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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
8f78e8de-7b48-44c9-a393-6a8c437233db | knowledge-guided-open-attribute-value | 2010.09189 | null | https://arxiv.org/abs/2010.09189v1 | https://arxiv.org/pdf/2010.09189v1.pdf | Knowledge-guided Open Attribute Value Extraction with Reinforcement Learning | Open attribute value extraction for emerging entities is an important but challenging task. A lot of previous works formulate the problem as a \textit{question-answering} (QA) task. While the collections of articles from web corpus provide updated information about the emerging entities, the retrieved texts can be noisy, irrelevant, thus leading to inaccurate answers. Effectively filtering out noisy articles as well as bad answers is the key to improving extraction accuracy. Knowledge graph (KG), which contains rich, well organized information about entities, provides a good resource to address the challenge. In this work, we propose a knowledge-guided reinforcement learning (RL) framework for open attribute value extraction. Informed by relevant knowledge in KG, we trained a deep Q-network to sequentially compare extracted answers to improve extraction accuracy. The proposed framework is applicable to different information extraction system. Our experimental results show that our method outperforms the baselines by 16.5 - 27.8\%. | ['Yanghua Xiao', 'Suo Feng', 'Rui Song', 'Sheng Zhang', 'Ye Liu'] | 2020-10-19 | null | https://aclanthology.org/2020.emnlp-main.693 | https://aclanthology.org/2020.emnlp-main.693.pdf | emnlp-2020-11 | ['attribute-value-extraction'] | ['natural-language-processing'] | [-1.72230303e-01 5.05964100e-01 -4.16388750e-01 -2.39735126e-01
-1.43915021e+00 -6.03056669e-01 6.80277869e-02 5.55377007e-01
-5.60871124e-01 1.34101820e+00 4.34316397e-01 7.17811510e-02
-2.38602147e-01 -1.24232137e+00 -8.21254075e-01 -2.72657394e-01
-1.05830403e-02 5.20188808e-01 2.96681076e-01 -4.32020992e-01
2.72602201e-01 -5.17464094e-02 -1.21154344e+00 2.70733535e-01
1.43326986e+00 1.17979288e+00 3.18526290e-02 3.43943506e-01
-7.43199587e-01 1.22169912e+00 -7.63130367e-01 -8.72618258e-01
-7.93162584e-02 3.95201752e-03 -1.39618170e+00 -4.69990015e-01
1.29338220e-01 -4.20708150e-01 -1.91033497e-01 1.27838898e+00
5.80755472e-01 8.21830332e-02 5.94017208e-01 -1.20432794e+00
-9.40498531e-01 1.09940970e+00 -4.91358370e-01 2.20190182e-01
5.43496013e-01 -2.25254729e-01 1.45859921e+00 -9.11119878e-01
1.00040388e+00 1.02807677e+00 3.91460568e-01 3.33657950e-01
-6.69493020e-01 -7.39422143e-01 2.51515955e-01 5.43142736e-01
-1.27329886e+00 -2.60198861e-01 6.57463610e-01 -9.26863030e-02
9.90288794e-01 1.43032268e-01 4.03887659e-01 8.73751640e-01
-1.51109993e-01 1.28284347e+00 7.44876564e-01 -1.57808989e-01
6.54513687e-02 1.08250789e-01 5.32999516e-01 4.02637959e-01
6.13771617e-01 -4.93569940e-01 -4.47702736e-01 -1.84625804e-01
1.46708176e-01 -1.74842238e-01 -2.16249883e-01 7.05146184e-03
-9.87718403e-01 8.38437557e-01 6.67412281e-01 9.52468291e-02
-6.58666790e-01 1.13019027e-01 5.34066439e-01 3.85503441e-01
4.35993820e-01 8.66957963e-01 -9.09719408e-01 -1.83197573e-01
-5.01613200e-01 6.72181845e-01 1.06074297e+00 1.13589895e+00
1.05751717e+00 -2.43977368e-01 -6.18423939e-01 7.33742058e-01
3.56272817e-01 6.32578671e-01 2.09761634e-01 -1.02661026e+00
1.11771071e+00 9.79964972e-01 5.28968036e-01 -1.05701911e+00
-4.41809505e-01 -5.08258283e-01 -7.00561762e-01 -5.99705100e-01
2.49125838e-01 -7.43781745e-01 -6.68689489e-01 1.59504509e+00
6.27122402e-01 -1.43092573e-01 3.78264159e-01 7.35369980e-01
1.54973602e+00 7.52386749e-01 4.83743787e-01 -3.91564220e-01
1.53227270e+00 -8.09415340e-01 -1.38744700e+00 -4.06098068e-02
5.58115065e-01 -5.46338260e-01 6.25666201e-01 2.71964282e-01
-9.05962050e-01 -1.56083226e-01 -9.01426017e-01 -2.94229358e-01
-6.04439080e-01 3.16481255e-02 4.07505780e-01 2.40385577e-01
-5.57552934e-01 4.75359201e-01 -3.20874304e-01 7.71985054e-02
6.67120636e-01 4.25572723e-01 -1.76950410e-01 3.65710966e-02
-1.96902883e+00 6.72684193e-01 6.09872043e-01 -2.13823300e-02
-3.59677166e-01 -5.96334279e-01 -8.32457781e-01 3.56552929e-01
1.03664958e+00 -7.11798668e-01 1.35863435e+00 -3.97910684e-01
-1.09060752e+00 4.12064672e-01 -1.35332808e-01 -6.15288615e-01
2.56963432e-01 -7.97798157e-01 -5.17201722e-01 2.31377080e-01
6.04571044e-01 3.88770014e-01 5.42218089e-01 -1.10891390e+00
-8.96029472e-01 -5.28470278e-01 1.66726753e-01 3.49402338e-01
-4.49544877e-01 -5.38966656e-02 -4.90732104e-01 -5.78326404e-01
-1.84515938e-01 -4.34506178e-01 -1.13912404e-01 -8.24493349e-01
-6.16763055e-01 -8.70611846e-01 5.17873764e-01 -8.99483025e-01
1.65428579e+00 -1.64730680e+00 -1.49341762e-01 1.06234513e-01
5.28874218e-01 1.63846567e-01 6.69670403e-02 6.74326479e-01
3.94105583e-01 4.35356647e-01 -6.45968392e-02 2.58724838e-01
1.98535219e-01 -2.06560716e-02 -5.06563365e-01 -3.08058232e-01
4.95378107e-01 1.34753454e+00 -1.12016559e+00 -8.02549601e-01
-4.63997275e-01 -7.18270382e-03 -2.38460004e-01 2.40632057e-01
-6.09903157e-01 1.08623512e-01 -1.13428843e+00 8.09600592e-01
7.73676693e-01 -4.82359409e-01 -9.08797234e-03 -1.66612595e-01
3.67078394e-01 2.51015276e-01 -1.38582802e+00 1.38113463e+00
-1.62076786e-01 2.50139117e-01 -1.99327290e-01 -8.63719642e-01
8.78339469e-01 3.02975893e-01 4.69738454e-01 -9.03817594e-01
1.72037795e-01 2.57563561e-01 -3.65001202e-01 -9.03886616e-01
8.92596066e-01 4.25169796e-01 -3.74869823e-01 2.05844760e-01
2.20337704e-01 1.41792610e-01 5.02934039e-01 6.42275274e-01
1.29604757e+00 3.90194394e-02 2.79287010e-01 6.73886314e-02
6.00360632e-01 6.47965148e-02 1.05593586e+00 7.42367506e-01
-2.13615373e-01 -2.36244183e-02 6.92132175e-01 -2.83721358e-01
-5.74183404e-01 -9.57926035e-01 -5.80551140e-02 9.03391540e-01
4.20457125e-01 -5.87837160e-01 -9.11697805e-01 -1.02954352e+00
1.81906864e-01 5.60507298e-01 -4.73169237e-01 -1.45628154e-01
-5.02401054e-01 -6.80511653e-01 3.82663876e-01 5.97218692e-01
8.38917196e-01 -1.40150869e+00 -1.70219049e-01 4.78322834e-01
-9.12151039e-01 -1.32184756e+00 -1.32399157e-01 1.61924198e-01
-6.06274247e-01 -1.29819047e+00 -5.72286665e-01 -7.44110167e-01
3.57484967e-01 -9.70297828e-02 1.57990932e+00 -1.40790744e-02
2.91909844e-01 1.55842468e-01 -8.17969143e-01 -7.25944519e-01
-5.61713576e-02 5.06351829e-01 -2.10493445e-01 -3.68040837e-02
8.98395300e-01 -4.41801250e-02 -8.69677782e-01 -5.15837036e-02
-6.79572165e-01 -4.50541586e-01 6.85466230e-01 9.83263910e-01
8.32303107e-01 -1.85804497e-02 1.45279658e+00 -1.50555170e+00
1.09144318e+00 -8.04697990e-01 -4.50887203e-01 4.96908545e-01
-8.91912043e-01 3.83537978e-01 7.88776338e-01 1.69188753e-02
-1.43802011e+00 -2.08735809e-01 -3.11466336e-01 2.48990253e-01
4.34131809e-02 8.41003597e-01 -3.43482584e-01 3.17050457e-01
5.23576677e-01 -3.87267210e-02 -7.14501202e-01 -4.06901926e-01
5.01250386e-01 8.14719260e-01 2.34424129e-01 -6.05679810e-01
7.07424819e-01 7.51699656e-02 -3.71736526e-01 -4.18331861e-01
-1.36539018e+00 -7.20575631e-01 -2.08175406e-01 4.99900281e-02
7.67579556e-01 -1.05166304e+00 -7.67325222e-01 6.00474440e-02
-1.16719413e+00 2.84230828e-01 -3.95921141e-01 1.55575857e-01
-2.80254811e-01 2.64590949e-01 -5.55484951e-01 -8.03988695e-01
-9.09839153e-01 -8.78490746e-01 1.04518127e+00 6.79558098e-01
2.38368586e-02 -4.34051841e-01 9.67850685e-02 7.88489521e-01
1.53326824e-01 2.49900639e-01 8.30930412e-01 -1.05531931e+00
-8.88498962e-01 -2.60360032e-01 -3.90362978e-01 1.20884217e-01
-3.63406092e-02 -3.10888797e-01 -7.89217770e-01 2.03400508e-01
-3.20545822e-01 -7.10454404e-01 1.16929817e+00 1.06010944e-01
1.07581913e+00 -5.88531673e-01 -3.60774487e-01 1.14555180e-01
1.28121340e+00 4.17339206e-02 5.98884642e-01 4.99724954e-01
6.79654539e-01 6.11123025e-01 1.11318743e+00 6.27011716e-01
9.38784540e-01 1.11719601e-01 3.78584415e-01 2.11995080e-01
2.92591363e-01 -4.41281706e-01 3.02117318e-04 8.14130485e-01
4.85814102e-02 -2.49983951e-01 -7.41717339e-01 8.32114220e-01
-2.09708619e+00 -1.00173128e+00 -2.22598732e-01 1.68368185e+00
1.28032351e+00 2.53925681e-01 -2.67853532e-02 7.99985901e-02
5.10437489e-01 2.68531311e-02 -7.66015053e-01 -8.16302076e-02
-9.73808542e-02 3.13686043e-01 3.83287460e-01 1.84969291e-01
-1.28000391e+00 9.92096245e-01 4.89368248e+00 1.04218161e+00
-2.72630125e-01 -8.29231367e-02 4.73860651e-01 3.05191875e-01
-5.91749966e-01 -3.04021444e-02 -1.12487960e+00 4.83161718e-01
7.71913826e-01 -5.61268449e-01 -1.40612246e-02 8.53337467e-01
-2.05524504e-01 -2.58264393e-01 -6.86571836e-01 8.73738706e-01
-2.26911947e-01 -1.24468076e+00 1.18328802e-01 -2.82304853e-01
9.39781606e-01 4.13275771e-02 -2.00834796e-01 9.09196854e-01
7.00592279e-01 -7.08652496e-01 2.42564440e-01 7.97064066e-01
3.44790578e-01 -1.03106630e+00 1.14913833e+00 4.20041084e-01
-1.36432111e+00 -2.57587403e-01 -4.27543074e-01 2.69498706e-01
-1.94740482e-03 8.13181460e-01 -7.26011753e-01 8.78789842e-01
1.08529699e+00 7.28198528e-01 -6.36743546e-01 7.60269403e-01
-5.64968646e-01 6.12477779e-01 -9.68949869e-02 -5.49130797e-01
7.69835487e-02 -9.12249386e-02 2.00922534e-01 9.24491584e-01
-5.94532304e-02 3.30822378e-01 7.53359199e-02 7.16464341e-01
-9.84197319e-01 6.36477470e-01 -5.02472878e-01 -1.67617753e-01
6.08630955e-01 1.49653780e+00 -4.31998640e-01 -5.04431665e-01
-5.67501962e-01 6.75385296e-01 8.11543882e-01 4.26129192e-01
-5.07643580e-01 -1.13093221e+00 2.21371248e-01 -2.76303262e-01
3.98508430e-01 2.48906612e-01 -6.72413334e-02 -1.41944182e+00
3.47839713e-01 -8.76302898e-01 8.54527950e-01 -6.43178761e-01
-1.52764320e+00 4.34354097e-01 -3.50785494e-01 -1.11314821e+00
-3.61062497e-01 -1.77773476e-01 -8.08795169e-02 6.06905401e-01
-2.04648376e+00 -9.09189284e-01 -2.90725350e-01 4.19630378e-01
3.87042284e-01 -6.90571144e-02 6.81314290e-01 3.99045914e-01
-4.89976645e-01 5.48919618e-01 3.73286724e-01 6.50062263e-01
7.46108830e-01 -1.73806334e+00 3.30935031e-01 5.07517099e-01
2.07016036e-01 4.68860537e-01 4.24772412e-01 -8.79776776e-01
-1.44006348e+00 -1.19149733e+00 9.74140942e-01 -6.78379714e-01
5.40503681e-01 -1.78127848e-02 -1.06800270e+00 4.55641508e-01
5.49688637e-01 -2.24631671e-02 8.51688683e-01 3.17594379e-01
-1.68484449e-01 -4.70933795e-01 -1.11317873e+00 4.18002814e-01
8.04936707e-01 -3.66065651e-01 -8.70674074e-01 1.96385980e-01
1.23156631e+00 -3.59404564e-01 -1.24905145e+00 5.35848796e-01
1.93382695e-01 -5.09396553e-01 8.29418898e-01 -7.02469528e-01
5.48428118e-01 -8.13587382e-02 1.31985411e-01 -1.33164895e+00
-6.82487339e-02 -4.71480459e-01 -8.58231723e-01 1.52066386e+00
6.87909305e-01 -2.67475516e-01 7.65474796e-01 7.12066650e-01
2.85470515e-01 -9.76015568e-01 -6.50545776e-01 -3.41227323e-01
1.06773771e-01 -1.02465481e-01 8.09798658e-01 9.13463414e-01
2.20860049e-01 8.67793560e-01 -1.00556649e-01 1.56328321e-01
5.67327678e-01 4.66250449e-01 6.04783475e-01 -1.42779219e+00
1.84067056e-01 -5.42157292e-02 3.11063994e-02 -9.80167389e-01
1.46380618e-01 -9.18983400e-01 -1.91253617e-01 -2.00387979e+00
2.91695684e-01 -3.01859051e-01 -5.16706049e-01 3.56898516e-01
-7.20569849e-01 -2.20037267e-01 -9.88142192e-03 -2.41794929e-04
-1.35649431e+00 8.87548447e-01 1.32660818e+00 -3.66361380e-01
-1.95417926e-01 -7.03139678e-02 -1.07996273e+00 5.55553675e-01
8.99165094e-01 -5.45300543e-01 -3.30919653e-01 -3.38284701e-01
9.79981005e-01 3.17940831e-01 -4.53754514e-02 -6.21625721e-01
4.30634767e-01 -8.38859752e-02 5.04736245e-01 -9.93632972e-01
-1.14678340e-02 -7.25194275e-01 -5.99061906e-01 3.72118019e-02
-3.84263992e-01 -7.03579783e-02 7.86377713e-02 8.41190577e-01
-5.77202499e-01 -3.80861551e-01 9.73987803e-02 -2.56840974e-01
-8.78249645e-01 5.41567683e-01 1.18099647e-02 1.02527559e+00
8.01928461e-01 3.41260999e-01 -6.10597670e-01 -4.22367543e-01
-6.95133507e-01 1.11820781e+00 -4.27828342e-01 4.58591372e-01
8.20393085e-01 -1.34769011e+00 -7.97242880e-01 -2.24873349e-01
3.40151101e-01 4.05632496e-01 2.73608267e-01 3.69128227e-01
-2.19742417e-01 2.80967742e-01 9.53953937e-02 -4.24801372e-04
-8.19086313e-01 5.85693777e-01 1.86920445e-02 -8.50586891e-01
-1.71314403e-01 7.58083820e-01 -2.54074603e-01 -5.36047399e-01
3.01910341e-01 -1.37936100e-01 -9.85814810e-01 4.67334121e-01
4.39054966e-01 4.40360308e-01 1.36610270e-01 -2.54501641e-01
-1.13587022e-01 2.76933104e-01 -5.36546290e-01 2.53519505e-01
1.43585622e+00 -3.58453602e-01 -1.97440431e-01 3.93594988e-02
9.47705567e-01 3.81038785e-02 -6.10718906e-01 -6.74675941e-01
5.18386900e-01 -3.74180712e-02 -8.70085414e-03 -9.46177185e-01
-8.90560389e-01 6.55175805e-01 2.12082580e-01 4.71595287e-01
9.95402336e-01 4.22832482e-02 1.26372159e+00 9.94122744e-01
3.51509064e-01 -1.69462740e+00 2.60008276e-02 6.53288364e-01
8.80243301e-01 -1.67352057e+00 4.42416370e-02 -4.43776995e-01
-7.81475306e-01 8.61426532e-01 1.08495033e+00 6.39419034e-02
7.04970539e-01 -9.26200449e-02 7.16050863e-02 -3.97091866e-01
-7.03925550e-01 -5.89797497e-01 2.94525057e-01 3.90811592e-01
2.82643527e-01 -1.45224854e-01 -5.94354808e-01 1.33578205e+00
-1.03736155e-01 5.91701865e-02 4.11357731e-01 9.09995079e-01
-5.45247197e-01 -1.01211417e+00 -8.82107988e-02 1.03020227e+00
-8.70266438e-01 -2.06645787e-01 -5.20017624e-01 4.90082502e-01
-1.95570383e-02 1.16999805e+00 -4.58705574e-01 -2.33006179e-01
5.19841611e-01 1.59075961e-01 3.59276421e-02 -5.58514595e-01
-6.86934769e-01 -2.45884538e-01 2.64688015e-01 -3.75937551e-01
-4.45198178e-01 -3.33123386e-01 -1.57764328e+00 3.48165333e-02
-6.85062706e-01 6.48090303e-01 1.56076804e-01 8.87062132e-01
6.41735137e-01 6.91283286e-01 6.63958490e-01 3.26729685e-01
-7.98209727e-01 -9.82976615e-01 -5.40337384e-01 5.59175491e-01
3.81453127e-01 -6.20113909e-01 1.04917496e-01 -1.10380404e-01] | [10.350749015808105, 8.102991104125977] |
f40ee9e2-4b20-460b-8114-b7549ac2fee9 | generating-high-quality-3dmpcs-by-adaptive | 2305.06777 | null | https://arxiv.org/abs/2305.06777v1 | https://arxiv.org/pdf/2305.06777v1.pdf | Generating high-quality 3DMPCs by adaptive data acquisition and NeREF-based reflectance correction to facilitate efficient plant phenotyping | Non-destructive assessments of plant phenotypic traits using high-quality three-dimensional (3D) and multispectral data can deepen breeders' understanding of plant growth and allow them to make informed managerial decisions. However, subjective viewpoint selection and complex illumination effects under natural light conditions decrease the data quality and increase the difficulty of resolving phenotypic parameters. We proposed methods for adaptive data acquisition and reflectance correction respectively, to generate high-quality 3D multispectral point clouds (3DMPCs) of plants. In the first stage, we proposed an efficient next-best-view (NBV) planning method based on a novel UGV platform with a multi-sensor-equipped robotic arm. In the second stage, we eliminated the illumination effects by using the neural reference field (NeREF) to predict the digital number (DN) of the reference. We tested them on 6 perilla and 6 tomato plants, and selected 2 visible leaves and 4 regions of interest (ROIs) for each plant to assess the biomass and the chlorophyll content. For NBV planning, the average execution time for single perilla and tomato plant at a joint speed of 1.55 rad/s was 58.70 s and 53.60 s respectively. The whole-plant data integrity was improved by an average of 27% compared to using fixed viewpoints alone, and the coefficients of determination (R2) for leaf biomass estimation reached 0.99 and 0.92. For reflectance correction, the average root mean squared error of the reflectance spectra with hemisphere reference-based correction at different ROIs was 0.08 and 0.07 for perilla and tomato. The R2 of chlorophyll content estimation was 0.91 and 0.93 respectively when principal component analysis and Gaussian process regression were applied. Our approach is promising for generating high-quality 3DMPCs of plants under natural light conditions and facilitates accurate plant phenotyping. | ['Haiyan Cen', 'Jiangpeng Zhu', 'Xuqi Lu', 'YuTao Shen', 'Mengqi Lv', 'Ruiming Du', 'Zhihong Ma', 'Pengyao Xie'] | 2023-05-11 | null | null | null | null | ['plant-phenotyping'] | ['computer-vision'] | [ 2.41337225e-01 -3.77496660e-01 2.05594957e-01 3.19507383e-02
-2.68652141e-01 -1.01240587e+00 -1.88496143e-01 2.55041689e-01
2.40107283e-01 5.69678605e-01 -7.28624701e-01 -4.42373931e-01
-5.69269180e-01 -1.15056086e+00 -2.04120368e-01 -1.12371862e+00
2.17550546e-01 1.92429081e-01 2.75714159e-01 -2.11848333e-01
1.21828876e-02 1.10167885e+00 -1.67402005e+00 4.81710769e-02
1.17674482e+00 9.21419561e-01 1.10452425e+00 8.76210272e-01
-7.37416521e-02 3.49260829e-02 -4.92047817e-01 1.25845566e-01
3.77356350e-01 -7.23034143e-02 -3.15944105e-01 5.95490277e-01
2.97855642e-02 -3.12255710e-01 4.45035696e-01 9.34297323e-01
6.38282776e-01 -1.29408970e-01 5.66407323e-01 -1.23138380e+00
-8.76499414e-01 2.37587079e-01 -1.35206926e+00 -5.22710383e-01
-1.32746309e-01 3.83066714e-01 6.25985503e-01 -4.43609267e-01
3.61591309e-01 8.36126983e-01 6.49168432e-01 1.61367118e-01
-1.59875834e+00 -1.23771586e-01 -7.76928365e-02 1.26639739e-01
-1.52680027e+00 -8.26663449e-02 4.91621137e-01 -5.65277755e-01
4.04552996e-01 4.71572161e-01 9.78095531e-01 1.63635388e-01
4.43504840e-01 4.79358584e-02 1.13725877e+00 -4.17275846e-01
1.16585810e-02 -3.87242325e-02 -1.59325793e-01 5.45321167e-01
4.75271106e-01 1.90424785e-01 2.63564110e-01 -3.72433886e-02
9.94589567e-01 -1.27232939e-01 -4.03128445e-01 -4.90225047e-01
-8.22054386e-01 5.35293996e-01 3.36205989e-01 2.59674102e-01
-8.03563595e-01 -3.96786213e-01 1.02805451e-01 -1.64393216e-01
1.60278514e-01 2.32668847e-01 -9.49057162e-01 3.34362775e-01
-6.24959707e-01 -1.95150793e-01 6.61412835e-01 1.03952861e+00
5.85604608e-01 1.77510515e-01 -2.07967058e-01 1.07066989e+00
4.27435905e-01 1.24292469e+00 -1.57042995e-01 -1.26145756e+00
-2.90310174e-01 9.49765205e-01 4.49267328e-01 -1.03831220e+00
-4.52254415e-01 -2.27991611e-01 -9.15549636e-01 5.32959223e-01
4.67694700e-01 -7.67134950e-02 -7.83673584e-01 1.19184279e+00
3.57221365e-01 -4.48782116e-01 2.49274909e-01 7.16645658e-01
6.64836407e-01 8.89429867e-01 -8.07072818e-02 -6.52259648e-01
1.50447607e+00 -4.10783559e-01 -5.91369808e-01 8.56407285e-02
5.91946185e-01 -1.09669542e+00 1.07524037e+00 5.09188175e-01
-1.11418176e+00 -6.00570977e-01 -7.57406294e-01 4.79202867e-01
-1.26208827e-01 1.15822697e+00 4.53908175e-01 7.41941452e-01
-9.66407418e-01 7.12872624e-01 -6.55077994e-01 -4.67695743e-01
4.81653035e-01 2.63444066e-01 -3.67110223e-01 -6.30339757e-02
-1.86028346e-01 8.35887611e-01 2.65580088e-01 6.64453387e-01
-8.96526754e-01 -8.19262147e-01 -1.94323305e-02 2.65215904e-01
3.29960674e-01 -2.87651211e-01 6.93720460e-01 -8.09971452e-01
-2.08993411e+00 1.06990767e+00 -6.88256370e-03 2.83035010e-01
-1.30456731e-01 1.29991710e-01 -3.63113910e-01 2.34768853e-01
-1.25078196e-02 3.44406098e-01 4.27915484e-01 -1.49408948e+00
-5.97271025e-01 -8.87794137e-01 -2.09849089e-01 1.03679158e-01
-1.32143781e-01 8.52423906e-02 2.25840673e-01 2.23698154e-01
8.18475366e-01 -1.09517932e+00 -3.46622705e-01 5.50916612e-01
-1.63703829e-01 4.46825624e-01 9.17735338e-01 -9.23039377e-01
3.85295391e-01 -2.03332400e+00 6.41184449e-02 2.95831021e-02
-1.59943730e-01 5.86770058e-01 -2.90664524e-01 3.11637707e-02
-1.22886099e-01 1.22343376e-01 -1.82968736e-01 5.05761683e-01
-6.78802252e-01 2.59196073e-01 1.52613774e-01 5.93189597e-01
9.71318558e-02 4.21630353e-01 -5.88881135e-01 -1.73074305e-01
3.29755336e-01 5.04500568e-01 -1.56133091e-02 2.04373077e-01
1.57633275e-02 2.56158918e-01 -2.82769650e-01 1.01911688e+00
1.47596478e+00 1.93598658e-01 1.55804798e-01 -5.94510913e-01
-9.13539231e-01 -6.38345778e-01 -1.28788340e+00 1.30001760e+00
-4.88133550e-01 2.30903506e-01 7.45542228e-01 -7.67666399e-01
1.50182307e+00 2.85279989e-01 7.51787305e-01 -9.45712179e-02
1.58414021e-02 2.49032319e-01 -1.76551864e-02 -3.78937215e-01
1.93695381e-01 1.24684826e-01 7.68926024e-01 -2.55195588e-01
-2.30057344e-01 -6.99939668e-01 3.74589935e-02 -2.91607738e-01
3.95535618e-01 5.46537936e-01 2.61379540e-01 -6.31204844e-01
8.07704806e-01 3.54754508e-01 8.80364537e-01 1.29662231e-01
-3.51667643e-01 3.26575369e-01 6.66178584e-01 4.51492965e-02
-1.05254316e+00 -9.33367074e-01 -2.40403339e-01 6.82501495e-01
1.03171378e-01 1.17064483e-01 -4.00994122e-01 -9.97285247e-02
1.52898550e-01 8.72988641e-01 5.83154988e-03 4.27349471e-02
4.40592654e-02 -1.15823448e+00 2.12756529e-01 2.06442595e-01
5.82525849e-01 -7.23968804e-01 -6.49523914e-01 1.24970987e-01
-1.71104148e-02 -1.02385998e+00 3.65423888e-01 1.25275373e-01
-1.21324539e+00 -1.04876399e+00 -6.70589745e-01 -3.26487362e-01
5.89361131e-01 8.72675955e-01 1.02520692e+00 -9.58180148e-03
-5.13764560e-01 1.57256141e-01 -5.76864541e-01 -5.48285723e-01
-5.26050687e-01 -3.55576962e-01 -2.87137210e-01 -2.92310596e-01
4.78326268e-02 -6.43697262e-01 -3.01329613e-01 6.57544255e-01
-4.86059815e-01 1.15796886e-02 4.69467193e-01 8.76583517e-01
9.59086180e-01 1.79745868e-01 6.64677843e-02 -4.23057944e-01
-7.00245425e-02 1.41326189e-02 -1.67275214e+00 5.26129127e-01
-6.37801707e-01 -6.77266002e-01 6.11055434e-01 -2.21760124e-01
-1.25011015e+00 3.85296226e-01 4.38337833e-01 2.11995225e-02
-5.58513403e-01 3.68771702e-01 -5.20950019e-01 -2.16057643e-01
7.94497788e-01 7.92326033e-02 2.33769596e-01 -3.18551451e-01
2.76069969e-01 5.81877291e-01 3.91015321e-01 -3.60731870e-01
8.82946193e-01 3.89551103e-01 6.73310518e-01 -1.28077841e+00
-3.97430748e-01 -4.03709322e-01 -1.05559349e+00 -4.51795161e-01
8.68417442e-01 -7.63800979e-01 -1.44330907e+00 9.08284426e-01
-1.15354943e+00 -2.23864481e-01 -2.32079208e-01 6.94363892e-01
-5.12078047e-01 4.76949334e-01 -2.32530802e-01 -1.01076126e+00
-4.44647491e-01 -1.15971768e+00 8.23008716e-01 5.25634825e-01
5.80897629e-01 -6.81039512e-01 -3.69807035e-01 1.97053283e-01
3.87432605e-01 2.98425734e-01 8.18757772e-01 3.17742616e-01
-6.19289577e-01 -2.77174413e-01 -7.57769227e-01 5.60317934e-01
3.26885253e-01 1.09012043e+00 -1.14881802e+00 -1.36524543e-01
2.29134075e-02 -6.12398610e-03 1.10054158e-01 1.04259503e+00
1.08480537e+00 3.27738464e-01 -2.75048345e-01 7.29613364e-01
2.01629543e+00 4.32063341e-01 5.99928498e-01 -1.40830055e-01
6.83296800e-01 8.15549374e-01 9.84047830e-01 6.38509572e-01
-1.50322571e-01 5.59918523e-01 1.10468173e+00 -2.19453245e-01
7.29111359e-02 3.31806302e-01 9.99020189e-02 4.18051153e-01
-6.75582945e-01 -4.41416860e-01 -8.09795260e-01 4.47128683e-01
-1.28974378e+00 -1.03521287e+00 -1.11819375e+00 2.34244204e+00
4.88815188e-01 -4.16478485e-01 -1.45206138e-01 2.62311369e-01
9.22715604e-01 -1.71514645e-01 -4.57213402e-01 -3.59487593e-01
-5.76493442e-01 -9.23141167e-02 1.11512089e+00 3.56613994e-01
-5.73555589e-01 8.39109600e-01 5.38301945e+00 3.16143930e-01
-9.46269929e-01 -1.61455631e-01 3.58624727e-01 3.75066370e-01
-1.31196544e-01 4.40968156e-01 -9.71825361e-01 -2.40199000e-01
8.68926823e-01 9.16118100e-02 5.22874415e-01 8.21182430e-01
8.03342998e-01 -5.97400129e-01 -1.89577103e-01 6.23275876e-01
-3.97137105e-01 -7.43274629e-01 -4.70713526e-01 2.67225713e-01
6.29413247e-01 -1.12418540e-01 -4.76854622e-01 -3.56651694e-01
4.46450979e-01 -4.58075076e-01 1.33150235e-01 5.72944403e-01
7.98004329e-01 -4.63596523e-01 6.07825220e-01 5.01312733e-01
-1.37516356e+00 -2.75256395e-01 -1.08242595e+00 1.71710581e-01
-7.24502355e-02 1.10421062e+00 -7.78578818e-01 9.51331556e-01
6.33540213e-01 4.85497862e-01 -6.25903964e-01 8.63692403e-01
-1.94231600e-01 5.46523094e-01 -5.93264878e-01 2.42378965e-01
-1.13608159e-01 -7.69177794e-01 5.69795549e-01 4.12899315e-01
8.63759100e-01 2.62723923e-01 -3.96203659e-02 1.19654977e+00
6.79700732e-01 4.18504357e-01 -2.71021605e-01 -1.15303077e-01
5.07510483e-01 1.76523554e+00 -9.32171643e-01 9.08692926e-02
-1.93932876e-01 8.38770688e-01 -4.00882006e-01 4.13678363e-02
-6.82102621e-01 -1.97307572e-01 2.67648935e-01 -9.49514434e-02
3.03334624e-01 -3.04218143e-01 -4.59927946e-01 -6.82863057e-01
-1.69209406e-01 -5.01091242e-01 -6.55654594e-02 -1.43660259e+00
-9.60754097e-01 1.02231376e-01 -1.19255722e-01 -1.01144099e+00
4.63917136e-01 -1.07875371e+00 -3.63427103e-01 1.61021423e+00
-1.26520729e+00 -1.33495581e+00 -1.08193994e+00 5.86037152e-02
3.74742627e-01 -1.39467075e-01 1.49638343e+00 -5.25045134e-02
-8.71720016e-01 -8.21014643e-02 6.98042810e-01 -7.13650286e-01
4.30662543e-01 -9.46068883e-01 -3.05985808e-01 8.93511713e-01
-4.29061323e-01 -2.40811542e-01 4.40093130e-01 -6.02188051e-01
-1.59237123e+00 -8.07858288e-01 6.08909607e-01 1.46909179e-02
2.92536974e-01 3.74187171e-01 -7.87588716e-01 2.11614549e-01
-1.21911429e-01 7.10236281e-02 4.92861748e-01 -1.71331286e-01
3.78902584e-01 -5.63394547e-01 -1.42919016e+00 4.47174877e-01
5.83022892e-01 -5.01727425e-02 5.35568774e-01 5.47723413e-01
4.61241812e-01 -2.73927391e-01 -1.37081087e+00 4.73418355e-01
8.01524043e-01 -9.21046019e-01 1.09683287e+00 -2.72472613e-02
1.95709720e-01 -6.59253359e-01 -5.70506275e-01 -1.35217953e+00
-8.93462896e-01 -4.85322148e-01 6.90072656e-01 1.51645529e+00
2.18823731e-01 -4.65313435e-01 4.70932573e-01 2.40765408e-01
-3.30445141e-01 -1.64475471e-01 -1.59764469e-01 -6.87153697e-01
-1.83987543e-01 7.37890601e-02 6.16592944e-01 5.56810200e-01
-7.78844774e-01 -8.78008828e-02 7.93996453e-02 1.00109518e+00
5.56999326e-01 3.41688275e-01 7.60767221e-01 -1.97671044e+00
-5.46350852e-02 -1.94325224e-01 -1.77968353e-01 -4.34949756e-01
-6.71524853e-02 -6.35837853e-01 -1.12524070e-01 -1.59138763e+00
7.13696778e-02 -4.47958052e-01 3.12205940e-01 3.99326533e-01
-3.30260098e-02 -3.91847938e-01 1.79252848e-01 1.42123565e-01
9.10275519e-01 2.30182320e-01 1.50437415e+00 7.26928264e-02
-5.70340931e-01 6.06322348e-01 -5.04322767e-01 8.51926327e-01
1.16459537e+00 -3.29633862e-01 -5.07110119e-01 -4.89253789e-01
-6.89019402e-03 3.95646662e-01 6.05173111e-01 -7.89917886e-01
-4.05644894e-01 -7.37617612e-01 4.80676085e-01 -1.06109178e+00
4.55059886e-01 -1.31830144e+00 5.33685744e-01 5.16639173e-01
2.80123800e-01 -1.59378201e-02 4.24095929e-01 1.63631767e-01
4.58367914e-01 -7.14329839e-01 1.09089434e+00 -1.90829128e-01
-4.41036761e-01 1.63977191e-01 -3.02743763e-01 -6.07543766e-01
1.27699518e+00 -5.00416875e-01 -4.48578835e-01 6.39276206e-02
-5.21764934e-01 -1.08298875e-01 5.51733851e-01 -1.75083175e-01
3.68303716e-01 -1.06497157e+00 -9.25134718e-01 3.21149826e-01
-9.87613797e-02 -4.63970341e-02 4.67391700e-01 9.87095535e-01
-1.14977682e+00 2.66415238e-01 -8.50354731e-01 -9.25725162e-01
-1.59384286e+00 2.64806509e-01 4.10702020e-01 1.71299502e-01
-1.88508213e-01 6.67200089e-01 -1.27614722e-01 -3.24285865e-01
-3.96190524e-01 -3.61040205e-01 -5.21248341e-01 9.70352441e-02
1.20745994e-01 9.01407480e-01 -4.23484063e-03 -6.69241726e-01
-2.49988720e-01 7.36152828e-01 6.25326514e-01 2.04584390e-01
1.43791008e+00 -3.23911875e-01 -4.43747729e-01 5.06470621e-01
4.77527916e-01 7.20020160e-02 -1.25366116e+00 4.60772336e-01
-4.20220822e-01 -7.35356688e-01 3.98119539e-01 -9.89195526e-01
-1.28411329e+00 7.48712659e-01 9.77068007e-01 6.74202800e-01
1.66183078e+00 -6.98983192e-01 -3.20733413e-02 1.76133364e-01
-1.58354864e-02 -8.77557993e-01 -7.65847921e-01 2.94443220e-01
9.28589642e-01 -1.05382466e+00 3.76152456e-01 -1.05162776e+00
-5.36577165e-01 1.43572760e+00 7.18339026e-01 1.96097031e-01
5.14807463e-01 2.89012432e-01 5.26996702e-02 -2.09418330e-02
-6.03986681e-01 -1.29219040e-01 -2.51073539e-01 1.19505382e+00
6.35160208e-01 3.47865701e-01 -3.19722176e-01 1.98790267e-01
2.42439523e-01 1.49041891e-01 7.69101918e-01 6.82060838e-01
-6.88036799e-01 -9.39504743e-01 -8.63903880e-01 2.57465452e-01
9.41419974e-02 1.67019978e-01 -9.85640511e-02 5.73374927e-01
6.38410375e-02 1.02132630e+00 -1.19890146e-01 -2.58934684e-02
6.33326530e-01 -6.24078810e-02 5.73994040e-01 -4.45795923e-01
-2.64774859e-01 3.94029677e-01 -2.08136097e-01 -4.20393378e-01
-5.71495652e-01 -8.86942327e-01 -8.28053594e-01 -4.85365510e-01
-9.93984282e-01 -2.69274652e-01 1.14492834e+00 3.06096822e-01
3.40477496e-01 6.13676608e-01 1.03265393e+00 -6.32281601e-01
-5.51170230e-01 -8.26135933e-01 -1.18669045e+00 -3.39433551e-01
-4.10426527e-01 -6.02846920e-01 -2.10596770e-01 3.65709633e-01] | [9.126110076904297, -1.6343046426773071] |
955e3825-a239-4f3f-81fa-3d8a1d6193ec | deep-semantic-ranking-based-hashing-for-multi | 1501.06272 | null | http://arxiv.org/abs/1501.06272v2 | http://arxiv.org/pdf/1501.06272v2.pdf | Deep Semantic Ranking Based Hashing for Multi-Label Image Retrieval | With the rapid growth of web images, hashing has received increasing
interests in large scale image retrieval. Research efforts have been devoted to
learning compact binary codes that preserve semantic similarity based on
labels. However, most of these hashing methods are designed to handle simple
binary similarity. The complex multilevel semantic structure of images
associated with multiple labels have not yet been well explored. Here we
propose a deep semantic ranking based method for learning hash functions that
preserve multilevel semantic similarity between multi-label images. In our
approach, deep convolutional neural network is incorporated into hash functions
to jointly learn feature representations and mappings from them to hash codes,
which avoids the limitation of semantic representation power of hand-crafted
features. Meanwhile, a ranking list that encodes the multilevel similarity
information is employed to guide the learning of such deep hash functions. An
effective scheme based on surrogate loss is used to solve the intractable
optimization problem of nonsmooth and multivariate ranking measures involved in
the learning procedure. Experimental results show the superiority of our
proposed approach over several state-of-the-art hashing methods in term of
ranking evaluation metrics when tested on multi-label image datasets. | ['Tieniu Tan', 'Liang Wang', 'Fang Zhao', 'Yongzhen Huang'] | 2015-01-26 | deep-semantic-ranking-based-hashing-for-multi-1 | http://openaccess.thecvf.com/content_cvpr_2015/html/Zhao_Deep_Semantic_Ranking_2015_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2015/papers/Zhao_Deep_Semantic_Ranking_2015_CVPR_paper.pdf | cvpr-2015-6 | ['multi-label-image-retrieval'] | ['computer-vision'] | [ 9.64344144e-02 -2.73911059e-01 -2.11255223e-01 -7.09766567e-01
-1.31005991e+00 -2.70611286e-01 4.05451268e-01 6.49332404e-01
-5.54266810e-01 4.40608174e-01 6.39394373e-02 3.08434129e-01
-3.35105628e-01 -8.29224706e-01 -6.86174572e-01 -9.09606636e-01
-3.15636285e-02 3.06750238e-01 1.78248525e-01 1.59530062e-02
4.34434474e-01 1.17183059e-01 -1.82487416e+00 2.55056053e-01
5.27294934e-01 1.37723935e+00 2.70466566e-01 1.24131046e-01
4.89376001e-02 7.53253758e-01 -2.33990788e-01 -5.19156933e-01
2.58926153e-01 -2.66043216e-01 -6.24729395e-01 -1.40513659e-01
6.11037612e-01 -3.50323766e-01 -6.39426053e-01 1.31670928e+00
7.37946928e-01 1.95310876e-01 6.92065954e-01 -1.18668914e+00
-9.38968480e-01 3.77748758e-01 -4.85570073e-01 -3.10164522e-02
2.48127177e-01 -2.70482928e-01 1.51147664e+00 -9.79672670e-01
3.06064129e-01 1.26074409e+00 7.99719632e-01 7.91915879e-02
-9.65487659e-01 -6.59125388e-01 -4.96316582e-01 6.01324618e-01
-1.94680059e+00 -1.18180834e-01 8.19801867e-01 -1.36115462e-01
4.50309932e-01 -6.25132304e-03 3.54906470e-01 4.36381400e-01
1.61749437e-01 6.81152344e-01 9.48270142e-01 -1.28123373e-01
2.25412883e-02 2.77101904e-01 -2.60815751e-02 1.11421967e+00
1.31795123e-01 -2.35428587e-01 -4.36689883e-01 -3.34555984e-01
4.87263590e-01 3.60562384e-01 -2.54173484e-02 -6.98465943e-01
-1.04173672e+00 1.39631820e+00 8.29139829e-01 1.37895927e-01
-2.83553272e-01 3.44769359e-01 7.26536512e-01 2.07103103e-01
2.50372559e-01 2.13464111e-01 -4.27129753e-02 4.01966155e-01
-9.53065932e-01 3.15710485e-01 3.25539589e-01 8.20806742e-01
1.13575113e+00 -6.17240012e-01 -2.66890168e-01 1.08420062e+00
4.28797483e-01 4.08939958e-01 6.78796113e-01 -7.99790442e-01
3.45314443e-01 7.52523601e-01 -1.60873801e-01 -1.53583169e+00
-4.38366294e-01 -2.31267348e-01 -8.94524872e-01 -3.37169468e-01
9.58701745e-02 7.08754659e-01 -5.29797614e-01 1.70870805e+00
4.13425267e-01 1.77467942e-01 -2.93806549e-02 9.97424185e-01
7.55315363e-01 7.83548832e-01 7.88656473e-02 4.73824292e-02
1.52772188e+00 -9.51528907e-01 -5.04776537e-01 3.21127594e-01
7.05038607e-01 -7.65151501e-01 1.00339293e+00 5.79263037e-03
-1.04487574e+00 -6.08924568e-01 -1.17161620e+00 -4.49353814e-01
-5.96612394e-01 -1.99958980e-02 5.37143469e-01 4.63802040e-01
-1.07497847e+00 6.12856030e-01 -3.68140668e-01 -1.95320398e-01
4.83988345e-01 6.15822732e-01 -3.57491285e-01 -4.84431833e-01
-1.67731631e+00 7.63267279e-01 5.79689562e-01 -1.04734093e-01
-9.25330818e-01 -3.44321817e-01 -1.16865969e+00 3.24334174e-01
-3.70586701e-02 -3.25632721e-01 9.92913127e-01 -5.43571711e-01
-1.05398667e+00 1.05172861e+00 2.10712045e-01 -3.28812659e-01
-1.82260554e-02 1.31071329e-01 -2.49122694e-01 6.18582785e-01
3.81202549e-01 9.60000157e-01 8.51460397e-01 -1.01382899e+00
-7.53313661e-01 -3.87638420e-01 8.55769888e-02 3.88381898e-01
-8.88968349e-01 1.18394196e-02 -2.75923491e-01 -6.07030034e-01
1.54531449e-01 -9.05772269e-01 1.91264264e-02 2.75000066e-01
-3.21762562e-02 -5.01864612e-01 7.16580093e-01 -4.56891805e-01
1.22405314e+00 -2.10610414e+00 7.63292983e-02 2.80781269e-01
-1.22106951e-02 5.87695763e-02 -2.89224118e-01 5.12354910e-01
2.98372537e-01 -2.25657657e-01 -2.55832016e-01 -3.95008355e-01
3.64456445e-01 6.54452071e-02 -2.49238446e-01 8.24880660e-01
1.71434190e-02 8.26313734e-01 -8.84192288e-01 -1.00247252e+00
1.25596121e-01 7.75750995e-01 -5.10104895e-01 3.42733562e-01
2.14188799e-01 -2.18358189e-02 -4.55781668e-01 7.58916020e-01
7.82398999e-01 -6.73714995e-01 -9.61176082e-02 -3.83091271e-01
7.63313621e-02 1.84900641e-01 -9.17760134e-01 2.03264451e+00
-4.52267855e-01 9.94171500e-02 -2.61018008e-01 -1.14634228e+00
9.12562013e-01 2.86854446e-01 6.29866660e-01 -7.52383232e-01
2.26731569e-01 5.13714254e-01 -6.16802156e-01 -2.82359719e-01
5.69655180e-01 -3.11654806e-01 -3.95505190e-01 4.41494614e-01
1.20115288e-01 -1.31416708e-01 1.42729685e-01 2.46729273e-02
8.26051652e-01 -3.85002106e-01 3.48895997e-01 -1.23495400e-01
9.58239019e-01 -3.07153672e-01 3.56809616e-01 5.19914150e-01
-2.20152840e-01 7.37703860e-01 -1.83092114e-02 -5.25831223e-01
-1.14801311e+00 -8.93896639e-01 -4.77855325e-01 1.23277640e+00
7.20415473e-01 -4.24609721e-01 -6.22920990e-01 -6.26900494e-01
1.85266823e-01 -1.36919573e-01 -5.81836760e-01 -4.32034314e-01
-4.11020130e-01 -7.01509714e-01 6.40918791e-01 2.41234034e-01
7.67201185e-01 -1.14860940e+00 -4.93883044e-01 2.04727292e-01
-3.15615505e-01 -8.29042673e-01 -7.96592176e-01 1.46827459e-01
-6.61634922e-01 -9.58954811e-01 -9.88268256e-01 -1.33828044e+00
5.49242258e-01 6.84236228e-01 7.39896238e-01 2.99040526e-01
-5.56152165e-01 2.53270894e-01 -4.02308524e-01 3.07197720e-01
-5.18779792e-02 2.75261849e-01 -1.75455555e-01 2.15991765e-01
4.09151942e-01 -3.26595128e-01 -9.46748853e-01 5.03693104e-01
-1.32578242e+00 -3.89974594e-01 8.21160018e-01 9.66347694e-01
7.33494759e-01 3.35321039e-01 4.99668092e-01 -4.53183532e-01
3.61428827e-01 -6.66370451e-01 -4.74772930e-01 3.96555930e-01
-6.47923172e-01 4.83067691e-01 6.05421066e-01 -3.01190823e-01
-4.74555135e-01 3.75795029e-02 6.40679523e-02 -3.17003995e-01
1.92318186e-01 5.15879810e-01 4.15395126e-02 -4.00417775e-01
8.54759142e-02 5.16321361e-01 -2.05835141e-03 -2.98387647e-01
4.14578885e-01 8.29579949e-01 4.14356947e-01 -5.74033201e-01
6.90300047e-01 6.12345994e-01 1.86710060e-01 -3.74723524e-01
-1.04994643e+00 -7.62289584e-01 -3.74906033e-01 5.00810817e-02
9.94929731e-01 -1.06665862e+00 -8.40992987e-01 5.14527917e-01
-9.89332318e-01 4.07099813e-01 1.77861333e-01 3.11376214e-01
-7.67518580e-01 5.69276810e-01 -8.73462856e-01 -3.47543269e-01
-4.62599546e-01 -1.33712471e+00 1.48555946e+00 2.10329488e-01
1.76214740e-01 -8.65266562e-01 3.65787536e-01 4.76911157e-01
4.54988062e-01 -1.35290861e-01 1.19264054e+00 -4.69950289e-01
-6.97533429e-01 -3.78825486e-01 -6.67231619e-01 2.78766483e-01
3.09786778e-02 -6.47307873e-01 -7.91967154e-01 -6.85735106e-01
-7.43052512e-02 -9.91179943e-01 8.68443072e-01 1.72663406e-01
1.33484662e+00 -4.50237840e-01 -8.37238282e-02 5.50174952e-01
1.69582081e+00 -2.43444610e-02 6.05346262e-01 5.13423622e-01
7.69000590e-01 6.81801498e-01 7.74233818e-01 7.36677051e-01
8.13180327e-01 7.94413090e-01 5.50392032e-01 5.28945662e-02
1.18507728e-01 -3.63135993e-01 1.34749323e-01 9.63953853e-01
6.06007576e-01 8.49125981e-02 -6.08084559e-01 5.71869433e-01
-1.77900231e+00 -8.58091772e-01 4.34855461e-01 2.16103101e+00
1.07596886e+00 -8.63445029e-02 -3.56910489e-02 1.76717058e-01
1.21219373e+00 3.72074097e-01 -4.39301372e-01 -1.21947855e-01
3.40504609e-02 2.77559971e-03 3.94077629e-01 1.94245771e-01
-1.58651304e+00 6.80751026e-01 4.96548891e+00 1.06430089e+00
-1.00027251e+00 1.10652320e-01 4.25145864e-01 2.62512147e-01
-2.53999889e-01 -1.61837399e-01 -8.76824021e-01 5.24867117e-01
8.03163648e-01 7.22338185e-02 3.50483596e-01 8.54927719e-01
-3.04622859e-01 1.41268536e-01 -8.79534125e-01 1.29588568e+00
2.63510853e-01 -1.07215130e+00 2.84908831e-01 -4.40618163e-03
7.05506563e-01 -2.77430624e-01 5.53966284e-01 3.22813988e-01
-2.17177514e-02 -9.54939246e-01 7.65905440e-01 2.01316923e-01
8.65579784e-01 -1.10344899e+00 7.98648298e-01 -1.19001068e-01
-1.55431759e+00 -3.15962672e-01 -9.20089364e-01 2.77456969e-01
-1.10658824e-01 4.06963527e-01 -6.55318737e-01 3.17755222e-01
6.12798691e-01 8.55142117e-01 -6.99051380e-01 1.06507325e+00
7.46576265e-02 -3.45385773e-03 3.37690152e-02 1.03084855e-01
5.37255943e-01 -2.01181602e-02 -1.71328560e-01 1.00872624e+00
4.88671005e-01 -1.17610589e-01 3.50476861e-01 4.19362992e-01
-3.04883152e-01 4.53239232e-01 -7.12845862e-01 -6.33939207e-02
5.10567129e-01 1.46240175e+00 -7.16218412e-01 -2.45770901e-01
-5.71603239e-01 9.58703995e-01 5.63192546e-01 -1.57950714e-01
-9.28306639e-01 -6.50600195e-01 6.09787166e-01 -9.87368971e-02
4.97411102e-01 3.25023793e-02 5.01525365e-02 -9.65772927e-01
-9.24719945e-02 -7.03092337e-01 7.37440407e-01 -4.81778502e-01
-1.39841259e+00 4.42261249e-01 -2.47968912e-01 -1.27287245e+00
4.58912812e-02 -2.76227206e-01 -1.55092329e-01 4.85300571e-01
-1.84469700e+00 -1.29364169e+00 -3.15596223e-01 8.54757428e-01
2.35497028e-01 -1.06039524e-01 9.29913402e-01 7.60189235e-01
-2.65769631e-01 8.32520366e-01 3.99066895e-01 -3.17894630e-02
8.67138267e-01 -1.10018075e+00 -1.53913975e-01 2.26904601e-01
3.32385600e-02 6.22189820e-01 3.56485188e-01 -2.32261211e-01
-1.30747807e+00 -1.19713783e+00 1.05906141e+00 1.12659946e-01
6.31079793e-01 -2.78900325e-01 -1.00861740e+00 1.00323319e-01
-1.30348178e-02 4.31957930e-01 8.90193880e-01 -3.54106337e-01
-8.16284180e-01 -4.41330910e-01 -1.21036839e+00 1.84228882e-01
5.14883280e-01 -9.25826907e-01 -3.50629717e-01 3.44591349e-01
7.25325942e-01 3.86946090e-02 -1.10125482e+00 4.52235341e-01
7.35357821e-01 -7.61166573e-01 1.21763885e+00 -2.17509866e-01
5.45318425e-01 -3.78531814e-01 -5.50426066e-01 -8.11434388e-01
-3.97061437e-01 -1.10026389e-01 1.50629461e-01 1.04017818e+00
-4.16891575e-02 -3.35431904e-01 6.04552269e-01 2.58728027e-01
4.34544198e-02 -8.18821728e-01 -8.90686333e-01 -6.39990628e-01
-1.34452343e-01 3.92009616e-01 5.27703464e-01 8.77255201e-01
-9.16893333e-02 1.77564830e-01 -4.68362898e-01 1.64894313e-01
9.56853032e-01 5.34914374e-01 2.90969253e-01 -1.23038983e+00
-5.67606091e-02 -3.07370603e-01 -8.44985664e-01 -7.52072692e-01
4.46700573e-01 -1.32276142e+00 2.49764264e-01 -1.22117436e+00
7.68353224e-01 -5.82543969e-01 -9.91750419e-01 5.19719779e-01
-1.92534447e-01 7.80227780e-01 6.99796602e-02 4.54359412e-01
-1.36192513e+00 9.94402409e-01 1.09009945e+00 -3.88673782e-01
4.16017830e-01 -3.52131963e-01 -6.19268000e-01 2.93118119e-01
5.60234010e-01 -8.86912584e-01 -4.79097962e-01 -2.32338190e-01
3.29492152e-01 6.31053150e-02 4.27819490e-01 -1.12010992e+00
3.43226284e-01 2.22463146e-01 1.51019633e-01 -5.11182308e-01
3.20933282e-01 -9.10816252e-01 -9.38201025e-02 4.99514222e-01
-7.96667099e-01 2.79417932e-01 -3.72146755e-01 5.68727732e-01
-6.41317189e-01 -4.71115798e-01 1.10277200e+00 -5.89492321e-02
-6.74891233e-01 5.96743047e-01 1.30146772e-01 7.04434663e-02
9.02658880e-01 1.18201546e-01 -1.01239502e-01 -2.30356842e-01
-1.95995748e-01 2.00347036e-01 5.55541635e-01 3.69681597e-01
9.17471766e-01 -1.85940766e+00 -4.78159100e-01 2.93433182e-02
5.04103065e-01 -1.49303138e-01 2.40828484e-01 4.76052374e-01
-5.25458157e-01 7.22652733e-01 -3.79382670e-01 -4.25665379e-01
-1.15021706e+00 7.88641930e-01 -9.49043259e-02 -4.49722379e-01
-4.30332303e-01 8.13549101e-01 2.89393544e-01 -4.06109184e-01
3.89397830e-01 1.08224146e-01 -1.62258163e-01 3.50904346e-01
6.39032781e-01 3.28612745e-01 8.50674957e-02 -7.80050397e-01
-4.27754343e-01 9.30076480e-01 -4.23935980e-01 1.26678228e-01
1.32901967e+00 -3.78231972e-01 -3.99014205e-01 2.31567249e-01
2.17231679e+00 -6.35716438e-01 -1.04994869e+00 -5.40628493e-01
1.92669824e-01 -4.34359133e-01 8.46560001e-02 -2.28372604e-01
-1.04239511e+00 9.65505481e-01 8.52807164e-01 -2.16497518e-02
1.10160959e+00 6.36585206e-02 1.40873015e+00 5.15717149e-01
5.82752407e-01 -1.09875083e+00 5.16189873e-01 3.24993521e-01
5.80707967e-01 -1.54328012e+00 -1.61470950e-01 -5.15792333e-02
-4.19027090e-01 9.34408665e-01 2.99905986e-01 -2.28586733e-01
5.79910874e-01 -4.01926756e-01 -9.57199410e-02 -3.46648097e-01
-3.31785470e-01 -2.00168118e-01 2.96285659e-01 1.46535173e-01
2.69518018e-01 -2.20282480e-01 -4.98758614e-01 2.01969057e-01
1.81162834e-01 -2.80786365e-01 1.27040260e-02 8.74821961e-01
-8.53634536e-01 -1.10043073e+00 -3.21051806e-01 2.39045307e-01
-4.75139052e-01 -2.50265241e-01 2.18894556e-01 3.63097638e-01
4.35295515e-02 6.64105654e-01 -1.74262628e-01 -3.05216044e-01
-7.10462630e-02 -1.02622144e-01 4.01424974e-01 -5.85915387e-01
-3.86138916e-01 -2.12848142e-01 -6.35215819e-01 -5.05326390e-01
-4.52252060e-01 -5.45389593e-01 -1.34600103e+00 -8.73021316e-03
-2.42781684e-01 3.38842839e-01 6.98885322e-01 6.36601627e-01
5.98329231e-02 -9.37906057e-02 1.08752847e+00 -8.29820156e-01
-9.42261219e-01 -5.59812129e-01 -8.92030478e-01 8.09437275e-01
3.83215308e-01 -9.70099449e-01 -3.22352558e-01 -2.20752150e-01] | [11.344804763793945, 0.9494827389717102] |
d393902b-6673-4e9e-893d-baeb7b898643 | joint-learning-of-answer-selection-and-answer | 1911.09801 | null | https://arxiv.org/abs/1911.09801v1 | https://arxiv.org/pdf/1911.09801v1.pdf | Joint Learning of Answer Selection and Answer Summary Generation in Community Question Answering | Community question answering (CQA) gains increasing popularity in both academy and industry recently. However, the redundancy and lengthiness issues of crowdsourced answers limit the performance of answer selection and lead to reading difficulties and misunderstandings for community users. To solve these problems, we tackle the tasks of answer selection and answer summary generation in CQA with a novel joint learning model. Specifically, we design a question-driven pointer-generator network, which exploits the correlation information between question-answer pairs to aid in attending the essential information when generating answer summaries. Meanwhile, we leverage the answer summaries to alleviate noise in original lengthy answers when ranking the relevancy degrees of question-answer pairs. In addition, we construct a new large-scale CQA corpus, WikiHowQA, which contains long answers for answer selection as well as reference summaries for answer summarization. The experimental results show that the joint learning method can effectively address the answer redundancy issue in CQA and achieves state-of-the-art results on both answer selection and text summarization tasks. Furthermore, the proposed model is shown to be of great transferring ability and applicability for resource-poor CQA tasks, which lack of reference answer summaries. | ['Yaliang Li', 'Yuexiang Xie', 'Daoyuan Chen', 'Min Yang', 'Ying Shen', 'Yang Deng', 'Wai Lam'] | 2019-11-22 | null | null | null | null | ['answer-selection'] | ['natural-language-processing'] | [-2.40823608e-02 9.78671536e-02 3.45895886e-02 -7.71759078e-02
-1.53727067e+00 -5.95994413e-01 4.04583961e-01 6.46257460e-01
-3.29072863e-01 9.52785790e-01 8.11652482e-01 -1.77083507e-01
-2.41390780e-01 -8.01258743e-01 -2.08724841e-01 -2.50694126e-01
3.11463356e-01 5.44079900e-01 4.51126605e-01 -6.44237041e-01
7.24386811e-01 -4.17380273e-01 -1.50629807e+00 5.62403321e-01
1.85960627e+00 6.48132920e-01 2.71278799e-01 9.36005771e-01
-7.63402998e-01 1.44683135e+00 -1.23006046e+00 -5.34163058e-01
-3.25280845e-01 -8.05986702e-01 -1.28797340e+00 -1.23777755e-01
5.20832479e-01 -3.46632212e-01 -3.42254996e-01 5.89810252e-01
9.91639078e-01 3.58328551e-01 4.39563692e-01 -1.08809114e+00
-9.12913322e-01 5.10206759e-01 -2.80634552e-01 3.75095993e-01
1.15456462e+00 2.45218098e-01 1.49347246e+00 -8.46477091e-01
3.74895930e-01 1.32177937e+00 3.65659446e-01 5.51679134e-01
-6.07169569e-01 -4.54967588e-01 3.85074690e-02 3.90179634e-01
-9.60069001e-01 -4.03885335e-01 6.55052006e-01 -2.11811379e-01
8.89648318e-01 4.70867932e-01 3.20023656e-01 4.78720933e-01
-6.39646873e-02 1.20682025e+00 6.54212117e-01 -3.58629882e-01
2.33248889e-01 -2.02795669e-01 2.78612882e-01 5.91798365e-01
1.76227719e-01 -9.87509608e-01 -7.14024484e-01 -4.85121161e-01
1.38970539e-01 -1.81611285e-01 -3.60467732e-01 2.66584337e-01
-1.10149753e+00 1.04823411e+00 2.52466112e-01 -2.57189907e-02
-3.65889400e-01 -1.83028892e-01 5.19027829e-01 3.91891479e-01
3.70834798e-01 9.61958170e-01 -1.31547749e-01 -1.13401450e-01
-6.73042357e-01 8.95856977e-01 1.09881485e+00 1.10408223e+00
6.20481908e-01 -3.12956244e-01 -1.22536027e+00 1.00447226e+00
1.36786550e-01 6.55875444e-01 5.77819228e-01 -1.35583842e+00
1.06646407e+00 1.10372961e+00 4.54694360e-01 -1.42914486e+00
-3.08210820e-01 -2.38225520e-01 -9.27782655e-01 -7.94909954e-01
3.34272116e-01 -3.37301999e-01 -3.91817652e-02 1.43225706e+00
2.95628399e-01 -2.18321979e-01 1.82758376e-01 6.61399782e-01
1.66619158e+00 8.32245708e-01 -1.30077034e-01 -4.62475419e-01
1.42799592e+00 -1.42677844e+00 -1.13101673e+00 -1.31457448e-01
7.66695976e-01 -9.06332016e-01 1.30556476e+00 -4.38795537e-02
-1.35161841e+00 -6.46191895e-01 -7.47642577e-01 -5.20679593e-01
1.64470762e-01 1.77273676e-01 2.89187551e-01 1.88263535e-01
-8.35624278e-01 1.58928961e-01 -2.05218587e-02 -1.01228461e-01
4.75282550e-01 1.56889126e-01 2.82238692e-01 -3.92665327e-01
-1.41111839e+00 5.12190580e-01 9.94113553e-03 -2.99622584e-02
-2.54857510e-01 -4.96571124e-01 -7.01242566e-01 9.71535444e-02
5.54137945e-01 -1.14586711e+00 1.61825216e+00 -1.62737146e-01
-1.38859129e+00 4.23964590e-01 -5.36639333e-01 -1.27891019e-01
2.90703297e-01 -2.23625332e-01 -2.88541704e-01 4.18296725e-01
6.43864095e-01 5.17762363e-01 5.30609429e-01 -9.00041401e-01
-8.76754165e-01 -9.81745645e-02 3.44670773e-01 5.39582312e-01
-3.04347157e-01 -1.55199379e-01 -3.18389833e-01 -3.74953270e-01
-1.28619857e-02 -3.12537521e-01 -2.21647903e-01 -6.11792624e-01
-5.41598380e-01 -1.05007827e+00 3.19546402e-01 -8.76543880e-01
1.82760108e+00 -1.55026293e+00 -3.54247205e-02 -2.13281527e-01
6.04398072e-01 3.67032498e-01 -5.67806721e-01 9.67496932e-01
7.65916586e-01 1.48659974e-01 -2.29667723e-01 -1.77964747e-01
-5.10241557e-03 1.31378381e-03 -5.82404435e-01 -1.82473525e-01
1.92087710e-01 1.03831065e+00 -1.50712276e+00 -8.76901507e-01
-4.56664026e-01 -3.10249031e-01 -5.77830613e-01 7.51434565e-01
-5.48065066e-01 4.80081588e-01 -1.04101419e+00 5.78211427e-01
4.02739376e-01 -5.27230859e-01 -2.75651485e-01 2.37472802e-01
2.00128391e-01 6.90494061e-01 -8.83740306e-01 1.63972712e+00
-4.60620761e-01 3.16102594e-01 -1.60418689e-01 -4.49713737e-01
1.10846972e+00 2.66453892e-01 2.05416113e-01 -1.15371037e+00
-1.75153837e-01 5.12864351e-01 -1.00186579e-01 -1.18183529e+00
1.07048118e+00 3.62357795e-01 -2.52422214e-01 9.08057213e-01
-1.67221829e-01 -5.89907467e-01 7.84642100e-01 7.70864785e-01
1.31767297e+00 -2.74284273e-01 3.37874204e-01 -9.76836160e-02
1.17357433e+00 2.11488366e-01 3.22283685e-01 1.05170369e+00
-2.32144386e-01 7.30313420e-01 5.85240424e-01 -3.43711898e-02
-6.35418892e-01 -8.33782554e-01 5.49970150e-01 1.30448008e+00
2.10869640e-01 -6.64007664e-01 -6.98065519e-01 -7.29919612e-01
-6.91795442e-03 6.27104104e-01 -1.56496793e-01 -2.93357342e-01
-8.60611081e-01 -3.10300618e-01 4.09818113e-01 3.78116608e-01
6.69256032e-01 -1.28652537e+00 -2.20719308e-01 3.36088777e-01
-1.08878160e+00 -7.63037682e-01 -9.44734633e-01 -6.65292501e-01
-8.52841616e-01 -1.31379950e+00 -8.43525410e-01 -9.22672331e-01
3.03647071e-01 8.35832179e-01 1.67342460e+00 3.89936328e-01
1.86154068e-01 6.05496049e-01 -6.43733025e-01 -3.21254253e-01
-3.21065515e-01 5.40549040e-01 -3.51563066e-01 -1.87276363e-01
4.38407063e-01 -3.91238153e-01 -1.01395643e+00 2.06864104e-01
-8.69883120e-01 -3.09978008e-01 2.70196885e-01 7.17671454e-01
2.48872489e-01 -3.62361401e-01 1.65062618e+00 -9.60784435e-01
1.75278449e+00 -6.96141064e-01 -2.56331060e-02 6.20529771e-01
-3.81153971e-01 -3.14166886e-03 7.08046079e-01 -1.12039983e-01
-1.10416734e+00 -5.90137899e-01 -1.96441010e-01 4.63289708e-01
4.01943415e-01 5.65115035e-01 -1.93520635e-02 4.33236957e-01
7.30760813e-01 3.00202042e-01 -4.31192964e-02 -2.99893081e-01
5.56148171e-01 9.12328660e-01 5.37895739e-01 -5.19382119e-01
5.04667222e-01 1.62754461e-01 -3.87224346e-01 -6.17688060e-01
-1.40502334e+00 -1.07651126e+00 -1.85626850e-01 -3.11117589e-01
7.21974194e-01 -9.51086521e-01 -8.42224240e-01 2.90175408e-01
-1.62934792e+00 2.49892414e-01 -3.53360265e-01 -5.95999546e-02
-4.39560354e-01 5.90819061e-01 -6.66486442e-01 -8.35839391e-01
-9.11425054e-01 -8.10667515e-01 9.40137029e-01 7.57216394e-01
-4.81325328e-01 -9.42105770e-01 4.61913764e-01 1.03082585e+00
4.06891227e-01 -1.66462466e-01 1.10587072e+00 -8.76222610e-01
-7.44164705e-01 -3.51125151e-01 -1.95107102e-01 2.93201972e-02
2.31844857e-01 -3.94790143e-01 -6.53829873e-01 -7.59331807e-02
1.32404611e-01 -8.15294027e-01 9.16029513e-01 1.77179219e-03
8.93176615e-01 -6.59903288e-01 1.55131742e-01 -3.48625094e-01
7.74420738e-01 -1.96536317e-01 4.17952389e-01 9.17675346e-02
7.96225429e-01 1.01439083e+00 7.32519269e-01 6.81164324e-01
1.11805046e+00 2.54585356e-01 1.43274590e-01 4.34043139e-01
-1.74362019e-01 -5.36364555e-01 1.29977256e-01 1.77741623e+00
3.14528972e-01 -5.57661772e-01 -6.62758291e-01 7.84598231e-01
-2.05828857e+00 -9.27865982e-01 -7.48097301e-01 2.03168273e+00
1.16776216e+00 -2.49061644e-01 2.37482712e-01 -7.63785541e-02
6.40509009e-01 4.48979497e-01 -5.14470935e-01 -1.83945701e-01
-2.10111111e-01 8.04159120e-02 -3.77135187e-01 5.61469555e-01
-6.04043663e-01 6.10090196e-01 5.91440868e+00 1.10121500e+00
-9.91909057e-02 1.94160372e-01 4.60420310e-01 9.78430510e-02
-8.04547966e-01 -1.15568221e-01 -6.39407635e-01 4.54464972e-01
5.90949595e-01 -5.53091586e-01 1.11096554e-01 4.54228044e-01
2.61957496e-01 -3.45889837e-01 -9.00191903e-01 8.08090448e-01
4.18442369e-01 -1.46148479e+00 3.77311528e-01 -5.46607792e-01
1.12067938e+00 -4.58620816e-01 -2.78070152e-01 8.04098487e-01
2.25373119e-01 -8.46000969e-01 2.27769002e-01 7.61267900e-01
2.75963694e-01 -6.88671291e-01 1.06935728e+00 6.67607248e-01
-1.09478819e+00 -2.29854599e-01 -5.20547330e-01 -4.48179662e-01
2.53307194e-01 6.00184858e-01 -6.25651658e-01 6.27213359e-01
2.85313636e-01 4.07635152e-01 -9.74518180e-01 1.30163777e+00
-5.79348445e-01 6.50372028e-01 7.65743554e-02 -6.13048017e-01
2.30031803e-01 -1.27573237e-01 3.63068581e-01 9.07526433e-01
1.72148049e-01 3.54757607e-01 2.78266221e-01 6.80422425e-01
-5.74639738e-01 6.00142717e-01 -2.83213586e-01 2.54260115e-02
9.45704222e-01 1.13845015e+00 -3.32296431e-01 -3.52610737e-01
-2.74927884e-01 7.15242445e-01 5.44554412e-01 4.03543621e-01
-1.93229169e-01 -7.00769067e-01 -4.59195711e-02 -9.58962813e-02
-1.85187496e-02 1.56124189e-01 -2.88320154e-01 -1.14611328e+00
4.62049782e-01 -1.11765265e+00 6.61981881e-01 -7.16175139e-01
-1.59525132e+00 2.95293123e-01 -2.78288811e-01 -1.22995913e+00
-5.16108751e-01 1.67151183e-01 -1.21451497e+00 9.13830876e-01
-1.64505315e+00 -7.42119670e-01 -3.56001318e-01 3.84699613e-01
8.82177830e-01 -1.84455857e-01 5.96293986e-01 2.51655847e-01
-3.70030791e-01 5.24267018e-01 1.44567996e-01 -1.40201265e-03
9.75049019e-01 -1.31364989e+00 2.64949530e-01 5.18590212e-01
-1.68173939e-01 6.95416331e-01 5.43423772e-01 -6.53495848e-01
-1.39404619e+00 -9.47699487e-01 1.61796618e+00 -7.12445676e-01
4.69289124e-01 -7.87141994e-02 -1.14019036e+00 -4.63584810e-02
5.26947558e-01 -5.42696476e-01 8.75254154e-01 1.53998062e-01
-1.23686999e-01 -1.35930553e-01 -8.31003189e-01 6.94766402e-01
6.37999237e-01 -7.74247706e-01 -9.73172188e-01 7.64230430e-01
1.27014303e+00 -2.16252208e-01 -4.65121567e-01 1.01860605e-01
9.53374058e-02 -8.45511079e-01 7.00926900e-01 -5.17943442e-01
7.73228645e-01 -3.63932550e-01 2.28098273e-01 -1.05698621e+00
-1.50147319e-01 -8.08444381e-01 -6.35603011e-01 1.62021625e+00
4.07377720e-01 -2.98063070e-01 5.93784928e-01 5.18367052e-01
-1.88578635e-01 -9.71361279e-01 -7.99785852e-01 -3.25754046e-01
2.35111102e-01 1.07641049e-01 7.16480792e-01 4.86017197e-01
2.76427239e-01 1.10507393e+00 -1.41970709e-01 -4.18062240e-01
2.58545250e-01 4.91841704e-01 8.62855017e-01 -1.07798362e+00
-1.23731472e-01 -5.53220272e-01 4.23634529e-01 -1.53268266e+00
3.64952954e-03 -7.27150381e-01 2.32549518e-01 -2.18981385e+00
3.37026745e-01 -1.32179260e-01 2.29663774e-01 -3.51718128e-01
-1.03289998e+00 -2.36569256e-01 1.55971467e-01 5.77545702e-01
-1.56120121e+00 9.91483092e-01 1.74947131e+00 -2.46396035e-01
-3.56105298e-01 3.46508801e-01 -9.62084651e-01 4.87550944e-01
5.89009941e-01 -3.84953201e-01 -7.51683772e-01 -5.76188326e-01
7.57399440e-01 6.13825142e-01 -6.86136484e-02 -7.06221402e-01
7.56946325e-01 6.31428929e-03 -2.21582904e-01 -9.12285268e-01
-1.85641035e-01 8.54992568e-02 -9.88385320e-01 2.08629206e-01
-8.84163439e-01 3.77832800e-01 -3.55926692e-01 8.55653644e-01
-4.75366384e-01 -6.26080990e-01 -1.86066274e-02 -3.12245369e-01
-1.57850742e-01 2.18534172e-01 -4.37127650e-01 1.04894030e+00
3.12943071e-01 4.76331264e-02 -6.49601996e-01 -9.53512311e-01
3.54195200e-02 1.20706725e+00 -1.32364735e-01 4.23782319e-01
7.41906166e-01 -1.43906605e+00 -1.29154885e+00 -4.54145044e-01
4.12838578e-01 5.14268994e-01 7.18529820e-01 6.45771801e-01
-4.12897855e-01 5.90893388e-01 3.68233711e-01 -3.87159765e-01
-8.62999618e-01 2.34891087e-01 -4.20888662e-02 -7.03017533e-01
-1.85010642e-01 9.25749838e-01 -2.69488811e-01 -7.41626561e-01
2.67902404e-01 -1.62656471e-01 -9.19519067e-01 4.03769523e-01
8.93969595e-01 8.16796422e-01 -1.07428320e-02 -2.11724743e-01
6.82363734e-02 2.36628219e-01 -1.74780861e-01 -3.84906270e-02
7.07938850e-01 -6.03534997e-01 -5.12262285e-01 3.89880359e-01
9.98514593e-01 2.79979795e-01 -6.97050989e-01 -5.69068670e-01
2.99164146e-01 -2.42069602e-01 -5.91828048e-01 -6.62479043e-01
-4.59741145e-01 9.02447581e-01 -1.13570899e-01 4.54122335e-01
8.78870964e-01 -1.12240557e-02 1.45394874e+00 9.82092261e-01
2.92485375e-02 -1.22197223e+00 8.07559013e-01 8.31251800e-01
1.03564072e+00 -1.35593367e+00 5.34103811e-02 -2.27905706e-01
-6.35710418e-01 9.14174318e-01 8.82535875e-01 -4.45987051e-03
-3.36458534e-02 -6.19856179e-01 9.40060019e-02 -2.85021216e-01
-8.82410049e-01 -1.92681044e-01 3.82686436e-01 5.86274087e-01
5.56259274e-01 -1.85334548e-01 -6.93049729e-01 8.82593155e-01
-3.57484132e-01 -2.31343821e-01 7.59993911e-01 8.94793153e-01
-8.45472753e-01 -9.20920908e-01 -2.72439331e-01 6.76377237e-01
-3.09786350e-01 -2.19373941e-01 -5.62776089e-01 1.90833643e-01
-2.95881867e-01 1.84207153e+00 -3.09919417e-01 -2.87019938e-01
4.13239628e-01 6.81684241e-02 1.33855268e-01 -8.59932780e-01
-9.34648395e-01 -3.87187392e-01 3.73526782e-01 -2.18823612e-01
-4.59689558e-01 -5.05668879e-01 -1.31644642e+00 -2.15914741e-01
-4.03515339e-01 7.20518470e-01 -1.17357701e-01 1.21994519e+00
5.81454694e-01 5.58461905e-01 8.30846906e-01 -1.14138842e-01
-9.27053571e-01 -1.28305244e+00 -2.26370126e-01 4.43698108e-01
4.67759520e-01 -2.81513073e-02 -1.72570661e-01 -2.79444307e-01] | [11.657644271850586, 8.254738807678223] |
b2e361e7-a9f8-4075-9c7a-7717b308a871 | the-whole-is-greater-than-the-sum-of-its | null | null | https://aclanthology.org/W18-0522 | https://aclanthology.org/W18-0522.pdf | The Whole is Greater than the Sum of its Parts: Towards the Effectiveness of Voting Ensemble Classifiers for Complex Word Identification | In this paper, we present an effective system using voting ensemble classifiers to detect contextually complex words for non-native English speakers. To make the final decision, we channel a set of eight calibrated classifiers based on lexical, size and vocabulary features and train our model with annotated datasets collected from a mixture of native and non-native speakers. Thereafter, we test our system on three datasets namely News, WikiNews, and Wikipedia and report competitive results with an F1-Score ranging between 0.777 to 0.855 for each of the datasets. Our system outperforms multiple other models and falls within 0.042 to 0.026 percent of the best-performing model{'}s score in the shared task. | ['', 'eep', 'S Mathias', 'Pushpak Bhattacharyya', 'Jayashree Aan Gajjam', 'Nikhil Wani'] | 2018-06-01 | null | null | null | ws-2018-6 | ['complex-word-identification'] | ['natural-language-processing'] | [-2.02792436e-01 -1.38025824e-02 -2.08157703e-01 -7.40710616e-01
-1.06611824e+00 -8.00072789e-01 7.82433093e-01 1.86010584e-01
-9.48989153e-01 8.34043264e-01 2.83112586e-01 -5.58465064e-01
2.30870008e-01 -4.65478987e-01 -3.22063208e-01 -2.06303000e-01
1.32744789e-01 3.42764229e-01 2.68202454e-01 -5.78778028e-01
1.70256317e-01 -1.75406456e-01 -1.48958302e+00 5.31503201e-01
1.23964393e+00 6.84730172e-01 1.96111634e-01 7.58504987e-01
-1.25773638e-01 5.64565241e-01 -7.22326636e-01 -8.79403770e-01
4.72807558e-03 -7.83782303e-02 -7.22683668e-01 -5.16209722e-01
7.99827158e-01 1.05103450e-02 -3.03893853e-02 9.70496655e-01
7.17103124e-01 5.45722730e-02 7.73526251e-01 -8.14802408e-01
-6.28279507e-01 1.14533448e+00 -3.21300000e-01 3.48405391e-01
6.29685640e-01 -1.90356389e-01 1.16356695e+00 -1.27259064e+00
7.56997049e-01 1.21342146e+00 7.29481339e-01 6.65920079e-01
-1.04035246e+00 -1.21024990e+00 4.81730610e-01 9.47591588e-02
-1.46194434e+00 -9.56313193e-01 3.43393058e-01 -3.29766273e-01
1.27384806e+00 3.40205103e-01 2.70019978e-01 1.44002974e+00
-2.06808150e-02 5.84461212e-01 1.57851517e+00 -8.73996377e-01
9.33406875e-02 6.01058066e-01 5.89183509e-01 6.41189814e-01
3.46483916e-01 -2.95648575e-02 -6.91846848e-01 -3.55684429e-01
-9.30437595e-02 -5.94629347e-01 -2.98063695e-01 2.68957496e-01
-1.07867539e+00 1.07208371e+00 1.28805473e-01 3.60395193e-01
3.06421574e-02 -2.77688235e-01 3.77755851e-01 4.95781511e-01
8.96037877e-01 4.36018050e-01 -9.99419153e-01 7.66908471e-03
-7.68942177e-01 2.96888918e-01 1.21472692e+00 8.68614674e-01
2.51747996e-01 -1.84289023e-01 9.95601118e-02 1.27787530e+00
4.55109268e-01 8.01347613e-01 7.82505870e-01 -3.69601667e-01
6.47868037e-01 3.12584192e-01 1.16111830e-01 -6.56820416e-01
-4.58226800e-01 -3.17471623e-01 -5.22446930e-01 -2.04926461e-01
4.30329502e-01 -4.36564296e-01 -1.11987710e+00 1.88776696e+00
1.33628517e-01 -1.26548618e-01 1.86972126e-01 3.06206763e-01
1.30050123e+00 6.41169727e-01 6.26304686e-01 -2.14081377e-01
1.62567651e+00 -8.86205494e-01 -6.74980879e-01 -6.28871143e-01
6.85464561e-01 -9.87164974e-01 1.14178801e+00 2.87502110e-01
-6.35831475e-01 -5.59843242e-01 -1.03613901e+00 1.93817452e-01
-6.84539914e-01 1.85674533e-01 2.46556327e-01 1.09421527e+00
-8.14913988e-01 5.21844998e-02 -2.67597735e-01 -4.28707600e-01
-1.14214821e-02 2.22622126e-01 -4.84963953e-01 5.96196987e-02
-1.59200227e+00 1.28680766e+00 2.96424598e-01 -3.41497779e-01
-5.01298428e-01 -6.02971077e-01 -6.96086168e-01 -3.45853418e-01
-6.27561733e-02 -3.09489399e-01 1.48229659e+00 -9.19699967e-01
-1.34884524e+00 1.13792598e+00 -2.22750485e-01 -1.55336931e-01
5.99038959e-01 -2.47063532e-01 -9.25173283e-01 -4.86234725e-01
2.49974906e-01 4.31982458e-01 5.16370296e-01 -1.02020347e+00
-8.62182796e-01 -2.63719022e-01 1.52895944e-02 9.68532637e-02
-2.45458513e-01 5.50252736e-01 -1.39814258e-01 -7.42725849e-01
-2.61301845e-01 -1.07464349e+00 -2.42984034e-02 -6.98457301e-01
-3.78002077e-01 -5.65120220e-01 3.68071735e-01 -9.34832394e-01
1.38603973e+00 -2.02163506e+00 -7.18639120e-02 2.71159381e-01
7.91104883e-02 3.96776944e-01 -8.71496871e-02 2.13761523e-01
3.17007191e-02 4.41889822e-01 -1.40944570e-01 -3.34877312e-01
-5.52466232e-03 -1.14737570e-01 -1.83743209e-01 2.15897098e-01
-1.45615190e-01 5.33482671e-01 -8.49793077e-01 -3.83386970e-01
7.64242560e-02 3.88745248e-01 -4.06869501e-01 1.15519874e-01
-6.97923601e-02 1.70747206e-01 -1.94916412e-01 3.52083564e-01
4.41928685e-01 1.32074907e-01 5.22323430e-01 1.92684665e-01
-1.76878735e-01 6.63101673e-01 -1.25842690e+00 1.17928994e+00
-8.97354722e-01 7.28334546e-01 -4.08765748e-02 -4.35670644e-01
8.98284018e-01 2.62648523e-01 -2.62362033e-01 -5.86032152e-01
2.06733301e-01 5.79471707e-01 3.31657976e-01 -3.22389722e-01
4.68534678e-01 1.19959608e-01 -5.51226735e-01 3.83717269e-01
1.00479394e-01 1.50509268e-01 1.46557540e-01 1.15033872e-02
8.55545878e-01 -2.53428549e-01 8.76048446e-01 -7.82347739e-01
7.63454378e-01 -2.60361463e-01 3.49888593e-01 9.33536708e-01
-4.43394810e-01 1.55218035e-01 7.38236830e-02 -5.58649838e-01
-8.09796929e-01 -9.21208322e-01 -5.27191281e-01 2.13317585e+00
-5.97047731e-02 -7.01032043e-01 -7.68966138e-01 -7.55805671e-01
8.49733420e-04 1.05881369e+00 -7.40778089e-01 9.09011066e-02
-6.45540297e-01 -6.94154382e-01 7.30229914e-01 4.82677907e-01
5.88141501e-01 -1.09678292e+00 -2.46315107e-01 1.63810804e-01
-4.44795549e-01 -1.16209590e+00 -5.90136945e-01 3.74952435e-01
7.04025337e-03 -8.08761954e-01 -3.27040255e-01 -1.09887493e+00
1.42014831e-01 8.11107010e-02 1.46752727e+00 1.30105644e-01
3.33940834e-02 7.43593574e-02 -5.65383792e-01 -9.69193339e-01
-5.19144535e-01 7.16965139e-01 2.29705811e-01 -3.16469550e-01
8.24956715e-01 -1.43071674e-02 -1.90123379e-01 3.15885007e-01
-1.81173190e-01 -1.99916705e-01 3.33639324e-01 7.83947587e-01
9.84481722e-02 -4.79006618e-01 6.06646657e-01 -1.16803598e+00
7.95291901e-01 -5.34794867e-01 -2.40511701e-01 4.33951914e-01
-6.52308285e-01 7.43115842e-02 4.59471434e-01 -7.46547341e-01
-1.28255403e+00 2.19356939e-01 -1.67684302e-01 5.21323562e-01
-1.42279863e-01 4.27819401e-01 -2.63038337e-01 2.76728183e-01
1.06460404e+00 -2.99919397e-02 -5.14946103e-01 -4.26638812e-01
6.21502161e-01 1.32372832e+00 3.04242402e-01 -5.19705772e-01
5.47639191e-01 -2.21457839e-01 -1.06455231e+00 -9.78375137e-01
-1.10717487e+00 -5.03598511e-01 -5.31144440e-01 -5.22746146e-02
7.45807290e-01 -1.20247591e+00 -2.52344489e-01 5.67818224e-01
-1.10173702e+00 -1.71657383e-01 3.30422908e-01 4.20559824e-01
1.05836391e-01 -3.63922454e-02 -4.07009065e-01 -7.51234353e-01
-6.89442396e-01 -9.47398961e-01 9.48958814e-01 -4.31083189e-03
-6.97255671e-01 -1.01353872e+00 3.99440944e-01 5.00280678e-01
6.97059810e-01 -6.67120442e-02 5.91746807e-01 -1.31691396e+00
2.38973141e-01 -1.43787071e-01 5.16412873e-03 2.44941890e-01
-1.60243735e-01 1.94317549e-01 -1.18891716e+00 -2.16374189e-01
-4.51492578e-01 -5.33445537e-01 1.08469653e+00 2.64404207e-01
8.29987943e-01 -1.84866369e-01 -5.40826380e-01 2.54629403e-01
1.01335323e+00 9.57059562e-02 2.72132486e-01 4.32471782e-01
2.84868777e-01 5.04280627e-01 2.12180406e-01 1.89030766e-01
6.77405715e-01 5.30364990e-01 -1.86911255e-01 3.40266705e-01
-2.62634712e-03 -2.89224863e-01 5.61938524e-01 9.16161478e-01
-6.56227395e-02 -3.20647657e-01 -1.15559065e+00 5.16294241e-01
-1.53853536e+00 -1.01893830e+00 -5.96721768e-02 2.02367187e+00
1.24738491e+00 3.19595158e-01 2.29196638e-01 -8.30889866e-02
8.91974211e-01 2.97166049e-01 -2.41676897e-01 -7.93344259e-01
-2.27757186e-01 6.21647656e-01 4.79868591e-01 8.25303257e-01
-1.56120455e+00 1.22791529e+00 6.84258366e+00 6.36078179e-01
-1.05251682e+00 2.52334028e-01 6.31693900e-01 7.43463784e-02
-2.69397199e-01 -2.55600989e-01 -1.23741591e+00 4.98535246e-01
1.45453596e+00 -4.22269583e-01 2.82644987e-01 1.01889241e+00
-4.89928782e-01 4.03784066e-02 -7.80529380e-01 8.67561162e-01
4.05135989e-01 -6.96758807e-01 -3.72422457e-01 -2.96182688e-02
9.16262209e-01 4.69429642e-01 3.16417739e-02 8.98472309e-01
7.31321692e-01 -1.31380224e+00 8.60603809e-01 5.42166531e-02
8.08415473e-01 -6.64003968e-01 8.27563822e-01 5.72043061e-01
-1.05713356e+00 -1.80699620e-02 -1.02631614e-01 -1.99177369e-01
-3.67860466e-01 4.21905607e-01 -7.91464150e-01 -1.97566181e-01
9.00961161e-01 1.14865884e-01 -6.95847452e-01 5.91377437e-01
-4.29105133e-01 8.57062399e-01 -4.40748274e-01 -5.54610312e-01
-2.30259709e-02 4.14895833e-01 3.63901332e-02 1.74145901e+00
6.47754893e-02 1.91571161e-01 2.47394025e-01 -2.19225660e-02
-5.19817114e-01 7.65035093e-01 -4.95590210e-01 4.87679601e-01
9.17241216e-01 1.33880568e+00 -2.72943676e-01 -4.75975096e-01
-5.83258927e-01 7.19357729e-01 7.30840445e-01 5.86637725e-05
-7.30543554e-01 -6.62737250e-01 6.13268554e-01 -1.72273949e-01
2.64266104e-01 8.39501619e-02 -2.58500338e-01 -1.22778773e+00
-9.07320902e-03 -1.08949459e+00 4.32028949e-01 -2.68766224e-01
-1.60692775e+00 1.02378631e+00 -2.03654543e-02 -6.34727299e-01
-3.91602874e-01 -9.56209719e-01 -4.57398951e-01 8.94379497e-01
-1.32595897e+00 -1.06678450e+00 -4.02972102e-01 3.43528837e-01
5.25274575e-01 -4.54423010e-01 1.13082695e+00 1.77187055e-01
-6.12967670e-01 9.55163956e-01 1.70281425e-01 4.91566390e-01
1.25120389e+00 -1.37917590e+00 6.96784019e-01 5.38859010e-01
2.67838895e-01 7.89938748e-01 6.78526521e-01 -4.63070542e-01
-5.60415983e-01 -1.04527044e+00 1.54643214e+00 -9.28390622e-01
6.12019777e-01 -5.02376199e-01 -6.98930264e-01 8.16297829e-01
5.46214759e-01 6.59264103e-02 1.09233963e+00 6.70266449e-01
-9.90794957e-01 -5.21007413e-03 -1.12854838e+00 3.83183897e-01
1.10025823e+00 -6.18983090e-01 -1.07014585e+00 3.93762171e-01
3.88713002e-01 -4.55128163e-01 -8.51274610e-01 2.89335579e-01
9.55726147e-01 -7.51225889e-01 7.03507960e-01 -8.04972649e-01
9.65962559e-02 1.13365121e-01 -4.67880040e-01 -1.73618603e+00
-3.26843411e-01 -2.13065952e-01 4.38115776e-01 1.12852752e+00
1.09431446e+00 -6.94260895e-01 1.67336032e-01 5.41087449e-01
8.52751881e-02 -3.77541035e-01 -1.06721389e+00 -5.03330946e-01
5.92678547e-01 -7.18818009e-01 4.19468015e-01 1.17223787e+00
2.18770757e-01 8.04102957e-01 -1.91667587e-01 -2.98829693e-02
4.55621719e-01 -1.13017626e-01 6.18936718e-01 -1.50249052e+00
1.25045002e-01 -4.30702627e-01 -2.49777101e-02 -6.29857481e-01
7.64690042e-01 -1.09420967e+00 8.41125771e-02 -1.06429684e+00
4.65019494e-01 -4.41982865e-01 -3.58960569e-01 4.72820103e-01
-5.46135187e-01 2.87431329e-01 -4.88521904e-02 4.66868980e-03
-5.65018713e-01 -2.89004855e-02 6.10767007e-01 -2.11190566e-01
-2.47823164e-01 -6.36579171e-02 -9.88387048e-01 8.71452808e-01
1.08449495e+00 -3.37549448e-01 -1.02501713e-01 -4.54448253e-01
2.03534082e-01 -4.03973520e-01 -2.09013879e-01 -6.26178324e-01
6.98846206e-02 -1.17626034e-01 5.46517670e-01 -2.45857745e-01
9.17719517e-05 -3.27649921e-01 -2.41538748e-01 4.53812838e-01
-7.13455796e-01 2.42717594e-01 1.56176344e-01 2.16778845e-01
4.72286604e-02 -1.58499286e-01 8.30813646e-01 -8.26754570e-02
-8.39122474e-01 -1.98895097e-01 -5.02277017e-01 4.85072494e-01
9.38200057e-01 2.60329098e-01 -5.14416873e-01 -2.17249110e-01
-7.64631152e-01 -3.56220976e-02 7.14111999e-02 9.18216765e-01
2.15158150e-01 -1.10659850e+00 -1.18144059e+00 2.29529664e-01
6.62298858e-01 -7.10546017e-01 -1.06791042e-01 2.16537625e-01
-4.73364770e-01 5.68011165e-01 1.04625367e-01 -2.92996079e-01
-1.49204481e+00 5.05933762e-02 4.29899395e-01 -2.53195345e-01
-2.06382759e-02 1.06511617e+00 -2.75348723e-01 -1.15353966e+00
1.98865205e-01 -2.56508172e-01 -3.21114063e-01 2.33607143e-01
6.99259818e-01 3.58070016e-01 2.30293304e-01 -1.03339672e+00
-8.23086202e-01 3.45570475e-01 -4.41632569e-01 -2.86900759e-01
9.50729191e-01 9.05938298e-02 9.43659618e-02 3.88844222e-01
1.12034988e+00 5.98218441e-01 -3.76534402e-01 -4.48564678e-01
2.30794400e-01 -6.61595762e-02 1.76173732e-01 -1.12038958e+00
-6.69555068e-01 3.36776435e-01 6.68974459e-01 7.38405809e-02
5.15238285e-01 1.74497411e-01 5.12312293e-01 5.38875818e-01
3.05139273e-01 -1.25249505e+00 -3.88617933e-01 7.70616651e-01
8.19885731e-01 -1.62076354e+00 2.66531575e-02 -4.11726713e-01
-6.81887925e-01 8.40855122e-01 7.18390763e-01 4.23647836e-02
1.11485255e+00 2.31346995e-01 5.75568140e-01 2.03793436e-01
-1.12622535e+00 -1.28787130e-01 5.86822867e-01 4.77734059e-01
1.06299007e+00 7.44445682e-01 -7.07515001e-01 8.12405467e-01
-7.94767857e-01 -4.82320994e-01 1.60543188e-01 7.21119404e-01
-7.35954762e-01 -1.13839626e+00 -3.35597903e-01 5.71098804e-01
-5.09870172e-01 -3.61338794e-01 -5.62166989e-01 1.04818666e+00
2.58695900e-01 1.33774364e+00 -9.90406889e-03 -5.36274076e-01
5.65389276e-01 5.33328772e-01 2.93094128e-01 -6.47186220e-01
-9.78709161e-01 -1.22360677e-01 6.52517021e-01 -1.85991585e-01
-4.10179585e-01 -8.33784282e-01 -7.61938632e-01 -3.56871009e-01
-4.21757013e-01 5.69407009e-02 7.07998157e-01 9.25825596e-01
-5.17076068e-03 -5.81451394e-02 5.79222381e-01 -3.56674641e-01
-9.26879048e-01 -1.49066675e+00 -2.53531188e-01 5.08072197e-01
1.87738732e-01 -5.00247240e-01 -6.12543404e-01 -1.68073669e-01] | [10.47040843963623, 10.494050979614258] |
f9520bd7-742f-4855-84ba-a848b534ff01 | atlas-a-dataset-and-benchmark-for-e-commerce | 1908.08984 | null | https://arxiv.org/abs/1908.08984v1 | https://arxiv.org/pdf/1908.08984v1.pdf | Atlas: A Dataset and Benchmark for E-commerce Clothing Product Categorization | In E-commerce, it is a common practice to organize the product catalog using product taxonomy. This enables the buyer to easily locate the item they are looking for and also to explore various items available under a category. Product taxonomy is a tree structure with 3 or more levels of depth and several leaf nodes. Product categorization is a large scale classification task that assigns a category path to a particular product. Research in this area is restricted by the unavailability of good real-world datasets and the variations in taxonomy due to the absence of a standard across the different e-commerce stores. In this paper, we introduce a high-quality product taxonomy dataset focusing on clothing products which contain 186,150 images under clothing category with 3 levels and 52 leaf nodes in the taxonomy. We explain the methodology used to collect and label this dataset. Further, we establish the benchmark by comparing image classification and Attention based Sequence models for predicting the category path. Our benchmark model reaches a micro f-score of 0.92 on the test set. The dataset, code and pre-trained models are publicly available at \url{https://github.com/vumaasha/atlas}. We invite the community to improve upon these baselines. | ['Venkatesh Umaashankar', 'Girish Shanmugam S', 'Aditi Prakash'] | 2019-08-12 | null | null | null | null | ['product-categorization'] | ['miscellaneous'] | [-9.90171134e-02 -4.24689829e-01 -7.19824612e-01 -6.90196633e-01
-1.95548698e-01 -1.09107304e+00 3.03632587e-01 4.08645153e-01
-8.66633728e-02 2.60301288e-02 3.16116214e-01 -2.69165218e-01
4.11255434e-02 -7.96778917e-01 -4.42129344e-01 -3.77485484e-01
-1.49467021e-01 3.58552754e-01 1.17576100e-01 -1.51896462e-01
3.92468691e-01 1.13520980e-01 -1.63446379e+00 6.94301844e-01
4.45307374e-01 1.50820041e+00 4.05730277e-01 2.82768577e-01
-3.05672050e-01 3.55129063e-01 -3.02714944e-01 -1.01781392e+00
3.61010820e-01 -3.22811991e-01 -9.93930161e-01 4.00663346e-01
6.53041542e-01 -1.74872413e-01 2.62811743e-02 1.36319518e+00
1.99938845e-02 -9.74571705e-02 6.63761079e-01 -1.36874413e+00
-1.18076348e+00 6.38311446e-01 -6.65920913e-01 2.25438327e-01
3.96485507e-01 -5.28306998e-02 1.43283451e+00 -8.48141134e-01
9.53475416e-01 1.07800388e+00 5.59274793e-01 2.54328400e-01
-1.09052098e+00 -6.35033906e-01 3.76414955e-01 5.93699217e-01
-1.47481704e+00 4.87139672e-02 7.56735206e-01 -6.66096270e-01
7.90286958e-01 8.33949223e-02 6.30245745e-01 1.19428229e+00
3.26331317e-01 6.39945745e-01 1.09431708e+00 -3.72417897e-01
2.97831059e-01 4.83765692e-01 8.90626669e-01 5.64383447e-01
1.15213551e-01 1.72710326e-02 -3.01321357e-01 1.95327610e-01
6.37066364e-01 1.13447860e-01 1.86912835e-01 -7.51776755e-01
-1.01982379e+00 1.06218171e+00 9.00517464e-01 3.44121397e-01
-3.29975665e-01 -2.56422251e-01 7.56634653e-01 4.54732120e-01
3.02392602e-01 2.43372709e-01 -5.65033913e-01 2.29169637e-01
-5.12486696e-01 2.75763065e-01 8.05487812e-01 1.28293324e+00
4.52523381e-01 -4.56147879e-01 9.28711593e-02 9.70599890e-01
3.94275904e-01 -1.37835756e-01 6.06644213e-01 -7.99245715e-01
1.74987286e-01 7.11578965e-01 6.33134916e-02 -1.10300314e+00
-3.27733129e-01 -5.71779132e-01 -6.95159078e-01 1.22625738e-01
3.01368684e-01 3.92773211e-01 -9.64614272e-01 1.38936388e+00
3.40067506e-01 -1.67379916e-01 -3.43978107e-01 1.15757442e+00
1.08038449e+00 4.95123595e-01 3.75477016e-01 1.67825911e-02
1.97895634e+00 -1.28912222e+00 -5.03811896e-01 3.10029071e-02
5.28894246e-01 -8.83759737e-01 1.33921027e+00 6.31703615e-01
-5.99850059e-01 -8.76944005e-01 -1.14801121e+00 -1.02195866e-01
-9.56489146e-01 -1.25593334e-01 1.10409462e+00 8.39604080e-01
-7.75649965e-01 3.34750801e-01 -3.55776995e-01 -7.92560697e-01
5.53483546e-01 9.04757231e-02 -3.16262215e-01 -3.52578878e-01
-1.14540410e+00 7.44225681e-01 4.18495655e-01 -9.76391658e-02
-8.75172019e-01 -6.65695965e-01 -8.44905913e-01 -9.80151370e-02
1.34517238e-01 -4.83545393e-01 1.28826141e+00 -1.22493732e+00
-9.87479925e-01 1.18595743e+00 1.25021026e-01 -3.77557725e-01
-1.65033713e-03 1.16559364e-01 -8.33652377e-01 -2.52702851e-02
3.70870292e-01 1.04146147e+00 5.16276062e-01 -1.35411656e+00
-1.04448557e+00 -5.41904449e-01 3.64527255e-01 9.80473086e-02
-2.70299554e-01 2.96802491e-01 -4.65326786e-01 -7.91207969e-01
2.07959518e-01 -9.06434655e-01 -5.04601225e-02 -1.90611407e-01
-2.30346322e-01 -4.45182472e-01 5.01821756e-01 -5.73077202e-01
1.15997267e+00 -2.18374515e+00 -2.22931087e-01 -8.10641795e-02
9.30936635e-03 -2.97575980e-01 -2.50151545e-01 4.27326441e-01
-1.88828334e-01 2.63056219e-01 8.24652463e-02 -1.16284981e-01
2.71808296e-01 3.68665829e-02 -3.87068450e-01 3.78775686e-01
-3.39643508e-01 1.11956048e+00 -7.42772043e-01 -4.26724344e-01
2.56446693e-02 2.42131948e-01 -4.78800654e-01 -1.25163317e-01
-3.08121979e-01 1.44616678e-01 -1.89521581e-01 1.10864937e+00
7.06334770e-01 -4.39192772e-01 2.28834257e-01 -4.56905872e-01
9.97285396e-02 4.05613303e-01 -1.08796191e+00 1.91778517e+00
-3.67039770e-01 2.93861419e-01 -3.03736068e-02 -9.03206587e-01
8.09654653e-01 5.77439144e-02 4.00192946e-01 -7.35497117e-01
2.39973500e-01 1.45379499e-01 8.88178945e-02 -2.65562892e-01
7.33393610e-01 4.72805725e-04 -4.03000534e-01 4.28028673e-01
2.59129107e-01 1.84242755e-01 5.33339143e-01 1.33126274e-01
7.43811190e-01 1.25305757e-01 2.67866641e-01 -3.75895441e-01
1.83562115e-01 4.03863758e-01 2.87865311e-01 4.81334627e-01
-3.87733757e-01 4.04460549e-01 1.67768851e-01 -8.05801690e-01
-1.10201240e+00 -1.16482985e+00 -5.42499185e-01 1.38740706e+00
5.13043284e-01 -4.44318622e-01 -4.20987695e-01 -9.16349053e-01
2.24272534e-01 6.00774586e-01 -8.33207428e-01 1.03245631e-01
-1.22306824e-01 -4.63036895e-01 -1.28447637e-01 5.11584342e-01
6.39805794e-01 -1.40150082e+00 -3.22983682e-01 9.38258544e-02
-2.21520305e-01 -8.57118368e-01 -7.37017989e-01 8.45997930e-02
-7.96192825e-01 -8.82776856e-01 -3.95504892e-01 -1.24438596e+00
5.40572941e-01 4.92560178e-01 1.47497690e+00 -5.04806116e-02
-3.41611832e-01 2.62183279e-01 -7.62151480e-01 -2.42843121e-01
-1.54597551e-01 3.59405279e-01 -4.25626896e-02 6.86228788e-03
1.06264973e+00 -2.67298400e-01 -7.21554458e-01 7.53014624e-01
-4.95092392e-01 -1.11242086e-01 3.91912997e-01 6.49025202e-01
7.15928733e-01 3.84564310e-01 5.42397261e-01 -7.04005063e-01
4.33500946e-01 -7.82173753e-01 -5.82258165e-01 1.50910050e-01
-7.87343323e-01 -5.45287251e-01 2.47178823e-01 -4.23553437e-01
-7.82300770e-01 1.35297418e-01 -1.16800860e-01 -5.45358397e-02
-5.64877927e-01 6.11597359e-01 -1.52081370e-01 1.80175975e-01
3.26703995e-01 2.56234384e-03 -3.54010016e-01 -8.13813269e-01
5.97843707e-01 8.40123475e-01 4.18734998e-01 -2.62798399e-01
3.39055002e-01 1.97113946e-01 -4.43756908e-01 -4.18967992e-01
-9.89703178e-01 -8.01404536e-01 -1.02268577e+00 -1.72364086e-01
8.10785115e-01 -8.51583421e-01 -8.15365016e-01 1.79141700e-01
-8.52483809e-01 -6.68612793e-02 -6.19909130e-02 3.06909621e-01
-3.17390114e-01 2.34135553e-01 -7.80935526e-01 -3.32857847e-01
-3.98708373e-01 -9.23872530e-01 7.57933617e-01 8.89109075e-02
-5.52302063e-01 -8.74110162e-01 -2.37576365e-01 5.51965773e-01
3.10205489e-01 -5.58040105e-02 9.98284101e-01 -5.81404388e-01
-5.68503439e-01 -7.30300918e-02 -2.49688968e-01 1.36435613e-01
1.88312605e-01 -3.37295860e-01 -6.89698398e-01 -3.41640085e-01
-8.98525044e-02 -3.79954636e-01 9.54867542e-01 4.69813555e-01
1.16309273e+00 -2.82183349e-01 -4.16459769e-01 3.68016958e-01
1.65262723e+00 4.06710893e-01 4.70846742e-01 6.66336596e-01
5.72570682e-01 9.50908661e-01 9.00023341e-01 1.32226527e-01
5.59014678e-01 9.92889166e-01 5.23701012e-01 1.21240355e-01
-1.20772369e-01 -3.58195931e-01 9.80347246e-02 6.70324564e-01
3.63923430e-01 -1.94428027e-01 -5.95787406e-01 8.77756417e-01
-1.57697845e+00 -9.08091664e-01 -2.48541310e-01 2.04266286e+00
7.06816435e-01 1.90162301e-01 5.24324358e-01 1.84767190e-02
6.88149154e-01 -2.31004044e-01 -4.62925792e-01 -6.87203467e-01
1.24084666e-01 -1.77478597e-01 4.62353408e-01 2.54412770e-01
-1.49288905e+00 9.50810254e-01 6.13676262e+00 8.42182338e-01
-8.73994946e-01 2.92346478e-01 7.34580159e-01 8.31345469e-02
-9.43081528e-02 -3.32198828e-01 -1.08421934e+00 6.90327525e-01
8.39110851e-01 -1.21126428e-01 2.70553946e-01 1.14656186e+00
-2.25527242e-01 1.79750789e-02 -1.13339138e+00 9.40694809e-01
4.34218384e-02 -1.15537071e+00 2.13378519e-02 2.42527872e-01
5.73255897e-01 -7.97649398e-02 3.68794531e-01 4.25555885e-01
5.92297316e-01 -9.59435284e-01 9.68394876e-01 -2.79219419e-01
8.23551774e-01 -5.86124063e-01 7.34477758e-01 -2.32157577e-02
-1.50704837e+00 -3.85462880e-01 -4.94732499e-01 -4.99290265e-02
2.35694900e-01 1.68014601e-01 -5.94962835e-01 4.94334131e-01
1.18614054e+00 9.72877920e-01 -7.09768713e-01 1.13961625e+00
-8.26387554e-02 5.52768409e-01 -4.37410362e-02 -9.09259915e-02
3.00800353e-01 -3.98820400e-01 -3.56490016e-02 1.30447185e+00
2.16169149e-01 1.44213522e-02 3.61035287e-01 6.18528366e-01
-4.22249734e-02 2.95380801e-01 -4.53692794e-01 7.13275671e-02
3.95735651e-01 1.54935098e+00 -1.26792955e+00 -1.62363559e-01
-7.17205405e-01 1.11038315e+00 1.57857109e-02 1.87717885e-01
-8.64388347e-01 -3.22686940e-01 9.24048364e-01 1.38438478e-01
6.74308836e-01 -1.71207741e-01 -5.81909597e-01 -8.35093975e-01
-5.84170781e-02 -8.03572536e-01 8.45641851e-01 -7.19362497e-01
-1.75420511e+00 6.52469456e-01 -1.75753087e-02 -1.26927626e+00
1.56433344e-01 -7.87847042e-01 -7.63391983e-03 6.22752011e-01
-1.24972284e+00 -1.45817244e+00 -2.95797437e-01 3.84457976e-01
9.28030849e-01 -2.15605870e-01 8.42309535e-01 5.44144928e-01
-3.21356535e-01 6.75192177e-01 -6.72886521e-02 2.72874385e-01
4.81878549e-01 -1.13715219e+00 6.22036338e-01 4.38997418e-01
4.49764013e-01 8.27458560e-01 6.00522161e-01 -6.34862006e-01
-1.16145873e+00 -1.02165639e+00 1.05318725e+00 -7.36987472e-01
8.93263876e-01 -7.67582357e-01 -8.24295223e-01 9.19124067e-01
4.55106944e-01 -3.69861394e-01 7.98915684e-01 3.87500167e-01
-7.44997144e-01 -2.44716898e-01 -1.30092072e+00 3.43597859e-01
1.24420261e+00 -2.31860712e-01 -4.99162287e-01 3.93550277e-01
8.70292723e-01 -6.60911798e-02 -1.16947663e+00 6.09998628e-02
8.33067894e-01 -7.47557938e-01 1.04996741e+00 -4.94149119e-01
6.79108799e-01 -1.82809368e-01 -3.63574654e-01 -1.13172567e+00
-7.77101874e-01 3.95153277e-02 9.96742845e-02 1.46950042e+00
5.09670436e-01 -4.38667178e-01 9.44574594e-01 3.32474917e-01
-6.35729507e-02 -6.63229585e-01 -5.63379824e-01 -9.33564961e-01
1.12199031e-01 -4.10271615e-01 8.14076781e-01 9.98132646e-01
1.44024014e-01 4.83676106e-01 -2.16312171e-03 1.02350526e-01
7.20815599e-01 7.35510290e-01 3.38206828e-01 -1.19316995e+00
-1.05052635e-01 -6.66788399e-01 -4.07104552e-01 -1.20873809e+00
-4.75197919e-02 -1.21625614e+00 -2.01961115e-01 -1.49470830e+00
4.85329092e-01 -6.24200165e-01 -3.69380206e-01 4.83118594e-01
3.84950250e-01 7.13535666e-01 1.66650951e-01 4.10712570e-01
-7.55580842e-01 2.86288979e-03 1.08918583e+00 -4.65335727e-01
-1.73358917e-02 -7.91461486e-03 -1.02681279e+00 3.21620762e-01
9.46984112e-01 -3.55095267e-01 -4.36136901e-01 -2.98117846e-01
1.34462506e-01 -5.72789252e-01 1.81462228e-01 -5.73632061e-01
3.01671177e-02 8.54225904e-02 4.46796387e-01 -8.35665703e-01
1.02705106e-01 -9.84205425e-01 3.52713585e-01 5.13714790e-01
-4.01613861e-01 1.64731994e-01 2.01850794e-02 3.74949634e-01
-2.98917055e-01 -4.73071843e-01 6.36028767e-01 -3.87226671e-01
-1.53788114e+00 3.70851547e-01 -4.40381840e-02 -4.69440877e-01
1.03396428e+00 -3.78654748e-01 -2.94609994e-01 -2.04058841e-01
-8.41203868e-01 1.98223129e-01 6.77274108e-01 1.10669816e+00
3.57352227e-01 -1.56132257e+00 -4.54657584e-01 1.30880207e-01
5.62166810e-01 -7.24073708e-01 2.60779530e-01 3.35361719e-01
-4.24136788e-01 7.72745788e-01 -5.64892232e-01 -5.52543163e-01
-1.38813221e+00 1.35293305e+00 6.94836751e-02 -1.35170981e-01
-4.83611763e-01 8.85373592e-01 2.86229968e-01 -4.72519636e-01
3.10301811e-01 -2.06441954e-01 -6.50564492e-01 2.44944662e-01
6.36005282e-01 5.71674742e-02 1.83685690e-01 -8.65069747e-01
-3.30354333e-01 7.16883242e-01 -3.99845332e-01 1.95270047e-01
1.04376030e+00 -5.37840962e-01 1.36597482e-02 3.80811125e-01
1.41624725e+00 -5.73916733e-01 -9.48867142e-01 -1.61985099e-01
2.34336406e-01 -6.41061902e-01 -6.46225438e-02 -1.14898777e+00
-1.14203954e+00 5.41146994e-01 9.75798070e-01 6.55067265e-01
1.19313502e+00 2.59814829e-01 7.36884296e-01 -2.11215869e-01
8.63384724e-01 -8.89334142e-01 2.93546282e-02 2.89070576e-01
1.01802206e+00 -1.40766776e+00 -6.13856949e-02 -8.22969437e-01
-9.02890265e-01 6.59304559e-01 5.62744677e-01 8.80554616e-02
8.43939245e-01 -1.90232638e-02 3.42333168e-01 -3.95108104e-01
-5.68868041e-01 -2.16466278e-01 1.83753669e-01 7.34972835e-01
5.48571944e-01 2.24020556e-01 -5.60285628e-01 8.64574611e-01
-4.93936300e-01 1.11231375e-02 2.98224837e-01 6.79733813e-01
-2.21609756e-01 -1.22368598e+00 -9.77922082e-02 4.22721773e-01
-5.14572799e-01 2.43797190e-02 -1.54257908e-01 6.02856636e-01
2.76174009e-01 1.19220793e+00 3.17322940e-01 -4.10196304e-01
2.72983670e-01 -1.06787749e-01 4.41700280e-01 -7.29970455e-01
-7.45446205e-01 -2.58023143e-02 7.79542699e-02 -4.94552195e-01
-3.25045258e-01 -7.50622749e-01 -8.54189456e-01 -5.09173632e-01
-5.72029762e-02 -3.52011919e-02 7.98759758e-01 2.48977095e-01
3.81256014e-01 2.69566979e-02 6.41757786e-01 -6.62276566e-01
-3.06562871e-01 -9.96173561e-01 -9.60504949e-01 7.99853206e-01
-1.01770520e-01 -8.54692042e-01 -4.33231145e-02 4.65098500e-01] | [9.906390190124512, 6.143749713897705] |
755048a8-90b6-4bff-92fc-3d75f1f04f27 | focusedcleaner-sanitizing-poisoned-graphs-for | 2210.13815 | null | https://arxiv.org/abs/2210.13815v1 | https://arxiv.org/pdf/2210.13815v1.pdf | FocusedCleaner: Sanitizing Poisoned Graphs for Robust GNN-based Node Classification | Recently, a lot of research attention has been devoted to exploring Web security, a most representative topic is the adversarial robustness of graph mining algorithms. Especially, a widely deployed adversarial attacks formulation is the graph manipulation attacks by modifying the relational data to mislead the Graph Neural Networks' (GNNs) predictions. Naturally, an intrinsic question one would ask is whether we can accurately identify the manipulations over graphs - we term this problem as poisoned graph sanitation. In this paper, we present FocusedCleaner, a poisoned graph sanitation framework consisting of two modules: bi-level structural learning and victim node detection. In particular, the structural learning module will reserve the attack process to steadily sanitize the graph while the detection module provides the "focus" - a narrowed and more accurate search region - to structural learning. These two modules will operate in iterations and reinforce each other to sanitize a poisoned graph step by step. Extensive experiments demonstrate that FocusedCleaner outperforms the state-of-the-art baselines both on poisoned graph sanitation and improving robustness. | ['Kai Zhou', 'Liang Tong', 'Yulin Zhu'] | 2022-10-25 | null | null | null | null | ['graph-mining'] | ['graphs'] | [ 2.87981510e-01 3.39770913e-01 -3.57749939e-01 3.52087468e-01
-4.82996225e-01 -8.90062809e-01 4.63064790e-01 4.60391968e-01
-9.56344977e-02 3.96752298e-01 -1.37857944e-01 -6.56124175e-01
-1.97252795e-01 -1.16648781e+00 -9.60245728e-01 -8.28093231e-01
-3.82760823e-01 3.21530521e-01 5.71111381e-01 -3.50676030e-01
2.78724998e-01 5.97072959e-01 -7.50776470e-01 -4.00543809e-02
8.18905771e-01 4.03392345e-01 -3.74098510e-01 6.62595689e-01
2.64575154e-01 1.07465422e+00 -6.90271914e-01 -9.62918758e-01
3.73708814e-01 -1.91891238e-01 -7.24937320e-01 -1.13171265e-01
3.25194687e-01 -1.15157306e-01 -9.14227724e-01 1.75739551e+00
2.89941430e-01 -4.85739745e-02 1.69204742e-01 -1.65006232e+00
-6.85712993e-01 1.23398960e+00 -8.20682704e-01 4.82105494e-01
3.65267396e-01 4.24023300e-01 9.76767361e-01 -1.60514072e-01
5.06072581e-01 1.49674928e+00 4.40888733e-01 4.74692613e-01
-1.09416795e+00 -8.13335359e-01 4.58647609e-01 6.42969832e-02
-1.38905394e+00 -2.43418496e-02 1.06658387e+00 -7.21278936e-02
5.60033798e-01 4.83781308e-01 1.29439116e-01 1.47666407e+00
3.39341521e-01 6.20682240e-01 6.41605854e-01 -1.93648323e-01
1.99542552e-01 9.66438055e-02 2.27369443e-01 1.07619083e+00
7.38042951e-01 3.55306685e-01 -2.51108378e-01 -6.19807720e-01
4.86465424e-01 -2.50905361e-02 -4.39649403e-01 -5.51917613e-01
-6.44212067e-01 1.11477804e+00 8.76115441e-01 1.44991670e-02
-3.19214731e-01 3.37444693e-01 6.06605589e-01 4.16338682e-01
3.65908682e-01 6.83464050e-01 -2.63574094e-01 4.45168376e-01
-3.62110794e-01 3.54305744e-01 1.18328035e+00 6.50796294e-01
4.68636423e-01 1.86307907e-01 9.03690830e-02 3.34003538e-01
2.23015800e-01 3.82066756e-01 -6.21223971e-02 -1.91849753e-01
6.52692735e-01 7.35955060e-01 -5.52369297e-01 -1.51643932e+00
-2.18790337e-01 -5.64105272e-01 -1.16229153e+00 1.53716624e-01
3.49407047e-01 -2.03895494e-01 -9.23020840e-01 1.66158295e+00
7.54906952e-01 2.91165322e-01 -1.31923392e-01 6.29572392e-01
4.18194950e-01 4.00645554e-01 3.59329224e-01 -1.14923976e-02
1.25855589e+00 -9.17659700e-01 -4.72764581e-01 -3.47238839e-01
4.99717534e-01 -2.09454536e-01 7.78685629e-01 7.05919564e-01
-9.50569868e-01 -6.42683432e-02 -1.28879583e+00 5.46912313e-01
-9.32182610e-01 -8.22623253e-01 6.41204357e-01 9.20378923e-01
-8.05537164e-01 8.60721529e-01 -8.25221419e-01 -2.65283734e-01
5.00395596e-01 3.71323884e-01 -5.43473184e-01 -1.06273845e-01
-1.58058691e+00 7.42203176e-01 5.65397203e-01 -2.51080036e-01
-1.48726892e+00 -5.92843354e-01 -8.81652117e-01 2.70145107e-02
1.08458066e+00 -4.71586198e-01 9.06670988e-01 -3.37984800e-01
-7.99530864e-01 6.27594531e-01 5.47624886e-01 -8.01124215e-01
4.68317866e-01 1.68420654e-02 -5.18168032e-01 2.31877431e-01
-1.45704508e-01 -8.32649693e-02 1.22345757e+00 -1.53128242e+00
-3.24330419e-01 -6.06056631e-01 2.48295829e-01 -6.58799857e-02
-4.94389504e-01 6.57858998e-02 -5.81696332e-01 -1.03862762e+00
-2.58307159e-02 -8.24050367e-01 -4.71952677e-01 -4.07078981e-01
-1.15091479e+00 -2.52784967e-01 9.97649848e-01 -6.52155399e-01
1.76302052e+00 -1.92359352e+00 3.89298260e-01 7.28465438e-01
7.19554842e-01 4.95497257e-01 -1.90370560e-01 6.01091444e-01
-5.02329409e-01 7.52288997e-01 -1.70805752e-01 -8.82497206e-02
-4.04487662e-02 -8.52108747e-02 -5.52141309e-01 8.16889942e-01
8.61533284e-02 9.16785657e-01 -9.57438529e-01 -2.72550672e-01
7.36627579e-02 1.33433267e-01 -4.43885177e-01 2.26187155e-01
-5.46038806e-01 2.01590419e-01 -5.67720115e-01 7.64294028e-01
9.12918508e-01 -2.28055313e-01 2.98842758e-01 -8.16520602e-02
4.85427707e-01 1.34860337e-01 -1.22189188e+00 1.36047077e+00
2.77909279e-01 -2.49392420e-01 4.65132743e-01 -8.79680216e-01
7.57233858e-01 -2.96064932e-02 3.28294724e-01 -4.65563118e-01
2.47964546e-01 -2.58836359e-01 -1.59129295e-02 -4.00166124e-01
1.34001985e-01 3.87562692e-01 -2.99383461e-01 6.44309044e-01
-1.83222890e-01 7.74920732e-02 2.48614743e-01 7.87314177e-01
1.84112298e+00 -3.20078909e-01 3.23078364e-01 -1.16742723e-01
7.01935649e-01 -2.40223575e-02 3.15611154e-01 1.04196668e+00
-2.57487327e-01 -2.80227009e-02 1.02198017e+00 -3.52335155e-01
-7.76765525e-01 -1.24875546e+00 4.87545848e-01 1.10296810e+00
3.12834054e-01 -6.75030112e-01 -1.06169868e+00 -1.58684301e+00
2.33969986e-01 5.87731421e-01 -9.14000094e-01 -8.40527594e-01
-6.79165244e-01 -5.75474143e-01 9.11410570e-01 9.61333364e-02
4.55564648e-01 -1.19832110e+00 2.43837804e-01 4.01929691e-02
2.15875059e-01 -7.60103822e-01 -6.45044386e-01 3.85184765e-01
-4.40121263e-01 -1.65929484e+00 1.83609992e-01 -6.58384442e-01
5.95060587e-01 4.14580524e-01 1.15378690e+00 5.60567975e-01
-3.60528380e-01 5.77338636e-02 -5.62806189e-01 -4.74316120e-01
-9.15398180e-01 4.98496622e-01 5.69464564e-02 -7.09912553e-02
1.75501123e-01 -8.76801312e-01 -3.13043505e-01 2.04475552e-01
-1.42310047e+00 -5.90816498e-01 5.89463830e-01 2.28949770e-01
2.82212615e-01 7.96611428e-01 4.73314643e-01 -1.53906167e+00
9.23548460e-01 -7.59251177e-01 -7.61061430e-01 4.31483209e-01
-5.66244841e-01 3.16354245e-01 9.02642488e-01 -6.44316494e-01
-3.32788676e-01 -2.48877645e-01 -2.22507268e-02 -7.64403701e-01
7.69902840e-02 5.38641691e-01 -7.66271949e-01 -3.92353743e-01
1.03217089e+00 1.51869655e-01 -2.85308510e-02 -2.52875894e-01
4.97032434e-01 5.48887812e-02 6.52182400e-01 -5.19861758e-01
1.77146935e+00 1.05339833e-01 1.12225756e-01 -4.75254059e-01
-6.92761362e-01 -2.21321568e-01 -2.27215886e-01 8.28266144e-02
6.12776160e-01 -3.21889639e-01 -9.55938160e-01 5.56459963e-01
-8.47513855e-01 -2.19141081e-01 -3.78005803e-02 -4.00072128e-01
-2.60691401e-02 7.96462715e-01 -6.17938459e-01 -7.13447988e-01
-6.33587241e-01 -1.13667417e+00 6.98979795e-01 -2.53804261e-03
1.75850049e-01 -1.04368782e+00 2.23709732e-01 1.83094904e-01
1.71538338e-01 6.71558321e-01 1.21760166e+00 -1.27101278e+00
-6.21079862e-01 -5.29767692e-01 -1.75435059e-02 2.78558224e-01
1.84416145e-01 -1.83226392e-01 -6.76751733e-01 -6.72332704e-01
6.12582304e-02 -3.55986565e-01 7.17125952e-01 -1.78358302e-01
1.45049524e+00 -8.14264357e-01 -5.51428914e-01 7.31661022e-01
1.36294222e+00 1.02242328e-01 7.50794291e-01 2.49095693e-01
1.07934296e+00 3.53630334e-01 2.97376245e-01 9.04174298e-02
9.89963263e-02 2.57063061e-01 1.23298717e+00 -9.92589071e-02
1.34423196e-01 -7.53696084e-01 2.93435186e-01 2.85785615e-01
5.23626626e-01 -7.34385252e-01 -7.56664753e-01 1.37075454e-01
-1.76986682e+00 -9.39500868e-01 1.34548038e-01 2.21685934e+00
7.00122893e-01 6.29120290e-01 3.65372837e-01 4.19320464e-01
1.02334511e+00 5.88497519e-01 -6.57816529e-01 -1.73191205e-01
9.55658481e-02 3.11215222e-01 7.96065986e-01 5.77176511e-01
-1.50848114e+00 1.27533042e+00 5.28737926e+00 1.00314784e+00
-7.42828846e-01 -1.57954231e-01 4.48925912e-01 3.75433266e-01
-8.62283856e-02 7.82759562e-02 -5.89084446e-01 3.74527961e-01
8.27613056e-01 -4.28212844e-02 8.97263527e-01 9.57929075e-01
-2.14884639e-01 2.85198927e-01 -1.03217983e+00 4.19344336e-01
2.50760645e-01 -1.22729015e+00 3.79127920e-01 2.62465656e-01
2.96348006e-01 -1.64873362e-01 6.87396107e-03 3.76872122e-01
1.03783476e+00 -1.11861491e+00 2.68691182e-01 2.45853111e-01
4.83197242e-01 -1.22124338e+00 1.80925071e-01 5.05167782e-01
-1.33410084e+00 -1.40207484e-01 -1.84160680e-01 5.12810409e-01
-1.46663263e-01 3.72220755e-01 -9.33412492e-01 8.99991393e-01
7.11903632e-01 3.04668576e-01 -7.92244375e-01 8.10041249e-01
-6.37185276e-01 6.52771890e-01 -1.48263305e-01 -2.29768883e-02
2.82729000e-01 -1.79279447e-01 9.09368694e-01 8.92888308e-01
-5.04363060e-01 -1.37187585e-01 6.10761106e-01 7.59603024e-01
-4.90708739e-01 -3.25858653e-01 -9.42273915e-01 -2.88168222e-01
7.06438959e-01 1.39915836e+00 -6.25580370e-01 1.68126673e-02
-3.27527039e-02 8.71271789e-01 7.12816119e-01 1.97318375e-01
-1.18037140e+00 -5.86443424e-01 4.66660172e-01 4.11302626e-01
2.39174455e-01 -8.35854635e-02 1.02143496e-01 -9.62926269e-01
-1.98565632e-01 -1.60202086e+00 9.52163279e-01 -3.81678075e-01
-1.59958148e+00 6.54065073e-01 -8.91192704e-02 -5.15069246e-01
2.21700028e-01 -5.83712757e-01 -8.88175189e-01 6.27109885e-01
-1.21558118e+00 -1.38639307e+00 -2.47905627e-01 1.08006549e+00
1.90899760e-01 -1.66524440e-01 5.79920411e-01 6.31929114e-02
-9.93374646e-01 1.04456830e+00 -5.09120226e-01 6.07838869e-01
4.07122254e-01 -1.37238717e+00 7.71637499e-01 1.51192200e+00
2.17093080e-01 6.51729405e-01 7.78335690e-01 -1.34068131e+00
-1.76930594e+00 -1.46834040e+00 2.25234240e-01 -5.38880944e-01
1.14479601e+00 -7.33412385e-01 -1.10223734e+00 8.29308569e-01
3.50732380e-03 2.99875215e-02 1.79652616e-01 -1.09035827e-01
-8.75027895e-01 -6.02604970e-02 -1.35624862e+00 8.18687320e-01
1.14685786e+00 -3.77249151e-01 -3.09003383e-01 6.73559129e-01
1.15886998e+00 -3.23021948e-01 -6.38206065e-01 4.68155742e-01
-2.69641191e-01 -6.58197641e-01 1.20050728e+00 -9.83233452e-01
-6.19401177e-03 -2.85319686e-01 1.61217257e-01 -1.17376018e+00
-5.25480807e-01 -1.28100002e+00 -6.91828370e-01 1.13723600e+00
2.75427669e-01 -6.35804057e-01 1.05231798e+00 2.45960996e-01
1.50796548e-01 -6.03043139e-01 -6.38526917e-01 -5.83891869e-01
2.24459264e-02 -2.34127745e-01 6.30615056e-01 1.17973936e+00
5.90749048e-02 3.89770269e-01 -4.16123688e-01 7.74511039e-01
9.88778055e-01 -3.83119404e-01 9.72891927e-01 -1.08356130e+00
-3.55055362e-01 -4.30832595e-01 -2.33740479e-01 -4.09169555e-01
1.63224310e-01 -1.01722348e+00 -2.52634704e-01 -1.02395868e+00
9.81817096e-02 -2.70663142e-01 -2.40932912e-01 7.31071472e-01
-5.91050446e-01 -1.46705404e-01 8.55762884e-03 -1.07807610e-02
-7.37051368e-01 1.76301435e-01 9.63134050e-01 -4.20032471e-01
-1.32645205e-01 1.73060954e-01 -1.16022992e+00 5.03927410e-01
8.48877728e-01 -8.12827766e-01 -6.18019640e-01 1.39236692e-02
1.97497442e-01 -1.33213863e-01 4.18472290e-01 -7.92997897e-01
5.08277118e-01 -1.92352936e-01 -9.66553092e-02 -4.00549799e-01
-1.55737832e-01 -7.14269876e-01 -4.47427668e-02 9.28125620e-01
-2.67894715e-01 4.26155984e-01 5.72519787e-02 1.02541232e+00
2.37559274e-01 -1.05484203e-01 1.07179320e+00 -3.13663721e-01
-2.42597282e-01 9.91035342e-01 1.96791701e-02 2.86732227e-01
1.31430900e+00 1.15159176e-01 -6.72423363e-01 -3.17889661e-01
-6.16975546e-01 4.05932754e-01 3.14088106e-01 5.43784916e-01
5.95119953e-01 -9.08668637e-01 -6.87086344e-01 1.88020244e-01
-2.88632782e-05 1.65897831e-02 1.65647119e-02 4.02349502e-01
-4.41571474e-01 -2.75905877e-02 2.75527775e-01 -6.05750680e-02
-1.17460573e+00 1.56573045e+00 4.55498934e-01 -1.02866030e+00
-5.28259456e-01 9.27071452e-01 7.06955492e-02 -3.45477700e-01
7.11844802e-01 2.59478062e-01 -1.00955874e-01 -3.50352824e-01
4.31357026e-01 4.02801126e-01 1.06439807e-01 -1.33655578e-01
-4.06729937e-01 -1.65459275e-01 -4.99065131e-01 4.84848738e-01
1.15390909e+00 3.20911109e-01 -3.89799356e-01 -4.30500895e-01
8.87807369e-01 1.43351257e-01 -8.74445558e-01 -4.35142606e-01
2.26631239e-01 -2.78210133e-01 -1.99598521e-01 -9.31080043e-01
-1.31159103e+00 3.43509018e-01 2.89597530e-02 8.94024611e-01
1.09422755e+00 -7.48073086e-02 6.48814380e-01 4.02761221e-01
3.04682046e-01 -6.19274318e-01 4.37600911e-01 3.38466048e-01
9.02687013e-01 -9.82956350e-01 1.90407604e-01 -7.61973977e-01
-4.45185840e-01 7.02539802e-01 9.20502126e-01 -6.04119360e-01
7.04486787e-01 4.11535650e-01 -4.51415569e-01 -6.28488779e-01
-5.15563488e-01 2.82535940e-01 1.28378928e-01 7.99825370e-01
-4.11114156e-01 -2.07832783e-01 2.35116646e-01 6.38272762e-01
-1.21270217e-01 -8.60332310e-01 3.36234957e-01 8.84680450e-01
-2.97170013e-01 -1.31494081e+00 -3.96664441e-01 4.27639157e-01
-6.41622424e-01 -1.80184692e-01 -9.32607710e-01 1.01047719e+00
-2.25833237e-01 1.09415543e+00 -5.20820677e-01 -8.68588090e-01
4.61996377e-01 -3.77609313e-01 1.35139510e-01 -6.11556172e-01
-1.08471620e+00 -2.80254811e-01 8.13901201e-02 -7.61523366e-01
3.71300459e-01 -2.69082755e-01 -8.26581478e-01 -8.23663116e-01
-5.84066212e-01 4.21447843e-01 4.02125269e-01 5.58210254e-01
1.91605553e-01 6.02863789e-01 9.76554990e-01 -3.17996293e-01
-1.05390024e+00 -8.28335524e-01 -6.30426884e-01 4.89534944e-01
2.21016064e-01 -3.90012532e-01 -8.22762132e-01 -5.05974650e-01] | [6.165030002593994, 7.30071496963501] |
aae5efcd-f9d2-46be-ac6b-ca425b227121 | inharmonious-region-localization-via | 2210.02036 | null | https://arxiv.org/abs/2210.02036v1 | https://arxiv.org/pdf/2210.02036v1.pdf | Inharmonious Region Localization via Recurrent Self-Reasoning | Synthetic images created by image editing operations are prevalent, but the color or illumination inconsistency between the manipulated region and background may make it unrealistic. Thus, it is important yet challenging to localize the inharmonious region to improve the quality of synthetic image. Inspired by the classic clustering algorithm, we aim to group pixels into two clusters: inharmonious cluster and background cluster by inserting a novel Recurrent Self-Reasoning (RSR) module into the bottleneck of UNet structure. The mask output from RSR module is provided for the decoder as attention guidance. Finally, we adaptively combine the masks from RSR and the decoder to form our final mask. Experimental results on the image harmonization dataset demonstrate that our method achieves competitive performance both quantitatively and qualitatively. | ['Liqing Zhang', 'Jing Liang', 'Li Niu', 'Penghao Wu'] | 2022-10-05 | null | null | null | null | ['image-harmonization'] | ['computer-vision'] | [ 4.71813351e-01 1.43265560e-01 1.79596677e-01 -1.05503373e-01
-5.04459977e-01 -5.00971258e-01 4.36231852e-01 -2.85943329e-01
-1.68481529e-01 3.97819012e-01 6.83717132e-02 3.35207582e-02
2.39389226e-01 -6.11312151e-01 -7.68630266e-01 -8.18098843e-01
5.66255033e-01 -6.16575871e-03 3.04291457e-01 -1.43189862e-01
2.87978530e-01 2.80087322e-01 -1.46712315e+00 4.40629452e-01
1.42519879e+00 9.32446122e-01 6.05431318e-01 2.95020014e-01
-8.02925751e-02 9.02163029e-01 -7.19795763e-01 -4.02218848e-01
5.36689281e-01 -8.63916516e-01 -3.63305986e-01 6.06428087e-01
2.23123297e-01 -5.44683561e-02 -2.69715250e-01 1.45135224e+00
4.78342116e-01 1.51779100e-01 3.80852997e-01 -1.16085589e+00
-7.52984583e-01 9.27394271e-01 -1.14755917e+00 3.90500054e-02
1.54713452e-01 4.08257633e-01 5.47562838e-01 -9.56546783e-01
5.61680079e-01 1.18354070e+00 1.45874858e-01 4.03518468e-01
-1.44460261e+00 -8.42305005e-01 4.22980815e-01 1.28826022e-01
-1.67718911e+00 -6.58224881e-01 9.91662204e-01 -6.25788569e-02
3.24384362e-01 4.14960295e-01 6.77490950e-01 9.21039164e-01
7.01217726e-02 8.08399200e-01 1.11659956e+00 -2.44930089e-01
1.34228319e-01 2.49603525e-01 -5.17163634e-01 6.56289279e-01
1.05742365e-01 -2.03422114e-01 -4.53654051e-01 3.02358240e-01
9.94789720e-01 -1.57725587e-01 -4.18511152e-01 -3.64609987e-01
-1.19282937e+00 5.53718448e-01 6.06821954e-01 1.03110410e-01
-4.95731950e-01 1.53213024e-01 -1.57558337e-01 2.61741012e-01
2.50344783e-01 5.02687037e-01 8.96587148e-02 3.84207010e-01
-1.06129754e+00 -3.30697931e-02 3.34937602e-01 1.05409300e+00
8.21607172e-01 1.53287008e-01 -4.37642008e-01 1.11724198e+00
8.33811909e-02 2.42049769e-01 3.04901481e-01 -1.41208363e+00
3.20260137e-01 7.74635494e-01 1.08615361e-01 -1.14909101e+00
2.86048092e-02 -5.86908996e-01 -1.04792225e+00 2.22114369e-01
3.49433422e-02 -8.04572180e-03 -9.19933736e-01 1.65367579e+00
3.75967622e-01 2.43487164e-01 -1.13112472e-01 1.07496822e+00
6.73300207e-01 6.32377863e-01 -2.17667475e-01 -3.61919791e-01
1.07437670e+00 -1.28038228e+00 -7.57307827e-01 -3.99361163e-01
7.28804842e-02 -9.03064251e-01 9.74223316e-01 2.29298413e-01
-1.48984241e+00 -6.33032978e-01 -9.95968282e-01 -4.31684852e-02
7.92245492e-02 2.80492246e-01 1.54548064e-01 2.76971847e-01
-1.05606616e+00 2.35871568e-01 -5.25494099e-01 -4.35131863e-02
5.73705792e-01 2.16160968e-01 -6.68457076e-02 1.61767635e-03
-9.07294452e-01 7.65873134e-01 4.16839093e-01 2.33047783e-01
-6.96247995e-01 -4.92942214e-01 -8.08772266e-01 -5.04573584e-02
3.97312909e-01 -6.58904135e-01 8.97033513e-01 -1.37181127e+00
-1.63762748e+00 1.01386166e+00 -2.31737733e-01 -3.30346495e-01
9.00357604e-01 1.44937560e-01 -2.57966429e-01 1.39426425e-01
2.53403425e-01 1.04819870e+00 1.08159578e+00 -1.84714866e+00
-7.20026851e-01 -1.55773520e-01 -1.83597594e-01 4.49135333e-01
-1.48567572e-01 -5.97698130e-02 -1.10165691e+00 -1.05334675e+00
2.58770257e-01 -9.15632665e-01 -3.56660396e-01 2.66732853e-02
-8.60303164e-01 2.41471112e-01 8.18150163e-01 -7.04748213e-01
1.28106022e+00 -2.46169090e+00 2.25705266e-01 3.73762190e-01
1.69857845e-01 1.01437785e-01 -2.84872532e-01 7.66923884e-03
-3.08655828e-01 -2.92471121e-03 -4.61781949e-01 -2.27988765e-01
-2.18201444e-01 6.70024976e-02 -3.60171795e-01 3.66906404e-01
3.37663591e-01 1.02230501e+00 -7.93666840e-01 -7.25532234e-01
2.63671070e-01 2.52952546e-01 -5.07958651e-01 2.75494874e-01
-1.83485106e-01 7.06510901e-01 -3.36924553e-01 6.45597875e-01
9.18706477e-01 -3.06619078e-01 4.75923903e-02 -3.61523241e-01
-1.37824744e-01 -1.66387372e-02 -1.15340006e+00 1.66658676e+00
-2.30273798e-01 5.55842161e-01 1.93646342e-01 -7.47974277e-01
9.30764675e-01 -8.96659866e-02 3.75568777e-01 -9.04935360e-01
2.28601098e-01 -1.55516848e-01 2.88661599e-01 -3.41306627e-01
5.01982570e-01 1.23570092e-01 -2.09555384e-02 4.16942805e-01
-4.36639577e-01 -3.05013448e-01 1.74689233e-01 2.08346680e-01
7.53245771e-01 5.57199344e-02 8.06315169e-02 -1.19415246e-01
5.73177993e-01 -2.78344210e-02 8.14865828e-01 5.91504574e-01
-2.70990878e-01 1.15995705e+00 5.54631650e-01 -8.15287828e-02
-9.95815754e-01 -1.26741385e+00 9.23438296e-02 8.02898824e-01
6.81542456e-01 -1.69952005e-01 -1.06387162e+00 -4.02624279e-01
-3.57567549e-01 8.15988362e-01 -7.95208454e-01 -5.02900124e-01
-5.70555031e-01 -6.37634039e-01 3.44629854e-01 3.42974693e-01
1.00911391e+00 -1.03273714e+00 -4.85651255e-01 1.60646319e-01
-4.72023010e-01 -1.13684165e+00 -8.48957360e-01 -5.92392422e-02
-3.76054853e-01 -7.60637760e-01 -6.97029650e-01 -9.07287598e-01
1.00786412e+00 6.48428440e-01 8.83675218e-01 2.64188528e-01
-2.07754359e-01 -2.21503615e-01 -2.45041952e-01 -1.06380239e-01
-6.85840607e-01 -2.59169966e-01 -3.94186020e-01 5.71150243e-01
-2.27043808e-01 -5.17660379e-01 -9.44670856e-01 6.17524981e-01
-9.55461383e-01 5.73376477e-01 6.41351998e-01 5.35558522e-01
6.57566488e-01 4.37957257e-01 3.22590441e-01 -7.25156784e-01
5.08496761e-01 -1.44456610e-01 -5.93028188e-01 2.80324966e-01
-2.58547723e-01 -1.58414036e-01 6.92291260e-01 -5.03778934e-01
-1.19285631e+00 2.39379629e-01 1.92656368e-01 -6.42875731e-01
1.94357306e-01 -1.29388317e-01 -5.57927728e-01 2.14898749e-03
4.65174913e-01 4.28210258e-01 1.00697145e-01 2.67066440e-04
5.12233138e-01 5.81827641e-01 9.21151340e-01 -1.88480824e-01
9.50565338e-01 6.78711832e-01 -5.25865674e-01 -5.63778222e-01
-6.31115079e-01 1.47502109e-01 -4.77730304e-01 -4.13420230e-01
1.02660441e+00 -9.22718048e-01 -4.76665407e-01 4.05381769e-01
-1.07053244e+00 -4.85781312e-01 -3.87196809e-01 -1.07120663e-01
-5.35367548e-01 2.29376242e-01 -3.44257057e-01 -6.18376255e-01
-1.06202066e-01 -1.41931629e+00 9.39093530e-01 5.71035385e-01
-1.93259090e-01 -4.95591521e-01 -4.78242785e-01 5.83468080e-01
3.71298850e-01 2.97718763e-01 7.90495217e-01 6.26422837e-03
-7.96542466e-01 5.29980175e-02 -4.66833472e-01 2.73012549e-01
2.26747662e-01 2.34166294e-01 -9.37168539e-01 -1.11217596e-01
-3.80567415e-03 -5.74899651e-02 1.11401844e+00 3.35991293e-01
1.48641562e+00 -1.56081483e-01 -3.30567122e-01 7.22551882e-01
1.10443163e+00 4.06491876e-01 9.22089636e-01 2.02019006e-01
7.75777996e-01 7.19367087e-01 5.19971848e-01 3.10421824e-01
4.03808653e-01 6.25692070e-01 3.34365040e-01 -4.75692362e-01
-5.97985387e-01 -3.56400430e-01 4.00980294e-01 7.60321796e-01
1.84019506e-01 -3.51630777e-01 -6.46558464e-01 3.40552479e-01
-1.88987505e+00 -9.77534652e-01 8.19967389e-02 2.04913282e+00
8.45925391e-01 1.35930315e-01 -1.40957028e-01 -2.02867351e-02
1.21564543e+00 3.20012838e-01 -8.57180119e-01 -1.97810367e-01
-5.41880012e-01 -2.35334143e-01 2.98896730e-01 3.14800769e-01
-9.71403420e-01 1.19420540e+00 6.32616138e+00 9.74215150e-01
-1.03318369e+00 -3.40404212e-02 1.03144872e+00 -2.66115427e-01
-4.07791734e-01 1.46567598e-02 -2.48842046e-01 7.06980228e-01
1.96751237e-01 -1.05807528e-01 7.67776072e-01 2.31532604e-01
6.00209773e-01 -3.55157107e-01 -7.18035042e-01 1.06400883e+00
2.13780224e-01 -1.25014150e+00 2.56109804e-01 -2.38159657e-01
1.07665789e+00 -2.73741663e-01 4.74652350e-01 -1.14094786e-01
5.50273836e-01 -9.51766372e-01 9.74136949e-01 3.05709243e-01
7.18284309e-01 -8.19663525e-01 1.47463605e-01 2.87349135e-01
-1.18842149e+00 -2.24622354e-01 -3.13937634e-01 3.50777566e-01
1.65707707e-01 4.47438627e-01 -4.88903344e-01 4.09754336e-01
6.84967995e-01 7.57523298e-01 -7.37601221e-01 9.24448013e-01
-4.67147827e-01 1.40487194e-01 1.99119598e-02 4.69209075e-01
-1.00438455e-02 -6.28103971e-01 6.17227435e-01 8.19469333e-01
1.18080616e-01 1.16695583e-01 -1.03819985e-02 1.43339038e+00
-2.87226737e-01 -1.26640648e-01 -4.52499092e-01 1.48980334e-01
6.17512643e-01 1.31413829e+00 -1.12970889e+00 -3.64175588e-01
-2.22510844e-01 1.47959745e+00 3.38679314e-01 5.16939938e-01
-1.14096236e+00 -8.09747353e-02 5.13163686e-01 7.83700719e-02
4.27997380e-01 1.58897981e-01 -6.32205665e-01 -1.22160578e+00
4.25033011e-02 -1.09407580e+00 1.73380390e-01 -1.10998404e+00
-1.01739323e+00 7.22904623e-01 -3.68925542e-01 -1.30859292e+00
2.37208664e-01 3.25434655e-02 -7.79369533e-01 7.10762739e-01
-1.23235500e+00 -1.05844474e+00 -5.65492928e-01 5.42921245e-01
6.22277319e-01 -3.87863512e-03 3.07247788e-01 2.36821473e-01
-1.16176450e+00 6.51387990e-01 -6.81599677e-02 8.22519138e-02
7.17538118e-01 -9.86225545e-01 3.86519432e-01 1.03174770e+00
2.28784643e-02 5.47671795e-01 6.67363346e-01 -6.60734355e-01
-1.15417576e+00 -1.48639619e+00 2.00971067e-01 -1.05729930e-01
2.76105970e-01 -4.30098981e-01 -9.68522906e-01 3.82045627e-01
5.85254550e-01 -1.74459413e-01 3.01071584e-01 -7.67144799e-01
-3.07291359e-01 -1.58133224e-01 -9.57181931e-01 1.16365325e+00
1.08963430e+00 -3.27530622e-01 -3.90327066e-01 1.79397374e-01
7.04879045e-01 -4.16303515e-01 -4.59140569e-01 3.24773490e-01
2.66810507e-01 -1.08159137e+00 9.42583799e-01 1.01449922e-01
6.27660036e-01 -7.88418293e-01 -3.89365107e-02 -1.37641156e+00
-3.59554321e-01 -9.53224480e-01 3.68380725e-01 1.38014257e+00
5.40967166e-01 -4.57750827e-01 6.17017210e-01 6.20359540e-01
-5.25039360e-02 -4.52162027e-01 -5.74653089e-01 -5.06122410e-01
-2.08931774e-01 -6.27864003e-02 5.26462078e-01 9.27981317e-01
-3.73367310e-01 2.85183758e-01 -2.20140204e-01 9.38397869e-02
7.08684623e-01 6.06191039e-01 8.83446872e-01 -7.70552158e-01
-2.55223840e-01 -7.77550220e-01 5.32064177e-02 -1.10641801e+00
1.90762475e-01 -8.41372311e-01 2.40719646e-01 -1.48140514e+00
4.73661631e-01 -4.29226130e-01 -2.40693390e-01 3.31217319e-01
-4.71918702e-01 7.28177547e-01 4.24492687e-01 2.73205519e-01
-6.27403200e-01 7.41178334e-01 1.62065029e+00 -3.13976794e-01
-3.40457767e-01 -2.74975777e-01 -1.08683312e+00 8.20663750e-01
7.22753704e-01 -3.88835371e-01 -4.38843966e-01 -6.15274310e-01
4.76747118e-02 -1.83182284e-02 2.64515519e-01 -7.82181203e-01
3.79653603e-01 -1.46640584e-01 6.63320720e-01 -6.19041204e-01
2.70711005e-01 -6.66508138e-01 3.29842031e-01 4.68290478e-01
-3.89001638e-01 -5.75919449e-02 2.04935577e-02 4.81543452e-01
-2.72355080e-01 1.55056700e-01 1.27868712e+00 -7.70573616e-02
-5.40705860e-01 2.38743529e-01 -2.80326545e-01 9.75649208e-02
1.28622651e+00 -5.47372460e-01 -3.07457268e-01 -5.17353058e-01
-5.05849183e-01 4.92358565e-01 9.91239250e-01 5.24513066e-01
7.66149521e-01 -1.29455066e+00 -7.29941130e-01 5.03191233e-01
1.90208796e-02 1.39102057e-01 3.29231381e-01 8.93410206e-01
-4.82049495e-01 -2.65380710e-01 -2.17311606e-01 -5.78259408e-01
-1.07833254e+00 6.42622471e-01 4.67963368e-01 3.20548058e-01
-6.07063413e-01 8.16847444e-01 6.45169616e-01 -1.62210584e-01
1.51397780e-01 -7.28003830e-02 -1.81642175e-03 -2.84429286e-02
4.29284960e-01 2.16415152e-01 -2.69561529e-01 -7.11766422e-01
-2.36483276e-01 6.05386794e-01 -1.55461535e-01 -2.80156374e-01
1.02939200e+00 -5.80065966e-01 -3.76046270e-01 2.11581379e-01
1.04383934e+00 -2.20398158e-02 -1.53630424e+00 -1.11804560e-01
-3.42172861e-01 -4.97121006e-01 -5.78405857e-02 -6.77870214e-01
-1.40342832e+00 5.87427497e-01 4.17144120e-01 4.93222848e-02
1.53583229e+00 -5.22399060e-02 7.45229244e-01 -2.11340755e-01
6.57490566e-02 -1.18373203e+00 2.34579995e-01 1.15718067e-01
1.14375806e+00 -1.07789254e+00 -1.09268717e-01 -7.44991243e-01
-9.73466456e-01 5.33457577e-01 8.59244287e-01 -3.41105998e-01
2.57985204e-01 4.65381712e-01 3.80403668e-01 2.44866367e-02
-7.23661602e-01 -1.16161726e-01 1.08632185e-01 5.40270090e-01
1.00093529e-01 -1.71997368e-01 -6.56821877e-02 3.77178311e-01
-2.94976950e-01 -4.21204120e-01 5.36260605e-01 6.58789098e-01
-3.77085388e-01 -8.59157681e-01 -4.79292423e-01 2.89224237e-01
-1.75417051e-01 -1.69219337e-02 -6.44655585e-01 4.80358034e-01
2.87177891e-01 1.11724317e+00 2.62102306e-01 -4.25631285e-01
2.84930021e-01 -3.94680589e-01 3.81776869e-01 -4.04146820e-01
-6.47376537e-01 4.86091614e-01 -3.47420573e-01 -7.48197794e-01
-3.37988615e-01 -4.57811356e-01 -1.36926115e+00 -8.14005435e-02
-2.88895696e-01 -6.50939047e-02 3.25572670e-01 6.63601577e-01
4.57103133e-01 6.81118906e-01 1.03121591e+00 -6.94881260e-01
-2.63763610e-02 -6.56244159e-01 -5.12504697e-01 4.67436343e-01
1.64761335e-01 -4.06556368e-01 -1.33332416e-01 3.51918221e-01] | [11.26550006866455, -1.163069725036621] |
04e45bab-027b-46b0-95a3-f109466f5747 | a-compact-neural-architecture-for-visual | 1910.06840 | null | https://arxiv.org/abs/1910.06840v3 | https://arxiv.org/pdf/1910.06840v3.pdf | A Hybrid Compact Neural Architecture for Visual Place Recognition | State-of-the-art algorithms for visual place recognition, and related visual navigation systems, can be broadly split into two categories: computer-science-oriented models including deep learning or image retrieval-based techniques with minimal biological plausibility, and neuroscience-oriented dynamical networks that model temporal properties underlying spatial navigation in the brain. In this letter, we propose a new compact and high-performing place recognition model that bridges this divide for the first time. Our approach comprises two key neural models of these categories: (1) FlyNet, a compact, sparse two-layer neural network inspired by brain architectures of fruit flies, Drosophila melanogaster, and (2) a one-dimensional continuous attractor neural network (CANN). The resulting FlyNet+CANN network incorporates the compact pattern recognition capabilities of our FlyNet model with the powerful temporal filtering capabilities of an equally compact CANN, replicating entirely in a hybrid neural implementation the functionality that yields high performance in algorithmic localization approaches like SeqSLAM. We evaluate our model, and compare it to three state-of-the-art methods, on two benchmark real-world datasets with small viewpoint variations and extreme environmental changes - achieving 87% AUC results under day to night transitions compared to 60% for Multi-Process Fusion, 46% for LoST-X and 1% for SeqSLAM, while being 6.5, 310, and 1.5 times faster, respectively. | ['Andrew B. Barron', 'Luis Hernandez-Nunez', 'Marvin Chancán', 'Ajay Narendra', 'Michael Milford'] | 2019-10-15 | null | null | null | null | ['sequential-place-learning', 'sequential-place-recognition'] | ['robots', 'robots'] | [-9.61506963e-02 -2.53191113e-01 1.30338609e-01 -4.00160030e-02
1.64242730e-01 -6.13832951e-01 1.20400870e+00 2.63947505e-03
-8.90435100e-01 6.79627955e-01 6.48197830e-02 -1.36507496e-01
-4.61479515e-01 -6.33307576e-01 -6.21614873e-01 -8.60436440e-01
-6.54780984e-01 3.02412868e-01 6.07459784e-01 -3.33742678e-01
2.84455299e-01 6.45707786e-01 -1.87917793e+00 -1.37589956e-02
5.65079927e-01 1.16167843e+00 3.18542808e-01 6.99412942e-01
2.90123038e-02 7.12525189e-01 -1.79229409e-01 1.85731903e-01
1.93235293e-01 -2.87423939e-01 -5.06663978e-01 -6.37151062e-01
3.16932261e-01 1.33031040e-01 -6.87317789e-01 7.29112566e-01
5.56495786e-01 3.10792238e-01 5.67430496e-01 -1.06074655e+00
-7.29289711e-01 2.00008214e-01 -2.16269746e-01 4.85661775e-01
2.32397050e-01 4.19516057e-01 6.07038677e-01 -7.64183164e-01
7.84754694e-01 1.11066270e+00 9.43788230e-01 4.85213012e-01
-1.34345555e+00 -2.84272462e-01 1.24923944e-01 2.15801239e-01
-1.33286798e+00 -5.82797170e-01 3.23432863e-01 -4.53965098e-01
1.43673778e+00 1.29505172e-02 1.12877381e+00 1.35346758e+00
7.74108708e-01 5.08953989e-01 1.17055833e+00 -1.98549777e-03
7.70128191e-01 -4.22572345e-01 -8.53708908e-02 8.18812490e-01
2.17980698e-01 4.54354405e-01 -8.69739056e-01 -1.22819401e-01
8.87255728e-01 2.50814021e-01 -3.37080806e-01 -6.40871584e-01
-1.62372315e+00 3.96749139e-01 9.20781851e-01 6.32093728e-01
-4.24004823e-01 5.03191531e-01 1.99605227e-01 9.64704752e-02
1.02796860e-01 3.14648002e-01 -2.51770824e-01 -1.07399926e-01
-1.27780390e+00 3.12200308e-01 8.30901742e-01 6.52333140e-01
6.88499749e-01 3.11500281e-01 -6.59660399e-02 5.03188014e-01
6.28146708e-01 5.40982068e-01 9.89815593e-01 -7.94542968e-01
-1.80993795e-01 4.51294482e-01 -7.60768726e-02 -1.08020961e+00
-9.28108096e-01 -4.22914147e-01 -1.18654346e+00 3.16728175e-01
2.45881066e-01 1.90423533e-01 -1.15497410e+00 1.85212862e+00
-1.25676781e-01 4.76782471e-01 8.14385042e-02 8.19858611e-01
7.46917844e-01 6.12645328e-01 -2.92354729e-03 2.59208053e-01
1.33149517e+00 -1.10630584e+00 -3.70201886e-01 -4.81566727e-01
3.17923635e-01 -1.48979336e-01 4.84782696e-01 1.79086506e-01
-8.32815349e-01 -5.77156067e-01 -1.29330969e+00 5.80186993e-02
-1.02667987e+00 1.99538976e-01 6.73203707e-01 3.43068779e-01
-1.87331378e+00 8.50211799e-01 -1.02847123e+00 -9.90965307e-01
4.67329204e-01 5.15146375e-01 -5.52214801e-01 2.31866494e-01
-9.51666296e-01 9.86452579e-01 3.91239703e-01 1.94320828e-01
-1.15063190e+00 -5.47478497e-01 -8.51269364e-01 4.17536870e-02
-1.63522959e-01 -9.46757019e-01 8.55826020e-01 -6.96170747e-01
-1.45922375e+00 9.02313232e-01 -3.53632271e-01 -1.13408589e+00
3.32760274e-01 -7.21763745e-02 -4.43384796e-01 4.84472141e-02
2.67127622e-02 1.23263681e+00 5.67616403e-01 -9.26029682e-01
-3.94426703e-01 -5.12803614e-01 -2.86368281e-01 -1.70192681e-02
-1.40158981e-02 -3.81039590e-01 -2.56757855e-01 -5.00118613e-01
2.31244430e-01 -1.01936936e+00 -5.65050542e-01 5.65310776e-01
-3.80474143e-02 9.88670290e-02 8.87185752e-01 -3.75443429e-01
8.87293756e-01 -2.18511033e+00 2.61746049e-01 6.59702271e-02
3.03794712e-01 4.19755727e-01 -4.12774563e-01 4.21383917e-01
4.23098952e-02 -2.06271917e-01 -5.64482212e-01 -3.57362300e-01
-1.38706431e-01 3.36195618e-01 -3.34262788e-01 7.97463059e-01
1.36148864e-02 1.18494713e+00 -1.03954542e+00 1.25158831e-01
3.36033791e-01 5.79051495e-01 -3.99798214e-01 -2.32496262e-01
-6.26839250e-02 2.61604548e-01 -6.50392622e-02 7.44459808e-01
4.23390925e-01 -3.03238392e-01 5.28194755e-02 2.22172856e-01
-6.11661196e-01 1.21212706e-01 -1.16983902e+00 2.27393126e+00
-1.95816368e-01 1.00231338e+00 4.84004820e-04 -7.26459205e-01
9.71414030e-01 8.92954543e-02 3.93407792e-01 -1.03771341e+00
1.51615646e-02 3.57314646e-01 1.68759301e-01 -5.71825132e-02
4.78561521e-01 2.54125029e-01 1.29669860e-01 1.45857051e-01
6.63228750e-01 2.66035169e-01 1.84122756e-01 -7.20982552e-02
1.43602288e+00 4.68057334e-01 4.05687898e-01 -7.79191911e-01
4.73783344e-01 -1.34471565e-01 3.50588918e-01 9.73851681e-01
-5.76982319e-01 6.22853458e-01 1.74964502e-01 -8.87768507e-01
-8.65247190e-01 -1.19508684e+00 -3.83294970e-02 8.47851515e-01
2.50942200e-01 -4.78331447e-01 -6.04689896e-01 1.39752459e-02
-8.94902647e-03 3.44826937e-01 -8.93495500e-01 -2.56974846e-01
-4.74900961e-01 -9.41940606e-01 1.02120376e+00 4.20660973e-01
8.37125003e-01 -1.25210083e+00 -1.32870007e+00 3.22681099e-01
1.74869746e-01 -6.72839284e-01 3.14863503e-01 6.65752888e-01
-7.55271673e-01 -1.03713727e+00 -8.80865514e-01 -7.17989326e-01
2.29664013e-01 1.78473905e-01 9.43453193e-01 -1.79299563e-01
-2.75370061e-01 2.63097793e-01 -6.41149934e-04 -1.76204056e-01
2.77699590e-01 1.92691162e-02 5.01182854e-01 -1.40102148e-01
1.78985059e-01 -1.08731675e+00 -7.37019420e-01 2.31519997e-01
-9.17601466e-01 -9.72621217e-02 5.97986162e-01 9.92950022e-01
5.80136776e-01 -4.87500846e-01 1.26711458e-01 -7.21381307e-02
3.86039853e-01 -6.45235538e-01 -7.65837312e-01 9.97312590e-02
-6.24077082e-01 3.22760880e-01 6.32025838e-01 -3.39271009e-01
-6.25655293e-01 3.33583266e-01 -2.23722026e-01 -4.20296848e-01
-5.26433349e-01 2.78091818e-01 3.28325421e-01 -5.96420944e-01
8.55243385e-01 9.10105288e-01 9.12745669e-02 -3.33375096e-01
4.13922817e-01 3.63805667e-02 7.77048886e-01 -2.37279490e-01
5.60141206e-01 9.47914660e-01 3.86255503e-01 -9.23519731e-01
-1.97583377e-01 -6.42913222e-01 -8.13397527e-01 -2.26049766e-01
9.58728492e-01 -8.48886311e-01 -7.01898694e-01 7.80878663e-01
-1.17271078e+00 -4.99284655e-01 -3.16163272e-01 3.16642940e-01
-8.55708361e-01 1.08033448e-01 -4.56109643e-01 -6.73663974e-01
-2.03565180e-01 -8.44523847e-01 1.00725555e+00 4.91406024e-01
-4.81239818e-02 -1.06511807e+00 7.55641937e-01 -5.64649701e-01
9.88609612e-01 2.84392506e-01 5.55342913e-01 -7.88367033e-01
-6.70969605e-01 1.82492346e-01 -9.77739021e-02 -2.01962516e-01
-2.37578034e-01 -8.23241994e-02 -1.13120246e+00 -2.22681046e-01
-1.87498659e-01 -1.62945434e-01 1.34844005e+00 4.72804755e-01
7.14237869e-01 -1.61930457e-01 -7.26807058e-01 9.77257729e-01
1.65419900e+00 3.37836951e-01 7.68303216e-01 7.61403024e-01
2.91520834e-01 3.60506147e-01 -5.56826517e-02 3.57319027e-01
3.20163995e-01 5.64691782e-01 7.24403322e-01 3.76038216e-02
-1.95895368e-03 -1.16745695e-01 3.41371953e-01 5.25613070e-01
-3.39330494e-01 -2.38826752e-01 -1.18742383e+00 6.79324448e-01
-2.18089175e+00 -1.32292390e+00 3.46622057e-02 1.98719037e+00
9.30290520e-02 -3.62447836e-02 -1.95719935e-02 -1.20926276e-01
2.24896178e-01 6.33833289e-01 -5.75487614e-01 -4.31274772e-02
-5.94273508e-01 1.84316635e-01 4.95232701e-01 5.79091944e-02
-1.41561997e+00 8.97738576e-01 6.73390579e+00 4.55105424e-01
-1.35376036e+00 -9.56541598e-02 3.66624326e-01 -1.38765946e-01
1.35922164e-01 -9.13830474e-02 -5.88246047e-01 4.04410839e-01
1.50619769e+00 6.77912682e-02 7.32478678e-01 8.08415055e-01
-1.69965513e-02 -2.70147741e-01 -9.78495181e-01 1.17213058e+00
1.95539549e-01 -1.81600308e+00 4.67625037e-02 1.46660954e-01
5.80124676e-01 8.17379177e-01 2.10102871e-01 1.50766850e-01
1.27092794e-01 -1.19806409e+00 9.04913366e-01 9.54123080e-01
2.58761823e-01 -3.54131401e-01 5.55734932e-01 4.92221624e-01
-1.46757233e+00 -3.23136419e-01 -4.57846850e-01 -2.18610123e-01
9.21957567e-02 2.67037362e-01 -1.73313037e-01 5.18843293e-01
1.08870101e+00 1.11027968e+00 -9.65659022e-01 1.67753255e+00
1.35028914e-01 1.74042672e-01 -5.72437406e-01 -2.25461453e-01
8.54796648e-01 -1.70919210e-01 7.19934821e-01 1.33573139e+00
5.74700832e-01 -2.61153311e-01 -3.07955027e-01 1.02207649e+00
2.07174972e-01 -3.19969296e-01 -1.08864689e+00 2.26414204e-01
3.53522748e-01 1.27710664e+00 -9.74232614e-01 -2.29207724e-01
-2.14977205e-01 1.17799437e+00 3.14825773e-01 4.27799881e-01
-6.60699427e-01 -2.89039671e-01 9.17896569e-01 -4.05221507e-02
5.65897882e-01 -4.26050276e-01 -4.31090556e-02 -1.13312352e+00
-2.47796044e-01 -4.05611902e-01 6.73636496e-02 -9.56483006e-01
-9.20048475e-01 8.82344842e-01 -2.82047093e-01 -1.08096886e+00
-4.11932021e-01 -1.10956776e+00 -6.49361372e-01 7.66444325e-01
-1.57726967e+00 -1.05251670e+00 -2.52192378e-01 7.08515525e-01
2.46751249e-01 -2.21165940e-01 1.06944573e+00 2.02041149e-01
-3.38444412e-01 8.63557905e-02 5.73894203e-01 4.79334034e-02
3.55204523e-01 -1.18316853e+00 8.19455028e-01 8.94489646e-01
4.79506373e-01 9.09858465e-01 4.19217139e-01 -2.44416356e-01
-1.43561065e+00 -1.16703856e+00 8.57152879e-01 -4.58719671e-01
7.35083222e-01 -5.63106596e-01 -8.05635273e-01 5.58124363e-01
3.45110089e-01 2.88031399e-01 4.46366221e-01 -2.08829314e-01
-3.39496315e-01 9.73203406e-02 -1.16303647e+00 8.01464558e-01
1.36719191e+00 -5.32737732e-01 -7.32210398e-01 5.08765839e-02
1.60358965e-01 -3.09569925e-01 -4.34879571e-01 2.96376675e-01
8.11060429e-01 -1.32158279e+00 1.12767112e+00 -3.61674845e-01
8.35181121e-03 -5.95612109e-01 -1.71606541e-01 -1.30937862e+00
-7.09663153e-01 -4.97560799e-01 -2.20330015e-01 5.42594075e-01
2.25328445e-01 -7.34195054e-01 5.99291027e-01 -6.57426342e-02
-1.93824604e-01 -7.34043062e-01 -1.48494279e+00 -8.91915560e-01
-1.35710761e-01 -2.05195040e-01 3.24313909e-01 6.35727346e-01
-2.05788434e-01 6.25428632e-02 -2.33911216e-01 3.91862318e-02
4.41094369e-01 -1.33007929e-01 3.14822048e-01 -1.49747515e+00
4.09607291e-02 -7.12216198e-01 -1.00989425e+00 -1.07036614e+00
1.11636691e-01 -7.55697429e-01 1.09633192e-01 -1.56889677e+00
-1.12060361e-01 -5.73441945e-02 -5.69617510e-01 6.38697684e-01
6.21100783e-01 2.47390911e-01 1.32770747e-01 3.72239113e-01
-8.60897899e-01 8.11731696e-01 5.73961556e-01 -1.00475460e-01
-1.75138801e-01 -3.87785643e-01 -2.28816420e-01 8.93072724e-01
5.39630711e-01 -3.83542299e-01 -1.61909536e-01 -3.33634704e-01
7.99878761e-02 -3.32627028e-01 9.43526268e-01 -1.86329496e+00
7.66270518e-01 1.53800622e-01 7.48577774e-01 -4.59358156e-01
4.27488774e-01 -7.64533579e-01 3.07506830e-01 9.99341309e-01
-3.65102366e-02 3.94200802e-01 5.71816325e-01 8.79732907e-01
-1.78880617e-01 3.41337353e-01 7.65068769e-01 -3.46937090e-01
-1.36222982e+00 3.15076411e-01 -8.60396683e-01 -3.25843215e-01
9.49747860e-01 -4.54208523e-01 -5.57268381e-01 -6.17501624e-02
-6.57124162e-01 -7.30007561e-03 5.34754872e-01 6.39316738e-01
6.94455862e-01 -1.28213263e+00 -1.94154605e-01 4.90049124e-01
1.22033291e-01 -4.08285707e-01 2.56950229e-01 9.52107608e-01
-6.54879928e-01 1.05402684e+00 -8.12314332e-01 -8.48144591e-01
-6.74806595e-01 4.73318279e-01 8.60426366e-01 -1.21693462e-01
-5.49905717e-01 6.41957283e-01 1.16167635e-01 -6.80520594e-01
6.83587268e-02 -6.67414427e-01 -2.41188154e-01 5.65023422e-02
5.15962720e-01 3.42521459e-01 -2.44572777e-02 -7.22365379e-01
-7.63087988e-01 5.01451790e-01 4.61074173e-01 -2.60918915e-01
1.40928483e+00 -3.26593108e-02 -2.64065802e-01 7.22265840e-01
7.73476481e-01 -5.29968023e-01 -1.26947832e+00 -5.97969745e-04
1.63071558e-01 1.56478658e-01 -7.31752589e-02 -8.64505291e-01
-6.51339650e-01 8.35729361e-01 1.09993672e+00 5.31588607e-02
9.79703248e-01 -1.47883102e-01 3.72735471e-01 9.26722586e-01
7.67813385e-01 -7.31273592e-01 -1.24289930e-01 9.65708435e-01
8.46651614e-01 -8.80589545e-01 -3.65737766e-01 4.29974973e-01
-4.26977463e-02 9.63845611e-01 4.82716888e-01 -5.05409479e-01
9.06723797e-01 2.40109831e-01 -2.09083349e-01 -3.04740787e-01
-9.26349699e-01 -2.54877865e-01 3.87175888e-01 7.52473474e-01
1.37061834e-01 -2.13954285e-01 1.47840023e-01 3.71108949e-01
-4.39140610e-02 1.38121992e-01 6.09158054e-02 1.17366087e+00
-6.23438060e-01 -5.41061103e-01 -7.68071339e-02 2.32423052e-01
-1.40779121e-02 -3.78219813e-01 -4.17384684e-01 8.54717851e-01
1.87485173e-01 3.81321937e-01 2.08545342e-01 -3.20759296e-01
2.05964997e-01 1.87979609e-01 2.90450960e-01 -1.80521235e-01
-7.34276354e-01 -3.31096321e-01 -3.36869478e-01 -1.11175132e+00
-5.69086015e-01 -6.10424995e-01 -1.08443177e+00 -2.51233160e-01
2.29728222e-01 -1.57063231e-01 6.06388271e-01 9.09489572e-01
7.83904433e-01 4.69838917e-01 -1.68962196e-01 -1.44047809e+00
-1.06146075e-01 -8.07451606e-01 -7.06827700e-01 -1.60530597e-01
6.42669439e-01 -8.35919678e-01 -3.26777369e-01 -1.44331098e-01] | [7.646579265594482, -1.7880743741989136] |
fa666467-4ae8-42f4-af34-093c7f6b48a3 | bi-matching-mechanism-to-combat-the-long-tail | null | null | https://openreview.net/forum?id=FvmGqUEuh7 | https://openreview.net/pdf?id=FvmGqUEuh7 | Bi-Matching Mechanism to Combat the Long Tail of Word Sense Disambiguation | The long tail phenomenon of word sense distribution in linguistics causes the Word Sense Disambiguation (WSD) task to face a serious polarization of word sense distribution, that is, Most Frequent Senses (MFSs) with huge sample sizes and Long Tail Senses (LTSs) with small sample sizes. The single matching mechanism model that does not distinguish between the two senses will cause LTSs to be ignored because LTSs are in a weak position. The few-shot learning method that mainly focuses on LTSs is not conducive to grasping the advantage of easy identification of MFSs. This paper proposes a bi-matching mechanism to serve the WSD model to deal with two kinds of senses in a targeted manner, namely definition matching and collocation feature matching. The experiment is carried out under the evaluation framework of English all-words WSD and is better than the baseline models. Moreover, state-of-the-art performance is achieved through data enhancement. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['word-sense-disambiguation'] | ['natural-language-processing'] | [-2.08657235e-01 -2.17563570e-01 -3.25746834e-01 -2.79126555e-01
-4.94812250e-01 -3.80423099e-01 5.10702193e-01 3.26353222e-01
-8.71027589e-01 6.26963973e-01 2.97095448e-01 -3.19366008e-01
-2.25666724e-02 -6.91620886e-01 2.75000900e-01 -7.01637685e-01
2.63600618e-01 3.01734120e-01 4.27674860e-01 -7.92149901e-01
3.98640186e-01 -1.29269019e-01 -1.77639496e+00 -4.87844460e-02
1.32091570e+00 6.85631633e-01 9.99552011e-01 -4.31827419e-02
-1.17956841e+00 -7.71489590e-02 -6.73094571e-01 -1.79686859e-01
1.06017448e-01 -7.50932097e-02 -8.34709585e-01 -3.46760541e-01
1.57403480e-02 3.42094451e-01 2.64772400e-02 1.41727936e+00
1.11532021e+00 2.29026288e-01 4.34152901e-01 -1.07314229e+00
-7.21476138e-01 8.96316290e-01 -8.10048044e-01 4.85498399e-01
4.95098740e-01 -1.76507488e-01 1.29249322e+00 -9.94942546e-01
4.34973538e-01 1.65600717e+00 4.84510005e-01 6.40793502e-01
-8.30396533e-01 -6.76942945e-01 4.38221067e-01 3.38104725e-01
-1.59278798e+00 -2.05996335e-01 5.19780636e-01 -1.41988903e-01
1.14782679e+00 1.25544459e-01 5.71137369e-01 1.05924106e+00
-6.06074259e-02 9.82578635e-01 1.23431385e+00 -6.55820727e-01
3.11118066e-01 3.69115435e-02 7.61622548e-01 1.07549481e-01
4.21877623e-01 -1.26548946e-01 -6.15154743e-01 -1.72031984e-01
1.68659464e-01 7.16866776e-02 -2.59056777e-01 1.23847142e-01
-1.10950029e+00 7.14181125e-01 1.70908403e-02 9.55430448e-01
-1.89511001e-01 -7.41157591e-01 6.97644293e-01 3.78392488e-01
4.53367859e-01 4.34189677e-01 -8.47629964e-01 -2.95551866e-01
-7.29771018e-01 3.12335789e-01 5.11716902e-01 9.61494565e-01
7.40031004e-01 3.92814074e-03 -2.95327961e-01 1.34057617e+00
2.42762148e-01 6.03171468e-01 1.37217140e+00 -5.72787113e-02
2.00497091e-01 8.18939388e-01 1.13336302e-01 -9.50085700e-01
-5.87106407e-01 -2.74744213e-01 -6.26873255e-01 -2.10495591e-01
5.15983105e-02 -2.07425103e-01 -1.00557578e+00 1.94843388e+00
4.86217767e-01 3.04503381e-01 2.33926401e-01 8.53856087e-01
1.07324088e+00 6.04288936e-01 4.90719289e-01 -5.08844733e-01
1.73894942e+00 -3.04101199e-01 -9.77731526e-01 -6.27852678e-01
6.85124099e-01 -1.03560448e+00 1.85451996e+00 -2.34777741e-02
-5.10754704e-01 -5.97819686e-01 -1.20008016e+00 1.72646940e-02
-7.99632072e-01 -2.11682484e-01 4.89938229e-01 6.23208880e-01
-6.18444622e-01 3.44636142e-01 -4.17557001e-01 -6.83523774e-01
1.58035345e-02 -1.75093025e-01 -1.65446401e-02 6.95285723e-02
-2.07596040e+00 1.02030444e+00 8.84245396e-01 -1.84774309e-01
-2.92550027e-01 -7.93725312e-01 -9.36979175e-01 -9.66250226e-02
4.46582615e-01 -3.48935723e-01 1.01960838e+00 -6.81954026e-01
-8.00976634e-01 1.09612155e+00 -2.73405850e-01 -1.15238391e-01
8.29533190e-02 -2.11537883e-01 -1.09721601e+00 -4.86869663e-01
6.58209443e-01 2.27222204e-01 4.37594920e-01 -1.19983327e+00
-1.06654012e+00 -3.70352387e-01 -1.23543553e-01 5.25983870e-01
-4.87345994e-01 -7.13710561e-02 -7.99330138e-03 -7.92931914e-01
3.53407860e-01 -5.71949482e-01 -1.70209318e-01 -7.58587182e-01
-4.36999977e-01 -9.60626900e-01 7.65620530e-01 -1.11709155e-01
1.78159416e+00 -2.23211837e+00 -3.45010012e-01 -8.46591741e-02
1.19912945e-01 7.29166269e-01 -2.25314796e-01 5.23413002e-01
-2.55565464e-01 1.71448231e-01 7.10226521e-02 -1.55243501e-01
2.04613045e-01 4.56460059e-01 -3.42748910e-01 -1.79068148e-02
-6.31963164e-02 4.46307451e-01 -1.45685661e+00 -7.79634416e-01
1.06909856e-01 -1.16126694e-01 4.73585166e-02 1.21832512e-01
2.53446810e-02 -1.92885488e-01 -8.29209268e-01 6.74136519e-01
8.53689134e-01 1.02688365e-01 2.48398751e-01 -3.32899332e-01
-2.63143778e-01 6.37088835e-01 -1.51970184e+00 2.04047871e+00
-4.68354404e-01 -1.43008009e-01 -1.92247719e-01 -6.81445479e-01
1.12735820e+00 3.68739188e-01 4.21960235e-01 -7.03243554e-01
9.33712944e-02 5.07726789e-01 7.31736943e-02 -8.14403057e-01
6.86986804e-01 -6.44875407e-01 -3.95663232e-01 3.41766119e-01
2.31773973e-01 1.13618188e-02 4.51967001e-01 2.59393156e-01
7.59352207e-01 -8.63583237e-02 6.28180087e-01 -8.75735044e-01
2.36917228e-01 -4.69834283e-02 1.10735774e+00 6.11725628e-01
-4.36496347e-01 3.04127306e-01 1.61361232e-01 -1.28939644e-01
-5.75834155e-01 -1.17619777e+00 -2.70205140e-01 1.27012157e+00
6.89361393e-01 -6.69317186e-01 -4.91602868e-01 -6.79105759e-01
-3.79290506e-02 1.00358915e+00 -2.12411553e-01 -3.28444809e-01
-1.03105865e-01 -8.08933437e-01 4.74868327e-01 2.05964029e-01
5.90720296e-01 -1.19646132e+00 -5.12939930e-01 3.30178231e-01
-3.36646289e-01 -7.59060502e-01 -5.77496111e-01 2.14298710e-01
-3.09138387e-01 -1.07352805e+00 -6.89739287e-01 -1.12162721e+00
1.19128540e-01 6.65672660e-01 1.35990536e+00 1.15000438e-02
-1.23964995e-01 -1.98835805e-01 -9.22097147e-01 -9.71472323e-01
1.78197622e-02 2.67832398e-01 6.02141261e-01 -2.01036468e-01
1.25142527e+00 -4.95743364e-01 -4.14007187e-01 8.39861706e-02
-9.10948753e-01 -4.73317891e-01 3.31944346e-01 9.57297742e-01
3.55915636e-01 1.30596057e-01 9.35382009e-01 -1.03568065e+00
9.36677217e-01 -5.53370357e-01 -1.45183474e-01 4.18852359e-01
-1.03164482e+00 -4.90892635e-05 2.64105141e-01 -5.96977413e-01
-1.32618308e+00 -2.67796576e-01 -2.88887739e-01 2.46833041e-02
-2.62686342e-01 7.74310350e-01 -4.06363428e-01 6.03029966e-01
6.78179979e-01 2.02742219e-01 -1.90943673e-01 -8.14046502e-01
1.37710810e-01 1.04078054e+00 1.01188481e-01 -6.03229582e-01
5.48873603e-01 -3.04753669e-02 -6.77463531e-01 -9.41044748e-01
-1.24035239e+00 -1.00952566e+00 -4.44775015e-01 1.50075138e-01
7.64227033e-01 -8.21189225e-01 -2.34969899e-01 6.23461962e-01
-1.09924161e+00 4.32066917e-01 -1.70682177e-01 5.15852928e-01
1.12595716e-02 5.07176757e-01 -3.82654846e-01 -1.03579557e+00
-5.18662035e-01 -8.17227900e-01 1.06400990e+00 5.54929554e-01
-4.65315610e-01 -8.75562429e-01 1.81281045e-01 -2.09032446e-01
2.98056275e-01 -1.80082470e-01 1.03096688e+00 -1.12762225e+00
2.67437130e-01 1.54633582e-01 7.53127411e-02 9.84049961e-02
4.84302253e-01 -3.23933333e-01 -1.06982994e+00 -1.57594994e-01
9.45730582e-02 -3.12148899e-01 8.51985455e-01 2.70884395e-01
5.22686601e-01 6.25558048e-02 -4.55427796e-01 9.28411260e-02
1.51212537e+00 3.05605233e-01 5.05252004e-01 2.96960056e-01
4.25813764e-01 5.80023527e-01 1.44344378e+00 4.20581281e-01
5.15247345e-01 2.92735845e-01 1.13555036e-01 -2.40183488e-01
7.06251338e-02 -3.53055954e-01 2.16548339e-01 1.21586120e+00
2.94892669e-01 -2.29943559e-01 -1.00231802e+00 9.18141127e-01
-1.77649915e+00 -9.84309435e-01 -1.36544555e-01 2.12787080e+00
1.22221911e+00 3.86215359e-01 5.66337444e-02 2.10710600e-01
1.14186800e+00 4.90640849e-01 -2.72560805e-01 -3.08452278e-01
-3.97737026e-01 1.21205606e-01 2.08888829e-01 3.73482555e-01
-1.00291073e+00 1.29815805e+00 5.43132877e+00 1.63120854e+00
-8.66784990e-01 2.21618250e-01 1.30682543e-01 3.51329237e-01
-5.67239046e-01 1.07029684e-01 -8.58807325e-01 8.03864300e-01
3.61978978e-01 -5.04418194e-01 -1.47884950e-01 7.25678742e-01
1.20308600e-01 -1.68094382e-01 -4.02962804e-01 1.33051252e+00
-1.97480723e-01 -7.88855910e-01 1.68002620e-01 -2.76698440e-01
4.25060004e-01 1.49549499e-01 -4.22410816e-01 5.43255150e-01
2.72963941e-01 -6.58162951e-01 6.16865218e-01 2.80529290e-01
9.09829557e-01 -7.68070579e-01 9.48961198e-01 5.38511157e-01
-1.46749091e+00 5.57097793e-02 -8.98232102e-01 -1.66628450e-01
2.41464987e-01 1.01238716e+00 -7.83134580e-01 8.43048751e-01
7.14337409e-01 7.60931194e-01 -2.81492025e-01 8.18364620e-01
-9.04762000e-02 6.09940708e-01 -2.09343180e-01 -5.34902990e-01
1.93052426e-01 -1.02695726e-01 8.08855236e-01 1.22490466e+00
4.26413000e-01 2.69758224e-01 5.92221260e-01 5.26210725e-01
5.26962757e-01 3.93484741e-01 -5.38895369e-01 -7.48546273e-02
1.15058208e+00 1.23520398e+00 -6.05277956e-01 -4.40696448e-01
-3.09866428e-01 7.25735247e-01 5.89880906e-02 1.72030464e-01
-2.87971169e-01 -7.70501912e-01 1.05250704e+00 -7.40787461e-02
8.84933844e-02 -7.80031979e-02 -3.70433152e-01 -1.27788091e+00
-6.43160865e-02 -6.24174476e-01 8.29223454e-01 -4.87937421e-01
-1.87597358e+00 6.18420899e-01 8.61521903e-03 -1.40813744e+00
1.86595600e-02 -5.61830461e-01 -8.78778934e-01 9.72297549e-01
-1.68571210e+00 -9.48255181e-01 2.06385162e-02 6.00845516e-01
7.75749981e-01 -2.71660596e-01 1.01848602e+00 4.63699967e-01
-3.33736956e-01 6.63957536e-01 -1.83742642e-01 -2.41074315e-03
8.17815423e-01 -1.52804530e+00 2.20579922e-01 7.69871593e-01
4.53209644e-03 8.65226686e-01 9.50359225e-01 -8.57460141e-01
-9.30762947e-01 -7.28210270e-01 1.56302214e+00 -4.39641662e-02
6.21123135e-01 -8.23851451e-02 -9.44855273e-01 -7.15172589e-02
1.15924679e-01 -3.87152702e-01 6.98578537e-01 6.61947846e-01
-3.15070987e-01 -2.38338932e-01 -1.04935241e+00 6.47814631e-01
1.18360817e+00 -4.61257011e-01 -1.41707301e+00 -2.23521783e-04
1.02774405e+00 -9.87751782e-02 -4.53020900e-01 3.74513388e-01
1.33448198e-01 -7.25786388e-01 9.25396800e-01 -6.28037810e-01
-1.19934455e-02 -4.85472679e-01 -2.59889126e-01 -1.63680828e+00
-1.90107256e-01 -5.24602592e-01 4.23868001e-01 1.87358952e+00
4.76770163e-01 -6.48778915e-01 3.48277651e-02 1.85525860e-03
-1.52333587e-01 -4.97117758e-01 -1.06289840e+00 -8.33884656e-01
3.78149329e-04 -4.45602059e-01 1.06748950e+00 1.43474364e+00
4.60849702e-01 9.29389060e-01 -1.08177617e-01 -1.63397223e-01
2.00593755e-01 2.67995596e-01 -3.19106467e-02 -1.24801409e+00
1.30234838e-01 -3.08587879e-01 -3.65134954e-01 -9.00016665e-01
2.08342969e-01 -8.67234886e-01 2.55386084e-01 -1.26050615e+00
2.25821972e-01 -7.41236687e-01 -8.61180961e-01 5.91030359e-01
-8.77723098e-01 -2.67810047e-01 5.47667481e-02 -5.97970262e-02
-5.54143369e-01 5.66537321e-01 1.18311596e+00 -4.45228629e-02
-1.25148788e-01 -1.72675207e-01 -9.06645596e-01 7.49438405e-01
6.78727090e-01 -3.92558634e-01 -6.04702532e-01 -3.33494335e-01
2.59609163e-01 -2.42979720e-01 -2.04015166e-01 -6.03739560e-01
1.61525071e-01 -3.47620934e-01 -1.19717322e-01 -5.30389726e-01
-1.44744394e-02 -4.36949611e-01 -5.39028943e-01 4.50163990e-01
-1.03482433e-01 2.92554915e-01 -8.23836178e-02 4.06796604e-01
-2.31290951e-01 -4.36223865e-01 5.81830502e-01 -4.37439173e-01
-1.34823155e+00 2.05091923e-01 -1.79084539e-01 7.19428897e-01
6.36878669e-01 -2.59462118e-01 -7.10503161e-02 7.83658549e-02
-4.39865321e-01 4.71167266e-01 1.78175509e-01 9.42607939e-01
4.60352719e-01 -1.55209756e+00 -7.76511014e-01 1.23016141e-01
5.54644763e-01 -1.30584091e-01 4.71414536e-01 3.42553318e-01
2.67498344e-01 -8.74878913e-02 8.11643004e-02 -3.90680522e-01
-1.07959330e+00 6.39664471e-01 4.24342789e-02 -2.23817602e-01
-3.43758881e-01 8.87461066e-01 1.05746612e-01 -7.18747258e-01
4.39648814e-02 1.71646494e-02 -6.51533961e-01 6.29861534e-01
7.37270892e-01 1.96168005e-01 3.89207751e-02 -6.83083236e-01
-6.14272833e-01 4.13940519e-01 -2.25150548e-02 1.35494694e-01
1.13214493e+00 -6.07009113e-01 -7.17340782e-02 8.49677742e-01
9.31151390e-01 1.18998080e-01 -5.50429523e-01 -4.78103280e-01
4.41083938e-01 -4.71022606e-01 -2.05408316e-02 -9.38771367e-01
-6.88950598e-01 7.40028739e-01 8.44166398e-01 6.22651875e-01
9.12632108e-01 7.70642832e-02 1.26242065e+00 1.45251364e-01
6.81898534e-01 -1.66821063e+00 -3.51739049e-01 8.62549245e-01
3.82744461e-01 -1.24334610e+00 -4.39087719e-01 -3.80714506e-01
-8.39989305e-01 8.85441184e-01 6.89012468e-01 -3.25794443e-02
8.76867712e-01 2.80547708e-01 4.69756037e-01 -2.67900348e-01
-6.02742195e-01 -8.61003160e-01 2.06680983e-01 7.12719500e-01
5.68766117e-01 2.87894279e-01 -1.13551831e+00 9.56242085e-01
-2.71128803e-01 -3.01191211e-01 1.38660312e-01 8.72111857e-01
-8.87174249e-01 -1.39663267e+00 -1.29539743e-01 5.11084080e-01
-4.55974638e-01 -5.65538764e-01 -1.11430943e-01 5.91213942e-01
5.37324369e-01 1.26019561e+00 2.16372050e-02 -5.87685704e-01
3.83124292e-01 2.65094340e-01 -2.40805466e-02 -8.93522918e-01
-2.82195181e-01 6.05962873e-02 1.72933772e-01 -3.86529177e-01
-5.32155931e-01 -6.55638635e-01 -1.55819786e+00 -1.43298969e-01
-6.10639989e-01 3.40227455e-01 9.82922167e-02 1.35031080e+00
2.48829126e-01 2.90760040e-01 6.85612798e-01 -1.86620489e-01
-8.87353718e-01 -1.16659701e+00 -1.26366830e+00 8.45776320e-01
1.38678983e-01 -9.37985897e-01 -2.09153742e-01 -4.10882354e-01] | [10.222604751586914, 9.077654838562012] |
7a73676e-9f9d-4a49-b683-6ed4a8d4d96b | cin-enhancing-topological-message-passing | 2306.03561 | null | https://arxiv.org/abs/2306.03561v1 | https://arxiv.org/pdf/2306.03561v1.pdf | CIN++: Enhancing Topological Message Passing | Graph Neural Networks (GNNs) have demonstrated remarkable success in learning from graph-structured data. However, they face significant limitations in expressive power, struggling with long-range interactions and lacking a principled approach to modeling higher-order structures and group interactions. Cellular Isomorphism Networks (CINs) recently addressed most of these challenges with a message passing scheme based on cell complexes. Despite their advantages, CINs make use only of boundary and upper messages which do not consider a direct interaction between the rings present in the underlying complex. Accounting for these interactions might be crucial for learning representations of many real-world complex phenomena such as the dynamics of supramolecular assemblies, neural activity within the brain, and gene regulation processes. In this work, we propose CIN++, an enhancement of the topological message passing scheme introduced in CINs. Our message passing scheme accounts for the aforementioned limitations by letting the cells to receive also lower messages within each layer. By providing a more comprehensive representation of higher-order and long-range interactions, our enhanced topological message passing scheme achieves state-of-the-art results on large-scale and long-range chemistry benchmarks. | ['Pietro Liò', 'Cristian Bodnar', 'Francesco Ceccarelli', 'Teodora Reu', 'Lorenzo Giusti'] | 2023-06-06 | null | null | null | null | ['graph-regression', 'graph-classification'] | ['graphs', 'graphs'] | [ 1.90330535e-01 2.71901965e-01 -1.10680051e-01 7.17250630e-02
2.48066232e-01 -5.99021196e-01 9.84347999e-01 8.45277607e-01
-3.48161429e-01 8.51264119e-01 5.17009050e-02 -6.15859509e-01
-4.77039546e-01 -1.32590318e+00 -1.15618217e+00 -7.55173206e-01
-6.85708165e-01 6.04953408e-01 5.05855918e-01 -4.96135086e-01
1.87773541e-01 8.24644625e-01 -1.37268138e+00 3.64576906e-01
8.81612778e-01 6.67790353e-01 5.89068867e-02 6.26892507e-01
-3.70351434e-01 9.09843326e-01 -1.59419209e-01 1.70500868e-03
-1.55268416e-01 -3.46973687e-01 -7.80784428e-01 -4.39488649e-01
4.26286697e-01 3.39087546e-01 -5.74654162e-01 6.49226248e-01
3.73637766e-01 1.12131275e-01 8.72660577e-01 -1.03933382e+00
-5.27093709e-01 6.07928991e-01 -1.31248757e-01 1.52685285e-01
2.05364510e-01 -4.62942272e-02 1.12028635e+00 -6.81016922e-01
1.02968013e+00 1.10700524e+00 5.93507826e-01 4.96990234e-01
-1.66532755e+00 -6.34641171e-01 3.41815501e-01 9.00124535e-02
-1.07873702e+00 -1.38927415e-01 6.84211969e-01 -5.16197979e-01
1.13950944e+00 2.16499224e-01 1.01283896e+00 7.73942947e-01
6.20759130e-01 7.98319429e-02 1.01943350e+00 -9.17841569e-02
3.31283450e-01 -5.96359253e-01 3.69673699e-01 9.16233718e-01
4.77991462e-01 -8.54049772e-02 -5.68285525e-01 -1.98847860e-01
1.05907369e+00 1.53793827e-01 -2.06278548e-01 -5.71733057e-01
-1.32339036e+00 7.16752231e-01 1.02414119e+00 6.07230902e-01
-8.69475007e-02 4.81418490e-01 3.06617826e-01 2.08636463e-01
1.80080310e-01 7.55403221e-01 -2.93734491e-01 3.92630786e-01
-3.97554457e-01 1.67001396e-01 9.26969528e-01 6.38882995e-01
1.01106763e+00 -3.14155221e-01 1.89006899e-03 5.24119556e-01
1.24449380e-01 -2.94922531e-01 1.53166763e-02 -5.75894356e-01
1.62704498e-01 1.04385734e+00 -4.98095274e-01 -1.04594719e+00
-8.80360305e-01 -6.56604886e-01 -1.47732031e+00 8.88449699e-02
5.58198869e-01 3.51417154e-01 -8.82615268e-01 1.77355516e+00
3.67149591e-01 3.74638677e-01 -5.27089424e-02 4.38874185e-01
1.15488398e+00 7.02981293e-01 3.76175672e-01 -8.83479230e-03
1.36700654e+00 -8.19854200e-01 -3.43154162e-01 1.93511337e-01
9.88246918e-01 -1.98754713e-01 6.47832274e-01 8.97319540e-02
-1.14853048e+00 -2.76270330e-01 -9.73005593e-01 -3.54171157e-01
-9.28968668e-01 -4.86143231e-01 1.29309189e+00 2.40208372e-01
-1.23921835e+00 9.82260823e-01 -5.44880807e-01 -3.45894605e-01
5.00249386e-01 9.50799465e-01 -5.95880687e-01 -1.20448451e-02
-1.22754109e+00 6.30852103e-01 2.90747315e-01 -2.01113243e-02
-5.31399429e-01 -1.04552102e+00 -8.22325706e-01 3.36345166e-01
2.86135912e-01 -9.12127852e-01 5.12585044e-01 -3.24450791e-01
-1.27100909e+00 7.12505102e-01 8.53884444e-02 -5.19818664e-01
2.60333836e-01 6.18895352e-01 -5.35305664e-02 6.89499676e-02
-4.35517162e-01 8.85557294e-01 2.34378129e-02 -1.03241646e+00
2.03115225e-01 -3.16971481e-01 3.72953832e-01 8.57486278e-02
-1.94369093e-01 -4.57779855e-01 -3.69169228e-02 -4.58329737e-01
3.31235111e-01 -9.45709467e-01 -3.98192555e-01 1.35003716e-01
-3.15712780e-01 -3.14755529e-01 6.54999793e-01 6.10893108e-02
9.64353681e-01 -1.77129149e+00 5.34157157e-01 3.87107491e-01
9.42764997e-01 9.85120609e-02 -2.62406141e-01 9.10326958e-01
-2.18161479e-01 3.42627019e-01 -2.03266189e-01 4.13054228e-02
-2.41652690e-02 2.03931153e-01 -7.01650186e-03 2.36964509e-01
1.64778084e-01 1.18001008e+00 -1.00295150e+00 -2.64611930e-01
1.89303622e-01 6.91148341e-01 -7.41225421e-01 -2.47308403e-01
-6.08359218e-01 6.41182780e-01 -2.62671918e-01 2.49447197e-01
6.12249196e-01 -7.76553690e-01 6.55149579e-01 -1.93733543e-01
-2.05893353e-01 4.32962596e-01 -9.30981636e-01 1.54664469e+00
-1.70802638e-01 3.77658069e-01 6.06638975e-02 -1.05406010e+00
7.18754470e-01 1.02841571e-01 7.42617726e-01 -7.56242156e-01
-6.97183162e-02 1.78561196e-01 4.44024980e-01 1.17493533e-01
1.58067808e-01 1.76702477e-02 2.31216401e-01 2.82837957e-01
1.00108139e-01 -2.69474811e-03 3.88589472e-01 6.44639313e-01
1.30202365e+00 -2.12678879e-01 1.03425607e-01 -7.37481833e-01
6.42916501e-01 -3.66759688e-01 2.41015464e-01 7.79796660e-01
2.83226758e-01 3.16417545e-01 9.81000483e-01 -7.38490105e-01
-9.74217415e-01 -9.10566568e-01 -3.65706421e-02 1.07028556e+00
1.01062328e-01 -8.36377144e-01 -6.61834896e-01 -1.14740908e-01
8.14946145e-02 -1.69391945e-01 -8.45550537e-01 -2.75009364e-01
-5.47735631e-01 -9.33610022e-01 6.27974868e-01 2.33656779e-01
3.36538136e-01 -8.37208033e-01 -9.27179158e-02 4.97619539e-01
2.28036463e-01 -1.02592003e+00 -3.46233815e-01 4.11309183e-01
-9.35184121e-01 -1.37674475e+00 -3.94363523e-01 -8.74315381e-01
9.10968840e-01 1.96813375e-01 1.10624099e+00 5.60770690e-01
-5.11179328e-01 -3.39341350e-02 2.21332729e-01 -1.44954339e-01
-4.56772476e-01 4.61457878e-01 -8.50715339e-02 -1.09114476e-01
-3.01179022e-01 -1.05613256e+00 -7.26408184e-01 2.34688625e-01
-1.21227515e+00 4.64106441e-01 3.95214587e-01 8.52082253e-01
5.16889155e-01 -1.60833299e-01 5.82753539e-01 -1.05577183e+00
4.77902114e-01 -4.68382269e-01 -6.26618564e-01 2.74114311e-01
-5.96699774e-01 4.42486435e-01 8.05817425e-01 -4.60635632e-01
-4.69280779e-01 -9.44413319e-02 -6.70260787e-02 3.22533339e-01
7.56255090e-02 5.75001121e-01 -2.00593635e-03 -6.12918437e-01
4.00725365e-01 1.13518998e-01 8.84235278e-02 -2.78123736e-01
1.42660871e-01 -8.27246904e-02 1.99351743e-01 -1.06365728e+00
4.52901274e-01 3.75538290e-01 1.00461400e+00 -9.82880712e-01
-2.63895154e-01 -1.87644109e-01 -7.67450511e-01 -1.96147542e-02
8.26271117e-01 -6.20084167e-01 -1.50616300e+00 6.92040265e-01
-1.13524091e+00 -4.78385508e-01 -4.46316935e-02 1.70046672e-01
-3.46276134e-01 2.86027431e-01 -9.47169602e-01 -3.68803948e-01
-2.15493202e-01 -1.04874706e+00 7.93519855e-01 2.06731949e-02
6.93232492e-02 -1.30017567e+00 2.39573464e-01 -7.03259408e-02
6.55205786e-01 6.13358259e-01 1.65493143e+00 -4.57111716e-01
-1.00952339e+00 2.35128462e-01 -2.74874181e-01 -5.07663488e-01
-7.08248094e-02 -3.48252654e-02 -5.49335897e-01 -2.20191523e-01
-8.69661093e-01 -1.70016736e-01 1.15226007e+00 1.43296510e-01
1.20177341e+00 -2.28230864e-01 -4.73142058e-01 5.31267226e-01
1.32617474e+00 -1.26634523e-01 8.57954681e-01 -3.14467736e-02
1.01584613e+00 7.45075524e-01 -6.44833505e-01 -8.96045119e-02
5.11364281e-01 6.07658923e-01 5.64790308e-01 -3.02445203e-01
-1.25717953e-01 -3.14899474e-01 6.73390329e-02 1.11150301e+00
-3.03552240e-01 -2.65389234e-01 -1.03765988e+00 4.30991352e-02
-1.83564258e+00 -9.22733724e-01 -5.22850335e-01 2.13477850e+00
8.63395333e-01 2.19327211e-01 -4.14452665e-02 -1.93657819e-02
5.41769564e-01 2.59615541e-01 -6.26018107e-01 -3.12170774e-01
-4.42304283e-01 3.83627415e-01 3.91081244e-01 5.52085102e-01
-7.09976137e-01 8.57316315e-01 6.58192730e+00 5.72727323e-01
-1.10777128e+00 -1.89947307e-01 6.30861104e-01 1.35672972e-01
-5.56324720e-01 2.30920434e-01 -6.48073554e-01 1.66866452e-01
9.47232842e-01 1.26315162e-01 6.88932598e-01 1.25587499e-02
-7.47907311e-02 4.12979014e-02 -1.24635887e+00 6.50111854e-01
-4.92918044e-01 -2.22664571e+00 3.99266034e-01 4.49896872e-01
7.38533437e-01 1.95371002e-01 -2.80263186e-01 8.91031921e-02
3.16252083e-01 -1.37611604e+00 3.42611074e-01 5.66239715e-01
7.45805323e-01 -5.08586526e-01 2.41115883e-01 2.33639836e-01
-1.46646070e+00 1.67132288e-01 -2.73962438e-01 -4.18143213e-01
-3.87464255e-01 7.39560485e-01 -3.92376184e-01 6.02943957e-01
1.08129330e-01 6.92336679e-01 -6.60522997e-01 7.66633809e-01
2.32358977e-01 3.10266793e-01 -3.59222412e-01 -2.96470553e-01
3.55725437e-01 -4.23831046e-01 1.47788107e-01 1.16844380e+00
-2.91260658e-03 1.93075746e-01 1.69783786e-01 9.26668108e-01
-5.71591556e-01 6.84814155e-02 -9.36949372e-01 -2.71880746e-01
2.91836888e-01 1.20803106e+00 -1.09597194e+00 -1.26029462e-01
-1.61866307e-01 3.07447016e-01 7.97446907e-01 3.33423704e-01
-5.20297050e-01 -2.65932113e-01 6.57608628e-01 5.77179193e-01
-9.85494349e-03 -6.16953313e-01 -8.21669325e-02 -1.07017684e+00
-2.37079218e-01 -5.41549265e-01 3.05855155e-01 -5.08967757e-01
-9.82858717e-01 2.42946565e-01 -2.92056859e-01 -5.15781343e-01
4.30675834e-01 -8.99307728e-01 -6.41615093e-01 5.16894877e-01
-1.51219440e+00 -1.20655966e+00 -2.47666061e-01 3.72941226e-01
-1.46032691e-01 3.32065672e-01 9.91880059e-01 2.74420679e-01
-3.88728857e-01 1.19264163e-01 1.38971105e-01 -7.68589750e-02
3.05078834e-01 -1.25769305e+00 4.75096345e-01 1.17788136e-01
-1.08360462e-02 1.01224256e+00 2.99824625e-01 -5.80091953e-01
-1.78306770e+00 -9.75790024e-01 8.88546526e-01 -2.68559992e-01
7.12254822e-01 -1.00229323e+00 -1.08652759e+00 4.02365178e-01
5.10145426e-02 2.16309786e-01 6.02040768e-01 3.02670568e-01
-6.38148189e-01 -1.04229607e-01 -7.95483351e-01 6.97245538e-01
1.54452145e+00 -6.41749263e-01 1.55939702e-02 5.00369728e-01
7.70598233e-01 -3.40429515e-01 -1.14324844e+00 3.38343710e-01
5.55752039e-01 -7.93245137e-01 1.06845140e+00 -8.00958216e-01
4.39546078e-01 -3.09754670e-01 2.22938105e-01 -1.13201189e+00
-5.61679602e-01 -7.91581929e-01 -1.78965911e-01 7.67672718e-01
4.10740018e-01 -1.03794181e+00 7.10476160e-01 2.53406316e-01
-1.95011750e-01 -1.08748293e+00 -8.85181427e-01 -5.68191886e-01
4.41481143e-01 1.85195401e-01 5.93534291e-01 1.07888663e+00
1.39021963e-01 5.54027796e-01 -2.72882227e-02 -1.68045402e-01
5.92283547e-01 -1.82115436e-01 5.42983830e-01 -1.80031776e+00
-1.92460924e-01 -6.66282594e-01 -5.41576266e-01 -8.27072263e-01
1.84514403e-01 -1.27302027e+00 -4.37990963e-01 -1.70681238e+00
2.69913614e-01 -5.33859015e-01 -4.27894413e-01 2.72630572e-01
2.65601844e-01 1.58777371e-01 1.36883721e-01 1.70613736e-01
-7.69238353e-01 5.23928583e-01 1.53457475e+00 -3.13413471e-01
3.53971384e-02 -6.38998866e-01 -5.72371960e-01 4.49734360e-01
7.39582658e-01 -3.77096564e-01 -2.65007406e-01 -3.34448248e-01
7.64886618e-01 -8.45448747e-02 6.02548063e-01 -1.13517499e+00
7.87062466e-01 -1.31220192e-01 3.83747876e-01 -3.12151253e-01
3.20518762e-01 -5.46837926e-01 5.72961450e-01 8.59272659e-01
-5.00928581e-01 1.15227602e-01 1.68399557e-01 7.47738481e-01
9.60775763e-02 3.91708463e-01 7.36971498e-01 -2.86376208e-01
-1.48678184e-01 5.72510540e-01 -5.65454245e-01 -1.58308819e-01
8.66711259e-01 -1.48803771e-01 -9.01179314e-01 -1.36141106e-01
-8.02864969e-01 2.33891383e-01 6.03814065e-01 1.71794996e-01
3.08672875e-01 -1.13158417e+00 -3.03508133e-01 1.99703723e-01
1.53766975e-01 1.41561866e-01 1.88098639e-01 7.71861136e-01
-6.71384454e-01 7.00829446e-01 -3.30708027e-01 -4.04954642e-01
-9.93769526e-01 3.80292535e-01 4.86546427e-01 -7.88846910e-01
-3.86790097e-01 4.21439439e-01 5.85125029e-01 -7.42681205e-01
6.71350509e-02 -5.36877513e-01 -1.63200557e-01 -8.34163204e-02
1.11675754e-01 2.68357426e-01 8.09663236e-02 -4.42925602e-01
-2.76567459e-01 6.75898671e-01 -1.31597325e-01 3.38864863e-01
1.45891893e+00 3.63973618e-01 -7.45069206e-01 3.68410438e-01
1.07151771e+00 -1.60779521e-01 -1.04664207e+00 -1.78713903e-01
1.59064438e-02 3.54630530e-01 -2.19749689e-01 -5.99413216e-01
-7.08959520e-01 9.23301876e-01 8.87569338e-02 3.44940573e-01
5.99981308e-01 1.15507148e-01 5.97092867e-01 8.38681161e-01
5.42011678e-01 -7.37447739e-01 2.37053618e-01 8.11854362e-01
5.79825699e-01 -6.06099427e-01 4.43673655e-02 -7.67054677e-01
3.28266919e-01 1.15486813e+00 5.82081854e-01 -1.60193726e-01
7.25170314e-01 1.78136319e-01 -6.33488238e-01 -5.96146762e-01
-1.01141441e+00 4.85679694e-03 1.99529082e-01 4.29205656e-01
6.01768196e-01 3.62217613e-02 -4.35455680e-01 5.71526065e-02
1.80661038e-01 -2.35370442e-01 5.77224672e-01 8.55601907e-01
-4.58170772e-01 -1.36821616e+00 1.17159702e-01 5.33383727e-01
-3.43428344e-01 -2.56651640e-01 -6.37023807e-01 7.07174182e-01
9.56667140e-02 4.03139919e-01 5.04399501e-02 -1.28439248e-01
1.25882044e-01 1.38049889e-02 7.77263582e-01 -6.04888439e-01
-6.11231089e-01 -2.75520265e-01 -7.88800791e-02 -3.76835555e-01
-3.94032866e-01 -2.04294428e-01 -1.62069952e+00 -6.21092439e-01
-1.31149322e-01 3.58828634e-01 6.99949324e-01 6.94691420e-01
9.12758827e-01 8.36718619e-01 1.77731439e-01 -9.34474409e-01
-1.34607479e-01 -5.06761432e-01 -5.55131137e-01 3.75082821e-01
3.42024714e-01 -6.78116679e-01 -3.43795046e-02 -2.86892831e-01] | [6.714847564697266, 6.0554890632629395] |
67246739-96b9-46ef-8736-85956e2d3785 | attention-mechanism-transformers-bert-and-gpt | null | null | https://osf.io/m6gcn | https://osf.io/m6gcn/download | Attention Mechanism, Transformers, BERT, and GPT: Tutorial and Survey | This is a tutorial and survey paper on the attention mechanism, transformers, BERT, and GPT. We first explain attention mechanism, sequence-to-sequence model without and with attention, self-attention, and attention in different areas such as natural language processing and computer vision. Then, we explain transformers which do not use any recurrence. We explain all the parts of encoder and decoder in the transformer, including positional encoding, multihead self-attention and cross-attention, and masked multihead attention. Thereafter, we introduce the Bidirectional Encoder Representations from Transformers (BERT) and Generative Pre-trained Transformer (GPT) as the stacks of encoders and decoders of transformer, respectively. We explain their characteristics and how they work. | ['Ali Ghodsi', 'Benyamin Ghojogh'] | 2020-11-17 | null | null | null | null | ['deep-attention', 'deep-attention'] | ['computer-vision', 'natural-language-processing'] | [ 1.80175155e-01 3.76505017e-01 -5.12840375e-02 -7.45869875e-02
-6.26882195e-01 -2.96975642e-01 5.00795126e-01 -3.01624328e-01
1.21401764e-01 6.63411617e-01 7.36314535e-01 -4.73476350e-01
2.34775111e-01 -7.61301994e-01 -6.68646276e-01 -7.36724019e-01
1.66603237e-01 4.61304218e-01 2.14087054e-01 -3.81855339e-01
4.23267096e-01 3.91986854e-02 -1.21887660e+00 3.27989072e-01
9.14888144e-01 8.77400696e-01 5.64675748e-01 1.05721343e+00
-4.44669306e-01 1.27682769e+00 -5.16955376e-01 -7.51243412e-01
-5.27646244e-01 -8.23531568e-01 -1.12119365e+00 -1.48467273e-01
-2.50709265e-01 -1.85859829e-01 -5.99580824e-01 8.19860339e-01
6.75777793e-01 7.69152641e-02 8.48334968e-01 -1.27445161e+00
-1.73584485e+00 1.01532352e+00 -6.06750548e-01 7.72806168e-01
3.29962283e-01 -1.39834836e-01 1.15682936e+00 -9.40704465e-01
2.63518184e-01 1.35549426e+00 7.99211204e-01 5.09485424e-01
-6.92887485e-01 -2.29982838e-01 3.16003144e-01 6.23746097e-01
-1.39179134e+00 -3.56588155e-01 3.38041335e-01 -5.67088604e-01
1.70389652e+00 1.82477534e-01 8.09049368e-01 8.84805202e-01
6.53548419e-01 1.04635048e+00 2.12334141e-01 -5.69000781e-01
-2.38443166e-01 -1.67023033e-01 3.29056293e-01 6.81441128e-01
-1.84520613e-02 -6.16149716e-02 -5.12895405e-01 1.60453647e-01
1.01903510e+00 -4.12936322e-02 -2.46087179e-01 4.03635614e-02
-1.00886619e+00 7.63792038e-01 4.48618710e-01 4.12658423e-01
-6.91029727e-01 3.80963027e-01 3.31309080e-01 8.69218558e-02
4.47021186e-01 -1.25036374e-01 -3.93314004e-01 -2.18418658e-01
-4.06770557e-01 -3.05923671e-01 3.61895919e-01 1.73970532e+00
7.04910994e-01 5.61913550e-01 -6.66461647e-01 8.33033979e-01
2.06869841e-01 3.75025630e-01 1.06427193e+00 -1.90975145e-01
5.02219737e-01 2.77364433e-01 -4.10640746e-01 -5.71341515e-01
-2.55728257e-03 -4.51034635e-01 -1.06003761e+00 -6.37867451e-01
-5.38570344e-01 -1.92972168e-01 -1.06257486e+00 1.64773941e+00
-2.83755064e-01 1.78848058e-01 2.52317656e-02 5.44134319e-01
1.06683528e+00 8.36737156e-01 8.68901461e-02 -1.39921054e-01
1.47267985e+00 -1.46457446e+00 -1.08473802e+00 -4.49438512e-01
3.71527374e-01 -6.39947236e-01 8.39107871e-01 -2.46890843e-01
-1.54758382e+00 -7.65020311e-01 -6.47854269e-01 -6.48586452e-01
-3.22403491e-01 1.02153651e-01 4.04481530e-01 2.97581822e-01
-1.40171504e+00 3.66263777e-01 -6.48858070e-01 -6.31077051e-01
3.45490456e-01 3.75558078e-01 -1.02256946e-01 2.95671195e-01
-1.48338425e+00 1.15582454e+00 3.70191127e-01 4.63165529e-02
-1.30432773e+00 -3.49024206e-01 -1.13416636e+00 5.71244001e-01
-1.04122289e-01 -9.17760670e-01 1.56860280e+00 -6.92677796e-01
-1.48848188e+00 6.34914994e-01 -7.06321120e-01 -5.96675396e-01
-2.11877659e-01 -5.23219228e-01 -2.45767489e-01 -3.56075794e-01
1.22288726e-01 5.81100464e-01 6.09253526e-01 -5.18393874e-01
-7.03001976e-01 5.20965718e-02 -3.70866172e-02 1.84943154e-01
-3.24030146e-02 1.95489705e-01 -7.51855195e-01 -6.41484976e-01
-1.54506251e-01 -5.36496937e-01 -3.65216751e-03 -7.08176732e-01
-5.63353479e-01 -3.34382296e-01 7.87780046e-01 -9.06168640e-01
1.64242899e+00 -2.32555842e+00 4.65725511e-01 -6.34862721e-01
1.61325172e-01 1.95986345e-01 -2.06570148e-01 9.18271840e-01
-5.07489920e-01 3.25414091e-01 -1.69916019e-01 -3.11013639e-01
-9.19003412e-02 4.65041280e-01 -4.65749949e-01 1.26156762e-01
3.49549413e-01 1.47524369e+00 -9.68827724e-01 -5.93195736e-01
1.11596920e-01 5.78704059e-01 -6.16009355e-01 7.36908138e-01
9.81110260e-02 3.04825813e-01 -2.45352507e-01 6.39616966e-01
2.81184822e-01 -4.75908190e-01 -2.05660671e-01 -1.44319385e-01
-3.25234264e-01 9.11917627e-01 -1.97169587e-01 1.33866286e+00
-5.14473915e-01 8.77195418e-01 -2.33790725e-01 -9.17324901e-01
8.34969282e-01 7.75146067e-01 -1.66075140e-01 -6.27607524e-01
2.62982279e-01 -1.76404417e-02 1.04949594e-01 -7.69952834e-01
6.18329644e-01 -1.88487098e-01 1.73738271e-01 5.33514977e-01
6.35846853e-01 2.77568877e-01 3.60903442e-01 3.13164085e-01
9.21982527e-01 9.42234918e-02 7.30025351e-01 -1.91704616e-01
7.10415423e-01 -5.80428779e-01 4.55888391e-01 6.75139070e-01
4.52660173e-02 7.48278022e-01 6.72895849e-01 -2.40379184e-01
-1.30819619e+00 -6.99129701e-01 3.76271635e-01 1.55599749e+00
-8.23204219e-02 -6.60458565e-01 -7.11613536e-01 -4.40924346e-01
-5.25980711e-01 7.75661469e-01 -8.92877758e-01 -3.99748772e-01
-4.57619011e-01 -6.16137326e-01 5.36517918e-01 1.21271718e+00
5.52825093e-01 -1.52274263e+00 -5.10729432e-01 1.72164097e-01
-5.02201200e-01 -5.73543787e-01 -8.18370283e-01 4.05371368e-01
-5.76390147e-01 -7.63012171e-01 -9.74639475e-01 -1.11540365e+00
5.00450730e-01 1.18507981e-01 1.31912434e+00 5.13549559e-02
-6.47833943e-02 1.04722187e-01 -3.57054949e-01 -3.62769872e-01
-1.14630915e-01 4.29494381e-01 -3.73235494e-01 -1.26901492e-01
3.12486410e-01 -6.92336023e-01 -1.99292347e-01 7.03990227e-03
-4.38234955e-01 2.22289622e-01 1.06013608e+00 9.54662442e-01
3.63597840e-01 -4.58445102e-01 1.48069426e-01 -7.59768009e-01
9.57965970e-01 -6.15466058e-01 -2.03904986e-01 4.45766240e-01
-2.67884851e-01 2.57395208e-01 2.99232215e-01 -1.63331538e-01
-9.90167141e-01 -3.22852045e-01 -7.25055516e-01 -4.83134806e-01
3.09527695e-01 5.32761633e-01 -3.64710331e-01 1.91033125e-01
1.81540728e-01 7.89282739e-01 -2.59855062e-01 -5.75123847e-01
5.18510580e-01 8.92809868e-01 3.83180201e-01 -2.41397575e-01
1.79043099e-01 -1.40590236e-01 -6.10974312e-01 -5.18203199e-01
-7.79738426e-01 -2.63803273e-01 -7.17636704e-01 8.95754173e-02
1.11875641e+00 -6.85318351e-01 -6.31965756e-01 4.10436839e-01
-1.59006870e+00 -3.08285624e-01 -3.92198205e-01 2.83923388e-01
-8.80999684e-01 7.29757026e-02 -1.08453178e+00 -9.64156389e-01
-7.83169508e-01 -1.39821482e+00 1.12564170e+00 2.33389437e-01
-2.35294342e-01 -1.16489077e+00 1.45555153e-01 -1.98448479e-01
7.81345248e-01 -4.58017290e-01 1.06982410e+00 -6.37887716e-01
-6.37123883e-01 3.46414685e-01 -2.17731014e-01 3.60955119e-01
1.89636409e-01 -1.00030735e-01 -9.03718710e-01 -3.63857858e-02
-3.86838503e-02 -3.14035304e-02 1.02053022e+00 4.98085856e-01
1.18436325e+00 -4.33004111e-01 -6.96272075e-01 7.92816103e-01
1.16475677e+00 7.61232257e-01 1.09129310e+00 2.68977769e-02
6.24103367e-01 1.89060956e-01 -1.26523197e-01 2.12147221e-01
5.60704350e-01 2.93877006e-01 3.90911847e-01 -2.19878322e-03
-1.75927728e-01 -4.92972285e-01 4.98098433e-01 1.55334377e+00
-3.92351538e-01 -7.48356819e-01 -7.42178142e-01 9.72852767e-01
-1.78396344e+00 -1.06757593e+00 -2.38627680e-02 1.85799551e+00
5.91443717e-01 -1.16692349e-01 -1.91371083e-01 1.61588043e-01
1.10918009e+00 1.56061530e-01 -2.48235002e-01 -9.05544400e-01
-8.23464394e-02 2.12288514e-01 4.48725790e-01 7.42168427e-01
-8.43648612e-01 9.24772024e-01 8.20745373e+00 6.30898833e-01
-7.77741790e-01 3.89359057e-01 5.26412785e-01 4.52150911e-01
-4.37937558e-01 5.58109395e-02 -8.91500711e-01 4.17822361e-01
1.08737242e+00 -5.31326473e-01 5.41410685e-01 7.83374786e-01
-3.88041228e-01 3.53844315e-01 -1.04010522e+00 7.32436419e-01
4.18750077e-01 -1.18104589e+00 4.34164912e-01 -2.40177214e-02
3.09038669e-01 8.78824666e-02 4.04968485e-02 6.42694652e-01
6.28995419e-01 -9.53905284e-01 6.46072507e-01 4.35609758e-01
8.33566427e-01 -5.41580021e-01 9.02208567e-01 1.73911706e-01
-1.59002757e+00 -2.19224676e-01 -4.99979734e-01 -1.97902307e-01
5.16118526e-01 2.11839154e-01 -4.99465108e-01 7.04802215e-01
6.63488209e-01 1.04762685e+00 -1.28330380e-01 8.15949559e-01
-7.49786198e-01 4.59364772e-01 2.45824665e-01 -1.70340210e-01
2.74964064e-01 3.55584244e-03 2.67782360e-01 1.38917732e+00
6.17506325e-01 2.07658812e-01 -4.99053597e-01 7.29765117e-01
-2.93369610e-02 -2.99504727e-01 -3.89703393e-01 -3.13948452e-01
5.33423364e-01 8.86516571e-01 -2.65655696e-01 -7.74784505e-01
-5.47478259e-01 1.20585775e+00 4.90747392e-01 3.58427703e-01
-1.21175539e+00 -6.77770197e-01 7.16805518e-01 8.32847506e-02
7.24549294e-01 5.74721247e-02 -2.45615598e-02 -1.00990152e+00
-5.09539843e-01 -4.47144985e-01 4.28323507e-01 -1.35993946e+00
-1.09317720e+00 9.79729116e-01 -1.07014880e-01 -7.91229486e-01
-3.38422596e-01 -4.15634096e-01 -7.29096413e-01 1.25697851e+00
-1.33544469e+00 -9.87878561e-01 -1.99939877e-01 5.25394559e-01
8.59287381e-01 -1.18942186e-01 6.95427060e-01 4.37285542e-01
-6.91876888e-01 5.54775000e-01 -8.16652998e-02 3.21206123e-01
4.09692340e-02 -1.18028224e+00 1.09720874e+00 7.52210021e-01
-2.36389078e-02 8.95314515e-01 2.42199093e-01 -5.56128502e-01
-1.16149616e+00 -1.26093602e+00 1.63302195e+00 -1.68108240e-01
6.07813239e-01 -3.66475135e-01 -8.45282316e-01 1.61103606e+00
1.10136533e+00 -3.52762520e-01 3.80211711e-01 1.84440300e-01
-9.96211320e-02 2.41809323e-01 -4.80984449e-01 5.66952169e-01
1.28211379e+00 -5.86895406e-01 -9.18519378e-01 2.04341367e-01
1.02170134e+00 -6.81903958e-01 -4.68164146e-01 6.32916903e-03
4.40846145e-01 -9.66568172e-01 9.02044356e-01 -5.44566214e-01
5.90465903e-01 -7.96957240e-02 -7.29479417e-02 -1.40769541e+00
-1.32240355e+00 -7.49261618e-01 -2.23417744e-01 1.43715405e+00
3.94652903e-01 -7.70450294e-01 1.77513197e-01 -9.56044272e-02
-8.63875031e-01 -7.54882514e-01 -7.06554353e-01 -5.68035960e-01
-4.17501740e-02 -1.29086673e-01 9.16998804e-01 5.79501152e-01
1.95468709e-01 1.04375041e+00 -6.40700042e-01 -1.27644464e-01
-3.01032588e-02 -1.06402941e-01 2.89562464e-01 -8.09801042e-01
-3.49477768e-01 -4.86429155e-01 -3.51500899e-01 -1.71370447e+00
-3.59579295e-01 -8.43130589e-01 2.12465242e-01 -1.95888174e+00
5.63722968e-01 7.67425299e-02 -2.36269802e-01 6.06117368e-01
-2.75012046e-01 3.76262590e-02 1.17136493e-01 1.50514930e-01
-4.79925215e-01 8.92335653e-01 1.06282341e+00 -1.12067245e-01
2.67848700e-01 -4.30293560e-01 -1.00311720e+00 3.41652691e-01
7.52950191e-01 -2.63687521e-01 -5.51167369e-01 -1.05076575e+00
3.12063377e-02 -1.13436431e-01 -3.92894633e-02 -8.24006796e-01
2.88260370e-01 1.01518527e-01 1.88661724e-01 -8.37439775e-01
2.58634895e-01 -3.60268712e-01 1.91484302e-01 4.14200455e-01
-4.56357181e-01 7.77242184e-01 1.21269353e-01 1.42604783e-01
-4.57339823e-01 -2.15444028e-01 4.12925243e-01 -4.98745590e-01
-6.38544142e-01 4.69165087e-01 -8.38537514e-01 1.03894129e-01
9.44302320e-01 -1.55694857e-02 -3.35130244e-01 -5.86559653e-01
-1.00401604e+00 2.12382704e-01 -7.29856119e-02 4.89657998e-01
5.73427200e-01 -1.50211036e+00 -6.31558597e-01 6.18712306e-01
-1.00025563e-02 -9.36665386e-02 3.41027796e-01 9.79306817e-01
-2.95784771e-01 1.13000524e+00 -7.65565410e-02 -4.18092728e-01
-1.11718047e+00 9.07846510e-01 4.61275667e-01 -3.34039181e-01
-4.26814139e-01 1.37968898e+00 9.15037751e-01 5.17936349e-02
9.70015302e-02 -4.63475674e-01 -3.66617143e-01 -2.74219155e-01
6.81225479e-01 1.40298471e-01 -6.40883669e-02 -6.98688447e-01
-5.62282443e-01 5.99576294e-01 -1.11248314e-01 5.42878583e-02
9.32882369e-01 -3.45107377e-01 -3.20896864e-01 5.57053804e-01
8.70178759e-01 -4.30313081e-01 -8.00104797e-01 -4.02741171e-02
-3.90893281e-01 -5.22016324e-02 -2.55723387e-01 -4.81968760e-01
-1.23298335e+00 1.61741507e+00 2.58733034e-02 5.39067924e-01
1.29118979e+00 2.12594986e-01 8.29480231e-01 -1.34079680e-01
1.38048418e-02 -5.67674041e-01 -6.96493313e-02 1.07183158e+00
1.21642733e+00 -2.99410850e-01 -5.22743404e-01 -3.29973370e-01
-7.26799846e-01 7.71057129e-01 7.28469193e-01 6.58972338e-02
6.41200960e-01 7.08787978e-01 -4.13013458e-01 4.80926633e-02
-1.49517906e+00 -5.86253583e-01 -9.53315496e-02 9.40592587e-01
1.01511574e+00 -3.11990172e-01 -1.91014528e-01 6.64690912e-01
-2.86370903e-01 2.40898386e-01 1.57635167e-01 8.49843144e-01
-5.86487770e-01 -8.99866581e-01 -1.79577202e-01 2.41985843e-01
-4.03290004e-01 -8.67328346e-01 -3.78427774e-01 5.65427601e-01
3.14585775e-01 5.43136477e-01 4.94548380e-01 -6.95913613e-01
3.17793965e-01 2.73594528e-01 4.71769840e-01 -6.63708448e-01
-6.43530071e-01 4.74192277e-02 -5.25029078e-02 -9.35662463e-02
-1.26396254e-01 -3.79320443e-01 -8.88840735e-01 -4.51043099e-01
-4.64006782e-01 4.82703328e-01 7.56374300e-02 9.36904073e-01
7.49549866e-01 1.11812019e+00 2.12837607e-01 -6.15028858e-01
-1.65454438e-03 -1.48950481e+00 -2.69467890e-01 -2.28925467e-01
4.89038616e-01 -5.25287628e-01 -6.88575804e-02 3.81231219e-01] | [11.072929382324219, 6.952631950378418] |
67f4d1c1-9fb3-4564-9319-b26e268e2d31 | towards-unified-prompt-tuning-for-few-shot-1 | 2205.05313 | null | https://arxiv.org/abs/2205.05313v1 | https://arxiv.org/pdf/2205.05313v1.pdf | Towards Unified Prompt Tuning for Few-shot Text Classification | Prompt-based fine-tuning has boosted the performance of Pre-trained Language Models (PLMs) on few-shot text classification by employing task-specific prompts. Yet, PLMs are unfamiliar with prompt-style expressions during pre-training, which limits the few-shot learning performance on downstream tasks. It would be desirable if the models can acquire some prompting knowledge before adaptation to specific NLP tasks. We present the Unified Prompt Tuning (UPT) framework, leading to better few-shot text classification for BERT-style models by explicitly capturing prompting semantics from non-target NLP datasets. In UPT, a novel paradigm Prompt-Options-Verbalizer is proposed for joint prompt learning across different NLP tasks, forcing PLMs to capture task-invariant prompting knowledge. We further design a self-supervised task named Knowledge-enhanced Selective Masked Language Modeling to improve the PLM's generalization abilities for accurate adaptation to previously unseen tasks. After multi-task learning across multiple tasks, the PLM can be better prompt-tuned towards any dissimilar target tasks in low-resourced settings. Experiments over a variety of NLP tasks show that UPT consistently outperforms state-of-the-arts for prompt-based fine-tuning. | ['Ming Gao', 'Songfang Huang', 'Qiuhui Shi', 'Fei Yang', 'Minghui Qiu', 'Chuanqi Tan', 'Fuli Luo', 'Chengyu Wang', 'Jianing Wang'] | 2022-05-11 | null | null | null | null | ['few-shot-text-classification'] | ['natural-language-processing'] | [ 3.86747599e-01 -2.55589969e-02 -3.73139799e-01 -7.02183723e-01
-8.95392299e-01 -5.38464367e-01 7.22929835e-01 1.89432949e-01
-9.27875280e-01 4.52556223e-01 5.67289829e-01 -1.85747787e-01
1.37444139e-01 -3.65257323e-01 -3.20474207e-01 -2.81680614e-01
4.59627360e-01 5.98145068e-01 4.41528797e-01 -4.96262372e-01
-1.39181346e-01 -1.70971695e-02 -1.18644297e+00 6.14038825e-01
7.96788812e-01 6.38833344e-01 6.98195517e-01 7.36437201e-01
-5.95618844e-01 6.90628231e-01 -4.75994766e-01 -3.03849280e-01
1.10407062e-01 -3.42544556e-01 -8.09932590e-01 -4.40660536e-01
2.88961142e-01 -3.21128666e-01 -2.20630288e-01 6.19016290e-01
6.92559898e-01 1.02373242e+00 7.94529200e-01 -1.07245731e+00
-1.09087253e+00 9.06542301e-01 -2.41352722e-01 5.57517052e-01
3.69650982e-02 4.90613520e-01 1.21197486e+00 -1.22714746e+00
4.93884444e-01 1.39486396e+00 4.73075479e-01 1.15226257e+00
-1.40872920e+00 -7.47899473e-01 6.11041129e-01 3.03696573e-01
-1.02993214e+00 -6.57351017e-01 5.26959777e-01 -3.22563410e-01
1.44013762e+00 -1.24049164e-01 4.09565121e-02 1.69507813e+00
-1.62544418e-02 1.12639904e+00 9.61464107e-01 -6.47535741e-01
2.89911777e-01 7.89265484e-02 7.08032906e-01 3.34500492e-01
-1.76378623e-01 1.46225363e-01 -9.42703784e-01 -2.57473379e-01
3.38900566e-01 1.44960016e-01 -2.10951269e-01 -4.68325540e-02
-1.34342778e+00 9.14341688e-01 1.56827286e-01 4.19641495e-01
-4.07731980e-01 -1.41909882e-01 6.88934207e-01 4.44831073e-01
7.17511594e-01 9.27363038e-01 -1.07219052e+00 -2.23117560e-01
-9.41419482e-01 -2.57144403e-03 7.70265996e-01 1.14559686e+00
7.14977682e-01 2.31740236e-01 -1.23522925e+00 1.15324712e+00
4.19383273e-02 2.98773974e-01 9.67293441e-01 -5.60017526e-01
4.24064159e-01 2.90477961e-01 6.82906806e-02 -7.36000985e-02
-5.31624019e-01 -1.77800328e-01 -5.02453327e-01 -4.10745382e-01
2.25410670e-01 -4.96043414e-01 -1.13747895e+00 1.99403727e+00
1.79895476e-01 3.48556787e-01 3.47624063e-01 6.71946883e-01
9.47324872e-01 9.44793344e-01 9.01812971e-01 -3.27550381e-01
1.50569034e+00 -1.28596866e+00 -7.36644924e-01 -7.84055710e-01
1.01573753e+00 -5.62695444e-01 2.02218962e+00 -3.47188786e-02
-5.31096101e-01 -7.30466306e-01 -6.18718982e-01 -4.18312162e-01
-3.99703711e-01 -1.25399649e-01 4.26378429e-01 1.82460293e-01
-6.54615939e-01 3.69172633e-01 -5.46147108e-01 -7.17569351e-01
4.44289684e-01 -1.47477210e-01 4.07764278e-02 -3.31367046e-01
-1.74638295e+00 1.09434426e+00 6.90280974e-01 -5.78241229e-01
-1.21341789e+00 -1.34827375e+00 -8.79493594e-01 4.16443974e-01
6.36596084e-01 -8.30082715e-01 1.99578857e+00 -6.37930930e-01
-1.70680737e+00 8.13685417e-01 -3.87025237e-01 -3.81417066e-01
1.49196401e-01 -4.08681214e-01 -3.90395731e-01 -1.40895784e-01
2.81086504e-01 8.59036803e-01 1.10171187e+00 -8.43471408e-01
-6.35475814e-01 8.03732648e-02 -1.29963487e-01 3.92549694e-01
-6.58511519e-01 3.42182964e-01 -4.95444424e-02 -7.18723357e-01
-6.44187033e-01 -5.48810482e-01 -1.98868424e-01 -1.71892777e-01
1.54234413e-02 -7.75647938e-01 7.23786056e-01 -2.26644292e-01
1.18301785e+00 -2.18138671e+00 -1.24479227e-01 -6.08841479e-01
-7.20325718e-03 8.11125696e-01 -8.28921914e-01 4.30063814e-01
1.51651902e-02 -1.12525664e-01 4.37923484e-02 -5.84603965e-01
1.50753871e-01 4.69779760e-01 -7.76073277e-01 -1.60323918e-01
3.12775701e-01 1.39202654e+00 -1.19087052e+00 -3.96715790e-01
8.64728317e-02 9.87828150e-02 -4.15702790e-01 6.64947569e-01
-7.54503906e-01 3.85101140e-01 -4.33487505e-01 3.35498154e-01
8.20403248e-02 -4.15375203e-01 -2.62037098e-01 1.94900110e-01
2.39724979e-01 3.74875724e-01 -7.32424557e-01 1.98592031e+00
-7.97532856e-01 2.96153814e-01 -1.19389087e-01 -9.10151601e-01
9.54249203e-01 5.97247422e-01 -2.25645839e-03 -6.94110215e-01
1.83072329e-01 -1.72118396e-01 8.49672966e-03 -6.14427269e-01
5.26462913e-01 -5.86691320e-01 -3.08188528e-01 7.86750555e-01
7.52027035e-01 -2.86137732e-03 1.45830348e-01 3.97370815e-01
1.02499759e+00 1.94247682e-02 7.53985763e-01 -2.77171642e-01
1.89443216e-01 1.09540232e-01 5.31653643e-01 1.21269727e+00
-7.46662319e-01 2.26630837e-01 -9.90834534e-02 -1.91210255e-01
-6.87228322e-01 -1.00597346e+00 5.46359494e-02 2.60821033e+00
-3.24606210e-01 -3.37474167e-01 -2.13177428e-01 -7.63998032e-01
7.58120939e-02 1.44278240e+00 -7.00938284e-01 -5.35716832e-01
-2.24691644e-01 -5.48298180e-01 5.04664361e-01 7.50668108e-01
1.20525785e-01 -1.21509385e+00 -5.26047170e-01 5.30144513e-01
-1.93880141e-01 -1.25702202e+00 -9.61244702e-01 8.16864908e-01
-7.37878740e-01 -6.26730859e-01 -9.31052089e-01 -7.75656939e-01
3.68660003e-01 5.92371345e-01 1.12466538e+00 -3.96665901e-01
-3.41292135e-02 6.56116664e-01 -7.14037955e-01 -7.10062563e-01
-4.20186549e-01 2.96319515e-01 4.50076699e-01 -9.57047865e-02
9.52962816e-01 -3.40047777e-01 -8.58902857e-02 2.95467764e-01
-6.85268521e-01 4.46772240e-02 4.65372115e-01 1.23008263e+00
2.75974542e-01 -5.43309093e-01 8.43360960e-01 -1.04757917e+00
1.09784889e+00 -6.11310780e-01 -7.99758881e-02 5.58540821e-01
-5.11699736e-01 2.71017402e-01 7.66307414e-01 -1.16484308e+00
-1.71438336e+00 -2.82379746e-01 2.39273816e-01 -6.04696035e-01
-3.94743949e-01 4.54147756e-01 9.12830159e-02 2.58714199e-01
1.35195792e+00 2.70390391e-01 -4.47087586e-01 -6.04112566e-01
8.58521760e-01 6.42373979e-01 3.68193358e-01 -8.62600029e-01
6.41591847e-01 2.01228000e-02 -6.20402038e-01 -8.64910603e-01
-1.67504597e+00 -9.16305602e-01 -6.43733978e-01 1.34305164e-01
7.37850070e-01 -1.19719052e+00 -1.71702489e-01 2.40816802e-01
-1.34825730e+00 -9.51911330e-01 -6.12231433e-01 3.94797176e-01
-4.66426283e-01 -5.35348393e-02 -6.52589500e-01 -6.42834246e-01
-6.03481054e-01 -6.37440741e-01 9.35405970e-01 3.59407187e-01
-7.59826183e-01 -1.32883966e+00 2.93293804e-01 1.08810090e-01
7.49290168e-01 -7.73626268e-01 1.11418045e+00 -1.64378607e+00
2.12693200e-01 1.48488760e-01 -2.47414529e-01 3.03052843e-01
2.49303803e-02 -5.01698911e-01 -1.26082098e+00 -1.68688416e-01
2.70248353e-01 -8.90596688e-01 1.00983310e+00 2.83678770e-01
7.78340042e-01 -3.99097234e-01 -4.44628298e-01 5.52760720e-01
9.67406094e-01 -8.73018652e-02 -1.89700201e-01 1.69399410e-01
5.78720987e-01 5.94078600e-01 9.88130927e-01 4.66661811e-01
2.93778777e-01 5.16158998e-01 -3.90889317e-01 2.27166906e-01
-2.72671074e-01 -5.30522764e-01 6.80624008e-01 6.70798600e-01
5.29733419e-01 -1.58111691e-01 -1.14671898e+00 6.22859836e-01
-2.00695419e+00 -7.62340665e-01 3.43476564e-01 1.66933620e+00
1.42011881e+00 1.20388065e-02 -1.70936018e-01 -6.42314732e-01
5.84997416e-01 4.67952102e-01 -8.05471003e-01 -4.34504777e-01
4.49361689e-02 3.25302899e-01 8.57157484e-02 6.01242483e-01
-9.60630000e-01 1.54921043e+00 5.87613344e+00 1.03797019e+00
-1.08276725e+00 7.59606361e-01 1.79411530e-01 -4.32914913e-01
-2.07549810e-01 -2.26734020e-02 -1.38947010e+00 2.97081977e-01
1.07297337e+00 -7.79040277e-01 2.63665289e-01 1.16084027e+00
1.20552883e-01 1.64557666e-01 -1.37332761e+00 7.52632022e-01
2.55807072e-01 -1.02012432e+00 1.08189322e-01 -6.82403684e-01
9.24794972e-01 4.69332069e-01 -2.19836101e-01 1.32529271e+00
8.55913043e-01 -7.12681234e-01 3.59902084e-01 3.35563481e-01
7.58482039e-01 -1.10885121e-01 3.03309292e-01 8.75735819e-01
-9.13937867e-01 -3.10324907e-01 -6.23296797e-01 -3.18082452e-01
4.03521895e-01 3.30027193e-01 -1.21739089e+00 6.88883811e-02
3.24577004e-01 5.94882846e-01 -5.84561884e-01 5.82825065e-01
-6.07010305e-01 8.52082253e-01 -6.09167330e-02 -5.25380552e-01
3.52312803e-01 4.87420738e-01 5.79139888e-01 1.52464747e+00
1.27592050e-02 4.70979452e-01 5.92681527e-01 8.03742766e-01
-1.43178567e-01 3.39749455e-01 -5.60198367e-01 -3.38224709e-01
7.17847645e-01 1.28495896e+00 -2.28893027e-01 -7.36954033e-01
-6.32319570e-01 9.63227987e-01 7.04479516e-01 7.39257157e-01
-3.90351295e-01 -1.38202056e-01 6.30886436e-01 -1.19855456e-01
1.88287437e-01 -1.82561859e-01 -2.33919367e-01 -1.37791264e+00
-5.26983082e-01 -6.92612946e-01 6.95330918e-01 -8.34710717e-01
-1.89778996e+00 4.48752403e-01 1.93537936e-01 -7.15257704e-01
-2.90693641e-01 -7.95361340e-01 -9.97127235e-01 1.05892313e+00
-1.67765760e+00 -1.27005255e+00 3.79042923e-02 7.77636528e-01
1.22220194e+00 -4.42673743e-01 1.16698635e+00 -2.25500822e-01
-4.34123725e-01 6.13836229e-01 -1.07500836e-01 -7.45999366e-02
1.54580939e+00 -1.06185639e+00 4.62989122e-01 9.44903016e-01
1.79146767e-01 7.18739212e-01 7.96715856e-01 -7.26882100e-01
-1.04580259e+00 -1.30813110e+00 1.10787487e+00 -7.74580181e-01
9.73584473e-01 -4.92196351e-01 -1.31353545e+00 1.07282853e+00
2.84209400e-01 1.90442547e-01 1.08398235e+00 7.02748120e-01
-7.60968029e-01 1.45062923e-01 -5.58429062e-01 7.09227324e-01
1.03774917e+00 -8.49115729e-01 -1.48352528e+00 5.39947331e-01
1.20310819e+00 -1.67675152e-01 -4.31105971e-01 8.03559572e-02
-2.72901002e-02 -1.21006668e-01 7.08380580e-01 -1.30723834e+00
1.56396717e-01 2.02387393e-01 -1.37797013e-01 -1.78957450e+00
-7.27502465e-01 -7.43915021e-01 -4.17366743e-01 1.29817879e+00
4.01765227e-01 -4.44346070e-01 2.49390125e-01 7.73269117e-01
-2.27938548e-01 -5.00087321e-01 -7.07307816e-01 -1.14353240e+00
2.75731474e-01 -5.37625551e-01 1.72691941e-01 1.21122420e+00
4.23461258e-01 1.28344083e+00 -3.69372576e-01 -2.96442360e-01
2.62117714e-01 -1.10043995e-01 7.50450790e-01 -1.44875300e+00
-3.98500979e-01 -3.11019450e-01 4.69635248e-01 -1.21470392e+00
6.99570239e-01 -1.30423641e+00 4.07328457e-01 -1.24714303e+00
3.32956225e-01 -2.06186518e-01 -5.44265330e-01 1.10692585e+00
-8.48233163e-01 -5.20079255e-01 3.31803918e-01 1.24373026e-01
-9.03455615e-01 8.78644764e-01 1.30105174e+00 -1.00504525e-01
-4.82807845e-01 -9.69023556e-02 -9.35721397e-01 6.21205688e-01
6.29369020e-01 -8.07906926e-01 -5.94812751e-01 -5.48239887e-01
-7.86964223e-02 -2.36859754e-01 5.11173978e-02 -5.25333047e-01
6.15544379e-01 -4.48721379e-01 1.71988413e-01 -6.76857606e-02
2.76441276e-01 -4.54119623e-01 -7.78666973e-01 1.64707869e-01
-9.44135010e-01 -4.03095007e-01 4.32819992e-01 6.50875926e-01
1.48456067e-01 -4.54643846e-01 9.30644214e-01 -3.12314749e-01
-1.28611314e+00 5.08000851e-01 -3.22227895e-01 7.77065992e-01
9.52029586e-01 2.11269453e-01 -5.70880413e-01 -1.78633124e-01
-7.59591401e-01 5.82645059e-01 1.82553023e-01 8.57337415e-01
3.93326640e-01 -1.15471435e+00 -8.30944061e-01 8.97478089e-02
8.24397266e-01 -5.01110964e-02 4.68397856e-01 7.53319025e-01
6.69773221e-01 5.60296059e-01 -1.37154728e-01 -4.58390385e-01
-1.07937002e+00 9.12098527e-01 3.11215338e-03 -4.56782728e-01
-6.54689431e-01 1.34860361e+00 5.69940388e-01 -6.38923168e-01
2.12011263e-01 -2.20752344e-01 -6.74633607e-02 3.75487983e-01
8.55916917e-01 5.05121797e-02 -1.95147976e-01 -8.62147734e-02
-1.55333936e-01 1.08091429e-01 -7.35302567e-01 -2.75211126e-01
1.10168254e+00 -2.41158679e-01 4.66258198e-01 1.00835907e+00
9.78939474e-01 -5.62657118e-01 -1.60548186e+00 -1.16389108e+00
1.86554149e-01 -1.56573266e-01 1.11900024e-01 -1.22131860e+00
-1.94332629e-01 1.23519814e+00 7.66284168e-02 -3.61041427e-01
7.52362788e-01 1.51215672e-01 7.93357432e-01 8.94794822e-01
2.39973694e-01 -1.38764405e+00 5.29719830e-01 1.24139643e+00
7.21215546e-01 -1.51454413e+00 -4.61400211e-01 1.30690992e-01
-1.24128079e+00 9.24151897e-01 9.33560014e-01 2.01373145e-01
6.04427814e-01 1.33815363e-01 3.53810668e-01 9.16104242e-02
-1.45028508e+00 -3.05413485e-01 3.90371233e-01 6.88085914e-01
4.51577753e-01 7.27531547e-03 -2.96519604e-02 1.33058226e+00
1.41808808e-01 6.91578090e-02 1.33399457e-01 9.48368430e-01
-8.85787189e-01 -8.60454321e-01 -4.55790898e-03 4.86931324e-01
4.03054553e-04 -6.89146817e-01 -1.63836583e-01 3.48842323e-01
-2.51145482e-01 7.86408305e-01 -6.96475208e-02 4.64765355e-02
3.81179631e-01 1.06382775e+00 2.75125593e-01 -1.53341627e+00
-5.75787008e-01 -2.72502095e-01 1.22026149e-02 -4.75159407e-01
6.87332600e-02 -4.87637162e-01 -1.10608447e+00 1.39700115e-01
-1.47583634e-01 1.07215449e-01 7.72035941e-02 1.30025852e+00
4.89931524e-01 7.26438701e-01 2.77146220e-01 -7.31213987e-01
-1.21338630e+00 -1.39522278e+00 -4.15675074e-01 4.85940307e-01
2.90487617e-01 -7.71856606e-01 -3.15469474e-01 8.89642760e-02] | [10.873193740844727, 8.054939270019531] |
0b99c98d-3cf3-45e8-b7fe-0f1206564ce8 | mastering-the-exploration-exploitation-trade | 2305.08624 | null | https://arxiv.org/abs/2305.08624v1 | https://arxiv.org/pdf/2305.08624v1.pdf | Mastering the exploration-exploitation trade-off in Bayesian Optimization | Gaussian Process based Bayesian Optimization is a well-known sample efficient sequential strategy for globally optimizing black-box, expensive, and multi-extremal functions. The role of the Gaussian Process is to provide a probabilistic approximation of the unknown function, depending on the sequentially collected observations, while an acquisition function drives the choice of the next solution to evaluate, balancing between exploration and exploitation, depending on the current Gaussian Process model. Despite the huge effort of the scientific community in defining effective exploration-exploitation mechanisms, we are still far away from the master acquisition function. This paper merges the most relevant results and insights from both algorithmic and human search strategies to propose a novel acquisition function, mastering the trade-off between explorative and exploitative choices, adaptively. We compare the proposed acquisition function on a number of test functions and against different state-of-the-art ones, which are instead based on prefixed or random scheduling between exploration and exploitation. A Pareto analysis is performed with respect to two (antagonistic) goals: convergence to the optimum and exploration capability. Results empirically prove that the proposed acquisition function is almost always Pareto optimal and also the most balanced trade-off between the two goals. | ['Antonio Candelieri'] | 2023-05-15 | null | null | null | null | ['bayesian-optimization'] | ['methodology'] | [ 1.40086770e-01 -7.57239908e-02 -1.21911354e-01 -9.38182026e-02
-5.04885316e-01 -6.64921343e-01 6.07449055e-01 4.36494768e-01
-7.82247603e-01 9.16903257e-01 -1.50638834e-01 -6.27371892e-02
-9.84186888e-01 -7.61363208e-01 -3.05672020e-01 -1.16746151e+00
-1.24798305e-01 1.05252111e+00 2.23403677e-01 -7.35476147e-03
8.05720985e-01 5.46021163e-01 -1.73470116e+00 -4.86526310e-01
1.19713736e+00 1.14708281e+00 4.36101258e-01 6.76244140e-01
-2.05916852e-01 4.89738584e-02 -5.83038926e-01 -2.02513665e-01
2.96482533e-01 -4.20879215e-01 -4.59956884e-01 -1.48615062e-01
-5.59867740e-01 4.45479482e-01 5.51339567e-01 1.18660450e+00
6.03373826e-01 3.34421068e-01 7.74510086e-01 -1.21383977e+00
-2.04845518e-01 5.01428008e-01 -4.42819923e-01 2.73626894e-01
1.36889026e-01 4.12112236e-01 8.43241692e-01 -4.19842511e-01
4.83317763e-01 1.17281950e+00 1.96547568e-01 3.10072243e-01
-1.28292286e+00 -3.53831381e-01 1.16180316e-01 2.00402632e-01
-1.33427215e+00 -1.22013651e-01 6.46018922e-01 -4.43880856e-01
3.87516409e-01 3.72647583e-01 7.97426760e-01 9.72876310e-01
5.55315018e-01 6.00062609e-01 1.42883658e+00 -4.36224520e-01
1.07912636e+00 4.27977979e-01 -6.80595189e-02 2.57921249e-01
5.25306702e-01 3.81450355e-01 -5.95479369e-01 -3.49943131e-01
5.48415065e-01 -3.39986324e-01 -4.45069313e-01 -7.48708725e-01
-7.59948552e-01 7.80114591e-01 -7.73156062e-02 4.64724928e-01
-8.47341955e-01 2.13034172e-02 -6.84509501e-02 9.58923027e-02
1.46135598e-01 8.44868004e-01 -3.38920146e-01 -6.15731180e-01
-1.06039941e+00 4.65319455e-01 1.04220808e+00 4.19249266e-01
5.84566414e-01 -4.02278490e-02 -4.33991134e-01 3.00140887e-01
5.72671592e-01 3.78805071e-01 6.03422105e-01 -6.08449340e-01
1.99642733e-01 5.64468563e-01 6.80792391e-01 -9.66552973e-01
-3.02326322e-01 -1.05603349e+00 -3.59713197e-01 6.39177859e-01
6.15239918e-01 -3.49551022e-01 -4.70346212e-01 1.58997488e+00
5.34707427e-01 -2.03121349e-01 -2.01860100e-01 8.44935179e-01
-8.92059430e-02 4.26607609e-01 -3.88049930e-02 -6.41162276e-01
1.15857244e+00 -7.47101486e-01 -7.64731705e-01 -1.89550081e-03
-3.40097919e-02 -5.19587576e-01 9.39549029e-01 8.52606356e-01
-1.13266969e+00 -3.55272502e-01 -1.17533875e+00 8.08486223e-01
-2.84202844e-01 -1.45881981e-01 3.13557416e-01 1.14548039e+00
-9.26612258e-01 8.28128874e-01 -9.07854557e-01 -1.73768818e-01
2.37989411e-01 3.43696535e-01 1.45768330e-01 3.60794604e-01
-7.87600458e-01 9.64583814e-01 6.71636999e-01 2.63400257e-01
-1.05778134e+00 -5.38344145e-01 -2.02277660e-01 3.06599587e-01
7.38985062e-01 -5.98764420e-01 8.96068156e-01 -9.42494750e-01
-2.03594232e+00 3.82523328e-01 -4.55447137e-02 -6.72329485e-01
1.06426620e+00 -2.66226530e-01 1.96309239e-02 -9.93243158e-02
-2.74165630e-01 2.68065453e-01 1.03678739e+00 -1.38925338e+00
-7.26548791e-01 -5.23910463e-01 -2.70596296e-01 3.45470428e-01
-3.45938802e-01 -4.64079082e-02 -1.18735820e-01 -3.48314255e-01
1.92197934e-02 -7.68406630e-01 -4.19048011e-01 -2.43486434e-01
-4.45220023e-01 -1.98280931e-01 4.42845255e-01 -2.36590281e-01
1.39511287e+00 -1.78449726e+00 4.98005211e-01 5.49496651e-01
-1.55070394e-01 -1.85254708e-01 2.65725464e-01 5.14715075e-01
1.87653825e-01 -7.06208637e-03 -3.77100199e-01 -4.69727308e-01
2.19495922e-01 1.06716253e-01 -1.48360789e-01 7.68489718e-01
-1.70799285e-01 4.52795208e-01 -1.06489849e+00 -4.27402765e-01
1.03418097e-01 1.66144311e-01 -2.60633856e-01 2.13882476e-01
