01
3.13834250e-01 1.00374386e-01 -2.02859461e-01 4.03515220e-01
5.72596788e-01 1.32777154e-01 1.61967538e-02 -4.34104614e-02
-1.01892471e+00 9.05672789e-01 7.78456748e-01 -1.25137389e-01
-1.94327340e-01 6.08126819e-01 3.46244156e-01 -2.57051736e-01
5.00073969e-01 9.36997682e-02 -5.21238685e-01 -2.35892713e-01
-8.97988737e-01 2.06301257e-01 4.69866365e-01 1.31573498e+00
1.21904397e+00 6.35461584e-02 -1.76753953e-01 7.53099918e-01
2.59435594e-01 6.49726391e-01 1.24573953e-01 -1.37533498e+00
5.12429535e-01 7.40098178e-01 -9.60194990e-02 -5.24406552e-01
-4.44880605e-01 -3.50712955e-01 -6.79198563e-01 2.13287726e-01
2.30744913e-01 -1.77071586e-01 -1.44482648e+00 1.35969126e+00
1.87501833e-01 4.35091496e-01 -9.67811719e-02 9.43936229e-01
1.15345454e+00 6.92257285e-01 2.87165910e-01 5.17713785e-01
1.34441721e+00 -1.33948064e+00 -3.13773751e-01 -6.14697754e-01
4.51999068e-01 -4.17677224e-01 1.25325108e+00 2.84866929e-01
-8.64762247e-01 -7.43401587e-01 -1.07966149e+00 -4.28902864e-01
-5.22858799e-01 3.07265729e-01 8.70100677e-01 3.37163389e-01
-9.29166734e-01 5.35555482e-01 -1.17599738e+00 -2.72411108e-01
1.08298707e+00 4.45209861e-01 -7.66714662e-02 -1.82233632e-01
-9.61596191e-01 6.36081934e-01 7.94331014e-01 2.91341454e-01
-9.23493445e-01 -6.56368434e-01 -1.09294116e+00 1.09716550e-01
5.27915061e-01 -6.28851295e-01 1.48656762e+00 -6.25314772e-01
-1.42397213e+00 7.65419483e-01 -2.59415239e-01 -5.97790778e-01
5.24765015e-01 -5.86799681e-01 2.67932229e-02 1.60319805e-01
2.76858568e-01 1.31900907e+00 4.31796789e-01 -1.22189212e+00
-8.15191209e-01 -2.60794073e-01 9.64487866e-02 1.24114744e-01
-4.10331376e-02 -1.48737729e-01 -8.92134368e-01 -2.27157876e-01
9.60199907e-02 -8.27604890e-01 -4.34068948e-01 4.87126643e-03
-5.50442278e-01 -4.20061648e-01 9.43893075e-01 -5.93328297e-01
8.90477717e-01 -2.15645957e+00 -2.79375911e-01 5.62803484e-02
2.78611124e-01 4.25553977e-01 1.17797463e-03 -8.68006721e-02
2.67695218e-01 1.24147445e-01 -6.34951711e-01 -5.20498335e-01
9.28864330e-02 5.05605042e-01 -1.38185933e-01 1.96904838e-01
3.17541778e-01 1.15750360e+00 -8.10367227e-01 -6.07583880e-01
5.84113121e-01 4.05550689e-01 -4.84946907e-01 4.67166267e-02
-4.93371576e-01 5.72951078e-01 -8.21687579e-01 5.95068216e-01
8.87236536e-01 -2.68876851e-01 -5.98046899e-01 -9.61384922e-02
-1.55990660e-01 3.31683785e-01 -8.62252176e-01 1.86743283e+00
-3.30698460e-01 5.27934968e-01 4.10206914e-02 -1.03914022e+00
8.69984388e-01 -2.21136272e-01 3.64452988e-01 -9.29969668e-01
5.02871990e-01 2.11012468e-01 -2.14459017e-01 -3.79402041e-01
4.93306339e-01 2.33696774e-01 -6.13995530e-02 -5.48501052e-02
4.49123681e-02 -1.85628578e-01 2.63313174e-01 5.16419783e-02
7.80934453e-01 3.58749032e-01 -1.98822841e-01 -3.66101444e-01
4.57012951e-01 2.97319680e-01 8.40387702e-01 6.03020728e-01
-2.37340108e-01 6.00746870e-01 2.73408234e-01 -3.14922929e-01
-8.01429689e-01 -1.05099678e+00 -3.51162404e-01 7.97611713e-01
7.31619656e-01 -2.32093796e-01 -9.69805241e-01 -5.52677929e-01
1.65383995e-01 1.01802862e+00 -4.03752714e-01 -1.50979340e-01
-5.83360314e-01 -6.00916862e-01 6.53280616e-01 9.54423249e-01
1.10039377e+00 -1.12326300e+00 -7.41769135e-01 6.55047074e-02
-1.46949723e-01 -1.33427846e+00 -2.62707651e-01 2.23008186e-01
-7.95201659e-01 -9.10012662e-01 -3.61935884e-01 -7.01690674e-01
4.80990380e-01 2.92635292e-01 1.01248395e+00 2.76780203e-02
-3.14004958e-01 -8.60551596e-02 -2.18607366e-01 -7.94444799e-01
2.12636858e-01 4.14370805e-01 -4.41179931e-01 -2.72204667e-01
6.17033064e-01 -5.25837481e-01 -7.92996466e-01 3.33533466e-01
-8.22732270e-01 4.40140277e-01 4.78887439e-01 5.85307300e-01
9.38177466e-01 -2.94586979e-02 2.97441393e-01 -8.62344444e-01
1.12384841e-01 -4.32307959e-01 -9.15078580e-01 -2.18948737e-01
-6.34401321e-01 -3.75299901e-02 4.78044242e-01 9.39105749e-02
-1.06037128e+00 1.50412023e-01 -5.15634477e-01 -5.57491660e-01
-3.37790877e-01 3.17868054e-01 -2.27036715e-01 -5.69323674e-02
3.27628851e-01 2.54724771e-01 -1.09524556e-01 -5.56897402e-01
3.65543336e-01 6.36890888e-01 6.15956128e-01 -5.77616811e-01
6.61355615e-01 5.84692776e-01 -2.51969665e-01 -6.28913581e-01
-1.30079544e+00 -5.75101733e-01 -5.92059493e-01 -3.83383632e-02
1.18037140e+00 -1.09377146e+00 -6.49102390e-01 6.46993101e-01
-1.05086553e+00 -8.05886745e-01 -1.55114532e-01 3.69623840e-01
-5.65219223e-01 1.02561414e-01 -4.53096718e-01 -5.94731748e-01
-3.81598264e-01 -1.43086231e+00 1.18089795e+00 6.51736259e-01
2.79113799e-01 -6.29581869e-01 -5.08913457e-01 7.67525494e-01
3.19841146e-01 1.56792864e-01 5.28613091e-01 -5.38328171e-01
-9.82526600e-01 -4.39171568e-02 -8.47759366e-01 6.34676158e-01
-1.45956874e-01 -3.16571072e-02 -1.13870108e+00 -7.51897320e-02
-4.25858796e-01 -3.19963276e-01 1.37758064e+00 5.33225000e-01
1.84860790e+00 -5.60220033e-02 -5.01003683e-01 1.13339686e+00
1.34066141e+00 1.52705356e-01 6.82981610e-01 4.03153241e-01
9.43087578e-01 3.94529641e-01 7.80870616e-01 1.33563191e-01
8.66491377e-01 2.68571675e-01 6.04408264e-01 -2.45751008e-01
-1.06209829e-01 -1.41609728e-01 1.00145182e-02 4.74710792e-01
6.73804730e-02 -1.85148522e-01 -1.04192960e+00 5.94590068e-01
-1.95979071e+00 -5.72028637e-01 -2.31040090e-01 1.81650722e+00
6.04551792e-01 4.13570702e-01 -1.15894422e-01 -1.31880924e-01
5.48786938e-01 1.75690114e-01 -9.25638735e-01 -3.43519747e-01
1.38667405e-01 3.70016366e-01 8.96000862e-01 5.11381924e-01
-1.32789862e+00 1.44431841e+00 4.96580267e+00 1.04803503e+00
-1.15188396e+00 2.16213375e-01 8.36508989e-01 2.12104749e-02
-9.71603841e-02 -2.21807808e-02 -1.04095066e+00 5.77245176e-01
9.11539435e-01 5.48530780e-02 3.96569550e-01 9.80309725e-01
1.71408981e-01 -3.06253821e-01 -6.84275627e-01 7.91774869e-01
-3.64674181e-01 -1.25691259e+00 -2.64890790e-01 -6.65421635e-02
5.80998302e-01 7.43156493e-01 -9.78312418e-02 4.62769061e-01
5.35417199e-01 -1.19070017e+00 9.03949499e-01 5.16389489e-01
7.92766333e-01 -7.66950607e-01 8.37549746e-01 3.42709929e-01
-1.35762131e+00 -6.69447184e-02 -5.19588470e-01 -1.76864471e-02
1.35080740e-01 6.31756008e-01 -4.66091067e-01 5.96906245e-01
1.12278509e+00 8.62590730e-01 -5.92081487e-01 1.10902846e+00
-4.22048301e-01 8.71369958e-01 -4.96165663e-01 1.29331231e-01
5.68393290e-01 -3.92864168e-01 2.50206172e-01 1.32023799e+00
2.25980908e-01 1.86405972e-01 5.42668521e-01 1.21684539e+00
-8.67027715e-02 -2.50640094e-01 -2.83690572e-01 6.42832275e-03
5.36824346e-01 1.42532337e+00 -8.92225623e-01 -4.51331735e-01
-3.41454834e-01 7.93173313e-01 2.38746732e-01 4.33681130e-01
-1.06828427e+00 -6.63974524e-01 8.90739501e-01 1.46893812e-02
3.61999333e-01 -3.95548731e-01 -7.76812732e-01 -8.83983612e-01
1.28643543e-01 -2.93381304e-01 2.12057501e-01 -8.08425844e-01
-9.60190475e-01 5.99176347e-01 1.88752875e-01 -7.91866481e-01
3.59062910e-01 -7.27054238e-01 -7.65620708e-01 1.05214369e+00
-1.97613883e+00 -1.29174125e+00 -8.84723008e-01 4.41492647e-01
7.14886785e-01 1.59020320e-01 3.83105338e-01 4.25900489e-01
-9.57386672e-01 4.35119748e-01 -1.17177211e-01 2.48652175e-01
2.88321316e-01 -1.22155929e+00 7.49002993e-01 7.86345780e-01
-3.96347553e-01 3.07626814e-01 3.87898654e-01 -4.68911588e-01
-9.71915305e-01 -1.66578948e+00 5.08295894e-01 -3.53080720e-01
5.12250125e-01 -3.95764232e-01 -9.25384521e-01 8.63111079e-01
7.34473690e-02 1.04943432e-01 1.70610949e-01 -2.27288693e-01
-1.81397676e-01 -3.51070166e-01 -9.83546436e-01 3.88850719e-01
1.26375961e+00 -1.36990353e-01 -3.55904996e-01 2.23383307e-01
1.14077532e+00 -6.46169364e-01 -5.23253322e-01 6.89583957e-01
2.63060957e-01 -9.42952573e-01 1.01643860e+00 -1.59053698e-01
4.38378662e-01 -3.92512918e-01 -2.22261921e-02 -8.59085262e-01
-1.82530776e-01 -1.13483958e-01 1.69822067e-01 1.07354057e+00
5.39157867e-01 -9.47744191e-01 9.43346441e-01 4.39071566e-01
-7.08349943e-01 -9.93010461e-01 -8.60728145e-01 -7.43615329e-01
1.48972109e-01 -7.14563131e-01 6.16041780e-01 4.10915554e-01
-7.61063337e-01 1.40182644e-01 2.25086734e-01 2.68812299e-01
6.19941115e-01 1.34658203e-01 7.79527187e-01 -1.22950840e+00
1.57024935e-01 -5.70006728e-01 -3.67751837e-01 -1.22428775e+00
2.94130802e-01 -1.00539851e+00 4.02641714e-01 -1.91619098e+00
7.59684518e-02 -7.13427305e-01 -3.83068264e-01 8.78095448e-01
-2.04352945e-01 3.25646430e-01 8.99788514e-02 -4.18088697e-02
-8.11721563e-01 8.35458517e-01 1.36662447e+00 -1.86087504e-01
-2.29883865e-02 5.89184947e-02 -7.80707121e-01 9.00601745e-01
1.08303154e+00 -4.64638829e-01 -4.21600908e-01 -7.17989266e-01
-1.54089212e-01 -4.01998073e-01 6.72906637e-01 -1.26762700e+00
2.02495575e-01 -2.95994461e-01 3.31114739e-01 -9.19885278e-01
5.15078604e-01 -5.73827386e-01 -1.17690898e-01 3.90779972e-01
-1.74012650e-02 -1.15757793e-01 5.13496757e-01 3.76144826e-01
-2.81370640e-01 -1.18718125e-01 8.67136717e-01 -1.82629615e-01
-1.10077071e+00 7.39675820e-01 2.66647022e-02 1.49710476e-02
9.59031701e-01 -4.26754266e-01 -4.26922679e-01 -5.20803668e-02
-3.37968290e-01 9.26051855e-01 3.16269308e-01 4.46852565e-01
5.84738553e-01 -1.01463056e+00 -5.43192863e-01 1.07779495e-01
1.87144384e-01 8.75374615e-01 2.29896754e-01 9.06982899e-01
-8.10958803e-01 5.34477055e-01 -1.21182598e-01 -7.66334713e-01
-8.59350085e-01 1.87905058e-01 3.74459267e-01 8.24104622e-02
-9.44163084e-01 1.20198369e+00 5.59587061e-01 -7.90282845e-01
2.74348706e-01 -6.34987593e-01 -1.87743008e-02 -3.22295219e-01
1.91093192e-01 3.97807032e-01 6.52604401e-02 -4.68077630e-01
-3.80158424e-01 5.77326298e-01 -1.80738606e-02 2.23067224e-01
1.18368816e+00 -1.38379976e-01 -5.83545119e-03 2.51690179e-01
1.04078901e+00 -6.31170809e-01 -1.85112929e+00 -2.01824307e-01
2.77359169e-02 -3.47549647e-01 3.30809206e-01 -9.51985359e-01
-1.38806200e+00 1.14275682e+00 6.34750783e-01 -3.05182189e-01
1.12549996e+00 -1.69412605e-02 1.26914477e+00 4.82539892e-01
4.08229738e-01 -1.04684114e+00 -2.66427666e-01 7.33089209e-01
6.32337451e-01 -1.43117774e+00 -1.40061185e-01 -6.36747539e-01
-5.93672454e-01 7.23392487e-01 9.95740592e-01 -3.03615659e-01
7.60304868e-01 1.61851913e-01 2.21359700e-01 -2.31235221e-01
-3.48781198e-01 -5.49500346e-01 2.91040510e-01 3.74862701e-01
6.75282776e-02 2.00223565e-01 -1.18865021e-01 7.76214957e-01
-5.00050843e-01 7.94995502e-02 5.40660769e-02 7.67455399e-01
-7.11888552e-01 -8.87223959e-01 4.70568053e-02 7.23386049e-01
-2.45289773e-01 -2.40840390e-01 -1.25331059e-01 7.11049199e-01
4.94616210e-01 8.48916888e-01 2.31880233e-01 -4.64136511e-01
3.98775429e-01 -4.80688512e-02 1.14757247e-01 -6.37795389e-01
-3.51005882e-01 -1.01568706e-01 5.36805242e-02 -9.37511384e-01
-1.36480749e-01 -4.96332943e-01 -1.90635490e+00 -2.63576031e-01
-2.47064993e-01 -1.46672368e-01 8.24679971e-01 1.04843771e+00
4.86323923e-01 7.88044453e-01 2.65700668e-01 -9.12151098e-01
-2.63977945e-01 -9.57191408e-01 -3.47482413e-01 7.46336952e-02
2.30099857e-01 -7.84166515e-01 9.75038018e-03 -8.76271650e-02] | [8.23227596282959, -2.6287498474121094] |
9e037aed-b882-4b0b-be4f-3c2602b0a9e4 | expanding-scope-adapting-english-adversarial | 2306.04874 | null | https://arxiv.org/abs/2306.04874v1 | https://arxiv.org/pdf/2306.04874v1.pdf | Expanding Scope: Adapting English Adversarial Attacks to Chinese | Recent studies have revealed that NLP predictive models are vulnerable to adversarial attacks. Most existing studies focused on designing attacks to evaluate the robustness of NLP models in the English language alone. Literature has seen an increasing need for NLP solutions for other languages. We, therefore, ask one natural question: whether state-of-the-art (SOTA) attack methods generalize to other languages. This paper investigates how to adapt SOTA adversarial attack algorithms in English to the Chinese language. Our experiments show that attack methods previously applied to English NLP can generate high-quality adversarial examples in Chinese when combined with proper text segmentation and linguistic constraints. In addition, we demonstrate that the generated adversarial examples can achieve high fluency and semantic consistency by focusing on the Chinese language's morphology and phonology, which in turn can be used to improve the adversarial robustness of Chinese NLP models. | ['Yanjun Qi', 'Chengyuan Cai', 'Hanyu Liu'] | 2023-06-08 | null | null | null | null | ['adversarial-attack', 'adversarial-robustness'] | ['adversarial', 'adversarial'] | [ 3.03086549e-01 1.02383837e-01 -5.36724105e-02 -1.62849709e-01
-1.02537704e+00 -1.28267241e+00 6.67974174e-01 -4.11289841e-01
-3.27551156e-01 6.03490174e-01 8.84113833e-02 -6.89666927e-01
4.18957591e-01 -7.96876073e-01 -7.72092342e-01 -3.85282904e-01
3.79820257e-01 5.82802474e-01 1.48393035e-01 -3.87538731e-01
6.63966686e-03 7.97266603e-01 -6.46553278e-01 3.96999300e-01
1.26069295e+00 1.70824349e-01 -3.60152870e-01 9.01683152e-01
6.33693263e-02 5.47342479e-01 -1.00701857e+00 -1.10669684e+00
4.53442067e-01 -4.27742392e-01 -8.96653175e-01 -5.08195639e-01
5.35507560e-01 -1.74598739e-01 -3.64597380e-01 1.57209682e+00
5.69633365e-01 -7.97171742e-02 5.92893839e-01 -1.32664764e+00
-1.08175361e+00 1.07852006e+00 -6.89710230e-02 1.10001527e-02
3.57863992e-01 5.21116972e-01 8.64230812e-01 -3.84909481e-01
5.38323402e-01 1.60904860e+00 6.52093649e-01 1.17161953e+00
-1.13240874e+00 -1.14493525e+00 7.99336582e-02 1.82341989e-02
-1.25407362e+00 -1.85889319e-01 7.99576879e-01 -6.48064762e-02
1.09513342e+00 5.23768783e-01 1.81227162e-01 1.74164116e+00
4.47189540e-01 8.82371247e-01 1.21905243e+00 -5.49837172e-01
-6.87385872e-02 2.29428157e-01 -3.15243840e-01 4.77116525e-01
1.78094804e-01 5.17982781e-01 -2.14064628e-01 -1.76210493e-01
2.39686370e-01 -8.77504528e-01 -1.60479948e-01 3.86735201e-01
-9.08639729e-01 1.02678955e+00 1.45709828e-01 2.51063824e-01
1.88383743e-01 2.47911379e-01 5.50408483e-01 2.53221124e-01
3.63858730e-01 1.04888809e+00 -7.54722714e-01 -1.39298499e-01
-6.99384570e-01 4.41106349e-01 1.01049840e+00 1.35656857e+00
6.33935481e-02 6.02289796e-01 6.17863238e-02 5.53243637e-01
1.90986216e-01 1.10435069e+00 4.46053356e-01 -8.71312082e-01
7.79505610e-01 7.62715787e-02 -1.53202906e-01 -9.93782282e-01
-6.02894947e-02 -7.15507939e-02 -5.33748627e-01 1.94465727e-01
4.33961481e-01 -6.37887836e-01 -8.41654897e-01 1.77350152e+00
1.07296959e-01 2.28811130e-02 5.33455610e-01 5.50810099e-01
4.65166122e-01 8.60465705e-01 5.27233243e-01 1.72102615e-01
9.50872123e-01 -9.53213692e-01 -5.45583904e-01 -5.28060555e-01
7.91734874e-01 -1.18026292e+00 1.50670135e+00 2.11531058e-01
-1.00784695e+00 -4.93853569e-01 -9.43259239e-01 1.27357803e-02
-6.26838148e-01 -2.67094642e-01 4.72800732e-01 1.55831909e+00
-7.05925465e-01 5.74830055e-01 -7.93262780e-01 -1.22748110e-02
3.38158280e-01 3.20221126e-01 -1.76544860e-01 -1.35954395e-01
-1.81398475e+00 1.10386419e+00 6.17536187e-01 -1.65928885e-01
-7.28082955e-01 -8.58496547e-01 -9.73408401e-01 -5.31868525e-02
1.42460853e-01 -3.86448652e-01 1.17500722e+00 -1.27433884e+00
-1.75037897e+00 7.01544762e-01 1.00863717e-01 -5.16327560e-01
6.57480836e-01 -4.31585670e-01 -8.13362896e-01 9.46009085e-02
-1.09797038e-01 7.25840211e-01 9.25655484e-01 -1.20189452e+00
-3.48052949e-01 -5.70808128e-02 -1.43789560e-01 1.19415991e-01
-3.52079242e-01 5.90762377e-01 -4.36728805e-01 -1.19568229e+00
-4.31944698e-01 -1.31717908e+00 -2.87434310e-01 -3.72862011e-01
-9.29129481e-01 -2.68316641e-02 8.81527364e-01 -9.19747531e-01
1.00740099e+00 -2.13581276e+00 1.38170078e-01 3.57678652e-01
-5.17537892e-01 7.98262477e-01 -3.56508613e-01 1.72593549e-01
-8.85530412e-02 1.04149747e+00 -3.42149377e-01 -1.77330002e-01
2.85983413e-01 3.57057869e-01 -8.82969499e-01 2.90009886e-01
3.60309064e-01 1.14062274e+00 -6.03890598e-01 -5.35235882e-01
2.88643222e-02 3.56200427e-01 -8.30112875e-01 5.90733476e-02
-5.43326855e-01 4.84910786e-01 -5.31282902e-01 6.85313284e-01
6.58673286e-01 5.62148809e-01 1.99686270e-02 3.01652640e-01
4.94828522e-01 2.74695307e-01 -6.00327790e-01 1.09328699e+00
-4.05340910e-01 5.78220189e-01 -1.21235549e-01 -3.82279664e-01
7.31232047e-01 2.40908831e-01 -2.52007514e-01 -3.59800339e-01
9.57489312e-02 4.03749287e-01 5.10775328e-01 -2.05403358e-01
4.83774632e-01 -3.64433914e-01 -6.33930981e-01 3.85765046e-01
-2.09904283e-01 -9.58749652e-01 -1.25528350e-01 1.28205478e-01
6.71574771e-01 2.80559510e-01 -1.14532746e-02 -2.35913545e-01
7.08364606e-01 1.91904455e-01 5.39472401e-01 9.20177817e-01
-4.82066125e-01 5.70042908e-01 4.96690601e-01 1.90112032e-02
-1.16083658e+00 -1.12325609e+00 5.74662797e-02 9.62810278e-01
-2.07590431e-01 -2.40214840e-01 -1.26311660e+00 -1.18946564e+00
-2.89066322e-02 1.49102092e+00 -4.12094206e-01 -5.33243954e-01
-9.68286157e-01 -5.07603586e-01 1.62455201e+00 6.47775710e-01
5.43374062e-01 -1.43457568e+00 2.05405816e-01 -7.10976729e-03
-4.85187955e-02 -1.44279957e+00 -7.37007082e-01 -7.66121224e-02
-6.07238531e-01 -9.46917892e-01 -4.43591028e-01 -8.91932845e-01
5.29234052e-01 -6.85843766e-01 9.72358942e-01 4.99149747e-02
1.27743170e-01 3.55864555e-01 -4.41029578e-01 -8.07634413e-01
-1.52491784e+00 4.34164971e-01 1.14295952e-01 -5.01269102e-01
4.98353571e-01 -1.40759841e-01 2.45360956e-01 2.99926311e-01
-1.09307420e+00 -3.10845137e-01 1.16835549e-01 5.50449431e-01
4.88631785e-01 3.78415942e-01 4.60594714e-01 -1.51823270e+00
7.94521451e-01 -8.16959962e-02 -7.36419737e-01 4.00250256e-01
-3.10602874e-01 2.33705118e-02 1.34977078e+00 -8.68525386e-01
-1.30825305e+00 2.64119003e-02 -7.06642509e-01 -3.05903018e-01
-3.33980888e-01 2.55240738e-01 -8.58511388e-01 -3.04057032e-01
8.40581596e-01 1.44711480e-01 -3.47755283e-01 -5.04953489e-02
7.13435411e-01 4.56296265e-01 5.06137431e-01 -1.02027464e+00
1.36326325e+00 1.02984667e-01 -2.63482690e-01 -7.67979681e-01
-5.81234574e-01 3.86386037e-01 -6.64734304e-01 2.59054601e-01
9.69133079e-01 -6.35008693e-01 -3.82401317e-01 9.26524639e-01
-1.32812631e+00 -6.86256588e-01 -1.52342945e-01 2.79705107e-01
-6.09424710e-01 5.98072886e-01 -1.01747561e+00 -2.89011449e-01
-4.86236781e-01 -1.28494394e+00 5.81319809e-01 -2.57860944e-02
-5.17269969e-01 -1.28467298e+00 -1.20932586e-01 4.30676669e-01
3.43549877e-01 5.20765185e-02 1.25473797e+00 -1.15849459e+00
-3.09305906e-01 -4.30708915e-01 2.16653347e-01 7.08327889e-01
-1.10391967e-01 2.93260366e-01 -1.01696682e+00 -1.00300238e-01
1.77611738e-01 -2.82807380e-01 3.80559266e-01 -3.69435847e-02
9.81426239e-01 -6.83405459e-01 -9.00103301e-02 9.73917067e-01
1.10574794e+00 1.88676506e-01 8.53679359e-01 2.88690358e-01
1.04456425e+00 5.56896985e-01 5.09167671e-01 -3.60678583e-01
-2.41901711e-01 3.72939467e-01 2.31638595e-01 6.72680438e-02
-1.05273567e-01 -5.46325147e-01 7.58346498e-01 7.56093919e-01
4.89616424e-01 -5.89745104e-01 -1.10699427e+00 2.71137059e-01
-1.20628309e+00 -7.37330675e-01 -6.84902817e-02 1.63321912e+00
1.24353182e+00 3.71750563e-01 -4.55384493e-01 -9.91671160e-02
7.38651335e-01 1.57921612e-01 -4.82996345e-01 -1.05526435e+00
-4.68450397e-01 4.62871075e-01 5.81279099e-01 7.33005762e-01
-1.28238881e+00 1.93057418e+00 6.94557142e+00 1.19028056e+00
-1.09349978e+00 5.34136668e-02 6.60355091e-01 2.50859588e-01
-6.49669766e-01 -8.40784982e-02 -1.04972124e+00 3.33030701e-01
1.05346513e+00 -3.36274624e-01 4.93761927e-01 8.88972223e-01
-1.04658857e-01 5.65267682e-01 -9.04493272e-01 2.11361587e-01
1.79477707e-01 -9.76313829e-01 4.04983431e-01 -8.06504041e-02
9.67227578e-01 -1.88331351e-01 4.11780626e-01 5.23032069e-01
8.42188120e-01 -1.44351125e+00 6.77339554e-01 -1.56757440e-02
8.33642185e-01 -1.26751065e+00 7.15464532e-01 3.17177862e-01
-6.53679371e-01 2.30817050e-01 -4.25315320e-01 3.59296262e-01
1.65294588e-01 -1.42443657e-01 -7.44034767e-01 1.85118467e-01
4.88817662e-01 3.25572938e-01 -7.38791347e-01 3.80678624e-01
-1.02846670e+00 1.11830819e+00 -3.48539591e-01 -1.23473331e-01
4.07615781e-01 6.92804381e-02 7.50411391e-01 1.36108553e+00
8.97260308e-02 1.02823423e-02 2.64845765e-03 9.83617544e-01
-1.38977334e-01 2.43860021e-01 -8.20996284e-01 -3.80751669e-01
5.14930665e-01 7.40560949e-01 -6.83600664e-01 -1.32196009e-01
-2.18172491e-01 1.27215850e+00 1.75084084e-01 4.35928732e-01
-1.03698075e+00 -3.80925715e-01 7.51876771e-01 -2.96743989e-01
1.95712775e-01 -2.94163287e-01 -6.93180382e-01 -1.03601837e+00
-2.51843244e-01 -1.69407833e+00 3.39844495e-01 -6.55045629e-01
-1.58207679e+00 6.76045120e-01 -2.13522553e-01 -7.65716553e-01
-2.67672420e-01 -8.50565434e-01 -7.17373013e-01 1.16679633e+00
-1.14067066e+00 -1.59816039e+00 6.38209939e-01 8.34378421e-01
5.32734990e-01 -5.19045055e-01 1.19520390e+00 -9.08962488e-02
-6.36906743e-01 1.36272299e+00 -4.52285632e-02 6.21622145e-01
7.50598788e-01 -1.14173400e+00 1.04078496e+00 1.46132433e+00
2.62127966e-01 6.19584620e-01 8.02001476e-01 -1.03126645e+00
-1.10919380e+00 -1.27942681e+00 9.74658608e-01 -8.49543869e-01
8.62995982e-01 -3.51821482e-01 -1.03189254e+00 9.13723826e-01
2.70545065e-01 -1.44019559e-01 7.48567164e-01 -1.71855390e-01
-5.75993657e-01 3.67825419e-01 -1.47476244e+00 1.16594923e+00
6.42543912e-01 -8.69905770e-01 -5.83559513e-01 3.61433744e-01
1.29244959e+00 -6.65446520e-01 -7.98799694e-01 2.10975945e-01
1.26007125e-01 -3.41710925e-01 9.52706456e-01 -1.21289456e+00
5.43886065e-01 -6.18937798e-02 -3.29215884e-01 -1.46340048e+00
2.88132988e-02 -1.01199329e+00 3.53569925e-01 1.46804738e+00
8.10852051e-01 -7.79487371e-01 7.62031317e-01 9.91532743e-01
-8.02824423e-02 -1.96776748e-01 -8.38509858e-01 -8.45124304e-01
1.18299162e+00 -9.50769603e-01 5.84739447e-01 1.10634470e+00
-1.32089928e-01 1.29899278e-01 -4.29849863e-01 6.30134463e-01
3.01923394e-01 -3.19198966e-01 6.44699097e-01 -4.88951832e-01
-3.21133643e-01 -4.77815837e-01 -1.57620668e-01 -4.78018314e-01
1.06903172e+00 -1.03219354e+00 1.46961272e-01 -6.73024654e-01
-3.12972575e-01 -5.82380116e-01 2.52384096e-01 5.45362413e-01
-6.07921422e-01 2.83502519e-01 6.59063876e-01 -1.01432512e-02
1.00599729e-01 4.18638736e-01 1.30570102e+00 -2.45190114e-01
2.84660514e-02 1.34035200e-01 -7.18270659e-01 8.84669423e-01
1.27759552e+00 -7.54462659e-01 -2.85777003e-01 -5.36413968e-01
1.15160361e-01 -5.46879292e-01 6.05084375e-02 -9.07730341e-01
1.08668447e-01 -3.49097461e-01 3.33170563e-01 -5.51826209e-02
1.99470278e-02 -8.96546602e-01 -2.07373053e-01 5.98496497e-01
-4.68774289e-01 7.80452937e-02 8.62626910e-01 3.12581062e-01
-1.23584181e-01 -6.35744810e-01 1.00176597e+00 -5.28269187e-02
-4.50095028e-01 1.52205318e-01 -5.69953978e-01 6.27091885e-01
1.04745865e+00 1.02259755e-01 -3.51977378e-01 -3.77223402e-01
-5.65448165e-01 6.75418675e-02 6.46749020e-01 6.68199122e-01
4.44737226e-01 -1.08952200e+00 -9.01419938e-01 3.60220551e-01
-3.78808260e-01 -3.05367053e-01 2.28658281e-02 3.54821943e-02
-1.00101113e+00 3.79717171e-01 -1.22925386e-01 -6.06903050e-04
-1.33146620e+00 9.60216939e-01 5.22582352e-01 -3.68278712e-01
-3.43056917e-01 9.67297614e-01 -1.70067865e-02 -9.83456671e-01
1.25176102e-01 4.03091721e-02 1.41976863e-01 -4.57185179e-01
2.94252902e-01 1.18872963e-01 -2.86290199e-01 -7.94928730e-01
-3.74329865e-01 3.75010461e-01 -1.65836558e-01 -3.43852013e-01
8.76958847e-01 2.82112360e-01 8.11449513e-02 -1.15905657e-01
1.09120107e+00 7.47305751e-01 -9.85291302e-01 5.37851080e-02
-3.98105346e-02 -2.65002966e-01 -4.32225168e-01 -1.00106215e+00
-1.05960834e+00 1.03921068e+00 2.30588272e-01 6.02023974e-02
9.27008629e-01 -2.77511179e-01 1.08874786e+00 4.36293036e-01
5.08646257e-02 -1.11943281e+00 -1.63143039e-01 1.00231946e+00
9.49359894e-01 -9.26281512e-01 -2.09813133e-01 -8.05342376e-01
-1.15005648e+00 1.04829311e+00 6.94490612e-01 -2.42074788e-01
6.57687664e-01 6.37699842e-01 3.89217824e-01 4.21958625e-01
-2.65210122e-01 3.47565949e-01 2.96706110e-01 1.08361185e+00
2.96739310e-01 2.04286262e-01 5.84904589e-02 7.51129746e-01
-8.71794403e-01 -6.25463009e-01 5.93301296e-01 6.25392973e-01
1.04693018e-01 -1.53657222e+00 -5.92989862e-01 3.48068550e-02
-1.03778028e+00 -6.15078986e-01 -8.69457126e-01 1.12803864e+00
-1.97800770e-02 9.09042418e-01 -3.00269872e-01 -3.33270490e-01
2.32123658e-01 3.77562135e-01 3.69758129e-01 -6.49996221e-01
-1.08608377e+00 -1.48748532e-01 3.19251418e-01 -4.39329714e-01
5.88490516e-02 -5.61916947e-01 -1.26893890e+00 -6.41335428e-01
-1.94581747e-01 1.10670477e-02 4.42686141e-01 8.33206654e-01
5.16383015e-02 3.03564876e-01 2.78881758e-01 -4.29239124e-01
-8.04835439e-01 -7.79673874e-01 -3.91136557e-01 5.27046680e-01
-1.72230616e-01 1.98803514e-01 -5.14306366e-01 5.61783127e-02] | [6.024596691131592, 8.156030654907227] |
e1ce86b3-4b38-41cb-a197-fc7271a49337 | detailed-2d-3d-joint-representation-for-human | 2004.08154 | null | https://arxiv.org/abs/2004.08154v2 | https://arxiv.org/pdf/2004.08154v2.pdf | Detailed 2D-3D Joint Representation for Human-Object Interaction | Human-Object Interaction (HOI) detection lies at the core of action understanding. Besides 2D information such as human/object appearance and locations, 3D pose is also usually utilized in HOI learning since its view-independence. However, rough 3D body joints just carry sparse body information and are not sufficient to understand complex interactions. Thus, we need detailed 3D body shape to go further. Meanwhile, the interacted object in 3D is also not fully studied in HOI learning. In light of these, we propose a detailed 2D-3D joint representation learning method. First, we utilize the single-view human body capture method to obtain detailed 3D body, face and hand shapes. Next, we estimate the 3D object location and size with reference to the 2D human-object spatial configuration and object category priors. Finally, a joint learning framework and cross-modal consistency tasks are proposed to learn the joint HOI representation. To better evaluate the 2D ambiguity processing capacity of models, we propose a new benchmark named Ambiguous-HOI consisting of hard ambiguous images. Extensive experiments in large-scale HOI benchmark and Ambiguous-HOI show impressive effectiveness of our method. Code and data are available at https://github.com/DirtyHarryLYL/DJ-RN. | ['Junqi Liu', 'Yong-Lu Li', 'Shiyi Wang', 'Jiefeng Li', 'Xinpeng Liu', 'Han Lu', 'Cewu Lu'] | 2020-04-17 | detailed-2d-3d-joint-representation-for-human-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Li_Detailed_2D-3D_Joint_Representation_for_Human-Object_Interaction_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Li_Detailed_2D-3D_Joint_Representation_for_Human-Object_Interaction_CVPR_2020_paper.pdf | cvpr-2020-6 | ['action-understanding'] | ['computer-vision'] | [-1.18087396e-01 -9.63016078e-02 -2.27970466e-01 -2.47733071e-01
-5.07975876e-01 -2.47331470e-01 2.57088631e-01 -4.61124986e-01
-3.11590880e-02 4.20423150e-01 3.56786937e-01 3.65507990e-01
-1.12546481e-01 -2.87127644e-01 -8.37745428e-01 -6.95820570e-01
1.73523337e-01 9.78561759e-01 2.52353877e-01 -3.88088301e-02
-1.64613158e-01 3.06282729e-01 -1.40868676e+00 3.92407961e-02
4.93105650e-01 1.03406942e+00 2.06447810e-01 3.91930699e-01
3.82101715e-01 1.43111303e-01 -4.38775212e-01 -1.57722846e-01
3.25469851e-01 -1.89902797e-01 -4.99214292e-01 5.07321477e-01
4.69553947e-01 -6.18719995e-01 -4.07601357e-01 8.47640336e-01
7.91386068e-01 2.32506499e-01 7.59204566e-01 -1.32406473e+00
-3.50975215e-01 2.08659753e-01 -9.95974183e-01 -6.19553439e-02
7.11021900e-01 4.44105506e-01 9.29707527e-01 -1.14576769e+00
6.62615716e-01 1.79990697e+00 2.88597912e-01 5.37775040e-01
-7.72278726e-01 -8.22370768e-01 4.22554195e-01 1.11239888e-01
-1.46874607e+00 -2.58405477e-01 9.87037420e-01 -4.65489000e-01
5.72252750e-01 1.07147843e-01 8.68893504e-01 1.29357529e+00
1.84286773e-01 1.32583916e+00 9.04878557e-01 -3.17597777e-01
-3.52676481e-01 -3.37080717e-01 8.63341093e-02 9.49760914e-01
3.51388395e-01 6.20352887e-02 -6.05375051e-01 6.79651201e-02
1.19866621e+00 1.81616694e-01 -2.94158101e-01 -6.05001926e-01
-1.37288523e+00 4.24655139e-01 3.20901394e-01 -7.09263012e-02
-2.32748345e-01 1.69028118e-01 3.04753006e-01 -2.89230198e-01
2.99242228e-01 -3.51759315e-01 -2.61165857e-01 1.10068813e-01
-1.40662417e-01 5.23371816e-01 3.79083484e-01 1.18242705e+00
4.97713178e-01 -2.20227554e-01 -3.37683707e-01 8.62896621e-01
8.45500290e-01 9.56437767e-01 1.43063650e-01 -9.80113029e-01
7.47737586e-01 7.08199859e-01 1.88117251e-01 -1.05898321e+00
-5.81032753e-01 -3.69830370e-01 -9.18387711e-01 1.69819206e-01
5.84030151e-01 1.54471874e-01 -9.86831009e-01 1.66497874e+00
8.35653186e-01 -1.69573694e-01 -4.62848961e-01 1.45992291e+00
9.84407365e-01 3.52670759e-01 1.82583928e-02 -1.77704357e-02
1.67513883e+00 -1.02517223e+00 -7.42375970e-01 -4.72521156e-01
2.45670199e-01 -6.82460964e-01 9.93439853e-01 3.41942310e-01
-1.12409067e+00 -7.64646292e-01 -7.38251805e-01 -2.23196030e-01
-1.06473438e-01 4.72318292e-01 6.95371747e-01 3.03798348e-01
-2.10926548e-01 -1.26122802e-01 -8.26942325e-01 -3.01747590e-01
3.29315573e-01 4.07964975e-01 -4.90813971e-01 -2.92651206e-01
-1.18961167e+00 7.71578074e-01 3.72701377e-01 5.14458358e-01
-9.77192760e-01 -2.90589452e-01 -1.16138697e+00 -3.56588006e-01
8.95896792e-01 -1.09465647e+00 1.16144300e+00 -2.07494378e-01
-1.35104454e+00 1.02850163e+00 -1.10433906e-01 2.04407007e-01
6.48522079e-01 -5.62826812e-01 -6.29725233e-02 1.22618645e-01
1.08072693e-02 6.84580088e-01 1.10258532e+00 -1.39039481e+00
-4.38636452e-01 -8.99771154e-01 4.76834513e-02 6.77788734e-01
2.81056136e-01 -1.75151691e-01 -1.06556523e+00 -7.10601211e-01
5.48543632e-01 -1.30947804e+00 1.21426754e-01 3.46604884e-01
-5.61402261e-01 -4.54794526e-01 6.12165511e-01 -8.89041603e-01
9.82626975e-01 -1.98814511e+00 5.31019092e-01 1.57923818e-01
1.97477236e-01 1.02506839e-01 1.03708312e-01 -9.59449559e-02
4.19745175e-03 -3.34305048e-01 7.07537914e-03 -4.43451375e-01
2.14034379e-01 3.44859868e-01 1.68703645e-01 6.87881947e-01
-3.67126800e-02 1.13634574e+00 -6.66660786e-01 -8.44339013e-01
3.16298246e-01 4.56339091e-01 -4.62389708e-01 2.83072561e-01
-5.50974496e-02 8.72042418e-01 -7.22695053e-01 1.11882973e+00
6.28186226e-01 -3.07219326e-01 -1.26066163e-01 -7.31139660e-01
3.33432794e-01 -2.19302505e-01 -1.39016294e+00 1.88614392e+00
-1.98170960e-01 -1.87684536e-01 3.41546595e-01 -7.97168791e-01
5.67753732e-01 3.12186688e-01 5.65997601e-01 -4.45349723e-01
2.88627505e-01 1.98998917e-02 -2.58348417e-02 -5.28734624e-01
-8.20186660e-02 3.51418257e-02 -1.05751410e-01 9.93259549e-02
-4.97322157e-02 -1.12977669e-01 -3.48023809e-02 -1.54200852e-01
3.98289829e-01 5.02878308e-01 4.42860156e-01 1.72686562e-01
3.81404400e-01 -4.67706770e-01 7.00732529e-01 5.01408100e-01
-3.80044788e-01 8.73280644e-01 1.26904190e-01 -3.01252276e-01
-7.23394990e-01 -1.23081970e+00 2.09851656e-02 1.01377821e+00
5.50922692e-01 -1.16849318e-01 -4.38948900e-01 -6.47590756e-01
1.96433961e-01 9.62260962e-02 -6.23498321e-01 -6.10497333e-02
-7.35211134e-01 -5.96973717e-01 -5.39321676e-02 8.14794362e-01
5.38854241e-01 -1.08051336e+00 -6.45561576e-01 -7.73362443e-02
-5.68369269e-01 -9.96992409e-01 -8.90696645e-01 -3.48535717e-01
-7.18262315e-01 -1.21970177e+00 -1.01771033e+00 -5.85013986e-01
6.61982715e-01 3.08433682e-01 9.57842946e-01 -1.67050119e-02
-6.44741714e-01 7.61159062e-01 -3.02226841e-01 -4.49045837e-01
2.05276147e-01 -2.09086344e-01 2.91984767e-01 -8.13567340e-02
1.64825357e-02 -4.35788840e-01 -8.56840372e-01 7.81765819e-01
-5.79835474e-01 1.51360020e-01 8.22452128e-01 7.30557919e-01
7.77443945e-01 6.84180185e-02 -1.15959339e-01 -4.05964226e-01
-8.38785917e-02 -1.76062256e-01 -3.40985149e-01 4.92864519e-01
8.65939483e-02 -2.50614345e-01 -2.29725882e-01 -6.26685441e-01
-1.31378567e+00 2.61605859e-01 8.50769784e-03 -6.82582080e-01
-4.49829608e-01 2.58654028e-01 -8.17277908e-01 1.62294686e-01
1.28098875e-01 4.45196219e-02 2.15517968e-01 -6.23173714e-01
1.77215233e-01 3.42130005e-01 5.11686087e-01 -8.39994669e-01
7.18509912e-01 6.00061476e-01 -1.72696710e-02 -6.12695932e-01
-1.09551609e+00 -6.69663191e-01 -8.76226664e-01 -4.15747732e-01
1.01392508e+00 -1.07092953e+00 -9.80664492e-01 7.07629859e-01
-1.20812881e+00 -2.79939890e-01 1.09495081e-01 7.52548814e-01
-7.13221908e-01 6.24631584e-01 -4.73142296e-01 -9.14984286e-01
-3.09462696e-01 -1.20545769e+00 1.77752340e+00 1.16574867e-02
-1.71622232e-01 -6.06150627e-01 -2.17408508e-01 8.41654241e-01
-3.40463430e-01 2.27493510e-01 6.22886121e-01 -1.74593568e-01
-5.51607907e-01 -1.85669422e-01 -2.21856356e-01 7.67194852e-02
4.36800569e-02 -4.11378264e-01 -6.17549598e-01 -3.26998919e-01
-8.60042796e-02 -4.68388170e-01 5.77134609e-01 7.54697144e-01
1.09695613e+00 -6.38176352e-02 -4.27612603e-01 5.01818538e-01
8.04386139e-01 -4.30331752e-02 3.70407254e-01 -7.63428807e-02
1.20378494e+00 8.91934097e-01 1.28142416e+00 8.71325076e-01
5.23548901e-01 1.08739829e+00 5.63012302e-01 1.02858417e-01
-3.88944089e-01 -3.24315935e-01 2.77492076e-01 6.84979200e-01
-5.35217404e-01 -8.63092467e-02 -9.52677488e-01 2.05876336e-01
-1.95381117e+00 -6.94319904e-01 -6.42828718e-02 2.08894515e+00
6.15071297e-01 1.14323176e-01 3.30117702e-01 -1.05094880e-01
7.95691192e-01 -4.62786444e-02 -7.87627518e-01 6.76415026e-01
7.80560821e-02 -3.35566521e-01 2.43191451e-01 4.47855473e-01
-1.17534423e+00 8.14982235e-01 4.87508345e+00 8.52591276e-01
-6.09875798e-01 1.28642723e-01 1.85779944e-01 -1.75242156e-01
-7.45331170e-04 -3.14827681e-01 -1.23445499e+00 4.36308622e-01
-7.93266594e-02 3.78543943e-01 2.22149715e-01 7.65391111e-01
8.87864977e-02 -1.89939365e-01 -1.12899625e+00 1.56397533e+00
2.29400188e-01 -5.04236162e-01 7.03980699e-02 2.54590660e-01
3.09295923e-01 -3.59514356e-01 8.28715339e-02 4.00202006e-01
-3.23847309e-02 -7.75812089e-01 8.63154709e-01 7.52457023e-01
7.56299496e-01 -3.77586752e-01 5.55804789e-01 5.27884185e-01
-1.56620026e+00 3.63538973e-02 -2.76824564e-01 1.12608142e-01
2.85007089e-01 4.31570649e-01 -4.30878192e-01 6.54390812e-01
9.26465094e-01 7.17933893e-01 -5.74423492e-01 9.08970296e-01
-2.79478222e-01 2.34142631e-01 -6.71080112e-01 4.84299392e-01
-2.16690943e-01 -5.64518757e-02 6.65647626e-01 7.71487236e-01
2.85119236e-01 5.49265742e-01 6.09820783e-01 6.21176302e-01
3.09055567e-01 -8.06473121e-02 -5.31787217e-01 1.92390248e-01
1.62204579e-01 9.69316065e-01 -5.82681477e-01 -2.04518333e-01
-3.76546949e-01 1.08733904e+00 2.15613991e-02 4.43169087e-01
-1.00459290e+00 1.92704871e-01 4.59892124e-01 1.27230421e-01
2.55809993e-01 -4.81309682e-01 1.68388218e-01 -1.43522751e+00
2.41506428e-01 -9.17377174e-01 5.78304112e-01 -9.08749104e-01
-1.33296835e+00 -2.01597810e-02 5.85266709e-01 -1.38986933e+00
-1.82087421e-01 -8.66491914e-01 -1.92532346e-01 6.48872614e-01
-9.69774842e-01 -1.61072028e+00 -6.39607012e-01 7.00388789e-01
7.43998647e-01 1.79571927e-01 4.02727902e-01 2.42028475e-01
-6.37132227e-01 5.64804137e-01 -4.57471371e-01 1.56267002e-01
8.88415515e-01 -9.13982749e-01 7.14038536e-02 2.20846042e-01
5.03730215e-02 6.81926370e-01 5.88873029e-01 -9.32487905e-01
-1.59013236e+00 -7.56923139e-01 2.80555010e-01 -9.39091802e-01
2.51637727e-01 -3.95246357e-01 -7.79845953e-01 8.31271648e-01
-3.11825663e-01 3.47682834e-01 3.33648592e-01 1.08344950e-01
-1.80601999e-01 -1.81216206e-02 -7.50398576e-01 5.77189803e-01
1.62972331e+00 -2.75106281e-01 -6.31405950e-01 4.98530090e-01
4.73296791e-01 -8.87774944e-01 -9.52490032e-01 7.57255793e-01
9.26839352e-01 -6.58383608e-01 1.31963611e+00 -2.50586510e-01
8.55988711e-02 -4.04463083e-01 -1.83422029e-01 -7.74789810e-01
-2.27336019e-01 -2.53036946e-01 -4.44288433e-01 1.00811338e+00
-1.96220309e-01 -3.49472612e-01 8.60620499e-01 5.38123071e-01
5.42704090e-02 -8.34870100e-01 -8.21618199e-01 -7.40581691e-01
-4.23723072e-01 -4.24711645e-01 2.89467007e-01 4.38243926e-01
-3.58011812e-01 4.38646168e-01 -7.68700063e-01 4.38850224e-01
1.01307666e+00 4.74035412e-01 1.11997092e+00 -1.26489878e+00
-5.00216961e-01 -1.65077507e-01 -3.51974756e-01 -1.43152702e+00
2.23371342e-01 -6.43448114e-01 8.32456499e-02 -1.35301316e+00
5.68032503e-01 -1.24638401e-01 -6.41656145e-02 5.78683794e-01
-2.96035200e-01 2.12955788e-01 1.84942648e-01 5.25777936e-01
-7.92984903e-01 9.13225174e-01 1.74541056e+00 -1.76648259e-01
1.17659196e-01 1.15328334e-01 -1.09530836e-01 9.96567428e-01
5.25246561e-01 -1.36581272e-01 -4.64187384e-01 -3.86986971e-01
-5.16591296e-02 2.34342843e-01 8.04753542e-01 -9.62465227e-01
-1.19460402e-02 -1.81124896e-01 6.49602830e-01 -1.08280969e+00
6.79749906e-01 -9.15281355e-01 1.46807194e-01 3.44995111e-01
-1.04048094e-02 -3.11521113e-01 -7.54735991e-02 8.05855215e-01
-4.83089536e-02 3.32041904e-02 5.35747111e-01 -3.88309538e-01
-7.62979507e-01 7.83466935e-01 1.58244029e-01 8.57772529e-02
8.62699747e-01 -1.95834890e-01 1.89176157e-01 -3.50169629e-01
-1.09668124e+00 5.56582570e-01 4.68275666e-01 5.69511235e-01
7.53779948e-01 -1.31945169e+00 -5.27448416e-01 2.00207040e-01
2.47091487e-01 4.50717300e-01 6.56501889e-01 1.17565691e+00
-6.47294745e-02 5.11680961e-01 -2.76637763e-01 -9.27129865e-01
-1.50411606e+00 5.26106834e-01 1.51879624e-01 -7.47734606e-02
-7.37229407e-01 7.93969393e-01 8.80675197e-01 -6.57173216e-01
5.20519674e-01 -2.89680898e-01 -6.34667501e-02 -9.38447565e-02
3.55471551e-01 3.27792674e-01 -3.74169141e-01 -9.60300624e-01
-5.08266807e-01 1.11252081e+00 8.86521041e-02 -6.52132705e-02
7.98447788e-01 -4.63628381e-01 2.77475286e-02 4.66223478e-01
1.03422952e+00 -2.54389793e-01 -1.45204842e+00 -4.05060649e-01
-4.38958704e-01 -6.28489614e-01 -3.91605347e-01 -5.68666875e-01
-1.00594103e+00 1.12880313e+00 7.02780426e-01 -6.24793768e-01
8.91617298e-01 4.40751761e-01 7.29890287e-01 3.66624385e-01
6.49912000e-01 -9.74677265e-01 4.47603256e-01 3.05270791e-01
1.44770217e+00 -1.54232156e+00 2.49078065e-01 -8.56521964e-01
-7.87892580e-01 7.77758956e-01 1.05026627e+00 2.44034722e-01
6.83222771e-01 -1.28769442e-01 -1.40476063e-01 -3.96053970e-01
-2.93149650e-01 -3.17951918e-01 7.80082047e-01 5.80197036e-01
2.20214739e-01 6.97611719e-02 -1.78580701e-01 7.17696190e-01
1.20889477e-01 -1.62939683e-01 -3.14556450e-01 8.92208934e-01
-2.09520742e-01 -8.33454847e-01 -7.41041183e-01 3.87348719e-02
-2.21572638e-01 5.26893497e-01 -1.95633128e-01 9.65247989e-01
2.60721743e-01 5.19317508e-01 -2.57483900e-01 -1.32364422e-01
4.73804295e-01 5.40381745e-02 9.99526501e-01 -7.59652853e-01
1.98206827e-01 4.79339719e-01 -1.71182171e-01 -8.48024786e-01
-4.32735622e-01 -6.56809926e-01 -1.37310088e+00 2.28624001e-01
-3.46073270e-01 -3.17960411e-01 5.09864807e-01 9.44857359e-01
4.00074758e-02 2.80077159e-01 3.65110002e-02 -1.44547367e+00
-6.07769072e-01 -9.41979051e-01 -7.83520699e-01 5.47680974e-01
1.31945044e-01 -1.45031440e+00 -1.95850834e-01 -7.59744924e-03] | [7.122542381286621, -0.8576562404632568] |
8376273e-d7a4-47c9-81e2-f47e4acc9111 | wmt-2016-multimodal-translation-system | null | null | https://aclanthology.org/W16-2362 | https://aclanthology.org/W16-2362.pdf | WMT 2016 Multimodal Translation System Description based on Bidirectional Recurrent Neural Networks with Double-Embeddings | null | ['Marta R. Costa-juss{\\`a}', "Sergio Rodr{\\'\\i}guez Guasch"] | 2016-08-01 | null | null | null | ws-2016-8 | ['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] | [-7.43734073638916, 3.630056381225586] |
0fc55e09-7c80-4c6c-b718-7b60b7c66256 | adversarial-attacks-and-defenses-in-machine | 2303.06302 | null | https://arxiv.org/abs/2303.06302v1 | https://arxiv.org/pdf/2303.06302v1.pdf | Adversarial Attacks and Defenses in Machine Learning-Powered Networks: A Contemporary Survey | Adversarial attacks and defenses in machine learning and deep neural network have been gaining significant attention due to the rapidly growing applications of deep learning in the Internet and relevant scenarios. This survey provides a comprehensive overview of the recent advancements in the field of adversarial attack and defense techniques, with a focus on deep neural network-based classification models. Specifically, we conduct a comprehensive classification of recent adversarial attack methods and state-of-the-art adversarial defense techniques based on attack principles, and present them in visually appealing tables and tree diagrams. This is based on a rigorous evaluation of the existing works, including an analysis of their strengths and limitations. We also categorize the methods into counter-attack detection and robustness enhancement, with a specific focus on regularization-based methods for enhancing robustness. New avenues of attack are also explored, including search-based, decision-based, drop-based, and physical-world attacks, and a hierarchical classification of the latest defense methods is provided, highlighting the challenges of balancing training costs with performance, maintaining clean accuracy, overcoming the effect of gradient masking, and ensuring method transferability. At last, the lessons learned and open challenges are summarized with future research opportunities recommended. | ['H. Vincent Poor', 'Ekram Hossain', 'Wei Ni', 'Xin Yuan', 'Shenghong Li', 'Tong Sun', 'Yulong Wang'] | 2023-03-11 | null | null | null | null | ['adversarial-defense'] | ['adversarial'] | [ 1.27533808e-01 -3.56223553e-01 -6.37217909e-02 -8.11405405e-02
-5.87932765e-01 -8.88734043e-01 5.17149866e-01 -1.87870488e-01
-3.41388941e-01 4.97291148e-01 -6.57248311e-03 -7.15659082e-01
-2.99011618e-01 -7.21675336e-01 -5.91179371e-01 -9.58373189e-01
-6.31677508e-01 -4.74400938e-01 4.79199663e-02 -5.50801873e-01
3.07161808e-01 1.13734341e+00 -8.84988189e-01 2.44656011e-01
5.91796756e-01 1.19433188e+00 -7.60382354e-01 8.39808464e-01
1.20251328e-01 8.54224980e-01 -1.40093541e+00 -1.00810599e+00
5.80545366e-01 -1.93016246e-01 -4.78438288e-01 -6.46604896e-01
6.37752354e-01 -6.45460010e-01 -1.19048405e+00 1.34631562e+00
1.05026090e+00 1.76265929e-02 4.58299041e-01 -1.44480872e+00
-9.52137530e-01 5.18932521e-01 -3.98469150e-01 5.20579219e-01
-1.98013574e-01 2.84548789e-01 4.80842143e-01 -4.05912489e-01
-9.75582097e-03 1.27888513e+00 8.99647951e-01 1.18114221e+00
-7.42871940e-01 -1.29885519e+00 5.50630271e-01 4.16797638e-01
-1.01503813e+00 -4.58464235e-01 1.08014894e+00 -1.48021579e-02
9.56118703e-01 4.68807340e-01 1.33469254e-01 1.73943281e+00
6.13796532e-01 5.81895590e-01 7.84795821e-01 -3.49728882e-01
3.13857585e-01 -4.26009819e-02 -5.16925473e-03 6.41621232e-01
3.83107245e-01 6.58902526e-01 -4.19274926e-01 -3.83540928e-01
3.98710698e-01 -1.80852294e-01 -1.55167088e-01 -2.26059377e-01
-2.78487444e-01 1.11480117e+00 7.14521766e-01 4.52189147e-02
-1.97483167e-01 2.33531758e-01 1.01199484e+00 4.70429808e-01
4.00546491e-01 3.45331907e-01 -4.75080431e-01 3.91364843e-01
-8.12354505e-01 3.16668153e-01 6.96727931e-01 6.62991524e-01
-1.38080254e-01 1.09308088e+00 -6.40798286e-02 7.12610722e-01
3.38092715e-01 4.89754111e-01 1.98453963e-01 -6.52370870e-01
5.71983814e-01 -2.02336103e-01 -4.84758198e-01 -1.22142184e+00
-2.80689836e-01 -5.61040103e-01 -1.08826780e+00 1.00136828e+00
-3.54035362e-03 -6.85915828e-01 -8.85716558e-01 1.63288903e+00
1.33347467e-01 1.70625746e-01 3.11372668e-01 4.21935976e-01
9.15240824e-01 2.94690371e-01 1.77397668e-01 1.44678112e-02
1.12688410e+00 -8.60691786e-01 -8.44930708e-01 -2.32183531e-01
9.81687307e-02 -6.90706551e-01 3.20775092e-01 5.25444984e-01
-8.81861687e-01 -4.50526983e-01 -1.51976919e+00 3.99150103e-01
-8.75143766e-01 -4.96360272e-01 6.40086532e-01 1.76001453e+00
-7.53246009e-01 8.96313250e-01 -7.22109318e-01 1.57810926e-01
9.29770529e-01 4.46063489e-01 -1.79297298e-01 1.72554940e-01
-1.80348957e+00 9.52854037e-01 1.31918654e-01 2.57527739e-01
-1.27868617e+00 -7.35717893e-01 -7.90038943e-01 -1.99629053e-01
-6.04074970e-02 -3.10592264e-01 9.01704371e-01 -6.37370765e-01
-1.49245262e+00 5.89125335e-01 6.77497864e-01 -9.75062013e-01
4.89714503e-01 -5.00119925e-01 -1.05639327e+00 2.83897549e-01
-5.41341960e-01 1.97512448e-01 1.13758552e+00 -1.17845881e+00
-5.03479898e-01 -3.94170672e-01 4.12689149e-01 -1.90308318e-01
-8.60805690e-01 5.01851141e-01 -9.89022776e-02 -1.17370903e+00
-2.64531881e-01 -6.96916938e-01 -3.52102578e-01 2.46061057e-01
-6.73990369e-01 3.47021163e-01 1.41883171e+00 -5.88696361e-01
1.41561842e+00 -2.24470401e+00 -2.66645253e-01 1.88641876e-01
3.09261382e-01 1.05361259e+00 -1.23077296e-01 7.40178764e-01
-4.22203600e-01 2.79412955e-01 -1.66808352e-01 -1.75949544e-01
9.40578356e-02 -8.59684963e-03 -8.90001714e-01 7.83365607e-01
1.10292248e-01 6.61809027e-01 -4.54554141e-01 5.80232739e-02
3.43222409e-01 7.05738127e-01 -4.00034845e-01 1.92497164e-01
3.01969111e-01 3.27743530e-01 -3.55230063e-01 9.14272249e-01
1.13442028e+00 5.72092593e-01 -6.29687160e-02 -3.52811635e-01
2.59350419e-01 1.87879413e-01 -1.04388213e+00 1.02524590e+00
-5.19122064e-01 7.80048311e-01 6.03877068e-01 -1.14704990e+00
7.91852415e-01 3.20333302e-01 2.71815002e-01 -5.75667918e-01
2.74667233e-01 2.03450650e-01 1.84600830e-01 -3.15827489e-01
2.33390592e-02 1.45566359e-01 -1.34291217e-01 2.44753629e-01
9.59223509e-03 1.83287174e-01 -2.58092254e-01 3.60317469e-01
1.32852054e+00 -1.90345764e-01 5.53174838e-02 -1.02841035e-01
6.91430748e-01 -4.37710285e-01 3.25446159e-01 1.03131175e+00
-7.57723272e-01 1.67750508e-01 3.60252857e-01 -5.07697999e-01
-5.11422396e-01 -1.19587386e+00 -2.99336910e-01 9.69260037e-01
2.31993377e-01 -3.00111234e-01 -9.97299552e-01 -1.16441429e+00
7.39129484e-02 4.24085736e-01 -8.86385381e-01 -7.56978393e-01
-6.98215663e-01 -1.09612894e+00 1.53404808e+00 6.90454066e-01
8.94562781e-01 -1.12326598e+00 -3.74394715e-01 -8.03180858e-02
2.58263171e-01 -1.07242596e+00 1.26772830e-02 4.73468840e-01
-8.03557813e-01 -1.05153465e+00 -5.07376075e-01 -6.40602708e-01
3.79598200e-01 1.40499994e-01 9.17152643e-01 1.99980438e-01
-4.89117533e-01 3.27165574e-01 -3.65899891e-01 -7.51766264e-01
-5.78557014e-01 -3.65814328e-01 5.39213717e-01 -1.94854990e-01
2.17299566e-01 -6.23357952e-01 -6.35094285e-01 1.78539813e-01
-9.41890180e-01 -1.04598331e+00 5.22456229e-01 9.49365556e-01
3.99123738e-03 2.96207964e-01 5.26767015e-01 -7.10922003e-01
8.28454375e-01 -3.57266456e-01 -4.80582833e-01 7.10292011e-02
-5.21399021e-01 -2.41971061e-01 9.66024995e-01 -5.96541345e-01
-8.71492445e-01 -4.70950246e-01 -9.11075771e-01 -5.66921413e-01
-1.15201987e-01 -6.74353093e-02 -3.51084799e-01 -9.98243213e-01
1.16582656e+00 -1.12693524e-02 -1.69451892e-01 -3.88830155e-01
3.40293378e-01 5.22594154e-01 4.42380965e-01 -5.58001339e-01
1.30495715e+00 5.66774487e-01 1.05348684e-01 -7.18361974e-01
-6.69201970e-01 1.98967695e-01 -4.65578258e-01 -1.65367305e-01
5.23481309e-01 -4.59092915e-01 -7.43485749e-01 1.02484953e+00
-9.76007700e-01 -2.29451671e-01 -1.39202848e-01 1.54500231e-01
-1.42609894e-01 6.79250777e-01 -1.12962031e+00 -8.70163560e-01
-7.94712484e-01 -1.33768427e+00 4.71962780e-01 1.55489713e-01
3.04236710e-01 -1.25809717e+00 -1.35163754e-01 1.31375581e-01
8.17997575e-01 7.33070374e-01 8.52513790e-01 -1.08611786e+00
-2.30032146e-01 -7.20313072e-01 -3.43574807e-02 8.42886627e-01
-3.28664258e-02 4.06808704e-02 -1.30840790e+00 -6.74487650e-01
3.00759733e-01 -4.88504201e-01 8.34154248e-01 4.25506651e-01
1.37226415e+00 -5.41684866e-01 -2.68858850e-01 1.09987581e+00
1.37791371e+00 4.95539933e-01 7.18626797e-01 6.28739297e-01
7.10736811e-01 5.88014185e-01 1.82989374e-01 1.32030100e-01
-7.01875746e-01 4.42329109e-01 1.16245759e+00 -2.12311611e-01
-5.41626438e-02 7.46078342e-02 4.55468327e-01 3.89388055e-01
2.18883917e-01 -4.87711728e-01 -5.00533819e-01 -9.11907628e-02
-1.18822718e+00 -1.04516065e+00 8.85690451e-02 1.99598789e+00
4.60757554e-01 6.56624079e-01 4.46372516e-02 6.14971101e-01
8.00499558e-01 6.61410868e-01 -8.54607642e-01 -8.76886725e-01
-1.48135290e-01 3.89716297e-01 8.52026045e-01 3.87006193e-01
-1.66643167e+00 8.65588069e-01 7.74970436e+00 1.23852479e+00
-1.20721781e+00 -8.30603614e-02 6.70773625e-01 -1.09866753e-01
3.57297629e-01 -6.34046733e-01 -7.71740377e-01 2.84157366e-01
1.04675603e+00 1.45254016e-01 3.09766173e-01 1.01963925e+00
-2.93979168e-01 6.74948812e-01 -6.42479956e-01 7.47508705e-01
2.46334359e-01 -1.50319386e+00 2.70895809e-01 6.44045174e-02
4.03451025e-01 1.83499947e-01 6.02468014e-01 3.54000598e-01
3.52217704e-01 -9.78690624e-01 5.54750502e-01 -1.68313578e-01
8.43158305e-01 -1.20269847e+00 6.15106463e-01 -1.76834911e-02
-1.00497639e+00 -5.69357932e-01 -3.33724588e-01 8.06922466e-02
-3.47085185e-02 2.40129337e-01 2.66990483e-01 7.35104442e-01
9.82527673e-01 4.02698725e-01 -4.30334985e-01 7.48458087e-01
-4.04622942e-01 6.82659090e-01 8.86447281e-02 1.76657028e-02
3.29899043e-01 3.00356805e-01 7.31683075e-01 1.42676032e+00
-2.66440660e-01 -3.53898630e-02 -1.33526186e-02 4.34341073e-01
-8.14667642e-02 -3.32150757e-01 -5.24566710e-01 8.51977989e-02
6.96823478e-01 1.14989281e+00 -4.29237247e-01 5.80866598e-02
-2.55350918e-01 8.20441902e-01 -2.93375775e-02 2.93325365e-01
-1.23736548e+00 -1.02100897e+00 1.16278780e+00 -4.10353303e-01
-8.39845315e-02 -2.41572440e-01 -5.01580298e-01 -7.50695527e-01
-8.80913213e-02 -1.45876920e+00 5.53558350e-01 -2.14000065e-02
-1.52174699e+00 1.04105258e+00 -1.43480673e-02 -1.24929762e+00
7.05030039e-02 -1.11940050e+00 -9.60459709e-01 8.91744077e-01
-1.30539238e+00 -8.79069984e-01 7.50637334e-03 6.82897866e-01
5.48475921e-01 -8.36137950e-01 1.09073746e+00 6.99690700e-01
-9.51068640e-01 1.43190682e+00 3.47157836e-01 6.25429988e-01
5.41539729e-01 -8.13187897e-01 1.15717208e+00 1.27602327e+00
7.28308856e-02 9.13349569e-01 7.44119763e-01 -4.26331192e-01
-1.27577388e+00 -1.11459053e+00 -3.39909680e-02 -3.45564246e-01
7.93410182e-01 -6.79396570e-01 -1.00853133e+00 2.71128416e-01
4.19830173e-01 1.19815029e-01 7.85829425e-01 -1.97891891e-01
-7.86876500e-01 -2.37862349e-01 -1.59090459e+00 8.80947113e-01
9.78615880e-01 -4.95842725e-01 -3.01372129e-02 4.96347070e-01
6.74428642e-01 -6.77133799e-01 -3.51564944e-01 4.59096938e-01
6.11653924e-01 -1.19347560e+00 1.53709292e+00 -9.38597679e-01