-3.18773210e-01 6.42710567e-01 -6.59899354e-01 6.82793617e-01
5.57067394e-01 9.50106606e-02 1.56587228e-01 2.76985884e-01
-4.43849772e-01 -4.31469858e-01 -1.69962549e+00 1.23087728e+00
-3.78239989e-01 2.70589650e-01 3.01030308e-01 -8.69718909e-01
1.23648536e+00 1.13519169e-01 5.19764125e-01 -2.70741403e-01
4.92446840e-01 4.92982566e-01 1.14601098e-01 -2.28782013e-01
4.15576845e-01 -2.34186172e-01 2.00507984e-01 4.40373719e-01
-8.81894305e-02 -1.64385617e-01 3.89958143e-01 -6.55948579e-01
9.29770350e-01 4.53035802e-01 7.14764774e-01 -8.94245446e-01
7.28353798e-01 -1.88025907e-01 4.57333088e-01 9.07414794e-01
-3.21346581e-01 3.79339069e-01 6.66810632e-01 -2.76549906e-01
-5.69061458e-01 -1.03073967e+00 -6.61105141e-02 9.41336274e-01
3.67800832e-01 1.17203057e-01 -6.98968112e-01 -4.97319132e-01
-8.85782242e-02 1.23391926e+00 -7.92151034e-01 -4.97604348e-02
-4.20970619e-01 -9.49251175e-01 1.11565098e-01 8.20433646e-02
2.95770735e-01 -9.83094394e-01 -1.65696895e+00 3.15289289e-01
2.36086041e-01 -3.93669188e-01 -2.73503195e-02 5.77587485e-01
-1.00562477e+00 -9.99801993e-01 -6.47853732e-01 5.62394485e-02
4.41916138e-01 -3.32630455e-01 1.05088305e+00 -3.56212586e-01
-5.92985898e-02 3.35110694e-01 -2.38055855e-01 -7.34575629e-01
-3.31935376e-01 4.24500555e-02 -5.83103709e-02 3.33353132e-01
7.92274624e-02 -7.52439559e-01 -5.89550912e-01 5.20244658e-01
-7.98385620e-01 -3.43310922e-01 7.15587735e-01 7.12903440e-01
6.59879148e-01 5.38268209e-01 1.49361238e-01 -1.91366389e-01
9.38043654e-01 -4.77054834e-01 -1.19737971e+00 3.33937883e-01
-8.67536426e-01 7.00398207e-01 4.32082325e-01 -5.25341630e-01
-1.14663231e+00 -1.71192452e-01 2.72002608e-01 -4.31295067e-01
-8.33762735e-02 2.55011827e-01 -3.18356991e-01 8.75424314e-03
5.06793201e-01 2.95545816e-01 -6.42520264e-02 -5.06095111e-01
1.53275400e-01 2.56141394e-01 3.41053545e-01 -9.51561689e-01
5.83066165e-01 6.11540914e-01 3.94541889e-01 -6.76816285e-01
-3.50662619e-01 -3.49021971e-01 -4.18764859e-01 -4.41157371e-01
7.77353704e-01 1.33452058e-01 -9.96925890e-01 2.58837193e-01
-9.26011860e-01 -1.55260280e-01 -7.39968419e-01 4.29761618e-01
-8.21313739e-01 2.28475764e-01 2.30734304e-01 -1.55425310e+00
-3.47747087e-01 -1.33075273e+00 9.55407262e-01 4.82411772e-01
-2.61460066e-01 -1.02336895e+00 3.31455916e-01 -5.40529415e-02
5.60629904e-01 5.31014085e-01 7.02614188e-01 -8.00563931e-01
-5.54521501e-01 -1.71032146e-01 3.86194766e-01 3.21838371e-02
-2.24142626e-01 6.17128015e-02 -5.74646950e-01 -2.15221569e-01
6.18197024e-01 1.76626772e-01 5.35351515e-01 5.46002805e-01
8.16347420e-01 -1.67724431e-01 -3.63692492e-01 4.29884523e-01
1.62959015e+00 6.84933722e-01 5.69708586e-01 7.75770247e-01
-1.95016339e-01 1.01366639e+00 8.83315921e-01 8.78015697e-01
-1.61715955e-01 7.64354229e-01 8.86124849e-01 6.63246512e-01
5.62312663e-01 -7.96103776e-02 1.81979239e-01 6.14515170e-02
-1.84026584e-01 -3.75163049e-01 -9.53410029e-01 4.01852965e-01
-1.92380643e+00 -7.74726927e-01 2.33122230e-01 2.74009800e+00
5.40121257e-01 4.35719579e-01 1.68748200e-01 2.19474792e-01
7.89101243e-01 -1.13276765e-01 -4.57002968e-01 -3.16835612e-01
-5.83801270e-02 2.73919314e-01 6.96734786e-01 6.30501986e-01
-9.74230349e-01 4.10839021e-01 5.95796824e+00 1.21706450e+00
-1.13508713e+00 5.08381762e-02 5.78605831e-01 -3.02448750e-01
-3.24324846e-01 1.03208192e-01 -8.26020241e-01 6.96600914e-01
8.18642318e-01 -4.17012393e-01 3.99049819e-01 8.87408316e-01
4.13711011e-01 -5.15393794e-01 -9.42185581e-01 7.87040651e-01
-1.63611650e-01 -9.39838171e-01 -2.82474697e-01 3.83673906e-01
7.41786659e-01 -4.83744293e-01 1.30265549e-01 -3.42871882e-02
1.81006655e-01 -1.18803954e+00 1.09216762e+00 9.48680282e-01
-1.54264197e-01 -9.23437059e-01 9.53905880e-01 5.40506661e-01
-9.35815871e-01 -5.25951207e-01 4.81079891e-02 1.66217685e-01
4.05342698e-01 5.93764126e-01 -5.18268645e-01 8.08263063e-01
6.05058074e-01 -6.88063353e-02 -3.26129824e-01 1.43366265e+00
-2.54164726e-01 3.42246175e-01 -5.99757195e-01 -6.59433722e-01
4.43047583e-01 -7.26508379e-01 1.35399175e+00 8.83780241e-01
5.71675897e-01 -5.22900045e-01 -4.47243229e-02 1.22832584e+00
8.24103296e-01 2.31576353e-01 -1.85020030e-01 -5.08210212e-02
5.42110562e-01 1.06833696e+00 -1.07858896e+00 1.67744271e-02
4.81602520e-01 4.15935129e-01 -1.34455919e-01 2.45919690e-01
-7.83026814e-01 -1.40247867e-01 4.06476617e-01 2.73926836e-02
4.51427698e-01 -2.92813718e-01 -6.57566190e-01 -5.13523281e-01
-4.25104722e-02 -6.78851485e-01 3.94293636e-01 -2.98380911e-01
-9.84159172e-01 5.22026122e-01 4.78408992e-01 -9.97523069e-01
-4.23669577e-01 -5.89403689e-01 -6.52571499e-01 9.99316812e-01
-1.18885100e+00 -5.39735436e-01 -2.12246016e-01 6.76164255e-02
5.56668699e-01 -2.02931330e-01 4.31582987e-01 -2.08373502e-01
-3.91593635e-01 1.26672253e-01 3.59611541e-01 -9.51713741e-01
1.44118801e-01 -1.40872478e+00 -4.22271609e-01 7.97959864e-01
-1.53673366e-01 6.81912780e-01 1.49285781e+00 -6.73147559e-01
-1.50312150e+00 -3.64585489e-01 2.11452037e-01 -3.07930380e-01
6.83822691e-01 -8.19445997e-02 -6.45835578e-01 -9.29655805e-02
1.19974472e-01 -5.77035725e-01 4.52727020e-01 -6.09150343e-02
4.97480243e-01 -1.08685836e-01 -1.26013017e+00 7.26433098e-01
6.19733691e-01 3.36420596e-01 -5.15474916e-01 -7.14681447e-02
2.07063869e-01 -1.56050712e-01 -6.29633784e-01 4.31593537e-01
5.99538624e-01 -1.24793780e+00 7.82266676e-01 -1.98195100e-01
5.02874479e-02 -3.03790748e-01 -1.26532212e-01 -1.31534076e+00
8.50609764e-02 -1.13519919e+00 -3.84859622e-01 1.00509286e+00
1.59845367e-01 -9.04553175e-01 9.62380052e-01 3.63082141e-01
2.33107880e-01 -1.34959936e+00 -1.32252228e+00 -1.06802440e+00
-1.28811225e-01 -3.37055922e-01 4.70964521e-01 2.01540709e-01
-4.25556630e-01 -6.88355342e-02 -9.65472981e-02 8.14868361e-02
8.79362047e-01 2.66553134e-01 7.54503608e-01 -1.32826769e+00
-7.16441333e-01 -1.03414810e+00 -3.01669329e-01 -6.73169076e-01
-2.99930990e-01 -2.00071737e-01 1.97879285e-01 -1.27829564e+00
-1.24936387e-01 -2.61209428e-01 -3.07986200e-01 -1.71450481e-01
-1.41115174e-01 -4.67361301e-01 4.55521531e-02 1.48598492e-01
-3.50012958e-01 8.62000048e-01 1.07132137e+00 2.52128571e-01
-6.41731918e-01 4.53806847e-01 -6.15966976e-01 7.33798444e-01
6.48548186e-01 -5.28103709e-01 -5.55602014e-01 1.93298772e-01
4.39145118e-01 4.73430119e-02 4.40421291e-02 -1.09450853e+00
3.29323620e-01 -4.27123189e-01 -1.39547950e-02 -6.76149130e-01
3.88986528e-01 -9.95832801e-01 4.24688995e-01 6.99714184e-01
-1.50456622e-01 1.15213245e-01 2.21415721e-02 9.66820359e-01
-6.98228776e-02 -8.36844027e-01 8.79658520e-01 -1.46202758e-01
-4.94050413e-01 -2.69109551e-02 -2.41587162e-01 -1.51240662e-01
1.42030036e+00 -6.91863358e-01 6.85422570e-02 -3.81627768e-01
-7.20057189e-01 3.60238224e-01 2.82855898e-01 1.71421662e-01
2.63202399e-01 -7.40509152e-01 -6.23240054e-01 -2.06539124e-01
-3.61551315e-01 -2.39394590e-01 1.10485218e-01 1.07263649e+00
-5.47673702e-01 3.81479710e-01 -2.89153159e-01 -6.44942462e-01
-1.01657677e+00 6.26002729e-01 4.25139129e-01 -8.10453892e-01
-8.58653039e-02 7.29321778e-01 -1.84272472e-02 -5.91528378e-02
3.27732176e-01 -1.46682039e-01 -3.22021365e-01 2.78860241e-01
2.94557929e-01 1.00795829e+00 -1.48218527e-01 -7.81790689e-02
-2.12063983e-01 4.73820150e-01 6.90497160e-01 -6.30592227e-01
1.17311513e+00 -1.81488290e-01 -1.91662312e-01 5.86639702e-01
5.46355665e-01 8.46545920e-02 -1.25377262e+00 3.28063697e-01
3.44954222e-01 -5.95663905e-01 -3.85725312e-02 -7.83339918e-01
-5.14629960e-01 6.81183040e-01 8.32092881e-01 4.88009036e-01
1.08071625e+00 -4.85426247e-01 8.31527412e-02 2.88051754e-01
5.70809841e-01 -1.22396588e+00 1.61835492e-01 2.18489677e-01
9.70654547e-01 -6.42800987e-01 4.17171270e-02 -1.33340210e-01
-4.19780821e-01 1.09858263e+00 3.66327435e-01 -2.31663451e-01
5.99896431e-01 2.88026899e-01 -3.93549949e-01 -1.30656987e-01
-6.50188267e-01 -1.32734805e-01 2.30763555e-01 4.11643595e-01
-1.05064288e-01 3.74747533e-03 -7.42495894e-01 5.34084380e-01
-1.76455557e-01 -2.09904388e-01 -3.38789411e-02 1.04646146e+00
-6.33827925e-01 -1.09056067e+00 -9.50246394e-01 1.25433892e-01
-3.47012043e-01 2.20078006e-01 -9.35888886e-02 1.00046122e+00
2.03383848e-01 8.66612375e-01 -1.82575554e-01 2.74280399e-01
1.69507831e-01 4.00625132e-02 4.53296065e-01 -1.62162945e-01
-6.32100582e-01 1.41690165e-01 -6.02240562e-02 -6.32962406e-01
-7.49803409e-02 -7.55012572e-01 -5.78322589e-01 1.85374349e-01
-4.08556461e-01 6.10608697e-01 1.02081358e+00 1.02696061e+00
1.69217363e-01 4.66495484e-01 4.44580942e-01 -9.83955920e-01
-1.28172755e+00 -8.50033522e-01 -5.22296607e-01 -1.71071932e-01
-1.78501979e-01 -1.06668115e+00 -4.87156868e-01 -5.70527792e-01] | [6.0307111740112305, 3.6468684673309326] |
d45f6ffb-f124-4e81-ad02-1cd98328309d | sparsifying-sparse-representations-for | 2112.09628 | null | https://arxiv.org/abs/2112.09628v1 | https://arxiv.org/pdf/2112.09628v1.pdf | Sparsifying Sparse Representations for Passage Retrieval by Top-$k$ Masking | Sparse lexical representation learning has demonstrated much progress in improving passage retrieval effectiveness in recent models such as DeepImpact, uniCOIL, and SPLADE. This paper describes a straightforward yet effective approach for sparsifying lexical representations for passage retrieval, building on SPLADE by introducing a top-$k$ masking scheme to control sparsity and a self-learning method to coax masked representations to mimic unmasked representations. A basic implementation of our model is competitive with more sophisticated approaches and achieves a good balance between effectiveness and efficiency. The simplicity of our methods opens the door for future explorations in lexical representation learning for passage retrieval. | ['Jimmy Lin', 'Xueguang Ma', 'Jheng-Hong Yang'] | 2021-12-17 | null | null | null | null | ['self-learning'] | ['natural-language-processing'] | [-2.78240621e-01 -2.64533460e-01 -9.27877605e-01 -3.49452421e-02
-1.50668383e+00 -4.62561071e-01 7.00310349e-01 3.94322366e-01
-5.96931696e-01 8.25644433e-01 8.88497293e-01 -2.82215863e-01
-2.13612810e-01 -9.83556569e-01 -5.19078016e-01 -3.79475057e-01
-1.92575201e-01 3.46967399e-01 3.43274564e-01 -6.15816891e-01
4.28275377e-01 3.26150894e-01 -1.54802144e+00 7.63083935e-01
3.20561081e-01 3.59871835e-01 -3.30046415e-02 6.45262182e-01
-4.26306605e-01 1.05027258e+00 -6.98513269e-01 7.05970125e-03
3.28359932e-01 -2.57565141e-01 -1.12048995e+00 -4.22729492e-01
5.64187229e-01 -1.19931422e-01 -9.26994801e-01 4.41615701e-01
8.07240665e-01 7.23972321e-01 7.07648695e-01 -4.44664180e-01
-6.02579892e-01 9.32313085e-01 -6.35904968e-01 1.02743030e+00
6.95395052e-01 -4.50122654e-01 1.33814144e+00 -1.03294373e+00
6.88818514e-01 1.06687856e+00 9.12390828e-01 2.67205447e-01
-1.02819610e+00 -5.83846569e-01 2.19304830e-01 2.35365450e-01
-1.90689635e+00 -6.21980250e-01 3.66105229e-01 3.04513462e-02
1.62324011e+00 5.59624314e-01 6.62902892e-01 8.62789333e-01
6.64244294e-02 1.13634515e+00 7.26677418e-01 -9.26732600e-01
8.96031328e-04 9.92657095e-02 4.13281560e-01 6.84752405e-01
2.15264723e-01 7.78259784e-02 -7.93511808e-01 -5.89776397e-01
8.33295107e-01 5.51595837e-02 -1.42450139e-01 -2.36713693e-01
-8.56163383e-01 1.05268741e+00 5.21440983e-01 5.60133040e-01
-2.89566547e-01 1.95032731e-01 6.62860990e-01 7.00842977e-01
4.18350041e-01 5.45799315e-01 -1.58248305e-01 2.10915655e-02
-1.54589474e+00 6.45991147e-01 7.07646787e-01 8.34325731e-01
6.35948539e-01 3.02996665e-01 -4.06710356e-01 1.20514166e+00
7.98525754e-03 1.18378036e-01 8.76307786e-01 -5.33280313e-01
3.99752706e-01 1.47349104e-01 -5.90593927e-02 -1.00049186e+00
-4.64003831e-01 -3.93317163e-01 -4.36169416e-01 -3.53780448e-01
-3.04392606e-01 2.21258774e-01 -1.05672765e+00 1.40349579e+00
-1.00985929e-01 1.80627763e-01 -2.07208712e-02 8.99341226e-01
1.20663178e+00 9.39912021e-01 3.63425553e-01 -2.29924116e-02
1.13512838e+00 -1.02902687e+00 -5.12829125e-01 -7.91979656e-02
8.74590099e-01 -1.06703794e+00 9.59056735e-01 4.67503443e-02
-1.49758518e+00 -4.39773858e-01 -1.11202908e+00 -3.79465610e-01
-4.44572687e-01 -6.98609324e-03 1.03845274e+00 7.76707172e-01
-1.21401703e+00 3.30804706e-01 -6.52369618e-01 -4.31904644e-01
2.40546688e-01 4.70063031e-01 -3.45454574e-01 -2.27448404e-01
-1.64140773e+00 9.94812250e-01 6.41466200e-01 -4.46382523e-01
-8.85790050e-01 -7.06389904e-01 -9.55632091e-01 1.86891854e-01
1.43356606e-01 -9.10239041e-01 1.18328357e+00 -2.41314605e-01
-1.11010373e+00 1.02587008e+00 -3.01384833e-02 -7.21658766e-01
-1.68139398e-01 -3.12164873e-01 -3.48980576e-01 3.68805617e-01
1.76488191e-01 7.40890503e-01 8.36071134e-01 -8.55388522e-01
-2.83690602e-01 1.82031348e-01 1.02676712e-01 7.04609036e-01
-4.86552387e-01 2.59742886e-01 -7.23785579e-01 -1.29917049e+00
5.33418506e-02 -5.41786909e-01 -4.44033802e-01 -3.31496507e-01
-2.86676213e-02 -2.11318687e-01 3.05630714e-01 -4.77298826e-01
1.87202668e+00 -2.03525066e+00 1.98207662e-01 4.55782503e-01
1.02958612e-01 2.15512827e-01 -3.80957007e-01 9.69686449e-01
-3.04668725e-01 1.13654628e-01 1.80969149e-01 -3.88466507e-01
-3.09356779e-01 2.97246575e-01 -7.33335137e-01 4.57456529e-01
-2.35760301e-01 9.43988562e-01 -7.87237048e-01 -5.65456152e-01
1.55153498e-01 4.02969301e-01 -7.70690858e-01 -1.75655648e-01
-1.38523076e-02 -3.45912993e-01 -4.21054035e-01 7.20288157e-01
1.71976984e-01 -1.90960735e-01 1.76564768e-01 1.90724820e-01
-1.73238534e-02 1.00717449e+00 -1.28315175e+00 2.01811552e+00
-5.03062189e-01 4.72125828e-01 -8.55344348e-03 -1.15788388e+00
7.09272325e-01 4.68590051e-01 4.49588150e-01 -6.95941687e-01
-1.03663281e-02 1.24128543e-01 -6.30187273e-01 -8.49506482e-02
1.21249735e+00 -5.53499758e-01 -3.54194343e-01 8.76303792e-01
2.58927077e-01 -2.21660167e-01 1.74849942e-01 6.67501271e-01
1.03859401e+00 -1.55221850e-01 5.46484828e-01 -3.38559210e-01
2.77503878e-01 1.15581542e-01 -3.62222083e-03 1.23136353e+00
2.46951059e-01 7.32342780e-01 -2.50295222e-01 -4.76381987e-01
-9.31634247e-01 -1.13987851e+00 -7.07858950e-02 1.56137800e+00
-9.77760255e-02 -1.02056968e+00 -1.92764103e-01 -3.81010085e-01
8.51318687e-02 3.06822270e-01 -6.26535296e-01 -4.67683285e-01
-7.41052449e-01 -8.55278432e-01 8.11655819e-01 7.38048017e-01
1.56692892e-01 -1.31001055e+00 -1.87051911e-02 1.86872423e-01
-3.52043092e-01 -5.49332142e-01 -5.34093201e-01 3.55143964e-01
-1.02623796e+00 -6.63646340e-01 -7.63216257e-01 -1.18706656e+00
2.57054001e-01 8.77996683e-01 1.34910727e+00 5.08202255e-01
-7.33540595e-01 4.88887608e-01 -5.00062406e-01 -1.75812185e-01
-4.61186357e-02 7.02569067e-01 1.95666313e-01 -7.97120452e-01
5.22044718e-01 -3.29165518e-01 -7.03722060e-01 -1.81929991e-01
-1.16769314e+00 -7.10283995e-01 4.26551789e-01 8.88797641e-01
7.67215133e-01 -2.29377393e-02 5.33680320e-01 -7.63846874e-01
1.01928401e+00 -8.84753346e-01 7.46458918e-02 -2.29635425e-02
-3.82242084e-01 -1.37055859e-01 2.05748200e-01 -3.35356623e-01
-6.50525331e-01 -4.35510427e-01 -4.49716657e-01 -4.13031608e-01
3.13912481e-01 9.37203467e-01 6.14678442e-01 -2.67318457e-01
1.16617692e+00 3.69179338e-01 -1.44172087e-01 -5.87655604e-01
7.29826868e-01 5.58662415e-01 1.50177836e-01 -7.88750052e-01
6.52235925e-01 4.38342422e-01 -6.23916864e-01 -1.24216127e+00
-8.47229838e-01 -1.09394181e+00 -1.70091704e-01 4.17930186e-01
1.90498739e-01 -1.35231733e+00 1.38216779e-01 -1.21610366e-01
-5.18338263e-01 -6.27734885e-02 -7.75345862e-01 4.87477213e-01
-4.45927083e-01 6.78828597e-01 -1.05961525e+00 -3.60955089e-01
-6.62397742e-01 -6.22463882e-01 9.81625259e-01 7.53466273e-03
-6.74002826e-01 -9.98540759e-01 4.13815618e-01 3.69154483e-01
6.00323319e-01 -5.17695129e-01 9.91406977e-01 -6.96362674e-01
-3.72143954e-01 -5.71291149e-01 -5.35028987e-02 1.32292369e-02
-5.22512523e-03 -4.25632030e-01 -5.57908773e-01 -7.27611244e-01
-2.28842601e-01 -7.53715158e-01 1.42546284e+00 4.66110110e-01
1.05382872e+00 -1.55371293e-01 -4.27119464e-01 4.47843969e-01
1.29722607e+00 -3.08339268e-01 7.96042502e-01 6.51886821e-01
2.66072601e-01 6.72241673e-02 2.99963772e-01 4.32366908e-01
3.11522543e-01 6.02902472e-01 -2.31628105e-01 -1.99941456e-01
-5.66101968e-01 -3.43134165e-01 1.55809373e-01 1.18381500e+00
1.05758384e-01 -1.53181121e-01 -7.48728931e-01 9.34369862e-01
-1.59632874e+00 -1.47015512e+00 4.00951147e-01 2.07409215e+00
9.56151903e-01 1.28567830e-01 3.02604854e-01 2.15989366e-01
2.61023521e-01 6.43119574e-01 1.20172285e-01 -5.19551456e-01
-3.26745242e-01 1.02518594e+00 2.79346645e-01 7.81117618e-01
-1.16838360e+00 1.22390413e+00 8.95900917e+00 1.24814403e+00
-6.78927362e-01 1.82640880e-01 2.02065930e-01 -5.91841459e-01
-5.75734973e-01 -2.16105074e-01 -9.21195447e-01 -1.63345803e-02
8.40196669e-01 -3.56579155e-01 2.57131577e-01 6.70362949e-01
-1.99306950e-01 4.72508930e-02 -6.05554521e-01 8.15091908e-01
5.61376274e-01 -2.01468205e+00 6.08942330e-01 -4.18945640e-01
8.33701730e-01 3.41767490e-01 2.01976404e-01 9.07040238e-01
4.24482316e-01 -1.02011549e+00 2.01692730e-01 2.66935676e-01
7.13279247e-01 -6.14195406e-01 4.59998339e-01 2.36635372e-01
-1.38715708e+00 -1.29447043e-01 -6.47191465e-01 -1.57419875e-01
-6.11841679e-03 -3.06948051e-02 -5.50409436e-01 5.57436168e-01
6.28393531e-01 5.88113487e-01 -3.22748542e-01 1.30038416e+00
-5.54820485e-02 6.44496024e-01 -5.34393072e-01 -2.71147918e-02
4.78401244e-01 3.30350131e-01 4.14672136e-01 1.75340915e+00
1.87598951e-02 3.21208507e-01 5.12216330e-01 1.07773349e-01
-2.71527886e-01 6.58695698e-01 -8.23870122e-01 7.23123997e-02
6.66817784e-01 6.23952508e-01 -7.46564209e-01 -5.90627491e-01
-5.18390238e-01 6.50129497e-01 4.99154955e-01 4.07847255e-01
-5.19592285e-01 -3.04874986e-01 5.05294919e-01 3.83837193e-01
6.08066201e-01 -1.45367414e-01 -3.06412905e-01 -1.41243267e+00
-2.78562844e-01 -1.06144917e+00 1.01287794e+00 -3.71253848e-01
-1.39676988e+00 4.63738412e-01 2.96637684e-01 -1.09350193e+00
-4.02405635e-02 -1.77566826e-01 -4.27454531e-01 9.73968863e-01
-1.55455935e+00 -8.64185452e-01 3.69087100e-01 9.54998136e-01
8.36847961e-01 -4.14702415e-01 1.26610398e+00 5.40594280e-01
-1.71523556e-01 9.29118276e-01 1.52846947e-01 6.56252429e-02
7.41408467e-01 -9.13620889e-01 2.66104788e-01 4.57372308e-01
7.88646936e-01 1.27247739e+00 3.45820874e-01 -5.76846838e-01
-1.19975328e+00 -5.76726198e-01 9.01513159e-01 -5.26267767e-01
7.98148215e-01 3.27423923e-02 -9.73431051e-01 8.62420380e-01
2.97872633e-01 -1.70335591e-01 1.26535213e+00 6.90730393e-01
-5.96715093e-01 1.41952530e-01 -7.59784877e-01 6.19922340e-01
8.14933717e-01 -1.02082121e+00 -1.25703335e+00 5.65321088e-01
4.83058184e-01 -5.30763447e-01 -7.57213593e-01 3.18984747e-01
4.85495567e-01 -4.77820545e-01 1.71581101e+00 -8.82051289e-01
1.31346807e-01 9.31160375e-02 -3.12830895e-01 -1.06482399e+00
-5.36377728e-01 -6.75841689e-01 -4.60767567e-01 5.20183563e-01
2.82614380e-01 -1.27790853e-01 1.14908087e+00 5.87747991e-02
-7.16697574e-02 -7.56736040e-01 -1.02023721e+00 -5.48824191e-01
5.56320906e-01 -3.32231760e-01 2.87303627e-01 9.99454916e-01
4.38074201e-01 3.16894948e-01 -5.17783642e-01 -1.10466279e-01
3.48412126e-01 2.62301207e-01 4.65612739e-01 -7.19079792e-01
-2.98059404e-01 -5.65386117e-01 -3.56022954e-01 -1.46147990e+00
1.17050633e-01 -1.42457020e+00 -2.98847795e-01 -1.51066518e+00
3.75163943e-01 -6.67354703e-01 -8.20715368e-01 5.48162282e-01
-3.73282641e-01 7.26525545e-01 2.46096086e-02 4.92736608e-01
-7.60695398e-01 5.17014563e-01 7.63536215e-01 -2.85589963e-01
-3.03069055e-01 -1.32294342e-01 -1.00168574e+00 4.56013381e-01
8.37166548e-01 -5.64617753e-01 -5.76595306e-01 -5.78546762e-01
2.50598013e-01 9.84990671e-02 -1.17846660e-01 -6.15060091e-01
3.64055008e-01 2.61179775e-01 3.51822853e-01 -6.81224585e-01
4.40986753e-01 -3.26054364e-01 -2.85450280e-01 3.30041945e-01
-5.83626270e-01 2.33867034e-01 5.68929136e-01 5.35960555e-01
-5.24207890e-01 -3.98749799e-01 5.08787334e-01 -5.22101521e-01
-8.85662675e-01 1.89449966e-01 -9.19622362e-01 2.35726938e-01
5.95662177e-01 -2.05720723e-01 -2.45085046e-01 -5.00681043e-01
-1.10553062e+00 1.77698836e-01 1.07609212e-01 5.85555553e-01
8.59943390e-01 -1.30937421e+00 -7.64749110e-01 3.36486787e-01
1.36591941e-01 -5.45847952e-01 2.03785807e-01 2.89244622e-01
-6.76076472e-01 6.52603388e-01 6.82501048e-02 -1.21760078e-01
-1.24619627e+00 6.12022996e-01 1.35917321e-01 -6.72724426e-01
-9.49236095e-01 1.28805685e+00 -3.01581591e-01 -3.68817717e-01
5.38935542e-01 1.04200475e-01 -2.54402846e-01 2.07888693e-01
8.06361914e-01 2.71223426e-01 2.43046135e-01 -3.22490156e-01
-3.26732397e-01 4.60206300e-01 -7.55362213e-01 -7.90731981e-02
1.19195509e+00 -4.43134531e-02 -1.56273576e-03 2.54614383e-01
1.11374962e+00 2.75911480e-01 -2.13421240e-01 -5.36279202e-01
-4.94957492e-02 -4.02926981e-01 3.98666233e-01 -6.60648584e-01
-7.52893984e-01 4.35705602e-01 1.82632908e-01 6.05727034e-03
9.63361144e-01 1.21353768e-01 9.16275442e-01 8.05601895e-01
5.96464515e-01 -1.02848446e+00 2.25883231e-01 5.80594957e-01
7.08236098e-01 -9.82803643e-01 6.03451908e-01 -2.07454845e-01
-3.16486627e-01 8.16883683e-01 2.58985013e-01 -7.01487958e-01
9.04708087e-01 3.08821917e-01 1.26442924e-01 -4.17652458e-01
-7.91432023e-01 -1.68258920e-01 2.39853427e-01 2.94570744e-01
6.48695886e-01 -1.25460774e-01 -5.31734169e-01 2.64118761e-01
-2.75073588e-01 -1.74819604e-01 2.34037980e-01 1.31229329e+00
-7.97030807e-01 -1.33496404e+00 -5.02234340e-01 4.71621543e-01
-7.73550868e-01 -7.17866063e-01 -2.38355547e-01 9.64093626e-01
-3.88283938e-01 6.41613126e-01 -6.47180751e-02 -4.93021682e-02
3.73970985e-01 2.47095391e-01 6.41195178e-01 -1.26000535e+00
-1.05145586e+00 2.20321342e-01 2.23062977e-01 -5.37016094e-01
-3.98684293e-01 -6.28602624e-01 -9.85353112e-01 -4.82105702e-01
-3.38417113e-01 6.68680012e-01 1.07807204e-01 6.24228716e-01
2.66449839e-01 3.83647978e-01 4.65905428e-01 -8.27894807e-01
-5.60222983e-01 -7.94064283e-01 -6.93501592e-01 7.24996477e-02
2.15282813e-01 -6.09727561e-01 -2.41534725e-01 -3.59553397e-01] | [11.47475814819336, 7.665626525878906] |
2f0be9f2-56b4-4517-846f-daca9ffe9c1f | cognifnn-a-fuzzy-neural-network-framework-for | 2009.11485 | null | https://arxiv.org/abs/2009.11485v2 | https://arxiv.org/pdf/2009.11485v2.pdf | CogniFNN: A Fuzzy Neural Network Framework for Cognitive Word Embedding Evaluation | Word embeddings can reflect the semantic representations, and the embedding qualities can be comprehensively evaluated with human natural reading-related cognitive data sources. In this paper, we proposed the CogniFNN framework, which is the first attempt at using fuzzy neural networks to extract non-linear and non-stationary characteristics for evaluations of English word embeddings against the corresponding cognitive datasets. In our experiment, we used 15 human cognitive datasets across three modalities: EEG, fMRI, and eye-tracking, and selected the mean square error and multiple hypotheses testing as metrics to evaluate our proposed CogniFNN framework. Compared to the recent pioneer framework, our proposed CogniFNN showed smaller prediction errors of both context-independent (GloVe) and context-sensitive (BERT) word embeddings, and achieved higher significant ratios with randomly generated word embeddings. Our findings suggested that the CogniFNN framework could provide a more accurate and comprehensive evaluation of cognitive word embeddings. It will potentially be beneficial to the further word embeddings evaluation on extrinsic natural language processing tasks. | ['Son Tran', 'Xinping Liu', 'Zehong Cao'] | 2020-09-24 | null | null | null | null | ['embeddings-evaluation'] | ['natural-language-processing'] | [-1.76352318e-02 -1.68642119e-01 7.20353350e-02 -4.60968286e-01
-8.26631635e-02 -3.75187010e-01 6.05487168e-01 4.77581888e-01
-1.07050550e+00 4.31268871e-01 3.29701096e-01 -2.38268867e-01
-4.98414397e-01 -9.11452949e-01 -1.56418636e-01 -2.11226106e-01
-8.39357674e-02 5.90862371e-02 2.64671952e-01 -3.92024100e-01
6.74892068e-01 1.99190721e-01 -1.74344337e+00 2.79988516e-02
1.26065922e+00 1.08979607e+00 5.01213431e-01 3.15902114e-01
-2.68035591e-01 4.01925474e-01 -5.24709046e-01 -3.70032549e-01
-1.77091643e-01 -1.78012028e-01 -7.49643445e-01 -4.48770493e-01
-1.09994240e-01 1.89220339e-01 -2.23159328e-01 1.52258480e+00
7.53683448e-01 6.40024245e-01 7.26354122e-01 -9.88004208e-01
-2.00670528e+00 3.23044002e-01 2.08440088e-02 8.20076585e-01
5.44784725e-01 2.41067773e-03 8.69326711e-01 -1.32941782e+00
1.63188979e-01 1.41737199e+00 4.84779596e-01 5.84873497e-01
-8.90728652e-01 -7.45349288e-01 1.30092055e-01 8.45361292e-01
-1.46349037e+00 -7.86673799e-02 5.93640745e-01 -3.40544164e-01
1.05450416e+00 8.43652114e-02 7.49647021e-01 1.16221273e+00
5.00843465e-01 4.89124298e-01 1.54630470e+00 -6.51410699e-01
2.95879483e-01 3.28736097e-01 9.05514419e-01 4.94261861e-01
3.79610449e-01 3.84095132e-01 -4.91424263e-01 5.98737150e-02
5.94242334e-01 1.17423467e-01 -5.03842771e-01 2.54834533e-01
-1.29787815e+00 8.45637143e-01 5.24654567e-01 5.62959015e-01
-4.35064077e-01 -1.17370307e-01 4.43203151e-01 2.18382686e-01
4.26100940e-01 6.36568487e-01 -3.83185416e-01 -1.40150011e-01
-4.54988301e-01 2.42853467e-03 1.08583853e-01 7.68445253e-01
5.05396962e-01 9.19966921e-02 -4.02333587e-01 9.32941616e-01
4.13102090e-01 5.69244027e-01 1.28085232e+00 -5.32410204e-01
-6.22921251e-03 6.95610225e-01 -1.97671518e-01 -1.29294538e+00
-5.61448753e-01 -5.66122867e-02 -4.54320073e-01 -2.44417358e-02
-1.48347586e-01 8.84138271e-02 -6.64983571e-01 1.78972924e+00
-1.54995099e-01 1.15152657e-01 1.95839569e-01 1.10680163e+00
1.11138391e+00 4.51444089e-01 4.49771494e-01 -2.41785869e-01
1.80583858e+00 -6.18734419e-01 -1.37126493e+00 -2.62673438e-01
2.83959746e-01 -4.93714362e-01 1.63647747e+00 1.68553889e-01
-8.75976741e-01 -1.00785911e+00 -1.24103785e+00 -1.13832660e-01
-9.19610739e-01 -1.04317553e-01 6.55140698e-01 7.93397665e-01
-1.06213975e+00 2.86994755e-01 -6.33857131e-01 -3.32059622e-01
3.23975950e-01 -3.25919054e-02 -3.51232350e-01 -1.37088904e-02
-1.86173522e+00 1.39582717e+00 8.24574351e-01 1.53577626e-01
-5.81510663e-01 -4.32682306e-01 -1.07429731e+00 1.03990674e-01
2.04137892e-01 -3.43706459e-01 6.43170834e-01 -5.97980857e-01
-1.06469548e+00 7.12701142e-01 -2.92138487e-01 -2.50234336e-01
-3.45337361e-01 -1.93249837e-01 -1.11635280e+00 3.54242511e-02
-6.38323128e-02 6.18545294e-01 4.44393754e-01 -7.69770086e-01
-6.14710078e-02 -3.87252688e-01 4.76334430e-02 1.75280631e-01
-1.06021941e+00 1.01832293e-01 6.83417544e-02 -9.19611454e-01
2.06509270e-02 -4.81616586e-01 2.18712002e-01 1.19297735e-01
2.94605136e-01 -8.15554202e-01 3.52371067e-01 -6.33078098e-01
1.44279051e+00 -2.16803861e+00 7.43985474e-02 2.57579852e-02
3.50752890e-01 6.18416309e-01 -4.28597093e-01 2.78924908e-02
-1.60779834e-01 2.88623422e-01 -3.24733816e-02 3.44762206e-01
2.89017230e-01 1.38569117e-01 4.43036901e-03 1.84858516e-01
4.13538754e-01 1.21642053e+00 -1.09584010e+00 -5.36934495e-01
2.27088630e-01 4.73510295e-01 -4.31055248e-01 9.63393375e-02
1.84059143e-01 -4.82085407e-01 -2.85227299e-01 4.61232662e-01
6.64168775e-01 1.62121236e-01 -1.90615043e-01 -3.37499499e-01
4.61994894e-02 6.85329735e-02 -9.90264595e-01 1.58125150e+00
-4.13558215e-01 9.57995176e-01 -6.10703290e-01 -9.77633178e-01
1.25362527e+00 1.83996961e-01 -2.18043342e-01 -1.23598444e+00
4.66155618e-01 -1.12886436e-01 1.82604834e-01 -8.71195674e-01
6.46590412e-01 -2.17778653e-01 1.54134668e-02 5.55160046e-01
4.25264925e-01 3.84847522e-01 8.92892927e-02 -9.45840329e-02
7.53277361e-01 -1.76934183e-01 5.74536920e-01 -9.09951806e-01
9.56876695e-01 -3.93302023e-01 4.63814080e-01 3.97943228e-01
-6.91129923e-01 3.70779008e-01 1.25300497e-01 -3.91065031e-01
-4.60457742e-01 -1.44151473e+00 -5.56465805e-01 1.04435325e+00
4.07543778e-01 -5.28822482e-01 -6.62733734e-01 -3.58600020e-01
-2.12555155e-01 1.21158707e+00 -6.95432603e-01 -7.88756788e-01
-1.05247103e-01 -8.47017348e-01 6.85617030e-01 9.30548787e-01
5.48961580e-01 -1.22876143e+00 -5.73556125e-01 6.75189570e-02
-1.07790969e-01 -8.52279961e-01 -7.07622230e-01 1.78438783e-01
-5.68818688e-01 -1.10722661e+00 -3.72190505e-01 -1.02522635e+00
5.00493646e-01 1.60437584e-01 9.32997286e-01 -2.47325953e-02
-3.21777940e-01 5.17809272e-01 -6.29977047e-01 -7.53480554e-01
2.23591533e-02 -3.48236173e-01 5.38364172e-01 -2.68078893e-01
1.27157235e+00 -2.11623341e-01 -5.26544213e-01 1.71176791e-01
-1.18704486e+00 -5.67402124e-01 3.42139900e-01 9.36462939e-01
4.31075603e-01 6.35702983e-02 7.77033567e-01 -1.61457047e-01
1.82709551e+00 -3.51799011e-01 -1.49001330e-01 5.54579437e-01
-1.14878416e+00 2.32764900e-01 5.79783261e-01 -6.93647325e-01
-9.49941576e-01 -8.56992662e-01 -4.90767770e-02 -2.03115657e-01
-1.83182105e-01 6.67305470e-01 -4.88090217e-02 -5.21464050e-02
8.65187824e-01 4.76292104e-01 -2.91143800e-03 -1.67089298e-01
3.59704047e-01 9.04513121e-01 4.48038995e-01 -7.57676601e-01
3.51196438e-01 -1.30426452e-01 -4.61703807e-01 -8.12381506e-01
-5.78782797e-01 -1.60005465e-02 -6.28918946e-01 -2.14492157e-01
1.25934994e+00 -6.76255465e-01 -8.52238357e-01 1.11466765e-01
-1.44162762e+00 4.38529700e-01 -9.88984033e-02 8.55555356e-01
-2.14908227e-01 5.22735596e-01 -4.23120320e-01 -8.27010512e-01
-4.47586954e-01 -1.11365962e+00 4.84827161e-01 3.73902082e-01
-5.27310371e-01 -1.39742923e+00 -8.38132389e-03 5.36465235e-02
6.19065225e-01 -1.83203459e-01 1.12822056e+00 -9.14035439e-01
2.01711953e-01 -7.30959475e-02 -4.75615978e-01 6.63219154e-01
2.45583445e-01 -3.53081375e-01 -9.99307692e-01 3.28748138e-03
3.48059863e-01 -4.15789962e-01 8.06539178e-01 1.82142064e-01
1.45281124e+00 5.77422231e-02 1.55504912e-01 2.86025465e-01
1.33312702e+00 6.59113109e-01 7.88037479e-01 4.71647501e-01
2.26923883e-01 4.05106425e-01 5.62768698e-01 1.92168429e-01
5.91542780e-01 2.41476357e-01 1.03897266e-01 3.07709724e-01
4.86986451e-02 1.03355981e-02 4.23111290e-01 1.18307769e+00
-2.20803246e-01 -2.01486290e-01 -9.34365749e-01 5.12944043e-01
-1.44056034e+00 -1.00710332e+00 4.04729927e-03 1.78994632e+00
7.66443729e-01 1.63897604e-01 -4.15918797e-01 2.46138856e-01
7.85046160e-01 1.68803349e-01 -4.36541766e-01 -6.94904387e-01
-2.44202882e-01 6.28935635e-01 1.35865752e-02 2.72007763e-01
-8.34917128e-01 7.52547443e-01 6.71426916e+00 7.91281879e-01
-7.03343809e-01 3.84364665e-01 -2.19395105e-02 4.07916047e-02
-6.37452066e-01 -4.94465619e-01 -4.59843129e-01 5.17878294e-01
1.24395740e+00 -6.65575981e-01 5.11390686e-01 4.76960272e-01
7.73700699e-02 1.40950069e-01 -9.57940638e-01 1.17763555e+00
4.33572143e-01 -1.12552166e+00 2.15329438e-01 -2.13340417e-01
2.66514868e-01 -3.58140647e-01 1.53975621e-01 6.22057498e-01
-2.68228054e-01 -1.27987015e+00 4.83389735e-01 1.11773813e+00
6.08907223e-01 -8.15134883e-01 9.47813690e-01 2.59501785e-01
-1.13030505e+00 -7.81643614e-02 -7.98360348e-01 -3.86307657e-01
-6.14015833e-02 3.68984759e-01 -2.51341194e-01 4.19980586e-01
6.12863004e-01 7.15957344e-01 -8.54326904e-01 5.43496311e-01
-1.28119305e-01 3.13872308e-01 1.92688987e-01 -7.20270336e-01
-1.07301436e-02 -2.25307018e-01 9.78867486e-02 1.15413761e+00
3.43777090e-01 3.57827604e-01 -9.01352763e-02 1.29764259e+00
7.13841468e-02 2.54566967e-01 -3.83669525e-01 -3.80401731e-01
8.15974057e-01 1.03629994e+00 -6.26062870e-01 -2.19011083e-01
-6.07754230e-01 8.35148096e-01 3.92986476e-01 2.31283754e-01
-1.06273544e+00 -8.59292865e-01 8.73437345e-01 -1.90689951e-01
-1.89801946e-01 -3.48097295e-01 -3.49312812e-01 -1.27725434e+00
4.55107689e-02 -4.84503418e-01 3.65611196e-01 -1.14296389e+00
-1.63820481e+00 8.00391972e-01 1.48616597e-01 -9.80531514e-01
1.68823853e-01 -1.18352926e+00 -5.91401577e-01 1.17873502e+00
-1.48315358e+00 -3.97200674e-01 -4.14380640e-01 6.40969694e-01
5.31244159e-01 -4.82316345e-01 1.22658932e+00 1.92588866e-01
-4.73242342e-01 6.56729162e-01 -2.18878388e-01 2.28098750e-01
6.57362223e-01 -1.08092809e+00 1.18385538e-01 7.41030693e-01
1.68085024e-01 1.20675957e+00 3.58630896e-01 -4.27974105e-01
-1.23395264e+00 -8.54696810e-01 9.31938231e-01 -5.87514162e-01
7.07581162e-01 -1.37146577e-01 -1.07872546e+00 4.19341564e-01
5.38066328e-01 -7.47216940e-02 1.09728193e+00 -6.04227465e-03
-1.38365000e-01 7.04210848e-02 -1.16359484e+00 6.27829611e-01
1.13682711e+00 -8.60329747e-01 -1.56726658e+00 7.98180327e-02
1.07079017e+00 1.62402287e-01 -1.12153590e+00 2.10386321e-01
3.48934054e-01 -7.23505795e-01 1.09413660e+00 -6.59301758e-01
4.02057111e-01 -2.36709222e-01 -4.53713894e-01 -1.62887120e+00
-8.18440616e-01 3.27889353e-01 -1.35997022e-02 1.25395274e+00
1.22777455e-01 -8.83974850e-01 3.33373100e-02 5.95612407e-01
-2.17104226e-01 -5.31221390e-01 -8.30900550e-01 -9.35400963e-01
3.00968260e-01 -8.07876885e-01 6.84910715e-01 9.56271291e-01
2.62560308e-01 2.79818594e-01 1.89872295e-01 -6.14244528e-02
4.37484026e-01 -4.18408930e-01 -3.62094194e-01 -1.46921730e+00
3.35509241e-01 -5.69225848e-01 -8.62613022e-01 -6.18042827e-01
7.08796799e-01 -1.38259768e+00 -2.57141829e-01 -1.54446352e+00
1.29756883e-01 2.28907868e-01 -9.48251843e-01 2.87980646e-01
-5.54185629e-01 -2.97896769e-02 4.04496007e-02 -2.10310653e-01
-3.72255594e-01 1.08445013e+00 1.34599721e+00 -3.13965678e-01
1.74857348e-01 -5.85483491e-01 -1.03149021e+00 6.22115850e-01
7.44807243e-01 -1.57159135e-01 -8.27040076e-01 -6.17589831e-01
5.23451082e-02 -3.65558743e-01 4.48799610e-01 -1.01861906e+00
2.72194952e-01 -1.11975536e-01 6.52668297e-01 -2.32288063e-01
1.96915895e-01 -8.38574648e-01 -4.84706879e-01 3.25205564e-01
-5.20121753e-01 6.07915580e-01 3.24773610e-01 5.13329208e-01
-3.49909157e-01 -4.34388846e-01 6.14519596e-01 -1.77802518e-01
-1.33703375e+00 3.57489437e-02 -3.77361208e-01 3.71714920e-01
1.11512327e+00 -5.06137371e-01 -4.83333141e-01 6.20424785e-02
-6.62416518e-01 1.95383742e-01 -2.08977517e-02 9.76454020e-01
1.29439294e+00 -1.85753667e+00 -4.51429188e-01 5.04581511e-01
3.88883173e-01 -6.63287640e-01 1.67417482e-01 5.66473961e-01
-2.71649241e-01 5.51658273e-01 -6.14361048e-01 -2.91166693e-01
-8.08698297e-01 8.34491372e-01 2.11149156e-01 4.71262306e-01
-2.75552481e-01 7.10752606e-01 -9.06439424e-02 -6.81919336e-01
1.02202050e-01 -4.79489207e-01 -6.62665546e-01 3.39426160e-01
9.66554105e-01 6.76632762e-01 1.27408653e-01 -5.47939122e-01
-6.73820674e-01 5.05844712e-01 2.32319278e-03 -2.54483782e-02
1.23036683e+00 -5.98363876e-02 -2.75342077e-01 6.80979848e-01
1.18656731e+00 -4.71556574e-01 -5.49516499e-01 -2.67768383e-01
2.22683549e-01 -4.11357313e-01 3.07486802e-01 -7.68508434e-01
-8.06197047e-01 1.07978463e+00 1.12358689e+00 1.48092270e-01
1.19182801e+00 -1.84875429e-01 6.51152194e-01 3.92080992e-01
5.25478959e-01 -1.53008103e+00 1.82187542e-01 6.93815172e-01
8.21686387e-01 -1.08050847e+00 -1.36172578e-01 -2.06479314e-03
-7.42798388e-01 1.42117310e+00 9.14716244e-01 -1.63170069e-01
9.59952712e-01 -4.94794734e-02 -8.57671499e-02 -3.53860587e-01
-6.86598063e-01 -2.77839065e-01 6.16659820e-01 9.02469456e-01
7.81342685e-01 1.80678383e-01 -9.66081798e-01 1.30101359e+00
-2.24406332e-01 6.80194870e-02 4.27603692e-01 7.50022411e-01
-7.39039838e-01 -8.27423275e-01 -4.30962801e-01 5.37753403e-01
6.10266887e-02 -3.70083600e-01 -1.42594576e-01 6.05536938e-01
3.31555903e-01 1.02516651e+00 2.64784396e-01 -6.97252929e-01
8.16244543e-01 4.71880227e-01 4.58807230e-01 -4.74538326e-01
-3.33788991e-01 -5.50556481e-01 -2.86233872e-01 -5.16992450e-01
-5.32510996e-01 -5.10093629e-01 -1.49636829e+00 -1.61590010e-01
-3.36606354e-01 1.04888998e-01 3.79472166e-01 1.08553505e+00
2.68237740e-01 8.37929070e-01 1.41552776e-01 -4.80723470e-01
-6.01091325e-01 -1.19954693e+00 -5.52003205e-01 5.37064910e-01
-1.72941282e-01 -1.14359272e+00 -4.58416522e-01 -1.86490133e-01] | [10.555442810058594, 8.75291633605957] |
0813e515-f436-48c1-9e61-9e7a11892e9d | l3cube-mahaner-a-marathi-named-entity | 2204.06029 | null | https://arxiv.org/abs/2204.06029v1 | https://arxiv.org/pdf/2204.06029v1.pdf | L3Cube-MahaNER: A Marathi Named Entity Recognition Dataset and BERT models | Named Entity Recognition (NER) is a basic NLP task and finds major applications in conversational and search systems. It helps us identify key entities in a sentence used for the downstream application. NER or similar slot filling systems for popular languages have been heavily used in commercial applications. In this work, we focus on Marathi, an Indian language, spoken prominently by the people of Maharashtra state. Marathi is a low resource language and still lacks useful NER resources. We present L3Cube-MahaNER, the first major gold standard named entity recognition dataset in Marathi. We also describe the manual annotation guidelines followed during the process. In the end, we benchmark the dataset on different CNN, LSTM, and Transformer based models like mBERT, XLM-RoBERTa, IndicBERT, MahaBERT, etc. The MahaBERT provides the best performance among all the models. The data and models are available at https://github.com/l3cube-pune/MarathiNLP . | ['Raviraj Joshi', 'Onkar Litake', 'Maithili Sabane', 'Aparna Ranade', 'Parth Patil'] | 2022-04-12 | null | https://aclanthology.org/2022.wildre-1.6 | https://aclanthology.org/2022.wildre-1.6.pdf | wildre-lrec-2022-6 | ['slot-filling'] | ['natural-language-processing'] | [-6.40937686e-01 -8.49405602e-02 9.72017720e-02 -4.17622805e-01
-8.49666119e-01 -6.28856778e-01 7.37080872e-01 9.54324603e-02
-9.12781298e-01 1.30599582e+00 5.66337883e-01 -3.59597802e-01
2.22193047e-01 -6.84019446e-01 -3.90629143e-01 -4.08553660e-01
-6.93457052e-02 9.36890900e-01 2.05727890e-01 -5.94056010e-01
2.32364908e-01 5.44324398e-01 -7.38975108e-01 4.84528512e-01
6.77025497e-01 5.37236094e-01 1.68548048e-01 6.42332613e-01
-6.87126040e-01 1.17838037e+00 -8.65494311e-01 -5.17569423e-01
-1.50001869e-01 -3.39955866e-01 -1.40632844e+00 -7.91490078e-01
3.14628296e-02 9.86078084e-02 -6.48829341e-01 6.51769638e-01
9.30500150e-01 7.04608619e-01 2.55868852e-01 -1.17695343e+00
-3.48014563e-01 1.01192343e+00 -2.00258225e-01 5.84694326e-01
2.73504674e-01 -3.91550571e-01 7.60412276e-01 -1.06669509e+00
9.72603500e-01 1.15836215e+00 8.02754223e-01 7.74133444e-01
-3.22674245e-01 -6.86830461e-01 -3.87248248e-01 4.98509645e-01
-1.52440405e+00 -5.91711879e-01 2.37906203e-01 1.07251644e-01
1.44768465e+00 2.05850378e-01 3.46369296e-01 1.02698874e+00
2.01642320e-01 1.07704759e+00 9.86635208e-01 -3.02878618e-01
1.32954670e-02 1.24495037e-01 2.89527684e-01 4.41841871e-01
-2.40633801e-01 -3.10597539e-01 -9.26121831e-01 -1.66453749e-01
2.52931982e-01 -2.51581788e-01 -2.59856224e-01 4.23225552e-01
-1.28668106e+00 7.09052920e-01 3.21521401e-01 6.55407965e-01
-4.57778543e-01 -4.29596268e-02 7.27337718e-01 3.61957192e-01
4.36086923e-01 3.61325055e-01 -8.11396420e-01 -6.75346076e-01
-7.39124417e-01 6.86542168e-02 1.14539325e+00 1.15257537e+00
4.34366584e-01 5.60860485e-02 -2.62681186e-01 1.32343185e+00
2.28837013e-01 4.56911653e-01 5.98171115e-01 -3.39216292e-01
6.96574867e-01 4.66628581e-01 1.86813161e-01 -5.77055871e-01
-5.77740431e-01 -1.11380011e-01 -7.55903900e-01 -4.88228559e-01
2.09117115e-01 -5.70711613e-01 -1.03696430e+00 1.35955358e+00
4.24293101e-01 9.80787426e-02 5.76276422e-01 6.48716986e-01
1.40848720e+00 1.14698935e+00 3.18649024e-01 2.32138261e-01
1.64245439e+00 -1.22649372e+00 -9.16053057e-01 -2.59398699e-01
7.45748878e-01 -1.02603483e+00 5.43284178e-01 -3.40688862e-02
-6.28645420e-01 -9.58512798e-02 -5.72898626e-01 -3.40496182e-01
-9.18005705e-01 4.04790521e-01 4.89159018e-01 5.66043139e-01
-1.02475488e+00 3.82177234e-01 -7.17402339e-01 -9.59021866e-01
2.19487324e-01 2.35451952e-01 -8.28462362e-01 1.14900634e-01
-1.71553898e+00 1.35418570e+00 6.61519945e-01 5.40254474e-01
-8.08046877e-01 -4.32736784e-01 -9.09917116e-01 -1.06706671e-01
1.00142129e-01 -5.65347634e-03 1.61508191e+00 -1.11212835e-01
-1.36016822e+00 9.61251318e-01 -4.74172324e-01 -7.22795844e-01
3.13636512e-01 -3.80335480e-01 -8.27475667e-01 -3.58436733e-01
2.22044259e-01 5.76087892e-01 -1.58318520e-01 -6.34533823e-01
-9.42134440e-01 -3.18139315e-01 -1.69751197e-01 3.03214461e-01
1.19749419e-01 6.55415058e-01 -3.07338357e-01 -3.17850828e-01
-2.25997344e-01 -8.48968387e-01 -1.47782475e-01 -8.43651474e-01
-6.32723093e-01 -6.75800204e-01 7.94081569e-01 -1.02587819e+00
1.21808600e+00 -2.01559782e+00 -4.45969045e-01 -1.04116261e-01
-2.10940570e-01 6.07464492e-01 -5.90821020e-02 1.14934063e+00
-7.89859369e-02 2.27503985e-01 -2.97456961e-02 -2.05720648e-01
-1.52156562e-01 2.53229469e-01 -2.18556195e-01 2.81819493e-01
1.89910114e-01 9.75478888e-01 -7.12418735e-01 -5.58781505e-01
4.03772295e-02 6.82546556e-01 2.40574196e-01 4.24885929e-01
1.19756013e-01 5.41638196e-01 -3.15234572e-01 5.48586011e-01
4.56928819e-01 1.10766374e-01 1.85214002e-02 -1.46794617e-01
-6.05987191e-01 8.92152905e-01 -1.08651924e+00 1.71947503e+00
-5.84811389e-01 9.38774109e-01 -6.51661977e-02 -8.15090477e-01
1.21645963e+00 7.56448507e-01 1.15077540e-01 -6.59929633e-01
1.86755776e-01 5.26877940e-01 -1.39910772e-01 -6.06489122e-01
9.17911649e-01 6.56443909e-02 -2.74899989e-01 2.89283246e-01
3.92192483e-01 4.59585696e-01 4.51125830e-01 2.29182281e-02
1.07929075e+00 -1.67269811e-01 5.50163746e-01 -8.96117091e-02
6.18702292e-01 3.84927243e-01 1.06087518e+00 7.40709543e-01
-4.71919596e-01 4.39636171e-01 1.40642539e-01 -4.84028369e-01
-1.07667136e+00 -7.47453272e-01 -8.46939981e-02 1.27102971e+00
-2.31055543e-01 -2.48875260e-01 -4.97505188e-01 -5.89907467e-01
-5.88081658e-01 9.18829978e-01 -4.29164320e-01 3.16436350e-01
-9.99612153e-01 -5.96789777e-01 1.22177994e+00 3.59434277e-01
1.17015696e+00 -1.77211618e+00 -1.91749543e-01 4.81587380e-01
-6.66401446e-01 -1.12845790e+00 -5.68286836e-01 2.32796893e-01
-5.80695689e-01 -7.55728006e-01 -9.77110982e-01 -1.14816213e+00
1.70399129e-01 -7.57495016e-02 1.22035158e+00 -4.24008250e-01
-2.59077847e-01 2.68319715e-02 -6.24271691e-01 -6.12244785e-01
-2.28174180e-01 5.62291861e-01 -1.93080053e-01 -2.44998068e-01
7.43581831e-01 -3.48114580e-01 -3.74493778e-01 2.98365712e-01
-3.91449094e-01 -1.03112765e-01 4.27685916e-01 8.74199033e-01
1.61236301e-01 -1.78538427e-01 6.09064817e-01 -1.21931219e+00
5.73794663e-01 -6.25210106e-01 -1.37888402e-01 4.55849975e-01
3.32440846e-02 -2.78047808e-02 2.18562722e-01 -1.13491066e-01
-1.52989566e+00 -7.46668503e-02 -7.60941267e-01 2.41293207e-01
-5.20315170e-01 6.06966674e-01 -2.69628465e-01 1.29426703e-01
5.99535763e-01 3.64823937e-01 -7.34198689e-01 -7.84216285e-01
2.78349757e-01 1.30786586e+00 4.78015363e-01 -3.06934774e-01
1.56135842e-01 1.58699632e-01 -7.63249159e-01 -1.21220088e+00
-7.03487098e-01 -7.80407310e-01 -5.61710835e-01 -1.86788663e-01
9.76584852e-01 -8.72285604e-01 -7.37037420e-01 6.42895579e-01
-1.62094092e+00 -1.98134348e-01 -1.12022333e-01 6.15525782e-01
-1.73415110e-01 1.18898535e-02 -1.09158814e+00 -1.12900996e+00
-1.04323268e+00 -6.03629231e-01 7.47282684e-01 7.88995743e-01
-2.06662923e-01 -8.63993347e-01 4.12128568e-01 4.80590880e-01
5.53510189e-01 -1.96473792e-01 5.83873332e-01 -1.68884027e+00
2.46369150e-02 1.90871041e-02 -7.40905255e-02 6.51934817e-02
5.18493578e-02 -2.59204715e-01 -9.28006351e-01 -4.10567671e-02
-2.31230289e-01 -1.97219923e-01 6.32908702e-01 -1.23979142e-02
2.89320707e-01 -3.26460451e-01 -3.13963860e-01 9.94774997e-02
1.08319938e+00 9.27881002e-01 9.04961348e-01 6.57240987e-01
5.13297796e-01 6.68399274e-01 7.38708258e-01 3.02379876e-01
7.00625837e-01 2.79328465e-01 -3.43578666e-01 -1.54526681e-01
-3.08727957e-02 -3.51553559e-01 3.97578776e-01 1.23434341e+00
3.10079828e-02 -4.65023071e-01 -1.48010588e+00 1.00494301e+00
-1.84229028e+00 -9.53071535e-01 -2.22677097e-01 1.89434671e+00
9.26947892e-01 -3.09365094e-01 -4.26424891e-02 -3.58898431e-01
1.09577167e+00 1.03161372e-01 -2.78659225e-01 -5.36098361e-01
-2.89616346e-01 3.62090856e-01 5.43888807e-01 2.00444043e-01
-1.24874115e+00 1.31674302e+00 5.26470375e+00 8.29314470e-01
-1.03840661e+00 4.71734792e-01 5.70016742e-01 4.52404946e-01
3.24177921e-01 1.53441489e-01 -1.27097678e+00 2.29682893e-01
1.44549251e+00 -2.25400314e-01 1.03076808e-01 8.99171174e-01
2.61023372e-01 6.78515211e-02 -6.78841829e-01 9.37477827e-01
-2.74516102e-02 -1.37532103e+00 -3.15521806e-01 -3.40597659e-01
4.62379098e-01 8.67432892e-01 -6.47383928e-01 6.65570319e-01
4.53482211e-01 -9.75088239e-01 3.22674543e-01 4.10151303e-01
4.25631016e-01 -9.28997457e-01 1.24531245e+00 2.38569334e-01
-1.19897234e+00 2.36401081e-01 -4.35797065e-01 5.17637491e-01
5.01855552e-01 3.07007253e-01 -1.06764710e+00 5.18864691e-01
9.23781812e-01 2.89593309e-01 -9.44209024e-02 1.44052911e+00
-2.74100751e-01 9.49038208e-01 -5.10356665e-01 -3.52997392e-01
3.52546751e-01 -5.01907757e-03 5.75350344e-01 1.69931257e+00
3.82195622e-01 2.30871990e-01 -8.41476619e-02 3.13010722e-01
-2.75380284e-01 6.27651572e-01 -5.50751984e-01 -3.56288403e-01
7.26313293e-01 1.27726638e+00 -5.40403426e-01 -3.30910802e-01
-3.60514760e-01 1.03406310e+00 4.19504881e-01 2.12036952e-01
-6.99979007e-01 -9.24802125e-01 5.24938524e-01 -2.95983195e-01
2.08768159e-01 -3.31018478e-01 2.20460117e-01 -1.30720663e+00
-2.64985561e-01 -9.46681201e-01 7.09117174e-01 -8.01940799e-01
-1.29352903e+00 1.19454944e+00 -2.52397358e-01 -9.30211186e-01
-2.24331945e-01 -6.41869605e-01 -9.13076162e-01 8.15853953e-01
-1.28807402e+00 -1.33132231e+00 -1.63747117e-01 5.61357379e-01
9.08515096e-01 -4.46612030e-01 9.90607738e-01 8.51330340e-01
-9.56971884e-01 4.77976620e-01 2.70169795e-01 8.95696998e-01
8.50706577e-01 -1.10854244e+00 1.04289448e+00 6.94770575e-01
3.93196434e-01 6.21383965e-01 5.07396042e-01 -7.50550866e-01
-1.06995082e+00 -9.78332520e-01 1.78279793e+00 -1.15037002e-01
6.41597748e-01 -3.97718430e-01 -8.71501207e-01 8.77601266e-01
7.02908278e-01 -3.14841539e-01 8.34459543e-01 1.45293251e-01
-4.34822254e-02 4.37143818e-02 -1.15643454e+00 4.92226154e-01
7.61578023e-01 -4.54938620e-01 -6.97549105e-01 4.39779133e-01
6.13734961e-01 -4.93301243e-01 -1.05461681e+00 3.08026634e-02
1.48756087e-01 -7.05070734e-01 4.46919382e-01 -9.27213192e-01
-2.18560427e-01 -2.03854829e-01 -1.37110362e-02 -1.32812297e+00
8.18895623e-02 -7.32185662e-01 3.03856671e-01 1.88161469e+00
1.01024270e+00 -8.54383528e-01 7.58999288e-01 5.05640388e-01
-1.31866947e-01 -3.52236480e-01 -1.25761735e+00 -5.18834770e-01
-3.95932123e-02 -4.33575600e-01 5.93343735e-01 1.12911463e+00
1.19824156e-01 5.98742723e-01 -5.56100309e-01 -1.11228442e-02
1.83608264e-01 -1.76099017e-01 5.29515386e-01 -8.78638208e-01
3.65311086e-01 1.77024588e-01 -3.87380213e-01 -9.53916848e-01
-2.20898986e-02 -7.90968716e-01 3.53248864e-01 -1.81363010e+00
-7.61870816e-02 -5.52934349e-01 -2.03062564e-01 1.04296744e+00
2.85868794e-01 1.58009976e-01 2.27412477e-01 2.56845683e-01
-5.11439204e-01 6.09200656e-01 5.41772902e-01 -1.44729778e-01
-3.57124895e-01 5.19790724e-02 -2.21120551e-01 4.55356658e-01
1.38491392e+00 -7.18279660e-01 1.46661267e-01 -5.12720585e-01
2.02439725e-01 1.04035422e-01 -2.43673071e-01 -9.22067165e-01
6.65536225e-01 2.29717102e-02 1.80350125e-01 -9.67978299e-01
2.74770737e-01 -3.38332415e-01 2.12203622e-01 2.52951860e-01
-3.87395084e-01 3.93162608e-01 2.59135962e-01 4.60439511e-02
-5.34852862e-01 -4.30349112e-01 6.01426303e-01 -3.06129813e-01
-1.23793101e+00 1.95331097e-01 -7.62687683e-01 3.35174829e-01
7.08753526e-01 2.32354581e-01 -5.00768721e-01 -2.79006749e-01
-7.72063911e-01 4.76128936e-01 -3.75578612e-01 6.72974527e-01
4.56790239e-01 -1.25126672e+00 -1.01090801e+00 -2.10495725e-01
1.60194308e-01 -3.22708398e-01 3.92885566e-01 8.22072744e-01
-9.71720397e-01 8.05267453e-01 -3.67093712e-01 -2.24609766e-02
-1.20603466e+00 -8.38051736e-02 3.67576480e-01 -4.04574722e-01
-5.85748911e-01 9.12991047e-01 -3.65394592e-01 -1.13810730e+00
5.99154383e-02 1.46076486e-01 -7.09755003e-01 1.29085422e-01
5.36129296e-01 6.44282043e-01 3.18878680e-01 -9.65832889e-01
-7.60762453e-01 -7.50261769e-02 -3.60931039e-01 -2.93910056e-01
1.32371211e+00 -3.22015405e-01 -3.49633098e-01 6.74874902e-01
1.09214461e+00 -8.45430791e-02 -1.21386439e-01 -1.97081372e-01
5.29220164e-01 3.55950035e-02 -2.23825097e-01 -9.19048607e-01
-9.24833775e-01 7.93995380e-01 6.52155340e-01 -1.44841984e-01
7.43731558e-01 -1.81232691e-02 1.30397773e+00 9.95741308e-01
5.20076334e-01 -1.18835676e+00 -7.64076233e-01 1.17732286e+00
7.03721166e-01 -1.17720151e+00 -5.88002563e-01 -1.29353136e-01
-8.96141768e-01 8.95036459e-01 7.44582117e-01 2.12556377e-01
5.77940702e-01 1.66048080e-01 6.64525211e-01 -1.00228436e-01
-7.32767582e-01 -2.44112760e-01 -4.51022908e-02 5.12681186e-01
8.23033810e-01 -5.30066565e-02 -6.67333305e-01 5.45609653e-01
-3.57839346e-01 -1.69566900e-01 5.24715364e-01 1.05481160e+00
-3.63891006e-01 -1.07352281e+00 -1.56292751e-01 3.00078630e-01
-7.53638387e-01 -4.82639700e-01 -4.94715065e-01 8.94461811e-01
-8.98987278e-02 1.15511143e+00 1.88698024e-02 -2.32552513e-01
6.44397020e-01 3.48620147e-01 -3.14736247e-01 -4.66620445e-01
-9.91272330e-01 -2.71314293e-01 7.60467649e-01 -2.86209792e-01
-2.72410333e-01 -3.09565574e-01 -1.61695588e+00 -4.15087461e-01
-3.73995036e-01 1.12169385e+00 9.61584449e-01 7.87800610e-01
5.44334829e-01 -8.79434422e-02 3.75152797e-01 -3.17305326e-01
5.49391434e-02 -1.32307124e+00 -5.08986652e-01 -1.39660873e-02
-1.01194352e-01 -3.91687512e-01 -2.82862633e-02 -2.04671398e-01] | [9.847987174987793, 9.753747940063477] |
07632cb1-f3d9-4f0f-9162-3516f29bca77 | joint-patch-and-multi-label-learning-for | null | null | http://openaccess.thecvf.com/content_cvpr_2015/html/Zhao_Joint_Patch_and_2015_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2015/papers/Zhao_Joint_Patch_and_2015_CVPR_paper.pdf | Joint Patch and Multi-Label Learning for Facial Action Unit Detection | The face is one of the most powerful channel of non-verbal communication. The most commonly used taxonomy to describe facial behaviour is the Facial Action Coding System (FACS). FACS segments the visible effects of facial muscle activation into 30+ action units (AUs). AUs, which may occur alone and in thousands of combinations, can describe nearly all-possible facial expressions. Most existing methods for automatic AU detection treat the problem using one-vs-all classifiers and fail to exploit dependencies among AU and facial features. We introduce joint-patch and multi-label learning (JPML) to address these issues. JPML leverages group sparsity by selecting a sparse subset of facial patches while learning a multi-label classifier. In four of five comparisons on three diverse datasets, CK+, GFT, and BP4D, JPML produced the highest average F1 scores in comparison with state-of-the art. | ['Fernando de la Torre', 'Wen-Sheng Chu', 'Jeffrey F. Cohn', 'Kaili Zhao', 'Honggang Zhang'] | 2015-06-01 | null | null | null | cvpr-2015-6 | ['action-unit-detection', 'facial-action-unit-detection'] | ['computer-vision', 'computer-vision'] | [ 2.86253035e-01 -4.92496639e-02 -6.58012748e-01 -5.81260622e-01
-7.54853606e-01 -4.45041984e-01 5.58849156e-01 -1.94959685e-01
-1.25967488e-01 4.41045523e-01 3.50224465e-01 4.31930423e-01
1.73776746e-01 -1.73532695e-01 -3.25517595e-01 -7.76859224e-01
-2.71054357e-01 5.98290488e-02 -3.66097003e-01 2.06524972e-03
-4.69332263e-02 5.64457357e-01 -1.90800929e+00 8.74968350e-01
-7.54752615e-03 1.37171102e+00 -6.90203846e-01 3.35668355e-01
2.88676079e-02 1.02008045e+00 -5.34501433e-01 -5.84097803e-01
2.08932012e-01 -5.59367418e-01 -7.42724240e-01 3.87997419e-01
8.80972981e-01 -3.50665808e-01 -3.09754349e-02 6.90099359e-01
5.72941005e-01 -2.26674333e-01 9.01225984e-01 -1.63426208e+00
-2.52720386e-01 2.90880948e-02 -9.91522908e-01 -9.89247784e-02
9.10059571e-01 -2.14943677e-01 1.08657956e+00 -8.79550517e-01
8.15490305e-01 1.77009749e+00 6.26645029e-01 8.26142490e-01
-1.57708645e+00 -1.21693897e+00 -7.63570294e-02 4.02787626e-02
-1.65456021e+00 -8.59035134e-01 6.49489284e-01 -5.45321524e-01
1.06193078e+00 2.66683221e-01 5.83334625e-01 1.33187819e+00
1.22689649e-01 8.80514205e-01 1.52688801e+00 -3.12174618e-01
-1.60264686e-01 -1.71401143e-01 -2.19998583e-01 9.20592427e-01
-4.32598352e-01 9.37317163e-02 -1.05294323e+00 -8.70549917e-01
5.92972934e-01 -2.18806759e-01 1.60347655e-01 2.63310850e-01
-4.88681972e-01 9.32767808e-01 -1.77288994e-01 1.19209431e-01
-3.73351812e-01 2.83449590e-01 5.12217879e-01 4.22421336e-01
5.84981859e-01 7.54691437e-02 -2.92007118e-01 -3.32858324e-01
-8.66400182e-01 3.82187515e-01 5.80209553e-01 4.93228257e-01
9.29531991e-01 -4.93790321e-02 -3.61583009e-02 1.26230872e+00
2.19362482e-01 2.99712747e-01 3.92223358e-01 -1.31764948e+00
-3.46620679e-01 4.78954971e-01 -3.65091085e-01 -1.14207780e+00
-6.46306396e-01 3.40224475e-01 -3.52404147e-01 3.63215238e-01
2.33052924e-01 -2.54825711e-01 -8.35274816e-01 1.86562884e+00
3.99342567e-01 3.08441401e-01 -3.33241761e-01 5.55755317e-01
1.14095986e+00 4.11059409e-01 5.82515895e-01 -4.45551336e-01
1.23939204e+00 -5.66062748e-01 -7.15737164e-01 -1.58234641e-01
9.86923695e-01 -7.32002914e-01 4.68116939e-01 6.76807463e-01
-7.70985544e-01 -2.64667958e-01 -4.56608057e-01 1.25938833e-01
-1.55947469e-02 2.26870373e-01 8.81225348e-01 6.49486244e-01
-1.09266973e+00 3.84203076e-01 -3.72555614e-01 -3.19353491e-01
7.11700261e-01 6.74298644e-01 -1.11929154e+00 -8.05462822e-02
-8.93326819e-01 8.55321229e-01 -2.89540201e-01 -3.20167869e-01
-9.07743275e-01 -4.71303910e-01 -9.21058893e-01 -3.77284914e-01
1.65590674e-01 2.53935698e-02 1.23219013e+00 -1.57701182e+00
-1.40334308e+00 1.47455847e+00 -3.93932968e-01 9.91295502e-02
-1.12941898e-01 2.06927173e-02 -5.04581273e-01 4.86101896e-01
-1.05496302e-01 1.30073011e+00 1.38144028e+00 -1.05764043e+00
-2.93341517e-01 -3.17560941e-01 -8.68181214e-02 1.61668062e-01
-2.28164211e-01 9.10261333e-01 -6.67462125e-02 -4.92780477e-01
-1.25548825e-01 -1.06348324e+00 3.39558899e-01 4.45190340e-01
-8.62221990e-04 -7.20381558e-01 9.15716946e-01 -5.38674295e-01
1.14750874e+00 -2.47026634e+00 2.67963290e-01 1.18014500e-01
2.88522065e-01 1.69745740e-02 -4.43623513e-01 3.06129366e-01
-5.10835290e-01 1.50865838e-01 2.92892486e-01 -6.46632552e-01
-1.68932065e-01 4.84627157e-01 2.84083128e-01 7.25630581e-01
3.93079937e-01 7.06046164e-01 -4.80814070e-01 -8.07947755e-01
-1.08566582e-01 4.89207983e-01 -5.59779525e-01 5.51840700e-02
-5.34547828e-02 4.08962667e-01 -2.40081668e-01 1.17080557e+00
4.96421993e-01 -5.16054220e-02 2.61805236e-01 -5.48041379e-03
2.37015232e-01 -1.22436307e-01 -9.26224887e-01 1.47811413e+00
-2.04984292e-01 4.36788797e-01 4.18103009e-01 -6.53995097e-01
8.23498011e-01 7.42964387e-01 1.05066621e+00 -4.65104550e-01
3.82129669e-01 9.78336930e-02 -1.20207198e-01 -6.04316950e-01
-2.64858335e-01 -5.50702095e-01 -7.47159943e-02 5.42049646e-01
5.42895496e-01 2.26212174e-01 -6.87913969e-02 -7.28983656e-02
1.23316967e+00 7.72227645e-02 7.24941611e-01 5.47773615e-02
3.66768062e-01 -5.60531557e-01 7.14797616e-01 8.17699581e-02
-3.72271538e-01 4.91799593e-01 9.85497534e-01 -4.50978249e-01
-3.04883569e-01 -6.39777064e-01 -2.67847687e-01 1.72940946e+00
-5.36236703e-01 -7.96928048e-01 -6.36727512e-01 -7.28545308e-01
3.62934113e-01 2.21154690e-01 -9.51166928e-01 -1.60197690e-01
1.37423510e-02 -5.73409021e-01 8.97512615e-01 2.56777644e-01
4.39012013e-02 -1.24816990e+00 -2.80639142e-01 -1.76969618e-01
-1.30605027e-01 -1.12641025e+00 -2.87500173e-01 1.41350880e-01
-2.56059200e-01 -1.09674966e+00 -4.36661631e-01 -5.82952917e-01
6.09310746e-01 5.09219207e-02 8.21219385e-01 2.08620563e-01
-5.53514719e-01 5.22061825e-01 -4.77440655e-01 -3.62390846e-01
-3.81427854e-01 -5.44699848e-01 2.70760775e-01 5.82727075e-01
7.63556778e-01 -5.30811489e-01 -3.53236608e-02 2.97857314e-01
-5.92260659e-01 -1.32481381e-01 6.29559875e-01 6.95464194e-01
4.13859993e-01 -3.37031662e-01 5.18998682e-01 -8.94531250e-01
5.64468980e-01 -6.48137152e-01 1.06101118e-01 5.05763963e-02
-9.69440937e-02 -4.06861007e-01 4.17571068e-02 -6.52836859e-01
-6.94762111e-01 4.56798345e-01 -1.75943494e-01 -7.98335016e-01
-6.47571206e-01 3.21171373e-01 1.25990674e-01 -6.87161267e-01
6.09642625e-01 -5.84012270e-02 3.24467123e-01 -2.37205759e-01
1.74398258e-01 8.47048581e-01 8.73884335e-02 -5.15916348e-01
7.13231787e-02 3.76460493e-01 3.19513112e-01 -1.17229009e+00
-8.96646261e-01 -6.24420822e-01 -7.57530570e-01 -7.66176820e-01
1.05549884e+00 -1.10369694e+00 -9.63656723e-01 7.11869299e-01
-1.03815067e+00 -2.28900045e-01 2.38696158e-01 2.07464889e-01
-6.37952387e-01 1.99232563e-01 -7.43041098e-01 -9.51894581e-01
5.19358814e-02 -1.00470388e+00 1.61964214e+00 -1.39148191e-01
-8.71619046e-01 -5.05700469e-01 -3.30135189e-02 4.02948707e-01
3.58817093e-02 6.23184443e-01 7.28016794e-01 -4.84102815e-01
3.88595253e-01 -2.15085521e-01 -1.96482614e-01 3.24375391e-01
3.92528713e-01 9.11131948e-02 -1.34116948e+00 -1.99732721e-01
-1.85169175e-01 -1.19212186e+00 7.45420098e-01 2.86100298e-01
1.27797270e+00 -4.73328620e-01 -1.89879701e-01 3.69453430e-01
9.59212244e-01 2.22564209e-02 5.05239248e-01 -2.46112987e-01
5.73080301e-01 8.44308853e-01 4.57048118e-01 6.82482779e-01
6.36082664e-02 9.51028466e-01 4.15892631e-01 3.70175950e-02
-1.38020605e-01 1.05237320e-01 6.46391451e-01 1.63855597e-01
-3.15969616e-01 2.34523728e-01 -6.63431346e-01 1.71641901e-01
-1.37761593e+00 -1.12776172e+00 1.38800189e-01 1.68759823e+00
9.46826100e-01 -3.38852674e-01 6.89279795e-01 2.05575272e-01
3.20043862e-01 3.04944605e-01 -2.78350592e-01 -8.12609494e-01
-2.37649247e-01 5.20387352e-01 1.42400220e-01 2.76077628e-01
-1.31298316e+00 9.55710769e-01 7.48120928e+00 9.35517251e-01
-1.29533434e+00 2.58424073e-01 6.82152390e-01 -3.94712627e-01
1.61005691e-01 -3.34702581e-01 -8.57072234e-01 3.33364516e-01
1.16258693e+00 3.05125356e-01 3.39431405e-01 7.96134114e-01
4.97092530e-02 -2.52134621e-01 -1.03638744e+00 1.41333437e+00
6.13688946e-01 -7.69507229e-01 -8.42832997e-02 1.59033045e-01
6.71269536e-01 -2.33092889e-01 1.05042271e-01 4.69532907e-02
2.98548210e-02 -1.57739484e+00 5.88235497e-01 4.40891773e-01
1.33154202e+00 -6.43917859e-01 3.41575801e-01 3.31309512e-02
-1.02178478e+00 -2.82733887e-01 -2.73420638e-03 -2.39860162e-01
2.82420069e-02 1.85338371e-02 -5.36684811e-01 -8.99509527e-03
6.61263704e-01 1.01535201e+00 -5.20608306e-01 4.29029346e-01
-2.71056175e-01 7.22516775e-01 -4.77664411e-01 5.12074269e-02
2.15804890e-01 5.73933758e-02 2.00330481e-01 1.23542404e+00
1.04989059e-01 3.95435572e-01 1.76016867e-01 2.21534848e-01
1.83682442e-02 3.29863369e-01 -7.06357360e-01 -3.74592751e-01
1.19999722e-01 1.44780099e+00 -2.80431092e-01 3.66995782e-02
-1.03905904e+00 1.06731462e+00 4.88944352e-01 1.19982557e-02
-5.88371933e-01 3.32024664e-01 1.03316748e+00 1.02098837e-01
1.18261978e-01 -3.50010134e-02 1.87893167e-01 -6.70663118e-01
-3.52016896e-01 -1.31449795e+00 5.92608154e-01 -7.00051367e-01
-1.39974391e+00 2.76734710e-01 -9.57089942e-04 -9.60432470e-01
-2.33601928e-01 -7.42386043e-01 -1.60309464e-01 5.18518031e-01
-1.16477764e+00 -1.64390397e+00 -2.19544321e-01 7.91288614e-01
2.82009214e-01 -1.16865948e-01 1.53225374e+00 3.55748296e-01
-5.27508616e-01 8.02199006e-01 -5.31339467e-01 1.16679192e-01
9.97745931e-01 -6.71471417e-01 -4.36801374e-01 -5.10945776e-03
1.83740288e-01 1.91022515e-01 3.76307040e-01 -4.15473610e-01
-1.08830082e+00 -8.39165509e-01 9.62229609e-01 -2.07585618e-01
5.68092704e-01 -4.52375174e-01 -6.48812354e-01 8.60922277e-01
-2.58128494e-02 1.94818780e-01 1.44602919e+00 2.27477118e-01
-9.18296337e-01 1.09244965e-01 -1.26187384e+00 3.16587746e-01
1.05447781e+00 -7.60453761e-01 -1.24516651e-01 4.23993617e-01
-3.41867357e-02 -2.12765798e-01 -1.12801754e+00 2.82517761e-01
9.72984970e-01 -1.09790075e+00 5.80977380e-01 -7.10927963e-01
5.51287472e-01 2.43261531e-01 -4.35679555e-01 -1.13233638e+00
-3.48834664e-01 -7.69282460e-01 -1.33243844e-01 1.06661236e+00
2.42023990e-01 -4.60891217e-01 7.95071721e-01 5.18661022e-01
2.04387575e-01 -9.74782884e-01 -1.18649209e+00 -4.36480731e-01
-1.31901249e-01 -3.10637087e-01 3.17008585e-01 1.11750925e+00
3.83725613e-01 2.18990788e-01 -8.01357329e-01 -4.29336548e-01
5.32125711e-01 -2.16012806e-01 6.48612142e-01 -1.42585683e+00
-1.91100314e-01 -5.26309192e-01 -1.00858736e+00 -4.68827575e-01
8.35805416e-01 -1.00875092e+00 -3.71605098e-01 -8.47633719e-01
4.44868058e-01 -1.56773835e-01 9.58454087e-02 1.35280001e+00
1.34847835e-01 6.88582063e-01 5.87546825e-02 1.73482701e-01
-5.62831640e-01 2.68293321e-01 1.05879223e+00 2.19791517e-01
1.66365609e-01 -2.89820015e-01 -6.45892680e-01 7.71082759e-01
5.97635269e-01 -5.19633591e-01 -7.25021586e-02 -2.73703784e-02
6.37510791e-02 1.10251181e-01 1.84692159e-01 -6.82467937e-01
-1.54127911e-01 -4.12750632e-01 3.81622732e-01 -3.20682116e-02
9.75073755e-01 -5.43007553e-01 1.83320090e-01 5.70362434e-02
-3.03705543e-01 -1.45768359e-01 3.14815402e-01 2.70632952e-01
-2.73860574e-01 8.70064124e-02 1.07762194e+00 -1.41460568e-01
-7.72005141e-01 3.43972713e-01 -7.11765766e-01 -3.27798158e-01
1.29274201e+00 -1.65539056e-01 -2.00932130e-01 -7.64234900e-01
-9.99696672e-01 -3.29718813e-02 2.36380890e-01 4.55790639e-01
5.77487171e-01 -1.57807124e+00 -7.38765538e-01 4.28896517e-01
3.73055577e-01 -7.06974089e-01 1.17133752e-01 1.04767680e+00
-7.71133751e-02 3.06750745e-01 -5.47087789e-01 -5.08533418e-01
-2.01040673e+00 1.83973312e-01 3.95880193e-01 2.64878660e-01
-7.91971460e-02 1.23305047e+00 1.48661017e-01 -1.83473125e-01
1.20680541e-01 4.67372775e-01 -3.54066759e-01 5.10026872e-01
6.90170884e-01 2.42489532e-01 -1.18202396e-01 -1.56433856e+00
-4.34523702e-01 7.65516341e-01 1.74522754e-02 4.11052927e-02
1.18647206e+00 2.02496320e-01 -4.24796879e-01 5.34765840e-01
1.60291922e+00 -7.49893785e-02 -8.86953592e-01 5.70769086e-02
-1.80433363e-01 -4.87188011e-01 -7.82458633e-02 -4.78079140e-01
-1.08877063e+00 6.45619452e-01 3.17393690e-01 -1.61394700e-01
1.30981874e+00 3.31905037e-01 4.18937504e-01 -1.02284588e-01
4.73262370e-01 -9.06495392e-01 3.24194133e-01 1.25210971e-01
9.78566229e-01 -1.19131088e+00 1.57904297e-01 -6.86606050e-01
-9.34797883e-01 1.07037330e+00 6.45778000e-01 3.70039828e-02
9.49581504e-01 3.88115913e-01 3.38741690e-01 -5.75036228e-01
-1.03036165e+00 -1.53362483e-01 2.83622921e-01 5.63034475e-01
7.35794723e-01 1.27178565e-01 -4.54441398e-01 4.27683204e-01
-2.14236323e-02 1.84666477e-02 8.03288147e-02 1.01929104e+00
-2.33225808e-01 -1.23558128e+00 -2.11150661e-01 8.63636613e-01
-8.60533834e-01 3.55033457e-01 -1.04299963e+00 5.75315237e-01
4.24619436e-01 9.93675649e-01 6.07292540e-02 -7.12767065e-01
-1.54652931e-02 5.73217392e-01 7.48584569e-01 -8.18695426e-01
-6.73273027e-01 3.84994417e-01 4.63875860e-01 -1.11036801e+00
-9.89108384e-01 -1.25753736e+00 -1.15541971e+00 -2.66570657e-01
-2.56746471e-01 -4.40884143e-01 2.82659918e-01 9.57885385e-01
2.61622369e-01 -1.45804763e-01 7.12407231e-01 -9.58401799e-01
-2.06157044e-01 -1.12512803e+00 -1.04236066e+00 7.67488241e-01
2.52902329e-01 -1.27929211e+00 -6.12623632e-01 -4.59797680e-02] | [13.547693252563477, 1.839769959449768] |
6ada8d7c-1e20-467c-9fcc-6b194f413d3f | crossfire-camera-relocalization-on-self | 2303.04869 | null | https://arxiv.org/abs/2303.04869v1 | https://arxiv.org/pdf/2303.04869v1.pdf | CROSSFIRE: Camera Relocalization On Self-Supervised Features from an Implicit Representation | Beyond novel view synthesis, Neural Radiance Fields are useful for applications that interact with the real world. In this paper, we use them as an implicit map of a given scene and propose a camera relocalization algorithm tailored for this representation. The proposed method enables to compute in real-time the precise position of a device using a single RGB camera, during its navigation. In contrast with previous work, we do not rely on pose regression or photometric alignment but rather use dense local features obtained through volumetric rendering which are specialized on the scene with a self-supervised objective. As a result, our algorithm is more accurate than competitors, able to operate in dynamic outdoor environments with changing lightning conditions and can be readily integrated in any volumetric neural renderer. | ['Arnaud de La Fortelle', 'Bogdan Stanciulescu', 'Dzmitry Tsishkou', 'Moussab Bennehar', 'Nathan Piasco', 'Arthur Moreau'] | 2023-03-08 | null | null | null | null | ['camera-relocalization'] | ['computer-vision'] | [ 0.50188833 -0.14466757 0.20843048 -0.42174032 -0.20004676 -0.83341914
0.6956419 0.12380278 -0.6176952 0.5150645 -0.42728972 -0.02642272
-0.12342906 -1.0724458 -0.7434122 -0.5731233 0.59831315 0.6511739
0.32128444 -0.3011477 0.24453445 0.9829918 -1.7531868 -0.07609244
0.5332249 1.1036916 0.6841884 0.70057446 0.12469608 0.74495625
-0.5815492 -0.17425184 0.3435204 -0.11964577 -0.30410793 0.18060274
0.7309151 -0.3339754 0.08693625 0.8308817 0.4213207 0.2633554
0.6391094 -0.6175282 0.02414247 -0.02277457 -0.39276648 -0.06446026
0.5550879 -0.23088053 0.8103649 -0.7475291 0.6954412 0.60720736
0.7026358 0.1610684 -1.335066 -0.16846661 0.42191446 0.10035445
-1.5369006 -0.20302151 0.9543282 -0.46845046 0.7712307 0.5226328
0.9726184 0.92986673 0.13117318 0.16003464 1.2234612 -0.4262117
0.5796801 0.29183748 -0.15394889 0.6393111 0.08220362 -0.15850107
-0.46690482 0.12850915 0.82330585 0.01755538 -0.6423815 -1.0325856
-1.0586058 0.5921318 0.7233366 0.28077692 -0.43249023 0.33395752
-0.0262718 0.11991015 0.54196835 0.3577685 -0.3255287 -0.14000763
-1.0325606 0.05868858 0.6681853 0.68927824 1.1014531 0.08127225
0.56071573 0.44953272 0.18411984 0.75868106 0.3084113 -1.0269349
0.28060782 0.49333262 0.08369294 -1.0865489 -0.5892095 -0.6002176
-0.6800389 0.48576808 0.06023052 0.19360079 -0.5964721 1.2484502
0.42135894 0.3531365 -0.02291138 0.86863434 0.42556834 0.29867157
-0.56147164 -0.05204979 1.0564566 -0.7247981 -0.35160086 -0.2707064
0.32323068 -0.6564647 0.909037 0.9422042 -0.6179468 -0.5160856
-1.2103157 -0.17498988 -0.61340684 0.23770465 0.512369 0.8886346
-1.349089 0.4737449 -0.805221 -0.4237224 0.00827881 0.42160898
-0.485363 0.1614971 -0.60971665 1.0376472 0.19918641 0.44847617
-0.68568057 -0.366876 -0.69135696 -0.17998973 0.30612314 -0.8687933
0.9325893 -1.2227567 -1.9625498 0.9328386 -0.24434036 -0.3834443
0.6479119 -0.58871996 -0.01198602 0.12515856 -0.25304288 0.2672143
0.84935534 -1.660256 -0.43100992 -0.46033183 0.5173997 0.4937974
-0.1477985 -0.5390681 -0.5640313 -0.2621452 0.44705546 -0.9341223
-0.4035789 0.1603036 -0.20870697 0.42859823 0.7834546 -0.4278196
0.6124024 -1.7123033 0.5738456 0.45898774 0.2391799 0.1614591
0.30877241 0.10304366 0.0856742 -0.3773594 -0.11742958 -0.7036761
-0.24455208 0.11888268 -0.30843103 0.7612192 -0.30757827 0.3871183
-0.795685 -0.02678038 1.0025177 0.9835418 -0.37491012 0.13653478
-0.14889246 0.8886202 -0.47952017 0.33390483 0.8463025 0.10548863
0.05965219 -0.2825585 -0.34929857 -0.0253724 -1.3184112 2.1929784
-1.2381364 0.8261341 0.04845305 -0.836814 1.2167424 -0.09854112
0.41515562 -0.85351664 0.3460355 0.05860871 -0.9132306 -0.08542565
0.82390857 0.17667624 0.1454601 0.33651495 -0.15187198 -0.43073314
-0.2870356 -0.15413639 0.8966406 0.69041485 0.48354036 0.01783626
0.7384343 -0.18932411 0.18005659 0.59548277 0.50898933 0.9356085
0.03897373 -0.4634781 -0.88898784 -1.238839 -0.2886873 0.7531421
0.52206564 -0.3569711 -0.8269624 -0.38805434 -0.3759585 0.6071814
-0.5624573 0.20858452 -0.5511518 -0.32832778 0.05137792 0.30541995
0.41757685 -0.66271424 -1.3822955 0.11604121 -0.11165135 -1.2862649
0.13966347 0.39269122 -0.93366456 -0.98340416 -0.56061554 -0.28257883
0.4977335 0.37409022 1.1941624 -0.28971323 -0.3302763 0.8833251
-0.3754571 -0.12561326 0.02350816 0.1672426 -0.08759317 0.23384742
-0.11587431 -0.93525004 -0.54667914 0.35855913 -0.58986056 0.14411509
0.14672509 0.36300686 0.9369776 -0.15734343 -0.29379374 -1.0588882
0.10829289 -0.15582615 -1.2979698 0.3970456 -0.5234137 -0.01450618
0.70103484 -0.10833687 -1.1784234 0.75405014 0.07191963 -0.5901761
-0.34069702 0.18442981 -0.33010915 -0.42264417 0.7695096 0.07657357
-0.5333466 -0.5587383 0.595937 0.42455408 0.6647646 -0.40159598
1.0065826 0.9308385 0.43685734 -1.0128974 -0.6034232 -0.5197039
-1.3121067 -0.6792033 0.82488996 -1.063585 -0.9301811 0.20590124
-1.2576016 -0.33517948 -0.22279307 0.4950765 -0.7512712 0.3836017
0.09139864 -0.9718858 -0.11542229 -1.0457139 1.461024 0.21224485
0.10395729 -1.110072 0.33945045 0.32391348 0.3224405 0.549302
0.02013087 0.28493285 -0.9487393 -0.14610262 -0.07523915 0.4051509
0.14954974 -0.01300247 -1.6124443 -0.15506427 0.27557907 -0.00980449
0.75467944 0.14826357 1.1674912 0.2459237 -0.20658107 1.2058213
1.8767251 -0.080075 0.57206446 0.5763102 1.0965133 0.70013255
0.62492186 0.5634529 0.43533117 1.1341598 1.0386782 -0.24230342
0.1529411 -0.03821695 0.13657375 0.5568993 -0.5840703 -0.2885447
-0.80846417 0.08138998 -1.9218249 -0.7025496 -0.38283715 2.5387182
0.05026719 -0.01123959 -0.3617749 0.09560856 0.2998006 0.27401653
-0.4448962 -0.37554672 -0.22197141 0.35213506 0.9853483 0.75754535
-0.9701948 0.90047294 5.7380757 0.20424637 -1.3916339 0.26914668
0.18943813 -0.09250801 -0.30327857 0.05519058 -0.539179 -0.02690038
0.77783334 0.3139031 0.70978945 0.96727794 0.12098987 -0.36382696
-1.0851582 1.1985567 0.45709234 -1.0734359 -0.2687292 0.15491828
0.54945475 0.27927095 0.23722292 -0.34525293 -0.05609053 -0.86298704
0.9836544 0.9111088 0.5572705 -0.79426324 0.5258818 0.43714306
-1.3167105 0.15263881 -0.44718042 -0.04440901 0.23932236 0.65419996
-0.8854157 0.8061506 0.70759237 0.6265072 -0.85763395 0.8755778
-0.3483717 -0.0471215 -0.49214524 0.20139621 -0.09219261 -0.5396654
0.30967015 0.8351259 0.48144487 -0.40542352 -0.05975434 0.645481
0.2525966 0.24008036 -0.87281984 0.57880557 -0.13813043 1.4300574
-0.968692 -0.02218973 -0.15859185 1.4372624 0.37770897 0.3018742
-0.9250216 -0.14722128 0.37315115 0.10115451 0.40004495 -0.43056914
-0.21244085 -1.3721241 0.36980504 -0.29175416 -0.2963157 -1.2578442
-0.62439615 1.1019542 -0.09230897 -1.4997844 -0.3645334 -0.8186499
-0.21678802 0.90821606 -1.6100569 -1.3760488 -0.8804537 0.790395
0.26960525 0.03240764 1.0813706 0.22926518 -0.09576391 0.04550051
0.39124817 -0.42292166 0.494682 -1.3974475 0.19697508 0.8200297
0.6975863 0.4606422 0.80276793 -0.3662586 -1.2267379 -0.9086387
0.26775253 -0.7528685 0.37625632 -0.8059678 -0.5268699 0.6259354
-0.03225332 0.15165748 0.36068115 0.2939505 -0.337853 -0.54196095
-0.853401 0.30544606 1.0737458 -0.8344603 -0.10472716 0.4452771
0.45509726 -0.83111614 -0.70809084 0.21071659 0.56181407 -1.3833354
1.1529272 0.21516675 -0.06814928 -0.59148103 -0.31311196 -1.1287531
0.13531922 -0.25137588 0.0501783 1.0283158 0.12837446 -0.62348896
0.8789077 0.15564486 -0.05442036 -0.3606 -1.1504198 -0.46233863
-0.5837171 -0.6908972 0.41147986 0.64352113 -0.7150964 0.25093284
-0.5012618 0.6127875 0.6820537 0.28936568 0.9320025 -1.5567797
-0.6518578 -0.30660152 -0.77768725 -0.9620399 0.29828724 -0.53439415
0.02508419 -1.350991 -0.25117177 -0.5596953 -0.11535058 0.07772255
0.38724166 0.78504586 0.07296035 0.18035205 -0.5794933 0.56464714
0.6814764 0.05137414 -0.3126802 0.24826564 -0.11185683 0.933701
0.7712182 -0.26990587 -0.5420378 -0.66563106 0.7045188 -0.19284028
0.5865191 -1.387483 0.29336366 0.12894371 0.42520696 -0.71639323
0.95241946 -1.4126992 0.5821483 0.127998 0.1530341 0.23144099
0.01484975 0.6379484 -0.08960039 -0.47654307 0.439735 -0.04236266
-0.84841657 -0.09605947 -0.29829854 -0.6302756 1.0937867 -0.582937
-0.07733352 -0.420141 -0.75428396 -0.3487489 1.0606431 0.29767197
0.53299606 -0.91686755 -0.06582808 0.19486824 0.28392023 0.3248869
0.37340647 0.55204886 -1.2023011 0.31526446 -0.19202319 -0.9876676
-1.3809733 0.6119086 0.62359846 0.0164848 -0.69756967 0.45314077
0.3436596 -0.51432496 0.15113917 -0.34254742 -0.35744798 0.13274576
0.32820553 0.12230568 0.42894918 -0.9629255 -0.34912282 1.1830497
0.4906955 -0.34874 1.3042308 -0.36911213 -0.11523486 0.9276183
1.0197221 0.37062448 -1.4183004 -0.18875629 -0.2804247 -0.58151644
0.5378699 -0.5120946 -1.1881307 0.9782967 1.0479695 -0.07339107
1.421839 -0.27793086 -0.03822659 0.6634329 0.93470114 -0.98371184
-0.22606921 0.29392302 0.7236661 -0.88196796 0.31950146 -0.59514356
-0.41889575 1.3820447 0.36233947 -0.0935541 0.48044872 0.2906162
0.1853087 -0.05384741 -0.22052917 -0.26666203 0.18065245 0.86337507
0.19364363 -0.04499083 0.32498646 -0.12586549 -0.31642666 -0.26067653
0.5593246 0.919396 -0.23651883 -1.1522827 -0.48057738 -0.04862868
0.11771972 0.09894529 -0.39730698 0.657453 0.41298944 0.7021634
0.16481766 -0.4623593 0.577313 -0.3179147 0.82766277 -0.5100891
-0.6647251 0.03855624 -0.20604205 -1.0182332 -0.6922294 -0.6882513
-0.7716208 -0.07625514 -0.3909132 -0.31099096 1.2683388 0.798975
0.05740757 0.6426258 0.8391502 -1.4025785 0.22293767 -0.38174516
-0.65997505 -0.19091628 0.42345008 -0.7599752 -0.22969839 0.10819843] | [9.164239883422852, -2.859923839569092] |
0be31ecc-4cb4-4793-9854-e5fa99147ab5 | acoustic-prosodic-and-lexical-cues-to | null | null | https://aclanthology.org/2020.tacl-1.14 | https://aclanthology.org/2020.tacl-1.14.pdf | Acoustic-Prosodic and Lexical Cues to Deception and Trust: Deciphering How People Detect Lies | Humans rarely perform better than chance at lie detection. To better understand human perception of deception, we created a game framework, LieCatcher, to collect ratings of perceived deception using a large corpus of deceptive and truthful interviews. We analyzed the acoustic-prosodic and linguistic characteristics of language trusted and mistrusted by raters and compared these to characteristics of actual truthful and deceptive language to understand how perception aligns with reality. With this data we built classifiers to automatically distinguish trusted from mistrusted speech, achieving an F1 of 66.1{\%}. We next evaluated whether the strategies raters said they used to discriminate between truthful and deceptive responses were in fact useful. Our results show that, although several prosodic and lexical features were consistently perceived as trustworthy, they were not reliable cues. Also, the strategies that judges reported using in deception detection were not helpful for the task. Our work sheds light on the nature of trusted language and provides insight into the challenging problem of human deception detection. | ['Julia Hirschberg', 'M', 'Xi (Leslie) Chen', 'Michelle Levine', 'Sarah Ita Levitan', 'Marko ic'] | 2020-01-01 | null | null | null | tacl-2020-1 | ['deception-detection'] | ['miscellaneous'] | [-2.48659968e-01 2.27625147e-01 8.40185285e-02 -8.02134633e-01
-9.72343445e-01 -1.02096987e+00 6.64145708e-01 1.30127937e-01
-5.23027718e-01 6.99656069e-01 3.88061941e-01 -3.94005120e-01
5.12164116e-01 -6.64888173e-02 -1.15678780e-01 -2.99777478e-01
3.28072667e-01 2.39494875e-01 -2.01545641e-01 -3.70497286e-01
6.97039545e-01 2.78867692e-01 -1.16826022e+00 5.26915252e-01
7.98017681e-01 9.83131111e-01 -5.87308705e-01 7.07004607e-01
3.95533949e-01 1.38366580e+00 -1.43858171e+00 -1.27418792e+00
1.19732276e-01 -6.05825126e-01 -7.42021263e-01 -5.25523163e-02
5.96571445e-01 -8.98634911e-01 -6.95885643e-02 1.19746327e+00
4.11464542e-01 -6.84680045e-02 5.69276690e-01 -1.14012420e+00
-1.00320458e+00 5.89151144e-01 -1.74678266e-01 6.28699839e-01
9.06906962e-01 1.95466533e-01 8.12850535e-01 -7.32224524e-01
3.85246277e-01 1.19110751e+00 5.82377434e-01 1.02591968e+00
-9.46723700e-01 -9.79982197e-01 -4.61057276e-01 2.77000517e-01
-1.27725279e+00 -1.35687828e+00 8.66278470e-01 -6.34810746e-01
6.40703142e-01 4.97050583e-01 4.83997405e-01 1.82789958e+00
3.55682433e-01 5.45902371e-01 1.42127013e+00 -1.84347913e-01
2.77673393e-01 7.94490159e-01 3.55297983e-01 5.22087157e-01
1.52553096e-01 4.29708451e-01 -9.86388564e-01 -1.01705110e+00
1.27466947e-01 -8.03062320e-01 -5.92165828e-01 1.79453656e-01
-7.62789011e-01 1.22076702e+00 -1.01672681e-02 3.67135376e-01
-2.46748134e-01 -1.95068002e-01 8.15098882e-01 3.89524430e-01
6.82816923e-01 8.68344426e-01 4.63695452e-02 -8.98428261e-01
-7.53129900e-01 1.87389538e-01 1.32388246e+00 4.25818026e-01
-9.33402777e-02 3.22099298e-01 2.39942104e-01 1.01000369e+00
-8.41918811e-02 5.07195592e-01 5.68030119e-01 -1.14101958e+00
2.30602503e-01 -5.06604277e-03 7.05740452e-01 -1.76042640e+00
-1.21541835e-01 -1.34564325e-01 -2.65669942e-01 2.63764948e-01
6.44897938e-01 -1.09296918e-01 -3.24440509e-01 1.34376371e+00
-2.25402594e-01 -5.89271665e-01 -1.48343444e-02 1.52823269e+00
5.68381667e-01 3.86148691e-01 1.74357399e-01 -3.03627670e-01
1.46066773e+00 -2.35427648e-01 -7.62253821e-01 -6.36210620e-01
5.44546962e-01 -7.69530892e-01 1.06250131e+00 6.80575728e-01
-1.07856965e+00 -2.48844385e-01 -1.28909159e+00 1.97554957e-02
6.55054748e-02 -2.18373358e-01 6.29757285e-01 1.36245608e+00
-8.60670030e-01 4.94655818e-01 -3.18007320e-01 3.97774391e-02
1.95428967e-01 -2.14017764e-01 -3.93106222e-01 3.00861806e-01
-1.45119822e+00 1.47704387e+00 8.70373473e-02 1.01119645e-01
-9.34426785e-01 -9.06871185e-02 -1.08960259e+00 -2.68210888e-01
3.39640290e-01 -6.79545403e-02 1.70237231e+00 -1.46847939e+00
-1.48550212e+00 1.47261119e+00 -1.18214250e-01 -4.60573435e-01
5.53087473e-01 1.89544652e-02 -8.02617371e-01 3.26666236e-01
1.42103344e-01 7.08834529e-02 1.21237302e+00 -1.41674113e+00
-9.47052017e-02 -5.82045674e-01 -1.14867292e-01 2.43209049e-01
1.40221283e-01 6.47548556e-01 9.67590928e-01 -3.98387581e-01
-6.82479367e-02 -4.01334435e-01 6.07643545e-01 -2.82061696e-01
-3.69919717e-01 -5.68673648e-02 5.06938696e-01 -1.07962739e+00
9.42291915e-01 -2.26929688e+00 -5.39939821e-01 1.51073873e-01
9.88604128e-01 4.82308716e-01 3.94944757e-01 4.51695651e-01
1.78132981e-01 7.12303758e-01 2.38121614e-01 -3.26798052e-01
2.42585540e-01 -1.76678635e-02 -3.50688279e-01 9.45167184e-01
-4.46904302e-01 8.99447918e-01 -9.29979146e-01 -4.25562382e-01
-8.44043270e-02 1.47769809e-01 4.58351634e-02 2.04877824e-01
3.96570921e-01 1.14431411e-01 -1.02396116e-01 6.78056061e-01
5.47186017e-01 4.41984504e-01 -8.49514008e-02 2.72820145e-01
1.02605566e-01 8.17908227e-01 -5.66624552e-02 7.85008311e-01
-1.54865101e-01 1.10083425e+00 8.56179118e-01 -4.41312492e-01
1.18991601e+00 3.72046709e-01 -4.39585596e-01 -5.05622983e-01
5.96678436e-01 4.89171296e-01 2.70675272e-01 -7.39042699e-01
7.82421887e-01 -7.37097979e-01 -6.27254009e-01 6.38928115e-01
-2.86131233e-01 -5.11390388e-01 -6.40655637e-01 2.53991395e-01
8.44995022e-01 -6.22607768e-01 6.07557178e-01 -2.42319018e-01
2.68349528e-01 1.07921332e-01 6.04866147e-01 8.24535429e-01
-1.14376295e+00 3.55553299e-01 7.56800950e-01 -5.30242085e-01
-6.42171741e-01 -9.26032424e-01 2.13093549e-01 1.03779304e+00
2.43281145e-02 -2.79548019e-01 -9.17460620e-01 -8.09980869e-01
-2.28856862e-01 1.48780906e+00 -4.43471938e-01 -6.79188192e-01
-5.89065515e-02 -2.36951753e-01 1.18878126e+00 9.18840393e-02
4.06115055e-01 -7.79618621e-01 -8.63629460e-01 -9.38115641e-02
-4.61460471e-01 -1.20528615e+00 -9.55230117e-01 -3.27836573e-01
-4.88447435e-02 -1.10167444e+00 1.37794414e-04 -3.88478130e-01
8.06016922e-02 3.58446300e-01 1.15839803e+00 2.86999345e-01
1.65637225e-01 2.00912073e-01 -5.16578972e-01 -4.51943874e-01
-1.14733243e+00 -6.84841812e-01 3.64788264e-01 -6.05610870e-02
8.45077872e-01 -3.53940666e-01 -2.35252380e-01 5.61296701e-01
-3.95521402e-01 -4.75732505e-01 5.31330630e-02 7.61577725e-01
-6.79947913e-01 -2.69779682e-01 5.45953512e-01 -7.08956420e-01
1.69719255e+00 -3.79824072e-01 -7.00175837e-02 -2.80586571e-01
-1.49106398e-01 -4.25603211e-01 6.97894692e-01 -3.73974264e-01
-8.74724984e-01 -6.53005779e-01 -1.14723563e-01 -1.92458242e-01
-3.52624863e-01 2.84832835e-01 1.36944428e-02 -3.26806009e-01
1.13464200e+00 1.33789390e-01 2.95324534e-01 6.98357373e-02
-7.84104392e-02 1.38539934e+00 6.65530324e-01 -5.01107752e-01
4.14596558e-01 1.08200103e-01 -1.02384031e+00 -8.58769238e-01
-8.23885798e-01 -3.96582186e-01 -1.69708177e-01 -2.97314048e-01
5.79201043e-01 -5.69403470e-01 -1.25231504e+00 6.81571901e-01
-1.53570354e+00 2.27597542e-02 2.89164782e-01 3.15873772e-01
-4.11420196e-01 8.74068379e-01 -1.05417407e+00 -1.55656803e+00
-4.17152971e-01 -8.80064189e-01 8.21841419e-01 -2.78245777e-01
-1.07068002e+00 -8.21124613e-01 -2.13710994e-01 1.16771710e+00
3.80015105e-01 1.76657990e-01 4.95495498e-01 -1.40049183e+00
4.61018533e-01 -6.02058649e-01 8.78520012e-02 5.84839880e-01
4.67796996e-02 -2.93778360e-01 -1.10068417e+00 3.04289497e-02
9.04695690e-01 -1.03228092e+00 2.65970528e-01 -1.28265977e-01
5.32808781e-01 -8.39305758e-01 1.26559064e-01 2.76230369e-02
6.28371835e-01 4.68618542e-01 5.31400502e-01 -1.89721048e-01
2.93025076e-01 8.26713622e-01 3.59576762e-01 3.38209480e-01
4.50339764e-01 7.57043958e-01 2.85679072e-01 6.60778999e-01
6.64105952e-01 -5.60783803e-01 7.81552613e-01 3.61614853e-01
2.82637626e-01 -2.86064118e-01 -9.20502484e-01 2.88827926e-01
-1.21634412e+00 -1.28825116e+00 -1.28815934e-01 2.05541015e+00
7.56278396e-01 4.75478977e-01 2.87434012e-01 1.93998829e-01
6.23394668e-01 3.68036807e-01 -1.40683755e-01 -1.38701987e+00
1.55557599e-02 -2.95711815e-01 2.31802478e-01 8.13685596e-01
-8.77389431e-01 9.01987314e-01 6.80908585e+00 7.04929292e-01
-8.67012620e-01 1.86552584e-01 7.36014724e-01 -1.15547003e-02
-1.79711610e-01 -4.69967455e-01 -1.71923444e-01 6.14684343e-01
1.24316621e+00 -2.39187166e-01 5.56591511e-01 8.17506194e-01
6.63306177e-01 -4.42438424e-01 -1.20816970e+00 1.11327899e+00
8.18949997e-01 -6.51897669e-01 -5.30287206e-01 -1.61745753e-02
-1.06603570e-01 -5.63983440e-01 -2.74066515e-02 1.72805458e-01
4.49502200e-01 -1.37514782e+00 1.22183692e+00 2.01350138e-01
4.10649568e-01 -5.60618222e-01 1.20258486e+00 9.30004895e-01
9.97868106e-02 1.41437888e-01 -2.91291207e-01 -6.44845128e-01
2.17755109e-01 2.79167503e-01 -1.35679328e+00 -1.18752949e-01
3.93633962e-01 2.23308638e-01 -4.41179305e-01 2.42070317e-01
-4.60065544e-01 7.13830888e-01 8.99936780e-02 -5.21325409e-01
8.05739090e-02 1.49231926e-01 8.29864383e-01 9.55540359e-01
-1.75459251e-01 3.41437280e-01 -3.57076600e-02 1.18897152e+00
-9.69629176e-03 -1.46938503e-01 -8.61493468e-01 -3.04664552e-01
7.60747969e-01 1.13398623e+00 -2.37209156e-01 -1.05865099e-01
1.33893117e-01 1.30902588e+00 2.66148359e-01 2.79720686e-02
-7.05588579e-01 -1.84227735e-01 7.47666121e-01 6.10041991e-02
-5.47911525e-01 -2.68189847e-01 -5.11819363e-01 -1.29885316e+00
2.40658239e-01 -1.37717772e+00 7.94401690e-02 -1.11422265e+00
-1.52476168e+00 7.21455991e-01 -1.14921428e-01 -6.65180624e-01
-6.40718162e-01 -5.54503620e-01 -3.99710000e-01 1.06806672e+00
-5.05312920e-01 -9.28297639e-01 -2.01000914e-01 9.79731232e-02
4.13862258e-01 -1.47385523e-01 9.48394358e-01 -3.87709677e-01
-1.96952730e-01 7.35796630e-01 -6.19764149e-01 5.04783034e-01
8.10790062e-01 -9.01716292e-01 3.83620292e-01 5.95234632e-01
6.41107261e-02 8.61083567e-01 1.13071156e+00 -5.61988711e-01
-8.13531339e-01 1.06696032e-01 1.02265382e+00 -8.30744863e-01
9.68197346e-01 -7.12744057e-01 -9.51718092e-01 4.68552321e-01
2.78666794e-01 -4.06521559e-01 8.51053059e-01 2.60509904e-02
-9.25867856e-01 4.67322022e-01 -1.69976187e+00 3.21795523e-01
9.23528433e-01 -1.15207469e+00 -1.05851948e+00 -5.75500876e-02
2.24675700e-01 -3.93665612e-01 -3.26807261e-01 -3.60219985e-01
6.77803814e-01 -1.51811051e+00 5.09850025e-01 -7.04533279e-01
3.92926157e-01 2.91075617e-01 -1.02139868e-01 -1.71372497e+00
-1.26277596e-01 -7.03265369e-01 2.90363610e-01 1.01215065e+00
2.47160882e-01 -8.88936460e-01 6.47130549e-01 1.14351487e+00
-1.04178250e-01 -2.89785892e-01 -1.23069990e+00 -5.03750265e-01
2.60942042e-01 -5.99449098e-01 7.32812062e-02 1.13015306e+00
9.06706989e-01 3.75088811e-01 -7.17795491e-01 -4.73394915e-02
4.32067513e-01 -4.95962113e-01 2.43347570e-01 -1.01906240e+00
-5.72949275e-02 -3.15989912e-01 -9.30737615e-01 -8.17646444e-01
7.52753794e-01 -5.02495825e-01 1.65552169e-01 -4.30631697e-01
1.16225988e-01 6.51531368e-02 5.96389532e-01 2.92570770e-01
-2.12815121e-01 2.14398935e-01 2.26318017e-01 3.69370878e-01
-3.04962039e-01 3.40412855e-01 7.52697766e-01 -1.30560383e-01
4.95386310e-02 1.95928752e-01 -1.32422733e+00 8.10852766e-01
9.81717288e-01 -2.71487683e-01 -1.96489617e-01 -2.99726903e-01
9.89865884e-02 4.16613519e-01 9.16748047e-01 -5.45523107e-01
2.10283771e-01 -9.26483497e-02 1.22654229e-01 2.23877668e-01
7.38845527e-01 -4.76560414e-01 -3.64012241e-01 7.46958181e-02
-4.57903773e-01 5.11692427e-02 9.79588330e-02 4.24496770e-01
-1.65807828e-01 -5.54744422e-01 8.62929881e-01 -3.88296008e-01
-4.46202993e-01 -6.43328547e-01 -9.97719944e-01 2.21816331e-01
8.47394526e-01 -5.05296826e-01 -5.75419664e-01 -1.43833411e+00
-5.82307756e-01 -7.32096583e-02 9.03854072e-01 3.51716071e-01
8.65831316e-01 -9.38521922e-01 -8.54449987e-01 8.13350230e-02
1.02208242e-01 -1.08903158e+00 -1.58144549e-01 6.76681101e-01
-4.86688286e-01 3.18110555e-01 -1.01507701e-01 -7.95365777e-03
-1.49502444e+00 3.66996795e-01 8.83176386e-01 4.98297721e-01
6.45171944e-03 9.03691053e-01 -2.80508280e-01 -2.28972882e-01
-5.50705083e-02 3.21038783e-01 8.25369954e-02 -1.19733848e-01
8.69668007e-01 4.43302661e-01 1.69656828e-01 -1.21393788e+00
-4.73678231e-01 -3.30215603e-01 -2.46212766e-01 -5.37391245e-01
5.53403020e-01 -2.73690522e-01 -1.97035670e-01 7.68318415e-01
1.14905190e+00 6.19747639e-01 -5.70257246e-01 3.99304807e-01
1.49540737e-01 -9.87314343e-01 -1.43516019e-01 -1.15864050e+00
-4.12571371e-01 5.26089787e-01 -2.83979028e-01 8.25176239e-01
4.62348521e-01 -9.16629434e-02 5.47586322e-01 4.49345149e-02
1.04998779e+00 -9.04094100e-01 2.83468273e-02 3.35127443e-01
1.13065529e+00 -1.31705976e+00 -1.97322100e-01 -6.37183666e-01
-1.24353147e+00 9.16602254e-01 5.65225661e-01 3.96154188e-02
1.60240188e-01 1.42807597e-02 6.31861150e-01 -4.90962982e-01
-5.84884405e-01 5.08216202e-01 1.64708510e-01 9.67611790e-01
6.11923635e-01 3.13520968e-01 -6.39442682e-01 1.03986752e+00
-1.05299044e+00 -3.20168972e-01 1.05068219e+00 5.23043692e-01
-4.58047241e-01 -5.67770123e-01 -6.16759002e-01 5.24701834e-01
-7.87345648e-01 -1.42931208e-01 -1.66458249e+00 6.09814763e-01
-2.21752197e-01 1.85881507e+00 -3.45897526e-01 -9.94316816e-01
2.25174069e-01 2.52862275e-01 1.59656405e-01 -6.92411900e-01
-8.94405603e-01 -5.12275517e-01 1.01242661e+00 -5.23878753e-01
6.98614866e-02 -7.53679931e-01 -4.57072854e-01 -9.37781572e-01
-4.27286863e-01 6.87211692e-01 3.27376783e-01 1.03339899e+00
6.87085241e-02 -4.90570426e-01 4.34725553e-01 -6.39305949e-01
-1.21244144e+00 -1.19248939e+00 -7.53536224e-01 4.48186040e-01
5.66451430e-01 -3.75186086e-01 -1.11327422e+00 -3.96024823e-01] | [8.195710182189941, 10.373002052307129] |
a6443cbe-b36e-4d5b-9d48-a3c64b39a315 | sire-separate-intra-and-inter-sentential | 2106.01709 | null | https://arxiv.org/abs/2106.01709v1 | https://arxiv.org/pdf/2106.01709v1.pdf | SIRE: Separate Intra- and Inter-sentential Reasoning for Document-level Relation Extraction | Document-level relation extraction has attracted much attention in recent years. It is usually formulated as a classification problem that predicts relations for all entity pairs in the document. However, previous works indiscriminately represent intra- and inter-sentential relations in the same way, confounding the different patterns for predicting them. Besides, they create a document graph and use paths between entities on the graph as clues for logical reasoning. However, not all entity pairs can be connected with a path and have the correct logical reasoning paths in their graph. Thus many cases of logical reasoning cannot be covered. This paper proposes an effective architecture, SIRE, to represent intra- and inter-sentential relations in different ways. We design a new and straightforward form of logical reasoning module that can cover more logical reasoning chains. Experiments on the public datasets show SIRE outperforms the previous state-of-the-art methods. Further analysis shows that our predictions are reliable and explainable. Our code is available at https://github.com/DreamInvoker/SIRE. | ['Baobao Chang', 'Yuting Wu', 'Shuang Zeng'] | 2021-06-03 | null | https://aclanthology.org/2021.findings-acl.47 | https://aclanthology.org/2021.findings-acl.47.pdf | findings-acl-2021-8 | ['document-level-relation-extraction'] | ['natural-language-processing'] | [-2.22884506e-01 8.00936699e-01 -6.46498621e-01 -5.64185023e-01
-1.70139465e-02 -5.32090247e-01 6.32689416e-01 4.38116103e-01
3.74236137e-01 8.05664182e-01 2.06275865e-01 -6.96143448e-01
-4.90271360e-01 -1.51656187e+00 -5.43287575e-01 1.79976776e-01
-1.52256116e-01 6.76245987e-01 6.27905965e-01 -3.39552134e-01
-7.24928603e-02 3.96908998e-01 -1.31682658e+00 8.74623001e-01
7.73619175e-01 7.72408724e-01 -6.01479888e-01 3.39012831e-01
-5.69073081e-01 1.60470164e+00 -5.55867493e-01 -1.05049813e+00
4.98406440e-02 -4.69498962e-01 -1.43273604e+00 -3.19027245e-01
3.10767382e-01 6.27314299e-02 -5.07143617e-01 8.45029593e-01
-2.74989098e-01 -4.65847433e-01 6.24711812e-01 -1.79870594e+00
-6.35024786e-01 1.26469326e+00 -5.27512848e-01 -1.31287336e-01
9.30878401e-01 -5.81450939e-01 1.66999304e+00 -7.01247931e-01
8.68728757e-01 1.26641679e+00 6.55543864e-01 3.36308807e-01
-9.34667706e-01 -6.94154382e-01 2.34627590e-01 7.45615661e-01
-1.45878792e+00 -1.25598684e-01 5.30789077e-01 -2.08027005e-01
1.35871470e+00 8.22423279e-01 6.00203097e-01 5.63937068e-01
3.88888836e-01 7.20840693e-01 8.50950897e-01 -4.55094635e-01
-1.67468265e-01 2.70363629e-01 6.53175831e-01 9.42694783e-01
8.25085282e-01 -5.00129461e-01 -6.90819740e-01 -2.10076928e-01
3.70063007e-01 1.91113167e-02 -1.95418492e-01 -1.74011797e-01
-1.05398929e+00 4.77159083e-01 5.38608849e-01 3.05023909e-01
-1.59477264e-01 -1.01624846e-01 1.58970401e-01 3.71723473e-01
1.77799523e-01 6.62046194e-01 -8.59511554e-01 1.86107069e-01
-5.22530913e-01 3.34072620e-01 1.52415276e+00 1.26729023e+00
4.89834726e-01 -7.71005869e-01 -1.24507979e-01 5.03802538e-01
4.24855739e-01 4.53087222e-03 -1.07189871e-01 -6.71925008e-01
6.85899734e-01 1.44680572e+00 -1.68829203e-01 -1.50121379e+00
-4.02742833e-01 -2.71495700e-01 -8.57316852e-01 -1.38449758e-01
2.71412879e-01 1.19735673e-02 -2.72910923e-01 1.21164596e+00
4.11890864e-01 3.98728922e-02 2.72707492e-01 5.25754511e-01
1.23576069e+00 4.61889863e-01 -2.03255881e-02 -1.21311001e-01
1.59348977e+00 -1.09926403e+00 -9.37604487e-01 -2.24023089e-01
7.90308952e-01 -6.25932574e-01 4.59471881e-01 4.40918654e-01
-8.94287109e-01 -1.42147139e-01 -1.04634273e+00 -1.49666861e-01
-8.02889943e-01 -9.48516503e-02 1.28401077e+00 3.79011989e-01
-7.14503169e-01 5.65793693e-01 -5.44371307e-01 -3.06731582e-01
4.23088610e-01 2.88343787e-01 -4.61754262e-01 -6.58653378e-02
-1.61524558e+00 1.05403566e+00 7.84107625e-01 3.23257106e-03
1.64562389e-01 -4.54675376e-01 -8.79077554e-01 3.16628397e-01
1.04394114e+00 -7.45098710e-01 1.00200129e+00 -1.76833391e-01
-8.00041020e-01 7.46334314e-01 -4.04970706e-01 -4.14724529e-01
3.48900944e-01 -1.89095572e-01 -8.88468742e-01 -1.35560840e-01
9.07573253e-02 2.23728806e-01 3.87103632e-02 -1.22695267e+00
-7.90961623e-01 -1.58796117e-01 5.05103946e-01 -1.29484877e-01
-1.87191844e-01 1.29079804e-01 -6.87725127e-01 -3.56516063e-01
5.06724358e-01 -5.72201252e-01 1.38645887e-01 -1.86411768e-01
-1.04049468e+00 -7.41098106e-01 6.89437628e-01 -4.42456573e-01
1.89928401e+00 -1.47909272e+00 -1.22520342e-01 4.66652423e-01
6.81212544e-01 5.59429564e-02 4.19860333e-01 7.70682991e-01
-2.93212503e-01 6.12051129e-01 -5.65236881e-02 2.23105565e-01
8.94831121e-02 4.56515878e-01 -3.97761434e-01 -2.41010264e-01
2.96851575e-01 1.04697645e+00 -8.63052845e-01 -8.58658850e-01
-2.39473596e-01 -6.96940627e-03 -3.05752993e-01 2.14101039e-02
-4.34411615e-01 -1.65480360e-01 -4.97037351e-01 7.25927711e-01
7.54496574e-01 -6.31246448e-01 9.04063880e-01 -2.64830083e-01
1.84627727e-01 8.26365650e-01 -1.29924631e+00 1.04569042e+00
-1.53811842e-01 5.31518281e-01 -4.97570515e-01 -8.06794047e-01
1.05288625e+00 2.21970707e-01 3.43360007e-01 -4.98718262e-01
-1.94120377e-01 1.54201612e-01 1.03851231e-02 -5.54798782e-01
4.47080404e-01 2.15269282e-01 -5.35993539e-02 3.99081856e-01
-1.42214760e-01 2.18527596e-02 6.63891554e-01 7.03853548e-01
1.41224825e+00 1.90198630e-01 8.02057862e-01 1.17996462e-01
7.53144443e-01 2.84679890e-01 8.35646152e-01 6.22622609e-01
4.19156462e-01 1.31082699e-01 1.18647552e+00 -5.85304201e-01
-4.99016583e-01 -7.70201504e-01 -1.00002736e-01 4.63682652e-01
3.60962063e-01 -1.55646229e+00 -2.91401565e-01 -1.14698088e+00
1.55493349e-01 7.76273370e-01 -3.51342320e-01 1.23527862e-01
-4.88449097e-01 -5.01395702e-01 5.59000731e-01 4.80885953e-01
4.43822563e-01 -9.27543342e-01 -8.16095993e-02 8.56014490e-02
-4.18274879e-01 -1.37428999e+00 4.81044114e-01 4.13160883e-02
-5.26582301e-01 -1.59016395e+00 4.94808823e-01 -5.16662538e-01
6.96855724e-01 7.09882006e-02 1.47285974e+00 6.15670443e-01
-6.77751973e-02 -1.63043290e-01 -5.30767739e-01 -2.36844167e-01
-3.68999898e-01 1.33854106e-01 -2.72817224e-01 -3.52344304e-01
9.17839527e-01 -5.18681288e-01 -1.57644257e-01 4.73027229e-01
-5.12276590e-01 5.35599947e-01 4.91156161e-01 4.13657963e-01
4.92947400e-01 7.79183626e-01 1.34381443e-01 -1.60481715e+00
6.82132125e-01 -7.13027298e-01 -2.91131943e-01 8.69345188e-01
-1.12148464e+00 3.77733797e-01 7.42623627e-01 1.04334012e-01
-1.02028501e+00 -3.61374497e-01 3.46062519e-02 2.38458753e-01
-1.38680562e-01 8.52281272e-01 -3.54611903e-01 4.24885631e-01
4.11594003e-01 -3.14806551e-01 -6.27843797e-01 -2.81563997e-01
4.33528095e-01 4.89164978e-01 2.93804675e-01 -7.00482011e-01
9.71598864e-01 2.28264779e-01 4.71784890e-01 -8.26583654e-02
-1.01730824e+00 -2.72947907e-01 -7.35477448e-01 -1.39584700e-02
3.91765982e-01 -6.67253792e-01 -1.04370809e+00 -1.29874758e-02
-1.50412011e+00 1.57189816e-02 -1.00184709e-01 1.62259579e-01
2.00120406e-03 1.89761564e-01 -7.00570166e-01 -7.13062286e-01
-2.30392903e-01 -6.64899766e-01 5.11704922e-01 1.69195339e-01
-8.25908601e-01 -8.29544306e-01 -1.06090121e-01 4.90682304e-01
-8.94097313e-02 2.61762589e-01 1.10433459e+00 -8.65061343e-01
-7.81432629e-01 -3.62275511e-01 -5.65346122e-01 -1.24281347e-01
3.60394150e-01 4.18323219e-01 -5.42665064e-01 6.02042556e-01
-6.27325654e-01 7.30521828e-02 6.29943848e-01 -4.04361755e-01
1.11462414e+00 -7.51141191e-01 -8.55890870e-01 2.25774229e-01
1.16733301e+00 1.61014780e-01 8.45176458e-01 3.72247249e-01
8.07573617e-01 8.56105864e-01 7.62661338e-01 9.35424864e-02
8.58240008e-01 6.62980676e-01 1.75947979e-01 3.34979966e-02
-1.28155321e-01 -5.97786427e-01 -5.92204072e-02 6.50318086e-01
-2.75917739e-01 -4.29317713e-01 -1.21921360e+00 2.63165057e-01
-2.23103333e+00 -1.25458312e+00 -9.82562840e-01 1.58588672e+00
1.15496016e+00 4.34985131e-01 -2.59155601e-01 4.96587008e-01
3.72252464e-01 -8.96318108e-02 -5.36476672e-02 -5.83247423e-01
-1.86076313e-01 1.66093975e-01 2.58446276e-01 5.38796425e-01
-8.62296104e-01 1.14219797e+00 5.74397135e+00 6.81622684e-01
-4.64145780e-01 -1.35640651e-01 4.81671244e-01 2.94856697e-01
-5.10705709e-01 4.62816060e-01 -1.10811579e+00 1.06316626e-01
7.17969894e-01 -2.57646143e-01 2.27962285e-02 8.29335928e-01
-2.29030460e-01 -1.77596480e-01 -1.36949325e+00 6.58475041e-01
-1.86639190e-01 -1.46671867e+00 1.13696039e-01 -8.70820209e-02
5.11457264e-01 -6.47611976e-01 -6.05280340e-01 3.22280943e-01
4.49751735e-01 -1.08193052e+00 5.27010560e-01 6.77400410e-01
4.99668241e-01 -5.90353549e-01 9.74953115e-01 2.38356486e-01
-1.51103306e+00 1.23176999e-01 -8.06505904e-02 -3.98402631e-01
-2.29900591e-02 9.51014757e-01 -9.04006302e-01 1.23050094e+00
6.58322811e-01 9.79864657e-01 -8.11777234e-01 6.22291625e-01
-1.07326472e+00 4.75690275e-01 -1.94905981e-01 -1.82867482e-01
-2.25554496e-01 -1.67110711e-01 1.64946273e-01 1.40747809e+00
7.79706910e-02 4.59054351e-01 -1.04762502e-01 8.94279420e-01
-2.39234701e-01 -4.71276827e-02 -5.58550596e-01 -1.22175619e-01
6.18597984e-01 1.26680529e+00 -8.00414324e-01 -5.53239703e-01
-6.86978877e-01 7.13529944e-01 4.86552685e-01 1.74110815e-01
-7.33004093e-01 -4.73663002e-01 5.31426132e-01 2.79264331e-01
-5.20336814e-02 2.12917831e-02 -6.08279049e-01 -1.34655917e+00
4.21322465e-01 -7.92197883e-01 7.38611281e-01 -8.97100091e-01
-1.24699938e+00 7.28070974e-01 1.41656265e-01 -1.11458409e+00
-2.04970106e-01 -5.54192066e-01 -4.52067167e-01 6.96982443e-01
-1.35695171e+00 -1.13298070e+00 -3.52150202e-01 4.97429699e-01
6.27185032e-02 1.28057852e-01 1.09579229e+00 2.51691461e-01
-6.46655262e-01 4.61868525e-01 -7.38729298e-01 4.54429388e-01
5.00511587e-01 -1.32450306e+00 4.67603385e-01 1.00100756e+00
4.16885734e-01 9.88935769e-01 6.29322827e-01 -8.89946878e-01
-1.25741160e+00 -8.80123496e-01 1.75782752e+00 -5.72828054e-01
1.02651811e+00 -3.19947839e-01 -1.01845002e+00 1.16721272e+00
2.51710027e-01 -3.89861129e-02 7.39760876e-01 7.04834282e-01
-7.80610263e-01 -2.11527169e-01 -8.87869477e-01 8.34613979e-01
1.48563230e+00 -3.21502179e-01 -8.45464826e-01 4.36031610e-01
7.21431673e-01 -3.02982122e-01 -9.63135898e-01 5.76520562e-01
5.75443447e-01 -1.05169499e+00 8.65430772e-01 -7.01103747e-01
9.32869732e-01 -5.27345061e-01 1.91444293e-01 -9.00755703e-01
-2.50186056e-01 -5.00885427e-01 -9.15942669e-01 1.48006332e+00
1.03346622e+00 -8.00187647e-01 6.67687833e-01 9.26863134e-01
3.15242380e-01 -1.24350393e+00 -2.86809564e-01 -7.79190719e-01
-4.62560952e-01 -5.08012593e-01 1.01805329e+00 1.14645028e+00
8.82812202e-01 7.47861505e-01 -1.49204850e-01 2.94312268e-01
4.69132543e-01 6.36074901e-01 7.79560864e-01 -1.53632867e+00
-2.40835637e-01 -4.78912413e-01 -5.24490714e-01 -9.20792103e-01
2.32531697e-01 -1.12703264e+00 -4.43688333e-01 -2.29070544e+00
2.60826409e-01 -6.45782113e-01 -9.26475152e-02 1.11620975e+00
-1.05930410e-01 -2.19045669e-01 -5.40851355e-02 2.33193472e-01
-6.77648365e-01 -2.07981206e-02 1.17410016e+00 -2.39750385e-01
1.64281130e-02 3.53652588e-03 -9.60418582e-01 7.53190875e-01
9.61095333e-01 -6.66153014e-01 -5.23356378e-01 -8.06745514e-02
9.81270850e-01 4.72370759e-02 1.59171999e-01 -7.10907280e-01
5.62439859e-01 -4.62381274e-01 1.75918594e-01 -6.44347906e-01
4.43981066e-02 -9.08485234e-01 5.11319697e-01 2.86473095e-01
-3.02009374e-01 -1.55763596e-01 -1.17705971e-01 1.84626698e-01
-5.18663347e-01 -9.24835354e-02 5.53867817e-02 2.03726627e-03
-7.19123960e-01 1.49422899e-01 1.24164827e-01 -1.51367456e-01
1.05169892e+00 -3.21475528e-02 -8.30868900e-01 -1.95762455e-01
-5.70486367e-01 4.58162755e-01 1.39539078e-01 4.14840341e-01
6.32061362e-01 -1.25559139e+00 -6.77342057e-01 -1.87832713e-01
3.33674431e-01 1.61953181e-01 -2.69660950e-01 8.72640550e-01
-6.08358860e-01 5.89081705e-01 7.10975677e-02 8.23682994e-02
-1.59754539e+00 5.53408802e-01 2.00260818e-01 -7.52832472e-01
-7.04321027e-01 7.14014411e-01 -3.17138255e-01 -4.01715279e-01
-4.20487076e-02 -4.73375112e-01 -5.40686965e-01 2.32334286e-02
5.20263612e-01 1.77571282e-01 7.68881440e-02 -3.26244563e-01
-7.29853272e-01 3.47861111e-01 -1.44540682e-01 3.04597676e-01
1.33663797e+00 8.16084817e-02 -8.58997345e-01 3.30521494e-01
7.82023132e-01 3.79946440e-01 -3.72885138e-01 -3.34940910e-01
4.94024098e-01 -4.51587349e-01 -4.20571297e-01 -1.00871730e+00
-9.69286561e-01 2.58231938e-01 -5.95285237e-01 9.09232616e-01
9.30151641e-01 4.16693509e-01 6.11207485e-01 4.52767223e-01
5.77271283e-01 -6.12643659e-01 -4.38263446e-01 5.35501957e-01
9.09103096e-01 -1.11552370e+00 4.80867684e-01 -1.45681751e+00
-6.95897102e-01 1.28998041e+00 7.45279729e-01 1.61387503e-01
5.73374748e-01 6.80753887e-01 -1.87794566e-01 -6.22900486e-01
-1.11957932e+00 -6.21621944e-02 4.58641082e-01 2.53970325e-01
8.15947711e-01 3.98265541e-01 -7.09298134e-01 7.79568493e-01
-5.87454379e-01 -1.04603721e-02 5.16809583e-01 9.06925499e-01
-4.44894321e-02 -1.65954053e+00 -1.97340459e-01 4.68169808e-01
-3.18409234e-01 -2.44263053e-01 -1.03568602e+00 8.83132517e-01
3.03571820e-01 1.25099695e+00 -1.92732185e-01 -6.89044893e-01
3.83967906e-01 1.63105831e-01 3.54289025e-01 -7.22279549e-01
-3.95190299e-01 -5.55351079e-01 7.69602299e-01 -6.96938992e-01
-3.89805853e-01 -3.06368440e-01 -1.75559509e+00 -7.42528856e-01
-2.63998300e-01 3.30667019e-01 1.22200422e-01 8.86309505e-01
3.89360607e-01 7.92618394e-01 2.25551084e-01 3.51856261e-01
2.42682457e-01 -6.77133381e-01 -5.15904605e-01 3.87401849e-01
-1.68434128e-01 -6.06722355e-01 -1.11867793e-01 -2.40536798e-02] | [9.168468475341797, 8.26472282409668] |
70adb4e8-f9a0-42d2-be96-71d1723a6347 | proceedings-37th-international-conference-on | 2109.07914 | null | https://arxiv.org/abs/2109.07914v1 | https://arxiv.org/pdf/2109.07914v1.pdf | Proceedings 37th International Conference on Logic Programming (Technical Communications) | ICLP is the premier international event for presenting research in logic programming. Contributions to ICLP 2021 were sought in all areas of logic programming, including but not limited to: Foundations: Semantics, Formalisms, Nonmonotonic reasoning, Knowledge representation. Languages issues: Concurrency, Objects, Coordination, Mobility, Higher order, Types, Modes, Assertions, Modules, Meta-programming, Logic-based domain-specific languages, Programming techniques. Programming support: Program analysis, Transformation, Validation, Verification, Debugging, Profiling, Testing, Execution visualization. Implementation: Compilation, Virtual machines, Memory management, Parallel and Distributed execution, Constraint handling rules, Tabling, Foreign interfaces, User interfaces. Related Paradigms and Synergies: Inductive and coinductive logic programming, Constraint logic programming, Answer set programming, Interaction with SAT, SMT and CSP solvers, Theorem proving, Argumentation, Probabilistic programming, Machine learning. Applications: Databases, Big data, Data integration and federation, Software engineering, Natural language processing, Web and semantic web, Agents, Artificial intelligence, Computational life sciences, Cyber-security, Robotics, Education. | ['Neng-Fa Zhou', 'Joost Vennekens', 'Gian Luca Pozzato', 'Paul Fodor', 'Carmine Dodaro', 'Veronica Dahl', 'Alex Brik', 'Bart Bogaerts', 'Yanhong Annie Liu', 'Andrea Formisano'] | 2021-09-15 | null | null | null | null | ['data-integration', 'automated-theorem-proving', 'automated-theorem-proving'] | ['knowledge-base', 'miscellaneous', 'reasoning'] | [-5.05414188e-01 3.78073305e-01 -2.09567592e-01 -4.69560742e-01
3.47748071e-01 -1.01903987e+00 5.22836030e-01 9.77803349e-01
2.08568469e-01 1.21925557e+00 -8.04623142e-02 -5.64950526e-01
-8.79243970e-01 -1.08922732e+00 -2.70348340e-01 -7.91401137e-03
-4.83214825e-01 1.27600098e+00 8.09365153e-01 -3.21222425e-01
4.34390068e-01 5.96425354e-01 -2.11289406e+00 4.40449566e-01
6.26649022e-01 1.05553401e+00 -6.95645571e-01 4.68789905e-01
-4.03317213e-01 1.45543563e+00 -9.05887261e-02 -2.55290627e-01
-1.37637883e-01 1.82720110e-01 -1.23091984e+00 -5.55092275e-01
-3.90554935e-01 2.89126754e-01 4.00340140e-01 1.43776882e+00
-1.16577797e-01 -1.21916644e-01 3.88916284e-01 -2.24980283e+00
-6.57164216e-01 9.84542906e-01 -3.90926093e-01 -9.81394425e-02
1.37019706e+00 -2.30607510e-01 7.16682374e-01 -3.93701047e-01
1.07293069e+00 1.75260448e+00 5.63398719e-01 3.29473555e-01
-9.86309230e-01 -4.43524837e-01 -8.69840607e-02 7.36677289e-01
-1.30784309e+00 2.99044430e-01 1.54868931e-01 -5.10975122e-01
1.42865622e+00 6.66938901e-01 5.02785921e-01 1.45600483e-01
2.07209617e-01 4.35034007e-01 1.23652029e+00 -8.23125482e-01
5.71244001e-01 1.27336907e+00 9.40451384e-01 8.29371989e-01
4.87936169e-01 1.11366585e-01 -5.77241004e-01 -6.17450774e-01
1.74454853e-01 -3.90928268e-01 2.28958800e-01 -4.23082352e-01
-9.65647578e-01 7.93667614e-01 -6.18569255e-01 4.89541858e-01
-3.31697881e-01 -2.04649761e-01 9.06925738e-01 6.69556916e-01
-5.78525841e-01 -1.43723916e-02 -1.11603975e+00 -1.54879928e-01
-2.72181749e-01 1.00428045e+00 1.67275572e+00 1.65093684e+00
6.51276112e-01 -2.82890737e-01 1.82978988e-01 1.72880441e-01
1.05425549e+00 4.22406524e-01 4.33497667e-01 -1.04825020e+00
1.22469924e-01 1.11023962e+00 2.52617329e-01 -8.82064998e-01
-6.22505248e-01 9.05433118e-01 -1.60037041e-01 2.22736105e-01
-2.66093999e-01 -1.54402569e-01 -4.41280335e-01 1.38680875e+00
9.14748967e-01 -4.28352803e-01 5.18361390e-01 5.45278907e-01
9.05579150e-01 4.99345720e-01 1.92744210e-01 -4.38798785e-01
1.67437494e+00 -4.05219644e-01 -7.68267274e-01 4.87823874e-01
2.54049689e-01 -6.54051423e-01 1.05721429e-01 1.03828740e+00
-1.41170919e+00 2.04426542e-01 -6.49823368e-01 2.82947779e-01
-1.40464807e+00 -4.59597468e-01 1.09727275e+00 6.53255641e-01
-9.05118406e-01 -6.58924282e-02 -6.72453761e-01 -5.51707268e-01
2.56619364e-01 6.48957551e-01 -4.50183809e-01 9.19678994e-03
-1.34630084e+00 1.36410415e+00 1.57648015e+00 -5.24862468e-01
-5.38931966e-01 -7.00347543e-01 -9.30940926e-01 -3.26397829e-02
4.75245476e-01 -1.28297761e-01 8.45640898e-01 -6.32081091e-01
-1.50292504e+00 9.27331805e-01 4.84481007e-02 -9.13353741e-01
2.24717587e-01 6.92274153e-01 -8.76178801e-01 -2.30495527e-01
2.34995455e-01 2.24929273e-01 -4.38932538e-01 -8.96347165e-01
-1.05691957e+00 -6.01081073e-01 3.46495271e-01 -4.50148582e-01
7.03188181e-01 7.04467773e-01 4.14497405e-01 3.76589656e-01
-3.77080552e-02 -5.23299277e-01 -1.16160713e-01 -4.91898268e-01
-3.01997989e-01 -7.16990650e-01 1.08618748e+00 -1.01140901e-01
8.13364208e-01 -1.85731268e+00 2.18240649e-01 1.04366481e+00
-2.38233730e-01 -1.43434569e-01 9.32406664e-01 1.01731384e+00
9.78990793e-02 3.05620968e-01 -1.00501314e-01 1.19619489e+00
8.63264263e-01 2.37669423e-01 -8.36878046e-02 9.95513350e-02
-1.07132241e-01 3.24107945e-01 -7.30999708e-01 -1.03387904e+00
7.15568662e-01 -1.93330020e-01 -6.28351748e-01 -3.11471701e-01
-1.15057898e+00 -2.83398837e-01 -6.43418908e-01 1.43812048e+00
7.79619575e-01 6.22672541e-03 7.80595124e-01 1.35146245e-01
-6.92557156e-01 -3.20581973e-01 -1.93251073e+00 1.39813316e+00
-5.81167221e-01 5.71784198e-01 1.62429839e-01 -1.00067437e+00
7.98275054e-01 8.74738455e-01 3.26986879e-01 -4.00828809e-01
6.87809777e-04 5.82034588e-01 -2.21268490e-01 -1.14449227e+00
6.74976632e-02 3.52687538e-01 -9.42173451e-02 5.27680218e-01
3.18351448e-01 -2.38908470e-01 9.23465312e-01 1.09041780e-01
6.95380270e-01 4.43143696e-01 8.36525798e-01 -4.63162541e-01
1.05722439e+00 6.36592805e-01 5.35057008e-01 2.11560950e-01
-6.71206340e-02 -4.97154891e-01 9.57293868e-01 -7.45911062e-01
-6.96885705e-01 -1.05272710e+00 -4.52451289e-01 1.03097570e+00
2.32563898e-01 -5.05967915e-01 -4.13508266e-01 -2.35034898e-01
2.81631529e-01 8.87085080e-01 -1.64885558e-02 -1.34004289e-02
-2.75234491e-01 -4.42249835e-01 6.04358613e-01 4.57201488e-02
3.75725865e-01 -1.37534690e+00 -1.21730626e+00 4.51836973e-01
4.58599806e-01 -9.11814034e-01 9.42913830e-01 1.68552101e-01
-5.58864236e-01 -1.52191031e+00 8.55380774e-01 -6.44499660e-01
3.19567531e-01 -8.59154403e-01 1.27833855e+00 1.30074933e-01
-5.89039743e-01 2.99444914e-01 -1.46714762e-01 -1.13161445e+00
-4.00149763e-01 -9.03306901e-01 5.96515089e-02 -8.48621607e-01
8.88055623e-01 -5.74244320e-01 3.20750356e-01 1.95005015e-01
-8.74031186e-01 -3.14624637e-01 1.67073593e-01 7.52775908e-01
4.19616282e-01 6.19250774e-01 3.45017999e-01 -1.24481106e+00
9.46216464e-01 -6.05319738e-01 -1.33055365e+00 8.55062187e-01
-8.18468571e-01 -2.30753087e-02 6.23968720e-01 -1.91188782e-01
-9.65104461e-01 -3.77613842e-01 4.03639853e-01 9.36393254e-03
-5.72935045e-01 7.77504086e-01 -6.71252191e-01 6.29615709e-02
6.77945256e-01 1.53192088e-01 1.69489495e-02 1.33196890e-01
4.65531707e-01 5.77678323e-01 2.29593784e-01 -1.45825005e+00
4.04774070e-01 3.11889887e-01 2.48404458e-01 -3.56312335e-01
-1.51330233e-01 -2.31043085e-01 -3.13796729e-01 5.91385663e-02
6.73890531e-01 -4.04910445e-01 -1.63039911e+00 -1.81876674e-01
-1.37051225e+00 2.64249265e-01 -4.12514955e-01 4.88950342e-01
-5.84214151e-01 -3.15061480e-01 -2.95433223e-01 -1.18426597e+00
-4.10017401e-01 -8.85270178e-01 3.22211742e-01 5.39291084e-01
-2.59205550e-01 -1.19472921e+00 1.68758720e-01 1.17883749e-01
2.83399045e-01 8.21462631e-01 1.37005639e+00 -1.33020937e+00
-6.33250833e-01 -3.25710565e-01 -8.64299536e-02 9.64484811e-02
-2.79420227e-01 8.84232640e-01 -4.14488822e-01 5.74259281e-01
-6.97571874e-01 -6.80782139e-01 -4.58396196e-01 -2.31397916e-02
8.03706467e-01 -8.55336666e-01 -4.35700476e-01 -2.07945615e-01
1.69361639e+00 6.11123860e-01 6.10112011e-01 5.05630672e-01
-2.79575169e-01 7.68891335e-01 8.74310017e-01 6.45754278e-01
6.47840083e-01 3.01013857e-01 3.98508608e-01 1.10284507e+00
5.28606236e-01 2.81872123e-01 -8.11340585e-02 1.01728879e-01
-5.27561665e-01 5.17904222e-01 -1.51440763e+00 2.45341077e-01
-2.36269236e+00 -1.36503780e+00 -4.82504010e-01 2.28947115e+00
9.75720346e-01 3.96601409e-01 1.39764339e-01 5.95932961e-01
5.78562915e-01 -5.81379652e-01 9.06689018e-02 -1.06534386e+00
2.79836189e-02 6.32182732e-02 3.16280425e-01 7.96643078e-01
-7.81308413e-01 1.05265582e+00 5.13044739e+00 6.55760467e-01
-9.59722340e-01 5.21589160e-01 -2.75110960e-01 5.51117837e-01
-5.99651396e-01 3.31412762e-01 -8.95774543e-01 3.95019859e-01
1.15934467e+00 -8.39224100e-01 1.04424918e+00 1.20668828e+00
-8.67627934e-02 -3.39221686e-01 -1.23227835e+00 4.32137042e-01
-3.53486426e-02 -1.79945493e+00 -1.46126509e-01 -3.02800417e-01
8.15180361e-01 -3.98069099e-02 -7.66341269e-01 4.74623322e-01
6.62453771e-01 -8.57977509e-01 1.20564222e+00 5.97532451e-01
-2.91938752e-01 -1.13664532e+00 1.10680640e+00 6.50845841e-02
-6.84598804e-01 -2.87624568e-01 -8.41500163e-02 -7.26461112e-02
3.86004858e-02 5.05049407e-01 -6.53539419e-01 1.18397534e+00
8.20028424e-01 4.70219441e-02 -4.07021999e-01 9.27542746e-01
1.09225087e-01 -1.01104781e-01 -7.56111741e-01 -6.61208153e-01
-1.97687715e-01 -2.60457397e-01 3.41888666e-01 1.17041779e+00
-3.77519459e-01 1.59355000e-01 4.04428750e-01 1.12091184e+00
6.06740177e-01 -4.43359762e-02 -5.85393310e-01 9.07054693e-02
9.12951946e-01 1.04088783e+00 -7.43371546e-01 -4.66006219e-01
-3.20132583e-01 -6.20678008e-01 -4.10684139e-01 3.46376628e-01
-7.45631218e-01 -9.38527942e-01 2.95532376e-01 -1.33498386e-01
-4.73968893e-01 8.37646201e-02 -4.23656166e-01 -7.32060909e-01
1.00174788e-02 -1.30819428e+00 8.42165291e-01 -5.22817373e-01
-9.26625967e-01 2.43496537e-01 8.40262651e-01 -5.27583241e-01
-6.34836674e-01 -9.61217523e-01 -5.13459027e-01 4.59782451e-01
-1.02410054e+00 -1.40466714e+00 -1.16490811e-01 8.92485023e-01
-2.40253434e-01 -6.55545354e-01 1.19010210e+00 5.79088986e-01
-8.85453522e-02 -2.72728831e-01 -4.38721389e-01 -3.81934881e-01
3.73342633e-02 -1.41643727e+00 -8.98072779e-01 4.02872741e-01
-2.19899774e-01 4.96867627e-01 8.42741489e-01 -3.66059870e-01
-1.95354283e+00 -6.09694183e-01 9.27798688e-01 -1.35101750e-01
1.11010945e+00 -2.78062016e-01 -4.14465874e-01 1.04576778e+00
8.77472833e-02 4.11893725e-01 5.08496940e-01 2.37618268e-01
-3.36074084e-01 -3.49490523e-01 -2.04452562e+00 2.39698395e-01
4.36234139e-02 -4.97287720e-01 -7.44175255e-01 1.04459250e+00
6.52871966e-01 -6.65325940e-01 -1.28914559e+00 2.59575322e-02
6.83816135e-01 -1.05031061e+00 7.81741321e-01 -9.49536800e-01
1.32862478e-01 -7.22946525e-01 -5.47338903e-01 -1.03452705e-01
4.09508675e-01 -5.06359041e-01 -3.52511346e-01 1.49099910e+00
8.45540166e-01 -1.05627596e+00 4.28710192e-01 1.24233913e+00
6.43012673e-03 -3.23975354e-01 -1.03654957e+00 -5.41306734e-01
-3.71719748e-01 -6.10168874e-01 1.07818902e+00 1.25718462e+00
1.09645069e+00 -2.15339392e-01 4.48209077e-01 4.84496862e-01
4.80514795e-01 4.16739911e-01 5.25810480e-01 -1.60325968e+00
-2.97302306e-01 -5.49821734e-01 -1.28474808e+00 6.16503537e-01
3.35017651e-01 -6.55827820e-01 -4.39219385e-01 -1.90023267e+00
-3.96984130e-01 -6.11265182e-01 1.45645961e-01 7.18554318e-01
9.32973266e-01 -5.53561807e-01 -2.53820419e-02 -4.13228542e-01
-1.02507091e+00 -4.55310792e-01 7.25716233e-01 -1.93293154e-01
-4.18183468e-02 -1.29181705e-02 -1.15788735e-01 8.78463149e-01
5.10446668e-01 -5.36161304e-01 -3.82924169e-01 1.43116906e-01
1.02176487e+00 3.11469734e-01 3.14957350e-01 -7.27701485e-01
8.59974384e-01 -1.09641993e+00 -2.33808473e-01 -4.22467649e-01
-1.71401948e-01 -1.44512665e+00 6.57512724e-01 7.23947287e-01
-2.55452693e-02 8.34720954e-03 2.68090248e-01 -3.32517270e-03
-5.49094319e-01 -1.06470954e+00 5.00960886e-01 -4.69439566e-01
-7.08354294e-01 -6.71453923e-02 -5.22864103e-01 2.64982611e-01
2.00431681e+00 1.16712853e-01 -5.09533763e-01 3.84808660e-01
-1.06877732e+00 6.29748642e-01 7.32254460e-02 1.05810143e-01
1.64980382e-01 -1.09216094e+00 -2.10129932e-01 -1.71593308e-01
6.92937002e-02 4.96309727e-01 -2.46129960e-01 8.53966594e-01
-1.14106643e+00 7.99471319e-01 -5.53265095e-01 -1.85270458e-01
-1.11550736e+00 7.19915807e-01 2.83558905e-01 -2.41779357e-01
7.70118460e-02 5.40727854e-01 -9.48886216e-01 -1.04968607e+00
5.05200803e-01 -5.20741045e-01 -2.97637165e-01 -2.60768116e-01
5.53233802e-01 5.89283884e-01 5.57060577e-02 5.04974416e-03
-1.01715326e+00 3.79736215e-01 3.54036450e-01 -2.87227988e-01
1.22754157e+00 4.78586912e-01 -1.11007750e+00 8.44340384e-01
1.03850313e-01 -1.18688583e-01 2.71520495e-01 2.71055311e-01
1.03021121e+00 -1.43279880e-01 -3.96144509e-01 -1.46641672e+00
-2.32861608e-01 2.95999765e-01 3.33120853e-01 1.12427282e+00
6.00930810e-01 -2.18891613e-02 -4.17982996e-01 7.01885283e-01
1.05939794e+00 -1.59605360e+00 -8.08110118e-01 7.53781378e-01
9.87982810e-01 -1.03584492e+00 2.86903590e-01 -3.51011276e-01
-5.66922903e-01 1.51715326e+00 6.13745511e-01 9.31121930e-02
1.12267983e+00 1.00594175e+00 -3.35962921e-01 -6.65406525e-01
-1.21872127e+00 1.35621801e-01 -3.79027098e-01 1.16928959e+00
5.40968180e-01 2.99504220e-01 -6.42459631e-01 8.45595121e-01
-3.74225974e-01 7.07244873e-01 7.17477441e-01 1.43283093e+00
-2.15669930e-01 -1.24916375e+00 -8.96842420e-01 2.57452667e-01
-6.14712834e-01 1.30616859e-01 -2.36715332e-01 1.26018655e+00
4.98768359e-01 9.52061474e-01 -1.67345420e-01 2.00356901e-01
5.79524100e-01 8.28668550e-02 7.22230732e-01 -5.62214017e-01
-7.23837018e-01 -8.25233877e-01 6.26394629e-01 -2.48749554e-01
-4.93524253e-01 -5.97001672e-01 -1.88683128e+00 -4.32114571e-01
-3.61582249e-01 9.29861009e-01 1.30761743e+00 1.07091606e+00
5.86199984e-02 1.89113185e-01 -1.27689123e-01 -1.23126455e-01
-3.97616476e-01 -7.97646582e-01 -5.64678669e-01 -3.92901808e-01
-2.41860896e-01 -6.68811500e-01 -1.48018122e-01 2.07286805e-01] | [8.667410850524902, 6.7690839767456055] |
386de307-1c50-43ee-8dbb-3c68ade85927 | rdcnet-instance-segmentation-with-a | 2010.00991 | null | https://arxiv.org/abs/2010.00991v1 | https://arxiv.org/pdf/2010.00991v1.pdf | RDCNet: Instance segmentation with a minimalist recurrent residual network | Instance segmentation is a key step for quantitative microscopy. While several machine learning based methods have been proposed for this problem, most of them rely on computationally complex models that are trained on surrogate tasks. Building on recent developments towards end-to-end trainable instance segmentation, we propose a minimalist recurrent network called recurrent dilated convolutional network (RDCNet), consisting of a shared stacked dilated convolution (sSDC) layer that iteratively refines its output and thereby generates interpretable intermediate predictions. It is light-weight and has few critical hyperparameters, which can be related to physical aspects such as object size or density.We perform a sensitivity analysis of its main parameters and we demonstrate its versatility on 3 tasks with different imaging modalities: nuclear segmentation of H&E slides, of 3D anisotropic stacks from light-sheet fluorescence microscopy and leaf segmentation of top-view images of plants. It achieves state-of-the-art on 2 of the 3 datasets. | ['Markus Rempfler', 'Prisca Liberali', 'Antoine H. F. M. Peters', 'Gustavo de Medeiros', 'Raphael Ortiz'] | 2020-10-02 | null | null | null | null | ['nuclear-segmentation'] | ['medical'] | [ 7.76482046e-01 4.82358754e-01 -5.54850698e-02 -4.73372400e-01
-9.88848507e-01 -7.25382090e-01 4.56220180e-01 -1.06628165e-02
-4.28494960e-01 6.69339359e-01 -2.80158311e-01 -6.01502657e-01
-1.05910271e-01 -3.85617018e-01 -8.66496623e-01 -9.55574453e-01
-1.81364566e-02 8.12198818e-01 4.16856587e-01 1.36905804e-01
2.44472370e-01 7.07239687e-01 -1.10104752e+00 2.97611564e-01
5.94989002e-01 9.66293335e-01 5.18045306e-01 1.19762850e+00
-4.41641314e-03 4.73516732e-01 -2.08089694e-01 -9.26636532e-02
5.14803231e-02 -3.80033672e-01 -1.25588775e+00 2.06674397e-01
1.68070063e-01 -1.60372138e-01 3.06500614e-01 7.60779440e-01
5.86055696e-01 -2.54835248e-01 8.31830680e-01 -8.09054673e-01
-7.95894206e-01 6.15898311e-01 -5.47404945e-01 2.53480464e-01
-3.53207856e-01 3.46803337e-01 1.09766018e+00 -6.94089413e-01
8.29679847e-01 1.09218335e+00 7.81713605e-01 6.94527566e-01
-2.02492881e+00 -6.90361410e-02 2.37298205e-01 -2.20604941e-01
-6.05732739e-01 -2.50817746e-01 4.06343192e-01 -4.55146044e-01
8.09301853e-01 1.32314056e-01 3.49940181e-01 8.53561878e-01
5.81066944e-02 8.05016577e-01 1.33977532e+00 -3.22654486e-01
4.18379217e-01 -1.65856063e-01 1.73919294e-02 6.50562286e-01
-2.24539712e-02 -1.95567727e-01 1.47652239e-01 -5.18125892e-02
1.00686014e+00 -2.60240406e-01 -3.53297710e-01 -3.71338278e-01
-1.31194317e+00 6.46485627e-01 4.48203415e-01 2.75580406e-01
-1.17660142e-01 2.88504273e-01 3.38030040e-01 6.12832010e-02
7.09787726e-01 5.00447989e-01 -9.88049328e-01 2.76212305e-01
-1.06881154e+00 2.11115003e-01 7.94951975e-01 5.44345319e-01
7.55718589e-01 -1.39668331e-01 -2.79609442e-01 5.85656226e-01
3.17126483e-01 3.25905025e-01 2.45904475e-01 -1.20669067e+00
-9.28000212e-02 7.85579562e-01 1.88839644e-01 -2.49324292e-01
-8.27961624e-01 -5.40024191e-02 -8.57944906e-01 5.44611275e-01
6.48736596e-01 -1.25824645e-01 -1.32710946e+00 1.65688980e+00
3.60302567e-01 1.61571637e-01 -2.68151835e-02 9.36366379e-01
9.99378502e-01 6.11737370e-01 2.13946626e-01 -1.97966412e-01
1.14700627e+00 -6.60817564e-01 -4.03584540e-01 -5.52150495e-02
6.92817271e-01 -5.90708077e-01 1.03148985e+00 7.99326450e-02
-1.33510864e+00 -3.23821008e-01 -6.84074640e-01 -2.82045454e-01
-6.68739855e-01 5.30220568e-01 5.34570932e-01 4.84210730e-01
-1.29416871e+00 1.03983009e+00 -1.17322576e+00 -3.53755504e-01
6.38639212e-01 6.63576484e-01 -1.52147487e-01 5.42156637e-01
-4.53885078e-01 7.40328908e-01 2.94140786e-01 4.58920360e-01
-1.02239299e+00 -1.06907856e+00 -3.72374356e-01 -2.46617403e-02
1.94335192e-01 -8.54583323e-01 1.15796649e+00 -5.39450943e-01
-2.07717967e+00 1.62299430e+00 -5.06499186e-02 -6.43927634e-01
3.47576261e-01 -3.99614312e-02 2.41561383e-01 1.95043504e-01
-1.94324732e-01 9.76865530e-01 6.72443211e-01 -1.36580825e+00
-3.38663429e-01 -6.05940342e-01 -1.11867964e-01 -2.55237013e-01
3.90885264e-01 1.04259834e-01 -5.16982675e-02 -3.36358190e-01
1.90717682e-01 -1.10810113e+00 -7.25463629e-01 3.24293524e-01
-8.18023980e-01 4.57249954e-03 1.19108689e+00 -4.95361000e-01
4.51532751e-01 -1.55871201e+00 6.26547813e-01 -2.73824427e-02
2.36064747e-01 5.39351642e-01 -2.62851149e-01 -8.26197490e-03
-1.44591168e-01 4.54818130e-01 -6.79914176e-01 -4.32072401e-01
2.90208608e-02 1.59191519e-01 -3.93460631e-01 5.42434514e-01
6.72492921e-01 1.17751789e+00 -6.49826467e-01 -3.24183643e-01
3.50621730e-01 7.96741605e-01 -3.96777213e-01 3.93880039e-01
-7.22355306e-01 9.50477660e-01 -4.91241068e-01 5.01847625e-01
7.67947614e-01 -8.41131926e-01 1.77154407e-01 -3.81638765e-01
-4.21453744e-01 1.78997219e-02 -6.28594041e-01 1.63061953e+00
-3.18959683e-01 4.53536838e-01 3.94504696e-01 -1.35218084e+00
7.53621519e-01 1.73388496e-01 4.40633208e-01 -1.03317045e-01
1.61817357e-01 3.60213846e-01 -1.77961946e-01 -2.99638599e-01
-7.89215267e-02 8.35639164e-02 4.28248018e-01 5.71253061e-01
3.86616856e-01 -6.84303045e-01 2.02037096e-01 -1.07235409e-01
1.09484911e+00 7.44759500e-01 -6.16645403e-02 -6.73656762e-01
6.36197507e-01 -2.31924534e-01 3.74122471e-01 5.10576487e-01
-5.50021641e-02 9.81922746e-01 8.54173243e-01 -6.17330730e-01
-1.43719292e+00 -9.20135498e-01 -2.64701307e-01 1.09976447e+00
-1.96561828e-01 -1.58681899e-01 -9.41611707e-01 -5.19148886e-01
-4.05674815e-01 2.32363477e-01 -7.78444588e-01 2.16775879e-01
-5.14310062e-01 -1.24571288e+00 5.35910308e-01 2.94213593e-01
3.65009964e-01 -1.28171206e+00 -9.49645996e-01 2.05258012e-01
2.35300809e-01 -1.31685627e+00 1.05381168e-01 7.20131397e-01
-1.06395912e+00 -1.18841016e+00 -9.07180727e-01 -6.10667408e-01
7.23809958e-01 -1.61699966e-01 1.49338841e+00 6.97284266e-02
-6.16158724e-01 1.16952345e-01 -4.23591509e-02 -3.37828577e-01
-4.77847666e-01 4.42415923e-01 -3.68251055e-01 -3.02837253e-01
-1.23361953e-01 -8.15011501e-01 -7.87418902e-01 2.90795267e-01
-1.01872945e+00 2.53879160e-01 4.99499083e-01 1.12313354e+00
1.13538182e+00 -4.71517652e-01 3.07888120e-01 -1.51092339e+00
2.68593758e-01 9.71634462e-02 -9.72063065e-01 4.35519785e-01
-4.28316891e-01 1.30668044e-01 8.56036007e-01 -1.62631795e-01
-8.17390263e-01 3.87210965e-01 -8.59929994e-02 -2.77711898e-01
-3.16852540e-01 3.06369990e-01 1.10812098e-01 -3.06445301e-01
5.75999439e-01 5.82358874e-02 1.22391663e-01 -2.52325326e-01
3.22819591e-01 9.14853886e-02 5.51735342e-01 -7.04126954e-01
7.18838394e-01 7.12126434e-01 5.03158510e-01 -7.87602246e-01
-1.17790580e+00 -9.25252885e-02 -9.79936600e-01 -6.13741053e-04
9.32717562e-01 -3.85502934e-01 -1.15921402e+00 5.10045767e-01
-1.10198665e+00 -1.08508909e+00 -3.75486851e-01 -4.29589674e-02
-1.03970623e+00 6.15581125e-02 -8.66995871e-01 -7.68069267e-01
-6.00319326e-01 -1.39109814e+00 1.58069801e+00 4.81280893e-01
1.63006559e-01 -1.26585627e+00 3.90180647e-02 -4.54855748e-02
4.96201664e-01 5.58985114e-01 1.27100432e+00 -4.39751625e-01
-7.14835644e-01 1.90395370e-01 -4.89812940e-01 2.71147907e-01
-4.58528176e-02 7.47635663e-01 -1.25036442e+00 -3.16226006e-01
-4.03035581e-01 -7.55105793e-01 1.11515033e+00 1.00334716e+00
1.70378566e+00 -1.29788190e-01 -4.70784754e-01 8.91411066e-01
1.42149496e+00 -3.76747340e-01 7.25039482e-01 8.86542350e-02
5.84219635e-01 8.11831892e-01 1.28204390e-01 2.80117780e-01
6.63170516e-02 3.83990735e-01 6.74322844e-01 -6.51993990e-01
-1.76524483e-02 2.66822577e-01 -7.96047971e-03 3.65401566e-01
-4.05398816e-01 -2.27861449e-01 -8.13792348e-01 3.18825692e-01
-1.86343920e+00 -6.93453789e-01 -1.53123662e-01 2.00395226e+00
9.67083156e-01 1.98016644e-01 1.86644644e-01 4.67898622e-02
5.89829266e-01 1.49963066e-01 -9.51863706e-01 -3.01395744e-01
-2.63569742e-01 4.83459473e-01 5.41375697e-01 4.15051013e-01
-1.04983342e+00 1.16086328e+00 6.91815710e+00 4.26975906e-01
-1.20383108e+00 -3.05887014e-01 1.41305888e+00 2.70619810e-01
-3.69036883e-01 1.15017854e-01 -8.94652426e-01 2.00079530e-01
9.96719360e-01 3.71226192e-01 3.72167379e-01 4.00045097e-01
4.12885457e-01 -4.83834073e-02 -1.13829398e+00 3.86947244e-01
-5.83938301e-01 -1.64079785e+00 -1.86398011e-02 -7.27715790e-02
7.00863004e-01 3.90696883e-01 3.76603931e-01 -4.97023063e-03
6.48605168e-01 -1.31492829e+00 1.77596167e-01 4.89280850e-01
8.88661981e-01 -4.53052849e-01 5.89090168e-01 3.15884262e-01
-7.90775895e-01 2.29714155e-01 -5.69376469e-01 5.79112232e-01
9.11673307e-02 7.65072167e-01 -7.76053190e-01 8.52128342e-02
7.65587866e-01 6.76904738e-01 -5.86147249e-01 6.27130747e-01
4.16397117e-02 7.98552215e-01 -5.67680478e-01 3.48730177e-01
4.72557455e-01 -4.43168640e-01 3.13333243e-01 1.34141207e+00
1.10449389e-01 7.20419735e-02 -7.40830787e-03 1.51239300e+00
-1.89378187e-01 -3.23015541e-01 -3.86727631e-01 -2.77084857e-01
2.26074103e-02 1.86891758e+00 -1.41399324e+00 -1.69468466e-02
-1.79001763e-02 7.16264069e-01 4.00416344e-01 4.45529044e-01
-6.16313636e-01 -1.77964345e-01 3.63071799e-01 6.87327459e-02
6.18645370e-01 6.15535527e-02 -1.17937617e-01 -8.46879005e-01
-3.63040328e-01 -3.73386115e-01 9.93072093e-02 -8.81585836e-01
-1.26078546e+00 3.75451773e-01 -3.92297655e-01 -6.32764459e-01
-3.10097218e-01 -1.21547496e+00 -7.45720327e-01 8.81067991e-01
-1.73131692e+00 -1.55306184e+00 -1.04572795e-01 1.82469130e-01
2.96475708e-01 1.74601287e-01 1.26171005e+00 -2.26188585e-01
-6.37137234e-01 3.57408635e-02 1.17978320e-01 -1.99311793e-01
3.63027394e-01 -1.61754620e+00 6.86696172e-01 5.09186268e-01
-1.65854901e-01 3.96277398e-01 8.98102343e-01 -2.31356084e-01
-1.05562305e+00 -1.17428541e+00 5.70780039e-01 -6.02928042e-01
6.28441036e-01 -2.64434904e-01 -9.82389390e-01 7.87715971e-01
8.85383263e-02 7.46616006e-01 6.14828527e-01 3.53306904e-02
-1.60173371e-01 3.02425146e-01 -1.16986716e+00 6.62777603e-01
8.71680558e-01 -4.93762583e-01 1.91745877e-01 6.24143124e-01
6.62372231e-01 -5.93239367e-01 -1.04083431e+00 6.71588540e-01
4.63066131e-01 -1.11467218e+00 9.50080156e-01 -8.80503058e-01
6.56540394e-01 -4.02920425e-01 1.33413732e-01 -1.27658224e+00
-2.88154989e-01 -9.21713531e-01 -8.72513652e-02 9.84297633e-01
7.76362181e-01 -2.72006959e-01 1.03641391e+00 4.59083200e-01
-2.74097681e-01 -1.09341931e+00 -6.60538971e-01 -2.38463730e-01
2.51169026e-01 6.87862784e-02 2.18084514e-01 4.33020353e-01
-3.91238421e-01 4.69517738e-01 8.72227177e-02 6.61018714e-02
5.94334006e-01 3.92363250e-01 4.95261490e-01 -1.46627545e+00
-9.71446708e-02 -4.34264630e-01 -1.13244206e-01 -1.08980584e+00
4.89964783e-01 -7.44961619e-01 7.55699947e-02 -1.34999967e+00
2.56513357e-01 -3.00102592e-01 -1.55968472e-01 4.27593976e-01
-6.67579770e-02 4.00159001e-01 -2.43206412e-01 1.18543290e-01
-7.25501001e-01 2.76897460e-01 1.47056496e+00 -1.41086206e-01
-1.84906557e-01 2.47737378e-01 -4.13540661e-01 8.15901756e-01
8.34923804e-01 -2.22117439e-01 -2.50855327e-01 -3.61655325e-01
3.08405727e-01 1.64733715e-02 7.23777592e-01 -7.17977464e-01
3.59303178e-03 -5.26454411e-02 4.75932717e-01 -7.00976729e-01
2.77625561e-01 -7.05110073e-01 -1.41710058e-01 3.78895164e-01
-8.12663972e-01 -1.66278720e-01 1.52692661e-01 4.74768162e-01
2.89966196e-01 -2.22963691e-01 1.20741224e+00 -4.25016105e-01
-1.87070474e-01 5.48998773e-01 -4.01344776e-01 1.02462746e-01
6.70659006e-01 -2.07556814e-01 -4.39751148e-01 -3.00045032e-02
-8.32081020e-01 1.69738941e-02 3.82350951e-01 -2.67122090e-01
4.02846187e-01 -7.74039984e-01 -7.61692703e-01 -5.18747605e-02
-2.07201377e-01 4.94645476e-01 -3.51807587e-02 9.08391714e-01
-6.49816453e-01 5.14625788e-01 -2.20721275e-01 -9.73301232e-01
-1.00116539e+00 1.01157770e-01 5.73294103e-01 -6.45119011e-01
-6.19737148e-01 9.52396810e-01 2.87069738e-01 -9.38254118e-01
-1.81095824e-02 -7.99179852e-01 -4.21569318e-01 -3.93579036e-01
3.50293577e-01 1.20321453e-01 9.44353640e-03 -3.26778322e-01
-1.03185676e-01 6.86050117e-01 -1.75353214e-01 1.91985697e-01
1.78512824e+00 -9.07652825e-02 -3.28170508e-01 6.78445697e-01
9.61211383e-01 -7.09324241e-01 -1.80171001e+00 -6.08829819e-02
2.03741029e-01 1.55353859e-01 1.12895876e-01 -9.00500596e-01
-1.17970037e+00 1.00136101e+00 6.68664455e-01 4.49921489e-01
1.00400174e+00 -3.54164913e-02 5.60582161e-01 5.55756629e-01
3.61540876e-02 -9.62294757e-01 1.59794185e-02 6.03023350e-01
4.66375142e-01 -1.54629374e+00 9.87396110e-03 -5.17667532e-01
-9.67534035e-02 1.28215325e+00 4.45517361e-01 -2.31284156e-01
5.90959013e-01 5.12278318e-01 2.77956817e-02 -4.01774287e-01
-1.02301562e+00 -3.03756505e-01 2.32600924e-02 7.38651216e-01
9.37546253e-01 -1.98513225e-01 1.59453437e-01 4.11165327e-01
1.73226073e-01 1.38518751e-01 3.36122036e-01 6.09879315e-01
-2.86294967e-01 -9.11228299e-01 2.05633834e-01 5.45866072e-01
-8.09119940e-01 5.14449291e-02 -3.83873075e-01 6.56964302e-01
-1.84841782e-01 3.34793121e-01 1.39459288e-02 1.29388928e-01
5.81478141e-02 -2.08481461e-01 7.26101577e-01 -6.51797950e-01
-6.24242902e-01 -2.81776078e-02 -2.68804401e-01 -5.78733087e-01
-9.04774547e-01 -4.53604162e-01 -1.35913396e+00 -2.00946763e-01
-4.29586232e-01 -8.99035186e-02 6.07224643e-01 1.03060257e+00
3.63329828e-01 6.79588735e-01 3.58074188e-01 -1.43871427e+00
-3.73336166e-01 -8.89894366e-01 -6.23324394e-01 1.76775590e-01
4.15567666e-01 -2.85878778e-01 -4.51637596e-01 5.55167139e-01] | [14.424206733703613, -3.0785269737243652] |
92f29b81-81ee-43b1-93ff-c6906ed8f894 | deep-attention-model-for-triage-of-emergency | 1804.03240 | null | http://arxiv.org/abs/1804.03240v1 | http://arxiv.org/pdf/1804.03240v1.pdf | Deep Attention Model for Triage of Emergency Department Patients | Optimization of patient throughput and wait time in emergency departments
(ED) is an important task for hospital systems. For that reason, Emergency
Severity Index (ESI) system for patient triage was introduced to help guide
manual estimation of acuity levels, which is used by nurses to rank the
patients and organize hospital resources. However, despite improvements that it
brought to managing medical resources, such triage system greatly depends on
nurse's subjective judgment and is thus prone to human errors. Here, we propose
a novel deep model based on the word attention mechanism designed for
predicting a number of resources an ED patient would need. Our approach
incorporates routinely available continuous and nominal (structured) data with
medical text (unstructured) data, including patient's chief complaint, past
medical history, medication list, and nurse assessment collected for 338,500 ED
visits over three years in a large urban hospital. Using both structured and
unstructured data, the proposed approach achieves the AUC of $\sim 88\%$ for
the task of identifying resource intensive patients (binary classification),
and the accuracy of $\sim 44\%$ for predicting exact category of number of
resources (multi-class classification task), giving an estimated lift over
nurses' performance by 16\% in accuracy. Furthermore, the attention mechanism
of the proposed model provides interpretability by assigning attention scores
for nurses' notes which is crucial for decision making and implementation of
such approaches in the real systems working on human health. | ['Ivan Stojkovic', 'Daniel Del Portal', 'Djordje Gligorijevic', 'Wayne Satz', 'Jelena Stojanovic', 'Zoran Obradovic', 'Kathrin Schreyer'] | 2018-03-28 | null | null | null | null | ['deep-attention', 'deep-attention'] | ['computer-vision', 'natural-language-processing'] | [-7.95222968e-02 1.68014929e-01 -2.92047262e-02 -2.92656749e-01
-7.01845229e-01 -8.11926201e-02 -2.54692614e-01 1.03904891e+00
-7.86071539e-01 6.88782394e-01 7.41841376e-01 -7.14352310e-01
-7.54887581e-01 -7.37592876e-01 -3.74419838e-02 -3.60411555e-01
-2.24099234e-01 1.06630635e+00 -5.22655785e-01 -1.64509972e-03
1.06722943e-01 4.29992408e-01 -1.11288583e+00 5.89054883e-01
7.74307311e-01 1.45458770e+00 2.72918820e-01 6.41794026e-01
-1.92385036e-02 1.14923179e+00 -4.80700999e-01 -9.77569818e-02
3.63935947e-01 -3.60615641e-01 -7.88437545e-01 -2.65863061e-01
-2.23974720e-01 -6.09861672e-01 -3.52108032e-01 5.96338332e-01
1.07136381e+00 7.88391232e-02 6.93794250e-01 -5.86348057e-01
-4.05099571e-01 6.33254468e-01 5.33485487e-02 6.21994495e-01
4.23599780e-01 3.22079033e-01 6.97968543e-01 -6.46960676e-01
-9.38561931e-02 8.17774355e-01 7.29794621e-01 3.65547627e-01
-4.66904879e-01 -5.56150138e-01 -3.54639679e-01 1.18827231e-01
-1.31292510e+00 -3.48578542e-01 7.59013295e-02 -6.10044062e-01
1.00615406e+00 3.75035644e-01 6.77731097e-01 7.56216645e-01
5.27513683e-01 2.13417709e-01 7.19528258e-01 -1.57010332e-01
1.49321377e-01 1.51802927e-01 2.91472405e-01 7.08745718e-01
4.50177431e-01 -2.97549635e-01 -1.62487924e-01 -4.32619065e-01
7.99316525e-01 9.81016695e-01 -3.74487668e-01 3.47848207e-01
-1.45124149e+00 6.52261913e-01 3.84400547e-01 1.58301011e-01
-9.42817390e-01 -3.93328726e-01 7.25511670e-01 -1.02604553e-01
-5.48658259e-02 7.65978515e-01 -6.44325495e-01 -4.28883255e-01
-6.50952995e-01 -1.40635192e-01 8.22966337e-01 8.87342751e-01
-9.70187690e-03 -1.10160813e-01 -9.03859556e-01 6.75119281e-01
-7.59979710e-02 6.43073916e-01 9.39990282e-01 -7.57492423e-01
5.86966157e-01 9.54618454e-01 4.55151737e-01 -9.66993213e-01
-1.19417262e+00 -4.72446412e-01 -1.44442618e+00 -5.69352508e-01
-3.68981175e-02 -4.08495545e-01 -7.48149335e-01 1.22673225e+00
-1.88564241e-01 -3.94838184e-01 1.75447941e-01 8.22776735e-01
9.25547719e-01 2.10933402e-01 3.22445780e-01 -5.57174742e-01
1.88793695e+00 -6.96840942e-01 -9.60136592e-01 1.11163281e-01
8.06691051e-01 -5.07775664e-01 9.83927071e-01 3.47579867e-01
-1.37040699e+00 -4.37071651e-01 -4.30394590e-01 1.92724854e-01
-1.58920571e-01 1.23750225e-01 3.48658592e-01 2.47316241e-01
-8.15432489e-01 5.30274153e-01 -6.76398396e-01 -3.21024150e-01
8.50945413e-01 5.18884778e-01 -1.49258628e-01 -1.33235782e-01
-1.20482218e+00 7.85991251e-01 2.92478919e-01 1.06646605e-01
-6.11322403e-01 -6.47502601e-01 -8.44800591e-01 7.47937143e-01
2.31913060e-01 -1.19002509e+00 1.17642784e+00 -2.38694757e-01
-8.89670014e-01 6.02879286e-01 -4.42715660e-02 -4.80005115e-01
3.63358915e-01 -4.82231051e-01 -4.00227875e-01 2.22397134e-01
8.88160467e-02 2.85149962e-01 2.78182626e-01 -6.07417047e-01
-7.81250238e-01 -5.10098040e-01 -1.56532079e-01 3.58506680e-01
-6.00219607e-01 -1.17851675e-01 -7.05926493e-02 -4.87680644e-01
-1.43870741e-01 -5.61506212e-01 -5.69345355e-01 -4.65288192e-01
-2.08813488e-01 -4.95678931e-02 1.62433043e-01 -8.74678373e-01
2.05043817e+00 -2.08795309e+00 -3.91450614e-01 1.54002294e-01
7.04430223e-01 3.96715105e-01 2.61505008e-01 4.93017048e-01
-1.12551793e-01 2.60143936e-01 -1.93945080e-01 -2.31598113e-02
-1.70950264e-01 8.22224542e-02 -2.64881402e-02 1.20299190e-01
2.60514468e-01 1.05899858e+00 -9.36036706e-01 -6.34344041e-01
2.99831420e-01 3.49638939e-01 -7.25413978e-01 6.71352923e-01
5.29310524e-01 5.29970050e-01 -4.72105980e-01 6.83025897e-01
1.18737839e-01 -7.72186518e-01 -1.84312258e-02 -6.19648136e-02
-1.79103483e-02 2.87669301e-01 -8.25182915e-01 1.14206028e+00
-5.57848454e-01 7.95397907e-02 -1.45883247e-01 -8.57627332e-01
7.54565060e-01 6.66223884e-01 1.06961429e+00 -6.43821299e-01
6.31796956e-01 -1.54098086e-02 4.00483072e-01 -9.93742943e-01
-3.15878238e-03 -1.92439910e-02 -2.39365205e-01 4.74865377e-01
-5.85220158e-01 3.56394470e-01 -1.76193684e-01 3.76535095e-02
1.68146908e+00 -7.03760564e-01 8.29084098e-01 -3.25634360e-01
4.70732063e-01 2.68783327e-02 5.17553806e-01 7.52148747e-01
-3.47323626e-01 5.49007177e-01 2.96927691e-01 -1.03383362e+00
-9.89536881e-01 -6.65275335e-01 -2.91247755e-01 7.02658415e-01
-2.92225748e-01 -1.40191421e-01 -5.84536970e-01 -4.88235533e-01
-5.25461603e-03 5.35092711e-01 -6.02717459e-01 -2.73413807e-01
-5.44888616e-01 -8.12099397e-01 3.00039470e-01 9.41290915e-01
3.63907635e-01 -1.38356924e+00 -1.43639278e+00 4.86890078e-01
-2.68621057e-01 -9.19157267e-01 -5.97450674e-01 5.20183444e-01
-8.29878688e-01 -1.28715336e+00 -8.43429089e-01 -5.95853209e-01
6.26637697e-01 -1.05712391e-01 1.17310262e+00 2.09728599e-01
-6.42509222e-01 4.76334989e-01 -3.36512923e-01 -6.42643213e-01
4.68541160e-02 7.70648122e-02 3.73062253e-01 -2.87409633e-01
6.70967400e-01 -5.41863404e-02 -1.33717310e+00 1.72934696e-01
-8.76529157e-01 -1.03408284e-01 8.53485703e-01 1.11538231e+00
4.88794655e-01 -1.20561555e-01 8.80556941e-01 -9.53228593e-01
1.01895344e+00 -8.77497137e-01 -1.69812500e-01 -6.88942820e-02
-1.02512586e+00 -2.55358256e-02 8.99956882e-01 1.05657622e-01
-4.15074766e-01 -2.34077066e-01 -1.72870621e-01 -3.72033834e-01
-2.94156104e-01 4.27723706e-01 2.14576229e-01 6.74693167e-01
3.70149970e-01 1.43855512e-01 -8.38821307e-02 -3.58513802e-01
-5.37549078e-01 1.14809334e+00 2.91190028e-01 -9.21880677e-02
3.48691526e-03 9.30895060e-02 1.60216257e-01 -2.32540771e-01
-1.08587849e+00 -8.32766831e-01 -5.48919201e-01 2.52410054e-01
1.18289518e+00 -8.91297102e-01 -1.46346223e+00 -2.50521570e-01
-1.07827020e+00 -1.03728421e-01 -3.93448859e-01 7.03533351e-01
-4.03067350e-01 1.78549439e-01 -5.20342350e-01 -9.52682734e-01
-9.55092192e-01 -1.21021330e+00 1.01297319e+00 -9.57821533e-02
-6.36178493e-01 -9.14162993e-01 -2.73046851e-01 4.31511790e-01
7.06141770e-01 2.45247960e-01 1.20557690e+00 -1.07401192e+00
-1.60559565e-01 -4.23832059e-01 -5.62249243e-01 3.76081079e-01
4.25586283e-01 -6.75999343e-01 -5.03350198e-01 -1.99086815e-01
2.79037148e-01 1.80768669e-02 4.57879812e-01 6.74394965e-01
1.86937201e+00 -6.17954731e-01 -1.85216561e-01 6.41388774e-01
1.27440488e+00 7.17164874e-01 5.08284628e-01 1.03449285e-01
6.28860414e-01 3.50104153e-01 5.73386371e-01 1.30137968e+00
5.71904182e-01 1.69365965e-02 5.21779716e-01 -3.13193411e-01
5.12439609e-01 3.52544844e-01 -2.28422076e-01 1.06358945e+00
-2.47042626e-01 -3.93644154e-01 -1.51375425e+00 4.41944987e-01
-1.78161204e+00 -7.53609180e-01 4.63926941e-02 2.30196905e+00
6.53299034e-01 -1.44336689e-02 -4.24499661e-02 9.35577303e-02
5.08405626e-01 -4.76070404e-01 -5.78411639e-01 -6.21330798e-01
3.55599225e-01 3.41714084e-01 7.30332315e-01 -5.56833856e-02
-8.15123141e-01 2.31385201e-01 6.30283213e+00 2.09014297e-01
-8.35878253e-01 -2.61526965e-02 1.11675107e+00 -2.98793346e-01
2.14532867e-01 -8.59928191e-01 -7.64545143e-01 8.39495480e-01
1.33942509e+00 8.02949145e-02 2.54967183e-01 7.33318508e-01
7.36022949e-01 -1.42368630e-01 -1.29815114e+00 1.42381120e+00
-4.62143272e-02 -1.23716331e+00 1.64944157e-01 9.25007612e-02
5.54046035e-01 -2.61095911e-01 6.67430535e-02 3.15125763e-01
2.14472950e-01 -1.42992806e+00 -2.90281381e-02 9.22747970e-01
1.19302249e+00 -7.01950073e-01 1.58498490e+00 3.28310162e-01
-8.66338074e-01 -6.54362023e-01 -2.36326054e-01 -2.52976209e-01
9.34396833e-02 3.87566268e-01 -1.23215270e+00 2.51908571e-01
8.99423838e-01 3.47705901e-01 -1.39398172e-01 1.02978683e+00
5.48857450e-01 4.89484400e-01 2.46528443e-02 2.23409072e-01
3.87833804e-01 8.08408950e-03 -3.42275528e-03 1.18978226e+00
6.06251299e-01 8.43884945e-01 3.31799507e-01 1.82779342e-01
-8.82921219e-02 3.31644475e-01 -6.90105438e-01 -1.07372798e-01
3.44396055e-01 1.18397534e+00 -6.05789781e-01 -6.69376731e-01
-2.55330443e-01 6.81332946e-01 3.01676616e-02 1.42071411e-01
-8.84253204e-01 -4.98127371e-01 4.38240170e-01 5.99045575e-01
-1.24276020e-01 1.47241011e-01 -6.35295868e-01 -6.28646135e-01
-3.90591502e-01 -8.19792926e-01 6.36719942e-01 -6.01351917e-01
-1.12201345e+00 9.27984834e-01 -4.29145217e-01 -1.41479957e+00
-2.33702540e-01 -6.61292136e-01 -2.07338884e-01 1.05203772e+00
-1.51225865e+00 -1.97609216e-01 -8.12046826e-01 5.88787377e-01
6.41193211e-01 -3.85618150e-01 1.17106187e+00 6.56314373e-01
-6.96949005e-01 4.78259832e-01 -7.21657872e-02 3.61306071e-01
8.42263520e-01 -1.16681743e+00 -3.38615716e-01 -1.60750717e-01
-9.46211159e-01 7.57097304e-01 1.84503987e-01 -5.90665758e-01
-1.20212615e+00 -1.30088663e+00 1.01534820e+00 -6.69200838e-01
2.32513189e-01 4.24326330e-01 -6.89139247e-01 4.03572202e-01
6.15339763e-02 -2.18564615e-01 1.36754310e+00 -3.62670571e-01
4.53169346e-01 1.19847665e-02 -1.19933224e+00 3.46785039e-01
9.85885262e-01 -1.57426875e-02 -5.88491857e-01 4.83209938e-01
8.28228116e-01 -2.97888696e-01 -1.30798841e+00 6.16685033e-01
4.01568592e-01 -7.48891056e-01 7.49415457e-01 -1.05624318e+00
7.74481177e-01 3.29539418e-01 -1.08711876e-01 -1.05413628e+00
-6.26083553e-01 -4.16080743e-01 -6.04390725e-02 3.46221566e-01
1.95759788e-01 -7.71695495e-01 3.57433945e-01 9.78253424e-01
-5.29205561e-01 -1.44601655e+00 -5.74828386e-01 -1.19663984e-01
-5.60900390e-01 -6.28384426e-02 7.93018878e-01 9.85618174e-01
1.00955600e-03 3.25386167e-01 -1.44598812e-01 2.80724950e-02
1.03048153e-01 -2.54983842e-01 2.52462029e-01 -1.41182649e+00
-6.79563731e-02 -5.42019308e-01 -3.17542523e-01 -6.10811412e-01
-5.10413051e-01 -7.84562230e-01 -1.87273264e-01 -2.14034510e+00
4.54102516e-01 -5.25041819e-01 -9.42329347e-01 5.06119072e-01
-2.56560922e-01 -1.34685457e-01 -2.67585725e-01 3.81999075e-01
-8.07173610e-01 3.21380675e-01 1.21752393e+00 1.64366782e-01
-3.16258132e-01 5.45215458e-02 -9.55697000e-01 7.92241096e-01
6.74900353e-01 -4.79140282e-01 1.86497364e-02 -5.13731539e-01
3.27722192e-01 6.13385677e-01 1.22071557e-01 -1.12580693e+00
1.43281862e-01 -2.60398090e-01 6.04556859e-01 -5.38515031e-01
4.11644317e-02 -1.25574422e+00 -1.98122248e-01 8.94790292e-01
-3.67808312e-01 8.31183791e-01 2.56816059e-01 3.59501094e-01
-3.09066445e-01 8.38615224e-02 5.07549286e-01 -4.04600859e-01
-1.36086851e-01 7.00890183e-01 -3.38852763e-01 2.32417792e-01
9.51382041e-01 -1.56901121e-01 -9.87539217e-02 -2.31261164e-01
-8.53675663e-01 5.47830284e-01 -5.88329472e-02 -8.10866337e-03
8.03005040e-01 -1.07091117e+00 -7.26885438e-01 2.24025667e-01
1.87160522e-01 3.70056212e-01 5.12735605e-01 1.36912179e+00
-9.46719229e-01 8.67751420e-01 -1.79497659e-01 -3.77773851e-01
-9.63700116e-01 7.87645221e-01 1.72348917e-02 -4.77845609e-01
-5.41873872e-01 5.04232347e-01 2.17994004e-01 1.43482774e-01
4.70510989e-01 -6.71323478e-01 -5.13528228e-01 1.83217019e-01
7.63331592e-01 4.98789161e-01 2.70902485e-01 -3.29656452e-01
-3.36628646e-01 3.60616326e-01 -2.96744648e-02 5.27536809e-01
1.45044744e+00 -1.80018455e-01 -5.97415008e-02 3.32096666e-01
9.20362771e-01 -3.08946788e-01 -5.97015917e-01 -1.35621116e-01
-1.90096751e-01 -1.53675318e-01 -5.53458445e-02 -8.38731050e-01
-6.47900999e-01 1.05809474e+00 8.03022921e-01 1.30174488e-01
1.30773127e+00 -2.89034605e-01 1.18385577e+00 5.40902376e-01
2.59085037e-02 -8.73427689e-01 1.65511330e-03 4.94300455e-01
6.31060719e-01 -1.68215203e+00 -3.00146401e-01 2.12638125e-01
-9.74310517e-01 1.03788030e+00 4.60801274e-01 1.62007123e-01
8.09546173e-01 1.33737653e-01 8.88885260e-02 -4.93374079e-01
-7.15757489e-01 -6.60055578e-02 1.00261226e-01 9.97204110e-02
5.50424218e-01 2.82979876e-01 -2.52454251e-01 1.16761279e+00
-4.14167270e-02 1.48889467e-01 3.34506482e-01 8.67305040e-01
-6.05414093e-01 -3.72896284e-01 -3.62526387e-01 1.19844437e+00
-8.68787527e-01 -7.05007493e-01 2.09886789e-01 3.02486956e-01
2.04077780e-01 6.84479415e-01 2.87197679e-01 -1.40407607e-01
6.15934730e-01 4.35022295e-01 -1.29328281e-01 -1.00180697e+00
-1.00518656e+00 -7.04580247e-02 -2.01130942e-01 -4.17462111e-01
2.13621289e-01 -2.58573145e-01 -1.56169307e+00 -6.37660623e-02
3.03658932e-01 2.84654289e-01 2.82363474e-01 6.82639778e-01
7.84764886e-01 1.28600049e+00 6.12399757e-01 -3.49731892e-01
-7.62939632e-01 -1.05713701e+00 -3.98213714e-01 5.62182009e-01
4.58116412e-01 -4.98231977e-01 -1.76387176e-01 -2.55764574e-02] | [7.944104194641113, 6.260335445404053] |
0a34cbb3-fc63-4709-bdc9-6985b4f11ac0 | investigating-prompting-techniques-for-zero | 2306.09996 | null | https://arxiv.org/abs/2306.09996v1 | https://arxiv.org/pdf/2306.09996v1.pdf | Investigating Prompting Techniques for Zero- and Few-Shot Visual Question Answering | Visual question answering (VQA) is a challenging task that requires the ability to comprehend and reason with visual information. While recent vision-language models have made strides, they continue to struggle with zero-shot VQA, particularly in handling complex compositional questions and adapting to new domains i.e. knowledge-based reasoning. This paper explores the use of various prompting strategies, focusing on the BLIP2 model, to enhance zero-shot VQA performance. We conduct a comprehensive investigation across several VQA datasets, examining the effectiveness of different question templates, the role of few-shot exemplars, the impact of chain-of-thought (CoT) reasoning, and the benefits of incorporating image captions as additional visual cues. Despite the varied outcomes, our findings demonstrate that carefully designed question templates and the integration of additional visual cues, like image captions, can contribute to improved VQA performance, especially when used in conjunction with few-shot examples. However, we also identify a limitation in the use of chain-of-thought rationalization, which negatively affects VQA accuracy. Our study thus provides critical insights into the potential of prompting for improving zero-shot VQA performance. | ['Aishwarya Agrawal', 'Le Zhang', 'Rabiul Awal'] | 2023-06-16 | null | null | null | null | ['visual-question-answering-1', 'image-captioning', 'question-answering'] | ['computer-vision', 'computer-vision', 'natural-language-processing'] | [ 1.55252770e-01 1.45320773e-01 1.23449109e-01 -3.54736507e-01
-7.20515490e-01 -6.69494748e-01 8.11958492e-01 2.20754460e-01
-4.61792648e-01 2.35082135e-01 5.28540313e-01 -6.23791397e-01
-1.21408165e-01 -5.50206363e-01 -6.66576207e-01 -2.02009425e-01
5.53722799e-01 1.67040378e-01 4.50105280e-01 -4.79915529e-01
4.78361964e-01 4.23989058e-01 -1.65609789e+00 5.01437664e-01
9.26237822e-01 6.81999028e-01 2.12559342e-01 6.45620465e-01
-6.29786432e-01 1.24316907e+00 -8.10095370e-01 -6.78427339e-01
2.54888892e-01 -4.71469015e-01 -8.24838638e-01 3.35860878e-01
1.05027282e+00 -7.36192822e-01 -2.60031372e-01 7.16900766e-01
4.96326029e-01 3.20852846e-01 5.46175897e-01 -1.43239701e+00
-1.15511608e+00 1.54917732e-01 -3.09577554e-01 6.86399102e-01
5.92325747e-01 1.00300431e+00 1.02801204e+00 -9.32405233e-01
7.53771961e-01 1.39432824e+00 5.17364025e-01 5.46910822e-01
-1.23340261e+00 -4.81817871e-01 3.04241359e-01 6.44533753e-01
-9.60288167e-01 -6.28820002e-01 7.76715100e-01 -5.28993189e-01
1.21424711e+00 1.43066747e-02 7.71257639e-01 1.05470157e+00
-3.61254299e-03 8.21551085e-01 1.27249491e+00 -5.19044936e-01
2.71130592e-01 1.71575740e-01 2.46482342e-01 6.49715006e-01
2.69818068e-01 1.15168691e-01 -6.21628046e-01 2.79865190e-02
6.49818063e-01 -2.02741235e-01 -2.74920404e-01 -2.91663706e-01
-1.11933482e+00 6.69755578e-01 6.94755852e-01 1.58806726e-01
-4.90301967e-01 2.39610076e-01 3.09938461e-01 8.29822123e-02
1.10127233e-01 7.87971079e-01 -8.50828271e-03 -1.25815436e-01
-8.49401116e-01 3.90810072e-01 4.49870884e-01 7.84955740e-01
6.84346735e-01 2.33371660e-01 -9.65346456e-01 6.59211695e-01
1.63324460e-01 5.14619052e-01 -2.31158193e-02 -1.18743920e+00
5.78643024e-01 7.73743868e-01 1.99830547e-01 -8.34016979e-01
-3.32511216e-01 -1.79338753e-01 1.08137242e-01 3.37072968e-01
7.30894744e-01 -1.99556053e-02 -1.36383271e+00 1.62140548e+00
2.20770106e-01 -3.72724175e-01 -5.55939153e-02 1.14665437e+00
1.10295570e+00 4.92962360e-01 6.35534763e-01 -1.13901393e-02
1.75427604e+00 -1.03022659e+00 -8.94737065e-01 -6.57236516e-01
4.11259383e-01 -7.79242754e-01 1.90067744e+00 3.96720581e-02
-1.16309679e+00 -5.04771650e-01 -9.06769991e-01 -5.53717852e-01
-4.14985538e-01 -2.82834083e-01 4.80141342e-01 7.15609074e-01
-1.10036480e+00 -1.72687098e-01 -4.17245418e-01 -5.85676730e-01
6.42074585e-01 -1.59033179e-01 -1.98633939e-01 -5.96580803e-01
-9.04740632e-01 1.20682716e+00 -5.64758480e-02 -1.55528903e-01
-8.83038878e-01 -7.88722396e-01 -8.18012059e-01 1.40056133e-01
6.92797601e-01 -1.10272098e+00 1.49351263e+00 -1.20704806e+00
-9.42299366e-01 6.20040715e-01 -3.42233777e-01 -2.73337811e-01
3.87454301e-01 -1.49363488e-01 -8.69186148e-02 6.68192565e-01
2.42282122e-01 1.04628563e+00 8.60810578e-01 -1.30286801e+00
-2.61878103e-01 -3.89222026e-01 5.56747675e-01 3.04967135e-01
-7.75785521e-02 1.72934905e-02 -3.93844634e-01 -5.59897780e-01
-2.81451643e-01 -6.26092732e-01 -4.23517376e-02 3.66119772e-01
8.49672109e-02 -3.71036381e-01 6.68017864e-01 -7.74362087e-01
8.96797597e-01 -2.14973688e+00 -1.70671657e-01 -3.11082035e-01
2.92311370e-01 2.80096680e-01 -3.54483664e-01 5.84070563e-01
2.97571093e-01 9.51484516e-02 6.08508997e-02 -2.80594882e-02
-6.91764951e-02 2.84400940e-01 -2.39569739e-01 -9.91660506e-02
6.71963573e-01 1.60054362e+00 -9.85265076e-01 -5.85418224e-01
3.19752574e-01 3.86878133e-01 -4.72272813e-01 1.13293439e-01
-6.52238071e-01 1.02571711e-01 -9.74543169e-02 6.56368554e-01
4.05041277e-01 -6.81709528e-01 -1.98888049e-01 -3.84283721e-01
-4.14694473e-02 1.37181967e-01 -6.02893651e-01 1.34174418e+00
-3.06573689e-01 9.57226753e-01 -2.22320020e-01 -3.36396843e-01
6.96610153e-01 1.06120810e-01 -1.10766828e-01 -1.11830354e+00
1.63100466e-01 -2.58770704e-01 3.04187596e-01 -9.87170517e-01
5.68019748e-01 -2.87925303e-01 4.33575839e-01 4.51341540e-01
3.54162641e-02 -3.45023781e-01 2.69179553e-01 6.25154316e-01
8.79929841e-01 -3.76233943e-02 4.04182434e-01 1.12965435e-01
1.06561504e-01 5.79422653e-01 -3.60911079e-02 1.06257188e+00
-6.55306816e-01 5.90024352e-01 4.36799169e-01 -2.37326488e-01
-1.11940062e+00 -9.75913584e-01 3.70236158e-01 1.39502239e+00
2.23086968e-01 -3.27427447e-01 -5.01878440e-01 -5.15823543e-01
7.36136362e-03 1.30645144e+00 -6.54551208e-01 -2.33229563e-01
-2.30875596e-01 -1.44784421e-01 4.35884833e-01 7.55805612e-01
5.87033987e-01 -1.14371383e+00 -1.27934229e+00 -1.78403389e-02
-3.08550000e-01 -1.30516756e+00 -3.46507847e-01 -1.90802410e-01
-7.10355163e-01 -1.10747123e+00 -8.04076374e-01 -5.33236265e-01
5.39454997e-01 9.72743750e-01 1.23744547e+00 1.33324087e-01
-3.14472497e-01 1.34801805e+00 -6.01102054e-01 -5.29144347e-01
-3.01671177e-01 -3.99770021e-01 -6.31923735e-01 -2.60989308e-01
4.83956516e-01 -2.18411107e-02 -8.34686220e-01 7.11665675e-02
-9.56072092e-01 2.60664105e-01 7.77960896e-01 6.72175467e-01
2.91776180e-01 -6.96426690e-01 5.13039231e-01 -7.21287608e-01
1.06703722e+00 -2.21463099e-01 -1.40390784e-01 6.35134339e-01
-6.62469029e-01 1.34289771e-01 1.20751686e-01 -6.52518511e-01
-1.33517790e+00 -4.34567034e-01 6.29361197e-02 -6.20135963e-01
-1.34150863e-01 2.01705769e-01 8.05288553e-02 -2.55918294e-01
9.64335740e-01 1.29021451e-01 1.19211972e-01 2.43496187e-02
9.85138953e-01 2.96729654e-01 4.01130229e-01 -5.71564972e-01
5.43875515e-01 4.90233064e-01 -2.98136503e-01 -7.87753224e-01
-9.19513524e-01 -3.38839859e-01 -1.29836813e-01 -6.19430423e-01
1.16821563e+00 -8.62533450e-01 -5.54520428e-01 1.27269283e-01
-1.18839633e+00 -4.64691281e-01 -4.31004763e-01 6.97579756e-02
-4.29861575e-01 4.09633338e-01 -3.63797873e-01 -8.68559659e-01
-3.48549277e-01 -1.13051367e+00 8.68931651e-01 4.68219638e-01
-2.24419579e-01 -5.96021116e-01 -2.97826469e-01 1.05138123e+00
4.79304731e-01 -2.22953148e-02 1.25797164e+00 -5.67840219e-01
-9.04998302e-01 2.37831235e-01 -7.13383198e-01 1.80507749e-01
-4.51805666e-02 -1.75563112e-01 -9.47927535e-01 1.44353151e-01
-1.09672263e-01 -6.28918350e-01 8.17296565e-01 1.56088024e-01
8.01393449e-01 -2.39669949e-01 1.68092847e-01 2.20037371e-01
1.27409554e+00 2.75700271e-01 6.07531548e-01 3.58902425e-01
6.60677552e-01 6.19610369e-01 7.06419587e-01 1.81246355e-01
8.97067368e-01 4.15097058e-01 5.26214242e-01 -1.75966825e-02
-7.05649674e-01 -2.27532148e-01 1.29562810e-01 3.09541021e-02
2.34150104e-02 -8.98151100e-02 -1.34246886e+00 6.09193921e-01
-1.62469864e+00 -1.00964987e+00 -6.72727674e-02 1.63043368e+00
6.78643107e-01 2.11130474e-02 9.08306837e-02 -1.16018876e-01
3.08015227e-01 2.61003196e-01 -6.48195922e-01 -4.22508180e-01
-1.47631511e-01 5.28986342e-02 2.01430365e-01 3.82924825e-01
-4.76858526e-01 1.03096640e+00 6.84918165e+00 2.68268913e-01
-1.03271472e+00 1.66011617e-01 4.21227187e-01 -2.59388953e-01
-6.33809865e-01 1.11058272e-01 -4.33473885e-01 7.43475705e-02
4.26996112e-01 1.43248895e-02 4.67163593e-01 5.45624912e-01
1.98322996e-01 -4.22309041e-01 -9.75935698e-01 8.44832599e-01
5.96888661e-01 -1.43630517e+00 4.06004786e-01 -4.38729167e-01
4.77632731e-01 -1.33156940e-01 1.90618783e-01 4.34906393e-01
-3.32020372e-02 -9.69533503e-01 8.91435742e-01 4.53515202e-01
6.00797653e-01 -1.84406161e-01 4.94769990e-01 6.42519370e-02
-6.99771106e-01 -2.53006399e-01 -1.10794362e-02 -3.44612628e-01
3.12979639e-01 -6.99899346e-02 -1.21938264e+00 9.52836424e-02
5.84872544e-01 1.97541192e-01 -1.14799786e+00 8.97429287e-01
-4.08581376e-01 5.16182005e-01 1.37549862e-01 -2.67351002e-01
2.68366158e-01 2.90114731e-01 3.86884451e-01 9.25069392e-01
-9.68863294e-02 5.72193801e-01 -1.06633246e-01 8.13994586e-01
2.65324682e-01 6.61793873e-02 -4.61401910e-01 -3.38439643e-01
5.57953835e-01 8.56503189e-01 -6.97432220e-01 -3.62294465e-01
-9.43393767e-01 7.06015944e-01 4.86633867e-01 8.33921552e-01
-5.19816160e-01 -5.82360327e-02 3.86908919e-01 4.98148620e-01
4.51118022e-01 -4.09229934e-01 -4.68342185e-01 -9.10448790e-01
2.05784533e-02 -1.15593457e+00 4.23828959e-01 -1.63072741e+00
-1.28084087e+00 2.76715487e-01 2.51314253e-01 -7.10933268e-01
-1.53173041e-02 -6.20642900e-01 -5.47742307e-01 6.48082316e-01
-1.57306635e+00 -1.38360643e+00 -6.82921112e-01 7.06637800e-01
8.13064456e-01 1.07524641e-01 3.91707629e-01 -9.53786671e-02
-1.56970724e-01 3.71042997e-01 -6.19815946e-01 -1.12223551e-01
9.32231724e-01 -9.54205751e-01 2.60280520e-01 9.03547645e-01
2.70022690e-01 7.51941025e-01 9.44290578e-01 -7.21291184e-01
-1.45295751e+00 -7.05338001e-01 5.06195962e-01 -6.72371984e-01
4.68647510e-01 -1.80119708e-01 -1.18710542e+00 6.70458376e-01
4.53554899e-01 -1.10619284e-01 6.04261398e-01 1.43124267e-01
-8.11734557e-01 8.04514159e-03 -1.06781864e+00 1.09043574e+00
9.70427454e-01 -7.77669668e-01 -1.17236483e+00 4.39248085e-02
1.00744796e+00 -1.50788883e-02 -4.85358745e-01 2.53344893e-01
5.26440501e-01 -1.05363429e+00 1.12657535e+00 -6.75242126e-01
5.21060586e-01 -2.57452756e-01 -6.06465749e-02 -1.00456512e+00
-4.96323258e-01 -1.02687836e-01 -1.43633917e-01 1.00697207e+00
2.12543443e-01 -4.07094687e-01 4.60811257e-01 9.87759948e-01
8.36099833e-02 -5.98665953e-01 -7.34438181e-01 -4.27170485e-01
-1.87182784e-01 -4.12748039e-01 3.22893023e-01 7.13085115e-01
-1.63782239e-01 6.88928068e-01 -2.22420201e-01 -5.21700419e-02
2.41664529e-01 -1.07451111e-01 9.10297513e-01 -6.71957850e-01
-2.24725172e-01 -4.58882630e-01 -2.86519945e-01 -8.17292690e-01
-1.51740015e-01 -6.07475877e-01 -6.91439137e-02 -2.12322831e+00
2.40411684e-01 1.17626891e-01 -3.10649686e-02 5.10531425e-01
-7.14563549e-01 6.19757064e-02 9.11042631e-01 1.65376812e-01
-7.29710877e-01 4.78198856e-01 1.65193391e+00 -2.09741220e-01
-1.73205137e-01 -5.12625098e-01 -1.03509867e+00 3.81388068e-01
5.75242400e-01 4.57741842e-02 -7.52531409e-01 -7.96829700e-01
3.77414286e-01 4.33349684e-02 7.95341492e-01 -7.36123681e-01
3.59264910e-01 -2.90986329e-01 5.73609293e-01 -5.13120294e-01
5.23008943e-01 -5.71608305e-01 -1.66569918e-01 3.06246251e-01
-3.93468380e-01 3.49576086e-01 6.80021882e-01 6.35192990e-01
-6.68554520e-03 -1.19998902e-01 5.97015500e-01 -3.40289891e-01
-1.20746148e+00 -2.56790936e-01 -3.65005910e-01 4.05833066e-01
1.04346085e+00 -5.21612167e-01 -8.44630420e-01 -6.06178701e-01
-4.92175281e-01 4.42269057e-01 6.70827091e-01 5.89692175e-01
8.40239346e-01 -1.02186573e+00 -5.34145772e-01 1.33573858e-03
6.46414638e-01 -2.70382613e-01 4.83027190e-01 7.83358753e-01
-3.92411441e-01 4.18987811e-01 -4.70909506e-01 -6.10707700e-01
-1.09772515e+00 8.31472039e-01 1.71131074e-01 3.23835135e-01
-6.02772892e-01 8.58657360e-01 3.86038572e-01 2.13427410e-01
2.36248985e-01 -2.81169146e-01 -2.76243597e-01 7.61033222e-02
5.46745777e-01 2.69562542e-01 -3.31149846e-01 -2.84769744e-01
-3.11715752e-01 3.41530174e-01 -2.27316618e-01 -4.36853379e-01
8.02144647e-01 -3.03224534e-01 2.25065961e-01 5.82987845e-01
4.14968312e-01 -1.87643483e-01 -1.42532194e+00 -2.04964310e-01
-1.00631498e-01 -6.16575718e-01 -1.43086970e-01 -1.32101858e+00
-4.99491721e-01 1.04358411e+00 4.61139470e-01 -1.65798530e-01
1.07189918e+00 3.68680686e-01 4.48266387e-01 2.96739548e-01
1.98113665e-01 -7.06493497e-01 7.57259548e-01 3.30121338e-01
9.49976861e-01 -1.46699286e+00 3.00473105e-02 -1.86591730e-01
-1.20613050e+00 8.44577730e-01 1.00261831e+00 1.90277874e-01
5.37701622e-02 -3.19833487e-01 4.62728143e-01 -3.92996311e-01
-1.04143250e+00 -5.91359496e-01 3.90786320e-01 7.24393964e-01
2.68503875e-01 -1.42380223e-01 -1.99814186e-01 4.03006047e-01
-6.70443997e-02 6.46743504e-03 5.47115564e-01 9.46304202e-01
-5.06988227e-01 -5.53907573e-01 -5.05283773e-01 6.12234473e-01
-8.21013004e-02 -2.50189304e-01 -6.53174639e-01 9.30775404e-01
3.84572148e-02 1.11829484e+00 1.34333983e-01 1.05948802e-02
4.34027046e-01 3.47410619e-01 7.97196746e-01 -6.69840276e-01
-7.46521175e-01 -3.75799328e-01 2.48008385e-01 -4.62562948e-01
-3.24872226e-01 -5.52000642e-01 -9.17867601e-01 -9.72947758e-03
-1.97030559e-01 -3.73628497e-01 5.35171270e-01 1.13664341e+00
5.65478086e-01 5.38183808e-01 -4.32560623e-01 -3.64996046e-01
-5.44378102e-01 -8.67365122e-01 1.14811294e-01 6.93098843e-01
5.39344013e-01 -7.78904855e-01 -2.07448140e-01 4.55923080e-02] | [10.753294944763184, 1.7335312366485596] |
e3e89c92-06be-4380-836f-0221b069c0ec | aiai-at-finsbd-task-sentence-boundary | null | null | https://aclanthology.org/W19-5514 | https://aclanthology.org/W19-5514.pdf | aiai at FinSBD task: Sentence Boundary Detection in Noisy Texts From Financial Documents Using Deep Attention Model | null | ['Zi Jun Peng', 'Ke Tian'] | 2019-08-01 | null | null | null | ws-2019-8 | ['deep-attention', 'deep-attention'] | ['computer-vision', '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] | [-7.316308498382568, 3.643272876739502] |
763ed97d-edbb-4d11-b745-c03a2853482e | cell-free-data-power-control-via-scalable | 2212.10299 | null | https://arxiv.org/abs/2212.10299v1 | https://arxiv.org/pdf/2212.10299v1.pdf | Cell-Free Data Power Control Via Scalable Multi-Objective Bayesian Optimisation | Cell-free multi-user multiple input multiple output networks are a promising alternative to classical cellular architectures, since they have the potential to provide uniform service quality and high resource utilisation over the entire coverage area of the network. To realise this potential, previous works have developed radio resource management mechanisms using various optimisation engines. In this work, we consider the problem of overall ergodic spectral efficiency maximisation in the context of uplink-downlink data power control in cell-free networks. To solve this problem in large networks, and to address convergence-time limitations, we apply scalable multi-objective Bayesian optimisation. Furthermore, we discuss how an intersection of multi-fidelity emulation and Bayesian optimisation can improve radio resource management in cell-free networks. | ['Hugo Tullberg', 'Gábor Fodor', 'Sergey S. Tambovskiy'] | 2022-12-20 | null | null | null | null | ['bayesian-optimisation'] | ['methodology'] | [ 4.01594728e-01 3.41598988e-01 -1.65699989e-01 1.58427745e-01
-3.05724770e-01 -2.87479699e-01 4.56067353e-01 -2.14213356e-01
-4.63293314e-01 1.68526256e+00 -9.21109924e-04 -5.10958195e-01
-7.93527126e-01 -9.69488978e-01 -2.58189682e-02 -1.08847582e+00
-3.75039816e-01 5.79807699e-01 -4.70509902e-02 -6.09585121e-02
-2.44909655e-02 9.16936338e-01 -1.01749969e+00 -5.17042160e-01
4.39903885e-01 1.05369532e+00 1.26542836e-01 1.04403913e+00
4.18332547e-01 2.39935845e-01 -7.96087861e-01 -8.04592972e-04
-1.56308599e-02 -4.55498844e-01 -5.70009887e-01 3.33031751e-02
-5.66720665e-01 -4.10405323e-02 -3.39126974e-01 4.02973682e-01
1.25734389e+00 2.20961303e-01 8.12805772e-01 -1.20654833e+00
2.51331806e-01 3.46025646e-01 -2.42666394e-01 4.60008532e-01
3.61284539e-02 1.07364006e-01 8.38896453e-01 -1.38324633e-01
2.27966100e-01 1.14942551e+00 3.98160547e-01 2.92072296e-01
-1.60144401e+00 -5.87559283e-01 -2.71871865e-01 -1.92140877e-01
-1.72457087e+00 -1.00803852e+00 -1.22000560e-01 -4.45386395e-02
1.01527607e+00 5.92985868e-01 6.30700588e-01 5.35836577e-01
6.01702072e-02 1.42061010e-01 7.87490845e-01 -7.58112073e-01
2.73742110e-01 -1.15909472e-01 -7.17680514e-01 3.97308439e-01
4.58880663e-01 1.29255250e-01 -1.62527055e-01 -3.12568814e-01
1.23551714e+00 -7.70144403e-01 -4.89030391e-01 -8.60760361e-02
-1.12919664e+00 5.62336385e-01 1.19957916e-01 4.43902045e-01
-4.37015235e-01 6.20431781e-01 -1.11478530e-01 2.32912436e-01
3.51285309e-01 4.70012635e-01 -2.57331640e-01 -1.13363139e-01
-1.24499476e+00 1.74708128e-01 1.00289440e+00 1.05632412e+00
3.23025584e-01 8.41416046e-02 -5.12091994e-01 9.05247152e-01
9.36540663e-01 6.15529597e-01 -2.75684267e-01 -1.41173100e+00
1.79593951e-01 -1.97166219e-01 6.03673123e-02 -5.64347565e-01
-7.48439372e-01 -1.30544686e+00 -1.21056700e+00 6.07259832e-02
3.93775225e-01 -8.23144853e-01 -4.60121214e-01 1.53759408e+00
1.44607797e-01 2.47974038e-01 3.53739709e-01 7.16459036e-01
1.65043727e-01 6.87475979e-01 -1.66337863e-01 -9.87670720e-01
1.13065445e+00 -6.27193630e-01 -8.87443066e-01 2.93489732e-03
4.80768949e-01 -6.54222369e-01 -3.55449289e-01 3.01530480e-01
-1.50688505e+00 9.86434147e-02 -1.21890020e+00 6.09355927e-01
1.32000586e-02 -2.41837397e-01 3.91532987e-01 1.36098623e+00
-1.22616184e+00 3.49430114e-01 -3.21571410e-01 -6.34305120e-01
5.14971554e-01 6.32733762e-01 3.13622892e-01 -2.25148395e-01
-1.15198326e+00 8.23189437e-01 6.11900449e-01 3.23232599e-02
-4.73342180e-01 -5.04746497e-01 -4.51130360e-01 2.86800295e-01
4.85918224e-01 -1.22911942e+00 1.11058676e+00 -3.98063749e-01
-1.75893521e+00 1.09501988e-01 -2.17853300e-02 -5.79829276e-01
3.00839424e-01 4.86836284e-01 -5.21140516e-01 5.41416034e-02
-3.03884238e-01 3.88937950e-01 5.46783134e-02 -1.09406614e+00
-8.36184025e-01 1.17709450e-01 4.70309071e-02 2.27241412e-01
5.78647219e-02 2.49444798e-01 -5.12948394e-01 -5.13369560e-01
-8.70122015e-02 -7.14361191e-01 -7.11963654e-01 -4.44287837e-01
-4.50108111e-01 3.30889493e-01 3.60383272e-01 -8.84176120e-02
1.36018109e+00 -1.73697901e+00 3.25148821e-01 9.21715975e-01
-1.74390133e-02 1.07713312e-01 1.60156906e-01 3.12248528e-01
3.29452842e-01 2.98634529e-01 5.81152961e-02 -5.29179052e-02
-2.10632935e-01 5.77746212e-01 5.89665830e-01 6.53656960e-01
-6.32380098e-02 5.93348503e-01 -7.72491336e-01 -4.39040095e-01
1.50809363e-01 3.51824343e-01 -4.94383067e-01 -1.93849042e-01
-8.48740265e-02 6.01502240e-01 -3.49393010e-01 6.81933939e-01
4.77290303e-01 -2.36701131e-01 4.76451635e-01 -1.17890788e-02
-1.77530408e-01 -2.98621148e-01 -1.36214483e+00 1.24754131e+00
-6.22257531e-01 6.37990415e-01 5.87072253e-01 -1.14840770e+00
3.12145084e-01 9.02742445e-01 8.62998188e-01 -6.60131454e-01
4.97275412e-01 4.53179032e-01 1.70504302e-01 -4.25036252e-02
2.60779202e-01 -6.37665391e-01 2.57195652e-01 3.27588499e-01
3.33829939e-01 9.82401744e-02 3.52715582e-01 -8.38182122e-02
1.20132363e+00 -2.05629081e-01 4.97720689e-01 -4.87301379e-01
4.55078751e-01 -8.07802677e-01 5.17026961e-01 9.18624580e-01
-2.06264123e-01 3.84070694e-01 4.39516068e-01 4.98672187e-01
-1.16189861e+00 -9.43634450e-01 -4.85497355e-01 6.97883010e-01
1.30558684e-01 1.28596025e-02 -4.95771855e-01 3.06225717e-01
-1.80084348e-01 6.37497783e-01 2.22925588e-01 2.20166475e-01
-1.22358441e-01 -1.55628300e+00 8.21096480e-01 -1.17489703e-01
2.69590229e-01 -3.21896315e-01 -2.60833055e-01 9.49086905e-01
-1.66636929e-01 -1.51740611e+00 7.09060729e-02 3.69951934e-01
-5.58699548e-01 -7.31599331e-01 -1.09637511e+00 -1.38598770e-01
2.19090059e-01 2.90962234e-02 1.00841665e+00 1.10740818e-01
-2.31598243e-01 5.73616028e-01 -2.55386159e-02 -3.60155910e-01
-4.85450476e-01 3.61933440e-01 3.07078868e-01 -1.18926968e-02
-5.03079772e-01 -8.30394685e-01 -5.27516186e-01 7.69212782e-01
-5.22843659e-01 -6.96800202e-02 6.64100289e-01 4.06236917e-01
3.33102107e-01 4.74155337e-01 1.04900908e+00 -2.46377617e-01
6.84295356e-01 -8.80427361e-01 -6.06721461e-01 1.07242972e-01
-4.57678437e-01 1.15860775e-02 2.38767505e-01 9.48521718e-02
-7.57575452e-01 -2.73823023e-01 -2.46411920e-01 3.74765605e-01
7.30233863e-02 5.54755270e-01 -3.38951409e-01 -3.44239950e-01
5.32469094e-01 -3.02251428e-01 2.21584111e-01 1.97491553e-02
1.75121710e-01 7.25718677e-01 6.85684457e-02 -5.86444378e-01
5.39743423e-01 4.24396813e-01 9.24118578e-01 -1.20971370e+00
-4.71469998e-01 -6.53300822e-01 -2.12756708e-01 -7.16973841e-01
5.70089042e-01 -1.01876438e+00 -7.16151536e-01 7.60692582e-02
-9.97875631e-01 -5.12742281e-01 -1.81512713e-01 7.61045516e-01
-1.02435207e+00 2.20177919e-01 -1.12402909e-01 -1.51134598e+00
8.76505375e-02 -9.92211401e-01 5.57191908e-01 4.03559506e-01
-9.07697454e-02 -1.09433413e+00 -1.91428393e-01 4.78916466e-02
8.05805981e-01 4.72143978e-01 5.53562164e-01 -8.44871998e-02
-8.48640800e-01 7.13778511e-02 -4.58429128e-01 5.19130938e-02
-1.69633403e-01 -1.31671354e-01 -9.60308135e-01 -6.31654322e-01
-5.20239890e-01 2.40508616e-01 5.49669981e-01 9.88494694e-01
8.38227630e-01 -2.14786783e-01 -7.67299294e-01 4.23254877e-01
1.65201199e+00 -2.57958490e-02 8.88170958e-01 2.32755080e-01
-6.43852875e-02 3.09812009e-01 4.36141253e-01 6.39510393e-01
7.19099194e-02 1.06724834e+00 7.63453245e-01 -1.44244954e-01
-1.31455958e-01 7.46745884e-01 -5.23510575e-02 3.17847699e-01
-3.66374463e-01 -1.24960101e+00 -5.16750515e-01 3.52066427e-01
-1.87048507e+00 -1.08936751e+00 -3.97940278e-01 2.10902834e+00
3.19593579e-01 4.09369946e-01 3.45262229e-01 3.79014164e-01
7.30186641e-01 -1.97056979e-01 -9.80428327e-03 -3.22687835e-01
-4.25669730e-01 -4.61943895e-02 1.32748950e+00 5.86605549e-01
-6.24266148e-01 8.09218511e-02 6.99587107e+00 1.29360211e+00
-3.22172046e-01 2.34193683e-01 6.70154393e-01 -5.74029446e-01
-9.22413543e-02 6.70746062e-03 -6.88048899e-01 2.83417940e-01
1.58590639e+00 -6.11257106e-02 6.38655603e-01 -2.06421927e-01
9.92628038e-01 -6.59976542e-01 -6.74528778e-01 6.40582263e-01
-4.60881174e-01 -1.49588406e+00 -4.00235504e-01 8.72748971e-01
6.49188519e-01 2.13070326e-02 -5.05268514e-01 -4.59916852e-02
2.03862593e-01 -1.18405592e+00 2.80287802e-01 1.20106924e+00
9.72084284e-01 -9.88826811e-01 9.33749259e-01 4.18118536e-01
-9.92691219e-01 -3.62514585e-01 -1.03293583e-01 -1.48112476e-01
1.05802715e+00 1.08212948e+00 -6.94391549e-01 1.00343525e+00
1.98903918e-01 1.05183713e-01 3.41140747e-01 1.96322823e+00
4.92124230e-01 5.50787330e-01 -8.64951432e-01 -6.59878924e-02
2.92337209e-01 -1.56235933e-01 9.76752043e-01 1.40780210e+00
6.78623974e-01 1.16476931e-01 1.87493768e-02 4.45255101e-01
-3.47672291e-02 2.21805573e-02 -3.41113240e-01 3.26203793e-01
4.86199349e-01 1.22174919e+00 -7.89996326e-01 -7.06185354e-03
-2.57215053e-01 3.14651757e-01 -3.25971425e-01 6.17280722e-01
-6.30229354e-01 -5.34760535e-01 8.88278067e-01 2.09123597e-01
8.04473162e-02 -4.50087816e-01 -3.01368147e-01 -3.20137739e-01
-6.02442861e-01 -3.11522573e-01 8.91927555e-02 -5.78203499e-01
-6.64108396e-01 -2.25971401e-01 3.08742579e-02 -7.83322930e-01
-9.35209394e-02 -4.92661774e-01 -3.31077933e-01 1.14819074e+00
-1.69013810e+00 -7.86804914e-01 1.28488526e-01 -2.08258390e-01
1.09220311e-01 -3.80669862e-01 8.94975543e-01 8.07388663e-01
-6.22769296e-01 3.83358032e-01 7.84575284e-01 -3.92512739e-01
1.79611057e-01 -1.07531238e+00 -2.38488287e-01 7.89590657e-01
-4.20617193e-01 3.22672054e-02 9.83781338e-01 -1.77378058e-01
-9.02414978e-01 -8.80887926e-01 3.42188030e-01 1.97273925e-01
3.44876707e-01 -8.51539448e-02 -2.19370916e-01 -1.06339045e-01
2.17998579e-01 -1.08331941e-01 1.01871753e+00 -4.36395369e-02
8.28506291e-01 1.82098314e-01 -1.23280072e+00 7.77849972e-01
9.61836934e-01 1.68487906e-01 6.35698497e-01 3.82577270e-01
2.62734592e-01 -1.97315231e-01 -1.41430748e+00 4.85734999e-01
7.35224068e-01 -6.73978448e-01 1.03262043e+00 -7.84983635e-02
-5.84788740e-01 -3.48760694e-01 -4.19816434e-01 -1.29525650e+00
-5.97350955e-01 -1.24029732e+00 -4.47619170e-01 1.27055800e+00
5.84536374e-01 -5.13605952e-01 6.60958648e-01 -1.11514971e-01
6.24108873e-02 -8.08375776e-01 -1.49228883e+00 -8.56144369e-01
-1.71310604e-01 -4.61713105e-01 4.71518278e-01 2.44162157e-01
1.43656256e-02 5.84871113e-01 -2.93809325e-01 4.33806837e-01
7.36942828e-01 -8.46877873e-01 5.69420457e-01 -1.40901935e+00
-3.82349849e-01 -9.19355512e-01 -4.66163158e-01 -1.14421475e+00
-3.01221218e-02 -6.82566106e-01 1.61191784e-02 -1.77820969e+00
-1.51955560e-01 -8.47787380e-01 -2.11426333e-01 -1.61291480e-01
2.57597387e-01 5.13215601e-01 -1.40062407e-01 -7.59590566e-02
-6.54229641e-01 2.29018599e-01 1.15459752e+00 3.79902691e-01
1.59635186e-01 6.34858012e-01 -6.77652001e-01 1.45870924e-01
1.00757647e+00 -2.85713911e-01 -2.91251481e-01 1.11148044e-01
4.42822784e-01 2.62080252e-01 3.31095248e-01 -1.30989742e+00
1.26867265e-01 -3.47612619e-01 2.80900151e-01 -2.28521079e-01
5.81157267e-01 -1.03605354e+00 2.76404560e-01 2.48634905e-01
2.20835939e-01 -7.75650442e-01 1.20064676e-01 8.89348686e-01
4.12191182e-01 -5.39710224e-01 1.08769071e+00 6.49665892e-02
-9.92410034e-02 1.97914585e-01 -1.33340895e+00 -1.15189999e-01
1.20760953e+00 -5.52653074e-01 -1.82146877e-01 -6.78603172e-01
-8.81410956e-01 7.14337945e-01 4.69071828e-02 -2.36602932e-01
2.82871611e-02 -1.19775665e+00 -7.40746617e-01 -3.16645533e-01
-2.23100811e-01 -1.13315970e-01 9.87027586e-02 1.12899661e+00
-3.57076854e-01 9.22426999e-01 -3.97708602e-02 -5.44833720e-01
-1.05304098e+00 -1.55192643e-01 1.16043007e+00 -2.65179694e-01
2.35107675e-01 5.09697735e-01 -6.04516268e-01 4.23281360e-03
1.44533381e-01 2.32868224e-01 -3.06570351e-01 2.64259335e-03
1.26240281e-02 8.48075211e-01 4.63112891e-02 -6.53410077e-01
4.81239557e-02 1.93754748e-01 7.18721271e-01 -5.70413888e-01
1.01132905e+00 -8.35146725e-01 -1.62322074e-01 1.61446750e-01
7.85091102e-01 -5.30522019e-02 -9.04562175e-01 -1.10538952e-01
8.43813494e-02 -2.49986202e-01 7.02011526e-01 -7.07386434e-01
-7.26093352e-01 1.87287882e-01 6.31377220e-01 8.32370341e-01
9.90247846e-01 -1.82713643e-01 2.11187243e-01 1.67960376e-01
4.18714315e-01 -1.38068879e+00 -3.92705947e-01 5.46962678e-01
3.62128437e-01 -7.54435718e-01 3.47612441e-01 -4.95585948e-01
3.49804819e-01 1.16289997e+00 1.00200124e-01 5.62181234e-01
7.24056602e-01 2.32174307e-01 -5.56904376e-01 -4.33914661e-02
-7.79244065e-01 -7.91082084e-01 -4.57160659e-02 8.27436626e-01
6.25260293e-01 7.38823116e-02 -3.10064048e-01 -1.93956103e-02
1.95359930e-01 -8.80740061e-02 5.93695462e-01 7.28596091e-01
-8.79230022e-01 -1.33716881e+00 -5.53324461e-01 8.44649971e-01
-7.80539036e-01 7.57919326e-02 3.57795507e-01 4.86216009e-01
-6.30991682e-02 1.35697389e+00 -5.87137640e-02 3.80262375e-01
3.37067127e-01 -2.82429695e-01 4.77589548e-01 -6.00135624e-01
-1.41565977e-02 4.79953527e-01 9.17041600e-01 -9.92219970e-02
-7.58383095e-01 -7.00637996e-01 -1.00668108e+00 -4.03728694e-01
-8.37606430e-01 2.11606339e-01 7.59587586e-01 1.30567288e+00
2.87932098e-01 1.28316808e+00 6.91084802e-01 -8.75079334e-01
-1.53191816e-02 -7.69658029e-01 -8.53401005e-01 -9.07626092e-01
4.10488635e-01 -6.15856767e-01 -1.60182983e-01 -4.35270578e-01] | [6.032399654388428, 1.59402334690094] |
49dc5bce-9d00-4f8c-b234-8a911f3aaddb | scalable-and-compact-3d-action-recognition | 1711.10290 | null | http://arxiv.org/abs/1711.10290v1 | http://arxiv.org/pdf/1711.10290v1.pdf | Scalable and Compact 3D Action Recognition with Approximated RBF Kernel Machines | Despite the recent deep learning (DL) revolution, kernel machines still
remain powerful methods for action recognition. DL has brought the use of large
datasets and this is typically a problem for kernel approaches, which are not
scaling up efficiently due to kernel Gram matrices. Nevertheless, kernel
methods are still attractive and more generally applicable since they can
equally manage different sizes of the datasets, also in cases where DL
techniques show some limitations. This work investigates these issues by
proposing an explicit approximated representation that, together with a linear
model, is an equivalent, yet scalable, implementation of a kernel machine. Our
approximation is directly inspired by the exact feature map that is induced by
an RBF Gaussian kernel but, unlike the latter, it is finite dimensional and
very compact. We justify the soundness of our idea with a theoretical analysis
which proves the unbiasedness of the approximation, and provides a vanishing
bound for its variance, which is shown to decrease much rapidly than in
alternative methods in the literature. In a broad experimental validation, we
assess the superiority of our approximation in terms of 1) ease and speed of
training, 2) compactness of the model, and 3) improvements with respect to the
state-of-the-art performance. | ['Jacopo Cavazza', 'Vittorio Murino', 'Pietro Morerio'] | 2017-11-28 | null | null | null | null | ['3d-human-action-recognition'] | ['computer-vision'] | [-1.62317261e-01 9.33415592e-02 -2.55670130e-01 -9.68973935e-02
-3.07910860e-01 -3.74407947e-01 5.68446457e-01 3.07551414e-01
-6.31447375e-01 8.38534892e-01 7.05605000e-02 -3.87915671e-01
-4.05081600e-01 -7.54393816e-01 -6.08515084e-01 -7.76280701e-01
-1.48044392e-01 3.55308145e-01 4.32349414e-01 1.05972648e-01
6.19899184e-02 7.74514079e-01 -1.54347610e+00 -1.21289119e-01
1.04295826e+00 1.21369374e+00 -2.69612908e-01 4.74593788e-01
2.03322526e-02 1.09252632e+00 -3.23036730e-01 -4.44380313e-01
1.69997722e-01 -2.50370950e-01 -8.74228954e-01 2.43679676e-02
4.38421279e-01 -4.78723496e-01 -4.50537890e-01 8.99128795e-01
1.82397932e-01 2.29361892e-01 8.81534994e-01 -1.24561930e+00
-6.43933773e-01 1.67581126e-01 -2.79386312e-01 -8.03985447e-02
1.13381237e-01 -4.25353259e-01 7.18474090e-01 -5.91142535e-01
3.44077855e-01 8.45037758e-01 1.02751851e+00 2.79252440e-01
-1.35929549e+00 -5.55213913e-02 -2.33331308e-01 3.86425793e-01
-1.50929856e+00 -2.27779225e-01 4.05346960e-01 -4.32628661e-01
8.85389388e-01 2.22287282e-01 6.25249803e-01 8.15827310e-01
1.59805387e-01 8.85863364e-01 1.21268404e+00 -6.04160964e-01
6.17669523e-01 5.36341310e-01 2.73399830e-01 8.23090971e-01
3.91520470e-01 -3.36829461e-02 -7.79684214e-03 -4.09340858e-01
1.00120735e+00 -7.69849261e-03 -2.79535621e-01 -9.76436377e-01
-9.66905296e-01 1.09187996e+00 2.56610066e-01 5.00401616e-01
-3.66009414e-01 2.43891507e-01 6.96281672e-01 2.26449639e-01
5.60618460e-01 5.13593145e-02 -2.49615446e-01 -4.43949103e-01
-9.38738585e-01 3.65460336e-01 1.13486695e+00 6.76606715e-01
4.62600827e-01 1.82922259e-02 9.74559933e-02 6.95456922e-01
-2.01079920e-01 3.04658383e-01 6.31919324e-01 -5.55954874e-01
-4.03471477e-02 6.64505839e-01 2.00593904e-01 -1.02654636e+00
-4.74139899e-01 -2.60525733e-01 -1.01459444e+00 3.93161416e-01
8.90546203e-01 2.17250854e-01 -3.55938315e-01 1.58942187e+00
3.64954084e-01 2.71340579e-01 6.81547523e-02 5.57769537e-01
4.35048528e-02 2.92705476e-01 -1.17507383e-01 -1.57210797e-01
1.17294359e+00 -7.45429575e-01 -5.75473905e-01 2.22096100e-01
9.23507392e-01 -5.19790530e-01 9.71149147e-01 5.36511064e-01
-5.25972188e-01 -4.02978987e-01 -9.01654363e-01 -6.91001192e-02
-6.05465710e-01 5.83970606e-01 1.16283345e+00 8.81304622e-01
-1.06553268e+00 9.21761930e-01 -7.96553373e-01 -6.10272408e-01
3.87028396e-01 3.64973247e-01 -7.27591395e-01 2.12495580e-01
-1.09303808e+00 1.17850137e+00 4.73488837e-01 -8.36526975e-02
-2.38916337e-01 -5.99848688e-01 -7.76225746e-01 9.55747515e-02
2.90261894e-01 -2.97648817e-01 1.06328285e+00 -8.24344993e-01
-1.67542589e+00 6.84905708e-01 1.39283270e-01 -8.92040849e-01
8.28651130e-01 -3.72448742e-01 -2.04179406e-01 4.28357840e-01
-2.91607559e-01 6.10694811e-02 9.65492070e-01 -5.36896884e-01
-4.54888612e-01 -2.66983986e-01 2.48850465e-01 -3.21081400e-01
-5.55924535e-01 -1.33907929e-01 -1.58245601e-02 -5.82045555e-01
-2.38693431e-01 -9.46552217e-01 -3.95867705e-01 2.03701541e-01
-1.11512348e-01 -4.22882646e-01 5.34228623e-01 -5.03836989e-01
1.24987268e+00 -2.20706058e+00 -2.20317133e-02 2.22717956e-01
1.57255411e-01 5.90585053e-01 2.69280970e-01 6.78549767e-01
-3.36283654e-01 -2.67399162e-01 -1.05227120e-01 -1.42638594e-01
2.57589072e-01 1.62080631e-01 -6.06491685e-01 9.57310796e-01
1.52283996e-01 8.28536630e-01 -7.22892642e-01 -3.17268342e-01
4.60476965e-01 5.96093774e-01 -2.64281392e-01 -7.19615119e-03
2.10145846e-01 -1.37319744e-01 -4.34640527e-01 3.93051535e-01
5.92249751e-01 -2.07673863e-01 2.02420056e-02 -3.83310646e-01
-1.90002382e-01 -1.44188941e-01 -1.46052039e+00 1.35640526e+00
-2.45646819e-01 5.85639179e-01 -3.68662089e-01 -1.54850805e+00
7.69604027e-01 2.06242844e-01 4.44025099e-01 -2.92251140e-01
1.33671924e-01 4.44783062e-01 -2.95824498e-01 -3.85684580e-01
2.69026458e-01 -2.92202771e-01 7.88917020e-02 5.39259374e-01
1.35158762e-01 2.37037733e-01 2.13360876e-01 -5.58220688e-03
1.15042663e+00 2.35066667e-01 7.68804133e-01 -5.49815774e-01
6.06257796e-01 -2.52438933e-01 -2.15158835e-02 6.51162624e-01
-2.35544160e-01 1.82692438e-01 7.17083752e-01 -6.44455671e-01
-9.29762602e-01 -9.78198290e-01 -4.58379596e-01 6.55697405e-01
-1.24788441e-01 -4.15160954e-01 -1.03009069e+00 -6.86660528e-01
3.08472782e-01 5.89485645e-01 -9.78153586e-01 -3.41782272e-01
-3.72195423e-01 -7.61126995e-01 9.06175971e-01 8.54922056e-01
3.86373281e-01 -5.88646233e-01 -9.03142154e-01 7.19092935e-02
3.66838634e-01 -1.26607406e+00 -1.43775837e-02 2.70190954e-01
-1.15061605e+00 -1.25764179e+00 -9.03746665e-01 -3.62637460e-01
4.76089507e-01 6.84777275e-02 7.09732771e-01 -2.46527717e-01
-4.96973217e-01 4.70935971e-01 -3.96771371e-01 -1.93494901e-01
-3.58234078e-01 -1.28753543e-01 5.19677937e-01 2.73367673e-01
6.15033805e-01 -8.05954278e-01 -3.83585304e-01 8.26339945e-02
-1.23618233e+00 -3.43024045e-01 7.92913556e-01 7.91138649e-01
2.42596880e-01 2.35168800e-01 4.83362168e-01 -7.62374818e-01
6.66410744e-01 -2.27433220e-01 -7.39715576e-01 2.33377576e-01
-8.10587764e-01 3.19553167e-01 9.23543870e-01 -6.33917570e-01
-6.28801703e-01 2.49245867e-01 6.20906092e-02 -4.94384289e-01
-3.92462015e-01 3.34134847e-01 2.06791922e-01 -4.87152278e-01
7.51198411e-01 4.67286468e-01 3.35415602e-01 -6.75610960e-01
5.56829035e-01 7.18731582e-01 3.43468666e-01 -5.00021935e-01
7.96908915e-01 7.25672305e-01 2.52128750e-01 -1.08052003e+00
-6.64119184e-01 -4.48351353e-01 -9.33577299e-01 -9.14115633e-04
6.82298601e-01 -5.69335938e-01 -8.74560118e-01 4.76420105e-01
-8.47753763e-01 -2.00217530e-01 -5.85566401e-01 7.58115411e-01
-9.56091523e-01 7.80186713e-01 -6.74975991e-01 -1.06352973e+00
-4.86479811e-02 -7.18772292e-01 7.76509166e-01 2.71238182e-02
-1.67968720e-01 -1.20762992e+00 1.62675276e-01 -2.26860289e-02
5.69684446e-01 3.87066692e-01 1.08185017e+00 -8.24064255e-01
-2.79063080e-02 -7.65303612e-01 -4.43828911e-01 9.18487489e-01
1.79393739e-01 -1.99392214e-01 -9.61114168e-01 -1.84711918e-01
9.43475142e-02 -3.67293894e-01 7.52734423e-01 2.10205197e-01
9.81570005e-01 -1.59916729e-01 -2.14117095e-01 3.93283546e-01
1.50828195e+00 -2.96734691e-01 7.85245121e-01 3.33069772e-01
4.08018500e-01 4.58185881e-01 3.76888573e-01 6.05986893e-01
-4.86729182e-02 7.51704156e-01 2.72156239e-01 -1.46047741e-01
1.45503983e-01 -1.38371317e-02 4.68915761e-01 4.56885546e-01
-3.87486905e-01 4.17385936e-01 -5.53395987e-01 4.14491713e-01
-2.10928774e+00 -9.53014195e-01 -2.02261522e-01 2.66101813e+00
6.65814042e-01 -3.94508839e-02 4.84065831e-01 4.39951956e-01
3.82679671e-01 -1.45508170e-01 -3.33847940e-01 -4.38293964e-01
-6.57828003e-02 5.74254632e-01 5.17546952e-01 2.57729262e-01
-1.25631309e+00 6.84791684e-01 6.93640709e+00 1.19134390e+00
-1.00835276e+00 2.49132682e-02 2.24192664e-01 1.01207606e-01
4.04992700e-01 -1.20701984e-01 -5.78242719e-01 4.11219805e-01
1.02291930e+00 -1.46760568e-01 3.06992948e-01 1.31537366e+00
-2.56003514e-02 -4.37557131e-01 -1.26328051e+00 1.12951624e+00
2.11577490e-01 -1.27337623e+00 2.81006675e-02 3.69522989e-01
2.56033540e-01 -3.22504461e-01 -9.40442681e-02 4.77615267e-01
-2.87689537e-01 -1.07766378e+00 5.43663681e-01 4.84792888e-01
4.77883995e-01 -1.04206884e+00 8.04981053e-01 4.36281383e-01
-9.78797257e-01 -5.40631153e-02 -8.98651898e-01 -3.36120099e-01
-1.75538346e-01 8.39775264e-01 -4.57909435e-01 7.10928202e-01
3.07672650e-01 5.04150271e-01 -6.29893422e-01 1.07813585e+00
3.69232520e-02 5.70509553e-01 -4.91534472e-01 -1.38266012e-01
3.82737160e-01 -5.74153304e-01 1.78564787e-01 1.41598976e+00
1.92668855e-01 -2.34103188e-01 -1.68549642e-02 7.50971019e-01
4.31082845e-01 5.40269792e-01 -7.30477214e-01 -2.46702671e-01
-9.10500661e-02 1.15303123e+00 -7.81811833e-01 -3.72610688e-01
-5.44077158e-01 1.38219047e+00 7.26611733e-01 2.74737269e-01
-9.02253151e-01 -5.97574472e-01 6.30381584e-01 1.38099104e-01
6.34381175e-01 -3.09659511e-01 1.52469173e-01 -1.24417257e+00
3.76583248e-01 -6.46723628e-01 1.60669699e-01 -1.71221569e-01
-1.23471737e+00 5.52053094e-01 1.47320479e-01 -1.38535798e+00
-4.22914714e-01 -1.16602647e+00 -2.11394414e-01 6.55805647e-01
-1.24183595e+00 -1.13721895e+00 1.95426151e-01 6.18551910e-01
-4.09767926e-02 -8.45570937e-02 1.35256529e+00 4.30183768e-01
-3.68621141e-01 7.34655201e-01 5.44179499e-01 2.19239011e-01
5.35535991e-01 -1.38922775e+00 -5.98260276e-02 5.19854784e-01
2.44927898e-01 6.61501229e-01 7.11350918e-01 -1.70928732e-01
-1.48159194e+00 -7.81218350e-01 7.16481328e-01 -4.47984308e-01
1.08030987e+00 -4.15310770e-01 -7.59100735e-01 4.96433288e-01
-3.65248710e-01 2.88259476e-01 9.87679303e-01 2.11115703e-01
-5.12879848e-01 -3.33685502e-02 -1.04617572e+00 4.62840319e-01
5.38912594e-01 -5.90240896e-01 -5.85399449e-01 3.95455986e-01
9.28725377e-02 -6.85986057e-02 -1.01560748e+00 2.23597795e-01
8.18902373e-01 -1.37541616e+00 8.67473066e-01 -9.38032627e-01
6.87295645e-02 -1.98582783e-01 -2.43807510e-01 -9.56847906e-01
-2.79210985e-01 -4.52635467e-01 -6.53812706e-01 8.49458814e-01
-3.29964012e-02 -8.56279194e-01 6.96392536e-01 5.65709293e-01
3.25433969e-01 -1.26075923e+00 -1.30038619e+00 -1.42982435e+00
1.63226426e-01 -5.34944832e-01 2.08683610e-01 8.53901565e-01
4.39188421e-01 1.27384454e-01 -5.93312144e-01 -8.30021799e-02
5.46868205e-01 2.13104323e-01 7.70122588e-01 -1.38934326e+00
-5.71440458e-01 -4.65161473e-01 -1.20606422e+00 -1.06559098e+00
1.82075843e-01 -5.90522230e-01 -3.63378495e-01 -1.11282468e+00
7.67954886e-02 -3.00803691e-01 -3.56372684e-01 4.58958328e-01
1.10852264e-01 3.23985130e-01 -9.40353945e-02 6.88705146e-02
-5.33378005e-01 5.62139630e-01 6.77282751e-01 2.98772782e-01
-9.02837515e-02 1.89374372e-01 -4.42510694e-01 9.88428533e-01
6.38788402e-01 -1.73102796e-01 -3.60951275e-01 1.90780118e-01
1.10006467e-01 -3.99269402e-01 4.81022477e-01 -1.39760315e+00
2.34817769e-02 1.40825480e-01 4.56111163e-01 -3.48876864e-02
3.11574101e-01 -8.82500350e-01 -1.48592651e-01 5.31553030e-01
-1.52085036e-01 -2.83446550e-01 1.83483779e-01 6.77506089e-01
-5.58540709e-02 -4.37750041e-01 7.87348986e-01 2.01053962e-01
-5.89210987e-01 1.71908036e-01 -4.06409413e-01 -2.36839265e-01
1.19026256e+00 -4.07892764e-01 8.61861631e-02 -3.63598228e-01
-5.88268757e-01 -5.46967864e-01 5.30557275e-01 4.29790542e-02
4.04821157e-01 -1.47403967e+00 -4.17130411e-01 1.07466780e-01
1.22073023e-02 -4.51949328e-01 2.68759038e-02 1.33539867e+00
-6.34268999e-01 6.23311341e-01 -1.21728949e-01 -3.32733661e-01
-1.05479145e+00 8.64577413e-01 2.13262096e-01 -4.52655107e-01
-9.37610447e-01 4.29575562e-01 1.46067575e-01 -6.89018071e-02
3.21275949e-01 -6.99270427e-01 1.18419835e-02 5.59389107e-02
1.01947308e+00 6.54125810e-01 9.57725421e-02 -4.80765641e-01
-3.54737461e-01 4.47062969e-01 -3.19041088e-02 2.12317288e-01
1.23813462e+00 3.63729715e-01 -2.71095812e-01 6.44648850e-01
1.15708637e+00 4.55857664e-02 -1.07495666e+00 -1.71301171e-01
6.44994304e-02 -4.59589034e-01 -5.72619960e-02 -3.54549438e-01
-6.34104013e-01 7.10725069e-01 6.20550573e-01 5.78254640e-01
9.84262168e-01 -5.16096279e-02 6.33620441e-01 7.14199126e-01
3.69816810e-01 -1.12664711e+00 -2.32214138e-01 4.74433213e-01
6.69817150e-01 -9.88913894e-01 2.21650183e-01 -4.92492765e-01
-2.86637545e-01 1.57411993e+00 7.00135306e-02 -5.13355196e-01
6.95302367e-01 2.86656559e-01 -1.30815446e-01 6.17110170e-02
-3.21498573e-01 -2.95162410e-01 2.05160812e-01 6.66091740e-01
3.48616600e-01 1.24596786e-02 -4.79999781e-01 4.48747367e-01
5.25732227e-02 3.65156859e-01 3.56451690e-01 9.05852139e-01
-4.17222589e-01 -1.09785783e+00 -1.87539756e-01 4.23914552e-01
-3.25008065e-01 2.01189648e-02 -1.36950165e-01 1.07343423e+00
-1.17720254e-01 5.74464083e-01 -2.52510369e-01 -1.59370258e-01
3.62409651e-01 2.53273606e-01 8.29938412e-01 -2.02867702e-01
-6.23355210e-02 -3.59769851e-01 -1.19624868e-01 -7.85222113e-01
-2.52800375e-01 -4.23633993e-01 -1.05450475e+00 -3.28161478e-01
-3.92634362e-01 2.95878559e-01 6.65761709e-01 1.17344880e+00
2.57595181e-01 3.11573483e-02 3.11197937e-01 -7.77680695e-01
-1.23029089e+00 -9.97263253e-01 -1.17175341e+00 4.09989893e-01
8.50015283e-02 -9.60576534e-01 -4.18639392e-01 -8.33831280e-02] | [7.691080093383789, 4.056509017944336] |
a71c307b-8969-4d1d-ba07-eb4b7607d3c9 | sar-image-despeckling-through-convolutional | 1704.00275 | null | http://arxiv.org/abs/1704.00275v2 | http://arxiv.org/pdf/1704.00275v2.pdf | SAR image despeckling through convolutional neural networks | In this paper we investigate the use of discriminative model learning through
Convolutional Neural Networks (CNNs) for SAR image despeckling. The network
uses a residual learning strategy, hence it does not recover the filtered
image, but the speckle component, which is then subtracted from the noisy one.