1.38402089e-01 2.08082780e-01 -1.36260927e-01 -1.17401576e+00
-4.77571219e-01 -9.73512471e-01 -5.03530443e-01 1.08324671e+00
1.50720522e-01 -7.61337340e-01 8.86182487e-01 3.11339825e-01
-2.57168561e-01 -1.04914057e+00 -1.04152799e+00 -1.11594307e+00
5.60823143e-01 -4.84900802e-01 2.16815606e-01 8.46864820e-01
-3.20565760e-01 -1.83320463e-01 -7.03383505e-01 5.70719004e-01
8.78098845e-01 -7.83432126e-01 5.24333060e-01 -6.98646188e-01
-2.47443631e-01 -5.38032651e-01 -7.37639844e-01 -7.51561403e-01
2.31342837e-01 -5.71098089e-01 -3.35626006e-01 -1.11303103e+00
-5.45795560e-01 -2.82511562e-01 -6.02798641e-01 3.37288439e-01
-7.06515014e-02 6.81828737e-01 1.10366665e-01 -5.55089787e-02
-8.44523683e-02 2.86102176e-01 6.12094343e-01 -3.98923725e-01
2.90955335e-01 2.79500604e-01 -7.93424726e-01 8.49432766e-01
1.12800002e+00 -6.11854672e-01 -4.27534312e-01 -4.65408415e-01
-2.93892354e-01 -4.21468228e-01 3.61814022e-01 -1.30647671e+00
9.36741456e-02 7.16531649e-02 6.04886949e-01 -1.69594496e-01
5.70514321e-01 -8.95166636e-01 -4.98511910e-01 1.10917413e+00
-5.09966493e-01 2.04739034e-01 6.69987023e-01 7.80448556e-01
-2.00515166e-02 -2.27074325e-01 1.42859709e+00 1.67362615e-01
-5.93394101e-01 5.02439082e-01 -4.43534464e-01 1.39585048e-01
1.13008463e+00 -1.67181522e-01 -5.73021352e-01 -3.65047008e-01
-5.99321723e-01 -1.24421373e-01 -3.05194676e-01 7.26976871e-01
7.25462437e-01 -1.22864282e+00 -6.79088175e-01 3.85769129e-01
-2.42386088e-01 -9.02298748e-01 5.46432853e-01 1.62780419e-01
-5.96098006e-01 3.24132144e-01 -5.23687661e-01 4.46897484e-02
-1.54453635e+00 9.93573666e-01 6.94582462e-01 -2.17474490e-01
-5.66364884e-01 1.27301908e+00 1.95034876e-01 -1.38682321e-01
9.80890036e-01 5.13123453e-01 -3.71932983e-01 -5.18296182e-01
8.60115290e-01 5.25769293e-01 4.03946400e-01 -3.94588917e-01
-6.57078862e-01 2.22152978e-01 -3.06991041e-01 1.81495070e-01
9.15982485e-01 2.39187062e-01 1.34759530e-01 -3.29000503e-01
1.18375933e+00 2.38011718e-01 -1.11947203e+00 -1.24362512e-02
-4.40225124e-01 -3.83491695e-01 -6.46475892e-05 -1.00427222e+00
-1.47108507e+00 1.22450066e+00 8.79975259e-01 3.92289460e-01
1.34061086e+00 -6.49839818e-01 9.82628584e-01 4.17586595e-01
1.95868060e-01 -7.70648062e-01 3.35470349e-01 6.35102630e-01
9.07158196e-01 -8.39370370e-01 1.36383161e-01 -4.15100545e-01
-3.12220514e-01 1.04616153e+00 7.12763071e-01 -5.63872814e-01
1.16913521e+00 6.82082117e-01 3.72564405e-01 1.08009912e-01
-2.70498008e-01 4.52967346e-01 2.04847991e-01 1.22888041e+00
1.46836579e-01 -5.00401378e-01 -4.48615514e-02 4.14611369e-01
-9.86577645e-02 -9.44640875e-01 8.29323456e-02 1.10013223e+00
-1.06004916e-01 -1.31171954e+00 -6.50451779e-01 1.97642595e-01
-1.25846684e+00 -3.21958065e-01 -5.01785159e-01 7.10904598e-01
-1.07075563e-02 1.22528398e+00 -4.31279659e-01 -8.59552741e-01
5.43765604e-01 -3.31220239e-01 2.36734763e-01 -2.84815103e-01
-9.76965010e-01 -4.39051926e-01 8.71308707e-03 -5.14049470e-01
8.05393010e-02 -9.87424478e-02 -7.36837924e-01 -6.51660085e-01
-3.67648065e-01 6.30511940e-02 9.27203178e-01 7.00915515e-01
3.65897119e-01 9.69193697e-01 8.46885443e-01 -1.03804696e+00
-1.09682751e+00 -5.21488667e-01 -4.59856302e-01 4.30968739e-02
6.42611146e-01 -5.86843312e-01 -9.38375533e-01 -4.50132787e-01] | [5.589046955108643, 7.8387556076049805] |
257901c5-2f17-4bee-9e27-0faa398c02e3 | inference-post-selection-of-group-sparse | 2012.15664 | null | https://arxiv.org/abs/2012.15664v4 | https://arxiv.org/pdf/2012.15664v4.pdf | Approximate Post-Selective Inference for Regression with the Group LASSO | After selection with the Group LASSO (or generalized variants such as the overlapping, sparse, or standardized Group LASSO), inference for the selected parameters is unreliable in the absence of adjustments for selection bias. In the penalized Gaussian regression setup, existing approaches provide adjustments for selection events that can be expressed as linear inequalities in the data variables. Such a representation, however, fails to hold for selection with the Group LASSO and substantially obstructs the scope of subsequent post-selective inference. Key questions of inferential interest -- for example, inference for the effects of selected variables on the outcome -- remain unanswered. In the present paper, we develop a consistent, post-selective, Bayesian method to address the existing gaps by deriving a likelihood adjustment factor and an approximation thereof that eliminates bias from the selection of groups. Experiments on simulated data and data from the Human Connectome Project demonstrate that our method recovers the effects of parameters within the selected groups while paying only a small price for bias adjustment. | ['Daniel Kessler', 'Peter W. MacDonald', 'Snigdha Panigrahi'] | 2020-12-31 | null | null | null | null | ['sparse-learning'] | ['methodology'] | [ 4.84890968e-01 2.70589530e-01 -7.64180660e-01 -7.96011269e-01
-7.67109096e-01 -3.75057817e-01 1.17264375e-01 2.90994853e-01
-3.96669596e-01 1.23344362e+00 4.31952715e-01 -3.60220373e-01
-5.11425376e-01 -6.10584378e-01 -6.95132554e-01 -6.34033263e-01
-3.27743143e-01 3.29176158e-01 -3.45404685e-01 4.23516065e-01
2.24658817e-01 9.54581946e-02 -8.90385747e-01 -4.10094470e-01
1.24268460e+00 4.64272410e-01 -3.42056662e-01 6.03687167e-02
4.33645695e-01 2.52245069e-01 -5.61637580e-02 -3.51642251e-01
1.73428804e-01 -4.20832247e-01 -2.06349269e-01 -1.97899610e-01
5.60206532e-01 -4.17433679e-01 -1.25105396e-01 1.01865900e+00
6.72282040e-01 2.91062146e-01 7.73183584e-01 -1.25545955e+00
-3.37796897e-01 1.02440250e+00 -7.17819214e-01 -1.43985311e-02
1.68531001e-01 -1.59370810e-01 1.06591666e+00 -7.95836329e-01
6.86027884e-01 1.42853296e+00 9.63960171e-01 3.50085735e-01
-1.83273184e+00 -8.35220516e-01 3.04248542e-01 -6.44667625e-01
-1.42724609e+00 -5.80196440e-01 3.01453561e-01 -6.75988197e-01
4.22001660e-01 4.24998015e-01 3.88409972e-01 1.17431021e+00
2.94348806e-01 3.33161891e-01 1.12890685e+00 -2.43809730e-01
3.93833935e-01 2.00133435e-02 5.37357211e-01 5.90553522e-01
7.97987640e-01 3.88559341e-01 -6.36137426e-01 -1.00734329e+00
7.99021482e-01 5.95846660e-02 -2.94378310e-01 -4.12080586e-01
-1.27591455e+00 1.15753806e+00 2.30623633e-02 -3.48186225e-01
-4.13793653e-01 2.55235463e-01 3.58938545e-01 -2.84378044e-03
9.08308387e-01 3.40712816e-01 -4.83728826e-01 2.86253124e-01
-1.17178833e+00 4.93871450e-01 6.56853914e-01 8.11760485e-01
5.26392043e-01 8.93036500e-02 -5.67944407e-01 5.93180299e-01
3.29906076e-01 6.14564240e-01 -1.27671629e-01 -1.04158473e+00
4.30215180e-01 5.02759404e-02 4.86157507e-01 -8.50125492e-01
-6.99249089e-01 -4.77812976e-01 -9.94664252e-01 -1.14504015e-02
5.80015004e-01 -7.85118520e-01 -7.62218893e-01 2.35440445e+00
6.28221869e-01 3.55027378e-01 -5.12674928e-01 7.35824645e-01
3.42532784e-01 1.34046525e-01 4.49909538e-01 -7.06024706e-01
1.12387264e+00 -3.51070911e-01 -8.38639438e-01 -2.49911949e-01
5.31980336e-01 -3.03648770e-01 8.21445167e-01 1.83166236e-01
-1.09279764e+00 -7.93389603e-02 -6.38611913e-01 6.53290302e-02
3.11514437e-01 -2.02081203e-01 1.00650752e+00 5.76123178e-01
-5.64614415e-01 5.00523210e-01 -9.53421056e-01 -3.43977869e-01
5.87928653e-01 5.37348092e-01 -2.33152404e-01 1.10688791e-01
-1.06700146e+00 7.15890527e-01 -1.63935736e-01 1.47602037e-01
-7.95516729e-01 -1.28821242e+00 -8.07988524e-01 2.73098946e-01
5.44433594e-01 -1.03014135e+00 6.55446827e-01 -7.82424986e-01
-8.52123976e-01 5.24859488e-01 -4.08756524e-01 -4.80187207e-01
7.70900309e-01 -2.15127856e-01 1.61286607e-01 -2.19876409e-01
4.75727439e-01 2.64971524e-01 8.39279234e-01 -9.41511154e-01
-3.68659735e-01 -6.10709071e-01 -2.97079951e-01 1.55851811e-01
7.64689222e-02 3.76017749e-01 -1.11815952e-01 -9.28698301e-01
2.17506215e-01 -1.08303523e+00 -7.81315863e-01 1.30012661e-01
-6.51512861e-01 -4.23816964e-02 -1.64754599e-01 -7.48848677e-01
1.35329497e+00 -2.30947661e+00 2.83696830e-01 4.63819057e-01
2.95534462e-01 -5.61652064e-01 2.06640795e-01 6.42804876e-02
-1.47612080e-01 4.32540178e-01 -4.99263376e-01 -3.58895183e-01
5.24407662e-02 -5.51787801e-02 -2.61308879e-01 1.03403068e+00
4.49403003e-02 5.86085081e-01 -6.67848945e-01 -5.92610478e-01
-1.15129746e-01 3.69175762e-01 -1.07692349e+00 -8.88124630e-02
3.14251721e-01 4.89704907e-01 -7.69315958e-01 4.11339432e-01
6.09876633e-01 -1.31295830e-01 4.41331923e-01 -6.57858467e-03
-1.23314917e-01 2.83078045e-01 -1.31563747e+00 1.17150927e+00
8.83214250e-02 2.78625607e-01 4.93080884e-01 -1.09393060e+00
5.88069677e-01 2.54035890e-01 6.66200161e-01 7.68126845e-02
-1.55618533e-01 2.40355417e-01 8.73993617e-03 -3.17731768e-01
6.54999763e-02 -4.50918227e-01 -3.10374528e-01 2.99636900e-01
-1.66891396e-01 -1.78405102e-02 1.06254583e-02 1.42971843e-01
8.30135942e-01 9.01843011e-02 4.60625321e-01 -5.61828434e-01
-1.55878022e-01 -1.24961399e-01 1.19892645e+00 1.34097683e+00
1.09897153e-02 7.06596196e-01 6.36698604e-01 -4.47069369e-02
-8.43425751e-01 -1.21208799e+00 -7.84682870e-01 9.06385005e-01
-2.09457397e-01 -5.40135875e-02 -4.09488708e-01 -5.67043960e-01
5.70818961e-01 8.58613253e-01 -7.36409843e-01 -2.45428190e-01
-3.94836426e-01 -1.58887708e+00 4.50996220e-01 5.10429204e-01
-2.93129325e-01 -4.91665334e-01 -4.48511392e-01 2.06515029e-01
-2.06667241e-02 -7.35676527e-01 -5.28665006e-01 1.54284120e-01
-9.94280815e-01 -9.79341567e-01 -4.50725049e-01 -3.39849919e-01
9.27903950e-01 -3.82758617e-01 8.62024188e-01 -1.74051985e-01
7.70178139e-02 1.58178732e-01 3.06442063e-02 -5.54006100e-01
9.77130756e-02 -3.55153643e-02 3.36980581e-01 6.81159832e-03
1.77490681e-01 -6.20737135e-01 -4.76570010e-01 1.37401477e-01
-4.73361373e-01 -1.58307686e-01 8.68870094e-02 1.09557950e+00
6.62681699e-01 -2.94859082e-01 1.07734525e+00 -1.37214553e+00
4.67740446e-01 -7.69865394e-01 -6.95544004e-01 1.09558687e-01
-8.85571957e-01 -2.08795875e-01 -5.86387888e-02 -8.05478454e-01
-1.10286105e+00 -1.64372832e-01 3.90371352e-01 -3.88939194e-02
2.29181245e-01 9.97153282e-01 -1.14736587e-01 7.39584640e-02
5.06896198e-01 -4.64794040e-01 -7.34801069e-02 -5.53256392e-01
4.71196100e-02 4.12997276e-01 1.99490175e-01 -8.16565990e-01
5.14426589e-01 6.44750595e-01 2.28089914e-01 -4.39318925e-01
-9.75207865e-01 -1.14975698e-01 -2.93463200e-01 3.41941983e-01
6.56027436e-01 -1.04592216e+00 -5.69261611e-01 -2.62696687e-02
-5.81343353e-01 -4.46948558e-01 -5.09155571e-01 1.10074496e+00
-5.69426715e-01 3.09973687e-01 -4.69456404e-01 -1.02277136e+00
-2.14309812e-01 -1.11203694e+00 7.10415840e-01 -5.32718608e-03
-5.80963433e-01 -1.02282751e+00 3.09828371e-02 2.71936148e-01
2.56483883e-01 4.55700517e-01 1.00934863e+00 -7.48191655e-01
-2.05740094e-01 -2.81197459e-01 -2.42551323e-02 -7.57674268e-03
8.82517081e-03 5.30970357e-02 -5.19681931e-01 -3.59240741e-01
2.88580786e-02 5.83793670e-02 7.94627666e-01 1.26691234e+00
1.22936106e+00 -4.15819883e-01 -5.09717822e-01 8.17427933e-01
1.20287788e+00 -1.65497616e-01 2.79341877e-01 -2.17423424e-01
5.70454836e-01 5.88360131e-01 5.48451364e-01 8.55124652e-01
3.17993253e-01 6.53121054e-01 -1.74417663e-02 -2.46352069e-02
3.28200847e-01 -3.37194473e-01 1.60197124e-01 2.04122007e-01
-5.82908653e-03 -1.14211716e-01 -6.54028952e-01 5.88128328e-01
-1.92864454e+00 -8.45701754e-01 -2.60236740e-01 2.64942789e+00
1.16547287e+00 -1.68559458e-02 1.60190001e-01 -3.56577843e-01
9.23188686e-01 -1.30951285e-01 -7.68152177e-01 -2.34343842e-01
-2.15436891e-01 6.55452386e-02 9.03082192e-01 9.08391356e-01
-9.52554345e-01 5.04741848e-01 7.71479559e+00 5.18504918e-01
-5.67184508e-01 1.73452184e-01 7.85045922e-01 -3.69976997e-01
-4.90921736e-01 5.26549816e-01 -8.01629007e-01 3.92704338e-01
7.69508421e-01 -3.39059353e-01 3.94949377e-01 3.79346341e-01
8.03691387e-01 -2.60888994e-01 -1.17131317e+00 4.05242622e-01
-1.65795520e-01 -8.36040676e-01 -3.31565261e-01 2.93789744e-01
8.58122587e-01 -5.86537048e-02 -5.74546866e-02 7.92535767e-02
5.76153934e-01 -1.12197852e+00 5.96239269e-01 6.88828945e-01
9.03940916e-01 -4.37268853e-01 5.96146584e-01 1.21033698e-01
-3.23080719e-01 -7.59509131e-02 -3.01375896e-01 -1.45440102e-01
4.72656548e-01 1.12681949e+00 -4.91850078e-01 2.71893352e-01
4.12206799e-01 6.04003489e-01 -7.56617114e-02 1.07698703e+00
-9.92587581e-02 1.19345701e+00 -6.13832951e-01 4.01259691e-01
-3.46489668e-01 -4.69615489e-01 9.37882245e-01 9.91505265e-01
3.85790229e-01 2.68878397e-02 1.23528793e-01 1.01320016e+00
1.39840534e-02 7.81561881e-02 -3.58561993e-01 -5.96629232e-02
7.23088741e-01 8.50950360e-01 -3.40133548e-01 -2.20528811e-01
-4.74299788e-01 2.10317716e-01 2.33989790e-01 8.20957899e-01
-7.65259743e-01 3.89046893e-02 5.13586998e-01 2.89416343e-01
1.61945764e-02 -6.56334832e-02 -7.20842481e-01 -1.30975974e+00
-1.77682996e-01 -1.00164998e+00 6.49198234e-01 -3.74991030e-01
-1.51291883e+00 -3.39665860e-01 4.71778691e-01 -4.80379641e-01
-1.48828760e-01 -1.75461680e-01 -3.15005511e-01 1.11033666e+00
-9.00160372e-01 -9.35398340e-01 2.22761542e-01 2.78600901e-01
1.05283499e-01 2.56503165e-01 5.98434627e-01 2.55918324e-01
-9.04371083e-01 6.65684223e-01 1.37861639e-01 -1.25191584e-01
1.06231296e+00 -1.11550343e+00 -2.22482219e-01 7.47653365e-01
-3.92391860e-01 1.10387242e+00 1.10933518e+00 -1.10333562e+00
-1.03539658e+00 -8.17582428e-01 8.47773075e-01 -3.94171268e-01
6.55811906e-01 -3.07549089e-01 -6.85758710e-01 1.26674318e+00
-4.60120678e-01 6.70632720e-02 8.09214175e-01 7.72946894e-01
-2.25888297e-01 -7.37031400e-02 -1.37228394e+00 5.84735990e-01
1.13233173e+00 -2.01119199e-01 -3.25056344e-01 4.36262220e-01
5.39928496e-01 -1.42092228e-01 -1.17160904e+00 7.25654542e-01
7.94183314e-01 -4.90993768e-01 1.09028876e+00 -1.14166224e+00
1.10051237e-01 -6.99106902e-02 -2.82501996e-01 -1.10915089e+00
-3.50336641e-01 -8.13871801e-01 6.34308457e-02 1.40018797e+00
4.94705290e-01 -8.76923382e-01 5.41768670e-01 1.16663933e+00
1.43024147e-01 -4.94273692e-01 -1.07113349e+00 -3.35332006e-01
3.30825508e-01 -2.03244776e-01 4.52085763e-01 1.23016834e+00
1.10246725e-02 1.93627566e-01 -6.23138070e-01 2.73090094e-01
1.10076749e+00 1.54578865e-01 5.20930409e-01 -1.31877756e+00
-3.89990985e-01 -1.15818577e-02 -1.53625801e-01 -3.94502014e-01
4.40948904e-01 -6.50823295e-01 -5.36257178e-02 -1.22710896e+00
6.62803829e-01 -9.07509506e-01 -1.23179980e-01 5.61137974e-01
-6.67238712e-01 -1.16297804e-01 -1.09384418e-01 -1.25153616e-01
-1.11773610e-01 4.36812609e-01 9.39081788e-01 4.42574210e-02
-4.06672120e-01 2.31204137e-01 -1.04757810e+00 8.64887416e-01
7.83833444e-01 -9.52154756e-01 -3.79582196e-01 -1.73357278e-01
6.13364756e-01 2.86822498e-01 5.51400542e-01 -2.71626711e-01
-4.29998115e-02 -6.38393283e-01 2.50698864e-01 -2.55065829e-01
1.06962107e-01 -4.45616215e-01 4.51237649e-01 3.10944349e-01
-7.92144120e-01 -1.67752311e-01 -1.20897009e-03 6.30960703e-01
2.60868341e-01 -2.42011383e-01 5.42591929e-01 1.95782572e-01
3.15074295e-01 2.02217579e-01 -2.98373580e-01 4.54798102e-01
5.46319246e-01 2.45763019e-01 -1.94349542e-01 -4.63677406e-01
-1.11114192e+00 4.85677570e-01 3.26882184e-01 -1.15215451e-01
2.66299725e-01 -1.14662242e+00 -1.32075465e+00 -8.39630887e-03
-3.23940039e-01 -1.32904081e-02 9.65765715e-02 1.26718783e+00
1.68147877e-01 5.46169057e-02 2.51773685e-01 -2.75602758e-01
-1.13530469e+00 4.16804999e-01 3.11514407e-01 -2.44906032e-03
-4.38537300e-01 5.01039386e-01 5.45307159e-01 -4.84430164e-01
1.02473199e-01 -2.33218342e-01 2.06318527e-01 2.03044802e-01
2.63203442e-01 6.41510010e-01 -5.24178207e-01 -3.90507281e-01
-4.61058557e-01 2.32966632e-01 3.00409436e-01 -1.52459174e-01
1.46215129e+00 -6.11829698e-01 -4.37019765e-01 6.59546971e-01
7.75084078e-01 4.43234742e-01 -1.29416716e+00 1.58941057e-02
-2.63467163e-01 -4.10774946e-01 1.08367085e-01 -6.48566186e-01
-7.17234790e-01 4.45580512e-01 2.67256051e-01 6.27193786e-03
8.15921485e-01 -1.79681718e-01 -6.80004656e-02 -7.80798495e-02
3.31357807e-01 -8.64294887e-01 -9.84520495e-01 -8.36511254e-02
7.64312685e-01 -1.16294801e+00 5.70249975e-01 -6.95944667e-01
-2.84248680e-01 5.30789912e-01 2.77778894e-01 -2.76375711e-01
7.94265807e-01 2.24747255e-01 -4.10548240e-01 -7.83790722e-02
-6.81051791e-01 2.87624359e-01 3.62731487e-01 4.91172761e-01
5.80075204e-01 2.92753071e-01 -1.11757541e+00 9.27027225e-01
5.11975586e-02 8.31386149e-02 5.10862827e-01 5.82986474e-01
3.23357359e-02 -6.58890724e-01 -4.53796357e-01 1.09961450e+00
-9.20060754e-01 -2.75958717e-01 9.06459615e-02 6.82943344e-01
6.25938326e-02 8.77384663e-01 1.37663200e-01 3.49377334e-01
5.95475286e-02 5.67730889e-02 2.94389546e-01 -7.18469501e-01
-4.33474243e-01 4.84596968e-01 4.58199888e-01 -5.74989736e-01
-6.12732530e-01 -1.28438342e+00 -8.66001666e-01 -2.42623657e-01
-7.12970436e-01 2.36029595e-01 1.72064424e-01 8.62517118e-01
2.14340001e-01 2.45713711e-01 3.77726048e-01 -5.88595808e-01
-1.06544590e+00 -9.13496733e-01 -7.60194063e-01 2.44460583e-01
4.67443734e-01 -7.69715130e-01 -8.24405015e-01 -1.48562118e-01] | [7.852373123168945, 5.1396613121032715] |
ee08b86e-71a8-4d45-a6a0-a7d0882d0ac7 | unsupervised-clinical-language-translation | 1902.01177 | null | https://arxiv.org/abs/1902.01177v2 | https://arxiv.org/pdf/1902.01177v2.pdf | Unsupervised Clinical Language Translation | As patients' access to their doctors' clinical notes becomes common, translating professional, clinical jargon to layperson-understandable language is essential to improve patient-clinician communication. Such translation yields better clinical outcomes by enhancing patients' understanding of their own health conditions, and thus improving patients' involvement in their own care. Existing research has used dictionary-based word replacement or definition insertion to approach the need. However, these methods are limited by expert curation, which is hard to scale and has trouble generalizing to unseen datasets that do not share an overlapping vocabulary. In contrast, we approach the clinical word and sentence translation problem in a completely unsupervised manner. We show that a framework using representation learning, bilingual dictionary induction and statistical machine translation yields the best precision at 10 of 0.827 on professional-to-consumer word translation, and mean opinion scores of 4.10 and 4.28 out of 5 for clinical correctness and layperson readability, respectively, on sentence translation. Our fully-unsupervised strategy overcomes the curation problem, and the clinically meaningful evaluation reduces biases from inappropriate evaluators, which are critical in clinical machine learning. | ['Yu-An Chung', 'Wei-Hung Weng', 'Peter Szolovits'] | 2019-02-04 | null | null | null | null | ['clinical-language-translation'] | ['natural-language-processing'] | [ 3.61064732e-01 3.75072956e-01 -5.97196698e-01 -2.43941680e-01
-1.43098783e+00 -6.10313177e-01 -7.38890693e-02 7.84282446e-01
-5.94942629e-01 9.26109672e-01 7.73356199e-01 -7.86955476e-01
-1.50141641e-01 -3.52407753e-01 -3.04081827e-01 -4.57051188e-01
8.16680253e-01 8.17951024e-01 -5.78288138e-01 -1.86781093e-01
9.90047380e-02 2.15718955e-01 -4.85627621e-01 4.41319197e-01
1.21982503e+00 3.36983681e-01 -4.80466150e-02 5.31491756e-01
-3.80823128e-02 6.77766800e-01 -4.18017358e-01 -7.10078657e-01
6.74148872e-02 -6.86162949e-01 -8.59799445e-01 7.16631189e-02
-6.37084022e-02 -3.23280811e-01 1.56142771e-01 9.57581639e-01
8.00685525e-01 -4.26271349e-01 3.85057718e-01 -2.87029773e-01
-1.26454079e+00 6.62343740e-01 -2.38546312e-01 1.95577033e-02
6.22185528e-01 9.83505920e-02 9.46568489e-01 -8.04599762e-01
7.56233275e-01 7.90606737e-01 8.15519333e-01 7.43153512e-01
-1.38220239e+00 -3.98610204e-01 -2.39906222e-01 -3.54788125e-01
-1.31205606e+00 -4.29897666e-01 2.65079886e-01 -6.53605103e-01
9.62788880e-01 4.25243229e-01 8.72379720e-01 8.10833097e-01
7.10006118e-01 1.77702203e-01 1.18189847e+00 -7.11871505e-01
5.75964600e-02 7.37435997e-01 -5.65918759e-02 2.98534125e-01
5.34063756e-01 -1.57979921e-01 -3.62450153e-01 -3.68643999e-01
6.12293363e-01 2.85264581e-01 -5.47843993e-01 1.95791379e-01
-1.41035295e+00 9.41171050e-01 -3.27207781e-02 5.15847683e-01
-5.01854658e-01 -4.04823840e-01 3.47470492e-01 4.69904006e-01
4.26012993e-01 9.50220287e-01 -4.72224951e-01 -3.56022149e-01
-7.80999601e-01 -1.46147773e-01 9.88565505e-01 8.72547865e-01
3.47471498e-02 -4.18174505e-01 -5.11576012e-02 8.55201542e-01
1.41596079e-01 9.59490955e-01 9.06294763e-01 -7.26773083e-01
1.88041538e-01 6.26295090e-01 2.07925722e-01 -9.34821606e-01
-3.20877105e-01 -7.01654792e-01 -7.05281973e-01 -6.59431517e-01
1.75958069e-03 -6.24517128e-02 -7.00089753e-01 1.49660349e+00
5.00678420e-02 -5.10843337e-01 5.08275688e-01 8.50825369e-01
8.90064538e-01 1.92026615e-01 1.96485281e-01 -6.89015627e-01
1.79090786e+00 -8.51625264e-01 -1.12175894e+00 -2.50363261e-01
1.17279208e+00 -1.29857945e+00 1.08073390e+00 3.14270288e-01
-1.13425457e+00 -7.93239549e-02 -6.55611455e-01 -3.39439102e-02
-4.51451959e-03 -6.28935813e-04 3.79858106e-01 7.02353358e-01
-1.04580832e+00 3.66753936e-01 -7.78764129e-01 -6.44555688e-01
4.66782570e-01 4.19121265e-01 -4.96071875e-01 -1.93527848e-01
-1.06978130e+00 1.33146000e+00 -7.36319050e-02 -1.09085165e-01
-2.92304635e-01 -6.91055834e-01 -6.89429760e-01 -2.93815076e-01
3.30109112e-02 -1.04448068e+00 1.26822221e+00 -9.61298287e-01
-1.14405024e+00 1.27358294e+00 -3.36508691e-01 -2.25962058e-01
2.34364927e-01 -1.60431519e-01 -6.70851946e-01 1.80411533e-01
3.29475105e-01 4.93226260e-01 3.02368760e-01 -8.45319211e-01
-3.34585577e-01 -2.70915926e-01 -2.43855134e-01 3.19238335e-01
-3.87952745e-01 1.67543530e-01 -7.87760317e-02 -7.18775749e-01
2.53044754e-01 -9.49158967e-01 -6.63336575e-01 4.27124538e-02
-1.34670347e-01 2.53597468e-01 -1.91533044e-01 -9.97897208e-01
1.42052138e+00 -2.15687704e+00 5.60697652e-02 2.16803104e-01
5.90029120e-01 3.36226106e-01 -4.99964952e-02 7.52261460e-01
-4.38494906e-02 4.58306670e-01 -1.92256019e-01 -2.22248640e-02
-4.02200103e-01 3.40881646e-01 -3.10674347e-02 3.19232523e-01
1.39780059e-01 1.26888847e+00 -1.13302779e+00 -6.80544555e-01
1.13959480e-02 5.13899446e-01 -6.93922579e-01 8.56987908e-02
3.89043599e-01 7.79496908e-01 -4.90477830e-01 6.10537171e-01
3.26471478e-01 -7.94730008e-01 5.91466010e-01 6.35771826e-02
1.83518469e-01 6.26711071e-01 -4.47419643e-01 1.76558006e+00
-5.91610730e-01 2.76673943e-01 -2.71256894e-01 -5.93857646e-01
7.80691326e-01 5.33179998e-01 6.10198021e-01 -6.41069949e-01
1.50879666e-01 6.70841336e-01 2.69195676e-01 -6.92642629e-01
2.48236969e-01 -7.48227954e-01 1.21027023e-01 4.48203355e-01
-5.11494279e-01 -1.79114401e-01 -9.15673152e-02 1.99906453e-01
9.14255381e-01 -4.13518578e-01 7.66273797e-01 -3.78643692e-01
2.17378810e-01 3.59942675e-01 6.21791482e-01 4.19759572e-01
-5.22142462e-02 7.89498210e-01 2.50217825e-01 -3.01723927e-01
-9.55189645e-01 -6.95123732e-01 -4.32060659e-01 3.04431766e-01
-2.36590415e-01 -4.48592484e-01 -8.35209310e-01 -2.69586772e-01
-9.74996090e-02 8.66513073e-01 -4.56611812e-01 -4.34563965e-01
-3.74245375e-01 -7.20739841e-01 3.83125305e-01 6.66265726e-01
-3.06272507e-01 -7.54476428e-01 -6.17631435e-01 7.34317005e-01
-5.38181424e-01 -1.22313738e+00 -9.18945968e-01 -1.34541139e-01
-9.64384854e-01 -9.75853920e-01 -8.48120749e-01 -9.43548203e-01
9.81218994e-01 2.48105913e-01 1.02200437e+00 6.75404817e-02
-3.61807555e-01 5.71789443e-01 -5.94513476e-01 -4.09783989e-01
-9.13544953e-01 -1.93735689e-01 1.48037329e-01 -3.87841493e-01
8.54541361e-01 -2.31062964e-01 -8.96626532e-01 1.26167163e-01
-1.01579702e+00 1.23810552e-01 8.11964571e-01 1.24976003e+00
7.30520546e-01 -7.70907044e-01 6.51788116e-01 -1.21309865e+00
1.06307697e+00 -2.52084076e-01 1.11989833e-01 2.94409543e-01
-1.26534688e+00 -1.23107977e-01 4.55250412e-01 -4.39898998e-01
-5.95441699e-01 -2.80750006e-01 -1.95724592e-01 3.93130556e-02
1.43172696e-01 8.32423091e-01 4.07617480e-01 1.10993877e-01
9.66928899e-01 1.89012140e-01 3.57458293e-01 -2.78615445e-01
1.80603132e-01 1.02950191e+00 1.10669523e-01 -1.84744477e-01
5.15291274e-01 3.12350065e-01 -5.20746529e-01 -4.93081301e-01
-8.57539773e-01 -6.00398183e-01 -4.23551023e-01 2.87409455e-01
1.04447651e+00 -1.23085713e+00 -3.40094030e-01 -2.09912017e-01
-1.11406898e+00 1.26664400e-01 -7.02499211e-01 9.63620067e-01
-2.92840481e-01 4.67936277e-01 -6.87408686e-01 -4.61224914e-01
-7.85791278e-01 -1.39694309e+00 1.13118947e+00 -4.26793993e-02
-1.03867280e+00 -1.13184345e+00 2.85073191e-01 7.46138215e-01
3.19252104e-01 9.32477564e-02 1.05137181e+00 -8.70192409e-01
-2.62239221e-02 -5.79512119e-01 -1.20142147e-01 5.55527389e-01
7.80795038e-01 -4.98869866e-01 -4.67822969e-01 -3.76143873e-01
4.02291447e-01 -1.51597098e-01 1.49822816e-01 3.31875980e-01
4.68475878e-01 -4.29358274e-01 -2.40897536e-01 2.56054640e-01
1.30197442e+00 4.58535075e-01 5.78862488e-01 1.48173258e-01
5.57831049e-01 5.45380056e-01 5.58003366e-01 4.85080600e-01
5.13225257e-01 4.37197655e-01 -3.84299845e-01 -4.79882091e-01
4.79385536e-03 -2.10692108e-01 1.21796966e-01 1.56078637e+00
-1.03872165e-03 -5.55639565e-02 -1.21269441e+00 7.28940845e-01
-1.53947186e+00 -3.92000556e-01 -7.99541175e-02 2.04293561e+00
1.36452913e+00 -2.47671697e-02 -3.02795887e-01 -1.54010296e-01
5.08099556e-01 -5.83793879e-01 -2.91506916e-01 -6.00362420e-01
-1.70532763e-02 4.54472601e-01 6.10586286e-01 5.46665490e-01
-3.02908421e-01 8.82365584e-01 6.45637035e+00 4.70717996e-01
-1.05410862e+00 3.62004608e-01 7.64263213e-01 2.44257469e-02
-7.48031795e-01 1.50066102e-02 -6.07508600e-01 1.89323261e-01
9.44182277e-01 -2.24665925e-01 3.39684099e-01 5.25703490e-01
6.32904172e-01 5.76518402e-02 -1.31559956e+00 1.18163919e+00
3.17124814e-01 -1.35599649e+00 9.28397998e-02 1.70086876e-01
8.37995410e-01 -4.46032047e-01 1.07310198e-01 -2.96223044e-01
7.22347125e-02 -1.28316271e+00 2.54960448e-01 5.26541591e-01
1.23311925e+00 -2.72515476e-01 1.20623517e+00 1.86677545e-01
-5.42805552e-01 2.80883819e-01 -2.14540660e-01 8.67419392e-02
2.47636691e-01 6.03238642e-01 -1.19458592e+00 3.25903833e-01
2.65872627e-01 5.73973417e-01 -9.76028964e-02 4.88186866e-01
-6.37110472e-02 5.38238108e-01 -7.72100389e-02 -7.85236061e-02
9.10703316e-02 -1.10073902e-01 2.75090784e-01 1.13378191e+00
2.54612863e-01 5.12701213e-01 1.79149777e-01 5.23295760e-01
-1.48717672e-01 8.52999806e-01 -6.11850977e-01 -6.28851891e-01
2.93536365e-01 9.74892735e-01 -6.84609413e-01 -4.35448050e-01
-5.08067369e-01 9.07102287e-01 -6.03617728e-02 1.28043354e-01
-6.15907490e-01 -2.74333835e-01 4.70797390e-01 4.24523830e-01
-1.46635741e-01 2.16917485e-01 -6.83831692e-01 -1.29577374e+00
2.01453149e-01 -1.77010202e+00 1.86352253e-01 -4.91308957e-01
-1.23103809e+00 7.20356941e-01 -5.92849255e-01 -1.54414582e+00
-1.47520334e-01 -5.21622062e-01 9.93058309e-02 9.52392101e-01
-1.29878771e+00 -1.10175717e+00 1.73077434e-01 2.61957109e-01
3.77370536e-01 -1.36637747e-01 1.34171236e+00 4.51868415e-01
-2.87116498e-01 9.05049264e-01 2.32144907e-01 1.18312046e-01
1.02911949e+00 -7.87175179e-01 2.78562568e-02 3.38601291e-01
-1.92164853e-01 1.26699948e+00 4.04251784e-01 -7.61240423e-01
-1.37933111e+00 -7.65565395e-01 1.70232403e+00 -7.11815894e-01
7.14274764e-01 1.00401595e-01 -8.30512762e-01 4.51691329e-01
1.67417616e-01 -6.85411394e-01 1.52407753e+00 -2.52089184e-02
-1.80496007e-01 -7.75460433e-03 -1.06469572e+00 8.34452152e-01
8.19633067e-01 -8.59750748e-01 -8.37603748e-01 9.13982451e-01
8.05834353e-01 -4.12642092e-01 -1.49921954e+00 2.00143993e-01
5.21426737e-01 -1.94917768e-01 7.34759986e-01 -1.12101614e+00
6.43754303e-01 -3.15516368e-02 -1.37875214e-01 -1.02288759e+00
-3.16431463e-01 -7.82626510e-01 7.04246163e-01 7.29636192e-01
9.55946386e-01 -9.37500000e-01 4.13619280e-01 9.61152077e-01
-1.59555495e-01 -1.06541133e+00 -6.50949597e-01 -3.31473112e-01
2.50273377e-01 -1.55813932e-01 3.60659987e-01 1.28844714e+00
6.09742522e-01 4.55161363e-01 -3.06005776e-01 -1.33540288e-01
-9.73183371e-04 5.24818264e-02 4.71339911e-01 -8.42159867e-01
-4.66558039e-01 -2.02474698e-01 -3.19699228e-01 -9.40625966e-01
-4.08416420e-01 -9.75542784e-01 -1.17569596e-01 -1.87234402e+00
6.46488607e-01 -3.64735395e-01 -2.11588219e-01 4.88498837e-01
-3.60374868e-01 2.30128497e-01 -1.83740050e-01 5.72162449e-01
-2.01731414e-01 1.76111698e-01 1.60804856e+00 -1.94319621e-01
-3.81250739e-01 -3.44668150e-01 -1.59095538e+00 4.61004078e-01
6.75359845e-01 -8.74831796e-01 -4.24249291e-01 -6.11232758e-01
5.46175420e-01 1.92304134e-01 1.92228667e-02 -3.48458290e-01
1.36149168e-01 -2.04719692e-01 7.95455649e-02 6.35299757e-02
-7.29473233e-02 -7.03061104e-01 3.86709899e-01 9.66103673e-01
-5.51907182e-01 4.30626005e-01 2.29048759e-01 3.17227155e-01
-2.09800154e-01 -1.88544139e-01 6.61924958e-01 -3.72177243e-01
1.88021645e-01 -2.47866940e-02 -5.34461498e-01 4.57926154e-01
8.07475448e-01 -2.12141261e-01 -6.37767240e-02 -5.21618724e-01
-1.05406260e+00 4.38947715e-02 6.49500132e-01 1.42742500e-01
7.79029548e-01 -1.13337278e+00 -1.03140008e+00 7.31371269e-02
2.64749885e-01 -1.65861487e-01 1.85690179e-01 1.48609281e+00
-9.39949751e-01 6.17370367e-01 2.10577995e-01 -5.86127341e-01
-1.07603228e+00 4.94816840e-01 1.14187859e-01 -4.92175162e-01
-6.00280881e-01 7.06940234e-01 1.32195070e-01 -2.18412757e-01
-1.28741711e-01 -4.69979167e-01 1.13638863e-03 -1.63648114e-01
4.55439687e-01 -2.20872685e-01 2.75038093e-01 -7.10492671e-01
-4.29624289e-01 7.77144670e-01 -6.66240752e-01 -1.72163755e-01
1.14049399e+00 -2.50131577e-01 -3.30199599e-01 2.54970014e-01
1.07643282e+00 2.99664825e-01 -1.04000457e-01 -2.59186655e-01
-1.05214536e-01 -4.04317230e-01 -1.19741499e-01 -1.02687371e+00
-5.99078178e-01 6.85579062e-01 5.74131429e-01 -1.97488040e-01
1.16504037e+00 -8.08752999e-02 1.06646335e+00 3.44888628e-01
2.53817260e-01 -1.01860619e+00 -1.57691494e-01 6.61829263e-02
6.43739104e-01 -1.33134747e+00 1.21598840e-01 -4.34161574e-01
-1.05804765e+00 8.35815430e-01 -6.42738640e-02 4.89507496e-01
6.24935031e-01 6.00333065e-02 7.18176663e-01 -2.24685073e-01
-6.20798349e-01 1.85128361e-01 1.31775141e-01 4.19236630e-01
8.27104688e-01 3.50597024e-01 -9.31111693e-01 6.61512792e-01
-2.17837781e-01 1.93524539e-01 5.30824542e-01 1.07684135e+00
-1.51048405e-02 -1.45001090e+00 -2.47776344e-01 5.98931849e-01
-8.55910778e-01 -7.44619191e-01 -3.99572641e-01 3.64251316e-01
-8.93811285e-02 1.04967737e+00 -3.96664441e-01 -2.45924816e-01
4.53677982e-01 1.30337849e-01 3.12211692e-01 -1.07461941e+00
-6.99693084e-01 4.50355440e-01 3.09713989e-01 -1.89754277e-01
-5.07842541e-01 -6.18561983e-01 -1.13650584e+00 -2.53053665e-01
-5.41027188e-01 5.71173549e-01 4.71580863e-01 9.24679339e-01
7.45495975e-01 5.34050107e-01 3.25184137e-01 4.49561238e-01
-6.90520287e-01 -7.20125794e-01 -1.83370486e-01 4.80555534e-01
1.78934693e-01 2.48588361e-02 -5.13044484e-02 3.56005609e-01] | [8.606769561767578, 8.49281120300293] |
fa814a33-e018-4871-8bdf-59ab2729226a | dream-a-challenge-dataset-and-models-for | 1902.00164 | null | http://arxiv.org/abs/1902.00164v1 | http://arxiv.org/pdf/1902.00164v1.pdf | DREAM: A Challenge Dataset and Models for Dialogue-Based Reading Comprehension | We present DREAM, the first dialogue-based multiple-choice reading
comprehension dataset. Collected from English-as-a-foreign-language
examinations designed by human experts to evaluate the comprehension level of
Chinese learners of English, our dataset contains 10,197 multiple-choice
questions for 6,444 dialogues. In contrast to existing reading comprehension
datasets, DREAM is the first to focus on in-depth multi-turn multi-party
dialogue understanding. DREAM is likely to present significant challenges for
existing reading comprehension systems: 84% of answers are non-extractive, 85%
of questions require reasoning beyond a single sentence, and 34% of questions
also involve commonsense knowledge.
We apply several popular neural reading comprehension models that primarily
exploit surface information within the text and find them to, at best, just
barely outperform a rule-based approach. We next investigate the effects of
incorporating dialogue structure and different kinds of general world knowledge
into both rule-based and (neural and non-neural) machine learning-based reading
comprehension models. Experimental results on the DREAM dataset show the
effectiveness of dialogue structure and general world knowledge. DREAM will be
available at https://dataset.org/dream/. | ['Jianshu Chen', 'Yejin Choi', 'Kai Sun', 'Dian Yu', 'Dong Yu', 'Claire Cardie'] | 2019-02-01 | null | null | null | null | ['dialogue-understanding'] | ['natural-language-processing'] | [ 1.75158530e-01 8.12444508e-01 1.08106337e-01 -4.41724807e-01
-1.06001294e+00 -8.81226122e-01 7.41331220e-01 5.23425460e-01
-5.04671454e-01 7.73525596e-01 8.10599685e-01 -9.13336337e-01
-1.75864473e-01 -1.03240812e+00 -5.81521988e-01 1.79579347e-01
4.58219290e-01 6.13090038e-01 3.22343320e-01 -1.05092096e+00
5.66335201e-01 -4.26406920e-01 -1.14012206e+00 6.31598830e-01
1.62069035e+00 6.84645951e-01 3.72533530e-01 9.39891160e-01
-3.70749950e-01 1.47781396e+00 -7.37214267e-01 -6.38582230e-01
-2.53037214e-01 -7.39547193e-01 -1.83283472e+00 -3.25457662e-01
5.47418654e-01 -7.57929206e-01 -2.93591022e-01 8.36038291e-01
3.72534424e-01 2.89481193e-01 4.99783456e-01 -5.30257583e-01
-1.02946985e+00 9.54887450e-01 3.23833674e-02 5.60941696e-01
1.23354459e+00 3.06573063e-01 1.13997257e+00 -6.86875284e-01
5.31127095e-01 1.34357917e+00 2.60829479e-01 8.08984399e-01
-9.57823575e-01 -2.07638547e-01 1.64178640e-01 5.95945954e-01
-6.80568218e-01 -6.04514360e-01 6.31213605e-01 -1.39422685e-01
1.28507161e+00 2.58320242e-01 6.37316942e-01 1.31585276e+00
2.48678967e-01 1.04047239e+00 1.59521973e+00 -6.62344158e-01
5.00346497e-02 -2.18834043e-01 8.41356814e-01 8.80760908e-01
-1.55803069e-01 -3.57579589e-01 -8.50464523e-01 1.45090684e-01
2.47868195e-01 -6.85660660e-01 -7.73013771e-01 3.72934490e-01
-1.23470783e+00 1.00067830e+00 1.55664355e-01 1.11872934e-01
-1.52867422e-01 -4.87112135e-01 3.14509392e-01 9.02662456e-01
3.04792553e-01 1.12421679e+00 -7.99141884e-01 -7.04758346e-01
-4.52189565e-01 2.34445304e-01 1.61420631e+00 8.00478458e-01
4.67958122e-01 -4.72732037e-01 -3.80781651e-01 9.78175581e-01
2.90276080e-01 4.75613952e-01 5.62779546e-01 -1.20631361e+00
8.23394895e-01 7.40650952e-01 -2.14778706e-01 -7.99817979e-01
-5.53981245e-01 -1.74333453e-01 -4.12594706e-01 -4.38186914e-01
9.38626289e-01 -3.92054588e-01 -3.46597761e-01 1.82033384e+00
-3.21942382e-02 -5.21723628e-01 3.93233091e-01 5.99416435e-01
1.57908690e+00 7.08430946e-01 -7.79396221e-02 -1.57535225e-01
1.69327843e+00 -1.21479177e+00 -8.77710938e-01 -4.51736957e-01
7.22356737e-01 -6.27527952e-01 1.57285833e+00 5.46845555e-01
-1.44437182e+00 -3.19408894e-01 -8.97828460e-01 -6.14754558e-01
-3.31827432e-01 -4.03136730e-01 2.69571215e-01 5.81217170e-01
-1.04847336e+00 4.70833592e-02 -2.25499317e-01 -1.90804228e-01
3.09613407e-01 -2.01966718e-01 -4.97717261e-02 -4.29120272e-01
-1.74522519e+00 1.42114556e+00 3.77818048e-01 -1.12209126e-01
-9.31063116e-01 -7.13655472e-01 -1.01688826e+00 1.46595582e-01
8.64872277e-01 -7.33858228e-01 1.88152099e+00 -6.38881266e-01
-1.92456436e+00 9.93139446e-01 -2.96996117e-01 -2.87420362e-01
2.02842757e-01 -5.88339686e-01 -9.02091041e-02 3.77972901e-01
1.94651097e-01 4.80569392e-01 2.09504783e-01 -7.35400200e-01
-2.46821687e-01 -2.50656039e-01 6.93213701e-01 6.52838826e-01
-7.79618137e-03 -2.10517794e-01 -1.57484904e-01 -1.86251074e-01
-7.64933228e-02 -5.56269169e-01 2.60809511e-01 -4.23400491e-01
-5.71122766e-01 -8.73891234e-01 2.34137580e-01 -1.12203693e+00
1.14515555e+00 -1.53588378e+00 1.82478607e-01 -3.76175106e-01
5.95047355e-01 1.42342225e-01 -2.80407935e-01 6.92430615e-01
3.06562483e-01 2.21827850e-01 -1.69732928e-01 7.08348304e-02
1.69404641e-01 5.85145727e-02 -3.58887434e-01 -4.67593260e-02
2.12632373e-01 1.34613621e+00 -1.10166764e+00 -3.48795533e-01
9.83571336e-02 -2.76979715e-01 -5.19847095e-01 5.85426509e-01
-7.59745598e-01 4.04652357e-01 -4.47074562e-01 3.84216964e-01
3.74438643e-01 -4.33886051e-01 -4.17384282e-02 3.86544496e-01
6.88845292e-02 1.27778828e+00 -3.26485664e-01 2.03753376e+00
-6.22757852e-01 8.90977740e-01 7.95622468e-02 -6.97546065e-01
5.16893864e-01 2.03951120e-01 -4.77890670e-01 -1.12605345e+00
2.37123147e-01 -8.55226815e-03 5.31499207e-01 -9.07767773e-01
5.74753582e-01 -7.88979977e-02 -2.02903509e-01 8.01512897e-01
1.74107045e-01 -4.40652221e-01 3.39846969e-01 6.81070745e-01
1.30057538e+00 -2.33601734e-01 4.86104369e-01 -5.29042304e-01
6.36448205e-01 2.37355798e-01 3.34035382e-02 1.11044443e+00
-1.60098583e-01 2.55864441e-01 8.62705112e-01 1.45629108e-01
-3.47017884e-01 -1.09774649e+00 8.24359953e-02 1.53893185e+00
1.17249109e-01 -4.48596090e-01 -1.01036453e+00 -5.15824318e-01
-6.25932992e-01 1.44287908e+00 -4.24056023e-01 -1.50173604e-01
-5.48435807e-01 -3.91993999e-01 8.01025510e-01 2.74268657e-01
1.10692489e+00 -1.12265241e+00 -7.89300323e-01 2.82779694e-01
-7.40757167e-01 -1.12280202e+00 -1.46517292e-01 4.78394181e-02
-6.33937061e-01 -1.32410634e+00 -3.21030706e-01 -8.06266487e-01
1.84904784e-01 1.05478684e-04 1.75437021e+00 5.33858538e-01
2.03208447e-01 8.42080235e-01 -7.84609616e-01 -5.44998288e-01
-6.87362015e-01 4.50737655e-01 -5.69060504e-01 -7.41248369e-01
6.34512663e-01 -1.41295418e-01 -3.74481440e-01 -3.64890508e-02
-5.46387672e-01 4.19916928e-01 5.80793262e-01 1.01128209e+00
-2.33309895e-01 -4.64725494e-01 8.25839877e-01 -9.64539111e-01
1.69892931e+00 -7.58867145e-01 -6.69139326e-02 6.41527772e-01
-3.68253231e-01 -5.82718030e-02 5.67696929e-01 -1.32654592e-01
-1.54455566e+00 -7.38852441e-01 -6.27997339e-01 6.43895745e-01
-4.53290612e-01 9.58141565e-01 -8.99560750e-02 4.41273332e-01
8.83863032e-01 4.64695364e-01 2.07635492e-01 -2.53192842e-01
5.44028282e-01 5.57548583e-01 6.48577929e-01 -1.06282079e+00
3.46963704e-01 -2.97130883e-01 -6.76128447e-01 -1.12637269e+00
-1.34134424e+00 -3.47377807e-01 -5.60897291e-01 -2.19506413e-01
1.08734429e+00 -8.68082106e-01 -8.76188099e-01 7.13151455e-01
-1.40300620e+00 -7.76872158e-01 -1.32522225e-01 7.79589638e-02
-5.85016668e-01 3.67607474e-01 -9.83102322e-01 -5.27579725e-01
-3.82970423e-01 -1.04697990e+00 5.22881866e-01 4.57025081e-01
-5.88956058e-01 -1.24047267e+00 -2.18481757e-02 1.02526104e+00
5.62504709e-01 -2.72041321e-01 1.51089251e+00 -1.22404611e+00
-3.73041868e-01 3.68661016e-01 -7.97349140e-02 2.50166208e-01
-8.25467259e-02 -5.56277275e-01 -8.89527380e-01 1.37805745e-01
5.82071841e-01 -1.28955948e+00 8.12352717e-01 3.85181829e-02
1.02231288e+00 -4.34915602e-01 2.73034692e-01 -1.65387556e-01
9.09917951e-01 2.16956548e-02 7.01909125e-01 2.29387298e-01
2.85646170e-01 1.00452006e+00 3.55707943e-01 4.54792120e-02
1.25278080e+00 1.10144548e-01 2.12927639e-01 3.58050883e-01
-1.08382829e-01 -4.43534166e-01 4.08809185e-01 1.47722137e+00
2.08398327e-01 -4.86597985e-01 -1.41957068e+00 6.65074050e-01
-1.45308959e+00 -9.49369609e-01 -3.80362689e-01 1.54523396e+00
1.51298583e+00 6.33516610e-02 -2.52334982e-01 -1.49016246e-01
2.61628658e-01 2.79518247e-01 -6.44611955e-01 -7.31214464e-01
-1.94983542e-01 4.69857365e-01 -2.32837260e-01 9.13973749e-01
-5.18204749e-01 1.08781075e+00 6.16066599e+00 7.09479451e-01
-4.45956647e-01 1.90034568e-01 5.54414630e-01 2.83072978e-01
-6.11458302e-01 -5.02869412e-02 -4.96607929e-01 3.20261717e-02
1.05504966e+00 -2.06132963e-01 4.67579037e-01 2.91444808e-01
-1.92187250e-01 -7.51609445e-01 -1.22632861e+00 5.19895792e-01
3.37664574e-01 -1.19663203e+00 8.98046866e-02 -4.86039996e-01
4.48189884e-01 -2.74217911e-02 -1.90475047e-01 7.94427574e-01
4.33141023e-01 -1.41570461e+00 4.29470032e-01 6.10126197e-01
3.22005063e-01 -4.56721336e-01 6.57223761e-01 1.02233946e+00
-4.36836809e-01 7.37505928e-02 -1.64327249e-01 -7.74675846e-01
1.54929250e-01 8.16783756e-02 -7.00733244e-01 2.42304474e-01
5.48567355e-01 4.14423943e-01 -9.48490739e-01 4.75303352e-01
-8.45698893e-01 1.01947093e+00 -1.85789376e-01 -7.64063478e-01
1.66839123e-01 5.89863583e-03 4.80977535e-01 1.01860058e+00
-3.07389975e-01 8.37951481e-01 -1.22612324e-02 8.99176836e-01
-3.79192650e-01 1.48007527e-01 -4.43799824e-01 3.16556506e-02
5.66335976e-01 7.45695412e-01 -3.25081915e-01 -2.90970713e-01
-7.38463044e-01 7.76868761e-01 6.43106401e-01 2.42044151e-01
-3.17622274e-01 -3.92727882e-01 2.25423932e-01 -1.98083997e-01
-1.84966207e-01 -2.91015804e-01 -3.62519681e-01 -1.39950693e+00
-7.53336996e-02 -1.57442725e+00 2.55597293e-01 -8.86544228e-01
-1.45115089e+00 3.46468180e-01 9.46023241e-02 -2.86653310e-01
-4.21536297e-01 -8.35227311e-01 -8.47702146e-01 8.76710415e-01
-1.49805057e+00 -6.58442914e-01 -3.50780249e-01 7.00246215e-01
1.25984132e+00 -7.09608346e-02 1.01569402e+00 -4.18062240e-01
-3.56606781e-01 4.01341170e-01 -2.76648521e-01 4.19548362e-01
6.44459903e-01 -1.54480982e+00 4.00133878e-01 5.29330015e-01
-1.15696132e-01 8.76547754e-01 5.87418318e-01 -5.23305535e-01
-1.60640264e+00 -3.26404601e-01 9.88812864e-01 -9.38848019e-01
7.43395805e-01 -2.59233296e-01 -1.27249324e+00 7.69609034e-01
1.31003630e+00 -1.04389226e+00 9.85261321e-01 5.47855914e-01
-3.13025415e-01 5.47976553e-01 -8.15035462e-01 7.94522107e-01
9.48669136e-01 -8.22114050e-01 -1.84315813e+00 3.73825729e-01
9.90499914e-01 -7.49100864e-01 -1.01095474e+00 1.60744563e-01
2.09419534e-01 -1.00321019e+00 4.99474972e-01 -6.80288374e-01
9.96468425e-01 3.00912738e-01 -1.56942476e-02 -1.64695561e+00
6.79095015e-02 -5.90515971e-01 -1.04025319e-01 9.31401670e-01
8.52584243e-01 -6.67654216e-01 2.23688781e-01 9.79316413e-01
-3.27718854e-01 -9.47575688e-01 -9.91758049e-01 -1.46696448e-01
9.97298181e-01 -4.22659069e-01 2.13177592e-01 9.67662930e-01
5.33758938e-01 1.12807918e+00 1.66685075e-01 -2.92432189e-01
1.75639108e-01 -5.19900620e-02 6.55467451e-01 -1.10434449e+00
-2.52747685e-01 -6.27685547e-01 4.96211439e-01 -1.80605543e+00
4.90502983e-01 -8.88790607e-01 1.96949735e-01 -1.88518393e+00
1.24673843e-01 1.65843606e-01 4.54098493e-01 3.97423625e-01
-4.66488749e-01 -6.20454490e-01 1.64623424e-01 -2.03556493e-01
-9.45588529e-01 7.93629110e-01 1.72588456e+00 -1.37470752e-01
1.12021873e-02 -1.66196376e-01 -1.16546297e+00 7.66408026e-01
9.48940873e-01 1.19821899e-01 -6.89427793e-01 -8.29378009e-01
5.36094427e-01 4.58317399e-01 1.57273978e-01 -5.13129056e-01
6.56238735e-01 -8.75035375e-02 1.37953430e-01 -4.15978581e-01
2.11114943e-01 -5.31596094e-02 -9.23120558e-01 3.68744075e-01
-9.33924258e-01 2.89925575e-01 3.31569761e-01 2.44614765e-01
-2.71491498e-01 -5.32640517e-01 4.27538306e-01 -5.52938282e-01
-8.20755839e-01 -5.00435472e-01 -1.02425468e+00 1.14981031e+00
6.05449498e-01 1.35111287e-01 -1.15803874e+00 -9.96476591e-01
-4.97972816e-01 7.97668755e-01 6.79339916e-02 5.85356653e-01
8.46201777e-01 -4.80751544e-01 -1.11934400e+00 -2.81791329e-01
6.73258677e-02 3.77688795e-01 1.89156055e-01 5.88338196e-01
-5.88591456e-01 7.17556596e-01 -1.21928066e-01 -2.96825647e-01
-9.82233882e-01 4.48915176e-02 4.08966064e-01 -3.77075016e-01
-4.46881026e-01 1.01143968e+00 1.75694562e-02 -9.10758376e-01
8.34257305e-02 -3.35514486e-01 -4.93950009e-01 1.16469152e-01
6.97713077e-01 4.15619999e-01 -1.00112684e-01 -1.74769327e-01
5.28075360e-02 2.46005431e-01 -2.50729412e-01 -3.44664067e-01
9.13723230e-01 -4.10011798e-01 -3.42076719e-01 8.16340625e-01
7.72729158e-01 3.90915712e-03 -6.36950314e-01 -3.65648538e-01
5.64625040e-02 1.53834805e-01 -2.13217020e-01 -1.53362870e+00
-2.99335927e-01 1.08843112e+00 -1.45983681e-01 3.43055010e-01
1.03970742e+00 2.54989535e-01 8.65216553e-01 1.09747016e+00
2.41710633e-01 -1.12830043e+00 4.06243593e-01 1.30520260e+00
1.23790717e+00 -1.38347900e+00 -1.85592219e-01 -3.78687382e-01
-6.97034359e-01 1.15198040e+00 1.32401466e+00 1.82071254e-01
4.05486017e-01 -1.42512888e-01 1.99198171e-01 -4.16567534e-01
-1.30410922e+00 -1.60402387e-01 3.54343802e-01 3.69324654e-01
6.79543197e-01 -1.36253417e-01 -4.87729788e-01 8.16961467e-01
-8.62204194e-01 -4.48342264e-01 9.08816278e-01 9.02052939e-01
-5.91758907e-01 -8.13359380e-01 -2.13535577e-01 7.94609427e-01
-2.90049791e-01 -6.07330322e-01 -8.43138278e-01 7.19342709e-01
-5.15926421e-01 1.66254377e+00 -1.08639590e-01 -8.29510391e-02
4.16488379e-01 3.78751963e-01 5.96819818e-01 -8.86821449e-01
-9.35526490e-01 -6.99996769e-01 5.89463472e-01 -4.35752779e-01
-2.84261376e-01 -5.23902357e-01 -1.21733975e+00 -4.46050256e-01
-3.66647273e-01 2.45527297e-01 1.51337922e-01 1.41133261e+00
6.74579740e-02 4.36162025e-01 -4.11534682e-02 1.12433508e-01
-8.53692174e-01 -1.42734778e+00 3.42917070e-02 1.73472136e-01
4.69791472e-01 -1.54979527e-01 -3.89075071e-01 -2.86407501e-01] | [11.313668251037598, 8.10371208190918] |
d4d354b2-9f67-459f-b3e8-1ebb37d70418 | plas-latent-action-space-for-offline | 2011.07213 | null | https://arxiv.org/abs/2011.07213v1 | https://arxiv.org/pdf/2011.07213v1.pdf | PLAS: Latent Action Space for Offline Reinforcement Learning | The goal of offline reinforcement learning is to learn a policy from a fixed dataset, without further interactions with the environment. This setting will be an increasingly more important paradigm for real-world applications of reinforcement learning such as robotics, in which data collection is slow and potentially dangerous. Existing off-policy algorithms have limited performance on static datasets due to extrapolation errors from out-of-distribution actions. This leads to the challenge of constraining the policy to select actions within the support of the dataset during training. We propose to simply learn the Policy in the Latent Action Space (PLAS) such that this requirement is naturally satisfied. We evaluate our method on continuous control benchmarks in simulation and a deformable object manipulation task with a physical robot. We demonstrate that our method provides competitive performance consistently across various continuous control tasks and different types of datasets, outperforming existing offline reinforcement learning methods with explicit constraints. Videos and code are available at https://sites.google.com/view/latent-policy. | ['David Held', 'Sujay Bajracharya', 'Wenxuan Zhou'] | 2020-11-14 | null | null | null | null | ['deformable-object-manipulation'] | ['robots'] | [ 1.32654145e-01 6.55482290e-03 -7.06061423e-01 -1.17115460e-01
-6.72097683e-01 -7.38950193e-01 6.70364738e-01 7.65579846e-03
-6.30186439e-01 1.02345002e+00 -3.22948545e-02 -3.43914241e-01
-3.14690083e-01 -5.49421966e-01 -9.50240374e-01 -6.92804337e-01
-3.65821719e-01 8.14967453e-01 1.83800533e-01 -7.95387775e-02
3.78698498e-01 4.41868395e-01 -1.37140322e+00 -1.09930031e-01
7.49394655e-01 8.16029966e-01 3.75039607e-01 5.84684551e-01
3.86921227e-01 6.70003176e-01 -3.60326707e-01 5.39692581e-01
6.36979580e-01 -1.60268828e-01 -6.25748515e-01 2.66149580e-01
2.09446877e-01 -7.36994505e-01 -4.75885063e-01 1.08383381e+00
4.03673381e-01 5.97455025e-01 4.24303681e-01 -1.33677018e+00
-3.17589909e-01 4.27179873e-01 -3.65019798e-01 -1.05096470e-03
2.84695774e-01 5.82406819e-01 8.24792743e-01 -3.88201952e-01
6.14294469e-01 1.32799554e+00 1.20822668e-01 7.24601150e-01
-1.29221535e+00 -6.22591853e-01 5.61711788e-01 1.64711960e-02
-8.43751073e-01 -3.91008258e-01 5.02864301e-01 -3.91811579e-01
8.14841211e-01 -1.47083789e-01 6.37878120e-01 1.31699300e+00
3.01647604e-01 8.83098125e-01 1.25709593e+00 -1.50400087e-01
7.46804535e-01 -4.26912636e-01 -5.11452138e-01 5.87996662e-01
-4.18748073e-02 5.40844381e-01 -5.09375632e-01 -3.55297238e-01
1.12221599e+00 1.34541452e-01 -1.63900480e-01 -1.03371811e+00
-1.24732244e+00 9.27505016e-01 1.10677540e-01 -2.53473759e-01
-5.45130670e-01 6.24936998e-01 3.89209300e-01 4.79382664e-01
9.79925692e-02 6.06456459e-01 -6.89810216e-01 -6.75141692e-01
-3.04069847e-01 7.86340594e-01 7.35715449e-01 1.01995778e+00
4.16193336e-01 1.64867267e-01 -1.30607679e-01 8.35787058e-01
2.08321348e-01 5.22225857e-01 4.90057588e-01 -1.48597741e+00
4.51611936e-01 2.00757056e-01 6.85737967e-01 -4.00246471e-01
-2.04346776e-01 -5.05738780e-02 -2.18727857e-01 8.10341239e-01
5.33937156e-01 -3.52860063e-01 -1.13176680e+00 1.80737054e+00
6.81685925e-01 2.51510441e-01 -5.21389320e-02 1.02024448e+00
-2.47476280e-01 4.99972403e-01 1.68729145e-02 -1.94976732e-01
7.11042643e-01 -8.46543908e-01 -5.97665727e-01 -3.83731395e-01
3.00348938e-01 -4.40345883e-01 1.29825246e+00 5.74448884e-01
-9.22587097e-01 -2.68061906e-01 -8.79146278e-01 2.54922479e-01
2.43879389e-03 -8.47426876e-02 7.31213927e-01 -5.32345586e-02
-6.42565608e-01 1.01518965e+00 -1.44714940e+00 -2.05537945e-01
5.09368062e-01 6.70629144e-01 -2.14128509e-01 9.79433134e-02
-8.16862583e-01 8.91902208e-01 3.99497420e-01 -9.29589942e-02
-1.63728499e+00 -5.99503815e-01 -6.74178302e-01 -2.31761977e-01
1.01555836e+00 -2.92675167e-01 1.91315842e+00 -8.25810313e-01
-1.96903646e+00 2.06435934e-01 4.03475463e-01 -5.75354159e-01
8.27834845e-01 -7.03939438e-01 1.63855582e-01 -4.82443906e-03
6.38326332e-02 6.38707757e-01 1.12629271e+00 -1.15475357e+00
-7.27950037e-01 -2.27454677e-01 3.37603629e-01 4.22726244e-01
-8.47411230e-02 -3.41159642e-01 -1.55841902e-01 -5.19638002e-01
-2.28468999e-01 -1.33889997e+00 -5.04932225e-01 2.49757156e-01
-3.21494117e-02 -4.65645492e-01 1.14139664e+00 -2.80786365e-01
7.35855639e-01 -2.01648140e+00 4.01598603e-01 1.46827683e-01
-1.01494715e-01 1.26338631e-01 -1.38340563e-01 5.55189967e-01
2.21173421e-01 -2.65908748e-01 -1.93033069e-01 -3.43425362e-03
2.41663605e-01 5.79974890e-01 -6.72152281e-01 7.12580383e-01
-4.02032956e-02 6.09730959e-01 -1.23693562e+00 -1.87346652e-01
3.52225363e-01 -1.27582863e-01 -8.10837805e-01 4.35229748e-01
-8.66801500e-01 8.65198493e-01 -8.89073730e-01 4.33254749e-01
1.36898205e-01 -8.66656303e-02 4.74228561e-01 5.45783997e-01
-1.08612753e-01 2.10006595e-01 -1.18922329e+00 1.86421561e+00