Training is carried out by considering a large multitemporal SAR image and its
multilook version, in order to approximate a clean image. Experimental results,
both on synthetic and real SAR data, show the method to achieve better
performance with respect to state-of-the-art techniques. | ['L. Verdoliva', 'G. Poggi', 'D. Cozzolino', 'G. Chierchia'] | 2017-04-02 | null | null | null | null | ['sar-image-despeckling'] | ['computer-vision'] | [ 6.60543501e-01 -1.86365351e-01 4.20752555e-01 -1.69095263e-01
-8.19527745e-01 -1.90934375e-01 6.03233457e-01 -3.66672784e-01
-7.98488617e-01 9.86692488e-01 2.93631136e-01 -9.34343040e-02
-3.02953959e-01 -7.56787896e-01 -4.38482076e-01 -1.05880451e+00
-4.12677079e-02 7.64991120e-02 -4.78613488e-02 -2.97077149e-01
-1.74411774e-01 7.52772927e-01 -1.25446236e+00 3.09286505e-01
8.16792965e-01 1.25756967e+00 2.68778890e-01 4.22668427e-01
3.74112874e-01 1.13829827e+00 -4.96571004e-01 3.00755743e-02
6.22094095e-01 -4.52558309e-01 -5.10113597e-01 3.07710737e-01
4.67780888e-01 -4.30845827e-01 -6.65753305e-01 1.38841426e+00
3.86844188e-01 1.70679986e-01 6.01152182e-01 -9.67564285e-02
-6.41545713e-01 3.44984800e-01 -4.09932405e-01 2.80555189e-01
-1.35543421e-01 4.61643226e-02 4.19863522e-01 -1.05346811e+00
6.72749817e-01 8.53199899e-01 8.61426830e-01 1.05116539e-01
-1.41402960e+00 -3.59856427e-01 -6.23611689e-01 1.93576425e-01
-1.42661488e+00 -6.13949895e-01 8.71224523e-01 -3.22301894e-01
5.57814002e-01 3.19871843e-01 4.87754911e-01 8.69123280e-01
4.36669707e-01 5.89313924e-01 1.73278010e+00 -6.09719217e-01
2.28296846e-01 -3.38017404e-01 2.08896995e-01 3.77781510e-01
1.79902121e-01 5.81062794e-01 2.86349617e-02 4.36579511e-02
6.96499050e-01 7.84067810e-02 -6.98045075e-01 -2.84510523e-01
-9.36632872e-01 8.15089762e-01 8.90842974e-01 7.84294665e-01
-8.64201427e-01 -1.76423445e-01 -9.33259130e-02 6.03520989e-01
6.07879877e-01 2.47538865e-01 -2.03964531e-01 6.25871897e-01
-1.63333666e+00 2.06137329e-01 6.16237462e-01 1.81169182e-01
8.86087358e-01 7.32882380e-01 -1.54761046e-01 7.28760540e-01
2.35594139e-02 7.34241128e-01 6.59725726e-01 -4.87090051e-01
1.85617089e-01 3.14001918e-01 2.20092267e-01 -1.02031064e+00
-3.87019128e-01 -9.53475535e-01 -1.55151284e+00 6.09719634e-01
2.39554197e-01 -2.66934603e-01 -1.50154388e+00 1.17686796e+00
-2.89651692e-01 2.94197172e-01 6.29329145e-01 1.02305734e+00
6.85114503e-01 7.05439448e-01 -2.64476657e-01 -3.19873750e-01
8.16817820e-01 -8.14436555e-01 -8.62814963e-01 -6.42880857e-01
1.41319698e-02 -9.69625950e-01 3.46365422e-01 6.81655228e-01
-8.14637482e-01 -8.97544622e-01 -1.28861618e+00 2.38083333e-01
-3.26100171e-01 7.45201349e-01 2.97271460e-01 3.63675416e-01
-1.00936389e+00 8.20556402e-01 -5.66097915e-01 -1.29762590e-01
4.50789481e-01 1.51432097e-01 -5.12909591e-01 -3.34922910e-01
-1.02728760e+00 1.10527825e+00 6.41923070e-01 6.79953814e-01
-9.80374813e-01 -1.81622446e-01 -7.87227631e-01 5.96232042e-02
1.53244838e-01 -3.44456434e-01 8.44945729e-01 -1.61903036e+00
-1.38113415e+00 5.87531924e-01 1.56862289e-01 -9.95261908e-01
3.99645716e-01 -2.76760429e-01 -8.52274120e-01 3.00073296e-01
-2.50344157e-01 -3.51724364e-02 1.44049370e+00 -1.50020337e+00
-1.67454362e-01 -2.07975686e-01 -3.21257383e-01 -2.11566135e-01
2.87356168e-01 -2.51551390e-01 2.69196749e-01 -9.04727697e-01
4.30442810e-01 -7.53757954e-01 -4.70562518e-01 -5.19878030e-01
-2.80605733e-01 7.55841076e-01 9.39034760e-01 -1.19509625e+00
6.99864864e-01 -2.26757479e+00 2.70780861e-01 3.73659998e-01
3.03682173e-03 8.43862057e-01 -3.69380683e-01 4.39148664e-01
-4.38363850e-01 -5.39575458e-01 -8.24758530e-01 -8.69805645e-03
-5.44358194e-01 6.53435960e-02 -4.50629413e-01 7.14627981e-01
1.78972900e-01 7.88851500e-01 -4.27110672e-01 3.62955779e-02
4.17348742e-01 6.15578711e-01 7.84644578e-03 1.91599965e-01
4.38251793e-02 6.66880012e-01 -2.70018160e-01 5.14266431e-01
1.26746821e+00 9.45406407e-02 2.06323922e-01 -7.37274051e-01
-2.50300407e-01 -4.07626390e-01 -1.08661294e+00 1.32948124e+00
-4.22661871e-01 8.01989079e-01 3.57731372e-01 -1.23530233e+00
1.14971840e+00 2.02047080e-01 1.74125671e-01 -9.85113680e-01
3.89513046e-01 4.54384357e-01 -1.26543520e-02 -4.63555723e-01
5.34365177e-01 -3.26076716e-01 3.59918267e-01 1.55173996e-02
8.49086493e-02 -1.48873165e-01 -1.66609716e-02 -1.72705680e-01
7.89564788e-01 -1.17725944e-02 4.95313406e-01 -3.33603770e-01
9.79569316e-01 6.36224747e-02 3.49350661e-01 7.83573329e-01
2.10966870e-01 6.55475080e-01 2.99037546e-02 -8.06950390e-01
-1.02515411e+00 -8.49659503e-01 -1.63955409e-02 6.15250394e-02
-7.40296617e-02 4.12191987e-01 -7.89040685e-01 -4.59414154e-01
-4.17901337e-01 6.29456937e-01 -5.31010211e-01 3.37198116e-02
-8.95148695e-01 -1.11132658e+00 2.81375140e-01 1.09514460e-01
1.09498715e+00 -1.00748599e+00 -7.14397907e-01 5.48233092e-01
-2.68941522e-01 -1.16041672e+00 -5.58189908e-03 3.87475193e-01
-1.01780009e+00 -1.03851736e+00 -8.15606594e-01 -7.15742350e-01
3.90505373e-01 4.28683400e-01 9.18101370e-01 -7.93464035e-02
-2.43795857e-01 -7.24782189e-03 -5.05709887e-01 -6.64776787e-02
-4.86023396e-01 -3.34236801e-01 -3.23945075e-01 5.73187232e-01
1.58385381e-01 -6.87902033e-01 -3.02813828e-01 -1.08356021e-01
-1.52177298e+00 -2.77358949e-01 1.30311275e+00 1.33279085e+00
5.03400266e-01 2.91446418e-01 1.85161397e-01 -6.90927744e-01
3.80073547e-01 -8.65510106e-02 -8.11702013e-01 1.55787691e-01
-3.84595811e-01 8.26014653e-02 8.03769886e-01 -2.61166364e-01
-1.26473129e+00 3.75304133e-01 -2.36723468e-01 -3.84067714e-01
-3.20881218e-01 8.71028781e-01 1.33215457e-01 -5.84144950e-01
7.57936120e-01 8.64942729e-01 6.67097718e-02 -8.77928197e-01
1.12749636e-01 7.39961267e-01 9.43805039e-01 2.45065853e-01
1.26168096e+00 6.84379041e-01 1.39655337e-01 -1.26344168e+00
-8.57337356e-01 -2.66256064e-01 -7.25161850e-01 -1.04325786e-01
8.49108458e-01 -1.14723372e+00 5.45914620e-02 7.65486419e-01
-1.02882171e+00 -3.24593842e-01 -3.35855782e-01 7.78046966e-01
-4.52064067e-01 3.48363101e-01 -6.98441267e-01 -6.02748990e-01
-3.59855145e-01 -9.03898597e-01 7.18024433e-01 3.09250802e-01
6.10823393e-01 -8.04812729e-01 1.47746176e-01 2.20608532e-01
8.01247716e-01 2.20444739e-01 6.10866666e-01 -5.88154912e-01
-7.61946201e-01 -3.28291655e-01 -3.11879575e-01 9.63592410e-01
-1.22045942e-01 -6.29492879e-01 -9.87177372e-01 -5.73791087e-01
4.92260486e-01 -2.67446756e-01 1.39233065e+00 5.13907135e-01
6.71731114e-01 -2.82541007e-01 1.21142104e-01 9.63616729e-01
2.16464663e+00 1.17922731e-01 1.31185663e+00 3.74308974e-01
1.98861450e-01 1.92853436e-01 5.11733770e-01 7.48476461e-02
-4.58985865e-01 5.80463231e-01 6.30988419e-01 -5.48237622e-01
-4.45461929e-01 2.55209029e-01 1.72491297e-01 6.12517774e-01
-2.92285562e-01 -1.76252961e-01 -6.55318081e-01 4.93270099e-01
-1.59824502e+00 -1.05585814e+00 -3.14466923e-01 1.94342589e+00
2.16388807e-01 -1.36678293e-01 -4.69071954e-01 3.18616927e-01
3.27576876e-01 7.31417418e-01 -3.75530124e-03 -3.18113267e-02
-7.19637752e-01 8.11405838e-01 9.22531664e-01 6.18022144e-01
-1.38889325e+00 9.09904897e-01 6.33755589e+00 9.55404937e-01
-1.50141823e+00 1.80447072e-01 5.29551923e-01 7.38661587e-01
1.62924990e-01 -1.21474847e-01 -1.02910742e-01 1.01303291e-02
9.72829938e-01 1.77887857e-01 5.08311570e-01 5.33102930e-01
2.47270450e-01 -3.06320012e-01 -2.44112700e-01 9.24353957e-01
3.98254395e-01 -1.30886424e+00 -8.82603228e-02 -1.06947042e-01
8.38504851e-01 1.55065373e-01 -7.10447729e-02 4.46925648e-02
1.83481649e-01 -1.05683303e+00 7.20623672e-01 1.05903804e+00
6.05587661e-01 -7.39972055e-01 1.42901957e+00 2.49875352e-01
-1.01030624e+00 -1.54471144e-01 -4.76660937e-01 -4.38883193e-02
5.51733002e-02 8.91255856e-01 -3.70247275e-01 1.17660916e+00
5.18832564e-01 7.19615698e-01 -5.77199221e-01 1.00484562e+00
-2.93703616e-01 5.24244130e-01 5.17980419e-02 4.69785601e-01
5.15073419e-01 -5.71794689e-01 7.63741553e-01 1.25474179e+00
5.93926728e-01 3.22701514e-01 2.08924681e-01 7.35223055e-01
1.94236457e-01 -1.08659565e-01 -6.77915215e-01 9.72242206e-02
-2.31599942e-01 1.31170917e+00 -6.15000606e-01 -3.63120168e-01
-1.26922518e-01 1.16164315e+00 -1.95111468e-01 6.74661279e-01
-2.86296934e-01 -3.55037093e-01 2.32702285e-01 -2.46300727e-01
8.84285808e-01 -4.56210166e-01 1.05918720e-01 -1.28490710e+00
-5.14538921e-02 -1.05885994e+00 3.22666913e-02 -9.45701718e-01
-1.14650655e+00 1.16670406e+00 -2.10857823e-01 -1.68156278e+00
-3.24412674e-01 -6.62057340e-01 -4.91077632e-01 1.17741692e+00
-1.84607649e+00 -1.33682775e+00 -4.41755354e-01 5.60708940e-01
4.48911339e-01 -5.34894109e-01 6.99527144e-01 1.74692288e-01
9.76370089e-03 -1.75365880e-01 5.10636926e-01 2.05856234e-01
3.21925521e-01 -7.97977567e-01 2.38336362e-02 1.41681409e+00
4.68361527e-02 1.20558187e-01 1.02334058e+00 -5.00785172e-01
-1.24261427e+00 -1.21346402e+00 8.51729035e-01 4.10465449e-01
4.98880893e-01 1.99614361e-01 -9.30751681e-01 7.21356094e-01
5.04436851e-01 2.16928318e-01 8.75449479e-02 -6.95481241e-01
-2.59557247e-01 -3.91466320e-01 -1.21519208e+00 1.40155837e-01
3.33511621e-01 -3.18855375e-01 -8.78210545e-01 1.37387052e-01
2.53153205e-01 -2.36968488e-01 -5.37899852e-01 6.21836841e-01
2.93971419e-01 -1.32511103e+00 1.17221940e+00 -3.55970897e-02
3.38313997e-01 -4.32089508e-01 -4.22772110e-01 -1.72935438e+00
-4.35081273e-01 -3.16901594e-01 1.37803435e-01 6.20654345e-01
1.62166446e-01 -6.63359106e-01 5.68581045e-01 -5.26591480e-01
-1.24341510e-01 -1.08702086e-01 -9.02615130e-01 -9.66406465e-01
-6.30264282e-02 7.34304786e-02 3.28116983e-01 7.71130264e-01
-8.63281429e-01 2.94257492e-01 -9.58576083e-01 4.13297355e-01
1.03613329e+00 3.59188735e-01 5.48757434e-01 -1.17998242e+00
-2.13326171e-01 8.34384114e-02 -4.93099868e-01 -7.29965389e-01
6.15602285e-02 -6.16941392e-01 2.28102580e-01 -1.34097922e+00
-2.96713203e-01 3.18980813e-02 -1.28745675e-01 2.26138741e-01
3.23823661e-01 8.75101089e-01 2.74409294e-01 2.54317284e-01
-1.07600473e-01 5.75950682e-01 9.79350924e-01 -4.70078051e-01
-9.53796227e-03 2.91916784e-02 -3.10280114e-01 7.65083551e-01
8.82337391e-01 -4.56922412e-01 1.08370550e-01 -4.34313595e-01
-4.69961733e-01 4.08287048e-01 8.31076920e-01 -1.75814939e+00
2.26278812e-01 8.60558450e-02 7.52267361e-01 -7.31315196e-01
4.33493018e-01 -1.16048729e+00 6.07650280e-01 8.81099582e-01
-9.30717736e-02 -2.26789564e-01 1.80332214e-01 5.10231376e-01
-7.88443327e-01 -3.84343058e-01 1.25044227e+00 -2.80871421e-01
-7.18472838e-01 3.62196267e-02 -4.89936590e-01 -5.07026374e-01
5.66702545e-01 -1.73414543e-01 -5.58553077e-02 -4.12706882e-01
-9.89242435e-01 -6.32604778e-01 2.17321694e-01 -2.69339591e-01
7.34239638e-01 -1.10280240e+00 -1.05380166e+00 5.43097734e-01
-1.39146045e-01 -5.47865093e-01 5.79030395e-01 8.88448119e-01
-7.72664011e-01 2.82257617e-01 -4.79886472e-01 -1.85015619e-01
-1.17357767e+00 5.89523375e-01 5.42611182e-01 -5.34855247e-01
-7.21024215e-01 2.32531711e-01 -3.06703568e-01 -1.59904152e-01
-2.80548751e-01 -6.39367402e-02 -4.90289748e-01 -3.68549787e-02
8.63623381e-01 1.67944074e-01 3.09550673e-01 -1.04047382e+00
-7.15964884e-02 7.55330801e-01 2.66772032e-01 -1.66084170e-01
1.82079661e+00 -1.73491687e-02 -3.95876676e-01 -6.45665675e-02
1.20750189e+00 2.37611935e-01 -1.02497184e+00 -6.36785746e-01
6.23890199e-02 -7.21191943e-01 5.27378678e-01 -9.27357376e-01
-1.36918378e+00 5.23632228e-01 1.20026731e+00 2.33119965e-01
1.62313223e+00 -5.91732800e-01 6.29540086e-01 5.70291162e-01
2.62042224e-01 -8.90347898e-01 -1.89292789e-01 6.62079036e-01
1.02935362e+00 -1.17892480e+00 2.80486345e-01 -2.11514249e-01
-2.49181718e-01 1.61551142e+00 -1.45048097e-01 -7.45224535e-01
8.22723746e-01 1.63177401e-01 2.64534861e-01 -2.44392574e-01
-1.64116636e-01 -5.96086442e-01 2.11602688e-01 6.14155173e-01
-1.34784998e-02 1.40569910e-01 -4.19062346e-01 1.47720009e-01
1.56344190e-01 4.28778946e-01 6.83315814e-01 8.15471113e-01
-6.17902875e-01 -1.01148045e+00 -8.10748339e-01 3.68274570e-01
-4.85988319e-01 -2.85096377e-01 -1.26191393e-01 9.85873401e-01
1.64817944e-01 9.11885142e-01 -3.07999343e-01 -3.78759414e-01
3.81767988e-01 -1.94949239e-01 3.08240026e-01 -1.63147986e-01
-6.29217863e-01 2.62684822e-01 8.71589929e-02 -4.53118443e-01
-9.29668665e-01 -4.98237669e-01 -4.29020256e-01 -5.09652402e-03
-1.66857362e-01 2.64246203e-02 6.28154218e-01 1.05314291e+00
-1.73289075e-01 6.14015341e-01 7.43740976e-01 -1.29088545e+00
-6.59299195e-01 -1.11002576e+00 -1.02551627e+00 2.79617518e-01
8.14790249e-01 -1.97233200e-01 -6.14549041e-01 2.73455560e-01] | [10.47329330444336, -2.2197964191436768] |
a33363cf-9db0-4bf7-ad16-a31f00a21012 | landscape-learning-for-neural-network | 2206.09027 | null | https://arxiv.org/abs/2206.09027v1 | https://arxiv.org/pdf/2206.09027v1.pdf | Landscape Learning for Neural Network Inversion | Many machine learning methods operate by inverting a neural network at inference time, which has become a popular technique for solving inverse problems in computer vision, robotics, and graphics. However, these methods often involve gradient descent through a highly non-convex loss landscape, causing the optimization process to be unstable and slow. We introduce a method that learns a loss landscape where gradient descent is efficient, bringing massive improvement and acceleration to the inversion process. We demonstrate this advantage on a number of methods for both generative and discriminative tasks, including GAN inversion, adversarial defense, and 3D human pose reconstruction. | ['Carl Vondrick', 'Hao Wang', 'Purva Tendulkar', 'Chengzhi Mao', 'Ruoshi Liu'] | 2022-06-17 | null | null | null | null | ['adversarial-defense'] | ['adversarial'] | [ 3.43767405e-01 1.50210530e-01 -1.21823363e-02 -4.45295274e-01
-7.78015792e-01 -5.35513163e-01 5.34274101e-01 -7.05929101e-01
-3.40917796e-01 7.98099935e-01 6.54741228e-02 -2.97629476e-01
2.39275441e-01 -7.64770627e-01 -9.68511224e-01 -5.46007276e-01
1.99843004e-01 7.13489354e-01 -3.31592292e-01 -2.55278319e-01
1.21343642e-01 6.43982887e-01 -9.66540098e-01 -1.25084996e-01
6.61986470e-01 8.93489659e-01 -2.47633949e-01 5.43521166e-01
3.60550046e-01 6.82352781e-01 -4.30188328e-01 -6.57573819e-01
4.16678786e-01 -7.17111170e-01 -5.84927857e-01 -1.38135418e-01
6.77245200e-01 -4.01979595e-01 -4.41021442e-01 1.00298834e+00
6.15820229e-01 2.58564383e-01 8.93567145e-01 -1.39632702e+00
-8.16200793e-01 -9.04922001e-03 -7.64244676e-01 -2.91442394e-01
3.95067036e-01 7.75296316e-02 1.00545645e+00 -9.12110209e-01
7.18339384e-01 1.40584707e+00 1.20172942e+00 7.03400552e-01
-1.46047068e+00 -7.04844177e-01 -2.51756251e-01 2.42739022e-02
-1.30175042e+00 -3.13025326e-01 1.02137744e+00 -2.41879120e-01
6.00561976e-01 1.92416310e-01 7.63000309e-01 1.28553486e+00
2.79394776e-01 8.47330093e-01 9.78938997e-01 -2.19891071e-01
2.55945474e-01 -2.22820118e-01 -7.28537440e-01 1.08031785e+00
-1.52324827e-03 2.38668010e-01 -4.05236840e-01 -7.91359171e-02
1.13604414e+00 3.23663875e-02 -1.61919057e-01 -5.65128803e-01
-9.33056891e-01 1.16544092e+00 9.20761883e-01 -3.87963682e-01
-3.71444225e-01 5.23476303e-01 3.16222012e-01 2.73753881e-01
4.72929984e-01 5.09356380e-01 -6.52684644e-02 -1.29743367e-01
-6.12119138e-01 3.80764604e-01 9.73273635e-01 6.23068929e-01
9.02664959e-01 2.42802083e-01 1.21669754e-01 7.86927164e-01
3.04165602e-01 6.55252516e-01 2.37028196e-01 -1.31053340e+00
3.38959694e-01 2.09945768e-01 -4.16281335e-02 -1.26692235e+00
-4.14314717e-02 -3.47378135e-01 -1.26202762e+00 7.65893877e-01
2.84774333e-01 -1.87291503e-01 -1.19326782e+00 2.05138898e+00
3.74169767e-01 2.48551100e-01 -2.96189100e-01 9.69983339e-01
4.93556201e-01 5.20641983e-01 -1.53811023e-01 2.54275888e-01
8.29335034e-01 -1.04952109e+00 -5.80307603e-01 -6.75359249e-01
2.91565508e-01 -5.84828973e-01 1.28084457e+00 3.24515492e-01
-1.13218832e+00 -2.80088305e-01 -1.07153976e+00 -4.62140173e-01
5.14259189e-02 -2.43109375e-01 8.70156109e-01 4.15132850e-01
-9.86806989e-01 7.15330005e-01 -9.51006413e-01 7.20279738e-02
7.02432454e-01 3.70361835e-01 -3.83382529e-01 -2.81836927e-01
-9.11204100e-01 9.66690063e-01 2.40195543e-03 2.81247884e-01
-8.78126442e-01 -6.30924821e-01 -1.01915526e+00 -1.30009502e-01
1.26897618e-01 -1.18134320e+00 9.84815776e-01 -9.25343871e-01
-1.79911411e+00 1.05019581e+00 3.54151018e-02 -3.64769518e-01
9.25873041e-01 -6.39913201e-01 1.03456803e-01 -2.53548801e-01
5.81988767e-02 6.96317673e-01 1.20067847e+00 -1.07381201e+00
2.37526167e-02 -5.44914067e-01 -5.89055158e-02 3.78038228e-01
1.03851527e-01 -4.89997685e-01 -4.18898582e-01 -8.06105554e-01
2.66912073e-01 -1.23621380e+00 -4.52594787e-01 4.49066013e-01
-4.38140988e-01 1.19938426e-01 9.88295019e-01 -8.91043901e-01
5.67277908e-01 -2.17520523e+00 5.95524371e-01 2.44531170e-01
1.72129706e-01 2.98516657e-02 -9.07014534e-02 -3.30400839e-02
1.38840929e-01 -1.01128601e-01 -5.21909773e-01 -5.65974951e-01
1.23590477e-01 4.76648986e-01 -5.01889408e-01 6.69432938e-01
9.87925902e-02 1.31602788e+00 -9.30139542e-01 -2.42606223e-01
2.58880794e-01 8.11337292e-01 -9.65815842e-01 4.05919224e-01
-1.55606106e-01 8.15970778e-01 -4.11382288e-01 5.14243126e-01
4.92516458e-01 -2.87228942e-01 -6.30272031e-02 -2.25172192e-01
4.34683800e-01 1.72802806e-01 -8.78104985e-01 2.12682343e+00
-7.02772617e-01 7.64809132e-01 1.93716571e-01 -1.27274895e+00
7.27823794e-01 -3.21682580e-02 2.31009543e-01 -5.04979491e-01
1.48020998e-01 1.80495605e-01 -4.16499406e-01 -2.64817446e-01
1.11372523e-01 -4.89022404e-01 6.28312826e-02 4.61995721e-01
-1.74777851e-01 -8.67675304e-01 -2.77924985e-01 -6.65233284e-02
9.63096023e-01 5.17515898e-01 -2.81948484e-02 1.28477648e-01
1.85073897e-01 -1.09080054e-01 5.78367174e-01 7.25934446e-01
1.41540676e-01 9.77300048e-01 2.90454537e-01 -7.49319911e-01
-1.23500729e+00 -1.34075820e+00 1.93520918e-01 8.30547810e-01
1.22281261e-01 8.76730829e-02 -7.24787533e-01 -6.45214438e-01
1.95679799e-01 6.34839535e-01 -6.63794994e-01 -5.54629803e-01
-8.58293235e-01 -6.28750861e-01 7.12708294e-01 6.79286778e-01
6.52165592e-01 -9.77127671e-01 -3.41959298e-01 -6.90753236e-02
-2.27099702e-01 -9.47049975e-01 -6.88347578e-01 -1.56780519e-02
-1.08694768e+00 -8.34457278e-01 -6.89184129e-01 -6.96114659e-01
9.23136234e-01 -1.57994017e-01 1.50039673e+00 1.41286803e-02
-5.03866553e-01 2.71722317e-01 1.83826670e-01 -3.13198566e-01
-4.16413963e-01 7.50575513e-02 -1.61986575e-01 -2.26197630e-01
-8.48563090e-02 -8.40161145e-01 -5.71153402e-01 2.52755225e-01
-7.26017892e-01 4.69942689e-02 5.33779919e-01 1.27396858e+00
8.36483896e-01 -3.38964909e-01 2.46870488e-01 -9.51208234e-01
5.93846500e-01 -2.57372230e-01 -6.17452562e-01 1.27720937e-01
-5.03296912e-01 3.70297819e-01 6.20531619e-01 -3.89979631e-01
-9.67599511e-01 1.65712953e-01 -5.74386239e-01 -7.10906208e-01
1.51069269e-01 3.92387867e-01 -1.41775087e-01 -5.11619925e-01
9.37038898e-01 1.00573368e-01 4.04566854e-01 -2.99958080e-01
4.94180024e-01 -2.42135059e-02 7.83776164e-01 -5.73004484e-01
1.03827858e+00 7.11189091e-01 3.94303799e-01 -5.17682076e-01
-1.11927629e+00 1.10710561e-01 -3.11124921e-01 -1.99282262e-02
8.55138898e-01 -8.02415013e-01 -5.93826175e-01 5.33881366e-01
-1.14342129e+00 -4.73516792e-01 -5.71670234e-01 3.89937162e-01
-9.42333162e-01 6.85937628e-02 -5.64032853e-01 -3.12862307e-01
-4.14941490e-01 -1.06885839e+00 1.24247956e+00 4.63806838e-02
-2.17333078e-01 -1.34956181e+00 2.67962098e-01 2.32742891e-01
3.71361881e-01 7.34382451e-01 7.47741461e-01 1.65688749e-02
-6.08222902e-01 -2.99114197e-01 -1.46024615e-01 3.66467118e-01
9.48915258e-02 -3.31565589e-01 -8.03227782e-01 -5.22091448e-01
1.27009213e-01 -7.46935487e-01 8.57343078e-01 3.64811599e-01
1.49840057e+00 -3.69334877e-01 -1.93432942e-01 1.55367017e+00
1.22308671e+00 -1.41510263e-01 7.89015770e-01 1.09009229e-01
9.82494473e-01 1.15348831e-01 1.43227831e-01 1.75711900e-01
1.62243754e-01 5.79560041e-01 6.54951692e-01 -4.47304696e-01
-6.45075217e-02 -6.16130531e-01 1.32583857e-01 5.47959983e-01
-3.52345049e-01 -3.57854329e-02 -8.28573227e-01 1.22114219e-01
-1.71692896e+00 -8.34026575e-01 5.28930724e-01 2.09820914e+00
8.07319105e-01 -3.52267884e-02 -1.36940867e-01 -2.46102512e-02
4.82803792e-01 2.71717817e-01 -9.41765785e-01 -3.34519386e-01
3.08638941e-02 4.81425256e-01 4.37024087e-01 5.05913079e-01
-9.93820369e-01 8.38835776e-01 7.65389776e+00 5.73227644e-01
-1.40142798e+00 8.74606222e-02 7.70448685e-01 6.62187934e-02
-4.73023623e-01 -1.68498561e-01 -2.86995411e-01 2.56250769e-01
3.24931622e-01 -3.01260352e-02 8.88393104e-01 1.00510943e+00
-2.44930536e-01 2.29228497e-01 -1.30147195e+00 1.27066660e+00
2.27401167e-01 -1.46846712e+00 1.25932992e-01 5.15510477e-02
9.11587059e-01 1.12379886e-01 3.52069199e-01 2.16634586e-01
5.21672964e-01 -1.35630774e+00 5.71834028e-01 2.51646727e-01
1.01122391e+00 -8.15232754e-01 2.86535889e-01 2.71581590e-01
-6.44589365e-01 1.77865803e-01 -1.41620234e-01 1.09521141e-02
5.14524758e-01 5.73380947e-01 -6.61938131e-01 5.73646724e-02
6.93159163e-01 7.09280908e-01 -6.42665476e-02 6.69549644e-01
-7.14877307e-01 3.64406854e-01 -5.08634448e-01 1.98164880e-01
1.72162667e-01 -7.17364013e-01 7.11355984e-01 8.46212983e-01
3.10167640e-01 -8.67083147e-02 1.41073450e-01 1.22355175e+00
-4.98281598e-01 -3.66467118e-01 -9.54470575e-01 -1.56369414e-02
1.21218979e-01 9.32476759e-01 -6.24169886e-01 7.34824240e-02
-9.60356072e-02 1.33769500e+00 3.93287271e-01 5.50123334e-01
-1.13584840e+00 -4.32756901e-01 7.70659566e-01 -5.03608957e-02
1.61953852e-01 -6.20641768e-01 -5.62198162e-01 -1.09107983e+00
3.86813097e-02 -6.64371014e-01 7.69756734e-02 -6.94355905e-01
-1.25850415e+00 3.22241068e-01 -2.61259019e-01 -9.67074096e-01
-7.76431322e-01 -6.82657123e-01 -6.92364931e-01 8.04205477e-01
-1.14722848e+00 -1.10918462e+00 -5.06841063e-01 6.71871662e-01
2.72986621e-01 -1.16006352e-01 5.96390605e-01 4.28981364e-01
-1.80376425e-01 7.01242387e-01 1.18494496e-01 9.32138711e-02
5.46962559e-01 -1.21059775e+00 7.76917636e-01 5.57890296e-01
2.92060167e-01 4.27046716e-01 6.82446778e-01 -3.72320116e-01
-1.65522063e+00 -9.81032431e-01 3.15656960e-01 -3.44503880e-01
3.95427734e-01 -4.11585063e-01 -6.71089172e-01 9.56005573e-01
-1.02808252e-01 2.77081549e-01 2.43645862e-01 1.28548101e-01
-4.26063538e-01 -5.91000775e-03 -1.29527247e+00 7.37954915e-01
1.23906672e+00 -8.74934733e-01 -2.88930863e-01 4.94681656e-01
5.16251087e-01 -9.94585156e-01 -4.57448691e-01 4.07044411e-01
7.33346581e-01 -8.66929054e-01 1.48359454e+00 -7.00043440e-01
6.70330167e-01 -8.52195080e-03 5.05314842e-02 -1.51218355e+00
-1.74939319e-01 -8.67685258e-01 -3.05843949e-01 5.44929862e-01
3.12661171e-01 -8.12374473e-01 9.88988340e-01 5.06905496e-01
-3.61860245e-02 -7.73052573e-01 -9.50649738e-01 -7.75195062e-01
1.20815404e-01 -1.67739138e-01 2.96883076e-01 8.36428523e-01
-8.95822287e-01 4.31936711e-01 -7.31422246e-01 -3.17428321e-01
8.07664335e-01 -2.34752763e-02 9.34685707e-01 -9.92062151e-01
-3.30446184e-01 -3.85313332e-01 -4.79249597e-01 -1.29189932e+00
4.40667510e-01 -8.34165215e-01 3.77863407e-01 -1.31432593e+00
-1.54563725e-01 -4.86323774e-01 -8.86213034e-02 3.24228823e-01
-1.08905502e-01 7.12039709e-01 -8.55377987e-02 1.07156865e-01
-1.51853353e-01 8.25505376e-01 1.53510761e+00 -2.75395900e-01
3.05254915e-04 2.13843778e-01 -5.21854937e-01 9.55276251e-01
5.24359882e-01 -4.66719925e-01 -4.78125095e-01 -8.70695889e-01
5.09413719e-01 -1.41082957e-01 6.84534013e-01 -7.79094636e-01
1.16147079e-01 -1.62088022e-01 5.55279136e-01 -1.33221865e-01
6.40704930e-01 -6.05447352e-01 4.25371915e-01 6.03007913e-01
-2.24680141e-01 2.85979331e-01 4.31792289e-02 5.09413719e-01
-2.87873566e-01 1.72167271e-02 1.02934670e+00 -2.44676918e-01
-4.77024049e-01 6.46922886e-01 3.67930949e-01 5.71229994e-01
6.51542366e-01 -9.87331644e-02 1.92498356e-01 -7.77642906e-01
-6.32133007e-01 -1.37912422e-01 6.17961466e-01 2.71352112e-01
8.52979541e-01 -1.61566722e+00 -7.50703454e-01 5.37259936e-01
-4.78932679e-01 3.38423103e-01 -1.19837366e-01 7.50563145e-01
-9.60303664e-01 -2.26016536e-01 -3.60687822e-01 -7.63616621e-01
-8.23541760e-01 3.29987794e-01 4.82869446e-01 -1.12962499e-01
-8.19819510e-01 1.17054021e+00 3.49000692e-01 -6.52554154e-01
1.94401905e-01 2.49827325e-01 4.33000445e-01 -4.22170192e-01
1.44166112e-01 4.46422875e-01 -1.62065029e-01 -4.41588461e-01
-1.91612825e-01 6.92850947e-01 2.05920041e-01 -2.43909493e-01
1.31229663e+00 1.12969361e-01 -2.55978912e-01 1.10678248e-01
1.36323345e+00 -1.38429523e-01 -1.52902615e+00 -6.49425983e-02
-3.85776907e-01 -6.67860448e-01 8.71012881e-02 -4.27522093e-01
-1.40969658e+00 9.45705116e-01 3.66284817e-01 -9.81319621e-02
1.05670834e+00 -1.94534764e-01 9.84629691e-01 6.63347006e-01
3.45102340e-01 -9.77700233e-01 3.39020491e-01 7.49721229e-01
1.22513032e+00 -1.36383069e+00 2.73166239e-01 -2.43779629e-01
-3.40057611e-01 7.77041256e-01 5.54159462e-01 -6.06712401e-01
7.86040902e-01 3.60390931e-01 5.46825007e-02 2.70616706e-03
-1.79904208e-01 3.75362456e-01 4.11192745e-01 5.67512512e-01
2.31407970e-01 -5.34308329e-02 -3.06233726e-02 -2.13428438e-01
-3.55090261e-01 -2.58790031e-02 -9.74147767e-02 9.47234333e-01
1.17918015e-01 -1.05910873e+00 -1.45095870e-01 2.49414176e-01
-3.59345853e-01 3.96776311e-02 -2.88414121e-01 7.67174065e-01
-1.29117951e-01 1.73838213e-01 8.53820071e-02 -2.10016906e-01
2.14493945e-01 -2.79184788e-01 8.26158166e-01 -3.24563086e-01
-2.11545989e-01 -3.04442018e-01 -1.72942311e-01 -8.52786303e-01
-2.79243857e-01 -4.24624652e-01 -1.09150422e+00 -4.05572176e-01
-1.51615471e-01 -3.41849290e-02 8.10139596e-01 9.28683758e-01
2.92804539e-01 4.08489823e-01 6.97385788e-01 -1.06805384e+00
-7.74842680e-01 -5.18266559e-01 -3.33648533e-01 7.57025778e-01
4.33974564e-01 -8.34649861e-01 -4.64936018e-01 2.40298975e-02] | [11.583789825439453, -0.541548490524292] |
9bf5285e-cb26-4104-ba2c-aebf185e0da9 | text-spotting-transformers | 2204.01918 | null | https://arxiv.org/abs/2204.01918v1 | https://arxiv.org/pdf/2204.01918v1.pdf | Text Spotting Transformers | In this paper, we present TExt Spotting TRansformers (TESTR), a generic end-to-end text spotting framework using Transformers for text detection and recognition in the wild. TESTR builds upon a single encoder and dual decoders for the joint text-box control point regression and character recognition. Other than most existing literature, our method is free from Region-of-Interest operations and heuristics-driven post-processing procedures; TESTR is particularly effective when dealing with curved text-boxes where special cares are needed for the adaptation of the traditional bounding-box representations. We show our canonical representation of control points suitable for text instances in both Bezier curve and polygon annotations. In addition, we design a bounding-box guided polygon detection (box-to-polygon) process. Experiments on curved and arbitrarily shaped datasets demonstrate state-of-the-art performances of the proposed TESTR algorithm. | ['Zhuowen Tu', 'Subarna Tripathi', 'Yongwen Su', 'Xiang Zhang'] | 2022-04-05 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Zhang_Text_Spotting_Transformers_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Zhang_Text_Spotting_Transformers_CVPR_2022_paper.pdf | cvpr-2022-1 | ['text-spotting'] | ['computer-vision'] | [ 5.79849303e-01 -2.88807541e-01 7.47542381e-02 -3.46063614e-01
-1.17223752e+00 -6.97788537e-01 5.89355588e-01 3.05182248e-01
-3.16393077e-01 2.31636375e-01 -2.20378816e-01 -6.50958359e-01
1.07473634e-01 -7.34222889e-01 -7.12040961e-01 -3.56817603e-01
1.50955528e-01 8.80564928e-01 4.74901825e-01 -2.25457788e-01
7.08523750e-01 5.48740983e-01 -1.29803336e+00 6.11444473e-01
8.39721024e-01 8.03735197e-01 2.28374258e-01 1.19548917e+00
-2.48845145e-01 3.04977596e-01 -6.76664114e-01 -7.32442558e-01
3.26034456e-01 -2.28304490e-02 -2.18367279e-01 3.51147056e-01
7.32438982e-01 -4.08944309e-01 -2.90562272e-01 7.51364708e-01
7.05826581e-01 -7.86528736e-02 7.08530426e-01 -6.65297151e-01
-4.23096955e-01 4.78058517e-01 -8.34211111e-01 -1.66919246e-01
5.04780233e-01 6.20283596e-02 1.17501044e+00 -1.44398642e+00
4.66143370e-01 1.18757260e+00 8.31570625e-01 2.47939080e-01
-1.01353157e+00 -2.50458062e-01 9.07534435e-02 -1.13777317e-01
-1.47414947e+00 -4.40951824e-01 5.09067535e-01 -4.40798998e-01
1.23501003e+00 6.77890480e-01 2.65842944e-01 8.61751258e-01
-1.09521728e-02 1.24382222e+00 5.22064626e-01 -7.07901478e-01
1.21292092e-01 -7.85936117e-02 5.94042689e-02 9.15777981e-01
1.44889221e-01 -3.47828478e-01 -4.45583820e-01 4.10231054e-02
8.47543836e-01 -1.49138138e-01 -2.86362112e-01 -3.69517624e-01
-1.28707981e+00 5.51626801e-01 -2.42076889e-01 -6.56820163e-02
1.25787333e-01 2.10807785e-01 8.21797550e-01 2.78485030e-01
6.06374800e-01 1.33737372e-02 -3.72533113e-01 -4.35450405e-01
-1.55213141e+00 1.80812940e-01 7.56195009e-01 1.55060077e+00
3.40643555e-01 3.65948193e-02 -4.88290668e-01 1.12306273e+00
3.42474669e-01 6.36884749e-01 3.34568322e-01 1.55834213e-01
1.20175374e+00 7.33340263e-01 -6.89250156e-02 -6.78371668e-01
-1.01323396e-01 -1.80212766e-01 -6.65266752e-01 1.79161444e-01
4.65725750e-01 2.01038010e-02 -1.10556149e+00 6.52280986e-01
1.61998183e-01 -1.08722486e-01 -1.02936909e-01 8.61952305e-01
5.83303094e-01 7.10364163e-01 -5.40490329e-01 3.75005096e-01
1.55461049e+00 -9.29832935e-01 -4.14798081e-01 -2.75068492e-01
9.37341988e-01 -9.81516957e-01 1.30281305e+00 6.00837350e-01
-9.97125566e-01 -1.81799382e-01 -1.13404679e+00 -6.46259785e-01
-4.75672990e-01 8.89614940e-01 9.88088027e-02 9.33934391e-01
-7.27672458e-01 4.39075828e-01 -6.35451138e-01 -3.26584727e-01
3.55908871e-01 5.75841010e-01 -1.41957313e-01 2.59641111e-01
-5.67103326e-01 6.94504559e-01 2.22824588e-02 2.78669834e-01
-6.39598429e-01 -3.11486721e-01 -7.61511445e-01 1.23306319e-01
6.92018449e-01 -2.40156561e-01 1.29343081e+00 -4.64044541e-01
-1.75042129e+00 1.00223935e+00 -3.83487612e-01 -3.66385907e-01
1.16701746e+00 -5.86331964e-01 -2.39370197e-01 7.98040926e-02
-1.39488906e-01 1.01793915e-01 1.33644724e+00 -8.56307149e-01
-6.43869638e-01 -3.88438165e-01 -5.13425887e-01 2.33206570e-01
-1.99983343e-01 3.42177600e-01 -1.10589015e+00 -1.03353381e+00
1.06995106e-01 -5.45994043e-01 1.11794516e-01 3.64302665e-01
-9.41162467e-01 -1.11149125e-01 1.22057545e+00 -8.16412866e-01
1.46880305e+00 -2.03542304e+00 -9.98046547e-02 4.67511952e-01
1.12572886e-01 2.05789119e-01 1.08288601e-01 4.85179722e-01
2.41948478e-03 2.02814847e-01 -3.00415069e-01 -6.99818671e-01
3.53704244e-01 -1.88180938e-01 -5.64809740e-01 7.93701768e-01
3.39026779e-01 5.47234297e-01 -3.98883134e-01 -5.97758055e-01
5.42246938e-01 3.55062008e-01 -4.41657931e-01 3.09180245e-02
-5.31107008e-01 -2.16913626e-01 -4.71896738e-01 1.05945945e+00
7.93528676e-01 1.28637984e-01 8.04592390e-03 7.89567642e-03
-4.49472070e-01 3.19295406e-01 -1.31335330e+00 1.30742955e+00
-3.58553916e-01 1.01090622e+00 -1.80170648e-02 -6.34916186e-01
1.33094716e+00 9.91669744e-02 6.53471984e-03 -4.81845677e-01
1.12368710e-01 4.82862681e-01 -4.51307297e-01 -2.57045865e-01
1.10330534e+00 5.07100105e-01 -1.77846149e-01 1.83835700e-01
-4.01526153e-01 -1.85027257e-01 2.50844389e-01 -4.90515865e-02
1.00639188e+00 4.57632691e-01 3.19330364e-01 -9.89535600e-02
6.97586417e-01 -3.08536720e-02 1.43111780e-01 7.17027247e-01
1.32463858e-01 9.48887706e-01 7.22491801e-01 -2.79545128e-01
-1.45469105e+00 -7.13876665e-01 -3.88852924e-01 1.31768310e+00
-1.61628544e-01 -7.01613367e-01 -6.97250783e-01 -5.53891778e-01
-1.16471881e-02 7.44376183e-01 -4.72212911e-01 5.10349989e-01
-9.39010680e-01 -4.32707578e-01 1.01971447e+00 6.34980679e-01
3.90114009e-01 -6.40486002e-01 -7.43717372e-01 1.46667108e-01
3.16695243e-01 -1.17424154e+00 -1.00487328e+00 5.22561729e-01
-9.22008932e-01 -9.38794434e-01 -1.12340868e+00 -8.76444101e-01
8.26687276e-01 5.57404645e-02 7.93080151e-01 -7.34951869e-02
-3.60863358e-01 1.01275176e-01 -5.18981040e-01 -1.27628624e-01
-3.93875510e-01 -3.10707241e-02 -5.81467569e-01 1.45836398e-01
2.47285753e-01 1.18583411e-01 -3.37248623e-01 7.09014416e-01
-6.68752372e-01 2.71086603e-01 4.45200861e-01 8.70049715e-01
6.83353126e-01 -2.12876156e-01 -1.06938608e-01 -8.10516417e-01
5.93654275e-01 1.42668515e-01 -9.59136605e-01 4.89532232e-01
-3.65796119e-01 2.89558135e-02 8.25156987e-01 -2.53592223e-01
-6.75379038e-01 2.47660220e-01 -5.22705652e-02 -3.48140419e-01
-4.75129522e-02 2.33692676e-01 -2.40103170e-01 -3.19847763e-02
5.11988342e-01 5.30699313e-01 -3.25972259e-01 -6.50953770e-01
3.92637491e-01 1.08448982e+00 4.99851257e-01 -6.17949843e-01
7.27202892e-01 3.31373632e-01 -1.63940966e-01 -1.26597118e+00
1.52958920e-02 -6.48897171e-01 -8.89247060e-01 -1.61360204e-01
6.34211779e-01 -6.51525021e-01 -6.88970864e-01 5.55692732e-01
-1.22203684e+00 -4.11474675e-01 -3.73945720e-02 -1.45330891e-01
-7.30007172e-01 7.98713624e-01 -3.90801489e-01 -1.03124034e+00
-6.92459464e-01 -1.43393624e+00 1.92020524e+00 -1.69692770e-01
2.01994344e-03 -7.20792174e-01 -1.56053871e-01 1.07876472e-01
-1.90505590e-02 -4.18395102e-02 8.84396672e-01 -7.96785116e-01
-5.98629713e-01 -4.91228938e-01 -3.99444580e-01 -9.34142843e-02
-4.11774755e-01 4.45273221e-01 -8.89874756e-01 -2.38274708e-01
-5.89560032e-01 -3.66259515e-02 8.82478118e-01 1.03648141e-01
1.35595536e+00 -2.37416029e-01 -4.18387860e-01 9.01002407e-01
1.41541898e+00 1.61619902e-01 8.41028571e-01 4.47296113e-01
8.89599800e-01 3.05667490e-01 5.64754903e-01 9.45964634e-01
1.48045748e-01 1.09091437e+00 1.45894825e-01 -9.70193446e-02
-1.67404916e-02 -3.58791262e-01 6.36843264e-01 4.58635598e-01
2.47737035e-01 -7.60107160e-01 -1.10280073e+00 3.82750869e-01
-1.77985990e+00 -6.67458594e-01 -6.55961514e-01 2.45959496e+00
6.46502316e-01 2.03792796e-01 1.86880305e-01 5.12598157e-01
1.05184746e+00 3.62455137e-02 -4.55752611e-01 -6.75311625e-01
-2.04664320e-01 2.79119223e-01 8.36155713e-01 4.01916176e-01
-1.32773793e+00 1.26418960e+00 5.82766151e+00 1.16602814e+00
-9.61436093e-01 -3.67713004e-01 4.96081233e-01 6.04315698e-02
-4.66526672e-02 -5.49573340e-02 -1.13478374e+00 3.15084726e-01
5.17674506e-01 2.77680308e-01 2.58182347e-01 8.38191450e-01
3.01812768e-01 -2.22407212e-03 -1.29359341e+00 1.14031422e+00
1.38108701e-01 -1.08495998e+00 4.91442438e-03 -1.70056850e-01
6.85497224e-02 -1.31744206e-01 6.83719218e-02 1.64100006e-01
-8.95980448e-02 -9.01820600e-01 1.13689029e+00 2.07631096e-01
1.56545365e+00 -5.64205229e-01 2.20564961e-01 8.89463201e-02
-1.35609424e+00 1.75238311e-01 -3.66629004e-01 4.71827686e-01
2.39812508e-02 6.35633409e-01 -1.31379938e+00 2.47372910e-01
2.69759238e-01 5.93725920e-01 -7.94385076e-01 1.23584628e+00
-1.17895097e-01 6.11505926e-01 -5.70911527e-01 -5.12889683e-01
1.08456805e-01 -1.58747196e-01 7.09287405e-01 1.99366772e+00
4.15671349e-01 -4.89838481e-01 2.58706510e-02 7.91964531e-01
-4.48591709e-02 4.96248841e-01 -5.09103596e-01 4.05530706e-02
3.34314615e-01 9.18149829e-01 -1.13178563e+00 -4.58481163e-01
-3.90793145e-01 1.40817380e+00 9.25541893e-02 1.44026443e-01
-9.43335712e-01 -7.97791600e-01 5.64418994e-02 1.96288601e-01
9.68806863e-01 -4.09152329e-01 -8.50111008e-01 -1.31540704e+00
3.43191773e-01 -1.01307869e+00 1.70487374e-01 -7.22742796e-01
-8.82621408e-01 3.55443656e-01 -2.56534904e-01 -1.31155324e+00
-1.91991180e-02 -1.05543494e+00 -6.43061459e-01 7.73213029e-01
-1.28481567e+00 -1.16699719e+00 -4.52710502e-02 5.18880665e-01
9.60442603e-01 -1.41200051e-01 7.69624710e-01 1.98088050e-01
-8.61341476e-01 1.16958487e+00 7.15351403e-01 5.24976790e-01
7.10889161e-01 -1.49415267e+00 1.18568265e+00 1.04347289e+00
1.04476117e-01 4.19025570e-01 5.91841817e-01 -8.83986652e-01
-1.90943813e+00 -1.04919887e+00 6.61851704e-01 -2.15549901e-01
5.40056407e-01 -1.12491632e+00 -7.09393084e-01 4.99015182e-01
-1.39943898e-01 -2.07204178e-01 2.86651880e-01 8.67903382e-02
-5.29573798e-01 7.69643262e-02 -1.01040006e+00 9.08935368e-01
6.35664284e-01 -5.79779804e-01 -4.47868258e-01 4.55494195e-01
2.12468594e-01 -8.19672942e-01 -4.91645306e-01 9.31746513e-02
7.10312128e-01 -8.11841309e-01 7.59437263e-01 -9.02597383e-02
4.13957685e-01 -4.78992492e-01 -1.85415223e-01 -7.49557674e-01
1.54493034e-01 -1.13737369e+00 -2.12754384e-01 1.01780355e+00
5.09527981e-01 -2.83888966e-01 9.83408689e-01 2.51468837e-01
-5.73556781e-01 -7.57477283e-01 -1.06847215e+00 -5.25601506e-01
-1.35571599e-01 -8.75462115e-01 2.90580302e-01 4.75515217e-01
1.30403116e-01 1.09451689e-01 -3.80535603e-01 1.22999094e-01
3.93068314e-01 1.22354850e-01 9.34608340e-01 -7.93248594e-01
-5.23370564e-01 -7.73478746e-01 -5.20639122e-01 -1.63946772e+00
-4.21213359e-01 -8.69292319e-01 2.41825327e-01 -1.29001391e+00
-1.89160213e-01 -2.65755236e-01 5.65437496e-01 5.50690651e-01
1.08055854e-02 3.86875197e-02 2.73751944e-01 5.72521500e-02
-5.60438693e-01 4.07269418e-01 1.05181539e+00 -3.06388617e-01
-1.79952949e-01 5.08228764e-02 -1.18868336e-01 5.94469845e-01
6.25076592e-01 -2.58672923e-01 -6.34501800e-02 -5.81938326e-01
3.06361228e-01 7.77891725e-02 2.03249410e-01 -8.68163407e-01
3.51602852e-01 2.85152406e-01 2.63608992e-01 -1.29149961e+00
2.38604680e-01 -8.03128183e-01 -6.16405785e-01 9.50007793e-03
-4.06200588e-01 1.96387634e-01 2.96088517e-01 4.90475714e-01
2.42753342e-01 -5.43368459e-01 7.65236497e-01 3.95372808e-01
-4.42627072e-01 -2.96116471e-02 -6.05876327e-01 -1.01961698e-02
9.88819540e-01 -8.43603015e-01 -3.79759043e-01 -1.45525619e-01
-4.45271313e-01 1.30595237e-01 4.42160606e-01 2.15576708e-01
1.03835130e+00 -7.34732628e-01 -7.67411530e-01 4.55244303e-01
2.46676713e-01 1.84562102e-01 -1.88731015e-01 6.04880750e-01
-1.12544644e+00 4.96899128e-01 3.78757834e-01 -6.44686103e-01
-1.53885734e+00 3.78766239e-01 3.86268228e-01 -2.75964081e-01
-1.12014997e+00 7.13974059e-01 -9.20270681e-02 -1.55788168e-01
6.23001337e-01 -8.78669262e-01 1.99230835e-01 -2.16760069e-01
4.81461912e-01 5.11258900e-01 3.86661708e-01 -3.61799240e-01
-2.78766036e-01 8.21989238e-01 -4.07492220e-01 -2.52687216e-01
1.02924049e+00 1.89945862e-01 1.43010557e-01 3.17098558e-01
8.98191869e-01 2.66774327e-01 -1.18116665e+00 6.86897635e-02
2.11396262e-01 -4.30898190e-01 8.69932026e-02 -9.74263251e-01
-5.18068016e-01 1.12397110e+00 4.16760713e-01 4.34061959e-02
9.07315016e-01 -5.44472873e-01 6.87203646e-01 6.74382210e-01
1.33260235e-01 -1.23550642e+00 -2.69467622e-01 9.39295888e-01
9.49549913e-01 -7.88945913e-01 1.81725845e-01 -6.16401851e-01
-4.40408260e-01 1.70774567e+00 4.10335034e-01 -1.87178746e-01
2.92288661e-01 7.86144793e-01 -3.31923008e-01 -4.23706658e-02
-7.40438521e-01 -8.64678845e-02 3.61348301e-01 3.56161267e-01
6.84081018e-01 1.83854252e-02 -1.23424716e-01 3.00424784e-01
-1.12918936e-01 -3.15180719e-01 5.72381377e-01 8.79411042e-01
-5.42230725e-01 -1.11765730e+00 -7.68735468e-01 5.54790258e-01
-3.68478119e-01 -4.52429712e-01 -5.58386326e-01 8.40377748e-01
-2.75578141e-01 5.42484462e-01 7.94441923e-02 -4.45995301e-01
4.72545594e-01 6.87273964e-02 4.24139708e-01 -6.34779572e-01
-7.90104449e-01 5.63497901e-01 2.24881470e-01 -2.24340111e-01
4.59433168e-01 -7.44998813e-01 -1.34818792e+00 -1.72645375e-01
-6.90754235e-01 -3.76015753e-01 8.23815525e-01 6.60131395e-01
4.25199747e-01 3.36230040e-01 4.38722551e-01 -7.13373244e-01
-7.44781971e-01 -1.00892103e+00 -5.02872407e-01 7.53948912e-02
2.70818621e-01 -2.86237925e-01 -1.36546616e-03 1.43253341e-01] | [11.959741592407227, 2.205212116241455] |
2b2e065a-d7a7-47a7-83d7-7d65325cbefc | monkey-business-reinforcement-learning-meets | 2202.13706 | null | https://arxiv.org/abs/2202.13706v1 | https://arxiv.org/pdf/2202.13706v1.pdf | Monkey Business: Reinforcement learning meets neighborhood search for Virtual Network Embedding | In this article, we consider the Virtual Network Embedding (VNE) problem for 5G networks slicing. This problem requires to allocate multiple Virtual Networks (VN) on a substrate virtualized physical network while maximizing among others, resource utilization, maximum number of placed VNs and network operator's benefit. We solve the online version of the problem where slices arrive over time. Inspired by the Nested Rollout Policy Adaptation (NRPA) algorithm, a variant of the well known Monte Carlo Tree Search (MCTS) that learns how to perform good simulations over time, we propose a new algorithm that we call Neighborhood Enhanced Policy Adaptation (NEPA). The key feature of our algorithm is to observe NRPA cannot exploit knowledge acquired in one branch of the state tree for another one which starts differently. NEPA learns by combining NRPA with Neighbordhood Search in a frugal manner which improves only promising solutions while keeping the running time low. We call this technique a monkey business because it comes down to jumping from one interesting branch to the other, similar to how monkeys jump from tree to tree instead of going down everytime. NEPA achieves better results in terms of acceptance ratio and revenue-to-cost ratio compared to other state-of-the-art algorithms, both on real and synthetic topologies. | ['Badii Jouaber', 'Hind Castel', 'Andrea Araldo', 'Massinissa Ait Aba', 'Maxime Elkael'] | 2022-02-28 | null | null | null | null | ['network-embedding'] | ['methodology'] | [-2.58445501e-01 3.38958740e-01 -7.05215812e-01 5.42113781e-02
-2.28137244e-02 -7.45696783e-01 8.77118111e-02 -2.62234539e-01
-2.76874989e-01 1.38672352e+00 -3.52600724e-01 -8.02036643e-01
-4.63591427e-01 -1.04371178e+00 -5.61501801e-01 -6.61018133e-01
-5.98361671e-01 1.03541791e+00 4.90838289e-01 -9.22021121e-02
-4.90792468e-03 8.62312317e-01 -7.93715119e-01 -2.38373235e-01
6.84293330e-01 9.78611887e-01 1.06332563e-01 4.85110074e-01
-2.86793441e-01 6.28049135e-01 -6.19474471e-01 -5.32057405e-01
5.06992161e-01 -6.40484273e-01 -1.01353872e+00 1.24879830e-01
-3.35062355e-01 -2.97635406e-01 -1.77420214e-01 9.02080476e-01
3.65251213e-01 -1.11079127e-01 1.85414299e-01 -1.74318957e+00
-4.98032421e-02 1.14294267e+00 -8.10590982e-01 6.13411248e-01
-1.97958335e-01 1.48583427e-01 1.12188351e+00 -4.96514849e-02
1.09861696e+00 1.16115630e+00 4.21539396e-02 5.06500483e-01
-1.50528216e+00 -7.10749805e-01 5.01297295e-01 2.98314393e-01
-1.00631189e+00 -3.62955987e-01 5.35408080e-01 4.44183424e-02
6.72504783e-01 2.73398072e-01 1.10336530e+00 8.49892676e-01
-8.37491453e-02 6.06226325e-01 8.11506867e-01 -3.89655173e-01
8.39339912e-01 3.25514257e-01 -4.91727889e-01 6.08533323e-01
4.40382183e-01 4.40261930e-01 -3.08359295e-01 -3.99261862e-01
1.08035219e+00 -4.67149854e-01 -2.56806850e-01 -9.90744650e-01
-9.71694648e-01 7.35681355e-01 6.31753206e-01 1.24377929e-01
-8.89047921e-01 4.88139957e-01 2.35066101e-01 5.38627505e-01
-2.11887620e-03 4.17917579e-01 -6.30122364e-01 -3.09571981e-01
-1.17798126e+00 -4.92810160e-02 1.10364151e+00 8.37868214e-01
4.70430315e-01 2.35989496e-01 -1.47645652e-01 3.39744955e-01
1.76183224e-01 3.59827012e-01 -2.25842416e-01 -1.14530039e+00
2.92965412e-01 1.68846071e-01 2.57945269e-01 -4.10231501e-01
-3.35848421e-01 -1.19813097e+00 -7.49151587e-01 2.91291505e-01
3.92403245e-01 -5.00472844e-01 -6.00058138e-01 1.95139349e+00
5.50882638e-01 6.74350262e-01 -2.14886308e-01 8.12994659e-01
-1.22579828e-01 7.32975781e-01 -1.87830478e-01 -7.06086218e-01
9.60641026e-01 -1.19962180e+00 -4.17261720e-01 4.60508093e-02
3.29045773e-01 -4.60932851e-01 5.84527016e-01 4.07771707e-01
-1.17486930e+00 8.68172571e-02 -1.14977479e+00 9.93640721e-01
-1.20961517e-01 -4.40537840e-01 6.90452158e-01 1.15524960e+00
-1.28997123e+00 6.83650553e-01 -6.48118019e-01 -5.57767212e-01
7.37129629e-01 4.84968424e-01 1.64792448e-01 1.45798661e-02
-8.83017719e-01 6.51236296e-01 1.96199492e-01 -3.05941671e-01
-1.21487343e+00 -7.64221668e-01 -3.70450653e-02 4.62426752e-01
1.21259236e+00 -7.30653942e-01 1.10064793e+00 -9.99610305e-01
-1.82996643e+00 2.23270535e-01 -1.44760832e-02 -8.56157064e-01
6.65691912e-01 4.08168674e-01 -5.91630161e-01 2.02387899e-01
-2.80474965e-02 4.61083055e-01 6.73419893e-01 -1.28926003e+00
-6.58474982e-01 -5.76755032e-02 3.77507955e-01 -2.83902198e-01
-8.35551247e-02 2.93232854e-02 -4.11891311e-01 -3.39871615e-01
5.02482057e-02 -9.00956511e-01 -7.13272095e-01 -2.47321781e-02
-4.83425260e-01 1.19910277e-01 6.58022583e-01 -1.73141629e-01
1.36570632e+00 -1.53034627e+00 2.65351236e-01 8.50676417e-01
3.06225687e-01 3.20166916e-01 -2.58642048e-01 5.59929371e-01
2.95060098e-01 4.30789948e-01 1.73103169e-01 1.43319353e-01
-6.71469653e-03 3.80970925e-01 -1.58677801e-01 1.19732313e-01
-4.37295288e-01 8.09530914e-01 -1.12823451e+00 -4.63412941e-01
4.80184779e-02 -1.83876783e-01 -9.00953114e-01 -2.72689372e-01
-3.87493551e-01 4.48617727e-01 -5.02182424e-01 6.41003191e-01
1.01543522e+00 -6.11104608e-01 9.12253141e-01 9.65147316e-02
6.27432615e-02 4.17731144e-02 -1.33133864e+00 1.19351327e+00
-6.75685644e-01 1.34221718e-01 1.87217981e-01 -1.02004111e+00
3.91184688e-01 2.08363503e-01 4.95557904e-01 -6.50761008e-01
8.97538960e-02 2.94841409e-01 5.92213683e-02 -1.23902276e-01
-6.75950050e-02 -2.10363775e-01 1.75241843e-01 5.84823847e-01
5.21013364e-02 4.43994820e-01 1.74676716e-01 4.40631300e-01
1.24061561e+00 -9.51472148e-02 4.98582363e-01 -1.98037252e-01
3.20205271e-01 -2.19938695e-01 9.84547496e-01 8.84973466e-01
-6.24993443e-01 -4.42243189e-01 1.12084568e+00 -2.31514841e-01
-9.32447672e-01 -1.46377563e+00 3.58657479e-01 7.52206862e-01
3.54829311e-01 -1.75149664e-01 -6.10892296e-01 -1.02094638e+00
6.80736229e-02 7.64176369e-01 -3.02308679e-01 1.47398919e-01
-4.04669821e-01 -4.07508314e-01 9.66933593e-02 1.38014883e-01
6.01497591e-01 -9.59487617e-01 -7.18614817e-01 6.83585227e-01
-6.79720193e-03 -1.17773688e+00 -4.01922166e-01 1.82825133e-01
-1.03559434e+00 -7.51066625e-01 -4.12462324e-01 -3.08397740e-01
3.02647531e-01 5.89546449e-02 1.23202777e+00 -7.49744533e-04
-1.12309799e-01 -2.81848181e-02 -2.17937171e-01 4.21619266e-02
-3.75471592e-01 3.73913914e-01 2.02082798e-01 1.41342014e-01
-1.76207736e-01 -1.22202492e+00 -5.98743737e-01 5.87956131e-01
-3.70346934e-01 -2.16061279e-01 6.76180661e-01 6.67251229e-01
6.49930239e-01 3.73666525e-01 9.00636673e-01 -8.62044752e-01
1.86343163e-01 -5.47444582e-01 -1.13530529e+00 4.87116694e-01
-6.89391613e-01 2.64156252e-01 7.79559135e-01 -2.51210779e-01
-6.59271955e-01 -4.05170143e-01 3.30874741e-01 -6.42392814e-01
4.69787180e-01 1.24003977e-01 -4.44671512e-01 -2.92426795e-01
3.56440932e-01 -9.76742804e-02 -8.22168663e-02 -3.05867136e-01
4.46492523e-01 3.20779532e-01 3.97342257e-02 -5.35313785e-01
1.08426750e+00 6.04205549e-01 4.90233332e-01 -4.18577790e-01
-4.64213341e-01 3.28391083e-02 -1.79869086e-01 -4.68372345e-01
2.24652216e-01 -3.59992653e-01 -1.18520808e+00 -2.87596703e-01
-8.58744502e-01 -4.67784047e-01 -5.67823112e-01 3.69395018e-01
-5.58517635e-01 9.94215682e-02 -3.62061650e-01 -9.88532782e-01
3.55243012e-02 -8.68004203e-01 9.25133377e-02 3.88011158e-01
1.57539800e-01 -8.03522408e-01 -9.91747454e-02 3.96279693e-02
8.38250756e-01 2.78274685e-01 1.03815687e+00 -2.38262445e-01
-1.31048703e+00 2.24930987e-01 -2.50749528e-01 -1.53189190e-02
-1.78163677e-01 -3.30631994e-02 -3.79373640e-01 -6.01290941e-01
-3.49100679e-01 2.01396003e-01 4.49776262e-01 7.15136170e-01
1.07572591e+00 -5.62646508e-01 -7.63600469e-01 6.48302436e-01
1.88568342e+00 4.62323695e-01 6.08395875e-01 3.85186434e-01
6.28791600e-02 2.43832767e-01 6.31235063e-01 7.41315961e-01
7.53078535e-02 9.44190741e-01 9.79613900e-01 4.48126942e-02
4.81308065e-02 -4.83836383e-01 2.50995725e-01 2.01246470e-01
3.90163437e-02 -7.36820102e-01 -4.15543914e-01 4.64062333e-01
-1.80464756e+00 -1.04539478e+00 5.02306104e-01 2.43346834e+00
1.32430524e-01 6.40802145e-01 5.99008441e-01 4.31802571e-02
7.83184469e-01 1.25993550e-01 -7.74130762e-01 -7.53341675e-01
9.53281894e-02 3.54575098e-01 1.04634070e+00 3.85883003e-01
-2.85592526e-01 1.15388274e+00 5.69190931e+00 1.14751470e+00
-1.00960207e+00 4.16182190e-01 6.42603457e-01 -4.70478088e-01
-3.06097180e-01 5.35963476e-01 -5.82402349e-01 5.69459856e-01
1.20537293e+00 -4.06468570e-01 1.27854717e+00 7.40643620e-01
5.87847888e-01 -2.30016693e-01 -9.22364831e-01 7.31232703e-01
-5.45329094e-01 -1.84299660e+00 7.14875981e-02 5.49522638e-01
6.50674284e-01 7.47555047e-02 -1.23984098e-01 2.22076669e-01
5.99254668e-01 -6.90000594e-01 6.95664406e-01 1.41759470e-01
5.93864441e-01 -1.07643831e+00 5.34268141e-01 2.83354491e-01
-1.18265116e+00 -4.92202014e-01 -7.76494965e-02 3.67352366e-01
7.24140763e-01 7.59166658e-01 -7.86884129e-01 7.52197087e-01
4.69689250e-01 1.05666872e-02 1.85101032e-02 1.39138186e+00
-2.79977232e-01 5.98721445e-01 -5.11580825e-01 -6.08121604e-02
3.96673918e-01 -4.44917649e-01 9.24030423e-01 3.87663454e-01
2.58433670e-01 -1.16782360e-01 2.59174347e-01 1.08432615e+00
-2.91233957e-01 -3.47492769e-02 -2.60544747e-01 2.40093359e-04
1.22781610e+00 1.13859868e+00 -1.15452552e+00 -3.75832796e-01
7.72586018e-02 8.97856534e-01 3.59927192e-02 4.68425870e-01
-1.18911231e+00 -3.33658725e-01 6.78759336e-01 3.36392045e-01
7.02890754e-01 4.43617292e-02 -1.57107368e-01 -7.60982394e-01
-6.47027194e-02 -6.68342054e-01 4.78288382e-01 -5.56351244e-01
-8.62713873e-01 6.88321412e-01 -1.63477302e-01 -1.32781076e+00
6.28987849e-02 -6.42073900e-02 -6.65416479e-01 5.02705455e-01
-1.61182058e+00 -6.93409443e-01 1.21797495e-01 3.79674703e-01
3.27625275e-01 -2.21860856e-01 5.08050561e-01 3.77865881e-01
-6.62960887e-01 8.97111833e-01 1.41986087e-01 -4.78739023e-01
9.79953855e-02 -8.50913644e-01 1.86133623e-01 1.01360393e+00
2.10111588e-01 1.09899864e-01 6.47530079e-01 -4.13995355e-01
-1.03394377e+00 -7.95091152e-01 6.18377328e-01 3.93907189e-01
6.42021537e-01 -1.24486960e-01 -1.78278923e-01 5.08050561e-01
6.76774308e-02 2.10239723e-01 3.09128791e-01 1.78217255e-02
8.73935372e-02 -4.30180609e-01 -1.59036911e+00 8.16611350e-01
1.46526253e+00 1.65937319e-01 3.83329540e-01 2.04785079e-01
8.69728506e-01 2.26990283e-02 -4.81072247e-01 9.86833721e-02
3.57173383e-01 -9.79564071e-01 8.05887699e-01 -9.40008223e-01
-2.25728825e-01 -3.61131221e-01 -2.47994885e-01 -1.52504897e+00
-3.66887212e-01 -1.30696630e+00 -4.49381888e-01 1.01145041e+00
6.69269621e-01 -1.02208686e+00 1.26281047e+00 -2.38226488e-01
3.40609729e-01 -9.25821424e-01 -1.36366594e+00 -1.47804582e+00
-1.74312294e-01 -3.76149788e-02 1.19587100e+00 7.68555582e-01
1.65933874e-02 2.98048437e-01 -3.86378735e-01 3.76916498e-01
1.01488483e+00 3.52898121e-01 6.76556408e-01 -1.11016178e+00
-8.29171479e-01 -4.88132358e-01 -2.70071238e-01 -1.03531468e+00
-2.54371539e-02 -8.52893174e-01 -5.70584834e-01 -1.43260431e+00
9.71146207e-03 -8.58444870e-01 -5.30120492e-01 1.13359094e-01
5.44676781e-01 -2.81456590e-01 4.72877711e-01 -2.99511105e-02
-6.22168303e-01 2.64550030e-01 1.33214223e+00 1.86677203e-01
-5.08083522e-01 4.84420270e-01 -5.60575068e-01 3.16265464e-01
9.91601884e-01 -7.51352966e-01 -7.24106252e-01 2.17279010e-02
1.18521690e-01 8.94490242e-01 2.59192139e-01 -9.00646806e-01
9.57825780e-02 -4.41637456e-01 -1.25153840e-01 -5.65531790e-01
2.12236151e-01 -9.41686094e-01 4.64350820e-01 8.81601751e-01
9.39184502e-02 -6.42484874e-02 -1.13320716e-01 8.54645967e-01
3.98887366e-01 -2.44672634e-02 1.09578872e+00 -1.83279529e-01
-5.02324820e-01 6.19509399e-01 -5.75628936e-01 7.95044750e-02
1.35203028e+00 -2.83209413e-01 -2.22238988e-01 -5.48646748e-01
-1.11283696e+00 6.28504097e-01 4.02767718e-01 3.55433412e-02
3.04423094e-01 -1.10148740e+00 -3.65888715e-01 -1.56373397e-01
-3.27495992e-01 -7.79776275e-01 3.66482198e-01 1.02212417e+00
-5.59706926e-01 3.94293100e-01 -4.44091529e-01 -3.40550244e-01
-9.74696696e-01 8.62836242e-01 7.12914467e-01 -9.27580774e-01
-4.16337281e-01 8.25682580e-01 -3.06696475e-01 -9.81316641e-02
1.71630055e-01 9.03910175e-02 1.31698102e-01 -9.18258056e-02
-3.28702815e-02 6.99741900e-01 -2.53083825e-01 5.10307029e-02
-5.54290891e-01 1.40986711e-01 -3.18539701e-02 -5.30631840e-01
1.18343544e+00 -2.85870999e-01 -1.55573294e-01 -2.02279165e-01
7.69092262e-01 6.49999157e-02 -8.87196779e-01 -3.78706962e-01
1.09004214e-01 -8.06755543e-01 3.09402972e-01 -1.09015095e+00
-1.61202502e+00 3.29605848e-01 7.73519456e-01 3.18402112e-01
1.15913033e+00 -8.84782001e-02 8.43465209e-01 2.88637076e-02
1.12710238e+00 -1.13019192e+00 4.88049677e-03 -1.16203837e-02
1.42153248e-01 -6.11300349e-01 -7.58173540e-02 -6.28561795e-01
-5.22521675e-01 7.46522605e-01 5.58084965e-01 8.46421719e-02
8.23194981e-01 1.97554067e-01 -4.23851967e-01 -1.52278274e-01
-1.03225720e+00 -2.99427599e-01 -8.89237881e-01 5.74948847e-01
-6.13642633e-01 5.59905887e-01 -2.33339980e-01 2.70137846e-01
4.20110896e-02 1.86383978e-01 8.90238285e-01 7.27195859e-01
-5.51780164e-01 -1.48872519e+00 7.13228807e-02 5.95757723e-01
-1.90009177e-01 4.84940261e-02 1.44726764e-02 7.66689301e-01
5.11233322e-02 1.03249717e+00 -9.20718089e-02 -2.23742291e-01
1.86515063e-01 -3.00807267e-01 5.67165136e-01 -3.23856056e-01
-3.28036934e-01 1.62750166e-02 4.13092315e-01 -7.97289193e-01
-1.91943161e-02 -4.86472756e-01 -9.62101877e-01 -8.46549332e-01
-4.94530261e-01 5.87873816e-01 5.99348128e-01 6.03927374e-01
5.12538135e-01 7.55274475e-01 1.28997159e+00 -3.95981282e-01
-5.97203434e-01 -3.36646020e-01 -9.15560126e-01 -5.46683550e-01
4.08734679e-02 -8.10878038e-01 -3.19246054e-01 -7.88365781e-01] | [5.8247809410095215, 1.745381474494934] |
2124bb21-cab7-4524-963f-a06a5a8b4423 | leveraging-reaction-aware-substructures-for | 2204.05919 | null | https://arxiv.org/abs/2204.05919v4 | https://arxiv.org/pdf/2204.05919v4.pdf | Leveraging Reaction-aware Substructures for Retrosynthesis Analysis | Retrosynthesis analysis is a critical task in organic chemistry central to many important industries. Previously, various machine learning approaches have achieved promising results on this task by representing output molecules as strings and autoregressively decoded token-by-token with generative models. Text generation or machine translation models in natural language processing were frequently utilized approaches. The token-by-token decoding approach is not intuitive from a chemistry perspective because some substructures are relatively stable and remain unchanged during reactions. In this paper, we propose a substructure-level decoding model, where the substructures are reaction-aware and can be automatically extracted with a fully data-driven approach. Our approach achieved improvement over previously reported models, and we find that the performance can be further boosted if the accuracy of substructure extraction is improved. The substructures extracted by our approach can provide users with better insights for decision-making compared to existing methods. We hope this work will generate interest in this fast growing and highly interdisciplinary area on retrosynthesis prediction and other related topics. | ['Li Tan', 'Junren Li', 'Lei Fang', 'Jian-Guang Lou', 'Ming Zhao'] | 2022-04-12 | null | null | null | null | ['retrosynthesis'] | ['medical'] | [ 8.11758876e-01 1.56491920e-01 -5.48994303e-01 -1.54118240e-01
-7.02611625e-01 -7.74356604e-01 9.87330914e-01 7.44430900e-01
-1.22730292e-01 1.25034225e+00 3.88534874e-01 -4.86751378e-01
4.05837804e-01 -9.96841431e-01 -7.09489226e-01 -9.74364817e-01
1.79732963e-01 3.06885421e-01 9.81548131e-02 -3.49317014e-01
6.56440496e-01 6.01597369e-01 -1.52490234e+00 7.44256675e-01
8.71666014e-01 6.20798647e-01 4.60642636e-01 6.80682659e-01
-5.94700277e-01 9.29656088e-01 -4.92956400e-01 -5.64872742e-01
2.33279597e-02 -7.85421669e-01 -7.28383243e-01 -4.40146744e-01
-3.07376742e-01 4.62618470e-01 4.06519361e-02 1.05948687e+00
4.36733067e-01 -1.61851626e-02 9.67152834e-01 -5.39764822e-01
-2.73978561e-01 1.10076177e+00 -1.90119356e-01 -2.97545254e-01
5.36242068e-01 3.91669944e-02 1.13234746e+00 -1.01639032e+00
8.34288657e-01 1.18331230e+00 1.79141924e-01 3.53571296e-01
-1.05894995e+00 -6.58884168e-01 2.82710552e-01 2.72041827e-01
-1.04114437e+00 -5.14755309e-01 6.71236813e-01 -5.42271376e-01
1.43708348e+00 3.76761526e-01 5.65619648e-01 1.07798994e+00
7.93183386e-01 5.00202417e-01 9.46540177e-01 -6.00043893e-01
3.49030644e-01 2.36222729e-01 -4.42245781e-01 5.64759374e-01
2.59742558e-01 3.80491139e-03 -7.08183289e-01 4.74891253e-02
3.03049028e-01 1.72434747e-01 -5.19759320e-02 3.00558861e-02
-1.52736473e+00 1.08272052e+00 2.48234987e-01 5.61067045e-01
-4.76516455e-01 4.02056836e-02 5.59232891e-01 1.79870084e-01
6.96483374e-01 9.00186419e-01 -6.29996061e-01 -6.47287518e-02
-1.33028185e+00 2.12281331e-01 1.03391683e+00 1.18248534e+00
4.61520165e-01 -2.50207949e-02 -1.92156971e-01 5.41647136e-01
2.63620973e-01 -6.57637641e-02 6.59487486e-01 -2.96554208e-01
3.26451331e-01 4.46675718e-01 -3.81452665e-02 -5.30729115e-01
-3.08962882e-01 -3.33718479e-01 -6.52815342e-01 -1.19563781e-01
2.82144159e-01 -5.65079832e-03 -1.03600216e+00 1.39944661e+00
1.97381839e-01 -3.90793502e-01 3.08239460e-01 4.90806401e-01
8.71867597e-01 1.27455556e+00 4.92427438e-01 -5.37221313e-01
1.38580251e+00 -9.96757150e-01 -7.69947290e-01 2.08186269e-01
8.73182774e-01 -1.30482841e+00 3.04795653e-01 9.12126303e-01
-1.07552075e+00 -3.46682817e-01 -1.27092922e+00 -1.48325428e-01
-9.07872200e-01 2.31154740e-01 8.70624959e-01 8.24464977e-01
-5.00180542e-01 1.14589262e+00 -6.94085181e-01 -3.10931474e-01
1.28951609e-01 4.68132973e-01 -1.56390041e-01 1.62498146e-01
-1.13570201e+00 8.57759774e-01 9.24009800e-01 -7.45202973e-02
-8.85893881e-01 -4.66441810e-01 -7.49993742e-01 -2.50458494e-02
6.44724965e-01 -6.20395541e-01 1.49310184e+00 -7.12246180e-01
-1.84380555e+00 4.65621144e-01 -4.82030630e-01 -4.78826255e-01
4.69403654e-01 2.64580607e-01 -4.80429292e-01 -9.77045894e-02
1.10397488e-01 6.66580498e-01 5.97349167e-01 -9.20853019e-01
-6.09124482e-01 -1.63781226e-01 -4.21552092e-01 -2.86429450e-02
-2.55786479e-02 1.28710389e-01 6.58446699e-02 -8.41720819e-01
1.44456044e-01 -7.44851410e-01 -5.14575005e-01 -3.62866342e-01
-5.84824145e-01 -3.17045093e-01 1.84051856e-01 -4.95153219e-01
1.15041947e+00 -1.33184266e+00 4.26216006e-01 1.88153729e-01
6.31042495e-02 2.13008046e-01 8.92966613e-03 1.18214929e+00
-3.99665534e-01 4.25293803e-01 -1.35819942e-01 4.90195788e-02
-1.04486465e-01 -2.16619581e-01 -3.92810136e-01 1.19188346e-01
3.70657831e-01 7.00396299e-01 -1.13609564e+00 -2.58696556e-01
2.00387657e-01 1.75248787e-01 -4.91456211e-01 5.21206595e-02
-8.59447956e-01 4.44911748e-01 -5.99879146e-01 8.76621187e-01
3.78125191e-01 -4.22401801e-02 4.98068959e-01 -2.69565850e-01
-6.69786334e-01 6.58614695e-01 -8.53517234e-01 1.81282461e+00
-2.80311853e-01 2.67180890e-01 -5.26708722e-01 -1.09072876e+00
1.04511905e+00 5.99291563e-01 4.14441288e-01 -4.91844773e-01
1.36426359e-01 3.79729867e-01 1.69826820e-01 -4.21567053e-01
1.08302438e+00 -6.94326580e-01 -1.29798755e-01 -1.94876119e-02
7.97894448e-02 -2.96778262e-01 6.27509832e-01 2.24993736e-01
6.26573563e-01 4.23656940e-01 8.58537793e-01 -2.22034946e-01
6.33049071e-01 3.88511866e-01 3.65003794e-01 6.66713655e-01
3.34850788e-01 2.79845834e-01 5.37941575e-01 -3.55446726e-01
-1.38289833e+00 -4.47608232e-01 -3.49659957e-02 9.98094678e-01
-3.57289910e-01 -1.01966619e+00 -5.33806384e-01 -5.15364230e-01
-2.76762694e-01 7.06367731e-01 -2.40113035e-01 -1.27335608e-01
-2.66816616e-01 -1.03222847e+00 3.35029006e-01 5.16920388e-01
-5.74588366e-02 -9.49040949e-01 2.70321190e-01 9.62525725e-01
9.63260084e-02 -8.50438297e-01 -1.34487063e-01 6.22439265e-01
-8.64764214e-01 -7.85613418e-01 -9.40145373e-01 -7.36936033e-01
5.28138578e-01 1.11704201e-01 6.77528441e-01 -3.77551496e-01
-3.59367818e-01 -5.07273793e-01 -5.29579043e-01 -8.36062968e-01
-1.10000014e+00 4.12825853e-01 -7.00072348e-02 -2.70436794e-01
1.94902733e-01 -4.29384440e-01 -5.85103512e-01 -1.22872531e-01
-9.47728336e-01 3.06839377e-01 7.01744437e-01 5.39190650e-01
7.18029618e-01 4.43588570e-02 7.80277669e-01 -1.22462356e+00
4.97096986e-01 -5.84143579e-01 -5.57038546e-01 1.99275613e-01
-7.62622774e-01 6.43950224e-01 1.08835220e+00 -2.77133703e-01
-1.11355019e+00 3.84822458e-01 -4.57911491e-01 3.97416621e-01
-1.82932064e-01 8.20025027e-01 -4.22743231e-01 1.15570739e-01
4.61628437e-01 4.49805260e-01 -2.20603764e-01 -4.40594167e-01
5.62616825e-01 7.11744130e-01 -5.11180284e-03 -8.69376719e-01
3.21166933e-01 9.73872021e-02 3.78720164e-01 -9.11675096e-01
-4.85089153e-01 -5.81324518e-01 -5.52528918e-01 2.02076808e-02
8.20514500e-01 -9.63484108e-01 -7.69789934e-01 1.69293880e-02
-1.30213106e+00 1.39297426e-01 5.73354438e-02 4.26185459e-01
-4.77266073e-01 6.67279899e-01 -5.39160311e-01 -8.38343501e-01
-4.74795789e-01 -1.30759859e+00 1.01508522e+00 7.39933476e-02
-4.60860014e-01 -8.75505805e-01 1.26792341e-01 3.16263974e-01
6.53729588e-02 2.83719063e-01 1.18723667e+00 -9.16131616e-01
-7.05876350e-01 -1.92344308e-01 1.67075351e-01 -4.06367816e-02
3.70623410e-01 1.50747985e-01 -8.31260979e-01 -8.24163705e-02
-4.12526488e-01 -1.71137601e-01 1.23143053e+00 1.17617510e-01
1.25616133e+00 -3.34108919e-01 -5.52590787e-01 5.73291630e-02
1.33021271e+00 6.18833542e-01 6.55651987e-01 1.78985462e-01
6.30314708e-01 7.23361075e-01 5.73426545e-01 5.20911813e-01
-1.71038285e-02 4.25006688e-01 3.17618877e-01 -3.04989181e-02
5.89570813e-02 -6.16106153e-01 5.81639111e-01 7.51084447e-01
-4.69015568e-01 -7.14075208e-01 -6.24914289e-01 2.30217487e-01
-1.74354875e+00 -1.22852743e+00 -4.02423024e-01 2.08358407e+00
1.11477864e+00 2.24746197e-01 1.91910982e-01 3.72384667e-01
4.82229918e-01 5.66741079e-02 -3.79815012e-01 -7.33469665e-01
5.99921532e-02 6.23455286e-01 7.05033362e-01 3.95838708e-01
-8.17766011e-01 1.36570036e+00 6.31774569e+00 1.07552636e+00
-1.34566212e+00 -2.29197204e-01 4.86384451e-01 1.49849877e-01
-2.71039158e-01 2.92614609e-01 -1.13546073e+00 5.27399778e-01
1.12912178e+00 -3.56603354e-01 8.74469504e-02 7.53101587e-01
5.39941311e-01 -7.84381032e-02 -1.36418200e+00 6.28474057e-01
-4.28055637e-02 -1.90359843e+00 3.45174372e-01 1.74195141e-01
7.08658218e-01 -4.36718374e-01 -2.06767157e-01 1.27766564e-01
1.58694148e-01 -1.08060396e+00 8.46967578e-01 4.47224826e-01
5.83427548e-01 -8.53334785e-01 4.57794905e-01 2.85013378e-01
-1.29708982e+00 1.59770489e-01 -4.40610915e-01 -2.91176569e-02
1.23271877e-02 8.20723236e-01 -1.30424559e+00 1.09771538e+00
-1.19456246e-01 9.80739653e-01 -1.73550487e-01 8.88617754e-01
-2.62257785e-01 5.84598243e-01 7.57352123e-03 -6.98458016e-01
3.61607492e-01 -4.75900531e-01 3.47540796e-01 1.63637137e+00
6.23019695e-01 3.68691497e-02 1.89626724e-01 8.69635761e-01
-2.49306887e-01 5.69477856e-01 -6.22142494e-01 -7.41533577e-01
1.60210893e-01 1.12431812e+00 -1.14141738e+00 -6.12412035e-01
-2.80220062e-01 9.31455970e-01 -2.64726609e-01 4.23386022e-02
-6.30771101e-01 -5.36702096e-01 3.50341648e-01 1.38457894e-01
4.21936750e-01 -3.71064126e-01 -1.98888451e-01 -1.19748402e+00
-4.77108359e-01 -8.57365489e-01 2.40787148e-01 -4.88911957e-01
-9.05510485e-01 3.85146379e-01 -2.12378874e-01 -1.03857577e+00
-2.60173887e-01 -9.05606210e-01 -4.44990247e-01 6.19623244e-01
-1.52903986e+00 -9.29432392e-01 3.84802252e-01 -7.82312304e-02
9.47562814e-01 -1.97380289e-01 9.75023985e-01 -3.28206569e-02
-5.83073020e-01 2.54520088e-01 4.81471181e-01 -3.41992080e-01
8.39391232e-01 -1.21564698e+00 2.46112600e-01 6.95085824e-01
3.30005020e-01 7.43386924e-01 9.28898394e-01 -7.80797601e-01
-1.44029415e+00 -1.27162981e+00 1.22190177e+00 -2.77303278e-01
6.29889548e-01 -6.62196279e-01 -4.88584667e-01 2.08730340e-01
3.00308436e-01 -7.40624845e-01 1.12991881e+00 -9.59269404e-02
-2.03788772e-01 2.15956151e-01 -8.85608494e-01 6.24474347e-01
7.88895249e-01 -2.00841650e-01 -3.84693682e-01 7.61091888e-01
6.47350013e-01 -3.07820529e-01 -1.10331118e+00 -9.30924788e-02
4.91571993e-01 -4.37253982e-01 9.10881996e-01 -8.04833055e-01
9.54857409e-01 -1.98376864e-01 5.17802022e-04 -1.28692365e+00
-1.37520239e-01 -8.28091502e-01 9.01306272e-02 9.72600758e-01
9.15589154e-01 -4.50063080e-01 7.40520954e-01 2.09256485e-01
-3.62165987e-01 -4.88514990e-01 -5.19330144e-01 -7.33504832e-01
3.67903620e-01 -1.62117660e-01 7.54502058e-01 7.66947865e-01
5.10229588e-01 6.81644380e-01 -4.72772449e-01 -3.51415932e-01
1.97045311e-01 2.57427275e-01 3.45832556e-01 -1.14019263e+00
-1.78128257e-01 -5.26750624e-01 -1.66070491e-01 -9.87293661e-01
5.08855656e-02 -1.54560781e+00 -4.76492234e-02 -1.45461512e+00
3.20613831e-01 5.36512360e-02 -9.99099761e-02 5.22945285e-01
1.11237556e-01 -6.59895018e-02 3.15612182e-02 -1.72625445e-02
-3.75021070e-01 6.18411720e-01 1.32057118e+00 -3.52591455e-01
-2.71494597e-01 -6.90203384e-02 -8.80679131e-01 3.06405455e-01
9.46845293e-01 -7.60406375e-01 -2.55734950e-01 3.53321135e-01
4.97042805e-01 8.46353918e-02 -2.38309041e-01 -5.72188914e-01
2.77932972e-01 -3.34688574e-01 5.15082419e-01 -6.36243522e-01
2.99433261e-01 -4.49860126e-01 4.02069449e-01 7.09969461e-01
-5.00272095e-01 -2.64698714e-01 4.37819287e-02 6.41618788e-01
-1.08505208e-02 -3.87027562e-01 4.17509884e-01 -6.10444427e-01
-2.71761626e-01 2.57893324e-01 -1.02962971e+00 -7.60692298e-01
9.23186004e-01 -2.60182887e-01 -5.66452146e-02 -1.49903283e-01
-7.80909657e-01 -3.29645902e-01 3.31949145e-01 3.30081880e-01
3.81675273e-01 -9.45203483e-01 -5.32051861e-01 -3.59093677e-03
2.92822301e-01 -2.39947334e-01 -3.17593902e-01 5.89014351e-01
-5.64578295e-01 9.65544760e-01 1.23555087e-01 -2.16570392e-01
-1.35708737e+00 6.77456439e-01 2.54573505e-02 -4.01211381e-01
-4.60981652e-02 4.55247730e-01 -9.24272463e-02 -1.22972853e-01
3.44287157e-02 -7.97935426e-01 -1.68011591e-01 4.21193719e-01
4.41165894e-01 9.16881934e-02 5.72318733e-01 -4.06188905e-01
-2.20701635e-01 2.22985744e-01 -7.39765346e-01 1.18600264e-01
1.42623615e+00 4.99296635e-01 -7.84474909e-02 1.78667083e-01
9.76141632e-01 3.02072223e-02 -7.41726696e-01 4.11749221e-02
1.60706282e-01 4.40604165e-02 4.77417819e-02 -8.52723360e-01
-4.24455225e-01 9.44727778e-01 5.51751181e-02 1.92847475e-01
5.97054303e-01 5.86989671e-02 6.38769746e-01 5.45794904e-01
2.21358567e-01 -1.11284435e+00 7.10709300e-03 3.56713861e-01
7.17026055e-01 -1.06025636e+00 1.93016961e-01 -6.66168332e-01
-3.65862489e-01 1.59539783e+00 -9.63622853e-02 4.21188653e-01
1.48454055e-01 1.10562749e-01 -5.30686796e-01 1.12486389e-02
-7.92056978e-01 -3.44196796e-01 9.56087932e-02 3.06962222e-01
1.23678100e+00 2.00821146e-01 -8.78499568e-01 6.52134776e-01
-3.57363701e-01 -1.17226303e-01 5.86001635e-01 1.24465334e+00
-8.13934624e-01 -2.09021354e+00 -2.69869715e-01 2.49308556e-01
-6.43003166e-01 -5.28628767e-01 -8.60366404e-01 3.86278957e-01
1.24307163e-01 1.16803288e+00 -6.55286372e-01 -2.83596724e-01
9.59206000e-02 4.08079296e-01 8.43145907e-01 -9.75668490e-01
-6.50219858e-01 2.63225943e-01 4.14630115e-01 -2.02657029e-01
-5.10563493e-01 -6.47707462e-01 -1.34247744e+00 -3.13651025e-01
-4.49404567e-01 6.58222020e-01 1.03261423e+00 8.25056255e-01
3.11338663e-01 6.42744541e-01 6.84752405e-01 -7.72499025e-01
-3.16735834e-01 -8.58839631e-01 -1.67648539e-01 -1.55825764e-01
-1.00498706e-01 -2.54074275e-01 9.57922414e-02 3.75733972e-01] | [4.545088291168213, 6.077699184417725] |
f5af63b4-1392-4f2c-8ae8-fda9db98a618 | phone-duration-modeling-for-speaker-age | 2109.01568 | null | https://arxiv.org/abs/2109.01568v1 | https://arxiv.org/pdf/2109.01568v1.pdf | Phone Duration Modeling for Speaker Age Estimation in Children | Automatic inference of important paralinguistic information such as age from speech is an important area of research with numerous spoken language technology based applications. Speaker age estimation has applications in enabling personalization and age-appropriate curation of information and content. However, research in speaker age estimation in children is especially challenging due to paucity of relevant speech data representing the developmental spectrum, and the high signal variability especially intra age variability that complicates modeling. Most approaches in children speaker age estimation adopt methods directly from research on adult speech processing. In this paper, we propose features specific to children and focus on speaker's phone duration as an important biomarker of children's age. We propose phone duration modeling for predicting age from child's speech. To enable that, children speech is first forced aligned with the corresponding transcription to derive phone duration distributions. Statistical functionals are computed from phone duration distributions for each phoneme which are in turn used to train regression models to predict speaker age. Two children speech datasets are employed to demonstrate the robustness of phone duration features. We perform age regression experiments on age categories ranging from children studying in kindergarten to grade 10. Experimental results suggest phone durations contain important development-related information of children. Phonemes contributing most to estimation of children speaker age are analyzed and presented. | ['Shrikanth Narayanan', 'Catherine Lord', 'Somer Bishop', 'Prashanth Gurunath Shivakumar'] | 2021-09-03 | null | null | null | null | ['age-estimation', 'age-estimation'] | ['computer-vision', 'miscellaneous'] | [ 5.74106947e-02 7.96337351e-02 -3.12154830e-01 -9.04854596e-01
-7.81202376e-01 -3.19562137e-01 3.91229421e-01 6.11763835e-01
-5.07244468e-01 3.48860741e-01 6.11949563e-01 2.23837444e-03
-3.30105945e-02 -4.32483226e-01 -2.37735823e-01 -7.52409816e-01
-1.91059664e-01 3.77206326e-01 -1.13214150e-01 1.57369435e-01
4.23930526e-01 3.32877159e-01 -2.05015969e+00 -4.74717543e-02
1.06639659e+00 7.19714820e-01 1.72951281e-01 9.64171648e-01
-3.52289647e-01 1.42664760e-01 -8.97296250e-01 -7.77061358e-02
-4.21450436e-01 -2.47339427e-01 -5.21653831e-01 -1.50338262e-01
5.14609933e-01 -4.34650809e-01 1.51576072e-01 7.31195152e-01
7.31408834e-01 9.67682824e-02 1.05912971e+00 -1.13264585e+00
-4.50463682e-01 1.44078648e+00 -3.81705642e-01 4.41543549e-01
6.78985476e-01 -4.96830642e-01 6.51800573e-01 -7.58866012e-01
-1.55190909e-02 1.41635513e+00 4.14144874e-01 9.29581344e-01
-8.07078600e-01 -8.20546985e-01 1.59354225e-01 8.86043757e-02
-1.47621047e+00 -7.87664413e-01 8.85337174e-01 -7.90615380e-01
6.36776984e-01 1.33561864e-01 8.63968849e-01 1.08143425e+00
-8.00995529e-02 5.80922127e-01 1.27906573e+00 -7.10289836e-01
-1.01772202e-02 2.14049354e-01 4.87363875e-01 3.07575822e-01
-2.26397887e-01 1.43422425e-01 -8.53088856e-01 3.63088369e-01
2.95750409e-01 -6.12932861e-01 1.60973564e-01 4.34681028e-01
-1.02400124e+00 5.99277675e-01 -2.99214751e-01 4.92302209e-01
-1.62164643e-01 -3.34579051e-02 2.75521189e-01 5.27141213e-01
1.05283904e+00 -8.50493927e-03 -6.03685796e-01 -7.28698015e-01
-9.93494689e-01 2.92458624e-01 6.83270872e-01 8.07787001e-01
4.26964432e-01 2.55467184e-02 5.90508357e-02 1.55509186e+00
5.31384885e-01 5.32690942e-01 6.78720951e-01 -6.64069116e-01
3.83613437e-01 3.07269961e-01 -4.60366726e-01 -3.43866974e-01
-5.00155032e-01 -5.15210740e-02 -5.11389852e-01 2.36190353e-02
7.91723788e-01 -2.79278681e-02 -9.33504343e-01 1.91718471e+00
5.47572613e-01 -1.41600192e-01 -1.75549775e-01 3.07544827e-01
1.02643871e+00 7.51046062e-01 4.23317194e-01 -6.86053574e-01
1.53219593e+00 -2.90473372e-01 -5.40656209e-01 -1.89777851e-01
3.10684264e-01 -7.31385827e-01 7.29511380e-01 6.07622862e-01
-1.30987036e+00 -5.25911987e-01 -6.86426997e-01 6.26484603e-02
-4.65660483e-01 -2.62397714e-02 6.06580257e-01 1.30772984e+00
-1.02117562e+00 4.57552224e-01 -8.55444312e-01 -4.11181718e-01
2.58005410e-02 6.34641469e-01 -3.33322883e-01 4.20333505e-01
-1.13896179e+00 8.24185729e-01 2.54289180e-01 -2.20885888e-01
-7.80859947e-01 -1.15384746e+00 -9.21852171e-01 -1.38954461e-01
-4.58113164e-01 -2.28268117e-01 1.45513964e+00 -7.60244668e-01
-1.69036365e+00 1.16120946e+00 -1.87994897e-01 -3.06778342e-01
1.82073176e-01 -2.02553645e-01 -6.00131929e-01 8.26414600e-02
4.72600991e-03 7.90395379e-01 1.10096419e+00 -6.10922515e-01
-8.92884254e-01 -9.01542842e-01 -6.27980709e-01 3.17662239e-01
-6.75174177e-01 7.59676993e-01 -1.35826971e-03 -7.40012825e-01
5.14028728e-01 -6.34351850e-01 2.00964943e-01 -3.33039165e-01
4.62669991e-02 -8.19548845e-01 3.78105849e-01 -1.26098299e+00
1.66919672e+00 -1.93418694e+00 -1.03913702e-03 1.08734600e-01
9.69706029e-02 -2.83113807e-01 2.39426076e-01 3.21139693e-01
-1.45612970e-01 -8.55986997e-02 3.17304470e-02 -6.04416549e-01
-8.07450432e-03 -7.08930492e-02 2.48659119e-01 4.57213759e-01
1.33379877e-01 3.83276671e-01 -6.40298843e-01 -6.89206541e-01
3.92099954e-02 3.49859297e-01 -3.78812462e-01 4.93399173e-01
1.99065432e-01 7.06627190e-01 -1.56501085e-01 7.44024098e-01
3.74314696e-01 8.92612576e-01 -4.40147191e-01 3.19003552e-01
-5.40039420e-01 6.58827960e-01 -7.58399069e-01 1.44411516e+00
-6.12541854e-01 5.00636995e-01 2.40044087e-01 -1.07918739e+00
1.14465845e+00 4.53701049e-01 3.01662683e-01 -3.24817449e-01
2.90362388e-01 1.98513642e-01 4.57597464e-01 -5.31234145e-01
3.29706728e-01 9.81916189e-02 -1.30779088e-01 3.20585042e-01
4.03417349e-02 -6.84269726e-01 2.34627619e-01 -2.33473629e-01
4.48676229e-01 -4.02269453e-01 1.75201356e-01 -4.54078287e-01
7.38486052e-01 -8.88480604e-01 3.48886728e-01 3.23093146e-01
-3.33229750e-01 3.36967736e-01 3.39903906e-02 -4.15803827e-02
-8.88092637e-01 -1.41723990e+00 -3.08050722e-01 1.75017154e+00
-8.71024966e-01 -2.32416898e-01 -1.06233108e+00 -2.33170286e-01
-7.27938935e-02 7.95469046e-01 -6.68845475e-01 -1.03435919e-01
-6.34202480e-01 -1.89744800e-01 4.57039773e-01 6.09278202e-01
-1.84550077e-01 -9.86771882e-01 -1.72661275e-01 1.49197027e-01
2.38624498e-01 -8.84562075e-01 -7.21558809e-01 -6.16966523e-02
-8.28078926e-01 -6.43701851e-01 -1.10854900e+00 -1.04082155e+00
4.69712764e-01 -3.17629397e-01 9.30812180e-01 -2.73125768e-01
-4.69250977e-02 7.79865026e-01 -4.69317347e-01 -1.04986978e+00
-1.03144681e+00 4.36148226e-01 6.69238210e-01 -4.01680619e-01
7.52183318e-01 -1.06238365e+00 -5.51778913e-01 -5.52315190e-02
-1.93571344e-01 -1.71266288e-01 2.48816267e-01 3.78622651e-01
-1.76420614e-01 -1.78864941e-01 1.01881194e+00 -4.60504264e-01
3.54298592e-01 -4.77534801e-01 -5.70766509e-01 -4.28259820e-02
-8.73359621e-01 -1.32802948e-01 1.82982087e-01 -7.70190597e-01
-1.00644922e+00 -8.04094821e-02 -7.11008728e-01 2.49818891e-01
-7.16030836e-01 3.77049565e-01 -2.57049978e-01 4.54967409e-01
3.13560635e-01 1.03463240e-01 2.28670850e-01 -5.49256802e-01
3.84051502e-02 1.32723033e+00 9.08206165e-01 -7.86560595e-01
5.43664336e-01 -3.68572801e-01 -3.98465246e-01 -1.41177094e+00
-6.00796282e-01 -4.32247818e-01 -6.07762158e-01 -5.25705218e-01
6.76109731e-01 -9.43957031e-01 -7.03560710e-01 1.23618710e+00
-9.45844233e-01 -4.71694991e-02 7.81339034e-02 9.25803781e-01
-5.18270075e-01 7.79851750e-02 -6.86233103e-01 -1.34067166e+00
-7.73519993e-01 -9.22851264e-01 7.85134792e-01 5.24881840e-01
-7.04327345e-01 -1.11578870e+00 1.43846318e-01 6.85993433e-01
7.88649995e-05 -3.42874937e-02 9.17037666e-01 -6.92336559e-01
2.71160901e-01 -1.60503253e-01 2.98062235e-01 4.49070334e-01
2.17594430e-01 3.68490577e-01 -1.27099967e+00 -1.56048685e-01
-2.97194928e-01 -4.94645201e-02 4.35725033e-01 9.18450415e-01
1.04324818e+00 9.57728624e-02 -5.04568629e-02 1.77124172e-01
6.02975011e-01 4.07591581e-01 2.19396293e-01 -2.80160367e-01
6.71482325e-01 1.45860088e+00 6.90221131e-01 6.29470170e-01
8.32668841e-01 4.35459912e-01 -1.16047867e-01 6.84687674e-01
-1.24476403e-01 -2.62629986e-01 7.00655758e-01 1.62033629e+00
-1.94042787e-01 1.41257346e-01 -1.06769872e+00 1.14048171e+00
-7.63014138e-01 -7.13861346e-01 -2.24958420e-01 2.41583490e+00
1.17094183e+00 1.31039068e-01 6.94230556e-01 8.09600234e-01
1.00090110e+00 -1.69556469e-01 -8.54695141e-02 -9.01674151e-01
3.30185652e-01 4.58112806e-01 -2.63323057e-02 3.92282456e-01
-8.00752938e-01 7.90504456e-01 6.26933002e+00 7.31232941e-01
-9.68419552e-01 -1.62605405e-01 7.34930456e-01 4.00176309e-02
-1.78695068e-01 -4.64967340e-01 -1.02889776e+00 6.33054972e-01
1.49871099e+00 -4.12708759e-01 3.53309453e-01 6.60314023e-01
5.50467372e-01 -4.25220966e-01 -1.51131582e+00 1.19417596e+00
4.76741269e-02 -3.84281158e-01 -5.98273933e-01 -3.66729237e-02
5.20450413e-01 -4.10410553e-01 1.84183523e-01 1.61520511e-01
4.73300554e-03 -9.79119599e-01 1.12020910e+00 2.22206742e-01
1.04527211e+00 -1.08912110e+00 1.05724305e-01 3.62751007e-01
-1.17071068e+00 -1.80354446e-01 -1.52742546e-02 -5.13072371e-01
-1.59364566e-02 7.64192104e-01 -1.21478534e+00 -1.51802585e-01
7.38931000e-01 4.75980043e-01 -5.55911660e-01 7.57893682e-01
-1.61492392e-01 1.15082467e+00 -3.53235364e-01 -2.15374798e-01
-4.62795526e-01 6.46362305e-02 2.28035629e-01 1.15884030e+00
9.34497476e-01 2.57155418e-01 -4.22363311e-01 3.72073740e-01
3.04178536e-01 5.07928431e-01 -4.14049983e-01 -4.98749971e-01
1.14911401e+00 8.37073624e-01 -5.95938087e-01 1.09231137e-02
-5.03633022e-01 5.81857562e-01 2.72974968e-01 -3.14299256e-01
-3.06278527e-01 -2.08728984e-01 9.15627122e-01 1.17239147e-01
-3.77527699e-02 -2.61383951e-01 -3.45606089e-01 -4.65362251e-01
-3.20595890e-01 -6.06930673e-01 2.67017096e-01 -2.75991142e-01
-1.18348312e+00 -8.64223614e-02 2.96582013e-01 -7.85433948e-01
-9.10984695e-01 -3.98755372e-01 -1.02803731e+00 9.86372292e-01
-6.24434769e-01 -1.21541166e+00 -1.45602077e-01 1.76207200e-01
9.99040663e-01 -2.04823032e-01 7.23181427e-01 2.69108593e-01
-6.48089111e-01 1.07966757e+00 -4.10577893e-01 -4.88133244e-02
6.74607038e-01 -1.59778297e+00 5.36423147e-01 4.66590375e-01
-1.45417824e-01 4.17340457e-01 1.07739711e+00 -4.80295181e-01
-9.36667919e-01 -4.56698388e-01 1.51882446e+00 -6.75347507e-01
6.05299175e-01 -3.98822844e-01 -7.85776377e-01 1.62097797e-01
7.67109692e-02 -8.80805969e-01 9.63752329e-01 4.24799144e-01
-1.21938266e-01 -2.20918894e-01 -1.02255011e+00 2.75195450e-01
1.05445278e+00 -5.75850666e-01 -7.77777612e-01 -3.44815165e-01
6.74207449e-01 -2.53283799e-01 -1.54965436e+00 2.96265066e-01
1.03944838e+00 -8.56837690e-01 9.57623661e-01 -6.48120372e-03
5.03156722e-01 3.92349035e-01 1.04120322e-01 -1.58346808e+00
1.37051672e-01 -4.60334361e-01 -1.44934103e-01 2.30108905e+00
3.46492469e-01 -3.89919162e-01 8.83602738e-01 7.37366796e-01
-3.74823600e-01 -7.52237082e-01 -8.14165831e-01 -5.99388838e-01
5.79227269e-01 -8.13031971e-01 8.10387909e-01 7.69932806e-01
1.46596106e-02 2.87227243e-01 5.20074517e-02 3.47044617e-01
6.45142674e-01 -4.25853729e-01 5.03305256e-01 -1.60629892e+00
6.60792142e-02 -5.66939175e-01 -5.78635573e-01 -4.99873221e-01
4.25678045e-01 -3.03860813e-01 2.33129516e-01 -1.13154471e+00
-1.72251731e-01 -3.82133693e-01 -5.53532280e-02 2.54590716e-03
-3.30679268e-01 -1.69701651e-01 -1.91404745e-01 -4.90852982e-01
4.44200307e-01 1.57769814e-01 8.84591460e-01 -1.17115453e-01
-4.48319495e-01 8.14855754e-01 -3.10620606e-01 8.54374945e-01
9.01069582e-01 -3.11356366e-01 -7.95783579e-01 -8.95052589e-03
-6.88428581e-02 2.13198885e-01 -3.80614161e-01 -9.14631605e-01
1.26680611e-02 -2.50102103e-01 3.09566468e-01 -8.57874453e-01
2.34768137e-01 -3.62040699e-01 -2.35858545e-01 2.92716712e-01
-4.07381237e-01 6.45361096e-02 -2.21364826e-01 -2.30648473e-01
-1.72975227e-01 -5.63957751e-01 7.54746497e-01 3.99364650e-01
-5.78897357e-01 4.08551604e-01 -6.09110594e-01 1.03942193e-01
7.38636255e-01 -3.65231782e-01 2.17973366e-01 -4.31223750e-01
-9.65970635e-01 2.43377358e-01 3.21349919e-01 6.00438416e-01
4.85329360e-01 -1.12542009e+00 -8.27145219e-01 3.48963827e-01
1.98680773e-01 -1.24136899e-02 3.71413261e-01 7.55382657e-01
-3.65558058e-01 1.97451830e-01 -9.75714698e-02 -5.74752986e-01
-2.01082444e+00 4.36173618e-01 -3.08783203e-01 3.12176585e-01
8.43329206e-02 1.38910854e+00 1.24538772e-01 -1.25901297e-01
7.26026833e-01 -6.69180393e-01 -6.47852719e-01 6.39460146e-01
8.61813724e-01 8.69107604e-01 -4.13001440e-02 -7.62768865e-01
-3.99798363e-01 6.27944112e-01 -3.80624413e-01 -4.69275057e-01
1.11609399e+00 -6.05266929e-01 -9.30750594e-02 9.89675999e-01
1.04258037e+00 1.81819797e-01 -9.67279971e-01 1.19391866e-01
1.65806010e-01 -1.95339665e-01 -6.98138326e-02 -4.57874715e-01
-8.22270811e-01 1.03432882e+00 7.67416954e-01 3.33719909e-01
1.20971942e+00 4.27578896e-01 6.07442200e-01 -2.30421111e-01
8.34830999e-02 -1.46796334e+00 -2.52521515e-01 4.63613033e-01
9.37509418e-01 -1.22035933e+00 -2.22478390e-01 -5.29454648e-01
-1.01790987e-01 9.66915548e-01 7.61770368e-01 4.98731345e-01
7.21757472e-01 8.27364251e-02 3.01294774e-01 3.16854686e-01
-6.40680134e-01 -2.61503100e-01 6.37248278e-01 1.04479992e+00
1.10034573e+00 4.38774973e-01 -7.89918542e-01 9.32255268e-01
-1.11246252e+00 -4.85517681e-01 3.92320514e-01 5.24380028e-01
-5.63681126e-01 -1.34143519e+00 -6.43026948e-01 6.76160455e-01
-8.22896898e-01 -1.62540644e-01 -1.99372947e-01 4.86344576e-01
4.15533453e-01 1.10602164e+00 2.40993798e-01 -2.70947665e-01
-4.97217802e-03 4.32812214e-01 8.69550467e-01 -9.40211236e-01
-5.69222093e-01 -1.20476879e-01 4.15172130e-01 1.40612900e-01
-2.65559047e-01 -1.27576506e+00 -1.09087479e+00 -3.04776073e-01
-2.66775727e-01 3.35671335e-01 1.27896357e+00 8.87435496e-01
-6.38393581e-01 2.50548780e-01 8.67188513e-01 -8.28415930e-01
-2.59868443e-01 -1.22915351e+00 -8.38802040e-01 -3.81789617e-02
4.00910348e-01 -6.79797351e-01 -5.62267542e-01 2.39425212e-01] | [14.237159729003906, 6.1906561851501465] |
98ad7f61-23cc-4ffb-84d5-32af83542f73 | lightweight-attentional-feature-fusion-for | 2112.01832 | null | https://arxiv.org/abs/2112.01832v3 | https://arxiv.org/pdf/2112.01832v3.pdf | Lightweight Attentional Feature Fusion: A New Baseline for Text-to-Video Retrieval | In this paper we revisit feature fusion, an old-fashioned topic, in the new context of text-to-video retrieval. Different from previous research that considers feature fusion only at one end, let it be video or text, we aim for feature fusion for both ends within a unified framework. We hypothesize that optimizing the convex combination of the features is preferred to modeling their correlations by computationally heavy multi-head self attention. We propose Lightweight Attentional Feature Fusion (LAFF). LAFF performs feature fusion at both early and late stages and at both video and text ends, making it a powerful method for exploiting diverse (off-the-shelf) features. The interpretability of LAFF can be used for feature selection. Extensive experiments on five public benchmark sets (MSR-VTT, MSVD, TGIF, VATEX and TRECVID AVS 2016-2020) justify LAFF as a new baseline for text-to-video retrieval. | ['Xirong Li', 'Jianfeng Dong', 'Fangming Zhou', 'Ziyue Wang', 'Aozhu Chen', 'Fan Hu'] | 2021-12-03 | null | null | null | null | ['ad-hoc-video-search'] | ['computer-vision'] | [ 4.06774096e-02 -6.37550831e-01 -3.41220766e-01 -4.42146391e-01
-1.27458692e+00 -5.93308449e-01 1.09324276e+00 2.30836064e-01
-6.07821882e-01 3.95333529e-01 4.60308999e-01 -2.98945755e-02
-3.60135108e-01 -6.16489872e-02 -3.88493627e-01 -5.83044946e-01
5.79584725e-02 3.01613510e-01 1.17658906e-01 -1.40689552e-01
3.70920926e-01 2.75553167e-01 -1.96450961e+00 6.29180729e-01
5.81475079e-01 1.33778131e+00 1.25631347e-01 6.02499723e-01
-1.80319175e-02 5.67822754e-01 -2.36395285e-01 -5.55070579e-01
3.94720823e-01 3.21464427e-02 -8.30316365e-01 -1.15789128e-02
7.87220836e-01 -2.61439264e-01 -4.93427128e-01 6.82281017e-01
5.65752149e-01 3.88361782e-01 8.05669963e-01 -1.48698926e+00
-3.96713555e-01 4.19200748e-01 -6.19282365e-01 3.19660932e-01
7.56081581e-01 -6.14233781e-04 1.43831313e+00 -1.45824897e+00
7.46162891e-01 1.40934157e+00 3.36693585e-01 2.88566381e-01
-8.26556385e-01 -3.25436354e-01 4.17855978e-01 5.43479204e-01
-1.57910466e+00 -7.15563536e-01 3.19125235e-01 -4.74024087e-01
9.96268153e-01 6.89348698e-01 6.00780547e-01 1.20864606e+00
3.39222729e-01 1.57622015e+00 7.74970710e-01 -3.84806663e-01
-1.88537836e-01 -1.89942226e-01 4.70622718e-01 4.27162558e-01
6.37481138e-02 -7.99459964e-02 -1.03745663e+00 -3.44087869e-01
1.30366907e-01 2.07654640e-01 -5.58504105e-01 -2.32689053e-01
-1.33779490e+00 7.72393703e-01 2.01623701e-02 2.19801769e-01
-4.25141692e-01 3.73952352e-02 6.62686825e-01 6.67694390e-01
5.56212664e-01 2.86216289e-01 -4.63438153e-01 -2.85610825e-01
-1.22859895e+00 4.85247910e-01 5.82242012e-01 9.05571997e-01
7.25712299e-01 -2.93814242e-01 -7.95798242e-01 7.26737678e-01
5.45923054e-01 4.24298614e-01 6.74793720e-01 -5.96951604e-01
4.29913968e-01 3.02358359e-01 1.44890323e-01 -7.29456067e-01
-2.76927888e-01 -9.46250856e-02 -3.44328344e-01 -4.13030744e-01
-2.66426783e-02 1.26545519e-01 -1.19072437e+00 1.31883442e+00
3.70213866e-01 1.80824831e-01 -2.00119764e-01 1.07343435e+00
1.09333515e+00 6.12237334e-01 -8.28302503e-02 -2.32490227e-01
1.32853687e+00 -1.25536001e+00 -7.74086475e-01 4.63810079e-02
6.38529241e-01 -9.89600360e-01 9.31571186e-01 4.12800997e-01
-1.03713059e+00 -3.80855858e-01 -7.78641582e-01 -3.18277121e-01
-4.68048811e-01 2.41053738e-02 5.40964961e-01 2.85367340e-01
-1.22370756e+00 5.94510913e-01 -6.53034389e-01 -4.66547519e-01
3.35802734e-01 3.81705999e-01 -5.52219450e-01 -2.50646740e-01
-1.19562519e+00 7.99681365e-01 2.13036075e-01 8.57040063e-02
-7.95622528e-01 -6.04788125e-01 -7.91964710e-01 1.04789190e-01
6.41192198e-01 -9.85758781e-01 1.17470050e+00 -9.21618581e-01
-1.15959847e+00 7.44650602e-01 -5.57856560e-01 -3.70402247e-01
5.81063092e-01 -8.36288214e-01 -3.46502453e-01 2.29844972e-01
1.49388656e-01 7.88008869e-01 1.49967802e+00 -8.45823944e-01
-7.83631325e-01 -3.59007001e-01 -1.13628231e-01 4.89389151e-01
-5.71014225e-01 3.22475404e-01 -9.68862176e-01 -6.70938790e-01
-7.47318491e-02 -7.84162223e-01 1.28153816e-01 -6.33634478e-02
-3.91019821e-01 -5.79415619e-01 1.15938914e+00 -3.37070823e-01
1.41083241e+00 -2.13631630e+00 3.20167273e-01 2.91051716e-01
4.54108924e-01 3.77759725e-01 -3.43168557e-01 5.71354747e-01
-1.15567133e-01 1.72679946e-01 3.24400514e-01 -7.30581939e-01
5.90908378e-02 -9.74448249e-02 -3.67581040e-01 6.14561558e-01
2.64469415e-01 1.18278170e+00 -7.87116528e-01 -7.70796180e-01
3.32874149e-01 5.72291136e-01 -4.92420673e-01 8.44178945e-02
9.01878700e-02 1.28719509e-01 -8.96341920e-01 9.69983697e-01
5.16054809e-01 -1.56433746e-01 -4.04144675e-01 -3.10396045e-01
-1.45432949e-01 3.16123143e-02 -8.79291713e-01 1.66877329e+00
-1.19624607e-01 7.56218791e-01 -1.79153681e-01 -5.20334721e-01
4.75777268e-01 4.42718476e-01 6.94524586e-01 -5.96998274e-01
2.74099499e-01 1.40036762e-01 -3.94148618e-01 -5.91776431e-01
1.06091976e+00 3.55680197e-01 -1.65263072e-01 2.69079059e-01
4.16049719e-01 1.82310924e-01 4.73263323e-01 3.99627894e-01
1.00439358e+00 1.56283528e-01 3.28363031e-01 -2.37433031e-01
5.21241665e-01 -2.53035307e-01 2.42261738e-01 9.82027948e-01
-2.37447336e-01 9.99644756e-01 1.88301131e-01 -4.10403162e-01
-7.19594240e-01 -8.17154109e-01 -1.53261259e-01 1.50796473e+00
6.15769513e-02 -9.73725438e-01 -4.56590056e-01 -9.24921751e-01
1.60620824e-01 3.61182630e-01 -7.33171582e-01 -2.29355663e-01
-2.31652573e-01 -3.63007009e-01 2.73372561e-01 2.20620289e-01
7.05330726e-03 -7.83914685e-01 -4.96829480e-01 -4.02476713e-02
-2.91113496e-01 -1.10614133e+00 -8.88866961e-01 1.47596732e-01
-3.68330985e-01 -9.61217403e-01 -9.51962113e-01 -2.49368235e-01
1.50563002e-01 7.10034609e-01 1.17480588e+00 2.65641600e-01
-3.25736195e-01 9.19994771e-01 -1.02336621e+00 -3.53780776e-01
2.21439824e-01 9.97433737e-02 1.93544000e-01 3.92146796e-01
4.13951010e-01 2.51029581e-02 -5.82483172e-01 2.44947881e-01
-1.00476420e+00 -1.62614465e-01 3.15651536e-01 9.54298079e-01
6.18429303e-01 -4.49217260e-01 1.75509945e-01 -5.13171554e-01
6.63346589e-01 -6.47677958e-01 -1.66268989e-01 5.23039281e-01
-5.26187420e-01 -3.11832484e-02 1.71750590e-01 -3.24168235e-01
-8.02566230e-01 -1.51190326e-01 5.53151406e-03 -1.02367115e+00
1.58540845e-01 7.35136092e-01 -1.05968798e-02 -3.59628387e-02
2.11318076e-01 3.63202244e-01 -1.91373780e-01 -5.12509704e-01
2.87957937e-01 6.51848137e-01 1.56259164e-01 -5.09360313e-01
6.74247921e-01 4.20133382e-01 -1.93985492e-01 -9.02150512e-01
-6.51558161e-01 -9.33695734e-01 -5.83849669e-01 -3.14177811e-01
6.40285671e-01 -1.12040114e+00 -5.84403217e-01 3.46287906e-01
-9.66015697e-01 2.81094611e-01 -2.57060587e-01 4.11834419e-01
-5.03945470e-01 5.16324162e-01 -2.35284820e-01 -7.26966083e-01
-6.47598863e-01 -1.32047677e+00 1.72139227e+00 5.36237434e-02
-2.85283864e-01 -8.52038622e-01 -8.33892226e-02 1.74542546e-01
3.63738537e-01 -6.73794076e-02 4.28797960e-01 -1.06804335e+00
-3.19567144e-01 -4.86241281e-01 -1.74124613e-01 5.52710257e-02
-8.49006400e-02 3.47255945e-01 -1.15134358e+00 -6.24943018e-01
-7.16040581e-02 -5.39311767e-01 1.58374596e+00 4.32476074e-01
1.10765648e+00 -4.91067134e-02 -4.25723016e-01 5.30669868e-01
1.17795599e+00 -9.13562030e-02 4.86480474e-01 3.15716058e-01
6.80471718e-01 3.18949044e-01 1.08335757e+00 7.73048818e-01
4.00081247e-01 9.80072320e-01 2.30322078e-01 1.33625463e-01
-1.37239665e-01 9.17469040e-02 4.92706209e-01 7.45068669e-01
-4.36448911e-03 -6.89592004e-01 -7.76618958e-01 4.69244152e-01
-2.08632779e+00 -8.92382145e-01 3.40067819e-02 2.14067101e+00
3.44499320e-01 -2.00138345e-01 2.04264492e-01 6.11324422e-02
6.07804418e-01 3.58908087e-01 -2.51479894e-01 -1.99357942e-01
-2.87663430e-01 1.30364308e-02 2.27498710e-01 2.15270981e-01
-1.43047237e+00 9.50239360e-01 6.21537876e+00 1.36360168e+00
-1.23443401e+00 1.48191139e-01 4.55336273e-01 -5.17984867e-01
-2.81157166e-01 -9.39638168e-02 -9.56537366e-01 3.50063622e-01
8.37325573e-01 -2.64281511e-01 1.86987638e-01 5.55510163e-01
-3.35491598e-02 -7.11798072e-02 -1.25738609e+00 1.14320052e+00
1.63146511e-01 -1.11664760e+00 4.50640917e-01 -1.06993094e-01
4.87463921e-01 3.59159201e-01 3.60871285e-01 3.95992994e-01
-1.04748547e-01 -8.56423974e-01 9.53736901e-01 6.96974099e-01
7.48064935e-01 -5.93385220e-01 7.03080833e-01 -1.06538430e-01
-1.45148659e+00 -2.89077461e-02 -5.53080775e-02 4.94412214e-01
2.22222477e-01 4.58696097e-01 -5.10439277e-01 1.05519104e+00
7.72633672e-01 1.28349602e+00 -9.35651064e-01 1.17963052e+00
3.64252895e-01 3.89703214e-01 -4.15203571e-01 8.55579600e-02
5.39432943e-01 1.92784637e-01 8.73206854e-01 1.37832284e+00
4.42889273e-01 -3.40668082e-01 4.77710813e-01 1.16876692e-01
-1.14121109e-01 4.08579230e-01 -4.48704302e-01 -2.43626267e-01
3.38396877e-01 1.35443783e+00 -5.15240669e-01 -4.20269489e-01
-5.86873591e-01 1.02296329e+00 1.31617501e-01 4.31448996e-01
-7.57881582e-01 -2.11386159e-01 6.97891057e-01 -1.62082314e-01
5.79906762e-01 1.50133193e-01 3.39291245e-01 -1.61660469e+00
2.11897716e-01 -8.98388922e-01 5.87098002e-01 -6.01039648e-01
-1.35500908e+00 7.53442347e-01 1.33844748e-01 -1.61164904e+00
-4.01191354e-01 -4.96015936e-01 -3.37525219e-01 6.07179880e-01
-1.82422996e+00 -1.23254442e+00 -6.88528493e-02 9.07207549e-01
9.28003252e-01 -1.96854219e-01 5.87893605e-01 4.60193485e-01
-4.90848333e-01 8.48055601e-01 3.99966896e-01 -2.83267170e-01
1.03970242e+00 -8.92660022e-01 3.09317559e-01 5.52920282e-01
4.53475535e-01 6.02942109e-01 6.75265253e-01 -5.29437065e-01
-1.95238233e+00 -1.04989243e+00 8.10116053e-01 -4.77479666e-01
6.97919130e-01 -1.87643871e-01 -7.38022029e-01 4.74406093e-01
3.68598074e-01 1.32237867e-01 6.48327529e-01 1.03408962e-01
-4.89824951e-01 1.28336087e-01 -9.14668798e-01 5.72142839e-01
8.36732447e-01 -6.60473406e-01 -4.36046451e-01 2.96646118e-01
7.19501317e-01 -3.30946952e-01 -9.10183668e-01 5.72482467e-01
9.15674806e-01 -8.47437263e-01 1.00461161e+00 -6.66108668e-01
2.90088594e-01 -3.64925675e-02 -5.25119364e-01 -1.06199193e+00
-1.95132628e-01 -8.20811510e-01 -4.64746892e-01 1.24548328e+00
1.45274490e-01 -3.50147039e-01 3.24975759e-01 4.60403591e-01
-2.37753749e-01 -9.90697920e-01 -1.07160759e+00 -8.31387520e-01
-2.05061510e-01 -5.82458079e-01 3.84225637e-01 7.56728888e-01
-1.28567576e-01 2.85052836e-01 -6.93540454e-01 -4.05121386e-01
2.92854041e-01 6.87031224e-02 6.89081252e-01 -1.41649485e+00
-8.36160704e-02 -6.36800051e-01 -5.42768180e-01 -1.12492788e+00
2.10477203e-01 -1.00595903e+00 2.56234035e-03 -1.19555187e+00
4.84566361e-01 5.65542988e-02 -5.47126174e-01 5.29613972e-01
-4.37729031e-01 3.46148193e-01 5.25931358e-01 3.84474844e-01
-1.29390955e+00 7.93724537e-01 1.00250900e+00 -9.22942460e-02
1.93677187e-01 -2.89459705e-01 -6.48458958e-01 4.89551783e-01
2.38186970e-01 -3.62216532e-01 -1.75535664e-01 -4.38467532e-01
1.20355256e-01 1.29026175e-02 2.20110595e-01 -6.93214238e-01
2.43136927e-01 -9.44841746e-03 3.07313472e-01 -8.45335960e-01
6.60647750e-01 -9.37872887e-01 -1.34509742e-01 -2.85883457e-03
-3.74493182e-01 1.29125908e-01 8.12987611e-02 6.80296004e-01
-5.24029791e-01 -1.76334888e-01 1.80676579e-01 6.77921548e-02
-8.35585475e-01 6.84707463e-01 -4.07764673e-01 9.46637541e-02
8.80437255e-01 -3.22963506e-01 -2.22425893e-01 -5.71867049e-01
-6.21624708e-01 6.36545241e-01 3.54904562e-01 9.27637160e-01
9.31037903e-01 -1.36270714e+00 -9.18911755e-01 1.36558712e-01
5.76276720e-01 -5.24258733e-01 3.79676342e-01 1.09732473e+00
1.36603400e-01 7.11291790e-01 1.38533831e-01 -8.44708800e-01
-1.67420876e+00 5.55208623e-01 -1.72327563e-01 -3.94170731e-01
-3.56613755e-01 8.72974098e-01 2.19613478e-01 1.20920599e-01
1.74635485e-01 6.94769174e-02 -3.18783045e-01 6.10450625e-01
7.30625093e-01 2.59953231e-01 3.09055984e-01 -1.07035387e+00
-4.37420785e-01 6.81457996e-01 -6.36991382e-01 2.01670323e-02
1.32108569e+00 -5.37997425e-01 4.89298580e-03 5.50396919e-01
1.64390123e+00 -2.81974375e-01 -8.82378399e-01 -4.14978385e-01
1.26637936e-01 -8.44230175e-01 3.15322340e-01 -4.57389534e-01
-1.12988913e+00 6.64951265e-01 5.92088878e-01 1.58893883e-01
1.08911812e+00 2.12988760e-02 7.52386570e-01 5.83472013e-01
3.19637328e-01 -1.01635730e+00 6.33612880e-03 7.18863666e-01
9.70361412e-01 -1.45477355e+00 2.65544742e-01 -2.16273561e-01
-6.96042836e-01 1.06476045e+00 2.55355537e-01 -1.80286970e-02
6.53647423e-01 -1.12183290e-02 -2.85856485e-01 -2.84619421e-01
-1.26963878e+00 -4.20177579e-01 9.35904384e-01 6.97460771e-02
5.59533119e-01 -2.94954300e-01 -4.26410437e-01 3.27196598e-01
1.93686932e-01 -5.28922565e-02 2.05777526e-01 1.26354373e+00
-3.16417962e-01 -1.12047720e+00 -3.66755277e-01 8.76490712e-01
-8.47659528e-01 -3.90023053e-01 -4.07203525e-01 7.52283037e-01
-2.36297339e-01 9.42771435e-01 -5.56577696e-03 -6.11156106e-01
1.22019283e-01 2.95325726e-01 3.61139655e-01 -3.55233043e-01
-9.98146236e-01 5.00845432e-01 3.77321355e-02 -9.08436239e-01
-4.43426162e-01 -9.73984361e-01 -6.92270100e-01 1.59891825e-02
-6.76246643e-01 2.49641061e-01 5.56431532e-01 1.15173233e+00
6.10344470e-01 3.40239882e-01 6.58887148e-01 -1.16081643e+00
-4.28429306e-01 -8.54632735e-01 -4.72591013e-01 5.40653706e-01
6.68033600e-01 -9.38848317e-01 -3.05272251e-01 -8.60216990e-02] | [10.434393882751465, 0.8952713012695312] |
1059480c-8623-40f8-a6ac-afd4a65f8cc3 | a-multi-camera-unsupervised-domain-adaptation | null | null | https://iplab.dmi.unict.it/OBJ-MDA/ | https://www.sciencedirect.com/science/article/abs/pii/S1077314222000911?CMX_ID=&SIS_ID=&dgcid=STMJ_AUTH_SERV_PUBLISHED&utm_acid=170381815&utm_campaign=STMJ_AUTH_SERV_PUBLISHED&utm_in=DM270343&utm_medium=email&utm_source=AC_ | A Multi Camera Unsupervised Domain Adaptation Pipeline for Object Detection in Cultural Sites through Adversarial Learning and Self-Training | Object detection algorithms allow to enable many interesting applications which can be implemented in different devices, such as smartphones and wearable devices. In the context of a cultural site, implementing these algorithms in a wearable device, such as a pair of smart glasses, allow to enable the use of augmented reality (AR) to show extra information about the artworks and enrich the visitors’ experience during their tour. However, object detection algorithms require to be trained on many well annotated examples to achieve reasonable results. This brings a major limitation since the annotation process requires human supervision which makes it expensive in terms of time and costs. A possible solution to reduce these costs consist in exploiting tools to automatically generate synthetic labeled images from a 3D model of the site. However, models trained with synthetic data do not generalize on real images acquired in the target scenario in which they are supposed to be used. Furthermore, object detectors should be able to work with different wearable devices or different mobile devices, which makes generalization even harder. In this paper, we present a new dataset collected in a cultural site to study the problem of domain adaptation for object detection in the presence of multiple unlabeled target domains corresponding to different cameras and a labeled source domain obtained considering synthetic images for training purposes. We present a new domain adaptation method which outperforms current state-of-the-art approaches combining the benefits of aligning the domains at the feature and pixel level with a self-training process. | ['Giovanni Maria Farinella', 'Antonino Furnari', 'Giovanni Pasqualino'] | 2022-09-01 | null | null | null | computer-vision-and-image-understanding-cviu-1 | ['multi-target-domain-adaptation'] | ['computer-vision'] | [ 4.96545881e-01 9.97934192e-02 3.42177413e-02 -2.10152537e-01
-4.16716069e-01 -6.26952410e-01 5.15836000e-01 -2.93818675e-02
-4.73240644e-01 5.89840233e-01 -2.94030279e-01 1.38574257e-01
1.23150893e-01 -8.37600946e-01 -9.07168269e-01 -4.11796689e-01
3.48571569e-01 7.01298177e-01 5.37248075e-01 -2.22104326e-01
-1.10104017e-01 7.26166368e-01 -2.09143281e+00 4.95466977e-01
7.15177536e-01 9.10475433e-01 7.91927755e-01 3.38601261e-01
-1.67209849e-01 8.86197016e-02 -6.70603335e-01 -4.56058949e-01
5.18962920e-01 -3.01457554e-01 -3.09455097e-01 6.92803562e-01
5.29372990e-01 -2.45813131e-01 1.81056380e-01 1.07257032e+00
4.07336950e-01 1.90799758e-01 4.39335883e-01 -9.60720778e-01
-4.93545271e-02 7.25126266e-03 -3.05946648e-01 -1.15187295e-01
6.98733985e-01 1.25051871e-01 5.38652062e-01 -5.88179648e-01
8.46752942e-01 9.31729078e-01 7.22916901e-01 6.38139665e-01
-1.36466932e+00 -3.68459523e-01 8.35124925e-02 7.87837803e-02
-1.22377861e+00 -3.36862594e-01 9.39954937e-01 -2.76647389e-01
4.58716094e-01 4.11270618e-01 1.00519943e+00 1.56063366e+00
-5.07138669e-01 9.07024741e-01 1.10766733e+00 -6.85658872e-01
5.58641911e-01 8.89187932e-01 -2.21936852e-01 3.81281823e-01
4.51963097e-01 -2.65261173e-01 -2.30684027e-01 1.10075280e-01
8.02324295e-01 1.63953826e-01 -3.22574854e-01 -8.50586057e-01
-9.87634301e-01 4.43604499e-01 4.71110433e-01 7.01331556e-01
-6.18620336e-01 -3.46193522e-01 2.02564120e-01 8.28876272e-02
2.20464855e-01 4.47455227e-01 -4.77663577e-01 4.25312817e-02
-8.78290594e-01 -1.01561494e-01 6.02823079e-01 1.03987515e+00
6.98178411e-01 -1.81894898e-01 3.23748916e-01 7.42746055e-01
2.42886007e-01 3.70094806e-01 5.71858883e-01 -5.51808476e-01
3.89971942e-01 1.00056672e+00 4.63335305e-01 -5.84647477e-01
-4.75265861e-01 -4.94277537e-01 -5.03802121e-01 2.38980711e-01
7.59358466e-01 2.11423546e-01 -8.72616947e-01 1.16743636e+00
5.82013309e-01 2.93911755e-01 1.44315651e-02 9.97476518e-01
4.77718651e-01 2.91513950e-01 -8.58668834e-02 -4.30651084e-02
1.46067166e+00 -7.06779599e-01 -5.35931766e-01 -4.81665254e-01
7.69700110e-01 -6.36722803e-01 1.61505890e+00 6.74997270e-01
-7.97533035e-01 -7.09039152e-01 -1.04219651e+00 2.17861041e-01
-6.33684218e-01 5.00858188e-01 3.66076440e-01 1.07686150e+00
-5.83598733e-01 2.33873487e-01 -7.05152273e-01 -1.01024878e+00
3.12620670e-01 3.73966455e-01 -5.49497128e-01 -1.44557387e-01
-6.84812248e-01 7.90345848e-01 4.62573439e-01 5.40203378e-02
-6.48221552e-01 -2.28999868e-01 -6.29677415e-01 -2.47333758e-02
4.79927778e-01 -5.59571385e-01 1.00021517e+00 -1.36361682e+00
-1.44653678e+00 1.18492210e+00 1.88517421e-01 -4.15518999e-01
7.66028285e-01 -2.91493922e-01 -5.58722973e-01 3.04692034e-02
6.43046275e-02 3.61739367e-01 8.34367037e-01 -1.49089932e+00
-6.65259600e-01 -6.45633638e-01 3.04001063e-01 2.67878920e-01
-6.10334933e-01 -3.48046392e-01 -6.66644752e-01 -3.21096510e-01
4.07509133e-02 -1.11512351e+00 -1.56063110e-01 -1.71439964e-02
-6.59705028e-02 4.14448768e-01 9.29835618e-01 -6.21998608e-01
6.68081343e-01 -2.30573010e+00 8.88340622e-02 2.52860487e-01
-5.14470518e-01 5.08654416e-01 8.21167901e-02 7.02310055e-02
1.73332155e-01 -2.42456898e-01 -1.40605152e-01 -5.08122325e-01
-3.16237956e-01 3.48868668e-01 3.96603942e-02 2.88102478e-01
-1.69764549e-01 3.44602644e-01 -8.18441451e-01 -4.45603877e-01
5.90744495e-01 5.44153631e-01 -3.19520921e-01 1.35380924e-01
-3.20206821e-01 9.96725678e-01 -3.52662593e-01 5.73637426e-01
5.87628067e-01 -1.39672449e-02 4.29003477e-01 -1.01823822e-01
1.17560491e-01 -6.97087944e-02 -1.73119283e+00 2.04028130e+00
-1.04523551e+00 3.42584580e-01 3.97514030e-02 -9.71435785e-01
1.06132865e+00 2.75697589e-01 3.40952486e-01 -7.94655859e-01
1.66242093e-01 4.10676420e-01 -4.56118584e-01 -6.93692207e-01
4.77894753e-01 -6.31056353e-02 9.94221270e-02 1.84153512e-01
3.86768356e-02 -1.63718507e-01 1.22770585e-01 -3.47187757e-01
9.28544223e-01 4.88682985e-01 2.08463445e-01 1.89776450e-01
7.00409055e-01 5.10249920e-02 2.58081794e-01 7.01532006e-01
1.57301024e-01 7.00793564e-01 -1.72775567e-01 -7.43471444e-01
-1.11599648e+00 -8.36911082e-01 -1.03360690e-01 9.22283113e-01
3.57454419e-01 5.24577573e-02 -8.02668214e-01 -9.89098728e-01
-3.69994968e-01 8.25975180e-01 -4.65272367e-01 3.15768607e-02
-4.22482848e-01 -6.30351424e-01 1.61264926e-01 3.84214073e-01
6.00162923e-01 -1.03168321e+00 -1.04777801e+00 1.44223660e-01
-2.23257378e-01 -1.45761180e+00 8.85615572e-02 1.61454290e-01
-1.21989441e+00 -9.67312157e-01 -7.44469047e-01 -6.05816722e-01
7.85434186e-01 2.81244457e-01 9.87307429e-01 -1.73975050e-01
-2.95965314e-01 5.87120414e-01 -6.17508173e-01 -4.02939171e-01
-2.69999743e-01 6.96036071e-02 3.83143015e-02 5.28714955e-01
3.73372674e-01 -6.02599680e-01 -3.93560618e-01 6.02926910e-01
-1.11084127e+00 -8.54290798e-02 6.18460298e-01 6.63450599e-01
6.22486949e-01 3.98499519e-02 1.35017976e-01 -8.25479090e-01
2.54899681e-01 -1.65319756e-01 -6.75355732e-01 4.83073264e-01
-1.24663189e-01 1.45582790e-02 5.09748638e-01 -7.65180290e-01
-1.29005969e+00 7.66089022e-01 8.97659212e-02 -2.90000975e-01
-6.51489317e-01 -6.43899143e-02 -6.24044359e-01 -2.76094638e-02
1.15373230e+00 9.41381976e-02 -3.84300470e-01 -7.03336954e-01
2.97948688e-01 9.25946891e-01 3.39047372e-01 -3.15649211e-01
6.02180958e-01 7.32498348e-01 -7.48713091e-02 -1.03654242e+00
-7.55704045e-01 -7.75344789e-01 -8.41340601e-01 -3.79769295e-01
6.82642579e-01 -8.59597683e-01 -2.55317718e-01 8.21773112e-02
-1.08598089e+00 -2.47866407e-01 -6.56693757e-01 6.27061188e-01
-6.40304923e-01 1.91300899e-01 1.81883574e-01 -1.05191731e+00
5.94026670e-02 -9.94343519e-01 1.46173072e+00 3.20815861e-01
6.33377433e-02 -8.06844890e-01 -9.15643647e-02 5.09576976e-01
7.56564885e-02 2.22269759e-01 2.81352997e-01 -5.02275765e-01
-4.98183608e-01 -5.15195489e-01 4.85868659e-03 4.63677168e-01
1.56938031e-01 -5.60858190e-01 -1.16696465e+00 7.28469342e-03
1.92572162e-01 -6.24603685e-03 3.40458930e-01 7.16580227e-02
9.94134247e-01 -3.06707807e-02 -5.13687491e-01 3.35966587e-01
1.43832624e+00 1.96535468e-01 7.93515086e-01 6.31731272e-01
4.21820104e-01 7.03861654e-01 8.26215565e-01 2.98375219e-01
-6.13603042e-03 1.23330808e+00 5.93275726e-01 -1.04314931e-01
-1.77097753e-01 -6.49467409e-02 2.99138606e-01 -8.69886056e-02
-4.04057056e-01 -1.70470759e-01 -8.11146677e-01 6.39930785e-01
-1.82672501e+00 -8.17565262e-01 -3.96586776e-01 2.68650413e+00
2.51428664e-01 1.23923831e-01 3.86191458e-01 4.26835448e-01
8.01275611e-01 -4.65435982e-01 -3.43369663e-01 -1.87355414e-01
-5.52647747e-02 1.35259256e-01 4.59672183e-01 4.75307144e-02
-1.08259690e+00 5.41336596e-01 4.92487383e+00 4.88170296e-01
-1.02881289e+00 4.11722213e-01 1.76182374e-01 -5.96376620e-02
8.04305151e-02 -1.25196710e-01 -7.34775960e-01 4.64326978e-01
7.49678254e-01 5.16658902e-01 4.12104160e-01 1.36086845e+00
3.18353027e-02 -2.28267491e-01 -1.07732069e+00 1.26511896e+00
2.11034805e-01 -9.67417240e-01 -2.95867234e-01 1.80783167e-01
7.21375585e-01 -1.41618431e-01 -1.04103349e-01 1.11281954e-01
-2.34819680e-01 -3.62568200e-01 6.05408549e-01 5.68941414e-01
5.82135797e-01 -3.59500974e-01 8.13153267e-01 6.23157859e-01
-9.41655517e-01 -3.10621530e-01 -2.67965674e-01 1.03468224e-01
2.51669973e-01 4.73576993e-01 -1.29582119e+00 4.64558870e-01
9.62327480e-01 2.75182217e-01 -7.15395629e-01 1.21628284e+00
-1.36506721e-01 2.77275089e-02 -6.38832331e-01 -2.36536451e-02
-1.63846985e-01 -2.05332607e-01 5.01560450e-01 1.00317037e+00
8.09807658e-01 -3.57527286e-01 -6.71838000e-02 4.81630743e-01
3.71545814e-02 2.01901644e-01 -9.54168737e-01 2.82302141e-01
6.05221912e-02 1.35880947e+00 -9.90310431e-01 -2.56592005e-01
-4.78628218e-01 1.47007263e+00 -8.53275061e-02 1.18783973e-01
-7.02833533e-01 -1.88688576e-01 3.07911169e-02 8.25221360e-01
4.04164344e-01 -2.24993616e-01 -2.41981372e-01 -1.36652625e+00
3.58075082e-01 -7.74990201e-01 2.07941636e-01 -9.46462512e-01
-9.48681355e-01 6.99216127e-01 -3.63316946e-02 -1.77676809e+00
-2.37113446e-01 -7.21737027e-01 -1.10871211e-01 3.77616704e-01
-1.11279571e+00 -1.38960266e+00 -8.19907665e-01 7.00294256e-01
6.58825576e-01 -9.06976759e-02 8.47846806e-01 5.85942090e-01
-1.93972409e-01 1.87116668e-01 1.37437165e-01 -2.76002586e-01
5.33572316e-01 -1.05940521e+00 9.50135887e-02 7.16908932e-01
7.78607905e-01 2.04460636e-01 7.51161456e-01 -4.59429115e-01
-1.20558751e+00 -8.06044519e-01 3.74658257e-01 -6.86073303e-01
2.04984739e-01 -6.40850782e-01 -7.91439593e-01 6.06731892e-01
-3.43224913e-01 2.10272759e-01 5.78027546e-01 8.47093463e-02
-3.79716349e-03 -2.72847086e-01 -1.45874357e+00 4.10389632e-01
1.24951816e+00 -3.56887609e-01 -5.15116632e-01 5.70205629e-01
9.12857503e-02 -4.50151175e-01 -5.84457994e-01 3.27901334e-01
6.45609319e-01 -1.12731576e+00 8.90511572e-01 -3.01928729e-01
-2.58800000e-01 -4.71803576e-01 -2.63101637e-01 -1.06217372e+00
4.52600569e-01 -1.78255975e-01 -1.87148526e-01 1.13107884e+00
1.81444719e-01 -3.17300737e-01 1.02977669e+00 6.74571157e-01
4.98071797e-02 -1.07499085e-01 -9.53893244e-01 -1.02271795e+00
-1.00847197e+00 -5.82923889e-01 6.60290241e-01 7.13470876e-01
-4.30377930e-01 2.69635290e-01 -2.48465344e-01 3.49936515e-01
4.24547940e-01 1.53470719e-02 1.33972442e+00 -1.63120043e+00
-4.72989082e-01 1.09756656e-01 -7.52811372e-01 -8.33919466e-01
-1.81965306e-01 -4.36550289e-01 -1.16648234e-01 -1.51835525e+00
-1.74755052e-01 -5.98855495e-01 1.32701918e-01 1.83272690e-01
3.44773948e-01 6.53866351e-01 1.25985920e-01 1.93886146e-01
-6.66793406e-01 3.56165230e-01 9.29827869e-01 -5.11858948e-02
-6.57012761e-01 3.81630480e-01 -2.62040347e-01 1.00029194e+00
6.10070109e-01 -4.57761973e-01 -1.78361490e-01 -3.23922068e-01
3.65303099e-01 -1.59015566e-01 4.96083498e-01 -1.52241778e+00
-3.85213234e-02 1.69519007e-01 5.01037896e-01 -2.07935303e-01
8.18216443e-01 -1.56767356e+00 5.35918117e-01 3.81672204e-01
2.15548292e-01 -2.96090215e-01 1.71634004e-01 6.56599760e-01
-1.66130409e-01 -6.94957435e-01 5.61141312e-01 -3.63183379e-01
-8.34044755e-01 -2.93195039e-01 -7.00377598e-02 -6.46183848e-01
1.41015184e+00 -8.17610025e-01 1.90899074e-01 -2.12227926e-01
-1.13129663e+00 -3.02917212e-01 8.85592818e-01 4.25983816e-01
3.76756996e-01 -1.11981881e+00 -1.54269710e-01 3.39877129e-01
4.24413413e-01 1.80979043e-01 3.75315607e-01 5.83070576e-01
-5.22798836e-01 9.82298851e-02 -3.84083807e-01 -8.36444139e-01
-1.35106397e+00 8.43835890e-01 2.57766932e-01 -2.49058321e-01
-5.04671097e-01 3.25741798e-01 8.80952775e-02 -5.88165998e-01
2.10084096e-01 -3.01695168e-01 -2.83464104e-01 1.71841234e-01
5.12358963e-01 3.38271856e-01 5.22622168e-01 -5.95974565e-01
-2.30706826e-01 6.59595370e-01 3.10547173e-01 -1.45780399e-01
1.31056011e+00 -1.90552980e-01 5.06682575e-01 3.90598387e-01
6.60133481e-01 1.18171252e-01 -1.20783639e+00 -2.56823748e-01
3.45492996e-02 -6.50059402e-01 -6.28160387e-02 -7.45297432e-01
-8.35544169e-01 7.71958530e-01 1.20825243e+00 3.15614909e-01
1.49640441e+00 1.08016439e-01 4.04198438e-01 5.67363262e-01
9.40313339e-01 -1.26083517e+00 2.99858212e-01 -1.75229520e-01
7.97156990e-01 -1.34577799e+00 -8.67296010e-02 -3.63819391e-01
-8.06147873e-01 1.07608569e+00 4.59557414e-01 3.53483520e-02
2.10409418e-01 -1.27529308e-01 -7.86019713e-02 -1.19413249e-01
1.40400246e-01 -6.66835368e-01 2.30850160e-01 9.24135983e-01
-1.05126556e-02 -3.20326574e-02 -1.89322889e-01 3.80960405e-01
1.49857715e-01 2.64064133e-01 5.35503745e-01 8.48709464e-01
-2.11095780e-01 -1.48138380e+00 -7.85602152e-01 2.18004495e-01
-3.18322659e-01 5.08034706e-01 -3.28427851e-01 9.34620261e-01
6.26827121e-01 7.88437665e-01 -3.96558903e-02 -2.77787566e-01
7.63427794e-01 1.35577381e-01 7.08605289e-01 -8.60139728e-01
-5.55381656e-01 2.31863521e-02 5.59596345e-02 -4.02056962e-01
-8.80767822e-01 -8.37085903e-01 -7.50424266e-01 4.78235006e-01
-6.30188286e-01 -1.18751645e-01 1.29173648e+00 7.43917227e-01
3.52870971e-01 3.36278081e-01 4.40602541e-01 -1.15083742e+00
-9.70385596e-02 -9.48437512e-01 -6.71765685e-01 6.83354378e-01
-4.52299323e-03 -7.86704481e-01 -1.30130127e-02 3.91692579e-01] | [7.808539867401123, -0.8371108770370483] |
230c71ec-dc69-46ac-b717-626d101abba0 | exploring-paracrawl-for-document-level-neural | 2304.10216 | null | https://arxiv.org/abs/2304.10216v1 | https://arxiv.org/pdf/2304.10216v1.pdf | Exploring Paracrawl for Document-level Neural Machine Translation | Document-level neural machine translation (NMT) has outperformed sentence-level NMT on a number of datasets. However, document-level NMT is still not widely adopted in real-world translation systems mainly due to the lack of large-scale general-domain training data for document-level NMT. We examine the effectiveness of using Paracrawl for learning document-level translation. Paracrawl is a large-scale parallel corpus crawled from the Internet and contains data from various domains. The official Paracrawl corpus was released as parallel sentences (extracted from parallel webpages) and therefore previous works only used Paracrawl for learning sentence-level translation. In this work, we extract parallel paragraphs from Paracrawl parallel webpages using automatic sentence alignments and we use the extracted parallel paragraphs as parallel documents for training document-level translation models. We show that document-level NMT models trained with only parallel paragraphs from Paracrawl can be used to translate real documents from TED, News and Europarl, outperforming sentence-level NMT models. We also perform a targeted pronoun evaluation and show that document-level models trained with Paracrawl data can help context-aware pronoun translation. | ['Josef van Genabith', 'Jingyi Zhang', 'Yusser Al Ghussin'] | 2023-04-20 | null | null | null | null | ['nmt'] | ['computer-code'] | [ 2.93665439e-01 -1.65648416e-01 -5.81502676e-01 -1.94878101e-01
-1.62497962e+00 -8.79989624e-01 8.75431657e-01 8.46867263e-02
-6.43788338e-01 1.23187816e+00 3.74394506e-01 -8.08565795e-01
1.99633762e-01 -5.59623480e-01 -1.16588521e+00 -2.21456692e-01
4.55386072e-01 1.23067462e+00 -1.13167241e-01 -6.86604917e-01
2.33886659e-01 1.17425337e-01 -4.25183505e-01 1.00770569e+00
1.23065078e+00 -6.63540959e-02 5.81798255e-01 5.65298915e-01
-5.45726955e-01 6.82700127e-02 -8.55529904e-01 -6.32752061e-01
3.40051144e-01 -8.27204525e-01 -1.01287937e+00 -2.42983535e-01
6.88592970e-01 6.63532168e-02 1.47485703e-01 8.67984295e-01
5.25999665e-01 -3.34095359e-01 6.92962050e-01 -6.08884513e-01
-8.07327330e-01 9.76502538e-01 -4.83981460e-01 3.90795887e-01
3.72334272e-01 -1.39771208e-01 1.31696904e+00 -1.11908698e+00
1.20287287e+00 1.48103106e+00 4.48916256e-01 7.87469327e-01
-1.31093109e+00 -4.81130332e-01 -2.86161482e-01 3.04893136e-01
-9.76267099e-01 -4.64069575e-01 5.76259375e-01 -9.01306421e-02
1.74962354e+00 3.15837413e-01 4.11723852e-01 1.67248130e+00
8.73468280e-01 9.90745246e-01 1.36609435e+00 -1.15745223e+00
-1.46089733e-01 -2.86281593e-02 -1.86425984e-01 1.98051602e-01
2.29357425e-02 -1.96471885e-01 -8.00070822e-01 -6.37647063e-02
3.91614556e-01 -6.67187333e-01 -2.60561984e-02 3.81037146e-01
-1.76741588e+00 8.60285938e-01 -1.32836223e-01 7.84060240e-01
-4.22191352e-01 -2.95271516e-01 7.18867421e-01 9.67483699e-01
7.32625067e-01 7.33432651e-01 -7.87996113e-01 -2.38874331e-01
-1.02115273e+00 2.61330664e-01 9.01513278e-01 1.24449193e+00
4.68702734e-01 -1.03666216e-01 -3.22732747e-01 1.27640617e+00
-2.00341970e-01 1.28119516e+00 9.50535238e-01 -5.40724814e-01
1.44417441e+00 2.07401693e-01 3.35085467e-02 -6.13138616e-01
-8.16531479e-02 -3.53002906e-01 -6.27277076e-01 -6.01187766e-01
3.31106722e-01 -3.78655970e-01 -6.31530344e-01 1.42682660e+00
-2.71276057e-01 -7.80824363e-01 6.28577232e-01 6.96433365e-01
4.29937869e-01 1.18223810e+00 -3.45333338e-01 -4.60639715e-01
1.27538800e+00 -1.33986223e+00 -7.73353279e-01 -5.73114932e-01
1.05383623e+00 -1.49909616e+00 1.24347520e+00 1.07785627e-01
-1.23020995e+00 -5.19465268e-01 -1.02737582e+00 -1.25849649e-01
-4.59925801e-01 3.93579602e-01 1.26604393e-01 2.12514982e-01
-1.01314330e+00 4.70037431e-01 -7.52858758e-01 -8.64003718e-01
-3.83325815e-02 1.95444688e-01 -4.24813747e-01 -2.89052933e-01
-1.47493935e+00 1.44795501e+00 4.09476042e-01 -1.36220783e-01
-4.65224326e-01 -7.45267197e-02 -4.95449036e-01 -3.63525212e-01
-8.05639103e-03 -6.80470407e-01 1.46431673e+00 -1.02930844e+00
-1.53397286e+00 9.00458336e-01 -4.76837009e-01 -6.63781106e-01
4.41982597e-01 -2.06271306e-01 -4.75592464e-01 -7.16573671e-02
4.02025610e-01 5.24940312e-01 6.29782200e-01 -7.74442077e-01
-8.11139882e-01 -2.48110518e-01 -3.89122218e-01 2.75265068e-01
-4.02774215e-01 2.93701231e-01 -2.28404760e-01 -6.58368230e-01
-3.49719748e-02 -1.09013021e+00 -3.71022373e-02 -1.03402710e+00
-3.80919874e-01 -4.46917862e-01 4.98428315e-01 -1.08662617e+00
9.86238360e-01 -1.44000506e+00 4.49724436e-01 -2.83697486e-01
-4.65043545e-01 4.37888026e-01 -8.28887701e-01 1.16278195e+00
3.71561438e-01 2.00866997e-01 -6.39204457e-02 -3.81188005e-01
3.20805460e-02 5.71212053e-01 -3.87250245e-01 3.63644585e-02
4.03629780e-01 1.37026930e+00 -8.32516909e-01 -5.88135242e-01
-1.44350931e-01 1.35310128e-01 -9.14518684e-02 5.67058288e-02
-4.80369270e-01 4.94655281e-01 -2.37965032e-01 4.90761399e-01
2.32625738e-01 2.92748213e-01 3.41692954e-01 2.25516498e-01
-1.79708943e-01 9.65552211e-01 8.41735862e-04 2.01186728e+00
-1.06565130e+00 9.04120743e-01 -4.56937283e-01 -8.34872127e-01
9.46052969e-01 5.98875642e-01 -7.06344545e-02 -1.22809613e+00
-2.21180677e-01 9.27923501e-01 3.40011060e-01 -4.03377563e-01
6.03254974e-01 -2.51596987e-01 -2.63714164e-01 5.44260442e-01
2.72570252e-01 -3.34147304e-01 6.29041910e-01 -7.97441602e-02
1.07057846e+00 2.45485231e-01 1.73478067e-01 -3.43527317e-01
5.39936244e-01 7.98366964e-01 3.96424651e-01 5.94456851e-01
3.84226263e-01 4.40306902e-01 1.12536184e-01 -2.99438149e-01
-1.46200347e+00 -8.87991250e-01 1.09664993e-02 1.03480136e+00
-5.45699060e-01 -5.66497087e-01 -1.05116785e+00 -8.43428850e-01
-3.45641226e-01 1.10618007e+00 -3.40691119e-01 2.81815797e-01
-1.46104276e+00 -9.17586088e-01 5.46431541e-01 1.51469260e-01
2.64049500e-01 -1.22793341e+00 2.74673104e-01 6.00124717e-01
-9.13715482e-01 -1.40691364e+00 -9.29466784e-01 8.87951553e-02
-1.16490066e+00 -5.70316195e-01 -8.69764209e-01 -1.12858570e+00
5.45784414e-01 1.09089531e-01 1.53394544e+00 -4.11792904e-01
4.15326774e-01 -1.77621201e-01 -5.21074235e-01 -5.07735848e-01
-1.30818689e+00 8.85460138e-01 1.02625750e-01 -5.52611589e-01
7.99809933e-01 -5.44723570e-01 9.88039747e-02 1.96046486e-01
-5.44255555e-01 4.23683882e-01 1.16592634e+00 1.03987086e+00
5.62074125e-01 -6.14438415e-01 8.68860722e-01 -1.01275206e+00
1.24843860e+00 -1.27712294e-01 -4.48066831e-01 3.25235724e-01
-5.99639833e-01 1.88116238e-01 1.12040079e+00 -5.34851670e-01
-8.41408491e-01 -5.88213325e-01 -4.07887191e-01 1.78541794e-01
5.91612682e-02 7.31516063e-01 7.90566858e-03 5.21079600e-01
8.55887532e-01 7.14632094e-01 -1.59305319e-01 -6.17603838e-01
2.53208071e-01 1.05815780e+00 3.29536080e-01 -8.55257511e-01
9.13640857e-01 -2.16092452e-01 -2.86572188e-01 -6.99148595e-01
-9.42565799e-01 -1.64610282e-01 -9.78097260e-01 1.88814715e-01
6.86064363e-01 -6.70051157e-01 1.23799466e-01 -4.07973789e-02
-1.85928917e+00 -3.33005875e-01 -2.41489336e-01 5.64579248e-01
-5.68467438e-01 3.97720374e-02 -1.05451000e+00 -1.02542751e-01
-1.03791010e+00 -9.97554362e-01 1.19324720e+00 -3.38358551e-01
-6.05447292e-01 -1.26673925e+00 4.90746975e-01 5.88284016e-01
3.70258361e-01 -1.83842272e-01 1.34377956e+00 -7.88910270e-01
-2.30544209e-01 -2.65282989e-01 8.88822079e-02 6.49200201e-01
2.42224231e-01 -4.91853058e-01 -1.91813618e-01 -4.48411733e-01
-2.93736393e-03 -2.35470951e-01 3.88907194e-01 4.05187905e-02
8.89848992e-02 -6.25802040e-01 -1.05441965e-01 1.88014939e-01
1.19902635e+00 5.86265624e-02 3.30915153e-01 8.56300175e-01
5.20849347e-01 4.68305349e-01 7.20077515e-01 -4.48245168e-01
3.33447665e-01 9.07716870e-01 -2.87952036e-01 -7.39148706e-02
-5.07662773e-01 -4.88223076e-01 9.75359380e-01 2.04026055e+00
-1.21683903e-01 -2.66172260e-01 -9.49108839e-01 6.48375034e-01
-1.81126177e+00 -6.13311052e-01 -2.98805624e-01 1.90601087e+00
1.31025422e+00 2.65381157e-01 -1.80059433e-01 -4.55178142e-01
4.36732680e-01 -8.96534398e-02 -1.94491401e-01 -1.09954834e+00
-5.82231760e-01 4.10290420e-01 5.53590655e-01 6.89574301e-01
-5.70913970e-01 1.41830039e+00 5.88192320e+00 1.22641277e+00
-1.01974666e+00 6.79628372e-01 1.84513986e-01 5.18273003e-02
-3.22451860e-01 1.08083807e-01 -1.14752507e+00 3.91980916e-01
1.48650432e+00 -4.84630674e-01 4.55960572e-01 4.61928308e-01
4.85891700e-01 3.71687919e-01 -1.37201869e+00 6.97394669e-01
2.49720931e-01 -1.46665275e+00 4.96456921e-01 1.10858336e-01
1.15409577e+00 5.91066480e-01 -2.29788825e-01 6.26919508e-01
2.10742593e-01 -8.66556585e-01 5.12106240e-01 2.18127277e-02
7.17302442e-01 -7.41722584e-01 1.09940219e+00 7.79337287e-01
-4.95146662e-01 5.63698888e-01 -7.91749477e-01 -8.56492370e-02
9.16597694e-02 3.43440741e-01 -1.22815943e+00 9.26598251e-01
1.22372225e-01 7.77154744e-01 -5.51055670e-01 1.95802286e-01
-4.86684769e-01 1.01945078e+00 -9.41572040e-02 -3.92249584e-01
6.70844138e-01 -4.14961338e-01 7.82655120e-01 1.74357891e+00
5.25864542e-01 -6.88272834e-01 5.65867275e-02 4.77624685e-01
-4.00155246e-01 8.01858008e-01 -5.05968451e-01 -2.79371589e-01
2.19004363e-01 1.01482379e+00 -2.08592430e-01 -3.90161395e-01
-5.89264512e-01 1.46478760e+00 4.06308919e-01 3.96054745e-01
-3.29714358e-01 -2.32345358e-01 3.75730753e-01 9.25218761e-02
-3.80098224e-02 -6.51438713e-01 -1.46048188e-01 -1.33653724e+00
3.75768602e-01 -1.42600048e+00 5.75949438e-03 -5.33423066e-01
-1.57183695e+00 1.21255314e+00 -1.73633948e-01 -1.42976499e+00
-7.12046325e-01 -6.05511189e-01 -3.77827972e-01 1.32893407e+00
-1.44459534e+00 -1.51481199e+00 6.10727847e-01 3.04264218e-01
1.29474485e+00 -6.93093836e-01 1.24284494e+00 1.29466310e-01
-3.24722230e-01 5.97013474e-01 7.50619531e-01 4.15926903e-01
1.05396473e+00 -1.24721777e+00 1.08825660e+00 8.44488978e-01
6.30888939e-01 7.39469409e-01 8.55741799e-01 -6.61495090e-01
-1.65814519e+00 -1.18983042e+00 1.96671629e+00 -7.58067787e-01
9.09370124e-01 -6.49194300e-01 -5.55904210e-01 7.75170624e-01
9.21604574e-01 -7.72582889e-01 5.43306172e-01 1.21517673e-01
-9.18294191e-02 -8.33358243e-02 -6.83867812e-01 1.00091493e+00
8.29481721e-01 -4.44986612e-01 -1.13716829e+00 9.47102606e-01
9.14175332e-01 -3.98985922e-01 -9.81481850e-01 1.93352953e-01
3.48078638e-01 -1.71044365e-01 5.32808185e-01 -8.02904367e-01
8.91392171e-01 6.65380061e-02 -2.77157515e-01 -1.94516206e+00
1.10597480e-02 -5.91231227e-01 1.02327116e-01 1.12921226e+00
1.00474274e+00 -7.90389955e-01 4.09724027e-01 -3.66397649e-01
-2.36555174e-01 -6.83826566e-01 -1.20950925e+00 -1.29346848e+00
1.03701663e+00 -1.73177928e-01 3.38974506e-01 8.63736093e-01
1.34136215e-01 1.21581113e+00 -3.37358177e-01 -3.80260438e-01
4.27163541e-01 4.30047303e-01 8.13692987e-01 -6.42254233e-01
-3.90819758e-01 -3.52890223e-01 1.20887145e-01 -1.19288433e+00
4.72316891e-01 -1.48268771e+00 -3.81035320e-02 -1.91089737e+00
2.08944365e-01 -3.47783752e-02 2.38326341e-01 4.57988501e-01
6.18725382e-02 3.31630379e-01 1.50539592e-01 6.48089349e-01
-1.47275433e-01 4.12055105e-01 1.68204141e+00 -4.00197595e-01
-2.35639215e-02 -5.46709411e-02 -1.80408925e-01 3.32807779e-01
9.44508255e-01 -7.88433194e-01 -3.84957753e-02 -9.98557150e-01
1.71324477e-01 1.81209520e-01 -5.11879027e-01 -6.17097914e-01
5.17582800e-03 -3.18816096e-01 2.51001030e-01 -6.15954280e-01
1.26266122e-01 -4.12995994e-01 -2.89885402e-01 4.02527988e-01
-4.79550779e-01 5.32002389e-01 3.02863002e-01 -1.02802627e-01
-4.31484252e-01 -3.22324157e-01 3.71219784e-01 -3.68433267e-01
-4.49139178e-02 -1.09319001e-01 -7.48340666e-01 2.73876786e-01
2.85704523e-01 5.60115539e-02 -4.24146354e-01 -2.57990807e-01
-2.27734417e-01 2.85164658e-02 4.71374393e-01 8.64994943e-01
2.88523823e-01 -1.32959425e+00 -1.48238873e+00 1.41101805e-02
1.26659408e-01 -4.43878472e-01 -4.39947367e-01 8.73784244e-01
-4.43049371e-01 1.08728564e+00 -3.18125457e-01 -7.31324852e-01
-1.19176161e+00 2.71658659e-01 -4.01201695e-02 -7.74461210e-01
-5.81326783e-01 2.99087971e-01 -4.87293035e-01 -1.00300944e+00
-4.42733079e-01 -1.85592383e-01 2.74871469e-01 -2.87693411e-01
3.70567262e-01 -9.95023996e-02 6.02819860e-01 -7.22802579e-01
-1.47799859e-02 3.95143986e-01 -5.55855095e-01 -6.26494765e-01
1.30965805e+00 -1.91601768e-01 -5.51555216e-01 5.92567742e-01
1.20484650e+00 2.79321074e-01 -3.72167945e-01 -3.68168682e-01
4.20292169e-01 -7.40934536e-02 -5.41465819e-01 -1.03347301e+00
-3.38909537e-01 9.59744573e-01 6.34278730e-02 -3.61666948e-01
8.36154819e-01 -1.52162045e-01 1.41335225e+00 9.67661560e-01
7.06121445e-01 -1.37739646e+00 -1.78067222e-01 1.09987462e+00
1.07326448e+00 -1.12300324e+00 -4.50707316e-01 -9.92681645e-03
-3.30888778e-01 1.21947443e+00 3.23621422e-01 5.20058610e-02
-1.46261856e-01 2.69774139e-01 5.48777223e-01 3.86251748e-01
-1.03978264e+00 3.18443775e-01 3.90256852e-01 4.76140171e-01
8.07489514e-01 1.81348577e-01 -1.18006361e+00 4.06554163e-01
-7.37326205e-01 -3.35098565e-01 5.61132133e-01 7.35418379e-01
-2.36577526e-01 -2.04496241e+00 -4.10726935e-01 4.12082583e-01
-6.22001588e-01 -7.04582393e-01 -5.98773599e-01 7.01483071e-01
-2.64463842e-01 1.03430438e+00 -2.42388934e-01 -2.42851719e-01
3.49780977e-01 4.31161255e-01 7.42355943e-01 -8.36691439e-01
-6.99619591e-01 2.47766674e-01 5.39311945e-01 -7.91490003e-02
-2.27393180e-01 -6.21280432e-01 -8.70848477e-01 -2.75374144e-01
-5.90596907e-03 6.60184264e-01 1.00816739e+00 1.30184114e+00
1.62623242e-01 2.84613669e-01 3.80955040e-01 -6.44363999e-01
-6.53230846e-01 -1.62972736e+00 4.88786064e-02 2.26172253e-01
-7.75483921e-02 2.27825969e-01 -7.85370320e-02 1.21486753e-01] | [11.580273628234863, 10.360323905944824] |
83379da4-dc38-4b00-9b4b-d5850cc1c5eb | unified-discrete-diffusion-for-simultaneous | 2211.14842 | null | https://arxiv.org/abs/2211.14842v1 | https://arxiv.org/pdf/2211.14842v1.pdf | Unified Discrete Diffusion for Simultaneous Vision-Language Generation | The recently developed discrete diffusion models perform extraordinarily well in the text-to-image task, showing significant promise for handling the multi-modality signals. In this work, we harness these traits and present a unified multimodal generation model that can conduct both the "modality translation" and "multi-modality generation" tasks using a single model, performing text-based, image-based, and even vision-language simultaneous generation. Specifically, we unify the discrete diffusion process for multimodal signals by proposing a unified transition matrix. Moreover, we design a mutual attention module with fused embedding layer and a unified objective function to emphasise the inter-modal linkages, which are vital for multi-modality generation. Extensive experiments indicate that our proposed method can perform comparably to the state-of-the-art solutions in various generation tasks. | ['Ponnuthurai N. Suganthan', 'DaCheng Tao', 'Zuopeng Yang', 'Chaoyue Wang', 'Tat-Jen Cham', 'Heliang Zheng', 'Chuanxia Zheng', 'Minghui Hu'] | 2022-11-27 | null | null | null | null | ['multimodal-generation'] | ['natural-language-processing'] | [ 3.54827225e-01 -1.25998193e-02 -8.99769459e-03 -1.85038671e-01
-1.15805507e+00 -3.48423749e-01 1.29382586e+00 -4.43249077e-01
-1.94822803e-01 6.88603759e-01 4.57205504e-01 -1.21641517e-01
-1.43241752e-02 -6.82634711e-01 -6.14888966e-01 -8.88033628e-01
4.88154978e-01 3.39367718e-01 -1.83785900e-01 -3.96215886e-01
-4.76613231e-02 4.96201217e-02 -1.10835755e+00 4.06813949e-01
9.67163801e-01 8.90173972e-01 2.67591417e-01 5.43551505e-01
-6.06830046e-02 7.92316020e-01 -3.10964465e-01 -9.13548470e-01
-2.47538928e-02 -6.85445070e-01 -6.55249119e-01 2.37303779e-01
3.63906711e-01 -3.42015475e-01 -5.07630646e-01 9.07162249e-01
9.02297974e-01 7.09091276e-02 9.90208268e-01 -1.28318954e+00
-1.53110564e+00 8.16234291e-01 -7.64385104e-01 -1.76553234e-01
6.52965128e-01 2.12820172e-01 1.06528640e+00 -1.16235375e+00
8.17522585e-01 1.28495657e+00 3.30386639e-01 7.39289463e-01
-1.39126563e+00 -4.50590760e-01 1.11302428e-01 7.41948783e-02
-1.11059868e+00 -5.04770219e-01 1.13716638e+00 -5.47404230e-01
6.54275835e-01 1.11352593e-01 4.57517713e-01 1.67893791e+00
2.83324927e-01 1.08758998e+00 1.08348680e+00 -4.07004952e-01
-2.59471953e-01 1.05723664e-01 -3.20578009e-01 7.11857498e-01
-1.75046727e-01 -1.41271399e-02 -7.75056779e-01 9.03342143e-02
7.68462956e-01 -1.57041714e-01 -2.35997692e-01 -1.65406913e-01
-1.94846153e+00 8.94018590e-01 2.16767564e-01 4.47602689e-01
-5.69351196e-01 3.20808321e-01 2.37414151e-01 2.92857379e-01
5.65195203e-01 2.28753075e-01 2.36039802e-01 3.08855586e-02
-9.59769607e-01 6.46503568e-02 3.33024442e-01 1.04947579e+00
4.68064547e-01 4.43236768e-01 -6.19918525e-01 8.90438437e-01
5.47008514e-01 7.66376257e-01 5.77298403e-01 -7.12374210e-01
6.44359171e-01 3.37812960e-01 -8.67949128e-02 -1.00373936e+00
-2.38758475e-01 -3.48680049e-01 -1.36301446e+00 -1.61194056e-01
3.98962246e-03 -3.63090247e-01 -1.00464952e+00 2.04367375e+00
2.03126594e-01 4.27313522e-02 3.49574178e-01 9.19901431e-01
1.03230715e+00 1.00651920e+00 1.35949641e-01 -1.40592054e-01
1.21993613e+00 -1.16518760e+00 -9.28412318e-01 -1.89394690e-02
7.29333386e-02 -1.00378418e+00 9.48201239e-01 -2.02312637e-02
-1.35949552e+00 -8.18259060e-01 -8.09645593e-01 -2.03559145e-01
-3.31407428e-01 4.30927575e-01 6.62730813e-01 3.51562023e-01
-1.29972756e+00 1.15424663e-01 -4.84427631e-01 -2.41487786e-01
1.35626048e-02 1.26485631e-01 -5.77409744e-01 -7.44430870e-02
-1.45429432e+00 8.68598878e-01 3.26921463e-01 3.12423617e-01
-1.03080797e+00 -4.66306776e-01 -9.77104127e-01 -3.02743822e-01
-4.93076444e-03 -1.46111608e+00 9.02296424e-01 -8.00571263e-01
-1.72899902e+00 7.38404989e-01 -6.83670193e-02 -1.46301612e-01
5.83123386e-01 3.83155383e-02 -5.02526462e-01 2.07150102e-01
1.27814516e-01 1.16256070e+00 1.28062105e+00 -1.44600272e+00
-3.31197500e-01 -1.95611551e-01 -6.31352589e-02 4.36087430e-01
-6.96314037e-01 -1.45698667e-01 -6.18190050e-01 -1.02077913e+00
4.01823819e-02 -9.54009831e-01 -8.42915401e-02 -2.61141926e-01
-6.63579166e-01 -6.75202236e-02 6.36560977e-01 -9.08252478e-01
1.19593513e+00 -2.19113684e+00 9.48293149e-01 -4.96732816e-02
2.83727348e-01 -2.73451060e-01 -5.94701350e-01 7.35360086e-01
5.11874221e-02 -7.37887248e-02 -3.28399330e-01 -8.25490594e-01
3.89512390e-01 1.08701177e-02 -2.37805709e-01 1.64350972e-01
4.25047874e-01 1.24272346e+00 -7.40451038e-01 -5.32296836e-01
9.09535736e-02 7.25947857e-01 -2.57066786e-01 2.60926217e-01
-1.01131588e-01 7.55682826e-01 -3.21210802e-01 6.96635425e-01
6.01231813e-01 -2.49921381e-01 3.32193933e-02 -5.37453353e-01
5.06624356e-02 -4.50347900e-01 -9.43247378e-01 2.21057010e+00
-4.33849037e-01 4.49628055e-01 8.78207535e-02 -6.57989621e-01
7.62157559e-01 5.62452197e-01 4.52080041e-01 -5.97662210e-01
5.99091984e-02 2.31735870e-01 -1.21333256e-01 -3.94106746e-01
8.85402262e-01 -4.34674293e-01 -2.50746638e-01 5.16110837e-01
5.75543821e-01 -2.76042104e-01 3.48801970e-01 4.29718196e-01
6.08092427e-01 3.02211016e-01 -3.31602812e-01 7.55341724e-02
5.10623097e-01 -3.46589446e-01 1.15106162e-02 5.92153132e-01
1.97987080e-01 7.93539524e-01 2.75907189e-01 1.14276670e-01
-1.06554556e+00 -1.27508831e+00 2.99179584e-01 9.88130212e-01
1.65972710e-01 2.86930660e-03 -6.42067730e-01 -4.99336839e-01
-2.02412903e-01 6.28592193e-01 -6.04137599e-01 -2.70217538e-01
-4.47741188e-02 -1.08287382e+00 7.13438749e-01 4.31441516e-01
5.08098185e-01 -9.20886099e-01 2.19914973e-01 2.15186179e-01
-5.94269872e-01 -1.21565187e+00 -6.48754120e-01 -2.80130774e-01
-6.52162552e-01 -4.24344420e-01 -1.57281315e+00 -8.73659253e-01
5.58321118e-01 1.57573909e-01 8.63260865e-01 -6.00644767e-01
8.42222199e-02 7.23929942e-01 -3.63637000e-01 7.92467967e-02
-5.75359941e-01 1.90174088e-01 -1.01192214e-01 6.02574468e-01
-2.44325653e-01 -6.17711067e-01 -5.17264724e-01 -2.09240858e-02
-1.31707597e+00 6.07218921e-01 8.55643630e-01 1.11069953e+00
4.28817570e-01 -1.84418842e-01 9.21321869e-01 -4.18299824e-01
1.32799602e+00 -5.61542988e-01 -7.49373436e-02 4.78645563e-01
-5.24232090e-01 1.02084905e-01 4.05037165e-01 -6.05452597e-01
-1.48387706e+00 1.51837543e-02 -1.18857697e-01 -4.76333082e-01
9.44184512e-02 7.95652211e-01 -1.84358433e-01 6.46207407e-02
3.09477717e-01 7.33143985e-01 8.90911818e-02 -4.21257108e-01
9.50464189e-01 5.22054911e-01 6.13116801e-01 -6.22203946e-01
8.66273999e-01 3.57137799e-01 -7.12750629e-02 -6.05656683e-01
-3.93364012e-01 -4.85604778e-02 -5.35743237e-01 -2.96702772e-01
1.11611962e+00 -1.22517633e+00 -1.78229570e-01 8.91801953e-01
-1.36075306e+00 -8.93256143e-02 -2.04239964e-01 4.82866496e-01
-5.57817101e-01 5.37064254e-01 -9.81673896e-01 -6.32437468e-01
-5.30322492e-01 -1.24746251e+00 1.36064076e+00 2.45070770e-01
1.01055145e-01 -1.31269038e+00 2.16743335e-01 5.86215198e-01
5.78884721e-01 1.56279713e-01 9.14488494e-01 -2.40779445e-01
-5.04617572e-01 -2.75991522e-02 -4.07203078e-01 3.43429685e-01
-4.98872623e-03 4.48725373e-03 -8.60842586e-01 -1.51175246e-01
-3.69595081e-01 -2.81720221e-01 1.07363009e+00 2.28478938e-01
5.22956431e-01 9.14695710e-02 -1.01035982e-01 3.84698659e-01
1.21840739e+00 2.14499030e-02 6.19482517e-01 3.36322300e-02
9.78437603e-01 4.95353639e-01 1.43754885e-01 3.97889912e-01
9.03762996e-01 6.03261948e-01 2.36761898e-01 -6.13258243e-01
-4.81209308e-01 -2.88928151e-01 5.36205530e-01 1.28704584e+00
-7.33188614e-02 -5.29560208e-01 -6.42188907e-01 6.89606845e-01
-2.01214194e+00 -1.07907903e+00 -1.91627190e-01 1.70134366e+00
9.06521022e-01 -3.34387988e-01 3.97374146e-02 -2.93183476e-01
6.41472757e-01 2.98260570e-01 -3.47915590e-01 -1.54270291e-01
-6.21791720e-01 -1.76215753e-01 -6.62250370e-02 4.25449759e-01
-1.11350441e+00 1.04825509e+00 6.83132744e+00 9.83007312e-01
-1.08725059e+00 2.95972526e-01 5.44226110e-01 1.25309542e-01
-8.55254412e-01 -3.42553705e-01 -4.97396320e-01 4.22030151e-01
6.76145196e-01 -4.95464057e-02 4.22540575e-01 2.36049697e-01
1.39737085e-01 7.25874081e-02 -9.90318239e-01 1.08809507e+00
3.54185313e-01 -1.31330240e+00 6.21046126e-01 1.10968715e-02
1.17651463e+00 -2.20505625e-01 5.31637073e-01 1.69171229e-01
9.85699967e-02 -7.94149458e-01 8.72781873e-01 9.08540666e-01
8.57430458e-01 -3.80099714e-01 5.06961226e-01 1.96989000e-01
-1.07119274e+00 1.01171404e-01 1.00694589e-01 4.70301449e-01
7.48864830e-01 6.81837916e-01 -2.79652715e-01 1.07133746e+00
1.12320349e-01 8.00374269e-01 -6.42115533e-01 6.58891857e-01
-9.33077261e-02 1.70383602e-01 -3.01583875e-02 2.85549611e-01
3.47122043e-01 -3.81665409e-01 6.00299299e-01 1.27048898e+00
7.36121476e-01 -3.96556199e-01 -2.69575231e-02 1.09791791e+00
-1.46278366e-01 7.38282800e-02 -7.14478195e-01 -4.84655917e-01
-3.31482738e-02 1.33966076e+00 -3.71246099e-01 -4.86528099e-01
-5.04607856e-01 1.57445288e+00 4.66547124e-02 6.17260873e-01
-1.05558479e+00 -2.47162864e-01 2.94440538e-01 -4.50981557e-01
1.66633397e-01 -3.24415833e-01 -1.20101571e-01 -1.52400112e+00
1.09889889e-02 -7.96193004e-01 1.03646077e-01 -1.15314293e+00
-1.59028447e+00 8.85630906e-01 4.83183488e-02 -1.16542780e+00
-4.63138044e-01 -4.35864747e-01 -4.40927118e-01 9.13262069e-01
-1.64688492e+00 -2.06972313e+00 -3.96721736e-02 8.96130025e-01
4.02582586e-01 -3.80436569e-01 7.16657341e-01 7.71235645e-01
-6.11353874e-01 6.46976709e-01 1.48253366e-01 -6.59053028e-02
8.18680823e-01 -1.07837129e+00 2.24454001e-01 9.74193454e-01
3.09182912e-01 5.16863286e-01 4.10636127e-01 -6.03032291e-01
-1.77135468e+00 -9.20836508e-01 7.81399310e-01 -2.12451130e-01
8.47537279e-01 -1.72658339e-01 -4.33096051e-01 5.11280596e-01
7.79788077e-01 -5.39517879e-01 7.23150313e-01 -1.90662965e-01
-2.67726332e-01 5.21718897e-02 -7.60987997e-01 8.57805252e-01
9.72266257e-01 -7.27790892e-01 -3.98714751e-01 2.26845741e-01
9.02738392e-01 -3.90525311e-01 -1.05486071e+00 3.15615714e-01
5.61691165e-01 -6.07175171e-01 1.06250668e+00 -3.03148121e-01
9.10534680e-01 -2.64485180e-01 -3.38243723e-01 -1.49867618e+00
-4.92088825e-01 -7.81500995e-01 -2.87783653e-01 1.58805442e+00
8.03115964e-01 -4.58120018e-01 1.69669315e-01 3.40590686e-01
-1.70632973e-01 -5.17477453e-01 -8.52725804e-01 -4.36622024e-01
4.28511985e-02 -2.53150940e-01 4.85612869e-01 9.43134606e-01
-1.55241638e-01 6.83853030e-01 -1.01592016e+00 -8.49886090e-02
5.87516487e-01 1.68594018e-01 6.11180127e-01 -8.49815965e-01
-3.43867630e-01 -6.50646329e-01 -9.29072350e-02 -1.06189084e+00
2.38485500e-01 -9.20830369e-01 -1.85637578e-01 -1.85523593e+00
4.20742661e-01 -2.38911156e-02 -3.95630151e-01 3.05780977e-01
-1.68780625e-01 5.53769052e-01 4.88444269e-01 1.47716105e-01
-5.67627132e-01 1.00813890e+00 1.74252009e+00 -4.29236859e-01
-1.02208462e-02 -3.52888316e-01 -9.43324745e-01 2.31614619e-01
3.51241708e-01 3.37600987e-03 -5.10926366e-01 -8.96803796e-01
3.06688786e-01 2.73311913e-01 3.07174176e-01 -7.49146760e-01
2.41307035e-01 -7.12810308e-02 2.59561688e-01 -3.89203817e-01
7.95184910e-01 -6.89832151e-01 3.98200154e-01 2.65696757e-02
-5.05720496e-01 1.54415727e-01 -2.44709812e-02 6.60605073e-01
-5.35087466e-01 1.29509911e-01 4.03263897e-01 1.23213433e-01
-7.11211860e-01 4.21875268e-01 -3.88799518e-01 -1.95324019e-01
9.15012121e-01 -1.05969146e-01 -3.12259823e-01 -7.07708061e-01
-8.98901463e-01 7.23940581e-02 2.14045793e-01 7.47032404e-01
8.56126249e-01 -1.85657084e+00 -1.02238655e+00 2.00661644e-01
1.67052075e-01 -5.36329210e-01 5.67709506e-01 1.20392370e+00
-4.44462933e-02 1.88684881e-01 -1.71462208e-01 -4.74271685e-01
-8.63094866e-01 3.91660213e-01 1.24037646e-01 -3.94539744e-01
-2.75006086e-01 7.72503316e-01 2.44257286e-01 -5.06837964e-01
-9.61942375e-02 1.99725591e-02 -1.34870216e-01 3.45067561e-01
2.79961407e-01 2.22592175e-01 -3.29247653e-01 -9.59281445e-01
-6.01651296e-02 6.40939593e-01 1.34924546e-01 -6.75498724e-01
1.18894267e+00 -3.78889680e-01 -2.64782101e-01 4.77692932e-01
1.08171880e+00 -1.47378385e-01 -1.03911674e+00 -2.92440236e-01
-4.65588093e-01 -5.66366315e-02 1.39523625e-01 -9.69973266e-01
-1.20231414e+00 1.05997121e+00 4.92137104e-01 2.46559054e-01
1.31083632e+00 -1.86659858e-01 1.01945722e+00 3.75335943e-03
1.80905476e-01 -1.03705251e+00 3.66499275e-01 4.06913519e-01
1.04307199e+00 -1.34119499e+00 -2.63212770e-01 -1.08929023e-01
-1.13083398e+00 8.34683716e-01 4.32070017e-01 2.96213448e-01
3.53685975e-01 3.18001658e-02 -8.07592422e-02 -6.66994974e-02
-7.91504800e-01 -2.12287679e-01 8.25323164e-01 6.31306708e-01
6.14258647e-01 1.46977469e-01 -3.55415434e-01 4.79782850e-01
1.36891812e-01 -7.01090023e-02 2.96304941e-01 5.20233810e-01
7.19833933e-03 -1.34905875e+00 -4.02936459e-01 5.32896072e-03
-2.13852182e-01 -3.19064379e-01 -4.76616830e-01 5.03199995e-01
-3.70038003e-02 9.46902037e-01 -2.99496233e-01 -6.24275684e-01
1.09668367e-01 2.22145289e-01 6.39121354e-01 -2.12268174e-01
-5.93825221e-01 5.47459543e-01 3.56294252e-02 -2.49080911e-01
-7.18556106e-01 -4.20575708e-01 -7.28872538e-01 -2.08023340e-01
-2.90454119e-01 -2.25219131e-01 7.09407568e-01 8.38753164e-01
7.24565685e-01 6.88657463e-01 5.61107874e-01 -1.07115090e+00
-3.81168753e-01 -1.09015167e+00 -5.03676713e-01 5.15697539e-01
2.01799273e-01 -4.51134264e-01 -6.59028664e-02 2.75203645e-01] | [11.316173553466797, 0.3204880654811859] |
fb8a1405-555a-43e1-a70c-b6278780c0af | few-shot-learning-with-siamese-networks-and | null | null | https://openreview.net/forum?id=za_XIJLkkB8 | https://openreview.net/pdf?id=za_XIJLkkB8 | Few-Shot Learning with Siamese Networks and Label Tuning | We study the problem of building text classifiers with little or no training data, commonly known as zero and few-shot text classification.
In recent years, an approach based on neural textual entailment models has been found to give strong results on a diverse range of tasks.
In this work, we show that with proper pre-training, Siamese networks that embed texts and labels are
a competitive alternative.
These models allow for a large reduction in inference cost: constant in the number of labels rather than linear.
Furthermore, we introduce label tuning, a simple and computationally efficient approach that allows to adapt the models in a few-shot setup by only changing the label embeddings.
While giving lower performance than model fine-tuning, this approach has the architectural advantage that a single encoder can be shared by many different tasks. | ['Anonymous'] | 2021-10-16 | null | null | null | acl-arr-october-2021-10 | ['few-shot-text-classification'] | ['natural-language-processing'] | [ 2.94546634e-01 4.74499464e-02 -3.83081466e-01 -6.65387869e-01
-7.87398100e-01 -4.90491807e-01 8.88801038e-01 4.90731567e-01
-9.19426978e-01 5.99722683e-01 1.35094956e-01 -1.65867895e-01
4.97572161e-02 -6.24046445e-01 -5.57527959e-01 -5.53739965e-01
3.15511405e-01 5.81581354e-01 3.36294770e-01 -1.43158570e-01
3.79875988e-01 8.20970684e-02 -1.66650426e+00 2.96294242e-01
4.90017563e-01 8.90193880e-01 -5.92564009e-02 8.86297345e-01
-3.17248791e-01 1.12175906e+00 -4.00332510e-01 -9.27430809e-01
6.26954064e-02 -4.00176287e-01 -1.02291358e+00 -1.17970780e-02
5.54658771e-01 -3.09966981e-01 -3.11600864e-01 8.93887341e-01
4.74416345e-01 4.90191132e-01 8.85356963e-01 -1.02064157e+00
-6.88885152e-01 9.35265243e-01 -3.51459742e-01 1.38349891e-01
2.76683830e-03 -1.82739571e-01 1.50774717e+00 -9.59567010e-01
6.01036429e-01 1.07747757e+00 8.54261994e-01 5.72268665e-01
-1.56851447e+00 -3.42042565e-01 8.86443034e-02 3.13593924e-01
-1.14870405e+00 -5.42719364e-01 5.37753761e-01 -3.42644632e-01
1.26724720e+00 9.70051959e-02 2.49942362e-01 1.26888299e+00
6.05792813e-02 6.96907163e-01 8.43661487e-01 -8.16552222e-01
4.24051821e-01 3.23494196e-01 4.25872475e-01 7.60459602e-01
2.16827646e-01 -3.22942019e-01 -4.29225415e-01 -1.76934555e-01
-4.49530873e-03 1.62240252e-01 2.16225218e-02 -4.99083817e-01
-8.73057246e-01 1.29792166e+00 3.05182606e-01 5.37666500e-01
1.06222011e-01 4.60353613e-01 8.87598574e-01 5.39059043e-01
8.79389822e-01 7.09866464e-01 -5.22774279e-01 -1.43595949e-01
-1.23066652e+00 1.00378163e-01 1.07541633e+00 7.63114691e-01
6.96476519e-01 -1.63153052e-01 -3.23705226e-01 8.58494878e-01
6.70681754e-03 3.18382010e-02 8.56951654e-01 -8.85353625e-01
4.21141952e-01 1.49007380e-01 6.03520088e-02 -6.51227891e-01
-4.56206232e-01 -2.62522906e-01 -5.65530956e-01 1.39920816e-01
4.34657902e-01 -1.99473023e-01 -9.56762075e-01 1.81906545e+00
9.83872190e-02 -1.83240871e-03 1.85008813e-02 3.32726806e-01
2.80592233e-01 6.46547616e-01 1.86748862e-01 -1.22081518e-01
1.35066199e+00 -1.27751851e+00 -7.95712888e-01 -4.75735039e-01
1.24469507e+00 -6.90321386e-01 1.28664863e+00 2.77826399e-01
-1.07646787e+00 -2.32023686e-01 -1.20389628e+00 -4.84781623e-01
-7.46399701e-01 -1.56499103e-01 4.82072949e-01 7.12502480e-01
-8.12689126e-01 9.74860191e-01 -7.33857155e-01 -5.65096259e-01
4.24806565e-01 2.16929689e-01 -2.46216193e-01 -2.91808814e-01
-1.29851842e+00 1.32799768e+00 4.98144031e-01 -5.00911832e-01
-7.26057827e-01 -5.87551832e-01 -9.53210831e-01 4.40492928e-01
6.28039777e-01 -4.97178346e-01 1.65283918e+00 -9.39103186e-01
-1.62392044e+00 9.83216643e-01 -1.90976217e-01 -6.71373904e-01
5.55744529e-01 -8.19591135e-02 4.99030575e-02 -1.47540309e-02
-7.08704367e-02 4.09862757e-01 8.14599991e-01 -6.89774215e-01
-5.91590941e-01 -2.47507542e-01 1.02928847e-01 1.61341533e-01
-7.83329129e-01 2.36825585e-01 -1.76016644e-01 -5.87988138e-01
-3.48590940e-01 -9.10565972e-01 -1.13180749e-01 1.93447486e-01
-1.09128281e-01 -6.08293295e-01 4.25795376e-01 -2.97593445e-01
1.19328129e+00 -2.00399327e+00 2.06947014e-01 -2.19099492e-01
1.06947012e-01 3.69856536e-01 -1.11513376e-01 5.76377809e-01
7.39338696e-02 3.75624955e-01 -3.42542470e-01 -8.02307606e-01
3.90725344e-01 2.86122113e-01 -1.35363266e-01 3.67662400e-01
2.17081070e-01 1.00961506e+00 -7.66288280e-01 -5.27593553e-01
1.78522598e-02 2.04096809e-01 -5.99376500e-01 8.05848539e-02
-3.74502152e-01 -3.41260374e-01 -5.31684272e-02 -4.20804136e-02
2.09372833e-01 -5.75494289e-01 2.07175314e-01 2.45658785e-01
1.43770784e-01 4.79364753e-01 -1.10486042e+00 1.91163242e+00
-6.99807882e-01 8.42617571e-01 -1.18395887e-01 -1.34451568e+00
4.92446274e-01 5.56711435e-01 8.02185982e-02 -3.78612787e-01
3.89667034e-01 1.69036105e-01 -2.07267568e-01 -6.07713461e-01
4.00880992e-01 -6.57785654e-01 -2.84036428e-01 8.90051186e-01
5.38041115e-01 1.46378614e-02 4.69752938e-01 2.04032257e-01
1.16958666e+00 -9.69985351e-02 5.56562603e-01 -8.57362151e-02
2.82811105e-01 -4.57295403e-02 9.71109048e-02 8.80714536e-01
-6.53934032e-02 3.97976816e-01 4.94740725e-01 -3.45845222e-01
-1.43040919e+00 -5.45931339e-01 -2.81521171e-01 1.83732116e+00
-2.18322679e-01 -5.22641659e-01 -7.67201483e-01 -6.10400259e-01
-7.89042190e-02 9.57951844e-01 -9.41780448e-01 -4.79946792e-01
-5.59905887e-01 -7.55782545e-01 6.45777762e-01 6.02837563e-01
-7.41257798e-03 -8.37031424e-01 -6.15997016e-01 2.35428661e-01
-1.39997657e-02 -1.03597450e+00 -4.96122301e-01 9.24573541e-01
-7.86411464e-01 -6.27607286e-01 -7.45449483e-01 -9.33475077e-01
4.21937108e-01 1.54586956e-01 1.15990412e+00 -2.84849890e-02
-3.82790804e-01 2.29369570e-02 -4.76254344e-01 -2.73922622e-01
-4.03081834e-01 6.37167633e-01 -1.85124189e-01 -1.07230082e-01
6.34788394e-01 -2.76534647e-01 -1.23299211e-01 -5.71407489e-02
-1.04950881e+00 -1.06546380e-01 3.63242388e-01 1.18303573e+00
1.10671207e-01 -8.44971240e-02 6.12213016e-01 -1.37214649e+00
7.11996675e-01 -5.29312611e-01 -2.83552945e-01 3.56439710e-01
-9.80967224e-01 4.97574329e-01 1.01320481e+00 -5.92120290e-01
-1.04085934e+00 -2.55820695e-02 -1.10568225e-01 -1.72428608e-01
-1.72464266e-01 5.57288170e-01 2.45847493e-01 4.70713377e-02
9.81324375e-01 -3.49898264e-02 -8.24689791e-02 -4.90915000e-01
7.89541364e-01 7.75799692e-01 1.63953871e-01 -3.68216395e-01
5.42946637e-01 3.86707246e-01 -1.52390003e-01 -5.71753979e-01
-1.47267485e+00 -5.45884430e-01 -8.42440367e-01 3.13264370e-01
8.80457163e-01 -6.82104886e-01 -2.17451170e-01 1.52124450e-01
-1.02680290e+00 -5.00402629e-01 -5.15594780e-01 4.32296902e-01
-6.30586088e-01 3.45828593e-01 -9.43801343e-01 -5.89945257e-01
-2.77664691e-01 -7.47750759e-01 7.46346176e-01 -7.40861893e-02
-3.94018620e-01 -1.21857369e+00 1.75572425e-01 -1.15760379e-02
5.46357870e-01 -1.87988192e-01 1.14129496e+00 -1.20168579e+00
-4.05127443e-02 -5.71459711e-01 -1.07742898e-01 4.50586528e-01
-1.06021762e-01 -1.36196882e-01 -1.28498316e+00 -2.93525696e-01
1.29682451e-01 -8.22329581e-01 1.04571736e+00 1.01544507e-01
9.33301568e-01 -2.60556996e-01 -1.58957452e-01 4.51250106e-01
1.49285948e+00 -8.48422348e-02 3.00176829e-01 2.06070185e-01
6.05754912e-01 6.25050128e-01 3.94596905e-01 4.24103528e-01
1.72481045e-01 5.83603442e-01 1.85954105e-02 9.87938643e-02
-4.35666926e-02 -1.89616174e-01 1.85305893e-01 8.23624611e-01
3.25624496e-01 -4.46568191e-01 -7.97240853e-01 5.09628236e-01
-1.98119509e+00 -1.05375385e+00 1.90559357e-01 2.12552905e+00
1.07815361e+00 3.11108351e-01 2.43541133e-02 2.91669130e-01
5.95466495e-01 1.06927611e-01 -3.69604975e-01 -6.62840009e-01
7.87844509e-02 6.05476201e-01 5.51544011e-01 6.68657303e-01
-1.17153156e+00 9.28199291e-01 6.99319506e+00 8.29523742e-01
-9.05821085e-01 5.25390923e-01 3.32712442e-01 -5.79514682e-01
-8.51494148e-02 1.55735752e-02 -8.13798130e-01 5.49072266e-01
1.42942607e+00 -3.73159677e-01 3.98416489e-01 8.98865402e-01
-2.82550424e-01 2.71242671e-02 -1.55783486e+00 6.77129209e-01
5.38745821e-01 -1.20215034e+00 -1.40222460e-01 -1.87656194e-01
8.42508972e-01 1.04490183e-01 -2.38505021e-01 5.81795394e-01
5.34212649e-01 -8.88069510e-01 6.23778999e-01 2.44553432e-01
7.04712868e-01 -6.91255569e-01 9.19557571e-01 5.87403893e-01
-7.91159928e-01 -2.45690525e-01 -5.95091283e-01 -2.08942726e-01
1.20115951e-01 3.16756666e-01 -6.06408596e-01 8.28036293e-02
3.81442875e-01 5.31711459e-01 -5.37984610e-01 8.74271989e-01
-2.38382712e-01 6.12320185e-01 -2.32946083e-01 -3.63220692e-01
4.45634782e-01 2.77588099e-01 4.69230376e-02 1.55145228e+00
2.02481821e-01 -1.98797479e-01 1.91467762e-01 4.78273690e-01
-3.69682878e-01 3.24815214e-01 -7.10815310e-01 2.44837496e-02
4.07546818e-01 1.13692594e+00 -6.30829453e-01 -6.64527714e-01
-6.50288224e-01 1.20674217e+00 8.15266371e-01 9.12172869e-02
-7.38941729e-01 -8.57200503e-01 2.00379357e-01 -6.03644317e-03
6.54715300e-01 -6.22354150e-02 -1.94012254e-01 -1.30548930e+00
-9.72146541e-02 -5.33423722e-01 4.52378839e-01 -4.71832544e-01
-1.48461986e+00 3.79341185e-01 -5.91570549e-02 -7.59211123e-01
-5.54370344e-01 -8.78806055e-01 -4.17123109e-01 6.13978267e-01
-1.54213393e+00 -8.95485759e-01 1.75581332e-02 2.49618113e-01
7.34847307e-01 3.67006138e-02 1.11156201e+00 3.67795885e-01
-5.85653126e-01 7.25831866e-01 5.68951428e-01 1.06714264e-01
9.45444107e-01 -1.42838550e+00 3.46152097e-01 5.89875400e-01
3.78373265e-01 4.37924087e-01 8.14298749e-01 -1.17461078e-01
-8.18848193e-01 -9.85501766e-01 1.41382039e+00 -4.89785761e-01
9.12546515e-01 -6.62474573e-01 -1.08573270e+00 8.94759655e-01
3.87986124e-01 5.80060259e-02 1.01407731e+00 4.06856596e-01
-6.51271880e-01 7.48028159e-02 -9.73844171e-01 5.01328349e-01
7.45104313e-01 -8.81431878e-01 -8.45257401e-01 5.75638950e-01
6.24413788e-01 6.86979964e-02 -7.19413459e-01 -1.08897530e-01
4.09278870e-01 -6.05448246e-01 6.24070287e-01 -9.07181799e-01
6.61473155e-01 1.95265621e-01 -1.37844056e-01 -1.40746450e+00
-4.25992906e-01 -3.66923809e-01 -3.13677609e-01 1.17195189e+00
5.84084213e-01 -4.84134257e-01 7.85405099e-01 7.25054801e-01
-5.59494365e-03 -6.43020988e-01 -8.70196581e-01 -9.27112341e-01
5.64438641e-01 -3.03034008e-01 7.73782283e-02 1.04906487e+00
4.59446520e-01 1.01413751e+00 -5.92011929e-01 -5.14757931e-01
4.09998953e-01 -1.38437241e-01 3.78204912e-01 -1.58449543e+00
-4.22816277e-01 -5.08578479e-01 -2.43273318e-01 -8.53254557e-01
4.91342813e-01 -1.35743797e+00 2.63071120e-01 -1.37175608e+00
4.39713329e-01 -2.39906356e-01 -4.15840536e-01 5.83582640e-01
-2.29961023e-01 3.08402061e-01 1.17416322e-01 1.40265107e-01
-7.32858896e-01 3.73888969e-01 6.66616678e-01 -1.43575877e-01
2.49746203e-01 -1.71104401e-01 -6.22323573e-01 7.03982770e-01
8.50090206e-01 -9.24700797e-01 -4.05606151e-01 -5.93746066e-01
4.47743088e-01 -2.94396460e-01 1.61802256e-03 -8.87153327e-01
4.23224032e-01 1.89435765e-01 1.79512665e-01 -1.12514846e-01
4.35977995e-01 -8.01262021e-01 -4.71716613e-01 4.01593387e-01
-1.13002646e+00 -1.77992955e-01 -1.08694263e-01 7.21961796e-01
-7.68645257e-02 -1.21629071e+00 1.12639678e+00 -3.43234748e-01
-4.26186591e-01 8.13911110e-02 -5.47874331e-01 2.50555903e-01
1.12648976e+00 2.60090828e-02 -2.18730405e-01 -2.86661416e-01
-7.24951029e-01 5.94235957e-03 4.24624175e-01 2.48962030e-01
2.49765422e-02 -1.22124660e+00 -6.01939797e-01 -1.14187971e-01
3.28982025e-01 -3.62187505e-01 -6.46526646e-03 5.26156604e-01
-2.17299432e-01 5.30541420e-01 8.58896226e-02 -2.87204236e-01
-1.13642848e+00 9.12549078e-01 2.63092518e-01 -5.96325040e-01
-7.28081465e-01 8.84539485e-01 -6.89630434e-02 -5.70021391e-01
3.67648661e-01 -1.69166803e-01 -4.78153564e-02 4.91900325e-01
7.26931095e-01 2.68246114e-01 2.67418325e-01 -2.87235290e-01
-5.24218157e-02 4.26575422e-01 -3.94451916e-01 -2.02642888e-01
1.19036579e+00 -1.57196656e-01 1.72378957e-01 1.08635855e+00
1.51619673e+00 -4.04118955e-01 -1.12537611e+00 -6.44759595e-01
4.45387006e-01 -2.24376515e-01 1.90829709e-01 -6.49148047e-01
-5.50317824e-01 1.26449060e+00 3.74760330e-01 5.01757681e-01
5.69801629e-01 -4.98299906e-03 6.91357195e-01 8.78748953e-01
9.24591646e-02 -1.45354021e+00 2.98910677e-01 7.57476151e-01
2.28319377e-01 -1.38427305e+00 -2.39940435e-02 2.47540474e-02
-6.78447306e-01 1.09226573e+00 2.84577131e-01 -1.59498066e-01
6.31228626e-01 4.11075443e-01 -1.73484460e-01 -5.80685921e-02
-1.19608402e+00 -2.36381724e-01 5.33188246e-02 2.28240550e-01
6.61415577e-01 -1.32390693e-01 -2.97494948e-01 3.82655144e-01
-3.36230658e-02 3.03018540e-02 5.00984490e-01 9.92441714e-01
-7.96684384e-01 -1.08545280e+00 1.85022309e-01 6.99453175e-01
-8.25479209e-01 -3.59098315e-01 -2.53111750e-01 6.37432456e-01
-7.00355694e-02 8.68247867e-01 2.52396345e-01 -7.74625465e-02
1.04954965e-01 9.57499683e-01 4.52804148e-01 -1.04753578e+00
-6.85024798e-01 -2.88831949e-01 2.14885339e-01 -1.00402899e-01
-2.47746691e-01 -3.67242634e-01 -1.05471826e+00 -3.46091002e-01
-5.98548949e-01 2.21166253e-01 6.32212818e-01 1.18705070e+00
9.95070115e-02 4.94909912e-01 5.22045672e-01 -7.39993989e-01
-1.23655331e+00 -1.15189445e+00 -7.11522937e-01 5.67112565e-01
2.76525587e-01 -4.78154123e-01 -6.47296309e-01 1.68801054e-01] | [10.68674087524414, 7.891918182373047] |
93b1bbc9-7a08-4744-ba45-7373ece3b9e0 | ai-based-software-for-lung-nodule-detection | 2206.10912 | null | https://arxiv.org/abs/2206.10912v1 | https://arxiv.org/pdf/2206.10912v1.pdf | AI-based software for lung nodule detection in chest X-rays -- Time for a second reader approach? | Objectives: To compare artificial intelligence (AI) as a second reader in detecting lung nodules on chest X-rays (CXR) versus radiologists of two binational institutions, and to evaluate AI performance when using two different modes: automated versus assisted (additional remote radiologist review). Methods: The CXR public database (n = 247) of the Japanese Society of Radiological Technology with various types and sizes of lung nodules was analyzed. Eight radiologists evaluated the CXR images with regard to the presence of lung nodules and nodule conspicuity. After radiologist review, the AI software processed and flagged the CXR with the highest probability of missed nodules. The calculated accuracy metrics were the area under the curve (AUC), sensitivity, specificity, F1 score, false negative case number (FN), and the effect of different AI modes (automated/assisted) on the accuracy of nodule detection. Results: For radiologists, the average AUC value was 0.77 $\pm$ 0.07, while the average FN was 52.63 $\pm$ 17.53 (all studies) and 32 $\pm$ 11.59 (studies containing a nodule of malignant etiology = 32% rate of missed malignant nodules). Both AI modes -- automated and assisted -- produced an average increase in sensitivity (by 14% and 12%) and of F1-score (5% and 6%) and a decrease in specificity (by 10% and 3%, respectively). Conclusions: Both AI modes flagged the pulmonary nodules missed by radiologists in a significant number of cases. AI as a second reader has a high potential to improve diagnostic accuracy and radiology workflow. AI might detect certain pulmonary nodules earlier than radiologists, with a potentially significant impact on patient outcomes. | ['Dirk Pickuth', 'Jūratė Dementavičienė', 'Artūras Samuilis', 'Jonas Ražanskas', 'Jonas Bialopetravičius', 'Vytautas Naujalis', 'Neringa Bielskienė', 'Darius Barušauskas', 'Emil Johnson Jeyakumar', 'Sebastian Huettinger', 'Naglis Ramanauskas', 'Susanne Ohlmann-Knafo'] | 2022-06-22 | null | null | null | null | ['lung-nodule-detection'] | ['medical'] | [ 1.97610676e-01 4.64771032e-01 -1.83792278e-01 1.24761440e-01
-1.00413954e+00 -6.69533432e-01 2.46653140e-01 2.67803937e-01
-6.20100021e-01 4.32660401e-01 -4.13505621e-02 -1.00461853e+00
-4.56804633e-01 -6.99347138e-01 -4.63532180e-01 -6.67331040e-01
-3.85599174e-02 8.91789615e-01 5.99333346e-01 6.12493575e-01
-1.08299606e-01 6.65264487e-01 -8.75942707e-01 2.89360315e-01
6.99165106e-01 9.36535656e-01 5.71061134e-01 9.16100800e-01
2.00041190e-01 1.21193516e+00 -4.96586740e-01 -2.52660096e-01
6.26031220e-01 -7.91712701e-01 -5.26434362e-01 1.51354507e-01
3.87245230e-02 -4.81934577e-01 -2.88676530e-01 4.82126802e-01
3.89980465e-01 -2.87898540e-01 9.93385315e-01 -6.16021693e-01
-2.71109372e-01 6.13913298e-01 -5.61439931e-01 6.82331204e-01
-1.04073763e-01 6.34225845e-01 5.46631277e-01 -5.87354124e-01
2.85231054e-01 3.46684009e-01 7.39613295e-01 4.37818855e-01
-7.53410816e-01 -4.78508919e-01 -5.50514460e-01 -1.17598616e-01
-1.24552286e+00 1.89266592e-01 -2.43742317e-01 -7.56161928e-01
5.78137577e-01 7.72835791e-01 9.10966277e-01 8.77217762e-03
4.14471805e-01 4.84498441e-02 8.99623871e-01 -3.36769819e-01
-4.79985811e-02 4.14298654e-01 -1.23192340e-01 8.91073167e-01
9.56912398e-01 8.66903514e-02 3.70609134e-01 -5.49349546e-01
1.12171721e+00 2.28207842e-01 -3.39699537e-01 3.01565472e-02
-1.71944165e+00 7.55486250e-01 4.55382973e-01 5.28586149e-01
-6.23815060e-01 -6.43871948e-02 2.14149475e-01 -1.49540901e-01
-2.50905961e-01 9.09565091e-01 -4.71605547e-02 6.91612661e-02
-6.31397069e-01 -2.71069586e-01 6.82075083e-01 4.31155056e-01
-7.64540285e-02 -1.27115697e-01 -4.86319572e-01 7.56221652e-01
1.36883065e-01 8.41407776e-01 8.73870909e-01 -7.91144073e-01
3.70080350e-03 7.20368564e-01 1.45819232e-01 -6.61798179e-01
-5.83584607e-01 -4.90715146e-01 -8.63741815e-01 -1.97236702e-01
4.88750488e-01 -2.73943722e-01 -1.17120135e+00 9.46717143e-01
-1.98909178e-01 -4.42702860e-01 -2.10307762e-01 9.86814022e-01
6.17160439e-01 3.20087433e-01 2.45331615e-01 -4.51287180e-01
1.77763045e+00 -8.55982363e-01 -3.37965101e-01 -5.77705763e-02
9.15802836e-01 -6.97516322e-01 8.21686924e-01 9.74900946e-02
-1.28228259e+00 -3.80031645e-01 -7.24951267e-01 7.90862978e-01
2.51025766e-01 5.64581811e-01 -4.35961671e-02 8.17271948e-01
-8.22034717e-01 3.05803984e-01 -1.05842197e+00 -4.33612972e-01
2.07100660e-01 5.55240452e-01 -4.92787920e-02 1.80800900e-01
-6.34942949e-01 1.24851370e+00 2.85890937e-01 -2.00554207e-01
-6.08807564e-01 -7.74803162e-01 -2.44413882e-01 1.73834190e-01
7.28283584e-01 -9.48774159e-01 1.55754447e+00 -7.33210146e-01
-6.82783127e-01 1.00637186e+00 7.45782331e-02 -6.24180198e-01
7.22580075e-01 1.49634287e-01 -2.38184854e-01 6.07958972e-01
3.87419641e-01 2.99090445e-01 3.60935092e-01 -9.25304890e-01
-9.34383214e-01 -3.22893053e-01 -3.28633398e-01 3.82702291e-01
2.00475022e-01 -9.90828425e-02 -2.55223721e-01 -4.99555022e-01
1.82176188e-01 -1.29836476e+00 -5.59934914e-01 7.16086179e-02
1.73359718e-02 -1.71380728e-01 4.96255398e-01 -7.83801079e-01
1.19925559e+00 -1.99300659e+00 -7.28828609e-01 4.76365834e-01
8.41104746e-01 5.50388336e-01 6.11699462e-01 -1.65446296e-01
-1.61552638e-01 5.42200625e-01 -2.67579198e-01 8.24300945e-01
-5.61255693e-01 -1.18310995e-01 4.67948377e-01 2.23919749e-01
1.45224422e-01 1.02172339e+00 -8.27827811e-01 -7.06519783e-01
4.86502379e-01 1.01462744e-01 -2.49306768e-01 2.04048067e-01
4.14163828e-01 3.88774008e-01 -3.08032960e-01 4.68024582e-01
1.89091042e-01 -1.00345814e+00 4.41979468e-01 1.11115023e-01
-2.12524399e-01 9.61309448e-02 -9.95545685e-01 4.55040812e-01
-2.78051704e-01 4.58494723e-01 -1.69485942e-01 -2.53929287e-01
7.87094951e-01 8.08305740e-01 6.54945672e-01 -3.42804790e-01
1.08026385e-01 5.86926877e-01 9.35828686e-01 -5.20099759e-01
-2.62088567e-01 -3.19479555e-01 4.47545677e-01 7.48754561e-01
-5.10209978e-01 -2.99772680e-01 1.52193904e-01 1.95589840e-01
1.41542745e+00 -8.97345483e-01 8.22665989e-01 -3.16425383e-01
4.96692032e-01 6.12897635e-01 -4.55974787e-02 1.43542278e+00
-2.86800981e-01 7.57554352e-01 4.47764874e-01 -3.59950632e-01
-8.94763291e-01 -1.18850267e+00 -3.63726765e-01 5.26519477e-01
-1.56871885e-01 2.05492690e-01 -4.14240092e-01 -7.91784585e-01
-6.51958957e-02 6.43158138e-01 -4.66949135e-01 -1.57707646e-01
-6.17355466e-01 -6.36310220e-01 3.69100362e-01 7.56017745e-01
3.34468395e-01 -1.03243828e+00 -1.23210979e+00 1.43302664e-01
-2.35135496e-01 -6.21197402e-01 -3.34420651e-01 1.52050957e-01
-8.68129015e-01 -1.46753335e+00 -1.28947854e+00 -6.15080655e-01
8.24344277e-01 4.62657720e-01 1.13007379e+00 5.57099640e-01
-7.05820560e-01 6.27263606e-01 -2.43713021e-01 -6.03759170e-01
-7.30868697e-01 -1.55053899e-01 -2.33944863e-01 -5.01581669e-01
2.03425631e-01 1.68783411e-01 -9.51224923e-01 6.39189124e-01
-8.29491735e-01 -1.24680609e-01 1.21135950e+00 7.32337713e-01
6.12191558e-01 -2.36971736e-01 4.28909928e-01 -1.09941423e+00
2.42342427e-01 -5.55470169e-01 -3.99216682e-01 1.80764690e-01
-7.45384216e-01 -3.65290225e-01 3.92558903e-01 -3.61453176e-01
-7.85463631e-01 -3.70723121e-02 5.23857713e-01 -2.82829285e-01
-8.20439234e-02 4.86239940e-01 8.94117713e-01 3.69678028e-02
1.02133346e+00 2.42023766e-02 2.18342096e-01 5.46431867e-03
-4.70702529e-01 5.91490149e-01 3.88770729e-01 1.20314114e-01
8.58708143e-01 2.66220063e-01 2.02757090e-01 -5.06970823e-01
-6.47022963e-01 -8.75150144e-01 -4.88819391e-01 -2.17973009e-01
9.67905700e-01 -7.05434382e-01 -2.46150285e-01 -2.38277420e-01
-4.50819731e-01 -5.30026220e-02 -5.80999970e-01 1.27365482e+00
-2.20732152e-01 3.51873755e-01 -3.06667924e-01 -5.60608089e-01
-7.08253324e-01 -1.42787647e+00 4.40927118e-01 1.81665242e-01
-5.89534461e-01 -6.69882774e-01 -1.44920051e-01 2.21476942e-01
6.16051376e-01 1.35108903e-01 1.06341863e+00 -1.16665542e+00
-5.59980273e-01 -5.11107802e-01 -6.29945695e-01 1.24925196e-01
6.23204708e-01 7.86041394e-02 -1.77481517e-01 8.29874501e-02
3.30930620e-01 2.33247042e-01 4.09777641e-01 9.13060784e-01
1.05594707e+00 -2.62881547e-01 -6.30218148e-01 3.06565333e-02
1.36670899e+00 1.14818311e+00 2.67289639e-01 2.33986944e-01
3.36122334e-01 4.88140620e-02 4.87976044e-01 2.42647275e-01
-3.34863633e-01 9.44303796e-02 2.52985477e-01 -1.83201775e-01
-1.80255368e-01 1.12341091e-01 -3.21535230e-01 4.39407945e-01
-5.09488106e-01 -6.98898658e-02 -1.62136817e+00 7.18378842e-01
-1.16226017e+00 -7.00563490e-01 -5.18699765e-01 2.17994142e+00
3.68584692e-01 2.70623893e-01 3.34399670e-01 -7.97501579e-02
1.03059089e+00 -3.38142723e-01 -3.06377828e-01 -1.95181221e-01
4.15525436e-01 1.60578325e-01 8.00851405e-01 1.18112713e-02
-6.18375242e-01 -1.39837295e-01 6.51901102e+00 2.85550505e-01
-1.11041737e+00 -3.36544365e-02 8.75814259e-01 -1.22861557e-01
1.63274884e-01 -1.07356176e-01 -3.53240490e-01 4.54716086e-01
9.26468432e-01 -5.50669372e-01 -1.11459650e-01 8.43078613e-01
6.73212409e-02 -6.74392700e-01 -8.32109034e-01 5.15820861e-01
-5.06905802e-02 -1.49004328e+00 -1.87958814e-02 2.46547386e-01
5.29817760e-01 3.71633992e-02 5.65998964e-02 6.19930550e-02
5.07133193e-02 -8.91887724e-01 2.09568053e-01 4.23900008e-01
1.19154918e+00 -2.08930433e-01 1.29355788e+00 2.89688587e-01
-1.05814171e+00 1.06347919e-01 2.07940042e-02 1.84578821e-02
-1.36183739e-01 2.60893196e-01 -2.02434564e+00 2.62120277e-01
7.25039184e-01 -1.63327381e-01 -7.39957154e-01 1.14753807e+00
1.09657012e-01 9.55900550e-01 -6.51285648e-01 -3.01919550e-01
3.34759951e-01 1.41835913e-01 5.36477029e-01 1.04093790e+00
3.05870712e-01 6.83969438e-01 -2.09475800e-01 5.35332263e-01
4.14477699e-02 2.59558469e-01 -3.86246651e-01 -1.53676838e-01
6.53759003e-01 1.17731822e+00 -1.24885547e+00 -7.04808891e-01
-6.81409240e-01 2.03764856e-01 -4.60477322e-01 -2.48017564e-01
-1.01763785e+00 -2.99110562e-01 -6.10281706e-01 9.44604576e-01
1.24282390e-02 5.70115864e-01 -5.95915198e-01 -5.52758873e-01
-1.31667241e-01 -7.30491757e-01 7.50044048e-01 -7.43780732e-01
-8.10159683e-01 5.45177996e-01 -2.50796409e-04 -1.29798174e+00
-4.19900417e-01 -6.22987509e-01 -5.34408212e-01 8.36323440e-01
-6.88651085e-01 -4.87286240e-01 -5.44533372e-01 5.35460236e-03
3.30116838e-01 -1.28836602e-01 6.03933334e-01 -1.74634337e-01
-5.31287007e-02 4.35937703e-01 1.17116198e-01 3.89807194e-01
6.25392318e-01 -1.28811467e+00 -2.79716313e-01 3.04640830e-01
-5.55456996e-01 4.37337905e-01 2.02972665e-02 -7.84287333e-01
-8.69616389e-01 -1.08237553e+00 6.94873035e-01 -6.15232527e-01
4.15051818e-01 8.30283344e-01 -6.38628662e-01 8.11703146e-01
-2.80105382e-01 -2.09115684e-01 8.38461518e-01 -5.30545235e-01
2.53322691e-01 2.70911455e-01 -1.33096123e+00 6.59973800e-01
4.50398296e-01 8.98132622e-02 -5.99501729e-01 4.40334171e-01
2.39613548e-01 -3.21795076e-01 -1.24422467e+00 8.68828297e-01
7.25685716e-01 -9.93366063e-01 8.73434007e-01 -1.23239711e-01
3.81283730e-01 -1.14335209e-01 3.28015476e-01 -5.75513065e-01
-8.36276352e-01 7.90808816e-03 3.47482204e-01 1.72943443e-01
8.38702440e-01 -7.67975807e-01 7.25545645e-01 7.75865495e-01
-4.73285355e-02 -9.96546328e-01 -7.32115149e-01 -3.01124990e-01
-1.34214193e-01 1.77036926e-01 5.37090935e-02 6.14378333e-01
-1.69623926e-01 -1.28938526e-01 4.17760104e-01 1.27382860e-01
1.00525118e-01 -2.61582434e-01 4.09465909e-01 -1.03907454e+00
-5.12184322e-01 -6.58311844e-01 -2.50599205e-01 -2.26742625e-01
-1.00020325e+00 -8.50338638e-01 -3.26791525e-01 -1.89990425e+00
8.12768698e-01 -2.98991501e-01 -1.92449629e-01 4.62391764e-01
-4.85036343e-01 9.74722281e-02 4.17203784e-01 8.84267390e-01
-1.04940727e-01 -5.74945629e-01 1.32956994e+00 2.52380729e-01
-3.01878810e-01 5.32956839e-01 -6.63250327e-01 8.82700861e-01
8.47439170e-01 -7.65839219e-01 -1.26282528e-01 -2.10275784e-01
-2.40930334e-01 5.73879004e-01 4.91056353e-01 -1.22771728e+00
8.00151154e-02 -2.68712193e-01 7.61950195e-01 -8.22576344e-01
-1.14848286e-01 -9.15377915e-01 5.50507367e-01 1.51406169e+00
-9.67120826e-02 2.41566524e-01 1.44752100e-01 3.79476577e-01
1.31443560e-01 -5.95350981e-01 9.89509284e-01 -7.49711812e-01
-3.47171910e-02 -1.12106018e-01 -8.70386899e-01 -3.59201164e-04
1.43241775e+00 -5.50626278e-01 -1.66369364e-01 -4.46224421e-01
-9.70009446e-01 -1.58431023e-01 2.93320507e-01 -2.60360450e-01
2.67827332e-01 -7.95041144e-01 -7.92562306e-01 -5.08253463e-02
2.28262320e-02 1.77490830e-01 1.98265716e-01 1.53967416e+00
-1.18660796e+00 7.63464808e-01 8.24906304e-02 -8.04606378e-01
-1.20000350e+00 2.34076053e-01 6.93913817e-01 -7.40508080e-01
-3.77576768e-01 7.32988954e-01 4.19975013e-01 -3.60775739e-02
-1.62174553e-02 -2.31444851e-01 2.57062912e-01 -2.39786223e-01
3.04235488e-01 7.29813278e-01 2.13874713e-01 -2.08947033e-01
-4.41352785e-01 4.26895022e-01 -3.24498385e-01 -7.70543292e-02
6.97310865e-01 1.62242875e-01 1.93638653e-02 4.17639464e-01
6.05412900e-01 7.09328055e-02 -4.85111326e-01 -1.68914095e-01
-1.39044657e-01 -2.38373265e-01 -2.26871282e-01 -1.34828973e+00
-6.85589135e-01 3.88281822e-01 8.21607232e-01 5.15258193e-01
1.00590324e+00 4.83071655e-01 2.46128261e-01 3.25987250e-01
-1.20710306e-01 -4.99512315e-01 7.76454136e-02 -2.66409963e-02
6.48637533e-01 -1.18539619e+00 2.18465954e-01 -2.79397696e-01
-9.88107979e-01 1.02109194e+00 6.46417022e-01 -1.30229890e-01
6.49776757e-01 9.15825292e-02 3.03865045e-01 -3.38767350e-01
-5.25269210e-01 1.01436283e-02 3.55600417e-01 3.00979644e-01
5.69240212e-01 4.58941817e-01 -7.57534206e-01 6.82314992e-01
4.24867356e-03 -3.62738669e-02 6.32499933e-01 1.12357688e+00
-7.84104347e-01 -2.52225220e-01 -7.15910017e-01 1.42725456e+00
-7.37035215e-01 1.14146441e-01 -5.67038774e-01 1.31091869e+00
8.76597390e-02 5.47317922e-01 2.37661839e-01 6.01477250e-02
3.44862968e-01 8.38488434e-03 1.99789047e-01 -8.52848589e-01
-9.79716420e-01 3.71965021e-01 -2.52554137e-02 1.57695845e-01
-2.68096000e-01 -6.08890653e-01 -1.41688240e+00 3.44002135e-02
-8.69547188e-01 3.82930607e-01 3.61512721e-01 5.80529451e-01
1.40372485e-01 7.14355707e-01 5.08485913e-01 -3.08805089e-02
-7.15810716e-01 -1.06611156e+00 -5.16653478e-01 1.43055255e-02
1.28408000e-01 -3.42390537e-01 -7.12917089e-01 9.91270244e-02] | [15.383484840393066, -2.0883796215057373] |
142ee38c-c394-40a8-9e77-d127430bf9cd | the-effects-of-noisy-labels-on-deep | 1706.02361 | null | http://arxiv.org/abs/1706.02361v3 | http://arxiv.org/pdf/1706.02361v3.pdf | The Effects of Noisy Labels on Deep Convolutional Neural Networks for Music Tagging | Deep neural networks (DNN) have been successfully applied to music
classification including music tagging. However, there are several open
questions regarding the training, evaluation, and analysis of DNNs. In this
article, we investigate specific aspects of neural networks, the effects of
noisy labels, to deepen our understanding of their properties. We analyse and
(re-)validate a large music tagging dataset to investigate the reliability of
training and evaluation. Using a trained network, we compute label vector
similarities which is compared to groundtruth similarity.
The results highlight several important aspects of music tagging and neural
networks. We show that networks can be effective despite relatively large error
rates in groundtruth datasets, while conjecturing that label noise can be the
cause of varying tag-wise performance differences. Lastly, the analysis of our
trained network provides valuable insight into the relationships between music
tags. These results highlight the benefit of using data-driven methods to
address automatic music tagging. | ['Kyunghyun Cho', 'Keunwoo Choi', 'Mark Sandler', 'George Fazekas'] | 2017-06-07 | null | null | null | null | ['music-classification'] | ['music'] | [ 3.91667545e-01 -1.63973123e-01 -1.12137780e-01 -3.22090775e-01
-6.85023367e-01 -9.48537946e-01 2.76873112e-01 1.32379413e-01
-5.86931884e-01 3.23284119e-01 4.28107053e-01 -6.08338267e-02
-4.03146625e-01 -3.97072285e-01 -5.25441587e-01 -5.20423532e-01
-2.28471130e-01 3.31038266e-01 -1.22151703e-01 -8.19360018e-02
2.28432894e-01 1.32552236e-01 -1.53884876e+00 3.71587783e-01
1.68387294e-01 9.46970820e-01 6.59136623e-02 4.67955530e-01
2.24577963e-01 7.49146640e-01 -8.88484478e-01 -4.49350059e-01
1.32662162e-01 -4.44365591e-01 -9.08150256e-01 -1.30800039e-01
8.28159809e-01 -1.00614734e-01 -6.61791638e-02 1.15057015e+00
7.86663115e-01 3.59085888e-01 4.28183347e-01 -1.09787440e+00
-4.95869607e-01 1.24210298e+00 1.16421692e-01 2.21735194e-01
-1.24929257e-01 -1.89562172e-01 1.49986196e+00 -2.96091139e-01
4.53392237e-01 7.78174341e-01 1.29034460e+00 5.02187788e-01
-1.10472083e+00 -8.93955112e-01 -2.32054383e-01 1.92207471e-02
-1.41684890e+00 -6.93931401e-01 6.49995863e-01 -6.51349604e-01
6.82288826e-01 7.78716654e-02 7.25778282e-01 1.12915027e+00
-2.58293629e-01 5.41258156e-01 8.46963942e-01 -5.45015931e-01
4.91019338e-02 -1.80982072e-02 1.20746955e-01 5.29040217e-01
2.18650520e-01 1.62548870e-01 -6.99831903e-01 -4.93089557e-02
8.62342238e-01 -4.08068269e-01 -1.24991044e-01 -9.30486545e-02
-1.41543686e+00 6.44795120e-01 4.24188346e-01 8.22414994e-01
-9.38182324e-02 5.96151650e-01 8.70555222e-01 2.12182537e-01
3.70034099e-01 1.14276493e+00 -6.47239029e-01 -3.99339885e-01
-1.37788320e+00 6.28413446e-03 5.76045692e-01 4.52145815e-01
3.18333685e-01 4.27632958e-01 -1.41979709e-01 1.17725623e+00
2.55017006e-03 2.80913264e-02 6.19140089e-01 -1.31045353e+00
5.18396385e-02 1.16291016e-01 -2.32562616e-01 -1.07805777e+00
-6.43305898e-01 -1.10898519e+00 -6.42764807e-01 -5.75857423e-02
4.61967945e-01 -1.39064239e-02 -5.10786355e-01 2.05966687e+00
-3.66291255e-01 3.48319560e-01 -2.64277369e-01 9.16334391e-01
8.46156240e-01 -2.60324329e-02 4.27656889e-01 1.87103793e-01
1.23202097e+00 -6.42623365e-01 -5.66271186e-01 -9.22633037e-02
8.96453857e-01 -8.47871065e-01 9.66689825e-01 4.73915309e-01
-8.58740211e-01 -9.25158441e-01 -9.78107691e-01 3.81465815e-02
-2.13814303e-01 5.93185425e-01 7.13206053e-01 6.89595342e-01
-9.73198712e-01 1.34475982e+00 -5.88905036e-01 -4.56951112e-01
4.27188724e-01 6.91296160e-01 -1.30912766e-01 5.62168241e-01
-1.23342180e+00 6.23946905e-01 6.68521464e-01 4.93278615e-02
-7.25276768e-01 -5.38869619e-01 -4.43791121e-01 1.52783453e-01
4.80577499e-02 -3.97666007e-01 1.45035315e+00 -1.17195404e+00
-1.07359064e+00 9.99770403e-01 4.80493039e-01 -4.22414601e-01
-1.44002452e-01 -1.13262832e-02 -6.61330640e-01 -1.59262210e-01
1.44317374e-01 9.30130959e-01 1.32904157e-01 -1.19139600e+00
-5.85599482e-01 -9.28725302e-02 -4.39519174e-02 -4.41175001e-03
-5.39542913e-01 -5.59122162e-03 -4.24266830e-02 -1.08267426e+00
1.58375666e-01 -1.12831521e+00 1.72948554e-01 -6.18713915e-01
-4.26492453e-01 -1.51224583e-01 -1.64652411e-02 -4.82912242e-01
1.21988118e+00 -2.46829343e+00 -3.17915827e-01 2.81881601e-01
4.32952046e-02 1.91465780e-01 -5.34844279e-01 2.44976282e-01
-2.21385494e-01 3.90627712e-01 1.46505326e-01 -2.95188338e-01
2.09119231e-01 1.56213716e-01 -3.54255557e-01 3.13191026e-01
-1.36499688e-01 9.18494046e-01 -7.33627260e-01 -2.44655102e-01
-3.81037965e-02 4.26956922e-01 -4.88255918e-01 -2.50593156e-01
-1.09662697e-01 3.65275353e-01 4.56267446e-02 5.98705947e-01
-4.41218764e-02 -1.04309343e-01 3.52939427e-01 -6.59776032e-01
-1.09176077e-01 8.59104633e-01 -8.69823217e-01 1.89924407e+00
-2.70246565e-01 1.03639162e+00 -2.37604126e-01 -9.24655437e-01
8.79493415e-01 3.99192065e-01 6.50680900e-01 -5.29850364e-01
6.01508141e-01 2.59147555e-01 3.79698753e-01 -1.62433088e-01
7.80691564e-01 -4.24617022e-01 -1.78461745e-01 6.35077059e-01
4.03106779e-01 3.11165243e-01 1.24646731e-01 -2.23346964e-01
1.08510303e+00 -3.48353642e-03 -1.86514810e-01 -2.66840011e-01
-2.72314548e-01 1.78321183e-01 5.91349840e-01 1.04996979e+00
-1.92658529e-01 7.38545358e-01 2.40594462e-01 -3.83138120e-01
-9.94260609e-01 -7.44977534e-01 -2.57164180e-01 1.63001704e+00
-1.56737417e-01 -7.29498923e-01 -5.46284139e-01 -4.86149698e-01
-7.51130879e-02 3.47818434e-01 -6.69923007e-01 -3.76064032e-01
-2.73925692e-01 -7.13064611e-01 1.32107961e+00 7.23415852e-01
3.45242113e-01 -1.26092672e+00 -4.63236213e-01 3.22222978e-01
-3.46572369e-01 -9.92553771e-01 -1.73331812e-01 6.52196527e-01
-9.87832725e-01 -9.48316157e-01 -3.86008382e-01 -9.47605789e-01
5.79887591e-02 8.17715079e-02 1.45000875e+00 1.77222610e-01
8.88957381e-02 3.79565001e-01 -4.37375307e-01 -3.99691522e-01
-5.32439888e-01 6.37553573e-01 2.74140716e-01 -6.00312173e-01
5.62704444e-01 -8.05538654e-01 -3.05975586e-01 4.25132066e-01
-7.07761586e-01 -3.45888704e-01 5.05964756e-01 7.23069608e-01
5.02109170e-01 3.78436357e-01 6.49880469e-01 -7.87648141e-01
9.16879952e-01 -1.60138458e-01 -4.39167172e-02 1.49414495e-01
-6.85270607e-01 -5.53584239e-03 7.48229474e-02 -6.07798815e-01
-3.00005704e-01 -4.01233509e-02 -1.47826090e-01 -3.85904998e-01
-2.03218356e-01 5.03369808e-01 2.88540572e-01 -1.29682515e-02
8.65268588e-01 -3.59266549e-01 -2.16386467e-01 -6.78855002e-01
1.81507051e-01 6.27292693e-01 8.06744695e-01 -6.58329546e-01
4.19275463e-01 2.46396407e-01 -9.70111042e-02 -3.28951031e-01
-1.33077312e+00 -4.85029757e-01 -6.67087793e-01 -3.02324235e-01
7.38791347e-01 -9.86874402e-01 -8.26682627e-01 1.57856926e-01
-6.57347143e-01 -5.74514985e-01 -5.96161306e-01 6.36141539e-01
-5.51399469e-01 -1.10618379e-02 -8.03855598e-01 -6.69676363e-01
-3.15865904e-01 -7.58829832e-01 9.41736579e-01 -1.12400062e-01
-9.77846324e-01 -1.04646599e+00 3.25922102e-01 4.88443673e-01
2.87602812e-01 -8.36646929e-02 8.46156180e-01 -9.43774819e-01
-8.76150653e-02 3.63225825e-02 -1.21098623e-01 4.57835943e-01
-1.96562130e-02 -7.63199925e-02 -1.41368163e+00 -1.59558296e-01
-5.67044497e-01 -5.11716664e-01 9.07343030e-01 5.38846254e-01
1.07728255e+00 3.78968269e-02 -1.48680294e-02 4.89787161e-01
1.08064568e+00 1.28290504e-02 3.19379449e-01 6.58999026e-01
6.98520601e-01 4.05089855e-01 2.85372227e-01 1.87498286e-01
1.89700723e-02 7.57421017e-01 1.53699815e-01 -3.70300487e-02
-3.38772684e-01 -3.78938198e-01 1.51573643e-01 1.07810390e+00
-1.58800840e-01 -1.79494724e-01 -7.77915657e-01 5.55522978e-01
-1.66265619e+00 -8.93903494e-01 -5.24064675e-02 2.15572000e+00
7.29141712e-01 1.77273214e-01 4.41041321e-01 5.99100828e-01
7.02298641e-01 2.74587106e-02 -1.97374150e-01 -2.28690386e-01
-2.77961612e-01 4.31229770e-01 4.64078337e-01 8.90326104e-04
-1.24698544e+00 9.86854672e-01 7.29144955e+00 7.44175434e-01
-1.04118335e+00 1.56534284e-01 2.82816261e-01 -2.49376521e-01
-1.15067184e-01 -1.58013344e-01 -4.30200994e-01 2.19328478e-01
1.02898240e+00 4.58253086e-01 5.58278918e-01 5.60493290e-01
1.15614235e-01 1.30150795e-01 -9.86057162e-01 9.18856978e-01
3.85788158e-02 -1.22777367e+00 -2.48032048e-01 4.68308851e-02
6.28640950e-01 4.06118035e-01 1.51120976e-01 4.09247488e-01
3.86341542e-01 -8.62418771e-01 1.00954676e+00 3.02912652e-01
7.96998501e-01 -7.77089536e-01 1.02925622e+00 -1.17202036e-01
-9.70101237e-01 -8.27082023e-02 -5.04320860e-01 -3.14409584e-01
7.88683072e-03 6.56681657e-01 -7.92537093e-01 1.11813322e-01
8.54983807e-01 8.14696848e-01 -6.73262656e-01 9.97917295e-01
-8.92583057e-02 1.00566554e+00 -1.97426602e-01 1.59244001e-01
1.52365327e-01 2.09375229e-02 2.78446496e-01 1.20498812e+00
5.78268290e-01 -3.97932529e-01 -7.45160729e-02 8.95155966e-01
-2.80215114e-01 6.77222610e-02 -3.40813130e-01 -5.05989850e-01
4.74893421e-01 1.23900318e+00 -1.04634058e+00 -9.52623785e-02
1.02714598e-01 7.79303372e-01 2.61153251e-01 1.24096684e-01
-4.68879104e-01 -1.23657204e-01 6.93118989e-01 3.84648032e-02
1.47989959e-01 -1.59525260e-01 -5.79292476e-01 -7.92252302e-01
-2.91787356e-01 -7.47609079e-01 4.29591745e-01 -1.01729298e+00
-1.29850709e+00 4.81756717e-01 -3.58417094e-01 -1.20563459e+00
-4.80665006e-02 -5.49172997e-01 -3.39903474e-01 3.72951001e-01
-8.53714705e-01 -8.90240192e-01 -1.03123866e-01 1.42478600e-01
-3.27055454e-02 -4.03795809e-01 1.08043063e+00 6.52673066e-01
-4.07001436e-01 7.09760070e-01 2.24800035e-01 5.26054025e-01
9.29391265e-01 -1.08656204e+00 3.29188198e-01 2.50158072e-01
9.79808450e-01 6.31312132e-01 4.52471077e-01 -5.05919933e-01
-4.52642798e-01 -1.04247260e+00 9.92485940e-01 -6.27018571e-01
5.96043587e-01 -2.52563685e-01 -5.98830760e-01 5.37623286e-01
1.48398215e-02 -4.54009145e-01 1.30257142e+00 8.92110467e-01
-4.77157980e-01 -3.09029799e-02 -7.51992285e-01 2.50506610e-01
1.25349116e+00 -7.42921352e-01 -3.98707926e-01 -1.00207999e-02
3.47043991e-01 -9.21858773e-02 -9.96091247e-01 5.67592382e-01
1.03627992e+00 -1.03416502e+00 9.01509225e-01 -6.20641589e-01
3.33611876e-01 -1.57109201e-01 -3.60046148e-01 -1.32668030e+00
-6.97433770e-01 -4.99861762e-02 5.26917517e-01 1.44389522e+00
4.05424505e-01 6.20597638e-02 8.90690565e-01 1.67731926e-01
-2.09825203e-01 -3.37219089e-01 -7.20268905e-01 -9.31751192e-01
2.75478270e-02 -9.13041830e-01 4.01012450e-01 1.49930358e+00
1.46975636e-01 3.54396433e-01 -4.24141139e-01 -1.62985742e-01
2.55911380e-01 -6.56658784e-02 3.66107762e-01 -1.71376240e+00
-3.48441392e-01 -6.01785183e-01 -6.79826200e-01 -5.84182441e-01
4.61254865e-01 -1.24811780e+00 2.60127746e-02 -1.34156644e+00
2.23394349e-01 -7.01271236e-01 -7.86024153e-01 7.99967647e-01
2.79246271e-01 1.03574026e+00 3.27494323e-01 4.71375465e-01
-7.68055677e-01 1.86283384e-02 8.08176637e-01 -6.75357133e-02
-1.60454392e-01 -2.11905129e-02 -8.38502347e-01 6.91937327e-01
1.20279551e+00 -6.57134056e-01 -3.29318106e-01 -5.30246258e-01
6.96850538e-01 -6.26805782e-01 4.48639393e-01 -1.41880918e+00
-5.39069949e-03 3.40140492e-01 3.23988825e-01 -3.54136109e-01
3.27946216e-01 -8.00609589e-01 2.46270284e-01 9.00614187e-02
-8.88066173e-01 2.95585059e-02 3.33334237e-01 2.13679641e-01
-2.28340983e-01 -4.58921105e-01 4.91828829e-01 -8.49744007e-02
-6.02358818e-01 -1.70880839e-01 -4.21222001e-01 1.77331701e-01
6.21321425e-02 -9.85081568e-02 -1.03883512e-01 -4.22316015e-01
-9.58814859e-01 -4.00972098e-01 3.92830640e-01 4.52823341e-01
-1.95688993e-01 -1.71633530e+00 -3.48067373e-01 -8.73162225e-02
1.73516959e-01 -5.33384562e-01 1.16311841e-01 6.96617901e-01
-1.15182698e-01 3.37883353e-01 -2.27078527e-01 -5.09403706e-01
-1.31518126e+00 2.07586557e-01 4.32704896e-01 5.04772067e-02
-2.77706772e-01 1.01500511e+00 -2.80257106e-01 -5.82284629e-01
5.53675294e-01 -3.81702214e-01 -1.77350536e-01 3.62315118e-01
5.46266958e-02 3.30397397e-01 2.48785689e-01 -5.74118435e-01
-1.95606530e-01 6.25519335e-01 3.78438979e-01 -2.51471192e-01
1.27483821e+00 1.12352289e-01 5.00509590e-02 7.89324701e-01
9.08601582e-01 -5.20986393e-02 -7.65201032e-01 -1.69486761e-01
1.73911706e-01 -1.34969642e-02 3.71999294e-01 -1.08313811e+00
-1.24938011e+00 8.31995606e-01 8.20309520e-01 2.71772802e-01
9.55836296e-01 1.45696133e-01 6.87449336e-01 5.42689800e-01
1.07141204e-01 -1.35990322e+00 -1.75676852e-01 5.60222924e-01
3.62375975e-01 -1.02536571e+00 -9.75278914e-02 4.04492319e-02
-4.72857744e-01 9.48564708e-01 2.85631478e-01 2.60846559e-02
4.76640284e-01 2.96427131e-01 4.10497487e-01 -4.11435038e-01
-3.68448704e-01 -4.96165127e-01 4.58746225e-01 4.24857885e-01
9.21813726e-01 2.02230603e-01 -1.85418036e-02 8.74994874e-01
-9.05673862e-01 1.46911621e-01 1.54328927e-01 5.66120923e-01
-3.24425101e-01 -1.21563160e+00 -1.89308405e-01 5.34890950e-01
-7.52708793e-01 -3.75417292e-01 -7.52033353e-01 6.43016160e-01
5.90095580e-01 9.79057431e-01 2.77822137e-01 -1.04837608e+00
2.25961655e-01 4.16934758e-01 5.58610320e-01 -6.33022606e-01
-9.68352616e-01 2.17689276e-01 4.48450208e-01 -1.10840686e-01
-8.39863420e-01 -6.30570114e-01 -9.46485400e-01 -1.77491114e-01
-6.52978837e-01 2.94248641e-01 7.40049660e-01 8.61159563e-01
3.04692626e-01 7.86637962e-01 1.43228963e-01 -6.11285388e-01
-2.29260713e-01 -1.24503160e+00 -8.84357870e-01 5.98426759e-01
-1.13470778e-01 -6.21131659e-01 -4.15401369e-01 9.00166109e-02] | [15.7643461227417, 5.240870475769043] |
95c00b43-46a1-4455-be2e-e2ec217e78e0 | a-confidence-based-partial-label-learning | 2305.12485 | null | https://arxiv.org/abs/2305.12485v1 | https://arxiv.org/pdf/2305.12485v1.pdf | A Confidence-based Partial Label Learning Model for Crowd-Annotated Named Entity Recognition | Existing models for named entity recognition (NER) are mainly based on large-scale labeled datasets, which always obtain using crowdsourcing. However, it is hard to obtain a unified and correct label via majority voting from multiple annotators for NER due to the large labeling space and complexity of this task. To address this problem, we aim to utilize the original multi-annotator labels directly. Particularly, we propose a Confidence-based Partial Label Learning (CPLL) method to integrate the prior confidence (given by annotators) and posterior confidences (learned by models) for crowd-annotated NER. This model learns a token- and content-dependent confidence via an Expectation-Maximization (EM) algorithm by minimizing empirical risk. The true posterior estimator and confidence estimator perform iteratively to update the true posterior and confidence respectively. We conduct extensive experimental results on both real-world and synthetic datasets, which show that our model can improve performance effectively compared with strong baselines. | ['Ying Shan', 'Jin Ma', 'Xuanjing Huang', 'Tao Gui', 'Qi Zhang', 'Yuanbin Wu', 'Xiao Wang', 'Qunxi Zhu', 'Jie zhou', 'Limao Xiong'] | 2023-05-21 | null | null | null | null | ['partial-label-learning', 'named-entity-recognition-ner'] | ['methodology', 'natural-language-processing'] | [-2.96737999e-01 1.59016281e-01 -9.26664397e-02 -6.70538068e-01
-1.47692907e+00 -8.76034617e-01 3.43174011e-01 2.39391088e-01
-9.96723354e-01 1.07447505e+00 1.68148473e-01 8.44872966e-02
5.14653563e-01 -3.94152105e-01 -5.12057662e-01 -5.23250759e-01
4.75753307e-01 6.60849988e-01 4.59971398e-01 4.48033363e-01
-2.37249341e-02 1.68270059e-03 -9.76619184e-01 -1.01203257e-02
1.22429872e+00 8.46298218e-01 1.91338494e-01 5.44329643e-01
-4.18189466e-01 1.01239598e+00 -8.67652178e-01 -8.11486006e-01
-1.32818567e-02 -2.27977946e-01 -9.26843286e-01 -2.47082138e-03
-4.91691474e-03 -1.60183683e-01 2.76478738e-01 1.40380383e+00
4.19285327e-01 2.66356766e-01 7.28462160e-01 -1.12698126e+00
-6.67799473e-01 6.46866024e-01 -5.89858532e-01 -2.90553838e-01
2.84752101e-01 -1.85788065e-01 1.07562304e+00 -1.16723323e+00
3.35669249e-01 1.22348702e+00 8.55215311e-01 6.30937636e-01
-9.57942188e-01 -6.65427864e-01 3.67136061e-01 -1.28932714e-01
-1.66429937e+00 -1.59515649e-01 2.99717844e-01 -2.38798767e-01
3.75807494e-01 -3.07544563e-02 1.00852931e-02 9.12099779e-01
-3.81258070e-01 9.74944413e-01 1.32972372e+00 -5.13968766e-01
5.63325465e-01 4.48842347e-01 -4.35381606e-02 6.73816323e-01
2.04777241e-01 -3.99277508e-01 -3.69975507e-01 -5.46474099e-01
5.03519356e-01 -1.75061151e-01 -2.23473266e-01 2.79798601e-02
-1.21979976e+00 6.93073392e-01 1.39522195e-01 1.66554525e-01
-3.23722392e-01 1.59755260e-01 2.28416055e-01 -2.39534870e-01
7.17409730e-01 3.43830995e-02 -8.30240488e-01 8.04365147e-03
-1.02020693e+00 -1.04033068e-01 1.17726088e+00 1.06024897e+00
8.10730278e-01 -3.64666998e-01 -5.29001296e-01 1.00661755e+00
8.05067897e-01 7.17332900e-01 2.59712070e-01 -1.02755475e+00
4.77783740e-01 4.47972804e-01 8.84503305e-01 -7.44937241e-01
-1.24544792e-01 -1.09919142e-02 -6.72417760e-01 -1.31153435e-01
6.41495645e-01 -7.01007724e-01 -7.34470248e-01 1.72568405e+00
7.86417961e-01 4.38092470e-01 -9.06475782e-02 8.90860975e-01
4.97406155e-01 5.49956858e-01 7.15726554e-01 -3.34669918e-01
1.47909331e+00 -1.15247512e+00 -9.36752141e-01 -2.62476385e-01
8.06744576e-01 -5.92885017e-01 8.27011645e-01 1.34630278e-01
-6.27353728e-01 -3.15803200e-01 -6.95825458e-01 2.53584087e-02
-9.42641422e-02 6.24887586e-01 1.99939370e-01 8.67598414e-01
-4.98610109e-01 3.89633775e-01 -8.90388250e-01 -7.40246698e-02
3.17100048e-01 1.59446552e-01 -4.26880628e-01 -1.38251454e-01
-1.24961197e+00 7.52824903e-01 4.64911878e-01 3.91771793e-01
-8.10739517e-01 -1.84210569e-01 -9.02750075e-01 -8.77087489e-02
6.35769784e-01 -2.69814879e-01 1.67356730e+00 -6.34848356e-01
-1.47179008e+00 7.66506851e-01 -4.82778132e-01 -1.18678436e-01
6.77868783e-01 -4.04576927e-01 -3.39585006e-01 -1.09167658e-01
5.48667669e-01 5.22267461e-01 3.38032037e-01 -1.54502130e+00
-8.83728445e-01 8.48596543e-02 -4.48393710e-02 9.54817832e-02
-2.82128632e-01 4.48238134e-01 -5.94896078e-01 -4.86828983e-01
3.50484587e-02 -7.67266572e-01 -5.13520956e-01 -5.29857613e-02
-5.10450840e-01 -5.91611862e-01 1.91447422e-01 -7.13874578e-01
1.12867808e+00 -1.95710611e+00 -4.25466835e-01 -3.02477162e-02
1.24657042e-01 3.59570533e-01 5.33362851e-02 9.73014459e-02
5.04473090e-01 4.60522830e-01 -3.87661815e-01 -7.29290903e-01
2.33393639e-01 2.93982714e-01 -1.73018619e-01 4.45533127e-01
3.21482271e-01 6.62291050e-01 -1.34221637e+00 -1.12520099e+00
-1.56309143e-01 3.22996736e-01 -7.77816176e-02 4.04635936e-01
-3.14872146e-01 5.57496130e-01 -6.58067703e-01 5.84256828e-01
8.56538296e-01 -5.33626676e-01 5.26065528e-01 -9.79038104e-02
-1.00477338e-01 4.93651330e-02 -1.68695724e+00 1.50576687e+00
-4.58517969e-01 3.42721380e-02 1.55351192e-01 -4.92329717e-01
9.67756450e-01 6.28983796e-01 2.44210996e-02 -4.67082448e-02
1.85561761e-01 4.44658965e-01 -7.13694930e-01 -3.39760303e-01
5.33730686e-01 -3.10923338e-01 -3.61933291e-01 6.33407950e-01
1.18562505e-01 2.19938457e-01 1.28076673e-01 1.52293295e-01
8.98636043e-01 2.15569094e-01 3.71609807e-01 7.04582827e-03
5.49290240e-01 -8.67376849e-02 1.29261518e+00 9.41126883e-01
-6.11514449e-01 5.73971748e-01 4.57159638e-01 -1.56311542e-01
-7.67837107e-01 -7.96895266e-01 -2.31279712e-02 1.19057691e+00
2.45240718e-01 -1.81752071e-01 -9.57774818e-01 -1.37772381e+00
-4.20566201e-01 6.36400044e-01 -4.93752152e-01 4.85225052e-01
-3.74991238e-01 -7.30713964e-01 7.70332754e-01 7.58267403e-01
6.90482855e-01 -8.73788416e-01 -2.89110676e-03 4.17458326e-01
-6.33126795e-01 -1.38642347e+00 -7.70376444e-01 1.24711439e-01
-3.77473027e-01 -9.97168064e-01 -8.15727651e-01 -7.11128712e-01
7.71236300e-01 -5.24951406e-02 9.95931745e-01 4.28738780e-02
3.62251431e-01 2.10270718e-01 -4.42632228e-01 -4.47770029e-01
-2.06146121e-01 8.52528140e-02 1.70400918e-01 2.58334912e-02
6.23448014e-01 -3.24198715e-02 -3.18497479e-01 7.65047371e-01
-7.92765975e-01 -3.91939610e-01 4.77818996e-01 8.47935021e-01
7.41476655e-01 -1.02106489e-01 7.95164943e-01 -1.11871469e+00
5.67820191e-01 -5.39276183e-01 -7.05779076e-01 8.55907738e-01
-4.72663164e-01 1.31245881e-01 3.83932143e-01 -5.35263956e-01
-1.39155662e+00 2.70023823e-01 -9.26403478e-02 -1.25809118e-01
-2.68965572e-01 3.71155560e-01 -3.60996842e-01 2.60161102e-01
4.57200676e-01 -1.14142567e-01 -7.78212786e-01 -6.20665193e-01
3.87353688e-01 1.04713404e+00 6.24395192e-01 -8.86647940e-01
6.58278525e-01 4.46203083e-01 -4.75736886e-01 -1.22115955e-01
-1.71741474e+00 -5.97456217e-01 -6.40948713e-01 -6.89682066e-02
1.06396079e+00 -1.28248048e+00 -6.36528969e-01 3.92224878e-01
-1.59810746e+00 -2.85735186e-02 -1.54834196e-01 6.77739143e-01
4.15477864e-02 7.32999802e-01 -7.01487184e-01 -1.34415865e+00
-3.12201679e-01 -8.15155983e-01 1.10359335e+00 7.27247596e-01
3.39914002e-02 -1.10797644e+00 2.94701576e-01 3.04561406e-01
8.64225402e-02 2.12170631e-02 1.78255245e-01 -1.05619168e+00
-2.50463694e-01 -5.92696965e-01 -5.07502735e-01 4.70247597e-01
-1.06099710e-01 -1.25925690e-01 -1.12808490e+00 2.01178063e-02
-1.19583093e-01 -7.61204958e-01 5.99353731e-01 -3.70195433e-02
8.65781546e-01 -1.36684433e-01 -3.42259586e-01 -1.56209752e-01
1.28511536e+00 -3.44127327e-01 2.81303883e-01 1.11915931e-01
6.27502918e-01 5.14300764e-01 9.64433670e-01 6.36255622e-01
7.63068497e-01 3.53796422e-01 5.49862720e-02 2.42984056e-01
3.91546845e-01 -6.26403093e-01 2.03816503e-01 1.00835598e+00
2.84726899e-02 -5.10451555e-01 -9.13060009e-01 5.61736226e-01
-2.20634460e+00 -6.62373662e-01 -1.71262965e-01 2.18825483e+00
1.24121797e+00 6.34122081e-03 -8.17202404e-02 -2.70668983e-01
1.44451475e+00 -1.37045220e-01 -4.41029698e-01 2.82248706e-01
-9.79554877e-02 -2.47257188e-01 4.42486197e-01 4.76096392e-01
-1.22539461e+00 1.03127432e+00 6.53421974e+00 1.06708741e+00
-2.97430754e-01 6.32624745e-01 6.35430396e-01 3.48616928e-01
-2.85667300e-01 2.27396309e-01 -1.21705496e+00 5.46912014e-01
9.25509095e-01 5.78099824e-02 3.44449878e-02 1.03767943e+00
-7.38565028e-02 -2.14334294e-01 -7.39412785e-01 8.17825556e-01
-1.13253124e-01 -8.60523522e-01 -5.58799684e-01 -1.14498846e-01
1.14268553e+00 1.53883919e-01 -7.65988231e-01 5.47399998e-01
1.10706460e+00 -6.18859172e-01 8.75283539e-01 6.64248109e-01
7.69993007e-01 -6.17026806e-01 1.16528082e+00 7.84578860e-01
-1.37224615e+00 1.04742214e-01 -6.09413624e-01 2.09793404e-01
5.91877043e-01 1.03218687e+00 -6.41696215e-01 3.79285216e-01
4.62935686e-01 1.78373009e-01 -1.49907500e-01 1.04535294e+00
-9.64688540e-01 7.34973252e-01 -4.32647705e-01 -3.88320386e-01
1.46201225e-02 1.23705521e-01 9.88117084e-02 1.47688937e+00
4.36576381e-02 1.50813445e-01 5.55397749e-01 8.69223595e-01
-6.84990585e-01 1.19068049e-01 1.32256910e-01 1.00285271e-02
1.09910595e+00 1.72885227e+00 -9.68101442e-01 -3.64008158e-01
-4.58922327e-01 1.06974840e+00 8.51947963e-01 3.52107406e-01
-9.13671196e-01 -4.56030369e-01 7.88017511e-02 -5.16199946e-01
3.76534969e-01 -2.74356574e-01 -1.68852478e-01 -1.35456932e+00
1.35804415e-01 -4.52912956e-01 4.12287176e-01 -5.56210935e-01
-1.75202513e+00 6.23434663e-01 -1.84214935e-01 -1.20767641e+00
2.07597036e-02 -5.28054535e-01 -6.48766458e-01 9.06662405e-01
-1.62516999e+00 -9.36807990e-01 -1.73262641e-01 1.76775396e-01
2.30996296e-01 2.72263408e-01 9.24127936e-01 3.23958457e-01
-6.67372942e-01 7.46750832e-01 5.27335182e-02 6.40956104e-01
1.03032565e+00 -1.41960251e+00 2.43217334e-01 7.47830272e-01
1.08048938e-01 4.89432901e-01 4.59288120e-01 -8.13667297e-01
-5.88070869e-01 -1.18184543e+00 1.33652878e+00 -8.87412786e-01
4.77319390e-01 -2.59123981e-01 -7.80794263e-01 5.98393023e-01
-6.38718531e-02 5.60594320e-01 9.45133746e-01 1.30104259e-01
-5.87018251e-01 3.09939861e-01 -1.36546063e+00 2.33053625e-01
7.76415408e-01 -6.18622601e-01 -5.10674715e-01 4.55518067e-01
8.61951530e-01 -3.97500068e-01 -9.98143137e-01 2.59394079e-01
3.55522692e-01 -4.96505588e-01 3.25631768e-01 -4.70097244e-01
4.53984551e-02 -8.84090483e-01 -1.79313615e-01 -1.20476961e+00
8.83143302e-03 -2.95616001e-01 -1.69431623e-02 1.91288209e+00
7.65547514e-01 -3.45332891e-01 4.95416909e-01 1.33788717e+00
2.07834259e-01 -3.02004874e-01 -9.13629651e-01 -7.84447908e-01
-1.62293345e-01 -5.36074519e-01 6.24318600e-01 1.06088698e+00
-5.28396219e-02 3.62818152e-01 -6.21383131e-01 6.21178210e-01
6.57149613e-01 -3.30159456e-01 5.15525222e-01 -1.27980578e+00
-2.70584702e-01 3.69768411e-01 -1.02872746e-02 -1.15904427e+00
3.89222294e-01 -3.79285514e-01 8.96737456e-01 -1.64878929e+00
5.49767196e-01 -5.99613070e-01 -3.30657959e-01 6.75552547e-01
-7.54329741e-01 8.10209513e-02 2.88396217e-02 4.19229835e-01
-1.60253334e+00 6.43576205e-01 8.92926931e-01 1.59125343e-01
1.72125086e-01 -7.91058466e-02 -6.52777970e-01 1.07748759e+00
5.36874533e-01 -1.30579972e+00 9.48197246e-02 -7.25730956e-01
5.85378647e-01 3.02951429e-02 3.66531014e-02 -8.13993931e-01
7.47501194e-01 -2.68255562e-01 2.88619965e-01 -4.24420834e-01
-1.80198714e-01 -6.68228447e-01 -1.83212504e-01 -1.79830007e-02
-4.68179613e-01 -2.05072209e-01 -2.88637459e-01 1.22006035e+00
-2.49279335e-01 -7.45364785e-01 6.15805030e-01 -2.49351084e-01
-3.59764963e-01 2.84151703e-01 -1.57228366e-01 5.01398027e-01
7.76787817e-01 3.46767873e-01 -2.18938842e-01 -2.75479078e-01
-6.67677581e-01 6.24681771e-01 2.82045454e-01 -2.58105695e-02
7.97480717e-02 -1.50536263e+00 -9.52141881e-01 -5.28832376e-01
2.72877812e-01 3.31477612e-01 1.31675243e-01 4.90756154e-01
-1.01413965e-01 -1.07962795e-01 5.77305079e-01 -3.18307281e-01
-7.82253623e-01 2.74473518e-01 3.01247865e-01 -6.13299370e-01
1.07191488e-01 1.04151535e+00 -2.36700669e-01 -1.06426263e+00
3.11384648e-01 9.36429426e-02 -2.76659399e-01 -1.72199249e-01
8.43147516e-01 4.55742598e-01 -2.53935874e-01 -5.44900954e-01
-4.04132962e-01 3.03427607e-01 2.98245903e-03 -4.74344224e-01
8.76818001e-01 -3.95369291e-01 -9.11629200e-02 4.85880822e-01
7.15051949e-01 3.71556312e-01 -1.40884709e+00 -7.20522702e-01
4.71601903e-01 -4.04225856e-01 -2.03549102e-01 -8.14458311e-01
-6.54572845e-01 6.07175112e-01 2.87141323e-01 8.33808482e-02
5.75413764e-01 3.08600534e-02 7.65219867e-01 7.51139522e-01
5.59405982e-01 -1.54000723e+00 -5.40115908e-02 5.53396225e-01
1.92852870e-01 -1.57867050e+00 -2.24140540e-01 -4.57749367e-01
-1.02813423e+00 7.37974048e-01 7.04597414e-01 1.41251475e-01
6.28952086e-01 1.81849331e-01 3.20401341e-01 2.45358825e-01
-6.28292739e-01 -2.32872427e-01 1.12826161e-01 5.29465735e-01
5.73189795e-01 1.91012740e-01 -4.85780567e-01 1.29238963e+00
5.36154866e-01 1.19417213e-01 2.13270873e-01 9.47582126e-01
-6.79799497e-01 -1.33668876e+00 -4.55840528e-01 1.10862695e-01
-7.38128185e-01 8.70106742e-03 -2.73507446e-01 1.44304603e-01
2.24229127e-01 1.56147122e+00 -3.87103498e-01 -2.47294515e-01
1.66094825e-01 2.37423196e-01 -2.82936022e-02 -8.23672175e-01
-4.47479248e-01 -6.53041601e-02 3.06940645e-01 -1.30819738e-01
-7.41970360e-01 -5.09055257e-01 -1.49005699e+00 -2.21958593e-03
-1.04619336e+00 7.20763385e-01 7.84062505e-01 1.38880026e+00
1.47983417e-01 8.07439536e-03 7.31508136e-01 -5.31529784e-01
-9.74166691e-01 -1.29491735e+00 -7.63880312e-01 3.02551180e-01
-3.03630948e-01 -4.86615717e-01 -5.06448507e-01 3.98615152e-02] | [9.6200532913208, 4.713494777679443] |
f3a475b2-dd59-424f-8b6f-c544e14fd13d | auxiliary-multimodal-lstm-for-audio-visual | 1701.04224 | null | http://arxiv.org/abs/1701.04224v2 | http://arxiv.org/pdf/1701.04224v2.pdf | Auxiliary Multimodal LSTM for Audio-visual Speech Recognition and Lipreading | The Aduio-visual Speech Recognition (AVSR) which employs both the video and
audio information to do Automatic Speech Recognition (ASR) is one of the
application of multimodal leaning making ASR system more robust and accuracy.
The traditional models usually treated AVSR as inference or projection but
strict prior limits its ability. As the revival of deep learning, Deep Neural
Networks (DNN) becomes an important toolkit in many traditional classification
tasks including ASR, image classification, natural language processing. Some
DNN models were used in AVSR like Multimodal Deep Autoencoders (MDAEs),
Multimodal Deep Belief Network (MDBN) and Multimodal Deep Boltzmann Machine
(MDBM) that actually work better than traditional methods. However, such DNN
models have several shortcomings: (1) They don't balance the modal fusion and
temporal fusion, or even haven't temporal fusion; (2)The architecture of these
models isn't end-to-end, the training and testing getting cumbersome. We
propose a DNN model, Auxiliary Multimodal LSTM (am-LSTM), to overcome such
weakness. The am-LSTM could be trained and tested once, moreover easy to train
and preventing overfitting automatically. The extensibility and flexibility are
also take into consideration. The experiments show that am-LSTM is much better
than traditional methods and other DNN models in three datasets. | ['Weijun Ji', 'Chunlin Tian'] | 2017-01-16 | null | null | null | null | ['lipreading', 'audio-visual-speech-recognition'] | ['computer-vision', 'speech'] | [-3.65068138e-01 -2.76924640e-01 -1.72513068e-01 -1.62565365e-01
-4.42758977e-01 -8.11369121e-02 8.09617400e-01 -4.09388125e-01
-5.22694051e-01 6.62202001e-01 4.58740711e-01 -5.39861977e-01
1.46520242e-01 -4.53925848e-01 -4.23182696e-01 -8.26821744e-01
2.51145095e-01 3.79967630e-01 1.55120715e-01 -4.52771544e-01
2.91033797e-02 6.13154233e-01 -1.37342846e+00 5.13350606e-01
5.14440358e-01 1.21662021e+00 4.36727196e-01 7.49973953e-01
-6.48071647e-01 1.14121997e+00 -6.21769607e-01 -3.48639667e-01
-2.83596426e-01 5.39797060e-02 -8.37797344e-01 -9.31928754e-02
-2.68993825e-01 -4.97098058e-01 -9.71790493e-01 9.15995300e-01
8.07563782e-01 4.81615037e-01 6.93917036e-01 -1.23831761e+00
-1.10391510e+00 4.83403891e-01 -5.91477811e-01 1.95172027e-01
3.37687135e-01 4.19308506e-02 2.91094244e-01 -1.11652780e+00
-4.16627415e-02 1.68825328e+00 5.36777794e-01 1.03591347e+00
-6.30500853e-01 -5.31117857e-01 4.54608463e-02 5.46908915e-01
-1.36575043e+00 -6.66081429e-01 6.09645784e-01 -1.71936154e-01
1.22961366e+00 4.28543746e-01 5.75849771e-01 1.62488866e+00
1.24737106e-01 1.14914525e+00 1.02423823e+00 -2.50838041e-01
9.06778872e-02 3.28294456e-01 3.42667066e-02 4.28844243e-01
-5.09325743e-01 1.11686625e-01 -5.33155978e-01 -4.97139394e-02
7.45660007e-01 4.93597627e-01 -1.85064912e-01 3.30594152e-01
-1.25342071e+00 6.77456498e-01 2.25422621e-01 5.56069613e-01
-2.59754658e-01 -3.42389494e-02 7.42043257e-01 4.21710908e-01
2.40232095e-01 -5.61303735e-01 -4.07645822e-01 -3.83598834e-01
-7.12573349e-01 -3.18720490e-01 3.58169317e-01 5.20805359e-01
5.39612114e-01 7.26167798e-01 2.12405413e-01 1.45303524e+00
6.98576570e-01 5.79263091e-01 1.31710255e+00 -5.00496745e-01
3.93026561e-01 5.13761044e-01 -4.38415468e-01 -7.62491226e-01
-2.33421937e-01 1.39089003e-01 -1.29295659e+00 1.93967581e-01
-2.89294273e-01 -5.19381687e-02 -1.53759217e+00 1.42989457e+00
-6.78358003e-02 -8.94991588e-03 6.03112578e-01 8.89680088e-01
1.64712942e+00 1.42157567e+00 2.55275905e-01 -1.75485477e-01
1.19241095e+00 -9.51546848e-01 -1.10086012e+00 -2.30405390e-01
4.22459364e-01 -1.01308250e+00 9.95416582e-01 6.48285329e-01
-9.37004864e-01 -7.60490417e-01 -9.18067694e-01 -2.07380071e-01
-6.81466401e-01 2.95629025e-01 5.59857607e-01 6.43989205e-01
-1.19347239e+00 8.51224810e-02 -7.58852959e-01 -4.25645053e-01
1.99183017e-01 5.96787930e-01 -8.18118155e-01 2.09427610e-01
-1.38300908e+00 1.18011343e+00 8.67617667e-01 4.70202148e-01
-1.18789899e+00 9.89914313e-02 -9.02094066e-01 -5.63304573e-02
9.89067405e-02 -3.82100165e-01 1.27012587e+00 -1.43757057e+00
-2.08785892e+00 4.91141766e-01 -2.01122314e-01 -2.32478499e-01
1.69671960e-02 -7.47822672e-02 -8.92422438e-01 -1.34349605e-02
-7.43150771e-01 9.37853813e-01 7.14588821e-01 -1.05261528e+00
-3.99565428e-01 -3.14799249e-01 -1.62019491e-01 2.85889179e-01
-5.33778906e-01 1.30905181e-01 -8.45956951e-02 -6.53267622e-01
1.84416965e-01 -5.68174362e-01 1.33580238e-01 -4.82026845e-01
-2.19590470e-01 -6.51379526e-01 1.50017631e+00 -1.02676177e+00
1.30510020e+00 -2.22771192e+00 2.45476231e-01 -1.10135540e-01
-4.36654128e-02 9.04239595e-01 -9.72806811e-02 6.14019513e-01
-2.88958639e-01 6.33731782e-02 -2.25964338e-02 -4.04163241e-01
-7.73668960e-02 6.86798275e-01 -3.81260663e-01 2.40271106e-01
-1.33467063e-01 8.70002210e-01 -3.54642779e-01 -5.91066360e-01
6.81450427e-01 7.94322014e-01 3.03939953e-02 2.76551455e-01
-6.46911189e-02 1.84015647e-01 -1.66049168e-01 8.80337059e-01
6.27208054e-01 2.52180129e-01 -3.45792890e-01 -3.66213143e-01
2.95043886e-02 -1.06022852e-02 -1.06556034e+00 1.66135967e+00
-3.39297116e-01 8.17083716e-01 1.49411522e-02 -1.35850608e+00
1.17249012e+00 9.17345524e-01 -6.76789656e-02 -7.42356241e-01
2.63501048e-01 2.73905516e-01 -1.84088815e-02 -1.10776258e+00
3.22595030e-01 -4.82593417e-01 3.49996388e-01 -4.88040000e-02
4.75424349e-01 3.60082746e-01 -4.72580969e-01 4.98067737e-02
5.79424322e-01 -1.07452057e-01 1.60332218e-01 2.45440587e-01
7.97818601e-01 -3.62567127e-01 5.56385279e-01 9.03927758e-02
-1.03299819e-01 5.04247367e-01 1.60486698e-02 -4.87145364e-01
-1.10239804e+00 -1.04481041e+00 1.00964688e-01 9.37927008e-01
-1.40378818e-01 2.86569428e-02 -3.52530420e-01 -5.00396490e-01
-3.50117952e-01 7.33221352e-01 -1.97540343e-01 -2.88411021e-01
-3.01736981e-01 -7.10548997e-01 9.48460758e-01 6.44629121e-01
8.65004122e-01 -1.14880371e+00 1.80297554e-01 -7.87623897e-02
-2.60979742e-01 -1.02574897e+00 -8.11619759e-02 -7.16849491e-02
-9.69496787e-01 -2.68554091e-01 -1.08245063e+00 -9.33129251e-01
3.01784277e-01 1.13837369e-01 5.14137268e-01 -2.19165057e-01
1.38323605e-01 3.83965462e-01 -4.98024821e-01 -2.43888617e-01
-4.71055776e-01 -2.14045882e-01 3.86363149e-01 1.46473750e-01
6.68980300e-01 -7.30715990e-01 -3.72201979e-01 1.93695962e-01
-1.01122391e+00 1.06565990e-01 9.48150337e-01 9.51026917e-01
2.61202246e-01 -9.51122716e-02 7.49236166e-01 -1.21975772e-01
7.35262990e-01 -5.25260448e-01 -1.89754456e-01 3.72805059e-01
-2.10833475e-01 -1.94964170e-01 5.49556673e-01 -7.57222831e-01
-1.33697927e+00 -3.29793870e-01 -7.39650488e-01 -9.45871532e-01
-3.99149239e-01 7.65168250e-01 -3.47456008e-01 5.28203137e-02
2.22142547e-01 8.01295459e-01 2.39805043e-01 -7.08247900e-01
1.62005380e-01 1.47267306e+00 3.88317227e-01 -2.96144396e-01
1.40944019e-01 1.08278178e-01 -3.61267895e-01 -1.18807220e+00
-5.17580621e-02 -1.48271129e-01 -1.85962036e-01 -3.68033081e-01
1.18298233e+00 -9.00968432e-01 -1.13080645e+00 5.93690038e-01
-1.38701642e+00 -1.29437447e-01 3.34877878e-01 1.07341099e+00
-1.94431841e-01 6.41726315e-01 -9.28996205e-01 -1.36301064e+00
-4.51953977e-01 -1.30811954e+00 6.45569086e-01 4.56470490e-01
2.30300799e-01 -1.06389093e+00 -1.57697834e-02 6.56773210e-01
4.94002372e-01 -2.88305134e-01 8.93098354e-01 -9.05226111e-01
-2.30679274e-01 -1.96677774e-01 -1.54719442e-01 9.81480241e-01
-5.30403852e-02 3.28566819e-01 -1.23185694e+00 -7.80753046e-02
8.00600722e-02 -5.23815095e-01 9.19774532e-01 2.50363141e-01
1.19531524e+00 -3.88865113e-01 -1.03627793e-01 3.01284790e-01
1.11335742e+00 7.86321998e-01 1.03990304e+00 2.46827185e-01
7.87596464e-01 4.18980509e-01 1.75855637e-01 -4.41990755e-02
3.93987566e-01 4.02424484e-01 4.65251297e-01 -1.63809150e-01
-6.81408718e-02 -1.79881632e-01 9.66426373e-01 1.70881665e+00
-3.53614628e-01 -4.10599977e-01 -8.83904457e-01 4.22892570e-01
-1.97803032e+00 -1.12326753e+00 1.21417262e-01 2.01440239e+00
5.03022313e-01 -1.19463257e-01 1.33484662e-01 2.40979359e-01
7.26672351e-01 1.53336421e-01 -1.81386679e-01 -8.50917161e-01
-3.77323121e-01 -5.33807911e-02 1.69340402e-01 4.29102331e-01
-7.89468050e-01 8.57579470e-01 5.73800707e+00 1.44344997e+00
-1.45874465e+00 5.16706288e-01 6.43679559e-01 2.98579223e-02
-2.76893049e-01 -2.81122535e-01 -6.42210186e-01 5.88882208e-01
1.11136866e+00 4.65991199e-01 5.28459907e-01 8.72887135e-01
1.27614021e-01 -2.99772201e-03 -8.44991148e-01 1.51906645e+00
1.77268431e-01 -1.00937629e+00 4.74531323e-01 9.95005015e-03
1.64342314e-01 6.90662712e-02 2.54641354e-01 7.02740252e-01
3.94801944e-02 -1.42399037e+00 4.52200681e-01 6.67821407e-01
5.60603499e-01 -1.13403535e+00 1.05986559e+00 4.64479297e-01
-9.18259561e-01 -2.19850242e-01 -6.50062859e-01 6.58210441e-02
2.46295378e-01 2.54927188e-01 -5.46503007e-01 6.32144213e-01
7.93309987e-01 3.41867387e-01 -1.58295482e-01 6.53059781e-01
1.07534692e-01 6.46661460e-01 -2.81775028e-01 -3.43238950e-01
5.83151996e-01 -1.89526528e-01 5.70243120e-01 1.25607157e+00
5.12225866e-01 1.15088075e-01 -1.39008621e-02 4.01258111e-01
1.09306879e-01 1.27205163e-01 -7.62706578e-01 -4.28862542e-01
3.65633637e-01 1.06091893e+00 -2.05471590e-01 -3.97758543e-01
-5.94975829e-01 8.86617184e-01 2.09649913e-02 6.09877944e-01
-8.11133027e-01 -2.74026752e-01 3.43871653e-01 -3.23560119e-01
1.22023430e-02 -3.32434416e-01 1.20020479e-01 -1.12006700e+00
-1.62355959e-01 -9.97872889e-01 2.90399551e-01 -1.20093608e+00
-1.35078740e+00 8.63862038e-01 1.28124896e-02 -1.09448302e+00
-4.84438613e-02 -9.29828942e-01 -5.39709985e-01 8.21824491e-01
-1.19279659e+00 -1.43878508e+00 -9.72814336e-02 8.32487583e-01
8.29600155e-01 -8.71823430e-01 8.66080821e-01 8.77353370e-01
-8.17165315e-01 5.43385267e-01 4.18985412e-02 3.21187884e-01
4.07900959e-01 -8.33828390e-01 -3.55763108e-01 6.70365155e-01
-1.71613265e-02 6.22696161e-01 3.61028492e-01 -3.91118407e-01
-1.49470150e+00 -5.87630510e-01 5.18048406e-01 -1.86548231e-03
4.79357421e-01 -1.23823896e-01 -9.44298148e-01 9.02505755e-01
6.92064583e-01 -2.43624821e-01 8.12663734e-01 -1.73744142e-01
-7.01873004e-02 -4.41268206e-01 -8.62471104e-01 5.76070368e-01
4.09093767e-01 -7.34634936e-01 -8.85248244e-01 1.28820792e-01
8.28805149e-01 -2.04717249e-01 -8.39730561e-01 5.84135890e-01
4.96849716e-01 -8.96716535e-01 1.01420236e+00 -5.80891132e-01
3.44724476e-01 -3.94055456e-01 -6.56493843e-01 -9.96823788e-01
1.39538292e-02 -3.26419562e-01 -2.83620447e-01 1.49860835e+00
2.93768376e-01 -7.80559003e-01 3.99732262e-01 3.26658815e-01
-3.95888269e-01 -7.93829262e-01 -1.40216756e+00 -6.26819670e-01
-2.00733840e-01 -4.87203866e-01 4.75988030e-01 9.69447315e-01
-1.04083009e-02 5.81446052e-01 -7.88498700e-01 2.45130546e-02
1.06967002e-01 -6.84423923e-01 4.75600809e-01 -9.12827313e-01
-9.83337536e-02 -5.35858095e-01 -6.46991670e-01 -1.24031866e+00
1.92996189e-01 -6.41054630e-01 -3.30705911e-01 -1.72384465e+00
1.58829272e-01 -1.89472865e-02 -4.47298765e-01 6.06087446e-01
2.46192798e-01 -1.21266522e-01 -1.23581104e-01 1.49246946e-01
-3.53989601e-01 9.99031365e-01 1.10292244e+00 -3.70122015e-01
-1.04487702e-01 -1.82041451e-02 -5.85917272e-02 7.93049097e-01
7.16122746e-01 -3.20173018e-02 -5.42467594e-01 -5.04212379e-01
5.26081771e-02 2.83153296e-01 3.56739074e-01 -7.51733959e-01
4.98436004e-01 -1.29734516e-01 6.47195816e-01 -1.11606038e+00
9.35772836e-01 -8.99386346e-01 2.86190003e-01 2.72966534e-01
5.63019188e-04 3.75718862e-01 2.91027457e-01 2.17983842e-01
-6.75991893e-01 -3.45329672e-01 7.07642376e-01 -3.08859825e-01
-1.04478121e+00 2.57053971e-01 -7.98416913e-01 -6.11378074e-01
9.30773735e-01 -4.23797190e-01 -2.86114693e-01 -5.81145704e-01
-8.52567852e-01 1.76026002e-02 -6.21340573e-02 4.56917465e-01
1.19242227e+00 -1.66945326e+00 -3.72484535e-01 2.66074501e-02
-2.47943088e-01 -2.55444311e-02 7.65928686e-01 9.99628007e-01
-6.63163543e-01 4.62349981e-01 -2.08860919e-01 -5.99714041e-01
-1.30882871e+00 7.21242487e-01 4.42763835e-01 1.81240365e-01
-3.85283142e-01 8.85579526e-01 1.21792644e-01 -7.31000245e-01
6.65091634e-01 1.52743990e-02 -5.92153013e-01 -2.34883912e-02
5.16183078e-01 4.86328989e-01 -6.87889904e-02 -8.29902411e-01
-3.78350079e-01 2.72197157e-01 -1.54991746e-01 -4.91829097e-01
1.22320902e+00 -1.79962099e-01 -4.33590919e-01 8.60277653e-01
1.28829026e+00 -4.66960907e-01 -5.17901123e-01 -1.08836576e-01
-4.37142193e-01 -1.73175320e-01 4.08465207e-01 -7.82266915e-01
-1.09350109e+00 1.52211547e+00 8.92592311e-01 2.72400737e-01
1.06135571e+00 -2.71916449e-01 1.21540189e+00 5.12962401e-01
-1.63311988e-01 -1.29488301e+00 1.73600152e-01 4.61070061e-01
9.01280463e-01 -1.23922825e+00 -3.83134305e-01 3.43813181e-01
-8.74595463e-01 1.53406930e+00 6.95805252e-01 2.98174620e-01
7.12030232e-01 1.50804490e-01 1.88907176e-01 1.30197391e-01
-8.85407031e-01 -4.76427637e-02 2.50804275e-01 5.95185637e-01
3.31319869e-01 -9.16619077e-02 -6.63415268e-02 7.55348325e-01
5.02010994e-03 1.35582611e-01 3.88824716e-02 7.77587712e-01
-6.36147499e-01 -9.10815239e-01 -5.79188645e-01 2.16901034e-01
-5.31874299e-01 -7.33977631e-02 -1.91556439e-01 5.42863786e-01
1.69300392e-01 1.01887202e+00 -3.27135846e-02 -9.53833103e-01
1.55458152e-02 1.62824482e-01 3.31678271e-01 -1.83433101e-01
-1.72590718e-01 1.45469666e-01 1.52318357e-02 -1.09742217e-01
-5.38078249e-01 -3.46087217e-02 -1.15328932e+00 -7.34844446e-01
-4.00509179e-01 1.37686610e-01 8.45747054e-01 1.27326250e+00
2.11409554e-01 4.61653858e-01 4.83277231e-01 -8.32957387e-01
-3.48241478e-01 -1.34144163e+00 -4.07601595e-01 -5.18115312e-02
4.10487443e-01 -6.24448121e-01 -2.99704909e-01 -7.79156461e-02] | [13.923480987548828, 5.414465427398682] |
0be6a72a-950c-448c-8d34-e629fd1e21cc | three-things-everyone-should-know-to-improve | null | null | https://ieeexplore.ieee.org/document/6248018 | https://ieeexplore.ieee.org/document/6248018 | Three things everyone should know to improve object retrieval | The objective of this work is object retrieval in large scale image datasets, where the object is specified by an image query and retrieval should be immediate at run time in the manner of Video Google [28]. We make the following three contributions: (i) a new method to compare SIFT descriptors (RootSIFT) which yields superior performance without increasing processing or storage requirements; (ii) a novel method for query expansion where a richer model for the query is learnt discriminatively in a form suited to immediate retrieval through efficient use of the inverted index; (iii) an improvement of the image augmentation method proposed by Turcot and Lowe [29], where only the augmenting features which are spatially consistent with the augmented image are kept. We evaluate these three methods over a number of standard benchmark datasets (Oxford Buildings 5k and 105k, and Paris 6k) and demonstrate substantial improvements in retrieval performance whilst maintaining immediate retrieval speeds. Combining these complementary methods achieves a new state-of-the-art performance on these datasets. | ['Andrew Zisserman', 'Relja Arandjelović'] | 2012-06-16 | null | null | null | cvpr-2012-6 | ['image-augmentation', 'image-matching'] | ['computer-vision', 'computer-vision'] | [ 2.12500647e-01 -7.11872339e-01 -3.70726466e-01 -1.94853008e-01
-1.20508850e+00 -8.53084743e-01 8.37382376e-01 2.01935604e-01
-5.54505050e-01 3.35753143e-01 1.31173963e-02 -3.42297144e-02
-4.48920071e-01 -6.92213595e-01 -6.02357149e-01 -4.17759269e-01
-4.31053728e-01 5.84587157e-01 5.52555561e-01 -3.88810456e-01
6.64593697e-01 8.99229467e-01 -2.04591560e+00 2.77026683e-01
4.17750597e-01 1.28792667e+00 2.74922550e-01 9.57370162e-01
2.30752364e-01 5.51563978e-01 -3.41079921e-01 -3.85931805e-02
7.64665663e-01 7.12876692e-02 -1.15670085e+00 1.71009198e-01
1.10500121e+00 -7.46261358e-01 -7.81641841e-01 6.80593431e-01
6.05152726e-01 3.75497043e-01 5.75192690e-01 -1.24008536e+00
-8.47381353e-01 -2.20621288e-01 -4.65065747e-01 3.57045650e-01
6.84165239e-01 -1.30010113e-01 1.00967515e+00 -1.18768799e+00
8.77963901e-01 1.13715720e+00 4.17883486e-01 -1.11481130e-01
-1.10920131e+00 -4.63653117e-01 -1.60795197e-01 5.03958821e-01
-2.03749681e+00 -4.96904999e-01 3.29032570e-01 1.69661478e-03
1.16750860e+00 4.40964490e-01 7.65825510e-01 2.93851227e-01
-1.20496795e-01 1.12524021e+00 7.38822222e-01 -7.18036890e-01
5.86675555e-02 -5.14444001e-02 1.06772825e-01 6.06661737e-01
-1.30623668e-01 2.96583205e-01 -6.28477097e-01 -5.10215938e-01
7.56146371e-01 1.02769241e-01 2.02264711e-02 -7.18558371e-01
-1.56744647e+00 7.68345058e-01 7.79156327e-01 7.85742328e-02
-4.51251298e-01 5.07983685e-01 6.16516173e-01 5.88640273e-01
3.60686034e-01 4.64645028e-01 -3.27296644e-01 -2.34083503e-01
-1.25564742e+00 6.51391387e-01 5.93274117e-01 1.40011954e+00
9.61599171e-01 -3.92114699e-01 -1.58949688e-01 6.95419073e-01
2.13827550e-01 8.72385204e-01 5.23382187e-01 -9.65683997e-01
3.81969273e-01 3.16122442e-01 2.76252866e-01 -1.01445675e+00
7.38706142e-02 6.32775649e-02 -2.72455335e-01 -1.10538043e-01
9.81487054e-03 7.27479935e-01 -1.20039737e+00 1.03239393e+00
3.66882771e-01 1.92378938e-01 8.53243750e-03 7.02928603e-01
8.08582544e-01 6.12592041e-01 -1.71645626e-01 4.08855349e-01
1.43439221e+00 -1.28438210e+00 -4.61462706e-01 1.71371009e-02
8.42010498e-01 -1.13342178e+00 9.26907659e-01 2.41721407e-01
-9.69333291e-01 -7.03750074e-01 -1.05760801e+00 -5.13061404e-01
-9.18729484e-01 4.56177443e-02 9.04861212e-01 4.08988833e-01
-1.64081562e+00 4.13666129e-01 -5.98755121e-01 -7.29395628e-01
1.37921259e-01 7.91149557e-01 -6.79736018e-01 -4.94862795e-01
-1.03265142e+00 7.02099204e-01 4.95713711e-01 -2.10784107e-01
-5.83112299e-01 -6.42526090e-01 -9.11297917e-01 -6.82828575e-02
2.28675619e-01 -4.17833418e-01 1.29313147e+00 -5.37937462e-01
-9.31674659e-01 1.13096082e+00 -1.34641767e-01 -3.61896604e-01
2.91896492e-01 -5.01727939e-01 -3.05247575e-01 7.56128788e-01
3.09233963e-01 1.29569173e+00 9.84165192e-01 -8.88793409e-01
-7.63202310e-01 -4.86072570e-01 3.43190059e-02 4.56268281e-01
-4.31273937e-01 1.71857342e-01 -1.17653775e+00 -8.31300795e-01
9.52070504e-02 -1.32172632e+00 -1.07746787e-01 4.15604889e-01
-1.57310322e-01 -3.18750441e-01 1.40612602e+00 -3.57474923e-01
1.11327708e+00 -2.20226407e+00 -9.31734592e-02 5.22897780e-01
-3.48094925e-02 4.16678101e-01 -5.52438200e-01 8.14496458e-01
1.27040958e-02 3.33377905e-02 2.81215161e-01 -2.46023759e-01
2.30289400e-02 2.51315743e-01 -5.68714797e-01 5.53876817e-01
1.02058187e-01 1.25695372e+00 -8.93595517e-01 -8.49592268e-01
5.38618863e-01 5.81346750e-01 -4.78632390e-01 -6.56945035e-02
2.28048354e-01 -1.26547202e-01 -5.89165449e-01 1.05894160e+00
6.19167387e-01 -3.62592906e-01 -4.11448419e-01 -2.53832817e-01
-1.42364472e-01 -4.46432643e-02 -1.29001153e+00 2.01714015e+00
-1.39504716e-01 5.81738472e-01 -1.07300229e-01 -7.11237669e-01
7.49765754e-01 1.43660262e-01 7.22010970e-01 -8.25697720e-01
-4.14205015e-01 3.48814368e-01 -7.21771836e-01 -2.86978006e-01
1.20988011e+00 6.71796739e-01 -1.19785755e-03 4.38229114e-01
-7.92045295e-02 -4.97054696e-01 4.64902550e-01 5.94978333e-01
1.15286982e+00 9.58111510e-02 2.26833448e-01 -4.70670462e-01
2.71744877e-01 2.75714755e-01 -4.06523615e-01 1.20151174e+00
-7.04257339e-02 7.14393377e-01 -3.24779332e-01 -6.49552047e-01
-1.07403433e+00 -9.52031076e-01 -3.83521587e-01 1.30681634e+00
4.80933964e-01 -7.91583240e-01 -1.08115204e-01 -4.59679097e-01
4.04758662e-01 -1.64756373e-01 -6.41955376e-01 3.81162651e-02
-4.48108584e-01 -1.85942143e-01 6.36354327e-01 6.93355322e-01
7.61291087e-01 -9.92238641e-01 -6.56333685e-01 -7.98315033e-02
-1.35754168e-01 -9.70062494e-01 -9.07731056e-01 1.23868011e-01
-9.17329490e-01 -9.97757792e-01 -1.03572416e+00 -9.55392599e-01
7.08507180e-01 8.30438197e-01 1.30248845e+00 6.50377214e-01
-7.66694605e-01 1.09654891e+00 -6.17271364e-01 -2.26438403e-01
2.82037556e-01 1.71483874e-01 -6.06877133e-02 -4.39200938e-01
5.29922903e-01 -2.44227257e-02 -7.26881146e-01 4.24476594e-01
-1.29868317e+00 -5.35986722e-01 5.53777218e-01 9.37237859e-01
9.36248183e-01 -8.94230753e-02 1.35235533e-01 -3.28998595e-01
3.24955821e-01 -1.90597191e-01 -7.70026028e-01 4.03403014e-01
-7.82131493e-01 8.88856873e-02 -4.23017219e-02 -3.42630506e-01
-5.54226220e-01 4.74927902e-01 1.10770002e-01 -4.99549031e-01
1.14883184e-01 3.98203760e-01 4.15808380e-01 -7.57500350e-01
5.21825314e-01 4.62609917e-01 -6.69331178e-02 -5.48321068e-01
6.37332499e-01 6.31214380e-01 6.30054832e-01 -6.19644403e-01
9.26805794e-01 7.12791383e-01 1.85572013e-01 -1.05451512e+00
-2.64406532e-01 -1.32000291e+00 -1.11941016e+00 1.04115456e-01
3.36513191e-01 -1.08936548e+00 -4.23789650e-01 4.16401327e-01
-7.12681592e-01 -8.70365724e-02 -5.48660934e-01 4.60368663e-01
-6.70996428e-01 5.67007661e-01 -5.80382705e-01 -3.58217776e-01
-4.88641888e-01 -1.19195318e+00 1.95501351e+00 2.80389562e-02
1.07430719e-01 -7.20378339e-01 1.59533083e-01 6.74293144e-03
5.20011067e-01 -1.81080803e-01 5.35740852e-01 -6.53862536e-01
-1.06401134e+00 -7.56529391e-01 -6.94866598e-01 5.07654510e-02
4.65421043e-02 2.70270649e-02 -8.19646120e-01 -6.93208635e-01
-7.97649264e-01 -7.72555947e-01 9.10516977e-01 2.52831012e-01
9.01716292e-01 3.18294428e-02 -6.58661425e-01 5.99136531e-01
1.62803483e+00 1.38333529e-01 1.01642048e+00 7.81369150e-01
2.60148644e-01 2.51629829e-01 1.03940117e+00 3.01607519e-01
3.17184657e-01 1.01742864e+00 2.89234996e-01 -2.07888514e-01
-1.41314417e-01 -1.87450275e-01 -2.05554180e-02 5.16435385e-01
-6.77730748e-03 -1.43694609e-01 -7.17870474e-01 9.18617725e-01
-1.83871710e+00 -9.93603885e-01 1.86851501e-01 2.35576916e+00
4.71491337e-01 -4.11725849e-01 4.74907570e-02 1.56676605e-01
3.30832303e-01 2.68270493e-01 -3.33101958e-01 3.79179679e-02
-1.46554515e-01 4.95009601e-01 7.46311188e-01 2.82774001e-01
-1.42158425e+00 9.95761752e-01 7.81003523e+00 9.99087393e-01
-9.24964964e-01 -2.60463297e-01 2.19804570e-01 -1.20480061e-01
1.49558991e-01 4.95675206e-03 -9.20527935e-01 -2.50312430e-03
7.36799181e-01 -1.84077680e-01 4.13826108e-01 1.03954697e+00
-3.79370570e-01 -2.91175902e-01 -1.13852406e+00 1.25755358e+00
2.66257614e-01 -1.29663444e+00 1.50358632e-01 3.80582720e-01
6.95394933e-01 3.75357002e-01 3.40762973e-01 4.23414886e-01
2.11192407e-02 -7.90059686e-01 5.57731867e-01 3.89737189e-01
9.70945358e-01 -5.91964960e-01 5.62215805e-01 -2.28083774e-01
-1.73677027e+00 1.10452965e-01 -4.60370064e-01 3.52068424e-01
-1.49163097e-01 -4.27297503e-02 -8.19333434e-01 5.36936820e-01
1.01436114e+00 8.45960975e-01 -1.13375723e+00 1.38189614e+00
2.25681096e-01 4.38171104e-02 -7.13455081e-01 1.34306088e-01
4.41334605e-01 1.83756948e-01 3.61674875e-01 1.26987028e+00
2.35906735e-01 8.93220204e-05 5.97858310e-01 1.24550592e-02
-9.10727829e-02 3.32935572e-01 -1.01789558e+00 -5.64562120e-02
5.83634317e-01 1.12390208e+00 -9.36051369e-01 -5.70192993e-01
-4.59656894e-01 1.22858548e+00 -1.02863699e-01 3.74291807e-01
-4.10911500e-01 -8.19946587e-01 4.55783129e-01 5.22303432e-02
7.63668835e-01 -1.30822554e-01 4.42185104e-01 -8.96570265e-01
2.11900607e-01 -6.48857355e-01 5.97341537e-01 -9.72657859e-01
-1.08494484e+00 5.08100867e-01 4.21595514e-01 -1.41045630e+00
-5.12492001e-01 -5.51995277e-01 6.71259463e-02 7.90621281e-01
-1.73765349e+00 -1.40167677e+00 -3.19071233e-01 9.87608969e-01
4.28361028e-01 6.78689331e-02 1.00303483e+00 6.56724691e-01
3.29195410e-01 5.73260665e-01 6.17631853e-01 2.45748218e-02
1.01906586e+00 -1.20204997e+00 6.18936539e-01 3.90493304e-01
4.77145165e-01 6.75478935e-01 2.20424443e-01 -5.46144724e-01
-1.73370278e+00 -9.41656291e-01 7.87525117e-01 -5.52508056e-01
5.47645330e-01 -1.11217603e-01 -6.30977511e-01 7.26326048e-01
6.45765960e-02 5.52428663e-01 3.66752505e-01 -1.51436850e-01
-3.92303437e-01 -1.60627425e-01 -1.20672429e+00 2.60841638e-01
8.78012121e-01 -9.72377837e-01 -3.97240251e-01 8.49835694e-01
7.54947305e-01 -5.73679030e-01 -9.24645185e-01 2.84139216e-01
8.63637269e-01 -6.12348795e-01 1.56406963e+00 -5.22671700e-01
-4.94861044e-02 -3.29614252e-01 -6.06424332e-01 -7.02021480e-01
-2.37374023e-01 -4.20818239e-01 -1.37423277e-01 8.46354008e-01
-3.43829989e-02 -3.41065645e-01 7.33464777e-01 6.00238502e-01
1.26399204e-01 -6.54536128e-01 -8.75613153e-01 -1.07704043e+00
-5.20855665e-01 -2.88309813e-01 6.52961791e-01 5.63232541e-01
-2.67207503e-01 -1.65722221e-01 -3.14462215e-01 -3.94738019e-02
4.35337007e-01 1.59210131e-01 9.19343948e-01 -1.04859352e+00
1.63610712e-01 -1.14979409e-02 -1.04564309e+00 -1.56919098e+00
-1.82451814e-01 -7.15414882e-01 -1.88765414e-02 -1.31713605e+00
2.20350876e-01 -5.02875388e-01 -4.43118244e-01 5.16435385e-01
2.49586597e-01 9.04060245e-01 3.20848376e-01 6.77022398e-01
-1.07404041e+00 4.97229278e-01 1.01668644e+00 -4.01644796e-01
-1.42752379e-01 -1.84630573e-01 -3.76144916e-01 2.19021395e-01
-1.67666539e-03 -2.91580081e-01 -5.66194713e-01 -4.31714714e-01
-6.93495497e-02 -1.35756344e-01 5.96464753e-01 -8.75186503e-01
5.05145550e-01 2.52605796e-01 4.72023576e-01 -1.18744707e+00
7.15039074e-01 -9.58898425e-01 -1.24462605e-01 3.21527898e-01
-3.42508942e-01 8.18880141e-01 3.19615602e-01 7.56679058e-01
-5.39643168e-01 -2.22727746e-01 4.73584980e-01 -1.14255324e-01
-1.33039832e+00 4.50347692e-01 -4.92029749e-02 -1.33877844e-01
9.55430031e-01 -3.56830478e-01 -3.25007200e-01 -4.92496103e-01
-4.27632809e-01 2.66151786e-01 6.40636921e-01 5.93862593e-01
7.65312672e-01 -1.70228815e+00 -4.88224208e-01 2.53882140e-01
5.59248686e-01 -2.65674531e-01 -3.05378083e-02 6.54615998e-01
-8.79842341e-01 1.00104249e+00 9.82739329e-02 -8.67961645e-01
-1.57503402e+00 8.89893889e-01 -1.57138437e-01 -3.51729929e-01
-7.58530617e-01 8.47292125e-01 -5.33904769e-02 -2.24456489e-01
3.31125438e-01 -1.67758271e-01 1.26307622e-01 5.05861305e-02
9.27410424e-01 3.21998149e-01 4.33361918e-01 -9.48346317e-01
-5.88953614e-01 7.35790670e-01 -5.49201727e-01 -1.17885746e-01
1.18616033e+00 -2.79671520e-01 -7.45091438e-02 -9.24652368e-02
1.64114690e+00 -1.29759192e-01 -6.48928583e-01 -5.59217751e-01
1.52952015e-01 -1.04338741e+00 2.27080196e-01 -4.75036114e-01
-9.61256683e-01 3.41249704e-01 1.07094455e+00 1.49352655e-01
1.27279663e+00 1.17757164e-01 7.63029754e-01 1.09170508e+00
6.05336964e-01 -1.08105588e+00 2.40125388e-01 4.08228695e-01
8.83127093e-01 -1.31862915e+00 4.63404030e-01 -7.89961666e-02
-2.49930471e-01 1.07956290e+00 1.37122065e-01 -3.30019176e-01
9.15592790e-01 -6.19947463e-02 -8.53590593e-02 -6.23470426e-01
-5.06855965e-01 -3.45519453e-01 6.33906186e-01 6.10155046e-01
6.44322932e-02 -3.93844754e-01 4.86396439e-02 -4.85398889e-01
5.95931262e-02 8.99308026e-02 -7.95408711e-02 1.62175930e+00
-1.91049799e-01 -1.15086746e+00 -6.44514441e-01 5.64182818e-01
-2.73898125e-01 -4.48770165e-01 -4.03327167e-01 1.20515430e+00
-4.47834402e-01 7.19169497e-01 9.42586288e-02 -2.25118771e-01
4.01335418e-01 -1.19287096e-01 5.09853423e-01 -5.01144946e-01
-4.50513989e-01 1.29841790e-01 -3.50149959e-01 -9.74193394e-01
-7.01178968e-01 -5.90789199e-01 -8.56612325e-01 -1.23544037e-01
-7.55591512e-01 3.47536385e-01 7.51077533e-01 6.47817433e-01
6.66578054e-01 -9.70508829e-02 4.81327802e-01 -1.28532195e+00
-5.84961295e-01 -6.99092627e-01 -8.86264682e-01 4.79814649e-01
5.20026028e-01 -6.85441673e-01 -1.37384832e-01 2.33678162e-01] | [10.651185989379883, 0.4356978237628937] |
00b9706e-6c44-4725-93b0-fc4e9539a0ef | mean-shift-mask-transformer-for-unseen-object | 2211.11679 | null | https://arxiv.org/abs/2211.11679v2 | https://arxiv.org/pdf/2211.11679v2.pdf | Mean Shift Mask Transformer for Unseen Object Instance Segmentation | Segmenting unseen objects from images is a critical perception skill that a robot needs to acquire. In robot manipulation, it can facilitate a robot to grasp and manipulate unseen objects. Mean shift clustering is a widely used method for image segmentation tasks. However, the traditional mean shift clustering algorithm is not differentiable, making it difficult to integrate it into an end-to-end neural network training framework. In this work, we propose the Mean Shift Mask Transformer (MSMFormer), a new transformer architecture that simulates the von Mises-Fisher (vMF) mean shift clustering algorithm, allowing for the joint training and inference of both the feature extractor and the clustering. Its central component is a hypersphere attention mechanism, which updates object queries on a hypersphere. To illustrate the effectiveness of our method, we apply MSMFormer to unseen object instance segmentation. Our experiments show that MSMFormer achieves competitive performance compared to state-of-the-art methods for unseen object instance segmentation. The video and code are available at https://irvlutd.github.io/MSMFormer | ['Yu Xiang', 'Nicholas Ruozzi', 'Yuqiao Chen', 'Yangxiao Lu'] | 2022-11-21 | null | null | null | null | ['unseen-object-instance-segmentation', 'robot-manipulation'] | ['computer-vision', 'robots'] | [ 7.63008818e-02 2.24587657e-02 -1.22069595e-02 -4.79324877e-01
-3.91936570e-01 -6.79103851e-01 3.56070817e-01 -2.12815732e-01
-5.20270050e-01 9.43846703e-02 -5.99482238e-01 -1.15955472e-01
-9.47566330e-03 -4.12376791e-01 -1.12890542e+00 -8.06363881e-01
1.75275072e-01 9.35677946e-01 2.97861248e-01 1.05766002e-02
2.99268126e-01 6.46776915e-01 -1.45568061e+00 7.56483600e-02
1.02010942e+00 1.09299564e+00 9.12167966e-01 6.04636014e-01
2.96319108e-02 3.22008103e-01 -4.33270425e-01 -2.02384740e-01
4.81246740e-01 -1.67480022e-01 -9.55625951e-01 1.86181739e-01
3.99459451e-01 -3.60071957e-01 -3.57182473e-01 1.33454466e+00
1.10700831e-01 6.77620828e-01 9.93496478e-01 -1.41932595e+00
-7.95144200e-01 6.97609246e-01 -7.54187703e-01 1.87843487e-01
-1.94666103e-01 1.44672960e-01 7.67088413e-01 -9.94180322e-01
5.32112420e-01 1.40502322e+00 1.16392769e-01 6.65175021e-01
-1.08957994e+00 -6.52173817e-01 3.39360654e-01 2.68553644e-01
-1.23617887e+00 -1.84346795e-01 7.51041472e-01 -5.63369274e-01
6.58272445e-01 2.49752812e-02 5.46872497e-01 7.70800591e-01
-1.59956906e-02 1.45232356e+00 4.89783227e-01 3.23237702e-02
2.76306719e-01 -2.53395475e-02 1.65383935e-01 6.94844306e-01
1.76388055e-01 -1.40445948e-01 -9.57188755e-02 2.47610331e-01
1.00195909e+00 4.29021597e-01 -4.11359191e-01 -8.61904979e-01
-1.36332977e+00 7.00951695e-01 1.13223290e+00 1.51050314e-01
-2.81035364e-01 5.26828110e-01 -3.53409946e-02 -1.24115877e-01
1.88631579e-01 5.16189456e-01 -3.27517122e-01 1.57224342e-01
-6.99929178e-01 8.41429681e-02 7.42185354e-01 1.24649918e+00
7.80612886e-01 -2.34497160e-01 -8.64656866e-02 6.25456929e-01
3.76683295e-01 5.90924323e-01 3.03012818e-01 -1.27537465e+00
-1.82578748e-03 5.81680298e-01 2.19775662e-01 -7.78578699e-01
-4.36275840e-01 -3.22951049e-01 -7.53071845e-01 1.74428508e-01
3.03811640e-01 8.00881628e-03 -1.33838558e+00 1.45633936e+00
5.21529078e-01 4.04824227e-01 -2.62584556e-02 1.50965178e+00
7.62753844e-01 5.94524562e-01 -1.79362923e-01 2.89817035e-01
1.12922382e+00 -1.38432264e+00 -4.30325896e-01 -4.44268465e-01
2.08976477e-01 -5.17770410e-01 9.48040664e-01 3.65479112e-01
-9.85294759e-01 -5.77386618e-01 -8.52844059e-01 -3.63424897e-01
-4.91309196e-01 1.90421343e-01 7.45492101e-01 1.62605010e-02
-8.44165742e-01 7.46477962e-01 -1.36768568e+00 -3.24322701e-01
6.70251548e-01 6.12703979e-01 -1.03836112e-01 4.25239317e-02
-6.72681212e-01 5.56726456e-01 7.08019912e-01 3.89299124e-01
-1.10820186e+00 -4.35661018e-01 -8.45373750e-01 1.02806635e-01
6.37500226e-01 -5.53903103e-01 1.46445072e+00 -7.09616065e-01
-1.47657835e+00 7.32633173e-01 -2.34788090e-01 -3.76298398e-01
4.29435700e-01 -4.36370999e-01 4.68136877e-01 3.15166414e-01
1.01891138e-01 1.34691405e+00 1.06887662e+00 -1.54654706e+00
-6.24096930e-01 -3.82869184e-01 -7.66838863e-02 3.31987500e-01
9.33748409e-02 -4.85637963e-01 -8.43995929e-01 -3.79270077e-01
4.79880452e-01 -1.03831637e+00 -3.28853160e-01 -8.40866566e-02
-6.38052821e-01 -4.96874124e-01 1.05095220e+00 -3.14952880e-01
4.77353245e-01 -2.27017760e+00 6.58125520e-01 9.30589885e-02
2.08212629e-01 1.23298883e-01 -1.24467313e-01 -2.53889598e-02
1.80447087e-01 -8.13799426e-02 -6.70107543e-01 -4.36254054e-01
2.67395020e-01 1.61221266e-01 -2.15192065e-01 6.68478310e-01
1.83949292e-01 1.13281071e+00 -9.92128968e-01 -2.86475837e-01
4.51616555e-01 4.83504921e-01 -5.78649759e-01 2.40935773e-01
-5.53126991e-01 6.70532346e-01 -4.55553800e-01 6.44604683e-01
7.38209307e-01 -4.78663325e-01 -2.27261543e-01 -2.36269996e-01
5.62491231e-02 -4.72942442e-01 -9.26191568e-01 2.04166865e+00
1.10521108e-01 6.18821740e-01 3.87352675e-01 -1.12523496e+00
7.64152944e-01 -1.69077635e-01 4.68140066e-01 -1.26214772e-01
5.53173423e-01 2.29962304e-01 -4.08007987e-02 -6.53959572e-01
4.37478125e-01 2.29266867e-01 2.94618364e-02 2.89924234e-01
1.32293403e-01 -5.51654577e-01 3.80440801e-01 3.15541685e-01
6.58599734e-01 1.20951794e-01 -2.56287724e-01 -3.06192070e-01
3.29856187e-01 2.25166202e-01 4.10896301e-01 6.83941543e-01
-2.71459043e-01 6.62381113e-01 1.64907157e-01 -2.50219941e-01
-7.47552812e-01 -1.18161619e+00 -1.58727854e-01 1.17010975e+00
8.63697112e-01 3.03827852e-01 -1.09586775e+00 -6.31132483e-01
1.46838129e-01 6.78327084e-01 -5.20757914e-01 -2.44886875e-01
-5.00342548e-01 -5.71338713e-01 -1.86858159e-02 5.84704578e-01
7.06070006e-01 -1.25794601e+00 -7.22853065e-01 1.94177707e-03
-3.32578838e-01 -1.14969778e+00 -6.52714550e-01 2.89245754e-01
-6.94107950e-01 -1.23561609e+00 -7.54393697e-01 -1.18390310e+00
8.82444263e-01 6.42394125e-01 6.89725280e-01 9.02474858e-03
-4.84638751e-01 5.45254409e-01 -3.31384331e-01 -4.85968620e-01
1.23084066e-02 3.32198054e-01 6.97391480e-02 -3.08629274e-02
5.04473627e-01 -1.66664094e-01 -8.93946290e-01 4.31437105e-01
-1.14019120e+00 -7.19079226e-02 7.34529495e-01 5.83600283e-01
8.22956145e-01 -4.16935645e-02 2.61421919e-01 -6.13214135e-01
2.85663575e-01 -4.13258463e-01 -8.25263381e-01 8.64965543e-02
-2.56575406e-01 -9.74301100e-02 2.31846049e-01 -4.94641006e-01
-7.43835270e-01 4.96401668e-01 2.08854839e-01 -9.39915359e-01
-2.47591212e-01 1.61804035e-01 -3.63630392e-02 -6.92267716e-02
2.26171926e-01 4.80948538e-02 1.70715913e-01 -4.52426851e-01
5.15780330e-01 7.53673077e-01 9.56773520e-01 -4.39944118e-01
7.92210877e-01 7.12977648e-01 -2.44610474e-01 -7.87730157e-01
-9.15079951e-01 -8.27105522e-01 -9.36608315e-01 -2.52509803e-01
1.36581707e+00 -6.98747277e-01 -1.06900096e+00 5.14597356e-01
-1.09051287e+00 -7.64162600e-01 -8.22622627e-02 5.06747544e-01
-8.25611711e-01 1.68865979e-01 -6.63849533e-01 -6.51078820e-01
-2.57024854e-01 -1.47849584e+00 1.40633821e+00 5.37682474e-01
2.32337102e-01 -7.35560954e-01 -5.56404889e-01 5.16964436e-01
1.23771481e-01 2.97539294e-01 5.95258534e-01 -7.46922553e-01
-1.07033467e+00 4.00831588e-02 -3.93934935e-01 3.19108129e-01
5.96703254e-02 -1.19230010e-01 -8.32567811e-01 -5.23636758e-01
2.66153395e-01 -2.22482488e-01 1.30693388e+00 7.04537749e-01
1.45779335e+00 -2.53464393e-02 -6.68627799e-01 8.21349680e-01
1.11597133e+00 2.88700432e-01 3.22111279e-01 1.90670133e-01
1.00450265e+00 5.36451280e-01 7.61374772e-01 3.90053466e-02
3.18373173e-01 3.23428959e-01 8.02038550e-01 -1.31907269e-01
2.07145661e-01 1.89260915e-01 1.39965460e-01 4.80879515e-01
1.91319019e-01 -2.64736444e-01 -9.48071480e-01 4.88843858e-01
-2.09063601e+00 -6.66717350e-01 -1.09379560e-01 1.80400610e+00
4.67292011e-01 1.72219425e-02 -1.97133750e-01 -3.31339598e-01
9.42021728e-01 -1.69896796e-01 -1.16062570e+00 6.53747236e-03
2.69777089e-01 -2.45985180e-01 4.45018381e-01 5.20081401e-01
-1.43773854e+00 1.51308429e+00 4.72958136e+00 6.48530841e-01
-1.02314758e+00 -6.64455444e-02 4.83294278e-01 -1.06002919e-01
1.70098484e-01 -3.50956023e-01 -5.81993937e-01 2.93497741e-01
2.26733401e-01 6.67527542e-02 8.15490186e-01 1.12835026e+00
-1.00702364e-02 -1.65462136e-01 -1.39358783e+00 1.09510887e+00
1.53191298e-01 -9.69963908e-01 -1.31748751e-01 -5.56642599e-02
7.75643110e-01 4.69446808e-01 3.16175163e-01 2.54073799e-01
2.59777218e-01 -9.13794041e-01 7.77085543e-01 4.90363926e-01
3.00001323e-01 -6.51623428e-01 5.82719922e-01 4.48435634e-01
-1.02293408e+00 -1.48053646e-01 -5.69757879e-01 4.17632103e-01
1.87562555e-01 3.73467326e-01 -8.80851805e-01 1.35325760e-01
9.53258872e-01 7.84335017e-01 -3.91083658e-01 1.18468118e+00
-2.18070164e-01 1.04348518e-01 -4.80351835e-01 -7.25516081e-02
5.06007373e-01 -5.56469619e-01 6.96901083e-01 9.53331411e-01
1.25758544e-01 2.37734437e-01 4.59335387e-01 1.28917146e+00
-2.48318613e-01 -3.41155022e-01 -3.50260496e-01 -1.67664260e-01
4.14702594e-01 1.40769482e+00 -1.33977187e+00 -3.94473076e-01
5.00884168e-02 1.24782133e+00 2.66808361e-01 6.22616768e-01
-8.85210276e-01 -6.00025892e-01 6.48298502e-01 -3.21855336e-01
7.22110033e-01 -4.51611519e-01 -3.88412811e-02 -1.08793998e+00
-7.89707452e-02 -5.43289840e-01 1.19006060e-01 -8.36616457e-01
-1.32564425e+00 4.75236982e-01 1.06023841e-01 -9.33213949e-01
7.61557296e-02 -9.12124515e-01 -4.19675440e-01 4.26189810e-01
-1.40610492e+00 -9.93053496e-01 -6.95641756e-01 4.53649729e-01
7.71649957e-01 1.84429377e-01 1.77352369e-01 -2.42038723e-02
-7.18995631e-01 -4.53914423e-03 1.56295747e-01 3.87145817e-01
4.55016017e-01 -1.49750626e+00 3.98119926e-01 7.00549304e-01
4.84563410e-02 5.92665851e-01 6.43803179e-01 -7.13726759e-01
-1.52068138e+00 -1.29671657e+00 -6.11206703e-02 -5.80157399e-01
5.50016403e-01 -5.34658253e-01 -1.11179280e+00 8.39794219e-01
2.14320108e-01 -3.11280023e-02 2.55762428e-01 -4.73723918e-01
1.79262891e-01 1.15935735e-01 -1.06382418e+00 6.52982116e-01
9.69514430e-01 -1.21610545e-01 -6.04464471e-01 5.24146140e-01
9.79754329e-01 -5.92574656e-01 -6.85839355e-01 6.28134608e-01
1.81999311e-01 -7.47943938e-01 8.11605752e-01 -2.38323927e-01
1.02048218e-01 -5.43347955e-01 -7.62140378e-02 -1.27767169e+00
-2.28531152e-01 -6.13131523e-01 -1.30519181e-01 8.61108005e-01
3.68164629e-01 -5.99379182e-01 9.30821061e-01 6.93753123e-01
-4.62415963e-01 -7.26380348e-01 -6.54708803e-01 -8.76779556e-01
1.18977018e-01 -2.21411809e-01 5.72657585e-01 7.07232893e-01
-1.13348067e-01 4.02947754e-01 2.36153096e-01 4.20331687e-01
8.43890250e-01 3.86229455e-01 6.89785779e-01 -1.37979448e+00
-8.86855423e-02 -7.84062624e-01 -2.09824234e-01 -1.52225554e+00
2.70650625e-01 -1.08924532e+00 7.74256647e-01 -1.83152616e+00
2.27598578e-01 -2.69822091e-01 -1.32549852e-01 3.85619372e-01
-2.14648962e-01 6.27879426e-02 2.94261545e-01 4.59180504e-01
-8.33152592e-01 7.63506830e-01 1.55352700e+00 -4.20719445e-01
-3.84938091e-01 6.58251345e-02 -3.85070533e-01 7.96475768e-01
9.85109746e-01 -4.24895525e-01 -4.43312883e-01 -7.30133593e-01
-2.65913457e-01 -3.42252523e-01 4.72368747e-01 -1.00111473e+00
5.53396881e-01 -1.02234855e-01 4.80272621e-01 -7.98798740e-01
5.07221043e-01 -7.82968998e-01 -3.25868934e-01 6.02133751e-01
-1.75234988e-01 -1.58401698e-01 -5.92154497e-03 7.38068521e-01
-9.43969563e-02 -3.81996125e-01 8.61578226e-01 -3.10306102e-01
-8.38654101e-01 6.19310617e-01 -1.89491615e-01 -7.68779367e-02
1.27764535e+00 -1.31811276e-01 -2.91846067e-01 -2.82858275e-02
-7.55487204e-01 8.10806692e-01 6.21590316e-01 5.43522358e-01
8.60572457e-01 -9.19054091e-01 -3.24989080e-01 1.37119740e-01
-1.34633094e-01 9.35350537e-01 2.49990486e-02 1.06713653e+00
-6.65824533e-01 2.99667776e-01 8.40081125e-02 -1.05495751e+00
-7.93111444e-01 8.42313111e-01 4.43592519e-01 5.13012409e-01
-6.73235595e-01 1.21275389e+00 6.17708087e-01 -5.43874800e-01
5.85201919e-01 -5.86773455e-01 -1.92859732e-02 -1.76500618e-01
2.78198779e-01 3.43164325e-01 -3.30411404e-01 -4.61776674e-01
-2.27635279e-01 5.43015480e-01 -1.72799006e-01 -5.20731620e-02
1.24982584e+00 -2.80915618e-01 -3.68503749e-01 4.78231460e-01
1.16415858e+00 -7.31653512e-01 -1.59281099e+00 -4.49287407e-02
1.20800231e-02 -1.72377169e-01 2.39295373e-03 -4.08380419e-01
-1.25275254e+00 9.60324287e-01 5.32269597e-01 1.64222494e-01
8.72586966e-01 4.59644765e-01 9.74709094e-01 9.60211456e-01
4.03458685e-01 -1.11233473e+00 3.52113277e-01 6.00325942e-01
9.56122398e-01 -1.59008157e+00 -3.01393390e-01 -4.97552633e-01
-7.24816203e-01 9.40384686e-01 9.00046766e-01 -3.64455283e-01
7.48266280e-01 4.47277725e-03 1.48117945e-01 -4.32224721e-01
-2.82769501e-01 -4.24100280e-01 4.15913492e-01 4.44797844e-01
-1.50013804e-01 1.75240580e-02 3.46094280e-01 4.00331795e-01
-2.65919238e-01 -3.80000234e-01 4.01417911e-01 9.67428803e-01
-8.33050787e-01 -5.44749856e-01 -1.88900247e-01 4.25324082e-01
-1.67820677e-01 1.70422658e-01 -5.66530943e-01 7.05885231e-01
2.85912529e-02 9.21013951e-01 3.82359833e-01 -3.26664686e-01
7.85861611e-02 -1.60653725e-01 3.79001588e-01 -7.45849431e-01
-1.73270226e-01 2.52743483e-01 -7.69557178e-01 -6.97003722e-01
-4.03803229e-01 -6.44312501e-01 -1.90167570e+00 9.23096016e-02
-4.59697247e-01 3.43281388e-01 8.94446552e-01 8.88900578e-01
3.24585289e-01 6.05933845e-01 4.20240939e-01 -1.45495665e+00
-3.04249495e-01 -9.90914047e-01 -4.53825593e-01 3.61314118e-01
4.77265984e-01 -8.15827489e-01 -3.69581074e-01 2.18727976e-01] | [7.906572341918945, -2.819173812866211] |
c9d9595b-9b26-4f65-8983-41255716114c | modelling-emotion-dynamics-in-song-lyrics | 2210.09434 | null | https://arxiv.org/abs/2210.09434v1 | https://arxiv.org/pdf/2210.09434v1.pdf | Modelling Emotion Dynamics in Song Lyrics with State Space Models | Most previous work in music emotion recognition assumes a single or a few song-level labels for the whole song. While it is known that different emotions can vary in intensity within a song, annotated data for this setup is scarce and difficult to obtain. In this work, we propose a method to predict emotion dynamics in song lyrics without song-level supervision. We frame each song as a time series and employ a State Space Model (SSM), combining a sentence-level emotion predictor with an Expectation-Maximization (EM) procedure to generate the full emotion dynamics. Our experiments show that applying our method consistently improves the performance of sentence-level baselines without requiring any annotated songs, making it ideal for limited training data scenarios. Further analysis through case studies shows the benefits of our method while also indicating the limitations and pointing to future directions. | ['Daniel Beck', 'Yingjin Song'] | 2022-10-17 | null | null | null | null | ['music-emotion-recognition'] | ['music'] | [ 1.92195177e-01 -4.62126881e-01 -1.13063157e-01 -4.57823515e-01
-8.28037202e-01 -9.01419938e-01 5.75794160e-01 -2.84510374e-01
-2.18759835e-01 5.39985478e-01 3.68594795e-01 7.87680447e-02
1.33813873e-01 -3.01752210e-01 -5.63915431e-01 -5.34738481e-01
-1.29533544e-01 1.40278593e-01 -2.66750723e-01 -2.67378151e-01
5.56870960e-02 2.36011878e-01 -1.70468509e+00 3.79586458e-01
4.04353619e-01 9.73499298e-01 -3.54049467e-02 1.02340543e+00
-1.09026143e-02 9.44327652e-01 -1.13298118e+00 -1.67454273e-01
3.13940883e-01 -8.60012472e-01 -6.76936626e-01 2.44984046e-01
3.40531170e-01 7.90435970e-02 7.34525099e-02 5.70102513e-01
7.32344866e-01 4.30520564e-01 4.75408733e-01 -1.18120420e+00
5.38460091e-02 4.30076957e-01 -1.63790599e-01 -2.29314491e-02
5.22299051e-01 8.10228512e-02 1.22679734e+00 -4.70254898e-01
6.49005830e-01 7.10016608e-01 8.71228874e-01 7.31733918e-01
-1.45365059e+00 -7.80500293e-01 1.19683988e-01 2.99900956e-02
-1.19412959e+00 -5.80969334e-01 1.18139136e+00 -4.51105118e-01
1.09320927e+00 5.60823202e-01 1.08273053e+00 1.35908449e+00
-2.70046685e-02 9.24933970e-01 1.32902193e+00 -3.77850503e-01
3.66537064e-01 4.45821695e-02 -1.39836252e-01 6.70377761e-02
-8.33100975e-01 -2.52653249e-02 -1.04334068e+00 -1.34649172e-01
5.55680633e-01 -3.55188698e-01 1.74857900e-02 -1.33965090e-01
-1.06772220e+00 5.96784353e-01 5.80187403e-02 4.77879137e-01
-5.17365038e-01 3.26881349e-01 6.79771662e-01 5.81212461e-01
6.61589324e-01 7.15368688e-01 -4.56612378e-01 -8.62932324e-01
-1.62074792e+00 3.23157996e-01 1.04991663e+00 5.25704801e-01
3.92026365e-01 3.05057943e-01 -2.46416196e-01 9.49756086e-01
-1.36626333e-01 1.63019434e-01 4.65874463e-01 -1.13217962e+00
-3.10160071e-02 2.20157728e-01 1.96538016e-01 -8.19347918e-01
-3.67008865e-01 -6.32397950e-01 -4.84861463e-01 -3.07714008e-02
3.61771673e-01 -3.06038797e-01 -4.85250950e-01 2.06313872e+00
8.89235511e-02 6.99584723e-01 -6.95860237e-02 9.43914950e-01
4.07485425e-01 7.71863997e-01 6.57112598e-02 -7.25262761e-01
9.33608472e-01 -9.86483395e-01 -8.52825344e-01 -9.45089459e-02
2.83854336e-01 -8.21991205e-01 1.52509117e+00 8.58935595e-01
-1.03691113e+00 -5.19238532e-01 -8.87951791e-01 3.86700630e-01
-8.93709883e-02 1.39742717e-01 9.32240546e-01 6.80178285e-01
-8.42702746e-01 7.90980458e-01 -9.55160022e-01 -3.07108700e-01
-1.48801655e-01 4.19877619e-01 1.00440443e-01 7.47959316e-01
-1.13394678e+00 4.91815835e-01 -7.14089163e-03 1.08065875e-02
-8.78129661e-01 -7.31367052e-01 -5.39175808e-01 -5.59546165e-02
1.50277615e-01 -4.80981171e-01 1.70271850e+00 -1.38289905e+00
-2.00593567e+00 5.89848816e-01 -3.27689171e-01 -4.26462620e-01
2.75625050e-01 -2.78958380e-01 -4.46967691e-01 -1.91635452e-02
-2.72319466e-01 5.96792519e-01 8.54242146e-01 -1.04343343e+00
-4.13113236e-01 4.77501713e-02 -3.06422394e-02 4.76725787e-01
-5.69182813e-01 4.39801484e-01 -4.62696105e-01 -8.69839430e-01
-3.34086329e-01 -1.13817263e+00 -1.64543048e-01 -7.59043634e-01
-2.58274913e-01 -1.29186139e-01 5.66656470e-01 -5.46990812e-01
1.83438408e+00 -2.22137427e+00 2.67909497e-01 1.03192829e-01
-5.59273541e-01 -2.95013994e-01 -1.21976040e-01 5.92305839e-01
-1.44519791e-01 -3.64517123e-02 -2.27899969e-01 -5.20025969e-01
1.90893725e-01 2.27042347e-01 -7.68719137e-01 8.04387107e-02
1.79305312e-03 8.17511916e-01 -7.74325609e-01 -2.43504956e-01
6.40097260e-02 5.38385570e-01 -6.91438198e-01 3.71974558e-01
-4.86209273e-01 7.74906695e-01 -7.00793054e-04 4.73531455e-01
-1.63727030e-02 1.66698843e-02 2.05635920e-01 -5.70861138e-02
-2.51086682e-01 7.47535288e-01 -1.48966634e+00 2.06425691e+00
-7.34450936e-01 6.81333303e-01 7.02529699e-02 -7.16229737e-01
8.51791799e-01 7.00679064e-01 7.63822496e-01 -3.75311881e-01
1.54143618e-02 4.55592461e-02 2.24747155e-02 -3.99152398e-01
5.29817641e-01 -5.36889970e-01 -6.77586794e-01 7.10365415e-01
1.44087300e-01 -6.21370673e-01 2.97168702e-01 -5.55803068e-02
1.11208069e+00 3.58020961e-01 6.05847389e-02 1.03744075e-01
2.29176819e-01 5.28734317e-03 6.43052101e-01 7.59870231e-01
-1.00949921e-01 5.41146815e-01 3.20972055e-01 -1.01205312e-01
-8.78390729e-01 -7.21320450e-01 -8.53413567e-02 1.31790924e+00
-2.73239970e-01 -8.94448817e-01 -7.78940856e-01 -5.27394354e-01
-3.90270919e-01 8.15722704e-01 -4.75412726e-01 -3.58198583e-02
-4.89604205e-01 -7.60170579e-01 7.15103328e-01 3.03504735e-01
-2.02914476e-02 -1.55828905e+00 -7.10875034e-01 4.16370600e-01
-3.78980547e-01 -7.23589659e-01 -5.61274946e-01 4.88795757e-01
-7.40750194e-01 -4.68451649e-01 -3.38438541e-01 -4.50959802e-01
-9.62838978e-02 -3.24957281e-01 1.33583570e+00 -3.80733371e-01
-3.09197396e-01 5.79503477e-01 -2.81101257e-01 -5.76906800e-01
-3.59121948e-01 1.04315929e-01 2.10470304e-01 1.10695511e-01
3.06559950e-01 -9.01933610e-01 -5.85222125e-01 1.14536822e-01
-7.48544514e-01 6.94127008e-02 2.43127123e-01 8.62283170e-01
6.28095031e-01 1.71378493e-01 7.78684199e-01 -7.39553869e-01
9.40385759e-01 -1.91361293e-01 -1.78091556e-01 -1.75621614e-01
-6.06297612e-01 -3.20900232e-01 6.57350063e-01 -9.21739459e-01
-8.67539585e-01 3.64601701e-01 -2.84918002e-03 -4.72658068e-01
-4.95399684e-01 5.84201932e-01 5.44613525e-02 3.33908200e-01
6.10422075e-01 2.67408729e-01 -2.88588762e-01 -5.46952605e-01
3.57015282e-01 8.31255615e-01 7.23871529e-01 -6.92199528e-01
5.44497728e-01 2.09084406e-01 -2.45693892e-01 -7.06488788e-01
-1.17702174e+00 -5.59984505e-01 -4.97786552e-01 -5.90856075e-01
6.37905538e-01 -9.57227945e-01 -7.81940460e-01 2.95116454e-01
-6.23596191e-01 -7.78296590e-01 -5.38210154e-01 4.40656602e-01
-1.01058149e+00 -2.68316790e-02 -9.87970352e-01 -1.11421430e+00
-3.94122750e-01 -6.86970592e-01 1.27703774e+00 1.24602415e-01
-8.68609309e-01 -8.26108158e-01 5.41628182e-01 2.88914263e-01
3.58628243e-01 2.36834347e-01 4.25869763e-01 -5.22799432e-01
-1.32919252e-01 -1.07080489e-01 6.69895291e-01 3.71540517e-01
1.08501501e-01 5.15265092e-02 -1.13288474e+00 -1.18794471e-01
1.98855624e-01 -6.69836640e-01 5.67683101e-01 2.94296473e-01
1.20196164e+00 -1.70503289e-01 3.08343619e-01 4.48818445e-01
1.08175170e+00 1.19832985e-01 2.99344987e-01 3.77471358e-01
3.96089345e-01 5.07666409e-01 9.30335045e-01 6.85604572e-01
1.67384371e-01 1.02278793e+00 7.71718696e-02 3.60687450e-02
-3.54969539e-02 -3.40166181e-01 9.00363684e-01 1.22563219e+00
2.03052331e-02 -1.71448588e-01 -6.25783741e-01 5.80048025e-01
-1.79955482e+00 -1.26251912e+00 7.77618438e-02 2.12084508e+00
1.30380881e+00 -6.29513413e-02 6.07155442e-01 5.12262046e-01
2.58115411e-01 3.12076628e-01 -6.41511858e-01 -6.54998600e-01
-1.30169019e-01 4.87535983e-01 -1.72654718e-01 4.27343488e-01
-1.18167138e+00 1.17112195e+00 7.09131622e+00 7.43615866e-01
-1.32845485e+00 -8.84095952e-02 2.78866470e-01 -8.50984335e-01
-1.74799070e-01 7.70014301e-02 -3.97728950e-01 3.58176112e-01
1.17209125e+00 -2.23987382e-02 9.96606231e-01 5.26544571e-01
6.10442698e-01 -2.32220814e-03 -1.04824030e+00 1.05661571e+00
8.34552273e-02 -8.22658896e-01 -4.08856213e-01 -1.36856601e-01
8.73418450e-01 -1.55948654e-01 2.97556282e-03 5.87611437e-01
-3.70530523e-02 -9.40499663e-01 8.58527839e-01 5.75240433e-01
6.63002610e-01 -8.40040803e-01 2.97470748e-01 5.37488997e-01
-1.12731111e+00 6.41782656e-02 1.35156825e-01 -4.51813459e-01
3.73182774e-01 2.91752011e-01 -8.19695175e-01 3.52702230e-01
6.87062204e-01 9.58805025e-01 -4.46582049e-01 7.42030561e-01
-3.10426533e-01 1.48326480e+00 -4.13863122e-01 -6.76130354e-02
1.70924999e-02 -2.50182033e-01 7.74739385e-01 1.50834215e+00
3.69503975e-01 -7.58893415e-02 2.26468354e-01 6.70545280e-01
1.08732194e-01 3.90766889e-01 -3.09737623e-01 -3.59217465e-01
2.55059451e-01 1.47764325e+00 -6.46255076e-01 -2.88265169e-01
-2.52875894e-01 1.34992814e+00 4.56140842e-03 3.90332103e-01
-7.34450221e-01 -1.74740463e-01 6.89811945e-01 -2.51150459e-01
1.29132405e-01 -1.79508030e-01 -3.02492291e-01 -1.34237325e+00
-1.48287922e-01 -1.24047923e+00 4.33257610e-01 -9.18286443e-01
-1.19947004e+00 5.37866533e-01 -2.79467344e-01 -1.37831545e+00
-6.95571005e-01 -2.04933897e-01 -6.72845781e-01 6.46559060e-01
-9.42299545e-01 -9.61464047e-01 3.06889936e-02 5.39236307e-01
6.90843880e-01 -1.65419467e-02 1.15169811e+00 1.80261314e-01
-6.07672334e-01 4.54783201e-01 -2.08433941e-01 -1.98869690e-01
1.07807696e+00 -1.65571606e+00 -6.64323494e-02 5.04778385e-01
8.83157134e-01 5.05287886e-01 1.17877400e+00 -4.86483514e-01
-1.34894514e+00 -6.88436270e-01 9.39546883e-01 -4.63386655e-01
8.58777344e-01 -7.42513239e-01 -7.57604063e-01 4.98562634e-01
4.68547255e-01 -5.29214084e-01 1.11990857e+00 6.22079253e-01
-1.33473292e-01 -8.83196965e-02 -5.74520588e-01 6.56439722e-01
8.34240019e-01 -8.02882969e-01 -5.23544550e-01 1.40098333e-01
3.70581686e-01 -4.12736505e-01 -1.11516857e+00 5.19156039e-01
7.15290368e-01 -8.75334382e-01 5.84594488e-01 -6.18875623e-01
2.91531712e-01 -5.30571938e-01 -2.47968122e-01 -1.52527428e+00
-4.48120832e-02 -8.46240819e-01 -3.38641256e-01 1.50575423e+00
3.27088565e-01 1.63379133e-01 8.71018231e-01 5.41427433e-01
-2.72063762e-02 -5.45764923e-01 -6.65970385e-01 -7.81286478e-01
-1.48976818e-01 -1.08442879e+00 3.25066447e-01 1.00981843e+00
2.91552216e-01 6.42337978e-01 -8.42251956e-01 -7.34150708e-02
9.91967395e-02 6.53685927e-01 1.00311267e+00 -1.02256978e+00
-8.34001422e-01 -5.94056129e-01 -1.66412696e-01 -8.45667243e-01
4.93482679e-01 -8.96439850e-01 2.65074700e-01 -1.10011637e+00
3.48695308e-01 -1.29207894e-01 -6.91996098e-01 5.72277606e-01
-6.45177662e-02 6.23194814e-01 2.95639724e-01 5.84481657e-02
-8.11430573e-01 5.93711436e-01 8.49659681e-01 -3.72169241e-02
-6.23745680e-01 2.50635266e-01 -4.75719243e-01 5.54839611e-01
9.49070632e-01 -6.01064503e-01 -4.07318860e-01 1.20690845e-01
3.50361109e-01 1.28060028e-01 2.02480301e-01 -9.72352564e-01
8.40679929e-02 -1.30486533e-01 2.00259000e-01 -5.16728699e-01
7.04384446e-01 -6.65970087e-01 5.19215703e-01 7.48078153e-02
-5.50767958e-01 -1.91519082e-01 4.91960883e-01 4.52662289e-01
-5.49888313e-01 -1.13357723e-01 3.83120269e-01 1.43248603e-01
-4.64826763e-01 -4.76398580e-02 -5.97663105e-01 -1.02157593e-01
6.30412340e-01 4.47904468e-02 5.13990998e-01 -9.34543908e-01
-1.05577755e+00 -4.89841215e-02 4.89144206e-01 6.32552922e-01
-3.33620189e-03 -1.47400522e+00 -4.58305925e-01 3.17467526e-02
3.18412408e-02 -5.14367700e-01 2.20518231e-01 6.64995492e-01
2.21930027e-01 1.47396490e-01 7.42455274e-02 -6.93763196e-01
-1.49892783e+00 4.15377736e-01 3.10463399e-01 -3.64523023e-01
-2.97286659e-01 5.89022160e-01 -1.70508981e-01 -6.32312000e-01
3.64608109e-01 7.87656158e-02 -2.02244580e-01 2.94181377e-01
3.64311248e-01 9.61764455e-02 -5.07482179e-02 -6.53825700e-01
-1.90595284e-01 5.45374990e-01 4.40447003e-01 -7.02632904e-01
1.41252255e+00 -1.25697911e-01 -1.38645647e-02 1.44672441e+00
7.70215869e-01 4.18517888e-01 -1.30656815e+00 1.10209346e-01
1.93218902e-01 -1.27126679e-01 -1.13381562e-03 -1.06590748e+00
-5.77810168e-01 8.61209452e-01 4.51053649e-01 2.72138834e-01
1.62641561e+00 -1.11817949e-01 6.86504424e-01 2.05666855e-01
2.53205270e-01 -1.25489748e+00 1.52666152e-01 5.45447171e-01
7.96144187e-01 -8.01309705e-01 -3.59261394e-01 3.78511995e-02
-9.23062861e-01 8.27198863e-01 4.73095715e-01 -6.82186112e-02
4.41304177e-01 4.46079195e-01 5.48026979e-01 2.33950950e-02
-1.30786538e+00 -1.62280664e-01 4.91986096e-01 -4.63927202e-02
9.11565602e-01 1.25137106e-01 -3.27791214e-01 1.02708375e+00
-7.11710691e-01 7.09561035e-02 3.17040443e-01 6.98979855e-01
-4.06978019e-02 -1.49951744e+00 -1.78086102e-01 3.19289446e-01
-8.83239746e-01 -1.07234776e-01 -7.62470543e-01 5.01385868e-01
2.28910074e-02 1.17723882e+00 6.06214739e-02 -6.15345597e-01
4.12741333e-01 7.18293190e-01 5.26447773e-01 -6.97531104e-01
-1.20017052e+00 8.41420829e-01 2.15058133e-01 -6.29183650e-01
-6.69619203e-01 -1.00556862e+00 -1.11164129e+00 2.01496035e-01
-2.54303902e-01 3.46336126e-01 7.08043516e-01 6.72745585e-01
3.29340696e-01 6.55539930e-01 9.15938377e-01 -9.03910697e-01
-2.17022315e-01 -1.12871385e+00 -7.25513935e-01 7.15083003e-01
1.63266703e-01 -2.51110822e-01 -4.22316253e-01 4.33591872e-01] | [15.897311210632324, 5.358955383300781] |
7ba2d491-7697-464d-8d3e-c0faa9c4f24d | workflow-discovery-from-dialogues-in-the-low | 2205.11690 | null | https://arxiv.org/abs/2205.11690v2 | https://arxiv.org/pdf/2205.11690v2.pdf | Workflow Discovery from Dialogues in the Low Data Regime | Text-based dialogues are now widely used to solve real-world problems. In cases where solution strategies are already known, they can sometimes be codified into workflows and used to guide humans or artificial agents through the task of helping clients. We introduce a new problem formulation that we call Workflow Discovery (WD) in which we are interested in the situation where a formal workflow may not yet exist. Still, we wish to discover the set of actions that have been taken to resolve a particular problem. We also examine a sequence-to-sequence (Seq2Seq) approach for this novel task. We present experiments where we extract workflows from dialogues in the Action-Based Conversations Dataset (ABCD). Since the ABCD dialogues follow known workflows to guide agents, we can evaluate our ability to extract such workflows using ground truth sequences of actions. We propose and evaluate an approach that conditions models on the set of possible actions, and we show that using this strategy, we can improve WD performance. Our conditioning approach also improves zero-shot and few-shot WD performance when transferring learned models to unseen domains within and across datasets. Further, on ABCD a modified variant of our Seq2Seq method achieves state-of-the-art performance on related but different problems of Action State Tracking (AST) and Cascading Dialogue Success (CDS) across many evaluation metrics. | ['Chris Pal', 'Pau Rodriguez', 'David Vazquez', 'Issam Laradji', 'Stefania Raimondo', 'Amine El Hattami'] | 2022-05-24 | null | null | null | null | ['workflow-discovery'] | ['natural-language-processing'] | [ 5.30750096e-01 1.57315552e-01 1.61793485e-01 -5.32968879e-01
-8.14179778e-01 -9.69019532e-01 1.09940958e+00 -5.63838519e-02
-5.05611956e-01 9.41545248e-01 6.28284335e-01 -3.97455245e-01
-2.89810926e-01 -4.65717673e-01 -2.90110558e-01 -4.00775999e-01
4.65502329e-02 9.39827800e-01 5.32722354e-01 -4.36737001e-01
4.09929663e-01 4.04402822e-01 -1.27136707e+00 7.18352973e-01
6.37368381e-01 3.96190375e-01 1.64609194e-01 1.07105100e+00
-5.09874046e-01 1.21012533e+00 -9.73961890e-01 -8.27875733e-02
1.62145287e-01 -7.16512382e-01 -1.68645096e+00 3.06547821e-01
-7.92557374e-02 -3.00480366e-01 -1.79330885e-01 5.60467362e-01
4.40141827e-01 4.92248267e-01 4.94614780e-01 -1.39108276e+00
4.18870039e-02 4.84448969e-01 -1.47870183e-01 2.76080877e-01
9.84409869e-01 8.25012207e-01 8.61733496e-01 -2.15435117e-01
1.11319542e+00 1.42826545e+00 4.06877905e-01 1.08000433e+00
-1.46395588e+00 -8.24942067e-02 2.84237802e-01 3.77639174e-01
-6.54604554e-01 -7.04801500e-01 2.38928795e-01 -5.19811392e-01
1.42650664e+00 3.85736495e-01 2.77131259e-01 1.33661211e+00
-2.25435987e-01 9.39511895e-01 1.10543394e+00 -5.20599723e-01
3.65169317e-01 -1.58592626e-01 2.02038735e-01 7.91722000e-01
-4.95581210e-01 -2.07574263e-01 -8.26990545e-01 -4.34870839e-01
2.58455664e-01 -2.32893348e-01 -4.11653787e-01 3.98150608e-02
-1.24591315e+00 7.19767153e-01 -3.16995740e-01 4.41347748e-01
-3.44027907e-01 -1.74221337e-01 5.01958013e-01 4.48378742e-01
1.65734336e-01 9.26071525e-01 -7.47445643e-01 -9.26641405e-01
-4.40025955e-01 6.33917987e-01 1.64354432e+00 8.10421228e-01
6.69584274e-01 -3.79454702e-01 -6.88126206e-01 6.63126171e-01
-2.19614848e-01 -6.57732338e-02 3.82142931e-01 -1.50439239e+00
4.17896837e-01 8.40134203e-01 7.32567608e-01 -2.06481397e-01
-5.91718376e-01 5.22896111e-01 -2.12635651e-01 1.69143975e-02
8.93502891e-01 -4.03641731e-01 -7.15372264e-01 1.70625222e+00
3.81535918e-01 3.00099760e-01 3.01509589e-01 7.60190606e-01
5.33717275e-01 6.03822708e-01 4.50186543e-02 -5.01041055e-01
1.14224875e+00 -9.51618969e-01 -6.95155680e-01 -2.57134855e-01
9.43290710e-01 -4.35989082e-01 1.28953898e+00 4.17512119e-01
-9.06481028e-01 -3.04557700e-02 -4.38486189e-01 2.46167570e-01
-1.50387779e-01 -4.25882638e-01 5.63848197e-01 3.63947421e-01
-9.94658649e-01 6.87637150e-01 -7.78903604e-01 -8.28528225e-01
1.95454478e-01 1.62549585e-01 -2.11616009e-01 -1.74763694e-01
-1.15120041e+00 1.16603756e+00 2.38101974e-01 -2.32588470e-01
-1.37790430e+00 -5.96387446e-01 -6.73971236e-01 2.49460638e-02
1.25374997e+00 -4.37135637e-01 2.07996774e+00 -6.94784760e-01
-1.80625737e+00 7.02656150e-01 -3.00545752e-01 -4.82007891e-01
5.92827082e-01 -5.37052564e-02 -1.68419152e-01 9.06132311e-02
4.23709601e-02 3.62249911e-01 3.50362986e-01 -9.12327647e-01
-1.01561940e+00 -1.82970598e-01 7.39218950e-01 1.35669142e-01
-7.77294189e-02 4.30792332e-01 -2.49573141e-01 1.17858015e-01
-5.05783260e-01 -1.04520190e+00 -2.32564583e-01 -1.91070393e-01
-4.59749371e-01 -7.08743274e-01 5.95799208e-01 -4.88008320e-01
1.07300138e+00 -1.70718980e+00 4.38118279e-01 -2.01494187e-01
6.14085756e-02 6.07333779e-01 -4.02197987e-01 9.98741269e-01
4.10101593e-01 1.85984150e-01 -5.45563400e-01 -1.09685048e-01
2.39447653e-01 7.65277147e-01 -1.39896214e-01 -5.51882759e-02
5.08980870e-01 7.42999554e-01 -1.08345783e+00 -3.32199365e-01
-6.87440159e-03 -1.33846432e-01 -4.26243514e-01 7.93102443e-01
-9.51793432e-01 6.79156542e-01 -5.15448034e-01 2.60682821e-01
1.13479726e-01 -3.86059731e-01 6.20429397e-01 2.92040557e-01
-1.59592077e-01 5.68059087e-01 -1.11641979e+00 1.92621040e+00
-5.31389892e-01 4.02452111e-01 1.01607218e-01 -8.55488181e-01
4.57733154e-01 4.39625919e-01 4.71559852e-01 -4.83573347e-01
-3.48031491e-01 -2.02782378e-01 2.08892822e-01 -8.60438108e-01
3.05334300e-01 -2.12405741e-01 -2.01733291e-01 9.72364426e-01
2.46928558e-01 9.45559219e-02 7.13190734e-01 4.63387758e-01
1.88830698e+00 9.83854979e-02 5.10474503e-01 9.39598233e-02
5.15683353e-01 6.70165718e-01 5.09789407e-01 9.69453990e-01
-1.77408457e-01 3.77260707e-02 1.11771619e+00 -5.60769320e-01
-7.28665233e-01 -4.89741147e-01 4.86918986e-01 1.48277271e+00
-2.02776462e-01 -6.13685191e-01 -7.30698228e-01 -1.07992983e+00
-1.05578534e-01 1.11563289e+00 -5.37367940e-01 5.54689765e-02
-6.95618927e-01 -5.06017387e-01 8.06791782e-01 2.81504869e-01
4.39629972e-01 -1.32509732e+00 -9.67888534e-01 4.89422441e-01
-4.27424461e-01 -1.19297040e+00 -5.29699504e-01 1.65919289e-01
-3.28836620e-01 -1.64345002e+00 -2.79683501e-01 -1.46950752e-01
1.49643049e-01 9.87597331e-02 1.26762342e+00 1.06402345e-01
-2.88295329e-01 8.23361874e-01 -5.12915671e-01 -2.50699610e-01
-9.98481750e-01 4.61594053e-02 -8.33399817e-02 -1.68841109e-01
5.21280408e-01 -2.57610291e-01 -1.00083863e-02 4.67971534e-01
-8.32896292e-01 1.08584478e-01 5.42933233e-02 7.84465969e-01
-8.97964910e-02 -1.25554472e-01 5.68816304e-01 -1.20764005e+00
1.26604748e+00 -3.62730592e-01 -5.66738665e-01 5.16878486e-01
-5.53214431e-01 3.78854096e-01 6.83821440e-01 -3.24999899e-01
-1.38138664e+00 7.22794831e-02 2.37148665e-02 -2.04107314e-01
-6.96949184e-01 5.16846776e-01 -5.33569828e-02 3.01664770e-01
7.47256696e-01 2.01394483e-01 1.52251735e-01 -5.65082073e-01
3.71261477e-01 6.83122575e-01 3.42597961e-01 -9.41244423e-01
3.06715995e-01 2.06729025e-01 -1.64967164e-01 -7.76892424e-01
-8.38119268e-01 -7.32544124e-01 -5.90822995e-01 -2.46746019e-01
8.04028153e-01 -2.98962563e-01 -1.36355662e+00 2.42870122e-01
-1.55596876e+00 -9.29567575e-01 -2.09670752e-01 -1.80710658e-01
-8.05554390e-01 2.73312211e-01 -4.91624385e-01 -1.16172004e+00
-1.10768273e-01 -1.02558434e+00 8.71597946e-01 1.41541079e-01
-6.46528006e-01 -9.49435711e-01 3.11062068e-01 2.05017507e-01
2.67727762e-01 1.18078336e-01 1.00088513e+00 -1.33039296e+00
-3.71714681e-01 3.11414659e-01 1.68148249e-01 1.84625998e-01
3.06083173e-01 -2.33141810e-01 -8.56075168e-01 -4.82240841e-02
-1.23918705e-01 -5.44911087e-01 4.23264593e-01 -1.35888129e-01
8.50964665e-01 -7.69093394e-01 -2.45866790e-01 -1.36699989e-01
9.42008138e-01 5.01075745e-01 5.88284850e-01 2.48172387e-01
2.22363263e-01 9.38581347e-01 6.88928664e-01 6.37612343e-01
3.24516863e-01 8.84640932e-01 1.35004986e-03 3.34005713e-01
2.14681059e-01 -2.97851190e-02 5.06009996e-01 6.14149347e-02
-1.30295947e-01 -5.40406346e-01 -1.30902898e+00 5.41094601e-01
-2.29810429e+00 -1.27437258e+00 2.19288431e-02 1.86430860e+00
1.06898606e+00 1.43850759e-01 5.48035622e-01 -1.22063577e-01
4.75279897e-01 4.26883101e-02 -6.63372815e-01 -5.65350115e-01
3.67673367e-01 1.58829734e-01 1.55935558e-02 6.87562585e-01
-7.68944144e-01 9.14386809e-01 6.05014372e+00 4.11862314e-01
-6.32792652e-01 1.64865375e-01 1.80185914e-01 -7.52537549e-02
1.55906215e-01 1.17756715e-02 -8.53621662e-01 2.47115761e-01
1.22756338e+00 -3.73363674e-01 8.22128594e-01 5.08578360e-01
4.55328017e-01 -4.99490440e-01 -1.60486948e+00 3.60536635e-01
-1.40892595e-01 -1.35401881e+00 -1.68408439e-01 -1.17146477e-01
4.66719121e-01 -1.45025939e-01 -6.75679684e-01 4.83362496e-01
1.08064413e+00 -9.87047493e-01 1.72522098e-01 6.11829340e-01
4.52914953e-01 -4.62079316e-01 5.42172670e-01 9.63428319e-01
-6.54976785e-01 -2.08868876e-01 1.45738572e-02 -2.38217011e-01
2.67965674e-01 -9.54073519e-02 -1.70378792e+00 6.26216233e-01
2.81687766e-01 5.69638371e-01 -1.45913035e-01 7.86543310e-01
-3.73997331e-01 5.83500206e-01 -1.75200030e-01 -2.73295224e-01
5.79532385e-01 6.13973066e-02 6.99554145e-01 1.34596074e+00
-7.67608434e-02 5.67190766e-01 5.02573431e-01 8.54846299e-01
6.61970153e-02 -3.92750233e-01 -6.09436095e-01 -3.60729724e-01
4.17042106e-01 1.11125469e+00 -4.76156652e-01 -5.62890947e-01
-3.87440264e-01 1.11754501e+00 2.60838628e-01 3.88698816e-01
-4.85215038e-01 -1.20231658e-01 1.00082517e+00 -9.19354409e-02
2.46829972e-01 -1.60174087e-01 3.39120328e-01 -9.69090879e-01
-2.20143601e-01 -1.41945744e+00 6.78673267e-01 -6.76652074e-01
-1.34859216e+00 4.90371853e-01 2.12340713e-01 -7.87912905e-01
-7.89133608e-01 -4.85414058e-01 -6.51877165e-01 9.10377324e-01
-1.35553646e+00 -6.94918156e-01 -3.25523376e-01 6.68885887e-01
1.00960076e+00 -1.34403616e-01 1.23749089e+00 -1.31225586e-01
-4.82673943e-01 -8.05175751e-02 -2.78219581e-01 -2.74659209e-02
7.56092846e-01 -1.41525960e+00 3.33593667e-01 7.08425224e-01
1.48995399e-01 5.76823175e-01 8.80746067e-01 -6.58103406e-01
-1.55772448e+00 -1.03200746e+00 8.27416658e-01 -6.72697604e-01
6.39257133e-01 -3.11517388e-01 -1.07737410e+00 8.78349245e-01
4.10904855e-01 -4.29594427e-01 5.78638852e-01 2.35278085e-01
-1.02265790e-01 2.61051089e-01 -1.16735220e+00 5.20153880e-01
1.34841537e+00 -4.64411467e-01 -9.03630972e-01 6.90979779e-01
7.11305976e-01 -4.26795542e-01 -5.24593651e-01 -1.24894902e-01
2.24396750e-01 -8.94717693e-01 5.59363365e-01 -1.63248265e+00
4.73828584e-01 -1.42654911e-01 -3.90956253e-02 -1.61612558e+00
3.69983055e-02 -1.24796081e+00 -1.68844849e-01 1.16087210e+00
3.16051334e-01 -4.01818812e-01 4.77443695e-01 1.12317896e+00
-1.28221020e-01 -4.82503206e-01 -7.82425702e-01 -8.95429373e-01
-3.86835128e-01 -2.90509164e-01 7.08286822e-01 8.31946135e-01
3.71994078e-01 6.00980937e-01 -4.14882690e-01 -2.81625539e-02
1.99309304e-01 1.37350619e-01 1.05462015e+00 -1.17044687e+00
-4.02045131e-01 -2.21108809e-01 8.43416154e-02 -1.07983148e+00
4.15034771e-01 -7.81530440e-01 4.28427428e-01 -1.82507408e+00
1.63931549e-01 -9.77566168e-02 6.18319586e-03 9.41021144e-01
1.20619964e-02 -5.83545744e-01 3.25087726e-01 1.44936308e-01
-1.02940571e+00 1.98208198e-01 1.16900814e+00 -1.49036095e-01
-4.14388418e-01 1.03045397e-01 -5.00815928e-01 6.30726039e-01
5.46975255e-01 -5.44975340e-01 -3.74628514e-01 -2.04700053e-01
-1.19092487e-01 5.87629080e-01 2.74163067e-01 -6.05038285e-01
5.66261470e-01 -7.99838781e-01 -4.53380287e-01 2.42813416e-02
2.68850178e-01 -5.56604028e-01 3.95656787e-02 4.71812725e-01
-8.35059822e-01 -1.11467175e-01 1.59501731e-01 5.94870746e-01
7.37925172e-02 -5.23423910e-01 4.58600849e-01 -4.83768731e-01
-1.08607042e+00 -6.84019625e-02 -9.11523461e-01 4.43334073e-01
1.12649846e+00 1.10118769e-01 -5.52428007e-01 -6.42125130e-01
-7.15010345e-01 5.81450701e-01 2.44631276e-01 3.82464200e-01
5.02227485e-01 -6.81212306e-01 -7.66362488e-01 -1.84489593e-01
2.09486410e-01 -2.00126454e-01 -1.28174171e-01 8.07527184e-01
-1.30011931e-01 5.27659595e-01 -2.21418932e-01 -4.21907187e-01
-1.55873346e+00 4.00904447e-01 4.24418479e-01 -4.75906223e-01
-5.54854214e-01 6.74360156e-01 -8.03334117e-02 -6.73210740e-01
3.29534233e-01 -1.96723416e-01 -2.54338443e-01 6.40863702e-02
8.08032393e-01 4.08412069e-01 1.65865198e-02 -1.43169805e-01
-5.58705747e-01 -2.07216129e-01 -1.80742636e-01 -2.84968048e-01
1.61730707e+00 8.81783366e-02 -7.46777793e-03 4.91377741e-01
6.34289503e-01 -4.06679422e-01 -1.38180244e+00 -4.13981557e-01
5.97449243e-01 -5.91924787e-01 -5.14620066e-01 -1.48608577e+00
-3.90211284e-01 7.85117507e-01 1.47637531e-01 5.99195361e-01
9.27656829e-01 1.28948599e-01 5.59824228e-01 9.45778430e-01
4.73922580e-01 -1.02732217e+00 3.33525538e-01 8.92411888e-01
8.96430969e-01 -1.12068975e+00 -3.83023977e-01 -9.48665142e-02
-1.00745237e+00 1.18681777e+00 7.17694283e-01 5.35732508e-01
-1.59610137e-01 5.23790598e-01 2.11933418e-03 -2.63879538e-01
-1.48071003e+00 -4.25681651e-01 -2.42630109e-01 7.10069776e-01
1.56950906e-01 -1.03482775e-01 -1.38620511e-01 5.42424977e-01
4.56519783e-01 2.54327536e-01 8.57307196e-01 1.31690776e+00
-5.34501255e-01 -1.37939930e+00 -1.68800846e-01 5.04611552e-01
-9.04910639e-02 1.67351708e-01 -9.72367167e-01 6.19598091e-01
-2.49319568e-01 1.23091424e+00 -1.72133669e-01 -2.93019891e-01
7.14054465e-01 7.22820759e-01 5.17057776e-01 -9.81477559e-01
-8.79596531e-01 -2.61776477e-01 9.80826914e-01 -8.03759575e-01
-6.45737052e-01 -6.84366167e-01 -1.32549167e+00 -4.49085176e-01
-7.24982843e-02 2.75728881e-01 2.81982452e-01 1.27206957e+00
3.22906464e-01 5.83128154e-01 3.95177752e-01 -4.75586921e-01
-8.50656509e-01 -1.10645592e+00 -1.31869227e-01 5.57295620e-01
7.62334839e-02 -6.03861272e-01 -6.59011602e-02 2.80401498e-01] | [12.870532989501953, 7.934365272521973] |
58cfa6d2-882d-4b6d-89f1-3c623992ecbf | interactive-molecular-discovery-with-natural | 2306.11976 | null | https://arxiv.org/abs/2306.11976v1 | https://arxiv.org/pdf/2306.11976v1.pdf | Interactive Molecular Discovery with Natural Language | Natural language is expected to be a key medium for various human-machine interactions in the era of large language models. When it comes to the biochemistry field, a series of tasks around molecules (e.g., property prediction, molecule mining, etc.) are of great significance while having a high technical threshold. Bridging the molecule expressions in natural language and chemical language can not only hugely improve the interpretability and reduce the operation difficulty of these tasks, but also fuse the chemical knowledge scattered in complementary materials for a deeper comprehension of molecules. Based on these benefits, we propose the conversational molecular design, a novel task adopting natural language for describing and editing target molecules. To better accomplish this task, we design ChatMol, a knowledgeable and versatile generative pre-trained model, enhanced by injecting experimental property information, molecular spatial knowledge, and the associations between natural and chemical languages into it. Several typical solutions including large language models (e.g., ChatGPT) are evaluated, proving the challenge of conversational molecular design and the effectiveness of our knowledge enhancement method. Case observations and analysis are conducted to provide directions for further exploration of natural-language interaction in molecular discovery. | ['Zhiyuan Liu', 'Guotong Xie', 'Maosong Sun', 'Xingzhi Sun', 'Haishen Yao', 'Cheng Yang', 'Jiarui Liu', 'Shipeng Wang', 'Bangchen Yin', 'Zheni Zeng'] | 2023-06-21 | null | null | null | null | ['property-prediction'] | ['medical'] | [ 3.40377957e-01 3.31688374e-02 -1.86607987e-01 -1.46070078e-01
-5.35107553e-01 -8.17708790e-01 6.73632085e-01 4.18680757e-01
-1.49824902e-01 1.35062313e+00 3.38506430e-01 -6.62350178e-01
-8.64110980e-03 -9.99965072e-01 -8.71465266e-01 -9.96223629e-01
-5.32451831e-02 2.95188040e-01 9.34802443e-02 -3.95705372e-01
4.12814349e-01 5.99974871e-01 -1.08450973e+00 5.53988159e-01
1.25293303e+00 4.41381693e-01 4.92193013e-01 3.26729953e-01
-5.70870161e-01 8.42293322e-01 -4.70066667e-01 -5.36412597e-01
-4.55910563e-01 -6.22236609e-01 -9.86310899e-01 -5.42777061e-01
-3.30356836e-01 3.34310800e-01 -4.09835465e-02 7.93484986e-01
6.77256882e-01 1.40297651e-01 5.69390297e-01 -7.24609435e-01
-8.19692731e-01 6.54704869e-01 -3.36761028e-01 -2.07226217e-01
6.92450523e-01 4.22469676e-01 7.72048116e-01 -7.86414027e-01
8.31181526e-01 1.35838509e+00 3.20739567e-01 6.52543902e-01
-1.03495967e+00 -6.36135042e-01 1.53621867e-01 3.83872449e-01
-1.31769335e+00 -2.98728496e-01 5.85196555e-01 -4.96983677e-01
1.10420489e+00 2.67947435e-01 5.47349870e-01 1.08804524e+00
5.10153770e-01 3.62731427e-01 7.58686125e-01 -2.74483681e-01
3.24872613e-01 2.52782315e-01 -6.08084500e-02 8.32687259e-01
1.99207943e-03 -2.03120515e-01 -7.89209843e-01 -2.93935269e-01
3.76506567e-01 1.34369373e-01 -3.25157225e-01 -1.13414116e-02
-1.26991773e+00 8.29331219e-01 4.69751954e-01 3.68430644e-01
-3.64141375e-01 -1.01037115e-01 3.33406717e-01 1.02019764e-01
4.28614974e-01 8.87848258e-01 -5.99948287e-01 2.98028849e-02
-1.42292574e-01 7.09765106e-02 9.14798915e-01 9.21006382e-01
7.28649855e-01 -2.85368294e-01 -1.70011550e-01 5.71496665e-01
2.66134143e-01 2.25767702e-01 2.58013725e-01 -2.96240747e-01
2.55364597e-01 7.97496498e-01 -6.84189573e-02 -8.05636585e-01
-4.28025544e-01 -2.53141314e-01 -9.46460783e-01 -3.61597389e-01
1.23363413e-01 4.19633090e-02 -7.56696403e-01 1.89655364e+00
6.21538103e-01 -3.42515670e-02 4.64306861e-01 5.66127121e-01
1.07544017e+00 1.10317659e+00 8.35117221e-01 -4.61458921e-01
1.66289949e+00 -6.82368159e-01 -6.72308266e-01 1.46727070e-01
8.84407759e-01 -7.44954705e-01 8.64298463e-01 1.74044237e-01
-6.70548260e-01 -4.74999994e-01 -8.41065288e-01 -2.45030299e-01
-1.03789198e+00 -1.95555076e-01 1.32960272e+00 5.24602890e-01
-5.07682443e-01 7.32729256e-01 -4.19826031e-01 -1.95530757e-01
5.94213009e-01 5.80537736e-01 -6.50245726e-01 -1.76091865e-01
-1.50610352e+00 9.76050496e-01 6.80211246e-01 2.64769524e-01
-7.58134425e-01 -9.48422968e-01 -6.26414895e-01 1.70653854e-02
4.65468794e-01 -9.41264808e-01 7.17950761e-01 -3.39914918e-01
-1.50717056e+00 5.01579404e-01 -3.07490557e-01 -1.59523338e-01
2.65108179e-02 1.18600324e-01 -3.31171691e-01 -1.51495114e-01
7.65056908e-02 5.60087144e-01 6.90134987e-02 -9.60227132e-01
-2.19463289e-01 -4.74930257e-01 2.43349671e-01 1.24330357e-01
-1.91284657e-01 -1.24647699e-01 -1.70336679e-01 -3.61641705e-01
-3.79444033e-01 -6.80097938e-01 -6.04665399e-01 -1.89121336e-01
-4.96960223e-01 -3.80064636e-01 5.01160026e-01 -4.82453495e-01
1.16330957e+00 -1.68309534e+00 5.75941443e-01 2.38746002e-01
4.82546955e-01 2.95352757e-01 -2.16628984e-01 9.65752482e-01
-2.33636081e-01 5.25000334e-01 -2.63321489e-01 2.89221376e-01
-2.90375501e-01 1.78462751e-02 -2.41826102e-01 8.43228921e-02
3.06382835e-01 1.23114550e+00 -1.07582641e+00 -3.57836485e-01
4.30069305e-02 5.91479897e-01 -6.46527350e-01 3.38046342e-01
-8.29289138e-01 8.70926023e-01 -9.33949471e-01 6.41975224e-01
4.17818099e-01 -3.93385053e-01 3.22930425e-01 -4.51239318e-01
-3.01450878e-01 2.49206305e-01 -6.82458937e-01 1.86908877e+00
-5.73958635e-01 1.34074077e-01 -4.71862525e-01 -7.58236825e-01
7.68818915e-01 4.10384446e-01 3.67457658e-01 -5.75756133e-01
-1.69125587e-01 1.22672319e-01 2.90459216e-01 -7.30889738e-01
1.02509864e-01 -4.04115945e-01 1.22543417e-01 2.38716915e-01
-8.35383758e-02 -1.20082565e-01 1.39590427e-01 3.85340452e-01
9.24354851e-01 3.22196275e-01 6.01835966e-01 -1.56685323e-01
8.03163528e-01 5.33519909e-02 2.57040471e-01 3.59919220e-01
3.48457366e-01 -1.60531193e-01 5.57238519e-01 -5.12456000e-01
-9.06411111e-01 -6.31576657e-01 5.56100756e-02 1.04850638e+00
6.02552854e-02 -7.70168424e-01 -7.39345193e-01 -4.70454991e-01
-3.77764642e-01 4.66863036e-01 -5.75491190e-01 -3.85459721e-01
-3.12166125e-01 -1.28698468e+00 5.29463410e-01 1.46434799e-01
4.09515321e-01 -1.13258600e+00 2.40704924e-01 4.01270956e-01
-1.25623615e-02 -8.74561489e-01 -3.56621444e-01 2.44938910e-01
-4.85070199e-01 -1.00991750e+00 -5.09999752e-01 -8.00838649e-01
5.38653731e-01 1.70842469e-01 9.20855880e-01 -1.29812453e-02
-5.13278365e-01 -1.83633789e-01 -2.75308162e-01 -5.34808040e-01
-7.19803751e-01 1.71784952e-01 -4.09234576e-02 -1.15171880e-01
2.59738266e-01 -8.14038694e-01 -5.58560669e-01 4.31686565e-02
-1.07205820e+00 4.68215376e-01 7.93199539e-01 8.76944363e-01
5.34195244e-01 -1.84592113e-01 8.76370490e-01 -1.04827738e+00
7.83485353e-01 -6.12539113e-01 -3.32479268e-01 5.41286290e-01
-2.63009459e-01 4.63012606e-01 8.69836271e-01 -5.81851721e-01
-1.10158038e+00 1.35662332e-02 -3.79262269e-01 3.55034530e-01
-5.41489869e-02 9.63470697e-01 -8.04639518e-01 -2.23047331e-01
5.61775923e-01 4.96652305e-01 -1.30790859e-01 -3.99111509e-01
7.16502607e-01 5.60439169e-01 -1.88120957e-02 -8.50304961e-01
4.66710836e-01 1.18487336e-01 3.67130280e-01 -9.32156920e-01
-6.00874841e-01 -2.98409522e-01 -6.05667770e-01 2.34126717e-01
1.02202022e+00 -7.72494256e-01 -1.33782601e+00 1.50671408e-01
-1.59998214e+00 2.14101765e-02 6.48072809e-02 3.05313587e-01
-3.13090354e-01 5.76379836e-01 -5.52907228e-01 -6.87810957e-01
-4.30717647e-01 -1.27204442e+00 1.03526986e+00 2.70210534e-01
-9.42838490e-02 -1.14932120e+00 1.68735445e-01 3.85697693e-01
1.81406900e-01 3.54096800e-01 1.52946615e+00 -6.73792362e-01
-9.93416011e-01 8.15728903e-02 -2.16331810e-01 -1.08147234e-01
5.00604987e-01 -2.43364140e-01 -9.32870567e-01 2.52211243e-01
-3.98683846e-01 -3.02339464e-01 7.08705604e-01 -1.17282122e-01
1.24157238e+00 -4.54652667e-01 -5.91924846e-01 3.86236131e-01
1.17426491e+00 6.67277873e-01 9.44240153e-01 -2.27129906e-02
1.07909334e+00 8.74993503e-01 3.51149797e-01 2.86820292e-01
1.24709137e-01 6.45516992e-01 1.36034131e-01 -3.30133706e-01
2.06013367e-01 -5.83657086e-01 2.76021183e-01 8.84416461e-01
-4.42312717e-01 -4.49823976e-01 -8.19893181e-01 -5.17979860e-02
-1.79961526e+00 -1.16655004e+00 -6.13914281e-02 1.90300047e+00
1.36625886e+00 -2.81263173e-01 -2.88329810e-01 -4.40806419e-01
4.21774983e-01 -5.28507791e-02 -6.84026122e-01 -2.71197528e-01
-2.71128207e-01 3.05412859e-01 1.41794160e-02 5.90755224e-01
-6.56516910e-01 1.32184124e+00 5.53772688e+00 1.30219400e+00
-1.02915990e+00 -1.37017176e-01 7.82745004e-01 5.85025012e-01
-5.59937418e-01 7.35712424e-02 -8.38483393e-01 3.50424707e-01
9.84964430e-01 -3.48820597e-01 3.95554751e-01 5.61500549e-01
4.52279657e-01 6.89486265e-02 -1.34720063e+00 1.00464404e+00
-2.59663194e-01 -2.07146955e+00 7.89765298e-01 5.65971099e-02
4.82768476e-01 -4.33978468e-01 -2.08514094e-01 1.48268208e-01
3.53203937e-02 -1.44908690e+00 6.44720048e-02 7.23992765e-01
6.70719445e-01 -7.24330485e-01 6.31059587e-01 4.46067035e-01
-1.22544754e+00 4.05063152e-01 -3.37837875e-01 8.18398148e-02
1.13237396e-01 5.26934862e-01 -8.95771682e-01 9.16765332e-01
6.75188098e-03 7.42309093e-01 -2.25223079e-01 6.78518236e-01
-9.01198536e-02 2.24962994e-01 -1.04951508e-01 -5.08850038e-01
4.20816392e-02 -4.75946903e-01 8.60595554e-02 1.26968253e+00
4.65667136e-02 5.97139478e-01 2.53276736e-01 1.04380107e+00
-3.51552486e-01 5.47082365e-01 -7.97294259e-01 -7.45537877e-01
2.38918468e-01 1.06479371e+00 -4.82847720e-01 -2.90929526e-01
-3.26694191e-01 9.09468770e-01 1.06729507e-01 3.45331132e-01
-8.28685105e-01 -3.42901796e-01 7.27396369e-01 3.68374661e-02
-3.30316812e-01 -2.87609190e-01 2.54143160e-02 -1.06222832e+00
-2.10945368e-01 -9.85015571e-01 -8.66321996e-02 -7.55189598e-01
-1.17288470e+00 4.26516116e-01 -2.90262699e-01 -9.30970430e-01
1.85469300e-01 -7.40314364e-01 -4.88105655e-01 1.07980454e+00
-1.34011984e+00 -1.20279241e+00 3.33582088e-02 3.50028813e-01
4.92963463e-01 -1.08583093e-01 1.24200499e+00 4.34324712e-01
-5.86439073e-01 2.30541065e-01 4.54243273e-03 -3.59671116e-01
6.50471807e-01 -9.44937944e-01 3.70789409e-01 1.26243949e-01
-1.39977932e-01 1.28904104e+00 5.35508037e-01 -7.26447165e-01
-1.86796188e+00 -1.16032934e+00 9.28775549e-01 -6.97470427e-01
6.46358728e-01 -7.06318080e-01 -1.05941904e+00 1.08681686e-01
7.61176925e-03 -5.69219589e-01 1.28395593e+00 1.28303945e-01
-2.59652972e-01 1.65940866e-01 -8.31801474e-01 9.14866209e-01
1.17343438e+00 -6.72683179e-01 -1.08549006e-01 9.24064040e-01
1.36964488e+00 -2.54561990e-01 -1.10958731e+00 2.17665628e-01
4.73130941e-01 -5.00138819e-01 1.18772197e+00 -1.23740149e+00
5.02277792e-01 -4.79069501e-01 3.07098068e-02 -1.21597385e+00
-1.72754973e-01 -8.85471702e-01 1.04826167e-01 1.12290537e+00
9.42253411e-01 -5.21181822e-01 6.93543077e-01 4.07590479e-01
-8.41580853e-02 -8.84885669e-01 -6.95571899e-01 -3.58407795e-01
9.64364037e-02 -2.51076162e-01 6.29133999e-01 1.01696503e+00
3.43331695e-01 1.08045924e+00 -4.99396682e-01 -4.09674942e-02
1.26215532e-01 1.37936950e-01 7.68046975e-01 -1.14272857e+00
-4.30378944e-01 -3.09290528e-01 -2.73299605e-01 -1.19814575e+00
4.72487099e-02 -1.02644539e+00 -3.27179193e-01 -1.61626351e+00
5.72386682e-01 -2.54438072e-01 -2.16507353e-02 4.12417918e-01
-2.83722192e-01 -2.96578914e-01 -1.13921449e-01 6.86929747e-02
-6.36900246e-01 8.28157306e-01 1.77796102e+00 -4.59694266e-01
-4.55004811e-01 -1.84138790e-01 -1.06170535e+00 3.42418909e-01
5.70303142e-01 -2.84358233e-01 -5.07515788e-01 1.41197681e-01
5.17319381e-01 1.42122433e-01 1.19116820e-01 -4.79484051e-01
3.52275610e-01 -5.74622214e-01 2.39560232e-01 -1.72626495e-01
3.90689909e-01 -5.53685308e-01 3.79284203e-01 4.81513530e-01
-3.61081719e-01 -1.99603975e-01 2.94538409e-01 7.86378264e-01
-2.04144910e-01 1.24020711e-01 5.66140950e-01 -3.88617575e-01
-5.94752312e-01 5.94523191e-01 -2.97120482e-01 -3.56578380e-01
9.59835231e-01 -4.88633700e-02 -5.15447795e-01 -1.34537891e-01
-7.57031977e-01 2.29309782e-01 1.75003707e-01 3.08408260e-01
6.70000196e-01 -8.71508181e-01 -4.05561566e-01 9.43408906e-02
3.73618931e-01 -6.59168791e-03 2.57577389e-01 7.24550843e-01
-4.78211552e-01 8.16835582e-01 1.64651826e-01 -2.61501074e-01
-1.28016710e+00 7.66139805e-01 3.03693056e-01 -4.90368366e-01
-6.36960799e-03 6.01211905e-01 8.24680626e-01 -3.02732766e-01
6.40124828e-02 -1.82079062e-01 -3.30663562e-01 -2.26874407e-02
7.09692836e-01 -1.25852540e-01 -5.94719909e-02 -1.99912146e-01
-3.45378458e-01 5.90482354e-01 -3.41574103e-01 4.09445614e-01
1.38606966e+00 2.72820026e-01 -5.24606526e-01 1.73766419e-01
1.08443701e+00 1.12031572e-01 -5.32680631e-01 -5.39598130e-02
7.01146349e-02 1.16944857e-01 -4.83273864e-01 -1.06564534e+00
-3.72193605e-01 9.51489627e-01 2.89435178e-01 -1.47163257e-01
7.26262450e-01 1.87891796e-01 6.47314548e-01 9.05075848e-01
5.08549809e-01 -5.25738895e-01 4.50151525e-02 4.03223574e-01
8.76230001e-01 -1.22599423e+00 1.42984524e-01 -7.86412477e-01
-4.75962609e-01 1.17936850e+00 3.84759098e-01 6.95740581e-01
4.31225777e-01 8.89356285e-02 -2.91836202e-01 -3.97906601e-01
-8.11682642e-01 -2.65797436e-01 6.27693906e-02 6.92524493e-01
8.35351825e-01 -7.49725774e-02 -4.28604782e-01 6.93569899e-01
2.10047707e-01 8.83919299e-02 4.72016409e-02 7.82296062e-01
-5.19070208e-01 -1.56178761e+00 -4.30591889e-02 -6.60476312e-02
-2.59901404e-01 -5.68575442e-01 -9.40407097e-01 4.48509693e-01
4.29728508e-01 8.86458457e-01 -7.15492129e-01 -3.97023559e-01
2.45453581e-01 1.42421812e-01 4.64148343e-01 -7.20849812e-01
-5.06604016e-01 -1.00006662e-01 2.08394602e-01 -2.54554451e-01
-5.99958956e-01 9.32079703e-02 -1.48651254e+00 -5.30435860e-01
-4.46149945e-01 7.49717116e-01 6.38540566e-01 9.83058512e-01
7.17382491e-01 6.73581123e-01 4.85185951e-01 -4.88672823e-01
-1.58895217e-02 -7.26894140e-01 -2.18957186e-01 9.04893801e-02
-2.64633051e-03 -3.39036673e-01 1.98413625e-01 3.54537725e-01] | [5.029621601104736, 5.87459659576416] |
1b3eaba3-052c-4abb-8312-953551b4d135 | mist-multiple-instance-spatial-transformer | 1811.10725 | null | https://arxiv.org/abs/1811.10725v5 | https://arxiv.org/pdf/1811.10725v5.pdf | MIST: Multiple Instance Spatial Transformer Network | We propose a deep network that can be trained to tackle image reconstruction and classification problems that involve detection of multiple object instances, without any supervision regarding their whereabouts. The network learns to extract the most significant top-K patches, and feeds these patches to a task-specific network -- e.g., auto-encoder or classifier -- to solve a domain specific problem. The challenge in training such a network is the non-differentiable top-K selection process. To address this issue, we lift the training optimization problem by treating the result of top-K selection as a slack variable, resulting in a simple, yet effective, multi-stage training. Our method is able to learn to detect recurrent structures in the training dataset by learning to reconstruct images. It can also learn to localize structures when only knowledge on the occurrence of the object is provided, and in doing so it outperforms the state-of-the-art. | ['Simon Kornblith', 'Yuhe Jin', 'Andrea Tagliasacchi', 'Kwang Moo Yi', 'Baptiste Angles'] | 2018-11-26 | null | null | null | null | ['anomaly-detection-in-surveillance-videos', 'anomaly-detection-in-surveillance-videos'] | ['computer-vision', 'methodology'] | [ 5.94236791e-01 3.14750999e-01 -1.99975103e-01 -2.32864380e-01
-1.03813469e+00 -4.17711049e-01 2.23849922e-01 -4.02499624e-02
-4.55183834e-01 4.05379087e-01 -1.20147526e-01 -1.73072144e-01
-3.78901362e-02 -5.57128966e-01 -1.23603928e+00 -7.75268078e-01
4.20439430e-02 4.94017571e-01 1.57009155e-01 1.65865123e-01
3.04020882e-01 4.87429589e-01 -1.46415627e+00 6.00408673e-01
5.24259090e-01 1.32754254e+00 6.33915186e-01 6.33353174e-01
1.42605618e-01 1.04284894e+00 -4.89340216e-01 -1.22852787e-01
1.65313721e-01 -4.40981030e-01 -8.61570179e-01 4.70558196e-01
6.90712988e-01 -3.61073494e-01 -1.42898068e-01 8.66554320e-01
3.64517987e-01 -1.87489972e-01 7.19493568e-01 -7.35766113e-01
-5.53869367e-01 5.40587723e-01 -5.35835207e-01 3.08585972e-01
-4.97680046e-02 -4.47432958e-02 1.26658046e+00 -1.18065906e+00
7.92006910e-01 9.35559928e-01 5.63814282e-01 5.26291728e-01
-1.65688539e+00 -2.81189114e-01 3.57480317e-01 1.87615052e-01
-1.22704971e+00 -4.14745510e-01 9.00106370e-01 -6.28402650e-01
9.06965852e-01 -8.24526027e-02 6.12662852e-01 1.09448850e+00
-1.89763427e-01 1.09795678e+00 8.76133204e-01 -4.93228734e-01
2.50427783e-01 1.92359746e-01 6.38116598e-02 9.06959116e-01
2.34043840e-02 -1.52758792e-01 -6.15453064e-01 4.97154007e-03
8.59347403e-01 -2.73679197e-02 -2.29975626e-01 -3.31645757e-01
-1.10964251e+00 7.35244036e-01 6.91975236e-01 3.39952648e-01
-5.49392402e-01 1.85993686e-01 1.96776256e-01 3.86816084e-01
4.01587665e-01 7.19994366e-01 -9.17028010e-01 3.66319031e-01
-9.66214597e-01 2.40810469e-01 5.74049354e-01 5.21041572e-01
1.02946937e+00 -3.42272595e-02 -1.40649706e-01 6.30636036e-01
-9.47154388e-02 1.11489914e-01 3.22334230e-01 -7.88972557e-01
4.95865345e-01 6.05541050e-01 -8.49401057e-02 -6.74866617e-01
-3.91099542e-01 -7.67133772e-01 -6.29570186e-01 3.56905818e-01
3.78253400e-01 -1.69670507e-01 -1.06080747e+00 1.77478933e+00
3.02972615e-01 4.63116348e-01 5.16840704e-02 9.76747453e-01
5.48220515e-01 6.89229727e-01 -2.30196804e-01 -1.62710786e-01
1.27647650e+00 -1.00496411e+00 -4.54491258e-01 -5.76000571e-01
5.78326046e-01 -3.20308834e-01 1.10119295e+00 3.87824208e-01
-1.11263776e+00 -4.73451972e-01 -9.58954692e-01 -1.36719987e-01
-1.34239390e-01 6.78797901e-01 3.67481202e-01 -2.60764956e-01
-9.35729504e-01 5.64229012e-01 -6.63743436e-01 -4.97603640e-02
6.50710762e-01 4.64916825e-01 -5.56784332e-01 5.51913232e-02
-7.74351716e-01 9.15102541e-01 5.05614698e-01 -1.67329349e-02
-1.28521812e+00 -7.89023578e-01 -7.96746612e-01 2.20346138e-01
8.38776708e-01 -6.78230643e-01 1.21248150e+00 -1.63546646e+00
-1.30898046e+00 1.04991937e+00 -2.06820577e-01 -5.79092562e-01
3.36760670e-01 -6.70579299e-02 4.90411222e-02 3.89183968e-01
2.08184212e-01 4.92649168e-01 1.41221428e+00 -1.29268277e+00
-7.12562382e-01 -4.33049887e-01 7.26770461e-02 1.86099216e-01
-2.13075563e-01 -8.99772123e-02 -6.17070496e-01 -7.31764078e-01
1.82026580e-01 -9.39208686e-01 -3.53766114e-01 1.41629338e-01
-5.70437372e-01 -1.64349928e-01 7.32609808e-01 -7.61301100e-01
7.95687735e-01 -2.37457204e+00 5.55776119e-01 7.31449947e-02
1.92728475e-01 2.87754953e-01 -3.09350133e-01 2.28774533e-01
-1.53823584e-01 -1.91161901e-01 -4.28078145e-01 -6.69838667e-01
-4.08052951e-01 1.96663126e-01 -3.44741017e-01 5.00049651e-01
7.95647681e-01 9.88757610e-01 -7.49816239e-01 -2.12064549e-01
-2.76216697e-02 4.78672087e-01 -6.12851083e-01 3.45382035e-01
-5.44837892e-01 5.33424675e-01 -4.83613104e-01 3.57355833e-01
3.69756401e-01 -7.43403673e-01 2.67538518e-01 -7.45185018e-02
3.48200351e-02 4.50887918e-01 -1.03293705e+00 1.77077603e+00
-5.87830961e-01 8.25062811e-01 1.61959365e-01 -1.63494158e+00
6.80365324e-01 1.84319034e-01 3.34172517e-01 -5.66635072e-01
-1.23094223e-01 2.22943783e-01 -7.28439689e-02 -6.94411039e-01
4.98131961e-02 -5.67976423e-02 1.04495086e-01 2.63168931e-01
4.39766943e-01 3.50101203e-01 -1.08347431e-01 1.34041756e-01
1.29771483e+00 -5.77133708e-02 1.49439171e-01 4.80647199e-02
4.75329041e-01 -1.31043628e-01 6.76256835e-01 7.87328005e-01
4.04088527e-01 6.43034756e-01 8.72307837e-01 -4.82959151e-01
-1.11294413e+00 -6.50122106e-01 -1.41169488e-01 1.11744559e+00
-5.17261997e-02 -1.92396700e-01 -7.14143813e-01 -9.84694481e-01
1.15424886e-01 4.15267378e-01 -9.13400650e-01 -1.55655548e-01
-8.60898137e-01 -3.23891938e-01 -5.35139777e-02 5.09090543e-01
2.37448230e-01 -1.34452939e+00 -7.02594995e-01 2.71997422e-01
-4.52721491e-02 -1.23161900e+00 -1.76777557e-01 6.34680867e-01
-9.36346471e-01 -1.11855817e+00 -7.54080534e-01 -1.15011907e+00
1.13144982e+00 -2.52219718e-02 1.15232432e+00 4.53960262e-02
-4.30883408e-01 1.25011906e-01 -1.65126875e-01 -2.00292677e-01
-1.25793666e-01 2.52727360e-01 -3.38265300e-01 4.03292775e-01
-1.62108049e-01 -5.54867566e-01 -5.82985878e-01 1.57216206e-01
-6.86533213e-01 6.28840104e-02 8.86230528e-01 1.16010964e+00
8.12954545e-01 2.30743542e-01 6.67728662e-01 -1.16021490e+00
2.52129287e-01 -1.28106803e-01 -7.37129331e-01 4.09477890e-01
-3.36807102e-01 3.01344663e-01 8.17589879e-01 -5.34867644e-01
-5.76013327e-01 5.41066945e-01 6.32524565e-02 -7.88297534e-01
-6.85867947e-03 7.05567539e-01 1.06572574e-02 -1.19376026e-01
5.86376369e-01 4.75705087e-01 -7.14787170e-02 -6.11888051e-01
1.98593587e-01 8.62199664e-02 6.49524570e-01 -2.85920173e-01
7.33948827e-01 4.04533178e-01 3.20321135e-02 -7.13862479e-01
-1.32033062e+00 -4.40403849e-01 -6.53702617e-01 -3.66050936e-02
7.79491603e-01 -1.09749854e+00 -4.51343954e-01 3.77528518e-01
-1.25507355e+00 -5.28136075e-01 -4.00469661e-01 2.24065408e-01
-6.58248782e-01 -1.54157385e-01 -5.62249541e-01 -7.58644521e-01
-9.07575041e-02 -1.07511389e+00 1.14363277e+00 -3.26831266e-02
3.05000663e-01 -7.44504094e-01 -2.31461212e-01 2.15364277e-01
7.41237327e-02 1.69369593e-01 8.75726104e-01 -5.75656891e-01
-9.86634254e-01 -2.47367531e-01 -1.32687762e-01 4.33950543e-01
-1.40796527e-01 -2.21488789e-01 -1.10001218e+00 -3.13852638e-01
1.91662595e-01 -5.86028755e-01 1.35793102e+00 5.26550055e-01
1.56979096e+00 -6.58263206e-01 -3.78561139e-01 6.13796592e-01
1.41628981e+00 -3.14508587e-01 3.98165911e-01 1.99822158e-01
5.46361864e-01 7.10731387e-01 4.89873171e-01 2.34128371e-01
3.53243887e-01 9.21003163e-01 6.90758765e-01 -2.77295917e-01
-1.42803490e-01 -3.10989827e-01 1.89275309e-01 3.14040691e-01
2.18396902e-01 9.15033668e-02 -6.95245028e-01 6.23222768e-01
-2.05882311e+00 -8.35449398e-01 2.68127114e-01 2.00529528e+00
8.70469987e-01 1.34276584e-01 1.37397259e-01 1.94605291e-01
5.34835100e-01 5.36151752e-02 -8.68593872e-01 1.95115522e-01
-1.11100100e-01 1.55396760e-01 3.34349751e-01 4.70892370e-01
-1.24263072e+00 1.09574294e+00 6.19284010e+00 6.92006767e-01
-1.49410474e+00 1.79444402e-02 6.81246161e-01 -1.76810436e-02
-8.16810951e-02 -1.26658333e-02 -8.15722048e-01 3.47353131e-01
6.26537740e-01 4.04266357e-01 6.01124108e-01 8.25226665e-01
4.88014519e-02 -1.17117062e-01 -1.33371222e+00 9.60285962e-01
2.10927516e-01 -1.57948923e+00 -2.31141105e-01 6.43617138e-02
6.51675105e-01 4.70441468e-02 4.16768312e-01 1.45947143e-01
-4.04774770e-02 -9.88273323e-01 8.46558928e-01 3.95133317e-01
6.63713098e-01 -5.14403880e-01 4.54372108e-01 7.02359319e-01
-8.32049429e-01 -3.97413433e-01 -3.47462475e-01 1.23871811e-01
-4.05192114e-02 8.33967745e-01 -9.12836492e-01 2.53731042e-01
6.24380589e-01 9.43132699e-01 -3.45009685e-01 9.75272059e-01
-5.46387792e-01 6.77180886e-01 -3.09635192e-01 2.64793128e-01
2.05444857e-01 6.63713962e-02 3.87526065e-01 1.04888356e+00
2.75341630e-01 5.43763489e-02 2.89721161e-01 1.03550065e+00
-2.89051831e-01 -1.19400956e-01 -4.88105178e-01 1.44162830e-02
1.33507341e-01 1.28906894e+00 -5.76197088e-01 -3.30732524e-01
-2.72150904e-01 1.19526958e+00 1.06958413e+00 4.27094877e-01
-4.44745272e-01 -1.04461409e-01 3.62403989e-01 2.71045595e-01
9.44058061e-01 -5.14657870e-02 -3.19585681e-01 -1.29738331e+00
3.58913630e-01 -9.05241787e-01 3.61995608e-01 -6.47199690e-01
-1.21116614e+00 6.38269186e-01 -4.04518902e-01 -9.59478617e-01
-2.04531148e-01 -8.80654156e-01 -5.39196253e-01 7.13584602e-01
-1.84394622e+00 -1.09736991e+00 1.15586504e-01 5.94954252e-01
3.81108046e-01 -1.46832928e-01 5.50098240e-01 2.29309663e-01
-6.58459961e-01 4.86142606e-01 -7.02608675e-02 2.44164720e-01
4.03675467e-01 -1.39103222e+00 1.02438375e-01 7.20284224e-01
3.51475924e-01 3.40454668e-01 5.19894540e-01 -2.93551326e-01
-1.36457682e+00 -1.14829624e+00 1.01788604e+00 -3.54836851e-01
5.37591696e-01 -6.71241403e-01 -9.43758070e-01 7.10898936e-01
-1.89349800e-01 6.14347517e-01 2.66054213e-01 2.59454906e-01
-4.98404562e-01 -1.80186659e-01 -9.50214982e-01 2.67281502e-01
8.90877366e-01 -8.11955452e-01 -4.29849416e-01 4.99024957e-01
6.09810710e-01 -4.89487708e-01 -3.75668466e-01 2.57612765e-01
2.36338884e-01 -8.10531080e-01 1.03204978e+00 -6.93184316e-01
7.14351296e-01 -2.89957970e-01 8.05719271e-02 -1.23582077e+00
-2.57730991e-01 -4.41650867e-01 -3.29369813e-01 8.60454082e-01
7.62037635e-01 -4.86763686e-01 8.17493677e-01 2.26478904e-01
-9.41955745e-02 -1.04306245e+00 -9.72687900e-01 -4.11376983e-01
-2.56078005e-01 -7.78831467e-02 5.79239018e-02 8.23227584e-01
-2.64171958e-01 5.62524438e-01 -6.34145319e-01 5.03178954e-01
5.15607476e-01 4.05587733e-01 6.15356505e-01 -1.20176959e+00
-6.68989420e-01 -2.97196895e-01 -3.63336354e-01 -1.21015882e+00
4.64610189e-01 -9.53109920e-01 1.58273503e-01 -1.37397230e+00
2.35711783e-01 -4.59313482e-01 -3.46896857e-01 8.36094201e-01
-3.09564173e-01 1.99626908e-01 1.87438414e-01 1.86976880e-01
-6.34338319e-01 3.74430269e-01 1.37067592e+00 -1.10036135e-01
-1.91880077e-01 2.06939533e-01 -7.25239694e-01 5.41076183e-01
5.23110330e-01 -6.95261180e-01 -2.62127876e-01 -4.06711072e-01
2.86470175e-01 1.72778532e-01 6.12179816e-01 -8.04393470e-01
2.79691398e-01 2.22111329e-01 5.16941547e-01 -5.50321281e-01
3.96387279e-01 -8.83828521e-01 -1.88717842e-01 3.07379484e-01
-8.11503232e-01 -2.75017947e-01 2.75231339e-02 5.83801568e-01
-3.43646795e-01 -3.64223450e-01 9.36022341e-01 -3.28375608e-01
-5.76335073e-01 3.80119234e-01 -2.49166414e-01 -5.18642515e-02
9.83700812e-01 2.38932092e-02 1.28503265e-02 -2.53792375e-01
-9.97597098e-01 2.00743362e-01 2.61334807e-01 6.52408749e-02
5.76176822e-01 -1.10529733e+00 -6.63897574e-01 4.28240836e-01
8.20657015e-02 4.18899059e-01 5.91404736e-02 6.81410670e-01
-5.32268360e-02 2.38977298e-01 1.10673144e-01 -8.17033887e-01
-1.07443631e+00 6.63501322e-01 6.48549259e-01 -4.76925999e-01
-9.51238811e-01 1.00155175e+00 3.49951327e-01 -2.49735862e-01
4.37034130e-01 -2.67709553e-01 -1.47855222e-01 9.65210423e-02
6.20913684e-01 -2.73261577e-01 2.10232690e-01 -4.04852659e-01
-3.26405406e-01 6.76823437e-01 -3.25777501e-01 -1.10332072e-02
1.69848537e+00 1.54482439e-01 -1.31545514e-01 4.78435010e-01
1.32129061e+00 -4.18504298e-01 -1.66887844e+00 -6.22357845e-01
6.14362992e-02 -3.24711233e-01 2.21507624e-01 -8.89763594e-01
-1.27932250e+00 8.37212384e-01 2.95156360e-01 9.57767665e-02
1.17856252e+00 2.55397022e-01 3.52726251e-01 6.23668253e-01
6.15015849e-02 -9.45462048e-01 5.76901436e-01 6.27281129e-01
9.85347986e-01 -1.39594877e+00 -3.49724352e-01 -2.64991879e-01
-6.13061011e-01 1.09166181e+00 5.97088993e-01 -4.45769906e-01
7.45530009e-01 3.47629458e-01 -3.68191563e-02 -2.81912863e-01
-1.04471993e+00 -2.29925334e-01 5.56553006e-01 3.83878201e-01
7.81092197e-02 -1.85995191e-01 2.45247811e-01 2.81996638e-01
1.23591587e-01 4.24409099e-02 7.73481876e-02 8.05664539e-01
-3.82748961e-01 -9.49833810e-01 -7.34372512e-02 6.55718923e-01
-6.30257726e-01 -1.29038513e-01 -4.57722694e-01 3.45393747e-01
1.03790849e-01 4.73705649e-01 5.53834215e-02 -1.92446679e-01
2.76638120e-01 -8.91443193e-02 4.83186275e-01 -7.43116856e-01
-6.17295742e-01 2.59314358e-01 -9.35497284e-02 -5.84629774e-01
-3.82446855e-01 -6.38882518e-01 -8.44008207e-01 2.98941165e-01
-4.78508800e-01 -2.34745629e-02 5.92210650e-01 9.59469736e-01
4.20836240e-01 6.44729912e-01 8.63064170e-01 -1.05993545e+00
-6.91863835e-01 -6.24422193e-01 -5.80263674e-01 3.84089798e-01
1.05721200e+00 -4.73257571e-01 -3.79908621e-01 -4.31945082e-04] | [9.567768096923828, 1.3605122566223145] |
41c5bfc4-ee2a-47de-89e6-dcf7374e2c51 | fusion-of-sentiment-and-asset-price | 2203.05673 | null | https://arxiv.org/abs/2203.05673v1 | https://arxiv.org/pdf/2203.05673v1.pdf | Fusion of Sentiment and Asset Price Predictions for Portfolio Optimization | The fusion of public sentiment data in the form of text with stock price prediction is a topic of increasing interest within the financial community. However, the research literature seldom explores the application of investor sentiment in the Portfolio Selection problem. This paper aims to unpack and develop an enhanced understanding of the sentiment aware portfolio selection problem. To this end, the study uses a Semantic Attention Model to predict sentiment towards an asset. We select the optimal portfolio through a sentiment-aware Long Short Term Memory (LSTM) recurrent neural network for price prediction and a mean-variance strategy. Our sentiment portfolio strategies achieved on average a significant increase in revenue above the non-sentiment aware models. However, the results show that our strategy does not outperform traditional portfolio allocation strategies from a stability perspective. We argue that an improved fusion of sentiment prediction with a combination of price prediction and portfolio optimization leads to an enhanced portfolio selection strategy. | ['Terence L. Van Zyl', 'Mufhumudzi Muthivhi'] | 2022-03-10 | null | null | null | null | ['portfolio-optimization', 'stock-price-prediction'] | ['time-series', 'time-series'] | [-5.52528165e-03 -8.74745473e-02 -3.49157065e-01 -5.39443016e-01
-7.36567438e-01 -6.19467616e-01 5.23525298e-01 -9.71487723e-03
-2.32318982e-01 4.67527211e-01 6.13073647e-01 -4.01769370e-01
-5.47186434e-01 -1.17315269e+00 -4.76846397e-01 -6.18964076e-01
3.73431176e-01 2.58496910e-01 -3.52458298e-01 -4.89473671e-01
9.07782853e-01 2.40790352e-01 -1.50725842e+00 4.36076403e-01
4.53630239e-01 1.72182631e+00 1.77628562e-01 1.46127596e-01
-4.69893098e-01 1.30394673e+00 -5.35169125e-01 -1.04779780e+00
7.43365049e-01 -3.11756134e-01 -5.33573031e-01 -6.09073751e-02
3.13918628e-02 -8.22277516e-02 3.55432481e-01 1.07578921e+00
5.09457231e-01 1.02356568e-01 3.83184910e-01 -8.54006469e-01
-8.64815772e-01 1.02920520e+00 -4.28653389e-01 3.78559589e-01
-2.94346303e-01 -2.14636788e-01 1.79639232e+00 -8.00307155e-01
1.93839267e-01 8.89551103e-01 6.40439451e-01 4.32650670e-02
-7.50592828e-01 -5.75073302e-01 3.85549158e-01 -4.75126784e-03
-5.14883101e-01 -9.86777321e-02 1.06199074e+00 -4.09408987e-01
1.15700924e+00 2.51148909e-01 8.75279427e-01 7.52932906e-01
7.56812215e-01 6.77864015e-01 1.27323520e+00 -2.16658831e-01
2.37887323e-01 7.68006444e-01 2.48814270e-01 -6.42335340e-02
5.02282739e-01 2.32379869e-01 -7.05774963e-01 -4.90051322e-02
4.03399527e-01 2.70311803e-01 1.52789369e-01 -1.32928370e-02
-7.36442566e-01 1.35071087e+00 2.77805865e-01 3.31510514e-01
-7.59190023e-01 1.23954624e-01 3.86622787e-01 8.61456454e-01
1.18168211e+00 1.10473704e+00 -9.97820973e-01 -1.72121320e-02
-1.07103240e+00 4.07610476e-01 1.01393950e+00 3.21362495e-01
3.74417275e-01 4.23728019e-01 -1.92169353e-01 5.39460897e-01
5.73574603e-01 7.33698547e-01 6.34777844e-01 -7.83910990e-01
5.74146509e-01 6.48467124e-01 1.87581971e-01 -1.22536063e+00
-3.95785302e-01 -1.00877595e+00 -1.59308642e-01 3.49895865e-01
6.52718320e-02 -5.60769081e-01 -1.63092181e-01 1.35477972e+00
-3.27664137e-01 -1.73288807e-01 3.49196464e-01 6.59843504e-01
3.64528686e-01 6.13045096e-01 1.24631435e-01 -2.00060025e-01
1.33232868e+00 -1.12871635e+00 -7.25189388e-01 -5.06361425e-01
5.89843214e-01 -8.34230065e-01 5.59591413e-01 2.99188197e-01
-1.13293850e+00 -2.60841101e-01 -1.06754136e+00 5.56463778e-01
-5.22632241e-01 -1.51546985e-01 8.86038363e-01 8.30988586e-01
-1.02836454e+00 9.34810579e-01 -2.21585125e-01 1.28618777e-01
3.54268909e-01 4.10949528e-01 3.41866851e-01 7.98554063e-01
-1.34818757e+00 1.38193905e+00 5.95420264e-02 -1.12121828e-01
2.99855024e-02 -7.39363432e-01 -6.71668887e-01 2.52998888e-01
3.53736341e-01 -7.47053981e-01 1.36834240e+00 -1.79526842e+00
-1.70852745e+00 4.65808272e-01 1.85524806e-01 -1.04751694e+00
2.71579057e-01 -4.38174605e-01 -3.11205834e-01 -1.31198749e-01
8.00465047e-02 -3.94166671e-02 9.03082252e-01 -6.03029311e-01
-9.69702721e-01 -2.88326710e-01 -9.00415611e-03 3.53632092e-01
-6.45325482e-01 3.43649328e-01 5.40857196e-01 -1.25626683e+00
-3.27558309e-01 -7.38656044e-01 -4.24196959e-01 -8.74375403e-01
-2.58144410e-03 4.18497287e-02 2.62365460e-01 -4.08425480e-01
1.32843649e+00 -1.66715765e+00 -3.49060819e-02 3.93189073e-01
-2.06099659e-01 -1.61297232e-01 1.09385140e-01 4.72546995e-01
-5.14149249e-01 3.70624036e-01 1.16991624e-01 -2.64041722e-01
3.00738662e-01 -3.67982328e-01 -1.02466655e+00 3.06975484e-01
1.81078762e-01 1.23985350e+00 -3.90065372e-01 1.31266102e-01
5.11008799e-02 1.40204117e-01 -4.68818456e-01 -3.95995714e-02
-2.04488099e-01 -1.85619310e-01 -7.70117760e-01 6.84898794e-01
2.57860154e-01 -5.02986848e-01 2.44803131e-02 1.01049645e-02
-3.26738685e-01 6.69221640e-01 -9.52138066e-01 8.37288558e-01
-4.77403551e-01 5.38068712e-01 -5.09975493e-01 -8.88287246e-01
1.16615665e+00 2.58867383e-01 5.18528759e-01 -9.55805123e-01
4.55779850e-01 6.16560817e-01 9.98853967e-02 -3.18971276e-02
8.32142293e-01 -8.23388577e-01 -2.61192501e-01 1.10221457e+00
-2.58841574e-01 4.09650393e-02 -6.50058091e-02 -4.08207268e-01
4.84072566e-01 4.10187729e-02 1.00836821e-01 -5.22225797e-01
3.59550744e-01 -1.22155271e-01 3.30826640e-01 6.25504076e-01
8.70066211e-02 3.18213522e-01 6.01476431e-01 -4.02322471e-01
-7.82789469e-01 -3.53059322e-01 2.17557847e-02 1.14451826e+00
-3.45320076e-01 9.45141912e-02 -5.07903099e-01 -5.36144972e-01
3.08995485e-01 1.08560669e+00 -6.73991621e-01 -1.01940855e-01
-3.07486832e-01 -1.07885337e+00 -2.27221936e-01 6.90260708e-01
3.25764805e-01 -1.41632009e+00 -9.36624050e-01 3.78928602e-01
4.69970524e-01 -4.56275672e-01 -3.19119871e-01 3.60070646e-01
-8.95807922e-01 -7.31252313e-01 -9.08739924e-01 -3.41766655e-01
8.74822140e-02 5.49265072e-02 1.16764820e+00 -3.09141994e-01
6.75116599e-01 4.09224033e-01 -5.03441870e-01 -1.06340945e+00
5.64761385e-02 3.50140810e-01 -1.13811895e-01 3.55129719e-01
6.23030663e-01 -1.60556227e-01 -6.29973233e-01 -4.59188893e-02
-5.62242627e-01 -2.99861282e-01 5.60364604e-01 6.89932704e-01
3.03068548e-01 2.78498083e-01 1.36027730e+00 -9.54408526e-01
9.99014258e-01 -6.94556594e-01 -8.51583719e-01 1.96579382e-01
-1.31704545e+00 6.62423372e-02 1.49720594e-01 -1.83836538e-02
-1.40792072e+00 -5.55729985e-01 4.69020149e-03 -8.04883540e-02
7.07785487e-01 9.99290109e-01 2.22605437e-01 5.16915321e-02
-5.09596011e-03 -5.61474748e-02 9.87574607e-02 -3.42652649e-01
5.72251491e-02 3.60074699e-01 -2.94650823e-01 -4.22381125e-02
5.18052995e-01 3.00691754e-01 -1.80597469e-01 9.98522118e-02
-1.44517159e+00 3.39104258e-03 -1.77779213e-01 -2.06116945e-01
7.04255819e-01 -9.32512760e-01 -4.76130009e-01 5.28568447e-01
-6.05979264e-01 -1.89784110e-01 -6.41730905e-01 6.40905678e-01
-7.30989337e-01 -3.48762214e-01 -6.52721941e-01 -1.14241159e+00
-9.44699585e-01 -1.18072462e+00 8.24139416e-01 2.59170711e-01
-5.02610028e-01 -1.44759560e+00 -2.05872320e-02 5.04335999e-01
9.51665580e-01 1.17772236e-01 5.64337254e-01 -1.38042021e+00
-4.62691844e-01 -5.02123833e-01 1.91639103e-02 3.83320540e-01
2.12641954e-01 -3.38180572e-01 -9.23708081e-01 1.05963737e-01
9.00812089e-01 8.88928324e-02 1.16206670e+00 6.99012160e-01
6.40419006e-01 -2.98456162e-01 2.53950685e-01 4.21302259e-01
1.60263813e+00 6.19886994e-01 2.91797042e-01 1.24374986e+00
2.61653960e-01 1.13192797e+00 9.95726883e-01 7.65338838e-01
2.99355626e-01 2.44560078e-01 2.58309603e-01 2.52723515e-01
6.72093451e-01 4.53498252e-02 5.61296701e-01 5.85558772e-01
1.77878231e-01 -3.08750361e-01 -7.35478044e-01 4.71962780e-01
-1.87337065e+00 -1.14889050e+00 2.66687930e-01 1.92516780e+00
4.56303418e-01 5.80837071e-01 1.43055573e-01 1.67584270e-01
5.26502967e-01 5.59993982e-01 -5.06535232e-01 -8.31547201e-01
-5.57887495e-01 2.94890910e-01 1.04578733e+00 2.62093276e-01
-1.23178196e+00 8.17548633e-01 6.57423306e+00 6.42541647e-01
-1.26804399e+00 -1.55846953e-01 1.08950031e+00 -5.34402966e-01
-9.51016247e-01 -1.32740855e-01 -1.02250993e+00 5.37689984e-01
1.26908946e+00 -6.83104336e-01 -3.96149158e-02 1.01915348e+00
2.02311650e-01 2.52887249e-01 -6.84262931e-01 4.73602563e-01
2.35028535e-01 -1.68329811e+00 1.45301213e-02 3.00281316e-01
7.94328630e-01 1.82496104e-02 7.82442212e-01 1.36853263e-01
2.79675603e-01 -9.89523172e-01 1.18704987e+00 1.10782611e+00
-2.12404415e-01 -1.18401146e+00 1.32280278e+00 -8.99508148e-02
-8.86551738e-01 -6.11391783e-01 -1.26040623e-01 -4.05076176e-01
4.91352081e-01 5.13025165e-01 -4.60414030e-02 4.08660412e-01
7.53611267e-01 1.07871115e+00 -4.76923704e-01 6.46911860e-01
5.85953705e-02 5.02769887e-01 9.99103263e-02 -3.10470313e-01
5.49776018e-01 -5.40905893e-01 3.68651420e-01 6.52953506e-01
5.73002815e-01 -1.32720724e-01 -2.56281167e-01 7.28208661e-01
7.26575851e-02 4.34159249e-01 -6.38816237e-01 -4.01016355e-01
3.54744047e-02 9.25035596e-01 -7.32693970e-01 -3.90196055e-01
-7.84584820e-01 4.24933493e-01 -1.16541557e-01 5.12558930e-02
-6.05432451e-01 -3.03226382e-01 5.75365305e-01 1.83252357e-02
6.84212804e-01 3.80511105e-01 -1.08156466e+00 -9.96017098e-01
5.19091543e-03 -7.18929112e-01 2.95276195e-01 -5.66979110e-01
-1.61916947e+00 4.36980903e-01 -2.22308472e-01 -1.13820195e+00
-3.68759841e-01 -7.33096659e-01 -8.94639671e-01 1.03066492e+00
-1.93961990e+00 -6.29880548e-01 5.72325051e-01 -6.79507703e-02
5.55950224e-01 -7.95448124e-01 5.57592332e-01 -3.64218913e-02
-4.42515999e-01 2.77156800e-01 2.00195059e-01 -1.79746538e-01
4.27787900e-01 -1.28871059e+00 4.34791774e-01 4.62919146e-01