-4.21723396e-01 1.87362790e-01 2.49438345e-01 -1.08211839e+00
7.34167814e-01 1.83199823e-01 9.30912435e-01 -6.54650867e-01
5.57195880e-02 8.44518542e-02 2.61400849e-01 -5.68423569e-01
2.81915396e-01 -9.25809741e-02 -1.47953227e-01 4.75942552e-01
5.54094613e-02 -3.49634051e-01 1.55091211e-01 -2.57060062e-02
1.42338383e+00 5.82883537e-01 2.28574097e-01 -3.03452998e-01
-7.36693889e-02 1.15563639e-01 7.15241373e-01 8.82761717e-01
-3.54902685e-01 -2.99041402e-02 5.94290257e-01 -3.39255184e-01
-1.07711697e+00 -1.13467145e+00 8.72170180e-02 1.04887938e+00
1.87027901e-01 -2.00607345e-01 -4.35262442e-01 -7.69594491e-01
5.07975876e-01 5.76415718e-01 -5.77297509e-01 -1.81744218e-01
-7.56756485e-01 -1.43144235e-01 1.00459671e-02 5.48815310e-01
1.62452579e-01 -1.34314382e+00 -1.05072129e+00 3.67789805e-01
2.40471080e-01 -1.01651073e+00 -3.84378552e-01 3.11265886e-01
-9.34581816e-01 -1.20052767e+00 -3.97052556e-01 -5.40456772e-01
6.80433512e-01 -4.29795049e-02 7.34730303e-01 -4.44696359e-02
-2.83803731e-01 7.86228597e-01 -2.00716138e-01 -3.40299278e-01
-4.13630694e-01 -8.33806619e-02 4.14650649e-01 -4.36133087e-01
-6.31251633e-02 -5.34473002e-01 -6.95230305e-01 3.31246763e-01
-7.53352344e-01 -5.65015115e-02 3.06276023e-01 1.05808294e+00
6.62101865e-01 1.54960863e-02 6.96216702e-01 -6.36549115e-01
6.29323244e-01 -5.25060236e-01 -1.21395588e+00 -9.77964327e-02
-3.95024329e-01 1.53583705e-01 8.10542405e-01 -8.68414342e-01
-8.06528986e-01 3.44361037e-01 2.48013660e-01 -6.98985815e-01
-9.98844802e-02 2.48271391e-01 1.72150619e-02 1.26478940e-01
5.38585961e-01 -2.47762799e-02 3.33574176e-01 -3.60611647e-01
1.37510538e-01 3.69253188e-01 2.43005112e-01 -1.24393177e+00
5.96709907e-01 3.93277079e-01 2.22739689e-02 -4.99964654e-01
-6.87163830e-01 -3.07033896e-01 -4.08996195e-01 -3.83714616e-01
4.59549755e-01 -7.05932081e-01 -9.08292353e-01 2.67592967e-01
-5.58598936e-01 -1.37284243e+00 -4.21174258e-01 6.47188604e-01
-1.29179788e+00 9.70642790e-02 -4.47994322e-01 -8.23051572e-01
9.83982459e-02 -1.28342485e+00 9.97361362e-01 2.16056556e-01
-4.07209247e-02 -8.17158401e-01 2.20261306e-01 1.98569037e-02
2.67351210e-01 3.10623020e-01 6.83614850e-01 -3.44005793e-01
-6.52567327e-01 1.06022626e-01 4.11096781e-01 3.45359296e-01
4.24430668e-01 -1.49269760e-01 -5.39665639e-01 -8.21982503e-01
-2.59660929e-01 -9.54101741e-01 6.30962610e-01 3.51981759e-01
1.55693138e+00 -4.40461516e-01 -3.20710599e-01 2.70353228e-01
1.24038923e+00 4.65623707e-01 2.33633190e-01 3.38131160e-01
3.43754828e-01 2.88321614e-01 1.26371217e+00 8.37630451e-01
-2.85085980e-02 8.47575903e-01 6.17865622e-01 4.01426584e-01
3.22370529e-01 -5.25148213e-01 6.70243859e-01 1.64960474e-01
1.82921112e-01 -8.74501392e-02 -9.28739846e-01 4.80328292e-01
-2.31253552e+00 -8.57013822e-01 6.15492523e-01 2.39305711e+00
1.01681459e+00 2.84978062e-01 2.57319659e-01 -2.53117919e-01
3.82657379e-01 -4.62714806e-02 -1.27798927e+00 -2.83559144e-01
6.48351312e-01 2.64652014e-01 6.76640630e-01 5.84407032e-01
-1.25901675e+00 9.27986979e-01 5.88307762e+00 6.77898467e-01
-1.32681561e+00 -4.58425581e-02 3.23755503e-01 -3.77909213e-01
-9.99633968e-03 4.89189476e-02 -6.75873578e-01 4.85402554e-01
8.08011591e-01 -5.02822064e-02 8.93105984e-01 9.98562455e-01
5.22511065e-01 -3.36078614e-01 -1.23300028e+00 5.72562695e-01
-5.23699760e-01 -1.05457866e+00 -4.49176431e-01 2.19497755e-01
8.14497590e-01 2.81853050e-01 1.53845280e-01 6.59371853e-01
8.51944745e-01 -8.88336420e-01 6.69540644e-01 2.54488468e-01
7.00376093e-01 -6.32419348e-01 1.07929166e-02 6.64506555e-01
-7.66131997e-01 -5.22359431e-01 -3.39041084e-01 -1.80926263e-01
9.37778354e-02 6.38938025e-02 -9.25258636e-01 -5.43163195e-02
4.79208052e-01 6.75671458e-01 1.09629072e-01 9.70359683e-01
-3.13061565e-01 6.04854703e-01 -4.89656419e-01 -1.35671541e-01
3.57419729e-01 -1.07707076e-01 5.39276600e-01 5.96484959e-01
1.27707735e-01 4.76293862e-02 9.87255156e-01 5.92198849e-01
-7.39403814e-03 -2.59998024e-01 -9.13426459e-01 -2.24920258e-01
5.50338745e-01 9.67487752e-01 -6.94280088e-01 -1.54727966e-01
-2.27876976e-01 6.76130056e-01 5.73281288e-01 3.51170361e-01
-9.27457273e-01 2.74556316e-02 8.89708221e-01 -6.89512566e-02
5.02132595e-01 -8.50536644e-01 2.41751596e-01 -1.01551056e+00
3.23093869e-02 -1.09229338e+00 3.05296332e-01 -2.62083471e-01
-1.08623993e+00 -1.76218271e-01 3.34482878e-01 -1.37599492e+00
-4.26915526e-01 -7.16909766e-01 -2.98345625e-01 3.02399099e-01
-1.35438633e+00 -5.89160264e-01 3.19758281e-02 6.00374520e-01
8.33698690e-01 -8.62742662e-02 7.82519639e-01 1.61516778e-02
-3.31259191e-01 1.85880542e-01 4.01588887e-01 -7.38344043e-02
7.58839369e-01 -1.40317857e+00 1.14318155e-01 3.83555502e-01
-6.80611283e-02 4.34333891e-01 7.97356248e-01 -7.87099123e-01
-1.72911155e+00 -9.81211305e-01 -3.19316655e-01 -2.70220697e-01
9.74665523e-01 -4.52781409e-01 -7.79343903e-01 7.91556954e-01
8.97246525e-02 3.90020221e-01 1.76654011e-01 6.90088868e-02
1.78791247e-02 9.68529806e-02 -1.02233493e+00 7.04855442e-01
1.01720691e+00 -2.34725729e-01 -2.73141176e-01 5.94665051e-01
6.41372442e-01 -9.12355840e-01 -9.41435516e-01 4.42402869e-01
5.56912780e-01 -3.23495150e-01 7.12991238e-01 -1.02725923e+00
4.26146418e-01 -2.71882594e-01 -5.72491027e-02 -1.55776668e+00
-7.15553761e-02 -9.76554215e-01 -5.89847744e-01 5.38890064e-01
1.14501148e-01 -6.01828873e-01 8.48070383e-01 4.68967676e-01
-1.09435387e-01 -1.08356392e+00 -1.00100935e+00 -1.09459305e+00
1.79839820e-01 -1.26026064e-01 2.40250230e-01 7.85257161e-01
1.32838443e-01 -4.43793787e-03 -4.62890387e-01 2.34934360e-01
6.06067955e-01 2.94589132e-01 9.33346868e-01 -7.13112056e-01
-6.36428356e-01 -2.30085596e-01 -9.32433084e-02 -1.33876920e+00
5.05874276e-01 -4.76091474e-01 5.63222229e-01 -1.33211386e+00
2.16441825e-02 -9.00135577e-01 -3.87584597e-01 7.38758504e-01
1.50376400e-02 -4.26666021e-01 2.75610685e-01 1.72256514e-01
-7.78366089e-01 1.03445709e+00 1.62703681e+00 2.65712896e-03
-2.66168684e-01 1.52590811e-01 -1.68066859e-01 7.26141393e-01
1.30613852e+00 -4.80989814e-01 -8.45338881e-01 -3.96548182e-01
-1.35308340e-01 3.86554241e-01 3.31411570e-01 -9.30508018e-01
2.81527936e-02 -9.81733561e-01 1.65112630e-01 -1.33074194e-01
4.40799803e-01 -8.66731584e-01 -3.04839104e-01 7.29110360e-01
-6.07374847e-01 1.58345893e-01 2.06787497e-01 9.55176592e-01
1.74745172e-01 -7.29747489e-02 8.51067483e-01 -1.08070143e-01
-7.28460252e-01 5.62058270e-01 -3.68144274e-01 3.67425501e-01
1.29564834e+00 2.64299482e-01 -3.14191043e-01 -3.70765179e-01
-6.91827297e-01 5.89839160e-01 6.07349277e-01 6.82663023e-01
6.05691135e-01 -1.18142545e+00 -3.16380143e-01 8.46513826e-03
1.07940650e-02 1.44438446e-01 -1.12197399e-01 6.45153284e-01
-2.25045800e-01 1.69219092e-01 -4.76046771e-01 -5.36189735e-01
-9.46666241e-01 5.79458356e-01 3.96320760e-01 -1.55319467e-01
-9.23037648e-01 4.41914022e-01 4.35146764e-02 -5.94435573e-01
4.29557353e-01 -4.65100467e-01 1.73994586e-01 -5.93323171e-01
2.44710892e-01 2.05419824e-01 -3.06289643e-01 2.38924269e-02
-7.24951625e-02 8.77926499e-03 -1.38763472e-01 -3.77302468e-01
1.42631710e+00 2.88796276e-01 3.97484720e-01 5.13716161e-01
7.12841630e-01 -1.35271177e-01 -2.15732598e+00 -1.89805493e-01
7.96594173e-02 -6.96351409e-01 -7.47550130e-02 -6.48255646e-01
-7.55052209e-01 4.44686770e-01 5.29783607e-01 2.19742768e-02
6.84480309e-01 -2.50915021e-01 5.74611664e-01 6.81831837e-01
8.85796487e-01 -1.64298236e+00 5.17802775e-01 6.81201577e-01
1.03706074e+00 -1.28469384e+00 2.28024855e-01 9.84911248e-02
-7.26271272e-01 9.50893164e-01 9.37470794e-01 -6.72650993e-01
6.41130090e-01 3.67001951e-01 -1.70773372e-01 4.35380712e-02
-1.15189826e+00 3.00150160e-02 -2.15406045e-01 5.05787075e-01
1.52185475e-02 2.70274848e-01 -7.74135739e-02 -5.70381060e-02
2.99722273e-02 1.19141906e-01 5.19394338e-01 1.67087853e+00
-6.11906052e-01 -1.36899376e+00 -1.39447451e-01 5.42662084e-01
-3.38397682e-01 5.28660238e-01 1.35611314e-02 8.65175307e-01
-2.66149640e-01 5.82897723e-01 -4.87150066e-02 5.39718904e-02
1.68425515e-01 -1.47896633e-01 8.31162512e-01 -8.67345631e-01
-1.01578139e-01 9.69872847e-02 5.89310266e-02 -9.95986700e-01
-2.67116308e-01 -1.03533792e+00 -1.58233011e+00 -6.04854226e-02
-4.49827947e-02 4.16379385e-02 5.79985738e-01 7.60595858e-01
3.47671121e-01 4.37228858e-01 6.61227882e-01 -9.19250786e-01
-1.46421731e+00 -7.77199447e-01 -4.17843759e-01 2.61474133e-01
6.92486465e-01 -1.26711452e+00 -9.79069099e-02 -5.39079309e-02] | [4.29074764251709, 1.5809839963912964] |
c0fbf2b6-611d-4624-9bba-a1ea47dffa74 | spatiotemporal-and-semantic-zero-inflated | 2304.01569 | null | https://arxiv.org/abs/2304.01569v1 | https://arxiv.org/pdf/2304.01569v1.pdf | Spatiotemporal and Semantic Zero-inflated Urban Anomaly Prediction | Urban anomaly predictions, such as traffic accident prediction and crime prediction, are of vital importance to smart city security and maintenance. Existing methods typically use deep learning to capture the intra-dependencies in spatial and temporal dimensions. However, numerous key challenges remain unsolved, for instance, sparse zero-inflated data due to urban anomalies occurring with low frequency (which can lead to poor performance on real-world datasets), and both intra- and inter-dependencies of abnormal patterns across spatial, temporal, and semantic dimensions. Moreover, a unified approach to predict multiple kinds of anomaly is left to explore. In this paper, we propose STS to jointly capture the intra- and inter-dependencies between the patterns and the influential factors in three dimensions. Further, we use a multi-task prediction module with a customized loss function to solve the zero-inflated issue. To verify the effectiveness of the model, we apply it to two urban anomaly prediction tasks, crime prediction and traffic accident risk prediction, respectively. Experiments on two application scenarios with four real-world datasets demonstrate the superiority of STS, which outperforms state-of-the-art methods in the mean absolute error and the root mean square error by 37.88% and 18.10% on zero-inflated datasets, and, 60.32% and 37.28% on non-zero datasets, respectively. | ['Haiyong Xie', 'Yong Liao', 'Pengyuan Zhou', 'Yao Lu'] | 2023-04-04 | null | null | null | null | ['crime-prediction'] | ['miscellaneous'] | [-2.68200964e-01 -4.12842989e-01 4.80390191e-02 -3.03350031e-01
-5.91361344e-01 1.74134389e-01 4.65264291e-01 2.37766176e-01
-2.79759258e-01 5.43832719e-01 2.18241334e-01 -4.05371606e-01
-4.51789796e-01 -9.69464958e-01 -4.81472045e-01 -6.82059288e-01
-3.38507295e-01 3.06664407e-01 4.49004769e-01 -3.28217566e-01
1.11156508e-01 2.84903020e-01 -1.55964601e+00 1.28777355e-01
1.24917948e+00 1.20854974e+00 -4.02695924e-01 7.52159134e-02
-2.63932884e-01 6.25394583e-01 -4.28004742e-01 -4.08611059e-01
2.08070949e-01 5.76101691e-02 -3.41522545e-01 2.80610006e-02
-3.22970119e-03 -2.38071218e-01 -6.30646169e-01 8.51523757e-01
4.20277327e-01 1.88925430e-01 7.13087678e-01 -1.59069514e+00
-4.30851579e-01 1.34098008e-01 -9.68356192e-01 7.37586141e-01
-5.12608662e-02 3.59066457e-01 7.71692157e-01 -8.45594645e-01
-2.43870482e-01 1.12514079e+00 7.48960257e-01 8.30681920e-02
-9.24106538e-01 -9.81460989e-01 5.41926086e-01 5.88885725e-01
-1.48059881e+00 -3.34632695e-01 7.23932207e-01 -7.22313344e-01
8.47382784e-01 1.47849977e-01 4.45004135e-01 1.10786688e+00
3.50043982e-01 9.20587420e-01 5.68429649e-01 3.36682141e-01
-2.33230237e-02 -2.25375101e-01 9.17622894e-02 3.26244086e-01
3.53966206e-01 2.54976511e-01 -1.26047716e-01 1.07092392e-02
3.06437433e-01 5.03953397e-01 2.67982930e-01 5.22944033e-02
-9.39979613e-01 8.29469681e-01 3.01067650e-01 1.99500680e-01
-4.83714581e-01 -2.32592613e-01 8.25431824e-01 1.29248291e-01
8.75448167e-01 -1.31955698e-01 -5.20345867e-01 -2.70192653e-01
-4.81955320e-01 3.49528193e-01 6.34489581e-02 6.90467477e-01
5.62621951e-01 3.64516050e-01 -3.28991920e-01 8.97997558e-01
3.50698322e-01 6.47013485e-01 4.71224308e-01 -1.54759377e-01
1.03456211e+00 7.81829059e-01 -8.19133669e-02 -1.67838228e+00
-6.83931589e-01 -6.04320884e-01 -1.42192507e+00 -2.56306380e-01
3.38094324e-01 -1.88699722e-01 -7.02494562e-01 1.73416400e+00
4.57579851e-01 8.52647364e-01 -1.36478961e-01 6.59448802e-01
6.37830496e-01 5.80138505e-01 3.12137634e-01 -6.58079237e-02
9.60336387e-01 -5.47817886e-01 -6.98186278e-01 -4.18394685e-01
1.06340623e+00 -5.67650259e-01 1.26078689e+00 4.47527878e-02
-6.13542318e-01 -2.60625005e-01 -5.08628428e-01 3.00939977e-01
-4.51969862e-01 5.81222922e-02 4.96590227e-01 3.52554888e-01
-3.18497568e-01 1.15187749e-01 -7.58902013e-01 -8.37927759e-02
7.76514769e-01 9.20752659e-02 -4.16842848e-02 -1.08776979e-01
-1.44674790e+00 3.62756431e-01 4.57021058e-01 7.88352042e-02
-3.23132515e-01 -1.05413389e+00 -8.03054154e-01 7.37514645e-02
3.03524673e-01 -2.70505339e-01 6.31801605e-01 -4.41287875e-01
-7.79512286e-01 5.09493351e-01 -1.87162310e-01 -3.22728515e-01
4.27033663e-01 -2.86523819e-01 -1.17044151e+00 -5.53557515e-01
6.31618977e-01 -1.87657937e-01 4.08989847e-01 -7.16975152e-01
-1.01479888e+00 -6.66472137e-01 -2.77844727e-01 -2.29891956e-01
-5.89272022e-01 -1.50551766e-01 -2.61712968e-01 -9.74828362e-01
2.56797671e-01 -6.83207214e-01 -3.37171674e-01 -3.53202254e-01
-5.21590769e-01 -4.29252446e-01 1.05001366e+00 -8.13724399e-01
1.85029542e+00 -2.34304023e+00 -4.48839426e-01 3.16340655e-01
1.26354188e-01 5.34984648e-01 -5.71724325e-02 5.64278662e-02
-1.99958667e-01 -1.03729963e-01 -4.16251212e-01 -1.76792458e-01
-1.04398683e-01 2.50885934e-01 -4.77087140e-01 4.72240806e-01
2.82129526e-01 8.55831623e-01 -8.58037472e-01 -3.20357651e-01
3.69752675e-01 2.36330181e-01 -4.91044432e-01 -2.71011256e-02
1.85558632e-01 5.78582644e-01 -8.87432456e-01 6.25731885e-01
8.02229524e-01 7.06897164e-03 -6.39694631e-01 1.96149126e-01
-6.80164248e-02 2.17744038e-01 -1.29138267e+00 1.20504761e+00
-5.08255422e-01 5.28806746e-01 -4.70180362e-01 -1.38550317e+00
9.27297294e-01 2.59857059e-01 8.88894439e-01 -1.34367311e+00
-1.06457978e-01 2.53796637e-01 -1.19750604e-01 -8.38080525e-01
1.93261579e-01 1.98280159e-02 -3.23167115e-01 5.07963419e-01
-6.40841186e-01 6.93394780e-01 8.72745365e-02 -8.97590164e-03
1.13540471e+00 -6.49146020e-01 3.09351869e-02 -7.02831522e-02
8.43184114e-01 -4.19518858e-01 9.50853348e-01 3.95929366e-01
-3.98540914e-01 3.14739585e-01 5.61061859e-01 -1.10334861e+00
-9.30111647e-01 -9.78425622e-01 -3.34633142e-01 8.56961727e-01
1.47550598e-01 -3.19822967e-01 -2.79166073e-01 -8.16403925e-01
9.38695371e-02 9.21889544e-01 -5.60674429e-01 -5.22827387e-01
-6.48075402e-01 -1.33726358e+00 5.83985925e-01 5.82075298e-01
6.25060439e-01 -9.38181758e-01 -3.42562139e-01 1.77707613e-01
-4.93110448e-01 -1.28031564e+00 -1.19742565e-01 -5.34844697e-01
-5.64972579e-01 -1.10789549e+00 -2.26573631e-01 -1.36172742e-01
4.27393675e-01 4.09773827e-01 1.08247614e+00 1.89220369e-01
-1.65161163e-01 -5.94766028e-02 -2.47893393e-01 -4.86684203e-01
3.01952958e-01 2.52121389e-01 8.94469768e-02 5.12551606e-01
7.49960840e-01 -8.43874991e-01 -6.27693355e-01 5.52728593e-01
-9.23562884e-01 -2.24141508e-01 5.04572928e-01 6.99049115e-01
6.01614475e-01 4.18831199e-01 8.10082316e-01 -5.68954825e-01
5.39690673e-01 -1.28173053e+00 -3.87049258e-01 -1.18432745e-01
-5.95829427e-01 -1.08740382e-01 7.11814702e-01 -2.97256082e-01
-8.68503571e-01 -2.81481713e-01 -2.35633343e-01 -3.51128250e-01
-5.52136958e-01 5.02616823e-01 -3.31455648e-01 6.38795137e-01
3.47077161e-01 3.63709688e-01 -5.16953588e-01 -3.78589839e-01
-1.76226884e-01 5.30563772e-01 2.70568132e-01 -4.34326738e-01
1.00269639e+00 4.31278825e-01 2.04677328e-01 -7.72354662e-01
-9.75287318e-01 -5.96560061e-01 -5.04931450e-01 -2.02720668e-02
7.24367619e-01 -9.70931590e-01 -5.22392333e-01 7.45264888e-01
-1.08202136e+00 -7.35105351e-02 -2.43826479e-01 4.14963692e-01
-2.22112000e-01 3.16744000e-01 -1.52599439e-01 -8.43331575e-01
-2.20394969e-01 -9.54858720e-01 1.02531338e+00 -2.92057395e-02
6.70403615e-02 -1.00117099e+00 3.54234539e-02 3.29780489e-01
5.14877200e-01 4.03480113e-01 1.03031242e+00 -7.01154768e-01
-4.65070307e-01 -3.21467310e-01 -5.54563046e-01 -2.12000106e-02
2.63620585e-01 -2.07432240e-01 -6.80743039e-01 2.86305565e-02
-3.61463755e-01 7.66984671e-02 9.01957929e-01 3.20763588e-01
1.83670580e+00 -4.43575770e-01 -3.69652092e-01 6.69191897e-01
8.67939532e-01 2.14001358e-01 7.90732026e-01 5.11593938e-01
9.38400626e-01 6.32909596e-01 7.61797488e-01 8.50889504e-01
8.19390714e-01 8.27258766e-01 6.41832173e-01 1.40433088e-02
3.89044881e-01 -2.17071488e-01 4.60959300e-02 5.51804483e-01
1.31112263e-01 -1.70087129e-01 -1.27980542e+00 8.61771047e-01
-2.18024516e+00 -1.27647674e+00 -6.48492932e-01 2.14617562e+00
1.45818278e-01 3.49342108e-01 3.04783493e-01 5.03782153e-01
5.45391202e-01 3.04865718e-01 -7.81396389e-01 -1.39589280e-01
-1.83080301e-01 -8.98308828e-02 3.52506101e-01 2.01983020e-01
-1.49976349e+00 9.10691679e-01 4.96342230e+00 1.18879950e+00
-1.16759896e+00 1.20930918e-01 1.02216709e+00 -2.40471214e-01
-1.72489211e-01 -4.76091862e-01 -6.30694270e-01 1.15931308e+00
9.20615315e-01 -1.04866043e-01 1.18134551e-01 8.49357903e-01
5.28858244e-01 2.93139607e-01 -4.60415781e-01 9.69059408e-01
-2.90791184e-01 -8.71562958e-01 -2.48490069e-02 2.47292340e-01
6.92136347e-01 1.33929804e-01 4.17246521e-01 5.84885776e-01
1.21832378e-01 -1.06944668e+00 2.92943448e-01 6.44226968e-01
5.16869962e-01 -1.12037230e+00 9.16215062e-01 5.88644028e-01
-1.61196208e+00 -4.29662317e-01 -2.87133425e-01 -2.11615458e-01
3.22810501e-01 9.85323489e-01 -2.20622480e-01 5.36505163e-01
9.50074136e-01 1.07578921e+00 -5.63603878e-01 1.03964102e+00
-4.49488163e-02 7.05378532e-01 -4.26430136e-01 4.16144669e-01
4.00950193e-01 -2.81747550e-01 5.76718986e-01 9.91782188e-01
6.81912601e-01 3.25670660e-01 2.21738607e-01 5.48453271e-01
3.47754151e-01 2.29654402e-01 -7.69258142e-01 4.49562699e-01
3.75654578e-01 8.93649518e-01 -1.90580994e-01 -9.15370882e-02
-5.45903921e-01 4.96670365e-01 1.53355867e-01 4.59491938e-01
-1.10548627e+00 -2.43736163e-01 1.18146574e+00 4.93796200e-01
1.59009755e-01 -1.08866826e-01 -7.30947018e-01 -1.22941196e+00
3.20737630e-01 -6.08891129e-01 6.75705969e-01 -2.09198996e-01
-1.43287385e+00 4.54792649e-01 -4.60701771e-02 -1.48695552e+00
-2.85959184e-01 -2.97748268e-01 -1.27906978e+00 6.05009973e-01
-1.54938102e+00 -1.14149690e+00 -3.76714498e-01 9.07544017e-01
5.86480558e-01 -5.76517284e-01 4.26186740e-01 9.14967418e-01
-1.32287085e+00 8.61070752e-01 1.21169038e-01 2.54088074e-01
2.18209445e-01 -7.40474045e-01 6.15171731e-01 1.08964884e+00
-1.36864856e-01 2.01186985e-02 3.75261992e-01 -4.82571006e-01
-7.16809452e-01 -1.67386663e+00 1.06936026e+00 -3.11469913e-01
7.93848932e-01 -7.34627172e-02 -1.20597696e+00 5.64423859e-01
-5.83373845e-01 3.85582805e-01 6.27040267e-01 4.12753135e-01
-2.11427510e-01 -3.89084578e-01 -1.12184513e+00 6.23523235e-01
1.15685534e+00 -2.59669989e-01 -3.82938609e-02 3.90001327e-01
6.34096324e-01 -6.37301952e-02 -6.23472393e-01 6.65963292e-01
1.68612793e-01 -1.06542718e+00 1.08813274e+00 -8.47516358e-01
4.42218065e-01 -1.76628470e-01 -2.39459127e-01 -1.26538527e+00
-4.53893751e-01 1.92882363e-02 -3.40682238e-01 1.22220206e+00
3.29927206e-01 -9.98971522e-01 7.41566539e-01 6.14660800e-01
-3.48833233e-01 -1.22683024e+00 -1.41666877e+00 -5.92911661e-01
1.38702154e-01 -1.06440711e+00 1.19704199e+00 1.07684016e+00
-3.81567568e-01 2.75688052e-01 -5.89967072e-01 4.82868880e-01
5.45328319e-01 -1.02525190e-01 9.18669343e-01 -1.52891016e+00
2.97620147e-01 -6.31606162e-01 -7.95732141e-01 -8.36194873e-01
3.64211917e-01 -4.60924327e-01 -4.69231367e-01 -1.18834829e+00
-2.07591370e-01 -7.59651244e-01 -6.65948927e-01 4.36370611e-01
-3.17765057e-01 -6.92830980e-02 -3.49234939e-01 9.11217555e-02
-7.17219651e-01 1.18975365e+00 7.09565163e-01 -2.02455893e-01
-1.74461305e-01 5.39239943e-01 -5.51362336e-01 9.69313562e-01
1.08712995e+00 -4.67093796e-01 -5.17070651e-01 -6.38569951e-01
2.11546063e-01 -2.41540760e-01 3.89553756e-01 -1.22748828e+00
1.36466026e-01 -5.60684621e-01 2.47628093e-01 -8.80256176e-01
2.30512723e-01 -9.12422776e-01 -1.90919355e-01 3.14159662e-01
1.45444646e-01 4.85413164e-01 2.91710168e-01 8.68207276e-01
-3.70219111e-01 6.11508787e-01 4.99547482e-01 3.30600739e-01
-7.79696524e-01 9.93739426e-01 -2.28227332e-01 3.50107193e-01
1.37269044e+00 -2.33800963e-01 -1.56767100e-01 -3.22899580e-01
-4.72147465e-01 5.96115887e-01 -1.19748913e-01 7.45788276e-01
7.89439499e-01 -1.82792413e+00 -8.69980395e-01 6.47053003e-01
4.03957307e-01 1.99863762e-01 7.84857333e-01 1.09550333e+00
1.65848862e-02 5.10616243e-01 -3.03397775e-02 -6.67560935e-01
-8.50071311e-01 5.07596314e-01 3.35753322e-01 -5.46026409e-01
-6.90658629e-01 4.09419388e-01 4.53602701e-01 -6.56120896e-01
9.11695361e-02 -1.91988155e-01 -2.92257041e-01 -1.88987330e-02
5.31406403e-01 7.49710441e-01 1.46344319e-01 -1.03486609e+00
-4.54750270e-01 6.72442138e-01 7.67907053e-02 4.65792030e-01
1.34076273e+00 -3.04369092e-01 1.76394090e-01 4.41278040e-01
9.86455083e-01 -3.47573757e-01 -1.01404881e+00 -6.60014808e-01
1.46127433e-01 -6.92394495e-01 2.80327648e-02 -4.41934794e-01
-1.49436438e+00 1.13278580e+00 6.24466360e-01 3.01305681e-01
1.26055658e+00 -1.37548074e-01 1.40771806e+00 1.71222568e-01
7.82736391e-02 -1.03499436e+00 1.85644746e-01 8.23442459e-01
6.33454144e-01 -1.50616705e+00 -2.90072411e-01 -3.10354084e-01
-7.08809733e-01 5.28418243e-01 9.72430944e-01 -3.52197587e-02
1.23048127e+00 -1.46513984e-01 -2.35297382e-01 -2.59703636e-01
-6.65578246e-01 -2.65821964e-01 4.71804202e-01 4.80035543e-01
1.90359831e-01 3.04541439e-01 -1.20442688e-01 8.13133240e-01
-1.90566242e-01 -3.21362019e-01 -1.03635177e-01 3.87033105e-01
-3.10644060e-01 -6.56987548e-01 -2.43010357e-01 7.21836269e-01
-5.77520430e-01 1.92379411e-02 1.50341287e-01 6.16157711e-01
4.43985105e-01 1.00518131e+00 4.86087561e-01 -6.64051712e-01
6.43701494e-01 -1.34782046e-01 -5.51215053e-01 -7.59519413e-02
-1.10933684e-01 -3.45243633e-01 -1.78744718e-01 -8.28148067e-01
1.90320089e-01 -8.32429290e-01 -1.05228424e+00 -7.56542623e-01
1.69116482e-01 -8.89078230e-02 1.82640389e-01 1.35125661e+00
7.28330255e-01 6.88472331e-01 8.77729893e-01 -2.48538136e-01
-2.44052142e-01 -9.31318939e-01 -4.94307399e-01 7.32374012e-01
4.16767806e-01 -9.97675657e-01 -3.16031367e-01 -6.10838532e-01] | [6.5339131355285645, 2.078439950942993] |
5ff1f282-be78-45ff-bd6f-fb59139603f9 | learning-from-high-dimensional-cyber-physical | 2303.08300 | null | https://arxiv.org/abs/2303.08300v1 | https://arxiv.org/pdf/2303.08300v1.pdf | Learning From High-Dimensional Cyber-Physical Data Streams for Diagnosing Faults in Smart Grids | The performance of fault diagnosis systems is highly affected by data quality in cyber-physical power systems. These systems generate massive amounts of data that overburden the system with excessive computational costs. Another issue is the presence of noise in recorded measurements, which prevents building a precise decision model. Furthermore, the diagnostic model is often provided with a mixture of redundant measurements that may deviate it from learning normal and fault distributions. This paper presents the effect of feature engineering on mitigating the aforementioned challenges in cyber-physical systems. Feature selection and dimensionality reduction methods are combined with decision models to simulate data-driven fault diagnosis in a 118-bus power system. A comparative study is enabled accordingly to compare several advanced techniques in both domains. Dimensionality reduction and feature selection methods are compared both jointly and separately. Finally, experiments are concluded, and a setting is suggested that enhances data quality for fault diagnosis. | ['Mehrdad Saif', 'Roozbeh Razavi-Far', 'Ehsan Hallaji', 'Hossein Hassani'] | 2023-03-15 | null | null | null | null | ['feature-engineering'] | ['methodology'] | [-2.57382765e-02 -4.36654538e-01 3.34557891e-02 -1.91788614e-01
-4.62265104e-01 -4.02399272e-01 2.42266223e-01 1.12006016e-01
4.46330100e-01 9.92254972e-01 -3.72103810e-01 -2.63485074e-01
-9.73011374e-01 -7.28295863e-01 -2.82576736e-02 -1.00049984e+00
-5.62284231e-01 5.45469224e-01 -3.17234725e-01 3.68976314e-03
2.32279599e-01 7.29557753e-01 -1.60689998e+00 -9.29432586e-02
1.15388417e+00 9.87579703e-01 -2.58394964e-02 5.70220888e-01
5.13201892e-01 3.85795385e-01 -1.28912759e+00 4.88795578e-01
5.64480007e-01 -4.18975145e-01 -4.14609730e-01 6.91621482e-01
-3.63489658e-01 -4.78230357e-01 -6.84885740e-01 1.26595926e+00
4.69198674e-01 1.17714241e-01 6.73345149e-01 -1.99597979e+00
-1.55462638e-01 3.71678174e-01 -3.26830417e-01 4.76386756e-01
3.25946301e-01 3.51700544e-01 5.46010315e-01 -8.98975790e-01
9.14334133e-02 9.62337136e-01 4.25953031e-01 4.70839255e-02
-1.11730397e+00 -3.92745197e-01 -1.60262212e-01 4.03133392e-01
-1.46503341e+00 -1.78906664e-01 1.05687034e+00 -4.62900251e-01
1.16631246e+00 4.95616257e-01 6.83626831e-01 6.25881612e-01
7.49555528e-01 4.14812267e-01 9.66218174e-01 -2.81315058e-01
6.96576178e-01 -1.06529854e-01 5.13694167e-01 2.61946410e-01
8.67184579e-01 5.03039300e-01 -1.81219533e-01 -5.50153494e-01
4.17990714e-01 3.13314706e-01 -4.94696438e-01 -3.33785027e-01
-6.68530762e-01 6.66208208e-01 1.04099698e-02 4.68085796e-01
-5.92636645e-01 -3.08640361e-01 5.61862350e-01 9.98101473e-01
1.53367430e-01 8.57894719e-01 -7.69413888e-01 -1.73272893e-01
-8.94448221e-01 2.99695909e-01 9.01601970e-01 9.52798665e-01
4.82122689e-01 8.04994583e-01 5.48541099e-02 4.36468899e-01
1.83773190e-01 6.91785812e-01 6.42441511e-01 -7.09601939e-01
3.68208885e-01 8.30624163e-01 1.29495993e-01 -1.02082455e+00
-6.78455591e-01 -7.52959609e-01 -9.48220193e-01 5.37579179e-01
4.88779396e-02 -4.51005250e-01 -6.70376241e-01 1.00406158e+00
2.28768483e-01 -2.08133142e-02 1.49861246e-01 9.61097896e-01
-9.24047232e-02 4.05712485e-01 -5.29698670e-01 -5.16170323e-01
1.10431468e+00 -3.40767592e-01 -1.24080551e+00 7.76195228e-02
6.81200266e-01 -4.82254237e-01 4.79816169e-01 1.02642512e+00
-8.69555891e-01 -4.04976338e-01 -1.54783964e+00 8.80760789e-01
-2.93681592e-01 1.63009465e-01 5.12679756e-01 7.23543882e-01
-6.61317229e-01 1.07547915e+00 -9.00504470e-01 -2.09257975e-02
2.78068274e-01 3.90469432e-01 -3.67888272e-01 -3.13309193e-01
-1.51923668e+00 1.29442096e+00 3.15015674e-01 3.29682261e-01
-9.20501769e-01 -7.09175825e-01 -7.64173746e-01 2.37628538e-02
3.39266032e-01 -5.58111787e-01 9.77729857e-01 -4.35343057e-01
-1.09347677e+00 -3.15155387e-01 4.51481849e-01 -6.33656323e-01
4.79085058e-01 -1.22082971e-01 -1.14765728e+00 3.74985814e-01
-5.96031547e-02 -5.63266814e-01 1.17225754e+00 -9.37890768e-01
-8.47655237e-01 -5.72517931e-01 -4.54514593e-01 8.46582875e-02
-7.36678019e-02 -4.86590594e-01 5.75322568e-01 -4.62423354e-01
3.20003241e-01 -5.50857008e-01 -3.59483629e-01 -4.89359111e-01
-7.44434178e-01 1.08718626e-01 1.55613399e+00 -5.93901277e-01
1.45166588e+00 -2.00374603e+00 -1.46289587e-01 7.20574200e-01
2.53066659e-01 2.12761834e-01 3.30987155e-01 9.42677855e-01
-5.97438753e-01 -1.34785846e-01 -1.25105947e-01 8.69806930e-02
1.81144789e-01 5.31992137e-01 -2.66829997e-01 1.04665184e+00
6.04724944e-01 2.47973189e-01 -6.69677675e-01 1.50207013e-01
6.70297265e-01 8.84479508e-02 -1.22105546e-01 1.17120236e-01
2.55161375e-01 2.83606440e-01 -6.28048480e-01 8.01811397e-01
8.11848581e-01 5.08131366e-03 1.14847846e-01 -3.03737104e-01
2.83647478e-01 1.12434931e-01 -1.71169078e+00 1.12635159e+00
-1.87302917e-01 2.70011425e-01 1.16082475e-01 -1.36948156e+00
1.04238188e+00 5.95408678e-01 8.62089872e-01 -4.22319889e-01
3.35922539e-01 1.01065591e-01 4.46640015e-01 -5.99511325e-01
2.72457868e-01 -2.27612797e-02 -3.28015119e-01 5.17295122e-01
2.51514018e-01 -4.53183532e-01 4.41328436e-01 -6.42946586e-02
1.63135183e+00 -6.46131396e-01 4.16438699e-01 -6.45500600e-01
3.43230277e-01 2.52974153e-01 1.07348454e+00 4.56698239e-01
-1.91190287e-01 2.51892507e-01 5.28698504e-01 -2.59713471e-01
-8.08762133e-01 -8.88229311e-01 -5.04854918e-01 -2.02543437e-01
-1.36888931e-02 -2.93887019e-01 -4.29996938e-01 -7.27601230e-01
5.82142591e-01 9.17360306e-01 -4.76705283e-01 -9.64905381e-01
1.00165932e-02 -1.19590259e+00 3.51216882e-01 3.57445538e-01
6.35243952e-02 -3.63995373e-01 -7.55599082e-01 4.72177148e-01
5.66605508e-01 -5.90274572e-01 1.99325651e-01 7.47615099e-01
-1.03483737e+00 -1.66186798e+00 -9.56287086e-02 -1.96243063e-01
9.94111061e-01 1.07645839e-01 1.04759097e+00 2.45861828e-01
-7.57331610e-01 1.72285616e-01 -4.88769412e-01 -2.67768264e-01
-5.48950851e-01 -4.98975933e-01 8.07155073e-01 -3.15672785e-01
3.53463918e-01 -4.95737255e-01 -2.62555093e-01 3.77002716e-01
-1.00680685e+00 -6.75845146e-01 5.97368240e-01 1.42519367e+00
5.54528683e-02 1.49531519e+00 1.11724043e+00 -6.95002139e-01
9.39628661e-01 -7.52065659e-01 -8.08237970e-01 9.45692137e-02
-1.14402318e+00 -1.34566024e-01 8.85300696e-01 -2.64325380e-01
-7.00490475e-01 -3.79619598e-02 2.71243930e-01 -6.89251542e-01
-4.58032727e-01 5.76736987e-01 -5.17484725e-01 -8.76099765e-02
5.27045369e-01 -8.21742117e-02 2.92146355e-01 -5.13158739e-01
-1.84054628e-01 7.70844698e-01 2.49562174e-01 -1.93024397e-01
1.02929699e+00 -6.72146827e-02 2.97299236e-01 -8.40835035e-01
-1.16466403e-01 -9.90153030e-02 -6.51920259e-01 3.77386920e-02
-2.73079246e-01 -7.57749617e-01 -3.78066510e-01 6.48545027e-01
-4.70899373e-01 1.30684674e-01 -6.65935874e-01 6.11754179e-01
-3.48593175e-01 3.38939220e-01 -7.59679198e-01 -8.84996951e-01
-8.53317007e-02 -1.26986086e+00 5.41225255e-01 7.93325678e-02
-2.31524080e-01 -1.02757311e+00 -8.36275741e-02 -2.99881965e-01
2.08933741e-01 4.66416627e-01 9.41147983e-01 -1.00102627e+00
2.78994851e-02 -8.87543261e-01 3.98362160e-01 6.67850912e-01
6.74720049e-01 2.58308440e-01 -8.40224683e-01 -7.95711696e-01
6.00504458e-01 1.88051000e-01 -3.05387359e-02 3.26990098e-01
6.73977852e-01 -2.28806153e-01 -3.75097662e-01 9.61280614e-02
1.39979601e+00 6.62008643e-01 2.84416348e-01 5.79531826e-02
4.09385085e-01 3.76585931e-01 1.26696396e+00 1.01174545e+00
-1.93774372e-01 2.60345936e-01 4.34540272e-01 4.47790995e-02
3.26134175e-01 4.20810342e-01 6.15961999e-02 8.86025488e-01
6.21877193e-01 -1.45858511e-01 -8.95019710e-01 4.73838806e-01
-1.56467021e+00 -6.28090084e-01 -4.06261802e-01 2.02328801e+00
3.71930897e-01 2.26251557e-01 -1.14355385e-02 1.41176426e+00
7.10759223e-01 -5.83773851e-01 -8.18546891e-01 -8.47553760e-02
-2.25832418e-01 -3.18186842e-02 2.84462094e-01 3.07117969e-01
-1.01538098e+00 -1.29551977e-01 6.70449686e+00 4.83612508e-01
-9.55718279e-01 -3.11599523e-01 3.95989031e-01 1.99176930e-02
2.41335005e-01 -1.66139022e-01 -3.76834422e-01 7.31326044e-01
1.29410732e+00 -6.52198553e-01 1.27257779e-01 9.09556925e-01
4.29349989e-01 -4.15449739e-01 -1.25936818e+00 9.54199851e-01
-1.37706578e-01 -7.00270653e-01 -3.23310167e-01 2.35258102e-01
8.34712923e-01 -3.58000010e-01 -8.57542530e-02 9.73215178e-02
2.93088555e-01 -7.09517181e-01 3.37851614e-01 5.18299103e-01
3.63992691e-01 -1.18157279e+00 1.15746069e+00 2.92900831e-01
-7.75286913e-01 -7.84707010e-01 -2.89235085e-01 -3.39797735e-01
2.35137120e-01 1.46948504e+00 -9.25738096e-01 1.21701658e+00
4.86371964e-01 8.13040018e-01 -5.45534134e-01 1.28125298e+00
-4.75621186e-02 5.13310134e-01 -2.29881808e-01 3.98812294e-01
-2.23538861e-01 -1.08384393e-01 5.87122500e-01 5.02364933e-01
5.49086988e-01 -9.42960456e-02 5.27095973e-01 6.12330496e-01
7.06094027e-01 -6.47487402e-01 -8.17798436e-01 -7.49596655e-02
1.07890046e+00 1.21981454e+00 -5.26997507e-01 -4.91860688e-01
-2.88516968e-01 5.55692613e-01 -4.82982993e-01 1.99220985e-01
-5.21726549e-01 -5.94241202e-01 9.17635739e-01 -1.72892049e-01
-2.40203992e-01 3.61561240e-03 -3.94904435e-01 -9.76040065e-01
-1.22887582e-01 -1.10188532e+00 5.13838232e-01 -3.72861356e-01
-1.85639119e+00 5.04248977e-01 1.30863348e-02 -1.92609644e+00
-8.47815514e-01 -5.26694357e-01 -5.24542570e-01 9.91704226e-01
-1.12393177e+00 -1.92762196e-01 -2.00772807e-02 5.63814342e-01
5.02758563e-01 -3.87870789e-01 8.50179851e-01 5.23727179e-01
-1.05990458e+00 1.69130027e-01 4.66779441e-01 -2.00793087e-01
2.48663247e-01 -1.44312215e+00 1.48595765e-01 1.13946331e+00
-4.17080283e-01 2.24865958e-01 1.15462565e+00 -9.13016319e-01
-2.10662937e+00 -1.11320889e+00 1.71074241e-01 -5.35370857e-02
7.30858266e-01 2.85413787e-02 -1.16214514e+00 2.18541279e-01
-7.65520036e-02 3.08427781e-01 6.19121969e-01 -3.05993170e-01
4.34844792e-01 -8.44678655e-02 -1.85916591e+00 6.67141154e-02
2.04893976e-01 -4.47953314e-01 -8.15260887e-01 3.09730470e-01
6.09439611e-03 -1.97992995e-01 -1.41399443e+00 4.06888008e-01
-3.09597522e-01 -5.72485149e-01 6.30710602e-01 -4.52337712e-01
-2.91326612e-01 -6.65733576e-01 -5.25180474e-02 -2.13644695e+00
-4.91438508e-01 -5.63973367e-01 -5.75087845e-01 1.13498604e+00
1.05044678e-01 -9.86758828e-01 5.11574507e-01 5.95428944e-01
-2.69022346e-01 -4.90479201e-01 -1.08894610e+00 -1.11893225e+00
-1.22552067e-01 -2.84132600e-01 9.44423735e-01 1.25656176e+00
6.93405330e-01 2.33415127e-01 -4.10763472e-02 7.20364809e-01
7.31907606e-01 8.56275037e-02 3.49033237e-01 -1.42874134e+00
-1.98569521e-01 -1.77515879e-01 -8.05519104e-01 -1.24230441e-02
-1.70459062e-01 -3.06760103e-01 6.58857450e-02 -1.13666511e+00
-5.23346543e-01 -5.21084726e-01 -3.73827845e-01 2.51141340e-01
-7.54814446e-02 -2.19213575e-01 -1.13758281e-01 1.15868136e-01
-1.09461807e-01 6.82241499e-01 7.61310220e-01 -6.10884987e-02
1.39680624e-01 1.27023458e-01 -1.39359191e-01 4.79416758e-01
9.69302237e-01 -4.16999161e-01 -7.13047206e-01 1.19851716e-01
-3.08911830e-01 8.17435443e-01 8.65333155e-02 -1.39795339e+00
1.23828083e-01 6.73118457e-02 7.58998811e-01 -8.02701056e-01
1.24710269e-01 -1.59004986e+00 4.48530644e-01 9.68674898e-01
3.23007554e-01 6.75120175e-01 5.26829325e-02 4.71869349e-01
-5.09648383e-01 -3.37204456e-01 4.77349401e-01 1.87774464e-01
-6.33581936e-01 1.05030276e-01 -7.43474543e-01 -4.35271084e-01
1.32949817e+00 -1.26053795e-01 -3.84774297e-01 -4.78171930e-02
-9.01030540e-01 5.60852110e-01 3.54626060e-01 4.72322673e-01
6.34585023e-01 -1.41903341e+00 -5.04672229e-01 6.82863772e-01
-1.87034309e-01 -1.64573580e-01 5.49918234e-01 8.56558383e-01
-1.25549451e-01 4.78186131e-01 -3.51034969e-01 -5.93253851e-01
-9.45313871e-01 7.26576746e-01 3.76195461e-01 -1.24993853e-01
-7.16028333e-01 7.99172223e-02 -6.47533834e-01 -1.84656873e-01
-1.56633221e-02 -5.32503188e-01 5.49968854e-02 1.20884374e-01
7.19769418e-01 9.85672176e-01 7.33998716e-01 -2.35210076e-01
-4.00539130e-01 -1.70012489e-01 1.40245423e-01 2.33805373e-01
1.28959203e+00 -3.06370616e-01 1.57235712e-01 5.45962214e-01
8.16453695e-01 -8.10571492e-01 -1.42358506e+00 5.34521341e-02
3.53661180e-01 -7.13319182e-01 3.07015300e-01 -1.00381517e+00
-1.35635805e+00 6.22877359e-01 7.93391645e-01 9.15894926e-01
1.42452693e+00 -7.24168420e-01 1.72780856e-01 3.91507626e-01
6.52211249e-01 -1.26919985e+00 -2.27140293e-01 2.63424516e-01
7.23675191e-01 -9.29024458e-01 2.01973155e-01 -3.17441851e-01
-4.47182775e-01 1.17428219e+00 4.64644790e-01 -7.28035688e-01
9.22105074e-01 1.03873241e+00 -1.18333988e-01 -3.67358387e-01
-1.14470208e+00 3.49970818e-01 -4.51414473e-02 8.11097682e-01
-1.62787527e-01 2.43029818e-01 -2.09826276e-01 7.84689724e-01
-1.83769334e-02 -1.42811045e-01 8.15871656e-01 1.60932720e+00
-2.28541747e-01 -9.82115924e-01 -9.59483087e-01 9.20927823e-01
-4.06702161e-01 4.92921650e-01 3.15040015e-02 1.01330578e+00
-1.16380304e-01 1.47106659e+00 1.20171219e-01 -5.35794079e-01
6.70588434e-01 -6.71741590e-02 1.41807452e-01 -4.72064793e-01
-4.47380066e-01 2.93081179e-02 1.98173895e-02 -5.00776350e-01
3.60980719e-01 -8.95755410e-01 -1.29384458e+00 -1.97164312e-01
-9.26335752e-01 6.69411600e-01 6.46383047e-01 9.59447861e-01
3.48783135e-01 1.26002324e+00 1.16635787e+00 -7.12811232e-01
-1.31639719e+00 -1.17050743e+00 -1.49790072e+00 1.40127450e-01
4.97314721e-01 -1.14831722e+00 -8.01265717e-01 -3.65083605e-01] | [6.197067737579346, 2.5535120964050293] |
856e3658-898f-4760-ba62-e59c15a76d6d | differential-equation-and-probability | 2202.13800 | null | https://arxiv.org/abs/2202.13800v2 | https://arxiv.org/pdf/2202.13800v2.pdf | Differential equation and probability inspired graph neural networks for latent variable learning | Probabilistic theory and differential equation are powerful tools for the interpretability and guidance of the design of machine learning models, especially for illuminating the mathematical motivation of learning latent variable from observation. Subspace learning maps high-dimensional features on low-dimensional subspace to capture efficient representation. Graphs are widely applied for modeling latent variable learning problems, and graph neural networks implement deep learning architectures on graphs. Inspired by probabilistic theory and differential equations, this paper conducts notes and proposals about graph neural networks to solve subspace learning problems by variational inference and differential equation. Source code of this paper is available at https://github.com/zshicode/Latent-variable-GNN. | ['Zhuangwei Shi'] | 2022-02-28 | null | null | null | null | ['stock-prediction'] | ['time-series'] | [-2.76588827e-01 2.73949534e-01 -8.03762913e-01 -1.65504485e-01
-2.67032087e-01 -5.23772597e-01 5.46214163e-01 -6.21490359e-01
2.98854560e-01 5.49569905e-01 3.66942793e-01 -3.29356492e-01
-3.08666021e-01 -7.43124068e-01 -3.90688270e-01 -1.07220840e+00
-4.36903536e-02 5.45757294e-01 -7.11508393e-01 4.05624211e-01
-6.38576671e-02 4.59727407e-01 -8.14289153e-01 -1.33074209e-01
5.85575461e-01 4.76926923e-01 2.50311829e-02 5.42443633e-01
-2.69767016e-01 4.50139791e-01 1.67011514e-01 -1.39133900e-01
6.04629442e-02 -5.48648000e-01 -4.37843949e-01 -4.86592390e-02
6.55916780e-02 -1.09364979e-01 -1.11233139e+00 1.32764447e+00
1.25043049e-01 2.97877342e-01 1.26573372e+00 -1.69139135e+00
-1.64709699e+00 8.53312731e-01 -4.79794234e-01 1.58697158e-01
5.74232079e-02 -7.97527537e-05 1.24752724e+00 -1.12989914e+00
7.18811810e-01 1.43975544e+00 3.89782578e-01 7.35303521e-01
-1.41177857e+00 -3.87211710e-01 3.86877432e-02 3.95785421e-01
-1.40993464e+00 -2.81922311e-01 1.37976193e+00 -8.88860524e-01
6.45606697e-01 1.73943788e-01 7.32073724e-01 1.61840892e+00
4.74759072e-01 1.17325258e+00 6.79115236e-01 -3.24077457e-01
2.50046998e-01 4.05600369e-02 4.24659759e-01 1.02251804e+00
2.38393933e-01 1.69637620e-01 -5.10058641e-01 -3.32329452e-01
1.12688208e+00 6.01452708e-01 -2.80229598e-01 -1.08698249e+00
-9.85465825e-01 1.49005640e+00 2.66583264e-01 1.32756323e-01
-4.56640810e-01 4.07658756e-01 7.30601698e-02 -3.03084701e-02
6.08849168e-01 -2.91625801e-02 -1.04683608e-01 6.88934773e-02
-8.19981813e-01 -3.38495821e-02 5.83716929e-01 1.13518238e+00
7.98638761e-01 6.71041250e-01 -9.43706483e-02 5.80868900e-01
1.00587404e+00 9.09073412e-01 3.82251203e-01 -1.20613682e+00
3.43960039e-02 5.92305005e-01 -3.48161370e-01 -1.03164506e+00
-2.86521137e-01 -4.34335053e-01 -1.28115940e+00 2.24442501e-02
9.55314934e-02 -3.07212263e-01 -8.72830868e-01 1.57436216e+00
3.99346054e-02 2.93487728e-01 8.87440797e-03 1.13361669e+00
8.36125612e-01 6.96552217e-01 -1.54309794e-01 -3.76541644e-01
8.51857781e-01 -8.34392250e-01 -1.11074007e+00 7.92209283e-02
5.16946614e-01 -3.97413522e-02 7.41217375e-01 3.90250795e-02
-7.46621013e-01 -2.49360666e-01 -6.97844684e-01 -2.67522991e-01
-2.26056546e-01 1.41032279e-01 9.53001916e-01 3.39388609e-01
-1.24691367e+00 3.59744817e-01 -1.20915830e+00 -2.25412652e-01
5.15061677e-01 1.76811427e-01 -3.62611145e-01 9.13645625e-02
-1.19596386e+00 3.35280269e-01 -7.74663910e-02 2.78164119e-01
-1.06609392e+00 -3.49060416e-01 -1.16324115e+00 7.50913695e-02
1.12833731e-01 -9.62557137e-01 8.61371756e-01 -6.18950427e-01
-1.53842902e+00 6.17747188e-01 -5.54919958e-01 -3.48823339e-01
1.89453661e-01 -2.62782961e-01 -2.96522528e-01 3.67053561e-02
-6.81940140e-03 1.63086712e-01 1.12682426e+00 -9.55168426e-01
2.73706228e-01 -6.81647062e-01 -2.59022385e-01 -3.25199636e-03
-3.13387185e-01 -3.24897856e-01 -3.88194680e-01 -3.84442955e-01
6.25689626e-01 -1.04831982e+00 -3.30792189e-01 9.79061518e-03
-5.04314125e-01 -4.37645137e-01 1.08276367e+00 -5.49008608e-01
1.00799668e+00 -2.03853297e+00 8.49614501e-01 1.31095707e-01
8.64955902e-01 -2.42273375e-01 -2.09087729e-02 4.56563354e-01
-3.76843750e-01 3.32273804e-02 -4.12989318e-01 -3.14993083e-01
2.36665621e-01 1.49492115e-01 -7.87709713e-01 1.02581418e+00
-1.00079156e-01 1.31947923e+00 -9.40913975e-01 -2.59791732e-01
5.26769936e-01 7.18852997e-01 -3.65695596e-01 8.92477334e-02
-2.56153166e-01 4.47991282e-01 -8.34773958e-01 4.97777522e-01
4.24660832e-01 -6.34368777e-01 6.31700009e-02 -5.19140530e-03
1.11687362e-01 -1.74442828e-01 -1.23149896e+00 1.88390350e+00
-1.99304353e-02 1.06104183e+00 -7.96470642e-02 -1.21960318e+00
8.15682054e-01 4.38544482e-01 5.27727306e-01 -1.66455910e-01
2.23600447e-01 -1.74245596e-01 -4.14529860e-01 -6.08867824e-01
1.82676613e-01 -5.91942435e-03 6.54071048e-02 4.62809324e-01
2.22829387e-01 6.38628006e-02 -1.82200387e-01 6.73527420e-01
5.98350465e-01 2.09165677e-01 1.87600374e-01 -4.76159990e-01
4.26883608e-01 -3.16904068e-01 6.10548377e-01 5.84284663e-01
-2.81024694e-01 3.32476437e-01 6.06078744e-01 -5.23961306e-01
-8.10070634e-01 -1.65743387e+00 -1.42328948e-01 9.05093968e-01
1.77395614e-05 -3.43029827e-01 -4.48626697e-01 -4.08740401e-01
6.88961372e-02 9.88668382e-01 -7.72860229e-01 -2.44994000e-01
-1.01951241e-01 -2.81415522e-01 2.89548278e-01 6.81649625e-01
-1.28074642e-02 -8.06825280e-01 1.83764040e-01 -1.70554340e-01
-1.21870697e-01 -4.80494261e-01 -3.19831014e-01 1.00475945e-01
-9.08674240e-01 -8.14592361e-01 -8.17144275e-01 -5.85738838e-01
8.22456658e-01 3.04342389e-01 8.17057133e-01 -2.57643163e-01
-1.92364156e-01 8.95482838e-01 1.48813305e-02 -2.69661218e-01
-6.74908459e-02 -7.53256679e-02 6.33148253e-01 -1.03163347e-01
7.96225131e-01 -7.72196889e-01 -4.05304164e-01 -5.54580241e-02
-7.01950908e-01 8.71152356e-02 2.44813129e-01 7.63287663e-01
6.80557013e-01 -2.88117468e-01 5.41411014e-03 -8.69142830e-01
6.10693157e-01 -7.64578700e-01 -9.65326011e-01 5.19266315e-02
-6.44129932e-01 4.65389729e-01 5.04142344e-01 -3.47231120e-01
-8.66830170e-01 -3.55153121e-02 4.33788389e-01 -1.11242497e+00
-1.14149243e-01 7.27811933e-01 -2.41647094e-01 4.26761001e-01
5.30698895e-01 3.02364975e-01 3.46623100e-02 -4.93238032e-01
8.48198771e-01 3.78391951e-01 1.46371841e-01 -2.75116026e-01
1.04857123e+00 6.60898864e-01 2.89135247e-01 -8.31287920e-01
-7.03306854e-01 -3.62338156e-01 -1.14049995e+00 -2.60319829e-01
1.07000470e+00 -9.70593512e-01 -4.05050635e-01 1.88020438e-01
-8.48107755e-01 -4.51314330e-01 -9.88782570e-02 8.05935502e-01
-7.55682886e-01 2.78883576e-01 -7.47194827e-01 -8.51999640e-01
-1.05232440e-01 -9.11322713e-01 9.38227296e-01 2.69100159e-01
-2.27150992e-01 -1.74248016e+00 3.88383240e-01 1.14176571e-01
1.08000636e-01 4.17676456e-02 8.17815542e-01 -1.56425938e-01
-7.63537943e-01 -1.54235557e-01 6.67342842e-02 4.71375585e-02
2.97825634e-02 2.71329343e-01 -8.94629478e-01 -3.28575701e-01
9.64606479e-02 1.15114003e-01 1.15219724e+00 1.08760929e+00
1.22314942e+00 -2.95950353e-01 -6.30138457e-01 1.13974917e+00
1.37501431e+00 -3.38210881e-01 2.53020704e-01 -2.63592750e-01
1.30626059e+00 4.02206689e-01 -2.60014236e-01 3.14049661e-01
4.41490710e-01 3.17426473e-01 2.34602883e-01 7.56963864e-02
1.69047639e-01 -2.35477194e-01 4.89949137e-01 1.17404926e+00
-2.08582240e-03 -2.15379000e-01 -1.19955194e+00 5.37311316e-01
-2.20954609e+00 -1.26837444e+00 -5.97896934e-01 1.65739655e+00
9.75781083e-02 -3.49678814e-01 -3.93612534e-02 -3.92489254e-01
8.91849399e-01 5.63840270e-01 -8.14452589e-01 -2.93608338e-01
-3.27104658e-01 -3.81392628e-01 4.96621341e-01 8.09251904e-01
-1.07274902e+00 1.15266573e+00 6.26817513e+00 5.57919025e-01
-1.06173384e+00 8.92316923e-02 1.95171580e-01 -1.54103234e-01
-7.86469281e-01 1.26982301e-01 -6.42944574e-01 1.55025378e-01
9.58797812e-01 -6.11247599e-01 7.51644850e-01 9.77175534e-01
4.94427472e-01 5.03538370e-01 -1.10978615e+00 1.15589201e+00
4.75988351e-02 -1.62607741e+00 3.17556024e-01 2.75509536e-01
9.06943083e-01 3.21893275e-01 5.53628266e-01 3.22175145e-01
5.51852167e-01 -1.08288121e+00 3.56514215e-01 8.07252944e-01
7.44246900e-01 -4.07968998e-01 2.86343455e-01 2.22990215e-01
-1.12083364e+00 7.32720345e-02 -7.28148520e-01 -3.35375547e-01
1.07132047e-01 6.28317952e-01 -4.88825262e-01 2.49829307e-01
2.62312561e-01 1.18531466e+00 -3.00850987e-01 6.54478133e-01
-7.60896266e-01 9.81505334e-01 8.60965177e-02 3.22236381e-02
1.60475492e-01 -8.42106819e-01 1.19315267e+00 1.00575089e+00
2.25978583e-01 5.48586585e-02 -7.93754384e-02 1.73746455e+00
4.00717044e-03 -1.52748823e-01 -1.00939596e+00 -6.07321501e-01
4.15332586e-01 1.30323648e+00 -7.61163056e-01 -8.90343413e-02
-3.08666915e-01 9.45730627e-01 3.82524759e-01 1.13943946e+00
-7.87043989e-01 2.28216429e-03 7.23673105e-01 -2.32932776e-01
2.90605575e-01 -8.64745438e-01 -6.38678670e-01 -1.67596722e+00
-2.43838683e-01 -2.59678572e-01 4.23668981e-01 -8.48661482e-01
-1.12018466e+00 1.69072405e-01 8.42622221e-02 -9.34248328e-01
-7.02659070e-01 -7.24714935e-01 -7.99475312e-01 1.24441099e+00
-7.06400514e-01 -1.14618099e+00 -6.55019134e-02 8.24091017e-01
3.37947488e-01 -5.54611385e-01 9.69281614e-01 -3.51642132e-01
-8.08251798e-01 3.40171963e-01 8.46852958e-01 4.28560793e-01
5.86977489e-02 -1.54141676e+00 4.85684693e-01 9.34342444e-01
8.31637621e-01 9.71010447e-01 9.80668664e-01 -7.63641536e-01
-2.05369568e+00 -9.30666685e-01 4.95276749e-01 -8.05155158e-01
1.05275428e+00 -6.32261276e-01 -9.83927727e-01 1.05727649e+00
6.06819429e-02 -4.98173982e-02 8.77730429e-01 6.23036623e-01
-3.11577141e-01 5.48534036e-01 -5.08227646e-01 7.84802020e-01
8.74974489e-01 -1.26506770e+00 -5.65419853e-01 5.50094664e-01
7.91788042e-01 -1.95028320e-01 -5.92158437e-01 -1.11223519e-01
4.92557883e-01 -4.63476658e-01 8.38365674e-01 -9.10278440e-01
4.06501830e-01 -8.22036341e-02 -3.16161126e-01 -1.29460716e+00
-9.32992518e-01 -6.38034701e-01 -8.93607199e-01 9.39053297e-01
3.66387457e-01 -5.85476935e-01 1.00873005e+00 7.61226356e-01
1.24355368e-01 -6.40560508e-01 -8.07895303e-01 -5.96938193e-01
3.66869032e-01 -5.82186401e-01 1.74723968e-01 1.26799190e+00
-9.03152376e-02 5.43178797e-01 -3.20355624e-01 4.10216659e-01
1.16643643e+00 3.67437512e-01 5.89481890e-01 -1.25347805e+00
-2.73748845e-01 -5.57304740e-01 -5.54862380e-01 -9.44726110e-01
8.37132037e-01 -1.30631590e+00 -4.01125044e-01 -1.85651863e+00
4.76869464e-01 3.98490578e-01 -4.90050167e-01 1.12885803e-01
-1.20317921e-01 1.42401475e-02 -1.13088889e-02 4.30444717e-01
-5.19972801e-01 9.89192188e-01 9.72332478e-01 -1.60847858e-01
-2.98232257e-01 2.62583755e-02 -4.48884130e-01 8.21477115e-01
8.58342052e-01 -3.86972606e-01 -4.77737755e-01 -6.79893672e-01
5.84143959e-02 1.72474623e-01 6.31875813e-01 -4.09289509e-01
2.39042819e-01 -4.41176713e-01 5.41261971e-01 -6.10254109e-01
3.38458389e-01 -4.94440943e-01 2.22861007e-01 2.54050732e-01
-4.98951942e-01 2.74335239e-02 -1.25902176e-01 8.52702498e-01
9.52548012e-02 -4.95382324e-02 3.94001871e-01 5.79527989e-02
-8.05064976e-01 8.94185126e-01 -5.72590709e-01 -4.12441455e-02
7.05341876e-01 -2.03219965e-01 -1.74434975e-01 -7.37458527e-01
-1.05626369e+00 1.61637247e-01 4.63994145e-01 5.44645905e-01
8.30649614e-01 -1.54382861e+00 -6.46852374e-01 5.13754427e-01
-9.03541669e-02 -2.48800129e-01 3.91298950e-01 6.93723023e-01
-1.80974141e-01 6.43782556e-01 -9.66080129e-02 -7.80896783e-01
-9.05066729e-01 7.34792650e-01 2.87991762e-01 2.96051830e-01
-9.15365696e-01 1.12868214e+00 5.73829532e-01 -4.62725341e-01
7.97766522e-02 4.70622927e-02 -7.35951960e-02 -4.42145690e-02
2.38266498e-01 4.29553360e-01 -8.99387419e-01 -7.86258996e-01
-3.90586853e-01 2.20888212e-01 2.48136476e-01 -2.72069454e-01
1.39324951e+00 -3.64547610e-01 -2.69650549e-01 1.15873027e+00
1.53987479e+00 -1.61018968e-01 -1.29380834e+00 -5.72661042e-01
-1.05294120e-02 -1.80429146e-01 5.40607452e-01 -3.13395001e-02
-1.06166267e+00 1.17723334e+00 5.42160571e-01 9.01922956e-02
6.10523224e-01 3.63773108e-01 1.75648883e-01 3.36200982e-01
1.62255228e-01 -9.13870037e-01 -8.61010477e-02 6.97873116e-01
9.23415482e-01 -1.17922127e+00 5.94868213e-02 -1.07841723e-01
-4.28428859e-01 1.21391392e+00 3.21754962e-01 -4.83391374e-01
1.06597412e+00 -1.12138532e-01 4.96590473e-02 -5.87334096e-01
-7.39701629e-01 1.91979960e-01 7.74349749e-01 6.47019684e-01
5.93883574e-01 5.79131424e-01 1.73961580e-01 5.32946289e-01
-2.17187285e-01 -3.77376854e-01 4.79463339e-01 1.69546828e-01
-1.13827795e-01 -8.18614066e-01 -3.48712236e-01 5.60630381e-01
-7.60018826e-02 -2.28114009e-01 -3.66558999e-01 4.31872100e-01
-5.30541718e-01 6.36312306e-01 7.13936985e-02 -1.55632332e-01
-5.43221295e-01 1.95146993e-01 2.91579157e-01 -4.93279397e-01
7.24838316e-01 2.45678559e-01 -4.92861509e-01 -5.70217192e-01
-2.51806170e-01 -1.00230086e+00 -1.21241832e+00 -2.37493113e-01
-2.35160276e-01 3.46547127e-01 5.19989669e-01 9.07501996e-01
2.60307819e-01 5.02512574e-01 4.07350004e-01 -6.86184466e-01
-4.87782508e-01 -9.17931616e-01 -1.06876028e+00 1.67767569e-01
3.63254130e-01 -9.27654624e-01 -6.47175133e-01 1.95451528e-01] | [7.928440570831299, 4.72159481048584] |