-1.22034192e-01 5.10045290e-01 7.11719573e-01 -9.00071859e-01
-1.11877120e+00 -1.09869516e+00 1.24711478e+00 -5.90256691e-01
1.15053976e+00 3.95631015e-01 -5.88919401e-01 7.12721705e-01
6.74630046e-01 -8.34051371e-01 1.18667829e+00 -1.02129832e-01
2.77215801e-02 -3.17183137e-01 -1.07031131e+00 5.70528924e-01
3.28530580e-01 -3.74646842e-01 -8.27673793e-01 8.46618041e-02
8.86870563e-01 1.94020510e-01 -1.18656278e+00 3.29612345e-01
6.26234353e-01 -1.20339155e+00 8.95421922e-01 -2.44734928e-01
6.41458631e-01 2.68896639e-01 -4.23261374e-01 -1.11137342e+00
-2.56043166e-01 -4.22926933e-01 1.09234661e-01 1.19851792e+00
8.20645809e-01 -1.28559852e+00 9.87228215e-01 9.14506376e-01
1.42439231e-01 -1.14137602e+00 -6.39328420e-01 -4.69519138e-01
4.15731639e-01 -5.48621356e-01 9.02506948e-01 6.26731932e-01
-5.66194020e-02 2.59537548e-01 -4.69901621e-01 -4.73444462e-01
4.72100317e-01 8.52342665e-01 2.01062992e-01 -1.28568435e+00
-3.07385325e-01 -1.14170146e+00 -1.09200412e-02 -4.50797260e-01
4.14493442e-01 -7.93746769e-01 -5.95237315e-01 -1.39431846e+00
-1.05540849e-01 -7.91576654e-02 -9.35027063e-01 1.14674337e-01
-3.91553342e-02 2.49075871e-02 4.90455598e-01 1.06878363e-01
-3.12609196e-01 6.68324888e-01 8.45062315e-01 -3.92704085e-02
-1.41584396e-01 4.25051630e-01 -1.68785691e+00 9.52391744e-01
1.27529752e+00 -4.41578001e-01 -2.90008456e-01 -1.30446672e-01
1.12378478e+00 8.01020190e-02 -1.64787844e-01 -2.50942200e-01
6.09161817e-02 -3.19543004e-01 4.29431736e-01 -8.52361798e-01
1.68488711e-01 -7.19950974e-01 1.85716227e-02 4.71837670e-01
-5.42736232e-01 4.68710423e-01 1.10975258e-01 3.86388242e-01
-3.83625865e-01 -5.88073373e-01 3.72782946e-01 -4.38806921e-01
-5.04938602e-01 5.88881932e-02 -5.31346440e-01 -2.72769749e-01
9.40108299e-01 -2.62041152e-01 -1.73082370e-02 -4.54733104e-01
-4.85471755e-01 2.60439694e-01 2.73477763e-01 6.29794419e-01
4.48906153e-01 -1.26861477e+00 -5.98549902e-01 2.12511197e-02
-2.40684256e-01 -6.58512771e-01 -7.42496625e-02 7.06329882e-01
-2.25586623e-01 8.13046455e-01 3.53616960e-02 1.98723555e-01
-7.72867858e-01 3.40453178e-01 5.72084248e-01 -6.14134789e-01
-2.94487268e-01 1.04645431e+00 7.00287297e-02 -1.48485035e-01
1.60344262e-02 -1.38156548e-01 -8.31840754e-01 9.64545190e-01
5.78759491e-01 3.79031032e-01 3.94918881e-02 -7.63843775e-01
-6.93551227e-02 8.19456816e-01 -1.64603572e-02 -3.09841514e-01
1.79306948e+00 -2.48441607e-01 -2.49702916e-01 7.56505847e-01
1.00024605e+00 5.39097153e-02 -9.68662441e-01 -1.35918230e-01
8.82742465e-01 -3.50114793e-01 2.05253944e-01 -8.08912814e-01
-1.66907072e+00 4.03931260e-01 2.23134786e-01 4.93959457e-01
1.13479698e+00 -3.14220697e-01 8.97883594e-01 3.59806567e-01
2.37623081e-01 -1.34217834e+00 -8.36220235e-02 5.42919219e-01
1.06072879e+00 -1.27535188e+00 9.89635438e-02 2.97843039e-01
-1.12121141e+00 1.07907677e+00 1.99991137e-01 -6.81784332e-01
1.27722442e+00 2.01372534e-01 3.04221272e-01 -4.07555640e-01
-1.17565989e+00 -1.86647519e-01 3.83642375e-01 -1.47284240e-01
5.05770445e-01 -1.78143103e-02 -2.70417094e-01 1.26571965e+00
-8.30374658e-01 -2.95997173e-01 5.32031357e-01 8.28113496e-01
-5.94925880e-01 -1.13477719e+00 -2.22390443e-01 1.03985107e+00
-1.37213039e+00 -4.30767328e-01 -4.02306437e-01 4.65183526e-01
-2.82999218e-01 9.26930249e-01 3.71989489e-01 -3.29854250e-01
4.29800570e-01 2.45693505e-01 -2.50825644e-01 -7.65318215e-01
-1.48269176e+00 4.27692324e-01 2.12993816e-01 -3.42098624e-01
-6.91681147e-01 -1.09153986e+00 -8.64678264e-01 -1.65605158e-01
-5.40276349e-01 2.08392292e-01 6.71893299e-01 1.01561952e+00
3.48440647e-01 7.57366598e-01 9.49612617e-01 -5.33899844e-01
-9.73858118e-01 -7.26981163e-01 -9.95444655e-01 2.99584065e-02
2.04261184e-01 -5.26448548e-01 -6.01913273e-01 -4.23985302e-01] | [4.408894062042236, 4.11757230758667] |
791dd8c6-28ce-43a4-b30f-57f02937cb13 | data-roaming-and-early-fusion-for-composed | 2303.09429 | null | https://arxiv.org/abs/2303.09429v1 | https://arxiv.org/pdf/2303.09429v1.pdf | Data Roaming and Early Fusion for Composed Image Retrieval | We study the task of Composed Image Retrieval (CoIR), where a query is composed of two modalities, image and text, extending the user's expression ability. Previous methods typically address this task by a separate encoding of each query modality, followed by late fusion of the extracted features. In this paper, we propose a new approach, Cross-Attention driven Shift Encoder (CASE), employing early fusion between modalities through a cross-attention module with an additional auxiliary task. We show that our method outperforms the existing state-of-the-art, on established benchmarks (FashionIQ and CIRR) by a large margin. However, CoIR datasets are a few orders of magnitude smaller compared to other vision and language (V&L) datasets, and some suffer from serious flaws (e.g., queries with a redundant modality). We address these shortcomings by introducing Large Scale Composed Image Retrieval (LaSCo), a new CoIR dataset x10 times larger than current ones. Pre-training on LaSCo yields a further performance boost. We further suggest a new analysis of CoIR datasets and methods, for detecting modality redundancy or necessity, in queries. | ['Dani Lischinski', 'Nir Darshan', 'Rami Ben-Ari', 'Matan Levy'] | 2023-03-16 | null | null | null | null | ['composed-image-retrieval'] | ['computer-vision'] | [ 5.19575715e-01 -3.14158112e-01 -2.84995288e-01 -2.99413830e-01
-1.40144801e+00 -7.50662386e-01 1.06331849e+00 8.24281424e-02
-6.79077446e-01 5.62409222e-01 3.66324306e-01 4.29631509e-02
-2.46729642e-01 -3.13828647e-01 -7.82996595e-01 -5.26343226e-01
2.52413243e-01 2.97991157e-01 2.12876111e-01 -5.01499534e-01
3.39265078e-01 3.52196723e-01 -2.04739118e+00 6.77491307e-01
4.94300097e-01 1.34434712e+00 1.48425996e-01 5.76517045e-01
1.50215641e-01 9.89334285e-01 -4.44682777e-01 -4.30338264e-01
6.38951957e-02 -3.17315280e-01 -1.00564778e+00 5.18049859e-02
7.36549199e-01 -6.25959694e-01 -6.79600298e-01 7.25855052e-01
9.31466401e-01 1.61665931e-01 6.20666265e-01 -1.27283096e+00
-1.23041570e+00 4.69145566e-01 -6.14401460e-01 1.74124822e-01
8.34021986e-01 -1.69369191e-01 1.06433511e+00 -1.08192039e+00
7.63708711e-01 1.38365304e+00 1.79541752e-01 6.86708450e-01
-1.05271745e+00 -5.77809751e-01 -2.32821833e-02 3.70372176e-01
-1.63929200e+00 -4.66642231e-01 7.08967745e-01 -1.91744328e-01
9.08423841e-01 6.75069153e-01 -1.16346270e-01 1.15185523e+00
-1.98573887e-01 1.14489198e+00 1.28151977e+00 -5.78276932e-01
-1.66455001e-01 3.75633776e-01 2.84664452e-01 3.56437147e-01
-3.99563581e-01 7.82879964e-02 -5.33130407e-01 -1.69739813e-01
4.10742074e-01 2.56676614e-01 -3.62715393e-01 -1.77636564e-01
-1.48521769e+00 5.96281052e-01 5.32513559e-01 4.99354392e-01
-2.97362626e-01 8.15113559e-02 4.78673279e-01 7.26198494e-01
2.61053562e-01 3.22600961e-01 -4.13661659e-01 2.83017755e-01
-8.17081630e-01 2.13624746e-01 3.59900087e-01 1.07293427e+00
7.80842602e-01 -5.44225395e-01 -6.95909917e-01 1.06076050e+00
1.00210689e-01 6.43536508e-01 7.10981309e-01 -8.14802110e-01
1.60646364e-01 4.43525791e-01 1.16004340e-01 -6.58637881e-01
-1.09090135e-02 -1.69384956e-01 -7.92934239e-01 -1.03510126e-01
-7.84929376e-03 3.07504117e-01 -9.70846236e-01 1.76858532e+00
-1.78808011e-02 -2.85998166e-01 2.03253269e-01 1.11619878e+00
1.24687767e+00 7.10091889e-01 8.24092403e-02 -2.55457498e-02
1.68782830e+00 -1.14844906e+00 -7.17258871e-01 1.25325263e-01
4.24527526e-01 -8.66535306e-01 1.19533789e+00 3.35526407e-01
-1.26529157e+00 -7.67695844e-01 -6.69389367e-01 -5.53362608e-01
-7.91967869e-01 3.55080456e-01 4.86313909e-01 2.29107887e-01
-1.41353798e+00 8.66525546e-02 8.93043280e-02 -5.28880000e-01
1.05733193e-01 4.60270077e-01 -8.71822417e-01 -4.18915093e-01
-1.42634356e+00 8.97960424e-01 2.87164152e-01 -1.12077348e-01
-7.89849162e-01 -4.52165574e-01 -8.91678035e-01 3.13275814e-01
5.93312860e-01 -5.93175054e-01 1.13006318e+00 -1.22812271e+00
-1.25159633e+00 1.22116542e+00 -1.97852850e-01 -1.58954933e-01
3.37947488e-01 -3.71915698e-01 -4.80485678e-01 2.51116246e-01
-2.52697691e-02 1.07856953e+00 1.09475458e+00 -1.54037273e+00
-4.83283192e-01 -4.61386472e-01 5.35569727e-01 3.50949764e-01
-4.02446568e-01 4.45640653e-01 -1.15589094e+00 -4.88556892e-01
-3.41337621e-01 -9.63445246e-01 1.47819385e-01 -5.92400767e-02
-2.79476345e-01 -6.23792768e-01 8.15568507e-01 -5.96498191e-01
1.44438469e+00 -2.43522859e+00 2.47309148e-01 -5.15916676e-04
2.23287895e-01 2.46048227e-01 -6.74237847e-01 6.60418808e-01
-9.08899382e-02 1.12146258e-01 -8.13610852e-02 -4.36672360e-01
1.80960614e-02 4.87228632e-02 -4.54255104e-01 2.04668000e-01
3.05604279e-01 1.24055171e+00 -8.18583429e-01 -7.16139674e-01
-1.08424956e-02 5.97526371e-01 -3.97780329e-01 2.78959692e-01
-7.29253609e-03 3.06706041e-01 -2.78331459e-01 9.47161078e-01
6.78244591e-01 -4.85863209e-01 -2.00129807e-01 -5.11012912e-01
2.29868642e-03 -3.40666950e-01 -7.07596123e-01 2.13783932e+00
-5.35649538e-01 6.67074919e-01 -4.72516157e-02 -8.70101810e-01
6.05279207e-01 3.87733459e-01 5.61258078e-01 -1.18601036e+00
2.65639812e-01 1.18668109e-01 -3.68405938e-01 -7.44446456e-01
9.39125359e-01 2.19163731e-01 -2.89953977e-01 4.49973762e-01
1.56875357e-01 1.16786316e-01 3.23356479e-01 4.80697095e-01
1.07511997e+00 1.98220126e-02 3.43018807e-02 1.46327600e-01
8.15310061e-01 -2.27563307e-01 -5.57926372e-02 9.92343187e-01
-2.10667133e-01 7.95411527e-01 2.59365112e-01 -4.88596670e-02
-9.30549204e-01 -8.71708214e-01 -8.23044330e-02 1.65821731e+00
3.84652168e-01 -3.44662845e-01 -3.24282736e-01 -7.04302013e-01
1.15960598e-01 2.82382101e-01 -8.11937809e-01 -2.88592130e-01
-3.83888781e-01 -5.13292968e-01 5.71133375e-01 5.16154468e-01
4.54802692e-01 -1.16767323e+00 -5.52754581e-01 -2.15218082e-01
-2.30392590e-01 -1.34928334e+00 -7.73428082e-01 -2.07705215e-01
-3.91791046e-01 -7.61034369e-01 -1.22481334e+00 -8.58096182e-01
4.86190468e-01 5.86336792e-01 1.26898634e+00 2.26172656e-01
-4.42739367e-01 1.03025627e+00 -5.92445850e-01 -8.38750303e-02
-5.36631458e-02 -1.56989366e-01 -2.65723616e-01 2.48644739e-01
3.28056365e-01 -2.26265248e-02 -6.95456505e-01 3.22870582e-01
-1.53272700e+00 -6.23321021e-03 1.05022490e+00 1.24350607e+00
5.14715791e-01 -2.30172232e-01 5.63513696e-01 -7.01036036e-01
7.58492470e-01 -5.65012932e-01 -3.38684022e-01 7.25675821e-01
-5.99896669e-01 1.81594059e-01 1.13787435e-01 -5.58165848e-01
-1.24703634e+00 -8.04383680e-02 1.66714594e-01 -6.96757495e-01
-1.45548925e-01 6.52252078e-01 2.08707601e-02 -7.09424242e-02
5.64637542e-01 4.01514977e-01 -9.22361463e-02 -5.10523081e-01
7.51294851e-01 9.38367963e-01 7.31714189e-01 -3.47846001e-01
4.40257132e-01 3.51569653e-01 -1.85534090e-01 -5.79403639e-01
-6.08495295e-01 -8.63316476e-01 -4.72677469e-01 -3.96728665e-02
7.26047754e-01 -1.16481102e+00 -6.95917904e-01 3.53592873e-01
-1.17270243e+00 8.32022429e-02 -1.52973562e-01 3.65166605e-01
-3.57968867e-01 4.68394190e-01 -6.87775791e-01 -7.42979825e-01
-4.52559710e-01 -1.06079578e+00 1.75144744e+00 1.02267437e-01
3.13156158e-01 -6.08631670e-01 4.65144329e-02 6.18464589e-01
6.12843573e-01 -1.71005532e-01 7.36899495e-01 -7.46080279e-01
-4.77054268e-01 -2.72407591e-01 -9.82518971e-01 4.32669222e-01
-6.39603361e-02 -2.44493008e-01 -1.24854493e+00 -5.05273581e-01
-3.73562694e-01 -9.27747071e-01 1.22652590e+00 -5.75874373e-02
1.36681437e+00 -1.21635720e-01 -3.04562777e-01 8.93101394e-02
1.52038467e+00 1.80731580e-01 8.18521619e-01 6.42140731e-02
3.35217655e-01 7.43063211e-01 7.21937239e-01 3.29273194e-01
4.44838017e-01 7.18192458e-01 2.12940782e-01 -2.92546391e-01
-4.06831831e-01 1.74349118e-02 4.04723406e-01 6.46445096e-01
-8.74701366e-02 -4.82890457e-01 -6.29881382e-01 5.96565187e-01
-1.89692712e+00 -8.87091041e-01 3.37485969e-01 2.13076639e+00
8.08569014e-01 -4.38393563e-01 -1.42766967e-01 9.39685479e-02
5.77575386e-01 7.86565244e-03 -4.27232623e-01 -2.96837650e-02
-4.30007607e-01 2.87540346e-01 3.03113043e-01 4.00534600e-01
-1.19618452e+00 9.28052843e-01 6.57244921e+00 9.97098386e-01
-1.16074991e+00 2.14234069e-01 4.95564908e-01 -1.15346105e-03
-4.85745728e-01 -1.58407465e-01 -5.76029241e-01 2.91243523e-01
6.52804255e-01 8.83553401e-02 4.97458100e-01 4.83816296e-01
-4.71192747e-01 -1.24840051e-01 -1.21073389e+00 1.43413663e+00
8.56575072e-01 -8.84493589e-01 2.32612446e-01 -2.94353724e-01
5.49644113e-01 1.08019061e-01 3.27612311e-01 6.45142734e-01
-1.31000325e-01 -8.42663169e-01 4.11221862e-01 6.73181951e-01
1.24356604e+00 -2.66859531e-01 8.87544394e-01 4.83343638e-02
-8.54230285e-01 -2.39861563e-01 -1.50683284e-01 3.03862751e-01
-1.81092113e-01 4.96904105e-02 -1.43553138e-01 8.01392615e-01
8.52382898e-01 6.17986500e-01 -9.93324220e-01 6.58292711e-01
2.80312449e-01 -1.01214123e-03 -2.19255343e-01 1.33672833e-01
1.64718792e-01 2.10220799e-01 2.72719175e-01 1.30081630e+00
2.34940469e-01 -5.40178344e-02 3.15846354e-02 5.61701238e-01
-4.29967761e-01 2.62737423e-01 -6.47641957e-01 -1.14520816e-02
1.91335976e-01 1.29924309e+00 -2.21637756e-01 -7.26431549e-01
-8.67453635e-01 1.47807217e+00 1.91349030e-01 6.97834194e-01
-7.19918251e-01 -4.21618849e-01 5.09498678e-02 -3.19050997e-01
3.37539136e-01 3.64312917e-01 4.93778676e-01 -1.52807999e+00
9.26318169e-02 -9.97546554e-01 6.95239663e-01 -1.29674983e+00
-1.46943402e+00 7.86653280e-01 1.88223511e-01 -1.32330775e+00
-3.48766685e-01 -5.00697970e-01 1.61895722e-01 6.85518563e-01
-2.06413627e+00 -1.47365344e+00 -4.07423705e-01 1.03157711e+00
6.23775065e-01 -1.31905913e-01 7.79390097e-01 7.56933033e-01
-2.22121209e-01 7.79595435e-01 1.04527459e-01 -1.16717210e-02
1.13975942e+00 -1.02652013e+00 -3.01474571e-01 2.58209288e-01
2.86694676e-01 3.71762574e-01 3.39643121e-01 -1.65982962e-01
-1.77816582e+00 -7.69906342e-01 1.08374405e+00 -5.45488179e-01
5.81134260e-01 -2.36439541e-01 -7.91822135e-01 7.00687349e-01
7.07303226e-01 3.26401144e-01 6.23275936e-01 -1.22036353e-01
-7.32004344e-01 -1.75897658e-01 -1.08310628e+00 3.84216756e-01
9.23932433e-01 -9.46960986e-01 -6.57979906e-01 2.80426741e-01
8.98557961e-01 -1.27950102e-01 -9.26029325e-01 6.43227994e-01
8.25526774e-01 -8.27093780e-01 1.16058290e+00 -7.25720525e-01
4.13843602e-01 -1.35664299e-01 -5.68329632e-01 -8.96101832e-01
-3.39104325e-01 -3.00994694e-01 -1.18834682e-01 1.16905713e+00
1.15392238e-01 -3.08620751e-01 3.03643107e-01 4.65078950e-01
1.06734909e-01 -5.24338126e-01 -8.46570194e-01 -6.67718649e-01
-2.15329394e-01 -2.48670548e-01 4.38593477e-01 9.15608466e-01
3.84417875e-03 6.28006935e-01 -6.17871284e-01 -1.43117458e-01
3.31610262e-01 2.43099779e-01 7.21503735e-01 -8.43696177e-01
-2.72073984e-01 -4.98444259e-01 -4.15535331e-01 -1.12775230e+00
2.08336011e-01 -7.42249846e-01 -1.08323656e-01 -1.25435233e+00
6.57195151e-01 -2.01825917e-01 -8.69870722e-01 6.00883901e-01
-2.34812737e-01 6.95590317e-01 3.86843979e-01 3.97230297e-01
-1.12589765e+00 6.06199086e-01 1.14685416e+00 -4.84841913e-01
1.44478101e-02 -5.42047381e-01 -6.21020317e-01 2.34924436e-01
3.66020232e-01 2.91008689e-02 -3.63395661e-01 -4.94644701e-01
2.76352793e-01 1.35236755e-01 5.17684996e-01 -6.53585851e-01
2.59101808e-01 2.17228979e-01 2.36436918e-01 -7.64124572e-01
4.94047403e-01 -1.07087362e+00 -6.76952824e-02 9.67929363e-02
-8.13280046e-01 1.95301920e-01 1.95485175e-01 4.27317321e-01
-7.55812287e-01 -1.82110574e-02 2.76148468e-01 -6.85922503e-02
-1.02658415e+00 2.82728165e-01 -1.73227012e-01 -3.02436441e-01
8.51756275e-01 3.48858595e-01 -5.12003839e-01 -6.15111411e-01
-6.04044259e-01 2.96214640e-01 2.27800012e-01 7.14444101e-01
7.81333745e-01 -1.50436604e+00 -7.99420774e-01 2.05950648e-01
8.17835748e-01 -5.13928533e-01 5.62783659e-01 8.46276641e-01
-9.52628255e-02 6.99577630e-01 -1.16438448e-01 -6.10547245e-01
-1.54166687e+00 1.26931429e+00 -8.50112066e-02 -4.82235104e-01
-1.21554233e-01 7.29124129e-01 4.65155602e-01 -2.54113078e-01
3.20933342e-01 -1.11763082e-01 -3.65097702e-01 2.36324266e-01
6.35345459e-01 1.84756413e-01 -5.40078655e-02 -9.10931170e-01
-2.53690153e-01 7.79892206e-01 -2.35638514e-01 -2.16376364e-01
9.04445648e-01 -5.48371255e-01 -3.78545016e-01 3.45496595e-01
1.48731816e+00 -1.09630682e-01 -6.54042304e-01 -8.40942621e-01
-2.21676156e-01 -5.39242387e-01 6.65497631e-02 -1.11766577e+00
-1.03110909e+00 8.21337998e-01 1.01766586e+00 1.52929276e-01
1.75354135e+00 5.09001315e-01 7.31264472e-01 5.94758809e-01
1.34526491e-01 -9.72469807e-01 2.73114413e-01 1.89894497e-01
1.34691167e+00 -1.85149193e+00 -7.24161863e-02 -1.41071528e-01
-6.23575866e-01 7.72224247e-01 5.06697476e-01 9.31864902e-02
6.30585849e-01 -2.17319354e-01 1.04447119e-01 -2.27395743e-01
-9.90613699e-01 -6.07056141e-01 7.35138655e-01 1.70720994e-01
4.88994688e-01 -2.34658286e-01 -3.30183059e-01 3.72388572e-01
3.90174001e-01 1.77790329e-01 -8.38203058e-02 1.15764737e+00
-3.68098691e-02 -1.06713986e+00 -5.13557076e-01 3.10940206e-01
-6.03947461e-01 -2.64572799e-01 -5.39894104e-01 7.37944663e-01
-1.73110596e-03 8.56747568e-01 1.45609140e-01 -4.25048947e-01
3.39089066e-01 5.62672019e-02 6.12724543e-01 -1.95854068e-01
-5.98446369e-01 2.33204871e-01 -7.67019093e-02 -6.73479140e-01
-8.34199667e-01 -3.42179805e-01 -5.69158018e-01 9.93662104e-02
-4.19642955e-01 -9.46976095e-02 5.39977670e-01 6.32639825e-01
6.76374733e-01 3.49006206e-01 7.26161480e-01 -7.44500518e-01
-4.76266623e-01 -1.09589386e+00 -5.74059069e-01 1.05257380e+00
5.38790941e-01 -5.37396610e-01 -1.85684696e-01 1.90901130e-01] | [10.900561332702637, 1.0861140489578247] |
06bd410f-af9d-42bb-be9b-ce0280a160f5 | guided-image-to-image-translation-with-bi-1 | 1910.11328 | null | https://arxiv.org/abs/1910.11328v1 | https://arxiv.org/pdf/1910.11328v1.pdf | Guided Image-to-Image Translation with Bi-Directional Feature Transformation | We address the problem of guided image-to-image translation where we translate an input image into another while respecting the constraints provided by an external, user-provided guidance image. Various conditioning methods for leveraging the given guidance image have been explored, including input concatenation , feature concatenation, and conditional affine transformation of feature activations. All these conditioning mechanisms, however, are uni-directional, i.e., no information flow from the input image back to the guidance. To better utilize the constraints of the guidance image, we present a bi-directional feature transformation (bFT) scheme. We show that our bFT scheme outperforms other conditioning schemes and has comparable results to state-of-the-art methods on different tasks. | ['Jia-Bin Huang', 'Badour AlBahar'] | 2019-10-24 | guided-image-to-image-translation-with-bi | http://openaccess.thecvf.com/content_ICCV_2019/html/AlBahar_Guided_Image-to-Image_Translation_With_Bi-Directional_Feature_Transformation_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/AlBahar_Guided_Image-to-Image_Translation_With_Bi-Directional_Feature_Transformation_ICCV_2019_paper.pdf | iccv-2019-10 | ['pose-transfer'] | ['computer-vision'] | [ 9.13714230e-01 2.38872059e-02 -2.72080809e-01 -5.51205218e-01
-6.96790814e-01 -6.34625256e-01 9.31699991e-01 -2.88636416e-01
-5.60480118e-01 4.97830063e-01 2.84328401e-01 -1.31776184e-01
1.73533991e-01 -4.85141963e-01 -8.18935394e-01 -6.47461414e-01
5.17962694e-01 5.76392561e-02 -1.55785242e-02 -3.62777002e-02
5.32881558e-01 3.85793924e-01 -1.09405518e+00 4.42354977e-01
6.48979366e-01 7.61678159e-01 4.19828564e-01 7.76338518e-01
1.61393732e-01 2.20514506e-01 -2.06243530e-01 -1.80186182e-01
3.46170157e-01 -6.70556843e-01 -6.97679341e-01 3.97627115e-01
7.61503816e-01 -4.88537490e-01 -5.10198653e-01 1.15696764e+00
3.38923663e-01 2.46895552e-01 7.67518282e-01 -1.46357417e+00
-1.14291322e+00 1.95177749e-01 -5.82076967e-01 1.01290435e-01
4.17043328e-01 1.15583830e-01 9.67714965e-01 -1.15432084e+00
8.89469147e-01 1.22093558e+00 9.89297871e-03 7.15857744e-01
-1.56620014e+00 -5.14060378e-01 5.84896028e-01 -3.48459542e-01
-9.81416523e-01 -4.45972919e-01 6.79235578e-01 -5.14042497e-01
9.11835313e-01 6.19494990e-02 5.28013885e-01 1.01777387e+00
3.16644102e-01 8.94812107e-01 1.28760576e+00 -6.00205064e-01
-1.54212102e-01 7.33071566e-02 8.26502498e-03 8.68962586e-01
-9.31560155e-03 4.82321709e-01 -8.21732402e-01 -3.26269530e-02
1.15680814e+00 -1.14093371e-01 -3.55717689e-01 -4.14196879e-01
-1.51577973e+00 8.15338612e-01 5.35480618e-01 8.43506083e-02
-1.84268355e-01 4.12697703e-01 1.14625417e-01 4.24483180e-01
3.69619668e-01 5.69116831e-01 -3.78535122e-01 1.59258798e-01
-9.00484264e-01 3.54534239e-01 3.97752374e-01 1.17314386e+00
9.69155192e-01 5.71418591e-02 -4.98451531e-01 2.62847602e-01
3.52513701e-01 5.98758340e-01 4.14257199e-01 -9.92828548e-01
7.12416708e-01 2.96093643e-01 3.75063159e-02 -9.10905778e-01
-2.30849728e-01 -2.67288685e-01 -6.85639024e-01 2.10952640e-01
2.25052580e-01 -1.46631703e-01 -1.34528577e+00 2.18018198e+00
7.95583501e-02 3.74850929e-02 -4.61485572e-02 1.04465926e+00
6.76856279e-01 6.61178112e-01 -6.59217462e-02 -6.08901195e-02
9.20098901e-01 -1.17251313e+00 -7.13467836e-01 -2.53950030e-01
4.43972766e-01 -8.31169903e-01 1.36927283e+00 1.14294231e-01
-1.10554910e+00 -6.80563629e-01 -1.08025002e+00 -3.13823342e-01
-2.27300748e-01 3.44160557e-01 6.63788021e-01 3.52745563e-01
-1.14831722e+00 4.36593026e-01 -9.12521482e-01 -4.28439468e-01
2.88884014e-01 4.12020862e-01 -7.75143504e-01 -1.97104320e-01
-8.14386845e-01 8.38687778e-01 3.55011062e-03 -8.45286250e-02
-1.00123227e+00 -6.28864646e-01 -1.07734430e+00 -8.70351270e-02
3.00771415e-01 -1.02776337e+00 1.32624722e+00 -1.06476033e+00
-1.45673990e+00 9.81107354e-01 -3.38583738e-01 -1.07920118e-01
3.96356642e-01 -6.53908372e-01 1.04356825e-01 7.62585178e-02
1.46636993e-01 1.12070048e+00 1.09947181e+00 -1.28481936e+00
-4.34419572e-01 -2.63451278e-01 2.16293752e-01 5.04796386e-01
-2.39072442e-01 5.88346422e-02 -7.68615007e-01 -7.66633213e-01
2.93405175e-01 -1.11121202e+00 -3.94756585e-01 3.55374992e-01
-8.30356598e-01 2.11260185e-01 9.61148024e-01 -2.35146493e-01
1.06561279e+00 -2.33276606e+00 4.76861447e-01 2.64292330e-01
5.23383021e-02 5.59867471e-02 -5.54776490e-01 3.93070936e-01
-1.47429839e-01 1.80862993e-01 -3.23609233e-01 -6.52099490e-01
2.60873809e-02 3.34026426e-01 -5.09062827e-01 4.65030044e-01
3.53211641e-01 1.18535793e+00 -9.02437150e-01 -3.28411579e-01
9.13965553e-02 5.01888990e-01 -8.03169966e-01 5.27483881e-01
-1.59922037e-02 8.01625252e-01 -5.60380399e-01 4.69961107e-01
4.31168169e-01 -8.96793082e-02 -3.00183415e-01 -3.56539249e-01
-6.42866194e-02 2.31933907e-01 -8.43473911e-01 2.05439258e+00
-3.04609954e-01 4.80061769e-01 -4.28129807e-02 -7.39000142e-01
8.33751917e-01 1.76538318e-01 2.68842489e-01 -6.55849934e-01
1.05617359e-01 -2.61208117e-02 -3.58815305e-02 -2.23198503e-01
3.54224473e-01 -1.51129901e-01 5.58319017e-02 6.89108849e-01
2.40029201e-01 -4.71305013e-01 1.10127203e-01 4.87298101e-01
7.80263126e-01 6.57047868e-01 2.39195287e-01 -2.82890916e-01
4.88242537e-01 -2.45043337e-01 5.04164934e-01 8.49579811e-01
-1.18778870e-01 1.05742645e+00 3.49989891e-01 -1.96429610e-01
-1.03287494e+00 -9.73546743e-01 2.94623137e-01 1.10058427e+00
1.03917599e-01 -4.60030526e-01 -9.02265906e-01 -7.89920807e-01
-1.65729839e-02 4.07426178e-01 -8.47870708e-01 -2.78335333e-01
-5.86880326e-01 -2.67063409e-01 4.06779379e-01 7.72390187e-01
7.32235551e-01 -9.00675893e-01 -6.70955360e-01 1.21534273e-01
-1.13356665e-01 -1.26211703e+00 -1.22493887e+00 4.51176539e-02
-1.04107857e+00 -8.59409928e-01 -6.94193125e-01 -7.02198088e-01
1.34222662e+00 5.05112171e-01 9.80983913e-01 1.49929792e-01
-1.44737318e-01 6.43795967e-01 -1.15757428e-01 -7.40332007e-02
-8.21408778e-02 -3.50716934e-02 3.17564979e-02 1.68056443e-01
-1.43344209e-01 -3.91216278e-01 -5.44294298e-01 4.49872404e-01
-1.05845904e+00 6.06322587e-01 6.63245440e-01 1.21818888e+00
9.92189646e-01 -5.76348662e-01 2.38677934e-01 -1.10361052e+00
6.43744230e-01 1.14601731e-01 -4.33769643e-01 2.65569985e-01
-5.98000944e-01 3.59026670e-01 3.96976113e-01 -6.46464109e-01
-1.23080194e+00 4.60011363e-01 6.26394153e-02 -5.06233633e-01
-7.65793025e-02 4.41438675e-01 -4.05001521e-01 -2.62313932e-01
6.00895762e-01 1.95320547e-01 -4.18465026e-02 -3.82043034e-01
8.52419317e-01 2.83859104e-01 8.64314616e-01 -9.23701286e-01
9.26266313e-01 5.87514102e-01 1.97891649e-02 -3.33610237e-01
-7.67304182e-01 -3.12812597e-01 -1.11333477e+00 9.11923200e-02
8.08715940e-01 -5.94865859e-01 -1.12583071e-01 4.49500322e-01
-1.23181713e+00 -3.83894056e-01 -2.53026873e-01 5.12575924e-01
-9.40793753e-01 1.70324087e-01 -4.88916457e-01 -4.27023858e-01
-2.74348378e-01 -1.43201363e+00 1.17944109e+00 2.14533612e-01
-2.09280267e-01 -9.33742523e-01 -5.10048866e-02 -3.19800004e-02
5.78429341e-01 2.99546629e-01 8.09795678e-01 -3.27922881e-01
-6.56142354e-01 -2.43954271e-01 -3.71963561e-01 4.18489128e-01
4.06493366e-01 -2.09004469e-02 -7.62437284e-01 -5.71322203e-01
-3.84240150e-02 -4.15275604e-01 9.67549086e-01 2.82755762e-01
8.77101600e-01 -2.16675296e-01 -3.92662346e-01 9.21479285e-01
1.22334003e+00 3.34486514e-02 6.15215898e-01 2.98330396e-01
6.67369664e-01 3.67267698e-01 7.17508495e-01 1.10158194e-02
1.30527347e-01 6.30558848e-01 3.55651706e-01 -5.18805504e-01
-3.08522463e-01 -4.89626199e-01 4.20305401e-01 3.53308141e-01
-1.07312784e-01 -1.72583997e-01 -3.56727451e-01 4.32669550e-01
-1.94825745e+00 -4.93285000e-01 1.02948420e-01 2.16395855e+00
6.98094904e-01 8.73142928e-02 -2.95156181e-01 -2.01000631e-01
4.65979487e-01 2.26735428e-01 -7.52974331e-01 -4.86139446e-01
-9.46402997e-02 1.33711016e-02 3.85083139e-01 6.07108653e-01
-1.11361897e+00 1.04268992e+00 7.49421215e+00 2.92919517e-01
-1.37502861e+00 8.71567614e-03 4.54495251e-01 1.06969737e-01
-4.32398915e-01 1.81089759e-01 -6.37139201e-01 -9.20521766e-02
2.97771126e-01 -1.71659157e-01 4.55442935e-01 5.99415004e-01
1.55510843e-01 -9.03292820e-02 -1.44693947e+00 8.46644282e-01
2.00770408e-01 -1.02214980e+00 3.36168498e-01 1.28188118e-01
9.82424557e-01 -1.12216398e-01 4.91275221e-01 -5.70924357e-02
1.74527600e-01 -9.29440081e-01 6.34287298e-01 5.54818034e-01
1.20622003e+00 -4.52657849e-01 3.45204860e-01 1.91692755e-01
-9.49054182e-01 3.22012097e-01 -1.00256614e-01 2.52512991e-02
2.61848152e-01 3.75451893e-01 -3.96109104e-01 5.20860076e-01
2.96691746e-01 6.19386554e-01 -5.59990942e-01 7.91700423e-01
-6.28407776e-01 4.12483186e-01 -2.56462842e-01 4.80678976e-01
4.93159622e-01 -1.02994852e-01 5.40672898e-01 1.31097400e+00
2.68860012e-01 1.47377193e-01 2.60588676e-01 9.79806840e-01
-3.05507183e-02 4.11665365e-02 -8.73253882e-01 2.84843836e-02
-1.17126629e-01 1.20026207e+00 -5.35350204e-01 -3.83049726e-01
-6.96107864e-01 1.41613472e+00 3.00673664e-01 6.92822099e-01
-6.18725896e-01 -2.75937021e-01 6.53488278e-01 -7.07649142e-02
1.61831886e-01 -5.09106278e-01 -3.52523208e-01 -1.12525284e+00
-7.10348785e-02 -6.88697755e-01 3.09394181e-01 -9.15488780e-01
-1.02201927e+00 7.48364627e-01 2.04733178e-01 -1.11477780e+00
-2.14999348e-01 -7.39500523e-01 -6.53024018e-01 1.10135627e+00
-1.57002115e+00 -1.34201956e+00 -3.31537962e-01 9.77081418e-01
3.27401042e-01 -1.08222507e-01 7.97795177e-01 -7.63458107e-03
-2.44914085e-01 7.30761170e-01 -3.62340420e-01 5.55146337e-02
9.21144247e-01 -1.20546925e+00 5.30592799e-01 1.16222119e+00
4.11042154e-01 9.33404446e-01 5.37288964e-01 -5.65693080e-01
-1.57677901e+00 -9.72606361e-01 6.88327610e-01 -3.39539766e-01
3.22531253e-01 -3.10880154e-01 -8.44967484e-01 9.40581560e-01
4.92215753e-01 2.55814850e-01 5.25128245e-01 -3.22318554e-01
-6.06912196e-01 5.69014437e-02 -9.69964504e-01 7.04847634e-01
1.17612195e+00 -7.62929142e-01 -5.23134828e-01 8.90547931e-02
7.73897588e-01 -6.86854482e-01 -3.57650280e-01 4.56738144e-01
5.12145042e-01 -6.34504855e-01 8.39125872e-01 -9.05973673e-01
4.64217007e-01 -5.60769022e-01 -4.86050427e-01 -1.58849478e+00
-5.69741845e-01 -8.16668868e-01 9.84245092e-02 9.85285878e-01
5.96005678e-01 -5.76526463e-01 5.78425527e-01 9.56307709e-01
-2.77953595e-01 -8.03239405e-01 -7.07464874e-01 -6.19163096e-01
-9.22775269e-03 -2.87989020e-01 4.88934904e-01 6.94144785e-01
2.42726021e-02 4.16672349e-01 -5.19121528e-01 2.64166091e-02
4.53698248e-01 2.41946578e-01 9.02393758e-01 -5.75243354e-01
-3.40975016e-01 -3.52007538e-01 -2.88105875e-01 -1.51953912e+00
7.62803480e-03 -9.28065717e-01 1.55960798e-01 -1.55004466e+00
3.89927775e-01 -2.04096019e-01 -3.67936522e-01 8.99004042e-01
-3.96567672e-01 5.13694346e-01 5.37439942e-01 2.01801106e-01
-2.17213601e-01 5.55956125e-01 1.85661936e+00 -1.45710364e-01
-1.56006247e-01 -1.47041261e-01 -9.13136125e-01 5.11027217e-01
6.22794509e-01 -3.74183387e-01 -5.28867960e-01 -7.90285170e-01
-5.66125028e-02 -1.37103960e-01 9.86190438e-02 -5.75518608e-01
4.26951766e-01 -3.28281432e-01 2.94560671e-01 -4.68719691e-01
3.37372094e-01 -6.16780102e-01 -1.35330975e-01 2.72461593e-01
-5.46694934e-01 2.95888782e-01 8.12255144e-02 4.15887147e-01
-4.23476964e-01 6.14851387e-03 7.78989613e-01 -5.25158271e-02
-5.70352435e-01 5.74156821e-01 -1.42837331e-01 -1.57727674e-01
7.47448087e-01 -2.03702495e-01 -3.64261538e-01 -5.57074249e-01
-7.60257125e-01 2.37619817e-01 4.31886017e-01 4.91014749e-01
9.25814629e-01 -1.51510072e+00 -8.46032917e-01 5.73132813e-01
2.02404827e-01 -9.61344466e-02 -2.41248563e-01 9.34872806e-01
6.37200326e-02 5.94803870e-01 -3.35746467e-01 -6.35448277e-01
-1.26057351e+00 6.45319402e-01 2.52999395e-01 -1.81950465e-01
-6.50787890e-01 7.59641707e-01 9.27819848e-01 -5.04003286e-01
1.87475190e-01 -4.25259501e-01 1.52811781e-01 -3.64393800e-01
5.12463570e-01 -1.82989523e-01 -1.08594790e-01 -8.07196140e-01
-1.94348335e-01 1.01045418e+00 -2.40858749e-01 -6.26611114e-01
9.81817305e-01 -1.50903955e-01 3.40787023e-02 2.05060199e-01
1.27708852e+00 -1.69213191e-01 -1.62342238e+00 -6.25494301e-01
-4.08393770e-01 -9.44459677e-01 -1.01315416e-02 -8.53937924e-01
-1.25948095e+00 9.61112976e-01 3.45343202e-01 -2.65398085e-01
1.34463346e+00 -1.62800416e-01 2.80361831e-01 3.61107528e-01
2.44664058e-01 -8.56821239e-01 4.23283249e-01 5.24258971e-01
1.34145665e+00 -1.09616458e+00 1.25155985e-01 -5.50402641e-01
-8.82719278e-01 9.20972466e-01 8.24107647e-01 -1.89869836e-01
6.16090834e-01 1.90046266e-01 2.28899181e-01 -5.48944920e-02
-7.42442727e-01 -2.95655906e-01 5.90755045e-01 7.65855610e-01
4.44139332e-01 -1.01017311e-01 -3.57390225e-01 3.53074729e-01
-5.84872365e-02 1.44409612e-01 2.46885821e-01 9.59982574e-01
-1.47191688e-01 -1.18913281e+00 -4.05786395e-01 3.17739695e-01
-3.02260101e-01 -2.53403902e-01 -5.26413977e-01 8.01153243e-01
-2.02835537e-02 7.26134002e-01 -1.04442030e-01 -3.47714156e-01
5.82658231e-01 3.36009488e-02 7.04893768e-01 -7.88754523e-01
-3.67503971e-01 4.15447503e-01 -1.71499848e-01 -8.31841528e-01
-6.23423278e-01 -6.83130026e-01 -1.26425397e+00 1.60727739e-01
-5.95125675e-01 -1.92725554e-01 7.27558732e-01 9.16451335e-01
4.50944334e-01 3.49528432e-01 6.76730692e-01 -1.19836617e+00
-2.46218547e-01 -9.43475962e-01 -2.45231315e-01 5.88757098e-01
4.62907732e-01 -7.08865941e-01 -2.16762394e-01 3.05568039e-01] | [11.484156608581543, -0.5198706388473511] |
7a428192-c4a8-483d-9c8d-24dae7d8c690 | improving-cloze-test-performance-of-language | null | null | https://aclanthology.org/C14-1091 | https://aclanthology.org/C14-1091.pdf | Improving Cloze Test Performance of Language Learners Using Web N-Grams | null | ['Anna Beyer', 'Benno Stein', 'Matthias Hagen', 'Martin Potthast'] | 2014-08-01 | improving-cloze-test-performance-of-language-1 | https://aclanthology.org/C14-1091 | https://aclanthology.org/C14-1091.pdf | coling-2014-8 | ['cloze-test'] | ['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] | [-7.267706871032715, 3.652627468109131] |
ac9a932b-77d0-4feb-8c70-b0baaa0692f6 | ontology-aware-token-embeddings-for | 1705.02925 | null | http://arxiv.org/abs/1705.02925v1 | http://arxiv.org/pdf/1705.02925v1.pdf | Ontology-Aware Token Embeddings for Prepositional Phrase Attachment | Type-level word embeddings use the same set of parameters to represent all
instances of a word regardless of its context, ignoring the inherent lexical
ambiguity in language. Instead, we embed semantic concepts (or synsets) as
defined in WordNet and represent a word token in a particular context by
estimating a distribution over relevant semantic concepts. We use the new,
context-sensitive embeddings in a model for predicting prepositional phrase(PP)
attachments and jointly learn the concept embeddings and model parameters. We
show that using context-sensitive embeddings improves the accuracy of the PP
attachment model by 5.4% absolute points, which amounts to a 34.4% relative
reduction in errors. | ['Waleed Ammar', 'Pradeep Dasigi', 'Chris Dyer', 'Eduard Hovy'] | 2017-05-08 | ontology-aware-token-embeddings-for-1 | https://aclanthology.org/P17-1191 | https://aclanthology.org/P17-1191.pdf | acl-2017-7 | ['prepositional-phrase-attachment'] | ['natural-language-processing'] | [-1.83323443e-01 1.03360474e-01 -4.92912799e-01 -5.18522859e-01
-4.98850584e-01 -6.26010120e-01 5.41056335e-01 7.88494468e-01
-9.72445965e-01 5.57465911e-01 6.41791403e-01 -2.61007309e-01
2.01771289e-01 -1.01278663e+00 -5.02623260e-01 -4.18752670e-01
-1.94367245e-01 4.75472450e-01 1.83522508e-01 -2.66327947e-01
2.11618483e-01 1.90988824e-01 -1.12330770e+00 1.75538838e-01
3.38916004e-01 8.23469162e-01 1.82171315e-01 3.35720569e-01
-6.10458732e-01 1.50773570e-01 -5.71409881e-01 -5.82641006e-01
-2.19124466e-01 2.72729039e-01 -8.43761206e-01 -4.12329972e-01
1.57795042e-01 1.07601121e-01 -5.97181395e-02 9.42954719e-01
7.43836462e-02 3.31180096e-01 8.22282195e-01 -9.87373948e-01
-1.43258393e+00 6.60424352e-01 -2.43550241e-01 4.96462643e-01
5.07011235e-01 -4.36618328e-01 1.84806061e+00 -1.43775237e+00
4.81885433e-01 1.26266277e+00 7.95331180e-01 5.91275394e-01
-1.37952530e+00 -3.45903337e-01 3.43643785e-01 2.13618338e-01
-1.46308863e+00 -1.23556264e-01 5.45444489e-01 -3.26342106e-01
1.71394765e+00 -2.31637105e-01 7.53587902e-01 1.30781698e+00
3.65845382e-01 2.51078814e-01 5.79895914e-01 -6.95915520e-01
4.28807735e-01 2.70678010e-02 6.44547522e-01 2.60619938e-01
7.33319581e-01 -8.80853385e-02 -3.64893168e-01 -4.45676088e-01
4.04692382e-01 4.82740998e-02 1.59998342e-01 -1.64533360e-03
-9.07807648e-01 1.28194189e+00 4.95294213e-01 3.93143475e-01
-2.89129496e-01 5.05081058e-01 3.46699625e-01 -1.13762826e-01
5.13229609e-01 7.14573085e-01 -8.81787062e-01 -2.01237146e-02
-1.07010834e-01 4.27901506e-01 7.99895167e-01 9.17672575e-01
9.11495447e-01 -5.09035029e-02 2.64900029e-01 1.21572185e+00
6.98718190e-01 3.21135759e-01 8.06354403e-01 -6.81079149e-01
1.39225751e-01 4.28195596e-01 1.02285869e-01 -9.37771559e-01
-2.67694801e-01 -9.90082547e-02 1.91640317e-01 -4.84973222e-01
2.79650874e-02 -1.49741203e-01 -7.96499550e-01 2.17729163e+00
1.75928399e-01 2.32388049e-01 2.75595844e-01 5.02461314e-01
5.91978312e-01 9.15127933e-01 9.35450315e-01 2.57867854e-02
1.94526374e+00 -5.16399682e-01 -6.50003374e-01 -8.89226079e-01
7.68182755e-01 -7.20173717e-01 1.21823323e+00 -3.62468213e-01
-7.33710766e-01 -2.08808154e-01 -1.19745648e+00 -3.56498718e-01
-9.32931244e-01 -6.27407670e-01 6.91052735e-01 6.07292593e-01
-8.96763682e-01 1.18240856e-01 -5.52363396e-01 -5.75066864e-01
1.01897500e-01 3.18535686e-01 -4.31021035e-01 -1.21192984e-01
-1.69927633e+00 1.26323926e+00 7.31727839e-01 -5.68894207e-01
-2.63944596e-01 -1.02454042e+00 -1.33571744e+00 2.83275992e-01
-7.64914602e-02 -5.94996631e-01 1.04672837e+00 -6.47163391e-01
-7.49881744e-01 1.03071582e+00 -4.31062728e-01 -3.84580076e-01
-7.61825323e-01 -3.07821214e-01 -8.83507490e-01 4.75751311e-02
3.78885806e-01 7.94986367e-01 4.43854839e-01 -1.13668561e+00
-5.76816976e-01 -1.99845314e-01 2.64358968e-01 2.18488455e-01
-7.88701534e-01 3.17152202e-01 -4.80616093e-02 -8.69199753e-01
1.06907226e-01 -6.70197248e-01 -1.30852327e-01 -3.46564292e-03
2.29303554e-01 -6.33343577e-01 5.04270136e-01 -3.83790195e-01
1.32958758e+00 -2.21770000e+00 -1.62501961e-01 1.08967768e-02
4.27638181e-02 7.92290494e-02 -2.98579544e-01 6.13363028e-01
-2.27000162e-01 5.22855341e-01 -3.68234754e-01 -4.56988156e-01
2.06615686e-01 9.23551023e-01 -4.29250836e-01 -7.72040104e-03
4.98873085e-01 8.35538089e-01 -1.17069674e+00 -3.27128232e-01
2.66076457e-02 5.45440435e-01 -9.15632129e-01 -6.28075823e-02
-2.51139402e-01 -6.79943264e-01 -3.06556314e-01 3.95740718e-01
4.32739347e-01 8.72983262e-02 6.21201873e-01 -3.96041162e-02
1.91197708e-01 8.70406866e-01 -7.96706378e-01 1.57656050e+00
-9.70943689e-01 3.97732586e-01 -4.18130368e-01 -8.53240430e-01
9.42760825e-01 4.93416905e-01 2.57010907e-01 -3.58452171e-01
-8.04150179e-02 3.19894820e-01 -2.13234369e-02 -1.91533327e-01
8.04222822e-01 -7.52153397e-01 -5.64597428e-01 3.93152177e-01
4.18244690e-01 1.53447583e-01 -7.87037367e-04 6.89645410e-02
9.08839226e-01 -2.23403379e-01 6.51286244e-01 -6.62921906e-01
2.13007271e-01 -1.67734131e-01 7.86735892e-01 2.09460646e-01
-1.50888100e-01 3.51177961e-01 4.82718021e-01 -5.60490787e-01
-9.82198656e-01 -1.59729695e+00 -6.89912677e-01 1.34224594e+00
-8.28649253e-02 -9.57015574e-01 -3.63962173e-01 -6.39228582e-01
3.33109558e-01 1.16298783e+00 -7.63539433e-01 -2.68849373e-01
-6.20973229e-01 -9.29133832e-01 3.59341353e-01 1.09130943e+00
-3.46456200e-01 -9.45150793e-01 -2.47828886e-01 4.13194895e-01
1.74894389e-02 -1.17570460e+00 -3.76496255e-01 3.31812739e-01
-6.12774730e-01 -8.00514519e-01 -1.09407201e-01 -1.11439180e+00
5.80196857e-01 -9.60243717e-02 1.38036394e+00 6.85829669e-02
9.59338900e-03 3.86385411e-01 -6.37897134e-01 -2.76679277e-01
-1.27530232e-01 -3.14811617e-02 4.62465823e-01 -3.53072673e-01
1.37600088e+00 -5.95843792e-01 -3.03928941e-01 -1.95648864e-01
-8.13564360e-01 -8.55700791e-01 -6.08540624e-02 1.06018913e+00
4.24845278e-01 -5.27795672e-01 7.76817203e-01 -7.03438222e-01
7.47050703e-01 -9.81331944e-01 -1.84939712e-01 1.37064746e-02
-6.54556036e-01 1.13867700e-01 3.82186085e-01 -3.60013008e-01
-8.00530434e-01 -3.34638655e-01 -4.79990274e-01 -2.58333422e-02
-3.34946774e-02 6.78701460e-01 -1.35809973e-01 4.88641441e-01
5.48349857e-01 -1.53539777e-01 -4.11688298e-01 -4.76744980e-01
6.64327502e-01 6.34357035e-01 2.48756945e-01 -1.09518814e+00
4.51507181e-01 1.26808956e-02 -3.36101234e-01 -8.20651889e-01
-1.12484622e+00 -4.76749271e-01 -5.89725971e-01 5.68905175e-01
1.21411693e+00 -9.20166790e-01 -3.03127766e-01 -3.21557939e-01
-1.42549884e+00 1.54467374e-01 -5.06868243e-01 5.11987984e-01
-2.25837797e-01 2.51667619e-01 -6.19305134e-01 -3.98230076e-01
-1.43955559e-01 -6.42887831e-01 6.76801860e-01 -5.97308017e-02
-7.57043660e-01 -1.61817467e+00 2.06791043e-01 -2.20657483e-01
2.40533873e-01 -2.09320724e-01 1.38219678e+00 -1.15552318e+00
7.24836588e-02 -1.58019274e-01 -1.41477272e-01 4.99052882e-01
3.39970827e-01 -8.54109451e-02 -9.38262403e-01 -5.93441247e-04
-4.23871353e-02 4.14686352e-02 7.29039788e-01 1.55413553e-01
8.11250806e-01 -3.78638655e-01 -4.33912337e-01 3.87691140e-01
1.80045164e+00 1.29360616e-01 3.35691899e-01 2.66155601e-01
4.53290403e-01 4.81549233e-01 3.46521735e-01 2.68157333e-01
4.99781191e-01 5.31202674e-01 1.02827720e-01 6.64828479e-01
8.56489092e-02 -6.36654377e-01 3.00921708e-01 9.22091663e-01
4.09657747e-01 -2.45421365e-01 -1.16830719e+00 1.19794238e+00
-1.50424814e+00 -7.09248304e-01 1.43281594e-01 1.94023097e+00
1.06275523e+00 2.27473497e-01 -2.36690879e-01 -1.39364988e-01
6.97670043e-01 2.75283366e-01 2.18067095e-02 -1.04940569e+00
3.43606062e-03 8.26876223e-01 3.78543645e-01 1.02887368e+00
-8.51009905e-01 1.35712254e+00 7.61159563e+00 5.08335531e-01
-6.40840232e-01 3.56063187e-01 8.40219930e-02 1.19470946e-01
-7.55526185e-01 2.61396915e-01 -1.02793467e+00 5.40356100e-01
1.22742331e+00 -4.68905360e-01 1.25116082e-02 9.27078664e-01
-3.93422008e-01 2.23139673e-01 -1.19428420e+00 5.41666389e-01
1.98484913e-01 -1.25837171e+00 4.03800577e-01 -1.76061869e-01
4.75826502e-01 -8.88439566e-02 -1.18511217e-02 3.50206673e-01
4.23891932e-01 -1.00292230e+00 4.62082654e-01 9.04770419e-02
6.08119428e-01 -8.68745685e-01 6.88681960e-01 -2.23372221e-01
-1.24467325e+00 -1.99736744e-01 -8.32703114e-01 -4.44087714e-01
3.06614548e-01 4.64046448e-01 -7.51982570e-01 -1.62361339e-02
5.48461378e-01 7.66247272e-01 -2.14130759e-01 3.59025061e-01
-7.22258270e-01 7.25959003e-01 -3.54308248e-01 -2.47175485e-01
3.30253869e-01 8.26111808e-02 4.61520374e-01 1.49478912e+00
1.62131235e-01 2.64474362e-01 7.98251331e-02 9.39781487e-01
-2.72774369e-01 9.96817723e-02 -4.49909359e-01 -9.28047597e-02
1.14929438e+00 1.03859937e+00 -2.64594972e-01 -3.47569704e-01
-7.57735014e-01 7.53983796e-01 5.69054961e-01 2.83849627e-01
-5.16911387e-01 -5.31706750e-01 1.59706664e+00 5.58506586e-02
3.91313672e-01 -4.69651669e-01 -3.42225969e-01 -1.00113177e+00
3.62690948e-02 6.94997748e-03 5.09265542e-01 -6.46799147e-01
-1.94256282e+00 3.21719319e-01 1.91413701e-01 -6.22219026e-01
-3.37812513e-01 -1.16526234e+00 -7.40881443e-01 1.08139682e+00
-1.56730974e+00 -9.99365509e-01 4.51006413e-01 1.37751579e-01
3.80762368e-01 1.88260190e-02 1.58135283e+00 1.12425193e-01
-2.15410337e-01 6.89876497e-01 -1.20961294e-01 3.89447004e-01
6.85523808e-01 -1.52734482e+00 8.01512659e-01 4.92027402e-01
2.80961543e-01 1.33314466e+00 6.35371327e-01 -4.26054120e-01
-9.66856480e-01 -9.59060311e-01 1.95518243e+00 -8.39794040e-01
1.11418915e+00 -6.00323617e-01 -1.11102045e+00 1.16594744e+00
2.94352680e-01 3.59414637e-01 1.38385677e+00 7.57861853e-01
-1.02038574e+00 -5.89286573e-02 -1.18041968e+00 6.50516450e-01
9.71414149e-01 -7.57078290e-01 -1.64233148e+00 1.47535279e-01
1.36893010e+00 2.25570947e-01 -1.17043638e+00 -2.21482851e-02
4.17686433e-01 -2.38910988e-01 1.33057237e+00 -1.06572175e+00
4.58470702e-01 -4.35607368e-03 -9.44086909e-01 -1.52828538e+00
-6.02167428e-01 1.37931809e-01 -1.43140152e-01 1.20919490e+00
5.99716306e-01 -9.68156338e-01 3.53375137e-01 7.10294068e-01
-2.32135937e-01 -8.11678767e-01 -1.07488501e+00 -7.66259372e-01
7.93592513e-01 -5.47811568e-01 7.59677112e-01 1.18551385e+00
5.50937295e-01 6.00755394e-01 3.61247808e-01 2.13046834e-01
3.42483521e-01 -2.42661417e-01 -3.40016961e-01 -1.25994921e+00
-1.24410979e-01 -2.31785536e-01 -7.45362103e-01 -8.18499029e-01
7.45538592e-01 -1.06724012e+00 -2.13985965e-01 -1.45698023e+00
-1.04706988e-01 -7.34503329e-01 -8.55466664e-01 5.66274881e-01
-3.31705540e-01 3.01152200e-01 1.09759226e-01 -2.63762385e-01
-9.11585018e-02 4.78792846e-01 3.88044775e-01 2.42884401e-02
3.10117006e-01 -6.65591180e-01 -1.10684383e+00 7.77455628e-01
8.13956797e-01 -6.98280752e-01 -4.55478430e-01 -8.59514058e-01
5.37369907e-01 -4.64030027e-01 2.78846264e-01 -4.96383458e-01
-5.87750822e-02 -2.53969014e-01 3.62557113e-01 -1.40812364e-03
6.97683513e-01 -7.06690967e-01 -4.47193295e-01 2.61311412e-01
-3.89638484e-01 5.02080321e-01 3.77874583e-01 5.95758855e-01
-2.46830940e-01 -6.38517261e-01 5.88034213e-01 -5.02306372e-02
-1.06314266e+00 4.49137650e-02 -5.34371853e-01 4.98783082e-01
9.24060941e-01 -8.58254805e-02 -1.15175076e-01 2.33253554e-01
-9.00900781e-01 -8.90678912e-02 5.84234476e-01 8.21761966e-01
6.62599623e-01 -1.60309768e+00 -4.92218852e-01 2.42659375e-01
6.20866239e-01 -3.83813024e-01 -1.76320493e-01 -1.52830526e-01
-2.30419680e-01 2.72105694e-01 -8.97034034e-02 -1.03684165e-01
-6.72530890e-01 6.41763151e-01 1.67811543e-01 2.38703400e-01
-4.48919803e-01 1.08713663e+00 6.51979372e-02 -5.17153800e-01
-3.20219457e-01 -4.81644690e-01 -1.54046685e-01 2.40105286e-01
4.44314986e-01 -1.50797486e-01 -3.39942910e-02 -7.24978745e-01
-8.38727951e-01 5.72450936e-01 -2.01244414e-01 -2.04493701e-01
1.24299765e+00 8.65079910e-02 -1.97458163e-01 7.10155845e-01
1.51528752e+00 7.45277479e-02 -6.99096859e-01 -3.38032305e-01
3.22933674e-01 -5.38925052e-01 -5.35802543e-02 -6.17929637e-01
-5.18961787e-01 9.40587640e-01 9.79244038e-02 -3.43712382e-02
4.59139496e-01 2.98953533e-01 9.95431721e-01 1.75214365e-01
3.08786392e-01 -1.04452312e+00 -1.69625178e-01 7.90980101e-01
5.83662271e-01 -7.55843818e-01 -2.03494936e-01 -4.34571266e-01
-4.88656312e-01 9.67830896e-01 4.47664589e-01 -6.15681231e-01
1.07038140e+00 3.59965533e-01 -3.17123532e-02 -9.11959186e-02
-1.01960409e+00 4.80498709e-02 -2.69233603e-02 9.12946284e-01
7.91240335e-01 4.25335228e-01 -6.44304156e-01 1.00580060e+00
-2.31398374e-01 -6.23855114e-01 4.74662125e-01 9.00455832e-01
-6.93240821e-01 -1.46456027e+00 -1.78606845e-02 3.11453372e-01
-4.88595009e-01 -6.70629025e-01 4.05284129e-02 5.86562455e-01
1.44910991e-01 8.99140656e-01 7.16784537e-01 -1.86593622e-01
2.83761472e-01 6.91593230e-01 2.80993640e-01 -1.13489544e+00
-4.28611070e-01 -5.31354427e-01 2.69527912e-01 -1.80172876e-01
-1.99859321e-01 -6.80243194e-01 -1.61489761e+00 -2.86246151e-01
1.64829921e-02 2.85106301e-01 6.35926664e-01 8.95739794e-01
3.83398652e-01 1.72467425e-01 2.64661551e-01 -2.80903369e-01
-5.71565509e-01 -8.36306810e-01 -4.33544368e-01 5.45846224e-01
1.63458765e-01 -9.33810830e-01 -4.39334124e-01 -1.87353775e-01] | [10.413700103759766, 8.92280101776123] |
4c59d790-0ab8-469f-bd30-a2fb5c379832 | mean-absorption-estimation-from-room-impulse | 2109.00393 | null | https://arxiv.org/abs/2109.00393v1 | https://arxiv.org/pdf/2109.00393v1.pdf | Mean absorption estimation from room impulse responses using virtually supervised learning | In the context of building acoustics and the acoustic diagnosis of an existing room, this paper introduces and investigates a new approach to estimate mean absorption coefficients solely from a room impulse response (RIR). This inverse problem is tackled via virtually-supervised learning, namely, the RIR-to-absorption mapping is implicitly learned by regression on a simulated dataset using artificial neural networks. We focus on simple models based on well-understood architectures. The critical choices of geometric, acoustic and simulation parameters used to train the models are extensively discussed and studied, while keeping in mind conditions that are representative of the field of building acoustics. Estimation errors from the learned neural models are compared to those obtained with classical formulas that require knowledge of the room's geometry and reverberation times. Extensive comparisons made on a variety of simulated test sets highlight different conditions under which the learned models can overcome the well-known limitations of the diffuse sound field hypothesis underlying these formulas. Results obtained on real RIRs measured in an acoustically configurable room show that at 1~kHz and above, the proposed approach performs comparably to classical models when reverberation times can be reliably estimated, and continues to work even when they cannot. | ['Diego Di Carlo', 'Antoine Deleforge', 'Cédric Foy'] | 2021-09-01 | null | null | null | null | ['room-impulse-response'] | ['audio'] | [ 5.56942403e-01 -7.79066095e-03 9.71722126e-01 -4.69808996e-01
-1.00955224e+00 -1.91415906e-01 8.58919844e-02 2.30184644e-02
-4.76944536e-01 6.58247828e-01 1.71703756e-01 -3.90074193e-01
-5.38239002e-01 -6.75031185e-01 -5.59846282e-01 -1.15861702e+00
-3.92478853e-01 1.03721701e-01 -2.60863602e-01 -4.24430341e-01
-1.10999100e-01 4.11075741e-01 -1.86771703e+00 -3.60626280e-02
5.97366393e-01 1.09397531e+00 3.59296113e-01 1.21043134e+00
1.95953935e-01 8.03159297e-01 -1.00845313e+00 4.57938880e-01
2.03648299e-01 -4.11103487e-01 -5.21183670e-01 -2.64877588e-01
4.47338074e-01 -2.97339022e-01 -2.29875833e-01 6.81549430e-01
9.12804723e-01 5.25006354e-01 7.53344357e-01 -3.76091540e-01
-4.77265745e-01 4.94633526e-01 2.37861142e-01 2.10167393e-01
3.82445097e-01 -1.46368593e-01 7.01939940e-01 -5.92166424e-01
-3.67551267e-01 7.62078941e-01 1.01275325e+00 4.64463949e-01
-1.22524035e+00 -3.53593290e-01 -2.31922820e-01 2.20297873e-02
-1.49524200e+00 -5.67642689e-01 1.00987291e+00 -2.91929513e-01
9.28334177e-01 6.76714182e-01 2.00447395e-01 7.29358196e-01
-2.24670500e-01 4.53729108e-02 1.44788694e+00 -1.09805965e+00
4.41558778e-01 5.35192311e-01 1.72302470e-01 5.85727930e-01
-2.52859056e-01 3.34737390e-01 -4.89891507e-02 -5.12827635e-02
4.84676152e-01 -5.77437520e-01 -8.05249453e-01 7.35720694e-02
-9.87677872e-01 5.12203693e-01 5.00651479e-01 6.14913881e-01
-1.41055062e-01 4.68767658e-02 1.64387934e-02 3.56612980e-01
4.55592215e-01 5.49659491e-01 -3.22336853e-01 2.30862364e-01
-6.53650105e-01 -3.76812294e-02 1.16556919e+00 7.86629379e-01
6.09057784e-01 5.70261180e-01 2.22882375e-01 1.28962064e+00
5.49106419e-01 1.00257885e+00 7.89920334e-03 -8.04734707e-01
2.35478222e-01 -4.76075619e-01 2.70512193e-01 -7.14971364e-01
-5.53052843e-01 -5.90020120e-01 -6.85072601e-01 2.99453169e-01
6.09714627e-01 -1.62110001e-01 -9.17986214e-01 1.46714556e+00
2.49835178e-01 3.13304782e-01 2.69466013e-01 7.13377416e-01
1.06033921e+00 9.34009790e-01 -3.39155972e-01 -4.81310278e-01
7.91572988e-01 -6.71689153e-01 -9.06658530e-01 -2.14263216e-01
3.85447592e-01 -8.19860220e-01 1.12830091e+00 6.42971337e-01
-1.06971800e+00 -8.39801550e-01 -1.23711860e+00 4.27642643e-01
-2.09605664e-01 -1.49283931e-01 -1.20703084e-02 1.37532020e+00
-1.13122058e+00 5.11784852e-01 -4.63793665e-01 2.17807025e-01
-3.89192879e-01 2.94366449e-01 8.11967701e-02 3.32930461e-02
-1.24512231e+00 7.89092958e-01 -2.72597402e-01 9.08518076e-01
-9.83186066e-01 -8.71017039e-01 -6.96428478e-01 -2.80505382e-02
1.46412820e-01 -4.06722069e-01 1.55348110e+00 -7.23109066e-01
-1.91893673e+00 3.53993028e-01 6.78539351e-02 -1.38345048e-01
1.58286944e-01 -4.40869927e-01 -9.09565985e-01 1.23973146e-01
-5.28623164e-01 -5.93598306e-01 7.39803314e-01 -1.92783785e+00
2.55211350e-02 1.64434776e-01 3.25514302e-02 -1.13366604e-01
-5.72219212e-03 -1.15475707e-01 4.03774977e-01 -2.44753420e-01
9.64221284e-02 -6.62233710e-01 -3.35135877e-01 -6.72396541e-01
-4.68493998e-01 2.83331186e-01 2.21222445e-01 -9.14113343e-01
1.09331894e+00 -1.98525703e+00 -2.22412214e-01 8.16044509e-01
-1.18515722e-01 1.43016517e-01 1.44158332e-02 6.29072905e-01
-4.92406487e-01 -2.14658007e-01 -6.33069217e-01 -2.17275053e-01
-8.73510912e-02 -1.71514377e-01 -3.19760948e-01 6.35124922e-01
-3.63015354e-01 -5.10190427e-03 -5.50385654e-01 -4.73695882e-02
4.91189867e-01 8.91368210e-01 -3.87546271e-01 6.68920040e-01
1.93571210e-01 7.41538584e-01 -3.62683237e-01 2.45743275e-01
7.26537108e-01 3.08218986e-01 -1.20433219e-01 -2.63786525e-01
-3.83921504e-01 2.92852402e-01 -1.27275109e+00 1.09436810e+00
-1.49588728e+00 6.09713435e-01 6.58977270e-01 -1.12050676e+00
1.22249591e+00 4.92549092e-01 1.49485171e-01 -7.90368080e-01
2.13816047e-01 3.48673314e-01 1.23100579e-01 -7.70005524e-01
8.43469705e-03 -5.50291419e-01 3.50674719e-01 5.49655855e-01
-2.68689264e-02 -5.91727376e-01 -6.08556032e-01 -3.65527779e-01
1.01977742e+00 -1.11554466e-01 1.41491545e-02 -5.12185633e-01
1.06770551e+00 -4.88524020e-01 -1.73696205e-01 9.36993659e-01
5.34790531e-02 5.70666432e-01 -3.79348129e-01 -3.76325935e-01
-8.39545250e-01 -1.13518631e+00 -4.40032303e-01 1.02450609e+00
-2.14517653e-01 1.56702861e-01 -1.10229790e+00 2.00622201e-01
-5.62761486e-01 1.33018208e+00 -6.35206044e-01 -1.42435534e-02
-1.02215004e+00 -8.84598017e-01 4.80888516e-01 1.89680874e-01
2.08818719e-01 -1.05586374e+00 -5.45131624e-01 3.17628384e-01
-3.80087823e-01 -1.17012680e+00 3.17066401e-01 6.48946941e-01
-4.23871756e-01 -6.36691928e-01 -5.59407175e-01 -6.37524903e-01
5.26126742e-01 2.71626711e-01 1.27956116e+00 2.62151212e-01
-6.93512201e-01 1.37662482e+00 -3.32942456e-01 -7.37777650e-01
-1.05427027e+00 -4.69590038e-01 1.70711920e-01 6.13922440e-02
-3.76770310e-02 -8.41840744e-01 -6.94738448e-01 5.23045540e-01
-6.93546474e-01 -5.06017148e-01 3.04237455e-01 4.72002059e-01
3.20703566e-01 3.11013669e-01 6.64354265e-01 -5.70337415e-01
4.92544651e-01 -1.38479665e-01 -5.28659642e-01 1.36799350e-01
-4.21144962e-01 -9.69688818e-02 7.78930962e-01 -2.34168246e-01
-1.67104006e+00 -3.16214204e-01 -6.59535646e-01 2.56813645e-01
-8.34296227e-01 9.25558507e-02 -4.52838928e-01 -2.73010254e-01
9.54841316e-01 2.43336082e-01 -2.20800862e-01 -4.18419898e-01
7.51343966e-02 6.64193511e-01 5.81883311e-01 -7.19844818e-01
1.23455846e+00 3.33226591e-01 4.97527272e-02 -1.60118580e+00
-7.91371465e-01 -5.55959344e-01 -5.77317834e-01 -4.70459044e-01
9.21781301e-01 -8.24062884e-01 -6.50599599e-01 4.64520901e-01
-1.00898266e+00 -7.87244976e-01 -3.75803053e-01 9.42319751e-01
-7.95410693e-01 1.77591786e-01 -4.78149503e-01 -1.59007657e+00
-1.62455171e-01 -7.76965261e-01 5.75733304e-01 -6.93125799e-02
9.05150920e-02 -1.47782838e+00 4.46941853e-01 4.90900606e-01
8.01865578e-01 1.22961812e-01 1.00699914e+00 -3.27080786e-01
-5.18978775e-01 -2.08212078e-01 3.01514834e-01 9.19702470e-01
2.75031775e-01 -2.25192800e-01 -2.03679657e+00 -1.49844483e-01
8.72858763e-01 -2.07955301e-01 7.38814294e-01 7.15435982e-01
1.30920839e+00 -1.06836088e-01 1.30024552e-01 4.72703964e-01
1.65386415e+00 1.83770850e-01 7.58845329e-01 3.98803324e-01
4.62024361e-01 9.38849628e-01 2.48362452e-01 3.95264983e-01
-1.82650238e-01 2.13786870e-01 7.47676849e-01 -3.48017514e-01
-9.27310735e-02 2.71931976e-01 1.66103035e-01 1.23077440e+00
-6.96633816e-01 -2.29862571e-01 -6.48239434e-01 2.13243902e-01
-7.82243013e-01 -9.64930415e-01 -2.41626665e-01 2.66430068e+00
5.54285407e-01 -2.69684270e-02 -3.46606195e-01 7.22507954e-01
4.53895599e-01 2.23773897e-01 1.67095080e-01 -6.84461653e-01
-1.72855135e-03 6.79805577e-01 3.83721143e-01 1.14457822e+00
-8.47497880e-01 -5.12070321e-02 7.34139490e+00 2.91571856e-01
-9.30828273e-01 7.61670694e-02 6.04333699e-01 1.19237088e-01
-4.19239372e-01 -5.58457971e-01 -5.22403061e-01 -2.26461872e-01
1.50718486e+00 3.78449559e-01 7.65293658e-01 5.59137166e-01
3.99023175e-01 -1.29813045e-01 -1.19171739e+00 7.00472832e-01
2.92922467e-01 -6.04658961e-01 -6.64619744e-01 -1.75089106e-01
5.01469314e-01 -1.91272095e-01 2.60282785e-01 2.79752463e-01
-1.19779885e-01 -1.35871220e+00 4.66881961e-01 6.95770741e-01
5.71598589e-01 -7.04538107e-01 6.88391864e-01 3.70852917e-01
-1.01685214e+00 -1.35571748e-01 -3.67416143e-01 6.12768997e-03
1.77944139e-01 7.91890442e-01 -1.17282450e+00 6.06181860e-01
8.44584703e-01 -2.43005708e-01 -1.85733706e-01 1.05720973e+00
-8.81254226e-02 1.18768263e+00 -5.28926373e-01 -3.57546471e-02
1.19944111e-01 -2.09452033e-01 3.77936333e-01 1.44062662e+00
5.73935926e-01 2.39639595e-01 -3.57267052e-01 9.51275229e-01
4.72280264e-01 2.41686195e-01 -5.63355505e-01 5.56972146e-01
-3.83950137e-02 1.27867532e+00 -4.53993887e-01 1.35090962e-01
-4.26397651e-01 4.05629158e-01 -2.68333763e-01 8.10934186e-01
-9.72030699e-01 -5.27805090e-01 7.40214139e-02 2.81561404e-01
2.71992564e-01 -2.31730267e-01 5.20107113e-02 -2.13035077e-01
-1.99426979e-01 -5.25604367e-01 -2.31920063e-01 -7.22262025e-01
-1.08009267e+00 9.38948631e-01 5.00218928e-01 -1.10916364e+00
-1.49764895e-01 -8.62215638e-01 -9.75299358e-01 1.09683537e+00
-1.64837921e+00 -6.88394070e-01 -5.71541667e-01 6.00373805e-01
3.77753407e-01 7.73159862e-02 1.24761486e+00 4.06914800e-01
-1.60188496e-01 3.07783544e-01 4.95728076e-01 -1.20668039e-01
5.12081981e-01 -1.37573922e+00 -3.29188146e-02 3.76012743e-01
-2.25707591e-02 5.47192395e-01 1.40953851e+00 1.89823195e-01
-1.15567303e+00 -7.92564631e-01 3.13232869e-01 -3.55475575e-01
4.83819455e-01 -5.06242871e-01 -1.21070850e+00 2.81760454e-01
2.62789905e-01 7.09986910e-02 1.14046454e+00 1.97500303e-01
-2.43826464e-01 -4.03013408e-01 -1.01601839e+00 8.30694437e-02
6.26284182e-01 -5.16805112e-01 -6.10996783e-01 5.69257796e-01
5.86528301e-01 -1.96084559e-01 -7.52819836e-01 1.85955003e-01
3.47017258e-01 -1.21095371e+00 1.42421877e+00 2.82190478e-04
9.71928164e-02 -1.42448589e-01 -6.18384600e-01 -1.43561685e+00
-1.12848841e-01 -5.97655356e-01 1.65469527e-01 1.13575387e+00
5.18396974e-01 -6.88926935e-01 2.32386336e-01 5.20555258e-01
-3.69320124e-01 -3.48526418e-01 -9.90030050e-01 -7.32254744e-01
4.82935868e-02 -7.62673318e-01 3.76673579e-01 5.56709647e-01
-6.16893947e-01 2.45901883e-01 -2.58122176e-01 8.82884979e-01
1.00564373e+00 -1.88861623e-01 6.08982801e-01 -1.23854256e+00
-6.32467031e-01 2.44545285e-02 1.48457438e-01 -9.22583461e-01
1.57176763e-01 -3.47592354e-01 8.34735930e-01 -1.43484712e+00
-4.97948736e-01 -8.40969265e-01 -5.72271764e-01 -3.38836104e-01
2.08742470e-01 -1.37377949e-02 -3.26169878e-01 -4.38172817e-01
1.13756359e-01 4.68928516e-01 1.16101527e+00 -1.21989332e-01
-3.33359152e-01 6.07275665e-01 -2.75323689e-01 1.00210631e+00
7.65787363e-01 -2.79634327e-01 -5.66897154e-01 -3.64911497e-01
1.74598515e-01 1.38078511e-01 5.26198566e-01 -1.49479270e+00
2.69277487e-02 1.25014722e-01 2.60234892e-01 -4.72920984e-01
6.47231340e-01 -1.17585504e+00 9.06739905e-02 2.28363663e-01
-5.08873880e-01 -4.61575955e-01 4.09811944e-01 6.97303057e-01
-9.85156819e-02 -6.36581481e-01 8.73697102e-01 -2.36232966e-01
-2.53120154e-01 -4.15174514e-01 -5.56783438e-01 -2.58210480e-01
5.31938493e-01 -2.58357853e-01 3.90952714e-02 -7.22931325e-01
-8.87730658e-01 -6.87265575e-01 -3.56208026e-01 -9.97579396e-02
6.04177773e-01 -8.33605170e-01 -7.69832015e-01 1.91935226e-01
-1.32988214e-01 1.44886941e-01 6.39784336e-01 7.27878928e-01
-6.12499595e-01 4.10082132e-01 2.75276870e-01 -6.65599585e-01
-1.28133523e+00 5.25202572e-01 9.18508708e-01 3.65332328e-02
-5.05225301e-01 8.44031334e-01 6.26993954e-01 -7.51974761e-01
2.95506984e-01 -5.63431263e-01 -2.68393338e-01 -5.05443573e-01
6.96827173e-01 4.37794060e-01 4.34380054e-01 -6.87839985e-01
-3.38169001e-02 6.85947895e-01 5.61394989e-01 -6.87263086e-02
1.45362771e+00 -4.00565416e-01 -3.51143442e-02 1.06288552e+00
1.19384420e+00 4.84752953e-01 -8.47626269e-01 -2.65234828e-01
-5.24870753e-01 -2.92486310e-01 3.34807396e-01 -9.46807861e-01
-6.82938576e-01 1.14433384e+00 9.70829487e-01 4.49233353e-01
1.34544110e+00 -1.80736274e-01 2.35609397e-01 7.81307101e-01
2.63165474e-01 -1.11724579e+00 3.01197976e-01 3.00585151e-01
1.09335518e+00 -9.07855630e-01 1.43816490e-02 -5.93493283e-01
1.93677973e-02 1.37186408e+00 3.95156264e-01 3.87142738e-03
1.09566665e+00 4.77122396e-01 4.34412628e-01 2.72850245e-01
-1.47648320e-01 1.70871377e-01 1.37940928e-01 1.01845098e+00
6.94337308e-01 -7.28220716e-02 6.42008364e-01 3.08960408e-01
-4.46537852e-01 -5.99225581e-01 6.73203290e-01 7.84612834e-01
-9.21445668e-01 -5.91899991e-01 -1.13861966e+00 -3.58177535e-02
-5.45259058e-01 -2.27303103e-01 3.05416405e-01 6.76956356e-01
-8.86477232e-02 1.33847737e+00 -1.75111458e-01 -9.73761007e-02
5.43869853e-01 -9.81265120e-03 6.30373478e-01 -3.42044473e-01
-5.47519445e-01 3.35823029e-01 2.67028689e-01 -1.48343846e-01
-8.69736254e-01 -3.13424498e-01 -1.11457789e+00 1.15336478e-01
-6.02993667e-01 4.25239056e-01 9.94292736e-01 8.76233876e-01
-5.85418463e-01 1.22243845e+00 1.00537586e+00 -9.96148467e-01
-6.64591968e-01 -1.00562525e+00 -9.19255972e-01 -1.17033564e-01
9.27916408e-01 -4.79306996e-01 -1.00066793e+00 1.18595377e-01] | [15.159894943237305, 5.767819404602051] |
c164e75f-aaef-4cac-9d24-a052e91e6fa0 | matching-entropy-based-disparity-estimation | 2210.15948 | null | https://arxiv.org/abs/2210.15948v2 | https://arxiv.org/pdf/2210.15948v2.pdf | Matching entropy based disparity estimation from light field | A major challenge for matching-based depth estimation is to prevent mismatches in occlusion and smooth regions. An effective matching window satisfying three characteristics: texture richness, disparity consistency and anti-occlusion should be able to prevent mismatches to some extent. According to these characteristics, we propose matching entropy in the spatial domain of light field to measure the amount of correct information in a matching window, which provides the criterion for matching window selection. Based on matching entropy regularization, we establish an optimization model for depth estimation with a matching cost fidelity term. To find the optimum, we propose a two-step adaptive matching algorithm. First, the region type is adaptively determined to identify occluding, occluded, smooth and textured regions. Then, the matching entropy criterion is used to adaptively select the size and shape of matching windows, as well as the visible viewpoints. The two-step process can reduce mismatches and redundant calculations by selecting effective matching windows. The experimental results on synthetic and real data show that the proposed method can effectively improve the accuracy of depth estimation in occlusion and smooth regions and has strong robustness for different noise levels. Therefore, high-precision depth estimation from 4D light field data is achieved. | ['Jun Qiu', 'Xing Zhao', 'Di He', 'Chang Liu', 'Ligen Shi'] | 2022-10-28 | null | null | null | null | ['disparity-estimation'] | ['computer-vision'] | [ 2.58504152e-01 -5.80581427e-01 -9.81719643e-02 -3.60470325e-01
-1.79323912e-01 1.80407956e-01 8.95817950e-02 -4.49888706e-02
-3.30622554e-01 4.90164340e-01 1.95325091e-01 1.12700559e-01
-1.40716165e-01 -1.13885832e+00 -7.26392791e-02 -9.45689201e-01
4.26219791e-01 -1.09291933e-01 7.56565392e-01 -7.08245933e-02
7.98147142e-01 5.59205532e-01 -1.67626214e+00 1.03454784e-01
1.19606483e+00 1.19711566e+00 6.02932274e-01 1.67684361e-01
-2.44604751e-01 4.85065937e-01 -2.62018114e-01 6.12084009e-02
4.44966912e-01 -3.83302778e-01 -1.65369555e-01 1.92980394e-01
3.55266362e-01 -7.16182888e-01 -2.54717916e-01 1.23719323e+00
7.39245057e-01 3.51823211e-01 4.66607034e-01 -6.91682160e-01
-4.78848442e-02 -1.97276339e-01 -7.44006813e-01 2.05699012e-01
4.37584817e-01 2.30116084e-01 6.04262054e-01 -9.39888537e-01
5.90970695e-01 1.25210309e+00 2.62958229e-01 3.31718326e-01
-1.00563812e+00 -6.96170628e-01 -6.46129297e-03 3.73877764e-01
-1.57863867e+00 -5.15309751e-01 1.10227668e+00 -4.28977758e-01
5.08926392e-01 2.44646460e-01 8.05337071e-01 1.57481223e-01
4.56241816e-01 3.67528826e-01 1.16891670e+00 -6.02157116e-01
1.72655299e-01 1.02121048e-01 -3.80031206e-02 6.90311015e-01
1.99267432e-01 3.70741695e-01 -3.88340652e-01 -9.66097116e-02
1.10260665e+00 -1.64527111e-02 -6.74668074e-01 -3.69814157e-01
-1.05994403e+00 6.79687262e-01 2.54142553e-01 1.95672348e-01
-1.52180851e-01 -2.95040339e-01 2.31364176e-01 7.07195792e-03
5.00181615e-01 -6.87009003e-03 -1.61538869e-01 1.37368813e-01
-5.57115078e-01 9.68453437e-02 2.22725615e-01 6.18985057e-01
1.05121171e+00 -1.95944905e-01 -1.61227345e-01 1.09217191e+00
5.10965228e-01 5.37511468e-01 3.14743817e-01 -9.93755519e-01
6.17487609e-01 7.14849591e-01 8.19115788e-02 -1.15874732e+00
-3.14364493e-01 5.41675789e-03 -7.49308288e-01 7.43026316e-01
2.82031149e-01 1.74565911e-01 -6.93655133e-01 1.33451998e+00
8.22390199e-01 6.94131181e-02 -6.63550645e-02 1.08766329e+00
1.05355430e+00 7.10602582e-01 -2.95477152e-01 -6.56769693e-01
1.31714964e+00 -3.98168802e-01 -9.60067630e-01 -1.22029968e-01
1.06886573e-01 -1.17866373e+00 9.36362326e-01 3.52372050e-01
-1.17534423e+00 -6.41461134e-01 -1.00515985e+00 -2.47346491e-01
2.10539475e-01 1.18332140e-01 3.64055127e-01 3.31466556e-01
-5.69453537e-01 4.35726285e-01 -4.27556187e-01 5.44990599e-02
1.22696728e-01 2.66838998e-01 -4.20608185e-02 -1.92040026e-01
-1.12267137e+00 7.70577729e-01 4.25071508e-01 1.02550313e-01
-1.19021878e-01 -4.11271155e-01 -8.53312135e-01 -2.21801311e-01
1.63029358e-02 -7.13750124e-01 5.16253710e-01 -5.77566087e-01
-1.59727085e+00 9.00088429e-01 -5.24544179e-01 1.42866939e-01
5.30009985e-01 2.66274095e-01 -3.21978986e-01 3.75014603e-01
1.52189685e-02 2.78836280e-01 8.05993855e-01 -1.16988337e+00
-9.58315313e-01 -5.70310056e-01 -1.66407317e-01 5.71024656e-01
-4.05351445e-02 -5.66276200e-02 -5.66300094e-01 -3.85948390e-01
9.18467641e-01 -3.67008924e-01 -1.60819352e-01 3.24032933e-01
-1.02361567e-01 -1.43868014e-01 7.37600267e-01 -4.92909640e-01
1.35050237e+00 -2.28953171e+00 -2.84240972e-02 3.99214208e-01
2.76484579e-01 1.81276761e-02 2.07894474e-01 -6.93085417e-02
2.36838654e-01 -3.66563857e-01 -1.36841968e-01 1.92907583e-02
-5.07060349e-01 3.42133455e-02 7.93088824e-02 6.61530256e-01
-1.21566333e-01 2.45257273e-01 -6.04925275e-01 -8.10292184e-01
8.25237691e-01 5.37612319e-01 -5.68482697e-01 3.92216742e-01
-5.72003350e-02 6.67839170e-01 -7.30622888e-01 7.39747643e-01
1.32220101e+00 1.43650770e-01 -2.96637446e-01 -5.04511237e-01
-5.81680298e-01 1.14269525e-01 -1.60943520e+00 1.32825649e+00
-5.29097795e-01 6.01454079e-01 2.49998808e-01 -4.93434012e-01
1.41827559e+00 6.40195310e-02 5.79571486e-01 -1.00688446e+00
3.69686842e-01 5.27556896e-01 -2.38217786e-01 -9.05973494e-01
3.61416072e-01 -2.84751505e-01 5.35704076e-01 -5.03119193e-02
-7.48439312e-01 -4.91980910e-01 -1.11882322e-01 -3.82532477e-01
6.06334865e-01 -1.44585401e-01 4.10930276e-01 -2.54722714e-01
1.11243534e+00 -4.16992188e-01 1.08979976e+00 1.91681728e-01
-3.68057996e-01 6.43821597e-01 2.15065237e-02 -6.90530956e-01
-8.33640039e-01 -8.10970604e-01 -7.36053407e-01 3.36312324e-01
1.09036815e+00 5.26005253e-02 -4.82322842e-01 -1.32960290e-01
-1.21256523e-01 3.18632543e-01 -3.10800195e-01 -4.44236100e-02
-8.60630274e-01 -6.24464512e-01 -3.91314417e-01 3.99273112e-02
9.53629255e-01 -1.02750540e+00 -6.76807404e-01 1.50210604e-01
-4.58804101e-01 -9.35561121e-01 -3.93224955e-01 -3.00278097e-01
-1.00904953e+00 -1.13895559e+00 -5.80385387e-01 -9.69470680e-01
8.31356883e-01 5.81887066e-01 8.85946512e-01 4.07560766e-01
-2.00833008e-01 -1.20314807e-01 -1.09516449e-01 5.01763523e-02
-4.59592305e-02 -4.71377492e-01 -3.55843455e-01 1.55624285e-01
2.55494863e-01 -6.02216542e-01 -1.13790333e+00 7.41507053e-01
-8.47564280e-01 2.37393931e-01 3.33907574e-01 8.27243984e-01
8.34422410e-01 2.99460471e-01 1.26933411e-01 -2.71741241e-01
2.85397500e-01 4.79851253e-02 -1.05440032e+00 2.11297199e-02
-5.77559769e-01 -1.09593920e-01 3.86573136e-01 -4.81918335e-01
-1.33080626e+00 -6.85105249e-02 -3.69912386e-01 -1.87316179e-01
-3.26413736e-02 3.03701106e-02 -4.67021793e-01 -5.62381625e-01
4.19062555e-01 3.67257774e-01 -7.92190135e-02 -2.70863891e-01
-1.31812051e-01 5.97956657e-01 2.43287072e-01 -4.68846828e-01
6.02608383e-01 8.97345483e-01 4.17381972e-01 -8.44740868e-01
-4.09190148e-01 -5.40710628e-01 -4.73784238e-01 -5.06103575e-01
9.60841954e-01 -7.55465031e-01 -1.04613960e+00 5.51754475e-01
-1.18371630e+00 3.16413417e-02 -6.96783736e-02 8.55544031e-01
-3.59341294e-01 7.27201045e-01 -5.17455399e-01 -9.22264099e-01
-3.74438733e-01 -1.33588755e+00 9.13984537e-01 5.18689394e-01
3.15834045e-01 -9.70825136e-01 -5.28163947e-02 2.44640902e-01
2.97696829e-01 1.96457326e-01 8.22239101e-01 4.65927571e-01
-1.18295681e+00 3.24703380e-02 -3.53123456e-01 2.16128722e-01
2.61418283e-01 1.94574401e-01 -8.76841664e-01 -4.44789492e-02
4.80665296e-01 1.20536737e-01 8.02789390e-01 8.70538473e-01
1.04238236e+00 -9.61624458e-03 -3.25263590e-01 1.16449630e+00
1.75259340e+00 7.06640780e-01 1.05095840e+00 5.16505361e-01
4.81613547e-01 7.67198145e-01 1.01548815e+00 5.25164783e-01
2.04565540e-01 8.21437001e-01 5.36905110e-01 -2.53608555e-01
-1.55196190e-01 -1.42957881e-01 -6.56168982e-02 7.79350817e-01
-5.68264499e-02 -3.29719111e-02 -3.74400705e-01 2.09516466e-01
-1.49610257e+00 -9.51768637e-01 -3.64893645e-01 2.67809677e+00
8.37165415e-01 2.51609325e-01 -3.42166960e-01 3.11104208e-01
9.30046976e-01 2.06665620e-01 -4.30123210e-01 -7.43746161e-02
-3.26278716e-01 -2.13873461e-02 3.45405906e-01 9.64266002e-01
-8.37463796e-01 6.72889948e-01 5.79181528e+00 1.25784504e+00
-1.14515126e+00 -2.60178715e-01 5.17585695e-01 1.37462974e-01
-6.14723861e-01 2.20111877e-01 -1.01085603e+00 8.07894588e-01
-1.62049949e-01 -5.12004690e-03 3.19957107e-01 5.64313710e-01
6.47781789e-01 -5.83157182e-01 -5.66592991e-01 1.27288997e+00
-1.54455185e-01 -1.11257446e+00 -1.83531553e-01 -4.93875816e-02
7.64240265e-01 -3.23797226e-01 -1.50277451e-01 -4.72278178e-01
-3.01500231e-01 -5.15142083e-01 3.94030035e-01 6.14844263e-01
7.17787325e-01 -6.64924324e-01 6.71019495e-01 4.30479586e-01
-1.46878648e+00 1.59916341e-01 -7.58237779e-01 -7.04567647e-03
2.47501150e-01 1.01423144e+00 -1.91076934e-01 3.67079973e-01
4.77857620e-01 7.58591115e-01 -7.41722435e-02 1.31138456e+00
-2.62582064e-01 -1.36171713e-01 -3.95179152e-01 7.62789920e-02
-3.43547165e-01 -8.38162065e-01 6.07842207e-01 6.63424730e-01
3.51350874e-01 6.38450444e-01 1.77485958e-01 7.80009627e-01
4.16518897e-01 5.09507477e-01 -4.50034261e-01 9.73238170e-01
6.24711096e-01 9.57227767e-01 -5.93680143e-01 -2.93144941e-01
-4.75270391e-01 6.04572594e-01 1.00832812e-01 3.39469463e-01
-5.73029101e-01 -5.60500920e-01 5.13247728e-01 4.10409003e-01
-6.32502139e-02 -1.07603602e-01 -6.55961275e-01 -1.17705107e+00
3.34493637e-01 -3.88875425e-01 3.03019106e-01 -7.91636586e-01
-9.92528021e-01 4.53108698e-01 -1.73307672e-01 -1.66989958e+00
4.03744191e-01 -3.22617620e-01 -8.14710855e-01 1.10231435e+00
-1.89602220e+00 -5.30712187e-01 -7.92989135e-01 7.80239403e-01
4.91929322e-01 2.14307740e-01 2.71465629e-01 4.91179705e-01
-5.84307551e-01 1.90228671e-01 9.64582860e-02 -2.63136268e-01
4.98602301e-01 -5.69176674e-01 -2.14962274e-01 8.54224741e-01
-5.53621352e-01 4.11830306e-01 4.95011061e-01 -6.56031311e-01
-9.73182261e-01 -4.48250532e-01 8.68073404e-01 1.58570260e-01
-1.55001553e-02 -7.70121887e-02 -1.14835501e+00 -1.61818042e-02
-2.50200391e-01 6.36926368e-02 1.01302855e-01 -4.78566080e-01
1.04908280e-01 -4.91037130e-01 -1.39667320e+00 5.69822013e-01
9.16967869e-01 -3.06950599e-01 -3.32857639e-01 3.48269671e-01
5.95397055e-01 -7.62062490e-01 -8.18534255e-01 8.42376709e-01
6.75063968e-01 -1.56436872e+00 1.14140999e+00 4.40760523e-01
4.16589051e-01 -6.18559361e-01 -1.78779978e-02 -7.79498339e-01
-3.22521329e-01 -4.02421325e-01 2.74365008e-01 1.07415915e+00
-9.05279368e-02 -9.62965488e-01 8.09705675e-01 6.49159074e-01
-7.69366249e-02 -8.34434986e-01 -1.01868260e+00 -4.83818948e-01
-4.44134027e-01 -8.82549509e-02 4.97030228e-01 7.27891564e-01
-2.52590448e-01 -2.04187334e-02 -2.23933548e-01 2.77433634e-01
9.07096744e-01 4.07095045e-01 8.00273418e-01 -1.27375400e+00
2.54194830e-02 -5.03625810e-01 -4.80296314e-01 -1.51767766e+00
-1.45031050e-01 -2.77249098e-01 -3.20144445e-02 -1.69464576e+00
5.70366494e-02 -9.12786543e-01 -2.47168709e-02 -1.50815129e-01
-4.03723866e-01 8.37677810e-03 -2.61440545e-01 6.54283047e-01
1.89665612e-02 7.30484128e-01 1.79444647e+00 1.13213152e-01
-4.50841069e-01 2.65165240e-01 -7.86070749e-02 7.85882294e-01
6.23844981e-01 -1.30081207e-01 -3.57862771e-01 -4.16323870e-01
6.25046492e-02 3.00625443e-01 8.97429660e-02 -9.85999942e-01
8.60104784e-02 -6.57476008e-01 4.86924320e-01 -7.44856060e-01
5.72240710e-01 -1.06403613e+00 2.82223653e-02 5.95249295e-01
-4.91654128e-02 -4.17256832e-01 -3.89181636e-02 3.08500290e-01
-2.92940438e-01 -3.55623186e-01 1.31203163e+00 3.01633086e-02
-8.16054106e-01 4.39594060e-01 2.12190419e-01 -3.63109969e-02
1.05633402e+00 -6.71099663e-01 -2.18132317e-01 -2.79737324e-01
-3.79056573e-01 9.11290795e-02 7.94413030e-01 -8.84967893e-02
1.04738629e+00 -1.40692604e+00 -7.32587278e-01 7.51967371e-01
1.40447780e-01 9.75084603e-02 5.44411004e-01 6.21274710e-01
-9.70518470e-01 -2.46604439e-02 -1.34676561e-01 -7.55454123e-01
-1.43329263e+00 3.47873628e-01 5.18265545e-01 4.31994274e-02
-6.93776786e-01 8.20381165e-01 4.46724564e-01 1.73284635e-01
8.92516300e-02 -3.96184564e-01 -3.60806108e-01 -3.28486264e-01
6.71827495e-01 4.25913483e-01 -6.12163395e-02 -5.06153286e-01
-2.76801229e-01 1.49183011e+00 2.57145345e-01 2.21194606e-02
7.78630555e-01 -6.21274650e-01 -2.78730690e-01 4.28873003e-02
1.28786886e+00 5.10437131e-01 -1.37526822e+00 -3.45244884e-01
-6.85958922e-01 -1.42159760e+00 3.45804274e-01 -1.40472278e-01
-1.29177570e+00 9.85437930e-01 7.43615627e-01 1.67283759e-01
1.57240427e+00 -3.33323926e-01 9.21151400e-01 -3.16334605e-01
4.06383276e-01 -1.10175574e+00 -1.35000020e-01 1.68285310e-01
6.67353630e-01 -1.38789654e+00 4.00447786e-01 -9.25743759e-01
-2.41877750e-01 1.26712763e+00 8.01526129e-01 3.18913534e-02
6.41845047e-01 4.86977212e-02 1.02512844e-01 -1.16674401e-01
-3.94601673e-01 -2.39959940e-01 3.49943310e-01 4.77285355e-01
1.99836671e-01 -2.98572063e-01 -7.56620646e-01 -2.26964012e-01
1.71510264e-01 -1.22639507e-01 2.43674636e-01 6.54667914e-01
-9.91362333e-01 -8.71810138e-01 -6.77945435e-01 9.46066603e-02
-2.56177068e-01 -7.38762170e-02 3.00981611e-01 5.91831148e-01
4.15005356e-01 1.08536446e+00 1.24531783e-01 -3.47309172e-01
3.63061637e-01 -5.94567716e-01 3.96062613e-01 -3.32397312e-01
-1.48178741e-01 3.93643796e-01 -2.65275061e-01 -5.29241920e-01
-5.40423751e-01 -3.22429061e-01 -1.32596016e+00 -2.92629898e-01
-7.66221285e-01 -2.17418019e-02 5.23687184e-01 6.27376676e-01
6.18862770e-02 7.88279697e-02 1.13099611e+00 -6.18460655e-01
-5.93217164e-02 -6.83691144e-01 -6.14807129e-01 4.96151567e-01
3.08419853e-01 -8.33124042e-01 -8.06097746e-01 -3.28024328e-01] | [9.265759468078613, -2.4644501209259033] |
aec32c4b-3a07-40ad-889b-9698ffad193d | improving-siem-for-critical-scada-water | 1904.05724 | null | http://arxiv.org/abs/1904.05724v1 | http://arxiv.org/pdf/1904.05724v1.pdf | Improving SIEM for Critical SCADA Water Infrastructures Using Machine Learning | Network Control Systems (NAC) have been used in many industrial processes.
They aim to reduce the human factor burden and efficiently handle the complex
process and communication of those systems. Supervisory control and data
acquisition (SCADA) systems are used in industrial, infrastructure and facility
processes (e.g. manufacturing, fabrication, oil and water pipelines, building
ventilation, etc.) Like other Internet of Things (IoT) implementations, SCADA
systems are vulnerable to cyber-attacks, therefore, a robust anomaly detection
is a major requirement. However, having an accurate anomaly detection system is
not an easy task, due to the difficulty to differentiate between cyber-attacks
and system internal failures (e.g. hardware failures). In this paper, we
present a model that detects anomaly events in a water system controlled by
SCADA. Six Machine Learning techniques have been used in building and
evaluating the model. The model classifies different anomaly events including
hardware failures (e.g. sensor failures), sabotage and cyber-attacks (e.g. DoS
and Spoofing). Unlike other detection systems, our proposed work focuses on
notifying the operator when an anomaly occurs with a probability of the event
occurring. This additional information helps in accelerating the mitigation
process. The model is trained and tested using a real-world dataset. | ['Xavier Bellekens', 'Hanan Hindy', 'David Brosset', 'Amar Seeam', 'Ethan Bayne'] | 2019-03-06 | null | null | null | null | ['cyber-attack-detection'] | ['miscellaneous'] | [ 2.60506839e-01 -1.91114590e-01 3.41987789e-01 4.34107631e-02
4.51359034e-01 -5.05861402e-01 6.97240293e-01 9.01232719e-01
4.98776659e-02 4.18271303e-01 -5.11770487e-01 -6.76468849e-01
-4.46246296e-01 -9.19066966e-01 -2.82558560e-01 -8.09628010e-01
-3.03714871e-01 2.35057309e-01 7.48548031e-01 -1.64559309e-03
4.03277159e-01 7.74102390e-01 -1.54563797e+00 -1.93192169e-01
4.22623992e-01 1.11758828e+00 5.97974211e-02 7.18624771e-01
-5.74545637e-02 6.30788088e-01 -1.09192455e+00 5.72785556e-01
2.96102822e-01 -2.57095873e-01 -6.82677627e-01 1.26713917e-01
-2.91520745e-01 -3.40403020e-01 3.40570770e-02 1.12609410e+00
1.22073472e-01 -3.77874635e-02 3.60871971e-01 -2.00348163e+00
2.28821248e-01 2.89945215e-01 -1.92476735e-01 3.32885683e-01
4.37604904e-01 4.14440334e-01 1.86520129e-01 -3.89797539e-01
3.97792161e-02 1.27417743e+00 2.67639965e-01 4.75433990e-02
-9.42671478e-01 -7.83279538e-01 -9.30109844e-02 1.98079914e-01
-9.62837994e-01 1.89015493e-01 6.44415915e-01 -5.14615119e-01
9.37403321e-01 7.72162229e-02 3.18830222e-01 1.01546967e+00
1.03537738e+00 9.62851644e-02 9.15176809e-01 -3.65264952e-01
7.66021669e-01 -1.62032604e-01 4.23379928e-01 9.79657099e-02
6.45061612e-01 1.79045662e-01 4.29585651e-02 -3.09328288e-01
6.66193008e-01 5.04617810e-01 6.89656287e-02 -5.61374761e-02
-1.07022846e+00 3.21601540e-01 -4.70976308e-02 5.51293790e-01
-8.00468445e-01 -7.54262954e-02 7.10143685e-01 6.65076733e-01
-4.36145723e-01 4.54063743e-01 -6.77236319e-01 -1.68093875e-01
-1.09298982e-01 -3.93967591e-02 1.09875715e+00 7.73143113e-01
3.29834312e-01 4.83199060e-01 3.62234086e-01 3.41279395e-02
4.25356865e-01 4.53519911e-01 4.74170923e-01 -4.09267098e-01
3.64943333e-02 6.78829610e-01 1.57768905e-01 -9.54797626e-01
-5.26698589e-01 9.45779756e-02 -9.93903995e-01 7.82366872e-01
1.30336046e-01 -2.78589487e-01 -1.00428808e+00 9.55886245e-01
4.74489450e-01 5.62134564e-01 1.35565728e-01 3.45249981e-01
-3.00346494e-01 7.77935565e-01 2.86047310e-01 -4.05575365e-01
1.24069214e+00 -1.58752188e-01 -1.31403410e+00 1.85299560e-01
3.22871536e-01 -1.05381215e+00 3.54633033e-01 9.82122421e-01
-4.21311051e-01 -6.53125405e-01 -1.25581813e+00 1.03372753e+00
-6.85075223e-01 -5.93937755e-01 3.49674344e-01 6.13460720e-01
-4.40054387e-01 8.20435286e-01 -1.20853567e+00 -6.05011106e-01
-1.58432558e-01 5.35222054e-01 -2.91615486e-01 1.99768603e-01
-1.11258340e+00 1.08706617e+00 7.13975251e-01 1.05512448e-01
-1.27980435e+00 -3.25935990e-01 -5.84913313e-01 -2.84169048e-01
5.40783882e-01 1.16390064e-01 1.16034603e+00 -2.27601916e-01
-1.28510118e+00 -1.08168520e-01 3.99048716e-01 -5.76097310e-01
1.54367685e-01 -3.62521797e-01 -1.08000743e+00 2.11732090e-02
-3.46141487e-01 -4.43199933e-01 8.76969874e-01 -1.00037074e+00
-8.72966111e-01 -3.06418300e-01 -5.34126610e-02 -7.73344696e-01
-1.64337624e-02 2.84303933e-01 1.04604805e+00 -2.07169294e-01
2.59021968e-01 -7.69500852e-01 -1.91778213e-01 -5.20298302e-01
-6.69766188e-01 -3.55920017e-01 1.78028572e+00 -4.95550126e-01
1.21708155e+00 -2.10916376e+00 -4.20889527e-01 8.55364323e-01
-1.83194026e-01 7.76231170e-01 4.99027014e-01 1.03283477e+00
-3.69807869e-01 -9.45706572e-03 -7.61074498e-02 3.79440129e-01
-6.82059079e-02 4.65094656e-01 -3.42781365e-01 5.00432432e-01
3.12930435e-01 -1.72379255e-01 -6.51769876e-01 7.89519474e-02
1.00055027e+00 -3.16017075e-03 1.27487853e-01 4.06542778e-01
-1.21449642e-02 6.76645219e-01 -8.01497936e-01 8.52964938e-01
3.73374999e-01 2.31895983e-01 3.58693041e-02 -1.46483630e-01
-3.78826976e-01 -3.32203917e-02 -1.78461766e+00 7.59541094e-01
-4.93054390e-01 2.07008962e-02 1.93600655e-01 -1.07618570e+00
1.12606657e+00 1.03182960e+00 5.34617186e-01 -7.02311695e-02
4.67406422e-01 5.25042832e-01 3.37027967e-01 -7.45910406e-01
-1.59550413e-01 1.74505979e-01 8.12946707e-02 4.86286461e-01
-6.89729601e-02 -1.18260808e-01 1.46740869e-01 -5.35135791e-02
1.82146585e+00 -2.90670544e-01 7.06461966e-01 -2.81832904e-01
9.34174776e-01 -8.93701687e-02 6.02960289e-01 4.08651143e-01
-3.62986624e-01 -2.79050142e-01 3.41289699e-01 -6.64508760e-01
-1.10573387e+00 -1.07654059e+00 3.27616632e-02 5.26336171e-02
2.58426100e-01 -1.44927099e-01 -3.74645561e-01 -5.65545976e-01
6.10546134e-02 9.41013992e-01 8.55087489e-02 -6.32733941e-01
-4.87336576e-01 -3.56777489e-01 4.07157362e-01 4.03993756e-01
4.88287121e-01 -1.31010091e+00 -8.31162930e-01 7.30656326e-01
6.76880300e-01 -1.11994886e+00 2.68356651e-01 4.96911913e-01
-8.94073009e-01 -1.54375589e+00 5.59119344e-01 -4.75804865e-01
6.50206566e-01 1.29058927e-01 6.14256203e-01 2.72498339e-01
-7.11884141e-01 3.92038018e-01 -5.24448276e-01 -8.96209955e-01
-9.59966660e-01 -5.26395440e-01 6.50130630e-01 -8.26162379e-03
7.60733783e-01 -6.73750281e-01 -3.04028571e-01 5.80268979e-01
-1.34338665e+00 -9.78952408e-01 5.75614274e-01 4.15215552e-01
1.90082207e-01 9.48583186e-01 7.27491260e-01 -5.73178947e-01
7.68016160e-01 -6.87660336e-01 -9.88901258e-01 5.52505478e-02
-8.26520205e-01 -2.45642021e-01 9.35467899e-01 -5.23424447e-01
-6.48273289e-01 2.02117145e-01 -6.68145716e-03 -3.75083596e-01
-1.26387346e+00 2.00503215e-01 -3.71259421e-01 8.58666897e-02
1.86286077e-01 -1.37726679e-01 1.99029461e-01 -5.14691114e-01
-5.54471254e-01 9.14523900e-01 3.25452477e-01 -1.94269463e-01
1.08825088e+00 1.46202609e-01 3.52989346e-01 -1.04315817e+00
-2.28350073e-01 -5.16443491e-01 -4.26301837e-01 -3.10638040e-01
8.22550476e-01 -4.00929123e-01 -1.11976898e+00 7.69813538e-01
-1.17187572e+00 1.11481778e-01 -1.99855730e-01 8.59289348e-01
-7.72332260e-03 2.74809778e-01 -5.87925732e-01 -1.15887570e+00
-2.24012062e-01 -1.01413918e+00 6.30235136e-01 2.46059135e-01
-4.05296504e-01 -7.64926076e-01 1.09821871e-01 -2.77547687e-01
5.75269282e-01 6.44095361e-01 8.39008868e-01 -1.20464265e+00
-6.26462936e-01 -6.25971854e-01 5.60535073e-01 7.48943269e-01
7.90204883e-01 5.47923982e-01 -6.67222738e-01 -4.60907221e-01
6.20987415e-01 1.26585156e-01 -2.06872344e-01 -4.78768758e-02
9.14853096e-01 -1.85679048e-01 -3.21859032e-01 -1.39740303e-01
1.35811806e+00 1.13804698e+00 7.30646074e-01 3.72866660e-01
5.21793425e-01 5.08294046e-01 1.07965028e+00 5.98924339e-01
-5.76344252e-01 3.11619490e-01 1.06997788e+00 1.86913416e-01
6.84726000e-01 1.39275298e-01 6.85985982e-01 5.69829822e-01
1.10014260e-01 -3.31988961e-01 -9.26753700e-01 1.96846619e-01
-1.60435045e+00 -8.83618474e-01 -5.40331125e-01 2.22799349e+00
1.55494958e-01 4.58204240e-01 -1.52633101e-01 1.11005759e+00
1.06122708e+00 -3.10165644e-01 -4.85692710e-01 -6.67434812e-01
3.93990517e-01 3.20942372e-01 4.00786281e-01 2.79624552e-01
-1.16529131e+00 3.55580896e-01 4.81955862e+00 2.88284749e-01
-9.61125374e-01 -1.23350032e-01 -1.01142325e-01 5.57072103e-01
5.84995925e-01 -2.79496685e-02 -6.49735391e-01 6.17685616e-01
1.16136885e+00 1.36280656e-01 1.13483489e-01 9.05665636e-01
6.00209236e-01 -3.84169035e-02 -1.21782351e+00 5.51043808e-01
-4.62588936e-01 -5.09732246e-01 -1.74685746e-01 1.52523115e-01
2.70380944e-01 -2.32167661e-01 -6.32172763e-01 -1.34578124e-01
5.41043937e-01 -7.38038301e-01 2.87427813e-01 2.04056367e-01
-1.96969017e-01 -8.20349991e-01 1.38211310e+00 1.61649957e-01
-1.06627011e+00 -4.69548345e-01 8.22500810e-02 -3.87707949e-01
6.92872822e-01 9.41303551e-01 -8.76795411e-01 7.12764442e-01
7.58120120e-01 2.59290159e-01 -1.14161253e-01 1.00938082e+00
-4.01149422e-01 6.65242910e-01 -5.81315994e-01 -9.68796108e-03
2.94328630e-02 -1.06257334e-01 7.07421601e-01 4.37801719e-01
5.30895412e-01 -7.86077157e-02 4.86332387e-01 5.21918118e-01
7.37538815e-01 -2.68314481e-01 -8.85262787e-01 -5.69602214e-02
7.92532027e-01 1.09603715e+00 -7.71060407e-01 7.82262813e-03
-2.21983612e-01 4.84828651e-01 -7.27503836e-01 8.07279721e-02
-5.97683132e-01 -8.55714381e-01 7.42574573e-01 8.21068436e-02
6.99823648e-02 -5.64208388e-01 1.80659354e-01 -3.39821398e-01
-7.75180161e-02 -8.38143945e-01 3.45552415e-01 -3.94027770e-01
-1.54439104e+00 3.72586459e-01 -2.24862527e-02 -1.42171097e+00
-3.03926378e-01 -8.55664134e-01 -1.06441689e+00 4.76276070e-01
-1.02417266e+00 -6.56042635e-01 -4.12412226e-01 8.26858461e-01
3.55734795e-01 -2.18216211e-01 8.72524977e-01 4.17062640e-01
-7.34818399e-01 -2.10781977e-01 -2.96208411e-01 3.59014757e-02
4.35398787e-01 -1.21374583e+00 1.43827364e-01 1.18579328e+00
-3.91725689e-01 5.78290880e-01 1.04734802e+00 -9.42785025e-01
-1.46339869e+00 -9.27090228e-01 3.08270305e-01 -8.35382938e-02
1.13469243e+00 -4.58410122e-02 -9.65195835e-01 8.83977771e-01
3.14221054e-01 1.56329662e-01 3.50719273e-01 -5.11766255e-01
3.33311290e-01 -3.30037951e-01 -1.46748209e+00 1.53311402e-01
2.88305014e-01 -1.07881166e-01 -8.03967059e-01 2.57727116e-01
5.43472588e-01 -5.73209897e-02 -1.07337308e+00 4.27075088e-01
-1.20827267e-02 -9.14193392e-01 6.50457561e-01 -5.73759675e-01
-1.27487615e-01 -8.60083938e-01 1.55852646e-01 -1.23426628e+00
1.88245974e-03 -7.47227550e-01 -3.08323771e-01 1.50223601e+00
-1.44001752e-01 -1.20568192e+00 2.05988556e-01 3.79465252e-01
-2.91249305e-01 -3.28922234e-02 -9.79766071e-01 -1.01000977e+00
-7.54225433e-01 -2.53890753e-01 8.60247612e-01 1.13856399e+00
2.74617165e-01 7.90968239e-02 -6.14980012e-02 1.09212303e+00
4.83176798e-01 -4.18615878e-01 7.20919430e-01 -1.56409740e+00
2.02637538e-02 9.49359015e-02 -9.80739415e-01 -5.93851879e-02
-4.37796354e-01 -3.00246533e-02 9.52883735e-02 -1.17838120e+00
-8.55558038e-01 -3.95333648e-01 -6.30267680e-01 5.29688835e-01
3.26001734e-01 -4.22700018e-01 -1.08815275e-01 -1.71474535e-02
-1.59114823e-01 1.92671940e-01 4.98703569e-01 6.98981509e-02
-1.60655230e-01 3.64582777e-01 3.28197032e-01 8.32389057e-01
1.37823462e+00 -5.97701371e-01 -3.69610488e-01 1.51770890e-01
-2.91634500e-02 6.56852871e-02 5.74037194e-01 -1.37890375e+00
2.76388377e-01 -4.02641505e-01 -2.71605160e-02 -6.61868870e-01
-2.09477320e-01 -1.90135539e+00 4.61904109e-01 1.37800312e+00
2.78740019e-01 6.07390821e-01 3.35135753e-03 8.01835597e-01
-5.01934707e-01 -3.49851012e-01 5.89429557e-01 4.29568402e-02
-6.23009562e-01 1.05788194e-01 -8.58599246e-01 -8.31259251e-01
1.76444924e+00 -1.71888590e-01 -5.23748457e-01 -4.35510790e-03
-6.46329284e-01 2.59629220e-01 1.06357291e-01 5.78123868e-01
5.89565575e-01 -9.83076751e-01 -4.99428771e-02 5.38559973e-01
7.57457269e-03 1.26645431e-01 -2.10030749e-02 8.10970426e-01
-6.61090195e-01 2.40590259e-01 -4.06349361e-01 -6.32966340e-01
-1.43132102e+00 6.79420590e-01 2.15012506e-01 -2.02101603e-01
-4.36902881e-01 -2.06336081e-02 -6.33846164e-01 -1.01046480e-01
3.09119374e-01 -6.03115499e-01 -3.99545640e-01 -3.84614348e-01
6.32934213e-01 9.39840317e-01 2.68894374e-01 -1.28906757e-01
-3.83142322e-01 2.50384480e-01 -6.22802414e-02 4.45239902e-01
1.13152027e+00 6.07353114e-02 -5.14807820e-01 6.14822030e-01
4.59783912e-01 -2.15967998e-01 -8.17710340e-01 1.52525887e-01
4.76400971e-01 -4.15755630e-01 -8.68272781e-02 -5.79203844e-01
-9.16228771e-01 7.61966050e-01 9.05152082e-01 1.12093270e+00
8.34899485e-01 -4.57691133e-01 7.57622719e-01 4.52168524e-01
7.20664918e-01 -1.06471229e+00 1.92495346e-01 4.21636254e-01
3.27153057e-01 -9.35487390e-01 1.79467741e-02 -5.74531317e-01
-2.56076276e-01 1.23455679e+00 6.82188094e-01 -3.08257610e-01
1.10855389e+00 6.62176490e-01 2.31955685e-02 -4.22234327e-01
-5.46483815e-01 4.11595732e-01 -6.13721251e-01 8.24900508e-01
-1.05246231e-01 7.96539262e-02 -1.62309512e-01 2.36270860e-01
1.61111057e-01 1.05261207e-01 8.24249864e-01 1.57279134e+00
-6.36731625e-01 -1.24485302e+00 -7.97039270e-01 4.82283741e-01
-6.93201065e-01 6.33031964e-01 -5.78683913e-02 8.06180239e-01
3.00469398e-01 1.57583797e+00 1.81874141e-01 -6.04628265e-01
7.20010161e-01 1.19496234e-01 -4.34250712e-01 -5.61069131e-01
-6.36827707e-01 -2.35187814e-01 -3.53962090e-03 -7.63095200e-01
-3.17160070e-01 -5.13181686e-01 -1.46701336e+00 -3.25111836e-01
-4.01637584e-01 7.78186321e-02 1.02743435e+00 1.12281895e+00
1.82213619e-01 1.11135960e+00 9.66078222e-01 -3.97408456e-01
-8.64493728e-01 -9.66235936e-01 -9.88338172e-01 5.19936562e-01
3.70502442e-01 -8.89988899e-01 -8.78977239e-01 -1.06839150e-01] | [6.304579734802246, 2.603804588317871] |
cea53249-c594-4f3d-ba9e-46fc48f86df1 | prototypes-as-explanation-for-time-series | 2307.01601 | null | https://arxiv.org/abs/2307.01601v1 | https://arxiv.org/pdf/2307.01601v1.pdf | Prototypes as Explanation for Time Series Anomaly Detection | Detecting abnormal patterns that deviate from a certain regular repeating pattern in time series is essential in many big data applications. However, the lack of labels, the dynamic nature of time series data, and unforeseeable abnormal behaviors make the detection process challenging. Despite the success of recent deep anomaly detection approaches, the mystical mechanisms in such black-box models have become a new challenge in safety-critical applications. The lack of model transparency and prediction reliability hinders further breakthroughs in such domains. This paper proposes ProtoAD, using prototypes as the example-based explanation for the state of regular patterns during anomaly detection. Without significant impact on the detection performance, prototypes shed light on the deep black-box models and provide intuitive understanding for domain experts and stakeholders. We extend the widely used prototype learning in classification problems into anomaly detection. By visualizing both the latent space and input space prototypes, we intuitively demonstrate how regular data are modeled and why specific patterns are considered abnormal. | ['Emmanuel Müller', 'Carsten Jentsch', 'Bin Li'] | 2023-07-04 | null | null | null | null | ['anomaly-detection', 'time-series-anomaly-detection'] | ['methodology', 'time-series'] | [ 3.81309092e-02 9.61806178e-02 -4.76983674e-02 -3.82604659e-01
2.85919756e-01 -6.28561437e-01 5.36553741e-01 5.05922139e-01
4.41452265e-01 4.40623425e-02 1.35893881e-01 -8.35586548e-01
-2.92910963e-01 -6.93959355e-01 -3.42872649e-01 -4.95804816e-01
-7.08779633e-01 2.04458162e-01 -4.18524677e-03 -2.82117844e-01
2.01698616e-01 6.17881119e-01 -1.62952256e+00 6.09087944e-01
7.73354471e-01 1.08650327e+00 -7.39305258e-01 3.36566120e-01
-3.96874070e-01 7.31831133e-01 -7.24537253e-01 5.69510944e-02
2.49587774e-01 -2.31829032e-01 -3.13918471e-01 1.32903799e-01
2.77470462e-02 -3.34122390e-01 -7.99150392e-02 9.68044639e-01
-7.83216730e-02 -4.04275209e-02 4.87729073e-01 -2.08777738e+00
-7.71163225e-01 3.38267505e-01 -4.32661533e-01 5.48752487e-01
2.86425054e-01 3.84071857e-01 1.14318764e+00 -6.21588826e-01
1.39582440e-01 1.11591196e+00 7.29338169e-01 6.40518129e-01
-1.27480531e+00 -5.68041146e-01 7.88616121e-01 5.29112041e-01
-8.05037439e-01 2.76754256e-02 9.64841485e-01 -8.75910759e-01
1.14156282e+00 6.32478714e-01 8.73578906e-01 1.39849448e+00
1.70460954e-01 7.36293197e-01 6.58672154e-01 -1.07908972e-01
5.69947481e-01 -9.36171114e-02 5.46481371e-01 2.85954326e-01
5.41233897e-01 4.37933087e-01 -2.51302332e-01 -4.60644424e-01
4.82034981e-01 9.94527578e-01 -2.01591500e-03 -4.91938055e-01
-1.00565708e+00 6.89857900e-01 2.19384685e-01 3.86735469e-01
-5.16571939e-01 -1.05780289e-01 6.37262046e-01 8.34011734e-01
4.54218268e-01 6.87077105e-01 -6.86083674e-01 -2.20468670e-01
-6.21747017e-01 1.75892726e-01 5.44482708e-01 5.56596875e-01
2.65796870e-01 5.66913366e-01 -2.40613505e-01 2.25413784e-01
1.76342517e-01 2.22100809e-01 5.61941922e-01 -4.97225702e-01
-2.02531312e-02 1.32699001e+00 2.69334346e-01 -1.20936954e+00
-5.79604626e-01 -6.46973670e-01 -7.45822251e-01 4.30608392e-01
3.09686780e-01 9.33918953e-02 -1.05676496e+00 1.50417507e+00
1.71468943e-01 2.78793544e-01 -4.35406864e-02 6.77312851e-01
2.07603928e-02 5.95759571e-01 6.05874322e-02 -4.80701178e-02
9.57977831e-01 -6.21868789e-01 -7.97721744e-01 -2.45930433e-01
1.07432950e+00 -2.15878144e-01 1.49244177e+00 5.68260312e-01
-5.09760737e-01 -4.07573193e-01 -9.73454475e-01 5.01966834e-01
-5.62095821e-01 -4.53308910e-01 6.89608395e-01 4.84800518e-01
-6.44900024e-01 5.81804097e-01 -1.21454453e+00 -4.71412420e-01
4.84336495e-01 1.07670009e-01 -1.57607436e-01 2.49599680e-01
-9.81333673e-01 7.34995782e-01 1.98449016e-01 1.77176639e-01
-9.01401520e-01 -9.11365092e-01 -6.87119067e-01 2.31863976e-01
2.19947934e-01 -2.62313336e-01 1.24760139e+00 -9.35737431e-01
-6.67718053e-01 5.81890225e-01 8.01408142e-02 -5.74775815e-01
7.46722102e-01 -4.79164630e-01 -1.24210095e+00 -3.40132803e-01
-1.38442755e-01 -1.13614887e-01 9.85739172e-01 -8.94643188e-01
-5.69277406e-01 -3.12756270e-01 -2.00791553e-01 -5.61116219e-01
-3.99939805e-01 -1.12881213e-01 4.92500156e-01 -5.33618212e-01
3.08349341e-01 -5.82215726e-01 -3.91957074e-01 1.26907483e-01
-5.51337183e-01 -2.00399563e-01 1.40493131e+00 -4.14068520e-01
1.64816928e+00 -2.40567780e+00 -5.28300107e-01 5.95347643e-01
3.93736541e-01 1.71462938e-01 -1.91656649e-02 6.57282650e-01
-6.78051412e-01 1.81284815e-01 -1.97193429e-01 2.29423121e-01
1.66789114e-01 2.54389793e-01 -1.06783581e+00 3.62119913e-01
4.39624369e-01 8.23370278e-01 -1.05736947e+00 2.44308352e-01
3.45341504e-01 3.98630351e-02 -5.58393657e-01 4.55918700e-01
-3.10176045e-01 5.76941192e-01 -4.20553774e-01 8.27884674e-01
3.11977893e-01 -4.23332244e-01 -1.66270673e-01 1.91851318e-01
-6.22322746e-02 2.05441371e-01 -9.70220864e-01 1.02408135e+00
-5.27994633e-02 6.83030367e-01 -5.04880369e-01 -1.37882197e+00
9.53023672e-01 3.35777193e-01 5.58651805e-01 -8.88060272e-01
-4.33466554e-01 1.33911133e-01 5.28083861e-01 -8.22880864e-01
9.49993432e-02 -2.55105309e-02 -3.17846909e-02 8.62875879e-01
-5.30349314e-01 5.55431724e-01 -6.20742179e-02 -4.99596447e-03
1.51341474e+00 -2.10279718e-01 2.76836544e-01 9.65192169e-02
8.31545964e-02 2.40464911e-01 8.00058246e-01 6.82057500e-01
-1.58140406e-01 4.38650101e-01 9.26110148e-01 -1.31736767e+00
-1.08954477e+00 -1.21663690e+00 -1.17878150e-02 9.78170514e-01
-3.29521388e-01 -5.51536858e-01 -1.67786658e-01 -8.71422172e-01
2.89485365e-01 1.04156864e+00 -8.44963908e-01 -8.04817498e-01
-4.83187735e-01 -6.93548501e-01 4.04551417e-01 7.75130808e-01
-4.68013659e-02 -1.25697994e+00 -1.12168634e+00 1.86893716e-01
5.46748340e-02 -7.26821482e-01 1.38736069e-01 1.11430921e-01
-1.15916252e+00 -1.15274203e+00 -3.30906846e-02 -5.64914681e-02
9.51173902e-01 5.38629331e-02 1.23821890e+00 2.91864544e-01
-5.42396426e-01 4.85053539e-01 -4.09063309e-01 -7.86082864e-01
-4.97846335e-01 -5.90048254e-01 2.97102541e-01 7.32216910e-02
9.68438745e-01 -7.53919840e-01 -7.75046051e-01 3.94071758e-01
-1.00526929e+00 -2.24010885e-01 3.45243305e-01 6.05959296e-01
2.45317176e-01 9.58332196e-02 7.76628435e-01 -8.77660632e-01
9.20801759e-01 -8.90010059e-01 -5.34479082e-01 1.95864245e-01
-1.14029419e+00 -7.29082823e-02 7.78848052e-01 -5.89025319e-01
-4.04486477e-01 -4.77084965e-01 1.64614797e-01 -6.77455604e-01
-5.58873415e-01 5.30420244e-01 1.31572962e-01 6.58605397e-01
8.91468883e-01 2.73515761e-01 1.95973925e-02 -5.20036161e-01
1.28705338e-01 3.84875476e-01 2.92016447e-01 -4.20092255e-01
6.96595371e-01 6.41112983e-01 -2.13825762e-01 -5.81199288e-01
-6.83363914e-01 -3.96233082e-01 -3.94525707e-01 -6.48486093e-02
3.17009449e-01 -5.05908549e-01 -6.31097257e-01 1.58611700e-01
-9.59589660e-01 -2.74668783e-01 -7.92620540e-01 -1.06459809e-02
-3.24177742e-01 3.17603707e-01 -2.13261306e-01 -1.00833464e+00
-1.76142320e-01 -5.76461911e-01 6.38880014e-01 -2.11966604e-01
-9.69601035e-01 -1.05431068e+00 2.23797560e-01 -3.73232663e-01
6.28733873e-01 5.29861152e-01 1.50951803e+00 -1.43555546e+00
-3.87476474e-01 -8.30367804e-01 -5.82162663e-03 2.90339023e-01
4.19544846e-01 1.93064019e-01 -1.05712831e+00 -2.65689343e-01
1.24715380e-01 2.10956350e-01 3.47739726e-01 1.40078470e-01
1.70604920e+00 -5.95987916e-01 -3.31465542e-01 3.87212873e-01
7.42400944e-01 4.89706337e-01 4.32715446e-01 4.28313613e-01
5.98308742e-01 8.44745040e-01 5.81178010e-01 7.90793955e-01
1.75976884e-02 2.34435961e-01 7.21679211e-01 -1.96864113e-01
5.78035057e-01 -2.52005577e-01 4.35994983e-01 4.02333170e-01
1.65873453e-01 2.72524245e-02 -1.49889505e+00 6.34958565e-01
-2.22058320e+00 -1.27974057e+00 -1.15987577e-01 2.27769136e+00
2.43692309e-01 4.47972983e-01 2.91001976e-01 5.14105976e-01
5.27323961e-01 9.78439301e-02 -9.91191328e-01 -6.86312318e-01
3.33276726e-02 -3.05161774e-01 -3.04007977e-01 -3.74575220e-02
-1.01821411e+00 3.76481295e-01 6.63983011e+00 -7.86689762e-03
-1.36688101e+00 -2.42403790e-01 5.69142103e-01 -1.90138698e-01
-4.53752339e-01 -1.02014609e-01 -2.10565537e-01 5.73886514e-01
1.04927313e+00 -3.96370023e-01 2.87255645e-02 1.12630725e+00
4.17234719e-01 5.30632973e-01 -1.89360690e+00 9.28567350e-01
-3.77904713e-01 -1.25075746e+00 3.08334798e-01 1.75057128e-01
3.93722415e-01 -2.69345846e-02 4.45471883e-01 3.92216980e-01
1.82942152e-01 -1.22966778e+00 5.83589315e-01 4.76854682e-01
3.20283175e-01 -3.22930247e-01 5.06660998e-01 4.03086096e-01
-7.25192308e-01 -8.66857052e-01 -1.62831262e-01 -5.39099276e-01
-1.92191098e-02 7.83750713e-01 -8.00310373e-01 3.24036777e-02
7.39645064e-01 8.78071189e-01 -5.04756451e-01 1.09675658e+00
-1.11361980e-01 9.21205461e-01 -2.59661138e-01 4.38598871e-01
1.90807238e-01 -5.34381084e-02 7.43257344e-01 9.77546871e-01
5.73571622e-01 -3.02110910e-01 1.79033071e-01 1.19925606e+00
5.52888095e-01 -3.00017864e-01 -1.06230152e+00 -4.42649245e-01
2.39891961e-01 1.02881587e+00 -5.83165705e-01 -3.06082666e-01
-5.54290354e-01 5.37983894e-01 -1.79156125e-01 6.09079063e-01
-5.82184136e-01 -9.68418196e-02 1.03997743e+00 4.06882644e-01
-2.12727994e-01 -1.10879377e-01 -5.92702389e-01 -1.19683945e+00
4.32371199e-01 -1.16349316e+00 7.95993805e-01 -4.26401764e-01
-1.55344558e+00 4.41655755e-01 -8.34138766e-02 -1.90313101e+00
-6.47721946e-01 -6.52964890e-01 -1.42418408e+00 5.61190963e-01
-1.11076868e+00 -8.39542210e-01 -2.91222364e-01 6.88655853e-01
5.11185110e-01 -2.32396707e-01 1.04523051e+00 6.88258857e-02
-6.82938337e-01 3.34219813e-01 4.75636125e-02 1.77345708e-01
2.55833477e-01 -1.32104897e+00 9.81548667e-01 1.02522171e+00
2.23230079e-01 7.87724137e-01 9.90828574e-01 -5.82453132e-01
-9.32871580e-01 -1.10439134e+00 4.97318059e-01 -7.78225183e-01
1.05827761e+00 -3.66177261e-01 -1.54399192e+00 8.40469658e-01
-2.48246819e-01 2.34600410e-01 1.08647227e+00 5.44382691e-01
-6.68045044e-01 -3.91017869e-02 -9.76485789e-01 8.37446272e-01
1.00366580e+00 -5.53679168e-01 -7.96060503e-01 3.17269653e-01
5.57589471e-01 1.13936186e-01 -2.81861037e-01 5.23619115e-01
4.33354706e-01 -1.07401490e+00 7.97565460e-01 -1.52845705e+00
2.00617269e-01 -2.72751123e-01 4.91440073e-02 -1.49516416e+00
-3.98011923e-01 -6.42555416e-01 -8.28649521e-01 6.41688585e-01
2.97637224e-01 -9.26253378e-01 6.66211367e-01 1.08895099e+00
-3.74314547e-01 -7.68112957e-01 -7.03567743e-01 -9.87955332e-01
-1.52030110e-01 -8.39259624e-01 1.06137872e+00 1.26642072e+00
4.29654598e-01 -2.40818813e-01 -1.49008542e-01 4.67264473e-01
4.92212415e-01 4.62384641e-01 6.06269479e-01 -1.75315607e+00
-6.42249063e-02 -5.98226428e-01 -7.47059822e-01 -4.43972498e-01
-2.06649438e-01 -5.92065036e-01 -5.57391226e-01 -1.22059643e+00
-3.30923855e-01 -3.58047873e-01 -1.01215124e+00 7.64499545e-01
1.02300189e-01 -3.20718408e-01 -1.81062251e-01 3.78295451e-01
-4.43078279e-01 3.51065040e-01 5.97727239e-01 -3.23158264e-01
-3.17602336e-01 2.56733209e-01 -6.59219325e-01 1.00679505e+00
9.62794781e-01 -4.08465683e-01 -6.95698798e-01 -3.88600945e-01
7.15627551e-01 -2.89299339e-01 7.24031210e-01 -9.37316895e-01
-1.90914907e-02 -4.16927993e-01 4.67488140e-01 -5.25244057e-01
8.28259438e-03 -1.13428795e+00 5.22662625e-02 8.20743859e-01
-4.50579852e-01 7.35866785e-01 3.12350899e-01 8.25761557e-01
-3.09181064e-01 3.93695146e-01 2.86050379e-01 5.06171398e-02
-1.00905728e+00 3.88461798e-01 -4.60393190e-01 -1.17964581e-01
1.21507084e+00 -5.56580305e-01 -4.67187107e-01 -4.71041024e-01
-1.08782470e+00 3.44344079e-01 4.34833467e-01 9.35751796e-01
7.79970109e-01 -1.26327968e+00 -3.12818974e-01 8.45774174e-01
7.16635108e-01 -1.37316525e-01 3.50669146e-01 8.61540735e-01
-1.92339107e-01 2.81419218e-01 -4.78707433e-01 -9.16990697e-01
-8.19023192e-01 8.33592534e-01 4.16117430e-01 6.10402822e-02
-1.15867066e+00 2.02585965e-01 4.30447698e-01 -2.78312534e-01
4.15841967e-01 -5.91011405e-01 -4.25795428e-02 -1.45030588e-01
7.72259593e-01 4.81774002e-01 1.26501769e-01 2.80861789e-03
-3.95794243e-01 3.60706225e-02 -3.48725349e-01 3.10632735e-01
1.30890477e+00 1.20982312e-01 -5.47055937e-02 9.91061985e-01
6.75027788e-01 -4.42653686e-01 -1.16859460e+00 -2.81632822e-02
7.06349730e-01 -5.53351939e-01 -4.44399476e-01 -8.88042569e-01
-7.49415576e-01 1.33778203e+00 8.83931875e-01 1.07385004e+00
9.42866564e-01 -4.58565541e-02 5.22103310e-01 6.41821325e-01
2.97554843e-02 -9.83095229e-01 2.59483308e-01 3.27180624e-01
9.66528118e-01 -1.37362027e+00 -4.31305081e-01 2.06805721e-01
-5.70414007e-01 1.05026233e+00 9.03909326e-01 -2.48679332e-02
7.74569571e-01 6.88157044e-03 4.97830093e-01 -5.63721359e-01
-1.12150586e+00 2.54532188e-01 2.56666183e-01 7.46702850e-01
2.73859769e-01 1.00205079e-01 2.80579388e-01 6.67851508e-01
-9.31831822e-02 -3.00791889e-01 4.91070449e-01 9.98787045e-01
-3.51155639e-01 -7.72965431e-01 -3.55082065e-01 7.73591697e-01
-5.28029740e-01 2.72875190e-01 -3.48399252e-01 6.42133653e-01
-1.33012697e-01 8.44789445e-01 5.40532947e-01 -3.95783216e-01
4.73555773e-01 4.53504294e-01 -2.69790232e-01 -7.06815124e-01
-2.75962532e-01 -2.94967741e-01 -3.01596075e-01 -8.22459638e-01
3.69090408e-01 -4.28429484e-01 -1.31823611e+00 -1.86482608e-01
1.55769676e-01 2.20661774e-01 4.51767981e-01 8.96530151e-01
8.70536506e-01 5.08293986e-01 7.37758398e-01 -1.59982875e-01
-8.21916044e-01 -7.91587293e-01 -4.76649851e-01 8.83108318e-01
8.54362845e-01 -6.12113416e-01 -5.72463334e-01 -1.24192290e-01] | [7.510058879852295, 2.562950849533081] |
76092173-e94f-4846-8407-ed76be14ce9a | how-self-supervised-learning-can-be-used-for | 2108.04893 | null | https://arxiv.org/abs/2108.04893v6 | https://arxiv.org/pdf/2108.04893v6.pdf | How Self-Supervised Learning Can be Used for Fine-Grained Head Pose Estimation? | The cost of head pose labeling is the main challenge of improving the fine-grained Head Pose Estimation (HPE). Although Self-Supervised Learning (SSL) can be a solution to the lack of huge amounts of labeled data, its efficacy for fine-grained HPE is not yet fully explored. This study aims to assess the usage of SSL in fine-grained HPE based on two scenarios: (1) using SSL for weights pre-training procedure, and (2) leveraging auxiliary SSL losses besides HPE. We design a Hybrid Multi-Task Learning (HMTL) architecture based on the ResNet50 backbone in which both strategies are applied. Our experimental results reveal that the combination of both scenarios is the best for HPE. Together, the average error rate is reduced up to 23.1% for AFLW2000 and 14.2% for BIWI benchmark compared to the baseline. Moreover, it is found that some SSL methods are more suitable for transfer learning, while others may be effective when they are considered as auxiliary tasks incorporated into supervised learning. Finally, it is shown that by using the proposed HMTL architecture, the average error is reduced with different types of initial weights: random, ImageNet and SSL pre-trained weights. | ['Seyedehsamaneh Shojaeilangari', 'Sasan Karamizadeh', 'Farzaneh Esmaili', 'Ebrahim Mousavi', 'Mahdi Pourmirzaei'] | 2021-08-10 | null | null | null | null | ['head-pose-estimation'] | ['computer-vision'] | [-1.00960601e-02 4.39159840e-01 -6.85250610e-02 -4.56627846e-01
-9.22265530e-01 1.72734596e-02 5.11205733e-01 4.51305099e-02
-9.16714728e-01 9.32158470e-01 3.93876493e-01 1.04647405e-01
-1.60383016e-01 -5.73450565e-01 -8.75869930e-01 -8.98825943e-01
4.23591062e-02 7.44496882e-01 3.65652353e-01 -4.26967829e-01
-1.01039276e-01 6.17207348e-01 -1.80094743e+00 -2.87761707e-02
7.96496987e-01 1.03842902e+00 2.79161870e-01 1.40296504e-01
-5.85814714e-02 4.26874250e-01 -5.51033854e-01 -4.25238311e-01
1.51399568e-01 1.28708035e-01 -7.28994966e-01 -1.01689182e-01
5.47204435e-01 -3.01426172e-01 1.51626706e-01 8.27792346e-01
1.09973192e+00 1.06390752e-01 5.70683599e-01 -1.35734439e+00
1.14718534e-01 5.93511820e-01 -5.59564292e-01 3.44830416e-02
2.58891255e-01 5.37545606e-02 7.91833162e-01 -9.72399533e-01
3.56945664e-01 1.35084581e+00 8.77846301e-01 7.98793375e-01
-9.96129155e-01 -8.42858791e-01 2.49386638e-01 3.81480515e-01
-1.46540117e+00 -6.75305843e-01 6.76074803e-01 -6.92713559e-02
9.78519559e-01 6.75306991e-02 4.10047710e-01 1.10322249e+00
1.55831486e-01 1.02248967e+00 1.42587292e+00 -5.75116098e-01
7.10595474e-02 2.08565608e-01 3.84638578e-01 6.26777649e-01
3.37915659e-01 1.74715623e-01 -6.78915143e-01 -5.94160222e-02
1.97686926e-01 -3.80046546e-01 -3.66920233e-01 -1.16711274e-01
-6.49846733e-01 8.89804184e-01 5.56004941e-01 4.07887518e-01
-4.17465150e-01 -1.89399078e-01 5.81846416e-01 8.11509863e-02
7.85433352e-01 2.38755062e-01 -5.76185703e-01 2.07786188e-01
-1.12215221e+00 3.45062107e-01 6.28118753e-01 7.31204450e-01
7.39800990e-01 -1.51191978e-02 -3.23544681e-01 1.06319880e+00
4.86546636e-01 4.71111119e-01 7.15708673e-01 -2.62792796e-01
5.25444806e-01 4.54371393e-01 -7.40732700e-02 -5.42013049e-01
-8.88901591e-01 -5.29573321e-01 -7.44261622e-01 1.09610900e-01
4.69934106e-01 -1.45648092e-01 -1.11391366e+00 2.01092029e+00
4.09383804e-01 1.42836183e-01 -2.03059673e-01 6.23562038e-01
1.15800202e+00 4.82970089e-01 4.65584576e-01 -1.76554650e-01
1.27743888e+00 -1.16141868e+00 -9.26157117e-01 -5.09098887e-01
6.06778920e-01 -6.92284107e-01 9.06253576e-01 1.88899547e-01
-1.13132489e+00 -5.97463310e-01 -1.04529881e+00 -4.15142952e-03
-5.24095774e-01 2.64161438e-01 4.53343540e-01 9.30443347e-01
-1.35704482e+00 4.29104090e-01 -7.05550969e-01 -4.02657449e-01
3.86874586e-01 7.58635998e-01 -3.89757544e-01 -6.28976300e-02
-1.30839980e+00 1.25689471e+00 5.60086071e-01 3.72881770e-01
-5.68576872e-01 -7.50740647e-01 -8.83626819e-01 1.21385232e-01
1.11165337e-01 -5.95938146e-01 1.14296603e+00 -4.26134646e-01
-1.66229057e+00 9.30746853e-01 -1.28676012e-01 -5.64010084e-01
7.79492497e-01 -4.52408195e-01 -1.62994727e-01 -1.02802798e-01
8.56518373e-02 1.02715647e+00 7.70019531e-01 -1.19862986e+00
-7.32005656e-01 -6.43142700e-01 -7.53812417e-02 2.91512191e-01
-4.15074468e-01 3.79906571e-03 -4.29089934e-01 -5.25197566e-01
-9.55393687e-02 -1.18648279e+00 -8.90625566e-02 -3.39201093e-01
-3.47680330e-01 -7.39018798e-01 7.65934348e-01 -8.45291257e-01
1.24761820e+00 -1.89499414e+00 -2.35334374e-02 1.79651350e-01
7.93798640e-02 5.98744690e-01 -8.69263187e-02 -2.19683126e-02
-8.40874314e-02 -2.46958107e-01 2.97628865e-02 -7.02768624e-01
9.35499072e-02 5.03598973e-02 2.63554096e-01 4.73935902e-01
-3.03247217e-02 9.87497866e-01 -4.47682649e-01 -5.29146969e-01
3.90198499e-01 5.84032536e-01 -5.62804580e-01 2.36475393e-01
-9.40524265e-02 4.52453375e-01 -1.11208640e-01 3.81549567e-01
9.45643723e-01 -1.75270960e-02 -2.50208303e-02 -5.21518469e-01
5.06738462e-02 2.71494478e-01 -1.27575696e+00 1.42358601e+00
-6.87560678e-01 2.22302169e-01 4.18883301e-02 -7.70357549e-01
7.09255099e-01 3.58062327e-01 1.89377442e-01 -8.61541450e-01
2.15221077e-01 3.03491384e-01 1.26014277e-03 -1.61130607e-01
4.27349597e-01 -1.41308874e-01 -4.36642133e-02 2.78455496e-01
3.98344934e-01 1.24763921e-01 1.09853126e-01 -2.33370915e-01
6.78035855e-01 1.41962931e-01 3.26528758e-01 -4.44672316e-01
6.68997347e-01 -5.12623072e-01 6.20889068e-01 5.27838826e-01
-3.34180057e-01 3.95264238e-01 -8.99814740e-02 -2.11757615e-01
-7.39396036e-01 -7.74939299e-01 -2.61286736e-01 1.57592118e+00
-1.34020939e-01 -1.31757081e-01 -1.07033968e+00 -7.16619194e-01
-1.11754246e-01 7.72456765e-01 -6.15561068e-01 -1.40805751e-01
-8.41290236e-01 -1.13636804e+00 5.46358764e-01 6.44688189e-01
9.50930536e-01 -1.27245998e+00 -4.59942728e-01 2.26721004e-01
-3.55029374e-01 -1.13663471e+00 -4.24055368e-01 4.52232152e-01
-7.35048234e-01 -7.57629037e-01 -1.11705971e+00 -7.32973516e-01
5.80520451e-01 5.12156598e-02 1.00909495e+00 8.58990103e-02
1.00968987e-01 1.65416017e-01 -2.97481000e-01 -5.03916025e-01
-2.02969119e-01 5.80802917e-01 8.36237371e-02 2.49313898e-02
4.56418067e-01 -3.69837195e-01 -6.45281732e-01 3.77685457e-01
-5.26008427e-01 -9.33250859e-02 8.29376340e-01 9.23462749e-01
2.92180270e-01 -2.66228557e-01 6.94251120e-01 -1.08681953e+00
4.78534102e-01 -2.64778256e-01 -3.48305643e-01 3.48105013e-01
-7.58062541e-01 2.30492264e-01 3.27314407e-01 -3.81715178e-01
-1.42717624e+00 -1.67831443e-02 -8.29273880e-01 -1.34978116e-01
-3.85912359e-01 1.68322816e-01 -3.42520922e-01 -2.85566032e-01
4.37967271e-01 1.96531937e-02 -6.33824989e-02 -6.07307613e-01
8.18531811e-02 6.55042052e-01 1.58861399e-01 -4.32806104e-01
4.74931300e-01 1.79067940e-01 -9.62096974e-02 -8.07304561e-01
-9.53607619e-01 -5.21966159e-01 -5.85556686e-01 -1.34282127e-01
8.87008369e-01 -9.17019248e-01 -7.36182988e-01 7.04420567e-01
-8.64278972e-01 -4.96363312e-01 -3.32530700e-02 3.53144526e-01
-3.66299778e-01 2.04540268e-01 -6.38792515e-01 -5.94865322e-01
-6.78131223e-01 -1.36040592e+00 1.36068738e+00 1.62428617e-01
-1.75870180e-01 -9.40614462e-01 -1.70928061e-01 5.60391784e-01
6.02765083e-01 -9.70308259e-02 7.93123007e-01 -7.93040097e-01
-1.34617671e-01 -1.62539050e-01 -1.82202548e-01 1.76609516e-01
-1.27156496e-01 -5.51876724e-01 -1.33580613e+00 -7.04328656e-01
-6.74242154e-02 -5.14492571e-01 8.58482301e-01 6.44865572e-01
8.44288528e-01 -1.58088818e-01 -3.23440224e-01 4.43787515e-01
1.19744658e+00 -3.72380093e-02 8.18360627e-01 6.95569396e-01
5.89330375e-01 7.69080520e-01 5.35457551e-01 2.82533884e-01
8.37331355e-01 1.11375046e+00 2.95077115e-01 -2.55547702e-01
-5.31010628e-01 -2.12095782e-01 2.92360693e-01 8.62574518e-01
-2.23482460e-01 -6.11449704e-02 -7.92890966e-01 2.89708853e-01
-1.64201391e+00 -6.55137420e-01 3.36720720e-02 2.27363110e+00
7.38601387e-01 3.46248150e-01 3.18954170e-01 4.72383976e-01
7.89871097e-01 1.12479188e-01 -4.08205330e-01 -1.73578590e-01
-4.42550108e-02 4.58339632e-01 6.31563485e-01 3.96692902e-01
-1.34678972e+00 1.06439757e+00 5.49634790e+00 1.07349122e+00
-1.22675788e+00 5.58348060e-01 4.47283566e-01 2.46885404e-01
2.08256483e-01 -5.90055048e-01 -1.48302829e+00 4.97878224e-01
1.09366930e+00 3.19489509e-01 -1.48198055e-02 8.00516546e-01
8.38109404e-02 -1.45141631e-01 -7.72403479e-01 9.53607678e-01
2.14360878e-01 -7.83210874e-01 -9.49168876e-02 -1.93365719e-02
4.77637231e-01 9.69737992e-02 -2.45535895e-02 6.39653683e-01
6.92925677e-02 -8.79383862e-01 8.62081885e-01 7.90858120e-02
6.36081040e-01 -7.33703077e-01 1.08236885e+00 4.26484793e-01
-1.28588402e+00 -2.34937165e-02 -1.28684491e-01 2.77974188e-01
2.16338545e-01 3.24006051e-01 -8.71378481e-01 3.76576245e-01
8.45316947e-01 1.09471411e-01 -8.38288426e-01 1.21354425e+00
-5.30431390e-01 5.77950716e-01 -6.05080843e-01 -5.72907031e-02
1.53564796e-01 4.23105717e-01 2.43030190e-01 1.28548133e+00
6.99042007e-02 -2.61438996e-01 1.55005306e-01 2.19993964e-01
3.23233120e-02 2.54781604e-01 -2.51882911e-01 6.58934116e-01
3.88901174e-01 1.24190712e+00 -6.74943328e-01 -4.39750254e-02
-4.13075089e-01 8.85974228e-01 4.64868605e-01 1.69183582e-01
-7.82910943e-01 -2.32647002e-01 3.44119757e-01 2.84083009e-01
2.63049573e-01 1.24810994e-01 -1.18829027e-01 -7.88181663e-01
-1.63031057e-01 -6.84324563e-01 4.33881700e-01 -5.23471236e-01
-1.11939216e+00 9.55925941e-01 2.29361281e-01 -9.60663855e-01
-4.21345860e-01 -6.25591874e-01 -3.06709677e-01 6.94830418e-01
-1.91436529e+00 -1.57340097e+00 -3.63313496e-01 6.88460469e-01
7.46679246e-01 1.37170022e-02 9.57840145e-01 7.48387754e-01
-6.65667951e-01 1.25801277e+00 -3.40403676e-01 -2.32802704e-01
6.64026201e-01 -1.21591687e+00 2.17471480e-01 5.20649254e-01
-2.23576829e-01 4.39184636e-01 8.21163654e-01 -4.95693207e-01
-9.59630489e-01 -1.22830951e+00 1.11053503e+00 5.06891757e-02
1.46739662e-01 -4.52699035e-01 -9.53911364e-01 7.19206393e-01
1.95136756e-01 2.19031312e-02 5.25590241e-01 2.32748225e-01
-1.48854941e-01 -1.50080293e-01 -1.47888005e+00 3.93890500e-01
8.15570891e-01 -1.78963244e-01 -7.37599075e-01 2.90145665e-01
5.26290417e-01 -5.59745431e-01 -8.35852325e-01 7.52528548e-01
4.93306786e-01 -8.90024722e-01 9.84535456e-01 -1.11272097e-01
1.08430078e-02 -7.06380457e-02 -4.87016030e-02 -1.60962510e+00
-2.67249048e-01 -1.87177017e-01 1.59978360e-01 1.27680326e+00
5.10325611e-01 -8.03286076e-01 1.04721057e+00 5.66353440e-01
-2.20256940e-01 -7.70243704e-01 -1.15262282e+00 -7.46470273e-01
9.65619311e-02 -3.27645898e-01 5.34714341e-01 5.68782210e-01
-4.87791389e-01 5.64049304e-01 -4.46051598e-01 5.21413945e-02
8.37758005e-01 -2.88191885e-01 7.12790489e-01 -1.32135332e+00
-2.04764176e-02 -3.22194785e-01 -2.94047266e-01 -8.40148211e-01
5.20818412e-01 -8.12154889e-01 2.30522856e-01 -1.43891025e+00
2.09365815e-01 -5.73621035e-01 -2.94595510e-01 6.71609759e-01
-3.34440947e-01 3.41468543e-01 4.22046959e-01 -3.78076322e-02
-4.56297159e-01 6.43638015e-01 1.01300895e+00 -1.04219792e-02
-2.07054675e-01 3.89721125e-01 -2.90389836e-01 1.00849056e+00
7.41239905e-01 -3.92649680e-01 -2.57572025e-01 -1.79986686e-01
-2.55336791e-01 -5.46186902e-02 3.94233353e-02 -1.05247331e+00
4.17723328e-01 4.04876173e-01 3.05557251e-01 -8.36936951e-01
4.94094729e-01 -7.41138339e-01 -8.10066164e-02 4.55055356e-01
-1.14028379e-01 -5.30186854e-02 5.04916489e-01 1.51913509e-01
-1.78620830e-01 -3.88222367e-01 1.03180695e+00 -7.48131201e-02
-8.99122238e-01 1.00297950e-01 -7.72068724e-02 4.31073904e-02
7.46205449e-01 -2.23054543e-01 -7.41530582e-02 -2.33934879e-01
-8.36369216e-01 1.55112103e-01 -1.58821136e-01 2.83140004e-01
3.57148290e-01 -1.27876484e+00 -6.61593854e-01 2.48769730e-01
1.26968041e-01 -1.77328750e-01 4.11581069e-01 8.99441421e-01
-1.96416467e-01 6.56431377e-01 -2.86288083e-01 -4.96997625e-01
-1.43538368e+00 1.45822614e-01 1.29710466e-01 -7.79282808e-01
-4.88257915e-01 1.09294474e+00 1.61646754e-01 -7.70400703e-01
7.32608318e-01 -1.54667720e-01 -5.81850350e-01 4.26508904e-01
4.23931330e-01 4.99897659e-01 6.47640884e-01 -9.84113336e-01
-4.88752753e-01 5.97455919e-01 -3.58614206e-01 9.31720361e-02
1.53024030e+00 -2.04106674e-01 1.36176333e-01 9.97373909e-02
1.24429965e+00 -2.09717840e-01 -1.01307237e+00 -2.30278596e-01
2.43882447e-01 -2.25018840e-02 2.72915810e-01 -8.95646691e-01
-1.23386133e+00 8.46682847e-01 1.02472460e+00 -3.01019639e-01
1.25779891e+00 -5.84107712e-02 9.15222406e-01 1.15602158e-01
8.54723573e-01 -9.77471173e-01 -2.16731206e-02 4.52897966e-01
8.64849031e-01 -1.32391977e+00 -8.55103601e-03 -4.72505242e-01
-4.98687863e-01 7.82927096e-01 7.55125880e-01 2.59880605e-03
7.00571418e-01 3.38475913e-01 -1.40603244e-01 -6.88926503e-02
-3.89378875e-01 -4.91526335e-01 4.28496569e-01 3.76876533e-01
5.21651387e-01 8.72162655e-02 -5.14919817e-01 6.75866425e-01
-3.90716493e-01 9.25043449e-02 6.98488057e-02 8.00944865e-01
-3.90879095e-01 -1.13350189e+00 -4.31320101e-01 5.38143277e-01
-6.17320478e-01 -3.73975746e-02 2.24588573e-01 9.65014577e-01
2.80983478e-01 8.23683858e-01 -2.60745615e-01 -2.06830278e-01
6.62964225e-01 1.71217158e-01 6.86833322e-01 -5.52916586e-01
-7.94894338e-01 2.20689997e-01 1.37327924e-01 -4.98717219e-01
-5.37165165e-01 -4.69307780e-01 -9.50259268e-01 -1.40630588e-01
-6.32190526e-01 7.29312971e-02 7.21402466e-01 1.16368937e+00
8.82063136e-02 5.28451800e-01 3.50176901e-01 -1.28527188e+00
-5.99206865e-01 -1.27569807e+00 -5.90454221e-01 2.73477882e-01
2.10073873e-01 -1.22499490e+00 -5.19446254e-01 -3.58223796e-01] | [13.699713706970215, 0.2798609137535095] |
1cb5fd30-8172-494d-aee0-b25d8768c861 | a-brief-summary-of-interactions-between-meta | 2103.00845 | null | https://arxiv.org/abs/2103.00845v2 | https://arxiv.org/pdf/2103.00845v2.pdf | A Brief Summary of Interactions Between Meta-Learning and Self-Supervised Learning | This paper briefly reviews the connections between meta-learning and self-supervised learning. Meta-learning can be applied to improve model generalization capability and to construct general AI algorithms. Self-supervised learning utilizes self-supervision from original data and extracts higher-level generalizable features through unsupervised pre-training or optimization of contrastive loss objectives. In self-supervised learning, data augmentation techniques are widely applied and data labels are not required since pseudo labels can be estimated from trained models on similar tasks. Meta-learning aims to adapt trained deep models to solve diverse tasks and to develop general AI algorithms. We review the associations of meta-learning with both generative and contrastive self-supervised learning models. Unlabeled data from multiple sources can be jointly considered even when data sources are vastly different. We show that an integration of meta-learning and self-supervised learning models can best contribute to the improvement of model generalization capability. Self-supervised learning guided by meta-learner and general meta-learning algorithms under self-supervision are both examples of possible combinations. | ['Huimin Peng'] | 2021-03-01 | null | null | null | null | ['unsupervised-pre-training'] | ['methodology'] | [ 4.11421478e-01 4.62725520e-01 -7.12215543e-01 -7.85260499e-01
-4.27842349e-01 -2.77391165e-01 8.15350235e-01 3.17684263e-01
-3.50028604e-01 8.04289877e-01 -1.43748417e-01 2.94710070e-01
-1.85518220e-01 -9.49356437e-01 -6.42274797e-01 -5.51861703e-01
-2.49079298e-02 7.87307978e-01 -1.82815775e-01 -2.36955374e-01
8.16081688e-02 3.08980227e-01 -1.90146053e+00 5.45892537e-01
1.29510319e+00 7.48454094e-01 4.01891172e-01 4.19494480e-01
-6.82758808e-01 7.50264466e-01 -6.41445160e-01 -2.29135290e-01
1.66194767e-01 -6.53361440e-01 -9.65580702e-01 4.63621438e-01
4.36304331e-01 3.06532055e-01 -1.06287636e-01 1.02158797e+00
1.04564995e-01 1.53055221e-01 1.06573176e+00 -1.44167936e+00
-6.72820270e-01 6.36125505e-01 -2.12566108e-01 2.94812590e-01
-9.26086009e-02 1.13441783e-03 7.89490938e-01 -7.88382351e-01
4.42570895e-01 1.19840825e+00 6.25956237e-01 8.75095904e-01
-1.28205776e+00 -4.86666590e-01 -5.76544506e-03 2.12460294e-01
-9.46751475e-01 -3.47753704e-01 1.00076580e+00 -2.69891113e-01
8.45896721e-01 2.09161662e-03 5.84316611e-01 1.17108428e+00
1.38519898e-01 1.13269949e+00 1.21103513e+00 -9.27130401e-01
3.89701873e-01 6.38476074e-01 8.93658698e-02 9.01472926e-01
9.37099010e-02 4.63235021e-01 -7.98648536e-01 -1.73461989e-01
4.47422296e-01 3.26654106e-01 2.90596545e-01 -5.80366731e-01
-1.26261652e+00 9.08842921e-01 4.91639197e-01 6.12830102e-01
-3.24667603e-01 -5.14273904e-02 4.55130458e-01 5.69431663e-01
7.64890909e-01 1.04763317e+00 -6.74826562e-01 3.36372346e-01
-1.14986801e+00 1.84944235e-02 4.71184045e-01 1.16354620e+00
1.34866321e+00 5.91055870e-01 7.66156763e-02 1.10879457e+00
2.81252265e-01 4.93475586e-01 1.15146089e+00 -9.16651309e-01
2.87962407e-01 9.06195164e-01 -3.71017784e-01 -3.38195920e-01
-4.86991614e-01 -1.05143034e+00 -7.95854986e-01 2.01539874e-01
-1.36014462e-01 -2.44062245e-01 -1.11721241e+00 1.65337694e+00
8.70143026e-02 -9.25922692e-02 4.26892221e-01 2.23512337e-01
8.28248739e-01 7.05421507e-01 3.12878728e-01 -5.08980870e-01
7.46226907e-01 -1.43302810e+00 -6.70142293e-01 -6.09757483e-01
9.62275743e-01 -1.66599289e-01 7.72164822e-01 2.89592147e-01
-1.20523036e+00 -9.30774152e-01 -1.14503598e+00 4.43494797e-01
-8.17514598e-01 5.01086004e-02 6.76199853e-01 6.07557535e-01
-8.49744022e-01 7.61893988e-01 -7.24796176e-01 -2.74800241e-01
6.26542330e-01 5.26993573e-01 -2.54540980e-01 1.69093296e-01
-1.05293190e+00 9.98404682e-01 8.48787904e-01 -2.83126980e-01
-1.04657066e+00 -6.40919268e-01 -1.01669216e+00 -7.08344728e-02
3.38091254e-01 -8.35779786e-01 1.16745305e+00 -1.75646400e+00
-1.30370641e+00 1.09454286e+00 -1.10671595e-01 -6.59154892e-01
2.28635609e-01 -4.06844579e-02 -6.57082140e-01 8.94920528e-02
2.38338441e-01 8.35943401e-01 1.29028761e+00 -1.45165038e+00
-8.31390679e-01 -5.25393128e-01 -5.23421586e-01 3.28824282e-01
-6.97023451e-01 -4.84038591e-01 2.01743156e-01 -8.38746846e-01
1.42730355e-01 -8.25092375e-01 -2.55406857e-01 -6.75139368e-01
-2.43072808e-01 -2.75198370e-01 8.85485649e-01 1.07713431e-01
8.27138245e-01 -1.84751463e+00 1.32806882e-01 6.59361184e-02
2.67295897e-01 4.97305751e-01 -4.20959830e-01 3.82713437e-01
-1.79690450e-01 -1.41590564e-02 -4.66828525e-01 -2.71101624e-01
-3.85221332e-01 3.39976102e-01 -1.90262437e-01 -7.40179280e-03
4.29913908e-01 1.20838261e+00 -1.13701117e+00 -4.74897295e-01
2.33206674e-01 -2.56280582e-02 -1.80727854e-01 5.05943775e-01
-5.12637854e-01 4.32695448e-01 -4.15016711e-01 5.17372072e-01
1.62695110e-01 -3.51441205e-01 8.69001299e-02 5.58115616e-02
1.41830713e-01 2.39283219e-01 -8.18047047e-01 1.82199526e+00
-7.81056404e-01 6.01102471e-01 -4.87188399e-01 -1.57763720e+00
1.19117224e+00 1.77844584e-01 5.27738094e-01 -8.35533202e-01
-1.63995102e-02 1.77427083e-01 -1.14817493e-01 -6.35571539e-01
3.07844430e-01 -3.12789649e-01 1.87815905e-01 6.22876346e-01
7.78869033e-01 -1.72755927e-01 2.58447319e-01 4.05251533e-02
6.32359505e-01 2.41574943e-01 5.48053443e-01 -3.28406900e-01
4.97441113e-01 4.67232913e-01 5.35083115e-01 9.45986092e-01
2.14090049e-01 1.76950708e-01 -2.65580028e-01 -6.23321176e-01
-9.50605690e-01 -9.82946396e-01 -1.11943699e-01 1.63784027e+00
-1.50587872e-01 -4.34010983e-01 -5.69822490e-01 -1.08614802e+00
-3.22401039e-02 7.85285354e-01 -8.02897334e-01 -5.92298329e-01
-2.45500699e-01 -9.49743986e-01 1.76593930e-01 7.17960238e-01
6.41725183e-01 -1.31320381e+00 -5.83766103e-01 4.24577981e-01
3.97038221e-01 -6.53209805e-01 1.35316029e-01 8.53510082e-01
-1.57233560e+00 -8.75323713e-01 -6.33058965e-01 -1.05979180e+00
1.01127601e+00 3.01801175e-01 1.28318870e+00 2.85032019e-02
-2.24109516e-01 6.38551235e-01 -4.48492616e-01 -6.99692428e-01
-1.01581419e+00 3.95416886e-01 3.50677699e-01 8.28446820e-02
4.00211096e-01 -6.53991401e-01 1.28936525e-02 1.85296655e-01
-1.03034806e+00 2.38860641e-02 6.86383545e-01 1.27458489e+00
5.34892082e-01 1.91005602e-01 1.22210217e+00 -1.50157785e+00
3.87469620e-01 -6.26948774e-01 -3.59933943e-01 5.04118204e-01
-1.18897188e+00 5.73566079e-01 7.59573460e-01 -5.98679006e-01
-1.33940136e+00 1.89988941e-01 2.79225409e-01 -4.48142469e-01
-5.09226084e-01 4.77344066e-01 -2.21291915e-01 -2.85492167e-02
1.12902844e+00 4.08835769e-01 -1.35947280e-02 -2.96153158e-01
5.29881954e-01 7.49752045e-01 4.24253404e-01 -5.58276832e-01
6.45757794e-01 3.93229514e-01 2.14121472e-02 -7.79500723e-01
-1.34262753e+00 -3.44935507e-01 -1.01754296e+00 -1.19304046e-01
4.71577853e-01 -7.80391395e-01 -2.15631002e-03 3.50542575e-01
-6.54605865e-01 -5.06114662e-01 -7.82203913e-01 4.04192150e-01
-8.59274387e-01 -9.35616088e-04 -1.10114045e-01 -7.24681377e-01
-2.97348708e-01 -7.79710054e-01 6.20832980e-01 3.92938554e-01
-1.51356116e-01 -1.66981089e+00 1.75012395e-01 2.47524500e-01
3.20879549e-01 -2.56881528e-02 7.89094448e-01 -1.26927292e+00
-1.48059698e-02 -1.14228427e-01 4.04591948e-01 6.06951058e-01
6.25736117e-01 -3.32721829e-01 -1.35381341e+00 -2.02533230e-01
2.07426846e-01 -8.24615419e-01 1.15619874e+00 3.27987969e-01
1.04603112e+00 -4.32838470e-01 -4.23994660e-01 5.51947117e-01
1.23925507e+00 2.90748507e-01 1.40807748e-01 5.94573736e-01
6.78339958e-01 9.68884349e-01 6.55058086e-01 1.79632813e-01
2.12022230e-01 1.83342949e-01 1.96036309e-01 -9.65815485e-02
-5.87009825e-02 -3.23746741e-01 1.74275592e-01 8.04523885e-01
9.95341763e-02 1.20490447e-01 -1.00891912e+00 5.49541533e-01
-1.81721139e+00 -1.05306602e+00 9.83741507e-02 2.10298228e+00
9.29300189e-01 2.01262400e-01 3.17669213e-01 2.02397302e-01
8.25787544e-01 2.74705291e-02 -8.85105550e-01 -3.60394686e-01
-3.84983033e-01 9.82366651e-02 3.36128414e-01 2.32903287e-01
-1.28176308e+00 1.07702303e+00 7.35258341e+00 7.82542408e-01
-1.12343562e+00 3.58038723e-01 7.06906557e-01 -4.30056192e-02
-1.87939584e-01 -3.83619517e-02 -8.49517167e-01 2.92014688e-01
1.13034213e+00 -1.34495616e-01 1.11652412e-01 1.05729878e+00
-3.43783230e-01 5.76821640e-02 -1.45495427e+00 1.00668681e+00
4.05724078e-01 -1.57343113e+00 2.08102912e-01 -1.56575609e-02
1.33465779e+00 1.10825017e-01 2.40100026e-01 4.71620411e-01
3.94628376e-01 -7.98109293e-01 3.51671129e-01 4.14850324e-01
4.17764097e-01 -7.61917353e-01 4.70991939e-01 7.32941270e-01
-6.64264739e-01 -4.05680597e-01 -4.37085956e-01 1.85016319e-01
-2.93576539e-01 5.54106712e-01 -8.44028294e-01 5.43970346e-01
3.85632396e-01 1.12186444e+00 -7.86288083e-01 8.22936952e-01
-2.21413150e-01 7.08949029e-01 4.64993902e-02 8.70781764e-02
2.43247524e-01 -2.52548382e-02 5.09830415e-01 1.23982084e+00
-7.64832720e-02 -2.62665957e-01 3.39447737e-01 1.01225924e+00
-1.98764373e-02 5.91075569e-02 -6.33972228e-01 -1.73895150e-01
3.30935687e-01 1.15349472e+00 -4.57183212e-01 -7.07681775e-01
-3.46894503e-01 9.23507392e-01 4.71243471e-01 3.13026696e-01
-1.55729309e-01 -1.44254282e-01 3.26447375e-02 6.33158609e-02
-2.92830672e-02 7.13244379e-02 -4.04920965e-01 -1.33146763e+00
-5.32015145e-01 -7.74762750e-01 6.07921124e-01 -6.57270372e-01
-1.54447687e+00 7.71720767e-01 1.86409876e-01 -1.35046327e+00
-1.02798581e+00 -4.41565663e-01 -5.61653852e-01 4.73972082e-01
-1.69187760e+00 -1.25512874e+00 -1.75019965e-01 5.93693614e-01
8.37518454e-01 -1.08038676e+00 8.89405370e-01 -3.22963148e-01
-2.92846024e-01 6.43552125e-01 4.10874903e-01 -5.97700812e-02
6.67603731e-01 -1.42697155e+00 1.41314507e-01 3.57923359e-01
6.61358237e-01 4.96906847e-01 5.28795421e-01 -6.98465168e-01
-1.04028916e+00 -1.32280076e+00 6.15068793e-01 -1.98533431e-01
3.28939259e-01 -1.26379773e-01 -1.11127388e+00 6.02776408e-01
2.03937471e-01 -2.47078994e-03 1.00114489e+00 2.27461085e-01
-3.61794055e-01 -2.67328918e-01 -1.16059458e+00 3.63292187e-01
9.28711593e-01 -4.30367708e-01 -1.06881046e+00 5.51097095e-01
3.20219189e-01 1.53104840e-02 -7.69606173e-01 4.63039905e-01
8.51577297e-02 -8.70983422e-01 8.11280012e-01 -1.15297532e+00
3.64192992e-01 2.30811596e-01 1.49340153e-01 -1.77167964e+00
-4.67782706e-01 -5.02750576e-01 -4.85279620e-01 1.31368113e+00
5.58960915e-01 -6.30497992e-01 9.22535717e-01 4.13323790e-01
-2.33544096e-01 -5.70856094e-01 -3.97875041e-01 -1.25961781e+00
2.78937191e-01 -2.41317078e-01 4.15907025e-01 1.33661604e+00
2.47959927e-01 5.04988968e-01 -2.72121221e-01 -2.28597254e-01
9.50796902e-01 2.84354210e-01 6.89698637e-01 -1.67915297e+00
-1.95417285e-01 -3.04690242e-01 -3.71640772e-01 -6.72909081e-01
8.58519256e-01 -1.40033686e+00 -1.39821619e-01 -1.38425374e+00
1.90781713e-01 -5.70657670e-01 -2.78552175e-01 8.19834173e-01
-1.32611692e-01 1.27913386e-01 -2.93005705e-02 5.35930336e-01
-5.70439696e-01 6.83852732e-01 1.11497188e+00 -3.96223545e-01
-6.46915734e-01 3.17712843e-01 -6.64265215e-01 8.17250431e-01
1.09803486e+00 -8.45663488e-01 -6.23513281e-01 -3.46663713e-01
1.31072268e-01 -1.59998044e-01 1.31162286e-01 -1.00777352e+00
2.94395894e-01 -3.69537055e-01 6.38908327e-01 -1.79752618e-01
2.21418187e-01 -7.74786353e-01 -3.28781277e-01 4.59061921e-01
-8.69573772e-01 -1.42246440e-01 8.69350359e-02 5.99198997e-01
-3.10856432e-01 -8.27246726e-01 9.92974818e-01 -5.44480979e-01
-8.28103542e-01 1.06522739e-01 -5.00688970e-01 2.25816980e-01
9.33724999e-01 -3.59167278e-01 -1.05957389e-01 -2.13426322e-01
-1.09682524e+00 1.14292130e-01 2.17487529e-01 4.91040796e-01
6.12786889e-01 -1.40396810e+00 -6.50262117e-01 4.26904589e-01
4.71278250e-01 -1.66651413e-01 -4.23960164e-02 2.77243525e-01
5.99166118e-02 4.37458664e-01 -4.84057158e-01 -7.03629553e-01
-8.52706790e-01 8.05407584e-01 3.93955916e-01 -1.40901595e-01
-2.69964784e-01 5.97965121e-01 1.25915080e-01 -6.77402020e-01
2.73378313e-01 3.66084352e-02 -2.76088923e-01 4.75346968e-02
5.16333461e-01 4.21695203e-01 -4.98716626e-03 -5.39223194e-01
1.36950221e-02 1.91759750e-01 -2.89707661e-01 -1.31176010e-01
1.49775791e+00 -1.20762236e-01 1.14636533e-01 9.33787644e-01
1.15503001e+00 -4.46175903e-01 -1.10485756e+00 -6.79128468e-01
3.55469525e-01 1.09367166e-02 2.01992467e-01 -9.25304234e-01
-1.12477732e+00 9.84857976e-01 5.83247721e-01 1.51167795e-01
1.14712667e+00 2.09827155e-01 3.58038515e-01 9.08519626e-01
2.38636926e-01 -1.44852602e+00 5.66041410e-01 4.27438617e-01
6.07334852e-01 -1.72858012e+00 5.27169276e-03 -7.17427656e-02
-7.66394496e-01 1.24059796e+00 9.29251492e-01 -1.61575869e-01
6.26147449e-01 4.69475351e-02 6.33681566e-02 -1.19900465e-01
-8.81109655e-01 -4.01473045e-01 6.45363033e-01 8.84423375e-01
2.97621518e-01 -2.69590169e-01 1.61705121e-01 3.64531696e-01
-7.10256100e-02 -1.89958349e-01 3.41042489e-01 1.11812532e+00
-6.73310339e-01 -1.39320457e+00 -3.19695055e-01 7.58176863e-01
-1.27208799e-01 -9.33983698e-02 -4.67939317e-01 4.32165593e-01
2.15210661e-01 9.93789673e-01 1.99081346e-01 -3.12629819e-01
-1.16298646e-01 4.23414379e-01 7.01421082e-01 -1.15196061e+00
-8.34318459e-01 -1.34613872e-01 -2.32477829e-01 -7.47871399e-03
-1.03278220e+00 -4.86594886e-01 -1.17133749e+00 2.48659074e-01
-5.36002576e-01 4.09199208e-01 8.06155801e-01 1.26794755e+00
2.74773508e-01 4.11704540e-01 9.44499612e-01 -6.80578351e-01
-5.25577009e-01 -9.74476874e-01 -5.97855508e-01 3.14732820e-01
3.58472407e-01 -6.28853202e-01 -3.00839782e-01 4.79242563e-01] | [9.852741241455078, 3.059251070022583] |
3836b908-d652-40e1-b36a-9fe1ed6df6aa | learning-generative-vision-transformer-with-1 | 2112.13528 | null | https://arxiv.org/abs/2112.13528v1 | https://arxiv.org/pdf/2112.13528v1.pdf | Learning Generative Vision Transformer with Energy-Based Latent Space for Saliency Prediction | Vision transformer networks have shown superiority in many computer vision tasks. In this paper, we take a step further by proposing a novel generative vision transformer with latent variables following an informative energy-based prior for salient object detection. Both the vision transformer network and the energy-based prior model are jointly trained via Markov chain Monte Carlo-based maximum likelihood estimation, in which the sampling from the intractable posterior and prior distributions of the latent variables are performed by Langevin dynamics. Further, with the generative vision transformer, we can easily obtain a pixel-wise uncertainty map from an image, which indicates the model confidence in predicting saliency from the image. Different from the existing generative models which define the prior distribution of the latent variables as a simple isotropic Gaussian distribution, our model uses an energy-based informative prior which can be more expressive to capture the latent space of the data. We apply the proposed framework to both RGB and RGB-D salient object detection tasks. Extensive experimental results show that our framework can achieve not only accurate saliency predictions but also meaningful uncertainty maps that are consistent with the human perception. | ['Ping Li', 'Nick Barnes', 'Jianwen Xie', 'Jing Zhang'] | 2021-12-27 | learning-generative-vision-transformer-with | http://proceedings.neurips.cc/paper/2021/hash/8289889263db4a40463e3f358bb7c7a1-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/8289889263db4a40463e3f358bb7c7a1-Paper.pdf | neurips-2021-12 | ['rgb-d-salient-object-detection', 'thermal-image-segmentation'] | ['computer-vision', 'computer-vision'] | [ 1.87002212e-01 1.26901120e-01 1.17828190e-01 -3.66568953e-01
-3.83016944e-01 1.51563548e-02 7.72685587e-01 -3.67783666e-01
-3.12298745e-01 6.78111911e-01 2.02201843e-01 1.34031564e-01
-1.80612579e-02 -8.25449765e-01 -8.59107912e-01 -1.00112557e+00
5.67403674e-01 4.49948311e-01 5.76591194e-01 1.93996638e-01
2.12924555e-01 -4.18722294e-02 -1.65299928e+00 -3.19375783e-01
1.23135066e+00 1.20489740e+00 8.26173842e-01 2.28333116e-01
6.57694489e-02 8.86535347e-01 -1.44526824e-01 -5.50515018e-02
-5.00342585e-02 -3.53643358e-01 -3.25299859e-01 1.49613306e-01
-2.18347996e-01 -5.26133120e-01 -1.61893070e-01 1.52507675e+00
1.45761803e-01 2.41876245e-01 9.00730550e-01 -1.32741547e+00
-9.52211022e-01 3.23362559e-01 -7.22079396e-01 -7.36134648e-02
4.64367010e-02 2.26449192e-01 7.80606866e-01 -9.35570478e-01
3.72215837e-01 1.47415078e+00 3.15429419e-01 3.30124885e-01
-9.87991273e-01 -3.99495155e-01 3.16212654e-01 4.70495552e-01
-1.34222472e+00 -2.41684079e-01 1.18182313e+00 -4.38955277e-01
4.32378411e-01 -5.39903855e-03 9.76855099e-01 1.04308558e+00
3.95091265e-01 1.05282867e+00 1.22342789e+00 -3.16774487e-01
5.67655265e-01 7.21156150e-02 -6.50428832e-02 7.55402446e-01
3.19119066e-01 1.64202482e-01 -5.24308741e-01 -7.93950111e-02
1.03280830e+00 2.76439160e-01 -2.45683581e-01 -8.24134290e-01
-1.33544242e+00 9.34136927e-01 5.87396979e-01 -1.65800393e-01
-6.58817649e-01 4.44483727e-01 -2.08485961e-01 -7.36299455e-01
4.46305990e-01 -1.78873241e-01 1.79420933e-01 1.56956762e-01
-9.97026801e-01 -3.02196667e-02 4.10705745e-01 1.01716375e+00
7.99673319e-01 2.99463421e-01 -4.61988419e-01 3.37662667e-01
1.22675157e+00 9.60276127e-01 3.83761734e-01 -1.11377656e+00
-2.12794915e-01 3.21643442e-01 6.09954119e-01 -9.55007493e-01
1.35115847e-01 -5.12549698e-01 -9.37393904e-01 4.24457192e-01
5.44797443e-02 7.32069388e-02 -1.19515765e+00 1.88874924e+00
3.16936821e-01 3.22494864e-01 -4.49476093e-02 1.20314860e+00
5.70456326e-01 8.58055055e-01 2.50155758e-02 -4.77596909e-01
1.28821194e+00 -9.21778381e-01 -9.24691379e-01 -2.62052804e-01
-5.52870214e-01 -5.86540997e-01 9.18638766e-01 1.65720597e-01
-1.24598074e+00 -5.77927709e-01 -1.06544518e+00 -2.40981206e-01
3.80140692e-02 2.49479383e-01 5.88709116e-01 2.20250443e-01
-1.11025465e+00 1.07180402e-01 -1.40130091e+00 -8.82156417e-02
3.87471795e-01 -1.40698165e-01 2.72920102e-01 2.18963623e-01
-1.00985122e+00 1.01052487e+00 3.49704862e-01 3.51560265e-01
-1.43131006e+00 -1.96187839e-01 -8.10886860e-01 5.78232333e-02
2.78523237e-01 -1.05457580e+00 1.28600371e+00 -6.47811592e-01
-1.70334315e+00 4.88410503e-01 -5.82162261e-01 -4.43376124e-01
5.44380903e-01 -2.49775216e-01 2.16871232e-01 5.77430464e-02
1.83869153e-01 8.20973516e-01 1.35294294e+00 -1.39145720e+00
-4.94592607e-01 -7.80674294e-02 -1.26182139e-01 3.44816595e-01
-3.74924578e-02 -1.63845748e-01 -4.87837970e-01 -3.74962986e-01
4.81438458e-01 -7.51048923e-01 -2.98566252e-01 2.88396120e-01
-4.63422626e-01 -2.38943368e-01 5.53847849e-01 -4.97024685e-01
7.43889868e-01 -1.95850933e+00 3.47650528e-01 -7.47934952e-02
4.00561154e-01 -1.97078034e-01 4.77989793e-01 -1.83227882e-01
5.00856817e-01 -2.59709805e-01 -3.26819658e-01 -5.13292253e-01
3.63481283e-01 2.81102896e-01 -6.38813794e-01 4.28527623e-01
1.59128681e-01 1.06165528e+00 -1.01788747e+00 -7.47746348e-01
4.96497601e-01 7.17454791e-01 -3.56419444e-01 1.59541488e-01
-4.48668242e-01 2.98333496e-01 -8.01374137e-01 4.83099937e-01
8.43166471e-01 -5.44652522e-01 -3.34652752e-01 -3.67441446e-01
-3.10730219e-01 2.13253163e-02 -1.06694961e+00 1.66549146e+00
-1.05462015e-01 4.62342292e-01 -2.24811770e-02 -6.37633622e-01
9.51134562e-01 -9.65418890e-02 1.74924627e-01 -3.45101178e-01
1.08240224e-01 -1.78205688e-02 -3.06824237e-01 -2.68115491e-01
6.17965758e-01 -2.28452101e-01 1.52979195e-01 3.91581088e-01
7.65229836e-02 -2.07097262e-01 -2.15921193e-01 1.64750665e-01
3.36319476e-01 5.74904025e-01 9.14245546e-02 -2.60421634e-01
3.14086109e-01 -2.48154074e-01 8.35034907e-01 7.49377310e-01
-2.42808372e-01 7.17889547e-01 1.06863029e-01 3.54549922e-02
-9.40138221e-01 -1.53496802e+00 -9.40851122e-02 4.40790743e-01
6.99399710e-01 2.94486415e-02 -7.48487055e-01 -1.97259724e-01
-3.16809982e-01 9.98703003e-01 -6.27895474e-01 -3.62275004e-01
2.69727246e-03 -8.12420189e-01 -2.27411658e-01 5.46068609e-01
9.53000128e-01 -9.15374398e-01 -1.04571927e+00 2.56201714e-01
-5.02947390e-01 -9.14330721e-01 -3.13728601e-01 -2.67001074e-02
-7.43673682e-01 -5.70091963e-01 -8.16145599e-01 -5.66644132e-01
7.88446367e-01 2.80833215e-01 8.10934305e-01 -4.83781308e-01
-1.75404757e-01 5.56255221e-01 1.49432914e-02 -6.55637860e-01
-7.99555257e-02 -6.02960169e-01 9.66300815e-02 1.64697081e-01
3.15203428e-01 -5.88757575e-01 -8.59116375e-01 2.34966144e-01
-7.24127889e-01 5.81929326e-01 6.34450436e-01 6.96840048e-01
1.10708785e+00 2.75109559e-02 1.29309684e-01 8.93550888e-02
3.73434067e-01 -4.26421851e-01 -8.24544966e-01 4.18637961e-01
-6.33024156e-01 3.37801695e-01 3.17594260e-02 -3.98340344e-01
-1.51529098e+00 1.37372762e-01 2.02348679e-01 -8.69776309e-01
9.72110685e-03 2.65672058e-01 -2.00037882e-01 1.38346210e-01
7.32946768e-02 6.69905484e-01 -1.79380495e-02 -4.06505525e-01
4.27402288e-01 2.32639953e-01 6.62438631e-01 -4.73719925e-01
8.03745568e-01 8.15706253e-01 2.23469868e-01 -4.50511813e-01
-8.81578505e-01 -4.82441336e-02 -4.09730166e-01 -3.44794035e-01
1.37379909e+00 -1.02437901e+00 -7.31637180e-01 7.63378561e-01
-1.34171069e+00 7.91565031e-02 -3.43450010e-01 7.47653186e-01
-8.84942710e-01 4.35618281e-01 -3.49442810e-01 -1.20218945e+00
-3.66932631e-01 -1.33829749e+00 1.14314544e+00 6.26601517e-01
3.14850211e-01 -9.99618828e-01 -9.84870791e-02 -1.01236682e-02
3.70732635e-01 2.51782417e-01 6.11785889e-01 8.15922096e-02
-1.19521987e+00 9.57801268e-02 -3.65943223e-01 2.08901331e-01
3.05128545e-02 1.60100430e-01 -1.03623593e+00 -1.81477377e-03
5.81155062e-01 -1.07716501e-01 1.14167404e+00 9.87505436e-01
8.74022722e-01 -1.53845191e-01 -4.28823113e-01 5.91240942e-01
1.43775165e+00 1.34696752e-01 5.33107638e-01 6.29964024e-02
7.58027673e-01 3.27191293e-01 8.56720030e-01 5.57147861e-01
5.68602264e-01 4.57241684e-01 8.25244784e-01 5.28623275e-02
2.81011388e-02 -7.14301229e-01 3.78779382e-01 5.67856014e-01
-2.34779537e-01 -1.74857110e-01 -6.84963644e-01 5.35490274e-01
-2.13610291e+00 -9.28158522e-01 -5.86232112e-04 2.08138943e+00
7.45975971e-01 2.24155128e-01 -1.70335799e-01 -2.81417996e-01
8.96909833e-01 1.24299064e-01 -9.06332374e-01 2.26537913e-01
-1.12148859e-02 -2.83523500e-01 2.96474308e-01 6.20462477e-01
-8.59504163e-01 8.15602183e-01 6.31908035e+00 8.53069365e-01
-9.51054156e-01 2.81561673e-01 3.90550941e-01 -6.53223917e-02
-7.38191962e-01 3.49439889e-01 -9.04480040e-01 7.80644178e-01
3.52920085e-01 -3.97577107e-01 2.44704664e-01 7.94125617e-01
4.17721540e-01 -4.52306688e-01 -7.77378857e-01 1.13074470e+00
-6.32319192e-04 -1.16078389e+00 1.46855399e-01 3.04103106e-01
6.80218101e-01 8.85784104e-02 4.45895821e-01 -1.30921364e-01
6.83979154e-01 -7.05617189e-01 1.35242105e+00 1.26728976e+00
4.11548823e-01 -5.58352888e-01 4.36179757e-01 6.45410895e-01
-8.78510296e-01 8.40040669e-02 -7.19577074e-01 -4.14778180e-02
4.92590576e-01 9.10803616e-01 -5.68232119e-01 1.85416237e-01
8.18816662e-01 8.10109735e-01 -3.84477049e-01 1.02266848e+00
-6.07091546e-01 4.18471843e-01 -4.62720424e-01 -2.94646710e-01
1.49759889e-01 -5.53051114e-01 8.68596554e-01 5.78523636e-01
5.32697022e-01 -3.02842762e-02 1.00161560e-01 1.86081469e+00
3.70410711e-01 -4.25542325e-01 -3.00679892e-01 2.49376714e-01
4.44553316e-01 1.20924461e+00 -8.12259316e-01 -2.03571245e-01
-1.03651084e-01 1.00278461e+00 9.60329100e-02 6.98884189e-01
-1.12768400e+00 -4.71036695e-02 2.80036181e-01 -3.14361423e-01
5.84433556e-01 -3.14302862e-01 -1.35563314e-01 -1.36029708e+00
9.07295868e-02 4.80222702e-02 -2.91083068e-01 -1.50848448e+00
-1.07105255e+00 3.61603945e-01 3.78557444e-01 -1.29437399e+00
-4.40641195e-01 -5.00527918e-01 -8.75364244e-01 1.19299936e+00
-1.74874091e+00 -1.28328419e+00 -4.31610167e-01 5.62747836e-01
2.50418931e-01 6.16169535e-02 5.43860555e-01 -4.66408223e-01
-3.75076652e-01 -4.73405793e-02 2.63650090e-01 -2.55067259e-01
3.83478910e-01 -1.25783205e+00 9.59319696e-02 1.09500790e+00
-2.43883096e-02 6.85189426e-01 1.17330801e+00 -6.40210092e-01
-1.14132309e+00 -9.04537737e-01 3.97497118e-01 -3.52188706e-01
5.78535080e-01 -2.79994339e-01 -6.54589474e-01 5.64204872e-01
3.61986935e-01 -6.66976860e-03 1.25003129e-01 -3.93735737e-01
1.57138079e-01 -1.40852444e-02 -1.00399852e+00 4.27335054e-01
9.14871693e-01 -3.79179597e-01 -7.87580550e-01 5.14093488e-02
8.63478959e-01 -2.99705446e-01 -4.99151409e-01 3.78315896e-01
4.18321818e-01 -9.87971842e-01 9.61487651e-01 6.28036633e-02
3.13765854e-01 -6.48514926e-01 -1.72837421e-01 -1.17919183e+00
-5.51724374e-01 -3.79413456e-01 -3.73286575e-01 1.25219774e+00
-1.14369854e-01 -5.81520617e-01 4.23441261e-01 6.21102095e-01
-1.33496344e-01 -5.56265175e-01 -9.69336689e-01 -4.96698886e-01
-2.63947606e-01 -4.34001684e-01 5.49139202e-01 2.46431813e-01
-3.64733815e-01 1.59201592e-01 -3.85610968e-01 2.89299726e-01
1.37212002e+00 4.18828189e-01 2.20464572e-01 -1.26369500e+00
-4.01940525e-01 -3.28514487e-01 -3.33055258e-01 -1.44641221e+00
1.65866211e-01 -4.56881702e-01 4.58164424e-01 -1.86731517e+00
6.79476500e-01 -1.06825933e-01 -4.37558532e-01 2.07106203e-01
-2.20213667e-01 -1.46416098e-01 -2.94403243e-03 4.72610801e-01
-7.96654880e-01 1.20751500e+00 1.45052254e+00 -1.58614442e-01
1.16047278e-01 5.56518175e-02 -5.76888740e-01 1.10214555e+00
4.38646466e-01 -2.24063888e-01 -7.37446547e-01 -4.62049425e-01
2.99420118e-01 -4.67973314e-02 9.60174561e-01 -8.76487255e-01
4.26965088e-01 -5.08024633e-01 4.99278814e-01 -9.55109358e-01
5.79364479e-01 -7.51542330e-01 4.89978008e-02 3.70106280e-01
-1.32153288e-01 -5.61702192e-01 -1.05875351e-01 9.63244915e-01
-9.81133580e-02 -1.90383732e-01 7.94543505e-01 -1.15433201e-01
-7.83359230e-01 3.51623178e-01 -2.60107845e-01 -1.99118122e-01
9.00994480e-01 -2.57677257e-01 -2.84325421e-01 -5.78604400e-01
-7.83141375e-01 1.43464163e-01 6.05903566e-01 2.57976562e-01
1.02822471e+00 -1.56994867e+00 -4.32389975e-01 2.78051406e-01
-6.22667857e-02 2.92750388e-01 1.10313989e-01 9.09405768e-01
-5.53502962e-02 2.54476875e-01 -3.23425472e-01 -1.09181380e+00
-6.39996409e-01 6.57432616e-01 4.65552360e-01 1.93096474e-01
-3.95033836e-01 8.10948670e-01 5.69231868e-01 8.87451023e-02
8.30234438e-02 -6.63608313e-01 -1.61530688e-01 -3.01642984e-01
2.80386955e-01 3.20470542e-01 -7.25259840e-01 -7.11010098e-01
-3.24414849e-01 6.11377656e-01 1.68213263e-01 -5.06380141e-01
1.18661892e+00 -5.89096785e-01 -1.19151130e-01 7.37033606e-01
7.25461423e-01 -2.96947896e-01 -1.99935889e+00 -3.05332035e-01
-4.94053960e-01 -3.61472547e-01 5.51373243e-01 -5.14296412e-01
-8.18973243e-01 1.13615108e+00 7.61750519e-01 9.62566435e-02
1.01288593e+00 1.64669320e-01 5.02767861e-01 -4.21358384e-02
6.13209307e-01 -1.03230727e+00 1.69792250e-01 2.73421824e-01
8.38277757e-01 -1.31978214e+00 -2.47341916e-02 -2.78144807e-01
-5.50223231e-01 6.47285938e-01 6.11840129e-01 -2.97019958e-01
7.26719975e-01 -7.21621737e-02 -3.84065866e-01 -5.12440726e-02
-7.25287497e-01 -2.75676548e-01 4.31296527e-01 6.58633769e-01
-1.07339248e-01 7.82458857e-02 -9.03385662e-05 8.03838968e-01
-2.07370259e-02 1.40134677e-01 4.30663228e-01 6.59439623e-01
-8.70660722e-01 -4.77585763e-01 -4.29421008e-01 7.32821897e-02
-3.51191685e-02 -1.75827280e-01 7.33478367e-02 2.69228667e-01
1.40160859e-01 8.44950974e-01 2.58211404e-01 2.03742483e-03
-1.51874661e-01 -7.42158741e-02 5.83347857e-01 -4.13191050e-01
6.68250740e-01 5.16166329e-01 -6.49864733e-01 -4.73642468e-01
-5.50125957e-01 -9.13436770e-01 -1.47830987e+00 1.66547194e-01
-4.90444511e-01 3.87080051e-02 7.44848609e-01 1.05348969e+00
2.05590889e-01 5.74049890e-01 4.42497253e-01 -1.00514662e+00
-6.99830949e-01 -9.95920837e-01 -6.66822493e-01 2.42924294e-03
3.64864230e-01 -1.14618504e+00 -6.38745904e-01 1.76376596e-01] | [10.034191131591797, -0.3265591263771057] |
6318534f-a1a6-4c70-a79b-9358e7683145 | learning-to-repeat-fine-grained-action | 1702.06054 | null | https://arxiv.org/abs/1702.06054v2 | https://arxiv.org/pdf/1702.06054v2.pdf | Learning to Repeat: Fine Grained Action Repetition for Deep Reinforcement Learning | Reinforcement Learning algorithms can learn complex behavioral patterns for sequential decision making tasks wherein an agent interacts with an environment and acquires feedback in the form of rewards sampled from it. Traditionally, such algorithms make decisions, i.e., select actions to execute, at every single time step of the agent-environment interactions. In this paper, we propose a novel framework, Fine Grained Action Repetition (FiGAR), which enables the agent to decide the action as well as the time scale of repeating it. FiGAR can be used for improving any Deep Reinforcement Learning algorithm which maintains an explicit policy estimate by enabling temporal abstractions in the action space. We empirically demonstrate the efficacy of our framework by showing performance improvements on top of three policy search algorithms in different domains: Asynchronous Advantage Actor Critic in the Atari 2600 domain, Trust Region Policy Optimization in Mujoco domain and Deep Deterministic Policy Gradients in the TORCS car racing domain. | ['Aravind Srinivas', 'Balaraman Ravindran', 'Sahil Sharma'] | 2017-02-20 | null | null | null | null | ['carracing-v0'] | ['playing-games'] | [-2.44930804e-01 5.51666282e-02 -3.99898589e-01 -1.17003970e-01
-3.75251412e-01 -6.27050936e-01 8.38198304e-01 2.04240650e-01
-9.83482718e-01 1.12615097e+00 9.77995321e-02 -4.90968555e-01
-1.11031272e-01 -7.15615690e-01 -6.92186177e-01 -7.94326127e-01
-4.23812211e-01 7.99679518e-01 2.01933146e-01 -3.30891073e-01
2.85265446e-01 4.61298972e-01 -1.30925453e+00 1.03895262e-01
4.80392218e-01 9.03572142e-01 1.29648864e-01 1.01686168e+00
1.90279260e-01 1.34549725e+00 -5.01708329e-01 3.38123858e-01
5.36919653e-01 -3.22225779e-01 -7.39653707e-01 1.60274535e-01
-3.99198234e-01 -7.62601137e-01 -2.10785404e-01 7.51578808e-01
3.28239441e-01 5.07550955e-01 4.22019720e-01 -1.35168087e+00
9.16250199e-02 9.63115513e-01 -4.32162374e-01 1.87516645e-01
1.02370322e-01 8.58115315e-01 7.45683014e-01 -1.62928388e-01
6.27009332e-01 1.47816014e+00 -4.72365245e-02 6.88082695e-01
-1.18311071e+00 -3.92225951e-01 6.96680725e-01 3.29910398e-01
-6.75656319e-01 -2.45377034e-01 3.56104225e-01 -1.23879150e-01
1.28294361e+00 -2.19546124e-01 8.41332257e-01 1.14568532e+00
4.19357657e-01 9.60341036e-01 1.32025290e+00 -2.50679553e-01
1.03027487e+00 -3.24333012e-01 -5.33304513e-01 7.14073122e-01
-2.52329648e-01 8.69441748e-01 -3.65507275e-01 -3.13105762e-01
1.04658175e+00 -3.30829769e-02 3.18301618e-01 -4.99447495e-01
-1.24300134e+00 7.15498149e-01 1.91355094e-01 -2.04619076e-02
-7.96295643e-01 7.83973634e-01 5.95596135e-01 8.55618954e-01
2.78419489e-03 6.09721899e-01 -6.21753573e-01 -6.27901614e-01
-3.09067130e-01 8.12233508e-01 8.30031455e-01 5.60750246e-01
7.82784104e-01 5.04047990e-01 -5.11822224e-01 3.67238343e-01
1.44244701e-01 5.13664603e-01 4.75267857e-01 -1.65397656e+00
2.33652711e-01 1.46534741e-01 8.91313493e-01 -2.09291548e-01
-3.91016275e-01 -5.02979979e-02 -2.28751570e-01 1.13814080e+00
3.93496335e-01 -7.82998800e-01 -9.29061294e-01 1.62585819e+00
6.60474718e-01 3.37866306e-01 2.22043946e-01 9.29463863e-01
-4.23360765e-01 5.35410583e-01 2.63779879e-01 -1.82039604e-01
8.09718192e-01 -9.88149822e-01 -2.95057893e-01 -1.69760510e-01
6.49532735e-01 -1.15052268e-01 8.54991734e-01 5.73799968e-01
-1.11402309e+00 -3.69328618e-01 -8.02690208e-01 5.68088353e-01
3.07879993e-03 -1.70809463e-01 4.96045619e-01 -3.80123891e-02
-1.15054488e+00 9.73719239e-01 -1.29603672e+00 9.01656523e-02
1.76927388e-01 4.80072170e-01 1.10806309e-01 3.68168890e-01
-9.73607123e-01 8.50249767e-01 6.12752080e-01 -1.68405682e-01
-1.98019719e+00 -4.59689707e-01 -4.44843918e-01 1.57916144e-01
9.83373821e-01 -4.80168104e-01 2.05388045e+00 -1.26636636e+00
-2.38302827e+00 2.88425207e-01 2.55909204e-01 -1.08629215e+00
6.94802284e-01 -2.89193630e-01 -6.08933121e-02 -8.19894522e-02
-4.48090881e-02 5.97430289e-01 1.22213137e+00 -9.26484108e-01
-1.24998760e+00 -2.57173330e-01 6.01031780e-01 5.55130184e-01
2.73730040e-01 -2.14239046e-01 2.65629649e-01 -3.03994089e-01
-8.46799135e-01 -1.21830964e+00 -9.21354234e-01 -4.05172795e-01
4.03663293e-02 -4.62588310e-01 6.67958736e-01 -1.40810728e-01
7.85662830e-01 -2.09217000e+00 4.62033182e-01 2.22937435e-01
1.28716394e-01 1.20987028e-01 -3.33890259e-01 4.87895936e-01
2.05829233e-01 -1.85100958e-01 4.21954133e-02 -2.09934041e-02
3.80242497e-01 4.67309147e-01 -4.27470297e-01 5.00357866e-01
4.17971537e-02 8.25527430e-01 -1.22010350e+00 -2.00391635e-02
2.67285794e-01 -1.41448766e-01 -6.42455697e-01 4.93277043e-01
-1.05237663e+00 6.00677788e-01 -9.71915722e-01 2.73927689e-01
6.49625994e-03 -6.19503716e-03 4.81624246e-01 8.10275078e-01
-6.02248549e-01 3.46432000e-01 -1.23284054e+00 1.54107213e+00
-5.98650575e-01 2.56501645e-01 3.80818367e-01 -9.77784336e-01
6.71162665e-01 2.97405779e-01 6.30091071e-01 -1.06900823e+00
2.96070427e-02 -1.81927364e-02 2.18824372e-01 -1.93241522e-01
4.06153560e-01 1.11237973e-01 -1.02770396e-01 8.83639395e-01
-3.13425124e-01 -1.03748672e-01 3.50271702e-01 7.73814991e-02
1.51865637e+00 4.78663653e-01 3.78706872e-01 -2.23928526e-01
3.09410065e-01 3.49037915e-01 5.70805788e-01 1.22544968e+00
-3.65477741e-01 -4.95743096e-01 7.62548923e-01 -8.08097064e-01
-1.16337645e+00 -9.41278636e-01 5.02814531e-01 1.53591120e+00
5.10408431e-02 -2.35083386e-01 -5.96445680e-01 -9.37944293e-01
3.39798272e-01 9.39510226e-01 -7.95265079e-01 -8.58582109e-02
-1.01639116e+00 -1.17980294e-01 7.05715343e-02 4.39348310e-01
6.24591172e-01 -1.60858428e+00 -1.36085403e+00 6.03235066e-01
5.84546626e-01 -7.72671700e-01 -5.24169564e-01 3.43008876e-01
-8.24784696e-01 -9.42766964e-01 -2.09951222e-01 -1.42069519e-01
3.15284580e-01 -1.03234440e-01 9.92129743e-01 -1.42140880e-01
1.57343559e-02 6.73842072e-01 -2.43640393e-01 -3.28787625e-01
-6.26820922e-01 1.36595815e-02 1.94272250e-01 -1.07242279e-01
9.77899414e-03 -4.78740990e-01 -7.61216104e-01 2.20079437e-01
-7.31336296e-01 -3.47788483e-02 5.15864432e-01 9.03792918e-01
5.59938014e-01 -6.43900633e-02 4.92653131e-01 -8.02245021e-01
1.01384628e+00 -4.79354858e-01 -1.20967710e+00 1.27539471e-01
-7.58853376e-01 6.38027787e-01 1.05230212e+00 -5.28914452e-01
-1.05342889e+00 2.95186453e-02 3.22874248e-01 -3.00569862e-01
-3.35080445e-01 1.14109315e-01 3.43612254e-01 3.13528657e-01
6.56857312e-01 2.95689642e-01 1.80689484e-01 -1.78015932e-01
4.25820947e-01 2.66761005e-01 1.63926631e-01 -1.11420465e+00
4.53370541e-01 3.38449717e-01 -1.19983017e-01 -2.43543208e-01
-2.27605745e-01 6.79209246e-04 -1.59680337e-01 -3.49057853e-01
6.09616041e-01 -7.65024900e-01 -1.30533147e+00 4.09567535e-01
-7.16572821e-01 -1.24219191e+00 -6.95228517e-01 3.27478409e-01
-1.00650048e+00 -2.68044382e-01 -5.63387871e-01 -8.04474950e-01
-2.04068217e-02 -1.34811223e+00 8.63938034e-01 4.58584130e-01
-4.54164334e-02 -8.74934077e-01 4.85262036e-01 -2.69128710e-01
5.62310934e-01 2.92707890e-01 7.84204662e-01 -4.96830940e-01
-6.59794509e-01 4.18366462e-01 4.12032455e-01 6.61407337e-02
7.46428370e-02 -1.04320548e-01 -5.43253362e-01 -6.96923554e-01
-1.75725356e-01 -6.44921958e-01 4.22842026e-01 3.87503475e-01
1.17394257e+00 -5.67897022e-01 -2.23638624e-01 2.12415352e-01
1.13910961e+00 7.65418887e-01 3.88917923e-01 8.09455931e-01
1.85521513e-01 1.02807358e-01 9.11815286e-01 1.00717723e+00
2.20169693e-01 7.26536572e-01 7.13327825e-01 1.93429753e-01
3.39103401e-01 -1.40038878e-01 8.52316141e-01 -1.30260676e-01
-1.84741214e-01 9.00608748e-02 -7.21237183e-01 4.35363740e-01
-2.27280402e+00 -1.04052293e+00 8.44388068e-01 2.19321513e+00
9.28388059e-01 2.82132119e-01 5.23461401e-01 -4.62670147e-01
2.85845816e-01 1.51766837e-01 -1.42135370e+00 -7.37835884e-01
4.74606723e-01 2.74575114e-01 7.19028413e-01 7.34348893e-01
-7.74177253e-01 1.24181163e+00 6.08628988e+00 6.16863668e-01
-1.30876815e+00 -9.04865339e-02 7.24327207e-01 -4.25913841e-01
-2.09883060e-02 9.48744714e-02 -6.35995626e-01 3.22895914e-01
1.28071189e+00 -2.96638250e-01 1.24517727e+00 1.04068220e+00
8.00381601e-01 -3.78153890e-01 -1.22202301e+00 4.56424981e-01
-9.14748549e-01 -1.39910698e+00 -4.94212806e-01 1.04725972e-01
7.52159595e-01 3.42264593e-01 6.57777637e-02 7.42036521e-01
1.61020696e+00 -1.01151109e+00 7.00610340e-01 1.85100392e-01
4.28766221e-01 -1.11422241e+00 1.80005595e-01 5.73633313e-01
-7.98793793e-01 -5.38209975e-01 -1.81118529e-02 -4.32220101e-01
-1.31405100e-01 -2.28497341e-01 -1.35883009e+00 2.34926753e-02
5.75082898e-01 5.86017370e-01 -2.60206126e-03 6.17909074e-01
-4.55517620e-01 7.39727020e-01 -3.29113632e-01 -3.69872153e-01
8.69688332e-01 -3.53280455e-01 5.81888139e-01 7.00488389e-01
8.71915277e-03 1.35372534e-01 6.81194901e-01 6.12691760e-01
1.04244068e-01 -3.07411999e-01 -4.78424400e-01 -7.02666342e-02
4.45118755e-01 9.70246434e-01 -5.57344794e-01 -4.18102562e-01
-4.63345386e-02 7.46002674e-01 6.96194172e-01 7.01522887e-01
-8.30147624e-01 1.26042068e-01 1.23452044e+00 -1.13626085e-01
5.54609954e-01 -4.89214718e-01 1.57693982e-01 -9.52630699e-01
-3.14269483e-01 -1.25518095e+00 2.79333323e-01 -2.82124877e-01
-6.59281015e-01 2.80434400e-01 -3.96064296e-02 -9.66766596e-01
-8.82236838e-01 -4.14455920e-01 -5.91150045e-01 4.98016357e-01
-1.44381404e+00 -1.58387363e-01 3.31612498e-01 6.82552338e-01
8.08257580e-01 -4.20859367e-01 6.06567383e-01 -3.65828663e-01
-5.50963342e-01 2.24803612e-02 4.91695881e-01 -1.17164448e-01
3.59575152e-01 -1.48052132e+00 4.08834159e-01 4.64199901e-01
-1.13360025e-01 1.44732505e-01 8.45770359e-01 -4.60059345e-01
-1.62985706e+00 -8.56084108e-01 -2.50081599e-01 1.37586705e-02
9.05425429e-01 -1.28028151e-02 -5.54993808e-01 6.87873363e-01
4.52262253e-01 -9.78912637e-02 -5.65829203e-02 -1.06561415e-01
8.62040371e-02 -2.36410394e-01 -1.08445907e+00 9.42028940e-01
7.27750838e-01 -1.78884983e-01 -2.09150299e-01 1.39942333e-01
7.45271385e-01 -7.39179671e-01 -6.63307130e-01 -6.73687756e-02
3.55856955e-01 -8.70129287e-01 5.41042149e-01 -1.21927643e+00
3.12443256e-01 -3.21736455e-01 8.49688277e-02 -1.91553104e+00
-2.92244464e-01 -1.11986363e+00 -4.13392127e-01 3.11272442e-01
2.23279536e-01 -6.74688995e-01 8.58780801e-01 5.54898977e-01
1.48982897e-01 -8.02643895e-01 -1.05415630e+00 -5.78758597e-01
1.40126541e-01 -1.60477445e-01 8.92052531e-01 3.97867292e-01
-5.83283529e-02 1.14249222e-01 -4.54493225e-01 1.35724589e-01
5.81923544e-01 1.42149523e-01 8.41655731e-01 -4.68957663e-01
-8.70101631e-01 -4.76980269e-01 1.69003233e-01 -1.13183892e+00
2.98654646e-01 -3.85791063e-01 2.79270440e-01 -1.07676029e+00
-2.67419636e-01 -5.31835973e-01 -5.55298567e-01 6.20455265e-01
6.32876828e-02 -9.27864432e-01 3.64139080e-01 9.01504382e-02
-1.00672483e+00 7.64370620e-01 1.42898810e+00 -1.73747256e-01
-4.60799932e-01 1.67733744e-01 -2.52027512e-01 5.90619743e-01
1.05680037e+00 -4.64058220e-01 -5.46624303e-01 -4.75595146e-01
6.62842467e-02 6.43871605e-01 1.90285981e-01 -8.79823804e-01
8.29607695e-02 -1.04274189e+00 1.44158572e-01 3.98139656e-02
2.00388599e-02 -7.40232110e-01 -1.02765754e-01 9.48401630e-01
-9.53326225e-01 6.18734598e-01 1.80109069e-01 8.18235278e-01
6.73661232e-02 1.51558578e-01 7.24846423e-01 -3.37978512e-01
-1.06635094e+00 3.95974219e-01 -8.99220407e-01 1.89399645e-01
1.32461834e+00 3.84451717e-01 -2.52534181e-01 -4.34314251e-01
-7.71069229e-01 8.36276948e-01 3.74154896e-01 3.91913444e-01
5.80353558e-01 -1.05126429e+00 -7.49807000e-01 -6.54001236e-02
-3.03356856e-01 -1.33792490e-01 -2.41516545e-01 4.00740057e-01
-3.17149818e-01 1.91921666e-01 -5.35849810e-01 -2.74775535e-01
-9.28894222e-01 5.00087976e-01 7.47236431e-01 -7.40142405e-01
-6.29588425e-01 4.00668561e-01 -3.89521755e-02 -3.02074850e-01
3.03676069e-01 -5.26787579e-01 -1.12710953e-01 -4.35451329e-01
7.15855539e-01 4.27719265e-01 -2.83271700e-01 2.26333052e-01
-7.32950345e-02 -2.40195915e-01 -3.29639673e-01 -7.02849925e-01
1.32932699e+00 4.93424274e-02 6.44950345e-02 3.38336140e-01
5.07183790e-01 -4.47323561e-01 -2.14324141e+00 -3.09017152e-01
2.28956729e-01 -2.98503280e-01 1.15410149e-01 -1.04220319e+00
-7.99306452e-01 3.23798329e-01 6.42520010e-01 2.39684433e-01
8.28688622e-01 -3.97478491e-01 5.04779458e-01 8.18169415e-01
7.63000011e-01 -1.69799125e+00 2.17731938e-01 7.96488166e-01
7.27587044e-01 -1.15172243e+00 -2.85645276e-01 7.12509096e-01
-1.08425939e+00 1.10514963e+00 7.79182613e-01 -5.52270234e-01
2.80922592e-01 5.34164250e-01 2.73691695e-02 -3.53984386e-02
-1.57403290e+00 -1.91824391e-01 -5.34919560e-01 4.39486384e-01
-1.32009044e-01 4.71826345e-01 3.04120723e-02 -2.25969583e-01
2.01029524e-01 1.48780808e-01 4.38392878e-01 1.15235376e+00
-7.38934875e-01 -1.30584490e+00 -2.88542420e-01 2.67841697e-01
-2.13426858e-01 3.29534888e-01 -1.94832087e-01 7.16636539e-01
-2.37539932e-01 6.46513164e-01 1.11288182e-01 -7.75376558e-02
1.66541785e-01 -1.98117077e-01 4.55885202e-01 -4.73998487e-01
-8.49350452e-01 -2.50628870e-02 1.94912285e-01 -1.14464259e+00
5.46033569e-02 -8.77013683e-01 -1.68303514e+00 -3.14142853e-01
6.06616974e-01 4.18826312e-01 4.90006238e-01 1.17508745e+00
5.07128179e-01 6.96609497e-01 9.22616541e-01 -8.04462194e-01
-1.35376549e+00 -7.04797924e-01 -4.60259259e-01 1.71197712e-01
6.90233707e-01 -6.38203681e-01 -2.00797920e-03 -3.00598234e-01] | [4.0495219230651855, 1.8154367208480835] |
73360ccd-88d3-4dd5-95fc-014b97017e22 | the-effect-of-balancing-methods-on-model | 2307.00157 | null | https://arxiv.org/abs/2307.00157v1 | https://arxiv.org/pdf/2307.00157v1.pdf | The Effect of Balancing Methods on Model Behavior in Imbalanced Classification Problems | Imbalanced data poses a significant challenge in classification as model performance is affected by insufficient learning from minority classes. Balancing methods are often used to address this problem. However, such techniques can lead to problems such as overfitting or loss of information. This study addresses a more challenging aspect of balancing methods - their impact on model behavior. To capture these changes, Explainable Artificial Intelligence tools are used to compare models trained on datasets before and after balancing. In addition to the variable importance method, this study uses the partial dependence profile and accumulated local effects techniques. Real and simulated datasets are tested, and an open-source Python package edgaro is developed to facilitate this analysis. The results obtained show significant changes in model behavior due to balancing methods, which can lead to biased models toward a balanced distribution. These findings confirm that balancing analysis should go beyond model performance comparisons to achieve higher reliability of machine learning models. Therefore, we propose a new method performance gain plot for informed data balancing strategy to make an optimal selection of balancing method by analyzing the measure of change in model behavior versus performance gain. | ['Przemysław Biecek', 'Mustafa Cavus', 'Adrian Stando'] | 2023-06-30 | null | null | null | null | ['explainable-artificial-intelligence', 'imbalanced-classification'] | ['computer-vision', 'miscellaneous'] | [ 1.05620913e-01 -1.69733495e-01 -6.38934016e-01 -6.39883697e-01
-4.23214793e-01 -1.41432822e-01 2.40458816e-01 5.30839562e-01
-2.42939487e-01 1.08333743e+00 -4.09382321e-02 -4.28280294e-01
-2.84055322e-01 -8.12883496e-01 -6.10385180e-01 -7.75936127e-01
2.76612699e-01 3.24774653e-01 9.01062600e-03 -2.52367184e-02
7.41350472e-01 4.46946740e-01 -1.74340510e+00 2.79215038e-01
1.31765544e+00 6.64600849e-01 -1.38048351e-01 1.85476188e-02
-4.32386577e-01 7.26893127e-01 -9.39821422e-01 -2.27050096e-01
1.69203088e-01 -4.97799605e-01 -3.69447768e-01 -2.09896848e-01
3.67931016e-02 -9.89622474e-02 4.51895714e-01 9.75443482e-01
6.39200747e-01 -1.02512643e-01 8.15114677e-01 -1.73159134e+00
-7.88516030e-02 7.43277550e-01 -8.13451827e-01 2.54833668e-01
1.58599764e-01 3.30377668e-01 6.77560329e-01 -3.92361760e-01
3.63733798e-01 1.28346860e+00 7.52402008e-01 2.33064219e-01
-1.49953592e+00 -1.16585064e+00 1.73548266e-01 4.96984661e-01
-1.32024550e+00 -1.23493046e-01 9.00848508e-01 -7.24862933e-01
6.94585145e-01 7.75858462e-01 6.69435322e-01 6.34962499e-01
5.57627559e-01 3.87913942e-01 1.54927754e+00 -5.53906679e-01
3.61844540e-01 7.26120472e-01 6.68329835e-01 5.02375625e-02
8.90477061e-01 7.25641623e-02 -5.96990824e-01 -2.01416329e-01
-3.44734713e-02 -1.71321929e-01 -2.38713473e-01 -3.12384993e-01
-5.61736643e-01 8.72050643e-01 4.68491644e-01 1.89333186e-01
-2.54031092e-01 -3.02132964e-01 3.97853523e-01 2.38736272e-01
7.53786981e-01 6.86034739e-01 -6.25758946e-01 -7.98099339e-02
-7.70707011e-01 4.99737799e-01 5.86882353e-01 3.12192649e-01
6.17486656e-01 -2.33861804e-02 -2.54830450e-01 9.07443285e-01
4.15380239e-01 2.48254657e-01 7.01646864e-01 -3.38151217e-01
5.36265910e-01 1.29909492e+00 -1.42600639e-02 -1.47206640e+00
-7.09464073e-01 -8.25956702e-01 -6.53530657e-01 2.86727071e-01
3.98683190e-01 2.09958464e-01 -7.45007098e-01 1.64803672e+00
5.92531264e-01 -2.70935714e-01 -2.12452903e-01 7.13707924e-01
6.04828298e-01 3.75607640e-01 3.69445384e-01 -4.21726197e-01
1.12339401e+00 -5.39635420e-01 -8.51364374e-01 -1.15431167e-01
9.14441943e-01 -5.56216478e-01 1.13697898e+00 3.27434272e-01
-8.01120877e-01 -4.69456136e-01 -1.13583314e+00 4.00243193e-01
-3.84133309e-01 -1.80492759e-01 4.11780387e-01 9.77992654e-01
-4.49324012e-01 7.21860707e-01 -7.17103183e-01 -2.48062208e-01
4.95210379e-01 4.39945191e-01 -3.81907485e-02 -1.15280552e-03
-1.15064621e+00 1.15641069e+00 4.88385260e-01 7.24989772e-02
-1.98854998e-01 -1.16848493e+00 -5.61481714e-01 5.97683080e-02
-2.08569225e-02 -5.05889356e-01 7.80605197e-01 -1.27874911e+00
-8.63843620e-01 6.86617792e-01 -1.12155713e-01 -5.13897777e-01
8.17375064e-01 8.25343654e-02 -2.77845293e-01 -5.82770348e-01
-4.85211313e-02 2.90135682e-01 3.58961791e-01 -1.44594085e+00
-4.82782543e-01 -7.68247962e-01 -3.70310456e-01 4.69095930e-02
-4.54953521e-01 -6.84866235e-02 1.07671522e-01 -5.33963203e-01
1.74569368e-01 -7.83143997e-01 -1.26160428e-01 -3.62647504e-01
-2.80516624e-01 9.37993079e-02 5.59010625e-01 -7.82747209e-01
1.80129254e+00 -1.83184874e+00 -3.99453998e-01 2.91285723e-01
1.28498241e-01 2.37403065e-01 1.59733251e-01 2.74812311e-01
-2.03251541e-01 4.57436591e-01 -2.24994555e-01 2.40984578e-02
-2.26715580e-01 1.89903099e-02 7.92371184e-02 4.56946492e-01