bde5281e-3634-411b-80de-03e29862262b | sega-structural-entropy-guided-anchor-view | 2305.04501 | null | https://arxiv.org/abs/2305.04501v2 | https://arxiv.org/pdf/2305.04501v2.pdf | SEGA: Structural Entropy Guided Anchor View for Graph Contrastive Learning | In contrastive learning, the choice of ``view'' controls the information that the representation captures and influences the performance of the model. However, leading graph contrastive learning methods generally produce views via random corruption or learning, which could lead to the loss of essential information and alteration of semantic information. An anchor view that maintains the essential information of input graphs for contrastive learning has been hardly investigated. In this paper, based on the theory of graph information bottleneck, we deduce the definition of this anchor view; put differently, \textit{the anchor view with essential information of input graph is supposed to have the minimal structural uncertainty}. Furthermore, guided by structural entropy, we implement the anchor view, termed \textbf{SEGA}, for graph contrastive learning. We extensively validate the proposed anchor view on various benchmarks regarding graph classification under unsupervised, semi-supervised, and transfer learning and achieve significant performance boosts compared to the state-of-the-art methods. | ['Ke Xu', 'Shangzhe Li', 'Bowen Shi', 'Xueyuan Chen', 'Junran Wu'] | 2023-05-08 | null | null | null | null | ['graph-classification'] | ['graphs'] | [ 4.28847820e-01 7.73968816e-01 -3.85095179e-01 -1.39511928e-01
-2.77554482e-01 -5.91308415e-01 6.70271814e-01 5.56015790e-01
5.24469167e-02 5.91328740e-01 6.20282814e-02 -1.19259723e-01
-4.24375176e-01 -8.84134114e-01 -8.01599860e-01 -9.99278307e-01
-1.91861942e-01 2.49808148e-01 4.91013601e-02 -2.54236553e-02
1.85888708e-01 2.07380056e-01 -1.30658293e+00 1.53418913e-01
9.29603159e-01 1.03459263e+00 9.28994194e-02 2.41648063e-01
-1.03671849e-01 9.85515356e-01 -4.59928125e-01 -5.95775306e-01
2.79577583e-01 -7.06979215e-01 -9.01745379e-01 1.69709980e-01
5.25318563e-01 1.65637374e-01 -3.42333585e-01 1.50631344e+00
2.56621867e-01 -3.11774015e-02 8.24366271e-01 -1.42453074e+00
-5.67291975e-01 8.42004299e-01 -5.54622233e-01 2.25301266e-01
2.39182636e-01 2.70070639e-02 1.35280025e+00 -7.05795050e-01
9.59328175e-01 1.20335770e+00 3.98094475e-01 3.64802361e-01
-1.46384108e+00 -3.18129331e-01 6.77357256e-01 1.86953351e-01
-1.18512428e+00 -3.21676731e-01 1.46747255e+00 -4.25359190e-01
4.14761096e-01 3.83399606e-01 8.19739938e-01 1.05238163e+00
4.22866702e-01 8.84319186e-01 1.50018775e+00 -4.77575570e-01
3.47936898e-01 2.89947450e-01 2.30024293e-01 1.23741829e+00
6.26724899e-01 7.66672418e-02 -7.07457185e-01 1.21362973e-02
1.77840874e-01 -1.43878609e-01 -6.05271995e-01 -9.79826629e-01
-7.95623720e-01 5.97284615e-01 6.84516966e-01 1.28944471e-01
-2.22953603e-01 -3.20113301e-02 3.51159871e-01 6.64843678e-01
6.57853723e-01 2.29575992e-01 -1.05812401e-01 4.07144904e-01
-6.43159866e-01 -1.69749543e-01 8.68745744e-01 9.32451129e-01
8.65915477e-01 1.05223238e-01 -5.92428213e-03 4.86775577e-01
4.73412663e-01 2.96176285e-01 6.63846955e-02 -4.37948704e-01
7.55044937e-01 8.56997252e-01 -5.98934591e-01 -1.12207794e+00
-3.41417819e-01 -9.69809532e-01 -1.19169903e+00 2.26929694e-01
2.96675801e-01 5.72607256e-02 -9.07149255e-01 1.75484443e+00
6.83818907e-02 -9.42728147e-02 2.08811119e-01 6.27497375e-01
9.40051913e-01 2.81352103e-01 -2.01132581e-01 -5.89679956e-01
9.86753166e-01 -9.03002858e-01 -7.43275106e-01 -2.95022309e-01
4.51948017e-01 -2.46606186e-01 1.05295813e+00 5.79508901e-01
-9.13266897e-01 -4.06151444e-01 -1.44197154e+00 4.65973198e-01
-3.92476916e-01 -2.84119636e-01 4.05762494e-01 6.33237660e-01
-9.18360651e-01 8.49318564e-01 -4.75174606e-01 -1.79036856e-01
6.56697631e-01 3.04679722e-02 -5.58618307e-01 -1.95597470e-01
-1.09966719e+00 7.10014522e-01 7.53373623e-01 1.04085140e-01
-9.33404148e-01 -3.81019592e-01 -8.28924716e-01 9.27342549e-02
8.16725671e-01 -6.81622088e-01 4.77580339e-01 -1.05384088e+00
-1.24252629e+00 7.88515031e-01 1.97793260e-01 -3.53440970e-01
8.64191711e-01 4.71816696e-02 -4.69449759e-02 3.51608723e-01
-2.64670163e-01 4.63228464e-01 1.31114030e+00 -1.84273934e+00
-1.57641470e-01 -5.34184217e-01 2.92444229e-01 4.39260870e-01
-3.06766003e-01 -7.80810952e-01 -2.98069358e-01 -6.76583529e-01
5.16179442e-01 -1.09356916e+00 3.96196768e-02 -2.19275072e-01
-7.19536841e-01 -1.65061235e-01 7.52563715e-01 -5.26320577e-01
1.23345947e+00 -2.03675556e+00 6.49484158e-01 4.70226884e-01
7.50461459e-01 -6.61420226e-02 1.63298279e-01 5.16787708e-01
-2.32421428e-01 3.86417866e-01 -2.39921257e-01 -1.37635231e-01
2.06482131e-02 1.32700577e-01 -9.80727822e-02 6.21654749e-01
2.56724041e-02 8.06304216e-01 -1.11593795e+00 -6.74006462e-01
2.30019078e-01 1.36744291e-01 -5.63265860e-01 1.96069509e-01
-1.08833641e-01 5.36577344e-01 -5.01791000e-01 5.30210853e-01
7.13994622e-01 -5.83277404e-01 5.76335251e-01 -4.23180282e-01
4.06948924e-01 8.96519199e-02 -1.11372900e+00 1.54207289e+00
-8.78911167e-02 2.70995229e-01 -6.62587732e-02 -1.15547597e+00
9.80824053e-01 -4.68564443e-02 3.08129668e-01 -5.53107202e-01
-9.11671445e-02 -9.46457088e-02 3.42461318e-02 -1.18833981e-01
2.26689443e-01 8.48072693e-02 1.36839673e-01 4.08732682e-01
2.51715124e-01 -1.43365636e-01 -4.52313535e-02 4.60702419e-01
1.01981878e+00 1.83785215e-01 6.33815587e-01 -4.48935747e-01
5.84836841e-01 -4.64282393e-01 3.52705210e-01 8.03385973e-01
-2.69999802e-01 3.13108742e-01 1.06477976e+00 -2.53260821e-01
-5.40491104e-01 -1.29664278e+00 1.85100615e-01 8.32894921e-01
2.94754177e-01 -6.32653594e-01 -7.86936879e-01 -1.38611877e+00
-2.30266288e-01 7.19761431e-01 -6.19942427e-01 -4.95957911e-01
-4.03182089e-01 -5.90147853e-01 1.90218955e-01 6.05676845e-02
6.37020946e-01 -8.13069940e-01 -2.22038463e-01 -1.48731142e-01
-4.62747924e-02 -7.87382364e-01 -2.88307756e-01 4.35263544e-01
-9.37642634e-01 -1.34789824e+00 -2.73088545e-01 -7.16757476e-01
1.09227300e+00 2.88886249e-01 1.18652487e+00 2.22887382e-01
3.00595999e-01 3.31135273e-01 -4.83827740e-01 -1.44867688e-01
-6.02215230e-01 1.59612745e-01 -3.60652506e-01 4.35597599e-02
-2.13974342e-01 -7.18792617e-01 -6.00745976e-01 -2.13906318e-02
-8.87284040e-01 3.69958699e-01 6.39833152e-01 9.18111086e-01
6.69489026e-01 3.72140765e-01 5.35842359e-01 -1.48789668e+00
5.14367759e-01 -3.13270867e-01 -4.80364233e-01 4.39603031e-01
-1.30398273e+00 4.80759710e-01 9.19625878e-01 -4.75296425e-03
-9.35858309e-01 -1.76572323e-01 2.62070507e-01 -5.05292594e-01
2.30127811e-01 8.27290416e-01 -5.83768249e-01 -1.37864172e-01
5.42160749e-01 3.83650273e-01 4.53318171e-02 -2.41416544e-01
5.34711719e-01 1.68749839e-01 1.01452164e-01 -4.74480927e-01
8.77034724e-01 4.19835001e-01 4.14084047e-01 -6.42420888e-01
-8.94804239e-01 -8.60650558e-03 -8.28135252e-01 -5.80005109e-01
4.49597895e-01 -5.91878474e-01 -4.50731069e-01 3.98991317e-01
-7.29645073e-01 2.81156856e-03 -2.64771283e-01 1.47316635e-01
-6.21074796e-01 5.83180308e-01 -2.95836300e-01 -6.63085580e-01
-3.28441411e-01 -1.07605243e+00 7.15932310e-01 -1.68895051e-01
1.22389771e-01 -1.33333850e+00 -1.51105449e-01 2.84690470e-01
-9.05136392e-03 5.93042552e-01 1.40586293e+00 -8.59258235e-01
-6.58534825e-01 -3.40959281e-02 -3.35688889e-02 4.71734881e-01
2.00246021e-01 -2.60803074e-01 -1.11464739e+00 -7.13168144e-01
-4.86043170e-02 -3.42227340e-01 1.12986147e+00 1.75287530e-01
1.12295651e+00 -6.54829919e-01 -3.66430163e-01 3.88645262e-01
1.56258798e+00 3.70813487e-03 5.89653134e-01 4.80468497e-02
1.03693390e+00 8.22842479e-01 5.25026500e-01 2.34291926e-01
3.82815182e-01 3.58553559e-01 7.62670338e-01 2.14006931e-01
-1.35736525e-01 -7.20628440e-01 3.26836973e-01 1.22389042e+00
-4.48290668e-02 -7.04885185e-01 -7.18386412e-01 1.28962845e-01
-1.78214037e+00 -7.88690686e-01 2.36697748e-01 2.42043614e+00
4.67270285e-01 6.72410667e-01 -2.39052162e-01 3.86104017e-01
8.12338829e-01 9.36150789e-01 -6.16654813e-01 -1.25841215e-01
-1.51227459e-01 -1.33246049e-01 3.95055890e-01 5.44044793e-01
-1.12221479e+00 7.58372188e-01 5.55828476e+00 7.96076119e-01
-1.01361334e+00 -4.39950228e-02 7.02439368e-01 3.14356595e-01
-7.00296223e-01 1.92819282e-01 -2.91666448e-01 4.13371176e-01
5.79130530e-01 -1.98371813e-01 4.58252102e-01 6.26323342e-01
-1.60189029e-02 1.33439273e-01 -1.37020218e+00 8.42832625e-01
2.91022480e-01 -1.13163269e+00 4.21178758e-01 2.11372331e-01
7.93218613e-01 -3.29478532e-01 1.28598675e-01 2.70559549e-01
2.88420111e-01 -8.49994421e-01 7.74749279e-01 7.03474045e-01
6.03161335e-01 -7.74493337e-01 6.83735013e-01 4.73133951e-01
-1.35633647e+00 1.29027158e-01 -1.07051440e-01 2.61458784e-01
-3.09686661e-01 6.51941180e-01 -6.65797532e-01 1.13808465e+00
3.82288903e-01 9.91954386e-01 -1.02910626e+00 4.01338160e-01
-4.25828934e-01 7.19337046e-01 1.72242820e-01 6.69323355e-02
5.41335624e-03 -4.16598767e-01 8.91936004e-01 8.57126236e-01
-1.04391426e-01 -2.88237154e-01 4.30128574e-01 6.88570857e-01
-4.02484000e-01 1.82957217e-01 -9.05027330e-01 -1.25642210e-01
5.29122114e-01 1.09797692e+00 -8.96194458e-01 -3.00232053e-01
-3.10706943e-01 7.31805980e-01 6.41111374e-01 2.66477883e-01
-5.92349410e-01 -5.74568696e-02 -1.76757157e-01 2.83149909e-02
2.24868819e-01 1.55465841e-01 1.23652746e-03 -1.17942643e+00
3.78665894e-01 -1.06906307e+00 7.00535476e-01 -4.24756885e-01
-1.34902477e+00 5.14851570e-01 2.28437647e-01 -1.42746115e+00
-1.02579296e-01 -4.22743559e-01 -4.57052350e-01 3.81030709e-01
-1.43357253e+00 -1.15142155e+00 -4.38591570e-01 3.92618239e-01
3.35361719e-01 -2.24874109e-01 5.20415604e-01 -6.20433129e-02
-4.24482793e-01 7.21748412e-01 4.61921021e-02 -8.62413496e-02
3.98903400e-01 -1.64316547e+00 1.67453751e-01 6.87553227e-01
2.78557748e-01 3.73302519e-01 7.26856947e-01 -6.98421180e-01
-1.59241736e+00 -1.25318265e+00 3.63158762e-01 -1.45192817e-01
5.73135853e-01 -3.26831967e-01 -9.42824006e-01 7.52273917e-01
2.55099893e-01 1.53008789e-01 4.57481802e-01 -3.76389548e-02
-5.32009959e-01 -3.43583852e-01 -1.24091125e+00 6.27631307e-01
1.33624101e+00 -6.33966506e-01 -5.61180055e-01 3.26064289e-01
7.56259561e-01 -2.04925984e-01 -9.49347258e-01 4.95397568e-01
3.81880403e-01 -1.09770656e+00 6.93448067e-01 -5.44586539e-01
2.62504727e-01 -1.08970903e-01 -1.02078594e-01 -1.69933331e+00
-3.69012356e-01 -5.16855776e-01 -4.54643548e-01 1.28558648e+00
3.46844047e-01 -8.67135346e-01 9.05718684e-01 -4.68811914e-02
-1.20847791e-01 -8.15000474e-01 -7.83503473e-01 -7.99178481e-01
1.73745472e-02 1.27091957e-03 3.59146953e-01 1.05620313e+00
1.82287004e-02 5.33556879e-01 -3.23694974e-01 3.56994867e-02
9.92013574e-01 1.74826577e-01 5.84171534e-01 -1.45301592e+00
-2.88884640e-01 -4.75866318e-01 -6.49861336e-01 -7.99643278e-01
4.00345385e-01 -1.38287461e+00 -2.29964584e-01 -1.45677543e+00
3.69518578e-01 -2.01022789e-01 -6.24662340e-01 1.74072787e-01
-2.64896244e-01 -1.57541722e-01 3.60651076e-01 3.31074625e-01
-7.97527373e-01 6.12476885e-01 1.45002329e+00 -4.14763868e-01
-6.21624216e-02 -1.00517549e-01 -7.54007936e-01 7.19459891e-01
7.02420950e-01 -4.61959004e-01 -8.51389050e-01 9.02278721e-02
4.31755871e-01 3.73658934e-03 2.71677673e-01 -8.07581007e-01
-2.31136102e-02 -3.73139940e-02 2.85183996e-01 -4.47902501e-01
5.88763542e-02 -9.16241348e-01 1.86216503e-01 7.55935192e-01
-6.17672026e-01 2.10614898e-03 -2.71888494e-01 1.03035760e+00
-1.36482492e-01 -3.47711951e-01 7.53627360e-01 -3.48869592e-01
-3.76305789e-01 3.85430545e-01 -1.13290995e-02 4.08139527e-01
8.26446474e-01 -1.09089665e-01 -6.04001760e-01 -4.30273175e-01
-7.77306080e-01 -2.75101569e-02 4.53916013e-01 3.06825161e-01
6.54199064e-01 -1.41568148e+00 -5.55347919e-01 6.17688522e-02
4.32012975e-01 -9.99752339e-03 1.16059363e-01 6.96108103e-01
-3.98966342e-01 2.11228564e-01 -9.51477289e-02 -6.69290364e-01
-1.17720366e+00 7.26245522e-01 3.06492209e-01 -6.48986220e-01
-7.89346516e-01 3.92938882e-01 2.64447570e-01 -3.21319312e-01
1.39236704e-01 -1.90429688e-01 -4.00068134e-01 3.39749873e-01
-4.05190103e-02 4.28541720e-01 1.27692834e-01 -5.83231509e-01
-2.44755536e-01 4.03440356e-01 -1.91132948e-01 2.22587243e-01
1.11123514e+00 -3.18473518e-01 -2.08486810e-01 6.33834064e-01
1.15870738e+00 6.44704700e-02 -1.42110336e+00 -3.67949277e-01
1.55186802e-01 -3.51749778e-01 1.62304178e-01 -5.83307207e-01
-1.22457576e+00 7.82128870e-01 6.35779202e-01 5.70075691e-01
9.42348599e-01 6.27561584e-02 2.86153257e-02 4.87858593e-01
5.47783375e-01 -8.75613570e-01 3.42106998e-01 7.24528059e-02
1.02434778e+00 -1.30958652e+00 3.25041622e-01 -6.02659047e-01
-5.46344161e-01 9.81103539e-01 5.78055084e-01 -4.60429005e-02
7.80828416e-01 -1.27610251e-01 -2.54566967e-01 -5.62340677e-01
-8.95861447e-01 -3.63168009e-02 5.79364479e-01 5.68788946e-01
9.02518257e-02 2.04634011e-01 -1.46583736e-01 2.34893069e-01
-2.03359127e-01 -4.44412947e-01 3.97924989e-01 8.53665471e-01
-3.68862987e-01 -6.86145782e-01 7.58963227e-02 5.99246204e-01
-2.51569211e-01 -3.49231586e-02 -7.30806053e-01 9.74816918e-01
-5.85643649e-02 8.66459131e-01 -2.94647783e-01 -4.43638951e-01
1.52911648e-01 1.59182757e-01 7.33568370e-01 -5.83124161e-01
-2.99429893e-01 -7.24989623e-02 1.56194389e-01 -3.34000081e-01
-3.91443759e-01 -4.04458821e-01 -1.08332157e+00 -2.48043984e-01
-4.98820573e-01 2.38197014e-01 2.28547603e-01 1.05881238e+00
1.00633822e-01 7.44920731e-01 8.32933307e-01 -3.80934298e-01
-6.59340560e-01 -8.66881549e-01 -6.52453423e-01 4.92109507e-01
2.81303674e-01 -7.12377369e-01 -7.30892062e-01 -1.47574350e-01] | [7.363410949707031, 6.2198686599731445] |
2d581471-772a-41fa-bbe8-534b36889c80 | gpv-pose-category-level-object-pose | 2203.07918 | null | https://arxiv.org/abs/2203.07918v2 | https://arxiv.org/pdf/2203.07918v2.pdf | GPV-Pose: Category-level Object Pose Estimation via Geometry-guided Point-wise Voting | While 6D object pose estimation has recently made a huge leap forward, most methods can still only handle a single or a handful of different objects, which limits their applications. To circumvent this problem, category-level object pose estimation has recently been revamped, which aims at predicting the 6D pose as well as the 3D metric size for previously unseen instances from a given set of object classes. This is, however, a much more challenging task due to severe intra-class shape variations. To address this issue, we propose GPV-Pose, a novel framework for robust category-level pose estimation, harnessing geometric insights to enhance the learning of category-level pose-sensitive features. First, we introduce a decoupled confidence-driven rotation representation, which allows geometry-aware recovery of the associated rotation matrix. Second, we propose a novel geometry-guided point-wise voting paradigm for robust retrieval of the 3D object bounding box. Finally, leveraging these different output streams, we can enforce several geometric consistency terms, further increasing performance, especially for non-symmetric categories. GPV-Pose produces superior results to state-of-the-art competitors on common public benchmarks, whilst almost achieving real-time inference speed at 20 FPS. | ['Federico Tombari', 'Nassir Navab', 'Xiangyang Ji', 'Fabian Manhardt', 'Zhiqiang Lou', 'Ruida Zhang', 'Yan Di'] | 2022-03-15 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Di_GPV-Pose_Category-Level_Object_Pose_Estimation_via_Geometry-Guided_Point-Wise_Voting_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Di_GPV-Pose_Category-Level_Object_Pose_Estimation_via_Geometry-Guided_Point-Wise_Voting_CVPR_2022_paper.pdf | cvpr-2022-1 | ['6d-pose-estimation-using-rgbd', '6d-pose-estimation-1', '6d-pose-estimation'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 9.77787841e-03 -3.88928145e-01 -2.38091871e-01 -3.69382828e-01
-1.11043942e+00 -8.90621901e-01 5.90186179e-01 3.33797455e-01
-2.58863419e-01 1.12073950e-01 -4.42281775e-02 1.69105101e-02
-1.85940146e-01 -5.02767265e-01 -8.22893023e-01 -7.57507205e-01
8.74564573e-02 8.01527143e-01 4.07837391e-01 9.42881107e-02
5.43914914e-01 8.93673658e-01 -1.71929026e+00 -3.53013910e-02
5.39335728e-01 1.42387033e+00 9.24078599e-02 4.31823373e-01
1.47004455e-01 7.45183751e-02 -3.82240295e-01 -3.34954977e-01
3.72940183e-01 2.67941326e-01 -4.94685650e-01 1.72417223e-01
1.05209708e+00 -4.45417792e-01 -1.15179136e-01 7.70637214e-01
3.98848772e-01 8.93522501e-02 8.27078223e-01 -1.17999494e+00
-1.31239489e-01 -2.38818005e-02 -6.52822912e-01 2.43024714e-02
4.46735710e-01 1.70948803e-01 1.09628415e+00 -1.10680962e+00
6.88009918e-01 1.35503995e+00 4.94848371e-01 1.54073164e-01
-1.29578865e+00 -6.58950269e-01 4.46111500e-01 1.01409756e-01
-1.53276622e+00 -1.31006077e-01 7.62853801e-01 -4.45831954e-01
8.44846487e-01 2.56815434e-01 6.04149997e-01 8.33772361e-01
6.09339960e-02 8.21788013e-01 8.86327267e-01 -4.05594371e-02
1.37102410e-01 -2.11079806e-01 -1.56861156e-01 4.58153307e-01
4.05203223e-01 -2.01993808e-01 -5.98600984e-01 -1.11811586e-01
6.45795822e-01 1.40438706e-01 -1.54746950e-01 -1.24083769e+00
-1.46113193e+00 5.78664482e-01 6.92970455e-01 -2.24770278e-01
-2.39666343e-01 1.58072904e-01 2.59599000e-01 -6.56455904e-02
4.75432992e-01 3.98951620e-01 -5.99739254e-01 -1.90623611e-01
-6.54475033e-01 7.26820588e-01 5.98079801e-01 9.89149213e-01
5.94063461e-01 -4.46827024e-01 -3.08568805e-01 6.74825788e-01
5.92614770e-01 8.12596738e-01 -1.60743594e-01 -7.36551464e-01
7.09780872e-01 8.16601336e-01 1.15078963e-01 -1.17491233e+00
-4.87066895e-01 -5.98209023e-01 -5.67401528e-01 9.73625779e-02
6.55129671e-01 6.58327460e-01 -8.89425933e-01 1.60207844e+00
9.03104126e-01 5.97962998e-02 -4.14659619e-01 1.17584503e+00
6.18093431e-01 4.05740321e-01 -6.36686534e-02 1.43571988e-01
1.56269050e+00 -7.01896608e-01 1.18486837e-01 -2.85398662e-01
3.56185228e-01 -8.37962031e-01 6.39519155e-01 5.53728044e-01
-8.65802169e-01 -3.26956570e-01 -9.55052435e-01 -2.14022875e-01
-3.30116332e-01 1.08847275e-01 6.70122981e-01 5.69853365e-01
-4.18039024e-01 3.27896476e-01 -8.95922244e-01 -1.24891154e-01
7.05945730e-01 4.30708468e-01 -5.26368856e-01 -3.96426022e-01
-5.34329414e-01 9.11781013e-01 3.04284632e-01 2.30460867e-01
-6.74729824e-01 -9.62803066e-01 -8.62129152e-01 -1.74006432e-01
6.96124196e-01 -6.42512381e-01 1.06516659e+00 -4.12836820e-02
-1.19647241e+00 1.01655304e+00 -2.05128919e-02 -1.89475507e-01
8.29191446e-01 -4.30478066e-01 8.30557272e-02 1.20570667e-01
7.30017498e-02 7.48120070e-01 1.09252548e+00 -1.25708103e+00
-5.76744616e-01 -9.97281909e-01 -8.26473981e-02 4.16064203e-01
-1.22590780e-01 -2.89713353e-01 -9.37362432e-01 -6.02103472e-01
7.16978073e-01 -1.13941717e+00 -3.41472104e-02 5.93306541e-01
-2.33122095e-01 -4.00274694e-01 8.83185506e-01 -2.89943188e-01
8.02829802e-01 -2.22503424e+00 3.42663884e-01 2.23402858e-01
1.69660196e-01 2.47760534e-01 2.54139416e-02 6.22000210e-02
2.98810810e-01 -1.73027575e-01 -1.16916023e-01 -5.56794405e-01
1.84865221e-01 9.90647003e-02 -3.91702056e-01 7.25777328e-01
4.50324893e-01 9.02380049e-01 -8.78393829e-01 -3.89959872e-01
4.84713107e-01 6.35441899e-01 -7.88178444e-01 8.34654942e-02
-2.61798888e-01 4.06480610e-01 -6.01226449e-01 9.40694273e-01
1.04826570e+00 -2.60315299e-01 -3.85532454e-02 -3.94293487e-01
1.31534025e-01 1.30967721e-01 -1.52885938e+00 1.88922548e+00
-3.99139941e-01 2.33647138e-01 -7.23195598e-02 -8.61186564e-01
9.94851053e-01 -1.73055291e-01 5.69722116e-01 -4.24049646e-01
2.55803853e-01 3.59833509e-01 -3.76738131e-01 -1.37962908e-01
4.76391166e-01 1.58895835e-01 -1.63301423e-01 2.57564127e-01
-8.85232762e-02 -6.70259774e-01 -4.92153689e-02 2.44647227e-02
8.23906481e-01 4.48454440e-01 2.92722017e-01 -4.61385287e-02
3.55599165e-01 -2.06450865e-01 3.99981052e-01 6.04773819e-01
-3.03898752e-01 1.07466161e+00 3.67546171e-01 -4.00231093e-01
-8.74978602e-01 -1.23061049e+00 -4.87274885e-01 8.81575584e-01
5.11056006e-01 -3.72696280e-01 -5.14050186e-01 -7.59470761e-01
6.22754216e-01 2.53059745e-01 -5.52660584e-01 -1.55778527e-01
-6.90234542e-01 -6.52421713e-01 -7.40886852e-03 5.85851192e-01
1.20603018e-01 -4.83984828e-01 -8.65953445e-01 1.58325836e-01
-9.65424180e-02 -1.41093600e+00 -3.88556421e-01 1.96498230e-01
-1.03611052e+00 -1.02807486e+00 -6.74286008e-01 -4.47057843e-01
5.74199140e-01 5.99014997e-01 1.14545572e+00 -1.22090682e-01
-4.77628618e-01 3.34130943e-01 -3.58061641e-01 -3.82455409e-01
1.87671959e-01 1.98603034e-01 1.65042281e-02 5.69506288e-02
3.30836803e-01 -4.31338787e-01 -9.52463925e-01 6.20545745e-01
-6.49066091e-01 -1.34908855e-01 6.54093027e-01 5.43567359e-01
9.58903968e-01 -3.13576907e-01 1.72739863e-01 -3.88420522e-01
-1.24162629e-01 -3.02043825e-01 -7.78153956e-01 6.30674437e-02
-4.73509282e-01 1.16787754e-01 2.05076322e-01 -4.79357243e-01
-6.30945146e-01 4.06213790e-01 5.89055791e-02 -6.78701818e-01
-6.97721466e-02 7.30469748e-02 -3.09758902e-01 -2.18004242e-01
4.23983961e-01 1.60757601e-02 8.16865340e-02 -5.06167650e-01
4.06375915e-01 5.26992321e-01 6.47417903e-01 -6.25794232e-01
1.00944602e+00 7.39937842e-01 3.42581242e-01 -6.11496568e-01
-1.05928147e+00 -8.47499013e-01 -8.20378244e-01 -2.91891515e-01
7.21874416e-01 -1.05294156e+00 -7.79033482e-01 5.25076866e-01
-1.04083395e+00 5.68670891e-02 -1.14698052e-01 5.12678981e-01
-5.55221617e-01 2.53183216e-01 -4.02918868e-02 -6.01907670e-01
-2.23944604e-01 -1.28203714e+00 1.88541174e+00 -1.85081661e-01
-2.19172202e-02 -3.63794148e-01 -1.17399514e-01 6.57484174e-01
8.43457058e-02 4.83573377e-01 6.85182571e-01 -5.12044370e-01
-1.05196249e+00 -6.09245658e-01 -4.87091482e-01 1.06231965e-01
-2.35579878e-01 2.28815563e-02 -8.42021823e-01 -4.50001419e-01
-1.85107261e-01 -3.69985640e-01 7.67128646e-01 1.88076630e-01
1.29778409e+00 2.18853086e-01 -4.78583038e-01 7.52962232e-01
1.26451147e+00 -4.76292104e-01 2.92452663e-01 2.52438068e-01
7.68401384e-01 4.41658378e-01 1.12165821e+00 6.20104969e-01
5.10136664e-01 1.09930301e+00 8.93504918e-01 3.89039636e-01
-1.18019991e-01 -2.53715694e-01 -1.70050785e-02 3.64473224e-01
-6.48636594e-02 -1.31284401e-01 -8.69994760e-01 3.12806249e-01
-1.60996068e+00 -4.77126598e-01 -2.42620662e-01 2.51500320e+00
4.79390472e-01 3.17411989e-01 1.89949512e-01 9.75069329e-02
4.41047072e-01 2.39968315e-01 -8.14597428e-01 3.68507951e-01
1.28741041e-01 1.06366105e-01 4.81469959e-01 6.95285127e-02
-1.25910568e+00 7.64281571e-01 5.20851898e+00 8.57322097e-01
-1.19256854e+00 -1.75495341e-01 4.08834219e-01 -1.57692730e-01
-3.25331613e-02 -4.82276045e-02 -1.16478002e+00 3.32865715e-01
2.59744078e-01 3.51758987e-01 4.50379290e-02 1.07006097e+00
-1.67790458e-01 -2.24479735e-01 -1.21663046e+00 1.22647667e+00
2.96677649e-01 -1.00489652e+00 4.70334962e-02 1.98350430e-01
5.86218596e-01 1.00490451e-01 1.07374534e-01 2.54540145e-01
-2.34730765e-01 -8.67498457e-01 1.03529942e+00 2.84842312e-01
7.93078780e-01 -8.88646066e-01 5.48992693e-01 3.02300751e-01
-1.36790895e+00 -5.29756956e-02 -3.69851708e-01 1.72069907e-01
-3.27473646e-03 5.83014250e-01 -8.28343570e-01 6.33088827e-01
8.68316412e-01 6.44170821e-01 -7.64158249e-01 1.23150039e+00
-1.26329601e-01 1.91626057e-01 -7.31891870e-01 3.77127193e-02
7.73763731e-02 1.71980754e-01 6.65333748e-01 9.76660669e-01
3.57678503e-01 5.50894849e-02 1.60402358e-01 6.66465700e-01
2.91244243e-03 2.40680110e-02 -2.72858799e-01 3.99508327e-01
6.52121365e-01 1.41910398e+00 -9.69202340e-01 6.65896460e-02
-2.17536360e-01 7.83068478e-01 4.01069194e-01 -7.14937747e-02
-8.95315170e-01 -2.21645925e-03 7.70889759e-01 2.08500311e-01
7.32174993e-01 -5.16716301e-01 -2.60315567e-01 -1.26665223e+00
4.54562813e-01 -7.56557941e-01 3.30138206e-01 -5.14063597e-01
-1.15972853e+00 9.54467431e-02 1.84462532e-01 -1.51044273e+00
-2.46553980e-02 -9.09636796e-01 -8.18039328e-02 4.06760216e-01
-1.58251655e+00 -1.34581995e+00 -5.83514512e-01 3.47952992e-01
4.60253060e-01 1.98276341e-01 5.00919282e-01 2.11192191e-01
-2.33751312e-01 6.27564192e-01 6.55258670e-02 -3.89972366e-02
9.22164857e-01 -1.07682765e+00 3.30077112e-01 4.98966306e-01
3.43196601e-01 4.49814856e-01 6.16430581e-01 -2.92594135e-01
-1.99686992e+00 -1.08963811e+00 6.12546206e-01 -9.56229270e-01
3.96815211e-01 -7.30948389e-01 -8.55453789e-01 2.98976690e-01
-7.98272312e-01 5.86763918e-01 2.46500418e-01 5.54918610e-02
-8.36671233e-01 -3.23843658e-01 -1.01100326e+00 3.72439176e-01
1.24956989e+00 -4.21947390e-01 -4.65796977e-01 2.99122930e-01
4.88881797e-01 -7.90428817e-01 -9.16768253e-01 8.02225888e-01
8.02563429e-01 -8.42594326e-01 1.42905211e+00 -2.08746180e-01
1.90924972e-01 -6.62488103e-01 -3.92659485e-01 -7.85601854e-01
-1.25648230e-01 -2.60885894e-01 -4.27769184e-01 1.05375385e+00
1.76048651e-03 -4.15770888e-01 9.17194545e-01 4.43251550e-01
-3.26848775e-02 -9.63824213e-01 -1.10361803e+00 -9.25951481e-01
-7.39181116e-02 -7.28940248e-01 5.81973553e-01 4.83487874e-01
-5.71477234e-01 1.14248820e-01 -1.23222858e-01 3.26413780e-01
7.71582186e-01 5.46452045e-01 1.18713629e+00 -1.40240836e+00
-1.79640830e-01 -5.25994241e-01 -9.18308318e-01 -1.48374665e+00
-1.57593358e-02 -9.55962420e-01 2.18154401e-01 -1.14063287e+00
2.58657873e-01 -6.10233366e-01 -1.21304333e-01 2.35734135e-01
-1.75375894e-01 7.86866307e-01 3.19001198e-01 1.79468110e-01
-8.26909482e-01 5.69275737e-01 1.30631387e+00 -6.15382120e-02
2.34057028e-02 5.79588823e-02 -4.77536440e-01 5.92097223e-01
3.80331755e-01 -4.11644131e-01 -3.13655473e-02 -3.69853973e-01
2.85256982e-01 -2.02866778e-01 7.56260276e-01 -1.17060292e+00
8.17161798e-02 -1.90400779e-01 4.90010202e-01 -1.08106649e+00
5.11601210e-01 -8.93265963e-01 -9.32343379e-02 3.10739636e-01
-4.23428603e-02 -3.58125150e-01 8.12673941e-02 9.07000542e-01
-5.62717952e-02 7.34575242e-02 8.26198697e-01 2.21751094e-01
-6.41293228e-01 6.69785261e-01 1.86405957e-01 8.75073001e-02
1.14029217e+00 -3.55741769e-01 9.73006804e-03 3.81809771e-02
-3.27288598e-01 2.96203017e-01 5.76286376e-01 6.35311484e-01
5.12105644e-01 -1.16270220e+00 -6.64405286e-01 2.43584156e-01
5.68349421e-01 7.33335078e-01 3.17951173e-01 9.61016595e-01
-4.01392460e-01 3.87182295e-01 1.10693969e-01 -1.29872811e+00
-1.31763685e+00 5.23621559e-01 5.15420996e-02 -2.09659487e-01
-5.07287562e-01 9.67573881e-01 1.98150501e-01 -6.08209729e-01
3.01096529e-01 -4.82077956e-01 4.74816784e-02 2.73135811e-01
3.56609613e-01 3.00257981e-01 4.16782230e-01 -7.41322875e-01
-6.57212913e-01 1.08128417e+00 -1.69388041e-01 2.86863863e-01
1.33607757e+00 -8.31657350e-02 1.18680820e-01 2.69431829e-01
1.28067994e+00 -1.06325582e-01 -1.67424607e+00 -2.90629834e-01
-1.26208603e-01 -9.20678914e-01 -5.23925163e-02 -5.99952757e-01
-8.93084228e-01 9.54127014e-01 6.20908380e-01 -2.69182861e-01
8.30989480e-01 2.64573842e-01 5.99627793e-01 4.96158212e-01
6.82880163e-01 -9.72680628e-01 3.49250406e-01 5.48314154e-01
9.29376423e-01 -1.54226708e+00 4.22141165e-01 -5.73751271e-01
-2.45294198e-01 1.13170183e+00 5.33720970e-01 -2.02411100e-01
5.52420795e-01 -5.92658184e-02 -2.54545510e-01 -1.51408762e-01
-4.76426125e-01 -3.05577647e-02 6.95168555e-01 4.50978488e-01
9.56234038e-02 -1.82059244e-03 1.62364036e-01 1.99827448e-01
-1.11517981e-01 -3.68578374e-01 -1.91177845e-01 9.39655602e-01
-4.36312288e-01 -8.75657558e-01 -5.26959419e-01 4.15628821e-01
-3.23965132e-01 3.48167062e-01 -1.75209135e-01 8.34843040e-01
5.74919675e-03 4.61703330e-01 1.22998923e-01 -1.54377222e-01
4.98412013e-01 -1.80839568e-01 7.79008746e-01 -5.61545789e-01
-1.51852623e-01 7.35075697e-02 -3.12479943e-01 -8.24315608e-01
-3.03139746e-01 -9.53800499e-01 -1.00963581e+00 -5.79469875e-02
-4.62793976e-01 -3.96721482e-01 8.37969303e-01 8.57240081e-01
4.26649958e-01 5.36716580e-02 5.30048013e-01 -1.38516045e+00
-7.74416208e-01 -6.65952682e-01 -3.44447225e-01 5.68626046e-01
2.74209559e-01 -1.10092700e+00 -3.83274853e-01 -3.28419149e-01] | [7.480472564697266, -2.638451337814331] |
a78ecc65-1d78-40d1-a83b-406d4adb7544 | k-rater-reliability-the-correct-unit-of | null | null | https://openreview.net/forum?id=35V2c0WUkuZ | https://openreview.net/pdf?id=35V2c0WUkuZ | k-Rater Reliability: The Correct Unit of Reliability for Aggregated Human Annotations | Since the inception of crowdsourcing, aggregation has been a common strategy for dealing with unreliable data. Aggregate ratings are more reliable than individual ones. However, many NLP datasets that rely on aggregate ratings only report the reliability of individual ones, which is the incorrect unit of analysis. In these instances, the data reliability is being under-reported. We present empirical, analytical, and bootstrap-based methods for measuring the reliability of aggregate ratings. We call this k-rater reliability (kRR), a multi-rater extension of inter-rater reliability (IRR). We apply these methods to the widely used word similarity benchmark dataset, WordSim. We conducted two replications of the WordSim dataset to obtain an empirical reference point. We hope this discussion will nudge researchers to report kRR, the correct unit of reliability for aggregate ratings, in addition to IRR. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['word-similarity'] | ['natural-language-processing'] | [-4.42314744e-01 1.32293984e-01 -1.84943795e-01 -3.51513118e-01
-1.14082134e+00 -8.52172613e-01 4.68579888e-01 6.73672557e-01
-6.91619039e-01 9.76226807e-01 4.28044051e-01 -1.19326681e-01
-2.77073625e-02 -5.77598810e-01 -2.83187240e-01 -2.77590960e-01
2.06657015e-02 2.75224715e-01 1.15977779e-01 -2.20897928e-01
6.33880913e-01 1.02325186e-01 -1.53874755e+00 2.77345031e-01
1.09841573e+00 8.95000756e-01 -2.74145424e-01 4.36698943e-01
-9.79612991e-02 8.77533197e-01 -1.30421007e+00 -8.16243649e-01
6.36454448e-02 -6.44617379e-02 -7.36195743e-01 -1.33855075e-01
-2.21466422e-01 -2.94899583e-01 3.14741433e-01 7.68708766e-01
4.75081056e-01 2.88485587e-01 5.91351867e-01 -1.50856078e+00
-1.23430490e+00 8.33044708e-01 -5.91853142e-01 1.76991165e-01
1.10477936e+00 -2.65978366e-01 1.34789217e+00 -1.10518456e+00
4.09335971e-01 1.01497436e+00 5.96434772e-01 1.77240863e-01
-7.90421367e-01 -5.45548141e-01 -1.21278450e-01 1.63374818e-03
-1.61956465e+00 -3.06022167e-01 1.99204504e-01 -4.47789729e-01
7.57562459e-01 1.87245637e-01 2.41756380e-01 8.74124587e-01
-8.46685842e-02 3.84790748e-01 1.54061592e+00 -3.83971989e-01
6.99181139e-01 3.85112822e-01 2.58531094e-01 -3.29712518e-02
8.24700892e-01 -2.57207066e-01 -8.83673847e-01 -8.21567833e-01
5.14435053e-01 -2.52946645e-01 -2.02667236e-01 4.96248960e-01
-9.62553084e-01 8.47190082e-01 -1.67028345e-02 2.40038335e-01
-3.75380129e-01 -2.10736141e-01 2.66532212e-01 4.97144669e-01
1.03218305e+00 5.50966620e-01 -3.96062046e-01 -5.67074776e-01
-8.01024139e-01 2.36865759e-01 1.11351681e+00 8.41681361e-01
5.39401054e-01 -2.66447991e-01 -1.10473886e-01 1.10842741e+00
1.91484228e-01 5.36897838e-01 7.16910362e-01 -1.23887873e+00
4.71395761e-01 6.90841496e-01 9.61548448e-01 -9.42186296e-01
-1.82582915e-01 1.05732568e-01 -3.89558226e-01 1.46504655e-01
3.80220711e-01 -9.06466246e-02 -4.86212283e-01 1.32631302e+00
2.59608120e-01 -3.41628283e-01 -2.97252717e-03 1.13367522e+00
9.59525108e-01 3.73253852e-01 1.78511471e-01 -3.54665518e-01
1.27058315e+00 -4.78012264e-01 -7.69576192e-01 1.51904687e-01
8.47630799e-01 -7.53935158e-01 1.27511179e+00 7.24651635e-01
-9.98444796e-01 -2.49061570e-01 -8.95823359e-01 -1.04595376e-02
-4.43045437e-01 -1.58832669e-01 4.03603405e-01 7.78864384e-01
-1.14037871e+00 4.98448491e-01 -1.94830418e-01 -2.15964526e-01
-1.27096102e-01 3.99569981e-02 -6.57702565e-01 1.40259400e-01
-1.45137012e+00 1.23791730e+00 -1.17889449e-01 -2.74322301e-01
-4.14739460e-01 -4.66037691e-01 -6.16307616e-01 -4.42451954e-01
3.63326699e-01 1.27035649e-02 1.25081587e+00 -6.92988098e-01
-1.11516452e+00 7.73046732e-01 -3.24028969e-01 -8.42106938e-02
2.37835288e-01 -2.47956812e-01 -6.03659093e-01 -2.52934583e-02
3.47211421e-01 2.51561850e-02 3.93051475e-01 -1.34433520e+00
-5.58113158e-01 -3.66976023e-01 1.09881036e-01 1.61781982e-01
-1.27666593e-01 5.18610001e-01 1.82766169e-01 -5.01749873e-01
-2.47860473e-04 -7.47895956e-01 -2.96150218e-03 -4.33103353e-01
-1.42292157e-02 -8.99418592e-01 1.29274027e-02 -7.37473786e-01
1.39093673e+00 -1.96026313e+00 -1.08597919e-01 1.82625115e-01
4.51449096e-01 2.44205490e-01 -1.94008067e-01 7.19911158e-01
1.05009846e-01 6.98255360e-01 -4.05902900e-02 -4.32291090e-01
-2.05983534e-01 4.84519973e-02 -3.33470821e-01 5.60691357e-01
1.04308680e-01 7.36003220e-01 -9.94335175e-01 -5.08038640e-01
-2.76767492e-01 1.50497898e-01 1.24464892e-01 2.67438650e-01
2.17224583e-01 1.69180874e-02 -3.06286961e-01 6.64966464e-01
4.52571720e-01 -2.53057808e-01 -1.97913826e-01 1.32091820e-01
-1.19953640e-01 2.90863395e-01 -1.00192428e+00 1.18648040e+00
-5.94410658e-01 3.65402520e-01 -2.38407433e-01 -3.46828163e-01
1.34051144e+00 5.61571419e-01 4.36852515e-01 -6.72009110e-01
-3.73887606e-02 4.48648900e-01 -2.88139433e-01 -4.44750637e-01
1.05117238e+00 -1.81889888e-02 -2.97529727e-01 9.59102988e-01
-4.04951483e-01 -3.85535210e-01 3.04420799e-01 3.98178279e-01
1.25791192e+00 -2.20627800e-01 6.67877853e-01 -1.28203392e-01
2.10385650e-01 -7.60875195e-02 4.18277621e-01 6.00629210e-01
-4.36621994e-01 8.80980074e-01 7.04146266e-01 -2.82007009e-01
-1.32122183e+00 -9.59869862e-01 2.83507835e-02 9.53915954e-01
1.29354656e-01 -5.48560679e-01 -4.26116645e-01 -7.10668266e-01
1.53762192e-01 8.16692352e-01 -6.99866951e-01 1.23114213e-01
1.69768333e-01 -5.31057715e-01 6.29597187e-01 4.73293245e-01
-2.95912027e-02 -8.61143351e-01 -3.87090683e-01 7.73343071e-02
-4.84348446e-01 -1.25790977e+00 -1.15594663e-01 -2.33351633e-01
-4.46517259e-01 -1.00492501e+00 -7.43909299e-01 1.42658800e-01
5.26765585e-01 5.83133996e-01 1.37372637e+00 5.84394693e-01
3.54857206e-01 3.74124676e-01 -1.17275548e+00 -4.63123143e-01
-3.94680560e-01 -1.00818910e-01 3.95447105e-01 -3.36308867e-01
6.61826789e-01 -4.84047413e-01 -3.92799586e-01 7.48365521e-01
-9.53016818e-01 -7.74112403e-01 2.07303300e-01 1.81375921e-01
2.63173699e-01 -2.51800299e-01 1.28028107e+00 -7.79883623e-01
1.45331454e+00 -9.12514687e-01 -3.22991073e-01 3.31583887e-01
-8.18074346e-01 -1.87966496e-01 5.76563478e-01 -3.73949349e-01
-6.77988589e-01 -6.54730022e-01 2.00239241e-01 1.53908893e-01
-5.64673217e-03 6.83666885e-01 4.44217831e-01 1.02854691e-01
1.00435174e+00 -3.15525115e-01 -9.43667591e-02 -2.62955219e-01
3.36397976e-01 1.31352127e+00 1.81964681e-01 -6.35351002e-01
4.88026142e-01 3.61283690e-01 -5.77600777e-01 -5.54587305e-01
-9.87464547e-01 -6.53239965e-01 -6.43300936e-02 -2.11799264e-01
3.69048029e-01 -1.12514579e+00 -7.99532473e-01 8.68718475e-02
-1.27037036e+00 -1.16162725e-01 -3.33319902e-01 3.85649443e-01
-2.75829136e-01 7.01786637e-01 -5.12081683e-01 -1.45259917e+00
-3.63166392e-01 -9.12681282e-01 1.13447964e+00 2.88252801e-01
-8.35171342e-01 -7.74237514e-01 3.04741114e-01 5.63115478e-01
4.70423311e-01 1.47868946e-01 2.56562650e-01 -1.02566051e+00
4.18843925e-01 -7.93343663e-01 -3.43113601e-01 4.04607624e-01
1.91972554e-01 4.21423733e-01 -8.79872203e-01 -5.57006197e-03
-1.49878925e-02 -6.01057291e-01 3.53450119e-01 -4.98312116e-02
8.61782074e-01 -2.58502007e-01 2.09668338e-01 -6.25505447e-01
1.24151480e+00 -7.07802251e-02 8.02588940e-01 2.19601393e-01
2.16782629e-01 7.71625400e-01 8.33812833e-01 8.93104017e-01
7.40337372e-01 6.01509631e-01 4.03868705e-02 2.51456976e-01
4.46410954e-01 4.00270009e-03 4.18183386e-01 1.03302097e+00
-4.95152861e-01 -3.18572372e-01 -9.17684495e-01 6.54154003e-01
-1.82591140e+00 -6.48530543e-01 -3.28879833e-01 2.54236221e+00
9.22824442e-01 -1.63969055e-01 4.31466639e-01 3.65064204e-01
8.67571950e-01 -1.25426829e-01 -3.97453420e-02 -7.40317225e-01
-1.25250861e-01 -1.22223489e-01 3.56622279e-01 4.83810008e-01
-3.65527570e-01 5.83900332e-01 7.28778839e+00 7.71303952e-01
-3.58919770e-01 4.65520114e-01 5.68237066e-01 -4.36738431e-02
-6.55388653e-01 1.32304803e-01 -5.37106454e-01 8.08372021e-01
1.17024422e+00 -4.02055353e-01 5.74577987e-01 8.21753144e-01
4.59322423e-01 -8.77121747e-01 -8.11045289e-01 7.59563327e-01
4.70950603e-02 -6.51905358e-01 -2.70912588e-01 1.48259416e-01
8.58226955e-01 -2.51893640e-01 -3.56116533e-01 1.09396264e-01
8.83712292e-01 -1.09780765e+00 7.16112912e-01 5.68884432e-01
8.84733915e-01 -6.33574665e-01 1.34401047e+00 4.11094129e-01
-7.67025769e-01 -8.02823678e-02 -6.32367015e-01 -3.65982741e-01
4.54083472e-01 1.49451780e+00 -5.45445621e-01 5.80713689e-01
6.67737246e-01 1.43471912e-01 -5.80362976e-01 8.31345260e-01
-3.78346741e-01 4.49904472e-01 -2.16709286e-01 -3.15366834e-01
-2.84864992e-01 -8.49683061e-02 9.91735086e-02 8.76897156e-01
4.22340393e-01 2.11234882e-01 -1.69230759e-01 5.93756735e-01
-2.20222265e-01 1.97594598e-01 -6.98236406e-01 -1.43131822e-01
1.10588908e+00 1.44222569e+00 -6.07555389e-01 -3.09698820e-01
-5.20268083e-01 5.15054047e-01 6.99831128e-01 2.84726143e-01
-7.30225623e-01 -3.99113953e-01 3.43174100e-01 -2.84654405e-02
-3.80086988e-01 -5.10362506e-01 -6.30096376e-01 -9.50656295e-01
3.42410356e-01 -9.70665455e-01 1.66484520e-01 -1.13443673e+00
-1.63408852e+00 6.21291101e-01 1.06932916e-01 -1.12733209e+00
-2.62599826e-01 -2.94714779e-01 -2.54910558e-01 9.82635856e-01
-1.07745957e+00 -4.58404034e-01 -4.78147209e-01 1.23373210e-01
1.86696023e-01 2.24841107e-02 6.96272016e-01 -2.92278677e-02
-1.65468514e-01 5.69341004e-01 -1.94734737e-01 -5.26477210e-03
1.05018890e+00 -1.10625088e+00 4.19482917e-01 3.89146209e-01
-6.07755147e-02 8.35298598e-01 9.96951461e-01 -9.69060242e-01
-9.20787513e-01 -5.98827481e-01 1.13640451e+00 -9.65849996e-01
9.37643468e-01 -4.56914958e-03 -9.99171913e-01 3.08886230e-01
-6.55104890e-02 -6.61011189e-02 9.42610860e-01 2.53030986e-01
-6.06564164e-01 -3.20315696e-02 -1.32478631e+00 3.92505199e-01
8.00113380e-01 -7.55234718e-01 -6.39245450e-01 3.71771455e-01
9.87490177e-01 -4.27711084e-02 -1.37538981e+00 2.24225074e-01
5.67859292e-01 -1.05732858e+00 3.63927722e-01 -3.15743238e-01
6.35764182e-01 -4.29013669e-01 -2.94382095e-01 -1.49495959e+00
-9.69841555e-02 -3.51515800e-01 -3.54206376e-02 1.49243414e+00
7.26759076e-01 -7.43052781e-01 2.61095256e-01 1.41983891e+00
2.69410193e-01 -7.54694819e-01 -9.78046298e-01 -9.92060900e-01
6.78071082e-02 -6.63697720e-01 7.96604633e-01 8.48458469e-01
4.36565578e-01 -1.55843839e-01 -3.05781603e-01 -1.19600914e-01
3.37609738e-01 -4.45770860e-01 7.10725248e-01 -1.26103961e+00
1.31115556e-01 3.02322786e-02 -3.95921022e-01 -2.91306973e-01
1.09139672e-02 -1.79320231e-01 1.23676993e-01 -1.59865320e+00
1.65846109e-01 -5.25828242e-01 3.35413702e-02 2.39200905e-01
-3.99671257e-01 5.07007182e-01 -3.16997431e-02 4.40035373e-01
-8.31755817e-01 3.62185240e-01 8.90790224e-01 4.02170241e-01
-1.25293478e-01 -8.66228640e-02 -1.01804924e+00 2.84954280e-01
8.83794725e-01 -5.92059791e-01 -2.20824137e-01 -2.90545434e-01
1.08185256e+00 -2.73856856e-02 9.59653556e-02 -6.59106910e-01
3.79701972e-01 -4.30774838e-01 -6.21881746e-02 -2.37644687e-01
1.14777580e-01 -5.65917015e-01 1.68318734e-01 -2.77955115e-01
-5.94500661e-01 5.51186085e-01 -4.26650673e-01 5.64170063e-01
-2.19799161e-01 -6.15369737e-01 2.98298061e-01 -2.43471220e-01
-1.91175729e-01 -9.34112370e-02 -4.15957123e-01 1.89607784e-01
7.27219224e-01 -2.18409210e-01 -8.77824306e-01 -7.53816664e-01
-2.96589077e-01 7.57178664e-02 8.72113109e-01 4.74345297e-01
6.51253819e-01 -1.23518014e+00 -1.03799486e+00 -5.81536829e-01
4.52741921e-01 -3.39023739e-01 -1.16481826e-01 8.24154198e-01
-2.06352785e-01 8.63589272e-02 2.53412366e-01 -3.03091686e-02
-7.49059618e-01 6.28325641e-01 -1.90935507e-01 -1.22743279e-01
-5.58172464e-02 4.21420217e-01 -5.25350749e-01 -5.75600341e-02
-2.34610699e-02 -5.62465042e-02 -4.07369465e-01 3.06562006e-01
8.42782974e-01 7.81589925e-01 2.78104335e-01 -7.63513207e-01
-3.61348152e-01 3.28995794e-01 1.91498429e-01 -7.25229740e-01
9.26577091e-01 -5.51859796e-01 -5.00135362e-01 9.90011036e-01
8.23600650e-01 3.91819566e-01 -4.36461210e-01 -1.04558639e-01
2.81592399e-01 -5.63800514e-01 -3.53285104e-01 -7.61729538e-01
-5.66885352e-01 3.09091151e-01 1.02034714e-02 9.97166097e-01
6.85656607e-01 -7.73602277e-02 4.57830280e-01 1.89531416e-01
9.57369804e-01 -1.61575401e+00 -7.23164761e-03 2.30847836e-01
1.05787027e+00 -1.44696641e+00 4.04438496e-01 -3.40205491e-01
-1.13973093e+00 7.78629005e-01 4.81141478e-01 -7.20078275e-02
4.96078283e-01 5.94628900e-02 8.88841078e-02 -1.08151753e-02
-8.77169371e-01 -1.23197407e-01 1.75059512e-01 6.89775765e-01
9.45688307e-01 3.64993781e-01 -1.03161538e+00 1.01540172e+00
-1.55074552e-01 1.15094990e-01 1.12864292e+00 7.61404395e-01
-5.47234058e-01 -8.82739007e-01 -5.16490877e-01 6.35624826e-01
-6.20040238e-01 -8.82704090e-03 -7.75148451e-01 4.03572053e-01
-3.44280213e-01 2.01192927e+00 -2.29143932e-01 -7.07102060e-01
2.89320171e-01 -1.06423490e-01 -1.00424677e-01 -6.35314703e-01
-6.76478267e-01 -7.83245206e-01 4.87349212e-01 -6.05525792e-01
-6.04789495e-01 -3.43812346e-01 -9.96871889e-01 -7.17406034e-01
-6.50793910e-01 6.65772021e-01 8.40826988e-01 1.20153296e+00
3.65360111e-01 -3.53461623e-01 7.96685994e-01 -4.41075951e-01
-6.83347762e-01 -1.19689369e+00 -9.12143111e-01 5.83418429e-01
2.01870613e-02 -7.07859933e-01 -8.38135779e-01 -3.62822056e-01] | [11.767109870910645, 8.8574800491333] |
250480df-dd2a-4f6c-8e14-cc4d9193aa5e | focalmix-semi-supervised-learning-for-3d | 2003.09108 | null | https://arxiv.org/abs/2003.09108v1 | https://arxiv.org/pdf/2003.09108v1.pdf | FocalMix: Semi-Supervised Learning for 3D Medical Image Detection | Applying artificial intelligence techniques in medical imaging is one of the most promising areas in medicine. However, most of the recent success in this area highly relies on large amounts of carefully annotated data, whereas annotating medical images is a costly process. In this paper, we propose a novel method, called FocalMix, which, to the best of our knowledge, is the first to leverage recent advances in semi-supervised learning (SSL) for 3D medical image detection. We conducted extensive experiments on two widely used datasets for lung nodule detection, LUNA16 and NLST. Results show that our proposed SSL methods can achieve a substantial improvement of up to 17.3% over state-of-the-art supervised learning approaches with 400 unlabeled CT scans. | ['Li-Wei Wang', 'Yuan Zhang', 'Kexin Zhang', 'Dong Wang'] | 2020-03-20 | focalmix-semi-supervised-learning-for-3d-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Wang_FocalMix_Semi-Supervised_Learning_for_3D_Medical_Image_Detection_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Wang_FocalMix_Semi-Supervised_Learning_for_3D_Medical_Image_Detection_CVPR_2020_paper.pdf | cvpr-2020-6 | ['medical-image-detection', 'lung-nodule-detection'] | ['computer-vision', 'medical'] | [ 2.90284783e-01 3.33702683e-01 -5.75333953e-01 -2.37795010e-01
-1.18060684e+00 -3.38868380e-01 3.20256233e-01 3.45056891e-01
-3.55520010e-01 3.38171035e-01 7.74388760e-02 -6.87592924e-01
-1.75404660e-02 -3.82607341e-01 -4.76863891e-01 -6.78888977e-01
-1.20508805e-01 8.40881646e-01 6.35871649e-01 3.49626124e-01
-2.02594191e-01 4.80516225e-01 -8.98169219e-01 3.36283207e-01
6.77432299e-01 8.32438827e-01 1.12213485e-01 5.49294949e-01
-1.58266947e-01 1.13089788e+00 -1.27484843e-01 -1.44226924e-01
2.41799250e-01 -4.95095253e-01 -1.00269175e+00 3.93609643e-01
1.12736672e-01 -2.06671894e-01 -2.02657446e-01 8.91497970e-01
6.43191338e-01 -4.32710230e-01 8.01178157e-01 -5.97780883e-01
-1.34499922e-01 8.41990352e-01 -7.66914606e-01 4.87053275e-01
1.37001604e-01 1.83766689e-02 7.44294405e-01 -7.87620723e-01
5.42669475e-01 7.11054862e-01 6.55536056e-01 5.40866017e-01
-8.00929725e-01 -4.87598091e-01 -2.70225316e-01 1.74627025e-02
-1.21157479e+00 -5.14137782e-02 5.49011111e-01 -5.85567296e-01
4.51955706e-01 2.41373524e-01 5.79096377e-01 5.95674872e-01
1.73038274e-01 1.18962443e+00 1.32210302e+00 -7.52364576e-01
2.73306906e-01 1.67387366e-01 1.31826118e-01 1.30930865e+00
4.04831916e-01 -4.16508466e-02 2.86221560e-02 -2.83556014e-01
8.59949529e-01 1.46542296e-01 -8.84842947e-02 -7.58728623e-01
-1.48347747e+00 8.02435994e-01 5.56751907e-01 4.52201009e-01
-2.81672657e-01 -1.27605185e-01 4.40518707e-01 -3.00424814e-01
5.02871335e-01 3.05393845e-01 -1.94923773e-01 1.54872969e-01
-8.60767066e-01 -3.15402567e-01 8.31162810e-01 6.47924125e-01
-5.21471798e-02 -3.43521953e-01 -1.18536927e-01 7.57505238e-01
5.41556418e-01 4.00693119e-01 8.01065207e-01 -5.12936950e-01
1.62945569e-01 8.13619554e-01 -2.00584948e-01 -2.56162047e-01
-6.72123313e-01 -4.86720651e-01 -8.53000641e-01 2.11468160e-01
4.51705992e-01 -8.80701393e-02 -1.33043015e+00 1.02454007e+00
4.33435321e-01 1.50721759e-01 -4.86888960e-02 9.39361095e-01
1.18369055e+00 9.50858518e-02 2.50563830e-01 -1.86342388e-01
1.43470478e+00 -1.28943133e+00 -4.99821395e-01 -1.14918254e-01
6.38126791e-01 -8.99091959e-01 8.10246170e-01 1.41010508e-01
-1.11842442e+00 -3.39470118e-01 -1.04055548e+00 2.96268702e-01
1.29206315e-01 3.37276012e-01 8.15477371e-01 7.78289557e-01
-5.80099881e-01 8.08433294e-02 -1.36685383e+00 -7.00151682e-01
7.42247701e-01 4.85185415e-01 -2.49793082e-01 -3.20161611e-01
-7.01263011e-01 9.90790248e-01 3.51307303e-01 -4.69377249e-01
-1.02503538e+00 -7.93107986e-01 -5.27957261e-01 -3.59202802e-01
8.56618047e-01 -7.07098365e-01 1.86432838e+00 -3.73459905e-01
-1.15260684e+00 1.35522962e+00 5.84851727e-02 -5.62263787e-01
6.76835299e-01 -8.37542787e-02 -3.02365363e-01 3.81504238e-01
1.34764850e-01 6.04725957e-01 5.85522473e-01 -1.15203619e+00
-6.06431782e-01 -3.86589527e-01 -3.67178321e-01 1.43985942e-01
-2.77951062e-01 9.20935348e-02 -7.51829863e-01 -6.38200164e-01
3.20928961e-01 -1.26105642e+00 -6.87587738e-01 2.90879399e-01
-5.51501215e-01 -3.46113205e-01 5.95161796e-01 -3.27298194e-01
1.21898401e+00 -1.86479020e+00 -1.23115517e-01 1.83321550e-01
4.66862291e-01 5.05503118e-01 4.91718829e-01 -1.57541409e-01
-4.71192151e-02 8.63488466e-02 -3.67168307e-01 -1.14948638e-01
-3.28569114e-01 3.81530613e-01 3.45887631e-01 4.02397603e-01
-1.10172994e-01 1.13091409e+00 -9.46886063e-01 -1.22169149e+00
4.74597394e-01 7.14618564e-02 -3.32098424e-01 4.44785058e-02
-9.47070792e-02 7.23136842e-01 -7.95239449e-01 1.05115020e+00
2.88347632e-01 -8.35649610e-01 2.88155138e-01 -2.45412394e-01
3.77618931e-02 1.79815348e-02 -8.05582702e-01 1.80023015e+00
-2.35641509e-01 1.08314157e-02 -3.12794060e-01 -8.94166827e-01
3.96585345e-01 6.88016593e-01 1.06909585e+00 -2.66176999e-01
3.49530637e-01 4.37816352e-01 1.77250713e-01 -8.30905914e-01
-3.28875303e-01 -2.63822168e-01 6.90929219e-02 6.96413517e-01
-8.57144222e-02 -2.81646103e-01 2.23815352e-01 1.93557262e-01
1.32809389e+00 -2.15594321e-01 9.79604661e-01 -2.92299509e-01
5.96059978e-01 3.91126275e-01 2.24102989e-01 9.23551679e-01
-5.76969087e-01 5.67284644e-01 9.71166417e-02 -5.50330818e-01
-9.08103585e-01 -1.07609272e+00 -3.92650336e-01 5.93260884e-01
-3.06055751e-02 -1.78583711e-01 -8.66496146e-01 -1.29926538e+00
-9.42396522e-02 2.53740162e-01 -5.87941706e-01 2.74361789e-01
-6.84477329e-01 -8.80845904e-01 4.89950478e-01 7.12453783e-01
5.35181105e-01 -1.08902371e+00 -7.83955097e-01 8.05156603e-02
-7.75336996e-02 -1.32226288e+00 -3.38129818e-01 3.23135853e-01
-1.33673561e+00 -1.30746067e+00 -9.92774725e-01 -9.96733248e-01
9.94696915e-01 4.62911904e-01 1.34196210e+00 2.02146068e-01
-8.62842739e-01 2.52829581e-01 -3.38452309e-01 -7.30521441e-01
-7.32875645e-01 3.03838730e-01 -2.13330865e-01 -4.35151607e-01
3.81640762e-01 3.08451559e-02 -6.12380505e-01 4.36249167e-01
-9.35359001e-01 1.84497163e-01 1.25992751e+00 8.70837450e-01
1.01275837e+00 1.55852988e-01 1.80738255e-01 -1.60392857e+00
6.90889955e-02 -3.73983234e-01 -2.53082663e-01 4.30746526e-01
-6.17981732e-01 1.29906669e-01 1.81568399e-01 -2.51002580e-01
-1.13349760e+00 5.88769257e-01 -1.35149077e-01 -1.75722703e-01
-5.12030005e-01 5.52832603e-01 3.91561925e-01 -4.35379088e-01
9.06065524e-01 -1.80951789e-01 -9.29904804e-02 -4.40583766e-01
1.71825677e-01 7.31039882e-01 4.48333800e-01 -2.62098312e-01
7.18069017e-01 8.23225319e-01 2.72909105e-01 -4.61857021e-01
-1.51877093e+00 -1.03863931e+00 -8.69534731e-01 -1.60862550e-01
9.19466078e-01 -6.92182958e-01 -1.33978382e-01 1.22893393e-01
-4.41903204e-01 -1.11488841e-01 -3.87942702e-01 8.53975356e-01
-3.65725398e-01 4.66890424e-01 -7.40567565e-01 -4.05273944e-01
-4.81159538e-01 -1.24108982e+00 1.04568720e+00 5.12679592e-02
-1.89661786e-01 -9.08964097e-01 1.55280590e-01 7.71931767e-01
2.23479927e-01 2.50300735e-01 7.16565609e-01 -8.21280897e-01
-5.52775264e-01 -3.89603615e-01 -3.00981164e-01 6.26328588e-02
5.25747955e-01 -3.26799542e-01 -7.20951080e-01 -1.54493019e-01
2.32395843e-01 -4.57413465e-01 1.08358443e+00 5.67630351e-01
1.25978684e+00 2.11592183e-01 -7.44868398e-01 2.15811789e-01
1.27789152e+00 3.23226392e-01 1.64101169e-01 1.49240792e-01
6.24888062e-01 4.68115546e-02 7.46344745e-01 2.92495370e-01
5.80791906e-02 3.61178815e-01 4.63720977e-01 -6.55856609e-01
-4.79744792e-01 -2.02195048e-02 -2.73011208e-01 8.50239396e-01
-2.84747630e-01 4.96648019e-03 -1.50074506e+00 4.50191587e-01
-1.72842193e+00 -5.13154566e-01 -2.82543618e-02 1.80883586e+00
1.00898612e+00 4.67553675e-01 -2.96580363e-02 2.68428117e-01
4.30963755e-01 -1.15745269e-01 -5.01487255e-01 3.21515411e-01
4.86663282e-01 4.43881959e-01 6.09738231e-01 1.12834796e-01
-1.67362869e+00 5.77175736e-01 6.72295618e+00 8.18719923e-01
-8.74919236e-01 3.75514269e-01 8.10645223e-01 7.65544623e-02
3.79147857e-01 -3.70528579e-01 -6.04600072e-01 1.52108997e-01
5.38987100e-01 1.44349799e-01 -1.95905283e-01 1.19097316e+00
5.81189319e-02 -3.01432341e-01 -1.05171251e+00 8.81175935e-01
4.06752408e-01 -1.21373057e+00 -2.07859531e-01 -9.31015015e-02
1.04467237e+00 2.32261807e-01 1.88033115e-02 7.75601491e-02
2.62838215e-01 -1.00506794e+00 8.49671811e-02 1.97343558e-01
6.79540873e-01 -3.23769987e-01 1.05408657e+00 3.96708518e-01
-1.13506651e+00 2.42114350e-01 -5.28444834e-02 5.74755013e-01
1.65824398e-01 8.28732431e-01 -1.43381059e+00 4.03233856e-01
7.10600495e-01 5.72192729e-01 -7.50105977e-01 1.45556295e+00
-3.93327653e-01 1.05116606e+00 -2.56431401e-01 1.35001540e-02
2.32323095e-01 4.81053978e-01 3.63285035e-01 1.19526398e+00
7.28767738e-02 3.86076093e-01 7.50231683e-01 1.97248697e-01
-1.67425439e-01 2.35156536e-01 -4.25830364e-01 -2.01747268e-02
-3.61016914e-02 1.44715738e+00 -1.18532872e+00 -4.86770421e-01
-7.47025609e-01 5.71365297e-01 -4.15291004e-02 -3.14343095e-01
-9.78061259e-01 3.34705472e-01 -4.76797402e-01 2.24768892e-01
1.43496796e-01 -7.05701485e-03 -2.67969429e-01 -9.71820116e-01
-2.11621508e-01 -7.68383443e-01 9.33944881e-01 -4.12713826e-01
-1.48966634e+00 5.44068396e-01 -5.76159284e-02 -1.46504021e+00
-8.54272991e-02 -6.71117544e-01 -2.54031301e-01 1.50889397e-01
-1.69478428e+00 -1.36894631e+00 -5.45776963e-01 5.42990923e-01
7.90650785e-01 -1.92760587e-01 9.68566597e-01 3.89279395e-01
-2.72681206e-01 2.64412791e-01 4.80718128e-02 3.46968919e-01
8.42723966e-01 -1.44524181e+00 2.81746909e-02 3.37252438e-01
4.03759688e-01 1.26764640e-01 1.80130526e-01 -5.03790319e-01
-1.28089738e+00 -1.06555367e+00 6.52939916e-01 -4.23137188e-01
5.70174396e-01 2.69812644e-01 -5.97817659e-01 7.02169180e-01
6.67202175e-02 6.45709276e-01 9.61948991e-01 -2.73918629e-01
8.58921185e-02 2.13267311e-01 -1.22347653e+00 3.80788773e-01
7.96294391e-01 -1.55011997e-01 -5.58194816e-01 8.27685654e-01
4.17797625e-01 -8.81376982e-01 -1.05782938e+00 9.14032459e-01
3.83514702e-01 -8.18583846e-01 1.16303742e+00 -5.02574980e-01
2.33596608e-01 -3.75127196e-01 1.39244467e-01 -8.33410501e-01
-2.77752012e-01 1.98524334e-02 -2.46356264e-01 5.36499858e-01
5.38379312e-01 -1.17456287e-01 1.32067657e+00 2.81359315e-01
-2.15181306e-01 -1.01056623e+00 -7.48710811e-01 -5.62961638e-01
2.54701357e-02 -1.97148636e-01 -9.88346413e-02 9.24981415e-01
-9.99084953e-03 1.39405742e-01 8.15147310e-02 3.95948552e-02
9.39867318e-01 2.14498386e-01 2.84224629e-01 -1.18394244e+00
-5.75197875e-01 -2.51518995e-01 -3.07301849e-01 -7.03701794e-01
-2.49920592e-01 -1.05559933e+00 1.28798097e-01 -1.70345211e+00
8.70769024e-01 -5.52941620e-01 -4.80045795e-01 6.32055521e-01
-2.74607271e-01 6.82340562e-01 -4.29276899e-02 5.31949282e-01
-9.70386684e-01 -1.29789576e-01 1.54877412e+00 -3.17465037e-01
1.76762380e-02 5.05589008e-01 -4.72113907e-01 1.22602570e+00
1.02079129e+00 -6.34945691e-01 -2.40646631e-01 -2.31097803e-01
-3.07904124e-01 1.20500579e-01 2.43315905e-01 -1.09882843e+00
3.95833343e-01 -6.31703958e-02 4.88640338e-01 -8.37664664e-01
-3.48003618e-02 -9.70038533e-01 -1.53574854e-01 8.89994323e-01
-3.31004381e-01 -4.22818452e-01 1.41926780e-01 5.82246125e-01
-3.24962109e-01 -3.81924510e-01 1.04838538e+00 -7.33740449e-01
-7.02590883e-01 3.34097475e-01 -3.54248434e-01 3.58498394e-02
1.42177761e+00 1.56440571e-01 -6.33950345e-04 3.91967557e-02
-1.07989061e+00 1.56490162e-01 2.12743580e-02 3.24306414e-02
3.96550298e-01 -1.13576257e+00 -7.58656025e-01 -1.13487095e-01
2.28504911e-01 2.81005621e-01 5.56284711e-02 1.12199485e+00
-6.09510899e-01 8.59180331e-01 1.88564122e-01 -1.06274557e+00
-1.54344523e+00 6.56856656e-01 2.68503845e-01 -9.50147331e-01
-7.20906854e-01 7.29335845e-01 -3.53330723e-03 -3.59758675e-01
3.86686355e-01 -3.46907943e-01 -1.87536642e-01 -3.86714846e-01
2.14483678e-01 2.52246439e-01 3.86287063e-01 -3.41813266e-01
-5.39847672e-01 5.52818060e-01 -3.62027824e-01 7.21733272e-02
1.12361670e+00 2.85449833e-01 9.85040739e-02 2.73875505e-01
7.33488619e-01 -1.21662894e-03 -7.18240798e-01 -6.16112351e-01
2.53462642e-01 -2.16017053e-01 2.70244628e-01 -1.10261571e+00
-1.10173190e+00 7.57393360e-01 8.99032891e-01 9.85579286e-03
1.09302557e+00 6.75472021e-01 7.41483092e-01 5.71651042e-01
4.04407412e-01 -6.22869134e-01 4.15715098e-01 1.54359639e-02
3.40417504e-01 -1.84675992e+00 4.94367123e-01 -9.10474598e-01
-7.60744393e-01 9.40393806e-01 5.58864057e-01 6.72720075e-02
9.35490847e-01 5.36845982e-01 3.90200585e-01 -4.36517864e-01
-4.71626043e-01 -3.53802264e-01 5.07270157e-01 2.25130349e-01
7.64372289e-01 1.80170327e-01 -2.61840135e-01 2.62200236e-01
3.39468837e-01 2.70856917e-01 8.76151249e-02 1.42153406e+00
-6.05169237e-01 -1.31570256e+00 -5.53648233e-01 6.92520142e-01
-8.73628318e-01 7.67805949e-02 -3.30236286e-01 1.05477571e+00
8.20733309e-02 5.49055517e-01 -4.09991443e-01 2.36774892e-01
2.03742221e-01 -8.93444270e-02 7.17267632e-01 -9.96393144e-01
-5.23479700e-01 4.15153414e-01 -1.48970902e-01 -2.11097345e-01
-9.88519788e-01 -7.37676561e-01 -1.69572377e+00 3.38809758e-01
-5.55264652e-01 -5.32465279e-02 5.24314106e-01 8.69763553e-01
-1.57347620e-01 7.16885626e-01 4.85365033e-01 -4.56241190e-01
-5.68277180e-01 -8.47958207e-01 -3.79391581e-01 5.25230289e-01
-1.33128762e-02 -6.66135073e-01 -1.21539421e-01 2.60925800e-01] | [15.04142951965332, -2.2441916465759277] |
403a6e53-9a66-4e54-b950-1813f08fbe43 | parallelized-acquisition-for-active-learning | 2305.19267 | null | https://arxiv.org/abs/2305.19267v1 | https://arxiv.org/pdf/2305.19267v1.pdf | Parallelized Acquisition for Active Learning using Monte Carlo Sampling | Bayesian inference remains one of the most important tool-kits for any scientist, but increasingly expensive likelihood functions are required for ever-more complex experiments, raising the cost of generating a Monte Carlo sample of the posterior. Recent attention has been directed towards the use of emulators of the posterior based on Gaussian Process (GP) regression combined with active sampling to achieve comparable precision with far fewer costly likelihood evaluations. Key to this approach is the batched acquisition of proposals, so that the true posterior can be evaluated in parallel. This is usually achieved via sequential maximization of the highly multimodal acquisition function. Unfortunately, this approach parallelizes poorly and is prone to getting stuck in local maxima. Our approach addresses this issue by generating nearly-optimal batches of candidates using an almost-embarrassingly parallel Nested Sampler on the mean prediction of the GP. The resulting nearly-sorted Monte Carlo sample is used to generate a batch of candidates ranked according to their sequentially conditioned acquisition function values at little cost. The final sample can also be used for inferring marginal quantities. Our proposed implementation (NORA) demonstrates comparable accuracy to sequential conditioned acquisition optimization and efficient parallelization in various synthetic and cosmological inference problems. | ['Jonas El Gammal', 'Nils Schöneberg', 'Jesús Torrado'] | 2023-05-30 | null | null | null | null | ['bayesian-inference', 'active-learning', 'active-learning'] | ['methodology', 'methodology', 'natural-language-processing'] | [ 1.86069623e-01 -4.47261706e-02 3.88333768e-01 -4.26776916e-01
-1.23725188e+00 -4.45555568e-01 8.56520176e-01 2.44115308e-01
-8.00179541e-01 1.02508450e+00 -1.70351475e-01 -2.96267807e-01
-3.44812907e-02 -9.41281736e-01 -5.15025020e-01 -1.20354092e+00
1.81762338e-01 1.33703589e+00 5.94733953e-01 2.74296016e-01
5.77379882e-01 5.87026179e-01 -1.52883732e+00 -3.49584818e-01
1.10688448e+00 7.04148412e-01 4.49944258e-01 8.35943699e-01
-8.28203335e-02 1.40951484e-01 -4.88876432e-01 -4.60744530e-01
1.08980261e-01 -4.85489666e-01 -3.51675957e-01 -2.72768855e-01
-6.73966890e-04 -2.45440081e-01 4.70881283e-01 1.12045622e+00
5.80632329e-01 3.51494730e-01 9.80624259e-01 -5.27777910e-01
3.63111198e-01 4.97880250e-01 -5.66108823e-01 -6.86515868e-02
1.53716937e-01 2.15561032e-01 8.22633922e-01 -1.06017840e+00
3.38523835e-01 1.19129145e+00 3.70547980e-01 -1.04815856e-01
-1.63457060e+00 -3.87831271e-01 -5.12323081e-01 -1.55393230e-02
-1.54096413e+00 -5.50701678e-01 4.11076069e-01 -3.38579655e-01
6.81762636e-01 3.35161746e-01 6.33942544e-01 7.20595241e-01
2.20851764e-01 3.11481416e-01 1.09521389e+00 -4.80404675e-01
8.26162994e-01 3.98516089e-01 -1.35452732e-01 5.21455586e-01
3.97978246e-01 -7.19927251e-02 -3.74540716e-01 -5.99681556e-01
6.89008832e-01 -4.56787765e-01 -1.30258128e-01 -2.88934290e-01
-1.15022826e+00 8.12020004e-01 -2.24663466e-01 -1.96403936e-02
-6.20957315e-01 9.68663320e-02 2.84394901e-02 -1.15049042e-01
5.40809810e-01 4.60365623e-01 -2.30494931e-01 -4.21800464e-01
-1.46030462e+00 6.98753357e-01 1.00195396e+00 3.94827157e-01
8.34871173e-01 -9.81477648e-02 -1.16328336e-01 7.51156688e-01
6.10860407e-01 7.71196663e-01 -3.35184373e-02 -9.52429950e-01
3.82968038e-01 1.13011815e-01 6.54731631e-01 -6.32660627e-01
-2.58578956e-01 -5.80657959e-01 -6.04300380e-01 5.25400817e-01
8.04805815e-01 -2.10846439e-01 -6.66303277e-01 1.33761239e+00
7.99283862e-01 -1.21900439e-01 -2.16546714e-01 7.44435430e-01
5.37901074e-02 9.37810659e-01 -3.74625064e-02 -4.83747959e-01
1.24381316e+00 -4.97652590e-01 -2.71999031e-01 3.68858501e-02
8.83542374e-02 -1.05351830e+00 6.81242704e-01 6.53735816e-01
-1.33646905e+00 -4.53562528e-01 -9.09771860e-01 2.27735654e-01
7.33097345e-02 1.48497179e-01 4.30440098e-01 9.39611256e-01
-8.62977266e-01 9.38266575e-01 -1.13293457e+00 4.25751843e-02
2.96181947e-01 5.42151988e-01 1.04760550e-01 2.46073738e-01
-5.04823864e-01 8.36989403e-01 6.06873095e-01 2.33486250e-01
-9.14010763e-01 -6.42769754e-01 -4.16706413e-01 1.11576132e-01
3.85329098e-01 -5.05745530e-01 1.14576125e+00 -5.60346067e-01
-2.00445342e+00 3.48768860e-01 -3.49561751e-01 -3.17072660e-01
7.56143212e-01 -1.42854303e-01 5.75784221e-03 1.35192156e-01
-1.28682718e-01 3.74561578e-01 1.12255287e+00 -1.08883202e+00
-3.90336335e-01 -3.64092261e-01 -4.74803597e-01 2.09105179e-01
3.13042313e-01 1.69225529e-01 -3.74405444e-01 -1.19990252e-01
4.34495121e-01 -8.02859783e-01 -4.24291700e-01 -3.72788459e-01
-2.45614454e-01 -1.56815514e-01 2.34227523e-01 -7.43557572e-01
9.50606525e-01 -1.82315147e+00 2.38765389e-01 5.48317850e-01
1.97626092e-02 1.46628721e-02 4.08859313e-01 4.20509607e-01
2.18477309e-01 -2.31633708e-01 -4.99115437e-01 -4.79184747e-01
3.88824083e-02 -2.26437729e-02 -3.04886848e-01 6.82213068e-01
2.17861399e-01 4.94875431e-01 -8.92011940e-01 -6.36521935e-01
2.62168527e-01 3.36288065e-01 -7.30451763e-01 3.02436382e-01
-5.19385576e-01 7.44184971e-01 -5.20745695e-01 3.21505755e-01
8.54917407e-01 -1.26509115e-01 1.44846290e-01 3.95569438e-03
-4.47478414e-01 3.02199215e-01 -1.47347462e+00 1.17863512e+00
-2.25954995e-01 2.75926054e-01 8.21664408e-02 -9.49828565e-01
1.05657244e+00 2.80930996e-01 1.30312011e-01 -2.25590274e-01
1.85629472e-01 5.09769917e-01 6.89815730e-02 -8.50094408e-02
7.17878699e-01 -3.47830623e-01 1.16874287e-02 4.94866908e-01
-6.87310845e-02 -6.96914971e-01 3.75482202e-01 1.43297330e-01
5.79048932e-01 6.18337274e-01 3.42891753e-01 -4.59273070e-01
6.47267103e-01 -1.01779036e-01 3.83210063e-01 9.43921745e-01
1.99923888e-01 6.95156991e-01 7.13857591e-01 -1.24353588e-01
-1.38772917e+00 -1.28594792e+00 -3.48414719e-01 6.02356732e-01
-2.77888119e-01 -1.27932414e-01 -7.32324898e-01 -2.57271379e-01
-2.40515456e-01 1.02221715e+00 -1.63475201e-01 2.66159266e-01
-6.00613356e-01 -1.30799556e+00 1.03473648e-01 -4.50632274e-02
1.34557739e-01 -9.25368369e-01 -7.52859473e-01 4.34157342e-01
-7.18575763e-03 -4.68504548e-01 1.16955549e-01 2.26540029e-01
-1.09177697e+00 -5.66018581e-01 -8.31824839e-01 -3.83490175e-02
8.20075095e-01 -2.70768523e-01 1.00897348e+00 -3.37923408e-01
-9.79949310e-02 -5.64264357e-02 -3.51099074e-02 -1.92555353e-01
-6.93839729e-01 -1.21511854e-01 -2.93137645e-03 2.48364657e-01
3.50922011e-02 -6.53216243e-01 -6.29538178e-01 1.76628113e-01
-6.72947764e-01 2.95304437e-03 4.60870266e-01 7.97826529e-01
6.12555087e-01 9.55497622e-02 3.34741950e-01 -5.86601019e-01
3.86242241e-01 -3.12085748e-01 -1.50425565e+00 5.74600436e-02
-2.71584630e-01 4.00581568e-01 6.55616522e-01 -2.17533246e-01
-1.62631750e+00 -2.43926831e-02 -2.66990662e-01 1.72082692e-01
3.74385230e-02 2.90509701e-01 -2.51793377e-02 5.13610169e-02
4.80866373e-01 1.73485816e-01 -9.38258320e-02 -5.63651919e-01
2.34884396e-01 5.59747100e-01 3.82218301e-01 -8.29044938e-01
6.83398247e-01 4.87370968e-01 3.16603631e-01 -1.10103428e+00
-5.07393479e-01 -3.14450771e-01 -4.48236108e-01 -3.91810656e-01
7.43484974e-01 -5.25580406e-01 -6.57412887e-01 4.79839116e-01
-1.00992906e+00 -8.85778069e-02 -3.15819800e-01 7.15363979e-01
-4.47254688e-01 6.59992516e-01 -1.17850222e-01 -1.35024607e+00
-1.99763298e-01 -1.30853713e+00 1.20727932e+00 4.23214078e-01
-2.30646193e-01 -6.38775289e-01 3.38990748e-01 3.70433152e-01
3.04207236e-01 -8.92472193e-02 7.70636737e-01 -4.85874444e-01
-8.25491130e-01 -4.66214001e-01 -1.60188377e-01 9.73928571e-02
-4.29876775e-01 2.74483949e-01 -1.05089164e+00 -2.39919975e-01
2.33701810e-01 -1.48307785e-01 6.96483016e-01 5.44663191e-01
9.41118896e-01 2.32406199e-01 -3.50618273e-01 2.52205372e-01
1.47057354e+00 2.06299394e-01 6.55304492e-01 -9.98318642e-02
1.20614454e-01 5.87312758e-01 6.45637393e-01 7.19551384e-01
-1.24725759e-01 7.19521403e-01 -2.94014765e-03 4.56128955e-01
4.25853133e-01 2.67414600e-02 2.96454251e-01 8.01469564e-01
-3.50359172e-01 -8.53806809e-02 -9.05217707e-01 3.60225737e-01
-1.63825071e+00 -9.81093168e-01 -3.62395465e-01 2.80582213e+00
9.98933077e-01 3.22647363e-01 2.67376974e-02 -8.48380774e-02
5.41684926e-01 -2.20520929e-01 -1.52961627e-01 -2.01818630e-01
2.98367918e-01 5.55954993e-01 4.39599454e-01 7.98736036e-01
-7.96572030e-01 5.29916227e-01 6.12595081e+00 1.15030146e+00
-7.30467260e-01 1.96521685e-01 6.55716121e-01 -1.77059114e-01
-3.06353718e-01 5.09455144e-01 -1.13643289e+00 7.99854994e-01
1.20193994e+00 2.45547835e-02 1.28828660e-01 5.81270099e-01
5.24126172e-01 -1.09798980e+00 -7.99749613e-01 6.95024073e-01
-3.40969712e-01 -1.07196724e+00 -2.63458788e-01 1.81501642e-01
6.62205815e-01 -2.31497660e-02 -2.82913327e-01 1.42438069e-01
2.54828900e-01 -4.90033239e-01 6.95103943e-01 8.37631762e-01
3.08484048e-01 -9.02698457e-01 7.23376155e-01 4.76909429e-01
-7.62737751e-01 2.72164404e-01 -6.99212670e-01 7.04381838e-02
6.96862578e-01 1.19502628e+00 -1.07006085e+00 5.00871360e-01
4.28677797e-01 -2.90216029e-01 -2.30451137e-01 1.28020608e+00
-2.09565442e-02 9.28098619e-01 -1.06976676e+00 -4.32293922e-01
8.71986151e-02 -9.50721204e-01 7.34293938e-01 1.03703582e+00
6.15414679e-01 -3.05466712e-01 -8.69860649e-02 1.17782867e+00
5.03292799e-01 7.30811656e-02 -6.47654310e-02 -1.15376957e-01
3.66358846e-01 1.18556225e+00 -1.27536917e+00 -5.60821116e-01
-1.33229598e-01 6.78166926e-01 2.77271211e-01 1.26530275e-01
-8.53868008e-01 -2.23468050e-01 5.67186903e-03 -6.58766776e-02
6.26830995e-01 -4.05985802e-01 -2.10953221e-01 -9.44815278e-01
-1.74828321e-01 -6.88056111e-01 1.37891471e-01 -5.74885130e-01
-9.65531945e-01 2.80987591e-01 2.86604851e-01 -7.19992757e-01
-5.47163069e-01 -3.78124267e-01 -6.70597017e-01 1.33100581e+00
-1.00069332e+00 -4.68616158e-01 -3.00044985e-03 9.51123983e-02
4.36029464e-01 -1.32028125e-02 7.08441794e-01 1.25896111e-01
-4.30249244e-01 3.79607975e-02 5.14046192e-01 -5.69681704e-01
4.62617487e-01 -1.35161328e+00 3.55039924e-01 8.10107112e-01
1.66406948e-02 6.11150444e-01 1.13759816e+00 -9.46391344e-01
-1.16435790e+00 -4.17853981e-01 9.02204692e-01 -3.23242396e-01
6.76470101e-01 -4.71813172e-01 -8.46524537e-01 4.34398986e-02
-1.38123604e-02 -5.54006159e-01 3.30131263e-01 4.56297062e-02
2.66963959e-01 -1.18524425e-01 -1.09642625e+00 7.08438933e-01
1.86307207e-01 -9.74639580e-02 -4.47409034e-01 2.93987006e-01
1.52136639e-01 -1.16929181e-01 -6.88585043e-01 1.89814925e-01
6.24044657e-01 -1.14374352e+00 8.39676380e-01 1.33509085e-01
1.05291530e-01 -5.81481814e-01 5.50447889e-02 -8.35345805e-01
1.50878895e-02 -8.60737801e-01 2.19690487e-01 1.04789269e+00
5.29619694e-01 -7.88929284e-01 8.51790369e-01 4.36783046e-01
2.75073349e-01 -5.73031545e-01 -1.10252738e+00 -5.44984579e-01
-1.92958564e-01 -4.01503921e-01 2.52922207e-01 2.96033412e-01
-3.55071425e-01 3.24397177e-01 -2.62055904e-01 3.62628281e-01
1.23198044e+00 2.35056356e-01 9.14594412e-01 -1.22320044e+00
-9.70237851e-01 -3.27815026e-01 -5.15503809e-02 -1.02302408e+00
-4.83852625e-01 -5.58408022e-01 3.27008843e-01 -1.19421399e+00
2.72117466e-01 -4.36240852e-01 2.66075701e-01 -3.24243069e-01
-1.78419948e-01 2.40690425e-01 -2.12220132e-01 1.67488515e-01
-3.80478680e-01 6.09192312e-01 9.76764083e-01 4.07190740e-01
-1.14868440e-01 3.29092622e-01 5.33150509e-02 8.01203191e-01
6.06323123e-01 -6.72680736e-01 -2.17889920e-01 9.79903117e-02
5.96815586e-01 3.32302332e-01 2.79937148e-01 -9.89548624e-01
1.59843415e-01 3.84417884e-02 4.10353482e-01 -1.24865615e+00
5.78092754e-01 -4.25797880e-01 5.23343563e-01 2.69743025e-01
1.30148605e-01 -2.38196582e-01 -5.08229919e-02 4.81886655e-01
1.15538925e-01 -1.03738678e+00 1.00080085e+00 -1.97424620e-01
6.75362274e-02 -7.25792274e-02 -6.13716900e-01 -2.60179043e-01
7.02255428e-01 -1.91668361e-01 3.90939325e-01 -3.56519550e-01
-6.85498655e-01 -3.82971130e-02 3.70989203e-01 -4.48123127e-01
2.36792520e-01 -7.11606443e-01 -9.07489121e-01 1.30834281e-01
-4.80577171e-01 1.96312979e-01 1.86017245e-01 1.03810465e+00
-9.02567863e-01 3.26455921e-01 2.80437052e-01 -8.58720303e-01
-1.00350785e+00 6.99466541e-02 1.46067396e-01 -4.55716342e-01
-5.67365944e-01 8.18578124e-01 2.00707000e-02 -3.80124420e-01
-8.31855610e-02 -1.08518325e-01 9.60523337e-02 8.19341093e-02
5.22231936e-01 7.31559038e-01 3.03158388e-02 -9.80266482e-02
-2.76268180e-02 2.81005800e-01 -2.36654226e-02 -8.60984147e-01
1.43247581e+00 -1.90254282e-02 -4.67654586e-01 4.80795324e-01
7.06173539e-01 3.12903523e-01 -1.27378643e+00 1.71919316e-01
1.65170282e-01 -4.54380870e-01 1.02632120e-01 -5.33954799e-01
-3.71435493e-01 6.64065540e-01 2.66352803e-01 1.95397690e-01
6.68811321e-01 -4.00761217e-02 3.12946856e-01 4.65052068e-01
5.39195418e-01 -1.20224392e+00 -2.95037985e-01 1.56222790e-01
5.92561662e-01 -9.67671692e-01 4.94277477e-01 -2.38086954e-01
-2.57698029e-01 1.09447575e+00 2.28877198e-02 -1.26834854e-01
5.01104712e-01 2.38395467e-01 -3.73968422e-01 -1.52026042e-01
-4.66976106e-01 1.65393576e-01 1.30451456e-01 1.32234961e-01
2.74623871e-01 9.18929204e-02 -5.40905476e-01 6.98306710e-02
-2.16470450e-01 -3.19012910e-01 3.84109974e-01 7.05204904e-01
-6.85483098e-01 -1.29961157e+00 -7.42834628e-01 4.87899333e-01
-3.92818332e-01 -1.76641539e-01 1.73695922e-01 4.68670875e-01
-1.34205177e-01 6.68627858e-01 1.35224223e-01 5.84955931e-01
-2.25537151e-01 3.55775476e-01 7.90230274e-01 -3.51514399e-01
-3.05627495e-01 5.69026887e-01 2.51055688e-01 -3.21901232e-01
-6.03322536e-02 -1.19743860e+00 -8.75680268e-01 -1.12958625e-01
-6.52286112e-01 5.49837291e-01 1.02351463e+00 1.01658177e+00
-1.43547609e-01 1.29238635e-01 3.88464183e-01 -1.29350853e+00
-9.18007553e-01 -1.01290250e+00 -5.67894995e-01 -5.64526096e-02
-1.10466450e-01 -5.71547389e-01 -4.35297251e-01 -2.98447371e-01] | [6.725743770599365, 3.915452480316162] |
48e4e999-5b74-41f5-b0d2-a65228022fec | centralized-cooperative-exploration-policy | 2301.02375 | null | https://arxiv.org/abs/2301.02375v1 | https://arxiv.org/pdf/2301.02375v1.pdf | Centralized Cooperative Exploration Policy for Continuous Control Tasks | The deep reinforcement learning (DRL) algorithm works brilliantly on solving various complex control tasks. This phenomenal success can be partly attributed to DRL encouraging intelligent agents to sufficiently explore the environment and collect diverse experiences during the agent training process. Therefore, exploration plays a significant role in accessing an optimal policy for DRL. Despite recent works making great progress in continuous control tasks, exploration in these tasks has remained insufficiently investigated. To explicitly encourage exploration in continuous control tasks, we propose CCEP (Centralized Cooperative Exploration Policy), which utilizes underestimation and overestimation of value functions to maintain the capacity of exploration. CCEP first keeps two value functions initialized with different parameters, and generates diverse policies with multiple exploration styles from a pair of value functions. In addition, a centralized policy framework ensures that CCEP achieves message delivery between multiple policies, furthermore contributing to exploring the environment cooperatively. Extensive experimental results demonstrate that CCEP achieves higher exploration capacity. Empirical analysis shows diverse exploration styles in the learned policies by CCEP, reaping benefits in more exploration regions. And this exploration capacity of CCEP ensures it outperforms the current state-of-the-art methods across multiple continuous control tasks shown in experiments. | ['Yu Liu', 'Xinwen Hou', 'Qiang He', 'Chen Gong', 'Chao Li'] | 2023-01-06 | null | null | null | null | ['continuous-control'] | ['playing-games'] | [-5.08499146e-01 7.39075989e-02 -6.07530057e-01 1.78753749e-01
-4.03644234e-01 -3.79778147e-01 6.17491782e-01 -2.31675580e-02
-6.42776012e-01 1.28739500e+00 7.95268789e-02 -3.75149071e-01
-3.12161595e-01 -7.61399329e-01 -5.96145689e-01 -1.15784347e+00
-7.83201456e-01 5.51640034e-01 6.72900528e-02 -4.72119242e-01
2.96141952e-01 1.41595513e-01 -1.38828850e+00 -3.37256640e-01
1.25205326e+00 8.73916924e-01 7.03434050e-01 3.99088651e-01
-2.05318362e-01 8.51382077e-01 -7.94698656e-01 4.30510938e-01
5.54683328e-01 -3.43524843e-01 -7.29458630e-01 -1.07171051e-01
-5.20781338e-01 -6.08636022e-01 -9.31943767e-03 9.81511056e-01
6.01604760e-01 5.86682677e-01 3.38419899e-03 -1.53180158e+00
-6.34741426e-01 1.00148785e+00 -9.11582768e-01 2.76265472e-01
2.39515036e-01 5.57172298e-01 7.54679263e-01 -4.61034954e-01
4.72733498e-01 1.44912875e+00 9.17152390e-02 7.28821933e-01
-1.00677454e+00 -8.40950906e-01 6.79267764e-01 4.61757034e-02
-8.90474558e-01 8.65440965e-02 7.21978962e-01 1.60052180e-01
7.69340813e-01 3.00092548e-02 1.08600986e+00 1.18414772e+00
3.40187639e-01 1.16628325e+00 1.49929595e+00 -2.32014328e-01
9.82226849e-01 1.77785441e-01 -5.90890408e-01 5.36176324e-01
3.61900449e-01 5.25908411e-01 -5.89505374e-01 -2.74727255e-01
1.09039724e+00 -1.87838674e-01 -3.65835845e-01 -6.73504889e-01
-1.35003090e+00 9.32666957e-01 5.72519243e-01 3.48466858e-02
-9.30373549e-01 3.35206836e-01 5.25726438e-01 6.40126765e-01
1.81488935e-02 1.04622889e+00 -4.11786228e-01 -6.87799990e-01
-1.88848391e-01 5.37158012e-01 5.89820266e-01 9.30745125e-01
6.53924167e-01 4.03195351e-01 -3.25080544e-01 7.56613612e-01
1.62280172e-01 5.04722774e-01 6.09674037e-01 -1.51768982e+00
4.47595686e-01 5.68872273e-01 5.58578491e-01 -7.05911994e-01
-2.76732832e-01 -7.52550364e-01 -6.85251594e-01 7.52201796e-01
7.76914954e-02 -6.73997104e-01 -4.07866359e-01 1.87622046e+00
5.81786931e-01 -2.35498399e-01 3.80786717e-01 1.02469015e+00
1.00478334e-02 7.82651186e-01 1.82276666e-01 -2.50664175e-01
1.04203105e+00 -1.16119683e+00 -8.09809029e-01 -2.55948782e-01
3.82185578e-01 -3.84079479e-02 1.17514622e+00 6.30378008e-01
-1.10769439e+00 -5.79314291e-01 -7.87716568e-01 7.86961854e-01
-1.41618118e-01 -2.72515994e-02 8.06872308e-01 3.06300014e-01
-1.18041205e+00 4.30881411e-01 -8.19529474e-01 -1.08473279e-01
4.91587102e-01 3.24338049e-01 2.21953586e-01 1.69409975e-01
-1.34016275e+00 8.30108643e-01 7.06526816e-01 -4.45425324e-02
-1.27111077e+00 -4.94536459e-01 -6.49520516e-01 1.46165177e-01
1.02514839e+00 -3.60774010e-01 1.44487262e+00 -7.85308421e-01
-1.82917690e+00 -1.88121229e-01 3.32553297e-01 -5.69448054e-01
7.08927929e-01 -2.94312179e-01 -4.57548872e-02 1.42849594e-01
8.14891383e-02 1.21026599e+00 8.02514672e-01 -1.72207463e+00
-9.54925060e-01 7.66881704e-02 3.65504742e-01 6.34460747e-01
-6.07014537e-01 -6.33013368e-01 -1.85080469e-01 -4.01956201e-01
-3.33378524e-01 -7.88372159e-01 -7.83062398e-01 -3.63460183e-02
2.32586022e-02 -5.53209066e-01 8.48839700e-01 -2.27853889e-03
1.20949650e+00 -1.80835485e+00 1.85588270e-01 2.90477604e-01
2.83907801e-01 2.21238360e-01 -4.08790648e-01 6.52842104e-01
5.37145197e-01 5.86039573e-02 5.38982973e-02 -2.00464666e-01
4.01896715e-01 6.90962851e-01 -4.82949555e-01 1.50430620e-01
-1.43058613e-01 1.01100826e+00 -1.49395001e+00 -3.35316092e-01
1.93126783e-01 1.57438785e-01 -6.09948754e-01 3.99725407e-01
-6.80514753e-01 7.10215390e-01 -1.07040584e+00 6.79717600e-01
6.94827557e-01 -3.40890527e-01 3.93059224e-01 7.41279721e-01
-3.43977183e-01 -1.01532884e-01 -1.00314867e+00 1.46295393e+00
-3.33700150e-01 3.35075170e-01 4.15646374e-01 -8.95138681e-01
1.16011775e+00 9.71959755e-02 6.63147211e-01 -1.32461488e+00
2.45396256e-01 1.05659641e-01 -9.10619274e-02 -3.98691237e-01
6.56310678e-01 5.08496046e-01 1.06659710e-01 7.57290184e-01
-5.72492599e-01 6.42725229e-02 3.53346825e-01 1.26796871e-01
9.88859057e-01 2.86315024e-01 2.52787143e-01 -4.39785659e-01
4.28686649e-01 1.25483111e-01 7.98816383e-01 1.12855923e+00
-6.77641451e-01 -5.44167757e-01 6.44199312e-01 -4.64144111e-01
-9.35267985e-01 -7.52960563e-01 2.04871550e-01 1.33391428e+00
6.32528484e-01 -9.69573036e-02 -5.59130847e-01 -6.41175091e-01
2.12175071e-01 8.18708956e-01 -8.56515050e-01 -7.10031316e-02
-6.80045366e-01 -5.41571736e-01 1.31670922e-01 5.17119467e-01
1.13033402e+00 -1.62632942e+00 -1.38020742e+00 4.33970988e-01
9.37881842e-02 -7.09197521e-01 -2.62519479e-01 2.90276676e-01
-7.44034886e-01 -9.75315571e-01 -8.39028776e-01 -6.37598395e-01
7.23930418e-01 4.00442153e-01 9.87967134e-01 1.05644077e-01
-1.62594393e-02 4.34982300e-01 -5.06017327e-01 -3.73420835e-01
-4.40930039e-01 2.02362835e-01 1.05905496e-01 -5.66721976e-01
2.65281290e-01 -4.50106323e-01 -7.53472686e-01 4.67050582e-01
-7.89476156e-01 -1.17091991e-01 7.01559067e-01 9.83562350e-01
4.71707463e-01 2.38197103e-01 9.95551467e-01 -2.72643894e-01
1.28446388e+00 -7.73125112e-01 -9.65151727e-01 6.89194873e-02
-9.10338461e-01 3.47296208e-01 8.03740323e-01 -7.44432330e-01
-1.20774496e+00 -6.08018875e-01 1.97047874e-01 -5.00334740e-01
1.78740919e-02 2.90732950e-01 3.55465502e-01 5.90165816e-02
4.07632470e-01 4.64390904e-01 4.12205100e-01 -9.15116072e-02
2.50706792e-01 3.05390030e-01 9.47532728e-02 -1.12721634e+00
4.74074811e-01 2.13296905e-01 -1.87071025e-01 -4.41347986e-01
-3.77780706e-01 -8.57392922e-02 7.07836077e-02 -4.15444404e-01
4.83583361e-01 -8.93177390e-01 -1.25475860e+00 2.36549467e-01
-6.08460248e-01 -9.74624395e-01 -5.37180901e-01 5.07568777e-01
-6.43597186e-01 4.85283621e-02 -2.91435421e-01 -1.21111786e+00
-2.24176556e-01 -1.28192616e+00 6.37540877e-01 5.96665084e-01
7.90736154e-02 -1.09435511e+00 4.41144645e-01 -2.57608175e-01
8.38262022e-01 2.50584513e-01 6.54391527e-01 -2.01489761e-01
-8.47196043e-01 5.58653891e-01 -2.33421326e-02 -9.87744927e-02
7.06196390e-03 -4.42495376e-01 -4.11867827e-01 -9.17428374e-01
-2.68931419e-01 -8.56951475e-01 6.12895846e-01 3.94652188e-01
1.40201187e+00 -4.09500808e-01 -4.65858698e-01 3.23817223e-01
1.23079860e+00 7.19297171e-01 2.99351096e-01 1.01872122e+00
-8.48007202e-02 3.39860827e-01 1.12734258e+00 1.12952518e+00
4.23898309e-01 3.00362885e-01 9.71144080e-01 -9.34859440e-02
4.03601617e-01 -3.57784480e-01 5.53442180e-01 3.28151673e-01
8.88188779e-02 -2.31766507e-01 -5.17896771e-01 6.26657069e-01
-2.10817862e+00 -1.06220698e+00 6.41088188e-01 1.87638927e+00
1.01927543e+00 1.36598334e-01 1.93157747e-01 -3.71155441e-01
4.59364772e-01 2.84353375e-01 -1.29846668e+00 -6.18854761e-01
9.00840238e-02 -1.97882175e-01 4.35651362e-01 3.95635039e-01
-6.47398174e-01 9.09005523e-01 6.23818731e+00 8.01418960e-01
-1.05878067e+00 -1.65606380e-01 5.83313882e-01 -2.22784907e-01
-4.96879905e-01 -3.14209759e-01 -6.83401048e-01 5.74939430e-01
4.60440338e-01 -1.63236126e-01 8.68378341e-01 1.26490211e+00
4.47332740e-01 -4.24260229e-01 -6.32918537e-01 6.14108682e-01
-5.08426726e-01 -1.25286674e+00 -2.77247906e-01 3.31394911e-01
1.15722525e+00 4.32487354e-02 2.71104097e-01 8.92213881e-01
1.09576416e+00 -7.41538346e-01 6.49561226e-01 1.45622715e-01
3.49026918e-01 -9.79527891e-01 5.56947351e-01 4.82297838e-01
-1.21133006e+00 -7.79561698e-01 -5.70802689e-01 -1.03776470e-01
8.03829357e-03 1.41002834e-01 -8.42882872e-01 2.57761955e-01
7.62228966e-01 3.89863312e-01 -2.48525038e-01 9.25022781e-01
-3.21508616e-01 4.92878318e-01 -1.17300250e-01 -6.07317209e-01
1.03875959e+00 -3.40001553e-01 6.90416396e-01 6.91412032e-01
1.55830026e-01 -3.86139564e-02 7.81623900e-01 1.25592220e+00
1.14514574e-01 -2.64820755e-01 -5.13782740e-01 -1.46041200e-01
1.04722691e+00 1.05400145e+00 -5.56785285e-01 -2.76276320e-01
1.41180828e-01 4.53599364e-01 5.71403563e-01 6.63942218e-01
-9.47948277e-01 -2.58856535e-01 8.52004290e-01 -5.62117636e-01
4.10985649e-01 -4.59164143e-01 -5.41084856e-02 -6.38953745e-01
-1.27642825e-01 -9.57466364e-01 1.25734374e-01 -3.93354863e-01
-9.57710445e-01 7.44453728e-01 1.13332391e-01 -1.15390432e+00
-4.48577702e-01 -1.73047170e-01 -6.66873693e-01 5.82672119e-01
-2.00246048e+00 -3.31830919e-01 -3.76699090e-01 4.36411798e-01
8.68690789e-01 -5.21895289e-01 3.70919019e-01 -3.08560640e-01
-7.01459765e-01 3.32145631e-01 4.55171198e-01 -3.81561726e-01
4.88806248e-01 -1.47155738e+00 1.43244848e-01 3.60030890e-01
-6.24609709e-01 5.58900952e-01 6.39035702e-01 -6.39666975e-01
-1.71063721e+00 -8.67838502e-01 -9.85720158e-02 1.39649495e-01
5.19866645e-01 -3.89205217e-02 -8.74484122e-01 2.54318357e-01
7.10379243e-01 -2.54605114e-01 1.58435330e-01 -1.11308470e-01
2.86140293e-01 2.59984192e-02 -1.00153685e+00 8.04133117e-01
7.87528276e-01 2.78560549e-01 -1.92719698e-01 1.80604551e-02
8.18674147e-01 -5.28454304e-01 -7.17764795e-01 1.06505774e-01
3.15667510e-01 -9.26461339e-01 8.29676867e-01 -4.99810010e-01
2.79489726e-01 -2.09630296e-01 1.91242233e-01 -1.87435758e+00
-1.71086401e-01 -9.88609374e-01 -4.71817255e-01 6.58227742e-01
5.47984578e-02 -9.68510151e-01 8.96705687e-01 3.73978049e-01
-8.77636671e-02 -1.38998055e+00 -7.23018706e-01 -1.08157313e+00
1.90967843e-01 1.09933347e-01 9.48184669e-01 8.04119766e-01
1.29958764e-01 -2.97529638e-01 -4.21419919e-01 -1.69467907e-02
8.78436625e-01 2.21507236e-01 9.48726416e-01 -7.51367390e-01
-2.66038626e-01 -6.17124677e-01 6.19546235e-01 -1.40107214e+00
1.82046875e-01 -2.74885952e-01 1.91663072e-01 -1.54119623e+00
-1.55693702e-02 -1.19483888e+00 -2.63722420e-01 6.58675253e-01
-2.16068745e-01 -4.87841636e-01 2.77067453e-01 3.34362626e-01
-1.03000689e+00 1.27261150e+00 1.95832884e+00 -6.25742450e-02
-6.22723758e-01 -2.83272296e-01 -8.28220487e-01 4.13875788e-01
1.27641106e+00 -1.32696912e-01 -9.09463704e-01 -4.71353501e-01
1.50595710e-01 2.52698243e-01 9.63916183e-02 -8.83587420e-01
1.69476315e-01 -7.39759564e-01 3.19327414e-01 -4.39702421e-01
1.83136299e-01 -7.13071108e-01 -2.56680042e-01 8.80557716e-01
-6.27361417e-01 3.68943244e-01 2.44389459e-01 7.83820510e-01
-6.81283697e-02 1.47195477e-02 6.73922598e-01 -3.42473567e-01
-9.84588563e-01 2.93846756e-01 -5.88562489e-01 2.04883367e-01
1.32672608e+00 -2.95937926e-01 -3.57959151e-01 -4.46388066e-01
-3.37356120e-01 1.38658261e+00 3.69388729e-01 5.50516903e-01
6.77571595e-01 -1.30424368e+00 -4.99198526e-01 2.00960934e-01
-2.53707796e-01 1.82032928e-01 3.25454958e-02 6.59153223e-01
-2.06354856e-01 3.21513981e-01 -6.17304921e-01 -5.21311045e-01
-6.58524811e-01 5.92105091e-01 2.66865373e-01 -6.10547900e-01
-7.07246006e-01 5.70899248e-01 -4.18425314e-02 -4.30900425e-01
7.49896467e-01 -2.94499069e-01 -2.50454187e-01 -2.06403628e-01
5.70960641e-01 7.07023442e-01 -6.58464551e-01 4.07383174e-01
-2.26701032e-02 1.46696687e-01 -1.78812653e-01 -7.79869482e-02
1.28440118e+00 -3.48582298e-01 1.97659001e-01 -1.16260499e-01
6.33025706e-01 -3.34385753e-01 -2.05289388e+00 -1.92033008e-01
-3.40668738e-01 -6.26468718e-01 2.08901465e-01 -1.05605626e+00
-1.08556247e+00 5.16025484e-01 4.46068138e-01 4.89217639e-01
1.09407520e+00 -3.01747233e-01 8.14873695e-01 6.71819627e-01
8.65105271e-01 -1.44735515e+00 6.52130604e-01 6.25604451e-01
9.26536798e-01 -1.31860101e+00 -1.01695880e-01 2.43656069e-01
-1.21595955e+00 9.11463380e-01 1.18263686e+00 -1.44860461e-01
7.93928802e-02 3.10370624e-01 -3.04055959e-02 -1.42245978e-01
-1.12401557e+00 -1.06766753e-01 -5.13331831e-01 7.34468758e-01
-1.45176485e-01 2.84882225e-02 -3.07520777e-01 -8.69681011e-04
1.04454644e-01 -1.39475793e-01 4.15469348e-01 1.23693061e+00
-9.80571151e-01 -1.19230592e+00 -5.07966101e-01 7.64110759e-02
8.30851868e-02 4.32968020e-01 4.70737927e-02 1.01638448e+00
-7.27433041e-02 8.02900255e-01 5.06389104e-02 7.63672441e-02
-5.83768822e-02 -6.93332613e-01 8.22918266e-02 7.13810250e-02
-6.52028859e-01 4.61468287e-02 -2.28949204e-01 -9.89701688e-01
-1.11031123e-01 -4.18823451e-01 -1.65345526e+00 -2.70758390e-01
2.04976345e-03 7.65894473e-01 4.71432328e-01 5.57744443e-01
5.68317354e-01 8.30074191e-01 9.54208016e-01 -8.94986093e-01
-1.14417088e+00 -6.12754583e-01 -5.61898470e-01 -1.69489868e-02
5.00921726e-01 -1.21452498e+00 -1.41128570e-01 -7.27007210e-01] | [3.90847110748291, 2.0250210762023926] |
2be972da-8586-4e8f-a7e6-f0a153483ddf | self-supervised-face-presentation-attack | 2208.13070 | null | https://arxiv.org/abs/2208.13070v4 | https://arxiv.org/pdf/2208.13070v4.pdf | Self-Supervised Face Presentation Attack Detection with Dynamic Grayscale Snippets | Face presentation attack detection (PAD) plays an important role in defending face recognition systems against presentation attacks. The success of PAD largely relies on supervised learning that requires a huge number of labeled data, which is especially challenging for videos and often requires expert knowledge. To avoid the costly collection of labeled data, this paper presents a novel method for self-supervised video representation learning via motion prediction. To achieve this, we exploit the temporal consistency based on three RGB frames which are acquired at three different times in the video sequence. The obtained frames are then transformed into grayscale images where each image is specified to three different channels such as R(red), G(green), and B(blue) to form a dynamic grayscale snippet (DGS). Motivated by this, the labels are automatically generated to increase the temporal diversity based on DGS by using the different temporal lengths of the videos, which prove to be very helpful for the downstream task. Benefiting from the self-supervised nature of our method, we report the results that outperform existing methods on four public benchmarks, namely, Replay-Attack, MSU-MFSD, CASIA-FASD, and OULU-NPU. Explainability analysis has been carried out through LIME and Grad-CAM techniques to visualize the most important features used in the DGS. | ['Mourad Oussalah', 'Usman Muhammad'] | 2022-08-27 | null | null | null | null | ['face-presentation-attack-detection'] | ['computer-vision'] | [ 2.04924807e-01 -5.76798797e-01 -1.89663321e-01 -2.67189443e-01
-4.50432569e-01 -7.58618057e-01 3.90136689e-01 -8.24221000e-02
-8.44776481e-02 5.73795676e-01 -2.67130248e-02 -4.97167021e-01
-8.31907764e-02 -4.76684004e-01 -6.24637961e-01 -1.00699043e+00
-4.24206346e-01 -3.07314694e-01 3.96936089e-02 -1.83490783e-01
2.53515065e-01 6.64012849e-01 -1.59121215e+00 4.84658748e-01
3.71018529e-01 1.37663829e+00 -2.07266271e-01 5.24921894e-01
-4.18121405e-02 8.84357512e-01 -6.80635810e-01 -5.64838231e-01
2.90036649e-01 -4.98216897e-01 -5.13979733e-01 3.13170761e-01
1.45959660e-01 -3.46986800e-01 -5.04567802e-01 1.01272833e+00
4.01073337e-01 4.48755957e-02 4.56340492e-01 -1.66718423e+00
-2.37115800e-01 1.95211291e-01 -9.76141214e-01 4.78826523e-01
5.86021364e-01 2.67864734e-01 6.68334424e-01 -7.05510199e-01
5.81345141e-01 1.42660296e+00 2.22680107e-01 6.70729458e-01
-9.86616671e-01 -1.02759707e+00 5.48473820e-02 8.77492428e-01
-1.38679361e+00 -3.33324105e-01 1.17457783e+00 -2.90227264e-01
3.56378883e-01 3.35340977e-01 4.59699512e-01 1.37323153e+00
3.02615557e-02 6.91714585e-01 1.25682056e+00 -3.72111589e-01
3.67573500e-01 -3.20934430e-02 1.38625979e-01 6.88065290e-01
-1.83195043e-02 1.75905630e-01 -7.85903692e-01 -9.59365964e-02
5.03692985e-01 2.06437021e-01 -5.09050488e-01 -1.30639166e-01
-8.64637256e-01 7.73957193e-01 3.53491098e-01 2.25214899e-01
-4.94446158e-02 -1.91259161e-01 4.70044017e-01 3.67369890e-01
3.48481238e-01 -1.70594886e-01 -2.14351907e-01 -1.64503485e-01
-8.28290343e-01 -2.86940306e-01 4.92272437e-01 5.03022254e-01
7.44573474e-01 2.49329373e-01 2.16376800e-02 4.74082828e-01
2.33030230e-01 4.53346550e-01 6.62924349e-01 -6.16730034e-01
4.82816458e-01 4.69113886e-01 -5.67299388e-02 -1.50639141e+00
-2.21705571e-01 -5.79382218e-02 -9.25402582e-01 3.05686772e-01
5.20334363e-01 -1.21204659e-01 -7.11159885e-01 1.64061999e+00
4.41782385e-01 8.52568567e-01 1.28068924e-01 8.78035069e-01
6.10305548e-01 7.71893203e-01 1.88156128e-01 -6.14896238e-01
1.19813418e+00 -6.62434757e-01 -7.54713476e-01 1.68437362e-01
6.66168094e-01 -6.54822290e-01 9.36655164e-01 6.71684027e-01
-5.08948982e-01 -6.38949096e-01 -1.12626374e+00 6.03661954e-01
-4.16090190e-01 1.01803683e-01 4.44535881e-01 1.08315814e+00
-7.36303866e-01 5.63610315e-01 -5.45683324e-01 -6.93349540e-02
7.25351512e-01 2.35871091e-01 -6.64272726e-01 -1.42612651e-01
-1.23050380e+00 3.56794298e-01 3.42213899e-01 3.22057962e-01
-9.35429096e-01 -2.10133493e-01 -7.46543407e-01 -3.65122147e-02
4.13146764e-01 3.11964422e-01 6.08196020e-01 -1.13554168e+00
-1.39748073e+00 6.89202785e-01 -7.98440874e-02 -5.41513085e-01
5.26077271e-01 -6.25096560e-02 -6.28437281e-01 6.89826727e-01
-2.73736000e-01 4.63233858e-01 1.58323669e+00 -1.16224718e+00
-4.62374836e-01 -4.39493418e-01 -5.04352115e-02 -6.34490922e-02
-8.22395623e-01 2.98078775e-01 -4.54408467e-01 -9.10964727e-01
-1.00414932e-01 -9.36667025e-01 2.99051523e-01 -8.77066329e-02
-4.53950167e-01 -1.61765665e-01 1.56567967e+00 -8.64088297e-01
1.41429770e+00 -2.48406911e+00 1.30299255e-01 3.40521574e-01
4.27542860e-03 5.95067799e-01 -1.25285566e-01 9.58559960e-02
-5.86283505e-01 5.76329157e-02 -1.16084002e-01 -1.66185349e-01
-3.28247696e-01 2.41381731e-02 -5.66442847e-01 6.17010236e-01
2.14454532e-01 4.91090000e-01 -7.79295743e-01 -5.96046031e-01
2.83951521e-01 4.16460097e-01 -2.98123091e-01 3.92905146e-01
-1.01742744e-01 8.04195344e-01 -2.95800596e-01 7.19171941e-01
8.84559631e-01 -9.29382145e-02 8.97680894e-02 -1.81497335e-01
2.71288276e-01 -3.14937264e-01 -1.36229110e+00 1.33591819e+00
-2.39536256e-01 7.09713757e-01 -3.03653836e-01 -1.11592627e+00
8.91200542e-01 3.42028558e-01 6.25580490e-01 -7.57547379e-01
1.84756905e-01 -9.07527879e-02 -1.26289055e-01 -5.73962033e-01
-6.14784248e-02 3.05022955e-01 1.67579040e-01 6.54432774e-01
-1.20474339e-01 3.91650647e-01 2.94865042e-01 3.51667851e-01
9.43908334e-01 7.15959221e-02 -5.30178249e-02 1.81248158e-01
9.25281227e-01 -4.00751740e-01 5.54239571e-01 3.00526410e-01
-3.74609292e-01 3.91340435e-01 8.74688387e-01 -5.64425647e-01
-6.40445232e-01 -7.88145542e-01 -1.72191299e-02 8.99101734e-01
2.16217935e-01 -6.20409787e-01 -9.20987606e-01 -1.03066540e+00
-2.44053334e-01 3.26957613e-01 -7.20142841e-01 -3.52777988e-01
-5.74384272e-01 -6.97570801e-01 5.46082616e-01 4.52961117e-01
6.16748333e-01 -1.21104157e+00 -6.08573437e-01 -9.97404158e-02
-1.63607970e-01 -1.35413837e+00 -3.45358133e-01 -1.21685617e-01
-4.84590083e-01 -1.32910442e+00 -5.78981638e-01 -5.63042581e-01
7.06521213e-01 5.55211544e-01 4.41867709e-01 2.86076456e-01
-5.36170840e-01 4.04012173e-01 -6.15735114e-01 -8.13565999e-02
-2.29935139e-01 -4.75470781e-01 2.26247415e-01 8.14450800e-01
1.51179478e-01 -4.27131385e-01 -5.87387443e-01 4.93160814e-01
-1.15794158e+00 -1.11116149e-01 3.25240195e-01 8.85260522e-01
3.80120128e-01 3.78028035e-01 5.02563596e-01 -6.85561776e-01
2.41465226e-01 -5.96961558e-01 -5.58050692e-01 3.26553583e-01
-5.17715365e-02 -2.10769370e-01 6.89076245e-01 -8.58869553e-01
-8.60623300e-01 2.06095785e-01 1.47443920e-01 -9.41687763e-01
-1.37790874e-01 1.22820966e-01 -3.60765338e-01 -1.65587440e-01
5.01705945e-01 3.07648212e-01 2.10601896e-01 -1.18175507e-01
3.66398573e-01 7.08981752e-01 4.10731435e-01 -2.55746841e-01
1.05974102e+00 5.61867535e-01 6.50773793e-02 -1.02978992e+00
-4.97980893e-01 -2.37703323e-01 -4.76771683e-01 -5.36385715e-01
8.38401675e-01 -6.04997516e-01 -1.03916371e+00 6.86089039e-01
-1.00846231e+00 -6.23105131e-02 1.74508065e-01 3.03536654e-01
-2.73335189e-01 9.22602057e-01 -6.36210263e-01 -9.20886219e-01
-2.74064749e-01 -1.23591971e+00 8.33012521e-01 4.08869058e-01
1.67990357e-01 -8.04838240e-01 -3.87875170e-01 3.50086093e-01
2.10415244e-01 4.02206242e-01 8.68520975e-01 -6.87708259e-01
-3.70442241e-01 -2.90400833e-01 -2.99765825e-01 5.67157626e-01
2.69760400e-01 1.49961084e-01 -9.79286194e-01 -4.34997618e-01
2.16188148e-01 -5.15401721e-01 6.17483616e-01 5.73196635e-02
1.58851898e+00 -4.15596306e-01 -2.04131797e-01 4.74927068e-01
1.07140756e+00 3.76151979e-01 7.51644790e-01 1.80835634e-01
6.94560289e-01 7.10653067e-01 7.40517139e-01 5.69771588e-01
-8.14419389e-02 7.77910352e-01 6.25335157e-01 1.42740697e-01
5.56308590e-02 -6.99442923e-02 6.55266821e-01 3.57425094e-01
1.23841256e-01 -9.83099863e-02 -5.81193388e-01 1.12811867e-04
-1.69018745e+00 -1.10392833e+00 1.33558124e-01 2.30574608e+00
6.04889333e-01 1.31754175e-01 2.68417209e-01 6.93056166e-01
8.85442853e-01 3.10881943e-01 -4.04990047e-01 -1.30283728e-01
-2.43595690e-01 2.89396018e-01 7.42552802e-02 9.95581970e-02
-1.22126639e+00 8.42346549e-01 4.84542322e+00 1.08524895e+00
-1.35107708e+00 -1.04776353e-01 1.07887459e+00 1.50139317e-01
2.86347181e-01 -5.71249016e-02 -7.42143095e-01 9.52147961e-01
9.34820712e-01 1.65797323e-01 3.75459582e-01 7.64233649e-01
1.74641758e-01 -1.64724141e-01 -9.12111282e-01 1.47031927e+00
2.32786149e-01 -1.13114738e+00 3.50991189e-02 -1.17559001e-01
4.58550394e-01 -8.30476463e-01 4.16853279e-01 8.00546631e-03
-1.29560381e-01 -9.79971051e-01 4.98550087e-01 9.83411446e-02
7.62028217e-01 -8.60359311e-01 4.70047832e-01 1.68041900e-01
-1.29419577e+00 -3.61872852e-01 -3.72845531e-01 3.93021256e-01
-9.37823653e-02 3.23629141e-01 -5.44277668e-01 5.54719448e-01
7.97500789e-01 8.46402586e-01 -7.39489615e-01 7.58336544e-01
-2.54714727e-01 7.17331171e-01 -1.94305107e-01 1.42339259e-01
5.51454313e-02 -1.14305943e-01 4.48759913e-01 1.03169596e+00
2.66047239e-01 5.82545325e-02 4.93727345e-03 2.53648132e-01
-3.22625716e-03 1.48250043e-01 -5.53262353e-01 -8.85277614e-02
3.85177046e-01 1.43526912e+00 -8.00262988e-01 -8.13773274e-02
-4.24249470e-01 1.08659029e+00 -1.23566963e-01 3.22544456e-01
-1.14861047e+00 -2.05787152e-01 4.01722133e-01 -6.80027604e-02
3.93525928e-01 -2.40267634e-01 3.01305950e-01 -1.11998177e+00
4.08719331e-02 -1.10251820e+00 7.34491527e-01 -5.86798191e-01
-1.09492218e+00 8.90309453e-01 2.25783195e-02 -1.65139389e+00
-4.52002466e-01 -6.71002626e-01 -5.76230884e-01 4.60044920e-01
-1.46999335e+00 -9.21194673e-01 -5.90964437e-01 1.11128116e+00
4.87415999e-01 -4.74987537e-01 6.79561138e-01 4.98881727e-01
-9.68274534e-01 8.09941649e-01 -2.41411880e-01 3.76909643e-01
6.89748526e-01 -7.75997162e-01 6.95061758e-02 8.15231740e-01
5.56019723e-01 2.62903929e-01 4.08166736e-01 -5.21110296e-01
-1.51844656e+00 -9.49397385e-01 1.18379183e-01 -1.06564164e-01
7.54687250e-01 -1.92765564e-01 -8.95716190e-01 2.92580277e-01
1.66167095e-02 4.24526364e-01 7.29232371e-01 -5.98366916e-01
-6.54992938e-01 -2.87033617e-01 -1.10079002e+00 5.48551679e-01
7.73041248e-01 -5.14237881e-01 -8.98594856e-02 4.43131208e-01
3.64834756e-01 -2.29057878e-01 -3.62128019e-01 2.08573252e-01
3.62757623e-01 -1.08324790e+00 8.84899497e-01 -7.28190601e-01
2.42728174e-01 -3.36196274e-01 -1.51698947e-01 -8.96448016e-01
3.08370590e-01 -8.00843418e-01 -3.19866896e-01 1.33848166e+00
-7.18323365e-02 -4.08151001e-01 9.71351981e-01 2.00222701e-01
5.23636520e-01 -6.69088781e-01 -1.14440620e+00 -8.11449707e-01
-4.35896099e-01 -3.94414663e-01 4.62587208e-01 9.28050101e-01
-2.67080460e-02 -1.15110166e-01 -7.17992663e-01 3.55620146e-01
7.38984525e-01 -1.10177472e-02 7.11104214e-01 -1.00156140e+00
-2.74055898e-01 -1.13244489e-01 -8.44548881e-01 -6.64006829e-01
3.73857081e-01 -5.49932003e-01 -4.18938458e-01 -6.28285944e-01
1.94082707e-02 -1.52082771e-01 -3.36458564e-01 6.40196800e-01
-9.69965011e-02 6.52075052e-01 3.00170630e-01 3.52680892e-01
-5.52990556e-01 3.98591459e-01 7.83700705e-01 -2.26358443e-01
-1.47960946e-01 5.31023927e-03 -2.71475852e-01 6.99003994e-01
6.84384048e-01 -2.69012898e-01 -4.26600486e-01 -3.25791934e-03
-3.47899616e-01 3.55455488e-01 3.37705880e-01 -1.07704067e+00
1.00592092e-01 -6.87879473e-02 6.24427795e-01 -3.60870540e-01
4.86092299e-01 -9.87098217e-01 -6.49038479e-02 6.80607975e-01
-1.32916898e-01 2.06952825e-01 2.26532817e-01 6.88232958e-01
-3.36233228e-01 -1.25909925e-01 9.32101846e-01 2.19494760e-01
-9.13709104e-01 3.98942053e-01 -6.91419393e-02 -2.76850194e-01
1.48290670e+00 -2.37996265e-01 -3.12455416e-01 -6.96786761e-01
-8.44682395e-01 -9.00051296e-02 6.01387247e-02 5.68282306e-01
9.62837875e-01 -1.37584710e+00 -4.24773842e-01 4.79589522e-01
1.81154087e-02 -5.27051330e-01 5.71465313e-01 6.82272255e-01
-4.30542886e-01 1.21530831e-01 -4.96345758e-01 -7.23734260e-01
-1.61751306e+00 7.88116217e-01 -1.46441598e-04 -1.72989428e-01
-5.46514153e-01 5.51875591e-01 -3.30335796e-02 4.76651579e-01
5.07257164e-01 3.15702319e-01 -5.59099317e-01 2.81228691e-01
9.17354703e-01 2.93233812e-01 1.80499516e-02 -9.69998777e-01
-3.59588981e-01 4.54310596e-01 -2.13467911e-01 -1.07944742e-01
1.14536870e+00 1.23751890e-02 2.88469270e-02 -9.86835919e-03
1.40230310e+00 -5.23552336e-02 -1.50921834e+00 -1.12484366e-01
9.00809243e-02 -7.34312594e-01 -3.08844954e-01 -2.61789262e-01
-1.44226146e+00 1.02222681e+00 9.74193335e-01 2.58641213e-01
1.36764622e+00 -3.22375745e-01 8.18455994e-01 1.95919067e-01
4.94779915e-01 -6.94136143e-01 8.16616416e-01 -8.20773914e-02
6.43815517e-01 -1.23697186e+00 -1.29878372e-01 -5.15125811e-01
-8.54979038e-01 1.24084651e+00 4.22136873e-01 1.21814504e-01
6.44604385e-01 6.84862509e-02 9.06190276e-03 2.38818228e-01
-3.95984322e-01 1.97587255e-02 1.19242862e-01 5.64076602e-01
-4.87104654e-02 -3.24978590e-01 -5.42414039e-02 4.46569294e-01
9.65045914e-02 -2.11417392e-01 3.76203209e-01 9.97550428e-01
-6.22779056e-02 -1.22891521e+00 -5.76637924e-01 1.12170532e-01
-5.25471389e-01 3.03669542e-01 -2.10355252e-01 5.47005594e-01
-6.90647960e-02 1.19584191e+00 -2.41349444e-01 -7.37766147e-01
-1.53692752e-01 3.93833481e-02 3.68926972e-01 -2.27172643e-01
-2.05844268e-01 6.55826628e-02 -4.85608160e-01 -5.75463891e-01
-3.80061954e-01 -4.33314145e-01 -1.13411939e+00 -3.38595122e-01
-1.92545280e-01 1.72471091e-01 6.16528392e-01 8.40331018e-01
2.01558664e-01 1.61304861e-01 1.43890059e+00 -8.41268122e-01
-4.73834217e-01 -6.35587811e-01 -5.05372524e-01 8.44701946e-01
2.50941783e-01 -8.71083736e-01 -6.64936602e-01 1.95092335e-01] | [13.03078842163086, 1.1780937910079956] |
eec3d538-c054-4c66-a1e8-74230a73ef69 | temporally-coherent-completion-of-dynamic | null | null | https://dl.acm.org/doi/10.1145/2980179.2982398 | https://dl.acm.org/doi/pdf/10.1145/2980179.2982398 | Temporally coherent completion of dynamic video | We present an automatic video completion algorithm that synthesizes missing regions in videos in a temporally coherent fashion. Our algorithm can handle dynamic scenes captured using a moving camera. State-of-the-art approaches have difficulties handling such videos because viewpoint changes cause image-space motion vectors in the missing and known regions to be inconsistent. We address this problem by jointly estimating optical flow and color in the missing regions. Using pixel-wise forward/backward flow fields enables us to synthesize temporally coherent colors. We formulate the problem as a non-parametric patch-based optimization. We demonstrate our technique on numerous challenging videos. | ['J. Kopf', 'N. Ahuja', 'S. B. Kang', 'J.-B. Huang'] | 2016-11-01 | null | null | null | null | ['video-inpainting'] | ['computer-vision'] | [ 1.33820415e-01 -6.97728038e-01 3.82439718e-02 -9.42776650e-02
-7.28933632e-01 -8.38971734e-01 2.44446367e-01 -7.58452773e-01
-2.40245759e-01 8.38285387e-01 2.81592328e-02 1.13114685e-01
2.74810672e-01 -1.97217748e-01 -9.29405153e-01 -6.16104364e-01
-9.24310386e-02 -1.92499515e-02 2.61529833e-01 1.24318503e-01
3.29202235e-01 4.46456313e-01 -1.38540471e+00 4.78286088e-01
6.82396710e-01 5.62311769e-01 2.61424452e-01 1.11432707e+00
1.47930846e-01 1.03168917e+00 -2.25917950e-01 1.36234567e-01
8.83080423e-01 -7.04757810e-01 -7.26090550e-01 7.84321129e-01
1.31374323e+00 -8.95439088e-01 -5.91880739e-01 1.05820072e+00
-8.45179781e-02 6.46721721e-01 2.33335555e-01 -1.39758670e+00
-4.43004489e-01 -1.05900250e-01 -8.31462443e-01 2.18840212e-01
7.55766630e-01 3.14074844e-01 6.77462101e-01 -9.45158064e-01
1.15270054e+00 1.19785655e+00 3.03545624e-01 5.19840002e-01
-1.37668240e+00 -3.48094881e-01 5.55151165e-01 3.92565429e-01
-1.33578992e+00 -7.71990776e-01 1.02264953e+00 -6.46596193e-01
5.95184088e-01 2.06770435e-01 9.37453210e-01 5.98809183e-01
3.44330698e-01 7.03345478e-01 9.55122054e-01 -2.89488584e-01
1.96896851e-01 -3.93725842e-01 -5.81409991e-01 8.13853025e-01
3.27980071e-02 2.52742022e-01 -7.45392740e-01 -7.01523125e-02
1.33575296e+00 1.09786391e-01 -7.08444417e-01 -6.33148253e-01
-1.61403716e+00 5.61968207e-01 1.29952818e-01 -5.77200809e-03
-4.64934856e-01 4.36803967e-01 -1.96936220e-01 2.37774804e-01
3.35559726e-01 3.20073307e-01 -2.95696020e-01 -7.67716765e-02
-1.34820807e+00 3.51701796e-01 6.95547700e-01 1.10504377e+00
8.99412572e-01 4.68011826e-01 1.55597981e-02 4.27815467e-01
2.99868554e-01 4.31367040e-01 -2.66126730e-02 -2.02720857e+00
2.74963319e-01 -2.37288013e-01 6.78698182e-01 -1.12099659e+00
1.92337409e-02 7.92259276e-02 -4.49491262e-01 5.15889347e-01
7.14052439e-01 -2.30824232e-01 -7.58958697e-01 1.67982054e+00
6.55899525e-01 7.55443513e-01 -1.48630202e-01 1.24736452e+00
2.64589965e-01 9.99043643e-01 -4.19199347e-01 -7.34423041e-01
7.79026449e-01 -1.32482445e+00 -1.08691549e+00 -1.64825693e-01
-8.60275254e-02 -1.10740709e+00 5.45732439e-01 6.52306378e-01