1.32098928e-01 2.80739665e-01 -3.08061451e-01 -4.43017840e-01
2.08979368e-01 2.80136526e-01 -7.24298179e-01 1.41750559e-01
9.46533456e-02 3.66356283e-01 -2.10083485e-01 7.02914476e-01
1.25173259e+00 -1.02319308e-01 3.77067149e-01 -1.92057788e-01
-2.02165581e-02 4.27101791e-01 -1.21372426e+00 5.00891507e-01
-3.20133835e-01 4.50961262e-01 -2.93475866e-01 -1.19240415e+00
1.21370912e+00 -1.49982393e-01 3.08985621e-01 -1.03661692e+00
2.46900499e-01 3.76191944e-01 4.65172201e-01 -4.49401796e-01
3.94151598e-01 -1.91778645e-01 2.03892007e-01 6.61483034e-02
-6.52582347e-01 -2.47856006e-02 6.00827560e-02 -3.10149610e-01
7.05413938e-01 -7.40142539e-02 4.92038369e-01 -5.47524035e-01
3.79648864e-01 3.33130509e-01 9.95446920e-01 5.66882730e-01
-3.07793528e-01 2.74050623e-01 6.72442436e-01 -4.90552217e-01
-7.41317809e-01 -5.34766972e-01 -6.27233386e-01 6.01333082e-01
2.61679560e-01 -1.81345731e-01 -7.10693061e-01 -4.31840628e-01
3.93836766e-01 1.14375281e+00 -7.08519876e-01 -7.06352830e-01
-2.85129309e-01 -1.49453866e+00 2.11084113e-01 3.87212396e-01
5.63992321e-01 -8.48184168e-01 -7.68841624e-01 1.36644840e-01
-6.93947375e-02 -4.63511735e-01 1.88946903e-01 2.41755083e-01
-1.26939356e+00 -1.29226613e+00 -2.61773288e-01 -1.36922300e-01
8.18287432e-01 2.22320892e-02 1.27853572e+00 5.21912098e-01
-2.26160139e-01 -3.34624082e-01 -2.47766703e-01 -7.86615908e-01
-5.99314332e-01 4.11140025e-01 -1.61400259e-01 -5.80746531e-02
5.37327111e-01 -2.31724679e-01 -5.84059179e-01 7.43666589e-01
-9.05941784e-01 6.76165894e-02 3.94127071e-01 8.69939566e-01
4.01485592e-01 1.04345873e-01 9.13530171e-01 -9.58615661e-01
7.75377095e-01 -7.08914876e-01 -7.61999726e-01 2.95431525e-01
-1.42325604e+00 4.14296947e-02 4.83681172e-01 -3.73536974e-01
-1.01540542e+00 -3.76530945e-01 1.32002831e-01 -3.78297269e-02
4.71199229e-02 6.87633276e-01 -3.66989642e-01 1.68371573e-01
7.89927542e-01 -2.68447012e-01 2.45126575e-01 -5.07096589e-01
-5.59966326e-01 9.87205565e-01 -3.43214631e-01 -2.88487077e-01
3.63578528e-01 1.68309435e-01 5.64608611e-02 -5.48379481e-01
-5.45702934e-01 -2.53255665e-01 -4.02781606e-01 -3.19674641e-01
2.51608461e-01 -5.99362850e-01 -5.61152101e-01 6.24722183e-01
-7.47746050e-01 -2.75788337e-01 9.40197632e-02 4.61288303e-01
8.38905349e-02 -5.46298623e-02 -3.48293153e-03 -1.05275524e+00
-2.49164492e-01 -1.23403680e+00 4.69018936e-01 4.52734530e-01
-5.22981346e-01 -9.63363349e-01 7.28131160e-02 7.85963476e-01
6.79460347e-01 3.82785916e-01 1.18616712e+00 -9.08565521e-01
-2.38150150e-01 -2.43485272e-01 -8.66175070e-02 2.76442081e-01
1.42218009e-01 3.63497674e-01 -1.03673446e+00 -3.68772656e-01
3.02958693e-02 1.54715152e-02 6.58776641e-01 5.79148412e-01
1.19991970e+00 -2.82981157e-01 -5.88192821e-01 3.97054553e-01
1.50439656e+00 7.22362459e-01 6.92407966e-01 7.45253623e-01
4.23000485e-01 8.46003532e-01 9.16496038e-01 5.16158342e-01
2.68403679e-01 8.64130795e-01 5.48931718e-01 -1.29110053e-01
3.55881192e-02 -1.13327205e-01 9.42758471e-02 5.66170692e-01
2.52178401e-01 -1.93259642e-01 -1.26133001e+00 3.61583829e-01
-1.53671408e+00 -6.84353054e-01 -4.94619548e-01 2.58756948e+00
7.63606071e-01 3.80662829e-01 7.93427005e-02 6.32136762e-01
7.66803086e-01 -1.57067925e-01 -6.85131907e-01 -6.88208997e-01
-6.95505133e-03 -3.46243262e-01 4.53743547e-01 3.75706375e-01
-5.51277220e-01 3.97547930e-01 6.38921976e+00 8.29306722e-01
-1.50924921e+00 -1.48798794e-01 1.28585768e+00 -2.00211823e-01
-5.78391671e-01 5.14181741e-02 -8.80461633e-01 8.71621966e-01
8.79060388e-01 -2.59870738e-01 -1.84058100e-02 7.13164210e-01
6.58048809e-01 -5.01722276e-01 -7.11947799e-01 7.93703675e-01
3.40026640e-03 -9.66243505e-01 2.38035340e-03 2.07882673e-01
7.30869412e-01 -3.87363613e-01 -2.13187218e-01 3.89192998e-01
8.24731961e-03 -1.18471563e+00 4.91287589e-01 4.23933029e-01
3.58815551e-01 -8.71873498e-01 1.14813232e+00 3.38165700e-01
-6.73410296e-01 -3.00827891e-01 -2.55578727e-01 -4.10745502e-01
-2.41101682e-01 1.13176286e+00 -1.11514080e+00 5.39928377e-01
7.04910755e-01 2.26344928e-01 -9.30567384e-01 1.30308759e+00
2.08470076e-01 1.04704916e+00 -2.25972295e-01 -2.86144406e-01
-5.83147824e-01 -1.62314206e-01 2.91241020e-01 7.25649774e-01
2.52379864e-01 -3.55847955e-01 -1.96063638e-01 8.96337211e-01
2.52086669e-01 4.79602754e-01 -4.85792339e-01 3.07946682e-01
7.14924753e-01 9.31040168e-01 -8.49240482e-01 -1.81235462e-01
-1.43549711e-01 1.48489147e-01 3.55117247e-02 -2.19289679e-02
-1.10767615e+00 -1.19828388e-01 5.49297869e-01 6.35451794e-01
-3.54641914e-01 3.89726698e-01 -1.06570637e+00 -7.12999761e-01
7.12494180e-02 -1.14311898e+00 4.85514194e-01 -5.30852437e-01
-1.14516163e+00 2.51667529e-01 1.92335248e-01 -1.22903621e+00
1.09186269e-01 -2.50787020e-01 -5.29350579e-01 9.40983415e-01
-1.36686218e+00 -7.11027384e-01 -6.91663027e-01 -2.49905046e-02
3.55079532e-01 2.56370567e-02 4.00772095e-01 4.38811690e-01
-8.69377732e-01 8.22949767e-01 1.29746795e-01 -4.46567237e-01
7.16056645e-01 -8.89611185e-01 -1.23242564e-01 4.70617980e-01
-4.28904444e-01 3.64250362e-01 8.73146594e-01 -8.65324736e-01
-8.23480010e-01 -8.26958776e-01 8.66761684e-01 -2.44207919e-01
1.43172532e-01 -2.30166148e-02 -1.22707891e+00 2.18066081e-01
-1.95101842e-01 -2.85606951e-01 9.19806838e-01 2.26731166e-01
7.68503174e-03 -5.74765384e-01 -1.49281108e+00 5.29228687e-01
6.37539148e-01 2.71678239e-01 -3.18156362e-01 -1.72406882e-01
2.85204947e-01 -2.58439690e-01 -6.82997882e-01 8.08537066e-01
5.82424521e-01 -1.17192066e+00 5.25799990e-01 -5.98840833e-01
4.53221262e-01 -2.01970294e-01 -9.45824981e-02 -1.50675070e+00
6.14593886e-02 7.31152222e-02 3.23169827e-01 1.52692437e+00
7.88014770e-01 -1.03023517e+00 7.54356325e-01 1.04404891e+00
3.15963745e-01 -9.98029113e-01 -8.65409732e-01 -6.65166318e-01
2.63932228e-01 -4.16773647e-01 9.56447065e-01 1.06118822e+00
-2.06132144e-01 4.18705791e-02 -1.64294049e-01 -1.58891737e-01
4.96190429e-01 -3.22535187e-02 1.04998541e+00 -1.36308551e+00
3.03827096e-02 -5.89754999e-01 -5.55840850e-01 -1.79571182e-01
1.44622261e-02 -8.45266223e-01 -2.89922118e-01 -1.29543865e+00
3.57320696e-01 -8.67787302e-01 -3.64403307e-01 3.93616855e-01
-5.29418826e-01 1.48739651e-01 2.57416517e-01 2.62703180e-01
6.56398311e-02 5.51259816e-01 1.02808034e+00 -1.10154659e-01
-4.13091451e-01 2.51807600e-01 -8.14810693e-01 6.32659554e-01
1.15190816e+00 -8.31597686e-01 -5.28791249e-01 -9.78480130e-02
4.41479355e-01 -4.99034338e-02 2.59446502e-01 -1.01797259e+00
-1.86386004e-01 -4.55941558e-01 4.53592539e-01 -5.44273496e-01
-2.28813097e-01 -1.02666903e+00 5.10619044e-01 8.07699740e-01
-3.44927311e-01 3.38831037e-01 5.60858369e-01 1.02225624e-01
-2.97328204e-01 -2.93274164e-01 9.42420065e-01 3.31434131e-01
-5.85946105e-02 -3.29255879e-01 -1.63765281e-01 -9.41757783e-02
1.32537115e+00 -4.24535036e-01 -4.18940932e-01 -1.80365756e-01
-3.85391027e-01 2.82680064e-01 5.13456166e-01 4.24268991e-01
8.30227882e-02 -1.16973102e+00 -6.24391079e-01 2.44226515e-01
1.52395532e-01 -2.44100884e-01 3.99021626e-01 9.93485510e-01
-6.51718199e-01 2.80914187e-01 -3.96782041e-01 -6.93033218e-01
-1.59701300e+00 3.78302872e-01 5.62343895e-01 -6.02831066e-01
3.03534474e-02 4.13681030e-01 -1.04037762e-01 -7.25431383e-01
1.20616436e-01 -4.55078185e-01 -3.51403117e-01 5.29395521e-01
3.61386210e-01 8.15057158e-01 4.93379802e-01 -3.10830265e-01
-6.54211462e-01 4.59962845e-01 -1.54662756e-02 2.56280869e-01
1.11259174e+00 -5.36938421e-02 -1.89194739e-01 6.10253274e-01
9.14432585e-01 4.74592634e-02 -9.92020130e-01 3.60435247e-01
1.31642967e-01 -8.13435614e-01 -1.10557511e-01 -1.13868654e+00
-1.17424536e+00 6.35634124e-01 9.19497728e-01 3.42469454e-01
1.26136613e+00 -5.71325839e-01 3.04616001e-02 -2.18134835e-01
2.71322817e-01 -1.07452416e+00 -3.38535637e-01 -4.20621075e-02
9.17473674e-01 -1.52829075e+00 3.07681501e-01 -2.64691621e-01
-5.52859843e-01 7.01970160e-01 9.81188715e-01 2.33517841e-01
7.19588578e-01 2.28603274e-01 3.68168682e-01 2.35570427e-02
-8.13613951e-01 5.03450572e-01 1.62940666e-01 3.65107954e-01
5.87928057e-01 1.44817948e-01 -1.07337177e+00 7.71617353e-01
-3.45350832e-01 1.49915844e-01 4.00929958e-01 7.89793849e-01
-4.62713875e-02 -1.16078305e+00 -8.73642504e-01 8.48021328e-01
-5.33668816e-01 2.05438495e-01 -4.09834087e-01 1.09400952e+00
2.40159288e-01 1.01624215e+00 2.89638340e-01 -5.57975054e-01
4.13204312e-01 2.14718655e-01 -4.90948111e-02 -2.03041047e-01
-9.00098979e-01 -3.06197196e-01 2.97867656e-02 -3.18456978e-01
-3.50959748e-01 -6.04248941e-01 -1.02760589e+00 -5.40582061e-01
-7.88167238e-01 2.61510551e-01 9.54939067e-01 6.91647410e-01
5.53346336e-01 8.64839852e-01 7.22255111e-01 -4.72334057e-01
-7.12796271e-01 -1.13688707e+00 -3.92978370e-01 6.52927220e-01
5.28547801e-02 -9.89271045e-01 -7.35577762e-01 -4.19716537e-01] | [8.730539321899414, 4.307992458343506] |
3e6a73cf-f83f-4989-9467-b63b78a093bc | tlpg-tracker-joint-learning-of-target | null | null | https://www.ijcai.org/proceedings/2020/0099 | https://www.ijcai.org/proceedings/2020/0099.pdf | TLPG-Tracker: Joint Learning of Target Localization and Proposal Generation for Visual Tracking. | Target localization and proposal generation are two essential subtasks in generic visual tracking, and it is a challenge to address both the two efficiently. In this paper, we propose an efficient two-stage architecture which makes full use of the complementarity of two subtasks to achieve robust localization and high-quality proposals generation of the target jointly. Specifically, our model performs a novel deformable central correlation operation by an online learning model in both two stages to locate new target centers while generating target proposals in the vicinity of these centers. The proposals are refined in the refinement stage to further improve accuracy and robustness. Moreover, the model benefits from multi-level features aggregation in a neck module and a feature enhancement module. We conduct extensive ablation studies to demonstrate the effectiveness of our proposed methods. Our tracker runs at over 30 FPS and sets a new state-ofthe-art on five tracking benchmarks, including La-SOT, VOT2018, TrackingNet, GOT10k, OTB2015. | ['Feng Du', 'Linglong Qiu', 'Anna Wang', 'Ziyu Liu', 'Zhi Zhang', 'Siyuan Li'] | 2020-06-01 | null | null | null | international-joint-conference-on-artificial-7 | ['visual-tracking'] | ['computer-vision'] | [-2.83040971e-01 -2.60633707e-01 -2.24001467e-01 -1.76113188e-01
-6.75760090e-01 -6.44724965e-01 6.04810834e-01 -4.77900133e-02
-4.88686889e-01 5.16524017e-01 9.21025649e-02 7.77446199e-03
2.10904121e-01 -3.24207693e-01 -5.07707834e-01 -4.35868323e-01
-1.33606285e-01 2.54230291e-01 9.34444547e-01 -2.51927637e-02
-3.38400304e-02 6.85963869e-01 -1.41269338e+00 2.27117185e-02
6.95391953e-01 1.25239241e+00 2.88266540e-01 5.17982900e-01
6.36416972e-02 4.92854446e-01 -4.56853986e-01 -3.80862892e-01
5.69022477e-01 2.03265280e-01 -3.28469485e-01 1.18390322e-01
7.60909796e-01 -3.86172295e-01 -3.27980012e-01 1.10105443e+00
6.23839378e-01 -1.17613360e-01 2.88297087e-01 -1.27665234e+00
-2.61923850e-01 3.57486635e-01 -8.67169082e-01 3.67749214e-01
8.72834772e-03 5.39682388e-01 7.55244970e-01 -1.09270990e+00
4.83080566e-01 1.19965601e+00 8.75881076e-01 5.86205840e-01
-1.00068080e+00 -9.61132705e-01 5.03224194e-01 -9.51976404e-02
-1.26981020e+00 -6.34981930e-01 2.73638606e-01 -5.55572331e-01
5.87147176e-01 -1.30890861e-01 7.39735425e-01 8.68458986e-01
2.83949852e-01 6.93238080e-01 8.19209814e-01 -3.68906260e-02
-9.98644531e-02 -7.95712695e-02 3.83282490e-02 7.91613281e-01
4.70340937e-01 6.17988527e-01 -5.54615676e-01 -3.56762297e-02
9.19257641e-01 1.60441011e-01 -1.05091974e-01 -5.70742965e-01
-1.55872512e+00 6.10903502e-01 9.45551991e-01 9.11690593e-02
-3.24802369e-01 4.57062930e-01 2.72448301e-01 -6.35599270e-02
2.94771671e-01 1.23672426e-01 -3.12284559e-01 8.74766037e-02
-1.13464367e+00 2.46197328e-01 2.73241848e-01 1.18552482e+00
5.33748209e-01 1.91396475e-02 -8.92650783e-01 4.02951598e-01
7.41096854e-01 7.76927948e-01 1.30541608e-01 -7.07161665e-01
3.67364943e-01 6.02054179e-01 5.42586446e-01 -8.56010497e-01
-4.26122844e-01 -8.84476006e-01 -4.31101054e-01 2.21493304e-01
3.56626242e-01 -2.41037071e-01 -1.21803510e+00 1.69414937e+00
8.59206796e-01 6.48000717e-01 -2.89924651e-01 1.15382910e+00
9.34708655e-01 3.21281165e-01 4.17124778e-01 6.55567423e-02
1.43454456e+00 -1.40173376e+00 -5.96360505e-01 -2.57527500e-01
4.45640028e-01 -1.06384110e+00 3.44854623e-01 -7.85611793e-02
-9.89722788e-01 -9.84263480e-01 -9.36109543e-01 1.06680058e-01
4.27416749e-02 7.75685728e-01 7.93366849e-01 5.03637195e-01
-8.83009553e-01 3.28356445e-01 -1.05088043e+00 -2.12998480e-01
6.28418505e-01 4.10412431e-01 -1.91115215e-01 1.49859324e-01
-8.16817880e-01 7.26166010e-01 3.39595377e-01 4.30621475e-01
-1.19013155e+00 -8.45567703e-01 -6.85780406e-01 -7.36727268e-02
3.74921620e-01 -7.54557192e-01 1.19401741e+00 -3.73115182e-01
-1.34810042e+00 6.23934507e-01 -1.49352610e-01 -6.32986128e-01
6.96276188e-01 -3.90769392e-01 -3.48421663e-01 -3.08318228e-01
3.48186284e-01 1.02777457e+00 1.01395428e+00 -1.01743603e+00
-1.14625382e+00 -3.53527486e-01 -1.32426545e-01 -3.25778089e-02
-1.51952654e-01 1.07585691e-01 -9.45653021e-01 -8.39430273e-01
-3.39315347e-02 -9.61310089e-01 -3.88061285e-01 3.90158802e-01
-3.74285340e-01 -2.76641071e-01 9.07426178e-01 -4.35212493e-01
1.07733941e+00 -2.10049462e+00 -4.69986303e-03 6.82085007e-02
4.78743196e-01 5.57555437e-01 -1.74930692e-01 -7.12143034e-02
3.11995655e-01 -3.88476610e-01 4.35916066e-01 -5.63641369e-01
5.37830181e-02 -2.85957754e-01 -3.59302580e-01 5.71428299e-01
3.08467567e-01 1.26076794e+00 -9.16580498e-01 -6.41446054e-01
3.58021140e-01 4.57743376e-01 -3.97091001e-01 1.63600639e-01
-1.94701836e-01 5.38456261e-01 -7.93673456e-01 8.40445876e-01
7.87949681e-01 -3.67712647e-01 -1.47184014e-01 -5.12558281e-01
-5.31115949e-01 -7.87251890e-02 -1.20914853e+00 1.75608754e+00
-7.85856694e-02 3.69778693e-01 1.38505816e-01 -3.26364398e-01
8.32046270e-01 6.13189451e-02 5.06169140e-01 -4.96526420e-01
3.23321551e-01 6.60169125e-02 1.19396575e-01 6.39217393e-03
7.19214261e-01 2.96593577e-01 3.09683178e-02 7.76579082e-02
1.18932985e-01 4.08976674e-01 1.40712425e-01 1.92889601e-01
1.09817481e+00 3.90293270e-01 7.61435181e-02 -1.45076275e-01
5.16017556e-01 1.57752410e-01 9.45363700e-01 8.31424952e-01
-5.71153283e-01 3.08292776e-01 -1.27468690e-01 -4.71323520e-01
-5.92946291e-01 -1.06954920e+00 8.82098749e-02 9.60464537e-01
5.65183699e-01 -5.33935249e-01 -2.93989569e-01 -9.77625906e-01
1.94174379e-01 5.43695390e-02 -6.34612203e-01 -2.06513464e-01
-6.08699143e-01 -3.12996387e-01 4.65891957e-01 7.55746603e-01
5.11406243e-01 -9.28745329e-01 -7.89223135e-01 3.00126672e-01
-7.07528293e-02 -1.36887097e+00 -8.82516980e-01 -1.81568801e-01
-8.34351957e-01 -1.09712124e+00 -7.49393821e-01 -6.49460316e-01
6.11513317e-01 5.70804119e-01 8.30374599e-01 1.72515720e-01
-4.63208199e-01 8.61422420e-02 -2.36766532e-01 -3.89548212e-01
-1.86811555e-02 6.26563951e-02 1.76625233e-02 3.35822068e-02
1.68786887e-02 2.75167241e-03 -7.84072518e-01 7.55607009e-01
-3.08895558e-01 5.47138527e-02 8.47199857e-01 5.41590631e-01
6.62989736e-01 -3.44242156e-01 2.72177726e-01 -2.32264772e-01
1.90798175e-02 -1.73648462e-01 -1.27066505e+00 4.26417083e-01
-1.56340107e-01 -3.83993387e-02 1.93946451e-01 -5.86720824e-01
-8.46789062e-01 5.87217152e-01 9.89550203e-02 -8.18746030e-01
2.72962242e-01 7.58438045e-03 -1.36344448e-01 -6.65008605e-01
5.19847095e-01 6.08877465e-03 -1.45537153e-01 -4.13569808e-01
4.29091007e-01 1.34240866e-01 6.99819207e-01 -3.36587310e-01
1.39271939e+00 5.55222392e-01 -1.66226879e-01 -2.24439055e-01
-9.03410017e-01 -5.77154815e-01 -5.60451627e-01 -5.03073514e-01
6.83831751e-01 -1.26034594e+00 -9.55607593e-01 3.99602503e-01
-1.08092630e+00 -1.81968957e-01 -3.31186891e-01 6.52378440e-01
-9.79580283e-02 2.48930007e-01 -2.82905430e-01 -5.73502839e-01
-6.76214516e-01 -1.26243627e+00 1.37464499e+00 6.79923415e-01
4.20686007e-01 -6.33706927e-01 9.38171819e-02 -2.53795600e-03
6.19706929e-01 1.57311276e-01 -1.67640537e-01 -3.37314278e-01
-1.01923275e+00 -2.70243675e-01 -5.08366823e-01 -1.62660196e-01
4.46930416e-02 2.08643265e-02 -6.51587307e-01 -7.78952181e-01
-5.27901411e-01 -2.73632914e-01 1.20613623e+00 5.04749894e-01
7.62787104e-01 1.24908231e-01 -9.53255355e-01 8.80812883e-01
1.17809689e+00 -1.09136008e-01 2.85731167e-01 1.68174297e-01
7.16685116e-01 3.21919881e-02 1.21307659e+00 4.33196634e-01
4.50633556e-01 1.17182732e+00 5.34583092e-01 -7.09946528e-02
-4.81201470e-01 -1.62535459e-01 3.67067546e-01 3.68948311e-01
-5.78286313e-02 4.83347215e-02 -7.02685297e-01 5.82092464e-01
-2.05360103e+00 -8.87871146e-01 -6.29478693e-02 2.17131233e+00
5.26586175e-01 2.64838427e-01 3.09907883e-01 -6.02331161e-01
9.77144659e-01 1.13001980e-01 -5.04332244e-01 7.10960984e-01
2.62568474e-01 1.81195140e-01 6.83627367e-01 2.52538502e-01
-1.51450455e+00 1.41649806e+00 5.96071196e+00 9.12558794e-01
-1.14156365e+00 2.01725259e-01 7.57835954e-02 -8.47132802e-02
3.52143377e-01 1.69677585e-01 -1.50080621e+00 5.27681172e-01
3.74532104e-01 1.48956748e-02 -4.59027179e-02 9.06865299e-01
7.51549751e-02 6.37973994e-02 -7.85369694e-01 8.22394848e-01
-3.65427434e-01 -1.52819002e+00 -2.28655979e-01 9.31013450e-02
5.67730129e-01 2.52581745e-01 1.30225107e-01 4.35145736e-01
5.19920290e-01 -7.06443489e-01 9.79176223e-01 4.91339386e-01
5.30673146e-01 -5.24263144e-01 5.20183980e-01 9.09810886e-02
-1.98801756e+00 -1.24640223e-02 -4.02557254e-01 3.72640610e-01
3.58281642e-01 4.31159496e-01 -7.03874171e-01 6.97309017e-01
7.86445558e-01 8.60374570e-01 -9.28748250e-01 1.75686169e+00
-3.29212010e-01 2.73406088e-01 -4.22320992e-01 2.23586112e-01
2.24320158e-01 4.10132974e-01 6.97387338e-01 1.03384757e+00
2.61902004e-01 -1.14145249e-01 6.78513706e-01 7.74525225e-01
5.35419397e-03 -2.71169543e-01 -2.23626550e-02 1.46726921e-01
6.59668565e-01 1.86327553e+00 -8.23634446e-01 -1.86075777e-01
-3.34656447e-01 5.38178504e-01 4.19454575e-01 2.53835805e-02
-1.53257680e+00 -2.23556068e-02 6.29810750e-01 2.78097074e-02
7.30734527e-01 -3.15222442e-01 2.54985750e-01 -1.22376049e+00
3.42029147e-02 -5.45682788e-01 2.74699241e-01 -4.21361655e-01
-1.09229743e+00 6.82830632e-01 -3.61902148e-01 -1.55009532e+00
1.22689046e-01 -3.28844517e-01 -6.71847761e-01 7.64421165e-01
-1.55169106e+00 -1.57179344e+00 -7.72664785e-01 6.86083257e-01
2.84000337e-01 -1.25634462e-01 2.46774390e-01 5.26351213e-01
-5.46180606e-01 7.36827850e-01 -4.09517735e-01 3.76449287e-01
8.87673080e-01 -9.35373843e-01 8.24007928e-01 1.05236185e+00
1.21429563e-01 5.57338595e-01 3.60276967e-01 -9.81707394e-01
-1.23548770e+00 -1.59304595e+00 4.75893825e-01 -3.89630616e-01
6.09414816e-01 -3.47893953e-01 -5.69911778e-01 7.48131156e-01
-6.95585981e-02 7.34153032e-01 6.23017512e-02 -1.29271567e-01
-2.41040289e-01 -8.09141845e-02 -7.83206582e-01 4.53858286e-01
1.23763323e+00 1.27879843e-01 -4.23816890e-01 2.65860170e-01
7.53723741e-01 -8.94391239e-01 -8.09630096e-01 6.68993115e-01
5.76650023e-01 -7.31106520e-01 1.20576108e+00 -4.53917384e-01
-3.10029536e-01 -9.13750052e-01 2.75058508e-01 -9.42168415e-01
-4.89403844e-01 -7.83647895e-01 -3.81385297e-01 1.15885735e+00
2.60189027e-01 -4.19043303e-01 8.54628503e-01 4.33904789e-02
-2.27789715e-01 -6.79437399e-01 -9.73720491e-01 -8.73034775e-01
-5.68471014e-01 -1.62334517e-01 4.90972370e-01 3.84247124e-01
-5.48340619e-01 1.76260546e-01 -5.00657499e-01 4.16642815e-01
9.64227140e-01 3.86686146e-01 1.19165289e+00 -1.20791006e+00
-2.11167142e-01 -3.11837703e-01 -5.91912270e-01 -1.51537657e+00
-9.55571383e-02 -7.18746901e-01 3.06084394e-01 -1.37638569e+00
2.41676912e-01 -8.20972264e-01 -2.99010903e-01 6.25542343e-01
-4.01071995e-01 3.41391772e-01 5.29690325e-01 3.81129265e-01
-1.22431386e+00 7.46769965e-01 1.23833549e+00 -5.53367473e-02
-2.41920099e-01 3.67759466e-01 -4.95506257e-01 6.38323069e-01
3.47261310e-01 -6.49448633e-01 2.02091690e-02 -4.83124852e-01
-3.41964066e-01 6.23427040e-04 8.30357134e-01 -1.39215815e+00
6.80103421e-01 -5.81560172e-02 6.96693122e-01 -1.07243490e+00
4.15341139e-01 -9.14315462e-01 7.71761388e-02 7.39034355e-01
8.16342533e-02 4.14275751e-02 4.68358070e-01 7.39510238e-01
-6.42274171e-02 3.30972075e-01 8.70959461e-01 2.31876001e-01
-9.90013003e-01 8.14752936e-01 2.56639391e-01 3.39430831e-02
1.33875942e+00 -1.19888693e-01 -4.29909557e-01 9.36492905e-02
-7.46895909e-01 7.61684954e-01 3.77557307e-01 8.34706545e-01
6.07585669e-01 -1.52062643e+00 -5.66679835e-01 9.51683242e-03
1.79057941e-01 -7.00908378e-02 1.67574912e-01 1.30126393e+00
-2.36739025e-01 5.10337472e-01 -2.01966062e-01 -9.09411192e-01
-1.31867182e+00 5.63008726e-01 4.38665390e-01 -4.86112714e-01
-6.91789985e-01 8.87733281e-01 3.35679799e-01 -1.53190032e-01
5.47372639e-01 -2.49238104e-01 -1.71916232e-01 -2.95728803e-01
6.70129359e-01 9.71801877e-02 -1.79433748e-01 -8.49476755e-01
-7.14031458e-01 7.92034149e-01 -3.65255445e-01 1.50026947e-01
9.42337215e-01 4.19732295e-02 2.73861766e-01 -4.17348534e-01
5.02216220e-01 5.98940775e-02 -1.78228354e+00 -3.80332440e-01
1.99698478e-01 -6.61725998e-01 1.72570888e-02 -7.63882220e-01
-1.41285014e+00 4.66603845e-01 8.29072535e-01 -3.10898334e-01
8.78671110e-01 5.51342405e-02 8.26567352e-01 7.42296353e-02
4.39582705e-01 -7.25601196e-01 2.47028291e-01 4.79152918e-01
6.37633145e-01 -1.24300826e+00 8.99624527e-02 -6.15919411e-01
-3.58344853e-01 9.09522891e-01 1.05363834e+00 -1.47022501e-01
4.84929204e-01 4.21499729e-01 -5.23685478e-02 -1.58158526e-01
-6.58203304e-01 -6.70476913e-01 6.57963634e-01 3.87800694e-01
1.84837684e-01 -1.61977649e-01 -1.42893210e-01 4.68762308e-01
2.38460466e-01 2.13627461e-02 -1.74589843e-01 1.05657840e+00
-5.70094705e-01 -1.02240491e+00 -4.17801470e-01 1.10897101e-01
-3.64059746e-01 1.41606629e-01 -2.95410067e-01 9.54104722e-01
2.73541033e-01 8.19381654e-01 -1.82926685e-01 -4.42863554e-01
4.28605199e-01 -5.57139814e-01 5.57703614e-01 -4.94841367e-01
-8.08200896e-01 3.37691873e-01 -1.91785350e-01 -1.03010321e+00
-6.31092429e-01 -6.22538447e-01 -1.19655144e+00 -7.57256523e-02
-8.17178249e-01 7.47711360e-02 5.99896550e-01 8.72264326e-01
6.64659381e-01 7.26873577e-01 3.75097066e-01 -1.06993318e+00
-5.17659009e-01 -8.41690838e-01 -9.65666398e-02 -5.28960442e-03
2.39901185e-01 -1.13295579e+00 1.66828334e-01 -1.61021784e-01] | [6.322630882263184, -2.129281997680664] |
a6c4527f-f3de-43f7-ae3d-7d25824a0e5e | a-fast-maximum-k-plex-algorithm-parameterized | 2306.13258 | null | https://arxiv.org/abs/2306.13258v1 | https://arxiv.org/pdf/2306.13258v1.pdf | A Fast Maximum $k$-Plex Algorithm Parameterized by the Degeneracy Gap | Given a graph, the $k$-plex is a vertex set in which each vertex is not adjacent to at most $k-1$ other vertices in the set. The maximum $k$-plex problem, which asks for the largest $k$-plex from a given graph, is an important but computationally challenging problem in applications like graph search and community detection. So far, there is a number of empirical algorithms without sufficient theoretical explanations on the efficiency. We try to bridge this gap by defining a novel parameter of the input instance, $g_k(G)$, the gap between the degeneracy bound and the size of maximum $k$-plex in the given graph, and presenting an exact algorithm parameterized by $g_k(G)$. In other words, we design an algorithm with running time polynomial in the size of input graph and exponential in $g_k(G)$ where $k$ is a constant. Usually, $g_k(G)$ is small and bounded by $O(\log{(|V|)})$ in real-world graphs, indicating that the algorithm runs in polynomial time. We also carry out massive experiments and show that the algorithm is competitive with the state-of-the-art solvers. Additionally, for large $k$ values such as $15$ and $20$, our algorithm has superior performance over existing algorithms. | ['Mingyu Xiao', 'Chunyu Luo', 'Yi Zhou', 'Zhengren Wang'] | 2023-06-23 | null | null | null | null | ['community-detection'] | ['graphs'] | [-8.09405074e-02 3.32856506e-01 -1.67290643e-01 2.28338629e-01
-4.97165352e-01 -7.87942052e-01 -4.42356199e-01 4.62879479e-01
-2.57362932e-01 5.68619192e-01 -8.28855395e-01 -5.94145834e-01
-6.00044906e-01 -1.31719065e+00 -7.67194331e-01 -6.13079011e-01
-9.02796328e-01 7.49758720e-01 5.59889078e-01 -5.91160133e-02
2.47286275e-01 2.59277344e-01 -1.20312262e+00 -5.85145712e-01
8.65808427e-01 9.39871132e-01 3.85783613e-02 5.60949087e-01
-1.52702942e-01 7.57051036e-02 -3.80760252e-01 -5.18388867e-01
6.42151296e-01 -6.76386654e-01 -1.04670644e+00 3.64682704e-01
5.14624901e-02 3.57066154e-01 -6.10126019e-01 1.46862912e+00
5.03664427e-02 -1.01828821e-01 1.65499225e-01 -1.52664149e+00
-1.98700979e-01 1.07782400e+00 -9.54827249e-01 2.33354375e-01
3.07008535e-01 2.31379643e-01 1.30135226e+00 -2.45064884e-01
4.55350846e-01 1.07655203e+00 3.97630692e-01 -3.60322110e-02
-1.30506718e+00 -9.02340174e-01 4.56477523e-01 1.30326137e-01
-1.86188245e+00 2.85561502e-01 7.14471459e-01 -3.15143198e-01
7.87319660e-01 6.64846539e-01 8.18557084e-01 -1.44782424e-01
-4.52650815e-01 4.56831157e-02 9.63409007e-01 -4.42151189e-01
2.99476415e-01 -2.24980161e-01 1.80497214e-01 1.02756238e+00
1.06133044e+00 -9.56867039e-02 -2.11847812e-01 -2.37417504e-01
6.07686043e-01 -2.84251869e-01 -4.40012574e-01 -3.12084645e-01
-8.10126603e-01 9.51525688e-01 5.84752500e-01 2.73262799e-01
7.25772977e-02 6.68259621e-01 1.92430988e-01 4.66711015e-01
-1.17776636e-02 5.70317686e-01 -4.05572921e-01 1.00995928e-01
-6.67240798e-01 3.26072276e-02 9.77878571e-01 1.15810943e+00
1.00992334e+00 -1.78086534e-01 4.93075669e-01 2.12217420e-01
2.19773012e-03 5.84245026e-01 -5.28816700e-01 -7.44530380e-01
7.08574712e-01 1.12503099e+00 -6.20136932e-02 -1.30475080e+00
-2.40914181e-01 -2.76575506e-01 -8.96109939e-01 -2.05882192e-01
7.65487373e-01 2.82379016e-02 -5.72701097e-01 1.89217103e+00
4.69611108e-01 1.06324054e-01 -5.93626857e-01 5.99027276e-01
3.97747755e-01 7.42658317e-01 -2.94389933e-01 -7.39030719e-01
1.51972294e+00 -7.68090189e-01 -1.44507185e-01 -5.22064865e-01
8.26490760e-01 -5.00181079e-01 9.11155462e-01 2.54815042e-01
-1.23752081e+00 2.02037528e-01 -9.17784989e-01 5.64558804e-01
-2.11524040e-01 -3.33044827e-01 8.57525945e-01 1.03562248e+00
-1.08077371e+00 2.63366550e-01 -5.12718439e-01 -2.94587880e-01
8.14605877e-03 5.75147986e-01 -3.16848069e-01 -4.45596844e-01
-6.45630956e-01 3.27845901e-01 4.42725033e-01 -5.74582331e-02
-4.72400427e-01 -3.27797949e-01 -6.60220563e-01 3.08879465e-01
1.16832376e+00 -4.46899146e-01 5.83033144e-01 -4.45177495e-01
-6.99143589e-01 8.27899277e-01 -9.29675698e-02 -1.32849351e-01
9.17857513e-02 7.30458677e-01 -6.74599484e-02 4.47094083e-01
-1.91628367e-01 -1.86114982e-01 3.76071960e-01 -1.01536059e+00
-4.59067613e-01 -6.55955136e-01 5.37415087e-01 -8.82396102e-02
-2.03854397e-01 3.28754306e-01 -6.76275849e-01 8.02885517e-02
6.37187004e-01 -1.14667845e+00 -8.22325766e-01 -4.70988080e-02
-4.25727010e-01 -4.46354717e-01 3.07286084e-01 -1.26204297e-01
1.61673462e+00 -2.11713409e+00 2.45524608e-02 9.63208199e-01
7.43163407e-01 1.71927437e-01 -1.77979067e-01 7.64801264e-01
-1.80009678e-02 4.30085003e-01 -1.47559837e-01 2.36140519e-01
9.78459418e-02 1.51519656e-01 1.78638607e-01 8.47091675e-01
-4.75606769e-01 5.14908254e-01 -7.49133110e-01 -2.30222985e-01
-3.36232364e-01 -3.15070570e-01 -6.13770664e-01 -1.72633491e-02
-2.21483633e-01 -4.34926897e-01 -4.64251071e-01 4.96704787e-01
1.07849431e+00 -9.94482815e-01 8.91144454e-01 2.93390632e-01
9.17675253e-03 1.45016640e-01 -1.92941010e+00 1.01686847e+00
1.74737409e-01 1.14263467e-01 4.77631867e-01 -1.34317780e+00
7.58293450e-01 -1.42256930e-01 3.95107716e-01 -5.12857080e-01
1.23280652e-01 4.55501467e-01 3.23655486e-01 -1.98429391e-01
1.36754483e-01 -4.53189239e-02 -2.37621963e-01 8.35961342e-01
-4.99544859e-01 1.89596340e-01 6.67838991e-01 6.72478080e-01
1.87128448e+00 -9.35697138e-01 9.16087180e-02 -4.82858479e-01
2.91496903e-01 -5.82273938e-02 6.20614707e-01 7.29426503e-01
-1.60332650e-01 6.58848956e-02 1.19224727e+00 -2.66541243e-01
-8.15250278e-01 -6.99094474e-01 3.90782833e-01 7.39122629e-01
7.03417420e-01 -8.67834687e-01 -9.71754193e-01 -3.32053721e-01
-2.53582909e-03 -1.76461518e-01 -6.70759916e-01 7.27702975e-02
-6.10574186e-01 -7.90218890e-01 5.29937930e-02 3.56706470e-01
2.52046525e-01 -6.46892488e-01 -3.17148715e-01 2.72302985e-01
-1.09279469e-01 -1.22720718e+00 -6.42373323e-01 2.21627787e-01
-8.97326350e-01 -1.66295016e+00 1.82835979e-03 -1.07825172e+00
1.25286722e+00 6.09499276e-01 9.60442245e-01 5.90210617e-01
-5.35195768e-01 9.19772908e-02 -3.35601032e-01 8.22655186e-02
9.31334347e-02 1.04659691e-01 -2.59958893e-01 -4.81498659e-01
-3.14158015e-02 -6.47833705e-01 -7.47834980e-01 3.35826576e-01
-8.65343392e-01 -2.02771619e-01 3.57766509e-01 4.64575499e-01
7.38697112e-01 7.80127347e-01 5.06574400e-02 -9.95780289e-01
3.26527417e-01 -4.28641349e-01 -1.26291609e+00 1.77714080e-01
-7.63929486e-01 4.88585187e-03 8.21825624e-01 -4.15581584e-01
9.54178944e-02 1.20301686e-01 2.39750624e-01 -1.47994682e-01
6.59857810e-01 1.01660681e+00 -2.74984986e-01 -3.62482786e-01
3.83084893e-01 1.97220057e-01 -1.03180997e-01 -2.31046587e-01
2.32224464e-01 -4.72683161e-02 2.27595568e-01 -5.56936026e-01
1.05112112e+00 4.68912899e-01 5.60444176e-01 -5.39954901e-01
-2.53955722e-01 -4.93956566e-01 7.10178465e-02 3.52235255e-03
-1.22514263e-01 -5.80476761e-01 -1.47857475e+00 1.73936009e-01
-6.54323518e-01 -3.35659772e-01 5.09797148e-02 1.45657137e-01
-1.50191873e-01 7.27782845e-01 -4.46545720e-01 -1.12297380e+00
-2.36734793e-01 -9.95483100e-01 3.55363846e-01 -2.66462788e-02
1.71671376e-01 -4.88934755e-01 -6.91345185e-02 1.94451213e-01
2.23304227e-01 3.95599335e-01 1.36245418e+00 -4.20112312e-01
-1.20402730e+00 -2.70490110e-01 -5.78335226e-01 -2.77682394e-01
-9.47697740e-03 -1.50512889e-01 4.51171137e-02 -7.44661331e-01
-4.40228552e-01 -8.28134939e-02 5.71202159e-01 7.32569322e-02
1.35364127e+00 -7.17260540e-01 -5.72246909e-01 4.60605532e-01
1.92848611e+00 1.17599577e-01 7.07995415e-01 1.10987993e-02
4.13495243e-01 2.53536463e-01 4.70763803e-01 6.14224195e-01
2.58442432e-01 2.90996313e-01 7.08329499e-01 9.99092683e-02
3.22052777e-01 -9.32675302e-02 -8.32179412e-02 9.10025775e-01
-1.02535143e-01 -5.59758425e-01 -9.36957181e-01 6.84721172e-01
-1.76242471e+00 -6.79711878e-01 -3.63051772e-01 2.51128793e+00
6.81051850e-01 3.42091203e-01 4.96186137e-01 4.77625042e-01
1.00730419e+00 -3.00916601e-02 -4.91728485e-01 -4.09839928e-01
4.10032980e-02 8.30566525e-01 9.47246492e-01 4.11533177e-01
-4.60754603e-01 8.64725649e-01 6.03266621e+00 1.00508761e+00
-8.86916161e-01 -2.13331550e-01 5.56836903e-01 -1.11256175e-01
-3.48162681e-01 5.31604588e-01 -3.59514922e-01 6.15238369e-01
9.51800346e-01 -8.34087074e-01 8.85524452e-01 9.31816876e-01
-3.33218351e-02 -4.45230365e-01 -9.65289712e-01 1.21888411e+00
-1.87252149e-01 -1.17327392e+00 -4.01914209e-01 6.94781482e-01
6.09892249e-01 -4.20190662e-01 -8.84676650e-02 -4.14305460e-03
4.72087145e-01 -1.18202436e+00 6.28967062e-02 -5.53009391e-01
9.08461273e-01 -1.14467084e+00 2.21599609e-01 5.56710303e-01
-1.81527519e+00 -2.33640075e-01 -3.17797869e-01 1.71959400e-02
-7.12060109e-02 5.83726883e-01 -4.26122695e-01 4.63948756e-01
8.64654481e-01 -2.08547309e-01 -1.07730776e-01 1.25987399e+00
8.93521085e-02 5.18754542e-01 -8.68347108e-01 -3.01087886e-01
3.40957105e-01 -4.58542079e-01 3.57226551e-01 7.03586817e-01
4.83032972e-01 8.32923234e-01 4.83068466e-01 5.78865945e-01
-5.12291133e-01 4.15391594e-01 -4.35115427e-01 -5.25859177e-01
7.98510790e-01 1.08796608e+00 -1.30610669e+00 -1.13913216e-01
-1.40261456e-01 5.85193336e-01 5.76663733e-01 6.52128011e-02
-9.93424177e-01 -7.83126891e-01 6.63178623e-01 6.04364753e-01
7.05185771e-01 -5.09857714e-01 -5.53351901e-02 -6.65086150e-01
3.56254071e-01 -9.24640417e-01 7.28733540e-01 -3.15140277e-01
-9.81547415e-01 5.86174130e-01 -2.83379108e-01 -6.92686617e-01
3.58423054e-01 -6.62864208e-01 -4.82003003e-01 6.41994298e-01
-9.84413624e-01 -4.02971029e-01 -2.42901266e-01 6.62645400e-01
-3.72854143e-01 5.67137718e-01 6.19433284e-01 2.20258087e-01
-4.52466190e-01 8.11732233e-01 5.94028533e-02 -4.88020927e-02
4.64589670e-02 -9.82817054e-01 9.62454602e-02 1.10470760e+00
-2.72258252e-01 6.11630559e-01 7.58911669e-01 -3.40854526e-01
-2.01561403e+00 -7.03772008e-01 7.92139828e-01 2.38641620e-01
7.53680706e-01 -3.61732274e-01 -6.44171238e-01 5.27572870e-01
-2.21954659e-01 3.25414121e-01 6.93142414e-01 2.48404950e-01
-5.88663578e-01 -4.34771150e-01 -1.31422138e+00 5.56349158e-01
1.57477641e+00 -3.55689704e-01 4.12874907e-01 3.42232108e-01
7.25399137e-01 -4.28226352e-01 -8.04559767e-01 2.46153891e-01
5.99889159e-02 -6.47424579e-01 7.99663782e-01 -3.87856364e-01
-3.84135805e-02 -4.20665413e-01 1.64673738e-02 -9.26000774e-01
-4.66935873e-01 -1.03801036e+00 -2.42248058e-01 5.86029112e-01
5.83058000e-01 -8.66200268e-01 1.18406713e+00 4.60633695e-01
3.64437670e-01 -9.99454021e-01 -1.25156963e+00 -1.04565752e+00
1.13771334e-01 -9.09119770e-02 9.00152922e-01 9.79771256e-01
4.36824024e-01 3.72860521e-01 -1.34873062e-01 4.78979707e-01
8.10550094e-01 7.21312046e-01 9.43103313e-01 -1.30631971e+00
-6.41211331e-01 -4.85337555e-01 -3.26487631e-01 -1.02571690e+00
-2.07773954e-01 -8.51904392e-01 -2.47087732e-01 -1.48267543e+00
7.12189376e-01 -1.00803292e+00 -2.20232103e-02 4.34066713e-01
-1.30409822e-01 8.45362842e-02 1.35699928e-01 -2.25223228e-01
-8.44537139e-01 1.86547078e-02 1.31355321e+00 -1.76573321e-02
-2.07767993e-01 -2.07333297e-01 -1.03056502e+00 3.88465762e-01
6.74343407e-01 -5.14629662e-01 -2.93792844e-01 -8.59830603e-02
7.79865444e-01 5.69697440e-01 -5.77714592e-02 -5.76849103e-01
1.92258403e-01 -6.01597726e-01 -5.78836441e-01 -4.32314873e-01
1.63924336e-01 -7.91544974e-01 6.44203305e-01 9.49632525e-01
9.47726890e-02 2.76418597e-01 -4.66564260e-02 6.27007008e-01
4.45264615e-02 -1.76157147e-01 8.30602109e-01 -8.62401500e-02
-2.57932931e-01 8.17707121e-01 2.83386167e-02 2.75705546e-01
1.30659997e+00 -5.80713034e-01 -8.29364419e-01 -3.28763932e-01
-3.75279158e-01 4.86073464e-01 4.61915731e-01 -2.42695987e-01
4.84023094e-01 -9.04884815e-01 -5.65391302e-01 -1.96512654e-01
3.34931731e-01 1.61466837e-01 2.72995681e-01 8.27213764e-01
-8.37732553e-01 2.14231938e-01 3.25599253e-01 -3.10208946e-01
-1.53426909e+00 9.47761059e-01 1.36777654e-01 -5.69054902e-01
-3.64679426e-01 1.00663185e+00 1.39061734e-01 1.64491579e-01
2.24646360e-01 -1.99872166e-01 5.82039535e-01 -5.97592354e-01
3.77343535e-01 5.30578732e-01 -2.45396927e-01 -2.76879787e-01
-5.45189619e-01 3.87191892e-01 -1.15340821e-01 4.79147471e-02
1.26292813e+00 -1.70574784e-02 -6.78613067e-01 -3.17819476e-01
1.08999848e+00 8.69556591e-02 -5.08904517e-01 -3.48290920e-01
-6.64838180e-02 -8.08368444e-01 -4.25292760e-01 -2.44599909e-01
-1.48317444e+00 3.36184323e-01 1.95450649e-01 7.68918037e-01
1.39831913e+00 5.34731448e-01 7.02186346e-01 4.11059618e-01
9.50792074e-01 -1.01550126e+00 5.92834651e-02 3.84925157e-01
3.38132799e-01 -6.48251891e-01 1.73089921e-01 -1.32683337e+00
1.16376944e-01 7.14826882e-01 8.29212368e-01 -2.90877879e-01
7.37700045e-01 1.26065359e-01 -7.62090445e-01 -4.52835917e-01
-7.53682077e-01 -3.39859903e-01 -2.98736215e-01 3.68491858e-02
-1.54731914e-01 4.48754996e-01 -8.03860605e-01 6.55611038e-01
-3.01144212e-01 -2.68028796e-01 5.38909256e-01 8.88670206e-01
-9.11227107e-01 -1.44404149e+00 -2.36107185e-01 5.05080402e-01
-3.23169112e-01 -1.38251543e-01 -5.22480309e-01 7.19239652e-01
-3.07985563e-02 9.87287402e-01 -2.79372185e-02 -1.82020143e-01
8.76241624e-02 -5.33414006e-01 8.68226469e-01 -5.01345754e-01
-4.33519751e-01 7.28823896e-03 9.91468877e-02 -5.33872485e-01
1.90471970e-02 -6.35565519e-02 -1.38383853e+00 -1.15104496e+00
-6.75108016e-01 5.13763309e-01 4.17208672e-01 5.95444202e-01
3.51288468e-01 -7.08286688e-02 9.34919477e-01 1.42970353e-01
-2.57504582e-01 -4.76906776e-01 -1.18703818e+00 -2.24452559e-02
-3.08731198e-01 -3.71742457e-01 -7.33680010e-01 -4.70841467e-01] | [6.865138053894043, 5.182799816131592] |
454a957a-d059-4717-a7db-fef34258fbf3 | glitch-in-the-matrix-a-large-scale-benchmark | 2305.01979 | null | https://arxiv.org/abs/2305.01979v2 | https://arxiv.org/pdf/2305.01979v2.pdf | "Glitch in the Matrix!": A Large Scale Benchmark for Content Driven Audio-Visual Forgery Detection and Localization | Most deepfake detection methods focus on detecting spatial and/or spatio-temporal changes in facial attributes. This is because available benchmark datasets contain mostly visual-only modifications. However, a sophisticated deepfake may include small segments of audio or audio-visual manipulations that can completely change the meaning of the content. To addresses this gap, we propose and benchmark a new dataset, Localized Audio Visual DeepFake (LAV-DF), consisting of strategic content-driven audio, visual and audio-visual manipulations. The proposed baseline method, Boundary Aware Temporal Forgery Detection (BA-TFD), is a 3D Convolutional Neural Network-based architecture which efficiently captures multimodal manipulations. We further improve (i.e. BA-TFD+) the baseline method by replacing the backbone with a Multiscale Vision Transformer and guide the training process with contrastive, frame classification, boundary matching and multimodal boundary matching loss functions. The quantitative analysis demonstrates the superiority of BA- TFD+ on temporal forgery localization and deepfake detection tasks using several benchmark datasets including our newly proposed dataset. The dataset, models and code are available at https://github.com/ControlNet/LAV-DF. | ['Munawar Hayat', 'Kalin Stefanov', 'Abhinav Dhall', 'Tom Gedeon', 'Shreya Ghosh', 'Zhixi Cai'] | 2023-05-03 | null | null | null | null | ['face-swapping'] | ['computer-vision'] | [ 4.37395014e-02 -4.43292439e-01 -8.68636593e-02 -1.50165334e-01
-6.50869429e-01 -4.83767778e-01 4.61927563e-01 -3.31001967e-01
-3.00651486e-03 2.55062252e-01 3.81546795e-01 5.96694350e-02
2.85980105e-01 -4.09005940e-01 -8.07363153e-01 -6.76060498e-01
1.54882640e-01 -4.05321538e-01 3.17338586e-01 -1.79527432e-01
2.71562099e-01 6.27431631e-01 -1.76432097e+00 6.76774561e-01
3.76483411e-01 1.56037676e+00 -1.50545344e-01 5.67931652e-01
1.23521857e-01 6.34428561e-01 -5.67774296e-01 -4.72833782e-01
3.21151674e-01 -3.64939868e-01 -5.13448536e-01 2.07404330e-01
8.79160047e-01 -8.99178565e-01 -6.50693536e-01 8.13312769e-01
9.39653933e-01 7.22907558e-02 3.82477194e-01 -1.71349144e+00
-6.24441743e-01 1.43169895e-01 -9.28158879e-01 4.58255351e-01
3.87078971e-01 4.03623223e-01 5.73919535e-01 -1.46687937e+00
5.85779905e-01 1.58128071e+00 1.04986203e+00 6.57453835e-01
-1.07711947e+00 -1.03061795e+00 6.34785071e-02 6.01258099e-01
-1.52696717e+00 -1.05554044e+00 1.17088664e+00 -4.86406356e-01
7.98399448e-01 2.00569838e-01 6.70021057e-01 1.46874177e+00
1.26601622e-01 8.38313222e-01 6.39844835e-01 -2.59637892e-01
1.57418894e-03 -4.64060992e-01 -1.59252971e-01 7.53902853e-01
-2.18600169e-01 4.58565623e-01 -8.56406331e-01 -4.11021188e-02
6.94936931e-01 -9.75615308e-02 -4.26157773e-01 -3.45817000e-01
-9.01246250e-01 4.70171422e-01 9.02920440e-02 1.01652354e-01
-5.85714802e-02 4.96444941e-01 8.72266471e-01 5.22418261e-01
3.61399233e-01 -1.94073886e-01 -8.83889347e-02 -9.45207700e-02
-1.10421574e+00 4.82485026e-01 5.52862212e-02 7.17677653e-01
5.30055285e-01 2.87095606e-01 -6.98887467e-01 8.71585906e-01
1.45731226e-01 2.78860003e-01 5.33174336e-01 -1.27500248e+00
2.31803000e-01 3.58774900e-01 8.19881782e-02 -1.26504016e+00
-4.51048166e-01 -1.74473032e-01 -8.30250382e-01 3.71164590e-01
3.99090886e-01 1.55643463e-01 -8.93933356e-01 1.80697429e+00
5.94779491e-01 5.09245276e-01 -5.75931907e-01 1.11895955e+00
1.19848084e+00 2.96077132e-01 -1.04846656e-01 -1.06258117e-01
1.18478823e+00 -9.61703718e-01 -1.09450173e+00 2.78535753e-01
4.88670707e-01 -9.53353763e-01 1.32080996e+00 5.39633334e-01
-1.32026732e+00 -7.50478685e-01 -8.43840420e-01 -4.82751489e-01
-2.45228946e-01 3.41670156e-01 2.20083535e-01 7.59116948e-01
-1.30226231e+00 4.37787920e-01 -4.82634753e-01 -2.92082489e-01
7.05437660e-01 1.64242119e-01 -4.68802184e-01 4.53801230e-02
-1.26865947e+00 4.09084707e-01 6.69901818e-02 4.03755695e-01
-1.49490857e+00 -8.76561582e-01 -8.87350738e-01 -4.69505526e-02
4.27579254e-01 -5.86123347e-01 1.38448358e+00 -1.15753686e+00
-1.45968151e+00 1.05269325e+00 -2.54220814e-01 -1.31937593e-01
7.76804745e-01 -6.84312880e-02 -3.87905627e-01 3.23981732e-01
-1.16573848e-01 8.44328701e-01 1.64608788e+00 -1.20515168e+00
-4.07530218e-01 -3.32760572e-01 5.37105612e-02 -1.67851746e-01
-3.85548174e-01 2.75981694e-01 -4.99752641e-01 -1.16215718e+00
-3.01153958e-01 -4.19116348e-01 5.98463356e-01 6.57086253e-01
-3.58947575e-01 -1.83892176e-01 1.45474613e+00 -8.61545503e-01
1.57040620e+00 -2.47476649e+00 1.88060608e-02 -2.05225334e-01
6.09006524e-01 4.74584788e-01 -4.11923856e-01 2.62963980e-01
-2.97631115e-01 7.25094527e-02 5.01397960e-02 -7.78633118e-01
6.21470250e-02 -1.17425896e-01 -3.09218109e-01 6.55511737e-01
1.70609578e-01 9.07634258e-01 -5.76880395e-01 -4.17848676e-01
4.22502011e-01 4.40814525e-01 -7.06790686e-01 -3.87870967e-02
-5.76895569e-03 2.42017999e-01 1.30368516e-01 1.23194814e+00
1.10555172e+00 1.79725900e-01 -4.09634352e-01 -7.29114115e-01
8.05979222e-03 -2.97602564e-01 -9.62871492e-01 1.96238637e+00
-2.73946494e-01 9.30829406e-01 3.94681722e-01 -8.44084680e-01
6.67068481e-01 3.43087912e-01 5.60649395e-01 -9.60886300e-01
3.78703773e-01 2.51815859e-02 -4.33554590e-01 -8.95032644e-01
5.03665984e-01 9.37335193e-02 1.50644541e-01 2.89872736e-01
9.89796966e-02 3.80520552e-01 -1.30686998e-01 1.20632134e-01
1.14609063e+00 1.02282912e-01 -1.19025044e-01 1.55465409e-01
6.17225647e-01 -4.93668973e-01 7.08904684e-01 4.87746626e-01
-9.18127120e-01 7.96824455e-01 6.43657506e-01 -3.27511758e-01
-1.00587475e+00 -1.14650679e+00 1.11366538e-02 1.34973443e+00
2.15454921e-01 -5.39605081e-01 -9.08326507e-01 -6.35282934e-01
2.36189961e-01 2.56083608e-01 -9.14639235e-01 -5.01097560e-01
-4.35409307e-01 -1.41202867e-01 1.07897007e+00 4.07802075e-01
8.35344434e-01 -9.66030836e-01 -3.89806777e-01 -1.08048446e-01
-5.42999208e-01 -1.07974100e+00 -1.00575292e+00 -5.91842294e-01
-3.56152266e-01 -1.03832424e+00 -7.05490708e-01 -8.27594638e-01
1.74004182e-01 5.18674135e-01 8.18979442e-01 5.62068038e-02
-7.13750243e-01 4.83734727e-01 -3.93990993e-01 2.03030203e-02
-2.29904547e-01 -5.08298218e-01 -4.17914763e-02 5.17388582e-01
2.24516734e-01 -7.57887483e-01 -9.00734484e-01 4.69252944e-01
-9.80635703e-01 -8.95862877e-02 2.97020406e-01 9.31747079e-01
3.82165849e-01 -1.71899542e-01 5.86327732e-01 -9.03910697e-02
6.74415648e-01 -2.77296901e-01 -2.22295806e-01 8.92882794e-02
-1.32928148e-01 -4.81779844e-01 2.37696916e-01 -7.02903748e-01
-9.83264446e-01 -1.45570561e-01 -1.56556711e-01 -1.29187286e+00
-2.37824142e-01 7.84283131e-02 -4.44768071e-01 -3.42879385e-01
5.31139076e-01 2.10375488e-01 3.63916419e-02 -6.16369426e-01
2.76434928e-01 7.08340764e-01 8.41604650e-01 -3.43592912e-01
4.03699756e-01 7.36985862e-01 -1.33508146e-01 -7.80989110e-01
-2.50249803e-01 -2.88623571e-01 -4.54473078e-01 -6.99400365e-01
6.16083086e-01 -8.85403931e-01 -9.16548491e-01 1.12880588e+00
-1.31488132e+00 -5.20163774e-01 -1.22458607e-01 1.05830237e-01
-6.03373289e-01 8.23397577e-01 -6.96076512e-01 -6.68287456e-01
-3.22322756e-01 -1.13098288e+00 1.43044090e+00 -2.44279370e-01
-1.00113839e-01 -6.63277805e-01 -5.98643459e-02 3.92020375e-01
3.93415928e-01 4.86731797e-01 8.14144373e-01 2.29857620e-02
-3.20019901e-01 1.12209894e-01 -4.24060255e-01 5.30502677e-01
1.75590307e-01 1.81430504e-01 -1.34286869e+00 -3.57360274e-01
-2.05935121e-01 -5.25351465e-01 1.06332850e+00 5.58714569e-01
1.50193834e+00 -2.96967566e-01 -8.21315050e-02 8.39718044e-01
1.06712568e+00 5.81560917e-02 7.37309456e-01 1.77624494e-01
6.12862110e-01 5.12832344e-01 5.06115079e-01 8.60796690e-01
2.55214661e-01 1.03873885e+00 7.14213073e-01 -1.77236676e-01
-6.85736001e-01 -2.50434756e-01 7.74314106e-01 2.68801332e-01
1.28254160e-01 -2.47156262e-01 -6.68227673e-01 6.06252849e-01
-1.73185527e+00 -1.00927043e+00 1.02465428e-01 1.86825490e+00
6.89108789e-01 -2.56404757e-01 3.91520321e-01 6.10746861e-01
1.01905155e+00 2.64360726e-01 -6.94872499e-01 -4.26378161e-01
-1.35953411e-01 1.25322685e-01 2.38633435e-02 4.07449365e-01
-1.23633039e+00 1.01409745e+00 5.35493040e+00 1.30575788e+00
-1.24820745e+00 3.78935754e-01 5.19969583e-01 -4.29775178e-01
-9.91472453e-02 -5.68296194e-01 -4.27487135e-01 6.05330348e-01
4.61068630e-01 4.40138206e-02 3.73779655e-01 3.98751616e-01
1.03256142e+00 1.58736911e-02 -1.07295787e+00 1.46181393e+00
2.96764821e-01 -1.33760023e+00 2.54822373e-01 -1.69518605e-01
2.06426650e-01 -4.95652050e-01 4.07447577e-01 1.36285827e-01
-2.90259510e-01 -1.02808917e+00 1.05064547e+00 5.19953668e-01
1.08588946e+00 -7.59570897e-01 3.29740793e-01 -1.82953835e-01
-1.38236094e+00 -2.31196404e-01 -8.96405652e-02 1.40577003e-01
-7.68325403e-02 3.50764990e-01 -1.83712721e-01 3.42612237e-01
1.22788525e+00 9.53370988e-01 -5.58953166e-01 1.02983963e+00
2.13041544e-01 4.49056476e-01 -1.53710440e-01 6.51402891e-01
-4.67101075e-02 4.13124532e-01 6.90194011e-01 1.23757112e+00
3.94195050e-01 -3.24538141e-01 -1.14933513e-01 8.78467083e-01
-2.16652170e-01 -1.23847239e-01 -5.74739277e-01 1.99849680e-01
5.79997003e-01 1.17041886e+00 -2.63425976e-01 -5.17580248e-02
-3.39676410e-01 1.15251505e+00 4.71002236e-02 3.60981166e-01
-1.23482847e+00 -4.27019626e-01 1.02096236e+00 1.84425667e-01
4.78009909e-01 4.21390161e-02 6.46157637e-02 -1.00914729e+00
2.39135310e-01 -9.23499167e-01 3.74684244e-01 -1.02387238e+00
-1.17072213e+00 2.79253721e-01 -1.40012145e-01 -1.23318672e+00
1.22709215e-01 -3.52352202e-01 -5.39011776e-01 2.75072813e-01
-1.43650937e+00 -1.35310614e+00 -7.61921346e-01 1.26024818e+00
7.02738464e-01 -2.70588845e-02 3.14820349e-01 8.01133215e-01
-8.27701509e-01 1.36517322e+00 -9.37658995e-02 2.38168418e-01
1.09377205e+00 -5.71921825e-01 2.06768632e-01 7.71885753e-01
-4.68284518e-01 2.85190362e-02 5.77870846e-01 -3.67112756e-01
-1.30182934e+00 -1.36452711e+00 5.40227950e-01 -1.41827509e-01
6.10960782e-01 -5.73240995e-01 -8.55437934e-01 5.34818292e-01
1.89191595e-01 3.22560489e-01 3.57186526e-01 -5.84189475e-01
-5.87186694e-01 -3.41390431e-01 -1.48411286e+00 4.42757159e-01
1.38254237e+00 -5.99539876e-01 -2.06327006e-01 1.19396210e-01
7.12736666e-01 -2.96009511e-01 -6.70713305e-01 5.01620412e-01
7.53630996e-01 -1.41651988e+00 1.06307054e+00 -4.13646936e-01
4.31151092e-01 -2.20268622e-01 -1.76307246e-01 -9.98488307e-01
-1.61526799e-01 -9.89662766e-01 -4.32499558e-01 1.43462431e+00
-4.37249273e-01 -3.06698859e-01 6.94344282e-01 9.46663842e-02
-3.75908494e-01 -5.22460878e-01 -1.26438785e+00 -8.44647288e-01
-5.67688383e-02 -6.13437235e-01 5.65041542e-01 1.08227694e+00
-2.37817705e-01 -2.44858652e-01 -6.26933157e-01 -3.08594592e-02
5.63099861e-01 -1.36843264e-01 7.04576433e-01 -8.02315652e-01
3.48976962e-02 -6.73558116e-01 -6.66352391e-01 -8.74227405e-01
1.01933859e-01 -5.49636960e-01 -2.46276438e-01 -9.19126809e-01
-3.34184952e-02 5.30377068e-02 -1.75033763e-01 7.38932014e-01
2.34174922e-01 5.64702272e-01 3.55537124e-02 6.12803884e-02
-5.33962488e-01 9.14716899e-01 1.35688019e+00 -1.57380491e-01
-1.12204142e-01 -3.63307536e-01 -2.26445302e-01 6.15315676e-01
7.21229196e-01 -1.24214068e-01 -3.07060957e-01 -4.56480384e-01
-4.04087864e-02 1.56843811e-01 1.01820207e+00 -9.86471236e-01
8.86646211e-02 -3.63866463e-02 4.65754092e-01 -6.01375461e-01
5.27816892e-01 -6.89612448e-01 3.40224467e-02 3.44292492e-01
-4.92204875e-01 1.21676959e-01 5.42461574e-01 6.45617366e-01
-3.56086493e-01 2.67211497e-01 9.37254906e-01 1.44102201e-01
-8.63623679e-01 4.41076934e-01 -6.09647512e-01 -5.54752443e-03
1.06378508e+00 -5.80250621e-01 -5.03723919e-01 -4.88450408e-01
-9.27284718e-01 -9.95159894e-02 3.28356504e-01 5.75437307e-01
1.06374037e+00 -1.73433042e+00 -5.76392114e-01 4.15267318e-01
1.68918759e-01 -3.70422214e-01 7.57961214e-01 1.03609419e+00
-4.75903749e-01 1.00116260e-01 -5.04682422e-01 -5.37976146e-01
-1.62517285e+00 6.90079927e-01 6.88276827e-01 3.56196284e-01
-5.30432045e-01 9.38502789e-01 4.73215938e-01 4.44893986e-02
7.54181027e-01 -5.95910609e-01 -5.91483302e-02 1.03853904e-01
5.70563018e-01 7.25553334e-01 1.72799915e-01 -7.36506283e-01
-3.42165083e-01 5.19259572e-01 -3.02548651e-02 8.25125426e-02
8.73425961e-01 -5.38486302e-01 6.50407225e-02 1.15884647e-01
1.39898384e+00 -1.04894906e-01 -1.45768535e+00 -2.73187816e-01
-4.91276413e-01 -7.82468140e-01 3.65316808e-01 -6.00502133e-01
-1.51241529e+00 1.14033759e+00 1.05113840e+00 -1.91867292e-01
1.68689585e+00 -1.14508249e-01 1.10122061e+00 -5.07414620e-03
-7.34968260e-02 -9.11496758e-01 6.69597149e-01 1.22033320e-01
1.26069045e+00 -9.32956696e-01 -3.49006742e-01 -4.08513367e-01
-3.40980828e-01 1.13759100e+00 7.73029029e-01 7.33546838e-02
8.45601082e-01 2.80272752e-01 -2.60097366e-02 -1.12809777e-01
-7.63078034e-01 -1.47082314e-01 1.15295395e-01 6.29365087e-01
-2.62934510e-02 -4.78718638e-01 1.01532504e-01 6.25789046e-01
9.87039655e-02 3.03978056e-01 3.02169412e-01 7.92224467e-01
-2.47746751e-01 -7.21950352e-01 -4.37873900e-01 1.51757626e-02
-4.48742956e-01 -1.21446326e-02 -6.83601022e-01 7.56834567e-01
4.58444566e-01 1.05378199e+00 1.24931961e-01 -7.76105642e-01
2.78346986e-01 -4.64771464e-02 6.22626126e-01 6.06517121e-02
-6.09019041e-01 3.06278050e-01 -1.63981304e-01 -1.10407472e+00
-3.71307403e-01 -5.79252541e-01 -9.10752237e-01 -7.51822114e-01
-1.07343003e-01 -5.80502629e-01 2.20795751e-01 6.26179695e-01
5.66537440e-01 4.15626973e-01 5.90082765e-01 -1.18206859e+00
-2.19277367e-01 -8.93529832e-01 -5.79432130e-01 5.45572102e-01
8.56028378e-01 -1.13440669e+00 -4.72866505e-01 2.29552969e-01] | [13.011922836303711, 1.2071019411087036] |
bc19ce5e-7b31-4d3f-b39f-7184ec9b4b41 | endtimes-at-semeval-2021-task-7-detecting-and | null | null | https://aclanthology.org/2021.semeval-1.172 | https://aclanthology.org/2021.semeval-1.172.pdf | EndTimes at SemEval-2021 Task 7: Detecting and Rating Humor and Offense with BERT and Ensembles | This paper describes Humor-BERT, a set of BERT Large based models that we used in the SemEval-2021 Task 7: Detecting and Rating Humor and Offense. It presents pre and post processing techniques, variable threshold learning, meta learning and Ensemble approach to solve various sub-tasks that were part of the challenge. We also present a comparative analysis of various models we tried. Our method was ranked 4th in Humor Controversy Detection, 8th in Humor Detection, 19th in Average Offense Score prediction and 40th in Average Humor Score prediction globally. F1 score obtained for Humor classification was 0.9655 and for Controversy detection it was 0.6261. Our user name on the leader board is ThisIstheEnd and team name is EndTimes. | ['Karan Mangla', 'Chirag Singh', 'Chandan Kumar Pandey'] | 2021-08-01 | null | null | null | semeval-2021 | ['humor-detection'] | ['natural-language-processing'] | [-5.63835800e-01 1.45065054e-01 1.34381935e-01 2.59916216e-01
-4.48167026e-01 -3.49162847e-01 7.92997777e-01 3.64793926e-01
-2.32142121e-01 1.04492974e+00 9.21050370e-01 -6.56834617e-02
-1.38352394e-01 -3.96166712e-01 1.39968172e-02 -1.48259282e-01
2.76773989e-01 6.22335315e-01 1.05381601e-01 -9.18667495e-01
1.33063376e+00 -7.87917376e-02 -1.26416278e+00 9.92257893e-01
7.08523035e-01 4.11784023e-01 -1.79328293e-01 1.14932871e+00
2.93743581e-01 2.11502290e+00 -1.09960151e+00 -3.97574782e-01
-3.36498246e-02 -7.14289367e-01 -1.10101914e+00 -1.73155397e-01
4.50471044e-01 1.84228599e-01 -2.10610956e-01 7.68607020e-01
8.04106057e-01 3.35913688e-01 6.22456729e-01 -1.31584620e+00
-3.47241163e-01 9.49128509e-01 -6.58291757e-01 5.42507172e-01
8.19339752e-01 -4.16426621e-02 1.07790554e+00 -1.32650495e+00
7.62543261e-01 1.17401457e+00 8.26137781e-01 2.70769179e-01
-8.37287843e-01 -6.01503193e-01 -8.58401835e-01 8.98536980e-01
-6.36869729e-01 -3.10768604e-01 7.65525043e-01 -1.10237753e+00
1.13975394e+00 3.10554147e-01 3.83734614e-01 8.59771729e-01
7.00876355e-01 6.61058843e-01 1.72458434e+00 -3.23353052e-01
1.84078619e-01 3.87868553e-01 8.06083798e-01 6.41733170e-01
-1.99085042e-01 -1.52443767e-01 -1.17337632e+00 -4.24446762e-01
2.57639170e-01 -5.82498550e-01 -1.05593443e-01 8.00480664e-01
-1.02289808e+00 1.10143387e+00 1.10368878e-01 6.10289395e-01
-4.52273309e-01 -2.70397693e-01 6.70945823e-01 7.71990180e-01
1.06587559e-01 9.81468618e-01 -1.16188891e-01 -4.07697290e-01
-1.09201777e+00 7.16613293e-01 1.24155080e+00 4.07202780e-01
1.20956421e-01 1.08881406e-01 -5.59810638e-01 1.05692971e+00
-1.78207368e-01 -1.56040248e-02 7.77791381e-01 -1.20757699e+00
2.44632602e-01 8.29271495e-01 4.00475919e-01 -1.21843040e+00
-7.73936987e-01 -5.86881280e-01 -5.71331739e-01 6.97767198e-01
1.18156314e-01 -1.56117544e-01 -5.06318390e-01 9.66413558e-01
-2.83077389e-01 -8.30690563e-02 -1.79197833e-01 9.41147625e-01
1.51929033e+00 6.22281849e-01 -2.00938731e-01 -5.24128675e-01
1.34603465e+00 -1.59579968e+00 -7.57639706e-01 1.20328739e-01
2.86019742e-01 -1.29814196e+00 9.25440431e-01 1.04428518e+00
-1.52765453e+00 -5.29114246e-01 -1.22760689e+00 -6.77772239e-02
-2.40916193e-01 -2.93099493e-01 -9.10267606e-02 3.18641067e-01
-7.78867602e-01 8.20447087e-01 3.60537440e-01 -2.30216533e-01
-6.30380139e-02 1.98965803e-01 -9.23053622e-02 5.26564598e-01
-1.33177269e+00 1.67019486e+00 2.30354637e-01 -6.98746860e-01
-1.03247392e+00 -4.23342288e-01 -1.82874277e-01 -1.22310124e-01
8.86289179e-02 -4.20804679e-01 1.39260185e+00 -8.05008948e-01
-1.22783077e+00 1.26681674e+00 2.42008120e-01 -6.29122972e-01
6.69283867e-01 -2.99920589e-01 -4.43330556e-01 -2.66097456e-01
1.58336684e-01 -1.69964299e-01 5.84950089e-01 -1.12156105e+00
-7.68814266e-01 -2.88503408e-01 -1.23811148e-01 2.99236774e-01
-2.03768965e-02 6.10019445e-01 5.83259046e-01 -3.31086844e-01
-2.47103125e-01 -4.64176863e-01 5.47904484e-02 -1.24670482e+00
-3.00483137e-01 -3.54308754e-01 8.31213653e-01 -1.01681411e+00
1.80954790e+00 -1.36772895e+00 4.90785062e-01 3.01323999e-02
6.51340961e-01 1.55994058e-01 3.48207235e-01 8.57545137e-01
-2.02997938e-01 -1.96101125e-02 2.50112534e-01 1.66427959e-02
-8.43664929e-02 -3.66033763e-01 -2.43190348e-01 3.19264084e-01
-6.66441917e-01 5.59393406e-01 -8.04228425e-01 -7.58853495e-01
1.78805947e-01 -1.38493314e-01 -1.17668517e-01 4.22583491e-01
8.60048980e-02 1.97740141e-02 2.15327963e-01 7.36008883e-01
2.69841492e-01 -1.31856292e-01 -1.76546663e-01 2.68470109e-01
-4.14503306e-01 3.27482462e-01 -7.53774226e-01 8.63378227e-01
-3.47451627e-01 9.42821503e-01 -8.73507634e-02 -3.70037228e-01
1.36019349e+00 4.79205698e-01 2.68085420e-01 -5.69470525e-01
5.92618704e-01 3.10327142e-01 1.39980480e-01 -7.38517880e-01
9.16703463e-01 -6.44256294e-01 -2.57322192e-01 4.70233470e-01
-8.93224329e-02 -1.35166407e-01 3.29739124e-01 4.66960967e-01
1.44125116e+00 -4.49499577e-01 9.10166919e-01 -4.86315668e-01
1.02001071e+00 2.59366602e-01 2.99140424e-01 7.55349100e-01
-6.51940584e-01 5.61269045e-01 8.21522832e-01 -5.99950194e-01
-1.26073658e+00 -6.76820815e-01 1.51742145e-01 1.52258825e+00
-2.86460042e-01 -7.42139637e-01 -4.27630514e-01 -5.06532073e-01
-5.35785183e-02 1.28448069e+00 -7.04422593e-01 1.57731816e-01
-4.74157125e-01 -6.18218124e-01 3.36774796e-01 6.94772601e-02
5.11006713e-01 -1.43639922e+00 -8.83510828e-01 4.23459381e-01
-2.73871332e-01 -6.95965111e-01 -7.83300549e-02 3.30580920e-01
-6.56443775e-01 -1.27354932e+00 -1.02504611e-01 -6.84425056e-01
-3.32335293e-01 1.14474580e-01 1.58962810e+00 4.51503903e-01
-2.44464248e-01 2.85292193e-02 -6.41130149e-01 -4.09215212e-01
-5.76199114e-01 -2.39623323e-01 -2.58700639e-01 -6.95708513e-01
5.09257853e-01 -7.04435408e-01 -3.39009762e-01 1.77271083e-01
-1.47524653e-02 6.24117255e-02 9.04123336e-02 1.22051680e+00
-4.69554424e-01 4.43743505e-02 7.98255503e-01 -9.93297935e-01
1.20045376e+00 -8.22318792e-01 5.09347618e-02 -1.62725717e-01
-8.52111459e-01 -6.71195626e-01 6.47218883e-01 -2.44961935e-03
-6.47422373e-01 -3.48919868e-01 2.56651372e-01 4.29606698e-02
1.18259475e-01 4.76856202e-01 6.11554384e-01 3.04103971e-01
1.50276697e+00 -2.95255870e-01 -2.32624263e-01 -2.49039307e-01
-3.52623872e-02 1.05878794e+00 7.15716898e-01 -2.55390525e-01
5.40576518e-01 -7.10078999e-02 -1.20985262e-01 -6.57794774e-01
-1.13141489e+00 -8.92989457e-01 -6.59709334e-01 -7.76221275e-01
6.73198223e-01 -8.41142356e-01 -7.49721289e-01 6.30960643e-01
-1.36993933e+00 4.24784981e-03 3.20110381e-01 -6.84513301e-02
-6.71661556e-01 1.14842832e-01 -8.56113195e-01 -1.33889008e+00
-9.34103072e-01 -4.07526731e-01 1.03607982e-01 3.79016250e-01
-9.31490660e-01 -7.92686999e-01 5.31990588e-01 1.18771422e+00
2.16320172e-01 5.78701258e-01 9.41245437e-01 -9.52179074e-01
3.52575064e-01 -9.45444480e-02 6.05424978e-02 2.02805832e-01
-3.82439911e-01 -2.43771300e-01 -7.75848031e-01 -2.87067499e-02
2.16365889e-01 -8.65933657e-01 1.00958323e+00 -1.54081909e-02
5.16858160e-01 -3.02749038e-01 2.66252667e-01 -3.98371696e-01
1.29274559e+00 1.72668651e-01 8.05593491e-01 9.80828404e-01
2.64963508e-01 5.16512930e-01 6.78296506e-01 1.04295063e+00
2.39303634e-01 4.75001335e-01 4.27117795e-01 5.47949016e-01
-2.29165629e-01 -1.16207190e-01 7.71767676e-01 8.71128738e-01
-4.16274279e-01 -4.78979610e-02 -1.24234676e+00 4.56560344e-01
-2.19349790e+00 -1.53498137e+00 -1.06702077e+00 1.86177671e+00
8.58219385e-01 3.32841873e-01 8.37592542e-01 4.68569756e-01
7.78856635e-01 1.68197483e-01 1.65296987e-01 -9.74424422e-01
-3.77708167e-01 1.93883076e-01 -1.89437449e-01 1.01830375e+00
-8.39345634e-01 1.02176499e+00 6.28525972e+00 5.32240748e-01
-3.99407893e-01 6.25356376e-01 4.59977031e-01 -5.32468736e-01
1.74575195e-01 1.01669192e-01 -6.16045594e-01 5.16349256e-01
1.08896327e+00 -4.14831817e-01 3.93663377e-01 9.79304969e-01
5.07050335e-01 -5.07889926e-01 -6.97179496e-01 1.07322800e+00
8.17054451e-01 -1.05151784e+00 -2.89476097e-01 -1.81077227e-01
1.12096083e+00 -2.25334242e-01 -2.53219455e-01 7.69027293e-01
3.37492406e-01 -1.52935851e+00 7.09976494e-01 4.76332814e-01
-1.38739794e-01 -9.69680667e-01 1.07572603e+00 7.26023197e-01
-2.98596710e-01 -6.59488499e-01 -5.00027776e-01 -7.77664006e-01
2.28811949e-01 6.38168514e-01 -1.08813810e+00 -9.34586227e-02
5.15524805e-01 5.20081401e-01 -8.23569655e-01 1.16666281e+00
-3.29901993e-01 7.15021908e-01 3.44614565e-01 -2.62424827e-01
1.48606241e-01 2.22147226e-01 9.14411664e-01 1.55572641e+00
1.36592060e-01 2.49585539e-01 2.32367069e-01 4.83676851e-01
-9.47342962e-02 4.54473913e-01 -3.31102490e-01 4.26402748e-01
4.68712449e-01 1.64115262e+00 -2.58767635e-01 -3.94664377e-01
2.69482166e-01 7.92693973e-01 2.06545591e-01 -3.66704077e-01
-9.21663463e-01 -5.15699983e-01 -2.39624098e-01 3.92607838e-01
-3.09625745e-01 1.39334187e-01 -1.00101757e+00 -7.83624470e-01
-7.96896815e-01 -1.46207714e+00 7.40030468e-01 -9.53425467e-01
-1.36276579e+00 6.70044601e-01 -2.64620364e-01 -6.65960729e-01
4.58671190e-02 -8.25657785e-01 -1.27755308e+00 6.00472629e-01
-6.33013844e-01 -1.11661661e+00 -6.39324963e-01 4.33198988e-01
1.01246679e+00 -9.16862786e-01 3.53054643e-01 -6.13934919e-03
-3.08992267e-01 1.85447857e-01 -2.35416815e-01 -2.36426935e-01
1.00897050e+00 -1.53239393e+00 -5.03304720e-01 1.31096005e-01
-3.07193547e-01 2.57720411e-01 1.51800871e+00 -7.96610475e-01
-6.58745110e-01 -3.92457008e-01 1.47049749e+00 -1.00421667e+00
1.03148758e+00 2.81352013e-01 -8.69682550e-01 1.22744963e-01
9.73155856e-01 -1.06360197e+00 7.16991186e-01 3.50734711e-01
-4.26358640e-01 4.14033949e-01 -1.15560365e+00 2.31599912e-01
4.07909334e-01 -2.86853939e-01 -1.19299018e+00 7.06880808e-01
-2.13915676e-01 -3.90237689e-01 -8.04077685e-01 1.69477776e-01
6.05334461e-01 -1.71200037e+00 4.01100993e-01 -9.26253974e-01
1.53256071e+00 -1.66727453e-01 -3.68656993e-01 -9.48969901e-01
-8.52548242e-01 -6.42929196e-01 -4.43077028e-01 1.05865896e+00
4.82361495e-01 1.30417317e-01 7.77277648e-01 2.09075555e-01
-1.00502692e-01 -6.65701330e-01 -8.32096040e-01 -3.96761566e-01
5.13737798e-01 3.35066579e-02 -3.26559633e-01 1.22073972e+00
7.58095920e-01 1.29240370e+00 -1.07019186e+00 -6.60716772e-01
5.55695057e-01 1.39102370e-01 7.41201043e-01 -1.40477014e+00
-4.89943057e-01 -8.20482790e-01 -3.83943468e-01 -1.60300463e-01
2.01219156e-01 -8.90392482e-01 -1.48210272e-01 -1.46571863e+00
1.18039274e+00 7.53600597e-01 -5.12077212e-01 3.30834091e-01
-1.98041618e-01 2.60383338e-01 5.73642373e-01 3.74030143e-01
-7.00212240e-01 8.60548913e-02 8.97019446e-01 7.24319592e-02
-1.89156443e-01 -1.33742496e-01 -6.23323917e-01 8.39226305e-01
1.11917317e+00 -4.13702518e-01 -4.65824679e-02 6.53415322e-01
5.41208148e-01 3.20079535e-01 4.76988167e-01 -1.00963497e+00
4.16312844e-01 -3.47904414e-01 3.29780668e-01 -1.06932616e+00
2.00115398e-01 -1.25751957e-01 -1.29542962e-01 6.67108893e-01
-3.49464357e-01 1.89156279e-01 -2.96406716e-01 -4.81302384e-03
-4.47278619e-02 -9.27272320e-01 1.30385661e+00 -6.04422748e-01
-5.09785056e-01 -8.02949011e-01 -7.52391398e-01 1.24064438e-01
1.00394392e+00 -4.31320965e-02 -9.29028451e-01 -9.15060580e-01
-8.39077413e-01 1.49831682e-01 3.55573654e-01 4.65475827e-01
6.57424867e-01 -1.12181818e+00 -1.36786962e+00 -8.60632420e-01
-2.16027260e-01 -1.24907219e+00 3.78136039e-02 1.25419474e+00
-4.29992914e-01 1.62142843e-01 -6.96214437e-01 8.61150101e-02
-1.89171124e+00 2.79532999e-01 3.86310965e-01 -6.16655648e-01
-3.90641510e-01 7.05180645e-01 -6.00552917e-01 -2.07224950e-01
-5.66802584e-02 1.03669369e+00 -8.92765939e-01 4.93606001e-01
6.10935628e-01 1.49173522e+00 -2.04112843e-01 -8.66291642e-01
-1.25244632e-01 2.68562138e-01 -1.20108597e-01 -2.68159419e-01
1.24257743e+00 2.05181986e-01 -6.78820729e-01 9.93542731e-01
6.99135959e-01 1.88487694e-01 -7.61652663e-02 3.27612936e-01
4.93642747e-01 -4.32745427e-01 2.18705207e-01 -1.59018099e+00
-3.26320052e-01 6.45516038e-01 3.92789960e-01 3.30940008e-01
7.34219670e-01 -2.17990309e-01 6.53143585e-01 5.35859287e-01
3.04581165e-01 -1.79739678e+00 8.40183079e-01 1.29110420e+00
1.59774089e+00 -1.34589541e+00 5.15571833e-01 6.42105043e-02
-1.05923092e+00 1.34126782e+00 1.07404757e+00 -4.25533354e-01
1.80597693e-01 1.60973817e-02 5.34880944e-02 -6.41862750e-01
-1.39121342e+00 3.53345871e-01 1.64949104e-01 2.32480355e-02
1.19205213e+00 -2.96476278e-02 -1.18923688e+00 9.77863252e-01
-7.42950737e-01 -1.86820760e-01 1.18183744e+00 5.58232009e-01
-1.41475332e+00 -6.02148771e-01 -7.67434299e-01 4.52404320e-01
-6.85146093e-01 -8.84931684e-02 -1.50337374e+00 7.16130853e-01
2.91700959e-01 1.37347221e+00 -7.27909863e-01 -1.15170419e+00
3.19411457e-01 3.07425946e-01 2.54972517e-01 -5.28107226e-01
-1.57083189e+00 -1.29371405e-01 8.48366380e-01 -7.25210831e-02
-1.31655872e-01 -5.39723992e-01 -1.22110009e+00 -9.36357558e-01
-3.78331125e-01 5.70452452e-01 1.76049143e-01 6.48656249e-01
-1.70099407e-01 2.84653783e-01 8.38663936e-01 -5.01379788e-01
-7.87834764e-01 -1.48161101e+00 -6.34407520e-01 4.53743577e-01
-1.04376830e-01 -6.08290017e-01 -7.50037193e-01 6.76506758e-02] | [8.87362289428711, 11.067334175109863] |
308126f5-bdb4-47a4-ba18-f364ec7981bd | vanishing-point-guided-natural-image | 2004.02478 | null | https://arxiv.org/abs/2004.02478v1 | https://arxiv.org/pdf/2004.02478v1.pdf | Vanishing Point Guided Natural Image Stitching | Recently, works on improving the naturalness of stitching images gain more and more extensive attention. Previous methods suffer the failures of severe projective distortion and unnatural rotation, especially when the number of involved images is large or images cover a very wide field of view. In this paper, we propose a novel natural image stitching method, which takes into account the guidance of vanishing points to tackle the mentioned failures. Inspired by a vital observation that mutually orthogonal vanishing points in Manhattan world can provide really useful orientation clues, we design a scheme to effectively estimate prior of image similarity. Given such estimated prior as global similarity constraints, we feed it into a popular mesh deformation framework to achieve impressive natural stitching performances. Compared with other existing methods, including APAP, SPHP, AANAP, and GSP, our method achieves state-of-the-art performance in both quantitative and qualitative experiments on natural image stitching. | ['Yahui Liu', 'Kai Chen', 'Jian Yao', 'Yinxuan Li', 'Jingmin Tu', 'Li Li'] | 2020-04-06 | null | null | null | null | ['image-stitching'] | ['computer-vision'] | [ 4.37643319e-01 -2.17707023e-01 -2.40029860e-02 5.23922034e-02
-3.60223353e-01 -5.53245127e-01 6.19935930e-01 -3.80659401e-01
-1.11024685e-01 3.70291501e-01 2.35100776e-01 1.36479497e-01
3.66474525e-03 -5.40695012e-01 -5.46850502e-01 -7.28312194e-01
3.02868336e-01 2.56348312e-01 5.56589365e-01 -5.42965949e-01
7.23660350e-01 2.50829935e-01 -1.30503809e+00 -2.80122936e-01
1.10940158e+00 6.14151537e-01 9.43999812e-02 3.23908865e-01
5.88429347e-02 2.30293781e-01 -3.29352260e-01 -5.26190817e-01
5.00472665e-01 -3.40739965e-01 -6.28348053e-01 4.28316712e-01
6.60166025e-01 -3.95569980e-01 -3.03210944e-01 1.31830347e+00
5.39063931e-01 5.78225181e-02 4.75899100e-01 -1.32528651e+00
-5.53418458e-01 2.98299819e-01 -1.31794155e+00 -7.48024061e-02
6.09321535e-01 1.49195686e-01 5.52647352e-01 -7.57864475e-01
6.96112692e-01 1.21308768e+00 8.53648305e-01 1.84900418e-01
-1.07692015e+00 -8.56396973e-01 -1.45464316e-01 -1.51755303e-01
-1.31168497e+00 -3.78651470e-01 1.18045986e+00 -2.51735538e-01
1.61834195e-01 3.31216633e-01 5.44190407e-01 9.21297252e-01
2.54899830e-01 5.20803094e-01 1.02948987e+00 -5.30930519e-01
-1.48017585e-01 -1.50930181e-01 -1.94988608e-01 6.30491257e-01
2.83138603e-01 2.59148479e-01 -3.21570992e-01 -2.16652244e-01
1.39980996e+00 2.94209093e-01 -4.42598283e-01 -5.01442015e-01
-1.59631312e+00 5.24608076e-01 3.87769997e-01 3.99825841e-01
8.49293843e-02 1.55882329e-01 2.47504160e-01 1.61018372e-01
3.57972771e-01 5.43070972e-01 5.82716875e-02 7.55344331e-03
-8.65346968e-01 9.10043418e-02 2.67867804e-01 9.95141447e-01
7.37021506e-01 3.17038633e-02 1.75831169e-01 7.91581988e-01
-1.58244427e-02 4.92958635e-01 3.48728448e-01 -1.00590384e+00
4.97614622e-01 3.65179777e-01 1.28581122e-01 -1.68467712e+00
-8.76798294e-03 -2.46569261e-01 -1.19834495e+00 7.11479187e-02
2.39304110e-01 7.29589760e-02 -7.63468027e-01 1.33750927e+00
4.45162565e-01 5.16731679e-01 -6.33562580e-02 9.05794799e-01
5.34472704e-01 4.93602872e-01 -4.73114848e-01 -2.80644178e-01
1.39808095e+00 -1.03729701e+00 -8.23072553e-01 -2.16364413e-01
1.82973683e-01 -1.34445643e+00 9.75257933e-01 5.44650853e-01
-9.69527781e-01 -7.05509841e-01 -1.04870784e+00 5.89629970e-02
1.69541493e-01 -1.24248110e-01 6.23107016e-01 6.06438875e-01
-9.19514894e-01 6.21284485e-01 -4.83944625e-01 -3.05052310e-01
5.08028343e-02 2.86162674e-01 -3.93103749e-01 -2.35508665e-01
-9.96351361e-01 6.90112472e-01 4.04506296e-01 7.82412142e-02
-3.55758458e-01 -4.60489541e-01 -7.46086657e-01 -3.16007465e-01
5.76509476e-01 -6.20764792e-01 7.66062796e-01 -8.79068494e-01
-1.54172850e+00 7.72982955e-01 1.40390456e-01 -3.10212336e-02
6.85773671e-01 -3.43404025e-01 -3.08375239e-01 2.45422006e-01
-6.41939975e-03 7.24794269e-01 1.16385150e+00 -1.48631549e+00
-2.71442086e-01 -3.34337890e-01 -1.91963881e-01 3.67377043e-01
-4.14873868e-01 -4.68509085e-03 -5.32398522e-01 -1.09253359e+00
4.39801037e-01 -8.27773273e-01 -4.55815256e-01 2.37014517e-02
-4.14483964e-01 1.07114032e-01 1.36206734e+00 -5.68140030e-01
1.10843277e+00 -2.22961235e+00 3.16090584e-01 5.76684847e-02
2.62053907e-01 3.65501493e-01 1.46686351e-02 6.07089400e-01
-2.62876078e-02 -3.04345638e-02 -2.29462072e-01 -2.18330607e-01
-3.64083290e-01 2.24144146e-01 -5.20812213e-01 6.16471529e-01
-9.50770751e-02 6.83969617e-01 -1.01962340e+00 -7.52942026e-01
4.31946695e-01 4.12301183e-01 -4.81436223e-01 9.30519328e-02
7.34712332e-02 7.55716980e-01 -5.51112533e-01 4.96481001e-01
1.25904655e+00 2.99021378e-02 -1.68278277e-01 -4.11947042e-01
-2.45762333e-01 -4.60661560e-01 -1.27921963e+00 1.86915267e+00
-7.92006254e-02 4.85795021e-01 -1.49833217e-01 -7.25197971e-01
1.08208704e+00 2.89407402e-01 5.63100696e-01 -1.62817881e-01
2.77097255e-01 4.60005432e-01 -3.96714985e-01 -5.49885690e-01
6.90349400e-01 -9.69165415e-02 3.56559716e-02 1.83594286e-01
-3.16907614e-01 -7.67783225e-01 -1.51249006e-01 1.52917221e-01
6.98454738e-01 1.46208748e-01 2.89337009e-01 -2.93401062e-01
7.87817359e-01 -2.86373764e-01 6.07021213e-01 2.00803891e-01
-7.08786547e-02 1.15105581e+00 3.01704884e-01 -6.00131810e-01
-1.29410052e+00 -7.80297220e-01 -5.20145930e-02 3.73309076e-01
9.37331259e-01 -4.43434387e-01 -8.10430884e-01 -4.78946805e-01
-3.01138431e-01 1.63785458e-01 -4.48901892e-01 -1.09265178e-01
-7.61210382e-01 -3.47322255e-01 6.08615577e-01 2.07902402e-01
1.20849979e+00 -8.44950140e-01 -4.60934848e-01 -4.77981567e-02
-3.16638976e-01 -1.24659824e+00 -1.00987899e+00 -6.21938646e-01
-1.04543221e+00 -1.07980120e+00 -1.18135118e+00 -1.12036252e+00
8.13722014e-01 9.79859114e-01 7.62697816e-01 2.87070155e-01
-1.09381989e-01 5.04092462e-02 -3.80482793e-01 4.11486849e-02
-2.70591527e-01 -2.71890104e-01 9.04345289e-02 8.87981728e-02
-2.48535320e-01 -7.94263661e-01 -8.65626454e-01 9.17830706e-01
-1.32193840e+00 4.33476269e-01 5.72803557e-01 8.74761581e-01
2.13596806e-01 2.12768212e-01 1.70424446e-01 -5.80719233e-01
2.78360903e-01 7.83220381e-02 -4.84456867e-01 2.20217139e-01
-4.15072471e-01 8.78608301e-02 4.76691067e-01 -5.01542389e-01
-1.05139792e+00 3.76712941e-02 1.71074865e-03 -6.46657348e-01
8.16444457e-02 2.00393736e-01 8.69890675e-03 -5.98575294e-01
4.99842525e-01 4.30624872e-01 2.13075474e-01 -4.16394413e-01
2.79685348e-01 4.59977567e-01 8.08836460e-01 -5.34552634e-01
1.28812051e+00 6.95385754e-01 -1.35428114e-02 -9.67621386e-01
-4.44198370e-01 -5.42963684e-01 -6.58792198e-01 -2.75087446e-01
6.63910270e-01 -5.76127470e-01 -5.27303755e-01 9.12717700e-01
-1.27495396e+00 1.73891261e-01 7.24974796e-02 4.25797403e-01
-4.89465863e-01 1.28594685e+00 -4.09729689e-01 -5.11158407e-01
-4.24534619e-01 -1.37859595e+00 1.25162625e+00 2.36954585e-01
1.15085140e-01 -8.86209667e-01 2.56526202e-01 4.31714505e-01
3.68401229e-01 5.54543614e-01 4.05363619e-01 1.34906635e-01
-6.06674433e-01 -1.05165191e-01 -3.32594186e-01 1.16965175e-01
3.68464112e-01 2.64550269e-01 -5.65734923e-01 -4.21808571e-01
3.76918912e-01 -5.26743419e-02 6.35742426e-01 2.61776119e-01
8.58197331e-01 -2.82289088e-01 -4.51122046e-01 9.16469216e-01
1.45084262e+00 5.64022064e-02 1.03964603e+00 3.37570518e-01
8.57402027e-01 6.14277184e-01 8.46774936e-01 3.48340392e-01
-2.49887034e-02 9.13955867e-01 5.50942540e-01 -2.80077308e-01
-2.62796823e-02 -3.17813247e-01 2.29432598e-01 7.72437096e-01
-4.22005504e-01 -1.10129930e-01 -6.30230486e-01 2.79382080e-01
-1.87289369e+00 -9.31340754e-01 -2.38910630e-01 2.41799855e+00
7.71961808e-01 5.27272522e-02 -5.62363379e-02 2.86537975e-01
1.09063828e+00 5.55753708e-01 -1.42086044e-01 5.94008598e-04
-1.04615562e-01 -2.28942707e-01 6.93062007e-01 4.20237958e-01
-9.29421425e-01 1.02980304e+00 5.93498278e+00 1.26402140e+00
-1.23019850e+00 -1.22846596e-01 4.58432376e-01 6.16392136e-01
-4.65744287e-01 2.78256118e-01 -3.83965492e-01 5.35781920e-01
-5.18737640e-03 1.49195477e-01 2.77299136e-01 4.35376137e-01
3.41865681e-02 -1.27306178e-01 -4.77766931e-01 1.23955011e+00
1.94731578e-01 -1.26125324e+00 6.75484538e-02 1.05764091e-01
9.45100188e-01 -4.10148859e-01 3.20904225e-01 -3.66779566e-01
5.32011315e-02 -7.78368175e-01 4.87327605e-01 1.66568518e-01
9.20359790e-01 -7.31928587e-01 6.62368774e-01 2.78053552e-01
-1.26869071e+00 2.72239357e-01 -4.25753146e-01 7.25127235e-02
3.28152657e-01 5.73789597e-01 -3.49154294e-01 8.93801868e-01
3.45678717e-01 9.79353905e-01 -4.09632474e-01 1.07014871e+00
-2.07018033e-01 2.57730693e-01 -3.23051065e-01 4.29618359e-01
3.47878873e-01 -5.14594972e-01 7.63426781e-01 7.41901755e-01
5.04407525e-01 3.73798490e-01 1.15992479e-01 5.22774279e-01
2.68852144e-01 8.48764926e-02 -8.99927378e-01 2.33771339e-01
2.74634063e-01 1.21795571e+00 -1.01942122e+00 -2.98385233e-01
-1.80039391e-01 1.15411294e+00 -2.96351433e-01 2.20292687e-01
-8.87226999e-01 -3.50620389e-01 3.02407682e-01 1.84957027e-01
1.08770572e-01 -4.81508434e-01 -2.44294614e-01 -1.47489202e+00
-4.26220782e-02 -1.02091527e+00 9.63875800e-02 -8.81036937e-01
-8.83215427e-01 5.57505488e-01 5.65575697e-02 -1.87038636e+00
3.53354141e-02 -2.56475627e-01 -9.15017724e-01 4.04926747e-01
-1.21625423e+00 -1.30640328e+00 -6.08646214e-01 6.46629512e-01
7.00274110e-01 7.92606398e-02 4.03818160e-01 1.82095602e-01
-4.28163022e-01 4.63651806e-01 7.80283511e-02 5.49893044e-02
1.14463806e+00 -8.02115440e-01 6.21127188e-01 1.01304090e+00
-1.59708858e-02 5.90384543e-01 9.67915118e-01 -8.35928500e-01
-1.44076824e+00 -6.33596301e-01 4.07968551e-01 -1.29604906e-01
4.06719983e-01 6.15266562e-02 -8.48603010e-01 4.30634201e-01
5.98457634e-01 -7.96386749e-02 2.23234762e-02 -4.69200104e-01
-3.60701978e-01 -5.00895344e-02 -1.27000225e+00 8.04137945e-01
1.23481774e+00 -7.37653896e-02 -6.56989753e-01 1.62001073e-01
8.22153747e-01 -6.11676753e-01 -6.31510139e-01 3.99108082e-01
6.05704546e-01 -1.20722854e+00 1.19180357e+00 1.33645564e-01
6.21652007e-01 -4.80186611e-01 2.38111958e-01 -1.20396376e+00
-2.00159118e-01 -1.30658591e+00 4.42635149e-01 1.17311847e+00
-3.64611804e-01 -8.14974248e-01 7.34594882e-01 1.24983393e-01
-1.69039309e-01 -4.26604897e-01 -6.77659273e-01 -6.71450257e-01
-3.17300260e-01 1.93158641e-01 6.59393370e-01 1.24218380e+00
8.75781849e-02 1.20521493e-01 -9.21112061e-01 1.00925989e-01
9.78720307e-01 1.38065472e-01 1.05383313e+00 -1.09040594e+00
-2.14014396e-01 -2.64639795e-01 -6.76645041e-01 -1.22481215e+00
-3.93524498e-01 -1.93920135e-01 -3.24686132e-02 -9.46183622e-01
1.29467994e-01 -4.63250667e-01 1.32633567e-01 6.07904680e-02
-3.18744212e-01 6.58294499e-01 1.24345109e-01 6.25172675e-01
-2.62822270e-01 6.11347377e-01 1.91216910e+00 9.69922692e-02
-1.33006319e-01 -8.61335546e-02 -4.50400531e-01 8.85581315e-01
5.68874180e-01 -1.32698968e-01 -4.57528681e-01 -6.13329470e-01
1.47168472e-01 4.09073353e-01 2.85408705e-01 -1.02044070e+00
2.84479439e-01 -4.73359615e-01 9.77707729e-02 -7.25306273e-01
3.37775648e-01 -8.97617221e-01 5.24852097e-01 6.71306431e-01
-6.93702400e-02 2.42861241e-01 -1.52124450e-01 6.41862392e-01
-4.63166416e-01 -2.19362766e-01 8.15651715e-01 -7.14482591e-02
-5.63360095e-01 4.27536160e-01 2.37635463e-01 -7.13296384e-02
9.96559858e-01 -6.58713639e-01 -4.28626776e-01 -4.49053079e-01
-1.96999200e-02 2.33183578e-02 9.91483271e-01 3.49920660e-01
8.52349997e-01 -1.45413768e+00 -5.50512433e-01 5.03103018e-01
-2.24047564e-02 1.69457346e-01 2.87840962e-01 1.06946039e+00
-9.52464879e-01 6.29071938e-03 -5.49621046e-01 -6.81601286e-01
-1.34566188e+00 7.35436440e-01 9.44224000e-03 -1.00642614e-01
-7.85582542e-01 6.38107836e-01 4.92958456e-01 -1.61390945e-01
-7.17303455e-02 -2.92029202e-01 -7.81409293e-02 -2.67304212e-01
3.55324447e-01 4.13234979e-01 -2.90869355e-01 -7.57900655e-01
-1.12958394e-01 1.42435789e+00 -1.35131180e-01 -2.37150937e-01
1.19490647e+00 -2.55292535e-01 -1.91005081e-01 -1.92977309e-01
1.20985270e+00 4.67844188e-01 -1.42186010e+00 -2.92478830e-01
-3.89422089e-01 -1.11028957e+00 -8.23841318e-02 -3.26017477e-02
-1.25713706e+00 7.91803777e-01 3.64507824e-01 1.88258439e-01
1.14375007e+00 -3.16464514e-01 1.31740677e+00 -1.25671282e-01
6.23533368e-01 -6.97624385e-01 5.17819643e-01 2.41717890e-01
9.84061420e-01 -1.23000300e+00 2.59839058e-01 -8.25549066e-01
-6.08928502e-01 1.29707408e+00 4.96498138e-01 -6.01536036e-01
2.68361151e-01 1.01840213e-01 1.43665122e-02 -1.76942959e-01
-4.91565987e-02 2.32027411e-01 3.86112899e-01 5.15676856e-01
7.75715858e-02 -3.44124377e-01 -4.80542690e-01 -2.49785170e-01
-5.62235080e-02 -5.51629514e-02 5.53349316e-01 8.73603880e-01
-6.54848576e-01 -1.27326691e+00 -8.43594015e-01 -8.07763934e-02
-3.30581546e-01 1.00333784e-02 -2.88161695e-01 7.43284225e-01
-7.53776133e-02 7.04121411e-01 -2.09977865e-01 -6.45968914e-01
1.57049879e-01 -6.27065003e-01 6.28703237e-01 -2.02662244e-01
-3.77781034e-01 4.02823925e-01 -2.87282109e-01 -4.26246405e-01
-6.57265246e-01 -4.98251677e-01 -9.21464145e-01 -4.50634271e-01
-6.27134383e-01 7.52039477e-02 2.87560552e-01 7.76089013e-01
8.79819393e-02 9.03485995e-03 9.41174328e-01 -1.18049634e+00
-4.62213695e-01 -7.88154542e-01 -5.14530122e-01 5.21949768e-01
3.28432500e-01 -7.80935824e-01 -4.78359222e-01 1.61325961e-01] | [9.355016708374023, -2.352206230163574] |
0fbc7dca-3f77-4eed-86cf-0b8f4a09fefc | an-end-to-end-ocr-framework-for-robust-arabic | 2208.11484 | null | https://arxiv.org/abs/2208.11484v2 | https://arxiv.org/pdf/2208.11484v2.pdf | An End-to-End OCR Framework for Robust Arabic-Handwriting Recognition using a Novel Transformers-based Model and an Innovative 270 Million-Words Multi-Font Corpus of Classical Arabic with Diacritics | This research is the second phase in a series of investigations on developing an Optical Character Recognition (OCR) of Arabic historical documents and examining how different modeling procedures interact with the problem. The first research studied the effect of Transformers on our custom-built Arabic dataset. One of the downsides of the first research was the size of the training data, a mere 15000 images from our 30 million images, due to lack of resources. Also, we add an image enhancement layer, time and space optimization, and Post-Correction layer to aid the model in predicting the correct word for the correct context. Notably, we propose an end-to-end text recognition approach using Vision Transformers as an encoder, namely BEIT, and vanilla Transformer as a decoder, eliminating CNNs for feature extraction and reducing the model's complexity. The experiments show that our end-to-end model outperforms Convolutions Backbones. The model attained a CER of 4.46%. | ['Amr S. Ghoneim', 'Anas Salah', 'Salma Jamal', 'Ahmed Elbehery', 'Ali Ashraf', 'Omar Mohamed', 'Aly Mostafa'] | 2022-08-20 | null | null | null | null | ['handwriting-recognition'] | ['computer-vision'] | [ 1.88797504e-01 -2.87976056e-01 4.75690395e-01 -3.37325662e-01
-3.93944919e-01 -6.02975368e-01 7.73699641e-01 -2.15799630e-01
-8.86325240e-01 1.90719873e-01 4.10525426e-02 -4.04514641e-01
4.41025943e-01 -7.75188982e-01 -7.50697851e-01 -4.85519260e-01
2.38656029e-01 3.30082148e-01 8.97281542e-02 -2.45150909e-01
8.49505842e-01 5.12573123e-01 -1.20849597e+00 5.35034180e-01
9.24109459e-01 1.03576267e+00 5.52760661e-01 1.06651974e+00
-8.91576558e-02 8.74235630e-01 -8.57979774e-01 -9.70060229e-01
-1.29686575e-02 -1.94975436e-01 -7.08225012e-01 3.19469541e-01
6.52099848e-01 -7.14512885e-01 -5.68570435e-01 9.55937445e-01
6.50712252e-01 -1.30089238e-01 4.84941542e-01 -7.07668543e-01
-1.44911015e+00 5.22763789e-01 -4.21498805e-01 1.07306316e-01
2.12681800e-01 1.83628142e-01 7.61025548e-01 -1.36382008e+00
6.04517043e-01 1.00418997e+00 7.29636490e-01 5.01707494e-01
-5.67124128e-01 -2.53485113e-01 5.88425659e-02 5.41541755e-01
-1.13861251e+00 -7.73193359e-01 3.95374954e-01 -1.72059864e-01
1.27029312e+00 7.07530677e-02 5.58624923e-01 1.08578730e+00
1.01399101e-01 9.83636796e-01 1.00897133e+00 -9.21961606e-01
2.67332215e-02 5.61637133e-02 4.33484674e-01 1.00750256e+00
1.03298910e-01 -2.87671030e-01 -4.07731384e-01 4.77443337e-01
6.53135598e-01 -3.77630293e-01 -7.98191130e-02 4.41729814e-01
-8.14814031e-01 5.38353086e-01 2.63626367e-01 3.51286352e-01
-2.03363314e-01 -4.83339243e-02 2.45321184e-01 2.93969750e-01
1.85684904e-01 2.90365636e-01 -3.07419598e-01 -3.26447427e-01
-1.25739527e+00 -8.75152200e-02 7.71626472e-01 1.02415621e+00
3.59150261e-01 2.15589479e-01 -1.27782360e-01 9.08593416e-01
3.71071219e-01 4.81837541e-01 6.29511118e-01 -4.48282719e-01
6.32285893e-01 5.92142344e-01 -1.42382428e-01 -8.39167774e-01
-2.09946796e-01 -3.81658316e-01 -4.20985252e-01 4.65577990e-02
7.04879522e-01 -2.03421041e-01 -1.37838244e+00 9.20612335e-01
-3.25238585e-01 -2.12365210e-01 -1.28432095e-01 1.01827359e+00
4.63070601e-01 8.43915880e-01 -1.72260076e-01 2.77465463e-01
1.67596877e+00 -1.49786782e+00 -7.07674325e-01 -5.05117476e-01
5.72808266e-01 -1.25141239e+00 1.25714958e+00 6.28730059e-01
-1.14173901e+00 -4.15793508e-01 -1.32938457e+00 -6.16333544e-01
-5.74869275e-01 7.83511579e-01 5.35829365e-01 8.74604225e-01
-1.06423295e+00 5.44329405e-01 -6.20587111e-01 -5.05196512e-01
4.56372768e-01 1.74498428e-02 -3.33778650e-01 -2.89670080e-01
-8.41343343e-01 1.27644205e+00 3.43699604e-01 4.60961163e-01
-8.38478744e-01 -2.70224810e-01 -5.69797814e-01 1.42947465e-01
1.68553948e-01 -1.08355721e-02 1.08555746e+00 -1.24267232e+00
-1.71854246e+00 7.08114684e-01 -6.00184985e-02 -4.64194387e-01
4.47971940e-01 -5.34364343e-01 -5.92202067e-01 2.42893666e-01
-2.67970353e-01 4.55164254e-01 9.28309500e-01 -1.01422930e+00
-6.06574118e-01 -3.73391479e-01 -7.36892596e-02 1.54861346e-01
-5.17049551e-01 2.78754711e-01 -1.11628151e+00 -7.89476514e-01
-2.16042735e-02 -8.92290294e-01 5.87207861e-02 -1.66462004e-01
-2.22440690e-01 1.89263642e-01 9.41543579e-01 -1.34277189e+00
1.18637371e+00 -2.00634551e+00 -1.11434683e-01 8.52528587e-02
-1.97657391e-01 6.07763052e-01 -5.79674065e-01 4.00976181e-01
1.48128346e-01 6.25236779e-02 -2.43073568e-01 -4.85746086e-01
-7.62736127e-02 -9.02879685e-02 -2.13595271e-01 3.79759490e-01
6.72114730e-01 8.86471033e-01 -4.64801759e-01 -3.75268251e-01
8.84591788e-02 5.39046288e-01 -4.69503671e-01 -7.63319200e-03
-3.13576907e-01 -2.82485306e-01 4.43526693e-02 9.55426991e-01
8.04803014e-01 7.76401116e-03 6.97086826e-02 -2.12029293e-01
-3.23564142e-01 3.69785875e-01 -9.56600845e-01 1.73647368e+00
-2.28439510e-01 1.16914690e+00 -1.83263700e-02 -7.28690684e-01
7.83595800e-01 2.20138073e-01 -2.57159006e-02 -1.13909733e+00
4.92875874e-01 2.65474826e-01 -4.51156795e-02 -4.91555214e-01
9.92715359e-01 4.94455665e-01 2.64666945e-01 4.73931104e-01
2.44635954e-01 1.82850018e-01 5.13708591e-01 1.49798155e-01
7.86682367e-01 3.20763499e-01 -3.13448966e-01 -3.59442607e-02
5.26965320e-01 2.22877994e-01 4.82892394e-02 6.80245519e-01
-1.17312975e-01 8.37170541e-01 3.56519014e-01 -4.42180157e-01
-1.27789557e+00 -6.83662117e-01 -6.15432933e-02 1.04464817e+00
-2.33669102e-01 -3.54239702e-01 -1.17110145e+00 -4.59961414e-01
-4.57012147e-01 8.07387829e-01 -5.68589211e-01 1.13389321e-01
-7.35566497e-01 -1.08052146e+00 1.17910957e+00 4.78400648e-01
8.88911724e-01 -9.98663723e-01 -6.54106617e-01 2.30177090e-01
-1.98566496e-01 -1.24469256e+00 -5.75361669e-01 1.71951473e-01
-8.49402487e-01 -9.38914776e-01 -7.95261145e-01 -1.02134347e+00
7.79034436e-01 1.76025465e-01 9.21774685e-01 1.78826049e-01
-4.02840823e-01 1.86000243e-02 -5.95052719e-01 -5.06215036e-01
-2.85833150e-01 1.30309060e-01 -3.91725391e-01 -8.38346034e-02
6.50562704e-01 5.37563711e-02 -3.76477957e-01 -5.49028777e-02
-8.81588221e-01 1.34527370e-01 7.29227543e-01 8.73953581e-01
1.38863713e-01 -3.04446697e-01 -2.64551584e-02 -7.01499760e-01
8.59495521e-01 -2.61457451e-02 -6.44035876e-01 5.82636654e-01
-7.82463968e-01 7.95646384e-02 4.25884604e-01 -4.06047016e-01
-1.04494631e+00 1.60010278e-01 2.31325608e-02 2.30712935e-01
4.48177680e-02 6.67252779e-01 1.01223454e-01 9.86050349e-04
5.24320662e-01 6.43774390e-01 6.37487322e-02 -5.65866768e-01
3.15001100e-01 1.10391188e+00 6.16336882e-01 -3.32824618e-01
5.04237235e-01 2.32081667e-01 -4.76105750e-01 -1.11091828e+00
-6.71025276e-01 -3.83391790e-02 -8.60045314e-01 -2.09354192e-01
7.16893196e-01 -1.10388446e+00 -5.54732442e-01 1.11366498e+00
-1.35945570e+00 -1.96112379e-01 2.35493571e-01 3.56367916e-01
-1.20618790e-01 5.80289721e-01 -1.04719162e+00 -6.57467723e-01
-7.69516945e-01 -1.18405807e+00 7.45544016e-01 5.25551856e-01
2.38869011e-01 -5.34622252e-01 -2.80969679e-01 5.21507561e-01
5.07973254e-01 -3.47357929e-01 8.18661630e-01 -4.09245461e-01
-5.98014534e-01 -4.18366581e-01 -6.08651757e-01 5.72897553e-01
-1.81418985e-01 4.68781352e-01 -1.19532883e+00 -3.24329674e-01
-1.58531591e-01 -2.31316671e-01 9.04114246e-01 1.62047029e-01
1.04499292e+00 -2.16228470e-01 2.47598261e-01 4.72611427e-01
1.38698733e+00 4.68175441e-01 1.41429877e+00 6.67610228e-01
7.25482881e-01 3.84079129e-01 4.05231744e-01 2.09163845e-01
4.87193227e-01 4.90870833e-01 3.12318623e-01 -9.69196633e-02
-3.75615478e-01 3.67269106e-02 8.40772510e-01 1.02810609e+00
-2.56410211e-01 -3.36975753e-01 -9.36556697e-01 5.67933023e-01
-1.42060745e+00 -8.66345346e-01 -2.67043114e-01 1.85932791e+00
9.09331799e-01 -3.59391123e-02 -1.09037712e-01 1.56024754e-01
5.42536438e-01 -1.26097932e-01 -1.51837751e-01 -7.89149284e-01
-3.49319130e-01 3.04343671e-01 6.36728466e-01 5.77239275e-01
-1.18127954e+00 1.31860137e+00 6.73794842e+00 7.00509965e-01
-1.35685658e+00 -3.00825745e-01 5.09700596e-01 -7.33268308e-03
1.92971289e-01 -3.19026061e-04 -7.87742257e-01 4.85204846e-01
1.12419093e+00 4.64511693e-01 6.01630270e-01 7.06701815e-01
2.76536550e-02 -2.01499239e-01 -8.41518641e-01 6.90189123e-01
5.61388671e-01 -1.38472998e+00 3.22662205e-01 -1.08159166e-02
3.06826890e-01 1.32252887e-01 -5.42069599e-03 1.98339541e-02
-4.13250690e-03 -1.24086356e+00 1.26820624e+00 5.25478184e-01
8.45714688e-01 -7.65150964e-01 7.10616648e-01 -1.47167454e-02
-7.49198377e-01 -8.42859149e-02 -4.24650520e-01 -2.17176273e-01
-2.02608898e-01 3.33014697e-01 -9.59718525e-01 1.47333220e-01
6.26135826e-01 4.52042013e-01 -1.21063030e+00 9.84153152e-01
-2.61595130e-01 7.02361286e-01 -1.16818354e-01 -4.02874947e-01
5.48882842e-01 -2.78813452e-01 1.20290257e-01 1.70795083e+00
2.58370817e-01 -1.25664756e-01 -4.78837550e-01 5.43965161e-01
-1.26390666e-01 -1.08230881e-01 -3.37717645e-02 -4.23906952e-01
4.84827220e-01 1.20868504e+00 -7.42822587e-01 -1.91758662e-01
-4.95225579e-01 1.49980950e+00 3.23382556e-01 1.59392729e-01
-8.81560087e-01 -8.72863770e-01 1.87560111e-01 -1.19031839e-01
4.99354988e-01 -4.69168365e-01 -7.11324692e-01 -1.20848095e+00
2.73443103e-01 -1.27252460e+00 -8.22887421e-02 -8.93501341e-01
-9.22018528e-01 5.15811324e-01 -7.22010076e-01 -8.20517659e-01
1.76241770e-01 -1.29634762e+00 -3.51496816e-01 9.42592144e-01
-1.56504440e+00 -1.42626691e+00 -2.97460973e-01 5.22109687e-01
6.93085968e-01 -4.64170098e-01 9.17270541e-01 6.57994568e-01
-7.94959307e-01 8.67601216e-01 2.28065819e-01 7.49291301e-01
8.07171941e-01 -1.19071448e+00 6.48161769e-01 1.48854232e+00
2.00309083e-01 7.33863533e-01 3.59065443e-01 -3.94877970e-01
-1.66541815e+00 -7.07856715e-01 1.35393322e+00 -4.77337420e-01
6.29271150e-01 -4.57260370e-01 -7.60376513e-01 5.14613688e-01
5.40706754e-01 -4.31442499e-01 4.01196748e-01 -1.79331467e-01
-5.79105854e-01 1.54905394e-01 -8.58579040e-01 9.41227674e-01
6.18853867e-01 -6.78134084e-01 -5.27690113e-01 9.03395638e-02
3.58532637e-01 -3.74971837e-01 -6.43045902e-01 -2.79189259e-01
9.29801822e-01 -7.37917066e-01 7.56998718e-01 -3.77485186e-01
1.07928181e+00 -3.07660043e-01 -1.81767404e-01 -1.00187373e+00
-3.48236471e-01 -3.45878065e-01 -1.49185181e-01 1.27190340e+00
6.87299609e-01 -8.03220421e-02 6.19028509e-01 5.17622709e-01
-2.95713991e-01 -3.76966596e-01 -6.20066583e-01 -3.28051209e-01
1.73481613e-01 -4.55016077e-01 4.26973015e-01 8.47348809e-01
-2.69388407e-01 3.05878937e-01 -4.41551328e-01 1.45725802e-01
2.36868426e-01 -1.63592845e-01 4.65161979e-01 -7.83202171e-01
-1.50016949e-01 -4.10813242e-01 -1.89153716e-01 -1.08890092e+00
-4.78994399e-01 -6.09204710e-01 1.09684326e-01 -1.44830573e+00
6.32275566e-02 -9.85678360e-02 7.84817785e-02 5.46457410e-01
-1.11938007e-01 2.98590750e-01 6.14304721e-01 9.92779657e-02
-1.18393436e-01 2.56987303e-01 1.21755254e+00 -3.12736541e-01
1.13369830e-01 -4.52282101e-01 -6.66207492e-01 5.86444318e-01
7.45086133e-01 -2.55439401e-01 -5.64812422e-02 -1.26919866e+00
8.56244788e-02 -4.42833781e-01 3.26593727e-01 -8.95709455e-01
4.81886417e-01 2.18872458e-01 8.32803428e-01 -7.34353483e-01
1.91001907e-01 -6.85757697e-01 -5.31396985e-01 4.29898143e-01
-1.43160880e-01 4.01646703e-01 3.37977171e-01 3.10703125e-02
-1.81784689e-01 -6.39440954e-01 7.97171235e-01 -5.87565303e-02
-9.83859479e-01 -2.28740320e-01 -5.67885816e-01 -2.46842369e-01
6.76430464e-01 -6.31105185e-01 -7.63596058e-01 -1.14973644e-02
-3.88975531e-01 -1.27562925e-01 4.04131025e-01 4.66591775e-01
7.26423085e-01 -8.56326997e-01 -8.71998191e-01 1.23255245e-01
-1.33210152e-01 -2.57045150e-01 -9.82983932e-02 5.52801669e-01
-1.33007276e+00 4.97051150e-01 -5.61900675e-01 -2.03931361e-01
-1.21689117e+00 1.14085309e-01 3.07527602e-01 -2.09618196e-01
-5.83321154e-01 8.97642970e-01 -7.06134737e-01 -2.74721652e-01
4.01481748e-01 -1.61134917e-02 -3.53931904e-01 3.13275069e-01
7.81798124e-01 4.19054329e-01 4.85974133e-01 -5.47441065e-01
-3.21326673e-01 4.80572581e-01 -5.08286059e-01 -4.34278280e-01
1.52875602e+00 -6.01591505e-02 -2.81485289e-01 -1.07354019e-02
8.20371449e-01 -3.34039740e-02 -1.26513040e+00 -1.36444226e-01
2.01159239e-01 -3.01667631e-01 1.99029535e-01 -1.26587582e+00
-1.06427407e+00 1.02706683e+00 7.68787146e-01 -1.94390222e-01
1.25816059e+00 -7.57323265e-01 7.26460636e-01 7.59636641e-01
-3.62154730e-02 -1.71211410e+00 1.05227426e-01 1.23276842e+00
8.92745733e-01 -1.13456202e+00 1.20561376e-01 -1.20774657e-01
-7.99264610e-01 1.61053514e+00 5.97739041e-01 -2.41335884e-01
3.54616731e-01 4.63554263e-01 2.06238121e-01 -1.06679551e-01
-5.55708528e-01 -2.16583647e-02 2.19478190e-01 7.44820952e-01
7.86552548e-01 -1.29608259e-01 -3.60795796e-01 6.44839108e-01
-3.02834809e-01 7.52039179e-02 9.44912195e-01 1.15446162e+00
-3.02501172e-01 -9.61804569e-01 -2.76250929e-01 2.05334887e-01
-6.09144151e-01 -6.27824068e-01 -6.23030365e-01 5.97978413e-01
-2.48292498e-02 1.07204735e+00 2.35749021e-01 -4.95199621e-01
3.30747485e-01 2.56021172e-01 5.69507539e-01 -3.12782764e-01
-8.35847497e-01 -7.51687214e-02 9.01469290e-02 -2.08393484e-01
-1.32423580e-01 -5.44747472e-01 -1.09637582e+00 -4.78817582e-01
-5.02596974e-01 -3.40770155e-01 1.05805409e+00 9.91034091e-01
4.13785338e-01 4.96313542e-01 4.40422267e-01 -6.19774640e-01
-5.00125408e-01 -1.23153675e+00 -2.91892767e-01 3.51198390e-02
1.31359845e-01 9.17806998e-02 1.38083547e-01 6.99126422e-01] | [11.8574857711792, 2.5076544284820557] |
18042053-1e00-4294-be5f-e148b163dcb5 | the-cross-evaluation-of-machine-learning | 2203.04686 | null | https://arxiv.org/abs/2203.04686v1 | https://arxiv.org/pdf/2203.04686v1.pdf | The Cross-evaluation of Machine Learning-based Network Intrusion Detection Systems | Enhancing Network Intrusion Detection Systems (NIDS) with supervised Machine Learning (ML) is tough. ML-NIDS must be trained and evaluated, operations requiring data where benign and malicious samples are clearly labelled. Such labels demand costly expert knowledge, resulting in a lack of real deployments, as well as on papers always relying on the same outdated data. The situation improved recently, as some efforts disclosed their labelled datasets. However, most past works used such datasets just as a 'yet another' testbed, overlooking the added potential provided by such availability. In contrast, we promote using such existing labelled data to cross-evaluate ML-NIDS. Such approach received only limited attention and, due to its complexity, requires a dedicated treatment. We hence propose the first cross-evaluation model. Our model highlights the broader range of realistic use-cases that can be assessed via cross-evaluations, allowing the discovery of still unknown qualities of state-of-the-art ML-NIDS. For instance, their detection surface can be extended--at no additional labelling cost. However, conducting such cross-evaluations is challenging. Hence, we propose the first framework, XeNIDS, for reliable cross-evaluations based on Network Flows. By using XeNIDS on six well-known datasets, we demonstrate the concealed potential, but also the risks, of cross-evaluations of ML-NIDS. | ['Mauro Conti', 'Luca Pajola', 'Giovanni Apruzzese'] | 2022-03-09 | null | null | null | null | ['network-intrusion-detection'] | ['miscellaneous'] | [ 1.06234536e-01 -7.17077684e-03 -3.44391376e-01 -5.95818043e-01
-3.01361442e-01 -7.72727907e-01 9.81358111e-01 2.81167403e-02
-5.23308635e-01 7.86048591e-01 -6.35496914e-01 -7.29909241e-01
-3.44850481e-01 -5.53489089e-01 -4.59162354e-01 -3.36153299e-01
-6.44617736e-01 7.07924068e-01 4.39665556e-01 -6.77916184e-02
1.09528884e-01 8.68290126e-01 -1.67547202e+00 3.79189909e-01
3.69987249e-01 9.91174817e-01 -6.60260618e-01 4.26886767e-01
-1.44195780e-01 6.87894702e-01 -1.17614269e+00 -8.94575298e-01
5.90344131e-01 -8.91148448e-02 -7.75739431e-01 -1.54981390e-01
3.96314859e-01 -3.68136138e-01 3.48677486e-03 8.62150848e-01
3.00027281e-01 -4.76581663e-01 3.02062124e-01 -2.20869946e+00
-3.03880095e-01 5.46266139e-01 -2.37952679e-01 1.43014356e-01
1.66867599e-01 6.13620639e-01 7.65550971e-01 2.60498766e-02
7.74523795e-01 1.08348739e+00 7.54416406e-01 8.64934385e-01
-1.28810644e+00 -7.73974895e-01 -3.14698927e-02 1.60207823e-01
-1.05847037e+00 -4.17645514e-01 5.81400037e-01 -3.39305431e-01
8.96345317e-01 4.59322780e-01 2.17641652e-01 1.74008584e+00
-4.25053149e-01 5.90943396e-01 1.50372291e+00 -4.62610543e-01
4.52115208e-01 8.11194003e-01 1.61880046e-01 4.32833642e-01
5.22175252e-01 6.12731338e-01 9.17498693e-02 -2.85039574e-01
2.23175153e-01 7.14354888e-02 -1.14030023e-04 -5.74729860e-01
-8.62644911e-01 7.28669584e-01 8.21767896e-02 5.45744538e-01
-1.14711598e-01 -2.77695745e-01 8.94244254e-01 5.81402957e-01
3.17566574e-01 5.90078950e-01 -8.02165270e-01 -3.93604070e-01
-1.04890072e+00 -1.14028431e-01 1.34213436e+00 5.67600965e-01
7.69114912e-01 2.78811324e-02 2.99116045e-01 4.49430972e-01
2.70923704e-01 2.18589485e-01 -6.07547350e-03 -5.90679824e-01
1.94372147e-01 7.35750794e-01 -1.32404000e-01 -8.09999406e-01
-4.63447899e-01 -5.07144928e-01 -6.40684843e-01 8.35402310e-01
5.88035107e-01 -2.69392192e-01 -5.85717857e-01 1.88974738e+00
2.89811105e-01 4.97287005e-01 1.44623816e-02 5.62726676e-01
1.90098137e-01 -1.18995264e-01 2.05893263e-01 -2.42951527e-01
1.12015915e+00 -7.25987375e-01 -3.11964065e-01 1.30297512e-01
7.90917397e-01 -4.59387273e-01 9.45570290e-01 6.36721671e-01
-5.35503387e-01 -1.44966796e-01 -1.09345126e+00 8.71963322e-01
-9.63303149e-01 -5.00009239e-01 9.35342968e-01 1.58292890e+00
-1.10907865e+00 8.53488028e-01 -6.00245178e-01 -5.17462671e-01
5.94811261e-01 4.53871340e-01 -4.67057467e-01 1.18428022e-01
-1.26539838e+00 9.26418006e-01 1.11032329e-01 -1.50104582e-01
-1.06446266e+00 -8.14133465e-01 -4.49489594e-01 -2.08303869e-01
5.97759783e-01 -1.72445759e-01 1.00325847e+00 -8.33179235e-01
-1.34686148e+00 9.80068326e-01 2.97864884e-01 -5.96432567e-01
9.51522171e-01 1.08577572e-01 -1.00868082e+00 1.11698508e-01
-1.04258642e-01 3.21428746e-01 8.23742270e-01 -1.64503968e+00
-5.04998088e-01 -2.08480492e-01 6.09651566e-01 -6.83096290e-01
-2.79994607e-01 3.85620594e-01 -4.97556254e-02 -1.01303034e-01
-6.66355133e-01 -6.80447817e-01 -2.59668887e-01 -2.63408482e-01
-6.33501768e-01 9.96862799e-02 1.37717390e+00 -5.70842996e-02
1.06574786e+00 -2.19732046e+00 -7.65753210e-01 3.91070157e-01
3.86097997e-01 9.07839298e-01 -1.67555183e-01 3.93034577e-01
-3.20696741e-01 7.77017593e-01 -2.69873500e-01 -4.41507012e-01
3.88823718e-01 2.07788184e-01 -2.64124215e-01 3.88138324e-01
6.11852527e-01 5.79743207e-01 -8.36660683e-01 -3.32500517e-01
5.56793511e-01 5.82060039e-01 -3.20789337e-01 1.17797770e-01
-1.66881442e-01 4.01267081e-01 -8.72145146e-02 8.76083255e-01
7.37150490e-01 -2.02610940e-01 1.46414503e-01 -6.77778125e-02
1.73676517e-02 1.02897972e-01 -1.26440358e+00 1.00393152e+00
-4.07198966e-01 4.03753251e-01 1.48204207e-01 -1.03369057e+00
8.21304381e-01 1.87232777e-01 5.94033599e-01 -8.65572751e-01
2.46158484e-02 3.25788170e-01 1.74631879e-01 -3.58600438e-01
-1.01927929e-01 -7.79635161e-02 9.03831720e-02 8.85568738e-01
7.24312291e-02 4.40224022e-01 3.68908048e-01 1.76117197e-01
1.46973181e+00 -4.03181054e-02 -3.03377328e-03 2.75561288e-02
5.91880202e-01 -1.55139551e-01 3.23184162e-01 9.38955009e-01
-7.50314176e-01 1.60737753e-01 8.93076360e-01 -4.96449471e-01
-8.30692172e-01 -9.36298907e-01 -3.85717362e-01 7.23006606e-01
-1.29620224e-01 -4.70987916e-01 -9.34248269e-01 -1.58438504e+00
-3.55008654e-02 6.92553818e-01 -5.94839334e-01 -1.58441857e-01
-4.71507668e-01 -8.89239848e-01 1.34905314e+00 -6.59386963e-02
3.63864690e-01 -1.37523699e+00 -8.85220766e-01 1.21724717e-02
4.63474184e-01 -1.36364329e+00 4.28849995e-01 4.65412587e-01
-5.74932098e-01 -1.50855839e+00 -4.74950559e-02 -1.08319268e-01
3.16826999e-01 -1.81494936e-01 1.36767638e+00 4.06991482e-01
-5.85851550e-01 3.93807262e-01 -4.41313386e-01 -2.52295136e-01
-7.50018954e-01 -4.39612661e-03 3.10917795e-01 1.49048045e-01
7.22735524e-01 -9.47926164e-01 -3.28595370e-01 7.06667840e-01
-9.85342920e-01 -7.61848748e-01 7.72240400e-01 4.94218588e-01
-1.03192337e-01 1.08088300e-01 1.00858223e+00 -1.35780609e+00
5.30076563e-01 -6.84324145e-01 -6.38837516e-01 3.47628295e-01
-1.06051981e+00 -8.19858685e-02 8.56734276e-01 -6.19321465e-01
-8.27467680e-01 -6.75878644e-01 -9.36717913e-02 -4.12562281e-01
-9.61899102e-01 1.36024468e-02 -4.15801734e-01 -3.17728341e-01
8.98482740e-01 -3.31340492e-01 -2.65261270e-02 -4.20841098e-01
3.72923434e-01 7.05925643e-01 1.98743135e-01 -7.36275852e-01
9.93240118e-01 5.57046235e-01 4.97046262e-02 -6.16597474e-01
-4.27016467e-01 -3.71974707e-01 -5.34555614e-01 -1.12737007e-01
2.54476726e-01 -2.26866424e-01 -9.72756684e-01 3.13671857e-01
-9.31232214e-01 -2.82510489e-01 -4.47519600e-01 3.43648791e-01
-9.31444764e-02 6.52587235e-01 -5.44439077e-01 -9.59006131e-01
-1.11011907e-01 -1.27330410e+00 5.53528547e-01 -1.11516595e-01
-2.52268255e-01 -1.28220034e+00 1.90900773e-01 2.23896652e-01
8.37099731e-01 4.68586862e-01 7.75236428e-01 -1.45063496e+00
-3.54875147e-01 -4.69564825e-01 -5.84660470e-01 7.04647839e-01
2.11712420e-01 3.44068527e-01 -1.44858146e+00 -2.93835431e-01
8.06587860e-02 -4.98586774e-01 3.59181464e-01 -2.21295580e-01
8.70075226e-01 -3.18856657e-01 -1.56000316e-01 4.32394862e-01
1.40701306e+00 2.58526087e-01 6.97211921e-01 7.22547054e-01
3.88449669e-01 8.19201410e-01 2.33461931e-01 4.60892111e-01
-1.05865963e-01 5.56877792e-01 8.46259773e-01 -3.26753616e-01
-5.57307713e-02 1.09655112e-01 2.50814766e-01 1.92896843e-01
1.24079384e-01 -1.20104618e-01 -8.90098095e-01 2.46229351e-01
-1.32460082e+00 -9.59444523e-01 -1.49362803e-01 2.29578328e+00
6.22051418e-01 6.60249650e-01 3.87771487e-01 5.07691383e-01
6.89061463e-01 1.38114557e-01 -5.63302934e-01 -6.03935063e-01
-9.71445590e-02 5.87442875e-01 2.69559175e-01 3.20530176e-01
-1.05838192e+00 6.47866309e-01 6.11624861e+00 8.55180144e-01
-1.27643836e+00 1.99414983e-01 6.08775795e-01 -3.66941355e-02
2.38555297e-02 3.03667516e-01 -7.66480863e-01 5.59600413e-01
1.45717239e+00 2.06678748e-01 3.72418731e-01 1.07049811e+00
7.07296794e-03 8.78375173e-02 -1.22782516e+00 7.15576589e-01
2.27929484e-02 -1.03756690e+00 -8.17397982e-02 4.04106200e-01
2.39804700e-01 1.48341238e-01 -4.92912531e-02 7.31133521e-01
2.91591465e-01 -9.04814184e-01 2.78869510e-01 1.74835503e-01
7.99159586e-01 -6.37734830e-01 8.89149249e-01 1.83302447e-01
-6.80904746e-01 8.36359933e-02 -1.79529581e-02 1.31971136e-01
5.76306470e-02 8.98532212e-01 -9.66036737e-01 6.42942429e-01
5.10430276e-01 3.93050730e-01 -9.67337072e-01 1.09669507e+00
-7.00966045e-02 7.85982609e-01 -6.03532493e-01 1.50058478e-01
3.11398417e-01 -1.32514555e-02 3.28141868e-01 1.52777386e+00
-2.31628850e-01 -6.14449978e-01 1.80946663e-01 8.04507494e-01
3.59787308e-02 -3.77010554e-02 -7.52041519e-01 -2.65776604e-01
3.23508501e-01 1.62669230e+00 -7.36508846e-01 -2.18074203e-01
-3.58891875e-01 6.18950546e-01 1.43761292e-01 2.29762971e-01
-1.01735067e+00 -1.94080263e-01 8.30123663e-01 1.41185567e-01
-2.47955784e-01 2.50083178e-01 -1.37232691e-01 -1.08821273e+00
1.12400025e-01 -1.21081913e+00 4.83431011e-01 1.76447406e-02
-1.70048559e+00 9.56085622e-01 7.53500685e-02 -1.13556683e+00
-3.39211047e-01 -7.82458067e-01 -4.53738928e-01 4.99509990e-01
-1.70610213e+00 -1.17784572e+00 -8.74820501e-02 6.35054708e-01
-3.26793566e-02 -3.53010386e-01 9.50072289e-01 8.29901099e-01
-8.13313782e-01 8.60503614e-01 -3.30178887e-01 5.76578118e-02
8.87036681e-01 -1.20186877e+00 3.36576462e-01 6.80236042e-01
2.48472556e-01 7.21710443e-01 4.65441048e-01 -5.11309385e-01
-7.88120985e-01 -8.93800318e-01 5.70060551e-01 -8.43010187e-01
9.03437316e-01 -5.28574109e-01 -1.05834854e+00 5.72776079e-01
1.49106132e-02 2.01257497e-01 7.46277273e-01 2.72649169e-01
-8.67931843e-01 -5.12867048e-02 -1.59907079e+00 4.39248532e-01
1.03273165e+00 -6.30097330e-01 -2.43130565e-01 2.73741841e-01
5.90919673e-01 3.86403412e-01 -7.88520813e-01 5.30107319e-01
3.93077791e-01 -1.63775969e+00 8.94712746e-01 -9.01919186e-01
-4.13191952e-02 -1.71943441e-01 -5.72779104e-02 -8.99359107e-01
4.09622341e-01 -6.54531121e-01 -2.02420115e-01 1.74056637e+00
5.72783768e-01 -9.33789849e-01 9.60249007e-01 7.38916278e-01
3.06996405e-01 -6.08614206e-01 -8.29322338e-01 -1.20600247e+00
-2.73414284e-01 -8.58430803e-01 8.15651655e-01 1.57813382e+00
-1.37183089e-02 7.30407685e-02 -3.26990128e-01 1.25484332e-01
8.48521352e-01 -4.82942283e-01 9.01498675e-01 -1.52092087e+00
-3.20351183e-01 -7.14714348e-01 -7.17611313e-01 -6.05492480e-02
3.02392840e-01 -6.82238400e-01 -6.33329093e-01 -7.72851229e-01
-4.64282334e-02 -9.39916492e-01 -3.51290911e-01 8.16737831e-01
2.57123083e-01 2.93203980e-01 2.35889569e-01 2.72939712e-01
-8.36791039e-01 -6.76544979e-02 5.17578840e-01 1.53590783e-01
-7.72360936e-02 2.30293199e-02 -4.27797228e-01 7.24332094e-01
1.02903652e+00 -5.71761549e-01 -4.20470536e-01 1.31828740e-01
-1.48662338e-02 -4.27167088e-01 5.79177856e-01 -1.28689134e+00
-9.74987727e-03 -2.48747785e-02 -1.30426493e-02 -1.03225432e-01
1.90423623e-01 -9.79817390e-01 5.87338172e-02 4.44675684e-01
-2.31263544e-02 8.73949751e-02 2.35974982e-01 3.33473951e-01
-9.65856761e-02 -3.59690070e-01 7.22431302e-01 -1.45318255e-01
-6.85088813e-01 3.27879459e-01 4.08348516e-02 2.26620689e-01
1.19806862e+00 -3.02007794e-01 -6.40556514e-01 -1.20950833e-01
-5.01205146e-01 -1.94489554e-01 6.42615676e-01 3.53138000e-01
2.55843788e-01 -7.31023073e-01 -4.06941950e-01 5.50589263e-01
1.71503276e-01 -6.29799068e-01 5.24471421e-03 6.93449378e-01
-3.97289723e-01 3.01901639e-01 -4.67388928e-01 -6.31552517e-01
-9.99215841e-01 9.49238658e-01 2.71286726e-01 -6.45140469e-01
-3.96197647e-01 3.38960409e-01 -1.93427771e-01 -8.96353662e-01
3.94157857e-01 2.65637338e-01 -6.40710294e-02 2.21269559e-02
6.53755486e-01 2.57818401e-01 4.28308427e-01 -2.59840131e-01
-6.01180732e-01 -5.94269633e-02 -2.44260266e-01 -9.97223258e-02
1.16323447e+00 1.98454782e-01 -2.43115529e-01 2.20938668e-01
1.21562278e+00 6.89258054e-02 -9.80061769e-01 2.59729475e-02
6.03143096e-01 -5.05671263e-01 -4.59787041e-01 -1.22441590e+00
-1.19582808e+00 9.10111427e-01 7.36168921e-01 6.47153139e-01
8.47717643e-01 -3.17933023e-01 3.14455479e-01 3.51405472e-01
7.73112416e-01 -6.13347709e-01 -3.38326371e-03 1.80604085e-01
1.73734024e-01 -1.38212991e+00 -2.48843119e-01 -3.57594818e-01
-4.15664792e-01 9.58850205e-01 7.76186645e-01 1.57706529e-01
7.00451672e-01 5.38975775e-01 4.20100987e-01 -2.73539364e-01
-8.02229106e-01 -1.98582351e-01 -3.85210395e-01 1.08280313e+00
6.16963655e-02 1.00667337e-02 -1.04102172e-01 3.25478613e-01
2.02725250e-02 -3.03452816e-02 5.34121037e-01 1.05706990e+00
2.11890385e-01 -1.62639976e+00 -3.85110378e-01 4.14148092e-01
-5.66514611e-01 8.89209379e-03 -6.70214057e-01 1.25265312e+00
3.48988146e-01 1.22963786e+00 -3.69215846e-01 -6.24136508e-01
4.49678034e-01 2.73820341e-01 1.21751621e-01 -3.74921948e-01
-9.40149426e-01 -5.64935088e-01 3.68177742e-01 -6.71369433e-01
-4.02970284e-01 -3.76873165e-01 -6.37668610e-01 -6.02193117e-01
-3.11589897e-01 3.72378916e-01 9.94873583e-01 1.00380278e+00
3.24635476e-01 4.31327432e-01 8.12251449e-01 -6.13491118e-01
-8.14078093e-01 -7.36796558e-01 -5.01977563e-01 9.22734797e-01
1.42063931e-01 -8.21887910e-01 -9.39630270e-01 -3.68841290e-01] | [5.272898197174072, 7.229483127593994] |
274afadb-92d7-4e8d-8a65-ff11573ea21e | to-fit-or-not-to-fit-model-based-face | 2106.09614 | null | https://arxiv.org/abs/2106.09614v3 | https://arxiv.org/pdf/2106.09614v3.pdf | Robust Model-based Face Reconstruction through Weakly-Supervised Outlier Segmentation | In this work, we aim to enhance model-based face reconstruction by avoiding fitting the model to outliers, i.e. regions that cannot be well-expressed by the model such as occluders or make-up. The core challenge for localizing outliers is that they are highly variable and difficult to annotate. To overcome this challenging problem, we introduce a joint Face-autoencoder and outlier segmentation approach (FOCUS).In particular, we exploit the fact that the outliers cannot be fitted well by the face model and hence can be localized well given a high-quality model fitting. The main challenge is that the model fitting and the outlier segmentation are mutually dependent on each other, and need to be inferred jointly. We resolve this chicken-and-egg problem with an EM-type training strategy, where a face autoencoder is trained jointly with an outlier segmentation network. This leads to a synergistic effect, in which the segmentation network prevents the face encoder from fitting to the outliers, enhancing the reconstruction quality. The improved 3D face reconstruction, in turn, enables the segmentation network to better predict the outliers. To resolve the ambiguity between outliers and regions that are difficult to fit, such as eyebrows, we build a statistical prior from synthetic data that measures the systematic bias in model fitting. Experiments on the NoW testset demonstrate that FOCUS achieves SOTA 3D face reconstruction performance among all baselines that are trained without 3D annotation. Moreover, our results on CelebA-HQ and the AR database show that the segmentation network can localize occluders accurately despite being trained without any segmentation annotation. | ['Adam Kortylewski', 'Bernhard Egger', 'Thomas Vetter', 'Andreas Morel-Forster', 'Chunlu Li'] | 2021-06-17 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Li_Robust_Model-Based_Face_Reconstruction_Through_Weakly-Supervised_Outlier_Segmentation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Li_Robust_Model-Based_Face_Reconstruction_Through_Weakly-Supervised_Outlier_Segmentation_CVPR_2023_paper.pdf | cvpr-2023-1 | ['3d-face-reconstruction', 'face-model', 'face-reconstruction'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-2.59519070e-01 3.36584151e-01 1.73653379e-01 -5.11084497e-01
-6.19200647e-01 -2.58568078e-01 1.70655504e-01 -4.33797777e-01
-1.16904244e-01 2.25283802e-01 -2.87979729e-02 3.06775987e-01
2.62478650e-01 -3.78776729e-01 -1.10693645e+00 -6.03776515e-01
3.37385893e-01 5.55252254e-01 -1.63653284e-01 1.47312567e-01
-2.17060059e-01 5.81728160e-01 -1.75317156e+00 7.30203316e-02
8.90529752e-01 9.02074933e-01 -1.40350580e-01 7.62826726e-02
-3.68783250e-02 2.23696291e-01 -6.41411483e-01 -5.36875010e-01
4.24796581e-01 -3.02010059e-01 -1.50497630e-01 3.64393771e-01
8.18454683e-01 -5.86727679e-01 -5.24292067e-02 1.08568799e+00
6.05234444e-01 -1.06274545e-01 6.06463015e-01 -1.21000290e+00
-3.98259103e-01 2.52433568e-01 -9.36311603e-01 -3.41889709e-01
3.23646605e-01 8.54708850e-02 6.38813913e-01 -1.18535864e+00
6.73063278e-01 1.44345152e+00 1.00066888e+00 8.10648441e-01
-1.52759635e+00 -7.61187077e-01 2.74701566e-01 -2.46903837e-01
-1.73818481e+00 -9.38727319e-01 9.11495090e-01 -5.40813982e-01
5.97240508e-01 8.29084665e-02 3.83195996e-01 1.29458332e+00
-1.15675040e-01 5.37280440e-01 6.85487509e-01 -2.42356762e-01
2.37183034e-01 1.52124874e-02 -2.77789980e-01 7.60429382e-01
2.58770347e-01 2.00989619e-01 -5.25272787e-01 -1.23242177e-01
8.82664084e-01 -7.75194541e-03 -2.84168541e-01 -3.94333392e-01
-4.18192118e-01 6.96937859e-01 3.58926833e-01 2.92540222e-01
-3.55673790e-01 1.71226308e-01 6.61980808e-02 1.33433170e-03
6.76693976e-01 1.33995384e-01 -5.50211251e-01 3.22797477e-01
-1.46358752e+00 6.37404099e-02 6.67219937e-01 7.22766876e-01
9.24110413e-01 3.52370739e-01 1.79100394e-01 1.04117680e+00
7.96917737e-01 4.22537386e-01 3.23015451e-01 -1.02103221e+00
-2.64985245e-02 6.04599178e-01 -7.11952671e-02 -1.00460160e+00
-4.12416279e-01 -3.16047132e-01 -6.18757367e-01 4.45460558e-01
6.51928246e-01 -3.31143625e-02 -1.23194110e+00 2.07090878e+00
6.15085959e-01 7.77335644e-01 -1.35219172e-01 9.15165305e-01
1.03048623e+00 1.49756283e-01 -1.25998482e-01 -1.93575561e-01
1.24124229e+00 -6.34045362e-01 -7.58807361e-01 -3.95644307e-01
5.50876558e-01 -8.31548512e-01 8.65310311e-01 1.11885756e-01
-1.06475937e+00 -5.77140510e-01 -9.10556674e-01 -9.68232229e-02
-4.32976447e-02 2.93689549e-01 3.71728778e-01 5.36574721e-01
-1.08839452e+00 5.60319185e-01 -1.02983165e+00 -2.65033096e-01
6.48132861e-01 7.05154300e-01 -6.07994497e-01 2.47969329e-02
-6.94355607e-01 7.71628201e-01 8.89666751e-02 4.13169295e-01
-8.11484039e-01 -6.16898239e-01 -1.17298830e+00 -3.70733291e-02
4.87882882e-01 -5.62019229e-01 9.52635229e-01 -1.32742965e+00
-1.42139089e+00 1.04114294e+00 -5.42218447e-01 -1.22990876e-01
5.72764695e-01 -1.51408911e-01 -3.26665819e-01 -2.33431533e-02
3.87056284e-02 8.63543928e-01 1.27833688e+00 -1.75020611e+00
-2.16989204e-01 -7.25111902e-01 -4.37073886e-01 -8.88391435e-02
6.79269508e-02 -4.14684340e-02 -9.81082857e-01 -5.00773847e-01
4.84184653e-01 -8.67134631e-01 -9.34106186e-02 6.28104284e-02
-3.94806772e-01 -6.07926510e-02 9.88814175e-01 -8.73511136e-01
9.78174686e-01 -2.42068863e+00 6.23981319e-02 4.03083235e-01
2.48739123e-01 1.38345078e-01 -2.79267550e-01 -2.45651126e-01
-3.59972268e-01 2.82850862e-01 -2.33286515e-01 -1.06805420e+00
2.57530063e-02 5.18637776e-01 -7.72226676e-02 6.39315784e-01
3.19963098e-01 8.29504430e-01 -2.99263656e-01 -4.68942016e-01
8.42174515e-02 7.96189845e-01 -9.08393979e-01 2.71092296e-01
-3.65404636e-01 7.94069290e-01 1.65318456e-02 8.62421930e-01
1.15117586e+00 1.03114679e-01 5.24579436e-02 -2.35839367e-01
1.61229923e-01 -9.11445171e-02 -1.43145406e+00 1.63606238e+00
-1.89627007e-01 1.61805779e-01 5.80031514e-01 -6.84426129e-01
8.31216156e-01 4.58689243e-01 5.72676480e-01 -4.77880925e-01
2.11422101e-01 3.01129460e-01 -1.19589604e-01 -5.18133581e-01
-7.58299930e-03 -1.17531084e-01 3.04581225e-01 2.35594556e-01
3.79492164e-01 -7.74777532e-02 -9.51822400e-02 -1.26535982e-01
6.72595441e-01 4.92116481e-01 -1.22045837e-02 -3.18344384e-02
3.03853542e-01 -6.46350622e-01 1.24233687e+00 3.75427634e-01
-1.63160801e-01 9.87597823e-01 7.18000829e-01 -3.43477577e-01
-8.04014623e-01 -8.82018864e-01 -3.13221186e-01 7.64043868e-01
-1.20390080e-01 -4.20975626e-01 -1.30219936e+00 -9.86712337e-01
2.14845628e-01 4.71438169e-01 -6.14439309e-01 -1.29732102e-01
-6.18481219e-01 -5.32042623e-01 5.71187913e-01 4.19745922e-01
9.89326164e-02 -8.73593569e-01 -6.67914897e-02 4.84666564e-02
-1.56774491e-01 -1.30986142e+00 -5.26133120e-01 1.94623228e-02
-8.12444985e-01 -1.17550313e+00 -3.19501042e-01 -6.02611899e-01
1.03758156e+00 -2.27775902e-01 1.17924035e+00 6.21081293e-01
-9.30903926e-02 2.85377562e-01 1.60100535e-01 -3.64628375e-01
-3.58070344e-01 -3.56439233e-01 3.93162847e-01 2.19325945e-01
5.92884600e-01 -6.51052713e-01 -2.87942678e-01 4.41812694e-01
-8.58813524e-01 -3.54028165e-01 2.48801619e-01 9.38796520e-01
8.10842633e-01 8.09127092e-02 4.58491623e-01 -8.26005340e-01
3.13501310e-04 -3.43246609e-01 -8.40211749e-01 -3.72212678e-02
-2.47572154e-01 -6.80943802e-02 4.10917222e-01 -3.96113843e-01
-1.06083965e+00 5.16102612e-01 -6.52542889e-01 -8.28425467e-01
-4.34025228e-01 1.89785868e-01 -7.57892787e-01 -2.49125585e-01
5.72003603e-01 -1.57607988e-01 2.00707749e-01 -5.58666110e-01
1.66443646e-01 5.83772898e-01 5.23878694e-01 -6.08257353e-01
8.56690586e-01 6.76184952e-01 -1.31205186e-01 -1.00052571e+00
-7.13169754e-01 -1.36674434e-01 -9.14143682e-01 -1.05919719e-01
7.38914013e-01 -1.04568589e+00 -7.51444638e-01 4.15441304e-01
-1.24224341e+00 -2.37481430e-01 -2.71706551e-01 3.29177231e-01
-4.84871715e-01 4.69123006e-01 -5.40199101e-01 -7.56663859e-01
6.81490824e-02 -1.47155690e+00 1.34418559e+00 2.75040239e-01
-2.66155958e-01 -8.45621467e-01 -2.83870041e-01 3.58817875e-01
9.03045386e-03 3.25102568e-01 6.33390009e-01 -7.02332199e-01
-5.31674445e-01 -1.23020232e-01 -3.26276794e-02 4.17652547e-01
-2.15978120e-02 3.88285965e-01 -1.47602105e+00 -2.61337638e-01
3.63303304e-01 -7.66908005e-02 7.83498228e-01 5.16549826e-01
1.11972380e+00 -1.48175791e-01 -3.32545519e-01 9.69336987e-01
1.10779393e+00 -1.28534362e-01 6.99296653e-01 -2.61445522e-01
8.27998042e-01 6.90918326e-01 2.09083796e-01 2.62046307e-01
3.05522561e-01 7.48718381e-01 5.96646070e-01 -2.23634183e-01
-1.21088237e-01 -3.50619078e-01 4.02006865e-01 5.65154731e-01
1.41663373e-01 5.81915639e-02 -7.30842471e-01 3.14500660e-01
-1.74918866e+00 -6.29634976e-01 -6.19306229e-02 2.19044113e+00
8.19533944e-01 -5.48475049e-02 -3.43424119e-02 3.61662358e-03
6.16800547e-01 -1.58994138e-01 -4.62984383e-01 -1.35528594e-01
-1.51908427e-01 7.54432231e-02 3.45000327e-02 7.55474269e-01
-9.64688063e-01 1.30276334e+00 6.06859732e+00 6.71554267e-01
-1.34601104e+00 1.31733626e-01 5.47696292e-01 -7.61413202e-02
-2.67053783e-01 -1.10188387e-02 -9.62514818e-01 6.44316316e-01
7.33821154e-01 6.43957794e-01 5.89439213e-01 8.69144857e-01
2.92866170e-01 -1.15429819e-01 -1.34640443e+00 1.01442730e+00
2.93612152e-01 -6.80742681e-01 -2.40610644e-01 2.04192430e-01
5.59880972e-01 -8.75336677e-02 9.57981721e-02 2.29682997e-01
2.90220995e-02 -1.31220710e+00 7.64723301e-01 6.39678478e-01
6.51978970e-01 -6.71303451e-01 7.15658963e-01 3.63791227e-01
-8.39890659e-01 5.64950667e-02 -5.97224951e-01 3.68952394e-01
-3.46900038e-02 8.75741184e-01 -7.21829355e-01 2.96304107e-01
8.45470846e-01 5.00668406e-01 -3.84646177e-01 9.09225643e-01
-5.09492874e-01 5.04791617e-01 -7.62854934e-01 8.79760742e-01
-3.36981982e-01 -3.27173918e-01 5.88500023e-01 7.73485065e-01
4.53014791e-01 -4.53424826e-02 1.40904859e-01 1.34682214e+00
-3.67739350e-01 -3.65251042e-02 -7.26109982e-01 1.44954875e-01
4.07822788e-01 1.19766164e+00 -5.02372921e-01 7.88449794e-02
-5.13793945e-01 1.04463196e+00 4.42404479e-01 5.50850332e-01
-8.99381638e-01 2.65515625e-01 8.92122269e-01 3.11756849e-01
4.08844948e-01 6.51302468e-03 -3.66010547e-01 -1.28778028e+00
2.16272354e-01 -1.04311812e+00 1.89463794e-01 -6.90628529e-01
-1.23380911e+00 6.00315690e-01 -3.69420975e-01 -8.00108552e-01
-2.42248654e-01 -5.33403099e-01 -4.82588738e-01 7.63735712e-01
-1.46989882e+00 -1.36645150e+00 -2.04544321e-01 5.00104845e-01
1.45164356e-01 1.04876719e-01 8.12138379e-01 5.06452799e-01
-1.00115085e+00 8.85481358e-01 -2.30472311e-01 1.36208475e-01
9.15833652e-01 -1.08180463e+00 1.17788441e-01 9.49299753e-01
3.63469005e-01 8.08193088e-01 5.05456150e-01 -8.26952875e-01
-1.03045404e+00 -1.07360470e+00 7.62237251e-01 -6.46972895e-01
1.09772868e-02 -6.49321556e-01 -1.16755188e+00 9.67888534e-01
-2.36561522e-01 4.71688718e-01 6.32583439e-01 2.35276893e-01
-5.53532422e-01 -1.43736735e-01 -1.41465199e+00 6.17984831e-01
1.00037324e+00 -5.68450570e-01 -4.66453075e-01 6.77804798e-02
3.63077015e-01 -6.71993971e-01 -7.01983809e-01 5.63433647e-01
3.27878356e-01 -1.23118913e+00 9.17303324e-01 -2.80896783e-01
5.35358377e-02 -4.87689823e-01 -9.66940299e-02 -1.29615486e+00
9.79169831e-02 -6.44963324e-01 -4.19524908e-01 1.54920852e+00
2.70536244e-01 -4.33672071e-01 9.09395754e-01 7.30722606e-01
-1.51385948e-01 -5.84203243e-01 -1.22261298e+00 -6.50692642e-01
-5.69594689e-02 -5.17589509e-01 8.85853171e-01 9.53391492e-01
-5.36502898e-01 1.24413587e-01 -3.60078931e-01 4.48238224e-01
6.73141003e-01 -3.37331951e-01 9.56705570e-01 -1.56193697e+00
-3.75641137e-02 -2.59270042e-01 -2.83659488e-01 -9.86033499e-01
6.85525954e-01 -7.54434824e-01 2.50449955e-01 -9.70408678e-01
-1.78939089e-01 -2.68955857e-01 1.41740873e-01 6.88265681e-01
-1.60119325e-01 4.33364540e-01 -1.30933169e-02 4.17277329e-02
-4.21437591e-01 5.60094774e-01 1.02325010e+00 2.19450980e-01
-2.98854560e-01 -5.37795201e-02 -7.04186797e-01 1.43380201e+00
4.41530198e-01 -6.97312832e-01 1.58942211e-02 -6.09691858e-01
5.52439550e-03 -9.76613611e-02 4.00976002e-01 -8.75778556e-01
1.53873414e-01 3.20131361e-01 8.25078309e-01 -5.13042569e-01
5.25822043e-01 -1.33350158e+00 2.29831502e-01 4.96851318e-02
3.06073040e-01 -2.41399571e-01 3.84842962e-01 4.59488869e-01
-1.02197409e-01 -3.25145245e-01 1.06005371e+00 4.23209555e-02
-2.85141557e-01 4.64367658e-01 -4.33843881e-02 -7.71290287e-02
8.94691765e-01 -3.88337076e-01 1.74737349e-01 -4.33104515e-01
-1.06609237e+00 1.68218285e-01 9.06076968e-01 2.33319983e-01
5.48015475e-01 -1.25040781e+00 -6.84295237e-01 9.29673195e-01
-5.73386811e-02 2.03437477e-01 2.43659556e-01 8.63427222e-01
-2.25004271e-01 -1.90823544e-02 2.56054640e-01 -9.02464807e-01
-1.26780570e+00 3.94927174e-01 7.31916010e-01 1.21037833e-01
-3.76234084e-01 9.20245767e-01 2.03724474e-01 -7.24391639e-01
5.60185611e-01 -3.03080767e-01 -3.39668244e-02 1.65088490e-01
3.35780025e-01 7.81571865e-02 1.82336360e-01 -1.09156692e+00
-3.71975303e-01 8.78088951e-01 1.70733973e-01 2.88527638e-01
1.31698132e+00 -1.60489023e-01 -4.20985371e-01 1.03432029e-01
1.02734447e+00 4.54431027e-01 -1.51606810e+00 -1.55426726e-01
6.07009716e-02 -5.45477569e-01 -4.97315302e-02 -6.23034418e-01
-1.41768491e+00 8.24002504e-01 5.68146050e-01 -1.15047857e-01
9.97609735e-01 -6.11297712e-02 6.61294878e-01 -6.17478564e-02
1.01619720e-01 -1.12971306e+00 2.82248086e-03 4.56122935e-01
9.69378412e-01 -1.43159461e+00 -2.14124322e-01 -6.32318974e-01
-3.75288635e-01 9.60220158e-01 8.99348319e-01 4.77021374e-02
6.38940275e-01 5.53411007e-01 3.33092272e-01 -2.95161724e-01
-2.40448475e-01 -2.33005494e-01 4.51763064e-01 5.63356161e-01
3.10463578e-01 -3.05920810e-01 2.04158410e-01 1.07394278e+00
-1.22411646e-01 -9.61369574e-02 8.83178338e-02 3.65437686e-01
-2.06794098e-01 -1.16477644e+00 -5.84937274e-01 2.15099826e-01
-7.28184640e-01 7.73790926e-02 -5.03525555e-01 7.35480428e-01
5.63134491e-01 8.42524588e-01 2.33632833e-01 -2.40155071e-01
3.31399769e-01 4.48561549e-01 3.61882001e-01 -6.69067562e-01
-4.59278822e-01 6.91219926e-01 -3.36492419e-01 -7.84651875e-01
-1.72880873e-01 -4.87551272e-01 -1.35125852e+00 -2.47215182e-01
-7.08141565e-01 5.95892146e-02 6.61686480e-01 8.63908947e-01
5.89251757e-01 4.57700431e-01 4.36554700e-01 -8.49359989e-01
-3.53473932e-01 -8.94465208e-01 -6.64812267e-01 5.48088074e-01
3.72273684e-01 -7.46595740e-01 -6.48538888e-01 -1.81274358e-02] | [13.352090835571289, 0.29314765334129333] |
cbfb145a-bed3-4ac0-a07c-3fe97e8d8e2c | are-random-decompositions-all-we-need-in-high | 2301.12844 | null | https://arxiv.org/abs/2301.12844v2 | https://arxiv.org/pdf/2301.12844v2.pdf | Are Random Decompositions all we need in High Dimensional Bayesian Optimisation? | Learning decompositions of expensive-to-evaluate black-box functions promises to scale Bayesian optimisation (BO) to high-dimensional problems. However, the success of these techniques depends on finding proper decompositions that accurately represent the black-box. While previous works learn those decompositions based on data, we investigate data-independent decomposition sampling rules in this paper. We find that data-driven learners of decompositions can be easily misled towards local decompositions that do not hold globally across the search space. Then, we formally show that a random tree-based decomposition sampler exhibits favourable theoretical guarantees that effectively trade off maximal information gain and functional mismatch between the actual black-box and its surrogate as provided by the decomposition. Those results motivate the development of the random decomposition upper-confidence bound algorithm (RDUCB) that is straightforward to implement - (almost) plug-and-play - and, surprisingly, yields significant empirical gains compared to the previous state-of-the-art on a comprehensive set of benchmarks. We also confirm the plug-and-play nature of our modelling component by integrating our method with HEBO, showing improved practical gains in the highest dimensional tasks from Bayesmark. | ['Haitham Bou-Ammar', 'Juliusz Ziomek'] | 2023-01-30 | null | null | null | null | ['bayesian-optimisation'] | ['methodology'] | [-2.01117188e-01 3.16851199e-01 -2.33580723e-01 -1.64793059e-01
-1.32194436e+00 -5.46037376e-01 5.91511607e-01 -3.89246047e-02
-3.24640006e-01 8.15130949e-01 2.28579506e-01 -4.14908051e-01
-6.68390572e-01 -5.30436218e-01 -7.98468411e-01 -1.09572613e+00
-2.08587706e-01 1.10349834e+00 2.15967819e-01 1.27082378e-01
4.64246655e-03 1.31831765e-01 -1.40992439e+00 2.07791790e-01
7.01707125e-01 1.12767255e+00 -4.93556399e-05 7.16979027e-01
1.30845249e-01 2.36097991e-01 -3.38359118e-01 -5.60712755e-01
6.94586158e-01 -3.17520469e-01 -7.08886623e-01 -6.41503856e-02
1.68179944e-01 -1.20801874e-01 3.11741412e-01 7.67793298e-01
4.53684479e-01 6.27584383e-02 9.14512873e-01 -1.05148232e+00
4.61406484e-02 5.66599309e-01 -6.48870826e-01 -2.05975771e-02
5.86431697e-02 2.79057831e-01 1.32710576e+00 -5.77511728e-01
5.18985271e-01 1.40965641e+00 1.02111518e+00 5.24476707e-01
-1.98834014e+00 -3.53621811e-01 3.22727889e-01 -6.60428032e-02
-1.10563385e+00 -5.90999722e-01 3.61340582e-01 -5.09086251e-01
8.42819631e-01 5.59373260e-01 8.11434746e-01 1.14218080e+00
-5.00072129e-02 1.01977491e+00 1.43357050e+00 -5.10434628e-01
5.93443155e-01 7.92928487e-02 2.05436438e-01 5.79926372e-01
4.20190156e-01 3.48445654e-01 -6.30838156e-01 -3.76653045e-01
4.44360584e-01 -3.05213213e-01 -2.91272968e-01 -9.89560068e-01
-9.02322233e-01 9.56343293e-01 6.38736933e-02 -1.26246661e-01
-4.17901337e-01 4.56301481e-01 3.72154593e-01 1.13282382e-01
6.96901798e-01 3.03036422e-01 -7.59901762e-01 -3.70006830e-01
-1.31817436e+00 5.43199778e-01 1.20479560e+00 6.30558610e-01
5.59814334e-01 -2.41472378e-01 -1.93096399e-01 6.17150307e-01
4.89353895e-01 4.34311442e-02 1.99248374e-01 -1.14305925e+00
4.24971551e-01 2.40513504e-01 1.54637590e-01 -3.83572608e-01
-4.10861790e-01 -8.56744051e-01 -7.71415293e-01 4.05536145e-01
8.59998047e-01 -1.82268605e-01 -5.02819002e-01 1.62351489e+00
7.90271759e-01 -2.68242806e-01 -1.08203284e-01 8.89047027e-01
-6.22612871e-02 3.04805458e-01 -3.15300584e-01 -4.27480608e-01
1.22072661e+00 -8.41296196e-01 -4.51425701e-01 -2.29239300e-01
5.06201208e-01 -2.83867776e-01 9.58990097e-01 1.10163152e+00
-1.21322083e+00 -4.39265847e-01 -1.15515137e+00 2.95427680e-01
1.12860069e-01 -1.77928075e-01 9.41718102e-01 1.24596858e+00
-8.76696348e-01 8.94776404e-01 -1.19238412e+00 4.49728221e-02
5.60226858e-01 3.11145097e-01 5.81739731e-02 1.13124475e-01
-6.93362415e-01 7.89693654e-01 3.76834184e-01 4.98585068e-02
-1.36706293e+00 -8.81673396e-01 -3.34164858e-01 -1.21905014e-01
7.38497674e-01 -9.88146007e-01 1.42050433e+00 -8.07102859e-01
-1.51881957e+00 5.05746067e-01 -1.03512868e-01 -8.09833169e-01
9.64055777e-01 -2.84674734e-01 3.42732579e-01 -5.16024642e-02
-1.68279052e-01 4.77966905e-01 1.07963634e+00 -1.34106433e+00
-6.51615679e-01 -4.89325196e-01 8.42170715e-02 -2.35384479e-02
-2.16596469e-01 -2.83143371e-01 -3.96879345e-01 -4.01588708e-01
2.07594350e-01 -5.38245559e-01 -4.11762893e-01 -5.00977635e-02
-2.80763954e-01 -2.92501688e-01 -4.88849506e-02 -5.92709839e-01
1.51606274e+00 -1.80413890e+00 5.18136919e-01 2.29864880e-01
2.50092357e-01 -2.19646215e-01 3.26207519e-01 3.76596063e-01
2.88999211e-02 3.77298929e-02 -3.17137837e-01 -7.56253421e-01
6.02188647e-01 4.21206743e-01 -2.75367320e-01 8.71523678e-01
-1.85442030e-01 6.46707714e-01 -7.78356850e-01 -3.58264714e-01
-2.49288566e-02 1.56661347e-01 -9.35058892e-01 2.86218803e-02
-5.38912952e-01 2.42997140e-01 -2.61689663e-01 3.74259055e-01
6.10782623e-01 -3.97362299e-02 2.85804719e-01 6.81131855e-02
7.31191086e-03 5.88624179e-01 -1.59033966e+00 1.70968080e+00
-5.61463535e-01 2.73017973e-01 5.97255588e-01 -1.33143032e+00
7.39884615e-01 2.14291066e-01 4.82668996e-01 -2.46477291e-01
-1.61611363e-01 3.52404922e-01 -5.68215959e-02 -9.14766788e-02
4.45613377e-02 -5.89448035e-01 1.74045026e-01 5.37687361e-01
1.81540951e-01 -1.05760626e-01 2.94342548e-01 -1.86299860e-01
1.05558205e+00 6.92136288e-01 3.72573733e-01 -6.09274864e-01
2.57727355e-01 -1.41410246e-01 5.13207018e-01 1.08890653e+00
1.97033450e-01 5.39044440e-01 8.77488792e-01 -4.30480033e-01
-8.12807739e-01 -9.88781154e-01 -6.08755946e-01 1.31211996e+00
-4.54133451e-01 -7.56338596e-01 -1.02178407e+00 -7.55807161e-01
6.92078844e-03 8.10588896e-01 -7.11327791e-01 2.30588078e-01
-1.91581145e-01 -1.15411174e+00 3.51728529e-01 3.73930603e-01
1.64575856e-02 -4.29127961e-01 -1.06662691e+00 4.27790314e-01
1.42531227e-02 -6.16090000e-01 -7.49320164e-02 9.02151167e-01
-1.30544686e+00 -1.08390892e+00 -5.93859196e-01 1.36616692e-01
2.66433477e-01 -4.11964148e-01 1.42969000e+00 -3.64519924e-01
-1.61624283e-01 4.32467163e-01 -9.00554955e-02 -5.45497239e-01
-4.32189375e-01 1.92974776e-01 -2.94129159e-02 -2.99719851e-02
2.54270643e-01 -8.21433961e-01 -5.67425489e-01 4.59450036e-01
-6.54681206e-01 1.79188475e-01 4.24057513e-01 1.00344729e+00
4.70032394e-01 2.19537109e-01 1.13617748e-01 -5.56406558e-01
5.27949989e-01 -3.33873540e-01 -1.05120873e+00 1.79290593e-01
-1.00474560e+00 6.88314319e-01 3.51229012e-01 -2.59741634e-01
-8.55653942e-01 9.09617245e-02 -6.69159368e-02 -2.03063250e-01
-4.45119403e-02 3.74796212e-01 -1.41568944e-01 3.43310088e-01
8.25252950e-01 -3.33625004e-02 -9.12085027e-02 -9.32687461e-01
3.90771270e-01 5.55006206e-01 4.13165867e-01 -1.24976873e+00
4.91238743e-01 7.34432578e-01 2.91471004e-01 -4.44213420e-01
-1.04385340e+00 -5.48278689e-01 -6.14972889e-01 -6.60029426e-02
5.35312355e-01 -7.72224784e-01 -1.01396918e+00 -1.33161945e-02
-1.06622648e+00 -4.42783833e-01 -5.58024943e-01 1.51749179e-01
-7.68768311e-01 3.56456548e-01 -1.73263811e-02 -1.48696637e+00
-1.89756408e-01 -1.03903210e+00 1.27170181e+00 -3.55652660e-01
-2.68518478e-01 -1.06935561e+00 4.79539186e-01 4.33503062e-01
-9.89161246e-03 1.94036469e-01 7.44527280e-01 -5.01355648e-01
-6.12395346e-01 -1.30496785e-01 1.69144168e-01 5.33941627e-01
-5.81400096e-01 -2.51580566e-01 -1.11561179e+00 -3.31327111e-01
3.54222715e-01 -2.88198948e-01 1.29366648e+00 6.99578345e-01
1.09034026e+00 -4.90459621e-01 -3.01553905e-01 8.20466161e-01
1.25307584e+00 -5.38186789e-01 3.91639411e-01 5.13160825e-01
1.12572417e-01 6.34794652e-01 3.42835993e-01 6.78009868e-01
3.04643035e-01 9.77591813e-01 5.53693235e-01 2.66245425e-01
1.01950802e-01 -2.29339853e-01 4.80941653e-01 5.63991547e-01
-2.88955659e-01 1.50677443e-01 -9.03710902e-01 5.66430151e-01
-2.12125969e+00 -5.80958545e-01 -8.32791179e-02 2.49204707e+00
1.25227439e+00 5.70474148e-01 6.36679828e-01 4.25368428e-01
1.38883010e-01 2.99956724e-02 -3.80190104e-01 -2.90387958e-01
3.10657769e-01 4.88806367e-01 5.85688293e-01 7.61383414e-01
-1.01213741e+00 2.75460571e-01 6.39362383e+00 1.15503466e+00
-3.58798325e-01 5.05218446e-01 6.14870608e-01 -3.28320533e-01
-3.16461653e-01 2.04129145e-01 -1.15237868e+00 3.72739136e-01
1.09580910e+00 3.19662958e-01 8.17094862e-01 1.00876105e+00
-9.55947209e-03 -2.49901295e-01 -1.60499692e+00 7.40502119e-01
-2.75880128e-01 -1.12703454e+00 -5.94441414e-01 4.17153716e-01
7.26919949e-01 -1.41001180e-01 -2.22464547e-01 3.52723092e-01
6.24877632e-01 -1.11028874e+00 9.69013929e-01 4.82703000e-01
4.82430160e-01 -6.44166827e-01 4.40427184e-01 8.38427305e-01
-9.15133536e-01 -3.27660024e-01 -5.19987047e-01 -2.35964730e-01
8.01863819e-02 9.13182139e-01 -6.92489803e-01 6.79786205e-01
6.68109298e-01 1.78185597e-01 -3.30731183e-01 1.03367674e+00
-3.46495688e-01 7.99024999e-01 -7.49838352e-01 -1.70133799e-01
2.49922425e-01 -3.48791480e-01 6.38559103e-01 1.21292329e+00
1.56155778e-02 -3.84128898e-01 8.75671431e-02 9.90528166e-01
3.25293124e-01 -2.79319018e-01 -6.83827102e-02 2.04005480e-01
2.23200191e-02 1.08647084e+00 -6.38698936e-01 -6.59366399e-02
-5.34702837e-02 3.67875338e-01 3.47704530e-01 2.31640965e-01
-6.04544818e-01 1.96875900e-01 6.84366822e-01 2.26543248e-01
8.37388933e-01 -1.47307292e-01 -6.13476217e-01 -9.08584237e-01
1.65586576e-01 -1.18535244e+00 5.67393184e-01 -2.29364589e-01
-1.34776270e+00 1.02630936e-01 5.41933775e-01 -7.36383915e-01
-4.93501186e-01 -7.79136360e-01 -1.06215335e-01 7.90885329e-01
-9.96692538e-01 -8.25342357e-01 1.89570233e-01 1.72814831e-01
5.64783931e-01 4.19089273e-02 9.01573122e-01 -1.17637813e-01
-3.15308869e-01 3.79791290e-01 5.50315738e-01 -3.11277032e-01
6.66657761e-02 -1.68258178e+00 3.50318909e-01 6.30603969e-01
5.47062933e-01 6.51453614e-01 9.24688339e-01 -3.67335558e-01
-1.45802367e+00 -2.87145823e-01 3.03572118e-01 -8.68676186e-01
7.49656618e-01 -7.23336101e-01 -4.59029347e-01 3.26489925e-01
1.43718403e-02 -8.18362683e-02 3.93444330e-01 8.47184062e-01
-4.52579975e-01 -4.17941391e-01 -9.26985681e-01 3.91317278e-01
1.01981270e+00 -1.29695237e-01 -7.01152861e-01 3.05557698e-01
2.39329666e-01 -4.29544955e-01 -6.54659688e-01 3.92636865e-01
8.60765398e-01 -1.46589696e+00 1.29467475e+00 -8.21041107e-01
3.90815109e-01 -3.26528959e-02 -3.12422186e-01 -1.17560577e+00
3.51454131e-03 -1.09714186e+00 -8.06842387e-01 9.43686247e-01
1.48536250e-01 -6.01971090e-01 7.60649145e-01 3.84682983e-01
7.65594468e-02 -1.14400184e+00 -1.45221949e+00 -9.57483530e-01
3.10554206e-01 -1.01999295e+00 4.24919784e-01 2.22810939e-01
-1.09093674e-01 3.71011376e-01 -2.35258251e-01 -9.62546244e-02
1.15931094e+00 3.81680243e-02 8.66863787e-01 -1.43698144e+00
-1.30827272e+00 -6.80309474e-01 -6.56329095e-02 -1.48073184e+00
-6.89046085e-02 -7.85731971e-01 4.90758345e-02 -1.24921334e+00
3.02403212e-01 -4.03591424e-01 -1.50399312e-01 2.63558686e-01
-6.53132871e-02 -6.87699690e-02 1.03990689e-01 4.95440736e-02
-7.51856565e-01 4.65413183e-01 1.09251118e+00 4.38025407e-02
4.60358337e-02 4.18092877e-01 -7.88763821e-01 8.70028317e-01
3.29298556e-01 -7.86883712e-01 -3.17321151e-01 -1.31792620e-01
6.61279082e-01 1.06618918e-01 5.35418451e-01 -8.99114430e-01
6.94419667e-02 3.08699589e-02 1.88769072e-01 -6.80793166e-01
3.56307089e-01 -8.23973358e-01 2.63577431e-01 6.09731257e-01
-3.73842865e-01 -4.00189221e-01 1.22560702e-01 7.58996189e-01
3.93994063e-01 -9.02235091e-01 6.34513080e-01 -2.73723453e-01
2.44219631e-01 -1.11348875e-01 -2.17588767e-01 4.97635990e-01
5.81948578e-01 -2.75489271e-01 1.04883537e-01 -3.04149032e-01
-6.58944607e-01 1.16522036e-01 1.64854482e-01 -1.84707329e-01
-1.72293670e-02 -9.82884228e-01 -7.00935721e-01 7.80025870e-02
-3.36651176e-01 3.79247963e-01 -1.70867532e-01 1.09535229e+00
-2.31731385e-01 5.78770220e-01 3.95734191e-01 -8.27489495e-01
-1.03414428e+00 4.96556252e-01 3.60742301e-01 -7.68777251e-01
-6.13142729e-01 9.78695035e-01 1.42409876e-01 -4.59443152e-01
8.31403136e-01 -4.97976065e-01 5.39729774e-01 3.14397514e-01
3.56537253e-01 7.67624617e-01 1.76662594e-01 3.72812927e-01
-2.19840527e-01 1.64654449e-01 3.28842670e-01 -5.75404346e-01
1.68640518e+00 6.98383823e-02 -8.43442827e-02 6.38791919e-01
7.96453059e-01 -1.20439537e-01 -1.81127620e+00 -9.41160247e-02
2.19768465e-01 -5.92018247e-01 2.99092323e-01 -9.23018456e-01
-7.21985579e-01 1.09206533e+00 4.76896852e-01 5.29851735e-01
9.60022032e-01 7.70682618e-02 2.34654918e-01 4.91122425e-01
5.05786777e-01 -1.17518544e+00 -5.52428328e-02 1.16774105e-01
8.95031452e-01 -8.07887614e-01 4.48626876e-01 7.87564218e-02
-3.68765235e-01 1.06720936e+00 9.02036056e-02 2.76198238e-03
5.69073737e-01 5.22539318e-01 -3.79547566e-01 -2.07633778e-01
-1.14522660e+00 -1.37673348e-01 3.99667323e-01 7.81815648e-01
9.84899402e-02 -3.48515878e-03 -2.10300088e-01 1.00543582e+00
-3.15717816e-01 -1.48341000e-01 -1.00017443e-01 5.25909543e-01
-3.89356405e-01 -1.22078276e+00 -7.51213670e-01 3.46750438e-01
-4.51458067e-01 -1.61703572e-01 -1.62253499e-01 8.41433167e-01
5.05662709e-02 7.19440818e-01 -2.91623354e-01 1.12762302e-01
8.66764318e-03 3.55535001e-01 9.22846735e-01 -4.14928228e-01
-6.47811830e-01 5.52407384e-01 2.97095269e-01 -6.01920903e-01
-3.06719214e-01 -1.07119763e+00 -5.18329084e-01 -1.33044317e-01
-4.79812562e-01 2.06431970e-01 7.75656521e-01 1.01799965e+00
-1.04076721e-01 5.50700068e-01 2.84208298e-01 -1.08133996e+00
-1.52511561e+00 -1.04200864e+00 -4.93271589e-01 -1.59536570e-01
3.61627728e-01 -8.05186570e-01 -5.59599161e-01 -1.79261133e-01] | [6.402853488922119, 3.8478591442108154] |
3b202a37-b9c8-4dc7-9b63-849bdeb50275 | benchmarking-the-human-brain-against | 2305.14363 | null | https://arxiv.org/abs/2305.14363v1 | https://arxiv.org/pdf/2305.14363v1.pdf | Benchmarking the human brain against computational architectures | The human brain has inspired novel concepts complementary to classical and quantum computing architectures, such as artificial neural networks and neuromorphic computers, but it is not clear how their performances compare. Here we report a new methodological framework for benchmarking cognitive performance based on solving computational problems with increasing problem size. We determine computational efficiencies in experiments with human participants and benchmark these against complexity classes. We show that a neuromorphic architecture with limited field-of-view size and added noise provides a good approximation to our results. The benchmarking also suggests there is no quantum advantage on the scales of human capability compared to the neuromorphic model. Thus, the framework offers unique insights into the computational efficiency of the brain by considering it a black box. | ['Philip Walther', 'Catherine Schuman', 'Céline van Valkenhoef'] | 2023-05-15 | null | null | null | null | ['novel-concepts'] | ['reasoning'] | [ 3.69424522e-01 1.68996483e-01 5.53246915e-01 -7.67984241e-02
-3.82003412e-02 -3.95472586e-01 8.59415591e-01 -2.23466530e-02
-1.04824984e+00 8.18828285e-01 -3.43352377e-01 -3.50655802e-03
-4.32472765e-01 -1.11535621e+00 -4.86243248e-01 -8.12551618e-01
-1.49309859e-01 3.51176560e-01 3.00416648e-01 -3.91847640e-01
4.90201622e-01 2.97055930e-01 -1.84772968e+00 1.82642072e-01
8.21764886e-01 9.55584526e-01 3.14598769e-01 6.79566681e-01
4.59503591e-01 3.84752810e-01 -5.71373701e-01 -6.17128611e-01
5.79905212e-01 -4.10440356e-01 -7.61077106e-01 -6.29925311e-01
1.25601441e-01 6.89589083e-02 -6.56759083e-01 1.36127245e+00
7.17672944e-01 1.78233951e-01 4.23970401e-01 -8.28377604e-01
-7.75593102e-01 5.07482052e-01 2.32672423e-01 7.44360983e-01
3.63960475e-01 4.21877891e-01 8.48133206e-01 -3.39245230e-01
6.40641272e-01 1.05180705e+00 7.25246072e-01 6.59931600e-01
-1.37219143e+00 -3.00166100e-01 -7.31465816e-01 6.64284647e-01
-1.41521049e+00 -3.97863328e-01 2.68387169e-01 1.04569374e-02
1.39410782e+00 1.18979275e-01 8.86511326e-01 1.04924858e+00
7.27356493e-01 -2.45550275e-01 1.64045978e+00 -6.56581879e-01
7.87560940e-01 1.03184141e-01 3.28814566e-01 6.05227709e-01
4.83037025e-01 6.35649741e-01 -8.65231097e-01 -6.02466092e-02
5.07449269e-01 -4.00295228e-01 -3.10373697e-02 6.97218850e-02
-1.23943305e+00 4.56839085e-01 4.59718823e-01 6.45952582e-01
-3.14373195e-01 4.87580210e-01 3.68483156e-01 4.13031459e-01
-3.26212913e-01 8.53168726e-01 -1.31711826e-01 -3.32609802e-01
-6.40209734e-01 2.29716584e-01 8.15356612e-01 7.22500026e-01
6.61049306e-01 -4.88993637e-02 2.48825867e-02 4.63369399e-01
-3.16552639e-01 5.34972489e-01 8.65578592e-01 -1.30285406e+00
-2.40492761e-01 3.22573781e-01 -1.81904316e-01 -6.34182513e-01
-7.91411281e-01 -3.54470491e-01 -7.52755702e-01 3.27610135e-01
6.15541279e-01 2.22757161e-01 -3.92790020e-01 1.70439696e+00
-2.64663696e-01 -2.75885522e-01 2.29344353e-01 8.97409916e-01
3.97360206e-01 4.89837974e-01 -1.22910663e-01 -1.25562698e-01
1.58462250e+00 -4.80148643e-01 -4.58338052e-01 -5.35154156e-02
4.15223509e-01 -3.33097816e-01 8.25556755e-01 7.52034307e-01
-1.24463272e+00 -6.06864870e-01 -1.25987601e+00 -1.46444052e-01
-6.86876059e-01 -4.30237919e-01 9.83050704e-01 1.48292911e+00
-1.36600530e+00 1.09318602e+00 -7.14968741e-01 -6.18626356e-01
2.52437085e-01 5.72288513e-01 -9.89582762e-02 4.14730877e-01
-1.28167975e+00 1.42802584e+00 8.08936536e-01 -2.79076487e-01
-6.58571243e-01 -3.44947666e-01 -1.32802606e-01 3.30901653e-01
-1.06715709e-01 -9.93006766e-01 1.03046715e+00 -6.20899022e-01
-1.74258876e+00 9.56409812e-01 1.95838049e-01 -8.37109566e-01
1.62255868e-01 4.63610083e-01 -3.37416083e-01 3.00948769e-01
-3.87239307e-01 6.29358649e-01 3.96924317e-01 -4.90507543e-01
-2.46330425e-01 -6.72943234e-01 2.10132897e-01 -2.34151185e-01
-3.93629223e-01 2.17803176e-02 4.37971205e-01 -2.49088351e-02
1.67473361e-01 -9.57317114e-01 -1.90826803e-01 -3.32397878e-01
2.97454774e-01 -2.06272438e-01 -1.13341913e-01 -1.25076413e-01
9.09300566e-01 -1.93963706e+00 4.55851592e-02 1.83726415e-01
3.74117345e-01 -5.77902161e-02 -2.99866814e-02 4.01507407e-01
1.05860382e-01 -6.26583397e-02 -2.39956066e-01 3.45262378e-01
4.26865846e-01 2.55906016e-01 -2.19417755e-02 5.60229063e-01
-2.30874225e-01 1.12541568e+00 -8.46224427e-01 -1.81511492e-01
-9.23452303e-02 2.59586662e-01 -7.16072857e-01 -1.88548401e-01
1.13287039e-01 1.94129512e-01 -1.10777244e-01 4.66294408e-01
6.68465436e-01 -2.57762611e-01 2.78117478e-01 7.05636516e-02
-3.48527312e-01 2.71213293e-01 -7.97495782e-01 1.78398979e+00
-2.87103474e-01 8.37274313e-01 -2.12676838e-01 -1.12706435e+00
6.79147303e-01 1.30887806e-01 -2.19886843e-02 -1.59798193e+00
3.30499411e-01 3.54688793e-01 8.43470633e-01 -3.61501455e-01
4.12203670e-01 -5.29036641e-01 9.45148990e-02 5.34856617e-01
7.44198382e-01 -2.11988762e-01 4.02775168e-01 -2.23862648e-01
1.45713937e+00 -6.77519888e-02 3.36756468e-01 -8.95148993e-01
4.14273977e-01 -3.24307173e-01 9.53326225e-02 1.13262379e+00
-5.71701467e-01 3.02238464e-01 3.99610192e-01 -5.32861292e-01
-1.30758870e+00 -1.48412931e+00 -6.77292645e-01 1.03944051e+00
1.69448599e-01 -5.57974815e-01 -1.24865949e+00 4.02917117e-01
-3.96370709e-01 4.70963061e-01 -5.72783887e-01 -3.06423277e-01
-3.33515525e-01 -1.31193125e+00 9.73289013e-01 2.41434053e-02
6.10753834e-01 -1.14852178e+00 -1.55092204e+00 1.37710841e-02
2.90292382e-01 -1.23355186e+00 4.85526443e-01 6.62263095e-01
-1.07529521e+00 -8.76174510e-01 -3.79766136e-01 -4.25164014e-01
2.76403695e-01 -1.01406306e-01 1.22632313e+00 3.75807807e-02
-7.08111465e-01 3.09167922e-01 -1.51391760e-01 -4.21119571e-01
-3.65683883e-01 -6.15089275e-02 5.38739204e-01 -5.89456797e-01
5.13182998e-01 -1.02959442e+00 -9.33707476e-01 -1.84441522e-01
-8.78838062e-01 -1.81454167e-01 6.38640702e-01 6.92778528e-01
1.77871421e-01 8.16000849e-02 3.61580521e-01 -4.05043870e-01
7.54674137e-01 -2.26849884e-01 -6.17645681e-01 1.33536011e-01
-9.62800443e-01 3.27242911e-01 9.99096870e-01 -1.10474974e-01
-7.35482514e-01 -2.72218883e-01 -6.33703405e-03 4.48862582e-01
-2.00791329e-01 1.92217156e-01 2.19555989e-01 -6.00662172e-01
1.21774399e+00 4.48558450e-01 -1.66198224e-01 -1.31344035e-01
2.85344779e-01 4.62545425e-01 5.07958174e-01 -7.34959126e-01
3.72370571e-01 5.35136402e-01 4.43064332e-01 -1.16118956e+00
-2.26945117e-01 9.98538435e-02 -6.58007324e-01 -1.33714393e-01
8.31915319e-01 -3.37519825e-01 -1.26170957e+00 3.16194415e-01
-1.29625094e+00 6.59437031e-02 -5.62492549e-01 6.16598248e-01
-1.00013828e+00 4.09866929e-01 -6.25375271e-01 -1.04682815e+00
-3.07165563e-01 -8.58357549e-01 5.77181280e-01 5.13870895e-01
-2.88044661e-02 -8.01972270e-01 2.31862158e-01 2.56668985e-01
6.25913620e-01 -1.14753202e-01 8.58811319e-01 -5.12731016e-01
-6.27294064e-01 6.61854297e-02 -2.34637365e-01 1.07730880e-01
-6.82096303e-01 -4.26145047e-01 -1.48253739e+00 -1.25365973e-01
3.97942126e-01 -4.45041269e-01 8.32740486e-01 3.95958424e-02
9.94370699e-01 -8.29501543e-03 1.33854210e-01 4.55911368e-01
1.60933959e+00 3.02501202e-01 9.91576314e-01 5.38685501e-01
-1.58374116e-01 7.44161963e-01 -3.39000195e-01 4.66259181e-01
7.60594383e-02 4.53161925e-01 1.70877039e-01 7.18486190e-01
1.54093817e-01 1.93378389e-01 2.76298642e-01 9.66724932e-01
-6.62938237e-01 1.07286185e-01 -9.97023761e-01 6.10696301e-02
-1.46991730e+00 -1.38542795e+00 9.47911888e-02 2.37483311e+00
4.51796949e-01 3.24443638e-01 1.93341002e-01 2.78826296e-01
5.76226890e-01 -3.44028115e-01 -2.44219482e-01 -8.81584287e-01
-4.31266785e-01 9.54478145e-01 3.69042486e-01 -8.97779763e-02
-3.57625097e-01 4.85923380e-01 8.05507851e+00 7.94272602e-01
-8.16294968e-01 6.23505890e-01 2.25524589e-01 -2.82651365e-01
8.93294439e-02 1.23385943e-01 -2.39976466e-01 6.50609374e-01
1.78210151e+00 -5.62560081e-01 1.13429463e+00 4.61618483e-01
-3.09253603e-01 -3.69353771e-01 -1.42536187e+00 1.26013219e+00
-4.85054180e-02 -1.19382334e+00 -5.49830385e-02 4.09810156e-01
4.63704556e-01 2.87145436e-01 6.83156550e-02 3.57747555e-01
-4.08344656e-01 -1.18928707e+00 7.97954321e-01 8.06879759e-01
3.26694012e-01 -4.80047375e-01 5.51905811e-01 4.96887952e-01
-5.05525291e-01 -4.22416270e-01 -9.85047460e-01 -8.15903306e-01
-3.28177601e-01 3.09705555e-01 -1.63835466e-01 3.01975191e-01
6.94864273e-01 -2.18687579e-01 -8.49884987e-01 1.25868738e+00
3.43021274e-01 1.97020799e-01 -5.14782250e-01 -6.51231647e-01
-1.10426312e-03 -4.76520568e-01 3.22389901e-01 1.18815053e+00
6.44959986e-01 3.41052651e-01 -7.56722391e-01 1.51671278e+00
9.19287931e-03 -3.76894698e-02 -6.35395288e-01 -2.39194319e-01
2.80594736e-01 1.29773998e+00 -1.07056010e+00 -1.59202874e-01
-3.31697822e-01 8.73918951e-01 2.76794881e-01 -1.00096643e-01
-6.18376791e-01 -5.62379062e-01 1.00497574e-01 -2.20940769e-01
-1.11252576e-01 -3.89956802e-01 -6.29701674e-01 -1.11321163e+00
-1.29922345e-01 -4.14396167e-01 1.21043082e-02 -7.44644761e-01
-1.00066578e+00 5.09974897e-01 -2.50265747e-01 -8.12340379e-01
-1.19939499e-01 -1.30553472e+00 -4.83574301e-01 6.69023871e-01
-8.97861481e-01 -3.09906781e-01 -2.21561000e-01 2.34152392e-01
-2.88201183e-01 -1.98730558e-01 1.22815156e+00 1.15990728e-01
-2.10067376e-01 5.08367598e-01 3.87593448e-01 -4.05446738e-01
-5.72315827e-02 -1.24250698e+00 2.95215726e-01 7.17566073e-01
2.46252000e-01 8.33168030e-01 8.18594754e-01 1.14080951e-01
-1.78319621e+00 -8.58630985e-02 6.33408964e-01 -7.23348796e-01
7.05578387e-01 -6.60815597e-01 -6.12313926e-01 -1.75714478e-01
4.52524364e-01 -7.48546422e-02 8.34519088e-01 2.28280127e-01
-3.05067420e-01 -7.00941756e-02 -1.25184977e+00 4.29449886e-01
1.38580835e+00 -8.83941174e-01 -8.89980495e-01 3.38737011e-01
3.22377145e-01 1.79477349e-01 -8.99706542e-01 6.32752404e-02
1.05529940e+00 -1.67327011e+00 8.78438652e-01 -2.61846185e-01
3.15256327e-01 1.77272186e-01 -3.30924869e-01 -1.09972787e+00
-4.01607633e-01 -5.91741204e-01 -3.29331905e-02 4.42264587e-01
6.25841916e-02 -9.19385910e-01 5.61071575e-01 7.10379064e-01
5.20845503e-02 -4.30195689e-01 -1.65400314e+00 -1.23501587e+00
4.53734189e-01 -4.12450254e-01 3.64720315e-01 4.84604001e-01
6.31953478e-01 2.91376621e-01 3.25884491e-01 -3.08729321e-01
8.43876362e-01 2.46728212e-02 1.46019951e-01 -1.36906505e+00
-5.01518965e-01 -9.44033206e-01 -1.13244176e+00 -3.49938184e-01
6.66515976e-02 -1.21306872e+00 -1.87880456e-01 -8.35792840e-01
5.94992459e-01 8.55417997e-02 -7.25985229e-01 -1.85773969e-01
3.82747471e-01 4.78554636e-01 3.99371535e-01 6.77228123e-02
-7.90808678e-01 3.55147779e-01 9.60782051e-01 1.17404059e-01
2.68886477e-01 -5.67776859e-01 -5.06003320e-01 5.71676373e-01
7.79335499e-01 -3.95788997e-01 2.90256515e-02 -2.74996668e-01
7.13975430e-01 -1.05621308e-01 5.63656390e-01 -1.81056929e+00
4.77268606e-01 1.91034511e-01 4.27963257e-01 3.19032192e-01
3.92066300e-01 -5.11414945e-01 4.59767170e-02 1.06333053e+00
-1.48898035e-01 8.21023136e-02 2.70042345e-02 3.54182065e-01
3.15959692e-01 -7.81854510e-01 1.04399991e+00 -6.10303998e-01
-6.50012255e-01 -1.62365526e-01 -7.04787970e-01 5.48258051e-02
8.05141985e-01 -6.67259991e-01 -7.71189213e-01 2.01855019e-01
-6.60035551e-01 -3.67003411e-01 5.02917409e-01 -5.23297824e-02
2.17919648e-01 -1.04422855e+00 -3.55549186e-01 1.21713229e-01
9.05176476e-02 -1.01014936e+00 2.66793430e-01 7.64104426e-01
-7.60695159e-01 9.08849835e-01 -1.08236504e+00 -2.82443941e-01
-6.74396932e-01 7.10665941e-01 6.31838262e-01 1.41410038e-01
-2.26198077e-01 6.73340917e-01 -3.05092949e-02 -3.53512496e-01
-1.00708805e-01 -2.62041926e-01 1.11842237e-01 -6.26425594e-02
5.44089496e-01 8.43417287e-01 2.57769018e-01 -3.66448581e-01
-4.40328509e-01 2.99696386e-01 5.03189921e-01 -3.84993285e-01
1.16564047e+00 7.78045952e-02 -6.40304327e-01 5.89900613e-01
8.94912779e-01 -5.70066333e-01 -4.66451943e-01 1.16860114e-01
2.14077964e-01 -1.69270769e-01 -3.67542207e-02 -7.38904536e-01
-2.66298294e-01 1.14108109e+00 1.17780340e+00 6.12252712e-01
1.19148219e+00 -1.50065407e-01 5.10871172e-01 1.19345665e+00
1.01484025e+00 -1.68239093e+00 -4.13297229e-02 7.14981318e-01
5.35873353e-01 -8.78323317e-01 1.00942245e-02 5.34694456e-02
2.12479368e-01 1.35523212e+00 4.40384001e-01 -5.18799126e-01
5.16112566e-01 1.70278132e-01 -5.87854266e-01 -2.35092968e-01
-9.11041498e-01 -4.66809571e-01 -7.40326270e-02 6.66134536e-01
4.82594192e-01 2.07603484e-01 -8.18000317e-01 7.80729234e-01
-8.50350857e-01 1.39995828e-01 8.95423949e-01 8.51888299e-01
-5.94968796e-01 -9.32721138e-01 -4.33729410e-01 4.32768315e-01
-4.90673780e-01 -2.48866096e-01 -3.37090611e-01 5.51822662e-01
4.12040174e-01 7.50855803e-01 9.67004970e-02 -5.31638026e-01
2.29255363e-01 4.25144196e-01 1.51609993e+00 -4.85639513e-01
-7.71330655e-01 -6.28732979e-01 -2.25208253e-01 -5.55166066e-01
-3.31698984e-01 -4.20852661e-01 -1.30950713e+00 -5.34400702e-01
-2.18509644e-01 1.82862207e-01 8.27625275e-01 8.72395277e-01
3.72412771e-01 4.12501991e-01 1.61191761e-01 -9.72756743e-01
-8.91863286e-01 -8.46748114e-01 -7.87894666e-01 2.58200645e-01
-1.45073310e-01 -5.60884893e-01 -3.04250598e-01 -2.08656117e-01] | [5.571905612945557, 4.973240852355957] |
76389909-4fac-493a-b3c6-a057cb637b3b | ns3-neuro-symbolic-semantic-code-search | 2205.10674 | null | https://arxiv.org/abs/2205.10674v2 | https://arxiv.org/pdf/2205.10674v2.pdf | NS3: Neuro-Symbolic Semantic Code Search | Semantic code search is the task of retrieving a code snippet given a textual description of its functionality. Recent work has been focused on using similarity metrics between neural embeddings of text and code. However, current language models are known to struggle with longer, compositional text, and multi-step reasoning. To overcome this limitation, we propose supplementing the query sentence with a layout of its semantic structure. The semantic layout is used to break down the final reasoning decision into a series of lower-level decisions. We use a Neural Module Network architecture to implement this idea. We compare our model - NS3 (Neuro-Symbolic Semantic Search) - to a number of baselines, including state-of-the-art semantic code retrieval methods, and evaluate on two datasets - CodeSearchNet and Code Search and Question Answering. We demonstrate that our approach results in more precise code retrieval, and we study the effectiveness of our modular design when handling compositional queries. | ['Xiang Ren', 'Luis Garcia', 'Christophe Hauser', 'Miltiadis Allamanis', 'Anna Hakhverdyan', 'Shushan Arakelyan'] | 2022-05-21 | null | null | null | null | ['code-search', 'code-search'] | ['computer-code', 'computer-vision'] | [ 2.22019732e-01 9.74179730e-02 -1.24554142e-01 -4.02052224e-01
-9.02395785e-01 -6.61543190e-01 4.99963135e-01 6.81948781e-01
-3.57506245e-01 -5.15659787e-02 5.60851634e-01 -6.22131228e-01
-1.85804337e-01 -6.05064273e-01 -5.71712971e-01 1.34412363e-01
2.93035805e-02 4.05448943e-01 4.78579044e-01 -1.77129060e-01
8.07423234e-01 -6.68532252e-02 -1.71897781e+00 5.62353015e-01
9.17301595e-01 9.08632636e-01 4.80853409e-01 5.78034222e-01
-6.31635845e-01 1.22794867e+00 -6.32110655e-01 -5.47412336e-01
-3.54015559e-01 -3.53116065e-01 -1.38837218e+00 -7.35769451e-01
5.54657757e-01 -2.66214192e-01 -1.67257622e-01 1.23422086e+00
2.41463467e-01 5.37802503e-02 7.27005541e-01 -1.08293676e+00
-1.15050638e+00 8.89398396e-01 -5.93503937e-02 2.08705202e-01
7.62951910e-01 -2.86022842e-01 1.30187249e+00 -8.35046232e-01
7.75867403e-01 1.09129691e+00 9.64761436e-01 7.90736794e-01
-1.20607173e+00 -4.92157698e-01 -2.77302433e-02 2.01558664e-01
-1.35436785e+00 -4.52066481e-01 4.28753763e-01 -7.15548277e-01
1.88610458e+00 -2.16746330e-02 3.27758878e-01 8.98108602e-01
7.92426392e-02 7.24798739e-01 2.77846545e-01 -3.53572190e-01
3.77672613e-01 1.08842120e-01 3.78428906e-01 1.10302603e+00
2.30224624e-01 -2.88561642e-01 -3.26430917e-01 -4.83004659e-01
1.15114398e-01 2.07014993e-01 -1.92420483e-01 -6.25770509e-01
-6.97831094e-01 1.00230277e+00 6.63864434e-01 5.47660053e-01
1.70520041e-02 7.48390555e-01 8.63115251e-01 4.27183032e-01
2.02878550e-01 1.01027918e+00 -4.60277289e-01 -3.25294405e-01
-1.30004442e+00 3.13535690e-01 1.31520438e+00 1.11742890e+00
7.85110533e-01 -3.86055112e-01 -2.42480487e-01 1.00162053e+00
4.39959258e-01 2.57388413e-01 7.90580809e-01 -1.06006360e+00
5.32892704e-01 8.72389555e-01 -2.12806016e-01 -1.01063085e+00
-2.42892921e-01 -1.42486259e-01 5.82152009e-02 -7.04772845e-02
-9.88880768e-02 3.26698214e-01 -5.57304919e-01 1.62920642e+00
-5.73735833e-01 -1.77669272e-01 2.51521558e-01 7.62659609e-01
9.12195742e-01 2.51286894e-01 2.45412260e-01 6.64193809e-01
1.29462898e+00 -1.30572748e+00 -2.77610093e-01 -4.18749064e-01
1.02273405e+00 -6.06138647e-01 1.13624072e+00 -8.17767456e-02
-9.85546947e-01 -2.58600414e-01 -1.06915510e+00 -4.74880844e-01
-7.41886020e-01 2.13478521e-01 4.88638461e-01 5.03632367e-01
-1.55099547e+00 6.52956724e-01 -7.68100142e-01 -8.39194596e-01
4.32573140e-01 8.61376002e-02 -1.23994529e-01 -1.33976266e-01
-9.52250063e-01 9.49363172e-01 2.96259165e-01 -6.08952045e-01
-1.02808940e+00 -7.89929390e-01 -1.17829370e+00 6.32929742e-01
3.65554452e-01 -7.39705741e-01 1.65936387e+00 -8.30672443e-01
-8.53515923e-01 1.07456982e+00 -3.50985587e-01 -4.60721135e-01
-1.78961426e-01 -2.01515555e-01 -3.92839611e-01 3.21922690e-01
3.02559853e-01 7.41617560e-01 5.59391618e-01 -1.08399951e+00
-2.49749228e-01 -1.09039150e-01 5.73074460e-01 -1.08377434e-01
-5.36084950e-01 2.65158594e-01 -6.04672730e-01 -6.22105598e-01
-2.16060936e-01 -8.42549086e-01 1.42600551e-01 1.18114688e-01
-1.62226081e-01 -3.71716887e-01 5.44596016e-01 -7.67931938e-01
1.67685091e+00 -2.31792140e+00 1.71372160e-01 3.16816941e-02
3.32358003e-01 1.13272034e-01 -2.35038444e-01 7.54258871e-01
-1.53591350e-01 4.69145566e-01 -4.97427702e-01 -5.10434091e-01
3.55062515e-01 2.58169957e-02 -5.93031585e-01 -5.43909408e-02
3.23657155e-01 1.12891781e+00 -1.05292392e+00 -5.13902128e-01
-3.75934809e-01 2.71485567e-01 -1.18709171e+00 -5.81474118e-02
-5.57633638e-01 -6.21542275e-01 -6.12235248e-01 8.12257767e-01
8.54704082e-02 -4.70266223e-01 -8.19833204e-03 1.23969309e-01
1.32971570e-01 6.27375126e-01 -4.23183024e-01 2.51128316e+00
-8.16305339e-01 9.52656686e-01 -8.40720758e-02 -8.78784001e-01
6.81795955e-01 1.09879918e-01 -2.52634510e-02 -8.65722597e-01
-1.35641113e-01 4.34352428e-01 -4.41072553e-01 -9.41170871e-01
7.20871866e-01 2.38625884e-01 -3.71837348e-01 7.00036824e-01
5.88562991e-03 -1.45351574e-01 1.17314883e-01 4.36238468e-01
1.74909663e+00 4.17240083e-01 -5.53429499e-02 -5.68031907e-01
6.54644787e-01 3.84029239e-01 -2.50776917e-01 9.34294462e-01
-9.19932034e-03 4.35636282e-01 7.13543296e-01 -2.23101065e-01
-9.49598551e-01 -8.39517236e-01 1.99627861e-01 1.24714041e+00
3.19382727e-01 -9.37655389e-01 -1.09656429e+00 -8.36105287e-01
2.82500714e-01 9.11700487e-01 -6.00035727e-01 -4.91859764e-01
-5.57451189e-01 -4.18660976e-02 1.00576961e+00 7.72124112e-01
2.95298159e-01 -1.06420100e+00 -1.06844258e+00 1.44076720e-01
-5.85500002e-02 -8.54986727e-01 -5.63561141e-01 3.09317082e-01
-6.29671454e-01 -1.18564498e+00 -3.40866476e-01 -1.07317591e+00
7.78869212e-01 7.99321383e-02 1.77386570e+00 7.91150928e-01
-4.12693322e-01 7.23789394e-01 -5.78575909e-01 1.35822922e-01
-7.75420368e-01 3.43797922e-01 -5.71478665e-01 -8.70294511e-01
7.00738728e-01 -3.33252221e-01 -6.12743974e-01 4.70323442e-03
-1.38825071e+00 -2.12117538e-01 5.07884979e-01 6.21377230e-01
-9.60124135e-02 -6.79925680e-02 2.48248279e-01 -6.58645213e-01
1.18468738e+00 -6.78014636e-01 -5.03851771e-01 5.86728334e-01
-7.93768942e-01 8.69877040e-01 7.10847199e-01 -1.48828641e-01
-7.33379841e-01 -1.71715677e-01 -9.50881764e-02 -4.49055135e-01
7.50220343e-02 9.36461747e-01 5.83295345e-01 -2.00065851e-01
9.40875053e-01 2.32113510e-01 3.74910831e-02 -5.81648529e-01
4.11428481e-01 7.26529181e-01 3.87781441e-01 -1.08770037e+00
5.27917862e-01 2.36533478e-01 -3.75489444e-01 -1.91399336e-01
-5.01095176e-01 -5.61659098e-01 -2.88559705e-01 2.27788419e-01
1.03562772e+00 -7.39535034e-01 -4.69173372e-01 -4.91321869e-02
-1.37475812e+00 -2.91443944e-01 1.59183312e-02 -5.12906574e-02
-5.99695444e-01 2.82152921e-01 -5.97279191e-01 -3.27616334e-01
-2.89296269e-01 -1.46146321e+00 1.48128283e+00 -2.16476455e-01
-5.50977111e-01 -1.00844967e+00 1.96672752e-01 1.65250346e-01
9.56861675e-01 -2.65323490e-01 1.55182970e+00 -1.05177629e+00
-6.58792019e-01 -1.08172119e-01 -5.74507773e-01 1.40386969e-01
-3.60947043e-01 -2.24348053e-01 -9.77865577e-01 -1.32881001e-01
-2.08568200e-01 -7.31695056e-01 1.03189170e+00 -3.23603868e-01
1.33744311e+00 -1.87836885e-01 -5.65637827e-01 4.49160457e-01
1.92903268e+00 1.87188655e-01 3.24327946e-01 4.31309730e-01
3.34372848e-01 4.40858066e-01 8.41887817e-02 1.40502468e-01
6.45770788e-01 6.38916671e-01 4.97075975e-01 4.12933975e-01
-2.72584438e-01 -4.53760117e-01 2.99901158e-01 8.56208742e-01
7.37689734e-01 -2.23014001e-02 -1.42080939e+00 7.76944220e-01
-1.84336722e+00 -8.65401745e-01 5.30116081e-01 1.80350423e+00
8.77734303e-01 -1.71004727e-01 -3.04606646e-01 -2.69399077e-01
3.25539798e-01 -5.18267602e-02 -6.17833078e-01 -7.18905926e-01
3.88945490e-01 2.97962159e-01 3.85159999e-01 4.36612993e-01
-8.49629164e-01 9.61734951e-01 6.65415668e+00 7.68324792e-01
-9.26042497e-01 1.27232924e-01 3.37362178e-02 1.06088929e-01
-6.70154750e-01 2.54409373e-01 -4.75628585e-01 3.18459839e-01
1.05617249e+00 -1.23408794e-01 9.92964029e-01 9.25074100e-01
-5.31594813e-01 -2.55928278e-01 -1.67243683e+00 9.51179385e-01
6.08176291e-01 -1.45378602e+00 2.12795421e-01 -5.75021088e-01
4.78283733e-01 2.23171607e-01 -2.17434824e-01 8.35624993e-01
3.74015331e-01 -1.15529346e+00 1.00044405e+00 5.14940619e-01
6.44134581e-01 -2.61868000e-01 4.39234376e-01 8.53964388e-02
-1.32813799e+00 -6.00086153e-01 -1.21865250e-01 4.04266268e-02
-4.28095073e-01 -1.89156886e-02 -6.56627834e-01 5.04446208e-01
8.33580077e-01 8.63979459e-01 -1.13082838e+00 9.07739878e-01
-1.32946521e-01 4.19768281e-02 1.30104244e-01 -5.28107107e-01
3.16392154e-01 3.61742496e-01 1.67944551e-01 1.51648235e+00
3.74222368e-01 -4.52746779e-01 4.44881618e-02 1.68469691e+00
-1.91053137e-01 -1.40062720e-01 -6.77774072e-01 -5.01045048e-01
6.02170587e-01 7.08731592e-01 -7.70361006e-01 -6.27354205e-01
-7.59703934e-01 1.16012418e+00 5.42354703e-01 4.92727816e-01
-7.98738182e-01 -8.92752945e-01 7.90604830e-01 -2.44540408e-01
5.23360252e-01 -9.58694518e-02 -4.49805371e-02 -1.23795176e+00
3.36722404e-01 -9.29947197e-01 3.52015793e-01 -1.22077405e+00
-9.21576679e-01 5.66321075e-01 1.60917327e-01 -6.88205123e-01
-3.30132902e-01 -5.71024537e-01 -3.15471262e-01 8.67254615e-01
-1.63450253e+00 -6.72921717e-01 -2.34568864e-01 1.61530823e-01
6.93660736e-01 -2.67336696e-01 1.10305870e+00 5.92410326e-01
-7.67812803e-02 5.21175861e-01 -2.26899683e-02 3.46315026e-01
3.84437084e-01 -1.19473541e+00 7.13579893e-01 5.50974846e-01
6.87684268e-02 1.20119917e+00 5.57820678e-01 -5.21914542e-01
-1.53704429e+00 -8.09614241e-01 1.11600840e+00 -7.98827469e-01
1.04961061e+00 -5.05221069e-01 -7.97207475e-01 4.86728281e-01
3.14227909e-01 -2.54868805e-01 5.24996758e-01 -4.88009043e-02
-9.44695294e-01 2.96086967e-01 -9.32311773e-01 5.22611797e-01
1.21894205e+00 -1.51509070e+00 -1.04526222e+00 5.83863445e-02
1.21425939e+00 -2.25404471e-01 -7.57693768e-01 1.69222593e-01
5.61079741e-01 -8.91109765e-01 1.04418337e+00 -5.69893777e-01
8.92312407e-01 -3.48672271e-01 -3.90713274e-01 -1.13841510e+00
-2.07038317e-02 -1.89158395e-01 9.26917419e-02 8.74769628e-01
5.30961454e-01 -3.92849624e-01 4.99066174e-01 7.20795453e-01
-2.91631222e-01 -7.95591831e-01 -5.80222309e-01 -7.80687571e-01
1.21521115e-01 -3.74995679e-01 7.14452207e-01 8.23579967e-01
4.94778812e-01 1.83774143e-01 5.78703701e-01 -6.76809400e-02
1.59194753e-01 3.48080546e-01 2.73091435e-01 -1.14084017e+00
-3.53173047e-01 -1.11378002e+00 -4.61577326e-01 -1.08058846e+00
7.12786853e-01 -1.50141227e+00 9.99010280e-02 -1.73262954e+00
4.93746907e-01 -1.47822022e-01 -1.78836808e-01 6.39951110e-01
1.10390685e-01 -2.10998461e-01 8.22032839e-02 2.89752096e-01
-9.61945355e-01 3.36330622e-01 5.08424997e-01 -5.23006678e-01
2.15854466e-01 -7.38059163e-01 -7.76805162e-01 2.92184949e-01
4.91067022e-01 -6.90135717e-01 -7.44253457e-01 -1.02165961e+00
7.88878262e-01 4.71995585e-02 5.32735050e-01 -1.20740843e+00
7.12435961e-01 5.50017416e-01 -2.27808088e-01 -1.47020459e-01
2.93347277e-02 -1.00881624e+00 -1.11915529e-01 5.89972556e-01
-9.61394250e-01 4.49520558e-01 4.55048800e-01 3.82865220e-01
-4.10084635e-01 -9.51596856e-01 2.06739679e-01 -2.69817799e-01
-8.20281684e-01 -1.66569322e-01 -3.20530713e-01 5.42211950e-01
4.80484873e-01 -8.20860565e-02 -6.63466454e-01 -8.17426145e-02
-3.86881053e-01 3.29729706e-01 9.24444258e-01 8.72956276e-01
6.52011037e-01 -1.04402936e+00 -3.31882872e-02 1.04509860e-01
6.85713053e-01 -4.36722666e-01 -3.37158561e-01 4.79241818e-01
-8.92970204e-01 7.96881557e-01 1.55424058e-01 -3.38363796e-01
-1.06336451e+00 7.64293790e-01 4.12513494e-01 -1.89403095e-03
-5.11191487e-01 7.83419132e-01 -1.38830140e-01 -6.16453648e-01
5.38681567e-01 -8.71001065e-01 -1.94453448e-01 -1.06630526e-01
3.19966018e-01 -2.62859482e-02 1.71432331e-01 -1.86927572e-01
-6.13574564e-01 7.00833201e-01 1.28183320e-01 7.95642957e-02
1.08007097e+00 8.87461826e-02 -4.72534090e-01 2.59645045e-01
1.76353157e+00 -1.08903304e-01 -7.01007485e-01 -2.43179068e-01
7.36933231e-01 -2.03444794e-01 -2.93081794e-02 -1.01971591e+00
-6.96504772e-01 7.32399344e-01 2.92783022e-01 3.08527768e-01
8.94012749e-01 3.51163536e-01 7.79586792e-01 1.00563228e+00
3.36043984e-01 -9.48941231e-01 2.33362585e-01 8.34307909e-01
6.04517281e-01 -1.00068891e+00 -1.85564369e-01 6.10955320e-02
-1.54780596e-01 1.31994534e+00 4.76994783e-01 -3.63352522e-02
6.10769033e-01 2.67956704e-01 -8.19506869e-02 -8.26220930e-01
-9.20342088e-01 -1.32880136e-01 1.87861338e-01 2.17464983e-01
5.51684201e-01 -3.80401254e-01 8.78156200e-02 4.98376787e-01
1.42290331e-02 1.35485560e-01 2.29061633e-01 1.33537304e+00
-4.98462021e-01 -9.97002423e-01 1.28879800e-01 4.34940845e-01
-4.54352826e-01 -7.25902915e-01 -5.66691637e-01 5.78971446e-01
-1.94558501e-01 7.50491977e-01 1.12937868e-01 -4.13103193e-01
3.15862179e-01 4.98047680e-01 2.16705278e-01 -9.12664354e-01
-9.57660317e-01 -7.40744770e-01 9.86610651e-02 -1.07208657e+00
-3.24537277e-01 -4.40265268e-01 -1.53476226e+00 5.43546714e-02
-1.17945224e-01 3.23704571e-01 1.02047610e+00 8.80159557e-01
6.95906579e-01 6.45994186e-01 -1.89437978e-02 -4.20479208e-01
-6.00675464e-01 -5.22658348e-01 2.07107011e-02 6.13003194e-01
5.17284393e-01 -5.16561270e-01 -4.31348979e-01 4.44191545e-02] | [7.521047115325928, 8.077773094177246] |
0ee63d0e-5f23-44c6-8ee5-af0ae28b782b | shape-robust-text-detection-with-progressive | 1806.02559 | null | http://arxiv.org/abs/1806.02559v1 | http://arxiv.org/pdf/1806.02559v1.pdf | Shape Robust Text Detection with Progressive Scale Expansion Network | The challenges of shape robust text detection lie in two aspects: 1) most
existing quadrangular bounding box based detectors are difficult to locate
texts with arbitrary shapes, which are hard to be enclosed perfectly in a
rectangle; 2) most pixel-wise segmentation-based detectors may not separate the
text instances that are very close to each other. To address these problems, we
propose a novel Progressive Scale Expansion Network (PSENet), designed as a
segmentation-based detector with multiple predictions for each text instance.