-1.53899348e+00 -6.13696754e-01 -8.78121734e-01 -1.91101670e-01
3.61991137e-01 -2.21849591e-01 2.33674765e-01 1.32899821e-01
-1.35555208e+00 4.77030218e-01 -8.99420440e-01 -1.81293245e-02
1.08909741e-01 1.88640133e-01 -5.19230723e-01 -6.44987702e-01
-7.32387125e-01 5.37383795e-01 3.39332335e-02 3.37418050e-01
-1.16063046e+00 -6.73493445e-01 -1.03621578e+00 -3.06435525e-01
5.18480361e-01 -1.00438654e+00 1.10309398e+00 -1.45880580e+00
-1.59440684e+00 5.66934288e-01 -7.40823507e-01 1.46792680e-01
6.10635102e-01 -3.45060408e-01 -2.39995137e-01 6.40745819e-01
1.66797206e-01 7.49072611e-01 1.29048121e+00 -1.51207376e+00
-6.93308413e-01 2.70779971e-02 -6.72376826e-02 4.91809458e-01
2.23837271e-01 -4.52111065e-02 -8.80466640e-01 -9.15001512e-01
2.55177379e-01 -1.15497756e+00 -3.07834595e-01 6.80337250e-01
-1.74480438e-01 5.15025258e-01 1.01461065e+00 -9.05984759e-01
9.33951318e-01 -1.97710311e+00 4.59516376e-01 -1.32147431e-01
1.89658687e-01 -2.40124971e-01 -3.00298423e-01 1.37812331e-01
1.43015251e-01 -6.01397634e-01 -3.47864985e-01 -3.24525267e-01
-5.14767289e-01 3.49475324e-01 -1.26719803e-01 8.92797649e-01
3.27870026e-02 5.65275252e-01 -1.26910329e+00 -5.41456342e-01
5.78689218e-01 6.81030154e-01 -9.19416308e-01 5.01492560e-01
-3.18728477e-01 9.31584120e-01 -2.08806366e-01 7.37636387e-01
9.75972354e-01 -6.79265857e-02 1.09646440e-01 -2.73133308e-01
-4.25482690e-01 -2.68250436e-01 -1.47558689e+00 2.23831153e+00
-1.93862572e-01 1.12822533e+00 5.88314772e-01 -6.11500621e-01
3.34715605e-01 1.34166807e-01 9.31689441e-01 -2.14880392e-01
-1.35550052e-01 -9.69235450e-02 -2.92774200e-01 -6.59171820e-01
8.05529237e-01 -1.64847076e-01 4.70559299e-01 2.80494928e-01
-1.45485595e-01 -2.93709815e-01 3.82401556e-01 2.72499979e-01
9.33181524e-01 6.85879290e-01 -1.01818033e-01 3.18262763e-02
4.90532935e-01 1.39667079e-01 9.36709404e-01 5.82713783e-01
-4.13100988e-01 1.11032426e+00 3.14570814e-02 -5.15469670e-01
-9.53084230e-01 -1.26953459e+00 2.50748992e-01 7.48943031e-01
5.13888597e-01 -2.45366797e-01 -4.67147648e-01 -2.74345547e-01
-1.71343818e-01 2.68887579e-01 -5.11293054e-01 4.79406953e-01
-8.69314730e-01 -7.20044151e-02 -2.92453021e-01 2.75066078e-01
3.04941475e-01 -6.99747920e-01 -6.50521517e-01 4.79315460e-01
-6.14463031e-01 -1.59389687e+00 -1.13357544e+00 -4.28170115e-01
-9.61695313e-01 -1.19160891e+00 -9.11787570e-01 -8.16082418e-01
8.52953553e-01 9.43250000e-01 1.20016515e+00 9.62123200e-02
-4.73116100e-01 7.20900476e-01 -1.55008763e-01 5.38174212e-01
-3.30632299e-01 -8.57254446e-01 -2.99899690e-02 4.73735273e-01
-3.33099365e-01 -3.43048275e-01 -8.27122688e-01 4.59441662e-01
-1.14570463e+00 2.64462888e-01 -9.40651353e-03 7.75055945e-01
8.76593292e-01 -1.66561566e-02 -2.23595470e-01 -6.43042624e-01
-7.03626825e-03 -2.22432479e-01 -9.08262432e-01 3.24145645e-01
8.34630430e-02 -1.11402415e-01 6.50809884e-01 -5.66277146e-01
-1.18713069e+00 6.43817484e-01 4.68851805e-01 -1.04350841e+00
-7.84970820e-03 -1.17749058e-01 3.48440148e-02 -3.16729009e-01
1.90732136e-01 2.79886931e-01 -3.27814370e-02 -1.21614553e-01
6.67911828e-01 -1.04897559e-01 9.00570929e-01 -5.38735032e-01
9.91854787e-01 1.02990174e+00 2.19475757e-02 -1.07186198e+00
-5.57117283e-01 -7.60499299e-01 -8.51474524e-01 -6.34605944e-01
1.07874787e+00 -1.06954956e+00 -4.12914127e-01 4.58659261e-01
-1.33623898e+00 -4.34288591e-01 -1.73101723e-01 6.80491328e-01
-7.66878486e-01 7.70982325e-01 -6.90229654e-01 -4.79848832e-01
2.13689610e-01 -1.19062757e+00 9.93833840e-01 1.79618210e-01
6.75523207e-02 -1.04938698e+00 4.13488269e-01 1.66551188e-01
1.93555236e-01 3.86674643e-01 7.78087303e-02 6.24466121e-01
-1.22873020e+00 2.89060563e-01 -3.03837862e-02 3.08750927e-01
2.78401226e-01 4.90484029e-01 -6.66005135e-01 -4.15782243e-01
5.09772599e-02 2.31318146e-01 7.29382634e-01 9.19329107e-01
8.14512551e-01 -3.74895185e-01 -3.85104865e-02 1.03599298e+00
1.72759461e+00 1.92411333e-01 5.20364523e-01 1.73438117e-01
1.03031075e+00 6.72297239e-01 6.63401425e-01 4.58886415e-01
3.25562090e-01 9.04748201e-01 3.42895925e-01 -5.43658249e-02
-2.88419783e-01 -7.67064989e-02 5.63269675e-01 6.60865903e-01
-1.20186619e-01 -2.71353990e-01 -3.53786975e-01 6.83725476e-01
-2.00595784e+00 -1.35275805e+00 -2.19565645e-01 2.04178214e+00
6.28148079e-01 -5.29059768e-01 1.83854718e-02 -2.41240844e-01
7.82854676e-01 3.25996786e-01 -4.41113025e-01 -5.32984361e-02
-1.54820934e-01 -2.32395470e-01 3.07498366e-01 9.41290319e-01
-1.10444832e+00 8.68944108e-01 7.05353498e+00 2.97738835e-02
-9.89467680e-01 4.69861645e-03 3.78899127e-01 -4.70029652e-01
-4.46104407e-01 2.69103080e-01 -2.25936115e-01 3.93635660e-01
2.47925311e-01 6.00864142e-02 9.13795829e-01 4.11323547e-01
5.38987339e-01 -5.30468822e-01 -1.10931444e+00 1.38162982e+00
4.93968844e-01 -1.44967580e+00 -6.92921206e-02 -6.12908266e-02
1.49535584e+00 -1.85767874e-01 4.92964126e-03 -5.56242466e-01
2.11247131e-01 -6.08991027e-01 8.94810021e-01 6.46968424e-01
7.09871113e-01 -3.32084715e-01 -4.36603203e-02 -1.64148360e-01
-1.21860349e+00 1.06938936e-01 -2.79895008e-01 -1.16090879e-01
8.17437053e-01 5.57786584e-01 -2.56394655e-01 4.00367081e-01
7.38362551e-01 1.26635969e+00 -2.02011958e-01 1.27739394e+00
-1.58770263e-01 1.04309596e-01 -2.60169506e-01 7.18597233e-01
1.31808624e-01 -8.73038828e-01 7.09489286e-01 8.81458879e-01
3.99671853e-01 3.92447263e-01 4.10329372e-01 8.78670335e-01
8.07478875e-02 -2.69074410e-01 -5.91486871e-01 3.62506866e-01
1.74318627e-01 1.13874626e+00 -6.55849993e-01 -4.16275054e-01
-7.56526947e-01 1.57781684e+00 -6.88217208e-02 8.91705930e-01
-7.53778160e-01 1.08975597e-01 8.45370471e-01 7.14635924e-02
4.31622803e-01 -7.27089942e-01 1.42968222e-01 -1.75694430e+00
6.10550418e-02 -6.85579240e-01 3.29928041e-01 -1.18869078e+00
-1.06752658e+00 4.29226398e-01 -5.82973249e-02 -1.69047666e+00
-3.30670416e-01 -3.54830772e-01 -4.68506277e-01 5.68009615e-01
-1.74705601e+00 -8.11125875e-01 -6.09364152e-01 1.08763993e+00
1.00733984e+00 1.95676044e-01 3.34757477e-01 3.04792285e-01
-3.37446809e-01 -6.69701472e-02 3.62449944e-01 -6.88921008e-03
1.01299572e+00 -1.08004105e+00 1.20081142e-01 1.40247321e+00
2.68511504e-01 4.22999084e-01 8.06504190e-01 -4.47996348e-01
-1.77410090e+00 -8.87228966e-01 3.80092382e-01 -3.66568893e-01
4.00958240e-01 -6.82531521e-02 -8.07379305e-01 7.29784667e-01
4.06638026e-01 5.23050189e-01 3.15050155e-01 -6.58417344e-01
-1.95547938e-01 -9.52901617e-02 -8.97321224e-01 4.27855849e-01
9.19681847e-01 -4.58985925e-01 -2.31788918e-01 2.10539088e-01
3.37890238e-01 -6.34377658e-01 -4.79305834e-01 -1.16482578e-01
7.87194669e-01 -9.97593284e-01 1.06972063e+00 -3.83744448e-01
4.36399907e-01 -8.94997418e-01 -1.37283504e-01 -1.30828285e+00
-1.90245837e-01 -1.18691230e+00 -2.86517531e-01 6.96279764e-01
-1.20785773e-01 -2.85971105e-01 7.32405365e-01 8.42892408e-01
-2.55128052e-02 -9.23093781e-02 -8.11923563e-01 -6.27321482e-01
-3.18023980e-01 -2.71376967e-01 -9.77880880e-02 1.03301048e+00
-2.09510952e-01 -1.82266414e-01 -9.10325944e-01 1.91761449e-01
1.01912332e+00 3.95084023e-01 8.64567518e-01 -7.17117310e-01
-5.03024101e-01 -1.40028566e-01 -2.08374619e-01 -1.48124731e+00
2.05881059e-01 -1.86671838e-01 5.17588556e-01 -1.44266343e+00
1.68380290e-01 -2.19879538e-01 3.07532772e-02 8.28274265e-02
-3.01584423e-01 4.42553729e-01 4.30758893e-01 2.40396589e-01
-7.32685387e-01 3.86251599e-01 1.60543668e+00 -1.82779744e-01
-2.67830133e-01 -4.41171855e-01 -6.59191534e-02 5.66714764e-01
4.85867023e-01 -3.81528139e-01 -4.13474292e-01 -8.08077812e-01
-2.44100671e-02 6.29837751e-01 2.64976263e-01 -8.34253013e-01
2.48465732e-01 -7.35027134e-01 6.67024553e-01 -4.65646207e-01
6.92805052e-01 -9.90148187e-01 5.34349978e-01 2.38857076e-01
-1.43996179e-01 3.76088202e-01 8.80651772e-02 8.06310952e-01
-3.16713482e-01 -2.54141656e-03 8.41916203e-01 -3.87007236e-01
-9.38713074e-01 4.63074982e-01 -5.86014688e-01 1.33750275e-01
1.18470502e+00 -4.06394273e-01 -1.42440647e-01 -7.01320648e-01
-7.12851584e-01 1.72802791e-01 1.02506483e+00 5.19268870e-01
9.74644423e-01 -1.20629430e+00 -7.32741177e-01 5.48283994e-01
-2.78464675e-01 -5.17304465e-02 6.26089513e-01 8.39873552e-01
-1.22850859e+00 -2.00486798e-02 -4.88570094e-01 -8.42753291e-01
-1.41343427e+00 7.44644046e-01 4.85215485e-01 2.62477040e-01
-6.24597907e-01 7.87159383e-01 4.51478869e-01 2.34662756e-01
3.12204231e-02 -2.61428356e-01 2.70085812e-01 -3.00520092e-01
6.74978137e-01 5.41480422e-01 -4.38035399e-01 -8.53457987e-01
-3.30734402e-01 9.84872937e-01 2.16248840e-01 -5.98293006e-01
1.15604854e+00 -5.30375361e-01 -1.86228171e-01 2.89577603e-01
1.41888022e+00 1.97442234e-01 -2.15086079e+00 3.47308256e-02
-7.60312796e-01 -1.31768501e+00 1.41391382e-01 -3.80884141e-01
-1.39855480e+00 5.94882667e-01 3.69693547e-01 -1.87382817e-01
1.24732041e+00 -3.94725740e-01 9.82702672e-01 -1.94583293e-02
3.20702076e-01 -1.11876452e+00 1.45578787e-01 3.25216651e-01
6.95267856e-01 -1.30664122e+00 3.27795446e-01 -5.88310599e-01
-5.26201010e-01 1.38649464e+00 4.99116004e-01 -2.27107465e-01
4.32130992e-01 2.07407549e-01 2.01462850e-01 1.68650329e-01
-8.44579995e-01 2.78958725e-03 4.50836003e-01 5.80827951e-01
2.68728673e-01 -2.23300442e-01 -4.63626869e-02 -5.97504020e-01
4.63907957e-01 -3.71411033e-02 1.02737212e+00 1.12344444e+00
-2.00936079e-01 -1.03191876e+00 -8.38979006e-01 -6.76814467e-02
-3.28506768e-01 7.84938857e-02 2.65946612e-02 4.99806195e-01
-1.26578674e-01 1.18116271e+00 1.45130605e-01 4.81317937e-02
-3.83444875e-02 -2.62664795e-01 1.06406868e+00 -3.53501827e-01
-8.11698213e-02 6.58621609e-01 -1.28998965e-01 -1.03924012e+00
-1.00302160e+00 -8.17206323e-01 -1.15143001e+00 -2.34636113e-01
-6.58771396e-02 -3.96732353e-02 3.97881240e-01 4.79991943e-01
1.35963842e-01 3.74704748e-01 9.28492963e-01 -1.31529355e+00
1.31646648e-01 -2.59413630e-01 -6.69935882e-01 6.99683368e-01
9.09199357e-01 -5.26949525e-01 -5.32175720e-01 7.80603766e-01] | [10.67921257019043, -1.5223337411880493] |
1421d382-832e-4265-8cff-840a0ddfc910 | dissecting-self-supervised-learning-methods | 2207.00449 | null | https://arxiv.org/abs/2207.00449v3 | https://arxiv.org/pdf/2207.00449v3.pdf | Dissecting Self-Supervised Learning Methods for Surgical Computer Vision | The field of surgical computer vision has undergone considerable breakthroughs in recent years with the rising popularity of deep neural network-based methods. However, standard fully-supervised approaches for training such models require vast amounts of annotated data, imposing a prohibitively high cost; especially in the clinical domain. Self-Supervised Learning (SSL) methods, which have begun to gain traction in the general computer vision community, represent a potential solution to these annotation costs, allowing to learn useful representations from only unlabeled data. Still, the effectiveness of SSL methods in more complex and impactful domains, such as medicine and surgery, remains limited and unexplored. In this work, we address this critical need by investigating four state-of-the-art SSL methods (MoCo v2, SimCLR, DINO, SwAV) in the context of surgical computer vision. We present an extensive analysis of the performance of these methods on the Cholec80 dataset for two fundamental and popular tasks in surgical context understanding, phase recognition and tool presence detection. We examine their parameterization, then their behavior with respect to training data quantities in semi-supervised settings. Correct transfer of these methods to surgery, as described and conducted in this work, leads to substantial performance gains over generic uses of SSL - up to 7.4% on phase recognition and 20% on tool presence detection - as well as state-of-the-art semi-supervised phase recognition approaches by up to 14%. Further results obtained on a highly diverse selection of surgical datasets exhibit strong generalization properties. The code is available at https://github.com/CAMMA-public/SelfSupSurg. | ['Antoine Fleurentin', 'Saurav Sharma', 'Nicolas Padoy', 'Alexandros Karargyris', 'Georgios Exarchakis', 'Idris Hamoud', 'Chinedu Innocent Nwoye', 'Luca Sestini', 'Aditya Murali', 'Tong Yu', 'Deepak Alapatt', 'Vinkle Srivastav', 'Sanat Ramesh'] | 2022-07-01 | null | null | null | null | ['action-triplet-recognition', 'surgical-tool-detection', 'surgical-phase-recognition'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 4.63874578e-01 3.89397353e-01 -7.85234094e-01 -1.44120827e-01
-1.07489824e+00 -5.16788125e-01 4.41958785e-01 4.61829901e-01
-6.99400246e-01 5.14627695e-01 2.19159320e-01 -4.13412571e-01
-2.60575473e-01 -2.59234756e-01 -4.98747468e-01 -7.50937581e-01
-2.59539455e-01 5.35242558e-01 1.62764281e-01 -1.04748346e-01
4.51098904e-02 4.35628057e-01 -1.35165620e+00 2.99672067e-01
5.13653100e-01 9.66138244e-01 9.77955982e-02 4.86042708e-01
1.62515372e-01 6.67768240e-01 -2.15831622e-01 5.85942343e-02
2.60818303e-01 -2.44270906e-01 -8.22775006e-01 1.06606819e-02
2.33999923e-01 -1.93730846e-01 -3.72924477e-01 7.91381478e-01
8.08662355e-01 -4.27463561e-01 5.27785420e-01 -8.26776922e-01
4.60917093e-02 5.88184178e-01 -3.73330891e-01 2.92517543e-01
2.04868481e-01 2.10602909e-01 6.87030315e-01 -7.33966112e-01
7.61573374e-01 4.57610190e-01 8.85213077e-01 7.76354015e-01
-1.05501258e+00 -5.55546165e-01 -2.55995572e-01 -2.29046331e-03
-1.14185691e+00 -3.76381010e-01 7.82580614e-01 -6.00129008e-01
6.04825616e-01 2.17403457e-01 7.32452869e-01 1.23651659e+00
3.38877887e-01 1.11514711e+00 1.19922054e+00 -6.47939444e-01
2.29261562e-01 2.33465999e-01 3.00158948e-01 1.06509566e+00
4.88755822e-01 3.01994860e-01 -6.07941747e-01 -5.74620031e-02
8.36255789e-01 1.17411181e-01 -4.25427645e-01 -7.31999695e-01
-1.40610576e+00 7.96027720e-01 4.88016754e-01 5.34734130e-01
-2.65219957e-01 1.05627723e-01 6.97363496e-01 1.01974860e-01
4.66836691e-01 6.97121143e-01 -3.73452157e-01 -2.35656962e-01
-1.01902306e+00 -1.51061788e-01 1.00944543e+00 9.56954837e-01
2.81932294e-01 -2.71677226e-01 -1.27235577e-01 7.59437323e-01
8.51276070e-02 -5.56041002e-02 8.13677549e-01 -6.97232962e-01
4.42947336e-02 6.38460100e-01 -2.79408395e-01 -4.28020507e-01
-8.81804168e-01 -8.39553118e-01 -9.79345024e-01 1.86509147e-01
4.28618163e-01 -3.99886742e-02 -1.15458405e+00 1.44244695e+00
6.82846010e-02 -1.18616065e-02 -8.22703689e-02 6.44143283e-01
1.04665327e+00 -2.49314800e-01 -1.88480318e-02 -4.02539492e-01
1.20150018e+00 -1.03314567e+00 -6.75576866e-01 -4.65249687e-01
1.12125206e+00 -6.08996093e-01 8.16424131e-01 5.41953564e-01
-8.38072479e-01 -1.50366053e-01 -9.80315864e-01 2.45735526e-01
-1.31560460e-01 4.50199097e-01 9.82202411e-01 6.34503722e-01
-9.59501028e-01 6.23996735e-01 -1.31522763e+00 -5.94252884e-01
7.69863665e-01 7.84444153e-01 -6.06659949e-01 -3.00276667e-01
-6.61879420e-01 1.01998055e+00 2.52345204e-01 1.62706256e-01
-1.17721438e+00 -7.57533848e-01 -9.06329989e-01 -3.29920501e-01
6.15820169e-01 -5.86849570e-01 1.43078935e+00 -8.05003345e-01
-1.15011466e+00 1.41933191e+00 1.47669733e-01 -5.91763914e-01
6.36969149e-01 -2.83406645e-01 -9.53995734e-02 2.00817063e-01
-6.78692311e-02 5.49969733e-01 3.76379818e-01 -1.12093723e+00
-2.72309154e-01 -4.65317130e-01 -4.71782461e-02 1.29698053e-01
-3.98584723e-01 -1.29493788e-01 -5.55112362e-01 -5.69516242e-01
1.27580002e-01 -1.30072618e+00 -5.75236738e-01 5.31248450e-01
-4.56420362e-01 5.49637564e-02 2.28535607e-01 -4.91527766e-01
1.12154841e+00 -2.16346264e+00 1.35962784e-01 -3.79946232e-02
2.89353728e-01 3.12208414e-01 1.90519035e-01 3.34975153e-01
-3.03961247e-01 -1.30170211e-01 -5.58127701e-01 -4.63313490e-01
-4.17549998e-01 5.30519664e-01 4.80483383e-01 8.41280580e-01
1.18367067e-02 8.00265133e-01 -1.00962222e+00 -7.34046876e-01
3.23767692e-01 1.59735009e-01 -6.79523468e-01 1.88257441e-01
2.51284480e-01 6.84953034e-01 -1.91407129e-01 8.95682454e-01
5.62868230e-02 -3.33404243e-01 3.57014447e-01 -2.68748254e-01
3.59176248e-02 -1.15095988e-01 -6.97314143e-01 2.33132625e+00
-5.89681745e-01 6.14356399e-01 5.79091534e-02 -1.23067808e+00
4.83148068e-01 3.71242583e-01 9.86645937e-01 -3.84244025e-01
4.48906273e-01 4.57881868e-01 3.12939525e-01 -6.01513088e-01
1.13807246e-02 -3.17285836e-01 6.75574541e-02 -9.06226039e-02
5.03004968e-01 1.76828038e-02 1.48536667e-01 2.89191119e-02
1.38083029e+00 -2.92895474e-02 6.94369674e-01 -4.61261541e-01
1.92730173e-01 3.84385735e-01 3.74498993e-01 6.57591939e-01
-4.54903245e-01 7.30571747e-01 4.96752590e-01 -2.64137685e-01
-5.94708622e-01 -9.07375515e-01 -4.12999839e-01 5.70511878e-01
1.44397750e-01 -3.29272509e-01 -6.08186364e-01 -8.81699264e-01
-3.95448543e-02 3.00520033e-01 -1.01920366e+00 -5.07982969e-01
-6.08562946e-01 -7.58924484e-01 3.35083455e-01 7.16657281e-01
4.03772108e-02 -1.04001558e+00 -7.78807163e-01 1.16859682e-01
9.98787433e-02 -1.27104795e+00 1.91518827e-03 6.67330146e-01
-1.20883965e+00 -1.41667187e+00 -9.95159149e-01 -9.14110363e-01
9.15315688e-01 4.62890491e-02 1.02397919e+00 1.03806354e-01
-9.16577935e-01 5.98962367e-01 -4.56192166e-01 -6.08646095e-01
-5.24807692e-01 2.18820810e-01 4.57172915e-02 -2.41574213e-01
1.54496610e-01 -2.66693205e-01 -7.82380939e-01 1.99844673e-01
-8.80069613e-01 1.59828961e-01 9.90453243e-01 1.18180895e+00
6.24181449e-01 -6.37555897e-01 2.34526187e-01 -1.32239056e+00
2.22067371e-01 -4.84827131e-01 -2.43333578e-01 7.58714741e-03
-7.98606217e-01 -5.77520952e-03 2.29801521e-01 -4.21938211e-01
-5.64127743e-01 3.19980890e-01 -1.65651664e-01 -5.29540360e-01
-2.43565261e-01 7.05596030e-01 3.75835180e-01 -3.66323173e-01
1.04131901e+00 1.50483638e-01 5.12001216e-01 -4.32123482e-01
8.03290009e-02 5.66851437e-01 5.32503188e-01 -2.99222291e-01
6.21994495e-01 6.38994455e-01 1.48086578e-01 -7.59762347e-01
-1.08856642e+00 -9.54728365e-01 -5.26959002e-01 -1.74083754e-01
6.44020677e-01 -8.40351164e-01 -3.59509945e-01 4.45593566e-01
-7.11218476e-01 -4.56683189e-01 -5.45152605e-01 6.66902006e-01
-7.91610837e-01 3.34499568e-01 -5.78860700e-01 -4.96376902e-01
-5.72432637e-01 -1.48311400e+00 1.19632792e+00 1.35674655e-01
-3.38664263e-01 -1.05604362e+00 2.00060666e-01 3.90080005e-01
3.98653835e-01 4.64444667e-01 6.35770977e-01 -9.12475944e-01
-2.24303141e-01 -6.12986624e-01 8.70140046e-02 4.68046010e-01
2.68820971e-01 -4.39258784e-01 -8.64055216e-01 -5.08918345e-01
-2.09213212e-01 -3.52280200e-01 8.74923587e-01 5.02112389e-01
1.18043256e+00 1.28459916e-01 -6.65480733e-01 6.91159129e-01
1.40425825e+00 -6.91273063e-02 3.51992100e-01 3.06729734e-01
4.74048555e-01 3.78256321e-01 7.69949019e-01 3.30708683e-01
-6.31151497e-02 5.25269449e-01 8.15692961e-01 -4.14862633e-01
-3.93578976e-01 4.30632755e-03 -1.19035639e-01 7.46584415e-01
-2.77635992e-01 2.50204414e-01 -1.30399179e+00 7.32574582e-01
-1.66523266e+00 -4.51712698e-01 1.42163321e-01 2.20949221e+00
9.54826653e-01 2.98093438e-01 -6.44488037e-02 3.89419049e-01
2.58828163e-01 6.30124239e-03 -4.29245144e-01 1.49607196e-01
3.19188386e-01 4.95520741e-01 7.45634258e-01 -4.72898372e-02
-1.18225753e+00 6.72809422e-01 6.06976938e+00 7.70957112e-01
-1.16251838e+00 2.85869211e-01 5.30699611e-01 -1.15558609e-01
3.46775204e-01 -1.57118887e-01 -5.35424173e-01 1.08139828e-01
7.76325941e-01 5.18952496e-02 6.56292401e-03 1.13382685e+00
8.30403417e-02 -4.30366367e-01 -1.41689003e+00 1.20151031e+00
2.35228568e-01 -1.60397303e+00 -3.94797415e-01 -1.26022855e-02
5.01035988e-01 2.86664456e-01 -1.30064443e-01 2.89103299e-01
-1.40731558e-01 -1.00989962e+00 1.70866624e-01 2.57549703e-01
1.15865743e+00 -2.54810750e-01 1.10324717e+00 3.45270932e-01
-7.24842429e-01 -1.43176660e-01 7.41343424e-02 3.13200265e-01
-1.79643169e-01 4.05028671e-01 -1.03430748e+00 5.77717543e-01
5.93754709e-01 8.73961926e-01 -5.00842333e-01 1.39569557e+00
-1.53670222e-01 7.68716872e-01 -2.84180015e-01 1.46360233e-01
1.56634733e-01 4.39868331e-01 4.11592245e-01 1.25217211e+00
2.42088363e-02 -1.14450492e-01 1.50211453e-01 3.04879278e-01
-1.31625803e-02 4.71229851e-02 -5.67475379e-01 -7.48917088e-02
-5.49079627e-02 1.41411102e+00 -7.62683034e-01 -1.24313660e-01
-3.02800745e-01 7.02052891e-01 1.92667216e-01 9.54679586e-03
-6.60701692e-01 -3.80987599e-02 2.86726058e-01 4.44578081e-01
-1.81211442e-01 -2.37297446e-01 -2.78612405e-01 -1.01358235e+00
-4.11930941e-02 -8.85915697e-01 5.63220799e-01 -3.54386687e-01
-8.89286995e-01 5.93748569e-01 3.54513968e-03 -1.75472081e+00
-2.44618431e-01 -9.95934010e-01 -2.57645160e-01 3.65060925e-01
-1.65166461e+00 -1.09829044e+00 -7.57656455e-01 5.38716078e-01
7.90781856e-01 -2.31555060e-01 1.24288702e+00 2.02476978e-01
-3.84216845e-01 6.93504989e-01 1.02854915e-01 3.88885677e-01
9.11770344e-01 -1.12747359e+00 -1.60168529e-01 4.54204440e-01
1.12586640e-01 5.09019017e-01 6.36669695e-01 -1.86136529e-01
-1.64212096e+00 -8.25537682e-01 2.86510795e-01 -4.48568135e-01
7.29350865e-01 -2.17168391e-01 -5.43194532e-01 6.58514738e-01
-9.11525264e-02 5.46556354e-01 9.67282712e-01 8.07763711e-02
-6.86857030e-02 -1.87477786e-02 -1.05433071e+00 4.31373656e-01
1.19886243e+00 -3.27593744e-01 -4.42445278e-01 7.09367216e-01
3.99062574e-01 -7.91450262e-01 -1.10945964e+00 9.12953317e-01
4.37955290e-01 -9.49496508e-01 7.40680993e-01 -6.07191622e-01
6.04280353e-01 1.41274869e-01 9.61191282e-02 -1.21679521e+00
2.90857013e-02 -5.11479378e-01 -1.65079266e-01 4.81123447e-01
4.56560701e-01 -4.12467092e-01 1.10249138e+00 2.11471424e-01
-5.64153254e-01 -1.39246154e+00 -1.00285292e+00 -6.76689088e-01
3.98018733e-02 -3.98673236e-01 -4.44187373e-01 9.45637882e-01
3.57415289e-01 1.24175381e-02 -1.19834043e-01 -5.56433201e-02
6.16201937e-01 -7.82387108e-02 5.73535442e-01 -1.16062760e+00
-4.72637475e-01 -4.05383855e-01 -7.58647144e-01 -5.36386967e-01
-8.47198889e-02 -1.08741009e+00 1.69325769e-01 -1.76125717e+00
3.30598354e-01 -4.16188717e-01 -4.88456994e-01 6.53069913e-01
-5.91659453e-03 3.29273522e-01 1.02995291e-01 3.78819674e-01
-5.46782196e-01 1.20178044e-01 1.13009369e+00 -1.98804319e-01
-2.17533782e-02 2.48829365e-01 -4.80340123e-01 8.99578214e-01
6.43166244e-01 -5.37842393e-01 -2.24054053e-01 -1.81618005e-01
-1.73209101e-01 2.29519054e-01 3.26440066e-01 -1.32745492e+00
3.15582544e-01 2.21531823e-01 1.21836282e-01 -1.05800107e-01
3.46357375e-01 -8.78574908e-01 -4.70447950e-02 9.75314736e-01
-3.37783664e-01 -4.21907961e-01 3.41480404e-01 4.75263059e-01
-4.86414880e-01 -2.48581022e-01 9.40761030e-01 -3.15279186e-01
-7.61862338e-01 4.15100008e-01 -2.18611658e-01 6.14462793e-02
1.16590154e+00 -3.54839712e-01 -2.00698540e-01 -1.05688266e-01
-1.10627913e+00 5.84571026e-02 2.51476735e-01 4.60315138e-01
4.88662273e-01 -8.91211987e-01 -6.48971856e-01 1.46130361e-02
5.45589387e-01 2.18472943e-01 3.46299112e-01 1.44458961e+00
-5.94367862e-01 4.65333670e-01 -1.79461911e-01 -8.48411500e-01
-1.31216681e+00 5.50671041e-01 4.21896309e-01 -6.31616056e-01
-7.25381911e-01 7.19707072e-01 1.51108786e-01 -2.68798769e-01
5.10021985e-01 -3.50697637e-01 -2.29095280e-01 -6.47002459e-02
1.90484494e-01 -6.98881373e-02 3.43179017e-01 -2.02418759e-01
-6.14383757e-01 5.31688571e-01 -1.75657570e-01 3.50632429e-01
1.34272623e+00 6.37457788e-01 2.79550582e-01 3.76602501e-01
1.23681784e+00 -3.74497771e-01 -7.95027316e-01 -3.05568486e-01
1.87948510e-01 1.24938898e-02 2.14763120e-01 -9.51222360e-01
-1.13389790e+00 8.53231370e-01 9.49129820e-01 -2.44353250e-01
1.16998720e+00 3.40657234e-01 3.95995408e-01 2.31992915e-01
7.07810283e-01 -6.41149282e-01 1.65561453e-01 1.85286760e-01
7.78970778e-01 -1.64054775e+00 9.96264815e-02 -6.66740894e-01
-5.17977655e-01 1.01468670e+00 4.11772192e-01 -1.95191786e-01
6.78878903e-01 4.98935848e-01 1.23031139e-01 -2.46438965e-01
-4.00124192e-01 -2.80616909e-01 2.53856272e-01 4.22854185e-01
6.86493754e-01 4.22725603e-02 -5.31246722e-01 6.28094733e-01
1.67593554e-01 3.89496446e-01 5.13237834e-01 1.43803012e+00
-1.59056619e-01 -9.12550092e-01 1.07508734e-01 7.31920838e-01
-6.31163061e-01 -1.64057508e-01 -1.12978242e-01 1.12556434e+00
-2.18804404e-01 5.11558771e-01 -3.16021383e-01 -2.28207588e-01
4.96639937e-01 -1.26075581e-01 5.67743123e-01 -9.28579092e-01
-7.71179080e-01 9.78993550e-02 3.75692636e-01 -7.80604601e-01
-6.10472023e-01 -5.30715823e-01 -1.26726890e+00 5.10961652e-01
-4.44000423e-01 -9.75633338e-02 7.78155267e-01 8.08034241e-01
7.56418779e-02 8.30986321e-01 2.88787603e-01 -8.73921514e-01
-7.49886990e-01 -1.00513673e+00 -4.36948240e-01 3.01038444e-01
2.67845631e-01 -8.67477059e-01 -4.44279611e-01 -1.50152907e-01] | [14.193161964416504, -3.167637586593628] |
975a6f74-f453-4a04-919e-0a48db535626 | large-scale-adversarial-representation | 1907.02544 | null | https://arxiv.org/abs/1907.02544v2 | https://arxiv.org/pdf/1907.02544v2.pdf | Large Scale Adversarial Representation Learning | Adversarially trained generative models (GANs) have recently achieved compelling image synthesis results. But despite early successes in using GANs for unsupervised representation learning, they have since been superseded by approaches based on self-supervision. In this work we show that progress in image generation quality translates to substantially improved representation learning performance. Our approach, BigBiGAN, builds upon the state-of-the-art BigGAN model, extending it to representation learning by adding an encoder and modifying the discriminator. We extensively evaluate the representation learning and generation capabilities of these BigBiGAN models, demonstrating that these generation-based models achieve the state of the art in unsupervised representation learning on ImageNet, as well as in unconditional image generation. Pretrained BigBiGAN models -- including image generators and encoders -- are available on TensorFlow Hub (https://tfhub.dev/s?publisher=deepmind&q=bigbigan). | ['Jeff Donahue', 'Karen Simonyan'] | 2019-07-04 | large-scale-adversarial-representation-1 | http://papers.nips.cc/paper/9240-large-scale-adversarial-representation-learning | http://papers.nips.cc/paper/9240-large-scale-adversarial-representation-learning.pdf | neurips-2019-12 | ['self-supervised-image-classification'] | ['computer-vision'] | [ 4.42190558e-01 6.82467222e-01 -9.06317979e-02 -3.97846192e-01
-8.49506795e-01 -5.86747825e-01 1.25972462e+00 -6.26952291e-01
6.13302961e-02 7.88677752e-01 6.43873155e-01 -3.53568494e-01
4.14347589e-01 -1.10777152e+00 -9.15561438e-01 -6.67979777e-01
1.90419912e-01 5.49601734e-01 -3.45758229e-01 -3.60418558e-01
-6.42935783e-02 1.58508390e-01 -1.31470501e+00 3.37488979e-01
7.27022409e-01 6.56182468e-01 -3.14269476e-02 9.20972705e-01
1.38723031e-01 1.31709588e+00 -6.94658518e-01 -5.90799689e-01
4.25490528e-01 -1.13331318e+00 -8.00885797e-01 -5.07485941e-02
3.79083395e-01 -6.09516740e-01 -6.52252018e-01 7.73943424e-01
5.86833298e-01 4.87738289e-02 7.50797093e-01 -1.31835508e+00
-1.23199916e+00 9.10045862e-01 -2.12590441e-01 7.76072070e-02
1.35841910e-02 5.45111895e-01 1.00467277e+00 -5.01161158e-01
9.10481274e-01 1.24335146e+00 4.58018959e-01 1.22453082e+00
-1.46923053e+00 -8.51650357e-01 -2.52598494e-01 -2.59199381e-01
-1.04517627e+00 -7.93374300e-01 6.35796428e-01 -4.49734807e-01
1.06121624e+00 8.58451705e-03 4.53182757e-01 1.51464391e+00
-1.39069585e-02 7.75418758e-01 1.11754549e+00 -4.47686672e-01
7.01619312e-02 -6.59134760e-02 -5.93322754e-01 6.66811109e-01
1.80247113e-01 5.34398317e-01 -3.08170766e-01 3.56575668e-01
1.31278348e+00 -1.10182554e-01 -1.65914103e-01 -8.78371820e-02
-1.18233585e+00 1.06571484e+00 7.66607940e-01 1.62979171e-01
-3.95193815e-01 1.02248359e+00 3.36706549e-01 4.45198745e-01
5.16778886e-01 5.90945661e-01 -5.78756630e-03 -1.82066441e-01
-9.52303648e-01 4.02729392e-01 6.04821086e-01 1.27859688e+00
8.89295578e-01 9.87207770e-01 -4.19542611e-01 5.61203539e-01
7.41572827e-02 3.79602939e-01 7.96085596e-01 -1.11195695e+00
3.38846982e-01 3.09356689e-01 -3.46416801e-01 -5.32077491e-01
2.58343220e-01 -4.25260663e-01 -9.13678527e-01 4.69662964e-01
1.05897315e-01 -4.06864882e-01 -1.45141637e+00 1.95571589e+00
-2.39502296e-01 1.63838774e-01 2.03143746e-01 4.36439395e-01
9.01432335e-01 7.63667524e-01 1.90788701e-01 3.40605766e-01
8.31971526e-01 -1.16697907e+00 -4.14828300e-01 -1.56122386e-01
5.71234107e-01 -6.78143799e-01 7.54887938e-01 1.73670441e-01
-1.35037112e+00 -8.32463145e-01 -1.01153755e+00 -1.04692005e-01
-2.62501717e-01 -2.12620884e-01 8.24483514e-01 8.05942178e-01
-1.45725000e+00 7.02289224e-01 -7.58796036e-01 -2.45502487e-01
8.18503857e-01 1.98368087e-01 -5.37955701e-01 -2.01233387e-01
-9.81744349e-01 7.60647833e-01 4.28613722e-01 -2.84766912e-01
-1.79274845e+00 -7.35663533e-01 -1.03602624e+00 -8.68050456e-02
-8.20590556e-02 -1.10765529e+00 1.36612701e+00 -1.28301060e+00
-1.62011063e+00 8.66443157e-01 1.58394009e-01 -1.03694952e+00
4.83060658e-01 -5.83809316e-02 -3.88707906e-01 1.37050614e-01
1.47876531e-01 1.35556173e+00 1.12637448e+00 -1.38375711e+00
-2.31815591e-01 2.92004853e-01 2.40881518e-01 -3.11203487e-02
-4.18070070e-02 -1.43791631e-01 1.11699738e-02 -9.58173335e-01
-4.38203454e-01 -9.32451248e-01 -3.97318065e-01 -4.48921084e-01
-5.71802139e-01 1.37857035e-01 7.28570163e-01 -6.33143544e-01
6.55714154e-01 -1.96082568e+00 -1.97636820e-02 1.61631126e-02
1.88345820e-01 3.05132359e-01 -5.48794866e-01 6.05075538e-01
-3.70397449e-01 4.96690750e-01 -2.38614574e-01 -6.24801159e-01
1.50833756e-01 3.67653161e-01 -6.69532359e-01 1.74618110e-01
5.78126192e-01 1.42729414e+00 -9.45212424e-01 -1.93705782e-01
2.57957369e-01 7.97194421e-01 -7.15136707e-01 4.39225554e-01
-4.30560589e-01 7.32182980e-01 -2.05170050e-01 4.43134904e-01
4.81570870e-01 -3.26698184e-01 4.39743325e-03 1.11673117e-01
2.23178670e-01 5.12594104e-01 -5.00501871e-01 1.94189382e+00
-5.83969951e-01 7.68153608e-01 -2.34044760e-01 -1.02407455e+00
9.97401536e-01 5.54790080e-01 2.48610407e-01 -7.02384830e-01
7.00422451e-02 1.31639421e-01 1.55493632e-01 2.03367099e-02
4.34185058e-01 -3.18018019e-01 1.05109401e-01 6.88917458e-01
8.67038012e-01 -5.78858137e-01 3.51269782e-01 5.04510403e-01
1.11027253e+00 7.02812612e-01 3.97154838e-02 -1.95731726e-02
1.34680063e-01 -9.36263576e-02 2.24608600e-01 7.57467985e-01
2.91030139e-01 1.03390515e+00 5.69489241e-01 -1.71596065e-01
-1.51258171e+00 -1.13468945e+00 2.48048738e-01 7.83393145e-01
-5.45477748e-01 -6.72658086e-01 -1.00734627e+00 -6.31685495e-01
-2.01959789e-01 9.52365935e-01 -9.37094569e-01 -3.81055892e-01
-5.64535916e-01 -5.02737045e-01 8.39066625e-01 7.22515881e-01
7.18057096e-01 -1.45370543e+00 -3.79093230e-01 2.20146418e-01
2.15379804e-01 -8.86695385e-01 -2.47441560e-01 5.27006425e-02
-7.74959683e-01 -6.71997666e-01 -8.35895836e-01 -5.08236766e-01
8.79006326e-01 -6.05503768e-02 1.55708063e+00 5.40406294e-02
-3.98016602e-01 5.90542257e-01 -4.86593723e-01 -5.48460901e-01
-1.05431771e+00 2.76913643e-01 -4.07747716e-01 -2.81711340e-01
-1.16687156e-01 -8.61116350e-01 -8.33507717e-01 -2.59897318e-02
-1.19191813e+00 3.77272099e-01 5.86760104e-01 8.77468348e-01
5.23755848e-01 -2.33791724e-01 7.83551931e-01 -1.32595301e+00
5.31790912e-01 -5.64459801e-01 -5.06526947e-01 -2.21717566e-01
-9.56541419e-01 1.90856785e-01 6.10834897e-01 -1.10453535e-02
-1.19754028e+00 -2.01941580e-01 -4.47447181e-01 -5.50082505e-01
-2.63555259e-01 2.91108876e-01 1.63919013e-02 5.49210012e-02
8.07588398e-01 3.61855447e-01 2.40986243e-01 -2.85059273e-01
9.68301833e-01 2.63205200e-01 6.29487336e-01 -6.37351274e-01
1.06518495e+00 4.55321878e-01 -5.40189222e-02 -3.88741106e-01
-5.58570445e-01 1.66327834e-01 -3.45398813e-01 5.49988523e-02
8.43755662e-01 -1.34187305e+00 -1.60902143e-01 3.74684066e-01
-9.44376230e-01 -8.52110565e-01 -1.00057447e+00 1.08660273e-01
-8.91859353e-01 -2.25327715e-01 -6.68986142e-01 -4.37145203e-01
-5.55978179e-01 -1.07913435e+00 8.12587380e-01 1.90475449e-01
-1.49850473e-01 -1.13540125e+00 4.12344754e-01 2.65353352e-01
8.98125529e-01 6.67704940e-01 5.71236312e-01 -4.68520433e-01
-7.98904181e-01 -9.52859297e-02 -8.95277858e-02 8.65839720e-01
1.12393461e-01 -1.13840528e-01 -1.13773835e+00 -3.44748974e-01
-3.29317421e-01 -5.11399090e-01 1.22158360e+00 3.74700159e-01
1.04357123e+00 -4.82960820e-01 4.56427373e-02 1.14959025e+00
1.52870512e+00 7.23075271e-02 1.15295362e+00 1.64563388e-01
8.79239798e-01 2.00823963e-01 -1.56584769e-01 2.70176262e-01
1.44104391e-01 2.88453192e-01 5.72337568e-01 -3.13778102e-01
-8.86958838e-01 -7.23679125e-01 4.30173874e-01 7.58737266e-01
-4.63928193e-01 -3.81624520e-01 -6.24425530e-01 6.50146008e-01
-1.65421975e+00 -1.23911178e+00 3.18727016e-01 1.72440267e+00
9.51750875e-01 3.64704663e-03 1.98416729e-02 -1.59656778e-01
3.77025694e-01 4.93805945e-01 -5.16626596e-01 -4.95229810e-01
-2.74970263e-01 9.23129082e-01 5.54685831e-01 3.54595929e-01
-8.72412503e-01 1.17701137e+00 6.70950747e+00 7.04561830e-01
-1.07377183e+00 3.10586691e-01 9.02695775e-01 5.77305481e-02
-7.16974258e-01 2.16215953e-01 -6.63837433e-01 3.48111659e-01
1.16937244e+00 -2.82791227e-01 5.98859549e-01 9.21184540e-01
-2.85703391e-01 3.61591876e-01 -1.08136451e+00 7.73847938e-01
2.16936529e-01 -1.78208172e+00 4.52128828e-01 2.25409016e-01
1.30557132e+00 2.70849407e-01 4.15475667e-01 4.87164110e-01
9.95087385e-01 -1.44727218e+00 7.05346882e-01 3.66971254e-01
1.26053584e+00 -6.72750711e-01 4.50076818e-01 -1.59176484e-01
-8.04124236e-01 1.27283186e-01 -4.01118994e-01 -1.38490358e-02
9.02985111e-02 3.67373079e-01 -8.81773770e-01 4.26723033e-01
2.88590550e-01 9.67726648e-01 -5.27586043e-01 5.49961507e-01
-9.02134001e-01 1.02760804e+00 1.77646652e-01 5.34087181e-01
2.53720850e-01 -1.43470377e-01 2.65383691e-01 1.07844591e+00
4.07116950e-01 -9.71449912e-02 -2.79208899e-01 1.38084424e+00
-6.21124744e-01 -3.12992245e-01 -9.47911739e-01 -5.44993043e-01
1.97063655e-01 1.33203232e+00 -4.40679491e-01 -5.24986744e-01
-2.41829425e-01 9.69916642e-01 2.24820018e-01 5.14673531e-01
-7.65098989e-01 -4.01659131e-01 5.66740036e-01 2.67873883e-01
4.94302303e-01 -1.59447268e-01 -1.31666541e-01 -1.21772814e+00
-5.28554320e-01 -1.14278734e+00 2.08686426e-01 -9.89099324e-01
-1.09676886e+00 7.94239342e-01 -5.44861928e-02 -1.11526167e+00
-1.00832105e+00 -4.32165772e-01 -8.17391038e-01 9.61504221e-01
-1.37510121e+00 -1.73751378e+00 -2.68322557e-01 7.33654380e-01
4.41152185e-01 -6.38894737e-01 1.18515694e+00 6.55977055e-02
-2.38303468e-01 6.59845352e-01 -6.48011938e-02 4.06291515e-01
4.60691720e-01 -1.29841316e+00 1.00118792e+00 1.02510667e+00
4.47293550e-01 6.18300021e-01 5.12011349e-01 -5.43220639e-01
-1.28195572e+00 -1.34082150e+00 4.42967564e-01 -5.01521289e-01
5.09944499e-01 -6.04188323e-01 -4.12709653e-01 1.26525581e+00
8.38554740e-01 1.64806500e-01 6.14059806e-01 -2.22620219e-01
-6.80395305e-01 4.58708443e-02 -1.04994380e+00 4.72543657e-01
1.03443682e+00 -6.33350015e-01 -2.02882871e-01 2.18525112e-01
8.78578126e-01 -2.59694934e-01 -7.98417687e-01 2.24130616e-01
3.37509841e-01 -9.90705311e-01 1.03566086e+00 -6.35937631e-01
1.18508744e+00 -1.14228807e-01 -7.62694031e-02 -1.56218326e+00
-3.93023610e-01 -1.08331800e+00 -2.52287567e-01 1.38209343e+00
3.80116642e-01 -7.32996285e-01 8.05481732e-01 2.45756701e-01
-4.02615398e-01 -5.16112924e-01 -5.61778128e-01 -8.01504195e-01
3.32367033e-01 -1.57604069e-01 9.31822956e-01 7.59221554e-01
-3.92968804e-01 3.74044001e-01 -5.16017139e-01 -4.39542651e-01
6.17217422e-01 -6.10532127e-02 1.13550842e+00 -6.20585382e-01
-6.17915750e-01 -4.89945084e-01 -5.16852856e-01 -7.35313773e-01
2.04137072e-01 -1.13769329e+00 -1.64651081e-01 -1.71580446e+00
3.15094031e-02 -3.68754119e-01 -3.00275862e-01 7.43856132e-01
9.47742388e-02 7.21958458e-01 4.30531025e-01 2.56948918e-01
-2.58015543e-01 5.49265981e-01 1.38579452e+00 -2.13705570e-01
2.58304954e-01 -2.93009579e-01 -1.07928586e+00 2.80600548e-01
9.88737702e-01 -2.93176204e-01 -8.80941689e-01 -6.57427609e-01
2.99828738e-01 -3.06731045e-01 5.37010252e-01 -1.25876999e+00
-2.18598887e-01 1.70072913e-01 5.57191670e-01 -8.91868398e-02
2.90805191e-01 -2.10134596e-01 5.23209393e-01 3.42320621e-01
-5.46586454e-01 -2.48307511e-02 1.48192734e-01 1.70009300e-01
-3.44756812e-01 -1.27408445e-01 7.38294005e-01 -5.29117584e-01
-6.62999034e-01 4.71571863e-01 -6.26998171e-02 2.96818465e-01
8.19288075e-01 -4.42024171e-02 -7.22864926e-01 -8.43667686e-01
-5.73612154e-01 -3.09547246e-01 5.58147669e-01 3.32440376e-01
6.62657738e-01 -1.44532561e+00 -1.11490011e+00 4.36550468e-01
1.27153108e-02 9.90403295e-02 1.14531204e-01 3.02931070e-01
-6.19011283e-01 2.08857626e-01 -5.69555640e-01 -1.89049199e-01
-5.18278420e-01 4.46763426e-01 2.05671296e-01 -3.80631506e-01
-5.35050392e-01 1.14596522e+00 4.78574872e-01 -3.08908254e-01
-3.42814952e-01 1.38021156e-01 2.11462766e-01 -3.69770378e-01
4.58748996e-01 1.17213309e-01 -1.53411165e-01 -6.34041488e-01
2.01672330e-01 3.58440056e-02 -1.57453045e-01 -3.35452795e-01
1.51784587e+00 3.42400759e-01 -6.17304929e-02 2.57077534e-02
1.09751129e+00 -2.35698327e-01 -1.50294483e+00 1.68271676e-01
-6.86019361e-01 -2.82015413e-01 -1.27746947e-02 -8.89043033e-01
-1.47949779e+00 1.13168406e+00 3.69203568e-01 3.08894534e-02
1.05055141e+00 -9.55101196e-03 8.41042280e-01 3.59744346e-03
3.43277454e-01 -6.85078561e-01 2.37741426e-01 4.13464516e-01
1.05742598e+00 -1.10297132e+00 3.34613882e-02 -3.00684404e-02
-6.52308643e-01 7.25290656e-01 5.29129207e-01 -6.77232563e-01
5.29793620e-01 2.93462634e-01 9.89482924e-02 -1.00971825e-01
-8.51787806e-01 -1.95528522e-01 5.12589738e-02 1.04209960e+00
7.08887160e-01 1.54919401e-01 1.76063851e-01 1.55031249e-01
-6.53640747e-01 8.05626437e-02 5.67004621e-01 8.43426108e-01
1.80240035e-01 -1.44233978e+00 2.11288072e-02 5.05776346e-01
-4.87988144e-01 -4.82114524e-01 -1.57544166e-01 7.80487597e-01
1.80240616e-01 7.04629362e-01 1.66067943e-01 -3.94853860e-01
-5.32396436e-02 -7.67375901e-02 7.47879267e-01 -7.97892630e-01
-7.29940414e-01 -2.50293519e-02 1.20102860e-01 -5.42702317e-01
-4.09508944e-01 -4.39443052e-01 -1.02136672e+00 -3.53926748e-01
2.11184393e-04 5.81932720e-03 7.40992129e-01 4.22107726e-01
6.14805281e-01 8.80934417e-01 5.65161407e-01 -1.01028919e+00
-5.11224747e-01 -1.12372482e+00 -5.56200266e-01 6.10921323e-01
1.72123656e-01 -1.76366776e-01 -1.59233242e-01 5.74660659e-01] | [11.549504280090332, -0.20997492969036102] |
0fec7b75-d5fc-4268-b09b-b91880264337 | high-fidelity-synthetic-face-generation-for | 2303.04839 | null | https://arxiv.org/abs/2303.04839v1 | https://arxiv.org/pdf/2303.04839v1.pdf | High Fidelity Synthetic Face Generation for Rosacea Skin Condition from Limited Data | Similar to the majority of deep learning applications, diagnosing skin diseases using computer vision and deep learning often requires a large volume of data. However, obtaining sufficient data for particular types of facial skin conditions can be difficult due to privacy concerns. As a result, conditions like Rosacea are often understudied in computer-aided diagnosis. The limited availability of data for facial skin conditions has led to the investigation of alternative methods for computer-aided diagnosis. In recent years, Generative Adversarial Networks (GANs), mainly variants of StyleGANs, have demonstrated promising results in generating synthetic facial images. In this study, for the first time, a small dataset of Rosacea with 300 full-face images is utilized to further investigate the possibility of generating synthetic data. The preliminary experiments show how fine-tuning the model and varying experimental settings significantly affect the fidelity of the Rosacea features. It is demonstrated that $R_1$ Regularization strength helps achieve high-fidelity details. Additionally, this study presents qualitative evaluations of synthetic/generated faces by expert dermatologists and non-specialist participants. The quantitative evaluation is presented using a few validation metric(s). Furthermore a number of limitations and future directions are discussed. Code and generated dataset are available at: \url{https://github.com/thinkercache/stylegan2-ada-pytorch} | ['Hossein Javidnia', 'Marija Bezbradica', 'Alistair Sutherland', 'Anwesha Mohanty'] | 2023-03-08 | null | null | null | null | ['face-generation'] | ['computer-vision'] | [ 3.04883808e-01 4.57579672e-01 1.03870228e-01 -4.07772243e-01
-7.72015393e-01 -3.59656125e-01 3.46654773e-01 -4.09796298e-01
-1.82946883e-02 7.51570702e-01 -3.79058011e-02 4.97753881e-02
1.21820763e-01 -6.19362235e-01 -3.96421194e-01 -8.90861869e-01
1.49984568e-01 -1.34431282e-02 -5.26930511e-01 -1.79352686e-01
-4.36307639e-02 8.59083891e-01 -1.56316721e+00 4.24859196e-01
8.20911050e-01 8.31951439e-01 -4.19158280e-01 6.08156979e-01
8.12032446e-02 4.35448825e-01 -7.54801929e-01 -7.84807146e-01
5.28441727e-01 -7.57330000e-01 -5.13100207e-01 2.32700214e-01
8.87471795e-01 -7.43743360e-01 -2.94933975e-01 8.96038830e-01
1.00858402e+00 -1.58226475e-01 5.61315715e-01 -1.40080225e+00
-6.65991247e-01 1.28252283e-01 -7.43053377e-01 -1.91109911e-01
3.51505309e-01 5.31999350e-01 4.49915260e-01 -8.15613449e-01
6.97250783e-01 1.24576819e+00 7.35160172e-01 1.31023037e+00
-1.16579437e+00 -1.04241157e+00 -4.34016138e-01 -4.71344404e-02
-1.48896599e+00 -6.81349993e-01 7.17756629e-01 -3.66534770e-01
3.64814579e-01 2.78548032e-01 5.92409313e-01 1.51279020e+00
4.44286875e-02 4.63119686e-01 1.52139902e+00 -3.58230919e-01
1.40932307e-01 4.26136643e-01 -5.37789822e-01 7.94823885e-01
2.78672308e-01 2.93604344e-01 -2.96082228e-01 -3.12910140e-01
8.72617781e-01 -2.10252643e-01 -2.47806802e-01 3.53299677e-02
-6.38225555e-01 9.52767134e-01 3.74200404e-01 6.99976608e-02
-3.57710540e-01 8.63559097e-02 2.75620401e-01 9.63124633e-02
4.51853871e-01 2.07163289e-01 4.87001278e-02 1.68124080e-01
-8.28388214e-01 1.77892044e-01 5.81099749e-01 6.16128623e-01
2.67259210e-01 4.43627775e-01 -4.85807657e-02 1.04780376e+00
2.37381309e-01 5.47478318e-01 2.14136750e-01 -9.29408252e-01
6.32032752e-02 3.18650544e-01 -2.06498668e-01 -1.13004422e+00
-2.03410074e-01 -1.46271467e-01 -1.03511131e+00 8.11050475e-01
5.72444141e-01 -3.68192434e-01 -1.26459312e+00 1.61967146e+00
4.96854722e-01 1.26581416e-01 -9.40352753e-02 9.73141134e-01
1.02060556e+00 2.39395529e-01 1.79172412e-01 -8.33310839e-03
1.30476415e+00 -4.87057954e-01 -7.18390048e-01 5.00038043e-02
1.80473179e-01 -8.56837809e-01 1.00642204e+00 4.88653272e-01
-1.04190981e+00 -2.72724718e-01 -7.10778296e-01 6.85120896e-02
-1.69849470e-01 2.69178867e-01 5.61763942e-01 1.11992133e+00
-1.11490965e+00 4.35232967e-01 -8.12021375e-01 -5.67649126e-01
1.02998364e+00 3.57727110e-01 -7.35832214e-01 -3.03689569e-01
-9.60831702e-01 7.15255558e-01 -2.26202935e-01 2.91399509e-01
-7.83322573e-01 -8.16650927e-01 -7.42305040e-01 -3.18450630e-01
4.61037606e-02 -7.43306637e-01 8.97020340e-01 -1.19417262e+00
-1.42476654e+00 1.17214131e+00 -1.54164154e-03 -2.68404067e-01
7.84079492e-01 7.84749091e-02 -4.49052483e-01 4.30534869e-01
-2.01379538e-01 9.86924887e-01 1.12091911e+00 -1.27918887e+00
-2.58799702e-01 -4.97585058e-01 -8.65891650e-02 -1.93069890e-01
-1.63576722e-01 1.13549739e-01 -1.21410728e-01 -8.20618093e-01
-3.86009008e-01 -1.05407393e+00 -1.63369700e-01 6.29825592e-01
-5.50179183e-01 2.18941778e-01 6.69523358e-01 -8.60400975e-01
7.90277839e-01 -2.07182956e+00 -1.79199785e-01 2.26392090e-01
2.89513797e-01 5.30871570e-01 -3.93214732e-01 3.71345073e-01
-3.25959444e-01 5.88938117e-01 -3.35958481e-01 -2.74187535e-01
-3.07268471e-01 -8.68274197e-02 6.96689859e-02 6.03047132e-01
4.44135875e-01 8.05779159e-01 -5.67968726e-01 -4.52898055e-01
2.41095290e-01 1.11844885e+00 -5.13305962e-01 6.79900870e-02
4.32541221e-02 6.80587649e-01 -3.18714768e-01 1.15566599e+00
9.00436819e-01 -4.76415679e-02 -7.89310932e-02 -2.80795842e-01
4.03051525e-01 -4.88253802e-01 -9.64847028e-01 1.34370601e+00
-3.56883854e-01 6.43259704e-01 2.20857680e-01 -5.22853494e-01
7.40971267e-01 5.36430001e-01 4.24697071e-01 -4.92905170e-01
3.76073360e-01 1.40667498e-01 1.08193137e-01 -6.86560333e-01
-1.24353200e-01 -3.89483035e-01 4.25689429e-01 8.80127251e-02
-1.43446892e-01 -1.80422962e-01 3.83252464e-02 -1.01764351e-01
9.28365171e-01 3.42888795e-02 1.79892480e-01 2.45271802e-01
3.22020829e-01 1.45755084e-02 4.32099819e-01 1.59774661e-01
-2.90865809e-01 9.67879176e-01 5.61501086e-01 -1.99015513e-01
-1.06792665e+00 -9.37449276e-01 -3.26463014e-01 3.03114444e-01
-4.11025465e-01 -1.32235825e-01 -1.08063233e+00 -6.64782643e-01
7.88701102e-02 5.56382537e-01 -1.01222551e+00 1.33534018e-02
-1.52712479e-01 -6.41087472e-01 9.06393945e-01 3.19858730e-01
4.54552501e-01 -1.09378541e+00 -3.10248733e-01 -2.92832583e-01
2.03792244e-01 -9.51644361e-01 -2.35468805e-01 -6.10917211e-01
-6.27910972e-01 -1.34882116e+00 -1.05187225e+00 -5.32299995e-01
1.22221911e+00 -8.07366148e-02 5.94202399e-01 1.83709428e-01
-1.07490265e+00 4.68577534e-01 -2.14170337e-01 -6.00493789e-01
-6.53086841e-01 -3.87274861e-01 6.37855530e-02 1.13896184e-01
2.69339681e-01 -5.16449928e-01 -8.81882906e-01 1.73573121e-01
-1.01652277e+00 -1.09115586e-01 6.94709420e-01 9.86195982e-01
5.64600646e-01 -1.95961688e-02 5.75095296e-01 -1.12201715e+00
7.75578678e-01 -4.33366776e-01 -4.23588812e-01 -1.85355414e-02
-4.80770677e-01 -4.31391031e-01 4.54069436e-01 -4.25072432e-01
-1.09467566e+00 9.09628998e-03 -2.61411607e-01 -7.88126230e-01
-5.01838207e-01 1.89200193e-01 -2.64892299e-02 -4.00852174e-01
1.01554358e+00 -1.95413306e-01 6.72483623e-01 -2.93787777e-01
7.28382394e-02 7.44086206e-01 1.56154945e-01 -2.15788618e-01
8.73876750e-01 6.49347544e-01 1.31277129e-01 -9.95195448e-01
-4.59684849e-01 2.40978181e-01 -2.20892519e-01 -3.32944483e-01
7.13247597e-01 -8.03967297e-01 -7.29661107e-01 7.72639453e-01
-7.96706438e-01 -2.51050204e-01 -2.80758739e-01 3.03278416e-01
-3.51832509e-01 1.40878856e-01 -4.47198570e-01 -7.44790614e-01
-5.03265738e-01 -1.14152062e+00 9.42352295e-01 5.18272936e-01
-3.69623274e-01 -9.55816805e-01 -3.74825075e-02 8.29094648e-01
4.76651371e-01 1.01500475e+00 6.20076835e-01 -1.64380148e-01
-3.07141274e-01 -4.98883694e-01 -1.23701908e-01 8.10581684e-01
5.28434038e-01 4.07219857e-01 -1.36755323e+00 -4.28405017e-01
-2.24705592e-01 -5.40074110e-01 4.89306778e-01 4.51102406e-01
1.34107852e+00 -4.72344607e-01 -7.34013841e-02 6.75396919e-01
1.34869933e+00 1.00883350e-01 9.72025156e-01 -1.56909153e-01
6.21737242e-01 8.29718590e-01 5.13625741e-01 4.39249754e-01
-6.12173826e-02 2.72634268e-01 4.20001715e-01 -3.75379384e-01
-5.11784315e-01 -2.38396272e-01 1.72459587e-01 1.25727579e-02
-3.55523974e-01 -2.56893665e-01 -8.20126474e-01 4.51375961e-01
-1.00366211e+00 -9.50638235e-01 7.94741288e-02 2.10280657e+00
7.87328899e-01 -4.48755622e-01 1.62089348e-01 1.08232632e-01
8.64292920e-01 -7.71623775e-02 -7.70667672e-01 -4.08213556e-01
-3.24565172e-02 6.86401248e-01 3.67208809e-01 2.58347005e-01
-7.60138035e-01 7.55529702e-01 6.17010546e+00 7.53495932e-01
-1.58071780e+00 -9.86869559e-02 9.39709365e-01 -3.45874310e-01
-3.33959430e-01 -3.80100071e-01 -5.68399906e-01 4.85529959e-01
7.18105435e-01 -4.80337534e-03 3.98783565e-01 7.59865105e-01
3.70501667e-01 -6.37668511e-03 -7.72162795e-01 1.04570878e+00
1.77631572e-01 -1.29184783e+00 7.36051574e-02 3.69178653e-01
6.32524312e-01 -2.57272899e-01 5.65366030e-01 -2.54422635e-01
-1.33834761e-02 -1.61816335e+00 1.07592426e-01 5.56102991e-01
1.53835952e+00 -7.61545181e-01 7.18636036e-01 -1.90760955e-01
-5.44167817e-01 7.09022731e-02 -5.12615405e-02 5.11619627e-01
-6.50046542e-02 5.14083922e-01 -1.08737421e+00 2.13461012e-01
6.38967752e-01 3.19550246e-01 -5.98676860e-01 8.40635598e-01
-3.29362273e-01 6.56599879e-01 -2.14277327e-01 1.70413896e-01
-1.22789070e-01 -6.15855493e-02 4.78139192e-01 8.20612967e-01
4.42025810e-01 2.56612062e-01 -5.53046763e-01 9.49347198e-01
-8.35136995e-02 1.89045042e-01 -7.19503701e-01 -3.82646829e-01
3.50845516e-01 1.54685271e+00 -3.29010069e-01 2.91809827e-01
-2.37064660e-01 8.76226962e-01 -1.83782000e-02 2.77227759e-01
-7.26913214e-01 -2.85625249e-01 8.34080338e-01 6.07790947e-01
-2.05335133e-02 1.84946924e-01 -2.15286121e-01 -8.30743372e-01
-4.99246307e-02 -1.38420725e+00 3.02248359e-01 -7.97149122e-01
-1.37237132e+00 6.65117741e-01 -2.35736057e-01 -1.24372113e+00
-3.43972832e-01 -5.64343214e-01 -6.56756401e-01 9.87319052e-01
-1.31609976e+00 -1.50907707e+00 -7.41304159e-01 6.20099306e-01
2.44792387e-01 -4.30764914e-01 1.15660024e+00 1.59035236e-01
-7.58663952e-01 1.11714005e+00 -1.52291492e-01 1.59947425e-01
8.83976579e-01 -8.64957154e-01 1.39459789e-01 5.34137070e-01
-3.59006196e-01 5.94519198e-01 6.29855275e-01 -6.05370164e-01
-1.28900766e+00 -1.05564773e+00 2.75594741e-01 -9.84353349e-02
2.99016714e-01 -1.88526839e-01 -7.70448983e-01 3.84603858e-01
3.45111042e-01 3.23224396e-01 1.07568848e+00 -4.35371131e-01
-3.06534857e-01 -1.72520816e-01 -2.09392309e+00 7.10555613e-01
7.94577479e-01 -4.02143329e-01 2.29890868e-01 2.75059760e-01
1.38092011e-01 -4.13090825e-01 -1.17494297e+00 5.05589247e-01
6.98037446e-01 -1.00131249e+00 7.94427216e-01 -5.04571915e-01
7.43593991e-01 -2.51083672e-02 1.07179672e-01 -1.34578371e+00
1.40649706e-01 -7.71003366e-01 4.75679636e-02 1.40592885e+00
1.85858414e-01 -7.57076919e-01 1.13659692e+00 8.50659251e-01
4.25556570e-01 -9.04421270e-01 -6.97942436e-01 -4.46774036e-01
1.60361499e-01 -2.84531433e-02 3.43927056e-01 1.02978849e+00
-4.20037538e-01 -1.86653957e-01 -3.72905672e-01 7.14488178e-02
8.30509841e-01 -4.01903450e-01 8.57241273e-01 -8.92306924e-01
1.85158569e-02 -2.04725638e-01 -6.28310144e-01 1.60219874e-02
2.07979977e-01 -7.85387337e-01 -4.66000170e-01 -1.22725821e+00
-1.39590338e-01 -5.46502113e-01 9.22657736e-03 8.11060131e-01
-1.20588075e-02 7.66484201e-01 4.55076844e-02 -2.15108201e-01
5.75450182e-01 2.17765942e-01 1.55114949e+00 -1.06314093e-01
8.48781317e-02 7.61952028e-02 -9.37494278e-01 5.48732162e-01
1.25956190e+00 -2.67582387e-01 -5.17227530e-01 -1.36515588e-01