These predictions correspond to different `kernels' produced by shrinking the
original text instance into various scales. Consequently, the final detection
can be conducted through our progressive scale expansion algorithm which
gradually expands the kernels with minimal scales to the text instances with
maximal and complete shapes. Due to the fact that there are large geometrical
margins among these minimal kernels, our method is effective to distinguish the
adjacent text instances and is robust to arbitrary shapes. The state-of-the-art
results on ICDAR 2015 and ICDAR 2017 MLT benchmarks further confirm the great
effectiveness of PSENet. Notably, PSENet outperforms the previous best record
by absolute 6.37\% on the curve text dataset SCUT-CTW1500. Code will be
available in https://github.com/whai362/PSENet. | ['Ruo-Ze Liu', 'Jian Yang', 'Wenhai Wang', 'Tong Lu', 'Xiang Li', 'Wenbo Hou'] | 2018-06-07 | null | null | null | null | ['curved-text-detection'] | ['computer-vision'] | [-7.63122458e-03 -7.59424493e-02 2.72828974e-02 -4.62704487e-02
-8.50061834e-01 -6.86972439e-01 5.62389731e-01 1.97221592e-01
-2.73698658e-01 7.99883008e-02 -1.59315695e-03 -1.76780269e-01
2.62407772e-02 -7.87337720e-01 -5.73340178e-01 -6.56846225e-01
1.16168573e-01 7.63171196e-01 8.11359048e-01 -1.52891889e-01
-4.72863391e-02 5.31535685e-01 -1.15203643e+00 2.17965022e-01
9.80907261e-01 1.11435783e+00 -5.70880342e-03 4.96753186e-01
-2.92474717e-01 6.02019206e-02 -3.69943649e-01 -5.24207652e-01
5.39599299e-01 -1.91662386e-02 -3.20370674e-01 1.17076091e-01
5.58813393e-01 -3.14150870e-01 -4.19838309e-01 1.11993158e+00
4.61022198e-01 -1.75876334e-01 9.26992297e-01 -1.00950634e+00
-6.49830401e-01 6.71647727e-01 -1.20642185e+00 3.44143412e-03
-1.48090571e-01 -2.09722921e-01 1.06507695e+00 -1.33884239e+00
4.95673150e-01 1.26971221e+00 9.89596188e-01 3.04840684e-01
-1.04866362e+00 -7.06153870e-01 2.17948124e-01 -5.55666909e-02
-1.78612542e+00 -7.90247545e-02 6.45625591e-01 -3.49134654e-01
5.22143364e-01 5.74504852e-01 4.51423705e-01 6.29374087e-01
7.21171796e-02 1.18784869e+00 8.24061275e-01 -1.94269970e-01
4.92054857e-02 3.99662331e-02 8.50835815e-02 6.19951308e-01
3.84726882e-01 -2.09646747e-01 -2.26563275e-01 -1.06718764e-01
7.21124530e-01 1.74306884e-01 -2.32691452e-01 -5.37584960e-01
-1.18189847e+00 7.97114193e-01 5.67846179e-01 3.66745681e-01
-8.43696892e-02 -1.96422622e-01 4.55216855e-01 -2.13286683e-01
5.86203337e-01 2.43847415e-01 -4.69569951e-01 3.33734542e-01
-1.15624893e+00 1.27369449e-01 5.06352186e-01 1.00972235e+00
4.01118040e-01 -1.14061095e-01 -3.75253201e-01 9.83567834e-01
2.91931927e-01 6.72853351e-01 2.47323483e-01 -2.44700387e-01
7.19878256e-01 1.04553235e+00 -1.37712106e-01 -9.84986007e-01
-5.56319416e-01 -4.32098836e-01 -1.10160661e+00 -1.36090936e-02
6.24832988e-01 -1.32519454e-01 -1.13157928e+00 1.06492674e+00
5.28996646e-01 -8.86217281e-02 -1.26235634e-01 8.09376836e-01
1.11531544e+00 7.28090405e-01 -3.41717973e-02 1.72138095e-01
1.66847408e+00 -7.76113451e-01 -3.64640057e-01 -2.94096053e-01
6.69703066e-01 -9.64863896e-01 7.33379841e-01 2.41446689e-01
-8.02697897e-01 -3.92512232e-01 -9.65154648e-01 1.51950335e-02
-4.65777814e-01 7.82090962e-01 1.99322358e-01 4.84975278e-01
-8.86202633e-01 2.65053004e-01 -7.22259998e-01 -3.77809823e-01
7.42154419e-01 2.65680224e-01 -2.59399246e-02 2.30397105e-01
-8.39912474e-01 4.86592859e-01 5.73521554e-01 2.02345908e-01
-2.33948961e-01 -5.50064683e-01 -6.63579643e-01 5.34752645e-02
6.71672702e-01 -2.74623960e-01 9.13037121e-01 -7.21052647e-01
-8.63328338e-01 8.88711751e-01 -1.01277851e-01 -6.03732765e-01
1.01204371e+00 -8.12048241e-02 -3.82940322e-01 1.13233551e-01
2.42665753e-01 8.24465573e-01 9.35284436e-01 -1.18680167e+00
-9.27695513e-01 -4.78040338e-01 -5.21113634e-01 2.83895671e-01
-4.09776300e-01 8.33187029e-02 -9.47323263e-01 -1.04400706e+00
5.43685317e-01 -9.26670671e-01 -1.27346590e-01 2.81389982e-01
-7.98872352e-01 -4.00409490e-01 1.17899609e+00 -6.45888865e-01
1.41027689e+00 -2.10501909e+00 -2.15434819e-01 1.93905115e-01
4.86171484e-01 3.50999326e-01 6.87453598e-02 1.63022175e-01
1.12872563e-01 1.68946296e-01 -1.75325900e-01 -3.07517588e-01
1.32363006e-01 -2.21769840e-01 -4.84826744e-01 5.00549376e-01
4.78853052e-03 1.11978459e+00 -4.24401700e-01 -7.63671756e-01
3.55001539e-01 3.85194659e-01 -5.78810051e-02 -2.85778910e-01
-1.59184650e-01 -5.19355834e-02 -6.56458557e-01 7.33263135e-01
1.10945725e+00 -2.33158469e-01 -6.52537197e-02 -2.19985574e-01
-2.61810571e-01 -2.17075080e-01 -1.28629935e+00 1.12963927e+00
3.09130102e-01 6.91309094e-01 9.49998572e-02 -7.25510776e-01
1.13699627e+00 1.28057301e-01 5.21763325e-01 -4.65765864e-01
3.40855062e-01 2.74667174e-01 -1.34785131e-01 -8.94601569e-02
5.11467993e-01 -9.03162919e-03 -1.14335559e-01 1.04413658e-01
-4.09985185e-01 -2.55535603e-01 2.96012223e-01 1.65730074e-01
8.29916179e-01 -1.78320885e-01 2.04460785e-01 -3.05211425e-01
4.93673354e-01 -1.15301453e-01 6.64401710e-01 7.44281173e-01
-4.29066330e-01 1.04214561e+00 5.07204056e-01 -5.51662445e-01
-1.14692879e+00 -1.11263716e+00 -5.87660193e-01 1.02229857e+00
2.72608548e-01 -4.30069476e-01 -7.68094778e-01 -8.57711196e-01
8.58578682e-02 4.02670532e-01 -6.54855192e-01 3.05346549e-01
-6.60797894e-01 -8.92594934e-01 6.83936000e-01 8.85676026e-01
6.93084121e-01 -8.94501984e-01 -3.99175256e-01 9.65662226e-02
-9.35155060e-03 -1.31581306e+00 -8.92765343e-01 2.03274071e-01
-8.05832446e-01 -9.54079151e-01 -1.04070663e+00 -1.02728462e+00
8.60692024e-01 3.46981585e-01 7.52065539e-01 -5.51444385e-03
-3.51146400e-01 -8.70726034e-02 -4.30844516e-01 -4.35969800e-01
-2.95063227e-01 4.65686582e-02 -2.04012707e-01 -4.26390063e-04
3.73147100e-01 1.75433923e-02 -6.71353638e-01 5.93603313e-01
-8.18428755e-01 2.20229208e-01 5.95916450e-01 6.90238357e-01
8.65302026e-01 3.57191592e-01 2.05254346e-01 -6.52812362e-01
4.65437084e-01 -8.35336074e-02 -6.72090411e-01 4.58578914e-01
-4.15849477e-01 -1.67919934e-01 7.15327621e-01 -5.23792863e-01
-8.23488355e-01 3.61269504e-01 -9.87985581e-02 -4.50310111e-01
-2.85916567e-01 4.96304780e-02 -2.02089231e-02 1.35439187e-02
6.09733582e-01 4.60250586e-01 -3.69332224e-01 -2.73957759e-01
2.79099464e-01 6.77383423e-01 3.80962372e-01 -3.17451954e-01
1.14446521e+00 7.37659395e-01 -3.01524475e-02 -8.99331927e-01
-6.66308761e-01 -6.70503974e-01 -8.53871047e-01 2.75432132e-02
9.36547995e-01 -7.61642396e-01 -3.29686463e-01 6.43147826e-01
-9.03222322e-01 -2.32722759e-01 -1.55011117e-01 8.14005136e-02
-6.58992082e-02 7.23547816e-01 -7.79320419e-01 -7.63413668e-01
-7.84966469e-01 -9.21136856e-01 1.37207639e+00 3.97797644e-01
-7.33463615e-02 -8.04098785e-01 -2.50806510e-01 2.79167026e-01
2.38084719e-02 1.43591747e-01 8.03481460e-01 -9.21938062e-01
-4.07218963e-01 -5.30291438e-01 -6.06460631e-01 7.20186234e-02
-4.87926183e-03 2.81409740e-01 -7.08949924e-01 -3.77310425e-01
-2.93976843e-01 -1.17634255e-02 1.10386205e+00 6.14484668e-01
1.12848532e+00 -6.88729882e-02 -7.21352577e-01 5.39730012e-01
1.24010170e+00 2.15504080e-01 3.89007449e-01 2.77334899e-01
8.40764880e-01 3.55077863e-01 6.35605872e-01 3.18013191e-01
1.29299924e-01 5.99173486e-01 2.29112506e-01 -4.12693888e-01
-7.69533068e-02 -2.70193011e-01 1.00125484e-01 4.22232270e-01
1.34829536e-01 -3.96308124e-01 -1.17303598e+00 4.96434242e-01
-1.83106017e+00 -5.41536570e-01 -6.03829443e-01 2.13796973e+00
4.34517920e-01 5.07163167e-01 1.48389220e-01 6.27496019e-02
1.01446247e+00 1.30093843e-01 -5.17328262e-01 2.13541780e-02
-4.17365372e-01 -2.81620063e-02 7.84222424e-01 9.40645300e-03
-1.67290604e+00 1.10964167e+00 5.37206841e+00 1.42981231e+00
-9.85186160e-01 -6.99841827e-02 7.83655763e-01 -3.77672003e-03
1.12484932e-01 -2.44897708e-01 -1.11480284e+00 4.14794803e-01
1.87153071e-01 6.12076819e-02 -5.01873456e-02 7.85933197e-01
2.15886563e-01 -8.56739655e-02 -6.71446502e-01 8.68515074e-01
-3.90494876e-02 -1.16053152e+00 8.87240469e-02 -5.64492680e-02
7.71351516e-01 1.83066696e-01 2.05017626e-01 2.39637852e-01
1.24445669e-01 -8.37261438e-01 8.33102286e-01 1.48972729e-02
8.38026583e-01 -7.20143855e-01 5.10591090e-01 4.77574974e-01
-1.78868508e+00 4.04081680e-02 -6.32170379e-01 6.94016993e-01
-6.01311438e-02 5.63416302e-01 -9.49509501e-01 4.22317892e-01
8.13921273e-01 5.37909448e-01 -7.83065259e-01 1.23912287e+00
-2.44092688e-01 5.83141148e-01 -6.28626645e-01 -1.61885202e-01
1.95026323e-01 -2.29617029e-01 7.45896339e-01 1.43360782e+00
3.40641499e-01 1.59272850e-02 2.81298608e-01 1.04916775e+00
-1.91445127e-01 5.00350773e-01 -3.24083388e-01 2.68138826e-01
4.91324097e-01 1.39152753e+00 -1.45446312e+00 -3.22412521e-01
-4.98681724e-01 8.94272327e-01 1.25331327e-01 1.95539907e-01
-1.04280436e+00 -3.88567239e-01 9.58811864e-02 2.66004831e-01
6.82366490e-01 -1.02952287e-01 -6.07360065e-01 -9.81823444e-01
3.32173318e-01 -7.35636830e-01 4.90475595e-01 -6.50645792e-01
-1.26135755e+00 7.57137299e-01 -2.92006671e-01 -1.45530713e+00
5.71865320e-01 -8.21192086e-01 -7.56111622e-01 5.84506989e-01
-1.04394281e+00 -1.43666017e+00 -2.22223267e-01 3.40237558e-01
7.56561756e-01 1.43094417e-02 5.00267625e-01 3.49921472e-02
-8.15665007e-01 8.07756960e-01 4.66520250e-01 6.44925416e-01
6.53783917e-01 -1.29123974e+00 7.03545809e-01 8.71975362e-01
2.57679909e-01 3.08417469e-01 4.53450739e-01 -9.72119749e-01
-9.86885190e-01 -1.30101240e+00 5.10956585e-01 -4.68443990e-01
7.61914492e-01 -6.83355033e-01 -1.16954446e+00 5.07887483e-01
-1.15768827e-01 1.28156200e-01 9.21992809e-02 -7.34903812e-02
-4.09874260e-01 3.72482613e-02 -9.90373135e-01 7.35540986e-01
7.72591770e-01 -6.37093186e-02 -4.71072942e-01 3.42877597e-01
3.32880676e-01 -6.20132148e-01 -6.27924979e-01 6.76554739e-01
4.76471692e-01 -8.48122478e-01 9.16638196e-01 4.42389101e-02
1.90527126e-01 -4.05142754e-01 1.88813269e-01 -8.18871856e-01
-1.99670792e-01 -3.22038472e-01 -9.42725688e-03 1.21396196e+00
5.41834414e-01 -7.83440828e-01 1.04163671e+00 2.57787019e-01
-3.59076145e-03 -9.99220967e-01 -1.14434528e+00 -8.87828946e-01
4.39684987e-01 -4.52481925e-01 4.91584241e-01 7.97274470e-01
-2.20683202e-01 1.53291270e-01 -1.17059186e-01 2.39132985e-01
5.11544764e-01 2.38754332e-01 5.01621485e-01 -1.38129878e+00
2.51627117e-02 -8.16638708e-01 -3.02352548e-01 -1.38342726e+00
-3.23141724e-01 -8.45480382e-01 2.38527954e-01 -1.47200489e+00
5.17058372e-01 -6.47596598e-01 3.79956514e-03 5.69950998e-01
-3.70018989e-01 2.72434890e-01 2.56620437e-01 4.34096426e-01
-6.75057173e-01 4.81183112e-01 1.30991757e+00 -1.52622178e-01
-2.89619297e-01 1.61602542e-01 -4.37130421e-01 1.00928903e+00
8.70565057e-01 -4.05975997e-01 5.63619239e-03 -1.17028490e-01
5.77391945e-02 -2.00659320e-01 2.89047301e-01 -9.41009104e-01
3.00862879e-01 1.66518390e-01 7.26916611e-01 -1.40229750e+00
-1.59436441e-03 -8.24271142e-01 -2.16463298e-01 3.56363684e-01
2.83822231e-02 -2.44233564e-01 4.31618601e-01 5.19613802e-01
5.50623722e-02 -1.75943688e-01 9.24786210e-01 2.27854744e-01
-5.83377302e-01 4.69927400e-01 -2.57492959e-01 1.46770075e-01
1.22963655e+00 -4.74383920e-01 -3.24222118e-01 -6.74694777e-02
-5.93220890e-01 4.65695202e-01 4.31200475e-01 4.74977046e-01
5.64301491e-01 -1.16295338e+00 -7.88098454e-01 1.28879815e-01
1.51044890e-01 4.31891352e-01 3.92574519e-01 1.01458883e+00
-6.08609140e-01 6.18934035e-01 3.79082531e-01 -7.95474529e-01
-1.62429404e+00 5.70282578e-01 5.75444043e-01 -3.99766207e-01
-1.03093565e+00 8.46627235e-01 6.81594253e-01 -4.79782045e-01
1.89459741e-01 -5.26209772e-01 -2.64673352e-01 1.02793567e-01
2.64691740e-01 3.70655924e-01 2.49746498e-02 -8.32359076e-01
-3.47660065e-01 9.28242445e-01 -3.07316184e-01 2.91421056e-01
1.00651753e+00 8.11693221e-02 1.72338247e-01 -2.25781873e-02
8.83257091e-01 1.81700349e-01 -1.42051625e+00 -4.08088267e-01
-2.95179840e-02 -2.60724902e-01 -9.87592265e-02 -6.96669042e-01
-1.13564825e+00 6.47899330e-01 6.26413703e-01 4.53220785e-01
1.13801265e+00 3.42624009e-01 8.44883502e-01 9.77979973e-02
-1.57664977e-02 -1.29128206e+00 2.55284682e-02 5.76554000e-01
9.78112340e-01 -1.21404552e+00 1.78684562e-01 -7.81096935e-01
-6.18904591e-01 1.34977746e+00 6.44604206e-01 4.31265123e-02
5.31032920e-01 4.10185486e-01 4.27446701e-02 -2.72574395e-01
-2.69432366e-01 -2.37382695e-01 8.25005352e-01 1.48631677e-01
2.81011879e-01 3.10675979e-01 -3.12189430e-01 6.09815180e-01
-1.39415726e-01 -5.44557810e-01 1.38963118e-01 6.92304790e-01
-6.38301492e-01 -7.22172558e-01 -7.85914600e-01 7.52256811e-01
-4.43100899e-01 -1.13979995e-01 -6.13440990e-01 1.03398883e+00
2.42469415e-01 6.50125146e-01 1.50101081e-01 -2.12846965e-01
3.48953128e-01 -6.91340566e-02 6.51904792e-02 -5.13881803e-01
-3.52649570e-01 6.65404141e-01 -2.96948373e-01 -5.76935671e-02
1.96262762e-01 -9.29694355e-01 -1.75951064e+00 -2.14529291e-01
-6.64899528e-01 -1.67684436e-01 2.95905530e-01 7.62790799e-01
1.82476506e-01 3.72378916e-01 4.35078591e-01 -5.29295504e-01
-5.02370119e-01 -9.25270021e-01 -6.68026209e-01 2.11414859e-01
3.14666256e-02 -3.62787008e-01 -3.93839210e-01 -1.65533394e-01] | [12.08283519744873, 2.2830724716186523] |
66b8ac2e-e65c-4e47-a39d-18cccd7695bb | uncertainty-aware-score-distribution-learning-1 | 2006.07665 | null | https://arxiv.org/abs/2006.07665v1 | https://arxiv.org/pdf/2006.07665v1.pdf | Uncertainty-aware Score Distribution Learning for Action Quality Assessment | Assessing action quality from videos has attracted growing attention in recent years. Most existing approaches usually tackle this problem based on regression algorithms, which ignore the intrinsic ambiguity in the score labels caused by multiple judges or their subjective appraisals. To address this issue, we propose an uncertainty-aware score distribution learning (USDL) approach for action quality assessment (AQA). Specifically, we regard an action as an instance associated with a score distribution, which describes the probability of different evaluated scores. Moreover, under the circumstance where fine-grained score labels are available (e.g., difficulty degree of an action or multiple scores from different judges), we further devise a multi-path uncertainty-aware score distributions learning (MUSDL) method to explore the disentangled components of a score. We conduct experiments on three AQA datasets containing various Olympic actions and surgical activities, where our approaches set new state-of-the-arts under the Spearman's Rank Correlation. | ['Jie zhou', 'Yansong Tang', 'Jiwen Lu', 'Danyang Zhang', 'Zanlin Ni', 'Ying Wu', 'Jiahuan Zhou'] | 2020-06-13 | uncertainty-aware-score-distribution-learning | http://openaccess.thecvf.com/content_CVPR_2020/html/Tang_Uncertainty-Aware_Score_Distribution_Learning_for_Action_Quality_Assessment_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Tang_Uncertainty-Aware_Score_Distribution_Learning_for_Action_Quality_Assessment_CVPR_2020_paper.pdf | cvpr-2020-6 | ['action-quality-assessment'] | ['computer-vision'] | [ 8.78368765e-02 -1.38204247e-01 -5.10020614e-01 -6.43628597e-01
-1.18297505e+00 -3.92827213e-01 2.13306516e-01 1.90823153e-01
-3.38665903e-01 7.86401212e-01 6.20395124e-01 2.17758119e-01
-6.46037877e-01 -5.58076084e-01 -3.11595410e-01 -7.31950283e-01
3.27410065e-02 2.06689924e-01 2.00018302e-01 -2.79637072e-02
4.60208058e-01 -1.97834060e-01 -1.30391538e+00 4.13655490e-01
9.42353845e-01 1.07098198e+00 -2.05142096e-01 4.62232232e-01
7.92170763e-02 9.60437715e-01 -7.67619431e-01 -7.13974714e-01
2.88723290e-01 -5.50368190e-01 -6.50724351e-01 3.38020138e-02
3.47351044e-01 -4.58539307e-01 -1.83767945e-01 1.17660344e+00
3.33893061e-01 2.41835658e-02 7.82220602e-01 -1.45275223e+00
-2.71663725e-01 5.34209728e-01 -6.82234704e-01 2.14808613e-01
6.59745693e-01 2.71493912e-01 1.31326926e+00 -4.40462619e-01
3.27780336e-01 1.22630322e+00 2.80292064e-01 3.75038415e-01
-1.07203531e+00 -5.92313588e-01 1.68662831e-01 7.77936757e-01
-1.11719000e+00 -1.36441082e-01 9.24757481e-01 -6.65324152e-01
-9.58239883e-02 4.39065434e-02 7.93238223e-01 1.03334141e+00
4.31586415e-01 1.06425977e+00 1.48134136e+00 -7.70821869e-02
5.75319588e-01 -3.60184520e-01 2.48902105e-02 5.50414920e-01
2.71497577e-01 1.60899177e-01 -8.86882246e-01 -3.27132083e-02
6.22677386e-01 -1.31553739e-01 -2.76964754e-02 -5.12658060e-01
-1.22017586e+00 5.48037171e-01 6.57828972e-02 -1.85969859e-01
-3.93755734e-01 8.70377719e-02 3.64680618e-01 2.38908261e-01
1.13487706e-01 3.50645989e-01 -4.39319015e-01 -5.42427659e-01
-9.50802922e-01 3.89914662e-01 5.06699920e-01 4.47495311e-01
5.21056771e-01 -3.72673422e-01 -7.03010142e-01 5.11870205e-01
3.37694108e-01 2.36841053e-01 4.01052892e-01 -1.38230765e+00
5.16171575e-01 6.28061950e-01 3.60708266e-01 -1.19942069e+00
-2.77587175e-01 -2.84230232e-01 -5.22862911e-01 4.58848387e-01
5.95978618e-01 -1.98462814e-01 -5.02184629e-01 1.89217353e+00
3.71278763e-01 3.11444014e-01 -1.67814106e-01 1.03376997e+00
4.92198974e-01 2.38622501e-01 1.94425732e-01 -5.30612290e-01
1.24310160e+00 -7.05058336e-01 -8.58560860e-01 1.06061704e-01
4.68352824e-01 -4.00996476e-01 1.08800447e+00 1.02554893e+00
-1.00229478e+00 -5.96387088e-01 -1.04922128e+00 2.56174624e-01
2.32733920e-01 1.58075944e-01 4.80515152e-01 5.73394477e-01
-3.72064650e-01 7.17974305e-01 -8.06537867e-01 2.46758372e-01
5.66379488e-01 2.74942994e-01 -3.66411448e-01 -2.17106745e-01
-1.12062132e+00 7.48191476e-01 2.31662393e-01 5.46614677e-02
-1.07777643e+00 -4.69132632e-01 -6.22904718e-01 -6.82756752e-02
7.90501595e-01 -4.60052043e-01 1.17841780e+00 -9.95890081e-01
-1.60507429e+00 4.87547040e-01 1.10015765e-01 -5.76043595e-03
4.72259969e-01 -5.11537313e-01 -5.15605390e-01 1.23291016e-01
1.61068290e-01 3.62200409e-01 6.34347320e-01 -8.61158669e-01
-5.12964785e-01 -6.56519771e-01 5.09320438e-01 3.70320588e-01
-2.03643486e-01 4.95801419e-02 -1.11966386e-01 -4.57381576e-01
3.23025845e-02 -8.25334430e-01 -2.27932200e-01 1.08518153e-01
-2.54465342e-01 -4.19373363e-01 -5.48018292e-02 -5.64112663e-01
1.68228912e+00 -2.08055401e+00 6.07375205e-01 7.72654787e-02
8.47667456e-02 8.43082219e-02 -1.61040500e-01 5.62095009e-02
1.29387677e-01 -4.55489047e-02 -4.36889604e-02 -1.86523739e-02
1.54721290e-01 2.31234282e-01 3.87888551e-01 4.31189179e-01
2.36219779e-01 3.93693328e-01 -1.15054584e+00 -1.04495287e+00
-2.47782804e-02 -2.15722416e-02 -7.07765579e-01 3.49483252e-01
-1.53929859e-01 6.59587145e-01 -9.24875319e-01 6.97519600e-01
3.74435186e-01 -5.91537915e-02 3.32596809e-01 -4.64835823e-01
-3.31420032e-03 -4.23919745e-02 -1.51232040e+00 2.10449553e+00
-2.71760792e-01 1.30899802e-01 -3.20139438e-01 -7.89217710e-01
7.46972620e-01 3.36178720e-01 7.51958489e-01 -5.63970268e-01
1.23932481e-01 1.02528349e-01 2.73808420e-01 -7.76629090e-01
2.23742425e-01 -2.44203866e-01 -4.22359377e-01 2.31159687e-01
1.65900856e-01 -3.94808687e-03 3.29271257e-01 1.23232156e-01
1.28206539e+00 3.66673231e-01 6.07143342e-01 3.30428556e-02
6.10664785e-01 -1.87032714e-01 1.06377900e+00 4.18134898e-01
-8.99747908e-01 6.80213332e-01 1.05850077e+00 -3.94229174e-01
-7.38779366e-01 -1.20288444e+00 4.23196368e-02 8.67863953e-01
2.86158651e-01 -7.18979239e-01 -5.90151548e-01 -9.42697406e-01
-2.99112052e-01 3.72849345e-01 -6.78030491e-01 -3.10662746e-01
-2.93676823e-01 -5.36264539e-01 2.72106647e-01 5.24534166e-01
5.36720216e-01 -9.17993128e-01 -3.19883019e-01 1.69880271e-01
-5.91608703e-01 -7.81749308e-01 -4.90235984e-01 -3.60209286e-01
-7.38912940e-01 -1.45217633e+00 -4.73762333e-01 1.52270589e-02
2.76472718e-01 -1.31901324e-01 1.22391939e+00 -3.83650392e-01
-4.93610501e-02 4.13447708e-01 -4.38080400e-01 -1.25086084e-01
-8.41164887e-02 -2.98375666e-01 2.22146466e-01 2.30581239e-01
5.22665441e-01 -4.48390931e-01 -9.83605802e-01 5.14830172e-01
-8.16888392e-01 -2.73833632e-01 8.91957939e-01 6.56142831e-01
7.28386641e-01 1.59001887e-01 8.26140106e-01 -5.43174386e-01
7.00806975e-01 -5.71637332e-01 -4.29333031e-01 4.24273461e-01
-6.19036436e-01 1.67013153e-01 1.45827770e-01 -4.23933476e-01
-1.09416962e+00 7.18289753e-03 2.46912483e-02 -2.87659228e-01
-2.92634368e-01 6.09698713e-01 -4.19226497e-01 3.24553877e-01
5.44682205e-01 -1.38914630e-01 -1.75089270e-01 -1.65628195e-01
2.36960068e-01 5.65196157e-01 3.68096292e-01 -8.22848141e-01
5.76722205e-01 2.86184818e-01 3.34264547e-01 -1.02146352e-02
-1.47421432e+00 -5.11956394e-01 -6.36465549e-01 -7.02084005e-01
1.18724418e+00 -1.01842451e+00 -8.02130163e-01 2.55797714e-01
-8.61005187e-01 9.34620500e-02 -3.25663894e-01 1.00190890e+00
-7.90144205e-01 4.64185029e-01 -3.18782389e-01 -7.97144532e-01
1.87975243e-01 -1.36724138e+00 9.18450773e-01 4.54263777e-01
-3.76933634e-01 -7.25787580e-01 4.75517541e-01 8.19076836e-01
-2.33014356e-02 4.77738410e-01 6.85589194e-01 -4.48495865e-01
-4.85619754e-01 -2.00238526e-01 -3.59901674e-02 6.07693434e-01
5.47582731e-02 -2.10205972e-01 -5.10783732e-01 3.61902378e-02
-1.04833111e-01 -6.67458713e-01 4.86471206e-01 4.59930301e-01
1.17575860e+00 -1.51016727e-01 2.30929866e-01 1.48774192e-01
1.21709192e+00 5.75541407e-02 7.11381376e-01 3.27773809e-01
4.40389901e-01 5.61950386e-01 1.12303770e+00 7.94491410e-01
3.67551297e-01 7.60888338e-01 4.82396543e-01 4.08577383e-01
1.37394369e-01 -3.13802928e-01 4.61031348e-01 8.99771214e-01
-4.87054914e-01 -3.56900208e-02 -3.49156559e-01 3.20691854e-01
-1.81457853e+00 -9.97395575e-01 -1.20940052e-01 2.31261539e+00
9.23263073e-01 3.24859798e-01 2.99886912e-01 2.15654194e-01
6.02244556e-01 3.41195345e-01 -6.15809321e-01 -1.25411242e-01
2.66960591e-01 1.95242290e-03 6.25804663e-02 1.56710207e-01
-8.91494036e-01 4.27431047e-01 6.25474405e+00 1.08366597e+00
-3.20370942e-01 4.34639715e-02 4.88213152e-01 -3.41176927e-01
-2.56134540e-01 3.81886028e-02 -3.39685321e-01 6.84117913e-01
6.26499712e-01 -1.91443056e-01 1.89202696e-01 7.74104297e-01
4.29787576e-01 -3.35255235e-01 -1.08028781e+00 9.22146618e-01
9.16232765e-02 -5.54923952e-01 -1.49193391e-01 8.35187063e-02
7.41559684e-01 -4.55193609e-01 5.15813790e-02 3.94748002e-01
4.74633962e-01 -8.63544464e-01 5.17148972e-01 9.84984457e-01
6.77892089e-01 -6.93037570e-01 8.50017309e-01 2.59775460e-01
-8.83822203e-01 -2.03005806e-01 -3.32904607e-01 -1.60290882e-01
1.35860041e-01 7.06825733e-01 -1.54428065e-01 7.20661998e-01
4.17289913e-01 6.66983604e-01 -5.34782469e-01 1.29316378e+00
-6.34761512e-01 7.73639619e-01 1.20702080e-01 2.64855623e-02
-2.63946988e-02 -1.19801849e-01 3.45793843e-01 5.63118100e-01
4.33197796e-01 4.36653167e-01 2.88369715e-01 6.25193894e-01
9.69053581e-02 1.71500966e-01 -3.08016837e-02 9.52877477e-02
2.11812139e-01 1.15242255e+00 -3.95977527e-01 -1.17280111e-01
-5.42619586e-01 6.78724587e-01 1.58673793e-01 3.53401415e-02
-8.69649589e-01 -3.09384670e-02 6.73750043e-01 -1.20588370e-01
-1.01722762e-01 -8.53147134e-02 -1.17417857e-01 -1.28306663e+00
5.84381148e-02 -9.41273451e-01 6.01121128e-01 -8.54880154e-01
-1.29382753e+00 2.03585863e-01 1.26911551e-01 -1.88175249e+00
-5.10632060e-02 -3.91663045e-01 -5.53517342e-01 3.64656508e-01
-1.34714174e+00 -6.56618178e-01 -2.45698437e-01 6.20371282e-01
6.70641363e-01 -4.89709862e-02 4.86215711e-01 1.91237479e-01
-5.04231751e-01 3.85376483e-01 -1.30783722e-01 5.11107072e-02
1.00203049e+00 -1.28136945e+00 -6.30720794e-01 5.73947608e-01
7.50708058e-02 1.83945954e-01 7.44991302e-01 -4.65590507e-01
-1.06021166e+00 -5.94020844e-01 6.07505500e-01 -5.65361440e-01
6.19357467e-01 2.57081956e-01 -6.07153952e-01 2.96630621e-01
-1.10026702e-01 4.73583750e-02 1.16889942e+00 3.86887759e-01
-3.71103287e-01 -3.24774712e-01 -1.10328984e+00 3.89280796e-01
9.96121645e-01 -2.66001016e-01 -8.07486594e-01 9.25851390e-02
4.34090972e-01 -2.38473669e-01 -1.08764422e+00 4.68427807e-01
8.32437873e-01 -1.25874019e+00 7.64235616e-01 -7.60371208e-01
9.10329401e-01 -4.77631658e-01 -3.14197183e-01 -1.45685494e+00
-4.23095405e-01 -1.34476334e-01 -1.51923969e-01 1.02560389e+00
5.60222939e-02 8.86327475e-02 7.20691323e-01 7.57184982e-01
-3.80279198e-02 -8.55229974e-01 -1.08029079e+00 -7.44079590e-01
-1.05512626e-01 -4.75598335e-01 6.14440620e-01 7.41085351e-01
3.06244612e-01 1.98205292e-01 -6.65050268e-01 1.29063815e-01
8.13354313e-01 -5.06828018e-02 5.68247259e-01 -1.42653883e+00
-6.52941346e-01 -3.15526634e-01 -8.10805142e-01 -6.81883514e-01
-2.23517157e-02 -4.07853097e-01 5.37793450e-02 -1.53249252e+00
4.76566404e-01 -6.95037842e-02 -8.18799973e-01 2.92411983e-01
-3.97031546e-01 1.14101041e-02 2.40950793e-01 2.47817457e-01
-1.41205430e+00 6.63179815e-01 1.53635728e+00 -5.56732826e-02
6.75650984e-02 2.52109230e-01 -7.03888655e-01 9.26428080e-01
5.19422591e-01 -6.91153288e-01 -6.32914543e-01 -2.59887695e-01
7.18860030e-01 6.15451276e-01 2.45341197e-01 -1.25363493e+00
7.97335580e-02 -6.44531786e-01 2.07372814e-01 -3.44550401e-01
1.33901596e-01 -6.65700853e-01 -5.96987410e-03 2.38810837e-01
-6.52074456e-01 -1.52689070e-01 -4.07186985e-01 9.61804390e-01
-5.37611961e-01 -2.50562757e-01 5.86506188e-01 -2.67618239e-01
-5.14858484e-01 4.26541120e-01 -1.41833380e-01 1.96307883e-01
9.52276468e-01 -1.44243240e-01 -1.56356484e-01 -3.46960634e-01
-1.07150710e+00 3.84800196e-01 9.34714600e-02 2.89339751e-01
4.85213667e-01 -1.57290769e+00 -8.37692320e-01 -3.29750061e-01
3.58597338e-01 -1.67385653e-01 6.61941171e-01 9.32510734e-01
4.09656540e-02 -1.29314903e-02 -4.76383388e-01 -3.83084863e-01
-1.15662110e+00 3.58431935e-01 1.04622252e-01 -6.58259928e-01
2.16262285e-02 4.90778327e-01 1.58376336e-01 -6.01874068e-02
1.70198992e-01 -1.45988941e-01 -6.50871575e-01 2.43172750e-01
5.37764907e-01 8.29463959e-01 -2.26705030e-01 -4.72072750e-01
-1.91817522e-01 6.02018356e-01 1.21289603e-01 -2.93995023e-01
1.13416648e+00 -1.10025480e-01 2.72539884e-01 6.75881028e-01
7.86582291e-01 4.53794375e-02 -1.52439737e+00 -2.01179698e-01
-4.92406674e-02 -7.90675402e-01 1.23465963e-01 -1.15523982e+00
-1.01673257e+00 7.43134320e-01 7.59729564e-01 -6.42663985e-02
1.14293647e+00 -8.91746432e-02 5.41941762e-01 1.60128251e-01
6.59064054e-01 -1.34921062e+00 6.72062397e-01 -6.01948164e-02
6.73718810e-01 -1.45945060e+00 1.22494154e-01 -1.62383780e-01
-9.11413312e-01 9.94911134e-01 7.93011606e-01 -2.08087023e-02
4.93224710e-01 -2.08530545e-01 2.19103917e-02 -9.66336057e-02
-6.91523850e-01 -1.16483532e-01 3.83245736e-01 5.34040570e-01
3.74271095e-01 3.86450291e-01 -8.26724231e-01 1.02003145e+00
5.86657710e-02 3.86136323e-01 7.48517454e-01 8.01877379e-01
-3.78337175e-01 -1.18415475e+00 -2.22422197e-01 4.14568037e-01
-5.16199470e-01 2.01611146e-01 -3.54487412e-02 3.85217398e-01
3.14503431e-01 1.05755937e+00 -3.52568597e-01 -5.69843590e-01
5.78948498e-01 -2.50941571e-02 6.63336217e-01 -4.74616885e-01
-2.12522000e-01 -4.53388132e-02 1.32554770e-03 -9.16556418e-01
-8.15963984e-01 -9.81588840e-01 -1.03837800e+00 4.60420027e-02
-3.12790602e-01 1.60945550e-01 3.95829558e-01 1.13391614e+00
1.03628203e-01 5.30346513e-01 6.67914331e-01 -2.53861517e-01
-8.18301558e-01 -9.27149236e-01 -9.65564072e-01 4.89810079e-01
4.98248525e-02 -9.62840319e-01 -4.42991704e-01 -2.67708413e-02] | [8.248050689697266, 0.6849004030227661] |
4c26e544-40e5-470e-8eed-8a37bee3d257 | uvosam-a-mask-free-paradigm-for-unsupervised | 2305.12659 | null | https://arxiv.org/abs/2305.12659v1 | https://arxiv.org/pdf/2305.12659v1.pdf | UVOSAM: A Mask-free Paradigm for Unsupervised Video Object Segmentation via Segment Anything Model | Unsupervised video object segmentation has made significant progress in recent years, but the manual annotation of video mask datasets is expensive and limits the diversity of available datasets. The Segment Anything Model (SAM) has introduced a new prompt-driven paradigm for image segmentation, unlocking a range of previously unexplored capabilities. In this paper, we propose a novel paradigm called UVOSAM, which leverages SAM for unsupervised video object segmentation without requiring video mask labels. To address SAM's limitations in instance discovery and identity association, we introduce a video salient object tracking network that automatically generates trajectories for prominent foreground objects. These trajectories then serve as prompts for SAM to produce video masks on a frame-by-frame basis. Our experimental results demonstrate that UVOSAM significantly outperforms current mask-supervised methods. These findings suggest that UVOSAM has the potential to improve unsupervised video object segmentation and reduce the cost of manual annotation. | ['Siyu Zhu', 'Zuozhuo Dai', 'Shengfan Zhang', 'Zhichao Wei', 'Zhenghao Zhang'] | 2023-05-22 | null | null | null | null | ['video-object-segmentation', 'video-semantic-segmentation', 'unsupervised-video-object-segmentation'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 6.41331136e-01 7.75042549e-02 -6.21434152e-01 -3.78462911e-01
-7.48347878e-01 -7.22636819e-01 5.91189682e-01 5.40345199e-02
-3.89269680e-01 3.73693407e-01 4.68288176e-02 -1.11160457e-01
9.58817080e-02 -2.47394145e-01 -6.43597126e-01 -5.06180346e-01
4.79920134e-02 3.18396866e-01 7.82053947e-01 3.87335926e-01
2.39355832e-01 4.25403148e-01 -1.60652173e+00 4.21298087e-01
6.98979795e-01 7.41512775e-01 3.27469289e-01 6.20003223e-01
-3.01901907e-01 1.03154373e+00 -4.88491952e-01 -7.66761750e-02
4.42789227e-01 -5.65361083e-01 -9.56804097e-01 7.46307194e-01
6.70059621e-01 -5.94590843e-01 -3.78460079e-01 1.02589786e+00
1.54721213e-03 4.28461045e-01 3.54361683e-01 -1.60372674e+00
-2.88915843e-01 7.46716022e-01 -7.14026928e-01 5.82425117e-01
1.54311433e-01 1.54824555e-01 9.16205883e-01 -9.41111088e-01
9.43388581e-01 1.03183854e+00 5.59614241e-01 7.05658793e-01
-1.30389845e+00 -4.88734066e-01 6.29741967e-01 7.39611313e-02
-1.29265225e+00 -6.11675084e-01 7.67757535e-01 -7.03696191e-01
6.99149311e-01 3.50507885e-01 7.38360405e-01 8.84690404e-01
-4.16005373e-01 1.38709736e+00 7.96322167e-01 -3.46086055e-01
1.91730455e-01 -9.42943171e-02 2.30393440e-01 8.41996491e-01
2.54083514e-01 -1.73041031e-01 -7.62718976e-01 -4.97061983e-02
8.20811808e-01 1.13676690e-01 -1.60351589e-01 -4.75339830e-01
-1.24602723e+00 6.86588049e-01 -4.17125039e-02 1.07407629e-01
-3.27474922e-01 2.87910014e-01 1.98322982e-01 -2.72282511e-01
5.97362518e-01 4.26145166e-01 -2.78511822e-01 -2.91273773e-01
-1.54663920e+00 2.90757060e-01 4.53167200e-01 1.19833648e+00
8.04708540e-01 1.96820408e-01 -3.37359726e-01 4.47073460e-01
3.29792917e-01 1.80930629e-01 1.75430000e-01 -1.39454162e+00
5.23417704e-02 7.84189045e-01 2.18518913e-01 -9.64724362e-01
-2.11381793e-01 7.44482502e-02 -8.31688717e-02 3.83077115e-02
4.39607143e-01 2.38328576e-02 -1.18540859e+00 1.47155190e+00
5.80883145e-01 5.51366568e-01 -2.51947135e-01 9.98351395e-01
7.00818479e-01 6.66562676e-01 4.62606370e-01 -3.70619476e-01
9.76040721e-01 -1.22911060e+00 -7.57380247e-01 -1.55739233e-01
3.45430762e-01 -6.30656421e-01 8.51862013e-01 3.22681218e-01
-1.14796615e+00 -5.38348258e-01 -6.88869298e-01 2.98937429e-02
-9.96171162e-02 5.29729994e-03 7.81696677e-01 5.20706296e-01
-9.91808593e-01 4.36313152e-01 -1.19368684e+00 -4.59653080e-01
8.91616344e-01 3.84400755e-01 -3.96749713e-02 -4.73395875e-03
-4.58945215e-01 3.11615586e-01 6.91216528e-01 -7.95150474e-02
-1.24191296e+00 -6.18771255e-01 -7.89213240e-01 -1.19256087e-01
9.32461083e-01 -3.73723269e-01 1.42039692e+00 -1.35697198e+00
-1.16785240e+00 7.51007259e-01 -3.17547321e-01 -7.28091300e-01
6.26273930e-01 -4.75722104e-01 -1.97180554e-01 6.47440732e-01
3.07636470e-01 1.32214725e+00 1.09787202e+00 -1.30941987e+00
-1.03628719e+00 7.10236728e-02 -7.47677460e-02 1.86555356e-01
-3.20103288e-01 4.43885267e-01 -1.05637383e+00 -9.41468954e-01
9.13267396e-03 -9.44115996e-01 -4.85138595e-01 -4.05760705e-02
-3.44133526e-01 -2.06408352e-01 1.40606308e+00 -4.11053509e-01
1.36552572e+00 -2.22464561e+00 1.13041610e-01 8.00612345e-02
3.62562507e-01 3.29865754e-01 -1.05507046e-01 2.19631940e-01
2.19736457e-01 1.55732140e-01 -4.64399159e-01 -4.09749568e-01
-3.23284328e-01 2.77929366e-01 -3.53415579e-01 3.29698950e-01
4.09389734e-01 1.02900207e+00 -1.12656021e+00 -9.00667131e-01
2.73783058e-01 1.52774185e-01 -6.41011655e-01 1.99124023e-01
-7.71340847e-01 5.68843424e-01 -3.73562574e-01 9.10370111e-01
4.64227140e-01 -3.57127011e-01 2.25658640e-01 -7.43269473e-02
-2.89767921e-01 -2.56320119e-01 -1.09481728e+00 1.63906991e+00
6.94366932e-01 9.77812052e-01 1.06945276e-01 -7.63019681e-01
4.56306994e-01 1.15785882e-01 1.05599236e+00 -5.47914356e-02
1.25399396e-01 -1.01886854e-01 -1.93494514e-01 -6.64738119e-01
8.12607706e-01 3.15557539e-01 1.79593027e-01 4.75425690e-01
1.12986922e-01 1.60918504e-01 5.63606620e-01 6.06455386e-01
1.04305768e+00 5.60294330e-01 4.66957986e-02 -2.90566981e-01
1.96584567e-01 5.79639375e-01 8.30221534e-01 9.17584836e-01
-5.35161614e-01 7.62020409e-01 1.82797432e-01 -3.23764622e-01
-7.73987889e-01 -1.02322066e+00 1.41764924e-01 1.28991878e+00
5.23607075e-01 -5.22095501e-01 -1.08820713e+00 -8.62996817e-01
1.90729983e-02 3.86548847e-01 -2.81130135e-01 1.73745319e-01
-6.79235518e-01 -5.63544929e-01 4.88191158e-01 6.75932646e-01
3.14539671e-01 -1.15764213e+00 -7.32928991e-01 2.68301785e-01
-2.46760383e-01 -1.59905910e+00 -7.72715747e-01 -1.62173703e-01
-8.87880087e-01 -1.29644072e+00 -4.99045879e-01 -8.90036523e-01
1.07669330e+00 8.63112748e-01 9.71372664e-01 1.63131043e-01
-3.88389975e-01 7.67529070e-01 -5.12019515e-01 -3.70657653e-01
-3.70655775e-01 -3.60244289e-02 2.08402336e-01 3.11141640e-01
4.87078398e-01 -1.24091730e-01 -6.30815029e-01 4.40355241e-01
-1.07646799e+00 4.25606430e-01 3.02141190e-01 1.79038525e-01
8.75971735e-01 -3.85177769e-02 5.34480333e-01 -1.10912859e+00
1.05465643e-01 -4.48557884e-01 -8.26283932e-01 4.00758870e-02
-4.08527553e-01 -3.85672718e-01 8.49018022e-02 -5.89886248e-01
-1.08053470e+00 5.08845627e-01 4.85689342e-01 -7.97505200e-01
-3.23561817e-01 3.38837117e-01 -4.60203663e-02 -1.14776723e-01
3.05542499e-01 6.92576766e-02 -9.89773646e-02 -5.50345004e-01
4.26568359e-01 4.72648799e-01 7.82053053e-01 -4.58995461e-01
7.75058627e-01 7.43636847e-01 -4.28663790e-01 -9.17850912e-01
-9.58691537e-01 -8.81420553e-01 -7.63271153e-01 -6.71979666e-01
1.21617281e+00 -9.91047323e-01 -2.48382583e-01 3.00034553e-01
-9.37976301e-01 -5.91524959e-01 -4.42440748e-01 3.81912470e-01
-5.20540833e-01 4.75185305e-01 -5.01194119e-01 -7.90854096e-01
-8.64062905e-02 -1.00973046e+00 9.57393467e-01 4.01635230e-01
-5.53302228e-01 -5.63057244e-01 -4.52605963e-01 5.73846281e-01
-1.58201367e-01 2.42092714e-01 3.87142092e-01 -6.66567802e-01
-1.25679266e+00 -9.34815556e-02 -2.73906767e-01 1.73351392e-01
1.35749489e-01 2.59137958e-01 -7.44951665e-01 -2.31552452e-01
-2.38033101e-01 3.48073430e-02 8.30234885e-01 5.79169571e-01
1.21787596e+00 -1.69469029e-01 -5.59632182e-01 5.41231930e-01
1.00629544e+00 4.83713299e-01 2.09036693e-01 2.73234010e-01
1.06225491e+00 5.14079928e-01 9.43666041e-01 2.52939999e-01
1.26950294e-01 5.05232871e-01 1.45498127e-01 -2.20469326e-01
-2.58866906e-01 -3.34409475e-01 4.15847361e-01 5.71817100e-01
-5.06821424e-02 -2.40301937e-01 -1.00435019e+00 8.24932635e-01
-2.25989699e+00 -1.04105449e+00 -1.01518132e-01 1.73762774e+00
6.11792505e-01 2.39392325e-01 5.15853286e-01 -1.69720098e-01
9.57034588e-01 2.13764504e-01 -5.55501163e-01 2.95101494e-01
-4.82558683e-02 -2.09961653e-01 4.26884383e-01 3.51047248e-01
-1.49410462e+00 1.41208351e+00 6.94817019e+00 6.22666895e-01
-7.00096071e-01 1.37793779e-01 5.57751000e-01 -2.62793064e-01
1.73750184e-02 1.61097080e-01 -8.78644764e-01 3.93563688e-01
4.45367038e-01 -6.90570697e-02 3.03881198e-01 1.00119936e+00
4.95459884e-01 -3.25293124e-01 -1.20892870e+00 9.52539682e-01
2.72956669e-01 -1.59610772e+00 2.71047592e-01 -1.03544869e-01
1.01096499e+00 3.00349984e-02 -7.10735098e-02 -1.09500304e-01
2.67903775e-01 -6.21293187e-01 8.94115031e-01 2.43175194e-01
5.08066833e-01 -5.52725434e-01 1.56232163e-01 1.64710879e-01
-1.32668245e+00 -6.56034350e-02 6.61481693e-02 5.15905907e-03
4.80984956e-01 2.07857773e-01 -9.25420940e-01 4.10286523e-02
7.04939961e-01 8.90681028e-01 -6.24267340e-01 1.18767214e+00
-7.34693697e-03 9.09274995e-01 -1.35802686e-01 2.57363200e-01
5.29376149e-01 -2.67759800e-01 7.85261214e-01 1.41168714e+00
-1.39040902e-01 2.16998830e-01 9.29525077e-01 8.55362833e-01
-2.08950892e-01 -1.40711054e-01 -5.01605093e-01 -4.12434340e-01
6.21546090e-01 1.31432343e+00 -1.69466507e+00 -6.39891505e-01
-4.40719187e-01 9.92683947e-01 -1.03319950e-01 5.92535794e-01
-1.01024270e+00 1.64723217e-01 6.15576804e-01 1.96477592e-01
5.13981700e-01 -4.57465708e-01 -9.05438587e-02 -1.04615998e+00
-1.63488612e-01 -8.82838070e-01 4.12298769e-01 -6.07516825e-01
-8.14645410e-01 1.72948778e-01 2.89513797e-01 -1.14806342e+00
1.04890309e-01 -1.99241325e-01 -4.25376028e-01 -3.10055837e-02
-1.12676132e+00 -1.12875903e+00 -3.01801860e-01 4.09195065e-01
1.19993854e+00 -2.92338729e-02 2.20050618e-01 2.35089794e-01
-8.53978872e-01 1.56998739e-01 -1.02267392e-01 4.13022101e-01
4.03996527e-01 -1.05872858e+00 4.60032195e-01 1.31185627e+00
6.96868837e-01 4.78423864e-01 6.10478282e-01 -9.43686843e-01
-1.30218184e+00 -1.50491619e+00 3.32752585e-01 -7.29896724e-01
3.92070502e-01 -4.32239234e-01 -7.76234388e-01 8.41267824e-01
7.31842145e-02 2.16240827e-02 6.78376555e-01 -3.02505821e-01
-4.22851332e-02 1.26857325e-01 -8.54050756e-01 7.81787336e-01
1.23227990e+00 -3.95040870e-01 -3.84348482e-01 2.98939526e-01
9.39706624e-01 -4.04285133e-01 -4.51213747e-01 3.81396055e-01
2.45846853e-01 -6.95249975e-01 9.42225635e-01 -6.64128602e-01
1.41013682e-01 -7.65958369e-01 -4.93293032e-02 -5.33604801e-01
-2.14203835e-01 -9.86246586e-01 -5.32581210e-01 1.50777328e+00
1.26399100e-01 3.88533734e-02 1.10805595e+00 8.73476863e-01
-3.07928741e-01 -5.71310759e-01 -5.99563003e-01 -7.96988428e-01
-6.69504941e-01 -5.89666426e-01 2.25269467e-01 9.93593335e-01
-2.62673646e-01 3.17554958e-02 -2.61775166e-01 1.63512602e-01
6.87791049e-01 1.96873948e-01 8.13608944e-01 -1.08637607e+00
-9.44543034e-02 -4.85945642e-01 -4.55056995e-01 -1.13525498e+00
1.92202568e-01 -8.56980622e-01 3.38245779e-01 -1.44410610e+00
4.21573937e-01 -2.78662473e-01 -3.10665280e-01 5.22176385e-01
-4.04886812e-01 6.40978694e-01 4.11237448e-01 3.90134215e-01
-1.30811787e+00 7.90587068e-02 8.44170094e-01 -2.27231413e-01
-4.95244771e-01 -2.58755118e-01 -5.88398695e-01 1.15124476e+00
6.69893086e-01 -5.71936667e-01 -5.51874042e-01 -4.20028359e-01
-3.78123879e-01 -2.36941725e-01 3.92408371e-01 -1.09654415e+00
3.07550669e-01 -4.67891186e-01 2.85825163e-01 -8.00349891e-01
1.73528984e-01 -7.35132754e-01 1.37578666e-01 2.41762400e-01
-2.72047669e-01 1.35901403e-02 2.30548337e-01 8.52906525e-01
-1.55708387e-01 -1.01414800e-01 6.42621815e-01 -8.08093473e-02
-1.33814907e+00 4.71217513e-01 -7.23358154e-01 2.17279479e-01
1.41611469e+00 -5.71731508e-01 -3.54088731e-02 -3.13736647e-02
-5.96714199e-01 2.83781052e-01 6.22992039e-01 6.17227912e-01
7.01745450e-01 -1.02613997e+00 -3.40847701e-01 1.15028307e-01
4.51114625e-02 2.01632157e-01 1.21498242e-01 7.65658915e-01
-4.84237909e-01 5.73020168e-02 1.11929543e-01 -9.47724462e-01
-1.63474178e+00 7.37067401e-01 -7.01442286e-02 3.26290041e-01
-1.00019765e+00 9.91737068e-01 2.90879220e-01 2.95574576e-01
3.66243958e-01 -2.22711489e-01 -9.88440141e-02 1.85166910e-01
5.22361696e-01 4.81313437e-01 -5.00494897e-01 -8.07642877e-01
-2.48226181e-01 3.26761127e-01 -3.34602147e-01 -1.19151071e-01
1.19167554e+00 -2.34041438e-01 -2.14943718e-02 3.90814990e-01
7.25206912e-01 -3.97812538e-02 -1.74709463e+00 -8.75140950e-02
5.38197517e-01 -6.54262900e-01 -5.09536974e-02 -4.07238871e-01
-9.99804378e-01 3.58167291e-01 2.98713505e-01 1.93832934e-01
9.82246459e-01 1.87672511e-01 8.89983058e-01 2.42750287e-01
3.26295853e-01 -1.28280926e+00 2.53688842e-01 2.84315675e-01
2.06274793e-01 -1.31511068e+00 3.61521058e-02 -7.09059417e-01
-7.03965008e-01 7.59504199e-01 8.90155911e-01 8.75485092e-02
2.56193608e-01 3.01656634e-01 1.55932456e-01 -2.92022467e-01
-3.58199984e-01 -4.19206500e-01 3.18124145e-01 6.24858737e-01
1.39497697e-01 -1.49368331e-01 -1.27465859e-01 3.56235206e-01
4.26834852e-01 5.42782024e-02 4.85568851e-01 1.20839870e+00
-6.52117014e-01 -1.01072848e+00 -3.98783445e-01 5.22132933e-01
-5.35230517e-01 -3.05627994e-02 -5.83191395e-01 6.61142051e-01
2.16400057e-01 9.87884521e-01 1.36715084e-01 -2.35180587e-01
-1.22368857e-01 1.08913653e-01 3.80483001e-01 -8.32968652e-01
-3.00565809e-01 3.74611199e-01 4.13750149e-02 -6.73872113e-01
-9.69120204e-01 -8.79389882e-01 -1.49605274e+00 2.04257771e-01
-3.62390697e-01 1.35200039e-01 2.17074841e-01 9.64679182e-01
5.29881358e-01 4.69238460e-01 1.91325933e-01 -1.11855507e+00
2.31701583e-01 -4.71422046e-01 -8.49962682e-02 5.00717402e-01
2.92826116e-01 -7.01412678e-01 1.39880255e-01 7.82807052e-01] | [9.105133056640625, -0.162195086479187] |
176df7a6-1982-4e7c-a9de-309c8bdd56cf | a-new-benchmark-for-group-distribution-shifts | 2211.00110 | null | https://arxiv.org/abs/2211.00110v1 | https://arxiv.org/pdf/2211.00110v1.pdf | A new benchmark for group distribution shifts in hand grasp regression for object manipulation. Can meta-learning raise the bar? | Understanding hand-object pose with computer vision opens the door to new applications in mixed reality, assisted living or human-robot interaction. Most methods are trained and evaluated on balanced datasets. This is of limited use in real-world applications; how do these methods perform in the wild on unknown objects? We propose a novel benchmark for object group distribution shifts in hand and object pose regression. We then test the hypothesis that meta-learning a baseline pose regression neural network can adapt to these shifts and generalize better to unknown objects. Our results show measurable improvements over the baseline, depending on the amount of prior knowledge. For the task of joint hand-object pose regression, we observe optimization interference for the meta-learner. To address this issue and improve the method further, we provide a comprehensive analysis which should serve as a basis for future work on this benchmark. | ['Gerard Lacey', 'Théo Morales'] | 2022-10-31 | null | null | null | null | ['hand-object-pose', 'mixed-reality'] | ['computer-vision', 'computer-vision'] | [ 1.38777837e-01 4.01551872e-02 -3.55602771e-01 -4.37057942e-01
-9.72035468e-01 -5.06150603e-01 4.09445554e-01 -2.93427795e-01
-4.99605268e-01 8.23256850e-01 1.15575112e-01 1.47032216e-01
-1.90668583e-01 -7.74737671e-02 -8.96407127e-01 -7.07305968e-01
-1.84754133e-01 9.70465302e-01 3.31780314e-01 -2.66222119e-01
2.26896808e-01 6.44277632e-01 -1.47632873e+00 3.46503764e-01
1.48114786e-01 6.52266562e-01 3.07359874e-01 6.95046008e-01
4.34170008e-01 6.00036204e-01 -5.49245715e-01 -1.35708466e-01
4.67209756e-01 -1.06159888e-01 -9.46867645e-01 4.17232886e-02
9.75684881e-01 -3.78849238e-01 -3.67594212e-01 8.49525928e-01
9.43238258e-01 1.93828911e-01 7.68835843e-01 -1.33117390e+00
-3.12930375e-01 4.95125741e-01 -7.93781996e-01 2.65693456e-01
2.92134076e-01 2.95128971e-01 1.03698289e+00 -8.91620398e-01
7.76616395e-01 1.49866629e+00 8.25346291e-01 7.60602355e-01
-1.21797764e+00 -4.84445900e-01 4.84666258e-01 3.52431417e-01
-1.19831312e+00 -4.26451713e-01 7.03228176e-01 -5.17867267e-01
8.33587289e-01 2.94838488e-01 1.57456532e-01 1.42413664e+00
-5.77202775e-02 1.06288099e+00 1.01722836e+00 -4.95865673e-01
-7.17279091e-02 1.84684411e-01 2.79440016e-01 5.35148203e-01
3.03143054e-01 2.36016974e-01 -5.87753057e-01 2.87239403e-02
6.37827158e-01 -2.24347234e-01 -3.88912499e-01 -9.86462712e-01
-1.44103205e+00 5.64897120e-01 6.84375703e-01 1.61975935e-01
-3.20967048e-01 4.49860066e-01 3.44260961e-01 9.37162116e-02
3.92600119e-01 5.39705873e-01 -9.60958719e-01 -8.17784742e-02
-5.46317816e-01 7.44067848e-01 8.45901251e-01 9.99847829e-01
3.94635111e-01 -3.90568167e-01 -2.89343357e-01 8.18628788e-01
2.74260044e-01 3.49651188e-01 1.39316872e-01 -1.07283092e+00
7.72496641e-01 5.96636161e-02 3.81225318e-01 -5.78547180e-01
-9.13493216e-01 -5.19118130e-01 -2.01132506e-01 3.87810916e-01
8.64539802e-01 -1.08075082e-01 -8.71057451e-01 1.69158256e+00
3.37092757e-01 -8.90463442e-02 -3.42038572e-01 1.10980761e+00
6.31642103e-01 1.72720566e-01 1.72640253e-02 -8.22977349e-02
1.11675644e+00 -1.20993018e+00 -4.85206127e-01 -4.80138719e-01
4.45743233e-01 -6.09252572e-01 1.18890083e+00 5.25665343e-01
-9.07252610e-01 -5.38861990e-01 -9.11775470e-01 -2.11154018e-02
-1.59752533e-01 1.61925122e-01 7.18403876e-01 6.35893226e-01
-5.73039830e-01 7.46046305e-01 -1.00896740e+00 -6.67886734e-01
5.22577643e-01 6.49805903e-01 -1.62272379e-01 9.61733684e-02
-5.68656862e-01 1.38189209e+00 3.17775846e-01 2.14454919e-01
-7.05753148e-01 -6.73457086e-01 -5.70440173e-01 -3.22339565e-01
6.11566246e-01 -6.69341445e-01 1.54984248e+00 -6.98754072e-01
-1.39631093e+00 1.00550663e+00 1.25697717e-01 -2.52616048e-01
7.72595763e-01 -4.87142593e-01 1.36816233e-01 -3.22141856e-01
-7.53566325e-02 8.82298708e-01 6.99739695e-01 -1.55772817e+00
-7.19071925e-01 -1.00993764e+00 -6.86232746e-02 2.68343985e-01
1.11855514e-01 -1.46533117e-01 -5.43321192e-01 -6.28230453e-01
2.15303436e-01 -1.32764924e+00 -1.91780940e-01 -2.29086597e-02
-1.86914802e-01 -3.41681093e-01 6.44793391e-01 -5.52243590e-01
6.51358247e-01 -1.82980824e+00 5.04665196e-01 8.91491957e-03
3.30951586e-02 -9.85044055e-03 -3.28788966e-01 8.04249421e-02
-1.10982172e-01 -2.62832671e-01 6.84070587e-02 -4.27890092e-01
8.25985745e-02 -3.27551402e-02 -2.78099716e-01 6.71584129e-01
-6.83461502e-02 1.00830007e+00 -7.07958817e-01 -3.04042578e-01
1.49676651e-01 3.36912036e-01 -6.80579603e-01 8.29839110e-02
-4.70241517e-01 7.52763093e-01 -2.04561025e-01 6.01118326e-01
4.59661573e-01 -1.49903476e-01 6.53359070e-02 -4.53784376e-01
2.52805799e-01 4.08557504e-02 -1.04649580e+00 1.86019814e+00
-4.22119677e-01 8.01343799e-01 -1.19885784e-02 -8.52433681e-01
4.16533798e-01 9.27765518e-02 5.87382793e-01 -3.89002562e-01
2.46539384e-01 1.29070476e-01 3.17572087e-01 -6.47254348e-01
3.94990355e-01 -2.28723228e-01 1.10373512e-01 1.45191476e-01
2.95333415e-01 -1.76842585e-01 -1.01190157e-01 -3.35563064e-01
8.81932199e-01 4.55942243e-01 1.21487543e-01 -1.22113623e-01
1.37409344e-01 6.32592440e-02 -4.17372398e-02 8.66713226e-01
-3.21138054e-01 8.49713445e-01 3.83043401e-02 -5.53391635e-01
-9.36687827e-01 -1.14871895e+00 -1.79775044e-01 1.63173068e+00
2.97556520e-01 4.88695279e-02 -5.03509283e-01 -8.66828084e-01
2.11477220e-01 6.35544956e-01 -6.15653157e-01 3.57868895e-02
-8.99379969e-01 -8.85876060e-01 8.19540322e-02 1.02678227e+00
-2.65610237e-02 -1.28604746e+00 -5.66416144e-01 9.49766785e-02
-1.59900665e-01 -1.12773693e+00 -3.53826016e-01 3.03677082e-01
-9.82234836e-01 -1.10014820e+00 -9.44783986e-01 -7.60197520e-01
3.64127755e-01 9.68771130e-02 1.15058732e+00 -2.29074091e-01
-4.82029885e-01 7.74472892e-01 -1.51826337e-01 -7.65388370e-01
-1.87745944e-01 5.25166810e-01 1.89935267e-01 -4.54462051e-01
2.22262919e-01 -4.92714673e-01 -5.28434098e-01 4.50545907e-01
-2.83818215e-01 -3.26077163e-01 6.08793378e-01 7.94156194e-01
1.68885231e-01 -5.08607447e-01 4.23436224e-01 -7.01125205e-01
4.71819252e-01 -1.62461191e-01 -6.14516973e-01 2.94347078e-01
-3.99936706e-01 2.96896189e-01 -5.69144869e-03 -6.98512197e-01
-8.87898743e-01 4.88681912e-01 3.11715640e-02 -3.79572451e-01
-5.50440401e-02 2.43779078e-01 -2.25405157e-01 -2.02971965e-01
9.86724913e-01 -2.97130436e-01 -3.65545750e-02 -6.28788114e-01
4.16412532e-01 5.76532483e-01 7.18340158e-01 -7.56472051e-01
6.69297576e-01 5.56471407e-01 5.37824668e-02 -6.43717349e-01
-1.08392739e+00 -5.67361951e-01 -8.99381220e-01 -6.79185018e-02
7.02504277e-01 -7.75256932e-01 -9.66261864e-01 1.83995917e-01
-1.32289302e+00 -6.53332353e-01 -1.38562381e-01 6.57797456e-01
-1.03621364e+00 1.59567744e-01 -1.95111766e-01 -7.18437672e-01
-1.89657025e-02 -1.16584551e+00 1.31003642e+00 -1.99834123e-01
-4.23195004e-01 -7.23354220e-01 2.43162319e-01 4.68084186e-01
-1.16115268e-02 -3.29295471e-02 5.97554803e-01 -6.83699489e-01
-6.08676493e-01 -2.37111360e-01 -1.82380781e-01 1.30828634e-01
9.52718332e-02 -3.79580587e-01 -1.04771614e+00 -5.75850308e-01
-1.69302717e-01 -4.46411848e-01 9.57335174e-01 5.70548773e-01
1.31685519e+00 -4.61250357e-02 -5.25118470e-01 3.89037281e-01
9.53360200e-01 -1.05483018e-01 4.84509528e-01 4.81180906e-01
8.69964421e-01 7.64638603e-01 6.12204790e-01 2.69488186e-01
4.53066409e-01 1.19707167e+00 4.21465188e-01 9.30479616e-02
-2.58448899e-01 9.09403712e-02 6.82372525e-02 7.88898245e-02
-4.93874341e-01 -2.29207888e-01 -9.85715449e-01 4.34632599e-01
-1.97820723e+00 -9.46684897e-01 1.29025161e-01 2.20875192e+00
6.05675757e-01 1.28609017e-01 5.78252673e-01 1.44175678e-01
6.44248843e-01 -1.14400320e-01 -6.68892324e-01 2.83393919e-01
3.15367550e-01 1.58987686e-01 6.73954606e-01 4.10937488e-01
-1.18040419e+00 8.49628448e-01 6.81932163e+00 3.78849328e-01
-1.12404656e+00 2.85247713e-01 1.59915835e-01 -2.69652665e-01
3.64472032e-01 -2.94860631e-01 -9.03412580e-01 4.42945845e-02
3.91942352e-01 4.61389691e-01 6.11642241e-01 9.01306510e-01
-1.78403199e-01 -1.95740256e-04 -1.87566173e+00 1.31429732e+00
2.58051097e-01 -8.20453823e-01 -3.34044337e-01 -2.53363885e-02
6.17808938e-01 3.67346704e-01 1.43662497e-01 5.74428201e-01
1.84950709e-01 -1.01674974e+00 7.75837660e-01 3.73794854e-01
4.22263354e-01 -3.37538242e-01 5.62550783e-01 4.35381383e-01
-7.35050857e-01 -2.20854267e-01 -7.04978481e-02 1.93670318e-02
2.11146139e-02 -2.07618207e-01 -1.06041682e+00 1.78527981e-01
6.70409441e-01 3.34581077e-01 -6.78387284e-01 1.11231148e+00
1.06849469e-01 2.29724064e-01 -3.50754082e-01 -2.98752217e-03
-1.79321378e-01 4.05369014e-01 6.83296621e-01 1.02211010e+00
-1.45283431e-01 -2.08556503e-01 3.35760981e-01 5.94691336e-01
7.88621381e-02 -1.75073817e-01 -2.56142080e-01 2.57332593e-01
1.38606459e-01 9.65058625e-01 -7.62226760e-01 2.08227336e-02
-3.10502052e-01 9.62998390e-01 5.02545118e-01 3.84868205e-01
-7.45430768e-01 -2.34905034e-02 3.50966275e-01 2.55864024e-01
3.57315630e-01 -2.90251672e-01 -4.93378133e-01 -1.07888162e+00
3.53615969e-01 -9.08211350e-01 4.03152943e-01 -7.52835572e-01
-1.47350311e+00 2.84289002e-01 3.96778136e-01 -9.77935612e-01
-4.94153917e-01 -1.05916142e+00 -3.11281700e-02 4.29934531e-01
-1.04201591e+00 -1.27885580e+00 -4.27932262e-01 2.80004323e-01
7.33960569e-01 -2.43884236e-01 6.43953264e-01 2.07654107e-02
-2.21887112e-01 6.83634043e-01 -7.25715831e-02 8.56323019e-02
9.31349993e-01 -1.20027149e+00 2.92563111e-01 1.82400107e-01
5.46740055e-01 4.81194645e-01 1.08193147e+00 -4.29223746e-01
-1.46757948e+00 -7.30058908e-01 1.28714040e-01 -1.25369561e+00
3.42923611e-01 -3.74002457e-01 -6.99603140e-01 1.03397107e+00
-5.77268265e-02 3.72978330e-01 1.39321461e-01 6.84771419e-01
-3.67535293e-01 -1.53539807e-01 -1.12790108e+00 5.30737936e-01
1.31845176e+00 -2.64215052e-01 -6.06932998e-01 6.14911735e-01
3.90768051e-01 -8.94195199e-01 -5.27097642e-01 8.26318860e-01
1.09369159e+00 -7.85113752e-01 1.17136514e+00 -9.14594233e-01
6.57200366e-02 -7.26686642e-02 -2.89852113e-01 -1.16423702e+00
-1.80662543e-01 -3.09635371e-01 -2.45292380e-01 7.45919645e-01
4.18173403e-01 -2.74570346e-01 1.04724526e+00 4.82025146e-01
1.00851491e-01 -7.52565145e-01 -8.11939418e-01 -1.01443958e+00
1.93877533e-01 -4.80260134e-01 2.12065354e-01 4.42539096e-01
-1.92358896e-01 5.96903980e-01 -1.94240466e-01 2.57893622e-01
6.80225015e-01 -7.33210444e-02 1.11722910e+00 -1.35817564e+00
-4.59804237e-01 -5.39404988e-01 -4.68657821e-01 -1.01371360e+00
3.73842508e-01 -7.05019295e-01 4.70003366e-01 -1.24185061e+00
4.40506071e-01 -3.00976098e-01 -1.41066372e-01 2.57135719e-01
-2.03685209e-01 4.09027100e-01 4.72697645e-01 3.04060966e-01
-7.38028467e-01 3.77818912e-01 1.24956691e+00 -2.91271150e-01
-3.47790152e-01 5.65944016e-01 -4.39256638e-01 8.56106758e-01
7.57982373e-01 -4.77098018e-01 -2.27939889e-01 -4.59500641e-01
-4.42525110e-04 -3.38047668e-02 7.12503910e-01 -1.00413740e+00
8.44801683e-03 -2.41810083e-01 4.14236724e-01 -7.18233049e-01
5.48215926e-01 -1.02459240e+00 -2.77956277e-01 3.53011191e-01
-4.42465365e-01 -1.11265458e-01 2.00516075e-01 6.19518280e-01
3.61616373e-01 -2.85637021e-01 8.27560663e-01 -2.09254280e-01
-5.72370768e-01 2.82320827e-01 9.17944908e-02 1.10096402e-01
1.01087558e+00 -2.27732614e-01 -3.47930193e-02 -4.92883682e-01
-1.06221092e+00 1.58140957e-01 2.06791207e-01 7.87926912e-01
1.47786215e-01 -1.20138228e+00 -7.34319091e-01 -2.26936564e-01
2.69352376e-01 8.67400393e-02 -1.11748658e-01 8.14836442e-01
-2.59407341e-01 4.73345578e-01 -2.23599210e-01 -9.24110472e-01
-1.35715222e+00 6.79062605e-01 5.14687598e-01 -5.10576107e-02
-3.01847994e-01 9.25650060e-01 1.95374876e-01 -6.59223616e-01
9.59625959e-01 -5.02120256e-01 7.67576545e-02 -1.77422874e-02
2.94115722e-01 6.50304496e-01 2.55032301e-01 -6.40066147e-01
-4.45498765e-01 5.76361418e-01 -1.03973165e-01 -3.71473432e-01
1.45884407e+00 -5.66255935e-02 1.67637035e-01 6.65881157e-01
1.07552731e+00 -1.65246353e-01 -1.37793612e+00 -1.66231439e-01
2.82347977e-01 -3.97070467e-01 -1.78333625e-01 -9.34899509e-01
-9.21931744e-01 8.01784754e-01 1.04613960e+00 -3.87637317e-01
5.59917867e-01 4.55031365e-01 2.12278485e-01 9.36855912e-01
5.96976161e-01 -9.62383330e-01 4.79785591e-01 4.17483777e-01
1.43541789e+00 -1.55100965e+00 2.72353142e-01 -2.70254046e-01
-6.12497628e-01 1.03760922e+00 9.75940466e-01 -9.77375731e-02
5.73292196e-01 3.25028837e-01 1.35124326e-01 -1.81594163e-01
-4.21753377e-01 -2.98824281e-01 5.72899103e-01 7.83018231e-01
5.11344314e-01 -5.78093380e-02 6.82162959e-03 5.44567287e-01
-2.43752033e-01 6.47927523e-02 -4.66817729e-02 9.42409217e-01
-2.18120471e-01 -1.00856709e+00 -6.68959260e-01 3.71931434e-01
-2.81646699e-01 3.69770497e-01 -4.51289564e-01 1.01620495e+00
6.41049221e-02 5.37596822e-01 1.62792224e-02 -2.12034881e-01
6.14436507e-01 6.27007708e-02 1.50068033e+00 -8.23785961e-01
-4.50263977e-01 5.00360467e-02 -2.22103626e-01 -4.77579206e-01
-5.74953377e-01 -8.48714888e-01 -1.05026734e+00 1.64539441e-01
-6.86341822e-01 -4.08776790e-01 7.20020890e-01 1.05104303e+00
-5.12985550e-02 4.26919013e-01 9.60253999e-02 -1.56911087e+00
-1.14646494e+00 -1.25371540e+00 -3.15963477e-01 2.96101391e-01
5.15606403e-01 -1.06121945e+00 -9.17557254e-02 -6.78233877e-02] | [6.917627334594727, -0.8929163217544556] |
6b9152ca-75aa-4e9d-8b2f-13b78c7ffef4 | inspherenet-a-concise-representation-and | 1912.11606 | null | https://arxiv.org/abs/1912.11606v2 | https://arxiv.org/pdf/1912.11606v2.pdf | InSphereNet: a Concise Representation and Classification Method for 3D Object | In this paper, we present an InSphereNet method for the problem of 3D object classification. Unlike previous methods that use points, voxels, or multi-view images as inputs of deep neural network (DNN), the proposed method constructs a class of more representative features named infilling spheres from signed distance field (SDF). Because of the admirable spatial representation of infilling spheres, we can not only utilize very fewer number of spheres to accomplish classification task, but also design a lightweight InSphereNet with less layers and parameters than previous methods. Experiments on ModelNet40 show that the proposed method leads to superior performance than PointNet and PointNet++ in accuracy. In particular, if there are only a few dozen sphere inputs or about 100000 DNN parameters, the accuracy of our method remains at a very high level (over 88%). This further validates the conciseness and effectiveness of the proposed InSphere 3D representation. Keywords: 3D object classification , signed distance field , deep learning , infilling sphere | ['Siyu Zhang', 'Shen Cai', 'Haikuan Du', 'Hui Cao'] | 2019-12-25 | null | null | null | null | ['3d-object-classification'] | ['computer-vision'] | [-5.10544837e-01 -1.24626331e-01 3.54869962e-01 -4.74051714e-01
-2.26199664e-02 -4.02374506e-01 7.54720688e-01 1.57827482e-01
-4.24643427e-01 6.79765165e-01 -2.36412615e-01 -1.54293254e-01
-1.84355840e-01 -1.07325852e+00 -7.94830918e-01 -4.83740032e-01
-2.42947951e-01 4.37082469e-01 5.42093515e-01 -1.81230843e-01
3.36736858e-01 1.21403968e+00 -1.65824068e+00 -1.40582055e-01
7.45339155e-01 1.60358560e+00 1.50348425e-01 1.32788137e-01
-2.99823672e-01 5.27021468e-01 -7.20341861e-01 -2.48177111e-01
4.48940843e-01 2.79495537e-01 -5.21753907e-01 -2.50560969e-01
6.62066579e-01 -5.61522424e-01 -3.35042983e-01 9.76030171e-01
5.89652896e-01 2.88023025e-01 7.45728910e-01 -1.19032216e+00
-5.81529021e-01 2.71378875e-01 -6.00329161e-01 2.33748734e-01
2.29004905e-01 -1.79847568e-01 8.71255279e-01 -1.25633895e+00
4.21128988e-01 1.22443712e+00 8.21842074e-01 1.84722394e-01
-6.44193411e-01 -7.24675775e-01 2.29973465e-01 7.07266554e-02
-1.50908875e+00 1.01397529e-01 7.37605691e-01 -3.47033381e-01
1.15818501e+00 3.67419302e-01 1.12947583e+00 6.26146674e-01
3.91902685e-01 6.77424431e-01 8.58066142e-01 -7.24623948e-02
3.05893928e-01 3.97465155e-02 4.23519999e-01 8.06516051e-01
3.60060006e-01 3.10691558e-02 -2.10335121e-01 -5.87184727e-02
1.32894230e+00 4.56601471e-01 -1.31075472e-01 -6.07116818e-01
-1.29959631e+00 7.84649432e-01 1.02855659e+00 2.62390852e-01
-3.03580582e-01 2.57799029e-01 3.91924739e-01 -4.79621626e-02
7.08464622e-01 2.27740571e-01 -2.97898233e-01 1.42138794e-01
-5.80256045e-01 3.89795780e-01 5.33501923e-01 1.09836161e+00
6.94021702e-01 1.81907326e-01 1.62191272e-01 8.12501311e-01
1.63265422e-01 7.25699782e-01 2.00227216e-01 -7.00222850e-01
3.61145824e-01 8.61280680e-01 2.71447189e-02 -1.32096970e+00
-6.84258223e-01 -8.27936113e-01 -1.17685270e+00 5.63240945e-01
2.45660722e-01 3.10220569e-01 -9.25742686e-01 1.16460967e+00
5.09617031e-01 -3.23818773e-02 -1.91719994e-01 1.06924033e+00
1.22227764e+00 6.36902153e-01 -3.25308412e-01 5.31509578e-01
1.13406706e+00 -7.46916652e-01 -1.12144336e-01 1.89219400e-01
4.75912511e-01 -4.18970197e-01 8.37938190e-01 3.98249239e-01
-1.28478098e+00 -7.73106158e-01 -1.21044815e+00 -1.66734636e-01
-5.10380685e-01 1.09455036e-02 9.34156060e-01 3.96868438e-01
-1.14896274e+00 8.53678644e-01 -7.84619272e-01 -1.79521173e-01
7.79016137e-01 5.35876036e-01 -4.57123548e-01 6.32054433e-02
-9.40126657e-01 8.30980897e-01 1.92446113e-01 1.31856650e-01
-6.38040841e-01 -1.00957716e+00 -8.25319767e-01 2.12380201e-01
-2.18897671e-01 -7.05384016e-01 1.01180363e+00 -2.18345016e-01
-1.17034042e+00 6.89898610e-01 1.72587540e-02 -3.17747205e-01
5.77123463e-01 -2.78927475e-01 1.91361513e-02 2.81056136e-01
1.35270432e-01 7.39759982e-01 4.81664479e-01 -1.19896293e+00
-6.62091076e-01 -7.62084365e-01 3.64838839e-01 3.18207443e-01
-3.07454407e-01 -2.46567383e-01 -1.57203078e-01 -4.31589127e-01
5.65695941e-01 -5.05123377e-01 -2.51926094e-01 5.67161024e-01
-3.72561127e-01 -7.70717680e-01 7.57750511e-01 -1.87770754e-01
4.98297244e-01 -2.29074454e+00 -1.53866053e-01 3.90214920e-01
8.35655749e-01 5.36124930e-02 7.53896683e-02 -7.62418211e-02
-9.66338590e-02 6.13809424e-03 -7.89662972e-02 -2.79354244e-01
5.33146113e-02 1.00980349e-01 -1.89115163e-02 7.99613893e-01
6.07227199e-02 8.28569233e-01 -8.99451137e-01 -3.81484121e-01
5.41982591e-01 6.60690665e-01 -4.20312047e-01 -1.62461311e-01
1.23334177e-01 2.69640740e-02 -6.89629078e-01 6.33824527e-01
1.17992461e+00 -3.96146655e-01 -7.68099368e-01 -3.72777790e-01
-1.79620802e-01 1.62085831e-01 -1.29577816e+00 1.67289925e+00
-4.78894025e-01 4.87117887e-01 -1.80175528e-01 -1.01839590e+00
1.40422595e+00 2.43424177e-02 7.02214420e-01 -7.57032275e-01
2.56251276e-01 3.36233020e-01 -2.61409283e-01 -2.35470328e-02
2.42249027e-01 -1.54888138e-01 2.18605131e-01 2.32846051e-01
2.44410802e-02 -3.31116676e-01 -2.13841777e-02 1.18849084e-01
9.55705822e-01 -2.91199386e-02 1.87518716e-01 -4.50443566e-01
6.22891784e-01 -1.85756877e-01 3.69646251e-01 5.31784892e-01
-2.15785131e-01 7.30499268e-01 3.47092539e-01 -9.33595777e-01
-1.00944173e+00 -1.35107517e+00 -5.92386603e-01 6.20183885e-01
5.44029713e-01 -8.69779214e-02 -3.67439747e-01 -5.76653421e-01
4.58116680e-01 2.91213632e-01 -6.27095163e-01 -6.95487310e-04
-5.42132318e-01 -3.99956077e-01 4.24289107e-01 9.19392526e-01
9.51514006e-01 -8.13754678e-01 -5.61002135e-01 8.65099579e-02
4.32489961e-01 -1.01898134e+00 -1.48364946e-01 1.84277296e-02
-1.23432577e+00 -1.02347052e+00 -1.10093939e+00 -1.00548017e+00
7.61479974e-01 5.48249900e-01 1.11983275e+00 -2.30293311e-02
-1.25574261e-01 1.35095045e-01 -2.01292738e-01 -4.30931121e-01
2.15924412e-01 -7.50294700e-02 1.86838523e-01 -3.61438632e-01
4.31110799e-01 -8.82665455e-01 -8.52707267e-01 2.31587842e-01
-5.97595513e-01 -4.93679456e-02 3.90943199e-01 4.87708658e-01
5.44935167e-01 -1.11396141e-01 4.74296957e-01 -5.50206721e-01
4.01939422e-01 -3.07679445e-01 -7.43930757e-01 -1.42658263e-01
-2.66527236e-01 -2.86685705e-01 8.25423539e-01 -1.28656834e-01
-6.30170763e-01 -9.34676602e-02 -3.37801069e-01 -7.26526499e-01
-2.10850969e-01 7.16499463e-02 5.63021749e-02 -4.48629797e-01
6.48709536e-01 3.39135587e-01 2.08192021e-01 -5.79936266e-01
3.38133365e-01 4.61308569e-01 2.13623479e-01 -3.52144897e-01
6.75134897e-01 8.54584873e-01 2.07891285e-01 -7.12412059e-01
-7.19965756e-01 -3.17994624e-01 -7.95486152e-01 -2.44769976e-01
6.59873366e-01 -9.03991401e-01 -1.12296116e+00 5.38902223e-01
-1.42123544e+00 8.27704072e-02 -3.92150819e-01 6.44260764e-01
-4.87618476e-01 2.13789225e-01 -5.39082825e-01 -5.96491873e-01
-4.67034787e-01 -1.12753618e+00 1.14221609e+00 6.72127903e-02
8.62190649e-02 -9.64257777e-01 -2.82099873e-01 7.64883533e-02
3.72460783e-01 4.71413195e-01 7.80828238e-01 -6.56445205e-01
-7.25333273e-01 -3.04835320e-01 -6.50685370e-01 4.07555968e-01
1.85693279e-01 -2.73359895e-01 -9.22516644e-01 -1.82861865e-01
2.10061982e-01 -1.15547992e-01 7.40092337e-01 4.46284801e-01
1.54026365e+00 -2.98802368e-02 -2.76678920e-01 9.39371824e-01
1.42599189e+00 3.78173888e-01 5.47014356e-01 4.13570553e-01
6.74016356e-01 2.88712412e-01 3.56894433e-01 5.31234443e-01
4.20725375e-01 1.74191713e-01 7.44749248e-01 -4.19630647e-01
-1.45896375e-01 1.18521072e-01 -2.70548791e-01 7.57577717e-01
-3.22317719e-01 6.75447807e-02 -8.80453706e-01 3.58830035e-01
-1.41474938e+00 -7.71733820e-01 -2.93548793e-01 2.04838824e+00
3.02680284e-01 3.59781921e-01 7.35222399e-02 3.44436288e-01
6.19740427e-01 7.89077505e-02 -7.73195267e-01 -1.67600483e-01
-1.10222235e-01 3.08570385e-01 3.91573519e-01 3.05385143e-01
-1.02925336e+00 6.43042982e-01 5.68516970e+00 8.27588856e-01
-1.17536187e+00 -1.16198152e-01 4.94036406e-01 -1.34118512e-01
-1.75415531e-01 -4.81023937e-01 -9.21816587e-01 3.74877065e-01
1.39976099e-01 5.60990116e-03 3.93263213e-02 1.06124341e+00
2.07573664e-03 2.96957828e-02 -1.17278957e+00 1.26383948e+00
1.67216942e-01 -1.54317212e+00 2.55306542e-01 -8.97670258e-03
5.23516059e-01 3.95485312e-01 4.97312956e-02 1.35465980e-01
4.16488275e-02 -1.05155683e+00 8.02035511e-01 3.32108289e-01
7.22638309e-01 -8.28829527e-01 7.65595913e-01 2.95436472e-01
-1.19581771e+00 1.33544892e-01 -9.03542042e-01 -7.17020333e-02
-1.00601435e-01 9.01687503e-01 -6.54131949e-01 5.55367768e-01
9.89890158e-01 1.08426380e+00 -3.92355114e-01 1.41981661e+00
1.79774329e-01 -1.05068740e-02 -7.67297626e-01 -4.95852143e-01
5.60585022e-01 -2.72679299e-01 4.76463109e-01 8.71039510e-01
4.21889454e-01 2.26875052e-01 -6.78342357e-02 1.02967763e+00
2.39230786e-02 1.02801614e-01 -8.63238454e-01 6.03407502e-01
4.27112639e-01 1.15598810e+00 -8.51234913e-01 -3.65760744e-01
-4.53885019e-01 6.93348110e-01 3.22583824e-01 2.54877090e-01
-7.13562131e-01 -7.06463575e-01 6.41332686e-01 2.45025545e-01
4.35637623e-01 -4.78465647e-01 -5.97111285e-01 -8.07851613e-01
2.78710455e-01 -2.91788876e-01 2.68794876e-02 -9.81986701e-01
-1.35301983e+00 8.65674973e-01 -7.76895657e-02 -1.42612600e+00
3.47568601e-01 -1.01601696e+00 -6.28366470e-01 8.43634605e-01
-1.58375156e+00 -9.65472877e-01 -6.21257186e-01 7.17034519e-01
2.86601484e-01 -2.07682937e-01 4.69178259e-01 2.66961962e-01
-3.90894338e-02 3.49274397e-01 2.91613877e-01 2.31496334e-01
8.63780826e-02 -1.32475793e+00 5.27433991e-01 3.72113213e-02
-9.25609469e-02 5.80147922e-01 1.81435809e-01 -4.40971196e-01
-1.13635004e+00 -1.04587984e+00 4.98723537e-01 -2.89957225e-01
4.29563284e-01 -5.05999207e-01 -8.14711154e-01 4.06992108e-01
-2.00013548e-01 2.83438236e-01 4.92433310e-01 -3.91550288e-02
-2.61660248e-01 -4.37680274e-01 -1.34615362e+00 2.99992591e-01
1.30213308e+00 -1.53866380e-01 -6.61326647e-01 6.01739109e-01
7.16392338e-01 -6.21619284e-01 -1.01880586e+00 4.82413620e-01
3.86272281e-01 -1.21678340e+00 1.36077476e+00 -4.10483062e-01
3.73356789e-01 -1.62058249e-01 -2.21665367e-01 -1.27722490e+00
-4.23782051e-01 -1.50753809e-02 -5.12663759e-02 7.27260768e-01
1.67536914e-01 -1.00190032e+00 9.91297185e-01 2.53586024e-01
-5.23206472e-01 -1.19943213e+00 -1.14953148e+00 -9.99199629e-01
3.48837942e-01 -2.95345992e-01 1.03379834e+00 7.62010455e-01
-3.55152369e-01 1.31292731e-01 3.57270449e-01 1.65618107e-01
8.31584334e-01 2.38612756e-01 6.37002826e-01 -1.67192173e+00
2.23529905e-01 -5.43754637e-01 -1.02734077e+00 -1.41614985e+00
-6.94586779e-04 -1.11000085e+00 -3.30856711e-01 -1.81669438e+00
6.78082407e-02 -9.86349404e-01 -3.30251396e-01 3.74259323e-01
2.42648587e-01 2.98164576e-01 1.35714754e-01 9.56574380e-02
-5.10287821e-01 9.20397520e-01 1.78123021e+00 -4.89965156e-02
2.99167745e-02 1.14506997e-01 -4.74315345e-01 8.96311820e-01
7.46469498e-01 -9.31278616e-02 -2.68770725e-01 -6.88023269e-01
-6.38571056e-03 -1.99205920e-01 6.72597349e-01 -1.25549030e+00
2.51013339e-01 1.19569622e-01 9.21820521e-01 -1.18850303e+00
6.91811204e-01 -8.46772373e-01 -2.50413150e-01 3.93641859e-01
2.17918754e-01 1.06239989e-01 2.24175259e-01 3.23012084e-01
-9.48155671e-02 -6.96678609e-02 7.55609393e-01 -3.78573388e-01
-4.93204981e-01 6.68825388e-01 7.97814783e-03 -1.35696650e-01
1.07092834e+00 -7.42794394e-01 -2.99519539e-01 -1.18120745e-01
-7.55255520e-01 2.29293227e-01 3.22230071e-01 1.85742661e-01
1.01741183e+00 -1.58332443e+00 -5.13745844e-01 4.66214865e-01
-1.82278752e-01 6.89533710e-01 3.82716209e-02 5.20782232e-01
-1.01307237e+00 5.87946534e-01 -2.92560369e-01 -1.03197217e+00
-7.38906980e-01 2.80439794e-01 5.02006829e-01 2.73442358e-01
-1.02543867e+00 1.06827867e+00 5.05259931e-01 -5.95891714e-01
5.12411237e-01 -6.13144398e-01 -1.73062533e-01 -1.92048162e-01
2.54913300e-01 5.80329001e-01 3.70922744e-01 -3.96000177e-01
-5.05634844e-01 8.15494478e-01 -5.62233031e-02 3.40936333e-01
1.79353952e+00 2.83867925e-01 -8.55889171e-02 2.45826498e-01
1.48381126e+00 -3.51987988e-01 -1.22616529e+00 -5.34624532e-02
-4.68990952e-01 -6.62551641e-01 1.89875573e-01 -3.64181548e-01
-1.28278768e+00 1.16940522e+00 5.57281196e-01 3.13698500e-01
6.95807338e-01 1.88443735e-01 8.70777130e-01 5.42569935e-01
4.97513533e-01 -7.04228997e-01 1.00654364e-01 5.93262911e-01
1.20052469e+00 -1.04968178e+00 6.70307651e-02 -4.73171264e-01
-6.94538802e-02 1.14803445e+00 9.63687420e-01 -6.88172281e-01
9.60276008e-01 1.19616330e-01 -1.27787605e-01 -5.35655916e-01
-1.65518016e-01 5.73870055e-02 2.09874794e-01 6.03958905e-01
2.06166089e-01 -2.62413323e-01 8.08183178e-02 3.67363095e-01
-5.49591005e-01 -2.01257855e-01 2.03983650e-01 8.59363914e-01
-6.69542670e-01 -4.10412192e-01 -3.85391951e-01 7.13033080e-01
-4.79669608e-02 2.20327191e-02 -1.06847428e-01 1.05491292e+00
4.98817600e-02 3.46566767e-01 6.92688465e-01 -2.10167974e-01
6.43161416e-01 -4.03778732e-01 5.80408454e-01 -4.52287674e-01
-3.46017182e-01 -2.94198394e-01 -4.17478889e-01 -6.14152074e-01
-2.58808911e-01 -3.33906502e-01 -1.49126208e+00 -5.11205256e-01
-3.99324775e-01 7.60810971e-02 7.59816825e-01 7.01090157e-01
3.01274955e-01 3.33649546e-01 8.28880429e-01 -1.16882670e+00
-7.00031996e-01 -7.54634559e-01 -9.01293576e-01 2.48465255e-01
3.00582916e-01 -1.06002808e+00 -4.24770772e-01 -5.20663738e-01] | [7.953796863555908, -3.5850234031677246] |
e7b07aea-814a-4920-ada5-1613d8c7502d | nfi-2-learning-noise-free-illuminance | 2305.10223 | null | https://arxiv.org/abs/2305.10223v1 | https://arxiv.org/pdf/2305.10223v1.pdf | NFI$_2$: Learning Noise-Free Illuminance-Interpolator for Unsupervised Low-Light Image Enhancement | Low-light situations severely restrict the pursuit of aesthetic quality in consumer photography. Although many efforts are devoted to designing heuristics, it is generally mired in a shallow spiral of tedium, such as piling up complex network architectures and empirical strategies. How to delve into the essential physical principles of illumination compensation has been neglected. Following the way of simplifying the complexity, this paper innovatively proposes a simple and efficient Noise-Free Illumination Interpolator (NFI$_2$). According to the constraint principle of illuminance and reflectance within a limited dynamic range, as a prior knowledge in the recovery process, we construct a learnable illuminance interpolator and thereby compensating for non-uniform lighting. With the intention of adapting denoising without annotated data, we design a self-calibrated denoiser with the intrinsic image properties to acquire noise-free low-light images. Starting from the properties of natural image manifolds, a self-regularized recovery loss is introduced as a way to encourage more natural and realistic reflectance map. The model architecture and training losses, guided by prior knowledge, complement and benefit each other, forming a powerful unsupervised leaning framework. Comprehensive experiments demonstrate that the proposed algorithm produces competitive qualitative and quantitative results while maintaining favorable generalization capability in unknown real-world scenarios. | ['Risheng Liu', 'Xin Fan', 'Ziyu Yue', 'Jiaxin Gao', 'Xiaofeng Liu'] | 2023-05-17 | null | null | null | null | ['image-enhancement', 'low-light-image-enhancement'] | ['computer-vision', 'computer-vision'] | [ 4.33326393e-01 -3.48948315e-02 4.46729958e-02 -4.51786935e-01
-8.70621204e-02 2.46981438e-02 2.44540259e-01 -5.72382510e-01
-3.37353528e-01 7.03629255e-01 3.05403639e-02 4.66152355e-02
-3.47192019e-01 -8.89028788e-01 -6.73650324e-01 -1.13139200e+00
4.14649189e-01 -3.21004480e-01 -3.63037288e-01 -3.98428559e-01
1.95789948e-01 4.02447701e-01 -1.67270315e+00 -2.80678153e-01
1.38838756e+00 9.27326202e-01 3.98364425e-01 1.54537618e-01
-9.69845727e-02 6.30110383e-01 -3.01017910e-01 -7.31261671e-01
5.99442542e-01 -6.08404279e-01 -4.21153545e-01 5.97659945e-01
2.17642650e-01 -2.74158001e-01 -3.68580669e-01 1.36995423e+00
5.10693908e-01 3.85723472e-01 4.40309733e-01 -7.70836115e-01
-9.75413263e-01 2.03025058e-01 -4.93347168e-01 -2.93557972e-01
1.06081471e-01 5.21009207e-01 7.64305949e-01 -6.02557838e-01
4.23386663e-01 8.69469643e-01 5.86948097e-01 5.67146242e-01
-1.17273080e+00 -3.24435174e-01 1.95906505e-01 2.22764388e-01
-1.41038573e+00 -4.41022366e-01 1.36620498e+00 1.42778484e-02
3.65557641e-01 4.32886243e-01 8.34900320e-01 9.20635939e-01
1.54977009e-01 5.10596395e-01 1.29503810e+00 -5.65904617e-01
1.50940016e-01 4.14160818e-01 -3.13404322e-01 6.97611392e-01
2.68686533e-01 9.79838446e-02 -3.99601758e-01 3.91192734e-01
9.13430095e-01 -3.10156122e-02 -6.64901018e-01 -4.28381711e-01
-8.01517665e-01 4.06202585e-01 6.98212624e-01 1.69752166e-01
-4.11704421e-01 -2.47892499e-01 5.88263348e-02 3.13090652e-01
5.01142442e-01 5.22565603e-01 2.01249011e-02 3.23000938e-01
-6.33861840e-01 -1.95677370e-01 4.70597148e-01 8.23951364e-01
1.03631759e+00 3.72943103e-01 1.11830004e-01 1.09555531e+00
2.61636853e-01 5.39554298e-01 2.51721025e-01 -1.13641763e+00
7.97507167e-02 7.06327736e-01 4.03502256e-01 -1.30674720e+00
-1.78388432e-01 -8.93377900e-01 -1.24074495e+00 5.19455373e-01
3.51352513e-01 5.84563836e-02 -7.22880065e-01 1.58969903e+00
4.49478984e-01 -1.57102257e-01 -8.37228596e-02 1.39300644e+00
2.92671025e-01 6.14715874e-01 -6.51542991e-02 -6.19143546e-01
9.79682088e-01 -8.11722994e-01 -8.40893149e-01 -1.99404154e-02
5.97241223e-02 -7.56032825e-01 1.58805108e+00 7.84277797e-01
-1.18652833e+00 -6.83205366e-01 -1.02813423e+00 -6.36236817e-02
-1.18684741e-02 2.29509071e-01 8.31585705e-01 7.24340558e-01
-8.91566455e-01 6.91199422e-01 -4.48216826e-01 -2.21636429e-01
3.05720568e-01 1.11124143e-02 -1.57974958e-01 -3.32967967e-01
-1.01237273e+00 9.43625093e-01 6.70927390e-02 7.76691020e-01
-6.44329846e-01 -5.01288176e-01 -5.12473822e-01 -2.47586936e-01
4.39985007e-01 -9.01641965e-01 6.08289599e-01 -1.44393086e+00
-2.20887637e+00 7.19923139e-01 4.38680425e-02 -6.69734553e-02
6.23791277e-01 -2.22817168e-01 -5.66740215e-01 1.33984715e-01
-2.41764486e-01 3.00370157e-01 1.20493579e+00 -1.62332547e+00
-2.79735506e-01 -6.38844445e-02 8.45726579e-02 5.57709813e-01
-5.73164344e-01 -2.34028220e-01 -4.54502881e-01 -5.00938058e-01
1.50530159e-01 -5.25862992e-01 -3.61925036e-01 2.00753316e-01
-3.59160632e-01 2.60826021e-01 4.01968271e-01 -6.01663053e-01
1.04576099e+00 -2.10331655e+00 1.09983474e-01 2.34648079e-01
3.85291725e-02 3.10075969e-01 -8.76700208e-02 2.02403262e-01
-8.65123495e-02 -2.52115071e-01 -4.89747554e-01 -1.31084710e-01
-7.46542364e-02 3.18302333e-01 -1.12263128e-01 7.02646315e-01
1.24869831e-01 6.37546301e-01 -1.04916167e+00 -3.02989006e-01
5.49119830e-01 7.90614784e-01 -5.54855347e-01 3.10658157e-01
-1.35900140e-01 7.25274086e-01 -3.39992702e-01 5.72452724e-01
7.78120458e-01 -2.82480903e-02 -5.56600280e-02 -6.08551502e-01
-1.88045606e-01 -6.44152611e-02 -1.32124972e+00 1.74121702e+00
-7.81568825e-01 3.76576245e-01 5.14784455e-01 -1.06659818e+00
1.23374689e+00 -1.36684954e-01 2.98585802e-01 -9.80734229e-01
3.32815140e-01 1.87842533e-01 -3.08650583e-01 -7.49391556e-01
2.62206852e-01 -4.11301762e-01 5.88192344e-01 -4.45056222e-02
-2.58444756e-01 -2.97122210e-01 -3.68008465e-02 -2.63845444e-01
5.66504717e-01 4.77955878e-01 8.60333890e-02 -2.32202008e-01
7.72551537e-01 -3.06843162e-01 6.99631870e-01 3.55140448e-01
-5.50097227e-02 5.66928685e-01 -1.23492576e-01 -5.87468326e-01
-1.11699021e+00 -1.06556916e+00 -2.04456612e-01 7.92537868e-01
4.50079411e-01 6.44622967e-02 -7.68237054e-01 -1.57301977e-01
-4.93654877e-01 6.91831172e-01 -4.04220760e-01 -2.37331912e-01
-5.52785218e-01 -9.80342865e-01 -3.81443650e-02 -5.14248833e-02
1.13754964e+00 -9.26891565e-01 -4.29873437e-01 -2.19125338e-02
-2.55734056e-01 -7.92117000e-01 -2.71520197e-01 7.87795559e-02
-6.46583855e-01 -9.79540646e-01 -6.33440673e-01 -6.56930566e-01
9.28501785e-01 4.54093516e-01 1.07066762e+00 9.99572277e-02
-4.60752517e-01 3.65635753e-01 -1.42412707e-01 -1.79310963e-01
-2.54987031e-01 -4.27705884e-01 -8.53566378e-02 5.32228529e-01
8.11166391e-02 -8.95136416e-01 -1.11229742e+00 4.46544081e-01
-9.55959022e-01 2.45334312e-01 8.22377205e-01 8.90589595e-01
6.04296148e-01 3.68273407e-01 2.78397262e-01 -5.70312858e-01
4.91420627e-01 -1.37754560e-01 -7.19072700e-01 1.30118206e-01
-9.47858155e-01 -2.05210343e-01 1.02573335e+00 -3.95582914e-01
-1.62275064e+00 6.49462640e-02 -1.68258756e-01 -3.23830754e-01
-1.98786914e-01 -5.43349702e-03 -5.30551493e-01 -4.18688685e-01
8.36439073e-01 4.84153211e-01 1.60663724e-01 -3.55398953e-01
6.60903692e-01 4.94094253e-01 5.93492389e-01 -5.13794959e-01
9.56461132e-01 7.87837327e-01 1.98530689e-01 -1.04712164e+00
-9.01865005e-01 -2.06549242e-01 -3.85289639e-01 -6.18315816e-01
6.71799421e-01 -8.36262405e-01 -8.75943601e-01 5.12678862e-01
-7.29484081e-01 -3.12783062e-01 -6.36004090e-01 4.21746582e-01
-4.91743177e-01 5.94532609e-01 -4.89649683e-01 -9.62647617e-01
-2.21203879e-01 -9.31116462e-01 4.73593324e-01 4.30693984e-01
3.48610818e-01 -9.64526951e-01 -1.09475985e-01 4.58738923e-01
6.07502580e-01 2.01360822e-01 6.87776864e-01 5.02354324e-01
-8.50929201e-01 -3.96901043e-03 -2.93312669e-01 1.04359341e+00
3.10745686e-01 -5.80286197e-02 -1.07062352e+00 4.31722309e-03
6.10319972e-01 -1.41251415e-01 7.37821281e-01 4.29711431e-01
1.31754386e+00 -4.38748121e-01 1.78083733e-01 1.08615851e+00
1.72656810e+00 8.14445168e-02 9.66357052e-01 3.93816411e-01
7.08556652e-01 9.39759016e-01 3.37226868e-01 3.73344183e-01
1.97695792e-01 2.91009605e-01 7.95837641e-01 -4.27807093e-01
-2.81876475e-01 -1.85497239e-01 1.08601436e-01 9.62016940e-01
-5.09280145e-01 -2.20088601e-01 -3.02086860e-01 4.19188917e-01
-1.50040329e+00 -8.91523480e-01 -9.66750681e-02 2.25123572e+00
9.71319795e-01 -3.19693349e-02 -1.33579612e-01 1.27533197e-01
7.48495221e-01 2.65497237e-01 -5.75813651e-01 -1.82613701e-01
-5.28724253e-01 -5.25460467e-02 5.75266719e-01 6.14328802e-01
-7.91039705e-01 6.38146758e-01 5.81538439e+00 8.21091533e-01
-1.22934592e+00 -1.26770824e-01 7.59737551e-01 6.83670714e-02
-6.56057954e-01 -3.04091722e-03 -3.97861600e-01 4.74671870e-01
5.44533491e-01 2.89912373e-01 9.18126762e-01 7.12750971e-01
5.78355849e-01 -1.02569096e-01 -4.88725036e-01 1.15047431e+00
1.62116006e-01 -9.31824327e-01 1.01649120e-01 -2.07167640e-01
7.15737343e-01 -3.46047491e-01 1.74645424e-01 -1.27047002e-01
1.72573388e-01 -8.72937143e-01 5.59023678e-01 9.09772456e-01
6.71707273e-01 -6.93091869e-01 4.26630944e-01 2.39750564e-01
-6.50199473e-01 -1.11146331e-01 -5.67870855e-01 -2.23494425e-01
1.39641657e-01 9.01661694e-01 -2.78710067e-01 7.76095688e-01
6.38838530e-01 7.51522779e-01 -1.57710299e-01 1.04282784e+00
-4.94474381e-01 4.13729310e-01 -2.03821421e-01 1.25340566e-01
2.42517609e-02 -9.48588550e-01 5.83209276e-01 8.39709461e-01
1.37711495e-01 2.17241749e-01 -1.32995456e-01 9.60909307e-01
-2.52334345e-02 3.28231514e-01 -3.94229919e-01 3.65536034e-01
4.58431765e-02 1.56882536e+00 -5.85555017e-01 4.42156494e-02
-3.80668759e-01 1.07422996e+00 1.03627749e-01 6.84098780e-01
-7.31648982e-01 -3.10401559e-01 5.00586867e-01 2.72932351e-01
2.76171099e-02 -3.67878675e-02 -5.29166043e-01 -1.20947480e+00
2.98087686e-01 -7.33647764e-01 -1.42933503e-01 -9.81828272e-01
-1.44508672e+00 4.42129314e-01 -2.74175435e-01 -1.41157413e+00
3.75593215e-01 -5.93092442e-01 -7.02435672e-01 8.83652091e-01
-2.07094216e+00 -1.07499325e+00 -7.89685190e-01 8.14526379e-01
3.65220398e-01 8.11128989e-02 5.50498605e-01 4.64449555e-01
-7.28330851e-01 3.44946653e-01 2.17175230e-01 -2.93013871e-01
7.83487678e-01 -9.92133737e-01 -3.26828420e-01 1.07120895e+00
-3.36819082e-01 7.04240024e-01 9.85534310e-01 -2.23720506e-01
-1.50582790e+00 -9.85946894e-01 2.70997077e-01 9.05682594e-02
5.12597263e-01 -2.12582752e-01 -9.18804109e-01 1.06273577e-01
3.40961784e-01 -3.02832257e-02 3.58816057e-01 -1.84698313e-01
-1.76427111e-01 -7.17622280e-01 -1.15115261e+00 8.57521296e-01
1.23074245e+00 -3.31541687e-01 -2.64166534e-01 4.27081764e-01
5.74948370e-01 -9.38019753e-02 -7.53274858e-01 2.98376054e-01
4.34651941e-01 -1.30787945e+00 1.02117169e+00 -1.02644689e-01
4.60454196e-01 -3.96097511e-01 1.52147710e-02 -1.28536987e+00
-3.70750964e-01 -1.13914442e+00 3.66820455e-01 1.30393422e+00
1.73681956e-02 -8.02135050e-01 8.13771546e-01 5.70593715e-01
-4.25916046e-01 -7.64807582e-01 -5.97751498e-01 -7.11655617e-01
-2.56620318e-01 -3.16377461e-01 3.26498717e-01 8.84531677e-01
-4.33521301e-01 4.94269021e-02 -6.20351732e-01 1.57203972e-01
1.02962339e+00 7.42723346e-02 7.62671113e-01 -9.96300101e-01
-2.40463853e-01 -5.03543079e-01 7.14336261e-02 -1.16519487e+00
-3.81224044e-02 -4.93092537e-01 1.58510149e-01 -1.39852679e+00
-5.93900420e-02 -5.06356597e-01 -3.48886311e-01 7.74236768e-02
-2.85445638e-02 3.59548479e-01 -1.15764409e-01 3.68204325e-01
-3.79856408e-01 9.30532813e-01 1.76500618e+00 -5.81927039e-03
-2.90557623e-01 -5.96343279e-02 -8.85849833e-01 1.01517761e+00
7.42193222e-01 4.72580455e-03 -7.12990046e-01 -5.16013265e-01
4.92821157e-01 -2.53499746e-01 4.29284900e-01 -8.79777670e-01
9.43166688e-02 -4.54766989e-01 3.05085003e-01 -6.49240389e-02
3.75241607e-01 -1.01457620e+00 1.77837923e-01 1.78592801e-01
-2.02564150e-01 -5.14158487e-01 -3.96214247e-01 6.51790082e-01
-2.19281480e-01 -8.85034651e-02 1.09252441e+00 -3.38830829e-01
-6.69552922e-01 1.34281456e-01 1.55181333e-01 -2.20651522e-01
7.98401773e-01 -5.45277059e-01 -2.57311255e-01 -2.94351250e-01
-4.02270555e-01 -1.81638636e-02 7.05331743e-01 4.43662480e-02
5.59791207e-01 -1.14624202e+00 -4.22651142e-01 3.64879072e-01
-1.60294250e-01 -8.48190673e-03 6.76471770e-01 7.82188058e-01
-6.61800563e-01 -7.46689597e-03 -3.94006640e-01 -3.21646988e-01
-6.87974513e-01 6.40362680e-01 5.07558644e-01 1.07776009e-01
-6.73537374e-01 7.28560984e-01 2.49755919e-01 -1.93631157e-01
1.58141285e-01 -1.64895684e-01 -7.48631358e-02 -2.84251124e-01
3.99087161e-01 4.84893858e-01 -6.24273084e-02 -3.72679502e-01
3.01639196e-02 7.25893497e-01 3.23092997e-01 2.74654359e-01
1.48392057e+00 -7.00298190e-01 -2.51983076e-01 1.82162166e-01
9.44802165e-01 1.65065229e-01 -1.59644735e+00 -1.22365892e-01
-5.13195693e-01 -8.05730402e-01 3.06230873e-01 -7.45807052e-01
-1.24696112e+00 6.04945242e-01 6.11451566e-01 2.17482850e-01
1.77971804e+00 -5.67091107e-01 6.80097938e-01 2.96390712e-01
2.86355317e-01 -1.27027917e+00 8.03883448e-02 -7.23550096e-02
8.76398385e-01 -1.26982164e+00 7.39673898e-02 -5.92399120e-01
-4.46663558e-01 1.01924443e+00 5.70050776e-01 -3.58474135e-01
3.50789279e-01 -5.25382198e-02 6.84000775e-02 -9.41243395e-03
-2.45171949e-01 -1.59519225e-01 1.48300648e-01 7.70802557e-01
1.68788970e-01 -2.81860709e-01 -3.81105423e-01 1.70881778e-01
5.54830302e-03 -4.70200777e-02 4.33386505e-01 3.38403374e-01
-5.26223063e-01 -8.40813041e-01 -3.18353295e-01 9.12569091e-02
-2.40460575e-01 -6.02299906e-02 1.88689694e-01 7.37872481e-01
3.45942914e-01 8.71432602e-01 -2.44978532e-01 -1.58543602e-01
3.42024386e-01 -2.49068603e-01 4.28592205e-01 -1.97400659e-01
-2.14734271e-01 2.42894933e-01 -1.06200203e-01 -6.00253046e-01
-7.07038224e-01 -2.41811797e-01 -7.69363582e-01 -2.91121483e-01
-3.91576499e-01 -3.23702022e-02 7.17764497e-01 7.36766338e-01
1.28739640e-01 4.15399343e-01 9.16086674e-01 -8.12902033e-01
-4.39949840e-01 -6.03258133e-01 -7.26674378e-01 4.98953193e-01
2.39551514e-01 -5.24038553e-01 -7.09213376e-01 1.53669059e-01] | [10.704581260681152, -2.55413556098938] |
cf49cd40-0067-4f1f-917a-2d1dbcd3ec87 | image-comes-dancing-with-collaborative | 2110.14147 | null | https://arxiv.org/abs/2110.14147v2 | https://arxiv.org/pdf/2110.14147v2.pdf | Image Comes Dancing with Collaborative Parsing-Flow Video Synthesis | Transferring human motion from a source to a target person poses great potential in computer vision and graphics applications. A crucial step is to manipulate sequential future motion while retaining the appearance characteristic.Previous work has either relied on crafted 3D human models or trained a separate model specifically for each target person, which is not scalable in practice.This work studies a more general setting, in which we aim to learn a single model to parsimoniously transfer motion from a source video to any target person given only one image of the person, named as Collaborative Parsing-Flow Network (CPF-Net). The paucity of information regarding the target person makes the task particularly challenging to faithfully preserve the appearance in varying designated poses. To address this issue, CPF-Net integrates the structured human parsing and appearance flow to guide the realistic foreground synthesis which is merged into the background by a spatio-temporal fusion module. In particular, CPF-Net decouples the problem into stages of human parsing sequence generation, foreground sequence generation and final video generation. The human parsing generation stage captures both the pose and the body structure of the target. The appearance flow is beneficial to keep details in synthesized frames. The integration of human parsing and appearance flow effectively guides the generation of video frames with realistic appearance. Finally, the dedicated designed fusion network ensure the temporal coherence. We further collect a large set of human dancing videos to push forward this research field. Both quantitative and qualitative results show our method substantially improves over previous approaches and is able to generate appealing and photo-realistic target videos given any input person image. All source code and dataset will be released at https://github.com/xiezhy6/CPF-Net. | ['Liang Lin', 'Haoye Dong', 'Yubei Xiao', 'Xiaodan Liang', 'Zhenyu Xie', 'Bowen Wu'] | 2021-10-27 | null | null | null | null | ['human-parsing'] | ['computer-vision'] | [ 2.92546779e-01 -1.85477994e-02 1.88926607e-02 -8.73012990e-02
-3.19885015e-01 -4.46487069e-01 5.14479578e-01 -7.54604459e-01
-1.12825446e-01 5.49726844e-01 9.18703899e-02 1.02091230e-01
5.20398557e-01 -7.42842734e-01 -7.11941421e-01 -8.25439811e-01
1.14410289e-01 1.32234842e-01 5.66527188e-01 -1.24105506e-01
-1.57179490e-01 4.51986611e-01 -1.50874043e+00 1.98532194e-01
6.53865039e-01 6.11922085e-01 3.10318172e-01 1.02057505e+00
2.90885836e-01 7.75377572e-01 -5.42284667e-01 -4.94578928e-01
5.36975920e-01 -8.05976331e-01 -5.09613991e-01 5.36412835e-01
6.96926057e-01 -7.11324036e-01 -5.04812896e-01 8.32700431e-01
6.57415211e-01 3.53437573e-01 4.00326997e-01 -1.47918069e+00
-3.49620402e-01 2.51349777e-01 -8.09809804e-01 3.78625616e-02
5.73320210e-01 5.95797837e-01 4.46780026e-01 -5.12829423e-01
8.95855963e-01 1.39961851e+00 4.67307210e-01 1.01952708e+00
-8.79495561e-01 -6.94502234e-01 3.91713768e-01 -7.95183107e-02
-1.04364562e+00 -4.46408331e-01 9.05878544e-01 -4.99112040e-01
3.03242177e-01 2.80935973e-01 1.14607835e+00 1.48301613e+00
2.15971500e-01 9.98088956e-01 6.66306794e-01 -3.46051067e-01
-5.14107719e-02 -1.88930824e-01 -1.67949408e-01 7.45114625e-01
1.79279193e-01 3.60598564e-01 -5.81694126e-01 2.15879440e-01
1.24623919e+00 4.78470735e-02 -5.38502276e-01 -4.24189419e-01
-1.46980107e+00 5.30997515e-01 1.34272963e-01 5.64035699e-02
-4.39024270e-01 2.69759417e-01 1.40870899e-01 -1.17171079e-01
2.63888180e-01 -1.49479747e-01 -1.34303495e-01 -8.44922382e-03
-1.14289594e+00 5.92823148e-01 5.80068231e-01 1.06933451e+00
4.04840589e-01 3.84022981e-01 -3.82155091e-01 4.65266794e-01
3.15826118e-01 7.41573215e-01 2.01083869e-01 -1.43616569e+00
3.63378972e-01 3.14683139e-01 3.37529689e-01 -1.19459271e+00
-1.76242962e-01 -2.56818742e-01 -8.34253609e-01 4.30804104e-01
6.83436513e-01 -5.08368790e-01 -9.19396937e-01 1.85231173e+00
8.89199734e-01 3.33534032e-01 -6.89814687e-02 1.24655664e+00
8.92963648e-01 8.67074490e-01 1.20766342e-01 -2.14644909e-01
1.35495675e+00 -1.35650325e+00 -7.62259841e-01 -4.66399431e-01
1.14615411e-01 -8.45080853e-01 8.64745021e-01 2.38338843e-01
-1.56196928e+00 -8.73956859e-01 -8.52531493e-01 -8.79180953e-02
2.11431786e-01 2.46873125e-01 5.29892385e-01 6.60837829e-01
-9.94065106e-01 4.35611010e-01 -8.72190714e-01 -3.93117011e-01
1.77430883e-01 2.48682395e-01 -5.12136161e-01 -1.51083320e-01
-1.20819080e+00 5.57224452e-01 3.53074074e-01 4.59620684e-01
-1.02254105e+00 -5.43897986e-01 -1.00804508e+00 -3.93544614e-01
3.38514358e-01 -1.42487025e+00 1.34342325e+00 -1.45635641e+00
-1.73839700e+00 7.66617060e-01 -1.86912596e-01 -1.27259195e-01
1.05860126e+00 -2.80464888e-01 -1.00888014e-01 2.73818344e-01
5.71612976e-02 9.43271160e-01 1.00414920e+00 -1.38944900e+00
-8.38626981e-01 -1.07547231e-01 3.27323042e-02 3.59205127e-01
3.16696055e-02 1.01136237e-01 -1.13335204e+00 -9.40453470e-01
-2.32449755e-01 -9.92463887e-01 -1.73868164e-01 1.97377175e-01
-3.74485940e-01 1.71713993e-01 9.46609676e-01 -1.07120860e+00
1.15877712e+00 -1.87182021e+00 5.21448851e-01 -2.45412260e-01
1.68615609e-01 3.35329354e-01 -1.90945834e-01 2.25199848e-01
-5.10736667e-02 -2.25345179e-01 -2.01867625e-01 -5.74054956e-01
-2.21494541e-01 7.68610835e-02 -4.82100360e-02 5.54533005e-01
4.50033098e-02 9.95533764e-01 -1.02237248e+00 -6.61784947e-01
3.96590233e-01 7.45481670e-01 -5.96945286e-01 5.05063176e-01
-1.34176150e-01 7.63899982e-01 -4.68893379e-01 7.22110391e-01
6.43736720e-01 -1.30885765e-01 9.96831134e-02 -3.14885110e-01
7.39960372e-02 -4.71648902e-01 -1.35580337e+00 1.73107767e+00
-1.05649438e-02 3.93050700e-01 3.38031530e-01 -5.02419591e-01
6.00072026e-01 3.41350675e-01 6.95287347e-01 -4.43872482e-01
2.16250017e-01 -7.64526948e-02 -5.34995012e-02 -6.16469204e-01
3.53947312e-01 -1.70431376e-01 -2.03987639e-02 1.82768673e-01
-1.07724488e-01 1.22814313e-01 2.72801697e-01 1.48292437e-01
7.46098161e-01 7.43495047e-01 1.07753426e-01 1.05660953e-01
6.28486335e-01 -2.61027049e-02 8.84219825e-01 4.05711114e-01
-4.59795237e-01 1.05916548e+00 1.37779519e-01 -4.90652502e-01
-1.13816786e+00 -1.13036752e+00 5.30957341e-01 8.88696074e-01
5.22380710e-01 -2.54331976e-01 -1.01742566e+00 -6.05967164e-01
-3.41745257e-01 4.58249241e-01 -5.54892898e-01 -1.00425042e-01
-1.06212747e+00 -5.38004279e-01 3.62853050e-01 4.96106207e-01
7.63850272e-01 -1.25657439e+00 -8.07253778e-01 1.29682109e-01
-5.91660142e-01 -1.15214384e+00 -9.56168532e-01 -6.40380442e-01
-6.98648870e-01 -1.11602008e+00 -1.21318138e+00 -8.93660784e-01
6.54861689e-01 3.74958575e-01 9.91635740e-01 5.27866669e-02
-2.60285139e-01 4.14361060e-01 -3.49144340e-01 -8.53935406e-02
-4.50547725e-01 -2.63597876e-01 -2.29374673e-02 3.38205546e-01
-1.78340524e-01 -4.53354836e-01 -8.96182120e-01 4.04683024e-01
-8.78339052e-01 7.51211941e-01 4.73338604e-01 6.19143665e-01
3.91536713e-01 1.76785216e-01 -2.51982287e-02 -6.34330332e-01
1.72830313e-01 -2.50487059e-01 -2.98322111e-01 2.71959156e-01
1.67336240e-01 -4.44890946e-01 5.51647842e-01 -7.06151307e-01
-1.37646401e+00 4.27395880e-01 -6.93813935e-02 -6.40289307e-01
-1.91194639e-01 -3.57721627e-01 -5.52297831e-01 2.25794062e-01
3.13816607e-01 2.61658043e-01 3.21797840e-02 -1.98498949e-01
4.45226729e-01 1.67023510e-01 9.06377316e-01 -6.22780263e-01
1.03686047e+00 6.20287418e-01 -1.15608282e-01 -7.42681026e-01
-5.63373387e-01 -1.96880817e-01 -7.73424149e-01 -7.15049326e-01
1.22294891e+00 -9.23226535e-01 -6.45632446e-01 9.36012208e-01
-1.31355023e+00 -6.57330036e-01 -7.22453594e-02 2.99656391e-01
-7.31219947e-01 5.93089163e-01 -7.23189414e-01 -7.88728893e-01
-4.08076972e-01 -1.06596982e+00 1.28946519e+00 4.60868210e-01
-1.95541993e-01 -1.06767440e+00 3.42374183e-02 5.76826751e-01
8.99855345e-02 7.08840311e-01 3.03673327e-01 1.16515875e-01
-8.08520079e-01 2.37629307e-03 5.52561209e-02 1.37463987e-01
2.41612181e-01 3.82728040e-01 -7.66751170e-01 -3.77248883e-01
-4.81713340e-02 1.56632498e-01 6.75045192e-01 6.68846071e-01
5.85994840e-01 -1.82669550e-01 -3.04159045e-01 7.20554471e-01
1.04289198e+00 3.49188209e-01 6.80700481e-01 2.59132564e-01
1.07935345e+00 8.97592306e-01 7.65710056e-01 2.42093027e-01
3.31195146e-01 8.99481237e-01 1.16919354e-01 -3.46825600e-01
-6.44876361e-01 -5.01910865e-01 6.90024257e-01 5.23243427e-01
-4.45756584e-01 -5.01392305e-01 -7.43907332e-01 3.55968326e-01
-2.03875065e+00 -1.28881979e+00 -5.36742918e-02 2.03407407e+00
6.22624695e-01 -1.50576055e-01 3.78645837e-01 -1.18828036e-01
9.95882213e-01 1.02837771e-01 -3.65760922e-01 2.62431085e-01
-1.66365489e-01 -2.98723578e-01 1.85050771e-01 6.27420425e-01
-1.20160103e+00 9.97184515e-01 5.54920340e+00 7.16246665e-01
-1.09566450e+00 8.44583213e-02 6.32711828e-01 -2.01217711e-01
-7.57487044e-02 -1.05638579e-01 -8.33081126e-01 6.12114012e-01
5.73025405e-01 6.52493956e-03 1.67712554e-01 4.79275614e-01
7.17911363e-01 -7.46394172e-02 -9.32521403e-01 1.08725429e+00
1.13625623e-01 -1.13640213e+00 2.06058636e-01 -4.17750552e-02
7.16968119e-01 -6.95719957e-01 2.55779177e-02 9.48119089e-02
9.75649804e-02 -8.59853268e-01 1.16381490e+00 7.17132568e-01
5.77571630e-01 -6.08525217e-01 4.56655741e-01 4.24566060e-01
-1.38491786e+00 8.92756879e-02 -1.53705264e-02 5.43812439e-02
8.27386677e-01 2.60346770e-01 -3.02004278e-01 6.91889882e-01
6.75891042e-01 7.51897097e-01 -5.08553445e-01 8.93852174e-01
-2.61418790e-01 2.48038143e-01 -8.71314183e-02 5.43734014e-01
-1.36762680e-02 -3.51248085e-01 6.66763306e-01 1.19238985e+00
3.89558852e-01 1.55222401e-01 4.32975024e-01 7.75249958e-01
3.56887668e-01 -2.62190253e-02 -5.16626120e-01 1.05286226e-01
1.71220347e-01 1.29860556e+00 -8.35233331e-01 -4.07377303e-01
-2.32690215e-01 1.32901943e+00 -8.59364495e-02 5.00328600e-01
-1.22924912e+00 8.17033499e-02 5.81226230e-01 3.85022998e-01
2.46475920e-01 -2.71084040e-01 5.59380949e-02 -1.40511870e+00
1.30219042e-01 -1.04883039e+00 3.20693582e-01 -9.01936591e-01
-1.03429317e+00 6.90687835e-01 2.35022277e-01 -1.34819639e+00
-4.62844461e-01 -3.11369240e-01 -7.13140965e-01 7.75469005e-01
-1.17441571e+00 -1.65747583e+00 -7.02183366e-01 6.72003269e-01
9.73218024e-01 7.98289776e-02 2.53121644e-01 4.92305368e-01
-9.35028970e-01 3.91408771e-01 -5.67704678e-01 2.87955582e-01
8.78201962e-01 -9.26552534e-01 5.39837897e-01 1.37368214e+00
-1.85424209e-01 4.41703439e-01 7.92976141e-01 -9.96983230e-01
-1.30301714e+00 -1.20952427e+00 5.79653680e-01 -5.50044298e-01
1.10310420e-01 -2.67023534e-01 -6.66700423e-01 6.47198021e-01
2.88585782e-01 3.63147110e-02 1.87358290e-01 -7.30098844e-01
1.58158943e-01 -9.22377105e-04 -8.78419042e-01 9.02589977e-01
1.20248675e+00 3.89947668e-02 -2.72716761e-01 3.98166850e-02
6.99704826e-01 -7.02409565e-01 -6.09450698e-01 3.75054896e-01
7.56013095e-01 -1.20042682e+00 1.16527939e+00 -3.57083142e-01
5.30000985e-01 -7.54098475e-01 1.40858158e-01 -8.55229020e-01
-3.10395390e-01 -8.80306840e-01 -3.75264913e-01 1.31338835e+00
-1.23690680e-01 -1.86610803e-01 9.46323991e-01 9.16775048e-01
8.59587938e-02 -5.82516730e-01 -4.17343229e-01 -5.57126522e-01
-2.49853842e-02 -2.23419040e-01 1.75453991e-01 6.07731044e-01
-5.85611701e-01 1.37057111e-01 -9.81116176e-01 2.80795842e-01
8.62078607e-01 1.24115184e-01 1.27548659e+00 -7.66378105e-01
-5.02002478e-01 -2.92982399e-01 -2.61209607e-01 -1.25454414e+00
1.17688343e-01 -4.58400935e-01 5.53847328e-02 -1.54855978e+00
3.00046146e-01 -8.12977776e-02 4.15957928e-01 2.34704211e-01
-4.87748951e-01 3.07635128e-01 5.95681489e-01 1.77762598e-01
-3.91007960e-01 5.41419625e-01 1.86302900e+00 3.33679132e-02
-3.11115175e-01 2.81079084e-01 -4.13114399e-01 7.99298167e-01
6.41153634e-01 -1.23888507e-01 -6.35052860e-01 -3.63797933e-01
-3.34351003e-01 4.03917938e-01 7.19937205e-01 -1.02597499e+00
2.76441365e-01 -3.40468079e-01 8.15356314e-01 -5.43385267e-01
4.70498592e-01 -7.14581132e-01 8.09560299e-01 6.31360173e-01
8.03316683e-02 1.48673862e-01 6.34011552e-02 6.13445163e-01
-3.47823426e-02 5.10759503e-02 8.81114185e-01 -3.69584084e-01
-9.83588040e-01 5.96101820e-01 -2.29795441e-01 -3.39779630e-03
1.33011651e+00 -5.95287085e-01 -9.87008289e-02 -5.72433650e-01
-7.90342391e-01 2.22483948e-01 7.47021198e-01 6.01514101e-01
7.33682275e-01 -1.36426413e+00 -7.99374044e-01 1.97145522e-01
-3.62145185e-01 1.74784184e-01 6.97096884e-01 8.05488229e-01
-8.99209797e-01 1.30172208e-01 -4.29476321e-01 -5.85583389e-01
-1.51615751e+00 6.38757467e-01 5.57130814e-01 -3.68321501e-02
-9.23954427e-01 7.31176674e-01 7.90046334e-01 -1.21203564e-01
1.87462121e-01 -7.67296478e-02 -3.86504903e-02 -1.84146553e-01
7.04849660e-01 5.04422069e-01 -5.76758444e-01 -1.03295386e+00
-1.69901997e-01 9.33465958e-01 2.88560897e-01 -2.39338115e-01
9.97561991e-01 -3.93389553e-01 1.34887710e-01 1.55214593e-01
8.55705619e-01 1.07992981e-02 -1.95712161e+00 1.06163673e-01
-6.18575990e-01 -7.36198604e-01 -2.95141339e-01 -5.02675295e-01
-1.47460485e+00 7.74245322e-01 3.86022270e-01 -3.36485326e-01
1.34698105e+00 -3.16088498e-01 1.10005558e+00 -2.76283979e-01
4.58649993e-01 -7.07757890e-01 3.31188023e-01 9.21608284e-02
8.10369849e-01 -9.70041394e-01 -8.20257068e-02 -6.98527694e-01
-8.70944202e-01 1.16473615e+00 9.56146479e-01 -5.75947054e-02
2.10215956e-01 2.32354477e-01 2.37184703e-01 9.11266953e-02
-4.79930878e-01 -1.00321867e-01 5.16414762e-01 8.55460703e-01
2.52686858e-01 -1.37491614e-01 -8.77795517e-02 5.26640117e-01
-1.50314227e-01 -5.21586724e-02 4.93135691e-01 8.00138891e-01
-1.94655582e-01 -1.23643744e+00 -6.33941889e-01 -3.31224680e-01
-3.43617886e-01 3.01228732e-01 -2.94905692e-01 1.00254667e+00
4.41508591e-01 8.69818151e-01 -1.69636071e-01 -1.65092319e-01
3.69608998e-01 -1.77434996e-01 6.24900699e-01 -4.47791964e-01
-5.60954332e-01 4.89595264e-01 -2.51368210e-02 -7.49666929e-01
-7.28143036e-01 -7.31085479e-01 -1.15353584e+00 -4.78559345e-01
7.17625841e-02 -2.52045035e-01 9.59366560e-02 5.42221427e-01
9.17470306e-02 6.43733561e-01 3.27639073e-01 -1.44235969e+00
1.51562970e-03 -4.81106162e-01 -4.39754874e-01 6.43563628e-01
3.92079234e-01 -6.05035603e-01 -2.38755077e-01 6.99109554e-01] | [10.911125183105469, -0.7910422682762146] |
0e4f3bc6-f88d-4be2-b30a-51704ee685c8 | toward-fast-and-accurate-neural-discourse | 1808.09147 | null | http://arxiv.org/abs/1808.09147v1 | http://arxiv.org/pdf/1808.09147v1.pdf | Toward Fast and Accurate Neural Discourse Segmentation | Discourse segmentation, which segments texts into Elementary Discourse Units,
is a fundamental step in discourse analysis. Previous discourse segmenters rely
on complicated hand-crafted features and are not practical in actual use. In
this paper, we propose an end-to-end neural segmenter based on BiLSTM-CRF
framework. To improve its accuracy, we address the problem of data
insufficiency by transferring a word representation model that is trained on a
large corpus. We also propose a restricted self-attention mechanism in order to
capture useful information within a neighborhood. Experiments on the RST-DT
corpus show that our model is significantly faster than previous methods, while
achieving new state-of-the-art performance. | ['Jingfeng Yang', 'Yizhong Wang', 'Sujian Li'] | 2018-08-28 | toward-fast-and-accurate-neural-discourse-1 | https://aclanthology.org/D18-1116 | https://aclanthology.org/D18-1116.pdf | emnlp-2018-10 | ['discourse-segmentation'] | ['natural-language-processing'] | [ 3.60231668e-01 5.43000221e-01 -3.85018647e-01 -3.70215893e-01
-8.08484316e-01 -3.47630918e-01 7.20340490e-01 1.53576761e-01
-6.93208337e-01 8.58353674e-01 7.40420997e-01 -6.12034976e-01
4.51752812e-01 -7.62034416e-01 -5.05581439e-01 -3.07042748e-01
1.95408955e-01 3.94112110e-01 6.03806973e-01 -3.87281626e-01
3.80196065e-01 -2.46394023e-01 -1.07416379e+00 6.42117083e-01
1.12334156e+00 4.33935374e-01 4.41474974e-01 7.53573000e-01
-3.93076479e-01 1.13590741e+00 -9.00131524e-01 -1.05849624e-01
-2.73490161e-01 -8.54694068e-01 -1.38714457e+00 1.05559871e-01
-1.20025866e-01 -4.10240978e-01 -2.51619488e-01 8.87142420e-01
2.74423927e-01 5.05982697e-01 5.08307874e-01 -4.09516960e-01
-6.03911817e-01 1.14416218e+00 -3.77377152e-01 3.54577750e-01
4.17534918e-01 -1.75509349e-01 1.07240009e+00 -5.20541072e-01
7.53387034e-01 1.16913211e+00 4.42664117e-01 5.94367087e-01
-9.34566796e-01 -2.18168303e-01 4.77669626e-01 3.96188021e-01
-7.68741250e-01 -4.28553820e-01 8.78721893e-01 -3.79954517e-01
1.30122030e+00 2.25865319e-01 3.30656111e-01 9.71545219e-01
-1.22930378e-01 1.26630032e+00 7.60031998e-01 -7.42263377e-01
1.05016679e-01 -2.68616676e-01 5.02385259e-01 2.98606157e-01
-3.70242804e-01 -4.72922176e-01 -1.26416519e-01 1.13236427e-01
6.82335019e-01 -4.15634364e-01 -3.57378989e-01 2.31183931e-01
-1.09857178e+00 1.07503247e+00 4.31977898e-01 8.48347127e-01
-3.40861231e-01 -1.94438502e-01 7.54865348e-01 2.70280123e-01
7.75123179e-01 4.11408573e-01 -4.49875854e-02 -5.79747975e-01
-1.06565797e+00 1.78794876e-01 7.81265676e-01 7.21412539e-01
2.67749071e-01 -8.41634274e-02 -5.96753478e-01 8.15364420e-01
2.21856773e-01 -2.85207517e-02 7.51326561e-01 -6.98729038e-01
7.12038219e-01 7.97888756e-01 4.98059671e-03 -5.40736079e-01
-5.47666907e-01 -9.18332860e-02 -7.39944577e-01 -8.30284134e-02
4.98553693e-01 -4.08940494e-01 -8.12101483e-01 1.62263632e+00
3.64588886e-01 -5.20849740e-03 1.63816020e-01 9.28997338e-01
8.82837474e-01 9.45033371e-01 2.03335047e-01 -4.35419947e-01
1.37118185e+00 -1.45397019e+00 -1.08911240e+00 -2.13087603e-01
8.29192579e-01 -8.03033352e-01 1.06417561e+00 3.08518320e-01
-1.26381791e+00 -6.55092299e-01 -9.86937582e-01 -5.08781731e-01
-7.32465610e-02 -4.86864941e-03 2.64327943e-01 5.09350300e-01
-6.18512511e-01 6.71148598e-01 -9.62855101e-01 -2.40867347e-01
3.25542837e-01 2.16007784e-01 -6.00372590e-02 2.64297605e-01
-1.30779207e+00 1.03332531e+00 8.98735821e-01 1.16730720e-01
-4.18790728e-01 -1.74172625e-01 -8.70681584e-01 -5.35346866e-02
5.71754932e-01 -4.57821697e-01 1.64350939e+00 -9.27350521e-01
-2.03056979e+00 5.51150143e-01 -3.88374716e-01 -6.05217755e-01
4.14610445e-01 -4.45814580e-01 -2.10790157e-01 1.64586335e-01
-1.53293192e-01 1.55765831e-01 5.23807466e-01 -9.01448846e-01
-7.08920062e-01 -1.08110525e-01 2.75485873e-01 4.20917928e-01
-3.64588976e-01 2.23068997e-01 -3.13662708e-01 -6.27284884e-01
-7.61226565e-02 -6.44919395e-01 -3.49614948e-01 -8.00872147e-01
-5.72322786e-01 -7.75310278e-01 8.27217638e-01 -1.12553668e+00
1.64552927e+00 -1.74313402e+00 3.39811057e-01 -1.83887124e-01
1.03580117e-01 9.40053463e-01 2.81819198e-02 4.75759983e-01
2.82633066e-01 3.39059834e-03 -3.45597237e-01 -4.68259037e-01
-7.41263777e-02 1.79479271e-01 -1.01884052e-01 3.26233685e-01
3.50190729e-01 9.42620873e-01 -1.01836133e+00 -6.08182073e-01
1.65455908e-01 2.88575113e-01 -4.75567073e-01 5.07850230e-01
-6.43117607e-01 6.40896082e-01 -7.27907479e-01 2.35840112e-01
3.40238780e-01 -2.32262686e-01 3.71552438e-01 2.07110003e-01
-3.79268497e-01 9.83227015e-01 -8.08671236e-01 2.11771631e+00
-3.40345740e-01 7.92146087e-01 -1.78309217e-01 -1.40066242e+00
9.12092388e-01 5.01616836e-01 -7.95894396e-03 -5.42026818e-01
6.34382904e-01 6.99687973e-02 2.49051496e-01 -7.84073174e-01
9.40138698e-01 -9.64851230e-02 -1.32182717e-01 6.42455757e-01
-1.06086582e-02 3.03384632e-01 6.44817233e-01 -5.75559624e-02
1.01805532e+00 2.85766751e-01 4.52663302e-01 -1.48855731e-01
6.97103441e-01 2.44092941e-01 4.91536379e-01 4.38171387e-01
-6.84779808e-02 7.97337890e-01 6.98774278e-01 -2.19411403e-01
-1.07236719e+00 -4.28217113e-01 -3.24777439e-02 1.37371826e+00
-5.71237728e-02 -3.31853777e-01 -1.14743364e+00 -8.40724468e-01
-4.84337896e-01 9.40180957e-01 -5.91989398e-01 2.74034232e-01
-1.25818443e+00 -6.82444394e-01 5.55424690e-01 7.04155564e-01
5.48447251e-01 -1.30175602e+00 -7.61160851e-01 7.03076184e-01
-4.82831955e-01 -1.00621164e+00 -3.59064162e-01 -3.39768603e-02
-7.55766690e-01 -9.44368422e-01 -8.91743720e-01 -1.03962159e+00
5.12080967e-01 1.47015691e-01 9.40276146e-01 1.95514783e-01
4.05500382e-01 -5.22540271e-01 -7.95877516e-01 -2.55129069e-01
-5.86257935e-01 7.07413018e-01 -5.27574480e-01 -3.37705761e-01
5.02193332e-01 -2.98825651e-01 -2.93295950e-01 -8.38128105e-02
-7.79015660e-01 2.87021250e-01 3.96547168e-01 1.08837032e+00
2.45866701e-01 -3.53288919e-01 8.56596172e-01 -1.31885445e+00
9.62429583e-01 -4.21500891e-01 -3.06741655e-01 3.13335210e-01
-1.22040093e-01 6.14649355e-02 6.91403329e-01 -5.01499057e-01
-1.41339159e+00 -1.91186234e-01 -5.28477609e-01 8.57615992e-02
-1.72304362e-01 8.58551562e-01 -1.66251719e-01 6.02563918e-01
6.16924763e-01 -1.44443199e-01 -6.66763410e-02 -7.06459999e-01
6.08006537e-01 1.01962996e+00 6.24597192e-01 -4.69816357e-01
3.14981610e-01 9.10491683e-03 -5.79855859e-01 -9.21323061e-01
-9.82679188e-01 -5.59097052e-01 -8.34101498e-01 3.26957740e-02
9.57193017e-01 -8.24760735e-01 -6.05463326e-01 2.57635504e-01
-1.50102627e+00 -6.32018805e-01 -3.03678960e-01 5.40960610e-01
-4.54510331e-01 6.27528846e-01 -8.64918292e-01 -8.35155666e-01
-4.62394059e-01 -9.83102441e-01 6.49597168e-01 5.16324699e-01
-4.73468125e-01 -1.01014292e+00 3.78394276e-01 4.23179567e-01
2.62258589e-01 3.60910028e-01 5.86418927e-01 -1.08779490e+00
-1.78433508e-01 9.85130146e-02 -3.47954869e-01 2.66580880e-01
2.27079485e-02 -9.41644907e-02 -1.01317155e+00 -2.28542745e-01
5.22350557e-02 -3.55673760e-01 1.08615291e+00 3.33332986e-01
8.32411408e-01 -3.31173390e-01 -3.15372229e-01 1.70642287e-01
9.07677174e-01 2.47367203e-01 6.27798319e-01 4.74357724e-01
7.67427504e-01 6.98687613e-01 7.38806069e-01 3.26001585e-01
6.42009974e-01 3.82438928e-01 6.63721338e-02 -1.57233551e-01
-1.74686909e-01 -2.12591499e-01 2.46793583e-01 1.28070283e+00
-2.93651968e-01 -5.67817986e-01 -9.68203962e-01 8.79893363e-01
-2.28025222e+00 -9.59430158e-01 -3.37044209e-01 1.71438789e+00
1.18311465e+00 5.20043254e-01 2.83616483e-01 1.61214367e-01
8.25353146e-01 3.54887933e-01 -2.78472513e-01 -6.12055361e-01
1.75464153e-02 1.81489214e-01 9.27469358e-02 6.75830960e-01
-1.21864736e+00 1.18387294e+00 6.23375320e+00 8.42833996e-01
-9.71831322e-01 2.05164403e-01 4.79549944e-01 1.61171973e-01
-2.62965947e-01 5.74133880e-02 -6.75884545e-01 3.60646933e-01
1.17733502e+00 -2.73948848e-01 -1.29239764e-02 7.87690163e-01
1.59679905e-01 -1.80779278e-01 -9.64748502e-01 3.34494531e-01
1.52106285e-01 -1.34570289e+00 -3.25266004e-01 -1.03170894e-01
6.87499762e-01 -1.50671437e-01 -4.70363826e-01 5.83559990e-01
3.26203197e-01 -1.02602041e+00 5.15266240e-01 2.68617004e-01
4.72904056e-01 -7.48762548e-01 1.00481987e+00 6.22196198e-01
-9.27397072e-01 2.21255168e-01 -4.64181840e-01 -6.12227321e-01
5.45451939e-01 4.11537230e-01 -1.10016048e+00 7.25439429e-01
2.95069832e-02 5.69758296e-01 -2.51265228e-01 8.12109768e-01
-5.52721977e-01 8.92385244e-01 -2.89910495e-01 -3.44671607e-01
5.22263348e-01 -3.61054018e-02 4.10614640e-01 1.48192811e+00
1.17466636e-02 3.04925501e-01 3.66653442e-01 6.42236531e-01
-2.19192818e-01 1.21889591e-01 -1.90926448e-01 -2.23494023e-02
3.64772379e-01 9.88602102e-01 -7.68669128e-01 -6.04103327e-01
-4.44923103e-01 7.59777427e-01 5.69945991e-01 1.29868746e-01
-6.78616524e-01 -6.39722347e-01 1.67038560e-01 -2.05013335e-01
5.47087431e-01 -2.54807174e-01 -2.61554718e-01 -1.11022890e+00
-2.95038894e-03 -8.37289751e-01 2.65900731e-01 -1.37624472e-01
-1.02087212e+00 9.09430683e-01 -8.36844146e-02 -9.56243455e-01
-8.51784766e-01 -3.67135763e-01 -9.82301235e-01 9.22330141e-01
-1.66812074e+00 -1.31272984e+00 -3.59848924e-02 3.32990378e-01
9.74680424e-01 -4.80840355e-03 8.95629108e-01 1.61077857e-01
-7.63810098e-01 4.38080579e-01 2.39437073e-01 5.05855918e-01
6.49490416e-01 -1.26995862e+00 7.82912493e-01 1.01733625e+00
-7.59817362e-02 5.08715987e-01 7.94428051e-01 -6.64817393e-01
-8.01787853e-01 -8.61565173e-01 1.19378555e+00 -2.16119841e-01
6.33313835e-01 -1.99964225e-01 -1.34361684e+00 7.60271013e-01
7.49903321e-01 -4.39986378e-01 8.33068848e-01 3.67500305e-01
2.43698116e-02 6.54809654e-01 -7.74174094e-01 4.74665731e-01
7.36589074e-01 -3.71633679e-01 -1.42230773e+00 1.72858953e-01
8.75833571e-01 -8.36223066e-01 -8.88961971e-01 2.66341060e-01
2.43275344e-01 -7.71631718e-01 5.88286042e-01 -6.45498991e-01
6.20556176e-01 -1.38154164e-01 2.04738855e-01 -1.20319188e+00
-2.53884077e-01 -7.92140067e-01 -2.27021605e-01 1.43299806e+00
4.27794307e-01 -2.57608742e-01 6.90034270e-01 4.52632844e-01
-5.16279817e-01 -6.13635719e-01 -9.57428932e-01 -5.97005367e-01
4.37182367e-01 -2.30535179e-01 5.62694848e-01 9.76993442e-01
8.29429209e-01 8.13269496e-01 -2.35698730e-01 -3.61849546e-01
1.62494928e-01 2.67714202e-01 7.30803370e-01 -1.04379225e+00
-1.25972405e-01 -6.43735647e-01 1.54293090e-01 -1.54182327e+00
4.27696109e-01 -5.37150919e-01 3.10108393e-01 -1.79366577e+00
-2.33767089e-02 -1.09992556e-01 -9.68334302e-02 2.83735454e-01
-5.45918524e-01 6.80540130e-02 2.57281035e-01 1.70632303e-01
-8.47052574e-01 7.23134816e-01 1.34903848e+00 -1.95473954e-02
-5.99976778e-01 -5.04786558e-02 -5.21384239e-01 8.46069932e-01
9.73966956e-01 -3.12782288e-01 -1.59445271e-01 -5.83943725e-01
-2.52598464e-01 1.35872275e-01 -2.65578747e-01 -7.28750706e-01
4.16680932e-01 -1.66975796e-01 7.70847872e-02 -1.02884781e+00
2.58984268e-01 -1.76210642e-01 -5.09686470e-01 1.80970833e-01
-5.73299766e-01 -2.99112916e-01 5.63744940e-02 1.95984542e-01
-4.45804983e-01 -7.21684277e-01 4.30455178e-01 -1.91925988e-01
-5.29586196e-01 -3.59922618e-01 -5.08983910e-01 2.65357763e-01
8.79583597e-01 9.96360630e-02 -4.91289407e-01 -1.19377382e-01
-4.61170673e-01 4.09194291e-01 3.14513892e-01 4.03371334e-01
3.42069298e-01 -1.18075156e+00 -9.55514848e-01 -1.19557381e-01
-1.63635314e-01 5.57426155e-01 1.82500377e-01 6.25278294e-01
-4.49823976e-01 5.22817314e-01 -1.14244625e-01 -3.67409021e-01
-1.22407413e+00 3.94235730e-01 -2.33235225e-01 -8.14000726e-01
-8.23725939e-01 7.09266007e-01 -8.38805586e-02 -1.94476590e-01
2.97394246e-01 -5.85286319e-01 -7.87546575e-01 3.68213713e-01
8.31966281e-01 3.24680269e-01 -1.16135977e-01 -7.81628609e-01
5.96448546e-03 3.06297779e-01 -2.80079991e-01 -8.36776197e-02
1.34874153e+00 -2.66280621e-01 6.15397990e-02 6.00403249e-01
9.82392907e-01 -1.97990716e-01 -1.25184846e+00 -4.96221066e-01
3.82617652e-01 -2.40814701e-01 8.32191706e-02 -4.07876194e-01
-5.50949335e-01 9.91764188e-01 -4.45255376e-02 6.21884048e-01
9.36098695e-01 -2.18826622e-01 1.24461520e+00 4.33890581e-01
-1.56190351e-01 -1.45119655e+00 -1.08157054e-01 1.00108755e+00
6.92167878e-01 -1.35540748e+00 8.89574364e-02 -2.13866159e-01
-6.70939445e-01 1.28446472e+00 5.50054491e-01 -3.00212055e-01
2.56637365e-01 -6.93517402e-02 1.14290781e-01 7.88361654e-02
-5.95841169e-01 -4.54877853e-01 3.08758259e-01 4.43454534e-01
9.44125891e-01 -8.70260000e-02 -7.00850785e-01 8.52657616e-01
-2.49248177e-01 9.12698060e-02 5.22206306e-01 1.03140056e+00
-7.46999264e-01 -1.36773682e+00 -3.27375561e-01 1.79171234e-01
-7.02857256e-01 1.02250442e-01 -2.27285698e-01 8.53012383e-01
-2.32809007e-01 1.18971789e+00 2.40032330e-01 -1.72189444e-01
4.38130885e-01 1.33801058e-01 2.74197280e-01 -7.56746888e-01
-9.25129473e-01 4.15792823e-01 5.74387848e-01 -1.29440933e-01
-8.06127310e-01 -7.37590492e-01 -1.52080846e+00 -2.99055099e-01
-5.60869694e-01 2.19733104e-01 3.64394426e-01 1.29068816e+00
-1.67701242e-03 1.02498209e+00 4.01194185e-01 -8.51129711e-01
-3.98285389e-01 -1.60452080e+00 -1.23359352e-01 4.46807563e-01
4.39437002e-01 -3.72238636e-01 6.37523755e-02 1.99756891e-01] | [10.777321815490723, 9.446059226989746] |
6dbf0897-a848-42b7-800a-41e105e01776 | towards-unified-text-based-person-retrieval-a | 2306.02898 | null | https://arxiv.org/abs/2306.02898v2 | https://arxiv.org/pdf/2306.02898v2.pdf | Towards Unified Text-based Person Retrieval: A Large-scale Multi-Attribute and Language Search Benchmark | In this paper, we introduce a large Multi-Attribute and Language Search dataset for text-based person retrieval, called MALS, and explore the feasibility of performing pre-training on both attribute recognition and image-text matching tasks in one stone. In particular, MALS contains 1,510,330 image-text pairs, which is about 37.5 times larger than prevailing CUHK-PEDES, and all images are annotated with 27 attributes. Considering the privacy concerns and annotation costs, we leverage the off-the-shelf diffusion models to generate the dataset. To verify the feasibility of learning from the generated data, we develop a new joint Attribute Prompt Learning and Text Matching Learning (APTM) framework, considering the shared knowledge between attribute and text. As the name implies, APTM contains an attribute prompt learning stream and a text matching learning stream. (1) The attribute prompt learning leverages the attribute prompts for image-attribute alignment, which enhances the text matching learning. (2) The text matching learning facilitates the representation learning on fine-grained details, and in turn, boosts the attribute prompt learning. Extensive experiments validate the effectiveness of the pre-training on MALS, achieving state-of-the-art retrieval performance via APTM on three challenging real-world benchmarks. In particular, APTM achieves a consistent improvement of +6.60%, +7.39%, and +15.90% Recall@1 accuracy on CUHK-PEDES, ICFG-PEDES, and RSTPReid datasets by a clear margin, respectively. | ['Zhedong Zheng', 'Li Zhu', 'Yujiao Wu', 'Yaxiong Wang', 'Yinan Zhou', 'Shuyu Yang'] | 2023-06-05 | null | null | null | null | ['pedestrian-attribute-recognition', 'person-retrieval', 'nlp-based-person-retrival', 'image-text-matching', 'text-matching'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'natural-language-processing'] | [ 5.91141954e-02 -2.89957613e-01 -2.79477000e-01 -5.92726290e-01
-1.28738999e+00 -3.64350080e-01 9.67678547e-01 1.43221870e-01
-5.48449755e-01 4.86165226e-01 3.99882466e-01 2.46260077e-01
-1.37917399e-01 -6.82732284e-01 -7.12828219e-01 -6.24689698e-01
1.73416287e-01 8.23518574e-01 -1.73669815e-01 5.70898131e-03
-2.09904194e-01 -7.65495747e-03 -1.50634193e+00 4.91138160e-01
7.60310888e-01 1.52311957e+00 -1.55810550e-01 3.14264953e-01
1.63311869e-01 4.93299067e-01 -4.03109878e-01 -1.14905226e+00
6.23924434e-01 1.74247667e-01 -6.83571935e-01 3.58865075e-02
8.83783937e-01 -7.58833766e-01 -9.00015831e-01 6.93839192e-01
8.73691201e-01 2.45688912e-02 8.66546929e-01 -1.45262170e+00
-1.07352412e+00 2.34726518e-01 -7.78209865e-01 -1.68124318e-01
5.06146252e-01 2.29155585e-01 1.19142878e+00 -1.13647079e+00
5.96216083e-01 1.25888538e+00 5.87798476e-01 5.46097457e-01
-9.36534882e-01 -1.03040791e+00 1.14126958e-01 2.33392000e-01
-1.89657235e+00 -4.94973332e-01 2.69967020e-01 -2.99956948e-01
5.86691380e-01 2.77134001e-01 6.38952076e-01 1.54729390e+00
6.21199831e-02 1.51788843e+00 1.24242091e+00 -1.30969122e-01
-1.96611434e-01 2.73301303e-01 8.14850181e-02 6.43123031e-01
1.69084579e-01 1.07840218e-01 -8.76465976e-01 -4.02130306e-01
4.48049635e-01 3.73745143e-01 -1.58682391e-02 -2.22000197e-01
-1.12511551e+00 7.52756715e-01 3.52990478e-01 -4.47692305e-01
-4.47824657e-01 1.00216962e-01 6.31277859e-01 2.39962965e-01
5.20592332e-01 6.87129349e-02 -1.73823804e-01 8.81558731e-02
-8.82814884e-01 4.00595963e-01 4.50553030e-01 1.44537377e+00
5.97683311e-01 -4.30748403e-01 -5.98313034e-01 1.03458619e+00
2.99996078e-01 1.23006511e+00 2.02081963e-01 -5.52060723e-01
7.30779946e-01 5.54024816e-01 -1.31932899e-01 -6.94724798e-01
6.42256346e-03 -3.51897329e-01 -1.11752951e+00 -6.18142784e-01
3.08298677e-01 -8.64195265e-03 -9.05465484e-01 1.56678748e+00
2.41963297e-01 1.64651752e-01 -2.81903885e-05 9.52006638e-01
1.04539418e+00 5.25221348e-01 2.86276191e-01 1.19269930e-01
1.68463206e+00 -1.03395653e+00 -5.26726782e-01 -2.84531176e-01
5.68959594e-01 -7.99621463e-01 1.15944695e+00 1.07820950e-01
-8.57890725e-01 -4.01672274e-01 -8.09286296e-01 -1.19525097e-01
-3.35027009e-01 3.83696258e-01 7.14177728e-01 4.11971986e-01
-9.16162610e-01 -1.96852796e-02 -3.06405246e-01 -4.39541638e-01
7.16924191e-01 3.93345892e-01 -6.43061757e-01 -5.72354138e-01
-1.40812385e+00 3.26664031e-01 -4.43927608e-02 -2.19682738e-01
-9.31012452e-01 -8.90203476e-01 -5.74220002e-01 6.56873137e-02
4.78818029e-01 -1.04730308e+00 9.10723686e-01 -5.27984023e-01
-8.03603530e-01 1.28980422e+00 -2.28781313e-01 -4.43986505e-01
8.37392688e-01 -5.47084987e-01 -5.70329905e-01 3.88296902e-01
3.85728508e-01 8.48304152e-01 1.04824138e+00 -1.05213022e+00
-6.20918989e-01 -6.71658635e-01 -2.40797877e-01 4.79410797e-01
-7.56892622e-01 3.29139456e-02 -1.25204206e+00 -9.19363797e-01
-3.28574240e-01 -1.06647933e+00 6.88969642e-02 2.10130185e-01
-6.69683635e-01 -3.45636070e-01 4.68024075e-01 -6.53767288e-01
1.03556681e+00 -2.20143127e+00 -2.95043588e-01 4.50020403e-01
4.30228949e-01 1.73301607e-01 -3.23005110e-01 3.36709470e-01
2.29433224e-01 -1.02066018e-01 3.10288697e-01 -8.22685242e-01
7.98282474e-02 7.88243189e-02 -4.63349819e-01 3.25056702e-01
-1.22908391e-01 1.34975159e+00 -5.75395882e-01 -8.06271195e-01
-6.00820184e-02 4.00526702e-01 -2.84502417e-01 2.82852829e-01
7.35583082e-02 9.49740037e-02 -8.38555634e-01 1.10402024e+00
6.37744546e-01 -3.97025496e-01 -1.41444474e-01 -2.46968508e-01
5.56901932e-01 -5.58185577e-02 -1.00774789e+00 1.66072965e+00
-1.37027025e-01 4.54945713e-01 -1.51968181e-01 -6.51752532e-01
8.18017006e-01 1.32489830e-01 7.29053319e-01 -1.24101615e+00
-1.71988189e-01 1.30618930e-01 -7.03111768e-01 -3.55185241e-01
7.22001910e-01 4.63462770e-01 -3.47905934e-01 6.20818913e-01
2.29237671e-03 6.11735821e-01 -1.12256207e-01 6.81751072e-01
7.62511373e-01 -1.60573989e-01 -7.92941526e-02 -1.21609904e-01
4.59248781e-01 -2.52505541e-01 4.29975212e-01 1.12904525e+00
-1.90288484e-01 7.95267224e-01 2.33633444e-02 -5.25311589e-01
-1.04536843e+00 -1.11680913e+00 -9.74757299e-02 1.28515697e+00
3.84192705e-01 -7.46817291e-01 -6.79996490e-01 -8.05937588e-01
3.58596474e-01 5.66372499e-02 -7.46749222e-01 -2.18864560e-01
-2.88273156e-01 -8.00225437e-01 9.25942898e-01 5.65596461e-01
9.97762680e-01 -7.12120771e-01 2.89611183e-02 -1.66838735e-01
-5.48769295e-01 -1.43982959e+00 -1.07021296e+00 -5.74939370e-01
-2.42940471e-01 -8.86140645e-01 -9.84982967e-01 -6.48078382e-01
6.63821042e-01 5.00429451e-01 1.19171071e+00 2.69423425e-01
-5.44282198e-01 9.02979493e-01 -2.91092604e-01 -3.31087917e-01
2.43367746e-01 2.65081316e-01 3.34725767e-01 4.25231069e-01
6.89057946e-01 -1.73060268e-01 -9.11686897e-01 4.54796225e-01
-5.83299816e-01 5.76358056e-03 1.02388692e+00 1.10624945e+00
9.80652630e-01 -3.95585358e-01 3.78424406e-01 -8.29375863e-01
5.55000603e-01 -4.71728891e-01 -3.90528113e-01 6.70065045e-01
-1.07810247e+00 -1.78555876e-01 1.57134101e-01 -6.06542647e-01
-9.90966201e-01 1.58768222e-01 5.21883718e-04 -5.11524856e-01
-2.13053189e-02 2.66928256e-01 -2.64404625e-01 -1.03691537e-02
2.82930374e-01 6.58180833e-01 5.34133203e-02 -3.81065905e-01
3.32233250e-01 1.01744199e+00 7.76527584e-01 -8.17952514e-01
1.02303994e+00 5.36672652e-01 -2.39310220e-01 -4.99237776e-01
-9.08424377e-01 -7.99476743e-01 -3.66009712e-01 -2.25989699e-01
7.40118980e-01 -1.53617263e+00 -9.34253514e-01 7.97343910e-01
-7.79510617e-01 -7.44153932e-02 -1.05068766e-01 3.13909322e-01
-3.78546774e-01 2.60547459e-01 -7.53358006e-01 -5.33676326e-01
-1.00206065e+00 -1.00266385e+00 1.40750074e+00 2.35069782e-01
-6.59039989e-03 -6.41644478e-01 -5.24975240e-01 9.25486386e-01
3.24432999e-01 -2.00582176e-01 6.20223403e-01 -1.08719015e+00
-8.16740811e-01 -4.29649204e-01 -6.17401600e-01 -8.79871771e-02
-1.91965252e-01 -5.20376325e-01 -9.94562745e-01 -5.62594295e-01
-5.73399484e-01 -7.00747609e-01 8.15726459e-01 1.32719561e-01
1.29739165e+00 -3.76794279e-01 -4.69828129e-01 7.99651384e-01
1.18113589e+00 -1.88158810e-01 6.66020989e-01 5.21623731e-01
7.21107304e-01 5.52939355e-01 1.10647047e+00 6.74336612e-01
8.26339722e-01 9.06889617e-01 3.58028337e-03 -2.38037273e-01
-3.20820332e-01 -6.24442220e-01 1.27938315e-01 4.72400010e-01
2.02756256e-01 -1.00841440e-01 -8.47493052e-01 5.36871910e-01
-1.96930766e+00 -9.00539041e-01 4.24954779e-02 2.35888553e+00
9.28487599e-01 -2.35878423e-01 1.48982257e-01 -4.35186148e-01
6.28383040e-01 8.69392157e-02 -6.95483088e-01 3.92368317e-01
-4.43562448e-01 -1.14786603e-01 7.41019547e-01 1.55915290e-01
-1.38449728e+00 1.08640969e+00 5.50136423e+00 1.14057875e+00
-7.37013280e-01 -1.34591041e-02 9.89485621e-01 -3.68301213e-01
-1.62652045e-01 -4.59128857e-01 -1.28532636e+00 5.89692414e-01
5.72519422e-01 -3.55126441e-01 2.91926563e-01 8.97829831e-01
-3.21359813e-01 1.90051034e-01 -1.21516967e+00 1.48920059e+00
2.99965143e-01 -1.39804065e+00 4.66325134e-01 2.54916906e-01
6.53172553e-01 3.02316318e-03 5.80989242e-01 4.66270387e-01
3.23736697e-01 -1.06041753e+00 4.51060116e-01 5.18731415e-01
1.19837999e+00 -6.63562298e-01 8.38102520e-01 9.69610363e-02
-1.33807254e+00 1.86762698e-02 -3.09308141e-01 5.89231908e-01
-1.08657479e-01 4.45260137e-01 -6.47668004e-01 7.68833816e-01
1.04634678e+00 7.11790442e-01 -1.01973772e+00 8.97613704e-01
9.39995274e-02 3.91866028e-01 -4.22400802e-01 8.01316351e-02
1.75089851e-01 -1.59397602e-01 1.62251621e-01 1.37208569e+00
1.58526242e-01 1.45558476e-01 3.37083161e-01 4.92954642e-01
-5.98348260e-01 2.97218740e-01 -3.96428496e-01 6.89484105e-02
9.39691901e-01 1.29100204e+00 -3.20505947e-02 -5.72125912e-01
-5.22862077e-01 1.23349035e+00 1.89759985e-01 5.81997991e-01
-8.18891883e-01 -1.62238941e-01 7.69317031e-01 9.87028107e-02
1.36058316e-01 3.41065496e-01 -1.70047209e-01 -1.21902561e+00
4.33364719e-01 -1.13355613e+00 8.83396327e-01 -6.75620198e-01
-1.80336344e+00 4.90117073e-01 5.06741693e-03 -1.01191115e+00
-2.60973036e-01 -1.59434080e-01 -1.03554092e-01 1.02596664e+00
-1.61987722e+00 -1.83077395e+00 -6.03962839e-01 1.12027156e+00
3.92377347e-01 -6.97368264e-01 7.76586950e-01 7.40718722e-01
-4.32095289e-01 1.55942273e+00 1.30361766e-01 5.86092949e-01
1.37489331e+00 -9.37857807e-01 5.30947864e-01 5.58065295e-01
1.88745648e-01 7.07251608e-01 2.34283686e-01 -7.22210109e-01
-1.77879333e+00 -1.34724581e+00 8.44450951e-01 -6.00818276e-01
4.53645140e-01 -7.56792009e-01 -8.66235375e-01 7.18016744e-01
2.73104757e-03 2.50236392e-01 8.31135869e-01 2.07127243e-01
-8.43268394e-01 -4.68237489e-01 -1.04588437e+00 5.71142554e-01
1.16436779e+00 -8.85881245e-01 -2.65535533e-01 6.05278313e-01
5.29803991e-01 -4.51574236e-01 -1.18556058e+00 2.98481107e-01
8.38047206e-01 -5.37977219e-01 1.44544578e+00 -5.58685124e-01
3.66476595e-01 1.16722941e-01 -2.78783798e-01 -7.64681041e-01
-3.75288427e-01 -4.86531436e-01 -1.55023858e-01 1.73777008e+00
2.37044543e-01 -6.37460530e-01 9.91981506e-01 1.01523495e+00
2.96216339e-01 -1.05918097e+00 -8.32298636e-01 -8.52015853e-01
-2.85541024e-02 -1.59393549e-01 8.56326938e-01 8.83143306e-01
-5.38614035e-01 2.52023846e-01 -8.46558273e-01 9.58654508e-02
9.92919147e-01 1.63920313e-01 1.04117644e+00 -1.21952057e+00
-2.52719104e-01 -1.58479095e-01 -2.14756221e-01 -1.45888329e+00
2.90973008e-01 -7.91625917e-01 -3.85776609e-01 -1.21003127e+00
6.95707977e-01 -8.27269197e-01 -2.40431041e-01 6.26362801e-01
-5.37028491e-01 3.90867174e-01 1.86171442e-01 6.95933163e-01
-1.00916100e+00 6.99238479e-01 9.49216664e-01 -4.47135150e-01
3.10012311e-01 -4.95908484e-02 -7.92584717e-01 2.74280548e-01
4.99069124e-01 -2.07104728e-01 -4.52666432e-01 -5.14123142e-01
2.42260732e-02 -1.75612748e-01 4.49723899e-01 -6.10144496e-01
6.39023185e-01 1.13293529e-01 7.62656987e-01 -6.71805501e-01
6.65680707e-01 -7.05991983e-01 -1.02281734e-01 1.51494935e-01
-5.13651252e-01 1.02701023e-01 -2.13751402e-02 7.46998966e-01
-1.82654187e-01 3.83453280e-01 2.72724628e-01 1.46360099e-01
-1.01964593e+00 1.05054700e+00 1.37139052e-01 3.28874111e-01
9.48915660e-01 -7.36370087e-02 -6.11628830e-01 -6.20395064e-01
-2.62931675e-01 7.92390347e-01 2.73934424e-01 6.75796390e-01
8.20507824e-01 -1.64000630e+00 -9.33838844e-01 2.45639846e-01
6.82524502e-01 -2.01093987e-01 4.24538434e-01 5.83956301e-01
1.30251199e-01 4.60736930e-01 -1.07088629e-02 -6.40828252e-01
-1.64162767e+00 5.98154426e-01 -1.68416798e-02 -2.87901342e-01
-7.41420448e-01 9.10273433e-01 3.93599570e-01 -3.22217137e-01
4.53076214e-01 7.18172550e-01 2.07930177e-01 1.21102659e-02
7.12577522e-01 1.61863446e-01 -9.40662175e-02 -6.29991412e-01
-3.98102582e-01 5.19502759e-01 -7.64562011e-01 -5.60327061e-02
1.09487820e+00 -3.47097695e-01 2.59449966e-02 -1.62981495e-01
1.33112848e+00 -1.12507150e-01 -1.24528921e+00 -8.45496535e-01
-1.44293427e-01 -7.02520072e-01 -2.42373586e-01 -9.79238153e-01
-1.20629680e+00 6.10525012e-01 6.92495346e-01 -5.05989373e-01
1.25450206e+00 2.55095541e-01 1.22004473e+00 5.78924477e-01
4.37468857e-01 -1.39589715e+00 3.16462964e-01 1.38022289e-01
6.86156750e-01 -1.54421616e+00 2.23364472e-01 -5.59183776e-01
-9.20061409e-01 5.70580363e-01 7.33301282e-01 2.68908322e-01
2.81601071e-01 1.15674734e-01 3.21454518e-02 -1.54683277e-01
-8.78567338e-01 4.91524823e-02 6.36787295e-01 6.80523217e-01
1.55271646e-02 7.71162882e-02 1.88575700e-01 7.17501760e-01
-9.63209718e-02 -1.80222362e-01 -1.04685023e-01 6.20118618e-01
4.30196598e-02 -1.10488605e+00 -3.58612925e-01 7.75614858e-01
-4.60062325e-01 -4.43156064e-01 -5.61470449e-01 6.86484814e-01
-1.42091662e-01 7.67253160e-01 1.84281226e-02 -6.34029746e-01
1.92948252e-01 -9.93021391e-03 1.72704041e-01 -2.13533826e-02
-4.91760910e-01 -1.47877410e-01 9.34283286e-02 -7.26540565e-01
5.82462475e-02 -8.16218019e-01 -9.62928593e-01 -6.38065040e-01
-2.32019611e-02 1.26248151e-01 3.16457778e-01 8.58782887e-01
8.35735202e-01 1.71861537e-02 5.77760041e-01 -7.97804296e-02
-6.92480624e-01 -7.03513205e-01 -3.72338951e-01 8.76754940e-01
6.37454018e-02 -5.11931717e-01 8.61348063e-02 -4.51944396e-02] | [14.631180763244629, 0.9267781972885132] |
923260b6-c75f-4cb2-aca1-08d83a897908 | on-the-usefulness-of-personality-traits-in | null | null | https://aclanthology.org/2021.ranlp-main.62 | https://aclanthology.org/2021.ranlp-main.62.pdf | On the Usefulness of Personality Traits in Opinion-oriented Tasks | We use a deep bidirectional transformer to extract the Myers-Briggs personality type from user-generated data in a multi-label and multi-class classification setting. Our dataset is large and made up of three available personality datasets of various social media platforms including Reddit, Twitter, and Personality Cafe forum. We induce personality embeddings from our transformer-based model and investigate if they can be used for downstream text classification tasks. Experimental evidence shows that personality embeddings are effective in three classification tasks including authorship verification, stance, and hyperpartisan detection. We also provide novel and interpretable analysis for the third task: hyperpartisan news classification. | ['Arjun Mukherjee', 'Dainis Boumber', 'Eduard Dragut', 'Marjan Hosseinia'] | null | null | https://aclanthology.org/2021.ranlp-1.62 | https://aclanthology.org/2021.ranlp-1.62.pdf | ranlp-2021-9 | ['news-classification', 'authorship-verification'] | ['natural-language-processing', 'natural-language-processing'] | [-4.62016374e-01 1.55556664e-01 -2.36531764e-01 -5.41797757e-01
-4.98725146e-01 -9.16578829e-01 7.48844385e-01 3.79154235e-01
-2.90463805e-01 6.03922486e-01 6.25025749e-01 -3.71257402e-02
-1.22267991e-01 -7.21544027e-01 -1.25576854e-01 -4.61469203e-01
1.05373509e-01 8.51180792e-01 -2.54911035e-01 -3.16688657e-01
2.82118052e-01 1.17607333e-01 -1.13649464e+00 4.63559926e-01
6.16253316e-01 1.06504619e+00 -6.60086274e-01 6.62469387e-01
3.47682893e-01 7.62828112e-01 -6.19954705e-01 -1.26106739e+00
1.26344740e-01 1.52468801e-01 -1.09291291e+00 -2.15346087e-02
3.93404663e-01 -2.14968503e-01 -2.94834465e-01 9.92541969e-01
5.81113577e-01 -2.69877821e-01 1.05328906e+00 -1.59402788e+00
-1.16195691e+00 1.03454149e+00 -7.51530111e-01 1.40638486e-01
3.13700736e-01 -1.26955360e-01 1.75432754e+00 -7.85998642e-01
6.53880179e-01 1.51292825e+00 7.82942653e-01 2.22407788e-01
-1.29121602e+00 -7.43443489e-01 -2.56729066e-01 1.09578306e-02
-9.69522655e-01 -1.98932216e-01 9.25856233e-01 -7.45541215e-01
4.05864269e-01 2.05762997e-01 4.48383838e-01 2.06370401e+00
2.78965145e-01 9.22617018e-01 1.91604614e+00 9.17780474e-02
-4.18549389e-01 5.79785228e-01 9.43383694e-01 9.43551600e-01
3.31094384e-01 -1.16416700e-01 -5.44504881e-01 -7.66077876e-01
3.60375196e-01 -4.11075324e-01 1.73670784e-01 3.25079769e-01
-1.30950201e+00 1.01865566e+00 -4.53770049e-02 1.05569195e-02
-2.38477573e-01 -1.36504233e-01 9.55641389e-01 5.43089569e-01
8.17407370e-01 7.81471908e-01 -4.13461804e-01 -2.23174140e-01
-6.99957252e-01 2.80123711e-01 1.02640283e+00 8.33732903e-01
5.48522711e-01 -3.89094144e-01 -5.55146456e-01 1.15896714e+00
1.29132196e-01 4.16730046e-01 6.41866088e-01 -8.52329552e-01
2.49864295e-01 7.10930943e-01 -7.13811954e-03 -1.22836721e+00
-7.53261626e-01 -4.61159796e-01 -8.57742727e-01 -3.56199414e-01
4.42340344e-01 -1.81047469e-01 -1.50561497e-01 1.46293581e+00
2.09398448e-01 -3.78998101e-01 -1.77181616e-01 6.76128328e-01
1.08150136e+00 1.89169392e-01 -9.54668298e-02 1.69951230e-01
1.94368017e+00 -9.30802763e-01 -4.58854079e-01 1.18380794e-02
5.17713845e-01 -3.95648569e-01 1.15773904e+00 3.93268496e-01
-7.99210429e-01 -4.18467700e-01 -7.32599616e-01 -4.91550475e-01
-4.63060915e-01 4.81954753e-01 6.35882616e-01 7.12819278e-01
-6.51615202e-01 6.74909532e-01 5.59371598e-02 -1.90938517e-01
5.87852180e-01 1.28276065e-01 -4.62277949e-01 3.44068348e-01
-1.54667377e+00 7.26777911e-01 3.58451456e-02 -3.41108769e-01
-4.90644395e-01 -3.69571567e-01 -4.53358322e-01 1.53365597e-01
-2.14575976e-01 -8.13821971e-01 8.34213257e-01 -9.49964345e-01
-1.39130425e+00 1.48351872e+00 8.62848833e-02 -1.99937388e-01
5.99674225e-01 2.94912726e-01 -4.29127246e-01 -8.41883048e-02
4.39336985e-01 1.85216546e-01 8.41696322e-01 -8.58253658e-01
-3.37381542e-01 -5.12396216e-01 -9.88381058e-02 -1.99351311e-01
-1.11313283e+00 4.93639886e-01 2.82575399e-01 -5.76988399e-01
-4.38784480e-01 -1.13426363e+00 4.41139340e-01 -6.25956357e-01
-9.58270609e-01 -8.57597947e-01 5.51871657e-01 -8.82510006e-01
1.18919969e+00 -1.90899956e+00 1.90662369e-01 1.39850244e-01
7.71501184e-01 -2.72843182e-01 -1.01782762e-01 2.94702768e-01
6.54251501e-02 5.65142572e-01 5.04531860e-01 -7.01430202e-01
6.00359857e-01 -3.30496401e-01 -3.35447252e-01 4.90191162e-01
1.26966596e-01 1.03604007e+00 -8.32042873e-01 -6.03011012e-01
-6.22290015e-01 1.25936151e-03 -3.10994029e-01 5.91380596e-02
1.49384752e-01 2.27446988e-01 -3.64138305e-01 7.57307410e-01
4.27610934e-01 -6.49348259e-01 3.43703866e-01 -2.50797480e-01
1.48747370e-01 4.23436999e-01 -3.90551955e-01 7.84576654e-01
-2.55518079e-01 8.89288187e-01 -3.88839096e-02 -3.54770541e-01
1.03642190e+00 8.73016492e-02 2.70816892e-01 -1.54193819e-01
7.70267010e-01 -7.54784197e-02 1.13826133e-01 -3.68707150e-01
9.60578918e-01 -1.04331322e-01 -8.11357856e-01 1.00922358e+00
-3.23639810e-02 6.00109577e-01 1.35577723e-01 4.35590446e-01
1.10476029e+00 -4.00051147e-01 2.91185409e-01 -6.27190173e-01
2.72537470e-01 -2.96383619e-01 7.45462835e-01 5.27193546e-01
-5.41060209e-01 1.75815672e-01 1.32984281e+00 -3.53860825e-01
-1.33509576e+00 -6.32364273e-01 -3.44935983e-01 1.84683979e+00
-4.17798102e-01 -6.72491968e-01 -2.80922979e-01 -1.06415331e+00
4.53288704e-01 2.18323186e-01 -1.06112504e+00 8.65782350e-02
-1.18757233e-01 -1.05080318e+00 1.16899025e+00 3.92681271e-01
3.13421875e-01 -7.57635355e-01 2.03707024e-01 -1.54924870e-01
-4.82403964e-01 -1.24154055e+00 -7.77443707e-01 -2.11675689e-01
-2.74202049e-01 -1.25599158e+00 -4.87499028e-01 -6.23658657e-01
3.59122276e-01 -2.21034765e-01 1.21190727e+00 -1.78379402e-01
3.10780972e-01 2.66990721e-01 -2.03600317e-01 -2.60430276e-01
-6.94709599e-01 4.34450597e-01 3.99058014e-01 2.65684038e-01
5.72422683e-01 -4.61492479e-01 -2.39870176e-01 4.39838916e-01
-2.66895205e-01 1.56068327e-02 4.12008256e-01 1.24151778e+00
-2.42679849e-01 -2.03914031e-01 7.96669424e-01 -1.13931620e+00
1.37614024e+00 -7.90694416e-01 -2.90906392e-02 7.53644854e-02
-8.15664053e-01 2.80702300e-02 8.54861796e-01 -6.76789343e-01
-7.26654887e-01 -4.93540585e-01 -1.55203817e-02 -2.81426072e-01
6.24467842e-02 4.34713930e-01 3.46288341e-03 1.28783196e-01
6.32919788e-01 -4.39838022e-02 -9.54172239e-02 -4.11168396e-01
9.59620252e-02 1.43047261e+00 1.58974215e-01 -8.90936136e-01
6.53779149e-01 2.74856925e-01 -6.82683736e-02 -6.59642875e-01
-1.17590344e+00 -3.36017877e-01 -7.39224851e-01 -7.96990395e-02
6.17104232e-01 -8.64825249e-01 -1.57407117e+00 5.50893784e-01
-1.22566688e+00 -9.30695562e-04 3.18310708e-01 -1.47916466e-01
-1.86192527e-01 6.05179906e-01 -1.23350966e+00 -6.17491603e-01
-7.20559597e-01 -7.90253758e-01 9.99227762e-01 -1.36752367e-01
-7.93164909e-01 -1.21442401e+00 7.85447508e-02 8.80253732e-01
-3.60982656e-03 1.90833628e-01 1.17199099e+00 -1.02251339e+00
1.27996951e-01 -1.24713898e-01 -5.41250885e-01 4.10499275e-01
-3.56851429e-01 2.19860509e-01 -1.17790842e+00 -2.41346732e-01
-6.04471743e-01 -7.87129164e-01 9.43849206e-01 -1.90174907e-01
1.36031413e+00 -6.83748305e-01 -1.86747685e-01 4.11441296e-01
7.93841004e-01 -8.08454931e-01 2.63845384e-01 4.60272878e-01
8.57382596e-01 7.93325424e-01 2.81978190e-01 8.34320843e-01
1.00216460e+00 3.66000801e-01 -2.75526475e-02 1.49223119e-01
2.76411384e-01 -3.74147147e-01 4.69881654e-01 9.80167508e-01
-2.27763783e-02 -2.10685268e-01 -1.08656645e+00 4.54729021e-01
-1.79040217e+00 -1.15637648e+00 -4.76610899e-01 1.60861802e+00
9.89448369e-01 1.85766086e-01 7.82029331e-01 1.40896022e-01
8.14620972e-01 6.08315617e-02 -4.99708980e-01 -6.97945654e-01
-3.49012434e-01 -1.68186307e-01 5.20446301e-01 8.24249536e-02
-1.22920966e+00 8.08247566e-01 6.52588844e+00 3.67039919e-01
-7.71850467e-01 3.63580883e-01 8.55109513e-01 1.90608110e-02
-2.53751874e-01 -3.63832921e-01 -9.33657467e-01 7.27163732e-01
1.08338869e+00 -4.17274445e-01 2.75123298e-01 8.29266727e-01
4.75275144e-03 3.73133957e-01 -1.27637720e+00 9.79002535e-01
2.55140305e-01 -9.84116852e-01 -1.93846583e-01 4.99063164e-01
6.12026393e-01 4.68760245e-02 3.21816355e-01 6.26685083e-01
5.92273951e-01 -9.30458009e-01 7.22241819e-01 2.57682383e-01
8.35111320e-01 -6.59191668e-01 5.69622695e-01 2.69884795e-01
-6.23905420e-01 -5.31920612e-01 -1.48533806e-01 -1.80202201e-01
-2.81272650e-01 8.35834980e-01 -9.08428729e-01 2.13843331e-01
5.63069344e-01 1.02425241e+00 -9.50760722e-01 2.62292087e-01
-2.10066989e-01 6.84707820e-01 1.33285150e-01 -4.37354088e-01
-4.67340909e-02 -1.54398605e-01 7.00107574e-01 1.21360350e+00
7.87601806e-03 -3.02067578e-01 1.15569957e-01 1.01285923e+00
-7.21457958e-01 8.66278186e-02 -3.99240077e-01 -5.67735970e-01
4.99316841e-01 1.96576762e+00 -4.92512256e-01 -4.73849356e-01
-1.55933321e-01 1.15787172e+00 8.01858664e-01 -7.35768080e-02
-9.21318829e-01 -3.56269568e-01 8.63408506e-01 4.87648211e-02
-1.61143884e-01 -8.14295635e-02 -5.50819099e-01 -1.45073771e+00
-4.31896180e-01 -9.33898807e-01 5.33944249e-01 -6.18107498e-01
-2.31719112e+00 4.83240843e-01 -3.60953659e-01 -8.52888703e-01
7.28437910e-03 -8.81907344e-01 -6.70710683e-01 8.00126374e-01
-1.24805164e+00 -1.75801241e+00 -2.32628182e-01 2.95164853e-01
-4.27602008e-02 -4.26454335e-01 8.86852682e-01 3.42131764e-01
-1.08018613e+00 9.20959055e-01 2.34637812e-01 8.11569571e-01
1.20973551e+00 -1.57974875e+00 4.14608389e-01 6.99943155e-02
-2.96577036e-01 6.73623919e-01 5.17501295e-01 -5.85167766e-01
-1.26981175e+00 -1.08614028e+00 1.08158696e+00 -1.03675270e+00
1.24101460e+00 -8.01079571e-01 -4.57403034e-01 9.34278727e-01
3.90574038e-01 -4.26251471e-01 1.34095848e+00 9.84110117e-01
-8.26606810e-01 1.19014740e-01 -1.04882050e+00 3.71377945e-01
8.60398948e-01 -7.00594902e-01 -5.59852123e-01 5.65130472e-01
2.48673245e-01 -5.40922629e-03 -1.54894471e+00 -1.75847501e-01
9.21323240e-01 -7.82484710e-01 6.40092015e-01 -1.09702373e+00
1.22021508e+00 5.78893840e-01 1.89765811e-01 -1.68816960e+00
-1.10409248e+00 -4.78706598e-01 -1.98102593e-01 1.59550250e+00
4.45331603e-01 -9.50389922e-01 5.50826371e-01 4.96066391e-01
2.17642486e-01 -6.74815893e-01 -5.77236056e-01 -5.86706758e-01
5.25160372e-01 1.96255043e-01 5.97935557e-01 1.64423299e+00
5.24877667e-01 1.13691235e+00 -6.11943305e-01 -1.64467618e-01
9.51895893e-01 4.32000667e-01 7.93081760e-01 -1.94949150e+00
-3.26607794e-01 -7.49701738e-01 -1.82147011e-01 -6.74480259e-01
9.86943126e-01 -1.31848681e+00 -6.83892369e-01 -1.31601834e+00
8.49809885e-01 -5.96639276e-01 -1.71756566e-01 5.37416279e-01
-1.12425126e-01 3.61708552e-01 -1.50364324e-01 3.06469023e-01
-4.68166828e-01 4.19191599e-01 1.34413266e+00 -4.15087312e-01
3.94586533e-01 -2.86955446e-01 -1.23110151e+00 5.19593894e-01
8.21081221e-01 -3.79997939e-01 1.45501226e-01 -1.44363632e-02
8.72207522e-01 -1.56636506e-01 4.16745216e-01 -1.65101573e-01
-1.57787591e-01 -4.99466769e-02 4.83589441e-01 -1.93642855e-01
4.48175550e-01 -1.71086743e-01 -5.27038097e-01 -1.11580797e-01
-6.30131781e-01 2.38723129e-01 -1.36285305e-01 5.13006628e-01
2.77392387e-01 4.72816937e-02 7.91834474e-01 -6.42985255e-02
-3.64710316e-02 4.26288188e-01 -4.18816864e-01 4.37083960e-01
8.21101069e-01 1.42692268e-01 -1.01185274e+00 -3.05572599e-01
-6.72908545e-01 2.56302744e-01 5.30659735e-01 6.17265403e-01
2.37800300e-01 -1.43979692e+00 -1.09681571e+00 1.35718346e-01
1.81872770e-01 -9.80558693e-01 -1.19927324e-01 1.16452217e+00
2.44757175e-01 1.80853888e-01 -5.33832490e-01 -2.21439123e-01
-1.61927712e+00 3.71803850e-01 1.50580645e-01 -4.12446767e-01
-1.81183726e-01 6.12871766e-01 5.91543794e-04 -6.54445589e-01
-2.76186526e-01 4.00881805e-02 -4.16886568e-01 5.73403060e-01
2.44934633e-01 7.55846977e-01 -2.26704553e-01 -7.99118757e-01
-1.93562597e-01 -2.00181410e-01 -1.65467471e-01 -7.86138140e-03
1.19502401e+00 -1.63329840e-01 -7.40316033e-01 8.10294926e-01
1.33796358e+00 3.17634046e-01 -5.02585351e-01 -2.94051543e-02
1.02521434e-01 -3.62704903e-01 3.58504825e-03 -7.27123022e-01
-5.59831440e-01 8.54346693e-01 -1.46484345e-01 7.70719707e-01
3.61846507e-01 1.06399879e-01 9.46119010e-01 1.73142314e-01
1.23608686e-01 -1.52604854e+00 3.24392617e-01 9.51124132e-01
7.80587375e-01 -1.30387557e+00 -5.53389080e-02 -1.59036458e-01
-9.96950090e-01 1.17892575e+00 7.04214394e-01 1.38750613e-01
6.86133325e-01 7.38248080e-02 -2.22562596e-01 -3.87529790e-01
-1.02155519e+00 1.45234808e-01 3.81144524e-01 1.82463005e-01
6.34242058e-01 3.86708319e-01 -3.20312083e-01 1.29656637e+00
-6.53911412e-01 -5.36107004e-01 9.39806402e-01 5.97869866e-02
-1.42178044e-01 -1.06865478e+00 -2.89884508e-01 1.06661308e+00
-5.72502077e-01 -8.06770325e-02 -1.06340802e+00 2.95428097e-01
-6.58089742e-02 7.09118068e-01 -2.36332323e-02 -8.92843843e-01
5.13459416e-03 2.46391177e-01 4.52901304e-01 -4.92479742e-01
-1.03822136e+00 -3.59109372e-01 8.26870859e-01 1.06448591e-01
-4.66669127e-02 -8.37933838e-01 -6.40543699e-01 -1.07997000e+00
-4.68150079e-02 1.75211549e-01 2.53534049e-01 6.26018584e-01
5.07904887e-01 2.89525211e-01 9.97333467e-01 -5.45301020e-01
-8.82256746e-01 -1.37274754e+00 -7.75817037e-01 8.90387058e-01
5.79286180e-02 -7.91802406e-01 -4.45846081e-01 -2.63153791e-01] | [9.3975830078125, 10.344913482666016] |
b3d5d6c2-e2c0-4913-9fe8-0e7ac3318fce | simultaneous-self-supervised-reconstruction | 2210.01696 | null | https://arxiv.org/abs/2210.01696v3 | https://arxiv.org/pdf/2210.01696v3.pdf | Simultaneous self-supervised reconstruction and denoising of sub-sampled MRI data with Noisier2Noise | Most existing methods for Magnetic Resonance Imaging (MRI) reconstruction with deep learning assume that a high signal-to-noise ratio (SNR), fully sampled sampled dataset exists and use fully supervised training. In many circumstances, however, such a dataset does not exist and may be highly impractical to acquire. Recently, a number of self-supervised methods for MR reconstruction have been proposed, which require a training dataset with sub-sampled k-space data only. However, existing methods do not denoise sampled data, so are only applicable in the high SNR regime. In this work, we propose a method based on Noisier2Noise and Self-Supervised Learning via Data Undersampling (SSDU) that trains a network to reconstruct clean images from sub-sampled, noisy training data. To our knowledge, our approach is the first that simultaneously denoises and reconstructs images in an entirely self-supervised manner. Our method is applicable to any network architecture, has a strong mathematical basis, and is straight-forward to implement. We evaluate our method on the multi-coil fastMRI brain dataset and find that it performs competitively with a network trained on clean, fully sampled data and substantially improves over methods that do not explicitly remove measurement noise. | ['Mark Chiew', 'Charles Millard'] | 2022-10-04 | null | null | null | null | ['mri-reconstruction'] | ['computer-vision'] | [ 7.44160891e-01 1.03466049e-01 4.97653931e-02 -5.81937611e-01
-1.02617216e+00 -1.10207491e-01 4.30101573e-01 -2.24110126e-01
-5.09941459e-01 8.76049697e-01 3.36113691e-01 -2.54292846e-01
-3.99091542e-01 -5.85310161e-01 -7.48008907e-01 -1.01111698e+00
-1.85312167e-01 3.59169155e-01 7.24573731e-02 -1.71408415e-01
-3.16083550e-01 3.66672605e-01 -1.17332876e+00 1.93628758e-01
6.76120579e-01 8.35845172e-01 3.61312509e-01 5.61966956e-01
2.70895958e-01 1.13804996e+00 -2.65740037e-01 7.93489367e-02
2.97941893e-01 -6.91569269e-01 -9.00251865e-01 2.06993092e-02
4.38318491e-01 -5.89466453e-01 -7.40535498e-01 1.07092631e+00
9.87185299e-01 7.73833245e-02 4.11307961e-01 -3.67679149e-01
-4.57103997e-01 8.67750764e-01 -3.75911206e-01 5.87504804e-01
-9.24610943e-02 1.44939497e-01 2.51513571e-01 -7.02166796e-01
6.40438855e-01 7.20059156e-01 1.05320871e+00 6.13749206e-01
-1.65118837e+00 -5.75117886e-01 -3.71920109e-01 3.65697704e-02
-1.00120592e+00 -8.11174870e-01 9.73563910e-01 -2.15870410e-01
5.94718456e-01 2.04641968e-01 6.01207018e-01 1.22464514e+00
3.38323206e-01 6.67814076e-01 1.90489852e+00 -4.74134743e-01
5.31593919e-01 -3.87311906e-01 -9.80570689e-02 3.31236809e-01
1.75288860e-02 4.23382163e-01 -3.82089019e-01 -7.98280910e-02
9.62297678e-01 1.99091919e-02 -4.30470467e-01 -5.74583948e-01
-1.51637113e+00 7.07513034e-01 4.63456005e-01 7.51741290e-01
-7.66463459e-01 1.35997906e-01 5.45838892e-01 4.14866924e-01
5.92352569e-01 4.33078200e-01 -1.25508115e-01 6.14519417e-02
-1.52084649e+00 -6.80446252e-02 6.82329834e-01 4.36375886e-01
3.40428412e-01 4.30145264e-01 1.73761308e-01 8.40211093e-01
-8.94508287e-02 4.92006004e-01 9.35572267e-01 -1.22387326e+00
-1.29227623e-01 -5.51829159e-01 -1.45297036e-01 -7.77310312e-01
-6.51176512e-01 -8.02944303e-01 -1.44264996e+00 2.32903898e-01
2.10112125e-01 -2.59094000e-01 -9.06101346e-01 1.70002067e+00
1.94059268e-01 4.32626516e-01 2.16068149e-01 1.18227363e+00
9.54464614e-01 1.67269513e-01 -2.02705890e-01 -6.22025371e-01
8.78079474e-01 -6.64271474e-01 -1.16628695e+00 -3.02205354e-01
1.32725388e-01 -7.21339226e-01 6.31532609e-01 8.57750714e-01
-1.37588489e+00 -5.26497543e-01 -1.20812428e+00 3.20618421e-01
1.56500846e-01 -2.44754136e-01 7.48291194e-01 7.13737905e-01
-1.32035637e+00 1.14086688e+00 -1.10653698e+00 -8.66597444e-02
4.49972242e-01 3.56918216e-01 -6.56116724e-01 -3.78524721e-01
-1.37245560e+00 1.19865096e+00 3.69669646e-01 2.62635380e-01
-1.13425803e+00 -6.33895636e-01 -8.74127984e-01 -4.76415068e-01
3.20348412e-01 -5.57691693e-01 1.27483594e+00 -1.13545346e+00
-1.66827631e+00 5.37371159e-01 1.92890674e-01 -7.95941114e-01
4.66182381e-01 8.49562809e-02 -5.44980526e-01 5.92642784e-01
3.64957552e-04 4.07867491e-01 1.19438398e+00 -1.31931436e+00
2.86637813e-01 -3.51521134e-01 -2.77867734e-01 -2.98432615e-02
2.77499240e-02 -1.50662765e-01 1.94494814e-01 -8.75921905e-01
6.08628452e-01 -7.26584613e-01 -5.72803795e-01 -2.71326989e-01
-1.69712424e-01 6.63087785e-01 3.34883332e-01 -9.54153299e-01
7.65948653e-01 -2.02785993e+00 -1.23581320e-01 2.79084533e-01
4.46085632e-01 3.64034116e-01 -1.12332709e-01 1.63685486e-01
-6.53761387e-01 -4.16163743e-01 -8.24605405e-01 -1.77533582e-01
-4.04732138e-01 4.03263062e-01 -4.20317762e-02 9.52711761e-01
-1.94458440e-02 8.55953693e-01 -1.33778036e+00 -3.55526686e-01
4.67829645e-01 7.97242582e-01 -4.34655219e-01 1.79636478e-01
4.64922547e-01 9.04382706e-01 -9.99087021e-02 1.36822447e-01
9.47211325e-01 -2.09361538e-01 4.68961298e-01 -5.18308699e-01
2.79169399e-02 1.57601506e-01 -1.24727535e+00 1.91094792e+00
-5.61180532e-01 3.55865389e-01 5.09432197e-01 -1.72637403e+00
7.28312910e-01 5.10068476e-01 9.54268456e-01 -8.75130177e-01
3.92933488e-02 5.42981923e-01 1.18958212e-01 -6.34305179e-01
-4.65162881e-02 -9.44931269e-01 3.05873036e-01 4.57009971e-01
4.40330118e-01 -3.15626770e-01 -1.53087005e-01 7.46223852e-02
1.53950715e+00 5.80066778e-02 3.78295600e-01 -4.72253233e-01
2.73288697e-01 -2.86220640e-01 3.94811779e-01 1.16926682e+00
-4.06473577e-01 1.00345552e+00 -1.89838763e-02 -4.29972827e-01
-1.18742418e+00 -1.06092560e+00 -6.12474740e-01 4.65075999e-01
-1.74402833e-01 8.98199454e-02 -9.72705007e-01 -3.95553201e-01
-4.30959344e-01 3.93423110e-01 -4.97470468e-01 -1.71786219e-01
-7.76432097e-01 -1.08045042e+00 5.59342265e-01 3.73358965e-01
4.36239094e-01 -1.11819029e+00 -4.99433905e-01 6.20328128e-01
-2.68441856e-01 -1.11142206e+00 -1.81657419e-01 7.04075098e-01
-1.23976564e+00 -9.73813474e-01 -9.02972043e-01 -6.29631698e-01
6.48205221e-01 1.94840342e-01 1.00354922e+00 -1.12160556e-01
-1.30030721e-01 2.62172341e-01 -2.59180009e-01 1.19148605e-01
-6.64245427e-01 -2.28864685e-01 2.21671909e-01 -6.66252747e-02
-3.20888199e-02 -1.01538551e+00 -6.88389778e-01 7.18287453e-02
-1.33008063e+00 7.04524368e-02 8.49437833e-01 1.34819567e+00
7.50107348e-01 3.07892323e-01 9.52285826e-01 -1.12816238e+00
4.50053811e-01 -5.09226799e-01 -5.22436574e-02 -2.66625166e-01
-4.89338338e-01 2.42731661e-01 8.69759679e-01 -3.55667681e-01
-9.64628100e-01 2.40426600e-01 -7.08522379e-01 -3.29092771e-01
-4.37913984e-01 5.39743185e-01 7.59503022e-02 -3.67692381e-01
1.02845252e+00 4.73132998e-01 5.09312391e-01 -5.04070997e-01
4.15289432e-01 6.95316672e-01 9.22199070e-01 -3.12586576e-01
6.80587709e-01 8.57750595e-01 1.15342453e-01 -7.23723352e-01
-8.27124536e-01 -3.43887120e-01 -6.26905322e-01 -1.48400798e-01
6.01531863e-01 -8.27506423e-01 -1.42879918e-01 5.41314065e-01
-6.41729116e-01 -4.73863721e-01 -5.18316269e-01 9.47913229e-01
-9.26426649e-01 7.83804595e-01 -8.50631416e-01 -4.58532989e-01
-5.15152276e-01 -1.27607822e+00 6.86158180e-01 -3.58318299e-01
2.95656342e-02 -9.24422443e-01 8.53009969e-02 2.68751949e-01
9.83569920e-01 4.07530248e-01 3.75668555e-01 -4.57998216e-01
-9.97053683e-02 -2.81508453e-02 1.29910082e-01 8.68639350e-01
3.41248631e-01 -9.83596683e-01 -1.02358150e+00 -5.49571872e-01
8.17798495e-01 -5.94636977e-01 9.59316432e-01 8.00429523e-01
1.39719903e+00 -1.17752872e-01 1.87199339e-01 7.43899643e-01
1.27339554e+00 -2.52885014e-01 8.56591582e-01 1.35666132e-01
3.68983507e-01 3.52854848e-01 -4.54468746e-03 3.23772281e-02
1.00152545e-01 3.75021070e-01 3.89837444e-01 -4.53886747e-01
-5.04833281e-01 6.82226568e-02 7.26471394e-02 1.39772761e+00
-2.39650942e-02 4.40239102e-01 -6.93967640e-01 5.82311690e-01
-1.46892774e+00 -1.02487850e+00 -8.52438509e-02 2.05541635e+00
1.27562261e+00 -4.91025066e-03 -7.51187205e-02 5.57614625e-01
6.06711447e-01 1.98532015e-01 -6.39335692e-01 1.99052200e-01
-2.46829063e-01 7.68942237e-01 6.82505786e-01 5.44127703e-01
-1.25029135e+00 3.78115356e-01 6.94352245e+00 7.35082030e-01
-1.29427159e+00 6.78602576e-01 6.57285094e-01 -8.95432755e-02
-9.00046602e-02 -2.54512668e-01 1.53267577e-01 4.01389360e-01
1.23861384e+00 4.06386048e-01 7.77646363e-01 5.80675721e-01
2.98115611e-01 -4.49481905e-02 -8.92812312e-01 1.10638869e+00
2.06396237e-01 -1.13753819e+00 -5.00141621e-01 -2.37856552e-01
8.05224538e-01 3.55536938e-01 -7.91600421e-02 -6.81418478e-02
1.13533787e-01 -1.17631686e+00 4.02850598e-01 6.86984360e-01
8.73082638e-01 -5.96861899e-01 8.07854235e-01 5.42465329e-01
-4.41793233e-01 1.89619988e-01 -2.76043266e-01 2.28359088e-01
3.27074260e-01 1.20409417e+00 -4.38649714e-01 6.08369410e-01
6.92805350e-01 7.08544016e-01 -2.20802188e-01 1.14032173e+00
-9.69115570e-02 7.94526756e-01 -2.07638696e-01 5.92842460e-01
1.07606575e-01 8.94924030e-02 5.41261792e-01 1.13230634e+00
1.62559748e-01 2.92135924e-01 1.06106587e-01 4.61284459e-01
1.25463724e-01 -1.22132942e-01 -4.67113942e-01 4.34994072e-01
1.43871056e-02 1.17857349e+00 -7.10208893e-01 -4.04747367e-01
-2.74365157e-01 9.17200327e-01 9.01280046e-02 2.44499251e-01
-4.23330992e-01 -2.06846863e-01 -9.20514613e-02 3.71616960e-01
2.73423821e-01 -1.66405648e-01 -2.78666466e-01 -1.37158883e+00
-6.72314391e-02 -1.20804679e+00 1.42035395e-01 -7.53474355e-01
-1.59392309e+00 7.91451454e-01 4.81533632e-02 -1.06616485e+00
-4.26515073e-01 -3.85718912e-01 -1.27655521e-01 7.46727586e-01
-1.86623740e+00 -6.92125559e-01 -1.49967387e-01 6.61905408e-01
1.44659102e-01 1.99901406e-02 8.77365410e-01 6.75773203e-01
8.79345760e-02 1.74421966e-01 5.16507685e-01 1.20248035e-01
6.43845201e-01 -1.31886315e+00 2.08002269e-01 8.51388276e-01
-1.81548133e-01 8.41960371e-01 8.14783335e-01 -5.65344155e-01
-1.59181082e+00 -9.56986666e-01 3.83609712e-01 6.85516000e-02
6.72062933e-01 -2.53816098e-01 -1.18477476e+00 3.71759266e-01
3.35078299e-01 7.27126956e-01 6.28480554e-01 -3.94093335e-01
5.57522178e-02 -2.32083946e-01 -1.42216158e+00 1.12591654e-01
9.63778675e-01 -6.00769520e-01 -8.78108382e-01 4.99333143e-01
2.35538960e-01 -6.92723572e-01 -1.24911916e+00 5.53216636e-01
3.01933557e-01 -1.01403451e+00 1.15726113e+00 -2.00800970e-01
2.80755162e-01 -3.08516502e-01 -1.62158289e-03 -1.69388545e+00
-3.23052704e-01 -5.87390602e-01 -2.78519094e-01 4.82761294e-01
2.57210084e-03 -6.55845523e-01 5.52530408e-01 1.82589427e-01
-3.17543685e-01 -6.33691967e-01 -1.26316381e+00 -7.45999277e-01
2.94473499e-01 -5.08220077e-01 4.70780343e-01 1.13258135e+00
-1.20768137e-01 -2.53796987e-02 -6.78135574e-01 -4.31187302e-02
1.18017495e+00 -1.32716656e-01 1.06492274e-01 -9.71814811e-01
-5.55109978e-01 -4.18178849e-02 -2.07658738e-01 -1.00518191e+00
2.41943732e-01 -1.02430165e+00 3.20120156e-01 -1.49668670e+00
1.41750321e-01 -5.30653954e-01 -4.30241704e-01 3.12318176e-01
1.09794602e-01 6.39202416e-01 -2.58658260e-01 3.25394601e-01
-4.15473104e-01 3.56292605e-01 1.43282259e+00 -6.41183183e-02
2.50845164e-01 -2.34469429e-01 -7.66264141e-01 6.27630889e-01
8.09924483e-01 -4.37521875e-01 -3.11038941e-01 -2.88762629e-01
-1.42969668e-01 3.23448122e-01 5.19463360e-01 -1.21367478e+00
2.04837769e-01 2.17727199e-01 6.46464050e-01 -3.18258315e-01
1.71470538e-01 -8.09767365e-01 4.94605631e-01 4.85848784e-01
-4.27800566e-01 -3.89306575e-01 -1.14307173e-01 3.36915940e-01
-2.00259060e-01 -3.99600714e-01 1.19505095e+00 -5.84320068e-01
-2.23118901e-01 6.67471886e-02 -3.27779740e-01 5.36096804e-02
4.20329899e-01 2.86665745e-02 1.04674585e-01 -3.95257354e-01
-1.10337114e+00 -4.06188548e-01 2.22871497e-01 -8.87005329e-02
7.59642720e-01 -1.43744671e+00 -8.01927924e-01 4.03442204e-01
-3.69456291e-01 4.63532237e-03 6.01604164e-01 1.10729313e+00
-3.65093857e-01 2.16958821e-01 -2.45766833e-01 -6.71906888e-01
-5.05067229e-01 6.26519144e-01 6.48551464e-01 -2.25068927e-01
-1.08070076e+00 2.89777040e-01 -1.98813304e-01 -8.02539885e-01
-4.35610414e-02 -1.91595539e-01 -9.75138098e-02 -3.05559337e-01
8.46836627e-01 5.21821491e-02 5.83984137e-01 -8.40456784e-01
-1.17692582e-01 2.65652061e-01 4.20203134e-02 -2.69833535e-01
1.77854419e+00 -5.28939674e-03 -2.25222424e-01 2.79209852e-01
1.25682616e+00 -3.51779580e-01 -1.29117668e+00 -6.31962895e-01
-2.39108935e-01 -3.46330822e-01 7.69992948e-01 -7.90583849e-01
-1.42477942e+00 8.07007253e-01 8.83505225e-01 2.03302115e-01
1.34847820e+00 -1.93080857e-01 8.71464491e-01 2.01595545e-01
4.87934351e-01 -1.09925914e+00 5.42830452e-02 3.37739259e-01
9.09615278e-01 -1.15964222e+00 7.83793256e-02 -8.51220265e-02
-3.70871872e-01 1.05506146e+00 1.06945530e-01 -3.68935108e-01
7.81240344e-01 6.16125882e-01 1.88998431e-01 -2.03719631e-01
-1.89531162e-01 -1.00402594e-01 5.16148806e-02 8.27553272e-01
4.06276196e-01 -6.82197437e-02 -3.29325318e-01 3.81418914e-01
-2.29584113e-01 4.81362224e-01 6.41186357e-01 1.09510696e+00
-2.69614369e-01 -9.83183265e-01 -6.74162090e-01 7.33859360e-01
-7.42597520e-01 -3.32941145e-01 4.41799760e-01 3.65344703e-01
5.60849570e-02 8.19487333e-01 -3.19194764e-01 4.91555547e-03
2.26924926e-01 -1.62091047e-01 7.64420450e-01 -3.85253668e-01
-4.52591836e-01 2.85763860e-01 -4.62857336e-02 -5.84407389e-01
-9.55920756e-01 -7.39820242e-01 -1.07882261e+00 -2.05965549e-01
-2.93641329e-01 1.98413413e-02 7.02560782e-01 9.63319361e-01
7.79574290e-02 6.97450757e-01 6.83066487e-01 -1.08126533e+00
-7.95554399e-01 -1.05882215e+00 -7.44229674e-01 4.67553198e-01
7.82896221e-01 -6.15470767e-01 -3.87451887e-01 -2.57044230e-02] | [13.370904922485352, -2.4508681297302246] |
33405817-1515-4760-abd1-ebb0e563fed5 | image-based-fire-detection-in-industrial | 2212.04786 | null | https://arxiv.org/abs/2212.04786v1 | https://arxiv.org/pdf/2212.04786v1.pdf | Image-Based Fire Detection in Industrial Environments with YOLOv4 | Fires have destructive power when they break out and affect their surroundings on a devastatingly large scale. The best way to minimize their damage is to detect the fire as quickly as possible before it has a chance to grow. Accordingly, this work looks into the potential of AI to detect and recognize fires and reduce detection time using object detection on an image stream. Object detection has made giant leaps in speed and accuracy over the last six years, making real-time detection feasible. To our end, we collected and labeled appropriate data from several public sources, which have been used to train and evaluate several models based on the popular YOLOv4 object detector. Our focus, driven by a collaborating industrial partner, is to implement our system in an industrial warehouse setting, which is characterized by high ceilings. A drawback of traditional smoke detectors in this setup is that the smoke has to rise to a sufficient height. The AI models brought forward in this research managed to outperform these detectors by a significant amount of time, providing precious anticipation that could help to minimize the effects of fires further. | ['Felix Nilsson', 'Fernando Alonso-Fernandez', 'Kevin Hernandez-Diaz', 'Joel Pålsson', 'Otto Zell'] | 2022-12-09 | null | null | null | null | ['fire-detection'] | ['time-series'] | [ 3.83917004e-01 -1.31799459e-01 2.88734496e-01 1.27003655e-01
-4.81286086e-02 -4.84173477e-01 6.62900746e-01 4.53432947e-01
-4.85986769e-01 4.95446444e-01 -1.95404723e-01 -7.29961842e-02
-9.00095850e-02 -1.15119159e+00 -3.39960158e-01 -6.83705866e-01
-1.49413019e-01 4.75590616e-01 4.86355633e-01 -2.26422161e-01
2.67696112e-01 6.94457531e-01 -1.84261191e+00 4.21949536e-01
3.34959894e-01 6.99719369e-01 8.12942147e-01 6.76660895e-01
2.51214743e-01 8.58949006e-01 -7.12413430e-01 1.98457047e-01
4.39291775e-01 -2.78742880e-01 -5.06527483e-01 -7.95853809e-02
1.00375548e-01 -3.04767787e-01 7.84506202e-02 6.48214519e-01
3.34818870e-01 4.25878055e-02 6.03202105e-01 -1.22423255e+00
-1.41738966e-01 5.50393879e-01 -3.08479667e-01 2.65157044e-01
1.60999194e-01 6.16921663e-01 6.58196747e-01 -4.35758412e-01
6.38290823e-01 1.06539607e+00 5.43411255e-01 2.40198046e-01
-1.17307448e+00 -7.19458163e-01 5.57668917e-02 1.44114450e-01
-1.34955180e+00 -1.00313328e-01 4.61628526e-01 -4.19354856e-01
1.26474369e+00 4.44623530e-01 9.14620936e-01 1.08589482e+00
1.18028887e-01 5.90641081e-01 1.18015361e+00 -6.91226423e-01
5.24898052e-01 2.02626213e-01 -3.62894952e-01 5.93290091e-01
5.53221703e-01 2.86409378e-01 -3.15542668e-01 1.64605111e-01
5.89303434e-01 3.48956794e-01 -1.31100863e-01 1.41318589e-01
-1.07413888e+00 7.45260894e-01 5.77038169e-01 8.07764053e-01
-7.31082380e-01 1.73265740e-01 1.08977176e-01 -4.56925109e-02
3.53814572e-01 8.96421075e-01 -3.49055856e-01 -2.01087482e-02
-9.97145772e-01 1.35198548e-01 8.77041221e-01 4.78604585e-01
4.32442844e-01 -2.08631337e-01 -1.11664049e-01 3.95052820e-01
1.32303730e-01 7.42037237e-01 3.88940461e-02 -7.15573072e-01
1.92780271e-01 8.47678363e-01 7.80191347e-02 -1.11029530e+00
-4.45162386e-01 -2.47343063e-01 -7.77131975e-01 8.62720430e-01
6.14539325e-01 1.17786359e-02 -9.49597061e-01 1.31164122e+00
1.12926953e-01 1.39358506e-01 -3.25829089e-01 1.00192344e+00
1.11736149e-01 7.59277880e-01 4.36246753e-01 -1.76488176e-01
1.42815113e+00 -4.89707083e-01 -4.38501418e-01 -4.29825515e-01
4.96557474e-01 -7.45039999e-01 8.35043252e-01 7.56504536e-01
-5.98472655e-01 -4.02860314e-01 -1.07468188e+00 6.72057629e-01
-7.07830667e-01 -1.18476517e-01 5.18059909e-01 6.70649886e-01
-4.96786475e-01 7.96972573e-01 -9.40955758e-01 -9.00001585e-01
4.80511934e-01 1.54829741e-01 -3.98521945e-02 -1.09862946e-01
-1.00777221e+00 1.31124938e+00 5.01543581e-01 3.54315788e-01
-1.00847888e+00 -5.26573122e-01 -2.24984080e-01 4.40376438e-02
6.69682145e-01 -4.11210537e-01 1.10807908e+00 -5.36223114e-01
-8.17035973e-01 7.87902117e-01 2.95033067e-01 -5.39362907e-01
8.06580842e-01 -2.43426502e-01 -3.45392972e-01 -8.24089497e-02
2.29161512e-02 5.04561305e-01 6.42820895e-01 -1.45352244e+00
-8.56263340e-01 -4.36476201e-01 1.04774334e-01 1.76518243e-02
-3.35518718e-01 1.58944115e-01 4.41072732e-01 -2.52769768e-01
-3.99872452e-01 -1.06792581e+00 -5.01331627e-01 -2.75438517e-01
-2.83744752e-01 9.71168131e-02 1.01583230e+00 -3.90385956e-01
1.03349030e+00 -1.79491472e+00 -2.16649368e-01 1.66098759e-01
1.49590205e-02 6.53709173e-01 1.44822478e-01 7.80554652e-01
1.41178995e-01 3.80198151e-01 -3.13524485e-01 1.24773256e-01
-1.70997098e-01 3.81977051e-01 -1.74311355e-01 2.91159302e-01
3.81901652e-01 4.15300608e-01 -1.06907105e+00 -3.89962465e-01
5.34418344e-01 3.84098649e-01 -8.83882344e-02 1.62895620e-01
-3.13760519e-01 5.65305166e-02 -3.77525538e-01 6.75178051e-01
5.35314441e-01 3.18879992e-01 1.17570944e-01 3.95290479e-02
-7.06555724e-01 -7.22872540e-02 -1.28615069e+00 1.14345539e+00
-4.70001489e-01 4.36025918e-01 1.38079852e-01 -7.59510398e-01
8.78605843e-01 3.55472952e-01 7.53419518e-01 -7.90499628e-01
1.21427678e-01 6.43136054e-02 -2.23057523e-01 -7.14135289e-01
2.35004947e-01 -3.77444357e-01 1.21796265e-01 4.39219207e-01
-6.20010376e-01 -4.14175510e-01 6.24540389e-01 -1.36421368e-01
1.64960146e+00 -4.45191227e-02 1.83359832e-01 -1.65002733e-01
3.61862153e-01 5.43324053e-01 3.36070687e-01 7.50036538e-01
-2.02034727e-01 4.67040241e-01 7.81050697e-02 -5.99194825e-01
-9.39148784e-01 -9.47258115e-01 8.61747637e-02 9.76719737e-01
-7.23798759e-03 -1.50417194e-01 -6.47885799e-01 -5.83157718e-01
8.79435688e-02 1.16668761e+00 -4.14845794e-01 -2.68622488e-01
-4.90305007e-01 -6.06298029e-01 4.50217515e-01 6.00923538e-01
5.30431569e-01 -1.67388415e+00 -1.45580959e+00 4.73109335e-01
-3.48582417e-02 -7.65950084e-01 2.79442787e-01 8.01753163e-01
-7.03145862e-01 -1.24235582e+00 -5.99270105e-01 -4.15647835e-01
3.73280615e-01 5.51704943e-01 1.03898311e+00 1.67794287e-01
-1.13272488e+00 2.06802472e-01 -4.93342191e-01 -1.22009778e+00
-6.70585275e-01 9.32724625e-02 7.66110513e-03 -5.32198787e-01
5.63496470e-01 -3.87664229e-01 -5.25355101e-01 3.35437030e-01
-1.27340400e+00 -1.53651461e-01 7.04604030e-01 1.67982206e-01
1.36256784e-01 6.63152158e-01 3.49792749e-01 -6.97552383e-01
6.30383372e-01 -2.70668834e-01 -5.95490217e-01 1.70093477e-01
-8.19747984e-01 -4.73765619e-02 4.65211183e-01 -2.41221905e-01
-9.00270641e-01 3.70619893e-01 2.24517584e-01 -1.80328235e-01
-6.95269406e-01 2.53973961e-01 1.68416902e-01 5.67974448e-01
7.93266058e-01 -2.48339444e-01 -1.74402639e-01 -4.29263890e-01
2.16978282e-01 6.03460968e-01 3.59822333e-01 -1.27331596e-02
1.14097464e+00 6.00780129e-01 1.57185435e-01 -1.15884221e+00
-5.90991616e-01 -7.76584327e-01 -8.85858357e-01 -6.84605598e-01
9.07296956e-01 -5.88115871e-01 -7.63856649e-01 4.35222596e-01
-1.11107683e+00 -4.04400885e-01 -4.89559025e-01 4.65478450e-01
-1.22522704e-01 -7.14963377e-02 -2.31594503e-01 -1.36283159e+00
-2.71111757e-01 -5.57009280e-01 7.64348030e-01 1.90689296e-01
-3.50988120e-01 -4.79057282e-01 2.31030986e-01 2.86196768e-01
5.76469183e-01 5.99830747e-01 4.29466635e-01 -2.22146511e-01
-6.33314073e-01 -3.06572616e-01 -6.24176636e-02 4.58185703e-01
3.22441459e-01 2.95073122e-01 -1.12227929e+00 -1.19475529e-01
1.54059425e-01 -1.29262656e-01 1.14824188e+00 1.05325438e-01
5.74564099e-01 6.37329295e-02 -5.61320364e-01 -5.68497553e-02
1.49101150e+00 4.70074236e-01 4.69187200e-01 8.28309596e-01
3.91595989e-01 5.86720943e-01 8.39845717e-01 5.55849493e-01
-3.02130163e-01 4.62012261e-01 1.03112268e+00 -3.34488064e-01
-1.43588617e-01 1.19005851e-01 3.81246388e-01 3.59450698e-01
-6.67746842e-01 -4.30742234e-01 -9.84698772e-01 3.83966208e-01
-1.76336002e+00 -1.33527720e+00 -4.61512566e-01 2.13791013e+00
3.48793447e-01 3.30209434e-01 2.44014531e-01 3.87258410e-01
5.97249448e-01 1.42795533e-01 -6.59650862e-02 -3.63653749e-01
1.62092924e-01 7.68572278e-03 6.97165132e-01 1.57630965e-02
-1.29500222e+00 5.81903338e-01 5.96822691e+00 2.54608691e-01
-1.01294935e+00 -1.45374969e-01 2.85489738e-01 -2.37491474e-01
2.72314399e-01 1.65389590e-02 -5.95075607e-01 3.72671902e-01
9.18740094e-01 9.82567444e-02 6.79530263e-01 8.86727214e-01
6.04294360e-01 -5.07140815e-01 -9.44302917e-01 3.36091846e-01
-9.03893039e-02 -9.88404572e-01 -5.42583108e-01 1.00066282e-01
2.17027396e-01 1.58999532e-01 -4.00410742e-01 2.09858328e-01
5.48475742e-01 -1.00803828e+00 6.81563914e-01 5.45600891e-01
3.06221366e-01 -7.38088608e-01 7.91808546e-01 7.19093680e-01
-1.14860034e+00 -5.08238435e-01 -4.39645857e-01 -4.55990791e-01
2.15945259e-01 7.18053699e-01 -1.41643476e+00 3.45851868e-01
7.70895541e-01 6.54618070e-02 -5.34972548e-01 1.13818085e+00
-3.29309523e-01 7.10763097e-01 -8.15588772e-01 -3.00257713e-01
2.92876333e-01 -2.06368923e-01 4.65198398e-01 1.44580054e+00
3.24419647e-01 5.01283854e-02 3.07754368e-01 9.68705893e-01
3.31547141e-01 -1.67266771e-01 -1.05632496e+00 -3.57149005e-01
1.78360969e-01 1.38068259e+00 -1.21868527e+00 6.55579343e-02
-1.14701673e-01 7.24110484e-01 -9.21226963e-02 -1.24680787e-01
-8.68889630e-01 -2.12542400e-01 3.03864568e-01 5.58081508e-01
1.97061047e-01 -2.94972748e-01 -1.97936803e-01 -4.70786005e-01
-1.25199154e-01 -6.35326505e-01 3.03662151e-01 -6.41267121e-01
-1.10678566e+00 3.13482195e-01 4.23618890e-02 -9.71398890e-01
-1.14827126e-01 -6.56218827e-01 -5.88534713e-01 5.10683954e-01
-1.14895964e+00 -1.20240879e+00 -6.98278069e-01 1.15276933e-01
6.16867900e-01 2.05262214e-01 1.07966304e+00 2.70667344e-01
-5.65661550e-01 -2.87423253e-01 -1.18561625e-01 -2.82271821e-02
5.15376389e-01 -1.01816177e+00 6.64838329e-02 7.68766940e-01
4.17440057e-01 1.19158275e-01 8.91218007e-01 -7.67383516e-01
-1.29048431e+00 -1.09555376e+00 8.21902215e-01 -4.70419377e-01
5.54703057e-01 -5.32185435e-01 -8.22942734e-01 4.17234302e-01
1.21875077e-01 -4.96499240e-01 3.57882716e-02 2.67357007e-03
-1.13064930e-01 -2.50406176e-01 -1.11696410e+00 3.84192735e-01
9.56905425e-01 -9.76551771e-02 -5.69650590e-01 6.51678503e-01
2.75931478e-01 2.20486596e-01 -4.35201466e-01 2.67242491e-01
4.17934030e-01 -1.11610556e+00 9.28017557e-01 -3.44705313e-01
5.36113679e-01 -2.83393651e-01 1.68406323e-01 -1.25773120e+00
-6.56512916e-01 -2.40733236e-01 2.63064682e-01 1.50299251e+00
2.78780103e-01 -3.47137332e-01 6.90639317e-01 6.31022274e-01
9.63459387e-02 -2.22440556e-01 -6.14254475e-01 -1.02448165e+00
-2.93273300e-01 -5.40562153e-01 3.19873661e-01 6.86881781e-01
-1.33546308e-01 1.50360346e-01 1.04710115e-02 2.21501112e-01
5.21146774e-01 -1.00122131e-02 6.40058577e-01 -1.48834980e+00
-1.20008981e-03 -4.19224143e-01 -4.08170015e-01 7.91350529e-02
-3.95127892e-01 -8.06780517e-01 3.49493116e-01 -1.73954391e+00
3.34138602e-01 -4.59713310e-01 -6.22512102e-01 8.50310147e-01
1.34938974e-02 2.39224672e-01 5.93553722e-01 1.82079703e-01
-2.41299048e-01 -5.58386408e-02 8.34804416e-01 -3.97355735e-01
-3.43524367e-01 1.05422266e-01 -3.64730269e-01 8.49996984e-01
9.39071655e-01 -9.06691253e-01 -9.80149880e-02 -1.88672736e-01
1.76218614e-01 -5.10273814e-01 5.48053861e-01 -1.40906000e+00
1.41776502e-01 -3.85147035e-01 4.37883139e-01 -4.97011155e-01
3.83699596e-01 -1.42096961e+00 3.71924669e-01 8.95611286e-01
-1.54142678e-01 -2.03473672e-01 2.02264518e-01 4.99720097e-01
1.94503725e-01 -5.61348438e-01 6.49362504e-01 -5.02287745e-01
-7.20854342e-01 -2.39870653e-01 -8.46079707e-01 -4.11930948e-01
1.49012756e+00 -1.92591295e-01 -1.94747940e-01 -1.73720531e-02
-3.54253143e-01 8.53549540e-02 4.30343062e-01 5.22798240e-01
1.77397877e-01 -7.54970908e-01 -6.96447968e-01 -8.29062015e-02
-1.42155110e-03 -1.11476786e-01 -1.24961756e-01 6.54218078e-01
-7.47857451e-01 4.18373972e-01 -4.15559709e-01 -5.38531721e-01
-1.41534805e+00 9.68313813e-01 -1.07051264e-02 -5.10096073e-01
-9.45238650e-01 6.52090788e-01 -2.97100842e-01 -3.67882960e-02
1.27630234e-01 -5.50783694e-01 -1.46150753e-01 4.02339756e-01
5.12350738e-01 6.49906993e-01 2.27784529e-01 -1.16097584e-01
-6.06742024e-01 1.07079715e-01 2.09629968e-01 8.46721232e-03
1.86489224e+00 3.55455488e-01 1.41418397e-01 5.39205015e-01
5.06358266e-01 -1.25745982e-01 -1.17862535e+00 4.66471076e-01
2.49794275e-01 -5.37775636e-01 2.97716707e-01 -1.04267895e+00
-9.19808745e-01 6.54854953e-01 8.09509039e-01 9.15665627e-01
1.19164252e+00 -1.38158007e-02 6.27554119e-01 5.66455364e-01
4.32921350e-01 -1.36464500e+00 1.40823036e-01 3.07604402e-01
8.28967750e-01 -1.23365784e+00 2.37305328e-01 -3.47981781e-01
-5.11185408e-01 1.09808242e+00 2.67364889e-01 -7.28777889e-03
1.61313921e-01 7.63972759e-01 1.43555179e-01 -4.14157540e-01
-6.10971570e-01 -6.12423837e-01 -4.16248083e-01 8.09638321e-01
1.55221254e-01 2.66327709e-01 -2.90954504e-02 -1.78612381e-01
2.20149964e-01 2.96300381e-01 3.35080981e-01 1.23024166e+00
-9.94619191e-01 -1.03659892e+00 -8.90059292e-01 3.34993094e-01
-3.38576704e-01 1.28798276e-01 -7.16932237e-01 1.14222479e+00
4.38825905e-01 1.35508311e+00 -1.35143176e-01 -3.54747653e-01
7.32858121e-01 9.78413224e-02 3.14569503e-01 -5.33002615e-01
-8.69373500e-01 -1.62281305e-01 2.04874575e-01 -3.93240422e-01
-3.29013228e-01 -7.81408966e-01 -1.28989041e+00 -2.10221812e-01
-4.72864777e-01 -2.26303726e-01 1.08430743e+00 7.60730684e-01
-6.36545196e-02 8.65107954e-01 4.89532888e-01 -1.02953649e+00
-6.57750845e-01 -9.68165278e-01 -7.64031827e-01 4.86427993e-01
-3.45000386e-01 -6.42393410e-01 -4.55409229e-01 1.32848516e-01] | [9.07221508026123, -1.1584419012069702] |
d67bfa89-00fe-4ae3-8baf-f0417bda9701 | tgglines-a-robust-topological-graph-guided | 2002.12428 | null | https://arxiv.org/abs/2002.12428v1 | https://arxiv.org/pdf/2002.12428v1.pdf | TGGLines: A Robust Topological Graph Guided Line Segment Detector for Low Quality Binary Images | Line segment detection is an essential task in computer vision and image analysis, as it is the critical foundation for advanced tasks such as shape modeling and road lane line detection for autonomous driving. We present a robust topological graph guided approach for line segment detection in low quality binary images (hence, we call it TGGLines). Due to the graph-guided approach, TGGLines not only detects line segments, but also organizes the segments with a line segment connectivity graph, which means the topological relationships (e.g., intersection, an isolated line segment) of the detected line segments are captured and stored; whereas other line detectors only retain a collection of loose line segments. Our empirical results show that the TGGLines detector visually and quantitatively outperforms state-of-the-art line segment detection methods. In addition, our TGGLines approach has the following two competitive advantages: (1) our method only requires one parameter and it is adaptive, whereas almost all other line segment detection methods require multiple (non-adaptive) parameters, and (2) the line segments detected by TGGLines are organized by a line segment connectivity graph. | ['Diane Oyen', 'Liping Yang', 'Catherine Potts', 'Vijayan K. Asari', 'Ming Gong', 'Brendt Wohlberg'] | 2020-02-27 | null | null | null | null | ['line-segment-detection', 'line-detection'] | ['computer-vision', 'computer-vision'] | [ 4.91358750e-02 -1.37857750e-01 -4.68396217e-01 -5.05980924e-02
-2.97854692e-01 -6.63090467e-01 5.55947781e-01 7.34848619e-01
-1.45307556e-01 4.13439780e-01 -5.20039618e-01 -6.62545443e-01
1.98795367e-02 -9.98074889e-01 -7.45960653e-01 -2.43098602e-01
-1.54816642e-01 3.73838335e-01 1.12697995e+00 -3.30996901e-01
6.65447295e-01 9.80016053e-01 -1.31766117e+00 -3.43647242e-01
1.14140868e+00 9.92385864e-01 -1.79325819e-01 5.55003583e-01
-5.19270182e-01 2.19889164e-01 -1.29177704e-01 -3.07941407e-01
2.37579212e-01 -3.64094347e-01 -3.95770609e-01 3.81586999e-01
7.15587974e-01 -1.78594843e-01 -4.20698225e-01 1.15092432e+00
3.88395116e-02 1.91773637e-03 6.16685569e-01 -1.49580133e+00
1.14409856e-01 2.93257147e-01 -1.02371192e+00 5.77041507e-02
2.24707529e-01 -1.13740459e-01 8.76337051e-01 -8.57208490e-01
6.49087369e-01 9.50709045e-01 8.61176074e-01 -1.85511217e-01
-1.27382553e+00 -4.60905105e-01 3.43900211e-02 3.11277080e-02
-1.50584924e+00 -3.03583026e-01 9.74779069e-01 -5.77324271e-01
5.17396867e-01 3.88733774e-01 8.75433028e-01 -2.05240510e-02
2.85503298e-01 7.72882640e-01 7.97888279e-01 -5.60306072e-01
8.95483941e-02 -1.77096743e-02 7.14975774e-01 1.28146589e+00
4.83500272e-01 2.48940848e-02 -1.33290529e-01 3.02240103e-02
1.23183143e+00 -2.68317491e-01 -1.41500324e-01 -9.54195917e-01
-1.28213882e+00 7.54219294e-01 5.81375241e-01 2.05537558e-01
-4.08470482e-02 1.37312457e-01 3.78075361e-01 7.92247355e-02
5.17591313e-02 1.52769044e-01 3.72512788e-01 -1.11513929e-02
-1.25963545e+00 3.14785056e-02 6.50551200e-01 1.26460469e+00
1.23595011e+00 -1.28227472e-01 -2.86496263e-02 6.63467646e-01
1.79100767e-01 5.47297478e-01 4.55224104e-02 -6.76435709e-01
6.46470547e-01 1.00465405e+00 4.32858877e-02 -1.51012278e+00
-8.43199432e-01 -4.07632679e-01 -7.79392660e-01 3.51186484e-01
4.69400585e-01 1.81374233e-03 -7.00216472e-01 9.52053189e-01
2.19122142e-01 -1.00685265e-02 -3.43120784e-01 4.59262818e-01
8.14853966e-01 6.81823373e-01 -4.04431760e-01 -1.90047726e-01
1.26960313e+00 -1.22982419e+00 -4.91866797e-01 -4.29106086e-01
8.16173255e-01 -7.70998955e-01 8.11426103e-01 9.92703363e-02
-1.04138839e+00 -6.25870764e-01 -1.18516672e+00 -5.39363846e-02
-2.06755087e-01 4.42019045e-01 4.93134975e-01 4.26905543e-01
-8.84616196e-01 6.86962530e-02 -7.03013122e-01 -4.02384281e-01
2.24441424e-01 2.10241780e-01 -4.22567338e-01 1.63241282e-01
-4.68866169e-01 5.61124265e-01 5.25225401e-01 1.30511329e-01
1.41451508e-01 -1.02448441e-01 -1.11946225e+00 2.94963084e-02
5.74880421e-01 -2.64440864e-01 8.60456407e-01 -4.25974756e-01
-1.11509240e+00 8.98180187e-01 -4.07606304e-01 -6.38355792e-01
9.46233869e-01 -1.07394941e-02 -4.13913101e-01 4.19470817e-01
9.58465189e-02 6.56964242e-01 5.52082777e-01 -1.48391747e+00
-9.35225487e-01 -5.75397797e-02 -3.07474166e-01 1.12362370e-01
2.79790193e-01 -2.79635429e-01 -9.07416642e-01 -2.92749524e-01
7.67328501e-01 -1.01544285e+00 -2.22590327e-01 3.15831065e-01
-9.84848738e-01 -3.41951936e-01 1.15125465e+00 -2.03644261e-01
1.62612534e+00 -2.26572680e+00 -5.29962480e-01 8.38586986e-01
4.44194615e-01 4.12317991e-01 1.46824628e-01 6.75323844e-01
7.86023736e-02 1.14664502e-01 -4.73561436e-01 -1.33892953e-01
-2.85157233e-01 -1.77200451e-01 -2.77516603e-01 6.04112267e-01
-6.01052046e-02 9.04612660e-01 -8.01930964e-01 -8.40567291e-01
6.68357790e-01 -1.14045233e-01 -1.41648665e-01 -2.65843272e-01
2.41491497e-02 5.70333302e-02 -4.65856194e-01 3.86640996e-01
8.24135780e-01 -4.02356274e-02 -2.89695919e-01 -4.22265083e-01
-8.82934153e-01 -1.75069109e-01 -1.31549549e+00 1.08796299e+00
2.57557064e-01 1.08693194e+00 -4.53800619e-01 -7.24827230e-01
1.57619286e+00 -2.78356522e-01 5.46587765e-01 -6.51492953e-01
-4.16328833e-02 3.17107767e-01 -8.84287134e-02 -2.11420044e-01
7.30741858e-01 6.12694204e-01 -2.59377033e-01 1.22614272e-01
-5.73768973e-01 -5.02176106e-01 5.33225417e-01 2.64699548e-01
6.42475188e-01 -1.45332128e-01 5.30857980e-01 -2.10850894e-01
6.10344529e-01 4.53094095e-01 4.92117316e-01 8.79613817e-01
-3.90353262e-01 4.85209197e-01 8.00699234e-01 -3.40418726e-01
-1.03861725e+00 -9.99271572e-01 -3.10529381e-01 3.07839692e-01
9.12548423e-01 -4.73979712e-01 -8.56195629e-01 -3.12223464e-01
7.92685598e-02 4.85803157e-01 -2.68826336e-01 5.96373118e-02
-9.05673862e-01 -2.16747168e-02 4.74452913e-01 4.93359566e-01
1.06347406e+00 -7.23768532e-01 -5.79031646e-01 2.69114107e-01
3.26815546e-02 -1.08382308e+00 -7.89115846e-01 -1.42712906e-01
-9.16607738e-01 -1.26083338e+00 -4.46834534e-01 -1.16473722e+00
8.68330181e-01 9.51744735e-01 8.31287503e-01 3.54082525e-01
2.39936449e-02 4.38719615e-02 -9.13246498e-02 -2.66931057e-01
-1.86882317e-01 -2.64646225e-02 -4.42152530e-01 1.26122937e-01
1.08625211e-01 2.69512460e-02 -6.05430186e-01 7.81006455e-01
-4.01697338e-01 4.30893213e-01 3.54943097e-01 5.22108436e-01
1.01186407e+00 2.01334491e-01 1.87817514e-01 -9.11077499e-01
4.71892983e-01 8.47985968e-02 -1.22093558e+00 2.19600439e-01
-4.31068033e-01 -2.18412936e-01 3.19108814e-01 8.76310468e-02
-4.69386756e-01 2.59632289e-01 -5.82466200e-02 -7.60678500e-02
-1.00516856e-01 5.02015412e-01 3.07808165e-02 -5.25982082e-01
5.18544078e-01 2.87904054e-01 -1.45672426e-01 -7.72230029e-02
4.55078691e-01 3.83613050e-01 9.08437371e-01 1.42397527e-02
9.93978322e-01 6.25864029e-01 4.35029119e-01 -1.27891707e+00
-1.95399106e-01 -7.90406227e-01 -1.15130937e+00 -6.44813120e-01
6.40409112e-01 -4.41152364e-01 -6.18143439e-01 6.65234923e-01
-1.05495131e+00 -1.88619524e-01 -1.64027303e-01 3.30166847e-01
-6.92405105e-01 8.46374750e-01 -4.38894689e-01 -8.36403608e-01
7.24661052e-02 -1.03259265e+00 7.79319704e-01 5.46318114e-01
-6.79638702e-03 -9.71025825e-01 1.15429319e-03 7.94330910e-02
-1.63623810e-01 7.22448707e-01 1.04793453e+00 -3.87368977e-01
-7.89830029e-01 -6.00264907e-01 -6.15632713e-01 -1.51834607e-01
-1.15625225e-02 4.87985045e-01 -2.83193827e-01 -5.28651178e-02
-6.93211675e-01 2.69788772e-01 7.93155730e-01 5.68452895e-01
7.94265091e-01 -3.11338976e-02 -8.63218486e-01 6.36955857e-01
1.27600253e+00 4.01459038e-01 7.56358922e-01 4.89561260e-01
9.24502552e-01 6.24026477e-01 8.03501666e-01 -4.37929444e-02
4.20327365e-01 7.54466951e-01 2.14025110e-01 -7.07130551e-01
-2.05852211e-01 -5.82054377e-01 3.00177541e-02 6.61084116e-01
1.63517937e-01 -2.68641442e-01 -9.78529096e-01 3.99708092e-01
-2.10187221e+00 -8.88188899e-01 -1.19195139e+00 2.47531128e+00
2.99499393e-01 6.92079902e-01 5.86576462e-01 4.46442157e-01
1.10749090e+00 1.32341653e-01 -6.43327475e-01 -4.59842443e-01
-2.96621561e-01 -4.09994841e-01 9.18561876e-01 4.89778996e-01
-1.26672351e+00 1.19128335e+00 6.28838968e+00 8.17838013e-01
-1.05762458e+00 -7.02280819e-01 4.36749071e-01 9.30584192e-01
-8.19992647e-02 1.29159793e-01 -8.61707866e-01 2.49895886e-01
9.92506593e-02 -3.41689140e-01 -8.59786719e-02 7.76774347e-01
3.80591720e-01 -5.53379178e-01 -7.65212595e-01 1.06775618e+00
-1.18085578e-01 -1.27766228e+00 -4.40097563e-02 1.30159736e-01
6.74473763e-01 -1.29738927e-01 -1.60348803e-01 -2.65768260e-01
-1.35704607e-01 -6.70701683e-01 8.55663061e-01 4.97605950e-01
7.47950792e-01 -8.25145662e-01 5.04579425e-01 3.81480455e-01
-1.67945027e+00 2.25301102e-01 -2.51247972e-01 3.91091615e-01
4.38255370e-01 6.39587820e-01 -8.14448357e-01 5.99481225e-01
9.09555890e-03 6.96240842e-01 -8.24428618e-01 1.99125218e+00
-2.04833522e-01 6.71796024e-01 -5.58115840e-01 8.20513293e-02
4.48115438e-01 -7.81591535e-01 7.60437787e-01 1.37462890e+00
1.97530791e-01 -3.23606730e-01 4.56557393e-01 6.79934800e-01
2.43758231e-01 3.08643639e-01 -7.52988875e-01 2.46527746e-01
7.27593601e-01 1.04578316e+00 -1.37672389e+00 -2.34568164e-01
-4.79938447e-01 5.38891554e-01 9.74572226e-02 2.99559563e-01
-6.55305922e-01 -9.42521214e-01 7.11566657e-02 4.01827544e-01
3.55660677e-01 -8.41600060e-01 -5.90850294e-01 -7.19661891e-01
-5.84132858e-02 -2.08852187e-01 1.34862021e-01 -6.36368394e-01
-5.89972973e-01 6.13326907e-01 -6.24480052e-03 -1.62313151e+00
-1.84608787e-01 -2.96729058e-01 -8.92590046e-01 4.35500383e-01
-1.57265341e+00 -9.82845962e-01 -6.62779987e-01 6.01858735e-01
3.62995803e-01 3.16618323e-01 3.03843021e-01 -1.21558504e-02
-7.32828021e-01 5.56970298e-01 2.37238333e-01 5.42851925e-01
3.41370583e-01 -1.11453915e+00 7.22728908e-01 1.09389973e+00
1.60757706e-01 3.85685146e-01 4.60559368e-01 -8.84873629e-01
-9.71017838e-01 -1.03755331e+00 7.80668378e-01 1.21982709e-01
5.65619409e-01 -3.17632616e-01 -1.04880774e+00 5.95950305e-01
-1.08255662e-01 -4.59108353e-02 2.80467153e-01 -3.81105930e-01
-7.72445202e-02 -9.88421217e-02 -8.23287368e-01 7.68032253e-01
1.01349485e+00 -1.01802111e-01 -4.08413231e-01 1.86327189e-01
2.93534040e-01 -4.73531902e-01 -4.03418303e-01 3.61225218e-01
5.00069618e-01 -1.33370364e+00 8.81642103e-01 1.32168293e-01
-1.99097365e-01 -8.91188860e-01 4.55929935e-01 -1.00486791e+00
-4.08873707e-01 -6.92894697e-01 2.00289011e-01 1.22456992e+00
4.64697510e-01 -8.27700913e-01 7.32546449e-01 1.66964039e-01
-3.93331081e-01 -6.59243107e-01 -7.75615275e-01 -9.07534420e-01
-4.99000758e-01 -2.86487520e-01 4.93775129e-01 6.54818177e-01
-1.02241717e-01 1.85400277e-01 -2.28847414e-02 1.47284865e-01
7.05701232e-01 4.12312835e-01 1.06334245e+00 -1.44859385e+00
3.97266567e-01 -9.37397242e-01 -7.83372283e-01 -1.61525989e+00
-2.87726283e-01 -6.48051739e-01 1.35016605e-01 -1.88420701e+00
-3.26167285e-01 -8.58902514e-01 1.97889060e-01 2.47057691e-01
1.03975475e-01 2.13097349e-01 2.05281094e-01 5.56836724e-01
-5.17327607e-01 2.29217559e-01 1.21707952e+00 4.32619452e-02
-5.75703084e-01 2.58396626e-01 -1.83748335e-01 8.86316836e-01
6.96026325e-01 -3.03337462e-02 -2.09129035e-01 7.47192204e-02
3.93926017e-02 8.87422264e-02 3.22477162e-01 -1.18618560e+00
6.00324571e-01 -1.86333835e-01 7.42242262e-02 -1.30762064e+00
1.20303564e-01 -5.81366241e-01 -1.47238404e-01 5.92019975e-01
2.31306981e-02 2.40660131e-01 2.36694872e-01 4.79098707e-01
-2.27011636e-01 -3.63005698e-01 8.45654428e-01 3.54294896e-01
-1.03946710e+00 2.19492674e-01 -5.12056947e-01 -2.50976920e-01
1.42388785e+00 -1.01806438e+00 -6.08536899e-01 -3.49190295e-01
-2.43558943e-01 4.77111548e-01 7.89589524e-01 1.38845965e-01
7.54071057e-01 -1.34812808e+00 -5.19303024e-01 4.31586951e-01
4.19448376e-01 2.64973342e-01 -1.54090673e-01 8.48983109e-01
-1.26070416e+00 6.49805486e-01 -1.52771696e-01 -9.53575730e-01
-1.28753030e+00 4.42509383e-01 3.09871197e-01 1.07243411e-01
-1.06505549e+00 5.59398472e-01 2.92134017e-01 4.75790314e-02
-2.33482942e-03 -5.69145143e-01 -3.44095916e-01 5.81988432e-02
1.54333293e-01 7.00557947e-01 1.92790735e-03 -1.02700436e+00
-1.78527802e-01 1.33467078e+00 4.98852879e-02 3.62197943e-02
7.65209556e-01 -2.66427666e-01 6.93559833e-03 5.47171950e-01
8.56111228e-01 4.04671252e-01 -1.22328532e+00 -1.77510694e-01
2.44514674e-01 -5.09827971e-01 2.33999919e-02 -2.23534256e-01
-7.99205542e-01 6.73981428e-01 3.13642561e-01 6.96942747e-01
7.74984539e-01 -7.35330358e-02 9.37271893e-01 3.19191337e-01
5.42451501e-01 -9.60676551e-01 -2.26087511e-01 4.88458991e-01
7.05746710e-01 -8.38912785e-01 -2.24056244e-02 -1.16240513e+00
-4.20645952e-01 1.53392959e+00 4.80561584e-01 -2.26797849e-01
3.68585676e-01 1.12082930e-02 -1.34346792e-02 -2.88546354e-01
1.22576214e-01 -4.34789896e-01 5.12738109e-01 3.77883285e-01
1.07992984e-01 1.33868456e-01 -4.26246822e-01 -3.16326439e-01
-4.88116115e-01 -2.90654004e-01 5.76735258e-01 7.92139947e-01
-9.89485741e-01 -9.11487222e-01 -5.56389332e-01 5.36197603e-01
4.82994080e-01 2.17510566e-01 -4.75673646e-01 1.12130487e+00
-1.53420135e-01 9.82297421e-01 3.46685886e-01 -2.27990106e-01
6.88449085e-01 -5.31038344e-01 1.32146895e-01 -3.79896671e-01
1.52528090e-02 1.51454732e-01 6.74746707e-02 -3.77144635e-01
-6.51392639e-02 -7.35753357e-01 -1.68279397e+00 -2.35646188e-01
-7.02942371e-01 2.37303495e-01 7.10321128e-01 6.99537873e-01
2.92910576e-01 8.68414342e-02 7.78632402e-01 -5.17165661e-01
3.61262321e-01 -3.54682982e-01 -6.03889108e-01 1.25437826e-01
4.67603505e-01 -6.99291945e-01 -9.88048390e-02 -1.02503844e-01] | [8.272261619567871, -1.6015883684158325] |
8d5e8c6d-2361-4ace-9bc1-587c58e124d8 | inductive-graph-unlearning | 2304.03093 | null | https://arxiv.org/abs/2304.03093v2 | https://arxiv.org/pdf/2304.03093v2.pdf | Inductive Graph Unlearning | As a way to implement the "right to be forgotten" in machine learning, \textit{machine unlearning} aims to completely remove the contributions and information of the samples to be deleted from a trained model without affecting the contributions of other samples. Recently, many frameworks for machine unlearning have been proposed, and most of them focus on image and text data. To extend machine unlearning to graph data, \textit{GraphEraser} has been proposed. However, a critical issue is that \textit{GraphEraser} is specifically designed for the transductive graph setting, where the graph is static and attributes and edges of test nodes are visible during training. It is unsuitable for the inductive setting, where the graph could be dynamic and the test graph information is invisible in advance. Such inductive capability is essential for production machine learning systems with evolving graphs like social media and transaction networks. To fill this gap, we propose the \underline{{\bf G}}\underline{{\bf U}}ided \underline{{\bf I}}n\underline{{\bf D}}uctiv\underline{{\bf E}} Graph Unlearning framework (GUIDE). GUIDE consists of three components: guided graph partitioning with fairness and balance, efficient subgraph repair, and similarity-based aggregation. Empirically, we evaluate our method on several inductive benchmarks and evolving transaction graphs. Generally speaking, GUIDE can be efficiently implemented on the inductive graph learning tasks for its low graph partition cost, no matter on computation or structure information. The code will be available here: https://github.com/Happy2Git/GUIDE. | ['Di Wang', 'Mengdi Huai', 'Cheng-Long Wang'] | 2023-04-06 | null | null | null | null | ['graph-partitioning'] | ['graphs'] | [ 4.16670203e-01 7.21046150e-01 -3.64603877e-01 3.11964191e-02
-1.18142188e-01 -7.63103366e-01 3.86272669e-01 3.08636665e-01
5.38284108e-02 8.53284478e-01 -3.45689833e-01 -5.70296466e-01
-4.34789598e-01 -1.06701195e+00 -8.75395060e-01 -7.90275097e-01
-2.15346843e-01 6.16004646e-01 2.51861006e-01 -1.94937974e-01
-1.50832951e-01 2.48173296e-01 -1.32737994e+00 7.35677257e-02
9.33256030e-01 5.71177483e-01 5.91687299e-02 6.15659177e-01
5.54772466e-03 1.03570223e+00 -3.90065163e-01 -6.17774367e-01
2.32621819e-01 -5.99309981e-01 -9.14715052e-01 1.16229437e-01
3.09083581e-01 -2.91962232e-02 -4.86565053e-01 1.14957750e+00
2.85089374e-01 2.79604010e-02 7.17366219e-01 -1.63845694e+00
-6.51703954e-01 1.04014122e+00 -6.79847896e-01 1.33795261e-01
1.67308822e-01 6.13644905e-02 1.32217598e+00 -7.01064169e-01
7.61361837e-01 8.93170536e-01 4.06186402e-01 4.76421922e-01
-1.20911646e+00 -6.82635367e-01 4.93944585e-01 9.28738043e-02
-1.25726724e+00 -1.23821028e-01 9.44230020e-01 -4.61191267e-01
6.94963992e-01 7.02353537e-01 7.09327221e-01 1.10546255e+00
9.91771817e-02 1.04344320e+00 8.70827079e-01 -5.39267480e-01
1.12465762e-01 4.72593680e-02 3.33601445e-01 1.09299660e+00
4.93229538e-01 1.92409307e-01 -2.70686001e-01 -3.33390757e-02
6.87907040e-01 4.24852856e-02 -3.25363725e-01 -6.06930971e-01
-8.71420205e-01 7.85125792e-01 4.50342536e-01 1.97802186e-01
1.39098048e-01 1.90142184e-01 2.99947023e-01 6.15693390e-01
7.03775346e-01 2.95782417e-01 -3.53228331e-01 1.83914915e-01
-5.78050971e-01 -1.32510692e-01 1.00386441e+00 1.17409956e+00
1.18617988e+00 1.66725084e-01 4.49078567e-02 8.61403942e-01
3.20064923e-04 2.99539179e-01 -5.89401787e-03 -7.28998601e-01
4.42243338e-01 1.00752532e+00 -5.06897926e-01 -1.08068419e+00
-3.77958089e-01 -5.45699179e-01 -1.07718289e+00 2.17325054e-02
3.63036811e-01 -2.17766792e-01 -1.00189459e+00 1.63542271e+00
4.38398123e-01 2.88292140e-01 -4.21899140e-01 6.19495332e-01
9.58005607e-01 3.51541042e-01 -2.35957116e-01 -3.45576882e-01
7.70103455e-01 -1.00441146e+00 -3.21411103e-01 -2.66162872e-01
9.72659707e-01 -4.39414591e-01 1.21029174e+00 4.20677066e-01
-7.79926598e-01 -3.28232348e-01 -1.00743079e+00 1.55861497e-01
-3.96028906e-01 -1.89458251e-01 7.55007148e-01 8.19669247e-01
-1.18735051e+00 6.73008919e-01 -5.64160645e-01 -2.90096313e-01
4.30170536e-01 4.61714953e-01 -4.07177180e-01 -4.74047601e-01
-1.02128196e+00 2.50737131e-01 6.81792080e-01 7.80015811e-02
-9.51463878e-01 -4.54928696e-01 -9.15959895e-01 1.75246801e-02
9.50245023e-01 -6.91353500e-01 5.85260570e-01 -1.20473862e+00
-1.04797220e+00 8.89267743e-01 3.38088304e-01 -1.03891358e-01
6.07801378e-01 9.46606919e-02 -3.18057001e-01 -7.11289272e-02
-1.15741566e-01 3.27697128e-01 8.99372220e-01 -1.54819417e+00
-3.01179200e-01 -6.19318366e-01 2.78629869e-01 2.40655337e-02
-6.34492040e-01 -3.29636335e-01 -6.78282857e-01 -8.94342363e-01
1.26377940e-01 -1.32494593e+00 -5.68164513e-03 -4.67712939e-01
-7.12934136e-01 -2.39099950e-01 9.29794788e-01 -4.04457480e-01
1.67074049e+00 -1.96717501e+00 2.09761038e-01 6.09841287e-01
7.89318502e-01 3.10462683e-01 -2.42337927e-01 5.89280069e-01
-1.92074686e-01 3.67032140e-01 -1.81019098e-01 1.77615508e-01
-9.76954550e-02 4.07277495e-01 -2.14379933e-02 4.02779967e-01
-1.49225950e-01 1.00077677e+00 -1.03119636e+00 -5.12968838e-01
4.63756882e-02 4.46495377e-02 -7.10017681e-01 -5.89431189e-02
-4.95972902e-01 3.44829381e-01 -4.58962649e-01 8.35596979e-01
5.97615361e-01 -6.56585753e-01 6.72703087e-01 1.30049571e-01
4.31521624e-01 -1.59169942e-01 -1.31451786e+00 1.31362426e+00
3.67382206e-02 2.46753395e-01 2.47758344e-01 -1.05753124e+00
8.72734845e-01 -4.40594777e-02 6.27094150e-01 -5.44306397e-01
1.25414401e-01 8.37503821e-02 1.34531692e-01 -3.07515740e-01
2.43957892e-01 3.74382883e-01 -6.71889782e-02 8.43091726e-01
1.41071109e-02 -1.79323032e-01 4.60250467e-01 6.08958721e-01
1.54646480e+00 1.09203205e-01 1.66590974e-01 -2.73553908e-01
4.18111742e-01 -1.81903362e-01 4.80256349e-01 8.41926932e-01
1.19217299e-02 3.44206214e-01 8.67947698e-01 -3.78245234e-01
-8.02906513e-01 -1.07952130e+00 2.79926121e-01 1.34296238e+00
2.27582097e-01 -7.44499981e-01 -7.41070509e-01 -1.16785920e+00
-3.78766954e-02 4.88855273e-01 -6.51580751e-01 -4.12000269e-01
-4.46825534e-01 -6.27070725e-01 3.79263043e-01 2.31221482e-01
9.70279798e-02 -1.09515369e+00 2.50808418e-01 -3.70935723e-02
-1.48717165e-02 -8.05787027e-01 -4.66009498e-01 1.68320149e-01
-7.51570225e-01 -1.39691269e+00 -2.34906986e-01 -7.32382238e-01
1.11030650e+00 2.10887015e-01 1.27575600e+00 7.29531705e-01
-1.14336550e-01 5.64997971e-01 -5.82282007e-01 -4.28333998e-01
-4.42552030e-01 3.93052161e-01 -1.32934585e-01 2.87166070e-02
9.90490243e-02 -8.05720329e-01 -4.32894409e-01 5.35859108e-01
-1.08510542e+00 2.19830751e-01 4.48192060e-01 9.19224977e-01
5.66493809e-01 2.40843445e-01 4.63500082e-01 -1.68170583e+00
5.53663731e-01 -5.52361369e-01 -3.71710956e-01 3.95880103e-01
-9.06399488e-01 9.77276359e-03 7.38042235e-01 -3.04578513e-01
-5.90137124e-01 -2.36374080e-01 5.80889992e-02 -4.75823492e-01
2.50033408e-01 6.44476831e-01 -3.15905809e-01 -1.92371942e-02
6.97158754e-01 1.07562900e-01 1.95209701e-02 -2.53873497e-01
2.83992529e-01 2.91153908e-01 1.39733851e-01 -7.34071434e-01
7.07781196e-01 3.72307330e-01 7.54093230e-02 -6.86803520e-01
-6.84971809e-01 -1.63584307e-01 -5.29187143e-01 -5.41019678e-01
3.48898500e-01 -5.10872543e-01 -5.61880827e-01 3.98769408e-01
-6.17707074e-01 -7.54997849e-01 -3.40618640e-01 2.03941911e-01
-3.30452383e-01 5.68132460e-01 -5.73146343e-01 -5.90396762e-01
-2.88974524e-01 -1.01242781e+00 5.76806426e-01 -1.32243638e-03
-1.94349810e-01 -1.20669651e+00 -1.31425977e-01 4.88810778e-01
-8.33077133e-02 4.13250476e-01 1.34659433e+00 -8.27759862e-01
-7.31169343e-01 -2.02263698e-01 -1.56993568e-01 5.16940534e-01
2.98288614e-01 1.29504338e-01 -6.79159999e-01 -6.99446619e-01
-5.17683744e-01 -4.34133798e-01 9.20944452e-01 4.31746766e-02
1.22752953e+00 -5.61392069e-01 -5.35894096e-01 4.72451627e-01
1.25809026e+00 -6.76165670e-02 5.47623456e-01 -8.32585245e-02
1.15584052e+00 5.51507950e-01 3.19885671e-01 3.82003665e-01
4.91709411e-01 2.46247873e-01 7.01476038e-01 -1.81471720e-01
-1.46725431e-01 -4.86972064e-01 4.09466565e-01 9.38319862e-01
-3.05078477e-01 -7.28160739e-01 -1.06592417e+00 3.25486124e-01
-1.96666884e+00 -7.66648412e-01 -3.14653248e-01 2.12573981e+00
6.44942045e-01 2.66348273e-01 -4.06169072e-02 2.17722699e-01
8.54303360e-01 1.40001178e-01 -7.08910763e-01 -2.59438336e-01
-3.66255119e-02 2.78480232e-01 2.76472479e-01 4.65715826e-01
-8.57640147e-01 1.02986181e+00 4.92247820e+00 1.05147994e+00
-8.96056950e-01 2.73234099e-02 7.63753057e-01 -1.90712363e-02
-7.38463938e-01 3.99370015e-01 -4.28279638e-01 4.02991027e-01
5.19673169e-01 -2.08403662e-01 5.89476585e-01 6.77277207e-01
-5.91071583e-02 -9.24723521e-02 -1.15934181e+00 6.66510463e-01
1.10186459e-02 -1.23338926e+00 2.76867688e-01 2.97906518e-01
8.60147715e-01 -4.53946479e-02 -1.04796700e-01 5.60306489e-01
8.25270414e-01 -9.64229286e-01 3.83371592e-01 3.93680573e-01
9.92388010e-01 -6.83189273e-01 3.38250816e-01 6.17848396e-01
-1.29751456e+00 -1.03041627e-01 -2.54501346e-02 -6.47276714e-02
-3.99650812e-01 9.37284708e-01 -8.98954213e-01 9.45009589e-01
6.86899543e-01 6.13743603e-01 -7.22946107e-01 4.90494788e-01
-4.69613105e-01 8.77446175e-01 -1.78717867e-01 -4.38700728e-02
-8.86805505e-02 -5.36918461e-01 7.75757372e-01 9.28817272e-01
1.42483398e-01 4.13013026e-02 5.83525181e-01 6.00409448e-01
-3.72335434e-01 3.10430795e-01 -8.76793563e-01 -2.53820717e-01
3.28748852e-01 1.17676938e+00 -1.06669223e+00 -1.67969644e-01
-2.51101345e-01 7.12667406e-01 6.69168055e-01 4.99584198e-01
-8.60038817e-01 -3.48300755e-01 2.06271559e-01 5.60024023e-01
1.21983521e-01 -7.22552463e-02 -1.31204173e-01 -1.12778461e+00
8.85053203e-02 -1.00778651e+00 1.00315988e+00 -5.34469128e-01
-1.19184804e+00 5.08932412e-01 -2.84512378e-02 -9.36754405e-01
-1.41685363e-03 -4.26622540e-01 -4.12055075e-01 2.95433462e-01
-1.01625717e+00 -1.31266451e+00 -4.72020537e-01 8.52626383e-01
1.40041932e-01 -2.02698126e-01 5.38966417e-01 2.84755498e-01
-7.09530532e-01 7.16470242e-01 -1.69261083e-01 2.68455297e-01
5.11537731e-01 -1.28937590e+00 5.29895760e-02 1.00111663e+00
3.04774076e-01 4.80992198e-01 4.59968805e-01 -8.96742046e-01
-1.41404486e+00 -1.28460479e+00 6.09289229e-01 -3.45371485e-01
7.20994413e-01 -6.78772688e-01 -9.05785978e-01 1.06407273e+00
-2.44787950e-02 2.32178152e-01 5.51762104e-01 2.77256012e-01
-3.60480666e-01 -3.17231178e-01 -9.35656607e-01 7.93963253e-01
1.63409126e+00 -2.54746765e-01 3.14970454e-03 5.62565446e-01
7.91468740e-01 -3.30676883e-01 -8.98273408e-01 6.91925824e-01
1.89637616e-01 -1.05893672e+00 6.14484727e-01 -4.85350937e-01
2.32054919e-01 -2.20127717e-01 2.89871246e-01 -1.23135293e+00
-3.83737952e-01 -1.08208370e+00 -4.74368036e-01 1.28235781e+00
6.08759344e-01 -9.29826438e-01 9.77702796e-01 3.99122357e-01
-1.87804610e-01 -7.19866335e-01 -7.60114610e-01 -6.91501558e-01
2.75622066e-02 -3.88704151e-01 5.58311105e-01 1.17910111e+00
-3.09015047e-02 5.26997209e-01 -4.16202605e-01 -5.44797517e-02
5.81303239e-01 2.74605691e-01 1.02498138e+00 -1.27914798e+00
-4.98439550e-01 -3.99988711e-01 -1.61440432e-01 -8.01726699e-01
2.01383591e-01 -1.51989532e+00 -5.01025021e-01 -1.52987206e+00
4.70387787e-01 -8.45391333e-01 -3.47095788e-01 8.49038422e-01
-2.02977717e-01 9.95590240e-02 1.39472149e-02 1.70135021e-01
-6.05646372e-01 3.75114053e-01 1.55963945e+00 -3.77723932e-01
-2.60469645e-01 1.24748178e-01 -8.09367061e-01 5.07676959e-01
8.16353858e-01 -5.14931619e-01 -8.60818386e-01 -2.61269867e-01
6.24050975e-01 -2.54209097e-02 3.53446931e-01 -6.56748652e-01
2.53502756e-01 -2.26971671e-01 5.29716238e-02 -3.17530513e-01
-8.77983272e-02 -5.96918225e-01 4.54457849e-01 3.79272461e-01
-4.61879745e-02 1.19684920e-01 -9.36550424e-02 7.17829704e-01
5.15931249e-02 -1.82627290e-02 6.04436755e-01 -3.65690380e-01
-5.62730491e-01 7.64281154e-01 8.64662677e-02 3.86832803e-01
1.22703278e+00 -3.29635948e-01 -6.03913128e-01 -3.67735058e-01
-8.43801558e-01 4.33717370e-01 4.98519540e-01 3.93355519e-01
5.29887795e-01 -1.05129790e+00 -7.71974742e-01 2.13661432e-01
2.69304127e-01 2.23734394e-01 3.71910989e-01 9.74310875e-01
-3.71146739e-01 -3.92633639e-02 1.04393490e-01 -5.02790689e-01
-1.24199450e+00 8.50216985e-01 1.51120052e-01 -6.05147004e-01
-4.64832664e-01 8.70391250e-01 4.06857193e-01 -6.55544639e-01
1.06018275e-01 -9.55869034e-02 9.95813385e-02 1.05216317e-01
-1.17057450e-01 4.69007343e-01 7.46597275e-02 -2.96170235e-01
-1.18442975e-01 3.12436838e-02 -3.15015107e-01 4.51491475e-01
1.29926026e+00 -7.55403787e-02 -5.36685228e-01 2.84010470e-01
8.54511023e-01 1.18053362e-01 -1.06718433e+00 -2.80901194e-01
6.48034662e-02 -2.60007679e-01 -2.54370302e-01 -7.49356329e-01
-1.27129316e+00 4.01364893e-01 -4.33373749e-02 5.20766616e-01
1.15278113e+00 2.07283616e-01 5.81786036e-01 2.02623948e-01
5.33758759e-01 -1.30697846e+00 4.34783489e-01 3.66058052e-01
8.86816561e-01 -1.08647358e+00 2.85001993e-01 -7.74273455e-01
-4.28320199e-01 9.26003754e-01 7.80487061e-01 5.38920378e-03
8.20990026e-01 1.79835916e-01 -3.60447258e-01 -4.93801057e-01
-6.78556740e-01 -2.97104150e-01 1.71398938e-01 5.70053935e-01
1.28312558e-01 2.38443419e-01 -4.02078241e-01 4.94700551e-01
-1.80840433e-01 -1.63121223e-01 5.17354012e-01 9.48491454e-01
-1.34447873e-01 -1.38497424e+00 2.36353371e-02 6.83827221e-01
-1.62166491e-01 -1.16180070e-01 -8.65152061e-01 9.82141078e-01
2.91184694e-01 9.09611344e-01 -3.81282181e-01 -6.00054204e-01
1.68651327e-01 -1.56431958e-01 5.51255345e-01 -8.32184970e-01
-4.98772115e-01 -2.14498695e-02 2.58568257e-01 -3.27266932e-01
-2.47844890e-01 -4.69955802e-01 -1.28361583e+00 -4.97273475e-01
-3.81856173e-01 1.10642657e-01 2.57106894e-03 7.07013369e-01
3.33203524e-01 5.59504628e-01 7.05572724e-01 -3.75191450e-01
-1.99970126e-01 -6.37672007e-01 -8.16059887e-01 3.81214231e-01
-6.48590401e-02 -5.39892077e-01 -4.74123985e-01 -6.19101636e-02] | [7.266981601715088, 6.184572219848633] |
8ef4b964-4694-4318-bd72-801b424b3efa | person-search-by-multi-scale-matching | 1807.08582 | null | http://arxiv.org/abs/1807.08582v1 | http://arxiv.org/pdf/1807.08582v1.pdf | Person Search by Multi-Scale Matching | We consider the problem of person search in unconstrained scene images.
Existing methods usually focus on improving the person detection accuracy to
mitigate negative effects imposed by misalignment, mis-detections, and false
alarms resulted from noisy people auto-detection. In contrast to previous
studies, we show that sufficiently reliable person instance cropping is
achievable by slightly improved state-of-the-art deep learning object detectors
(e.g. Faster-RCNN), and the under-studied multi-scale matching problem in
person search is a more severe barrier. In this work, we address this
multi-scale person search challenge by proposing a Cross-Level Semantic
Alignment (CLSA) deep learning approach capable of learning more discriminative
identity feature representations in a unified end-to-end model. This is
realised by exploiting the in-network feature pyramid structure of a deep
neural network enhanced by a novel cross pyramid-level semantic alignment loss
function. This favourably eliminates the need for constructing a
computationally expensive image pyramid and a complex multi-branch network
architecture. Extensive experiments show the modelling advantages and
performance superiority of CLSA over the state-of-the-art person search and
multi-scale matching methods on two large person search benchmarking datasets:
CUHK-SYSU and PRW. | ['Shaogang Gong', 'Xiatian Zhu', 'Xu Lan'] | 2018-07-23 | null | null | null | eccv-2018 | ['person-search'] | ['computer-vision'] | [ 2.13098690e-01 -3.41136009e-01 3.51216137e-01 -2.69237906e-01
-8.79197598e-01 -3.31712753e-01 7.41561651e-01 5.61189801e-02
-1.14151263e+00 4.74678934e-01 2.39749521e-01 5.02181649e-01
-3.29900235e-01 -6.90435827e-01 -7.22631812e-01 -3.57997924e-01
1.81002125e-01 8.74074876e-01 3.05141300e-01 -2.66036242e-01
-7.41537735e-02 4.29442763e-01 -1.99754322e+00 3.41841161e-01
6.73279881e-01 9.22460556e-01 6.60258755e-02 6.07209563e-01
4.21531528e-01 2.47536883e-01 -8.11298668e-01 -9.35959220e-01
5.98078191e-01 -1.17248341e-01 -7.84226716e-01 -3.64778899e-02
1.25185406e+00 -2.80729413e-01 -3.89501184e-01 1.11525071e+00
1.18916261e+00 2.10998788e-01 3.61145824e-01 -1.21895111e+00
-6.38424993e-01 1.33052677e-01 -5.30423999e-01 3.95935386e-01
6.29813373e-01 3.86576533e-01 9.47495043e-01 -8.69996190e-01
5.25571883e-01 1.75839543e+00 1.17781770e+00 5.25311708e-01
-1.34687436e+00 -5.87651491e-01 3.03341132e-02 4.65662062e-01
-1.73518765e+00 -2.73623079e-01 3.29043716e-01 -3.16490084e-01
1.04952645e+00 3.63733858e-01 5.74158430e-01 1.41043520e+00
-3.35264504e-01 1.03454041e+00 9.97453868e-01 -5.59664249e-01
-3.09685051e-01 1.94415510e-01 1.60666630e-01 8.43329430e-01
5.21198928e-01 4.63270307e-01 -6.36154652e-01 -3.15686613e-01
7.19797313e-01 9.21200553e-04 -9.55737159e-02 -4.74168360e-01
-1.06946945e+00 7.76464045e-01 7.81051278e-01 4.71713573e-01
-5.19100428e-01 1.71136394e-01 6.06631875e-01 1.86598644e-01
1.86794192e-01 4.39366311e-01 -3.06427538e-01 2.46864915e-01
-1.21908796e+00 8.90687227e-01 5.41117966e-01 6.78972960e-01
3.37854475e-01 -2.04326674e-01 -7.72822201e-01 1.01185477e+00
3.36109847e-02 5.77553213e-01 5.11336863e-01 -5.25347888e-01
3.73599201e-01 6.56985521e-01 1.98630512e-01 -9.04056311e-01
-6.10002697e-01 -9.97665167e-01 -7.84465730e-01 1.71153396e-01
7.54241347e-01 2.14916334e-01 -9.19238567e-01 1.79579568e+00
2.46309608e-01 1.38425738e-01 -2.10212916e-01 1.27688456e+00
8.60170484e-01 -9.12210345e-02 2.98444092e-01 5.51245928e-01
2.01885486e+00 -1.21093726e+00 -3.32497448e-01 -4.29607540e-01
2.48447567e-01 -6.77386761e-01 9.61537600e-01 1.01005271e-01
-1.18435383e+00 -1.09333813e+00 -8.67068291e-01 -1.64877191e-01
-7.41915345e-01 4.51675177e-01 4.58282232e-01 9.89331007e-01
-1.11253679e+00 4.78670925e-01 -2.75621951e-01 -6.43125296e-01
6.37933314e-01 5.27480721e-01 -5.03382504e-01 -1.26942113e-01
-1.31044447e+00 1.13060021e+00 3.42622548e-01 1.73398212e-01
-4.36105430e-01 -7.67174900e-01 -8.33122969e-01 1.92126483e-01
4.08065975e-01 -1.32077849e+00 1.06866670e+00 -9.07361984e-01
-8.28390718e-01 1.41992199e+00 -6.68076724e-02 -6.53961837e-01
1.07092786e+00 -6.05053425e-01 -3.70773733e-01 1.16726987e-01
4.88625467e-01 7.45769083e-01 7.95044184e-01 -1.02088130e+00
-8.13460290e-01 -6.37213290e-01 -1.62332520e-01 1.94564015e-01
-1.74405694e-01 6.17697716e-01 -7.85510778e-01 -8.12533259e-01
-3.90416294e-01 -8.98576498e-01 -3.18472207e-01 6.04487322e-02
-1.84185863e-01 -4.37290162e-01 3.24774683e-01 -9.03724730e-01
9.17414725e-01 -1.69634819e+00 1.05865747e-01 2.42209986e-01
-1.58432692e-01 5.21351278e-01 -3.58593643e-01 7.95490965e-02
4.44551222e-02 -4.28330302e-01 9.68506262e-02 -7.89366424e-01
2.16135159e-01 -1.03353433e-01 1.51796222e-01 6.52812481e-01
3.62719558e-02 1.17214727e+00 -6.63091600e-01 -4.26201999e-01
2.60215431e-01 6.36210859e-01 -4.32402253e-01 -3.11166886e-03
3.32216084e-01 1.84804782e-01 -7.72354528e-02 9.27277923e-01
5.43909788e-01 -2.74083227e-01 -3.23832452e-01 -2.70867348e-01
1.13323756e-01 -1.09521158e-01 -1.51162481e+00 1.89715910e+00
-2.47941807e-01 2.93377578e-01 3.89524102e-02 -8.34156334e-01
6.10805213e-01 1.45064712e-01 2.04771712e-01 -9.58631158e-01
1.12924345e-01 2.08138749e-01 -1.89391807e-01 -2.30768427e-01
4.80906785e-01 3.86023939e-01 -2.65200973e-01 -8.27802531e-03
1.70936108e-01 5.02136469e-01 1.69351891e-01 -2.37736236e-02
8.38851929e-01 9.05696079e-02 2.20854074e-01 -4.69617575e-01
9.86286104e-01 -1.31104305e-01 3.35290939e-01 1.39805686e+00
-5.04143178e-01 6.81276798e-01 -2.29853421e-01 -8.36336315e-01
-1.09110856e+00 -1.00445223e+00 -6.80048466e-02 1.44115460e+00
2.19408289e-01 -3.08323592e-01 -8.54769766e-01 -4.64228511e-01
3.84096175e-01 2.34157249e-01 -6.38825297e-01 -1.15084596e-01
-8.66511524e-01 -1.01427805e+00 1.05841780e+00 6.19130790e-01
7.77580619e-01 -1.12379146e+00 -8.28154981e-01 2.63470918e-01
-2.61134297e-01 -1.39514554e+00 -6.69844627e-01 -2.43550748e-01
-1.21120185e-01 -1.17200506e+00 -1.11419225e+00 -8.92755151e-01
4.09542024e-01 9.46950614e-02 1.36219418e+00 1.13658525e-01
-1.11631966e+00 5.79858840e-01 -1.45505304e-02 -3.39914173e-01
1.73161533e-02 1.17239520e-01 2.84681290e-01 1.07200798e-02
7.94389844e-01 -2.53148645e-01 -9.60341692e-01 3.66208702e-01
-3.86186957e-01 -2.57814258e-01 7.22598314e-01 9.84214604e-01
3.10094863e-01 -1.81762248e-01 3.22476447e-01 -1.93755984e-01
6.37067676e-01 3.91361788e-02 -6.40372455e-01 5.15372634e-01
-5.72866976e-01 -1.54244397e-02 1.08046383e-01 -3.86461049e-01
-8.94995511e-01 2.09831566e-01 -1.95178509e-01 -2.68291712e-01
-4.79000062e-01 -2.90129662e-01 7.98866600e-02 -3.86758059e-01
8.14660251e-01 3.64382654e-01 -1.52769417e-01 -5.51669061e-01
3.17867488e-01 4.65062261e-01 1.06120574e+00 -5.34455776e-01
8.63789976e-01 5.82880795e-01 -2.75701378e-02 -4.91818517e-01
-7.86758900e-01 -1.08087468e+00 -8.50975096e-01 -1.91231206e-01
9.34341311e-01 -1.33851039e+00 -9.80857670e-01 7.27943480e-01
-1.16554308e+00 1.33776903e-01 -1.18588217e-01 1.55299217e-01
-3.52133363e-01 5.35230100e-01 -5.46216905e-01 -8.14909101e-01
-7.74763048e-01 -9.83557224e-01 1.60730529e+00 4.18834239e-01
-2.95652747e-01 -5.90628684e-01 -6.02040626e-02 8.35031927e-01
4.61707652e-01 2.22295046e-01 9.40720215e-02 -8.19705248e-01
-5.22256851e-01 -5.83957434e-01 -6.13708138e-01 4.10575084e-02
-3.04461002e-01 -9.52546120e-01 -1.12435436e+00 -5.84850430e-01
-5.21744370e-01 -3.99953336e-01 1.22021365e+00 2.57102609e-01
9.80564356e-01 -3.42568918e-03 -5.20599723e-01 6.50592566e-01
1.44043088e+00 -4.82968271e-01 3.29404742e-01 6.53549671e-01
6.97813153e-01 6.13495827e-01 3.14535856e-01 3.02104324e-01
5.33568978e-01 1.40110719e+00 7.66922683e-02 -3.41168553e-01
-5.23254514e-01 -1.32568449e-01 -3.46897170e-02 -4.69338655e-01
-3.36810380e-01 3.64927389e-02 -8.70458782e-01 7.37644315e-01
-2.22400093e+00 -1.29745412e+00 -8.23131353e-02 2.09539676e+00
6.46769702e-01 -1.25964116e-02 6.81229651e-01 1.19686536e-02
8.10718596e-01 -8.88433009e-02 -1.63227946e-01 8.34644139e-02
-5.33791006e-01 3.78048152e-01 6.73795581e-01 2.99909800e-01
-1.63831902e+00 9.24024343e-01 5.43125057e+00 1.08776307e+00
-3.85097355e-01 3.86831731e-01 2.37129748e-01 -3.38585615e-01
5.18891454e-01 -5.74027002e-01 -1.28792512e+00 4.33429927e-01
4.48424101e-01 3.59319657e-01 3.10457319e-01 9.32128012e-01
-1.31164566e-01 -5.77676035e-02 -1.18370497e+00 1.62580979e+00
5.00283837e-01 -1.14317334e+00 -8.73698592e-02 -1.31392166e-01
4.42306042e-01 2.25166008e-02 5.67847900e-02 4.62529510e-01
7.60764480e-02 -9.91837382e-01 8.24360311e-01 4.63771611e-01
4.20567870e-01 -6.13177061e-01 8.89341891e-01 1.88919768e-01
-1.42392397e+00 -4.34498101e-01 -2.73405313e-01 1.40795097e-01
2.58849353e-01 1.97368473e-01 -4.44531649e-01 6.24901950e-01
1.25554550e+00 1.15411907e-01 -1.09021711e+00 1.40255308e+00
1.72764510e-01 -1.20216042e-01 -5.35809398e-01 1.55537009e-01
2.67293155e-01 2.05015376e-01 5.46777308e-01 1.75483048e+00
4.91592921e-02 -3.86445463e-01 4.05458778e-01 8.82559836e-01
1.24036744e-01 -3.55836749e-02 -1.75007924e-01 6.95598841e-01
1.66867599e-01 1.27688336e+00 -5.04370034e-01 -4.94227678e-01
-4.39724863e-01 1.49119079e+00 3.46095055e-01 1.29331231e-01
-8.11296701e-01 2.82763317e-03 7.71050274e-01 1.54767513e-01
5.19177318e-01 2.24765122e-01 -1.67901933e-01 -1.04685533e+00
3.02609295e-01 -9.77393031e-01 8.63638222e-01 -4.43074107e-01
-1.62534189e+00 4.87927139e-01 -9.94529128e-02 -6.95370197e-01
-2.44827613e-01 -5.70880949e-01 -3.84909600e-01 1.15587318e+00
-1.61964619e+00 -1.81426787e+00 -4.38096970e-01 8.93291593e-01
6.10094786e-01 -5.11524737e-01 8.33167255e-01 7.75879502e-01
-3.76183152e-01 1.25782669e+00 -3.58183682e-01 4.06864941e-01
8.11786771e-01 -1.18343091e+00 6.53663695e-01 9.35254991e-01
2.07775384e-01 5.16613841e-01 6.83258176e-01 -6.26641929e-01
-9.70770657e-01 -8.66489172e-01 1.16624415e+00 -8.13323677e-01
3.08278531e-01 -5.61147273e-01 -6.54733777e-01 2.59154409e-01
3.79680693e-02 1.03336334e-01 5.03874242e-01 2.51158237e-01
-6.78106248e-01 1.06150582e-02 -1.20089555e+00 4.84168500e-01
1.52099013e+00 -5.91526687e-01 -6.34226859e-01 4.86332476e-01
2.35855967e-01 -2.84267783e-01 -5.28877556e-01 4.68684912e-01
7.80634284e-01 -1.07800674e+00 1.95415461e+00 -5.94589889e-01
-2.58031368e-01 -2.35392630e-01 -3.78412046e-02 -8.02207768e-01
-6.22633755e-01 -3.58483136e-01 9.80704129e-02 9.14953291e-01
4.76346817e-03 -3.12550455e-01 8.08475852e-01 6.16298378e-01
2.62848556e-01 -3.71604919e-01 -1.21000481e+00 -1.10787499e+00
-3.38648826e-01 -1.45845506e-02 4.01034504e-01 5.63697517e-01
-5.46390593e-01 1.71978384e-01 -6.50358498e-01 4.41153407e-01
1.23443615e+00 3.79210263e-02 8.82140279e-01 -1.41459525e+00
-5.10287046e-01 -7.78598487e-01 -6.53326094e-01 -6.47917867e-01
7.43008107e-02 -7.62053728e-01 -1.74303398e-01 -1.34232807e+00
4.66828912e-01 -1.71193287e-01 -3.87790382e-01 2.37518549e-01
-5.97690761e-01 5.21572471e-01 4.75874752e-01 2.13066027e-01
-8.81886899e-01 3.63162041e-01 7.27770448e-01 -1.40838772e-01
4.31364290e-02 1.42450854e-01 -4.67746228e-01 6.08573258e-01
4.55582023e-01 -2.72023290e-01 1.53331056e-01 -3.78955752e-01
-9.64904390e-03 -4.87690747e-01 1.29495645e+00 -1.32053077e+00
5.97934842e-01 6.70552433e-01 8.88337076e-01 -5.64984024e-01
6.49748385e-01 -7.09141493e-01 1.21021755e-02 8.75141740e-01
-1.95080101e-01 1.33138731e-01 1.89316556e-01 4.76798773e-01
-1.27716988e-01 -3.62451673e-02 8.72232258e-01 -4.33397591e-01
-8.77792537e-01 2.53004640e-01 1.89172849e-01 -1.31989062e-01
7.41715729e-01 -4.71870780e-01 -1.87288299e-01 -6.39035087e-03
-6.67401314e-01 3.25871319e-01 1.13372304e-01 8.31626654e-01
2.96694010e-01 -1.29871452e+00 -9.54751551e-01 2.66797543e-01
1.96386084e-01 -4.53641623e-01 3.39719892e-01 5.89976251e-01
-1.58649191e-01 5.94261646e-01 -2.83368915e-01 -6.00784600e-01
-1.66621530e+00 4.79317486e-01 6.39625192e-01 -6.39732778e-01
-5.86226463e-01 1.27842820e+00 6.84230477e-02 -5.96232116e-01
4.79615450e-01 2.88739473e-01 -1.32636458e-01 8.31387714e-02
6.37950420e-01 4.85905528e-01 5.71412370e-02 -7.96892524e-01
-7.20395207e-01 7.82072008e-01 -1.55136526e-01 1.44507676e-01
1.10662651e+00 1.14880232e-02 1.54443190e-01 -4.20657426e-01
8.27456474e-01 -3.52134138e-01 -9.15256083e-01 -5.64283907e-01
1.66437551e-01 -4.44787323e-01 -2.88146168e-01 -1.03720248e+00
-6.56665683e-01 4.36396450e-01 1.14677477e+00 -1.73221275e-01
8.90992999e-01 2.49267310e-01 8.43027294e-01 4.51321393e-01
3.41406405e-01 -1.50048542e+00 1.28161743e-01 1.77188337e-01
9.55482423e-01 -1.54518902e+00 4.82086949e-02 -2.88744509e-01
-3.82290035e-01 6.53797984e-01 7.42173791e-01 -2.01272845e-01
2.59656966e-01 -1.15514852e-01 -9.20228362e-02 -3.09885532e-01
-4.16389219e-02 -7.73390532e-01 6.93090737e-01 6.54942930e-01
-8.14700276e-02 -1.41594941e-02 -3.73470969e-02 7.04231918e-01
-2.62502432e-01 -5.40559217e-02 -3.46792936e-01 5.71043015e-01
-2.46280089e-01 -1.04302895e+00 -8.12149823e-01 1.43597081e-01
-5.58915198e-01 -1.70112222e-01 -3.24253350e-01 7.84963191e-01
5.80110908e-01 7.25458682e-01 -2.17900947e-02 4.41277996e-02
7.11352944e-01 2.18326449e-01 7.69612551e-01 -1.76712334e-01
-1.24521160e+00 -2.20493019e-01 1.72505453e-01 -7.94291496e-01
-5.93303859e-01 -8.86294901e-01 -5.04507244e-01 -1.02620721e-01
-2.51372606e-01 -3.78798932e-01 4.49673980e-01 9.13660467e-01
3.88033956e-01 4.59205687e-01 -1.78850457e-01 -9.45022821e-01
-8.22437882e-01 -8.74533951e-01 -1.13794699e-01 9.54833329e-01
2.33523905e-01 -8.03857625e-01 9.02903378e-02 -1.90248936e-01] | [14.837800025939941, 0.8079224228858948] |
b60ff512-0105-46fd-89d6-42a756806e33 | sentiment-analysis-in-the-era-of-large | 2305.15005 | null | https://arxiv.org/abs/2305.15005v1 | https://arxiv.org/pdf/2305.15005v1.pdf | Sentiment Analysis in the Era of Large Language Models: A Reality Check | Sentiment analysis (SA) has been a long-standing research area in natural language processing. It can offer rich insights into human sentiments and opinions and has thus seen considerable interest from both academia and industry. With the advent of large language models (LLMs) such as ChatGPT, there is a great potential for their employment on SA problems. However, the extent to which existing LLMs can be leveraged for different sentiment analysis tasks remains unclear. This paper aims to provide a comprehensive investigation into the capabilities of LLMs in performing various sentiment analysis tasks, from conventional sentiment classification to aspect-based sentiment analysis and multifaceted analysis of subjective texts. We evaluate performance across 13 tasks on 26 datasets and compare the results against small language models (SLMs) trained on domain-specific datasets. Our study reveals that while LLMs demonstrate satisfactory performance in simpler tasks, they lag behind in more complex tasks requiring deeper understanding or structured sentiment information. However, LLMs significantly outperform SLMs in few-shot learning settings, suggesting their potential when annotation resources are limited. We also highlight the limitations of current evaluation practices in assessing LLMs' SA abilities and propose a novel benchmark, \textsc{SentiEval}, for a more comprehensive and realistic evaluation. Data and code during our investigations are available at \url{https://github.com/DAMO-NLP-SG/LLM-Sentiment}. | ['Lidong Bing', 'Sinno Jialin Pan', 'Bing Liu', 'Yue Deng', 'Wenxuan Zhang'] | 2023-05-24 | null | null | null | null | ['sentiment-analysis'] | ['natural-language-processing'] | [ 5.98624870e-02 -1.90534249e-01 -2.28479624e-01 -6.84110403e-01
-8.16657007e-01 -6.13287866e-01 7.18658745e-01 4.88523901e-01
-6.94859326e-01 4.78037894e-01 4.38223600e-01 -3.75377625e-01
3.76853466e-01 -5.11267424e-01 -4.59486805e-02 -5.09559512e-01
2.48144135e-01 2.34440416e-01 2.99723167e-03 -8.38874817e-01
4.57927406e-01 -7.49447569e-02 -1.44821346e+00 5.58794916e-01
7.08302140e-01 1.03439128e+00 4.82182950e-02 6.39736831e-01
-6.22246683e-01 1.01436031e+00 -8.17532241e-01 -9.19337511e-01
-1.28984556e-01 -1.32321075e-01 -8.92920673e-01 -3.44604515e-02
8.63623470e-02 3.51929277e-01 4.65797067e-01 9.17611837e-01
6.71257019e-01 3.38608891e-01 3.98148119e-01 -1.17357028e+00
-6.24978423e-01 6.23585403e-01 -4.43560541e-01 3.84622663e-01
5.53020418e-01 -7.12008327e-02 1.40169501e+00 -9.19977427e-01
4.59002227e-01 1.06480396e+00 7.60330677e-01 4.51498479e-01
-7.46642590e-01 -5.09354115e-01 1.79943562e-01 -3.74862924e-02
-7.63420880e-01 -5.91350734e-01 6.40451849e-01 -3.92968088e-01
1.39817774e+00 1.43648401e-01 2.66489357e-01 1.16447794e+00
3.81372631e-01 1.10977221e+00 1.30324376e+00 -6.41075253e-01
3.77149165e-01 5.87123275e-01 6.25084341e-01 3.52241009e-01
7.28713423e-02 -7.31703222e-01 -9.32845473e-01 -2.27709129e-01
-1.34652689e-01 -2.62241103e-02 1.07201159e-01 1.02929220e-01
-1.07688868e+00 1.09931290e+00 -1.72346868e-02 5.94967604e-01
-3.21671098e-01 -2.45682865e-01 9.42673922e-01 5.39769471e-01
1.12451029e+00 7.96939492e-01 -9.95541036e-01 -5.85643172e-01
-7.69021153e-01 1.20633505e-01 1.19138479e+00 7.17225015e-01
5.32504082e-01 2.65285466e-02 -2.97906902e-03 1.24846685e+00
2.31653839e-01 4.02001739e-01 9.50928032e-01 -5.18036306e-01
5.89878917e-01 7.91525066e-01 4.64781374e-02 -9.90577579e-01
-5.67498326e-01 -1.99018270e-01 -4.95636106e-01 -2.13916842e-02
1.75837487e-01 -3.63843650e-01 -5.93778014e-01 1.35922384e+00
6.98965639e-02 -3.07435721e-01 3.24676186e-01 5.24304926e-01
1.05524826e+00 6.80164456e-01 4.43064123e-01 -5.56970760e-02
1.54996920e+00 -1.13167596e+00 -6.93440914e-01 -8.77243280e-01
1.05878735e+00 -1.13507843e+00 1.58396375e+00 4.86917406e-01
-9.19745207e-01 -2.21555054e-01 -8.01708817e-01 -1.20916918e-01
-9.50603843e-01 -1.36881473e-03 9.16961133e-01 9.83166993e-01
-1.01547909e+00 3.47512484e-01 -6.97985709e-01 -6.35478377e-01
4.69488770e-01 2.07982302e-01 -1.49006516e-01 6.30611032e-02
-1.37603343e+00 1.11116636e+00 -9.84433815e-02 -5.61559689e-04
-2.11315066e-01 -3.64340425e-01 -1.08838546e+00 -5.59638180e-02
4.26967144e-01 -2.47967839e-01 1.70932972e+00 -1.14326704e+00
-1.46966708e+00 1.02671611e+00 -6.19095564e-01 -2.91658789e-01
3.22388415e-03 -3.46851557e-01 -6.78093910e-01 -2.36820862e-01
4.09893960e-01 2.21009120e-01 5.16465425e-01 -8.17330360e-01
-6.12455189e-01 -3.23192567e-01 3.31846356e-01 2.69967943e-01
-8.69031012e-01 7.72620320e-01 -3.08463067e-01 -5.49257994e-01
-5.35016537e-01 -9.16233659e-01 -5.25858939e-01 -6.19455278e-01
-2.50440121e-01 -4.50135171e-01 4.96453941e-01 -3.39262903e-01
1.41948426e+00 -1.96349216e+00 -2.47059494e-01 -7.86753297e-02
-1.21122316e-01 4.82819766e-01 -1.28143057e-01 8.33539546e-01
1.57035172e-01 3.65528077e-01 -2.34666973e-01 -9.35675621e-01
1.05841033e-01 1.88664868e-01 -2.67457485e-01 1.71035230e-01
1.10210843e-01 1.05735934e+00 -9.04850304e-01 -3.04293901e-01
2.21952766e-01 3.24219316e-01 -3.55796039e-01 -4.64335941e-02
-1.92553550e-01 1.03566036e-01 -6.86563253e-01 8.75033915e-01
1.41283512e-01 -5.28773904e-01 -5.34989825e-03 1.74343303e-01
-1.53236553e-01 4.50752646e-01 -7.90087640e-01 1.51753604e+00
-9.02050376e-01 6.70780361e-01 -5.20996489e-02 -9.83255327e-01
8.21712196e-01 3.58577400e-01 2.42467836e-01 -6.16029441e-01
5.56893528e-01 1.09833881e-01 -1.50620356e-01 -5.10766447e-01
8.02870870e-01 -5.00089109e-01 -5.97719908e-01 6.95910335e-01
1.80361480e-01 -4.08243716e-01 6.11040592e-01 3.19269240e-01
8.08628261e-01 -2.67652422e-01 5.86719215e-01 -3.09751153e-01
6.89494073e-01 2.60035872e-01 2.78028756e-01 5.09956717e-01
-3.30400974e-01 4.21189398e-01 5.95632136e-01 -4.28344458e-01
-4.88853812e-01 -3.81789058e-01 -4.78211641e-02 1.82300651e+00
-9.26993713e-02 -9.01078880e-01 -4.75271195e-01 -6.15799725e-01
-2.41338134e-01 9.91971433e-01 -6.20975077e-01 1.23588800e-01
2.78530978e-02 -1.11831737e+00 3.89157772e-01 4.33237225e-01
2.57219702e-01 -1.53790903e+00 -5.88021278e-01 6.10058345e-02
-2.14844540e-01 -1.29085302e+00 -2.39913359e-01 2.28982940e-01
-7.46898234e-01 -8.13628972e-01 -3.63671303e-01 -7.40861416e-01
3.82525682e-01 3.86536270e-01 1.42582488e+00 -1.55289724e-01
6.95605800e-02 4.64812815e-01 -7.43478239e-01 -1.17979884e+00
-3.89414668e-01 4.08278078e-01 1.20588496e-01 -1.29478917e-01
1.08584917e+00 -1.85315803e-01 -4.24896389e-01 2.57236421e-01
-9.92253900e-01 -2.36846715e-01 3.14866334e-01 5.14934599e-01
1.63032472e-01 -1.02598220e-01 8.39238465e-01 -1.39208019e+00
1.31641340e+00 -6.95013881e-01 -1.38246775e-01 6.74195364e-02
-6.53336108e-01 -3.79643798e-01 7.30192959e-01 -9.30303782e-02
-1.24364626e+00 -4.20362383e-01 -4.89175022e-01 2.10038841e-01
-2.06639871e-01 1.00878954e+00 2.05811471e-01 1.27384558e-01
7.36566663e-01 -6.23702332e-02 -1.74546055e-02 -3.31944913e-01
2.19620988e-01 8.85018826e-01 -1.29178077e-01 -5.16429543e-01
2.05543607e-01 5.83086669e-01 -5.82721353e-01 -1.01934242e+00
-1.53030658e+00 -9.79319632e-01 -4.52678800e-01 -1.38526067e-01
7.86300600e-01 -9.97099459e-01 -5.68800092e-01 4.90855604e-01
-7.14145243e-01 -3.40521276e-01 -1.71678886e-01 2.27815047e-01
-2.35730156e-01 3.58602524e-01 -8.18790138e-01 -8.58793259e-01
-8.15481603e-01 -1.08061123e+00 1.19630122e+00 3.18432450e-01
-9.10990357e-01 -1.60139465e+00 2.36196890e-01 8.09222937e-01
5.47577500e-01 -9.14537311e-02 5.74814856e-01 -1.12516153e+00
4.14594352e-01 -7.15556026e-01 2.62494534e-01 7.30765462e-01
2.52502173e-01 -2.53967922e-02 -1.20098746e+00 -1.48965433e-01
1.32132292e-01 -8.43156576e-01 6.97840393e-01 3.31123382e-01
7.63400495e-01 5.22714760e-03 2.71477234e-02 1.77344251e-02
1.24246144e+00 -4.33887821e-03 3.15029383e-01 7.68374801e-01
4.50326681e-01 8.51615667e-01 8.66344154e-01 5.42284667e-01
6.85952365e-01 8.27886164e-02 7.65827000e-02 2.97703426e-02
4.62622195e-01 1.94499582e-01 7.42223740e-01 1.16148973e+00
-7.53662512e-02 -4.65517133e-01 -1.12537253e+00 6.63798392e-01
-1.74811196e+00 -7.57409036e-01 -1.35352224e-01 1.85072577e+00
7.36779034e-01 4.68742073e-01 -8.43223482e-02 3.55632901e-02
2.50763208e-01 6.78450346e-01 -3.34850550e-01 -1.15762472e+00
-2.74441898e-01 3.42901409e-01 7.92121887e-02 3.26592296e-01
-1.14359522e+00 1.24442959e+00 5.96959877e+00 7.86342204e-01
-9.76198614e-01 2.29496360e-01 8.12991142e-01 -2.02558532e-01
-3.23092490e-01 -1.33905321e-01 -9.07919884e-01 2.45505497e-01
1.26922202e+00 -2.76746511e-01 -1.47397444e-01 1.05484498e+00
3.96803916e-01 -3.32347035e-01 -6.58085287e-01 5.58568060e-01
3.48369658e-01 -1.27279413e+00 -5.94900325e-02 -3.29718053e-01
9.34459627e-01 4.84792739e-01 3.36644910e-02 7.37610459e-01
4.24516112e-01 -8.21716070e-01 4.49313998e-01 3.87239084e-02
3.91987741e-01 -5.87788641e-01 1.26753521e+00 3.48540157e-01
-1.00035310e+00 -3.37045975e-02 -4.15184706e-01 -4.96374428e-01
4.26393211e-01 5.41222751e-01 -6.37293637e-01 3.53696316e-01
8.71337831e-01 1.07401836e+00 -6.21299028e-01 4.36323583e-01
-2.83930987e-01 9.09028709e-01 1.01473171e-03 -4.62956309e-01
5.68354309e-01 -2.44118929e-01 2.55122304e-01 1.55724883e+00
-1.06096547e-02 -4.37838547e-02 1.17423229e-01 1.75328851e-01
-1.03444522e-02 6.06931567e-01 -5.13513803e-01 -4.76496190e-01
2.59779077e-02 1.63597834e+00 -8.94230068e-01 -4.27533835e-01
-1.02449572e+00 6.33161724e-01 2.47744307e-01 2.45882213e-01
-4.03140783e-01 -2.93079138e-01 8.73756468e-01 -2.57724136e-01
3.27419564e-02 -1.43179342e-01 -5.70505619e-01 -1.48802865e+00
-1.11183628e-01 -1.09544671e+00 5.87024510e-01 -7.80850410e-01
-1.51059377e+00 7.38083839e-01 -2.97980905e-01 -1.01121271e+00
-2.31436551e-01 -9.06388938e-01 -9.99462724e-01 8.00608099e-01
-1.64931595e+00 -1.13317895e+00 -3.05007417e-02 5.17119646e-01
1.14153242e+00 -2.99574822e-01 1.08578539e+00 1.49595514e-02
-6.27120435e-01 3.34347844e-01 3.95485684e-02 2.29520667e-02
1.00062788e+00 -1.19977665e+00 6.37992382e-01 6.01788819e-01
1.19542517e-01 8.00706148e-01 8.35815251e-01 -4.99195129e-01
-1.24976861e+00 -7.15209723e-01 1.39729559e+00 -9.22820926e-01
1.24375272e+00 -4.41231221e-01 -7.28546143e-01 6.80480361e-01
3.83168459e-01 -4.50832844e-01 1.46082437e+00 6.18452787e-01
-1.91082641e-01 2.04544634e-01 -8.32886398e-01 7.00074196e-01
4.34955567e-01 -6.35128975e-01 -8.35017502e-01 4.17372406e-01
6.29643023e-01 -8.67020711e-03 -8.26105356e-01 1.57705009e-01
4.83463436e-01 -9.16282892e-01 6.80814862e-01 -7.51293123e-01
8.24063420e-01 1.20541289e-01 -1.44769326e-01 -1.53381073e+00
1.63001031e-01 -4.79717731e-01 2.15593114e-01 1.29050839e+00
7.95044661e-01 -8.11947286e-01 6.92121506e-01 8.63672495e-01
-1.21655673e-01 -9.73262608e-01 -3.71194541e-01 -3.87025952e-01
2.54829973e-01 -1.04198265e+00 1.66477606e-01 1.16921318e+00
5.25716186e-01 7.11109638e-01 -1.45955592e-01 -2.07680613e-01
1.12346157e-01 2.27833226e-01 6.77437305e-01 -1.11860991e+00
2.30791233e-02 -5.52344441e-01 -7.94146881e-02 -6.34004772e-01
3.49474251e-01 -7.87745416e-01 -3.32116634e-02 -1.69659555e+00
3.22854258e-02 -1.71509340e-01 -4.54860687e-01 5.41889429e-01
-3.56972188e-01 3.71483654e-01 1.68492585e-01 3.01775560e-02
-9.64391351e-01 4.41403836e-01 1.02155709e+00 8.64981562e-02
-2.01147616e-01 3.06251168e-01 -1.20629537e+00 1.14085948e+00
1.10923171e+00 -2.97077656e-01 -3.37917507e-01 -2.41862684e-01
7.91801929e-01 -3.84217292e-01 -3.68716031e-01 -4.74945188e-01
2.49016121e-01 -7.53326565e-02 -2.65558138e-02 -2.08284423e-01
3.62885982e-01 -4.27516848e-01 -7.53322124e-01 -8.86823535e-02
-4.39592808e-01 2.61216640e-01 4.10467982e-01 2.56120265e-01
-5.80755770e-01 -3.83274496e-01 2.98470259e-01 -3.50325376e-01
-1.05741668e+00 1.52561411e-01 -7.63542533e-01 1.61390468e-01
6.76387668e-01 -8.72363448e-02 -2.77129173e-01 -5.98515153e-01
-6.03530526e-01 4.65213448e-01 3.67048442e-01 8.39110374e-01
3.65558624e-01 -7.21079469e-01 -4.81492400e-01 -5.01892716e-02
4.94094372e-01 -1.92146942e-01 1.85381249e-01 9.04306173e-01
-2.85425842e-01 4.79711473e-01 1.73127994e-01 -1.85824960e-01
-1.34484839e+00 2.59967417e-01 -1.15858033e-01 -4.37670916e-01
-1.35169163e-01 9.87130225e-01 4.14866842e-02 -7.58567214e-01
-6.15733787e-02 -2.07659841e-01 -7.24159718e-01 5.48668563e-01
7.55278528e-01 2.34449819e-01 1.75101370e-01 -8.64271581e-01
-3.65469068e-01 4.34627593e-01 -2.51660079e-01 1.51654016e-02
1.56411457e+00 -3.93487573e-01 -2.49317557e-01 9.90124583e-01
1.11119354e+00 2.40048468e-01 -5.87162912e-01 -2.96198696e-01
4.41620111e-01 -1.78841978e-01 -7.64232129e-03 -7.69201458e-01
-7.20155060e-01 8.83013010e-01 -8.71526822e-02 5.86493611e-01
1.02370322e+00 1.29537955e-01 7.70834923e-01 5.68120599e-01
3.33170265e-01 -1.29803681e+00 1.28409937e-01 9.37913120e-01
6.19320750e-01 -1.81968439e+00 5.46850711e-02 -1.77162722e-01
-1.46721709e+00 9.72620249e-01 5.02944708e-01 -4.11293283e-02
8.58545244e-01 2.33409002e-01 6.10016584e-01 -4.96903569e-01
-9.40170109e-01 -2.98211694e-01 2.44798496e-01 8.97406191e-02
9.86971319e-01 1.22897699e-02 -3.27389091e-01 8.99241209e-01
-6.00177050e-01 -4.04285155e-02 6.65737212e-01 1.17954826e+00
-4.25052881e-01 -1.12403536e+00 -1.17315557e-02 7.11863637e-01
-1.00270319e+00 -4.47309226e-01 -5.43002546e-01 5.45500934e-01
-2.92263210e-01 1.43317115e+00 -1.86491549e-01 -2.15507641e-01
3.32479507e-01 3.55474144e-01 -2.92506933e-01 -1.17316794e+00
-9.98551965e-01 -6.84093237e-02 5.63441634e-01 -5.00611186e-01
-8.52970958e-01 -7.61939049e-01 -1.01314092e+00 -2.94065207e-01
-3.54688436e-01 3.49186838e-01 7.57559776e-01 1.14105272e+00
2.48066619e-01 3.06884885e-01 4.14096057e-01 -6.84823632e-01
-3.08444738e-01 -1.10656095e+00 -6.11019969e-01 3.81911129e-01
1.27850041e-01 -2.01600403e-01 -6.07920527e-01 4.06980738e-02] | [11.302867889404297, 6.899796009063721] |
e5229c6c-9ff3-4e12-9557-0f0e7ff992c5 | low-resource-quadratic-forms-for-knowledge | null | null | https://aclanthology.org/2021.sustainlp-1.1 | https://aclanthology.org/2021.sustainlp-1.1.pdf | Low Resource Quadratic Forms for Knowledge Graph Embeddings | We address the problem of link prediction between entities and relations of knowledge graphs. State of the art techniques that address this problem, while increasingly accurate, are computationally intensive. In this paper we cast link prediction as a sparse convex program whose solution defines a quadratic form that is used as a ranking function. The structure of our convex program is such that standard support vector machine software packages, which are numerically robust and efficient, can solve it. We show that on benchmark data sets, our model’s performance is competitive with state of the art models, but training times can be reduced by a factor of 40 using only CPU-based (and not GPU-accelerated) computing resources. This approach may be suitable for applications where balancing the demands of graph completion performance against computational efficiency is a desirable trade-off. | ['Glenn Fung', 'Devin Conathan', 'Jeffery Kline', 'Zachary Zhou'] | null | null | null | null | emnlp-sustainlp-2021-11 | ['knowledge-graph-embeddings', 'knowledge-graph-embeddings'] | ['graphs', 'methodology'] | [ 1.09638259e-01 2.19227806e-01 -5.65882087e-01 -2.45820180e-01
-7.99046040e-01 -6.52222931e-01 4.32726145e-01 5.77838957e-01
-2.94880748e-01 6.73624635e-01 -1.15144506e-01 -6.39395118e-01
-4.18298572e-01 -1.00827384e+00 -7.92943180e-01 -2.37118214e-01
-5.23384690e-01 8.66960585e-01 3.43099982e-01 -3.94378245e-01
1.20560296e-01 5.03504872e-01 -1.36129439e+00 1.05911657e-01
1.01234066e+00 8.02394152e-01 -2.37045005e-01 6.77919090e-01
-3.68656293e-02 7.43327618e-01 -1.70954950e-02 -7.94659793e-01
3.49170089e-01 1.61270708e-01 -1.12561965e+00 -1.59081578e-01
6.48741841e-01 6.07390441e-02 -4.56901163e-01 9.49251354e-01
1.68411508e-01 2.01005876e-01 5.32465518e-01 -1.45279765e+00
-2.50993520e-01 3.84777486e-01 -7.25488365e-01 1.69622317e-01
6.16623759e-01 -6.18383586e-01 1.63233292e+00 -1.01757514e+00
6.30096316e-01 9.38246548e-01 8.55078816e-01 -1.88943073e-01
-1.76789749e+00 -1.94095820e-01 3.82273048e-02 4.60186988e-01
-1.61185884e+00 -4.90720659e-01 4.62357134e-01 -5.31551123e-01
1.14585853e+00 6.51863396e-01 7.13820279e-01 3.19979608e-01
9.26854163e-02 5.65320313e-01 6.36921883e-01 -3.10879678e-01
3.63234356e-02 2.02761844e-01 3.32958192e-01 1.13066792e+00
7.19860017e-01 -1.34247020e-01 -5.68063200e-01 -7.07078695e-01
6.01946354e-01 -3.69883507e-01 -2.81962842e-01 -8.73360395e-01
-1.07665503e+00 1.01134539e+00 3.87063414e-01 -1.62541822e-01
-2.02713206e-01 1.83107913e-01 3.90401155e-01 5.36525488e-01
5.87609529e-01 4.03562516e-01 -5.59340775e-01 6.46140799e-02
-9.61120963e-01 1.67617306e-01 1.40635884e+00 9.77247179e-01
8.14525425e-01 -2.76781321e-01 1.13880977e-01 7.44317591e-01
2.24975303e-01 8.69242772e-02 -2.22624883e-01 -7.53459811e-01
7.19635427e-01 6.13102555e-01 1.03202514e-01 -1.50523937e+00
-5.92697263e-01 -4.92123991e-01 -5.58470547e-01 -1.36348486e-01
4.83752131e-01 -1.26635926e-02 -2.73578852e-01 1.29283702e+00
5.95481396e-01 3.47061723e-01 -9.79544073e-02 6.83969915e-01
6.40149951e-01 7.06750989e-01 -6.20398149e-02 -3.40116113e-01
1.30558085e+00 -1.11159766e+00 -2.89216697e-01 -2.13944316e-01
1.16522050e+00 -7.82863438e-01 8.15008700e-01 1.90239787e-01
-1.16833889e+00 -6.15349747e-02 -1.06344831e+00 -2.65021741e-01
-2.14610741e-01 1.86414659e-01 1.24737144e+00 7.10455596e-01
-1.19863319e+00 7.60039091e-01 -8.05198133e-01 -4.15024191e-01
1.02140605e-01 7.33998835e-01 -5.04074633e-01 -3.75518352e-02
-8.98631036e-01 1.19563985e+00 4.60273415e-01 -2.83042509e-02
-1.73077315e-01 -9.62542892e-01 -8.34826529e-01 3.87861401e-01
6.75890744e-01 -7.60919154e-01 8.81193936e-01 -4.70678687e-01
-1.07902610e+00 9.09255803e-01 -9.62234065e-02 -4.12761331e-01
3.78707767e-01 -3.57654365e-03 -3.90625268e-01 -2.24624425e-02
-4.34054956e-02 4.67723422e-02 5.65264404e-01 -1.00849581e+00
-4.91438627e-01 -2.15717107e-01 7.82727748e-02 1.89251781e-01
-5.82883835e-01 1.71421226e-02 -7.58724391e-01 -3.43418419e-01
1.00454070e-01 -1.22692895e+00 -3.81699085e-01 2.82820351e-02
-4.51195061e-01 -2.27086723e-01 4.97639060e-01 -7.23394096e-01
1.29458928e+00 -1.68205011e+00 2.47302845e-01 9.21269596e-01
3.71095449e-01 2.13877991e-01 -7.02896565e-02 6.56100214e-01
-1.01260573e-01 -5.15698493e-02 -1.26001969e-01 -1.54281586e-01
-1.09307617e-01 1.99394271e-01 -5.80148175e-02 8.26291263e-01
-3.00363693e-02 8.45398545e-01 -8.38139117e-01 -5.91267109e-01
-9.76317227e-02 3.45050365e-01 -7.22257078e-01 -1.66841615e-02
-1.96567550e-01 -1.49631098e-01 -5.44235706e-01 4.87994850e-01
4.36369181e-01 -8.21494222e-01 7.86907434e-01 -3.75788331e-01
1.05347328e-01 2.52335936e-01 -1.51857710e+00 1.38784623e+00
-4.02962625e-01 5.85333884e-01 2.75703520e-01 -1.38810980e+00
7.62376785e-01 -1.21624693e-02 6.21999800e-01 -8.17144513e-02
-1.38545455e-02 2.59152651e-01 -1.12776630e-01 -1.82704166e-01
6.12791240e-01 2.06582919e-01 2.80745506e-01 2.93310046e-01
-2.05387726e-01 3.90648320e-02 4.69494164e-01 5.12609422e-01
1.35178077e+00 -4.77398336e-02 3.99506837e-01 -6.70298398e-01
4.53138649e-01 4.42305952e-01 4.52882409e-01 4.68133211e-01
3.62423927e-01 4.34922390e-02 8.45756948e-01 -5.16857982e-01
-9.76058662e-01 -7.29157329e-01 -1.52145043e-01 1.29591131e+00
9.87150222e-02 -8.18869054e-01 -3.82673144e-01 -5.82998812e-01
3.48200262e-01 3.27021003e-01 -3.03873479e-01 1.17757037e-01
-5.70457339e-01 -7.58700252e-01 1.56280547e-01 4.81525868e-01
-2.78132081e-01 -2.65142918e-01 1.34869181e-02 2.93850601e-01
7.58299753e-02 -1.29977059e+00 -4.58675027e-01 8.13826621e-02
-9.58950996e-01 -1.36019456e+00 -3.62751216e-01 -9.17049408e-01
8.99315000e-01 3.56077909e-01 1.46888399e+00 4.27745223e-01
-3.21159840e-01 4.71617669e-01 -6.56305179e-02 -5.22561222e-02
-4.81843390e-02 3.12748760e-01 1.22842938e-02 -1.12128124e-01
8.87900870e-03 -4.83947843e-01 -3.82607579e-01 1.46394476e-01
-5.62830806e-01 1.19916856e-01 4.35976386e-01 9.64353979e-01
7.86581576e-01 6.89682141e-02 2.72065550e-01 -1.38220453e+00
7.23760843e-01 -6.02756679e-01 -9.15208220e-01 5.81034780e-01
-9.67091978e-01 2.32517242e-01 4.77668196e-01 -9.40414295e-02
-5.56903660e-01 3.99305969e-01 1.95884421e-01 -2.30666220e-01
5.96627533e-01 1.12628865e+00 1.96341857e-01 -7.63559699e-01
6.47580802e-01 -1.59321666e-01 4.05906903e-04 -3.13103586e-01
4.30549324e-01 1.59757704e-01 3.17790240e-01 -6.74117088e-01
9.86891091e-01 3.56733948e-01 6.03791058e-01 -7.65889823e-01
-1.01234615e+00 -7.31920958e-01 -4.19773638e-01 2.01826543e-02
1.76586494e-01 -9.54478502e-01 -8.71192873e-01 -4.90565032e-01
-9.40380871e-01 -4.94151413e-02 -9.83713288e-03 2.96296269e-01
-5.59758067e-01 4.60777134e-01 -5.53388000e-01 -5.58610618e-01
-3.32744807e-01 -7.24155128e-01 7.86817014e-01 -1.59290224e-01
-2.64559478e-01 -1.16989076e+00 1.71877667e-01 6.42480373e-01
2.40013286e-01 3.83345455e-01 9.92540061e-01 -6.84286535e-01
-6.70267224e-01 -4.79206920e-01 -5.40413618e-01 -1.50718525e-01
-1.94496915e-01 6.12221695e-02 -3.37388068e-01 -5.26431978e-01
-5.92621326e-01 -1.02004506e-01 8.45226884e-01 1.07363977e-01
1.02072215e+00 -5.69729805e-01 -7.55910575e-01 5.62188923e-01
1.61536860e+00 -4.71363842e-01 2.56582141e-01 1.49893418e-01
9.60733533e-01 6.62616491e-01 7.42717206e-01 3.79875600e-01
7.31675386e-01 9.35773671e-01 2.71977097e-01 -1.84573680e-01
2.26200655e-01 -1.88539013e-01 7.00711310e-02 9.57519710e-01
-3.62885118e-01 -3.69591080e-02 -1.25047994e+00 6.95421338e-01
-2.48437619e+00 -8.12148690e-01 -6.16214991e-01 2.15061951e+00
8.15716267e-01 9.26478133e-02 3.03192168e-01 1.26421124e-01
5.58232486e-01 -9.93371233e-02 -3.06164473e-01 -4.35915411e-01
1.05890356e-01 2.14375541e-01 8.11173141e-01 7.89612293e-01
-1.12022758e+00 7.96482623e-01 6.94355869e+00 7.50313103e-01
-6.51199758e-01 4.75395322e-02 4.06674206e-01 -1.89986691e-01
-4.05492991e-01 1.87695071e-01 -6.69154167e-01 8.51428807e-02
1.06292570e+00 -6.04690194e-01 6.20482624e-01 9.82613325e-01
-8.18065181e-02 4.88651246e-02 -1.38911569e+00 1.08575010e+00
7.12776463e-03 -1.66181314e+00 -4.04787332e-01 3.29605520e-01
9.80829000e-01 1.02227159e-01 -1.42599329e-01 1.50286227e-01
4.88675714e-01 -1.05715358e+00 2.52354711e-01 3.30434203e-01
6.49205983e-01 -8.87667179e-01 7.00340688e-01 1.62756428e-01
-1.55492830e+00 1.02098204e-01 -4.10168380e-01 -8.02997202e-02
1.45849258e-01 8.85067761e-01 -8.28073740e-01 6.86175644e-01
4.71652955e-01 6.74541533e-01 -5.42426050e-01 1.16212177e+00
-5.07805571e-02 5.32557309e-01 -6.81699872e-01 -1.26045838e-01
-4.75730114e-02 -3.93598855e-01 4.41774935e-01 1.22713006e+00
2.45964572e-01 1.97241619e-01 4.62791383e-01 2.45586187e-01
-3.06995898e-01 5.73116183e-01 -6.66319311e-01 1.76808070e-02
5.34971535e-01 1.36444080e+00 -7.26031959e-01 -1.80281907e-01
-6.77145660e-01 7.17468381e-01 9.97102320e-01 1.52246982e-01
-7.54681230e-01 -2.27818131e-01 6.48647070e-01 2.71481186e-01
3.73722255e-01 -4.84919637e-01 -5.51262498e-01 -1.26729095e+00
2.47470930e-01 -6.50519431e-01 7.09484339e-01 -2.30810881e-01
-1.24725914e+00 2.55314052e-01 -1.51355028e-01 -8.63650680e-01
-9.86652151e-02 -6.14902556e-01 -3.19584906e-01 5.99329174e-01
-1.37533593e+00 -9.79618430e-01 -9.58743021e-02 5.61125934e-01
-1.72547638e-01 -3.00121065e-02 9.52552378e-01 4.16594893e-01
-6.47842348e-01 6.74625635e-01 3.48200917e-01 -1.00854747e-01
2.44059965e-01 -1.30748296e+00 2.40690574e-01 6.54808879e-01
6.17800891e-01 4.82280225e-01 7.82011390e-01 -6.13648057e-01
-2.00349236e+00 -1.05191398e+00 1.22099745e+00 -4.13751066e-01
1.16953862e+00 -1.56408563e-01 -7.83884346e-01 8.36113393e-01
-2.61505276e-01 4.75903630e-01 1.02127612e+00 1.00443172e+00
-3.57094675e-01 -1.39804184e-01 -1.01280379e+00 5.47458351e-01
1.19741428e+00 -5.59905767e-01 -1.21264331e-01 8.11615467e-01
3.12176853e-01 -5.04357219e-01 -1.41449630e+00 2.84583926e-01
4.74843293e-01 -5.45315266e-01 1.28434002e+00 -9.38231945e-01
1.34000644e-01 -1.85435653e-01 -7.97510222e-02 -1.20554948e+00
-5.15999675e-01 -8.81247580e-01 -8.92944455e-01 7.17770338e-01
7.00893402e-01 -5.26469648e-01 1.21529114e+00 8.72232378e-01
4.24093217e-01 -1.13212144e+00 -8.60220611e-01 -9.41364765e-01
-2.97903150e-01 -2.66160876e-01 2.94596940e-01 1.09291124e+00
4.27644521e-01 6.28380299e-01 -4.66349721e-01 3.34343046e-01
7.62923241e-01 3.97245944e-01 8.26286554e-01 -1.61151791e+00
-3.90971839e-01 -2.36568645e-01 -6.36052430e-01 -7.63246715e-01
3.99325728e-01 -1.19052124e+00 -5.70074975e-01 -1.65241885e+00
2.10918784e-01 -7.04831898e-01 -8.93471017e-02 6.41898274e-01
-1.86995745e-01 2.08925426e-01 1.31289080e-01 1.94800377e-01
-5.66711783e-01 3.12951088e-01 7.57965326e-01 -3.58256996e-02
-2.98904568e-01 8.11865274e-03 -7.05986619e-01 6.69686854e-01
7.00250983e-01 -4.74768788e-01 -4.03109729e-01 -4.07215804e-01
8.18896532e-01 3.40998352e-01 2.00534895e-01 -5.67680776e-01
7.06587136e-01 -2.24525124e-01 -3.64303589e-04 -4.31158304e-01
5.10092378e-01 -9.30639684e-01 2.63621867e-01 3.99427295e-01
-2.15564847e-01 1.33348638e-02 -3.74340080e-02 9.66286957e-01
-9.76409540e-02 -1.96677417e-01 5.11160553e-01 3.57656002e-01
-5.19508719e-01 5.21962702e-01 4.82266955e-02 1.00042284e-01
1.22633123e+00 -2.38526543e-03 -2.77878672e-01 -4.90673482e-01
-7.16972709e-01 3.51791859e-01 2.70820022e-01 8.14406648e-02
4.44072604e-01 -1.40324354e+00 -8.81799281e-01 -3.98850352e-01
9.57405940e-02 -2.86136836e-01 -2.63190240e-01 1.01118684e+00
-7.54706621e-01 5.00107646e-01 3.27980667e-01 -2.99979478e-01
-1.62811399e+00 7.74673581e-01 -9.07284245e-02 -6.93249404e-01
-5.72084427e-01 8.67939055e-01 -4.20857877e-01 -3.11194956e-01
1.34946793e-01 1.45404935e-01 -8.75571221e-02 9.28490534e-02
2.34885037e-01 6.95412397e-01 2.91350693e-01 -4.89069581e-01
-6.27193689e-01 4.20854211e-01 -6.84125200e-02 3.93725306e-01
1.57015228e+00 1.64292857e-01 -3.33324701e-01 -3.66762117e-03
1.36110461e+00 1.86136499e-01 -6.56440020e-01 -4.14735079e-01
4.24046874e-01 -4.86211091e-01 3.30453902e-01 -2.09040925e-01
-1.14404702e+00 1.77182660e-01 1.26288217e-02 6.02554798e-01
8.80187213e-01 2.02194408e-01 6.74859643e-01 8.10940027e-01
5.27001977e-01 -1.13269341e+00 -3.70175719e-01 4.29353654e-01
6.72577262e-01 -1.35570061e+00 6.83710635e-01 -1.41066730e+00
-3.00882965e-01 9.93366301e-01 3.80606651e-01 -4.77271974e-01
9.72672641e-01 2.74875700e-01 -5.70621073e-01 -3.41999650e-01
-1.02401614e+00 -2.55350918e-01 7.29634345e-01 3.81020874e-01
4.60830867e-01 3.41743201e-01 -5.71561992e-01 3.56710762e-01
-4.53611732e-01 -3.89999151e-01 3.75442326e-01 8.45701814e-01
-2.39456460e-01 -1.22261214e+00 -1.85506523e-01 8.96352828e-01
-3.39144677e-01 -2.30973184e-01 -2.00026795e-01 5.05643368e-01
-3.86765957e-01 1.09726846e+00 -1.02976792e-01 -3.07398081e-01
3.93642545e-01 -2.41273254e-01 4.96188611e-01 -7.45068908e-01
-3.81773114e-01 -3.71099710e-01 8.80530715e-01 -8.41626585e-01
-2.60720402e-01 -5.68278134e-01 -1.09357405e+00 -8.23721349e-01
-5.08913338e-01 2.20363423e-01 5.24890125e-01 8.52027953e-01
4.72857267e-01 1.70490906e-01 5.53043425e-01 -5.56834698e-01
-5.65560400e-01 -3.27283442e-01 -6.78521395e-01 6.99915783e-03
-5.85337915e-02 -6.92804754e-01 -3.98306921e-02 -1.03724405e-01] | [7.155229568481445, 5.5803680419921875] |
e7d6d345-16dc-473a-b82b-9c45ac157d7a | tight-and-fast-generalization-error-bound-of | 2305.07971 | null | https://arxiv.org/abs/2305.07971v1 | https://arxiv.org/pdf/2305.07971v1.pdf | Tight and fast generalization error bound of graph embedding in metric space | Recent studies have experimentally shown that we can achieve in non-Euclidean metric space effective and efficient graph embedding, which aims to obtain the vertices' representations reflecting the graph's structure in the metric space. Specifically, graph embedding in hyperbolic space has experimentally succeeded in embedding graphs with hierarchical-tree structure, e.g., data in natural languages, social networks, and knowledge bases. However, recent theoretical analyses have shown a much higher upper bound on non-Euclidean graph embedding's generalization error than Euclidean one's, where a high generalization error indicates that the incompleteness and noise in the data can significantly damage learning performance. It implies that the existing bound cannot guarantee the success of graph embedding in non-Euclidean metric space in a practical training data size, which can prevent non-Euclidean graph embedding's application in real problems. This paper provides a novel upper bound of graph embedding's generalization error by evaluating the local Rademacher complexity of the model as a function set of the distances of representation couples. Our bound clarifies that the performance of graph embedding in non-Euclidean metric space, including hyperbolic space, is better than the existing upper bounds suggest. Specifically, our new upper bound is polynomial in the metric space's geometric radius $R$ and can be $O(\frac{1}{S})$ at the fastest, where $S$ is the training data size. Our bound is significantly tighter and faster than the existing one, which can be exponential to $R$ and $O(\frac{1}{\sqrt{S}})$ at the fastest. Specific calculations on example cases show that graph embedding in non-Euclidean metric space can outperform that in Euclidean space with much smaller training data than the existing bound has suggested. | ['Kenji Yamanishi', 'Feng Tian', 'Jing Wang', 'Taiji Suzuki', 'Atsushi Nitanda', 'Atsushi Suzuki'] | 2023-05-13 | null | null | null | null | ['graph-embedding'] | ['graphs'] | [-1.64754510e-01 5.35644710e-01 6.51634634e-02 -1.23669684e-01
-4.07128602e-01 -6.11453950e-01 5.05280774e-03 1.87383458e-01
-5.57103515e-01 4.55956489e-01 -1.48154184e-01 -7.14289129e-01
-7.04124033e-01 -1.13216794e+00 -5.21507442e-01 -9.30864692e-01
-6.44919276e-01 3.29716474e-01 4.06982541e-01 -4.07821864e-01
3.72584403e-01 5.46802342e-01 -1.29598963e+00 -4.70806599e-01
5.45634329e-01 6.55064166e-01 -1.05996862e-01 9.56080973e-01
-3.69249247e-02 2.04760641e-01 -5.33075809e-01 -3.35580766e-01
3.41155201e-01 -4.56718475e-01 -8.91617894e-01 -2.18618065e-01
1.16249613e-01 2.14852300e-02 -1.02145708e+00 1.30282879e+00
3.68838936e-01 1.04592979e-01 6.31660461e-01 -1.50541019e+00
-9.80321467e-01 5.57912827e-01 -3.26739073e-01 2.29223326e-01
1.48611054e-01 -4.16685671e-01 1.06799471e+00 -6.23975992e-01
2.60227323e-01 1.01526320e+00 8.34623873e-01 4.25825506e-01
-1.07709789e+00 -6.25350654e-01 -9.90937129e-02 2.54325837e-01
-1.91766810e+00 3.08716983e-01 7.52633035e-01 -2.63840228e-01
7.82407701e-01 5.19906640e-01 5.65877855e-01 3.38991910e-01
8.80299658e-02 1.81788087e-01 6.49050474e-01 -6.62482977e-01
1.84152618e-01 2.91000865e-02 2.32908681e-01 1.08046710e+00
5.97896457e-01 8.51190910e-02 9.09188613e-02 -1.65728152e-01
8.07378769e-01 2.35121287e-02 -3.16863835e-01 -4.63226020e-01
-1.16544020e+00 1.18679976e+00 7.81200707e-01 6.87852204e-01
2.66318142e-01 4.84896004e-01 2.36984625e-01 5.50251365e-01
1.99131891e-01 6.76376343e-01 -2.95246780e-01 -1.62741333e-01
-3.84017766e-01 -6.81247562e-02 8.02512050e-01 1.32357299e+00
7.57286191e-01 -6.19724616e-02 6.04509711e-01 4.57016885e-01
1.39688522e-01 4.65313226e-01 3.37816000e-01 -6.26998127e-01
5.36313236e-01 8.41385841e-01 -1.59672409e-01 -1.43972015e+00
-5.68345487e-01 -1.96453005e-01 -9.49202836e-01 -8.98272358e-03
6.83914483e-01 -1.68346260e-02 -3.36242080e-01 1.69262254e+00
3.39681983e-01 -2.22067279e-03 1.37356788e-01 7.87663937e-01
5.19137919e-01 6.79184437e-01 -4.81184483e-01 -4.25657928e-02
1.37977755e+00 -5.82684100e-01 -4.29765403e-01 7.31416270e-02
1.42087936e+00 -4.29676294e-01 1.29535270e+00 1.85916081e-01
-6.38691247e-01 -1.92481175e-01 -1.52225518e+00 -6.82463944e-02
-6.60410404e-01 -2.31501326e-01 7.89275765e-01 1.10586083e+00
-1.04045486e+00 7.74828136e-01 -5.77272475e-01 -3.44242483e-01
3.88046727e-03 6.28840268e-01 -6.10497236e-01 -2.29754880e-01
-1.31395113e+00 6.55963480e-01 5.20349860e-01 2.01398432e-01
-1.63789690e-01 -4.07076299e-01 -8.92027974e-01 5.66130169e-02
2.84662873e-01 -1.16671525e-01 4.49236780e-01 -2.00945780e-01
-1.09409845e+00 5.39745152e-01 3.83482784e-01 -2.40446776e-01
9.40963998e-02 4.03155178e-01 -4.96754020e-01 2.18818203e-01
-3.90594989e-01 1.26162499e-01 3.80846053e-01 -6.97569549e-01
-2.08896905e-01 -7.59892821e-01 5.24251997e-01 1.98930260e-02
-7.94486046e-01 -1.64065391e-01 -5.23183718e-02 -3.56232703e-01
7.01970994e-01 -1.22923493e+00 -2.32725918e-01 8.00353214e-02
-2.05674954e-03 -4.40424830e-01 8.11758816e-01 -2.91662633e-01
1.54820764e+00 -2.51969647e+00 2.12378547e-01 3.36932689e-01
5.92200637e-01 1.64625093e-01 -1.16293006e-01 7.38497078e-01
-1.65397719e-01 5.38147628e-01 -1.48950741e-01 2.39515692e-01
2.20815033e-01 3.84979278e-01 -1.89230274e-02 8.79432142e-01
-1.99826360e-01 4.91933525e-01 -1.02737176e+00 -4.77436721e-01
-1.24313263e-02 4.72613245e-01 -5.42679608e-01 -1.60797730e-01
4.62168545e-01 -2.48503253e-01 -4.53102708e-01 -2.24920474e-02
6.74035549e-01 -3.43602985e-01 2.33204797e-01 -1.70640692e-01
2.12390527e-01 -5.89352101e-03 -1.56883132e+00 1.48487139e+00
-1.89327642e-01 7.20980287e-01 -2.61500955e-01 -1.24653220e+00
1.25037575e+00 2.12398425e-01 5.00882506e-01 -4.11786914e-01
1.83069427e-02 2.76947618e-01 1.25829384e-01 -2.60273457e-01
4.33196336e-01 -1.82081059e-01 -1.60710156e-01 6.57519460e-01
-2.46711224e-01 5.56693859e-02 9.21602398e-02 3.97076428e-01
1.54277658e+00 -4.56782818e-01 1.06588617e-01 -4.99965996e-01
5.54218113e-01 -3.79764289e-01 2.92138368e-01 4.37204063e-01
-2.11168423e-01 4.21875924e-01 9.22044218e-01 -4.92386997e-01
-1.22673655e+00 -1.02858102e+00 -3.14573944e-01 9.89035666e-01
5.02297223e-01 -8.65375996e-01 -8.40911865e-01 -8.91352952e-01
-2.80223414e-02 4.03492212e-01 -8.88251364e-01 -6.83891654e-01
-5.63206613e-01 -7.88359404e-01 7.78394103e-01 6.37223125e-01
3.30407292e-01 -5.80577552e-01 -2.02751532e-01 -1.77028969e-01
3.41973007e-01 -9.45146501e-01 -6.67824328e-01 1.23137310e-01
-1.00370276e+00 -1.35268199e+00 -3.48382980e-01 -9.16275680e-01
1.03597844e+00 4.73389894e-01 4.66946095e-01 4.72784817e-01
-2.90555656e-01 4.28668529e-01 -4.78109330e-01 2.47273780e-02
-3.56000870e-01 2.66801298e-01 1.94750458e-01 -3.43064606e-01
6.19006157e-01 -6.67210877e-01 -8.35182250e-01 5.44731617e-01
-1.14441836e+00 -2.96798080e-01 3.92265916e-01 7.36332059e-01
2.33125255e-01 4.47513402e-01 5.75620055e-01 -6.07033491e-01
5.10469556e-01 -3.89622629e-01 -7.32602298e-01 1.39659941e-01
-9.96908128e-01 5.17391384e-01 8.56869876e-01 -3.38877529e-01
-4.39487118e-03 -2.82229722e-01 3.15443933e-01 -2.42687419e-01
4.18000013e-01 4.69827384e-01 2.59126648e-02 -3.96128565e-01
9.36458588e-01 1.93479970e-01 3.21939103e-02 -4.17100400e-01
3.88562351e-01 7.81196535e-01 1.82451218e-01 -4.92093295e-01
8.91034782e-01 1.61669254e-01 5.86844921e-01 -8.36412728e-01
-3.70527834e-01 -3.65113556e-01 -8.45327795e-01 2.65590191e-01
5.51204801e-01 -4.09851313e-01 -1.01004922e+00 -2.24787384e-01
-7.68936753e-01 1.02230735e-01 -1.65930510e-01 7.56636322e-01
-5.18843651e-01 5.96637726e-01 -5.19126654e-01 -7.74825811e-01
-1.80337533e-01 -8.82412314e-01 5.74162185e-01 -1.30238935e-01
1.40711918e-01 -1.22102201e+00 1.34382278e-01 1.00616530e-01
1.39526784e-01 3.20244461e-01 1.28954089e+00 -7.84812391e-01
-4.57554758e-01 -7.22844481e-01 -4.35944200e-01 5.04825056e-01
1.34410590e-01 -9.35216248e-02 -3.90037745e-01 -6.33058131e-01
-4.64643203e-02 2.43264571e-01 2.51160651e-01 -3.34180892e-01
1.11693168e+00 -5.14725685e-01 -3.17606032e-01 5.33053339e-01
1.64381838e+00 7.55330846e-02 8.09316397e-01 1.01138480e-01
7.13136435e-01 3.14746737e-01 5.14108539e-01 2.30052352e-01
1.83965400e-01 4.86054897e-01 3.93428326e-01 4.69365083e-02
8.55720863e-02 -1.81924045e-01 2.24723697e-01 1.30372655e+00
-6.02687672e-02 3.64359282e-02 -8.76027822e-01 5.69474697e-01
-1.63522267e+00 -7.04757869e-01 -1.87117144e-01 2.61729503e+00
5.83203554e-01 1.42400384e-01 2.55054325e-01 6.79004490e-01
8.38183820e-01 -1.92557164e-02 -3.23845804e-01 -7.86921382e-01
1.59565851e-01 2.86762476e-01 7.50075698e-01 5.93793392e-01
-6.54326558e-01 6.96480870e-01 5.96902132e+00 6.86585724e-01
-8.04424822e-01 3.18999253e-02 -1.56422362e-01 3.37105058e-02
-2.62416482e-01 1.88108116e-01 -4.98310655e-01 4.56279516e-01
1.09276700e+00 -7.27504849e-01 6.71639502e-01 9.26180482e-01
-4.01334226e-01 3.18478018e-01 -1.33551693e+00 1.16362476e+00
9.58059952e-02 -1.00733769e+00 -1.34586141e-01 5.84418356e-01
3.27511370e-01 -1.75502151e-01 -1.19873933e-01 3.20229918e-01
7.05955848e-02 -1.14358425e+00 -9.17381980e-03 -2.86403182e-03
8.10385883e-01 -1.14604402e+00 7.59047210e-01 3.78176361e-01
-1.49958491e+00 -1.90091923e-01 -8.56449544e-01 2.92305332e-02
-4.50447410e-01 2.91012049e-01 -1.00043285e+00 6.53733253e-01
3.69981110e-01 3.40546638e-01 -6.74613655e-01 8.12923551e-01
1.03895918e-01 2.75587767e-01 -4.71712679e-01 -4.34185982e-01
4.60777164e-01 -5.37921309e-01 4.89731669e-01 1.06598747e+00
5.06695628e-01 4.88214403e-01 -6.23418279e-02 4.63252485e-01
-4.80607487e-02 4.09390360e-01 -9.24747348e-01 -3.49888414e-01
6.26654804e-01 9.43840802e-01 -7.81181633e-01 -5.02536912e-03
-3.61448467e-01 8.38879049e-01 4.27108169e-01 1.89788416e-01
-1.01571012e+00 -1.28598356e+00 5.22740722e-01 2.50982016e-01
3.01174343e-01 -6.60011232e-01 4.98400964e-02 -8.39344323e-01
1.95831388e-01 -3.77442300e-01 5.08511901e-01 -4.36259240e-01
-8.47589970e-01 7.59616554e-01 3.65863219e-02 -1.11095798e+00
-2.83498149e-02 -9.21146512e-01 -3.13800154e-03 6.38724923e-01
-7.41802335e-01 -7.66888440e-01 -1.73918620e-01 6.86049342e-01
-3.21384162e-01 -1.24907285e-01 1.16713572e+00 3.75880957e-01
-4.65139806e-01 1.01979315e+00 3.90908927e-01 3.85053575e-01
2.30727017e-01 -1.46821821e+00 2.83368438e-01 5.78938603e-01
3.44451368e-01 4.78797197e-01 6.20062113e-01 -2.23997459e-01
-1.81682003e+00 -7.87959397e-01 8.14169049e-01 -4.74574804e-01
8.91732037e-01 -5.64433277e-01 -1.00711656e+00 7.53928185e-01
-5.13109565e-01 3.92877907e-01 8.50246727e-01 2.15586245e-01
-7.35779047e-01 -3.29363883e-01 -1.40222025e+00 7.18458116e-01
1.36561036e+00 -7.12817192e-01 -4.32754099e-01 6.04769468e-01
1.00305092e+00 -1.55803770e-01 -1.50118625e+00 2.76015967e-01
4.18651700e-01 -6.09193683e-01 8.70578825e-01 -9.07813907e-01
-8.02079514e-02 -3.61690700e-01 -6.28307581e-01 -8.23495924e-01
-4.41733688e-01 -7.99146295e-01 -2.41576597e-01 6.84633493e-01
3.35699975e-01 -1.01065028e+00 7.06302285e-01 4.65036541e-01
6.73347786e-02 -1.10380781e+00 -1.24509680e+00 -1.22287333e+00
5.10609508e-01 -2.11101979e-01 8.39300394e-01 9.72607434e-01
3.98235261e-01 2.20242292e-01 1.66409984e-02 3.09262127e-01
5.94719470e-01 -1.72129408e-01 6.89549088e-01 -1.34685993e+00
-6.68444037e-02 -4.12263811e-01 -1.35830092e+00 -7.73758173e-01
4.21302058e-02 -1.13073003e+00 -4.83016133e-01 -1.33207083e+00
-1.89811513e-02 -8.63248110e-01 -4.67601836e-01 2.48689562e-01
1.41591772e-01 1.49029434e-01 1.45235034e-02 -9.67009515e-02
-3.62499297e-01 4.78204638e-01 1.24375570e+00 1.40553683e-01
-6.52211010e-02 -3.28552425e-01 -6.56860888e-01 4.87179548e-01
6.64978921e-01 -6.07324839e-01 -7.26598680e-01 -2.70604670e-01
5.68463802e-01 -1.00870438e-01 9.15742740e-02 -9.79264081e-01
2.90794164e-01 -3.40317097e-03 -2.37898052e-01 -1.72818452e-01
2.78153062e-01 -7.95942724e-01 7.83452913e-02 4.87530380e-01
-1.67113796e-01 2.92939961e-01 -1.32454574e-01 7.62603581e-01
1.67886838e-01 -4.60627347e-01 5.93570888e-01 3.28539908e-01
-2.08516091e-01 5.40130079e-01 2.32152775e-01 1.40665591e-01
1.24702203e+00 -5.14787018e-01 -4.71306950e-01 -2.26833850e-01
-6.96575403e-01 -1.28519580e-01 4.62941378e-01 2.28952631e-01
6.85429215e-01 -1.71426344e+00 -4.89539921e-01 1.63263246e-01
1.66766614e-01 -8.90248790e-02 4.11819145e-02 8.11901808e-01
-7.45155692e-01 4.93792921e-01 1.78918958e-01 -4.86852229e-01
-1.22567880e+00 1.08285809e+00 3.09813976e-01 -5.62805906e-02
-7.89728165e-01 5.76030493e-01 -9.29948222e-03 -4.05886739e-01
2.27247298e-01 -5.05349755e-01 2.89021581e-01 -3.91661614e-01
4.91889626e-01 7.08001435e-01 5.71050867e-02 -3.76355588e-01
-3.66377383e-01 6.91485763e-01 -8.57730657e-02 3.23778279e-02
1.07968926e+00 2.27781758e-02 -3.15779060e-01 5.08355319e-01
1.79183507e+00 -1.47228479e-01 -7.04353511e-01 -2.53395051e-01
-9.63298306e-02 -6.00795865e-01 -1.81684159e-02 -2.09862571e-02
-7.67334163e-01 8.95728707e-01 7.04869211e-01 8.41137946e-01
8.35889280e-01 1.71791315e-01 6.59079254e-01 7.36644030e-01
7.01279640e-01 -1.06513309e+00 8.06110725e-02 3.28075290e-01
9.85805631e-01 -7.80117333e-01 1.98013902e-01 -4.67041135e-01
-1.38281569e-01 1.16361165e+00 3.99424762e-01 -3.74907017e-01
1.01367807e+00 -2.08609402e-01 -4.96941924e-01 -2.67447829e-01
-4.26966548e-01 1.81960300e-01 3.06106091e-01 5.59632599e-01
3.97405088e-01 2.83078998e-01 -4.48776662e-01 5.36367416e-01
-6.82392895e-01 -4.97680694e-01 5.62440038e-01 7.15903461e-01
-6.90630317e-01 -1.02027392e+00 -1.36904746e-01 1.50795534e-01
-1.21295527e-01 9.59421247e-02 -3.16704392e-01 1.18109727e+00
-1.23534627e-01 8.13083470e-01 9.96806845e-02 -5.67588329e-01
3.50872368e-01 1.98268387e-02 8.58409882e-01 -4.45936859e-01
1.27530266e-02 -6.06551468e-01 -1.96264401e-01 -4.47930872e-01
1.34273186e-01 -1.88407853e-01 -1.62081492e+00 -8.30290258e-01
-4.95686680e-01 6.01247489e-01 7.82024682e-01 6.78995252e-01
3.75970572e-01 1.69836178e-01 9.58509386e-01 -1.49181381e-01
-7.98067391e-01 -7.85402060e-01 -1.02617919e+00 2.57678181e-01
5.07911853e-02 -6.34102404e-01 -8.70608449e-01 -5.66923857e-01] | [7.155984878540039, 6.0192084312438965] |
29db3411-098d-495b-94b3-b2f385eba97c | short-duration-speaker-verification-sdsv | 1912.06311 | null | https://arxiv.org/abs/1912.06311v3 | https://arxiv.org/pdf/1912.06311v3.pdf | Short-duration Speaker Verification (SdSV) Challenge 2021: the Challenge Evaluation Plan | This document describes the Short-duration Speaker Verification (SdSV) Challenge 2021. The main goal of the challenge is to evaluate new technologies for text-dependent (TD) and text-independent (TI) speaker verification (SV) in a short duration scenario. The proposed challenge evaluates SdSV with varying degree of phonetic overlap between the enrollment and test utterances (cross-lingual). It is the first challenge with a broad focus on systematic benchmark and analysis on varying degrees of phonetic variability on short-duration speaker recognition. We expect that modern methods (deep neural networks in particular) will play a key role. | ['Lukas Burget', 'Kong Aik Lee', 'Hossein Zeinali', 'Jahangir Alam'] | 2019-12-13 | null | null | null | null | ['text-independent-speaker-verification', 'text-dependent-speaker-verification'] | ['speech', 'speech'] | [-4.69229892e-02 -4.27862316e-01 -2.09576078e-02 -9.48840201e-01
-1.36486435e+00 -7.90556669e-01 8.81375849e-01 -2.66713321e-01
-1.59468845e-01 4.35882896e-01 3.49612862e-01 -6.56942248e-01
2.08998457e-01 2.63716072e-01 -2.62358099e-01 -7.60395288e-01
-4.98075970e-02 5.71582139e-01 -2.46127903e-01 -2.66306162e-01
-2.85026748e-02 6.93650007e-01 -1.65336990e+00 1.98741719e-01
4.66317624e-01 9.01715636e-01 -9.83375963e-03 8.64208460e-01
8.63410011e-02 1.54697537e-01 -9.61334586e-01 -2.08282799e-01
2.59175543e-02 -1.54646307e-01 -8.55769277e-01 -2.48276025e-01
8.32620502e-01 -1.81251124e-01 -2.23513067e-01 7.86480963e-01
1.22476268e+00 2.91401476e-01 7.36018181e-01 -1.27001250e+00
-4.64838803e-01 9.81543303e-01 9.53631252e-02 6.07137084e-01
3.03199023e-01 8.68615285e-02 6.20183468e-01 -1.01291454e+00
3.34549665e-01 1.35419869e+00 9.12027597e-01 7.36620963e-01
-9.34683979e-01 -7.83471048e-01 2.09765226e-01 5.60170710e-01
-1.56056309e+00 -1.37538385e+00 7.16781735e-01 -4.81526315e-01
1.13740313e+00 4.96041626e-01 -8.16182718e-02 1.65857887e+00
-6.06948920e-02 7.87457943e-01 1.10638833e+00 -5.69063604e-01
2.12133154e-01 3.57530296e-01 6.33410990e-01 -4.60015796e-03
-3.43592107e-01 4.71800119e-01 -8.16139042e-01 7.31549487e-02
-2.72555530e-01 -8.46601486e-01 -2.03620732e-01 1.69505328e-01
-9.07875359e-01 7.49108851e-01 -5.13756990e-01 6.01185024e-01
-5.74543774e-02 -1.45531803e-01 9.83712435e-01 6.08212769e-01
3.52757394e-01 -2.26185054e-01 -6.86956286e-01 -4.92822587e-01
-1.32182360e+00 6.27524257e-02 7.83285022e-01 1.05224288e+00
-2.66471319e-02 9.13983107e-01 -5.57953775e-01 1.07138729e+00
5.03219843e-01 8.32119763e-01 9.28495705e-01 -3.46569657e-01
4.05096710e-01 -4.64459062e-01 1.71486381e-03 -2.00187415e-01
-2.05202457e-02 -5.03975093e-01 -7.30813742e-01 1.13852985e-01
3.16447765e-01 -2.60841638e-01 -1.08279777e+00 1.71394145e+00
1.25232607e-01 1.50033236e-01 3.63383353e-01 3.44040900e-01
1.44693387e+00 5.29657602e-01 4.61061625e-03 -2.10316852e-01
1.40370381e+00 -7.41593242e-01 -1.01717567e+00 -1.36933565e-01
4.03921694e-01 -1.16359520e+00 9.62340474e-01 3.57268751e-01
-8.66334796e-01 -8.64349306e-01 -1.16242933e+00 7.87905157e-02
-5.24471819e-01 2.27228180e-01 -2.97829688e-01 1.52601576e+00
-1.46604562e+00 1.67064324e-01 -4.41076636e-01 -3.12026381e-01
-2.32899189e-02 3.55499476e-01 -3.97089094e-01 1.72659144e-01
-1.51262164e+00 1.06159258e+00 -7.72636905e-02 2.74673074e-01
-1.45296955e+00 -7.56950796e-01 -1.00255406e+00 -1.75588250e-01
-1.77962556e-01 1.70110926e-01 1.60503590e+00 -4.42616016e-01
-1.92426777e+00 1.06182265e+00 -7.37456858e-01 -6.58108354e-01
4.73901153e-01 2.36242749e-02 -1.10974395e+00 -3.46630573e-01
-3.26094389e-01 3.44150364e-01 1.02838409e+00 -1.03619409e+00
-3.38934213e-01 -4.90810513e-01 -7.66864419e-01 4.16028276e-02
-8.82813185e-02 8.10535967e-01 -7.26953615e-03 -4.39010501e-01
-1.84075668e-01 -8.76507640e-01 4.82423306e-01 -7.27678716e-01
-7.82310545e-01 -5.70759296e-01 1.29488349e+00 -1.20902550e+00
8.68250966e-01 -2.37940788e+00 -1.84309140e-01 -2.04964146e-01
-3.84492397e-01 6.57013655e-01 -1.59518898e-01 4.25533623e-01
-9.80435759e-02 2.89594643e-02 1.91410631e-02 -1.31489432e+00
5.16429543e-01 -4.70131673e-02 -4.56008852e-01 5.80242693e-01
-2.49476343e-01 9.10530925e-01 -8.55141878e-02 -4.14364964e-01
3.43264580e-01 6.18549705e-01 2.56137282e-01 1.89960808e-01
8.23411122e-02 5.92555285e-01 2.07907319e-01 7.87904561e-01
1.08612633e+00 6.47882104e-01 9.78193991e-03 -6.20301673e-03
-4.13563639e-01 7.63981700e-01 -1.10893583e+00 1.42124927e+00
-5.92896223e-01 1.21553373e+00 4.78915393e-01 -9.29248035e-01
1.08625245e+00 1.02688074e+00 1.03324139e-02 -4.23726648e-01
3.30265462e-01 3.18874866e-01 1.60172582e-02 -3.79454821e-01
4.64566946e-01 -2.24430650e-01 -1.98913544e-01 2.84729064e-01
2.95895875e-01 -1.76206324e-02 -2.74285197e-01 -2.26265445e-01
5.00577331e-01 -4.31948811e-01 1.36905998e-01 -3.81705314e-01
8.55319321e-01 -4.84591275e-01 3.98311079e-01 6.63034797e-01
-1.17186940e+00 5.93813062e-01 -5.95004670e-02 1.15581304e-01
-9.07845855e-01 -1.05036831e+00 -4.52856511e-01 1.13711071e+00
-4.33638841e-01 1.45111620e-01 -8.56924415e-01 -5.11836231e-01
-2.81246863e-02 9.47978497e-01 -5.27441919e-01 1.19790882e-01
-5.89443266e-01 -2.57235080e-01 1.42154491e+00 5.05621672e-01
4.17789370e-01 -1.05091715e+00 2.46717587e-01 2.03145921e-01
-1.09577402e-01 -1.27370286e+00 -8.21453273e-01 4.97332126e-01
-4.45755899e-01 -3.14094722e-01 -8.55401278e-01 -1.26572847e+00
-2.51921654e-01 -4.71944585e-02 7.24266827e-01 -6.93887889e-01
1.46652952e-01 3.27810556e-01 -1.74442932e-01 -4.65572864e-01
-8.81575048e-01 1.69528022e-01 5.07352412e-01 6.11950969e-03
7.42903650e-01 -2.84195691e-01 -1.44297034e-01 6.20361745e-01
-2.56721407e-01 -8.78686726e-01 2.30715908e-02 7.32944191e-01
2.58017749e-01 -6.50320575e-02 9.98832762e-01 -2.26548612e-01
7.10318804e-01 -1.81870818e-01 -7.46058464e-01 4.47109729e-01
-4.27178025e-01 -1.68211877e-01 4.10984129e-01 -6.41103983e-01
-1.27000284e+00 1.16520999e-02 -8.64269197e-01 -4.22777265e-01
-7.85637677e-01 3.88395935e-01 -5.72818637e-01 -5.15055433e-02
4.91820246e-01 8.76743495e-01 -2.47974247e-02 -9.10599709e-01
1.74533010e-01 1.43266451e+00 7.13009238e-01 -2.14770138e-01
6.22018516e-01 8.37305635e-02 -6.86017334e-01 -1.16710258e+00
-1.72962278e-01 -6.47485018e-01 -8.49612355e-01 -1.63316309e-01
6.62875772e-01 -1.08830667e+00 -8.73770893e-01 1.02324665e+00
-1.18412888e+00 -4.44195181e-01 -2.40121894e-02 5.85840762e-01
-1.69226706e-01 4.33221459e-01 -6.53825939e-01 -1.05861223e+00
-7.09056675e-01 -1.42781425e+00 1.16466689e+00 -1.34966552e-01
-1.79489106e-01 -9.26079154e-01 3.23571593e-01 5.49487531e-01
8.98874104e-01 -4.40291256e-01 4.76831704e-01 -1.26821530e+00
-3.67557444e-02 -3.41484874e-01 2.65624881e-01 9.94835198e-01
1.82366878e-01 2.34793261e-01 -1.75848854e+00 -4.97921109e-01
2.39231318e-01 -1.57637119e-01 4.85282481e-01 8.87120128e-01
6.50538564e-01 -3.75180617e-02 -1.09161027e-01 3.60198617e-01
1.00174975e+00 5.33112526e-01 4.87650871e-01 -9.59680006e-02
3.64036739e-01 6.60453260e-01 4.15387660e-01 1.53395668e-01
3.49718511e-01 1.20780540e+00 -1.81544274e-01 2.49698609e-01
-5.71156502e-01 1.43017247e-01 8.99739623e-01 1.28388679e+00
6.63042367e-01 -5.17617047e-01 -1.11852813e+00 6.42902851e-01
-1.00879872e+00 -1.08970022e+00 -3.83542180e-01 2.31188726e+00
6.95771575e-01 -1.15375914e-01 3.49955201e-01 5.49540281e-01
1.11763203e+00 2.89332509e-01 -4.58850324e-01 -1.04676723e+00
-4.01409149e-01 1.25702053e-01 1.11036286e-01 8.05940807e-01
-1.26435137e+00 8.03337753e-01 7.46036530e+00 1.05584908e+00
-1.50787890e+00 5.04092872e-01 5.18128037e-01 7.62357488e-02
1.54567033e-01 -5.73312998e-01 -1.62283790e+00 5.08627117e-01
1.85654056e+00 -2.66631037e-01 -9.95984599e-02 9.23339546e-01
4.60868597e-01 4.86301452e-01 -1.16760898e+00 9.69150186e-01
3.69873494e-01 -9.94042993e-01 -4.67315286e-01 1.39673101e-02
6.51033461e-01 5.12964487e-01 5.99227071e-01 6.39618039e-01
3.00522577e-02 -1.06952798e+00 1.03470278e+00 -4.66303192e-02
1.28188944e+00 -5.96251011e-01 9.39489961e-01 1.03360936e-02
-1.44721091e+00 1.12268843e-01 -1.15236133e-01 5.30208409e-01
4.60376590e-01 3.28278393e-01 -1.14355290e+00 4.08130169e-01
4.31479990e-01 4.36013788e-01 -2.70830631e-01 9.55079913e-01
1.31498486e-01 9.35994267e-01 -1.01402022e-01 4.67127748e-02
5.89354895e-03 3.43464166e-01 7.36940324e-01 1.60892630e+00
3.04811686e-01 -5.21058619e-01 -2.31361449e-01 3.14341486e-01
-7.76354074e-02 7.98528716e-02 -5.24968505e-01 8.45488012e-02
8.57426167e-01 6.47880375e-01 -1.24663450e-01 -5.46997428e-01
-1.84170250e-02 8.68154645e-01 -1.75961778e-01 3.80957335e-01
-6.93971753e-01 -1.94940761e-01 9.83388186e-01 -4.27992016e-01
6.23193741e-01 -3.68824601e-01 -4.41615283e-01 -7.08629966e-01
-3.61152105e-02 -1.10733712e+00 3.02391291e-01 -2.16058165e-01
-1.20900118e+00 1.00525963e+00 -1.62259340e-01 -1.13474584e+00
-5.11524141e-01 -3.27709734e-01 -6.18439794e-01 1.40424240e+00
-1.69687867e+00 -1.16791368e+00 -2.97919419e-02 5.83691239e-01
1.00210118e+00 -6.77908838e-01 1.09008765e+00 7.81804800e-01
-5.85872710e-01 1.40551734e+00 6.80585325e-01 1.13727972e-02
9.29958463e-01 -1.04707158e+00 9.90890801e-01 8.27611387e-01
2.12859157e-02 4.00960624e-01 8.47158134e-01 -3.41599703e-01
-1.29966068e+00 -1.08257306e+00 1.35168481e+00 -6.42243803e-01
8.29028860e-02 -8.16559434e-01 -7.79461503e-01 7.39373386e-01
3.12116921e-01 -2.81431407e-01 7.30375350e-01 1.78048640e-01
-5.36137462e-01 -1.83642030e-01 -1.22006214e+00 6.26199916e-02
5.88198543e-01 -1.08968961e+00 -6.06499076e-01 1.81404904e-01
8.47618520e-01 -2.88632631e-01 -7.01815069e-01 3.96006703e-01
5.95794022e-01 -7.20606327e-01 9.75806415e-01 -3.27586174e-01
-4.69629586e-01 -9.34207663e-02 -5.36995471e-01 -1.12549520e+00
6.58934116e-02 -6.48957670e-01 3.42263319e-02 1.76817429e+00
5.40900886e-01 -6.94668710e-01 8.22317898e-01 2.08744720e-01
-6.33368731e-01 -9.29479972e-02 -1.62434995e+00 -1.62111318e+00
2.82340050e-01 -8.35166872e-01 7.77701914e-01 9.29340959e-01
-2.42120281e-01 2.06248257e-02 -5.68470120e-01 4.74089921e-01
4.16777283e-01 -2.38976806e-01 5.71073532e-01 -1.01056063e+00
1.25923142e-01 -5.42517304e-01 -4.76335675e-01 -8.52797389e-01
3.41767818e-01 -7.35249877e-01 4.99758422e-01 -9.84087706e-01
-3.14192921e-01 -1.49214447e-01 -3.35028261e-01 1.99032739e-01
1.32697091e-01 5.30273542e-02 -1.18502758e-01 -1.73463196e-01
-1.29159540e-01 8.70718896e-01 4.51189339e-01 -4.25434262e-01
-2.23390564e-01 4.52571929e-01 -1.25183314e-01 3.56975012e-02
9.22887623e-01 -4.10507470e-01 -3.34676832e-01 -1.57252073e-01
-1.04495049e+00 6.42731413e-02 1.60628892e-02 -1.02495205e+00
1.93164706e-01 2.33598173e-01 -2.79366195e-01 -1.06101167e+00
6.74011886e-01 -2.69232064e-01 2.35378325e-01 4.66124594e-01
-5.44455826e-01 -1.03975862e-01 5.86735547e-01 3.37066241e-02
-5.79567194e-01 -2.23418549e-01 9.68957543e-01 4.96208340e-01
-5.75740099e-01 1.27985492e-01 -7.70343363e-01 -9.88205373e-02
6.37347460e-01 -1.74486354e-01 -2.52903968e-01 -4.04766083e-01
-9.02649403e-01 -3.11661988e-01 4.96732555e-02 6.56220853e-01
4.72557217e-01 -1.23236763e+00 -1.20254612e+00 2.24270925e-01
2.07449555e-01 -6.99054062e-01 6.37952387e-01 7.18491137e-01
-3.80831920e-02 1.15053988e+00 1.20944402e-03 -7.09924221e-01
-2.05715656e+00 2.76847929e-01 7.20810771e-01 1.05936281e-01
-2.93049216e-01 1.42960787e+00 -1.70696065e-01 -7.94048011e-01
7.98112750e-01 -1.46724164e-01 -1.65066525e-01 1.02649339e-01
7.18916118e-01 3.00529093e-01 6.82564318e-01 -1.32520413e+00
-7.66306281e-01 1.66247547e-01 -3.14051807e-01 -5.63642681e-01
1.00034964e+00 -3.23980808e-01 4.51161683e-01 6.76790416e-01
1.53712916e+00 1.38001487e-01 -7.78791070e-01 -3.56510609e-01
2.01716125e-02 -8.73097405e-02 2.80981749e-01 -9.46497321e-01
-6.62424922e-01 1.17220342e+00 1.15307593e+00 5.09539321e-02
6.41880214e-01 -1.00290939e-01 9.17547107e-01 -6.22337917e-03
1.34251500e-02 -9.76217389e-01 -4.26129162e-01 6.51947021e-01
1.13137257e+00 -1.22079706e+00 -5.79934478e-01 -1.26020297e-01
-7.20568478e-01 7.68475592e-01 1.33574277e-01 6.53165281e-01
8.80968869e-01 2.85342574e-01 6.26192391e-01 8.88287369e-03
-6.52801335e-01 5.39434813e-02 5.63152015e-01 9.76875007e-01
5.80622792e-01 2.69372910e-01 9.77696106e-02 4.54413503e-01
-5.17941952e-01 -6.09358251e-01 2.44479701e-01 7.36126661e-01
-1.52226195e-01 -1.49474859e+00 -6.42652810e-01 1.37606874e-01
-3.04029584e-01 -3.06630135e-01 -3.84165108e-01 3.86967927e-01
-8.06618705e-02 1.49307644e+00 -3.28912407e-01 -5.47325253e-01
1.96636543e-01 7.54046619e-01 1.13178149e-01 -3.99012327e-01
-7.91420341e-01 -4.18307260e-02 4.97079790e-01 -1.47145316e-01
-1.35769276e-03 -1.17358541e+00 -6.41959608e-01 -4.92446065e-01
-3.17824572e-01 1.44542024e-01 1.68971145e+00 9.06181753e-01
4.40594494e-01 3.12129229e-01 8.27944696e-01 -6.31583095e-01
-1.00108361e+00 -1.39205241e+00 -7.91792870e-01 2.84791216e-02
8.28692257e-01 -4.65937674e-01 -8.08939219e-01 7.57026374e-02] | [14.336138725280762, 6.102877616882324] |
f0eed058-5337-4b17-af0d-a21727fedb8d | computational-technologies-for-fashion | 2306.03395 | null | https://arxiv.org/abs/2306.03395v1 | https://arxiv.org/pdf/2306.03395v1.pdf | Computational Technologies for Fashion Recommendation: A Survey | Fashion recommendation is a key research field in computational fashion research and has attracted considerable interest in the computer vision, multimedia, and information retrieval communities in recent years. Due to the great demand for applications, various fashion recommendation tasks, such as personalized fashion product recommendation, complementary (mix-and-match) recommendation, and outfit recommendation, have been posed and explored in the literature. The continuing research attention and advances impel us to look back and in-depth into the field for a better understanding. In this paper, we comprehensively review recent research efforts on fashion recommendation from a technological perspective. We first introduce fashion recommendation at a macro level and analyse its characteristics and differences with general recommendation tasks. We then clearly categorize different fashion recommendation efforts into several sub-tasks and focus on each sub-task in terms of its problem formulation, research focus, state-of-the-art methods, and limitations. We also summarize the datasets proposed in the literature for use in fashion recommendation studies to give readers a brief illustration. Finally, we discuss several promising directions for future research in this field. Overall, this survey systematically reviews the development of fashion recommendation research. It also discusses the current limitations and gaps between academic research and the real needs of the fashion industry. In the process, we offer a deep insight into how the fashion industry could benefit from fashion recommendation technologies. the computational technologies of fashion recommendation. | ['Tat-Seng Chua', 'P. Y. Mok', 'Zhihui Lai', 'Yujuan Ding'] | 2023-06-06 | null | null | null | null | ['product-recommendation', 'information-retrieval'] | ['miscellaneous', 'natural-language-processing'] | [ 2.07789615e-01 -5.41827023e-01 -8.24192286e-01 -5.82302392e-01
-2.11330578e-01 -7.06484139e-01 1.13935307e-01 1.07880428e-01
9.90106761e-02 8.36192816e-02 6.24630332e-01 -2.14421317e-01
-4.20105517e-01 -7.19110310e-01 -2.11779460e-01 -5.28602719e-01
2.57610440e-01 -4.55645546e-02 -3.16089898e-01 -3.74585271e-01
2.55455941e-01 1.72524303e-01 -1.67891479e+00 5.00942111e-01
3.31639349e-01 1.00450110e+00 2.92542756e-01 4.74545389e-01
1.64215669e-01 3.34875971e-01 -7.71290902e-03 -9.44242656e-01
2.75681585e-01 -3.94931406e-01 -4.15864497e-01 5.64968467e-01
5.49743950e-01 -3.28369319e-01 -6.63963318e-01 7.74831653e-01
5.60903192e-01 6.39095306e-01 4.48093891e-01 -8.73686433e-01
-1.80755103e+00 7.18224049e-01 -2.94023514e-01 1.33517534e-01
4.31822598e-01 2.37234849e-02 1.16886032e+00 -6.62115395e-01
3.63835096e-01 9.00689900e-01 7.07060099e-01 3.31753492e-01
-9.58935738e-01 -3.32902580e-01 9.34610426e-01 1.36392042e-01
-9.71413076e-01 -3.65023106e-01 1.20017529e+00 -3.11590463e-01
4.66111422e-01 7.59094775e-01 9.96247888e-01 1.06815028e+00
2.12320104e-01 1.11459315e+00 1.08765149e+00 -3.55324984e-01
1.66110024e-02 1.35111734e-01 6.49884582e-01 3.52295369e-01
1.38320312e-01 6.97425455e-02 -2.64657587e-01 2.14367151e-01
1.10302305e+00 2.72262663e-01 4.43997055e-01 -7.64799565e-02
-9.59532619e-01 1.08373201e+00 2.63394654e-01 2.98356414e-01
-5.24887681e-01 5.06561138e-02 5.54302573e-01 8.09348106e-01
9.22837615e-01 4.24523979e-01 -4.28918540e-01 2.99245089e-01
-7.44652331e-01 6.92167699e-01 6.44112647e-01 1.08052123e+00
-8.21635723e-02 2.83919841e-01 -2.21138656e-01 1.18426752e+00
7.62550056e-01 2.75637090e-01 7.77601004e-02 -6.81118548e-01
-1.14724703e-01 -3.24026942e-02 6.51042089e-02 -1.55302334e+00
-2.70007193e-01 -7.13162005e-01 -6.93715274e-01 -3.64623785e-01
2.18977049e-01 1.00274965e-01 -5.51663637e-01 1.08007956e+00
3.10722321e-01 -2.85695624e-02 -3.32998872e-01 1.42581224e+00
1.70490468e+00 4.72686917e-01 4.03676242e-01 -3.22000772e-01
1.85070229e+00 -1.41242158e+00 -8.65833700e-01 -2.23214805e-01
6.49838373e-02 -1.58636236e+00 9.78495240e-01 3.17011029e-01
-1.01942861e+00 -1.16852844e+00 -7.81277120e-01 -1.46954060e-01
-3.52669030e-01 4.76785392e-01 1.34127426e+00 9.22170460e-01
-7.14217126e-01 4.77044612e-01 -3.98553967e-01 -8.95319283e-01
1.31494462e-01 3.60154241e-01 2.22103387e-01 -5.32702982e-01
-1.02304947e+00 6.00755870e-01 -4.10912693e-01 4.22914386e-01
-3.14225048e-01 -7.78504968e-01 -7.49093831e-01 -6.12231255e-01
3.28768492e-01 -1.19600511e+00 1.20929384e+00 -8.60529006e-01
-1.65926850e+00 8.96316111e-01 1.74663156e-01 1.07963510e-01
4.76349369e-02 -4.12484974e-01 -9.98598814e-01 -4.35802937e-01
-2.12762669e-01 3.04061860e-01 7.40518928e-01 -1.21213651e+00
-6.98305786e-01 -3.34361255e-01 4.82160300e-01 1.16916344e-01
-1.35810554e-01 7.86934555e-01 -6.64333224e-01 -1.49932170e+00
-1.02534048e-01 -1.14922154e+00 -6.59958839e-01 -3.24347526e-01
-3.01081181e-01 -4.26393688e-01 3.30477089e-01 -6.85845137e-01
1.53410959e+00 -2.02681637e+00 1.12818763e-01 -5.11217937e-02
1.70889035e-01 -4.57015522e-02 -2.93391258e-01 6.62369072e-01
1.74848273e-01 -2.82976151e-01 6.43475890e-01 -2.58766562e-01
1.69300377e-01 1.29643872e-01 -3.28649580e-01 6.50682509e-01
-5.60697377e-01 1.20626783e+00 -1.00472260e+00 1.21976659e-01
6.42357111e-01 6.00339949e-01 -5.21021545e-01 -7.79322872e-04
-2.58618206e-01 5.96332967e-01 -1.01783943e+00 9.64095771e-01
5.47042966e-01 -9.23706368e-02 4.36085552e-01 -9.52665925e-01
-3.87953907e-01 9.82847065e-02 -1.09897363e+00 1.57643366e+00
-3.65500331e-01 2.61459470e-01 -1.06900282e-01 -8.92973304e-01
1.04985666e+00 1.17885426e-01 5.53728282e-01 -6.46567404e-01
3.21836114e-01 7.53692235e-04 -2.55369991e-01 -6.16293907e-01
1.18687999e+00 -9.86026376e-02 -3.80165905e-01 3.75679284e-01
-2.49927744e-01 3.10535818e-01 1.04094915e-01 -2.01536328e-01
4.23422664e-01 3.12863529e-01 1.93199381e-01 -1.83300644e-01
3.23946327e-01 6.01680614e-02 6.48441985e-02 6.07503474e-01
1.22421227e-01 2.13843733e-01 -4.67367887e-01 -8.25053096e-01
-8.52104247e-01 -9.76865828e-01 1.15949519e-01 1.81214035e+00
4.21780407e-01 -6.95687950e-01 -5.21530621e-02 -7.14188755e-01
1.58335075e-01 5.27340293e-01 -5.25100708e-01 -4.65620272e-02
-7.84729838e-01 -8.40724289e-01 -1.66035518e-01 7.26995111e-01
1.73791975e-01 -9.90643859e-01 1.03656575e-01 2.83010960e-01
6.87325373e-02 -8.51740479e-01 -8.49102616e-01 -4.51144695e-01
-1.11687338e+00 -7.16276884e-01 -7.45024085e-01 -1.17665291e+00
6.69601917e-01 1.06553471e+00 1.36022902e+00 7.18054771e-02
1.54315567e-04 1.02063477e+00 -7.95775294e-01 -2.23381110e-02
-5.25455587e-02 -1.00764364e-01 1.84728205e-01 -7.85436258e-02
6.70817614e-01 -2.95438111e-01 -1.02655125e+00 7.34881461e-01
-5.82938313e-01 -9.04441066e-03 3.90820861e-01 4.04142857e-01
9.69049275e-01 2.87212551e-01 8.56861115e-01 -8.50292325e-01
9.64698911e-01 -7.18927443e-01 -6.46457970e-02 7.55246431e-02
-6.20114148e-01 -7.29824483e-01 4.62710947e-01 -6.14459217e-01
-9.84227002e-01 -1.14787400e-01 -4.44857448e-01 1.02064814e-02
-2.90287763e-01 6.63726568e-01 -2.51134541e-02 -1.96835607e-01
-1.63515192e-02 3.04957796e-02 -2.36406878e-01 -1.16357958e+00
9.97055948e-01 4.95902807e-01 1.62793010e-01 -7.27395713e-01
4.34977561e-01 2.33292773e-01 -3.32453609e-01 -7.83095479e-01
-1.13232517e+00 -5.10679007e-01 -3.45677316e-01 -7.79195070e-01
6.88468158e-01 -7.21571445e-01 -7.05551267e-01 2.92288631e-01
-5.20112574e-01 -1.15642892e-02 -5.33273220e-01 6.93779886e-01
-4.14258748e-01 3.11426103e-01 -1.14469302e+00 -6.25723779e-01
-4.50583071e-01 -1.25700164e+00 7.47848570e-01 3.28443766e-01
-4.98993695e-01 -1.17009449e+00 -1.42980218e-01 9.44885910e-01
6.15026772e-01 -6.47070780e-02 7.28685081e-01 -4.04831231e-01
-3.49276781e-01 -4.19941485e-01 -5.65918647e-02 1.18104607e-01
3.36657047e-01 -3.00355285e-01 -5.81645012e-01 -2.01836184e-01
-1.64061517e-01 2.26522550e-01 8.27428997e-01 9.84412909e-01
9.63593841e-01 -6.46923333e-02 -2.93935716e-01 1.95546746e-01
1.21045327e+00 3.48575324e-01 3.02685738e-01 3.41904342e-01
7.40668058e-01 8.43941450e-01 1.11381078e+00 1.78933501e-01
8.39249969e-01 9.68085706e-01 2.69476008e-02 -1.21184483e-01
-6.12494528e-01 -3.74424130e-01 1.48656189e-01 1.46210480e+00
-7.69491076e-01 -3.22484702e-01 3.59232813e-01 3.12122822e-01
-2.12026882e+00 -7.69878626e-01 -3.92388493e-01 2.01095128e+00
2.84625769e-01 -3.93833041e-01 7.65949249e-01 -6.40904531e-02
6.23120308e-01 2.22763032e-01 -4.42304999e-01 -7.10467696e-01
6.33655563e-02 -8.54508132e-02 4.86128718e-01 -8.73606279e-02
-1.47319007e+00 9.78816211e-01 7.12433767e+00 7.71168888e-01
-7.77954340e-01 -5.43300919e-02 7.54284024e-01 8.41045901e-02
-3.64515454e-01 -4.65440094e-01 -7.71380424e-01 4.29625690e-01
4.68029022e-01 1.59542114e-01 6.19774401e-01 8.65305901e-01
4.02735472e-01 3.83652985e-01 -1.08983934e+00 9.68428910e-01
3.28673780e-01 -1.34263325e+00 -1.22502327e-01 1.02004237e-01
7.78269410e-01 -4.65229481e-01 7.52869427e-01 2.93068111e-01
3.07371110e-01 -6.63166344e-01 7.92313099e-01 4.80970234e-01
3.85668248e-01 -5.81115186e-01 6.59651101e-01 -3.72376025e-01
-1.61976302e+00 -4.73929830e-02 -5.83872855e-01 -5.14086902e-01
6.41139567e-01 6.50377810e-01 3.55565846e-01 9.42510784e-01
5.57553530e-01 1.24426246e+00 -5.96801750e-02 1.15068996e+00
1.33364275e-01 6.02514207e-01 9.72158164e-02 -7.19380006e-02
8.74174386e-02 -7.53529906e-01 2.69567341e-01 1.28972757e+00
3.22150409e-01 3.33647013e-01 6.96871281e-01 8.34163204e-02
7.67051578e-02 4.89283085e-01 -2.36545995e-01 -1.16213486e-01
-2.65258066e-02 1.18985641e+00 -1.09422994e+00 -4.21688147e-02
-9.06373382e-01 7.78240979e-01 -3.54683340e-01 2.78587282e-01
-7.88216054e-01 2.42661148e-01 1.03763115e+00 5.90818077e-02
4.96783108e-01 -3.58766586e-01 -4.25688714e-01 -8.91432047e-01
-4.61496353e-01 -9.44025218e-01 6.00585759e-01 -3.35458964e-01
-2.14230633e+00 3.68483365e-01 1.27521947e-01 -1.35316992e+00
5.00470400e-01 -5.67795992e-01 -2.80922174e-01 3.41198951e-01
-1.27548778e+00 -1.82467794e+00 4.25951555e-02 2.11683527e-01
1.21214879e+00 -1.69745386e-01 7.50618279e-01 8.48731101e-01
-6.50876820e-01 6.40115738e-01 2.14831635e-01 -2.99102813e-01
7.35269904e-01 -5.15808642e-01 4.93415952e-01 2.11136878e-01
-5.34015857e-02 9.14970934e-01 1.03588259e+00 -7.40907967e-01
-2.26287341e+00 -1.14922619e+00 8.61691117e-01 -2.92282581e-01
6.87870324e-01 2.28777137e-02 -7.78665319e-02 7.66086876e-01
2.56846845e-01 -4.21623081e-01 1.31740308e+00 5.76204360e-01
-1.86256438e-01 -2.70206124e-01 -1.27757001e+00 8.12141299e-01
1.30570245e+00 1.28229242e-02 -3.86752218e-01 4.06807423e-01
5.91698229e-01 -1.38886884e-01 -1.38332653e+00 -3.09905056e-02
1.35180342e+00 -3.60846639e-01 1.57531214e+00 -8.66260767e-01
3.86943489e-01 -2.56983995e-01 -5.60694098e-01 -1.14708054e+00
-1.23793006e+00 -3.47828060e-01 -1.86021462e-01 1.46298766e+00
-1.55745268e-01 -1.12931564e-01 6.71941936e-01 7.13010609e-01
-3.47855717e-01 -9.57208037e-01 -1.27750263e-01 -5.31558454e-01
-2.51345605e-01 -8.20408642e-01 6.84417009e-01 9.22590077e-01
-1.16858520e-01 6.60574734e-01 -1.24972379e+00 -2.43063331e-01
7.62955606e-01 9.02187407e-01 6.33049607e-01 -1.03215480e+00
-3.39229375e-01 -4.87891614e-01 -2.15227902e-01 -1.79544175e+00
-2.48298585e-01 -1.03902590e+00 -4.57353592e-01 -1.84630013e+00
1.57168061e-01 -4.98148561e-01 -5.49146950e-01 -9.64773744e-02
6.67671263e-02 9.81283903e-01 3.59368950e-01 1.80333597e-03
-7.45207310e-01 5.76671176e-02 1.68429923e+00 -2.85255432e-01
-4.04168695e-01 4.73548293e-01 -1.61033738e+00 7.98480332e-01
9.15918648e-01 -2.97128614e-02 -4.31997448e-01 -5.01806855e-01
4.36037302e-01 -2.59075284e-01 4.08552848e-02 -3.02228220e-02
2.51747877e-03 -4.85462725e-01 4.68039751e-01 -6.17278039e-01
3.73459727e-01 -1.04545116e+00 6.13149822e-01 -2.26637758e-02
-4.63607460e-01 1.64043486e-01 -4.49007861e-02 6.62976623e-01
4.39190239e-01 -3.84693444e-02 3.01625013e-01 -1.40904188e-01
-1.21902752e+00 3.54150206e-01 -7.99313366e-01 -8.77521217e-01
9.61468935e-01 -4.02296662e-01 -4.98044752e-02 -2.57790864e-01
-1.08936155e+00 7.97244608e-02 9.86670554e-02 1.00081897e+00
7.40940511e-01 -1.71608508e+00 -6.26234174e-01 2.30473150e-02
3.08704644e-01 -1.18531609e+00 8.58930290e-01 7.51052678e-01
3.92231047e-02 4.01266754e-01 -2.03226432e-02 1.18670188e-01
-1.49105048e+00 1.13638043e+00 -4.48515028e-01 -1.53865680e-01
-7.00906038e-01 6.17980301e-01 1.74400955e-01 -3.66554797e-01
1.57011852e-01 -2.94313610e-01 -7.66474068e-01 -4.68381308e-02
6.26186252e-01 6.87763929e-01 -2.84485340e-01 -8.05297196e-01
-2.09787309e-01 9.94081438e-01 -1.16359644e-01 8.75709832e-01
1.44152141e+00 -7.00699270e-01 2.21692827e-02 3.46675336e-01
6.88831449e-01 1.53139174e-01 -5.44086277e-01 -2.92680413e-01
-3.19943368e-01 -8.55566680e-01 3.47622186e-01 -9.50381041e-01
-1.47856116e+00 -8.75038933e-03 5.36921620e-01 5.22574663e-01
1.27571678e+00 4.97292668e-01 1.24944031e+00 -2.04670683e-01
3.88294756e-01 -1.00140464e+00 -1.93748832e-01 -2.50422508e-02
9.05829668e-01 -9.74325538e-01 4.83541101e-01 -9.46685135e-01
-7.74064183e-01 1.04248416e+00 1.41831890e-01 -5.26497900e-01
1.25821102e+00 -9.39547569e-02 -4.50310856e-02 -2.56606787e-01
-2.12766290e-01 -2.64252156e-01 8.86716545e-01 6.85529530e-01
1.05564249e+00 4.77204293e-01 -8.62749815e-01 1.43581963e+00
-1.91049591e-01 2.88463503e-01 -2.24558428e-01 7.13533223e-01
-3.65225315e-01 -1.51125836e+00 -2.51871884e-01 8.64585340e-01
-4.21399623e-01 4.06949036e-02 -4.13453169e-02 2.57355809e-01
1.71462953e-01 1.63502765e+00 -1.33712903e-01 -7.55504906e-01
7.63358414e-01 -7.61975408e-01 9.44769502e-01 -3.72499406e-01
-9.15629387e-01 5.56886673e-01 6.22097015e-01 -1.97046995e-01
-7.76861310e-01 -4.52263534e-01 -2.01503381e-01 -9.51502144e-01
-3.61375898e-01 3.18732522e-02 6.18569851e-01 8.41060281e-01
2.53867149e-01 8.50452721e-01 4.33846653e-01 -8.81913841e-01
-6.09532632e-02 -4.66461062e-01 -8.30591500e-01 1.11792400e-01
-1.88527346e-01 -6.59078956e-01 4.12785023e-01 3.51777583e-01] | [11.031770706176758, 0.25499558448791504] |
37e1f558-8b4c-4e00-bb0e-86940dfebf98 | client-selection-for-federated-policy | 2305.10978 | null | https://arxiv.org/abs/2305.10978v3 | https://arxiv.org/pdf/2305.10978v3.pdf | Client Selection for Federated Policy Optimization with Environment Heterogeneity | The development of Policy Iteration (PI) has inspired many recent algorithms for Reinforcement Learning (RL), including several policy gradient methods, that gained both theoretical soundness and empirical success on a variety of tasks. The theory of PI is rich in the context of centralized learning, but its study is still in the infant stage under the federated setting. This paper explores the federated version of Approximate PI (API) and derives its error bound, taking into account the approximation error introduced by environment heterogeneity. We theoretically prove that a proper client selection scheme can reduce this error bound. Based on the theoretical result, we propose a client selection algorithm to alleviate the additional approximation error caused by environment heterogeneity. Experiment results show that the proposed algorithm outperforms other biased and unbiased client selection methods on the federated mountain car problem by effectively selecting clients with a lower level of heterogeneity from the population distribution. | ['S. H. Song', 'Zhijie Xie'] | 2023-05-18 | null | null | null | null | ['policy-gradient-methods'] | ['methodology'] | [-1.65313348e-01 6.23460300e-02 -7.34532714e-01 -1.32618964e-01
-7.77195394e-01 -2.65029699e-01 4.94223684e-01 -6.11352287e-02
-6.95250094e-01 1.25493371e+00 2.41449311e-01 -4.60574746e-01
-3.66321832e-01 -6.86391771e-01 -8.34085703e-01 -1.01065838e+00
-3.02097291e-01 7.43637800e-01 3.93271297e-02 -1.39956534e-01
3.18507880e-01 1.58522129e-01 -1.79909444e+00 -2.40523592e-01
1.23983991e+00 9.51484323e-01 3.48932445e-01 3.31336707e-01
-3.13121825e-01 8.63385916e-01 -7.28591502e-01 -1.61280632e-01
4.64747995e-01 -5.49693823e-01 -7.70215631e-01 -2.35178694e-01
-1.05472365e-02 -5.84736109e-01 -5.26265018e-02 1.17056203e+00
8.19255829e-01 2.65811265e-01 3.34403098e-01 -1.53839111e+00
-1.22673556e-01 1.04039812e+00 -6.20747983e-01 9.48600993e-02
1.55928001e-01 1.61664963e-01 1.01014376e+00 -3.12454253e-01
4.41151619e-01 1.42582428e+00 3.91500860e-01 6.95133209e-01
-1.07904434e+00 -6.88938379e-01 6.01082146e-01 3.56676280e-01
-8.83061349e-01 -3.11618508e-03 5.30424118e-01 1.19973972e-01
6.21530175e-01 1.62239581e-01 5.93285263e-01 1.20034969e+00
-2.28558928e-01 1.24846375e+00 1.47066510e+00 -6.39495730e-01
8.61045659e-01 4.05114442e-01 -1.90227881e-01 5.42216420e-01
4.07238692e-01 6.23009741e-01 -5.59931517e-01 -8.25328648e-01
6.32087708e-01 -9.74358618e-02 -2.90459096e-01 -7.76705325e-01
-6.77791238e-01 1.16286743e+00 1.12473361e-01 -6.32794900e-03
-6.94759727e-01 2.13165000e-01 3.14741105e-01 4.99632806e-01
3.39213818e-01 9.75231826e-02 -5.26714861e-01 -3.86650473e-01
-5.66826284e-01 5.79517961e-01 1.21777081e+00 8.55710030e-01
6.77443862e-01 2.61830837e-01 -4.11382675e-01 7.75877595e-01
1.83352649e-01 7.46446192e-01 6.19128168e-01 -1.26297903e+00
4.19989675e-01 2.94058293e-01 6.22571647e-01 -3.51171702e-01
-2.33657192e-02 -5.35655439e-01 -3.35702360e-01 5.39066553e-01
5.82398057e-01 -5.49312353e-01 -2.02696338e-01 2.04158998e+00
7.65924990e-01 1.66501820e-01 -2.00399254e-02 8.80551398e-01
-2.02754408e-01 3.52715462e-01 4.37834300e-02 -5.71092844e-01
8.33888471e-01 -8.91695380e-01 -5.63667119e-01 2.32159749e-01
4.36097831e-01 -1.97151661e-01 1.23396158e+00 4.84942526e-01
-8.53928924e-01 1.33335844e-01 -7.03112066e-01 8.32343698e-01
8.61331746e-02 -3.01579207e-01 7.35222280e-01 1.04930615e+00
-9.64732409e-01 7.23790348e-01 -6.44240797e-01 -5.11639893e-01
5.71820974e-01 3.59702647e-01 3.83396834e-01 -5.41546568e-02
-1.04407549e+00 7.98761308e-01 5.28238900e-02 -4.65382755e-01
-1.06621373e+00 -5.89734793e-01 -1.58689454e-01 1.47087589e-01
8.36493552e-01 -7.88627803e-01 1.70896530e+00 -1.18898058e+00
-2.09667420e+00 1.80002321e-02 -5.51801585e-02 -7.70685732e-01
9.96328890e-01 -2.61023879e-01 1.24893814e-01 1.02268904e-01
8.52520857e-03 1.24706319e-02 9.28666174e-01 -1.32676387e+00
-1.44987321e+00 -5.48531294e-01 7.40163727e-03 5.32939494e-01
-5.15390515e-01 -7.57494941e-02 2.84506530e-01 -2.19172761e-01
-6.62932158e-01 -6.39147341e-01 -6.54295862e-01 -3.55409503e-01
1.76066294e-01 -5.87741673e-01 7.08781064e-01 -2.16307402e-01
8.99109304e-01 -1.72537231e+00 -1.64095983e-01 4.25342768e-01
-2.76770350e-02 1.36561207e-02 -3.80083881e-02 5.44629693e-01
6.59659326e-01 -1.57809168e-01 -7.09714741e-02 -8.69084522e-02
3.78581703e-01 3.27710152e-01 -5.48362136e-01 5.81502020e-01
-5.55867553e-01 3.26615304e-01 -1.25678241e+00 -3.98532480e-01
-3.06119043e-02 5.57465218e-02 -8.06186259e-01 3.30055863e-01
-5.09962976e-01 2.81365871e-01 -8.83733809e-01 5.35331488e-01
5.36085010e-01 5.54807624e-03 4.79587108e-01 5.23346543e-01
-3.76445264e-01 1.51915655e-01 -1.17084682e+00 1.10375106e+00
-5.90651333e-01 -9.72476602e-03 6.56109989e-01 -1.21338952e+00
7.57753611e-01 3.70738119e-01 9.30544436e-01 -6.69985950e-01
2.08971322e-01 4.58413988e-01 -1.47006571e-01 -4.33587432e-01
3.41276109e-01 -1.01320200e-01 2.34305680e-01 7.89231539e-01
-2.49130204e-01 4.20057297e-01 -6.52707890e-02 -3.72329019e-02
1.12187958e+00 4.02018279e-01 1.70072019e-01 -4.64103520e-01
3.21236223e-01 1.33140199e-02 7.25826383e-01 1.42607617e+00
-6.87301219e-01 -3.86199206e-01 3.19235057e-01 -2.82657236e-01
-8.80661011e-01 -7.84820795e-01 -2.75103264e-02 1.49844325e+00
2.56547958e-01 6.48727864e-02 -7.76493073e-01 -8.94672632e-01
5.74445665e-01 7.23192811e-01 -5.54724038e-01 -7.52580464e-02
-5.06110609e-01 -7.81736910e-01 2.73740381e-01 2.26836786e-01
6.89605892e-01 -1.01888323e+00 -9.59521532e-01 4.11788195e-01
-1.07662916e-01 -5.77261508e-01 -1.72582805e-01 1.75383106e-01
-8.59167695e-01 -1.08466494e+00 -8.49603832e-01 -4.26188082e-01
2.81825572e-01 1.71105087e-01 8.06654513e-01 -1.95709139e-01
1.55661777e-01 8.58107150e-01 -3.65533233e-01 -6.63085222e-01
-3.22021097e-01 1.73760071e-01 3.66696537e-01 7.27094635e-02
4.31097329e-01 -6.05829179e-01 -6.80233359e-01 4.19456303e-01
-6.23942375e-01 -4.56980735e-01 4.86343384e-01 1.17410541e+00
4.04413790e-01 2.69004535e-02 9.84471381e-01 -8.49150300e-01
9.87565994e-01 -5.88474035e-01 -9.38738883e-01 3.66355032e-01
-1.08882916e+00 3.92747432e-01 7.82469749e-01 -6.16499245e-01
-1.45665061e+00 -2.14130059e-01 2.46339753e-01 -3.29118878e-01
-4.90373857e-02 1.46727428e-01 -9.98264775e-02 -9.66110304e-02
6.84931219e-01 2.22880840e-01 3.61545771e-01 -7.57572889e-01
4.24240589e-01 7.89787054e-01 2.04530954e-01 -1.30676031e+00
4.82104480e-01 6.41776919e-01 -8.36474374e-02 -5.45519888e-01
-4.79816049e-01 -2.44556516e-01 2.89525002e-01 -2.60559618e-01
4.19599637e-02 -7.29771018e-01 -1.18834221e+00 1.20780535e-01
-5.02018273e-01 -6.85110986e-01 -5.20549059e-01 6.93692446e-01
-1.02593768e+00 2.87669003e-01 -3.80450606e-01 -1.35563862e+00
-4.43334132e-01 -9.05178607e-01 5.32185316e-01 3.34630847e-01
3.54832739e-01 -8.21811736e-01 5.50001740e-01 -2.05600262e-02
6.88676298e-01 -1.37733137e-02 7.82779276e-01 -4.69237059e-01
-4.78545457e-01 2.05631167e-01 1.84463277e-01 -2.58539803e-02
-3.96352075e-02 -4.24188107e-01 -7.58122265e-01 -6.42300248e-01
1.55279592e-01 -3.20803165e-01 6.93086028e-01 4.86944705e-01
9.86669898e-01 -7.60080874e-01 -3.67548913e-01 4.18941408e-01
1.57550514e+00 1.90488681e-01 8.33188742e-02 8.17123592e-01
1.72220320e-01 4.74227428e-01 7.86239088e-01 1.19898868e+00
3.10463041e-01 5.24084508e-01 7.41853654e-01 2.11880863e-01
3.15625101e-01 -4.29410428e-01 5.81913173e-01 1.52437940e-01
-1.96901232e-01 1.52951658e-01 -3.58427882e-01 5.13160110e-01
-2.31765532e+00 -1.18133962e+00 4.81119275e-01 2.64120722e+00
8.99972022e-01 -3.27585399e-01 8.71598601e-01 -1.51250660e-01
8.60120416e-01 -2.61459053e-01 -1.05825424e+00 -4.93184179e-01
-1.77048668e-01 1.34093776e-01 8.43180239e-01 4.20301616e-01
-5.33864796e-01 8.24640453e-01 6.58048105e+00 8.99873793e-01
-9.38755572e-01 1.67314872e-01 5.67962468e-01 -2.46986806e-01
-3.15510392e-01 8.86426643e-02 -8.20683718e-01 4.50165004e-01
9.29137945e-01 -5.49307227e-01 1.00497556e+00 1.37403822e+00
4.32590604e-01 -2.20087826e-01 -9.00960207e-01 6.45264387e-01
-3.48241478e-01 -9.18554485e-01 -1.97538421e-01 2.83923537e-01
9.38224256e-01 3.50639462e-01 1.27354488e-01 6.26428306e-01
9.53521132e-01 -6.07358038e-01 6.96940839e-01 7.96242431e-02
4.76999968e-01 -1.18328547e+00 5.03410757e-01 6.57699585e-01
-7.21373975e-01 -8.35166395e-01 -4.59301680e-01 -2.28815258e-01
-1.30666390e-01 2.35021248e-01 -7.86152899e-01 3.84153098e-01
7.69797266e-01 -4.71802913e-02 2.99093183e-02 1.18734729e+00
1.59280207e-02 8.34778786e-01 -5.90955555e-01 -4.92761344e-01
3.59994441e-01 -4.19439822e-01 6.03272557e-01 6.11620605e-01
2.36301512e-01 -2.87435263e-01 3.64384264e-01 6.41632795e-01
2.97931973e-02 4.37782943e-01 -4.67581302e-01 3.51070911e-01
7.31154442e-01 9.49748993e-01 -2.70268947e-01 -2.36104250e-01
-3.33037019e-01 5.84292173e-01 5.50751746e-01 5.69937825e-01
-6.55099809e-01 -2.25596339e-01 9.27963197e-01 -1.10399134e-01
5.81611574e-01 3.07772994e-01 -6.70269877e-02 -8.02764118e-01
6.44367859e-02 -1.13593030e+00 5.81936777e-01 2.13860556e-01
-1.29211354e+00 1.76780850e-01 1.59021318e-01 -8.79767954e-01
-5.94558001e-01 -3.01005006e-01 -5.46578169e-01 5.18295527e-01
-1.58303487e+00 -5.53546727e-01 2.65096217e-01 7.50535429e-01
2.34592676e-01 -3.52773845e-01 6.91121459e-01 1.92082211e-01
-6.35007441e-01 7.81218946e-01 9.63912070e-01 -4.83047277e-01
5.20931721e-01 -1.31660354e+00 -4.65755165e-01 3.55056822e-01
-3.93409729e-01 3.95684332e-01 8.34260941e-01 -4.46868330e-01
-1.67005157e+00 -8.49406302e-01 4.61797535e-01 -6.32318249e-03
4.92708474e-01 -2.95956563e-02 -4.35559541e-01 3.01230729e-01
1.98362365e-01 -2.65713334e-01 2.60364205e-01 2.30984211e-01
-1.78529635e-01 -3.86710763e-01 -1.52353585e+00 8.04940820e-01
1.04635370e+00 -1.68476298e-01 -4.80222590e-02 3.04257542e-01
5.27249098e-01 3.82863218e-03 -5.85918367e-01 1.01043344e-01
5.77964008e-01 -1.03376281e+00 5.63753188e-01 -7.17536926e-01
-2.42668971e-01 9.73458290e-02 -1.58240035e-01 -1.43656576e+00
-1.46652326e-01 -1.23508370e+00 -5.50619900e-01 1.00236416e+00
4.42876332e-02 -9.90465760e-01 1.05155706e+00 4.72031325e-01
2.77275443e-01 -9.23209131e-01 -1.24762297e+00 -1.21608531e+00
3.99305344e-01 -4.06895019e-02 9.66553092e-01 4.87856001e-01
3.42669010e-01 5.41613102e-02 -4.59638715e-01 -1.02924697e-01
1.16606832e+00 4.42869961e-01 7.85405755e-01 -1.00836337e+00
-6.84214532e-01 -7.30811357e-01 3.15606475e-01 -1.04427636e+00
3.96098584e-01 -3.37418437e-01 1.42920196e-01 -1.17375982e+00
1.50124371e-01 -6.46758080e-01 -3.80122453e-01 2.73998380e-01
-2.50283599e-01 -6.77241504e-01 1.49191633e-01 1.89846799e-01
-6.90220296e-01 9.40726221e-01 9.87628400e-01 2.51766015e-02
-4.05816078e-01 4.91484463e-01 -7.18123198e-01 4.68072057e-01
1.18773949e+00 -5.32903910e-01 -5.18655479e-01 -7.13132173e-02
-9.59327444e-02 1.39312848e-01 5.34008024e-03 -6.77798808e-01
-1.00765184e-01 -7.02498972e-01 -2.00079665e-01 -1.49579331e-01
-1.66391149e-01 -8.89738321e-01 -2.01433197e-01 8.65889966e-01
-5.45527995e-01 6.79530203e-04 -3.00308406e-01 9.02832747e-01
2.51587510e-01 -2.23085448e-01 7.13370919e-01 -2.48302877e-01
-3.61038029e-01 4.31031048e-01 -5.18864810e-01 2.65919417e-01
8.50444555e-01 1.19373128e-01 -2.20392182e-01 -7.42775321e-01
-5.64996377e-02 4.78833646e-01 4.36910152e-01 3.02912463e-02
2.58009076e-01 -1.11147952e+00 -5.97056150e-01 -1.53251067e-01
-2.44959980e-01 -4.56999958e-01 -1.48586720e-01 7.79316664e-01
-4.83339466e-02 2.69931406e-01 -9.42424461e-02 -3.01997960e-01
-8.80645931e-01 5.65380454e-01 4.18497384e-01 -4.21821415e-01
-6.60080016e-01 3.90760690e-01 -1.82743937e-01 -3.27194214e-01
8.43928695e-01 2.61799851e-05 7.38366917e-02 -4.21714425e-01
6.06796384e-01 8.73982012e-01 -2.86027044e-01 -1.00623913e-01
-1.06984250e-01 -1.96847524e-02 4.11453657e-03 -5.77841043e-01
1.30335569e+00 -2.17797995e-01 2.56766677e-01 5.97913824e-02
6.36961222e-01 -1.81705371e-01 -1.65039551e+00 -5.59854746e-01
2.75273025e-01 -6.91345453e-01 5.12184128e-02 -6.77718997e-01
-9.49590325e-01 3.47970396e-01 9.51598227e-01 2.57536441e-01
9.06115353e-01 -3.04975599e-01 5.60800195e-01 5.93232930e-01
1.06855810e+00 -1.52954471e+00 -2.03122839e-01 3.51490587e-01
4.31996256e-01 -1.03082097e+00 -1.09309584e-01 2.16926917e-01
-7.45401978e-01 6.01576209e-01 6.98573291e-01 -1.23534508e-01
4.33131725e-01 1.32088840e-01 -6.08034879e-02 2.43867785e-01
-9.48526382e-01 -4.58191097e-01 -7.69764662e-01 9.89577472e-01
-1.87781185e-03 2.12570280e-01 -7.32761741e-01 4.56804484e-01
2.93205976e-02 1.66388467e-01 1.66460469e-01 1.13037884e+00
-7.83993304e-01 -1.35368264e+00 -5.41280031e-01 3.62232327e-01
-5.41425526e-01 3.60863775e-01 -1.66404694e-01 6.36182249e-01
-2.15041205e-01 9.58998621e-01 -1.88938484e-01 1.05112813e-01
6.22448660e-02 -6.15212955e-02 8.02961349e-01 1.02335341e-01
-7.39041567e-01 1.83368430e-01 2.33244002e-02 -6.75355792e-01
-4.01278287e-01 -7.17043340e-01 -1.13875580e+00 -4.87676889e-01
-2.70414382e-01 7.28448153e-01 6.77257657e-01 7.60056555e-01
3.96495640e-01 -1.99293494e-02 1.16071022e+00 -6.57891512e-01
-1.82320714e+00 -7.76201069e-01 -8.75728965e-01 1.43708929e-01
2.54259795e-01 -9.31279898e-01 -5.15834391e-01 -6.93266749e-01] | [4.080258369445801, 2.3819589614868164] |
83025f4d-7838-4a57-8bd7-064f04992a2d | a-graph-convolution-for-signed-directed | 2208.11511 | null | https://arxiv.org/abs/2208.11511v3 | https://arxiv.org/pdf/2208.11511v3.pdf | A Graph Convolution for Signed Directed Graphs | A signed directed graph is a graph with sign and direction information on the edges. Even though signed directed graphs are more informative than unsigned or undirected graphs, they are more complicated to analyze and have received less research attention. This paper investigates a spectral graph convolution model to fully utilize the information embedded in signed directed edges. We propose a novel complex Hermitian adjacency matrix that encodes graph information via complex numbers. Compared to a simple connection-based adjacency matrix, the complex Hermitian can represent edge direction, sign, and connectivity via its phases and magnitudes. Then, we define a magnetic Laplacian of the proposed adjacency matrix and prove that it is positive semi-definite (PSD) for the analyses using spectral graph convolution. We perform extensive experiments on four real-world datasets. Our experiments show that the proposed scheme outperforms several state-of-the-art techniques. | ['Chong-Kwon Kim', 'Taewook Ko'] | 2022-08-23 | null | null | null | null | ['link-sign-prediction'] | ['graphs'] | [ 3.06192100e-01 2.02650413e-01 7.53326789e-02 -4.24507141e-01
4.30001795e-01 -6.54580534e-01 4.48828220e-01 -7.57040130e-03
-8.38699117e-02 5.98942459e-01 -1.18039632e-02 -6.81701481e-01
-4.42849547e-01 -9.12363231e-01 -4.53692704e-01 -6.45765781e-01
-6.91205978e-01 -1.07914574e-01 2.53848344e-01 -2.38586158e-01
3.24235976e-01 3.94775063e-01 -9.74704325e-01 -2.12917954e-01
9.81341362e-01 9.46970344e-01 -1.37974739e-01 4.81472522e-01
4.41616103e-02 7.07396448e-01 -9.11855474e-02 -2.78614551e-01
9.84340981e-02 -4.96816397e-01 -4.84603345e-01 2.05889180e-01
2.10650325e-01 1.54979184e-01 -8.07846785e-01 1.58380449e+00
2.89635599e-01 -2.37331092e-01 6.95761621e-01 -1.45607924e+00
-9.81372237e-01 7.39497483e-01 -8.18544567e-01 -4.99921367e-02
3.30229491e-01 -1.01455249e-01 1.07054853e+00 -6.63909018e-01
5.71707487e-01 9.03032482e-01 6.64573371e-01 -8.56099203e-02
-1.09065711e+00 -8.22391808e-01 3.76068354e-02 4.60007608e-01
-1.36575520e+00 8.64500925e-02 1.42679501e+00 -1.77985325e-01
6.32639468e-01 3.99412185e-01 9.59819317e-01 4.97255534e-01
1.17559545e-01 1.88521549e-01 1.24349451e+00 -4.45287943e-01
-8.16613510e-02 -1.94560468e-01 4.02617455e-01 1.47021461e+00
8.01393390e-01 -1.85693428e-01 -3.05632949e-01 -8.01801085e-02
7.13580966e-01 -2.05869928e-01 -8.44619632e-01 -8.33804846e-01
-1.43167818e+00 7.63590395e-01 8.54178190e-01 3.17226499e-01
-4.12242472e-01 3.64133954e-01 6.46489486e-02 1.23656191e-01
7.79970363e-02 -1.23964444e-01 3.13465446e-01 1.86301246e-01
-3.63391012e-01 -5.04991055e-01 1.06949174e+00 9.98284757e-01
8.67117822e-01 1.38379067e-01 2.55297810e-01 4.77532059e-01
5.37086308e-01 9.06826079e-01 -6.68927580e-02 -5.98330259e-01
3.84628713e-01 8.68677676e-01 -3.52887779e-01 -1.67385364e+00
-7.76999354e-01 -4.99268711e-01 -1.45554185e+00 -3.53446901e-02
1.76053256e-01 -2.03582108e-01 -5.79174697e-01 1.62782454e+00
1.73592393e-03 4.25139308e-01 -1.71671808e-01 1.01532614e+00
9.52661753e-01 4.09402966e-01 -3.59104276e-01 -1.66887909e-01
1.03124070e+00 -5.14609218e-01 -8.74160051e-01 -1.78788155e-01
2.35187307e-01 -7.83663154e-01 6.80460572e-01 2.18649745e-01
-8.79580855e-01 -7.64043331e-02 -1.35183299e+00 3.05157214e-01
-3.53689253e-01 3.37127745e-01 1.22724295e+00 8.04570615e-01
-1.20843267e+00 5.82266331e-01 -5.01099288e-01 -4.46578175e-01
-1.07848793e-01 3.09493840e-01 -5.17071068e-01 8.82622004e-02
-1.11605346e+00 3.37736100e-01 1.76680371e-01 5.60611427e-01
2.01414317e-01 -1.43144608e-01 -9.43627834e-01 1.71286881e-01
1.26927141e-02 -3.96128744e-01 6.09679103e-01 -6.93757892e-01
-1.14916348e+00 7.95333564e-01 2.02715144e-01 -1.96164608e-01
1.75530881e-01 7.63480783e-01 -8.99272799e-01 4.47648078e-01
-2.59133935e-01 5.18026436e-03 6.98357463e-01 -1.14409769e+00
-2.90847141e-02 -1.67250067e-01 8.69471133e-02 -1.05823740e-01
-6.40562773e-01 -2.82314539e-01 -3.48306954e-01 -6.40070200e-01
8.31576347e-01 -9.78007793e-01 7.06378147e-02 1.65138721e-01
-6.70591056e-01 3.49790782e-01 9.02393818e-01 -4.77792054e-01
1.64391518e+00 -2.17451882e+00 -4.52583805e-02 9.86649096e-01
6.37229562e-01 1.89282358e-01 -1.32335916e-01 6.89323902e-01
-4.30846989e-01 5.48319258e-02 -5.17799020e-01 4.31397021e-01
1.14973426e-01 5.55638969e-02 1.94419771e-01 8.81792486e-01
2.11597413e-01 9.64532673e-01 -9.84018326e-01 -3.47057402e-01
1.36747658e-01 7.40155280e-01 -4.24089849e-01 -3.12119335e-01
4.92972165e-01 5.15137613e-02 -4.61105943e-01 4.75454479e-01
1.17284119e+00 -7.75844514e-01 7.41652131e-01 -8.72513533e-01
-4.91296984e-02 5.68867140e-02 -1.57694447e+00 1.18387699e+00
3.93726416e-02 7.70904064e-01 4.87416118e-01 -1.32115233e+00
1.00399804e+00 -5.25939651e-02 4.00426894e-01 -6.82381511e-01
2.48822823e-01 3.16505373e-01 4.59202498e-01 -2.70434588e-01
3.34723234e-01 1.30871117e-01 5.10463417e-02 5.15061975e-01
-4.14139293e-02 2.76013883e-03 4.12158191e-01 7.99216211e-01
1.38617921e+00 -4.84511465e-01 1.69735819e-01 -5.08886397e-01
7.96986997e-01 -5.49204409e-01 2.45953143e-01 2.10567176e-01
-1.16984196e-01 3.11919868e-01 6.86345637e-01 1.45670876e-01
-3.93407047e-01 -1.44477940e+00 6.53250888e-02 -2.94284765e-02
3.98567557e-01 -3.89150023e-01 -5.83078086e-01 -4.37797457e-01
9.69672296e-03 3.75476666e-02 -4.42036033e-01 -3.80806297e-01
-3.96378905e-01 -9.41578567e-01 4.12678659e-01 2.51234949e-01
9.87826109e-01 -6.90812767e-01 4.33916077e-02 -6.92731217e-02
-4.31418121e-02 -9.87472236e-01 -8.05138290e-01 1.02678254e-01
-6.70745194e-01 -1.51298201e+00 -3.69171292e-01 -1.21251845e+00
1.22415376e+00 5.63724935e-01 9.03205395e-01 1.59854889e-01
-2.50127286e-01 6.34548724e-01 -4.22718585e-01 -5.10098296e-04
-8.05176646e-02 -3.35984796e-01 -4.22044918e-02 4.74355668e-01
5.15100658e-02 -9.31278825e-01 -6.46526814e-01 1.52878940e-01
-8.72189701e-01 8.00114498e-02 7.95077443e-01 8.60533118e-01
3.56569052e-01 1.68845743e-01 4.56584066e-01 -1.03613305e+00
1.02179992e+00 -1.93896353e-01 -6.87444985e-01 3.14539105e-01
-6.49049580e-01 3.86097133e-01 6.71527326e-01 -2.57893384e-01
-7.67547965e-01 -2.83736568e-02 5.06532073e-01 -6.71322197e-02
5.07261455e-01 9.83762383e-01 -1.04326777e-01 -9.39216912e-01
2.69495785e-01 1.59093380e-01 7.64754042e-02 -1.68151423e-01
4.19582039e-01 5.44842780e-01 6.23018801e-01 -3.08419973e-01
9.69997108e-01 5.61246634e-01 6.47967398e-01 -9.45380509e-01
-3.50068361e-02 -3.66188318e-01 -5.67557931e-01 -5.10571718e-01
4.28681582e-01 -3.78463477e-01 -1.22270763e+00 6.54754579e-01
-1.16115844e+00 1.98828727e-01 2.52560377e-01 7.58429110e-01
-1.19798325e-01 8.88635695e-01 -7.76736796e-01 -6.76674962e-01
-2.50671446e-01 -6.85778856e-01 6.03444993e-01 1.21894725e-01
2.23406672e-01 -1.28679204e+00 -2.51140058e-01 -2.52669871e-01
5.30560374e-01 2.47098535e-01 9.54040051e-01 4.44921316e-04
-7.75381982e-01 -4.35337067e-01 -8.02911043e-01 1.73540741e-01
1.21647656e-01 8.22897255e-02 -3.86796445e-01 -1.72078073e-01
-2.05931783e-01 2.03863978e-01 8.51529121e-01 5.47310039e-02
1.15255356e+00 -1.71173975e-01 -2.84050375e-01 7.86019266e-01
1.42923069e+00 7.09558353e-02 5.78624427e-01 -2.18539178e-01
8.26530159e-01 1.36119217e-01 2.44550273e-01 4.19412374e-01
2.74215132e-01 2.74831522e-02 5.06762624e-01 -2.72301137e-01
-5.47254272e-02 -2.31681645e-01 7.20552355e-02 1.69117761e+00
-3.33509386e-01 -2.88600475e-01 -6.62416875e-01 6.23205751e-02
-1.84776652e+00 -9.83613312e-01 -9.66956615e-01 2.08311820e+00
4.33704436e-01 1.45650402e-01 -3.62708747e-01 4.90202636e-01
9.85668957e-01 4.12153393e-01 -3.80282193e-01 -2.38358259e-01
-5.01405537e-01 5.58136106e-01 9.91328895e-01 4.69031572e-01
-9.21522617e-01 3.86268169e-01 6.19197845e+00 3.72622490e-01
-1.20694542e+00 -2.68385291e-01 6.71393611e-03 6.31613195e-01
-6.22352123e-01 1.04809247e-01 1.18631341e-01 3.85557652e-01
4.61927891e-01 -4.33459461e-01 6.66199565e-01 3.11484158e-01
-1.13618094e-02 -3.36375870e-02 -7.64209092e-01 1.34452403e+00
8.99153352e-02 -1.20051038e+00 -5.28056808e-02 1.14627331e-01
5.79549253e-01 -2.31241494e-01 5.52806854e-02 -2.22283036e-01
8.91463310e-02 -7.62805521e-01 4.68456537e-01 8.54823947e-01
7.53409147e-01 -6.51682079e-01 6.08983934e-01 -1.12821311e-01
-1.68324256e+00 2.22053081e-01 -1.05765805e-01 -2.20971435e-01
2.48828322e-01 9.85703588e-01 -4.07641917e-01 7.71683574e-01
2.63900757e-01 9.71568584e-01 -6.32466018e-01 1.18481219e+00
-3.47493798e-01 7.55280197e-01 -2.08765686e-01 -5.78997731e-01
2.39977598e-01 -9.88472700e-01 4.99351442e-01 1.00496650e+00
4.12757158e-01 2.41761461e-01 2.08668038e-02 1.02483857e+00
-2.05440074e-01 1.39614731e-01 -7.78646111e-01 -6.70830965e-01
3.56800824e-01 1.54110360e+00 -9.64356899e-01 -3.03100310e-02
-7.54306853e-01 1.26339805e+00 1.44940376e-01 7.26894855e-01
-8.19962084e-01 -1.07227266e+00 3.36617380e-01 -3.37551944e-02
1.11875311e-01 -5.88733375e-01 -2.06358079e-02 -1.07134044e+00
2.71280348e-01 -3.23721528e-01 1.76658198e-01 -9.19486821e-01
-1.37538493e+00 3.28517139e-01 -3.70055288e-01 -1.33234143e+00
1.73680514e-01 -1.02126324e+00 -6.10031724e-01 8.78945947e-01
-1.32685852e+00 -9.84825134e-01 -6.03445232e-01 6.00559294e-01
-7.95625150e-01 5.47247492e-02 6.55676246e-01 6.30300462e-01
-2.55784869e-01 3.54130745e-01 1.78905249e-01 5.18060207e-01
3.28080535e-01 -1.20259380e+00 2.09825605e-01 8.02662909e-01
6.47615865e-02 7.71661401e-01 3.07349831e-01 -7.13800967e-01
-2.08665323e+00 -7.71829963e-01 7.23424375e-01 5.20065725e-01
1.15390372e+00 -3.18163872e-01 -6.61415935e-01 6.64657712e-01
3.00319076e-01 4.56463367e-01 4.35876817e-01 -2.54467487e-01
-4.66036707e-01 -1.58877626e-01 -1.12741780e+00 5.61699450e-01
1.56910431e+00 -7.10215926e-01 5.51471040e-02 2.04745620e-01
3.64631474e-01 -6.10224046e-02 -6.19374752e-01 4.60028917e-01
4.76507246e-01 -1.09251845e+00 8.52599025e-01 -1.67652816e-01
-1.77307710e-01 -5.99825621e-01 -8.64886716e-02 -1.56646931e+00
-6.82984054e-01 -6.96583211e-01 9.26403925e-02 1.00802207e+00
3.39531183e-01 -1.20778310e+00 6.29393339e-01 -1.28519228e-02
-3.51540036e-02 -5.63506961e-01 -6.03809834e-01 -7.55173385e-01
-4.97619808e-01 -2.83709705e-01 4.35829252e-01 1.23200893e+00
5.12103975e-01 5.05534112e-01 -2.47073341e-02 3.02541137e-01
8.88279378e-01 3.37456793e-01 1.04656398e-01 -1.50219274e+00
-1.52374834e-01 -6.88197196e-01 -1.09729791e+00 -7.66397119e-01
1.52977571e-01 -1.42854929e+00 -4.08680379e-01 -1.73467553e+00
7.04987124e-02 -3.92127007e-01 -1.51841909e-01 1.18294753e-01
2.59755075e-01 4.32877362e-01 -1.97503746e-01 -5.95837295e-01
-4.21117723e-01 4.10395890e-01 1.38668418e+00 -3.79211843e-01
9.42972749e-02 -3.94162118e-01 -3.70061547e-01 7.05637515e-01
7.82917142e-01 1.52945057e-01 -5.24697781e-01 -4.18214127e-02
5.83167493e-01 1.01294056e-01 4.50005442e-01 -1.08760285e+00
3.61617863e-01 1.27947792e-01 -4.18812260e-02 -5.70999682e-01
1.73627406e-01 -8.87900591e-01 3.24109346e-01 6.10414207e-01
1.54471070e-01 1.79797426e-01 -1.95908040e-01 7.11532414e-01
-2.94291496e-01 4.26924340e-02 6.31702304e-01 1.36636645e-01
-7.72362769e-01 2.73823321e-01 -5.36441058e-02 1.78489683e-03
8.29885662e-01 -1.96183279e-01 -5.74650824e-01 -7.15072870e-01
-6.64358139e-01 8.15192983e-02 4.11658734e-01 -6.06825203e-02
9.45380151e-01 -1.69842172e+00 -2.55664319e-01 4.28077787e-01
1.14066966e-01 -6.84881628e-01 2.27115732e-02 1.19730318e+00
-8.01395178e-01 4.29249316e-01 -1.02741323e-01 -3.56810391e-01
-1.44399190e+00 3.29925805e-01 3.76409203e-01 1.21683031e-01
-4.27936137e-01 5.08118093e-01 -1.63036034e-01 -4.99195009e-01
1.52490929e-01 -3.92472476e-01 -3.03104315e-02 -1.18041553e-01
1.11283764e-01 4.45008814e-01 5.84048219e-02 -8.90063584e-01
-4.70850676e-01 8.05217981e-01 7.45882809e-01 1.02413404e-04
1.10987639e+00 -1.24910638e-01 -7.57004857e-01 2.48985380e-01
1.46686387e+00 2.28954226e-01 -4.32647616e-01 -4.44473252e-02
-2.60778278e-01 -3.85653138e-01 3.82926613e-02 -4.45619345e-01
-1.44909990e+00 9.65251982e-01 4.56683874e-01 6.70407534e-01
1.06606030e+00 -2.03569621e-01 6.28885984e-01 5.40772974e-01
6.01479888e-01 -8.17495883e-01 -1.53044949e-03 5.41581392e-01
7.83624291e-01 -8.32498550e-01 -5.87409958e-02 -1.29322231e+00
-2.42942408e-01 1.22116113e+00 3.14040154e-01 -1.56458661e-01
1.29275334e+00 1.59667119e-01 -3.38192761e-01 -4.64417100e-01
1.53491469e-02 -5.04348278e-01 4.88793194e-01 7.22263455e-01
6.75240040e-01 4.42291886e-01 -8.14561367e-01 4.14288163e-01
-1.79918483e-01 -1.68519840e-01 7.49830723e-01 9.04897451e-01
-4.53179210e-01 -1.03745365e+00 -3.16761464e-01 7.42693841e-01
1.16044730e-01 -1.51082829e-01 -7.67796934e-01 5.26189089e-01
-3.44604611e-01 1.02438533e+00 5.66677526e-02 -6.66536570e-01
2.73103863e-01 -1.73759341e-01 8.24742734e-01 -3.08186561e-01
6.84905648e-02 -3.62037063e-01 2.38119423e-01 -4.16485697e-01
-6.00174308e-01 -3.53801191e-01 -1.86879170e+00 -6.63587749e-01
-3.45525801e-01 1.79333523e-01 6.07070208e-01 2.25713909e-01
3.10584992e-01 6.63324714e-01 6.31103814e-01 -5.78838408e-01
-1.16266675e-01 -6.39652371e-01 -1.30343521e+00 5.02156138e-01
2.35954951e-02 -8.36059928e-01 -5.24414599e-01 -3.28150749e-01] | [7.056813716888428, 5.960743427276611] |
d71bc929-1022-47be-92f4-d7906470144e | self-play-reinforcement-learning-for-fast | 2010.00909 | null | https://arxiv.org/abs/2010.00909v1 | https://arxiv.org/pdf/2010.00909v1.pdf | Self-Play Reinforcement Learning for Fast Image Retargeting | In this study, we address image retargeting, which is a task that adjusts input images to arbitrary sizes. In one of the best-performing methods called MULTIOP, multiple retargeting operators were combined and retargeted images at each stage were generated to find the optimal sequence of operators that minimized the distance between original and retargeted images. The limitation of this method is in its tremendous processing time, which severely prohibits its practical use. Therefore, the purpose of this study is to find the optimal combination of operators within a reasonable processing time; we propose a method of predicting the optimal operator for each step using a reinforcement learning agent. The technical contributions of this study are as follows. Firstly, we propose a reward based on self-play, which will be insensitive to the large variance in the content-dependent distance measured in MULTIOP. Secondly, we propose to dynamically change the loss weight for each action to prevent the algorithm from falling into a local optimum and from choosing only the most frequently used operator in its training. Our experiments showed that we achieved multi-operator image retargeting with less processing time by three orders of magnitude and the same quality as the original multi-operator-based method, which was the best-performing algorithm in retargeting tasks. | ['Toshihiko Yamasaki', 'Xueting Wang', 'Satoshi Kosugi', 'Nobukatsu Kajiura'] | 2020-10-02 | null | null | null | null | ['image-retargeting'] | ['computer-vision'] | [ 4.73019332e-01 -4.10885997e-02 -1.20041125e-01 7.97095522e-02
-4.13235486e-01 -4.38514322e-01 1.34291932e-01 2.17509151e-01
-8.74399245e-01 6.01841331e-01 -3.10738325e-01 -1.37445495e-01
-3.04868728e-01 -5.82974494e-01 -6.87859356e-01 -8.28299880e-01
5.91888912e-02 8.12068954e-02 7.56872058e-01 -2.83460587e-01
7.68352926e-01 7.57085741e-01 -1.67553842e+00 1.13636121e-01
1.02754378e+00 7.73577869e-01 6.12140775e-01 8.20499718e-01
1.47090867e-01 4.82310534e-01 -6.22537434e-01 -2.25842297e-01
5.18188596e-01 -7.71214843e-01 -7.93779314e-01 4.32013482e-01
2.93116033e-01 -1.81639269e-02 -1.51387781e-01 1.20625103e+00
6.95469379e-01 5.86540401e-01 4.54842657e-01 -1.12801433e+00
-6.24472082e-01 4.89722461e-01 -9.67355371e-01 6.55796111e-01
3.06844890e-01 7.41611049e-02 8.09373915e-01 -5.98755658e-01
6.17996275e-01 1.06025743e+00 3.89181942e-01 6.08427644e-01
-1.43329442e+00 -4.42651719e-01 2.23710895e-01 3.66999388e-01
-1.59747338e+00 -2.84053266e-01 6.58281565e-01 -2.64263362e-01
8.36684763e-01 5.46044648e-01 6.63438261e-01 4.87332076e-01
6.69865012e-01 7.12119460e-01 1.10767853e+00 -9.45477664e-01
3.10318917e-01 2.11416408e-01 -3.46301287e-01 8.56214404e-01
1.02954480e-04 1.90568194e-01 -3.85031551e-01 6.41950965e-02
8.09254408e-01 -1.58274919e-01 -2.73933202e-01 -6.05249465e-01
-1.20620108e+00 8.96993697e-01 4.72871691e-01 5.03114820e-01
-4.43977982e-01 -2.00149883e-02 2.21911192e-01 4.62226182e-01
3.21742475e-01 9.70345259e-01 -3.43881011e-01 -3.86419543e-03
-6.31699860e-01 2.33239293e-01 2.48433173e-01 5.06199598e-01
7.86774397e-01 -2.39445269e-01 -3.24905157e-01 9.47950721e-01
-1.21133447e-01 9.89327282e-02 9.10119712e-01 -1.03970015e+00
3.95330578e-01 5.97735167e-01 8.22099373e-02 -1.18916392e+00
-3.79575312e-01 -2.51039028e-01 -5.13299525e-01 5.82473576e-01
2.65473604e-01 -1.76361382e-01 -9.95949030e-01 1.54960740e+00
3.61897826e-01 -4.94215749e-02 -1.46338940e-01 1.00447631e+00
1.68522969e-01 4.79031712e-01 1.25248998e-01 -5.46534479e-01
1.21060348e+00 -1.22271538e+00 -5.57254732e-01 -9.79665667e-02
5.09574175e-01 -8.85305166e-01 1.32651103e+00 3.31696987e-01
-1.14109421e+00 -6.20116234e-01 -1.08201551e+00 5.48691094e-01
-3.48704368e-01 2.12121233e-02 3.27341646e-01 7.14835405e-01
-1.13006330e+00 7.62388110e-01 -5.40796459e-01 -5.90197027e-01
8.33016038e-02 5.36531329e-01 -2.47275338e-01 1.76732212e-01
-1.05351937e+00 1.03722787e+00 7.66638100e-01 -4.81662452e-02
-5.52222311e-01 -4.71472770e-01 -5.64537048e-01 2.92746965e-02
8.29262018e-01 -4.97053802e-01 1.29103541e+00 -1.58876181e+00
-1.75538456e+00 5.25310934e-01 -1.56565048e-02 -4.51155037e-01
5.02417505e-01 1.17009051e-01 -5.10320425e-01 2.06985816e-01
-7.58336112e-03 9.55189347e-01 1.19059825e+00 -1.28891230e+00
-1.00605679e+00 -1.95027679e-01 6.60398975e-02 6.84567630e-01
-4.85963047e-01 -2.77579315e-02 -6.72912359e-01 -7.97045052e-01
8.40070918e-02 -1.27022076e+00 -5.97819924e-01 -1.63254708e-01
-1.79280460e-01 2.30303500e-02 5.50746262e-01 -2.44477063e-01
1.55071616e+00 -2.27526760e+00 3.76570791e-01 1.59964785e-01
1.93310916e-01 3.79092723e-01 -3.17231506e-01 4.46529657e-01
-1.25150919e-01 1.61118150e-01 -1.32158116e-01 9.26802233e-02
-4.17565495e-01 1.71021298e-01 -8.52298513e-02 3.21075261e-01
-5.03348410e-02 7.57444263e-01 -9.43674743e-01 -6.94746077e-01
2.49841020e-01 1.06451869e-01 -4.84623432e-01 1.14088610e-01
-4.91469689e-02 2.76118338e-01 -3.59143138e-01 2.29257777e-01
4.40547079e-01 -4.83852141e-02 -3.84284407e-02 -2.46889472e-01
-2.21896872e-01 -5.68574548e-01 -1.28908050e+00 1.23245561e+00
-5.71659744e-01 5.29788256e-01 -3.79706293e-01 -7.36163974e-01
8.20323110e-01 1.04861036e-01 7.37565219e-01 -9.06399190e-01
7.78679401e-02 8.52237865e-02 1.50198415e-01 -6.02849722e-01
7.96320975e-01 -1.27748949e-02 1.71031296e-01 4.46123272e-01
-5.25304019e-01 -4.16736677e-02 5.21842837e-01 4.48465906e-02
1.00049388e+00 -5.34628257e-02 6.83593452e-01 -1.39424339e-01
6.44601405e-01 -1.89505350e-02 4.67526495e-01 8.48041952e-01
-4.08024430e-01 6.36596382e-01 3.48903894e-01 -6.43249214e-01
-8.75772953e-01 -7.57556379e-01 1.42794490e-01 1.35708928e+00
5.08248210e-01 -1.58189282e-01 -8.24940801e-01 -8.97177160e-01
-2.92744134e-02 6.74010634e-01 -8.46520066e-01 -3.70568216e-01
-6.28379881e-01 -9.24397051e-01 1.95906475e-01 2.02471465e-01
5.53086758e-01 -1.29118383e+00 -1.12037563e+00 2.14546725e-01
-3.91874239e-02 -6.35124266e-01 -1.15043962e+00 2.12237462e-01
-9.64812756e-01 -1.09871578e+00 -9.57821429e-01 -7.86384523e-01
1.17157221e+00 4.85032469e-01 5.89294434e-01 1.23281382e-01
-2.69265503e-01 1.74303174e-01 -6.01429582e-01 -1.71263203e-01
-5.14869928e-01 2.43472368e-01 -1.33017272e-01 2.30643556e-01
-1.32049143e-01 -1.57303616e-01 -6.32125497e-01 6.69585109e-01
-1.15694356e+00 4.37263995e-02 8.45267177e-01 8.01849246e-01
6.71130598e-01 5.90317905e-01 4.54923362e-01 -8.60815942e-01
8.93519700e-01 -6.07185215e-02 -6.81730807e-01 4.99050885e-01
-1.06114006e+00 3.33542973e-01 6.82303250e-01 -8.12338710e-01
-9.27550554e-01 4.50174719e-01 2.39030078e-01 -3.63033324e-01
2.11519912e-01 2.07885474e-01 1.65711328e-01 -4.29060310e-01
8.63825381e-01 2.40153074e-01 -7.50173181e-02 -1.77347183e-01
3.56689870e-01 3.82395774e-01 3.11529726e-01 -1.97194636e-01
5.61435759e-01 8.89791101e-02 2.78588161e-02 -6.59505785e-01
-4.78520781e-01 -4.74377245e-01 -6.60840452e-01 -3.90098542e-01
7.98113167e-01 -2.91607469e-01 -6.95100605e-01 3.82030189e-01
-9.06335950e-01 -3.88710648e-01 -4.92237836e-01 3.50329161e-01
-6.70876324e-01 5.76274991e-01 -1.82082191e-01 -6.43310785e-01
-1.20516278e-01 -1.35499763e+00 6.47544384e-01 4.26934391e-01
-1.45858496e-01 -7.71366537e-01 1.07124016e-01 1.23953328e-01
3.04768145e-01 3.01806033e-02 7.91600287e-01 -2.85746515e-01
-4.01062429e-01 -2.86187101e-02 -1.59785256e-01 8.56360197e-02
3.40587616e-01 -1.63446087e-02 -4.14225817e-01 -4.74044383e-01
-2.07457587e-01 -5.72359562e-02 7.55120814e-01 4.35509413e-01
1.23234689e+00 -1.93298370e-01 -3.81556571e-01 3.61898601e-01
1.56931937e+00 7.56989419e-01 7.84092784e-01 7.50423431e-01
4.50785249e-01 4.10381317e-01 9.03197527e-01 3.19546074e-01
-5.30034788e-02 1.11031079e+00 3.59124631e-01 -2.05317676e-01
-1.68494537e-01 -1.74364254e-01 2.10582450e-01 3.45189333e-01
-6.74736425e-02 -4.31683481e-01 -6.32408023e-01 5.45995533e-01
-2.15967727e+00 -8.98472011e-01 2.34573126e-01 2.52338839e+00
6.86825216e-01 2.06616700e-01 2.88312465e-01 1.42514795e-01
9.22599852e-01 -7.86018209e-04 -6.27644598e-01 -6.29511774e-01
-2.67067109e-03 -7.81854466e-02 7.96079040e-01 6.21678412e-01
-1.16631782e+00 9.13617015e-01 6.73862076e+00 8.93717289e-01
-1.14133358e+00 -1.92559555e-01 5.73127627e-01 -5.92021644e-02
5.37794363e-03 -1.14159182e-01 -5.16614735e-01 6.04622483e-01
8.13891888e-01 -3.71410966e-01 6.85291171e-01 5.70842385e-01
4.29036140e-01 -5.33234000e-01 -6.65166259e-01 8.16958606e-01
2.36193821e-01 -1.00815451e+00 6.27904013e-02 -7.39734471e-02
8.60896647e-01 -4.97473985e-01 5.40728495e-02 -2.04645637e-02
1.44191608e-01 -6.58146083e-01 7.21949399e-01 3.79952580e-01
4.16652709e-01 -9.26345170e-01 5.02147913e-01 1.42755747e-01
-1.10185134e+00 -4.02561188e-01 -2.22823516e-01 2.22341329e-01
3.19871195e-02 3.73088688e-01 -8.24580491e-01 3.60173762e-01
5.96872330e-01 1.73197597e-01 -7.86968648e-01 1.29195476e+00
-1.08088344e-01 9.76103097e-02 9.01138932e-02 -1.17251098e-01
8.88264030e-02 -3.67556140e-02 6.27944589e-01 8.89143169e-01
2.96938807e-01 4.18221988e-02 4.20838416e-01 2.34484136e-01
1.55835077e-01 4.49318022e-01 -3.97527337e-01 1.97632074e-01
3.38359535e-01 1.08876634e+00 -1.24586558e+00 -1.51497737e-01
-1.75638005e-01 1.46874738e+00 3.97766113e-01 3.09784055e-01
-8.81787181e-01 -6.90578997e-01 6.72921240e-02 2.06429094e-01
5.06009281e-01 3.59581155e-03 -2.75859386e-01 -7.31479287e-01
-7.09078014e-02 -8.46624076e-01 6.35736823e-01 -6.62186205e-01
-9.29021657e-01 8.05878758e-01 1.84613109e-01 -1.20005035e+00
-1.63191825e-01 -1.77747652e-01 -2.71414638e-01 7.24840879e-01
-1.23438203e+00 -6.51365697e-01 -1.56946242e-01 3.70416343e-01
8.43083441e-01 -2.32403055e-01 4.89906371e-01 2.70109683e-01
-7.02272534e-01 6.90064430e-01 4.19160128e-02 -4.16482240e-01
8.23069990e-01 -1.23295176e+00 1.43829808e-01 1.01676440e+00
2.03082189e-01 2.07337379e-01 9.51357484e-01 -5.70042849e-01
-1.02527487e+00 -8.36219728e-01 7.35916495e-01 -1.43244434e-02
4.14193958e-01 1.66892588e-01 -7.56216109e-01 2.52572179e-01
1.72606722e-01 -1.35510147e-01 2.26715684e-01 -3.42452794e-01
1.49272695e-01 -2.77195573e-01 -1.25028419e+00 1.11439586e+00
8.55731726e-01 1.68439094e-03 -2.93368816e-01 2.30906874e-01
6.54067814e-01 -5.64814031e-01 -5.31771541e-01 1.54250532e-01
4.60369796e-01 -9.20818925e-01 7.55124688e-01 -2.84868628e-01
1.20394737e-01 -4.97960746e-01 1.67681724e-01 -1.57243562e+00
-6.88718021e-01 -7.15860903e-01 4.27825190e-02 6.89704001e-01
5.68483353e-01 -6.13917291e-01 7.85227656e-01 5.00106394e-01
1.23646945e-01 -8.86988699e-01 -9.12612677e-01 -8.52781892e-01
-3.97564262e-01 9.44443047e-02 2.62295902e-01 5.55195451e-01
-1.26944080e-01 2.36937568e-01 -3.92603159e-01 9.81689394e-02
3.64825428e-01 3.43754888e-02 4.94075775e-01 -6.59498394e-01
-4.33013082e-01 -4.94276494e-01 -4.82874632e-01 -8.88697207e-01
-3.05259656e-02 -5.58550358e-01 3.00692677e-01 -1.30400932e+00
2.38354206e-01 -3.37880671e-01 -4.23268646e-01 5.79456866e-01
-5.00881433e-01 4.25494164e-01 4.10033196e-01 2.86627322e-01
-4.53896612e-01 3.88382614e-01 1.45869100e+00 -1.99611858e-01
-8.07830334e-01 2.48926640e-01 -8.18480372e-01 3.68506283e-01
8.61068726e-01 -5.63009262e-01 -7.44019985e-01 -2.61036664e-01
5.30086681e-02 5.98078258e-02 -7.66564510e-04 -9.86748159e-01
3.11022550e-01 -4.59749371e-01 2.04022199e-01 -1.48163021e-01
4.40996364e-02 -9.90857899e-01 9.70566720e-02 7.11494565e-01
-5.83362818e-01 5.36885619e-01 3.28160703e-01 5.49110234e-01
2.53317114e-02 -6.58582687e-01 1.10713542e+00 -7.37003330e-03
-1.14547431e+00 -6.20919932e-03 -5.70132196e-01 -3.17278236e-01
1.64150047e+00 -5.11266768e-01 1.39667213e-01 -1.99020877e-01
-6.50383055e-01 2.19120085e-02 4.82737362e-01 4.96611595e-01
6.12641752e-01 -1.10030305e+00 -4.44769770e-01 1.07465662e-01
7.25129619e-02 -5.27768016e-01 4.51595694e-01 6.89876735e-01
-6.38805747e-01 2.27614731e-01 -3.93339306e-01 -3.21650803e-01
-1.59809113e+00 1.05975342e+00 4.09127653e-01 -3.76858562e-01
-6.28723860e-01 6.12436533e-01 -1.54946027e-02 4.89449650e-02
2.27702871e-01 -1.68808326e-02 -3.00552249e-01 -8.33277777e-02
4.90426719e-01 5.42660415e-01 1.39110103e-01 -4.69005466e-01
-2.56164700e-01 6.91795528e-01 -4.55496460e-01 -2.82682836e-01
8.73469293e-01 -4.29174215e-01 5.36023006e-02 1.73936695e-01
9.40382063e-01 -2.89574545e-02 -1.30773568e+00 5.93906678e-02
-9.78225097e-02 -8.73895705e-01 2.78479725e-01 -8.51482809e-01
-1.20918703e+00 2.45461658e-01 1.02070296e+00 3.58783901e-01
1.69988775e+00 -3.46823603e-01 6.02657497e-01 2.82460988e-01
3.04543853e-01 -1.23398149e+00 3.21025550e-01 9.01140943e-02
8.93975317e-01 -1.05074608e+00 2.23304480e-01 -3.94791842e-01
-9.43692446e-01 1.04929793e+00 8.89265120e-01 9.27009583e-02
3.80273193e-01 -4.39986698e-02 1.86280653e-01 -2.01719608e-02
-5.00193954e-01 -1.45057946e-01 2.56221563e-01 5.44199944e-01
7.23658055e-02 -1.04973793e-01 -7.09093273e-01 -3.57107848e-01
1.87703297e-01 2.33088844e-02 6.49784267e-01 1.02209949e+00
-6.24368250e-01 -1.29640341e+00 -5.32608688e-01 4.06328470e-01
-3.23036224e-01 1.99117988e-01 -1.89112514e-01 6.59248173e-01
2.27867514e-01 9.62425828e-01 -3.34720537e-02 -4.24490690e-01
6.36753678e-01 -3.17033261e-01 5.83631456e-01 -3.45841736e-01
-6.61041260e-01 -1.39563754e-01 -3.37237805e-01 -5.42243063e-01
-4.23406094e-01 -4.93046284e-01 -1.06688535e+00 -2.15849370e-01
-3.63269866e-01 2.06794053e-01 4.79811341e-01 7.02940226e-01
4.06037152e-01 6.18871272e-01 9.44842517e-01 -7.68786907e-01
-5.36055207e-01 -4.95599210e-01 -5.96234620e-01 3.95587295e-01
1.22072875e-01 -6.32674396e-01 -1.92245185e-01 2.14525238e-01] | [11.161088943481445, -1.0390865802764893] |
804f7b10-08fa-40e8-b13b-0ce9462b2fa1 | findings-on-conversation-disentanglement | 2112.05346 | null | https://arxiv.org/abs/2112.05346v1 | https://arxiv.org/pdf/2112.05346v1.pdf | Findings on Conversation Disentanglement | Conversation disentanglement, the task to identify separate threads in conversations, is an important pre-processing step in multi-party conversational NLP applications such as conversational question answering and conversation summarization. Framing it as a utterance-to-utterance classification problem -- i.e. given an utterance of interest (UOI), find which past utterance it replies to -- we explore a number of transformer-based models and found that BERT in combination with handcrafted features remains a strong baseline. We then build a multi-task learning model that jointly learns utterance-to-utterance and utterance-to-thread classification. Observing that the ground truth label (past utterance) is in the top candidates when our model makes an error, we experiment with using bipartite graphs as a post-processing step to learn how to best match a set of UOIs to past utterances. Experiments on the Ubuntu IRC dataset show that this approach has the potential to outperform the conventional greedy approach of simply selecting the highest probability candidate for each UOI independently, indicating a promising future research direction. | ['Jianzhong Qi', 'Jey Han Lau', 'Rongxin Zhu'] | 2021-12-10 | null | https://aclanthology.org/2021.alta-1.1 | https://aclanthology.org/2021.alta-1.1.pdf | alta-2021-12 | ['conversation-disentanglement'] | ['natural-language-processing'] | [ 6.62942708e-01 5.63197374e-01 -3.73915106e-01 -6.58770382e-01
-1.43669999e+00 -6.78226709e-01 9.43359375e-01 2.70973623e-01
-1.04674073e-02 7.88289368e-01 1.08027005e+00 -4.76384014e-01
5.03364392e-02 -3.92246753e-01 -5.10475338e-01 -6.27956986e-01
-1.81695029e-01 1.02236104e+00 1.96979917e-03 -5.42633891e-01
3.15677643e-01 -1.41216323e-01 -1.23909426e+00 1.00965595e+00
4.15345490e-01 6.39090300e-01 -1.97060257e-01 1.15206993e+00
-2.65169114e-01 1.36996377e+00 -7.06646025e-01 -5.51472366e-01
-1.86752051e-01 -7.31459320e-01 -1.70495987e+00 4.47472334e-01
4.64250028e-01 -1.71901599e-01 -2.92956799e-01 6.59763098e-01
3.50596786e-01 3.18234980e-01 5.38015366e-01 -1.32536077e+00
8.53083655e-02 1.01370895e+00 -3.09656292e-01 5.54144025e-01
1.03497052e+00 -1.96234420e-01 1.81946015e+00 -4.26713169e-01
6.28306746e-01 1.56664574e+00 4.73466516e-01 3.60947996e-01
-1.35765350e+00 -2.22330585e-01 1.79201305e-01 1.47267893e-01
-6.87044859e-01 -7.48951972e-01 8.16141486e-01 -3.29452991e-01
1.23556018e+00 6.78633928e-01 1.45319343e-01 1.20566046e+00
2.30030775e-01 1.23010242e+00 9.35779691e-01 -5.37432373e-01
-6.08313978e-02 -6.56580999e-02 6.45529747e-01 5.38913369e-01
-6.47401154e-01 -5.16108036e-01 -8.77684057e-01 -9.23618734e-01
6.47284612e-02 -3.79232287e-01 -4.31765407e-01 2.70392094e-02
-1.29520941e+00 9.46205735e-01 5.50561994e-02 4.69667405e-01
-3.68013591e-01 -1.51827689e-02 6.72630906e-01 8.00175130e-01
5.97738385e-01 5.18596172e-01 -4.31644440e-01 -3.05561721e-01
-7.03860879e-01 6.17015243e-01 1.38869691e+00 6.87105656e-01
9.05651331e-01 -5.08158624e-01 -3.95535439e-01 9.26117063e-01
-9.15350839e-02 -2.16849282e-01 3.19327146e-01 -1.03152251e+00
7.88449347e-01 5.27988255e-01 1.87493056e-01 -7.18422115e-01
-5.55259287e-01 1.03965260e-01 -5.11046290e-01 -6.62317514e-01
5.78572989e-01 -3.61056983e-01 -3.22575539e-01 1.61026239e+00
4.47006911e-01 1.41041860e-01 3.41660082e-01 7.69081295e-01
8.05805743e-01 8.08997929e-01 -2.25991502e-01 -5.06242335e-01
1.72052109e+00 -1.20043206e+00 -7.82666028e-01 -4.63003814e-01
7.70569265e-01 -9.47305918e-01 7.59033620e-01 1.60686195e-01
-8.85133803e-01 -1.74022466e-01 -7.43165433e-01 -1.04141802e-01
2.64547914e-01 -1.61460862e-01 6.69632018e-01 4.56444323e-01
-1.00516438e+00 4.33524251e-01 -4.30458069e-01 -4.28779572e-01
-2.47428656e-01 2.54579723e-01 -3.00835371e-01 -1.27330333e-01
-1.28377497e+00 9.60263312e-01 2.30333462e-01 -1.04964592e-01
-6.50852501e-01 -1.35540247e-01 -8.78945410e-01 7.29951113e-02
7.44522095e-01 -6.96205676e-01 1.86259985e+00 -9.66314793e-01
-1.71420252e+00 7.44516075e-01 -7.46605694e-01 -6.09397829e-01
1.38622880e-01 6.36891052e-02 -1.93115130e-01 1.82119414e-01
1.37773857e-01 2.31237322e-01 8.54441583e-01 -1.04211140e+00
-9.68001366e-01 -3.19422275e-01 4.80778128e-01 5.59072733e-01
1.69223368e-01 3.07638735e-01 -2.29422137e-01 -5.07323816e-03
3.67882609e-01 -1.17927670e+00 -8.24410021e-02 -1.00821304e+00
-5.00251889e-01 -9.03139114e-01 6.16899431e-01 -6.01468146e-01
9.15634274e-01 -1.56605291e+00 4.41868663e-01 1.87353157e-02
3.51020813e-01 -3.07404190e-01 -8.14400688e-02 8.94685030e-01
4.27762493e-02 -4.23313826e-02 6.38250262e-02 -6.06877506e-01
-4.18844298e-02 1.27008125e-01 -3.54281992e-01 4.48335499e-01
2.13481393e-03 7.31973469e-01 -1.14327025e+00 -4.27552223e-01
8.65044910e-03 -4.88300361e-02 -3.45530391e-01 5.07408977e-01
-5.90086699e-01 5.83675921e-01 -5.57020903e-01 2.26032168e-01
6.87502921e-02 -4.38843846e-01 6.59354627e-01 -3.29218693e-02
2.60869145e-01 1.04053462e+00 -7.91155815e-01 1.54605925e+00
-5.38291335e-01 8.80359828e-01 4.59892213e-01 -1.06633139e+00
6.56673014e-01 8.06924105e-01 4.08233553e-01 -3.41644078e-01
2.43562847e-01 1.42742805e-02 1.43590942e-01 -5.92008591e-01
8.58876705e-01 -2.71767974e-01 -4.77013320e-01 9.09919560e-01
2.51357347e-01 -8.70561153e-02 4.30297889e-02 7.05984294e-01
1.34257066e+00 -3.59067202e-01 4.22885925e-01 -2.25126758e-01
7.32278109e-01 6.68263584e-02 3.06729764e-01 8.53204250e-01
-1.24946304e-01 4.56014693e-01 1.10617447e+00 -3.20361167e-01
-8.51218343e-01 -4.75670487e-01 2.49914363e-01 1.70786691e+00
-7.02483132e-02 -4.32413727e-01 -6.47182524e-01 -8.15729618e-01
-2.96434015e-01 9.54516351e-01 -4.64497119e-01 2.39082232e-01
-6.60107493e-01 -6.56322300e-01 3.47096145e-01 6.01195991e-02
7.10083246e-02 -9.27681386e-01 -8.49548876e-02 3.92666698e-01
-1.03011549e+00 -1.31775951e+00 -6.93972886e-01 1.74047828e-01
-5.14058828e-01 -1.01430845e+00 -4.01721656e-01 -7.59375334e-01
2.46189773e-01 4.25218880e-01 1.41166651e+00 -2.65488774e-02
3.69087487e-01 5.06394386e-01 -5.34095705e-01 -5.62002920e-02
-9.58806932e-01 2.60436028e-01 -1.12811320e-01 3.95879567e-01
3.12212080e-01 -4.58160698e-01 -4.08847302e-01 5.07578850e-01
-3.77226621e-01 1.48437530e-01 1.48166686e-01 1.01093602e+00
-8.06607455e-02 -3.59368533e-01 8.79253924e-01 -1.26505411e+00
1.15613616e+00 -5.53689718e-01 -7.20918849e-02 4.60481167e-01
-6.97385371e-02 1.05705813e-01 4.28147793e-01 -1.37689486e-01
-1.15192282e+00 -1.59951627e-01 -2.06062898e-01 1.26746558e-02
-1.91243470e-01 5.48410356e-01 -1.00010015e-01 3.36553037e-01
5.09854019e-01 2.88306594e-01 1.12597220e-01 -2.57447928e-01
4.67521667e-01 1.02715826e+00 5.01451194e-01 -7.43182480e-01
2.49745399e-02 1.47210166e-01 -4.85690892e-01 -1.01379585e+00
-1.20850766e+00 -1.08182240e+00 -3.96206081e-01 -4.13027793e-01
9.63398993e-01 -8.99661958e-01 -1.19847035e+00 1.07007712e-01
-1.47604036e+00 -1.40462071e-01 2.86722898e-01 2.87171662e-01
-6.38202310e-01 5.50335050e-01 -7.09454775e-01 -1.05611253e+00
-5.04445136e-01 -1.07913184e+00 1.14977038e+00 3.51108015e-02
-7.97695279e-01 -9.70842361e-01 1.84282094e-01 1.15736496e+00
3.57668847e-02 2.14966275e-02 1.05258548e+00 -1.39407730e+00
-3.89933020e-01 -3.14836800e-01 2.37504300e-02 -6.29502162e-02
9.68292654e-02 -3.57855141e-01 -1.09744251e+00 -3.99782717e-01
3.09606045e-01 -6.10739470e-01 7.82993197e-01 1.60756871e-01
6.33074760e-01 -6.85466707e-01 -3.73246282e-01 -2.30950460e-01
8.60263288e-01 2.31568944e-02 3.44491571e-01 7.89102260e-03
5.57606936e-01 9.55240726e-01 4.53017026e-01 1.79356471e-01
8.49227309e-01 8.03067744e-01 1.26634806e-01 3.68157476e-01
1.33445561e-01 -1.39930755e-01 3.67215544e-01 1.01938212e+00
2.73182541e-01 -6.91818714e-01 -9.08171654e-01 6.77062511e-01
-2.25006247e+00 -1.12792122e+00 -2.05285624e-01 1.94747770e+00
8.06435466e-01 3.76117349e-01 3.54133755e-01 -2.08866909e-01
6.82433367e-01 6.13131821e-01 -1.37751937e-01 -6.26288593e-01
4.00127172e-02 -1.73031747e-01 1.50002062e-01 8.88061106e-01
-1.04508173e+00 8.09524119e-01 6.10278034e+00 5.17166615e-01
-7.46947527e-01 2.08164081e-01 6.42824650e-01 1.30927727e-01
-3.57814342e-01 2.68503577e-01 -7.69325316e-01 3.83612849e-02
1.26596177e+00 -3.47985893e-01 4.51223254e-01 4.48091537e-01
2.11931601e-01 -3.22784245e-01 -1.55234277e+00 6.39165938e-01
4.04224813e-01 -1.59635842e+00 -9.86481160e-02 -1.64318055e-01
6.13581181e-01 1.88702941e-01 -6.25364959e-01 5.09269416e-01
6.71612799e-01 -7.42865026e-01 4.22638565e-01 9.55052599e-02
1.60442349e-02 -5.04867673e-01 9.07702684e-01 7.85592079e-01
-9.74170029e-01 -2.38455180e-02 -7.01796189e-02 -3.40600312e-01
3.95925015e-01 2.07146317e-01 -1.62796855e+00 7.91253448e-01
1.55975655e-01 3.24342877e-01 5.38004525e-02 5.13591409e-01
9.09052044e-03 9.32630956e-01 -1.89412206e-01 -3.44164282e-01
4.12762344e-01 -9.17262957e-02 8.53405297e-01 1.34374738e+00
-2.00180322e-01 3.30919772e-01 6.06833220e-01 2.17998430e-01
-4.32854354e-01 1.06002077e-01 -5.35436809e-01 1.00933537e-01
4.05297846e-01 1.12631035e+00 -6.18599832e-01 -5.60768306e-01
-5.78059494e-01 8.56775165e-01 5.13220251e-01 1.84495598e-01
-3.53518397e-01 1.04152039e-02 5.28145134e-01 -2.98453093e-01
7.21979737e-02 -2.87184678e-02 -1.09750740e-02 -1.24974132e+00
-9.69310030e-02 -1.28778458e+00 6.52096450e-01 -5.19661963e-01
-1.22861791e+00 6.93676710e-01 -3.64980176e-02 -9.16564524e-01
-7.87515044e-01 -3.04933880e-02 -1.07890975e+00 8.37010503e-01
-1.10793507e+00 -1.09804976e+00 1.97577298e-01 3.11215580e-01
1.13348401e+00 6.97486624e-02 8.23398888e-01 -1.50501832e-01
-6.00744486e-01 4.30745453e-01 4.57633324e-02 1.95213661e-01
5.97856104e-01 -1.28804266e+00 4.86351758e-01 5.57224989e-01
4.57613558e-01 3.57826173e-01 1.19737089e+00 -3.66106242e-01
-1.43360174e+00 -6.96400523e-01 1.48498106e+00 -5.38701952e-01
9.06448662e-01 -5.20895064e-01 -1.03261864e+00 1.08710766e+00
6.93777025e-01 -7.42021918e-01 8.29954565e-01 6.62125707e-01
-1.14011541e-01 1.91438869e-01 -6.91324472e-01 5.02855301e-01
6.04655206e-01 -8.91597569e-01 -1.08077884e+00 8.47655952e-01
8.22747886e-01 -6.19986773e-01 -7.87784815e-01 -5.67930229e-02
3.05499852e-01 -9.42724288e-01 6.35912001e-01 -9.04874623e-01
3.72102231e-01 2.22375914e-01 -1.22882903e-01 -1.31915975e+00
-1.83194818e-03 -1.25773048e+00 -3.26978080e-02 1.34219801e+00
7.22821534e-01 -6.15541637e-01 7.92725921e-01 8.34545910e-01
-1.38832808e-01 -5.96197307e-01 -1.13414574e+00 -2.72388995e-01
-2.84117788e-01 -3.07824135e-01 4.48490441e-01 8.48857045e-01
7.33506262e-01 1.36360788e+00 -7.22329378e-01 1.14329785e-01
3.93469185e-01 4.75295275e-01 1.19611347e+00 -9.63685155e-01
-3.90395701e-01 -3.31518978e-01 7.32273236e-02 -1.62655711e+00
5.83884180e-01 -8.88286293e-01 3.43423575e-01 -1.60621524e+00
2.70136535e-01 2.09859898e-03 1.38297290e-01 2.73623794e-01
-3.77152532e-01 -4.16188836e-01 3.69048230e-02 4.13009286e-01
-8.60102355e-01 4.26662058e-01 1.05597472e+00 -3.45619887e-01
-2.52600700e-01 6.31914437e-01 -6.42850280e-01 5.87844074e-01
5.49581587e-01 -5.94378352e-01 -2.57559210e-01 -6.09379858e-02
-2.17180066e-02 1.18942893e+00 -6.63848966e-02 -2.25198641e-01
4.42812979e-01 -4.92898636e-02 -4.44201618e-01 -6.21770263e-01
5.87345660e-01 -2.92016327e-01 -2.69577414e-01 1.24644548e-01
-8.81277263e-01 -1.12418503e-01 -1.09150477e-01 8.63727391e-01
-3.52372378e-01 -3.48950088e-01 1.97965533e-01 -1.38210520e-01
-5.13543665e-01 -1.44814923e-01 -7.71602511e-01 2.41667002e-01
6.29042149e-01 1.12992980e-01 -4.52696472e-01 -1.06574857e+00
-7.17120886e-01 6.70387447e-01 -1.00594528e-01 5.50028682e-01
1.37919098e-01 -7.35732198e-01 -1.12862945e+00 -3.29533428e-01
1.34501025e-01 -1.96071267e-01 1.84477657e-01 8.43088925e-01
6.70338050e-02 6.50717795e-01 2.85715163e-01 -7.79254138e-01
-1.60960674e+00 1.04913108e-01 2.12967381e-01 -7.72945523e-01
-5.12035668e-01 8.98816168e-01 -1.06930397e-01 -5.17763555e-01
2.04544410e-01 -1.56377435e-01 -3.66954565e-01 3.76702100e-01
4.35762703e-01 1.85555398e-01 2.69447982e-01 -7.91550219e-01
-3.63214135e-01 -1.47470698e-01 -4.18062419e-01 -2.92939126e-01
1.02687132e+00 -4.13350224e-01 -3.64812970e-01 6.56292200e-01
1.47229302e+00 -9.06249657e-02 -7.65644252e-01 -6.85084999e-01
4.10543710e-01 -3.77659947e-01 -1.01537891e-01 -4.75872189e-01
-2.98728466e-01 4.62689728e-01 -3.04430604e-01 7.64455318e-01
4.93594408e-01 4.70725596e-01 8.11432719e-01 7.46621907e-01
3.30597758e-01 -7.50868380e-01 1.31080151e-01 8.15553725e-01
9.73480940e-01 -1.40305793e+00 3.17685492e-02 -3.34866196e-01
-9.75588143e-01 1.13275433e+00 3.58359814e-01 1.49551690e-01
3.55195493e-01 -1.07771307e-01 2.03814805e-01 -4.09106016e-01
-1.40225053e+00 -1.19353339e-01 3.13239992e-01 3.39827202e-02
6.45878255e-01 1.40805066e-01 -3.38298202e-01 1.58907756e-01
-4.43186283e-01 -4.53654379e-01 7.71734953e-01 8.37340295e-01
-5.11746526e-01 -1.05602002e+00 -2.78019667e-01 7.12404430e-01
-4.89660680e-01 6.94686174e-02 -7.39676833e-01 2.95668244e-01
-7.57103801e-01 1.64355850e+00 1.33981362e-01 -4.60151196e-01
2.16870010e-01 4.87841159e-01 1.96492136e-01 -8.23180616e-01
-1.13492346e+00 1.38672590e-01 1.04839063e+00 -3.74117762e-01
-6.24221921e-01 -8.91468346e-01 -9.38080728e-01 -3.35570544e-01
-6.95539713e-01 6.78332746e-01 4.09266323e-01 1.41951525e+00
2.00674459e-01 4.53000933e-01 9.21572983e-01 -8.97817791e-01
-7.20353425e-01 -1.08662474e+00 -1.09398521e-01 4.66887087e-01
7.26976991e-01 -2.59466887e-01 -4.19430435e-01 -5.95489778e-02] | [12.61271858215332, 7.80245304107666] |
ed1d76d1-5373-40a5-9f4b-29f7e5bf42d1 | a-lexical-simplification-tool-for-promoting | null | null | https://aclanthology.org/2020.readi-1.11 | https://aclanthology.org/2020.readi-1.11.pdf | A Lexical Simplification Tool for Promoting Health Literacy | This paper presents MedSimples, an authoring tool that combines Natural Language Processing, Corpus Linguistics and Terminology to help writers to convert health-related information into a more accessible version for people with low literacy skills. MedSimples applies parsing methods associated with lexical resources to automatically evaluate a text and present simplification suggestions that are more suitable for the target audience. Using the suggestions provided by the tool, the author can adapt the original text and make it more accessible. The focus of MedSimples lies on texts for special purposes, so that it not only deals with general vocabulary, but also with specialized terms. The tool is currently under development, but an online working prototype exists and can be tested freely. An assessment of MedSimples was carried out aiming at evaluating its current performance with some promising results, especially for informing the future developments that are planned for the tool. | ["Maria Jos{\\'e} Bocorny Finatto", 'Liana Braga Paraguassu', 'Laura Berwanger', 'Gabriel Ponomarenko', 'Leonardo Zilio', 'Luis Antonio Leiva Hercules'] | 2020-05-01 | null | null | null | lrec-2020-5 | ['lexical-simplification'] | ['natural-language-processing'] | [-1.84559748e-02 6.20508492e-01 -3.64399195e-01 -2.54303545e-01
-3.65174949e-01 -2.91662604e-01 4.89285618e-01 1.13099122e+00
-8.49023283e-01 5.98774135e-01 7.02849627e-01 -6.75217867e-01
-2.01843023e-01 -6.35602772e-01 1.40405193e-01 1.35557381e-02
3.34855258e-01 8.48983169e-01 2.63140768e-01 -5.57414532e-01
4.50842381e-01 4.06860530e-01 -1.70926368e+00 3.77265453e-01
1.21077764e+00 2.09051371e-02 7.68685818e-01 5.45747221e-01
-8.28398466e-01 4.45244670e-01 -7.73827970e-01 -4.93721128e-01
-2.65112787e-01 -4.14091915e-01 -9.97935176e-01 -2.38071054e-01
2.00011939e-01 -1.99122190e-01 5.25377095e-01 9.60341632e-01
5.79812169e-01 -1.49691999e-01 3.73357892e-01 -2.63443202e-01
-2.94552863e-01 8.92049968e-01 2.60604143e-01 2.68991023e-01
9.68258202e-01 -2.30500922e-01 5.01133919e-01 -5.04934072e-01
9.74374056e-01 1.47091591e+00 5.42385936e-01 6.61163568e-01
-8.39764059e-01 -4.19426054e-01 -1.82204861e-02 8.54356512e-02
-1.21879077e+00 -4.79232222e-01 2.85406321e-01 -4.40441489e-01
1.41383529e+00 5.81183136e-01 8.52211416e-01 8.71770620e-01
3.83814335e-01 2.87415683e-01 1.14495838e+00 -1.26860869e+00
1.11004323e-01 7.18884885e-01 4.09180343e-01 6.62470937e-01
7.12917507e-01 -4.47435856e-01 -3.85342151e-01 8.48369673e-02
3.14489216e-01 -5.15554905e-01 -1.73436731e-01 9.68255624e-02
-9.58863020e-01 8.17218781e-01 -3.54330212e-01 1.13309848e+00
-4.16968346e-01 -7.67877042e-01 7.13016748e-01 2.41447821e-01
6.06713533e-01 6.94588780e-01 -4.29582566e-01 -4.90457207e-01
-1.01074255e+00 5.38626015e-01 1.24372375e+00 8.49340796e-01
1.57580376e-01 -5.96874118e-01 -1.53625637e-01 7.84539282e-01
6.19428873e-01 3.74176949e-01 7.25304425e-01 -5.35033643e-01
3.98410022e-01 8.79996538e-01 -1.61067247e-01 -8.43042314e-01
-8.91658962e-01 -1.76761732e-01 -4.34476614e-01 -3.55582014e-02
5.49430013e-01 -1.75759241e-01 -4.14473176e-01 1.21987689e+00
1.56029776e-01 -7.49348581e-01 1.29612520e-01 4.27159995e-01
1.59705579e+00 5.15429258e-01 4.70952928e-01 -5.71788490e-01
1.83822417e+00 -4.62574542e-01 -1.27235150e+00 -1.60300165e-01
1.04528439e+00 -1.15735090e+00 1.10147750e+00 5.36002100e-01
-1.27967584e+00 -2.70223051e-01 -9.98993278e-01 -4.43514764e-01
-8.76912117e-01 1.72884837e-01 3.20104897e-01 1.17135596e+00
-1.06552696e+00 5.33575773e-01 -7.12729335e-01 -1.03130257e+00
-2.42696449e-01 4.79709245e-02 -1.86073124e-01 1.12088688e-01
-1.56830847e+00 1.42473423e+00 7.32387006e-01 -3.42183799e-01
4.54992205e-01 -5.67923129e-01 -1.13417721e+00 -2.20011204e-01
2.59121746e-01 -8.25201392e-01 1.32798684e+00 -6.64888144e-01
-1.66628528e+00 1.18787634e+00 -1.47281334e-01 -3.92129153e-01
6.21658266e-01 -2.60879733e-02 -6.55458033e-01 2.72093356e-01
6.35947764e-01 2.73092419e-01 2.19476029e-01 -3.72665584e-01
-8.15009475e-01 -2.99422204e-01 1.26070844e-03 4.99476373e-01
-3.35558593e-01 7.83174813e-01 -7.13485599e-01 -9.02800798e-01
-4.45007503e-01 -2.68773228e-01 -6.27360716e-02 -1.09861866e-01
-2.51452625e-01 -5.67399800e-01 3.16635938e-03 -1.17959213e+00
1.81734025e+00 -1.68234265e+00 2.29125544e-02 2.77229190e-01
6.60637766e-02 5.92876613e-01 1.76055312e-01 1.01466715e+00
1.74642816e-01 4.95881468e-01 1.63006708e-01 -5.59189916e-02
-1.08839810e-01 3.74000162e-01 2.77440935e-01 2.08625440e-02
-6.33567721e-02 4.42874134e-01 -1.01965976e+00 -7.84144461e-01
4.02574599e-01 4.45239484e-01 -1.68271199e-01 -2.14361444e-01
-1.12862982e-01 2.45283052e-01 -5.33929765e-01 3.02885175e-01
3.79535705e-01 2.89647162e-01 2.60254055e-01 1.96432665e-01
-6.60420775e-01 5.89216948e-01 -1.19683552e+00 1.57113540e+00
-7.47391820e-01 4.71155018e-01 -8.53985846e-02 -6.75561070e-01
7.97359467e-01 3.22734058e-01 1.01001196e-01 -7.40371466e-01
2.70772964e-01 4.13584501e-01 -7.68835694e-02 -1.16789210e+00
4.86743420e-01 -8.52800086e-02 2.20585376e-01 2.60028720e-01
-4.21162210e-02 -4.57145460e-02 9.56064463e-01 2.03739107e-01
7.09352136e-01 2.84722060e-01 1.02023327e+00 -7.49084055e-01
1.00571561e+00 2.23587036e-01 -1.09002506e-02 5.89951098e-01
2.36207172e-01 -1.34487942e-01 4.60234672e-01 -2.77906328e-01
-9.98593152e-01 -4.19741213e-01 -4.83369410e-01 7.89143562e-01
-3.69439960e-01 -1.04543614e+00 -1.15807498e+00 -3.45495343e-01
-4.90774006e-01 1.26594818e+00 -5.44891298e-01 3.02432299e-01
-3.59489232e-01 -3.28340143e-01 2.46954679e-01 9.35739372e-03
2.53515989e-01 -1.22754049e+00 -1.03550947e+00 4.55368489e-01
-1.97239831e-01 -8.24362099e-01 -1.68143705e-01 -1.57846242e-01
-7.84720659e-01 -1.07496524e+00 -9.39802527e-01 -9.31202769e-01
5.07591665e-01 -2.95371085e-01 1.06844342e+00 3.41260999e-01
-3.13214451e-01 3.63730520e-01 -8.57824385e-01 -9.62417841e-01
-1.16526806e+00 3.99210036e-01 -3.03393811e-01 -9.07389283e-01
6.16334319e-01 1.36309773e-01 -9.03758705e-02 -2.73987204e-01
-1.06319964e+00 2.44666174e-01 5.15193105e-01 5.10208309e-01
2.49267876e-01 -5.94213698e-03 4.86660808e-01 -1.30797088e+00
1.17585361e+00 -1.26485974e-01 -2.69657105e-01 2.59959012e-01
-9.28660989e-01 2.37075627e-01 6.44074559e-01 -1.23331204e-01
-1.32183337e+00 -5.26959538e-01 -9.82436955e-01 8.20470870e-01
-4.58647549e-01 9.58193898e-01 -2.03812778e-01 -2.11264603e-02
9.32428479e-01 -1.72919348e-01 2.25264326e-01 -7.54216313e-01
2.32041255e-01 8.88164043e-01 2.11854190e-01 -1.62633389e-01
2.45412618e-01 -2.53934443e-01 -2.84103572e-01 -1.30935264e+00
-3.66759330e-01 -7.09302425e-01 -6.48492098e-01 -1.21499456e-01
7.20936716e-01 -4.19516265e-01 -2.71087050e-01 6.76015988e-02
-1.03505576e+00 -1.54853716e-01 -4.02501315e-01 5.18940270e-01
-1.61851048e-01 5.48041403e-01 -2.29169577e-01 -6.43949032e-01
-6.58839345e-01 -1.03037298e+00 7.73163080e-01 3.12702388e-01
-8.31910074e-01 -1.40360129e+00 1.29194483e-01 3.65142614e-01
3.14495057e-01 2.10360326e-02 1.26837039e+00 -1.10348141e+00
4.22759742e-01 -4.11494672e-01 2.47300088e-01 1.15457565e-01
5.52152805e-02 4.56203073e-02 -3.11957747e-01 7.81564191e-02
-9.12962332e-02 7.45357424e-02 2.18425512e-01 1.96176708e-01
7.07395256e-01 -5.08975744e-01 -4.70766336e-01 6.54189214e-02
1.29538715e+00 1.74431086e-01 8.06036174e-01 9.14694309e-01
1.26694664e-01 1.02361369e+00 6.84115350e-01 2.58162379e-01
5.28082907e-01 7.91255653e-01 -1.58738181e-01 -8.51716772e-02
-2.55616367e-01 -1.15238819e-02 2.30280742e-01 7.93850601e-01
7.51074636e-03 -3.70337293e-02 -9.38211083e-01 2.82199413e-01
-1.43904459e+00 -7.17725992e-01 -3.97682160e-01 2.28311491e+00
1.01409340e+00 3.67432415e-01 3.77406925e-01 1.76442027e-01
4.20046985e-01 -2.40155742e-01 1.26906499e-01 -1.22617590e+00
6.52683154e-02 7.91245997e-01 3.14261049e-01 7.13611364e-01
-7.59148657e-01 7.85741925e-01 6.44509983e+00 8.55612338e-01
-1.08491075e+00 6.57439008e-02 2.76009709e-01 3.47851038e-01
-2.37618059e-01 -1.94647133e-01 -9.84360516e-01 3.55374396e-01
1.21957386e+00 -7.82649636e-01 7.40276352e-02 4.59112912e-01
7.68407166e-01 -4.01900381e-01 -6.85361385e-01 5.87613463e-01
3.39202642e-01 -1.32262290e+00 7.36553743e-02 -2.63480872e-01
-1.25291809e-01 -3.02976131e-01 -4.59491670e-01 2.73034543e-01
-4.06148851e-01 -7.85492539e-01 6.59851491e-01 6.18910372e-01
6.49133325e-01 -6.85213745e-01 1.23284793e+00 5.74896216e-01
-8.19821060e-01 3.70987296e-01 -1.61941662e-01 -3.91670495e-01
6.26839474e-02 3.09158295e-01 -9.87314105e-01 7.26064801e-01
4.19390917e-01 4.82508808e-01 -1.09967530e+00 1.18448544e+00
-3.92850965e-01 2.55867809e-01 -2.82626450e-01 -7.16020525e-01
2.20087711e-02 -2.04687372e-01 6.66398764e-01 1.70920300e+00
3.48129362e-01 1.30044475e-01 3.28111425e-02 5.13655305e-01
5.59644699e-01 1.52351773e+00 -4.35337782e-01 -7.80340135e-02
3.37066203e-01 1.17784429e+00 -1.00342405e+00 -5.75024068e-01
-5.68567693e-01 7.95443714e-01 -8.48333910e-02 -7.05912560e-02
-8.03238600e-02 -7.62804329e-01 -6.81742951e-02 5.91931224e-01
-1.48368925e-01 1.83047593e-01 -3.94643337e-01 -9.53004658e-01
2.87068803e-02 -1.20442820e+00 4.49402511e-01 -5.92815995e-01
-6.87610745e-01 6.47068441e-01 6.41891599e-01 -6.73865378e-01
-5.46184719e-01 -8.92161667e-01 -4.25967932e-01 1.26521766e+00
-9.45930421e-01 -1.07188237e+00 5.96576259e-02 7.14551806e-02
6.49396598e-01 -5.15930057e-02 1.33409142e+00 2.84258455e-01
-4.93322372e-01 4.76508886e-01 8.45504366e-03 -2.76081026e-01
6.74971223e-01 -1.37660563e+00 2.69811988e-01 8.35655868e-01
-2.66105115e-01 9.29290771e-01 1.20484221e+00 -9.26727653e-01
-6.39561355e-01 -6.20821893e-01 1.95684326e+00 -9.45074782e-02
7.47859240e-01 -1.93390548e-01 -9.23854291e-01 -3.52408621e-03
6.39563322e-01 -1.07063639e+00 9.74565148e-01 -9.87495109e-03
2.90406406e-01 2.22138241e-01 -1.30897081e+00 8.24452400e-01
6.39706194e-01 -9.58421528e-02 -1.19131613e+00 6.47962272e-01
2.62211353e-01 -6.91694140e-01 -1.00726104e+00 4.02709916e-02
5.27795672e-01 -7.39666939e-01 5.47282219e-01 -2.84735769e-01
1.40947238e-01 8.08285177e-02 6.03240669e-01 -1.14321065e+00
3.62753645e-02 -8.31458926e-01 4.62359786e-01 1.18079865e+00
8.06222558e-01 -8.21047127e-01 2.97465742e-01 7.52892017e-01
-1.52111545e-01 -4.93104041e-01 -9.27609622e-01 -4.91296381e-01
2.90165067e-01 -4.62825924e-01 3.99138391e-01 6.71515822e-01
6.75404429e-01 4.38052714e-01 1.42797828e-01 -2.74676263e-01
1.99548736e-01 -6.43356442e-01 4.80237901e-01 -1.47616196e+00
1.22495489e-02 -7.37918198e-01 -2.87820667e-01 -5.00467718e-01
4.41491492e-02 -9.01655197e-01 -3.36230695e-01 -2.08068871e+00
-1.30243942e-01 7.68714249e-02 5.55812180e-01 6.19171083e-01
-1.20995834e-01 -9.34077576e-02 8.83659720e-02 -1.42157555e-01
-9.78750959e-02 -2.33896926e-01 9.37557697e-01 1.13157623e-01
-6.51426494e-01 1.43190339e-01 -1.14613485e+00 7.50835896e-01
8.14825594e-01 -3.88605356e-01 -2.56921589e-01 4.49578315e-02
4.98421758e-01 -3.60551089e-01 -1.56335369e-01 -8.01538527e-01
9.55020562e-02 -7.28411302e-02 1.59657717e-01 -4.75639313e-01
-4.63761777e-01 -7.14972258e-01 2.28039980e-01 8.24035048e-01
-3.81984383e-01 2.51223266e-01 4.49432939e-01 -2.66302228e-01
1.40908033e-01 -1.31123626e+00 4.69529420e-01 -2.47379020e-01
-6.49388492e-01 -2.97934711e-01 -9.46828187e-01 3.31306644e-02
9.50877845e-01 -2.59289116e-01 -9.46874768e-02 -3.28646809e-01
-9.15348351e-01 2.14387581e-01 5.39791763e-01 4.27990198e-01
3.32007766e-01 -7.89379537e-01 -6.30560577e-01 1.23923190e-01
3.91029179e-01 -2.91226923e-01 7.74328411e-02 7.23653018e-01
-1.38906491e+00 8.70807767e-01 -3.00217777e-01 3.27064246e-02
-1.60042500e+00 6.76951051e-01 1.15729652e-01 -5.72683692e-01
-1.01769114e+00 2.27167115e-01 -3.95232111e-01 -1.59081355e-01
3.61229986e-01 -7.48443544e-01 -1.28641427e+00 3.80961925e-01
1.03910577e+00 3.88709188e-01 3.59085262e-01 -7.47754097e-01
-3.47587049e-01 4.37398493e-01 -2.07804829e-01 -2.41165638e-01
1.17708182e+00 -5.01594126e-01 -4.35357392e-01 4.88431782e-01
5.66036046e-01 4.46879417e-01 -8.44426453e-03 -1.19024679e-01
3.53619158e-01 -6.01144806e-02 8.49675313e-02 -1.08910465e+00
-3.28590035e-01 5.55363178e-01 6.74845874e-01 4.45845336e-01
1.19103777e+00 -8.48917961e-02 4.18040633e-01 4.10281926e-01
9.08741653e-02 -1.45255578e+00 -6.45326078e-01 4.62845683e-01
1.05221033e+00 -6.89003408e-01 3.85480791e-01 -5.70547998e-01
-5.45347631e-01 1.79691827e+00 -2.63436865e-02 5.22619903e-01
6.98210180e-01 1.02787651e-01 3.21277320e-01 -3.26858610e-01
-2.37499386e-01 -4.47841853e-01 5.40337443e-01 9.56717968e-01
9.15785074e-01 -1.49281211e-02 -1.80080795e+00 7.11165667e-01
-3.69114548e-01 3.30674976e-01 7.06769884e-01 7.57896364e-01
-4.17188644e-01 -1.83355749e+00 -5.89821398e-01 6.40976191e-01
-8.85700166e-01 -3.43698084e-01 -5.46461821e-01 1.12566912e+00
4.37461376e-01 1.03657663e+00 -2.44525969e-01 2.06534341e-01
7.25891232e-01 1.37939319e-01 4.95366901e-01 -1.16887641e+00
-1.02307177e+00 3.70714039e-01 8.03578913e-01 -2.26422474e-01
-6.35762036e-01 -9.11636710e-01 -1.13716102e+00 -1.99524015e-01
-2.88818013e-02 3.69634926e-01 9.03907657e-01 1.21122301e+00
2.08947230e-02 5.21965265e-01 -2.40663081e-01 -4.01346356e-01
-1.92608491e-01 -1.11602056e+00 -3.60871315e-01 1.11648135e-01
-1.43792167e-01 -4.48658317e-01 1.91238880e-01 6.47789314e-02] | [10.584894180297852, 10.226747512817383] |
a94362af-4568-432f-b78d-9cce58d71d06 | weakly-supervised-joint-whole-slide | 2301.02933 | null | https://arxiv.org/abs/2301.02933v1 | https://arxiv.org/pdf/2301.02933v1.pdf | Weakly Supervised Joint Whole-Slide Segmentation and Classification in Prostate Cancer | The segmentation and automatic identification of histological regions of diagnostic interest offer a valuable aid to pathologists. However, segmentation methods are hampered by the difficulty of obtaining pixel-level annotations, which are tedious and expensive to obtain for Whole-Slide images (WSI). To remedy this, weakly supervised methods have been developed to exploit the annotations directly available at the image level. However, to our knowledge, none of these techniques is adapted to deal with WSIs. In this paper, we propose WholeSIGHT, a weakly-supervised method, to simultaneously segment and classify WSIs of arbitrary shapes and sizes. Formally, WholeSIGHT first constructs a tissue-graph representation of the WSI, where the nodes and edges depict tissue regions and their interactions, respectively. During training, a graph classification head classifies the WSI and produces node-level pseudo labels via post-hoc feature attribution. These pseudo labels are then used to train a node classification head for WSI segmentation. During testing, both heads simultaneously render class prediction and segmentation for an input WSI. We evaluated WholeSIGHT on three public prostate cancer WSI datasets. Our method achieved state-of-the-art weakly-supervised segmentation performance on all datasets while resulting in better or comparable classification with respect to state-of-the-art weakly-supervised WSI classification methods. Additionally, we quantify the generalization capability of our method in terms of segmentation and classification performance, uncertainty estimation, and model calibration. | ['Orcun Goksel', 'Maria Gabrani', 'Behzad Bozorgtabar', 'Kevin Thandiackal', 'Zeineb Ayadi', 'Guillaume Jaume', 'Pushpak Pati'] | 2023-01-07 | null | null | null | null | ['whole-slide-images'] | ['computer-vision'] | [ 7.27887452e-01 6.08357251e-01 -4.95179981e-01 -2.96578646e-01
-1.12483859e+00 -8.49612951e-01 4.85480845e-01 7.89656281e-01
-2.80329525e-01 5.23400187e-01 -3.58676940e-01 -4.59557503e-01
-9.71529037e-02 -6.60187066e-01 -4.63005573e-01 -1.10203898e+00
2.34559109e-03 7.19558537e-01 6.28885090e-01 2.81020045e-01
1.30532086e-02 5.42008698e-01 -9.12294447e-01 1.34309411e-01
9.50047851e-01 1.13166308e+00 1.45547196e-01 6.88774049e-01
-1.97333515e-01 4.62466329e-01 -1.68801904e-01 -3.03646922e-01
5.11371493e-02 -2.27056891e-01 -1.00502074e+00 2.10467756e-01
3.41079623e-01 2.42660061e-01 1.47681432e-02 1.19822991e+00
8.47107023e-02 -4.26140904e-01 8.16053212e-01 -1.11147463e+00
1.30044902e-02 6.37220562e-01 -7.25668728e-01 -2.69300900e-02
-1.42044118e-02 2.22375408e-01 1.06529284e+00 -4.48206991e-01
8.56387496e-01 4.55835342e-01 8.62030387e-01 4.92546171e-01
-1.61761010e+00 -4.13390815e-01 -3.74901891e-02 -3.46381336e-01
-1.52693594e+00 -2.16978326e-01 7.36606717e-01 -3.40618104e-01
5.91453135e-01 5.70329309e-01 6.41243458e-01 5.91386199e-01
3.30205441e-01 1.08356094e+00 1.14841366e+00 -3.61214042e-01
3.36897254e-01 2.63492376e-01 2.95351952e-01 9.21153843e-01
1.32312894e-01 -2.65592784e-01 -1.03528000e-01 -1.78282768e-01
6.56846344e-01 -1.97152585e-01 -1.39564022e-01 -4.58316743e-01
-1.24949300e+00 5.04156828e-01 7.26358891e-01 3.83754551e-01
-6.77316412e-02 -7.19637498e-02 4.10160482e-01 -1.56237930e-01
7.24054158e-01 2.91517168e-01 -1.75837010e-01 2.10753769e-01
-1.28493536e+00 -2.29552761e-01 7.76398242e-01 6.42086625e-01
6.27039790e-01 -6.12447798e-01 -2.49313354e-01 6.92280829e-01
3.63811493e-01 2.48218223e-01 3.19372326e-01 -5.76673269e-01
4.20486033e-02 1.15034378e+00 -3.66608679e-01 -7.28064537e-01
-9.10588145e-01 -7.67621994e-01 -1.06694996e+00 6.69501424e-02
7.80976534e-01 3.79929990e-01 -1.24325979e+00 1.38667953e+00
4.69020277e-01 2.32403710e-01 -7.49400184e-02 6.51187539e-01
7.99305916e-01 1.66145757e-01 2.57750869e-01 -1.46446556e-01
1.32785511e+00 -8.59899521e-01 -3.62285346e-01 -2.88886100e-01
9.52243924e-01 -4.88611877e-01 8.73826087e-01 1.29683748e-01
-9.53852057e-01 -1.18411802e-01 -6.23245597e-01 1.52922362e-01
-3.09664398e-01 4.73451972e-01 6.07821286e-01 6.02940559e-01
-1.06417608e+00 5.03519833e-01 -1.26082695e+00 -4.26822960e-01
9.16243494e-01 6.00065172e-01 -3.78573626e-01 1.91121727e-01
-5.60218573e-01 6.01622462e-01 2.86077827e-01 1.27344280e-01
-8.49441111e-01 -8.76753926e-01 -9.20341313e-01 -4.65055294e-02
5.09649456e-01 -3.54787052e-01 1.09561610e+00 -8.67140710e-01
-1.25817049e+00 1.39799070e+00 -2.31166065e-01 -3.76180142e-01
6.40137076e-01 8.06666911e-01 -4.81179208e-02 2.42930233e-01
1.80609971e-01 7.48825371e-01 5.57183564e-01 -1.40077221e+00
-6.61121428e-01 -6.25388145e-01 -1.97581708e-01 -3.05329170e-02
-1.90771520e-01 -5.95964789e-01 -9.15547132e-01 -5.25627255e-01
4.51241881e-01 -1.17287147e+00 -4.73030806e-01 4.48215127e-01
-9.86679256e-01 -1.58670351e-01 7.91631997e-01 -5.62930107e-01
1.05502534e+00 -2.31565690e+00 1.28104359e-01 7.92727828e-01
4.74317908e-01 2.51629978e-01 -3.66543420e-02 -1.94468856e-01
1.42462775e-01 2.94057757e-01 -6.96819842e-01 -5.66548169e-01
-4.28291261e-02 4.43521857e-01 3.04795265e-01 7.30393410e-01
1.96780801e-01 1.15182185e+00 -1.01002955e+00 -1.01975369e+00
2.95499802e-01 1.23960234e-01 -3.89827222e-01 1.97223142e-01
-1.97735757e-01 5.54252982e-01 -3.81367147e-01 9.96630311e-01
5.11153519e-01 -6.73204899e-01 3.95877868e-01 -3.57810915e-01
3.20898265e-01 -2.09377125e-01 -8.56240451e-01 1.62516820e+00
-2.50462234e-01 3.75948608e-01 3.22724789e-01 -1.20506001e+00
6.54898882e-01 5.93547300e-02 6.36932194e-01 -1.79511920e-01
2.03319862e-01 3.46173495e-01 -5.99119589e-02 -2.94471264e-01
6.38806298e-02 -1.88490361e-01 -1.32582933e-01 8.85685086e-02
9.86044016e-03 -4.27389145e-01 3.28964114e-01 2.69499600e-01
1.24642062e+00 3.86472791e-02 4.44811910e-01 -4.60776031e-01
7.02868462e-01 2.25291252e-01 5.45154333e-01 6.00534618e-01
-2.89652109e-01 6.66014731e-01 7.74254799e-01 5.89148998e-02
-6.36747718e-01 -1.23523521e+00 -4.10081804e-01 7.51725197e-01
2.79468447e-01 -1.07280590e-01 -7.75274038e-01 -1.14220512e+00
5.53486012e-02 4.02313262e-01 -9.62199450e-01 7.99155235e-03
-2.33505949e-01 -8.19525063e-01 6.05008245e-01 6.41351402e-01
2.89863229e-01 -6.78334355e-01 -3.42391074e-01 1.53593421e-01
-4.67914417e-02 -1.10220373e+00 -2.87040144e-01 4.68254179e-01
-9.80032146e-01 -1.33446491e+00 -7.14349389e-01 -9.44566905e-01
1.38237846e+00 -1.78322241e-01 9.63030040e-01 2.79723644e-01
-8.11753333e-01 1.05156988e-01 -2.25815214e-02 -1.35109738e-01
-7.35642254e-01 2.87924945e-01 -5.86450398e-01 4.65525351e-02
8.69323388e-02 -1.60344690e-01 -6.75294816e-01 4.33077425e-01
-1.07576072e+00 1.75713867e-01 7.61548042e-01 9.55072880e-01
1.15974462e+00 9.77732316e-02 9.37089548e-02 -1.39357054e+00
1.25917584e-01 -3.28313470e-01 -6.51273608e-01 4.53940988e-01
-4.84034717e-01 -2.24638628e-04 5.56156337e-01 -1.67596653e-01
-9.11957741e-01 5.12881517e-01 -2.11973310e-01 1.28976896e-01
-2.79133707e-01 8.19079876e-01 1.07837811e-01 -4.41600084e-01
6.03991687e-01 7.13844448e-02 2.80535787e-01 -4.63676872e-03
3.64795700e-02 5.69532633e-01 5.63678443e-01 -3.02046508e-01
6.93797171e-01 6.86939299e-01 2.74369210e-01 -6.44449413e-01
-8.07331085e-01 -6.78218126e-01 -7.61954963e-01 -2.44751900e-01
8.50637615e-01 -4.05335844e-01 -5.26551783e-01 5.81384480e-01
-4.60924923e-01 -7.25875914e-01 -4.90245521e-01 7.53858015e-02
-4.15686309e-01 3.56115043e-01 -7.77644694e-01 -4.81880605e-01
-1.72125444e-01 -1.37389743e+00 1.46998262e+00 2.05030814e-01
-2.41174445e-01 -1.27884412e+00 -7.24770576e-02 5.00263512e-01
4.21092540e-01 5.32640874e-01 1.01658285e+00 -7.00557530e-01
-3.79206181e-01 -5.79871655e-01 -4.17566001e-01 4.68452498e-02
1.73596710e-01 3.05708736e-01 -1.03000069e+00 -2.32782155e-01
-7.24892437e-01 -2.81487495e-01 9.32193339e-01 5.65195262e-01
1.50273776e+00 3.14837322e-02 -1.04069722e+00 6.48345768e-01
1.52940261e+00 -1.20480180e-01 2.34703317e-01 1.89860940e-01
4.99767303e-01 6.45005286e-01 5.44612050e-01 9.21611786e-02
3.00568402e-01 4.26497191e-01 4.95179564e-01 -6.33376002e-01
-1.86748073e-01 -6.00678474e-02 -2.21606016e-01 2.10510209e-01
3.08215357e-02 -2.86856741e-01 -1.29676962e+00 6.50720477e-01
-1.66769373e+00 -2.15355948e-01 -2.15316817e-01 1.95807076e+00
1.04407477e+00 3.48074436e-01 -1.46284744e-01 3.04010183e-01
4.03556883e-01 1.97041053e-02 -7.75096118e-01 1.90354809e-01
2.06776604e-01 1.49015546e-01 6.61000848e-01 4.21730042e-01
-1.30268121e+00 8.67343783e-01 5.54526854e+00 9.92750227e-01
-1.05953562e+00 5.31104058e-02 1.24623966e+00 1.98883638e-01
-2.02484891e-01 -1.56992733e-01 -6.18746877e-01 1.84551522e-01
7.01024473e-01 2.55113780e-01 7.29923993e-02 7.17333794e-01
-8.01562592e-02 -3.58681887e-01 -1.21174049e+00 5.27475893e-01
-1.88264504e-01 -1.48592269e+00 -2.89336026e-01 3.23795974e-01
8.22276950e-01 -5.66844121e-02 -6.39784113e-02 1.39272764e-01
2.26023808e-01 -1.03861690e+00 2.18770057e-01 3.37234974e-01
9.50401723e-01 -5.47678649e-01 1.15999484e+00 3.64939511e-01
-1.19490314e+00 1.70850500e-01 -3.44284773e-02 6.17253184e-01
-1.67827085e-01 7.99804151e-01 -1.15714407e+00 5.04514754e-01
2.81321824e-01 5.53888857e-01 -9.26245391e-01 9.59493041e-01
-1.56537339e-01 8.14114571e-01 -4.37162191e-01 2.38243472e-02
2.47665733e-01 8.66147280e-02 2.10500285e-01 1.37873268e+00
3.94426472e-02 3.19351591e-02 4.22074735e-01 8.17932546e-01
-1.06868267e-01 -6.86553679e-03 -1.37429507e-02 -9.05167684e-02
2.15965524e-01 1.75853622e+00 -1.53244853e+00 -2.86741167e-01
-1.54590040e-01 8.96160841e-01 3.18941712e-01 3.22231770e-01
-6.33197784e-01 3.60537991e-02 6.15070201e-02 2.47227415e-01
-1.73291653e-01 5.21863624e-02 -5.41922629e-01 -8.45691502e-01
-1.75318673e-01 -3.71235549e-01 6.35913014e-01 -2.16118619e-01
-1.35439265e+00 4.59229439e-01 -1.31458625e-01 -1.18615067e+00
2.38680374e-02 -7.24514961e-01 -6.33634627e-01 4.00239110e-01
-1.84502220e+00 -1.54028440e+00 -6.76736355e-01 2.73770094e-01
2.88917750e-01 3.12312216e-01 1.01507711e+00 -1.33063793e-01
-6.23607457e-01 6.80982649e-01 5.22005698e-03 2.83736110e-01
4.98252720e-01 -1.75259566e+00 -4.55439799e-02 4.67109770e-01
-1.32353604e-01 2.87390947e-01 3.72839451e-01 -6.06779873e-01
-1.29458582e+00 -1.31191385e+00 5.71399748e-01 -1.07679650e-01
8.29504967e-01 -2.97721803e-01 -8.47523391e-01 5.60045719e-01
-2.28329659e-01 7.33998895e-01 1.00012624e+00 -1.47119895e-01
1.66584343e-01 -7.32826665e-02 -1.47381473e+00 4.90928710e-01
8.18289638e-01 -3.53828132e-01 -1.22785777e-01 5.76471269e-01
1.44905314e-01 -7.90777862e-01 -1.11600244e+00 4.75762844e-01
3.54248613e-01 -6.20927334e-01 9.07433569e-01 -9.11890529e-03
2.89880484e-01 -3.68487418e-01 2.29556218e-01 -1.18170011e+00
-1.46217853e-01 -5.00330143e-02 2.62594312e-01 1.24244606e+00
8.80366743e-01 -6.30918920e-01 1.29216254e+00 7.96392083e-01
-3.07731211e-01 -9.90554988e-01 -9.82556105e-01 -6.17277503e-01
-9.17041078e-02 -2.59446621e-01 2.03188255e-01 7.95880497e-01
1.57207757e-01 -2.10164309e-01 3.95696580e-01 3.37458760e-01
8.89057279e-01 2.22396493e-01 4.53486830e-01 -1.32024169e+00
-2.69298822e-01 -7.37730622e-01 -6.51070952e-01 -5.01463830e-01
1.63792267e-01 -1.39638066e+00 2.61273980e-01 -1.68630660e+00
4.32851315e-01 -8.02492023e-01 -3.11418444e-01 8.23518455e-01
-3.18363905e-01 7.07422912e-01 -2.42825180e-01 3.36539775e-01
-5.41790128e-01 6.03076965e-02 1.33506274e+00 -4.69714344e-01
-1.34873718e-01 4.33605686e-02 -4.48497355e-01 8.98142457e-01
7.02089489e-01 -2.66522020e-01 -2.43730187e-01 9.08554122e-02
-8.51164237e-02 1.89075932e-01 5.49355686e-01 -9.48171794e-01
3.22717875e-01 -6.00683503e-02 5.60872793e-01 -5.75806260e-01
3.61740068e-02 -9.06823516e-01 2.62432769e-02 5.28093874e-01
-5.47699809e-01 -8.74717116e-01 2.65318066e-01 6.09465301e-01
-2.00455412e-01 -3.42537224e-01 8.74706388e-01 -1.25824288e-01
-3.43976200e-01 4.13143009e-01 -1.95650309e-01 -1.02143608e-01
1.36020124e+00 -2.99553722e-01 -2.28732258e-01 1.34853199e-01
-1.03108692e+00 4.39695418e-01 5.40023208e-01 -2.16316327e-01
4.07698750e-01 -7.72498906e-01 -6.36545539e-01 2.45814577e-01
4.41934317e-01 5.56246996e-01 2.41637096e-01 1.11242139e+00
-5.54165840e-01 2.03860268e-01 2.32134789e-01 -1.12485468e+00
-1.40916598e+00 2.64333248e-01 3.31355184e-01 -9.08042729e-01
-3.96801054e-01 1.14189434e+00 2.64846742e-01 -6.65913284e-01
2.71483719e-01 -7.26667881e-01 -1.01244152e-01 -1.27083957e-01
1.05441339e-01 -8.16424042e-02 1.83354676e-01 -4.20596749e-01
-4.74034995e-01 5.25717556e-01 -1.35527834e-01 2.43938237e-01
1.14605856e+00 6.18361216e-03 -3.34175736e-01 2.77639210e-01
1.14055693e+00 -2.23429784e-01 -1.30609322e+00 -3.89003217e-01
2.84454763e-01 -2.59453863e-01 2.89958090e-01 -1.09128022e+00
-1.24520743e+00 6.77940547e-01 6.57970726e-01 3.11226338e-01
1.14913297e+00 3.64092797e-01 5.43714464e-01 -1.12717293e-01
2.35026881e-01 -8.99776340e-01 -2.22151473e-01 -1.47347316e-01
3.62143368e-01 -1.46900749e+00 -3.92062776e-02 -8.97822797e-01
-5.07023275e-01 1.11876559e+00 4.42215592e-01 6.15253262e-02
5.96544087e-01 5.51139653e-01 1.61374584e-02 -3.97337258e-01
-3.69024754e-01 -1.53639227e-01 5.35667837e-01 4.14013058e-01
3.70297462e-01 3.38070959e-01 -1.53605342e-01 4.60377365e-01
2.45502889e-02 6.53837249e-02 1.59833387e-01 9.56417680e-01
-3.71705920e-01 -8.99103940e-01 -1.17950462e-01 8.04315627e-01
-4.04329717e-01 1.07469104e-01 -4.86952305e-01 8.27179909e-01
-4.88352636e-03 7.21187294e-01 -5.42103462e-02 -1.24451689e-01
1.09387420e-01 -2.24874750e-01 2.84827054e-01 -6.60753429e-01
-5.77855766e-01 2.26991653e-01 3.78688774e-03 -4.93792653e-01
-4.26005453e-01 -5.92330217e-01 -1.66511559e+00 3.33340168e-01
-4.28926587e-01 1.03635050e-01 7.39778280e-01 9.67726409e-01
-6.97420388e-02 7.66910136e-01 4.52693671e-01 -7.02583134e-01
-3.44959825e-01 -7.25533247e-01 -6.58374667e-01 4.01811957e-01
2.81160355e-01 -2.80199587e-01 -3.75975728e-01 9.88739431e-02] | [14.979242324829102, -2.888782024383545] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.