-2.45632395e-01 5.27991802e-02 5.83052874e-01 -8.59744847e-01
-9.22526047e-02 -2.74645209e-01 6.25514507e-01 -1.05015799e-01
6.32178009e-01 -6.58726275e-01 4.63325381e-01 5.14288247e-01
-1.98030725e-01 -3.66298854e-01 4.71695602e-01 2.98657149e-01
-2.03426659e-01 -4.87112850e-02 1.26302874e+00 -1.87983200e-01
-4.12044883e-01 4.84549224e-01 -2.03325450e-01 -5.50022423e-02
1.25390744e+00 -3.55343431e-01 -2.26176947e-01 -6.04155838e-01
-7.30980992e-01 -9.34567377e-02 6.82588995e-01 3.42064083e-01
8.92762065e-01 -1.23719227e+00 -1.01783133e+00 3.37129623e-01
8.92439559e-02 4.48303893e-02 6.98669314e-01 7.63068438e-01
-8.11452389e-01 1.22643327e-02 -5.98888457e-01 -3.53749841e-01
-1.71124768e+00 2.40744993e-01 5.42773664e-01 3.69832993e-01
-3.98413599e-01 9.53298092e-01 7.43492171e-02 -2.17302442e-01
1.83131516e-01 1.60488904e-01 6.81497753e-02 6.33225068e-02
6.76524282e-01 4.09697056e-01 3.50526981e-02 -5.81413627e-01
-2.01329023e-01 4.59029883e-01 -2.40504891e-01 7.53779113e-02
1.33552945e+00 1.63516343e-01 -8.29248428e-02 -2.49483287e-01
1.08685231e+00 1.74567446e-01 -1.19103467e+00 1.93657979e-01
-6.72909796e-01 -8.22405338e-01 -6.56513870e-02 -9.60637152e-01
-1.70333946e+00 8.50741923e-01 9.88510907e-01 -1.94613099e-01
1.40512764e+00 -2.38200009e-01 6.02789819e-01 -9.51875374e-02
2.61647731e-01 -7.14309692e-01 1.04689980e-02 -3.27512771e-01
1.23938715e+00 -1.24954677e+00 -1.10480823e-02 -6.42204344e-01
-8.23087394e-01 9.78128135e-01 7.52272367e-01 -5.06122969e-02
5.14240742e-01 4.63134497e-01 5.32736778e-01 -1.70287922e-01
-4.93328124e-01 7.79763237e-02 1.20292470e-01 9.21116889e-01
4.69548464e-01 -1.01698749e-01 -2.54411012e-01 1.90899476e-01
-3.18687618e-01 1.57302916e-01 5.62967002e-01 8.14045072e-01
2.63611436e-01 -1.33199835e+00 -5.32249749e-01 6.18367672e-01
-8.85501146e-01 9.72334743e-02 -7.02108800e-01 9.63873565e-01
1.66327477e-01 8.45202148e-01 -4.28591259e-02 -2.34063670e-01
1.87628180e-01 -9.90677029e-02 6.87517464e-01 -4.92108405e-01
-5.82059264e-01 -1.45150885e-01 1.01592720e-01 -5.47486365e-01
-3.56075108e-01 -6.39196277e-01 -1.05243456e+00 -6.74236596e-01
-1.60711467e-01 -2.44465306e-01 9.33889508e-01 3.34806412e-01
4.97046918e-01 2.93402761e-01 7.39678621e-01 -5.21810830e-01
-3.98452520e-01 -9.28024709e-01 -6.81029499e-01 4.51855004e-01
2.68109024e-01 -5.65655291e-01 -4.99909371e-01 2.37721339e-01] | [12.910597801208496, 0.30817967653274536] |
96669b71-3bef-4f86-bc2b-c5e06f3ed9f4 | an-investigation-of-interpretability | 2002.09192 | null | https://arxiv.org/abs/2002.09192v1 | https://arxiv.org/pdf/2002.09192v1.pdf | An Investigation of Interpretability Techniques for Deep Learning in Predictive Process Analytics | This paper explores interpretability techniques for two of the most successful learning algorithms in medical decision-making literature: deep neural networks and random forests. We applied these algorithms in a real-world medical dataset containing information about patients with cancer, where we learn models that try to predict the type of cancer of the patient, given their set of medical activity records. We explored different algorithms based on neural network architectures using long short term deep neural networks, and random forests. Since there is a growing need to provide decision-makers understandings about the logic of predictions of black boxes, we also explored different techniques that provide interpretations for these classifiers. In one of the techniques, we intercepted some hidden layers of these neural networks and used autoencoders in order to learn what is the representation of the input in the hidden layers. In another, we investigated an interpretable model locally around the random forest's prediction. Results show learning an interpretable model locally around the model's prediction leads to a higher understanding of why the algorithm is making some decision. Use of local and linear model helps identify the features used in prediction of a specific instance or data point. We see certain distinct features used for predictions that provide useful insights about the type of cancer, along with features that do not generalize well. In addition, the structured deep learning approach using autoencoders provided meaningful prediction insights, which resulted in the identification of nonlinear clusters correspondent to the patients' different types of cancer. | ['Peter Bruza', 'Chun Ouyang', 'Catarina Moreira', 'Renuka Sindhgatta', 'Andreas Wichert'] | 2020-02-21 | null | null | null | null | ['interpretability-techniques-for-deep-learning'] | ['miscellaneous'] | [ 2.42609978e-01 8.16467404e-01 -3.51066947e-01 -6.81519508e-01
9.63644311e-02 -1.14780970e-01 4.55771089e-01 3.67150396e-01
1.84716024e-02 6.83293581e-01 6.32707655e-01 -7.46275008e-01
-6.99424088e-01 -1.04724026e+00 -5.83883405e-01 -7.73830175e-01
-2.17177927e-01 8.91942322e-01 -2.98516303e-01 -1.17738709e-01
7.28633255e-02 4.20856237e-01 -1.26480830e+00 1.20163143e+00
4.50305611e-01 1.17123663e+00 -4.44709286e-02 6.63216174e-01
-3.60154897e-01 1.36990631e+00 -4.13558602e-01 -1.87811762e-01
-4.36398238e-02 -3.18324655e-01 -1.25807905e+00 -1.50320098e-01
-1.81127474e-01 -2.31596738e-01 1.66178897e-01 5.66259384e-01
-1.62720367e-01 -4.34833646e-01 9.83302295e-01 -1.08915472e+00
-6.22374237e-01 8.48272026e-01 4.74789329e-02 8.19648057e-02
2.34861106e-01 1.66581422e-01 9.13867772e-01 -3.76174659e-01
5.97629786e-01 1.24401009e+00 1.17627883e+00 7.18422830e-01
-1.03127348e+00 -2.99250424e-01 1.40443793e-03 4.31471616e-01
-8.77727687e-01 -1.69829875e-01 4.56834435e-01 -8.08838427e-01
8.70368659e-01 5.44853389e-01 9.06706631e-01 1.24058020e+00
7.40423441e-01 4.85522926e-01 1.39680088e+00 -6.33992434e-01
2.95760661e-01 4.42868292e-01 6.15390301e-01 8.32199037e-01
1.86295569e-01 4.34278786e-01 -1.79909110e-01 -2.53981560e-01
5.11179924e-01 6.28437221e-01 -4.06592302e-02 7.44094476e-02
-1.11843097e+00 9.75355506e-01 9.05639768e-01 6.31081641e-01
-6.56184494e-01 -6.93622604e-02 2.79094249e-01 1.05867177e-01
3.46410543e-01 6.57746792e-01 -8.91949713e-01 4.33158606e-01
-7.23641872e-01 -6.59067333e-02 9.68149722e-01 1.93836525e-01
6.84275150e-01 -3.46944094e-01 -2.34069645e-01 2.88318992e-01
4.26952660e-01 -2.63377428e-01 9.64714110e-01 -6.58930242e-01
5.41757867e-02 1.03974187e+00 -2.30530977e-01 -9.93762553e-01
-7.88749754e-01 -6.25434458e-01 -1.17004359e+00 3.16754490e-01
5.61526597e-01 -1.74163476e-01 -1.24330485e+00 1.20859206e+00
-3.10031809e-02 -1.95368156e-02 4.30831850e-01 5.23462713e-01
6.08179808e-01 3.91746610e-01 2.24538609e-01 6.65936545e-02
1.45912969e+00 -8.23668301e-01 -6.32295251e-01 -1.97071850e-01
7.94649422e-01 -5.21760806e-02 6.41839921e-01 4.41170365e-01
-4.64591622e-01 -5.67716300e-01 -9.10882473e-01 1.42279221e-02
-8.52965355e-01 -2.96212826e-02 8.40106368e-01 3.47527802e-01
-1.13220978e+00 1.07013214e+00 -9.27104592e-01 -5.18958449e-01
6.68261707e-01 6.06327415e-01 -2.15946093e-01 1.60859734e-01
-1.16307092e+00 1.24339342e+00 5.28885484e-01 2.64498413e-01
-7.36369312e-01 -3.30104530e-01 -5.09603918e-01 2.67241389e-01
-1.80360079e-02 -1.12392533e+00 9.80614305e-01 -1.67479503e+00
-9.48589206e-01 8.50515962e-01 -2.80665278e-01 -9.37887609e-01
3.63936275e-01 1.80275589e-02 -4.64905292e-01 -2.28385970e-01
-2.99206516e-03 6.01282835e-01 5.47513783e-01 -1.10327733e+00
-5.80977261e-01 -6.84206605e-01 -1.76738083e-01 -1.68855220e-01
-1.78369761e-01 -3.60323340e-01 4.24310148e-01 -2.96258867e-01
1.41890764e-01 -8.10700655e-01 -5.68629265e-01 -1.28895998e-01
-7.59628236e-01 -2.14154094e-01 6.36717737e-01 -8.14844549e-01
1.30400944e+00 -1.79365253e+00 -1.72777340e-01 4.40071821e-01
6.91985548e-01 -9.31785703e-02 4.44592744e-01 2.80694604e-01
-4.65430021e-01 5.00366390e-01 -1.76041201e-01 1.52565449e-01
-4.35947180e-01 7.05206394e-01 -1.83582082e-01 -4.77230065e-02
2.75941491e-01 9.13900137e-01 -5.16810596e-01 -3.21013898e-01
2.89847963e-02 3.12371701e-01 -3.80429000e-01 2.05927491e-01
-2.29798138e-01 2.73583353e-01 -5.81614196e-01 4.58562732e-01
2.17161626e-01 -4.74302530e-01 3.07557076e-01 -5.89491054e-02
1.89420626e-01 3.62995237e-01 -5.01547694e-01 9.58169341e-01
-4.29712206e-01 7.08641946e-01 -4.55809981e-01 -1.37488770e+00
6.80764854e-01 4.38918978e-01 3.08501303e-01 -1.00351699e-01
4.33151908e-02 2.78606098e-02 3.35251153e-01 -7.13721633e-01
-1.45367369e-01 -6.71937823e-01 3.18965822e-01 3.51456881e-01
-7.28667825e-02 5.89231491e-01 -4.90977705e-01 -3.74664962e-01
1.28467131e+00 -1.19989745e-01 9.07587171e-01 -3.16678196e-01
4.19724107e-01 3.34210455e-01 5.87916791e-01 7.58593500e-01
2.83061638e-02 4.45557952e-01 6.35192573e-01 -1.51100099e+00
-6.76134586e-01 -7.45814621e-01 -2.75725484e-01 7.83298373e-01
-4.21803325e-01 -1.20239131e-01 -6.29354775e-01 -9.18709278e-01
-2.94607133e-02 1.00391543e+00 -1.48502791e+00 -1.65915027e-01
-2.92309344e-01 -8.30914140e-01 3.80244166e-01 6.38832867e-01
1.97316229e-01 -1.35553110e+00 -8.45314920e-01 1.30437836e-01
1.61812380e-02 -6.04322910e-01 3.36230636e-01 1.00015378e+00
-1.43848634e+00 -1.30967939e+00 -1.56779617e-01 -6.38369024e-01
9.12822366e-01 -5.50230443e-01 1.34757757e+00 2.99736917e-01
-1.59371540e-01 -1.00985430e-01 -1.63768917e-01 -9.82175231e-01
-8.96606982e-01 1.44527569e-01 -1.63402602e-01 -4.02192213e-02
7.24994481e-01 -2.86505669e-01 -5.23424506e-01 1.08736761e-01
-7.17234552e-01 3.75708014e-01 8.19167256e-01 1.05670059e+00
3.64375889e-01 2.91059315e-01 2.06861600e-01 -1.42951393e+00
6.61733449e-01 -9.00102675e-01 2.14488536e-01 3.69189501e-01
-1.03677344e+00 6.83812261e-01 7.77875841e-01 -4.70753871e-02
-8.54808569e-01 1.68609634e-01 -2.94340432e-01 -2.00081125e-01
-7.26671994e-01 6.41487300e-01 1.30324900e-01 5.15605271e-01
9.74977672e-01 1.68240651e-01 8.79120976e-02 -3.50428909e-01
-8.54720175e-02 7.08553374e-01 4.67188619e-02 -1.42868608e-01
2.01636150e-01 4.18852031e-01 8.63588303e-02 -2.44806722e-01
-9.52826023e-01 -1.13665767e-01 -8.49338353e-01 3.52756716e-02
1.11597800e+00 -4.64529335e-01 -6.21747375e-01 6.52007908e-02
-1.07226586e+00 -2.03942701e-01 -4.47619677e-01 2.32637882e-01
-5.14310956e-01 -3.37972105e-01 -4.77287024e-01 -7.01685369e-01
-3.93131047e-01 -1.04928660e+00 7.00533628e-01 5.76158389e-02
-8.74925077e-01 -1.52176690e+00 1.40907895e-02 3.45183074e-01
3.83053929e-01 3.95095259e-01 1.51939309e+00 -1.25145376e+00
-2.34527901e-01 -2.70987898e-01 6.18624799e-02 2.32768282e-01
2.68854529e-01 -1.33974925e-01 -1.10618412e+00 3.60986084e-01
2.16302127e-01 1.50797054e-01 8.98707151e-01 7.00879276e-01
1.64014113e+00 -1.02538419e+00 -7.32306480e-01 6.13141179e-01
1.33114624e+00 5.38573205e-01 6.38660848e-01 4.98994678e-01
4.27746534e-01 7.90951371e-01 1.16108045e-01 5.43669090e-02
3.86287302e-01 3.82006198e-01 5.27113795e-01 -3.91190976e-01
2.09517092e-01 -2.75362432e-01 2.42865741e-01 2.78471768e-01
-4.97214019e-01 8.63747522e-02 -1.24942887e+00 3.15668732e-01
-1.95835829e+00 -9.17969704e-01 -1.19243741e-01 1.88212228e+00
6.36977732e-01 2.67744869e-01 2.80765481e-02 3.63185167e-01
3.34229976e-01 -3.25522065e-01 -4.79572952e-01 -8.99644613e-01
9.08994824e-02 2.85984129e-01 3.80674392e-01 3.44406247e-01
-1.01785803e+00 4.52905983e-01 6.76429987e+00 4.06574547e-01
-1.29258621e+00 -6.29124511e-03 1.45063829e+00 4.08112496e-01
-3.01111966e-01 -1.96255650e-02 -5.04474938e-01 3.44475836e-01
1.37719083e+00 3.00526291e-01 1.64521113e-01 9.87359703e-01
3.82828534e-01 -1.47971228e-01 -1.58415830e+00 3.76428396e-01
-3.09047699e-01 -1.71782279e+00 2.80831993e-01 3.42167020e-01
4.47513551e-01 -1.99839696e-01 -1.14719480e-01 8.73395205e-02
5.91395557e-01 -1.72303843e+00 3.20694447e-01 1.09281147e+00
3.40406209e-01 -4.13542897e-01 1.18225849e+00 6.10196173e-01
-4.38579857e-01 -5.74474335e-01 -1.03360556e-01 -5.35606205e-01
-6.00961804e-01 6.84682608e-01 -1.72267222e+00 2.18557805e-01
8.18350434e-01 7.01877058e-01 -6.41597807e-01 4.92850065e-01
-2.20505878e-01 9.41594124e-01 -9.80721042e-02 -2.07453504e-01
2.30735824e-01 2.03799263e-01 4.68986742e-02 1.14821148e+00
2.56222606e-01 6.24463670e-02 -1.67161375e-01 8.67915273e-01
3.60526681e-01 -8.83007720e-02 -7.79425204e-01 6.26253039e-02
-2.30594680e-01 9.82295036e-01 -6.79305971e-01 -4.15066808e-01
-2.20570043e-02 7.56066382e-01 1.88192174e-01 1.56337082e-01
-4.60301667e-01 1.30154982e-01 3.90916824e-01 6.06325686e-01
-1.12783886e-01 4.35783654e-01 -1.01355565e+00 -9.39696848e-01
-3.95509183e-01 -8.11256170e-01 6.75092816e-01 -8.42076421e-01
-1.28797424e+00 8.50351334e-01 -2.67661393e-01 -1.00843203e+00
-7.13063240e-01 -9.37001944e-01 -8.53915513e-01 8.81785512e-01
-9.58103180e-01 -1.26272857e+00 -1.95158854e-01 5.55653393e-01
4.04523611e-01 -2.79939413e-01 1.46172011e+00 -5.15279830e-01
-2.81934351e-01 1.17234275e-01 2.14319583e-02 3.80156845e-01
1.25953227e-01 -1.46104455e+00 -9.81274173e-02 1.58754572e-01
1.91254809e-01 7.24884987e-01 6.53271675e-01 -4.09215599e-01
-7.40022957e-01 -1.04414248e+00 1.11008048e+00 -6.43533945e-01
4.12976831e-01 1.82717070e-01 -9.23673868e-01 8.99381459e-01
2.29372606e-01 -2.01217890e-01 1.25183809e+00 5.49624205e-01
9.79195237e-02 -2.42621019e-01 -1.35817802e+00 2.43013084e-01
5.56647301e-01 -1.94607213e-01 -8.77370596e-01 4.39348280e-01
3.01442713e-01 -4.48424667e-02 -7.29785919e-01 3.27226162e-01
7.70728707e-01 -1.27146232e+00 8.59202266e-01 -1.28712940e+00
1.13184345e+00 1.77380200e-02 6.27171546e-02 -1.50217950e+00
-6.42364383e-01 1.86756968e-01 -1.05487444e-01 5.09066224e-01
8.50509286e-01 -7.45361686e-01 1.02354705e+00 8.72345150e-01
1.17652103e-01 -1.36240160e+00 -6.27709091e-01 3.47698182e-02
2.03936517e-01 -3.55290949e-01 6.18055701e-01 9.82534707e-01
-6.86794966e-02 3.37391257e-01 -1.55010119e-01 8.12571794e-02
1.34082034e-01 1.36474431e-01 2.52273470e-01 -1.42511737e+00
-5.02317786e-01 -3.54697466e-01 -6.45528078e-01 -2.58614510e-01
-5.65001294e-02 -1.02636838e+00 -2.94898957e-01 -1.72585058e+00
2.53308773e-01 -4.26653981e-01 -5.80983758e-01 9.92831469e-01
-9.35110599e-02 -3.14177960e-01 6.91715032e-02 2.67081529e-01
9.21944082e-02 -2.14806452e-01 9.01866257e-01 -3.09646875e-01
-9.63591486e-02 7.54489124e-01 -1.09453368e+00 1.25774944e+00
8.88744116e-01 -6.37815177e-01 -2.04802766e-01 -3.25372875e-01
3.37145209e-01 2.57670552e-01 5.46969414e-01 -1.06703317e+00
1.24566488e-01 -2.87135690e-01 1.18045914e+00 -3.14263076e-01
-2.59914398e-02 -1.17782927e+00 5.13640523e-01 1.04491746e+00
-8.10341954e-01 -3.70402150e-02 -5.09492010e-02 5.22711873e-01
-2.49082968e-01 -3.56453836e-01 5.05832076e-01 -5.91770232e-01
-6.04132533e-01 -1.29054353e-01 -6.39871716e-01 -5.36909044e-01
1.00113916e+00 -4.21118349e-01 2.06535786e-01 -4.85859036e-01
-1.47042048e+00 -4.64325584e-03 -3.19130793e-02 1.94931477e-01
5.93199015e-01 -9.04602647e-01 -4.91217107e-01 2.21077815e-01
-1.34363860e-01 6.80799335e-02 -1.09480150e-01 5.99875629e-01
-7.68937349e-01 6.61337376e-01 -3.83643895e-01 -6.06191814e-01
-1.26721168e+00 6.19608223e-01 7.83804238e-01 -8.51807237e-01
-2.98131883e-01 6.39769912e-01 2.65462667e-01 -4.92817879e-01
-1.23550698e-01 -8.23499620e-01 -5.43816745e-01 6.66825324e-02
3.49754363e-01 1.70839176e-01 8.28633457e-02 -2.46655941e-01
-4.24385726e-01 2.03675181e-01 -9.20211524e-02 4.01060253e-01
1.65880966e+00 3.25520724e-01 -1.98268548e-01 7.03963995e-01
9.93256569e-01 -4.39703554e-01 -6.91115201e-01 8.50500762e-02
4.32622224e-01 7.71929324e-02 -4.92491201e-02 -1.50626421e+00
-1.01445270e+00 1.20105898e+00 7.63059497e-01 3.41151536e-01
1.38523817e+00 6.17877766e-03 2.27559775e-01 5.07571757e-01
1.48827508e-01 -6.96590424e-01 -4.72927749e-01 3.78476471e-01
9.36483860e-01 -1.37200892e+00 4.95577231e-02 -9.55822170e-02
-6.66887164e-01 1.81287408e+00 3.45430404e-01 -8.55698436e-03
9.32363689e-01 2.46973440e-01 3.14630926e-01 -3.64168078e-01
-1.01647055e+00 2.15040147e-01 4.06872094e-01 6.03549480e-01
4.79297042e-01 6.71441555e-01 8.88031274e-02 1.07731009e+00
-5.10169864e-01 4.94414657e-01 2.98761338e-01 4.82898206e-01
-3.10578316e-01 -8.32343817e-01 -5.31051636e-01 9.78165269e-01
-5.45858800e-01 -1.38352409e-01 -7.23699689e-01 6.46946132e-01
7.07243264e-01 5.85076153e-01 7.68567026e-02 -6.73848808e-01
-5.84600680e-02 5.32060802e-01 -4.92408536e-02 -5.70216477e-01
-8.79409254e-01 -3.85417819e-01 8.57951120e-02 -3.72107208e-01
-3.80346894e-01 -5.28069258e-01 -1.35357046e+00 -1.29084513e-02
-4.43439111e-02 7.96927363e-02 5.50405979e-01 1.31854379e+00
3.35009396e-01 7.88086534e-01 3.86367202e-01 -4.45253551e-01
-3.55580598e-01 -1.02706742e+00 -2.07606211e-01 3.95580560e-01
5.28692544e-01 -2.64467984e-01 -2.23866388e-01 3.08295965e-01] | [8.608857154846191, 5.743265151977539] |
40f3f5be-2b9f-4c07-a2a1-bb87e67eb563 | siamese-networks-with-location-prior-for | 1901.08109 | null | http://arxiv.org/abs/1901.08109v1 | http://arxiv.org/pdf/1901.08109v1.pdf | Siamese Networks with Location Prior for Landmark Tracking in Liver Ultrasound Sequences | Image-guided radiation therapy can benefit from accurate motion tracking by
ultrasound imaging, in order to minimize treatment margins and radiate moving
anatomical targets, e.g., due to breathing. One way to formulate this tracking
problem is the automatic localization of given tracked anatomical landmarks
throughout a temporal ultrasound sequence. For this, we herein propose a
fully-convolutional Siamese network that learns the similarity between pairs of
image regions containing the same landmark. Accordingly, it learns to localize
and thus track arbitrary image features, not only predefined anatomical
structures. We employ a temporal consistency model as a location prior, which
we combine with the network-predicted location probability map to track a
target iteratively in ultrasound sequences. We applied this method on the
dataset of the Challenge on Liver Ultrasound Tracking (CLUST) with competitive
results, where our work is the first to effectively apply CNNs on this tracking
problem, thanks to our temporal regularization. | ['Weiye Li', 'Orcun Goksel', 'Christine Tanner', 'Ece Ozkan', 'Alvaro Gomariz'] | 2019-01-23 | null | null | null | null | ['landmark-tracking'] | ['computer-vision'] | [ 1.94993570e-01 2.22025737e-01 -2.30846852e-01 -1.03989609e-01
-1.01191318e+00 -9.87224460e-01 4.52993453e-01 1.12115443e-01
-4.78016019e-01 2.16743633e-01 4.43182230e-01 -2.34518662e-01
-4.62185204e-01 -3.74630183e-01 -7.74898648e-01 -9.39561307e-01
-5.15150547e-01 3.87972206e-01 1.77124396e-01 2.18430594e-01
-6.11553714e-02 4.97114301e-01 -3.09373111e-01 8.17989334e-02
5.36142349e-01 8.86068583e-01 3.12872440e-01 6.31892204e-01
3.09228510e-01 9.07127023e-01 2.05724761e-02 8.33414588e-03
1.28515497e-01 -5.38899720e-01 -1.00560832e+00 7.71578075e-03
6.31935239e-01 -1.35542288e-01 -6.68420792e-01 1.00385869e+00
5.31268954e-01 2.12883011e-01 6.23791456e-01 -5.24107039e-01
-5.74250877e-01 7.86306500e-01 -3.85543317e-01 4.57768708e-01
1.27161995e-01 4.14137959e-01 5.79912305e-01 -5.55028617e-01
8.79242420e-01 5.04404366e-01 1.10401559e+00 7.50150084e-01
-1.00967705e+00 -1.25183582e-01 -6.07943274e-02 -1.73287556e-01
-1.22029948e+00 -2.86689311e-01 4.73561347e-01 -5.55708945e-01
3.70281011e-01 4.41905171e-01 6.83771670e-01 9.59602773e-01
6.97474003e-01 8.14817846e-01 6.18788779e-01 -2.46012390e-01
2.91495882e-02 -3.77601385e-01 3.24496552e-02 1.18770611e+00
-2.38529369e-01 3.23876530e-01 -1.81525633e-01 -1.04423910e-01
8.66894007e-01 1.30873904e-01 -9.65114236e-01 -8.97658110e-01
-1.49949265e+00 6.23676717e-01 1.24029839e+00 9.92490053e-01
-4.91926432e-01 6.36514902e-01 2.38954097e-01 -3.31087172e-01
3.16613674e-01 4.89545345e-01 -3.71819675e-01 -2.74353893e-03
-9.02496159e-01 -2.22676575e-01 6.92529380e-01 5.84027410e-01
1.30063295e-01 -2.83558905e-01 -8.01880300e-01 2.54934430e-01
5.03457665e-01 1.37959212e-01 6.08278811e-01 -7.91123629e-01
1.08856663e-01 1.15607366e-01 2.05024302e-01 -7.24175930e-01
-8.55779469e-01 -8.05198729e-01 -1.00326753e+00 -4.51061316e-02
7.52654672e-01 -1.32504806e-01 -1.32522202e+00 1.97124577e+00
4.69634980e-01 7.81498313e-01 -1.97454348e-01 1.21310127e+00
7.75782764e-01 2.21906453e-01 2.84227759e-01 -3.17364901e-01
1.51264918e+00 -1.13059962e+00 -6.99755728e-01 -8.46266225e-02
1.24079227e+00 -6.53590798e-01 6.20974123e-01 -9.20276940e-02
-1.24191535e+00 -1.63312897e-01 -6.62389040e-01 4.46862429e-01
1.27246097e-01 2.91725159e-01 5.44095874e-01 5.73070884e-01
-1.09517682e+00 1.05872667e+00 -1.65395653e+00 -3.03243667e-01
6.94987059e-01 4.64933127e-01 -3.08543205e-01 -1.38700843e-01
-9.76917326e-01 1.08947647e+00 1.93954930e-01 4.41577017e-01
-9.45339322e-01 -1.57302117e+00 -1.09792459e+00 6.08667433e-02
2.53029168e-01 -9.67114568e-01 1.23116255e+00 -6.74586356e-01
-1.36620557e+00 8.18305194e-01 -2.94107608e-02 -6.27550304e-01
8.76016796e-01 -2.69501768e-02 -7.67432973e-02 2.57606447e-01
3.60109657e-03 3.25574905e-01 5.97451150e-01 -9.52086747e-01
-2.11107522e-01 -2.61398315e-01 -1.93533659e-01 -4.69788734e-04
-1.12327956e-01 -1.48146689e-01 -5.78421474e-01 -7.05562711e-01
2.73757428e-01 -1.38650262e+00 -6.17688954e-01 2.74358511e-01
-4.00543600e-01 -1.31168559e-01 6.44851506e-01 -8.30340624e-01
9.43821073e-01 -2.05278778e+00 3.95918876e-01 2.81943440e-01
4.15029705e-01 5.40512009e-03 -6.90564886e-02 -2.90215284e-01
-2.37186015e-01 -1.50255086e-02 -3.21826428e-01 -3.12534034e-01
-9.47592184e-02 6.14882559e-02 1.31186754e-01 1.08216453e+00
-1.09169908e-01 1.54314399e+00 -1.28986633e+00 -3.75752121e-01
3.10887277e-01 6.06291890e-01 -4.13633823e-01 1.92215592e-01
-1.34779140e-01 1.25307488e+00 -5.15768409e-01 3.56061459e-01
6.48352623e-01 -5.86272180e-01 3.11145514e-01 -6.66114926e-01
-1.28886938e-01 -6.66792020e-02 -5.75765669e-01 2.56933999e+00
-7.00022161e-01 3.24981004e-01 1.98659539e-01 -6.97561324e-01
1.38915077e-01 6.52983904e-01 1.27955925e+00 -5.24044096e-01
3.03178549e-01 2.95384973e-01 8.79958570e-02 -5.65532327e-01
-1.00098655e-01 -1.47594333e-01 2.03002870e-01 3.44510674e-01
1.46340564e-01 6.15184242e-03 -2.05636248e-01 7.94309080e-02
1.45503950e+00 3.24789733e-01 -8.54739025e-02 -5.43690085e-01
4.25570220e-01 4.88696136e-02 3.33427608e-01 9.18072462e-01
-3.66248757e-01 8.77505958e-01 2.91482173e-02 -7.45385826e-01
-7.04202116e-01 -9.29668903e-01 -3.73567253e-01 5.97879946e-01
2.33821243e-01 8.28496739e-02 -6.07236862e-01 -1.25338781e+00
-6.40458986e-02 2.42822424e-01 -1.05987668e+00 -1.46161914e-01
-1.10325038e+00 -6.31286263e-01 1.70270905e-01 4.83176500e-01
-8.35973993e-02 -9.31106389e-01 -8.38134408e-01 4.88341033e-01
-6.90552294e-02 -1.02043235e+00 -1.00133789e+00 4.06070381e-01
-8.56547356e-01 -1.23735178e+00 -1.26294899e+00 -7.95345604e-01
8.03347886e-01 -7.64908865e-02 1.19860637e+00 3.85573626e-01
-4.99667734e-01 5.32993436e-01 -1.45532623e-01 1.82055533e-01
-5.49712062e-01 2.90947795e-01 -2.70886689e-01 -1.49008676e-01
-3.18411022e-01 -3.59765738e-01 -1.10239339e+00 1.84889391e-01
-7.51121283e-01 -6.48602471e-02 5.20042002e-01 1.02835846e+00
4.60657984e-01 -4.30488676e-01 6.62794933e-02 -8.37574363e-01
8.66853595e-02 -4.90301251e-01 -6.65388227e-01 5.52529275e-01
-3.03613432e-02 -2.29452159e-02 4.13307369e-01 -4.22093809e-01
-6.69410884e-01 5.77604353e-01 -1.30058020e-01 -7.19365120e-01
8.13109651e-02 6.13655150e-01 5.62242866e-01 -8.01543951e-01
6.41643822e-01 3.63993794e-01 -1.04491808e-01 -9.67687666e-02
3.78749639e-01 -2.95035958e-01 7.36508727e-01 -2.44526520e-01
8.03052485e-01 6.28341615e-01 3.70688885e-01 -1.29215971e-01
-1.11205339e+00 -5.17690778e-01 -8.07475746e-01 -2.79542178e-01
1.12816906e+00 -5.74885607e-01 -1.06198466e+00 -2.39052754e-02
-1.10224092e+00 -6.25106871e-01 -2.84844786e-01 8.50547612e-01
-5.72489083e-01 5.29091537e-01 -9.68908608e-01 -4.15289462e-01
-4.88478273e-01 -1.28035796e+00 1.15067124e+00 3.28478128e-01
-1.16198234e-01 -1.56775975e+00 1.37871727e-01 -2.39978209e-01
6.60440624e-01 5.55175602e-01 5.56750298e-01 -4.99542326e-01
-9.59812522e-01 -1.70965672e-01 -8.21660459e-02 -2.17477649e-01
4.62318450e-01 -4.09949839e-01 -6.65730596e-01 -7.68035412e-01
7.18121082e-02 -3.76017280e-02 9.51387286e-01 1.15483022e+00
1.27720857e+00 -1.34695768e-01 -7.69037127e-01 1.17019105e+00
1.25088656e+00 -6.37365356e-02 2.04979837e-01 1.24691740e-01
8.66685331e-01 1.24935210e-01 4.60465580e-01 2.26737395e-01
4.45732921e-02 8.40576172e-01 6.34979069e-01 -4.88199919e-01
-3.33566785e-01 1.13086954e-01 -2.95287758e-01 7.05394208e-01
4.03474569e-02 -1.75641045e-01 -1.05986547e+00 6.80863202e-01
-1.84244049e+00 -6.79981709e-01 -1.05225720e-01 1.92183125e+00
7.88061380e-01 -3.54976773e-01 -4.36241955e-01 -7.17064261e-01
4.23886895e-01 8.91278237e-02 -3.57809842e-01 4.87038761e-01
3.67833138e-01 3.40696305e-01 7.06170917e-01 6.72859669e-01
-1.62678492e+00 5.90635896e-01 5.90634775e+00 5.66477358e-01
-1.26725757e+00 4.78036851e-01 5.50226092e-01 1.42601162e-01
-1.29500568e-01 -1.28795400e-01 -2.90495485e-01 3.42853427e-01
4.93772358e-01 4.20147888e-02 1.85085461e-01 4.67282474e-01
1.45150766e-01 2.17936244e-02 -1.38885033e+00 6.65296316e-01
-6.96994141e-02 -1.76509142e+00 -5.24647295e-01 -2.75427513e-02
9.33768809e-01 1.38448417e-01 3.01153064e-01 -2.24286746e-02
3.32153350e-01 -1.15169168e+00 3.96846980e-01 6.63464367e-01
4.81854528e-01 -2.62774259e-01 6.51361108e-01 2.02537447e-01
-1.08735836e+00 2.18122572e-01 3.78202996e-03 8.41236413e-01
2.80806959e-01 2.47628391e-01 -1.00472188e+00 9.50022995e-01
4.86385375e-01 9.06579494e-01 -2.06376240e-01 1.58867335e+00
-5.15380763e-02 3.67289782e-01 -3.19146574e-01 4.73479986e-01
4.99098212e-01 -6.27966076e-02 5.97255230e-01 1.23787189e+00
6.99043989e-01 6.61030114e-02 3.20576817e-01 8.35131526e-01
-1.73268393e-01 -1.27576828e-01 -2.94352621e-01 4.72449422e-01
1.24935590e-01 1.59390998e+00 -9.68425512e-01 -1.71903208e-01
-1.37694821e-01 1.03268898e+00 1.67576283e-01 2.61416137e-01
-1.05300498e+00 3.42677772e-01 1.44360617e-01 -9.17384680e-03
2.59571254e-01 -2.43730560e-01 -5.29261529e-02 -1.17336881e+00
-1.65285125e-01 -3.06495756e-01 5.62641740e-01 -6.25271738e-01
-1.06995082e+00 6.72204912e-01 -6.46257699e-01 -1.31634820e+00
-2.21837789e-01 -4.50122833e-01 -6.89263761e-01 1.03333807e+00
-1.61669683e+00 -1.32474089e+00 -6.03921115e-01 5.02698243e-01
1.23889901e-01 3.61404628e-01 8.17708433e-01 4.31130081e-01
-2.08570421e-01 5.49858332e-01 1.13844089e-01 4.46283579e-01
6.49837017e-01 -1.41868436e+00 1.36031523e-01 6.93083942e-01
1.80664256e-01 7.03465223e-01 4.87809479e-01 -5.46570599e-01
-1.69281375e+00 -1.30523264e+00 3.25536191e-01 -9.86427724e-01
9.14934695e-01 -9.75324586e-02 -9.12483037e-01 9.75970924e-01
1.93585530e-01 9.27076757e-01 2.91610748e-01 -3.21220160e-01
2.96771377e-01 4.18826550e-01 -1.08266068e+00 3.87696326e-01
1.03756762e+00 -2.50394374e-01 -1.56176150e-01 9.05586302e-01
6.67510211e-01 -1.35223246e+00 -1.38405716e+00 6.73175097e-01
3.78816962e-01 -3.99587095e-01 1.16084564e+00 -5.85774481e-01
2.48396277e-01 -3.07993621e-01 5.76284647e-01 -1.44491076e+00
-4.28292841e-01 -6.69078290e-01 -8.73920247e-02 5.72739899e-01
2.69784540e-01 -2.51009256e-01 1.16154575e+00 4.98099029e-01
-4.86505657e-01 -7.24913895e-01 -1.39691770e+00 -5.49381614e-01
1.38297915e-01 -1.23790152e-01 9.13097709e-02 1.24516606e+00
-1.82338670e-01 -3.75356406e-01 -2.76167065e-01 6.06718898e-01
8.25987935e-01 1.37758374e-01 1.06494486e-01 -7.22134113e-01
-2.96966046e-01 -5.40571630e-01 -2.44989946e-01 -1.17723525e+00
1.77314460e-01 -1.38816488e+00 2.92724222e-01 -1.47087753e+00
2.61387914e-01 -7.24565029e-01 -5.10089874e-01 4.75108355e-01
-3.01956773e-01 4.20183837e-01 1.33945674e-01 2.95979440e-01
-6.55479252e-01 5.11182845e-01 1.89898586e+00 -2.45981723e-01
-2.76034381e-02 2.98906773e-01 -7.26278052e-02 4.59780633e-01
2.83629417e-01 -8.43956411e-01 7.24888369e-02 -4.99718726e-01
-2.01099589e-01 6.73712373e-01 7.10801005e-01 -9.37465131e-01
6.27056777e-01 1.70384035e-01 6.42583311e-01 -2.38580450e-01
5.22070043e-02 -1.19897759e+00 1.33649275e-01 1.05430532e+00
-3.44038218e-01 -2.76044726e-01 1.92086697e-01 4.37313378e-01
-9.11394879e-02 -2.80914187e-01 9.63421524e-01 -9.15165916e-02
-3.98716420e-01 8.10499072e-01 -1.80801749e-01 -1.26542389e-01
9.44326043e-01 2.51684427e-01 -8.60327259e-02 -1.12640053e-01
-1.41513753e+00 3.17635268e-01 1.08543523e-01 2.69138455e-01
4.02320117e-01 -1.44692361e+00 -6.50800824e-01 -9.23984349e-02
-1.10139072e-01 -1.44110795e-03 5.86420953e-01 1.69168258e+00
-5.89639664e-01 5.04963100e-01 1.60517618e-01 -1.18453979e+00
-9.09651518e-01 6.44765735e-01 1.18683326e+00 -6.67351425e-01
-1.01631749e+00 9.77942646e-01 3.62223834e-01 -4.68718141e-01
1.94469154e-01 -7.04551280e-01 -1.49254248e-01 -5.87236226e-01
2.53106773e-01 -4.69987303e-01 2.22230747e-01 -3.88873160e-01
-5.59492648e-01 8.47964108e-01 1.39594050e-02 2.24064499e-01
1.16860509e+00 2.04711594e-02 5.64008951e-02 -1.54188007e-01
1.22948682e+00 -5.03420178e-03 -1.52280807e+00 -5.32059610e-01
3.02479118e-01 -4.71705854e-01 2.98965365e-01 -9.09631908e-01
-1.48247504e+00 5.35104632e-01 1.01170135e+00 1.79987013e-01
8.62153113e-01 1.91010147e-01 6.88021064e-01 -1.31807327e-01
2.13488176e-01 -6.61765262e-02 8.03728029e-02 4.80716825e-01
7.77574480e-01 -1.25050461e+00 -2.10581794e-01 -4.83924121e-01
-3.42716306e-01 1.23090518e+00 3.77577573e-01 -8.66385773e-02
6.61678433e-01 4.28229541e-01 2.58948475e-01 -3.80263239e-01
-2.95390248e-01 -1.34951517e-01 6.22699976e-01 3.99165809e-01
7.73152590e-01 -6.78399131e-02 -7.12828487e-02 2.68924505e-01
3.44298929e-01 1.41029194e-01 1.26485735e-01 8.86167526e-01
-1.29863899e-02 -8.73223782e-01 -7.50601441e-02 4.82547581e-02
-7.24451840e-01 -1.74305677e-01 4.29243416e-01 6.71554863e-01
-1.95648596e-01 3.07660401e-01 -1.59890622e-01 1.38531819e-01
3.58127981e-01 -2.98603743e-01 8.36114585e-01 -5.33739328e-01
-9.87419784e-01 1.92168117e-01 -4.44511831e-01 -7.38922775e-01
-3.92768949e-01 -5.86366236e-01 -1.24660683e+00 1.58645198e-01
-2.44792163e-01 2.69150853e-01 7.12739050e-01 8.73948336e-01
2.91706696e-02 1.10473132e+00 6.51609719e-01 -9.76534188e-01
-6.32471323e-01 -6.96136713e-01 -3.29605907e-01 5.43219209e-01
5.63404441e-01 -3.68074596e-01 -1.65410027e-01 -4.36909795e-02] | [14.391539573669434, -2.6482324600219727] |
cd2549d6-be89-48de-ae0f-b51924b032b3 | physics-informed-deep-diffusion-mri | 2210.11388 | null | https://arxiv.org/abs/2210.11388v1 | https://arxiv.org/pdf/2210.11388v1.pdf | Physics-informed deep diffusion MRI reconstruction: break the bottleneck of training data in artificial intelligence | In this work, we propose a Physics-Informed Deep Diffusion magnetic resonance imaging (DWI) reconstruction method (PIDD). PIDD contains two main components: The multi-shot DWI data synthesis and a deep learning reconstruction network. For data synthesis, we first mathematically analyze the motion during the multi-shot data acquisition and approach it by a simplified physical motion model. The motion model inspires a polynomial model for motion-induced phase synthesis. Then, lots of synthetic phases are combined with a few real data to generate a large amount of training data. For reconstruction network, we exploit the smoothness property of each shot image phase as learnable convolution kernels in the k-space and complementary sparsity in the image domain. Results on both synthetic and in vivo brain data show that, the proposed PIDD trained on synthetic data enables sub-second ultra-fast, high-quality, and robust reconstruction with different b-values and undersampling patterns. | ['Xiaobo Qu', 'Di Guo', 'Zhigang Wu', 'Ran Tao', 'Boyu Jiang', 'Taishan Kang', 'Qingrui Cai', 'Xinlin Zhang', 'Zi Wang', 'Chen Qian'] | 2022-10-20 | null | null | null | null | ['mri-reconstruction'] | ['computer-vision'] | [ 1.54818445e-01 -3.01318467e-01 6.92197606e-02 -3.22512239e-01
-5.85269332e-01 1.03244983e-01 4.66595918e-01 -4.17624116e-01
-3.77332926e-01 7.38586605e-01 3.49856317e-01 1.00331753e-01
-4.47114468e-01 -6.28323615e-01 -5.11180222e-01 -1.05839038e+00
-1.12121284e-01 3.75460327e-01 4.84527946e-01 4.78754053e-03
4.21357714e-02 5.40870130e-01 -8.56551230e-01 8.27083811e-02
8.19016516e-01 7.02315867e-01 7.16677189e-01 3.92252743e-01
8.33843425e-02 1.19058418e+00 -2.44348675e-01 3.88210505e-01
1.55923858e-01 -6.00165427e-01 -8.07294250e-01 2.51159444e-02
-2.05193609e-01 -7.89349318e-01 -9.09369051e-01 8.18165243e-01
9.79029179e-01 3.15864235e-01 6.31577253e-01 -6.69401824e-01
-6.16928279e-01 5.56325376e-01 -6.48009241e-01 7.00706542e-01
-1.03059644e-02 6.25228703e-01 1.08993858e-01 -9.29544747e-01
1.27193356e+00 6.78927183e-01 5.55449665e-01 6.41971886e-01
-1.26296425e+00 -4.39610958e-01 -5.76876640e-01 4.79558349e-01
-1.04424262e+00 -4.07373905e-01 1.21411705e+00 -6.52849317e-01
6.33500218e-01 -1.73544839e-01 7.97570169e-01 1.04657352e+00
6.83506250e-01 7.04349637e-01 1.30065584e+00 -3.24634343e-01
3.19912463e-01 -4.74621266e-01 3.47690552e-01 4.90053594e-01
3.62874642e-02 2.89417267e-01 -5.08209527e-01 9.30916443e-02
1.20337057e+00 8.21112394e-02 -6.94499373e-01 -2.41347834e-01
-1.64561009e+00 4.11277503e-01 3.22440177e-01 5.74825644e-01
-7.65689611e-01 6.79648072e-02 4.21882361e-01 1.85524821e-01
3.58063906e-01 1.87264845e-01 5.50971515e-02 1.04321972e-01
-1.07941687e+00 1.66359290e-01 4.27475601e-01 5.75729311e-01
4.09635723e-01 4.82772678e-01 -4.40769225e-01 7.39122570e-01
2.55966634e-01 5.57141066e-01 9.88452375e-01 -1.03405964e+00
6.27994025e-03 -1.42083317e-01 2.01439738e-01 -8.65003109e-01
-7.12496221e-01 -2.22704470e-01 -1.21048307e+00 1.80468410e-01
4.50858623e-01 -2.06079349e-01 -9.56706166e-01 1.69036806e+00
4.87833112e-01 5.77108324e-01 -1.13884032e-01 1.38687241e+00
9.14516926e-01 6.42781496e-01 -1.54665157e-01 -8.12837422e-01
1.18688035e+00 -7.86716938e-01 -1.23259449e+00 1.84690759e-01
3.88108671e-01 -3.63041490e-01 7.87442803e-01 2.65814692e-01
-1.37953031e+00 -5.01347065e-01 -1.25354767e+00 8.72296318e-02
3.83062392e-01 -2.55271465e-01 3.92141134e-01 2.36619432e-02
-8.17400813e-01 1.02000201e+00 -1.09826660e+00 2.17849970e-01
4.48187053e-01 7.56123960e-02 -3.10378671e-01 -4.50981289e-01
-1.35611069e+00 8.15378249e-01 2.59198308e-01 -1.18599487e-02
-1.20361662e+00 -1.35655975e+00 -5.62487960e-01 -3.84293228e-01
-2.18338743e-02 -6.86288416e-01 1.06982362e+00 -6.45994067e-01
-1.77085721e+00 7.56467164e-01 7.08671112e-04 -4.57392097e-01
5.28147697e-01 2.04730287e-01 -5.62553585e-01 8.27302277e-01
-8.35483603e-04 2.85341024e-01 1.28645539e+00 -9.76107121e-01
1.79622605e-01 -2.90461630e-01 -4.20983613e-01 -5.17079532e-02
2.42685407e-01 -2.28381027e-02 4.36061248e-02 -7.90914237e-01
2.67201602e-01 -7.12214828e-01 -2.44525850e-01 2.39448473e-01
-1.11579545e-01 5.24665952e-01 5.94607353e-01 -1.12271881e+00
9.81260061e-01 -2.15410161e+00 3.58207971e-01 -1.39224738e-01
7.35743821e-01 2.80091107e-01 -8.00859034e-02 4.96861711e-02
-4.37955856e-01 -6.06568873e-01 -5.61769903e-01 1.85905453e-02
-4.89283055e-01 1.78632438e-02 -1.63639873e-01 8.60982120e-01
8.58633406e-03 1.15131176e+00 -1.22990870e+00 -4.12303269e-01
3.56153011e-01 5.73948801e-01 -3.77256393e-01 3.40273798e-01
1.18632041e-01 1.21515679e+00 -4.21341419e-01 2.30863318e-01
1.01386201e+00 -4.28473473e-01 2.00728804e-01 -6.30542278e-01
-2.90167302e-01 -7.16730803e-02 -9.42272246e-01 2.21500134e+00
-2.53735811e-01 4.12741363e-01 2.02952355e-01 -1.21907830e+00
5.37296176e-01 6.57061815e-01 1.19158137e+00 -1.05738938e+00
1.39485285e-01 4.83979881e-01 7.76512250e-02 -9.13451433e-01
-2.86664814e-02 -5.65772414e-01 5.96743703e-01 8.13725889e-01
3.59631687e-01 -1.25267774e-01 -1.73668214e-03 4.20032106e-02
1.12203193e+00 1.02172762e-01 -6.79442510e-02 -5.63234985e-01
4.76751924e-01 -5.40568382e-02 5.39877832e-01 5.19960761e-01
-3.58311206e-01 7.34922230e-01 1.75814047e-01 -6.72306418e-01
-1.57680726e+00 -9.73332047e-01 -3.78828526e-01 3.79872806e-02
2.52654940e-01 1.40044019e-01 -7.94952095e-01 -4.32162546e-02
-4.71867144e-01 5.09445727e-01 -4.72006828e-01 -3.71774405e-01
-1.00933921e+00 -1.14000857e+00 7.30772391e-02 1.72668025e-01
5.92120886e-01 -1.21852827e+00 -7.88057446e-01 7.08193243e-01
-4.21037555e-01 -1.05650687e+00 -4.51546997e-01 -2.11284142e-02
-9.91556704e-01 -9.55890179e-01 -1.28504574e+00 -6.66014075e-01
5.39605021e-01 5.12317181e-01 7.06633151e-01 -3.53791416e-01
-6.80868268e-01 1.67302385e-01 -2.10548803e-01 1.65565595e-01
-4.13040251e-01 -6.04219258e-01 1.60595283e-01 1.59333318e-01
-1.20145395e-01 -1.06662023e+00 -1.11427379e+00 1.30910063e-02
-9.45000052e-01 4.63241100e-01 3.83545995e-01 1.04938293e+00
6.91810787e-01 -1.22958906e-01 6.19358897e-01 -3.31167072e-01
6.32824063e-01 -7.24364519e-01 -3.58857840e-01 1.23039447e-01
-3.67259473e-01 3.21092010e-01 5.62638879e-01 -8.88257384e-01
-1.13145876e+00 -2.89302856e-01 -1.81589216e-01 -7.15688109e-01
7.10137412e-02 3.17546666e-01 2.31250584e-01 -3.12856108e-01
7.47008145e-01 7.21560180e-01 4.93514806e-01 -3.10519367e-01
2.30398446e-01 2.43980706e-01 5.97640574e-01 -4.54272032e-01
5.62119484e-01 9.17244792e-01 6.91730678e-02 -1.12453806e+00
-5.78189611e-01 -1.60383403e-01 -7.43550718e-01 -5.13334453e-01
1.08994317e+00 -6.07074916e-01 -5.01902044e-01 1.08718371e+00
-1.15455639e+00 -7.16325521e-01 -6.65351927e-01 9.10168648e-01
-8.12057257e-01 6.34015918e-01 -1.09684372e+00 -2.33335599e-01
-7.25323796e-01 -1.53308105e+00 5.89139998e-01 4.97548468e-02
1.17529981e-01 -8.41174245e-01 3.39202076e-01 2.26336643e-02
6.26384795e-01 3.69631618e-01 9.29162800e-01 -8.99580494e-02
-7.56798089e-01 2.54726410e-01 -1.34358823e-01 1.81142718e-01
2.06516385e-02 -4.32530254e-01 -6.21980608e-01 -2.38509133e-01
8.17155302e-01 -2.38453344e-01 5.22542179e-01 1.13487744e+00
1.05318558e+00 -8.74268711e-02 -3.95227298e-02 1.09697771e+00
1.40088212e+00 3.61496836e-01 8.11293185e-01 3.29105780e-02
5.98447621e-01 3.83917481e-01 2.40813226e-01 5.51194966e-01
6.37662113e-02 5.80919862e-01 -1.82227008e-02 -2.30995938e-02
-5.59590340e-01 1.16707608e-01 6.35947213e-02 1.35981882e+00
-1.87292293e-01 2.57065862e-01 -8.22047830e-01 4.83148634e-01
-1.62700284e+00 -1.19159782e+00 -3.29648316e-01 2.05684781e+00
9.35853541e-01 -1.98091596e-01 -1.30831972e-01 1.28668934e-01
5.14938951e-01 3.20521325e-01 -7.93944180e-01 4.55021590e-01
-9.32050794e-02 3.47224206e-01 2.62902617e-01 5.39715588e-01
-5.95826566e-01 5.22491276e-01 6.37239742e+00 7.23276258e-01
-1.61383951e+00 6.55788898e-01 4.10025060e-01 -1.23763122e-01
-4.60013479e-01 -2.61321992e-01 -1.84603110e-01 7.63973296e-01
9.71929252e-01 -3.17003280e-01 5.86889744e-01 2.82128543e-01
5.97674787e-01 -4.84501012e-02 -5.18451810e-01 1.18502212e+00
-2.77909994e-01 -1.78589809e+00 -1.74149156e-01 -1.62104443e-01
5.39494634e-01 3.00547451e-01 -1.81311101e-01 -1.95420653e-01
-1.06525704e-01 -8.01783681e-01 7.48279333e-01 1.01174140e+00
1.13717103e+00 -4.39614505e-01 1.67017207e-01 5.22415936e-01
-7.13702142e-01 1.31002650e-01 -2.87931025e-01 3.04138929e-01
6.20769441e-01 8.45159233e-01 -2.64765143e-01 6.83870137e-01
3.60252529e-01 9.14851010e-01 1.05542526e-01 9.68105793e-01
-4.50881124e-02 4.78488654e-01 -1.91253945e-02 4.40492332e-01
8.98444504e-02 -4.77900475e-01 7.24052727e-01 8.25433850e-01
3.28768045e-01 6.42000556e-01 -2.37031311e-01 1.10211027e+00
2.50278771e-01 -2.36797914e-01 -3.74968350e-01 1.30775958e-01
1.95811883e-01 1.17936718e+00 -5.31491756e-01 -4.16217864e-01
-2.91178077e-01 1.08687127e+00 8.62126276e-02 6.02920532e-01
-7.73129940e-01 -1.12754360e-01 2.34772846e-01 3.83990824e-01
-2.81677488e-02 -5.51258087e-01 6.49765730e-02 -1.51221502e+00
-9.25943628e-02 -5.42057514e-01 -1.85301453e-01 -9.64766026e-01
-1.06693602e+00 5.23796380e-01 1.73440620e-01 -1.21731281e+00
-2.10196435e-01 -3.73312116e-01 -7.38547742e-01 9.31659341e-01
-1.68527114e+00 -6.96925461e-01 -4.50025588e-01 7.59399176e-01
4.59425598e-01 9.15390104e-02 6.20295644e-01 5.52832425e-01
-4.23290044e-01 -1.31540805e-01 2.75294989e-01 1.36793293e-02
5.42694032e-01 -7.24701703e-01 2.92882830e-01 7.96742380e-01
-4.95695949e-01 4.72787112e-01 5.88890135e-01 -7.25404263e-01
-1.57971859e+00 -8.29742610e-01 2.45203450e-01 2.10472390e-01
7.81014025e-01 2.13809431e-01 -1.26334727e+00 3.38672727e-01
6.62247017e-02 6.37131214e-01 2.30463743e-01 -1.10306263e+00
2.28631884e-01 -1.13482038e-02 -1.44854534e+00 3.37454736e-01
8.72632623e-01 -4.53104466e-01 -7.16912210e-01 4.88419145e-01
8.63928378e-01 -4.78002429e-01 -1.26550937e+00 2.45891571e-01
5.04119456e-01 -8.09398174e-01 1.20894730e+00 -4.49688017e-01
6.43497229e-01 -1.23643406e-01 2.32501149e-01 -1.42424786e+00
-5.69654346e-01 -7.86084950e-01 -5.03493369e-01 4.12709594e-01
-2.30613783e-01 -4.97552603e-01 4.37695175e-01 4.50895220e-01
-3.74039948e-01 -7.24927068e-01 -9.45907593e-01 -6.13880277e-01
4.71010506e-02 -2.91900724e-01 4.30162728e-01 1.22503114e+00
-1.12059891e-01 8.93644243e-03 -5.65109253e-01 -3.28022540e-02
1.26297545e+00 -4.04703878e-02 1.06302291e-01 -8.00735772e-01
-3.49222988e-01 -1.03394754e-01 -9.70415920e-02 -9.91262972e-01
-5.23177534e-02 -8.84867013e-01 -1.64498165e-01 -1.41354287e+00
1.47468150e-01 -5.54834306e-01 -3.22166622e-01 -1.34671211e-01
3.15121189e-02 -8.01461935e-02 -9.59458351e-02 7.41095006e-01
3.62399295e-02 6.18930757e-01 2.03932643e+00 -9.88870487e-02
-1.35837495e-01 -3.62383127e-01 5.67092896e-02 3.57667536e-01
5.18284440e-01 -3.61692309e-01 -5.79477429e-01 -5.32228410e-01
-2.22725630e-01 8.62191737e-01 4.26439285e-01 -1.15952051e+00
2.79087991e-01 -1.69576839e-01 2.78668761e-01 -4.99676526e-01
2.56519049e-01 -5.81327081e-01 4.98224676e-01 8.45215023e-01
-9.87303406e-02 -4.65956628e-01 -1.79000258e-01 3.44832987e-01
-9.62521210e-02 -2.55600780e-01 1.34290159e+00 -3.87517631e-01
-6.34157360e-01 8.88790131e-01 -1.94751039e-01 1.97555974e-01
9.88359869e-01 -1.49451807e-01 -1.05403244e-01 1.52063472e-02
-9.74989593e-01 -2.12719321e-01 2.19694182e-01 -6.27744496e-02
9.52710509e-01 -1.46278811e+00 -7.54323423e-01 5.22309959e-01
-3.19281131e-01 -5.17019304e-03 1.04406488e+00 1.53358400e+00
-8.19759667e-01 3.37167382e-02 -7.50436425e-01 -6.33677125e-01
-2.97234803e-01 6.91645145e-01 7.17737794e-01 -3.52192491e-01
-1.56229639e+00 5.40743828e-01 -7.41053745e-02 -1.14828050e-01
-2.58163512e-01 -3.47520530e-01 -9.18330550e-02 -1.26593858e-01
1.10059536e+00 2.94736385e-01 -6.16718978e-02 -5.33548176e-01
-1.62359044e-01 6.01946473e-01 1.22508176e-01 -2.95680255e-01
1.84996438e+00 -8.21713880e-02 -3.42874788e-02 4.37843531e-01
1.12637389e+00 -4.87417519e-01 -1.64073896e+00 -4.62590814e-01
-2.94948429e-01 -1.93861470e-01 3.73073399e-01 -3.86699945e-01
-1.26584864e+00 9.79832113e-01 6.49396658e-01 -2.11147726e-01
9.71727610e-01 -3.14873904e-01 1.26311326e+00 -1.04730584e-01
5.59428751e-01 -8.89096737e-01 2.92955995e-01 2.86566824e-01
8.94061506e-01 -9.37460840e-01 -7.83592090e-02 2.56005581e-02
-5.52154005e-01 1.28326285e+00 1.20414138e-01 -4.23396558e-01
8.66069376e-01 3.90168846e-01 -2.68640220e-01 -4.30376232e-01
-4.33012366e-01 4.06821012e-01 3.04670874e-02 7.42996693e-01
6.52377680e-02 -1.78736702e-01 -4.74541634e-01 4.57215309e-01
4.50098813e-01 6.57615244e-01 6.17044866e-01 8.69272292e-01
-4.48961943e-01 -6.14332497e-01 -2.06428781e-01 2.59756476e-01
-2.59464264e-01 3.41201052e-02 6.48582995e-01 4.44346726e-01
-1.53820500e-01 3.11650008e-01 -7.01760203e-02 -4.79767695e-02
1.66569501e-01 8.23370833e-03 1.00795031e+00 -1.83987662e-01
-1.18357264e-01 9.38420277e-03 -3.48564178e-01 -7.54402041e-01
-5.35993993e-01 -3.61999273e-01 -1.55511820e+00 -2.60530829e-01
1.41604811e-01 -1.48299098e-01 6.82738304e-01 1.02931595e+00
2.95413136e-01 8.98113489e-01 7.09319413e-01 -1.08912587e+00
-6.68825626e-01 -1.02327228e+00 -9.20024931e-01 5.53142786e-01
4.66411859e-01 -6.98494136e-01 -3.38434398e-01 4.22596596e-02] | [13.544145584106445, -2.3991639614105225] |
0adcaefe-32de-4175-80a0-46a92fa9696b | d-2-nerf-self-supervised-decoupling-of | 2205.15838 | null | https://arxiv.org/abs/2205.15838v4 | https://arxiv.org/pdf/2205.15838v4.pdf | D$^2$NeRF: Self-Supervised Decoupling of Dynamic and Static Objects from a Monocular Video | Given a monocular video, segmenting and decoupling dynamic objects while recovering the static environment is a widely studied problem in machine intelligence. Existing solutions usually approach this problem in the image domain, limiting their performance and understanding of the environment. We introduce Decoupled Dynamic Neural Radiance Field (D$^2$NeRF), a self-supervised approach that takes a monocular video and learns a 3D scene representation which decouples moving objects, including their shadows, from the static background. Our method represents the moving objects and the static background by two separate neural radiance fields with only one allowing for temporal changes. A naive implementation of this approach leads to the dynamic component taking over the static one as the representation of the former is inherently more general and prone to overfitting. To this end, we propose a novel loss to promote correct separation of phenomena. We further propose a shadow field network to detect and decouple dynamically moving shadows. We introduce a new dataset containing various dynamic objects and shadows and demonstrate that our method can achieve better performance than state-of-the-art approaches in decoupling dynamic and static 3D objects, occlusion and shadow removal, and image segmentation for moving objects. | ['Cengiz Oztireli', 'Forrester Cole', 'Andrea Tagliasacchi', 'Fangcheng Zhong', 'Tianhao Wu'] | 2022-05-31 | null | null | null | null | ['shadow-removal'] | ['computer-vision'] | [ 6.17495060e-01 -3.41541439e-01 1.34764493e-01 -3.77383471e-01
-2.33690500e-01 -7.15938687e-01 4.67483342e-01 -5.03827691e-01
-3.80878299e-01 6.29608631e-01 -1.85456634e-01 -2.51845449e-01
2.57832527e-01 -6.64054155e-01 -9.35137510e-01 -1.26701713e+00
-2.55378690e-02 1.67261109e-01 9.72811818e-01 -4.48221080e-02
-9.81507152e-02 6.42310739e-01 -1.69775188e+00 2.34702051e-01
8.50948453e-01 9.36341226e-01 6.99004829e-01 1.09299529e+00
-1.88045576e-02 1.14077640e+00 -7.98375070e-01 1.33980721e-01
7.39369392e-01 -3.11601311e-01 -3.75110894e-01 1.50946498e-01
9.81691360e-01 -7.11404860e-01 -7.83001661e-01 8.85077119e-01
3.51816446e-01 3.63983959e-01 3.91394913e-01 -1.26867831e+00
-5.27855217e-01 -5.23113925e-03 -6.83392644e-01 4.85153198e-01
5.41436598e-02 9.16335881e-02 4.69873995e-01 -4.54980493e-01
6.28630936e-01 1.23716366e+00 5.80980599e-01 5.06927729e-01
-1.22975791e+00 -5.52303374e-01 7.40581036e-01 2.22759992e-01
-1.04762018e+00 -3.16951811e-01 9.74153757e-01 -4.06210393e-01
7.49051154e-01 5.36476552e-01 5.73553681e-01 1.13504529e+00
2.89568692e-01 1.06732416e+00 1.11705697e+00 -1.62813365e-01
2.24612713e-01 2.03127995e-01 3.80738616e-01 4.74217117e-01
2.35884905e-01 2.79940665e-01 -2.64129579e-01 -5.86210079e-02
5.27441025e-01 1.65280551e-01 -7.51008213e-01 -7.80704319e-01
-7.77595460e-01 2.24425972e-01 3.26281071e-01 5.88159598e-02
-2.15200968e-02 3.75215381e-01 -6.31039590e-02 6.27811775e-02
5.89622557e-01 -7.35781016e-03 -6.18855715e-01 1.93642035e-01
-9.62717772e-01 3.61169785e-01 6.01579070e-01 6.67143643e-01
7.91241050e-01 2.50526190e-01 -1.04523905e-01 6.11950517e-01
1.65995043e-02 1.00829899e+00 2.33456790e-01 -9.78146315e-01
5.00459783e-02 4.53088522e-01 3.99581283e-01 -1.14467359e+00
-3.66151303e-01 -4.05900121e-01 -6.23803377e-01 5.09961069e-01
4.07091767e-01 -9.22013894e-02 -1.41459513e+00 1.83677971e+00
6.65216625e-01 6.75045550e-01 5.55362925e-02 1.04786062e+00
9.11216378e-01 8.77482057e-01 -2.56619453e-01 -1.21937744e-01
1.01722741e+00 -1.20105994e+00 -5.92327297e-01 -6.64406419e-01
1.74474511e-02 -5.17704427e-01 7.12095380e-01 3.78984690e-01
-8.87964666e-01 -6.14454448e-01 -1.10686362e+00 -1.89325422e-01
-5.20015657e-01 8.66008457e-03 6.91123724e-01 5.08121967e-01
-8.82565737e-01 5.19176126e-01 -9.62567449e-01 -5.23197614e-02
2.67417043e-01 3.48915040e-01 -1.28135055e-01 -1.52789265e-01
-1.03612804e+00 7.44664073e-01 2.20877036e-01 2.78303802e-01
-1.00033379e+00 -7.43230283e-01 -7.98986733e-01 -7.32480064e-02
7.43184447e-01 -5.83546281e-01 1.04695380e+00 -1.42293322e+00
-1.37335086e+00 6.98680222e-01 -3.18725914e-01 -3.01511765e-01
6.13224030e-01 -5.79781711e-01 -2.85191357e-01 2.29030058e-01
-1.38255864e-01 3.86457086e-01 1.11479020e+00 -1.85463345e+00
-6.80835664e-01 -3.56235206e-01 3.93564939e-01 4.90885019e-01
3.38013582e-02 -2.74913579e-01 -7.12821603e-01 -4.88379300e-01
1.26454562e-01 -1.08487606e+00 -1.15614958e-01 1.90006152e-01
-2.11096197e-01 5.15541732e-01 1.37194049e+00 -7.60875821e-01
9.92112637e-01 -2.00887966e+00 2.80836552e-01 -2.59427875e-01
3.17884326e-01 4.07378435e-01 -7.82909244e-02 -2.54140168e-01
-6.65231422e-02 -5.19359291e-01 -5.13931155e-01 -2.83896416e-01
-2.69652098e-01 4.52259988e-01 -4.92175311e-01 6.94795847e-01
-9.35585350e-02 7.22070217e-01 -1.05918050e+00 -2.76072681e-01
3.35004330e-01 7.92292655e-01 -2.94604391e-01 3.15706253e-01
-3.74909788e-01 5.01984656e-01 -3.18486065e-01 6.33223712e-01
1.26028669e+00 -7.23500699e-02 1.44818321e-01 -1.39083147e-01
-9.61063281e-02 -1.00872390e-01 -1.42096782e+00 1.21681488e+00
-3.97469401e-01 1.03940105e+00 5.26210427e-01 -6.62638605e-01
6.27251863e-01 -1.77073423e-02 6.47955477e-01 -8.14692318e-01
1.19840547e-01 4.00069281e-02 -2.60662884e-01 -6.03925169e-01
6.28652096e-01 -1.59781277e-01 2.01558888e-01 2.62899876e-01
-2.64534235e-01 -1.86963543e-01 -8.93527642e-02 8.32625404e-02
1.12405002e+00 6.19165957e-01 -1.79891899e-01 -1.23279177e-01
4.72239345e-01 -1.96475968e-01 8.84730279e-01 9.00697470e-01
-3.01714897e-01 7.86387861e-01 1.19123727e-01 -5.06240964e-01
-4.32137549e-01 -1.14305949e+00 5.16267307e-02 1.15468109e+00
9.43412840e-01 2.08371058e-01 -7.39194989e-01 -7.04430044e-01
3.22394282e-01 5.88317156e-01 -6.48572922e-01 -2.30234802e-01
-1.00052094e+00 -1.15182436e+00 2.34536812e-01 3.35915327e-01
5.44345915e-01 -7.93853700e-01 -1.40639365e+00 -4.43532057e-02
-5.20373583e-01 -1.40265357e+00 -3.04609835e-01 6.40985906e-01
-5.88465035e-01 -1.10538852e+00 -6.58925056e-01 -4.68208671e-01
4.69574690e-01 1.03715587e+00 1.23496461e+00 -1.48147624e-02
-6.10198319e-01 4.08523351e-01 1.09283067e-02 -3.56296122e-01
-2.14774877e-01 -3.42201799e-01 -7.89219365e-02 3.25720251e-01
1.54069886e-01 -6.54134572e-01 -8.36726248e-01 3.19409311e-01
-1.20488465e+00 2.74806201e-01 4.00059283e-01 3.60807359e-01
4.31455761e-01 1.61499739e-01 -3.46548140e-01 -9.65194702e-01
-3.02663267e-01 -3.93197060e-01 -8.29710484e-01 1.65418699e-01
-2.17245728e-01 -1.06332786e-01 6.51110768e-01 -5.37149370e-01
-1.52348900e+00 3.21002543e-01 4.54266429e-01 -6.59803927e-01
-3.62851977e-01 -3.10190529e-01 -4.69408512e-01 -2.23192140e-01
5.94171107e-01 3.23840946e-01 -5.33268511e-01 -5.48447907e-01
2.75298953e-01 3.86217237e-01 9.38107848e-01 -3.79816890e-01
1.15134239e+00 1.18675685e+00 4.25438583e-02 -1.10208595e+00
-1.10926497e+00 -7.27526188e-01 -7.89314985e-01 -4.31398511e-01
9.17186439e-01 -9.10017192e-01 -2.47777969e-01 7.82685995e-01
-1.24292755e+00 -8.57856154e-01 -2.54945427e-01 6.25325814e-02
-4.06434685e-01 4.27559823e-01 -3.71497005e-01 -9.12090242e-01
1.03099719e-02 -1.03732514e+00 1.49955928e+00 2.69233078e-01
4.05755073e-01 -9.40193057e-01 3.25722784e-01 2.62875915e-01
2.61160403e-01 6.38350010e-01 7.64514446e-01 -1.11638950e-02
-1.09645844e+00 1.72010317e-01 -4.26512271e-01 3.87317777e-01
1.92663074e-01 -8.37740079e-02 -1.43552053e+00 -2.66477227e-01
4.42555577e-01 1.09358214e-01 1.27615619e+00 5.93019724e-01
9.53621864e-01 -2.27951054e-02 -5.81583619e-01 1.04721212e+00
1.71476376e+00 5.09483635e-01 4.57835913e-01 3.03928554e-01
1.21576679e+00 6.72290623e-01 4.75553989e-01 1.73790827e-01
2.15890959e-01 8.33862484e-01 5.13880789e-01 -4.92315620e-01
-4.93046403e-01 2.49641120e-01 5.06536067e-01 3.75705242e-01
5.18273227e-02 -6.09289765e-01 -7.28246391e-01 4.80884731e-01
-1.95984519e+00 -9.67008471e-01 -1.76757708e-01 1.94177568e+00
5.91065586e-01 1.22743748e-01 -2.43755519e-01 -5.45794666e-02
4.70210820e-01 6.20564461e-01 -9.01606143e-01 7.56803676e-02
-5.46667039e-01 -6.26301244e-02 7.62322903e-01 8.20573032e-01
-1.36900294e+00 1.02560914e+00 5.98099518e+00 3.82486850e-01
-1.41615891e+00 5.24377152e-02 4.47713017e-01 -3.24684292e-01
-3.01110208e-01 4.43122908e-02 -8.34274411e-01 2.88505077e-01
6.07669353e-01 4.15369302e-01 4.74585265e-01 6.59413099e-01
1.60755813e-01 -5.53404868e-01 -1.01970506e+00 9.79641378e-01
3.89510721e-01 -8.57537031e-01 -1.36385024e-01 -1.89913020e-01
9.47123110e-01 1.51463076e-01 -2.46032383e-02 8.49403143e-02
2.03429222e-01 -8.21981370e-01 9.61218715e-01 7.16572106e-01
3.74753147e-01 -2.30672956e-01 3.51453722e-01 3.90321374e-01
-1.23749793e+00 -2.51242012e-01 -1.67912841e-01 2.79567260e-02
1.94561973e-01 7.10453689e-01 -2.17015475e-01 5.19215107e-01
1.01052499e+00 6.20361686e-01 -6.42220259e-01 7.37197876e-01
1.32941678e-01 3.21886241e-01 -4.27198470e-01 4.48168337e-01
1.07538827e-01 -3.48773479e-01 8.83320570e-01 1.38396955e+00
-2.18134299e-01 2.02257290e-01 3.75552922e-01 7.27372885e-01
1.44047350e-01 -3.78524661e-01 -5.01615763e-01 4.10520345e-01
-1.62111402e-01 1.06617653e+00 -1.08769965e+00 -3.36116582e-01
-4.40511554e-01 1.32301426e+00 1.45626351e-01 7.94950545e-01
-1.16535234e+00 -1.84441403e-01 7.32334077e-01 9.68301445e-02
5.99129856e-01 -3.58708829e-01 -9.12413597e-02 -1.52143753e+00
2.33599365e-01 -6.51077688e-01 -2.82449853e-02 -9.35702443e-01
-7.05265641e-01 6.23391747e-01 2.28811547e-01 -9.20471430e-01
1.79822311e-01 -8.56158018e-01 -3.06059331e-01 5.37058115e-01
-1.97293878e+00 -8.96848321e-01 -9.07393694e-01 6.10143125e-01
6.87304378e-01 3.80544245e-01 2.46295676e-01 4.58495975e-01
-4.99484479e-01 8.72229785e-02 3.69644672e-01 -1.48947567e-01
7.80396402e-01 -1.42209458e+00 3.02754194e-01 1.23349011e+00
2.89386678e-02 1.34773001e-01 9.21423674e-01 -7.47611940e-01
-1.44615734e+00 -1.19350505e+00 3.00216168e-01 -7.62656987e-01
2.93822080e-01 -7.92519033e-01 -1.06448901e+00 5.18294394e-01
-2.18689471e-01 1.73502639e-01 2.35250250e-01 -4.54950809e-01
-3.39228868e-01 -3.09994698e-01 -9.73012984e-01 7.86823452e-01
1.21198583e+00 -3.81746769e-01 -2.72260666e-01 3.15484315e-01
8.26350808e-01 -6.20388329e-01 -1.39442787e-01 4.87925410e-01
5.65229595e-01 -1.39214087e+00 1.22368252e+00 -1.75670043e-01
1.87121689e-01 -5.51591456e-01 -4.16821629e-01 -9.44345713e-01
6.38377992e-03 -6.00304425e-01 -6.79908097e-01 7.63637662e-01
-1.91734433e-01 -2.83497214e-01 1.01959920e+00 4.81904984e-01
-1.31028339e-01 -5.41529357e-01 -6.77256882e-01 -7.22015500e-01
-1.00554809e-01 -3.56408864e-01 2.99279809e-01 7.05451250e-01
-9.83573973e-01 9.69264731e-02 -5.19288719e-01 5.81607640e-01
7.53828943e-01 6.60213232e-01 8.59221518e-01 -1.26652467e+00
-5.67006111e-01 -1.50754467e-01 -1.50137514e-01 -1.41644239e+00
2.91281343e-01 -4.81735617e-01 6.18732095e-01 -1.33595800e+00
3.75572056e-01 -2.61817515e-01 -1.69356823e-01 2.28555039e-01
-3.36662054e-01 3.81318152e-01 2.19873860e-01 1.13623045e-01
-6.13830388e-01 6.01503313e-01 1.18336940e+00 -1.62846088e-01
-4.62768167e-01 1.87447853e-02 -3.07132304e-01 1.01544046e+00
4.51093644e-01 -3.61836433e-01 -6.18610382e-01 -7.56679296e-01
-1.81777805e-01 -1.58724606e-01 6.72468662e-01 -9.71161306e-01
5.91603331e-02 -3.79819304e-01 4.59047377e-01 -7.15260923e-01
6.56391978e-01 -1.06664300e+00 2.44542584e-01 2.88771451e-01
1.41952604e-01 -3.29160243e-01 4.31338489e-01 8.49773467e-01
1.13013476e-01 5.86894266e-02 9.72732604e-01 -2.99419731e-01
-8.67743015e-01 1.87100768e-01 -3.40717047e-01 3.25984918e-02
1.00225890e+00 -4.07840937e-01 -3.68820012e-01 -3.29756200e-01
-5.09897172e-01 9.62529704e-02 5.55659235e-01 4.95026797e-01
5.32767653e-01 -7.77598381e-01 -2.96718329e-01 1.56818047e-01
-3.03518981e-01 1.77471966e-01 4.35295850e-01 6.78220749e-01
-7.31415272e-01 2.54262328e-01 8.33946629e-04 -6.95085227e-01
-1.52869236e+00 5.76175690e-01 8.30907285e-01 -1.01326935e-01
-8.40292931e-01 7.32413769e-01 1.09674144e+00 -3.93380821e-01
4.55899209e-01 -5.16455472e-01 5.36276214e-02 -1.25248685e-01
4.26572740e-01 5.58294892e-01 -8.84417742e-02 -8.19704294e-01
-2.77462214e-01 8.59123111e-01 9.70701128e-02 3.91518734e-02
1.25310171e+00 -4.71886426e-01 2.62400531e-03 7.31561720e-01
1.24772334e+00 7.60693401e-02 -2.00224853e+00 -3.84897083e-01
-3.14093053e-01 -5.99091053e-01 3.13510716e-01 -7.92646110e-01
-1.23908901e+00 8.02637815e-01 7.72850931e-01 5.13358638e-02
1.34655702e+00 -1.82253987e-01 1.01215398e+00 2.75909483e-01
1.63080752e-01 -9.36156452e-01 -1.94972809e-02 5.89273632e-01
4.44622993e-01 -1.21274304e+00 1.96592048e-01 -7.03005612e-01
-3.47039968e-01 1.05175853e+00 7.60563433e-01 -2.07560584e-01
6.88210011e-01 6.03409886e-01 2.99407393e-01 1.99924572e-03
-5.20237982e-01 -2.51516372e-01 3.42744112e-01 6.01556599e-01
-1.82389840e-01 -1.42411113e-01 5.06372929e-01 2.50985831e-01
1.20747030e-01 -4.26364750e-01 5.51895440e-01 1.13380134e+00
-3.16909462e-01 -6.01884365e-01 -5.72905481e-01 -3.81907411e-02
-3.54162753e-01 -5.71977235e-02 -5.38370550e-01 7.76345670e-01
5.38632095e-01 7.28494346e-01 8.05875361e-02 -9.88468453e-02
3.18182260e-01 -9.86749828e-02 4.86787468e-01 -3.93542975e-01
-4.26293463e-01 3.07880729e-01 -2.96422631e-01 -7.62608826e-01
-6.66749060e-01 -6.61571860e-01 -1.07509243e+00 4.23959680e-02
-3.62967253e-01 -2.17628643e-01 5.88606000e-01 8.67917180e-01
1.76939234e-01 8.21017683e-01 5.31110883e-01 -1.32046163e+00
-1.02171749e-01 -2.63150930e-01 -6.03065848e-01 3.88146281e-01
1.00497270e+00 -9.19598997e-01 -5.98202467e-01 1.60138309e-01] | [10.824285507202148, -4.069308757781982] |
2e633fd4-28ab-41a8-8ec8-7bcdf0e73509 | lmsim-computing-domain-specific-semantic-word | null | null | https://aclanthology.org/W14-5116 | https://aclanthology.org/W14-5116.pdf | LMSim : Computing Domain-specific Semantic Word Similarities Using a Language Modeling Approach | null | ['Sachin Pawar', 'Swapnil Hingmire', 'Girish K. Palshikar'] | 2014-12-01 | null | null | null | ws-2014-12 | ['text-clustering'] | ['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] | [-7.258103370666504, 3.826150417327881] |
dcfba564-18a1-410a-815e-5ee9d60afc11 | large-scale-genealogical-information | 2304.14044 | null | https://arxiv.org/abs/2304.14044v1 | https://arxiv.org/pdf/2304.14044v1.pdf | Large Scale Genealogical Information Extraction From Handwritten Quebec Parish Records | This paper presents a complete workflow designed for extracting information from Quebec handwritten parish registers. The acts in these documents contain individual and family information highly valuable for genetic, demographic and social studies of the Quebec population. From an image of parish records, our workflow is able to identify the acts and extract personal information. The workflow is divided into successive steps: page classification, text line detection, handwritten text recognition, named entity recognition and act detection and classification. For all these steps, different machine learning models are compared. Once the information is extracted, validation rules designed by experts are then applied to standardize the extracted information and ensure its consistency with the type of act (birth, marriage, and death). This validation step is able to reject records that are considered invalid or merged. The full workflow has been used to process over two million pages of Quebec parish registers from the 19-20th centuries. On a sample comprising 65% of registers, 3.2 million acts were recognized. Verification of the birth and death acts from this sample shows that 74% of them are considered complete and valid. These records will be integrated into the BALSAC database and linked together to recreate family and genealogical relations at large scale. | ['Christopher Kermorvant', 'Hélène Vézina', 'Eugénie Capel', 'James McGrath', 'Mélodie Boillet', 'Martin Maarand', 'Solène Tarride'] | 2023-04-27 | null | null | null | null | ['line-detection'] | ['computer-vision'] | [ 2.00557083e-01 3.01623017e-01 -1.34784788e-01 -4.11825180e-01
-6.70572400e-01 -5.75323641e-01 9.07209694e-01 8.04782808e-01
-7.20897853e-01 1.14582062e+00 5.15031636e-01 1.35165825e-01
-1.56194568e-01 -1.06951165e+00 -2.39030793e-01 -3.11165422e-01
1.37654409e-01 1.17523098e+00 9.90790278e-02 -1.35729715e-01
4.32996154e-01 6.25824809e-01 -1.38965273e+00 3.61734450e-01
8.97084415e-01 4.81010169e-01 -3.95503700e-01 8.73448431e-01
-1.67373210e-01 4.75399643e-01 -7.79169083e-01 -1.11130214e+00
8.62061046e-03 -3.99172455e-01 -1.00754869e+00 -1.34971082e-01
4.74104911e-01 -4.94992226e-01 -1.19510397e-01 7.24912524e-01
4.45567280e-01 -3.57384324e-01 8.61064613e-01 -6.30471587e-01
-1.44424647e-01 9.38237369e-01 -8.89767185e-02 -1.62853479e-01
7.77405918e-01 -2.72093713e-01 6.62594318e-01 -6.38846159e-01
1.10093808e+00 9.83001351e-01 9.25967872e-01 3.85422975e-01
-1.10341585e+00 -5.06723166e-01 -6.17959499e-01 5.44154570e-02
-1.57252991e+00 -7.17430770e-01 2.80317456e-01 -7.24240363e-01
6.53050125e-01 3.98246974e-01 1.17633533e+00 7.60518849e-01
-8.98750350e-02 5.15835941e-01 1.07313120e+00 -6.34427905e-01
4.54273298e-02 2.94054925e-01 3.66809338e-01 8.22490931e-01
5.44998944e-01 -1.85809135e-01 -5.81369460e-01 -5.03164113e-01
5.66664219e-01 -2.52199173e-01 7.96949789e-02 3.33069265e-01
-9.16335404e-01 4.13850695e-01 -1.33241281e-01 4.81494039e-01
-4.97324705e-01 -2.95616716e-01 5.16919017e-01 -2.04504672e-02
3.71332288e-01 1.47685900e-01 -2.29393095e-01 -6.17928743e-01
-1.47877395e+00 3.40502679e-01 8.89194846e-01 9.08273757e-01
8.05757523e-01 -4.16926503e-01 1.21336207e-02 1.00510907e+00
4.00805503e-01 8.11124742e-01 1.68761611e-01 -5.67208171e-01
4.78810847e-01 1.18254006e+00 1.06554523e-01 -1.04907238e+00
-6.27158940e-01 6.82214350e-02 -7.15607285e-01 -6.62528649e-02
6.96061075e-01 -1.00928567e-01 -9.35978591e-01 8.05871546e-01
4.37235266e-01 -4.15789634e-01 -1.86799124e-01 2.70781875e-01
9.36373770e-01 2.05364123e-01 -7.70619214e-02 -2.28121221e-01
1.61508775e+00 9.43909883e-02 -7.87511945e-01 2.16191351e-01
3.63801956e-01 -7.69613504e-01 2.13213414e-01 1.62466079e-01
-1.24871635e+00 -2.30467230e-01 -9.91227388e-01 6.73624128e-02
-5.87532878e-01 4.79467958e-01 3.06416869e-01 9.38178003e-01
-6.22463524e-01 7.39294827e-01 -8.86478961e-01 -6.69453859e-01
7.26556599e-01 5.12755632e-01 -5.18968880e-01 3.67163092e-01
-1.13288236e+00 7.15678453e-01 6.04714096e-01 1.17582597e-01
-3.49429220e-01 -4.62981999e-01 -9.68241215e-01 -2.57217586e-01
-1.37126356e-01 -1.23608314e-01 7.29064703e-01 -5.38836062e-01
-1.18165255e+00 1.51361752e+00 -3.84041443e-02 -4.22921121e-01
9.97298598e-01 -1.83906816e-02 -6.97715521e-01 2.06810176e-01
3.52023154e-01 7.66173676e-02 5.01570463e-01 -9.06990826e-01
-8.96697044e-01 -7.09507406e-01 -7.48952568e-01 -4.04859006e-01
-2.53087103e-01 3.56518298e-01 -9.02005136e-01 -7.40833163e-01
1.96184024e-01 -8.39289904e-01 4.23105061e-01 -4.82100070e-01
-3.90559405e-01 2.15178262e-02 6.59545735e-02 -1.57006741e+00
1.71088850e+00 -1.78037369e+00 -2.77282000e-01 6.73093259e-01
1.70198858e-01 3.33712608e-01 3.76560599e-01 8.77848446e-01
2.72444487e-01 5.76856397e-02 -4.99406129e-01 -1.74636170e-01
2.26132106e-03 1.71743631e-02 1.17019750e-01 8.08196664e-01
-1.78811103e-02 8.01443815e-01 -8.42230976e-01 -8.31830442e-01
1.22699149e-01 3.80467474e-01 -1.78817749e-01 -1.71321511e-01
1.13233708e-01 2.38086507e-01 -4.24783826e-01 9.86680567e-01
7.71162808e-01 2.93669343e-01 6.28548682e-01 7.75891319e-02
-1.61034212e-01 3.17105323e-01 -1.04420304e+00 1.17576158e+00
9.30510759e-02 7.18270898e-01 1.22562639e-01 -4.18343961e-01
1.33848703e+00 1.50493100e-01 4.72929597e-01 -5.12207508e-01
1.76689744e-01 4.94695842e-01 -3.19782376e-01 -5.64001620e-01
8.55808914e-01 2.08135158e-01 -1.46679267e-01 4.68539357e-01
-1.46092206e-01 2.43893802e-01 7.53557622e-01 1.59247652e-01
9.93769348e-01 2.79558122e-01 4.13106829e-01 -9.50274095e-02
9.41497087e-01 3.11824441e-01 6.25276685e-01 5.33377409e-01
-1.68846712e-01 5.07603288e-01 5.40597916e-01 -6.20443165e-01
-1.32499957e+00 -9.61454332e-01 -5.06269753e-01 5.73163688e-01
-4.34490591e-01 -8.44603539e-01 -8.69246542e-01 -6.90589249e-01
2.14999192e-03 3.78275275e-01 -9.00964439e-01 2.86809206e-01
-7.91808724e-01 -7.03311265e-01 1.16376019e+00 6.70348033e-02
9.66886163e-01 -9.14842010e-01 -6.24535859e-01 3.04223448e-01
-1.56891897e-01 -6.30199432e-01 9.98438820e-02 -4.62700993e-01
-5.63627481e-01 -1.45186853e+00 -9.94627476e-01 -4.78342921e-01
8.53726804e-01 -9.72740054e-01 9.26524043e-01 2.69867629e-01
-4.89473373e-01 2.99194813e-01 -3.28759581e-01 -2.59403259e-01
-1.10639989e+00 2.74978876e-01 -1.23015970e-01 1.47331446e-01
4.96585220e-01 -2.78899640e-01 -1.64663792e-01 1.50576815e-01
-6.13235712e-01 -2.67057210e-01 3.42543721e-01 5.42003989e-01
3.14534307e-02 -1.74514830e-01 2.62242675e-01 -1.16891003e+00
4.54064488e-01 -1.76765785e-01 -7.00600207e-01 6.69969320e-01
-5.51404774e-01 -1.21364094e-01 3.31496835e-01 1.45314604e-01
-1.07319903e+00 2.84876555e-01 -3.68917346e-01 5.82042933e-01
-4.82823789e-01 5.19074798e-01 -6.53730193e-03 5.63869953e-01
4.19340938e-01 3.88852805e-01 3.04394197e-02 -6.32251680e-01
-7.12889582e-02 1.31688654e+00 7.49381602e-01 -2.81258404e-01
5.37026227e-01 4.91072953e-01 -1.52408510e-01 -1.07675743e+00
2.09984872e-02 -4.50395674e-01 -1.43206024e+00 -5.71813405e-01
9.58616376e-01 -5.09575844e-01 -8.54542255e-01 1.00429714e+00
-9.25116181e-01 1.42421550e-03 7.59257749e-02 4.47685033e-01
-2.83279836e-01 3.54013503e-01 -6.55589461e-01 -1.23475122e+00
-3.54685426e-01 -4.57264841e-01 9.78228509e-01 1.70042947e-01
-6.83331251e-01 -8.45359981e-01 4.10295337e-01 5.51297784e-01
-2.00494662e-01 3.03990871e-01 7.72733450e-01 -8.77464354e-01
-7.76664168e-02 -6.74668968e-01 -8.50922521e-03 -2.25693017e-01
2.01145694e-01 4.59817797e-01 -7.08290458e-01 6.30922541e-02
-5.73273778e-01 1.78746521e-01 8.20911944e-01 -1.05387978e-01
-2.03780040e-01 -4.73435879e-01 -4.67799813e-01 2.99608260e-01
1.11113691e+00 2.84143299e-01 1.13745236e+00 4.79846090e-01
2.87737548e-01 7.36850500e-01 4.00283903e-01 5.70076704e-01
3.73640954e-01 6.82454526e-01 -1.78320482e-01 1.85065493e-01
5.13331033e-02 -2.96708167e-01 4.96246397e-01 4.88449991e-01
-6.46530330e-01 2.76972502e-01 -1.49229002e+00 5.66738725e-01
-1.66289878e+00 -1.24016702e+00 -3.99001300e-01 2.43558264e+00
8.24661791e-01 6.99030310e-02 6.13367617e-01 4.72964257e-01
1.01131976e+00 -2.09448427e-01 4.91807200e-02 -3.17303509e-01
-2.22708896e-01 1.91458225e-01 8.46747220e-01 4.88348275e-01
-1.16075230e+00 8.47509384e-01 6.57084179e+00 5.12124062e-01
-7.28975773e-01 -3.64512831e-01 4.88018811e-01 4.55160551e-02
6.07686713e-02 -1.95998609e-01 -1.12715554e+00 5.46570420e-01
8.53340626e-01 -2.68653668e-02 3.41267765e-01 1.06214985e-01
1.55462965e-01 -4.44975197e-01 -8.26173306e-01 9.26874459e-01
2.96717703e-01 -1.62412643e+00 -1.81004032e-01 4.64569151e-01
4.58034337e-01 -1.21924132e-01 -6.73504710e-01 2.35884562e-02
3.94275337e-01 -8.18387866e-01 9.67458129e-01 1.07684934e+00
9.83588755e-01 -7.20905066e-01 7.95527339e-01 7.52154663e-02
-9.20464933e-01 -3.90183665e-02 -1.26259327e-01 1.84677392e-01
3.78952205e-01 6.19990051e-01 -1.05333531e+00 8.40938985e-01
5.17697155e-01 7.91781008e-01 -9.57416058e-01 8.91704381e-01
-2.52448499e-01 5.20077169e-01 -4.19884652e-01 -1.41734391e-01
-1.33970693e-01 -4.72746044e-01 4.08890873e-01 1.25807619e+00
2.19429925e-01 -5.09361215e-02 -4.23636734e-01 6.15105093e-01
-1.19957440e-01 3.73380691e-01 -2.13829920e-01 -2.06135020e-01
3.91768217e-01 1.01037955e+00 -1.05470932e+00 -6.17585242e-01
-1.04249649e-01 1.02655268e+00 3.05185616e-01 -1.04976073e-01
-5.09505689e-01 -3.78709406e-01 2.99438149e-01 5.07399380e-01
1.86734825e-01 -1.36271223e-01 -1.82477713e-01 -9.15149868e-01
4.37736511e-03 -7.72871315e-01 8.52743745e-01 -3.15463960e-01
-7.85850763e-01 4.29019660e-01 -1.44712999e-01 -7.63187945e-01
-2.57924139e-01 -3.87470871e-01 -2.54938543e-01 6.85089111e-01
-4.76312757e-01 -1.31560016e+00 -2.05779634e-02 2.55012125e-01
-1.25551373e-02 -4.01842356e-01 9.90575373e-01 5.45547843e-01
-6.69299245e-01 4.55285996e-01 3.97903144e-01 9.97676075e-01
5.59706688e-01 -1.15245831e+00 2.55886257e-01 5.57726622e-01
8.54468420e-02 6.20674074e-01 2.70715773e-01 -1.59621942e+00
-9.38200712e-01 -7.84127355e-01 1.59881628e+00 -4.96596038e-01
5.35443842e-01 -6.56712234e-01 -6.24116659e-01 5.91056883e-01
-1.18082032e-01 -6.50339842e-01 7.57359505e-01 2.79873777e-02
1.75230894e-02 -2.48220582e-02 -1.25892019e+00 3.70253652e-01
7.27311611e-01 -4.31094080e-01 -6.38722062e-01 6.87305927e-02
-6.10834301e-01 -1.24951072e-01 -1.27151561e+00 3.31567600e-02
1.16136181e+00 -9.55173671e-01 5.24723828e-01 -4.54429299e-01
3.62687826e-01 -2.00480223e-01 2.69700289e-01 -5.70433974e-01
5.56642897e-02 -7.76874363e-01 4.21594024e-01 1.83283663e+00
7.15065539e-01 -6.79175556e-01 8.66855979e-01 8.72874260e-01
2.92878896e-01 -9.02293846e-02 -1.18002594e+00 -5.24755716e-01
-8.32393020e-02 -1.82708576e-01 8.74994934e-01 9.14969444e-01
2.76110083e-01 1.52218476e-01 -4.56578732e-01 -1.79718196e-01
3.98675501e-01 -2.89848477e-01 9.39630687e-01 -1.41804326e+00
2.83171624e-01 -3.94710183e-01 -7.39378095e-01 -1.88951865e-01
-2.52537519e-01 -1.06765687e+00 -3.77436221e-01 -1.78561246e+00
1.54602006e-01 -5.41954577e-01 4.26450640e-01 4.77697551e-01
1.71703756e-01 3.08125973e-01 1.21024035e-01 6.05245113e-01
-1.68676540e-01 -1.21949213e-02 5.18116176e-01 -1.21877529e-01
-3.23935539e-01 7.57218823e-02 1.72135904e-02 7.55081415e-01
7.63985634e-01 -4.62045699e-01 5.77834249e-01 1.02696218e-01
4.23747659e-01 -1.32711828e-01 1.50671199e-01 -1.18790209e+00
1.98329747e-01 5.41262366e-02 7.72935927e-01 -1.12963414e+00
1.54828638e-01 -8.26547205e-01 5.96144676e-01 8.29638720e-01
-1.45862386e-01 -1.78737834e-01 -9.36184358e-03 2.24674419e-01
7.98023194e-02 -4.14100319e-01 6.02531195e-01 -2.25284502e-01
-1.64043158e-01 -1.36197865e-01 -7.86191165e-01 -2.89622605e-01
9.82371211e-01 -4.42923486e-01 -2.45530888e-01 -2.30228662e-01
-1.03733158e+00 6.32285327e-02 7.29016721e-01 -3.65835205e-02
3.38512510e-01 -1.12859070e+00 -1.28261864e+00 5.10433614e-01
1.83588803e-01 -5.96662939e-01 6.64105192e-02 8.01600873e-01
-1.29351544e+00 2.61292517e-01 -2.98806459e-01 -3.26367289e-01
-1.71382499e+00 2.90768713e-01 9.40157995e-02 -3.58354390e-01
-6.35495305e-01 7.94638470e-02 -9.45984244e-01 -3.20038408e-01
8.41666088e-02 7.94062242e-02 -6.58855796e-01 7.62544036e-01
7.69019544e-01 8.07934523e-01 1.32008508e-01 -1.21898472e+00
-6.33045197e-01 6.15201414e-01 2.94271484e-03 -4.95144159e-01
1.61150813e+00 -3.47770490e-02 -3.71347249e-01 3.69632661e-01
7.82626271e-01 6.38213515e-01 -5.92962801e-01 2.78008014e-01
4.68567103e-01 -4.09492582e-01 -6.40440822e-01 -9.42738950e-01
-7.46384799e-01 2.75924146e-01 3.37502986e-01 3.91438425e-01
7.69825518e-01 6.12879060e-02 4.38916922e-01 5.02882227e-02
1.10482238e-01 -1.49074745e+00 -9.98539686e-01 6.43115163e-01
8.82963538e-01 -7.59975135e-01 4.71395254e-01 -3.34938437e-01
-2.52307415e-01 1.41939616e+00 -2.95529794e-02 2.93907493e-01
5.08173168e-01 2.28932977e-01 9.02136490e-02 -3.00946325e-01
-1.62346482e-01 -2.40102470e-01 2.61282325e-01 7.25375652e-01
4.20665950e-01 7.34572411e-02 -1.09760964e+00 4.71174747e-01
-6.33597612e-01 2.05808893e-01 5.29317975e-01 1.03353453e+00
-2.48358011e-01 -1.36448920e+00 -8.55964243e-01 7.75905073e-01
-6.12394929e-01 -1.13218114e-01 -1.15416849e+00 9.37584579e-01
2.59187907e-01 6.86970234e-01 4.01176065e-01 -1.59483626e-01
1.56899184e-01 2.98034906e-01 4.87949848e-01 -2.89709628e-01
-1.21278346e+00 8.78664181e-02 8.85309637e-01 4.89303134e-02
-5.73575437e-01 -1.03197753e+00 -1.35252309e+00 -5.60993314e-01
1.41980657e-02 2.86278844e-01 7.31685877e-01 7.50692725e-01
9.71093774e-02 -9.78325531e-02 1.70276701e-01 -3.76623333e-01
8.56951550e-02 -9.88214791e-01 -7.54803002e-01 5.67726851e-01
-1.03377037e-01 -1.83078542e-01 4.54520248e-02 4.67880785e-01] | [10.122241020202637, 10.282012939453125] |
f6f984c0-fae2-4a8a-9038-02125ec4b838 | compressing-deep-neural-networks-on-fpgas-to | 2003.06308 | null | https://arxiv.org/abs/2003.06308v2 | https://arxiv.org/pdf/2003.06308v2.pdf | Compressing deep neural networks on FPGAs to binary and ternary precision with HLS4ML | We present the implementation of binary and ternary neural networks in the hls4ml library, designed to automatically convert deep neural network models to digital circuits with FPGA firmware. Starting from benchmark models trained with floating point precision, we investigate different strategies to reduce the network's resource consumption by reducing the numerical precision of the network parameters to binary or ternary. We discuss the trade-off between model accuracy and resource consumption. In addition, we show how to balance between latency and accuracy by retaining full precision on a selected subset of network components. As an example, we consider two multiclass classification tasks: handwritten digit recognition with the MNIST data set and jet identification with simulated proton-proton collisions at the CERN Large Hadron Collider. The binary and ternary implementation has similar performance to the higher precision implementation while using drastically fewer FPGA resources. | ['Nhan Tran', 'Sheila Sagear', 'Sergo Jindariani', 'Philip Harris', 'Kevin Pedro', 'Jennifer Ngadiuba', 'Giuseppe Di Guglielmo', 'Edward Kreinar', 'Zhenbin Wu', 'Vladimir Loncar', 'Mia Liu', 'Maurizio Pierini', 'Javier Duarte', 'Sioni Summers', 'Dylan Rankin', 'Duc Hoang'] | 2020-03-11 | null | null | null | null | ['handwritten-digit-recognition'] | ['computer-vision'] | [ 2.15677932e-01 1.54434457e-01 -2.40663409e-01 -7.33629465e-01
-9.11110118e-02 -4.12818015e-01 4.24126208e-01 4.14618589e-02
-8.64090979e-01 7.88157761e-01 -6.99524581e-01 -6.19881690e-01
-2.80095458e-01 -9.70723748e-01 -8.75306785e-01 -4.98051643e-01
2.82452822e-01 7.13258922e-01 2.89164960e-01 1.25521556e-01
-1.76352993e-01 1.11683249e+00 -1.47059906e+00 7.37241924e-01
-2.64637321e-02 1.38565218e+00 -2.10919186e-01 8.88508499e-01
2.54411727e-01 9.76101518e-01 -7.73098767e-01 -4.51723844e-01
6.16126001e-01 -2.73740739e-02 -5.53280890e-01 -6.95735693e-01
3.58460784e-01 -1.31956622e-01 -5.63966215e-01 1.21603847e+00
3.75379890e-01 -8.50307941e-02 5.68400502e-01 -1.45040047e+00
1.11468449e-01 9.98476923e-01 2.27761284e-01 2.74888426e-01
-4.40158427e-01 -4.90759639e-03 5.90725720e-01 -4.16126102e-01
3.80319625e-01 1.32207716e+00 1.07080555e+00 3.75254720e-01
-1.09749019e+00 -9.01998699e-01 -4.23022330e-01 2.19820678e-01
-1.46488202e+00 -5.03536701e-01 2.44878516e-01 -3.22234601e-01
1.72833157e+00 2.67682076e-01 9.30680037e-01 9.26572204e-01
5.71867526e-01 1.59395874e-01 5.14067769e-01 -6.27101481e-01
4.54410732e-01 3.18150520e-01 7.19305158e-01 7.60626495e-01
8.58699501e-01 5.09432912e-01 -3.14808577e-01 -1.39020890e-01
9.37913954e-01 -3.24757874e-01 6.75925100e-03 -7.91930556e-02
-8.75846744e-01 6.64425731e-01 3.99115831e-01 3.07089269e-01
-2.61680186e-01 7.23624229e-01 8.10952246e-01 4.47018355e-01
-2.87702471e-01 5.11206329e-01 -9.18429554e-01 8.40491131e-02
-8.14717889e-01 2.30721369e-01 1.04520476e+00 1.17059791e+00
4.78910387e-01 4.17746335e-01 2.38397479e-01 5.48356771e-01
4.15697843e-01 1.37736842e-01 6.29877865e-01 -7.64299273e-01
2.50595421e-01 3.55112642e-01 -3.57279480e-01 -5.36847651e-01
-6.93034530e-01 -4.98276114e-01 -8.60160947e-01 4.82204795e-01
4.79822487e-01 -4.40454453e-01 -1.04257655e+00 1.08635867e+00
-3.81293260e-02 -2.28289396e-01 3.90733659e-01 6.27136886e-01
7.85555303e-01 6.89000547e-01 1.84653431e-01 1.66336179e-01
1.33218431e+00 -9.47336912e-01 -3.25115263e-01 -3.63266289e-01
8.58626962e-01 -4.30786937e-01 1.89284116e-01 6.66462421e-01
-1.01781249e+00 -7.25631416e-01 -1.82727873e+00 -2.40474835e-01
-6.61086142e-01 5.95323980e-01 8.00534248e-01 9.50270891e-01
-9.67569709e-01 1.17038417e+00 -1.21223772e+00 -2.73866262e-02
1.81653008e-01 1.07146215e+00 -6.56508505e-02 2.64448643e-01
-1.26774359e+00 1.03490853e+00 1.05426443e+00 1.55250385e-01
-3.85624915e-01 -8.81850183e-01 -7.78734803e-01 3.20405364e-01
-4.26882923e-01 -5.22825241e-01 1.67624533e+00 -1.22315502e+00
-1.57739973e+00 4.42885995e-01 4.86894190e-01 -1.16844022e+00
1.96265921e-01 3.62145364e-01 -6.22825146e-01 -6.54024035e-02
-6.72314167e-01 1.15665245e+00 4.65624362e-01 -2.89326429e-01
-6.41493261e-01 -2.66341325e-02 -1.22857265e-01 -1.95197359e-01
-2.62296796e-01 -3.27836461e-02 -6.54194802e-02 -2.67075300e-01
-2.56439596e-02 -9.18509424e-01 -2.39511669e-01 8.79799277e-02
-1.65862456e-01 8.04971606e-02 7.12469220e-01 -3.37778658e-01
6.57130063e-01 -2.08094025e+00 -2.62053937e-01 4.49196458e-01
-1.40851766e-01 4.56548721e-01 3.48849297e-02 -1.97615325e-01
-3.50101590e-01 -2.04095542e-01 2.61903018e-01 2.49126136e-01
4.32173423e-02 3.55953664e-01 -8.52438733e-02 2.80092925e-01
1.69731632e-01 6.90494955e-01 -3.79179418e-01 -8.78415778e-02
1.01612940e-01 7.07223177e-01 -5.90418398e-01 -6.88716710e-01
-2.42325857e-01 -4.57314700e-01 -2.15451807e-01 5.23221970e-01
4.82317924e-01 -6.87433928e-02 5.63068330e-01 -6.43519223e-01
1.76938958e-02 5.39289236e-01 -1.03862667e+00 1.05575812e+00
-4.24433976e-01 1.01253092e+00 1.28371716e-01 -8.03108573e-01
9.79686081e-01 3.66118699e-01 2.53339261e-02 -7.48638988e-01
7.39840388e-01 4.38854069e-01 5.85694432e-01 1.86598331e-01
5.20909429e-01 -2.77285069e-01 3.73171293e-03 -4.59247641e-02
4.66542602e-01 -6.91363961e-02 2.60983914e-01 -4.32930321e-01
1.00519335e+00 -1.84421748e-01 3.23643178e-01 -4.24145967e-01
2.42869571e-01 3.04639786e-01 4.25029248e-01 6.54977322e-01
5.93105480e-02 1.82424501e-01 6.92368031e-01 -6.28420234e-01
-1.20034587e+00 -6.44584775e-01 -4.69913721e-01 6.13475204e-01
-5.42150140e-01 -3.58403683e-01 -8.71342480e-01 -2.68223345e-01
2.09916253e-02 7.45821714e-01 -1.71301484e-01 -3.16388160e-01
-5.58708668e-01 -1.06403089e+00 1.11513257e+00 9.09771204e-01
5.43861568e-01 -1.02920270e+00 -1.02475893e+00 2.36316279e-01
9.81733441e-01 -1.22412348e+00 2.91491866e-01 1.14790320e+00
-9.73271608e-01 -9.68328834e-01 -1.50931492e-01 -8.71394813e-01
4.42925036e-01 -5.89017868e-01 8.30615222e-01 7.08392262e-02
-5.14631808e-01 -1.70707360e-01 1.44879907e-01 -5.93511224e-01
-6.30463898e-01 3.03076804e-01 2.35213965e-01 -6.81818366e-01
4.18813556e-01 -2.18697116e-01 2.40695104e-02 -4.78180274e-02
-7.98516333e-01 1.49596825e-01 6.76449060e-01 8.47674191e-01
4.68873233e-01 6.10410690e-01 2.72450536e-01 -6.26173556e-01
3.37957889e-01 3.59029025e-02 -1.14911473e+00 1.11848563e-01
-5.23190379e-01 3.03866655e-01 8.73324215e-01 -4.32123959e-01
-6.27550006e-01 8.29157233e-01 -3.76234651e-01 -3.70821059e-01
-1.18727991e-02 4.35636550e-01 -2.69050181e-01 -6.15164816e-01
7.78943956e-01 -3.55193198e-01 -1.13157377e-01 -3.27586323e-01
-9.32193473e-02 5.37532866e-01 6.61878049e-01 -3.63375306e-01
9.51422378e-02 -1.41684785e-01 3.13238204e-01 -7.81571150e-01
-2.66403615e-01 1.76360473e-01 -7.62533009e-01 9.89593938e-02
5.42193353e-01 -8.36585343e-01 -8.25186312e-01 6.40592158e-01
-1.44341445e+00 -5.41503906e-01 -3.17385882e-01 7.45937586e-01
-2.99786717e-01 -2.41300896e-01 -1.00202906e+00 -3.65036309e-01
-4.59604949e-01 -1.30212963e+00 6.49772823e-01 3.19224864e-01
-2.64170170e-01 -7.72527695e-01 -4.14363682e-01 -4.38575119e-01
1.36592820e-01 9.29306075e-03 1.29096878e+00 -8.86906326e-01
-3.78587425e-01 -5.03384650e-01 -4.07283068e-01 6.86896980e-01
-4.11829293e-01 1.53900817e-01 -1.07734966e+00 -1.07431211e-01
1.45519748e-02 -8.72934312e-02 7.61663198e-01 3.93171281e-01
1.08704507e+00 -2.85821795e-01 -4.30888444e-01 7.22705722e-01
1.36386144e+00 8.65587473e-01 6.42701864e-01 4.47541267e-01
4.42812115e-01 2.62423158e-01 3.45858932e-01 8.25312808e-02
-2.87502348e-01 5.73831618e-01 1.07093595e-01 1.71968594e-01
-1.40320584e-01 9.42920223e-02 2.43789494e-01 6.09495878e-01
2.39167273e-01 -1.67212799e-01 -1.22070539e+00 -6.79739518e-03
-1.55006731e+00 -5.23331523e-01 -2.64696062e-01 2.02032113e+00
6.78083777e-01 7.13990629e-01 -2.04524145e-01 6.79471433e-01
5.37961423e-01 -3.95591676e-01 -3.16929638e-01 -7.99984217e-01
1.50187165e-01 7.16023505e-01 1.20350826e+00 5.15244901e-01
-9.22105670e-01 7.73732603e-01 6.92299509e+00 7.24923372e-01
-1.45892704e+00 -6.02856325e-03 7.36736596e-01 -4.76885140e-01
5.30320108e-01 -3.46266001e-01 -1.33864868e+00 2.12932050e-01
1.93826723e+00 1.49542559e-02 2.66419053e-01 1.11865020e+00
-5.96303046e-01 7.50202835e-02 -1.54999316e+00 9.02995825e-01
-3.83553892e-01 -1.56502986e+00 -1.02024607e-01 -1.42337129e-01
1.30683467e-01 2.05814615e-01 -2.50144333e-01 6.55951619e-01
7.32340664e-02 -1.21377909e+00 1.02548385e+00 3.01608622e-01
9.15651023e-01 -1.22043681e+00 8.41957211e-01 1.32809520e-01
-8.73022199e-01 -3.02301258e-01 -6.20062947e-01 -3.12251508e-01
-2.68936604e-01 4.85252827e-01 -1.23339462e+00 1.26131251e-02
7.15961695e-01 3.04708958e-01 -5.73283255e-01 1.03026247e+00
2.72124019e-02 3.75233471e-01 -5.24023890e-01 -4.17427748e-01
2.65084028e-01 9.28243995e-02 -1.72440842e-01 1.29917336e+00
4.03975636e-01 8.37364420e-02 -4.95996684e-01 4.75537956e-01
-9.98491049e-02 -5.26720405e-01 -3.10261905e-01 -5.77927530e-02
6.05015993e-01 1.25578976e+00 -1.08335912e+00 -5.47241092e-01
-1.30925318e-02 4.87972498e-01 1.16391532e-01 -6.85828105e-02
-8.66288245e-01 -7.96764612e-01 4.52223837e-01 -5.84139302e-02
3.36326778e-01 -1.77274480e-01 -6.49811089e-01 -5.09085655e-01
-3.50271344e-01 -8.22275579e-01 -1.46294851e-02 -8.04209650e-01
-4.50571835e-01 7.91276813e-01 6.09114021e-02 -6.46885872e-01
-4.38471705e-01 -1.56263351e+00 -2.46229798e-01 6.42772377e-01
-9.52520013e-01 -6.15170658e-01 1.36433214e-01 -1.84105840e-02
1.35144129e-01 -4.55721825e-01 8.26905191e-01 6.71190739e-01
-7.47176826e-01 7.86638498e-01 1.61446899e-01 1.77135304e-01
1.35472268e-01 -8.06194961e-01 8.53227258e-01 4.54103708e-01
-2.03374419e-02 5.33000886e-01 6.46350622e-01 -4.73874748e-01
-1.44169009e+00 -1.45377874e+00 8.90821636e-01 4.29638177e-02
5.41956544e-01 -5.88004231e-01 -6.85430348e-01 9.61534858e-01
-1.76475137e-01 1.88103184e-01 1.68315873e-01 -2.07063004e-01
-1.66657284e-01 -2.58690417e-01 -1.48733461e+00 4.30349141e-01
5.77124536e-01 -2.88824022e-01 -1.72669858e-01 5.76788664e-01
5.31572580e-01 -6.57370269e-01 -9.19678032e-01 5.59292674e-01
8.26659441e-01 -6.78327560e-01 1.07216656e+00 -3.73133570e-01
1.68564841e-01 -4.82780784e-02 -2.52060980e-01 -9.96269107e-01
-1.48442626e-01 -9.61430147e-02 -8.02633762e-02 7.47274339e-01
5.43654561e-01 -7.15613782e-01 1.09287786e+00 6.31544590e-01
-1.95233405e-01 -7.57870853e-01 -1.22886193e+00 -1.00923300e+00
1.48075506e-01 -4.84849274e-01 7.14591682e-01 5.88759720e-01
-1.80776432e-01 3.52044016e-01 1.05798051e-01 1.76432505e-01
1.86599270e-01 -5.44516921e-01 3.28221321e-02 -1.43414700e+00
-5.60928226e-01 -4.90162253e-01 -8.40024173e-01 -6.03482902e-01
2.99702108e-01 -1.05986059e+00 2.75188778e-02 -8.62294078e-01
-3.08313131e-01 -6.05948031e-01 -2.71688074e-01 8.25504899e-01
9.95587528e-01 3.95151109e-01 9.33473855e-02 -2.15455875e-01
-6.53612167e-02 -9.30018499e-02 4.58153039e-01 -2.39546344e-01
2.67415829e-02 -4.60317358e-02 -1.80980608e-01 9.89891231e-01
9.51774716e-01 -7.40561366e-01 -2.77891308e-01 -4.63677406e-01
2.29407191e-01 2.41903007e-01 5.26904464e-01 -1.68594778e+00
5.40057302e-01 3.41272563e-01 9.34468865e-01 -5.59287369e-01
6.43681705e-01 -8.97066355e-01 4.29622233e-01 1.05891490e+00
-4.15812850e-01 2.74274141e-01 8.25101435e-01 4.01097089e-02
-2.39811279e-02 -9.36863005e-01 1.12441492e+00 1.07514851e-01
-5.57362378e-01 -5.62168397e-02 -7.73124218e-01 -7.86317468e-01
8.45679939e-01 -4.59464528e-02 -5.38929403e-01 3.57012540e-01
-8.06154609e-01 -2.15920389e-01 1.09173439e-01 3.08789074e-01
3.44593853e-01 -1.30951715e+00 1.06742943e-03 6.16069973e-01
-4.85398561e-01 -2.90002227e-01 -7.03305230e-02 3.77036870e-01
-1.22901785e+00 1.14492321e+00 -6.73139572e-01 -3.28938574e-01
-1.46746111e+00 3.11376512e-01 9.48299468e-01 -7.79236704e-02
-3.61027509e-01 8.81469369e-01 -4.41703022e-01 -5.91328502e-01
5.33601522e-01 -1.02903426e+00 1.19973823e-01 -1.32456854e-01
4.51048821e-01 4.12234038e-01 8.98518801e-01 -1.34659395e-01
-4.47076321e-01 5.63017465e-02 -1.85443744e-01 -1.30354360e-01
1.06707764e+00 8.75595987e-01 -5.78878298e-02 1.54788077e-01
1.25971067e+00 -8.60218883e-01 -8.27706397e-01 3.02709848e-01
1.88283131e-01 4.15482908e-01 4.51596618e-01 -9.48677957e-01
-1.25969827e+00 8.25430930e-01 8.15140069e-01 -8.71998444e-02
8.96106243e-01 -6.10831201e-01 5.32836676e-01 1.37643635e+00
3.54996830e-01 -1.05443847e+00 -6.62974536e-01 9.81141388e-01
6.12287343e-01 -5.29736638e-01 2.46718615e-01 -3.85271132e-01
-8.43851343e-02 1.73500800e+00 7.89910555e-01 -3.32485825e-01
7.21049607e-01 1.08006012e+00 -2.27888957e-01 -9.18907151e-02
-8.68229032e-01 7.21941769e-01 1.06495939e-01 3.86855453e-01
3.17331403e-01 -5.07593080e-02 -1.34690732e-01 9.22929049e-01
-7.46710479e-01 2.67499149e-01 6.49941385e-01 1.04001212e+00
-4.20358956e-01 -1.08911479e+00 -3.57993335e-01 7.55466938e-01
-3.74812484e-01 -2.35429198e-01 -3.92585665e-01 1.09811568e+00
2.53922582e-01 3.14796805e-01 4.71975833e-01 -5.31895757e-01
2.45423377e-01 2.90478349e-01 9.58338499e-01 -4.72497731e-01
-1.14774513e+00 -3.35904777e-01 4.00488138e-01 -1.38509825e-01
1.23440914e-01 -4.66801107e-01 -1.54667354e+00 -2.59426892e-01
-1.42064616e-01 -8.62113573e-03 1.23821521e+00 7.72213221e-01
2.57236332e-01 9.66618776e-01 -2.40205646e-01 -1.15316832e+00
-6.58410668e-01 -7.32859433e-01 -5.77401996e-01 -8.77766609e-01
1.57749444e-01 -4.48478609e-01 -7.83244744e-02 -2.49205828e-01] | [8.355791091918945, 2.9334046840667725] |
cfbf28b6-a663-4255-9de5-853329d83c95 | ridiculously-fast-shot-boundary-detection | 1705.08214 | null | http://arxiv.org/abs/1705.08214v1 | http://arxiv.org/pdf/1705.08214v1.pdf | Ridiculously Fast Shot Boundary Detection with Fully Convolutional Neural Networks | Shot boundary detection (SBD) is an important component of many video
analysis tasks, such as action recognition, video indexing, summarization and
editing. Previous work typically used a combination of low-level features like
color histograms, in conjunction with simple models such as SVMs. Instead, we
propose to learn shot detection end-to-end, from pixels to final shot
boundaries. For training such a model, we rely on our insight that all shot
boundaries are generated. Thus, we create a dataset with one million frames and
automatically generated transitions such as cuts, dissolves and fades. In order
to efficiently analyze hours of videos, we propose a Convolutional Neural
Network (CNN) which is fully convolutional in time, thus allowing to use a
large temporal context without the need to repeatedly processing frames. With
this architecture our method obtains state-of-the-art results while running at
an unprecedented speed of more than 120x real-time. | ['Michael Gygli'] | 2017-05-23 | null | null | null | null | ['camera-shot-boundary-detection'] | ['computer-vision'] | [ 3.47511053e-01 -2.82218367e-01 -1.39994472e-01 -4.51797694e-02
-4.52423662e-01 -4.79420960e-01 7.19215870e-01 3.86472374e-01
-5.69335461e-01 4.02809620e-01 1.66219547e-01 -1.31619141e-01
3.79478365e-01 -7.07750976e-01 -8.24952483e-01 -2.97921687e-01
-3.78037900e-01 1.56152859e-01 6.95302248e-01 7.97472671e-02
2.84923524e-01 4.02575165e-01 -1.75341475e+00 4.32936370e-01
4.25182551e-01 1.13068295e+00 -1.22844152e-01 1.05136192e+00
-1.26472741e-01 1.13129568e+00 -4.36057895e-01 -1.44497216e-01
3.67549360e-01 -4.51217800e-01 -6.48072004e-01 3.43459696e-01
5.55013716e-01 -7.46430814e-01 -7.11012363e-01 8.64588618e-01
2.35197365e-01 4.12510514e-01 5.33026576e-01 -1.11263907e+00
-1.48589835e-01 3.90839815e-01 -6.72273099e-01 4.97571111e-01
4.35777277e-01 4.40057129e-01 8.06055963e-01 -7.58700073e-01
9.46330011e-01 9.41002607e-01 6.22063160e-01 5.02948105e-01
-1.13728416e+00 -2.30086863e-01 1.64419740e-01 6.11367047e-01
-1.17042947e+00 -6.66398823e-01 5.97460151e-01 -5.71911395e-01
1.05659771e+00 1.32138044e-01 1.01769042e+00 9.69975829e-01
1.76560789e-01 1.05503714e+00 5.33273995e-01 -4.66370821e-01
5.57957053e-01 -5.15284777e-01 -8.75544399e-02 8.90568793e-01
-1.06782071e-01 -3.57308596e-01 -6.17128968e-01 1.20863102e-01
7.11807072e-01 2.42010981e-01 -1.86215132e-01 -3.41682792e-01
-1.23969626e+00 6.08018994e-01 1.52291715e-01 4.34316434e-02
-4.21935558e-01 4.83225703e-01 7.85208523e-01 2.26088107e-01
5.05840659e-01 1.41314894e-01 -1.27488270e-01 -5.40785432e-01
-1.39797008e+00 3.44302595e-01 5.89218080e-01 5.92646539e-01
7.47725964e-01 -2.80547813e-02 -4.57698375e-01 5.15945375e-01
-1.69052124e-01 1.21454000e-01 3.72571677e-01 -1.39598227e+00
1.38998210e-01 5.05790651e-01 1.09100968e-01 -8.72327805e-01
-3.27132493e-01 1.29392281e-01 -7.27256536e-01 5.15551209e-01
4.28528875e-01 -1.02144942e-01 -1.28042758e+00 1.57128358e+00
4.67408687e-01 6.24562800e-01 -3.07954520e-01 8.46711159e-01
2.82436669e-01 6.91560268e-01 -1.15697995e-01 -2.94859320e-01
1.31506157e+00 -1.12597561e+00 -6.37268782e-01 -1.26180753e-01
5.51998794e-01 -4.45852667e-01 9.93471622e-01 6.33830726e-01
-1.23927617e+00 -4.55424994e-01 -1.20050132e+00 -2.47998253e-01
-3.35918099e-01 -1.09654032e-01 5.42907536e-01 2.38983750e-01
-1.03257537e+00 1.04632068e+00 -1.29453659e+00 -5.86340725e-01
6.88047886e-01 7.72490427e-02 -4.02149737e-01 -2.33570654e-02
-8.06846857e-01 6.94358587e-01 4.10086542e-01 -1.95266917e-01
-1.06586254e+00 -6.34482086e-01 -1.00655818e+00 2.34230056e-01
6.81760132e-01 -6.14573896e-01 1.27108300e+00 -1.01465273e+00
-1.51011419e+00 7.70556808e-01 -1.43713742e-01 -9.50743496e-01
8.01050782e-01 -4.57027078e-01 -2.62247119e-02 7.03066111e-01
-1.44523263e-01 7.97631681e-01 9.56576526e-01 -5.43121874e-01
-7.26295292e-01 -1.93634331e-01 1.97459936e-01 -1.58983860e-02
-3.71096551e-01 2.71302223e-01 -8.55468571e-01 -5.94570220e-01
-2.28499711e-01 -8.35664928e-01 -2.61136800e-01 2.31834009e-01
-3.13887119e-01 -1.25242621e-01 8.85486424e-01 -7.81960845e-01
1.32675970e+00 -2.09637904e+00 2.29105577e-01 -9.12606716e-02
3.75676334e-01 3.64223361e-01 1.10914029e-01 3.21219593e-01
1.24031246e-01 -9.80247706e-02 -3.09639364e-01 -5.57087123e-01
-2.36651078e-02 -1.14351511e-01 -2.01396644e-01 5.12814164e-01
3.88328224e-01 8.23536575e-01 -8.85195911e-01 -5.65195501e-01
5.84915519e-01 3.91703963e-01 -6.99306130e-01 1.74973339e-01
-4.48501199e-01 9.07028392e-02 6.14627451e-02 4.75956321e-01
2.89173663e-01 -2.22068742e-01 1.08821407e-01 2.56461720e-03
-2.50608742e-01 1.51044922e-02 -1.10037172e+00 2.10361767e+00
-2.86771595e-01 1.03881848e+00 -2.08389789e-01 -9.90806758e-01
4.05710727e-01 2.39881009e-01 7.63956964e-01 -6.13995194e-01
2.88556427e-01 -8.59289989e-02 -3.48789245e-01 -5.02146006e-01
6.16573274e-01 2.36389354e-01 5.49216606e-02 5.50568640e-01
2.36135870e-01 3.17659587e-01 7.43477762e-01 2.86451280e-01
1.65357053e+00 2.77815342e-01 3.10210943e-01 1.04586594e-01
1.24533258e-01 8.17678496e-03 4.79145944e-01 5.91752529e-01
-2.38175482e-01 8.73807669e-01 8.66525590e-01 -6.44255519e-01
-1.28570139e+00 -8.89158726e-01 3.73006165e-01 1.04506648e+00
3.89139280e-02 -5.93462706e-01 -1.21615040e+00 -3.97426158e-01
-2.57571042e-01 3.70745450e-01 -7.10647106e-01 -2.76731819e-01
-8.28806520e-01 -1.77517340e-01 3.94313365e-01 6.29793525e-01
4.46086526e-01 -1.15018058e+00 -1.36588550e+00 4.83015776e-01
6.13105781e-02 -1.09435380e+00 -5.06339252e-01 2.21208334e-01
-7.84469426e-01 -1.05737329e+00 -7.41131425e-01 -2.69636184e-01
3.54884177e-01 2.90493608e-01 1.00660276e+00 -4.96599004e-02
-8.60617876e-01 1.90919742e-01 -3.89524907e-01 -1.09871954e-01
-7.42298886e-02 -4.99903336e-02 1.81667563e-02 2.28628591e-02
3.07230622e-01 -7.16595709e-01 -7.48304367e-01 -1.53682664e-01
-1.11187327e+00 1.41572565e-01 3.79887223e-01 5.79895437e-01
5.79894781e-01 -2.34975159e-01 2.25886211e-01 -6.76981091e-01
2.46071547e-01 -2.53466487e-01 -5.98555088e-01 2.46736854e-01
-1.62453249e-01 -7.15438649e-02 7.41096914e-01 -4.84233588e-01
-7.83089578e-01 1.81256101e-01 -6.94649518e-02 -1.05484581e+00
-3.81748050e-01 2.65316814e-01 1.36885613e-01 2.21683949e-01
5.45379460e-01 2.63070554e-01 -3.03657129e-02 -4.34390247e-01
5.41326880e-01 4.40628141e-01 8.00204694e-01 -1.82239920e-01
4.27246332e-01 7.41231978e-01 -1.60840601e-01 -1.04619586e+00
-7.57881224e-01 -6.48347795e-01 -7.99986124e-01 -5.37902951e-01
1.00226462e+00 -8.50594640e-01 -6.46520853e-01 6.27230167e-01
-1.04140198e+00 -6.74278498e-01 -5.25146544e-01 3.52154940e-01
-8.33664596e-01 4.43893075e-01 -8.84727061e-01 -7.83762693e-01
-3.18015903e-01 -8.52796376e-01 1.08513820e+00 3.59026790e-01
-3.49796206e-01 -6.15387499e-01 8.93475488e-02 1.40711218e-01
2.62719512e-01 6.45820439e-01 6.65273070e-01 -3.59448135e-01
-6.83711886e-01 -2.04132438e-01 -7.17017800e-02 2.67521739e-01
-1.15997657e-01 5.10168374e-01 -8.98195148e-01 -1.75047919e-01
-2.58866578e-01 -4.01376992e-01 1.24765110e+00 5.53674579e-01
1.38698876e+00 -1.86344564e-01 -1.99959084e-01 6.78121924e-01
1.26055253e+00 1.47377461e-01 8.50911021e-01 1.95471987e-01
5.67108452e-01 4.13814396e-01 5.24445593e-01 7.68958986e-01
8.28432888e-02 6.39735103e-01 3.92672181e-01 -1.77219287e-02
-2.02422440e-01 -1.13764465e-01 4.32224959e-01 3.28932494e-01
-1.75601631e-01 -3.03592831e-01 -8.15930665e-01 6.44932151e-01
-2.03537703e+00 -1.35917556e+00 1.62001163e-01 2.36407185e+00
6.58931315e-01 6.02834523e-01 3.69396567e-01 2.46256441e-01
6.60412133e-01 6.10018730e-01 -6.55568480e-01 -3.05693716e-01
2.19960079e-01 2.42149934e-01 3.98808122e-01 2.39378363e-01
-1.49961162e+00 1.08872068e+00 6.20485115e+00 7.42349625e-01
-1.37539160e+00 -9.57540646e-02 6.14826500e-01 -5.53903580e-01
1.96922183e-01 -6.84006214e-02 -2.64502108e-01 4.56378102e-01
8.46650302e-01 3.78500298e-02 4.83088911e-01 8.80058050e-01
3.64723980e-01 -3.60597700e-01 -1.14002490e+00 1.05987263e+00
1.33005187e-01 -1.68095863e+00 -1.58871606e-01 -4.57865931e-02
6.59518003e-01 1.22585334e-01 -2.35870659e-01 1.99889332e-01
1.52821586e-01 -8.45854104e-01 1.00593889e+00 6.22960329e-01
8.13248217e-01 -6.45902991e-01 3.55136871e-01 1.54955879e-01
-1.17225349e+00 -2.29291692e-02 -3.16752642e-01 -2.57401168e-01
3.64311606e-01 6.47256494e-01 -5.66613913e-01 2.49891505e-01
6.28268778e-01 9.20621395e-01 -4.66656804e-01 1.28493047e+00
5.25565594e-02 5.12719095e-01 -3.96926165e-01 2.05118343e-01
2.24053711e-01 -6.11464456e-02 3.53984386e-01 1.40588856e+00
3.34306866e-01 8.26188549e-02 3.42679143e-01 3.94996554e-01
-1.68333232e-01 -1.22315802e-01 -5.40318191e-01 -1.98230445e-01
1.77359015e-01 1.20750880e+00 -1.26830697e+00 -6.92714751e-01
-4.24589425e-01 1.25761127e+00 2.38134116e-01 1.86932176e-01
-1.03672910e+00 -6.40271485e-01 8.31579268e-01 2.20807761e-01
6.28405035e-01 -4.85996902e-01 -7.89265558e-02 -1.34296799e+00
1.73256248e-01 -6.30264223e-01 2.79604852e-01 -5.60858965e-01
-6.80008650e-01 2.06130385e-01 -1.24054506e-01 -1.24909973e+00
-2.50979573e-01 -4.15274531e-01 -9.74169135e-01 6.50271997e-02
-1.25779390e+00 -7.26730347e-01 -4.45167691e-01 5.42789519e-01
9.32912588e-01 2.32384905e-01 4.06357944e-01 4.56878185e-01
-7.23005414e-01 1.03184491e-01 6.41801283e-02 3.27436179e-01
6.55919731e-01 -1.22671628e+00 6.27489746e-01 1.14204001e+00
4.11177546e-01 2.52964616e-01 7.74994671e-01 -4.86676663e-01
-1.30023968e+00 -1.01960611e+00 5.58545649e-01 -9.73672792e-02
6.57452941e-01 -3.86417150e-01 -8.12450230e-01 4.67602581e-01
2.83489227e-01 3.13142002e-01 4.52015638e-01 -1.49604633e-01
-2.19845936e-01 -7.74393901e-02 -6.46777332e-01 6.56916559e-01
1.18537295e+00 -4.59171891e-01 -4.54924285e-01 2.16547757e-01
5.55422664e-01 -3.64398509e-01 -4.59068537e-01 7.08510056e-02
5.28542638e-01 -1.21042871e+00 8.92258167e-01 -7.88668513e-01
7.84757376e-01 -3.54297191e-01 1.99490815e-01 -9.00909007e-01
-7.22204894e-02 -5.78094840e-01 -6.63121819e-01 9.55617309e-01
-1.28177376e-02 3.62760834e-02 7.91591048e-01 3.61106724e-01
-2.81967998e-01 -7.44641423e-01 -8.77371192e-01 -6.51491642e-01
-5.35777390e-01 -4.05886710e-01 2.19772220e-01 6.42741024e-01
9.75243524e-02 1.03511617e-01 -5.03297567e-01 -2.13258475e-01
5.59553027e-01 9.77746248e-02 7.85061896e-01 -1.09523952e+00
-1.79994479e-01 -5.10499179e-01 -6.78782105e-01 -1.02395785e+00
1.09801233e-01 -2.97175169e-01 2.72242337e-01 -1.54086494e+00
2.54283130e-01 7.13928416e-02 -4.06869590e-01 5.82292199e-01
-9.69179422e-02 5.70519745e-01 2.62435138e-01 -6.35384172e-02
-1.11691964e+00 4.34076756e-01 8.14830959e-01 -2.15754479e-01
-2.61183918e-01 -3.47158104e-01 -1.00692719e-01 8.75267923e-01
5.82334638e-01 -3.96845192e-01 -2.65690893e-01 -3.71725470e-01
2.16885619e-05 1.61701158e-01 4.64066744e-01 -1.44719851e+00
3.14422935e-01 -2.46458441e-01 4.88253862e-01 -4.60961908e-01
5.34481585e-01 -4.11649913e-01 4.94058616e-02 5.16464412e-01
-4.02699888e-01 -2.18291789e-01 -4.21489552e-02 5.90396225e-01
-3.39267284e-01 -1.65467396e-01 7.56435037e-01 -2.98861086e-01
-1.00588536e+00 4.10961539e-01 -5.74092627e-01 1.22241795e-01
1.37927210e+00 -3.67144287e-01 -1.27718210e-01 -2.49769509e-01
-7.17974842e-01 9.75056216e-02 8.15231025e-01 3.39499772e-01
4.90610182e-01 -1.12956941e+00 -4.56358820e-01 -4.40873429e-02
1.33643830e-02 -2.37867478e-02 5.44211447e-01 9.30093288e-01
-1.01448298e+00 1.90480784e-01 -4.95840460e-01 -6.28679454e-01
-1.29287326e+00 6.72985196e-01 -8.87598656e-03 -1.26456216e-01
-9.98588324e-01 7.47206688e-01 -1.02306649e-01 5.11650562e-01
2.62802780e-01 -3.53550553e-01 -1.91063404e-01 4.37747270e-01
9.03850973e-01 3.66080344e-01 1.92123186e-02 -2.63531119e-01
-3.21613133e-01 3.76278847e-01 -4.62321937e-02 -1.29415244e-01
1.33032489e+00 8.75646099e-02 2.06852823e-01 4.71389651e-01
1.19372547e+00 -3.44596088e-01 -1.93720675e+00 -6.57423884e-02
1.35993510e-01 -6.17723405e-01 5.75453788e-03 -9.11527220e-03
-9.59283948e-01 1.11793828e+00 5.44030190e-01 2.46386319e-01
1.12365091e+00 -1.95340976e-01 1.06855297e+00 4.42378074e-01
2.86392540e-01 -1.44032657e+00 3.66971791e-01 4.60305661e-01
3.80112976e-01 -1.14896154e+00 -1.00669056e-01 -1.57040041e-02
-5.05852997e-01 1.29621089e+00 2.82676041e-01 -3.31327885e-01
3.02855968e-01 3.58176231e-01 -1.50587529e-01 1.76871996e-02
-1.00719762e+00 -3.83768022e-01 -2.14925762e-02 3.38374019e-01
1.20391324e-01 -1.70928493e-01 -1.27455279e-01 -3.11446339e-02
1.90808192e-01 3.07487816e-01 7.42055655e-01 1.20246494e+00
-7.39675343e-01 -7.39626586e-01 3.18526104e-02 5.47430336e-01
-5.35552323e-01 4.81574610e-02 -1.39115065e-01 4.20758247e-01
-3.10833752e-02 6.58396602e-01 4.83530104e-01 -3.83001149e-01
1.29379839e-01 3.10534805e-01 4.72993553e-01 -5.54884791e-01
-3.45986068e-01 6.96901307e-02 -1.01284407e-01 -1.01299250e+00
-4.45036232e-01 -7.14384556e-01 -1.15478015e+00 -3.55938226e-01
3.54520977e-02 -2.56723821e-01 5.54402947e-01 9.07407403e-01
3.49132001e-01 6.57155573e-01 3.50410223e-01 -1.10995746e+00
-1.75465435e-01 -8.15008700e-01 -4.88818884e-01 4.98608768e-01
3.13931167e-01 -5.77782333e-01 -7.92582855e-02 6.09789014e-01] | [8.51487922668457, 0.23042486608028412] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.